Dr. G.S Shephard

Determination of fumonisins in maize by High Performance Liquid

Chromatography with fluorescence and ultraviolet detection of o-

phthaldialdehyde, naphthalene-2,3-dicarboxaldehyde and dansyl chloride

derivatives

by
Ncediwe Ndube
A thesis submitted in partial fulfilment of the requirements for the degree
Magister Scientiae in the Department of Chemistry, University of the
Western Cape
Supervisors: Professor I.R Green
Dr. G.S Shephard
Co-supervisor: Dr. L. van der Westhuizen
May 2011
Abstract

Determination of fumonisins in maize by High Performance Liquid

derivatives
N. Ndube
MSc Chemistry Thesis, Department of Chemistry, University of the Western Cape
Keywords: Fumonisins; fluorescence detector; ultraviolet detector; o-
phthaldialdehyde; naphthalene-2,3-dicarboxaldehyde; dansyl chloride; high-
performance liquid chromatography; strong anion extraction; immunoaffinity;
diode array detector.
Fumonisins, carcinogenic mycotoxins produced by various Fusarium
species, occur naturally in maize and maize-based food products. They are
hazards for animal and human health as they cause cancer in rodents and have
been associated with oesophageal cancer and neural tube defects in humans.
The most abundant naturally occurring fumonisins analogues in maize are
ii
fumonisin B1, B2 and B3 (FB1, FB2 and FB3). For analytical determination, they

mostly require suitable extraction, clean-up and pre or post-column

derivatization together with reversed-phase HPLC separation. o-

Phthaldialdehyde (OPA) had been adopted as the most widely used
derivatization reagent for fumonisins as they lack useful chromophores or
fluorophores. Alternative derivatization reagents, naphthalene-2,3-
dicarboxaldehyde (NDA) and dansyl chloride (DnS-Cl), were investigated in this
study. The HPLC system used was equipped with diode array (DAD) and
fluorescence detectors to determine UV detection as an alternative following
derivatization with OPA, NDA and DnS-Cl. Optimization of the NDA derivatives
with working standards resulted in limits of detection (LOD) of FB1, FB2 and FB3
with FLD of 0.11 ng, 0.50 ng and 0.27 ng, respectively, and with DAD 13.8 ng,
12.5 ng and 6.6 ng, respectively. Subsequently naturally contaminated maize
samples, collected from subsistence farmers in the Eastern Cape, were cleaned-
up with strong anion exchange (SAX) solid phase extraction (SPE) cartridges. The
coefficient of variation (CV) for FB1, FB2 and FB3 in maize samples (n=6) were
2.6%, 1.8% and 5.3%, respectively, with FLD compared to 6.0%, 3.4% and 9.5%,
respectively, with DAD. Subsequently the NDA derivatization was compared to
the OPA derivatization, as well as an alternative sample clean-up with
immunoaffinity column (IAC) by analyzing naturally contaminated maize samples
(n = 15) ranging in total fumonisin (TFB = FB1 + FB2 + FB3) levels from 106 - 6000
μg/kg. After IAC clean-up of extracted samples, the recoveries for NDA-FLD of
FB1, FB2 and FB3 were 62%, 94% and 64%, respectively. NDA proved to be an
iii
effective derivatization reagent for fumonisin in naturally contaminated maize

samples following IAC clean-up, except for DAD at TFB levels below 1000 μg/kg.

In contrast NDA derivatization following SAX clean-up produced results

comparable to OPA only for levels below

1000 μg/kg. FLD and DAD produced
comparable results irrespective of the clean-up method or the derivatization
agent. The investigation of DnS-Cl as a derivatization reagent resulted in LOD for
FB1, FB2 and FB3 of 4.3 ng, 3.9 ng and 2.1 ng, respectively, with FLD and 17.2 ng,
15.6 ng, 15.6 ng, respectively, with DAD. Although sensitive and reproducible
derivatives were formed with fumonisin working standards, matrix interferences
from maize samples with DnS-Cl rendered this derivatization reagent unsuitable
for fumonisin analysis in naturally contaminated maize. In conclusion this study
has shown that UV detection can be utilized as an alternative to FLD for
fumonisin analysis in naturally contaminated maize irrespective of the clean-up
method or the derivatization agent.
13 May 2011
iv
DECLARATION

I declare that ‘Determination of fumonisins in maize by High Performance Liquid

derivatives’ is my own work, that it has not been submitted for any degree or
examination in any other university, and that all the sources I have used or
quoted have been indicated and acknowledged by complete references.
Full name............................................................. Date..................................................
Signed...................................................................
v
ACKNOWELEDGEMENTS

To my supervisors Dr Gordon Shephard

and Dr Liana van der Westhuizen, thank

you for affording me the opportunity to further my studies, your support,

encouragement and your guidance has made it possible for me to complete this
project. Prof Ivan R Green, thank you for the opportunity to work with you, your
advice and valuable contributions to the study are much appreciated.
To my mother Teboho Julia Ndube, thanks mom for your encouragement, your
support and love through-out my studies, you my rock, love you. My sister and
brother, Nonceba and Mcebisi Ndube thank you for your support and
understanding, Asamahle and Chumani, your aunt loves you dearly. My family,
your support has kept me going; life would be difficult without you.
To my fiancé, Thami Tsolekile, thank you for your patience, constant
encouragement, your support and faith in me. To the Tsolekile family, my friends
and relatives thank you for your encouragement. My colleagues at the PROMEC
Unit, MRC thank you for your support.
The National Research Foundation for funding the study.
Lastly, I want to thank God for all the blessings he has bestowed on me. I am who
I am because I know the Lord “Know the Lord, He is God, it is He who has made
us, and not we ourselves; we are His people and the sheep of His pasture- Psalm
100: 3”.
vi
TABLE OF CONTENTS

Abstract ii
Declaration v
Acknowledgements vi
Table of Contents vii
List of Tables ix
List of Figures xi
Abbreviations xv
Contributions of thesis xvi
CHAPTER 1: Introduction
1.1 Introduction 2
1.2 Aims and objectives 5
1.3 Research approach 6
1.4 Research structure 7
References 10
CHAPTER 2: Literature review
2.1 Introduction 17
2.1.1 Background to fumonisins 17
2.1.2 Occurence of fumonisins 18
2.1.3 Impact of fumonisins 20
2.2 Analysis of fumonisins 22
2.2.1 Introduction 22
2.2.2 Extraction of samples 23
2.2.3 Clean-up 25
2.2.4 Derivatization 26
2.3 Detection of fumonisins 35
2.3.1 Introduction 35
2.3.2 Chromatographic methods 35
2.3.3 Immunological methods 37
2.3.4 Mass spectrometry 38
2.3.5 Recent developments in fumonisin detection 38
2.4 Overview of literature review 39
References 41
CHAPTER 3: Ultraviolet (UV) detection of fumonisin B analogues as

OPA derivatives
3.1 Introduction 59
vii
3.2 Material and methods 59
3.3 Chromatography 63
3.4 Results and Discussion 63
3.5 Method Application 72
3.6 Conclusion 77

References 78

CHAPTER 4: Optimization of naphthalene-2,3-dicarboxaldehyde
(NDA) reagent for fumonisin derivatization and its applicability to
fluorescence (FLD) and ultraviolet (UV) detection
4.1 Introduction 81
4.5 Robustness 93
4.6 Recoveries 103
4.7 Method Application 105
4.8 Conclusion 110
References 111
CHAPTER 5 : An evaluation of dansyl chloride (DnS-Cl) for fumonisin
derivatization analysed by HPLC with fluorescence (FLD) and
ultraviolet (UV) detection
5.1 Introduction 116

5.5 Robustness 125
5.6 Recoveries 130
5.7 Optimization of recoveries using naturally 133
contaminated samples
5.8 Conclusion 139
References 140
CHAPTER 6: Comparisons of methods, General Discussion,

Recommendations and Conclusion
6.1 Comparison of OPA and NDA 144
6.2 General discussion 149
6.3 Recommendations 152
6.4 Conclusion 152
References 154
viii
LIST OF TABLES

Chapter 3

Table 3.1 Intra-day precision of fumonisin working standards 68
(n=3) using standard peak
areas
Table 3.2
Inter-day precision of fumonisin working standards 69
using standard peak areas, (n=18, 3 x std injected/day)
Table 3.3 Determination of the detection limits using the amount 70

(ng) injected into the column
Table 3.4 Recoveries with IAC clean-up 72
Table 3.5 Fumonisin levels (µg/kg) in naturally contaminated 73

maize samples cleaned-up with SAX
Table 3.6 Fumonisin levels (µg/kg) in naturally contaminated 74

maize samples cleaned-up with IAC
Chapter 4
Table 4.1 Intra-day precision of fumonisin working standards 88

(n=3) reported as peak areas
Table 4.2 Inter-day precision of fumonisin working standards 89

(n=12) reported as peak areas
Table 4.3 Amount (ng) injected into HPLC Column 90
Table 4.4 Limits of detection (LOD) and quantification (LOQ) 90

expressed as levels in sample (µg/kg) following IAC
clean-up.
Table 4.5 Stability of FB-NDA standard after six consecutive 92

injections (~ 120 min)
Table 4.6 Solvent extraction efficiency for NDA derivatization of 102

maize samples (20 g / 100 mL)
Table 4.7 Fumonisin recoveries (%) from maize samples cleaned 104
up with SAX
Table 4.8 Fumonisin recoveries (%) from maize samples cleaned- 104
up with IAC
ix
Table 4.9 Comparison of HPLC-FLD and DAD following SAX clean- 106
up (µg/kg)
Table 4.10 Comparison of HPLCFLD

and DAD following IAC clean- 107
up (µg/kg)

Chapter 5
Table 5.1 Intra-day precision using peak areas of working 122
standards (n=3)
Table 5.2 Inter-day precision using peak areas of working 123

standards injected (n=12)
Table 5.3 Limits of detection (LOD) and quantitation (LOQ) in 124

terms of amount injected (ng) onto HPLC column
Table 5.4 Effect of wavelength on working standard peak areas 127

for FLD
Table 5.5 Effect of wavelength on working standard peak areas 127

for DAD
Table 5.6 Determining the recovery of the maize extraction using 131
IAC
Table 5.7 Determination of the recovery of the derivatization 131

procedure
Table 5.8 Effect of reaction temperature on maize derivatized for 135

15 minutes with DnS-Cl
Table 5.9 Effect of reaction time on maize derivatized at 60°C 135

with DnS-Cl
Chapter 6
Table 6.1 Total fumonisin levels (FB1+FB2+FB3; µg/kg) in naturally 146

contaminated maize cleaned-up with SAX and IAC,
derivatized with OPA or NDA
Table 6.2 Chromatographic parameters for determination of 149

fumonisins as OPA, NDA and DnS-Cl derivatives
x
LIST OF FIGURES

Chapter 2

Figure 2.1 Schematic diagram of the structures of Fumonisin B1 18
(FB1), Fumonisin B2 (FB 2) and Fumonisin B3 (FB3)
Figure 2.2 Maize infected with fumonisin producing fungi, F. 19

Verticillioides
Figure 2.3 Structure of o-phthaldialdehyde (OPA) 27
Figure 2.4 Reaction mechanism for the formation of FB-OPA 29

complex
Figure 2.5 Structure of Napthalene-2, 3-dicarboxaldehyde (NDA) 30
Figure 2.6 Reaction of NDA with FB1 to form 1- 31

