2006-Detecting Allergens in Food - ISBN-10 0-8493-2574-9

Detecting allergens in food
Edited by
Stef J. Koppelman and Sue L. Hefle
CRC Press
Boca Raton Boston New York Washington, DC
WOODHEAD PUBLISHING LIMITED

Cambridge England
Published by Woodhead Publishing Limited, Abington Hall, Abington,
Cambridge CB1 6AH, England
www.woodheadpublishing.com
Published in North America by CRC Press LLC, 6000 Broken Sound Parkway, NW,
Suite 300, Boca Raton, FL 33487, USA
First published 2006, Woodhead Publishing Ltd and CRC Press LLC
© 2006, Woodhead Publishing Ltd
The authors have asserted their moral rights.
This book contains information obtained from authentic and highly regarded sources.
Reprinted material is quoted with permission, and sources are indicated. Reasonable
efforts have been made to publish reliable data and information, but the authors and
the publishers cannot assume responsibility for the validity of all materials. Neither
the authors nor the publishers, nor anyone else associated with this publication, shall
be liable for any loss, damage or liability directly or indirectly caused or alleged to be
caused by this book.
Neither this book nor any part may be reproduced or transmitted in any form or by
any means, electronic or mechanical, including photocopying, microfilming and
recording, or by any information storage or retrieval system, without permission in
writing from the publishers.
The consent of Woodhead Publishing Limited does not extend to copying for
general distribution, for promotion, for creating new works, or for resale. Specific
permission must be obtained in writing from Woodhead Publishing Limited for such
copying.
Trademark notice: Product or corporate names may be trademarks or registered

trademarks, and are used only for identification and explanation, without intent to
infringe.
British Library Cataloguing in Publication Data

A catalogue record for this book is available from the British Library.
Library of Congress Cataloging in Publication Data

A catalog record for this book is available from the Library of Congress.
Woodhead Publishing ISBN-13: 978-1-85573-728-0 (book)

Woodhead Publishing ISBN-10: 1-85573-728-0 (book)
Woodhead Publishing ISBN-13: 978-1-84569-055-7 (e-book)
Woodhead Publishing ISBN-10: 1-84569-055-9 (e-book)
CRC Press ISBN-10: 0-8493-2574-9
CRC Press order number: WP2574
The publishers’ policy is to use permanent paper from mills that operate a sustainable
forestry policy, and which has been manufactured from pulp which is processed
using acid-free and elementary chlorine-free practices. Furthermore, the publishers
ensure that the text paper and cover board used have met acceptable environmental
accreditation standards.
Typeset in India by Replika Press Pvt Ltd.

Printed by TJ International, Padstow, Cornwall, England
Contributor contact details
(* = main point of contact)
Chapter 1 Chapter 3
Professor Steve L. Taylor Professor Sue L. Hefle
Food Allergy Research & Resource Food Allergy Research & Resource
Program Program
University of Nebraska University of Nebraska
143 Food Industry Bldg 255 Food Industry Bldg
Lincoln, NE 68583-0919 Lincoln, NE 68583-0919
USA USA
E-mail: staylor2@unl.edu E-mail: shefle1@unl.edu
Chapter 2 Chapter 4
Professor Dr Heimo Breiteneder Dr Robert G. Hamilton
Department of Pathophysiology Johns Hopkins University School
Medical University of Vienna of Medicine
AKH-EBO-3Q Johns Hopkins DACI Reference
Waehringer Guertel 18-20 Laboratory
1090 Vienna Room 1A20-JHAAC
Austria 5501 Hopkins Bayview Circle
Baltimore, MD 21224
E-mail: USA
heimo.breiteneder@meduniwien.ac.at
E-mail: rhamilto@jhmi.edu
xii Contributor contact details
Chapter 5 Chapter 9
Dr Ronald van Ree Dr Ingrid Malmheden Yman
Department of Immunopathology Swedish National Food
Sanquin Research Administration
PO Box 9190 PO Box 622
NL-1006 AD SE-751 26 Uppsala
Amsterdam Sweden
The Netherlands
E-mail:
E-mail: r.vanree@sanquin.nl ingrid.malmheden.yman@slv.se
Chapter 6 Chapter 10
Dr Jupiter Yeung Dr Sabine Baumgartner*
Food Products Association BOKU – University of Natural
Suite 300 Resources and Applied Life
13501 Street NW Sciences, Vienna
Washington, DC 20005 C/O Department for
USA Agrobiotechnology, IFA – Tulln
Centre for Analytical Chemistry
E-mail: jyeung@fpa-food.org Konrad-Lorenz-Strasse 20
A-3430 Tulln
Austria
Chapter 7
Dr Stefan Vieths E-mail:
Paul-Ehrlich-Institut sabine.baumgartner@boku.ac.at
Paul-Ehrlich-Strasse 51-59
D-63225 Langen Dr Rob van Herwijnen
Germany
E-mail: info@evlonline.nl (att. Dr
E-mail: viest@pei.de Rob van Herwijnen)
Professor Michael G. Zeece Professor Sue L. Hefle
Department of Food Science & Food Allergy Research & Resource
Technology Program
University of Nebraska University of Nebraska
354 Food Industry Bldg 255 Food Industry Bldg
Lincoln, NE 68583-0919 Lincoln, NE 68583-0919
USA USA
E-mail: mzeece@unlnotes.unl.edu E-mail: shefle1@unl.edu

Contributor contact details xiii
Dr Stef Koppelman Dr Stef Koppelman*
Department of Dermatology/ Department of Dermatology/
Allergology Allergology
University Medical Centre Utrecht University Medical Centre Utrecht
NL-3508 GA Utrecht NL-3508 GA Utrecht
The Netherlands The Netherlands
E-mail: stefkoppelman@zonnet.nl E-mail: stefkoppelman@zonnet.nl
Professor Sue L. Hefle

Chapter 13 Food Allergy Research & Resource
Dr Claude Demeulemester* and Program
Dr Ines Giovannacci University of Nebraska
CTSCCV 255 Food Industry Bldg
7 avenue du General de Gaulle Lincoln, NE 68583-0919
F-94704 Maisons-Alfort USA
France
E-mail: shefle1@unl.edu
E-mail: cdemeulemester@ctsccv.fr
Dr Virginie Leduc Chapter 16

Allerbio Dr Chris Cordle
67 rue de Dunkergue Ross Products Division
75009 Paris Abbott Laboratories
3300 Stelzer Road
E-mail: v.leduc@allerbio.fr Columbus OH 43219-3034
USA
Chapter 14 E-mail: chris.cordle@abbott.com

Frederik W. Janssen, M.Sc.
Fascoda FAS Food Consultancy
Coehoornsingel 44 Chapter 17
NL-7201 AD Zutphen Dr René Crevel
The Netherlands Safety & Environmental Assurance
Centre
E-mail: fwjanssen@chello.nl Unilever Research and Development
Colworth
Bedford MK44 ILQ
UK
E-mail: rene.crevel@unilever.com
xiv Contributor contact details
Dr Ulrike Immer Heereluurt Heeres LL.M., B.Sc.
R-Biopharm AG TNO Quality of Life
Landwehrstrasse 54 PO Box 360
D-64293 Darmstadt NL-3700 AJ Zeist
Germany The Netherlands
E-mail: u.immer@r-biopharm.de E-mail: heereluurt.heeres@tno.nl
Dr Roland E. Poms* Dr Stef Koppelman*
International Association for Cereal Department of Dermatology/
Science and Technology (ICC) Allergology
Marxergasse 2 University Medical Centre Utrecht
A-1030 Vienna NL-3508 GA Utrecht
Austria The Netherlands
E-mail: roland.poms@icc.or.at E-mail: stefkoppelman@zonnet.nl
Dr Hendrik Emons and Dr Elke Professor Sue L. Hefle

Anklam Food Allergy Research & Resource
European Commission Joint Program
Research Centre University of Nebraska
Institute for Reference Materials 255 Food Industry Bldg
and Measurements (IRMM) Lincoln, NE 68583-0919
Retieseweg 111 USA
B-2440 Geel
Belgium E-mail: shefle1@unl.edu
Chapter 20
Mr Martin J. Hahn
Hogan & Hartson LLP
555 13th Street NW
Washington, DC 20004
USA
E-mail: MJHahn@HHLaw.com
Preface
With food allergies on the increase in Western societies, and new legislation
on allergen labeling in both Europe and the USA, the food industry is confronted
with the issue of providing safe foods for the food-allergic consumer. While
most food-allergic reactions occur after ingestion of non-packaged food
products, the food industry has been subjected to increasing scrutiny of its
allergen controls; the resulting impact on the industry has been remarkable.
In the past 15 years, the food industry has made significant investment,
effort, and improvements in allergen control. In the past eight years, tests for
some allergenic foods have been commercialized and have proven useful to
the industry in controlling allergens, and also to regulatory agencies
investigating food-allergic consumer complaints. There are many strategies
food manufacturers can exercise in controlling allergens in their plants, from
changing raw materials to improving cleaning procedures and using
precautionary labeling indicating allergens that might be present. However,
measuring the content of particular allergenic residues on processing equipment,
in ingredients or in final products, provides information that can also be used
for risk assessment, enabling the food industry to provide the food-allergic
consumer with practical information. The inspiration for this book originated
in the early years of this millennium when increasing numbers of commercial
test kits for measuring allergenic food residues were launched. The purpose
of the book is to serve as a resource for experts from different backgrounds
such as biochemistry, food chemistry, food legislation, and allergology with
regard to the technical possibilities and limitations of food allergen detection
methods.
The first part of the book deals with the nature of food allergy and introduces
the reader to the range of allergenic foods. What causes an individual to be
at risk from food allergy? What makes a protein an allergen? How much
allergenic residue is too much for a food-allergic individual? What is the
nature of the immune response to food that causes allergic reactions? These
questions will be discussed as a basis for the remainder of the book.
Part II deals with methods designed for measuring food allergens, based
xvi Preface
on either immunochemical or DNA techniques. How specific and sensitive

does a method need to be, and does one need quantitative ability, or just
qualitative? Should a method detect the allergen itself or marker/indicator
molecules? What is a technically feasible detection limit, and what is a
practically useful detection limit? When do we need high throughput screening,
and how much time will analysis take? And what will allergenic residue
detection methods look like in the future?
Part III deals with published detection methods and those that are
commercially available at the time this book went to print. Methods for the
main food allergens such as peanut, milk, egg, tree nuts and seeds, fish and
crustacean shellfish, soy and cereals are discussed by experts in those particular
areas. Detection limits are discussed in view of clinical food allergen threshold
data as well as case reports of allergic individuals who experienced reactions
after inadvertent ingestion of a certain amount of allergen.
Part IV of the book deals with the challenges and issues facing the food
industry in the struggle to deal with allergenic food residues in production
environments. Methods for allergen detection are available, with certain
specificities and sensitivities, and practical limitations. New ingredient-labeling
legislation is pending in many countries, and the outcome of these efforts
may in the future give rise to even more specific legislation on inadvertent
allergen contamination of food products. The food industry often uses shared
equipment and facilities for the production of processed foods with a variety
of formulations. This is often out of necessity rather than just fueled by
economic reasons. What are the tools and information available to the industry
in risk assessment, and where do detection methods fit into these discussions?
In the concluding chapter, we will summarize the problems involved in
the detection of allergenic residues in food, and will provide an outlook for
the future of allergen detection as we envision it.
We wish to express our gratitude to all of the contributors to this book for
their excellent work and for their cooperation and patience during the reviewing
process. We are proud of the result, and we hope many colleagues will enjoy
using this book.
Stef Koppelman and Sue Hefle

Contents
Contributor Contact Details .................................................... xi
Preface ................................................................................... xv
Part I. The Basics of Food Allergy

1. The Nature of Food Allergy ................................................... 3
1.1 Introduction: Defining Food Allergy ....................... 3
1.2 Mechanisms of Food Allergy ................................. 7
1.3 Avoidance Diets and Treatment for
IgE-mediated Food Allergies ................................. 10
1.4 Future trends ........................................................ 15
1.5 Sources of Further Information and Advice ........... 16
1.6 References ........................................................... 17
2. Classifying Food Allergens ................................................... 21
2.1 Introduction ........................................................... 21
2.2 Plant Food Allergens ............................................. 23
2.3 Animal Derived Food Allergens ............................. 37
2.4 Future Trends ....................................................... 43
2.6 Acknowledgement ................................................. 45
2.7 References ........................................................... 45
Part II. Types of Detection Method

3. Antibodies .............................................................................. 65
3.1 Nature of Antibodies ............................................. 65
This page has been reformatted by Knovel to provide easier navigation. v

vi Contents
3.2 Immunogens and Antigens ................................... 68

3.3 Antibody Production .............................................. 68
3.4 Choice of Producing Monoclonal or Polyclonal
Antibodies ............................................................. 75
3.5 References ........................................................... 77
4. Allergen-specific Human IgE Antibody-based Analysis
of Food .................................................................................. 79
4.1 Introduction ........................................................... 79
4.2 IgE Antibody-based in Vivo Assay ........................ 80
4.3 IgE Antibody-based in Vitro Assay for Food
Allergen Detection ................................................. 85
4.4 Applications .......................................................... 91
4.5 RAST Inhibition Assay Strengths and
Weaknesses ......................................................... 94
4.6 Future Trends ....................................................... 95
4.7 References ........................................................... 96
5. Immunoblotting in Allergen Detection ................................... 98
5.1 Introduction ........................................................... 98
5.2 Mono-specific Antibody Reagents ......................... 100
5.3 Critical Assessment of (Mono-)specificity .............. 102
5.4 Food Processing and Antibody Specificity ............ 104
5.5 Future Trends ....................................................... 104
5.6 References ........................................................... 105
6. Enzyme-linked Immunosorbent Assays (ELISAs) for
Detecting Allergens in Foods ................................................ 109
6.1 Introduction ........................................................... 109
6.2 Principles of ELISAs ............................................. 110
6.3 Applications .......................................................... 115
6.4 Future trends ........................................................ 120
6.5 Conclusions .......................................................... 121
6.6 Acknowledgements ............................................... 121
6.7 References ........................................................... 121
This page has been reformatted by Knovel to provide easier navigation.

Contents vii
7. Polymerase Chain Reaction (PCR) Methods for the

Detection of Allergenic Foods ............................................... 125
7.1 Introduction ........................................................... 125
7.2 PCR Principles ...................................................... 125
7.3 Application of PCR for the Detection of
Allergenic Foods ................................................... 133
7.4 Advantages and Disadvantages of PCR
Compared to ELISA .............................................. 138
7.5 Future Trends ....................................................... 140
7.6 References ........................................................... 141
8. Proteomic Assessment of Allergens in Food ........................ 144
8.1 Introduction ........................................................... 144
8.2 Key Issues in Proteomic Assessment of
Allergens ............................................................... 145
8.3 Applications of Proteomics for Detection of
Allergens ............................................................... 149
8.4 Future Trends ....................................................... 155
8.5 Conclusions .......................................................... 155
8.6 References ........................................................... 156
9. Detecting Food Allergens with a Surface Plasmon
Resonance Immunoassay .................................................... 158
9.1 Introduction ........................................................... 158
9.2 Biosensors and SPR Technology .......................... 159
9.3 Developing a Food Allergen SPR
Immunoassay ....................................................... 160
9.4 Published Methods ............................................... 164
9.5 Experimental Data ................................................ 165
9.6 Conclusions .......................................................... 172
9.8 References ........................................................... 173
10. The Use of Lateral Flow Devices to Detect Food
Allergens ................................................................................ 175
10.1 Introduction ........................................................... 175

viii Contents
10.2 Antibodies ............................................................. 176

10.3 Constructing a Lateral Flow Device (LFD) ............. 177
10.4 Running a Sample ................................................ 178
10.5 Methods Available ................................................. 180
10.6 Future Trends ....................................................... 180
10.7 References ........................................................... 181
Part III. Detection Methods for Particular Allergens

11. Methods for Detecting Peanuts in Food ............................... 185
11.1 Introduction: Peanut Allergy .................................. 185
11.2 Allergenic Peanut Proteins .................................... 186
11.3 Peanut Detection Methods .................................... 186
11.4 Appropriate Detection Limits for Peanut
Methods ................................................................ 195
11.5 Future Trends ....................................................... 195
11.6 References ........................................................... 197
12. Detecting Tree Nuts and Seeds in Food .............................. 201
12.1 Introduction ........................................................... 201
12.2 Prevalence of Nut and Seed Allergies ................... 202
12.3 Thresholds ............................................................ 203
12.4 Allergenic Proteins in Nuts and Seeds .................. 204
12.5 Effect of Food Processing on Allergenicity ............ 204
12.6 Detecting Nut and Seed Residues in Food:
Selecting a Method ............................................... 205
12.7 Conclusions .......................................................... 212
12.8 References ........................................................... 214
13. Detecting Dairy and Egg Residues in Food ......................... 219
13.1 Introduction ........................................................... 219
13.2 Milk ....................................................................... 220
13.3 Egg ....................................................................... 225
13.4 Types of Detection Methods ................................. 228
13.5 Future Trends ....................................................... 235

Contents ix
13.6 Acknowledgements ............................................... 236

13.7 References ........................................................... 236
14. Detecting Wheat Gluten in Food ........................................... 244
14.1 Introduction ........................................................... 244
14.2 Key Requirements for Detection and
Quantization .......................................................... 249
14.3 Types of Detection Methods ................................. 249
14.4 Non-antibody-based Techniques .......................... 262
14.5 Selecting a Method ............................................... 264
14.6 Future trends ........................................................ 265
14.8 References ........................................................... 268
15. Detecting Soy, Fish and Crustaceans in Food ..................... 273
15.1 Introduction ........................................................... 273
15.2 Soy ....................................................................... 273
15.3 Crustaceans .......................................................... 281
15.4 Fish ....................................................................... 283
15.5 Future trends ........................................................ 285
15.6 References ........................................................... 286
Part IV. Issues in Allergen Detection Methods

16. Allergen Quality Assurance for Hypoallergenic
Formula ................................................................................. 293
16.1 Introduction ........................................................... 293
16.2 Key Terms and Clinical and Analytical
Performance Targets ............................................ 294
16.3 Analytical Methods ................................................ 297
16.4 Applications .......................................................... 306
16.5 Summary and Future Trends ................................ 311
16.6 References ........................................................... 313
17. Common Issues in Detecting Allergenic Residues on
Equipment and in Processed Foods ..................................... 315
17.1 Introduction ........................................................... 315

x Contents
17.2 Food Allergy and Product Safety ........................... 316

17.3 Management of Food Allergy Risks ...................... 318
17.4 Role of Allergen Detection and Other
Considerations ...................................................... 321
17.5 Future Trends ....................................................... 326
17.6 References ........................................................... 328
18. Factors Affecting the Effectiveness of Allergen
Detection ............................................................................... 330
18.1 Introduction ........................................................... 330
18.2 Factors Affecting the Determination of
Allergenic Residues .............................................. 331
18.3 Troubleshooting .................................................... 340
18.4 Future Trends ....................................................... 344
18.5 Summary .............................................................. 346
18.6 References ........................................................... 346
19. Reference Materials and Method Validation in Allergen
Detection ............................................................................... 348
19.1 Introduction ........................................................... 348
19.2 Quality Assurance for the Analysis of
Allergens ............................................................... 349
19.3 Towards Validated Methods for Allergen
Determination ....................................................... 350
19.4 Characteristics and Use of Reference
Materials ............................................................... 351
19.5 Towards Reference Materials for Allergens ........... 353
19.6 Future Trends ....................................................... 354
19.8 References ........................................................... 355
20. US Regulation of Undeclared Allergens in Food
Products ................................................................................ 357
20.1 Introduction ........................................................... 357
20.2 Regulatory Liability ............................................... 358

Contents xi
20.3 Legal Grounds for Product Liability

Actions .................................................................. 364
20.4 Future Trends ....................................................... 372
20.5 Conclusion ............................................................ 373
20.7 References ........................................................... 373
Appendix I: Summary of Cases Involving Known
Allergens ............................................................... 375
Appendix II: Summary of Cases Involving
Ingredients That Are Not Allergens ....................... 377
21. EU Regulation of Undeclared Allergens in Food
Products ................................................................................ 378
21.1 Introduction ........................................................... 378
21.2 Food Legislation Concerning the Labelling of
Ingredients ............................................................ 379
21.3 Food Legislation Concerning the Labelling of
Allergens ............................................................... 382
21.4 Legislation Concerning General Product
Safety (Directive 2001/95/EC) ............................... 386
21.5 Legislation Concerning Food Safety
(Regulation (EC) No. 2002/178/EC) ...................... 388
21.6 Legislation Concerning Product Liability ................ 390
21.7 Key Issues in Labelling of Allergens,
Undeclared Allergens, Food Safety and
Product Liability .................................................... 395
21.8 Future Trends ....................................................... 401
21.10 References ........................................................... 403
22. Conclusions ........................................................................... 405
22.1 Recent Literature and Trends ............................... 405
22.2 Relating Detection Limits to Clinically
Relevant Doses .................................................... 407

xii Contents
22.3 Reference Materials, Extraction and

Recovery .............................................................. 409
22.4 Developing Realistic and Practical Detection
Methods ................................................................ 409
22.5 Summary .............................................................. 410
22.6 References ........................................................... 412
Index ...................................................................................... 413

Part I
The basics of food allergy

1
The nature of food allergy

S. Taylor, University of Nebraska, USA
1.1 Introduction: defining food allergy

While eating is necessary to sustain life, most people find eating to be an
enjoyable experience given the variety and abundance of food available to
them. However, for individuals with food allergies and related illnesses,
consuming certain foods can be a debilitating, and possibly even life-
threatening, experience. Consequently, the joy of eating is diminished by the
ever-present fear of consuming a food or food ingredient that will cause an
adverse reaction. Food-allergic consumers must assiduously avoid the offending
foods and/or food ingredients because strict avoidance diets are the only
available preventive strategy. For such consumers, food selection often becomes
a tedious task requiring meticulous reading of ingredient lists on labels,
dependence on food manufacturers to maintain accurate labels, and a continual
search for more knowledge about food composition. For these individuals,
food preparation requires careful attention to detail, cooking ‘from scratch’,
and seeking alternative recipes for many dishes. Because very small amounts
of the offending food can elicit allergic reactions in some affected individuals,
these consumers live in constant fear that, despite their caution, trace amounts
of the offending food, sufficient to elicit an adverse reaction, might still exist
in the foods that they consume. They are concerned about ingredients derived
from the offending food because such ingredients might contain residual
allergenic proteins from the source food. This fear is compounded by the fact
that declaration of the source of ingredients used in foods is not always
required on food labels.
4 Detecting allergens in food
1.1.1 Definition and classification

Food allergies can be defined as adverse, immune-mediated reactions to
foods that occur in certain individuals. Often, the public and even some
within the medical community categorize all individualistic reactions to foods
as food allergies. However, true food allergies should be restricted to those
individualistic reactions to foods that are mediated by the immune system.
The term ‘food sensitivity’ can be used to refer to all types of individualistic
adverse reactions to foods. These food-related illnesses are individualistic
because they affect only a few people in the population. Food intolerances
are individualistic adverse reactions to foods that occur through non-
immunological mechanisms. Knowing the difference between immunological
food allergies and non-immunological food intolerances is critical to proper
management of these illnesses. Food intolerances are often controlled by
limiting the amount of food eaten; with food allergies, much more strict
avoidance of the offending food is usually necessary. Table 1.1 provides a
classification scheme for the various illnesses that are known to occur as
individualistic adverse reactions to foods.
Food allergy is an abnormal immunological response to a food or food
component; food allergens are almost always proteins.1 Examples include
allergic reactions to common foods such as peanuts and milk. Within this
category are immediate hypersensitivity reactions where symptoms ensue
within minutes to an hour after ingestion of the offending food and delayed
hypersensitivity reactions where the onset of symptoms occurs 6–24 or more
hours after ingestion of the offending food.
Immediate hypersensitivity reactions are mediated by immunoglobulin E
(IgE) antibodies. Exercise-induced food allergies are a subset of food allergies
involving immediate reactions that occur only when the specific food is
ingested just before or after exercise,2 although many cases of exercise-
induced allergies are not associated with foods.3
Delayed hypersensitivity reactions are cell-mediated, normally involving
sensitized immune cells in the small intestine, usually lymphocytes, that are
sensitized to the specific substance that triggers the reaction.4 The ultimate
result is tissue inflammation often restricted to certain sites in the body with
Table 1.1 A classification scheme for food allergies and intolerances
Food allergy IgE-mediated food allergy (immediate hypersensitivity)

e.g. peanut allergy, cows’ milk allergy
Cell-mediated food allergy (delayed hypersensitivity)
e.g. celiac disease
Food intolerance Anaphylactoid reactions
e.g. strawberry-induced urticaria (unproven)
Metabolic food disorders
e.g. lactose intolerance
Idiosyncratic reactions
e.g. sulfite-induced asthma
The nature of food allergy 5
symptoms appearing on a more delayed basis, as much as 24 or more hours

after consumption of the offending food.
Food intolerances, in contrast to true food allergies, do not involve abnormal
responses of the immune system.5 Anaphylactoid reactions involve the release
of the chemical mediators (mostly histamine) of allergic reactions into the
body without the intervention of IgE antibodies. Foods such as strawberries
and chocolate are thought to allegedly induce such reactions, but definitive
proof for this type of food intolerance does not exist. Metabolic food disorders
are genetically determined metabolic deficiencies that result in adverse reactions
to a food component. Lactose intolerance serves as a good example of a
metabolic food disorder. In lactose intolerance, the affected individual has a
deficiency of the intestinal enzyme, β-galactosidase, which is essential for
the metabolism of the lactose in milk.6 Consequently, lactose cannot be
absorbed from the intestinal lumen leading to bacterial fermentation of the
lactose in the colon with resultant flatulence and frothy diarrhea. Food
idiosyncrasies are adverse reactions to foods or a food component that occur
through unknown mechanisms. Examples include sulfite-induced asthma
and tartrazine-induced asthma.7 In many cases, the cause-and-effect relationship
between the food or food component and the particular adverse reaction
remains unproven; this would be the situation with tartrazine-induced asthma.
Psychosomatic illnesses are included in this category.
Allergy-like intoxications are worth some mention here because these
illnesses can be confused diagnostically with food allergies. Unlike food
allergies, everyone in the population is probably susceptible. This reaction
occurs as a result of the ingestion of histamine, one of the primary mediators
of allergic disease. Histamine is released from cells within the body in true
food allergies and anaphylactoid reactions but is ingested in the case of
allergy-like intoxications. Histamine poisoning (also known as scombroid
fish poisoning) is commonly associated with the ingestion of spoiled tuna,
mackerel, mahi-mahi, and other fish and also occasionally with cheese.8 The
symptoms mimic some of the most common symptoms encountered in true
food allergies.
1.1.2 Prevalence and health impacts

The overall and worldwide prevalence of IgE-mediated food allergies is not
precisely known. In the US, the overall prevalence of IgE-mediated food
allergies would be estimated at 3–4% of the population. The prevalence is
highest in infants and young children ranging from 4–8% in that age group,9
while that in adults is probably about 2–3%. Recently, the estimate of the
prevalence of IgE-mediated food allergies in adults in the US has been
increased, based upon a survey that indicates that 1.9% of the total population
has shrimp/crustacean shellfish allergy and 0.4% of the total population has
fish allergy;10 both fish and crustacean shellfish allergies affect many adults.
These percentages are added to earlier estimates that peanut and tree nut
Table 1.2 Symptoms of IgE-mediated food allergies
Gastrointestinal Nausea, vomiting, diarrhea, abdominal cramps

Respiratory Asthma, wheezing, rhinitis
Cutaneous Urticaria (hives), eczema or atopic dermatitis, pruritis, rash,
angioedema
Other Anaphylactic shock, hypotension, swelling of the tongue, laryngeal
edema, oral allergy syndrome
allergies affect an estimated 1.1% of the US population, including many

adults.11 Cows’ milk, eggs, and peanuts are among the most common allergenic
foods for infants and young children,12 while crustacean shellfish, fish, and
peanuts are among the most common allergenic foods for adults in the US,10,11
although the situation can vary in other countries depending upon dietary
eating patterns.
IgE-mediated allergic reactions involve numerous symptoms ranging from
mild to life-threatening (Table 1.2).13 Food-allergic individuals experience
quite varied symptoms and it is likely that no one suffers from all of the
symptoms mentioned in Table 1.2. The nature and severity of the symptoms
may also vary from one occasion to another in the same individual as a result
of the amount of the offending food ingested and the length of time since the
last previous exposure. Anaphylactic shock is the most severe symptom for
IgE-mediated individuals. Fortunately, comparatively few food-allergic
individuals are susceptible to such severe reactions. Systemic anaphylaxis
involves many organ systems and numerous symptoms including those also
noted in Table 1.2 together with cyanosis, chest pain, and shock. Anaphylactic
shock and asthma are the most common causes of death in the occasional
fatalities associated with true food allergies.14,15
Although severe, life-threatening reactions are definitely the most worrisome
manifestations of IgE-mediated food allergies, mild symptoms are much
more likely to occur. One of the more common and typically most mild
forms of IgE-mediated food allergy is the so-called oral allergy syndrome
(OAS).16 In this condition, symptoms are usually confined to the oropharyngeal
area of the mouth and throat and include pruritis, urticaria, and angioedema.
OAS is most frequently associated with the ingestion of various fresh fruits
and vegetables.16 Even though fresh fruits and vegetables typically contain
low amounts of proteins, OAS is an IgE-mediated reaction to specific
proteinaceous allergens present in these foods.16 These fruit and vegetable
allergens are apparently quite susceptible to digestive proteases in the
gastrointestinal tract.17 Systemic reactions are not typically encountered in
OAS, but can be with certain patients and certain foods, e.g. celery. These
allergens are also apparently heat-labile,17 since the heat-processed versions
of these foods are not typically involved in initiation of OAS. With OAS,
affected individuals are initially sensitized to pollens in the environment,
such as birch and mugwort pollens, that cross-react with related proteins
found in the fresh fruits and vegetables.18 Apparently, sensitization to these

pollens can increase the likelihood of sensitization to specific foods.
Occasionally, food allergies occur in conjunction with exercise.2 The
prevalence of exercise-associated food allergies is unknown. In these cases,
the exercise must be preceded or followed by the ingestion of specific foods
in order to elicit an allergic reaction. Shellfish,19 wheat,20 and celery2 are
among the foods that have been incriminated in food-dependent, exercise-
associated anaphylaxis. The symptoms in this type of food allergy are
individualistic and similar to those involved in other food allergies. With
increased awareness of the existence of this syndrome and the emphasis on
physical fitness in many countries, reports of this condition may continue to
increase.
1.2 Mechanisms of food allergy

Two types of allergic or hypersensitivity reactions occur as basic immunological
mechanisms involved in food allergies.5 IgE-mediated reactions, also known
as immediate hypersensitivity reactions, involve the formation of IgE antibodies
that specifically recognize certain allergens in foods. IgE-mediated reactions
are the most important type of food allergy because these reactions involve
a wide variety of different foods and the reactions can be severe in some
individuals. IgE-mediated mechanisms are also responsible for allergic reactions
to pollens, mold spores, animal danders, insect venoms, and certain drugs;
only the source of the allergen differs. Cell-mediated reactions or delayed
hypersensitivities probably play an important, although as yet undefined,
role in food hypersensitivity.5 Celiac disease, which will be discussed later,
may be a form of cell-mediated delayed hypersensitivity.21
1.2.1 IgE-mediated allergic reaction (immediate hypersensitivity)

Although Hippocrates was the first to document the occurrence of food
allergies,22 the involvement of the immune system in food allergies was not
recognized for many years thereafter. Prausnitz and Kustner23 were the first
clinicians to recognize that the blood contained some substance that conferred
allergic hypersensitivity. They subcutaneously injected a non-allergic individual
with a fish extract and noted no adverse reaction. However, when this same
normal individual was first inoculated, under the skin, with serum from a
fish-allergic person and then subsequently injected with the fish extract,
there was an inflammatory skin reaction at the serum-injected site. This
passive sensitization experiment provided the first evidence that the blood
contained some substance that sensitized the allergic individual to the fish.
For some years thereafter, this blood factor remained unidentified and was
referred to as reaginic factor. The reaginic factor involved in allergic reactions
was first recognized as an antibody in 1966, when Ishizaka et al.24,25
demonstrated that this reaginic activity was associated with a unique

immunoglobulin and tentatively called this protein γE. The protein was officially
named immunoglobulin E or IgE by the World Health Organization in 1968.
The identification of IgE as a reaginic antibody provided immunochemical
approaches to analyze the mechanisms involved in hypersensitivity reactions.26
Immunoglobulin E is one of five classes of antibody that are present in the
human immune system (the others being IgG, IgM, IgA, and IgD).
The normal function of IgE antibodies is protection from parasitic infections.
Although all humans have low levels of IgE antibodies, individuals predisposed
to the development of allergies are most likely to produce IgE antibodies that
are specific for and recognize certain environmental antigens.5 These antigens
are typically proteins, although only a few of the many proteins in nature are
capable of stimulating the production of specific IgE antibodies in susceptible
individuals.27 The allergens eliciting IgE antibody formation can be found in
pollens, mold spores, venoms, dust mites, and animal danders in addition
to foods.28
The mechanism involved in IgE-mediated reactions is now well understood
and is depicted in Figure 1.1.5 In IgE-mediated food allergies, allergen-
specific antibodies are first produced in response to stimulus of the antibody-
forming B cells in response to the immunological stimulus created by exposure
of the immune system to a specific food allergen, usually a naturally-occurring
protein present in the food. The immune response in the small intestine
which is responsible for the dominance of the IgE antibody generation is
quite complex and involves T helper type 2 cells, interleukin-4 (IL-4), and
other factors.29 The IgE antibodies bind to the surfaces of mast cells in the
tissues or basophils in the blood in a process known as sensitization. The
sensitization phase of the allergic reaction is symptomless. In fact, sensitization
can occur without the development of clinical reactivity29 so the demonstration
of IgE antibodies directed against a particular food in human blood serum is
Stimulates
Allergen production
of
IgE Sensitized
Mast cell
basophil cell
Sensitized cell
Allergen
Degranulation
Release:
histamine and
other mediators
Fig. 1.1 Mechanism of IgE-mediated allergic reaction.

insufficient evidence for the diagnosis of a food allergy unless it is coupled

to a strong history of food allergy or a positive double-blind, placebo-controlled
food challenge.
Once sensitized, exposure to the same food allergen on a subsequent
occasion can result in an allergic reaction.5 When this happens, the allergen
associates with the mast cell- or basophil-bound IgE, and cross-links two of
the IgE molecules. A series of biochemical events is initiated which causes
cell membrane disruption and the release of a variety of mediators contained
within granules existing in the mast cells and basophils. The granules in mast
cells and basophils contain most of the important mediators of the allergic
reaction.30 While several dozen substances have been identified as chemical
mediators emanating from mast cell and basophils, histamine is responsible
for most of the immediate effects of allergic reactions. The histamine-related
effects include inflammation, pruritis, and contraction of the smooth muscles
in the blood vessels, gastrointestinal tract, and respiratory tract.31 Other
important mediators include a variety of prostaglandins and leukotrienes;
these particular mediators are associated with some of the slower-developing
responses observed in some cases of food allergy (e.g. late-phase asthmatic
reactions).22
While those individuals prone to the development of food allergies may
form specific IgE antibodies upon dietary exposure to a particular food,
other individuals will not. Even among those individuals predisposed to
allergies, exposure to food proteins does not usually result in formation of
allergen-specific IgE. Food-allergic individuals will normally be sensitive to
only a few of the wide variety of foods in the typical human diet. In normal
individuals, and even in those who are susceptible to the development of
food allergies, exposure to a food protein most often results in oral tolerance
through induction of T-cell anergy, deletion of reactive T cells, the generation
of suppressor T cells, the formation of protective secretory IgA antibodies,
and other immunological responses.29,32 Heredity and other physiological
factors are significant in predisposing individuals to the development of IgE-
mediated food allergies and also other environmental allergies.33 Approximately
65% of patients with clinically documented allergy have first degree relatives
with allergic disease.33 Conditions that increase the permeability of the intestine
to macromolecules, such as viral gastroenteritis, premature birth, and cystic
fibrosis, may increase the risk of development of food allergy.22 Although
food allergies may also involve other types of immunological mechanisms,
the IgE-mediated mechanism is, by far, the most well-documented and
understood.
1.2.2 Exercise-associated food allergies

The mechanism for food-dependent, exercise-induced anaphylaxis is unknown,
but enhanced mast cell responsiveness to physical stimuli may be involved.3
IgE sensitization appears to be involved in many cases.34
1.2.3 Cell-mediated reactions (delayed hypersensitivity)

In contrast to IgE-mediated reactions, the symptoms of cell-mediated allergic
reactions do not begin to appear until 6–24 hours after ingestion of the
offending food. 5 These reactions develop slowly, reaching a peak at
approximately 48 hours and subsiding after 72–96 hours. The mechanisms
of cell-mediated food allergies are not nearly as well understood. They involve
an interaction between specific food allergens and sensitized T lymphocytes.
Lymphocyte stimulation initiates the release of cytokines and lymphokines
which produces a localized inflammatory response.35 Antibodies are not
involved in these reactions.
T lymphocytes are a major component of the gut-associated lymphoid
tissue.36 Except for celiac disease, evidence for the involvement of cell-
mediated immune reactions in food allergies remains incomplete. However,
cell-mediated reactions appear to be involved in some cases of cows’ milk
allergy occurring especially in infants and with symptoms confined primarily
to the gastrointestinal tract.37 No estimates of the prevalance of cell-mediated
food allergies have been made.
Celiac disease, also known as celiac sprue or gluten-sensitive enteropathy,
appears to be an example of a cell-mediated reaction.21 Celiac disease is a
malabsorption syndrome occurring in sensitive individuals upon the
consumption of wheat, rye, barley, triticale, spelt, and kamut.5,38 The role of
oats in celiac disease is much less well-defined. Apparently, oats that are
totally free of wheat, rye, and barley are safe for celiac sufferers to consume.39
The consumption of wheat or other offending grains or products made from
these grains elicits inflammatory damage to the absorptive epithelial cells in
the small intestine.38 The absorptive function of the small intestine is
compromised as a result. Nutrients are not properly absorbed and water can
leak out. The loss of absorptive function along with the ongoing inflammatory
process results in a severe malabsorption syndrome characterized by diarrhea,
bloating, weight loss, anemia, bone pain, chronic fatigue, weakness, muscle
cramps, and, in children, failure to gain weight and growth retardation.13,40
The inflammatory mechanism involved in celiac disease is mediated by
intestinal T lymphocytes.21 The gliadin fraction of wheat and related fractions
in barley and rye are associated with initiation of celiac disease in susceptible
individuals.40
1.3 Avoidance diets and treatment for IgE-mediated

food allergies
Typically, the prevention of IgE-mediated food allergies primarily involves
avoiding the offending foods.41 Certain drugs, such as epinephrine (adrenaline)
and various anti-histamines, can be used to treat the symptoms that develop
during an allergic reaction if inadvertent ingestion of the offending food
occurs.42 However, no pharmaceutical approaches are presently available for

the prevention of food allergies. Anti-IgE therapy may hold some promise as
initial trials have resulted in an increase in tolerance to peanuts among peanut-
allergic individuals.43
With the lack of pharmaceutical approaches, specific avoidance diets remain
the preventive strategy of choice for most food-allergic individuals. Thus,
food-allergic individuals are forced to become avid label readers in an attempt
to avoid their offending foods and certain ingredients derived from these
foods. Their efforts are fraught with difficulty because individuals with IgE-
mediated food allergies can react to mere traces of the offending food in their
diet.44,45 The construction of safe and effective avoidance diets and the
difficulties faced by consumers who must adhere to such diets have been
extensively reviewed elsewhere.46,47
Consumers with IgE-mediated food allergies often face several challenging
questions as they attempt to implement a safe and effective avoidance diet.
(1) Will trace levels of the food elicit reactions or increase sensitization?
(2) Do all foods and food ingredients made from the offending food contain
the allergens?
(3) Are cross-reactions likely to occur between closely related species?
Clear answers to these questions are not always available, in part because of
a lack of research data and in part because of the lack of reliable analytical
information. For example, many ingredients are derived from commonly
allergenic foods; soybeans alone are used to generate hydrolyzed soy protein,
soy sauce, soy oil, soy lecithin, soy isoflavones, tocopherol (vitamin E), soy
phytosterols, soy fiber, and others. As noted later in Chapter 15, not all of
these ingredients will be expected to contain the responsible allergens but
clinical data on safety are often lacking. Also, from an analytical perspective,
detection of proteins or allergens can be particularly challenging in some of
these ingredient matrices. For example, the detection of protein residues in
oils can be difficult. Until recent years, reliable methods for the detection of
residues of allergenic foods in other foods or in such ingredients did not
exist. Even now, the ability of existing methods to assess the allergenic
potential of ingredients containing hydrolyzed or partially hydrolyzed proteins
is problematic.
1.3.1 Threshold doses (minimal elicting doses)

Practical experiences indicate very clearly that trace levels of the offending
food can elicit adverse reactions. Anecdotal reports exist of reactions from
touching utensils or bottles contaminated with the offending food, kissing
the lips of someone who has recently eaten the offending food, opening
packages of the offending food, the inhalation of vapors from cooking of the
offending food, and the transfer of food allergens from lactating mothers to
breast-feeding infants.48 Although the amount of the offending allergen involved
in eliciting such reactions must be rather low, analytical information on
minimal eliciting doses from such experiences is not available. However, a

few episodes of reactions to residues of allergenic foods hidden in other
foods have been well investigated and lend credibility to the anecdotal
reports.44,45 For all practical purposes, complete avoidance must be maintained.
Analytical evidence can be used, in part, to obtain clues to the minimal
eliciting doses for various allergenic foods. However, great care must be
taken in the use of this information to establish such doses. First of all,
existing evidence suggests that food-allergic individuals may vary in their
minimal eliciting doses by several orders of magnitude.49,50 This fact
complicates the investigation of case reports because the affected individual
will not necessarily be representative of other individuals in the population
allergic to the same food. A second uncertainty factor surrounds whether or
not a representative food sample can be obtained for analysis. On many
occasions, several foods are consumed, and the affected consumer may not
necessarily incriminate the food that actually caused the reaction. If the
analyst finds a very low level of allergenic residue in an incriminated food,
it is very important to determine if this food was the only one consumed at
the implicated meal because other foods might have even higher levels of the
particular allergen. Additionally, the consumer may have consumed all of the
incriminated food sample or discarded the remainder. In those situations, the
analyst can attempt to obtain similar samples from the marketplace, but the
analysis of these samples creates uncertainties due to the questions about
whether this sample is representative of the actual product that may have
caused the illness. Analysis of samples from actual episodes is always
complicated if the allergen contamination may not have been uniform. The
bite that the consumer eats may have higher levels of allergen contamination
than the remaining portion in some cases. Finally, until recently, reliable
analytical methods for the detection of residues of allergenic foods were not
readily available. Even today, it is uncertain whether analysis of the same
sample by several different methods would yield a similar result. When
assembling data from the existing literature, the analytical results presented
in case reports may be somewhat questionable and direct comparison with
results from current methods may be difficult.
Despite uncertainties involved in the investigation of consumer complaints,
threshold doses or minimal eliciting doses do exist below which allergic
individuals will not experience adverse reactions.49,51 These minimal eliciting
doses are likely to be very low and variable from one allergic individual to
another.49 For example, the minimal eliciting doses for peanut needed to
provoke mild, objective adverse reactions in one particular group of peanut-
allergic individuals ranged from 2 to over 50 mg.50 Although a rather large
number of low-dose challenge trials have now been conducted for peanut,
milk, and egg in particular, consensus does not yet exist on the minimal
eliciting dose for the most sensitive individuals. For example, other investigators
have occasionally identified individuals with lower minimal eliciting doses
than 2 mg.52 While consensus on minimal eliciting doses has not been achieved,
dose-response modeling for peanut, milk, and egg (where the largest number
of challenges has been obtained) indicate that only 1% or less of the affected
population would react to residual levels of the allergenic food of 0.26 mg
for egg, 0.76 mg for peanut, and 7.76 mg for milk.51 Additional controlled
challenge studies conducted with standard challenge protocols will likely
reduce the uncertainty of these estimates in coming years.
Existing test methods for residues of allergenic foods appear to be in the
ideal range to detect potentially hazardous residues of undeclared allergens
in foods. Current immunoassay methods have lower limits of about 1 ppm
(1 mg/kg). If minimal eliciting doses are 200 μg or higher and typical serving
sizes are 50–100 g or less, then potentially hazardous foods would be easily
detected with these immunoassay methods. If a highly sensitive individual
ingested a typical 50 g serving of a food containing 200 μg of undeclared
residues of an allergenic food, that food would have contained 4 ppm of the
offending food. Previously, the advice given was that foods containing less
than 10 ppm of undeclared residues of an allergic food would present little
risk to food-allergic consumers.53 That advice remains fairly sound now that
much more evidence exists on minimal eliciting doses. The development of
analytical procedures that are more sensitive than the current ones does not
seem to be justified.
Food-allergic consumers, especially those susceptible to comparatively
low threshold levels, can occasionally experience reactions to packaged food
products even when the ingredient list indicates that the particular item
should be safe. In manufacturing situations, foods may become contaminated
with trace amounts of other foods through various means including the use
of rework and the use of shared equipment.54,55 No avoidance diet provides
absolute safety, but careful adherence to an effective avoidance diet will
minimize the chances of a reaction.
1.3.2 Allergenicity of ingredients derived from known allergenic

sources
Many ingredients are derived from sources that are known to be allergenic.
Very common examples would include peanut and soybean oils, soybean
lecithin, and hydrolyzed milk, wheat, or soybean proteins. Less well known
examples would include fish gelatin used to encapsulate certain vitamins and
xanthan gum, a fermentation product made using wheat or soybean substrates.
When considering ingredients derived from an allergenic food source, the
presence of the allergenic protein is an important consideration.
Refined edible oils contain extremely low levels of residual proteins.56
Clinical trials have documented that refined peanut and soybean oils are safe
for individuals allergic to the source food.57,58 However, less highly refined
oils including cold-pressed peanut oil59 and sesame seed oil60 can contain
sufficient protein residues to elicit allergic reactions in individuals sensitive
to the source foods.
Lecithin can be derived from several sources, although soy lecithin is

probably the most commonly used commercially. Soy lecithin definitely
contains residual amounts of soy protein and detectable soy allergens.61,62
Extensive surveys of the amount of protein in commercial soy lecithin have
not been conducted although some variability is likely to exist because several
different commercial processes are used. Soy lecithin does not appear to
elicit adverse reactions in the majority of soybean-allergic individuals, although
clinical trials have not been conducted to critically examine the allergenicity
of soy lecithin to such individuals.
If the proteins have been extensively hydrolyzed, the allergenicity of the
protein is likely to be eliminated.17 For example, extensively hydrolyzed
casein serves as the basis for hypoallergenic infant formulae recommended
for milk-allergic infants. However, protein hydrolysates used in the food
industry can vary widely in terms of degree of hydrolysis, and only those
that are extensively hydrolyzed would be safe for individuals allergic to the
source food. Of course, hydrolysis of the protein can dramatically affect the
ability to detect residual proteins, especially using immunochemical techniques,
even in cases where residual allergenicity may remain.
In other cases, the nature of the protein may have some impact on
allergenicity. Although fish gelatin is derived from fish, this ingredient is
primarily composed of fish collagen while the fish allergens are primarily
parvalbumins.63 Recently, fish gelatin derived from codfish skins was
documented to be safe for the vast majority of codfish-allergic individuals at
levels up to 3.61 g of cumulative intake.64 If foods derived from allergenic
sources contain detectable protein residues, the safety of these foods must be
established by clinical trials in sensitive individuals. Alternatively, the foods
should be labeled to declare the source of the ingredient.
1.3.3 Cross-reactions
Ubiquitous statements cannot be made about cross-reactions between closely-
related foods. Cross-reactivity to closely-related foods occurs in some situations,
but not others. For example, individuals with a shrimp allergy are often
sensitive to all species of shrimp and to other crustacean shellfish such as
crab and lobster.65 Tropomyosin is a pan-allergen existing in all crustacean
shellfish66 and possibly also in some species of molluscan shellfish.66 Fish
have a distinctly different pan-allergen, parvalbumin, that exists in all species
of fish,63 although fish may also contain certain other minor allergens and
cross-reactivity has not been universally observed.63 Some peanut-allergic
individuals are allergic to other legumes, such as soybeans,67 although this is
not a frequent occurrence. Clinical hypersensitivity to one legume, such as
peanuts or soybeans, does not warrant dietary elimination of the entire legume
food family unless allergy to each legume is individually confirmed by double-
blind, placebo-controlled food challenges (DBPCFC).68 In contrast, it is
known that cross-reactions commonly occur between different species of
avian eggs69 and between cows’ milk and goats’ milk.70
As noted previously, cross-reactions also occur between some types of

pollens and certain foods. These include ragweed pollen and melons
(watermelon, cantaloupe, honeydew), mugwort pollen and celery, and
birch pollen and various foods such as carrots, apples, hazelnuts, and
potatoes.71–74 Cross-reactions also occur between latex and certain fruits,
particularly banana, chestnut, and avacado.74 Patients with a history of allergic
reactions to latex should be aware of the potential for allergic reactions to
certain fruits.
The clinical observation of cross-reacting IgE antibodies is occasionally
unexpected and confusing, but does not invariably imply the existence of an
allergy to each food. For the interpretation of IgE antibody assays, it is
important to appreciate that finding IgE antibodies to an allergen does not
imply that the patient has ever been exposed to that allergen or that they will
react after ingestion of that food.75
1.4 Future trends

Research in food allergies has been rapidly expanding in recent years with
the recognition that the prevalence and severity of food allergies is increasing.
Research is progressing on multiple fronts. From the clinical perspective,
research will continue on the prevalence of food allergies including specific
food allergies; improved diagnostic methods, particularly focusing on strategies
that circumvent the need for DBPCFC in all cases; an improved understanding
of the role of delayed hypersensitivity reactions in food allergy; and better
treatment options for food allergy, ranging from anti-IgE therapy to vaccines
to ‘cure’ allergies to probiotic approaches that will aid in the prevention of
food allergies in infants. However, perhaps the most important research will
be to better define threshold doses. While this is clinical research, there is
currently no more important research than this for the food industry. How
much is too much? That is a critical question that will drive food industry
labeling and sanitation (and other allergen control) strategies as well as
improved clinical approaches oriented toward the needs of individual patients.
From the food industry perspective, research should also continue on assessment
of the potential allergenicity of ingredients derived from allergenic sources
and on the effectiveness of various sanitation practices. How clean is clean
enough? That question is also related to the threshold issue and cannot
effectively be answered until more data are available on minimal eliciting
doses.
From an analytical perspective, more research is needed to develop methods
for the detection of residues of allergenic foods that are not commercially
available yet. Such analytical methods will provide important tools for the
food industry and for regulatory agencies. With respect to regulatory agencies,
the detection of allergenic food residues can be of tremendous assistance in
their role as protectors of the health of food-allergic consumers. Obviously,
these methods will be important quality assurance tools for the food industry
because they will allow the validation of sanitation practices on shared
equipment. Shared equipment is an economic necessity in the food industry
and is common in a wide variety of food industry sectors (e.g. ice cream,
confectionary, bakery, pasta). Proper sanitation of that shared equipment is
critical to mitigation of potential allergen hazards, and analytical tools will
facilitate the validation of sanitation practices. Clearly, analytical methods
must be specific, allowing the detection of an allergenic food in the presence
of others, and must be appropriately sensitive. The ideal range of sensitivity
of the methods should be driven by knowledge of minimal eliciting doses.
The goal should be to protect allergic consumers while allowing them the
greatest variety of foods in their diets rather than the detection of miniscule
levels of allergenic food residues that have no adverse health consequences
on even the most sensitive individuals.
1.5 Sources of further information and advice

The latest clinical information on food allergies is often available from medical
societies such as the American Academy of Allergy, Asthma and Immunology
(www.aaaai.org) and the European Academy of Allergy and Clinical
Immunology (www.eaaci.org). Consumer groups can be sought out for
information from the consumers’ perspective, but some care must be taken to
assure the accuracy of this information. Consumer groups do not exist in all
countries. The Food Allergy and Anaphylaxis Network in the US
(www.foodallergy.org), the Anaphylaxis Campaign in the UK
(www.anaphylaxis.org.uk), Nederlands Anafylaxis Network in the Netherlands
(www.anafylaxis.net), Anaphylaxis Australia (www.allergyfacts.org.au),
Allergy New Zealand (allergy.org.nz), and Association Quebecoise des Allergies
Alimentaires in Quebec Province of Canada (www.aqaa.qc.ca) serve as
examples of such consumer groups. Regulatory agencies should be consulted
for information on labeling and other regulations that may exist in various
locales to protect food-allergic consumers. Regulations vary greatly from
one country to another. The Codex Alimentarius Commission
(www.codexalimentarius.net) and its Committee on Food Labeling are a
good source for general guidance. For regulations specific to individual
countries, the websites of the countries of interest should be examined – for
Europe, the European Food Safety Authority (www.efsa.ev.int) would be the
most appropriate agency. Specific advice can also be sought from various
research groups including the University of Nebraska (Lincoln, Nebraska,
USA) Food Allergy Research and Resource Program (www.farrp.org).
1.6 References
1. Bush, R K and Hefle S L (1996) ‘Food allergens’, Crit Rev Food Sci Nutr, 36
(suppl.), S119–S163.
2. Kidd, J M, Cohen, S H, Sosman A J and Fink J N (1983) ‘Food-dependent exercise-
induced anaphylaxis’, J Allergy Clin Immunol, 71, 407–411.
3. Castells, M C, Horan, R F and Sheffer, A L (2003) ‘Exercise-induced anaphylaxis’,
Curr Allergy Asthma Rpts, 3, 15–21.
4. Knicker, W T (1999) ‘Delayed and non-IgE-mediated reactions’, in Frieri, M and
Kettelhut, B, Food Hypersensitivity and Adverse Reactions – A Practical Guide to
Diagnosis and Management, New York, Marcel Dekker, 165–217.
5. Taylor, S L and Hefle, S L (2001) ‘Food allergies and other food sensitivities’, Food
Technol, 55(9), 68–83.
6. Suarez, F L and Savaiano, D A (1997) ‘Diet, genetics, and lactose intolerance’, Food
Technol, 51, 74–76.
7. Bush, R K, Taylor S L and Hefle S L (2003) ‘Adverse reactions to food and drug
additives’, in Adkinson, N F, Yunginger, J W, Busse, W W, Bochner, B S, Holgate,
S T and Simons, F E R, Middleton’s Allergy Principles and Practice, Vol. 2, 6th edn,
St Louis, MO, Mosby, 1645–1663.
8. Taylor, S L, Stratton, J E and Nordlee, J A (1989) ‘Histamine poisoning (scombroid
fish poisoning): an allergy-like intoxication’, Clin Toxicol, 27, 225–240.
9. Sampson, H A (1990) ‘Food allergy’, Curr Opinion Immunol, 2, 542–547.
10. Sicherer, S H, Munoz-Furlong, A and Sampson, H A (2004) ‘Prevalence of seafood
allergy in the United States determined by a random telephone survey’, J Allergy
Clin Immunol, 114, 159–165.
11. Sicherer, S H, Munoz-Furlong, A, Burks, A W and Sampson, H A (1999) ‘Prevalence
of peanut and tree nut allergy in the US determined by a random digit dial telephone
survey’, J Allergy Clin Immunol, 103, 559–562.
12. Sampson, H A and McCaskill, C C (1985) ‘Food hypersensitivity and atopic dermatitis:
evaluation of 113 patients’, J Pediatr, 107, 669–675.
13. Taylor, S L and Hefle, S L (2002) ‘Allergic reactions and food intolerances’, in
Kotsonis, F N and Mackey, M A, Nutritional Toxicology, 2nd edn, London, Taylor
and Francis, 93–121.
14. Sampson, H A, Mendelson, L and Rosen, J P (1992) ‘Fatal and near-fatal anaphylactic
reactions to food in children and adolescents’, N Engl J Med, 327, 380–384.
15. Bock, S A, Munoz-Furlong, A and Sampson, H A (2001) ‘Fatalities due to anaphylactic
reactions to foods’, J Allergy Clin Immunol, 107, 191–193.
16. Pastorello, E A and Ortolani, C (2003) ‘Oral allergy syndrome,’ in Metcalfe, D D,
Sampson, H A and Simon, R A, Food Allergy – Adverse Reactions to Foods and Food
Additives, 3rd edn, Malden, MA, Blackwell Science, 169–182.
17. Taylor, S L and Lehrer, S B (1996) ‘Principles and characteristics of food allergens’,
Crit Rev Food Sci Nutr, 36 (Suppl), S91–S118.
18. Calkhoven, P G, Aalbers, M, Koshte, V L, Pos, O, Oei, H D and Aalberse, R C
(1987) ‘Cross-reactivity among birch pollen, vegetables and fruits as detected by
IgE antibodies is due to at least three distinct cross-reactive structures’, Allergy, 42,
382–390.
19. Maulitz, R M, Pratt, D S and Schocket, A L (1979) ‘Exercise-induced anaphylactic
reaction to shellfish’, J Allergy Clin Immunol, 63, 433–434.
20. Kushimoto, H and Aoki, T (1985) ‘Masked type I wheat allergy. Relation to exercise-
induced anaphylaxis’, Arch Dermatol, 121, 355–360.
21. Strober, W (1986) ‘Gluten-sensitive enteropathy: a nonallergic immune hypersensitivity
of the gastrointestinal tract’, J Allergy Clin Immunol, 78, 202–211.
22. Taylor, S L, Hefle, S L and Gauger, B J (2001) ‘Food allergies and sensitivities’, in
Helferich, W and Winter, C K, Food Toxicology, Boca Raton, FL, CRC Press, 1–36.
23. Prausnitz, C and Kustner, H (1921) ‘Studien uber die Ueberempfindlichkeit’, Centralbl
fur Bakt, 86, 160–169.
24. Ishizaka, K, Ishizaka, T and Hornbrook, M M (1966) ‘Physico-chemical properties
of reaginic antibody. V. Correlation of reaginic activity with gamma-E-globulin
antibody’, J Immunol, 97, 840–853.
25. Ishizaka, K, Ishizaka, T and Hornbrook, M M (1966) ‘Physico-chemical properties
of human reaginic antibody. IV. Presence of a unique immunoglobin as a carrier of
reaginic activity’, J Immunol, 97, 75–85.
26. Ishizaka, T (1982) ‘IgE and mechanisms of IgE-mediated hypersensitivity’, Ann
Allergy, 48, 313–319.
27. Taylor, S L (2002) ‘Protein allergenicity assessment of foods produced through
agricultural biotechnology’, Ann Rev Pharmacol Toxicol, 42, 99–112.
28. Esch, R E and Bush, R K (2003) ‘Aerobiology of outdoor allergens’, in Adkinson,
N F, Yunginger, J W, Busse, W W, Bochner, B S, Holgate, S T and Simons, F E R,
Middleton’s Allergy Principles and Practice, Vol. 1, 6th edn, St Louis, MO, Mosby,
529–572.
29. Sicherer, S H (2002) ‘Food allergy’, Lancet, 360, 701–710.
30. Church, M K, Shute, J K and Sampson, A P (2003) ‘Mast cell-derived mediators’,
in Adkinson, N F, Yunginger J W, Busse, W W, Bochner, B S, Holgate, S T and
Simons, F E R, Middleton’s Allergy Principles and Practice, Vol 1, 6th edn, St Louis,
MO, Mosby, 189–212.
31. Bush, R K and Taylor, S L (2003) ‘Histamine’, in Caballero, B, Trugo, L and
Finglas, P, Encyclopedia of Food Sciences and Nutrition, 2nd edn, Oxford, Elsevier
Science, 3108–3111.
32. Holt, P G and Björksten, B (2003) ‘Development of immunological tolerance to
food antigens’, in Metcalfe, D D, Sampson, H A and Simon, R A, Food Allergy –
Adverse Reactions to Foods and Food Additives, 3rd edn, Malden, MA, Blackwell
Science, 81–90.
33. Chandra, R K (1987) ‘Food allergy: setting the theme’, in Chandra, R K, Food
Allergy, Nutrition Research Education Foundation, St. John’s, Newfoundland, 3–5.
34. O’Connor, M E and Schocket, A L (2003) ‘Exercise- and pressure-induced syndromes’,
in Metcalfe, D D, Sampson, H A and Simon, R A, Food Allergy – Adverse Reactions
to Foods and Food Additives, 3nd edn, Malden, MA, Blackwell Science, 262–269.
35. Sampson, H A (1991) ‘Immunologic mechanisms in adverse reactions to foods’,
Immunol Allergy Clin North Am, 11, 701–716.
36. Stern, M and Walker, W A (1985) ‘Food allergy and intolerance’, Pediatr Clin North
Am, 32, 471–492.
37. Ford, R P K, Hill, D J and Hosking, C S (1983) ‘Cows’ milk hypersensitivity:
immediate and delayed onset clinical patterns’, Arch Dis Child, 58, 856–862.
38. Ferguson, A (1997) ‘Gluten-sensitive enteropathy (celiac disease)’, in Metcalfe, D
D, Sampson, H A and Simon, R A, Food Allergy – Adverse Reactions to Foods and
Food Additives, 2nd edn, Cambridge, Blackwell Science, 287–301.
39. Janatuinen, E K, Pikkarainen, P H, Kemppainen, T A, Kosma, V M, Jarvinen, R M
K, Uusitupa, M I J and Julkunen, R J K (1995) ‘A comparison of diets with and
without oats in adults with celiac disease’, N Engl J Med, 333, 1033–1037.
40. Skerritt, J H, Devery, J M and Hill, A S (1990) ‘Gluten intolerance: chemistry,
celiac-toxicity, and detection of prolamins in foods’, Cereal Foods World, 35, 638–
644.
41. Taylor, S L (1990) ‘Food allergies and related adverse reactions to foods: a food
science perspective’, in Perkin, J E, Food Allergies and Adverse Reactions, Rockville,
MD, Aspen Publishers, 189–206.
42. Furukawa, C T (1988) ‘Nondietary management of food allergy’, in Chiaramonte, L
T, Schneider, A T and Lifshitz, F, Food Allergy – A Practical Approach to Diagnosis
and Management, New York, Marcel Dekker, 365–375.
43. Leung, D Y M, Sampson, H A, Yunginger, J W, Burks, A W, Schneider, L C, Wortel,

C H, Davis, F M, Hyun, J D and Shanahan, W R (2003) ‘Effect of anti-IgE therapy
in patients with peanut allergy’, N Engl J Med, 348, 986–993.
44. Laoprasert, N, Wallen, N D, Jones, R T, Hefle, S L, Taylor, S L and Yunginger, J W
(1998) ‘Anaphylaxis in a milk-allergic child following ingestion of lemon sorbet
containing trace quantities of milk’, J Food Prot, 61, 1522–1524.
45. Gern, J E, Yang, E, Evrard, H M and Sampson, H A (1991) ‘Allergic reactions to
milk-contaminated ‘non-dairy’ products’, N Engl J Med, 324, 976–979.
46. Taylor, S L, Hefle, S L and Munoz-Furlong, A (1999) ‘Food allergies and avoidance
diets’, Nutr Today, 34, 15–22.
47. Taylor, S L, Bush, R K and Busse, W W (1986) ‘Avoidance diets – how selective
should we be?’, N Engl Reg Allergy Proc, 7, 527–532.
48. Taylor, S L, Nordlee, J A and Rupnow, J H (1989) ‘Food allergies and sensitivities’
in Taylor, S L and Scanlan, R A, Food Toxicology – A Perspective on the Relative
Risks, New York, Marcel Dekker, 255–295.
49. Taylor, S L, Hefle, S L, Bindslev-Jensen, C, Bock, S A, Burks, A W, Christie, L, Hill,
D J, Host, A, Hourihane, J O’B, Lack, G, Metcalfe, D D, Moneret-Vautrin, D A,
Vadas, P A, Rance, F, Skrypec, D J, Trautman, T A, Malmeheden Yman I and Zeiger,
R S (2002) ‘Factors affecting the determination of threshold doses for allergenic
foods: how much is too much?’, J Allergy Clin Immunol, 109, 24–30.
50. Hourihane, J O’B, Kilburn, S A, Nordlee, J A, Hefle, S L, Taylor, S L and Warner,
J O (1997) ‘An evaluation of the sensitivity of subjects with peanut allergy to very
low doses of peanut protein: a randomized, double-blind, placebo-controlled food
challenge study’, J Allergy Clin Immunol, 100, 596–600.
51. Bindslev-Jensen, C, Briggs, D, Osterballe, M (2002) ‘Can we determine a threshold
level for allergenic foods by statistical analysis of published data in the literature?’,
Allergy, 57, 741–746.
52. Wensing, M, Penninks, A H, Hefle, S L, Koppelman, S J, Bruijnzeel-Koomen, C A
and Knulst, A C (2002) ‘The distribution of individual threshold doses eliciting
allergic reactions in a population with peanut allergy’, J Allergy Clin Immunol, 110,
915–920.
53. Taylor, S L and Nordlee, J A (1996) ‘Detection of food allergens’, Food Technol,
50(5), 231–234, 238.
54. Taylor, S L and Hefle, S L (2000) ‘Good manufacturing practices for allergenic
foods: use of shared equipment’, Food Allergy Intolerance 1, 47–50.
55. Taylor, S L and Hefle, S L (2000) ‘Good manufacturing practices for the food
industry: use of rework’, Food Allergy Intolerance, 1, 114–117.
56. Crevel, R W R, Kerkhoff, M A T and Koning, M M G (2000) ‘Allergenicity of
refined vegetable oils’, Food Chem Toxicol, 38, 385–393.
57. Hourihane, J O’B, Bedwani, S J, Dean, T P and Warner, J O (1997) ‘Randomised,
double-blind, crossover challenge study of allergenicity of peanut oils in subjects
allergic to peanut’, Br Med J, 314, 1084–1088.
58. Bush, R K, Taylor, S L, Nordlee, J A and Busse, W W (1985) Soybean oil is not
allergenic to soybean-sensitive individuals’, J Allergy Clin Immunol, 76, 242–245.
59. Hoffman, D R and Collins-Williams, C (1994) ‘Cold-pressed peanut oils may contain
peanut allergen’, J Allergy Clin Immunol, 93, 801–802.
60. Kanny, G, de Hauteclocque, C and Moneret-Vautrin, D A (1996) ‘Sesame seed and
sesame seed oil contain masked allergens of growing importance’, Allergy, 51, 952–
957.
61. Gu, X, Beardslee, T, Zeece, M, Sarath, G and Markwell, J (2001) ‘Identification of
IgE-binding proteins in soy lecithin’ Int Arch Allergy Immunol, 126, 218–225.
62. Paschke, A, Zunker, K, Wigotzki, M and Steinhart, H (2001) ‘Determination of the
IgE-binding activity of soy lecithin and refined and non-refined soybean oils’ J
Chromatogr B, 756, 249–254.
63. Taylor, S L, Kabourek, J L and Hefle, S L (2004) ‘Fish allergy: fish and products
thereof’ J Food Sci, 69, R175–180.
64. Hansen, T K, Poulsen, L K, Skov, P S, Hefle, S L, Hlywka, J J, Taylor, S L,
Bindslev-Jensen, U and Bindslev-Jensen, C (2004) ‘A randomized, double-blinded,
placebo-controlled oral challenge study to evaluate the allergenicity of commercial,
food-grade fish gelatin’ Food Chem Toxicol, 42, 2037–2044.
65. Lehrer, S B (1986) ‘The complex nature of food antigens: studies on cross-reacting
crustacea allergens’, Ann Allergy, 57, 267–272.
66. Reese, G, Ayuso, R and Lehrer, S B (1999) ‘Tropomyosin: an invertebrate pan-
allergen’, Int Arch Allergy Immunol, 119, 247–258.
67. Herian, A M, Taylor, S L and Bush, R K (1990) ‘Identification of soybean allergens
by immunoblotting with sera from soy-allergic adults’, Int Arch Allergy Appl Immunol,
92, 193–198.
68. Bernhisel-Broadbent, J and Sampson, H A (1989) ‘Cross-allergenicity in the legume
botanical family in children with food hypersensitivity’, J Allergy Clin Immunol, 83,
435–440.
69. Langeland, T (1983) ‘A clinical and immunological study of allergy to hen’s egg
white. VI. Occurrence of proteins cross-reacting with allergens in hen’s egg white as
studied in egg white from turkey, duck, goose, seagull, and in hen egg yolk, and hen
and chicken sera and flesh’, Allergy 38, 399–412.
70. Bellioni-Businco, B, Paganelli, R, Lucenti, P, Giampietro, P G, Perborn, H and
Businco, L (1999) ‘Allergenicity of goat’s milk in children with cow’s milk allergy’,
J Allergy Clin Immunol, 103, 1191–1194.
71. Van Ree, R and Aalberse, R C (1993) ‘Pollen-vegetable food crossreactivity: serological
and clinical relevance of crossreactive IgE’, J Clin Immunoassay, 16, 124–130.
72. Ballmer-Weber, B K, Vieths, S, Luttkopf, D, Heuschmann, P and Wuthrich, B (2000)
‘Celery allergy confirmed by double-blind, placebo-controlled food challenge: a
clinical study in 32 subjects with a history of adverse reactions to celery root’, J
Allergy Clin Immunol, 106, 373–378.
73. Enberg, R N, Leickly, F E, McCullough, J, Bailey, J and Ownby, D R (1987)
‘Watermelon and ragweed share allergens’, J Allergy Clin Immunol, 79, 867–875.
74. Blanco, C, Carrillo, T, Castillo, R, Quiralte, J and Cuevas, M (1994) ‘Latex allergy:
clinical features and cross-reactivity with fruits’, Ann Allergy, 73, 309–314.
75. Aalberse, R C (1992) ‘Clinically significant cross-reactivities among allergens’, Int
Arch Allergy Immunol, 99, 261–264.
2
Classifying food allergens

H. Breiteneder, Medical University of Vienna, Austria
2.1 Introduction
Plant tissues including those used for human consumption contain thousands
of different proteins. Proteins are clustered together into families if they have
residue identities of 30% and greater, or if they have lower sequence identities
but their functions and structures are very similar.1 Families whose proteins
have low sequence identities but whose structures and functional features
suggest a probable common evolutionary origin are placed together in
superfamilies.1 The Structural Classification of Proteins (SCOP) database
(http://scop.mrc-lmb.cam.ac.uk/scop/count.html) to date includes 2164 protein
families and 1232 protein superfamilies.
However, most plant food allergens are included in only a few protein
families and superfamilies. They belong to the cupin superfamily (7S and
11S seed storage proteins), the prolamin superfamily (2S albumins, non-
specific lipid transfer proteins (nsLTPs), alpha-amylase/trypsin inhibitors,
prolamin storage proteins of cereals), or the papain superfamily of cysteine
proteinases (Table 2.1). The pathogenesis-related (PR) proteins represent a
heterogeneous collection of 14 plant protein families that are all involved in
plant resistance to pathogens or adverse environmental conditions.2 Many
plant food allergens are homologous to such PR proteins (Table 2.2).3,4
Storage proteins are the cause of well-known allergic reactions to peanuts
and cereals.3 PR proteins are responsible for pollen–fruit or latex–fruit cross-
reactive syndromes.3,4 In addition, there are some structural and metabolic
plant protein families which harbour allergenic proteins such as the profilins
or isoflavone reductase homologues (Table 2.3).3 In animal-derived food
products one finds a much smaller variety of allergens. Seafood contains
Table 2.1 Allergenic cupins, prolamins and cystein proteinases
Protein superfamily/family Representative allergenic proteins
Cupin superfamily
Vicilins (7S seed storage proteins) Ara h 1 (peanut), alpha-subunit of beta
conglycinin (soybean), Jug r 2 (English
walnut), Len c 1 (lentil), Ana o 1
(cashew), Ses i 3 (sesame)
Legumins (11S seed storage proteins) Ara h 3 and Ara h 4 (peanut), glycinin
subunits (soybean), Cor a 9 (hazelnut),
AMP (almond)
Prolamin superfamily
2S albumins Sin a 1 (yellow mustard), Ber e 1 (Brazil
nut), Jug r 1 (English walnut), Ses i 2
(sesame), Ara h 2, Ara h 6, Ara h 7
(peanut)
Non-specific lipid transfer proteins Pru p 3 (peach), Mal d 3 (apple), Pru ar 3
(apricot), Cor a 8 (hazelnut), Aspa o 1
(asparagus), Lac s 1 (lettuce)
Alpha-amylase/protease inhibitors Hor v 15 (barley), Sec c 1 (rye), RAPs
(rice allergenic proteins)
Cereal prolamins Tri a 19 (wheat), Sec c 20 (rye), Hor v
21 (barley)
Papain superfamily of cysteine proteases
Papain-like cysteine proteases Act c 1 (kiwi), papain (papaya), bromelain
(pineapple), P34/Gly m Bd 30K (soybean)
Table 2.2 Allergenic pathogenesis-related proteins
Plant pathogenesis-related Representative PR-like allergens

protein families
PR-2 (beta-1,3-glucanases) beta-1,3-glucanases from banana, potato,

and tomato
PR-3 (class I chitinases) Pers a 1 (avocado), Cas s 5 (chestnut)
PR-4 (hevein- and Win-like chitinases) Bra r 2 (turnip)
PR-5 (thaumatin-like proteins) Pru av 2 (cherry), Mal d 2 (apple), Cap a 1
(bell pepper),
PR-9 (peroxidases) Tri a Bd 36K (wheat)
PR-10 (Bet v 1 homologues) Mal d 1 (apple), Pru av 1 (sweet
cherry), Api g 1 (celery), Cor a 1.04
(hazelnut), Gly m 4 (soybean)
PR-14 (non-specific lipid transfer Pru p 3 (peach), Mal d 3 (apple), Pru ar 3
proteins) (apricot), Cor a 8 (hazelnut), Aspa o 1
(asparagus), Lac s 1 (lettuce)
Classifying food allergens 23
Table 2.3 Allergenic members of other plant protein families
Other protein families Examples of allergens
Profilins Mal d 4 (apple), Pru av 4 (sweet cherry),

Apig 4 (celery), Cor a 4 (hazelnut), Lyc e
1 (tomato), Ara h 5 (peanut), Gly m 3
(soybean), Ana c 1 (pineapple)
Kunitz-type protease inhibitors 20 kDa Kunitz soybean trypsin inhibitor,
Sola t 2, Sola t 3, Sola t 4 (potato)
Lectins 31 kDa peanut agglutinin
Patatin-like proteins Sola t 1 (potato)
Phenylcoumaran benzylic ether Pyr c 6 (pear)
reductases
Oleosins peanut oleosin
Beta-fructofuranosidases Lyc e 2 (tomato)
Subtilisin-like serine proteases Cuc m 1/cucumisin (melon)
FAD1 containing oxidases Api g 5 (celery)
1
Flavin adenine dinucteotide
basically either the calcium-binding parvalbumins or the tropomyosins that

play a key regulatory role in muscular contractions, and avian and mammalian
food products contain the well-characterized egg and milk allergens (Table
2.4).
2.2 Plant food allergens

2.2.1 The cupin superfamily
The cupins form a large and functionally highly diverse superfamily of proteins
whose evolution can be traced from bacteria to eukaryotes including animals
and higher plants.5 The term cupin (from the Latin term cupa for a small
barrel or cask) has been given to a beta-barrel structural domain identified in
these proteins.6 These proteins have been divided into two categories, one
being the single-domain cupins that contain only one conserved cupin domain.
The two-domain bicupins that include the 7S and 11S globular seed storage
proteins, major components of the human diet, are thought to have evolved
from the duplication of this single microbial sequence.7 Seed storage globulins
from various legumes and tree nuts are highly abundant proteins comprising
up to 50% of the total seed protein. They provide resources for the growth of
the seedling. The globulins can be divided into two groups based on their
sedimentation coefficient, the 7S vicilin-type globulins and the 11S legumin-
type globulins. They have been studied in the most detail in legumes,
particularly in soy.
Table 2.4 Animal-derived food allergens
Protein family Examples of allergens
Calcium-binding EF hand proteins

Parvalbumins Gad c 1 (cod), Gad m 1 (cod), Sal s 1
(salmon), Cyp c 1 (carp)
Tropomyosins Pen i 1 (Indian shrimp), Par f 1
(Taiwanese shrimp), Pen a 1 (brown
shrimp), Cha f 1 (common crab), Pan s
1 (spiny lobster), Hom a 1 (American
lobster), Cra g 1 and Cra g 2 (Pacific
oyster), Tur c 1 (a gastropod), Tod p 1
(squid), Per v 1 (tropical green mussel)
ATP:guanido phosphotransferases
Arginine kinases Par f 1 (Parapenaeus fissurus shrimp)
Pen m 2 (Penaeus monodon shrimp)
Glycoside hydrolase family 22
Alpha-lactalbumins Bos d 4 (cows’ milk)
Lysozymes Gal d 4 (hen’s egg)
Lipocalins
Beta-lactoglobulins Bos d 5 (cows’ milk)
Serum albumins Bos d 6 (cows’ milk and meat)
Gal d 5 (hen’s egg and chicken meat)
Immunoglobulins Bos d 7 (bovine IgG)
Alpha/beta-caseins Bos d 8 (cow’s milk)
Transferrins lactoferrin (cow’s milk)
Gal d 3 (ovotransferrin, hen’s egg)
Kazal-type serine protease inhibitors
Ovomucoids Gal d 1 (hen’s egg ovomucoid)
Serpins
Ovalbumins Gal d 2 (hen’s egg ovalbumin)
Vicilins
Molecular characteristics
The mature 7S globulins are homotrimeric proteins of about 150–190 kDa
but may undergo reversible aggregation into hexamers, depending on the
ionic strength of the ambient medium.8 The molecular weights of the subunits
range from about 40–80 kDa. Vicilins lack cysteines and, therefore, contain
no disulfide bonds. Their detailed subunit compositions vary considerably
due to differences in proteolytical processing and glycosylation.9 Among the
vicilins are two major variants in size, the regular vicilins with subunits of
about 50 kDa and the large vicilins of the convicilin and beta-conglycinin
type which have additional N-terminal insertions of about 80–190 residues
in length.10 The three-dimensional structures of three 7S globulins have been
determined: canavalin from jack bean,11,12 phaseolin from French bean,13
and the beta subunit of beta-conglycinin from soybean.14 These structures
show that the trimeric vicilins are disk shaped.
Allergenic vicilins
Probably the best characterized allergenic vicilin is the major peanut allergen
Ara h 1, one of the main storage proteins of the peanut, and recognized by
serum IgE from over 90% of peanut allergic patients.15 Its linear IgE-epitopes
have been mapped and shown to consist of 23 independent binding sites.16
Mutational analysis of the immunodominant epitopes revealed that single
amino acid changes within these peptides had dramatic effects on IgE-binding
characteristics. The 63.5 kDa Ara h 1 forms highly stable homotrimers, a
characteristic that may be important for its allergenic properties.17 The majority
of the IgE-binding epitopes are located in the area of the subunit–subunit
contacts. These sites are protected from protease degradation, indicating that
the protein structure may play a significant role in overall allergenicity.18
Ara h 1 purified from peanuts and subjected to dry heat treatments at different
temperatures exhibited IgE-binding properties similar to those found for
native Ara h 1, indicating that the allergenicity of Ara h 1 is heat-stable
although the conformation of native Ara h 1 undergoes significant heat-
induced change.19 However, compared with dry roasted peanuts, the relative
amount of Ara h 1 was reduced in fried and boiled preparations as practised
in China which resulted in a significant reduction of IgE-binding intensity.20
Clinically relevant cross-reactivity between pea and peanut does occur. Vicilin
homologues in pea and peanut (Ara h 1) are the molecular basis for this
cross-reactivity.21
Beta-conglycinin, a glycoprotein of 180 kDa, is a major soybean globulin
that comprises about 50% of the 7S fraction. It forms trimers that are composed
of three subunits, alpha, alpha′, and beta. These have molecular weights of
76, 72, and 53 kDa, respectively.22 Beta-conglycinin is a heterogenous mixture
of different molecular species resulting from the various combinations of the
three subunits. One of the major allergenic proteins in the soybean 7S-
globulin fraction was identified as the alpha subunit of beta-conglycinin,
also known as Gly m Bd 60K.23
To date, several other 7S globulin family members have been identified as
food allergens. These include Jug r 2, a 47 kDa allergenic vicilin from
English walnuts;24 Len c 1, a gamma-vicilin subunit from lentils;25 Ana o 1
from cashew;26 and Ses i 3 from sesame seeds.27
Legumins
The mature 11S globulins that are initially assembled as intermediate trimers
are hexameric.28 They comprise six subunit pairs that interact non-covalently
and are arranged in an open ring structure of approximately 360 kDa. Each
subunit pair consists in turn of an acidic 30–40 kDa polypeptide linked by a
disulfide bond to a basic polypeptide of ~ 20 kDa. Each subunit pair is
synthesized as a single precursor that is post-translationally cleaved after
disulfide bridge formation into an acidic and a basic polypeptide chain. The
basic or C-terminal chain of the 11S legumins is related to the C-terminal
region of the 7S vicilins. Legumins are rarely glycosylated. The three-

dimensional structure of proglycinin from soybean, an 11S globulin precursor,
has been determined.29
Allergenic legumins
Two important peanut allergens belong to the 11S legumin-like seed storage
proteins. The 14 kDa Ara h 3 represents the N-terminal part of a 57 kDa
peanut glycinin subunit. A cDNA clone encoding the full-length protein was
isolated and the linear IgE-binding epitopes of Ara h 3 were mapped.30 The
recombinant Ara h 3 was expressed in E. coli and was recognized by serum
IgE from about 50% of a peanut allergic patient population.31 The cDNA
clone of Ara h 4 encodes a protein of 35.9 kDa with a sequence similarity of
70% to the glycinin family of proteins.32 The structure of Ara h 3 is similar
to that of soybean glycinin, and both the basic and the acidic subunit can
bind IgE from peanut allergic individuals.33
The 11S fraction of soybean proteins consists almost entirely of glycinin,
the principal soy globulin. Soybean glycinin is a major seed storage protein
that makes up more than 20% of the soybean seed by dry weight and 35–
40% of total soy protein. Native glycinin is a 350 kDa hexamer composed of
six non-identical subunits. The five known subunits (G1–G5) of soybean
glycinins are encoded by a small gene family. IgE epitopes for the soybean
glycinin G1 acidic chain have been determined and found to be similar to
IgE epitopes of the peanut glycinin Ara h 3.34 Moreover, each basic chain
from the five glycinin subunits reacted with IgE from soybean allergic
individuals to a similar extent.35 The linear IgE epitopes of the basic chain of
the glycinin G2 subunit were mapped and a three-dimensional model was
developed using the known structure of the 7S phaseolin.36 Additional 11S
allergenic plant food globulins were identified as Cor a 9 in hazelnut,37 in
coconut and walnut,38 and as AMP (almond major protein) in almond.39
2.2.2 The prolamin superfamily

The prolamin superfamily includes the major storage proteins of cereal grains,
with the exception of oats and rice (in which the major storage proteins are
11S globulin-like), and also several groups of low molecular mass cysteine-
rich proteins. Prolamins were originally characterized by their solubility in
alcohol–water mixtures and their high content of proline and glutamine,
hence the name prolamin. Allergies to prolamins in the strict sense do not
occur very frequently. Gluten is the protein component of wheat flour. It
consists of numerous proteins. Two different types are responsible for different
physical properties of dough: the glutenins, which are primarily responsible
for the elasticity, and the gliadins, which contribute to the extensibility.40
The glutenins are of two different types, termed low (LMW) and high molecular
weight (HMW) glutenins. Although IgE reactivity to LMW glutenin exists,
coeliac disease is not an IgE-mediated type I allergy but a genetically-
determined chronic inflammatory intestinal disease induced by wheat gluten.41

The interested reader is referred to Chapter 14 by F.W. Janssen. Alpha-
gliadin, gamma-gliadin and low molecular weight subunits have been shown
to bind IgE from patients with dietary allergy to wheat.42,43
While true prolamins are restricted to grass seeds, they form part of a
larger protein superfamily that contains proteins with much wider distribution.
The broader definition of the prolamin superfamily now includes several
groups of proteins that contain many important plant food allergens: 2S
albumin seed storage proteins of dicotyledonous plants, ns LTPs and inhibitors
of alpha-amylase and trypsin from cereal seeds. All of these proteins can be
described as low molecular weight cysteine-rich proteins. They have similar
three-dimensional structures that are rich in alpha helices and are defined in
the SCOP structural database as an all alpha class of proteins with a structure
made up of four helices, a folded leaf, and a right-handed superhelix (scop.mrc-
lmb.cam.ac.uk/scop).1 Interestingly, all three protein families also have
defensive roles against pathogens or pests, at least in some species. The
soybean hydrophobic protein that is responsible for respiratory allergy to
soybeans possesses the characteristic eight cysteine residue skeleton forming
four intra-chain disulfide bridges, illustrating the relationship to the similar
folds of the 2S albumins, nsLTPs and the cereal trypsin/alpha-amylase
inhibitors.44,45
2S albumins
The 2S albumins constitute a family of structurally related homologous proteins
and form a major group of storage proteins in many dicotyledonous plant
species.46 Many 2S albumins possess high levels of sulfur-containing amino
acid residues. Typical 2S albumins, such as the napins from the Brassicaceae
or the Brazil nut 2S albumin, are heterodimeric proteins consisting of two
polypeptide chains with molecular weights of about 4 and 9 kDa. They are
synthesized as single precursor proteins that are proteolytically cleaved with
the loss of a linker peptide and short peptides from both the N- and the C-
terminus.47 These proteins are rich in alpha-helices and are held together by
four disulfide bonds involving eight conserved cysteine residues.46 The major
role of 2S albumins is as seed storage proteins for the developing seed. They
appear to be stable to proteolysis and can bind lipids. In addition, antifungal
activity against a number of plant-pathogenic fungi has been shown for the
napins from radish.48
Allergenic 2S albumins
Water-soluble allergenic 2S albumin storage proteins occur in many cultivated
species of dicotyledonous plants and include Sin a 1 from yellow mustard
seeds (Sinapis alba),49 Bra j 1 from oriental mustard seeds (Brassica juncea),50
the napin BnIII/Bra n 1 from oilseed rape (Brassica napus),51 Ber e 1 from
Brazil nut (Bertholletia excelsior),52 Jug r 1 from the English walnut (Juglans
regia),53 and SFA-8/SSA from sunflower seeds (Helianthus annuus).54 For

Jug r 1 one major linear IgE epitope and its critical core amino acid residues
could be identified.55 Allergenic 2S albumins were also identified in cashew
nuts56 and as Ses i 2, the clinically most important allergen of sesame seeds.27,57
Among the peanut allergens, the 17.5 kDa glycoprotein Ara h 2,58 the 14.5
kDa Ara h 6,32 and the 15.8 kDa Ara h 732 belong to the conglutin protein
family that is related to the 2S albumin family of seed storage proteins. Ara
h 2 has significant homology to trypsin inhibitors and bifunctional trypsin/
alpha-amylase inhibitors. It was shown that Ara h 2 functions as a trypsin
inhibitor and that dry roasting caused a 3.6-fold increase in this activity.59
Non-specific lipid transfer proteins (ns LTPs)

The LTP family is made up of low molecular weight (7–9 kDa) monomeric
proteins. They are able to catalyze the transfer of lipids between liposomes
and membranes in vitro and there is increasing evidence that their in vivo
role may be in cutin biosynthesis.60 LTPs usually accumulate in the outer
epidermal layers of plant organs, thus explaining the stronger allergenicity
of peels as compared to the pulps of Rosaceae fruits. LTPs are made up of a
bundle of four alpha-helices with a lipid binding cavity in the center. LTPs
are resistant to harsh pH changes or thermal treatments and can refold to
their native structure on cooling. Plant LTPs have been identified as important
food allergens. LTPs have also been listed as family 14 of the pathogenesis-
related proteins.2 Their common structural features, such as eight conserved
cysteines forming four disulfide bridges, basic iso-electric point, and high
similarity in amino acid sequence, are the basis of their allergic clinical
cross-reactivity.61 LTPs are potentially severe food allergens because of their
high resistance to pepsin digestion.62
Allergenic nsLTPs
The nsLTPs have an even wider distribution than the 2S albumins with
sequences being available from seeds, fruits, and vegetables. In the
Mediterranean area, Rosaceae fruits, particularly peach, are among the most
frequent causes of food-induced allergic reactions. LTPs represent major
plant food allergens in populations in these regions that are virtually free of
Fagales trees, in particular birch. LTPs have been identified as major peach
(Pru p 3),63,64 apple (Mal d 3),65 and apricot (Pru ar 3)66 allergens in
Mediterranean populations. In general, the LTP allergens of the Prunoideae
subfamily of the botanical family of the Rosaceae whose sequence identities
are between 88 and 98% are highly cross-reactive. In addition to peach and
apricot, this subfamily includes sweet cherry and the European plum, whose
LTPs have been described as the allergens Pru av 367 and Pru d 3,68 respectively.
LTPs seem to be able to sensitize subjects who have not been previously
sensitized by pollen allergens. The full spectrum of cross-reactivity with
different foods containing LTPs has yet to be investigated. Some of the
patients sensitized to peach LTP are also clinically allergic to tomato and
corn, whereas the hazelnut LTP Cor a 8 was described as highly cross-
reactive with the peach LTP.69 The chestnut LTP, Cas s 8, was found to share
only some IgE epitopes with the corresponding peach allergen.70 The list of
allergenic nsLTPs is continuously expanding, encompassing more seeds (Zea
m 14 from corn),71 vegetables (Aspa o 1 from asparagus),72 and fruits (Vit v
1 from grape).73 Although allergic reactions to lettuce are not frequently
reported, its nsLTP was shown to cause anaphylaxis in susceptible individuals
and, therefore, received the nomenclature designation Lac s 1.74
The cereal superfamily of alpha-amylase and protease inhibitors

Inhibitors of this family are present in a range of cereals including wheat,
barley, rye, rice, and corn. They generally have subunits of about 12–16 kDa,
exist in monomeric, dimeric, and tetrameric form, and possess inhibitory
activity against trypsin or various types of alpha-amylases from insects and
fungi, or they may be bifunctional.75 These allergens are capable of sensitizing
susceptible atopic patients by ingestion or by inhalation.76 The inhibitors of
this family present in wheat, barley, and rye are selectively extracted in
chloroform/methanol (CM) mixtures which is reflected in their designation
as CM proteins. They include monomeric inhibitors of trypsin and dimeric
and tetrameric inhibitors of alpha-amylase.
Allergenic alpha-amylase/trypsin inhibitors

Allergens within the cereal superfamily of inhibitors include the glycosylated
subunits of the tetrameric CM16* (CM, soluble in chloroform/methanol; 16,
molecular weight in kDa; *, glycosylated) inhibitor from wheat,77 the
homologous glycoproteins CMb* (b, barley),77 Hor v 15 (Hor v 1/BMAI-1;
barley monomeric alpha-amylase inhibitor),78 and BDP (barley dimeric
protein)79 from barley, and three members of the alpha-amylase/trypsin inhibitor
family from rye flour, Sec c 1,79 RDAI-1 (rice dimeric alpha-amylase inhibitor),
and RDAI-3.80 The best characterized allergens of this group are the alpha-
amylase inhibitors of rice.81,82 cDNA and genomic clones coding for 14–16
kDa rice allergenic proteins have been obtained.83–85
Cereal prolamins
The ethanol-soluble cereal prolamins, gliadins in wheat, secalins in rye, and
hordeins in barley, are the major storage proteins found in the endosperm of
cereal grains. They are unusually rich in proline and glutamine. Several
sulfur-rich members of the prolamin superfamily were identified as allergens
from wheat. The highest IgE reactivity was found for low molecular mass
(LMM) glutenin followed by alpha-gliadin and gamma-gliadin.86 The allergenic
LMM glutenin was shown to contain a number of pentapeptide repeat motifs,
Gln-Gln-Gln-Pro-Pro.87 Alpha-gliadin, gamma-gliadin, and LMM glutenin
subunits have similar sequences comprising a repetitive N-terminal domain

and a non-repetitive C-terminal domain that contains the conserved cysteine
residues typical for the prolamin superfamily. In contrast to the sulfur-rich
alpha- and gamma-gliadins, omega-gliadins consist almost entirely of repeats
and are characterized by a low content of sulfur-containing amino acid residues
and a lack of cysteine residues. Nevertheless, omega-5 gliadin (Tri a 19) has
been shown to be a significant allergen in young children with immediate
allergic reactions to ingested wheat.88 Omega-5 gliadin from wheat cross-
reacts with gamma-70 and gamma-35 secalins from rye (Sec c 20)89 and
with gamma-3 hordein from barley (Hor v 21).89 In addition to IgE-mediated
type I allergy, cereal prolamins are also of crucial importance in coeliac
disease. Coeliac disease is a T-cell mediated chronic inflammatory bowel
disorder with an autoimmune component. This disease is precipitated by the
gluten storage proteins of wheat. The alcohol-soluble fraction, gliadin, has
been most studied, but most or all gluten proteins are likely to be involved,
as are similar proteins of barley (hordeins) and rye (secalins).41
2.2.3 Papain superfamily of cysteine proteinases

This superfamily of proteins contains three groups of cysteine proteases that
are distantly related: the papain group, the bleomycin hydrolase group, and
the calpain group.90 Proteases of the papain superfamily possess a nucleophilic
cysteine residue in their active site that attacks the peptide bonds of proteins
to be cleaved. They are synthesized as preproenzymes and are located in
lysosomes or analogous organelles. These proteases appear to have arisen
early during eukaryotic evolution.
Peptidases C1A, papain-like cysteine proteases

Peptidases are grouped into clans and families. Clans are groups of families
for which there is evidence of common ancestry. Families are grouped by
their catalytic type, the first character representing the catalytic type, so C
stands for cysteine. Cysteine proteases are divided into clans that consist of
proteins which are evolutionary related.91 Clan CA contains the families of
papain (C1), calpain (C2), streptopain (C10), and the ubiquitin-specific
peptidases (C12, C19). The subfamily C1A of cysteine proteases consists of
papain and related plant proteinases such as chymopapain, caricain, bromelain,
actinidin, ficin, and aleurain. These proteases occur widely in plants where
they display a variety of functions including digestion of seed storage proteins,
ageing of flowers and leaves, or processing of enzyme precursors.92 Cysteine
proteases have also been associated with resistance to lepidopteran pests.93
The high levels of cysteine proteases in fruits is consistent with such a
protective role. The most widely studied cysteine protease is papain from the
papaya fruit. Mature cysteine proteases have a molecular weight of about
25–35 kDa. They are typically stabilized by three disulfide bonds and tend
to be stable against denaturation.
Allergenic papain-like cysteine proteases

The family of cysteine proteases includes several proteolytic plant enzymes
such as papain from papaya, ficin from fig, bromelain from pineapple, and
actinidin from kiwi. Patients with allergic reactions to kiwi fruit produced
IgE that was also reactive to bromelain and papain.94 The cross-reactivity of
papain-like allergens was further shown by cross-inhibition experiments
including ficin and papain.95 Actinidin, a thiol protease of kiwi (Actinidia
chinensis) that accounts for about 50% of the soluble protein of the fruit, was
identified as its major allergen and designated Act c 1.96 P34/Gly m Bd 30K,
a soybean seed storage vacuole protein, belongs to the plant thiol protease
family and is an outlying member of the papain superfamily. P34 has lost its
protease activity and as a glycine has replaced the cysteine residue in the
active site. It has been shown to be associated with the storage vacuoles of
soybean.97 P34 has been identified as a major soybean allergen in patients
with atopic dermatitis and food allergy.98 As P34/Gly m Bd 30K is a major
soybean allergen, it was also referred to as Gly m 1 by several research
groups.99,100 However, the Allergen Nomenclature Sub-Committee of the
International Union of Immunological Societies (www.allergen.org) lists a
LMW respiratory allergen from soybean hull as Gly m 1.101,102 The activity
of Gly m Bd 30K as a cysteine protease could not be demonstrated and an
alternative function of binding syringolide elicitor has been proposed.103
2.2.4 Plant pathogenesis-related proteins

Pathogenesis-related proteins (PRs) are defined as proteins that are encoded
by the host plant and induced specifically in response to infections by pathogens
such as fungi, bacteria, or viruses, or by adverse environmental factors. PRs
do not form a superfamily of proteins but represent a collection of unrelated
protein families which function as part of the plant defense system. To date,
PRs have been classified into 14 families.2 Many plant food allergens are
homologous to proteins that are included in PR families (Table 2.2).
PR-2: beta-1,3-glucanases
O-Glycosyl hydrolases are a widespread group of enzymes that hydrolyze
the glycosidic bond between two or more carbohydrates, or between a
carbohydrate and a non-carbohydrate moiety.104 A classification system for
glycosyl hydrolases, based on sequence similarity, has led to the definition
of 85 different families.105 Plant beta-1,3-glucanases are monomers with a
molecular weight in the 25–35 kDa range. Most plant beta-1,3-glucanases
are endoglucanases with the potential to partially degrade fungal cell walls
by hydrolyzing the beta-1,3-glucan fibers of the growing hyphae of filamentous
fungi.
PR-2-like allergens
The latex–fruit syndrome describes allergy to Hevea brasiliensis latex products
and an associated allergy to certain plant foods, especially avocado, banana,
chestnut, fig, and kiwi.106 It is speculated that the inhalation of PR-2 proteins
is one of the causes for the cross-reactivity of latex with plant foods.
Homologous proteins present in these fruits and vegetables are regarded as
responsible for the observed cross-reactivity. The basic beta-1,3-glucanases
from Hevea latex were recognized by specific IgE from food-allergic patients
suffering from hypersensitivity to banana, potato, and tomato – rather than to
latex products.107 Certain isoforms of the allergenic beta-1,3-glucanases from
H. brasiliensis are glycosylated. The carbohydrate moieties seem to harbour
the major IgE epitopes and may be responsible for the observed cross-
reactivities to beta-1,3-glucanases from fruits and vegetables.108 Banana is a
commonly reported cross-reactive food included in the latex–fruit syndrome.
An abundant beta-1,3-endoglucanase was isolated from ripe bananas and its
cDNA was cloned.109 Its allergenic potential has not yet been determined.
PR-3: class I chitinases

Chitinases are enzymes that catalyze the hydrolysis of the beta-1,4-N-acetyl-
D-glucosamine linkages in chitin polymers. Chitinases belong to glycoside
hydrolase families 18 or 19.105 Chitinases of family 19 (also known as classes
IA or I and IB or II) are endochitinases from plants that degrade chitin, a
major structural component of the exoskeleton of insects and of the cell
walls of many pathogenic fungi.110 Class I chitinases contain an N-terminal
so-called hevein domain with putative chitin-binding properties.111
PR-3-like allergens
Relevant allergens of chestnut and avocado have been identified as class I
chitinases.112 Pers a 1, an endochitinase and major allergen from avocado,
was cloned and expressed in the yeast Pichia pastoris.113 Two major IgE-
binding proteins from banana were identified as class I chitinases.114 Class
I chitinases have an N-terminal domain that is homologous to hevein, one of
the major allergens of latex. Hevein (Hev b 6.02) occurs in Hevea latex
cleaved from prohevein and shares high sequence similarities with the hevein
domain of class I chitinases. Hevein contributes to the increased prevalence
of fruit allergies in individuals allergic to latex. Sensitization to Pers a 1
seems to be caused by hevein in the majority of patients allergic to latex.115
Class I chitinases are now regarded as the major pan-allergens in fruits
associated with the latex–fruit syndrome. These enzymes are inactivated by
heat treatment which may explain why only freshly consumed fruits are
associated with the latex–fruit syndrome. The hevein-like domains of allergenic
class I chitinases seem to include all the main IgE epitopes.116
PR-4: Hevein- and Win-like chitinases

The 4.7 kDa hevein (Hev b 6.02), a small antifungal protein that is found in
high concentrations in the latex of Hevea trees, is one of the most important
latex allergens.117 The wound-induced proteins Win1 and Win2 from potato
were found to include a hevein-like domain and revealed high similarity to
the hevein precursor prohevein.118 Two PR-4-type proteins were isolated
from tobacco mosaic virus infected leaves119 and found to constitute another
family of chitinases.2 Their primary structure indicated 50% sequence similarity
to Win1 and Win2. However, these tobacco PR-4 proteins lack the N-terminal
hevein-like domain of the Win proteins and Hev b 6.120
PR-4-like allergens
Wounding and chemical treatment of turnip (Brassica rapa) plants induced
the expression of an 18.7 kDa allergen that was recognized by IgE of 80%
of 60 sera of natural rubber latex allergic individuals.121 It seems likely that
prohevein and hevein were the primary sensitizers and the IgE binding to the
turnip allergen was a result of allergen cross-reactivity. Peptide sequencing
of the turnip allergen revealed 70% identity to prohevein and high similarities
to wound-induced proteins from tomato (74%) and potato (71%). Cloning of
a hevein-like fruit protein from mature elderberry (Sambucus nigra) fruits
demonstrated the occurrence of a hybrid gene encoding a protein consisting
of the N-terminal hevein-like domain of PR-4 proteins and the C-terminal
domain of class V chitinases.122 The allergenic properties of this elderberry
protein remain to be investigated.
PR-5: thaumatin-like proteins (TLPs)

The family of PR-5 proteins comprises unique proteins with diverse functions.
Due to the sequence homologies between PR-5 proteins and thaumatin, an
intensely sweet tasting protein isolated from the fruits of the West African
rain forest shrub Thaumatococcus daniellii, members of this family of proteins
are referred to as thaumatin-like proteins (TLPs). TLPs were classified by
Dudler et al.123 into three groups: (i) those produced in response to pathogen
infection, (ii) those produced in response to osmotic stress, also called osmotins;
and (iii) antifungal proteins present in cereal seeds. TLPs are generally resistant
to proteases and pH- or heat-induced denaturation. This may be due to the
presence of 16 conserved cysteines that form eight disulfide bridges. The
crystal structure of three PR-5-type proteins, PR-5d from tobacco,124 zeamatin
from Z. mays,125 and thaumatin from T. daniellii 126, are known to date.
PR-5-like allergens
Mal d 2 is an important allergen of apple that is associated with IgE-mediated
symptoms in apple-allergic individuals. The cDNA sequence of Mal d 2 was
expressed in Nicotiana benthamiana plants using a recombinant tobacco

mosaic viral vector.127 Purified recombinant Mal d 2 displayed ability to
bind IgE from apple-allergic individuals equivalent to natural Mal d 2. In
addition, the recombinant thaumatin-like Mal d 2 exhibited antifungal activity
against Fusarium oxysporum and Penicillium expansum, implying a function
in plant defense against fungal pathogens. In sweet cherry (Prunus avium),
a 23.3 kDa thaumatin-like protein was identified as a major allergen, designated
Pru av 2, and its cDNA was cloned.128 The N-terminal sequence of a 23 kDa
bell pepper allergen was found to be identical to a corresponding portion of
the osmotin-like protein P23 from tomatoes129 and the complete coding
sequence of Cap a 1 was obtained (EMBL Accession No. AJ297410).130
Recently, a TLP was identified as a minor allergen of grape with an amino
acid sequence highly similar to Mal d 2 and Pru av 2,131 and a TLP from kiwi
was described as the allergen Act c 2.132
PR-9: peroxidases
Peroxidases are heme-containing enzymes that utilize H2O2 for a series of
oxidative reactions. Heme peroxidases include two superfamilies, one found
in bacteria, fungi, and plants, and the other found in animals. The first one
consists of three major classes.133 Class III includes the secretory plant
peroxidases that are monomeric glycoproteins and contain four conserved
disulfide bridges and two calcium ions. They share the same architecture,
two all alpha domains between which the heme group is located. Peroxidases
have been thoroughly researched in higher plants where their activities are
correlated with a large number of growth, developmental, and defense processes.
Specific lignin-forming peroxidases, induced by pathogens and involved in
plant defense against pathogens, have been designated PR-9.134
PR-9-like allergens
A prominent 36 kDa allergen was isolated from wheat flour and identified by
N-terminal and internal amino acid sequences as a peroxidase.135 This wheat
allergen, also referred to as Tri a Bd 36K, was characterized as a glycoprotein.
The carbohydrate moiety seemed to be at least partially involved in the IgE
binding.136 IgE of food-allergic patients bound to a glycosylated peroxidase
from tomato whose cDNA coded for seven potential N-linked glycosylation
sites.137 In general, the role of glycosylation in the allergenicity of plant
peroxidases has not yet been elucidated.
PR-10: proteins homologous to Bet v 1

The PR-10 family is a ubiquitous group of defense-related and intracellular
proteins that are over expressed after fungal or bacterial infections. Bet v 1,
the major birch pollen allergen, was the first allergen described that was
homologous to PR-10 family members.138 PR-10 family members have been
identified in a large number of plant species, both mono- and dicotyledons.139

Many of the allergenic PR-10 proteins exist in various isoforms that differ in
their IgE-binding capacities.140–142 Crystal structures have been determined
for Bet v 1.2801,143 the hypoallergenic isoform Bet v 1l (also designated Bet
v 1.1001),144 and for Pru av 1, the Bet v 1 homologue from cherry.145 These
proteins function most likely as plant steroid hormone transporters.144,145
PR-10-like allergens
The oral allergy syndrome (OAS), an association of food allergies to fruits,
nuts, and vegetables in patients with pollen allergy, is the most frequent
clinical syndrome caused by cross-reactive IgE antibodies. In the majority of
the cases, OAS in individuals allergic to tree pollen is caused by IgE cross-
reactivity between the major birch pollen allergen Bet v 1 and its homologous
proteins. Bet v 1-homologous food allergens have been identified in fruits of
Rosaceae species (Mal d 1 in apple146, Pru av 1 in sweet cherry147, Pru ar 1
in apricot148, Pyr c 1 in pear149), or somatic tissues, vegetables of Apiaceae
species (Api g 1 in celery150, Dau c 1 in carrot151, pcPR1 and pcPR2 in
parsley152). Additional Bet v 1-homologous proteins capable of binding anti-
Bet v 1 IgE were described as pSTH-2 and pSTH-21 from potato.152 Recently,
the Bet v 1-related major allergen of hazelnuts Cor a 1.04 has been characterized
in detail.153 The Bet v 1 homologous allergen SAM22/Gly m 4 of soybean
was found to be responsible for inducing an OAS of extraordinary severity
and severe systemic reactions in some birch pollen allergic individuals.154
PR-14: non-specific lipid transfer proteins

See prolamin superfamily, pp. 28–29.
2.2.5 Allergens from other plant protein families

Profilins
The actin cytoskeleton of a cell is composed of a network of actin filaments
whose organization is regulated by a number of actin-binding proteins. One
of these proteins is profilin, a 12–15 kDa monomeric actin-binding protein.
The allergenic nature of profilin was first discovered by the ability of the birch
pollen profilin Bet v 2 to bind IgE from birch pollen-allergic individuals.155
Today, profilins are well-known ubiquitous cross-reactive plant allergens.156
Allergenic profilins
In most cases, patients who are sensitized to pollen profilins characteristically
react with a wide range of profilins from nutritive allergen sources. Profilins
are involved in the celery–mugwort–spice syndrome.157 A hazelnut profilin,
Cor a 2, was identified as a relevant IgE-binding protein for a minority of
pollen–nut allergic individuals.158 Grass pollen profilin-allergic patients also
reacted to homologues in celery and carrots, Api g 4 and Dau c 4, respectively.159
IgE of food-hypersensitive tree pollen-allergic individuals cross-reacted with

profilins from apple (Mal d 4), pear (Pyr c 4), carrot (Dau c 4), celery (Api
g 4), and potato.158 An anaphylactic reaction to lychee fruit was mediated by
a profilin, Lit c 1.160 Furthermore, profilins were identified as causing allergic
reactions to tomato (Lyc e 1)161 and to pumpkin seeds.162 The cDNAs coding
for profilins from peanut,32 soybean,163 and celery164 were cloned, produced
as recombinant allergens, and designated Ara h 5, Gly m 3, and Api g 4,
respectively. Furthermore, the allergenicity of profilins from pear, cherry (Pr
u av 4), and celery,165 and from banana (Mus xp 1) and pineapple (Ana c
1)166 have been studied in detail.
Kunitz-type protease inhibitors

The soybean trypsin inhibitor (Kunitz) family is one of the numerous families
of proteinase inhibitors.167 It comprises plant proteins which have inhibitory
activity against serine proteinases from the trypsin and subtilisin families,
thiol proteinases, and aspartic proteinases, as well as some proteins that are
probably involved in seed storage. All the actively inhibitory members contain
two disulfide bridges. The Kunitz family of trypsin inhibitors is present in a
range of legume species and has been characterized in most detail from
soybean. The soybean inhibitors have a molecular weight of about 21 kDa
and a single reactive site for trypsin. Soybean trypsin inhibitor inhibits the
growth of lepidopteran and coleopteran larvae.168
Allergenic Kunitz protease inhibitors

The 20 kDa Kunitz soybean trypsin inhibitor bound IgE of 20% of soy
challenge positive patients, indicating that this protein is a minor allergen.169
However, the Kunitz soybean trypsin inhibitor was also reported to induce
food anaphylaxis.170 IgE-binding potato proteins with molecular weights
from 16–20 kDa were identified as protease inhibitors belonging to the
family of soybean trypsin inhibitors and designated Sola t 2, Sola t 3, and
Sola t 4.171
Lectins
Lectins, also known as plant agglutinins, are a class of proteins that bind to
specific sequences of sugar determinants on glycoproteins. Lectins are found
in seeds, especially those of legumes. Some lectins react unspecifically with
the carbohydrate moieties of IgE, induce histamine release and can thus
induce allergy-like symptoms.172
Allergenic lectin
The 31 kDa peanut agglutinin was identified as a lectin that is specifically
recognized by IgE from a minority of peanut allergic patients.173
2.2.6 Emerging plant food allergens

Several of the now well-accepted allergen families, such as the class I chitinases,
the nsLPPs or, more recently, the thaumatin-like allergens have started out
with descriptions of a single allergenic member of the respective protein
family. Over time, evidence has been collected that resulted in properly
establishing these allergen families. As the field of allergen identification
progresses, more allergens are discovered. Whether these allergens will remain
the subject of case studies or serve as the building blocks for further allergen
families, only time will tell. Sola t 1, a patatin storage protein, was described
as a novel allergen of potato tuber.174 Pyr c 6, a Bet v 5-related food allergen
from pear, was identified as a phenylcoumaran benzylic ether reductase.175
A peanut oleosin that belongs to a family of proteins involved in the formation
of oil bodies was suggested as a new allergen.176 Lyc e 2, a glycosylated
allergen from tomato, was characterized as a beta-fructofuranosidase.177 Cuc
m 1/cucumisin, a subtilisin-like allergenic serine protease of melon, was
proposed as a potential pan-allergen.178 Api g 5, a glycoprotein allergen from
celery with homology to flavin adenine dinucleotide-containing oxidases,
was used to show that cross-reactive carbohydrates were capable of eliciting
allergic reactions in vivo.179 Genetically modified plants may become a potential
source of allergens.180 The classic example that has received extensive publicity
was that of the introduction of a Brazil nut 2S albumin allergen in transgenic
soy.52
2.3 Animal derived food allergens

Seafood allergy is a serious food allergy, especially in coastal regions and
fish processing communities where seafood is a common constituent of the
diet. Besides peanuts and tree nuts, fish and shellfish are among the most
frequent causes of IgE-mediated allergic reactions in adolescents and adults.181
The major allergen of crustaceans such as shrimp, crab, and lobster, has been
identified as the muscle protein tropomyosin.182 Invertebrate tropomyosins
are pan-allergens with significant sequence similarity identified in seafood
and insects such as storage and house dust mites and cockroaches.183
Consequently, tropomyosins are also found as aeroallergens, which raises
the possibility of sensitization by the respiratory route.182 Parvalbumins have
been identified as the major allergens of fish that are recognized by IgE of
more than 95% of fish-allergic individuals.184
Cows’ milk is the first foreign antigen source ingested in large quantities
in early infancy. Consequently, cows’ milk allergy (CMA) is a common
disease of infancy and childhood. The incidence of CMA in infancy is
approximately 2–3% in developed countries.185 Normally, children outgrow
their milk allergy. After acidification or chymosin treatment, cows’ milk
proteins can either be found in the lactoserum (whey) or the coagulum (curd).
Whey contains essentially globular proteins, the major ones being beta-
lactoglobulin, alpha-lactalbumin, and bovine serum albumin. The curd fraction

comprises the caseins.186 Beef allergy is rare. A higher prevalence of beef
allergy might be expected among children allergic to milk, because both
foods contain bovine serum albumin, bovine gamma globulin, and other
proteins in significant quantities. Patients clinically reactive only to rare beef
had decreased IgE binding to beef fractions compared with patients reactive
to well-cooked beef. Therefore, patients reacting only to rare beef would not
necessarily have to maintain a complete beef elimination diet.187
Hen’s egg is another food most frequently reported to elicit allergic reactions
in children.188 Egg white protein contains 23 different glycoproteins.
Ovomucoid (Gal d 1), ovalbumin (Gal d 2), ovotransferrin (Gal d 3), and
lysozyme (Gal d 4) have been identified as the major allergens, but ovomucoid
has been shown to be the dominant allergen.189,190 Cross-reactions among
various avian eggs have been described191 and reactions to duck and goose
egg in the absence of hen’s egg allergy have also been reported.192 Reports
of allergy to bird meats are not common. Most cases have been observed in
patients with ‘bird-egg syndrome’ that is based on the presence of alpha-
livetin in egg yolk, feathers, and serum (chicken serum albumin).193 A
subset of the patients with ‘bird-egg syndrome’ are also allergic to chicken
meat possibly due to the serum proteins present in the meat.194 Another
category of patients is allergic to chicken meat but not to egg.195 Patients
allergic to one bird’s meat may be allergic to others, including game birds.
However, allergy to chicken meat is quite rare and allergy to turkey is even
less common.
2.3.1 Calcium-binding EF-hand proteins

Many calcium-binding proteins belong to the same evolutionary family and
share a type of calcium-binding domain known as the EF-hand.196 This type
of domain consists of a twelve residue loop flanked on both sides by a 12-
residue alpha-helical domain.
Parvalbumins
Parvalbumins constitute a class of calcium-binding proteins characterized by
the presence of several helix-loop-helix (EF-hand) motifs. They are present
in the white muscle of many fish species in relatively high amounts of up to
5 mg/g fresh weight. Parvalbumin is assumed to be important for the relaxation
of muscle fibers by binding free calcium in the cells.197 Parvalbumins are
heat stable and resistant to denaturation and proteolytic digestion. The major
allergenic protein of cod (Gadus callarias) is a 12.3 kDa parvalbumin. It was
named allergen M or Gad c 1 and has been intensively studied both structurally
and immunologically.198 Recently, two distinct parvalbumin transcripts were
identified in Atlantic cod suggesting that isotypic variants are generally present
in fish.199 Allergenic parvalbumins have also been described from a second
cod species Gadus morhua as Gad m 1,200 from the salmon Salmo salar as
Sal s 1,201 from the carp Cyprinus carpio as Cyp c 1,202 from tuna,203 from
the edible frog Rana esculenta,204 and from mackerel.205
2.3.2 Tropomyosins
Tropomyosins are a family of closely related proteins present in muscle and
non-muscle cells.206 Tropomyosin plays a key regulatory role in muscle
contraction together with actin and myosin. In non-muscle cells, tropomyosin
is believed to play a role in the regulation of cell morphology and motility.
In muscle cells, two alpha-helical tropomyosin molecules are wound around
each other forming a parallel dimeric alpha-helical coiled-coil structure.
Tropomyosins have a sequence (284 amino acid residues long in most isoforms)
with a seven-residue (‘heptad’) repeat of the form a-b-c-d-e-f-g, where a and
d are generally apolar residues. Most isoforms have been shown to have an
unbroken series of 40 continuous heptads. Tropomyosin molecules bond
head-to-tail with a short overlap to form an unbroken coiled-coil cable that
winds around the actin helix.
Allergenic tropomyosins
Two invertebrate groups, Crustacea and Mollusca, are generally referred to
as shellfish and are common constituents in the diet of many populations.
Tropomyosins are heat stable cross-reactive food allergens present in
crustaceans and molluscs. Tropomyosins were identified as the major allergens
of shrimp by several laboratories. They include the allergenic tropomyosins
Pen i 1 from the Indian shrimp Penaeus indicus,207 Par f 1 from the most
common Taiwanese shrimp Parapenaeus fissurus ,208 Pen a 1 from the brown
shrimp Penaeus aztecus,209 and Met e 1 from Metapenaeus ensis.210 Additional
allergenic crustacean tropomyosins were described as Cha f 1 from the common
crab Charybdis feriatus,211 as Pan s 1 from the spiny lobster Panulirus
stimpsonii, 212 and as Hom a 1 from the American lobster Homarus
americanus.213 Allergenic tropomyosins are also common in molluscs including
Cra g 1214,215 and Cra g 2216 of the Pacific oyster Crassostrea gigas, Tur c 1
of the gastropod Turbo cornutus,217 Tod p 1 from the squid Todarodes
pacificus,218 Per v 1 from the tropical green mussel Perna viridis,219 Hel as
1 from the brown garden snail,220 and the allergenic tropomyosins of the
abalone Haliotis diversicolor and the scallop Chlamys nobilis.221
2.3.3 ATP:guanido phosphotransferases

ATP:guanido phosphotransferases are a family of structurally and functionally
related enzymes that reversibly catalyze the transfer of phosphate between
ATP and various phosphogens.222 The enzymes belonging to this family
include arginine kinases.
Arginine kinases
Arginine kinases catalyze the transfer of phosphate from ATP to arginine. An
arginine kinase of the shrimp species Parapenaeus fissurus was discovered
as a major and novel allergen of crustaceans and named Par f 1.223 The
homologous allergen was also isolated from the crab Portunus
trituberculotus.223 Another allergenic arginine kinase was identified from
the shrimp species Penaeus monodon by two-dimensional immunoblotting,
designated Pen m 2, and its cDNA was cloned.224 Additional allergenic
crustacean arginine kinases from the sand shrimp Metapenaeus ensis, the
lobster species Homarus gammarus, the crawfish species Metanephrops
thomsoni, and the crab species Scylla serrata were described.224
2.3.4 Glycoside hydrolase family 22

O-Glycosyl hydrolases hydrolyze the glycosidic bond between two or more
carbohydrates, or between a carbohydrate and a non-carbohydrate moiety.
Glycosyl hydrolases have been classified into 85 different families based on
sequence similarity.105,225,226 Glycoside hydrolase family 22 comprises alpha-
lactalbumins and lysozymes of the type C. Lysozymes type C and alpha-
lactalbumins probably evolved from a common ancestral protein.
Alpha-lactalbumin, Bos d 4
Lactalbumin, which attaches to beta-galactosyltransferase to create the lactose
synthetase complex, is essential for milk production.227 Alpha-lactalbumin is
a monomeric globular calcium-binding metalloprotein of 14.4 kDa, with
four disulfide bridges.228 The calcium is bound in a loop that is superficially
similar to the classic EF-hand motif. Bovine alpha-lactalbumin is a major
cows’ milk allergen. Alpha-lactalbumin and hen’s egg white lysozyme are
closely related, having evolved from a common ancestral gene.229
Lysozyme, Gal d 4
Lysozyme is a muramidase that catalyzes the hydrolysis of beta-1,4-links
between N-acetyl-muramic acid and N-acetyl-D-glucosamine in the
peptidoglycan of bacterial cell walls. Egg white lysozyme, Gal d 4, is one of
the main allergens of hen’s egg.230
2.3.5 Lipocalins
The lipocalins are a diverse family of proteins comprising extracellular ligand-
binding proteins with high specificity for small hydrophobic molecules.231
These proteins transport pheromones or nutrients, control cell regulation, or
play a role in cryptic coloration or the enzymatic synthesis of prostaglandins.
The crystal structures of several lipocalins have been elucidated and show a
novel eight-stranded anti-parallel beta-barrel fold well conserved within the
family.
Beta-lactoglobulin, Bos d 5
Beta-lactoglobulin (Bos d 5) is the major whey protein in the milk of ruminants
and many other mammals. It binds a wide variety of hydrophobic ligands,
but its function remains unknown. Bos d 5, a major cows’ milk allergen, is
absent from human breast milk. Bos d 5 occurs as a 36 kDa dimer. Each
subunit consists of 162 amino acid residues and possesses two disulfide
bonds and one free cysteine. The relative resistance of Bos d 5 to acid
hydrolysis as well as to proteases allows some of the protein to remain intact
after digestion. There are two main isoforms which differ only in two amino
acid positions.232 The crystal structure of Bos d 5 has been determined,
confirming its membership of the lipocalin protein family.233
2.3.6 Serum albumins

Albumins are the main proteins of plasma. They bind water, cations (such as
Ca2+, Na+, and K+), fatty acids, hormones, bilirubin, and drugs. Their main
function is to regulate the colloidal osmotic pressure of blood. Structurally,
the serum albumins are similar, each of the three homologous domains
containing five or six internal disulfide bonds. The three-dimensional structure
of human serum albumin has been determined by X-ray crystallography to a
resolution of 2.8A.234 Bovine serum albumin, also referred to as Bos d 6,
although present in milk in low quantities, reacted with IgE from 50% of
milk allergic patients.235,236 Chicken serum albumin (Gal d 5, alpha-livetin)
contains three albumin domains and has been implicated as the causative
allergen of the bird-egg syndrome.237 Gal d 5 is a partially heat-labile allergen.
2.3.7 Immunoglobulins
IgE directed against bovine IgG, Bos d 7, was detected in raw beef in 83%
of beef-allergic subjects but in only 24% of beef-tolerant subjects. Complete
inhibition of the IgE reactivity to the bovine IgG was obtained with lamb,
venison, and milk. Bovine IgG appears to be a major cross-reacting meat
allergen that could predict beef allergy.238 The role of Bos d 7 in cows’ milk
remains to be studied.
2.3.8 Alpha/beta-caseins
Caseins are the major protein constituent of milk. The biological function of
the caseins is considered as modulating the precipitation of calcium phosphate
from solution.239 The major allergens of the casein fraction are the calcium-
sensitive alpha-s1-, alpha-s2-, and beta-caseins (whole casein fraction = Bos
d 8).240,241 Alpha-s1 and alpha-s2 caseins from cows, goats, and sheep share
87–98% identical amino acids. The alpha-caseins from these animal species
are highly cross-reactive and consequently milk of sheep and goat is not a
suitable alternative for individuals suffering from cows’ milk allergy
(CMA).235 Oral challenge studies clearly showed that goats’ milk is not an
appropriate cows’ milk substitute for children with IgE-mediated CMA.236
In addition, IgE from children allergic to cows’ milk were capable of recognizing
most of the milk proteins from ewe, goat, and buffalo while no serum contained
IgE that reacted with camels’ milk proteins.242 Interestingly, mares’ milk can
be regarded as a good substitute for cows’ milk, having produced a positive
oral challenge in only 1/24 children with CMA.243
2.3.9 Transferrins
Transferrins are eukaryotic iron-binding glycoproteins that control the level
of free iron in biological fluids. The proteins have arisen by duplication of a
domain, each duplicated domain binding one iron atom. Members of the
family include milk lactotransferrin (lactoferrin) and egg white ovotransferrin
(conalbumin). The structural comparison of allergenic sites in alpha-lactalbumin
and beta-lactoglobulin with the structure of lactoferrin has shown that lactoferrin
also possesses allergenic sites.244 Lactoferrin reacted with IgE from 45% of
patients allergic to cows’ milk.241 Ovotransferrin from hen’s egg, also known
as Gal d 3 or conalbumin, has been identified as another major egg white
allergen.245
2.3.10 Kazal-type serine protease inhibitors

Kazal inhibitors, which inhibit a number of serine proteases (such as trypsin
and elastase), belong to a family of proteins that includes pancreatic secretory
trypsin inhibitor, avian ovomucoid and elastase inhibitor. These proteins
contain between 1 and 7 Kazal-type inhibitor repeats.246 The structure of the
Kazal repeat includes a large quantity of extended chain, two short alpha-
helices and a three-stranded anti-parallel beta sheet.247
Ovomucoids
Avian ovomucoids contain three Kazal-like inhibitory domains.248 Chicken
ovomucoid has been shown to be the dominant hen’s egg white allergen Gal
d 1.190 Gal d 1 comprises 186 amino acid residues that are arranged in three
tandem domains (Gal d 1.1, Gal d 1.2, Gal d 1.3). Each domain contains
three intradomain disulfide bonds. Gal d 1.1 and Gal d 1.2 contain two
carbohydrate chains each, and about 50% of the Gal d 1.3 domains contain
one carbohydrate chain.189 Significantly more IgE from hen’s egg-allergic
patients reacts with the second ovomucoid domain. It has been further suggested
that conformational B-cell epitopes play a significant role in ovomucoid
allergenicity and that the carbohydrate moieties have a minor affect on
allergenicity.189
Serpins
The serpins (serine proteinase inhibitors) and related proteins constitute one
of the earliest described protein superfamilies.249 Serpins are a group of

structurally-related proteins of high molecular weight (400–500 amino acids),
and they are extracellular irreversible serine protease inhibitors. The ovalbumin
family of serpin proteins includes both active protease inhibitor molecules as
well as molecules lacking this activity, such as ovalbumin.250
Ovalbumin, Gal d 2
It has become clear that ovomucoid is the immunodominant protein fraction
in egg white and that the use of commercially purified ovalbumin has led to
an over-estimation of the dominance of ovalbumin as a major egg allergen in
human beings.190
2.4 Future trends

The presently ongoing classification of food allergens by protein families
provides an instrumental framework for the characterization and analysis of
allergenic proteins. In addition, a protein can be evaluated in terms of its
association with a certain protein family that is already known to contain a
high number of allergens such as the cupin or prolamin superfamily. This
knowledge is useful for the safety assessment of genetically modified organisms,
especially genetically engineered crops.
Plant biotechnology and food technology, albeit by different means, aim
to reduce the allergenicity of foods by modification of the allergen structure
or the complete removal of the allergen. An antisense RNA strategy was used
to markedly reduce the mRNA and protein content of a family of 14–16 kDa
rice allergens that belong to the alpha-amylase/trypsin inhibitors.251 Transgene-
induced gene silencing was also applied to completely suppress the expression
of Gly m Bd 30 K, a major soybean allergen and member of the plant thiol
protease family.252 The Gly m Bd 30 K silenced plants and their seeds were
developmentally, structurally and compositionally identical to the control
plants. Future developments in plant biotechnology may allow targeted gene
mutation or a broader application of gene silencing or replacement. This
technology may be limited by the presence of multiple allergens. Moreover,
the long-term stability of these systems needs to be evaluated.
The use of certain processing methodologies to reduce the allergenicity of
foods so far has been largely empirical. One needs to differentiate between
thermolabile allergens and allergens that are more resistant to heat treatment,
decrease in pH, or proteolytic degradation. It is not possible to describe the
full range of literature available on the effects of processing on the allergenicity
of certain proteins within this chapter. Generally, the different allergens of
the pollen/fruit syndrome such as members of the Bet v 1 family (PR-10
proteins) are less resistant to heating and digestion, a fact that is connected
to their ability to cause only oral allergy symptoms. Sensitization occurs
predominantly via inhalation of the pollen allergens. Heating reduces the
allergenicity of apple, hazelnut, and celeriac.253,254 However, the Bet v 1

homologous allergen SAM22/Gly m 4 of soybean was able to induce severe
systemic reactions in birch pollen-allergic individuals.154 Allergens of the
cupin, prolamin or cysteine protease superfamilies predominantly sensitize
via the gastrointestinal tract. Their stability to heating and digestion is attributed
to the presence of several disulfide bonds in the molecules, as has been
shown for example for nsLTPs255 or for the 2S albumin-related conglutin
Ara h 2.256 The increase in the knowledge of allergen and epitope structures
will in the future allow the design of more rational and efficient processing
strategies. Currently the only available therapy for food-allergic individuals
is dietary exclusion. This is the reason why reliable detection methods for
food allergens are of extreme importance to allow accurate food labeling.
Therefore, in the search for recombinant food allergy therapeutics several
molecular biology techniques are being explored.257–260 In general, these
genetically engineered immunotherapeutics will need to possess greatly reduced
IgE epitopes while preserving sequences necessary for T-cell recognition
and for induction of IgG antibodies reactive with the natural allergen.261

There are several databases of allergen sequences on the internet. The Allergen
Nomenclature Subcommittee of the International Union of Immunological
Societies (IUIS) makes the official list of peer-reviewed allergens available
on the net (www.allergen.org). The Allergome (www.allergome.org) is based
on the literature published since the early sixties and provides information
on allergenic molecules for people working in allergy and immunology. The
Food Allergy Research and Resource Program (FARRP) at the University of
Nebraska-Lincoln contains a list of publicly known allergens
(www.allergenonline.com). The biotechnology information for food safety
database contains accession numbers for food allergens, non-food allergens,
and wheat gluten proteins (www.iit.edu/~sgendel/fa.htm). The Structural
Database of Allergenic Proteins (SDAP, fermi.utmb.edu/SDAP/index.html)
is a web server that integrates a database of allergenic proteins with various
bioinformatics tools for performing structural studies related to allergens
and characterization of their epitopes. Nine specialized allergen databases
have been critically reviewed and the need for a centralized allergen reference
database has been pronounced.262
PROTALL is a European network of scientists with expertise relevant to
studying the problems of food allergy. The PROTALL database contains
biochemical and clinical information about plant food allergens
(www.ifr.bbsrc.ac.uk/protall). The SAFE EU project studies apple allergy in
great detail (www.akh-wien.ac.at/safe). InformAll (www.informall.eu.com),
a concerted action project funded under the 5th Framework programme of
the EU, was created to promote the provision of visible, credible food allergy
information sources to consumers, the agri-food industry, allergic consumers,

health professionals, and regulators.
The information that led to the structure- and function-based classification
of food allergens can be found in several databases that specialize in protein
structure, families, and superfamilies. The Protein Databank (PDB) currently
holds almost 32 000 structures of biological macromolecules including crystal
and solution structures of many important allergens (www.rcsb.org/pdb/).
SWISS MODEL is a fully automated protein structure homology-modeling
server that makes protein modeling accessible worldwide (www.expasy.org/
swissmod/SWISS-MODEL.html). The PredictProtein server is a service for
sequence analysis and structure prediction. After submission of a protein
sequence, PredictProtein retrieves similar sequences in the database and
predicts aspects of protein structure (www.embl-heidelberg.de/predictprotein/
predictprotein.html). The CATH protein structure classification
(www.biochem.ucl.ac.uk/bsm/cath/) uses a novel hierarchical classification
of protein domain structures, which clusters proteins at four major levels,
class (C), architecture (A), topology (T) and homologous superfamily (H).
The SCOP database aims to provide a detailed and comprehensive description
of the structural and evolutionary relationships between all proteins whose
structures are known (scop.mrc-lmb.cam.ac.uk/scop/). Pfam, a protein family
database, is a large collection of multiple sequence alignments covering
many common protein domains and families (www.sanger.ac.uk/Software/
Pfam/index.shtml).
2.6 Acknowledgement
This work was supported by the Austrian Science Fund grant SFB F018-02.
2.7 References
1. Murzin, A G, Brenner, S E, Hubbard, T and Chothia, C (1995) ‘SCOP: a structural
classification of proteins database for the investigation of sequences and structures’,
J Mol Biol, 247(4), 536–540.
2. Van Loon, L C and Van Strien, E A (1999) ‘The families of pathogenesis-related
proteins, their activities, and comparative analysis of PR-1 types proteins’, Physiol
Mol Plant Pathol, 55(1), 85–97.
3. Breiteneder, H and Ebner, C (2000) ‘Molecular and biochemical classification of
plant-derived food allergens’, Allergy Clin Immunol 106(1 Pt 1), 27–36.
4. Hoffmann-Sommergruber, K (2000) ‘Plant allergens and pathogenesis-related
proteins. What do they have in common ?’, Int Arch Allergy Immunol 122(3), 155–
166.
5. Dunwell, J M, Khuri, S and Gane, P J (2000) ‘Microbial relatives of the seed
storage proteins of higher plants: conservation of structure and diversification of
function during evolution of the cupin superfamily’, Microbiol Mol Biol Rev, 64(1),
153–179.
6. Dunwell, J M (1998) ‘Cupins: a new superfamily of functionally diverse proteins

that include germins and plant storage proteins’, Biotechnol Genet Eng Rev, 15, 1–
32.
7. Dunwell, J M and Gane, J P (1998) ‘Microbial relatives of seed storage proteins:
conservation of motifs in a functionally diverse superfamily of enzymes’, J Mol
Evol, 46(2), 147–154.
8. Thanh, V B and Shibasaki, K (1979) ‘Major proteins of soybean seeds. Reversible
and irreversible dissociation of beta-conglycinin’, J Agric Food Chem, 27(4), 805–
809.
9. Shewry, P R, Napier, J A and Tatham, A S (1995) ‘Seed storage proteins: structures
and biosynthesis,’ Plant Cell, 7(7), 945–956.
10. Häger, K P and Fischer, H (1999) ‘Molecular phylogenies and structural diversification
of gymnosperm and angiosperm storage globulins’, in Shewry, P R and Casey, R
(eds) Seed Proteins, Dordrecht, Kluwer Academic Publishers, 499–515.
11. Ko, T P, Day, J and McPherson, A (2000) ‘The refined structure of canavalin from
jack bean in two crystal forms at 2.1 and 2.0 A resolution,’ Acta Crystallogr D Biol
Crystallogr, 56(Pt4), 411–420.
12. Ko, T P, Kuznetsov, Y G, Malkin, A J, Day, J and McPherson, A (2001) ‘X-ray
diffraction and atomic force microscopy analysis of twinned crystals: rhombohedral
canavalin’, Acta Crystallogr D Biol Crystallogr, 57(6), 829–839.
13. Lawrence, M C, Izard, T, Beuchat, M, Blagrove, R J and Colman, P M (1994)
‘Structure of phaseolin at 2.2 A resolution. Implications for a common vicilin/
legumin structure and the genetic engineering of seed storage proteins,’ J Mol Biol,
23(5), 748–76.
14. Maruyama, N, Adachi, M, Takahashi, K, Yagasaki, K, Kohno, M, Takenaka, Y,
Okuda, E, Nakagawa, S, Mikami, B and Utsumi, S (2001) ‘Crystal structures of
recombinant and native soybean beta-conglycinin beta homotrimers,’ Eur J Biochem,
268(12), 3595–3604.
15. Burks, A W, Williams, L W, Helm, R M, Connaughton, C, Cockrell G and O’Brien
(1991) ‘Identification of a major peanut allergen, Ara h 1, in patients with atopic
dermatitis and positive peanut challenge,’ J Allergy Clin Immunol, 88(2), 172–179.
16. Burks, A W, Shin, D, Cockrell, G, Stanley, J S, Helm, R M and Bannon G A (1997)
‘Mapping and mutational analysis of the IgE-binding epitopes on Ara h 1, a legume
vicilin protein and a major allergen in peanut hypersensitivity,’ Eur J Biochem,
245(2), 334–339.
17. Shin, D S, Compadre, C M, Maleki, S J, Kopper, R A, Sampson, H, Huang, S K,
Burks, A W and Bannon, G A (1998) ‘Biochemical and structural analysis of the
IgE binding sites on Ara h1, an abundant and highly allergenic peanut protein’, J
Biol Chem, 273(22), 13753–13759.
18. Maleki, S J, Kopper, R A, Shin, D S, Park, C W, Compadre, C M, Sampson, H,
Burks, A W and Bannon, G A (2000) ‘Structure of the major peanut allergen Ara
h 1 may protect IgE-binding epitopes from degradation’, J Immunol, 164(11),
5844–5849.
19. Koppelman, S J, Bruijnzeel-Koomen, C A, Hessing, M and de Jongh, H H (1999)
‘Heat-induced conformational changes of Ara h 1, a major peanut allergen, do not
affect its allergenic properties’, J Biol Chem 274(8), 4770–4777.
20. Beyer, K, Morrow, E, Li, X M, Bardina, L, Bannon, G A, Burks, A W, Sampson,
H A (2001), ‘Effects of cooking methods on peanut allergenicity’, J Allergy Clin
Immunol 107(6), 1077–1081.
21. Wensing, M, Knulst, A C, Piersma, S, O’Kane, F, Knol, E F and Koppelman, S J
(2003) ‘Patients with anaphylaxis to pea can have peanut allergy caused by cross-
reactive IgE to vicilin (Ara h 1)’, J Allergy Clin Immunol 111(2), 420–424.
22. Burks, A W, Brooks, J R and Sampson, H A (1988) ‘Allergenicity of major component
proteins of soybean determined by enzyme-linked immunosorbent assay (ELISA)
and immunoblotting in children with atopic dermatitis and positive soy challenges’,
J Allergy Clin Immunol, 81(6), 1135–1142.
23. Ogawa, T, Bando, N, Tsuji, H, Nishikawa, K and Kitamura, K (1995) ‘Alpha-
subunit of beta-conglycinin, an allergenic protein recognized by IgE antibodies of
soybean-sensitive patients with atopic dermatitis’, Biosci Biotechnol Biochem, 5(5),
831–833.
24. Teuber, S S, Jarvis, K C, Dandekar, A M, Peterson, W R and Ansari, A A (1999)
‘Identification and cloning of a complementary DNA encoding a vicilin-like
proprotein, Jug r 2, from English walnut kernel (Juglans regia), a major food
allergen’, J Allergy Clin Immunol, 104(6), 1311–1320.
25. Sanchez-Monge, R, Pascual, C Y, Diaz-Perales, A, Fernandez-Crespo, J, Martin-
Esteban, M and Salcedo, G (2000) ‘Isolation and characterization of relevant allergens
from boiled lentils’, J Allergy Clin Immunol, 106(5), 955–961.
26. Wang, F, Robotham, J M, Teuber, S S, Tawde, P, Sathe, S K, Roux, K H (2002)
‘Ana o 1, a cashew (Anacardium occidental) allergen of the vicilin seed storage
protein family’, J Allergy Clin Immunol, 110(1), 160–166.
27 Beyer, K, Bardina, L, Grishina, G and Sampson, H A, ‘Identification of sesame
seed allergens by 2-dimensional proteomics and Edman sequencing: seed storage
proteins as common food allergens’, J Allergy Clin Immunol, 110(1), 154–159.
28. Müntz, K, Jung, R and Saalbach, G (1993) ‘Synthesis, processing and targeting of
legume seed proteins’, in Shewry, P, Prand, W and Stobart, A K (eds), Seed Storage
Compounds: Biosynthesis, Interactions, and Manipulation, Oxford University Press,
128–146.
29. Adachi, M, Takenaka, Y, Gidamis, AB, Mikami, B and Utsumi, S (2001) ‘Crystal
structure of soybean proglycinin A1aB1b homotrimer’, J Mol Biol, 305(2), 291–
305.
30. Rabjohn, P, Helm, E M, Stanley, J S, West, C M, Sampson, H A, Burks, A W and
Bannon, G A (1999) ‘Molecular cloning and epitope analysis of the peanut allergen
Ara h 3’, J Clin Invest, 103(4), 535–542.
31. Rabjohn, P, West, C M, Connaughton, C, Sampson, H A, Helm, R M, Burks, A W
and Bannon, G A (2002) ‘Modification of peanut allergen Ara h 3: effects on IgE
binding and T cell stimulation’, Int Arch Allergy Immunol, 128(1), 15–23.
32. Kleber-Janke, T, Crameri, R, Appenzeller, U, Schlaak, M and Becker, W M (1999)
‘Selective cloning of peanut allergens, including profilin and 2S albumins, by
phage display technology’, Int Arch Allergy Immunol, 19(4), 265–274.
33. Koppelman, S J, Knol, E F, Vlooswijk, R A, Wensing, M, Knulst, A C, Hefle, S L,
Gruppen, H and Piersma, S (2003) ‘Peanut allergen Ara h 3: isolation from peanuts
and biochemical characterization’, Allergy, 58(11), 1144–1151.
34. Beardslee, T A, Zeece, M G, Sarath, G and Markwell, J P (2000) ‘Soybean glycinin
G1 acidic chain shares IgE epitopes with peanut allergen Ara h 3’, Int Arch Allergy
Immunol, 123(4), 299–307.
35. Helm, R M, Cockrell, G, Connaughton, C, Sampson, H A, Bannon, G A, Beilinson,
V, Livingstone, D, Nielsen, N C and Burks, A W (2000) ‘A soybean G2 glycinin
allergen. 1. Identification and characterization’, Int Arch Allergy Immunol, 123(3),
205–212.
36. Helm, R M, Cockrell, G, Connaughton, C, Sampson, H A, Bannon, G A, Beilinson,
V, Nielsen, N C and Burks, A W (2000) ‘A soybean G2 glycinin allergen. 2.
Epitope mapping and three-dimensional modeling’, Int Arch Allergy Immunol,
123(3), 213–219.
37. Beyer, K, Grishina, G, Bardina, L, Grishin, A and Sampson, H A (2002) ‘Identification
of an 11S globulin as a major hazelnut food allergen in hazelnut-induced systemic
reactions’, J Allergy Clin Immunol, 110(3), 517–523.
38. Teuber, S S and Peterson, W R (1999) ‘Systemic allergic reaction to coconut
(Cocos nucifera) in 2 subjects with hypersensitivity to tree nut and demonstration
of cross-reactivity to legumin-like seed storage proteins: new coconut and walnut

food allergens’, J Allergy Clin Immunol, 103(6), 1180–1185.
39. Pasini, G, Simonato, B, Giannattasio, M, Gemignani, C and Curioni, A (2000) ‘IgE
binding to almond proteins in two CAP-FEIA-negative patients with allergic
symptoms to almond as compared to three CAP-FEIA-false-positive subjects’,
Allergy, 55(10), 955–958.
40. Tatham, A S and Shewry, P R (2000) ‘Elastomeric proteins: biological roles, structures
and mechanisms’, Trends Biochem Sci 25(11), 567–571.
41. Green, P H and Jabri, B (2003) ‘Coeliac disease’, Lancet 362(9381), 383–393.
42. Maruyama, N, Ichise, K, Katsube, T, Kishimoto, T, Kawase, S, Matsumura, Y,
Takeuchi, Y, Sawada, T and Utsumi, S (1998) ‘Identification of major wheat allergens
by means of the Escherichia coli expression system’, Eur J Biochem, 255(3), 739–
745.
43. Tanabe, S, Arai, S, Yanagihara, Y, Mita, H, Takahashi, K and Watanabe, M (1996)
‘A major wheat allergen has a Gln-Gln-Gln-Pro-Pro motif identified as an IgE-
binding epitope’, Biochem Biophys Res Commun, 219(2), 290–293.
44. Codina, R, Lockey, R F, Fernandez-Caldas, E and Rama, R (1997) ‘Purification
and characterization of a soybean hull allergen responsible for the Barcelona asthma
outbreaks. II. Purification and sequencing of the Gly m 2 allergen’, Clin Exp
Allergy, 27(4), 424–430.
45. Baud, F, Pebay-Peyroula, E, Cohen-Addad, C, Odani, S and Lehmann, M S (1993)
‘Crystal structure of hydrophobic protein from soybean; a member of a new cysteine-
rich family’, J Mol Biol, 231(3), 877–87.
46. Shewry, P R and Pandya, M J (1999) ‘The 2S albumin storage proteins’, in Shewry,
P R, Casey, R (eds) Seed Proteins, Dordrecht, Kluwer Academic Publishers, 563–
586.
47. Shewry, P R, Napier, J A and Tatham, A S (1995) ‘Seed storage proteins: structures
and biosynthesis’, Plant Cell, 7(7), 945–956.
48. Terras, F R, Schoofs, H, De Bolle, M F, Van Leuven, F, Rees, S B, Vanderleyden,
J, Cammue, B P and Broekaert, W F (1992) ‘Analysis of two novel classes of plant
antifungal proteins from radish (Raphanus sativus L.) seeds’, J Biol Chem, 267(22),
15301–15309.
49. Menéndez-Arias, L, Moneo, I, Domínguez, J and Rodríguez, R (1988) ‘Primary
structure of the major allergen from yellow mustard seeds, Sin a 1’, Eur J Biochem,
177(1), 159–166.
50. Monsalve, R I, Gonzáles de la Peña, M A, Menéndez-Arias, L, López-Otín, C,
Villalba, M and Rodríguez, R (1993) ‘Characterization of a new oriental-mustard
(Brassica juncea) allergen, Bra j 1E: detection of an allergenic epitope’, Biochem
J, 293(Pt3) 625–632.
51. Monsalve, R I, Gonzáles de la Peña, M A, López-Otín, C, Fiandor, A, Fernández,
C, Villalba, M and Rodriguez, R (1997) ‘Detection, isolation and complete amino
acid sequence of an aeroallergenic protein from rapeseed flour’, Clin Exp Allergy,
27(7), 833–841.
52. Nordlee, J A, Taylor, S L, Townsend, J A, Thomas, L A and Bush, R K (1996)
‘Identification of a Brazil-nut allergen in transgenic soybeans’, New Engl J Med,
334(11), 688–692.
53. Teuber, S S, Jarvis, K C, Dandekar, A M, Peterson, W R and Ansari, A A,
‘Identification and cloning of a complementary DNA encoding a vicilin-like
proprotein, Jug r 2, from English walnut kernel (Juglans regia), a major food
allergen’, J Allergy Clin Immunol, 104(6), 1311–1320.
54. Kelly, J D, Hlywka, J J and Hefle, S L (2000) ‘Identification of sunflower seed
IgE-binding proteins’, Int Arch Allergy Immunol, 121(1), 19–24.
55. Robotham, J M, Teuber, S S, Sathe, S K and Roux, K (2002) ‘Linear IgE epitope
mapping of the English walnut (Juglans regia) major food allergen, Jug r 1’, J
Allergy Clin Immunol, 109(1), 143–149.
56. Teuber, S S, Sathe, S K, Peterson, W R and Roux, K H (2002) ‘Characterization of

the soluble allergenic proteins of cashew nut (Anacardium occidentale L.)’, J
Agric Food Chem, 50(22), 6543–6549.
57. Pastorello, E A, Varin, E, Farioli, L, Pravettoni, V, Ortolani, C, Trambaioli, C,
Fortunato, D, Giuffrida, M G, Rivolta, F, Robino, A, Calamari, A M, Lacava, L and
Conti, A (2001) ‘The major allergen of sesame seeds (Sesamum indicum) is a 2S
albumin’, J Chromatogr B Biomed Sci Appl, 756(1–2), 85–93.
58. Burks, A W, Williams, L W, Connaughton, C, Cockrell, G, O’Brien T and Helm,
R M (1992) ‘Identification and characterization of a second major peanut allergen,
Ara h II, with use of sera of patients with atopic dermatitis and positive peanut
challenge’, J Allergy Clin Immunol, 90(6Pt1), 962–969.
59. Maleki, S J, Viquez, O, Jacks, T, Dodo, H, Champagne, E T, Chung, S Y and
Landry, S J (2003) ‘The major peanut allergen, Ara h 2, functions as a trypsin
inhibitor, and roasting enhances this function’, J Allergy Clin Immunol 112(1),
190–195.
60. Kunst, L and Samuels, A L (2003) ‘Biosynthesis and secretion of plant cuticular
wax’, Prog Lipid Res 42(1), 51–80.
61. Pastorello, E A, Pompei, C, Pravettoni, V, Brenna, O, Farioli, L, Trambaioli, C and
Conti, A (2001) ‘Lipid transfer proteins and 2S albumins as allergens’, Allergy,
56(Suppl 67), 45–57.
62. Asero, R, Mistrello, G, Roncarolo, D, de Vries, S C, Gautier, M F, Ciurana, C L,
Verbeek, E, Mohammadi, T, Knul-Brettlova, V, Akkerdaas, J H, Bulder, I, Aalberse,
R C and van Ree, R (2001) ‘Lipid transfer protein: a pan-allergen in plant-derived
foods that is highly resistant to pepsin digestion’, Int Arch Allergy Immunol, 124(1–
3), 67–69.
63. Sanchez-Monge, R, Lombardero, M, Garcia-Selles, F J, Barber, D and Salcedo, G
(1999) ‘Lipid-transfer proteins are relevant allergens in fruit allergy’, J Allergy
Clin Immunol, 103(3Pt1), 514–519.
64. Pastorello, E A, Farioli, L, Pravettoni, V, Ortolani, C, Ispano, M, Monza, M,
Baroglio, C, Scibola, E, Ansaloni, R, Incorvaia, C and Conti A (1999) ‘The major
allergen of peach (Prunus persica) is a lipid transfer protein’, J Allergy Clin Immunol,
103(3Pt1), 520–526.
65. Pastorello, E A, Pravettoni, V, Farioli, L, Ispano, M, Fortunato, D, Monza, M,
Giuffrida, M G, Rivolta, F, Scibola, E, Ansaloni, R, Incorvaia, C, Conti, A and
Ortolani, C (1999) ‘Clinical role of a lipid transfer protein that acts as a new apple-
specific allergen’, J Allergy Clin Immunol, 104(5), 1099–1106.
66. Pastorello, E A, D’Ambrosio, F P, Pravettoni, V, Farioli, L, Giuffrida, G, Monza,
M, Ansaloni, R, Fortunato, D, Scibola, E, Rivolta, F, Incorvaia, C, Bengtsson, A,
Conti, A and Ortolani, C (2000) ‘Evidence for a lipid transfer protein as the major
allergen of apricot’, J Allergy Clin Immunol, 105(2Pt1), 371–377.
67. Scheurer, S, Pastorello, E A, Wangorsch, A, Kastner, M, Haustein, D and Vieths, S
(2001) ‘Recombinant allergens Pru av 1 and Pru av 4 and a newly identified lipid
transfer protein in the in vitro diagnosis of cherry allergy’, J Allergy Clin Immunol,
107(4), 724–731.
68. Pastorello, E A, Farioli, L, Pravettoni, V, Giuffrida, M G, Ortolani, C, Fortunato, D,
Trambaioli, C, Scibola, E, Calamari, A M, Robino, A M and Conti, A (2001)
‘Characterization of the major allergen of plum as a lipid transfer protein’, J
Chromatogr B Biomed Sci Appl, 756(1–2), 95–103.
69. Pastorello, E A, Vieths, S, Pravettoni, V, Farioli, L, Trambaioli, C, Fortunato, D,
Luttkopf, D, Calamari, M, Ansaloni, R, Scibilia, J, Ballmer-Weber, B K, Poulsen,
L K, Wutrich, B, Hansen, K S, Robino, A M, Ortolani, C and Conti, A (2002)
‘Identification of hazelnut major allergens in sensitive patients with positive double-
blind, placebo-controlled food challenge results’, J Allergy Clin Immunol, 109(3),
563–570.
70. Diaz-Perales, A, Lombardero, M, Sanchez-Monge, R, Garcia-Selles, F J, Pernas,

M, Fernandez-Rivas, M, Barber, D and Salcedo, G (2000) ‘Lipid-transfer proteins
as potential plant panallergens: cross-reactivity among proteins of Artemisia pollen,
Castanea nut and Rosaceae fruits, with different IgE-binding capacities’, Clin Exp
Allergy, 30(10), 1403–1410.
71. Pastorello, E A, Farioli, L, Pravettoni, V, Ispano, M, Scibola, E, Trambaioli, C,
Giuffrida, M G, Ansaloni, R, Godovac-Zimmermann, J, Conti, A, Fortunato, D and
Ortolani, C (2000) ‘The maize major allergen, which is responsible for food-
induced allergic reactions, is a lipid transfer protein’, J Allergy Clin Immunol,
106(4), 744–751.
72. Diaz-Perales, A, Tabar, A I, Sanchez-Monge, R, Garcia, B E, Gomez, B, Barber, D
and Salcedo, G (2002) ‘Characterization of asparagus allergens: a relevant role of
lipid transfer proteins’, J Allergy Clin Immunol, 110(5), 790–796.
73. Pastorello, E A, Farioli, L, Pravettoni, V, Ortolani, C, Fortunato, D, Giuffrida, M G,
Perono Garoffo, L, Calamari, A M, Brenna, O and Conti, A (2003) ‘Identification
of grape and wine allergens as an endochitinase 4, a lipid-transfer protein, and a
thaumatin’, J Allergy Clin Immunol, 111(2), 350–359.
74. San Miguel-Moncin, M, Krail, M, Scheurer, S, Enrique, E, Alonso, R, Conti, A,
Cistero-Bahima, A and Vieths, S (2003) ‘Lettuce anaphylaxis: identification of a
lipid transfer protein as the major allergen’, Allergy, 58(6), 511–517.
75. García-Olmedo, F, Salcedo, G, Sanchez-Monge, R, Gomez, L, Royo, J and Carbonero,
P (1987) ‘Plant proteinaceous inhibitors of proteinases and alpha-amylases’, Oxford
Surv Plant Mol Cell Biol 4(3), 275–334.
76. James, J M, Sixbey, J P, Helm, R M, Bannon, G A and Burks, A W (1997) ‘Wheat
alpha-amylase inhibitor: a second route of allergic sensitization’, J Allergy Clin
Immunol, 99(2), 239–244.
77. Sánchez-Monge, R, Gomez, L, Barber, D, Lopez-Otín, C, Armentia, A and Salcedo,
G (1992) ‘Wheat and barley allergens associated with baker’s asthma. Glycosylated
subunits of the alpha-amylase-inhibitor family have enhanced IgE-binding capacity’,
Biochem J, 281(3), 401–405.
78. Mena, M, Sanchez-Monge, R, Gomez, L, Salcedo, G and Carbonero, P (1992) ‘A
major barley allergen associated with baker’s asthma disease is a glycosylated
monomeric inhibitor of insect alpha-amylase: cDNA cloning and chromosomal
location of the gene’, Plant Mol Biol, 20(3), 451–455.
79. García-Casado, G, Armentia, A, Sánchez-Monge, R, Sánchez, L M, Lopez-Otín,
C, Salcedo, G (1996) ‘A major baker’s asthma allergen from rye flour is considerably
more active than its barley counterpart’, FEBS Lett, 364(1), 36–40.
80. García-Casado, G, Armentia, A, Sánchez-Monge, R, Malpica, J M and Salcedo, G
(1996) ‘Rye flour allergens associated with baker’s asthma. Correlation between in
vivo and in vitro activities and comparison with their wheat and barley homologues’,
Clin Exp Allergy, 26(4), 428–435.
81. Nakase, M, Usui, Y, Alvarez-Nakase, A M, Adachi, T, Urisu, A, Nakamura, R,
Aoki, N, Kitajima, K and Matsuda, T (1998) ‘Cereal allergens: rice-seed allergens
with structural similarity to wheat and barley allergens’, Allergy, 53(46 Suppl), 55–
57.
82. Nakase, M, Adachi, T, Urisu, A, Miyashita, T, Alvarez, A M, Nagasaka, S, Aoki,
N, Nakamura, R and Matsuda, T (1996) ‘Rice (Oryza sativa L.) alpha amylase
inhibitors of 14–16 kDa are potential allergens and products of a multi gene family’,
J Agric Food Chem, 44(8), 2624–2628.
83. Izumi, H, Adachi, T, Fujii, N, Matsuda, T, Nakamura, R, Tanaka, K, Urisu, A and
Kurosawa, Y (1992) ‘Nucleotide sequence of a cDNA clone encoding a major
allergenic protein in rice seeds. Homology of the deduced amino acid sequence
with members of alpha-amylase/trypsin inhibitor family’, FEBS Lett, 302(3), 213–
216.
84. Adachi, T, Izumi, H, Yamada, T, Tanaka, K, Takeuchi, S, Nakamura, R and Matsuda,

T (1993) ‘Gene structure and expression of rice seed allergenic proteins belonging
to the alpha-amylase/trypsin inhibitor family’, Plant Mol Biol, 21(2), 239–248.
85. Alvarez, A M, Fukuhara, E, Nakase, M, Adachi, T, Aoki, N, Nakamura, R and
Matsuda, T (1995) ‘Four rice seed cDNA clones belonging to the alpha-amylase/
trypsin inhibitor gene family encode potential rice allergens’, Biosci Biotechnol
Biochem, 59(7), 1304–1308.
86. Maruyama, N, Ichise, K, Katsube, T, Kishimoto, T, Kawase, S, Matsumura, Y,
Takeuchi, Y, Sawada, T and Utsumi, S (1998) ‘Identification of major wheat allergens
by means of the Escherichia coli expression system’, Eur J Biochem, 255(3), 739–
745.
87. Tanabe, S, Arai, S, Yanagihara, Y, Mita, H, Takahashi, K and Watanabe, M (1996)
‘A major wheat allergen has a Gln-Gln-Gln-Pro-Pro motif identified as an IgE-
binding epitope’, Biochem Biophys Res Commun, 219(2), 290–293.
88. Palosuo, K, Varjonen, E, Kekki, OM, Klemola, T, Kalkkinen, N, Alenius, H and
Reunala, T (2001) ‘Wheat omega-5 gliadin is a major allergen in children with
immediate allergy to ingested wheat’, J Allergy Clin Immunol, 108(4), 634–638.
89. Palosuo, K, Alenius, H, Varjonen, E, Kalkkinen, N and Reunala, T (2001) ‘Rye
gamma-70 and gamma-35 secalins and barley gamma-3 hordein cross-react with
omega-5 gliadin, a major allergen in wheat-dependent, exercise-induced anaphylaxis’,
90. Berti, P J and Storer, A C (1995) ‘Alignment/phylogeny of the papain superfamily
of cysteine proteases’, J Mol Biol, 246(2), 273–283.
91. Barrett, A J and Rawlings, N D (2001) ‘Evolutionary lines of cysteine peptidases’,
Biol Chem 382(5), 727–733.
92. Rawlings, N D and Barrett, A J ‘Families of cysteine peptidases’, Methods Enzymol
244, 461–486.
93. Pechan, T, Cohen, A, Williams, W P and Luthe, D S (2002) ‘Insect feeding mobilizes
a unique plant defense protease that disrupts the peritrophic matrix of caterpillars’,
Proc Natl Acad Sci USA 99(20), 13319–13323.
94. Gall, H, Kalveram, K J, Forck, G and Sterry, W (1994) ‘Kiwi fruit allergy: a new
birch pollen-associated food allergy’, J Allergy Clin Immunol, 94(1), 70–76.
95. Diez-Gomez, M L, Quirce, S, Aragoneses, E and Cuevas, M (1998) ‘Asthma caused
by Ficus benjamina latex: evidence of cross-reactivity with fig fruit and papain’,
Ann Allergy Asthma Immunol, 80(1), 24–30.
96. Pastorello, E A, Conti, A, Pravettoni, V, Farioli, L, Rivolta, F, Ansaloni, R, Ispano,
M, Incorvaia, C, Giuffrida, M G and Ortolani, C (1998) ‘Identification of actinidin
as the major allergen of kiwi fruit’, J Allergy Clin Immunol, 101(4Pt1), 531–537.
97. Kalinski, A, Weisemann, J M, Matthews, B F and Herman, E M (1990) ‘Molecular
cloning of a protein associated with soybean seed oil bodies that is similar to thiol
proteases of the papain family’, J Biol Chem, 265(23), 13843–13848.
98. Ogawa, T, Tsuji, H, Bando, N, Kitamura, K, Zhu, Y L, Hirano, H and Nishikawa,
K (1993) ‘Identification of the soybean allergenic protein, Gly m Bd 30K, with the
soybean seed 34-kDa oil-body-associated protein’, Biosci Biotechnol Biochem,
57(6), 1030–1033.
99. Ogawa, T, Bando, N, Tsuji, H, Okajima, H, Nishikawa, K and Sasaoka K (1991)
‘Investigation of the IgE-binding proteins in soybeans by immunoblotting with the
sera of the soybean-sensitive patients with atopic dermatitis’, J Nutr Sci Vitaminol
(Tokyo) 37(6), 555–565.
100. Helm, R, Cockrell, G, Herman, E, Burks, A, Sampson, H and Bannon, G (1998)
‘Cellular and molecular characterization of a major soybean allergen’, Int Arch
Allergy Immunol 117(1), 29–37.
101. Gonzalez, R, Polo, F, Zapatero, L, Caravaca, F and Carreira, J (1992) ‘Purification
and characterization of major inhalant allergens from soybean hulls’, Clin Exp
Allergy 22(8), 748–755.
102. Gonzalez, R, Varela, J, Carreira, J and Polo, F (1995) ‘Soybean hydrophobic

protein and soybean hull allergy’, Lancet 346(8966), 48–49.
103. Ji, C, Boyd, C, Slaymaker, D, Okinaka, Y, Takeuchi, Y, Midland, S L, Sims, J J,
Herman, E and Keen, N (1998) ‘Characterization of a 34-kDa soybean binding
protein for the syringolide elicitors’, Proc Natl Acad Sci USA, 95 (6), 3306–3311.
104. Davies, G and Henrissat, B (1995) ‘Structures and mechanisms of glycosyl
hydrolases’, Structure 3(9), 853–859.
105. Henrissat, B (1991) ‘A classification of glycosyl hydrolases based on amino acid
sequence similarities’, Biochem J 280 (Pt 2), 309–316.
106. Brehler, R, Theissen, U, Mohr, C and Luger, T (1997) ‘Latex–fruit syndrome’:
frequency of cross-reacting IgE antibodies’, Allergy, 52(4), 404–410.
107. Yagami, T, Sato, M, Nakamura, A, Komiyama, T, Kitagawa, K, Akasawa, A and
Ikezawa, Z (1998) ‘Plant defense-related enzymes as latex antigens’, J Allergy
Clin Immunol, 101(3), 379–385.
108. Yagami, T, Osuna, H, Kouno, M, Haishima, Y, Nakamura, A and Ikezawa, Z (2002)
‘Significance of carbohydrate epitopes in a latex allergen with beta-1,3-glucanase
activity’, Int Arch Allergy Immunol, 129(1) 27–37.
109. Peumans, W J, Barre, A, Derycke, V, Rouge, P, Zhang, W, May, G D, Delcour, J A,
Van Leuven, F and Van Damme, E J (2000) ‘Purification, characterization and
structural analysis of an abundant beta-1,3-glucanase from banana fruit’, Eur J
Biochem, 267(4), 1188–1195.
110. Kasprzewska, A (2003) ‘Plant chitinases – regulation and function’, Cell Mol Biol
Lett 8(3), 809–824.
111. Beintema, J J (1994) ‘Structural features of plant chitinases and chitin-binding
proteins’, FEBS Lett 350(2–3), 159–163.
112. Diaz-Perales, A, Collada, C, Blanco, C, Sánchez-Monge, R, Carrillo, T, Aragoncillo,
C and Salcedo, G (1998) ‘Class I chitinases with hevein-like domains, but not class
II enzymes, are relevant chestnut and avocado allergens’, J Allergy Clin Immunol,
102(1), 127–133.
113. Sowka, S, Hsieh, L-S, Krebitz, M, Akasawa, A, Martin, B, Starrett, D, Peterbauer,
C K, Scheiner, O and Breiteneder, H (1998) ‘Identification and cloning of Prs a 1,
a 32 kDa major allergen of avocado and its expression in Pichia pastoris’, J Biol
Chem, 273(43), 28091–28097.
114. Sanchez-Monge, R, Blanco, C, Díaz-Perales, A, Collada, C, Carrillo, T, Aragoncillo,
C and Salcedo, G (1999) ‘Isolation and characterization of major banana allergens:
identification as fruit class I chitinases’, Clin Exp Allergy, 29(5), 673–680.
115. Chen, Z, Posch, A, Cremer, R, Raulf-Heimsoth, M and Baur, X (1998) ‘Identification
of hevein (Hev b 6.02) in Hevea latex as a major cross-reacting allergen with
avocado fruit in patients with latex allergy’, J Allergy Clin Immunol, 102(3), 476–
481.
116. Diaz-Perales, A, Sanchez-Monge, R, Blanco, C, Lombardero, M, Carillo, T and
Salcedo, G (2002) ‘What is the role of the hevein-like domain of fruit class I
chitinases in their allergenic capacity?’, Clin Exp Allergy, 32(3), 448–454.
117. Karisola, P, Alenius, H, Mikkola, J, Kalkkinen, N, Helin, J, Pentikainen, O T,
Repo, S, Reunala, T, Turjanmaa, K, Johnson, M S, Palosuo, T and Kulomaa, M S
(2002) ‘The major conformational IgE-binding epitopes of hevein (Hev b6.02) are
identified by a novel chimera-based allergen epitope mapping strategy’, J Biol
Chem, 277(25), 22656–22661.
118. Stanford, A, Bevan, M and Northcote, D (1989) ‘Differential expression within a
family of novel wound-induced genes in potato’, Mol Gen Genet, 215(2), 200–208.
119. Kauffmann, S, Legrand, M and Fritig, B (1990) ‘Isolation and characterization of
six pathogenesis-related (PR) proteins of Samsun NN tobacco’, Plant Mol Biol,
14(3), 381–390.
120. Linthorst, H J M, Danhash, N, Brederode, F T, Van Kan, J A L, De Wit, P J G M
and Bol, J F (1991) ‘Tobacco and tomato PR proteins homologous to win and pro-
hevein lack the ‘hevein’ domain’, Mol Plant Microbe Inter, 4(6), 586–592.
121. Hanninen, A R, Mikkola, J H, Kalkkinen, N, Turjanmaa, K, Ylitalo, L, Reunala, T
and Palosuo, T (1999) ‘Increased allergen production in turnip (Brassica rapa) by
treatments activating defense mechanisms’, J Allergy Clin Immunol, 104(1), 194–
201.
122. Van Damme, E J, Charels, D, Roy, S, Tierens, K, Barre, A, Martins, J C, Rouge, P,
Van Leuven, F, Does, M and Peumans, W J (1999) ‘A gene encoding a hevein-like
protein from elderberry fruits is homologous to PR-4 and class V chitinase genes’,
Plant Physiol, 119(4), 1547–1556.
123. Dudler, R, Mauch, F and Reimmann, C (1994) Thaumatin-like proteins, in Witty,
M, Higginbotham, J D (eds), Thaumatin, Boca Raton, FL, CRC Press Inc, 93–199.
124. Koiwa, H, Kato, H, Nakatsu, T, Oda, J, Yamada, Y and Sato, F (1999) ‘Crystal
structure of tobacco PR-5d protein at 1.8 A resolution reveals a conserved acidic
cleft structure in antifungal thaumatin-like proteins’, J Mol Biol, 286(4), 1137–
1145.
125. Batalia, M A, Monzingo, A F, Ernst, S, Roberts, W and Robertus, J D (1996) ‘The
crystal structure of the antifungal protein zeamatin, a member of the thaumatin-
like, PR-5 protein family’, Nat Struct Biol, 3(1), 19–23.
126. Ogata, C M, Gordon, P F, de Vos, A M and Kim, S H (1992) ‘Crystal structure of
a sweet tasting protein thaumatin I, at 1.65 A resolution’, J Mol Biol, 228(3), 893–
908.
127. Krebitz, M, Wagner, B, Ferreira, F, Peterbauer, C, Campillo, N, Witty, M, Kolarich,
D, Steinkellner, H, Scheiner, O and Breiteneder, H (2003) ‘Plant-based heterologous
expression of Mal d 2, a thaumatin-like protein and allergen of apple (Malus
domestica), and its characterization as an antifungal protein, J Mol Biol, 329(4),
721–730.
128. Inschlag, C, Hoffmann-Sommergruber, K, O’Riordain, G, Ahorn, H, Ebner, C,
Scheiner, O and Breiteneder, H (1998) ‘Biochemical characterization of Pru a 2, a
23-kD thaumatin-like protein representing a potential major allergen in cherry
(Prunus avium)’, Int Arch Allergy Immunol, 116(1) 22–28.
129. Jensen-Jarolim, E, Santner, B, Leitner, A, Grimm, R, Scheiner, O, Ebner, C and
Breiteneder, H (1998) ‘Bell peppers (Capsicum annuum) express allergens (profilin,
pathogenesis-related protein P23 and Bet v 1) depending on the horticultural strain’,
Int Arch Allergy Immunol, 116(2), 103–109.
130. Fuchs, H C, Hoffmann-Sommergruber, K, Wagner, B, Krebitz, M, Scheiner, O and
Breiteneder, H (2002) ‘Heterologous expression in Nicotiana benthamiana of Cap
a 1, a thaumatin-like protein and major allergen from bell pepper (Capsicum annuum)’,
J Allergy Clin Immunol, 109(S), 134–135.
131. Pastorello, E A, Farioli, L, Pravettoni, V, Ortolani, C, Fortunato, D, Giuffrida, M G,
Perono Garoffo, L, Calamari, A M, Brenna, O and Conti, A (2003) ‘Identification
of grape and wine allergens as an endochitinase 4, a lipid-transfer protein, and a
thaumatin’, J Allergy Clin Immunol, 111(2), 350–359.
132. Gavrovic-Jankulovic, M, Cirkovic, T, Vuckovic, O, Atanaskovic-Markovic, M,
Petersen, A, Gojgic, G, Burazer, L and Jankov, R M (2002) ‘Isolation and biochemical
characterization of a thaumatin-like kiwi allergen’, J Allergy Clin Immunol 110(5),
805–810.
133. Welinder, K G (1992) ‘Superfamily of plant, fungal and bacterial peroxidases’,
Curr Opin Struct Biol, 2(3), 388–393.
134. Van Loon, L C, Pierpoint, W S, Boller, T and Conejero, V (1994) ‘Recommendations
for naming plant pathogenesis-related proteins’, Plant Mol Biol Rep, 12(2), 245–
264.
135. Sánchez-Monge, R, García-Casado, G, López-Otín, C, Armentia, A and Salcedo,
G (1997) ‘Wheat flour peroxidase is a prominent allergen associated with baker’s
asthma’, Clin Exp Allergy, 27(10), 1130–1137.
136. Yamashita, H, Nanba, Y, Onishi, M, Kimoto, M, Hiemori, M and Tsuji H (2002)

‘Identification of a wheat allergen, Tri a Bd 36K, as a peroxidase’, Biosci Biotechnol
Biochem, 66(11), 2487–2490.
137. Weangsripanaval, T, Nomura, N, Moriyama, T, Ohta, N and Ogawa, T (2003)
‘Identification of suberization-associated anionic peroxidase as a possible allergenic
protein from tomato’, Biosci Biotechnol Biochem, 67(6), 1299–1304.
138. Breiteneder, H, Pettenburger, K, Bito, A, Valenta, R, Kraft, D, Rumpold H, Scheiner
O and Breitenbach M (1989) ‘The gene coding for the major birch pollen allergen
Betv1, is highly homologous to a pea disease resistance response gene’, EMBO J,
8(7), 1935–1938.
139. Wen, J, Vanek-Krebitz, M, Hoffmann-Sommergruber, K, Scheiner, O and Breiteneder,
H (1997) ‘The potential of Betv1 homologues, a nuclear multigene family, as
phylogenetic markers in flowering plants’, Mol Phylogenet Evol, 8(3) 317–333.
140. Hoffmann-Sommergruber, K, Ferris, R, Pec, M, Radauer, C, O’Riordain, G, Laimer
Da Camara Machado, M, Scheiner, O and Breiteneder, H (2000) ‘Characterization
of Api g 1.0201, a new member of the Api g 1 family of celery allergens’, Int Arch
Allergy Immunol, 122(2), 115–123.
141. Arquint, O, Helbling, A, Crameri, R, Ferreira, F, Breitenbach, M and Pichler, W J
(1999) ‘Reduced in vivo allergenicity of Bet v 1d isoform, a natural component of
birch pollen’, J Allergy Clin Immunol, 104(6), 1239–1243.
142. Son, D Y, Scheurer, S, Hoffmann, A, Haustein, D and Vieths, S (1999) ‘Pollen-
related food allergy: cloning and immunological analysis of isoforms and mutants
of Mal d 1, the major apple allergen, and Bet v 1, the major birch pollen allergen’,
Eur J Nutr, 38(4), 201–215.
143. Gajhede, M, Osmark, P, Poulsen, F M, Ipsen, H, Larsen, J N, Joost van Neerven,
R J, Schou, C, Lowenstein, H and Spangfort, M D (1996) ‘X-ray and NMR structure
of Bet v 1, the origin of birch pollen allergy’, Nat Struct Biol, 3(12), 1040–1045.
144. Markovic-Housley, Z, Degano, M, Lamba, D, von Roepenack-Lahaye, E, Clemens,
S, Susani, M, Ferreira, F, Scheiner, O and Breiteneder, H (2003) ‘Crystal structure
of a hypoallergenic isoform of the major birch pollen allergen Bet v 1 and its likely
biological function as a plant steroid carrier’, J Mol Biol, 325(1) 123–133.
145. Neudecker, P, Schweimer, K, Nerkamp, J, Scheurer, S, Vieths, S, Sticht, H and
Rosch, P (2001) ‘Allergic cross-reactivity made visible: solution structure of the
major cherry allergen Pru av 1’, J Biol Chem, 276(25), 22756–27763.
146. Vanek-Krebitz, M, Hoffmann-Sommergruber, K, Laimer da Camara Machado, M,
Susani, M, Ebner, C, Kraft, D, Scheiner, O and Breiteneder, H (1995) ‘Cloning and
sequencing of Mal d 1, the major allergen from apple (Malus domestica), and its
immunological relationship to Bet v 1, the major birch pollen allergen’, Biochem
Biophys Res Comm, 214(2), 538–551.
147. Scheurer, S, Metzner, K, Haustein, D and Vieths, S (1997) ‘Molecular cloning,
expression and characterization of Pru a 1, the major cherry allergen’, Mol Immunol,
34(8–9), 619–629.
148. Scheurer, S, Pastorello, E A, Wangorsch, A, Kastner, M, Haustein, D and Vieths,
S (2001) ‘Recombinant allergens Pru av 1 and Pru av 4 and a newly identified lipid
transfer protein in the in vitro diagnosis of cherry allergy’, J Allergy Clin Immunol,
107(4), 724–731.
149. Karamloo, F, Scheurer, S, Wangorsch, A, May, S, Haustein, D and Vieths, S (2001)
‘Pyr c 1, the major allergen from pear (Pyrus communis), is a new member of the
Bet v 1 allergen family’, J Chromatogr B Biomed Sci Appl, 756(1–2), 281–293.
150. Breiteneder, H, Hoffmann-Sommergruber, K, O’Riordain, G, Susani M, Ahorn, H,
Ebner, C, Kraft, D and Scheiner, O (1995) ‘Molecular characterization of Api g 1,
the major allergen of celery (Apium graveolens) and its immunological and structural
relationships to a group of 17-kDa tree pollen allergens’, Eur J Biochem, 233(2),
484–489.
151. Hoffmann-Sommergruber, K, O’Riordain, G, Ahorn, H, Ebner, C, Laimer da Camara

Machado, M, Pühringer, H, Scheiner, O and Breiteneder, H (1999) ‘Molecular
characterization of Dau c 1, the Bet v 1 homologous protein from carrot and its
cross-reactivity with Bet v 1 and Api g 1’, Clin Exp Allergy, 29(6), 840–847.
152. Hoffmann-Sommergruber, K, Vanek-Krebitz, M, Radauer, C, Wen, J, Ferreira, F,
Scheiner, O and Breitenender, H (1997) ‘Genomic characterization of members of
the Bet v 1 family: genes coding for allergens and pathogenesis-related proteins
share intron positions’, Gene, 197(1–2), 91–100.
153. Luttkopf, D, Muller, U, Skov, P S, Ballmer-Weber, B K, Wüthrich, B, Skamstrup
Hansen, K, Poulsen, L K, Kastner, M, Haustein, D and Vieths, S (2002) ‘Comparison
of four variants of a major allergen in hazelnut (Corylus avellana) Cor a 1.04 with
the major hazel pollen allergen Cor a 1.01’, Mol Immunol, 38(7), 515–525.
154. Kleine-Tebbe, J, Vogel, L, Crowell, D N, Haustein, U F and Vieths, S (2002)
‘Severe oral allergy syndrome and anaphylactic reactions caused by a Bet v 1-
related PR-10 protein in soybean, SAM22’, J Allergy Clin Immunol, 110(5), 797–
804.
155. Valenta, R, Duchene, M, Pettenburger, K, Sillaber, C, Valent, P, Bettelheim, P,
Breitenbach, M, Rumpold, H, Kraft, D and Scheiner, O (1991) ‘Identification of
profilin as a novel pollen allergen; IgE autoreactivity in sensitized individuals’,
Science, 253(5019), 557–560.
156. Valenta, R, Duchene, M, Ebner, C, Valent, P, Sillaber, C, Deviller, P, Ferreira, F,
Tejkl, M, Edelmann, H, Kraft, D and Scheiner, O (1992) ‘Profilins constitute a
novel family of functional plant pan-allergens’, J Exp Med, 175(2), 377–385.
157. Ebner, C, Jensen-Jarolim, E, Leitner, A and Breitender, H (1998) ‘Characterization
of allergens in plant-derived spices: Apiaceae spices, pepper (Piperaceae), and
paprika (bell peppers, Solanaceae)’, Allergy, 53(46 Suppl), 52–54.
158. Hirschwehr, R, Valenta, R, Ebner, C, Ferreira, F, Sperr, W R, Valent, P, Rohac, M,
Rumpold, H, Scheiner, O and Kraft, D (1992) ‘Identification of common allergenic
structures in hazel pollen and hazelnuts: a possible explanation for sensitivity to
hazelnuts in patients allergic to tree pollen’, J Allergy Clin Immunol, 90(6Pt1),
927–936.
159. Van Ree, R, Voitenko, V, van Leeuwen, W A and Aalberse, R C, (1992) ‘Profilin
is a cross-reactive allergen in pollen and vegetable foods’, Int Arch Allergy Immunol,
98(2), 97–104.
160. Fah, J, Wüthrich, B and Vieths, S (1995) ‘Anaphylactic reaction to lychee fruit:
evidence for sensitization to profilin’, Clin Exp Allergy, 25(10), 1018–1023.
161. Petersen, A, Vieths, S, Aulepp, H, Schlaak, M and Becker, W M (1996) ‘Ubiquitous
structures responsible for IgE cross-reactivity between tomato fruit and grass pollen
allergens’, J Allergy Clin Immunol, 98(4), 805–815.
162. Fritsch, R, Ebner, H, Kraft, D, Ebner, C (1997) ‘Food allergy to pumpkin seed –
characterization of allergens’, Allergy, 52(3), 335–337.
163. Rihs, H P, Chen, Z, Rueff, F, Petersen, A, Rozynek, P, Heimann, H and Baur, X
(1999) ‘IgE binding of the recombinant allergen soybean profilin (rGly m 3) is
mediated by conformational epitopes’, J Allergy Clin Immunol, 104(6), 1293–
1301.
164. Scheurer, S, Wangorsch, A, Haustein, D and Vieths, S (2000) ‘Cloning of the minor
allergen Api g 4 profilin from celery (Apium graveolens) and its cross-reactivity
with birch pollen profilin Bet v 2’, Clin Exp Allergy, 30(7), 962–971.
165. Scheurer, S, Wangorsch, A, Nerkamp, J, Skov, P S, Ballmer-Weber, B, Wuthrich,
B, Haustein, D and Vieths, S (2001) Cross-reactivity within the profilin panallergen
family investigated by comparison of recombinant profilins from pear (Pyr c 4),
cherry (Pru av 4) and celery (Api g 4) with birch pollen profilin Bet v 2’, J
Chromatogr B Biomed Sci Appl, 756(1–2), 315–325.
166. Reindl, J, Rihs, H P, Scheurer, S, Wangorsch, A, Haustein, D and Vieths, S (2002)
‘IgE reactivity to profilin in pollen-sensitized subjects with adverse reactions to

banana and pineapple’, Int Arch Allergy Immunol, 128(2), 105–114.
167. Laskowski, M Jr and Kato, I (1980) ‘Protein inhibitors of proteinases’, Ann Rev
Biochem 49, 593–626.
168. Mazumdar-Leighton, S and Broadway, R M (2001) ‘Identification of six chymotrypsin
cDNAs from larval midguts of Helicoverpa zea and Agrotis ipsilon feeding on the
soybean (Kunitz) trypsin inhibitor’, Insect Biochem Mol Biol 31(6–7), 633–644.
169. Burks, A W, Cockrell, G, Connaughton, C, Guin, J, Allen, W and Helm, R M
(1994) ‘Identification of peanut agglutinin and soybean trypsin inhibitor as minor
legume allergens’, Int Arch Allergy Immunol, 105(2), 143–149.
170. Moroz, L A and Lang, W H (1980) ‘Kunitz soybean trypsin inhibitor. A specific
allergen in food anaphylaxis’, New Engl J Med, 302(20), 1126–1128.
171. Seppala, U, Majamaa, H, Turjanmaa, K, Helin, J, Reunala, T, Kalkkinen, N and
Palosuo, T (2001) ‘Identification of four novel potato (Solanum tuberosum) allergens
belonging to the family of soybean trypsin inhibitors’, Allergy, 56(7), 619–626.
172. Shibasaki, M, Sumazaki, R, Isoyama, S and Takita, H (1992) ‘Interaction of lectins
with human IgE: IgE-binding property and histamine-releasing activity of twelve
plant lectins’, Int Arch Allergy Immunol, 98(1), 18–25.
173. Burks, A W, Cockrell, G, Connaughton, C, Guin, J, Allen, W and Helm, R M
(1994) ‘Identification of peanut agglutinin and soybean trypsin inhibitor as minor
legume allergens, Int Arch Allergy Immunol, 105(2), 143–149.
174. Seppala, U, Alenius, H, Turjanmaa, K, Reunala, T, Palosuo, T and Kalkkinen, N
(1999) ‘Identification of patatin as a novel allergen for children with positive skin
prick test responses to raw potato’, J Allergy Clin Immunol, 103(1Pt1), 165–171.
175. Karamloo, F, Wangorsch, A, Kasahara, H, Davin, L B, Haustein, D, Lewis, N G
and Vieths, S (2001) ‘Phenylcoumaran benzylic ether and isoflavonoid reductases
are a new class of cross-reactive allergens in birch pollen, fruits and vegetables’,
Eur J Biochem, 268(20), 5310–5320.
176. Pons, L, Chery, C, Romano, A, Namour, F, Artesani, M C and Gueant, J L (2002)
‘The 18 kDa peanut oleosin is a candidate allergen for IgE-mediated reactions to
peanuts’, Allergy, 2002 57(Suppl 72), 88–93.
177. Westphal, S, Kolarich, D, Foetisch, K, Lauer, I, Altmann, F, Conti A, Crespo, J F,
Rodriguez, J, Enrique, E, Vieths, S and Scheurer, S (2003) ‘Molecular characterization
and allergenic activity of Lyc e 2 (beta-fructofuranosidase), a glycosylated allergen
of tomato’, Eur J Biochem, 270(6), 1327–1337.
178. Cuesta-Herranz, J, Pastor, C, Figueredo, E, Vidarte, L, De las Heras, M, Duran, C,
Fernandez-Caldas, E, de Miguel, J and Vivanco, F (2003) ‘Identification of cucumisin
(Cuc m 1), a subtilisin-like endopeptidase, as the major allergen of melon fruit’,
179. Bublin, M, Radauer, C, Wilson, I B H, Kraft, D, Scheiner, O, Breiteneder, H and
Hoffmann-Sommergruber, K (2003) ‘Cross-reactive N-glycans of Api g 5, a high
molecular weight glycoprotein allergen from celery, are required for immunoglobulin
E binding and activation of effector cells from allergic patients’, FASEB J, 17(12),
1697–1699.
180. Lehrer, S B and Reese, G (1997) ‘Recombinant proteins in newly developed foods:
identification of allergenic activity’, Int Arch Allergy Immunol 113(1–3) 122–124.
181. Sampson, H A (2000) ‘Food anaphylaxis’, Br Med Bull 56(4), 925–935.
182. Reese, G, Ayuso, R and Lehrer, S B (1999) ‘Tropomyosin: an invertebrate pan-
allergen’, Int Arch Allergy Immunol 119(4), 247–258.
183. Sicherer, S H (2001) ‘Clinical implications of cross-reactive food allergens’, J
Allergy Clin Immunol 108(6), 881–890.
184. Bugajska-Schretter, A, Grote, M, Vangelista, L, Valent, P, Sperr, W R, Rumpold, H,
Pastore, A, Reichelt, R, Valenta, R and Spitzauer, S (2000) ‘Purification, biochemical,
and immunological characterisation of a major food allergen: different
immunoglobulin E recognition of the apo- and calcium-bound forms of carp

parvalbumin’, Gut 46(5), 661–669.
185. Host, A (2002), ‘Frequency of cow’s milk allergy in childhood’, Ann Allergy Asthma
Immunol 89(6 Suppl 1), 33–37.
186. Bordin, G, Cordeiro Raposo, F, de la Calle, B and Rodriguez, A R (2001)
‘Identification and quantification of major bovine milk proteins by liquid
chromatography’, J Chromatogr A, 928(1), 63–76.
187. Werfel, S J, Cooke, S K and Sampson, H A (1997) ‘Clinical reactivity to beef in
children allergic to cow’s milk’, J Allergy Clin Immunol, 99(3), 293–300.
188. Sampson, H A (1997) ‘Food sensitivity and the pathogenesis of atopic dermatitis’,
J R Soc Med 90(Suppl 30), 2–8.
189 Cooke, S K and Sampson, H A (1997) ‘Allergenic properties of ovomucoid in
man’, J Immunol, 159(4), 2026–2032.
190. Bernhisel-Broadbent, J, Dintzis, H M, Dintzis, R Z and Sampson, H A (1994)
‘Allergenicity and antigenicity of chicken egg ovomucoid (Gal d III) compared
with ovalbumin (Gal d I) in children with egg allergy and in mice’, J Allergy Clin
Immunol 93(6), 1047–1059.
191. Langland, T (1983) ‘A clinical and immunological study of allergy to hen’s egg
white. VI. Occurrence of proteins cross-reacting with allergens in hen’s egg white
as studied in egg white from turkey, duck, goose, seagull, and in hen egg yolk, and
hen and chicken sera and flesh’, Allergy, 38(7), 399–412.
192. Anibarro, B, Seoane, F J, Vila, C and Lombardero, M (2000) ‘Allergy to eggs from
duck and goose without sensitization to hen egg proteins’, J Allergy Clin Immunol,
105(4), 834–836.
193. Szepfalusi, Z, Ebner, C, Pandjaitan, R, Orlicek, F, Scheiner, O, Boltz-Nitulescu, G,
Kraft, D and Ebner, H (1994) ‘Egg yolk alpha-livetin (chicken serum albumin) is
a cross-reactive allergen in the bird-egg syndrome’, J Allergy Clin Immunol, 93(5),
932–942.
194. Bausela, B A, Garcia-Ara, M C, Martin Esteban, M, Boyano Martinez, T B, Diaz
Pena, J M and Ojeda Casas, J A (1997) ‘Peculiarities of egg allergy in children
with bird protein sensitization’, Ann Allergy Asthma Immunol, 78(2), 213–216.
195. Kelso, J M, Cockrell, G E, Helm, R M and Burks, A W (1999) ‘Common allergens
in avian meat’, J Allergy Clin Immunol, 104(1), 202–204.
196. Lewit-Bentley, A and Rety, S (2000) ‘EF-hand calcium-binding proteins’, Curr
Opin Struct Biol 10(6), 637–643.
197. Pauls, T L, Cox, J A and Berchtold, M W (1996) ‘The Ca2+(–)binding proteins
parvalbumin and oncomodulin and their genes: new structural and functional findings’,
Biochim Biophys Acta, 1306(1), 39–54.
198. Elsayed, S and Apold, J (1983) ‘Immunochemical analysis of cod fish allergen M:
locations of the immunoglobulin binding sites as demonstrated by the native and
synthetic peptides’, Allergy, 38(7), 449–459.
199. Van Do, T, Hordvik, I, Endresen, C and Elsayed, S (2003) ‘The major allergen
(parvalbumin) of codfish is encoded by at least two isotypic genes: cDNA cloning,
expression and antibody binding of the recombinant allergens’, Mol Immunol,
39(10), 595–602.
200. Das Dores, S, Chopin, C, Villaume, C, Fleurence, J and Gueant, J L (2002) ‘A new
oligomeric parvalbumin allergen of Atlantic cod (Gad m I) encoded by a gene
distinct from that of Gad c I’, Allergy, 57(Suppl72), 79–83.
201. Lindstrom, C D, van Do, T, Hordvik, I, Endresen, C and Elsayed, S (1996) ‘Cloning
of two distinct cDNAs encoding parvalbumin, the major allergen of Atlantic salmon
(Salmo salar)’, Scand J Immunol, 44(4), 335–344.
202. Swoboda, I, Bugajska-Schretter, A, Verdino, P, Keller, W, Sperr, W R, Valent, P,
Valenta, R and Spitzauer, S (2002) ‘Recombinant carp parvalbumin, the major
cross-reactive fish allergen: a tool for diagnosis and therapy of fish allergy’, J
Immunol, 168(9), 4576–4584.
203. Yamada, S, Nolte, H and Zychlinsky, E (1999) ‘Identification and characterization

of allergens in two species of tuna fish’, Ann Allergy Asthma Immunol, 82(4), 395–
400.
204. Hilger, C, Grigioni, F, Thill, L, Mertens, L and Hentges, F (2002) ‘Severe IgE-
mediated anaphylaxis following consumption of fried frog legs: definition of alpha-
parvalbumin as the allergen in cause’, Allergy 57(11), 1053–1058.
205. Hamada, Y, Tanaka, H, Ishizaki, S, Ishida, M, Nagashima, Y and Shiomi, K (2003)
‘Purification, reactivity with IgE and cDNA cloning of parvalbumin as the major
allergen of mackerels’, Food Chem Toxicol, 41(8), 1149–1156.
206. MacLeod, A R (1987) ‘Genetic origin of diversity of human cytoskeletal
tropomyosins’, Bioessays 6(5), 208–212.
207. Shanti, K N, Martin, B M, Nagpal, S, Metcalfe, D D and Rao, P V (1993)
‘Identification of tropomyosin as the major shrimp allergen and characterization of
its IgE-binding epitopes’, J Immunol, 151(10), 5354–63.
208. Lin, R Y, Shen, H D and Han, S H (1993) ‘Identification and characterization of a
30 kd major allergen from Parapenaeus fissurus’, J Allergy Clin Immunol, 92(6),
837–845.
209. Daul, C B, Slattery, M, Reese, G and Lehrer, S B (1994) ‘Identification of the
major brown shrimp (Penaeus aztecus) allergen as the muscle protein tropomyosin’,
Int Arch Allergy Immunol, 105(1), 49–55.
210. Leung, P S, Chu, K H, Chow, W K, Ansari, A, Bandea, C I, Kwan, H S, Nagy, S
M and Gershwin, M E (1994) ‘Cloning, expression, and primary structure of
Metapenaeus ensis tropomyosin, the major heat-stable shrimp allergen’, J Allergy
Clin Immunol, 94(5), 882–890.
211. Leung, P S, Chen, Y C, Gershwin, M E, Wong, S H, Kwan, H S and Chu, K H
(1998) ‘Identification and molecular characterization of Charybdis feriatus
tropomyosin, the major crab allergen’, J Allergy Clin Immunol, 102(5), 847–852.
212. Leung, P S, Chen, Y C, Mykles, D L, Chow, W K, Li, C P and Chu, K H (1998)
‘Molecular identification of the lobster muscle protein tropomyosin as a seafood
allergen’, Mol Mar Biol Biotechnol, 7(1), 12–20.
213. Mykles, D L, Cotton, J L, Taniguchi, H, Sano, K and Maeda, Y (1998) ‘Cloning of
tropomyosins from lobster (Homarus americanus) striated muscles: fast and slow
isoforms may be generated from the same transcript’, J Muscle Res Cell Motil,
19(2), 105–115.
214. Ishikawa, M, Ishida, M, Shimakura, K, Nagashima, Y and Shiomi, K (1998)
‘Tropomyosin, the major oyster Crassostrea gigas allergen and its IgE-binding
epitopes’, J Food Sci, 63 44–47.
215. Leung, P S and Chu K H (2001) ‘cDNA cloning and molecular identification of the
major oyster allergen from the Pacific oyster Crassostrea gigas’, Clin Exp Allergy,
31(8), 1287–1294.
216. Ishikawa, M, Nagashima, Y and Shiomi, K (1998) ‘Identification of the oyster
allergen Cra g 2 as tropomyosin’, Fish Sci, 64(5), 854–855.
217. Ishikawa, M, Ishida, M, Shimakura, K, Nagashima, Y and Shiomi, K (1998)
‘Purification and IgE-binding epitopes of a major allergen in the gastropod Turbo
cornutus’, Biosci Biotechnol Biochem, 62(7), 1337–1343.
218. Miyazawa, H, Fukamachi, H, Inagaki, Y, Reese, G, Daul, C B, Lehrer, S B, Inouye,
S and Sakaguchi, M (1996) ‘Identification of the first major allergen of a squid
(Todarodes pacificus)’, J Allergy Clin Immunol, 98(5Pt1), 948–953.
219. Chu, K H, Wong, S H and Leung, P S (2000) ‘Tropomyosin is the major mollusk
allergen: reverse transcriptase polymerase chain reaction, expression and IgE
reactivity’, Mar Biotechnol, 2(5), 499–509.
220. Asturias, J A, Eraso, E, Arilla, M C, Gomez-Bayon, N, Inacio, F and Martinez, A
(2002) ‘Cloning, isolation, and IgE-binding properties of Helix aspersa (brown
garden snail) tropomyosin’, Int Arch Allergy Immunol 128(2), 90–96.
221. Chu, K H, Wong, S H and Leung, P S (2000) ‘Tropomyosin is the major mollusk
allergen: reverse transcriptase polymerase chain reaction, expression and IgE
reactivity’, Mar Biotechnol (NY), 2(5), 499–509.
222. Strong, S J and Ellington, W R (1995) ‘Isolation and sequence analysis of the gene
for arginine kinase from the chelicerate arthropod, Limulus polyphemus: insights
into catalytically important residues’, Biochim Biophys Acta 1246(2), 197–200.
223. Lin, R Y, Shen, H and Han, S H (1993 ‘Identification and characterization of a 30
kd major allergen from Parapenaeus fissurus’, J Allergy Clin Immunol 92(6), 837–
845.
224. Yu, C J, Lin, Y F, Chiang, B L and Chow, L P (2003) ‘Proteomics and immunological
analysis of a novel shrimp allergen, Pen m 2’, J Immunol 170(1), 445–453.
225. Henrissat, B and Bairoch, A (1993) ‘New families in the classification of glycosyl
hydrolases based on amino acid sequence similarities’, Biochem J 293 (Pt 3), 781–
788.
226. Henrissat, B and Bairoch, A (1996) ‘Updating the sequence-based classification of
glycosyl hydrolases’, Biochem J 316 (Pt 2), 695–696.
227. Shewale, J G, Sinha, S K and Brew, K (1984) ‘Evolution of alpha-lactalbumins.
The complete amino acid sequence of the alpha-lactalbumin from a marsupial
(Macropus rufogriseus) and corrections to regions of sequence in bovine and goat
alpha-lactalbumins’, J Biol Chem 259(8), 4947–4956.
228. Brew, K, Castellino, F J, Vanaman, T C and Hill, R L (1970) ‘The complete amino
acid sequence of bovine alpha-lactalbumin’, J Biol Chem 245(17), 4570–4582.
229. Nitta, K and Sugai, S (1989) ‘The evolution of lysozyme and alpha-lactalbumin’,
Eur J Biochem 182(1), 111–118.
230. Anet, J, Back, J F, Baker, R S, Barnett, D, Burley, R W and Howden, M E (1985)
‘Allergens in the white and yolk of hen’s egg. A study of IgE binding by egg
proteins’, Int Arch Allergy Appl Immunol 77(3), 364–371.
231. Pervaiz, S and Brew, K (1985) ‘Homology of beta-lactoglobulin, serum retinol-
binding protein, and protein HC’, Science 228(4697), 335–337.
232. Jamieson, A C, Vandeyar, M A, Kang, Y C, Kinsella, J E and Batt, C A (1987)
‘Cloning and nucleotide sequence of the bovine beta-lactoglobulin gene’, Gene
61(1), 85–90.
233. Brownlow, S, Morais Cabral, J H, Cooper, R, Flower, D R, Yewdall, S J, Polikarpov,
I, North, A C and Sawyer, L (1997) ‘Bovine beta-lactoglobulin at 1.8 A resolution
– still an enigmatic lipocalin’, Structure 5(4), 481–495.
234. He, X M and Carter, D C (1992) ‘Atomic structure and chemistry of human serum
albumin’, Nature 358(6383), 209–215.
235. Spuergin, P, Walter, M, Schiltz, E, Deichmann, K, Forster, J and Mueller, H (1997)
‘Allergenicity of alpha-caseins from cow, sheep, and goat’, Allergy, 52(3), 293–
298.
236. Bellioni-Businco, B, Paganelli, R, Lucenti, P, Giampietro, P G, Perborn, H and
Businco, L (1999) ‘Allergenicity of goat’s milk in children with cow’s milk allergy’,
237. Quirce, S, Maranon, F, Umpierrez, A, de las Heras, M, Fernandez-Caldas, E and
Sastre, J (2001) ‘Chicken serum albumin (Gal d 5) is a partially heat-labile
inhalant and food allergen implicated in the bird-egg syndrome’, Allergy 56(8),
754–762.
238. Ayuso, R, Lehrer, S B, Lopez, M, Reese, G, Ibanez, M D, Esteban, M M, Ownby,
D R and Schwartz, H (2000) ‘Identification of bovine IgG as a major cross-reactive
vertebrate meat allergen’, Allergy 55(4), 348–354.
239. Holt, C and Sawyer, L (1988) ‘Primary and predicted secondary structures of the
caseins in relation to their biological functions’, Protein Eng 2(4), 251–259.
240. Wal, J M (2001) ‘Structure and function of milk allergens’, Allergy, 56(Suppl67),
35–38.
241. Wal, J M (1998) ‘Cow’s milk allergens’, Allergy, 53(11), 1013–1022.

242. Restani, P, Gaiaschi, A, Plebani, A, Beretta, B, Cavagni, G, Fiocchi, A, Poiesi, C,
Velona, T, Ugazio, A G and Galli, C L (1999) ‘Cross-reactivity between milk
proteins from different animal species’, Clin Exp Allergy, 29(7), 997–1004.
243. Businco, L, Giampietro, P G, Lucenti, P, Lucaroni, F, Pini, C, Di Felice, G, Iacovacci,
P, Curadi, C and Orlandi, M (2000) ‘Allergenicity of mare’s milk in children with
cow’s milk allergy’, J Allergy Clin Immunol, 105(5), 1031–1034.
244. Sharma, S, Kumar, P, Betzel, C and Singh, T P (2001) ‘Structure and function of
proteins involved in milk allergies’, J Chromatogr B Biomed Sci Appl 756(1–2),
183–187.
245. Holen, E and Elsayed, S (1990) ‘Characterization of four major allergens of hen
egg-white by IEF/SDS-PAGE combined with electrophoretic transfer and IgE-
immunoautoradiography’, Int Arch Allergy Appl Immunol 91(2), 136–141.
246. Laskowski, M, Jr, Kato, I, Ardelt, W, Cook, J, Denton, A, Empie, M W, Kohr, W
J, Park, S J, Parks, K, Schatzley, B L, Schoenberger, O L, Tashiro, M, Vichot, G,
Whatley, H A, Wieczorek, A and Wieczorek, M (1987) ‘Ovomucoid third domains
from 100 avian species: isolation, sequences, and hypervariability of enzyme-
inhibitor contact residues; Biochemistry, 26(1), 202–221.
247. Williamson, M P, Marion, D and Wuthrich, K (1984) ‘Secondary structure in the
solution conformation of the proteinase inhibitor IIA from bull seminal plasma by
nuclear magnetic resonance’, J Mol Biol 173(3), 341–359.
248. Kato, I, Schrode, J, Kohr, W J and Laskowski, M, Jr (1987) ‘Chicken ovomucoid:
determination of its amino acid sequence, determination of the trypsin reactive
site, and preparation of all three of its domains’, Biochemistry 26(1), 193–201.
249. Hunt, L T and Dayhoff, M O (1980) ‘A surprising new protein superfamily containing
ovalbumin, antithrombin-III, and alpha 1-proteinase inhibitor’, Biochem Biophys
Res Commun 95(2), 864–871.
250. Remold-O’Donnell, E (1993) ‘The ovalbumin family of serpin proteins’, FEBS
Lett 315(2), 105–108.
251. Tada, Y, Nakase, M, Adachi, T, Nakamura, R, Shimada, H, Takahashi, M, Fujimura,
T and Matsuda, T (1996) ‘Reduction of 14–16 kDa allergenic proteins in transgenic
rice plants by antisense gene’, FEBS Lett, 391(3), 341–345.
252. Herman, E M, Helm, R M, Jung, R and Kinney, A J (2003) ‘Genetic modification
removes an immunodominant allergen from soybean’, Plant Physiol, 132(1), 36–
43.
253. Vieths, S, Hoffmann, A, Holzhauser, T, Muller, U, Reindl, J and Haustein, D
(1998) ‘Factors influencing the quality of food extracts for in vitro and in vivo
diagnosis’, Allergy, 53(46 Suppl), 65–71.
254. Ballmer-Weber, B K, Hoffmann, A, Wüthrich, B, Lüttkopf, D, Pompei, C, Wangorsch,
A, Kastner, M and Vieths, S (2002) ‘Influence of food processing on the allergenicity
of celery: DBPCFC with celery spice and allergy’, Allergy, 57(3), 228–235.
255. Asero, R, Mistrello, G, Roncarolo, D, de Vries, S C, Gautier, M F, Ciurana, C L,
Verbeek, E, Mohammadi, T, Knul-Brettlova, V, Akkerdaas, J H, Bulder, I, Aalberse,
R C and van Ree, R (2000) ‘Lipid transfer protein: a pan-allergen in plant-derived
foods that is highly resistant to pepsin digestion’, Int Arch Allergy Immunol, 122(1),
20–32.
256. Sen, M, Kopper, R, Pons, L, Abraham, E C, Burks, A W and Bannon, G A (2002)
‘Protein structure plays a critical role in peanut allergen stability and may determine
immunodominant IgE-binding epitopes’, J Immunol, 169(2), 882–887.
257. Ferreira, F, Wallner, M, Breiteneder, H, Hartl, A, Thalhamer, J and Ebner, C (2002)
‘Genetic engineering of allergens: future therapeutic products’, Int Arch Allergy
Immunol 128(3), 171–178.
258. Bannon, G A, Cockrell, G, Connaughton, C, West, C M, Helm, R, Stanley, J S,
King, N, Rabjohn, P, Sampson, H A and Burks, A W (2001) ‘Engineering,
characterization and in vitro efficacy of the major peanut allergens for use in
immunotherapy’, Int Arch Allergy Immunol, 124(1–3), 70–72.
259. Rabjohn, P, West, C M, Connaughton, C, Sampson, H A, Helm, R M, Burks, A W
and Bannon, G A (2002) ‘Modification of peanut allergen Ara h 3: effects on IgE
binding and T cell stimulation’, Int Arch Allergy Immunol, 128(1), 15–23.
260. Cocco, R R, Jarvinen, K M, Sampson, H A and Beyer, K (2003) ‘Mutational
analysis of major, sequential IgE-binding epitopes in alpha-s1-casein, a major
cow’s milk allergen’, J Allergy Clin Immunol, 112(2), 433–437.
261. Valenta, R and Kraft, D (2002) ‘From allergen structure to new forms of allergen-
specific immunotherapy’, Curr Opin Immunol, 14(6) 718–727.
262. Brusic, V, Millot, M, Petrovsky, N, Gendel, S M, Gigonzac, O and Stelman, S J
(2003) ‘Allergen databases’, Allergy 58(11), 1093–1100.
Part II
Types of detection method

3
Antibodies
S. Hefle, University of Nebraska, USA, J. Yeung, Food
Products Association, USA and R. Helm, University of Arkansas,
USA
3.1 Nature of antibodies

The basic role of the humoral immune system is to eliminate foreign invaders.
This is accomplished through the production of antibodies, proteins produced
by the immune system in response to the presence of a foreign antigen. In
order for this to occur, it is necessary that the immune system recognizes a
large range of non-self antigens but responds well to foreign antigens while
not over-responding to self or switching to IgE responses (allergy).
Antibodies can be divided into five classes: IgG, IgM, IgA, IgD and IgE,
based on the number of Y units and the type of heavy chain. IgG is most
often used in immunoassays. The basic structure of an antibody
(immunoglobulin) was discovered by Porter in 1959 (Porter, 1959); using
the enzyme papain, he isolated one fragment from rabbit 7S immunoglobulin
that was subsequently crystallized and named Fc, for ‘fragment crystallizable’.
The two remaining fragments (which bound antigen) were designated Fab,
or ‘fragment antibody binding’. Based on these initial studies, in this same
research, Porter (1962) described a four-chain structural model for
immunoglobulins; antibodies exist as one or more copies of a Y-shaped unit,
and the typical structure is as shown in Fig. 3.1. Each Y contains two identical
copies of a ‘heavy chain’ (about 50 kD each), and two identical copies of a
‘light chain’ (about 25 kD each), named for their relative molecular weights.
The four chains are held together by non-covalent forces and disulfide bonds.
Both heavy and light chains contain peptide sequences which show little
dissimilarity between antibodies, and which are termed the ‘constant’ regions;
other regions vary considerably among individual immunoglobulins, and
thus are called the ‘variable’ regions. The Fc region comprises the ‘tail’ of
Antigen binding site
F(ab′)2
Fab
Papain
Pepsin
Fc
Fc′
Fig. 3.1 Antibody (immunoglobulin) structure showing heavy and light chains.
Papain and pepsin cleavage sites are also indicated. Digestion by papain gives Fc and
Fab fragments. Digestion by pepsin separates the molecule into two parts, a bivalent
fraction called F(ab′)2 that contains both antigen binding sites, but also the Fc into
small parts called Fc′.
the antibody and plays a role in immune response, as well as serving as a

useful ‘handle’ for manipulating the antibody during some immunochemical
procedures. Digestion by pepsin separates IgG into two parts, a bivalent
fragment that contains both antigen binding sites, F(ab′)2, but also the Fc
into small parts called Fc′.
Antibodies are produced by plasma B lymphocytes (B cells). They are
flexible molecules that on the specific ends can bind their antigens and on
the other end can bind to effector cells (phagocytes) of the immune system;
the antigen–antibody complex is then removed by phagocytosis, thereby
eliminating invaders (Deshpande, 1996a). The specific ends display a great
deal of flexibility and thus can operate independently: each monomeric antibody
unit contains two antigen-binding sites and is thus bivalent. The primary job
of immunoglobulins is to bind antigen, but they can have other effects depending
on the type of antibody involved, and these can include complement activation,
placental transfer, reactivity with staphylococcal proteins, and other functions
(Deshpande, 1996a).
Antigen–antibody interactions occur through multiple non-covalent bonds
between the antigen and the amino acids in the binding site of the antibody,
including van der Waals forces, electrostatic attractions, and hydrophobic
and hydrogen bonds. Individually, these forces are weak compared to covalent
bonds (Deshpande, 1996b). However, together they form a tight bond, leading
to stabilization. The strength of this bond is called the antibody’s ‘affinity’
(Harlow and Lane, 1999). Typical affinity constants are 10–5 to 10–12 M
(Feldkamp and Carey, 1996; Harlow and Lane, 1999).
Antibodies 67
Hydrogen bonding is the result of hydrogen bridges being created between

atoms, and does not often play a role in the primary binding of the antigen
with the antibody (Ag–Ab binding), although it does significantly assist in
secondary binding (Jefferis and Deverill, 1991). Electrostatic forces consist
of the attraction between oppositely-charged groups on proteins, and cause
primary Ag–Ab binding (Davies and Padlan, 1990). Van der Waals forces
occur by the interaction of electron clouds of two polar groups and contribute
to primary Ag–Ab binding. Hydrophobic interactions are based on the
association of non-polar, hydrophobic regions or side chains of amino acids.
They have a high stability owing to structural alteration of the aqueous
environment when these groups come into contact, excluding water (Deshpande,
1996b). Hydrophobic interactions strengthen existing primary and secondary
Ag–Ab bonds, and the act of excluding water from the bond increases the
binding energy considerably, serving to further stabilize the bond. While
both electrostatic and van der Waals forces are factors in primary Ag–Ab
bonding in the early phases (first few milliseconds), they contribute only a
small portion of the total (Jefferis and Deverill, 1991). Secondary Ag–Ab
binding involves the creation of hydrogen bonds and hydrophobic interactions,
continuing over a longer time-frame, and contributes significantly to the
stabilization of the Ag–Ab bond. The overall stability of the Ag–Ab complexes
is referred to as ‘avidity’ (Harlow and Lane, 1999).
For forces to become substantial, the interacting moieties must be in close
proximity to each other. Compared to the attractive forces described above,
steric repulsion is much more sensitive (Steward and Steensgaard, 1983).
Repulsive influences between non-bonded atoms come from the
interpenetration of their electron clouds, and determine the level of ‘fit’
between antigen and antibody. A better ‘fit’ of the electron clouds between
the antigenic determinant and the binding site of the antibody will lower the
repulsive force, and vice versa. Non-complementarity will result in high
repulsive forces, and minimize any small forces of attraction, resulting in
poor ‘fit’. Because of the varied nature of the forces involved in Ag–Ab
bonding, electrostatic forces are very dependent on the pH. Outside the
range of pH 6–8, Ag–Ab binding can be affected adversely (Hughes-Jones et
al., 1964). Temperature and ionic strength can also affect binding in a negative
way (Deshpande, 1996b).
There are two types of antibodies used in immunoassays; monoclonal and
polyclonal. The productions of both types is discussed at length later in this
chapter; however, basic definitions are made here. Polyclonal antibodies are
produced in animals in response to injections of a substance that is capable
of inducing the immune system (antigen) to produce antibodies. Since several
antibody-producing B cell clones in the immunized animals synthesize these
polyclonal antibodies, polyclonal antibodies recognize a variety of specific
binding sites (epitopes) on the antigen. Monoclonal antibodies are usually
produced using rats or mice, and are created by the fusion of immunized
spleen cells and myeloma cells; the spleen cells convey antibody secretion
and the myeloma cells impart immortality. Since monoclonal antibodies

allow for individual B cells to combine with a myeloma cell, antibody very
specific for a certain part of the antigen can be selected and produced.
3.2 Immunogens and antigens

Traditionally, the term ‘antigen’ was used to describe any molecule that
produced a specific immune response such as antibody production, tolerance,
or cell-mediated immunity (Deshpande, 1996a). Nevertheless, to differentiate
molecules that make an immune response and those that are antibody-binding
targets, ‘immunogen’ is used for the former and ‘antigen’ for the latter (Catty,
1988). Molecules that can act as immunogens are normally > 1 kD and
usually > 5 kD in molecular mass. Antibodies do not recognize the whole
antigen, merely portions of it. These portions are called ‘epitopes’ or ‘antigenic
determinants’. Each individual antibody is specific for a particular antigenic
determinant of the antigen. Even when a single pure protein is used as an
immunogen, the resulting polyclonal antibodies may be heterogeneous and
possess a range of different affinities (Hawcroft et al., 1987). Antibodies are
produced by injection of immunogens that are different from self antigens.
For allergens, antibodies can be produced against any protein or peptide that
has at least one antigenic determinant (generally 4–6 amino acids) (Deshpande,
1996a). Like small chemicals, synthetic peptides can serve as haptens with
carrier proteins as immunogens. Once an immune response to an antigen is
made, the plasma B cells make large amounts of antibody. For proteins, T
cells must also be involved in order to get antibody response, and for these
T-dependent responses, at least two antigenic determinants are required (one
for the T cells and one for the B cells) (Raff, 1970).
Because antibodies themselves can act as immunogens, ‘second antibodies’
– or, antibodies that detect antibodies – can be produced. Second antibodies
are made against specific sections of the antibody of interest and they bind
without destroying the binding of the primary antibody. Some of these are
available commercially (examples include rabbit anti-goat IgG, goat anti-
rabbit IgG, rabbit anti-mouse IgG, and mouse anti-human IgE second
antibodies), with or without enzyme- or radio-label.
3.3 Antibody production

3.3.1 Choice and form of immunogen
The choice of immunogen, and also the form of the immunogen, is particularly
important for success in making antibodies for allergenic residues. This is
beyond a simple choice of raw or heat-processed immunogen. Every aspect
of the immunogen preparation process, from the choice of variety to the
Antibodies 69
washing of the starting materials to remove agricultural contaminants,

contributes to the selectivity and usefulness of the resulting antibodies. One
can manipulate the immune response by injecting processed or denatured
antigen, but parameters around how to process and how much to process are
still mostly determined by trial and error and previous experience. Processing
can result in alteration of proteins such as partial or complete denaturation,
aggregation, and formation of Maillard reaction products, some of which
can serve as immunodominant determinants. Even age-old recommendations
such as defatting may not be advisable for production of antibodies for
allergenic residues, as important oil-body associated proteins can be lost in
the process. A severely underestimated cause of cross-reactivity in polyclonal
antiserum preparations is agricultural contamination, which is more often
the norm rather than the exception, even with unrelated materials; simple
extensive washing of the source materials can circumvent this problem. The
debate of using raw or processed source material for producing antibodies
for allergens continues; a good approach to start with is to use processed
immunogen in trying to make antibodies that can detect processed forms of
the allergenic food; however, success depends on many factors and properties
beyond this solitary consideration.
While the approach of using a crude extract of allergenic food to make
antibodies has been criticized for having the potential to result in highly
cross-reactive antisera, high-quality monospecific polyclonal antisera against
peanut and other allergenic foods have been very successfully produced
using crude extracts (Yeung and Collins, 1996; ‘Holzhauser and Vieths,
1999; Hlywka et al., 2000). In fact, many of the successful commercial
enzyme-linked immunosorbent assay (ELISA) kits for allergenic residue
detection use antibodies generated against crude extracts of allergenic foods,
and the methods suffer few cross-reactivity issues. It is possible to make
antibodies against a single isolated protein, but processing or denaturation of
that protein can mean reduced activity. In most cases the food industry really
only cares about whether any peanut protein is present, not just if one allergenic
peanut protein is present. The view of regulatory agencies may differ from
this in that they may need to prove a hazard exists, and hence the presence
of the allergen, but indicator proteins can be a successful approach. In fact,
the commercial peanut protein test kits can detect the major peanut allergens
Ara h 1 and Ara h 2 (Nogueira et al., 2004).
3.3.2 Adjuvants
Often the use of adjuvants is necessary for antibody production. The function
of an adjuvant is to enhance the immunogenicity of the antigen by increasing
the immune response in several ways; it increases the efficiency of antigen
presentation and the number of antibody-secreting B cells, acts as an
immunogen depot (for extended antigen stimulation), protects the immunogen
from rapid removal and catabolism, and increases the affinity and avidity of
the antibody response (Spier and Griffiths, 1985; Catty, 1988; Roitt et al.,
1989). Adjuvants are available in many forms (Harlow and Lane, 1988). The
best known adjuvant is complete Freund’s adjuvant (CFA), a suspension of
killed mycobacteria in mineral oil which was first described over 50 years
ago. Incomplete Freund’s adjuvant (IFA) is the mineral oil minus the bacteria;
sometimes IFA can be used successfully by itself (Campbell, 1996). There
are many other types of adjuvants that can be used besides CFA and IFA
(Hefle, 1995; Campbell, 1996).
3.3.3 Route of administration

The route of administration of immunogen varies according to the volume
being injected, the buffers and other components that will be injected with
the immunogen, and how quickly the immunogen should be released into the
circulation (Harlow and Lane, 1988). The choice of route used also depends
on the physical nature of the immunogen, the species, and the stage of
immunization (Deshpande, 1996c). Larger volumes are usually given to animals
via the subcutaneous route, but in rodents, intraperitoneal injection is the
most-used method. Intravenous injections can be used, but care is necessary,
as there is a risk of both anaphylaxis and also pulmonary embolism.
Immunogens which contain particulate matter or adjuvants should not be
given intravenously. Intramuscular or intradermal injections are used for
slower immunogen release. Subcutaneous and intramuscular routes are most
often used with Freund’s adjuvant. Both of these are proficient in forming
depot sites and slow immunogen release. Subcutaneous injection releases
immunogen slower than other routes, and is also preferred for booster injections,
due to a decreased chance of anaphylaxis. Adjuvant can be safely used with
intraperitoneal injections, but should not be used with intravenous injections.
Subcutaneous and intradermal routes are often used in rabbits, but intradermal
or intramuscular routes are used in larger animals.
3.3.4 Dose
The dose of immunogen depends on the its nature and also on the animal
being used. Large doses may induce tolerance, but in some cases
hyperimmunization schedules utilizing a relatively high amount of immunogen
have been successful (Cordle et al., 1991). The desired result is an antibody
with high avidity and affinity; therefore, use of the lowest amount of immunogen
necessary to achieve antibody production is usually the best approach. A
typical dose for rabbits is 100 μg for the primary injection, with 50–100 μg
used for booster injections (Harlow and Lane, 1988), although amounts well
above that have been reported and can be used. For larger animals, 500–
1000 μg is usual to start with. For booster injections, the traditional amount
used is 10–50% of the primary dose. Smaller booster doses promote clonal
selection for high-affinity Ab. For antigens in precious supply, one or two
Antibodies 71
IgM IgG IgG IgG

low affinity low affinity medium affinity high affinity
Serum titer
Primary Secondary Tertiary Multiple

injection injection injection injection
Fig. 3.2 Kinetics of a typical immune response.
injections with a low dose to prime the animals followed by a larger booster
injection can be successful (Harlow and Lane, 1988).
A specific antibody in serum can usually be detected about a week after
the initial immunization. The first type of antibody produced is IgM with
high avidity but low affinity (Deshpande, 1996c), but then a switch to IgG of
lower avidity, but higher affinity, occurs (Fig. 3.2). Responses to booster
injections are an increase in titer; as the time of immunization extends,
antibodies of the IgG class dominate. In general, a three to four week interval
between the primary injection and a booster injection is recommended
(Deshpande, 1996c). Usually, high titer is attained in about two to four
weeks after booster injections. The immune response matures, producing
higher-affinity and avidity antibodies as time goes on. For this reason, it is
possible to pool polyclonal antibodies after a certain amount of time, rather
than including early bleeds. Although intervals between booster injections
can be varied, there should be adequate time allowed for the circulating
antibody level to drop low enough to prevent prompt clearance of the injected
immunogen. For IgG, the half life is 20–25 days in the blood circulation, and
the clearance is 10–15% per day (Cruse and Lewis, 1999). As a result, to
avoid this problem, booster injections should be done every four to six weeks.
The quality of the antiserum is more important in immunoassays than in
other techniques. During immunization, antibody production and quality
(titer) needs to be monitored. Serum samples are collected 7–14 days after
booster injection, which corresponds to peak IgG production rate. For polyclonal
antibody production, test bleeds (small volume) are taken until sufficient
titer is reached, at which time production bleeds (larger volume) are done.
Serum samples from individual animals should not be pooled in the initial
stages; once a good titer and affinity have been developed, individual bleeds
can then be pooled to form a larger quantity of homogeneous antiserum.
Titers are usually monitored using indirect ELISA, in which the antigen is
coated onto microtiter plates and the antiserum is diluted serially and added
to plates. Titer can be defined in many ways, but one common method is the
mid-linear point of the titration curve. Antigen-specific IgG of approximately
8–14 mg/mL serum can be obtained.
Antibody affinity can also be ascertained using indirect ELISA, but using
the analyte of interest as an inhibitor. The dilution of antibody used in the
ELISA is the dilution identified by the titer. The horizontal displacement of
the inhibitor curve is an indirect measure of the affinity of the antibody
(Morris, 1985), as the greater the displacement, the greater the affinity of the
antibody. Cross-reactivity of the antibody can also be determined using a
similar technique; the curve generated from other substances should be able
to be superimposed on the antibody dilution curve.
3.3.5 Engineered antibodies

Progress in phage antibody display technology has revolutionized the ability
to select and engineer monoclonal antibodies. Usually these are human
monoclonal antibody fragments with desirable specificities that have therapeutic
potential. They are selected from extremely large libraries consisting of
engineered phage particles, each expressing an antibody fragment with a
unique specificity, such as for a tumor cell. Antibody fragments with relevant
binding are converted into intact IgG or IgA antibodies and could have
future potential in therapy for some types of diseases. While engineering
production of antibodies for therapeutic purposes is a promising new field,
the costs for such an approach preclude the use of this technique for routine
production of specific antisera for many applications.
Tolerance to a particular immunogen can occur for several reasons. Certain
conditions promote tolerance, and should be avoided in the production of
high-affinity, quality antibodies (Howard, 1979). Tolerance is mounted against
specific antigenic determinants, not the antigen as a whole. T-independent
immunogens (those which induce antibody synthesis, usually only IgM, in
the absence of lymphokines released by T-cells) may act as tolerogens if
injected at a high dose (100–1000 × normal) (Parks et al., 1979). In contrast,
low doses may cause immature B-cell clones to terminate, and immature B
cells are more prone to tolerance than mature B-cells. Excess T-independent
immunogen and absence of T-cell assistance prevent plasma B cells from
their normal function and, if repeatedly injected, can function to remove
functional B-cell clones. In addition, excess T-independent immunogen can
hinder secretion of antibody by plasma B cells. Weak immunogens may
induce tolerance if given in very small doses. Slowly released immunogens
preserve tolerance longer than those that are quickly catabolized.
3.3.6 Cross-reactivity
Antigen–antibody reactions show a high degree of specificity when the binding
sites of the antibodies are not complementary and thus do not recognize or
Antibodies 73
cross-react with the determinants of another antigen. However, the specificity

of an antiserum echoes the many specificities of its component antibodies
(Catty, 1988). A polyclonal antiserum does have the capability to
monospecifically bind to an antigen if the range of its constituent specificities
is only to epitopes that are restricted to that antigen. A cross-reactive antigen
combines with antibodies induced in response to a different antigen but
which has shared antigenic determinants. Cross-reactivity comes to pass
either from the binding of structurally different determinants of the same
antigen by the same antibody, or because of the existence of common epitopes
(Deshpande, 1996b).
It is important to assess cross-reactivity with many food ingredients when
developing immunoassay methods for detection allergenic residues, due to
the vast variety and type of foods and food ingredients used in the manufacture
of packaged food products. Simply reporting cross-reactivity of low levels of,
or diluted, ingredients, is not sufficient, and test substances for cross reactivity
should be extracted in concentrated form (at full extraction concentration)
and analyzed; in other words, cross-reactivity should be approached such
that testing substances are assumed to comprise 100% of future samples.
3.3.7 Production of monoclonal antibodies

The ability of each individual B lymphocyte to produce an individual, unique
antibody was discovered in the 1950s. It is estimated that a typical BALB/
c mouse that has 2 × 108 B-cells can produce 10–40 × 106 distinct clonotypes
(Klinman, 1972); therefore, for any given animal, hypothetically, several
different clonotypes could be produced. Kohler and Milstein (1976) were the
first to produce monoclonal antibodies. Antibody-producing lymphocytes
are isolated and then fused with myeloma cells (mutant cells that produce no
antibodies) to produce perpetual hybridomas, which produce many copies of
the same antibody. Detailed procedures for production of monoclonal antibodies
can be found elsewhere (Goding, 1986; Harlow and Lane 1988). There are
five major parts to monoclonal antibody production; immunization of the
rodents, fusion of immunized spleen cells with myeloma cells, selection of
appropriate clones, cloning, and production of large amounts of antiserum
(Fig. 3.3). The hybrid cells are selected using selective culture media and
then screened for the specific antibody produced. The result is a cell line
clone that only produces one type of monoclonal antibody. These cells can
be maintained over long periods of time. Large amounts of antiserum can
either be collected through the production of ascites in vivo or by in vitro cell
culture. Ascites is intraperitoneal fluid obtained from mice that have been
injected intraperitoneally with a hybridoma clone, causing tumors in the gut
that produce the monoclonal antibody. In the past, ascites generation was
most often used to collect large amounts of antiserum from monoclonal cell
lines, but due to animal welfare concerns, production of ascites is banned or
severely limited in many countries today. Ascites fluid can contain up to 5–
Antigen
Polyclonal serum
antibodies have
variable affinity
for epitope
Myeloma PEG B cells from
cell line spleen
Hybridomas Assay for clone

secreting high-
affinity antibody
to specific epitope
Propagate
Growth and cloning
Fig. 3.3 Monoclonal antibody production.
15 mg/mL of the monoclonal antibody (Deshpande, 1996c), and as much as

40 mL can be obtained per mouse. Alternatives to ascites include recovery of
the antibodies from stationary cell culture supernatant solutions, use of roller
bottles or stirred tanks, gas-permeable tissue culture bags, mini-fermentors,
and airlift (packed-bed) reactors. Spinner flasks are usually not used for
antibody harvest becaue the shear forces can damage the antibodies. The
process of producing hybridomas can take from two months to up to a year.
Methods for ascites and cell culture production of monoclonal antibodies are
can be found elsewhere (Goding, 1986; Harlow and Lane, 1988).
3.3.8 Purification
Antiserum either from polyclonal antiserum or monoclonal culture supernatant
or ascites can be further purified. Often, crude serum or semi-purified serum
can be used, but in some cases further purification is recommended or necessary.
Antibodies 75
For polyclonal antiserum, antibodies can be semi-purified using as crude a

scheme as enriching the IgG content through ammonium sulfate precipitation,
or as sophisticated as using antigen immunoaffinity to purify it. The source
material, nature of possible contaminants, and the final application of the
antiserum are all considerations for the choice of purification approach
(Deshpande, 1996c). While the total immunoglobulin concentration in a
crude polyclonal antiserum preparation can be as high as 10 mg/mL, the
antibody of interest may comprise only about 10% of this (Deshpande, 1996c).
Monoclonal antibodies in culture supernatant are fairly dilute (5%) and therefore
usually need concentration before use; one can use ammonium sulfate
precipitation for this, although substances in the culture media can be co-
precipitated, such as fetal calf serum proteins. Ascites can contain 0.9–9 mg/
mL specific antibodies, but may also have other mouse antibodies present
(Harlow and Lane, 1988). Approaches for purification of antibodies are reviewed
elsewhere (Goding, 1986; Harlow and Lane, 1988; Deshpande, 1996c).
3.4 Choice of producing monoclonal or polyclonal

antibodies
An ideal antibody will possess high titer and also high affinity and specificity
for the analyte of interest. Affinity is the main factor in achieving a low
detection limit for an immunoassay (Deshpande, 1996b). For commercial
interests, large scale production of antibody is a major consideration; while
monoclonal antibodies are good from this aspect, they are expensive to initially
produce and may not be as good as polyclonal antibodies at detecting processed
forms of the antigen. Usually, a specific polyclonal antibody, once produced,
is available in quantities that are large enough for commercial or research
interest, and if a goat or sheep is used, large amounts of antiserum are
available. With antigens that are good immunogens, producing more antiserum
is ordinarily not a problem.
Polyclonal antibodies often recognize multiple epitopes, making them
more tolerant of small changes in the nature of the antigen. Therefore, polyclonal
antibodies may be preferred when the antigen of interest is denatured or
altered in some way, such as in food processing (Goding, 1996). Antigen–
antibody bonds are usually stable under pH conditions of 4–9 and salt
concentrations of 0.1–1 M NaCl, although occasionally, monoclonal antibodies
can be separated from their antigen under these mild conditions (Hermann
and Mescher, 1979). By definition, a monoclonal antibody binds with only
one epitope on the antigen. The popular myth is that monoclonal antibodies
have an advantage over polyclonal antibodies because they are supposedly
not cross-reactive; however, monoclonal antibodies can possess cross-reactivity
to unrelated antigens (Lane and Koprowski, 1982). In one report, monoclonal
antibodies were reported to cross-react with two muscle proteins, tropomyosin
and vimentin (Blose et al., 1982). Lane and Koprowski (1982) have suggested
two possible causes for monoclonal antibody cross-reaction; that the antibody
detects structural similarities in the two antigens, or that the binding site of
the antibody could combine with unrelated antigens in a multispecific way,
and recognize two or more entirely different epitopes (Richards et al., 1975).
An antigen–antibody complex occurs whenever there are enough and the
right strength of inter- (and intra-) molecular forces. Goding (1996) states
that it is not surprising that monoclonal antibodies can have multispecificity
and, indeed, Hefle et al. (1994) noted cross-reactivity among their developed
monoclonal antibodies, to different peanut proteins. Unusual cross-reactions
are almost never seen with polyclonal antibodies because the odd cross-
reaction is ‘diluted out’ in the enormity of the rest of the response and the
‘consensus’ of different clones (Goding, 1996). Specificity also depends on
this consensus and polyclonal antibodies bind to determinants that cover
almost the entire external surface of the antigen (Benjamin et al., 1984).
Therefore, minute changes in the structure of the antigen due to genetic
polymorphism, heterogeneity of glycosylation, or slight denaturation will
usually have little or no effect on polyclonal antibody binding. A subset of
polyclonal antibodies will usually bind to antigen that has been modified or
denatured, even if this was not the form of the immunogen used for the
immunization (Burnette, 1981). In contrast, monoclonal antibodies most
often bind to a single unique epitope – if for any reason this site is altered
(due to denaturation, unfolding, aggregation, formation of Maillard reaction
products, etc.), the monoclonal antibody may not bind (Goding, 1996). Another
potential drawback to monoclonal antibodies is that they may only be able to
cross-link antigen molecules with two or more binding sites (Goding, 1996),
although this requirement is usually not an issue for allergenic proteins. If
cross-reactivity is present with monoclonal antibodies, it is hard to remove,
unlike for polyclonal antibodies. Monoclonal antibodies have sensitivity to
only part of the total antigenicity of the immunogen, resulting in poorer
performance compared to that of polyclonal antibodies in many applications.
Polyclonal antibodies may be generated in a great variety of species, including
rabbit, goat, sheep, donkey and chicken, giving the immunoassay developer
many choices in design of assay; this is in contrast to monoclonal antibodies,
which are usually only produced in rodents due to practical considerations.
It can be difficult to obtain polyclonal antisera with identical properties
from different animals. Even from the same animal, polyclonal antisera
collected at different times can have different properties. These limitations
of polyclonal antibodies can justify producing monoclonal antibodies, as the
latter have identical physical, biochemical and immunological properties.
Due to their specificity, monoclonal antibodies are good for use as a primary
antibody in immunoassays. In addition, the specificity of monoclonal antibodies
spurs efficient binding of antigen in a mixture of related molecules, and this
property can be exploited in such techniques as affinity purification. A
disadvantage of monoclonal antibodies is that their preparation can be time-
Antibodies 77
consuming and expensive. In addition, they can be difficult to label (Deshpande,

1996c).
3.5 References
Benjamin, D C, Berzofsky, J A, Esat, I J, Gurd, F R N, Hannum, C, Leach, S J, Margoliash,
E, Michael, J G, Miller, A, Prager, E M, Reichlin, M, Sercarz, E E, Smith-Gill, S J,
Todd, P E and Wilson, A C (1984) ‘The antigenic structure of proteins: a reappraisal’,
Ann Rev Immunol, 2, 67–101.
Blose, S H, Matsumara, F and Lin, J J C (1982) ‘Structure of vimentin 10-nm filaments
probed with a monoclonal antibody that recognizes a common antigenic determinant
on vimentin and tropomyosin’, Cold Spring Harbor Symp Quant Biol, 46, 455–463.
Burnette, W N (1981) ‘“ Western blotting”; electrophoretic transfer of proteins from
sodium dodecyl sulfate-polyacrylamide gels to unmodified nitrocellulose and
radiographic detection with antibody and radioiodinated protein A’, Anal Biochem,
112, 195–203.
Campbell, A M (1996) Production and purification of antibodies, in Diamands, E P and
Christopoulos, T K (eds), Immunoassay, San Diego, CA, Academic Press, 95–115.
Catty, D (1988) Antibodies, a Practical Approach, Vol 1, Oxford, UK, IRL Press.
Cordle, C T, Mahmoud, M I and Moore, V (1991) ‘Immunogenicity evaluation of protein
hydrolysates for hypoallergenic infant formulae’, J Pediatr Gastroenterol Nutr, 13,
270–276.
Cruse, J M and Lewis, R E (1999) ‘Atlas of Immunology’, Washington, DC, CRC Press,
127–142.
Davies, D R and Padlan, E A (1990) ‘Antibody–antigen complexes’, Ann Rev Biochem,
59, 439–473.
Deshpande, S S (1996a) ‘Antibodies: biochemistry, structure, and function’, in Deshpande,
S S, Enzyme Immunoasssays: From Concept to Product Development, New York, NY,
Chapman and Hall, 24–51.
Deshpande, S S (1996b) ‘Antigen-antibody reactions’, in Deshpande S S, Enzyme
Immunoasssays: From Concept to Product Development New York, NY, Chapman and
Hall, 52–71.
Deshpande, S S (1996c) ‘Antibody production’, in Deshpande S S, Enzyme Immunoasssays:
From Concept to Product Development, New York, NY, Chapman and Hall, 117–154.
Feldkamp, C S and Carey, J S (1996) ‘Immune function and antibody structure’, in
Diamands, E P and Christopoulos, T K (eds), Immunoassay, San Diego, CA, Academic
Press, 5–24.
Goding, J W (1986) Monoclonal Antibodies: Principles and Practice, 2nd edn, London,
UK, Academic Press, 3–5, 45–103.
Harlow, E and Lane, D (1988) Antibodies: A Laboratory Manual, Cold Spring Harbor,
NY, Cold Spring Harbor Laboratory Press.
Harlow, E and Lane, D (1999) Antigen-antibody interactions, in Harlow, E and Lane, D,
Using Antibodies, A Laboratory Manual, Cold Spring Harbor, NY, Cold Spring Harbor
Laboratory Press, 23–37.
Hawcroft, D, Hector, T and Rowell, F (1987) Quantitative Bioassay, New York, NY, John
Wiley & Sons, Inc.
Hefle, S L (1995) ‘Immunoassay fundamentals’, Food Technol, 49, 102–107.
Hefle, S L, Folgert, J P, Chu, F S and Bush, R K (1994) ‘Monoclonal antibodies against
selected peanut allergens: production and use as affinity agents, Food Agric Immunol,
6, 197–208.
Hermann, S H and Mescher, M F (1979) ‘Purification of the H-2K k molecule of the
murine major histocompatibility complex’, J Biol Chem, 254, 8713–8716.
Hlywka, J J, Hefle, S L and Taylor, S L (2000) ‘A sandwich enzyme-linked immunosorbent

assay for the detection of almonds in foods’, J Food Prot, 63, 252–257.
Holzhauser, T and Vieths, S (1999) ‘Indirect competitive ELISA for determination of
traces of peanut (Arachis hypogaea L.) protein in complex food matrices’, J Agric
Food Chem, 47, 603–611.
Howard, J G (1979) Immunoglobulin tolerance, in Lennox, E S (ed), Defence and
Recognition – Cellular Aspects, Vol 22, Baltimore, MD, University Park Press.
Hughes-Jones, N C, Gardner, B and Telford, R (1964), ‘Effect of pH and ionic strength
on the reaction between anti-D and erythrocytes’, Immunology, 7, 72–81.
Jefferis, R and Deverill, I (1991) The antigen–antibody reaction, in Paice, C P and
Newman, D J (eds), Principles and Practice of Immunoassay, New York, NY, Stockton
Press, 1–18.
Lane, D and Koprowski, H (1982) ‘Molecular recognition and the future of monoclonal
antibodies’, Nature, 296, 200–202.
Klinman, N R (1972) ‘Mechanism of antigenic stimulation of primary and secondary
clonal precursor cells’, J Exp Med, 136, 241–260.
Kohler, G and Milstein, C (1976) ‘Derivation of specific antibody-producing tissue culture
and tumor lines by cell fusion’, Eur J Immunol, 6, 511–519.
Morris, B A (1985) Principles of immunoassay, in Morris, B A and Clifford, M N (eds),
Immunoassays in Food Analysis, London, UK, Elsevier Applied Science Publishers,
21–52.
Nogueira, M C L, McDonald, R, Westphal, C D, Maleki, S J and Yeung, J M, (2004), ‘Can
commercial peanut assay kits detect peanut allergen?’, J AOAC Int, 87, 1480–1484.
Parks, D R, Bryan, V M, Oi, V T and Herzenberg, L A (1979) ‘Antigen-specific identification
and cloning of hybridomas with a fluorescent-activated cell sorter’, Proc Natl Acad
Sci USA, 76, 1962–1966.
Porter, R R (1959) ‘Hydrolysis of rabbit γ-globulin and antibodies with crystalline papain’,
Biochem J, 73, 119–126.
Porter, R R (1962) Basic Problems of Neoplastic Disease, New York NY, Columbia
University Press.
Raff, M S (1970) ‘Role of thymus-derived lymphocytes in the secondary humoral immune
response in mice’, Nature (London), 226, 1257–1258.
Richards, E F, Konigsberg, W H, Rosenstein R W and Varga, J M (1975) ‘On the specificity
of antibodies, Science, 187, 130–137.
Roitt, I M, Brostoff, J and Male, D (1989) Immunology, London, UK, Gower Medical
Publishing.
Spier, R and Griffiths, B (1985) Adjuvants in Animal Cell Biotechnology, Vols 1 and 2,
New York, NY, Academic Press.
Steward, M W and Steensgaard, J (1983) Antibody Affinity: Thermodynamic Aspects and
Biological Significance, Boca Raton, FL, CRC Press.
Yeung, J M and Collins, P G (1996) ‘Enzyme immunoassay for the determination of
peanut proteins in food products’, J AOAC Int, 79, 1411–1416.
4
Allergen-specific human IgE antibody-

based analysis of food
R. Hamilton, Johns Hopkins University School of Medicine, USA
4.1 Introduction
Allergens are foreign proteins or immunogens that, when introduced in an
immunocompetent and predisposed host, elicit the formation of IgE antibodies.
Once induced, IgE circulates in the blood and binds on to high-affinity IgE
Fcε receptors on mast cells and basophils. This process leads to a state of
sensitization that is evidenced by the detection of allergen-specific IgE antibody
on skin mast cells and blood basophils. Upon further allergen exposure by
ingestion, injection or inhalation, allergen can cross-link receptor bound IgE
antibody, causing degranulation of vasoactive mediators from mast cells and
leading to a spectrum of allergic reactions. Allergen-specific IgE antibody is
thus the key analyte in both in vivo – skin test, basophil meditator release –
and in vitro radioallergosorbent (RAST)/fluorescent enzyme immunoassay
(FEIA)/enzyme-linked immunosorbent assay (ELISA) – immunoblot –
inhibition assays that permits qualitative identification and quantitative
measurement of allergens in complex biological substances such as foods.1
In vivo, allergen that is injected into the skin of a sensitized individual
induces a wheal and flare reaction that can be measured and used to identify
the presence and quantify the allergen’s potency. Alternatively, controlled
ingestion of foods in a double-blind, placebo-controlled manner allows in
vivo food challenges which many allergists consider a definitive diagnostic
test for an individual’s sensitization to a particular food allergen. Due to its
relative hazard, however, this method is rarely used to identify or quantify
the levels of allergenic protein in particular foods. In vitro, allergen can
competitively inhibit the binding of allergen-specific IgE antibody to solid
phase allergen in the RAST2 and its newer non-isotopic counterparts such as
the FEIA-based CAP SystemTM (Pharmacia Diagnostics AB, Kalamazoo,

MI)3,4 and the Immulite® System (Diagnostic Products Corporation, Los
Angeles, CA).5 This chapter focuses on the design, performance characteristics
and strengths and weaknesses of allergen-specific IgE antibody-based in
vivo and in vitro assay methods.
4.2 IgE antibody-based in vivo assay

The allergenic potency of an extract of raw or cooked food can be measured
by the ability of allergenic proteins in this extract to elicit a reaction in the
skin of an individual who is known to be sensitized to that food. Administration
of allergen into the sensitized skin produces a wheal (sharply circumscribed
localized area of edema) and flare (area of erythema surrounding the wheal).
As briefly noted above, the principle of this in vivo assay involves food
allergen-specific IgE antibodies that are bound to Fcε receptors on skin mast
cells. They are cross-linked by allergen once it has been injected into the
skin of a sensitized individual. Within 15 minutes of this event, the cells are
activated to secrete histamine, leukotrienes and other mediators that produce
the observed wheal and flare reaction in the skin. By injecting different
dilutions or doses of calibrated diagnostic reference allergen extracts of
known potency into the skin, a dose-response curve can be generated and the
relative potency of a test food extract can be determined. The larger the skin
reaction, the more allergen is present in the extract.
4.2.1 Reagents
Food allergen extracts
While virtually any food can cause allergic symptoms, relatively few foods
reportedly induce allergic reactions with any frequency in humans. The
commonly implicated allergenic foods include cows’ milk, egg, wheat, soy,
peanut, tree nuts, fish and crustacean shellfish.6 Diagnostic skin test extracts
are commonly prepared from fresh or cooked foods in a defined weight
versus physiological extraction buffer ratio (e.g. 1:20 w/v). They are sometimes
defatted, concentrated and put into 50% glycerin to minimize degradation
during storage. Except in the case of recombinant and some purified research
food allergens, diagnostic food allergen extracts tend to be crude allergenic
mixtures that display multiple protein and glycoprotein bands on
polyacrylamide gel electrophoresis and isoelectric focusing. Food allergens
contain linear allergenic epitopes that bind to IgE antibody and appear highly
resistant to denaturation by heat and proteases and conformational epitopes
that require three-dimensional folding of the allergenic protein to permit
binding to IgE antibodies. A comprehensive listing of food allergen specificities
together with their genus and species is presented elsewhere.7
Allergen-specific human IgE antibody-based analysis of food 81
Reference materials
Since there are few established diagnostic food allergen reference preparations,
most materials that are used as reference or calibration reagents in IgE-based
quantitative assays have been through a previous allergenic potency assessment
by the in vivo or in vitro methods described in this chapter. They are often
first characterized by separation and identification of their component proteins
using sodium dodecyl sulphate (SDS) polyacrylamide gel electrophoresis
and isoelectric focusing. Their total protein concentration is also determined
by one of several assays. They are then analyzed in an in vitro inhibition
assay for the purpose of determining the relative allergenic potency of test
allergen extracts, presumably with the same allergenic protein profile.
4.2.2 In vivo assay design

The US Food and Drug Administration has adopted the ‘parallel line bioassay
skin test method’ to assay the allergenic potency of diagnostic skin test
extracts in vivo.8 In this assay, 18 serial three-fold dilutions of reference and
test extracts are prepared from a concentrated extract (0 dose). A volume of
0.05 mL of each extract dilution is then injected intradermally or applied by
puncture onto the skin (volar area of the arm and/or back) and monitored at
15 minutes for swelling and redness. The circumferences of the wheal and
erythema are outlined with a pen and transferred onto tape for a permanent
record. The sum of the longest and midpoint orthogonal diameters of the
erythema and wheal is then used to construct titration curves. Preliminary
skin tests on the arm are used to determine the doses that result in response
with a sum of the erythema diameters from 0–120 mm. Responses at the first
four dilutions that produce a positive reaction are used to construct dose–
response curves on the back that exhibit steep slopes and good correlation
coefficients. The relative potency (RP) of a test extract is interpolated from
the dose–response curve of the reference extract that has been calibrated in
either mg of total protein, protein nitrogen units or arbitrary biological RP
units. While the parallel line bioassay skin test method is useful in assessing
the potency of diagnostic skin test reagents, it is not practical for the routine
quantification of the amount of allergen that may be present in an extract
prepared from a food.
Figure 4.1 illustrates the typical immediate-type hypersensitivity skin
reaction that occurs 20 minutes following the administration of four ten-fold
dilutions of an allergen extract into the skin of an individual sensitive to the
allergen. The diameters of both the wheals and erythemas get larger with
increasing concentrations of the allergen. The highest concentration of allergen
(1:10 dilution) produces the largest wheal and erythema response. The figure
illustrates the importance of administering the different skin test doses at
sufficient distances from each other to ensure that the erythema from the
individual skin tests do not overlap and thus become difficult to measure
accurately. At five hours, a late-phase skin reaction associated with cellular
Immediate and late skin reactions

Late response Immediate response
(at 5 hours) (at 20 minutes)
1:10
1:100
1:1000
1:10 000
Fig. 4.1 Intradermal skin test titration in which the photograph was taken 20 minutes
(right side) and five hours (left side) after the administration of four ten-fold
concentrations of allergen to the forearm of a sensitized individual. The swelling
(wheal) and redness (erythema) on the right side of the arm after 20 minutes identify
the titer or the most dilute concentration of allergen that produces a positive skin
reaction in comparison to a saline control. Reproduced with permission from
Roitt et al.25
infiltrates can also be observed; however, this is generally overlooked in the

assessment of diagnostic skin test extracts for allergen potency.
4.2.3 Human subject considerations

Atopic subjects participating in the assessment of the in vivo potency of
diagnostic food allergen extracts must first have a definitive sensitivity to the
allergenic food stuff in question, based minimally on a positive clinical
history of food allergy and a positive skin test. Second, they must provide
informed consent. Where possible, the subject should have a positive double-
blind placebo-controlled food challenge9 to confirm their sensitivity to the
particular food(s) being evaluated. Approximately 30–50% of individuals
with a clinical history of an adverse food reaction and a positive puncture
skin test to the identified food have a positive confirmatory oral food
challenge.10,11 At the time of the study, the individual should not have ingested
medications such as antihistamines that might interfere with a wheal and
flare response and they should not be on immunotherapy. Atopic individuals
who are chronically exposed to allergen will often produce IgE antibody to
different allergens within a given complex protein mixture. Thus, to obtain
a reliable estimate of the relative allergen content of test extracts, it is preferable
to perform a potency bioassay on more than 20 sensitized individuals.8
4.2.4 Calibration strategy

The relative potency of an allergen extract is determined by interpolation of
test extract response data from a dose–response curve generated with a reference
extract of defined allergenic potency. Often the allergen content in the reference
material is reported in arbitrary units such as AUs (allergen units), BAUs
(biological allergen units) or RP units. Occasionally, the reference extract’s
total protein has been quantified using the modified ninhydrin assay12 or the
primary allergen content of the extract. Allergen levels in the test extracts are
interpolated and reported in relationship to these arbitrary units in the reference
extract.
4.2.5 Quality control

The controls used in this assay include histamine as a positive control to rule
out medication (anti-histamine) interference or suppression of the wheal and
flare reaction and a negative diluent control such as saline to rule out
dermagraphism that can occur from skin trauma. Intradermal injections are
performed in duplicate to maximize repeatability and reproducibility.
4.2.6 Performance characteristics

Analytical sensitivity
The analytical sensitivity of the in vivo bioassay is dependent upon the
overall sensitivity of the subject for the allergen specificity of interest and
the criteria used for defining a positive skin test result (e.g. 3 mm above the
saline control). Sensitivity ultimately depends upon the amount and specificity
of IgE antibody bound to mast cells and the releasability of the individual’s
mast cells following allergen challenge. The reported sensitivity in the parallel-
line skin test assay using common ragweed (Ambrosia artemisiifolia) as a
model antigen was 5 × 10–5 μg/ml of Amb a 1 [AgE] for subjects sensitive
to labile antigens and 3 × 10–6 μg/ml of Amb a 1 for subjects sensitive to
stable allergens.8 Using erythema data, the intradermal skin test application
required an average of 30 000-fold lower dose of allergen than the puncture
skin test application to obtain a comparable skin test response.
Analytical specificity
The allergenicity of a substance depends on the extent of its foreigness (size,
stability, 1°, 2°, and 3° protein structure), the extent of the individual’s
exposure (concentration) and the genetic (atopic) predisposition of the exposed
individual. Plant-derived food allergens have been classified into discrete
allergen groups that explain the molecular basis of the cross-reactivity seen
among different food groups13 (see Chapter 2). Some individuals, for instance,
who develop an IgE antibody response to proteins in natural rubber latex
from Hevea brasiliensis (Hev b) trees also exhibit an oral allergy syndrome
when they eat certain foods such as avocados, kiwi, bananas and chestnuts.
This has been called the latex–fruit syndrome and it stems from cross-reactivity
of IgE antibodies to proteins in these foods and natural rubber latex14. For
example, the class I chitinases from the pathogenesis-related (PR) group 4
have been shown to be involved in the IgE antibody cross-reactivity observed

between avocado, chestnut and natural rubber latex.15 The consequence of
this cross-reactivity is that individuals who are selected for allergen potency
assessment by in vivo bioassay will need to have their IgE antibody tested for
confounding cross-reactivities to structurally similar allergenic epitopes in
other foods.
Variability
Both repeatability (intra-assay) and reproducibility (inter-assay) should be
maximized by applying each skin test dilution in duplicate. The reproducibility
reported in the parallel-line skin test assay using the intradermal method of
application was 14.2% coefficient of variation (mean +/– 2SD = 102 +/–
14.8 RP).8 Since the slope of the puncture skin test dose–response line is
significantly flatter than the intradermally applied skin test results, there is
increased variability of an assay using the puncture skin testing application
method.
4.2.7 Dynamic range and data analysis

The dynamic range of this bioassay extends over 4–5 logs depending upon
the concentration of allergen-specific IgE and the overall skin sensitivity of
the test subjects. Figure 4.1 illustrates the increasing size of the wheal and
erythema with increasing allergen concentrations from 1:10 000 to 1:10 at 20
minutes following administration of the allergen to skin. By administering
the reference and test extracts each at four dilutions in duplicate, side by side
in the same subjects, the mean wheal and erythema data for the test extract
can be interpolated from the reference curve. In the original report, Turkeltaub
et al.8 performed a best-fit linear regression of the sum of the erythema or
wheal diameters obtained from four intradermal three-fold serial dilutions
near the endpoint. Their original expectation was a correlation coefficient of
> 0.85 and slopes of the reference and test extract regression lines that did
not significantly differ from each other.
4.2.8 Assay limitations and applicability

Because of a number of limitations associated with the in vivo potency
assessment of food allergen extracts, methods such as the parallel-line bioassay
are rarely performed today for allergen quantitation. First, the bioassay involves
direct challenge of the skin of sensitized individuals with allergen. This
places an allergic individual at risk for a possible severe systemic allergic
reaction, especially when test extracts of unknown potency are being
administered. The possibility of a systemic reaction makes obtaining informed
consent most difficult. Second, when the bioassay is performed, it is limited
to sensitized adults. Since there is a loss of symptomatic reactivity to most
food allergens over time in adults who avoid exposure, it is difficult to obtain
a number of sufficient individuals with the appropriate skin reactivity to

evaluate the potency of food extracts. This can even be the case with food
allergen sensitivities such as peanut, tree nut, fish and shellfish that extend
into adulthood. Third, even if an individual is willing and has the appropriate
skin reactivity, the individual’s actual degree of sensitivity is often not sufficient
to permit the bioassay to actually be performed. Fourth, there is a general
concern that a potency assessment dependent upon skin reactivity does not
accurately reflect the actual potency of food allergens that are ingested and
adsorbed through the gastrointestinal (GI) tract. Finally, the quality of the
bioassay can be compromised by prior anti-histamine use, dermagraphism,
the variability of skin test procedures and the diverse criteria used for defining
positive results. These limitations, especially the dependence on consented
human subjects with sufficient sensitivity to the food specificity of interest,
have forced investigators interested in food allergen potency assessment
toward the use of human IgE antibody-based in vitro assays.
4.3 IgE antibody-based in vitro assay for food allergen

detection
The radioallergosorbent test or RAST was the first immunoassay developed
to detect allergen-specific human IgE antibody.2,16 It employed a paper disc-
based solid-phase allergen or allergosorbent to bind specific antibody from
human serum and radiolabeled anti-human IgE to detect bound IgE antibody.
A minor modification involving pre-incubation of the IgE antibody containing
human serum with an allergen preparation (such as food extract) prior to the
addition of the allergosorbent produced a concentration-dependent inhibition
of the IgE antibody binding to the allergosorbent. The quantitative nature of
this inhibition permitted the amount of allergen in the extract to be assessed.
This competitive inhibition assay format (RAST inhibition assay) rapidly
became a widely used tool for identifying the presence and quantifying the
amount of allergen in a test extract. More recently, a plate-based enzyme
immunoassay (EIA) version of the RAST inhibition assay has been developed
as discussed elsewhere in this volume. Since allergen is typically adsorbed
on plastic surfaces of microtiter plate wells, the EIA inhibition assay works
best with purified allergens. Its IgE antibody binding capacity is generally
more limited than assays that use allergens covalently coupled to an
allergosorbent since less than 1 microgram of total protein typically adsorbs
on the plastic well surface. Both the RAST and EIA inhibition assays are
especially useful for assessing the allergenic potency of extracts containing
mixtures of allergenic proteins because the allergosorbent can be prepared in
a manner to ensure that all the different proteins in the mixture are insolubilized
in molar excess concentrations to IgE antibody levels in serum. However, as
with the in vivo method, the in vitro competitive RAST inhibition assay still
depends on human serum with sufficient quantities and appropriate specificities

of the allergen-specific IgE antibody.
4.3.1 Reagents
There are a number of required reagents that are needed to perform the
competitive RAST inhibition analysis. These include the allergosorbent that
is composed of a solid-phase matrix and an allergen component, allergen-
specific IgE antibody containing human serum, a radioisotope-, enzyme- or
fluorophor-labelled anti-human IgE detection reagent and characterized allergen
calibrators.
Allergen
There are several hundred food allergen specificities that are known to induce
IgE antibody responses in humans. Of these, there are a handful of allergen
specificities that produce much of the food allergy symptoms involving the
GI tract (vomiting, diarrhoea), respiratory tract (pneumonitis, asthma) and
skin (chronic urticaria, atopic dermatitis). These include cows’ milk, chicken
egg, wheat, soybean, fish, crustacean shellfish, peanuts, tree nuts, cereal
grains and citrus fruits. In vitro assay manufacturers have classified them as
a group with the letter F (for foods) and a sequential number that has no
significance other than when it was identified as an allergen source. For
instance F13 is the peanut specificity that is prepared from raw and shelled
Arachis hypogaea. Within the food group, the allergens can be segregated
into subgroups, such as dairy, fish, grains/grass, seed/nut, legume, crustacean,
vegetable, meat, fowl, mollusk, fungi, fruit, spice and stimulant (chocolate,
coffee). A number of complex food allergen extracts have been further
characterized with immunochemical methods so that individual allergenic
proteins have been identified. For instance, five allergens have been identified
in chicken egg white: ovomucoid (Gal d 1), ovoalbumin or conalbumin (Gal
d 2), ovotransferrin (Gal d 3), lysozyme (Gal d 4) and chicken serum albumin
(Gal d 5).17,18
For the purposes of this discussion, food extracts should be considered
complex protein mixtures. They are prepared by selecting fresh or cooked
food, grinding it up and preparing a physiological extract that is sometimes
defatted and then centrifuged and sterile filtered. Initial quality control of the
crude extract begins with analysis in one of several immunochemical methods
(isoelectric focusing immunoblot, SDS-polyacrylamide gel electrophoresis
separation followed by Western blot analysis, competitive binding
immunoassay) to ensure that it contains allergenic proteins. Allergen extracts
are subject to contamination, inherent biological variation and laboratory
misclassification. The use of fresh or cooked food, the precise method and
buffer used in extraction and the subsequent processing (column
chromatography, precipitation steps) can lead to highly variable extracts.
Stability during storage and following chemical manipulations such as coupling
onto a solid support or labelling with biotin can add to the heterogeneity.
Allergosorbent
The solid phase allergen or allergosorbent is one of the principal components
of the competitive RAST inhibition assay. The earliest RAST assay used
paper discs that were activated with cyanogen bromide to covalently couple
proteins. Because the paper disc was a defined size that fit into a 12 × 75 mm
test tube, it was difficult to vary the allergen amount bound to the disc. This
led researchers to covalently couple allergenic proteins to a variety of mobile
cyanogens/bromide-activated carbohydrate-based particles such as cellulose
(Avicel®/FMC BioPolymer, Philadelphia, PA) and cross-linked agarose
(SepharoseTM CL-4B; GE Healthcare AB, Uppsala, Sweden).19,20 Allergenic
proteins have also been coupled to cellulose threads, immobilon or nitrocellulose
membranes or more recently to a high binding capacity immobile sponge
that is used in the Pharmacia CAP SystemTM.3 Diagnostic Products Corporation
has chosen to biotinylate allergen mixtures and then insolubilize them on an
avidin-coated solid phase.5 In each of these cases, the covalent binding
process has permitted the immobilization of proteins of widely varying
molecular weights and isoelectric points to a single solid phase. Ideally, the
allergen concentration on the solid phase should be in molar excess to the
amount of allergen-specific IgE antibody present in the human serum.
Some commercially-available allergosorbents have also been prepared
with food allergen mixes. For instance, the F × 5E from Pharmacia contains
major allergens from the five foods (chicken egg white, cows’ milk, fish
[Gadus morhua], wheat [Triticum aestivum], peanut [Arachis hypogaea] and
soybean [Glycine max]) that are known to produce most food allergies in
young children. While these food allergen mixes containing allergosorbents
are useful in screening children for confirmation of sensitization to foods,
they are not generally useful in inhibition assays for assessing allergenic
potency of food extracts and thus they are not discussed further in this
chapter.
A different allergosorbent preparation approach has been used in microtiter
plate-based non-isotopic assays. While the inhibition format of ELISAs is
discussed elsewhere in this volume, it is important to note here that adsorption
of proteins onto a plastic microtiter plate surface has severe limitations for
complex protein mixtures such as those present in food extracts. Adsorption
is a random process in which proteins of select isoelectric points and molecular
weights adhere to the plastic surface by weak ionic and hydrophobic
interactions. This adsorption process in allergosorbent preparation functions
best for recombinant or native allergens that have been purified from other
‘irrelevant’ proteineous material in the extract. Unfortunately, it has not been
possible to ensure the binding of all the relevant allergens from crude allergen
extracts, especially those prepared from foods. Therefore, it is recommended
that a covalent coupling method be employed if mixtures of proteins from
foods are being used to prepare the allergosorbent.
Allergen-specific IgE antibody source

The second critical reagent in the competitive allergosorbent inhibition assay
is the allergen-specific human IgE containing serum. Most atopic individuals
produce IgE antibodies to multiple allergen specificities. Only a small number
of these specificities are relevant to any given RAST inhibition assay for
food allergen detection and quantification. For this reason, pools containing
serum from 10–100 individuals who are naturally sensitive to the allergen of
interest are the most useful in competitive inhibition assays. Use of individual
sera in inhibition assays is uncommon (unless units of plasma are available)
because IgE antibody is typically present in ng/ml levels and this generally
requires the use of undiluted serum. The individuals providing serum should
have a history of allergic symptoms following ingestion of the food in question.
Ideally, the subjects should also have a double-blind placebo-controlled food
challenge to verify their sensitivity to the food specificity in question. Use of
serum pools increases the volume of serum available and the probability that
the IgE antibody-containing serum will be representative of the major allergen
specificities relevant to the particular allergen that is being detected in the
inhibition assay. Use of a serum pool also minimizes the possibility of an
over-representation of idiosyncratic minor IgE antibody specificities. These
theoretical advantages are, however, sometimes offset by the fact that mixing
many sera tends to concentrate the most common IgE antibody specificities
and dilute out the minor IgE antibody specificities.
When human serum containing specific IgE antibodies is in short supply,
well-characterized polyclonal antibodies from hyper-immunized animals or
monoclonal antibodies that are known to bind to epitopes on defined food
allergens may be attractive alternatives. These can be used either individually
or in mixtures to replace human IgE containing serum; however, care should
be taken in their use to make sure that these antibodies bind specifically to
allergen epitopes and not just antigenic proteins in the food extract.
Anti-human IgE reagent

Bound IgE antibody in the RAST or its non-isotopic counterpart may be
detected with I125 radioisotope or enzyme-labelled anti-human IgE Fc. A
number of labelled polyclonal anti-human IgE reagents are commercially
available. Alternatively, biotinylated monoclonal antibodies specific for human
IgE Fc such as clone HP6061, murine IgM and clone HP6029, murine IgG1
are available from companies such as EMD Biosciences (La Jolla, CA).
These polyclonal and monoclonal antibodies are usually prepared by
immunization of animals with the Fc fragment of a human IgE myeloma and
screened for IgE reactivity with a second intact IgE myeloma. Specificity for
human IgE needs to be confirmed using well-characterized human
immunoglobulin-related proteins, including IgE, IgG1, IgG2, IgG3, IgG4,
IgM, IgA1, IgA2, IgD and kappa and lambda light chains and secretory
component. The material safety data sheet for the anti-human IgE should
indicate its immunizing antigen source, species, clonality (polyclonal or
monoclonal), pH, preservatives, label type, buffer, shelf-life, optimal

temperature for storage and percentage cross-reactivity to human polyclonal
and monoclonal myeloma proteins in comparison to its reactivity with human
IgE. This can be performed by dilutional analysis using the consensus protocol
from the National Committee on Clinical Laboratory Standards.7
Calibrators, standards and controls