Cyanobenzisoindole (CBI) stable derivative
Figure 2.7 Structure of Dansyl Chloride (DnS-Cl) 33
Chapter 3
Figure 3.1 Chromatogram of fumonisin working standard detected 66

by FLD

by DAD
Figure 3.3 Comparison of FLD and DAD using SAX celan-up, Total 75
Fumonsins = FB1 + FB2 + FB3
Figure 3.4 Comparison of FLD and DAD using IAC celan-up, Total 76
Fumonsins = FB1 + FB2 + FB3
Chapter 4
Figure 4.1 Chromatogram of fumonisins working standard 86
detected by FLD

by DAD
Figure 4.3 Stability of FB-NDA derivatives (µg/kg) over period of 92

three days (72 hours). Results calculated from mean
xi
(n=6) ± standard deviation. (A) Stability of FLD, (B)
stability of DAD
Figure 4.4 Effect of methanol concentration

on retention time and 93
interference with reagent peaks

Figure 4.5 Optimum wavelength selections for UV absorbance 95
relative to peak area, results calculated from mean
(n=3)
Figure 4.6 Effect of reaction time on peak areas. Results reported 96

as mean (n=5) ± standard deviation. (A) Comparison of
FLD, (B) Comparison of DAD
Figure 4.7 Effect of reaction temperature on peak areas. Results 97

reported as mean (n=5) ± standard deviation for FLD
Figure 4.8 Effect of reaction temperature on peak areas. Results 98

reported as mean (n=5) ± standard deviation for DAD
Figure 4.9 Effect of buffer concentration on peak areas for mean 99
(n=5) ± standard deviation for FLD
Figure 4.10 Effect of buffer concentration on peak areas for mean 99

(n=5) ± standard deviation for DAD
Figure 4.11 Comparison of FLD and DAD following SAX clean-up, 108
Total Fumonisins = FB1 + FB2 + FB3
Figure 4.12 Comparison of FLD and DAD following IAC clean-up, 109
Chapter 5
Figure 5.1 Chromatogram of combined fumonisins working 120

standard detected by FLD
Figure 5.2 Chromatogram of combined fumonisins working 121

standard detected by DAD
Figure 5.3 Comparison of acetonitrile and acetone as DnS-Cl 126

reagent solvents. Results reported as mean (n=6) ±
standard deviation. (A) Comparison of FLD, (B)
xii
Comparison of DAD

Figure 5.4 Effect of reaction temperature on peak areas of FLD, 128
results calculated on mean
(n=5) ± standard deviation
Figure 5.5
Effect of reaction temperature on peak areas of DAD, 128
results calculated on mean

(n=5) ± standard deviation
Figure 5.6 Effect of reaction time at 40 °C on peak areas for FLD. 129
Results reported as mean (n=5) ± standard deviation
Figure 5.7 Effect of reaction time at 40 °C on peak areas for DAD. 130
Results reported as mean (n=5) ± standard deviation
Figure 5.8 Purity peak check results 132
Figure 5.9 Chromatogram of naturally contaminated maize 133

sample with DAD
Figure 5.10 Effect of reagent volume (DnS-Cl) on reaction yield, 136

results reported on mean (n=6) ± standard deviation
Figure 5.11 Naturally contaminated maize sample with DAD 137
Figure 5.12 Fumonisin working standard derivatized with DnS-Cl 138

and Na2CO3 as buffer
Chapter 6
Figure 6.1 Comparison of OPA with NDA following SAX clean-up 147
for FLD and DAD, Total Fumonisins = FB1 + FB2 + FB3
Figure 6.2 Comparison of OPA and NDA following IAC clean-up for 147
FLD and DAD, Total Fumonisins = FB1 + FB2 + FB3
Figure 6.3 Comparison of IAC and SAX with OPA derivatization for 148
Figure 6.4 Comparison of IAC and SAX with NDA derivatization for 148
xiii
ABBREVIATIONS

FB Fumonisins

FB1 Fumonisin B1

FB2 Fumonisin B2

FB3 Fumonisin B3
OPA o-phthaldialdehyde
NDA Naphthalene-2,3-dicarboxaldehyde
DnS-Cl Dansyl Chloride
SAX Strong anion exchange
IAC Immunoaffinity column
FBTest FumoniTest
HPLC High - performance liquid chromatography
FLD Fluorescence detector
DAD Diode array detector
UV Ultraviolet detection
RSD Relative standard deviation
CV Coefficient of variation
LOD Limit of detection
LOQ Limit of quantification
Std Standard
The language and style used in the thesis are in accordance with the
requirements of Mycotoxin Research, Journal of Chromatogrphy B and the
PROMEC Unit, MRC. Each chapter is an individual entity and some repetition
between chapters has, therefore, been unavoidable.
xiv
CONTRIBUTIONS OF THESIS

Data presented in this thesis have already been incorporated into the following

scientific papers and presentations:
Published: Ndube N, Van der Westhuizen

L and Shephard GS (2009)
Determination of fumonisins in maize with ultraviolet detection of
o-phthaldialdehyde. Mycotoxin Research 25: 225-228
In-Press: Ndube N, Van der Westhuizen L, Ivan RG and Shephard GS. HPLC
determination of fumonisin mycotoxins in maize: A comparative study
of naphthalene-2,3-dicarboxaldehyde and o-phthaldialdehyde
derivatization reagents for fluorescence and diode array detection.
(Journal of chromatography B)
Conference contributions from thesis:
1. Analitika Conference 2010 (Stellenbosch University, South Africa)
Poster: N. Ndube, L. van der Westhuizen, Green IR and G.S. Shephard. UV
detection as an alternative to fluorescence detection for HPLC
determination of fumonisins in maize.
2. Mycored International Conference 2011
(CTICC, Cape Town, South Africa)
Poster: N. Ndube, L. van der Westhuizen, Green IR and G.S. Shephard.
HPLC analysis of fumonisins in maize using ultraviolet detection
as an alternative to fluorescence detection.
xv

CHAPTER 1
Introduction
1.1 Introduction

Toxins are poisonous substances that are produced by living cells or organisms
(Medical Dictionary). They are synthesised

by plant species, animal or micro-
organisms and are generally harmful to a different organism. Mycotoxins are
toxic secondary metabolites produced by fungi growing on a range of cereal and
food matrixes and many are produced by plant pathogens (Turner et al., 2009).
Since the discovery of aflatoxins in 1960, mycotoxins have been found to be
responsible for a variety of human and animal diseases (Shephard, 2008).
Fumonisins are carcinogenic mycotoxins produced by fungi of the Fusarium
species, primarily by F. verticillioides and F. proliferatum and were first isolated
by Gelderblom et al., 1988. They are an economically important group of
mycotoxins that occur primarily in maize and maize-based products (Shephard et
al., 1996).
There are at least 28 chemical analogues of fumonisins, but the most important
are the fumonisin B’s (FBs) which occur naturally in contaminated maize. The
most important FBs are fumonisin B1 (FB1), fumonisin B2 (FB2) and fumonisin B3
(FB3), with FB1 being the most prevalent (Rheeder et al., 2002). Fumonisins have
been reported to show cancer promoting properties in rats (Gelderblom et al.,
1996). Fumonisins are not mutagenic (Gelderblom et al., 1991; Knasmuller et al.,
1997) nor genotoxic in primary rat hepatocytes (Norred et al., 1992) but FB1 is
hepatocarcinogenic in male BD IX rats (Gelderblom et al., 2001) and B6C3F 1
female mice and nephrocarcinogenic in male Fischer 344 rats (Howard et al.,
2
2001). Fumonisins are known to cause leukoencephalomalacia in horses and

pulmonary edema syndrome in pigs (Howard et al., 2001; Marasas, 2001).

Fumonisins have been associated with the high prevalence of human
former Transkei region in South Africa

oesophageal cancer in areas like the
(Rheeder et al., 1992) and Santa Catarina State, Brazil (Van der Westhuizen et al.,
2003), where high levels of fumonisin contaminated foods are part of the normal
diet (Rheeder et al., 1992; Sun et al., 2007; Wang et al., 2008). It has been
suggested that fumonisins are a risk for inducing liver cancer in humans (Ueno et
al., 1997). There has also been a proposed link between fumonisin exposure and
neural tube defects in humans (Missmer et al., 2006; Hendricks, 1999; Marasas
et al., 2004).
Based on information available at that time, the International Agency for
Research on Cancer classified FB1 as a possible human carcinogen (group 2B)
(IARC, 2002). The US Food and Drug Administration (FDA) have recommended
that cautionary levels be put in place to reduce human exposure to fumonisins
intended for human and animal consumption (FDA, 2001). In 2007, The European
Commission (EC) regulated fumonisin exposure at various levels from 4000 µg
fumonisins/kg for unprocessed maize, 1000 µg/kg for maize intended for direct
human consumption and 200 µg fumonisins/kg for processed maize-based foods
and baby foods (EC, 2007).
Due to the problems and risks associated with fumonisin contamination of
animal feed and human food, there is a growing need to develop reliable and
3
sensitive methods for the determination of fumonisins in maize and maize-based

foods (Shephard, 1998). Various different methods that have been reported for

the analysis of fumonisins include: thin-layer chromatography (TLC) (Gelderblom

et al., 1988; Shephard and Sewram

2004), liquid chromatography-mass
spectrometry (LC-MS) (Zollner et al., 2006), gas chromatography-mass
spectrometry (GC-MS) (Plattner et al., 1990) and liquid chromatography with
reversed-phase high performance liquid chromatography (RP-HPLC) being the
most widely used current method for fumonisin determination (Shephard et al.,
1996). Samples are extracted either with aqueous methanol or acetonitrile and
mostly cleaned-up on strong anion exchange (SAX) solid phase extraction (SPE)
cartridges or immunoaffinity columns (Sydenham et al., 1996). Fumonisins lack a
suitable chromophore that enables them to be detected by UV and consequently
need to be derivatized prior to HPLC separation (Shephard, 1998).
Different derivatization reagents have been reported for fumonisin analysis such
as maleyl (Sydenham et al., 1990), fluorescamine (Ross et al., 1991), 4-fluoro-7-
nitrobenzofurazan (Scott and Lawrence 1992), but o-phthaldialdehyde (OPA) still
remains the derivatization reagent of choice for most laboratories as it produces
highly fluorescent derivatives readily separated by HPLC (Shephard, 1998).
Detection of OPA fumonisin derivatives is mainly fluorescence-based at
excitation wavelength of 335 nm and emission wavelength of 420 nm (Shephard
1990). However, many laboratories requiring infrequent fumonisin analysis are
equipped only with HPLC with an ultraviolet detector (UV) (Ndube et al., 2009).
4
1.2 Aims and objectives

Fumonisin analysis has been based almost
exclusively on fluorescence detection
(FLD) and liquid chromatography mass

spectrometry (LC-MS). However, there
analytical method for fumonisin analysis

has been a growing need to develop an
by laboratories requiring occasional fumonisin analysis and equipped only with a
UV detector. Limited work has been done on the analysis of fumonisins by HPLC-
UV detection. Maleyl derivatives analysed by HPLC-UV gave detection limits of 10
µg/g which is inappropriate for naturally contaminated maize samples
(Sydenham et al., 1990) and fluorescent derivatives of 4-fluoro-7-
nitrobenzofurazan (NBDF) gave higher detection limits of 100 µg/g and also
showed limited stability (Scott and Lawrence 1994). OPA, naphthalene-2,3-
dicarboxaldehyde (NDA) and dansyl chloride (DnS-Cl), the reagents studied here,
have previously only been used as fluorogenic reagents.
The aim of this work was to investigate the degree to which fumonisins present
in maize can be determined by high-performance liquid chromatography (HPLC)
with both FLD and UV detection of their OPA, NDA and DnS-Cl derivatives. The
objective of the study was to determine the extent to which UV may be used as
an alternative to FLD as well as the extent to which NDA or DnS-Cl may be
employed as alternatives to the widely used OPA derivatization reagent. The
specific objectives of this study were:
5
1. To investigate the scope to which fumonisins in maize could be determined

by high-performance liquid chromatography (HPLC) with ultraviolet (UV) and

fluorescence (FLD) detection.