Commercially-available standardized allergen extracts in the US are limited
to short ragweed pollen, Hymenoptera venom, house dust mites and cat hair,
dander or pelt. There are a number of purified food allergen reagents available
from researchers: peanut (Ara h 1, 2 and 3), chicken egg (Gal d 1 [ovomucoid],
Gal d 2 [ovalbumin], Gal d 3 [ovotransferrin] and Gal d 4 [lysozyme]), cows’
milk (α-casein, β-lactoglobulin, α-lactalbumin), Brazil nut (Ber e 1), rice
(Ory s 1: α amylase inhibitor), codfish (Gad c 1 [antigen M]), shrimp
(tropomyosin, Pen a 1, Pen i 1, Met e 1). No standardized crude food allergen
extracts are available for use as reference proteins. Thus, any calibrator that
is to be used in the competitive inhibition assays to study a food extract will
have to have a previously defined potency from researchers in the area. It can
be used to assess the relative potency of new food extracts or to interpolate
competitive inhibition data for purposes of identifying the presence and
quantifying the amount of the food allergen in a test extract.
Preparation of reference allergen calibrators is challenging, especially
when the extracts being testing are prepared from foods with widely different
biomass matrices. For instance, highly purified chicken egg white lysozyme
(Gal d 4) can be weighed out and used to prepare a precise calibrator for a
Gal d 4 allergosorbent-based inhibition assay. However, the matrix into which
the calibrator lysozyme is dissolved must vary as a function of the matrix of
the material being tested. For instance, various lots of red wine, extracts of
cheese and xanthan gums may each be tested for the presence of Gal d 4 that
is occasionally used in processing steps for these foods. However, the biomass
matrix of wine, cheese and 0.25% xanthan gum preparations vary in pH,
lipid, protein and carbohydrate content and general consistency from each
other. Thus each will have a different non-specific binding profile that can
lead to non-parallelism in the inhibition assay unless the calibrator has the
same matrix. This issue is addresed by preparing appropriate calibrators with
known quantities of Gal d 4 in each of the different matrices using predefined
wine, cheese and xanthan gum preparations that are known not to contain
Gal d 4 in any of the processing steps.
4.3.2 Assay design

The general assay protocol involves the addition of 0.05 mL of fluid phase
test or reference allergen (inhibitor) at several three-fold dilutions into their
respective 12 × 75 mm plastic test tubes containing 0.1 mL of the human IgE
antibody serum pool, 0.4 mL of assay buffer (phosphate buffered saline with
1% bovine serum albumin) and 0.5 mL of particulate allergosorbent (typically

1–3% v/v). The tubes are capped and rotated overnight at room temperature.
The following day, the tubes are uncapped and centrifuged (1000 × g,
10 minutes, room temperature) and the supernatant is aspirated. The pelleted
allergosorbent–antibody complexes are washed three times with assay buffer
followed by centrifugation and aspiration of the supernatant. For the RAST
assay, radiolabeled anti-human IgE is added to the allergosorbent–antibody
pellet (0.5 mL, typically 100 000 cpm, immunoreactivity 50–60%). The tubes
are re-capped and rotated (one hour or overnight), uncapped and washed
four times followed each time with centrifugation and supernatant aspiration.
The amount of radiolabeled anti-human IgE bound to the allergosorbent is
determined in an automated gamma spectrometer. Tubes are counted to
statistical accuracy (at least 10 000 counts).
A number of quantitative, second and third generation, non-isotopic RAST-
type assays for allergen-specific IgE antibody have been introduced
commercially. The most widely used assay is the Pharmacia ImmunoCAP
SystemTM that employs a sponge matrix onto which allergenic proteins are
covalently coupled and an enzyme-labelled anti-human IgE to detect bound
IgE antibody.3,4 The CAP SystemTM, like most of the other new generation
RAST-type assays, is calibrated with a total serum IgE dose–response curve
so IgE antibody results are reported in kIU/L, based on a World Health
Organization Total Serum IgE reference preparation. Moreover, the CAPTM
allergosorbents tend to be stable. Most are cleared by the US Food and Drug
Administration; however, there are some that are labelled as ‘analyte-specific
reagents’ because they have not been FDA-cleared due to the unavailability
of sufficient numbers of sera from individuals sensitized to that food. Each
new batch of allergosorbent is quality controlled by the manufacturer with
IgE-containing sera from subjects with documented allergic disease. Prior to
coupling, the food allergen extract is generally subjected to immunochemistry
studies by the manufacturer that involve Western blot analysis. Thus, the
performance of the allergosorbent (and the IgE antibody assay) remains
highly reproducible and potent for most allergen specificities, especially
when testing challenging crude food allergen extracts. For this reason, use of
these commercial allergosorbents and newer generation IgE antibody assays
has become an attractive alternative in food allergen detection and quantification
to the use of research-based particle allergosorbents that have been routinely
used in the past in RAST-type assays.
4.3.3 Data analysis

All analyses are performed in replicates. The extent of non-specific binding
is assessed by the analysis of sera from non-food allergic (atopic and non-
atopic) individuals instead of the food-specific IgE antibody-containing serum
pool. The extent of maximum IgE antibody binding (B0) is determined by the
addition of buffer (0 dose) instead of test or reference allergen extracts
(inhibitor). The percentage of inhibition is computed by subtraction of the

non-specific binding from the total binding obtained with the test or reference
extract inhibitor and this is then divided by B0.
The percent inhibition (linear scale) is plotted on the ordinate versus the
quantity or volume of inhibitor (log10 scale) on the abscissa. Since the test
allergen extract has been analyzed at multiple dilutions, least squares regression
lines for each of the test extract results from 10–90% inhibition should be
plotted. Semi-log plots in this region should be linear. Adolphson, Yunginger
and Gleich21 proposed the following criteria for validity: a minimum of four
inhibition points; these inhibition points must bracket the 50% inhibition
point; the linear correlation coefficient for each curve is > 0.9; and the slope
of the test extract inhibition curve must not differ from that of the standard
inhibition curve at p < 0.01, as determined by a t-test.
Identity of the allergen distribution between two extracts may be surmised
if the inhibition curves are non-overlapping or parallel. Parallel but displaced
curves would suggest identity of the allergen distribution but differences in
allergen quantity (potency). Non-parallel (intersecting) curves suggest partial
identity and differential quantities of allergens. Inhibition curves with a slope
of 0 or a negative slope would suggest no identity or cross-reactivity.
4.3.4 Assay quality control

Quality control is performed by analyzing dilutions of several control extracts
that have previously determined potency estimates. This permits the
computation of a precision profile in which the coefficient of variation (CV)
of the interpolated dose is plotted against the allergen dose. The precision
profile plot may be used to statistically define the dynamic range of the
allergen assay that provides the desired precision (10–15% CV). It can also
be used to determine the sensitivity of the assay or the minimum detectable
concentration of allergen inhibitor that can be reproducibly and statistically
distinguished from the variance of the background noise produced in the
absence of allergen inhibitor. The upper limit of the dynamic working range
of the assay is that concentration of inhibitor that saturates the system and
above which the assay estimates of allergen quantity become increasingly
imprecise.
4.4 Applications
A number of studies provide unique insights into the practical use of the
RAST inhibition assay in the assessment of foods for the presence and
relative quantity of allergens. As an illustration, the RAST inhibition assay
has been used to investigate the allergenicity of commercially-available peanut-
containing foods and to identify peanut contamination of non-peanut derived
foods.19,21,22 In one such protocol, 100 gm of each test product is defatted
once with 250 mL of acetone and then five times with 250 ml of ethyl ether.
The residual, defatted product is then dried and mixed for 20 hours at 25 °C
with 300 mL of 0.1 M ammonium bicarbonate adjusted to pH 8.0. The
extract is clarified by centrifugation of the supernatant for 30 minutes at
24 000 × g at 2 °C and sterile filtration. A total protein is performed using the
Folin phenol reagent described by Lowry et al.23 Only the peanut oil was
tested directly without defatting or extraction.
In these studies, the competitive RAST inhibition analysis was performed
using a human IgE antibody pool comprising sera from multiple peanut
allergic individuals with highly positive peanut specific IgE antibody levels.
The peanut allergosorbent was prepared by coupling crude raw peanut extract
to cyanogen bromide-activated microcrystalline cellulose particles. RAST
inhibitory activity of the test extracts was compared to the competitive inhibition
reference curve produced by a reference peanut extract.
Qualitatively, 17 of the 20 peanut products tested were shown to contain
peanut allergen by the competitive RAST inhibition assay. Two were raw
peanuts and the others were processed foods containing peanut that had
involved shelling, blanching, dry roasting, oil roasting, toasting, grinding,
defatting, extracting and/or combining them with other ingredients. Three
figures have been reproduced from the Nordlee et al. paper19 to illustrate
several points. Figure 4.2 displays similar RAST inhibition curves produced
J I L
STD
K
70 K
J LK
60 J LK
I
J LI L
% Inhibition
50 J L K
I I
J I L
40
J L
30 I K
20 I
J
10
0
0.1 0.6 1.1 1.6 2.1 2.6 3.1 3.6
Log μg
Fig. 4.2 RAST inhibition curves produced by extracts of four peanut butter products.
The percent inhibition of the binding of peanut-specific IgE-containing serum to raw
peanut extract allergosorbent that is induced by the various extracts is plotted as a
function of the log of the protein concentration in the extract in micrograms. Extract I
= peanut butter powder; extract J = peanut butter; extract K = peanut butter syrup;
extract L = peanut butter flavoured chips. The inhibition curves are essentially
parallel, with slopes that are analogous to the RAST inhibition curve slope produced
by the standard raw peanut extract. Only the slope of the peanut butter extract (J)
differs significantly from the standard curve (p < 0.025) and extract J was more potent
than extracts from the other peanut butter products that contained a mixture of peanut
and whey proteins. Reproduced with permission from Nordlee et al.19
O Q M STD
P
M PP N
70 P P
Q P P
60 N
M P N
O QQQ
% Inhibition
M M N N
50 Q
M P
O N
40 Q
Q N
N M
30 O
Q
MO
20
O
10 O
0
0.1 0.6 1.1 1.6 2.1 2.6 3.1 3.6
Log μg
Fig. 4.3 RAST inhibition curves produced with raw and roasted peanut extracts.
Extract M = deflavored peanuts; extract N = raw Virginia peanuts; extract O = oil
roasted Virginia peanuts; extract P = raw Florunner peanuts; extract Q = dry roasted
Florunner peanuts. These curves suggest differences in the allergenic potency among
the different varieties of peanuts. Moreover, other factors are most likely involved
because the inhibition curves produced by extracts of the deflavored peanuts (M) and
dry roasted Florunner peanuts (Q) had approximately the same degree of allergenicity
in comparison to the standard peanut extract. Reproduced with permission from
Nordlee et al.19
with the peanut butter product extracts that appear to have a similar quantity
and specificity distribution of peanut allergens. In contrast, Fig. 4.3 displays
inhibition curves with significantly different slopes suggesting that the extracts
from Florunner peanuts are less allergenic than the deflavored and dry roasted
peanuts. This also may suggest varietal differences in the nature of the allergenic
determinants among the different sources of peanuts. Figure 4.4 illustrates
RAST inhibition curves for peanut oil and the acid hydrolyzed peanuts that
contained no detectable peanut allergen. The negative slope of the inhibition
curve with the peanut oil suggests non-specific interference in the RAST
inhibition assay. In Fig. 4.4, the peanut hull flour contained low levels of
peanut allergen in comparison to the reference peanut extract as evidenced
by the large shift of the inhibition curve (albeit parallel) from the reference
curve. The authors conclude that peanut hypersensitive individuals should
avoid all peanut-containing products with the exception of peanut oil and
possibly acid hydrolyzed peanut where the allergens appear to have been
destroyed. The authors also caution that some non-peanut foodstuffs can
become cross-contaminated with peanut allergens as detectable by competitive
RAST inhibition due to inadequate cleaning of common processing
equipment.21
STD
70
T
60 T
T
% Inhibition
50
T
T
40
T
30
T
R R
20
T S S S
R S R S S
10 R R S
S
R
0 R R
0.1 0.6 1.1 1.6 2.1 2.6 3.1 3.6
Log μg (R is Log μl)
Fig. 4.4 RAST inhibition curves produced by a variety of peanut containing foods
and peanut oil. Extract R = peanut oil (analyzed per microliter volume of oil); S =
hydrolysed peanut protein; extract T = peanut hull flour. The flat or inverse slope of
extracts S and R, respectively indicate that these materials do not contain any
detectable peanut allergen. In contrast, the peanut hull flour extract contains only
small amounts of peanut allergen as evidenced by the parallel slope that is displaced
extensively to the right. Reproduced with permission from Nordlee et al.19.
4.5 RAST inhibition assay strengths and weaknesses

There are a number of assay protocols that employ human IgE antibodies in
a competitive inhibition format to detect and quantify the presence of residual
food allergens in aqueous extracts. These methods differ in minor aspects,
such as in the solid-phase matrix to which the allergen has been coupled
(covalent coupling to particles or nitrocellulose paper versus adsorption to
plastic microtiter plates) and the type of label used in the assay (radiolabel
versus enzyme with colormetric or fluorescent substrates). Each of these
protocols has their strengths and weaknesses.
4.5.1 Strengths
The competitive RAST inhibition’s strength is its ability to detect and quantify
cross-reactivity among allergenic proteins from structurally similar allergens
in different foods and between pollens and food.24 Because allergen-specific
human IgE antibody is used, clinically relevant allergens are detected that
may induce human IgE antibody responses in some clinically sensitive
individuals. The slope and displacement of the test extract inhibition curve
from the reference curve provides useful information about the relative
allergenic potency and specificity distribution of complex protein allergen
mixtures, especially when specific allergens have not yet been identified.
4.5.2 Weaknesses
All human IgE-based methods suffer from a dependence on the unique
allergenic specificity of the IgE antibody that is naturally present in individual
sera or serum pools. Second, inhibition assays that use IgE antibody containing
human sera consume large amounts of serum because the IgE antibody is
normally present in small (ng/mL) quantities. For this reason, IgE antibody-
based inhibition assays are rarely used in food manufacturing facilities where
IgE containing sera are not readily available. Third, the heterogeneity of the
IgE antibody specificity among different sensitized individuals can be a
problem for reproducibility of the RAST inhibition assay. Cross-validation
of serum pools is important to verify consistency of the IgE antibody specificity
and quantity in new serum pools which are being prepared for use. Fourth,
radioisotopes such as I125 in the RAST assay need special care and skill and
specialized laboratory licences ensuring appropriate surveying programs to
permit their handling. Some laboratories prefer to use non-isotopic methods
such as enzymes or fluorophors that are safer and easier to handle and
dispose of after the assay is completed. This has been largely addressed by
the development of commercial RAST-like assays that employ an enzyme
label and colorimetric or fluorescent substrates to detect IgE bound to the
allergosorbent following the serum incubation step.
4.6 Future trends

The competitive inhibition immunoassay format that uses human serum
containing allergen-specific IgE antibody will remain a principal tool for
detecting and quantifying food allergen residues in extracts of test foodstuffs,
especially in cases where the major food allergens have not been identified.
In fact, in the case of food containing fermented ingredients or hydrolyzed
proteins, the competitive inhibition assay may be the only assay that can
adequately detect and quantify residual allergenic proteins. The use of the
human IgE antibody provides an element of specificity to the assay since it
detects clinically relevant allergenic proteins that can elicit allergic symptoms
in sensitized individuals. However, once the major food allergens have been
characterized, polyclonal or monoclonal animal antisera, specific for known
(often cloned and sequenced) allergenic epitopes, will begin to be used in
two-site immunoenzymetric assays to replace the use of human IgE antibody
containing sera in competitive inhibition assays. Future trends will also involve
the increased use of non-isotopic labels coupled to the anti-human IgE reagent,
improved and better characterized reference allergen extracts and the use of
solid phases that increase the binding capacity of allergosorbents.
4.7 References
1. Hamilton, R G and Adkinson, N F Jr (2004) ‘In vitro assays for the diagnosis of IgE
mediated disorders’ J Allergy Clin Immunol, 114, 213–225.
2. Schellenberg, R R and Adkinson, N F Jr (1975) ‘Measurement of absolute amounts
of antigen-specific human IgE by a radioallergosorbent test (RAST) elution technique’,
J Immunol, 115, 1577–1583.
3. Ewan, P W, and Coote, D (1990) ‘Evaluation of a capsulated hydrophilic carrier
polymer (the ImmunoCAP) for measurement of specific IgE antibodies’, Allergy,
45, 22–29.
4. Paganelli, R, Ansotegui, I J, Sastre, J, Lange, C E, Roovers, M H W M, de Groot, H,
Lindholm, N B and Ewan, P W (1998) ‘Specific IgE antibodies in the diagnosis of
atopic disease. Clinical evaluation of a new in vitro test system, UniCAP, in six
European allergy clinics’, Allergy, 53, 763–768.
5. Li, T M, Chuang, T, Tse, S, Hovanec–Burns, D, Said El and Shami, A (2004)
‘Development and validation of a third generation allergen-specific IgE assay on the
continuous random access IMMULITE® 2000 analyzer’, Ann Clin Lab Sci, 34, 67–
74.
6. Food and Agriculture Organization (1995) ‘Report of the FAO technical consultation
on food allergies’, Rome, FAO.
7. Matsson, P, Hamilton, R G, Adkinson, N F, Jr, Esch, R, Homburger, H, Maxim, P
and Williams, P (1998) Evaluation Methods and Analytical Performance Characteristics
of Immunological Assays for Human Immunoglobulin E (Ige) Antibodies of Defined
Allergen Specificities, Vol 16-LA20-A, 19, Wayne, PA, National Committee on Clinical
Laboratory Standards.
8. Turkeltaub, P C, Rostogi, S C, Baer, H, Anderson, M C and Norman, P S (1982) ‘A
standardized quantitative skin-test assay of allergen potency and stability: studies on
the allergen dose-response curve and effect of wheal, erythema, and patient selection
on assay results’, J Allergy Clin Immunol, 70, 343–352.
9. Taylor, S L, Hefle, S L, Bindslev-Jensen, C, Atkins, F M, Andre, C, Bruijnzeel-
Koomen, C, Burks, A W, Bush, R K, Ebisawa, M, Eigenmann, P A, Host, A, Hourihane,
J O, Isolauri, E, Hill, D J, Knulst, A, Lack, G, Sampson, H A, Moneret-Vautrin, D
A, Rance, F, Vadas, P A, Yunginger, J W, Zeiger, R S, Salminen, J W, Madsen, C and
Abbott, P (2004) ‘A consensus protocol for the determination of the threshold dose
for allergenic foods: how much is too much?’ Clin Exp Allergy, 34, 689–695.
10. Sampson, H A and Scanlon, S M (1989) ‘Natural history of food hypersensitivity in
children with atopic dermatitis’, J Pediatr, 115, 23–27.
11. Jansen, J N, Kardinaal, A F M, Huijbers, G, Vlieg-Boerstra, B J, Martens, B P M and
Ockhuizen, T (1994) ‘Prevalence of food allergy and intolerance in the adult Dutch
population’, J Allergy Clin Immunol, 93, 446–456.
12. Richman, P G and Cissel, D S (1988) ‘A procedure for total protein determination
with special application to allergen extract standardization’, J Biol Stand, 16, 225–
238.
13. Breiteneder, H. and Radauer, C (2004) ‘A classification of plant food allergens’, J
14. Blanco, C, Carrillo, T, Castillo, R, Quiralte, J and Cuevas, M (1994) ‘Latex allergy:
clinical features and cross-reactivity with fruits’, Ann Allergy, 73, 309–314.
15. Diaz-Perales, A, Collada, C, Blanco, C, Sanchez-Monge, R, Carrillo, T, Aragoncillo,
C and Salcedo, G (1998) ‘Class I chitinases with hevein-like domains, but not class
II enzymes are relevant chestnut and avocado allergens’, J Allergy Clin Immunol,
102, 127–33.
16. Wide, L, Bennich, H and Johansson, S G O (1967) ‘Diagnosis of allergy by an in
vitro test for allergen-antibodies’, Lancet, 2, 1105–1108.
17. Langeland, T A (1982) ‘A clinical and immunological study of allergy to hen’s egg
white. III. Allergens in hen’s egg white studies by crossed radio-immunoelectrophoresis

(CRIE)’, Allergy, 37, 521–530.
18. Quirce, S, Maranon, F, Umpierrez, A, de las Heras, M, Fernandez-Caldas, E and
Sastre, J (2001) ‘Chicken serum albumin (Gal d 5) is a partially heat-labile inhalant
and food allergen implicated in the bird-egg syndrome’, Allergy, 56, 754–762.
19. Nordlee, J A, Taylor, S L, Jones, R T and Yunginger, J W (1981) ‘Allergenicity of
various peanut products as determined by RAST inhibition’, J Allergy Clin Immunol,
68, 376–82.
20. Hamilton, R G, Wisenauer, J A, Golden, D B, Valentine, M D and Adkinson, N F Jr
(1993) ‘Selection of Hymenoptera venoms for immunotherapy based on patient’s
IgE antibody crossreactivity’, J Allergy Clin Immunol, 92, 651–659.
21. Adolphson, C A, Yunginger, J W and Gleich, G J (1980) ‘Standardization of allergens’,
in Rose, N R and Friedman, H (eds)’ Manual of Clinical Laboratory Immunology,
2nd ed, Washington DC, American Society for Microbiology, 778.
22. Keating, M U, Jones, R T, Worley, N J, Shively, C A and Yunginger, J W (1990)
‘Immunoassay of peanut allergens in food-processing materials and finished foods’,
23. Lowry, O H, Rosenbough, N H, Farr, A L and Randall, R J (1951) ‘Protein measurement
with the Folin phenol reagent’, J Biol Chem, 193, 265–275.
24. Kondo, Y, Tokuda, R, Urisu, A and Matsuda, T (2002) Assessment of cross-reactivity
between Japanese cedar (Cryptomeria japonica) pollen and tomato fruit extracts by
RAST inhibition and immunoblot inhibition’, Clin Exp Allergy, 32, 590–594.
25. Roitt, I, Brostoff, J and Male, D (1997) ‘Hypersensitivity Type I’, in Immunology, 5th
edn, London, Mosby, Inc, 310, Fig. 23.15.
5
Immunoblotting in allergen detection

R. van Ree, J. Akkerdaas and L. Zuidmeer, Sanquin Research,
The Netherlands
5.1 Introduction
Since its introduction in health sciences (Burnette, 1981; Glass, Briggs and
Hnilica, 1981; Karcher et al., 1981; McMichael, Greisiger and Millman,
1981; Reiser and Wardale, 1981; Symington, Green and Brackmann, 1981),
immunoblotting or Western blotting has proven to be a very powerful research
tool and one that has spread rapidly. A search in PubMed using the keywords
‘immunoblotting’, ‘food’ and ‘allergens’ resulted in almost 500 hits, of which
the first go back to the second half of the 1980s (Theobald et al., 1986;
Enberg et al., 1987; Burks, Jr, Brooks and Sampson, 1988; Bush et al., 1988;
Enberg, McCullough and Ownby, 1988; Vallier et al., 1988 Naqpal et al.,
1989;). In most cases, immunoblotting (Kurien and Scofield, 2003) is used
in conjunction with sodium dodecyl sulphate-polyacrylamide gel
electrophoresis (SDS-PAGE) under reducing conditions (Fig. 5.1).
Separation of proteins prior to electrophoretic transfer to nitrocellulose or
polyvinylidene difluoride (PVDF) membranes can also be two-dimensional,
in which case SDS-PAGE is preceded by isoelectric focusing (Celis and
Gromov, 2000; Lilley, Razzaq and Dupree, 2002). Two-dimensional
electrophoresis allows separation of proteins with similar molecular mass
but different pI. In food allergen research, immunoblotting has been used for
many different purposes. Most frequently, the technique has been applied for
qualitative purposes, i.e. the identification of allergenic (IgE-binding) molecules
and the establishment of IgE cross-reactivity between foods or between
foods and inhalant allergens (Bauer et al., 1996; Teuber and Peterson, 1999;
Kazemi-Shirazi et al., 2000; Asturias et al., 2002; Lüttkopf et al., 2002;
Reindl et al., 2002; Miguel-Moncin et al., 2003; Wensing et al., 2003; Bolhaar
Immunoblotting in allergen detection 99
Slot for sample loading
W
M
h
ig
H
W
M
L ow
MW
markers
(a)
Nitrocellulose
Transfer of SDS-PAGE separated proteins to nitrocellulose
(b)
Secondary labeled or
Allergen-specific conjugated antibody
antibody
Contaminating
allergen
Immunodetection of separated proteins on nitrocellulose
(c)
Fig. 5.1 Schematic representation of SDS-PAGE in conjunction with

immunoblotting: (a) Proteins in extracts of food samples are separated by SDS-PAGE
(MW = molecular weight); (b) Separated proteins are electroblotted from the gel onto
nitrocellulose (or PVDF) membrane; (c) By incubation with allergen-specific antibody
reagents, contamination can be traced.
et al., 2004; Comstock et al., 2004). In the former case, serum samples of
food-allergic patients are used to identify proteins that are recognized by IgE
antibodies. A critical issue in such analyses is of course patient selection. In
the latter case, IgE binding to immunoblots is inhibited by allergen extracts
or purified natural or recombinant allergens with suspected cross-reactivity
to the allergens on immunoblot. Immunoblot inhibition with purified allergens
as an inhibitor is a powerful tool to identify or confirm the nature of an
allergen recognized on immunoblot.
Detection of allergenic molecules in food products is another area where
immunoblotting can be applied. Specific monoclonal antibodies (mAb) or
monospecific polyclonal antibodies against the major peanut allergen Ara h
1 could, for example, be used to trace contamination of chocolate products
with peanut. Also human sera from allergic patients have been used for the
detection in chocolate of trace amounts of allergens from, for example,
hazelnut and almond (Scheibe et al., 2001). Prerequisite for such an application
is selective recognition of the food contamination under investigation. Although
antibodies do not necessarily have to be directed to allergens (any protein
will do), we will limit ourselves to allergen-reactive antibody reagents.
5.2 Mono-specific antibody reagents

Which antibodies should be used for detecting allergens in food products?
Most food-allergic patients have IgE antibodies that are far from food-specific,
let alone mono-specific. Even when they are clinically allergic to a single
food, their IgE antibodies almost always cross-react to some other foods.
This property makes IgE antibodies less suitable for specific detection of
traces of allergens in food products. Mouse mAbs against many important
food allergens have been produced but are usually not commercially available
(Kahlert et al., 1992; Tsuji et al., 1993; Burks et al., 1994, 1995; Fahlbusch
et al., 1995; Gonzalez De La Pena et al., 1996; Miyazawa et al., 1996; Reese
et al., 1996; Becker 1997; Bando et al., 1998; Karamloo et al., 2001; Yamashita
et al., 2001; Duffort et al., 2002; Restani et al., 2002; Rozenfeld et al., 2002;
Pomes et al., 2003; Pomes, Vinton and Chapman, 2004). These antibodies
can potentially be used to detect food allergens in extracts of food (products).
An example is given in Fig. 5.2. A mAb raised against the major birch pollen
allergen Bet v 1, mAb 5H8, was found to be cross-reactive to the major
allergen in apple, Mal d 1 (Akkerdaas et al., 1995). Semi-quantitative detection
of Mal d 1 in extract of apple is possible using a standard curve of purified
Mal d 1. Similar approaches can be used for many other food allergens.
Polyclonal antisera raised against purified food allergens can be used
instead of mAbs. Usually rabbits are used for this purpose, but also guinea
pigs, sheep, goats and even chicken are sources for mono-specific antibody
reagents. The advantage of mono-specific polyclonal antisera is that they
usually have higher affinity than mAbs, resulting in a lower detection limit.
1 2 3 4 5 6 7
Fig. 5.2 Semi-quantitative detection of Mal d 1 in apple. Monoclonal antibody 5H8

raised against Bet v 1 was used for the detection of Mal d 1 in extract of Golden
Delicious apple. Lanes 2–7 contain a standard curve of affinity purified Mal d 1
(highest concentration 30 μg/ml; two-fold dilutions), lane 1 an apple extract. On the
basis of this experiment a concentration of ~ 10 μg/ml was assigned to the apple
extract.
1 2
38
28
17
14
Fig. 5.3 Detection of apple LTP (Mal d 3): monoclonals versus polyclonals. Mouse
monoclonal antibodies (lane 1) and rabbit polyclonal antibodies (lane 2) raised against
purified Mal d 3 both exclusively detect the 9 kDa LTP band, demonstrating their
mono-specificity.
In addition, loss of reactivity as a result of processing steps used in highly-

processed foods is more likely for mAbs. Moreover, production of such
antibodies is much cheaper and easier than mAbs. Most commercial enzyme-
linked immunosorbent assay (ELISA) kits for allergenic residue detection
utilize polyclonal antibodies. Of course, polyclonals lack the infinite availability
of mAbs. In Fig. 5.3 rabbit antibodies and mAbs against Mal d 3 have been
used to detect this non-specific lipid transfer protein (LTP) in extract of
apple. Both reagents exclusively detect the same band in apple extract.
5.3 Critical assessment of (mono-)specificity

As described above, mAbs reactive with a major food allergen like Mal d 1
are very useful tools to (semi-)quantify such allergens in extracts of the food
itself, in this case, apple. What if they are used for tracing the presence of a
food (allergen) in compound food products like candy bars, desserts or cookies?
Mal d 1 is member of a large family of so-called pathogenesis-related (PR)
proteins, i.e. the PR10 proteins (Breiteneder and Ebner, 2000). mAb 5H8
was raised by immunization with Bet v 1, the birch pollen representative of
the PR10 protein family, resulting in mAbs that are cross-reactive to Mal d
1 (Akkerdaas et al., 1995). This cross-reactivity is, however, not limited to
apple, but covers a broad spectrum of fruits and nuts (Fig. 5.4). This implies
that detection of a protein band with mAb 5H8 in an extract of a compound
food product establishes the presence of a PR10-like allergen, but its origin
remains unclear due to cross-reactivity of the antibody reagent. Similarly,
rabbit polyclonal antibodies against apple LTP (Mal d 3) recognize LTPs in
several related fruits and nuts.
In the above described examples, cross-reactivity results in detection of
related allergens. IgE antibodies in food-allergic patients tend to have similar
degrees of cross-reactivity. A patient with severe LTP-induced peach and
apple allergy (Fernandez-Rivas et al., 2003) will most likely want to know
whether a food product contains fruit or nut LTP in general, even though not
every LTP will induce clinical allergy (Asero et al., 2002). Antibodies reactive
to a family of related allergens therefore certainly have their value. Polyclonal
antisera can, however, also be (cross-) reactive across the border of protein
1 2 3 4
Fig. 5.4 Detection of Bet v 1 homologues in different fruits. Mouse monoclonal

antibody 5H8 detects Bet v 1 homologues in a broad spectrum of fruits. Lane 1:
control (birch pollen extract), lane 2: sharon fruit extract, lane 3: jackfruit extract, and
lane 4: apple extract.
families. In that case, the interpretation of immunoblot data becomes more

complex. Rabbits or other laboratory animals are not always immune-naïve
when they are immunized to a food allergen, since they may have been
(orally) exposed to foods before. Serum samples taken at the start of an
immunization protocol frequently contain IgG antibodies that react to the
food from which the immunogen originates. If this reactivity causes recognition
of proteins in the same molecular mass range as the immunogen, the antiserum
will have to be affinity-purified using the same immunogen before it can be
used for its specific detection in immunoblotting. Alternatively, binding to
immunoblot can be inhibited by the purified immunogen to confirm specificity
of the interaction.
Another source of disturbing cross-reactivity resides on the immunogen
itself. Immunization of rabbits with plant glycoproteins usually results in an
immune response that is partly directed to plant carbohydrate moieties, in
particular complex N-glycans (Kaladas, Goldberg and Poretz, 1983; Prenner
et al., 1992; Faye et al., 1993; Wilson et al., 1998; Bardor et al., 2003). Plant
complex N-glycans carry two substitutions that are not seen in mammals, a
fucose linked α(1,3) to the proximal N-acetylglucosamine and a xylose linked
β(1, 2) to the core mannose (Lerouge et al., 1998, 2000; Fotisch and Vieths,
2001; van Ree, 2002). These carbohydrate moieties are, therefore, highly
immunogenic in mammals. Since these typically linked fucose and xylose
are found on most plant glycoproteins, antisera raised against them tend to
recognize a broad spectrum of mainly high molecular mass glycoproteins in
virtually any plant food. This is illustrated in Fig. 5.5, where an antiserum
against the major grass pollen allergen Lol p 1 is used in immunoblots of
several foods. There is a relatively simple way to remove these cross-reactive
antibodies that severely lower specificity, namely absorption with a source
rich in plant N-glycans. Essentially, any plant (food) extract coupled to a
solid phase, for example SepharoseTM (GE Healthcare AB, Uppsala, Sweden),
1 2 3 4 5 6 7 1 2 3 4 5 6 7
(a) (b)
Fig. 5.5 Lack of specificity of polyclonal rabbit antisera. Rabbit antisera against two
grass pollen allergens, Lol p 1 (a) and Lol p 12 or profilin (b) detect multiple bands in
plant food extracts. Although anti-Lol p 12 indeed detects profilin (indicated by an
arrow) this antiserum is far from monospecific. The fact that both antisera recognize
very similar (glyco)proteins suggests that these antibodies were already present prior
to immunisation and are directed to highly cross-reactive (carbohydrate?) structures.
can be used for depletion. It is important to choose a plant (food) that is

phylogenetically not too closely related to the source from which the
immunogen originates. This will decrease the risk that antibodies against the
peptide backbone of the immunogen will also be depleted on the basis of
cross-reactivity. Alternatively, protease-digested plant (food) extracts can be
used for depletion. This is efficiently achieved using proteinase K (van Ree
et al., 2000). Finally, purified plant-derived glycoproteins that carry the
typical plant N-glycans are candidates for use as immunosorbent. A good
example of such a glycoprotein is pineapple stem bromelain (Bouwstra et
al., 1990).
5.4 Food processing and antibody specificity

An alternative to depletion of the antiserum is confirmation of the specificity
of the interaction by immunoblot-inhibition with a non-glycosylated
recombinant version of the allergen obtained by expression in E coli. Many
food products undergo processing steps like heating, fermentation or high
pressure before they are marketed. These processes potentially impact on the
integrity of food proteins. It is, for example, well established that PR10
proteins like Mal d 1 lose their allergenicity during processing (Björksten et
al., 1980; Asero et al., 2000). During extraction, polyphenol oxidases and
peroxidases modify the protein in such way that IgE antibodies lose their
reactivity. Subsequent heating will also result in loss of structural integrity
(Breiteneder and Ebner, 2000). In contrast, lipid transfer proteins like Mal d
3 are extremely stable and survive most processing steps in a fully immune-
reactive fashion (Asero, et al., 2000; Scheurer et al., 2004). When using
antibodies for detection of allergens in processed foods, these reagents will
have to be evaluated for their reactivity after processing of food. One could
of course argue that an allergen that loses its reactivity after processing is not
worth being detected because it poses no allergenic risk anymore. It should
be realized though that if detection of an allergen is used as a marker for the
presence of a certain food as an ingredient of a compound food, processing-
sensitive allergens are not the optimal marker proteins. A choice for stable
allergens such as LTP makes more sense. Alternatively, antibodies can be
raised against processed allergens.
5.5 Future trends

Provided that the right precautions are taken considering specificity of antibody
reagents, immunoblotting for the detection of allergens in food will remain
a powerful qualitative and semi-quantitative tool. With the increasing availability
of purified recombinant food allergens, it will become easier to produce
good mono-specific reagents. Disadvantages for a broad application of

immunoblotting in a production setting are the requirement for specialized
equipment and trained staff, and special handling if used in combination
with human serum. For accurate quantitative analysis other techniques such
as sandwich ELISA seem more appropriate. Although the same precautions
apply for antibodies used in ELISA, the technique is more rapid and better
adapted to use in a food processing environment.
5.6 References
Akkerdaas, J H, van Ree, R, Aalbers, M, Stapel, S O and Aalberse, R C (1995) ‘Multiplicity
of cross-reactive epitopes on Bet v I as detected with monoclonal antibodies and
human IgE’, Allergy, 50(3), 215–220.
Asero, R, Mistrello, G, Roncarolo, D, de Vries, S C, Gautier, M F, Ciurana, C L, Verbeek,
E, Mohammadi, T, Knul-Brettlova, V, Akkerdaas, J H, Bulder, I, Aalberse, R C and
van Ree, R (2000) ‘Lipid transfer protein: a pan-allergen in plant-derived foods that
is highly resistant to pepsin digestion’, Int Arch Allergy Immunol, 122(1), 20–32.
Asero, R, Mistrello, G, Roncarolo, D, Amato, S, Caldironi, G, Barocci, F and van Ree, R
(2002) ‘Immunological cross-reactivity between lipid transfer proteins from botanically
unrelated plant-derived foods: a clinical study’, Allergy, 57(10), 900–906.
Asturias, J A, Eraso, E, Arilla, M C, Gomez-Bayon, N, Inacio, F and Martinez, A (2002)
‘Cloning, isolation, and IgE-binding properties of Helix aspersa (brown garden snail)
tropomyosin’, Int Arch Allergy Immunol, 128(2), 90–96.
Bando, N, Tsuji, H, Hiemori, M, Yoshizumi, K, Yamanishi, R, Kimoto, M and Ogawa, T
(1998) ‘Quantitative analysis of Gly m Bd 28K in soybean products by a sandwich
enzyme-linked immunosorbent assay’, J Nutr Sci Vitaminol. (Tokyo), 44(5) 655–664.
Bardor, M, Faveeuw, C, Fitchette, A C, Gilbert, D, Galas, L, Trottein, F, Faye, L and
Lerouge, P (2003) ‘Immunoreactivity in mammals of two typical plant glyco-epitopes,
core alpha(1,3)-fucose and core xylose’, Glycobiology, 13(6), 427–434.
Bauer, L, Ebner, C, Hirschwehr, R, Wüthrich, B, Pichler, C, Fritsch, R, Scheiner, O, and
Kraft, D (1996) ‘IgE cross-reactivity between birch pollen, mugwort pollen and celery
is due to at least three distinct cross-reacting allergens: immunoblot investigation of
the birch-mugwort-celery syndrome’, Clin Exp Allergy, 26(10), 1161–1170.
Becker, W M (1997) ‘Characterization of Ara h 1 by two-dimensional electrophoresis
immunoblot and recombinant techniques: new digestion experiments with peanuts
imitating the gastrointestinal tract’, Int Arch Allergy Immunol, 113(1–3), 118–121.
Björksten, F, Halmepuro, L, Hannuksela, M and Lahti, A (1980) ‘Extraction and properties
of apple allergens’, Allergy, 35(8), 671–677.
Bolhaar, S T, Ree, R, Bruijnzeel-Koomen, C A, Knulst, A C and Zuidmeer, L (2004)
‘Allergy to jackfruit: a novel example of Bet v 1-related food allergy’, Allergy, 59(11),
1187–1192.
Bouwstra, J B, Spoelstra, E C, De Waard, P, Leeflang, B R, Kamerling, J P and Vliegenthart,
J F (1990) ‘Conformational studies on the N-linked carbohydrate chain of bromelain’,
Eur J Biochem, 190(1), 113–122.
Breiteneder, H and Ebner, C (2000) ‘Molecular and biochemical classification of plant-
derived food allergens’, J Allergy Clin Immunol, 106(1 Pt 1), 27–36.
Burks, A W, Jr, Brooks, J R and Sampson, H A (1988) ‘Allergenicity of major component
proteins of soybean determined by enzyme-linked immunosorbent assay (ELISA) and
immunoblotting in children with atopic dermatitis and positive soy challenges’, J
Allergy Clin Immunol, 81(6), 1135–1142.
Burks, A W, Cockrell, G, Connaughton, C and Helm, R M (1994) ‘Epitope specificity and

immunoaffinity purification of the major peanut allergen, Ara h I’, J Allergy Clin
Immunol, 93(4), 743–750.
Burks, A W, Cockrell, G, Connaughton, C, Karpas, A and Helm, R M (1995) ‘Epitope
specificity of the major peanut allergen, Ara h II’, J Allergy Clin Immunol, 95(2), 607–
611.
Burnette, W N (1981) “ ‘Western blotting”: electrophoretic transfer of proteins from
sodium dodecyl sulfate – polyacrylamide gels to unmodified nitrocellulose and
radiographic detection with antibody and radioiodinated protein A’, Anal Biochem,
112(2), 195–203.
Bush, R K, Schroeckenstein, D, Meier-Davis, S, Balmes, J and Rempel, D (1988) ‘Soybean
flour asthma: detection of allergens by immunoblotting’, J Allergy Clin Immunol,
82(2), 251–255.
Celis, J E and Gromov, P (2000) ‘High-resolution two-dimensional gel electrophoresis
and protein identification using western blotting and ECL detection’, EXS, 88 55–67.
Comstock, S S, McGranahan, G, Peterson, W R and Teuber, S S (2004) ‘Extensive in
vitro cross-reactivity to seed storage proteins is present among walnut (Juglans) cultivars
and species’, Clin Exp Allergy, 34(10), 1583–1590.
Duffort, O A, Polo, F, Lombardero, M, Diaz-Perales, A, Sanchez-Monge, R, Garcia-
Casado, G, Salcedo, G and Barber, D (2002) ‘Immunoassay to quantify the major
peach allergen Pru p 3 in foodstuffs. Differential allergen release and stability under
physiological conditions’, J Agric Food Chem, 50(26), 7738–7741.
Enberg, R N, Leickly, F E, McCullough, J, Bailey, J and Ownby, D R (1987) ‘Watermelon
and ragweed share allergens’, J Allergy Clin Immunol, 79(6), 867–875.
Enberg, R N, McCullough, J and Ownby, D R (1988) ‘Antibody responses in watermelon
sensitivity’, J Allergy Clin Immunol, 82(5 Pt 1), 795–800.
Fahlbusch, B, Rudeschko, O, Müller, W D, Schlenvoigt, G, Vettermann, S and Jäger, L
(1995) ‘Purification and characterization of the major allergen from apple and its
allergenic cross-reactivity with Bet v 1’, Int Arch Allergy Immunol, 108(2), 119–126.
Faye, L, Gomord, V., Fitchette-Laine, A C and Chrispeels, M J (1993) ‘Affinity purification
of antibodies specific for Asn-linked glycans containing alpha 1 →3 fucose or beta
1→2 xylose’, Anal Biochem, 209(1), 104–108.
Fernandez-Rivas, M, Gonzalez-Mancebo, E, Rodriguez-Perez, R, Benito, C, Sanchez-
Monge, R, Salcedo, G, Alonso, M D, Rosado, A, Tejedor, M A, Vila, C and Casas, M
L (2003) ‘Clinically relevant peach allergy is related to peach lipid transfer protein,
Pru p 3, in the Spanish population’, J Allergy Clin Immunol, 112(4), 789–795.
Fötisch, K and Vieths, S (2001) ‘N- and O-linked oligosaccharides of allergenic
glycoproteins’, Glycoconj J, 18(5), 373–390.
Glass, W F, Briggs, R C and Hnilica, L S (1981) ‘Identification of tissue-specific nuclear
antigens transferred to nitrocellulose from polyacrylamide gels’, Science, 211(4477),
70–72.
Gonzalez De La Pena, M A, Monsalve, R I, Batanero, E, Villalba, M and Rodriguez, R
(1996) ‘Expression in Escherichia coli of Sin a 1, the major allergen from mustard’,
Eur J Biochem, 237(3), 827–832.
Kahlert, H, Petersen, A, Becker, W M, and Schlaak, M (1992) ‘Epitope analysis of the
allergen ovalbumin (Gal d II) with monoclonal antibodies and patients’ IgE’, Mol
Immunol, 29(10), 1191–1201.
Kaladas, P M, Goldberg, R and Poretz, R D (1983) ‘Rabbit anti-carbohydrate antibody
elicited by the lymphocyte mitogenic glycoprotein from Wistaria floribunda seeds’,
Mol Immunol, 20(7), 727–735.
Karamloo, F, Wangorsch, A, Kasahara, H, Davin, L B, Haustein, D, Lewis, N G and
Vieths, S (2001) ‘Phenylcoumaran benzylic ether and isoflavonoid reductases are a
new class of cross-reactive allergens in birch pollen, fruits and vegetables’, Eur J
Biochem, 268(20), 5310–5320.
Karcher, D, Lowenthal, A, Thormar, H and Noppe, M (1981) ‘Serological identification

of viral antigens after electrophoretic transfer’, J Immunol Meths, 43(2), 175–179.
Kazemi-Shirazi, L, Pauli, G, Purohit, A, Spitzauer, S, Froschl, R, Hoffmann-Sommergruber,
K, Breiteneder, H, Scheiner, O, Kraft, D and Valenta, R (2000) ‘Quantitative IgE
inhibition experiments with purified recombinant allergens indicate pollen-derived
allergens as the sensitizing agents responsible for many forms of plant food allergy’,
J Allergy Clin Immunol, 105(1 Pt 1), 116–125.
Kurien, B T and Scofield, R H (2003) ‘Protein blotting: a review’, J Immunol Meths,
274(1–2), 1–15.
Lerouge, P, Cabanes-Macheteau, M, Rayon, C, Fischette-Laine, A C, Gomord, V and
Faye, L (1998) ‘N-glycoprotein biosynthesis in plants: recent developments and future
trends’, Plant Mol Biol, 38(1–2), 31–48.
Lerouge, P, Bardor, M, Pagny, S, Gomord, V and Faye, L (2000) ‘N-glycosylation of
recombinant pharmaceutical glycoproteins produced in transgenic plants: towards an
humanisation of plant N-glycans’, Curr Pharm Biotechnol, 1(4), 347–354.
Lilley, K S, Razzaq, A and Dupree, P (2002) ‘Two-dimensional gel electrophoresis:
recent advances in sample preparation, detection and quantitation’, Curr Opin Chem
Biol, 61, 46–50.
Lüttkopf, D, Müller, U, Skov, P S, Ballmer-Weber, B K, Wüthrich, B, Skamstrup, H K,
Poulsen, L K, Kastner, M, Haustein, D and Vieths, S (2002) ‘Comparison of four
variants of a major allergen in hazelnut (Corylus avellana) Cor a 1.04 with the major
hazel pollen allergen Cor a 1.01’, Mol Immunol, 38(7), 515–525.
McMichael, J C, Greisiger, L M and Millman, I (1981) ‘The use of nitrocellulose blotting
for the study of hepatitis B surface antigen electrophoresed in agarose gels’, J Immunol
Methods, 45(1), 79–94.
Miguel-Moncin, M, Krail, M, Scheurer, S, Enrique, E, Alonso, R, Conti, A, Cistero-
Bahima, A and Vieths, S (2003) ‘Lettuce anaphylaxis: identification of a lipid transfer
protein as the major allergen’, Allergy, 58(6), 511–517.
Miyazawa, H, Fukamachi, H, Inagaki, Y, Reese, G, Daul, C B, Lehrer, S B, Inouye, S and
Sakaguchi, M (1996) ‘Identification of the first major allergen of a squid (Todarodes
pacificus)’, J Allergy Clin Immunol, 98(5 Pt 1), 948–953.
Naqpal, S, Rajappa, L, Metcalfe, D D and Rao, P V (1989) ‘Isolation and characterization
of heat-stable allergens from shrimp (Penaeus indicus)’, J Allergy Clin.Immunol,
83(1), 26–36.
Pomes, A, Helm, R M, Bannon, G A, Burks, A W, Tsay, A and Chapman, M D (2003)
‘Monitoring peanut allergen in food products by measuring Ara h 1’, J Allergy Clin
Immunol, 111(3), 640–645.
Pomes, A, Vinton, R and Chapman, M D (2004) ‘Peanut allergen (Ara h 1) detection in
foods containing chocolate’, J Food Prot, 67(4), 793–798.
Prenner, C, Mach, L, Glossl, J and März, L (1992) ‘The antigenicity of the carbohydrate
moiety of an insect glycoprotein, honey-bee (Apis mellifera) venom phospholipase
A2. The role of alpha 1,3-fucosylation of the asparagine-bound N-acetylglucosamine’,
Biochem J, 284(Pt 2), 377–380.
Reese, G, Tracey, D, Daul, C B and Lehrer, S B (1996) ‘IgE and monoclonal antibody
reactivities to the major shrimp allergen Pen a 1 (tropomyosin) and vertebrate
tropomyosins’, Adv Exp Med Biol, 409, 225–230.
Reindl, J, Rihs, H P, Scheurer, S, Wangorsch, A, Haustein, D and Vieths, S (2002) ‘IgE
reactivity to profilin in pollen-sensitized subjects with adverse reactions to banana and
pineapple’, Int Arch Allergy Immunol, 128(2), 105–114.
Reiser, J and Wardale, J (1981) ‘Immunological detection of specific proteins in total cell
extracts by fractionation in gels and transfer to diazophenylthioether paper’, Eur J
Biochem, 114(3), 569–575.
Restani, P, Beretta, B, Fiocchi, A, Ballabio, C and Galli, C L (2002) ‘Cross-reactivity
between mammalian proteins’, Ann Allergy Asthma Immunol, 89(6 Suppl 1), 11–15.
Rozenfeld, P, Docena, G H, Anon, M C and Fossati, C A (2002) ‘Detection and identification

of a soy protein component that cross-reacts with caseins from cow’s milk’, Clin Exp
Immunol, 130(1), 49–58.
Scheibe, B, Weiss, W, Rueff, F, Przybilla, B, and Gorg, A (2001) ‘Detection of trace
amounts of hidden allergens: hazelnut and almond proteins in chocolate’, J Chromatogr
B Biomed Sci Appl, 756(1–2), 229–237.
Scheurer, S, Lauer, I, Foetisch, K, Moncin, M S, Retzek, M, Hartz, C, Enrique, E,
Lidholm, J, Cistero-Bahima, A and Vieths, S (2004) ‘Strong allergenicity of Pru av 3,
the lipid transfer protein from cherry, is related to high stability against thermal
processing and digestion’, J Allergy Clin Immunol, 114(4), 900–907.
Symington, J, Green, M and Brackmann, K (1981) ‘Immunoautoradiographic detection
of proteins after electrophoretic transfer from gels to diazo-paper: analysis of adenovirus
encoded proteins’, PNAS, 78(1), 177–181.
Teuber, S S and Peterson, W R (1999) ‘Systemic allergic reaction to coconut (Cocos
nucifera) in 2 subjects with hypersensitivity to tree nut and demonstration of cross-
reactivity to legumin-like seed storage proteins: new coconut and walnut food allergens’,
Theobald, K, Thiel, H, Kallweit, C, Ulmer, W and Konig, W (1986) ‘Detection of proteins
in wheat flour extracts that bind human IgG, IgE, and mouse monoclonal antibodies’,
J Allergy ClinImmunol, 78(3 Pt 1), 470–477.
Tsuji, H, Bando, N, Kimoto, M, Okada, N and Ogawa, T (1993) ‘Preparation and application
of monoclonal antibodies for a sandwich enzyme-linked immunosorbent assay of the
major soybean allergen, Gly m Bd 30K’, J Nutr Sci Vitaminol (Tokyo), 39(4), 389–
397.
Vallier, P, Dechamp, C, Vial, O and Deviller, P (1988) ‘A study of allergens in celery with
cross-sensitivity to mugwort and birch pollens’, Clin Allergy, 18(5), 491–500.
van Ree, R (2002) ‘Carbohydrate epitopes and their relevance for the diagnosis and
treatment of allergic diseases’, Int Arch Allergy Immunol, 129(3), 189–197.
van Ree, R, Cabanes-Macheteau, M, Akkerdaas, J, Milazzo, J P, Loutelier-Bourhis, C,
Rayon, C, Villalba, M, Koppelman, S, Aalberse, R, Rodriguez, R, Faye, L and Lerouge,
P (2000) ‘Beta(1,2)-xylose and alpha(1,3)-fucose residues have a strong contribution
in IgE binding to plant glycoallergens’, J Biol Chem, 275(15), 11451–11458.
Wensing, M, Knulst, A C, Piersma, S, O’Kane, F, Knol, E F, and Koppelman, S J (2003)
‘Patients with anaphylaxis to pea can have peanut allergy caused by cross-reactive
IgE to vicilin (Ara h 1)’, J Allergy Clin Immunol, 111(2), 420–424.
Wilson, I B, Harthill, J E, Mullin, N P, Ashford, D A and Altmann, F (1998) ‘Core
alpha1,3-fucose is a key part of the epitope recognized by antibodies reacting against
plant N-linked oligosaccharides and is present in a wide variety of plant extracts’,
Glycobiology, 8(7), 651–661.
Yamashita, H, Kimoto, M, Hiemori, M, Okita, M, Suzuki, K and Tsuji, H (2001) ‘Sandwich
enzyme-linked immunosorbent assay system for micro-detection of the wheat allergen,
Tri a Bd 17 K’, Biosci Biotechnol Biochem, 65(12), 2730–2734.
6
Enzyme-linked immunosorbent assays

(ELISAs) for detecting allergens in foods
J. Yeung, Food Products Association, USA
6.1 Introduction
Immunoassays such as Enzyme-Linked Immunosorbent Assays (ELISAs)
have been found to have considerable application in clinical diagnostics.
They are in fact the method of choice, but have had thus far little impact on
food analysis. However, since 1990 there have been an increasing number of
reports on the use of ELISA techniques for environmental contaminants in
foods. Unfortunately, only a few have been reported for food allergens.
Until 1988, when Yunginger et al.1 reported a series of eight accidental
deaths due to food-induced anaphylaxis, allergic reactions to food were not
considered a significant public health risk. Subsequent reports of fatalities and
severe allergic reactions2–4 due to food allergies heightened the awareness of
both regulators and the food industry that a small segment of the population
could experience serious reactions to food allergens. With the increased attention
to allergic reactions to food and, especially, allergen-related recalls, there was
a need for quick and accurate methods to detect allergenic residues in food.
Prior to the current emphasis on immunochemical methods for the detection
of allergens, most of the detection used organoleptic techniques, such as
subjective sensory (smell and taste) tests, or chromatographic fatty acid
profile analyses. Unfortunately, these sensory and chromatographic methods
were neither specific to the allergenic proteins, nor sensitive enough to test
for allergen residues. Methods currently used to test for allergens, including
commercially available test kits, are primarily based upon immunochemical
procedures, although some polymerase chain reaction (PCR) kits have also
been developed. These allergen detection methods have recently been reviewed
by Besler5 and Poms et al6; an update is provided in the second part of this
book.
6.2 Principles of ELISAs

Immunochemical methods for the analysis of environmental contaminants
are relatively new in the analytical chemistry arena.7,8 These methods are
based on the use of a specific antibody as a detector for the analyte of
interest, such as a food allergen. In this context, we do not differentiate
between allergen and antigen, and they may be used interchangeably.
Immunoassays are rapid, sensitive and selective, and are generally cost-
effective. Immunoassays can be designed as rapid, field-portable, semi-
quantitative methods, or as standard quantitative laboratory procedures. They
are well suited to the analysis of large numbers of samples, and often do not
require lengthy sample preparation as is the case in chromatography.
ELISA tests are based on the use of an enzyme linked to an antibody to
detect the binding of antigen (Ag) and antibody (Ab). The enzyme converts
a colorless substrate (chromogen) to a colored product, indicating the presence
of Ag:Ab complex. In the food industry, ELISA tests are usually used to
detect antigens such as allergens, pesticides, mycotoxins or pathogens in a
sample. There are two techniques for antigen measurement, the ‘sandwich
technique’ and the ‘competitive technique.’ Almost all commercial allergen
ELISA test kits use the sandwich technique.
6.2.1 Antibodies
Immunoglobulins (as discussed in Chapter 3) are a group of closely related
glycoproteins composed of 82–96% protein and 4–18% carbohydrate. An
antibody is an immunoglobulin, a protein produced by the immune system in
response to the presence of an antigen. Antibodies exist as one or more
copies of a Y-shaped unit, composed of four polypeptide chains. Each Y
contains two identical copies of a ‘heavy chain’ (H), and two identical copies
of a ‘light chain’ (L), named as such by their relative molecular weights,
linked together by inter-chain disulfide bonds (Fig. 6.1a). Intra-chain disulfide
en te bi An
nd tig
ig si
nt ng Variable
in en
g
A di si
n constant te
bi
Light chain
Heavy chain
(a)
Fig. 6.1 (a) Antibody structures. (b) VH and VL regions on an antibody. (c) Pepsin
and papain digestion of an IgG antibody.
Enzyme-linked immunosorbent assays (ELISAs) 111
+H 3N
+H 3N
+H 3N
+H 3N
VH
S
S
S
S
S
S
Lc
ch in
VL
S
H
L
ain
ch ain
ha
C H1
ch
ai
S
n
S
S
Fab
Fa
S
S
b
C
S
L
S
SS
S
S
S
S
Hin
C H2
ge
S
S
reg
H chain
H chain
ion
Fc
S
S
C H3
S
S
COO–
COO–
(b)
Fab
F(ab′)2
Pa
in pa
ps dig in
Pe tio
n es
es tio Fab
d i g IgG n
Peptides
Fc
pFc′
(c)
Fig. 6.1 continued.

bonds are responsible for the formation of loops, leading to the compact,
domain-like structure of the molecule. The amino terminal portions of the
heavy and light chains, characterized by a highly variable amino-acid
composition among different antibodies, are referred to as VH and VL,
respectively. The constant parts of the light chain are designated as CL, while
the constant parts of the heavy chains are further divided into three distinct
subunits: CH1, CH2 and CH3 (Fig. 6.1b). Functionally, the V regions are
involved in antigen binding. The C regions interact to hold the molecule
together and are involved in several biological activities, such as complement
binding and binding to cell membranes.
Antibodies are divided into five major classes: IgM, IgG, IgA, IgD and
IgE, based on their constant region structure and immune function. The most
commonly used antibody is IgG, which can be cleaved into three parts, two
F(ab) regions and one Fc, by the proteolytic enzyme papain, or into two
parts, one F(ab′)2 and one Fc by the proteolytic enzyme pepsin. While both
F(ab) and F(ab′)2 fragments can be used in immunoassays to enhance sensitivity,
F(ab′)2 has higher avidity (Fig. 6.1c). When designing procedures, it is important
to differentiate between monoclonal and polyclonal antibodies, as these
differences are the foundation of both advantages and limitations for their
use.
6.2.2 ELISA formats

Sandwich ELISA
Sandwich ELISA is a sensitive test that can detect and quantify the concentration
of specific soluble proteins. A sandwich ELISA can be more specific because
antibodies directed against two or more distinct epitopes are required. The
basic sandwich ELISA method uses excess of highly purified, specific
antibodies (capture antibodies), which are adsorbed or ‘coated’ onto plastic
microwell plates. The immobilized antibodies serve to specifically capture
their corresponding antigens, such as food allergens, that are present in
samples. After washing away unbound material, the captured antigens are
detected using enzyme-conjugated antibodies (detector antibodies). Following
the addition of a chromogenic substrate-containing solution, the level of
colored product generated by the bound, enzyme-linked antibodies can be
conveniently measured spectrophotometrically using an ELISA-plate reader
at an appropriate wavelength. While sandwich ELISA results can be quantified
against a standard curve, the intensity of the color change is roughly proportional
to the concentration of allergen in the sample, i.e. the more intense the color
change, the higher the amount of allergen in the sample (Fig. 6.2). Alternatively,
qualitative results can be visualized against the standard used.
Standard curves are generally plotted as the standard protein concentration
versus the corresponding mean optical density (OD) value of replicates. The
concentrations of the analyte-containing samples can be interpolated from
the standard curve. This process is made easier by using a computer software
Legend
A Capture antibody
A Antigen or allergen
E Enzyme-linked detector antibody
E S Substrate
S
E
Fig. 6.2 A typical sandwich ELISA.
curve-fitting program, which is either a stand-alone or part of the ELISA

reader operating software. Although a typical dose–response curve or standard
curve is sigmoid shaped, depending on the design of the ELISA, investigators
may choose to apply different curve-fitting analysis to their data, including
linear regression, point-to-point, linear-log, logit log, four-parameter logistic
or cubic spline transformations.7,8
Competitive ELISA
The competitive inhibition ELISA is a technique that uses a one-epitope
approach for the Ab to recognize allergenic residue in a sample. In a competitive
assay, the Ag is coated on the wells, and a solution containing a limited
amount of first Ab along with the Ag or analyte is added. The assay is based
on the principle that an Ag in the sample will bind to an Ab and then compete
for the binding of the Ag coating on the wells. After the unbound Ab is
washed off, a second Ab-enzyme conjugate is used to detect the bound
Ag:Ab complex in the wells. Then a substrate of the enzyme is added (Fig.
6.3). In this format, the color produced is inversely proportional to the
concentration of the analyte, i.e. the higher the color the lower the concentration
of the protein.
A A
Legend
First antibody
A
A
A Antigen or allergen
A A
E Enzyme-linked second antibody
S Substrate
A A
S
E
A A
Fig. 6.3 A typical competitive inhibition ELISA. First antibody reacts with bound
antigen, and a labelled second antibody reacts with the primary antibody.
General characteristics of ELISA platforms

Given the formats described previously the critical components needed to
build a good immunoassay system to test for the presence of the antigen
include: a supply of the appropriate antigen (used to raise antisera and as an
antigen standard for the assay); a supply of high-quality antigen-specific
antisera or antibodies; a stable, high-turnover enzyme detection system; and
a substrate for the reaction. The single most effective way to enhance sensitivity
in ELISA is to find a higher-affinity Ab, although increasing the amount of
captive Ab in a sandwich ELISA may be possible. On the other hand, decreasing
the amount of high-affinity first Ab in a competitive ELISA is an option. For
quantitative ELISA, since the signals of unknown samples are compared to
those of a standard curve, it follows that standards must be run with each
assay to ensure accuracy.
A trained analyst in a modestly equipped laboratory can do a typical
sandwich ELISA test in about 30 minutes–4 hours, while a competitive
inhibitory ELISA takes about 4–8 hours to complete. It should be noted that
it is more complicated to develop a sandwich ELISA than an inhibitory

ELISA. Sandwich ELISA requires a large amount of purified captive Ab and
a successful conjugation of an enzyme to the signal Ab items are not necessary
for inhibitory ELISA. The common buffers used for proteins, including Ab
and Ag, in both ELISA systems are usually phosphate buffer or Tris buffer.
Tris is temperature-sensitive, while phosphate is ionic strength-dependent.
Although a number of enzymes have been used with immunoassays, alkaline
phosphatase and horseradish peroxidase have received the most attention.
The latter is usually preferred because of its low cost. Common substrates
used with peroxidase include o-phenylenediamine (OPD) and 3,3′5,5′-
tetramethylbenzidine (TMB) base. Both substrates require the use of different
filters for reading the respective stopped and non-stopped OD values. Unlike
OPD, TMB is not considered a hazardous material.
Advantages and disadvantages of ELISA

Advantages:
• ease of use, simple, fast, and can be automated
• convenient and standardized 96-well format, which may come in strips of
microwells
• sensitive (in low ppm range)
• selective to the allergenic residues
• availability of labeled reagents
• rapid data reduction
• low initial cost
• portability
Disadvantages:
• lengthy developmental time
• cross-reactivity possible
• matrix effects
• potential false positives from noise or matrix
• confirmation requirement for positives
• no multi-residue analysis yet
• difficult to diagnose problems when assay does not meet the quality
assurance specifications.
6.3 Applications
6.3.1 Methodologies
Numerous immunochemical methods for the detection of allergens have
been described in the literature9–46 and are discussed in detail elsewhere in
the book. These include ELISAs, dipsticks (see Chapter 10), biosensors (see
Chapter 9) and immunoblot (see Chapter 5) and immunoaffinity columns.
6.3.2 Commercial allergenic residue test kits

An allergenic residue test kit is a packaged system of the key components for
detecting or measuring specific allergen or marker proteins in a food matrix
within a laboratory or non-laboratory environment. The key components
include standard calibrators, antibody-coated wells, antibody-enzyme
conjugate, color substrate, stopping solution, extraction buffer and washing
buffer that may be readily prepared by the user of the kit. Test kits include
directions for use and are often self-contained, complete analytical systems,
but they may also require additional supplies and equipment.
Availability
In the mid-1990s, commercial kits came to the market offering laboratories
the opportunity to analyze peanut proteins in processed food products or in
samples taken from various places in a processing facility, and others followed.
Currently, most of the available allergen test kits are ELISA-based methods,
but PCR-type kits are emerging. At the present time, there are five allergen
kit manufacturers (Table 6.1); it is expected that in the near future more
manufacturers will be producing a wider range of allergenic residue kits for
use by the food industry. Since detection limits declared by manufacturers
change in time, especially with the change of regulatory climate, readers are
referred to their respective kit manufacturers for details.
6.3.3 Strengths and weaknesses

Issues with commercial kits
Allergenic residue test kits are simple yet sensitive and specific assays
conveniently packaged in one box. However, a major difficulty facing the
manufacturers of allergenic residue test kits is the extent of assay validation
necessary. The intended use of a test kit is usually not defined for specific
matrices, but rather is given in general terms such as ‘to detect peanuts in
foods.’ However, the potential uses of kits by the food industry, regulatory
agencies and research community are extensive, and they may be used for
tests associated with a variety of complex food matrices. This makes full
validation covering all possible applications virtually impossible. In instances
where a defined market need is known, specific support studies can be
performed, e.g. validation studies on the detection of peanuts in chocolate.
Otherwise, the responsibility for validating the kit in a specific food matrix
is left to the investigator. Furthermore, even if the kit is properly validated,
end users should still do their own in-house evaluation of the limit of detection
(LOD), limit of quantification (LOQ), recovery, precision and ruggedness,
according to AOAC Research Institute recommendations (http://www.aoac.org/
testkits/app_packet_3/Ops.pdf). Users should be aware that immunoassay
kits intended for research, including allergenic residue test kits, do not require
licensing by the US Food and Drug Administration or other national bodies.
Some national bodies are in the process of carrying out validation trials on
the test kits themselves.
Table 6.1 Commercially available kits for food allergen testing (at time of printing)
ELISA r-Biopharm Neogen Pro-Lab Tepnel

Systems Corporation Diagnostics BioSystems
ELISA Almond Almond Almond Peanut Egg

Crustacea B-lactoglobulin Egg Egg Gluten (wheat and rye only)
Egg Egg Gluten (wheat, rye, and Milk (casein, B-lactoglobulin, BSA)
Hazelnut Gluten (wheat, barley) Peanut
Milk (casein and rye, and barley) Milk (casein and whey) Seasame
β-lactoglobulin) Hazelnut Peanut Soy
Peanut Peanut
Sesame
Soy
Dipstick Gluten Peanut Gluten (wheat and rye only)
Peanut
PCR Almond Milk
Hazelnut Peanut
Peanut Soy
Soy
PCR-ELISA Almond
Hazelnut
Peanut
Soy
Contact:
ELISA Systems www.elisasystems.net
Neogen Corporation www.neogen.com
Pro-Lab Diagnostics www.pro-lab.com
r-Biopharm www.r-biopharm.com
Tepnel www.tepnel.com
Other than the newly introduced PCR kits, all five commercial kit
manufacturers use a similar ELISA technology for the detection of allergenic
residues; hence, they all share the same pitfalls. Using peanut as an example,
one manufacturer claims a detection limit of 0.5 ppm peanut proteins while
the other four claim detection limits ranging from 1–5 ppm peanuts. Please
note, the reporting units are not equivalent. Since the LOD can vary greatly
depending on food matrix interferences, extraction efficiency, specificity of
antibodies and variation of peanut protein standards, should be using proper
validation suitable for its intended use is essential. However, none of the
commercial methods have been published in a peer-reviewed journal and
only test kits for peanuts are validated by inter-laboratory collaborative studies
to date, although work on this is ongoing at the time of writing.
In the real world, analytical sensitivity has limited practical value if
uncertainty is not defined, since precision decreases very rapidly as
concentration decreases. A question that remains to be answered is how
sensitive the detection methods should be. It is commonly recognized that
the required analytical sensitivity will vary, depending on the purpose of use
of the assay result. Allergen test results are ultimately used to protect allergic
consumers. In the absence of de minimis threshold levels for food allergens,
sensitivity of allergen methods should aim at detecting any amount of allergenic
food sufficient to elicit objective allergic reactions in allergic individuals
established by double-blind placebo-control food challenge (DBPCFC) studies.
Although allergen threshold levels are finite, and not zero,47 unfortunately,
regulatory guidance on action levels on undeclared allergens is still lacking.
Any positive ELISA result can only be treated as a presumptive positive,
and confirmation or further testing has to be performed. There is an urgent
need for confirmatory tests for the presence of allergenic residues in food.
The only unambiguous confirmatory test for peanut ELISA has demonstrated
the presence of a major peanut allergen, Ara h 1, in ice cream using liquid
chromatography/tandem mass spectrometry (LC/MS/MS). To identify potential
Ara h 1 biomarkers, Shefcheck and Musser48 digested the peanut proteins
with trypsin into their component peptides and then identified the four abundant
peptides that are unique to Ara h 1 as specific peptide biomarkers.
Finally, there is no consensus agreement on acceptable reference materials
by the analytical community, even though some standard reference materials
are commercially available, e.g. National Institute of Standards and Technology
(NIST), https://srmors.nist.gov/pricerpt.cfm; or The Institute for Reference
Materials and Measurements (IRMM), http://www.irmm.jrc.be/. Because
there is no uniform peanut protein standard known, kit manufacturers use
their own peanut protein extracts as standards and as immunogens. As a
result, quantitative results of identical samples will differ when they are
analyzed by different manufacturers’ kits. For example, our data showed that
when 10 ppm of commercial medium roasted peanut flour was added to
melted milk chocolate and mixed to homogeneity, results obtained from four
different commercial kits ranged from 2 to 40 ppm of peanuts or peanut
Table 6.2 Quality control samples analyzed by different commercial peanut kits
Commodities Spikes Tepnel Neogen Pro-Lab r-Biopharm

(ppm flour)
Chocolate 0 nd nd nd nd
10 positive1 26 2.3 9.2
50 positive 207 18.8 27.8
Cookies 0 nd nd nd nd
10 positive 10.9 0.9 4.8
50 positive 160 7.0 13.1
Cereal 0 nd nd nd nd
10 positive 43.5 2.3 6.7
50 positive 171 24.9 44.8
Ice cream 0 nd nd nd nd
10 38.5 40.2 2.0 8.7
50 154 222 7.5 56.0
Detection 1.0 2.5 0.5 2.5
limit (ppm)
nd: below the detection limit

1
As defined by manufacturer at the time of analysis, i.e ≥ 1 ppm
proteins according to the reporting unit of the test kit used. Similar variations
were also observed in different matrices in our quality control samples that
were prepared similarly (Table 6.2). In order to resolve questions regarding
commercial test kit methods, reference standards must be available globally
from a recognized body such as the NIST or IRMM, reporting units need to
be harmonized and collaborative validation studies should be done. Without
reference standards, other issues such as false positives, false negatives,
sensitivity, matrix interference and recovery cannot be adequately assessed.
Sampling
Even when the best analytical methodology and quality assurance schemes
are in place, large errors can be introduced into allergenic residue analysis by
inadequate sampling. This is due, in part, to the inadvertent presence of
allergens being unevenly distributed in solid samples, such as finished food
products. Therefore, obtaining a representative sample is a way of minimizing
false results and increases the chances of accurate determination of allergens
in a batch or lot.
Different organizations/analysts develop different sampling protocols to
ensure that ‘representative’ samples are taken. Some ensure that ‘false negative’
results are minimized; others are more effective at reducing ‘false positive’
results. The ideal sampling plan minimizes the risks associated with both
these errors to protect consumer safety and reduce unnecessary product waste.
Nevertheless, it is important to have effective sampling plans to satisfy both
government and industry acceptable limits.
6.4 Future trends

6.4.1 Rapid methods and new technologies
Automation and rapid methods are dynamic areas of interest for the detection
of allergens in foods, both in the laboratory environment and in food processing
plants. Assays can be automated to reduce hands-on manipulations in
conventional tests, speed up analyses and reduce human errors. The difficulty
in adapting test kits for use in food plants is to enable the assays to be
conducted quickly and accurately under operational conditions and with
non-technical staff. Nevertheless, several rapid methods have been developed
that can be used to analyze allergenic residues with accuracy, simplicity of
operation and acceptable operational speed. Rapid methods usually do not
require a laboratory and results can be obtained on site. This type of procedure
may be vital for allergen screening at remote sites, such as food processing
plants. The results from rapid tests can be displayed immediately and also
archived for documentation. Such tools can be of great benefit to the food
processor, enabling management decisions to be made early in the production
cycle.
Dipstick technology
A promising approach in rapid food allergen detection is the use of a dipstick
format (discussed at length in Chapter 10). Such tests are inexpensive, rapid
and can be done anywhere. They can be used on-site rather than in the
laboratory, hence permitting early detection of potential problems, such as
the ineffective cleaning of equipment. Antibody-based dipsticks (lateral flow
devices) for peanut and gluten are now commercially available. Similar
developments for other allergens are expected to emerge shortly.
Biosensors
Biosensors offer the possibility of in-line detection of allergens. These devices
provide real-time data and require only minimal technical training to operate.
Biosensors are based on the coupling of two components: a sensor chip
containing a bioactive receptor such as an Ab that captures the analyte of
interest, and a transducer that converts the biochemical recognition step into
a quantifiable optical signal.49 For example, surface plasmon resonance (SPR)
biosensors use a direct sensing technique that can detect refractive index
changes that occur in the vicinity of a thin metal film surface where Ab
complexes with Ag. The change in refractive index is proportional to the
concentration of the analyte under investigation. For this type of assay, the
whole process takes less than five minutes. The SPR-based biosensor has
been successfully used for the determination of proteins, mycotoxins, drug
residues, pesticides and peanut proteins.50,51 It should be possible to use SPR
technology for the detection of all food allergens since they are all proteins.
6.5 Conclusions
Immunoassays, such as ELISA, have clearly had a major impact on food
allergen detection, but their potential in other areas of allergen research is
only now just beginning to be realized. ELISA-based methods can help
manufacturers to validate their allergen prevention programs, including
sanitation Standard Operation Procedures (sSOP), to prevent cross contact in
food processing operations and help government agencies to support regulatory
actions.
While allergenic residue testing techniques are continuing to improve
with respect to accuracy, reliability and speed, it should be stressed that their
performance is strongly affected by sampling strategies. It is important to
harmonize internationally accepted sampling protocols, validated
methodologies and approved reference standards in this global economic
climate. Validation and harmonization of quantitative ELISA is needed to
address regulatory compliance for food processing and food services industries
to ensure the effectiveness of their food safety programs.
There is an urgent need for the development of ELISA-based rapid methods,
such as dipsticks or biosensors and multi-residue procedures, which may
provide attractive tools for field monitoring of the integrity of good
manufacturing practices (GMP), whereby non-specialized personnel can
employ them in a cost-effective manner. Such tests can be used on-site rather
than in the laboratory, hence permitting early detection of potential problems,
such as the ineffective cleaning of line equipment before shifting to another
product. This early availability of critical data translates into early management
decisions to avoid the inadvertent introduction of allergens into the production
lines due to cross-contact.
Clearly, ELISA is a very powerful tool in food allergen detection. In fact,
it is the method of choice. The field of ELISA technology shows dynamic
growth driven by growing interest in allergenic residue analyses, and increasing
legislative and regulatory attention to the health risk of food allergens. As a
technical platform, current ELISA technology offers a significant improvement
in versatility and performance advantages over other detection technologies.
6.6 Acknowledgements
The author wishes to thank Regina McDonald, Dr Cecilia Fernandez and Dr
Mara Nogueira for providing some of the peanut kit evaluation data.
6.7 References
1. Yunginger, J W, Sweeney, K G, Sturner, W Q, Giannandrea, L A, Teigland, J D,
Bray, M, Benson, P A, York, J A, Biedzycki, L, Squillance, D L and Helm, R M
(1988) ‘Fatal food-induced anaphylaxis’, JAMA, 260, 1450–1452.

reactions to foods in children and adolescents’, New Engl J Med, 327, 380–384.
3. Bock, S A, Muñoz-Furlong A and Sampson H A (2001) ‘Fatalities due to anaphylactic
4. Wüthrich, B (2000) ‘Lethal or life-threatening allergic reactions to food’, J Investing
Allergol Clin Immunol, 10, 59–65.
5. Besler, M (2001) ‘Determination of allergens in foods’, Trends Anal Chem, 20, 662–
672.
6. Poms, R E, Klein, C L, and Anklam, E (2004) ‘Methods for allergen analysis in
food: a review’, Food Addit Contam, 21, 1–31.
7. Kurtz, D A, Skerrit, J H and Stanker, L ed. (1995) New Frontiers in Agrochemical
Immunoassay, Arlington, Virginia, AOAC International.
8. Wild, D ed. (2001) The Immunoassay Handbook, 2nd edn, New York, NY, Nature
Publishing Group.
9. Hefle, S L, Folgert, J P, Bush, R, and Chu, F S (1994) ‘Monoclonal antibodies
against selected peanut allergens: production and use as affinity agents’, Food Agric
Immunol, 6,197–208.
10. Hefle, S L, Bush, R K, Yunginger, J W and Chu, F S (1994) ‘A sandwich enzyme-
linked immunosorbent assay (ELISA) for the quantification of selected peanut proteins
in foods’, J Food Prot, 57, 419–423.
11. Yeung, J M and Collins, P G (1996) ‘Enzyme immunoassay for determination of
peanut proteins in food products’, J AOAC, 79, 1411–1416.
12. Holzhauser, T and Vieths, S (1999) ‘Indirect competitive ELISA for determination
of traces of peanut (Arachis hypogoea L.) protein in complex food matrices’, J Agric
Food Chem, 47, 603–611.
13. Clare Mills, E N, Potts, A, Plumb, G W, Lambert, N and Morgan, M R A (1997)
‘Development of a rapid dipstick immunoassay for the detection of peanut
contamination of food’, Food Agric Immunol, 9, 37–50.
14. Stephan, O, Moller, N, Lehmann, S, Holzhauser, T and Vieths, S (2002) ‘Development
and validation of two dipstick type immunoassays for determination of trace amounts
of peanut and hazelnut in processed foods’, Eur Food Res Technol, 215, 431–436.
15. Newsome, W H and Abbott, M (1999) ‘An immunoaffinity column for the determination
of peanut proteins in chocolate’, J AOAC, 82, 666–668.
16. Blais, B W and Phillippe, L M (2000) ‘A cloth-based immunoassy for detection of
peanut proteins in foods’, Food Agric Immunol, 12, 243–248.
17. Pomés, A, Helm R M, Bannon, G A, Burks, A W, Tsay, A and Chapman, M D (2003)
Immunol, 111, 640–645.
18. Yeung, J M, Newsome, W H and Abbott, M A (2000) ‘Development of an enzyme
immunoassay for the determination of egg proteins in food products’, J AOAC Int,
83, 139–143.
19. Hefle, S L, Jeanniton, E and Taylor, S L (2001) ‘Development of a sandwich enzyme-
linked immunoassay for the detection of egg residues in processed foods’, J Food
Prot, 64, 1812–1816.
20. Baumgartner, S, Steiner, I, Hirmann, D, Kloiber, S, Krska, R and Yeung, J (2002)
‘Towards the development of a dipstick immunoassay for the detection of trace
amounts of egg proteins in food’, Eur Food Res Technol, 214, 168–170.
21. Mariager, B, Solve, M, Eriksen, H and Brogren, C-H (1994) ‘Bovine ß-lactoglobulin
in hypoallergenic and ordinary infant formulas measured by indirect competitive
ELISA using monoclonal and polyclonal antibodies’, Food Agric Immunol, 6, 73–
83.
22. Zahur, U H and Pruett, S B (1993) ‘Development of an enzyme linked immunoassay
for ß-lactoglobulin in dairy products’, Cul Dairy Prod J, February, 23–26.
23. Mäkinen-Kiljunen, S and Palosuo, T (1992) ‘A sensitive enzyme-linked immunosorbent
assay for detection of bovine ß-lactoglobulin in infant feeding formulas and in

human milk’, Allergy, 47, 347–352.
24. Negroni, L, Bernard, H, Clement, G, Chatel, J M, Brune, P, Frobert, Y, Wal, J M and
Grassi, J (1998) ‘Two-site enzyme immunosorbent assays for determination of native
and denatured ß-lactoglobulin’, J Immunol Methods, 220, 25–37.
25. Plebani, A, Restani, P, Naselli, A, Galli, C L, Meini, A, Cavagni, G, Ugazio, A G and
Poiesi, C. (1997) ‘Monoclonal and polyclonal antibodies against casein components
of cow milk for evaluation of residual antigenic activity in ‘hypoallergenic’ infant
formulas’, Clin Exp Allergy, 27, 949–956.
26. Anguita, G, Martin, R, Garcia, T, Morales, P, Haza, A I, Gonzalez, I, Sanz, B and
Hernandez, P E. (1997) ‘A competitive enzyme-linked immunosorbent assay for
detection of bovine milk in ovine and caprine milk and cheese using a monoclonal
antibody against bovine beta-casein’, J Food Prot, 60, 64–66.
27. Brandon, D L and Friedman, M (2002) ‘Immunoassays of soy proteins’, J Agric
Food Chem, 50, 6635–6642.
28. McNeal, J E (1988) ‘Semiquantitative enzyme-linked immunosorbent assay of soy
protein in meat products: summary of collaborative study’, J AOAC, 71, 443.
29. Yeung, J M and Collins, P G (1997) ‘Determination of soy proteins in food products
by enzyme immunoassay’, Food Tech Biotech, Special Issuue: Immunochemical
Techniques and Methods, 36, 209–214.
30. Bando, N, Tsuji, H, Hiemori, M, Yoshizumi, K, Yamanishi, R, Kimoto, M and
Ogawa, T (1998) Quantitative analysis of Gly m Bd 28K in soybean products by a
sandwich-linked immunosorbent assay’, J Nutr Sci Vitaminol, 44, 655–664.
31. Haasnoot, W, Olieman, K, Cazemier, G and Verheijen, R (2001) ‘Direct biosensor
immunoassays for the detection of nonmilk proteins in milk powder’, J Agric Food
Chem, 49, 5201–5206.
32. Skerrit, J H and Hill, A S (1990) ‘Monoclonal antibody sandwich enzyme immunoassays
for detection of gluten in foods’, J Agric Food Chem, 38, 1771–1778.
33. Skerrit, J H and Hill, A S (1991) ‘Enzyme immunoassays for detection of gluten in
foods: collaborative study’, J AOAC, 74, 257–264.
34. Sorell, L, Lopez, J A, Valdes, I, Alfonso, P, Camafeita, E, Acevedo, B, Chirdo, F,
Gavilondo, J and Mendez, E (1998) ‘An innovative sandwich ELISA system based
on an antibody cocktail for gluten analysis’, FEBS Lett, 439, 46–50.
35. Garcia, E, Llorente, M, Wieser, H, Mothes, T, Dermuhl, L, Immer, U, Thorell, L,
Vasallo, I and Méndez, E (2002) ‘Mini ring test with test kit prototypes based on a
specific monoclonal antibody’, in Stern M, Proceeding of the 16th Meeting: Working
Group on Prolamin Analysis and Toxicity, 8–11 November, 2001, Sitges (Barcelona),
Spain, Verlag Wissenchaftilche Scripten, Zwickau, Germany.
36. Koppelman, S J, Knulst, A C, Koers, W J, Penninks, A H, Peppelman, H, Vlooswijk,
R, Pigmans, I, van Duijn, G and Hessing, M (1999) ‘Comparison of different
immunochemical methods for the detection and quantification of hazelnut proteins
in food products’, J Immunol Methods, 229, 107–120.
37. Holzhauser, T and Vieths, S (1999) ‘Quantitative sandwich ELISA for determination
of traces of hazelnut (Corylus avellana) protein in complex food matrixes’, J Agric
Food Chem, 47, 4209–18.
38. Holzhauser, T, Stephan, O and Vieths, S (2002) ‘Detection of potentially allergenic
hazelnut (Corylus avellana) residues in food: a comparative study with DNA PCR-
ELISA and protein sandwich-ELISA’, J Agric Food Chem, 50, 5808–15.
39. Blais, B W and Phillippe, L (2001) ‘Detection of hazelnut proteins in foods by
enzyme immunoassay using egg yolk antibodies’, J Food Prot, 64, 895–8.
40. Blais, B W, Omar, M and Phillippe, L (2002) ‘Detection of Brazil nut proteins in
foods by enzyme immunoassay’, Food Agric Immunol, 14, 163–168.
41. Scheibe, B, Weiss, W, Rueff, F, Przybilla, B and Gorg, A (2001) ‘Detection of trace
B Biomed Sci Appl, 756, 229–37.
42. Hlywka, J J, Hefle, S L and Taylor, S L (2000) ‘A sandwich enzyme-linked

immunosorbent assay for the detection of almonds in foods’, J Food Prot, 63, 252–
257.
43. Roux, K H, Teuber, S S, Robotham, J M and Sathe, S K (2001) ‘Detection and
stability of the major almond allergen in foods’, J Agric Food Chem, 49, 2131–2136.
44. Acosta, M R, Roux, K H, Teuber, S S and Sathe, S K (1999) ‘Production and
characterization of rabbit polyclonal antibodies to almond (Prunus dulcis L.) major
storage protein’, J Agric Food Chem, 47, 4053–4059.
45. Jeoung, B J, Reese, G, Hauck, P, Oliver, J B, Daul, C B and Lehrer, S B (1997)
‘Quantification of the major brown shrimp allergen Pen a 1 (tropomyosin) by a
monoclonal antibody-based sandwich ELISA’, J Allergy Clin Immunol, 100, 229–
234.
46. Taylor, W J and Jones, J L (1992) ‘An immunoassay for verifying the identity of
canned sardines’, Food Agric Immunol, 4, 169–175.
47. Taylor, S L, Heflé, S L, Bindslev-Jensen, C, Bock, S A, Burks, A W Jr, Christie, L,
Hill, D J, Host, A, Hourihane, J O, Lack, G, Metcalfe, D D, Moneret-Vautrin, D A,
Vadas, P A, Rance, F, Skrypec, D J, Trautman, T A, Yman, I M and Zeiger, R S
(2002) ‘Factors affecting the determination of threshold doses for allergenic foods:
how much is too much?’ J Allergy Clin Immunol, 109, 24–30.
48. Shefcheck, K J and Musser, S M (2004) ‘Confirmation of the allergenic peanut
protein, Ara h 1, in a model food matrix using liquid chromatography/tandem mass
spectrometry (LC/MS/MS)’, J Agric Food Chem, 52, 2785–2790.
49. Patel, P D (2002) ‘(Bio)sensors for measurement of analytes implicated in food
safety: a review’, Trends Anal Chem, 21, 96–115.
50. Mohammed, I, Mullett, W M, Lai, E P C and Yeung, J M (2001) ‘Is biosensor a
viable method for food allergen detection? Anal Chim Acta, 444, 97–102.
51. Mullett, W, Lai, E P C and Yeung, J M (2000) ‘Surface plasmon resonance-based
immunoassays’, Methods, 22, 77–91.
7
Polymerase chain reaction (PCR)

methods for the detection of allergenic
foods
T. Holzhauser, O. Stephan and S. Vieths, Paul-Ehrlich-Institut,
Germany
7.1 Introduction
For verification of allergen labeling of foods and in order to identify hidden
allergens in processed food it is important to provide analytical methods that
are able to detect very low amounts of allergenic residue in processed foods.
In general, these methods are based on the detection of species-specific
proteins by enzyme-linked immunoassay (ELISA).1–8 In addition, the detection
of species-specific DNA molecules by the polymerase chain reaction
(PCR) for the identification of allergenic foods at trace levels has been
demonstrated.9–18
PCR technology has already been established as the DNA-based method
for the identification of genetically modified organisms, pathogens and food-
related plant and animal species. Furthermore, PCR has been widely used for
many approaches in the field of molecular biology and routinely applied in
clinical chemistry. Since the mid 1990s this methodology has attracted
increasing attention for the detection of allergen traces in food products; this
is reflected by an increasing number of published and commercially available
PCR assays (summarized by Poms et al.19). This chapter will review published
data on the applicability of PCR as a possible alternative to, and as a tool for
the verification of, ELISA methods and discuss the potential and restrictions
of PCR used in the detection of allergenic foods.
7.2 PCR principles

Each PCR analysis comprises three consecutive steps: the DNA extraction
and purification, the amplification of a specific DNA sequence, and the
detection of the amplified PCR products (amplicons). This section highlights

the particular steps in PCR with special regard to the detection of allergenic
foods.
7.2.1 DNA extraction and purification

For analytical methods that detect hidden allergens, a sensitivity of 10 ppm
(mg/kg) or less of the allergenic food in a composed food has been suggested.20
The need for detection of allergenic foods at this level is supported by
double-blind placebo-controlled food challenge (DBPCFC) studies that have
shown that as little as 100 μg of peanut can elicit a mild allergic reaction in
peanut-sensitive individuals.21,22 To reach an appropriate sensitivity in PCR,
the DNA extraction method is extremely important because an insufficient
purification procedure can drastically reduce the assay sensitivity. A variety
of compounds in animal and plant tissues, such as proteases and polyphenols,
can slow down or even inhibit the amplification reaction in PCR. Hence,
DNA preparations of high purity are necessary. A common and approved
method for the extraction of PCR-grade DNA is a combination of the CTAB
(cetyltrimethylammonium bromide) or the GTC (guanidine thiocyanate)
protocol and a subsequent purification of the extracted DNA with commercially
available silica columns.23,24 An alternative to this relatively time-consuming
method is the use of commercially available kits that usually provide a silica-
based DNA extraction and are optimized for the analysis of processed foods.
Figure 7.1 displays the principle of a silica-based DNA extraction and
purification.
7.2.2 General principle of PCR

The basis for PCR is a thermo-stable DNA polymerase which is able to
amplify a specific DNA fragment that is flanked by two specific
oligonucleotides (primers). In most applications the Taq DNA polymerase
from Thermus aquaticus is used. In general, the PCR is a temperature-
dependent reaction that consists of a series of 25–45 cycles with each cycle
consisting of three phases, the denaturation, the annealing, and the extension
phase (Figure 7.2). All phases are performed in one reaction vial at defined
temperatures.
During the denaturation phase the reaction temperature is increased to
94–96 °C for 30–60 seconds to melt double-stranded DNA into single strands
by breaking up hydrogen bonds. During the annealing phase the temperature
is decreased to 45–65 °C for 30–60 seconds depending on the melting
temperature of the primers. In this phase the primers hybridize to their
complementary sequences on the single-stranded template DNA. A suitable
annealing temperature depends on the length and the nucleotide composition
of the primers (for details, see Newton and Graham25). A wrong temperature
during the annealing process can result in the occurrence of unspecific PCR
Polymerase chain reaction (PCR) methods 127
1. Sample lysis/ 2. DNA binding to

DNA extraction silica membrane
Sample lysis Silica spin filter

buffer, heat centrifuge
3. DNA purification 4. DNA elution from

on silica membrane silica membrane
Wash buffer Elution buffer

centrifuge centrifuge
Fig. 7.1 DNA extraction and purification based on DNA-binding silica membranes:
characteristic steps in DNA-isolation and necessary reagents and consumables.
(a) Cycle 1: Denaturation phase (b) Cycle 1: Annealing phase
95°C 45–65°C
Double-stranded
DNA Single-stranded
DNA
= Primer
(c) Cycle 1: Extension phase (d) Cycle 1–n: DNA amplification

Start Cycle 1 Cycle 2 Cycle n
72°C
2n
= Taq-polymerase = dNTPs
Fig. 7.2 Principle of the PCR. After heat denaturation of the genomic DNA (a),
sequence-specific primers bind to the single-stranded DNA during the annealing phase
(b). Subsequently, the Taq DNA polymerase extends the primers by filling up the
missing strand with desoxynucleotidetriphosphates (c). The exponential amplification
of the PCR products is displayed in (d).
products if the annealing temperature is too low, or in a failure of the whole

reaction if the temperature is too high. The extension phase is performed at
72 °C, the optimum temperature for the enzymatic activity of the Taq DNA
polymerase. During this step the polymerase binds at the 3′-OH end of the
primers and fills up the missing strands with desoxynucleotidetriphosphates
(dNTPs) which are part of the PCR reaction mix. With a well-optimized and
highly efficient PCR, the amount of the generated PCR product is ideally
doubled within each cycle which results in an exponential amplification.25,26
Subsequently, the PCR product can be visualized and its sequence verified as
described in Section 7.2.3.
7.2.3 Detection of PCR products

With special regard to the verification of the amplified sequence, the commonly
used size-dependent detection of the PCR product by agarose-gel
electrophoresis is not sequence-specific and may cause false positive results
as described below. Therefore, the German Institute for Standardization (DIN)
and the American Society for Testing and Materials (ASTM) have emphasized
the sequence verification of a PCR product if used for analytical purposes.
This can be done either by Southern blotting with sequence-specific
hybridization probes, by nucleotide sequencing, or with endonuclease restriction
cuts.27,28 However, these methods are very time-consuming and can be
overcome by the use of a real-time PCR or a PCR-ELISA if the sequence of
the generated PCR product is verified by a detectable probe that binds sequence-
specific to the amplified sequence.
Agarose gel electrophoresis

The classical method for the detection of amplified PCR products is agarose-
gel electrophoresis that provides a size-dependent separation of the negatively
charged DNA molecules in an electric field where small amplicons migrate
faster than larger ones. After electrophoresis the gel is stained with a fluorescent
dye, e.g. ethidium bromide or Sybr®Green (Molecular Probes Inc, Eugene,
OR), that binds to nucleic acids, and the PCR products can be visualized on
a UV transilluminator. The amplicon size can be roughly determined by
comparison with a marker DNA that contains DNA fragments of defined
sizes. For analytical approaches such as the detection of hidden allergens in
food, it is advisable not to use this procedure as the only method of detection
because false positive results can be obtained from PCR products of comparable
size that may be caused by a cross-reactivity. Consequently, the identity of
the PCR products should be verified by a sequence-specific method. Another
disadvantage of agarose-gel electrophoresis is the use of fluorescent dyes
that are potentially mutagenic because of their ability to intercalate with
DNA.
Detection of PCR products by PCR–ELISA

After PCR amplification, the products are detected sequence-specific with
an ELISA-like technique. This method requires a modification of the PCR
protocol. Several protocols are available of which the following is an example
of a recently published hazelnut-specific PCR–ELISA.13 Biotinylated primers
are used in the PCR which leads to biotinylated PCR products that are
coupled onto a streptavidin-coated microtiterplate. Subsequently, the PCR
product is denatured resulting in a surface-bound single-stranded DNA that
is accessible for the binding of a sequence-specific DNA probe. The probe,
which is linked to a label is detected by an enzyme-conjugated antibody that
is directed against the label. A colorimetric reaction is catalyzed by the
enzyme after the addition of a chromogenic substrate. The light absorption
of the developed color product is measured in a microplate reader. The
principle of this post-PCR detection of amplicons is shown in Figure 7.3.
Even though the ELISA-like detection of amplicons can be used in a quantitative
manner, the PCR–ELISA can only be used as a qualitative test for the absence
or presence of the allergenic food within the validated range of detectability.
The reason for this is the qualitative nature of the PCR as an endpoint
method.
In comparison to the agarose-gel electrophoresis, a higher sample throughput
can be achieved with the PCR–ELISA, and the amplicon sequence is verified
with a sequence-specific probe. Further benefits of the PCR–ELISA are the
avoidance of mutagenic dyes such as ethidium bromide and cheaper equipment
compared to real-time PCR applications. As in agarose-gel electrophoresis,
the PCR–ELISA makes it necessary to perform manipulations with amplified
PCR products which can lead to DNA contaminations and therefore to the
occurrence of false positive results in future PCR experiments, especially if
adequate safety precautions are not complied with (see page 133).
Detection of PCR products in real-time

The principle of this state-of-the-art method for quantitative detection of
amplifiable DNA is based on the measurement of a fluorescent signal that is
increased during the amplification of PCR products. For sequence-specific
real-time detection of PCR products, a fluorogenic probe that is designed to
bind to the PCR product is used. Several probe formats are described such as
‘TaqMan’™ Probes (Roche Molecular Systems Inc, Pleasanton, CA), ‘LC
Hybridization®’ Probes (Roche Diagnostics, Indianapolis, IN, and Idaho
Technology Inc, Salt Lake City, UT) ‘Scorpions’® (DxS Ltd, Manchester,
UK) and ‘Molecular Beacons’. Most of the commonly used probe formats
are based on a physical process called fluorescence resonance energy transfer
(FRET). Details of the various probe formats and their characteristics are
given elsewhere.29 Due to the proportional relation between the fluorescent
signal and the amount of the generated PCR product it is possible to record
a PCR amplification plot. Based on the so-called threshold cycle (Ct), the
PCR cycle number at which the fluorescent signal can be distinguished from
(a) Amplicon-binding to the solid phase (b) Denaturation of double-stranded amplicons
Biotinylated
PCR amplicon
Single-stranded
amplicon bound
at the solid
Streptavidin phase
Microplate well Microplate well
(c) Probe hybridization (d) Probe detection
Sequence-specific Enzyme-
labeled probe conjugated
antibody
Enzyme
substrate
Microplate well Microplate well
Fig. 7.3 Principle of the PCR-ELISA. Biotinylated PCR amplicons are bound to streptavidin-coated wells in a microtiterplate (a). After
denaturation of the double-stranded amplicons (b) a sequence-specific labelled probe can hybridize (c). The detection of the probe is done by an
enzyme-conjugated antibody specific for the probe label (d). The color development is measured as absorbance in a microplate reader.
the background noise, it is possible to quantify the number of DNA copies

present in the sample before amplification. For the determination of the
DNA copy number in the unknown sample, the Ct value from a sample is
compared with the Ct values from a DNA standard calibration curve, similar
to the protein standards in ELISA tests (Figure 7.4). In the case of the
quantitative analysis of allergenic foods in composed food products, the
availability of standard reference materials is essential to cover the influence
of matrix and processing effects.
3.5
3.0
2.5
Fluorescence
2.0
1.5
1.0
0.5
Threshold
0.0
–0.5
10 20 30 40 50 60
Cycle number
(a)
45
40 y = –3.9647x + 42.13
R2 = 0.985
35
Ct-value
30
25
20
100 101 102 103 104 105
Copy number
(b)
Fig. 7.4 Real-time PCR. (a) Amplification plots of serially diluted DNA. (b) Cycle
threshold (Ct) values of the amplification plots from (a) plotted against the logarithm
of the copy numbers of amplifiable DNA prior to PCR analysis. These result in linear
correlation.
Besides the possibility of DNA quantification, real-time PCR methods do

not need further post-PCR steps, which overcomes problems encountered
with post-PCR contaminations. However, it has to be considered that the
equipment for this PCR technology is still expensive (> $30 000). For the
quantification by PCR, competitive PCR assays with amplicon detection in
agarose gels have also been applied. An example is the quantification of
wheat, barley, and rye in gluten-free food for coeliac patients.14 However,
because of easier handling, a lower risk of post-PCR contamination, and a
larger dynamic range of quantification, real-time PCR assays favoured.
7.2.4 Particular considerations in PCR

Choice of primers
The choice of the target gene and the design of the primers have a great
impact on the sensitivity and the specificity of a PCR method. DNA is often
degraded in processed food. As a consequence, for PCR to detect hidden
allergens, the amplified DNA fragment that is defined by the position of
primers on the target gene should have a relatively small size of the some
60–150 basepairs.28 Moreover, the primer pair needs to be tested for possible
cross-reactivity with the most important food matrices and with all relevant
food ingredients that are related to the analyte. For example, a peanut-specific
PCR should not detect relevant legumes, nuts, cereals, milk, and cocoa to
allow analysis of chocolate, cereals, and cookies.
Furthermore, very sensitive PCR assays can be established if the primer
target is a multicopy gene (e.g. mitochondrial genes, chloroplast genes) as
was described for the quantitative detection of wheat, barley, and rye in
gluten-free foods.14 However, due to highly conserved nucleotide sequences
of these genes in different species, strong cross-reactions can occur.15 For
this reason, most of the PCR assays for identification of allergenic foods
have been based on the amplification of the more characteristic, and thus
more specific, single-copy genes. With regard to the detection of allergenic
foods, a detection of the gene encoding an allergen is not necessary. Any
gene that allows an unequivocal identification of the allergenic food species
can be used as an indicative marker.
No matter which gene sequence is chosen for amplification, it is essential
that the primers bind specifically to the target sequence to avoid unspecific
by-products and a reduction of PCR efficiency. Hence, the primers and
probes have to be designed and optimized to meet these needs. Several
rules for primer and probe design have to be considered.25,26 Briefly, the
annealing temperature of the primers should be within the range 45–65 °C
and the content of the pyrimidine bases guanine and cytosine should be
between 35% and 65%. Self-complementary sequences should be avoided,
especially in the 3′ region of the primers, to avoid the formation of hairpins
or primer dimers that can drastically reduce the efficiency of the amplification
reaction.
Contamination and how to avoid it

The very high sensitivity of the PCR is one of the major advantages of this
method, but at the same time it is a great challenge because of the risk of
persistent DNA contamination that may lead to false positive results in future
PCR experiments. Theoretically, one copy of amplifiable DNA can be
multiplied and detected. For this reason, negative controls are extremely
important. Post-PCR steps are, in particular, very problematic. Significant
copy numbers of amplified DNA fragments can be dispersed in the laboratory
due to the formation of aerosols when handling the PCR products.
Consequently, post-PCR steps need to be performed in a dedicated room that
is separated from the other PCR working areas. To minimize the risk of
carry-over contamination, further precautions, such as the use of separate
pipettes, aliquoting of reagents, use of filter-tips, and frequent changing of
gloves, should be considered. 30 Moreover, the use of a chemical
decontamination reaction during each PCR run helps to avoid carry-over
contamination. This system is based on an enzyme called uracil-DNA-
glycosylase (UNG) that catalyzes the removal of uracil from single- and
double-stranded DNA. If dTTP is substituted by dUTP in the dNTP mix,
amplicons that contain uracil will be generated in PCR. If amplicons that
contain uracil are carried over into a following amplification reaction, these
amplicons will be degraded by UNG in contrast to the natural DNA template
that contains thymidine.31
7.3 Application of PCR for the detection of allergenic foods

7.3.1 Published PCR applications
So far five PCR assays for the detection of hidden hazelnut and peanut traces
at a level of 10 mg/kg have been published in the scientific literature.12,13,15,16,18
A PCR for the detection of celery as an allergen was presented11, and another
one recently reported.32 Furthermore, a PCR specific for soy was developed
with special regard to meat adulteration, but has potential to be applied in the
detection of allergenic soybean in composed foods.10 In addition, for the
detection of cereals that are toxic for subjects with coeliac disease, a PCR for
wheat,9 and a competitive PCR for the quantification of the gluten-containing
cereals wheat, barley, and rye in gluten-free food14 were published, and
recently a real-time PCR with melting curve analysis for the detection and
discrimination of wheat, barley, and rye was developed and applied to a
selection of final food products.17
The first qualitative PCR method for the detection of hazelnut as an
allergen in processed foods was published in 2000.12 The primers were specific
for the gene of Cor a 1.0401, the major hazelnut allergen, and the 182 bp
PCR product was detected in agarose-gel electrophoresis. The LOD was
determined as 0.001% (10 mg/kg) in the investigated food matrices (chocolate,
cookies, cereals/muesli, potato snacks, puffed corn) and detection of hazelnut
traces correlated well with a previously published and in-house validated

hazelnut-specific ELISA. Cross-reactivities with 33 relevant foods were not
observed. The specific hazelnut PCR was further improved and developed as
a PCR–ELISA application.13 Herein, a significantly shorter DNA extraction
method was used and led to a shortened total assay duration of approximately
six hours in contrast to the previously described method. For the PCR–
ELISA, the sensitivity was also determined at a level of 0.001% or 10 mg/
kg of hazelnut in various foods. In this study the PCR–ELISA was used to
analyse 41 food products for the presence of undeclared hazelnut traces and
was compared with an in-house validated hazelnut-specific sandwich ELISA
having a limit of detection of < 1 mg/kg of hazelnut protein (corresponding
to < 10 mg/kg hazelnut). Both tests showed a very good correlation for foods
having ≥ 1 mg/kg of hazelnut protein as was quantified by ELISA. Sporadic
discrepancies in the detection of hazelnut were only found for foods having
< 1 mg/kg of hazelnut protein (Fig. 7.5).
Recently, another hazelnut-specific PCR was published that is based on a
primer set specific for an intron of the mitochondrial encoded gene nad1.15
The specificity of this method was tested with a variety of closely related
plants. Cross-reactivities were found with Carpinus turczaninovii (rock
hornbeam) and Ostrya carpinifolia (hop hornbeam), belonging to the Betulaceae
family as does hazelnut (Corylus sp.). The sensitivity of this method was
reported as 10 mg/kg of hazelnut in foods.
The first peanut-specific PCR assay was published in 2003.16 With this
real-time PCR using the TaqMan™ format, the authors were able to detect
peanut down to 2 mg/kg spiked into a biscuit. However, the PCR results
14
Total
12
ELISA
10 PCR-ELISA
Sample number
0
00 10
0 10 >
1
<
1
tiv
e
10 > > a
>
N eg
mg/kg hazelnut protein
Fig. 7.5 Investigation of commercial food samples for hazelnut residue with hazelnut
specific protein sandwich-ELISA and DNA PCR-ELISA.13 Of the 41 investigated
chocolate, creme desert, breakfast flakes, cereal bars, and cookie samples, only three
discrepancies were observed for samples having less than 1 mg/kg hazelnut protein
(two samples were negative in ELISA but positive in PCR; one sample was positive in
ELISA but negative in PCR).
were not compared with those of a validated immunological assay, nor was
an unspiked biscuit analyzed as negative control. As reported by the authors,
the peanut PCR may be used for analysis if reference standards for peanut
become available.
The performance of a new peanut-specific real-time PCR was compared
with an in-house validated peanut ELISA.18 Thirty-three different food samples
were analyzed for the presence of peanut traces with both methods allowing
a detection of peanut at a level of ≤ 10 mg/kg. The results of ELISA and PCR
were in good correlation. There was only one discrepancy found in a chocolate
that was negative in ELISA (< 0.5 mg/kg of peanut protein), but positive in
PCR. The major characteristics of the published PCR methods are summarized
in Table 7.1.
7.3.2 PCR as possible alternative for immunological methods

So far, the results of several different PCR assays specific for wheat, soya,
hazelnut, gluten-containing cereals (wheat, barley, and rye), and peanut have
been compared with the results obtained from in-house validated or commercial
ELISAs.9,10,12–14,17,18 The data presented indicated that both DNA-based and
protein-based assays are applicable for the detection of hidden wheat, soy,
hazelnut, gluten-containing cereals, and peanut traces in the investigated
matrices, and that there is a good correlation of the results from both methods.
Furthermore, it was shown that some matrices may be more suitable for
PCR methods than for ELISA analysis and vice versa. In particular, two
creme deserts containing hazelnut, as stated in the list of ingredients, were
tested negative for hazelnut by protein ELISA but not by PCR–ELISA.13
This phenomenon was supposed to be related to an acidic or microbial
degradation of the proteins that had no or little influence on the DNA stability.
By contrast, a white chocolate containing only 0.5 mg/kg of hazelnut protein
tested negative by the PCR–ELISA but positive by the protein-ELISA.13 The
analysis of bread, pastry, and baby food for gluten-containing cereals
demonstrated that some food matrices may inhibit the PCR and therefore
lead to false-negative results. By contrast, the applied ELISA seemed to be
less sensitive in other food samples.14 Studying the detection of wheat, barley,
and rye in gluten-free oat products, the results of PCR and a gliadin-specific
ELISA were in complete concordance above a level of 50 mg/kg gluten.17
However, the ELISA was slightly more sensitive in some foods. In another
study, a commercial soya-specific ELISA was compared with a soybean-
specific PCR. The PCR was able to detect a soya product having approximately
70 mg/kg of soya protein, whereas the ELISA was negative with a limit of
detection of 3500 mg/kg of soya protein,10 which may be not sensitive enough
for the detection of soybean as an allergen. Similar findings were obtained
by comparing a gluten-specific ELISA and a more sensitive wheat-specific
PCR for the detection of wheat adulterations in foods with special regard to
coeliac disease.9 Moreover, it should be mentioned that the PCR assays for
Table 7.1 Major characteristics of published PCR methods suitable for food allergen detection
Food allergens See Target gene Amplicon Sensitivity

reference size [bp]
pg DNA Gene mg/kg allergenic
copies food
Hazelnut 12 Cor a 1.0401 182 n.p. n.p. ≤ 10

Hazelnut 13 Cor a 1.0401 152 2–4 1–2 ≤4
Hazelnut 15 Nad1 294 30–300 n.p. 10
(mitochondrial)
Peanut 16 Ara h 2 663 n.p. n.p. 2
Peanut 18 Ara h 2 86 n.p. 10 ≤ 10
Soy1 10 Lectin Le1 118 n.p. n.p. n.p.
Wheat1 9 RNA gene 109 1 n.p. < 3004
interspacer
between 25S and
18S (ribosomal)
Gluten-cereals wheat, 14 trnL (chloroplast) 201 (wheat, rye) 2–20 n.p. < 2004
barley, rye1 196 (barley)
Gluten-cereals wheat, 17 Wheat ω-gliadin 181 (wheat, rye) < 50 ≤5 100–10005
barley, rye1 Rye ω-secalin 164 (barley)
Barley hordein 104 (oats)
Oats avenin
Celery 32 Api g 1 145 30–250 n.p. n.p.
n.p. not published
1
the soybean and gluten-cereals specific PCR methods were not developed for food allergen detection but for soy adulteration and gluten analysis, respectively
2
validated versus an ELISA method
3
derived from published sequences
4
derived from positive sample with lowest detected amount of wheat
5
derived from a detection level of approximately 50 mg/kg gluten
soy and wheat were not optimized for sensitivity as the aim of the studies
was either the identification of meat adulterations at the percent level or the
detection of gluten-containing wheat at a level of 200–400 mg/kg (10–20
mg/kg gliadin). Hence, even better sensitivities may have been achieved with
optimized PCR protocols. Most of the published and commercial
immunological methods for soya have been developed for the detection of
meat adulteration with soybean products having high limits of detection of
approximately 1000 mg/kg of soy protein (summarized by Poms et al.19). In
contrast to the numerous publications about the detection of peanut, which is
a legume like soybean, only two scientific references about the detection of
soybean with ELISA at a suitable level of 1–2 mg/kg have been reported so
far.33,34
One reason why soya has been difficult to target at the protein level using
ELISA methods is the great number of different soya products, such as
defatted soybean flakes, texturized soybean protein, protein isolates and
concentrates, and protein hydrolysates. The differences in processing can
result in a variety of altered peptides and protein breakdown products with
differing structural properties and therefore differing detectabilities. Therefore,
soya detection by PCR methods may be a possible alternative to immunological
methods if the detection of soybean DNA correlates with the presence of
soybean protein and peptides, respectively. So far, no celery-specific and
sensitive ELISA has been published. One reason for this may be the difficulty
in generating a celery-specific antiserum that allows sensitive detection of
celery in foodstuffs without cross-reacting to other species of the Apiaceae
family. By contrast, the PCR has great potential for the differentiation between
phylogenetically closely related species. In this context, celery, fish, crustaceans,
and various tree nuts may be interesting candidates to detect with PCR
methods in addition to the application of PCR for verification of ELISA
results or even as a substitute for ELISA.
For hazelnut, peanut, wheat, soya, and celery, PCR has already been
shown to be an alternative to ELISA assays. However, more investigations
have to be done to support these findings. Additionally, the applicability of
both methodologies has to be thoroughly investigated for each allergenic
food and in the most important food matrices. Thereafter, a final conclusion
can be drawn about the potential use of both techniques depending on the
allergenic food and the food matrix.
7.3.3 Commercially available PCR assays

Besides the published DNA-based methods for the detection of hidden allergens
in processed foods, several applications of PCR–ELISA and real-time PCR
are commercially available (R-Biopharm, Darmstadt, Germany; Eurofins
Scientific, Memphis, TN; Tepnel Biosystems Ltd, Deeside, UK; Genetic ID
NA, Fairfield, IA). Currently, assays for the detection of peanut, almond,
hazelnut, soy, wheat, chicken (for hen’s egg), beef (for cow’s milk), fish,
celery, sesame, and mustard are available (summarized by Poms et al.19).

The detection limit of the tests was specified by the manufacturers to be
between 10 and 20 DNA copies. However, without having standard reference
samples, a determination of copy numbers may not be correlated with the
actual amount of the allergenic food that is present in the investigated sample.
Detailed studies that show the performance characteristics of most of these
assays have not been published, and publications that investigate the
applicability of these methods are mainly lacking.
7.4 Advantages and disadvantages of PCR compared to

ELISA
The majority of commercial and published allergen-specific assays are protein-
based, whereas only a few DNA-based assays have been developed so far.
Reasons for this are quick performance and easy handling, and the rather
simple equipment which is required for ELISA. Furthermore, experience in
the development of ELISA techniques has been gained over more than 30
years, whereas the analysis of foods by PCR began only in the mid 1990s.
However, the PCR methodology presents several interesting features for the
application in routine analysis of food allergens. In contrast to ELISA tests
that are based on polyclonal antibodies, the whole PCR chemistry and
recombinant thermostable DNA polymerase can be synthesized and
manufactured in unlimited amounts and constant quality. Furthermore, the
specific reagents, such as primers and probes, that are needed for PCR can
be designed to meet the needs for specificity and sensitivity. Similarly, the
use of monoclonal antibodies in ELISA can overcome these limitations of
availability and consistency, but ELISA based on monoclonal antibodies
may be more susceptible to matrix effects because of a lower complexity in
protein-specific recognition sites. This can lead to false-negative results if
the target protein or the target epitope is affected during food processing or
because other proteins that are not detected are important allergens for a
subgroup of allergic individuals. Moreover, the antigenic epitopes that are
detected with animal antibodies do not necessarily reflect the allergenic
epitopes that are recognized by the allergic individual. Similarly, the PCR
cannot detect the allergenic molecules because the amplified and detected
target molecule is the DNA. Therefore, neither currently available ELISA
tests nor the PCR applications can be considered to be direct detection methods.
Both methodologies are mainly based on the detection of specific markers
that allow extrapolation for the presence of the allergenic food of interest.
Nevertheless, the identification of protein from an allergenic food gives
some more evidence for allergenic potential as does the DNA analysis by
PCR. Hence, it is a prerequisite in allergen detection with PCR to verify a
positive correlation of the marker DNA in comparison to the presence of
potentially allergenic food proteins. As DNA can be found in any tissue of
plant or animal origin, the species-specific detection of DNA gives evidence

for the presence of the allergenic food if food technological processing does
not separate the marker DNA from the allergenic protein fraction. Consequently,
allergenic residues in food ingredients such as protein isolates, sugars, starch,
aromas, and highly-processed vegetable oils may be difficult to detect and
results obtained by PCR difficult to interpret. However, residues in sugars,
starch, aromas, and highly-processed vegetable oils may also be difficult to
detect with ELISA tests. Moreover, the sensitivity of the PCR strongly depends
on the amount and on the quality of the template DNA. In some cases the
DNA isolated from food products can be highly degraded, making amplification
of a specific PCR product impossible. Especially in products having a low
pH, such as soured or fermented foods, the relevant DNA can be hydrolyzed,
which makes a PCR analysis very difficult. This matrix effect was obvious
when celery DNA from celery juice could not be amplified and detected with
a 241 basepair celery-specific PCR.11 However, such matrix effects can also
occur with protein-based methods if the target proteins are susceptible to
degradation, as was described for hazelnut.13
Other problematic matrices for PCR analysis are foods or food products
having low amounts of DNA, such as milk, egg white, vegetable oils, or
animal fats. In these cases, protein ELISA may be more suitable than PCR
assays. Moreover, the PCR is not able to distinguish between egg and chicken
meat, or cow’s milk and beef, because DNA is not tissue-specific. Nevertheless,
PCR offers unparalleled specificity in species identification making it a
favourable method for the detection of allergenic food components of animal
or plant origin with high homology to other species used in food manufacture.
By the use of a single protocol, DNA from various species is extracted
simultaneously from a complex food matrix. Therefore, a multi-analyte
detection that is based on the same sample preparation is possible. For this
reason, PCR has great potential to become a screening method because
currently available ELISAs are still single-analyte tests and require optimized
extraction conditions for each allergenic food that is to be detected. However,
it has to be considered that the performance of PCR, at least with current
technology, does still need specialized laboratories. Besides the issues with
possible cross-contamination, PCR does not seem to be practical for direct
analysis in an average food production line, whereas dipstick ELISA or other
rapid ELISAs are favoured because of easier and faster handling. Obviously,
it is very important to consider the pros and cons of both technologies before
starting to analyse foods for the presence of hidden allergens. Moreover, any
method used for the detection of allergens in foods needs to be validated for
its sensitivity, specificity, and detectability in the food matrix of interest to
guarantee its applicability. Ideally, the correlation between the presence of
allergenic proteins and the successful detection of the allergenic food by
protein ELISA or PCR should be demonstrated.
So far, validation on allergen-specific tests has been carried out mainly
with peanut-specific ELISAs. Taking into account the results of these studies,
it was shown that commercially available peanut-specific ELISAs have the

potential for correct qualitative and semi-quantitative but not yet for quantitative
analysis.35–37 Similarly, the detection of peanut with real-time PCR methods
indicated that the relative Ct values obtained by the real-time PCR reflected
the relative amount of peanut present in a sample.16,18 Moreover, the comparison
of a real-time PCR and an ELISA specific for peanut revealed that the
relative Ct values obtained by the real-time PCR are correlated to the amount
of protein measured by the sandwich ELISA.18 This underlines the potential
of PCR to be used as a semi-quantitative method for the detection of peanut
traces in processed foods if suitable reference materials with defined amounts
of peanut are available. Currently, several approaches to generating such
standard material and to improve the standardization of allergen specific
assays are in progress (European Institute of Reference Materials and
Measurements, IRMM; German Institute for Standardization, DIN; European
Committee for Standardization, CEN; US Food and Drug Administration,
FDA). In spite of the availability of standard reference materials, it will
nevertheless not be possible to quantify the exact amount of an allergenic
food in processed foods because of unknown processing grades (e.g. different
roasting times and temperatures for nuts) of the foods and the vast number
of different food matrices that lead to a varying detectability of the biological
target molecules protein and DNA. However, the generation of standard
reference materials is necessary to make the results of allergen-specific assays
comparable and to achieve reproducible inter-laboratory results.
7.5 Future trends

A few scientific studies suggest that DNA-based methods can be applied to
the analysis of hidden allergens in finished food products and can complement
the application of ELISA tests. Some allergenic foods and food matrices
may be more suitable for PCR and others more suitable for ELISA analysis.
In both cases the developed methods have to be thoroughly validated with
particular regard to their applicability in the target food matrices. For example,
tests developed for peanut or tree nut need to perform well in food matrices
such as cookies or chocolate, whereas a crustacea-specific assay needs to be
validated for products potentially containing fish or crustacea residues. ELISA
tests may be advantageous for the detection of eggs and milk as allergenic
foods whereas the differentiation between phylogenetically closely related
tree nuts, fish, and crustaceans may be easier to achieve with PCR-based
methods. In the future, the development of methods to isolate amplifiable
DNA at high yields and of high purity will further help to improve the
applicability of PCR in the detection of hidden allergenic foods. Moreover,
with the use of a single DNA extraction, several allergenic food compounds
can be screened at once. This makes PCR technology an ideal multi-analyte
screening tool.
In conclusion, the development of sensitive and specific analytical methods

for the quantification of allergens in finished food products is still a great
challenge due to the lack of suitable standard reference materials. The
availability of either DNA or protein standards will help to make quantification
with protein- and DNA-based assays comparable. With the availability of
standard reference materials, it is to be expected that the development of
PCR assays for allergen detection will further proceed, and the development
of quantitative PCR assays will be made possible.
7.6 References
1. Hefle, S L, Bush, R K, Yunginger, J W and Chu, F S (1994) ‘A sandwich enzyme-
linked immunosorbent assay (ELISA) for the quantitation of selected peanut proteins
in foods’, J Food Prot, 57, 419–423.
2. Yeung, J M and Collins, P G (1996) ‘Enzyme immunoassay for determination of
3. Holzhauser, T and Vieths, S (1999) ‘Indirect competitive ELISA for determination
of traces of peanut (Arachis hypogaea L.) protein in complex food matrices’, J Agric
Food Chem, 47, 603–611.
4. Holzhauser, T and Vieths, S (1999) ‘Quantitative sandwich ELISA for determination
of traces of hazelnut (Corylus avellana) protein in complex food matrixes’, J Agric
Food Chem, 47, 4209–4218.
5. Koppelman, S J, Knulst, A C, Koers, W J, Penninks, A H, Peppelman, H, Vlooswijk,
R, Pigmans, I, van Duijn, G and Hessing, M (1999) ‘Comparison of different
immunochemical methods for the detection and quantification of hazelnut proteins
in food products’, J Immunol Methods, 229, 107–120.
6. Hlywka, J J, Hefle, S L and Taylor, S L (2000) ‘A sandwich enzyme-linked
immunosorbent assay for the detection of almonds in foods’, J Food Prot, 63, 252–
257.
7. Hefle, S L, Jeanniton, E and Taylor, S L (2001) ‘Development of a sandwich enzyme-
linked immunosorbent assay for the detection of egg residues in processed foods’, J
Food Prot, 64, 1812–1816.
8. Pomes, A, Helm, R M, Bannon, G A, Burks, A W, Tsay, A and Chapman (2003)
Immunol, 111, 640–645.
9. Allmann, M, Candrian, U, Höfelein, C and Luthy J (1993) ‘Polymerase chain reaction
(PCR): a possible alternative to immunochemical methods assuring safety and quality
of food’, Z Lebensm Unters Forsch, 196, 248–251.
10. Meyer, R, Chardonnens, F, Hübner, P and Lüthy, J (1996) ‘Polymerase chain reaction
(PCR) in the quality and safety assurance of food: detection of soya in processed
meat products’, Z Lebensm Unters Forsch, 203, 339–344.
11. Jankiewicz, A, Hübner, P, Bögl, K W, Dehne, L I, Vieths, S, Baltes, W and Lüthy, J
(1997) ‘Celery Allergy: PCR as a tool for the detection of trace amounts of celery in
processed foods’, Proceedings of Euro Food Chem IX, Interlaken, Switzerland, Sept
24–26.
12. Holzhauser, T, Wangorsch, A and Vieths, S (2000) ‘Polymerase chain reaction (PCR)
for detection of potentially allergenic hazelnut residues in complex food matrixes’,
Eur Food Res Technol, 211, 360–365.
13. Holzhauser, T, Stephan, O and Vieths, S (2002) ‘Detection of potentially allergenic
hazelnut (Corylus avellana) residues in food: a comparative study with DNA PCR-
ELISA and protein sandwich-ELISA’, J Agric Food Chem, 50, 5808–5815.
14. Dahinden, I, von Büren, M and Lüthy, J (2001) ‘A quantitative competitive PCR
system to detect contamination of wheat, barley or rye in gluten-free food for coeliac
patients’, Eur Food Res Technol, 212, 228–233.
15. Herman, L, Block, J D and Viane, R (2003) ‘Detection of hazelnut DNA traces in
chocolate by PCR’, Int J Food Sci Tech, 38, 633–640.
16. Hird, H, Lloyd, J, Goodier, R, Brown, J and Reece, P (2003) ‘Detection of peanut
using real-time polymerase chain reaction’, Eur Food Res Technol, 217, 265–268.
17. Sandberg, M, Lundberg, L, Ferm, M and Malmeheden Yman, I (2003) ‘Real time
PCR for the detection and discrimination of cereal contamination in gluten free
foods’, Eur Food Res Technol, 217, 344–349.
18. Stephan, O and Vieths, S (2004) ‘Development of a real-time PCR and a sandwich
ELISA for detection of potentially allergenic trace amounts of peanut (Arachis
hypogaea) in processed foods’, J Agric Food Chem, 52, 3754–3760.
19. Poms, R E, Klein, C L and Anklam, E (2004) ‘Methods for allergen analysis in food:
a review’, Food Addit Contam, 21, 1–31.
20. Taylor, S and Nordlee, J (1996) ‘Detection of food allergens’, Food Technol, 50,
231–234.
21. Hourihane, J O B, Kilburn, S A, Nordlee, J A, Hefle, S L, Taylor, S L and Warner,
J O (1997) ‘An evaluation of the sensitivity of subjects with peanut allergy to very
low doses of peanut protein: a randomized, double-blind, placebo-controlled food
challenge study’, J Allergy Clin Immunol, 100, 596–600.
915–920.
23. CEN, ‘Foodstuffs – Methods of analysis for the detection of genetically modified
organisms and derived products – Nucleic acid extraction’ (ISO/DIS 21571) prEN
ISO 21571, Brussels, CEN.
24. SLMB, Schweizerisches Lebensmittelbuch, Kapitel 52B, Molekularbiologische
Methoden, Bern, Bundesamt für Gesundheit Facheinheit Lebensmittelsicherheit (2003).
25. Newton, C R and Graham, A (1997) PCR (Introduction to biotechniques), Oxford,
Bios Scientific Pub Ltd.
26. Sambrook, J and Russels, D W (2001) Molecular Cloning: A Laboratory Manual,
Cold Spring Harbor Press, NY.
27. DIN, ‘Part 60: Polymerase chain reaction (PCR): Terminology, general method-
specific requirement’, Diagnostics of infectious diseases and diseases of the immune
system in serology and molecular biology, DIN 589667-60, Berlin, Beuth, 2001.
28. CEN, ‘Foodstuffs – Methods of analysis for the detection of genetically modified
organisms and derived products – Qualitative nucleic acid based methods’, (ISO/
DIS 21569) prEN ISO 21569 Brussels, CEN.
29. Meuer, S, Wittwer, C and Nakagawara, K (eds.), (2003) Rapid Cycle Real-time PCR
– Methods and Applications, Berlin, Springer.
30. CEN, ‘Foodstuffs – Nucleic acid based methods of analysis for the detection of
allergens – General requirements and definitions’, (ISO/DIS 24276) prEN ISO 24276,
Brussels, CEN.
31. Longo, M C, Berninger, M S and Hartley, J L (1990) ‘Use of uracil DNA glycosylase
to control carry-over contamination in polymerase chain reactions’, Gene, 93, 125–
128.
32. Stephan, O, Weisz, N, Weiser, T, Rabe, B, Vatterott, W and Vieths, S (2004) ‘Protein
quantification, sandwich ELISA, and real-time PCR used to monitor industrial cleaning
procedures for contamination with peanut and celery allergens’, J AOAC Int, 87,
1448–1457.
33. Yeung, J M and Collins, P G (1997) ‘Determination of soy proteins in food products
by enzyme immunoassay’, Food Technol Biotechnol, 35, 209–214.
34. Koppelman, S J, Lakemond, C M M, Vlooswijk, R and Hefle, S L (2004) ‘Detection

of soy proteins in processed foods: Literature overview and new experimental work’,
J AOAC Int, 87, 1398–1407.
35. Hurst, W J, Krout, E R and Burks and W R (2002) ‘A comparison of commercially
available peanut ELISA test kits on the analysis of samples of dark and milk chocolate’,
J Immunoassay Immunochem, 23, 451–459.
36. Koch, P, Schäppi, G F, Poms, R E, Wüthrich, B, Anklam, E and Battaglia, R (2003)
‘Comparison of commercially available ELISA kits with human sera-based detection
methods for peanut allergens in foods’, Food Addit Contam, 20, 797–803.
37. Poms, R E, Lisi, C, Summa, C, Stroka, J and Anklam, E (2003) ‘In-house validation
of commercially available ELISA test kits for peanut allergens in food’, Report from
the Institute for Reference Materials and Measurements, Joint Research Centre,
European Commission, EUR 20767 EN.
8
Proteomic assessment of allergens in

food
M. Zeece, J. Markwell, G. Sarath and X. Gu,
University of Nebraska, USA
8.1 Introduction
Proteomics has recently received much attention. The word ‘proteomics’
encompasses a family of technologies with which to determine the complement
of proteins in a cell, tissue or organism. Currently, these technologies include
multi-dimensional separations via electrophoresis or liquid chromatography;
various methods for protein identification including mass spectrometry; arrays
to map protein–protein interactions; and bioinformatics to analyze massive
data sets.
Proteomics is evolving from its roots in two-dimensional electrophoresis.
However, at present, the electrophoretic approach still represents a valuable
technique for separating proteins. Proteomics is especially useful for the
identification of IgE-binding proteins for several reasons. First, conditions
for solubilization of proteins are very aggressive, involving combinations of
urea, thiourea, non-ionic detergents and strong reducing agents. These reagents
increase the likelihood of detecting allergens. Second, spots in two-dimensional
separations are more likely to be homogenous compared to protein bands in
one-dimensional sodium dodecyl sulphate polyacrylamide gel electrophoresis
(SDS-PAGE) separations. In addition, two-dimensional separation provides
substantial information in the form of isoelectric point (pI), molecular weight
and heterogeneity (charge and/or size variants) of the IgE-binding component.
This information makes it possible to substantially narrow the search for the
protein’s identity. Finally, the value of a proteomics approach can be seen in
the number of newly discovered allergens (discussed below). While the number
of reports using proteomics to investigate IgE-binding proteins is low, most
have uncovered ‘novel’ allergens.
Proteomic assessment of allergens in food 145
8.2 Key issues in proteomic assessment of allergens

8.2.1 Separations
Two-dimensional electrophoresis (2-DE) consisting of isoelectric focusing
(IEF) followed by SDS-PAGE separation represents the core of proteomics
technology since its establishment a quarter century ago (O’Farrell, 1975).
Recent advances have greatly improved the efficiency of resolution, speed
and reproducibility of the technology. Most notable has been development of
immobilized ampholine technology. Strips containing immobilized ampholines
(IPG strips) result in greater resolution by enabling the application of higher
electric field strengths while also preventing cathodic drift of the pH gradient.
IEF power supplies with vastly improved performance are now available.
These units are capable of routinely running at field strengths of 8–10 kV.
They also employ efficient cooling systems to control Joule heating and
apply power in a programmed sequence. The new generation of IEF power
supplies can run 8–12 samples at once. Typically, IEF separations are conducted
for a specified number of volt-hours (Vh), e.g. 50 kVh. Thus relatively high
sample throughput can be achieved when compared with older equipment.
IEF separation
Complex mixtures of proteins can be separated by IEF. The unique balance
of positive and negatively charged amino acids determines a protein’s isoelectric
point (pI). A protein’s pI corresponds to that pH at which the positive and
negative charges sum to zero. When the pH is equal to a protein’s pI, it is
electrically neutral and its migration in an electric field is halted. Typically,
IEF is conducted in the presence of high concentrations of urea (6–8 M),
non-ionic detergents (e.g. (3-[3-cholamidopropylammonio]-1-propanesulfonate
– CHAPS) and reducing agent (e.g. dithiothreitol – DTT) to promote solubility.
A pH gradient is established typically from 3.0–10.0 using immobilized
ampholines in a polymer-based IPG strip.
The second dimension separation is usually performed via SDS-PAGE
following appropriate equilibration. The combination of separations based
on different mechanisms of selectivity (pI vs relative mass) results in enhanced
resolution of component proteins. Thus 2-DE has distinct advantages for
characterizing the IgE-binding components in complex mixtures of proteins.
8.2.2 Quantitation
Significant improvements in staining methods and computer-driven image
analysis routines have greatly improved the ability to quantitatively assess
differences between samples. However, while the image analysis packages
may provide numerical evaluations, careful use is needed to ensure accuracy
(Miller et al., 2001; Miura, 2001).
Quantitation of stained proteins in gels requires several steps to establish
the validity of the measurement. Determination of linear range of detection
is an often overlooked evaluation criterion. The linear dynamic range can be

established by comparing (plotting) response measurements for a series of
protein standards. Differential dye absorption should also be expected. Proteins
can vary significantly in the amount of dye absorbed per unit of protein and
this results in values that over-or under-estimate the amount present.
Additionally, accurate quantitation requires control over separation parameters
such as: voltage, separation time, temperature and gel staining/destaining
procedures. Finally, appropriate analytical protocols such as triplicate sampling
and determining the variability (standard error) of measurements, should be
employed.
Image analysis programs have advanced features that facilitate recording
and measuring the spot position and intensities. No two gels (especially 2-D)
will have identical x-y coordinates for all component spots. Most image
analysis programs can adjust patterns such that highly accurate overlays and
subtractive analyses can be performed. In addition, these programs usually
contain statistical evaluation routines to aid in the accuracy of the assessment.
Image analysis programs generally determine the amount of protein in a
spot based on measurement of its size and intensity. The intensity of a spot
is determined from a Gaussian representation of its optical density (OD).
The size of the spot is determined from appropriate area integration algorithms.
Spot values are then expressed as OD times area units. This value is called
spot quantity in some image analysis programs. In applications where the
goal is to compare the amount of an individual component between two or
more samples, spot quantities are usually expressed as a percentage of the
total. In this way, small differences in load amounts are normalized. Thus,
image analysis programs can be useful for determining overall differences in
the protein profile of samples. For example, image analysis can be used to
assess whether there has been alteration in potential allergen content between
conventional and genetically modified (GM) varieties.
8.2.3 Methods of detection and identification

Proteins in 2-DE separations are commonly visualized by staining with
Coomassie Blue (R250). This tried and true standard has one major
disadvantage, namely its lack of sensitivity. The limit of detection is
approximately 0.1 μg for most proteins. Enhanced sensitivity can be achieved
by use of silver staining. There are several slightly different silver stain
chemistries that result in varied colors (brown to black) but most have a 10-
fold lower limit of detection. Recently, fluorescent dyes such as SYPRO®
(Molecular Probes Inc, Eugene, OR) with even lower (1–10 ng) detection
limits have become available. Staining with SYPRO® is much faster than
silver staining and has a greater dynamic range of measurement in image
analysis (Berggren et al., 2000; Lopez et al., 2000; Nishihara and Champion,
2002). Fluorescent dye is also more compatible with subsequent mass
spectrometry (MS) analysis procedures than is silver stain.
Proteins in 2-DE separations can be identified following transfer to

polyvinylidene difluoride (PVDF) membrane via immunoblotting or Edman-
based N-terminal sequence analysis. However, the amount of material required
for Edman analysis and/or the presence of N-terminal modifications, often
limits its use for identification of proteins separated by 2-DE. Immunoblotting
can also be used to assess the level of specific IgE-binding proteins in a
sample, but this of course requires prior knowledge of the protein’s identity
and an appropriate detection antibody.
Mass spectrometry has become the method of choice for the detection and
identification of proteins in 2-DE gels because of its greater sensitivity and
ability to determine sequence information. Mass spectrometry can also be
used to characterize protein modifications such as phosphorylation and/or
glycosylation. This section will only briefly describe some of the mass
spectrometry techniques that are useful in the application of allergen
identification. The reader is referred to the recent review by Mann et al.
(2001) for a wider survey of various MS analytical approaches and their
mechanism of operation.
Most MS analyses are performed on proteins after they have been digested
with trypsin into constituent peptide fragments. In this analysis, a gel plug
containing the protein is excised, destained and digested with an excess
amount of trypsin. The resultant peptides are extracted and can then be
analyzed by several MS techniques.
The most commonly used MS analytical approach is matrix-assisted laser
desorption ionization (MALDI). In MALDI analysis, the mixture of peptides
is combined with an excess of matrix compound such as α-cyano 4-
hydroxycinnamic acid and 1–2 μL is spotted onto a plate. The spots are
irradiated by nanosecond laser pulses, during which excited matrix molecules
donate protons to the peptides. The charged molecules are differentially
accelerated (based on mass/charge ratio) in a strong electric field. This results
in lighter ions arriving at the detector before heavier ones. This technique for
differentiation is referred to as time-of-flight (TOF) analysis. Accurate masses
obtained from each peptide are collected into a mass profile that is compared
to theoretical mass profiles based on predicted trypsin cleavage sites of
known protein sequences. Thus the identity of an unknown protein can be
established from its mass profile. Difficulties in the MALDI–TOF method of
identification (called mass fingerprinting) are encountered when digestion
of the protein is incomplete or when peptide recovery is low. Hydrophobic
proteins often present this type of problem. Some Arg and Lys residues may
be inaccessible to the enzyme and/or peptides containing a high proportion
of non-polar residues which are poorly soluble (Mann et al., 2001).
A second major type of MS analysis employs the use of tandem mass
spectrometers to fragment an individual peptide and obtain its sequence
(MS/MS analysis). In this technique, a peptide is selected from the mixture
in the first mass spectrometer and is then dissociated by collision with an
inert gas. Collision-activated dissociation causes a single cleavage to occur
randomly and predominantly at various amide bonds in the peptide. The

resulting fragments (a series differing by the mass of one amino acid) are
analyzed in the second mass spectrometer. Tandem mass spectrometry also
employs other techniques to refine the analysis. For example, the range of
the masses to be analyzed can be selected by using a magnetic sector
(quadrupole) coupled with a time-of-flight analyzer. The MS/MS approach
has several advantages in that multiple sequences can be obtained for the
protein thus enabling more precise identification. Additionally, MS/MS analysis
can also characterize protein modifications such as phosphorylation and
glycosylation.
Mass spectrometry analysis can also be employed in combination with
liquid chromatography (LC) separation. The LC-MS/MS approach provides
information in the form of peptide sequences and represents an excellent tool
for the detection and identification of allergens in foods. Shefcheck and
Musser (2004) recently demonstrated the usefulness of this approach for
detecting the peanut allergen Ara h 1 in ice cream. In this report, extracts of
ice cream matrix were enzymatically digested followed by LC separation
tandem MS analysis. The authors found the method to be capable of detecting
as little as 10 ppm of Ara h 1 in spiked samples. Thus MS shows promise for
being an important tool for the detection and characterization of allergens.
8.2.4 Limitations of the technology

Two-dimensional electrophoresis still represents the major separation
technology used to analyze complex samples. There are numerous examples
in which this technology has enabled characterization of the complement of
proteins in plant and animal tissues. Databases can now be accessed to
provide presumptive identification of proteins based on their positions (pI &
Mr) in the separation (http://www.lecb.ncifcrf.gov/EP/table2Ddatabases.html).
However, 2-DE also has a number of limitations that must be considered.
First, unlike SDS-PAGE, the technique requires more operator skill and
time. Second, almost every sample system presents unique challenges for
development of appropriate preparation and separation protocols. The
optimization of resolution, i.e. reduction of horizontal and/or vertical streaking,
must be achieved by empirical adjustment of variables such as surfactant
type and concentration and reductive alkylation reagents. Also, run conditions
such as pH range employed and volt-hours of separation in IEF must be
optimized.
Third, the nature of the proteins to be analyzed impacts 2-DE separations.
Large and/or very hydrophobic proteins do not resolve well in IEF. Typically,
proteins with Mr greater than 100 kD are not observed in 2-DE separations.
While the advent of IPG strip technology has greatly enhanced the efficiency
and reproducibility of IEF separations, hydrophobic proteins remain
problematic. Thus, membrane-associated proteins represent a very difficult
group of proteins to resolve via 2-DE.
Fourth, low abundance proteins represent a challenge to 2-DE-based

analyses. This is of concern to investigation of IgE-binding proteins because
major allergens may be minor components of the total protein complement.
Enhanced detection of proteins in 2-DE separations can be achieved by use
of silver or fluorescent (e.g. SYPRO®) staining protocols. Typically, a 10–
100 fold increase in sensitivity can be achieved by these methods. However,
lower protein level in the spot also represents a challenge to protein
identification, even by MS. Most MS methods require the protein to be
digested to peptides. However, the extent of digestion may be low and the
subsequent recovery of peptides (especially those with a significant percentage
of non-polar residues) may be very low. Thus the ability to identify a protein
diminishes significantly as its level in the gel decreases.
A common strategy to overcome the issue of low abundance is to fractionate
the sample prior to separation. This approach is referred to as a targeted
proteome and may employ a number of methods such as differential
sedimentation (e.g. 2S, 7S & 11S fractions) or ‘pull downs’ using affinity
probes such as antibodies. The compromise in a targeted proteome approach
is that other proteins of interest may be excluded from the analysis.
8.3 Applications of proteomics for detection of allergens

While the number of reports using a proteomics approach for the investigation
of allergens is small, substantial growth is expected as the advantages of the
technology become known. The examples discussed below describe some of
the relevant applications of note.
8.3.1 Wheat allergens

Investigation of wheat proteins associated with baker’s asthma represents
one of the first applications of proteomics. Two-dimensional gel electrophoresis
and immunoblotting demonstrated numerous IgE-binding proteins in wheat
(Weiss et al., 1993; Posch et al., 1995; Weiss et al., 1997).
Two-dimensional electrophoretic separation and detection of IgE-binding
proteins was performed on fractionated wheat flour. Specifically, the albumin/
globulin fraction was the focus of investigation because it was judged to
contain a majority of the IgE-binding proteins (Weiss et al., 1993). The study
used sera from a pool of four patients (bakers) experiencing asthmatic symptoms
for immunoblotting of 2-DE separations. Approximately 30 polypeptide spots
were observed to bind IgE from these allergic patients. Major IgE binding
was observed for a cluster of spots in the 27 kD region. This group contained
four to five charge variants with approximately equal IgE-binding intensity.
Additional IgE-binding polypeptides were observed in the separation including
groups in the 15–20 kD and 30–35 kD regions.
Subsequent proteomic investigations of IgE-binding wheat proteins

associated with baker’s asthma implemented N-terminal sequencing for
identification (Posch et al., 1995). Nine major IgE-binding spots in the 2-DE
separation of the globulin fraction were selected for analysis. The 27 kD
family IgE-binding proteins were identified from Edman sequencing data as
isoforms of Acyl-CoA oxidase. Identification of IgE-binding spots with Mr
in the 29–35 kD range was not successful. Finally, two IgE-binding spots of
14–18 kD were identified from their N-terminal sequence as isoforms of α-
amylase inhibitor.
Most recently this group used a collection of 20 sera from bakers
experiencing asthmatic symptoms to further characterize the profile of IgE-
binding wheat proteins (Weiss et al., 1997). The salt-soluble fraction of
several wheat cultivars containing albumins and globulins was the focus of
this study. Numerous IgE-binding proteins were again observed. IgE binding
was found for albumins/globulins in the 70, 55, 35, 26–28 and 14–18 kD
regions. However, major IgE binding was found to 26–28 and 14–18 kD
proteins. IgE-binding spots corresponding to these spots were identified as
Acyl CoA oxidase, and α-amylase inhibitors, respectively.
8.3.2 Hazelnut allergens

Hazelnuts represent a significant cause of allergic reactions ranging from
oral allergy syndrome to anaphylaxis. The source of oral allergy syndrome
may be linked to cross-reactivity of the profilin protein, termed Cor a 1 in
hazelnuts (Hirschwehr et al., 1992). However, most other IgE-binding proteins,
thought to be responsible for more severe reactions, have not been characterized.
Recently, a major allergen was identified using a proteomic approach (Beyer
et al., 2002a). Two-dimensional electrophoresis and immunoblotting with
sera from 14 hazelnut-allergic individuals was used to detect IgE-binding
proteins. While IgE binding was observed to numerous spots, a 40 kD
polypeptide was recognized by 12 of the 14 allergic patient sera. The 40 kD
protein exhibits several isoforms. The protein corresponding to this spot was
excised, digested and the sequence of two fragments obtained by the Edman
method.
The sequences were found to have significant homology with the legumin
precursor of English oak and pruning 2 precursor of almond. Both proteins
are 11S seed storage proteins. The identity of this protein as an 11S globulin
was confirmed by expressing the sequence-specific clone from a hazelnut
cDNA library. The full-length clone was found to encode for a 59 kD protein.
Protein expressed by this clone was positive for IgE-binding. Like other 11S
globulins of soybean and peanut (glycinins), the 40 kD portion represents
the post-translationally processed acidic subunit that has also been shown to
be an allergen in these species (Rabjohn et al., 1999; Xiang et al., 2002).
This 11S globulin (glycinin acidic chain) represents a novel IgE-binding
protein in hazelnuts and has been designated in allergen nomenclature as Cor
a 9. This protein shares a considerable sequence similarity with other allergens,

based on it its origin in the cupin superfamily of proteins. Interestingly, when
sequences were aligned with the corresponding peanut allergen (Ara h 3)
IgE epitope, nine of 15 residues (from G237 to A251) were found to be
identical (Beyer et al., 2002a).
8.3.3 Sesame seed allergens

There has been a rise in the number of reported cases of allergic reaction to
sesame seed (Sporik and Hill 1996). The incidence of sesame allergy is
higher for young children. It is thought that the incidence of sesame allergy
is linked to increased consumption of sesame seed used in desserts and as
toppings on baked goods. Numerous proteins with Mr from 8–80 kD bind
IgE from sesame-allergic patients. Among these proteins, a 2S albumin was
reported to be immunodominant and has been designated in the allergen
nomenclature as Ses i 1 (Pastorello et al., 2001).
Recently Beyer et al., (2002b) used 2-DE in combination with MS and
Edman methods of analysis to investigate IgE-binding proteins in extracts of
sesame. This study used sera from 20 sesame-allergic patients to characterize
the IgE-binding protein pattern in both 1-D and 2-DE separations. IgE-
binding proteins ranged from 7–78 kD. Four IgE-binding regions at 7, 34,
45, and 78 kD were recognized by most (> 50%) of the allergic-patient sera
used in these immunoblots of 2-D separations The spots corresponding to
these IgE-binding polypeptides were excised, digested and sequenced.
Peptides from 78 and 34 kD spots were identified as embryonic abundant
protein (soybean) and seed maturation protein (glucose and ribitol
dehydrogenases from soybean), respectively. These findings represent novel
allergens, not previously described in the literature. Their significance will
have to await future investigation.
The 45 kD IgE-binding polypeptides were identified from sequence analysis
as a 7S vicilin-type globulin. Comparison of the full-length protein sequence
for this allergen with the much more allergenic peanut vicilin (Ara h 1)
showed only 36% homology. Interestingly, examination of the sequence in
the region corresponding to a reported IgE epitope of Ara h 1 revealed a
much closer match with the 45 kD sesame protein. Seven of 10 residues
corresponding to T294 to P303 were identical. The newly discovered 7S
sesame vicilin has subsequently been designated as Ses i 3 (Beyer et al.,
2002b).
The 7 kD polypeptide was found to be a 2S albumin. Several examples of
2S albumin allergens have been characterized from Brazil nut (Ber e 1) and
walnut (Jug r 1). They are heterodimeric proteins with approximately 8 and
14 kD subunits, linked by disulfide bonds. The sequence of the polypeptide
isolated by Beyer et al. (2002b) showed a high degree of homology (38 and
40% respectively) to Ber e 1 and Jug r 1. A comparison of the deduced amino
acid sequences of the sesame 2S albumin reported by Beyer et al. (2002b)
with that reported by Pastorello et al. (2001) shows differences sufficient for
them to be regarded as distinct allergens. The sesame 2S albumin reported
by Beyer et al. (2002b) is higher in sulfur-containing amino acids and is
deemed to be a novel allergen. It has been designated in the allergen
nomenclature as Ses i 2.
8.3.4 Shrimp allergens

Shellfish (shrimp) are a common cause of food allergy. The major IgE-
binding protein in shellfish has been designated in the allergen nomenclature
as Pen a 1. This allergen has been extensively researched and is an invertebrate
form of the contractile protein tropomyosin (Reese et al., 1999). Allergic
reactions to this protein have been noted in cockroaches and dust mites as
well as crustaceans (Asturias et al., 1998; Santos et al., 1999).
Recently, proteomics has been applied to further characterize the IgE-
binding proteins from shrimp. In particular, the black shrimp (Penaeus mondon)
of southeast Asia was the focus of this investigation (Yu et al., 2003). A
crude shrimp extract was separated by 2-DE and immunoblotted using a
pool of six shrimp-allergic patient sera. Numerous IgE-binding polypeptides
were observed in the 2-DE separation using a pool of six shrimp-allergic
sera. Approximately 10 IgE-binding spots were found in the 20–40 kD (pI
4.0–7.0) Mr range. The strongest IgE binding was found at 40 kD (pI 6.0)
and 34 kD (pI 4.6). They were excised, digested with trypsin and the resultant
peptides analyzed by MS. Spots 2 and 3 were, not surprisingly, identified as
the well-known allergen, tropomyosin (Pen a 1). However, analysis of spot
number 1 revealed a sequence that matched the enzyme arginine kinase, a
novel finding. To further examine the allergenicity of arginine kinase, the
enzyme was purified from other crustaceans (crab, lobster and crawfish),
tested for IgE binding, and analyzed by dot blot and competitive ELISA
using 13 shrimp-allergic sera. Arginine kinase from the other crustaceans
showed significant cross-reactivity, suggesting that it is also a common seafood
allergen. Arginine kinase was designated in the allergen nomenclature as Pen
m 2.
8.3.5 Bird’s nest allergen

Consuming the Chinese delicacy known as bird’s nest is a long practised
tradition. While some claims have been made for its medicinal value, it has
more recently become a suspect cause of food-induced anaphylaxis, especially
in children (Shek and Lee, 1999). Proteomics was used to analyze the proteins
from an aqueous extract of this material (Ou et al., 2001). The 2-DE separation
contained relatively few spots (< 30). Immunoblotting with the sera from a
bird’s nest-allergic patient showed that a single family of spots with a Mr of
approximately 66 kD was positive for IgE binding. Mass spectrometry analysis
of several internal peptide fragments corresponding to the 66 kD protein
resulted in its identification as a Kazal-type serine protease inhibitor, analogous

to egg white ovoinhibitor (Ou et al., 2001). Ovoinhibitor is a well known
egg-associated allergen and therefore linked to adverse reactions resulting
from consumption of this food.
8.3.6 Monitoring genetic modification

Agricultural biotechnology is well recognized for its potential benefits to
both producers and consumers. Bacillus thuringiensis-transformed varieties
express a Bt protein that inhibits insect damage. This transformation offers
the potential for increased yield while reducing chemical application. Similarly,
enhanced nutritional value is possible in golden rice, a variety that produces
the vitamin A precursor, beta carotene.
However, agricultural biotechnology has also come under intense criticism
because of the perceived potential for causing unintended consequences. A
major concern is that insertion of a novel gene or removal (silencing) of an
existing gene may alter the expression of undesirable components such as
allergens. Thus in many parts of the world, genetically modified crops are
not allowed in food production.
Proteomics represents an excellent technology for objectively assessing
the changes in crop varieties resulting from genetic modification or conventional
breeding as well. Recently, Herman et al., (2003) used proteomic analysis
together with immunoblotting and localization to assess the effectiveness of
gene silencing technology on the expression of a major soy allergen. The
allergen targeted for silencing in soybean was Gly m Bd 30K, also referred
to as P34. This low-abundance protein has been shown to be a major allergen
(Helm et al., 1998, 2000). P34 protein is associated with other globulins and
can be found after processing as a component of soy protein isolates, a
widely used food ingredient. Thus silencing the expression of this protein
would remove a major allergen from soybean.
Herman et al., (2003), used transgene-induced gene silencing to produce
soybeans that expressed little or no P34 protein. Two-dimensional
electrophoresis and image analysis of control and P34-silenced seed lines
showed very few changes. A comparison of control and silenced lines revealed
only five polypeptide spots with significantly altered expression. These spots
were down-regulated in the transgenic line two of the polypeptides with
positions in the gel corresponding to the pI and Mr range of P34 were positively
identified by MS analysis as the P34 allergen. This report serves as an
example of how proteomics can used to provide objective evaluation of
changes (or the lack thereof) induced by genetic modification. In this case,
proteomic analysis clearly showed that silencing technology had achieved its
desired objective of allergen reduction. The proteomic approach also
demonstrated that no unintended consequences had occurred.
8.3.7 Detection of IgE-binding proteins in foods

In general, a proteomic approach might not be suited to routine identification
and quantitation of IgE-binding proteins in foods because of the amount of
time required to perform the analyses. Additionally, appropriate protein
extraction protocols would be needed for food matrices. However, there are
examples when the time and effort required by the proteomic approach are
of importance.
Once such case can be found in the food ingredient lecithin. Lecithin,
derived from soybeans, is composed principally of phosphatidyl choline,
phosphatidyl inositol and phosphatidyl ethanolamine. Phospholipids are
excellent emulsifiers and are often used for that function in food and
pharmaceuticals. However, soybean lecithin has been shown to contain protein
at levels ranging from 50–1000 ppm. Several of these proteins also bind IgE
from soy-allergic patients (Awazuhara et al., 1998; Muller et al., 1998; Gu
et al., 2001).
Previous studies using 1-D SDS-PAGE and immunoblotting with soy-
allergic sera identified several IgE-binding bands (Awazuhara et al., 1998;
Muller et al., 1998; Gu et al., 2001). Recently Gu et al. (unpublished) used
2-DE, image analysis and immunoblotting to further examine the profile of
IgE-binding proteins associated with lecithin.
The presence of IgE-binding proteins in the separation was determined by
immunoblotting with the same soy allergic sera pool as used by Gu et al.,
2001. Numerous spots exhibited IgE binding and are summarized as follows.
Several spots in the 57 kD range (pI 4.2–5.0) bound IgE but were low in
abundance. Little or no protein can be seen in the 2-DE separation. However,
IgE binding was detected. Based on their pI and Mr, these spots were tentatively
identified as conglycinins.
All spots in the 39 kD range were strongly IgE-binding. They represented
a significant proportion of the total lecithin-associated protein. Collectively
the 39 kDa spots (pI 4.2–5.2) constitute about 23% of the sample protein.
However the 39 kD protein is a novel IgE-binding protein whose function
and location is unknown (Gu et al., 2001). The 39 kDa protein may be
associated with oil bodies as our recent findings with a polyclonal antibody
suggest (Xiang et al., unpublished).
The most diverse group of proteins was found in the 20 kD region. At
least eight individual polypeptides with pIs ranging from 4.25–9.7 were
observed with this Mr. Collectively, polypeptides in the 20 kD group make
up the largest fraction of lecithin-associated proteins. They comprise about
37% of the total sample protein. The IgE-binding, 20 kD spot (pI 4.4) was
tentatively identified (based on pI and Mr) as soybean Kunitz-type inhibitor
(SKTI).
Strong IgE-binding spots can also be found at 12 kD. They range in pI
from 4.38–5.3. The 12 kD (pI 4.38) spot is much greater in abundance than
the others in this Mr range and represents approximately 26% of the sample
protein. This 12 kD spot was tentatively identified (based on pI and Mr) as
a soybean 2S albumin (napin type) protein. This work further illustrates

advantages associated with proteomic analysis for allergen identification.
The approach provides uniquely definitive information in regard to the profile
and amount of proteins present.
8.4 Future trends

The future promises to bring many changes for proteomic technology. A
major focus of developmental effort is to replace 2-DE gels with other separation
technologies. Microfluidic devices are currently receiving considerable attention
for achieving this goal. Microfluidic devices are small platforms constructed
from glass or polymeric substances, typically equivalent in size to a microscope
slide. Separation channels with nanoliter internal volumes are etched into the
surface. Microfluidic separations are driven electrophoretically because of
separation efficiency and the difficulty in developing miniaturized mechanical
pumps. Separation times vary with the analyte but are typically accomplished
in seconds. Detection can be via laser-induced fluorescence or MS. It is
anticipated that a complete separation device may eventually be handheld,
i.e. the size of a pocket calculator. Numerous applications for microfludic
separations of proteins, peptides, nucleic acids, drugs, etc. have been reported
(Figeys and Pinto, 2001; Verpoorte 2002).
The separation of peptides from tryptic digests of complex (tissues or
cells) samples represents an immediate goal for development. A model for
the direction of future separation technology can be seen in ongoing research
of multi-dimensional CE-based methods (Chen et al., 2002; Shen and Smith,
2002). These separations employ various combinations of IEF, reversed-
phase, affinity or ion exchange to achieve resolution of complex peptide
mixtures. A further goal of the technology is to couple separation with MS
analysis to obtain several types of information, including profile (mass
fingerprint), sequence and modifications of the parent protein. This analytical
approach generates enormous amounts of data, requiring development of
appropriate bioinformatic programs.
8.5 Conclusions
Proteomics is an increasingly valuable tool in allergy research. Presently, it
offers enhanced potential for the detection and identification of IgE-binding
proteins in complex mixtures. All of the examples included in this review
showed that proteomic analysis resulted in new and more detailed information
regarding the profile of allergens in the materials examined. Proteomics was
also shown to be of value in the quantitative assessment of allergen content
in genetically modified soybeans. In the future, the development of new
separation (microfluidics) and detection (aptamers) technologies will further

enhance the value of proteomics to allergen investigations.
8.6 References
Asturias, J A, Arilla, M C, Gomez-Bayon, N, Martinez, A, Martinez, J and Palacios, R
(1998) ‘Sequencing and high level expression in Escherichia coli of the Tropomyosin
allergen (Der p 10) from Dermatophagoides pteronyssinus’, Biochim Biophysic Acta,
1397, 27–30.
Awazuhara, H, Kawai, H, Baba, M, Matsui, T and Komiyama, A (1998) ‘Antigenicity of
the proteins in soy lecithin and soy oil in soybean allergy’, Clin Exp Allergy, 28, 1559–
1564.
Berggren, K, Chernokalskaya, E, Steinberg, T H, Kemper, C, Lopez, M F, Diwu, Z,
Haugland, R P and Patton, W F (2000) ‘Background-free high sensitivity staining of
proteins in one- and two-dimensional sodium dodecyl sulfate polyacrylamide gels
using a luminescent ruthenium complex’, Electrophoresis, 21, 2509–2521.
Beyer, K, Grishina, G, Bardina, L, Grishin, A and Sampson, H A (2002a) ‘Identification
reactions’, J Allergy Clin Immunol, 110, 517–523.
Beyer, K, Bardina, L, Grishina, G and Sampson, H A (2002b) ‘Identification of sesame
seed allergens by 2-dimensional proteomics and Edman sequencing: seed storage
proteins as common food allergens’, J Allergy Clin Immunol, 110, 154–159.
Chen, J, Lee, C L, Shen, Y, Smith, R D and Baehrecke, E H (2002) ‘Integration of
capillary isoelectric focusing with capillary reversed-phase liquid chromatography for
two-dimensional proteomics separation’, Electrophoresis, 23, 3143–3148.
Figeys, D and Pinto, D (2001) ‘Proteomics on a chip: promising developments’,
Electrophoresis, 22, 208–216.
Gu, X, Beardslee, T, Zeece, M G, Sarath, G and Markwell, J (2001) ‘Identification of
IgE-binding proteins in lecithin’, Int Arch Allergy Immunol, 126, 218–225.
Helm, R M, Cockrell, G, Herman, E, Burks, A W, Sampson, H A and Bannon, G A (1998)
‘Cellular and molecular characterization of a major soybean allergen’, Int Arch Allergy
Immunol, 117, 29–37.
Helm, R M, Cockrell, G, West, C M, Herman, E, Burks, A W, Sampson, H A, Bannon, G
A and Burks, A W (2000) ‘Mutational analysis of the IgE-binding epitopes of P34/Gly
m 1’, J Allergy Clin Immunol, 105, 378–384.
Herman, E M, Helm, R M, Jung, R and Kinney, A J (2003) ‘Genetic modification removes
an immunodominant allergen from soybean’, Plant Physiol, 132, 36–43.
Hirschwehr, R, Valenta, R, Ebner, C, Sperr, F, Valent, W R, Rohac, M, Rumpold, H,
Scheiner, O and Kraft, D (1992) ‘Identification of common allergenic structures in
hazel pollen and hazelnuts: a possible explanation for sensitivity to hazelnut in patients
allergic to tree pollen’, J Allergy Clin Immunol, 909, 927–936.
Lopez, M F, Berggren, K, Chernokalskaya, E, Lazarev, A, M, Robinson, T H and Patton,
W F (2000) ‘A comparison of silver stain and SYPRO ruby protein gel stain with
respect to protein detection in two-dimensional gels and identification by peptide
mass profiling’, Electrophoresis, 21, 3673–3683.
Mann, M, Hendrickson, R C and Pandey, A (2001) ‘Analysis of proteins and proteomes
by mass spectrometry’, Annl Rev Biochem, 70, 437–474.
Miller, M D, Acey, R A, Yu-Tzu-Lee, L and Edwards, A J (2001) ‘Digital imaging
considerations for gel electrophoresis analysis systems’, Electrophoresis, 22, 791–
800.
Miura, K (2001) ‘Imaging and detection technologies for image analysis in electrophoresis’,
Muller, U, Weber, W, Hoffman, A, Franke, S, Lange, R and Vieths, S (1998) ‘Commercial

soybean lecithins: a source of hidden allergens?’, Z Lebens Unter Forsch A, 207, 341–
351.
Nishihara, J C and Champion, K M (2002) ‘Quantitative evaluation of proteins in one-
and two-dimensional polyacrylamide gels using a fluorescent stain’, Electrophoresis,
23, 2203–2215.
O’Farrell, P H (1975) ‘High resolution two-dimensional electrophoresis’, J Biol Chem,
250, 4007–4021.
Ou, K, Seow, T K, My Liang, R C Goh, D I L Chua, K Y and Chung, M C M (2001)
‘Identification of a serine protease inhibitor homologue in bird’s nest by an integrated
proteomics approach’, Electrophoresis, 22, 3589–3595.
Pastorello, E, Varin, E, Farioli, L, Pravettoni, V, Ortolani, C, Trambaioli, C, Fortunato, D,
Giuffrida, M G, Rivolta, R, Calamari, A M, Lacava, L and Conti, A (2001) ‘The major
allergen of sesame seeds (Sesame indicum) is a 2S albumin’, J Chromatogr B, 756,
85–93.
Posch, A, Weiss, W, Wheeler, C, Dunn, M J and Gorg, A (1995) ‘Sequence analysis of
wheat grain allergens separated by two-dimensional electrophoresis with immobilized
pH gradients’, Electrophoresis, 16, 1115–1119.
Rabjohn, P, Helm, R M, Stanley, J S, West, C M, Sampson, H A, Burks, A W and Bannon,
G A (1999) ‘Molecular cloning and epitope analysis of peanut allergen Ara h 3’, J Clin
Invest, 103, 535–542.
Reese, G, Ayuso, A and Lehrer, S B (1999) ‘Tropomyosin: an invertebrate panallergen’,
Int Arch Allergy Immunol, 119, 247–558.
Santos, A B, Chapman, R D, Aalberse, R C, Vailes, L D, Ferriani, V P L and Oliver, C
(1999) ‘Cockroach allergens and asthma in Brazil: identification of Tropomyosin as a
major allergen with potential cross-reactivity with mite and shrimp allergens’, J Allergy
Shefcheck, K J and Musser, S J (2004) ‘Conformation of the allergenic protein Ara h1 in
a model matrix using liquid chromatography/tandem mass spectrometry’, J Agric
Food Chem, 52, 2785–2790.
Shek, L P C and Lee, B W (1999) ‘Food allergy in children in the Singapore story’, Asian
Pac J Allergy Immunol, 17, 203–206.
Shen, Y and Smith, R D (2002) ‘Proteomics based high-efficiency capillary separations’,
Sporik, R and Hill, D (1996) ‘Allergy to peanut, nuts and sesame seed in Australian
children’, Brit Med J, 313, 1477–1478.
Weiss, W, Vogelmeier, C and Gorg, A (1993) ‘Electrophoretic characterization of wheat
grain allergens from different cultivars involved in baker’s asthma’, Electrophoresis.
14, 805–816.
Weiss, W, Huber, G, Engel, K, Pethran, A, Dunn, M J, Gooley, A A and Gorg, A (1997)
‘Identification and characterization of wheat grain albumin/globulin allergens’,
Xiang, P, Beardslee, T A, Zeece, M G, Markwell, J and Sarath, G (2002) ‘Identification
and analysis of a conserved IgE-binding epitope in soybean G1 a, G2 a and peanut Ara
h 3 glycinins’, Arch Biochem Biophys, 408(1), 51–57.
Verpoorte, E (2002), ‘Microfludic chips for clinical and forensic analysis’, Electrophoresis,
23, 677–712.
Yu, C J, Chia-Jung Lin, Y, Chiang, B L and Chow, L P (2003) ‘Proteomics and immunological
analysis of a novel shrimp allergen (Pen m 2)’, J Immunology, 170, 445–453.
9
Detecting food allergens with a surface

plasmon resonance immunoassay
H. Jonsson, A. Eriksson and I. Malmheden Yman,
National Food Administration, Sweden
9.1 Introduction
During recent years, awareness of food allergens has increased, resulting in
a growing demand for reliable, rapid and sensitive analytical methods to
detect them (Poms et al., 2004). Tracing and quantifying allergens in food is
often a challenging task. Allergens can be present in a wide variety of food
types, with different compositions and chemical properties. Some food
components may interfere with the assays. Food processing of different
kinds makes the extraction difficult. The sensitivity requirement is also high
since even very low doses may elicit allergic reactions. Tracing allergens in
food could very much be compared to finding a needle in a haystack. To be
able to reliably analyze as many allergens as possible in as many sample
matrixes as possible, a whole battery of methods is needed. Different analytical
techniques may complement each other.
As a research project, the Swedish National Food Administration has
started to investigate a novel analytical technique based on a surface plasmon
resonance (SPR) biosensor for food allergen detection. The objective is to
develop rapid and easy-to-use assays for common food allergens that would
complement existing techniques already employed, such as rocket
immunoelectrophoresis (RIE), enzyme-linked immunoassays (ELISAs) and
real-time PCR. Besides enhanced sensitivity, increased sample throughput is
a major driving force behind the development of the automated platform.
The high speed, ease-of-use and high degree of automation should also be
attractive features for food manufacturers who want to analyze ingredients
or final food products.
This chapter summarizes the literature on SPR technology related to food
Detecting with a surface plasmon resonance immunoassay 159
allergen detection and describes the work performed so far on developing

SPR biosensor immunoassays on the Biacore™ SPR system. (Biacore AB,
Uppsala, Sweden).
Section 9.2 gives a short introduction to biosensor technology, followed
in Section 9.3 by a description of the process of setting up an allergen SPR
assay. Published methods are summarised in Section 9.4 and then in Section
9.5 details of the assay are given together with a description of the specific
analysis conditions used and presentation and discussion of the results obtained.
9.2 Biosensors and SPR technology

Biosensors are analytical devices consisting of a biological recognition element
(e.g. cells, proteins or oligonucleotides) in direct contact with a transducer
that produces a signal, which is further processed to give an output that is
proportional to the concentration of a specific analyte. Transducers can be
optical, amperometrical, potentiometrical or acoustical. Surface plasmon
resonance employed in this study is an affinity-based optical transduction
principle that detects the binding between molecules through local changes
in the refractive index close to a surface. Currently, food analysis using SPR
is a rapidly growing field. So far most development has been in the area of
vitamins and veterinary drug residues, with ready-to-use kits commercially
available. Examples of protein SPR food immunoassays are detection of
Staphylococcal enterotoxin B (SEB) (Rasooly, 2001; Medina, 2003),
Salmonella antibodies in chicken (Jongerius-Gortemaker et al., 2002) and
fungal aflatoxin B1 (Daly et al., 2000). Feriotto et al. (2002, 2003) developed
an assay for genetically modified soy oligonucleotides. Detection of various
milk proteins has been reported by several groups (Haasnoot et al., 2001;
Indyk and Filonzi, 2003; Nygren et al., 2003; Müller-Renaud et al., 2004).
Analysis of peanut allergens using a miniaturized SPR system has been
demonstrated by Mohammed et al. (2001).
SPR instruments are available from different commercial sources.
Approximately 90% of all work published in the optical biosensor field has
used instruments from Biacore (Myszka, 1999; Rich and Myszka, 2000,
2001, 2002). In Biacore’s SPR technology, detection takes place in a flow
cell on an interchangeable sensor chip (Fig. 9.1). To construct a SPR biosensor,
the recognition element is immobilized on the surface which consists of a
gold film deposited on a glass support. The surface is covered by a dextran
layer which provides covalent binding sites when chemically activated and a
protein-friendly hydrophilic environment (Löfås, 1995). Samples are introduced
through an integrated microfluidic cartridge (Sjölander and Urbaniczky, 1991).
Built-in microvalves are used to accurately switch between sample and buffer
flow. Opposite the flow cell side of the sensor surface, a prism is optically
coupled to the sensor surface. Polarized light from a light-emitting diode is
reflected in the glass and detected by a charge-coupled diode (CCD) array.
Optical
detection Intensity
Light- unit
source
I
I II
Polarised II Angle
light Reflected
light Resonance
Prism II signal
Sensor chip
with gold film Time
I
Sensorgram
Flow channel
Fig. 9.1 Principles of surface plasmon resonance detection. Due to the SPR
phenomenon the reflected light will have an intensity minimum at a specific angle that
is continuously monitored. Binding of biomolecules to the sensor surface results in a
shift in the resonance angle.
At a specified resonance wavelength and angle, surface plasmons, i.e. clouds

of free electrons in the metal film, interact with the photons resulting in a
drop in the reflected light detected by the CCD. The resonance angle is
sensitive to the refractive index close to the surface. Liedberg et al. (1983)
first showed that biomolecules binding to the surface result in a refractive
index change, which in turn alters the resonance angle. The shift is proportional
to the mass bound to the surface. In the Biacore instrument, the resonance
angle is continuously monitored and converted to ‘resonance units’ (RU).
One RU corresponds to an angular shift of 0.0001° (Jönsson et al., 1991).
Plotting the resonance angle against time yields a sensorgram that describes
the interaction taking place on the surface in real-time.
Typically in SPR immunoassays samples are analyzed serially. Each sample
cycle consists of sample injection followed by a regeneration pulse that
washes away any bound sample molecules making the surface ready for the
next cycle (Fig. 9.2). Between injections a continuous flow of buffer is
maintained. In the food allergen assays developed, an extra injection of
antibodies was usually added after each sample injection. Binding of the
antibodies confirms the identity of the sample and can also be used to determine
the amount bound indirectly. This is similar to sandwich ELISA, but with the
great difference that all steps can be followed in real-time, without the need
to label any of the interactants. In contrast to ELISA, the regeneration,
however, allows the same antibody coated surface to be used for about 100–
200 times.
9.3 Developing a food allergen SPR immunoassay

This section describes the major steps in developing an SPR immunoassay
for food allergens.
Sensor response
Time
Buffer Buffer Buffer Buffer
Sample Secondary Regeneration
antibodies
Fig. 9.2 Sensorgram for one analysis cycle. This example shows a typical injection
sequence for a sandwich format assay, with secondary antibody injection after the
sample injection. Switching of solutions is indicated by arrows. The pictures above
illustrate schematically the surface coverage at different stages.
9.3.1 Preparation of the sensor surface: immobilizing the antibody

The first step in developing an SPR food allergen immunoassay is to prepare
a sensor surface coated with specific antibodies that bind the allergen in
question. Immunopurification on an affinity column coupled with the allergen
is highly recommended in order to obtain high responses and a specific
assay. In Biacore instruments, preparation of a surface covered by antibodies
is easily performed through a standardized and automated procedure that
covalently couples the antibodies via free primary amine groups to the
carboxylated dextran layer on the sensor surface with the sensor chip mounted
in the instrument (Löfås et al., 1995). With SPR detection the entire
immobilization procedure can be monitored in real-time, and the immobilized
level of antibodies can be read out as the increase in sensor response. For a
sensitive assay, high levels of immobilization are required. The dextran layer
on the Biacore CM5 chip extends the surface into a third dimension and
enhances the surface binding capacity, compared to a bare gold surface. In
the allergen assays developed, immobilization levels of 10 000–13 000 RU
of antibodies were typically achieved on the standard CM5 chip, which
corresponds to surface densities of 10–13 ng mm–2. For many of the experiments
a sensor chip CM3, with a thinner dextran layer than CM5, was used. For
CM3 chips the immobilization level was between 3000 and 5000 RU. These
levels were sufficient to allow detection in the ng mL –1 range.
9.3.2 Binding and regeneration

When an antibody-coated sensor surface has been prepared, the activity can
be checked by injecting a solution of the allergen. An increase in sensor
response of several hundred RU should be observed during the injection of
food allergens (two minutes, 10 μg mL–1). The next essential step is to find
a suitable regeneration solution able to remove all the bound analyte without
damaging the surface. This solution is injected after each sample. When a
good regeneration solution has been found, the same surface can be used to
analyze 100 or more samples, resulting in a very economical use of antibodies
and sensor chips. In the allergen assays developed so far, 0.1 M hydrochloric
acid has been proved to regenerate the surface quite well and can be used
initially for newly-developed assays. To further optimize the regeneration
for the specific antibody used, it is advisable to test the regeneration efficiency
with 10 mM glycine-HCl with pH ranging from 3 down to 1.5 in a series of
analysis cycles run on a newly immobilized surface. In each cycle, a high
concentration of allergen (e.g. 10 μg mL–1) is injected followed by the
regeneration solution. The mildest regeneration solutions (highest pH) should
be run first, and each condition should be repeated for five cycles before
switching to the next. A plot showing the response increase during the allergen
injection versus cycle number can then be used to determine the optimal
regeneration solution. The regeneration solution that gives the highest and
most stable analyte response should be chosen for the assay. Software guides
are available for this type of assay development in the Biacore Q system.
9.3.3 Sandwich assay

To enhance sensor signal and increase the specificity of the assay, a sandwich
format can be employed. In this assay format an injection of secondary
antibodies is performed after the injection of sample and before the regeneration
step (see Fig. 9.2). The regeneration step removes both the second antibody
as well as the allergen and allows repeated use of the sensor surface. Binding
of an allergen-specific antibody, readily seen as an increase in the sensor
output during injection, confirms the identity of the bound material on the
surface. Specific binding of allergen can in this way be distinguished from
non-specific binding from the sample matrix. With SPR the secondary antibody
needs no labelling to be detected. Furthermore, with polyclonal antibodies it
is possible to use the same antibody as for the immobilization. This convenient
strategy was found to work well for all tested allergens. Complete removal
of allergen during the regeneration is especially crucial in the sandwich
assays, because any remaining allergens on the surface in one cycle can bind
secondary antibody in the next. The secondary antibody injection step should
therefore be included when regeneration optimization is performed.
9.3.4 Calibration of the sensor response

For the determination of the concentration of allergens in unknown samples,
the sensor output has to be calibrated before unknown samples are run. This
is achieved by injecting a series of standard solutions with known concentrations
of allergens. Each standard is run exactly the same way as samples with
unknown concentrations, with regeneration of the surface at the end of the

cycle. For direct measurements the increase in response during the sample
injection is used for quantification. The response increment is measured
after the end of the injection when buffer flow has been re-established and
non-bound sample has been washed away. In this way the measurements
become independent of the refractive index of the sample extract. In the
sandwich assay the response increment during injection of the secondary
antibody is used instead. The calibration curves in the food allergen assays
were generally linear for low concentrations (< 1 μg mL–1) but flattened out
for higher concentrations due to saturation of binding sites on the surface. A
four-parameter non-linear function was used to fit the points in the calibration
curve.
9.3.5 Recovery studies

During assay development, recovery studies with spiked samples can be
undertaken to study the effect of the sample matrix on the assay. Spiking, i.e.
addition of known amounts of allergen, can be done in different stages
depending on the purpose of the experiment. In the development of the SPR
allergen assays two types of recovery studies were carried out: spiking of
extracts and spiking of food before extraction. Spiking of blank extracts,
made from samples known to be free from the allergen, was used to reveal
possible matrix inhibition of the binding to the surface. In general, matrix
inhibition turned out not to be a problem. However, for a dark chocolate
extract the specific binding to an anti-hazelnut protein surface was reduced
by 50%. The subsequent binding of secondary antibodies was also inhibited
by 50% resulting in a recovery of only 25% in the sandwich assay. Apparently,
this extract contained something, probably tannins, that bound to the hazelnut
proteins and inhibited the binding to the antibodies. When spiking was
performed before the extraction, the recovery of hazelnut proteins from dark
chocolate was even lower, indicating precipitation by tannins. In the hazelnut
assay, addition of 1% polyvinylpyrrolidone (PVP) during extraction was
found to restore the recovery in dark chocolate to 97–104%.
9.3.6 Analysis of food products

Once the sensor output has been calibrated, single samples can be analysed
very rapidly, within minutes. In the Biacore Q software the calculations of
concentration in unknown samples, including any dilutions made, are performed
in real-time and the results are presented immediately as soon as the sample
cycle is finished. The instrument can also be programmed to analyze up to
180 samples (two microplates) subsequently without any need of human
interaction. Control samples and recalibrations can be included at intervals.
In this way it is convenient to load the instrument in the evening, let the
instrument analyze the samples overnight and on the following day all results
are summarized in a software-generated report, ready to be printed out.
9.3.7 Advantages of real-time detection

SPR detection is not dependent on radioactive, fluorescent or enzyme labels.
All binding events on the surface can thus be followed in real-time, which is
very useful during assay development as well as during routine analysis.
Compared to endpoint analysis such as ELISA and rocket
immunoelectrophoresis, analysis by SPR gives additional information. For
instance, the amount of antibodies immobilized on the surface of a chip can
be quantified. ELISA does not offer similar direct control of the coating
levels. Although only a single point is used to construct the calibration curve
and to determine the concentration, the reviewing of the binding curves for
each sample cycle is an important quality control operation. Operational
errors can be identified by reviewing the response curves. For instance, an
empty vial will result in injection of air, which is readily revealed as a great
shift in the response during the injection period. As discussed below, non-
specific binding to the surface can be a problem for certain types of food
samples. Such conditions can often be identified by analyzing the binding
curve. Generally, non-specifically bound material dissociates faster than
allergens specifically bound to the antibodies. Thus, an unusually high negative
slope after injection is a warning signal of non-specific binding. A response
not returning to the base level after regeneration is another warning that not
only allergens have been bound to the surface. A trend plot of the baseline
response levels in the beginning of each cycle during run is thus an efficient
tool to reveal assay anomalies.
9.4 Published methods

No fully validated method for food allergen detection with SPR detection
has yet been published. Besides the results presented in Section 9.5, Mohammed
et al. (2001) have demonstrated food allergen analysis with SPR. They carried
out detection of peanut proteins on a miniature SPR system that could be
used to provide immediate, on-line detection and quantification during food
production. The limit of detection was 0.7 μg mL–1 of purified peanut protein
in buffer solution, which is about 100 times higher than that obtained for the
peanut assay developed on the Biacore (see below).
A few other SPR food protein assays have been published to date. Haasnoot
et al. (2001) developed a rapid (five minutes per sample) SPR immunoassay
for the detection of soy, pea and soluble wheat proteins. Although these
analytes are allergens, the main purpose with this assay was to reveal
adulteration of dairy products with non-milk proteins. A four-channel system
was employed, enabling simultaneous detection of all three species (the
fourth channel was used as a reference). The detection limits for the three
plant proteins in solution were 20–50 ng mL–1 and below 0.1% of plant protein
in total milk protein content. Prepared sensor chips could be used for more
than 650 sample injections, resulting in a very economical use of antibodies.
Rasooly (2001) reported a sensitive SPR immunoassay for the detection

of SEB, which employed a sandwich format. SEB was detected down to
10 ng mL–1 within five to eight minutes per sample. Little interference from
the food matrix was observed.
A quantitative SPR assay for beta casein in milk and cheese was developed
by Muller-Renaud et al. (2004). A two-step sandwich approach was used
involving secondary antibodies binding to the casein protein was used. In
this way native casein could be discriminated from its degradation products.
The analysis time was less than 10 minutes per sample and the same surface
could be used for more than 250 determinations. The detection limit was
85 ng mL–1.
9.5 Experimental data

This section presents the results from the development of SPR food allergen
assays at the National Food Administration, Sweden.
9.5.1 Assay details

Instrumentation and Reagents
The SPR system Biacore Q from Biacore AB was used in most analyses.
However, the initial experiments with beta-lactoglobulin, hazelnut and
ovomucoid were performed on a Biacore 2000.
CM5 (Research Grade) and CM3 sensor chips, Surfactant P20, HBS-EP
buffer (10 mM HEPES, pH 7.4, 0.15 M NaCl, 3 mM EDTA, 0.005% v/v
Surfactant P20), Amine Coupling kit containing 100 mM NHS (N-hydroxy-
succinimide), 400 mM EDC (N-ethyl-N′-(3-ethylaminopropyl) carbodiimide)
and 1 M ethanolamine, pH 8.5, were all bought from Biacore AB.
Polyvinylpyrrolidone (PVP, 150 kD) was obtained from Sigma (Sigma-Aldrich,
Stockholm, Sweden). All other chemicals were of analytical grade.
Protein fractions from hazelnuts, peanuts and sesame seeds were purified
at the National Food Administration according to Klein et al. (1985) and
crab protein according to Byun et al. (2000). Beta-lactoglobulin and ovomucoid
were purchased from Sigma. Polyclonal rabbit antibodies against beta-
lactoglobulin and ovomucoid were obtained from Pharmacia Diagnostics
(Uppsala, Sweden). Rabbit antibodies against hazelnut proteins, peanut proteins,
sesame seed proteins and crab tropomyosin were produced for the National
Food Administration by Agrisera AB (Vännäs, Sweden).
Surface preparation
Allergen-specific rabbit IgG from the immune sera was isolated by
immunoaffinity chromatography on a NHS-activated HiTrap™ column
(GE Healthcare, Uppsala, Sweden) coupled with the respective allergens
(affinity purification). In some of the initial experiments total IgG, as obtained
by purification with Protein G (GE Healthcare, Uppsala, Sweden), was

used.
The affinity-purified antibodies were covalently bound to the
carboxymethylated dextran layer of a CM5 or a F1 chip using the amine
coupling method pre-programmed in the instrument (Löfås et al., 1995). A
flow rate of 5 μL min–1 was used during immobilisation. The sensor chip
surface was activated with a mixture of NHS and EDC (1:1 v/v) for seven
minutes prior to incubation of the ligand solution. The ligand solution, consisting
of purified antibodies, diluted to a concentration of about 50 μg mL–1 in
10 mM sodium acetate, pH 4, was injected for seven minutes, followed by
a seven minute injection of ethanolamine to block remaining active sites.
Newly prepared surfaces were washed with three short pulses or more of
regeneration solution (0.1 M HCl) prior to use.
Sample extraction and analysis

Between 0.5 and 3.0 g homogenized sample was extracted in 10 mL extraction
buffer (20 mM Tris, pH 8.7, 0.5 M NaCl) in an ultrasonic bath at 45 °C for
one hour, followed by centrifugation at 30 800 g for 10 minutes at 4 °C. One
percent (1%) PVP was added to the extraction buffer when analyzing dark
chocolate. The supernatant solutions were collected in 1.5 mL vials.
Immediately before analysis, extracts were centrifuged to remove particles.
Biosensor analysis was performed using a running buffer composed of the
extraction buffer with addition of 0.05% Surfactant P20 to reduce the non-
specific binding. Samples were injected for two minutes at a flow rate of
40 μL min–1, followed by a two minute washing period with running buffer.
Secondary antibodies, diluted to about 20 μg mL–1 in running buffer, were
injected for 60 seconds at a flow rate of 10 μL min–1. The surface was
typically regenerated with a 60 second pulse of 0.1 M hydrochloric acid.
9.5.2 Results and discussion

Evaluation of the SPR system for food allergen analysis has been made with
several important food allergens, including egg ovomucoid, milk beta-
lactoglobulin, hazelnut proteins, peanut proteins, sesame seed proteins and
crab tropomyosin, a common allergen in crustaceans. The results obtained so
far are briefly summarized below.
Initial proof-of-principle experiments with ovomucoid, beta-lactoglobulin

and hazelnut proteins
Ovomucoid, beta-lactoglobulin and hazelnut proteins were included in the
first round of tests to evaluate the versatility of the method. Antibodies
against ovomucoid, beta-lactoglobulin and hazelnut protein were successfully
immobilized on sensor surfaces. These antibodies were the IgG fraction
from antiserum purified on a Protein G affinity column. Figure 9.3 shows the
calibration curves for the three allergens based on total binding of pure
250 Beta-lactoglobulin (18 kDa)

Ovomucoid (28 kDa)
Total sample binding (RU)
Hazelnut proteins (21–57 kDa)

200
150
100
50
0
0 200 400 600 800 1000 1200
Concentration (ng/mL–1)
Fig. 9.3 Calibration curves for beta-lactoglobulin, ovomucoid and hazelnut proteins.
Two-minute injections were used.
solutions. This initial investigation showed that the sensitivity of the SPR
detector is sufficient for detection of food allergens. Compared to rocket
immunoelectrophoresis using the same antibodies the sensitivity was at least
ten times more sensitive.
Analysis of food extracts, however, showed that the assays were not yet
specific enough. Some types of food extracts bound non-specifically to the
surface. Since everything that binds to the surface will contribute to the
sensor signal in SPR, non-specific binding will result in false-positive results
in measurements based on total sample binding. By comparing the response
on surfaces with or without immobilized antibodies, it was shown that most
of the non-specific binding occurred to the antibodies, not to the surface. The
dextran-coated surface was quite inert. Extracts from bread and other bakery
products never gave high non-specific binding. Chocolate extracts, on the
other hand, were very important for the hazelnut assay, often bound to a very
high degree making sensitive measurements of hazelnuts impossible.
One reason for the non-specific binding to the immobilized antibodies is
that the Protein G-purified antiserum contains a mixture of IgG towards all
food proteins that rabbits have encountered during their life. One way to
enhance the specificity is to isolate only specific antibodies (see below).
Fighting non-specific binding

In the further development of the hazelnut assay, many different means of
reducing the non-specific binding were tested (Jonsson and Hellenäs, 2001).
Firstly, the sensor surface was made more specific by isolation of antibodies
on an affinity column coupled with hazelnut proteins. This resulted in 50%
higher sensor response, at the same level of immobilization, compared to
antibodies purified on the Protein G column. Switching from the most common
type of sensor chip (CM5) to a variant with a thinner dextran layer (CM3)
decreased the non-specific binding of a dark chocolate extract by 55% while
the specific binding of hazelnut proteins remained almost the same. Still, the
non-specific binding was too high for several types of sample extracts.
Increasing the concentration of salt and increasing the pH of the buffer
further reduced the non-specific binding from chocolate extracts. However,
these buffer modifications also reduced the specific binding. Increasing the
amount of surfactant (P20) reduced the non-specific binding for dark chocolate
extracts (Jonsson and Hellenäs, 2001). In contrast, increasing the surfactant
concentration increased the non-specific binding of milk chocolate extracts.
The best way of reducing the high non-specific binding for dark chocolate
extracts was found to be addition of PVP during the extraction. This also
greatly increased recovery of hazelnut proteins in spiked samples.
This examination indicates that it is possible to optimize the assay conditions
in the SPR assays so that sensitive measurements can be performed even in
food matrices that contain many interfering substances. When the composition
of the analyzed food extract is known and similar from time to time, this can
be used to obtain a very rapid analysis of samples. For instance, this could
be used in continuous monitoring of a manufacturing process. In most situations,
however, the compositions of the samples are not very well known and differ
much from sample to sample. Under such assay circumstances, it is very
difficult to find conditions that will give very little non-specific binding for
all types of sample extracts. The best way of circumvent this problem was
found to be extending the assay slightly by adding an injection of antibodies
after the sample, creating a ‘sandwich’ format.
Increasing the specificity through the sandwich format

In the sandwich format, injection of a secondary antibody is used to increase
the specificity of the assay. Binding of the secondary antibody is used to
confirm the identity of the material bound during the sample injection. As
SPR detects everything that binds to the surface, the secondary antibodies do
not have to be labelled in any way to be detected. Binding of the secondary
antibody can be followed second by second, just like the binding of samples.
The binding will be proportional to the amount of the allergen bound on the
surface during the sample injection. Specific antibodies will not bind to non-
specifically bound material from the sample extracts. Thus, using the response
increase during injection of the secondary antibody instead of the increase
during sample injection to determine the concentration will give a much
more specific assay.
Table 9.1 shows a comparison of cross-reactivity obtained in the hazelnut
assays utilizing both direct and indirect measurements. While virtually all
samples cross-reacted to some degree in the direct binding assay, only a few
species cross-reacted in the sandwich assay using affinity-purified antibodies.
The minor cross-reactivity that was observed in the latter approach was
towards walnut, coconut, cashew and pecan, similar to what was earlier
described for affinity-purified antibody-based sandwich ELISA for hazelnut
(Koppelman et al., 1999). These results show that very specific SPR assays
can be achieved by using a second antibody in a sandwich approach.
Table 9.1 Cross-reactivity expressed as ppm hazelnut protein in solid food sample. All
samples were diluted a hundred times before analysis. Concentrations assigned as >100
indicate that the level was above the measurement range when normal assay conditions
were applied. Only hazelnut extract was further diluted to allow higher concentrations to
be determined.
Sample Direct binding assay Sandwich assay
Hazelnut 98 700 97 200

Walnut > 100 2.2
Pecan > 100 2.9
Cashew n.d. 1.3
Pistache 55 n.d.
Brazil nut 76 n.d.
Nutmeg >100 n.d.
Almond 35 n.d.
Sesame 54 n.d.
Peanut 45 n.d.
Coconut 83 n.d.
Pine kernel 24 n.d.
Dark chocolate 0 n.d.
Milk chocolate 10 0
n.d. = not detected (< 10 ppm)
Precision and sensitivity for the hazelnut assay

Replicate injections of different concentrations of hazelnut proteins showed
that the response was repeatable. For concentrations between 0.05 and
1.0 μg mL–1 the variation (relative standard deviation, four replicates) was
less than 1.5%. Based on the response of buffer injections, the limit of
detection in purified solutions of hazelnut proteins was calculated to 5 ng
mL–1 (three standard deviations above the average response for buffer
injections). This is the same order of magnitude that was achieved using
hazelnut ELISA (Holzhauzer and Vieths, 1999; Koppelman et al., 1999).
Compared with a previous report of analysis of peanut proteins on a
miniaturized SPR system, this represents a hundredfold improvement
(Mohammed et al., 2001). However, analysis of purified allergens in buffer
is an idealized situation. The actual limit of detection will be determined by
the background response of real blank food extracts. The dilution of extracts
also influenced the sensitivity. Preliminary results for the hazelnut assay
showed that the limit of detection is about 0.5 mg protein per 100 g in
chocolate (5 ppm), when samples were diluted 100 times before analysis.
Later results with a peanut assay showed that it is possible to analyze samples
that were diluted only ten-fold. This could improve the limit of detection
even more. As the response was still increasing when the sample injection
was ended, the maximum surface binding capacity had not been reached.
The limit of detection may therefore also be improved by increasing the
contact time. In the present study two minute injections were used to optimize
the assay for speed. Compared to rocket immunoelectrophoresis assays
currently used for routine analysis at the Swedish National Food Administration,
the SPR hazelnut assays were ten-fold more sensitive. The limit of detection
was in the same range as commercially available kits for hazelnuts.
Assay for peanuts

A peanut assay was established using the same conditions as for hazelnut.
For comparison between the SPR immunoassay and the RIE and commercially
available enzyme immunoassays (EIAs), model chocolate samples, spiked
with peanut butter, were analyzed. The model chocolates were made by
melting different amounts of peanut butter in a peanut-free milk chocolate at
70 °C. The peanut butter contained about 25% protein according to the
ingredient list. Chocolate samples with peanut protein levels from 12 500
down to 0.375 ppm were produced (Table 9.2). The RIE assay (Malmheden
Yman et al., 1994) detected peanut proteins down to 70 ppm. The SPR
immunoassay extended the measurement range down to 1 ppm, giving a
response significantly above the response for blank samples. The Ridascreen®
enzyme immunoassay (R-Biopharm, Darmstadt, Germany) had a similar
sensitivity, while the BioKits™ peanut assay (Tepnel BioSystems, Deeside,
UK) was even more sensitive. A good correlation between the SPR assay and
both the ELISA kits was seen. The recovery of the SPR assay was similar to
that of the Ridascreeen assay, but lower than that of the BioKits™ assay.
Egg protein assay

Conalbumin (ovotransferrin) (Sigma-Aldrich, Sweden) was used for
immunization of a rabbit. The protein was also coupled to a HiTrap™ column
as described above for hazelnut, for the immunopurification of specific rabbit
IgG to conalbumin. The purified IgG was immobilized on a CM3 chip. Pasta
samples were extracted with Tris-HCl buffer, pH 8.7. The extracts were
further diluted in HBS-EP buffer, before being applied to the chip. The result
for positive samples ranged from 0.29 to 6.8 ppm conalbumin. The results
were compared to those obtained by analyzing the same samples for another
egg protein, ovalbumin, by RIE. All samples positive with the RIE were also
positive in the SPR assay. Due to the low level of ovalbumin in the pasta
samples, near or below the quantification limit of RIE, a quantitative comparison
between the RIE and the SPR assay was not possible.
Sesame seed assay

Antiserum to purified sesame seed proteins was produced in a rabbit. The
antiserum achieved was not mono-specific towards a single sesame seed
protein, but showed reactivity to several proteins with molecular mass ranging
from 10 up to 70 kDa. The dominating bands had a molecular mass of 21 and
30 respectively, as revealed by immunoblotting. The antiserum was
immunoaffinity purified on a HiTrap™ column, coupled with the protein.
Analysis by SPR in an indirect (sandwich) assay showed high sensitivity and
selectivity for sesame seed proteins. Sesame seed protein could be detected
Table 9.2 Comparison of four different antibody-based methods for peanut protein analysis
Sample Peanut butter Chocolate Calculated RIE peanut SPR BioKits™ EIA Ridascreen
amount of protein immunoassay peanut protein EIA peanut
peanut protein (ppm) peanut protein (ppm) protein
(ppm) (ppm) (ppm)
1 0.5 g 9.5 g 12 500 13 000

2 0.1 g 9.9 g 2500 3250
3 0.05 g 9.95 g 1250 1630
4 0.025 g 9.975 g 625 740
5 0.013 g 9.9875 g 310 350 226
6 0.0062 g 9.9938 g 150 70 75 133 80
7 5 g of sample 6 5g 75 38 53 42
8 5 g of sample 7 5g 37.5 17 26 19
9 1 g of sample 8 5g 3.75 1.0 2.35 1.7
10 1 g of sample 9 5g 0.375 0.125
at a level of 0.125 μg mL–1, corresponding to 12.5 ppm in solid food in both

a direct assay as well as in a sandwich assay. In comparison, the immunoblotting
technique used could detect 8 ppm sesame seed protein in purified solution.
However, the latter technique is quite laborious and time-consuming. It is
only semi-quanitative. The established RIE, based on the precipitation between
the antigen and the corresponding antibody, gave very diffuse and faint
staining rockets with the antiserum produced and was regarded as not suitable
for the purpose of analyzing sesame seed protein in food samples.
Tropomyosin assay
Recently, crab tropomyosin was purified and antisera raised in a rabbit as
well as in eggs (egg yolk) by immunizing two hens. The rabbit IgG as well
as the IgY from egg yolks was immunopurified on a HiTrap™ column,
coupled with tropomyosin. Rabbit IgG or hen IgY was coupled to CM3
chips according to the established procedure. The chips were tested both in
a direct and in a sandwich assay with different solutions of tropomyosin,
with crab, shrimp and model foods, composed of surimi containing different
amount of crab meat. The lowest concentration detected in a tropomyosin
solution was 0.2 μg mL–1. In model products, a level of 0.063% crab meat
in surimi could be detected, corresponding to a tropomyosin concentration
of 13 ppm tropomyosin in solid food. The response both in the direct as well
as in the sandwich assay was virtually the same. By using IgY bound to the
surface and rabbit IgG as the secondary antibody, the non-specific binding to
the chip surface was reduced.
9.6 Conclusions
The SPR technique has proven to be a useful tool for the analysis of allergens
in food. Analysis of six important food allergens has been demonstrated.
SPR provides non-destructive detection that requires no labeling or any other
type of modification of the analytes detected. This gives unsurpassed control
during assay development and routine analysis, not available in established
methods of food allergen detection. Limits of detection in the low ppm level
can be obtained and with the sandwich format, very specific assays can be
achieved.
The main advantage of the SPR methodology is the short analysis time
per sample. This is important for large series of samples and for real-time
process monitoring. In the future, SPR analysis of food allergens may be
extended to analyze several allergens in the same extract simultaneously by
injection over several surfaces immobilized with antibodies raised against
different species.

• More information on SPR analysis in general, and food analysis in particular,
can be found on the Biacore company homepage www.biacore.com/food.
• FoodSense was a EU sponsored project with the goal of exploring the
capability of SPR biosensors in the food industry. More information on
the FoodSense project can be found on the project’s homepage www.slv.se/
foodsense
9.8 References
Byun, M W, Kim, J H, Lee, J W, Park, J W, Hong, C H and Kang, I J (2000) ‘Effects of
gamma radiation on the conformational and antigenic properties of a heat-stable major
allergen in brown shrimp’, J Food Prot, 63, 940–944.
Daly, S J, Keating, G J, Dillon, P P, Manning, B M, O’Kennedy, R, Lee, H A and Morgan,
M R (2000) ‘Development of surface plasmon resonance-based immunoassay for
aflatoxin B(1)’, J Agric Food Chem, 48, 5097–5104.
Feriotto, G, Borgatti, M, Mischiati, C, Bianchi, N and Gambari, R (2002) ‘Biosensor
technology and surface plasmon resonance for real-time detection of genetically modified
roundup ready soybean gene sequences’, J Agric Food Chem, 50, 955–962.
Feriotto, G, Gardenghi, S and Gambari, R (2003) ‘SPR-based assays for real-time detection
of genetically modified organisms’, Biacore J, 2, 5–8.
Haasnoot, W, Olieman, K, Cazemier, G and Verheijen, R (2001) ‘Direct biosensor
immunoassays for the detection of nonmilk proteins in milk powder’, J Agric Food
Chem, 49, 5201–5206.
Holzhauser, T and Vieths, S (1999) ‘Quantitative sandwich ELISA for determination of
traces of hazelnut (Corylus avellana) protein in complex food matrixes’, J Agric Food
Chem, 47, 4209–4218.
Indyk, H E and Filonzi, E L (2003) ‘Determination of immunoglobulin G in bovine
colostrum and milk by direct biosensor SPR-immunoassay’, J AOAC Int, 86, 386–393.
Jongerius-Gortemaker, B G M, Goverde, R L J, van Knapen, F and Bergwerff, A A (2002)
‘Surface plasmon resonance (BIACORE) detection of serum antibodies against
Salmonella enteritidis and Salmonella typhimurium’ J Immunol Methods, 266, 33–44.
Jonsson, H and Hellenäs, K-E (2001) ‘Optimizing assay conditions in the detection of
food allergens with Biacore’s SPR technology’, Biacore J, 1(2), 16–18.
Jönsson, U, Fägerstam, L, Ivarsson, B, Johnsson, B, Karlsson, R, Lundh, K, Löfås, S,
Persson, B, Roos, H and Rönnberg, I (1991) ‘Real-time biospecific interaction analysis
using surface plasmon resonance and a sensor chip technology’, Biotechniques, 11,
620–627.
Klein, E, Baudner, S and Günther, H O (1985) ‘Immunochemische Bestimmung des
Hasselnussproteins mit Hilfe der Elektroimmundiffusion nach Laurell. I. Metteilungen
der optimierten Methode’, Z Lebensm Unters Forsch, 180, 30–35.
Koppelman, S J, Knulst, A C, Koers, W J, Penninks, A H, Peppelman, H, Vlooswijk, R,
Pigmans, I, Van Duijn, G and Hessing, M (1999) ‘Comparison of different
immunochemical methods for the detection and quantification of hazelnut proteins in
food products’, J Immunol Methods, 229, 107–120.
Liedberg, B, Nylander, C and Lundström, I (1983) ‘Surface plasmon resonance for gas
detection and biosensing’, Sensors and Actuators, 4, 299–304.
Löfås, S (1995) ‘Dextran modified self-assembled monolayer surfaces for use in
biointeraction analysis with surface plasmon resonance’, Pure Appl Chem, 67, 829–
834.
Löfås, S, Johnsson, B, Edström, Å, Hansson, A, Lindquist, G, Müller Hillgren, R and

Stigh, L (1995) ‘Methods for site controlled coupling to carboxymethyldextran surfaces
in surface plasmon resonance sensors’, Biosens & Bioelectron, 10, 813–822.
Malmheden Yman, I, Eriksson, A, Everitt, G, Yman, L and Karlsson, T (1994) ‘Analysis
of food proteins for verification of contamination or mislabeling’, Food Agric Immunol,
6, 167–172.
Medina, M B (2003) ‘Detection of staphylococcal enterotoxin B (SEB) with surface
plasmon resonance biosensor’, J Rapid Methods Autom in Microbiol, 11, 225–243.
Mohammed, I, Mullet, W M, Lai, E P C and Yeung, J M (2001) ‘Is biosensor a viable
method for food allergen detection?’, Anal Chim Acta, 444, 97–102.
Muller-Renaud, S P, Dupont, D and Delieu, P (2004) ‘Quantification of beta-casein in
milk and cheese using an optical immunosensor’, J Agric Food Chem, 52, 659–664.
Myszka, D G (1999) ‘Survey of the 1998 optical biosensor literature’, J Mol Recognit,
12, 390–408.
Nygren, L, Sternesjö, A and Björck, L (2003) ‘Determination of folate-binding proteins
from milk by optical biosensor analysis’, Int Dairy J, 13, 283–290.
Poms, R E, Klein, C L and Anklam, E (2004) ‘Methods for allergen analysis in food: a
review’. Food Addit Contam, 21, 1–31.
Rasooly, A (2001) ‘Surface plasmon resonance analysis of staphylococcal enterotoxin B
in food’, J Food Prot, 64, 37–43.
Rich, R and Myszka, D G (2000) ‘Survey of the 1999 surface plasmon resonance biosensor
literature’, J Mol Recog, 13, 388–407.
Rich, R and Myszka, D G (2001) ‘Survey of the year 2000 commercial optical biosensor
Rich, R and Myszka, D G (2002) ‘Survey of the year 2001 commercial optical biosensor
Sjölander, S and Urbaniczky, C (1991) ‘Integrated fluid handling system for biomolecular
interaction analysis’, Anal Chem, 63, 2338–2345.
10
The use of lateral flow devices to detect

food allergens
R. Van Herwijnen, European Veterinary Laboratory, The
Netherlands and S. Baumgartner, Department for
Agrobiotechnology, IFA – Tulln, Austria
10.1 Introduction
In recent years, the number of solid phase-based immunoassays available
has further increased. Most of these are enzyme-linked immunosorbent assays
(ELISAs). Among the newly introduced tests, a significant number have
been developed either to identify the presence of allergenic substances in
e.g. food preparations, such as peanuts or hazelnuts in cookies, or to measure
in sera the antibodies (IgE) that patients may have developed in response to
repeated exposure to such allergens. These types of tests show a very high
specificity and a high sensitivity.
In general, an analytical methodology is sought where the detection of
offending foods or allergenic proteins needs to be managed with high specificity,
sensitivity and rapidity. In the following, typical demanded characteristics
for the methodology are listed.
• Sufficient specificity – should allow the detection of the analyte in question
in a wide variety of matrices including processed foods.
• Satisfactory sensitivity – detection in the range of 1–5 mg/kg is necessary.
• Rapid detection of contamination – short reaction time for correction
before large volumes of mislabelled products have been made or to give
the consumer a yes/no decision within an adequate time span.
• Ease-of-use – no specially trained employees needed or less experienced
persons can perform the assay.
• For industrial purposes – continuous sampling could be advantageous.
Most of the immunochemical methods mentioned above are considered
rapid, but there is still great demand for easy-to-use immunochemical tests
for the detection of traces of allergenic proteins in food. The faster commercial
ELISA kits for allergens on the market take only 30 minutes, and some of
them (qualitative determinations) do not require a high level of expertise, as
they have dropper bottles and visual comparisons. However, the more sensitive,
quantitative tests take a trained technician to carry them out; the desire to
improve on speed and ease-of-use, especially in non-research settings, has
prompted researchers to look for faster and simpler procedures. Thus, new
rapid test platforms have been developed, such as the chromatographic test
strip (lateral flow test). This is widely used for example in so-called ‘home
tests’, such as hCG measuring pregnancy tests, which do not require assistance
from trained technicians, but can be carried out by the consumers themselves.
The first generation of such tests was more or less based on variations on the
classic ‘dot blot’ test. From the experience gained with these early rapid
assays, much was learned regarding attainable sensitivities, and suitable
read-out systems. However, they did not take full advantage of the increased
kinetics, due to close contact of high concentrations of the analyte with
antibody. More recently, a second generation of immunomigration media
has been developed which exploit the micropores in nitrocellulose or nylon
membranes. The success of this new generation of so-called ‘lateral flow’
tests is based on the fast flow of fluids running through the membrane
enabling the application of various immunoreactants at different locations
along the membrane strip. There are currently four dipstick-type tests
commercially available, two for detection of peanut and two for detection of
gluten – these will be discussed later in the chapter.
10.2 Antibodies
10.2.1 Antibody requirements
There is more to designing one-step tests than meets the eye. Components
proven to be useful in ELISA are not necessarily adequate for one-step tests.
The environment in which they have to bind in ELISA is totally different
from that in one-step tests – in the latter all components are acting
simultaneously and within a very short time-frame. Thus, selecting antibodies
for a one-step test requires their affinities to be sufficient and yet mutually
tolerant so that they will allow each other to participate in forming a visible
complex within the short time that they flow across the test and control lines.
Also, it is more critical that they do not compete for the same epitopes. In
addition, gold labeling has a greater interference with antibodies, where
binding affects overall charge and secondary and tertiary structure more than
labeling with enzymes or biotin. As an example, rabbit antibodies do not
retain gold for more than a few days, while gold-labeled goat antibodies are
very stable.
The use of lateral flow devices to detect food allergens 177
10.2.2 Subclass influences

Suitable antibodies can be found among all IgG subclasses. In contrast, IgM-
type antibodies may be suitable as capture antibodies on the test or control
lines (see Section 10.3 and Figs 7.1 and 7.3); however, attaching gold label
to them is cumbersome and, due to their pentameric structure, they tend to
envelope the allergen tested, preventing their capture at the test and control
lines.
10.3 Constructing a lateral flow device (LFD)

A lateral flow test comprises five different basic components.
1. Sample filter. The sample filter is a paper-like material with two main
functions: filtering out any solid material (food or serum particles), and
buffering the sample after extraction/pretreatment (see Fig. 10.1).
2. Conjugate pad (gold pad). The conjugate pad is a fibreglass-like pad.
This can either be sprayed by, or bathed in, gold-, latex- or carbon-conjugated
solutions:
• spraying gives a well-defined, reproducible amount applied per mm2;
• bathing gives a non-controllable amount applied per mm2, which results
in variable intensities of both test and control lines.
The conjugate particles can be of various sizes (5–200 nm), which will
influence test sensitivity.
3. Membrane. The membrane is composed of nylon or nitrocellulose, usually
glued upon a backing material (made of polypropylene, polyethylene or
Sample filter
Gold pad
Test line
Control line
• Sample filter
• Gold pad
• Membrane
• Test line
• Control line Membrane
• Absorption filter
• Store at 10–28 °C
Absorption filter
Fig. 10.1 Components of the one-step assay (store at 10–28 °C).

polystyrene). These membranes are made to display a maze (fishnet)

structure with various pore sizes. The size of the pores will influence both
the speed and sensitivity of the test.
4. Reservoir (absorption pad). The reservoir has only one purpose: to adsorb
the liquid at the end of the strip. It is usually made of paper-like material.
The material can basically be any material with high wetting capacity.
The adsorption capacity of the reservoir causes the sample to continue to
flow over the nitrocellulose or nylon membrane.
5. Test line and control lines. The test and control lines are sprayed onto
the nitrocellulose/nylon membrane with an instrument delivering a precise
volume per mm2. The system and speed used determine the quality of the
lines.
10.4 Running a sample

A sample, suspected to contain the analyte in question (extracted or not), is
applied to the sample filter, housed inside the device casing. The sample runs
through the sample filter and conjugate pad. This conjugate pad contains the
affinity-purified and labelled (gold/latex/carbon) antibody, specific for the
analyte under test. The analyte, if present, will form a complex with the
conjugate and migrate further along the membrane to the test zone. This test
zone contains an immobilised antibody, specific for the analyte, but preferably
not competing with the conjugated antibody for the same or adjacent epitopes.
This test line will thus capture the migrating analyte–conjugate complex.
The intensity of the test line correlates well with the amount of analyte in the
sample.
Thus these tests can also be read semi-quantitatively by use of reader
equipment, based on change-coupled diode (CCD) cameras and analysing
software. These readers use calibration curves to calibrate and calculate the
concentration of the analyte in the sample. To ensure that the test has run to
completion a control system is included in it. This will normally be a control
line, immobilised at the end of the membrane (control zone), usually composed
of anti-conjugate antibodies, e.g. in the case where mouse monoclonal
antibodies are used as the basis for the conjugate, antibodies to mouse
immunoglobulins will be used for the control line at the end of the strip.
In general, the following results will be obtained (Fig. 10.2):
• positive result – two coloured lines will become visible;
• negative result – only one, the control line, will become visible;
• invalid result – all other possibilities (no line, only test line).
In this type of test several test formats can be used (Fig. 10.3):
• antigen detection – measuring the presence of the analyte as described
above;
C Control zone C Control zone C Control zone
T Test zone T Test zone T Test zone
Sample zone Sample zone Sample zone
S S S
(a) Positive (b) Negative (c) Not valid
Fig. 10.2 One-step assay results (in general).
• Antigen-detection
• Antibody-detection
• Inhibition assay
Fig. 10.3 Assay principles.
• antibody detection – measuring the antibodies reacting with a certain

antigen (e.g. IgE against peanut or wheat proteins);
• inhibition assay – usually used to detect small analytes (hormones,
pesticides), but proven useful in some cases for larger molecules like
allergens. In this type of test only one control line means a positive result
(the analyte is present in the sample and blocks the reaction), two lines
means a negative result. Some commercially available LFDs for allergenic
residues use this format.
10.5 Methods available

At the time of writing, there is little information and published data to be
found concerning commercially available dipsticks for use in the food industry.
In Europe, the ability to detect peanut residues is in great demand. With the
combination of rabbit anti-conarachin IgG, biotinylated rabbit anti-conarachin
IgG, avidin-horseradish peroxidase and a paddle-type dipstick, a rapid test
was achieved with a detection limit of 0.01% of unprocessed peanut in
marzipan (Mills et al., 1997), which corresponds to a detection limit of
0.0007–0.001% of conarachin. A commercially available LFD is the Gluten
Home Test Kit, a dry-strip immunochemistry format from Tepnel Biosystems
Ltd (Deeside, UK) for the qualitative determination of ω-gliadin. Detection
limits of 50–200 ppm were claimed to be reached in cakes for example
and 200–1200 ppm in highly-processed flour. Also, R-Biopharm AG
(Darmstadt, Germany) has an LFD device (RIDA®QUICK Gliadin) for gluten
detection that has a detection limit of 5 ppm gluten (corresponding to 2.5
ppm gliadin).
Additionally, two LFDs for the detection of peanut residues were launched
in 2004, the first by Neogen Corporation, Lansing, MI (Reveal® for Peanut),
followed some months later by Tepnel Biosystems (BioKits® Rapid Peanut).
Both tests work according to the single-step lateral flow
immunochromatographic principle. The Tepnel method uses antibodies against
conarachin and the Neogen antibodies are directed against multiple peanut
proteins. In terms of detection limits or working ranges, both tests refer to a
low ppm level. The Neogen Reveal® test screens samples at 5 ppm peanut
and can be used for all foods, rinse waters, etc., whereas the instructions for
the Tepnel method indicate that the test can be conducted in defined matrices
– milk and dark chocolate, cereals, biscuits, ice cream and nuts. Although
much academic research is currently directed towards the development of
dipsticks for the determination of allergenic residues, only the few mentioned
above are commercially available. Even though fast test systems make it a
lot easier to monitor adulteration and unintentional inclusion or contamination,
with the exception of the gluten ‘home’ test, they should be restricted to
manufacturing/industrial use or food labelling enforcement testing – this is
also indicated in the test kit inserts. It is not the test itself or the handling
which make at home or restaurant use risky; rather, it is problems of accurate
sampling and efficient extraction of the protein in question from food that
make this infeasible.
10.6 Future trends

Many commercial systems that allow fully automated testing, mostly based
on ELISA, are already on the market and commercial LFDs for allergenic
residue detection are now starting to appear. The use of new recombinant
antibodies and molecular imprinting techniques will improve the sensitivity

and versatility of the technique.
In summary, immunoassays are in general specific, sensitive and rapid
methods which enable the detection and quantification even of trace amounts
of allergens in raw materials as well as in intermediate and finished food
products, the success of which always depends on how the antibodies were
raised. However, routine application is still limited to the detection of only
some of the allergenic residues of concern. Only a few test kits have been
validated in third-party ring trials (AOAC, 2003). Moreover, standardised
tests according to CEN (European Committee for Standardisation) are
completely lacking. There is still a great need for cost-effective screening
methods which can rapidly detect ppm amounts of allergenic residues.
There are three main reasons for using lateral flow testing devices:
• in order to test outside the laboratory (food processing line, rinse waters,
home testing);
• when results have to be reported quickly;
• if a qualitative result is desired, LFDs are preferred because they are
packaged individually and so there is no need to use an entire strip of
micro-wells for just a few samples.
However, in some cases LFDs are not suitable, for example, if quantitative
results are required or if an extract is too viscous to be processed as is the
case with cereals or food gums. Here ELISA remains the method of choice.
In the future, these single-step tests will be used more and more in allergen
control programs, such as in validating sanitation procedures (line cleaning
in food production), checking for absence of contaminants in raw materials
and in-process products and, in the clinical arena, checking for presence of
diseases (doctor’s office/airports/hospitals).
10.7 References
AOAC Research Institute (2003) ‘Performance tested methods’, Inside Laboratory
Management, Sept/Oct, 19–22.
Mills C E N, Potts, A, Plumb, G W, Lambert, N and Morgan, R A (1997) ‘Development
of a rapid dipstick immunoassay for the detection of peanut contamination of food’,
Food Agric Immunol, 9, 37–50.
Part III
Detection methods for particular

allergens
11
Methods for detecting peanuts in food

S. Hefle, University of Nebraska, USA
11.1 Introduction: peanut allergy

Peanuts are the most allergenic food known and, in sensitive individuals, can
cause adverse reactions ranging from mild urticaria to life-threatening
anaphylaxis (Yunginger et al., 1988; Bock and Atkins, 1990; Sampson et al.,
1992). The estimated prevalence of food allergy in the population is estimated
to be 4% (Sampson, 2004), while the number of people allergic to peanuts
and tree nuts has been estimated at 0.6% in the US (Sicherer et al., 1999) and
0.48% in the UK (Emmett et al., 1999). A recent report indicated that peanut
allergy doubled in the years between 1999 and 2003 in children in the US
(Sicherer et al., 2003). While the threshold level necessary to cause a reaction
is not known with certainty, minute amounts may trigger anaphylaxis
(Hourihane et al., 1997; Wensing et al., 2002); most studies to date have
shown that low mg amounts (1–3 mg) of peanut protein are sufficient to
cause objective reactions in peanut-allergic subjects (Taylor et al., 2002;
Morisset et al., 2003). Therefore, trace amounts of undeclared peanut present
in food can be hazardous to peanut-allergic individuals if consumed. Persons
suffering from food allergy carefully avoid consuming their allergenic food,
but inadvertent ingestion still occurs. Bock et al. (2001) found that peanut
was responsible for more than half of 32 food anaphylaxis fatalities they
investigated, and it is recognized as the most common cause of food-induced
anaphylactic death (Ewan 1996; Burks et al., 1999).
11.2 Allergenic peanut proteins

Peanut seeds contain an average of 29% protein (Freeman et al., 1954).
Approximately 20% of the total protein can be attributed to the major allergen
Ara h 1 (vicilin-like protein, molecular weight 63–64 kD), and about 10% is
made up of another major allergen Ara h 2 (conglutin-like protein, molecular
weight about 17 kD) (Burks et al., 1998; Koppelman et al., 2001).
Ara h 2 is emerging as the most important allergen in terms of a link to
severity of reactions and number of peanut-allergic individuals that possess
IgE for it (Clarke et al., 1998; Koppelman et al., 2004), but other allergens
in peanut include Ara h 3 and 4 (glycinin proteins) (Bannon et al., 2000;
Koppelman et al., 2003; Scurlock and Burks, 2004), Ara h 5 (profilin), 6 and
7 (conglutin-like proteins; Scurlock and Burks, 2004), and Ara h 8 (Mittag
et al., 2004). There are several excellent overviews on peanut allergens that
the reader can consult for a full review (Burks et al., 1998; Sampson, 2004;
also Chapter 2).
The major peanut allergens have been found to be resistant to food processing
techniques and stable to digestion (Burks et al., 1998). Koppelman et al.
(1999) showed that roasting had no effect on allergenicity of Ara h 1, while
Malecki et al. (2000) found in another study that roasting enhanced allergenicity
of Ara h 1. A key factor in the success of immunodetection methods for
peanut residue is how well animal-derived antibodies recognize roasted/
heat-treated peanuts, since these are the most common forms consumed.
11.3 Peanut detection methods

Several types of methods have been used over the years to detect peanut
residue in other foods, including radioallergosorbent assay (RAST) inhibition,
IgE immunoblotting, rocket immunoelectrophoresis (RIE), enzyme-linked
immunosorbent assay (ELISA), and polymerase chain reaction (PCR) methods.
These methods are described more fully in other chapters of this book.
11.3.1 Rocket immunoelectrophoresis (RIE)

Rocket immunoelectrophoresis was first used by Barnett and Howden in
1984 (Barnett and Howden, 1984) to detect heat-treated peanut protein, and
this marked the first time an immunological method was used to detect
peanut protein. The authors at the time stated that the results of their work
‘may extend the usefulness of immunological techniques to include the
detection of peanut protein in processed foods.’ They used RIE to detect
peanut protein in a vegetarian meat loaf, but no levels were reported. Rocket
immunoelectrophoresis was also used by Malmheden-Yman et al. (1994) to
detect the undeclared presence of peanuts in a small number of food products.
The detection limit of the method was approximately 30 ppm, and no cross-
Methods for detecting peanuts in food 187
reactivity evaluations were reported. Holzhauser et al. (1998) a few years

later used a commercially available anti-raw peanut antibody in developing
a RIE method and were able to obtain a 10 ppm detection limit. The method
was optimized for the detection of peanut in chocolate, achieving recoveries
of 85–101% in this matrix. More than 20 types of foods were analyzed in
cross-reactivity studies, with no observed issues. Successful use of a commercial
antibody and also attaining a 10 ppm detection limit in a difficult matrix like
chocolate were significant achievements of this work. One disadvantage of
this technique was that the amount of antiserum used in the gel has to be
optimized; this is dependent upon whether or not the sample contains a large
or small amount of peanut, details of which are seldom known in advance.
The investigators also performed a limited survey of 11 commercial food
samples that did not declare the presence of peanut as an ingredient but four
of which contained appreciable amounts (0.0006–1.5% peanut). Notably,
these investigators used model chocolates with specific amounts of
peanuts manufactured into them under industrial conditions, an approach
that shows the true extraction efficiency of a method, as compared to simple
‘spiking’.
11.3.2 Radioallergosorbent assay (RAST) inhibition

The first report of RAST inhibition being used to detect peanut contamination
of a non-peanut-containing food was described by Yunginger et al. (1983).
The technique was used to document the presence of peanut butter in sunflower
butter (likely made on shared production lines) in investigating a reaction in
a peanut-allergic consumer. The investigators found peanut butter at levels
ranging from 0.3–3.3% (3000–33 000 ppm) in six of eight samples of sunflower
butter from one company. The protein used on the solid phase and for standards
was a crude extract of raw unsalted peanuts.
This method was expanded upon by Keating et al. (1990). In contrast to
the work discussed above, a roasted defatted peanut meal was used for the
solid phase and standards in a RAST inhibition method. However, test samples
required a long extraction and handling process, including lyophilization.
The investigators reported that their recoveries for peanut protein were 31–
94% and the detection limit was 87.5 ppm peanut; most experts agree at this
time that allergen detection methods should have an ability to detect down to
at least 10 ppm. These investigators also showed that peanut allergens can be
transferred to clean vegetable oil after roasting, an important point that has
not been studied by any other group to date. While one case report has linked
transfer of allergens in cooking oil to a fatality in a fish-allergic person
using RAST (Yunginger et al., 1988), this area will remain a concern
until more work is done to demonstrate the true risks from this type of
processing.
11.3.3 IgE-immunoblotting
An IgE-dot-immunoblotting method using peanut-allergic sera was described
by Schappi et al. (2001) to detect the presence of peanut proteins in foods.
The standards were made from roasted peanuts. The detection limit was
approximately 50 ppm, too high for practical allergen management. The use
of human sera and of urea in the extraction procedure limits the practical use
of this type of method in food-processing facilities. The authors found that
19 of 46 commercial food products had undeclared peanut residues, ranging
from 0.05 – >1% peanut (500 ppm – >10 000 ppm).
While IgE-based techniques such as RAST and enzyme allergosorbent
test (EAST inhibition) and IgE-immunoblotting are useful techniques that
can be used in many clinical and research settings (see Chapters 4 and 5), it
must be stressed that they require well-characterized peanut-allergic human
sera, a substance which needs special handling, and the use of this reagent is
precluded in food processing environments. Also, the specificity of IgE from
allergic individuals can vary considerably, making standardization difficult.
11.3.4 Enzyme-linked immunosorbent assays (ELISAs)

The first ELISA method for detection of peanut residues was described by
Hefle et al. (1994a). The method used monoclonal antibodies against selected
peanut proteins with molecular weights [in sodium dodeyl sulphate
polyacrylamide gel electrophoresis (SDS-PAGE)] of 14–44 kD (Hefle et al.,
1994b) as capture antibody and a polyclonal anti-peanut rabbit antibody as
detector. The rabbit antibody was raised against a crude roasted peanut extract.
The method could detect < 40 ppm peanut in most food matrices. Dry-
roasted peanut was spiked into vanilla ice cream at levels of 0.01–5.0%, then
tested in the ELISA, but also skin prick tested in seven peanut-allergic
subjects. In 6/7 patients, the skin test was more sensitive than the ELISA at
the lowest levels tested (0.01% = 100 ppm). The samples were also analyzed
by RAST inhibition using the sera from the seven peanut-allergic individuals.
The ELISA and the RAST inhibition results were proportional and correlated
well (r value = 0.95). This new method opened the door for other ELISA
formats with increased sensitivity (Poms et al., 2004). These authors suggested
that the use of amplification systems, improved antibodies, and assessment
of food matrix interference problems could result in increased sensitivity,
and this transpired in the form of many ELISA methods for peanut being
developed and published since the mid-1990s.
In 1996, both Yeung and Collins (Yeung and Collins 1996) and Koppelman
et al. (1996) published ELISA methods for detection of peanut proteins in
foods. Yeung and Collins developed a competitive format ELISA using a
rabbit polyclonal antibody raised against a mixture of defatted, roasted, raw,
denatured, and unfolded raw peanut. Their ‘denatured’ raw immunogen was
prepared by treatment with sodium dodecyl sulfate and dithiothreitol at 4 °C
for 60 minutes. In the described ELISA, roasted peanut was used to coat the
plate, and the authors indicated that they could store coated plates for
24 months. Yeung and Collins also spiked food samples at 2.5, 5, 10, and
20 ppm; the spiking resulted in observed recoveries of 68–90%. They reported
no cross-reactivity but tested samples for this in very low protein amounts
(< 40 ug/mL). They found that their rabbit anti-roasted peanut antibody
produced the most effective ELISA. They analyzed 15 non-peanut chocolate
bars and three types of sesame crackers and found that none of them had
undeclared peanut residues. This ELISA method reportedly had a detection
limit of 400 ppb.
Koppelman et al. (1996) described a sandwich-type ELISA using rabbit
polyclonal antibodies against partially purified Ara h 1 derived from raw
peanuts. This sandwich ELISA utilized the same antibody for detector and
capture; part of the antibody was labeled with horseradish peroxidase to be
the detector antibody. In studies, it was observed that fried peanuts were
somewhat less reactive in the ELISA, but the method did not show cross-
reactivity with the majority of legumes and other foods tested. The highest
concentration of almond and cashew tested (500 ug/mL) showed a slightly
positive signal, but a ten-fold dilution removed it. Recoveries in various
matrices ranged from 35–75%. The investigators also found that of 25 non-
peanut foods surveyed, none had detectable peanut content. The assay was
capable of measuring raw or roasted peanut, and the reported detection limit
was 0.1 ppm.
Much later, Pomes et al. (2003) published another sandwich ELISA based
on monoclonal antibody against Ara h 1. In this study, Ara h 1 could not
always be recovered in spiked chocolate samples nor always detected in the
presence of chocolate. However, it was found that in general, Ara h 1 content
correlated with a commercial ELISA method (Veratox® for Peanut – Neogen
Corporation, Lansing, MI), but the results were not compared to any other
commercial ELISA kits. The investigators ‘spiked’ ground chocolate with
ground peanut flour, but without using industrially manufactured standards,
the true extraction efficiency of the method cannot be truly ascertained. The
best extraction of Ara h 1 from chocolate was found to be with 5% non-fat
dry milk in phosphate-buffered saline for 2.5 hours at room temperature, an
approach that is too long for the rapid testing needs of the food industry, but
may be applicable elsewhere.
Keck-Gassenmeier et al. (1999) used a commercial sandwich-type ELISA
kit for peanut (Cortecs, now Tepnel Biosystems, Deeside, UK) and found
that adding 12.5% fish gelatin to the extraction buffer allowed recoveries in
chocolate of 60–90% and other food products of 48–88%; foods other than
chocolate in most cases did not require the use of the fish gelatin for improved
recoveries. A detection limit of 2 ppm could be obtained in chocolate. The
commercial test was qualitative only, but the investigators could make it
semi-quantitative by making their own standards; dark chocolate was melted
and then spiked with peanut butter. Cross-reactions were only assessed for
almonds, hazelnut, and milk and dark chocolate. Raw peanuts were found to
have a higher response in the assay. The detection limit of the modified assay
for detection of peanut proteins in dark chocolate was better than that claimed
by the manufacturer for the commercial kit.
Newsome and Abbott (1999) coupled an immunoaffinity column to capture
peanut proteins from foods, with a subsequent ELISA. The antiserum used
was raised against roasted peanuts. The antiserum was coupled to a commercial
gel preparation and they then did an ELISA according to the protocol of
Yeung and Collins (1996). The authors said that spiked chocolate gave
recoveries of 72–84%, and that the minimum detection limit was 0.1 ppm,
but they did not provide many details on the spiking.
Holzhauser and Vieths (1999) also developed an indirect ELISA method
using commercial anti-raw peanut rabbit antisera (absorbed to remove possible
soy reactivity) and ground roasted peanuts as a standard. This method was
reported to have a detection limit of 2 ppm and showed no cross-reaction
concerns with the food ingredients tested. In five of 17 commercial food
products not declaring peanut, levels of peanut ranging from 2–18 ppm were
found.
Later, the same research group (Stephan and Vieths, 2004) reported both
a sandwich ELISA and a PCR method for detection of peanut residues. Both
sheep and rabbit anti-peanut antibodies were used in the development of the
ELISA and, while the authors did indicate that the antibodies were raised
against peanut extract, it was not reported whether the peanuts used were
raw or roasted. The methods were developed primarily for confection products,
and so cross-reactivity with some types of food ingredients, such as spices,
was not done. Industrially-manufactured standards of 10, 40, and 200 ppm
peanut were made in semi-sweet and whole milk chocolate. For the sandwich
ELISA, recoveries from whole milk chocolate were 80.6–141.9%, and for
semi-sweet, 64.3–110.9%. Fish gelatin was added to the extraction buffer as
recommended by Keck-Gassenmeier et al. (1999); however, interference
still was observed with chocolate even when this was employed. In a survey
of 33 products, two were found to have undeclared peanut residue, in the
range of 6.3–74 ppm.
A cloth-based immunoassay was described by Blais and Phillippe (2000).
The method utilized polyester cloth as the solid phase and the capture and
detector (peroxidase-labeled) antibody was chicken anti-peanut IgY. Peanut
immunogen prepared according to the Yeung and Collins (1996) method was
used to immunize chickens, followed by a fairly lengthy purification of IgY
from egg yolks. The authors felt that the cloth solid phase provided the
advantages of conferring a high surface area for rapid immunoreactions and
ease of washing between steps. Food samples connected with consumer
illnesses were analyzed using the cloth-ELISA; of 11 samples not declaring
peanut, nine contained peanut residues and were positive in the cloth-ELISA.
Levels were reported to be 1.2–116 ppm (determined using the Yeung and
Collins (1996) quantitative ELISA method). Cross-reactivity data was not
shown; chocolate seemed to give significant interference in the method. The
investigators found that skim milk powder alleviated the problem, but some
samples were positive when analyzed in a commercial peanut kit but not
detected using the Yeung and Collins (1996) method – however, results from
the cloth-ELISA and the commercial peanut kit did match. The authors
reported that the detection limit of the cloth-ELISA was less than 1 ppm
peanut protein for some matrices.
As an outgrowth of Koppelman et al.’s (1996) work (discussed above),
R-Biopharm AG (Darmstadt, Germany) has developed a faster version of
their peanut ELISA assay (RIDASCREEN®), which can be completed in
about 30 minutes (Immer et al., 2004). Other commercial kit companies
have also reduced the assay time of their peanut ELISA peanut kits to a
similar level.
11.3.5 Dipsticks/lateral flow methods

The first peanut dipstick method was described by Mills et al. (1997). Dipstick
or lateral flow methods (described in Chapter 10) are only semi-quantitative
or qualitative in nature. In this study, polyclonal rabbit antisera were raised
to conarachin, the 7S globulin of peanut, which had been isolated from
defatted raw peanuts. The same serum was used as both capture and detector
antibody (detector was labeled with biotin). No cross-reactions among several
foods and food ingredients were noted but, since the antibody was made
against a purified protein, no cross-reactions were necessarily expected. The
detection limit was reported as being 100 ppm in marzipan and 1000 ppm in
chocolate; roasted peanut could be detected down to 1000 ppm in marzipan
and chocolate. The format used, that of using cryotubes coated with antisera,
limited matrix interference issues with confection products. Only one food
product of six tested gave a positive result in the study; roasting did not
appear to affect the assay a great deal. The detection limit of this assay
(0.01% of raw peanut in marzipan) is much too high for allergen concerns.
Stephan et al. (2002) developed a dipstick method for the detection of
trace amounts of peanut and hazelnut in foods. The reported detection limit
was 1 ppm protein. These tests took 3–4 hours, which is long for dipstick
methods (the current commercial dipsticks on the market are performed
within 10 minutes). Various food samples were tested using the dipstick and
results were compared with a previously described validated quantitative
ELISA for peanut (Holzhauser and Vieths, 1999). Dipsticks are a useful tool
for rapid and convenient testing of food and environmental samples. The
investigators reported some cross-reactivity with some food ingredients with
this method (walnut in particular). Also employed was the use of 10% fish
gelatin in the extraction buffer, which improved the extraction of the peanut
from chocolate-containing foods. Of 28 foods examined in this study that
did not declare peanut on their labels, eight tested positive for residues.
There are two commercial lateral flow devices on the market for peanut,
made by Neogen (Reveal®) and Tepnel Biosystems (Biokits® Rapid Peanut
Test), and other diagnostic kit companies are planning to produce kits in this
type of format also.
Two multiplex methods have been described, based on earlier work by the
authors. Ben Rejeb et al. (2002) developed a multi-analysis semi-quantitative
ELISA for the detection of peanut, almond, cashew, Brazil nut, and hazelnut.
The reported detection limit was 2 ppm for all allergens. Blais et al. (2003)
continued with their use of IgY and cloth in a multiple immunoassay method
(as described above) for peanut, Brazil nut, and hazelnut. The claimed detection
limit for peanut for this multiplex method was 0.1 ppm. The multiplex approach
offers the advantages of simple and rapid testing for various allergens
simultaneously, allowing for an initial preliminary screening to identify which
allergenic residues are present, to be followed up with a confirmatory test,
such as a specific ELISA method. The practicality of and need for this type
of approach for food manufacturers have yet to be evaluated fully.
In the only study to date addressing simply chocolate contamination,
Vadas and Perelman (2001) performed a limited survey of retail chocolate
bars using the Neogen Veratox® for Peanut test kit; results showed that eight
of 46 samples contained undeclared peanut residues, ranging from 7.9–69.6
ppm peanut.
11.3.6 Comparisons/validation of commercial peanut ELISA kits

Hurst et al. (2002) compared commercial test kits for peanut on samples of
dark and milk chocolate. Dark and milk chocolate samples were made in the
levels of 0, 10, 40, and 200 ppm peanut, although details of how this was
done and how homogeneity was determined were not disclosed. Four
commercial test kits for the detection of peanut – Neogen Veratox®, Pro-Lab
Prolisa Peanut Pak® (Pro-Lab Diagnostics, Richmond Hill, Ontario), Tepnel
Biokit®, and R-Biopharm RIDASCREEN®) were evaluated using the
chocolate samples. In general, the kits varied in their ability to quantitate the
levels of spiked peanut. However, with no details given on the nature of the
peanut material used to make the spiked chocolate, etc., it is difficult to truly
compare one kit to another. In addition, refinements have been made to most
of these kits since this work was published.
In one study, Koch et al. (2003) compared spiked samples using some
commercial peanut ELISA kits and IgE-immunoblotting using peanut-allergic
sera. Samples of peanut-free cookie matrix were spiked with various levels
of ground raw or roasted peanuts. The protein extractions for blotting were
done with urea, a powerful disrupting agent. However, when samples were
analyzed in the commercial ELISAs, manufacturer instructions were followed,
and none of these include urea in their extraction buffer. Dot blotting was
used to estimate the amount of peanut in the samples, whereas IgE-
immunoblotting was used to analyze the extracts further and identify specific
peanut allergens. The levels of peanut spiked ranged from 500–2500 ppm,
well above allergen concern levels. In addition, extracts of spiked samples
were diluted in extraction buffers to 10 ppm in the comparison of the three

kits; it must be pointed out that this is not how the kits are designed to be
used. The authors made some assumptions that test kit companies used raw
peanuts as immunogen to explain their results showing that raw peanuts
were detected at higher levels, but in fact some kit manufacturers use roasted
peanuts as immunogen for producing their antibodies. The authors also
concluded that roasted peanuts were less detectable than raw peanuts in
commercial ELISA kits, but only used one of the three commercial kits to
test this hypothesis. A direct comparison of IgE-blotting using urea extraction
and use of commercial test kits (with their non-urea extraction methods)
casts some doubt on the validity of some of the conclusions made in this
study.
The Association of Official Analytical Chemists (AOAC) – Research Institute
and the US Food and Drug Administration has undertaken a performance
evaluation of three (Neogen Veratox®, Tepnel Biokits®, and r-Biopharm
RIDASCREEN®) commercial peanut quantitative ELISA kits; matrices of
breakfast cereal, cookies, ice cream, and milk chocolate were ‘spiked’ at
levels of 0 and 5 ppm peanut all three kits have passed the performance
evaluation (Park et al., 2005).
In another effort, the analytical quantitation performance of five different
peanut ELISA kits was validated in an international collaborative study by
the Joint Research Centre of the European Commission. The study involved
industrial dark and milk chocolate samples containing peanut in the low ppm
range (Poms et al., 2003, 2005).
Using dark chocolate ‘manufactured’ to contain known amounts of peanut
in preliminary work for a more in-depth study, we evaluated five commercial
peanut detection kits. Individual reference dark chocolate standards were
made in an industrial setting (Barry Callebaut, St Hyacinthe, Quebec, Canada).
Dark chocolate standards containing 0, 0.5, 1, 2, 10, 20, 50, and 100 ppm
peanut were prepared. The chocolate was prepared using chocolate liquor,
butter oil, cocoa butter, soy lecithin, sugar, and vanilla. Tempered chocolate
standards were thoroughly mixed (Hobart Corp., Troy, OH) before molding.
Kits were purchased from four manufacturers: R-Biopharm (RIDASCREEN®
Peanut); Pro-Lab Diagnostics (Prolisa Peanut Pak®), Neogen Corporation
(Alert® and Veratox® for Peanut) and Tepnel BioSystems (Tepnel BioKits®
Peanut Assay), and run on the peanut-in-chocolate dark chocolate model
food (three trials, triplicates for each trial) strictly according to manufacturer
instructions. In general, four of the five kits performed well in their ability to
quantitate the levels of peanut in the dark chocolate (Table 11.1), even though
their target analytes are different (some are for total peanut, some for peanut
protein). However, one commercial kit did not perform well – this kit was
purchased twice, six months apart, and the result was the same both times.
Many more studies like this one using ‘manufactured’, real-life model foods
is needed, not only for peanut, but also other allergenic foods (see Section 11.6).
Table 11.1 Comparison of four commercial quantitative peanut kits using manufactured
peanut in dark chocolate samples
Peanut in Neogen R-Biopharm Tepnel Pro-Lab

dark Veratox® RIDASCREEN® Biokits Prolisa
chocolate for Peanut Peanut Peanut Peanut Pak®
0 ppm < 2.5 ppm < 2.5 ppm 0.33 ± 0.1 ppm 0 ppm
0.5 ppm < 2.5 < 2.5 0.93 ± 0.25 0.16 ± 0.02
1 ppm < 2.5 < 2.5 1.57 ± 0.29 0.11 ± 0.11
2 ppm < 2.5 < 2.5 2.82 ± 0.38 0± 0
10 ppm 12.87 ± 5.6 6.67 ± 0.31 12.1 ± 2.8 0.16 ± 0.1
20 ppm 23 ± 6.1 12.0 ± 0.67 25.1 ± 7.1 0.27 ± 0.1
50 ppm 52.5 ± 16.7 66.4 ± 15.0 63.5 ± 9.7 1.62 ± 1.4
100 ppm 121.0 ± 27.6 117.4 ± 34.2 127.1 ± 28.5 3.65 ± 3.9
11.3.7 Non-immunological methods for detection of peanut

A recent publication by Shefcheck and Musser (2004) describes the detection/
confirmation of Ara h 1 in a food matrix using a liquid chromatography/
tandem mass spectrometry (LC/MS/MS) method. This technique provides
confirmation of allergenic proteins in food systems as a confirmatory test for
immunoassays, which has been identified by some regulatory agencies as a
significant need. The method can be adapted for use with any well-characterized
protein in almost any matrix, and does not require specific antibodies. It
does, however, require specialized equipment and training, and it is likely to
be best suited to regulatory agencies or academic labs rather than food
manufacturers. It is felt that these types of methods could possibly have
detection limits approaching 1 ppm protein.
In addition to protein/immunochemical methods, some PCR (discussed in
Chapter 7) methods for peanut have been described. Two types of formats
are currently commercially available; real-time PCR and DNA-ELISA.
Reported detection limits are below 10 ppm. Hird et al. (2003) published a
real-time PCR method for detection of peanut residues. The amplification
consisted of a part of the DNA encoding Ara h 2, a major peanut allergen.
Several commercial DNA extraction methods were evaluated. The investigators
reported that the method could detect 2 ppm lightly roasted peanut flour
spiked into cookies. Raw peanuts were spiked into pastry, sausage, chocolate,
or cake at a 10% level (much higher than is of concern for allergen residue
detection) and, to attempt to mimic cross-contact, a single piece of peanut
was tumbled with these negative food matrices for 30 minutes, then analyzed.
The spiked peanut was detected in the foods as was the presence of peanut
after being tumbled with the foods. Stephan and Vieths (2004), as noted
above, also developed a PCR method for the detection of peanut. A commercial
DNA extraction method was used and the primer was also specific for a
coding region on Ara h 2. This real-time PCR method was able to detect
down to 10 ppm peanut in semi-sweet chocolate and whole milk chocolate.
In analyzing selected products for undeclared peanut, the PCR method detected
one more positive sample than their sandwich ELISA did. The authors
concluded that PCR and ELISA gave similar results for detection of peanut
residues in foods, but pointed out that in some cases, such as pickled foods,
oils, and canned foods, the results will not be comparable due to DNA
degradation during processing. There are two commercial PCR methods
currently on the market for the detection of peanut residues (Table 11.2).
Indirect approaches such as PCR-based methods detect a marker for the
allergenic fraction rather than allergen itself. For food ingredients that are
only slightly processed, such as peanuts as topping on cookies, these indirect
approaches can be useful. However, when food ingredients undergo
fractionation steps such as concentration of the protein fraction as is done for
defatted peanut meal, indirect approaches run the risk that the outcome of
the analytical results could be underestimated. On the contrary, for peanut
PCR, results may be overestimated for food products from which the protein
fraction is removed, such as refined, bleached, deodorized peanut oils, which
are essentially devoid of protein, but can have DNA potentially present.
11.4 Appropriate detection limits for peanut methods

As discussed in Chapter 1 of this book, existing methods appear to be in the
ideal range to detect potentially hazardous residues of undeclared allergens
in foods. Detection limits based on clinical relevance are important; good
manufacturing practices and regulatory limits imposed by governmental
agencies should be based on knowledge of clinical elicitation levels. Poms
et al. (2003) indicates that detection limits for peanut allergens probably
need to be 1 ppm or lower, but support for this level is not borne out by
double-blind clinical oral threshold studies. Morisset et al. (2003), based on
their threshold study data, recommends a detection limit for peanut of 24
ppm for foods and 5 ppm for oil. The action level of 1000 ppm put forth by
Schappi et al. (2001) is certainly not appropriate, and could leave some
factions of the peanut-allergic population at great risk. At this time, it appears
that methods with detection limits of 1–5 ppm are best for peanut, and there
is currently no need to take detection abilities below this level. The currently
commercially available methods have detection limits in this range, or
somewhat higher (Table 11.2).
11.5 Future trends

With so many peanut detection systems already developed, one particular
need in this area is the production of peanut reference materials for multi-
laboratory validation trials. This work is of particular importance due to the
plethora of types of peanuts and the wide variety of processing methods used
Table 11.2 Commercially available peanut residue detection kits
Company Method Name of kit Type LOD1 LLOQ2 Time3 Contact information
4
ELISA Systems ELISA Peanut Protein Residue Quantitative 0.5 ppm – 30 min www.elisasystems.net
ELISA Systems ELISA Peanut Residue Mass Quantitative 1 ppm 2.5 ppm 60 min www.elisasystems.net
Neogen ELISA Veratox® Quantitative5 1 ppm 2.5 ppm 30 min www.neogen.com
Neogen ELISA Alert® Qualitative 5 ppm – 30 min www.neogen.com
Neogen LFD6 Reveal® Qualitative 5 ppm – 10 min www.neogen.com
Pro-Lab ELISA Quantitative www.pro-lab.com
Diagnostics
R-Biopharm ELISA RIDASCREEN® Quantitative < 2.5 ppm 3.3 ppm 1.5 hr www.r-biopharm.com
Peanut
R-Biopharm ELISA RIDASCREEN® Quantitative 1.5 ppm 2.5 ppm 30 min www.r-biopharm.com
FAST Peanut
R-Biopharm PCR7 SureFood® Peanut Quantitative < 10 ppm – 2–3 hr www.r-biopharm.com
PCR-ELISA
R-Biopharm PCR SureFood® Peanut Qualitative 10–50 ppm – 30 min www.r-biopharm.com
Real-time PCR
Tepnel ELISA BioKits Quantitative < 0.1 ppm 1 ppm 1.25 hr www.tepnel.com
BioSystems Peanut Assay
Tepnel LFD BioKits RAPID Qualitative Low ppm < 50 ppm 6 min www.tepnel.com
BioSystems Peanut Test
Tepnel PCR BioKits Peanut Qualitative < 10 ppm – < 2 hr www.tepnel.com
BioSystems PCR Mastermix Pod
1
LOD = limit of detection; LOD is defined in different ways (whole food or protein); check with manufacturers for specifics
2
LLOQ = lower limit of quantitation
3
Time = excludes sample and reagent preparation times
4
ELISA = enzyme-linked immunosorbent assay
5
Any quantitative ELISA can be run qualitatively or semi-quantitatively
6
LFD = lateral flow device
7
PCR = polymerase chain reaction
Notes: Claims are as specified by manufacturers; all test performance is sample-dependent
Units of measurement are different for different kits; please check with manufacturers for specifics
world-wide. Another important research need is for a comparison of different

methods with regard to their ability to detect peanuts in a variety of matrices
using industrially-manufactured standards. Simple ‘spiking’, i.e. putting a
peanut extract into a food extract, or peanut flour into a dry food powder, is
no longer an appropriate means of assessing performance other than for the
purpose of initially probing possible matrix interference. Making model
foods according to pilot plant or true industrial conditions is not necessarily
an easy task, but some research groups (as discussed in Chapter 17) are
doing work in this area, and these materials will be key to assessing peanut
detection methods now and in the future. A listing of the currently available
commercial peanut detection methods is given in Table 11.2.
11.6 References
Bannon, G A, Besler, M, Heflé, S L, Hourihane, J O’B and Sicherer, S H (2000) ‘Allergen
data collection: peanut (Arachis hypogea)’, Internet Symposium on Food Allergens, 2,
87–122.
Barnett, D and Howden, M E H (1984) ‘A rocket immunoelectrophoretic method for
the detection of heat-treated peanut protein’, Food Technology in Australia, 36, 510–
511.
Ben Rejeb, S, Abbott, M, Davies, D, Cleroux, C and Streng, D (2002) ‘Multi-screening
immunoassay for the detection of nut proteins in chocolate’, Proc. TNO International
Food Allergy Forum, Noordwijkerhout, The Netherlands, April 15–16, 58.
Blais, B W and Phillippe, L M (2000) ‘A cloth-based enzyme immunoassay for detection
of peanut proteins in foods’, Food Agric Immunol, 12, 243–248.
Blais, B W, Gaudreault, M and Phillippe, L M (2003) ‘Multiplex enzyme immunoassay
system for the simultaneous detection of multiple allergens in foods’, Food Control
14, 43–47.
Bock, S A and Atkins, F M (1990) ‘Patterns of food hypersensitivity during sixteen years
of double-blind, placebo-controlled food challenges’, J Pediatrics, 117, 561–567.
Bock, S A, Munoz-Furlong, A and Sampson, H A (2001) ‘Fatalities due to anaphylactic
Burks, A W, Sampson, H A and Bannon, G A (1998) ‘Peanut allergens’, Allergy, 53, 725–
730.
Burks, A W, Bannon, G A, Sicherer, S and Sampson, H A (1999) ‘Peanut induced
anaphylactic reactions’, Int Arch Allergy Immunol, 119, 165–172.
Clarke, M C A, Kilburn, S A, Hourihane, J O’B, Dean, K R, Warner, J O and Dean, T P
(1998) ‘Serological characteristics of peanut allergy’, Clin Exp Allergy, 28, 1251–
1257.
Emmett, S E, Angus, F J, Fry, J S and Lee, P N (1999) ‘Perceived prevalence of peanut
allergy in Great Britain and its association with other atopic conditions and with
peanut allergy in other household members’, Allergy, 54, 380–385.
Ewan, P W (1996) ‘Clinical study of peanut and nut allergy in 62 consecutive patients:
new features and associations’, Br Med J, 312, 1074–1078.
Freeman, A K, Morris, N J and Willich, R K (1954) ‘Peanut butter’, U.S. Dept. of
Agriculture Bulletin AIC–370.
Hefle, S L, Chu, F S, Yunginger, J W and Bush, R K (1994a) ‘A sandwich-type enzyme-
linked immunosorbent assay (ELISA) for quantitation of selected peanut proteins in
foods’, J Food Prot, 57, 419–423.
Hefle, S L, Folgert, J P, Chu, F S and Bush, R K (1994b) ‘Monoclonal antibodies against

selected peanut allergens: production and use as affinity agents’, Food Agric Immunol,
6, 197–208.
Hird, H, Lloyd, J, Goodier, R, Brown, J and Reece, P (2003) ‘Detection of peanut using
real-time polymerase chain reaction’, Eur Food Res Technol, 217, 265–268.
Holzhauser, T and Vieths, S (1999) ‘Indirect competitive ELISA for determination of
traces of peanut (Arachis hypogaea L.) protein in complex food matrices’, J Agric
Food Chem, 47, 603–611.
Holzhauser, T, Dehne, L I, Hoffmann, A, Haustein, D and Vieths, S (1998) ‘Rocket
immunoelectrophoresis (RIE) for determination of potentially allergenic peanut proteins
in processed foods as a simple means for quality assurance and food safety’, Z Lebens
Unters Forsch A, 206, 1–8.
Hourihane, J O, Kilburn, S A, Nordlee, J A, Hefle, S L, Taylor, S L and Warner, J O
(1997) ‘An evaluation of the sensitivity of subjects with peanut allergy to very low
doses of peanut protein: a randomized, double-blind, placebo-controlled food challenge
study’, J Allergy Clin Immunol, 100, 596–600.
Hurst, W J, Krout, E R and Burks, W R (2002) ‘A comparison of commercially available
peanut ELISA test kits on the analysis of samples of dark and milk chocolate’, J
Immunoassay Immunochem’, 23, 451–459.
Immer, U, Reck, B, Lindeke, S and Koppelman, S (2004), RIDASCREEN® FAST PEANUT,
a rapid and safe tool to determine peanut contamination in food’, Intl J Food Sci
Technol, 39, 869–871.
Keating, M U, Jones, R T, Worley, N J, Shively, C A, Yunginger, J W (1990) ‘Immunoassay
of peanut allergens in food-processing materials and finished foods’, J Allergy Clin
Immunol, 86, 41–44.
Keck-Gassenmeier, B, Benet, S, Rosa, C and Hischenhuber, C (1999) ‘Determination of
peanut traces in food by a commercially-available ELISA test’, Food Agric Immunol,
11, 243–250.
Koch, P, Schappi, G F, Poms, R E, Wüthrich, B, Anklam, E and Battaglia, R (2003)
Koppelman, S J, Bleeker-Marcelis, H, van Duijn, G and Hessing, M (1996) ‘Detecting
peanut allergens: the development of an immunochemical assay for peanut proteins’,
The World of Ingredients, 7195, 35–38.
Koppelman, S J, Bruijnzeel-Koomen, C A, Hessing, M and de Jongh, H H (1999)’, Heat-
induced conformational changes of Ara h 1, a major peanut allergen, do not affect its
allergenic properties’, J Biol Chem, 274, 4770–4777.
Koppelman, S J, Vlooswijk, R A A, Knippels, L M J, Hessing, M, Knol, E F, van Reijsen,
F C and Bruijnzeel-Koomen, C A F M (2001) ‘Quantification of major peanut allergens
Ara h 1 and Ara h 2 in the peanut varieties Runner, Spanish, Virginia, and Valencia,
bred in different part of the world’, Allergy, 56, 132–137.
Koppelman, S J, Knol, E F, Vlooswijk, R A A, Wensing, M, Knulst, A C, Hefle, S,
Gruppen, H and Piersma, S (2003) ‘Peanut allergen Ara h 3:isolation from peanuts
and biochemical characterization’, Allergy, 58, 1144–1151.
Koppelman, S J, Wensing, M, Ertmann, M, Knulst, A C and Knol, E F (2004) ‘Relevance
of Ara h 1, Ara h 2 and Ara h 3 in peanut-allergic patients, as determined by
immunoglobulin E Western blotting, basophil-histamine release and intracutaneous
testing: Ara h 2 is the most important peanut allergen’, Clin Exp Allergy, 34, 583–590.
Malecki, S J, Chung, S Y, Champagne, E R and Raufman, J P (2000) ‘The effects of
roasting on the allergenic properties of peanut proteins’, J Allergy Clin Immunol, 106,
763–768.
Malmheden Yman, I, Eriksson, A, Everitt, G, Yman, L and Karlsson, T (1994) ‘Analysis
of food proteins for verification of identification of contamination or mislabeling’,
Mills, C E N, Potts, A, Plumb, G W, Lambert, N and Morgan, R A (1997) ‘Development

of a rapid dipstick immunoassay for the detection of peanut contamination of food’,
Mittag, D, Akkerdaas, J, Ballmer-Weber, B K, Vogel, L, Wensing, M, Becker, W-M,
Koppelman, S, Helbling, A, Hefle, S L, van Ree, R and Vieths, S (2004) ‘Ara h 8; a
Bet v 1-homologous allergen from peanut, is a major allergen in patients with combined
birch pollen and peanut allergy’, J Allergy Clin Immunol, 114, 1410–1417.
Morisset, M, Moneret-Vautrin, D A, Kanny, G, Guenard, L, Beaudouin, E, Flabbee, J and
Hatahet, R (2003) ‘Thresholds of clinical reactivity to milk, peanut, and sesame in
immunoglobulin E-dependent allergies: evaluation by double-blind or single-blind
placebo-controlled oral challenges’, Clin Exp Allergy, 33, 1046–1051.
Newsome, W H and Abbott, M (1999) ‘An immunoaffinity column for the determination
of peanut protein in chocolate’, J AOAC Int, 82, 666–668.
Park, D L, Coates, S, Brewer, V A, Garber, E A, Abouzied, M, Johnson, K, Ritter, B and
McKenzie, D (2005) ‘Performance Tested Method multiple laboratory study of ELISA-
based assays for the detection of peanuts in food’, J AOAC Int, 88, 156–160.
Pomes, A, Helm, R M, Bannon, G A, Burks, A W, Tsay, A and Chapman, M D (2003)
Immunol, 111, 640–645.
Poms, R E, Anklam, E, Klein, C L and Ulberth, F (2003) ‘Peanut allergens in food
products’, G.I.T. Laboratory Journal, 3, 132–135.
Poms, R E, Klein, C L and Anklam, E (2004) ‘Methods for allergen analysis in food: a
review’, Food Addit Contam, 21, 1–31.
Poms, R E, Agazzi, M E, Bau, A, Brohee, M, Capelletti, C, Norgaard, J V and Anklam,
E (2005) ‘Inter-laboratory validation study of five commercial ELISA test kits for the
determination of peanut proteins in biscuits and dark chocolate’, Food Addit Contam,
22, 104–112.
Sampson, H A (2004) ‘Update on food allergy’, J Allergy Clin Immunol, 113, 805–819.
Sampson, H A, Mendelson, L and Rosen, J P (1992) ‘Fatal and near-fatal anaphylactic
reactions to food in children and adolescents’, N Engl J Med, 327, 380–384.
Schappi, G F, Konrad, V, Imhof, D, Etter, R and Wuthrich, B (2001) ‘Hidden peanut
allergens detected in various foods: findings and legal measures’, Allergy, 56, 1216–
1220.
Scurlock, A M and Burks, A W (2004) ‘Peanut allergenicity’, Ann Allergy Asthma Immunol,
93(Suppl. 3), S12–S18.
Shefcheck, K J and Musser, S M (2004) ‘Confirmation of the allergenic peanut protein,
Ara h 1, in a model food matrix using liquid chromatography/tandem mass spectrometry
(LC/MS/MS)’, J Agric Food Chem, 52, 2785–2790.
Sicherer, S H, Munoz-Furlong, A, Burks, A W and Sampson, H A (1999) ‘Prevalence of
peanut and tree nut allergy in the U.S. determined by a random digit dial telephone
Sicherer, S H, Munoz-Furlong, A and Sampson, H A (2003) ‘Prevalence of peanut and
tree nut allergy in the United States determined by means of a random digit dial
telephone survey: a 5-year follow-up study’, J Allergy Clin Immunol, 112, 1203–1207.
Stephan, O and Vieths, S (2004) ‘Development of a real-time PCR and a sandwich ELISA
for detection of potentially allergenic trace amounts of peanuts (Arachis hypogaea) in
processed foods’, J Agric Food Chem, 52, 3754–3760.
Stephan, O, Moller, N, Lehmann, S, Holzhauser, T and Vieths, S (2002) ‘Development
and validation of two dipstick type immunoassays for determination of trace amounts
of peanut and hazelnut in processed foods’, Eur Food Res Technol, 215, 431–436.
Taylor, S L, Hefle, S L, Bindslev-Jensen, C, Bock, S A, Burks, A W, Christie, L, Hill, D
J, Host, A, Hourihane, J O’B, Lack, G, Metcalfe, D D, Moneret-Vautrin, D A, Vadas,
P A, Rance, F, Skrypec, D J, Trautman, T A, Yman, I M and Zeiger, R S (2002)
‘Factors affecting the determination of threshold doses for allergenic foods: how
much is too much’? J Allergy Clin Immunol, 109, 24–30.
Vadas, P and Perelman, B (2001) ‘Presence of undeclared peanut proteins in chocolate

bars purchased in the Toronto area’, Canadian J Allergy Clin Immunol, 6, 116–117.
Wensing, M, Penninks, A H, Hefle, S L, Koppelman, S J, Bruijnzeel-Kooman, C A and
Knulst, A C (2002) ‘The distribution of individual threshold doses eliciting allergic
reactions in a population with peanut allergy’, J Allergy Clin Immunol, 110, 915–920.
Yeung, J M and Collins, P G (1996) ‘Enzyme immunoassay for the determination of
Yunginger, J W, Gauerke, M B, Jones, R T, Dahlberg, M E and Ackerman, S J (1983) ‘Use
of radioimmunoassay to determine the nature, quantity and source of allergenic
contamination of sunflower butter’, J Food Prot, 46, 625–628.
Yunginger, J W, Sweeney, K G, Sturner, W Q, Giannandrea, L A, Teigland, J D, Bray, M,
Benson, P A, York, J A, Biedrzycki, L, Squillace, D L and Helm R M (1988) ‘Fatal
food-induced anaphylaxis’, JAMA, 260, 1450–1452.
12
Detecting tree nuts and seeds in food

S. Koppelman, AllerTeQ and University Medical Centre Utrecht,
The Netherlands
12.1 Introduction
Food allergy is an increasing health problem, and tree nuts belong to the ‘big
eight’ food allergens related to 90% of the food allergies (FAO, 1995). Tree
nut allergens are stable towards food processing and digestion. Similar stable
allergens have been identified in several seeds, and both seeds and nuts are
often used by the food industry owing to their taste. Food product diversification
since the 1990s has led to a vast number of recipes with nuts and seeds. This
was further stimulated by the observation that nuts have the beneficial health
effect of lowering cholesterol levels (Hu and Stampfer, 1999). Products that
do not contain nuts or seeds may be manufactured in facilities that also
process nuts and seeds, potentially forming a risk of cross-contact. Traces of
these allergens present in assumed nut- or seed-free products have led to
several severe allergic reactions. This indicates that methods to detect nut
and seed allergens are important tools for the food industry to control their
manufacturing and cleaning procedures. This chapter deals with the prevalence
of nut and seed allergy, threshold levels of allergens, the types of allergens
found in nuts and seeds, the effect of food processing on allergenicity, and
with tools available for detecting nuts and seeds. An overview of individual
methods will be given as well as a list of commercially available methods
that are presently used by the food industry. Since hazelnuts have been used
for a long period of time, especially in Western Europe, a lot of effort was
put into hazelnut detection. Pros and cons of the hazelnut detection methods
will be discussed, and we will look ahead to how detection of nut and seed
allergens may help the food industry in the future.
12.2 Prevalence of nut and seed allergies