2.
To optimize the derivatization of fumonisins with NDA and DnS-Cl.
3. To determine and compare differences between FLD and DAD responses of
the different derivatization reagents.
4. To evaluate the applicability of the derivatization reagents for the analysis of
naturally contaminated maize samples using different clean-up methods.
1.3 Research approach
Derivatization of fumonisins is necessary as they lack a suitable chromophore or
fluorophore for ultraviolet and/or fluorescence detection. This is best achieved
by pre-column rather than a post-column on-line approach, as pre-column
derivatization requires less equipment and is much easier to handle.
Consequently the initial step of the study was to identify suitable derivatization
reagents which might offer a suitable system for UV detection. The well
characterized method using OPA was investigated first, where after the reagents
NDA and DnS-Cl were evaluated on a comparative basis. As these latter two
reagents have found limited use in fumonisin analysis, the derivatization protocol
and chromatographic separation were first optimized. RP-HPLC determination of
fumonisins mostly follows SAX or IAC clean-up of a suitable extract. Both clean-
up methods were used in these studies to test which clean-up method works
6
best for each derivatization reagent. The HPLC system was equipped with both

UV and fluorescence detectors connected in series, to determine the extent to

which UV offers an alternative to fluorescence detection of fumonisins and to

allow for direct comparison between the

two detectors.
1.4 Research structure
The current chapter includes the introductory overview on the research theme,
the aims and objectives, research approach and structure of the study. The
succeeding chapters will be as follows:
Chapter 2: Literature Review
The background, occurrence and impact of fumonisins will be
presented. This is followed by an in-depth description of the analysis of
fumonisins which includes extraction of maize samples, clean-up and
derivatization. All the derivatization reagents relevant to the study are
reviewed in detail including their mechanism or reactions to form
derivatives. The review concludes with a literature survey of HPLC
detection methods for fumonisins which are used in the study as well
as a discussion of other detection methods.
Chapter 3: Ultraviolet detection (UV) of fumonisin B analogues as o-
phthaldialdehyde derivatives
Chapter 3 investigates the degree to which the fumonisin method of
Sydenham et al., 1996 could be combined with UV detection of the FB-
OPA derivatives. The FLD and developed DAD method is applied to
7
naturally contaminated maize samples. Comparison of SAX with IAC

and FLD and DAD concludes the chapter.

Chapter 4: Optimization of naphthalene-2,3-dicarboxaldehyde

(NDA) derivatives
for fumonisin derivatization and its applicability to fluorescence (FLD)
and ultraviolet (UV) detection
This chapter discusses the extent to which the optimized NDA
derivatization method could be used for analysis of South African
subsistence maize samples. It aims to provide stable fumonisin
derivatives to allow for automated detection of fumonisins with HPLC.
Chapter 5: An evaluation of dansyl chloride (DnS-Cl) for fumonisin derivatization
analysed by HPLC with fluorescence (FLD) and ultraviolet (UV)
detection
Chapter 5 investigates the extent to which DnS-Cl can be used for
fumonisin derivatization in maize samples. The method is optimized to
obtain best conditions for pre-column derivatization and HPLC analysis
of the derivatives.
Chapter 6: Comparisons of methods, General Discussion, Recommendations and
Conclusion
The study is concluded with a concise discussion of the most significant
points from all the preceding chapters. A comparison of the
derivatization reagents are discussed and as well as how each method
8
has either advantages or not over the current. Finally, directions for

future work are suggested.

9
References

1. European Commission Regulation

(EC) No 1126/2007 of 28 September

2007. Official Journal of the European Union L255: 14-17
2. FDA, Food and Drug Administration (2001) Fumonisin levels in human
foods and animal feeds. Final guidance revised 9 November 2001
3. Gelderblom WCA, Jaskiewicz K, Marasas WFO, Thiel PG, Horak RM,
Vleggaar R and Krieg NP (1988) Fumonisins-Novel mycotoxins with cancer-
promoting activity produced by Fusarium moniliforme. Applied and
Environmental Microbiology 54: 1806-1811
4. Gelderblom WC, Kriek NP, Marasas WF and Thiel PG (1991) Toxicity and
carcinogenicity of the Fusarium moniliforme metabolite, fumonisin B1 in
rats. Carcinogenesis 12: 1247-1251
5. Gelderblom WCA, Synman SD, Abel S, Lebepe-Mazur S, Smuts CM, van der
Westerhuizen L, Marasas WFO, Voctor TC, Knasmuller S and Huber W
(1996) Hepatotoxicity and carcinogenicity of the fumonisins in rats: A
review regarding mechanistic implications for establishing risk in humans.
In Fumonisins in Food: Jackson LS, Devries JW and Bullerman LD, Eds. New
York: Plenum Press 24: 279-296
6. Gelderblom WCA, Abel S, Smuts CM, Marnewick J, Marasas WFO, Lemmer
ER and Ramljak D (2001) Fumonisin-induced hepatocarcinogenesis:
10
Mechanisms related to cancer initiation and promotion. Environmental

Health Percetives 109: 291-300

7. Hendricks K (1999) Fumonisins and neural tube defects in South Texas.

Epidemiology 10: 198-200
8. Howard PC, Eppley RM, Stack ME, Warbritton A, Voss KA, Lorentzen RJ,
Kovach RM and Bucci TJ (2001) Fumonisins B1 carcinogenicity in a two-year
feeding study using F344 rats and B6C3F1 mice. Environmental Health
Perspectives 109: 277-282
9. IARC, International Agency for Research on Cancer (2002) Fumonisin B1. In:
IARC monographs on the evaluation of the carcinogenic risks to humans:
Some traditional herbal medicines, some mycotoxins, naphthalene and
styrene. IARC, Lyon, France, 82: 301-366
10. Knasmuller S, Bresgen N, Kassie F, Mersch-Sundermann V, Gelderblom W,
Zohrer E and Eckl PM (1997) Genotoxic effects of the three Fusarium
mycotoxins, fumonisin B1, moniliforme and vomitoxin in bacteria and in
primary cultures of rat hepatocytes. Mutatation Research 391: 39-48
11. Marasas WFO (2001) Discovery and occurrence of the fumonisins: a
historical perspective. Environmental Health Perspectives 109: 239-243
12. Marasas WFO, Riley RT, Hendricks KA, Stevens VL, Sadler TW, Waes JG,
Missmer SA, Cabrera J, Torres O, Gelderblom WFO, Allegood J, Martinez C,
Maddox J, Miller JD, Sullards CM, Roman AV, Voss KA, Wang E and Merrill
AH (2004) Fumionisins Disrupt Sphingolipid metabolism, folate transport
11
and neural tube development in embryo culture and in vivo: a potential

risk factor for human neural tube defects amoung populations consuming

fumonisin-contaminated maize. American Society for Nutritional Sciences

134: 711-716
13. Medical Dictionary, Saunders WB; Dorland Illustrated: Medical Dictionary
25 Edition
14. Missmer SA, Suarez L, Felkner M, Wang E, Merrill AH Jr, Rothman KJ and
Hendricks KA (2006) Exposure to fumonisins and the occurrence of neural
tube defects along the Texas-Mexico border. Environmental Health
Perspective 114: 237-241
15. Ndube N, Van der Westhuizen L and Shephard GS (2009) Determination of
fumonisins in maize with ultraviolet detection of o-phthaldialdehyde.
Mycotoxin Research 25: 225-228
16. Norred WP, Platter RD, Vesonder RF, Bacon CW and Voss KA (1992) Effects
of selected secondary metabolites of Fusarium moniliforme on
unscheduled synthesis of DNA by rat primary hepatocytes. Food and
Chemical Toxicology 30: 233-237
17. Plattner RD, Ross PF, Stedelin J and Rice LG (1990) Analysis of corn and
cultural corn for fumonisins B1 by HPLC and GC-MS by four laboratories.
Journal of Veterinary Diagnostic Investigations 3: 357-358
12
18. Rheeder JP, Marasas WFO, Thiel PG, Sydenham EW, Shephard GS and Van

Schalkwyk DJ (1992) Fusarium moniliforme and fumonisins in corn in

relation to human cancer in Transkei. Phytopathology 82: 353-357

19. Rheeder JP, Marasas WFO and Vismer HF (2002) Production of fumonisin
analogues by Fusarium species. Applied and Environmental Microbiology
68: 2101-2105
20. Ross PF, Rice LG, Plattner RD, Osweller TM, Wilson TM, Owens DL, Nelson
HA and Richard JL (1991) Concentration of fumonisin B1 in feeds associated
with health problems. Mycopathologia 114: 129-135
21. Scott PM and Lawrence GA (1992) Liquid chromatographic determination
of fumonisins with 4-fluoro-7-nitrobenzofurazan. Journal of AOAC
International 75: 829-834
22. Scott PM and Lawrence GA (1994) Stability and problems in recovery of
fumonisins added to corn based foods. Journal of AOAC International 77:
541-545
23. Shephard GS and Sewram V (1994) Determination of the mycotoxin
fumonisin B1 in maize by reversed-phase thin-layered chromatography: a
collaborative study. Food Additives and Contaminant 21: 498-505
24. Shephard GS, Thiel PG, Stockenström S and Sydenham EW (1996)
Worldwide survey of fumonisin contamination of corn and corn-based
products. Journal of AOAC International 79: 671-687
13
25. Shephard GS (1998) Chromatographic determination of fumonisin

mycotoxins. Journal of Chromatography A 815: 31-39

26. Shephard (2000) Liquid Chromatographic Method for fumonisins in corn,

methods in molecular biology. Mycotoxin

Protocols 157: 147-158
27. Sun G, Wang S, Su J, Huang T, Tang L, Gao W and Wang JS (2007)
Fumonisins B1 contamination of home-grown corn in high-areas for
oesophageal and liver cancer in China. Food Additives and Contaminants
24: 181-185
28. Sydenham EW, Gelderblom WCA, Thiel PG and Marasas WFO (1990)
Evidence for the natural occurrence of fumonisins B1, a mycotoxin
produced by Fusarium moniliforme in corn. Journal of Agricultural and
Food Chemistry 38: 285-290
29. Sydenham EW, Shephard GS, Thiel PG, Stockenström S, Snijman PW and
Van Schalkwyk DJ (1996) Liquid chromatographic determination of
fumonisins B1, B2 and B3 in corn. Journal of AOAC International 79: 688-696
30. Turner NW, Subrahman S and Piletsky SA (2009) Analytical methods for
determination of mycotoxins: A review. Analytica Chimica Acta 632: 168-
180
31. Ueno Y, IIjima K, Wang SD, Sugiura Y, Sekijima M, Tanaka T, Chen C and Yu
SZ (1997) Fumonisins as a possible contributory risk factor for primary liver
cancer: A 3-year study of corn harvested in Haimen China by HPLC and
Elisa. Food and Chemical Toxicology 35: 1143-1150
14
32. Van der Westhuizen L, Shephard GS, Scussel VM, Costa LLF, Vismer HF,