In 1996, Tariq et al. published a cohort study that included 1218 newborns
of the Isle of Wight in the UK on the prevalence of tree nut allergy (Tariq et
al., 1996). 1.2% of the cohort was IgE-positive for tree nuts as demonstrated
by skin prick test, and 0.16% reacted clinically. This is in accordance with
the study of Sicherer et al. who estimated the prevalence of tree nut allergy
at 0.5% using a random digit dial telephone survey that included over 12 000
individuals in the US (Sicherer et al., 1999). Nut-induced allergic reactions
vary from mild, local reactions like oral allergy syndrome, to severe, systemic
reactions, and a number of fatal and near-fatal reactions have been described
(Yunginger et al., 1988; Bock et al., 2001). The consumption of nuts differs
between the US and Europe, and this may have consequences for sensitization
patterns. In Europe, hazelnut allergy is common and often associated with
birch pollen allergy (Andersen and Lowenstein, 1978) whereas in the US
walnut leads, followed by cashew, almond, pecan, and pistachio (Sicherer
et al., 1999). Other nuts that are considered as important allergens are Brazil
nut and chestnut. Since Brazil nuts are rich in sulfur-containing amino acids,
one of its proteins was cloned into transgenic soy in order to improve the
nutritional value for animal feed. The fact that this transgenic soy was allergenic
for nut-allergic individuals (Nordlee et al., 1996) stopped the development
and marketing of this crop. A nut that is gaining increasing attention is the
pine nut and cases of pine nut-induced allergic reactions have been reported
(Beyer et al., 1998). Some other, more exotic nuts have been reported to be
allergenic, for example the Micronesian nut Nangai (Sten et al., 2002), although
these cases are rare.
Table 12.1 summarizes the tree nuts involved in allergic reactions. Also
included in Table 12.1 are seeds that allergenic. Because of similarities in
protein composition, stability, and area of application, allergenicity of tree
nuts and seeds are often discussed together. Peanuts are legumes rather than
nuts and therefore discussed elsewhere in this book. Sesame seed and sunflower
seeds have been reported as causing allergic reactions (Axelsson et al., 1994;
Kanny et al., 1996; Sporik and Hill, 1996; Asero et al., 1999) while reaction
Table 12.1 Allergenic nuts and seeds
Nuts Seeds
Almond Celery
Brazil nut Linseed
Cashew Mustard
Chestnut Poppy seed
Hazelnut Rape seed
Macadamia Sesame seed
Pecan Sunflower seed
Pine nut
Pistachio
Walnut
Detecting tree nuts and seeds in food 203
to poppy seed, rape seeds, and linseed are less frequently described (Axelsson
et al., 1994; Kolopp-Sarda et al., 1997; Lezaun et al., 1998; Frantzen et al.,
2000; Galloway, 2000). Reactions vary for seeds, as for tree nuts, from mild
to severe, with anaphylaxis often described. Interestingly, the majority of
studies on the allergenicity of seeds and reports of allergic reactions come
from Europe. The European Commission published data on the prevalence
of seed allergies in 1998, and seeds were involved 1–2% of food-induced
anaphylaxis. There are no studies reporting prevalence of seed allergy in a
cohort of randomly selected individuals as has been done for tree nuts.
12.3 Thresholds
A lot of cases of unintentional ingestion of tree nuts and seeds have been
described, and traces of nuts and seeds are able to induce allergic reactions.
In some cases, the amount of offending allergen has been estimated. However,
double-blind, placebo-controlled food challenge studies are required to judge
how much allergen is too much (Hefle and Taylor, 2002). For hazelnut,
several researchers have investigated threshold in this way. Ortolani et al.
(2000) described thresholds of 168 mg–1.8 g hazelnut protein (Ortolani et
al., 2000) and Wensing et al. (2002) found thresholds as low as 1 mg hazelnut
protein (Wensing et al., 2002), indicating a considerable difference between
patients. The published thresholds are in line with case reports concerning
allergic reaction to hazelnut traces; milligram amounts of hazelnut protein
have been found to be enough to induce allergic reactions (Malmheden
Yman et al., 1994; Wensing et al., 2001a). Wensing et al. (2001b) showed
that the threshold for roasted hazelnuts was higher compared than that for
raw hazelnuts (Wensing et al., 2001b). This was further investigated by
Hansen et al. (2003) who found that only five of 17 hazelnut-allergic patients
that reacted clinically to raw hazelnut reacted to roasted hazelnut. Differences
in threshold levels were not determined (Hansen et al., 2003). Table 12.2
gives a brief overview of threshold data for some nuts and seeds obtained
using double blind, placebo-controlled food challenges. Note that for most
tree nuts and seeds, thresholds have not been established. Considering
thresholds in the mg range, and an average food intake of 50 g, as in the case
of a chocolate bar, a relevant limit for food allergen detection is in the order
of 10 ppm (Taylor and Nordlee, 1996). It should be noted, however, that
there may be food-allergic individuals who react to smaller amounts of
allergenic food. Clearly, there is an urgent need for threshold data, and the
Food Allergy Research and Resource Program of the University of Nebraska
(FARRP) coordinates multi-center studies on this topic (Hefle and Taylor,
2002; Taylor et al., 2002). A standardized protocol for performing threshold
studies has also been developed (Taylor et al., 2004). An overview of threshold
studies and the progress thereof is described in Chapter 1.
Table 12.2 Thresholds for nuts and seeds determined with double-blind, placebo controlled
food challenges
Allergen Reported thresholds Reference

(amount of protein)
Hazelnut 1 mg–1.8 g Wensing et al., 2002

168 mg–1.8 g Ortolani et al., 2000
10–18.2 g (roasted) hazelnut Hansen et al., 2003
> 30 mg (roasted) Wensing et al., 2001b
Pecan 93 mg Bock et al., 1978
Pistachio 88 mg Bock et al., 1978
Sesame 18 mg–1.8 g Kanny et al., 1996
18 mg–1.8 g Kolopp-Sarda et al., 1997
12.4 Allergenic proteins in nuts and seeds

It is clear that allergenic compounds in nuts and seeds lie in the protein
fraction, as in virtually all food allergies. However, the carbohydrate moieties
on proteins may play an important role as well, especially for plant-derived
allergens (van Ree, 2002). Allergenicity of nut oils has been described, but
the allergen in these oils is residual protein that is present in a concentration
that depends on the refining process (Teuber et al., 1997; Crevel et al.,
2000). Since the 1990s, a number of allergens from nuts and seeds have been
identified using IgE-immunoblotting combined with N-terminal protein
sequencing and cloning techniques. An excellent review on plant allergens
was published by Breiteneder and Ebner (2000). An important aspect of nut
and seed allergens with respect to clinical reactions is related to IgE cross-
reactivity with pollen from the corresponding tree or plant (Andersen and
Lowenstein, 1978; Caballero and Martin Esteban, 1998; Etesamifar and
Wuthrich, 1998). For example, allergens homologous to pollen allergen Bet
v1 often lead to oral allergy syndrome, while the more severe reactions are
associated with other allergens (Schocker et al., 2000; Beyer et al., 2002). In
the case of hazelnut, this phenomenon was investigated in detail and reviewed
recently (Pastorello et al., 2001). In particular, storage proteins which are
abundant in nuts and seeds have been identified as allergens. Because of the
often high stability towards heating and digestion, storage proteins can be
potent allergens (Astwood et al., 1996).
12.5 Effect of food processing on allergenicity

Heat treatment is often applied in food processing for obtaining better taste
and smell (roasting, frying) or to extend shelf-life (pasteurization). Typical
processes for nuts and seeds are drying at ambient temperatures, roasting at
140–170 °C and frying at around 180 °C, and further processing in the
preparation of final consumer products, like baking for cookies. Wigotzki et

al. (2000) investigated the IgE-binding properties of hazelnut proteins after
heating and found no effects up to 155 °C. At 170 °C and higher some IgE-
binding was lost, but other allergens remained able to bind IgE (Wigotzki et
al., 2000). In another study, pollen-related hazelnut allergens were found to
be heat-labile, resulting in strongly diminished IgE-binding using serum
from pollen-allergic individuals co-sensitized for hazelnut (Muller et al.,
2000). Using sera from hazelnut allergic subjects who recognized non-pollen
related allergens, on the other hand, showed that part of the hazelnut allergen
fraction was stable to heat treatment (Schocker et al., 2000). True allergenicity
can only be judged in vivo, and data from Hansen et al. (2003) show that
roasted hazelnuts elicited reaction in fewer subjects compared to raw hazelnuts
(Hansen et al., 2003), and Wensing et al. (2001b) showed that thresholds for
roasted hazelnuts were higher compared to raw hazelnuts, indicating a heat-
labile nature of hazelnut allergens.
For the majority of allergenic nuts and seeds, the effect of heat treatment
on allergenicity has yet to be established. Tree nuts and seeds are applied
both in raw and heat-treated form. It is therefore fair to say that the majority
of nut- or seed-allergic individuals will also react, to some extent, on the
heated form of the allergen. It is not expected that food processing will lead
to hypoallergenic foods containing nuts and seeds. Although the general
belief is that heating reduces allergenicity to some extent, there is one report
that describes the contrary. An atopic child that tolerated raw pecan nuts
experienced an anaphylactic shock after ingestion of pecan nuts in cookies
(Malanin et al., 1995). Apparently, baking the cookie resulted in the formation
of neo-allergens or neo-epitopes, probably due to Maillard cross-linking of
sugars to protein (Berrens, 1996). However, this remains the only report of
its kind.
Performing oral challenges and comparing minimum provoking doses is
the best approach to studying the effect of heat treatment on nut and seed
allergens. However, it is difficult to compare such studies from different
medical centers since there are often small deviations in experimental design.
In some studies, it is difficult to discern whether raw or heat-treated nuts or
seeds are used (Hefle and Taylor, 2002).
12.6 Detecting nut and seed residues in food: selecting a

method
Reports of severe allergic reactions in nut- or seed-allergic individuals have
led to the development of methods for detecting residues of specific nuts and
seeds in other foods. The techniques that can be used for developing such
methods are described elsewhere in this book. The next section will start
with an overview on the development of hazelnut detection methods, chosen
because hazelnuts are widely used in food manufacturing. Following this, an

overview of described methods for the detection of other nuts and seeds will
be given. The percentage of hazelnut used in a recipe is considered to be a
mark of quality, for example in chocolate, chocolate products, and nougats.
The first studies on hazelnut detection were therefore related to quantification
of content of hazelnuts rather than to detecting traces.
12.6.1 Detecting hazelnuts

Product adulteration has always been a matter of concern for the food industry
and retailers, especially where it is suspected that expensive ingredients
have been replaced by cheaper alternatives. Hazelnuts in chocolate and
chocolate products are appreciated by many consumers, and the hazelnut
content determines the cost price of certain consumer products to a large
extent. Typically 5–15% hazelnut is added in order to get the desired hazelnut
flavor. In the mid 1980s, the first method for quantifying the amount of
hazelnut in food products was reported. The method was based on
electroimmunodiffusion (Laurell, 1972) using antiserum from laboratory
animals and detected 0.01–0.1% of hazelnut in food products corresponding
to 100–1000 ppm (Klein et al., 1985). The investigated matrix was nougat
for which the recovery was > 95% (Klein et al., 1985). Since roasting hazelnut
is a common practice to improve the flavor, the reactivity of roasted hazelnuts
was investigated with this method as well. Up to 135 °C, the reactivity
decreased to 70–80% of the original value, and above 160 °C, the reactivity
was too low for proper analysis (Klein and Guenther, 1985).
A similar technique was used by Malmheden-Yman et al. (1994) who
analyzed several foods for undeclared presence of hazelnut. This was the
first time that hazelnut detection was used in relation to the safety of hazelnut-
allergic individuals. A complaint sample of chocolate, supposedly free of
hazelnut, contained 0.2% of hazelnut, leading to an asthmatic reaction
(Malmheden Yman et al., 1994). Thresholds for hazelnuts were not known in
the 1990s, but for peanut, thresholds as low as 100 μg were reported (Hourihane
et al., 1997). Using this figure, our group aimed at a detection limit of 1–10
ppm when developing a new immunological assay for the detection of hazelnuts.
A polyclonal antibody was raised in rabbits, and its reactivity was compared
to a pool of serum from hazelnut-allergic patients. The rabbit serum detected
proteins on immunoblots that bound IgE, indicating that relevant proteins
were detected (Koppelman et al., 1999). To improve the sensitivity of the
method, several ELISA formats were evaluated, and a sandwich ELISA
using the rabbit antiserum gave the best results. Around 30 ng/mL hazelnut
protein could be detected (both raw and roasted) and, taking into account the
required extraction and dilution steps, this detection limit corresponds to
0.1 ppm in food samples. As shown in Fig. 12.1, the assay detected both raw
and roasted hazelnuts. This assay was applied to the analysis of several
suspected samples, and one case report on an allergic reaction to chocolate
2.5
2
A 490
1.5
0.5
0
1 10 100 1000 10 000
Hazelnut protein (ng/ml)
Fig. 12.1 Detection of hazelnut proteins by sandwich ELISA using polyclonal

antibodies. Circles: dilutions of an extract of raw hazelnuts; squares: dilutions of an
extract of roasted hazelnuts. Reprinted with permission from the Journal of
Immunological Methods, Koppelman et al., 1999).
spread that inadvertently contained hazelnut was recently published (Wensing

et al., 2001a). Holzhauser published a similar sandwich ELISA, with a
comparable sensitivity (Holzhauser and Vieths, 1999). Since 1999, several
other immunochemical assays for the detection of hazelnut have been reported.
In addition to ELISA-based methods, dipstick immunoassays, IgE and IgG
immunoblots, and PCR techniques have been described. Table 12.3 gives an
overview of the methods published to date.
Although hazelnut detection via DNA is reasonably sensitive, one should
keep in mind that such an approach is indirect. It is not the allergenic fraction
but a marker molecule which is measured. Hazelnuts are often used as whole
nuts, with DNA and proteins probably present in a constant ratio, but when
hazelnut oil, hazelnut meal, or extracted hazelnut flavor are used, it is not
clear whether DNA and protein remain together; test results may become
false positive or, worse, false negative. An advantage of detection via DNA
is the specificity of the assay. Primers can be chosen that uniquely react with
hazelnut DNA (Holzhauser et al., 2000), while immunoassays dependent on
antibodies can show some degree of non-specificity. Specificity is of particular
importance when assessing the content of hazelnut in products like hazelnut-
containing chocolate, or nougats (Klein et al., 1985), related to product
adulteration. On the other hand, hazelnut-allergic consumers are often also
sensitized to other tree nuts (Roux et al., 2003). False positive test results of
hazelnut-free products caused by the presence of other tree nuts or peanuts
will not change the conclusion of the hazelnut-allergic consumers, because
Table 12.3 Detection methods for hazelnut residues (in order of publication date)
Analyte Technique Key reagent Detection limit Reference
Hazelnut protein Electroimmunodiffusion Polyclonal animal serum 100–1000 ppm Klein and Guenther, 1985;
Klein et al., 1985;
Malmheden Yman et al., 1994
Hazelnut allergens Immunoblot Hazelnut-allergic serum Not reported Teuber et al., 1997
Hazelnut protein Immunoblot Polyclonal rabbit serum 300 ppm Koppelman et al., 1999
Hazelnut allergens Inhibition ELISA Hazelnut-allergic serum 0.1 ppm Koppelman et al., 1999
Hazelnut protein Sandwich ELISA Polyclonal rabbit serum 0.1 ppm Koppelman et al., 1999
Hazelnut protein Sandwich ELISA Polyclonal rabbit serum 0.2 ppm Holzhauser and Vieths, 1999
Hazelnut DNA PCR-hybridization Specific DNA primers 10 ppm Holzhauser et al., 2000
Hazelnut protein Dot blot Egg yolk IgY 1 ppm Blais and Phillippe, 2000
Hazelnut protein Immunoblot Polyclonal rabbit serum 5 ppm Scheibe et al., 2001
Hazelnut protein Sandwich ELISA Egg yolk IgY 1 ppm Blais and Phillippe, 2001
Hazelnut protein Biacore immunoassay Polyclonal rabbit serum 2 ppm Jonsson and Hellenas, 2001
Hazelnut protein Dipstick (sandwich) Polyclonal rabbit serum 1 ppm (only Stephan et al., 2002
qualitative)
Hazelnut 2S albumin Basophil histamine release Hazelnut-allergic serum < 0.1 ppm Akkerdaas et al., 2002a
Hazelnut protein Sandwich ELISA Polyclonal rabbit serum < 0.1 ppm Akkerdaas et al., 2002b
food products that contain nuts other than hazelnut will also be avoided by
hazelnut-allergic consumers. Nevertheless, specificity is an important aspect
and, since tree nuts are phylogentically related, immunochemical cross-
reactivity leading to non-specificity is expected. The theoretical explanation
for immunochemical cross-reactivity is found in the antibodies recognizing
specific parts of the target proteins. Amino acid sequences of nut proteins are
sometimes very similar and homologous to a high extent as in the case for
the family of 2S storage proteins (Ampe et al., 1986; Teuber et al., 1998).
Antiserum may recognize epitopes on the immunogen as well as on homologous
proteins in other species. Cross-reacting epitopes usually have lower affinities
compared to the original epitopes, and the recognition of other species can
be orders of magnitude less. It is interesting to note that four different research
groups who raised hazelnut antisera independently describe the same cross-
reacting species; walnut appeared to cross-react in all raised antibodies,
while almond and cashew cross-reacted with three and two antisera, respectively
(Holzhauser and Vieths, 1999; Koppelman et al., 1999; Blais and Phillippe,
2001; Scheibe et al., 2001). One antiserum recognized pumpkin seed, and
another reacted with pine nut and peanut, although the latter two were a
100 000 fold less compared to hazelnut (Koppelman et al., 1999) and therefore
considered to be negligible. Other tree nuts, seeds, and legumes were tested
for only two of the sera, and appeared to be negative (Holzhauser and Vieths,
1999; Koppelman et al., 1999), while for the other sera an extensive
investigation of cross-reactivity was not performed (Blais and Phillippe,
2001; Scheibe et al., 2001). One antiserum was not tested for possible cross-
reactivity (Jonsson and Hellenas, 2001).
Recovery of analytes, in this case hazelnut protein, from food matrices is
sometimes difficult because of food processing steps. Heat treatment especially
can denature proteins, thereby decreasing the solubility. This aspect is discussed
in detail elsewhere in this book. For hazelnut detection, the main areas of
application are chocolates and bakery products. Extraction from chocolate
can be particularly difficult due to binding of tannins to proteins, limiting the
protein solubility. Holzhauser and Vieths investigated the recovery of hazelnut
protein from various spiked samples as analyzed with their assay. They
found an average recovery of 106 ± 17%, although for samples spiked with
lower amounts of hazelnut, the recovery varied between 62 and 132%
(Holzhauser and Vieths, 1999). Recoveries of commercial kits are optimized
by adding tannin-binding components such as skimmed milk powder or fish
gelatin (Keck-Gassenmeier et al., 1999). Another aspect that may influence
the quantification of hazelnut in food is roasting. Roasting of hazelnut has
two effects on immunochemical analysis: (i) changing the immunochemical
properties of the target protein, for example by protein denaturation and
Maillard reactions; and (ii) limitation of solubility. It is therefore important
to assess the effect of roasting on the quantitative determination of hazelnuts.
One report shows that dilution curves of raw and roasted hazelnut protein
were almost identical with an estimated maximum difference of about 10%
(Koppelman et al., 1999), and another report shows a difference between

raw and roasted of about 20% (Holzhauser and Vieths, 1999). For the IgY-
based assay it was shown that roasting did not significantly change the
reactivity (Blais and Phillippe, 2001). For the PCR method it was demonstrated
that the PCR led to amplified DNA for both raw and roasted hazelnut, but the
possible differences were not quantified (Holzhauser et al., 2000). In the
case of detecting hazelnut residues related to food safety of hazelnut-allergic
consumers, true quantification is not necessary. Differences between detection
of raw and roasted hazelnut would be important if the roasted form is only
recognized with a sensitivity that is orders of magnitude less, or if roasted
hazelnuts are not detected at all, but this is not the case in the published
methods.
In summary, there are a number of methods published for the detection of
hazelnuts in foods. The majority are based on specific antibodies, and all
have rather similar characteristics with respect to sensitivity and specificity.
A commercial method for detection of hazelnut is available, and several
laboratories offer an analytical service. A list with website addresses for
commercially available methods for tree nuts and seeds is provided in Table
12.5 on p. 214.
12.6.2 Overview of methods for other tree nuts and seeds

While a number of assays have been described for hazelnut, the published
methods for other nuts and seeds are very limited. Radioallergosorbent assay
(RAST) inhibition techniques with defined patient sera may provide a tool
for allergen-specific measurement when no animal sera are available (see
Chapter 4). The section below describes published methods for specific nuts
and seeds.
Almond
For almond, polyclonal antibodies were raised in rabbits, and successfully
applied in immunoblot and competition ELISA (Acosta et al., 1999). The
reactivity was strongly influenced by heat treatment and pH treatment, and
some cross-reactivity with other nuts and legumes was observed (Acosta
et al., 1999). The same group reported an optimized ELISA for almond and
recovery from different matrices, and in addition the effect of heat treatment
was studied (Roux et al., 2001). The detection limit of the method was 5 ppm
in processed foods, and neither blanching, roasting, nor autoclaving markedly
decreased the reactivity (Roux et al., 2001). Meanwhile Hlykwa et al. published
a method to detect almond residues in food as well (Hlywka et al., 2000).
This sandwich ELISA detected down to 1 ppm, and did not cross-react
substantially with other food ingredients, although some reactivity was observed
for sesame seeds, and to a lesser extent for walnuts and macademia, (Hlywka
et al., 2000). The assay was used in a survey of 20 brands of dry breakfast
cereal, and in these products no almond cross-contact was observed.
Interestingly, some products did not contain detectable amounts of almond

although the manufacturer included almond on the ingredient list. It is likely
that the reason for labeling was that the manufacture could not exclude the
possibility of cross-contact with almonds (Hlywka et al., 2000). The recovery
of almond from difficult matrices such as chocolate was 86–100%, possibly
owing to the heating step during extraction (Hlywka et al., 2000).
Brazil nut
Blais et al. (2002) developed an ELISA for the detection of Brazil nut. This
assay was based on egg yolk antibodies (IgY) and reached a sensitivity of
well below 1 ppm. Cross-reactivity with other ingredients was investigated
and appeared to be negligible for peanut, hazelnut, sunflower seed, cucumber
seed, and castor beans (Blais et al., 2002). The antibody was later applied in
a multiplex enzyme immunoassay for the detection of peanut, hazelnut, and
Brazil nut with a sensitivity of around 1 μg protein per gram (1 ppm) (Blais
et al., 2003). The advantage of this test is that important food allergens that
are often used in chocolate manufacturing are assayed simultaneously. Clemente
et al. (2004) developed an inhibition ELISA for Brazil nut 2S albumin, with
a limit of detection of 1 ppm. Cross-reactivity with other nuts and legumes
including peanuts was negligible, and the food products with Brazil nut
declared were judged positive (Clemente et al., 2004).
Walnut
An ELISA for walnut detection has been described by Niemann and Hefle
(2003). The antibodies were raised against a mixture of several varieties of
walnut, both raw and roasted, in sheep and rabbits. Levels of 1 ppm were
detected, and recovery from different food matrices including chocolate was
between 85 and 96% even when low concentrations of walnut were spiked.
Of a set of 50 main food ingredients, only pecan which is a close relative of
walnut cross-reacted significantly (Niemann and Hefle, 2003).
Cashew
Recently, the first assay for cashew nut detection was published. Antibodies
were raised in rabbits and goats against major storage protein. The resulting
antisera were immunoadsorbed with immobilized protein of several nuts,
legumes, and seeds in order to decrease possible cross-reaction (Wei et al.,
2003). Immunoblotting studies demonstrated that the polyclonal sera recognized
proteins similar to those recognized by IgE from cashew nut-allergic patients.
Cross-reactivity of other food ingredients was low, although pistachio and
sunflower seed showed significant cross-reaction. A sensitivity of 0.02 ppm
was reached and recoveries from divers food matrices varied from 64 to
136%, while heating did not affect the reactivity of cashew in the test (Wei
et al., 2003).
2
1.8
1.6
1.4
1.2
A 490
1
0.8
0.6
0.4
0.2
0
0.001 0.01 0.1 1 10 100
Mustard protein (μg/ml)
Fig. 12.2 Calibration curve of mustard seed protein in an inhibition ELISA for
mustard using rabbit polyclonal antibodies. The line represents serial dilutions of
mustard seed protein.
Sesame
Brett et al. (1998) purified the 13S protein fraction of sesame seeds and
raised polyclonal antibodies in rabbits towards it. IgG-immunoblots indicated
that the serum was specific although the authors did not provide a complete
list of tested food ingredients. Using ELISA, detection of sesame was possible
(no detection limit described) (Brett et al., 1998).
Mustard
EU legislation on labeling of food allergens also includes mustard and products
thereof (EU, 2003). Up to now, no methods are either commercially available
or published. Recent work from our group led to a polyclonal antiserum for
mustard seed proteins, and an inhibition ELISA was constructed. The test
performance has been sorted out for some food applications, and Fig. 12.2
shows a calibration curve for mustard seed protein. Mustard seed protein at
a concentration of around 50 ng/ml results in significant inhibition. Taking
into account the necessary dilution steps, the sensitivity of this ELISA is
around 1–2 ppm.
Table 12.4 gives an overview of the published methods for nuts and seeds.
12.7 Conclusions
For hazelnut detection a large number of (ELISA) tests has been described,
and there are test kits commercially available. For other nuts and seeds, there
remains a lot of work to do. Some tests are already commercially available
and used by the food industry (Table 12.5). The assays described up to now
have detection limits around 1 ppm. This is suitable for allergen detection, as
clinical case reports and threshold studies indicate that allergen levels of
Table 12.4 Detection methods for nuts and seeds other than hazelnut
Analyte Technique Key reagent Detection limit Reference
Almond protein Inhibition ELISA Polyclonal rabbit serum 5 ppm Acosta et al., 1999;
Roux et al., 2001
Almond protein Sandwich ELISA Polyclonal rabbit and 1 ppm Hlywka et al., 2000
sheep serum
Brazil nut protein Sandwich ELISA Egg yolk IgY < 0.1 ppm Blais et al., 2002
Walnut protein Sandwich ELISA Polyclonal rabbit and 1 ppm Niemann and Hefle, 2003
sheep serum
Cashew storage protein Sandwich ELISA Polyclonal rabbit and < 0.1 ppm Wei et al., 2003
goat serum
Sesame 13S protein Inhibition ELISA Polyclonal rabbit serum Not specified Brett et al., 1998
Table 12.5 Commercially available detection methods for nut and seed allergens1
Target Type of Kit Sensitivity Web site

allergen method manufacturer
Hazelnut Sandwich ELISA R-Biopharm 10 ppm www.r-biopharm.com

Almond Sandwich ELISA Neogen 2.5 ppm www.neogen.com
Almond Sandwich ELISA R-Biopharm 1.7 ppm www.r-biopharm.com
Sesame Sandwich ELISA Tepnel 2.5 ppm www.tepnel.com
Biosystems
1
Note that peanut detection is described in Chapter 11 and is therefore not included in this table
10 ppm are relevant (discussed in Chapter 1). A major problem with test
development for allergen detection from nuts and seeds is that nuts and seeds
are closely related phylogenetically. Therefore, raising antibodies that are
specific is a challenge. On the other hand, products tested for certain analytes,
for example traces of hazelnut, should in most cases also be free of other
nuts. In the case that nut speciation is desired, PCR techniques may be
helpful as they are usually more specific compared to immunochemical
techniques.
Commercially available ELISA kits have been improved with respect to
performance. Most optimized tests use three incubation steps of 10 minutes
only (Immer et al., 2004). Compared to sample preparation which may take
considerable time (milling, mixing, heating, and centrifugation) the test itself
is no longer the rate-limiting step. We foresee a future for ELISA-based
methods in the detection of nut and seed allergenic residues. With new
polyclonal antibodies produced and characterized by several groups around
the world, the availability of new kits will increase. It is, however, difficult
to predict which nuts and seeds will be considered the most important allergens
in the future as clinical reactions for a wide variety of nuts and seeds have
been described. For the nuts and seeds that are common food ingredients in
Western diets, we believe that a complete set of ELISA-based kits will be
available in the coming years.
12.8 References
Acosta, M R, Roux, K H, Teuber, S S and Sathe, S K (1999) ‘Production and characterization
of rabbit polyclonal antibodies to almond (Prunus dulcis L.) major storage protein’, J
Agric Food Chem, 47, 4053–4059.
Akkerdaas, J, Schocker, F, Alcocer, M J, Schilte, P, Knulst, A C, Hefle, S L, Aalberse, R
and van Ree, R V (2002a) ‘Detection and immunological characterization of hazelnut
2S albumin’, J Allergy Clin Immunol, 109, s328.
Akkerdaas, J, Wensing, M, Knulst, A C, Hefle, S L, Aalberse, R and van Ree, R (2002b)
‘Sensitive and specific measurement of crude and processed hazelnut protein by a
sandwich-type ELISA’, J Allergy Clin Immunol, 109, s302.
Ampe, C, Van Damme, J, de Castro, L A, Sampaio, M J, Van Montagu, M and
Vandekerckhove, J (1986) ‘The amino-acid sequence of the 2S sulphur-rich proteins

from seeds of Brazil nut (Bertholletia excelsa H.B.K.)’, Eur J Biochem, 159, 597–
604.
Andersen, K E and Lowenstein, H (1978) ‘An investigation of the possible immunological
relationship between allergen extracts from birch pollen, hazelnut, potato and apple’,
Contact Dermatitis, 4, 73–79.
Asero, R, Mistrello, G, Roncarolo, D, Antoniotti, P L and Falagiani, P (1999) ‘A case of
sesame seed-induced anaphylaxis’, Allergy, 54, 526–527.
Astwood, J D, Leach, J N and Fuchs, R L (1996) ‘Stability of food allergens to digestion
in vitro’, Nat Biotechnol, 14, 1269–1273.
Axelsson, I G, Ihre, E and Zetterstrom, O (1994) ‘Anaphylactic reactions to sunflower
seed’, Allergy, 49, 517–520.
Berrens, L (1996) ‘Neoallergens in heated pecan nut: products of Maillard-type degradation’?
Allergy, 51, 277–278.
Beyer, A V, Gall, H and Peter, R U (1998) ‘Anaphylaxis to pine nuts’, Allergy, 53, 1227–
1228.
Beyer, K, Grishina, G, Bardina, L, Grishin, A and Sampson, H A (2002) ‘Identification
reactions’, J Allergy Clin Immunol, 110, 517–523.
Blais, B W and Phillippe, L M (2000) ‘A cloth-based enzyme immunoassay for detection
of peanut proteins in foods’, Food Agric Immunol, 12, 243–248.
Blais, B W and Phillippe, L (2001) ‘Detection of hazelnut proteins in foods by enzyme
immunoassay using egg yolk antibodies’, J Food Prot, 64, 895–898.
Blais, B W, Omar, M and Phillippe, L (2002) ‘Detection of Brazil nut proteins in foods
by enzyme immunoassay’, Food Agric Immunol, 14, 163–168.
Blais, B W, Gaudreault, M and Phillippe, L (2003) ‘Multiplex enzyme immunoassay
system for the simultaneous detection of multiple allergens in foods’, Food Control,
14, 43–47.
Bock, S A, Lee, W Y, Remigio, L K and May, C D (1978) ‘Studies of hypersensitivity
reactions to foods in infants and children’, J Allergy Clin Immunol, 62, 327–334.
Bock, S A, Munoz-Furlong, A and Sampson, H A (2001) ‘Fatalities due to anaphylactic
Breiteneder, H and Ebner, C (2000) ‘Molecular and biochemical classification of plant-
derived food allergens’, J Allergy Clin Immunol, 106, 27–36.
Brett, G M, Bull, V J and Morgan, M R (1998) ‘Identification of hidden allergens within
foods’, Allergy, 53, 109–110.
Caballero, T and Martin Esteban, M (1998) ‘Association between pollen hypersensitivity
and edible vegetable allergy: A review’, J Investig Allergol Clin Immunol, 8, 6–16.
Clemente, A, Chambers, S J, Lodi, F, Nicoletti, C and Brett, G M (2004) ‘Use of the
indirect competitive ELISA for the detection of Brazil nut in food products’, Food
Control, 15, 65–69.
Crevel, R W R, Kerkhoff, M A T and Koning, M M G (2000) ‘Allergenicity of refined
vegetable oils’, Food Chem Toxicol, 38, 385–393.
Etesamifar, M and Wüthrich, B (1998) ‘IgE-mediated food allergies including oral allergy
syndrome in 383 patients’, Allergologie, 21, 451–457.
EU (2003) ‘Directive 2003/89/EC amending Directive 2000/13/EC as regards indication
of the ingredients present in foodstuffs’, Official Journal of the European Union, L
308/15.
FAO (1995) Report of the FAO Technical Consultation of Food Allergies, Rome, FAO.
Frantzen, B, Brocker, E B and Trautmann, A (2000) ‘Immediate-type allergy caused by
poppy seed’, Allergy, 55, 97–98.
Galloway, D (2000) ‘Oilseed rape – allergen or irritant?’ Clin Exp Allergy, 30, 308–309.
Hansen, K S, Ballmer-Weber, B K, Luttkopf, D, Skov, P S, Wüthrich, B, Bindslev-Jensen,
C, Vieths, S and Poulsen, L K (2003) ‘Roasted hazelnuts – allergenic activity evaluated
by double-blind, placebo-controlled food challenge’, Allergy, 58, 132–138.
Hefle, S L and Taylor, S L (2002) ‘How much food is too much? Threshold doses for
allergenic foods’, Curr Allergy Asthma Rep, 2, 63–66.
Hlywka, J J, Hefle, S L and Taylor, S L (2000) ‘A sandwich enzyme-linked immunosorbent
assay for the detection of almonds in foods’, J. Food Prot 63, 252–257.
Holzhauser, T and Vieths, S (1999) ‘Quantitative sandwich ELISA for determination of
traces of hazelnut (Corylus avellana) protein in complex food matrixes’, J Agric Food
Chem, 47, 4209–4218.
Holzhauser, T, Wangorsch, A and Vieths, S (2000) ‘Polymerase chain reaction (PCR) for
detection of potentially allergenic hazelnut residues in complex food matrixes’, Eur
Food Res Technol, 211, 360–365.
Hourihane, J O, Bedwani, S J, Dean, T P and Warner, J O (1997) ‘Randomised, double
blind, crossover challenge study of allergenicity of peanut oils in subjects allergic to
peanuts’, Br Med J, 314, 1084–1088.
Hu, F B and Stampfer, M J (1999) ‘Nut consumption and risk of coronary heart disease:
a review of epidemiologic evidence’, Curr Atheroscler Rep 1, 204–209.
Immer, U, Reck, B, Lindeke, S and Koppelman, S (2004) ‘RIDASCREEN ® FAST
PEANUT, a rapid and safe tool to determine peanut contamination in food’, Intl J
Food Sci Tech, 39, 869–871.
Jonsson, H and Hellenas, K E (2001) ‘Optimizing assay conditions in the detection of
food allergens with Biacore’s SPR technology’, Biacore J, 2, 16–18.
Kanny, G, de Hauteclocque, C and Moneret-Vautrin, D A (1996) ‘Sesame seed and
sesame seed oil contain masked allergens of growing importance’, Allergy, 51, 952–
957.
Keck-Gassenmeier, B, Benet, S, Rosa, C and Hischenhuber, C (1999) ‘Determination of
11, 243–250.
Klein, E and Guenther, H O (1985) [Determination of hazelnut proteins by
electroimmunodiffusion. II. Dependence of protein yield on the degree of roasting.]
‘Bestimmung von Haselnussprotein (Corylin) mit Hilfe der Elektroimmundiffusion
nach Laurell. II. Abhaengigkeit der Proteinausbeute vom Roestgrad der Produkte’, Z
Lebens Unters Forsch, 180, 36–40.
Klein, E, Baudner, S and Guenther, H O (1985) ‘[Determination of hazelnut proteins by
electroimmunodiffusion. I. Optimized methodology.] ‘Immunchemische Bestimmung
des Haselnussproteins mit Hilfe der Elektroimmundiffusion nach Laurell. I. Mitteilung
der optimierten Methode’, Z Lebens Unters Forsch, 180, 30–35.
Kolopp-Sarda, M N, Moneret-Vautrin, D A, Gobert, B, Kanny, G, Brodschii, M, Bene, M
C and Faure, G C (1997) ‘Specific humoral immune responses in 12 cases of food
sensitization to sesame seed’, Clin Exp Allergy, 27, 1285–1291.
Koppelman, S J, Knulst, A C, Koers, W J, Penninks, A H, Peppelman, H, Vlooswijk, R,
Pigmans, I, van Duijn, G and Hessing, M (1999) ‘Comparison of different
immunochemical methods for the detection and quantification of hazelnut proteins in
food products’, J Immunol Methods, 229, 107–120.
Laurell, C B (1972) ‘Electroimmuno assay’, Scand J Clin Lab Invest Suppl, 124, 21–37.
Lezaun, A, Fraj, J, Colas, C, Duce, F, Dominguez, M A, Cuevas, M and Eiras, P (1998)
‘Anaphylaxis from linseed’, Allergy, 53, 105–106.
Malanin, K, Lundberg, M and Johansson, S G (1995) ‘Anaphylactic reaction caused by
neoallergens in heated pecan nut’, Allergy, 50, 988–991.
Malmheden Yman, I, Eriksson, A, Everitt, G and Yman, L (1994) ‘Analysis of food
proteins for verification of contamination or mislabelling’, Food Agric Immunol, 6,
167–172.
Muller, U, Luttkopf, D, Hoffman, A, Petersen, A, Becker, W M, Schocker, F, Niggeman,
B, Altmann, F, Kolarich, D, Haustein, D and Vieths, S (2000) ‘Allergens in raw and
roasted hazelnuts (Corylus avellana) and their cross-reactivity to pollen’, Eur Food
Res Technol, 212, 2–12.
Niemann, L and Hefle, S L (2003) ‘Validated ELISA for detection of undeclared walnut
residues in food’, J Allergy Clin Immunol, 109, s248.
Nordlee, J A, Taylor, S L, Townsend, J A, Thomas, L A and Bush, R K (1996) ‘Identification
of a Brazil-nut allergen in transgenic soybeans’, N Engl J Med, 334, 688–692.
Ortolani, C, Ballmer-Weber, B K, Hansen, K S, Ispano, M, Wüthrich, B, Bindslev-
Jensen, C, Ansaloni, R, Vannucci, L, Pravettoni, V, Scibilia, J, Poulsen, L K and
Pastorello, E A, (2000) ‘Hazelnut allergy: a double-blind, placebo-controlled food
challenge multicenter study’, J Allergy Clin Immunol, 105, 577–581.
Pastorello, E A, Pompei, C, Pravettoni, V, Brenna, O, Farioli, L, Trambaioli, C and Conti,
A (2001) ‘Lipid transfer proteins and 2S albumins as allergens’, Allergy, 56 (Suppl
67), 45–47.
Roux, K H, Teuber, S S, Robotham, J M and Sathe, S K (2001) ‘Detection and stability
of the major almond allergen in foods’, J Agric Food Chem 49, 2131–2136.
Roux, K H, Teuber, S S and Sathe, S K (2003) ‘Tree nut allergens’, Int Arch Allergy
Immunol, 131, 234–244.
Scheibe, B, Weiss, W, Rueff, F, Przybilla, B and Gorg, A (2001) ‘Detection of trace
B Biomed Sci Appl, 756, 229–237.
Schocker, F, Luttkopf, D, Muller, U, Thomas, P, Vieths, S and Becker, W M (2000) ‘IgE
binding to unique hazelnut allergens: Identification of non pollen-related and heat-
stable hazelnut allergens eliciting severe allergic reactions’, Eur J Nutr, 39, 172–180.
peanut and tree nut allergy in the US determined by a random digit dial telephone
Sporik, R and Hill, D (1996) ‘Allergy to peanut, nuts, and sesame seed in Australian
children [letter]’, Br Med J, 313, 1477–1478.
Sten, E, Stahl, S P, Andersen, S B, Torp, A M, Olesen, A, Bindslev-Jensen, U, Poulsen,
L K and Bindslev-Jensen, C (2002) ‘Allergenic components of a novel food, Micronesian
nut Nangai (Canarium indicum), shows IgE cross-reactivity in pollen allergic patients’,
Allergy, 57, 398–404.
Stephan, O, Moller, N, Lehmann, S, Holzhauser, T and Vieths, S (2002) ‘Development
and validation of two dipstick type immuno assays for the determination of trace
amounts of peanut and hazelnut in processed food’, Eur Food Res Technol, 215, 431–
436.
Tariq, S M, Stevens, M, Matthews, S, Ridout, S, Twiselton, R and Hide, D W (1996)
‘Cohort study of peanut and tree nut sensitisation by age of 4 years’, Br Med J, 313,
514–517.
Taylor, S L and Nordlee, J A (1996) ‘Detection of food allergens’, Food Technol, 50, 231–
238.
Taylor, S L, Hefle, S L, Bindslev-Jensen, C, Bock, S A, Burks, Jr, A W, Christie, L, Hill,
D J, Host, A, Hourihane, J O, Lack, G, Metcalfe, D D, Moneret-Vautrin, D A, Vadas,
much is too much?’ J Allergy Clin Immunol, 109, 24–30.
Taylor, S L, Hefle, S L, Bindslev-Jensen, C, Atkins, F M, Andre, C, Bruijnzeel-Koomen,
C, Burks, A W, Bush, R K, Ebisawa, M, Eigenmann, P A, Host, A, Hourihane, J O,
Isolauri, E, Hill, D J, Knulst, A, Lack, G, Sampson, H A, Moneret-Vautrin, D A,
Rance, F, Vadas, P A, Yunginger, J W, Zeiger, R S, Salminen, J W, Madsen, C and
Abbott, P. (2004) ‘A consensus protocol for the determination of the threshold doses
Teuber, S S, Brown, R L and Haapanen, L A (1997) ‘Allergenicity of gourmet nut oils
processed by different methods’, J Allergy Clin Immunol, 99, 502–507.
Teuber, S S, Dandekar, A M, Peterson, W R and Sellers, C L (1998) ‘Cloning and
sequencing of a gene encoding a 2S albumin seed storage protein precursor from
English walnut (Juglans regia), a major food allergen’, J Allergy Clin Immunol, 101,
807–814.
van Ree, R (2002) ‘Carbohydrate epitopes and their relevance for the diagnosis and
treatment of allergic diseases’, Int Arch Allergy Immunol, 129, 189–197.
Wei, Y, Sathe, S K, Teuber, S S and Roux, K H (2003) ‘A sensitive sandwich ELISA for
the detection of trace amounts of cashew (Anacardium occidentale L.) nut in foods’,
J Agric Food Chem, 51, 3215–3221.
Wensing, M, Koppelman, S J, Penninks, A H, Bruijnzeel-Koomen, C A F M and Knulst,
A C (2001a) ‘Hidden hazelnut is a threat to allergic patients’, Allergy, 56, 191–192.
Wensing, M, Penninks, A H, Hefle, S L, Akkerdaas, J, Ree, R V, Koppelman, S J,
Bruijnzeel-Koomen, C A F M and Knulst, A C (2001b) ‘Determination of threshold
levels of patients with hazelnut allergy using double blind placebo controlled food
challenges (DBPCFC’s) 8th International Symposium on Immunological and Clinical
Problems of Food Allergy, Venice, March 11–13, 63.
Wensing, M, Penninks, A H, Hefle, S L, Akkerdaas, J H, van Ree, R, Koppelman, S J,
Bruijnzeel-Koomen, C A and Knulst, A C (2002) ‘The range of minimum provoking
doses in hazelnut-allergic patients as determined by double-blind, placebo-controlled
food challenges’, Clin Exp Allergy, 32, 1757–1762.
Wigotzki, M, Steinhart, H and Paschke, A (2000) ‘Influence of varieties, storage and heat
treatment on IgE-binding proteins in hazelnuts’, (Corylus avellana), Food Agric Immunol,
12, 217–229.
Yunginger, J W, Sweeney, K G, Sturner, W Q, Giannandrea, L A, Teigland, J D, Bray, M,
Benson, P A, York, J A, Biedrzycki, L and Squillace, D L (1988) ‘Fatal food-induced
anaphylaxis’, JAMA, 260, 1450–1452.
13
Detecting dairy and egg residues in food

C. Demeulemester and I. Giovannacci, Centre Technique de la
Salaison, de la Charcuterie et des Conserves de Viandes
(CTSCCV), France, and V. Leduc, Allerbio, France
13.1 Introduction
The major sources of food allergens having animal origin are milk, egg
(especially egg white), crustaceans and fish. Allergies to meat are less frequent.
Crustaceans and fishes are less often used as processed ingredients, except in
surimi, flavors and spices using shrimp powders. They are generally consumed
as clearly recognizable and labelled ingredients. Moreover, they are usually
eaten later in life. In contrast, milk and egg are widely consumed as basic
food products, as milk is the first food for babies. They constitute a good
source of animal protein for a modest financial cost. They are also traditionally
used around the world as food ingredients due to their useful technological
properties such as foaming, jellification, thickening, emulsifying and binding.
They are used in bakery and cakes, delicatessen and processed meat products,
soups, dressings and sauces, among other applications.
Food allergies to cows’ milk and hen’s egg white are frequent in man, and
they are the two prevalent food allergies in children. Symptoms of egg and
milk allergy are identical to those for other food allergies (see Chapter 1).
Many studies have been published concerning the prevalence of egg and
milk allergies; results may differ depending on the selection (food-allergic
patients or unselected population) or on the age of the population studied,
but they show that egg and milk allergies are the most prevalent among food
allergy. In studies on unselected (general) populations, milk allergy could
represent about 2.0% in Australia (Hill et al., 1997, 1999), 1.9% in Finland
(Saarinen et al., 2000), 3.2% in France (Kanny et al., 2001), 3.9% in Germany
(Schäfer et al., 1999), 4.0% in Japan (Iikura et al., 1999) and 1.4% in the US
(Rhim and McMorris, 2001) and egg allergy could represent about 3.2% in
Australia (Hill et al., 1999) and 2.8% in Germany (Schäfer et al., 1999).
Allergic reactions due to the presence of egg white proteins in vaccines were
also shown by Davies and Pepys (1976). Pichler and Campi (1992) reported
an allergy case due to the presence of lysozyme contained in vaginal
suppositories.
The amount of allergenic food that can provoke allergic reactions is not
well known. There is variability in patient sensitivity and in allergen specificity.
Food ingredients can be enriched with allergens due to the production methods
or processing. Heat treatment may decrease or increase the allergenic potential
of food allergens. For all these reasons, according to studies performed by
clinicians, it is considered that detection techniques for allergenic sources
should be around 1–10 ppm, i.e. 1–10 mg/kg (Moneret-Vautrin et al., 2003;
Morisset et al., 2003; Osterballe and Bindslev-Jensen, 2003). Most commercial
methods have detection limits encompassing this range.
Taken together, these reasons indicate the need for quick, sensitive,
economical and available detection methods for milk and egg. Analyses
using these methods are carried out in order to ensure the fairness of commercial
transactions and the protection of food-allergic consumers.
13.2 Milk
13.2.1 Cows’ milk allergy
Cows’ milk allergy (CMA) is a common disease of infancy and early childhood,
with a prevalence of 8% reported in a population of 1121 people with food
allergies (Kanny et al., 2001). CMA affects 2.5% of children under two
years of age, but about 80% become clinically tolerant within the first three
years of life (Høst, 1997). Symptoms of CMA are the usual symptoms of
food-allergic reactions. They are mainly urticaria (76%), atopic dermatitis
(28%), vomiting (9%) appearing in less than two hours (68%) at cumulative
reactive doses below 20 mL (59%) and between 20 mL and 200 mL (32%)
(Saarinen et al., 2000).
More recent studies have tried to evaluate the threshold doses of clinical
reactivity to milk allergens. Clinical allergists have published existing data
on threshold doses for several food allergens such as milk, egg, peanut, fish
and mustard (Hefle and Taylor, 2002; Taylor et al., 2002; and Taylor et al.,
2004). For milk, the lowest amount which provoked symptoms by double-
blind placebo-controlled, single-blind or open challenge ranged from 0.02–
5 mL corresponding, respectively, to 0.6–180 mg of protein. Morisset et al.
(2003) presented results of 59 positive oral challenges to milk. A low threshold
inferior or equal to 0.3 mL of milk was observed in 1.7% of milk allergies
and a cumulative reactive dose inferior or equal to 0.8 mL of milk characterized
5% of milk allergies. The lowest reactive dose found was 0.1 mL of milk.
Sicherer et al. (1999) reported that 25% of milk allergics reacted to doses of
100 mg.
Detecting dairy and egg residues in food 221
13.2.2 Cows’ milk allergens

The protein content of cows’ milk is about 30–35 g/L, divided into 20%
whey and 80% caseins (see Table 13.1). Caseins precipitate under acidic pH
(pH 4.6) or with the action of the chymosin and are the major components of
cheese. Caseins (allergen nomenclature: Bos d 8) are a family of related
proteins, divided into α, β, κ and γ caseins and ranging from 19–33 kDa.
Casein subunits associate in solution forming complexes and ordered aggregates
of micelles. Sequence homologies between caseins αS1 and αS2 are about
22%. One study of 58 children with CMA showed that 85% possessed IgE
against each casein with different IgE levels: αS1 > β >> αS2 = κ (Bernard
et al., 1998). Casein is one of the major allergens responsible for CMA and
has been shown to play an important role in persistent allergy. Using 96
overlapping synthesized decapeptides representing the entire length of
αS1-casein, Chatchatee et al. (2001) identified two amino acid regions (AA
69–78 and AA 173–194) recognized by the majority of patients over nine
years of age with persistent CMA. None of children under three years of age
who are likely to outgrow CMA recognized these epitopes.
β-lactoglobulin (Bos d 5) is the most abundant protein from the whey
fraction. It is a protein of 18 kDa belonging to the lipocalin family (which
bind retinol). The amino acid sequence of bovine β-lactoglobulin shows
90% homology with ovine β-lactoglobulin. This allergen is not present in
human milk, but in one study was recognized by approximately 60% of 92
CMA patients (Wal et al., 1995).
α-lactalbumin (Bos d 4) is composed of two variants of 14.2 and 13 kDa.
Ιt belongs to the family of glycosyl hydrolase (lysozyme c superfamily) and
shows 95% sequence homology with α-lactalbumin from goat and sheep and
78% with α-lactalbumin from human milk. α-lactalbumin also shows 46%
sequence homology with egg lysozyme. Frequency of sensitisation to
α-lactalbumin was reported to range from 0–80% of patients (Besler et al.,
2002a).
Bovine serum albumin (Bos d 6) is a 67 kDa molecule containing nine
disulfide bridges. It is the main plasma protein and binds water, ions, fatty
acids and hormones and regulates osmotic blood pressure. Sequence homology
with mammalian serum albumins ranges from 74–79%. Recognition frequency
was reported as ranging from 0% in six children using sodium dodecyl
sulphate polyacrylamide gel electrophoresis (SDS-PAGE) and IgE
immunoblotting (Restani et al., 1999) to 75% with 21 children using crossed
radio-immunoelectrophoresis (CRIE) (Høst et al., 1992). Bovine
immunoglobulins and lactoferrin have been described as minor allergens.
Maillard reaction products, which are glycan-protein conjugates, may also
act as allergens (Wal, 1998, 2001).
13.2.3 Hydrolyzed infant formulae

Breast-feeding is certainly the best way to prevent children from becoming
Table 13.1 Cow’s milk allergens
IUIS Relative kDa pl Sequence Structure Function/

nomenclature content database (post-traductional family
SwissProt modifications)
Whey 20%
α-lactalbumin Bos d 4 5% 14.2 4.8 P00711 4 disulfide bonds Glycosyl
glycosylated hydrolase
β-lactoglobulin Bos d 5 10% 18.3 5.3 P02754 2 disulfide bonds, Lipocalin
dimer, not Retinol
present in binding
human milk protein
(RBP)
Bovine serum Bos d 6 1% 67 4.9–5.1 P02769 9 disulfide bonds Plasma
albumin protein
Immunoglobulins Bos d 7 3% 160 6–8
Lactoferrin < 1% 80 8.7 P24627 Iron
transport
Caseins Bos d 8 80%
αS1-casein 32% 23.6 4.9–5 P02662 Phosphorylated Calcium
transport
αS2-casein 10% 25.2 5.2–5.4 P02663 Phosphorylated +
1 disulfide bond
β-casein 28% 24.0 5.1–5.4 P02666 Phosphorylated
and glycosylated
κ-casein 10% 19.0 5.4–5.6 P02668 Phosphorylated + Miscelles
1 disulfide bond stability
γ-casein
sensitized to cows’ milk proteins. However, cows’ milk proteins have been
demonstrated to be present in breast milk (Axelsson et al., 1986). An increasing
population of infants receives hydrolyzed cows’ milk derived formulae for
the prevention and treatment of CMA. However, for primary prevention in
non-sensitized infants and protection from CMA, partially hydrolyzed
hypoallergenic formulae can be recommended. These formulae have reduced
allergenic activity (Restani et al., 1996).
Evaluation of the residual antigenicity and allergenicity of cows’ milk
substitutes was performed using several in vitro tests (Docena et al., 2002).
Immunoenzymatic methods such as enzyme-linked immunosorbent assay
(ELISA) using anti-cows’ milk protein antibodies were used to detect the
presence of residual antigenic epitopes. Direct EAST (enzyme allergosorbent
test) enables the detection of a potential decrease of IgE-binding to peptides
from moderately and extensively hydrolyzed milk preparations. SDS-PAGE
followed by IgE-immunoblotting using sera from CMA patients or cows’
milk protein-specific IgG antibodies showed that extensively hydrolyzed
preparations do not contain residual components from cows’ milk proteins.
While extensively hydrolyzed milk substitutes produced from whey or caseins
were reported to be safe for some CMA consumers (Halken et al., 1993;
Martin-Esteban et al., 1998), adverse reactions were also reported (Ragno et
al., 1993; Nilsson et al., 1999; Sotto et al., 1999; Carroccio et al., 2000). In
these cases of extreme sensitivity, an elemental amino acid formula should
be used.
13.2.4 Milk protein denaturation

However heat stable milk proteins are, heat treatments cause modifications
in their structure. Heat destroys conformational epitopes and can thus reduce
the allergenicity of some proteins. Boiling milk for ten minutes eliminates
skin prick test reactivity to bovine serum albumin and β-lactoglobulin, whereas
caseins were found to be heat stable after testing in eight CMA patients
(Norgaard et al., 1996). When using CRIE, Gjesing et al. (1986) showed a
reduction of IgE binding to α-lactalbumin (about 50%), to the casein fraction
(> 66%) and abolishment of IgE binding to β-lactoglobulin, bovine serum
albumin and bovine immunoglobulins after ten minutes of boiling. γ-irradiation
was shown to reduce IgE reactivity to α-lactalbumin and β-lactoglobulin
(Lee et al., 2001).
13.2.5 Hidden milk allergens

Cows’ milk proteins have often been described as hidden (undeclared or
contaminated) in food products (Gern et al., 1991; Laoprasert et al., 1998;
Giovannacci et al., 2004) and even in human milk due to ingestion of cows’
milk by the mother. By radioimmunoassay, detectable amounts of β-
lactoglobulin (from 5–800 ng/mL) were found in 93 out of 232 (40%) human
milk samples (Axelsson et al., 1986). On the other hand, Restani et al.
(2000) did not detect any β-lactoglobulin or casein in human milk from
breast-feeding mothers using SDS-PAGE, immunoblotting, and β-lactoglobulin
or caseins-specific monoclonal antibodies. However, monoclonal antibodies
might not be able to detect the proteins due to proteolytic changes.
Using RAST inhibition, α-lactalbumin was found to contaminate food-
grade lactose at levels of 1–5 ppm (Frémont et al., 1996). A 30 year old
woman with CMA and without fish allergy experienced anaphylaxis after
ingestion of reconstituted salmon containing undeclared casein, which was
used as a gelation agent combined with microbial transglutaminase, which
serves to cross-link casein to meat proteins for restructuring (Koppelman et
al., 1999).
Cows’ milk proteins can be found in other types of products. Ylitalo et al.
(1999) detected the presence of cows’ milk casein in 13 of 30 commercial
natural latex glove brands by rocket immunoelectrophoresis (RIE) and RAST
inhibition with casein-specific IgE from pooled human sera. Caseins can
also be an aero-allergen: a case of occupational asthma and rhinoconjunctivitis
has been described in a chocolate candy worker in contact with dried cows’
milk. α-lactalbumin was implicated as the cause of the sensitization (Bernaola
et al., 1994). Mäkinen-Kiljunen and Mussalo-Rauhamaa (2002) showed the
presence of caseins in 90 out of 91 house dust samples. The median casein
level was 30 μg/g (range 3–3300 μg/g), which was higher than those of pets
or mites (0.11 μg/g of Fel d 1 and < 0.01 μg/g of Der p 1 or Der f 1). The
authors explain these findings by the use of casein in indoor plasters to
improve their handling properties in practice since the 1960s. Some paints
can also contain casein proteins. In one case, inhalation of milk proteins was
reported to cause death (Bosetti et al., 1997).
13.2.6 Cross-reactions
Cross-reactions between ovine (sheep or goat) and bovine milk have been
studied by several authors (Dean et al., 1993; Sabbah et al., 1997; Bellioni-
Businco et al., 1999; Restani et al., 1999). People allergic to cows’ milk
proteins will often have serum IgE to goat or sheep’s milk proteins (Besler
et al., 2002b,c). β-lactoglobulin and caseins have been described as the main
cross-reacting allergens. Comparison between milk from the different ruminant
species reveals high homologies in both protein composition and structure,
particularly for the four caseins (αS1, αS2, β and κ) that are found in cows’,
sheep’s and goat’s milk. Moreover, important sequence homologies are
observed. Using CMA patient sera by ELISA with purified caseins, Bernard
et al. (1999) showed that important levels of cross-reactivity occur between
bovine, caprine and ovine caseins.
In a study on cows’ milk and human α-lactalbumin, proteins showing a
high degree of sequence homology demonstrated IgE cross-reactivity by
ELISA inhibition (Maynard et al., 1999). The authors suggested low clinical
relevance of these cross-reactive IgE. Other reports describe specific allergy

to ovine milk with good tolerance to cows’ milk (Wüthrich and Johansson,
1995; Umpierrez et al., 1999). Businco et al. (2000) investigated, by in vitro
and in vivo methods, the allergenicity of mare’s milk in a population of 25
children with severe CMA. Only two patients showed positive skin prick
tests to mare’s milk and one reacted to it in double-blind placebo-controlled
food challenge (DBPCFC). Therefore, in some cases, CMA patients may be
able to ingest mare’s milk.
13.3 Egg
13.3.1 Egg allergy
The prevalence of egg allergy, as studied by skin prick tests, was reported to
be about 3.2% in a group of 620 Australian children at risk of atopy (Hill et
al., 1999), 53% by labial food challenge in a population of 142 food-allergic
children (Rancé and Dutau, 1997), and 0.4% in an unselected population of
502 adults (Gislason et al., 1999). Cumulative reactive doses were reported
as ranging from 0.13 mg (raw whole egg) to 200 mg of proteins (dried whole
egg) (Taylor et al., 2002). A recent report (Morisset et al., 2003) showed that
a cumulative reactive dose inferior or equal to 65 mg was observed in 16%
of egg allergic patients with 125 oral challenges performed. The lowest
reactive dose was observed at less than 2 mg of egg. The conclusion of the
study was that the detection tests for egg should ensure a sensitivity of at
least 10 ppm taken from minimal quantities as a 95% guarantee for the
safety of patients allergic to egg on the basis of consumption of 100 g of
food.
13.3.2 Egg allergens

Ovomucoid (Gal d 1), is the dominant allergen in hen’s egg. It is a glycoprotein
with a molecular weight of 28 kDa and a pI of 4.1 exhibiting trypsin inhibitor
activity (see Table 13.2). One study (Bernhisel-Broadbent et al., 1994) suggested
that the use of commercially purified ovalbumin led to the impression that
ovalbumin was the major allergen from egg white, because ovalbumin was
contaminated with less than 1% of ovomucoid, resulting in false positive
results. The recognition frequencies reported in this study were omucoid
89%, ovalbumin 78% and lysozyme 61% by skin prick test in 18 egg-allergic
children. Urisu et al. (1997) classified egg white allergen IgE binding prevalence
in decreasing order: ovomucoid > ovalbumin >> ovotransferrin and lysozyme
by RAST using 72 egg allergic-patients.
Ovomucoid is a highly glycosylated protein comprising 186 amino acids
arranged in three tandem domains (Gal d 1.1, 1.2, and 1.3). The three ovomucoid
domains were isolated and evaluated with sera from egg-allergic patients to
determine B-cell domain specificity, B-cell epitopes and the relative importance
Table 13.2 Egg allergens
IUIS Relative kDa pl Sequence Structure Function/

nomenclature content database (post-traductional family
SwissProt modifications)
Egg white
Ovomucoid Gal d 1 11 28 4.1 P01005 glycoprotein trypsin (serine
protease)
inhibitor
Ovalbumin Gal d 2 54 43 4.5 P01012 phosphoglyco- serpin
protein
Ovotransferrin Gal d 3 12 77.7 6.0 P02789 glycoprotein (15 iron
disulfide bonds) transport
Lysozyme Gal d 4 3.4 14.3 10.7 P00698 (4 disulfide bonds) glycosidase
bacteriolytic
function
Ovomucin 1.5 0.2–8 106
Egg yolk
α-livetin Gal d 5 67 P19121 glycoprotein (17 serum
disulfide bonds) albumin
of linear and conformational structures and carbohydrate chains to B-cell

epitopes (Cooke and Sampson, 1997). There was significantly more IgE
activity against the second ovomucoid domain (median percentage of
ovomucoid-specific IgE: Gal d 1.2, 40%; Gal d 1.1, 23%; Gal d 1.3, 26%).
Five IgE and seven IgG binding regions were identified and IgE antibodies
binding to reduced ovomucoid and IgG binding to oxidized ovomucoid were
significantly decreased as compared with that to native ovomucoid (28 and
69%, respectively). The authors showed that conformational B cell epitopes
play a significant role in ovomucoid allergenicity and that carbohydrate
moieties have a minor effect on allergenicity. In this paper, the authors suggested
that persistent egg allergy is related to reactivity to linear sequences and that
egg allergy that is outgrown is associated with conformational epitopes.
Ovalbumin (Gal d 2) is a monomeric phosphoglycoprotein with a molecular
mass of 43–45 kDa and a pI of 4.5. As noted above, at first it was thought to
be the major egg white allergen, but this has been shown subsequently to be
less frequently recognized by IgE from egg-allergic patients than ovomucoid.
Ovotransferrin (conalbumin, Gal d 3) has a molecular weight of 77 kDa and
a pI of 6.0, and has iron-binding properties. Its frequency of recognition in
egg-allergic patients was reported in one study to be 22% (Djurtoft et al.,
1991). Lysozyme (Gal d 4) is a 14.3 kDa protein with a basic pI of 10.7.
Lysozyme (1,4-β-N-acetylmuraminidase C) has a bacteriolytic function and
hydrolyzes the peptidoglycan polymer of prokaryotic cell walls. The recognition
prevalence for lysozyme is reported to range from 6–67% (Djurtoft et al.,
1991; Yamada et al., 1993; Frémont et al., 1997).
Specific IgE to egg yolk were first reported by Carrillo Diaz et al. (1986).
Alpha-livetin (Gal d 5) is the chicken egg yolk protein implicated in ‘bird-
egg’ syndrome. This syndrome occurs in patients in contact with birds and
showing respiratory symptoms. This respiratory allergy precedes in all cases
the onset of food allergy to eggs. Respiratory symptoms in the presence of
certain fowl by sensitization to α-livetin from egg yolk is related to food
allergy to egg (de Blay et al., 1994; Szepfalusi et al., 1994).
13.3.3 Stability of egg allergens

In contrast to milk proteins, egg proteins are very heat-labile. When cooked,
egg yolk and white coagulate. Eigenmann (2000) described two cases of
anaphylactic reactions after eating raw eggs whereas patients tolerated cooked
eggs. Heating egg white for ten minutes at 90 °C significantly decreased
RAST for 50% of 16 egg-allergic patients (Anet et al., 1985). In a population
of 38 subjects with positive response to freeze-dried egg white, 21 had
negative challenge responses to heated egg white (Urisu et al., 1997). In the
same report, ovomucoid was shown to be a heat stable allergen. Oral challenges
in 17 (94%) of egg-allergic patients with positive challenges to heated egg
white had negative challenges to ovomucoid-depleted egg white.
13.4 Types of detection methods

13.4.1 Diffusion-in-gel methods
Diffusion-in-gel methods are cheap and easily performed but are not very
sensitive, especially in processed or heat-treated food products; they are being
replaced by immunoenzymic techniques. Langeland (1982) used CRIE and
Hoffman (1983) utilized RIE to study egg white allergens. Results showed
that ovalbumin and ovomucoid were strong allergens and conalbumin
(ovotransferrin) was a less important allergen, while lysozyme was shown to
be a weak allergen. Sajdok et al. (1990) determined the egg and egg white
content in food products using radial immunodiffusion and RIE. Ylitalo et al.
(1999) detected cows’ milk casein in natural rubber latex gloves using RIE.
The NMKL (Nordic Committee on Food Analysis) performed a multicenter
collaborative study on food allergen detection using double immunodiffusion
(Malmheden-Yman et al., 2002). The results showed detection limits of 300 ppm
casein in sausages and 30 ppm ovalbumin in pasta. However, technical expertise
on the part of the laboratory enhances the possibility of obtaining correct
results. The level of detection varies somewhat with antigen and with the titer
of antiserum but lies normally at mg levels per 100 g food samples. This level
of detection is generally not low enough, as it is usually recommended that
methods have detection limits of 1–10 ppm for allergenic residues (Moneret-
Vautrin et al., 2003; Morisset et al., 2003). Using radial immunodiffusion,
Bertrand-Harb et al. (2003) followed the concentrations of β-lactoglobulin
and α-lactalbumin during yoghurt fermentation. A RIE method was developed
using commercially available anti-casein antiserum (Moen et al., 2003).
Quantification of unknown amounts of casein in foodstuffs resulted in a linear
correlation between the height of the precipitates and the applied amount of
casein. In this study, RIE had a detection limit for casein of 10 ppm and was
described as being less expensive than commercial ELISA kits. Nevertheless,
RIE is time- and antibody-consuming and requires trained people specialized
in this field. Furthermore, ELISA kits are more practical for the food industry
than RIE or CRIE.
13.4.2 RAST and EAST inhibition

RAST and EAST (see Chapter 4) are based on the binding of specific IgE to
immobilized allergen molecules. RAST or EAST methods have been used for
egg or milk allergens (Walsh et al., 1987; Adams et al., 1991; Rugo et al., 1992;
Dean et al., 1993; Frémont et al., 1996; Ylitalo et al.,1999). Frémont et al. (1996)
reported a detection limit of 1 ppm for α-lactalbumin. These techniques using
egg-allergic serum are more suitable for clinicians than for the food industry.
13.4.3 Bioactivity of allergens

In order to quantify the allergenic activity of diagnostic allergen extracts, the
techniques of RAST or EAST are currently used, as discussed in Chapter 4.
When the biological activity of a diagnostic allergen extract has to be

determined, the capacity of the extract to degranulate IgE-sensitized basophils
or mast cells reflects the presence of allergenic molecules inducing important
cellular changes and finally degranulation (as discussed in Chapter 1). This
degranulation is quantified by measuring the mediators (histamine, serotonine)
or enzymes (β-hexosaminidase) specifically released by the cells. A choice
of cellular systems for measuring such bioactivity of diagnostic allergen
extracts is available, and human as well as animal models have been described
(May, 1976; Norgaard et al., 1992; Prouvost-Danon et al., 1994). Nowadays,
use of human basophils or mast cells is reserved only for diagnostic or
research purposes. Animal mast cells are triggering units allowing the study
of allergenic bioactivity, as they may be coated with allergen-specific IgE,
after either active or passive sensitization of the cells with the immunoglobulins.
In practice, serial ten-fold dilutions of an allergen extract are incubated in
the presence of sensitized cells. Bell-shaped curves are obtained when the
percentage of total mediator or enzyme release is plotted against the allergen
concentration. This indicates that the release observed is the consequence of
membrane-bound IgE triggering and not of toxicity on the cells of another
component present in the extract. The lowest concentration of the allergen
extract inducing the release of a well-defined percentage of the total content
of enzyme or mediators represents the bioactivity of the extract.
The newly described transfectants of RBL (rat basophilic leukocytes)
cells presenting chimeric receptors fixing human IgE (Marchand et al., 2003;
Taudou et al., 1993) is a technique suitable for the detection of allergenic
molecules encountered in an allergic pathology. The cells can be sensitized
with a diluted pool of food-allergic sera and afterwards triggered with the
allergen extract being studied. The latter techniques may be used for the
determination of the bioactivity of food allergens, such as allergenic proteins
present in milk or egg.
13.4.4 Dot blot technique

The dot blot technique initially described by Hawkes et al. (1982) was first
used by Janssen et al. (1987) to detect egg and milk proteins in heat-treated
meat products. The principle of this technique is to bind the antigen, from a
complex mixture supposed to contain it, onto a membrane which is further
incubated with a specific antibody enzymatically labeled. Finally, the membrane
is incubated with a substrate, the product of which precipitates on the site of
enzymic reaction. This precipitation allows the visualization of the presence
of the antigen. The method was improved using cyanogen bromide-activated
(CNBr-activated) nitrocellulose membranes (Demeulemester et al., 1991)
which have high binding capacities (Demeulemester et al., 1987; Becker,
1989).
13.4.5 Electrophoretic methods, blotting methods, immunodetection

These techniques, based on the separation of samples in an electric field
followed by immunodetection using either human allergic sera or animal
antibodies, are discussed at length in Chapter 5. They are able to separate
and to reveal the proteic components in raw or slightly cooked food products,
but their use is not efficient enough for highly cooked and sterilized products
because denaturation and aggregation of molecules disturb their separation
(King, 1984; Hitchcock and Crimes, 1985; Bonnefoi et al., 1986).
Electrophoretic transfer after polyacrylamide gel electrophoresis

The protein unfolding caused by the use of SDS in PAGE may lead to the
loss of conformational epitopes with sequential epitopes being preserved;
this is the main disadvantage of this technique which was used to study egg
and milk antigens and allergens (Quirce et al., 1998; Añibarro et al., 2000;
Kim et al., 2002 ; Rupa and Mine, 2003).
Using this technique, Leduc et al. (1997) analyzed a few milk- and egg-
based ingredients. These ingredients were chosen because they are commonly
used in processed foods for their very good protein-binding properties. Results
obtained with milk-based ingredients are presented in Fig. 13.1. Silver staining
revealed five major proteins in skimmed milk: lactoferrin, bovine serum albumin,
caseins, β-lactoglobulin and α-lactalbumin. Bovine serum albumin,
β-lactoglobulin and α-lactalbumin are the major constituents of whey, while
caseins and lactoferrin compose the other proteins. The polyclonal anti-milk
and anti-whey rabbit antisera used in the study had a wide specificity: they
recognized all proteins revealed by silver staining, with α-lactalbumin only
slightly recognized. These results confirm that the casein fractions can contain
1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4
kDa
148
Lf
BSA 60
42
Cas
30
22
β-Lg
17
α-Lac
A B C D E
Fig. 13.1 SDS-PAGE of milk products (Leduc et al., 1997). A: silver staining; B:
antigens revealed with rabbit antimilk antiseruml C: antigens revealed with rabbit
antiwhey antiserum; D: allergen revealed with patient’s serum H[15] × M; E: allergen
revealed with patient’s serum H[17] × M; 1: proteic binder milk by-product; 2:
skimmed milk powder; 3: whey; 4: sodium caseinate; Lf: lactoferrin; BSA: bovine
serum albumin; Cas: caseins; β-Lg: β-lactoglobulin; α-Lac: α-Lactalbumin.
whey proteins, and that the whey fractions contain caseins, as also previously
observed by Janssen et al. (1987). The use of two milk-allergic patient sera
showed that their IgE recognized allergens in the four milk ingredients studied.
For these two sera, major allergens were caseins and, to a lesser extent, lactoferrin.
Blotting transfer after isoelectric focusing in agarose gels

In this method, proteins are separated in agarose gels by isoelectric focusing
(IEF) as described by Peltre et al. (1982) and are then probed using
immunodetection techniques. In contrast to SDS-PAGE, this technique
preserves conformational epitopes. Demeulemester et al. (1997) used this
technique to screen milk and egg white allergens. Briefly, extracts from milk
and egg ingredients were focused in agarose gels and then transferred onto
CNBr-activated nitrocellulose membranes (Desvaux et al., 1989). After
incubation of the membranes with milk- or egg-allergic patient sera, IgE-
binding was revealed by alkaline phosphatase labeled anti-human IgE and
chromogenic substrate. Results showed that IgE recognized caseins in six
out of 17 tested sera. β-lactoglobulin and α-lactalbumin, which are not clearly
separated during IEF because their pI are very close and because they have
scores of isoforms (Holen and Elsayed, 1990), bound IgE from seven sera.
With regard to egg white, lysozyme was slightly recognized by five out of 18
sera, ovotransferrin was substantially recognized by IgE in seven sera ovalbumin
by five sera and ovomucoid by 11 sera.
A multiple successive immunoprinting after IEF in agarose gel was described
by Desvaux et al. (1990). A protein extract is focused in an agarose gel, and
then subsequently transferred to different CNBr-activated nitrocellulose
membranes. Hence, different prints which are copies of the same initial gel
are obtained; therefore, they have the same electrophoretic pattern except for
molecules present at very low amounts. Leduc et al. (1999) performed this
technique on raw, heat-pasteurized or heat-sterilized meat paste extracts with
or without egg white (see Fig. 13.2). The first print was used to reveal egg
white antigens using a rabbit anti-egg white antiserum (see Fig. 13.2a). The
second and third prints were each incubated with an egg-allergic serum.
Results showed that egg white antigens and allergens can be visualized in
raw and pasteurized products using this technique. The IgE of patient 4
recognized ovotransferrin as allergen (see Fig. 13.2b); that of patient 1
recognized ovalbumin and another acidic molecule close to ovalbumin, probably
ovomucoid (see Fig. 10.2c). None of the egg white antigens or allergens
were detected in heat-sterilized products.
2D-electrophoresis
Two-dimensional (2D)-electrophoresis, initially described by O’Farrell (1975)
combines IEF and SDS-PAGE. Holen and Elsayed (1990) used this technique
to characterize egg white allergens and showed that in their population of
egg-allergic patients, ovalbumin, ovomucoid, ovotransferrin and lysozyme
were the major egg white allergens. Brodard et al. (1995) used this method
Paste C Paste E
Paste C Paste E R P S R P S
R P S R P S E3
E3
pH 9
Lys pH 9
AP
Ot Ot
Ova
pH 3
pH 3
(a) (b)
Paste C Paste E
R P S R P S
E3
pH 9
Ova
pH 3
(c)
Fig. 13.2 Immunoblots from IEF separation of meat extracts (Leduc et al., 1999). (a)
egg white antigens revealed with rabbit antiserum; (b) allergens revealed with human
serum H [4] × M; (c) allergens revealed with human serum H [1] × M; paste C:
negative control meat product; paste E: meat product containing egg white; R: raw; P:
pasteurized; S: sterilized; E3: frozen pasteurized egg white; Lys: lysozyme; AP:
alkaline phosphatase; Ot: ovotoransferrin; Ova: ovalbumin; 䉰: sample application.
to describe a 2D-map of milk allergens detected by IgE from milk-allergic

patients.
13.4.6 ELISA
Many ELISA methods have been described for detection of milk proteins
(Gern et al., 1991; Mäkinen-Kiljunen and Palosuo, 1992; Mariarger et al.,
1994; Venien et al., 1997; Hefle and Lambrecht, 2004) and egg white proteins
(Breton et al., 1988; Rauch et al., 1990; Turin and Bonomi, 1994;
Demeulemester and Guizard, 1996; Leduc et al., 1999; Yeung et al., 2000;
Hefle et al., 2001; Baumgartner et al., 2002; Immer et al., 2003). Only one
ELISA was reported for yolk proteins (Pressi et al., 1994). The specific
antibodies used in these ELISA methods are IgG; thus, they do not detect the
allergenic epitopes, but the antigenic epitopes. Some commercial kits are
available. A list of suppliers is available in Table 6.1 in Chapter 6. Currently
available ELISA methods are generally based on the detection of caseins
and/or β-lactoglobulin, while commercial egg ELISA kits are usually based
on the detection of ovalbumin and/or ovomucoid.
Limits of detection in processed food products depend on various parameters,
such as fat content, severity of heat processing, muscle origin, state of meat
maturation, etc. Hence, detection limits might be different from one product
to another. From a theoretical point of view, ELISA methods are quantitative
but, for the same reasons as the detection limit, results can be only
semi-quantitative or qualitative unless validated appropriately.
In a collaborative study performed by two independent laboratories (Immer
et al., 2003), 43 food products were assayed using a commercial ELISA kit
for egg white detection (see Table 13.3). Of the 43, 32 had labels which
mentioned ‘egg’, ‘egg white’ or ‘egg yolk’, egg white was detected in 27 of
these 32 products and was not detected in five (505, 512, 515, 539 and 543).
Among the 11 products whose labels did not mention egg, egg white was
detected in one product (528) at 5 ppm. The discrepancy between egg levels
found and the levels expected could be due to a number of reasons: (i) the
products do not contain egg; (ii) they contain less egg than mentioned on the
labels or less than the detection limits of the ELISA kit; (iii) the egg-based
ingredients used or the final products were submitted to high levels of heat
or other processing that led to the denaturation of egg molecules and to the
lack of detection by the commercial kit.
Hefle and Lambrecht (2004) developed an ELISA for casein determination
with a detection limit of less than 0.5 ppm. They analyzed retail non-milk-
containing foods such as chocolates, sorbets, fruit drinks and fruit juices and
found in many of them undeclared casein from 0.5 to more than 10 000 ppm.
They also analyzed products associated with milk-allergic consumer complaints:
casein was detected in all of them (5 500–44 500 ppm).
Detection of bovine milk proteins is also useful in quality control to
differentiate milk from other species. ELISA methods were published to
specifically detect cows’ milk in goats’ milk (Castro et al., 1992) or in ovine
milk (Garcia et al., 1991; Rodriguez et al., 1993) as cows’ milk is cheaper
than ovine and caprine milk.
13.4.7 DNA methods

As discussed in Chapter 7, polymerase chain reaction (PCR) methods can be
used for the detection of food allergens (Holzhauser et al., 2000; Dahinden
Table 13.3 Detection of egg white proteins in food products using ELISA (Immer et al.,
2003). ELISA method used was the Ridascreen® Egg Protein kit (R-Biopharm, Darmstadt,
Germany). 501–505: cheese and cheese delicatessens; 506 and 507: rice puddings; 508 and
509: custards; 510 and 511: vegetables purees; 516–520: vacuum-packed meat products; 521–
524: fresh pasta; 525–533: dried pasta; 537–541: biscuits; 542 and 543: breads. *: aberrant
results. A and B are the laboratories where assays were performed. Some of the foods are
typical French foods.
Product Designation Animal species animal Determination of egg

No proteic binders (label) white (ppm)
A B
501 Roquefort Milk (ewe) <2 <2

502 Cheese delicatessens Milk (ewe, cow)/ > 5 400 35 000
egg white
503 Cheese delicatessens Milk (cow)/egg white > 5 400 24 300
504 Cheese delicatessens Milk (cow) <2 <2
505 Cheese delicatessens Milk (cow)/egg white <2 <2
506 Rice pudding Milk/whole egg (7 %) 5.4 4
507 Rice pudding Milk <2 <2
508 Caramel custard Milk/whole egg (19 %) 83 64
509 French custard Milk/whole egg 8.8 6
510 Houmous – <2 <2
(chickpea purée)
511 Aubergine salad Whole egg > 540 830
512 Fish rillettes Tuna/egg white/milk <2 *
513 Surimi Fish/crab/egg white 450 370
514 Shortcrust pastry – <2 <2
515 Shortcrust pastry Egg <2 <2
516 Garlic sausage Pork/egg white 480 440
517 Cocktail sausages Turkey <2 <2
518 Pike quenelles Whole egg/pike/milk 91 72
519 Liver pâté Pork/milk <2 <2
520 Liver mousse Pork/milk/egg 16 28
521 Gnocchi – <2 <2
522 Fresh tagliatellas Whole egg (25 %) 35 21
523 Fresh tagliatellas Whole egg (14 %) 660 500
524 Fresh tagliatellas Whole egg (11 %) 3.8 3
525 Nids d’Alsace Whole egg (30 %) > 5400 19 200
526 Pâtes d’Alsace Whole egg (24 %) > 5400 37 100
527 Crozets de Savoie Whole egg (15 %) > 5400 5700
528 Farfalle – 5.4 5
529 Papillons d’Alsace Whole egg (30 %) > 5 400 32 500
530 Farandelles de Savoie Whole egg (21 %) > 5 400 25 100
531 Macaroni Whole egg (13.5 %) > 5 400 10 300
532 Spaghetti – <2 <2
533 Spaghetti Whole egg (19.36 %) ≥ 5 400 4,100
534 Mayonnaise Whole egg > 5 400 *
535 Mayonnaise Egg yolk 1680 850
536 Mayonnaise Whole egg > 5 400 8700
537 Palets Saint Michel Whole egg 82 74
538 Shortbread Milk/whole egg 190 122
539 Galettes bretonnes Whole egg (3.7 %), milk <2 <2
540 Biscuits cuiller Whole egg (32 %), milk 1620 720
541 Biscuits Whole egg (2 %), milk 380 200
542 Bread – <2 <2
543 Brioche Milk/egg <2 <2
et al., 2001; Hird et al., 2003). As milk contains DNA from epithelial cells
(Lipkin et al., 1993), López-Calleja et al. (2004) described a PCR method to
detect cows’ milk in sheep and goat’s milk. Eggs contain DNA which can be
found in commercial egg white ingredients. DNA methods could therefore
be used for detection of milk and egg. But in practice, they are not adapted
for detection of egg or milk allergens because the presence of chicken meat
and beef, respectively, could give false positive responses.
13.4.8 Biosensors
Surface plasmon resonance (SPR) technology (Biacore™, Biacore AB,
Uppsala, Sweden) is discussed in Chapter 9. Allergen detection by SPR
immunoassay is based on the measurement of total sample binding to
allergen-specific antibodies coupled to the sensor surface. Purified antibodies
raised against ovomucoid (egg) and β-lactoglobulin (milk) were covalently
bound to a gold-dextran chip (Jonsson and Hellenäs, 2001). Calibration
curves established with purified allergens showed a detection limit of
approximately 10 ng/mL. No data was given for detection limits in solid
food matrices. However, non-specific binding can cause false positive results
when allergens are analyzed in untreated samples and total sample binding
is used for quantification (Mohammed et al., 2001).
13.5 Future trends

Some quick test kits and qualitative methods for milk and egg are on the
market now, and some kit companies plan on developing more rapid methods
in the form of lateral flow devices (see Chapter 10). Commercial quantitative
tests are already on the market but, as quantitation requires more time, these
methods are not as ‘quick’ as the qualitative methods. Use of these techniques
by the food industry has allowed the monitoring of egg and milk residues in
the manufacturing environment, making immunoassays a useful tool for
sanitation assessment and verification/validation. While to date, threshold
levels for milk and egg are not known with certainty (see Chapter 1), as far
as the appropriate detection limits are concerned, the methods used by kit
companies for milk and egg appear to be adequate for allergic consumer
protection.
Quantitative tests are unreliable unless appropriately validated, as the
precise nature of ingredients used and technological treatments undergone
are sometimes hard to determine. The main obstacle to the quantification in
some food matrices is low extraction yield; this may be due to the denaturation
or aggregation of proteins by processing and to food matrix effects (influence
of lipids, Maillard reactions, etc.).
The recent development of array technologies is promising for rapid and
multiparametric analysis for allergen detection in food products in the future
(Weinberger et al., 2000; Bashir et al., 2001; Lin et al., 2001; Zlatanova and
Mirzabekov, 2001; Talapatra et al., 2002; Moreno-Bondi et al., 2003) but, at
the present time, these are impractical for use in most manufacturing
environments.
Quality assurance (ISO, 1999) has become widely used in laboratories.
Good practice guidelines, acknowledged or normalized methods, collaborative
studies, reference materials and accreditation of analysis by independent
organisations are very useful tools for reliable results.
Development of milk and egg with low allergenic properties may be
useful (Besler and Mine, 1999). However, they present certain disadvantages:
the low allergenic products are quite expensive and do not always protect
from possible reactions. Indeed, low allergenic products, prepared to decrease
the number of certain allergenic proteins/peptides, may fail to protect people
allergic to other proteins/peptides.
13.6 Acknowledgements
The authors thank Dominique Bornhauser, Gilles-Antoine Couëtte (Allerbio),
Laurence Guérin (Allerbio), Cécile Guizard (CTSCCV), Pravina Nallatamby
(CTSCCV), Céline Tattegrain and Anne Weyer (Pasteur Institute, Paris) for
helpful discussions.
13.7 References
Adams, S L, Barnett, D, Walsh, B J, Pearce, R J, Hill, D J and Howden, M E (1991)
‘Human IgE-binding synthetic peptides of bovine beta-lactoglobulin and alpha-
lactalbumin. In vitro cross-reactivity of the allergens’, Immunol Cell Biol, 69, 191–
197.
Anet, J, Back, J F, Baker, R S, Barnett, D, Burley, R W and Howden, M E (1985)
‘Allergens in the white and yolk of hen’s egg. A study of IgE binding by egg proteins’,
Int Arch Allergy Appl Immunol, 77, 364–371.
Añibarro, B, Seoane, F J, Vila, C and Lombardero, M (2000) ‘Allergy to eggs from duck
and goose without sensitization to hen egg proteins’, J Allergy Clin Immunol, 105,
834–836.
Axelsson, I, Jakobsson, I, Lindberg, T and Benediktsson, B (1986) ‘Bovine beta-lactoglobulin
in the human milk. A longitudinal study during the whole lactation period’, Acta
Paediatr Scand, 75, 702–707.
Bashir, R, Gomez, R, Sarikaya, A, Ladisch, M R, Sturgis, J and Robinson, J P (2001)
‘Adsorption of avidin on microfabricated surfaces for protein biochip applications’,
Biotechnol Bioeng, 73, 324–328.
Baumgartner, S, Steiner, I, Kloiber, S, Hirmann, D, Krska, R and Yeung, J (2002) ‘Towards
the development of a dipstick immunoassay for the detection of trace amounts of egg
proteins in food’, Eur Food Res Technol, 214, 168–170.
Becker, W M (1989) ‘Reactivities of immunoglobulin E and immunoglobulin G subclasses
identified by isoelectric focusing-immunoprint in allergic patients’, Electrophoresis,
10, 633–639.
Bellioni-Businco, B, Paganelli, R, Lucenti, P, Giampetro, P G, Perborn, H and Businco,

L (1999) ‘Allergenicity of goat’s milk in children with cow’s milk allergy’, J Allergy
Clin Immunol, 103, 1191–1194.
Bernaola, G, Echechipia, S, Urrutia, I, Fernandez, E, Audicana, M and Fernandez de
Corres, L (1994) ‘Occupational asthma and rhinoconjunctivitis from inhalation of
dried cow’s milk cauzed by sensitization to alpha-lactalbumin’, Allergy, 49, 189–191.
Bernard, H, Créminon, C, Yvon, M and Wal, J M (1998) ‘Specificity of the human IgE
response to the different purified caseins in allergy to cow’s milk proteins’, Int Arch
Allergy Immunol, 115, 235–244.
Bernard, H, Créminon, C, Négroni, L, Peltre, G and Wal, J M (1999) ‘IgE cross-reactivity
with caseins from different species in humans allergic to cow’s milk’, Food & Agric
Immunol, 11, 101–111.
Bernhisel-Broadbent, J, Dintzis, H M, Dintzis, R Z and Sampson, H A (1994) ‘Allergenicity
and antigenicity of chicken egg ovomucoid (Gal d III) compared with ovalbumin (Gal
d I) in children with egg allergy and in mice’, J Allergy Clin Immunol, 93, 1047–1059.
Bertrand-Harb, C, Ivanova, I V, Dalgalarrondo, M and Haertlle, T (2003) ‘Evolution of β-
lactoglobulin and α-lactalbumin content during yoghurt fermentation’, Int Dairy J,
13, 39–45.
Besler, M and Mine, Y (1999) ‘The major allergen from hen’s egg white: ovomucoid (Gal
d 1)’, Internet Symposium on Food Allergens, 1, 137–146.
Besler, M, Eigenmann, P and Schwartz, R H (2002a) ‘Allergen data collection – update:
cow’s milk (Bos domesticus)’, Internet Symposium on Food Allergens, 4, 19–106.
Besler, M, Eigenmann, P and Schwartz, R H (2002b) ‘Allergen data collection: goat’s
milk (Capra spp.)’, Internet Symposium on Food Allergens, 4, 119–124.
Besler, M, Eigenmann, P and Schwartz, R H (2002c) ‘Allergen data collection: sheep’s
milk (Ovis spp.), Internet Symposium on Food Allergens, 4, 125–130.
Bonnefoi, M, Bénard, G and Labie, C (1986) ‘Gel electrophoresis: a qualitative method
for detection of duck and goose liver in canned foie gras’, J Food Sci, 51, 1362–1363.
Bosetti, M, Ispano, M and Rotondo, F (1997) ‘Anaphylaxis resulting in death after inhalation
of milk proteins’, Allergy, 37, 121.
Breton, C, Phan Thanh, L and Paraf, A (1988) ‘Immunochemical identification and
quantification of ovalbumin additive in canned mushrooms’, J Food Sci, 53, 226–230.
Brodard, V, Bernard, H, Wal, J M, David, B and Peltre, G (1995) ‘Two-dimensional
analysis of cow’s milk allergens with the IPG-DALT technique’, Adv Exp Med Biol,
371 B, 875–878.
Businco, L, Giampietro, P G, Lucenti, P, Lucaroni, F, Pini, C, Di Felice, G, Iacovacci, P,
Curadi, C and Orlandi, M (2000) ‘Allergenicity of mare’s milk in children with cow’s
milk allergy’, J Allergy Clin Immunol, 105, 1031–1034.
Carrillo Diaz, T, Cuevas Agustin, M, Moneo Goiri, I, Ibanez Sandin, M D and Urena
Vilardell, V (1986) ‘Detection of IgE specific for egg yolk by enzyme immunoassay,
description of a case’. Allergol Immunopathol (Madr), 14, 9–14.
Carroccio, A, Cavataio, F, Montalto, G, D’Amico, D, Alabrese, L and Iacono, G (2000)
‘Intolerance to hydrolysed cow’s milk proteins in infants: clinical characteristics and
dietary treatment’, Clin Exp Allergy, 30, 1597–1603.
Castro, C, Martin, R, Garcia, T, Rodriguez, E, Gonzalez, I, Sanz, B and Hernandez, P E
(1992) ‘Indirect enzyme-linked immunosorbent assay for detection of cow’s milk in
goat’s milk’, Food Agric Immunol, 4, 11–18.
Chatchatee, P, Järvinen, K M, Bardina, L, Beyer, K and Sampson, H A (2001) ‘Identification
of IgE- and IgG-binding epitopes on αS1-casein: differences in patients with persistent
and transient cow’s milk allergy’, J Allergy Clin Immunol, 107, 379–383.
Cooke, S K and Sampson, H A (1997) ‘Allergenic properties of ovomucoid in man’, J
Immunol, 159, 2026–2032.
Dahinden, I., von Büren, M and Lüthy, J (2001) ‘A quantitative competitive PCR system
to detect contamination of wheat, barley, or rye in gluten-free food for coeliac patients’,
Davies, R and Pepys, J (1976) ‘Egg allergy, influenza vaccine, and immunoglobulin E
antibody’, J Allergy Clin Immunol, 57, 373–383.
De Blay, F, Hoyet, C, Candolfi, E, Thierry, R and Pauli, G (1994) ‘Identification of alpha
livetin as a cross reacting allergen in a bird-egg syndrome’, Allergy Proc, 15, 77–78.
Dean, T P, Adler, B R, Ruge, F and Warner, J O (1993) ‘In vitro allergenicity of cows’
milk substitutes’, Clin Exp Allergy, 23, 205–210.
Demeulemester, C and Guizard, C (1996) ‘Détection des constituants protéiques par
ELISA (Enzyme-Linked ImmunoSorbent Assay)’, Bull Liaison CTSCCV, 6, 26–29.
Demeulemester, C, Peltre, G, Laurent, M, Panheleux, D and David, B (1987) ‘Cyanogen
bromide-activated nitrocellulose membranes: a new tool for immunoprint techniques’,
Demeulemester, C, Lajon, A, Abramowski, V, Martin, J L and Durand, P (1991) ‘Improved
ELISA and dot-blot methods for the detection of whey proteins in meat products’, J
Sci Food Agric, 56, 325–333.
Demeulemester, C, Leduc, V, Polack, B, Huneau, J F, Le Guern, L, Guizard, C and Peltre,
G (1997) ‘Antigènes et allergènes de lait, de blanc d’œuf, de soja. 2 – Dans les
produits carnés’, Ann Fals Exp Chim, 90, 235–246.
Desvaux, F X, Peltre, G and David, B (1989) ‘Characterization of grass pollen specific
IgE, IgA, IgM classes and IgG subclasses in allergic patients’, Int Arch Allergy Appl
Immunol, 89, 281–287.
Desvaux, F X, David, B and Peltre, G (1990) ‘Multiple successive immunoprinting: a fast
blotting technique of a single agarose isoelectric focusing gel’, Electrophoresis, 11,
37–41.
Djurtoft, R, Pedersen, H S, Aabin, B and Barkholt, V (1991) ‘Studies of food allergens:
soybean and egg proteins’, Adv Exp Med Biol, 289, 281–293.
Docena, G, Rozenfeld, P, Fernandez, R and Fossati, C A (2002) ‘Evaluation of the
residual antigenicity and allergenicity of cow’s milk substitutes by in vitro tests’,
Allergy, 57, 83–91.
Eigenmann, P A (2000) ‘Anaphylactic reactions to raw eggs after negative challenges
with cooked eggs’, J Allergy Clin Immunol, 105, 587–588.
Frémont, S, Kanny, G, Bieber, S, Nicolas, J P and Moneret-Vautrin, D A (1996) ‘Identification
of a masked allergen, alpha-lactalbumin, in baby-food cereal flour guaranteed free of
cow’s milk protein’, Allergy, 51, 749–754.
Frémont, S, Kanny, G, Nicolas, J P and Moneret-Vautrin, D A (1997) ‘Prevalence of
lysozyme sensitization in an egg-allergic population’, Allergy, 52, 224–228.
Garcia, T, Martin, R, Rodriguez, E, Azcona, J I, Sanz, B and Hernandez, P E (1991)
‘Detection of bovine milk in ovine milk by a sandwich enzyme-linked immunosorbent
assay (ELISA)’, J Food Prot, 54, 366–369.
Gern, J E, Yang, E, Evrard, H M and Sampson, H A (1991) ‘Allergic reactions to milk-
contaminated ‘nondairy’ products’, N Engl J Med, 324, 976–979.
Gislason, D, Björnsson, E, Gislason, T, Janson, C, Sjöberg, O, Elfman, L and Boman, G
(1999) ‘Sensitization to airborne and food allergens in Reykjavic (Iceland) and Uppsala
(Sweden) – a comparative study’, Allergy, 54, 1160–1167.
Giovannacci, I, Guizard, C, Carlier, M, Duval, V, Martin, J L and Demeulemester, C
(2004) ‘Species identification of meat products by ELISA’, Int J Food Sci Tech, 39,
863–867.
Gjesing, B, Osterballe, O, Schwartz, B, Wahn, U and Lowenstein, H (1986) ‘Allergen-
specific IgE antibodies against antigenic components in cow milk and milk substitutes’
Allergy, 41, 51–56.
Halken, S, Høst, A, Hansen, L G and Osterballe, O (1993) ‘Safety of a new, ultrafiltrated
whey hydrolysate formula in children with cow milk allergy: a clinical investigation,’
Pediatr Allergy Immunol, 4, 53–59.
Hawkes, R, Niday, E and Gordon, J (1982) ‘A dot-immunobinding assay for monoclonal
and other antibodies’, Anal Biochem, 119, 142–147.
Hefle, S L and Lambrecht, D M (2004) ‘Validated sandwich enzyme-linked immunosorbent

assay for casein and its application to retail and milk-allergic complaint foods’, J Food
Prot, 67, 1933–1938.
Hefle, S L and Taylor, S L (2002) ‘How much food is too much? Threshold doses for
allergenic foods’, Curr Allergy Asthma Rep, 2, 63–66.
Hefle, S L, Jeanniton, E and Taylor, S L (2001) ‘Development of a sandwich enzyme-
linked immunosorbent assay for the detection of egg residues in processed foods’, J
Food Prot, 64, 1812–1816.
Hill, D J, Hosking, C S, Zhie, C Y, Leung, R, Baratwidjaja, K, Iikura, Y, Iyngkaran, N,
Gonzalez-Andaya, A, Wah, L B and Hsieh, K H (1997) ‘The frequency of food allergy
in Australia and Asia’, Environ Toxicol Pharmacol, 4, 101–110.
Hill, D J, Hosking, C S and Heine, R G (1999) ‘Clinical spectrum of food allergy in
children in Australia and South-East Asia: identification and targets for treatment’,
Ann Med, 31, 272–281.
Hird, H, Lloyd, J, Goodier, R, Brown, J and Reece, P (2003) ‘Detection of peanut using
real-time polymerase chain reaction,’ Eur Food Res Technol, 217, 265–268.
Hitchcock, C H S and Crimes, A A (1985) ‘Methodology for meat species identification:
a review’, Meat Sci, 15, 215–224.
Hoffman, D R (1983) ‘Immunochemical identification of the allergens in egg white’, J
Holen, E and Elsayed, S (1990) ‘Characterization of four major allergens of hen egg
white by IEF/SDS-PAGE combined with electrophoretic transfer and IgE-
immunoautoradiography’, Int Arch Allergy Appl Immunol, 91, 136–141.
Holzhauser, T, Wangorsch, T and Vieths, S (2000) ‘Polymerase chain reaction (PCR) for
detection of potentially allergenic hazelnut residues in complex food matrixes’, Eur
Høst, A (1997) ‘Cow’s milk allergy’, J R Soc Med, 90 (Suppl. 30), 31–39.
Høst, A, Husby, S, Gjesing, B, Larsen, J N and Løwenstein, H (1992) ‘Prospective
estimation of IgG, IgG subclass and IgE antibodies to dietary proteins in infants with
cow milk allergy. Levels of antibodies to whole milk protein, BLG and ovalbumin in
relation to repeated milk challenge and clinical course of cow milk allergy’, Allergy,
47, 218–229.
Iikura, Y, Imai, Y, Imai, T, Akasawa, A, Fujita, K, Hoshiyama, K, Nakura, H, Kohno, Y,
Koike, K, Okudaira, H and Iwasaki, E (1999) ‘Frequency of immediate-type food
allergy in children in Japan’, Int Arch Allergy Immunol, 118, 251–252.
Immer, U, Dupont, A, Giovannacci, I, Guizard, C, Demeulemester, C and Dreher, R M
(2003) ‘Détection des protéines d’œuf dans les produits alimentaires’, Bull Liaison
CTSCCV, 13 (4), 17–20.
ISO (1999) ‘General requirements for the competence of testing and calibration laboratories’,
ISO/CEI 17025, Geneva, ISO.
Janssen, F W, Voortman, G and de Baaij, J A (1987) ‘Rapid detection of soya protein,
casein, whey protein, ovalbumin and wheat gluten in heat-treated meat products by
means of dot blotting’, Fleischwirtsch, 67, 577–580.
Jonsson, H and Hellenäs, K E (2001) ‘Optimizing assay conditions in the detection of
food allergens with Biacore’s SPR technology’, Biacore J, 2, 16–18.
Kanny, G, Moneret-Vautrin, D A, Flabbée, J, Beaudoin, E, Morisset, M and Thévenin, F
(2001) ‘Population study of food allergy in France’, J Allergy Clin Immunol, 108,
133–140.
Kim, T E, Park, S W, Noh, G and Lee, S (2002) ‘Comparison of skin prick test results
between crude allergen extracts from foods and commercial allergen extracts in atopic
dermatitis by double-blind placebo-controlled food challenge for milk, egg, and soybean’,
Yonsei Med J, 43, 613–620.
King, N L (1984) ‘Species identification of cooked meats by enzyme-staining of
isoelectricfocusing gels’, Meat Sci, 11, 59–72.
Koppelman, S J, Wensing, M, de Jong, G A and Knulst, A C (1999) ‘Anaphylaxis caused

by the unexpected presence of casein in salmon’, Lancet, 354, 2136.
Langeland, T (1982) ‘A clinical and immunological study of allergy to hen’s egg white.
III. Allergens in hen’s egg white studied by crossed radio-immunoelectrophoresis
(CRIE)’, Allergy, 37, 521–530.
Laoprasert, N, Wallen, N D, Jones, R T, Hefle, S L, Taylor, S L and Yunginger, J W (1998)
‘Anaphylaxis in a milk-allergic child following ingestion of lemon sorbet containing
trace quantities of milk’, J Food Prot, 61, 1522–1524.
Leduc, V, Peltre, G, Guizard, C, Polack, B, Huneau, J F and Demeulemester, C (1997)
‘Antigènes et allergènes de lait, de blanc d’œuf, de soja. 1 – Dans les ingrédients
alimentaires’, Ann Fals Exp Chim, 90, 139–149.
Leduc, V, Demeulemester, C, Polack, B, Guizard, C, Le Guern, L and Peltre, G (1999)
‘Immunochemical detection of egg white antigens and allergens in meat products’,
Allergy, 54, 464–472.
Lee, J W, Kim, J H, Yook, H S, Kang, K O, Lee, S Y, Hwang, H J and Byun, M W (2001)
‘Effects of gamma radiation on the allergenic and antigenic properties of milk proteins’,
J Food Prot, 64, 272–276.
Lin, S, Tornatore, P, King, D, Orlando, R, Weinberger, S R (2001) ‘Limited acid hydrolysis
as a means of fragmenting proteins isolated upon proteinchip array surfaces’, Proteomics,
1, 1172–1184.
Lipkin, E, Shalom, A, Khatib, H, Soller, M and Friedmann, A (1993) ‘Milk as a source
of deoxyribonucleic acid and as a substrate for polymerase chain reaction’, J Dairy
Sci, 76, 2025–2032.
López-Calleja, I, González, I, Fajardo, V, Rodríguez, M A, Hernández, P E, García, T and
Martín, R (2004) ‘Rapid detection of cows’ milk in sheeps’ and goats’ milk by a
species-specific polymerase chain reaction technique’, J Dairy Sci, 87, 2839–2845.
Mäkinen-Kiljunen, S and Mussalo-Rauhamaa, H (2002) ‘Casein, an important house dust
allergen’, Allergy, 57, 1084–1085.
Mäkinen-Kiljunen, S and Palosuo, T (1992) ‘A sensitive enzyme-linked immunosorbent
assay for determination of bovine β-lactoglobulin in infant feeding formulas and in
human milk’, Allergy, 47, 347–352.
Malmheden-Yman, I, Kruse, B and Demeulemester, C (2002) ‘Une étude interlaboratoire
du NMKL: Détection de lait, de blanc d’œuf, d’arachide et de noisette dans les
produits alimentaires’, Bull Liaison CTSCCV, 12(1), 3–7.
Marchand, F, Mecheri, S, Guilloux, L, Iannascoli, B, Weyer, A and Blank, U (2003)
‘Human serum IgE-mediated mast cell degranulation shows poor correlation to allergen-
specific IgE content’, Allergy, 58, 1037–1043.
Mariarger, B, Solve, M, Eriksen, H and Brogen, C H (1994) ‘Bovine beta-lactoglobulin
in hypoallergenic and ordinary infant formulas measured by an indirect competitive
ELISA using monoclonal and polyclonal antibodies’, Food Agric Immunol, 6, 73–83.
Martin-Esteban, M, Garcia-Ara, M C, Banque-Molas, M, Boyano-Martinez, M T, Martin-
Munoz, F and Diaz-Pena, J M (1998) ‘Evaluation of an extensively hydrolyzed casein-
whey protein formula in immediate cow’s milk protein hypersensitivity’, J Pediatr
Gastroenterol Nutr, 26, 398–401.
May, C D (1976) ‘Objective clinical and laboratory studies of immediate hypersensitivity
reactions to foods in asthmatic children’, J Allergy Clin Immunol, 58, 500–515.
Maynard, F, Chatel, J M and Wal, J M (1999) ‘Immunological IgE cross-reactions of
bovine and human α-lactalbumins in cow’s milk allergic patients’, Food Agric Immunol,
11, 179–189.
Moen, L H, Plassen, C, Miller, I, Gutleb, A C and Egaas, E (2003) ‘Rocket immuno-
electrophoresis (RIE) as a tool to identify trace amounts of casein in food – an old
method that out-passes a modern commercial ELISA’, XXII Congress Eur Acad Allergol
Clin Immunol, Paris, June 7–11, 603.
Mohammed, I, Mullet, W M, Lai, E P C and Yeung, J M (2001) ‘Is biosensor a viable
method for food allergen detection?’, Analytica Chimica Acta, 444, 97–102.
Moneret-Vautrin, D A, Kanny, G and Frémont, S (2003) ‘Laboratory tests for diagnosis

of food allergy: advantages, disadvantages and future perspectives’, Allerg Immunol
(Paris), 35, 113–119.
Moreno-Bondi, M C, Alarie, J P and Vo-Dinh, T (2003) ‘Multi-analyte analysis system
using an antibody-based biochip’, Anal Bioanal Chem, 375, 120–124.
Morisset, M, Moneret-Vautrin, D A, Kanny, G, Guénard, L, Beaudouin, E, Flabéee, J and
Hatahet, R (2003) ‘Thresholds of clinical reactivity to milk, egg, peanut and sesame
in immunoglobulin E-dependent allergies: evaluation by double-blind or single-blind
placebo-controlled oral challenges’, Clin Exp Allergy, 33, 1046–1051.
Nilsson, C, Öman, H, Hallden, G, Lilja, G, Lundberg, M and Härfast, B (1999) ‘A case
of allergy to cow’s milk hydrolysate’, Allergy, 54, 1322–1326.
Norgaard, A, Skov, P S and Bindslev-Jensen, C (1992) ‘Egg and milk allergy in adults:
comparison between fresh foods and commercial allergen extracts in skin prick test
and histamine release from basophils’, Clin Exp Allergy, 22, 940–947.
Norgaard, A, Bernard, H, Wal, J M, Peltre, G, Skov, P S, Poulsen, L K and Bindslev-
Jensen, C (1996) ‘Allergenicity of individual cow milk proteins in DBPCFC-positive
milk allergic adults’, J Allergy Clin Immunol, 97(1Pt3), 237.
O’Farrell, P H (1975) ‘High resolution two-dimensional electrophoresis of proteins’, J
Biol Chem, 250, 4007–4021.
Osterballe, M and Bindslev-Jensen, C (2003) ‘Threshold levels in food challenge and
specific IgE in patients with egg allergy: is there a relationship?’ J Allergy Clin
Immunol, 112, 196–201.
Peltre, G, Lapeyre, J and David, B (1982) ‘Heterogeneity of grass pollen allergens (Dactylis
glomerata) recognized by IgE antibodies in human patients sera by a new nitrocellulose
immunoprint technique’, Immunol Lett, 5, 127–131.
Pichler, W J and Campi, P (1992) ‘Allergy to lysozyme/egg white-containing vaginal
suppositories’, Ann Allergy, 69, 521–525.
Pressi, G, Curioni, A, Peruffo, A D B L, Furegon, L and Zamorani, A (1994) ‘Determination
of egg content in egg pasta by an indirect ELISA procedure’, J Sci Food Agric, 64,
163–169.
Prouvost-Danon, A, Weyer, A, Silva Lima, M, Lapeyre, J, Marchand, F, Expert-Besançon,
N, Delpierre, A and David, B (1994) ‘Histamine release by Dolichos lablab lectin’, in
van Driessche, E, Fischer, J, Beeckmans, S and Bøg-Hansen, T C, Lectins: Biology,
Biochemistry, Clinical Biochemistry, Hellerup (Denmark), Textop, 10, 338–344.
Quirce, S, Diez-Gomez, M L, Eiras, P, Cuevas, M, Baz, G and Losada, E (1998) ‘Inhalant
allergy to egg yolk and egg white proteins’, Clin Exp Allergy, 28, 478–485.
Ragno, V, Giampietro, P G, Bruno, G and Businco, L (1993) ‘Allergenicity of milk
protein hydrolysate formulae in children with cow’s milk allergy’, Eur J Pediatr, 152,
760–762.
Rancé, F and Dutau, G (1997) ‘Labial food challenge in children with food allergy’,
Pediatr Allergy Immunol, 8, 41–44.
Rauch, P, Hochel, I and Kas, J (1990) ‘Sandwich enzyme immunoassay of hen egg
lysozyme in foods’, J Food Sci, 55, 103–105.
Restani, P, Plebani, A, Velona, T, Cavagni, G, Ugazio, A G, Poiesi, C, Muraro, A and
Galli, C L (1996) ‘Use of immunoblotting and monoclonal antibodies to evaluate the
residual antigenic activity of milk protein hydrolysed formulae’, Clin Exp Allergy, 26,
1182–1187.
Restani, P, Gaiaschi, A, Plebani, A, Beretta, B, Cavagni, G, Fiocchi, A, Poiesi, C, Velona,
T, Ugazio, A G and Galli, C L (1999) ‘Cross-reactivity between milk proteins from
different animal species’, Clin Exp Allergy, 29, 997–1004.
Restani, P, Gaiaschi, A, Plebani, A, Beretta, B, Velona, T, Cavagni, G, Poiesi, C, Ugazio,
A G and Galli, C L (2000) ‘Evaluation of the presence of bovine proteins in human
milk as a possible cause of allergic symptoms in breast-fed children’. Ann Allergy
Asthma Immunol, 84, 353–360.
Rhim, G S and McMorris, M S (2001) ‘School readiness for children with food allergies’,
Ann Allergy Asthma Immunol, 86, 172–176.
Rodríguez, E, Martín, R, García, T, González, I, Morales, P, Sanz, B and Hernández, P E
(1993) ‘Detection of cow’s milk in ewes’ milk and cheese by a sandwich enzyme-
linked immunosorbent assay ELISA’, J Sci Food Agric, 61, 175–180.
Rugo, E, Wahl, R and Wahn, U (1992) ‘How allergenic are hypoallergenic infant formulae?’
Clin Exp Allergy, 22, 635–639.
Rupa, P and Mine, Y (2003) ‘Immunological comparison of native and recombinant egg
allergen, ovalbumin, expressed in Escherichia coli’, Biotechnol Lett, 25, 1917–1924.
Saarinen, K M, Juntunen-Backman, K, Jarvenpaa, A L, Klemetti, P, Kuitunen, P, Lope, L,
Renlund, M, Siivola, M, Vaarala, O and Savilahti, E (2000) ‘Breast-feeding and the
development of cows’ milk protein allergy’, Adv Exp Med Biol, 478, 121–130.
Sabbah, A, Drouet, M, Lauret, M G, Giffard, G and Minkhar, M (1997) ‘Étude de
réactivité croisée entre le lait de vache (LV) et le lait de chèvre (LC)’, Allergie Immunol
(Paris), 29, 203–214.
Sajdok, J, Rauch, P, Paluska, E and Kas, J (1990) ‘Determination of egg and egg white
content of food products by means of immunochemical assessment of ovalbumin’, J
Sci Food Agric, 53, 253–259.
Schäfer, T, Kramer, U, Dockery, D, Vieluf, D, Behrendt, H and Ring, J (1999) ‘What
makes a child allergic? Analysis of risk factors for allergic sensitization in preschool
children from East and West Germany’, Allergy Asthma Proc, 20, 23–27.
peanut and nut allergy in the US determined by a random digit dial telephone survey’,
Sotto, D, Tounian, P, Baudon, J J, Pauliat, S, Challier, P, Fontaine, J L and Girardet, J P
(1999) ‘Allergy to cow’s milk protein hydrolysates: apropos of 8 cases’, Arch Pediatr,
6, 1279–1285.
Szepfalusi, Z, Ebner, C, Pandjaitan, R, Orlicek, F, Scheiner, O, Boltz-Nitulescu, G, Kraft,
D and Ebner, H (1994) ‘Egg yolk alpha-livetin (chicken serum albumin) is a cross-
reactive allergen in the bird-egg syndrome’, J Allergy Clin Immunol, 93, 932–942.
Talapatra, A, Rouse, R and Hardiman, G (2002) ‘Protein microarrays: challenges and
promises’, Pharmacogenomics, 3, 527–536.
Taudou, G, Varin-Blank, N, Jouin, H, Marchand, F, Weyer, A and Blank, U (1993)
‘Expression of the alpha chain of human Fc epsilon RI in transfected rat basophilic
leukemia cells: functional activation after sensitization with human mite-specific IgE’,
Int Arch Allergy Immunol, 100, 344–350.
Taylor, S L, Hefle, S L, Bindslev-Jensen, C, Bock, S A, Burks, A W, Jr, Christie, L, Hill,
D J, Høst, A, Hourihane, J O, Lack, G, Metcalfe, D D, Moneret-Vautrin, D A, Vadas,
P A, Rancé, F, Skrypec, D J, Trautman, T A, Yman, I M and Zeiger, R S (2002)
much is too much?’ J Allergy Clin Immunol, 109, 24–30.
Taylor, S L, Hefle, S L, Bindslev-Jensen, C, Atkins, F M, André, C, Bruijnzeel-Koomen,
C, Burks, A W, Bush, R K, Ebisawa, M, Eigenmann, P A, Høst, A, Hourihane, J O,
Isolauri, E, Hill, D J, Knulst, A, Lack, G, Sampson, H A, Moneret-Vautrin, D A,
Rancé, F, Vadas, P A, Yunginger, J W, Zeiger, R S, Salminen, J W, Madsen, C and
Abbott, P (2004) ‘A consensus protocol for the determination of the threshold doses
Turin, L and Bonomi, F (1994) ‘Immunochemical detection of ovalbumin in dairy-based
foods’, J Sci Food Agric, 64, 39–45.
Umpierrez, A, Quirce, S, Maranon, F, Cuesta, J, Garcia-Villamuza, Y, Lahoz, C and
Sastre, J (1999) ‘Allergy to goat and sheep cheese with good tolerance to cow cheese’,
Clin Exp Allergy, 29, 1064–1068.
Urisu, A, Ando, H, Morita, Y, Wada, E, Yasaki, T, Yamada, K, Komada, K, Torii, S, Goto,
M and Wakamatsu, T (1997) ‘Allergenic activity of heated and ovomucoid-depleted
egg white’, J Allergy Clin Immunol, 100, 171–176.
Venien, A, Levieux, D, Astier, C, Briand, L, Chobert, J M, Haertle, T (1997) ‘Production

and epitopic characterization of monoclonal antibodies against bovine beta-lactoglobulin’,
J Dairy Sci, 80, 1977–1987.
Wal, J M (1998) ‘Cow’s milk allergens’, Allergy, 53, 1013–1022.
Wal, J M (2001) ‘Structure and function of milk allergens’, Allergy, 56(Suppl 67), 35–38.
Wal, J M, Bernard, H, Créminon, C, Hamberger, C, David, B and Peltre, G (1995) ‘Cow’s
milk allergy: the humoral immune response to eight purified allergens’, Adv Exp Med
Biol, 371B, 879–881.
Walsh, B J, Elliott, C, Baker, R S, Barnett, D, Burley, R W, Hill, D J and Howden, M E
(1987) ‘Allergenic cross-reactivity of egg-white and egg-yolk proteins. An in vitro
study’, Int Arch Allergy Appl Immunol, 84, 228–232.
Weinberger, S R, Morris, T S and Pawlak, M (2000) ‘Recent trends in protein biochip
technology’, Pharmacogenomics, 1, 395–416.
Wüthrich, B and Johansson, S G (1995) ‘Allergy to cheese produced from sheep’s and
goat’s milk but not to cheese produced from cow’s milk’, J Allergy Clin Allergy, 96,
270–273.
Yamada, T, Yamada, M, Sasamoto, K, Nakamura, H, Mishima, T, Yasueda, H, Shida, T
and Likura, Y (1993) ‘Specific IgE antibody titers to hen’s egg white lysozyme in
allergic children to egg’, Arerugi, 42, 136–141.
Yeung, J, Newsome, H and Abbott, M (2000) ‘Determination of egg proteins in food
products by enzyme immunoassay’, J AOAC Int, 83, 139–143.
Ylitalo, L, Mäkinen-Kiljunen, S, Turjanmaa, K, Palosuo, T and Reunala, T (1999) ‘Cow’s
milk casein, a hidden allergen in natural rubber latex gloves’, J Allergy Clin Immunol,
104, 177–180.
Zlatanova, J and Mirzabekov, A (2001) ‘Gel-immobilized microarrays of nucleic acids
and proteins, production and application for macromolecular research’, Methods Mol
Biol, 170, 17–38.
14
Detecting wheat gluten in food

F. Janssen, Fascoda FAS Food Consultancy,
The Netherlands
14.1 Introduction
Although allergy to wheat is not uncommon, especially the occupational
aspect of it (baker’s asthma), the proteins and peptide motifs involved in
IgE-mediated allergy have not been identified in detail. It is known that a
broad spectrum of wheat proteins is reactive with sera from wheat-allergic
persons. Battais et al. (2003) found, by using radioallergosorbent tests (RAST),
that 60% of sera from their wheat-allergic subjects had IgE and IgG antibodies
against α /β-gliadins and low molecular weight (LMW) glutenin subunits,
55% to γ-gliadins, 48% to ω-gliadins and 26% to high molecular weight
(HMW) glutenins. These results were confirmed with immunoblotting, by
which it was also shown that 67% of patients have IgE antibodies to the
albumin/globulin fraction. Heat processing of wheat dough seems to render
these proteins resistant to breakdown. It has been shown by in vitro digestion
that, while unheated wheat doughs lost their IgE-binding properties during
digestion, the IgE-binding potential was retained in heat-processed dough
(Simonato et al., 2001).
Even although antibodies to whole wheat proteins (including albumins)
are commercially available, most of the research to detect wheat proteins is
directed towards proteins involved in the pathogenesis of quite another
affliction: coeliac disease, a non-IgE-related hypersensitivity in which the
toxic principle is the storage protein (gluten) from some cereal species.
Although this disease is different from IgE-mediated allergy, the eliciting
proteins are shared to a large extent and, as long as these proteins stick
together, testing for one of them will be proof that wheat is present. Because
the prevalence of coeliac disease is increasing tremendously, primarily due
Detecting wheat gluten in food 245
to increased awareness and improvements in diagnosis, emphasis is placed

on the determination of proteins involved in the pathogenesis of this disease.
Wheat-allergic persons may, however, use the same sources of food in
complying with a wheat-free diet that coeliac sufferers do.
14.1.1 Coeliac disease

In this introduction, some aspects of coeliac disease, such as history, aetiology,
pathology, prevalence, toxic cereals, dietary aspects and developing legislation
(Codex, EU), will be covered. The first description of coeliac disease or
coeliac sprue was given in 1888 by Samuel Gee. Our modern knowledge
concerning the exacerbating factors, the cereal prolamines from wheat, rye
and barley and oats, dates back to the years that followed World War II. It
was then that the Dutch paediatrician Willem Karel Dicke (1950) confirmed
his suspicion about the presumptive toxicity of wheat flour with carefully
designed experiments which led ultimately to the conclusion that wheat, rye
and barley, and possibly oats were toxic to some of his coeliac patients. Soon
afterwards it was discovered that the deleterious factor resided in the protein
fraction of the wheat, notably in the alcohol-soluble part, the gliadin.
In 1954 Paully presented the first description of the coeliac lesion, being
an atrophy of the villi of the small bowel caused by gluten ingestion, and
after that a biopsy of the jejunum, the proximal part of the small bowel,
became the hallmark of the diagnosis. Recently, non-invasive diagnostic
methods have been developed based on serum parameters like antiendomysial
IgA, anti-gliadin IgA and anti-tissue transglutaminase (reviewed by Dieterich
et al., 2000). These tests have made screening of non-suspect populations
feasible, which has revealed that the prevalence of this disease is much
higher than was previously thought. A prevalence of 1:300 was found in Italy
by Catassi et al. (1994) and afterwards similar screening showed that in
other European countries, the prevalence was much higher than previously
thought. In the US, Fasano et al. (2003) reported that 1:133 persons had
coeliac disease. Many of these screening-diagnosed persons, however, had
no clinical symptoms, although this ‘silent’ disease is believed to become
overt later in life and may, if it remains undiagnosed, cause osteopenia and
iron deficiency, among other things.
Coeliac disease manifests itself in children with an array of symptoms,
among which the most prominent are poor growth and weight loss, diarrhoea
and increased fat excretion in the stool (steatorrhoea). This is the ‘classical’
picture. Given the large number of patients who are detected by serological
screening without ever presenting with these classical symptoms, this picture
of coeliac disease is now subject to change. There is a strong association of
coeliac disease with some other diseases, in particular, dermatitis herpetiformis
(DH) which has been named by some as ‘coeliac disease of the skin’, even
although this particular condition does not improve solely by compliance to
a gluten-free diet, but requires further medication.
Coeliac disease is currently considered to be an autoimmune disease and

it is believed that it predisposes one to other autoimmune disorders such as
type I diabetes, autoimmune thyroid and other endocrine diseases. One of
the most serious side effects is osteopenia, while increased risk of malignant
diseases, especially intestinal lymphomas derived from mucosal T-cells,
reproductive failures, neurological disorders and autoimmune diseases, has
also been reported (James and Scott 2001; Pengiran Tengah et al., 2002;
Vestergaard P, 2003). As there are indications that a gluten-free diet protects
coeliac patients against the development of these complications, the importance
of compliance to such a diet is stressed. As there is no cure, a lifelong
compliance to a gluten-free diet is the only way to alleviate symptoms.
Coeliac disease is believed to have a genetic component because there is
a strong association of the disease with the presence of specific HLA class
II antigens, predominantly HLA DQ2 and (to a lesser extent) DQ8 (reviewed
by Louka and Sollid, 2003). Either of these is needed to develop the disease,
although the majority of DQ2 and DQ8 positive persons are coeliac-negative
which indicates that more genes are involved. It is believed that the proteins
implicated in the disease, the prolamins, bind to the DQ2 or DQ8 molecule,
giving rise to T-cell activation and triggering a sequence of events leading to
full-blown disease. Questions arose as to how this occurrs, because the
prolamines, given their hydrophobic nature, should not bind very strongly to
the DQ2/DQ8 peptide binding groove. It was found by Sjöstrom et al. (1998)
that T-cell reactivity is enhanced by the action of tissue transglutaminase,
which deamidates specific glutamine residues of gliadin and glutenin peptides.
Extensive research has shown that many sequences (deamidated or not) react
with susceptible T-cells, and it has now been found that these reactions are
to some extent idiotypic, i.e. different between patients.
A role for the innate immune system has also been hypothesized. The
innate immune system consists of a variety of immune responses that constitute
the first line of defence against the invasion of foreign substances, e.g.
pathogens. Its response to challenge is ahead of the induction of the adaptive
immune response, is not antigen-specific and is without memory. Activation
at inappropriate sites and/or to an excessive degree can exacerbate tissue
damage in several diseases, and a possible role in the first stages of coeliac
disease has been suggested (Maiuri et al., 2003).
14.1.2 Gluten proteins

The storage protein of wheat, gluten, is composed of monomeric gliadins
and polymeric glutenins. Wheat gluten has the unique capacity of forming a
highly cohesive, elastic dough which prevents the escape of gas bubbles
formed by the yeast cells during leavening. Wheat gluten can be prepared by
kneading an amount of wheat flour into dough and washing this dough under
running tap water to remove the starch particles until the remaining, chewing-
gum-like residue is starch-free. In industry, more sophisticated procedures
are used. Traditionally, cereal proteins have been classified into four types
according to their solubility (Osborne, 1907):
• albumins – soluble in water or dilute salt solutions;
• globulins – insoluble in water, but soluble in dilute salt concentrations
and insoluble at high salt concentrations;
• prolamins – soluble in aqueous alcohol;
• glutelins – soluble in neither water nor dilute salt solutions but soluble in
dilute acid or alkali.
An alternative classification to that described above has been proposed based
on chemical similarity of reported amino acid sequences of individual
components rather than on differential solubility.
A comparison between these nomenclatures is given by Shewry et al.
(1986). In fact the LMW glutenins can also be considered as aggregated
gliadins, because their composition is comparable to the α-gliadin fraction.
The difference is that they are present in a polymeric form.
The gliadins and glutenins are characterized by a high content of the
amino acids glutamine (Q) and proline (P), ca. 30 and 17 mol%, respectively.
They lack nutritionally important amino acids like tryptophan, lysine and
methionine which makes their nutritional score rather low. They also lack
any enzyme activity and are generally considered as a nitrogen source for the
germinating seed. Gliadins are classified into fractions, according to their
mobility in an electrophoretic field, as α-, β-, γ- and ω-gliadins: the differences
between the first two fractions are small, and they are commonly designated
as α/β. A unique property of ω-gliadin is that it does not contain cysteine and
is not involved in disulfide exchange.
Which part of the prolamin is toxic?

When the coeliac toxicity of wheat was established, a quest to establish
which part of the wheat contained the toxic principle began. Older research
confined the toxicity to the prolamin fraction of the wheat protein and not
the glutenin portion. More recently, however, T-cell toxicity of glutenin has
been described (Van de Wal, et al., 1999). Extensive research has been
performed to elucidate the peptide structures that precipitate the disease in
susceptible persons. As mentioned before, coeliac-toxic cereals are wheat,
rye, barley and their cross-bred varieties, while the toxicity of oats is debated.
Oats can be easily contaminated by wheat or barley since the these crops are
often grown in close proximity. Quite recently it has also been found that
some coeliacs have avenin-reactive mucosal T-cells pointing to intrinsic toxicity
of oats. Clinical follow-up of coeliac disease patients eating oats is consequently
advisable (Arentz-Hansen et al., 2004).
Buckwheat, maize, rice, sorghum and millet are considered to be non-
toxic. Research has been focused on wheat because this is the most abundant
species. Clinical testing of gliadin fractions showed that all fractions
(α-, β-, γ- and ω-gliadin) were toxic (Ciclitira et al., 1984). Subsequently, in
a series of experiments Wieser et al. (1983) isolated gliadin fractions by

reversed-phase HPLC and prepared peptides by enzymatic digestion. Further
research on isolated peptides showed that those derived from the N-terminal
part of the protein (residues 25–55, 1–30, 31–55) were toxic (De Ritis et al.,
1988). Much work has since been done with panels of synthetic peptides to
narrow the search for the precipitating motif, and many sequences have been
found to be toxic (reviewed in Stern et al., 2001). Recently, much emphasis
has been placed on T-cell toxicity, and it has been shown that many gliadin
and gluten sequences are toxic to T-cell populations of coeliac patients (Koning,
2003).
14.1.3 Regulatory framework for gluten (gliadin) in gluten-free

dietary products
A Codex Standard for gluten-free foods was adopted by the Codex Alimentarius
Commission in 1976 and published in 1981 as Codex Stan 118, 1981. This
standard defines gluten as ‘those proteins, commonly found in wheat, triticale,
rye, barley or oats to which some persons are intolerant.’ In the context of the
standard, ‘gluten-free’ means that the total nitrogen content of the gluten-
containing cereal grains used in the product does not exceed 0.05 g per
100 g on a dry matter basis. It is easy to become tangled up by this definition,
however, because, although the nitrogen content can be translated into a
gluten content by applying the conventional nitrogen conversion factor and
correcting for non-gluten protein, the level seems to have been validated
only in wheat starch. In wheat starch the aforementioned conversion is not
valid because most of the protein is present as ‘starch granule protein’. The
gluten content of a wheat starch product containing 0.05 g/100 g product is
(was) thought to be about equivalent to 200 mg/1000 g and this was formerly
supposed to be a safe limit. Nowadays more is known about nitrogen/gluten
ratios in wheat starch and it has become clear that the measurement of
nitrogen in this matrix can be used only as an approximation of the gluten
content. Because most gluten-free products consist of mixtures of non-coeliac-
toxic cereals like maize or rice, it is mandatory to measure specifically the
gluten content and not the nitrogen content of a product, as the latter would
be meaningless in this situation.
The 118–1981 Codex Alimentarius standard is nowadays considered to
be outdated, because the threshold is too relaxed for foods which are gluten-
free by nature (i.e. maize or rice flours). In addition, the confusing definition
of ‘gluten-free’ needs to be revised. A new standard has been elaborated by
the Codex Committee on Nutrition and Foods for Special Dietary Uses (CC
NFSDU) but as of this writing this project has not been finalized. A value
of 20 mg gluten/kg product has been proposed as a statutory limit for
gluten-free food that is ‘naturally gluten free’, e.g. maize flour and rice
flour, and this proposal seems to be universally endorsed. No consensus
has yet been reached about food rendered gluten free, like wheatstarch,
where the proposed limit of 200 mg/kg has been rejected by coeliac
organizations.
14.2 Key requirements for detection and quantization

14.2.1 Accuracy and broad scope
Gluten-free foods will probably be the first type of food where threshold
values for the toxic principle will be established. This implies that litigation
is likely to occur and, consequently, high standards of reliability are required.
It is imperative that methods are highly accurate and sufficiently sensitive to
measure well below the levels being proposed as thresholds or already laid
down in food regulations. In addition, the method should allow the
determination of gluten in a wide range of foods that may have been processed
under a variety of conditions (i.e. the method should have a broad scope). In
the past, one of the major concerns has been to find a method able to measure
gluten in heat-processed food. There are no reports that gluten toxicity is
abolished by heat processing and, consequently, the method should be able
to measure gluten in food that has been prepared under a wide variety of
conditions. Wheat gluten is a key factor in achieving good texture in bakery
products. During kneading, leavening and baking of dough, there is an exchange
of disulfide bonds which (helped by the action of oxidizing agents) results in
insolubilization of the gliadin fraction.
14.2.2 Reference material

Determination of gluten (gliadin) is subject to several constraints, one of
which is the reference material used. All the methods published so far are
relative methods, and the analysis is conducted with the help of a calibration
curve made by measuring known concentrations of a reference substance. It
has been shown that the origin and type of gliadin used for calibration has a
tremendous impact on the results of the analysis (Van Eckert et al., 1998).
This fact prompted the European Working Group on Prolamin Analysis and
Toxicity (WGPAT – see Section 14.7) to initiate the development and production
of a reference material produced from a mixture of the 28 most common
European wheat cultivars (Van Eckert et al., 2002; 2005, Klein et al., 2003).
This standard is currently being further evaluated by the EU Institute for
Reference Materials and Methods (IRRM) in Belgium, but it is already
available from this institute (see Section 14.7).
14.3 Types of detection methods

The vast majority of methods which have been published so far are of the
enzyme-linked immunosorbent assay (ELISA) type (Table 14.1). A key

requirement in detecting and quantifying gluten is specificity. The method
should detect (coeliac) toxic cereals and produce negative results with cereals
considered non-toxic to coeliacs such as rice, maize, sorghum, millet, etc. Of
all methods published so far, ELISA methods, which have been developed
from the early 1980s on (Windemann et al., 1982; Meier et al., 1984), form
the bulk, due to unrivalled specificity and ease-of-use. In the following,
several methods are discussed with the emphasis on those having been made
commercially available as kits.
14.3.1 ELISA with antibodies to ω-gliadins

Work of sterling quality has been done by John Skerritt’s group at the
Commonwealth Scientific and Industrial Research Organisation (CSIRO) in
Australia. They developed the first method of determining gluten to be made
available in kit form, and it was also the first method designed to assess the
gluten content of heat-processed food (Skerritt et al., 1984; Skerritt, 1985;
Skerritt and Smith, 1985; Skerritt and Martinuzi, 1986; Skerritt and Hill,
1990, a, b, c).
Events occurring during heat processing of foods are quite complicated.
It has been found by Schofield et al. (1983) among others that baking quality
decreased with heating and was completely destroyed at 75 °C. This was
parallelled by a decreased extractability in a sodium dodecyl sulphate (SDS)
containing buffer. With gel filtration techniques it was found that predominantly
glutenin proteins were destroyed. Extractability of gliadins remained intact
until 75 °C, but decreased rapidly when the dough was heated at higher
temperatures, and at 100 °C only the ω-gliadins remained soluble, presumably
because this fraction did not contain sulfhydryl groups. When a reducing
agent such as dithiothreitol (DTT) was added, the extractability could be
restored. It is assumed that unfolding of glutenin at elevated temperatures
promotes sulphydryl/disulfide exchange not only between glutenin units but
also between glutenin and gliadin, and that this exchange becomes fixed
upon cooling or heating at temperatures higher than 75 °C. This decrease in
extractability was also found by Weegels et al. (1994a, b) and confirmed by
Rumbo et al. (1996, 2001), who found that the effect of heat was almost
absent when flour was heated in dry conditions; Wieser (1998) found this
also.
The novel aspect of the method developed by Skerritt et al. is that it uses
a selected monoclonal antibody (mAB) against the most heat-tolerant gliadin
fraction, the ω-gliadin. The method is available as a kit and a collaborative
trial has been carried out. The method has been approved by the AOAC
(Skerritt and Hill, 1991; AOAC, 1995).
Over time it was found that the method had some shortcomings which
were difficult to overcome. The first one was the sensitivity. As ω-gliadin is
a relatively minor fraction, the method has limited sensitivity compared to
Table 14.1 Overview of immunochemical assays for gluten quantification in food
Year Number Reference Layout Antibody1 Protein Source Epitope Sensitivity2 Reactivity3 Cross- Extraction Remark
reactivity4
1982 1 Windemann Sandwich pAB A-gliadin Wheat Unknown 1 ng/ml 70% ethanol,
et al., (A-gliadin) 40 °C
1982* 10 ng/ml
(gliadin)
1983 2 Ciclitira pAB α-gliadin Wheat Unknown 1 ng/ml α-gliadin, 5
and Lennox no reaction

1983 with γ- and
ω-gliadin,
neither with
barley,
rye and oats
1986 4 Troncone Sandwich pAB Gliadin Wheat Unknown 5 ng/ml Maize, rice
et al., 1986
1987 5 Freedman pAB/mAB Gliadin Wheat Unknown 15 ng/ml
et al., 1987
1988 6 Friis 1988 Competitive pAB Gliadin Wheat Unknown 1 ng/ml Wheat >>> 70% ethanol
rye >> barley
> oats
1988 7 Ayob et al., pAB Gliadin Wheat Unknown 30 ng/ml
1988
6
1989 8 Mills et al., Capture: Gliadin Wheat Unknown
1989 IgY,
Detection:
mAB
1990 9 Skerritt Sandwich mAB ω-gliadin Wheat Unknown 100 ng/ml Wheat > 40% ethanol 7
et al., rye >> barley

1984–1990, > oats
a,b,c
Table 14.1 Continued
reactivity4
1994 10 Ellis et al., pAB/mAB Gliadin Wheat 54 amino- 15 ng/ml Wheat = rye
1994 acid >> barley
peptid >> oats
N-terminal
1995 11 Chirdo Competitive pAB Gliadin Wheat Unknown 1 ng/ml Wheat >> 70% ethanol
et al., 1995 barley > rye after pre-
>> oats extraction
of albumins/
globulins by
0.15 M-NaCl
1995 12 Albrecht Sandwich Gliadin Wheat Unknown 25 ng/ml Maize
and Toth,
1995
1998 13 Ellis et al., Sandwich pAB/mAB Gliadin Wheat A-gliadin 4 ng/ml Wheat >> 40% ethanol
1998 31–49 rye > barley or 50%
= oats propane-1-ol
with 1%
M.E in buffer
1998 14 Chirdo Sandwich mAB Gliadin Wheat Unknown 20 ng/ml 70% ethanol
et al.,1998 after pre-
extraction of
albumins/
globulins by
0.15 M-NaCl
1998 15 Chirdo Competitive mAB Gliadin Wheat Unknown 5 ng/ml
et al., 1998
1998 16 Chirdo Capture mAB Gliadin Wheat Unknown 1 ng/ml
et al., 1998
Table 14.1 Continued
reactivity4
1998 17 Sorell Sandwich mAB Secalin Rye / Unknown 3 ng/ml Rye > wheat Ethanol 60%
et al., 1998 2mAB wheat = barley >>>
oats
8
2003 18 Valdez Sandwich mAB Secalin Rye QXPW/FP 1,5 ng/ml Barley = rye Ethanol 60%
et al. 2003 = wheat
>>> oats
2003 19 Henterich Direct, mAB Secalin Rye As 18 0,16 ng/ml As 18 Reducing
et al., 2003 signal agent with
amplified guanidine
with PCR hydrochloride
2004 20 Spaenij- Competitive mAB α-gliadin Wheat α-59-71, 12 ng/ml Barley = Ethanol 40% 9
Dekking and γ- γ142-153 wheat = rye

et al., 2004 gliadin and γ147- >>> oats
159
* in German
1
pAB means polyclonal antibody; mAB means monoclonal antibody
2
sensitivity is given in ng/ml. Actual detection level in food is depending on amount of food extracted
3
reactivity: the sensitivity with respect to the various prolamins is stated
4
cross-reactivity: if unspecific, cross-reacting cereal species are stated
5
radioimmunoassay, Staphylococcus aureus used to separate bound and free125 iodinated antigen
6
designed to detect conformation-dependent epitopes in wheat
7
ringtested in AOAC framework
8
alternatively an extraction is carrried out in a reducing environment containing guanidine hydrochloride
9
uses ABs detecting T-cell epitopes
methods using antibodies against a broader spectrum of wheat gliadins. This

shortcoming could be alleviated by modifying the reporter system (replacing
the initially proposed ABTS substrate by the more sensitive
tetramethylbenzidine–TMB). Another option would be to dilute the initial
extract less than prescribed, but this would have negative effects on the
specificity of the assay.
More serious, however, has been the finding that the ω-gliadin content
varied considerably in European wheat varieties. Wieser et al. (1994) found
that the relative contribution of ω-gliadin to the gliadin pool could vary
between 6.2 and 20%, depending on the wheat variety, which means that,
related to the ω-gliadin content of the standard, the values could be over- or
underestimated by a factor of 2–3.
Recent research by Seilmeier and Wieser (2003) showed that other factors
also contributed and that the between-cultivar variation (as determined with
the anti-ω kit) was even higher than could be accounted for by ω content as
determined by reversed-phase HPLC. It is suggested that frequency or spacing
of ω epitopes or non-specific binding of α-gliadin could be the cause of
these discrepancies.
Because wheat products in trade are often not present as pure varieties,
the effect of variable ω-gliadin/total gliadin ratios might not be as detrimental
as it seems at first sight. Using a polyclonal competitive assay, Rumbo et al.
(2001) found a good correlation between the standard used in the ω-based
commercial kit and a gliadin purchased through Sigma Chemical Company
(St Louis, MO). The variable ω content was refuted by Skerritt (personal
communication) with the argument that in some instances the ω-gliadin in
the reference material had been washed out during gliadin preparation. Omega-
gliadin is the most soluble gliadin, especially in water at room temperature
and above, and is also poorly extracted in 70% ethanol. This view was
endorsed by Kasarda (personal comment in Ellis et al., 1998).
14.3.2 ELISAs in a competitive format

Principally, ELISAs in a competitive format have the advantage that no
multiple or repetitive epitope is needed, in contrast to the majority of sandwich
ELISAs. An assay in a competitive format would thus be ideal to determine
enzymatically degraded gluten in those situations where such degradation is
to be expected, e.g. in malted cereals or ‘solubilized’ wheat protein. Friis
(1988) developed a sequentional competitive assay based on a polyclonal,
Protein A-purified antiserum against gliadin. The specificity of the antibody
was checked by immunoblotting and dot blot. The antibody reacted with
α-, β-, and γ- as well as ω-gliadin.
Antibody and sample extract diluted in bovine seralbumin (BSA)/ Tween
20 ® (ICI America Inc, Wilmington, DE) mix containing 0.02 mmol DTT
was incubated for 16 hours. Polystyrene plates were coated with an optimized
amount of gliadin, washed and incubated for 30 minutes with the antibody/
antigen mix for 30 minutes at 20 °C. After washing, horseradish peroxidase

(HRP) conjugated swine anti-rabbit IgG, 1:700 diluted in phosphate buffered
saline (PBS) containing 0.05% Tween 20 ® and 0.5% BSA was added and
the plate was incubated again for 30 minutes at 20 °C. O-phenylenediamine
(OPD) was used as the peroxidase substrate. The assay had a detection limit
of 1 ng/ml. The antibodies, however, showed poor reactivity against rye
prolamins and even lower towards barley and oat prolamins. A gliadin-like
response was found in several commercial flours, e.g. buckwheat (may have
been contaminated) and wheat starch (which always contains residual gluten).
A different, sequentional assay was developed by Chirdo et al. (1995) using
a polyclonal antibody against gliadin. A pre-incubation in the liquid phase was
carried out by mixing standard or sample extracts at appropriate dilutions with
antibody dilutions for a predetermined time and temperature (optimal 2 hours
at 37 °C). An aliquot of the mixture was subsequently incubated in microtiter
wells precoated with an optimized amount of gliadin. Binding of the antibody
to gliadin on the wall was visualized by incubating with goat-anti-rabbit IgG
HRP conjugate. The optimized assay had a sensitivity of 1 ng/ml which enabled
determination of 1 ppm gluten at a sample dilution of 1/50.
It was found that the effect of the extraction solvent (ethanol) was
pronounced. Elevated ethanol concentrations in the final incubation mix
(0.8–4% when 40% ethanol had been used) resulted in a marked decrease in
signal, especially when sample dilutions of 1/50 or 1/10 were used. This
could result in an under-estimation of the gliadin content by 2–40%, an
outcome which stresses the importance of using high working dilutions.
Consequently, a test with high sensitivity allowing for high sample dilutions
is preferred in enforcing low statutory gluten threshold values.
The same authors (Chirdo et al., 1998) investigated the use of the biotin/
streptavidin amplification system using three different mABs developed against
commercial gliadin. Three formats were compared: a sequentional competitive,
a capture ELISA and a competitive assay using biotin-labelled gliadin. The
first had a detection limit of 20 ng gliadin/ml with the most sensitive mAB; the
capture format had a sensitivity of 1 ng/ml with a different mAB – the mAB
which showed the best performance in the sequentional competitive format
gave only poor results. Although there are data suggesting that binding to plastic
surfaces is often the cause of structural changes, this seemed not the case here,
as all antibodies were equally able to bind biotinylated gliadin. The authors
suggest that the cause might have been that the mAB react with epitopes present
at a low density. The competitive assay using biotin-labelled gliadin showed
a sensitivity of 5 ng/ml. The authors also suggest that the latter type of assay
may be used in determining heated or enzymatically degraded prolamin peptides.
14.3.3 ELISAs with mABs and extraction under reduced conditions

Basically, two means are used to circumvent impaired extraction efficiency
in products that have been heat processed: the use of an antibody against
heat stable gliadin fractions or the use of reducing agents and/or chaotropic
agents to reverse the disulfide interactions that take place during dough
kneading and baking.
The latter approach was investigated by Sorell et al. (1998) and García et
al. (2005), and a patent has been applied for this extraction procedure (WO
02/092633). An amount of 0.25 g of a food is extracted with 2.5 ml of a
solvent containing 250 mM 2-mercaptoethanol and 2 M guanidine
hydrochloride in PBS. The extract is kept at 50 °C for 40 minutes and is then
diluted with 7.5 ml 80% ethanol, after which it is vortexed and incubated for
one hour at room temperature in a rotary shaker. After centrifugation, the
extracts are ready to be used. A panel of antibodies was generated against rye
prolamins. Out of this panel, two, R3 and R5, were chosen for the assay in
addition to another, 13B4, which has been described elsewhere (Ellis et al.
1998). A mixture of R5 and 13B4, which have complementary selectivities
(R5 reacting with secalins and hordeins and 13B4 with gliadins), was used
to coat the wells. This multiple coating consequently allowed the binding of
all three cereal species. A third antibody (R3) was conjugated with horseradish
peroxidase and used as a reporter antibody. The assay thus developed was
able to detect gliadins, hordeins and secalins at about the same sensitivity.
The detection level as reported by the authors was 1.5 ng/ml. No combination
of the R-monoclonals was sensitive enough to extend the sensitivity to oats:
a separate method was developed to detect this cereal.
A modification of this assay was described by Valdés et al. (2003). In this
assay (sandwich format) just one monoclonal (the R5) antibody was used for
capturing as well as detection. The molecular recognition pattern of this
antibody has been investigated by Osman et al. (2001) by phage display and
pepscan. With both techniques a pentapeptide consensus sequence was found
(QXPW/FP, resp. QQPFP). This sequence is one of the amino acid motifs
believed to be involved in the pathogenicity of coeliac disease. It is present
in a high number of copies in γ- and ω-gliadin, and in a smaller number in
α-gliadin and LMW glutenin. The presence of an epitope in LMW glutenin
poses problems with respect to correct calibration of the assay because,
when comparing to a gliadin standard, part of the glutenin will be reported
as gliadin. In this assay, however, using the same mAB for capture and
reporting, this might not lead to substantial errors because multiple epitopes
would be needed for detection of glutenin, and LMW glutenin contains only
one. In addition, the number of epitopes on γ- and ω-gliadin outweighs the
number of epitopes on the LMW glutenin. The authors report a sensitivity of
1.56 ng gliadin/ml corresponding to 1.5 mg/kg in food using a low dilution
of 1:25. The reproducibility and repeatability was stated to be 8.7% and the
repeatability 7.7% in a small scale investigation.
A collaborative trial was organized in 2002 by the WGPAT (see Section
14.7), the provisional results of which have been published (Immer et al.,
2003; Immer and Haas-Lauterbach, 2004). Twelve foods were analysed, of
which four were spiked with the reference gliadin developed by the WGPAT
in amounts between 35 and 168 mg/kg. The performance of two R5 antibody-

based kits from two different manufacturers was compared. Mean recoveries
were quite good, ranging from 71–104% with one kit and from 79–111%
with the other. The relative standard deviations (RSDs) varied between 22
and 52. Compared to the method as developed by Skerritt, these methods
have the advantage that the reactivity shows a better correlation with coeliac
toxicity and that they have a broader specificity, allowing the determination
of wheat, as well as rye and barley. Given the high abundance of the QQPFP
repeat, a strong dependence on ω content of the cultivar might be present.
This ‘R5’ method has been endorsed by the Codex Alimentarius Committee
Methods of Analysis and Sampling (CC MAS) on a temporary basis pending
publication of a full statistical report (see report of the twenty-sixth session of
the Codex Committee on methods of analysis and sampling, Budapest, Hungary,
4–8 April 2005). A publication of the statistical report of the collaborative trial
is in preparation (Méndez et al., 2005, submitted). The method is available as
a kit produced by two manufacturers, R-Biopharm (Darmstadt, Germany) and
Ingenasa (Madrid, Spain) and through other distributors.
14.3.4 ELISAs with monoclonal antibodies against toxic peptide

motifs
Polyclonal antibodies for use in gluten detection methods are regularly prepared
by immunizing a rabbit with a gliadin preparation. The antibody can afterwards
be purified with standard biochemical techniques (e.g. Protein A-like affinity
chromatography) and absorbed to remove unspecific antibodies. The resulting
antibody quite often doesn’t have the required profile. By producing mABs
against whole gliadin, there is a greater possibility of selecting antibodies
with the required specificity, but there is no guarantee that this selectivity
reflects the toxicity of the distinct cereals.
To overcome this bias, Ellis et al. (1998) raised mABs against a synthetic
19-mer of the toxic α-gliadin peptide LGQQQPFPPQQPYPQPQPF (31–49
of α-gliadin) and to point-substituted peptides in which amino acids 31, 33,
36, 38, 39, 41, 42, 44, 46 and 48 were replaced by alanine. A polyclonal anti-
unfractionated gliadin antibody was used as a capture antibody in a sandwich
ELISA. Gliadin fractions were prepared and standard solutions were made
from these fractions as well as from whole gliadin cv. Rektor and cv. Timgalen.
The standards were prepared in 40% ethanol and diluted to 100 ug/ml with
water. After filtration, the protein content was determined.
The sensitivity of the assay as determined with the prepared prolamins
varied from 4 ng/ml (whole gliadin and α-gliadin), 16 ng/ml (β- and γ-
gliadin) and 1000 ng/ml for ω-gliadin. Sensitivity against rye, barley and oat
prolamin was respectively 500, 1000 and 1000 ng/ml. The assay was tested
for ruggedness with respect to several aspects: incubation time, repeatability
and reproducibility, and storage stability. The amino acid substitutions revealed
that substitution of amino acid Q (33), P (36) and P (38) with alanine reduced
binding substantially, suggesting that the binding was in the region QQQPFP.
This region is present in α- and ω-gliadin. The truncated motif is present in
avenin (QQQPF) and the more truncated motifs, QQQP and QQPFP, are
present in γ-gliadin, hordein and secalins. The reactivity to ω-gliadin is
lower than expected given that it is present in this fraction. The authors
suggest that this might be due to a lower reactivity of the capture antibody to
ω-gliadins.
Wheat starches, gluten-free ingredients and gluten-free products as well
as spiked samples were analyzed. The initial extract was made by mixing
50 mg product with 1 ml 40% ethanol for one hour at room temperature,
after preliminary defatting twice with four volumes in n-butanol. In a separate
experiment, cooked crust and crumb of a loaf prepared from flour spiked
with 1% var. Rektor gliadin (spiked before the stage of dough formation)
and which had been baked at 230 °C for 10 minutes were analyzed. This loaf
was extracted at 60 °C under nitrogen atmosphere for one hour with a reducing
buffer containing 50% propane-1-ol, 1% mercaptoethanol, 0.08 mol/l Tris-
HCl (pH 7.5) and 2 mol/l urea. In this case, the calibration curve was made
up in the reducing buffer diluted 100-fold. Prolamins from coeliac non-toxic
rice, maize, millet and sorghum as well as presumably gluten-free ingredients
of miscellaneous origin did not cross-react in the assay. The assay could
detect gluten in cooked foods, although at reduced sensitivity, the recovery
being about 17%. Use of reducing agents was not successful, presumably
because they denatured the capture antibodies.
This method might offer good prospects as long as non-reducing conditions
are used. The motif QQQPFP is, however, also present in LMW glutenins
(which become solubilized under reducing conditions) and this raises the
question of what should be taken as a reference material: gliadin or whole
gluten. However, at present this issue is not relevant; it appears that the
antibodies do not tolerate the reducing agent mercaptoethanol at a concentration
of 0.01% and the more efficient extraction is traded off by this phenomenon.
A puzzling phenomenon is the difference between the antibody used in
this assay (against QQQPSP) and the R5 used by Sorell et al. (1998) and
Valdés et al. (2003) (against QQPSP) with respect to the difference in reactivity
to γ- and ω-gliadin and cereals like rye and barley. What a difference a Q
makes! A variety of commercially available gluten-free foods were analyzed
and small quantities of gluten were detected in some products.
Of course it would be feasible to design antibody-based assays against
other peptides as well. Fraser et al. (2003) investigated the toxicity of a
sequence containing residues 56–75 of α-gliadin, which proved to be toxic
in vivo. This peptide would consequently also be a candidate to design an
ELISA around. A problem not yet solved is: how many toxic motifs will be
found in the future and how should they be weighed in the final result to get
a realistic representation of toxicity?
14.3.5 ELISA of native, heat processed and modified gluten after

limited hydrolysis
To improve the extraction of gluten, a procedure with a limited hydrolysis
was investigated by Denery-Papini et al. (2000, 2002). Food or reference
material was hydrolyzed with pepsin, which led to about 90% protein
extraction in a saline buffer. Monoclonal and polyclonal antibodies were
used against repetitive motifs of α-, β- and γ-gliadins. In a competitive-
type assay, 35 ppm could be determined with the mAB and 150 ppm with
the pAB.
The assay was improved by using anti-γ-gliadin developed against the
repetitive domain of γ-gliadins. This domain was isolated and purified with
size-exclusion chromatography and reversed-phase HPLC and was found to
correspond to a polypeptide of 28 kD. The mAB reacted strongly with ω-
gliadins and with α-, β- and γ-gliadins and only weakly with LMW and
HMW glutenins. No difference in reaction between native and heated products
was found. Detection of hordeins was at same level as gliadins, secalins
could be detected at a level ten times lower, and avenins 100 times higher,
than gliadin. No reaction occurred with maize and rice. Sensitivity was
0.3 ug/ml, allowing determination at the 150 ppm level.
14.3.6 Other antibody-based methods

As a quantitation of the gliadin is necessary, given the proposed threshold
values mentioned in the Codex Alimentarius Standard, immunotechniques
which do not produce quantitative results like dot blot, counter-
immunoelectrophoresis or immunodiffusion are inadequate. Combining an
electrophoretic separation with immunodetection yields a method with very
powerful features which can be used conveniently to confirm the results of
ELISA tests. The method is very weak with respect to quantitation, as the
process of densitometric quantitation of the separated (immunochemically)
stained bands in relation to a standard is very complicated. Depending on
the electrophoretic method used, the band pattern is often very variable.
Acid polyacrylamide gel electrophoresis (PAGE) has been used frequently
in cultivar identification. In SDS-PAGE this variation is less, as the separation
is based on molecular weight, which is a more conserved attribute, but
even this type of electrophoresis often produces a cultivar-dependent band
pattern. SDS-PAGE with a general staining like Coomassie Blue has been
used extensively in breeding programs, especially because the HMW subunits
of glutenin can be solubilized and separated. In SDS-PAGE the sample is
prepared by heating in a buffer consisting of buffered SDS with a reducing
agent like mercaptoethanol or DTT. Almost all proteins of a food matrix,
even when heat-processed, will be solubilized, and consequently, this method
offers good potential to determine gluten in heat-processed samples. The
variable band pattern, however, quite often precludes an accurate quantitation,
although samples may be grouped into classes (e.g. contains between 100
Fig. 14.1 Allelic variance between cereal cultivars. (From left to right): lane 1–5
wheat, 6–8 rye, 9–11 barley, 12, 13 wheat gluten reference. SDS-PAGE (gradient gel
and immunochemical staining with a polyclonal anti-wheat gluten antibody).
and 150 ppm gluten). The technique is very efficient at the qualitative
level, i.e. it can be used to confirm the presence of gluten in a gluten-free
product with a high degree of confidence if the band pattern obtained with
the suspected sample matches that of a standard. Due to this cultivar
dependence an exact match is, however, very difficult to achieve (see
Fig. 14.1).
14.3.7 Comments on antibody-based techniques

The reference standard
There has been much debate about which standard should be used. Van
Eckert et al. (1998) showed that the results of ELISA methods varied
considerably when using different reference standards. This is not unexpected.
An antibody reacts with a set of epitopes in the food analyzed. If this set of
epitopes is different in term of abundancy and/or affinity of the individual
constituting components there will not be a match between the analyte and
the reference protein and, consequently, the result is either under- or over-
estimated. This mismatch differs from antibody to antibody, causing differences
between the various assays. The development of a standard produced from a
mixture of cultivars which are important in trade has the advantage of
minimizing this bias. However it is impossible to exclude it completely.
Format
The composition of the food is very relevant to the format of the assay. A
capture ELISA in sandwich format can be very specific, but it may miss a
small size peptide containing just one epitope. Designing antibodies against
specific peptide motifs and using these in competitive format might circumvent
these shortcomings. Quantitation might be impaired anyway in situations
where stoichiometry is disturbed and if there is no prior knowledge about
size of the peptides. Ideally (but hardly feasible), quantitation of toxic peptides
should be performed with a gliadin standard degraded to the same extent and
in the same manner (sites of hydrolytic cleavage and deamidation) as the
sample. Another option would be to include a predigestion of the sample to
level out differences between them and the reference protein. It remains
questionable whether this is feasible. A patent has been applied for using this
approach (Denery-Papini et al., 2001).
Specificity of antibodies
Increasingly, antibodies are mapped on the molecular level. However,
unexpected reactions may occur, dependent on binding on e.g. plastic surfaces
(Chirdo et al., 1995), temperature (Brett et al., 2002) or extraction method.
We have repeatedly found that antibodies which behaved quite specifically
in a sandwich or competitive assay produced unspecific reactions (a positive
response to maize, sorghum, millet and rice) when used in Western blotting.
Apparently cross-reacting protein fractions which are not extracted by alcoholic
solutions were extracted when using SDS buffer and a reducing environment.
Surprisingly, the antibody, which had been prepared with an ethanol-extracted
gliadin as immunogen and which had been absorbed to remove all unwanted
specificities (and which should thus be specific for gliadin), behaved
unspecifically when brought in contact with proteins extracted differently.
Detection of deamidated protein

Wheat gluten is sometimes solubilized by partial hydrolysis. The functional
properties of the protein are then modified: the protein becomes less
hydrophobic and consequently the functional properties of the protein are
changed, which opens the door to wider application of the protein into foods.
Although reactivity of the protein with antibodies sometimes remains partially
intact, the quantitative aspects are corrupted. Again, quantitation can be
achieved only if the appropriate reference material is available. Whether or
not modified protein has been added to a food may be detected by performing
SDS-PAGE, blotting and immunodetection (Janssen et al., 1994). An
immunoreactive smear is indicative for the presence of modified gluten;
however, the results can be interpreted only in a qualitative way and even
then precautions should be taken because severe heating may produce the
same effect.
14.4 Non-antibody-based techniques

A number of non-antibody-based techniques have been used in the past.
Some of these (nitrogen determination or luminescence) have become obsolete
now as their scope is limited and accuracy insufficient.
14.4.1 Polymerase chain reaction

PCR methods are increasingly used in food analysis. They share one significant
advantage: the analyte itself is amplified and, principally, extremely low
amounts of reporter DNA can be detected (see Chapter 7 for an in-depth
discussion of PCR methods). The first PCR method to determine wheat in
gluten-free products was described by Allmann et al. (1993). The assay
consisted of a double PCR: a eukaryote PCR, amplifying a 137 bp fragment
of a highly conserved eukaryotic sequence as a control for the condition of
the extracted DNA; and a wheat-specific PCR, amplifying a 109 bp segment
in the major repeat unit of the intergenic between the 25S and 18S wheat
ribosomal RNA genes. This intergenic segment was chosen because it is
present in a high number of copies, allowing sensitive detection which would
not be possible using a single copy gene. The eukaryotic PCR serves as a
control to detect whether false negative results did occur due to DNA
degradation. DNA extraction and visualization of the amplified bands were
carried out using standard methodology. The assay was specific for wheat:
rye produced a very weak signal. The sensitivity of this analysis was 1 pg
wheat DNA, which equalled 50 ng wheat. Several commercial products were
analyzed with this PCR method as well as with a commercial anti-ω kit as
developed by Skerritt. It proved that some thickening agents, such as dextrin,
carob starch and guar flour, were refractory to DNA extraction. Other processed,
heated or fermented food yielded good-quality DNA. Of all food tested,
tapioca, curry and malt extract didn’t produce a positive signal with the
eukaryotic PCR, while other food gave a positive signal. Comparison of the
results obtained with the ELISA kit and the PCR method showed agreement
in most cases; in some foods, however (instant soups), they were discordant.
There are many inhibitory substances present in DNA extraction reagents
used to extract a crude matrix like food which may disturb the amplification.
Rossen et al. (1992) found that the use of NaOH/SDS inhibited the reaction
significantly, while other components, such as detergents, lysozyme, NaOH,
alcohols and (not surprisingly) EDTA and EGTA, had a negative effect on
amplification. A modified DNA extraction was proposed by Szamos et al.
(1998).
As mentioned before, the toxicity of oats is debated. Data about its toxicity
(especially older data) are often corrupted by a high level of contamination
of commercial oat products (e.g. rolled oats) with wheat and/or barley. Using
ELISA, it is difficult to detect whether oats are contaminated with wheat,
because in most ELISA methods, oat itself shows a low, but distinct reaction.
Detection of a low degree of contamination by ELISA is consequently not

possible. Köppel et al. (1998) investigated the suitability of the PCR method
described above to detect wheat in oat products. With a slightly modified
DNA extraction protocol, the method proved to be very sensitive and (at
higher contamination levels) a good correlation was found between the results
of the ELISA and PCR.
More recently, a quantitative competitive PCR system to detect wheat,
barley or rye in gluten-free food was described by Dahinden et al. (2001).
This assay is based on simultaneous detection of wheat, rye and barley by
the amplification of a chloroplast trnL (UAA) intron. Good correlations were
found between the PCR and values obtained by ELISA. As stated by the
authors, deviations from an expected ratio between PCR signal and an immuno-
based response are indicative for wheat starch addition (PCR positive, ELISA
negative) or addition of gluten to a food (PCR low or absent, ELISA high)
and both cases warrant further investigation.
Real-time PCR using melting curve analysis (LightCycler® FastStart DNA
MasterPLUS SYBR Green 1 – Roche Diagnostics GmbH, Mannheim, Germany)
for identification of wheat, rye, barley and oats was recently described by
Sandberg et al. (2003). The length of the amplicons varied between 104 and
181 bp and the melting points between 81.2 and 85 °C. DNA was extracted
with the Dneasy® tissue kit (Qiagen, Hilden, Germany) according to
manufacturer protocol. Target sequences were chosen encoding the ω-gliadin,
ω-secalin, hordein and avenin sequences. Though these sequences have a
high degree of homology, it proved possible to design species-specific primers
by choosing at least one of the primers in the non-coding region of the gene.
Wheat species (spring and autumn wheat), durum wheat, spelt and kamut
(triticum polonicum) were all detected. Several products were analyzed and
a good correlation between the PCR assay and an ELISA was found. Of the
17 oat samples, nine were contaminated, some containing two different cereal
species as contaminants.
PCR methods offer an attractive alternative to ELISA methods in screening
or confirmation. As the technique is completely orthogonal to ELISA, a
positive result with both techniques gives almost irrefutable proof that the
incriminated cereal species is present. In some matrices it might be difficult
to extract DNA of good quality, e.g. thickening agents or emulsifiers. In
acid-treated food DNA is often degraded substantially. Thermal treatment of
the food poses no problem if short DNA sequences are amplified and
consequently the method is ideal to confirm ELISA findings in cooked
foods.
14.4.2 Maldi-TOF
A mass-spectrometric method was explored by Camafeita et al. (1997, 1998)
and Camafeita and Méndez (1998). In this method, the proteins of a food are
extracted and measured on a MALDI-TOF mass spectrometry (MS) system.
Mass spectra were recorded in a linear positive mode at 30 kV acceleration

voltage and 2 kV linear detector by accumulating 70 spectra of single laser
shots under threshold irradiance. Only highly intense, resolved mass signals
from 2–3 target spots were considered. The apparatus was calibrated with
bovine serum albumin. Patterns of several wheat, barley, rye and oat cultivars
were made and were compared to band patterns as revealed by SDS-PAGE
and stained with Coomassie Blue. The MALDI-TOF spectra from the four
wheat cultivars yielded very similar mass patterns in the 30–40 kD range,
which encompasses the molecular weights of α/β- and γ-gliadin. Mass spectra
from the other cereal species also showed very consistent patterns, but different
from each other, which allowed the unscrambling of mixtures. The authors
suggest that this method could be worked out to detect total prolamin from
various sources by measuring the mass profile in the 20–45 kD range. No
data with respect to non-cereal matrices were presented, however. As the
method itself is non-specific, the discrimination between cereals and non-
cereals depends completely on differences in mass pattern and on the selectivity
of the extraction solvent (in this case 70% ethanol). Compared to SDS-
PAGE (separation also based on mass) this method has the advantage of
speed. It is open to debate as to whether the intrinsic possibility to
measure oats at the same sensitivity as wheat, barley and rye is a substantial
advantage. If toxic at all, oats are most probably much less toxic than wheat,
rye and barley. The ability to measure wheat, rye and barley contamination
in oats by shifting the mass window to higher masses might prove to be
useful.
To improve the performance in the maize and rice matrices, the method
was modified by changing the extraction procedure (Hernando et al., 2003).
The extracts, as made in 60% ethanol, were dried and resuspended in 0.5 ml
1M acetic acid and mixed for 15 minutes at room temperature. The work
proved that the prolamins from rice and maize remained insoluble (confirmed
by taking a mass spectrum of the pellet) while gliadin went into solution
(which could be confirmed by preparing a mass spectrum of the supernatant).
Sensitivity of the method, however, was not high enough to allow determination
of gliadin at the 20 ppm level.
14.5 Selecting a method

The choice of a method strongly depends on the position of the laboratory
that is doing the analysis. Whereas, for example, an in-house quality control
laboratory would have insight into the formulation of the product analyzed,
its ingoing raw materials, additives and the way the food is processed, a
laboratory of an official food authority would not have direct access on a
regular basis to this information. Having access to these data greatly facilitates
the interpretation of the results as a ‘baseline’ response for the food analyzed
is known. Laboratories of official food authorities and external private
laboratories doing this type of analysis on an ad hoc basis have no knowledge

of the baseline response of their test material and would have to discuss
whether their results are indeed caused by gluten contamination or are just a
false positive result caused by non-specific binding of a quite unrelated
substance. Especially in case of trade disputes or litigation, it might be
imperative to analyse by two completely unrelated methods. Each method
should be validated with respect to trueness (recovery and absence of matrix
effects) as well as reproducibility. The latter can be achieved by participating
in collaborative trials and proficiency studies as organized, e.g. by any taking
part in a proficiency scheme [e.g. the rounds organized by the UK Food
Analysis Performance Assessment Scheme (FAPAS)]. A FAPAs round to
determine gluten in a dry bread mix was carried out in 2004.
Major flaws in current methodologies are summarized below.
• Methods produce results expressed as ‘wheat gluten equivalents’, because
wheat gluten (either the European reference material certified by IRMM
or any other commercially available material) is used in preparing calibration
curves. The actual amount of non-wheat gluten (barley, rye, secale, oats,
etc.) can relate to this result in a variable and irreproducible way.
• Even if, by using standardized methods, the reproducibility and repeatability
might be acceptable, there is no guarantee that the results are true. Only
when there is a perfect match between the epitopes of the contaminating
cereal (in terms of relative contribution) and the standard will the result be
true.
14.6 Future trends

Until now, methods to detect and determine wheat gluten and related proteins
from coeliac-toxic cereal species have focused on the determination of the
cereal species, without taking into account its toxicity. Presently, much research
is being carried out to determine the precipitating factor at the molecular
level, i.e. the toxic sequences. Work is also being done to produce non-toxic
wheat gluten expressed by yeast cells to be added to non-toxic cereals like
maize or rice to improve baking quality. In addition, there is much interest in
selecting wheat cultivars which lack the toxic sequences, either by screening
of gene databases for attributions naturally lacking these toxic sequences or
by removing those sequences by genetic engineering. If this strategy succeeds,
it will be a prerequisite for a method that it distinguishes between wheat with,
and wheat without, toxic sequences (the same will hold for other cereal species).
It is now generally accepted that coeliac disease is caused by inflammatory
T-cell responses to gluten peptides bound to HLA-DQ2 or -DQ8 molecules.
There is overwhelming evidence that coeliac disease patients can mount T-
cell responses to discrete amino acid sequences found in α- and γ-gliadin
and LMW and HMW glutenin.
Assays that would detect the presence or absence of such peptide motifs
would thus be accurate indicators of safety of the food for consumption by
coeliac disease patients. A novel method to detect such gluten peptides in
food has been developed at the Department of Immunohaematology and
Blood Transfusion of the Leiden University Medical Centre (LUMC), Leiden,
The Netherlands (Spaenij-Dekking et al., 2004). Monoclonal antibodies were
generated that are specific for T-cell stimulatory epitopes in α-gliadin, γ-
gliadin or LMW glutenin molecules. These mABs are used in competition
assays in which peptides present in the sample to be measured compete for
binding to the peptide-specific mAB with a biotinylated indicator peptide.
These assays are highly specific and detect up to 50 ng gluten per ml of the
European gliadin standard. This corresponds to detection of 0.5 ppm in
extracted gluten preparations. The advantages of the new method are multifold:
(i) the assay measures specifically the presence of T-cell stimulatory epitopes
known to be involved in the development of CD; (ii) as a consequence of the
format, this type of assay requires only one mAB per peptide, allowing the
detection of small gluten peptides corresponding to the size of T-cell stimulatory
epitopes – this is impossible with the currently available sandwich ELISA
systems; (iii) the assay detects gliadin and glutenin molecules simultaneously,
allowing a more fair judgement of the potential toxicity of the foods tested
for coeliac disease patients. This type of assay offers good possibility to
screen novel cereal sources, like tef (Eragrostis tef), or other cereals with
disputed coeliac toxicity. However, given the almost idiotypic reaction of
coeliacs, an antibody mimicking the reactivity of a panel of T-cells might be
necessary.
Without any doubt, another trend is the development of rapid screening
tests, by which a production lot can be screened rapidly for gluten contamination
without requiring access to laboratory facilities. In verifying the performance
of hazard analysis of critical control point (HACCP) systems an analysis of
the end-product is the ultimate method to check whether all critical points
during production are under control. At the time of writing, dipstick assays
are commercially available from Tepnel (Deeside, UK), R-Biopharm
(Darmstadt, Germany) and Operon (Cuarte de Huerva, Spain) based on the
immunochemical capture of coloured microparticles on a blotting medium;
however, only limited data about their performance in practice is available
yet. Drawbacks for home use include the need to extract the protein from a
(most often) solid food, which might be cumbersome and prone to mistakes;
however, these methods might have good prospects in HACCP programs.
A biochip is being developed in the framework of an EU funded research
project: Quantitation of coeliac disease toxic gluten in foodstuffs using a
chip system with integrated extraction, fluidics and biosensoric detection
(QLK1-CT-2002-02077) http://www.etseq.urv.es/BBG/dinamic/Cd-chef/. This
project, which is coordinated by the Department of Chemical Engineering,
Universitat Rovira i Virgili, Tarragona, Spain, includes the development of
immuno- and apta-sensor generic platforms with optical/electrochemical/
quartz crystal nanobalance detection. It is aimed at integrating extraction

and detection on a disposable microsystem with the time required for total
assay being less than 15 minutes at a cost of less than EUR 15, meeting the
product design requirements listed both for extraction and for detection. It
will be based on either antibodies against the toxic peptide motif or on
aptamers (i.e. synthetic RNA/DNA constructs) showing high affinity to the
toxic peptide motifs. This project is in its preliminary stages and no performance
data are yet available.

(Prospective) threshold values for gluten in gluten-free food in national food
legislations are increasingly derived from what has been agreed upon in
international institutions such as Codex Alimentarius.
Codex Alimentarius is a worldwide organization and its standards are the
basis of several national food legislations. Codex standards will have to be
implemented in national legislation in order to come into force. Important
committees within Codex are: the Codex Committee on Nutrition and Food
for Special Dietary Uses (CC NFSDU), the committee on Food Labelling
(CC FL) and the committee on methods of analysis and sampling (CC MAS).
Current discussions between producers of gluten-free food and consumer
organizations have not yet reached consensus about the proposed statutory
limit for ‘gluten-free’. Of course many producers prefer to produce under
relaxed conditions and favor a high limit value while in the eyes of the
consumer organizations, a low limit is to be preferred based on the precautionary
principle. Ultimately the limit will be a compromise between political (risk
management) and scientific (risk analysis) arguments. The reports of the
meetings of these committees as well of the commission itself are published
on the Codex Alimentarius website: http://www.codexalimentarius.net.
The Working Group on Prolamin Analysis and Toxicity (WGPAT) has
brought together a group of scientists of various disciplines, gastroenterologists,
cereal/food chemists, pediatricians, with an interest in coeliac disease. The
WGPAT has an observer status in Codex Alimentarius. This working group
organizes annual meetings, which are appreciated as a platform for
multidisciplinary exchange of ideas and may be attended by invitation.
Numerous national coeliac patient organizations exist around the world.
If one is interested in issues that are considered important in the coeliac
world one should visit one of the national coeliac sites. A list of these is
available at www.aoecs.org.
Another item to look at is the recent amendment on the European labelling
directive 2000/13/EG (discussed at length in Chapter 21 of this book). This
amendment makes the labelling of allergens (and also coeliac-toxic cereals)
compulsory. Depending on the outcome of research with respect to the coeliac
toxicity of so-called second and third generation products (i.e. products
derived from wheat starch) this amendment will be likely to have dramatic
effects on the labelling of many products. Wheat starch-derived products
like glucose syrup may be used as a raw material for the production of many
food ingredients and additives like polyols, ascorbic acid, acetic acid, etc. In
the near future it may thus become mandatory to label food containing these
ingredients as derived from wheat unless an exemption is allowed (wheat-
based glucose syrups and maltodextrins, glucose syrups based on barley and
cereals used in distillates for spirits are (by directive 2005/26/EC) provisionally
excluded from this mandatory labelling, because, most probably their gluten
content is very low). In the US and other countries, newer labeling laws
require, or will soon require, this type of labeling.
Regarding the gliadin standard: the reference material (IRRM-480)
developed on the initiative of the WGPAT is available from the IRRM in
Geel, Belgium. If one is aiming at producing reproducible results, the use of
a worldwide accepted standard is of utmost importance. The IRRM website
is http://www.irmm.jrc.be/.
14.8 References
Albrecht, E and Toth, A (1995) ‘Investigation of gliadin content of wheat flour by ELISA
method’, Acta Alimentaria, 24, 23–29.
Allmann, M, Candrian, U and Lüthy, J (1992) ‘Detection of wheat contamination in
dietary non-wheat products by PCR’, Lancet, 339, 309.
Allmann, M, Candrian, U, Höfelein and Lüthy, J (1993) ‘Polymerase chain reaction
(PCR) a possible alternative to immunochemical methods assuring safety and quality
of food. Detection of wheat contamination in non-wheat products’, Z Lebens Unters
Forsch, 196, 248–251.
AOAC official methods of analysis: supplement (1995) AOAC official method 991.19,
chap 32:13 [32.31.24].
Arentz-Hansen, H, Fleckenstein, B, Molberg, Ø, Scott, H, Koning, F, Jung, G, Roepstorff,
P, Lundin, KEA and Sollid, L M (2004) ‘The molecular basis for oat intolerance in
patients with celiac disease’, Plos Med 1, Epub, Oct 19.
Ayob, M K, Rittenburg, J, Allen, J C and Smith, C J (1988) ‘Development of a rapid
enzyme-linked immunosorbent assay (ELISA) for gliadin determination in food’,
Food Hydrocolloids, 2, 39–49.
Battais, F, Pineau, F, Popineau, Y, Aparicio, C, Kanny, G, Guerin, L, Moneret-Vautrin, D
A and Denery-Papini, S (2003) ‘Food allergy to wheat: identification of immunoglobulin
E and immunoglobulin G-binding proteins with sequential extracts and purified proteins
from wheat flour’, Clin Exp Allergy, 33, 962–70.
Brett, G M, Mills, E N, Bacon, J, Wellner, N, Husain, R D, Tatham, A S, Shewry, P R and
Morgan, M R (2002) ‘Temperature-dependent binding of monoclonal antibodies to C
hordein’, Biochim Biophys Acta, 1594, 17–26.
Camafeita, E and Méndez, E (1998) ‘Screening of gluten avenins in foods by matrix-
assisted laser desorption/ionization time-of-flight mass spectrometry’, J Mass Spectrom,
33, 1023–1028.
Camafeita, E, Alfonso, P, Mothes, T and Méndez, E (1997) ‘Matrix-assisted laser desorption/
ionization time-of-flight mass spectrometric micro-analysis: the first non-immunological
alternative attempt to quantify gluten gliadins in food samples’, J Mass Spectrom, 32,
940–947.
Camafeita, E, Solís, J, Alfonso, P, Lopez, J A, Sorell, L and Méndez, E (1998) ‘Selective

identification by matrix-assisted laser desorption/ionization time-of-flight mass
spectrometry of different types of gluten in foods made with cereal mixtures’, J
Chromatogr A, 823, 299–306.
Catassi, C, Rätsch, I-M, Fabiani, E, Rossini, M, Bordicchia, F, Candela, F, Coppa, G V
and Giorgi, P L (1994) ‘Coeliac disease in the year 2000: exploring the iceberg’,
Lancet, 343, 200–203.
Chirdo, F G, Añón, M C and Fossati, C A (1995) ‘Optimization of a competitive ELISA
with polyclonal antibodies for quantitation of prolamines in foods’, Food Agric Immunol,
7, 333–343.
Chirdo, F G, Añón, M C and Fossati, C A (1998) ‘Development of high-sensitive enzyme
immunoassays for gliadin quantification using the streptavidin-biotin amplification
system’, Food Agric Immunol, 10, 143–155.
Ciclitira, P J and Lennox, E S (1983) ‘A radioimmunoassay for alpha- and beta-gliadins’,
Clin Sci (Lond), 64, 655–659.
Ciclitira, P J, Evans, D J, Fagg, N L K, Lennox, E S and Dowling, R H (1984) ‘Clinical
testing of gliadin fractions in coeliac patients’, Cin Sci, 66, 357–364.
Dahinden, I, Von Büren, M and Lüthy, J (2001) ‘A quantitative competitive PCR system
to detect contamination of wheat, barley or rye in gluten-free food for coeliac patients’,
Denery-Papini, S, Boucherie, B, Larré, C, Pineau, F, Popineau, Y, Guéguen, J, Lauriere,
M, Bouchez, I and Delaunay, T (2000) ‘Extraction and immunochemical measurement
of raw and heated gluten in food’, Proceedings of the Working Group on Prolamin
Analysis and Toxicity, Merano, M Stern ed. 139–142.
Denery-Papini, S, Popineau, Y, Guéguen, J and Larré, C (2001) ‘Immunoassay for cereal
reserve protein’, Patent application WO 01/75451.
Denery-Papini, S, Boucherie, B, Larré, C, Pineau, F, Popineau, Y, Guéguen, J, Lauriere,
M, Bouchez, I and Delaunay, T (2002) ‘Measurement of raw, heated and modified
gluten after limited hydrolysis’, Proceedings of the 16th Meeting of the Working
Group on Prolamin Analysis and Toxicity, M Stern (ed), Zwickau, Germany, Verlag
Wissenschaftliche Scripten, 71–75.
De Ritis, G, Auricchio, S, Jones, H W, Lew E J-L, Bernardin, J E and Kasarda, D D
(1988) ‘In vitro (organ culture) studies of the toxicity of specific A-gliadin peptides in
coeliac disease’, Gastroenterology, 94, 41–49.
Dicke, W K (1950) ‘Coeliac disease. Investigation of the harmful effects of certain types
of cereal on patients suffering from coeliac disease’, Thesis, Utrecht 1950, 1st English
edition, April 1993.
Dieterich, W, Storch, W B and Schuppan, D, (2000) ‘Serum antibodies in celiac disease’,
Clin Lab, 46, 361–364.
Ellis, H J, Doyle, A P, Day, P, Wieser, H and Ciclitira, P J (1994) ‘Demonstration of the
presence of coeliac-activating gliadin-like epitopes in malted barley’, Int Arch Allergy
Immunol, 104, 308–310.
Ellis, H J, Rosen-Bronson, S, O’Reilly, N and Ciclitira, P J (1998) ‘Measurement of
gluten using a monoclonal antibody to a coeliac toxic peptide of A-gliadin’, Gut, 43,
190–195.
Fasano, A, Berti, I, Gerarduzzi, T, Not, T, Colletti, R B, Drago, S, Elitsur, Y, Green, P H,
Guandalini, S, Hill, I D, Pietzak, M, Ventura, A, Thorpe, M, Kryszak, D, Fornaroli, F,
Wasserman, S S, Murray, J A and Horvath, K (2003) ‘Prevalence of celiac disease in
at-risk and not-at-risk groups in the United States: a large multicenter study’, Arch
Intern Med, 163, 286–292.
Fraser, J S, Engel, W, Ellis, H J, Moodie, S J, Pollock, E L, Wieser, H and Ciclitira, P J
(2003) ‘Coeliac disease: in vivo toxicity of the putative immunodominant epitope’,
Gut, 52, 1698–1702.
Freedman, A R, Galfre, G, Gal, E, Ellis, H J and Ciclitira, J (1987) ‘Detection of wheat
gliadin contamination of gluten-free foods by a monoclonal antibody dot immunobinding

assay’, Clin Chim Acta, 166, 323–328.
Friis, S U (1988) ‘Enzyme-linked immunosorbent assay for quantitation of cereal proteins
toxic in coeliac disease’, Clin Chim Acta, 178, 261–270.
Garcia E, Llorente M, Hernando A, Kieffer R, Weiser H, Mendez E (2005) ‘Development
of a general quantitative gliadin extruction procedure for unheated and heat-processed
foods’, European Journal of Gastroenterology and Hepatology, 17(5), 529–539.
Gee, S (1888) ‘On the coeliac affection’, St. Barth’s Hosp Rep, 24, 17–20.
Henterich, N, Osman, A A, Méndez, E and Mothes, T (2003) ‘Assay of gliadin by real-
time immunopolymerase chain reaction’, Nahrung, 47, 345–348.
Hernando, A, Valdes, I and Méndez E (2003) ‘New strategy for the determination of
gliadins in maize- or rice-based foods matrix-assisted laser desorption/ionization time-
of-flight mass spectrometry: fractionation of gliadins from maize or rice prolamins by
acidic treatment’, J Mass Spectrom, 38, 862–871.
Immer, U and Haas-Lauterbach, S (2004) ‘Statistical evaluation of gliadin ring trial’,
Proceedings of the 18th Meeting of the Working Group on Prolamin Analysis and
Toxicity, M, Stern (ed), Zwickau, Germany, Verlag Wissenschaftliche Scripten, 23–36.
Immer, U, Vela, C, Méndez, E and Janssen, F W (2003) ‘PWG collaborative trial of
gluten in gluten-free food through “Cocktail ELISA” ’, Proceedings of the 17th Meeting
of the Working Group on Prolamin Analysis and Toxicity, M Stern (ed), Zwickau,
Germany, Verlag Wissenschaftliche Scripten, 23–33.
James, M W and Scott, B B (2001) ‘Coeliac disease: the cause of the various associated
disorders? Eur J Gastroenterol Hepatol’, 13, 1119–1121.
Janssen, F W, De Baaij, J A and Hägele, G H (1994) ‘Detection of modified gluten by
SDS-electrophoresis, Western blotting and immunochemical detection’, Die
Fleischwirtschaft, 74, 176–178 (in German), 168–170 (in English).
Klein, K, Franchini, F and Schimmel, H, (2003) ‘Towards a certified reference material
for ‘Gliadin from European wheat’: study design and methodological approaches’,
Toxicity, M Stern (ed), Zwickau, Germany, Verlag Wissenschaftliche Scripten, 113–
117.
Koning, F (2003) ‘The molecular basis of celiac disease’, J Mol Recognit, 16, 333–336.
Köppel, E, Stadler, M and Lüthy, J (1998) ‘Detection of wheat contamination in oats by
polymerase chain reaction (PCR) and enzyme-linked immunosorbent assay (ELISA)’,
Z Lebens Unters Forsch, 206, 399–403.
Louka, A S and Sollid, L M (2003) ‘HLA in coeliac disease: unravelling the complex
genetics of a complex disorder’, Tissue Antigens, 61, 105–107.
Maiuri, L, Ciacci, C, Ricciardelli, I, Vacca, L, Raia, V, Auricchio, S, Picard, J, Osman, M,
Quaratino, S and Londei, M (2003) ‘Association between innate response to gliadin
and activation of pathogenic T cells in coeliac disease’, Lancet, July 5, 362, 3–4.
Meier, P, Windemann, H and Baumgartner, E (1984) ‘Zur Bestimmung des A-Gliadin-
Gehaltes in glutenhaltigen und ‘glutenfreien’ erhitzten Lebensmittel’, Z Lebens Unters
Forsch, 178, 361–365.
Méndez, E, Vela, C, Immer, U and Janssen, F W (2005) ‘Report of a collaborative trial
to investigate the performance of the R5 enzyme linked immunoassay to determine
gliadin in gluten-free food’, Eur J Gastroenterol Hepatol, 17, 1053–1063.
Mills, E N C, Spinks, C A and Morgan, M R A (1989) ‘A two-site enzyme-linked
immunosorbent assay for wheat gliadins’, Food Agric Immunol, 1, 333–343.
Osborne, T B (1907) ‘The proteins of the wheat kernel’, Washington DC, Carnegie
Institute, Publication No. 84.
Osman, A A, Uhlig, H H, Valdes, I, Amin, M, Méndez, E and Mothes, T (2001) ‘A
monoclonal antibody that recognizes a potential coeliac-toxic repetitive pentapeptide
epitope in gliadins’, Eur J Gastroenterol Hepatol, Oct, 13, 1189–1193.
Paully, J W (1954) ‘Observation on the aetiology of idiopathic steatorrhoea: jejunal and
lymph node biopsies’, Br Med J, 2, 1318–1312.
Pengiran Tengah, D S, Wills, A J and Holmes, G K (2002) ‘Neurological complications

of coeliac disease’, Postgrad Med J, 78, 393–398.
Rossen, L, Norskov, P, Holmstrom, K and Rasmussen, O F (1992) ‘Inhibition of PCR by
components of food samples, microbial diagnostic assays and DNA-extraction solutions’,
Int J Food Microbiol, 17, 37–45.
Rumbo, M, Chirdo, F G, Fossati, C A and Añón, M C (1996) ‘Influence of thermal
treatment of food on the immunochemical quantitation of gliadin’, Food Agric Immunol,
8, 195–203.
Rumbo, M, Chirdo, F G, Fossati, C A and Añón, M C (2001) ‘Analysis of the effects of
heat treatment on gliadin immunochemical quantification using a panel of anti-prolamin
antibodies’, J Agric Food Chem, 49, 5719–5726.
Sandberg, M, Lundberg, L, Ferm, M and Malmheden Yman, I (2003) ‘Real time PCR for
the detection and discrimination of cereal contamination in gluten free foods’, Eur
Schofield, J D, Bottoml, R C, Timms, M F and Booth, M R (1983) ‘The effect of heat on
wheat gluten and the involvement of sulphydryl-disulphide interchange reactions’, J
Cereal Sci, 1, 241–253.
Seilmeier, W and Wieser, H (2003) ‘Comparative investigations of gluten proteins from
different wheat species. IV Reactivity of gliadin fractions and components from different
wheat species in a commercial assay’, Eur Food Res Technol, 217, 360–364.
Shewry, P R, Tatham, A S, Forde, J, Kreis, M and Miflin, B J (1986) ‘The classification
and nomenclature of wheat gluten proteins: a reassesment, J Cereal Sci, 4, 97–106.
Simonato, B, Pasini, G, Giannattasio, M, Peruffo, A D, De Lazzari, F and Curioni, A
(2001) ‘Food allergy to wheat products: the effect of bread baking and in vitro digestion
on wheat allergenic proteins. A study with bread dough, crumb, and crust’, J Agric
Food Chem, 49, 5668–5673.
Sjöstrom, H, Lundin, K E, Molberg, O, Korner, R, McAdam, S N, Anthonsen, D, Quarsten,
H, Noren, O, Roepstorf, P, Thorsby, E and Sollid, L M (1998) ‘Identification of a
gliadin T-cell epitope in coeliac disease: general importance of gliadin deamidation
for intestinal T-cell recognition’, Scand J Immunol, 48, 111–115.
Skerritt, J H (1985) ‘A sensitive monoclonal antibody-based test for gluten detection:
quantitative immunoassay’, J Sci Food Agric, 36, 987–994.
Skerritt, J H and Hill, A S (1990a) ‘Monoclonal antibody sandwich enzyme immunoassays
for determination of gluten in foods’, J Agric Food Chem, 38, 1771–1778.
Skerritt, J H and Hill, A S (1990b) ‘Homologies between grain storage proteins of different
cereal species. 2. Effects of assay format and grain extractant on antibody cross-
reactivity. J Cereal Sci, 11, 123–141.
Skerritt, J H and Hill, A S (1990c) ‘Monoclonal antibody sandwich enzyme immunoassays
for determination of gluten in foods’, J Agric Food Chem, 38, 1771–1778.
Skerritt, J H and Hill A S (1991) ‘Enzyme-immunoassay for the determination of gluten
in foods: collaborative study’, J AOAC, 74, 257–264.
Skerritt, J H and Martinuzzi, O (1986) ‘Effect of solid phase and antigen solvent on the
binding and immunoassay of water soluble flour proteins’, J Immunol Methods, 88,
217–224.
Skerritt, J H and Smith, R A (1985) ‘A sensitive monoclonal antibody based test for
gluten detection: studies with cooked or processed foods’, J Sci Food Agric, 36, 980–
986.
Skerritt, J H, Smith, R A, Wrigley, C and Underwood, P A (1984) ‘Monoclonal antibodies
to gliadin proteins used to examine cereal grain protein homologies’, J Cereal Sci, 2,
215–224.
Sorell, L, Lopez, J A, Valdes, I, Alfonso, P, Camafeita, E, Acevedo, B, Chirdo, F, Gavilondo,
J and Méndez, E (1998) ‘An innovative sandwich ELISA system based on an antibody
cocktail for gluten analysis’, FEBS Lett, 43, 46–50.
Spaenij-Dekking, E H, Kooy-Winkelaar, E M, Nieuwenhuizen, W F, Drijfhout, J W and
Koning, F (2004) ‘A novel and sensitive method for the detection of T cell stimulatory
epitopes of alpha/beta- and gamma-gliadin’, Gut, 53, 1267–1273.
Stern, M, Ciclitira, P J, Van Eckert, R, Feighery, C, Janssen, F W, Méndez, E, Mothes, T,
Troncone, R and Wieser, H (2001) ‘Analysis and clinical effects of gluten in coeliac
disease’, Eur J Gastroenterol, 13, 741–747.
Szamos, J, Aubrecht, E and Gelencsér, É (1998) ‘Detection of wheat by adapted polymerase
chain reaction (PCR methodology)’, Acta Alimenta, 27, 87–95.
Troncone, R, Vitale, M, Donatiello, A, Farris, E, Rossi, G and Auricchio, S (1986) ‘A
sandwich enzyme immunoassay for wheat gliadin’, J Immunol Methods, 92, 21–23.
Valdes, I, Garcia, E, Llorente, M and Méndez, E (2003) ‘Innovative approach to low-
level gluten determination in foods using a novel sandwich enzyme-linked
immunosorbent assay protocol’, Eur J Gastroenterol Hepatol, May, 15(5), 465–74.
Erratum in: Eur J Gastroenterol Hepatol, Jul, 15, 839.
Van de Wal, Y, Kooy, Y M, Van Veelen, P, Vader, W, August, S A, Drijfhout, J W, Peña,
S A and Koning, F (1999) ‘Glutenin is involved in the gluten-driven mucosal T cell
response’, Eur J Immunol, 29, 3133–3139.
Van Eckert, R (2002) ‘The PWG gliadin, a new reference material’, Proceedings of the
16th Meeting of the Working Group on Prolamin Analysis and Toxicity, M Stern (ed.),
Zwickau, Germany, Verlag Wissenschaftliche Scripten, 25–27.
Van Eckert, R, Scharf, M, Wald, T and Pfannhauser, W (1998) ‘Determination of proteins
with ELISA-Methods: Doubtful quantitative results?’ Proceedings of the 12th Meeting
of the Working Group on Prolamin Analysis and Toxicity, M Stern (ed.), Zwickau,
Germany, Verlag Wissenschaftliche Scripten, 35–40.
Van Eckert, R, Berghofer, E, Ciclitira, P J, Chirdo, F S, Denery-Papini, S, Ellis, H J,
Ferranti, P, Goodwin, P, Immer, U, Mamone, G, Méndez, E, Mothes, T, Novalin, S,
Osman, A, Rumbo, M, Stern, M, Thorell, L, Whim, A and Wieser, H (2005) ‘Towards
a new gliadin reference material – isolation and characterisation’, J Cereal Sci, submitted.
Vestergaard, P (2003) ‘Bone loss associated with gastrointestinal disease: prevalence and
pathogenesis’, Eur J Gastroenterol Hepatol, 15, 851–856.
Weegels, P L, De Groot, A M G, Verhoek, J A and Hamer, R J (1994a) ‘Effects on gluten
of heating at different moisture contents. II Changes in physico-chemical properties
and secundary structure’, J Cereal Sci, 19, 39–47.
Weegels, P L, De Groot, A M G, Verhoek, J A and Hamer, R J (1994b) ‘Effects on gluten
of heating at different moisture contents. I Changes in functional properties’, J Cereal
Sci, 19, 31–38.
Wieser, H (1998) ‘Investigations on the extractability of gluten proteins from wheat
bread in comparison with flour’, Z Lebens Unters Forsch, 207, 128–132.
Wieser, H, Belitz, H-D, Ashkenazi, A and Idar, D (1983) ‘Isolation of coeliac active
peptide fractions from gliadin’, Z Lebens Unters Forsch, 176, 85–94.
Wieser, H, Seilmeier, W and Belitz, H-D (1994) ‘Quantitative Determination of Gliadin
Subgroups from different Wheat cultivars’, J Cereal Sci, 19, 149–155.
Windemann, H, Fritschy, F and Baumgartner, E (1982) ‘Enzyme-linked immunosorbent
assay for wheat a-gliadin and whole gliadin’, Biochim Biophys Acta, 709, 110–121.
15
Detecting soy, fish and crustaceans in

food
S. Koppelman, AllerTeQ and University Medical Centre Utrecht,
The Netherlands, and S. Hefle, University of Nebraska, USA
15.1 Introduction
Detection of peanuts, tree nuts, wheat, milk and egg as members of the ‘Big
Eight’ allergenic foods is described in the preceding chapters. In this chapter
we describe the tests available for the remaining allergens: soy, crustaceans
and fish. Compared to the other food allergens, detection of the present
group of allergens is relatively unexplored, due to several reasons related to
these particular foods. Soy is a food stuff that is used in a vast number of
applications, undergoing a wide variety of processing steps, with consequences
for the detection of allergenic residues. Crustaceans and fish count a large
number of species, and for detection purposes it is necessary to choose
whether to test for one species, for a subset or for the entire group.
Considerations for such choices are based on the allergic patient’s
characteristics: do they react to one of the species, or even one allergen, or
to more species and more allergens? From the food industry perspective, the
issue is to detect fish and/or crustacean residue, regardless of the species and
the nature of the allergens. The sections below will cover the published and
commercially available methods for soy, fish and crustaceans, and will discuss
the advantages and disadvantages of the chosen approaches.
15.2 Soy
Soy is a food ingredient with both technofunctionality and biofunctionality.
The beneficial effects on health of soy isoflavones have been recently reviewed,1
and food products containing 25 g of soy per intake may claim to be healthy
(see US Food and Drug Administration (FDA) regulations at www.FDA.gov).

Soy is used in a number of applications, as consumer product diversification
since the 1980s has required improved technological functionality of
ingredients. In terms of authenticity and product adulteration, such as for
high-quality meat products,2 there is a need for methods to detect the presence
of soy protein in foodstuffs. Many methods were published in the 1980s,
when the use of soy in foods expanded enormously. Test requirements then
related to specificity, with respect to other food ingredients, mainly meat,
and recovery for processed foods. The sensitivity was of less importance; 0.1
or 1% (w/w) of soy was a suitable detection limit. With food allergies on the
increase since the 1990s, it has become increasingly clear that only minute
amounts of soy, in particular soy protein, may threaten soy-allergic consumers.
Since the minimal provoking doses of food allergens are usually low, in the
high μg to mg level, the sensitivity of soy tests is now an important issue. For
peanut, another member of the Legumae family, thresholds for subjective
reactions of 1003 and 30 μg4 have been described. It is expected that the
threshold for objective symptoms and more severe reactions will be significantly
higher,5 in the 1–3 mg range, as demonstrated by several case reports on
inadvertent intake of peanut-containing foods. There is only limited information
on the threshold doses of soy protein for soy-allergic individuals. In fact, a
multi-centre study coordinated by the Food Allergy Research and Resource
Program at the University of Nebraska (www.farrp.org) to investigate the
minimal provoking dose for soy flour has recently commenced. Some case
reports show that minor amounts of soy, well below 0.1 and 1% (w/w), can
lead to allergic reactions.6,7 This section aims to review tests that have been
described in the literature for the detection of soy, and to discuss advantages
and disadvantages of the different methods.
15.2.1 Soy protein solubility and consequences for selecting a

detection method
Foods that contain soy proteins are, usually, subjected to some form of heat
treatment (e.g. pasteurization, sterilization) during processing to obtain (i) a
certain shelf-life, (ii) an inactivation of the anti-nutritional properties, and
(iii) certain functional properties. Proteins generally aggregate as a result of
heat denaturation and this is also observed for soy proteins. Solubility data
on soy proteins are plentiful, but their value in obtaining an overview is
frequently limited because of the large variation in heat treatment given,
soybean variety used, conditions of determination (pH, ionic strength), solvent
used, etc. It is known that these factors all affect solubility.8 The solubility
behavior of the major proteins in soy (glycinin, beta-conglycinin and trypsin
inhibitors) was recently reviewed.9 In general, native soy protein is poorly
soluble at neutral pH, while either lowering (< pH 4) or increasing (> pH 8)
increases solubility.8 For heat-treated or otherwise processed soy protein,
adjusting the pH will not be sufficient for solubilizing the protein fraction.
Detecting soy, fish and crustaceans in food 275
Denaturing extraction buffers containing high concentrations of urea or

guadinium, or disulphide bridge-reducing agents like beta-mercaptoethanol,
detergents like sodium dodecyl sulphate (SDS), or combinations thereof
may be necessary to obtain satisfactory solubility levels.10–16 Recent work
showed that at extremely high pH most commonly used soy ingredients are
soluble to a large extent.9 Only some ingredients are soluble at neutral pH,
under native conditions. Harsh conditions are required to substantially improve
solubility, particularly for processed ingredients (isolates, texturized soy).
Considering the difficulties with soy protein solubility, it could be suggested
that the protein fraction may not be the optimal analyte for testing for the
presence of soy. DNA-based methods may also be used to detect soy. However,
it should be kept in mind that the protein fraction causes allergic responses,
and the presence of DNA does not necessarily coincide with that of the
protein fraction. This is especially important when processed products, like
refined soy oil and lecithin, and fermented products, such as soy sauce, are
considered.
15.2.2 Detection of soy proteins

Many different approaches for detection of soy proteins and other soy
components in meat have been reported and reviewed.15,17 Most tests were
developed for measuring comparatively larger quantities of soy protein
(percents of total weight) in food products in order to detect the use of soy
as a protein fortifier/binder. There are only a small number of tests developed
for the detection of traces of soy cross-contamination that threaten soy-
allergic consumers. These are mainly immunochemical tests based on antibody–
antigen interactions, and are usually more sensitive when compared to chemical
or biochemical methods. The different techniques used for soy detection are
summarized below.
With regard to the safety of foods for soy-allergic consumers, the sensitivity
of the detection method is of utmost importance. It is clear that concentrations
well below 0.01% (100 ppm) for some allergens can cause allergic reactions,
although this has not been proven for soy to date. For peanut, another member
of the Legumae family, minimal doses that elicited subjective reactions of
304 and 100 μg3 have been described. This corresponds to 0.3–1 ppm when
a serving size of 100 g is considered. For soy, there are no accurate data on
minimal provoking doses. Based on information on threshold levels of other
food allergens, and taking into account case reports, Taylor et al.18 postulated
that a detection limit of 10 ppm is a useful figure. When threshold information
for soy becomes available, this number should be reconsidered in light of the
new data.
Detection of soy components other than protein; indirect analysis

Since direct protein measurement is sometimes difficult due to solubility
issues, indirect methods based on appropriate non-protein markers associated
with soy may be useful. Suitable markers have been chosen for their stability
and ease of assay under the processing conditions used. Examples of markers
that have been used for assay development include insoluble polysaccharides,
oligosaccharides, protein-bound sugars, free amino acids, free peptides, phytate,
saponins, sterols and metals.15 Another approach has used histological
techniques. A microscopic examination of calcium oxalate crystals and plant
cell components indicate the presence of soy. Staining for cell wall
polysaccharides could also be useful in this respect.15 The detection levels of
these methods are too high for quantification of traces of soy for allergenic
concerns.
Meyer et al.19 evaluated a DNA-based PCR method for detecting soy
residue in food products, and achieved a detection limit of 70 ppm. More
recent public discussion on genetically modified organisms (GMO) and foods
originating from such crops led to the development of assays for GMO soy,
with a reasonable sensitivity (for GMO concerns; 0.9%, EU legislation, for
more information see http://europa.eu.int/comm/food/fs/gmo/gmo_
ongoinit_en.html). Methods based on DNA, most often polymerase chain
reaction (PCR) tests, are applied to distinguish between wild-type and GMO
soy. As a test control, a gene present in wild-type soy was detected as well.
Van Duijn et al.,20 for example, measured the soy lectin gene as a positive
control. This test has also been used to detect soy in processed food with a
reported detection limit of 0.01% (100 ppm).
The sensitivities of the described indirect methods are between 0.01%
and 1% (100–10 000 pm). These are appropriate for detecting soy added as
a functional ingredient, but not for detecting traces originating from cross-
contamination. Besides this practical disadvantage, there is a theoretical
concern. In attempting to protect soy-allergic consumers, the allergic
component, i.e. protein, should be measured rather than a marker. This is of
key importance because industrial soy processing results in a variety of soy
protein products that may have lost other (marker) components during
processing.
Antibody-based tests
Enzyme-linked immunosorbent assay (ELISA) is a powerful analysis tool
for detection of specific proteins. It has the advantage of simultaneously
testing a larger series of samples at a high level of sensitivity. A large number
of investigators have used ELISA methods for the detection of soy protein in
meat products. Immunochemical methods are generally limited to the qualitative
screening of raw or mildly processed products, since protein denaturation
often alters the antigen–antibody interaction.15 This has been studied by
different authors15,21–32 and the most important results are discussed below.
Antibodies against soy can be raised in laboratory animals, usually rabbits,
mice, goats, sheep or donkeys, as described in Chapter 3. Individual animals
may react differently to immunization protocols, and it is difficult to predict
which species is the best host for raising antibodies against soy. In addition,
there is another issue that influences the success of raising antibodies against
soy. Soybean protein is a common substance in animal food, so the antigenic
responses obtained after immunization are not expected to be high. This may
lead to antibodies with low affinity and, consequently, ELISAs constructed
with such antibodies have poor sensitivities. This may be overcome by using
animals kept on soy-free diets.24 Aiming for a sensitive method to detect soy
protein, it seems to be logical to choose glycinin or β-conglycinin as antigenic
protein since these are the major soy proteins and they are abundantly present
in soy. However, using a single-allergen method creates the risk of false
negative test results in the case that the analyte is not present or does not
react while other allergens may still be present.
ELISAs for soy glycinin

Ravenstein and Driedonks11 raised antibodies against the SDS-denatured
acidic polypeptides of glycinin in rabbits. An ELISA constructed with these
antibodies had a reported detection limit of 0.5% (5000 ppm). Applying the
antibodies in an immunoblotting experiment decreased the detection limit
enormously, but this was not considered relevant since soy is usually used in
concentrations at 1% and higher. Meisel27 immunized chickens with SDS-
denatured glycinin and obtained specific IgY antibodies from egg yolk. With
these antibodies, an ELISA with a sensitive range between 0.5 and 256 μg/
ml for dilutions of soy protein isolate was constructed. The detection limit
was 0.5 μg/ml, corresponding with 0.1% soy protein spiked in a food matrix.
IgY antibodies used in immunoblotting even allowed the detection of the
acidic polypeptide of glycinin at nanogram levels.27 Plumb et al.28 raised
antibodies against native glycinin. They found that heating glycinin at pH
7.6 caused the immunoreactivity to decrease to around 50% of the original
value but it increased sharply above 92 °C. These results are in agreement
with Demonte et al.30 but differ from the results of Iwabuchi and Yamauchi,22
who found that the antigenic activity disappeared after heating. Huang et
al.31 found that the epitope identified by Plumb et al.28 corresponded to
residues 86–104 of the acidic polypeptides of glycinin A1aB1b and lies on
the C-terminus of the proteolytic intermediate known as glycinin-T. The
epitope seems to be continuous in nature. Iwabuchi and Yamauchi22 studied
the effect of ionic strength (I) on thermal denaturation of soybean glycinin.
Up to I = 0.7, no effects of ionic strength were found. A reduction in
immunogenicity also occurred when glycinin was taken to pH 2.0 and pH
11.0 and exposed to high temperatures.21
ELISAs for soy beta-conglycinin

Plumb et al.29 raised antibodies that were specific for the native α and α′
subunits of β-conglycinin. β-conglycinin immunoreactivity was increased as
the protein was heated, reaching a maximum at the denaturation temperature
of 65 °C. This phenomenon is unusual as most thermally denatured proteins
have low immunoreactivity when probed with antibodies raised against native
protein. This was also found to occur at pH 7.6 at different ionic strengths.21,33
The epitope of the antibody used by these researchers corresponds to the
residues 78–84 in the acidic extension present in the α′ subunit of β-
conglycinin,29 and seems to be continuous in nature.31 A linear epitope, in
contrast to a conformational epitope, is expected to be more heat stable and
may become more exposed after denaturation. This may explain the increased
detectability of heated β-conglycinin.
ELISAs for soy trypsin inhibitors

Brandon et al.23 worked with antibodies raised against native and heat denatured
soybean Kunitz trypsin inhibitor.23 Antibodies raised against the denatured
inhibitor were specific for the denatured conformation of the protein. In
contrast, the native inhibitor elicited antibodies that selectively recognized
determinants in both native and heat-treated proteins. Barkholt et al.25 raised
an antibody against Kunitz-type soybean trypsin inhibitor that worked equally
well against native and denatured protein. Because Kunitz-type soybean
trypsin inhibitor is an anti-nutritional factor, soybeans lacking this protein
were bred.26 At present, there are no commercial cultivars lacking the Kunitz-
type soybean trypsin inhibitor, but these may become available in the future.
From that perspective, using the Kunitz-type soybean trypsin inhibitor as a
marker for residual soy protein is not a good choice. Brandon et al.24 used
antibodies that bind the Bowman-Birk protease inhibitor. They identified an
epitope that was destroyed by heat and developed an ELISA for specific
recognition of native Bowman-Birk inhibitor in the presence of denatured
forms. None of these tests were optimized for sensitivity.
ELISAs for total soy protein

Several methods use antibodies raised not against one particular protein but
against the whole soy protein fraction. For example Janssen et al.34 detected
soy proteins in meat products up to 0.1% by gel electrophoresis followed by
blotting and dot blot. All major soy fractions were recognized by the antibody.
Hitchcock et al.15 raised antibodies against soy protein which had been
‘renatured’ from a hot urea solution by removing or diluting the denaturant.
This method could recognize heat-treated soy fractions as well as native
ones. Some specialized products, however, gave little or no response,
presumably because the proteins added were hydrolyzed (it should be noted
that hydrolyzed soy proteins can retain allergenicity). It was found that the
method of Hitchcock et al.15 could also be used with commercially available
anti-soy antibodies35 and a commercially available soy protein assay kit
(Cortecs, presently Tepnel Biosystems, Deeside, UK) was developed. The
antibody that was used is not specified in the kit insert. This test is meant for
quantitation of soy protein levels between 1 and 10% of total weight of raw
or processed mixed meat products. The relatively high detection limit is
partially due to the fact that the extracted samples need to be diluted in order
to prepare an environment that is compatible with ELISA. In this case the
required dilution factor was 20. Recently a method with a high sensitivity
(around 1 ppm) was described.9 This method uses an easy extraction protocol,
a one-step extraction at pH 12. Table 15.1 shows that this extraction has high
recoveries for a variety of different soy ingredients as compared to extraction
under native condition or extraction with urea and reducing agents. A mixture
of soy protein isolate, soy protein concentrate and texturized soy protein was
extracted at pH 12, and the extract used for raising antibodies in goats and
rabbits. A classical protocol for immunization was used and, prior to that, the
animals were kept on a soy-free diet for at least three months. An inhibition
Table 15.1 Extractability of processed soy ingredients using three different methods
Sample description Type Extraction recovery

(% related to product specification)
TRIS pH 8.2 Urea/DTT pH 12
Soy 200/90 Meal 39 52 67

(own supply)
light toasted
Central Soya Flour2 46 45
Centex 4220
Texturized Soy
Protein
Central Soya Isolate 39 43 60
Pro Fam 981
ADM Pro Isolate 33 46 51
Fam 891
ADM Pro Isolate 28 36 37
Fam 781
Central Soya Isolate 28 46 66
Pro Fam 646
ADM Arcon S Concentrate 34 47 64
ADM Arcon Concentrate 41 50 67
SM 066-405
ADM Arcon F Concentrate 3 48 52
Central Soya Concentrate 5 35 52
Response
4400 Textured
Procon 2000
Promine
DS Functional
ADM Arcon T Flakes ND ND ND
Mean (± SD) 20 (± 15) 45 (± 5) 55 (± 10)
36
Protein concentration measured by Bradford method.
2500
2000
1500
mAbs
1000
500
0
0.0001 0.001 0.01 0.1 1 10 100 1000 10 000
Soy flour concentration (μg protein/ml)
Fig. 15.1 Standard curve for soy protein ELISA. (ELISA plates were coated with pH
12-extracted soy and incubated with mixtures of rabbit antibody against pH 12-
extracted soy with calibrator (pH 12-extract of lightly toasted soy flour).
ELISA using rabbit antibodies was constructed and the detection limit is
0.02 μg/ml (Figure 15.1) which corresponds to 0.4 ppm for solid food samples.
Table 15.2 shows that the recovery of this test is high (mean 73%) and that
the variation between individual samples is only moderate (35%). These
characteristics are promising as they allow soy detection with sensitivity
Table 15.2 Recovery of soybean detection
Type of soy Percent measured Listed on product Recovery

soy protein1 specification sheet (%)
Meal Used as standard – –

Texturized Flour 35 54 65
Isolate 87 90 97
Isolate 79 90 88
Isolate 27 90 30
Isolate 123 90 136
Concentrate 61 72 86
Flakes 36 Not specified –
Mean (± SD) – – 73 (± 35)
1
Protein concentration measured by Bradford method36
Table 15.3 Commercially available methods to detect soy
Analyte Type of Kit Sensitivity Web site

method manufacturer
Soy protein Inhibition Tepnel 5000 ppm www.tepnel.com

ELISA
Soy trypsin Sandwich Elisa 1 ppm www.
inhibitor ELISA Systems elisasystems.net
Soy lectin Real-time Congen 10 ppm www.congen.de
gene PCR
Soy protein Sandwich Neogen (under 1–5 ppm www.neogen.com
ELISA development)
relevant for soy-allergic patients. Assay validation is now in progress and

commercialization is expected in 2005 (Neogen Corporation,
www.neogen.com).
Available methods for soy detection

Since the 1990s several tests to detect soy in food have become commercially
available. Most of these are ELISAs, but more recently DNA-based methods
have been launched. Table 15.3 gives an overview of the available methods
together with their most important characteristics. The ELISA for total soy
protein from Tepnel is meant for detecting large additions of soy protein in,
for example, meat products (p 278). The relatively high detection limit is not
suitable for allergen detection. Recently, ELISA Systems Ltd launched a
new soy protein screening assay. The detection limit is 1 ppm, and it is
therefore well suited to allergen detection. The kit is suitable for screening
the presence of soy protein and is based on the detection of Kunitz-type
soybean trypsin inhibitor. It is not applicable for samples that have been
subjected to extensive hydrolysis, fermentation and heat treatment (ELISA
Systems, personal communication). As mentioned above, if commercially
available cultivars lacking the Kunitz trypsin inhibitor were to become available,
this type of method would no longer be useful.
15.3 Crustaceans
Shrimp is the most commonly consumed crustacean shellfish and causes
severe allergic reactions, but other members of the Crustacean family such
as lobster, crab and crawfish can also elicit severe reactions.37,38,39,40 Estimates
of the number of shrimp-allergic individuals range from 0.6–2.8% of the
population. In the US, allergy to crustacean shellfish is more prevalent than
peanut allergy.
15.3.1 Crustacean allergens

The major allergen of shrimp is tropomyosin, a muscle protein41 which is
readily isolated from the boiling water and meat of cooked shrimp. Monoclonal
antibodies directed against this 36 kD shrimp allergen reacted to a 36 kD
protein in crayfish, crab and lobster extracts41. The allergen nomenclature
for this protein is Pen a 1. The molecular weight of various crustacean
tropomyosins has been reported as 34–39 kD; this protein is the target of
some of the detection methods that will be discussed next. At least 80% of
shrimp-allergic subjects have IgE-binding to tropomyosin42, making it a
major allergen. Another shrimp allergen (Pen m 2) was identified as having
extensive similarity to crustacean arginine kinase and to possess arginine
kinase activity41. Yu et al.42 made specific polyclonal antibodies to Pen m 2
and also developed a competitive ELISA using shrimp-allergic patient sera
to show that Pen m 2 was a cross-reactive crustacean allergen. In addition, a
yet-unidentified and unnamed allergen of 16.5 kD has been reported to cause
IgE-binding in a significant majority of shrimp-allergic subjects in one study42.
Many other minor IgE-binding proteins have been described in several reports,
but their significance is unknown.
15.3.2 Cross-contamination and cross-reacting allergens

The seafood processing, soup and frozen snack and meal industries have the
most interest in detecting the undeclared presence of shrimp residues in
other foods. Of concern is the transfer of shrimp allergens through frying oil
in restaurants and in battering operations in the food industry, and also in the
manufacture of shrimp-containing flavors and other ingredients on shared
lines with non-shrimp ingredients. There are also significant reactions reported
as a consequence of occupational exposure to aerosolized crustacean
allergens. 44,45 In some cases, cross-reactivity was found using IgE
immunoassays between shrimp tropomyosin and non-crustacean shellfish
(in a restaurant seafood handler and also in seafood processing workers),
even though these shellfish are of different phyla.46,47
15.3.3 Detecting crustacean allergens

There are few methods that have been described for the detection of crustacean
residues. ELISA or radioallergosorbent testing (RAST) utilizing sera from
shrimp-allergic patients have been used for research purposes and to investigate
and characterize crustacean allergens, but only two methods have been reported
to date to quantitate shrimp residues. Jeoung et al.48 made monoclonal
antibodies to shrimp tropomyosin and developed a sandwich ELISA for the
purpose of quantitating the protein in commercial skin test extracts. In these
clinical extracts, the ELISA could detect 4–125 ng tropomyosin/ml. In a
method designed for detecting shrimp in foods, Ben Rejeb et al.49 described
an ELISA made with antibodies against shrimp tropomyosin that possessed
a detection limit of approximately 2.5 ppm. However, up to now, this method

has been described only in the abstract and without many details, but it was
apparently used to determine the presence of tropomyosin in some food
matrices. There is one crustacean ELISA method on the market currently,
made by ELISA Systems Pty Ltd, Windsor, Queensland, Australia (http://
www.elisas.com.au). The claimed limit of detection of this sandwich
tropomyosin assay is 0.05 mg/kg (0.05 ppm). Although the assay is positioned
as a screening assay for tropomyosin, the reactivity of tropomyosin from
crustaceans other than shrimp is not shown. While some efforts to raise
specific polyclonal or monoclonal antibodies have resulted in antibodies that
have high specificity for the species of interest,48,49 one group reported
producing a monoclonal antibody against abalone tropomyosin that had
surprisingly high reactivity to chicken and crustacean tropomyosins.50
Therefore, any anti-tropomyosin antibody/detection method should be checked
against various types of tropomyosins for possible cross-reactivity issues. In
particular, due to their being in the same phyla, anti-crustacean tropomyosins
should be checked for cross-reactivity with insect tropomyosins. Sometimes,
storage mites contaminate food ingredients such as breading materials, and
this has the potential to lead to cross-reactions from the mite tropomyosin.
Tropomyosins from different species/phyla do share highly conserved regions
but, interestingly, crustacean-allergic subjects do not react to chicken or
other non-seafood tropomyosin, which means that the IgE-binding sites are
somewhat specific.
There is also one report in which direct ELISA and dot-immunoblotting
was used to detect crab residues in heated and sterilized surimi-based products.51
Polyclonal anti-lobster arginine kinase antibody was utilized and resulted in
an ELISA with a detection limit of 10–25 g of crab per kg of surimi-based
product; however, this detection limit is too high for allergen detection
requirements. Work on detection systems for other crustacea besides those
mentioned above has not been reported to date.
15.4 Fish
Fish allergy is relatively common, especially in geographic regions where
fish is an important part of the diet. Fatal reactions have occurred as a result
of allergic reactions to fish.39,52 Numerous studies have attempted to estimate
the prevalence of fish allergy, usually ranging from 0.3–0.5%.38,53,54 Although
the exact prevalence of fish allergy is not well established, it is listed among
the most common of all food allergens.55 Fish can also cause occupational
reactions. While asthma is the most prominent symptom, anaphylaxis has
been reported, and physical contact with seafood can also elicit symptoms in
allergic individuals.56,57
15.4.1 Fish allergens

The best-characterized fish allergen is the major allergen of codfish, named
Gad c 1, which belongs to a group of muscle tissue proteins known as
parvalbumins58. Parvalbumins are responsible for mediating the concentration
of calcium in muscle. Gad c 1 is a member of the parvalbumin family found
in muscle tissue of amphibians, fish and other animals. Parvalbumins are
found in high amounts in lower vertebrates and in smaller amounts in higher
vertebrates.59 Gad c 1 is an acidic protein of 12 kD which is stable to heating,
extremes of pH and mild proteolysis.60 All codfish-allergic individuals react
to Gad c 1.61,62 Another codfish allergen was identified, in an in vitro study
with cod-allergic patient sera, to have a molecular size of 41 kD, in addition
to six other bands at 13, 21, 27, 47 and 58 kD. The 13 kD band was a major
allergen for all subjects in the study, and the 13 and 41 kD bands were
recognized by anti-parvalbumin monoclonal antibodies, as were bands at 28
and 49 kD. Although the 41 kD protein was dissimilar to Gad c 1, cod-
allergic sera demonstrated IgE binding to this protein.63 In another study
investigating IgE binding in sera of fish-allergic patients, pre-rigor cod muscle
was found to contain IgE-binding bands of 12, 22, 30, 45, 60, 67, 104 and
130 kD.64 A major IgE-binding band for all sera was the 12 kD band; the 30
and 67 kD bands also possessed good IgE-binding activity. It was interesting
to note that when the cod had been dead for several days, new IgE-binding
bands of 18, 41 and 80 kD appeared, and the relative content of IgE-binding
proteins was increased. The action of naturally occurring proteases could
have cleaved the allergen and opened up sections of the peptides making
them more accessible to IgE-binding during this period. Anti-codfish
parvalbumin monoclonal antibody was also found to bind to all IgE-binding
bands identified in this study, indicating that there was shared similar structures
in all peptides.
While many fish-allergic patients are allergic to multiple fish species,
monospecificity has been reported to tuna and cod65, and swordfish.66 In
these studies, subjects with multiple allergies to fish species showed IgE
binding to proteins at 12–13 kD bands, while the monospecifically-sensitive
subjects showed IgE-binding to unique bands at 40 kD in tuna and 25 kD in
swordfish.
15.4.2 Pan-allergens among fish species

The existence of structurally-related parvalbumins in divergent fish species
may offer an explanation of cross-reactivity to fish species in certain allergic
individuals. Hansen et al.62 noted that in adults with clinical sensitivity to
cod, reactions were reported to mackerel, herring and plaice, and in vitro
results showed IgE-binding to a single band in the 11–14 kD region of these
fish extracts. An IgE-binding protein from salmon, named Sal s 1, was identified
by in vitro analyses, and was shown to be salmon parvalbumin.67 In one
study, almost 50% of cod-allergic patients had clinical sensitivity to salmon68.
Sal s 1 is described as having two bands, at 12 and 14 kD, and therefore has
at least two isotypic variants. The allergens from other fish species are also
parvalbumins including horse mackerel (Tra j 1) and bigeye tuna (Thu o
1).69–71 While tuna-allergic individuals have been noted to have IgE-binding
to Gad c 1, one study noted that raw tuna extracts seemed to lack IgE-
binding bands in the parvalbumin size range that were present in extracts of
catfish, cod and snapper.65 However, Park et al.72 described IgE-binding to
tuna proteins at 12–13 kD using sera from fish-allergic persons and IgE-
binding bands at 12–13 kD in mackerel, pollock and cod. Also, unique
binding to a band at 19 kD from saury and 37 kD from tuna were observed.
In one study, canned tuna did not present problems for five fish-allergic
children;73 canning may decrease the allergenicity of cod. However, amino
acid sequencing has shown that the 12–13 kD band from tuna does not have
the same sequence as does cod parvalbumin; this may explain why tuna did
not cross-react extensively with other species in this study.
Bernhisel-Broadbent et al.74 found that in 11 subjects demonstrating in
vitro allergy to fish, positive oral challenge only occurred to one species in
seven subjects, to two species in one subject, and to three in two others. This
observation has been also seen in other studies.62,75,76 However, some studies
have found that subjects clinically react to multiple fish.65,77 A recent survey
found that the rate of reactions to multiple fish among those with fish allergy
was 67%.39
15.4.3 Detection of fish allergens

Because of the complexity of fish allergy, attempts to develop detection
systems for fish residues have been lagging behind efforts for other allergens.
Several commercial monoclonal antibodies against different invertebrate
parvalbumins are available; a few methods have been described for detection
of fish parvalbumins, and most use monoclonal antibodies. In one recent
study,78 a rapid method using a surface plasmon resonance biosensor was
developed that could detect parvalbumins from carp, sardine and skipjack
tuna (katsuonut) in five minutes. This work built upon previous
accomplishments by this group in developing monoclonal antibody (against
bluefin tuna parvalbumin) that was cross-reactive to all vertebrate
parvalbumins.79 No detection limit in food for the biosensor technique was
reported, but the authors felt the method would be useful in detecting undeclared
fish residues in foods and that the biosensor chip has the advantages over
ELISA methods of regeneration and continuous use. However, it is unknown
whether this method will detect many different types of fish.
15.5 Future trends

For soy, some tests with good sensitivity have been described, and it is
expected that more sensitive ELISAs will soon be commercially available.

Together with DNA-based methods, tools are available to test for the presence
of soy residues in a wide variety of ingredients and food matrices in a
sensitive way. There is, however, a strong need for soy reference materials
in order to enable quantification and comparison of test results between
different laboratories. In addition, highly-processed soy products will still be
difficult to analyze, especially when the processing history of the ingredients
is not known.
More research in the area of detection of fish and also crustacea residues
in foods needs to be conducted to provide a variety of methods that are
sufficiently easy and relatively inexpensive for use by the food industry in
protecting seafood-allergic consumers. A main challenge for detection of
crustacea and fish is related to specificity. Are all relevant species within the
large family of crustaceans and fish recognized to the same extent? An
additional issue concerns proteins that are potent allergens in crustacea/fish
while their homologues in other species are harmless; appropriate assessment
of cross-reactivity issues will be important in this respect.
The ‘Big Eight’ allergenic foods are the most relevant for the food industry
to measure as they represent the foods that sensitive consumers are most
often allergic to. This chapter deals with three of them while the others were
discussed in earlier chapters. The list of allergenic foods in the EU was
recently extended with adoption of legislation on food allergen labelling.
The added allergens are celery, mustard and sesame. For these allergenic
foods no commercial tests are available. A polymerase chain reaction (PCR)
for celery has been described, and for mustard a polyclonal antibody-based
ELISA was recently developed (Chapter 12). Future challenges for these
foods are numerous: for some of the allergenic foods, methods need to be
prepared from scratch by raising antibodies and constructing ELISAs. Other
allergens are further along and opportunities exist for commercialization.
For all of them, even including the allergenic foods dealt with earlier, reference
materials are required, and methods that are available now need to be validated
in interlaboratory studies. All in all, for this latter group of allergenic foods,
much effort is still required in order to have the tools needed to protect
affected consumers.
15.6 References
1. Fitzpatrick, L A (2003) ‘Soy isoflavones: hope or hype?’ Maturitas, 44, S21–S29.
2. Hoogenkamp, H, Vegetable Protein, 3rd edn, St Louis, MO, Protein Technologies
International.
3. Hourihane, J O, Kilburn, S A, Nordlee, J A, Hefle, S L, Taylor, S L and Warner, J O
(1997) ‘An evaluation of the sensitivity of subjects with peanut allergy to very low
doses of peanut protein: a randomized, double-blind, placebo-controlled food challenge
study.’ J Allergy Clin Immunol, 100, 596–600.

allergic reactions in a population with peanut allergy.’ J Allergy Clin Immunol, 110,
915–920.
5. Sampson, H A (1990) ‘Peanut anaphylaxis.’ J Allergy Clin Immunol, 86, 1–3.
6. Foucard, T and Malmheden-Yman, I (2001) ‘Food-induced anaphylaxis.’ Pediatr
Allergy Immunol, 12 (Suppl 14), 97–101.
7. Foucard, T, Edberg, U and Malmheden Yman, I (1997) ‘Severe, even fatal reactions
to foodstuffs: peanut and soya underestimated allergens.’ Lakartidningen 94, 2635–
2638.
8. Kinsella, J E (1979) ‘Functional properties of soy proteins.’ J Am Oil Chem Soc, 56,
242–258.
9. Koppelman, S J, Lakemond, C M M, Vlooswijk, R and Hefle, S L. (2004) ‘Detection
of soy proteins in processed foods: literature overview and new experimental work.’
J AOAC Int. 87, 1398–1407.
10. Prudencio Ferreira, S H and Areas, J A (1993) ‘Protein–protein interactions in the
extrusion of soya at various temperatures and moisture contents.’ J Food Sci, 58,
378–384.
11. Ravenstein, P and Driedonks, R A (1986) ‘Quantitative immunoassay for soya in
raw and sterilized meat products.’ J Food Technol, 21, 19–32.
12. Molander, E (1982) ‘Determination of soya protein in meat products by standard curves
obtained from SDS-gel electrophoresis.’ Z Lebens Unters Forsch 174, 278–281.
13. Parsons, A L and Lawrie, R A (1977) ‘Quantitative identification of plant proteins in
food products. A procedure based on laser densitometry of proteins extractable in
10 M urea and resolved by thin layer polyacrylamide gel-electrophoresis.’ Annales
de la Nutrition et de l’Alimentation, 31, 201–205.
14. Marsman, G J P, Gruppen, H, de Groot, J and Voragen, A G (1998) ‘Effect of
toasting and extrusion at different shear levels on soy protein interactions.’ J Agric
Food Chem, 46, 2770–2777.
15. Hitchcock, C H S, Bailey, F J, Crimes, A A, Dean, D A G and Davis, P J (1981)
‘Determination of soya proteins in food using an enzyme-linked immunosorbent
assay [ELISA] procedure.’ J Sci Food Agric, 32, 157–165.
16. Janssen, F W, Voortman, G and de Baaij, J A (1987) ‘Detection of wheat gluten,
whey protein, casein, ovalbumin, and soy protein in heated meat products by
electrophoresis, blotting, and immunoperoxidase staining.’ J Agric Food Chem (USA)
35, 563–567.
17. Daussant, J and Massaue, J P (1977) ‘Conference Proceedings.’ Annales de la Nutrition
et de l’Alimentation 31, 139–270.
18. Taylor, S L and Nordlee, J A (1996) ‘Detection of food allergens.’ Food Technol, 50,
231–234, 238.
19. Meyer, R, Chardonnens, F, Hubner, P and Lüthy, J (1996) ‘Polymerase chain reaction
(PCR) in the quality assurance and safety of food: detection of soya in processed
meat products.’ Z Lebens Unters Forsch, 203, 339–344.
20. van Duijn, G J, van Biert, R, Bleeker-Marcelis, H, Van Boeijen, I, Adan, A J, Jhakrie,
S and Hessing, M (2002) ‘Detection of genetically modified organisms in foods by
protein- and DNA-based techniques: bridging the methods.’ J AOAC Int, 85, 787–
791.
21. Iwabuchi, S and Shibasaki, K (1981) ‘Immunochemical studies on the effect of ionic
strength on thermal denaturation of soybean 11S globulin.’ Agric Biol Chem, 45,
285–293.
22. Iwabuchi, S and Yamauchi, F (1984) ‘Effects of heat and ionic strength upon
dissociation-association of soybean protein fractions.’ J Food Sci, 49, 1289–1294.
23. Brandon, D L, Bates, A H and Friedman, M (1988) ‘Enzyme-linked immunoassay of
soybean Kunitz trypsin inhibitor using monoclonal antibodies.’ J Food Sci, 53, 102–
106.
24. Brandon, D L, Bates, A H and Friedman, M (1989) ‘Monoclonal antibody-based

enzyme immunoassay of the Bowman-Birk protease inhibitor of soybeans.’ J Agric
Food Chem, 37, 1192–1196.
25. Barkholt, V, Ebbehoj, K, Sorensen, S, Arentsen, C, Dahl, M and Frokier, H (1994)
‘Production of monoclonal antibodies for detection of antigens in both their native
and denatured states.’ Food Agric Immunol, 6,163–166.
26. Hymowitz, T (1986) ‘Genetics and breeding of soybeans lacking the Kunitz trypsin
inhibitor’. Adv Exp Med Biol, 199, 291–298.
27. Meisel, H (1993) ‘Enzyme-linked immunosorbent assay and immunoblotting using
IgY antibodies against soybean glycinin A.’ Int Dairy J, 3, 149–161.
28. Plumb, G W, Mills, E N C, Tatton, M J, D’Ursel, C C M, Lambert, N and Morgan,
M R A (1994) ‘Effect of thermal and proteolytic processing on glycinin, the 11S
globulin of soy (Glycine max): a study utilizing monoclonal and polyclonal antibodies.’
J Agric Food Chem, 42, 834–840.
29. Plumb, G W, Lambert, N, Mills, E N C, Tatton, M J, D’Ursel, C C M, Bogracheva,
T and Morgan, M R A (1995) ‘Characterisation of monoclonal antibodies against
beta-conglycinin from soya bean (Glycine max) and their use as probes for thermal
denaturation.’ J Sci Food Agric, 67, 511–520.
30. Demonte, A, Carlos, I Z, Lourenco, E J and Dutra de Oliveira, J E (1997) ‘Effect of
pH and temperature on the immunogenicity of glycinin (Glycin max L.).’ Plant
Foods Hum Nutr, 50, 63–69.
31. Huang, L, Mills, E N, Carter, J M and Morgan, M R (1998) ‘Analysis of thermal
stability of soya globulins using monoclonal antibodies.’ Biochim Biophys Acta,
1388, 215–226.
32. Hashizume, K and Watanabe, T (1979) ‘Heat denaturation of soybean proteins. II
Influence of heating temperature on conformational changes of soybean proteins.’
Agric Biol Chem, 43, 683–686.
33. Iwabuchi, S and Shibasaki, K (1981) ‘Immunochemical studies of the effect of ionic
strength on thermal denaturation of soybean 11 S globulin.’ Agric Biol Chem, 45,
285–293.
34. Janssen, F W, Voortman, G and de Baaij, J A (1986) ‘Detection of soya proteins in
heated meat products by ‘blotting’ and ‘dot blot’.’ Z Lebens Unters Forsch, 182,
479–483.
35. Griffiths, N M, Billington, M J, Crimes, A A and Hitchcock, C H S (1984) ‘An
assessment of commercially available reagents for an enzyme-linked immunosorbent
assay (ELISA) of soya protein in meat products.’ J Sci Food Agric, 35, 1255–1260.
36. Bradford, M M (1976) ‘A rapid and sensitive method for the quantitation of microgram
quantities of protein utilizing the principle of protein-dye binding.’ Anal Biochem,
72, 248–254.
37. Steensma, D P (2003) ‘The kiss of death: a severe allergic reaction to a shellfish
induced by a good-night kiss.’ Mayo Clin Proc, 78, 221–222.
38. Yunginger, J W, Sweeney, K G, Sturner, W Q, Giannandrea, L A, Teigland, J D,
Bray, M, Benson, P A, York, J A, Biedrzycki, L, Squillace, D L and Helm, R M
(1988) ‘Fatal food-induced anaphylaxis.’ JAMA, 260, 1450–1452.
39. Sicherer, S H, Munoz-Furlong, A and Sampson, H A (2004) ‘Prevalence of seafood
allergy in the United States determined by a random telephone survey.’ J Allergy
40. Besler, M, Daul, C B and Leung, P S B (2001) ‘Shrimps (Natantia).’ Internet Symposium
on Food Allergens, 3, 37–53 (www.food-allergens.de).
41. Daul, C B, Slattery, M, Reese, G and Lehrer, S B (1994) ‘Identification of the major
brown shrimp (Penaeus aztecus) allergen as the muscle protein tropomyosin.’ Int
Arch Allergy Immunol, 105, 49–55.
42. Yu, C J, Lin, Y F, Chiang, B L and Chow, L P (2003) ‘Proteomics and immunological
analyses of a novel shrimp allergen Pen m 2.’ J Immunol, 170, 445–453.
reactions to food in children and adolescents.’ N Engl J Med, 327, 380–384.
44. Cartier, A, Malo, J.-L, Ghezzo, H, McCants, M and Lehrer, S (1986) ‘IgE sensitization
in snow crab-processing workers.’ J Allergy Clin Immunol, 78, 344–348.
45. Malo, J-L and Cartier A. ‘Occupational reactions in the seafood industry.’ Clin Rev
Allergy, 11, 223–240.
46. Desjardins, A, Malo, J L, L’Archeveque, J, Cartier, A, McCants, M and Lehrer, S B
(1995) ‘Occupational IgE-mediated sensitization and asthma caused by clam and
shrimp.’ J Allergy Clin Immunol 96, 608–617.
47. Goetz, D W and Whisman, B A (2000) ‘Occupational asthma in a seafood restaurant
worker: cross-reactivity of shrimp and scallops.’ Ann Allergy Asthma Immunol, 85,
461–466.
48. Jeoung, B J, Reese, G, Hauck, P, Oliver, J B, Daul, C B and Lehrer, S B (1997)
‘Quantification of the major brown shrimp allergen Pen a 1 (tropomyosin) by a
monoclonal antibody-based sandwich ELISA.’ J Allergy Clin Immunol, 100, 229–
234.
49. Ben Rejeb, S, Davies, D, Querry, J, Cleroux, C, Langlois, D and Delahaut, P (2002)
‘Enzyme immunoassay for the detection of crustacean proteins in food.’ AOAC
Annual Meeting Abstracts, Los Angeles, CA, Abstract C-124, p. 102.
50. Lu, Y, Oshima, T, Ushio, H and Shiomi, K (2004) ‘Preparation and characterization
of monoclonal antibody against abalone allergen tropomyosin.’ Hybrid Hybridomics,
23, 357–361.
51. Verrez-Bagnis, V and Escriche-Roberto, I (1993) ‘The performance of ELISA and
dot-blot methods for the detection of crab flesh in heated and sterilized surimi-based
products.’ J Sci Food Agric, 63, 445–449.
52. Bock, S A, Muñoz-Furlong, A and Sampson, H A (2001) ‘Fatalities due to anaphylactic
reactions.’ J Allergy Clin Immunol, 107,191–93.
53. Bjornsson, E, Janson, C, Plaschke, P, Norrman, E and Sjoberg, O (1996) ‘Prevalence
of sensitization to food allergens in adult Swedes.’ Ann Allergy Asthma Immunol, 77,
327–332.
54. de Martino, M, Peruzzi, M, de Luca, M, Amato, A G, Galli, L, Lega, L, Azzari, C
and Vierucci, A (1993) ‘Fish allergy in children.’ Ann Allergy, 71, 159–165.
55. Food and Agriculture Organization of the United Nations. (1995) Food allergies,
Report of the Technical Consultation of the Food and Agriculture Organization of
the United Nations, FAO, Rome, November 13–14.
56. Beck, H I and Nissen, B K (1983) ‘Contact urticaria to commercial fish in atopic
persons.’ Acta Derm Venereol (Stockh), 63, 257–260.
57. Cronin, E (1987) ‘Dermatitis of the hands in caterers.’ Cont Derm 17, 265–269.
58. Elsayed, S and Bennich, H (1975) ‘The primary structure of allergen M from cod.’
Scand J Immunol, 4, 203–208.
59. Lehky, P, Blum, H, Stein, E and Fischer, E (1974) ‘Isolation and characterization of
parvalbumins from skeletal muscle of higher vertebrates.’ J Biol Chem, 249, 4332–
4334.
60. Aas, K and Elsayed, S (1975) ‘Physio-chemical properties and specific activity of a
purified allergen (codfish).’ Dev Biol Standard, 29, 90–98.
61. Elsayed, S and Aas, K (1971) ‘Characterization of a major allergen (cod). Observations
on effect of denaturation on the allergenic activity.’ J Allergy, 47, 283–291.
62. Hansen, T K, Bindslev-Jensen, C, Skov, P S and Poulsen, L K (1997) ‘Codfish
allergy in adults: IgE cross-reactivity among fish species.’ Ann Allergy Asthma Immunol,
78, 187–194.
63. Galland, A V, Dory, D, Pons, L, Chopin, C, Rabesona, H, Gueant, J L and Fleurence,
J (1998) ‘Purification of 41 kDa cod-allergenic protein.’ J Chromatogr B Biomed Sci
Appl, 706, 63–71.
64. Dory, D, Chopin, C, Aimone-Gastin, I, Gueant, J L, Guerin, L, Sainte-Laudy, J,
Moneret-Vautrin, D A and Fleurence, J (1998) ‘Recognition of an extensive range of

IgE-reactive proteins in cod extract.’ Allergy, 53, 42–50.
65. James, J M, Helm, R M, Burks, A W and Lehrer, S B (1997) ‘Comparison of
pediatric and adult IgE antibody binding to fish proteins.’ Ann Allergy Asthma Immunol,
79, 131–137.
66. Kelso, J M, Jones, R T and Yunginger, J W (1996) ‘Monospecific allergy to swordfish.’
Ann Allergy Asthma Immunol, 77, 227–228.
67. Lindstrom, C D, van Do, T, Hordvik, I, Endresen, C and Elsayed, S (1996) ‘Cloning
of two distinct cDNAs encoding parvalbumin, the major allergen of Atlantic salmon
(Salmo salar).’ Scand J Immunol, 44, 335–344.
68. Aas, K (1987) ‘Fish allergy and the codfish allergen model’, in Brostoff, J and
Challacombe, S J. (eds), Food Allergy and Intolerance, London, Bailliere Tindall,
356–366.
69. Van Do, T, Hordvik, I, Endresen, C and Elsayed, S (1999) ‘Expression and analysis
of recombinant salmon parvalbumin, the major allergen in Atlantic salmon (Salmo
salar).’ Scand J Immunol, 50, 619–625.
70. Shiomi, K, Hayashi, S, Ishikawa, M, Shimakura, K and Nagashima, Y (1998)
‘Identification of parvalbumin as an allergen in horse mackerel muscle.’ Fisheries
Sci, 64, 300–304.
71. Shiomi K, Hamada, Y, Sekiguchi, K, Shimakura, K and Nagashima, Y (1999) ‘Two
classes of allergens, parvalbumins and higher molecular weight substances, in Japanese
eel and bigeye tuna.’ Fisheries Sci, 65, 943–948.
72. Park, K H, Kim, H Y, Jeong, B J, Kim, K Y, Yoo, T J and Lee, K Y 1998. ‘Pediatric
IgE antibody binding to the most common seafood proteins in Korea.’ J Allergy Clin
Immunol, 101, S90.
73. Bernhisel-Broadbent, J, Straus, D and Sampson, H A (1992) ‘Fish hypersensitivity.
II. Clinical relevance of altered fish allergenicity caused by various preparation
methods.’ J Allergy Clin Immunol, 90, 622–629.
74. Bernhisel-Broadbent, J, Scanlon, S M and Sampson, H A (1992) ‘Fish hypersensitivity
I. In vitro and oral challenge results in fish-allergic patients.’ J Allergy Clin Immunol,
89, 730–737.
75. Haydel, R, El-Dahr, J, McCants, M, Helbling, A, Musmand, J and Lehrer, S (1993)
‘Food allergy: challenge studies of fish allergic subjects.’ J Allergy Clin Immunol,
91, 344.
76. Bugajska-Schretter, A, Grote, M, Vangelista, L, Valent, P, Sperr, W R, Rumpold, H,
Pastore, A, Reichelt, R, Valenta, R and Spitzauer, S (2000) ‘Purification, biochemical,
and immunological characterization of a major food allergen: different immunoglobulin
E recognition of the apo- and calcium-bound forms of carp parvalbumin.’ Gut, 46,
661–669.
77. Lin, H Y, Shyur, S D, Fu, J L, Lai, Y C and Lin, J S (1998) ‘Fish induced anaphylactic
reaction: report of one case.’ Zhonghua Min Guo Xiao Er Ke Yi Xue Hui Za Zhi, 39,
200–202.
78. Lu, Y, Ohshima, T and Ushio, H (2004) ‘Rapid detection of fish major allergen
parvalbumin by surface plasmon resonance biosensor.’ J Food Sci, 69, 652–658.
79. Kawase, S, Ushio, H, Ohshima, T, Yamanaka, H and Fukuda, H (2001) ‘Preparation
of monoclonal antibodies against tuna parvalbumin.’ Fisheries Sci, 67, 559–561.
Part IV
Issues in allergen detection methods

16
Allergen quality assurance for

hypoallergenic formula
C. Cordle, Abbott Laboratories, USA
16.1 Introduction
The development and production of hypoallergenic infant formulas (HIF)
present some unique food allergen control challenges. Almost all infants
using HIF have already developed strong allergies to cows’ milk and soy
proteins. Some of these infants will react violently to extremely small amounts
of allergen.1 Formulas made for this population represent a severe test of
allergen content control in manufactured foods.
HIF are usually made using extensively hydrolyzed proteins selected based
on their almost completely diminished immunologic reactivities.2 Current
HIF use enzymatically hydrolyzed cow milk casein or whey proteins.
Manufacturing hydrolysates for HIF is a challenging biochemical task. As
the amount of hydrolysis increases, the functional properties, organoleptic
characteristics, and biochemical stability of the hydrolysate deteriorate. During
hydrolysate development these factors must be balanced while achieving the
required reduction in immunologic reactivity. Once a satisfactory hydrolysis
procedure is set, it must be validated and controlled, and each hydrolysate
ingredient batch must be tested against strict allergen content specifications.
HIF are often produced in facilities that also manufacture other products
containing intact or partially hydrolyzed cows’ milk and/or soy protein allergens,
often using at least some shared equipment. In this environment, HIF allergen
quality assurance requires careful control of ingredients, effective and consistent
equipment cleaning procedures, processing and packaging controls, and valid
testing methods to verify product identity and quality. These methods and
procedures must be executed in an appropriately documented total quality
environment that includes allergen awareness training for plant personnel.
Procedures and training must be targeted to the needs of the specific

manufacturing process. Quality systems used to successfully manufacture
HIF may be applied, where appropriate, to aid in allergen control in more
general food production settings.
This chapter will present strategies and technology used to successfully
produce formula intended for the most severely food-allergic populations.
Section 16.2 will define key terms and discuss some important assumptions
that simplify the study of the immunology and immunochemistry of food
allergens. Relationships between clinical allergen reactivity thresholds and
analytical method’s sensitivity requirements for HIF will also be addressed.
Section 16.3 will provide a description of allergen quantitation methods used
to test HIF. Enzyme-linked immunosorbent assay (ELISA) methods are the
most valuable, and strengths and weaknesses of various ELISA procedures
will be discussed. Section 16.4 will present examples of how the analytical
tools can be applied to validate allergen-cleaning protocols and verify product
quality. Section 16.5 will summarize the current state of allergen quality
control technology, identify unresolved issues and uncertainties, and attempt
to integrate the analytical capabilities with operational procedures and plant
worker training to develop a comprehensive allergen management quality
system.
16.2 Key terms and clinical and analytical performance

targets
16.2.1 Definitions and characteristics
‘Hypoallergenic’ literally means ‘less allergenic’. This description has no
specific quantitative limits and can accurately represent comparisons between
many cooked and raw foods, processed versus unprocessed foods, or hydrolyzed
and unhydrolyzed proteins. Management of food allergy in infants requires
a more rigorous definition that relates to the clinical performance of
hypoallergenic formulas. American and European Pediatric Societies and
regulatory agencies now agree that ‘clinically hypoallergenic’ formulas must
not cause allergic symptoms when fed to 90% of infants allergic to the base
protein used to manufacture the formula. This definition also requires that
sufficient clinical research using double-blinded, placebo-controlled food
challenges (DBPCFC) must be done to have at least a 95% statistical confidence
in establishing the > 90% negative allergy reaction frequency.3,4 It is important
to note that some clinical allergic reactivity should be expected for HIF (up
to 10%) and that these products are not ‘non-allergenic’.
Protein hydrolysates used in HIF are selected for their minimal allergenic
reactivity. Successful development of these products requires an understanding
of the structure/function and dose/response relationships between food allergens
and the immune system. Research in this area is complicated by the highly
Allergen quality assurance for hypoallergenic formula 295
variable clinical manifestations of food allergy and the inability to

comprehensively identify all food allergen epitopes. This makes it difficult
to study food allergens directly. However, it is possible to predict the allergenic
potential of modified protein systems using much more standardized animal
models of immunogenicity (all allergens are immunogens) and
immunochemical assessments of antigen content (allergens and immunogens
must be antigens). Data demonstrating very low formula immunogenicity in
hyperimmunized animals and very low antigen content by sensitive
immunochemical methods are important prerequisites for clinical challenge
studies in allergic patients.2 DBPCFC in specific food-allergic patients are
still required to establish the true hypoallergenic clinical performance of
HIF.5
16.2.2 Identifying hypoallergenic ingredients

The first step in HIF development is to identify a suitable hypoallergenic
protein hydrolysate ingredient. This involves an immunochemical assessment
of the degree of reduction in protein antigenic reactivity following hydrolysis.
An important question is, ‘How much reduction in immunologic reactivity is
sufficient to achieve hypoallergenic clinical performance?’ This question is
analogous to the key allergen question in the general food industry, ‘How
much unlabeled food allergen cross-contamination is too much?’ Both questions
are addressed by exploring clinical data describing allergic reaction threshold
doses. Data from HIF clinical studies are not comprehensive but indicate that
200 μg of milk allergen / feeding might be well in the ‘hypoallergenic range’
with actual formula reactivity rates of 2–3% of milk-allergic infants.6 However,
there are isolated cases where less than 2 μg of allergen protein has created
strong reactions in the most highly allergic individuals.1
The allergen content data described here were obtained using sensitive
inhibition-format ELISA (discussed in Section 16.3 and Chapter 6). ELISA
results can readily differentiate clinically hypoallergenic products (and
hydrolysates) from more allergenically active products based on partially
hydrolyzed proteins (Table 16.1). As a general rule for formulas based on
cows’ milk protein hydrolysates, hypoallergenic clinical performance can be
achieved at residual allergen concentrations below 15 μg/mL, with the most
effective formulas containing less than 1 μg allergen/mL. (Testing is performed
at ready-to-feed concentrations on liquid formula or reconstituted powdered
formula, serving size is approximately 250 mL.) Therefore, to be useful in
quality assurance testing of hypoallergenic formula, analytical methods to
measure the casein and whey allergen components must have quantitative
sensitivity limits in the 100 ng/mL range. This sensitivity limit also seems
adequate for general food industry applications. Key complicating issues for
adapting these methods to other products are variability in serving size and
the need to test non-liquid products. The most important challenge in setting
allergen content limits is the difficulty in obtaining clinical data from blinded,
Table 16.1 Antigen concentration of hydrolysate infant formulas1
Formula Casein Whey Total Reactions2
Alimentum (H) 0.14 0.20 0.34 3

Good Start 4.40 520 524 45
Nutramigen (H) 0.19 0.43 0.62 2
Pregestimil (H) 0.19 NA NA NA
Profylac (H) 0.50 13.5 14.0 4
1
By inhibition format ELISA based on rabbit antisera to bovine casein or whey, reported as μg
immunologically active antigen/mL, at ready-to-feed formula concentration (approximately 15–18
gm protein/L)
2
Clinical reaction rate (%) in milk-allergic infants (from reference 6)
H = hypoallergenic clinical performance
controlled, progressive allergen challenges in highly allergic patients. Some

of the most sensitive patients are also at the greatest risk during challenges.
Current data also indicate that sensitivity limits are not the same for all food
allergens (Figure 16.1).7 Dose–response data have been reported for peanut,8
hazelnut,9 egg,10 milk,11 and soy.11 Interestingly, recent data from soy-allergic
infants and children suggest a substantial difference in reaction thresholds
between cows’ milk and soy allergens.11 If confirmed, this might ease the
sensitivity requirements of soy allergen quantitation methods.
When selecting method sensitivity performance targets for product allergen
control applications, the most (clinically) meaningful data are derived by
measuring the allergen content in a single serving of the food product. There
is no agreed allergen content limit for food products at this time.12 The best
performing hypoallergenic formulas contain less than 200 μg of
immunologically active antigen per 250 mL serving. This could be considered
an upper limit for acceptable serving doses of other strong food allergens.
100
Peanut
90 Hazelnut
80 Egg
70 Soy
% Responders
Milk
60
50
40
30
20
10
0
0.01 0.1 1 10 100 1000 10 000
Allergen dose (mg Protein)
Fig. 16.1 Food allergen reaction thresholds (by DBPCFC). Ingested allergen dose
(mg protein) vs % allergic responses in challenged patients. Adapted from reference 7.
Method sensitivity requirements for a specific food product can then be

estimated by dividing this amount by the product serving size. These results
should be modified based on more extensive information on the acceptable
dosage limits for specific allergens as this information becomes available.
Finally, HIF allergen doses are delivered in liquid form, while most general
food allergens are solids. Liquid allergen doses may be more immediately
active than solid allergens. If so, liquid allergen reactivity thresholds may be
lower than the same allergen ingested in solid form.
16.2.3 Hypoallergenic formula production controls

Once the hypoallergenic hydrolysate has been selected and the analytical
methods are in place, commercial production of HIF presents its own challenges.
HIF must be free of allergen contamination. HIF allergen quality assurance
programs begin with ingredient qualification and include manufacturing
equipment cleaning validation, manufacturing process control, and finished
product release testing to meet very low antigen content specifications. The
ELISA methods used during HIF product development are also useful in
quality assurance. The analytical testing goal is to assure HIF products are
free of ‘immunologically significant’ levels of contaminating antigens. How
much antigen is ‘immunologically significant’ in the context of preventing
allergic symptoms in already-sensitized patients? As discussed above, there
is no agreed standard. Manufacturing HIFs that successfully achieve
hypoallergenic clinical performance provides an interesting model for food
allergen control. Analysis of this process illustrates methods to avoid contamina-
tion, confirm production equipment cleaning, and judge product quality.
16.3 Analytical methods

16.3.1 Method applications
Methods for food allergen quantitation have three main applications in
hypoallergenic formula manufacturing. First, quantitation of residual
immunologically active allergen fragments in hydrolysate ingredients. Second,
detecting and quantitating potential protein contaminates from other non-
protein ingredients, shared processing equipment, and the manufacturing
environment. This includes validating and monitoring production equipment
and environmental cleaning procedures. Finally, allergen content of the finished
formula must be measured as a fitness for use specification. Each application
has its own requirements and technical challenges. In addition to allergen
quantitation and detection of contamination, it is also sometimes important
to have rapid and/or inexpensive methods to identify hypoallergenic ingredients
and formula with non-immunologic tests.
Measuring residual allergen content in the hydrolysate ingredient is
complicated because the analytical method must accurately measure fragmented
antigen that may contain only one or two antibody-binding epitopes. Even
these small fragments can trigger allergic reactions in some patients. This
makes assay formats based on antigen capture or ‘sandwich’ approaches
unusable. Measuring hydrolysate allergen content is made easier because
these ingredients are usually highly soluble powders that can be dissolved at
relatively high concentrations to achieve desired method sensitivity. The
assay may also be carried out in buffer systems that are optimized for method
performance. Data are best reported in units of μg immunologically active
antigen/g total protein. Hypoallergenic hydrolysates usually range from 20–
100 μg immunologically active antigen/g protein. Assuming that these
ingredients are tested at a concentration of 10 mg protein/mL, the ELISA
method should be sensitive in the range of < 200 ng/mL.
Testing for contaminants from non-protein ingredients, carry-over from
production equipment, and the manufacturing environment can be accomplished
using sandwich-type assays since the contaminants will usually be intact
proteins from other ingredients and products. These capture assays are highly
sensitive and are generally less time-consuming. Capture assays also operate
with less interference from sample matrix anomalies and can be validated for
use with a number of liquid and solid product forms. Currently, a number of
commercial ELISA kits are offered to measure allergenic food residues
(Neogen, Lansing, MI; ELISA Systems Ltd, Brisbane, Australia; R-Biopharm,
Darmstadt, Germany; Tepnel Life Sciences PLC, Deeside, UK).
Samples for these tests include product contact surfaces and environmental
samples collected by swabbing techniques. The swabbing methods are similar
to those used to collect samples for microbiological testing. A standard
surface area is swabbed, then the swab is placed in a set volume of solvent
(usually a neutral buffer with dilute detergent) and the dissolved allergen is
measured. A second sample type is rinse water following a clean-in-place
(CIP) cycle used to clean closed liquid processing systems. Care must be
exercised in collecting these samples. The samples must be a representative
rinse water sample and not a diluted cleaning fluid. Chemicals in cleaning
fluids in many cases will interfere with the assay immunochemistry. The
preferred protocol is to perform the CIP cycle, then flush the system with a
volume of clean water similar to the normal volume of liquid product that is
used to flush the system prior to initiating product collection during normal
production. Sensitivity requirements for assays used in this application are
difficult to assess since there is usually not a clear relationship between
antigen levels found on product contact surfaces or in rinse water samples
and antigen content of the initial amounts of formula product produced at
start-up. This will be discussed in more detail in Section 16.4.
Measuring residual allergen in finished hypoallergenic products has
immunochemical and sensitivity requirements similar to hydrolysate
characterization methods. Capture format assays are not useful. The tests are
also made more difficult by interference from matrix effects associated with
the emulsified formula. Similar problems in non-hypoallergenic food
applications can be largely overcome by using capture format assays. Non-

hypoallergenic foods are not routinely tested for allergen content. However,
selectively applying these tests to the first product samples made after an
equipment cleaning trial is a useful way to help interpret data from the trial
and to better understand relationships between equipment cleaning trial results
and product allergen content (see Section 16.4).
16.3.2 ELISA methods for hypoallergenic formula

Currently the most widely used analytical method for allergen quantitation is
the ELISA. Appropriate ELISAs can be used to measure fragmented antigens
(characterize hydrolysates and HIF) and intact antigens (detect HIF
contamination and monitor equipment/environmental cleaning). Successful
ELISA development requires knowledge of specific food allergen
immunochemistry. There are a number of complicating issues: food allergens
are complex mixtures of immunologically active proteins; specific allergen
reactivity will be affected differently by various food processing methods,
including hydrolysis; the allergen epitope recognition pattern will vary widely
among allergic patients and may also be different from the antigen epitope
specificity of the anti-food protein antisera used in the assay (usually animal-
derived); allergens are not measured directly, antigen content (measured
with IgG antibodies) is used as a surrogate; individual allergens are also
complex with many antibody recognition sites (epitopes) per allergen; data
are reported as ‘immunologically active (parent protein system)’ based on
standards and control samples; there are no ‘certified’ food allergen standards
so the potency, specificity, and stability of assay standards and control samples
must be individually established and be internally consistent from batch to
batch of standard.
Food antigen ELISA methods use either direct (sandwich) or inhibition
formats and are easily sensitive to 10 ng antigen / mL, with typical accuracies
of ± 5–20%. These methods are covered in detail in Chapter 6 so the discussion
here will be limited to elements of ELISA validation for testing hypoallergenic
formula. Quality elements for ELISA methods used to test hypoallergenic
formula include: potency, specificity, and stability of the anti-parent protein
antisera used as the key reagent in the test; solubility, consistency, and stability
of the selected standards and control samples; test sample integrity; and
optimization of operating conditions including validation of electronic data
collection, calculation, and reporting systems.
16.3.3 ELISA quality elements

Anti-food allergen antibody potency
ELISA quality starts with selection or generation of the anti-food protein
antisera. This antibody is the most important reagent, determining specificity
and sensitivity of the test. Polyclonal antisera generated in hyperimmunized
laboratory animals seem to be the best choice. These antisera can be obtained
in relatively large amounts using standardized food allergen samples as
immunogens. This yields reproducible potency and specificity, especially
when more than 10 animals are immunized and the processed and qualified
antisera are pooled. Animal antisera generated in this way are preferable to
food allergen-specific human IgE pools because they can be reproducibly
generated, have a broader and more consistent specificity profile, and are
cheaper and more widely available. The disadvantage of using animal antisera
is that allergen epitopes recognized by allergic humans are not necessarily
the same as immunogen epitopes recognized by hyperimmunized animals.
Food allergy in humans tends to have limited and individually variable reactivity
patterns to the complex mixture of proteins contained in a food allergen.
Establishing a pool of human IgE antisera that reacts strongly with all proteins
in a selected food allergen (and only those proteins) has not been feasible to
this point. Use of IgG antibodies from animal antisera means that the ELISA
is technically measuring food antigens, not food allergens. However, while
antigens may or may not be allergens, all allergens must be antigens. Using
animal antisera to measure food antigens may overestimate food allergen
content, but this theoretical error is on the side of product safety. Operationally,
the use of animal antisera has not been a problem in interpreting food allergen
information in the hypoallergenic formula industry for more than 15 years.
Having selected hyperimmunized animals as the source of antibody for
food antigen analysis, interest then turns to optimizing antibody potency and
specificity. Both are controlled by the hyperimmunization protocol.9 Animal
immunizations should be given at relatively high antigen doses (5 mg protein/
immunization in 4 mL of immunogen), in a potent oil-based adjuvant (Complete
Freund’s Adjuvant, CFA) processed mechanically to form a stable emulsion
(two volumes CFA: one volume antigen solution). Animals are given a booster
immunization at three weeks (5 mg protein, one volume CFA: one volume
antigen solution) with additional boosters at four to six week intervals
(substituting Incomplete Freund’s Adjuvant for CFA). Antisera collection is
started at four weeks and can continue for several months. Typically, at least
10 animals are immunized. This immunization protocol assures antisera potency
and can generate antibody titers greater than 10 million.13 Antigen-specific
antibody titers are obtained from individual animals at all time points and
evaluated for sample pooling. After the desired serum pool is made and
mixed thoroughly, the pool is aliquoted into cryo storage vials and stored at
< –70 °C. Antibody activity stability under these conditions is > 10 years.
Upon thawing, the cryo vial contents are realiquoted into smaller freezer
tubes that are used as working sources. These are stored frozen (~ –20 °C)
in a non-frost-free freezer and are stable for at least 12 months.
Anti-food allergen antibody specificity

With antisera potency established by hyperimmunization, specificity must
be determined. Hyperimmune antisera specificity will largely be controlled
by the purity of the immunogen used for hyperimmunization. A second

factor is the animal’s diet. While the hyperimmunization protocol will be
sufficiently vigorous to break antigen-specific oral tolerance, it is still advisable
to restrict the food antigens of interest from the animal’s diet, through the
previous generation if possible, or at least from birth.
Since all food allergens are complex protein mixtures, it is important to
establish that the antisera will detect (bind to) each antigen contained in the
mixture. This can be accomplished qualitatively using immunoblotting or
immunoelectrophoresis, and quantitatively using protein-specific ELISAs.
Quantitative specificity analysis requires the availability of purified individual
protein components of the food allergen. In the quantitative tests purified
individual antigens are coated onto ELISA plates and the antiserum antibody
titer (reciprocal antiserum dilution generating a specified ELISA signal) to
each allergen component is established. These titers indicate the utility of the
antisera and can be used to quantitatively compare its relative specificity
(Table 16.2, column 3: ELISA titer). The titer testing profile used to establish
specificity should also be used to show equivalence of new antibody
preparations to specificity ranges of previous antisera pools. If antisera
specificity does not meet acceptance criteria, the antibody must be regenerated,
usually using an immunogen supplemented with antigens that gave sub-
potent responses in the original immunization.
When using an inhibition format ELISA, antisera specificity is only half
of the assay specificity equation. The protein content and purity of the plate
coating antigen are other important determiners of overall assay specificity.
Note, in Table 16.2, for the bovine casein and soy ELISAs, high antibody
titers to the desired antigen and low concentrations (absence) of other proteins
in the coating antigens combine to give these assays very high specificities.
Table 16.2 Relative antisera and assay specificity (from ELISA data)
Antibody to Antigen ELISA titer Coat1 % Reactivity2
Bov. casein Bov. casein 1 930 000 99.7 100

Bov. β-lactoglobulin 9470 0.29 0
Bov. α-lactalbumin 9500 0.01 0
Bov. serum albumin 16 200 0 0
Bov. whey Bov. whey 6 670 000 70 89.0
Bov. casein 1 930 000 30 11.0
Soy protein Soluble soy protein 318 000 100 100
Bov. casein 6400 0 0
Bov. β-lactoglobulin 543 0 0
Bov. α-lactalbumin 272 0 0
Bov. serum albumin 149 0 0
1
Individual protein concentration in plate coating antigen (g/100 g)
(antigen titer) × (antigen [coat])
2 Assay specific reactivity (%) = × 100 %
Σ[(antigen titer) × (antigen [coat])]
For the whey ELISA, specificity is lower (89%). This is an unavoidable

consequence of the nature of commodity whey protein concentrates.
Approximately 70% of the protein in these ingredients is made up of the
expected whey proteins, β-lactoglobulin, α-lactalbumin, bovine serum albumin,
and immunoglobulin. However, the remaining 30% of protein is κ-casein
macropeptide and proteose peptone, both soluble casein fragments generated
during the cheese manufacturing process. Therefore, common food ingredient
whey protein concentrates will contain casein antigens and immunochemical
methods to detect ‘whey’ will also usually have some level of casein reactivity.
In some cases, selecting the optimal coating antigen can reduce the
complexity of ELISA analysis. By incorporating acceptable protein ingredients
used in formula production as sources of coating antigens and antigen standards
it is possible to achieve close alignment between the assay’s performance
with standards and unknown samples. This is appropriate since contamination,
if present, will be derived from the protein ingredients used in the production
facility. The success of this approach relies on the lot-to-lot consistency of
composition and character of the commodity ingredients. These usually vary
more than defined biochemical standards. However, careful comparative
characterization of commodity protein ingredient lots has allowed this strategy.
Another consideration in measuring complex food antigens is that single
protein components in purified form should not be used as coating antigens
or antigen standards to assess total protein antigen content or residual antigen
in protein hydrolysates. For example, the four major bovine whey proteins
have different heat and enzymatic digestion stabilities.14 Removal of one
specificity by heat treatment or enzymatic digestion does not guarantee the
same level of removal of the other whey protein antigens.
Judging ELISA method specificity requirements should be done with the
manufacturing facility’s ingredient inventory in mind. Assay specificity can
be custom selected for each location. Key questions are, ‘What allergen-
containing ingredients are present in the facility?’ and ‘Is the cleaning efficiency
of each allergen the same?’ (answer to the latter question is usually ‘no’ or
‘not sure’). Example: if a manufacturer used cows’ milk protein as an ingredient,
this could come in the form of condensed skimmed milk, non-fat dried milk,
caseinates, or whey protein concentrates. Each contains ‘milk allergens’.
However, these ingredients have different impacts on processing equipment
cleaning requirements since whey proteins are sometimes more difficult to
clean from equipment than caseinates. If the plant ingredient inventory shows
both whey protein concentrates and caseinates, then separate assays for casein
and whey may give a clearer picture of processing equipment cleaning validation
trials. Specificity will then be judged separately to establish that each antiserum
adequately recognizes the different protein components of casein or whey
antigens contained in the ingredients.
Specificity of capture format ELISAs is determined by the specificities of
the capture antibody and the detecting antibody, plus the composition and
character of the ‘sandwich’ antigen. Specificities of the two antibodies must
be determined independently if a different antibody preparation is used for

each. At least one of the antibodies should be polyclonal, but several
combinations of monoclonal and polyclonal antigen-specific antibodies can
be successfully employed. Use of appropriate protein commodity ingredients
as antigen standards in capture format ELISAs can also increase the value of
these ELISA data in assessing food allergen control strategies.
Establishing food antigen standards and control samples