Rheeder JP and Marasas WFO (2003) Fumonisin contamination and

Fusarium incidence in corn from Santa Catarina, Brazil. Journal of

Agricultural and Food Chemistry 51: 5574-5578
33. Wang J, Zhou Y, Liub W, Zhu X, Du L and Wang Q (2008) Fumonisin level in
corn-based food and feed from Linxian County a high-risk area for
oesophageal cancer in China. Food Chemistry 106: 241-246
34. Zollner P and Helm MB (2006) Trace mycotoxin analysis in complex
biological and food matrices by liquid chromatography-atmospheric
pressure ionization mass spectrometry. Journal of Chromatography A 123:
123-169
15

CHAPTER 2
Literature review
2.1 Introduction

2.1.1 Background to fumonisins

Fumonisins are mycotoxins that are produced by several species of Fusarium

mainly F. verticillioides and F. proliferatum (Marasas, 2001). Fumonisins are
known to be an epidemiologically and economically important group of
mycotoxins (Shephard, 2000). They were first isolated in 1988 from F.
verticillioides strain MRC 826 at the Programme of Mycotoxin and Experimental
Carcinogenesis (PROMEC) of the Medical Research Council for South Africa
(MRC) by Gelderblom et al., 1988. The structure of fumonisins was determined in
1988 in a collaborative study between the PROMEC Unit of the MRC and the
Council for Scientific and Industrial Research (CSIR) in Pretoria (Marasas, 2001)
and is based on a long eicosane hydrocarbon chain substituted with methyl,
hydroxyl and amino groups. Fumonisins are diesters of propane-1,2,3-
tricarboxylic acid and 2-amino-12,16-dimethylpolyhydroxyeicosanes in which the
C14 and C15 hydroxyl groups are esterified with the terminal carboxyl group of
tricarballylic acid (Bezuidenhout et al., 1988; Figure 2.1). There are at least four
closely related series of fumonisins viz., A, B, C and P that have been isolated
(Rheeder et al., 2002). The fumonisin B has three analogues that occur most
abundantly in naturally contaminated maize are termed FB1, FB2 and FB3
(Shephard et al., 1996). FB1 is the most predominant of the fumonisins in the
range of 70-80 %, followed by FB2 and FB3 which occur at between 15-25 % and
3-8 % respectively (Rheeder et al., 2002).
17

Figure 2.1 Schematic diagram of the structures of Fumonisin B1 (FB1),
Fumonisin B2 (FB2) and Fumonisin B3 (FB3)
[Published in: Shephard, 1998 J. Chromatogr. A 815: 31-39]
2.1.2 Occurrence of fumonisins
Mycotoxins are produced by one or more specific fungal species, with some
species forming more than one mycotoxin (EMAN). FB1 is a secondary metabolite
and therefore its occurrence is caused entirely by the existence of fungal
contamination (Miller, 2001). Fumonisins are natural contaminants of cereal
grains worldwide (Weidenborner, 2001) but have been found to occur
predominately in maize and maize-based products (Shephard et al., 1996; Figure
2.2). Fumonisin contamination has been reported in a variety of food
commodities, which include sweet maize at low concentrations of 4-82 µg/kg
(Trucksess et al., 1995) and maize beer samples at concentrations of 43-1329
18
µg/kg for total fumonisins (Shephard et al., 2005). Other commodities where

evidence of fumonisin contamination occurs include wheat, rice and cereal-

based food (Shephard et al., 1996). FB2 has recently been reported to be present

in red wine (Logrieco et al., 2010), black

tea and medicinal plants (Martins et al.,
2001). One of highest fumonisin levels in maize intended for human
consumption was reported by Rheeder et al., 1992 in the Transkei region, South
Africa with a FB1 level of 117.5 mg/kg. Other areas of high contamination include
Santa Catarina State, Brazil and Huainan and Fusui, China (Van der Westhuizen et
al., 2003; Sun et al., 2007). The highest levels ever reported for animal feed (330
mg/kg) were found in the US maize screening (Ross et al., 1991).
Figure 2.2 Maize infected with fumonisin producing fungi, F. Verticillioides.
[Picture obtained from WFO Marasas, PROMEC Unit]
19
Exposure assessment studies have been performed in the rural former Transkei

region of South Africa. Human exposure in the region of Bizana, an area of

relatively low oesophageal cancer, was found to be 3.43 ± 0.15 µg/kg body

-1
weight day which was lower than in Centane, an area with high oesophageal
cancer, which was reported to have a mean exposure of 8.67 ± 0.18 µg/kg body
weight day-1. Both areas reported results higher than 2 µg/kg body weight day-1
which is the provisional maximum tolerable daily intake set by the Joint FAO /
WHO Expert Committee on Food Additives (Shephard et al., 2007).
2.1.3 Impact of fumonisins
Fusarium verticillioides MRC 826 culture material is highly hepatotoxic and
cardiotoxic in rats (Kriek et al., 1981) and was later found to be
hepatocarcinogenic in rats and to cause primary hepatocellular carcinoma and
cholangiocarcinoma in rats (Marasas 1984). Fumonisins have been reported to
cause equine leukoencephalomalacia (ELEM) in horses orally dosed with
fumonisin B1 (Kellerman et al., 1990), and to cause porcine pulmonary edema
(PPE) in pigs (Harrison et al., 1990). Laboratory studies have shown FB 1 to be
hepatocarcinogenic and nephrocarcinogenic in male rats and hepatocarcinogenic
in female mice (Gelderblom et al., 1991; Howard et al., 2001).
The high rates of human esophageal cancer have been associated with high
intake of maize contaminated with fumonisins (Rheeder et al. 1992). Due to
increased rates of neural tube defects (NTD) in populations along the Texas-
Mexico border (Hendricks et al., 1999; Marasas et al., 2004), and typical of areas
20
where maize forms part of the diet, an investigation was conducted to correlate

NTD with the consumption of fumonisin contaminated maize by the mothers.

The findings suggested that “fumonisin exposure increases NTD risk,

proportionate to dose and up to the threshold

level at which death may be more
likely to occur” (Missmer et al., 2006). It has also been suggested that fumonisins
are a risk factor for liver cancer in humans (Ueno et al., 1997; Marasas et al.,
2004). Further elaboration of the human health effects of fumonisins requires
the availability of a suitable biomarker of exposure. The biochemical mechanism
of action of fumonisins is disruption of de novo sphingolipd biosynthesis leading
to an accumulation of sphinganine and an increase in the sphinganine :
sphingosine ratio in cells and physiological fluids (Van der Westhuizen et al.,
2008; Silva et al., 2009 b). Attempts to use this as a biomarker in humans have
been unsuccessful but recently a urinary FB1 biomarker has been validated in the
population of the former Transkei (Van der Westhuizen et al., 2011).
Economically, fumonisins impact directly on animal loss, health and veterinary
care costs as well as regulatory and research costs which focus in order to
determine the impact and severity of mycotoxin problems (Hussein et al., 2001;
Wu, 2004; Shephard, 2000). A few countries worldwide have legislated maximum
tolerated levels, including Bulgaria, Cuba, France, Iran and Switzerland (FAO,
2003). The Joint FAO/WHO Expert Committee on Food Additives (JECFA) has
recommended a provisional maximum tolerable daily intake (PMTDI) of 2 µg/kg
body weight for FB1, FB2 and FB3 alone or in combination (Bolger et al., 2001).
The US Food and Drug Administration (FDA) has guidance levels for fumonisins in
21
human food and animal feed and has recommended that levels be put in place to

reduce the exposure of fumonisins in maize products intended for human and

animal consumption (FDA, 2001a). The European Commission (EC) has regulated

fumonisins at various levels from 4000 µg/kg for unprocessed maize to 200 µg/kg
for baby foods (EC, 2007).
Analytical methods for all fumonisin analogues in maize and maize-based foods
generally rely on reversed-phase high-performance liquid chromatography (RP-
HPLC) separation after suitable extraction and clean-up. As fumonisins lack a
useful chromophore or fluorophore for HPLC, detection is facilitated by suitable
derivatization followed by sensitive fluorescence detection (Shephard et al.,
1996). There are currently several methods used to measure the concentration
of fumonisins in various matrixes (Shephard, 2008). One of the most commonly
used methods for quantitative analysis in maize involves solid-phase extraction
(SPE) of solvent extracts, followed by strong anion exchange (SAX) SPE and o-
phthaldiadehyde (OPA) derivatization prior to HPLC separation and
quantification of fluorescent OPA-FB1, FB2 and FB3 derivatives (Shephard, 1998).
2.2 Analysis of fumonisins
2.2.1 Introduction
The majority of analytical methods employed in the analysis of fumonisins
include sampling and sub-sampling, appropriate extraction, clean-up and
concentration and then derivatization prior to instrumental analysis (Shephard et
al., 2011). Fumonisin analysis by HPLC requires sample clean-up in order to
22
remove matrix impurities and to then concentrate the fumonisin (Shephard,

1998). There are different clean-up methods that could be used; these include

SAX, immunoaffinity columns (IAC), QUeChERS (Quick, Easy, Cheap, Rugged and

Safe) and C18 columns, and each clean-up

has its own advantages (Bennett et al.,
1994; Visconti et al., 1996; Stockenström et al., 1994; Shephard et al., 2011;
Zachariasova et al., 2010).
2.2.2 Extraction of samples
The fumonisin structure is quite polar due to the four carboxylic acid groups and
an amine group, which makes them readily soluble in polar solvents (Wilkes et
al., 1998) and hence amenable to extraction using polar solvents such as
methanol : water, acetonitrile : water, methanol : acetonitrile : water and
acetonitrile : sodium dihydrogen phosphate and by using different combinations
and proportions followed by a clean-up step using SPE on a reversed phase C18
column, SAX or IAC (Cortez-Rocha et al., 2003; Shephard, 2000) it is reasonable
to anticipate a good clean product will result.
Food matrixes are generally extracted either by acetonitrile : water (1:1, v/v)
(Rice et al., 1995; Bennett et al., 1994) or methanol : water at 70-80 % methanol
with optimum results being obtained when using methanol : water (3:1)
(Shephard, 1998). Different extraction efficiencies have been reported when
using either blending or homogenization. Increased efficiencies were reported by
Sydenham et al., 1992 by using methanol : water (3:1) combined with
homogenization between 1-5 minutes. In contrast, Bennett et al., 1994 later
23
obtained better extraction efficiencies with acetonitrile : water compared to

methanol : water when shaking was employed for 30-60 minutes.