Antigen standard preparations are used to establish the accuracy of ELISA
antigen quantitation methods while the precision of the tests and their stability
over time are monitored using antigen control samples. An antigen standard
is simply a stable antigen solution of known concentration. However, preparing
antigen standards for ELISA methods is more complicated than simply
dissolving food proteins in appropriate buffers and determining protein
concentration. The objective of the assay is to measure the amount of
immunologic reactivity of the sample, its antigen content. This is not always
the same as protein content, especially for commodity food proteins that
have been exposed to heat treatment and other food processing stresses.
Processing treatments tend to decrease the amount of immunologic reactivity
for a set amount of food protein. For some commodity ingredients there can
be more than an 80% lot-to-lot difference in immunologic reactivity per unit
protein. It is therefore important to select antigen standards that are as
undenatured (immunologically active) as possible. Unfortunately, at this time
there are no commercially available ‘gold standard’ food antigen preparations.
Therefore, careful selection of the best materials available and complete
characterization of each lot of these assay standards is important. Key quality
assessment targets for standards and control samples are purity, compositional
completeness of protein elements, high relative immunological reactivity,
consistency, and solubility.
Once selected and characterized, the stability of the standard must be
established. Stability of common powdered food antigens stored at –70 °C is
excellent (> 10 years), while reconstituted standards are usually stable when
stored as concentrated solutions (1–3 mg protein/mL) at –20 °C in non-frost-
free freezers for 12–24 months. Assay control samples are usually made
from different batches of the same materials used for assay standards. Control
samples are stored under the same conditions as standards and will demonstrate
similar stabilities.
Sample integrity
Sample integrity is critical in food allergen testing, especially with the high
sensitivity of the test methods. Sample contamination must be avoided. Special
care must be used in sample collection where contaminated collection
equipment and poorly cleaned equipment sampling ports can cause problems.
Another concern is that very low but significant levels of antigen can be lost
from sample solutions if the antigenic proteins adsorb to the surface of the
sample container. For rinse water samples this can be avoided by using EPA
(US Environmental Protection Agency)-approved glass vials designed for
water testing. Place 5 mL of a 10X concentrated buffer containing Tween®
20 (ICI Americas Inc, Wilmington, DE) detergent into the vial, then add 45
mL of the rinse water sample. The resulting solution is a 10% dilution of the
sample that is properly buffered while any protein is kept in solution by the
dilute detergent. Sample integrity issues should be resolved on a case-specific
basis during method validation.
Optimizing assay operating conditions

A wide variety of elements within an ELISA can (must) be optimized to
achieve the desired accuracy, precision, sensitivity range, and ruggedness.
Assay conditions can also be changed to achieve some time and cost savings
goals. There is not a single answer to the question ‘How to best perform an
ELISA?’ Each situation must be addressed based on its unique needs. Evaluation
of these needs should be a key element in prospective assay validation planning.
Unfortunately, unlike testing hypoallergenic formula, the analytical targets
for allergen testing of general foods to control allergen contamination have
not been defined by extensive clinical research, and are therefore poorly
understood. Without a clear target for ‘How much is too much?’ it is impossible
to know with certainty the meaning of food allergen test results, and also
impossible to set firm analytical performance requirements. On the other
hand, when applied in an integrated fashion, by testing environmental and
equipment cleaning samples and linking these results to antigen testing on
first-produced product, it is possible to identify environmental and equipment
locations that show much higher than normal antigen levels. These ‘hot
spots’ can then be lowered or eliminated by environmental or equipment
cleaning modifications. This is a major goal of food allergen testing of
shared processing equipment.
Data handling
The final element of analytical quality is data manipulation and reporting. As
in all cases, the automated data collection and calculation hardware and
software must be justified and validated. Standard curves will be used and
the fit parameters for standard curve acceptance must be established. Acceptable
control sample performance parameters will be established and a data quality
acceptance tree can be constructed (Figure 16.2). Consistent data reporting
in meaningful units of measurement is also important. Since agreed food
allergen standards do not exist, antigen content data will not be traceable to
officially designated standard materials. The chosen assay standards must
therefore be internally consistent to establish data quality. Linking data collected
using internal standards with the clinical performance of formula in DBPCFC
trials establishes the utility of the assay data as a product quality control tool.
Optimal reporting units for hypoallergenic formula quality assurance testing
are situational: for hydrolyzed protein ingredients and hypoallergenic formula,
1. Standard curve fit performance:
Is the correlation coefficient (R 2) of the standard curve fit > 0.98?
Yes: Go to step 2.
No: Discard the data, repeat the test.
2. Control sample performance:
Are results for the control sample(s) within acceptable limits?
Yes: Sample data are acceptable.
No: Discard the data, repeat the test.
Fig. 16.2 ELISA data quality acceptance decision tree.
μg immunologically active parent protein/g total protein (multiplied by g

protein/serving = total antigen dose/serving, the key indicator of clinical
reactivity); for rinse water samples, μg immunologically active parent protein/
mL of sample; for swab samples, μg immunologically active parent protein/
mL (assumes a standard swab solvent volume [usually 5.0 mL] which can be
converted to μg immunologically active parent protein/m2, based on the area
of the equipment surface that was swabbed).
16.3.4 Non-immunologic testing

Hypoallergenic formulas must undergo all normal infant formula nutrient
testing and microbiological/sterility testing to demonstrate quality. These
formula and their unique ingredients must also be properly identified. The
specificity that makes food antigen immunochemical assays so valuable can
also limit their utility. Biochemical methods have been developed for use in
these situations.
Biochemical characterization of hypoallergenic hydrolysates

Hypoallergenic protein hydrolysates cannot be identified and clearly
differentiated by ELISA from intact proteins from sources other than the
hydrolysate parent protein. For example, a hypoallergenic casein hydrolysate
and intact soy proteins both will give very low to negative results in a milk
protein or casein ELISA test. Low/negative ELISA data must be interpreted
as ‘no/low milk antigen’, but should not be interpreted ‘hypoallergenic
hydrolysate’ since intact soy proteins, protein-free carbohydrates, and a number
of other ingredients will also be devoid of milk allergens and yield negative
ELISA results. A separate identification method is required. As hydrolyzed
protein, the hypoallergenic hydrolysate will have a characteristically high
amino nitrogen to total nitrogen ratio (AN/TN). Hypoallergenic hydrolysates
are easily identified by a combination of high AN/TN (matching set
specifications) plus a very low (also to specification) antigen content by
ELISA. Both tests are needed to establish the hypoallergenic ingredient’s
identity.
Biochemical identification of hypoallergenic formula

Most current hypoallergenic formulas are based on hydrolyzed cows’ milk
casein or whey proteins. These formulas must meet rigorous casein and/or
whey antigen content lot release specifications. However, as above, other
formulae based on non-cows’ milk proteins will also show negative results
in casein and whey protein ELISA methods. Polyacrylamide gel electrophoresis
(PAGE), when visualized using sensitive silver staining, will show strong
banding patterns individually characteristic of non-hypoallergenic formula.
PAGE analysis of hypoallergenic formula does not yield staining.15 Most
non-hypoallergenic formulas manufactured in our facilities contain significant
amounts of intact milk proteins. The presence of these formulas as contaminants
will be easily detected using a casein ELISA. Soy-based formula is not
detected by casein ELISA, but is easily detected with PAGE. Using both
methods, from a food allergen content perspective, it can be shown that a
formula sample contains ‘very low casein antigen, according to hypoallergenic
formula specification’ and that it is ‘not intact soy-based formula’; therefore,
‘positive identification as hypoallergenic formula’.
16.4 Applications
16.4.1 Ingredient identification
Quality assurance for hypoallergenic formula begins with selection of protein
hydrolysates with the capacity to achieve hypoallergenic clinical performance.
Whether these ingredients are self-manufactured or purchased from a vendor,
a careful analysis of the hydrolysate manufacturing process is required. The
goal of this effort is to reproducibly manufacture a hydrolysate with an
extremely consistent and well-controlled degree of hydrolysis that is free of
contaminants. Hypoallergenic utility of the hydrolysate is best predicted
using animal models of immunogenicity1, 13 and established in blinded food
challenge clinical studies with food-allergic patients. Before this research is
started, it is critical that the process for manufacturing the hydrolysate be
consistent and fixed. This is not a straightforward task. Variables include the
nature and consistency of the intact protein substrate, the consistency and
composition of the enzyme and/or chemical systems used for hydrolysis, the
reproducibility and control of the hydrolysis process (time, temperature profiles,
concentrations of reactants over time, pH and ionic character control),
microbiological control, reproducibility of post-hydrolysis purification, drying
conditions, packaging, storage conditions, and transportation. In this process
the ability to measure very low levels of all allergenic protein systems present
in the hydrolysate manufacturing facility is important. The inhibition format
ELISA is clearly the best method available for this purpose. The end result
of this research is a rigorously defined and validated set of ingredients,
procedures, and specifications, along with a quality system that will ensure
their execution, which produces a consistent hydrolysate with very low residual
antigen content that is free of contamination.
The use of single protein-dedicated or ‘functionally dedicated’ facilities
for hypoallergenic hydrolysate production is most helpful and allows monitoring
of only one protein allergen source to establish adequate cleaning. If additional
proteins must be present in the production environment, additional allergen
testing will be necessary. Environmental controls deserve special attention
when configuring these facilities. Special care must be taken to control allergen-
containing dust if the protein substrate or any of the hydrolysis components
are powdered products (as is most often the case). Creative air handling and
filtration can be of great value in controlling airborne allergens. Protein dust
may be vacuumed or washed clean but should never be blown with compressed
air. The analytical goal is to perform environmental testing to identify and
eliminate allergen hot spots, then correlate the cleaned environment with
production of hydrolysate batches that are free of contamination. Controlled
repetition of the cleaning process with appropriate ELISA monitoring of
antigen content of the environment and hydrolysate produced after cleaning
will establish validity of the cleaning process for both environment and
production equipment. These data will also validate the allergen integrity of
the production processes and their controls.
16.4.2 Ingredient receipt, testing, and storage

Obviously, care must be taken to avoid ingredient contamination during
transportation and at receipt. Usually the ingredients are powdered products
in a container that may not be airtight. This allows the generation of some
dust that, if not controlled, can lead to significant levels of contamination.
Strategies to minimize dust risk include air movement and filtration
optimization, a firm policy for spill cleanup, and sequestering ingredients for
hypoallergenic products away from other protein ingredients. It is important
to note that all ingredients used in the hypoallergenic formula require this
treatment.
Ingredient acceptance testing is carried out to establish the quality of the
ingredients. Hydrolysates will be tested against biochemical, microbiological,
and antigen content specifications (by inhibition format ELISA). Non-
hydrolysate ingredients will receive normal characterization testing, but special
care must be taken to ensure that these ingredients are free of protein. Both
are then stored under conditions that will ensure stability and prevent
contamination.
16.4.3 Allergen awareness in the manufacturing environment

A number of strategies can be employed to ensure that hypoallergenic products
are free of contaminating allergens. All these start with food allergen awareness
training for production managers and operators. The goal of this training is
to raise awareness of the serious nature of food allergies in all who contribute
to formula quality. It is vital to gain commitment from operators, analysts,
and managers to proactive implementation of food allergen safety programs.
Training should foster continuous improvement in procedures and equipment
to ensure the food allergy safety of products. There is no substitute for
informed, committed, and empowered employees in food allergen control.
Generally, food allergen awareness training should include the what, why,
and how of food allergen control. This will begin with a description of food
allergies, what causes the allergic reactions, and the serious consequences
including death that can result from uncontrolled allergic reactions to food.
The regulatory environment should also be addressed, describing the increased
emphasis on food allergen regulation and FDA inspection, and the growing
food allergen product recall pattern. The potentially life-threatening
consequences for consumers should be re-emphasized and product liability
risks should be covered. Training should raise awareness in the context of
the regulations and establish personal commitment to proactively managing
food allergens in the manufacturing environment. With that accomplished,
training on allergen management can begin. This will include training on
plant-specific allergen control procedures for ingredient receipt, acceptance
testing and storage, equipment and environment cleaning procedures, allergen
testing and data interpretation, allergen control in formulation and processing,
and the importance of labeling information and packaging. Finally, the quality
systems that ensure compliance with food allergen management procedures
should be described and the trainees should be invited to suggest further
improvements as opportunities arise.
Training for operators and production supervisors is usually most meaningful
when it is custom designed for individual production environments. Training
for middle and upper managers should be more global and take advantage of
combined information from a variety of food industries. The University of
Nebraska’s Food Allergy Research and Resource Program (Lincoln, NE,
USA; www.farrp.org) provides an excellent short course on food allergen
management in industry and is recommended for this group.
16.4.4 Production environment and equipment cleaning

Production environment and equipment cleaning will involve dry cleaning,
wet washing, and clean-in-place approaches. In all cases the objective is to
reproducibly remove allergens using controlled, documented cleaning
procedures that have been demonstrated to be effective by repeated testing
(validated). Use of sensitive antigen detection methods is critical to success
here. As mentioned earlier, the capture format ELISA is probably the most
rugged and useful method in this application and a number of commercial
ELISA kits can be used effectively. An important element of using commercial
kits is to validate kit performance in detecting the specific antigens present
in the facility prior to initiating cleaning trials.
Advice on allergen cleaning procedures will be equipment-specific and is

beyond the scope of this chapter. However, there are some common principles
that deserve mention. First, contamination can be ‘systematic’ or ‘spot’.
Systematic contamination can be detected, causes identified, and procedural
corrections made to avoid the problem. Spot contamination is usually from
environmental sources and is much more difficult to address. Spot
contamination appears randomly and is especially hard to detect in powdered
or solid products where homogeneity of the formula is not assured. Second,
the relationship between equipment/environment allergen test result and level
of product allergen contamination is almost always unknown. Just as there is
no firm answer to the clinical question, ‘how much is too much?’, there is no
firm answer when the same question is asked of cleaning/manufacturing
data. This is especially challenging in interpreting environmental test results.
It is possible to produce uncontaminated products (sometimes) in the presence
of relatively high levels of antigen in the production environment. However,
inability to interpret the data in absolute terms should not be used to justify
not performing the testing.
Environmental testing should be conducted as a risk management exercise
designed to lower risks by identifying antigen hot spots and modifying cleaning
procedures to minimize environmental antigen content. While the antigen
results have limited absolute meaning, less is better, and ‘undetectable’ is an
assurance of very low risk. Finally, antigen results from swabbing food
contact surfaces and from rinse waters of production equipment will tend to
be most closely related to formula contamination potential. These tests will
provide the most valuable information in identifying hot spots and establishing
validated cleaning systems. In this effort, special care should be given to
processing equipment that heats liquid products during manufacturing since
‘burn on’ of food allergens is a common cleaning problem. ‘Dead spots’ in
liquid processing systems (valves, sampling ports, joints, etc.) where turbulent
flow is not maintained during cleaning also represent special cleaning challenges
and should be included in the sampling plan.
Antigen content data from environmental and equipment cleaning trials
should be analyzed for correlation with antigen testing of products made
following the various cleaning efforts. A future goal is to better understand
the relationship between the more abstract data from cleaning trials and
clinically-relevant data on product antigen content.
A theoretical example of trials designed to measure and optimize allergen
cleaning is shown in Table 16.3. Trial 1 represents the starting point before
efforts to optimize allergen cleaning. After identifying hot spots on the dump
room wall and air duct, environmental cleaning procedures are changed to
assure cleaning of these areas. For trial 2 the CIP cycle that cleans the surge
tank, homogenizer 4, and the heat exchange loop including the valve, is
modified by increasing circulation time. This improves surge tank and
homogenizer antigen results to undetectable levels, but there is still antigen
in the heat exchange loop. Also in this trial, vacuum cleaning the can filler
Table 16.3 Antigen cleaning optimization (example)
Sample ELISA data (ng antigen/mL or ng antigen/100 cm2
Type Trial 1 Trial 2 Trial 3
Dump room wall Swab > 3000 < 10 < 10

Blend tank 7 Rinse water < 10 < 10 < 10
Surge tank 1 Rinse water 350 < 10 < 10
Homogenizer 4 Rinse water 30 < 10 < 10
Heat exchange loop Rinse water 1250 175 50
Heat exchange valve Swab > 3000 1850 301
Air duct 8 Swab > 3000 < 10 50
Can filler 3 Swab 465 375 < 102
Can sealer 3 Swab < 10 < 10 < 10
1
After disassembly and hand cleaning
2
After warm water rinse
has reduced but not eliminated antigen. In trial 3, modifying the CIP cleaning
fluids has effectively cleaned the heat exchanger while acceptable cleaning
of the valve requires disassembly and hand cleaning. The can filler is also
effectively cleaned using a warm water rinse. This scenario, while theoretical,
is based on actual cleaning trial data and is presented to illustrate data
interpretation. Once the cleaning processes are optimized they should be
fully documented and replicated to yield an allergen cleaning validation
package and standard operating procedures (SOPs) for allergen cleaning.
16.4.5 Manufacturing controls

All of the normal quality assurance procedures and methods used for non-
hypoallergenic formula are also employed to assure the quality of
hypoallergenic formula. These are extensive but will not be discussed here.
Instead, the additional controls and testing required to manufacture
hypoallergenic formula will be detailed. Once hypoallergenic ingredient testing
methods have been developed and validated, ingredient specifications set,
allergen cleaning procedures validated and SOPs written, the next step is to
put in place procedures to control manufacturing to eliminate the risk of
releasing contaminated formula.
The best way to manufacture hypoallergenic formulas is to use facilities
that are physically separated and completely dedicated for that purpose. If
only hypoallergenic ingredients enter the facility and only hypoallergenic
formula are produced by the facility, there are no normal circumstances
whereby the formula can become contaminated. Because hypoallergenic
formulas tend to be small-volume products, use of completely separate
production facilities is not usually economically feasible. However, it is
highly advantageous to analyze production facilities to determine what
equipment can be dedicated for use only for hypoallergenic formula. Dump
stations, batching tanks, transfer pumps and lines, heat exchangers, filling
lines, sampling and cleanup tools might all be considered for dedication.
Some disadvantages of equipment dedication in a mixed product manufacturing
setting are maintaining operator training to use only the right equipment and
keeping accurate records of compliance. Clearly, dedicated equipment in a
mixed manufacturing environment does not ease the requirements to execute
validated equipment and environmental cleaning.
Assuming a mixed manufacturing or partially dedicated facility, the first
step in manufacturing control is to optimize product production scheduling.
This is a simple but powerful tool that takes advantage of a known hierarchy
in product allergenic reactivity: hypoallergenic products are least allergenically
active and are made first following the required cleaning, soy-based products
have relatively low allergenic reactivity7 and are made next, cows’ milk-
based products are most allergenic and are manufactured last in the cycle. It
is also helpful to schedule longer run times, producing a number of formula
batches and thereby minimizing the costs and down time associated with
cleaning between product change-overs.
In-process testing of the formula for antigen content is not required but
may be useful, especially when starting new production lines or incorporating
new production equipment. The advantage of in-process testing is that it
allows savings of sterilization and packaging costs of contaminated batches
that must be discarded, and it provides information useful in investigating
possible sources of contamination.
16.4.6 Finished formula release testing

In addition to the normal nutritional and microbiological release testing
performed on all formula, hypoallergenic formula must be tested to confirm
their identity and low antigen content. This is accomplished using an inhibition
format antigen ELISA as discussed in Section 16.3. Formula antigen content
specifications are tied to the antigen content of the formula batches used in
the clinical food challenge studies that established the hypoallergenic clinical
performance of the formula. If non-cows’ milk based formulas are manufactured
in the same facility, a PAGE test (described in Section 16.3.4) is used to
identify the formula by showing that it does not contain intact non-milk
proteins that are not detected by the casein ELISA.
16.5 Summary and future trends

Successful food allergen control in sophisticated food manufacturing facilities
has been practised to very exacting standards in the production of HIF for an
extended period of time. Experience gained with HIF can be used in the
general food industry. Many of the allergen management techniques directly
apply. These include: allergen control guidelines and inspections for suppliers
to ensure ingredient integrity; ingredient inventory control to provide allergen
segregation during storage; dust control and air handling optimization to

minimize contamination; production scheduling to segregate allergens; and,
most important, employee allergen awareness training to obtain broad
commitment to food allergen management. Information gained from equipment
and environmental cleaning research is also useful, and it is likely that new
research programs in this area will provide additional valuable data on food
allergen cleaning efficiencies.
The quixotic solution to food allergen contamination is to build individual
manufacturing plants for each food allergen. While this is not realistic, it is
important to remember that segregation of allergens is the best strategy for
food allergen control. It should also be noted that, although most food
manufacturing plants are not new, they are often undergoing change. Substantial
progressive improvements in food allergen management can be achieved if
allergen management becomes a design goal in ongoing plant renovations.
Operator creativity is also a powerful force in identifying allergen management
improvements. This enhances the value of allergen awareness training and
fosters ownership of product allergen quality among those closest to production.
Most of the specialized and expensive immunology tools used in
manufacturing HIF do not seem to have a place in daily mainstream food
manufacturing. It should not be necessary to establish immunochemistry
laboratories in all food manufacturing facilities. Instead, the procedures for
sampling and data analysis used to control HIF can be referenced and coupled
with commercially available food antigen ELISA kits that are used with
plant-specific food allergen ingredients to investigate food antigens in the
production environment and processing equipment. In a relatively short time
environmental and equipment antigen-cleaning procedures can be developed
and validated. If needed, some of this work can be accomplished using
consultants and contract laboratories. Once food allergen management practices
are validated and in place, maintaining control becomes an exercise in recurrent
training and taking advantage of equipment changes over time to enhance
allergen segregation.
A number of food allergen management issues are currently without solution.
The most basic unresolved question is ‘How much is too much?’ or, more
directly, ‘What is an actionable limit?’ Without a clear answer to this question
food allergen management programs are left to the uncertainties of clinical
reaction thresholds. If a product causes a clinical reaction, the allergen level
was too much. However, field reports of allergic reactions are not controlled
clinical trials and tend to represent unusual cases with an ‘n’ of one. How
will the balance between food cost contributed by extraordinary allergen
control methods and allergic reaction frequency be determined? Is this issue
scientific or political? When and by whom will it be resolved? While these
questions are being debated it is important to maintain focus on efforts to
improve allergen management and lower risks using cost-effective measures.
Another key weakness is in the analytical area. A number of useful food
allergen ELISA kits are on the market, but the list of available assays does
not include many key allergens, and there is a general lack of standardization
and validation. There are no widely accepted standards for allergens, antigens,
or antisera. Also lacking are standard assay protocols, sample collection and
preparation procedures, data calculation and reporting tools, and assay
validation data for food quality assurance applications. Substantial progress
is being made to fill these gaps and new options, better understanding, and
new partners will change the future.
A postscript regarding allergen quality assurance of ‘non-allergenic’
‘elemental’ diets: Recall that by definition, a ‘clinically hypoallergenic’ formula
can have an allergic reaction failure rate of up to 10%. Patients who react to
hypoallergenic formula must use diets that are devoid of immunological
reactivity. The protein components of these formulas are based on purified
amino acids and the products must be completely free from food protein
contamination. Amino acid-based diets were initially developed to manage
digestive difficulties. In these applications small amounts of intact protein
contamination did not have a significant negative metabolic effect on the
diet’s performance. However, when the diets are used to control extremely
severe food allergies, additional allergen quality assurance requirements must
be observed. HIF testing technology can be used with some slight modifications:
first, product contamination will only be with intact proteins from the production
environment, therefore commercially available capture-format ELISAs can
be used; second, testing will need to include all protein sources present in the
manufacturing facility; finally, while amino acid-based diets can be shown
to be clinically hypoallergenic, there are (can be) no testable clinical
performance standards for ‘nonallergenic’ formulas. Operationally, the modified
quality assurance programs have enabled us to successfully manage the
allergen quality of amino acid-based diets used in highly allergic patients for
more than six years.
16.6 References
1. Amonette, M S, Schwartz, R H, Mattson, L, Peers, L B and Eldredge, D M (1991)
‘Double-blind, placebo-controlled food challenges (DBPCFC) demonstrating acute
IgE-mediated allergic reactions to good start, ultrafiltered good start, alfare, nutramigen,
and alimentum in a seven-year-old’, Pediatr Asth Allergy Immunol, 5, 245–251.
2. Cordle, C T (1994) ‘Control of food allergies using protein hydrolysates’, Food
Techno, 48, 72–76.
3. American Academy of Pediatrics, Committee on Nutrition, ‘Hypoallergenic infant
formula’, Pediatrics, 2000, 106, 346–349.
4. Host, A, Koletzko, B, Dreborg, S, Muraro, A, Wahn, U, Aggett, P, Bresson, J L,
Hernell, O, Lafeber, H, Michaelsen, K F, Micheli, J L, Rigo, J, Weaver, L, Heymans,
H, Strobel, S and Vandenplas, Y (1999) ‘Dietary products used in infants for treatment
and prevention of food allergy. Joint Statement of the European Society for Paediatric
Allergology and Clinical Immunology (ESPACI) Committee on Hypoallergenic
Formulas and the European Society for Paediatric Gastroenterology, Hepatology
and Nutrition (ESPGHAN) Committee on Nutrition’, Arch Dis Child, 81, 80–84.
5. Bock, S A, Sampson, H A, Atkins, F M, Zeiger, R S, Lehrer, S, Sachs, M, Bush, R

K and Metcalfe, D D (1988) ‘Double-blind, placebo-controlled food challenge
(DBPCFC) as an office procedure: a manual’, J Allergy Clin Immunol, 82, 986–997.
6. Zeiger, R S (2000) ‘Dietary aspects of food allergy prevention in infants and children’,
J Pediatr Gastroenterol Nutr, 30, S77–S86.
7. Cordle, C T (2004) ‘Soy protein allergy: incidence and relative severity’, J Nutr, 134,
1213S–1219S.
F M, and Knulst, A C (2002) ‘The distribution of individual threshold doses eliciting
915–920.
9. Wensing, M, Penninks, A H, Heflé, S L, Akkerdaas, J H, van Ree, R, Koppelman, S
J, Bruijnzeel-Koomen, C A F M and Knulst, A C (2002) ‘The range of minimum
provoking doses in hazelnut-allergic patients as determined by double-blind, placebo-
controlled food challenges’, Clin Exp Allergy, 32, 1757–1762.
10. Morisset, M, Moneret-Vautrin, D A, Kanny, G, Guenard, L, Beaudouin, E, Flabbee,
J and Hatahet, R (2003) ‘Thresholds of clinical reactivity to milk, egg, peanut, and
sesame in immunoglobulin E-dependent allergies: evaluation by double-blind or
single blind placebo-controlled oral challenges’, Clin Exp Allergy, 33, 1046–1051.
11. Zeiger, R S, Sampson, H A, Bock, S A, Burks, A W, Harden, K, Noone, S, Martin,
D, Leung, S and Wilson, G (1999) ‘Soy allergy in infants and children with IgE-
associated cow’s milk allergy’, J Pediatr, 134, 614–622.
12. Taylor, S L, Hefle, S L, Bindslev-Jensen, C, Bock, S A, Burks, A W, Jr, Christie, L,
Hill, D J, Host, A, Hourihane, J O, Lack, G, Metcalfe, D D, Moneret-Vautrin, D A,
Vadas, P A, Rance, F, Skrypec, D J, Trautman, T A, Yman, I M and Zeiger, R S
How much is too much?’, J Allergy Clin Immunol, 109, 24–30.
13. Cordle, C T, Mahmoud, M I and Moore, V (1991) ‘Immunogenicity evaluation of
protein hydrolysates for hypoallergenic infant formulae’, J Pediatr Gastroenterol
Nutr, 13, 270–276.
14. Jost, R, Monti, J C and Pahud, J J (1987) ‘Whey protein allergenicity and its reduction
by technological means’, Food Tech, 1987, 41, 118–121.
15. Sampson, H A and Metcalfe, D D (1991) ‘Immediate reactions to foods’, in Metcalfe,
D, Sampson H and Simon, R, Food Allergy: Adverse Reactions to Food and Additives,
Boston, Blackwell Scientific Publications, 99–112.
17
Common issues in detecting allergenic

residues on equipment and in
processed foods
R. Crevel, Unilever Research and Development, Colworth, UK
17.1 Introduction
Food allergy has been long recognised as a clinical phenomenon, with numerous
reports in the 20th century medical literature.1,2,3 However, while it was
known that patients could suffer extremely severe and sometimes fatal reactions
following ingestion of minute amounts of the offending food, food allergy
was perceived as a problem for the individual sufferers. Since the late 1980s,
however, this perception has changed, and food allergy is now recognised as
an important public health problem. A major factor in this increased concern
is probably the rise in the prevalence of atopic disease,4 of which it can be
considered a manifestation. The prevalence and incidence of food allergy
and the number of severe reactions5 may be increasing, although the lack of
sound baseline epidemiological data precludes firm conclusions. The new
perception of food allergy has been accompanied by the recognition that the
solution to the problem lies with collaboration between all the stakeholders,
including patients and those who look after them, clinicians, public authorities
and the food industry.
The ultimate aim for all stakeholders is to avoid food allergy sufferers
reacting to the allergens to which they are sensitised. This can be achieved
in two ways. One is to ensure accurate allergen declaration through labelling,
the other is to ensure that where a specific allergen is not declared, the
product does not contain it in an amount that would pose a risk and food
allergy sufferers can assume it is safe for them. Both these requirements can
only be fulfilled by detailed knowledge of the composition of products. Food
manufacturing processes are extremely complex. This complexity derives
from several factors including material sourcing, processing, efficient use of
equipment and other resources, and product formulation. Managing allergen

risks requires an integrated approach, which takes into account all these
factors throughout the supply chain, from ingredient suppliers through to
retailers, and ultimately the consumer.
Total avoidance of cross-contact and therefore absence of specific allergens
from products where they are not part of the formulation is often not practicable.
Such circumstances require an analysis of the risk arising from residual
allergen, and subsequently a quantitative risk assessment. Although knowledge
of minimum provoking doses for many allergens is inadequate, knowing
how much allergen is present in a product is a key element in this assessment,
and the subsequent management of the allergen risk. Allergen detection
methods can play an important role at several points in the analysis of this
risk. These include the initial analysis phase, in which the current risk is
determined, and the validation of specific risk management procedures, such
as line cleaning. Upstream of the food manufacturer, analytical methods can
be used as part of the supplier audit process. Detection of allergen also plays
a role in investigations of incidents and compliance with process standards.
This chapter will discuss the role of allergen detection methods, but will not
address details of individual methods, except where these could have a bearing
on the specific use of such methods.
17.2 Food allergy and product safety

Before considering the role of detection methods in ensuring the safety of
food-allergic consumers, it is useful to take a broader view of food allergy as
a safety issue. A first consideration with food allergy is that the nature of the
hazard differs from that of other toxicants. Indeed, schemes for the classification
of adverse effects of food usually distinguish it from toxic reactions, which
potentially affect anyone who eats the food.6 Specifically, food allergy only
affects a defined section of the population, and food allergens present no risk
to non-allergic persons, irrespective of the level of intake. However, while
risk assessment in food allergy must therefore focus on the specific population
at risk, it is still valuable to analyse it through the accepted framework of
hazard identification, hazard characterisation, exposure assessment and risk
characterisation.
The first step, hazard identification, is encompassed in the definition of
the problem, namely food allergy. The hazard under consideration is any
reaction to a food mediated by the immune system, although for practical
purposes it extends only to those responses in which antibodies of the IgE
class are implicated. Intrinsic in the accepted definition of allergy is the
concept of clinical reactivity. It thus excludes situations where people are
only sensitised, as revealed by skin prick testing or measurement of specific
IgE, but do not otherwise react to contact with the food.
Issues in detecting on equipment and in processed food 317
Risk characterisation consists of establishing the relationship between the

dose of a material and the response it produces. In conventional toxicology,
this is usually achieved by experiments in suitable animal species. These
provide information about dose–response relationships and result in the
definition of a no observed adverse effect level (NOAEL) from which a safe
dose for man is calculated. Hazard characterisation in food allergy differs
from this situation, firstly because animal experimentation is irrelevant, and
secondly because it can be viewed at both an individual level and a population
one. The population dimension is probably the most relevant from the public
health point of view, and consequently for the food manufacturer, while the
individual dimension is most relevant to the clinician advising a patient on
management of the condition. For ethical reasons it is extremely difficult to
obtain information about individual responsiveness to different doses of
allergen. Full characterisation, up to the point of the most severe reaction,
can only occur inadvertently, where the dose increment used proves to be too
large. Characterisation of the population response to food allergens is, however,
feasible and ethical, in the context of helping patients manage their condition.
This is achieved through studies using double-blind placebo-controlled food
challenges (DBPCFC) with increasing doses of allergen up to the lowest
dose producing an objective response. Such experiments provide information
on the frequency of response to particular doses in the population tested.
However, such studies have a number of limitations as tools to identify a
precise NOAEL. Firstly, because of their logistics, they can only be performed
in relatively small numbers of allergic individuals, which limits their statistical
power. Typically 29 or 58 individuals would be challenged, and it can be
shown that statistically, such numbers give 95% confidence that fewer than
10 or 5% respectively of the population from which they are drawn would
react to the dose identified as the NOAEL.7 A further limitation is that
individuals who have experienced a severe reaction to the allergen of interest
are excluded from the studies by some clinicians. A modelling approach to
overcome such problems has recently been proposed.8 Although this approach
is proving promising, much validation work is still required before it can be
used in risk management.
Once the hazard has been characterised, and a NOAEL defined, exposure
assessment is required. For individuals with a food allergy, the risk arises
from acute ingestion of often quite small amounts of allergen, rather than
ingestion over a period of time. The usual measure to consider for purposes
of risk assessment is the amount of allergen that could be present in a portion
of food. An important unresolved issue exists, however, with respect to the
period over which the intake of allergen should be summed, as well as the
possible effect on the provoking dose of exposure to small amounts, unable
by themselves to stimulate an allergic reaction. Indeed some publications
have attributed increased reactivity to exposure to such doses via routes
other than the oral one.9
17.3 Management of food allergy risks

17.3.1 Aims of food allergen management
The first aim of food allergen management must clearly be to protect food-
allergic consumers, while not limiting their food choices unnecessarily. This
implies a risk assessment, as described above, as the alternative would be to
use precautionary labelling for almost all products in some instances. Indeed
the role of detection methods for allergenic residues is predicated upon a
risk-based approach to management of food allergens in food manufacturing.
If allergens were declared under all circumstances, irrespective of the risk
they pose, there would be no need for methods to detect residual allergen.
The allergic consumer would simply be informed of the presence of particular
allergens, and left to manage the risk individually. However, this approach is
viewed by allergic consumers as an abdication of responsibility, and is much
disliked.10,11 Furthermore, rather than safeguarding allergic consumers, it
can actually place them at increased risk by leading them to erroneous
conclusions about the safety of products. From the manufacturer’s point of
view, managing allergens on a risk basis means effectively taking a view as
to what proportion of the allergic population it is feasible to protect, based
on knowledge of NOAEL for individual allergens, coupled with an assessment
of achievable residual allergen content for particular products. Defining the
aims of allergen management is important not only in setting process control
objectives, but also in providing a basis for clear communication with
stakeholders, such as allergic consumers and health practitioners who they
consult for advice. For instance, decisions must be made about whether the
policy aims to avoid all reactions in allergic individuals or just severe ones.
The implications of achieving this aim must also be evaluated in a wider
socio-economic context. For instance, more thorough cleaning procedures
may result in undesirable environmental consequences, or the introduction
of water into dry systems may introduce a microbiological hazard. The range
of responsiveness of allergic patients is extremely wide, and some react to
extremely low doses. Protection of such consumers may only be achieved by
advising them not to consume manufactured foods. If so what does that
mean for assays?
17.3.2 Integrated approach

Current approaches to the management of the allergen risk in the food industry
recognise that it has to be integrated into the whole product life-cycle, from
its design right through to the point at which the consumer eats it. It is within
that integrated approach that the role of allergen detection methods must fit.
Major food manufacturers have devised specific corporate policies for the
handling of allergens, supplemented by guidelines which provide practical
advice to individual manufacturing units. These methods ensure that a high
minimum standard exists for the handling of allergens throughout the company.
For instance, Unilever has a policy for dealing with allergens which states
that it shall declare the presence in its products of any allergen which is a
common cause of allergic reactions. At a minimum, any allergen required by
local regulations will be declared. However, beyond that, the allergenic risk
from foods not commonly known to be allergenic may be assessed if clinical
or epidemiological data indicate the need. If then classed as a common cause
of allergic reactions, this food component would be declared on labels and
included in hazard analysis of critical control points (HACCP) plans. Unilever
also undertakes to inform any consumer on request about the presence of
uncommon allergens in specific products.
Allergen management guidelines need to ensure that allergens are correctly
and intelligibly declared in products, but also to make sure that allergen is
not present inadvertently at levels likely to cause adverse effects on health.
Such guidelines specifically need to address all stages in the product life-
cycle, from its design, through the sourcing of ingredients, to manufacture,
labelling and distribution. Specifically, they need to deal with:
• Innovation. Is the use of the allergenic ingredient necessary for the
functionality of the product or could an equivalent non-allergenic ingredient
serve as well?
• Supply chain. Control of allergens in the supply chain requires a close
relationship between suppliers and manufacturers so that the needs and
requirements of the latter can be met. Typically, the starting point of the
supplier assessment will be a questionnaire about allergens handled and
precautions in place to avoid cross-contact, including the existence of a
HACCP plan. This is backed up by periodic audits of the suppliers’ facilities.
Additionally, suppliers are required to seek agreement to any change in
the formulation of the ingredient they supply.
• Manufacturing protocols. Main considerations are the inclusion of common
allergens in HACCP plans, production scheduling to minimise cross-contact,
validated cleaning procedures and clear labelling and separation of
specific allergenic ingredients within the factory. Procedures need to cover
rework, where sound product is not packaged but ‘recycled’. Staff training
to understand the importance of allergen control procedures is vital and
improves support for what can be additional procedures in the production
process. Finally, the same degree of attention is needed whether the
company’s own manufacturing facility is concerned or that of co-packers.
• Packaging, promotion and advertising. Packaging carries the label and
therefore the allergen information. Care is required to ensure that information
remains with the product until it reaches the consumer. Other considerations
include warnings if the formulation has changed to include an allergenic
ingredient previously not present.
• Retailers. Generally, the manufacturer’s allergen information will be
sufficient. However, situations such as in-store promotions require care to
ensure that the consumer is fully informed. Sound product, which fails to
meet all standards for general sale, may be repackaged and sold in specialised
outlets or even in a different market. The manufacturer needs to ensure

that appropriate allergen information is retained and is available to the
ultimate consumer.
• Food professionals. Most allergic reactions to foods occur outside the
home, in conditions where the product is often not labelled and even when
asked, food professionals fail to provide correct information. Where pre-
prepared food is provided to that sector, the manufacturer has a responsibility
to ensure that accurate allergen information is provided and conveyed to
the consumer.
17.3.3 Role of allergen detection in the integrated approach

Consideration of the various phases of the product life-cycle reveals a number
of points where the detection of allergenic residues has a role. These will
include all the points at which there is uncertainty about the presence of
allergens or their levels. By definition, detection only applies to allergens
which could be present inadvertently and will therefore apply to the
manufacturing stages and any upstream of those stages. Detection of allergenic
residues will have a relatively limited role at the innovation stage, although
it may be useful in making a choice between suppliers of an ingredient. As
discussed, assessing the risk arising from inadvertent allergen presence begins
with ingredient suppliers, and clearly measurement of residual allergen levels
can provide valuable information.
As mentioned earlier, the use of detection methodology in allergen risk
management must be guided by the objectives of the policy. Total elimination
of the allergen risk, in the sense of a guarantee that no allergic individual
may be affected, irrespective of their sensitivity or the severity of reactions
they experience, is rarely, if at all, possible unless specific allergens are
excluded from manufacturing facilities. A key element in deciding what
needs to be achieved in order to afford a specified level of protection to the
allergic population is the minimum dose which provokes a reaction in such
individuals (threshold). Data on such doses are unfortunately still scarce,
and subject to much debate,9 even in the case of the most common food
allergens. They can be difficult to use confidently in management of allergen
risks, particularly since the uncertainties surrounding their derivation are
difficult to quantify. Recently attempts have been made to investigate the
distribution of such doses in the population, and to estimate by mathematical
modelling below what level in food a residual allergenic protein must be
kept in order to protect a specified proportion of the allergic population.8 As
work progresses on defining such levels, they will provide more effective
ways of monitoring the success of risk management measures. Methods for
the detection of residues will thus increase in importance in roles such as
confirming that products and ingredients meet set specifications, and the
validation of risk measures such as cleaning. They will also provide, of
course, the basis for assessment of compliance by authorities.
17.4 Role of allergen detection and other considerations

17.4.1 Why do we need detection methods?
Methods for the detection of allergenic residues can be deployed for a variety
of uses. In industry, these will include what is effectively the exposure
assessment part of risk assessment. Typical activities would be assessment of
the extent of cross-contact at different points, as part of a HACCP study, and
subsequently validation of the measures put in place to control the extent of
low level continuous cross-contact. Extending up the supply chain, such
methods could also be used to confirm suppliers’ statements about their
ingredients, as part of the audit of their processes, while downstream, product
analysis could be envisaged where incidents have occurred, or there is a
suspicion that allergenic residues may exceed specification. Similarly,
confirmation that residues are present and in what amount would be an
important starting point for investigation of incidents. However, industry is
not the only potential user of detection methods. Public authorities need to
provide evidence to support compliance activities, and demonstration of the
presence of residual allergen in products which are not supposed to contain
them can form a strong part of such evidence. Allergic consumers may also
be potential users of such methods, although none are currently suitable for
this type of application.
Different users are likely to require different methods with different
characteristics, with respect to detection limits, quantitation, robustness and
ease-of-use. Risk assessment activities imply quantitative evaluations and
require methods which measure accurately and reliably the residues of interest,
even in complex matrices. (In HACCP studies, although desirable, it is probably
not essential for the method to be easy to use.) In contrast, enforcement
authorities will only in practice be interested in quantitation if the relevant
regulations specify an action level. If no level is specified, it would presumably
be sufficient for a method to have an adequate detection limit and to be
known not to produce false positives. For potential allergic users, a key
requirement is no false negatives, as well as an adequate limit of detection.
17.4.2 What should assays for allergenic residues detect?

The allergenic activity of a food usually depends on a range of proteins, and
it has been shown many times that the pattern of response of allergic people
to the different proteins can differ considerably.12 It has also been shown
more recently both in allergic patients13 and in experimental animals14 that
the overall response to an allergenic food is a summation of the responses to
the individual proteins. The implication of those observations is that
immunoassays for food allergens should essentially be considered as means
of quantitating the relevant protein(s), rather than measuring allergenic activity
in the food, which will differ for each allergic patient. Another implication
is that quantitation of single allergenic proteins may be valuable if one is
trying to monitor the effect of processing on such proteins, but may give
highly misleading results if used in an assay intended for other purposes,
such as the estimation of the extent of cross-contact or establishing whether
a product contains more allergenic material than a set limit. The key
consideration with respect to assay development should therefore be the
purpose of the assay. The allergenic protein, although the obvious candidate,
may not always be the optimal choice. However, detecting the protein is
probably the most common approach and it is therefore appropriate to discuss
the options available.
In developing an assay based on protein detection, two main choices of
methodology exist: monoclonal and polyclonal antibody technology. Both
have advantages and drawbacks. As monoclonal antibodies recognise single
epitopes on proteins, this technology usually results in a highly specific
assay, with relatively low incidence of cross-reactivity, even with closely
related proteins, provided the antibodies have been correctly screened.
Theoretically there is also an endless supply of antibody with exactly the
same performance characteristics as the original antibody. However, the narrow
specificity of the monoclonal antibody can also be its Achilles’ heel, as
detection of the protein of interest will only take place if the antibody binding
site remains intact and accessible in the various food matrices in which the
protein may be present. Polyclonal antibody technology relies on the production
of a range of antibodies, either to a single protein of interest, or to all the
proteins within the food, depending on the preparation which is used. This
provides a detection system which is less likely to fail completely to identify
the presence of proteins of interest, although quantitation may remain
problematical if processing alters the relative proportions of the different
immunochemically active proteins in a food. However, the main drawbacks
of polyclonal technology lie in the need for extensive purification procedures
that may need to be applied to the protein(s) of interest, as well as to the
resulting antiserum to ensure specificity and absence of unwanted cross-
reactivity, together with the need to develop procedures to ensure batch to
batch reproducibility.
Once the technology itself has been selected, there remain several
possibilities in developing immunoassays based on protein. As discussed,
monoclonal technology results in a highly specific detection system, but it
can nevertheless be broadened by using a combination of detection antibodies
against different epitopes of the same protein, and/or different proteins.
However, beyond a few proteins, this becomes complex to optimise. Polyclonal
technology leaves open the choice of the material against which the antiserum
can be raised. Thus it can be as general as a protein extract of the whole food,
or as specific as a highly purified protein.
However, detection of the protein, or proteins, may not be necessary or
even be the method of choice in all instances where detection of allergenic
residues is sought. Instead, a marker molecule, for which a robust and sensitive
analytical method exists and which is always found in a known ratio to the
proteins, can be used. An example would be lactose in milk which could be

used as a tracer for estimating the amount of milk protein left on a line by
cross-contact. Similarly, a marker compound could be used in supplier audits.
However, only measurement of the protein could help an allergic individual
decide whether a particular product is safe to eat. Similarly, compliance
activities by food safety authorities are likely to be required to demonstrate
directly the presence of the offending allergen, rather than a marker material.
The reverse transcriptase polymerase chain reaction (RT-PCR) assay has
recently become the subject of considerable interest, and is described elsewhere
in this book. It takes advantage of amplification of any relevant (species)
DNA present in the food of interest to the point where it can be detected. It
can also be used in a mode which can be considered semi-quantitative.
However, in relation to allergens, it relies on an implicit assumption that the
presence of DNA in a processed food denotes the presence of (allergenic)
protein, a contention that needs experimental justification.
17.4.3 Limit of detection

The limit of detection of any assay is an important parameter, but it needs to
be considered together with the other parameters that make up the assay.
Again, the purpose of the assay should dictate this factor, as it does the
others. For most purposes, such as monitoring the effectiveness of allergen
control measures or verifying compliance with set limits, it would be reasonable
for the detection limit to be such that the assay could detect allergenic
material in an amount in a portion of food that was close to the lowest
amount shown to provoke some reaction under controlled clinical conditions.
This, of course, begs the question as to what those levels are, and to what
extent any uncertainty in determining minimum provoking doses should also
be allowed for. Minimum provoking doses vary considerably between
individuals15,16 and Bindslev-Jensen et al.8 have recently described their
cumulative frequency versus log-normal dose in the population of allergic
patients as a sigmoidal log-normal plot. The implication of this distribution
is that there is a small proportion of individuals who will respond to very
small amounts of allergen. Assays could be developed to detect such small
amounts but, even leaving aside possible technical issues of signal-to-noise
ratio, it is questionable whether such sensitivity is actually required, except
for forensic purposes. A recent suggestion, based on clinical findings17 proposed
that measures to minimise the inadvertent presence of allergen in industrial
food manufacture should aim to protect 95% of the allergic population. On
that basis, the lowest levels that residual allergen should not exceed were of
the order of 5 ppm protein, a requirement which most current commercially
available assays meet, assuming minimal losses during extraction. It should
also be borne in mind that the concept of a valid lower detection limit is only
meaningful in the context of an assay developed to detect a representative
range of proteins in a food product. Assays developed to detect a single
protein could be very difficult to interpret in these circumstances, for reasons

already discussed.
17.4.4 Characteristics of the ideal allergen detection assay

Having considered various parameters of assays for the detection of allergenic
residues, it is appropriate to examine what an ideal assay might look like.
However, simply in formulating such a question, one is inevitably drawn to
question whether there can be a single ‘ideal assay’ equally suitable for all
purposes. The ‘ideal assay’ concept will therefore be examined in some of
the different contexts in which it might be used.
Assays for monitoring effectiveness of allergen risk management measures

A typical application of such an assay would be to assess effectively different
line cleaning measures, for instance. A primary requirement would be that it
is sensitive enough to detect relevant levels of allergen and that it possesses
adequate accuracy, providing a true measure of the analyte. A high degree of
precision is not as important as accuracy, however. High specificity is also a
requirement, as food products will often contain a number of proteins from
different sources. Detection of irrelevant proteins could lead to inappropriate
decisions with regard to allergen management, e.g. implementing more stringent
measures than necessary or defensive labelling, if the analysis suggests the
problem cannot be overcome. However, in addition to the appropriate technical
parameters, the assay also needs to be designed for those who will be the
primary users. In the factory environment, these will not necessarily be
people with specialised laboratory training. The assay design must therefore
be sufficiently robust for use by non-laboratory personnel, and formats such
as calibrated test strips, for instance, are worth investigating. Similarly,
experience of the development of in-home diagnostic assays used for clinical
monitoring may hold valuable lessons.
Assays for measuring residual allergen in finished products

Although measuring residual allergen in finished products is not a common
application of assays, there are occasions when it is necessary. These may
be, for instance, where a process failure leads to the conclusion that there is
a high probability of residual allergen being present at a level likely to
endanger allergic individuals, or where there have been reports of reactions
to the product. While this type of application requires of an assay many of
the same properties as described for monitoring assays, the assay is likely to
be performed by laboratory personnel, rather than in a factory, and therefore
‘user-friendliness’, while always desirable, is of a lower priority.
Assays for investigating compliance

Where limits have been set on the residual allergen amounts in food products,
the assay must meet more stringent requirements in some respects than in the
previous instances. Sensitivity and specificity are key elements as before,

while precision is more important than for other assays, given the potential
legal implications of the test detecting allergen. In particular, the assay should
have a low incidence of false positives.
Assays for measuring single allergenic proteins

Many assays have been developed for single allergenic proteins. While these
can sometimes serve for the purposes described above, they have many
shortcomings in this regard, as discussed earlier. However, they could have
a role in evaluating the effects of processing, for instance, on the protein of
interest. Such assays will essentially be highly specific, and will probably
also be sensitive. However, they will be used almost exclusively in the
laboratory, and therefore simplicity of operation will not be a primary design
consideration.
17.4.5 Common limitations

Most of the assays discussed above have a number of limitations, some of
which are inherent in the methodology, while others result from particular
combinations of methodology and the substrate in which the analyte is sought.
The most significant limitations relate to extraction of the analyte from the
food for analysis, interference by other components of the matrix which
cannot be readily be separated and changes in the analyte itself which reduce
the ability of the method to detect it.
Variability in extracting the analyte from the food

Most common methods, including many of those for total protein analysis
(e.g. Bradford,18 BCA19) and ELISA methods, operate in an aqueous
environment and require extraction of the protein prior to analysis. The
efficiency of this extraction will depend on the solubility of the protein(s) of
interest in the aqueous buffer used for extraction. Many foods and food
products include lipids as part of their formulation, and many proteins, including
allergens, are associated in the food or product with the lipid component. In
experiments to measure the total residual protein content of edible oils, we
consistently recovered only 50% by extraction into phosphate-buffered saline,
based on a comparison with the content measured by excited nitrogen analysis,
which does not require extraction. This effect can be difficult to detect. In the
example quoted, recovery of protein spiked into oils was virtually quantitative,
presumably because of differences in their physicochemical properties,
compared with the proteins remaining in oils after refining. In a different
context, Keck-Gassenmeier et al.20 found very low recoveries (2–3%) of
peanut protein added to chocolate products, but were able to improve this to
near-quantitative recovery by the addition of fish gelatine to the extraction
buffer. These experiences indicate the need for a thorough knowledge of the
physicochemical characteristics of both the matrix and the protein(s) of interest
in order to obtain reliable results.
Matrix interference
As well as interfering with the recovery of the analyte(s) of interest, the food
matrix, or some of its components, may actually interfere with the subsequent
assay, if those components are co-extracted in sufficient amounts. For instance,
we have found that on occasions, solutions with very high sugar content
(although within the range used in several foods) reduced considerably the
recovery of β-lactoglobulin (unpublished results). Other materials commonly
used in foods, such as colours, could obviously also interfere with the
performance of assays based on colorimetric endpoints, depending on their
fate during extraction.
Changes to proteins due to processing

Food processing probably poses the greatest challenges with respect to allergen
detection, particularly for the most common type of assay, namely
immunoassays. Processing can alter either the allergenicity of a protein, or
its ability to be detected in the food matrix, either because of changes in
immunoreactivity or in the interactions between the protein and the matrix,
or indeed both. Thus, fermentation of milk with certain strains of Lactobacilli
reduces the IgE-binding capacity of the product compared to native milk,
suggesting a reduction in its ability to provoke reactions. Under these
circumstances detection of lower amounts of milk protein would be a true
reflection of a reduction in hazard to the allergic individual. A similar situation
occurs in the case of the apple allergen, which is known to be heat-labile21.
However, assays can also significantly underestimate the content of heat-
treated milk proteins, as a result of what must be assumed to be altered
recognition of the protein analyte, since a total protein assay yielded total
recovery (unpublished results). Similar findings were reported by Koch
et al.22 for roasted peanut proteins. Clinical data on reactivity to heated milk
and peanut proteins suggests that under those circumstances the apparent
reduced protein content does not reflect a reduction in hazard to the allergic
patient. These examples illustrate the need for a thorough understanding of
the pitfalls of an assay before it is used to generate data which will be used
in risk assessment.
17.5 Future trends

The need for detection of allergenic residues has now been established as the
importance of food allergy as a public health problem has become
acknowledged. Several current trends are likely to influence the development
and application of allergen detection. One is the developing legal framework,
which will ultimately lead to defined action levels. Another is the determination
of NOAELs for many of the main allergenic foods. A third may be the need
for ways of monitoring particular foods as they are modified to reduce their
allergenicity. Finally, while it not currently feasible, pressure from allergic
patients and their support groups for the means of monitoring foods for the
presence of cross-contact allergens may lead to development of some rapid
assays. The likely influence of each of these trends will be examined separately.
The legal framework with respect to food allergens is developing fast,
with Switzerland, Japan, Australia/New Zealand, the United States and the
European Union bringing in legislation specifying which allergenic foods
must be declared. The lists are usually based on the list of allergens in the
Codex General Standard on Labelling, but extend it to cover allergens of
regional importance such as celery and mustard in the European Union. This
legislation can be anticipated to drive food manufacturers to use test kits
much more extensively to demonstrate for legal purposes that their allergen
risk management procedures are effective. Although allergen testing has not
proved to be the primary mechanism of enforcement, in some legislatures
where it has a longer history such as the USA, enforcement authorities will
undoubtedly seek to use them to support other evidence. Except for Switzerland,
current legislation does not address the issue of allergen presence through
cross-contact, and action levels have not yet been set. However, as allergen
test kits become used to a greater extent, pressure is likely to grow, particularly
from manufacturers, for defined action levels, below which the presence of
the allergen would not constitute an infringement of the law. If action levels
are not set by the agencies or the legislators, they will likely be defined by
case law, which is probably not an ideal mechanism for this type of issue.
Determination of NOAELs and their use will provide manufacturers with
defined targets for their allergen management policies, in terms of what
amounts constitute a risk to what proportion of food allergy sufferers. They
will also provide manufacturers with information for improved control of
allergen hazards. Such control will, however, require that they know what
level of allergenic residues is present in their products. Measurement of
allergenic residues at appropriate points during the manufacturing process
will be one way to obtain this information and could therefore increase
considerably from its current relatively limited use.
Monitoring the allergenicity of certain foods or food products is another
area where detection of allergenic residues could play an increasing role, as
manufacturers seek to provide foods with reduced allergenicity. However,
this area is probably likely to have a lower impact than the previous two, as
looking for residual allergenicity by protein quantification is only one of
several steps in defining reduced allergenicity.
Food allergy significantly impairs quality of life for sufferers.23,24 Greater
control over their condition by food allergy sufferers would undoubtedly
help restore some of this quality. Demand for means to do so could spur an
extension of the measurement of allergenic residues to this totally new area.
This prospect is probably still quite distant, inasmuch as it requires methods
which are simple to use and robust. A critical question will be the extent of
the test manufacturer’s legal liability in the event of an allergen not being
detected and producing a reaction in a sufferer.
17.6 References
1. Prausnitz, C and Küstner, H (1921) ‘Studien über die Ueberempfindlichkeit’, Centralbl
Bakteriol Abt Orig, 86, 160–169.
2. Loveless, M H (1950) ‘Milk allergy: a survey of its incidence: experience with
masked ingestion test’, J Allergy, 21, 489.
3. Tuft, L and Blumstein, G I (1942) ‘Studies in food allergy. Sensitization to fresh
fruits: clinical and experimental observations’, J Allergy, 13, 574–581.
4. Lewis, S, Butland, B, Strachan, D, Bynner, J, Richards, D, Butler, N and Britton, J
(1996) ‘Study of the aetiology of wheezing illness at age 16 in two national British
birth cohorts’, Thorax, 51, 670–676
5. Sheikh, A and Alves, B (2000) ‘Hospital admissions for acute anaphylaxis: time
trend study’, Br Med J. 320, 1441.
6. Bruijnzeel-Coomen, C, Ortolani, C, Aas, K, Bindslev-Jensen, C, Bjorksten, B, Moneret-
Vautrin, D, Wuthrich, B (1995) ‘EAACI position paper. Adverse reactions to food’
Allergy, 50, 623–635.
7. American Academy of Pediatrics (2000) ‘Hypoallergenic infant formulas’, Pediatrics,
106, 346–349.
8. Bindslev-Jensen, C, Briggs, D and Osterballe, M (2002) ‘Can we determine a threshold
level for allergenic foods by statistical analysis of published data in the literature?’
Allergy, 57, 741–746.
9. Steensma, D P (2003) ‘The kiss of death: a severe allergic reaction to a shellfish
induced by a good-night kiss’, Mayo-Clin-Proc, 78, 221–222.
10. Said, M and Weiner, J M (2004) ‘May contain traces of…’: hidden food allergens in
Australia’, MJA, 181, 183–184.
11. Kosa, K M, Cates, S C, Post, R C and Canavan, J (2004) ‘Consumers’ attitudes
toward labelling food products with possible allergens’, Food Prot Trends, 24, 605–
611.
12. de Jong, E C, Van Zijverden, M, Spanhaak, S, Koppelman, S J, Pellegrom, H and
Penninks, A H (1998) ‘Identification and partial characterization of multiple major
allergens in peanut proteins’, Clin Exp Allergy, 28, 743–51.
13. Lewis, S A, Warner, J O and Hourihane, J (2003) ‘Promiscuity of IgE binding to
peanut allergens correlates with clinical reactivity to whole peanut during double-
blind placebo controlled challenge’, Poster 136, presented at the XXIInd Congress
of the European Academy of Allergy and Clinical Immunology, Paris, June 7–11.
14. Knippels, L M J, Penninks, A H and Bannon G A (2003) ‘The sensitising potential
of peanut proteins in four different mice strains’, Poster 681, presented at the XXIInd
Congress of the European Academy of Allergy and Clinical Immunology, Paris,
June 7–11.
15. Taylor, S L, Hefle, S L, Bindslev-Jensen, C, Bock, S A, Burks, A W, Christie, L, Hill,
D J, Host, A, Hourihane, J O, Lack, G, Metcalfe, D D, Moneret-Vautrin, D A, Vadas,
P A, Rancé, F, Skrypec, D J, Trauman, T A, Malmheden, Yman, I and Zeiger, R S
How much is too much?’ J Allergy Clin Immunol, 109, 24–30.
16. Crevel, R W R, Kerhoff, M A T and Koning, M M G (2000) ‘Allergenicity of refined
vegetable oils’; Food Chem Toxico, 38, 385–393.
17. Morisset, M, Moneret-Vautrin, D A, Kanny, G, Guénard L, Beaudouin E, Flabbée,
J and Hatahet, R (2003) ‘Thresholds of clinical reactivity to milk, egg, peanut and
sesame in Immunoglobulin E-dependent allergies: evaluation by double-blind or
single-blind placebo-controlled oral challenges’, Clin Exp Allergy, 33, 1046–1051.
18. Bradford, M M (1976) ‘Rapid and sensitive method for quantification of microgram
quantities of protein utilizing principle of protein-dye binding’, Anal Biochem, 72,
248–254.
19. Smith, P K, Krohn, R I, Hermansson, G T, Mallia A K, Gartner, F H, Provenzano,
M D, Fujimoto, E K, Goeke, N M, Olson, B J and Klenk, D C (1985) ‘Measurement

of protein using bicinchoninic acid’, Anal Biochem, 150, 76–85.
20. Keck-Gassenmeier, B, Benet, C R and Hischenhuber, C (1999) ‘Determination of
peanut traces in food by a commercially-available ELISA test’. Food Agric Immunol,
11, 243–250.
21. Vieths, S, Aulepp, H, Becker, W-M and Buschmann, L (1996) ‘Characterization of
labile and stable allergens in foods of plant origin’, in Food allergies and intolerances:
proceedings of a symposium, Bonn, May 1995, Weinheim, VCH Publishers, 130–
149.
22. Koch, P, Schappi, G, Poms, R E, Wüthrich, B, Anklam, E and Battaglia, R (2003)
‘Comparison of commercially avilable ELISA kits with human sera-based detection
23. Avery, N J, King, R M, Knight, S and Hourihane, J O (2003) ‘Assessment of quality
of life in children with peanut allergy’, Pediatr Allergy Immunol, 14, 378–382
24. Sicherer, S H, Noone, S A and Munoz-Furlong, A (2001) ‘The impact of childhood
food allergy on quality of life’, Ann Allergy Asthma Immunol, 87, 461–464.
18
Factors affecting the effectiveness of

allergen detection
U. Immer, R-Biopharm AG, Germany
18.1 Introduction
More than 40 years have passed since the first publication of the
radioimmunoassay technique by Yalow and Berson.1 Despite pessimistic
prophecies about its efficacy, it has survived, developed, and recently been
augmented by other diagnostic tests to become a powerful tool to investigate
very small amounts of substances in the nanogram to picogram range.
Developed originally to analyze hormones, the method is nowadays used for
a wide range of applications, for example, in the area of clinical diagnosis,
food and feed diagnostics and to test for environmental pollutants.
The effectiveness of any immunoassay depends directly on the quality of
the antigen used as a target and the quality of the antibody used for capture
and detection but in addition to this, the performance of the assay itself is
important. Allergens are proteins, and therefore immunoassays are appropriate
tools to measure them. Often more than one allergenic protein exists in a
complex matrix amongst many other proteins. The nature of food can vary
over an extremely wide range and sample preparations are often very difficult
to work with. Many food products are affected by heating, which can alter
the allergenic and non-allergenic proteins. So there are two components
which can influence the assay performance; one comes from the extraction
of the allergen and the other from the assays themselves.
Factors affecting the effectiveness of allergen detection 331
18.2 Factors affecting the determination of allergenic

residues
18.2.1 Factors resulting from the assay
Before the implementation of an immunoassay, whether it is an in-house
method or a commercial kit, it is necessary to become familiar with the basic
principle of the test. Therefore, the protocol, (or instructions/kit inserts if a
commercial test) have to be read thoroughly. As discussed in Chapter 6, there
are two main types of enzyme-linked immunosorbent assay (ELISA) design:
competitive and sandwich ELISAs. In the case of competitive (inhibition)
ELISAs, the more allergenic residue in the sample, the fewer antibodies are
bound and the less color is developed (indirect ratio). In the case of a sandwich-
type ELISA the amount of developed color is proportional to the amount of
allergenic residue in the sample (direct ratio). The two types of curves
represented by these two assays are illustrated in Fig. 18.1 for egg protein
and beta-lactoglobulin.
Two kinds of errors can be found in immunoassays: one is systematic
(which is manifested in the form of bias) and the second is random error
(which is reflected in poor reproducibility or ‘imprecision’). Bias and
imprecision are two statistical parameters which describe an assay. Systematic
errors leading to bias are often difficult to perceive. It implies that the ‘true’
value of a sample is known. The assessment and minimization of random
errors of measurement, which lead to imprecision and to reduced accuracy
of quantitative measurement, will now be discussed. Important parameters
are:
• performance – the precision (intra-/interassay) reproducibility/recovery,
sensitivity and reliability;
2.5 100
Sandwich ELISA Competitive ELISA
90
2 80
70
1.5 60
OD
OD
50
1 40
30
0.5 20
10
0 0
0 2 6 18 54 10 30 90 270 810
egg protein (ppm) β-Lactoglobulin (ppb)
Fig. 18.1 RIDASCREEN® Egg ELISA (sandwich) and RIDASCREEN®

ß-Lactoglobulin ELISA (competitive assay).
• convenience – the analysis time from sample preparation to final results,

the environment, the quality of equipment (pipettes and instruments); and
• skill – of technical staff.
Test performance
Because it is desirable that results should be available as fast as possible, the
tendency is to shorten the assay performance time more and more. Speed,
however, comes at a price. Antigen–antibody reactions are subject to the
laws of chemical equilibrium. If the reaction is interrupted before the steady
state is reached, the assay becomes subject to greater variability. Usually it
is required that assays should not take longer than a few minutes, at most one
to three hours. To obtain consistent results in commercial kits, the manufacturer-
recommended temperature and incubation times have to be strictly adhered
to. During each moment of incubation, the result is changing. Times and
temperatures are optimized in such a way that the recovery of the analyte
should be nearly 100%. The manufacturer has to check the robustness of the
assay concerning variation of time and temperature (in most cases a variation
of about 5% is tolerated) to guarantee the correctness of results within certain
limits.
It is necessary to understand the basic principle of the type of test being
used. In competitive assays, the reaction is initiated by adding the antibody
to the wells, which allows for slow and deliberate pipetting of samples and
standards. In the sandwich-type format, the reaction is more immediate, and
pipetting must take place without delay, since the capture antibody on the
plate binds the allergenic residues on initial contact. In the latter type of
assay, an analyst has a little bit more time if a reaction step covers about 30
minutes or more. But if the assay time takes only a few minutes (5–10) the
technician has a challenge to pipette quickly. Furthermore, for commercial
kits, it is clear that more assay wells can be processed if the assay takes a
longer time. Hence, it is important that the same velocity of pipetting is done
throughout the entire test procedure.
It is clear that the maximum number of possible samples which may be
investigated depends on the reaction time as well. An assay with a total of
five minutes incubation time will be designed with only a few standards,
with the standards and samples running as single determinations and the
number of samples not greater than 10 so that the technician is able to pipette
them within a minute. Otherwise the delay between the first and the last
pipetting is too long. If an incubation time of about 30 minutes or more is
included, it is possible to prepare half a plate or more, provided that the
pipetting is finished in less than five minutes.
For commercial kits, all of the included test components have to be handled
carefully according to the manufacturer instructions. Before using, kit
components should reach room temperature and they need to be mixed
thoroughly without creating foam within the solutions.
One of the most important points for either in-house methods or commercial
kits is the avoidance of contamination. Two types of contamination are possible:

bacterial contamination (washing buffer, sample dilution buffer as well as
standards or conjugates) and contamination due to the allergenic residue
itself. If, for example, the conjugate is delivered as a concentrate, and it is
going to be used more than once for dilution, it is essential that bacterial
contamination is avoided. On the other hand, it is also necessary to make
sure that the medium (buffer or water) which is used for diluting the concentrate
is not contaminated bacterially or by the analyte itself. This is also a
consideration for in-house methods where buffers and conjugates may be
used for several days or by several technicians for other assays also. Another
example would be that if the conjugate concentrate is diluted with the same
buffer in which the sample is to be diluted, a sufficient amount of buffer for
diluting the conjugate should be set aside separately before the buffer is used
for sample preparation. Otherwise, a contamination of sample dilution buffer
by the analyte can occur and may lead to false positive results or uninterpretable
standard curves.
Absolutely clean containers should be used for the preparation of washing
buffer, as well as sample dilution buffer. Allergenic residues sometimes adsorb
strongly onto certain surfaces, which can result in contamination of solutions.
Such contamination can also lead to uninterpretable results (high absorbances
over the whole standard curve without differentiation, see Table 18.1).
Dust containing allergenic residues present in the air or on laboratory
equipment can contaminate solutions too. In these cases, the assay should be
carried out in rooms separated from the sample preparation room. The analyst
should ensure that there is no contamination of laboratory equipment (tables,
pipettes, tubes) as well as hands (wash hands, wear gloves) before starting
the assay. The microtiterplate should be covered during each reaction step to
avoid contamination. In methods with very low detection limits, there can be
problems if dust and hand contamination are not controlled. An example of
contamination of components of a commercial kit by a customer’s laboratory
is shown in Table 18.2. The first run had results in accordance with the
standard curve which was done in the kit manufacturer’s quality assurance
laboratory. But the second run shows very high absorbance in the lower
range of the standard curve, a typical sign of contamination.
Table 18.1 RIDASCREEN ® FAST Gliadin ELISA absorbances: there was wheat flour
contamination in the cap of the sample dilution buffer bottle in which the buffer concentrate
was diluted, and afterwards this buffer was used for diluting the conjugate-concentrate
Standards Contaminated sample dilution buffer, Fresh prepared buffer,

(ng/ml) used for diluting the conjugate clean lab, clean equipment
0 3.067 0.075
10 3.067 0.594
20 3.240 1.118
40 3.064 1.931
80 2.882 2.651
Table 18.2 RIDASCREEN ® Gliadin ELISA absorbances: contamination of the second

test run in a customer’s lab
Standards Fresh opened kit Second run Curve of the quality

(ng/ml) assurance lab
0 0.040 0.685 0.056

5 0.145 0.767 0.196
10 0.303 0.790 0.440
20 0.612 0.888 0.746
40 1.123 0.938 1.235
80 1.815 1.703 1.943
For both commercial methods and in-house methods, the washing of the
ELISA plate between the reaction steps counts as an additional factor which
may influence the results. It is important to use the right washing buffer, the
right amount of buffer and the right number of washing steps. It is also
important to guarantee that no cross-contamination from well to well occurs
during washing. It is better to use more washing steps than too few. The plate
should be inverted strongly against very absorbent paper two or three times
after each washing step. It is necessary to ensure that the wells are clearly
empty between the washing steps and before each new reagent addition
begins, but the wells should not be allowed to dry. Improper washing often
leads to a higher background level, which results in higher absorbance values
in the blank control and, in both types of ELISA formats, gives rise to a
poorer differentiation between the lower standards. Often, a high coefficient
of variation (CV) indicates improper washing. Table 18.3 shows high
absorbance at the low standards in a sandwich-type commercial method. The
very high CV values highlighted are due to incomplete washing.
Technicians should make sure that pipettes are calibrated and take care to
pipet without splashing into neighbouring wells. In addition, the use of a
calibrated ELISA reader is absolutely necessary and should be maintained.
Moreover, some chromogens/substrate solutions should be kept in the dark
and for these, of course, the color reaction should be run in the dark as well.
Parallelism between standard and analyte is essential to producing an accurate
Table 18.3 RIDASCREEN® Gliadin ELISA: very high OD in the lower area of curve
connected with high CV due to incomplete washing
Standards (ng/ml) Standard Curve Coefficient of Variation (CV %)
0 0.639 13.1
5 0.988 20.1
10 1.247 1.5
20 1.832 1.0
40 2.126 6.2
80 2.682 7.1
concentration estimate of an analyte in a sample using an immunoassay.

Unfortunately, pure standards or defined standards are seldom available in
the field of allergenic residue determination. For commercial assays, each
includes its own standard. For in-house methods, investigators make their
own standards, too. The degree to which the standard is adapted to reality is
a quality parameter. Often proteins are altered during food processing. How
can the standards be adapted to all of these possibilities? This issue will be
discussed later in Chapter 19.
What does correct pipetting imply?

It is well-known that proteins often adsorb onto surfaces, so pipette tips
should be rinsed several times in standard or sample before pipetting can
begin. Standards can be pipetted from zero to the highest amount with one
tip if the tip is rinsed several times in each standard before pipetting into the
wells. If more than one dilution from a sample must be pipetted, the technician
should start with the lowest concentration (the highest dilution) and go to the
most concentrated, using the same tip rinsing technique as with the standards.
For each sample a new tip should be used. Spreading from well to well
during pipetting should be avoided since it gives rise to bubbles.
After pipetting, the plate should be covered and gently shaken (either
manually or with a microplate shaker). For commercial kits, during the
incubation steps, the plate should only be shaken if the manufacturer
recommends this step; otherwise the optical density (OD) may become too
high for photometric measurements. In connection with this, the incubation
temperature is important. Most commercial assays are developed at what is
considered room temperature in the US and middle Europe (22–24 °C).
Usually higher temperatures lead to higher OD values. Most modern ELISA
readers can measure OD ranges between 4.0 and 6.0, but sometimes old
photometers are not capable of reading more than OD 2.5. In hotter climates,
the incubation period for color development may be shortened, if this is in
agreement with the manufacturer. For in-house methods, developers/users
can adapt their conditions to their laboratory temperatures by changing
incubation times or other parameters. ELISA readers read high ODs in a less
accurate fashion than low ODs, and this may result in unreliable results. In
addition, the standard curve can often end up sigmoidal. For commercial
assays, sigmoidal curves are not ideal, but for in-house methods sigmoidal
curves can elicit good results if the developer/user uses the linear portion of
the curve. Samples which fall outside linear portions of curves should be
repeated by using a further dilution step to yield an OD which does fall on
the appropriate part of the curve. In the same way, extrapolation below the
first and above the highest standard of curve should be avoided, as these lead
to inaccurate values, which may be misleading.
Test characteristics
Whether using a commercial ELISA kit or an in-house method, the ELISA
must be robust and reliably employ a procedure that ideally has a minimum
number of manipulations.
For ELISA methods:
• For commercial kits, the standard curve should be nearly identical to the
one documented by the manufacturer. Most manufacturers deliver a quality
control data sheet for comparison.
• For in-house methods, internal quality control methods should be used to
make sure that standard curves run similarly on different days, with different
technicians, etc.
• Sandwich ELISAs should start with very low OD at the zero standard, and
maximal OD should be similar to that reported by the manufacturer; in the
case of an in-house method, again, internal quality control should be done
to ensure that the assay is running identically from day to day.
• Competitive ELISA curves should begin in a steep manner between the
first and second standard to guarantee accurate determination of samples
in the low range of the curve.
• Benchmarks of 80%, 50%, and 20% values can be used to characterize
the curve as well. These are the points of intersection of the curve at 80/
50/20% binding of the analyte and can be used to describe the reproducibility
of the standard curve. For commercial kits, most manufacturers indicate
the 50% level in the kit insert, which is that concentration reached at
50% binding of the analyte.
• It is advisable to establish quality control data, which should indicate the
80/50/20% levels (or whatever benchmark in-house methods users define)
from run to run as well as the maximum and minimum OD to evaluate the
method performance. Over time the analyst has the possibility to compare
the data from run to run and this will make it easier to judge a run that
experienced problems.
• Of additional importance is the CV of samples and standards, which is
calculated by dividing the standard deviation by the mean of replicate
results and multiplying by 100. A CV of greater than 10% within the assay
can indicates insufficient washing or pipetting. The CV calculated within
one run refers to the intra-assay reproducibility, while the CV calculated
between runs refers to the inter-assay reproducibility. Both terms help to
evaluate the assay performance.
• To evaluate the accuracy of test runs, known negative and positive controls
should be included within each run. The results can be included additionally
in the quality control data. This also helps to promptly recognize outlying data.
• To control the reproducibility of a run, the suggestion would be to run one
to three recovery samples with known contents of analyte (one negative
and two positive). Intra-assay and inter-assay reproducibilities as well as
the recoveries (measured mean value of the sample divided by the theoretical
value multiplied by 100) can then be calculated.
All of these parameters help to identify outlying data and enhance the accuracy
and reliability of the results obtained.
18.2.2 Factors resulting from the extraction method

The reliability of an ELISA method depends not only on the assay itself but
also on the development of sample extraction procedures. Extraction procedures
ideally should give a complete extraction of an analyte from all foods containing
it. Specific difficulties in determining the presence of allergenic residues in
foods can be related to the complexity of ways in which the proteins are
presented in foods. The allergenic residue should be measurable in any food
matrix, or the limits of application have to be defined. Food matrices are
very complex and range from liquids and pastes to solids and powders.
Allergenic residues can be found in raw materials or in processed food. The
components of food may have been processed by a variety of heat treatments
or extractive procedures. The food may have been stored raw, frozen or even
pasteurized or sterilized at over 100 °C. The sensitivity and selectivity of the
immunoassay must be combined with a selective extraction procedure. Much
work is necessary to ensure the robustness of extraction. For commercial
kits, the duration of the extraction method should ideally not be longer then
the incubation time of the kit, but for in-house methods the user has a lot
more flexibility in this regard. For commercial methods, the extraction should
be rapid and simple and acceptable by the user. For both commercial and in-
house methods, the extraction must be validated. Validation for allergenic
residue methods can include spiking experiments or use of ‘manufactured’
standards, such as peanut-in-chocolate (discussed in Section 18.3.3).
In commercial kits, the manufacturer describes in the kit instructions the
matrices for which the assay is validated, and also which extraction procedures
are recommended for different matrices. The buffer used as extraction buffer
is one of the most important elements of successful immunoassays for food
analysis, to provide the efficient extraction of the allergenic proteins from
the sample into a liquid phase and to minimize background effects due to
non-specific binding. Sometimes special extraction buffers or extraction
procedures for complicated matrices are used. If a matrix of interest is not
mentioned in the commercial kit instructions, the manufacturer can often
help with special supplements for special matrices or with developing a new
extraction procedure.
In most cases, samples are not homogeneous. Therefore it is recommended
to collect a solid sample from different areas of the food of interest. A greater
amount (100–200 g) of a solid sample should be collected, which then has to
be ground and homogenized to a fine powder or homogeneous mixture.
The samples are then weighed. Instruments and containers used have to be
clean to avoid traces of the analyte (especially peanut which adsorbs very
effectively or dusty materials, such as cereal flours in the case of gliadin and
milk powders), which can lead to cross-contamination. If a lot of samples must
be weighed, the spatula should be properly cleaned after each sample (especially
peanut samples or gliadin-containing samples) or disposable spatulas can be
used to avoid cross-contamination. If the concentrations of an analyte in a
series of samples are known, weighing should be performed starting from the
smallest to the largest concentration. Dust should be avoided, especially when

weighing flours (for example to determine gliadin). Sample preparation and
implementation of assay should be carried out in different rooms and the
ELISA procedure should then take place away from the weighing area.
The extraction buffer should be applied, if using a commercial kit, strictly
following the manufacturer’s instructions. Ice cream should be thawed and
homogenized before weighing. Meat, sausages and canned products have to
be finely ground or minced. It can be helpful to work with deeply frozen
samples (–20–80 °C). Often the allergen extraction needs special additives
to extract the allergen more completely. Fish gelatin, skim milk powder or
other proteins can be helpful in extraction, especially from chocolate.
Chocolate is one of the most difficult matrices. This matrix should be
heated to approximately 30–40 °C, and afterwards well homogenized and
divided into portions (for repeated analysis) by weighing. It is convenient to
heat the extraction buffer also in order to produce completely homogenized
samples. Before each repeating analysis can begin, the chocolate must be
melted and homogenized again. All of the different types of matrices have to
be mixed very well after adding the extraction medium to homogenize the
solution or suspension.
In the case of commercial ELISAs, extraction time and temperature are
optimized by the manufacturer. The former should be as short as possible.
Often an additional centrifugation step is needed to get clear supernatant
solutions. If further fine particles in the extract are still discovered, a filtration
step should be carried out afterwards. It is important to get a solution as clear
as possible to minimize background effects. Fat-containing samples, such as
meat products, ice cream or chocolate, can be centrifuged in the cold. The fat
layer is then situated on the top and can be discarded easily with a laboratory
spoon or spatula.
In addition, it can be helpful to freeze such solid samples at –20 or –80 °C
(before homogenizing) which contain a high amount of fat or cocoa (for
example chocolate bars, cream-containing bakery goods). Alternatively,
defatting with acetone can be performed before the extraction (acetone dried
powder), but this process can modify the allergenic protein itself. Food
matrices can be composed of a lot of different substances, and sometimes
one or more of these may disturb the extraction procedure. We distinguish
between two types of problem:
• substances which disturb the extraction of an analyte and
• substances which disturb the reaction between the antibody and the analyte
due to cross-reactivities or masking processes.
In some cases accompanying substances themselves disturb the determination
of the allergen. For example, chestnut can decrease the determination of
gliadin considerably. A high content of cocoa can hinder the extraction of
allergenic residues because the tannins from cocoa can mask the allergens.
Therefore additives should be used which can displace the tannins from the
allergens. Such additives, as mentioned above, should be added in relatively

high concentration (10 –> 25%).
In addition, the samples themselves are of interest. We distinguish between
raw materials (nuts, cereals, seeds, flours) and processed food products (bakery
goods, meat products, canned products). Processing of food in most cases
leads to partial denaturation of proteins (hazelnut, egg white and gliadin are
especially sensitive to heating), which can affect the ability of animal antiserum
to recognize them; however, they can retain their allergenicity after processing.
The extent to which processing can be recognized depends on the antibody;
different commercial kits use different antibodies. Nuts are used mostly
roasted, but can also be used in a dried, non-heat-treated state. If roasted,
they can be processed at different times and temperatures. An assay should
be able to detect both raw and roasted material. The correct result also
depends on what protein standards are used in the methods. Unfortunately, to
date internationaly approved standards for the different kinds of nuts do not
exist, but Chapter 19 deals with reference matertials in detail.
The detection of nut residues depends on the extent of roasting (peanut,
hazelnut, almond) and, for egg white proteins or milk proteins, depends on
the extent of heating processes (for example pasteurizing or sterilizing of
canned products). Peanut proteins are changed during the roasting process
only after prolonged times (> 30 minutes) or high temperature (> 140 °C),
methods which are not used in normal roasting of peanuts for consumption.
Hazelnut proteins are much more sensitive to roasting and can denature
much earlier. Some of the egg white allergenic proteins (e.g. ovalbumin)
denature during the process of pasteurizing or sterilizing whereas others
(ovomucoid) are stable. Prolamins from wheat, rye or barley can change
during bakery processes. A specialized extraction method is therefore
recommended to extract heat-processed prolamins.2 This procedure is also
useful if processed soy milk is evaluated for gliadin, as these types of samples
can give false positive results if done without a special extraction additive
(cocktail). All of these examples should show how important it can be to
know a great deal about the samples being tested. The more knowledge the
analysts have, the better they can judge the result.
The quality of the antibody used in ELISA methods is most important.
The main question is whether the antibody is able to recognize heat-modified
proteins from allergenic foods in the same way as the raw form. Often, in
commercial kits, the manufacturer indicates which kind of proteins the antibody
can detect. In many kits the antibody recognizes denatured proteins but not
completely, while occasionally only a qualitative result can be obtained.
In many cases, cross-reactivities may be encountered. It is important to
know how specific the antibody is. A wide range of food ingredients which
can occur in food products should be investigated (cereals, nuts, legumes,
seeds, food additives, spices, processing aids, etc.). With most commercial
kits for the detection of allergenic residues, antibodies are very specific for
the analyte of interest, but it could be the case that antibodies raised against
peanut or hazelnut could also detect walnut, or antibodies against almond

could detect hazelnut also. Very often roasted nuts show much higher cross-
reactivity than raw. When using a commercial kit, the insert should be read
thoroughly to see what types of substances cross-react in the method, or one
should call the manufacturer and inquire. Substances which have similar
species structures in relation to the analyte proteins to be determined can
also be recognized by antibody (e.g. peas in the case of peanut, apricots in
the case of almond, sunflower seeds in the case of hazelnut). Again, check
the insert of the kit or contact the manufacturer regarding this.
Special problems: prolamins

Gluten is the term for a protein mixture found in wheat cereals which contains
glutelins and prolamins (for example gliadins) in approximately equal
quantities. (Chapter 14 covers gluten detection methods in detail). One
exception is wheat starches.3 Their ratio of prolamins/gluten depends on the
washing process used (ratios exist between 1.6 and 2.6). Prolamins are ethanol-
extractable protein fractions of wheat (gliadins), rye (secalins) and barley
(hordeins) which are responsible for coeliac disease. Currently, the Codex
Alimentarius definition for ‘gluten-free’ means ‘that the total amount of the
gluten from wheat, rye, barley and oat in the products or those crossed
species in food or ingredients is not more than 200 ppm (mg/kg) on the dry
substance basis’4. But today a lower cut-off (10 ppm gliadin/20 ppm gluten)
is in discussion. Most of the manufacturers of ‘gluten-free’ food apply this
lower limit even now. Therefore, very sensitive and specific immunoassays
are necessary to find gluten contamination of these products or to find hidden,
undeclared prolamins. In the past the dominant method for gluten analysis
were Skerritt and Hill’s monoclonal antibodies.5 However, due to its higher
sensitivity and specificity the R5-mAB based ELISA6 is now recommended
and was endorsed as ‘Method One’ by Codex Alimentarius in 2005.7 On the
other hand, special extraction procedures are necessary to extract heat-treated
prolamins. A new gliadin material has been purified which aims to make
gluten analysis safer and more efficient,8 and based on it the Institute for
Reference Materials and Measurements (IRMM) has issued a gliadin standard.
Thanks to this, a new level of quality has been achieved in the field of gliadin
analysis with assays capable of detecting 1.5 ppm gliadin. The new, lower
sensitivity of the assays means that, when handling gliadin-containing samples
in the laboratory, there are some issues needing greater care than previously,
for example minimizing airborne contamination, etc. One consideration which
is not often discussed is that equipment should be tested for contamination
by swabbing from time to time.
18.3 Troubleshooting
This part of chapter is dedicated to troubleshooting in case something goes
wrong with an assay.
18.3.1 The standard curve does not match the manufacturer’s

The recommendation would be to look at the concentration of the standard
curve which represents 50% binding (50% dose). If the concentration values
differ +/– 20% from the 50% value, the assay should be repeated. The
preparation of the components as well as the incubation time and temperature
should be checked. Be sure that there is no contamination of conjugate,
sample dilution buffer or washing buffer. In addition, control of the calibration
of pipettes and of the photometer as well as the quality of substrate and
chromogen is necessary.
If the curve of a sandwich ELISA starts with an OD at the zero standard
point (0 ppm) much higher than expected from the manufacturer’s kit insert,
or if the CV of the curve and/or samples is over 10%, the washing process
should be evaluated. It is important to have a clear blank value at standard
point 1 and the second standard should be far enough from zero, otherwise
false positive results can occur.
If the analyst finds nearly the same OD for two neighbouring standards,
perhaps one of them is contaminated with the analyte, but also be sure that
a single standard is not pipetted twice. If the OD is less than the ODmax
expected, check the dilution of components and the photometer, and be sure
that the reagents do not show deterioration (cloudy or fluffy solution). The
chromogen and substrate should also be tested, and there exists a simple
method to do this. If conjugate and chromogen/substrate are pipetted together
into a test tube, the color reaction should start immediately, otherwise one of
the components has deteriorated. However, two standards with the same OD
could also mean a stability issue with the standards and, if using a commercial
kit, the manufacturer should be contacted if the problem recurs.
If the ODmax of a sandwich ELISA standard curve is too high, the incubation
time and temperature should be checked, and this should also be done if the
ODs are too low. Also, check the washing procedure, as incomplete washing
can lead to an increase in non-specific binding. It can be important, especially
for analytes which are present in a very low concentration, to minimize
issues by good performance of the assay. If the conjugate has to be diluted,
make sure the proper dilution was made.
If a standard curve shows nearly the same, very high OD over the whole
curve, it is an unmistakable sign of contamination. All components of the
assay should be prepared fresh in new, absolutely clean containers, with new
water (pure and sterilized water may be bought from a pharmacy). The
laboratory should be cleaned before starting the next assay. Sometimes (for
example, prolamin contamination) an ethanolic solution should be used for
cleaning. In some cases, water purification systems can have components
that contain milk residues, so this should be considered if running casein or
whey ELISAs. Components of a test kit from different lots or different
manufacturers should not be mixed, because the components of each lot and
kit are adjusted to each other.
18.3.2 Bad accuracy of known samples

The extraction procedure should be checked for the correct extraction buffer
and additive and the right ratio of sample to buffer. Is it necessary to use an
extraction additive? Does the analyst have to test a heat-processed sample?
Could it be that the sample is modified significantly enough through processing
such that the antibody cannot detect the protein residues appropriately?
The measured OD for the samples should be found in the middle and
steepest part of the curve. If the OD is in that part of standard curve which
is not linear, the sample determination should be repeated with a further
dilution. The analyst should not extrapolate below the first or over the last
standard of the curve. Extrapolations are not precise enough and are usually
not guaranteed by commercial test kit manufacturers. If the analyst has in
addition used a high dilution of the sample, the extrapolated mistake will be
multiplied. If the evidence shows that the extraction process was less than
perfect, check the procedure and repeat the extraction.
Check the CV for these samples and the run of the standard curve. If
reagents of a commercial kit have to be diluted, check whether the dilution
was in the right order and check the washing steps. Eventually, the measurement
of samples or even the extraction procedure of the samples should be repeated
if results continue to be poor.
18.3.3 Bad recovery of spiked samples

The analyst should be sure that the spiking of samples was done as exactly
as possible. The ELISA method can only be run with correctly spiked samples.
For spiking experiments two possibilities exist:
• weigh the required amount of a blank sample and then add the spiking
analyte solution or solid to the flask; or
• produce a higher concentration of the analyte in the desired blank matrix
and then gradually dilute it with the blank matrix until the desired
concentration is reached. This method can lower the error made by weighing.
Stir it very well over a long period of time (0.5–1 hour) at each step to be
sure that the step is homogeneous. Some matrices take more than four
hours of mixing to achieve homogeneity.
The best way to approach this depends on the matrix itself and the desired
concentration of the analyte. Spiking experiments with extracts of the analyte
provide a correct measure of the ability of the extraction buffer to do its job,
because the extraction procedure is then used twice (first for the analyte and
afterwards again for the analyte within the matrix), and gives important data
on possible matrix interferences.
If matrices are to be spiked with a known content of protein/food, one
should be sure that the matrix which will be used is free of the analyte. For
example, some types of European chocolate can contain hidden hazelnuts.
Therefore, it can be necessary to produce chocolate in-house for spiking
experiments. Peanut can cause cross-contamination quite easily due to its

tendency to be adsorbed onto vessels and spatulas, so precautions should be
taken.
Gliadin contamination can occur by airborne means, if flours or a powder
of a product are used. Solid matrices like cereals or cookies should be ground
to a fine powder before starting the spiking experiments, in a room separate
from the assay procedure laboratory in certain cases. If it is certain that the
matrix is free of the analyte, spiking can begin. Concentrations of interest
rank in the μg/g (ppm) range. The procedure cannot be carried out with
sufficient precision by weighing on a balance, because micrograms must be
weighed into a large amount of the blank matrix (for example 100 μg/10 g
sample). Therefore it is better to start with a higher concentration of analyte
(1.0–0.5% of analyte) and mix it down by repeating dilutions with the blank
matrix until the desired concentration is reached. After mixing and strong
homogenizing, the sample is further diluted step by step with the blank
matrix, for example from 1.0 to 0.1%, from 0.1% to 0.01% and then further
down to 0.001%. Appropriate mixing and homogenizing at each step is
imperative. This is the best method to get the correct concentrations in the
case of spiking solid food matrices. In the case of chocolate, the matrix
should be heated at 30–40°C while stirring during spiking.
For spiking experiments of dry solid matrices (cereals, cookies) it may be
better to use a suspension of the analyte (for example hazelnut, peanut or
almond) for spiking, which allows weighing of a higher amount of the analyte
(depending on the amount in the suspension) in a homogeneous solution.
Otherwise pockets or clumps of analyte may result.
18.3.4 Poor precision (reproducibility and variation (CV))

Intra-assay problems
In a test, the OD has a very high variance between duplicates. The analyst
should test the components for bacterial contamination (cloudy and flaky
solutions). Inspection and evaluation of the pipetting procedure should be
done. Did the pipette tips get rinsed several times before pipetting? The
washing process should be checked again. Incomplete washing and inadequate
evacuation of wells after washing can influence the zero standard and the
CV, and therefore the detection limit of the assay as well as the precision of
the sample determination. Also, incomplete mixing of reagents in the microwells
can cause this problem. Ensure that the components are mixed by gently
rocking the plate after adding the reagents or, in some cases, a microplate
shaker can be employed. Check background effects with an extracted blank
sample or with wells into which only conjugate and the reagents for color
reaction are pipetted.
Inter-assay problems
In a test, the ODs between runs of the same lot or different lots show a large
range of variance. Pipetting errors or reagent deterioration could be the
reason. Further, the stability of sample extracts should be tested. Also, the
reproducibility of the extraction itself should be inspected by a reference
sample (known concentration) by repeating the extraction. The homogeneity
of the sample should be examined. The analyst should increase the sample
size and should repeat the extraction again.
18.3.5 Inadequate color development

If all of the wells show the same OD, one of the reagents could have deteriorated,
the washing buffer could be contaminated, the reagents could be mixed
incompletely or the conjugate or standards could be contaminated. If there is
no color development in the wells, possibly the conjugate or color reagent
are missing. The analyst should check the color development by mixing
color reagents and conjugate. Color development should start immediately,
otherwise it is likely that one of the components has deteriorated.
18.4 Future trends

18.4.1 Fast assays
The goal in the future will be to achieve shorter immunoassays which are
nevertheless able to guarantee the detection limit that is demanded for allergenic
residue detection. Concentration limits of allergens in food should be defined
according to appearance of clinical reaction in food-allergic individuals (see
Chapter 2 for a discussion of thresholds).
In the commercial kit market, more rapid assays, which possess, in general,
incubation times of 3 × 10 minutes, are already available for some of the
most important allergenic residues. Incubation times will be further decreased
in relation to a low detection limit, which can be reached, assuming that the
quality of antibodies is high enough. Also, extraction procedures will become
faster, easier, and more effective at getting complete recovery. One day
internationally-approved reference materials for all kinds of food allergens
will exist.
18.4.2 Dipsticks
Dipstick formats have been developed for allergenic residues – the antibodies
or antigens are immobilized to a microporous surface on a strip. The assay
can be easily performed by transferring the stick into a sample after extraction
from food. For a complete discussion of dipsticks/lateral flow devices, see
Chapter 10 of this book.
There are a few commercially available dipsticks which allow determination

of allergenic residues in the range of a reasonable cut-off point to get a yes/
no answer. A decision within 5–10 minutes is possible. Two types of design
are available:
• determination of a single allergen; or
• determination of a group of allergens (e.g. tree nuts) on one dipstick.
For example, a very sensitive dipstick is now available for detection of
gliadin in food (R-Biopharm AG, Darmstadt, Germany), based on the
monoclonal antibody R56 which is able to recognize prolamins from wheat,
rye and barley at 100% and fails to recognize prolamins from maize and rice.
The antibody recognizes a toxic peptide, which is repeatedly found in gliadins
and glutenins, and is also able to recognize heat-denaturated gluten provided
a special extraction procedure is conducted.2 In a basic design common to all
lateral flow devices, the test principle is based on an immunological capture
of colored microparticles (which are bound to the monoclonal antibodies
against gliadin) that migrate through a membrane, in which also the monoclonal
antibody against gliadin has been immobilized. The sample flows through
the porous membrane. If there are gliadins in the sample, the colored
microparticles are bound and the complex moves across to the zone of
immobilized gliadin capture antibodies, where it is immobilized. A red colored
line will appear if the sample contains gliadin.
This assay works quickly (within five minutes a visible result is obtained)
and is easy to perform. The use of the stick is very safe because, additionally,
a blue safety control band appears if the stick works correctly. This gliadin
dipstick is able to detect less than 10 ppm gliadin in a sample. Further
dipsticks are under development by various companies for various allergenic
foods. Two manufacturers have had dipsticks for peanut on the market for
the past year. It is expected that commercial dipsticks for milk, egg, almond
and hazelnut will be developed in the next few years.
Dipsticks are helpful in evaluating incoming raw material quickly, to
check finished goods and to inspect the production process and equipment
for allergenic residues. Often the production of allergen-containing and non-
allergen-containing food occurs on the same line or on parallel lines. It is
very important to assess the production process to guarantee that cross-
contamination of allergenic residue is avoided or to indicate if the processing
line is free of residues. The dipstick can be used directly by swabbing
processing/packaging equipment. Usually, an area of 10 × 10 cm2 is swabbed.
Then the stick is dipped into a buffer solution and the migration process
begins. In the case of the R-Biopharm AG gliadin test, as little as 1.5 μg/100
cm2 can be detected. Dipsticks are also very useful for analyzing rinse waters
from clean-in-place systems.
18.5 Summary
The success of any immunoassay depends directly on the main components
(antigens, antibodies), but the performance of the assay is also important in
obtaining a reliable result.
Two kinds of errors can be found – a systematic (in the form of bias) and
a random error (reflected in poor reproducibility or ‘imprecision’). Bias and
imprecision are two statistical parameters which describe each assay. Some
suggestions to minimize random errors of an assay are made in this chapter.
The performance of a test influences the precision (intra-/interassay),
reproducibility, recovery, sensitivity and reliability of the system. It is neccessary
to understand the basic principle of the assay type one is using. The assay
conditions and the handling on the one hand and the analyte extraction
procedure on the other are important parameters influencing the result. Before
starting, the user should be clear about the assay format, the time requirements,
the number of possible samples with one run, and the standardizing/calibration
of the assay. Further, the careful handling of test components and mistakes
that can occur (adsorption, contamination, degradation) and the factors
influencing the procedure (time, temperature, washing steps, starting point
of the reaction) are covered. Today’s immunoassays are very sensitive. It is
clear that care must be taken to avoid cross-contamination in the laboratory
and some examples are given concerning possible origins of contamination.
The evaluation of results should come from a standard curve which fits the
measured values in the best way; often a cubic spline method is needed.
Extrapolations above and below the standard range of the curve should not
be made. Hints are given for an internal quality control regime, which should
be introduced into each laboratory that deals with immunoassays, to secure
the results.
The sensitivity and selectivity of the immunoassay has to be combined
with a selective and effective extraction procedure of the analyte. Factors of
interest are the incubation time and temperature during the extraction, but
most important is the suitability of the extraction buffer. Optimization is
absolutely neccessary. Tips are given on how to spike different kinds of
matrices effectively to control the assay and the extraction procedure. Finally,
recommendations are made concerning troubleshooting in the case of problems
in obtaining good analytical results.
18.6 References
1. Yalow, R S and Berson, S A (1960) ‘Immunoassay of endogenous plasma insulin in
man’, J Clin Invest, 39, 1157–1161.
2. Lopez, E, Reyes, E, Hernandez, M, Llorente, M, Mendez, E and Wieser, H (2000) ‘A
quantitative gluten “cocktail extraction” procedure for heat-processed foods’,
Proceedings of the 15th meeting of the Working Group on Prolamin Analysis and
Toxicity, Nov. Meran, 129–138.
3. Wieser, H and Antes, S (2001) ‘Development of a non-immunological method for the

quantitative determination of gluten in wheat starch’, Proceedings of the 16th meeting
of the Working Group on Prolamin Analysis and Toxicity, Nov, Sitges, Verlag
Wissenschaftliche Scripten, Zwickau, 19–23.
4. Codex Alimentarius Commission, Standards for gluten-free foods, Rome, 1981 FAO/
WHO, 181.
5. Skerritt, J H and Hill, A S (1990) ‘Monoclonal antibody sandwich enzyme immunoassays
for determination of gluten in food’, J Agric Food Chem, 38, 1771–1778.
6. Lopez, E, Reyes, E, Valdes, I, Llorente, M, Iglesias, J and Mendez, E (1999) ‘Massive
analysis of food samples by a sandwich ELISA based on a unique monoclonal antibody’,
Toxicity, Nov, Neuhausen, 53–61.
7. Codex Alimentarius Commission (2005) 26th Session of the Codex Committee on
Methods of Analysis and Sampling, Budapest, 4–8 April, in press.
8. van Eckert, R (2001) ‘The PWG gliadin, a new reference material’, Proceedings of
the 16th Meeting of the Working Group on Prolamin Analysis and Toxicity, Nov,
Sitges, Verlag Wissenschaftliche Scripten, Zwickau, 25–27.
19
Reference materials and method

validation in allergen detection
R. Poms, International Association for Cereal Science and
Technology, Austria, H. Emons and E. Anklam, Institute for
Reference Materials and Measurements, Belgium
19.1 Introduction
Reliable detection and quantification methods for food allergens are necessary
in order to improve consumer protection and to ensure compliance with food
labelling regulations. However, the analysis of allergenic components in
food products can be very difficult, because their chemical identity is often
not sufficiently known and it is suggested that they are frequently only
present in trace amounts. Another question yet to be answered is the required
detection power and selectivity of corresponding analytical methods, as
scientifically sound threshold levels have not yet been established, e.g. by
human oral challenge studies. Threshold levels for specific allergic reactions
determined until now by double-blind placebo-controlled food challenges
(DBPCFC) range between less than 1 mg and more than 1 g of allergenic
protein depending on the food concerned and the selected individuals (Taylor
et al., 2002). There seems to be a general assumption that the detection limits
for different food products need to be around 10 ppm [mg allergen (protein)/
kg food] or lower, depending on the particular food (Koppelman et al., 1996,
Poms and Anklam, 2004). But one should keep in mind that mass is usually
not an appropriate quantification and measurement parameter for analytes
with varying chemical/biological activity such as proteins.
Currently, there are several analytical approaches for the detection of
potential allergens in food products (reviewed in Poms et al., 2004). The
methods employed are targeting either the allergen (protein) itself or a marker
that indicates the presence of the offending food. At present, detecting the
allergen per se is not always feasible, as the crucial chemical properties may
not be well characterized and understood or the detection limits of the methods
Reference materials and method validation in allergen detection 349
used are insufficient. Typically, specific proteins or DNA fragments are targeted
as markers for the presence of potentially allergenic food products or
ingredients. Protein-based methods usually involve immunochemical detection
protocols such as the radio-allergosorbent test (RAST), enzyme-allergosorbent
test (EAST), enzyme-linked immunosorbent assay (ELISA) and
immunoblotting (discussed in other chapters of this book), whereas only the
ELISA technique is presently used in routine food analysis. Methods operating
on the DNA level are usually based on an amplification of a specific DNA
fragment by the polymerase chain reaction (PCR) (see Chapter 7). Currently
the employment of DNA-based methods for the detection of allergens in
food products is a matter of controversy, since proteins are the causative
agents in allergy and PCR results cannot be linked to any allergen/protein
content. However, the choice of method is still mainly dependent on the food
concerned (availability of specific antibodies/DNA-primers and the achievable
detection limit) and on the effects of food processing during production.
Therefore, protein-based and DNA-based methods each have their own
characteristic merits and drawbacks concerning their applicability in the
detection and quantification of allergens in various food products (Poms and
Anklam, 2004).
19.2 Quality assurance for the analysis of allergens

Whatever approach is used for the analysis of allergen-containing food
components, it must be assured that the measurement results are reliable and
relevant for the assessment of the allergic potential of the investigated food.
Therefore the whole, so-called analytical process (Fig. 19.1) has to be properly
Problem definition
Sampling
Processing, conservation
Quality assurance
Analytical sample preparation
Analyte identification
Quantitative measurement
Evaluation
Assessment
Fig. 19.1 Principal steps of an analytical process.

designed. For instance, technologies used in food processing such as roasting

and extrusion can have a significant influence on the availability (solubility)
of specific proteins for allergic reactions. This must be taken into account
when antibodies are raised for immunochemical techniques.
Especially during the 1990s, general concepts and tools have been developed
for assuring the reliability and comparability of measurement results in
quantitative (bio)chemical analysis. It seems to be generally accepted now
that analytical quality assurance (AQA) is based on three main components,
namely the systematic use of validated methods, reference materials and
proficiency testing. Consequently, the general prerequisite for the
implementation of any AQA system is the availability of problem-adjusted
combinations of reference methods and reference materials.
19.3 Towards validated methods for allergen determination

The urgent need for method standardization in this field has been recognized
by the international analytical community and has recently led to the
establishment of a new working group (WG 12) in the technical committee
on food horizontal methods (TC 275) of the European Committee for
Standardisation (CEN). American and European institutions e.g. the AOAC
International Research Institute (AOAC-RI) the European Commission’s
Joint Research Centre (EC-JRC) the German Institute for Standardization
(DIN) Health Canada, and others) have taken the initiative to validate the
analytical performance of several test kits for various food allergens.
However, recent studies showed that quantitative results obtained by the
application of different test methods varied significantly (Hurst et al., 2002;
Koch et al., 2003; Poms and Anklam 2004). These findings underline the
fact that the detection and particularly the quantitative determination of
allergenic residues in food products can be impaired by: (i) interactions with
compounds from the food matrix (e.g., polyphenols and tannins); (ii) reduced
solubility and reactivity of heat-denatured proteins; (iii) differences in antibody
affinity/recognition of allergenic proteins from different species and
geographical origin (Hurst et al., 2002; Hischenhuber, 2002; Keck-Gassenmeier
et al., 1999). A recent inter-laboratory validation study for peanut allergen
detection methods involving more than 30 international laboratories concluded
that quantitative results varied strongly with differences of up to 300% (Poms
et al., 2005).
Only a very few inter-laboratory validation studies for allergen detection
methods have been completed to date, and even their status and conclusions
suffer from the lack of international guidelines for validation of test kits and
of corresponding reference materials.
19.4 Characteristics and use of reference materials

In recent years the harmonization of terminology with respect to reference
materials and the understanding of underlying scientific principles has
considerably improved in the various analytical communities. Nevertheless,
there still exists some confusion in corresponding communications. One
should keep in mind that ISO Guide 30 (ISO, 1992) and the current VIM
(BIPM et al., 1993) list the following definitions with respect to reference
materials:
• Reference material (RM). Material or substance one or more of whose
property values are sufficiently homogeneous and well-established to be
used for the calibration of an apparatus, the assessment of a measurement
method, or for assigning values to materials.
• Certified reference material (CRM). Reference material accompanied
by a certificate, one or more of whose property values are certified by a
procedure which establishes its traceability to an accurate realization of
the unit in which the property values are expressed, and for which each
certified value is accompanied by an uncertainty at a stated level of
confidence.
According to these definitions, CRMs form a subgroup of RMs, namely
those RMs that possess additional characteristics, namely a certificate and
traceable assigned values with an uncertainty statement. It should be noted
that traceability is defined in this context as a property of the result of a
measurement or the value of a standard, whereby it can be related to stated
references, usually national or international standards, through an unbroken
chain of comparisons all having stated uncertainties (BIPM et al., 1993).
There are many different terms in use for the other subgroup, the non-
certified RMs, with a tendency to call them quality control materials (QCM)
(Emons et al., 2004). Certified and non-certified reference materials are used
for method development and validation, calibration, statistical quality control
within a laboratory (mainly for establishing control charts) and to assess the
performance of laboratories by third parties, especially by proficiency testing
schemes. All these applications allow detection of pitfalls of the respective
analytical methodologies and/or their implementation, thus increasing
comparability of measurements between laboratories and improving confidence
in their analytical data.
Any use of a material must be preceded by a definition of its characteristics,
as they determine whether a material is ‘fit for the purpose’. The most
important characteristics of materials for analytical quality control are minimum
sample size to assure in-sample homogeneity, between-bottle homogeneity,
stability during transport, stability during storage, commutability and, if
necessary, the assignment of traceable property values with an uncertainty
statement and the presence of a certificate. In this course, certification of a
reference material is only part of an integrated process (Fig. 19.2) comprising
Feasibility study
Planning Material selection (collection, processing,
characterisation)
Homogeneity Interim transport Material

Storage Processing
study and storage collection
Short-term Long-term Interlaboratory Value assignment

stability study stability study characterisation (certification)
Follow-up stability Assessment

by experts Documentation
monitoring of CRM
CRM distribution and sales
Fig. 19.2 Process of CRM development and production at IRMM.
correct preparation, homogeneity and stability demonstration, and accurate

material analysis that leads to a certified value together with its total uncertainty.
The term ‘commutability’ originates from clinical analysis and describes
the similarity of the analytical response obtained for a given material to the
response obtained from routine samples (ISO, 2003). This term covers all
‘matrix effects’ and all effects that arise from the processing of the candidate
reference material. For example, processing such as fine grinding can facilitate
the extraction of an analyte from a material. Analysis of such a RM would
then not allow a correct assessment of the extraction efficiency of a method
with respect to routine samples.
Most of the parameters required for method validation and for the estimation
of measurement uncertainty can be derived without assigned values. But for
the assessment of the trueness – and consequently accuracy – of a method,
assigned values with a stated uncertainty which are traceable to the same
reference as the analytical results of the method used are required. CRMs
provide exactly this traceable assigned value. Such a certified value is also
required if the material is to be used for calibration. In addition, a minimum
sample size is defined beforehand for CRMs, enabling analysts to know
without testing of their own whether the material is homogeneous enough
for method development. An additional advantage is the ensured stability of
the materials. All these advantages must be balanced against the potential
mismatch between the particular CRM and the routine sample.
Disadvantages of using CRMs for certain purposes result from the
compromises which have to be accepted because of additional material
manipulations to achieve the necessary homogeneity and stability for a CRM.

It has to be decided by the user whether the resulting deviation from the real
sample to be analyzed can be accepted within the QA process.
Though usually overlooked by technical staff, legal issues can play an
important role in the discussion about CRMs and QCMs. The situation is
well established for CRMs: the CRM producer guarantees the accuracy of
the certified property value, including homogeneity and stability for each
unit (bottle) of material. Therefore, a CRM producer who negligently brings
insufficiently characterized materials on the market could be held liable for
any costs arising. For example, if a laboratory validates a method using a
falsely certified CRM, the producer is at least morally responsible for wrong
assessments on real samples based on this validation.
19.5 Towards reference materials for allergens

The availability of reference materials for allergens in food would offer the
potential to harmonize, standardize and calibrate corresponding methods
and to achieve international traceability, i.e. worldwide comparability, of
such measurement results. Furthermore, RMs could offer the basis for further
development and production of antibodies and detection systems. Several
food matrix CRMs are available from producers of reference materials (e.g.
National Institute of Standards and Technology, NIST, from the USA, Institute
for Reference Materials and Measurements, IRMM, from the European
Commission) that may also contain allergens, such as milk powder, egg
powder and peanut butter; however, these materials are certified for components
other than allergens. But so far no material for food allergens which fulfils
the quality criteria described above for CRMs or QCMs has become
commercially available. Moreover, it is important to have real food matrix
materials (not only spiked with extracts of allergenic food) available. Obviously
the development and production of such reference materials for the analysis
of food allergens face several challenges.
19.5.1 The analyte

Firstly, the analyte for food allergen determination is in many cases not
clearly defined. Typically, a number of key proteins and epitopes are known
and described, which cause corresponding reactions. In addition, hypersensitive
patients react to one or more different epitopes and, furthermore, the severity
of reaction may vary between different epitopes of the same protein. Thus
the analyte may be a certain protein or a variety of different proteins specific
for a particular species. Moreover, the spatial structure of these allergenic
proteins may vary, thus exhibiting different in vivo and in vitro reactivity.
Consequently, measurement results of food samples can be influenced by
intrinsic factors (i.e. biological variability such as protein content and food
composition due to geographical and seasonal variability) and extrinsic factors,

such as food processing history and different protein denaturation; but also
the time of harvest and the duration of storage may affect the characteristics
of food allergens, particularly for plant allergens. Additionally, the state of
the extracted analyte also often depends on the extraction method used.
19.5.2 The method

Secondly, there is no appropriate reference method available to date to which
procedures and results are traceable. The currently employed methods in
allergen determination involve either an immunochemical detection or an
indirect species-specific DNA analysis. Protein-based immunochemical
methods are typically very specific for the food commodity concerned or for
a specific allergenic protein; however, the detected analyte is often not well-
defined (binding sites of employed antibodies are mostly unknown and usually
numerous different binding sites are targeted). On the other hand, results
from DNA-based methods cannot be linked to protein/allergen quantities in
a certain food. Promising developments are aiming at specific peptide mapping
employing high-performance liquid chromatography (HPLC) and mass
spectrometry (Shefcheck and Musser, 2004). Current problems associated
with these techniques are the lack of sensitivity and specificity. Moreover,
the matrix of the foodstuffs analyzed may have a strong effect on the reliability
of the method used, which has to be taken into account.
19.5.3 The material

Thirdly, a reference material for a specific food allergen or a particular
allergenic food component needs to fulfill certain quality criteria depending
on its intended use as outlined above. Even non-certified RMs have to be
sufficiently homogeneous, depending on the required sample size of the
method, and stable with respect to the intended time-frame from preparation
to final application. Such parameters have to be checked before an RM can
be made available and used. Therefore, the target properties of ‘allergen
materials’ must be defined in the beginning, and appropriate analytical methods
have to be available for their control. In general, certification of a candidate
reference material will only be possible for a well-defined property based on
either molecular structure of the analyte or a method-defined parameter.
19.6 Future trends

Several attempts are ongoing to break the ‘vicious circle’ concerning the
interrelation between lack of methods and materials for allergen analysis.
For instance, the IRMM in co-operation with Health Canada and the Prolamin
Working Group of Codex Alimentarius has developed a candidate reference
material ‘gliadin from European wheat’ to support quality assurance for

gluten determination. The material is in the process of characterization and
its certification is envisaged for the end of 2005. In addition, IRMM has
recently developed processing techniques for the homogenization of peanuts.
This should allow the preparation of a ‘reference matrix’ for peanut analysis
in the near future. It is foreseen that defined powders will be made available
originating from several peanut varieties which have undergone different
food processing. Further the International Association for Cereal Science
and Technology will make testing materials available that incorporate these
well-defined powders in various food products (e.g. in dark chocolate, cookies,
breakfast cereal flakes). In addition, the US Food and Drug Administration
and the Food Allergy Research and Resource Program at the University of
Nebraska have similar efforts ongoing in the area of manufactured standards.
All those activities should pave the way into the future of appropriate quality
assurance tools in allergen analysis.