The European Intercomparison Study

showed that the use of higher
solvent/sample ratios improved the recoveries of fumonisins (Visconti et al.,
1996). The extraction of highly contaminated samples is difficult compared to the
extraction of spiked maize which can be attributed to matrix components in the
sample (Bennet et al., 1994). The extraction of processed baby foods has been
found to be particularly difficult. This could be caused by different factors such as
matrix interference. Optimum results for extraction of cornmeal-based infant
foods were achieved with 70 % methanol at pH 4 (Sewram et al., 2003).
Comparison of different extraction procedures showed that acetonitrile : water
(1:1) gives higher recoveries than methanol : water (3+1, v/v) for all infant
formulae even though phase separation during the extraction step with
acetonitrile : water showed it to be an inappropriate mixture (De Girolamo et al.,
2001). Other approaches that have been reported for the improvement of
extraction of fumonisins in maize include the use of extraction solvent mixtures
at alkaline pH (Scott and Lawrence 1994), the use of EDTA as extraction solvent
(Kim et al., 2002) and increasing the temperature of the solvent used
(Lawrence et al., 2000).
24
2.2.3 Clean-up

Fumonisin sample extracts are normally purified in order to remove unwanted

and interfering matrix impurities and to facilitate the concentrate of the

fumonisins (Shephard, 1998). The most

widely applied purification methods (SAX
and IAC column cartridges) will be investigated.
2.2.3.1 Strong anion exchange clean-up (SAX)
The anionic nature of fumonisins is the reason why SAX is the most widely used
method for clean-up of maize samples (Maragos et al., 1996). Purification of
maize samples with SAX is dependent on the pH or ionic strength of the sample
(Shephard, 1998). For optimum recovery results, the pH of the sample extracts
must be monitored at 5.8-6.2 with elution flow rate < 2.0 mL/min (Sydenham et
al., 1992). SAX cartridges have been reported to provide superior purification
over C18 clean-up (Visconti et al., 1996).
2.2.3.2 Immunoaffinity column clean-up (IAC)
IACs are composed of mycotoxin-specific antibodies bound to the sol-gel
material and packed in small cartridges. The specificity of the antibody ensures a
relatively clean final product (Kruger et al., 1999). Once bound to the antibody,
the mycotoxin is eluted by denaturing the antibody using an organic solvent
(Trebstein et al., 2008), usually methanol. For this reason, it is not generally
recommended that IACs be used more than once. IAC provides much cleaner
analyte eluates when compared to SPE clean-up although not completely
25
selective to the analyte (Kruger et al., 1999). These have been reported to be

more robust for fumonisin analysis and are less likely to present matrix

interferences compared to other SPE methods (De Girolamo et al., 2001). IAC

clean-up is used mainly for mycotoxins

from varying diverse matrices (Krska et
al., 1998) including fumonisins in highly contaminated maize samples (Kim et al.,
2004).
After extraction and clean-up, samples are often analyzed using different
chromatographic methods varying from thin layer chromatography (TLC) to high-
performance liquid chromatography, which is the most widely used analytical
method for the detection of fumonisins.
2.2.4 Derivatization
For sensitive detection of fumonisins by spectrometric methods, derivatization is
required to form suitable derivatives that can be easily isolated, separated and
detected (Shephard 1998). Since they lack a suitable chromophore or
fluorophore they are derivaterized prior to HPLC injection using fluorogenic
reagents like o-phthaldialdehyde (OPA), naphthalene-2,3-dicarboxaldehyde
(NDA) or dansyl chloride (DnS-Cl), with OPA the most commonly used
derivatization reagent as it yields highly fluorescent compounds that can be
easily separated by HPLC (Shephard, 2000; Bennett and Richard 1994).
The analytical result of fumonisin analysis is greatly affected by the
derivatization reagent used viz., fluorescamine formed two reaction products
with fumonisins (Sydenham et al., 1990) and maleyl derivatives produced
26
detection limits of 10 µg/g with HPLC-UV which is unsuitable for naturally

contaminated maize (Scott and Lawrence 1994). Other derivatization reagents

that have been investigated for fumonisin analyses include 4-fluoro-7-

nitrobenzofuran (NBDF) (Scott

and Lawrence 1992), 9-
fluorenylmethylchloroformate (FMOC) (Holcomb et al., 1995) and 6-amino-
quinolyl-N-hydroxysuccinimidylcarbamate (AccQ.Fluor) (Velazquez et al., 2000).
The derivatization reagents that will be investigated in this study are OPA, NDA
and DnS-Cl.
2.2.4.1 O-phthaldialdehyde (OPA) derivatization reagent
Figure 2.3 Structure of o-phthaldialdehyde (OPA)
OPA is the most widely used derivatization reagent for pre-column
derivatization of fumonisins in most laboratories (Shephard et al., 1996).
Fumonisins react with OPA in the presence of a sulphur containing nucleophile,
most usually 2-mercaptoethanol (ME) to form a highly fluorescent isoindole, the
reaction occurring at the free amine moiety as illustrated by figure 2.4.
The combination of OPA with other nucleophiles like OPA/N-acetyl-L-cysteine
OPA/ethanethiol and OPA/3-mercaptopropionic acid have been reported, but
27
were found to form either two reaction products or side reactions with aliphatic

amines (Hanczko et al., 2004). Stroka et al., 2002 tested the percentage decay for

2-mercaptoethanol (ME) compared with other nucleophiles and found that 94 %

of ME decayed within 2 hours but nucleophiles

like 2-thioglycerol showed no
decay within the tested time. The disadvantage of using OPA is its instability at
room temperature (Williams et al., 2004), even though the use of ME improves
the stability of the derivative and thereby its fluorescence peak area and peak
height (Stroka et al., 2002). However, OPA/ME is not stable enough as a
derivative for overnight or auto-injection analysis (Williams et al., 2004). OPA
derivatives have been reported to be stable for up to 4 minutes at room
temperature after which they decrease by 5 % after 8 minutes and 52 % after 64
minutes in their specific fluorescence response of the FB1 derivative. The time-
degradation of fumonisin derivatives can be overcome by standardizing the time
(< 4 minutes) between the reagent addition and the HPLC injection (Shephard,
1998). Despite its instability problems, the AOAC (Association of Official
Analytical Chemists International) approved the OPA derivatization of fumonisins
as an official method (Sydenham et al., 1996).
Blank maize samples spiked with FB1, FB2 and FB3 standards at concentrations of
100 to 8000 ng/g produce mean recoveries of between 75% and 85% for
individual toxins and detection limits of 50 ng/g have been reported for OPA
analysis (Sydenham et al., 1996).
28

Figure 2.4 Reaction mechanism for the formation of FB-OPA complex
[Published in: Samapundo et al., 2006. J. Chromatog. B 1109: 312-
316]
29
2.2.4.2 Naphthalene-2,3-dicarboxaldehyde (NDA) derivatization reagent

Figure 2.5 Structure of Naphthalene-2,3-dicarboxaldehyde (NDA)
NDA is known to be a useful derivatization reagent for primary amines, amino
acids and small peptides and is the second most widely used fluorogenic reagent
for the detection of amines (Carlson et al., 1986). It was screened by De
Montighy et al., (1987) using several nucleophiles like ME, HSO3- and CN- with
alanine as a primary amine. Under fluorescence spectroscopy, CN- was the most
suitable reagent to form derivatives with both high fluorescence intensity and
good chemical stability (De Montighy et al., 1987). The NDA/CN - reaction was
then further tested with other primary amines, amino acids, peptides and
proteins at room temperature to produce highly fluorescent and stable 2-
substituted 1-cyanobens [f] isoindole derivatives, which are often used to
measure trace levels of biogenic amines in biological matrixes (Bennet et al.,
1994; Figure 2.6).
30

Figure 2.6 Reaction of NDA with FB1 to form 1-Cyanobenzisoindole (CBI)
stable derivative
[Published in: Bennett and Richard 1994 J. AOAC Int. 77: 501-506]
The above reaction is for FB1; other fumonisin B analogues follow the same
reaction sequence.
Most analytical methods derivatizing fumonisins with NDA and other amines are
based on the method by Ware et al., 1993 with minor modifications. Extraction
methods used vary as per laboratory, sample type and instrument of analysis.
For chromatographic separation, researchers have used similar conditions, viz.,
sodium borate pH 9, NDA dissolved in acetonitrile or methanol and the addition
of cyanide to drive the reaction (Carlson et al., 1986; De Montigny et al., 1987).
Chromatographic separation of NDA derivatives is generally achieved with
acetonitrile : water : acetic acid (~ 66 : 38 : 1) mobile phase at 1 mL/min using
isocratic elution (Bennett et al., 1994; Lino et al., 2007) and fluorescence
31
detection at an excitation wavelength of 420 nm and emission wavelength of

520 nm (Bennett et al., 1994; De Montigny et al., 1987). This thesis describes for

the first time a comparison of NDA with other derivatization reagents using both

diode array detection and fluorescence

detection on South African maize
samples for the most naturally abundant fumonisin B analogues (FB1, FB2 and
FB3).
When compared with other derivatization reagents, NDA gives comparable
results and can be used as an alternative to OPA as it produces more stable and
highly fluorescent derivatives of fumonisins (Ware et al., 1993). NDA is a stable
derivatization reagent and it has been reported that after 24 hours the NDA
fluorescent signal decreases to 86.4±5.9 % as compared to 63.2±4.8 % of OPA
and to 82.8±7.1 % after 48 hours for NDA compared to 57.2±8.3 % for OPA (Cho
et al., 2002).
Detection limits of 20 µg/kg for FB1 and 15 µg/kg for FB2 were reported by Silva
et al., (2009 a) and 23.3 nmol/L and 34.4 nmol/L for amines (Lamba et al., 2008).
Bennett et al., 2004 reported recoveries of 92-95 % for FB1 and FB2, respectively,
at levels of 10 µg using SAX columns and 83-88 % for FB1 and FB2 at 10 µg levels
using RP C18 columns, whereas Silva et al., (2009 a) reported recoveries of 79 %
to 102 % for FB1 and FB2, respectively, at spiking levels of 150 µg/kg and 250
µg/kg using LC -MS.
32
2.2.4.3 Dansyl Chloride (DnS-Cl) derivatization reagent

Figure 2.7 Structure of Dansyl Chloride (DnS-Cl)
Dansyl chloride (DnS-Cl) (5-dimethylaminonaphthalene-1-sulfonyl chloride) is a
fluorescent labelling reagent for primary, secondary and tertiary amines
(Bartzatt, 2001). It was introduced in 1952 by G. Weber to prepare conjugated
proteins (Blau et al., 1978). It is often used for the quantitation of polyamines in
biological samples (Khuhawar et al., 2001) and in the presence of amino acids,
dansyl chloride forms stable fluorescent sulfonamide adducts (Walker et al.,
1994) and in the presence of sodium carbonate it allows for the detection of 1 µg
amounts of analyte (Bartzatt, 2001).
Sodium carbonate (1 M) has sufficient ionic strength and high pH to enable the
binding of DnS to tertiary amines where elimination of either alkyl or aryl
substituents is possible (Bartzatt 2001). DnS-Cl derivatives are more stable than
OPA and more suitable for detection at pico molar ranges (Minocha et al., 2004)
33
and are stable for over two hours (Dasko et al., 2006). It has been reported to be

a good derivatization reagent for fumonisin analysis, but with the disadvantage

of forming analytical interferences with maize samples (Scott et al., 1992; Arranz

et al., 2004). The derivatization of amine

components with DnS-Cl can produce
intense overlapping HPLC peaks due the reaction of DnS-Cl with water to
produce hydrolysis products (Kang et al., 2006). Sample clean-up with IAC has
been used to eliminate interferences for determination of FB 1 in beer samples
(Dasko et al., 2006). Cleaner chromatograms are also obtained by reacting
minimum stoichiometric amounts of DnS-Cl or adding triethylamine and
tetrabutylammonium hydroxide to separate the hydrolysis peaks from the
analyte peak (Kang et al., 2006). Environmental conditions like temperature
(Dasko et al., 2006), reaction time, pH and concentration of DnS-Cl can affect the
reaction yield (Kang et al., 2006). Hence DnS-Cl methods should be optimized as
per laboratory and sample matrix.
Dansyl chloride derivatives can be measured using both ultraviolet and
fluorescence detection with UV detection at 286 nm being the most sensitive
(Minocha et al., 2004). Recoveries between 89.47-97 % have been reported for
the separation of di- and polyamines as their dansyl derivatives using RP-HPLC
and methanol : water or acetonitrile : water mobile phase (Marce et al., 1995).
Recoveries of amines of 73.8 to 114 % in wine samples with detection limits of
0.08 ng and quantization limit of 0.16 ng where reported by Loukou et al., 2003.
34
2.3 Detection of fumonisins

2.3.1 Introduction

There are various separation techniques used for fumonisin analysis which vary

from chromatography, enzyme-linked immunoassay (ELISA) to electrophoresis
methods. Alternative methods for the detection of fumonisin which are often
used for confirmation of the presence of fumonisins include gas chromatography
(GC), HPLC, liquid chromatography- mass spectrometry (LC-MS), and thin-layer
chromatography (TLC) which provides faster analysis time for rapid screening of
samples (Shephard 1998) with the ELISA and TLC often being relegated to
screening methods. Fumonisins are polar molecules that are soluble in water and
polar solvents are thus suited for analysis by RP-HPLC (Shephard, 1998). This
study uses RP-HPLC for chromatographic separation of the fumonisin B
analogues.
2.3.2 Chromatographic methods
2.3.2.1 High-performance liquid chromatographic
A world-wide survey found that 90 % of the laboratories that reported results on
fumonisin analysis used pre-column derivatization and quantification by HPLC
(Shephard et al., 1996). Since fumonisins do not fluorescence or contain a UV
absorbing chromophore, most HPLC methods measure fumonisins after
derivatization of their free amino group (Shephard, 2000). Shephard et al., 1990
reported for the first time on the quantitative, sensitive and simultaneous
35
detection of FB1 and FB2 in naturally contaminated maize samples using RP-HPLC.

The method was based on MeOH : H2O extraction, SAX clean-up, fluorescence

(FLD) detection of OPA derivatives. Collaborative studies using the method

resulted in the AOAC International approving

it an official method for fumonisin
analysis in maize (Sydenham et al., 1996). HPLC-FLD detection of fumonisins is
specific and sensitive with fluorescence detection often achieved at an excitation
wavelength of 335 nm and emission wavelength of 440 nm (Shephard et al.,
1990).
2.3.2.2 Gas Chromatography
Sydenham et al., 1990 confirmed the presence of fumonisins in maize using GC.
The method involved acid hydrolysis of the fumonisins to cleave the ester bond
and the tricarballylic acid thus formed was confirmed by GC-MS. A direct method
for fumonisin analysis involved the production of the fumonisin backbone
(aminopolyol) by alkaline hydrolysis which was isolated on XAD-2 resin and then
converted to the trimethylsilyl derivative for GC analysis (Plattner, 1990).
Plattner et al., 1994 later reported that the accuracy and precision of the GC-MS
method can be improved by adding deuterium-labelled FB1 as an internal
standard to the sample extract prior to hydrolysis. Fumonisin analysis using GC
requires multiple time consuming sample handling steps such as sample
hydrolysis, clean-up and derivatization prior to analysis and so has found little
application for fumonisin detection (Shephard, 1998).
36
2.3.2.3 Thin Layer Chromatography

TLC provides a fast and reliable means
of screening contaminated samples. The
first method developed for fumonisins

involved reversed-phase TLC on C18-
modified silica plates developed with methanol : water (3:1, v/v) as a solvent
(Cawood et al., 1991). This method was later improved by the use of p-
anisaldehyde solutions or spraying with ninhydrin to visualize the fumonisins
(Shephard, 2000). However, the detection limits of 0.5 mg/g which were
obtained were not suitable for naturally contaminated maize (Sydenham et al.,
1990). When used with fluorescamine reagent spray under UV light, TLC gave
better selectivity and sensitivity for fumonisin analysis in naturally contaminated
maize samples (Rottinghaus et al., 1992). Improved detection limits were
obtained with the use of SAX clean-up rather than reversed-phase C18 cartridges
(Stockenström et al., 1994). TLC detection limits were further enhanced with IAC
clean-up and scanning fluorodensitometry to 0.1 mg/kg in maize samples (Preis
et al., 2000). A TLC method based on pre-derivatization before TLC separation
was reported in a collaborative study for FB1 analysis in maize and gave recovery
results of 74.5 % (Shephard et al., 2004).
2.3.3 Immunological methods
Immunological assays have been used to successfully detect mycotoxins since
the late 1970s (Pestka et al., 1995). These methods rely on the immunological
principle which is based on the interaction between an antigen (analytes of
interest) and an isolated antibody raised against the antigen (Shephard, 2008).
37
These methods include enzyme-linked immunosorbent assays (ELISA) and

immunoaffinity columns (IAC). ELISA has been validated for measuring total FB in

maize at levels greater than 0.1 µg/g with acceptable precision (Bird et al., 2002).

IAC have been developed with specific

antibodies for different mycotoxins with
recovery results averaging between 99.7 % for FB1 at fortification levels of 250
µg/kg and 74.8 % for FB2 at fortification levels of 200 µg/kg (Lino et al., 2007).
Immunoassays are still being used for screening commodities and food for
fumonisins with new developments in antibodies and immunoassays reported
(Shephard et al., 2011).
2.3.4 Mass Spectrometry
2.3.4.1 Liquid chromatography-mass spectrometry
LC-MS is a combination of HPLC with MS and is a powerful technique for
identification of fumonisin B analogues (Shephard, 1998). The use of LC-MS has
enabled sensitive and specific fumonisin methods to be developed (Silva et al.,
2009 a). LC-MS/MS has recently been used for multi-mycotoxin screening of 87
mouldy foods sampled from individual homes. The method involves acetonitrile :
water : acetic acid extraction with LC-MS/MS-ESI (electrospray ionization) and
HPLC-MS/MS detection (Sulyok et al., 2010).
2.3.5 Recent developments in fumonisin detection
The use of multi-mycotoxin analysis has drawn much attention within the
toxicology industry with the use of sophisticated instruments such as UHPLC-
38
MS/MS and LC-MS/MS-ESI. These have recently involved multi-component

methods for the simultaneous detection of mycotoxins and pesticides (Romero-

Gonzalez et al., 2011; Sulyok et al., 2010). Ofitserova et al., 2005 provided a

screening method for five families of

toxins, a method suitable for screening
beverages, grains and feeds. It involved chromatographic separation with
MYCOTOXTM reversed-phase C18 column and post-column separation
instrument, Pinnacle PCX (Pickering laboratories) (Ofitserova et al., 2005). A LC-
MS/MS method for multi-mycotoxin has recently been reported by Sulyok et al.,
2010. The method involves semi-quantitative screening of 87 mouldy samples.
From the analysis results, 49 different fungal metabolites were identified,
showing the usefulness of multi-mycotoxin analysis. Other rapid screening
methods for mycotoxin analysis include fluorescence polarization immunoassay
(FPIA) (Maragos, 2009), lateral flow devices (LFD; dipstick) and biosensors
(Maragos and Busman 2010).
2.4 Overview of literature review
The fumonisin B analogues where first isolated in 1988 at the PROMEC Unit of
the MRC and occur in maize and maize-based foods. They cause ELEM in horses,
porcine pulmonary oedema in pigs, and are hepatotoxic and cardiotoxic in rats.
They have been linked to the high incidence of esophageal cancer in the rural
Transkei area of South Africa where maize is the stable diet. In 2002, the IARC
evaluated the carcinogenic risk of FB1 to humans and classified it as a Group 2B
carcinogen (as a possibly carcinogen to humans). The current status of existing
39
methods available for fumonisin analysis includes extraction, clean-up,

derivatization and chromatographic separation. HPLC with fluorescence, MS or

tandem MS are still the most used in laboratory-based methods.

The method developed by Shephard et al., 1990 was approved by the AOAC
International as an official method for fumonisin analysis in maize. Derivatization
reagents that have been tested for fumonisin analysis include OPA, NDA and
DnS-Cl which will be studied in this thesis. HPLC with fluorescence detection is
the most widely used detection method. The thesis will investigate the
applicability of UV as a possible detection method in HPLC alternative to
fluorescence.
40
References

1. Arranz I, Baeyens WRG and Van der Weken G (2004) Review: HPLC

determination of fumonisin mycotoxins. Critical Reviews in Food Science
and Nutrition 44: 195-203
2. Bartzatt R (2001) Fluorescent labelling of drugs and simple organic
compounds containing amine functional groups, utilizing dansyl chloride in
Na2CO3 buffer. Journal of Pharmacology Toxicology Methods 45: 247-253
3. Bennett GA and Richards JL (1994) Liquid chromatographic method for
analysis of naphthalene dicarboxaldehyde derivative of fumonisins. Journal
of AOAC International 77: 501-506
4. Bezuidenhout SC, Gelderblom WCA, Gorst-Allman CP, Horak RM,Marasas
WFO, Spiteller G and Vleggaar R (1988) Structure elucidation of the
fumonisins mycotoxins from Fusarium moniliforme. Journal of the
Chemical Society Chemical Communications 783: 743-745
5. Bird CB, Malone B, Rice LG, Ross PF, Eppley R and Abouzied MM (2002)
Determination of total fumonisins in corn by competitive direct enzyme
linked immunosorbent assay: collaborative study. Food Chemical
Contaminants 85: 404-410
6. Blau K and King G (1978) Handbook of derivatives for chromatography pp.
104-151 Heyden London
41
7. Bolger M, Coker RD, DiNovi M, Gaylor D, Gelderblom W, Olsen M, Paster N,