• Codex Alimentarius
http://www.codexalimentarius.net/web/index_en.jsp
• IRMM (Institute for Reference Materials and Measurements)
http://www.irmm.jrc.be/
• Health Canada
http://www.hc-sc.gc.ca/
• ICC (International Association for Cereal Science and Technology)
http://www.icc.or.at/
• FDA (US Food and Drug Administration)
http://www.fda.gov/
• FARRP (Food Allergen Research and Resource Program) at the University
of Nebraska http://www.farrp.org
19.8 References
BIPM-IEC-IFCC-ISO-IUPAC-IUPAP-OIML (1993) International Vocabulary of Basic
and General Terms in Metrology (VIM), 2nd edn, Geneva, ISO.
Emons, H, Linsinger, T P J and Gawlik, B M (2004) ‘Reference materials: terminology
and use. Can’t one see the forest for the trees? Trends Anal Chem, 23, 442–449.
Hischenhuber, C, (2002) ‘Analytical methods for the detection of hidden allergens – their
use and limitations’, Minutes and Proceedings of an International Workshop on Food
Allergy: Chemical and Technological Aspects (Ispra, Italy, 2001), EUR 20241 EN
2002.
Hurst, W J, Krout, E R and Burks, W R (2002) ‘A comparison of commercially available
peanut ELISA test kits on the analysis of samples of dark and milk chocolate’,
Immunoassay Immunoch, 23, 451–459.
ISO (1992) ISO Guide 30, Geneva, ISO,

http://www.iso.org/iso/en/CatalogueDetailPage.CatalogueDetail?CSNUMBER=21638.
ISO (2003) ISO 17511, Geneva, ISO,
http://www.iso.org/iso/en/CatalogueDetailPage.CatalogueDetail?CSNUMBER=
30716&COMMID=&scopelist=.
Keck-Gassenmeier, B, Benet, S, Rosa, C and Hischenhuber, C, (1999) ‘Determination of
11, 243–250.
Koch, P, Schäppi, G F, Poms, R E, Wüthrich, R, Anklam, E and Battaglia, R (2003)
Koppelman, J, Bleeker-Marcelis, H, Duijn, G and Hessing, M (1996) ‘Detecting peanut
allergens. The development of an immunochemical assay for peanut proteins’, World
of Ingredients, 12, 35–38.
Poms, R E and Anklam, E (2004) ‘Current trends in detecting food allergens’ GIT Laboratory
Journal Europe, 1, 43–46.
Poms, R E, Ulberth, F, Klein, C L and Anklam, E (2003) ‘Peanut allergens in food
products’, GIT Laboratory Journal Europe, 3, 132–135.
Poms, R E, Klein, C L and Anklam, E, (2004) ‘Methods for allergen analysis in food –
a review’, Food Addit Contams, 21, 1–31.
Poms, R E, Agazzi, M-E, Bau, A, Brohee, M, Capelletti, C, Norgaard, J and Anklam, E
(2005) ‘Inter-laboratory validation study of five commercial ELISA test kits for the
determination of peanut proteins in biscuits and dark chocolate’, Food Addit Contam,
22(2), 104–112.
Shefcheck, K J and Musser, S M (2004) ‘Confirmation of the allergenic peanut protein,
Ara h 1, in a model food matrix using liquid chromatography/tandem mass spectrometry
(LC/MS/MS)’, J Agric Food Chem, 52, 2785–2790.
Taylor, S L, Hefle, S L, Bindslev-Jensen, C, Bock, S A, Burks, A W, jr, Christie, L, Hill,
D J, Host, A, Hourihane, J O, Lack, G, Metclafe, D D, Moneret-Vautrin, D A, Vadas,
much is too much?’, J Allergy Clin Immunol, 109, 24–30.
20
US regulation of undeclared allergens in

food products
M. Hahn, Hogan & Hartson LLP, USA
20.1 Introduction
A company selling food products in the US is vulnerable to potential regulatory
and product liability when the food product causes injury due to the presence
of an undeclared allergen. There is potential regulatory liability because the
US Food and Drug Administration (FDA) has long taken the position that
foods should be recalled when they are found to contain an undeclared major
allergen. In addition, the Food Allergen Labeling and Consumer Protection
Act, signed into law in August 2004, deems a food product misbranded if the
label fails to declare a major allergen by a common English name such as
‘milk,’ ‘egg,’ or ‘peanut.’ Once a company discovers that one of its products
contains an undeclared major allergen, either through its own due diligence,
in response to a consumer complaint, or in response to an FDA inquiry, the
company generally will recall the offending product. Given the severity of
reaction that can occur when a food-allergic consumer is exposed to a major
allergen, companies generally must issue press releases announcing the recall.
A company can easily spend hundreds of thousands of dollars recalling
products with an undeclared allergen.
The sale of a food product containing an undeclared allergen also presents
potential product liability. Americans have long been known for their fascination
with lawsuits and their propensity to resolve disputes through the court
system. The media carries numerous stories of consumers who have been
harmed from food and are seeking huge damages from restaurants or food
manufacturers. Indeed, customers have sued food restaurants for serving hot
coffee that caused a serious burn and for serving food that allegedly contributed
to obesity. There have also been hundreds of cases that have been brought
against the food industry for selling foods that contained pathogens such as
Salmonella or E. coli that have caused consumer injury. Numerous cases
have also been filed against the food industry for selling a food that contained
an undeclared major allergen. While the vast majority of cases are settled out
of court, the few reported cases available indicate that the potential liability
for selling a food with an undeclared allergen can range from a few thousand
to over a million dollars.
The food industry has invested considerable time and resources since the
1990s in attempting to reduce the presence of undeclared allergens in the
food supply. Other chapters in this publication focus on the various ways in
which undeclared allergens can become integrated into the food supply and
the manufacturing controls that can be implemented to reduce them. Indeed,
many companies are now using allergen test kits to monitor for the presence
of undeclared allergens that could be incorporated into foods by use of
shared equipment. These test kits can be incorporated effectively into a
company’s allergen control practices by identifying the presence of an
undeclared allergen that otherwise is not visible to the naked eye. Companies
must realize that the results from the test kit may one day be available to
regulators. If the company becomes involved in a product liability action
resulting from an alleged injury due to an undeclared allergen, the allergen
test results would also need to be made available to lawyers that are
bringing the action against the company. Although a powerful tool for
incorporation into an allergen control program, companies must give careful
consideration to how they will incorporate test kits into their allergen control
program.
20.2 Regulatory liability

20.2.1 Misbranding provisions
The Federal Food, Drug, and Cosmetic Act (FFDCA) prohibits the introduction
into interstate commerce of foods products that are misbranded.1 The FFDCA
deems a food misbranded unless its label bears an ingredient statement
identifying each ingredient that has been added to the product by an
appropriately descriptive common or usual name.2 Spices, flavors, and colors
that do not have to be certified are exempt from the ingredient labeling
requirements and may be declared by generic terms such as ‘spice,’ ‘flavor,’
‘natural flavor,’ ‘artificial flavor,’ or ‘artificial color,’ as appropriate.3
The FDA regulations exempt incidental additives from the ingredient labeling
requirement.4 To be eligible for regulation as an incidental additive, and
therefore be exempt from the ingredient labeling requirements, a substance
must be present at an insignificant level and have no technical or functional
effect in the food.5 Examples of incidental additives include substances that
become incorporated into a food because they are present in a sub-ingredient
(such as carriers for flavors and colors), processing aids (such as enzymes
US regulation of undeclared allergens in food products 359
that are used during the manufacturing process but then later removed), and
substances migrating to food from equipment and packaging materials.
The Food Allergen Labeling and Consumer Protection Act establishes
new requirements for the labeling of major allergens. The President signed
the law in August 2004 and its provisions become effective for products
labeled after January 1, 2006.6 At the time of this article, FDA has not yet
provided any guidance on how it intends to interpret this new law. The new
law amends the FFDCA in two significant ways. First, it defines the term
‘major food allergen’ and second, it deems a food misbranded unless the
major food allergens are declared by a ‘common English name’ that is easily
recognizable by consumers. In large part, the legislation is limited to what
are commonly called the ‘Big 8 allergens.’ ‘Major food allergen’ is defined
as ‘milk, egg, fish (e.g. bass, flounder, or cod), crustacean shellfish (e.g.
crab, lobster, or shrimp), tree nuts (e.g. almonds, pecans, or walnuts), wheat,
peanuts, and soybeans’ and food ingredients that contain proteins derived
from the major food allergens.7 These proteins nevertheless may be excluded
from this definition if they fall under one of two exceptions.
First, highly refined oils, and ingredients derived from these oils, are not
a ‘major food allergen.8 These “highly refined oils” are refined, bleached,
deodorized oils. In its report, the Committee notes that the legislation would
not change the common or usual name of these highly refined oils. Thus,
oils, such as peanut oil, would still be labeled as such in ingredient statements.
Second, a manufacturer may obtain an exemption for ingredients derived
from major allergens by submitting a premarket notification or petition
demonstrating that such ingredient should not be considered a major allergen.9
The premarket notification procedure would be available when (i) scientific
evidence establishes that the food ingredient does not contain allergenic
protein or (ii) FDA has made a determination under a Section 409 premarket
approval or premarket notification program that the ingredient does not cause
an allergic response that poses a risk to human health.10 The Committee
explains that the Generally Recognized as Safe (GRAS) notification process
is not included as part of this exception.
It is difficult to predict how the agency will implement the allergen law,
particularly when it comes to its application to processing aids and other
incidental additives that are derived from major allergens. Ingredients such
as soy lecithin, fish gelatin, and wheat starch have been used for decades as
processing aids and have been exempt from ingredient labeling requirements.
While these ingredients may have detectable levels of protein, they are
frequently used at very low levels in finished foods. In many instances the
level of protein in the finished food contributed by the use of an ingredient
such as soy lecithin can be well below 1 part per million.
It would be reasonable for FDA to take the position that allergenic proteins
present at de minimis, or very low levels, essentially are not present in
finished foods and should not be subject to the allergen labeling requirements.
Indeed, the courts have applied a de minimis concept with regard to the
regulation of food additives that may be present at very low levels in foods.11
At the time of this article, FDA has not offered guidance on whether it would
apply a de minimis analysis and exempt from the allergen labeling requirements
those food ingredients derived from major allergens that contribute insignificant
levels of allergenic protein in finished foods.
Moreover, even if FDA does agree that a de minimis analysis is appropriate,
it remains unclear what level of allergenic protein will be considered a de
minimis level. The absence of an established threshold below which allergens
will not cause an allergic reaction complicates the agency’s ability to establish
a de minimis level. There are factors, however, that would support the
establishment of 10 parts per million or a similar level as the de minimis
level. This is the level identified by some experts as being unlikely to trigger
an allergenic reaction in most food-allergic consumers. It also is the level
that has been used informally by many in the food industry for years in
determining whether an allergen should be declared. In addition, 10 parts per
million is reasonably close to the one to five parts per million analytical level
found in some of the commercially available allergen test kits.
The de minimis concept not only is important when determining whether
a food ingredient should be subject to the allergen labeling requirements, but
also could apply when determining if an ingredient is eligible for the notification
or petition process. The notification process is available in those instances
when a manufacturer can demonstrate that an ingredient derived from a
major allergen, although containing protein, does not contain an allergenic
protein. The law does not comment, however, on what is meant by ‘containing
allergenic protein.’ The law does not comment on whether the agency should
consider the limit of detection of the most sensitive analytical method or the
intended use of the ingredient and the level of allergenic protein that would
theoretically be present in the finished food through such use.
It presumably would be reasonable for the agency to focus on the manner
in which the ingredient is used in the finished food and the theoretical levels
of allergenic proteins that would be present through such use. In instances
when the ingredient would contribute a de minimis level of allergenic protein
under its intended conditions of use, it would be reasonable for the agency
to take the position that the ingredient does not contain allergenic protein
and is eligible for the notification process. If the agency instead focuses on
whether the ingredient contains a detectable level of allergenic protein, there
would be very few ingredients that would be eligible for the notification
program. Under such an interpretation, an ingredient such as soy lecithin
that may contain less than 50 parts per million of total protein would be
ineligible for the notification process.
A petition process is established for those ingredients derived from major
food allergens that are ineligible for the notification program.12 An exemption
would be obtained by submitting a petition with data demonstrating that the
ingredient does not cause an allergic response that poses a risk to human
health. The petitioner would bear the burden of proof. The petition would be
deemed denied after 180 days unless the petitioner and FDA mutually agree
to an extension. FDA is required to post to a public site the notifications and
petitions that have been received and any responses or objections thereto.
Once within the definition of ‘major food allergen,’ ingredients would be
subject to the provisions of the law regardless of whether they had previously
been treated as ‘incidental additives.’ If the agency is unwilling to apply a de
minimis analysis or find some other legal means to interpret the allergen law
flexibly so it is possible to obtain an exemption for processing aids that
contribute insignificant levels of allergenic proteins in finished foods, there
will be many food products that will be subject to the allergen labeling
requirements. Indeed, products containing soy lecithin, fish gelatin, wheat
starch, and other common processing aids would suddenly be subject to
allergen labeling. The food-allergic consumer in many instances has been
consuming these products without incident and will suddenly be presented
with a label disclosing the presence of the major allergen in the food.
In addition to defining major food allergen, the new law amended Section
403 of the FFDCA to provide three new misbranding provisions. The first
provides FDA the authority to require by regulation the appropriate labeling
of any spice, flavoring, coloring, or incidental additive that includes, as a
constituent, a food allergen that is not one of the major food allergens. The
second requires colors, flavors, and incidental additives that contain major
allergens to be subject to the labeling requirements for major allergens. The
third requires the eight major food allergens to be labeled on foods that are
not raw agricultural products.
Manufacturers will have the following two options for identifying the
eight major food allergens on the food label:13
1. By placing the statement ‘Contains ’ with the blank filled in with
the name of each major food allergen, at the end of, or immediately
adjacent to, the ingredient statement; or
2. By using a parenthetical following the name of the ingredient that identifies
the name of the major allergen, such as ‘casein (milk).’ The parenthetical
would not be required when the ingredient name uses the name of the
major food allergen (such as ‘milk’ or ‘soy protein’) or when the name
of the major food allergen appears elsewhere in the ingredient statement
(unless the other listing is for a food ingredient that the Secretary has
determined does not cause an allergenic response).
Regardless of the option selected, the major food allergens would need to be
identified by the name of the food source from which the major allergen is
derived, such as milk, egg, wheat, peanut, or soybean. The specific type of
tree nut (e.g., almonds, pecans, or walnuts), crustacean shellfish (e.g., crab,
lobster, or shrimp) or fish (e.g., bass, flounder, or cod) would also need to be
identified.
The agency has for several years had an allergen labeling policy that
applies to the same ‘major allergens’ that are covered by the allergen labeling
law.14 The Allergen Labeling Policy is of lesser significance today because

it has been superseded by the Food Allergen Labeling and Consumer Protection
Act. It, nonetheless, is appropriate to review the policy because it remains in
effect until the allergen legislation becomes effective in January 2006.
The FDA Allergen Policy applies to the labeling of ingredients that are
added directly to the food and to ingredients that may otherwise be eligible
for an ingredient labeling exemption. FDA states:
Ingredients Added Directly to Foods: Products which contain an allergenic

ingredient by design must comply with 21 U.S.C. 343(i)(2). Where
substances that are, bear, or contain allergens are added as ingredients or
sub-ingredients (including rework), the Federal Food, Drug, and Cosmetic
Act (the Act) requires a complete listing of the food ingredients (section
403(i)(2); 21 U.S.C. 343(i)(2); 21 C.F.R.101.4) unless a labeling exemption
applies.
Exemptions from Ingredient Labeling: Section 403(i)(2) of the Act
provides that spices, flavors, and certain colors used in a food may be
declared collectively without naming each one. In some instances, these
ingredients contain sub-components that are allergens. FDA’s regulations
(21 CFR 101.100(a)(3)), provide that incidental additives, such as processing
aids, which are present in a food at insignificant levels and that do not
have a technical or functional effect in the finished food are exempt from
ingredient declaration. Some manufacturers have asserted to FDA that
some allergens that are used as processing aids qualify for this exemption.
FDA, however, has never considered food allergens eligible for this
exemption. Evidence indicates that some food allergens can cause serious
reactions in sensitive individuals upon ingestion of very small amounts;
therefore, the presence of an allergen must be declared in accordance with
21 CFR 101.4. The exemption under 21 CFR 101.100(a)(3) does not
apply to allergenic ingredients.15
FDA has clarified that it considers products that do not comply with this
policy misbranded. FDA reasons ‘the article was misbranded when introduced
into and while in interstate commerce and is misbranded while held for sale
after shipment in interstate commerce, within the meaning of the section
403(i)(2) of the Act, in that it is fabricated from two or more ingredients, and
its label fails to bear the common or usual name of each such ingredient,
namely (specify the undeclared allergenic ingredient).’16
Since 1996, FDA has stated ‘processing aids that contain allergenic
ingredients must be declared in accordance with 21 CFR 101.4(a)(1).’17
Under this FDA policy, a substance that may be present at very low levels
and have no technical or functional effect on the food would be subject to
ingredient labeling if it contains a protein from a major allergen.
20.2.2 Adulteration provisions

The FFDCA also prohibits the introduction into interstate commerce of foods
products that are adulterated.18 A food is deemed adulterated for many reasons,
including if it has been prepared, packed, or held under insanitary conditions
whereby it may have become contaminated with filth, or whereby it may
have been rendered injurious to health.’19
The FDA Allergen Policy addresses those allergens that may become
inadvertently incorporated into food during the manufacturing process. FDA
recognizes that allergens may be unintentionally added to food ‘as a result of
practices such as improper rework addition, product carry-over due to use of
common equipment and production sequencing, or the presence of an allergenic
product above exposed product lines.’20 FDA notes ‘such practices with
respect to allergenic substances may be insanitary conditions that may render
the food injurious to health and adulterate the product under section 402(a)(4)
of the Act [21 U.S.C. 342(a)(4)].’21 FDA states such a product is ‘adulterated
when introduced into and while in interstate commerce and is adulterated
while held for sale after shipment in interstate commerce, within the meaning
of the section 402(a)(4) of the Act, in that it has been prepared, packed and
held under insanitary conditions whereby it may have been rendered injurious
to health.’22
The food industry routinely uses the same equipment to manufacture
different products in much the same way that consumers will use the same
mixers, cake pans, cookie trays and utensils when baking cookies and cakes.
Given the ability of allergens to produce an allergic reaction when present at
very low levels – and certainly levels below what can be seen visibly –
allergenic proteins can adhere to mixers, utensils, and baking pans, and other
food contact surfaces and be released in subsequently manufactured products.
The unintentional contamination can occur in the food manufacturing facility,
in restaurant kitchens, and in homes.
The courts have never been asked to address whether FDA can invoke
section 402(a)(4) (21 U.S.C. § 342(a)(9)) to require the labeling of allergenic
substances that may become incorporated into foods through inadvertent
cross-contact during the food production process. Manufacturers that fail to
implement practices to control the inadvertent cross-contact of major allergens
face potential regulatory action. In general, the food industry has implemented
numerous practices and procedures to reduce the likelihood that allergens
will be introduced through inadvertent cross-contact. There may be some
instances, however, when cross-contact cannot be avoided even when following
state-of-the-art manufacturing and cleaning practices. In instances when cross-
contact cannot be avoided when companies are following good manufacturing
practices (GMPs), FDA will allow companies to use precautionary labeling
such as ‘may contain ______’ with the blank filled in with the name of the
allergen that may be present in the food through inadvertent cross-contact.
FDA has made it clear that it will allow ‘may contain’ statements provided
manufacturers are not using them in lieu of GMPs.
20.2.3 Product recalls

A company that has introduced into interstate commerce a food product that
contains an undeclared allergen will be subject to potential FDA regulatory
action. Although the agency does not have the legal authority to mandate
product recalls, the agency does have the legal authority to seek judicial
intervention through an injunction against the company or a seizure of the
products.23 The Bioterrorism Statute promulgated in June 2002 expanded
FDA’s authority to include administrative detentions as well.24 FDA can also
generate tremendous publicity in the event that the company refuses to recall
a product that contains an undeclared allergen. Given the powerful enforcement
tools and the power of the press available to the agency, it would be
unprecedented for a company to refuse to recall a food product when requested
to do so by the agency.
Under FDA’s administrative procedure regulations, a recall is defined as
‘a firm’s removal or correction of a marketed product that the Food and Drug
Administration considers to be in violation of the laws it administers and
against which the agency would initiate legal action, e.g., seizure.’25 There
are three classifications of recalls, Class I, Class II, Class III, with Class I
being the most serious. A Class I recall is defined as ‘a situation in which
there is a reasonable probability that the use of, or exposure to, a violative
product will cause serious adverse health consequences or death.’26 Almost
all recalls involving major food allergens, with the exception of wheat, will
result in a Class I recall. Undeclared wheat generally will result in a Class II
recall because, although wheat can trigger an anaphylactic reaction, it has
not been reported to cause death, as has been reported with exposure to the
other major allergens.
During a Class I recall, FDA ordinarily will require the company to issue
a press release notifying the consumers about the recall. These press releases,
when picked up and broadcast by the media, are an effective tool in notifying
the food-allergic consumer that he or she needs to return the product. The
publicity generated by a recall can adversely impact the company’s image
and the cost of recovering foods from the marketplace can easily exceed
$ 100 000, depending on the quantity of food introduced into commerce and
the breadth of the distribution. The publicity generated by recalls also notifies
lawyers that there may be a harmed consumer or consumers that could serve
as the basis of a legal challenge against the company.
20.3 Legal grounds for product liability actions

The presence of undeclared allergens in the food supply will present potential
product liability. It is difficult to quantify the potential product liability risk,
however, because there are relatively few reported cases. Many companies
will settle potential product liability actions involving undeclared allergens
before cases are filed. These settlements will be done quietly and generally
will include as a condition of settlement a clause prohibiting the injured

consumer from disclosing the terms of the settlement.
Of the few cases that are filed, many are settled before the trial and once
again the terms of the settlement generally prohibit disclosure of the terms.
These confidentiality clauses make it difficult to fully assess the damages
that a company may incur. The difficulty is further exacerbated by the nature
of product liability laws which generally are brought under state law. The
liability, therefore, potentially may vary from state to state, depending on the
product liability laws of the jurisdiction.
Although state product liability laws differ, there are similarities in product
liability actions. Under most state laws, a consumer will be able to sue a
company to recover damages for injuries allegedly caused by food allergens,
which, when consumed, have caused severe or life-threatening adverse
reactions. These plaintiffs sue under a product liability theory, which refers
to a variety of legal claims, including negligence, breach of warranty, and
failure to warn. These claims, which are not mutually exclusive, are used by
plaintiffs who are seeking damages for injuries suffered because of alleged
‘defects’ in the goods they have purchased.
20.3.1 Overview of product liability cases and negligence

An entity that markets a product in the US faces potential product liability
exposure to the extent that the product causes harm to consumers. Generally,
the aggrieved party must bear the burden of proof in demonstrating that the
product was ‘defective’ and that the defect caused the individual’s injuries.
There are numerous causes of action under which a plaintiff can bring a
product liability action, including negligence, breach of express and implied
warranty or misrepresentations. In negligence actions, an aggrieved party
would need to demonstrate that the company marketing a product failed to
meet the standard of care when making and marketing the defective product.27
In an action alleging breach of express or implied warranty, the aggrieved
party alleges that he or she had an express or implied contract with the
manufacturer of the allegedly defective product and that the company breached
the terms of that contract.
The traditional theories of product liability would present potential hurdles
in the instance of undeclared allergens in the food supply. In many instances,
companies could allege that they were in fact following standard industry
practice and that the injury resulted from no fault of the company. For example,
a chocolate manufacturer could claim that it should not be held negligent for
selling a chocolate bar that contained trace levels of peanuts in an instance
when the company implemented state-of-the-art allergen controls, had data
demonstrating that its allergen controls were effective, that it adhered to the
allergen controls and that due to an unforeseeable circumstance, a trace level
of peanut found its way into the product. Under these circumstances, it could
be difficult to hold the food company liable under a theory of negligence or
breach of warranty.
20.3.2 Strict liability

The rules changed in the 1950s, and for the first time the courts recognized
that a company could and should be held strictly liable for harm caused by
its product even in the absence of negligence on the company’s part.28 The
theory of strict liability is applied to many industries, including the food
industry. Under a theory of strict liability, a plaintiff essentially has to prove
one of three key elements (i) there was a mistake in the manufacturing
process, (ii) the manufacturer formulated or designed the product improperly,
or (iii) the manufacturer placed the product into commerce without adequate
warnings. The element that likely will be involved in most allergen cases is
the third element regarding the company’s failure to provide adequate warnings
about the potential danger.
Perhaps most troubling for the food industry is that liability can attach in
the absence of any negligence. There are numerous theories of ‘strict liability’
that can be used by plaintiffs to recover damages even when the defendant is
found to have exercised due care in the conduct of his or her business. In
most strict liability actions, the issue is not whether the food manufacturer
acted negligently, but whether the manufacturer provided its customers with
sufficient warnings about the product. For example, in the prior example
regarding the chocolate company, although that manufacturer could not be
held liable under a theory of negligence, the company could be held liable
under a theory of strict liability.
In product liability actions against the food industry, the issue is generally
whether the company had a responsibility to provide an adequate warning
and, if so, whether the company provided an adequate warning. The Restatement
Second of Torts, a treatise that many judges use when determining whether
liability should attach in certain instances, specifically addresses the warning
requirements for food allergens. The Restatement provides
In order to prevent the product from being unreasonably dangerous, the

seller may be required to give directions or warning, on the container, as
to its use. The seller may reasonably assume that those with common
allergies, as for example to eggs or strawberries, will be aware of them,
and he is not required to warn against them. Where, however, the product
contains an ingredient to which a substantial number of the population are
allergic, and the ingredient is one whose danger is not generally known,
or if known is one which the consumer would reasonably not expect to
find in the product, the seller is required to give warning against it, if he
has knowledge, or by the application of reasonable, developed human
skill and foresight should have knowledge, of the presence of the ingredient
and the danger.29
The Restatement makes it clear that there will be circumstances in which

the seller of foods should provide a warning that the product contains an
allergen. It is important to note that the restatement applies to all sellers of
foods and, as such, imposes potential liability on manufacturers of packaged

foods as well as to restaurants and others that provide food directly to
consumers. A warning is not required under the Restatement when the presence
of the allergen is obvious. For example, a restaurant that serves a chef salad
with large chunks of egg does not need to provide a warning that the entrée
contains eggs because the food-allergic consumer would see the eggs and be
aware of their presence.
A warning arises when the presence of an allergen is not obvious to the
consumer. First, the Restatement requires a substantial portion of the population
to be allergic to the ingredient. It is safe to assume that the Restatement
would apply to warnings for the major allergens. It is less clear, however,
whether the Restatement would apply to warnings for those foods which are
known to be allergenic to only a very limited number of individuals.
Unfortunately, there is no guidance in the Restatement or in the case law
whether a warning would be required for allergenic ingredients to which
very few are allergic. A seller reasonably could argue that a warning is not
required to the extent that a consumer has an allergy that affects only a very
small number of individuals. There can be no assurance, however, that the
courts would agree and the issue ultimately may need to be treated as a
factual issue that must be decided by a judge or jury.
If there are a substantial number of people in the population that are
allergic to the ingredient, a warning would be required when the consumer
reasonably would not expect to find the allergen in the product and the seller
has knowledge, or by the application of reasonably developed human skill
and foresight, should have knowledge of the presence of the ingredient and
the danger. A warning would be required, for example, in instances when a
manufacturer develops a new ingredient that contains an allergenic component.
For example, if a company develops a new fat substitute from peanut protein,
the marketer of this product should warn the peanut-allergic consumer of the
presence of the peanut protein. This could be accomplished through the
product name (e.g., Acme Peanut Derived Fat Substitute) or through a warning
statement on the food label (e.g., Attention Peanut Allergic Community:
Acme Fat Substitute is Made from Peanuts), or through an ingredient statement
disclosing the presence of peanuts in the product (e.g., Acme Fat Substitute
(contains peanuts)).
A warning may also be required to inform a consumer of the presence of
an allergen in a food that would not ordinarily be expected to be in such
food. For example, assume that a restaurant adds peanut butter to its chili.
The restaurant should consider providing notice to consumers even if peanut
butter is the ‘secret ingredient,’ because the consumer would not ordinarily
expect to find peanut butter in a chili. In addition, a warning such as ‘may
contain peanuts’ could be required when an allergen becomes a component
of food through inadvertent cross-contact.
As of the time of this article, there are few cases that provide any guidance
on when a warning should be required. In one case, a plaintiff sued McDonald’s
claiming that the restaurant had a duty to warn of the presence of carageenan
in its McLean hamburger.30 McDonald’s filed a summary judgment motion
requesting that the case be dismissed, in part, because the company made
available to consumers upon request brochures that identified the carageenan
in the product. The trial court granted the motion without issuing a decision.
The plaintiff appealed and the appellate court reversed the trial court decision
finding issues of ‘material fact upon which reasonable minds could differ.’
While acknowledging that McDonald’s listed ‘carageenan’ as an ingredient
in a flier that it made available to consumers, the appellate court noted that
the flier ‘does not inform the reader that carageenan is derived from seaweed,
nor that persons who are allergic to seaweed may experience an adverse
reaction to that ingredient.’31 The court also noted that the plaintiff purchased
the sandwich at a drive-through window and there was no evidence as to
whether she received a flyer or that it was offered to her at that time. Under
the applicable Ohio statute, a jury reasonably could find, for instance, that a
manufacturer should have known of a risk to even a small number of consumers.
The statute asks
whether a manufacturer exercising reasonable care would warn of the risk

in light of both the likelihood and the seriousness of the potential harm.
Within this framework, whether the plaintiff’s harm was unusual or not
would be a factor in calculating whether a manufacturer exercised reasonable
care in its decision not to warn. The incidence of the kind of harm at issue
in this case is only one factor a jury would consider in finding a duty to
warn.32
McDonald’s also claimed that there was no evidence indicating that the
carageenan in its hamburger caused the plaintiff’s reaction. The appellate
court rejected this line of argument noting that the plaintiff had alleged that
the product caused the reaction and that she supplemented this evidence with
a note from a treating physician identifying the hamburger as the cause of
her reactions. Because the plaintiff offered evidence ‘tending to prove causation,
the element remains a question for the jury.’33 Noting that the plaintiff had
raised genuine issues of material fact, the appellate court held that it would
be up to a jury to decide whether McDonald’s should be liable for the
plaintiff’s harm.
The McDonald’s case is recognizably troubling because McDonald’s did
in fact identify the presence of carageenan as an ingredient in flyers made
available to consumers and because carageenan is not known or recognized
as an allergen. McDonald’s likely reached an out-of-court settlement after
losing the appeal. Had the case gone to trial, McDonald’s then could have
presented evidence establishing that (i) its flyer constituted adequate warning
and (ii) that its hamburger did not cause the plaintiff’s injury. It would have
been up to the judge or jury, however, to decide these issues of fact.
20.3.3 Damages
The failure to provide adequate warnings to allergens exposes the food and
restaurant industries to potential product liability lawsuits. Plaintiffs that
have suffered injury from consuming the allergen may decide to file lawsuits
seeking damages against the food manufacturer or restaurant. Damage awards
can include both compensatory damages and punitive damages. Compensatory
damages are intended to compensate the consumer for injuries, which may
include medical expenses and the amorphous ‘pain and suffering.’
A food company or restaurant can also be liable for punitive damages if
its failure to warn is judged to be the result of what is known as ‘actual
malice’ or a ‘reckless disregard’ for the safety of customers. Although the
definitions of ‘actual malice’ and ‘reckless disregard’ vary by jurisdiction,
the general concept describes a situation where the defendant was aware of
the dangers posed by his conduct but went ahead anyway. For example,
punitive damages might be imposed where a server had serious doubts about
the truth of his answer when he told a customer a dish did not contain a
particular ingredient, but he chose to ignore those doubts, thereby endangering
the customer. Punitive damages also could be awarded, for example, if a
company becomes aware that it failed to include eggs in the ingredient
statement, but fails to recall the product. Although no cases could be found
when a company has been held to pay punitive damages, these are the types
of activities that in some jurisdictions may expose a company to a punitive
damages award. Importantly, punitive damages could end up being many
times the amount of actual damages sustained.
Other damages that should not be overlooked are the adverse publicity,
the lost business that may result from a lawsuit, and the legal fees associated
with defending the action. It is the potential for adverse publicity that likely
is responsible for the large number of cases that are settled out of court. With
regard to lost business, in some instances the loss of business has caused
companies to go out of business.
20.3.4 Summary of product liability actions

There have been numerous product liability actions in the US resulting from
the presence of undeclared allergens in foods. Many of the actions are brought
against restaurants because, in many instances, the waiting staff have failed
to provide accurate information about the allergens in the food. Other actions
have been brought against packaged food manufacturers for marketing products
with labels that do not identify all of the allergens. The tables in the appendices
track many, although not all, of the product liability actions that have been
filed. Appendix I summarizes cases that have been filed involving known
allergens. Appendix II summarizes cases involving ingredients that are not
known allergens, but nonetheless have been alleged to cause an adverse
reaction. A review of the cases reveals that damages can range from zero to
well over one million dollars.
20.3.5 Food and restaurant industry responsibility and due diligence

Many of the actions brought against restaurants are the result of misinformation
that has been communicated by the waiting staff to the customers. All restaurants
should educate their employees about the serious adverse reactions that can
result if an individual with a food allergy is exposed to an allergen. The staff
must realize that allergic reactions can be severe and in many instances
produce life-threatening adverse reactions. If the waiting staff has a better
understanding of the severity of the reaction, they be more inclined to take
seriously customer comments about allergies to food ingredients.
Restaurants and food service providers should also designate at least one
person, preferably a manager, as a ‘point person’ to respond to customer
questions about food allergies. This person should be well-informed of the
ingredients that are used to prepare the foods and of the severe reactions that
can occur when the food-allergic consumer is exposed to an allergen. The
waiting staff must be instructed to direct all questions about food allergies or
intolerances to the point person. The point person must be instructed to
provide accurate information about the ingredients in the food, even if it
means disclosing a secret ingredient.
There are also numerous steps that food companies can take to reduce
their risk of liability. First, companies should make certain that the ingredient
statement accurately reflects the composition of the product. Companies
should also become familiar with the requirements of the Food Allergen
Labeling and Consumer Protection Act and with FDA policy on the labeling
of allergens. Companies should develop an allergen control policy that identifies
the manufacturing, procurement, processing, packaging, labeling, and other
issues that will be implemented to address allergens. In instances when
inadvertent cross-contact with allergens cannot be controlled through GMPs,
companies should use a label statement such as ‘may contain ___’ with the
blank filled in with the major allergen that may be present in the food. This
type of warning should provide the food-allergic consumer with adequate
notice of the potential for the food to contain an allergen that needs to be
avoided.
Regular audits of the labels are advisable to ensure that all ingredients are
declared. This can be a daunting task because many food products have
labels with 10, 20, 30, or even more ingredients. The finished food manufacturer
needs to review carefully the ingredients that it adds to the product as well
as the ingredient decks on products that it purchases from outside vendors.
This audit should be performed whenever a change is made to the formulation
of the product or when new labels are designed.
Companies should also conduct an internal audit of the manufacturing
process. The allergenic ingredients in the manufacturing process must be
identified and the process should be carefully analyzed to identify any instances
where cross-contamination could occur. If the same machinery is used to
manufacture products with and without allergenic ingredients, there is a
potential for allergen residues to end up in subsequently manufactured product.
There also is a potential for cross-contamination when product lines cross,

unless they are covered. In the event that cross-contamination can be avoided
by good manufacturing practices, the manufacturer should institute procedures
to address the issue. If not, the manufacturer should consider labeling that
will inform the consumer of the presence of the allergen, such as ‘may
contain peanuts.’
Manufacturers should consider whether they will incorporate allergen test
kits into their allergen control program. Before using the test kit, however, it
is imperative that the company develop a firm policy on how the test kits will
be used. This policy should address, for example, whether the test kits will
be used to test finished product, products not intended for commercial sale,
or the manufacturers’ cleaning operation. If the manufacturer tests finished
product, it would be advisable to establish procedures that identify when the
product will be tested and specify that the product should be retained until
the test results are received. The manufacturer should also consider the level
of detection of the test kits. While it may be tempting to use the most
sensitive method of analysis, there may be little value in using a method that
can detect allergenic proteins below one to five parts per million, a level that
may very well be well beyond the threshold for a safety concern.
In the event that manufacturers test finished product, find detectable levels
of allergenic protein, and allow the products to enter commerce, the analytical
results could be made available to regulators or plaintiff lawyers in any
product liability action resulting from any consumer harm from the consumption
of the product. Although it is difficult to imagine that any reputable company
would release products under this factual situation, if a company did find
itself in litigation it could be facing not only compensatory but potential
punitive damages as well.
Conducting an analysis on the finished product is the best way to ascertain
whether a company has adequate cleaning procedures to ensure that allergens
do not end up in subsequently manufactured foods. Identification of the
allergen in the finished product, however, would preclude the company from
introducing that product into interstate commerce. To minimize the adverse
consequences of a positive test, companies may want to scale back their
batch sizes so that the consequences of a positive test can be minimized.
Alternatively, companies could test product that is not intended for introduction
into interstate commerce. Other companies may limit allergen testing to
cleaning materials (such as push through materials in dry cleaning operations)
or swabbing of food contact surfaces.
It could be difficult to protect the confidentiality of the test results by
having the allergen test results conducted by an in-house or outside lawyer.
While most states will maintain confidentiality of materials developed by
lawyers in anticipation of litigation, including analytical results, few
jurisdictions may be inclined to extend the protection in an instance when
the lawyer routinely analyzes the product and the product is released after
the lawyer receives results from the analysis. It could, therefore, be difficult
to maintain the confidentiality of the test kit results by having the analysis
performed by an in-house or outside lawyer. A company, therefore, should
assume that the allergen analytical results would need to be made available
to a plaintiff that is suing the company for injuries resulting from the alleged
presence of undeclared allergens in the food.
Allergen test kits provide the industry with a useful tool in controlling the
presence of undeclared allergens. Careful thought and consideration, however,
must be given to how the company will incorporate the test kits into its
overall allergen management program.
20.4 Future trends

The issues presented by allergens in foods are continuing to receive a great
deal of attention in the US by regulators and legislators. This issue received
the attention of the US Congress and legislation has been passed that requires
the declaration of major allergens on food labels by common English names.
The legislation directs the FDA to prepare numerous reports on major food
allergens. By February 2006, the FDA is required to submit reports that: (i)
analyze the ways in which foods are unintentionally contaminated with major
food allergens (cross-contact) and estimate how common these practices are
in the food industry; (ii) advise whether good manufacturing practices or
other methods can be used to reduce or eliminate cross-contact of foods with
the major food allergens; (iii) describe the various types of advisory labeling,
the conditions of manufacturing that are associated with the various types of
advisory labeling, and the extent to which advisory labels are being used on
food products; (iv) determine how consumers with food allergies would
prefer that information about the risk of cross-contact be communicated on
food labels; (v) state the number of inspections of food manufacturing and
processing facilities conducted in the previous two years and discuss the
findings of these inspections; and (vi) assess the extent to which the Secretary
and the food industry have effectively addressed cross-contact issues.
The legislation also requires the FDA, within two years, to issue a proposed
rule that would define and permit the use of ‘gluten-free’ on food labels. The
reports that must be prepared by the FDA and the mandate to establish gluten
labeling will ensure that the FDA will continue to focus its resources on
allergen labeling issues.
The FDA will also continue to conduct allergen inspections of the food
industry and will take regulatory action when appropriate. Manufacturers
that have introduced foods containing undeclared allergens will also continue
to face potential recalls.
In the event that consumers are injured by an undeclared allergen in the
food, the industry should anticipate potential product liability. One of the
unintended consequences of the recent media attention on the legislative and
regulatory allergen initiatives is that it increases the awareness in the plaintiff’s
bar of the potential for a new and potentially lucrative cause of action to be
brought against the food industry. Because test kits are available to the
general public, savvy plaintiff’s lawyers could begin testing products for the
presence of undeclared allergens and then attempt to bring class action lawsuits
against companies marketing products with undeclared allergens. The number
of product liability actions brought against the food industry for undeclared
allergens may very well increase, unless the industry is able to implement
procedures to reduce the release into commerce of products with undeclared
allergens.
20.5 Conclusion
The presence of undeclared allergens in the food supply is proving to be an
issue that subjects the food industry to potential product liability. Restaurants
can reduce their potential liability by educating their waiting staff of the
importance of providing consumers with accurate information about products.
Restaurants also should consider ways in which they can inform the food-
allergic community of the presence of allergens in products where they
would not ordinarily be expected.
Manufacturers of finished food can also take numerous steps to reduce the
potential regulatory and product liability. The labeling of products with
ingredient statements that identify allergens is the best way to minimize
product liability. Although there are no magic formulas that will guarantee
immunity from liability for an allergic reaction or sensitivity to food, there
are numerous steps that can be taken to minimize potential liability.

Additional information on the regulatory and legal developments on food
allergens can be found on FDA’s website at http://vm.cfsan.fda.gov/list.html.
20.7 References
1. 21 United States Code (U.S.C.) § 331(a); FFDCA 301(a).
2. 21 U.S.C. § 343(i)(1); FFDCA § 403(i)(1).
3. 21 U.S.C. § 343(i)(1); FFDCA § 403(i)(1).
4. 21 U.S.C. § 101.100(a)(3).
5. 21 C.F.R. § 101.100(a)(3).
6. Food Allergen Labeling and Consumer Protection Act, Pub.L. No. 108–282, 118
Stat. 905 (August 2, 2004).
7. 21 U.S.C. § 321(gg); FFDCA § 201(gg).
8. 21 U.S.C. § 321(gg)(2)(A); FFDCA § 201(gg)(2)(A).
9. 21 U.S.C. § 321(gg)(2)(B); FFDCA § 201(gg)(2)B).

10. 21 U.S.C. § 343(w)(7); FFDCA § 403(w)(7).
11. Monsanto Company v. Kennedy, 613 F.2d 947 (D.C. Cir. 1979) (the Commissioner
may determine based on the evidence before him that the level of migration into
food of a particular chemical is so negligible as to present no public health or safety
concerns, even to assure a wide margin of safety).
12. 21 U.S.C. § 343(w)(6); FFDCA § 403(w)(6).
13. 21 U.S.C. § 343(w)(1); FFDCA § 403(w)(1).
14. Statement of Policy for Labeling and Preventing Cross-contact of Common Food
Allergens, Compliance Policy Guide 555.250 (April 19, 2001) (copy accessed in
October 2004 at
http://www.fda.gov/ora/compliance_ref/cpg/cpgfod/cpg555-250.htm).
October 2004 at
16. Questions and Answers on Allergen Guide, U.S. Food and Drug Administration,
Center for Food Safety and Applied Nutrition (May 3, 2001) (accessed October 2004
at http://www.cfsan.fda.gov/~dms/alrgpgtp.html).
17. Notice to Manufacturers on Label Declaration of Allergenic Substances in Food,
from Fred R. Shank, Ph.D., Director, Center for Food Safety and Applied Nutrition,
U.S. Food and Drug Administration (June 10, 1996) (copy accessed in October 2004
at http://www.cfsan.fda.gov/~lrd/allerg7.html).
18. 21 U.S.C. § 331(a); FFDCA § 301(a).
19. 21. U.S.C. § 342(a)(4); FFDCA § 402(a)(4).
October 2004 at
October 2004 at
22. Questions and Answers on Allergen Guide, U.S. Food and Drug Administration,
Center for Food Safety and Applied Nutrition (May 3, 2001) (accessed in October
2004 at http://www.cfsan.fda.gov/~dms/alrgpgtp.html).
23. 21 U.S.C. § 332 (injunctions), § 334 (seizures); FFDCA §§ 302, 304.
24. 21 U.S.C. § 334(h); FFDCA § 304(h).
25. 21 C.F.R. § 7.3(g).
26. 21 C.F.R. § 7.3(g).
27. Larsen v. General Motors Corp., 391 F.2d 495 (8th Cir. 1968).
28. See, e.g., Greenman v. Yuba Power Products, Inc. 377 P.2d 897 (Cal. 1962).
29. Comment j to Section 402 of 2 Restatement of the Law 2d, Tort (1965).
30. Brown et. al. v. McDonald’s Corporation et. al.,101 Ohio App.3d 294, 655 N.E. 440
(OH Court of Appeals 1995); see also, Livingston v. Marie Callender’s, Inc. 72 Cal.
App. 4th 830, 85 Cal.Rptr.2d 528 (Cal. Court of Appeals, 1999).
31. McDonald’s, 655 N.E. at 443.
32. McDonald’s 655 N.E. at 444 (emphasis original).
33. McDonald’s 655 N.E. at 444 (emphasis original).
Appendix I: Summary of cases involving known allergens
Style of case Damages Comments
Thompson vs. -0- Plaintiff went into anaphylactic shock, suffered

East Pacific Case respiratory arrest and heart attack. Dish did not
Enterprises, dismissed contain peanuts; trace amounts of peanut
Inc. et al.1 product were present in food as result of cross-
contamination. Plaintiff did not inform
restaurant of her allergy to peanut products.
Abbhi vs. Not reported Nine year old child died from anaphylactic
AMI et al.2 shock after eating pastry.
Knight vs. Just Not reported Defendants asserted idiosyncratic allergic
Born, Inc. et al.3 response defense.
Mitchell vs. $10 000 Allergic reaction caused anaphylaxis requiring
Eddie’s Place on epinephrine shot, resulting in scarring at
Chagrin, Inc.4 injection site on leg and emotional distress.
Marshall $57 137 Plaintiff regularly bought cookies at store, was
Randall vs. Trans told no nuts in cookie. Store substituted new
Pacific Stores, cookie with nuts; had told employee to change
Ltd.5 label but employee had not done so. Allergic
reaction resulted in anaphylactic reaction and
angioedema with respiratory failure and
acidosis; was unconscious and close to death
upon arrival in emergency room and admitted to
intensive care unit.
Vosko vs. Wing $8096 Restaurant mistakenly put shrimp, instead of
Hong, Inc.6 chicken, in egg foo yung. After eating meal,
plaintiff took Benadryl and went to ER with
symptoms of anaphylactic shock. Was given
adrenalin, observed for two hours and released.
Dobeus vs. $500 Allergic reaction resulted in asthmatic episodes
The Chef’s Place, and hospitalization.
Inc.7
Plaintiff vs. $1 000 000 Allergic reaction resulted in stroke with
Trattoria permanent brain lesion and inability to return
Spago8 to work. Had informed waitress of allergy.
Restaurant denied that dessert contained nuts.
Bocon vs. -0- Plaintiff informed waitress of dairy allergy;
Pizzeria Verdict for after starting to eat, plaintiff developed nausea,
Uno, Inc.9 defense diarrhea and vomiting, then went to emergency
room with irritable bowel syndrome (IBS).
Plaintiff had pre-existing IBS.
McGowan vs. $3500 Plaintiff suffered allergic reaction and
Public House, emotional distress; was told cookie contained
Inc.10 no nuts.
Krueger vs. $5000 Allergic reaction caused pain and swelling in
Camaraderie face and increased heart rate. Plaintiff alleged
Restaurant, that waitress stated that dessert was nut-free.
Inc.11
Appendix I: Continued

Castiello vs. WLH $172 500 Anaphylactic attack resulting in permanent
Management asthmatic condition.
Corp.12
Balison vs. JB’s $12 287 Allergic reaction resulting in anaphylactic
Restaurant, shock. Disputed whether plaintiff asked about
Inc.13 cooking methods or defendant made
representations regarding same.
Ficken vs. Edy’s $2654 Allergic reaction resulted in sleeplessness,
Grand Ice vomiting, and headaches.
Cream, Inc.14
MacLeod vs. $350 Student suffered psychological trauma
Univ. following allergic reaction to mayonnaise on
of Mass. sandwich. $500 award reduced by $150 for
Health contributory negligence for not checking food
Services15 before eating.
Yost vs. -0- Allergic reaction resulted in permanent
Chapter Verdict for asthmatic condition.
Eleven defense
Restaurant16
O’Brien vs. $100 000 Plaintiff suffered acute respiratory arrest,
Moishs anaphylactic shock and temporary coma.
Addison Fiancée had been told by sales clerk that there
Bakery17 were walnuts, but not peanuts, in brownies.
1
Unpublished, 2003 WL 352914, Washington Ct of Appeals No. 49924-6-I (Feb. 18, 2003). (Appeal
from trial court grant of summary judgment for defendants on all counts; decision affirmed.)
2
1997 WL 325580, Conn. Superior Court, No. CV-96032195-S (June 3, 1997). (Motion to strike
non-products liability claims granted.)
3
2000 WL 924624, US District Court, Oregon (Mar. 28, 2000). (Cross motions for summary judgment.
Summary judgment denied as to manufacturing defect and negligence claims. Summary judgment
granted as to failure to warn claim because inconsistent with manufacturing defect claim.)
4
Settlement, Cuyahoga County, Ohio, Court of Common Pleas, No. 396582 (Oct. 17, 2000).
5
Unpublished Jury verdict, San Francisco County, Cal., No. 995772 (Sept. 10, 1999).
6
Unpublished, Jury verdict, Oakland County, Michigan Circuit Court.
7
Unpublished Settlement, Lake County, Illinois Circuit Court, No. 96-AR-2078 (Nov. 1997).
8
Unpublished Settlement, LA County Superior Court, No. BC-105-476 (April 1996).
9
Unpublished Jury verdict, Norfolk County, Mass. Superior Court, No. 94-494 (Sept. 25, 1995).
10
Unpublished Settlement, Worcester County, Mass. Superior Court, No. 93-0252-A (Feb. 1995).
11
Unpublished Settlement, Eau Claire County, Wis. Circuit Court, No. N/A (Feb. 1994).
12
Unpublished Settlement, York County, Maine Superior Court, No. CV-92-610 (Dec. 1993).
13
Unpublished Settlement, Kootenai County, Idaho District Court, No. 88757 (March 1993).
14
Unpublished Verdict, Jackson County, Mo. Circuit Court, No. CV-90-22738 (Sept. 1992).
15
Unpublished Verdict, Hampshire County, Mass. Superior Court, No. 87-173 (Feb. 1990).
16
Unpublished Verdict, Spokane County, Wash. Superior Court, No. 85-2-01946-1 (Oct. 1988).
17
Unpublished Settlement, Philadelphia County, Penn., No. 1996-12-918 (1999).
Appendix II: Summary of cases involving ingredients that are not allergens
Livingston vs. Not reported Plaintiff suffered MSG Symptom Complex

Marie including respiratory arrest, hypoxia, cardiac
Callendar’s arrest and brain damage after eating soup
Inc.1 containing MSG. Waitress had told plaintiff
that soup did not contain MSG.
Brown vs. Not reported Carageenan is seaweed-derived product to which
McDonald’s people allergic to seafood may have an allergic
Corp.2 reaction. After eating hamburger plaintiff
developed rash, tight chest blue lips and hives,
required five-hour hospital stay.
Trapnell vs. Not reported Chronic asthmatic died after eating food
Sysco containing sulfites. Decedent had been told that
Food Services, no sulfites had been used in the food.
Inc. et al.3
McPike vs. -0- Ten year old asthmatic girl died after allergic
Enciso’s Cocina Verdict for reaction to sulfite in guacamole.
Mejicana, defense
Inc.4
Caravia vs. $1 500 000 Female died after asthma attack caused by
Sbarros sulfites in salad. Jury found defendant 60%
Licensing liable and decedent 40% negligent.
Inc.5
1
85 Cal.Rptr.2d 528, Cal. Court of Appeal, 2d District, No. B-115078, June 3, 1999. (Appeal from
trial court dismissal of strict liability claim; Reversed and remanded.)
2
655 N.E.2d 440, Ohio App. 1995, No. 94-CA-005904. (Appeal from trial court grant of summary
judgment for defendants; affirmed in part and reversed in part; Case remanded to trial court.)
3
850 S.W.2d 529, Tex. App. 1992, 890 S.W.2d 796, Tex. 1994. (Appeal from trial court grant of
summary judgment. Reversed by Court of Appeals; reversal affirmed by Texas Supreme Court. Case
remanded.)
4
762 P.2d 315, Oregon Ct. Appeals 1988. (Appeal from jury verdict for defendants; affirmed.)
5
Unpublished Jury verdict, Broward County, Fla. Circuit Court, No. 86-31404 (Sept. 1990).
21
EU regulation of undeclared allergens in

food products
H. Heeres, TNO Quality of Life, The Netherlands
21.1 Introduction
The European Union promotes the free movement of goods between Member
States through the EC Treaty, which prohibits quantitative restrictions on
imports between them (Article 28). This principle of free movement was
affirmed in the Cassis de Dijon case.1
In the late 1990s, the BSE crisis led to major changes in the regulatory
regime for foodstuffs. The White Paper on Food Safety, presented by the
Commission, stated that ‘assuring that the EU has the highest standards of
food safety is a key policy priority for the Commission,’ and that ‘a radical
new approach is proposed.’ The White Paper recommended the establishment
of an independent European Food Authority to address the need to guarantee
a high level of food safety. A range of other measures aimed at improving
and standardising legislation covering all aspects of food products, from
‘farm to table’, is needed to support this recommendation.
The supporting measures proposed by the Commission are presented in
the Annex, with an indication of priority and likely timing. The ‘Action Plan
on Food Safety’ (Action Number 3) proposes a General Food Law Directive.
The objectives are to establish food safety as the primary objective of EU
food law and to lay down the common principles underlying food legislation,
in particular the scientific basis for legislation, responsibility of producers
and suppliers, traceability along the food chain, efficient controls and effective
enforcement. Action Number 16 concerns a proposal to amend the existing
directive on the labelling, presentation and advertising of foodstuffs. The
objectives of this action are to ensure that the components of compound
ingredients forming less than 25% of the final product are indicated, and to
EU regulation of undeclared allergens in food products 379
lay down a list of allergenic substances. The objectives of Action Number 3

were laid down in 2002 in Regulation 178/2002/EC,2 and those of Number
16 in 2000, in Directive 2000/13/EC.3 Recently, this directive has been
amended by Directive 2003/89/EC,4 which introduces the obligation to indicate
a number of allergen ingredients on the label.
Regulation 258/97/EEC5 concerns the placement of novel foods or novel
food ingredients on the market within the Community. Foods and ingredients
are novel if they have not hitherto been used for human consumption to a
significant degree within the Community. Those responsible for placing novel
foods on the market must submit a request to the Member State where the
product is to be introduced.
The European Commission also has a number of legal instruments concerned
with compensation for victims of defective products, including the Product
Liability Directive, Directive 85/374/EEC,6 and the General Product Safety
Directive, Directive 92/59/EEC.7 The Product Liability Directive was amended
by Directive 1999/34/EC8 to include primary agricultural products. The General
Product Safety Directive was repealed by Directive 2001/95/EC.9 On the
basis of these instruments, producers are obliged to place only safe (food)
products on the market, and are liable for damage caused by any defective
(food) products.
21.2 Food legislation concerning the labelling of

ingredients
Directive 2000/13/EC10 addresses the labelling, presentation and advertising
of foodstuffs, and has superseded the older Directive, 79/112/EEC.11 The
major elements of the new directive are:
• ‘the need to inform and to protect the consumer’;
• ‘to enact Community rules of a general nature applicable horizontally to
all foodstuffs put on the market’; and
• to promote the use of ‘detailed labelling, in particular giving the exact
nature and characteristics of the product, which enables the consumer to
make his choice in full knowledge of the facts’.
The most important requirements of this Directive are discussed below.
21.2.1 Definitions
Labelling shall mean any words, particulars, trade marks, brand name, pictorial
matter or symbol relating to a foodstuff and placed on any packaging, document,
notice, label, ring or collar accompanying or referring to such foodstuff. Pre-
packaged foodstuffs are those where the contents cannot be altered without
opening or changing the packaging.
21.2.2 Misleading information

One of the most important requirements is that the labelling and methods
used must not mislead the purchaser. Foodstuffs should not be labelled,
presented or displayed in a way that suggests they possess properties or
special characteristics that are either not present, or are in fact typical for all
similar foodstuffs.
21.2.3 Information required on the label

The following particulars are compulsory on the labels of foodstuffs:
(1) the name under which the product is sold;
(2) the list of ingredients;
(3) the quantity of certain ingredients or categories of ingredients;
(4) in the case of pre-packaged foodstuffs, the net quantity;
(5) the date of minimum durability or, in the case of foodstuffs that are
highly perishable (microbiologically), the ‘use by’ date;
(6) any special storage conditions or conditions of use;
(7) the name or business name and address of the manufacturer or packager,
or of a seller established within the Community;
(8) particulars of the place of origin or provenance where failure to give
such particulars might mislead the consumer to a material degree as to
the true origin or provenance of the foodstuff;
(9) instructions for use when it would be impossible to make appropriate
use of the foodstuff in the absence of such instructions; and
(10) the actual alcoholic strength by volume for beverages containing more
than 1.2% by volume of alcohol.
21.2.4 Name
The name under which a foodstuff is sold shall be the name provided for in
the Community provisions applicable to it or, where there are no Community
provisions, the name provided for in the laws, regulations and administrative
provisions applicable in the Member State in which the product is sold.
Failing this, the name or description of the foodstuff and, if necessary, its use
must be clear enough to let the purchaser know its true nature and distinguish
it from other products.
21.2.5 List of ingredients

An ingredient is any substance, including additives, used in the manufacture
or preparation of a foodstuff and still present in the finished product, even if
in an altered form. If an ingredient of the foodstuff is itself the product of
several ingredients, the latter shall be regarded as ingredients of the foodstuff
in question. Additives that serve no technological function in the finished
product are not ingredients, nor are solvents, media or other processing aids.
In general, the list of ingredients shall include all the ingredients of the
foodstuff, in descending order of weight, as recorded at the time of their use
in the manufacture of the foodstuff. Exceptions include added water and
volatile products, ingredients used in concentrated or dehydrated form,
concentrated or dehydrated foods, mixtures of spices or herbs, and ingredients
constituting less than 2% of the finished product (may be listed in a different
order after the other ingredients). Ingredients may be listed by their specific
name, or by category name where applicable (see Annex I of the Directive).
Additives that belong to one of the categories listed in Annex II of the
Directive must be designated by the name of that category, followed by their
specific name or EC number. Flavourings shall be designated in accordance
with the rules laid down in Annex III of the Directive. There are some cases
where ingredients do not need to be listed, for instance single-ingredient
foodstuffs such as fresh fruit and vegetables, carbonated water, fermentation
vinegars, cheese, butter, fermented milk and cream.
21.2.6 Quantitative ingredients declaration (QUID) regulation

The QUID regulation specifies the circumstances under which the quantity
of an ingredient or category of ingredients used in the manufacture or
preparation of a foodstuff must be stated. The quantity indicated, expressed
as a percentage, must correspond to the quantity of the ingredient or ingredients
at the time of its/their use, and must appear either in or immediately next to
the name under which the foodstuff is sold, or in the list of ingredients in
connection with the ingredient or category of ingredients in question.
21.2.7 Net quantity

The net quantity of pre-packaged foodstuffs must be expressed in units of
volume in the case of liquids and in units of mass in the case of other
products, using the litre, centilitre, millilitre, kilogram or gram, as appropriate.
Additional regulations concern other aspects of labelling net quantity, such
as the weight of pre-packaged items that are sold in combination, foodstuffs
that are normally sold by number, and use of drained net weight for foodstuffs
normally sold in a liquid medium.
21.2.8 Date of minimum durability

The date of minimum durability of a foodstuff shall be the date until which
the foodstuff retains its specific properties when properly stored. The date
shall be preceded by the words: ‘Best before …’ when the date includes an
indication of the day, and ‘Best before end …’ in other cases. Either the date
itself or a reference to where the date is given on the labelling shall accompany
these words. If need be, these particulars shall be followed by a description
of the storage conditions which must be observed if the product is to keep for
the specified period. The date of minimum durability does not apply to all
foodstuffs. In the case of highly perishable foodstuffs, which are likely to
constitute an immediate danger to human health after a short period, the date
of minimum durability is replaced by the ‘use by’ date.
21.2.9 Instructions for use

Instructions for use of a foodstuff shall be indicated so that it is used
appropriately.
21.2.10 Foodstuffs without pre-packaging

The Member State itself shall adopt detailed rules concerning the labelling
of foodstuffs offered for sale to the ultimate consumer or to mass caterers
without pre-packaging, or where foodstuffs are packaged on the sales premises
at the consumer’s request or pre-packaged for direct sale.
21.2.11 Language
The information on the label should be written in a language easily understood
by the consumer, unless the consumer is informed by other means. Member
States may specify the language(s) used to label products marketed within
their own territory, in accordance with the rules of the Treaty.
21.3 Food legislation concerning the labelling of

allergens
Directive 2003/89/EC12 amends Directive 2000/13/EC13 concerning the
information requirements for foodstuffs, and is designed to achieve a high
level of health protection for consumers and to guarantee their right to
information through accurate labelling of ingredients. The amendment includes
the requirement to label all the ingredients, additives, processing aids and
other substances that may cause allergic or intolerance reactions in consumers.
21.3.1 Considerations
Allergic or intolerance reactions constitute a danger to the health of the
consumer. The Scientific Committee on Food has stated that food allergies
affect the lives of many people, and reactions range from very mild to potentially
fatal. Common food allergens include cows’ milk, fruits, legumes, especially
peanuts and soybeans, eggs, crustaceans, tree nuts, fish, vegetables (celery
and other foods of the Umbelliferae family), wheat and other cereals. The
most common food allergens are found in a wide variety of processed foods,
so labelling rules are necessary to ensure that consumers suffering from food
allergies receive appropriate information.
A list of allergenic substances has been drafted, and includes those foodstuffs,
ingredients and other substances recognised as causing hypersensitivity. The
list of allergens may be revised if new allergens are identified, or if technological
developments permit the removal of allergenicity in ingredients and other
substances, in order to avoid unnecessary obligations on labelling.
21.3.2 List of allergenic ingredients

Annex IIIa shows a list of allergenic substances that must be indicated in the
list of ingredients, and includes:
• cereals containing gluten (i.e. wheat, rye, barley, oats, spelt, kamut or
their hybridised strains) and products thereof;
• crustaceans and products thereof;
• eggs and products thereof;
• fish and products thereof;
• peanuts and products thereof;
• soybeans and products thereof;
• milk and products thereof (including lactose);
• nuts, i.e. almond (Amygdalus communis L.), hazelnut (Corylus avellana),
walnut (Juglans regia), cashew (Anacardium occidentale), pecan nut (Carya
illinoesnis (Wangenh.) K. Koch), Brazil nut (Bertholletia excelsa), pistachio
nut (Pistacia vera), macadamia nut and Queensland nut (Macadamia
ternifolia) and products thereof;
• celery and products thereof;
• mustard and products thereof;
• sesame seeds and products thereof; and
• sulphur dioxide and sulphites at concentrations of more than 10 mg/kg or
10 mg/litre expressed as SO2.
The words ‘allergenic ingredient(s)’ and ‘allergenic substance(s)’mean an
allergenic substance(s) named in Annex IIIa, and include the substance itself
and also the products thereof. There are other foodstuffs that can cause
adverse reactions, but on the basis of this Directive, the presence of such an
ingredient or product thereof in a foodstuff does not have to be indicated on
the label. However, the Directive does indicate that it is important to be able
to revise the list of ingredients rapidly, when necessary by including or
deleting certain ingredients or substances. Recently the Directive 2005/26/
EC14 has established a list of food ingredients or substances which are
provisionally excluded from Annex IIIa of Directive 2000/13/EC (see Table
21.1). These ingredients or substances will be excluded from this annex until
25 November 2007 and so do not have to be indicated on the label.
Table 21.1 List of food ingredients and substances provisionally excluded from Annex
IIIa of Directive 2000/13/EC
Ingredients Products thereof provisionally excluded
Cereals Wheat-based glucose syrups including dextrose1

containing Wheat-based maltodextrins1
gluten Glucose syrups based on barley
Cereals used in distillates for spirits
Eggs Lysozyme (produced from egg) used in wine
Albumin (produced from egg) used as fining agent in wine and cider
Fish Fish gelatine used as carrier for vitamins and flavours
Fish gelatine or isinglass used as fining agent in beer, cider and wine
Soybean Fully refined soybean oil and fat (1)
Natural mixed tocopherols (E306), natural D-alpha tocopherol,
natural D-alpha tocopherol acetate, natural D-alpha tocopherol
succinate from soybean sources
Vegetable oils derived phytosterols and phytosterol esters from
soybean sources
Plant stanol ester produced from vegetable oil sterols from soybean
sources
Milk Whey used in distillates for spirits
Lactitol
Milk (casein) products used as fining agents in cider and wines
Nuts Nuts used in distillates for spirits
Nuts (almonds, walnuts) used (as flavour) in spirits
Celery Celery leaf and seed oil
Celery seed oleoresin
Mustard Mustard oil
Mustard seed oil
Mustard seed oleoresin
1
And products thereof, in so far as the process that they have undergone is not likely to increase the
level of allergenicity assessed by the European Food Safety Authority for the relevant product from
which they originated
21.3.3 Labelling of allergenic ingredients of beverages containing

more than 1.2% by volume of alcohol
Any ingredient listed in Annex IIIa shall be indicated on the labelling where
it is present in the beverage. This indication shall comprise the word ‘contains’
followed by the name of the ingredient(s) concerned. However, an indication
is not necessary when the ingredient is already included under its specific name
in the list of ingredients or in the name under which the beverage is sold. When
necessary the EU can adopt detailed rules for the presentation of the indication.
21.3.4 Labelling of allergenic ingredients in other foodstuffs

Any ingredient listed in Annex IIIa or originating from an ingredient listed
in Annex IIIa that is used in production of a foodstuff and still present in the
finished product, even in altered form, must be indicated on the label with a
clear reference to the name of the ingredient. This means that any allergenic
ingredient used in production and present in the foodstuff has to be indicated
on the label even if a list of ingredients is not needed. This requirement also
applies to the ingredients of a compound ingredient, food additives, a component
of a food additive, flavourings, a component of a flavouring, processing aids,
and carriers, solvents or media for additives, flavours or processing aids.
Indication is not required if the name under which the foodstuff is sold
clearly refers to the ingredient concerned.
21.3.5 Method of labelling an allergenic ingredient

The directive states that ‘the ingredient shall be indicated on the label with
a clear reference to the name of this ingredient,’ and that ‘the indication shall
not be required if the name under which the foodstuff is sold clearly refers
to the ingredient concerned.’ This open formulation for labelling allows
some leeway in deciding whether an indication is needed, the text of the
indication itself and positioning on the label. In practice, it can lead to
differences regarding the labelling of allergenic substances, and it is crucial
that the considerations described above are taken into account when drawing
up the indication. It is also important to assess the knowledge of consumers,
because, for example, there are consumers who do not know that yoghurt is
made of (cows’) milk.
21.3.6 Indication and labelling of an allergenic ingredient

If a list of ingredients is not needed for a particular foodstuff, the allergenic
ingredient should be labelled as ‘Contains’ followed by the name of the
allergenic substance. If a list of ingredients is required, then the presence of
allergenic ingredient(s) can be stated in the list of ingredients. The allergenic
substance should be stated after the ingredient concerned, preceded by the
word ‘contains’, and followed by the name of the allergenic substance, e.g.
Ingredients: yoghurt, contains milk, and so on. A second possibility is to
indicate the presence of an allergen in a sentence under the list of ingredients,
e.g. ‘This foodstuff contains the allergenic ingredients’ followed by the name
of the allergenic substance, or ‘Contains’ followed by the name of the allergenic
substance. In these cases, it is advisable that the allergenic substance(s) in
the list of ingredients and in the sentence under the list are joined, e.g. by a
character or figure between brackets in front of or after the allergenic ingredient
concerned.
This Directive has to be implemented in accordance with the associated
national laws of each Member State, and, under the Directive, it is possible
for a Member State to define another system of indication and/or wording
for labelling allergenic substances in foodstuffs. For example the Netherlands
describes coercively the labelling of allergenic ingredient(s) in the list.
21.3.7 Guidance documents

Since publication of the Directive, organisations in the food industry have
set up guidelines for the practical application of the new labelling rules with
respect to ingredients and allergens. These guidelines do not have legal
status.
In the Netherlands, the ‘Centraal Bureau Levensmiddelenhandel’ and
‘Federatie Nederlandse Levensmiddelen Industrie (FNLI) have drawn up
‘Wijziging van de etiketteringvoorschriften ingevolge de publicatie van
RICHTLIJN 2003/89/EG VAN HET EUROPEES PARLEMENT EN DE
RAAD van 10 November 2003 tot wijziging van Richtlijn 2000/13/EG met
betrekking tot de vermelding van de ingrediënten van levensmiddelen; Een
Handleiding’.
The Confederation of the Food and Drink Industries of the EU (CIAA)
developed the ‘Final Guidance Document on the practical application of
Directive 2003/89/EC on the ingredient and allergen labelling’.
21.3.8 Limits of detection, standards

The directive does not include any limit of detection or standard. If an
allergenic ingredient has been used in the preparation of a foodstuff, it
should be labelled. If it is not used but is present, then it does not have to be
labelled.
21.4 Legislation concerning general product safety

(Directive 2001/95/EC)
Directive 2001/95/EC15 superseded Directive 92/59/EEC16 pertaining to the
safety of products and ensuring that products placed on the market are safe.
The most important requirements of this Directive are discussed below.
Where Member States have different legislation on product safety, this not
only affects the levels of protection provided to consumers, but also tends to
create barriers to trade and distortion of competition within the European
Community market. The Community aims to protect the health and safety of
consumers, and Community-wide legislation introducing general product
safety requirements and containing provisions on the general obligations of
producers and distributors should contribute to that aim. Because it is very
difficult to adopt Community-wide legislation for every product, there is a
need for a broad-based, legislative framework that establishes general safety
requirements for any product.
21.4.2 Product definitions

A product is defined as any product, including the provision of a service,
which is intended for consumers or likely, under reasonably foreseeable
conditions, to be used by consumers even if not intended for them, and is
supplied or made available, whether for consideration or not, in the course of
a commercial activity, and whether new, used or reconditioned.
A ‘safe product’ is defined as any product which, under normal or reasonably
foreseeable conditions of use including duration and, where applicable, putting
into service, installation and maintenance requirements, does not present any
risk or only the minimum risks compatible with the product’s use, considered
to be acceptable and consistent with a high level of protection for the safety
and health of persons.
Important aspects of this definition for foodstuffs are:
• the characteristics of the product, including its composition, packaging,
instructions for assembly and, where applicable, for installation and
maintenance;
• the presentation of the product, the labelling, any warnings and instructions
for its use and disposal and any other indication or information regarding
the product; and
• the categories of consumers at risk when using the product, in particular
children and the elderly.
A ‘dangerous product’ means any product, which does not meet the definition
of ‘safe product’. The feasibility of obtaining higher levels of safety or the
availability of other products presenting a lesser degree of risk shall not
constitute grounds for considering a product to be ‘dangerous’.
21.4.3 General safety requirement, conformity assessment criteria

and European standards
Producers shall be obliged to place only safe products on the market. A
product shall be deemed safe, when, in the absence of specific Community
provisions governing the safety of the product in question, it conforms to the
specific rules under national law of the Member State in whose territory the
product is marketed. A product shall be presumed safe as far as the risks and
risk categories covered by relevant national standards are concerned when it
conforms to voluntary national standards in place of European standards.
21.4.4 Other obligations of producers and distributors

Within the limits of their respective activities, producers shall provide consumers
with information that enables them to assess the risks inherent in a product
throughout the normal or reasonably foreseeable period of its use, and
precautions against those risks. This obligation applies if such risks are not
immediately obvious without adequate warnings.
21.4.5 Product recall

A recall means any measure aimed at achieving the return of a dangerous
product that has already been supplied or made available to consumers by
the producer or distributor. The producer is obliged to monitor the products
made, and producers and distributors are obliged to recall products from the
market, and consumers, if they present risks. Member States shall ensure
that producers and distributors comply with their obligations under the directive
in such a way that products placed on the market are safe. Governments can
oblige producers and distributors to carry out a product recall.
21.4.6 Guide to corrective action including recalls

Intertek Research and Testing Centre published ‘Product safety in Europe, A
guide to corrective action including recalls’ on behalf of the UK Consumers
Association. This is a voluntary guide to carrying out corrective actions
regarding product safety, and covers all types of corrective action aimed at
removing a safety risk arising from a non-food product placed on the market.
It also gives information for corrective actions concerning foodstuffs if they
pose a risk. The text is available on the Internet (see Section 21.9).
21.5 Legislation concerning food safety (Regulation (EC)

No. 2002/178/EC)17
This regulation provides the basis for assuring a high level of protection of
human health and consumer interests in relation to food. The regulation lays
down the general principles and requirements of food law, establishes the
European Food Safety Authority and defines procedures in matters of food
safety.
The free movement of safe and wholesome food can be achieved only if food
and feed safety requirements do not differ significantly from Member State
to Member State. Experience has shown that measures are needed to guarantee
that unsafe food is not placed on the market, and that systems exist to
identify and respond to food safety problems. To ensure the safety of food,
all aspects of the food production chain must be considered, from and including
primary production and the production of animal feed, up to and including
sale or supply of food to the consumer, because each element may have a
potential impact on food safety. Community policies aim to ensure a high
level of protection of human life and health, and Community food policies
form a common basis for measures governing food and feed taken by individual
Member States and at Community level.
21.5.2 Definition of food

Food or foodstuff means any substance or product, whether processed, partially
processed or unprocessed, intended to be, or reasonably expected to be,
ingested by humans. Food includes drink, chewing gum and any substance,
intentionally incorporated into the food during its manufacture, preparation
or treatment. It includes water. Food shall not include feed, live animals
unless they are prepared for placing on the market for human consumption,
plants prior to harvesting, medicinal products, cosmetics, tobacco and tobacco
products, narcotic or psychotropic substances, residues and contaminants.
21.5.3 Food safety requirements

Unsafe foodstuffs must not be placed on the market. Foodstuffs are deemed
unsafe if they are considered to be injurious to health or unfit for human
consumption. In determining whether any food is unsafe, normal conditions
of use of the food by the consumer and at each stage of production, processing
and distribution, and all information provided to the consumer should be
taken into account. In determining whether any food is injurious to health,
the probable immediate and/or short-term and/or long-term effects of that
food on the health of a person consuming it (and subsequent generations),
the probable cumulative toxic effects, and the particular health sensitivities
of a specific category of consumers should be taken into account.
21.5.4 Responsibilities of food and feed business operators

Food and feed business operators at all stages of production, processing and
distribution within the businesses under their control shall ensure that foods
or feeds satisfy the requirements of food law that are relevant to their activities
and shall verify that such requirements are met.
21.5.5 Traceability
Food and any other substance intended to be, or expected to be, incorporated
into a food shall be traceable through all stages of production, processing
and distribution. Traceability is of great importance for product liability. In
the case a defective product, the plaintiff can make each producer in the
chain liable, from raw material to end-product. Producers therefore need to
have a traceability system in place that enables them to check whether the
cause of a defective product lies in their own factory, is due to a defective
raw material, or arose later with a user in the chain. In the latter two cases the
producer can take recourse against the supplier or user. Food which is placed
on the market or is likely to be placed on the market in the Community shall
be adequately labelled or identified to facilitate its traceability, through relevant
documentation or information.
21.5.6 Responsibilities of food business operators for

non-compliant foodstuffs
If a food business operator has reason to believe that a food which they have
imported, produced, processed, manufactured or distributed is not in compliance
with food safety requirements, then they must immediately initiate procedures
to withdraw the food, whether from other food business operators or from
the consumers themselves. Food business operators are also obliged to inform
and collaborate with the relevant authorities when a food has been recalled,
and provide reasons for withdrawal to consumers, where applicable. Food
business operators that are not producers also have obligations in the withdrawal
of unsafe foods, including passing on information necessary to trace a food,
and co-operating in the actions taken by producers, processors, manufacturers
and/or the competent authorities. Food business operators shall not prevent
or discourage any person from co-operating, in accordance with national law
and legal practice, with the competent authorities, where this may prevent,
reduce or eliminate a risk arising from a food.
21.5.7 Liability
The provisions stated in Chapter 2 of the Regulation (EC) No 2002/178/EC
concerning General Requirements of Food Law, Food Safety Requirements,
Responsibilities and Traceability, shall be without prejudice to Directive 85/
374/EEC18 concerning liability for defective products. This means that the
rules concerning Food Safety Requirements, Responsibilities and Traceability,
and those concerning defective products, exist side by side.
21.6 Legislation concerning product liability

The liability for defective products in the European Community is laid down
in Directive 85/374/EEC19 and is based on strict liability for the producer,
‘liability without fault’. Each Member State also has national rules concerning
liability, which are generally based on fault and can be used if the case meets
conditions laid down in the national rules.
Different laws among Member States concerning the liability of the producer
for damage caused by a defective product may distort competition and affect
the free movement of goods. National laws also result in differing levels of
consumer protection against damage to health or property caused by a defective
product. Standardisation of laws across the Community is not and cannot be
total, but this Directive lays down the principle of strict liability for the
producer. Even in the absence of fault, the producer may be held responsible for
damage caused by a defective product that the producer has put on the market.
21.6.2 Optional possibilities for Member States

The Directive contains three optional possibilities for Member States. Firstly,
that in some cases a producer shall not be liable if he proves that the state of
scientific and technical knowledge at the time when the product was put into
circulation was such that the existence of the defect could not be discovered.
Secondly, that a Member State might limit total liability for damage, so that
damage resulting from death or personal injury and caused by identical
items with the same defect can be limited to an amount not less than 70
million ECU (Euro 70 million). Thirdly, that a Member State might legislate
that the term ‘product’ also includes agricultural products and game. The
Member States Greece, France, Luxembourg, Finland and Sweden used this
option, but it has now been abrogated by the implementation of Directive
1999/34/EC,20 which was aimed at restoring consumer confidence in the
safety of agricultural products following the BSE crisis.
21.6.3 Requirements for liability

Article 1 lays down the requirements for a producer to be liable for a defective
product. The four requirements are:
• the person claimed is the producer;
• there is damage;
• the product is defective; and
• there is a causal relationship between the defect and the damage.
The four requirements, opportunities for defence, and the limitation or expiry
period are discussed below.
21.6.4 Producer
On the basis of Article 3, ‘Producer’ means the manufacturer of a finished
product, the producer of any raw material or the manufacturer of a component
part and any person who, by putting his name, trademark or other distinguishing
feature on the product presents himself as its producer. In addition, any
person who imports a product into the Community for sale, hire, lease or any
form of distribution in the course of his business shall be deemed to be a
producer within the meaning of the Directive and shall be responsible as a
producer. The producers are liable jointly and severally, and the victim of
damage may make a claim against any of these persons for compensation,
whether they are the actual producer or not. If the producer of the product
cannot be identified, each supplier of the product shall be treated as its
producer unless he informs the injured person, within a reasonable time, of
the identity of the producer or of the person who supplied him with the
product.
21.6.5 Damage
The scope of the damage to be compensated is restricted according to Article
9, to include damage caused by death or by personal injuries, and damage to,
or destruction of, any item of property other than the defective product itself,
with a lower threshold of 500 ECU (Euro 500). The item of property should
be of a type ordinarily intended for private use or consumption. This directive
does not cover damage to professional property, where national laws concerning
liability are applicable. National laws are also needed for damage that is not
covered by the scope of product liability, such as the lower threshold of Euro
500. As described above, damage resulting from death or personal injury and
caused by identical items with the same defect can be limited to an amount
not less than 70 million ECU (Euro 70 million).
21.6.6 Defective product

A product is defined in Article 2 and described in Section 21.4.2. For the
purpose of this Directive, ‘product’ also means all movables even if incorporated
into another movable or into an immovable, and foodstuffs are products in
the sentence of this directive.
Article 6 shows the requirements for a defective product. A product is
defective when it does not provide the safety which a person is entitled to
expect, taking all circumstances into account, including:
• the presentation of the product;
• the use to which it could reasonably be expected that the product would
be put; and
• the time when the product was put into circulation.
The presentation of the product includes the way the product is advertised,
as well as the information concerning the product on the label, and instructions
and directions for use. All these have to point out the dangers and risks of
using the product, and the consumer may expect that there are no other
dangers and risks than those indicated.
The producer has to take into account intended and reasonably expected
use of the product, and allow that, in practice, a user will not take all safety
and precautionary measures. The safety standards in force at the time when
the product was put into circulation are deciding factors in qualifying the
product as defective or not.
The production chain comprises the stages a product passes through on its
route from raw material to finished consumer product. Typically a product
could go from the raw material producer to a processor to be combined with
other raw materials into a semi-product before passing to a manufacturer
who might combine several semi-products into an end product for sale and
delivery to the consumer. It is generally accepted that each time a raw material,
semi-product or end product moves along the chain it has been ‘put into
circulation’. Hence the same commodity can be put into circulation several
times during the production and distribution process. It is only when the
commodity is in circulation (and if it is defective) that the producer will be
liable. During storage and processing the product is deemed not to be in
circulation and, based on the legislation concerning product liability, the
producer cannot be liable. However, producer liability is possible based on
other legislation such as, for example, labour laws. Processing aids are not
deemed to have been put into circulation if they are used for the preparation
of foodstuffs, but only if they have been passed along the chain.
The expectations of the consumer play an important part in determining
whether a product is defective or not. Hence other circumstances under
which a product is considered defective relate to the nature of the product
(for instance, a medicine has a different nature to a foodstuff), the advantages
and the usefulness of the product, the chance or frequency of appearance of
the danger, the appearance of the harmful additional workings, and the
knowableness of the danger. A product is not considered defective if there is
a mere lack of conformity to specifications, stemming from regulations or
from a contract, and the safety of the product is not at stake.
The cause of a product defect can be reduced to faults in production,
where there are three kinds of faults.
• Design defects. This kind of fault leads to categorical defects, and therefore
to a series of defective products. The fault only can be removed by a new
or changed design.
• Manufacturing faults. This kind of fault occurs during production, and
results in one or more products which deviate from other products made
according to the same specifications. Producers have a duty to check
products against specifications and ensure that defective products are not
put into circulation.
• Inadequate instructions or warnings. If the instructions or warnings
included in the advertisement of the product, on the label, in the instructions
and on the directions for use do not give any or enough information about
the safe use of the product, then they are inadequate.
The appearance of one of these faults does not mean that the product is
defective, unless the safety of the product is at stake.
Products may become unsafe after they are put into circulation, which is
not covered by this Directive. If a product becomes defective after being put
into circulation, the claim for damages must comply with national rules
based on fault. The producer has a duty of care as laid down in Directive
2001/95/EC21, and must monitor products accordingly, warning consumers
if the product, or its use, becomes unsafe.
21.6.7 Causal relationship

The plaintiff has to prove the damage, that the product is defective, and the
causal relationship between the defect and the damage. In practice, this is
very difficult for the plaintiff to prove, and the Directive does not allow a
conversion of the burden of proof. But it is within the scope of the Directive
for a judge to charge the producer with contradicting the thesis of the plaintiff,
if the plaintiff has provided a sufficient part of the proof.
21.6.8 Means of defence

There are a number of means for disclaiming liability, although all are very
difficult to fulfil in practice. Article 7 lays down that a producer shall not be
liable if they prove the existence of one or more of the following circumstances.
• The producer charged did not put the product into circulation, e.g. if the
product was stolen from the factory site, or the product has been
counterfeited.
• It is probable that the defect which caused the damage did not exist at the
time when the product was put into circulation, or that this defect came
into being afterwards. Wear of a product does not make it a defective
product.
• The product was neither manufactured by the producer for sale or any
form of distribution for economic purpose, nor manufactured or distributed
by the producer in the course of business. Both conditions must be fulfilled
for a successful appeal.
• The defect is due to compliance with mandatory regulations issued by the
public authorities. In this case, the government has prescribed the
composition or labelling, or certain conditions needed for permissibility
and approval. In the case of a successful appeal, the product is still a
defective product but the producer is not liable for it.
• The state of scientific and technical knowledge at the time when the
product was put into circulation was not such as to enable the existence of
the defect to be discovered. As described earlier, each Member State is
permitted to disallow this means of defence. In practice, an appeal under
this provision is not likely to be successful. The producer has a duty to
research the objective state of scientific and technological knowledge at
the time when the product was put into circulation.
• In the case of a manufacturer of a component, that the defect is attributable
to the design of the product into which the component has been fitted, or
to the instructions given by the manufacturer of the product.
21.6.9 Limitation period

Article 10 states that a limitation period of three years shall apply to proceedings
for the recovery of damages as provided for in this Directive. The limitation
period shall begin to run from the day on which the plaintiff became aware,
or should reasonably have become aware, of the damage, the defect and the
identity of the producer. These conditions are cumulative. The laws of the
Member States concerning suspension and interruption are unhindered by

these provisions.
21.6.10 Period of expiry

The period of expiry is laid down in Article 11, ‘the rights conferred upon
the injured person pursuant to this Directive shall be extinguished upon the
expiry of a certain period.’ This period is ten years from the date on which
the producer put the actual product which caused the damage into circulation,
unless the injured person has in the meantime instituted proceedings against
the producer.
21.6.11 Product recall

Rules concerning product recall are laid down in Directive 2001/95/EC,22
which has repealed Directive 92/59/EEC.23 Regulation 178/2002/EC 24 also
shows rules concerning recalls
21.7 Key issues in labelling of allergens, undeclared

allergens, food safety and product liability
21.7.1 Possibilities leading to liability under amended Directive
2000/13/EC
There are a number of possibilities that can lead to a producer being liable
for a defective product when it contains one or more allergenic ingredients.
Those that come under the scope of amended Directive 2000/13/EC25 include:
• an allergenic ingredient is listed in Annex IIIa of the amended Directive
2000/13/EC26 and has been consciously used in production and is still
present in a foodstuff but has not been indicated on the label.
• an allergenic ingredient is listed in Annex IIIa of the amended Directive
2000/13/EC27 and has been used in production and is still present in a
foodstuff, and has been indicated on the label.
These situations are discussed further within the framework of Directive
2001/95/EC,28 Regulation 178/2002/EC,29 Directive 85/374/EEC,30 Directive
2000/13/EC31 and Directive 2003/89/EC32.
Directive 2001/95/EC33
Product definitions
Foodstuffs with one or more allergenic ingredients are products for the purpose
of this directive, under the definitions listed in Section 21.4.2. A foodstuff
containing one or more allergenic ingredients which are not indicated on the
label is not a safe product under the definitions listed in Section 21.4.2.
Important considerations for foodstuffs containing one or more allergenic

ingredients are:
• the characteristics of the product, including its composition, packaging,
instructions for assembly and, where applicable, for installation and
maintenance;
• the presentation of the product, the labelling, any warnings and instructions
for its use and disposal and any other indication or information regarding
the product;
• the categories of consumers at risk when using the product, in particular
children and the elderly.
A foodstuff which contains one or more allergenic ingredients that are not
indicated on the label will be considered a dangerous product.
General safety requirement, conformity assessment criteria and

European standards
There are specific Community provisions governing product safety for
foodstuffs containing one or more allergenic ingredients (Section 21.4.3). A
foodstuff containing allergenic ingredients which are not indicated on the
label is not a safe food and may not be placed on the market.
Other obligations of producers and obligations of distributors

Other obligations (Section 21.4.4) apply to foodstuffs containing allergenic
ingredients because they present a risk to some consumers.
Product recall
Producers of foodstuffs have an obligation to monitor their products and to
recall products from the market if they present risks (Section 21.4.5). If an
allergenic ingredient is not indicated on the label of a foodstuff, its presence
forms a risk for some consumers and hence the product is dangerous. Once
the producer is aware of this risk, the foodstuff must be withdrawn from the
market. Producers and distributors are also obliged to recall dangerous products
from consumers. In some cases, other appropriate actions may include
adequately and effectively warning consumers. Governments can oblige
producers and distributors to recall a product, from the market and/or consumers.
Regulation 178/2002/EC34
Definition of food
Foodstuffs with one or more allergenic ingredients belong to food under the
definition of food for this regulation (Section 21.5.2).
Food safety requirements

Under the definition of unsafe foods (Section 21.5.3), foodstuffs with allergenic
ingredients that are not indicated on the label are unsafe for a specific category
of consumers and should not be placed on the market.
Responsibilities of food and feed business operators

Producers of foodstuffs must ensure that the foodstuffs do not contain allergenic
ingredients if these are not indicated on the label (Section 21.5.4).
Traceability
Producers of all foodstuffs, with and without allergenic ingredients, indicated
on the label or not, must meet traceability requirements as listed in Section
21.5.5
Responsibilities of food business operators for non-compliant foodstuffs

Producers of foodstuffs with allergenic ingredients which are not indicated
have the ultimate responsibility for placing safe foods on the market (Section
21.5.6). They are obliged to withdraw such foodstuffs as soon as possible
after they become aware of requirements to indicate the allergenic ingredients.
Directive 85/374/EEC35
Requirements for liability
The producer shall be liable for damage caused by a defect in his product
under the requirements described in Section 21.6.3.
Producer
The definition ‘Producer’ as described in Section 21.6.4 applies to the producer
of a foodstuff containing an allergenic ingredient that is not indicated on the
label.
Product
Foodstuffs with and without allergenic ingredients are products in the sentence
of this directive (Section 21.6.6).
Damage
A plaintiff can claim damage as described in Section 21.6.5.
Defective product
A foodstuff containing an allergenic ingredient is a defective product if the
allergenic ingredient is not indicated on the label, because the allergenic
ingredient presents a risk to some consumers (Section 21.6.6). A product
containing an allergenic ingredient may also be defective if there are faults
during production, such as defects in the design of the product and the label.
If the instructions or warning concerning the indication on the label do not
give enough or any information about one or more allergenic ingredients,
then the warnings or instructions are inadequate.
Causal relationship
The plaintiff has to prove the damage, that the product is defective, and the
causal relationship between the defect and the damage (Section 21.6.7). The
proof of the damage and the defect will be relatively straightforward. The
causal relationship will be more difficult to prove, because it will be possible
that the plaintiff will have consumed other foodstuffs containing the same
allergenic ingredients.
Means of defence
There are a number of means of defence for a producer to prove that he is not
liable (Section 21.6.8). But, in the case of foodstuffs containing allergenic
ingredients, the defect which caused the damage will have existed at the time
when the product was put into circulation. In addition, it will be clear that the
product was manufactured for sale and distribution by the producer, that the
defect is not due to compliance with mandatory regulations issued by the
public authorities, that the state of scientific and technical knowledge at the
time when the product was put into circulation made it possible to discover
the existence of the defect, and that the producer is not the manufacturer of
a component. The defect is therefore attributable to the design of the product
made by the producer.
Limitation period
The limitation period of three years applies in the case of foodstuffs containing
allergens (Section 21.6.9).
Period of expiry
The 10-year period of expiry applies in the case of foodstuffs containing
allergens (Section 21.6.10).
Directive 2000/13/EC36 and Directive 2003/89/EC37

Definitions
The indication of the presence of an allergenic ingredient in a foodstuff
comes under the requirements for labelling (Section 21.2.1).
Misleading information
Omitting to indicate the presence of allergenic ingredients could be qualified
as misleading because it concerns the characteristics of the foodstuff (Section
21.2.2).
Information on the label

The following particulars on the labelling of foodstuffs concern allergenic
ingredients (Section 21.2.3):
• the name under which the product is sold;
• the list of ingredients; and
• instructions for use when it would be impossible to make appropriate use
of the foodstuff in the absence of such instructions.
An allergenic ingredient must be indicated in order to comply with the
requirements.
Name of the foodstuff

A foodstuff may be sold under a name which clearly refers to the allergenic
ingredient. If this is the case, further indication is not needed (Section 21.2.4).
List of ingredients
Ingredients must be indicated under the requirements to list ingredients (Section
21.2.5). In addition to this, allergic ingredients which are listed in the Annex
to Directive 2003/89/EC must be indicated in a prescribed manner on the
label (Sections 21.2.5 and 21.3.2).
Labelling of allergenic ingredients of beverages containing more than 1.2%

by volume of alcohol
If a foodstuff is a beverage and one or more allergenic ingredients are not
indicated, then the labelling does not meet the requirements (Section 21.3.3).
Labelling of allergenic ingredients of other foodstuffs

Any allergenic ingredient used in production and present in the foodstuff
must be indicated on the label even if a list of ingredients is not needed
(Section 21.3.4). The lack of an indication does not meet the requirements.
If the name under which the foodstuff is sold clearly refers to the ingredient
concerned, the indication is not required.
Conditions for labelling an allergenic ingredient

The Directive has open specifications for indication, stating that ‘the ingredient
shall be indicated on the label with a clear reference to the name of this
ingredient’, and that ‘the indication shall not be required if the name under
which the foodstuff is sold clearly refers to the ingredient concerned’ (Section
21.3.5). But if a required indication is missing, then the label does not meet
the requirements.
Researching the need to indicate an allergenic ingredient

The producer is expected to have researched the need to indicate any allergenic
ingredients in their products (Section 21.3.2). If the producer has judged
incorrectly that an indication is not needed, whereas in fact it is required,
then the labelling will not meet the requirements.
QUID regulation
The QUID regulation is relevant if the quantity of an allergenic ingredient is
such that it must be stated (Section 21.2.6).
Net quantity
The requirements concerning the net quantity are not relevant for this subject
(Section 21.2.7)
Date of minimum durability

The requirements concerning the date of minimum durability are not relevant
for this subject (Section 21.2.8).
Instructions for use

The presence of one or more allergenic ingredients listed in Annex IIIa has
to be indicated under instructions for use requirements (Sections 21.2.5 and
21.3.2).
Language of the information on the label

Language requirements for labels apply to foodstuffs containing allergenic
ingredients (Section 21.2.11).
21.7.2 Other possibilities for liability

There are other possibilities for liability. If there is no legal requirement to
indicate the allergenic ingredient, then the producer may indicate the presence
of an allergenic ingredient by the indication ‘contains’ or ‘may contain’.
Examples include:
• an allergenic ingredient is not (yet) listed in Annex IIIa of the amended
Directive 2000/13/EC,38 has been consciously used in production and is
still present in a foodstuff. This ingredient has not been indicated on the
label because listing it is not mandatory.
• an allergenic ingredient, which may or may not be listed in Annex IIIa of
the amended Directive 2000/13/EC,39 has not been used in production of
a foodstuff, but it is present because of cross-contamination. It has not
been indicated on the label, but it is known or one could reasonably
expect that this contamination could occur.
• an allergenic ingredient, which may or may not be listed in Annex IIIa of
the amended Directive 2000/13/EC,40 has not been used in production of
a foodstuff, but it is present because of cross-contamination. It has not
been indicated on the label and one could not reasonably expect that this
contamination should occur.
In these instances, the claim of the plaintiff has to be based on national rules
based on fault, and the plaintiff has to prove the fault. The circumstances
have to be taken into account, such as the prevailing knowledge about the
danger of allergenic ingredients in general, and the specific knowledge regarding
the particular allergenic ingredient concerned. An increasing awareness of
the dangers of allergenic ingredients increases the responsibility of the
producer.
21.8 Future trends

21.8.1 Labelling in general
Additional information will be required on the label, e.g. information about
nutritional facts. Currently, including nutritional information is voluntary,
but if a producer makes a claim such as ‘light’, then supporting information
must be provided on the label. The type and content of nutritional and health
claims will be regulated in the near future.
21.8.2 Labelling of allergens

At present there are 12 allergens that must be labelled according Directive
2003/89/EC. However, there are other ingredients used in foods that can
cause allergies or intolerances in consumers which do not come under
mandatory labelling requirements. In order to protect consumers, it is anticipated
that additional allergens will be listed in the future. The opposite also applies,
in that ingredients which have been scientifically established not to cause
adverse reactions will be deleted from the list of allergens.
21.8.3 General product safety

Public authorities can oblige producers (and distributors) to carry out a product
recall, from the market and from consumers where necessary, on the basis of
Directive 92/59/EEC41. An extension of this obligation may be to authorise
consumers to obtain injunctions, in order to have dangerous products be
withdrawn from the market.
21.8.4 Product liability

The supplier of foods may be assimilated with the producer where product
liability is concerned. At present, the liability of a supplier can only be
established based on fault under national law. The strict liability regime may
also be extended to cover damages to professional products.

21.9.1 Additional information on regulatory and legal developments
Additional information on the regulatory and legal developments on labelling
of foods, labelling of allergens, product liability and product safety can be
found at the following web-sites:
• CBL
Centraal Bureau Levensmiddelenhandel
http://www.cbl.nl/
• CIAA
Confédération des Industries Agro-Alimentaires de l’EU
http://www.ciaa.be/
• European Community
http://europa.eu.int/eur-lex/nl/index.html
• FNLI
Federatie Nederlandse Levensmiddelen Industrie
http://www.fnli.nl/
• TNO Quality of Life
http://www.voeding.tno.nl/
• VAI
Nederlandse Voedingsmiddelen Industrie
http://www.vai-voeding.nl/
• Europe
— Labelling
http://europa.eu.int/comm/food/food/labellingnutrition/foodlabelling/
index_en.htm
— Consumer affairs/product safety/product liability
http://europa.eu.int/comm/consumers/index_en.htm
— Public health
http://europa.eu.int/comm/health/horiz_publications_en.htm
— European Food Safety Authority (EFSA)
http://www.efsa.eu.int/index_en.html
21.9.2 Tabling of the main legislation of the European Union

concerning establishing community, product liability, labelling and
product safety
• EC Treaty42
Treaty establishing the European Community.
• Directive 79/112/EEC43
Council Directive 79/112/EEC of 18 December 1978 on the approximation
of the laws of the Member States relating to the labelling, presentation
and advertising of foodstuffs. This directive has been repealed by Directive
2000/13/EC,44 so it is no longer in force.
Council Directive of 25 July 1985 on the approximation of the laws,
regulations and administrative provisions of the Member States concerning
liability for defective products (85/374/EEC).
Council Directive 92/59/EEC of 29 June 1992 on general product safety.
This directive has been repealed by Directive 2001/95/EC,47 so it is no
longer in force.
• Directive 1999/34/EC48
Directive 1999/34/EC of the European Parliament and of the council of
10 May 1999 amending Council Directive 85/374/EEC on the approximation

of the laws, regulations and administrative provisions of the Member
States concerning liability for defective products.
Directive 2000/13/EC of the European Parliament and the Council of 20
March 2000 on the approximation of the laws of the Member States
relating to the labelling, presentation and advertising of foodstuffs.
Directive 2001/95/EC of the European Parliament and of the Council of
3 December 2001 on general product safety.
• Regulation 178/2002/EC, General Food Law or GFL51
Regulation (EC) No 178/2002 of the European Parliament and of the
Council of 28 January 2002 laying down the general principles and
requirements of food law, establishing the European Food Safety Authority
and laying down procedures in matters of food safety.
Directive 2003/89/EC of the European Parliament and of the Council of
10 November 2003 amending Directive 2000/13/EC as regards indication
of the ingredients present in foodstuffs,
• White Paper food safety
White Paper on Food Safety, COM (1999) 719 final, Brussels, 12.1.2000.
21.10 References
1. Rewe-Central A G vs. BundesmonopolVerwaltung fúr Branntwein (Cassis de Dijon)
(1979) ECR 469.
2. OJEC 2002 L31/1
3. OJEC 2000 L109/29
4. OJEC 2003 L308/15
5. OJEC 1997 L043/1
6. OJEC 1985 L210/29
7. OJEC 1992 L228/24
8. OJEC 1999 L141/20
9. OJEC 2001 L011/4
10. OJEC 2000 L109/29
11. OJEC 1979 L33/1
12. OJEC 2003 L308/15
13. OJEC 2000 L109/29
14. OJEC 2005 L75/33
15. OJ 2001 L11/4
16. OJ 1992 L228/0024
17. OJ 2002 L031/001
18. OJ 1985 L210/29
19. OJ 1985 L210/29
20. OJEC 1999 L141/20
21. OJ 2001 L11/4
22. OJ 2001 L11/4
23. OJEC 1992 L228/24
24. OJ 2001 L31/1

25. OJEC 2000 L109/29
26. OJEC 2000 L109/29
27. OJEC 2000 L109/29
28. OJEC 2001 L011/4
29. OJEC 2002 L31/1
30. OJEC 1985 L210/29
31. OJEC 2000 L109/29
32. OJEC 2003 L308/15
33. OJEC 2001 L011/4
34. OJEC 2002 L31/1
35. OJEC 1985 L210/29
36. OJEC 2000 L109/29
37. OJEC 2003 L308/15
38. OJEC 2000 L109/29
39. OJEC 2000 L109/29
40. OJEC 2000 L109/29
41. OJ 1992 L228/0024
42. OJEC 2002 C325/35
43. OJEC 1979 L33/1
44. OJEC 2000 L109/29
45. OJEC 1985 L210/29
46. OJEC 1992 L228/24
47. OJEC 2001 L011/4
48. OJEC 1999 L141/20
49. OJEC 2000 L109/29
50. OJEC 2001 L011/4
51. OJEC 2002 L31/1
52. OJEC 2003 L308/15
22
Conclusions
S. Hefle, University of Nebraska, USA and S. Koppelman,
AllerTeQ and University Medical Centre Utrecht, The Netherlands
22.1 Recent literature and trends

While this book was in production, several new articles were published
concerning allergen detection methods. Some methods for other allergenic
foods were developed as were several novel experimental approaches for
detection of allergenic residues or markers of allergenic residue. The majority
of the new publications dealt with validation trials of commercially available
test kits and the optimizing recoveries for these tests. We believe that all
these developments are important and we hereunder summarize the main
conclusions of the papers published in 2004–05 on allergen detection methods.
Several methods have been developed for allergenic foods that could
previously not be detected. A sandwich enzyme-linked immunosorbent assay
(ELISA) for lupin (Holden et al., 2005) utilizing polyclonal antibodies was
reported to have a 1 ppm detection limit, with high specificity for lupin and
only minor cross-reactivities for other legumes. Recoveries in spiking studies
were 60–115% in various food matrices, but model foods were not utilized
in this study. Additionally some polymerase chain reaction (PCR) methods
for some allergens not previously described were published. A PCR method
for mackerel contamination was developed by Aranishi and Okimoto (2004).
Another one for crustacean shellfish was described by Brzezinski (2005),
but it had a reported detection limit for shrimp of <0.1%, which is not
sensitive enough for allergen concerns. Also, a buckwheat PCR method was
described by Hirao et al. (2005) and had a detection limit of 1 ppm (w/w) of
buckwheat DNA spiked in wheat DNA. Before these newly developed methods
can be applied on a routine basis in the food industry, the methods should be
validated in terms of precision, recovery, sensitivity and specificity, etc.
Such validation studies have been undertaken recently for some allergen
detection kits and several publications became available in 2005:
• A validation trial designed by the U.S. Food and Drug Administration
(Park et al., 2005) utilized three laboratories in a performance validation
of three commercial peanut ELISA kits at 0 and 5 ppm to be used for
screening purposes using peanut butter spiked into breakfast cereal, cookies,
ice cream, and milk chocolate spiked at 0 and 5 ppm peanut. They found
when two validated kits are used in parallel, the probability of detecting
5 ppm is >95% peanut protein in four different food matrixes.
• Another interlaboratory trial of five commercial ELISA kits for peanut in
biscuits (cookies) and dark chocolate model foods containing peanut at 0–
10 ppm was published by Poms et al. (2005). They found that, in general,
these five quantitative kits performed better in the 5–10 ppm range than at
lower ranges of 2–2.5 ppm, but in some cases, 2 ppm was below the
official detection limit of the kit, so the authors were pushing the technology
in some cases beyond what the kit was designed to do. Impressively, 34
laboratories participated in this trial, and results were obtained from 12–
14 laboratories for each commercial kit. For the biscuits (cookies), peanut
flour was mixed and then baked in household-size portions and equipment.
For dark chocolate, peanut-free chocolate was melted at 60 °C and then
peanut flour was mixed in. Average recoveries varied with all types of test
kits and matrices, and both the type of kit and the matrix seemed to be
important for recovery.
• More work has been performed on interlaboratory validations for allergen
detection methods. Akiyama et al. (2005) described two interlaboratory
trials on peanut and buckwheat ELISAs: extracts of biscuit, sauce, chocolate
and butter spiked with peanut standard protein at a level of 5–20 ng/mL
were found to be reproducible in the study. Mean recoveries of the peanut
standard protein from the food extracts were over 40% using two ELISA
methods. The detection limits of both ELISA methods were 2–2.5 ng/mL
in sample solutions. Inter-laboratory evaluation studies were conducted
for the ELISA methods for buckwheat. Extracts of snacks, buns and udon
noodles spiked with buckwheat protein at a level of 5–20 ng/mL as sample
solutions were analyzed in replicate at 10 laboratories. Mean recoveries
of the buckwheat standard protein from the food extracts were over 40%
in the two ELISA methods. The detection limits of both ELISA buckwheat
methods were 1 ng/mL.
With the data of the validation studies available, we believe that the detection
and quantification of peanut residue in food products, as desired by the food
industry and food regulators, is possible for the majority of food products.
One should, however, always be aware of specific situations that may disturb
proper analysis. Recovery from the food matrix may be the most important
issue here, and some food matrices are particularly notorious. Chocolate, for
example, is a difficult matrix to extract peanut residue from. Akkerdaas et al.
Conclusions 407
(2004) designed a protocol to improve the extraction of hazelnut allergens

from dark chocolate. This assay took advantage of the protein-solubilizing
properties of pepsin, a proteolytic enzyme. Pepsin-digested hazelnut was
used to make polyclonal antibodies. The reported detection limit was 1 ppm
and good recovery (97%) was obtained even from a normally difficult dark
chocolate matrix. The study employed model foods manufactured to contain
different amounts of hazelnut. The requirement to digest the food matrix for
each individual sample with pepsin makes it impractical to use in a food
production setting in most cases, however.
In addition to new methods and validation studies of commercial test kits,
we observed several trends in the detection of allergenic residues. Stephan et
al. (2004) found that the use of celery PCR and a peanut ELISA, together
with more inexpensive total protein measurement techniques, to detect protein
traces can be a fast and cost-effective method for monitoring the effectiveness
of wet cleaning procedures. Regulatory agencies and some food companies
could make use of multiple allergen detection in one assay; one paper by Ben
Rejeb et al. (2005) described a competitive multi-assay for peanut, hazelnut,
almond, cashew and Brazil nuts in a single run. The assay was adapted for
screening purposes applied to chocolate samples. The limit of detection was
reported to be 1 ppm. Another interesting approach was to use a lipid transfer
protein (LTP), a highly conserved protein present in plants, to design a pan-
allergen ELISA (Zuidmeer et al., 2005); a single sandwich ELISA method
can be used to detect LTP from apple, cherry, nectarine, and hazelnut. However,
this might not be practical for food companies that are only interested in one
of these allergenic substances (example hazelnut) in the presence of other
LTP-containing ingredients in the food matrix (example apple). Nevertheless,
such approaches can be further explored for allergens from sources that are
closely related such as fish or crustaceans.
There may be novel methods in the future for the quick and reliable
assessment of the content of allergenic residue in food, either based on
immunochemical methods, DNA techniques, or based on an entirely new
methodology. One unconventional new approach to peanut PCR was described
by Sforza et al. (2005): circular dichroism on directly stained peanut DNA
amplified by PCR. The detection limit was very low, reported to be a few
picomoles. However, both the low detection limit and the technical aspects
of the method do not make it very practical for the food industry to use. We
should be aware, though, of emerging technologies that are complementary
to or may replace conventional methods in the future.
22.2 Relating detection limits to clinically relevant doses

Some state and national regulatory agencies have used allergen detection
methods to investigate consumer complaints from food-allergic consumers.
The availability of methods for allergenic residue detection will make
investigation of consumer complaints and gathering of data easier for risk

assessment for all of these stakeholders, but the reliability of some analytical
results may be questioned until methods are validated and collaboratively
studied. In exquisitely allergic individuals, the threshold dose for elicitation
of such reactions is often considered to be zero. However, some food-allergic
patients report that they can tolerate small quantities of allergenic food.
Studies covered in the various chapters in this book include some that have
used allergen detection methods to discover that some packaged food products
contain undeclared allergen residues at levels that could be potentially hazardous
to allergic individuals. These findings suggest that residues do exist in retail
food products, with no report of allergic reactions. This could be because
either the eliciting doses are higher than the levels present or perhaps that
food-allergic individuals do not buy certain products that they feel are vulnerable
to contamination with undeclared allergenic food residues. Some uncertainty
is likely to remain regarding whether the threshold dose has been determined
for the most sensitive individuals. Clinicians will never be completely certain
that the most sensitive individual has been identified and tested. However, an
analogy can be drawn to the infant formula industry where hypoallergenic
infant formula based upon extensively hydrolyzed casein is safe for the vast
majority of milk-allergic infants. Yet, hypoallergenic infant formula is known
to elicit adverse reactions in a very small number of milk-allergic infants.
This analogy illustrates the difficulty in developing threshold doses for all
allergic consumers. Instead, the goal should be to identify threshold doses
that protect the vast majority of allergic consumers. Allergic individuals can
(and probably do) ingest foods, on occasion, containing lower amounts of
their offending food without any adverse reactions.
Threshold doses are currently reported to be in the low mg range, with
subjective (not observable) symptoms reported in some studies of peanut to
be 100 ug. As discussed in this book, detection limits of current commercial
methods (1–2.5 ppm, or 1–2.5 mg/kg) for residues of allergenic foods appear
to be in the ideal range to detect potentially hazardous residues of undeclared
allergens in foods. Some noteworthy publications over the years have suggested
that detection methods be 5–10 ppm, and that advice remains fairly sound
now that more evidence exists on threshold doses. The development of analytical
procedures that are more sensitive than the current ones does not seem to be
justified clinically. The ideal range of sensitivity of detection methods for
allergenic residues should be driven by knowledge of minimal eliciting dose
levels. Good manufacturing practices in the food industry and regulatory
limits imposed by governmental agencies should also be based on minimal
eliciting doses. The goal should be to protect allergic consumers while allowing
them the greatest variety of foods in their diets rather than the detection of
ever-vanishing levels of allergenic food residues that have no adverse health
consequences to even the most sensitive consumers. Use of low detection
limit methods may actually contribute to unjustified proliferation of ‘may
contain’-type labeling and therefore decrease the quality of life for food-
Conclusions 409
allergic consumers. An increase in ‘may contain’-type labeling may also

contribute to food-allergic consumers being driven to exhibit risky behavior
(especially teenagers, risk-takers by nature), and make dangerous choices to
disregard this type of labeling.
22.3 Reference materials, extraction and recovery

Short-term needs in the allergen detection area include the critical need for
reference materials and validation trials, and also the development of assays
for some commonly allergenic foods for which none exist. Any method used
for the detection of allergens in foods needs to be validated for its sensitivity,
specificity, and detectability in the food matrix of interest to guarantee its
applicability. In many studies, no information on the allergenic source material
used to spike or even to make model foods is provided. Even in validation
trials that have already been published, little or no detail is given on the
nature of the material used to make the spiked or model (manufactured to
contain specific amounts of allergen) matrices. It is therefore difficult to
make a true comparison between one kit and another. Currently, efforts to
generate such reference materials are in progress at various centers. The use
of standard reference materials is necessary to make the results of methods
comparable and to be able to attain reproducible interlaboratory results. The
development of quantitative allergenic methods can still be challenging due
to the lack of suitable standard reference materials.
In addition, simple ‘spiking’, i.e. putting a peanut extract into a food
extract, or dry milk into a dry food powder, is no longer considered appropriate
in assessing the performance of allergen detection methods, except when
investigating possible matrix interference purposes. The true extraction
efficiency of a method and the true recovery from a food matrix can only be
ascertained with the use of model foods. The development and production of
such reference materials for the analysis of food allergens face several
challenges, and is it not practical to make model foods for every matrix.
While efforts to make model foods should be applauded, making them in a
home kitchen or on a small scale is of limited use, as they do not mimic true
processing conditions or practical situations. Making model foods according
to pilot plant or true industrial conditions is not necessarily an easy task, but
some research groups are doing work in this area, and these materials will be
key in the evaluation of allergen detection methods.
22.4 Developing realistic and practical detection methods

Before 1998, no commercial methods were available for the detection of
allergenic residues in foods or on food processing equipment. The methodology
for detection of allergenic residues in foods improved with the advent of

ELISA and PCR methods. As new methods have been developed and
commercialized, the food industry has increasingly used them. However,
there is a lack of easy-to-use methods for some of the most commonly
allergic foods (for example crustaceans and fish). For the majority of users
(food companies), useful methods will include rapid, simple (easy-to-use)
kits, with on-(production) line monitoring. Rapid ELISAs and lateral flow
devices/dipsticks will, therefore, continue to be the most-used format in the
food industry for some time. Complicated, long extraction protocols and
lengthy methods are not workable. Recombinant antibodies are starting to be
used in developing ELISA methods for phytotoxin analysis, and the ability
to engineer specificity of antibodies may offer unique opportunities in the
future in allergenic residue detection. PCR does not seem to be practical for
analysis on food production lines, whereas lateral flow devices/dipsticks are
favored because of easier and faster use. However, other techniques such as
PCR can be very useful for regulatory agencies, companies which possess
the appropriate equipment and space, or through contract laboratories in
providing additional information about a sample. In addition, newer methods
such as surface plasmon resonance (SPR) on automated platforms can provide
rapid, efficient analysis of large quantities of samples without direct supervision,
and might therefore be feasible in some production settings. Other technologies
may become more feasible in the future. An example is microfluidic separation
where it is anticipated that separation devices may eventually be handheld.
The more recent developments in assay technologies are promising for rapid
and multiple analysis for allergen detection in food in the future. However,
at the present time, these techniques are impractical for use by the food
industry. The food industry’s need for multiplex methodologies remains to
be fully met.
22.5 Summary
The unintended presence of allergenic residue in foods is, and will be in the
future, an issue that affects the food allergic consumer, and thereby the food
industry. Data on threshold levels of an allergen to elicit an allergic reaction
in food-allergic patients are available for a selected number of allergens, and
based on these threshold levels, a sensible detection limit for food allergen
detection assays can be set. We propose that this detection limit should be
between 1 and 10 ppm, providing that the recovery of the analytical procedure
is at least 50% for the food products where the test is applied, and preferably
higher. Since the early 1990s many methods have been published to detect
allergenic residue in food and food ingredients. Some assays have been
validated in multi-laboratory trials (peanut) and have been shown to be useful
for assessing the amount of allergenic residue in foods. For some allergens
Conclusions 411
other than peanut such validation studies are in progress, and for other allergens,
methods need to optimized before validation studies can be considered.
Trace contamination can arise from practices in the food industry, restaurants,
and in homes and schools. Food-service employees and food processors
alike must be very vigilant about the use of shared utensils and equipment.
Most of the deaths from inadvertent exposure to allergenic foods have occurred
in food-service situations (restaurants, schools, homes). However, in many
countries, food-service establishments do not have specific regulations about
handling or labeling allergens. The increased demand among the general
consuming public for a wide variety of food products, and the ever-increasing
consumption of food prepared outside the home, can assist in producing
reactions to hidden food allergens. While the majority of reactions occur
after ingestion of a food that is prepared in a food-service environment rather
than packaged food products, the food processing industry has been subjected
to increasing scrutiny of its allergen controls. The resulting impact has been
remarkable: the food industry has made significant investment, effort and
improvements in allergen control in recent years. Food processors often use
shared equipment and facilities for the production of processed foods with a
variety of formulations. This is often out of necessity rather than just fueled
by economic reasons. Due to various limitations and practices in food
processing, foods may occasionally contain trace residues of other foods that
may not be declared on the ingredient label. The food industry began investing
a significant amount of time, money, and effort to attain better allergen
control in manufacturing facilities in the late 1980s. Since the late 1990s, the
commercialization of some test methods for the detection of some allergenic
foods has allowed food manufacturers to use these tests to create important
data to assess and design allergen control strategies and protocols.
Debate and discussion about the effects of processing on allergenicity and
allergenic residue detection in foods will be ongoing. In addressing new
labeling legislation in the EU and the US, analytical methods are being used
to assess allergenicity of some ingredients. However, analysis of some allergen-
derived ingredients can pose serious challenges. Problematic ingredients
include oils and lecithin due to their oily nature, and also hydrolyzed proteins
and residues of allergenic growth substrates in gums, starter cultures, enzymes,
etc. Standard ELISAs cannot be used in the case of hydrolyzed/proteolyzed/
fermented ingredients and PCR cannot be used to assess how allergenic the
ingredient is. Methods involving IgE obtained from individuals allergic to
the particular food are the only way to determine allergenicity of these types
of ingredients with the exception of oral challenge studies in allergic individuals.
Although there is much work to do to deal with the issue of detecting
allergenic residues in food, we believe that the enormous output of many
investigators and test kit manufacturers, and the use of test kits by the food
industry, has already succeeded in providing better protection for the food-
allergic consumer. This is a solid basis for the further development of tools
for the detection of allergenic residues.
22.6 References
Akkerdaas, J H, Wensing, M, Knulst, A C, Stephan, O, Hefle, S L, Aalberse, R C, van
Ree, R (2004) ‘A novel approach for the detection of potentially hazardous pepsin
stable hazelnut proteins as contaminants in chocolate-based food’, J Agric Food Chem
52, 7726–7731.
Aranishi, F, Okimoto, T (2004) ‘PCR-based detection of allergenic mackerel ingredients
in seafood’, J Genetics 83, 193–195.
Akiyama, H, Nakamura, K, Harikai, N, Watanabe, H, Iijima, K, Yamakawa, H, Mizuguchi,
Y, Yoshikawa, R, Yamamoto, M, Sato, H, Watai, M, Arakawa, F, Ogasawara, T, Nishihara,
R, Kato, H, Yamauchi, A, Takahata, Y, Morimatsu, F, Mamegoshi, S, Muraoka, S,
Honjoh, T, Watanabe, T, Sakata, K, Imamura, T, Toyoda, M, Matsuda, R, Maitani, T
(2004) ‘Inter-laboratory evaluation studies for establishment of notified ELISA methods
for allergic substances (peanuts)’, (in Japanese) Shokuhin Eiseigaku Zasshi 45, 325–
331.
Akiyama, H, Nakamura, K, Harikai, N, Watanabe, H, Iijima, K, Yamakawa, H, Mizuguchi,
Y, Yoshikawa, R, Yamamoto, M, Sato, H, Watai, M, Arakawa, F, Ogasawara, T, Nishihara,
R, Kato, H, Yamauchi, A, Takahata, Y, Morimatsu, F, Mamegoshi, S, Muraoka, S,
Honjoh, T, Watanabe, T, Sakata, K, Imamura, T, Toyoda, M, Matsuda, R, Maitani, T
(2004) ‘Inter-laboratory evaluation studies for establishment of notified ELISA methods
for allergic substances (buckwheat)’, (in Japanese) Shokuhin Eiseigaku Zasshi 45,
313–318.
Ben Rejeb, S, Abbott, M, Davies, D, Cleroux, C, Delahaut, P (2005) ‘Multi-allergen
screening immunoassay for the detection of protein markers of peanut and four tree
nuts in chocolate’, Food Addit Contam 22, 709–715.
Brzezinski, J L (2005) ‘Detection of crustacean DNA and species identification using a
PCR-restriction fragment length polymorphism method’, J Food Prot 68, 1866–1873.
Hirao, T, Imai, S, Sawada, H, Shiomi, N, Hachimura, S, Kato, H (2005) ‘PCR method for
detecting trace amounts of buckwheat (Fagopyrum spp.) in food’, Biosc Biotechnol
Biochem 69, 724–731.
Holden, L, Faeste, C K, Egaas, E (2005) ‘Quantitative sandwich ELISA for the determination
of lupine (Lupinus spp.) in foods’, J Agric Food Chem 53, 5866–5871.
Park, D L, Coates, S, Brewer, V A, Garber, E A, Abouzied, M, Johnson, K, Ritter, B,
McKenzie, D (2005) ‘Performance Tested Method multiple laboratory validation study
of ELISA-based assays for the detection of peanuts in food’, J AOAC Int 88, 156–160.
Poms, R E, Agazzi, M E, Bau, A, Nrohee, M, Capelletti, C, Norgaard, J V, and Anklam,
E (2005) ‘Inter-laboratory validation study of five commercial ELISA test kits for the
determination of peanut proteins in biscuits and dark chocolate’, Food Addit Contam
22, 104–112.
Sforza, S, Scaravelli, E, Corradini, R, Marchelli, R (2005) ‘Unconventional method based
on circular dichroism to detect peanut DNA in food by means of a PNA probe and a
cyanine dye’, Chirality 17, 515–521.
Stephan, O, Weisz, N, Vieths, S, Weiser, T, Rabe, B, Vatterott, W (2004) ‘Protein
quantification, sandwich ELISA, and real-time PCR used to monitor industrial cleaning
procedures for contamination with peanut and celery allergens’, J AOAC Int 87, 1448–
1457.
Zuidmeer, L, van Leeuwen, W A, Budde, I K, Cornelissen, J, Bulder, I, Rafalska, I,
Besoli, N T, Akkerdaas, J H, Asero, R, Rivas, M F, Mancebo, E G, van Ree, R (2005)
‘Lipid transfer proteins from fruit: cloning, expression and quantification’, Int Arch
Allergy Immunol 137, 273–281.
Index
Index terms Links
A
adjuvants 69
adulteration regulations 363
advertising and promotion of foods 319
agricultural biotechnology 153
albumins 41 221 247
alpha-lactalbumin 40 221
bovine serum albumin 221
ovalbumin 43 227
parvalbumins 38 284
serum albumins 41
2S albumins 27
allergen labelling directive 267 382 398 401
allergen management 318
Allergen Nomenclature Subcommittee 44
allergosorbents 87
almond detection 210
alpha-amylase and protease inhibitors 29
alpha-caseins 41
alpha-lactalbumin 40 221
analytes 325 353
anaphylactoid reactions 5
anaphylaxis 150 152
This page has been reformatted by Knovel to provide easier navigation. 413
414
Index terms Links
animal derived allergens 37
ATP: guanido phosphotransferases 39
calcium-binding EF-hand proteins 38
caseins 41 221 224
glycoside hydrolases 40
immunoglobulins 41
Kazal-type serine protease inhibitors 42
lipocalins 40
serum albumins 41
transferrins 42
tropomyosins 14 37 39 172
282 283
annealing process 126 132
anti-food allergen antibody specificity 300
anti-food protein antisera 299
antibodies 65 110 176
antigen-antibody interactions 66
see also cross-reactions
classes of 65 112
engineered antibodies 72
IgA class 9
IgE class see IgE antibodies
IgG class 65 71
IgM class 71
immobilization 161
immunogen tolerance 72
immunoglobin structural model 65 110
monoclonal antibodies 67 73 75 257
one-step tests 176
polyclonal antibodies 67 75 257
production of 66 68
recognition of antigens 68
second antibodies 68

415
Index terms Links
antigen-antibody interactions 66
see also cross-reactions
antigenic determinants 68
antisense RNA strategy 43
antiserum purification 74
apricots 28
AQA systems see quality assurance
aragose gel electrophoresis 128 129
arginine kinases 40
ascites 73
assay development 321 324
for allergen risk management effectiveness 324
for compliance 324
marker compounds 275 322 348
matrix interference 326
and protein changes during processing 326
for residual allergen measurement 324
for single allergenic proteins 325
variability in extracting analyte 325
asthma, baker’s asthma 149 244
atopic disease 315
atopic human subjects 82
ATP: guanido phosphotransferases 39
audits 370
avoidance diets 10
cross-reactions 14
ingredient allergenicity 13
minimum eliciting doses 11
B
baker’s asthma 149 244
beef allergy 38 41

416
Index terms Links
best before dates 381 400
Bet v 1 allergens 34 43
beta-1,3-glucanases 31
beta-caseins 41
beta-conglycinin 25 277
beta-lactoglobulin 41 221
bias in testing methods 331
binding allergens 161
non-specific binding 167
biosensors 120 159
dairy and egg residue detection 235
bird’s nest allergy 152
bird-egg syndrome 227
blotting transfer 231
bovine serum albumin 221
Brazil nut detection 211
C
calcium-binding EF-hand proteins 38
carageenan 368
caseins 41 221 224
cashew nut detection 211
CATH protein structure classification 45
celery detection 133
cell-mediated reactions 4 7 10
cereals
and coeliac disease 10 30 244 247
265
oats 247 262
PCR detection 133
prolamins 26 29

417
Index terms Links
cereals (Continued)
see also wheat gluten
certified reference materials (CRM) 351
CFA (complete Freund’s adjuvant) 70
chitinases 32
chocolate matrix 338
chocolate products 206
classes of antibodies 65 112
classification of allergies 4
cleaning procedures 308
and contamination of test components 333
cloth-ELISA 190
Codex Alimentarius standards, gluten-free foods 248 267
coeliac disease 10 30 244 265
coeliac-toxic cereals 247
collagen 14
color development 344
commercial allergenic residue test kits 116
availability 116
for dairy and egg residues 235
sampling errors 119
strengths and weaknesses 116
commercial assays 137
commutability 352
competitive ELISA 113 114 254 331
confidentiality of test results 371
confirmatory tests 118
conjugate pads 177 178
contamination 133 333
cows’ milk allergies 37 219
allergens 221 223

418
Index terms Links
cows’ milk allergies (Continued)
heat treatment of milk 223
hydrolyzed infant formulae 221
ovine and bovine milk cross-reactions 224
prevalence of 219
protein content of cows’ milk 221
protein denaturation 223
residue detection see dairy and egg residues
threshold doses 220
cross-reactions 14 72 76 102
339
crustaceans 282
latex-fruit syndrome 83
ovine and bovine milk 224
crustaceans 281
allergens 37 282
cross-reactivity 282
ELISA testing 282
prevalence of allergies 281
and proteomics 152
cupin superfamily 21 23
legumins 25
vicilins 24
cysteine proteases 30
D
dairy and egg residues 219
bioactivity of allergens 228
biosensors 235
blotting transfer 231
commercial tests 235
diffusion-in-gel detection 228

419
Index terms Links
dairy and egg residues (Continued)
dot blot detection 229
EAST inhibition 228
egg allergies 38 219 225
electrophoretic detection 230
ELISA detection 232
milk allergies 37 219
PCR detection 233
prevalence of allergies 219 225
quality assurance 236
quantitative tests 235
RAST inhibition 228
surface plasmon immunoassay (SPR) 235
two-dimensional electrophoresis 231
damage awards 369 376 392
de minimis concept 359
defective products 392 397
delayed hypersensitivity reactions 4 7 10
detection limits 195
differential dye absorption 146
diffusion-in-gel detection 228
dipstick detection 120 344
peanuts 191
DNA
extraction 126
low DNA foods 139
purification 126
dot blot detection 229
E
EAST inhibition 228
edible oils 13

420
Index terms Links
effectiveness of allergen detection 330
accuracy of known samples 342
assay factors 331
color development 344
contamination of test components 333
dipsticks 344
fast assays 344
inter-assay problems 344
intra-assay problems 343
optical density (OD) 335 341
and pipetting 334 335
poor precision 343
random errors 331
sample extraction procedures 337
spiked samples 342
standard curve accuracy 341
systematic errors 331
test characteristics 335
test performance 332
time of tests 332
egg allergies 38 219 225
allergens 225 227
bird-egg syndrome 227
heat treatment of egg proteins 227
prevalence 219 225
residue detection see dairy and egg residues
respiratory symptoms 227
yolk protein 227
electroimmunodiffusion 206
electrophoresis
aragose gel 128 129
dairy and egg residue detection 230

421
Index terms Links
electrophoresis (Continued)
SDS-PAGE 98 144 259
two-dimensional 98 145
ELISA (enzyme-linked immunosorbent assay) 68 72 109
advantages 115
cloth-ELISA 190
compared to PCR analysis 138
competitive ELISA 113 114 254 331
components of 114
confirmatory tests 118
crustacean tests 282
disadvantages 115
and heat processed foods 255
hypoallergenic infant formulas (HIF) 299
nut and seed tests 214
PCR-ELISA 129 135
peanut tests 139 188 192
sandwich ELISA 112 114 331
soy detection 276
test times 114 176
wheat gluten tests 250
see also effectiveness of allergen detection
engineered antibodies 72
environmental testing 309
epitopes 68
equipment
cleaning 308
shared equipment 16
European Union regulations 378
on food safety 388 395
on labelling 267 379 395
on product liability 390 395

422
Index terms Links
European Union regulations (Continued)
on product safety 386 395
exercise-associated food allergies 7 9
extraction procedures 337
F
fast assays 344
fish 283
allergens 14 37 284
collagen 14
detection methods 285
gelatin 14
monospecificity of allergies 284
fluorescent dyes 146 149
food allergies, definition 4
food idiosyncrasies 5
food intolerances 4 5
food processing see processed foods
food safety 388 395
definition of food 389
product recalls 357 364 388 395
396
responsibilities of business operators 389 390
traceability requirements 389 397
unsafe foods 389
see also product safety
food sensitivity 4
G
Gee, Samuel 245
gelatin 14

423
Index terms Links
genetically modified crops 153
gliadins 247 340 343 345
globulins 247
glutelins 247
gluten-free food standards 248 260 267
see also wheat gluten
glycinin 26 277
glycoside hydrolases 40
grains see cereals; wheat gluten
H
hazard characterisation 317
hazard identification 316
hazelnuts
in chocolate products 206
detection 206
proteomics 150
PCR analysis 133 210
surface plasmon immunoassay (SPR) 166 168 169
threshold levels 203
heat treatments
and egg proteins 227
and impaired extraction efficiency 255
and milk 223
and nut and seed allergenicity 204
and peanuts 339
and wheat gluten 250
Hevein-like chitinases 33
histamine 5
hydrogen bonding 67
hydrolysates 14 293 294
hydrolyzed infant formulae 221

424
Index terms Links
hydrophobic interactions 67
biochemical characterization 305
biochemical identification 306
definitions 294
ELISA tests 299
anti-food allergen antibody specificity 300
anti-food protein antisera 299
assay operating conditions 304
data handling 304
sample integrity 303
standards and control samples 303
equipment cleaning 308
hydrolysate manufacture 293
ingredient acceptance testing 307
ingredient identification 295 306
manufacturing controls 297 310
non-immunologic testing 305
non-protein contaminant measurements 298
production environment 308
release testing 311
residual allergen content measurement 297
sampling techniques 298
sensitivity targets 295 303
training in allergen awareness 307
I
IEF power supplies 145
IFA (incomplete Freund’s adjuvant) 70
IgA antibodies 9
IgE antibodies 8 9 79 88

425
Index terms Links
IgE antibody-based in vitro assays 79 85
allergosorbents 87
calibration 89
data analysis 90
design of 89
quality control 91
RAST (radioallergosorbent test) 85 88 91
reagents 86
standards and controls 89
IgE antibody-based in vivo assays 79 80
analytical sensitivity 83
analytical specificity 83
applicability 84
atopic human subjects 82
calibration strategy 82
data analysis 84
design of 81
dynamic range 84
limitations 84
performance characteristics 83
quality control 83
reagents 80
reference materials 81
relative potency of extract 81
skin test extracts 80
variability 84
IgE-binding proteins 154
IgE-immunoblotting 188
IgE-mediated reactions 4 7 10
IgG antibodies 65 71
IgM antibodies 71
image analysis programs 146

426
Index terms Links
immediate hypersensitivity reactions 4 7
treatment 10
immunoblotting 98
immunogens 68 68
administration 70
choice of 68
doses 70
mono-specific antibody reagents 101
polyclonal antisera 100
tolerance 72
immunoglobin structural model 65 110
see also antibodies
immunoglobulins 41
imprecision in testing methods 331
ingredient lists 380 399
allergenic ingredients 383
exemptions from 359 362
ingredients
acceptance testing 307
added directly to foods 362
allergenicity 13
identification 295 306
inhibition assay 179
innovation 319
instructions for use 382 393
inter-assay problems 344
intra-assay problems 343
K
Kazal-type serine protease inhibitors 42
kiwi fruit 31
Kunitz-type protease inhibitors 36 278

427
Index terms Links
L
labelling 357 370 379 395
of allergenic ingredients 267 382 398 401
best before dates 381 400
definitions 379
foods without pre-packaging 382
information required 380
ingredient lists 380 399
allergenic ingredients 383
exemptions from 359 362
ingredients added directly to foods 362
instructions for use 382 393
language 382
misbranding regulations 358
misleading information 380
name of product 380
net quantity 381 399
premarket notification 359
quantitative ingredients declaration (QUID) 381 399
lactalbumin 40 221
lactoglobulin assay 166
lactose intolerance 5
lateral flow devices 175
conjugate pad 177 178
control line 178
inhibition assay 179
membrane 177
peanut residue detection 180 191
reservoir 178
sample analysis 178
sample filter 177
test line 178

428
Index terms Links
latex-fruit syndrome 83
lawsuits see product liability
lecithin 14 154
lectins 36
legal framework see regulations and standards
legumins 25
lipid transfer proteins (LTPs) 28
lipocalins 40
liquid chromatography (LC) separation 148 194
lymphocytes 10 73
lysozyme 40 227
M
McDonald’s 367
MALDI analysis 147 263
manufacturing controls 297 310 319
marker compounds 275 322 348
mass spectrometry analyses 147 194 263
matrix interference 326
metabolic food disorders 5
method validation 350 352
microfluidic devices 155
milk allergies see cows’ milk allergies
minimum eliciting doses 11
misbranding regulations 358
misleading information 380
mono-specific antibody reagents 101
monoclonal antibodies 67 73 75 257
mustard detection 212

429
Index terms Links
N
negligence actions 365
net quantity 381 399
NOAEL (no observed adverse effect levels) 317 326 327
non-specific lipid transfer proteins 28
nut and seed allergies 201
allergenic proteins 204
almond detection 210
Brazil nut detection 211
cashew nut detection 211
ELISA detection methods 214
food processing effects 204 339
hazelnut detection 206
heat treatment and allergenicity 204
mustard detection 212
prevalence 202
sesame seed detection 212
threshold levels 203
walnut nut detection 211
see also peanuts
O
oats 247 262
one-step tests 176
optical density (OD) 335 341
oral allergy syndrome (OAS) 6 35 150
organoleptic techniques 109
osteopenia 246
ovalbumin 43 227
ovine and bovine milk cross-reactions 224
ovomucoids 42 166 225

430
Index terms Links
P
packaging 319
papain superfamily 21 30
parvalbumins 38 284
pathogenesis-related proteins 21 31 102
PCR (polymerase chain reaction) analysis 125
annealing process 126 132
aragose gel electrophoresis 128 129
celery detection 133
cereal detection 133
commercial assays 137
compared to ELISA 138
and contamination 133
DNA extraction 126
DNA purification 126
hazelnut detection 133 210
and low DNA foods 139
peanut detection 134 194
primer choice 132
real-time PCR 129
soy detection 133 137 276
temperature of reaction 126
wheat gluten detection 262
PCR-ELISA 129 135
peaches 28
peanuts 185
detection limits 195
dipstick detection 191
ELISA detection 139 188 192
and heat treatment 339
and IgE-immunoblotting 188

431
Index terms Links
peanuts (Continued)
lateral flow detection 180 191
liquid chromatography/tandem mass spectrometry
detection 194
non-immunological detection methods 194
PCR (polymerase chain reaction) analysis 134 194
prevalence of peanut allergy 185
protein content 25 26 186
radioallergosorbent assay (RAST) inhibition 187 188
rocket immunoelectrophoresis (RIE) detection 186
threshold levels 274 275
peptidases C1A 30
peptides 151 155
peroxidases 34
pipetting 334 335
plant food allergens 23
cereal superfamily 29
Kunitz-type protease inhibitors 36 278
lectins 36
pathogenesis-related proteins 21 31 102
profilins 35
prolamin superfamily 21 26
storage proteins 21
pollens 15
polyclonal antibodies 67 75 257
polyclonal antisera 100
polypeptides 151
premarket ingredient notification 359

432
Index terms Links
crustaceans 281
dairy and egg 219 225
fish 283
nut and seed 202
peanut 185
primer choice 132
processed foods 104 315 339
advertising and promotion 319
detection of allergenic residues 320
assay development 321 324
limits of detection 323
need for 321
hazard characterisation 317
hazard identification 316
and innovation 319
manufacturing protocols 297 310 319
NOAEL (no observed adverse effect levels) 317 326 327
and nut and seed residues 204 339
packaging 319
and protein changes 326
and retailers 319
risk characterisation 317
risk management 318
supply chain 319
see also food safety; heat treatments; product safety
product liability 357 364 390 395
adequate warning responsibilities 366
burden of proof 365
causal relationship 393 397
confidentiality of test results 371
damage awards 369 376 392
defective products 392 397

433
Index terms Links
product liability (Continued)
defences to 394 398
expiry period 395 398
limitation period 394 398
negligence actions 365
optional possibilities 391
producer liability 391
restaurant responsibilities 369 370
settlements out of court 364
strict liability 366
use of test kits 371
product recalls 357 364 388 395
396
product safety 316 386 395
see also food safety
production of antibodies 66 68
adjuvants 69
immunogen administration 70
immunogen choice 68
immunogen doses 70
monoclonal antibodies 73 75
monoclonal/polyclonal choice 75
purification 74
serum samples 71
profilins 35
prolamin superfamily 21 26 247 339
340
cereal prolamins 26 29
non-specific lipid transfer proteins 28
2S albumins 27
PROTALL database 44
protein denaturation 223

434
Index terms Links
proteins 21 43
cereal proteins 247
changes during processing 326
in cows’ milk 221
hydrolysates 14 293 294
lipid transfer proteins (LTPs) 28
pathogenesis-related 21 31 102
in peanuts 25 26 186
prolamin superfamily 21 26 247 339
340
residue detection 11
in wheat gluten 244 246
proteomics 144
bird’s nest allergen detection 152
detection and identification methods 146
differential dye absorption 146
and genetically modified crops 153
hazelnut allergen detection 150
IEF power supplies 145
and IgE-binding proteins 154
image analysis programs 146
lecithin detection 154
mass spectrometry analyses 147
quantitation 145
separation methods 144 145
sesame seed allergen detection 151
shrimp allergen detection 152
soy allergen detection 153
targeted proteome method 149
technologies 144
wheat allergen detection 149

435
Index terms Links
punitive damages 369
see also damage awards
purification 74
Q
IgE antibody-based in vitro assays 91
IgE antibody-based in vivo assays 83
quality control materials (QCM) 351 353
see also effectiveness of allergen detection
quantitative ingredients declaration (QUID) 381 399
quantitative tests 145 235
R
random errors 331
rapid screening tests 266
RAST (radioallergosorbent test) 85 88 91 187
188
strengths 94
weaknesses 95
see also IgE antibody-based in vitro assays
reaginic factor 7
real-time PCR 129
recognition of antigens 68
reference materials 81 351
reference matrix 355
regeneration solutions 161
regulations and standards 89 303 326
gluten-free foods 248 267

436
Index terms Links
regulations and standards (Continued)
see also European Union regulations; United States
regulations
respiratory symptoms 227
restaurant responsibilities 369 370
retailers 319
risk characterisation 317
risk management 318
rocket immunoelectrophoresis (RIE) detection 186
Rosaceae fruits 28
S
SAFE EU project 44
samples
accuracy of known samples 342
extraction procedures 160 166 298 337
integrity 303
serum samples 71
spiking 163 342
sandwich ELISA 112 114 331
sandwich surface plasmon immunoassay (SPR) 162 168
SCOP database 45
SDS-PAGE 98 144 259
seafood see crustaceans; fish
second antibodies 68
seed allergies see nut and seed allergies
sensitivity targets see threshold levels
separation methods 144 145
serpins 42
serum albumins 41
serum samples 71

437
Index terms Links
sesame seed detection 212
proteomics 151
shared equipment 16
shellfish see crustaceans
silver staining 146
skin test extracts 80
soy 273 285
allergens 11 13
beta-conglycinin 277
detection of soy proteins 275
ELISA testing 276
glycinin 26 277
health benefits 273
non-protein markers 275
PCR analysis 133 137 276
and proteomics 153
solubility of soy proteins 274
trypsin inhibitors 278
spiked samples 163 342
standard curve accuracy 341
standardization of tests 350
standards see regulations and standards
storage proteins 21
strict liability 366
supply chain 319
surface plasmon resonance (SPR) 158
advantages 164 172
antibody immobilization 161
binding allergens 161
non-specific binding 167

438
Index terms Links
surface plasmon resonance (SPR) (Continued)
calibration 162
and dairy and egg residues 235
egg protein assay 170
food product analysis 163
hazelnut assay 166 168 169
instrumentation 159 165
lactoglobulin assay 166
ovomucoid assay 166
peanut assay 170
published assays 164
recovery studies 163
regeneration solutions 161
sample extraction and analysis 160 166
sandwich assay 162 168
sensor surface preparation 161 165
sesame seed assay 170
spiking 163
tropomyosin assay 172
Swiss Model 45
symptoms 6
systematic errors 331
T
T-cells 9 68 248 265
T lymphocytes 10
tandem mass spectrometers 147 194
targeted proteome method 149
test characteristics 335
test performance 332
test times 114 176 332
thaumatin-like proteins (TLPs) 33

439
Index terms Links
cows’ milk allergens 220
hazelnuts 203
hypoallergenic infant formulas (HIF) 295 303
nut and seed allergens 203
peanuts 274 275
soy 274 275
wheat gluten 249
time of tests 114 176 332
time-of-flight (TOF) analysis 147
traceability of foods 389 397
training in allergen awareness 307
transferrins 42
tree nuts see nut and seed allergies
tropomyosins 14 37 39 172
282 283
trypsin inhibitors 278
tuna allergy 285
turnips 33
2S albumins 27
two-dimensional electrophoresis 98 145
U
Unilever 319
United States regulations 357
adulteration provisions 363
de minimis concept 359
misbranding provisions 358
product recalls 357 364
unsafe foods 389

440
Index terms Links
V
validated methods 350 352
vicilins 24
W
walnut detection 211
Western blotting see immunoblotting
wheat gluten 26 244 340
choice of detection method 264
and coeliac disease 10 30 244 247
265
electrophoretic separations 259
ELISA detection 250
gluten-free food standards 248 260
and heat processing 250
Maldi-TOF detection 263
PCR detection 262
proteins in 244 246
and proteomics 149
rapid screening tests 266
threshold values 249
Win-like chitinases 33
Y
yolk protein 227

2006-Detecting Allergens in Food - ISBN-10 0-8493-2574-9

Uploaded by

Copyright:

Available Formats

2006-Detecting Allergens in Food - ISBN-10 0-8493-2574-9

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

2006-Detecting Allergens in Food - ISBN-10 0-8493-2574-9

Uploaded by

Copyright:

Available Formats

Detecting allergens in food

WOODHEAD PUBLISHING LIMITED

Trademark notice: Product or corporate names may be trademarks or registered

British Library Cataloguing in Publication Data

Library of Congress Cataloging in Publication Data

Woodhead Publishing ISBN-13: 978-1-85573-728-0 (book)

Typeset in India by Replika Press Pvt Ltd.

E-mail: staylor2@unl.edu E-mail: shefle1@unl.edu

E-mail: mzeece@unlnotes.unl.edu E-mail: shefle1@unl.edu

E-mail: stefkoppelman@zonnet.nl E-mail: stefkoppelman@zonnet.nl

Professor Sue L. Hefle

Dr Virginie Leduc Chapter 16

Chapter 14 E-mail: chris.cordle@abbott.com

E-mail: u.immer@r-biopharm.de E-mail: heereluurt.heeres@tno.nl

E-mail: roland.poms@icc.or.at E-mail: stefkoppelman@zonnet.nl

Dr Hendrik Emons and Dr Elke Professor Sue L. Hefle

on either immunochemical or DNA techniques. How specific and sensitive

Stef Koppelman and Sue Hefle

Contributor Contact Details .................................................... xi

Part I. The Basics of Food Allergy

Part II. Types of Detection Method

This page has been reformatted by Knovel to provide easier navigation. v

3.2 Immunogens and Antigens ................................... 68

This page has been reformatted by Knovel to provide easier navigation.

7. Polymerase Chain Reaction (PCR) Methods for the

This page has been reformatted by Knovel to provide easier navigation.

10.2 Antibodies ............................................................. 176

Part III. Detection Methods for Particular Allergens

This page has been reformatted by Knovel to provide easier navigation.

13.6 Acknowledgements ............................................... 236

Part IV. Issues in Allergen Detection Methods

This page has been reformatted by Knovel to provide easier navigation.

17.2 Food Allergy and Product Safety ........................... 316

This page has been reformatted by Knovel to provide easier navigation.

20.3 Legal Grounds for Product Liability

This page has been reformatted by Knovel to provide easier navigation.

22.3 Reference Materials, Extraction and

Index ...................................................................................... 413

This page has been reformatted by Knovel to provide easier navigation.

The basics of food allergy

The nature of food allergy

1.1 Introduction: defining food allergy

1.1.1 Definition and classification

Table 1.1 A classification scheme for food allergies and intolerances

Food allergy IgE-mediated food allergy (immediate hypersensitivity)

symptoms appearing on a more delayed basis, as much as 24 or more hours

1.1.2 Prevalence and health impacts

Table 1.2 Symptoms of IgE-mediated food allergies

Gastrointestinal Nausea, vomiting, diarrhea, abdominal cramps

allergies affect an estimated 1.1% of the US population, including many

found in the fresh fruits and vegetables.18 Apparently, sensitization to these

1.2 Mechanisms of food allergy

1.2.1 IgE-mediated allergic reaction (immediate hypersensitivity)

demonstrated that this reaginic activity was associated with a unique

Fig. 1.1 Mechanism of IgE-mediated allergic reaction.

insufficient evidence for the diagnosis of a food allergy unless it is coupled

1.2.2 Exercise-associated food allergies

1.2.3 Cell-mediated reactions (delayed hypersensitivity)

1.3 Avoidance diets and treatment for IgE-mediated