Riley RT, Shephard G and Speijers GJA (2001) Fumonisins In: Safety

evaluation of certain mycotoxins in food. Prepared by the Fifty-sixth

Meeting of the Joint FAO/WHO

Expert Committee on Food Additives
(JECFA), WHO Food Additives Series No. 47. FAO Food and Nutrition Paper
No 74, WHO, Geneva, Switzerland, 74: 103–279
8. Carlson RG, Srinvasachar K, Givens RS and Matuszewski BK (1986) New
derivatizing agents for amino acids and peptides: facile synthesis of N-
substituted 1-cyanobenz [f] isoindokes and their spectroscopic properties.
Journal of Organic Chemistry 51: 3978-3983
9. Cawood ME, Gelderblom WCA, Vleggaar R, Behrend Y, Thiel PG and
Marasas WFO (1991) Isolation of fumonisin mycotoxins: A quantitative
approach. Journal of Agricultural and Food Chemistry 39: 1985-196
10. Cho YH, Yoo HS, Min JK, Lee EY, Hong SP, Chung YB and Lee YM (2002)
Comparative study of naphthalene-2,3-dicarboxaldehyde and o-
phthaldialdehyde fluorogenic reagents for chromatographic detection of
sphingoid bases. Journal of Chromatography A 977: 69-76
11. Cortez-Rocha MO, Ramirez WR, Sanchez RI, Rosas EC, Wong FJ, Borboa J,
Castillion LG and Tequida M (2003) Fumonisins and fungal species in corn
from Sonora. Bulletin of Environmental Contamination and Toxicology 70:
668-673
42
12. Dasko L, Rauova D and Belajova E (2006) Comparison of the suitability of

derivatization agents in HPLC-fluorescence detection analysis of

fumonisins. Journal of Food and Nutrition Research 45: 127-133

13.
De Girolamo A, Solfrizo M, von Hoist C and Visconti A (2001) Comparison of
different extraction and clean-up procedures for the determination of
fumonisins in maize and maize based food products. Food Additives and
Contaminants 18: 59-67
14. De Montigny P, Stobaugh JF, Givens RS, Carlson RG, Srinivasachar K,
Sternson LA and Hihuchi T (1987) Naphthalene-2,3-dicarboxaldehyde/
cyanide ion: A rationally designed fluorogenic reagent for primary amines.
Analytical Chemistry 59: 1096-1101
15. European Commission Regulation (EC) No 1126/2007 of 28 September
2007, Official Journal of the European Union L255:14-17
16. European Mycotoxin Awareness Network (EMAN), www.mycotoxins.org/
17. Food and Agriculture Organization of the United Nations, 2004. Worldwide
Regulations for Mycotoxins in Food and Feed in 2003. FAO Food and
Nutrition Paper 81. FAO, Rome
18. Food and Drug Administration (2001a) Guidance for Industry: Fumonisin
levels in human foods and animal feeds. Final Guidance. Centre for Food
Safety and Food Safety and Applied Nutrition, Centre for Veterinary
Medicine (http://www.cfsan.fda.gov/~dms/fumongu2.html)
43
19. Gelderblom WCA, Jaskiewicz K, Marasas WFO, Thiel PG, Horak RM,

Vleggaar R and Krieg NP (1988) Fumonisins-Novel mycotoxins with cancer-

promoting activity produced by Fusarium moniliforme. Applied and

20. Gelderblom WCA, Kriek NPJ, Marasas WFO and Thiel PG (1991) Toxicity and
carcinogenicity of the Fusarium moniliforme metabolite, FB1, in rats.
Carcinogenesis 12: 1247-51
21. Hanczko R, Kutlan D, Toth F and Molnar-Perl I (2004) Behaviour and
characteristics of the o-phthaldialdehyde derivatives of n-C16-C8 amines
and phenylethylamine with four additive SH-containing reagents. Journal of
Chromatography A 1031: 51-66
22. Harrison LR, Colvin BM, Greene JT, Newman LE and Cole JR (1990)
Pulmonary edema and hydrothorax in swine produced by fumonisin B 1, a
toxic metabolite of Fusarium moniliforme. Journal of Veterinary Diagnostic
Investigations 2: 217-221
23. Hendricks K (1999) Fumonisins and neural tube defects in South Texas.
Epidemiology 10: 198-200
24. Holcomb M and Thompson C (1995) Analysis of fumonisin B1 in corn by
capillary electrophoresis with fluorescence detection of the FMOC
derivative. Journal of Microcolumn Separations 7: 451-454
44
25. Howard PC, Eppley RM, Stack ME, Warbritton A, Voss KA, Lorentzen RJ,

Kovach RM and Bucci TJ (2001) FB1 carcinogenicity in a two-year feeding

study using F344 rats and B6C3F1 mice. Environmental Health Perspectives

109: 277-82
26. Hussein HS and Jeffrey MB (2001) Toxicity, metabolism and impact of
mycotoxins on humans and animals. Toxicology 167: 101-134
27. IARC, International Agency for Research on Cancer (2002) Fumonisin B1. In:
IARC monographs on the evaluation of the carcinogenic risks to humans:
Some traditional herbal medicines, some mycotoxins, naphthalene and
styrene. IARC, Lyon, France, 82: 301-366
28. Kang X, Xiao J, Huang X and Gu Z (2006) Optimization of dansyl
derivatization and chromatographic conditions in the determination of
neuroactive amino acids of biological samples. Clinica Chimica Acta 366 :
352-356
29. Kellerman TS, Marasas WFO, Thiel PG, Gelderblom WCA, Cawood M and
Coetzer JAW (1990) Leukoencephalomalacia in two horses induced by oral
dosing of fumonisin B1. Onderstepoort Journal of Veterinary Research 57:
269-275
30. Khuhawar MY and Qureshi GA (2001) Polyamines as cancer markers:
applicable separation methods. Journal of Chromatography B 764: 385-407
45
31. Kim EK, Scott PM, Lau BPY and Lewis DA (2002) Extraction of fumonisin B1

and B2 from white rice flour and their stability in white rice, corn starch,

cornmeal and glucose. Journal of Agricultural and Food Chemistry 50:

3614-3620
32. Kim EK, Maragos CM and Kendra DF (2004) Liquid chromatographic
determination of fumonisin B1, B2 and B3 in corn silage. Journal of
33. Kriek NPJ, Kellerman TS, Marasas WFO, Thiel PG (1981) Hepato- and
cardiotoxity of Fusarium verticillioides (F. moniliforme) isolates from
southern African maize. Food and Cosmetic Toxicological 19: 447-456
34. Krska R (1998) Performance of modern sample preparation techniques in
the analysis of Fusarium mycotoxins in cereals. Journal of Chromatography
A 815: 49-57
35. Kruger SC, Kohn B, Ramsey CS and Prioli R (1999) Rapid immunoaffinity-
based method determination of zearalenone in corn by fluorometry and
liquid chromatography. Journal of AOAC International 82: 1364-1369
36. Lamba S, Pandit A, Sanghi SK, Gowe VS, Tlwari A, Baderia VK, Singh DK and
Nigam P (2008) Determination of aliphatic amines by high performance
liquid chromatography-amperometric detection after derivatization with
naphthalene-2,3-dicarboxaldehyde. Analytica Chimica Acta 614: 190-195
46
37. Lawrence JF, Niedzwiadek B and Scott PM (2000) Effect of temperature and

solvent composition on extraction of FB1 and FB2 from corn products.

Journal of AOAC International 83: 604-611

38.
Lino CM, Silva LJG, Pena A, Fernandez M and Manes J (2007) Occurrence of
fumonisins B1 and B2 in broa, typical Portuguese maize bread. International
Journal of Food Microbiology 118: 79-82
39. Logrieco A, Ferracane R, Visconti A and Ritieni A (2010) Natural occurrence
of fumonisin B2 in red wine from Italy. Food Additives and Contaminates
27: 1136-1141
40. Loukou Z and Zotou A (2003) Determination of biogenic amines as dansyl
derivatives in alcoholic beverages by high-performance liquid
chromatography with fluorimetric detection and characterization of the
dansylated amines by liquid chromatography-atmospheric pressure
chemical ionization mass spectrometry. Journal of Chromatography A 996 :
103-113
41. Maragos CM, Bennett GA and Richard JL (1996) Analysis of fumonisins B1 in
corn by capillary electrophoresis. Advances in Experimental Medicine and
Biology 392: 105-112
42. Maragos CM (2009) Fluorescence polarization immunoassay of mycotoxins
: A review 1: 196-207
47
43. Maragos CM and Busman M (2010) Rapid and advanced tools for

mycotoxin analysis: A review. Food Additives and Contaminants 27: 688-

700

44. Marasas WFO, Kriek NPJ, Fincham

JE and van Rensburg J (1984) Primary
liver cancer and oesophageal basal cell hyperplasia in rats caused by
Fusarium moniliforme. International Journal of Cancer 3: 383-387
45. Marasas WFO (2001) Discovery and occurrence of fumonisins: A historical
perspective. Environmental and Health Perspectives 109: 239-243
46. Marasas WFO, Riley RT, Hendricks KA, Stevens VL, Sadler TW, Waes JGV,
Missmer SA, Cabrera J, Torres O, Gelderblom WFO, Allegood J, Martinez C,
Maddox J, Miller JD, Starr L, Sullards MC, Roman AV, Voss KA, Wang E and
Merrill AH Jr. (2004) Fumonisins disrupt sphingolipid metabolism, folate
transport and neural tube development in embryo culture and in vivo: a
potential risk factor for human neural tube defects amoung populations
consuming fumonisin-contaminated maize. American Society for
Nutritional Sciences 134: 711-716
47. Marce M, Brown DS, Capell T, Figueras X and Tiburcio AF (1995) Rapid high-
performance liquid chromatography method for the quantitation of
polyamines as their dansyl derivatives: application to plant and animal
tissue. Journal of Chromatography B 666: 329-335
48. Martin ML, Martins HM and Bernando F (2001) Fumonisin B1 and B2 in
black tea and medicinal plants. Journal of Food Protection 64: 1268-1270
48
49. Miller JD (2001) Factors that affect the occurrence of fumonisin.

Environmental Health Perspectives 109: 321-324

50. Minocha R and Long SJ (2004) Simultaneous

separation and quantitation of
amino acids and polyamines of forest tree and cell cultures within a single
high-performance liquid chromatography run using dansyl chloride. Journal
of Chromatography A 1034 : 63-73
51. Missmer SA, Suarez L, Felkner M, Wang E, Merrill AH Jr, Rothman KJ and
Hendricks KA (2006) Exposure to fumonisins and the occurrence of neural
tube defects along the Texas-Mexico border. Environmental Health
Perspectives 114: 237-241
52. Ofitserova M, Siantar D, Nerkar S and Pickering M (2005) Multiple-Residue
Mycotoxin Analysis: Method abstract for post-column liquid
chromatography. Pickering Laboratories (Method abstract for post column
liquid chromatography)
53. Pestka JJ, Abouzied MN and Sutikno (1995) Immunological assay for
mycotoxin detection. Food Technology 49: 120-128
54. Plattner RD, Norred WP, Bacon CW, Voss KA, Peterson R, Shackelford DD
and Weisleder D (1990) A method for detection of fumonisins in corn
samples associated with field cases of leukoencephalomalacia. Mycologia
82: 698-702
49
55. Plattner RD and Branham EB (1994) Labelled fumonisins: Production and

use of fumonisin B1 containing stable isotopes. Journal of AOAC


56. Preis RA and Vargas EA (2000) A method for determining fumonisin B 1 in
corn using immunoaffinity column clean-up and thin layer
chromatography/densitometry. Food Additives and Contaminants 17: 463-
468
57. Rheeder JP, Marasas WFO, Thiel PG, Sydenham EW, Shephard GS and Van
Schalkwyk DJ (1992) Fusarium moniliforme and fumonisins in corn in
relation to human esophageal cancer in Transkei. Phytopathology 82 : 353-
357
58. Rheeder JP, Marasas WFO and Vismer HF (2002) Production of fumonisin
analogues by Fusarium species. Applied and Environmental Microbiology
68: 2101-2105
59. Rice LG, Ross PR and De Jong J (1995) Evaluation of a liquid
chromatography method for the determination of fumonisins in corn,
poultry feed and Fusarium culture material. Journal of AOAC International
78: 1002-1009
60. Romero-Gonzalez R, Garrido Frenich A, Martinez Vidal JL, Prestes OD and
Grio SL (2011) Simultaneous determination of pesticides, biopesticides and
mycotoxins in organic products a quick, easy, cheap, effective, rugged and
safe extraction procedure and ultra-high performance liquid
50
chromatography-tandem mass spectrometry. Journal of Chromatography A

1218: 1477-1485

61. Ross PF, Rice LG, Plattner GD, Osweiller

TM, Wilson DL, Owens HA, Nelson

JL and Richard JL (1991) Concentrations of fumonisin B1 in feeds associated
with animal health problems. Mycopathalogia 114: 129-135
62. Rottinghaus GE, Coatney CE and Minor HC (1992) A rapid, sensitive thin
layer chromatography procedure for detection of fumonisin B1 and B2,
Journal Veterinary Diagnostic Investigation 4: 326-329
63. Samapundo S, De Meulenaer B, De Muer N, Debevere J and Devlieghere F
(2006) Influence of experimental parameters on the fluorescence response
and recovery of the high-performance liquid chromatography analysis of
fumonisin B1. Journal of Chromatography A 1109: 312-316
65. Scott PM, Yeung JM, Lawrence GA and Prelusky DB (1997) Evaluation of
enzyme-linked immunosorbent assay for analysis of beer for fumonisins.
Food Additives and Contaminants 14: 445-450
66. Sewram V, Shephard GS, Marasas WFO and de Castro MFPM (2003)
Improving extraction of fumonisin mycotoxins from Brazilian corn-based
infant foods. Journal of Food Protection 66: 854-85
51
67. Shephard (2000) Liquid Chromatographic Method for fumonisins in corn,

methods in molecular biology. Mycotoxin Protocols 157: 147-158

68. Shephard GS, Thiel PG and Sydenham EW (1995) Liquid chromatographic

determination of the mycotoxin

fumonisin B2 in physiological samples.
Journal of Chromatography A 692: 39-43
70. Shephard (1998) Chromatographic determination of fumonisin mycotoxins.
Journal of Chromatography A 815: 31-39
71. Shephard (2000) Liquid chromatographic method for fumonisins in corn.
Methods in Molecular Biology 157: 147-158
72. Shephard GS and Sewram V (2004) Determination of the mycotoxin
fumonisin B1 in maize by reversed-phase thin-layer chromatography: a
collaborative study. Food Additives and Contaminants 21: 498-505
73. Shephard GS, Van der Westhuizen L, Gatyeni PM, Somdyala NIM, Burger
HM and Marasas WFO (2005) Fumonisin mycotoxins in traditional maize
beer in South Africa. Journal of Agricultural and Food Chemistry 53: 9634-
9637
74. Shephard GS, Marasas WFO, Burger HM, Somdyala NIM, Rheeder JP, van
der Westhuizen L, Gatyeni P and Van Schalkwyk DJ (2007) Exposure
assessment for fumonisins in the former Transkei region of South Africa.
Food Additives and Contaminants 24: 621-629
52
75. Shephard GS (2008) Determination of mycotoxins in human foods.

Chemical Society Reviews 37: 2468-2477

76. Shephard GS, Berthiller F, Burdaspal

P, Crews C, Jonker MA, Krska R,
C, Sabino M, Solfrzzo M, Van Egmond HP

MacDonald S, Malone B, Maragos
and Whitaker TB (2011) Developments in mycotoxin analysis: an update for
2009-2010. World Mycotoxin Journal 4: 3-28
77. Silva L, Fernandez-Frazon M, Font G, Pena A, Silveira I, Lino C and Manes J
(2009 a) Analysis of fumonisins in corn-based food by liquid
chromatography with fluorescence and mass spectrometry. Food
Chemistry 112: 1031-1037
78. Silva LJ, Lino CM and Pena A (2009 b) Sphinganine-sphingosine ratio in
urine form two Portuguese populations as biomarkers to fumonisins
exposure. Toxicology 54: 390-398
79. Stockenström S, Sydenham EW and Thiel PG (1994) Determination of
fumonisins in corn: Evaluation of two purification procedures. Mycotoxin
Research 10: 9-14
80. Stroka J, Capelletti C, Papadopoulou-Bouraoui A, Pallaroni L and Anklam E
(2002) Investigation of alternative reagents to 2-mercaptoethanol for the
pre-column derivatization of fumonisins with o-phthaldialdehyde, for HPLC
analysis. Journal of Liquid Chromatography and Related Technologies 25:
1821-1833
53
81. Sulyok M, Krska R and Schuhmacher R (2010) Application of an LC-MS/MS

based multi-mycotoxins method for the semi-quantitative determination of

mycotoxins occurring in different type of food infected by moulds. Food

Chemistry 119: 408-416
82. Sun G, Wang S, Hu XU, Su J, Huang T, Yu J, Tang L, Gao W and Wang JS
(2007) Fumonisin B1 contamination of home-grown corn in high-risk areas
for esophageal and liver cancer in China. Food Additives and Contaminants
24: 181-185
83. Sydenham EW, Gelderblom WCA, Thiel PG and Marasas WFO (1990)
Evidence for the natural occurrence of fumonisin B1, a mycotoxin produced
by Fusarium moniliforme in corn. Journal of Agricultural and Food
Chemistry 38: 285-3920
84. Sydenham EW, Shephard GS and Thiel PG (1992) Liquid chromatographic
determination of fumonisins B1, B2 and B3 in food and feeds. Journal of
AOAC International 75: 313-318
85. Sydenham EW, Shephard GS, Thiel PG, Stockenström S, Snijman PW and
Van Schalkwyk DJ (1996) Liquid chromatographic determination of
fumonisins B1, B2 and B3 in corn: AOAC-IUPAC collaborative study. Journal
86. Trebstein A, Seefelder W, Lauber U and Humf HU (2008) Determination of
T-2 and HT-2 toxins in cereals including oats after immunoaffinity clean up
54
by liquid chromatography and fluorescence detection. Journal of


87. Trucksess MW, Stack ME, Allen S and Barrion N (1995) Immunoaffinity

column coupled with liquid chromatography

for the determination of
fumonisin B1 in canned and frozen sweet corn. Journal of AOAC
88. Ueno Y, IIjima K, Wang SD, Suguira Y, Sekijime M, Tanaka T, Chen C and Yu
SZ (1997) Fumonisins as a possible contributory risk factor for primary liver
cancer: A 3-year study of corn harvested in Haimen China by HPLC and
ELISA. Food and Chemical Toxicology 35: 1143-1150
89. Van der Westhuizen L, Shephard GS, Scussel VM, Costa LLF, Vismer HF,
Rheeder JP and Marasas WFO (2003) Fumonisin contamination and
Fusarium incidence in corn from Santa Catarina, Brazil. Journal of
90. Van der Westhuizen L, Shephard GS, Rheeder JP, Somdyala NIM and
Marasas WFO (2008) Sphingoid base levels in humans consuming
fumonisin contaminated maize from rural areas in the former Transkei,
South Africa: A cross sectional study. Food Additives Contaminants Part A
25: 1385-1391
91. Van der Westhuizen L, Shephard GS, HM Burger HM Rheeder JP,
Gelderblom WCA, Wild CP, Gong YY (2011) Fumonisin B1 as a urinary
biomarker of exposure in a maize intervention study among South African
55
subsistence farmers. Cancer Epidemiology Biomarkers Prevention 20: 483–

489

92. Velazquez C, Llovera M, Plana J and Canela R (2000) Effect of solvents on

fumonisins analysis by high-performance liquid chromatography with
AccQ.Fluor as the derivatizing reagent. Journal of Chromatography A 870:
469-472
93. Visconti A, Boenke A, Solfrizzo M, Pascale M and Doko MB (1996) European
intercomparison study for the determination of fumonisin content in two
maize materials. Food Additives and Contaminants 13: 909-927
94. Walker JM (1994) The dansyl method for identifying N-terminal amino
acids. Methods of Molecular Biology 32: 321–328
95. Ware GM, Francis O, Kaun SS, Umrigar SSKP, Carmen A, Carter L and
Bennett GA (1993) Determination of fumonisin B1 in corn by high
performace liquid chromatography with fluorescence detection. Analytical
Letters 26: 1751-1760
96. Weidenborner M (2001) Foods and fumonisins. European Food Research
Technology 212: 262-273
97. Wilkes JG and Sutherland JB (1998) Sample preparation and high-resolution
separation of mycotoxins possessing carboxyl groups. Journal of
Chromatography B 717: 135-156
56
98. Williams LD, Meredith FI and Riley RT (2004) Fumonisin-ortho-

phthaldaldehyde derivative is stabilized a low temperature. Journal of


99. Wu F (2004) Mycotoxins risk assessment for the purpose of setting
international regulatory standards. Environmental Science and Technology
38 : 4049-4055
100. Zachariasova M, Lacina O, Malachova A, Kostelanska M, Poustka J, Godula
M and Hajslova J (2010) Novel approaches in analysis of Fusarium
mycotoxins in cereals employing ultra performance liquid chromatography
coupled with high resolution mass spectrometry. Analytica Chimica Acta
66: 51-61
57

CHAPTER 3
Ultraviolet (UV) detection of fumonisin B

analogues as OPA derivatives
3.1 Introduction

Determination of fumonisins in maize is widely achieved using a validated

method involving methanol : water extraction followed by strong anion

exchange (SAX) clean-up and derivatization

prior to separation by reversed-
phase high-performance liquid chromatography (RP-HPLC) with fluorescence
detection (Sydenham et al. 1996). As fumonisins lack a useful chromophore or
fluorophore, HPLC detection is achieved by suitable derivatization (Shephard et
al., 1990). Although a number of fluorogenic derivatizing reagents have been
investigated, OPA remains widely used for sensitive and specific analysis of
fumonisins as it yields strongly fluorescent derivatives easily separated by HPLC
(Shephard, 1998).
On occasion, laboratories equipped with HPLC and ultraviolet (UV) detectors
seek to undertake limited fumonisin determinations without the purchase of
further instrumentation in the form of a fluorescence detector. The aim of the
study reported in this chapter was to investigate the degree to which the
fumonisins could be determined using UV detection of the fumonisin-OPA
derivatives.
3.2 Materials and Methods
3.2.1 Chemicals
All chemicals used were of analytical grade. Methanol, acetone, sodium
hydrogen carbonate (NaHCO3), acetonitrile, o-phosphoric acid (H3PO4),
potassium dihydrogen phosphate (KH2PO4), sodium hydroxide (NaOH), o-
59
phthaldialdehyde (OPA), disodium tetraborate (Na2B4O7.10H2O), disodium

hydrogen phosphate (Na2HPO4.2H2O), potassium chloride (KCl), sodium

dihydrogen phosphate (NaH2PO4), hydrochloric acid (HCl), sodium chloride

(NaCl), 2-mercaptoethanol (ME) were

purchased from Merck. Phosphate
buffered saline (PBS) was prepared by dissolving 8.0 g sodium chloride, 1.2 g
disodium hydrogen phosphate, 0.2 g potassium dihydrogen phosphate and 0.2 g
potassium chloride in a litre distilled water. The pH was adjusted to 7. The OPA
derivatization reagent was prepared by dissolving OPA (40 mg) in 1 mL methanol
and adding 0.1 M sodium tetraborate (5 mL) solution, 2- mercaptoethanol (50
µL) and vortexing after each solvent addition. The OPA derivatization reagent
was used up to 7 days following preparation.
3.2.2 Fumonisin Standards
FB1, FB2 and FB3 standards were isolated at the PROMEC Unit according to the
method of Cawood et al. (1991). The FB3 standard contains approximately 21 -
42% epi-FB3 (Gelderblom et al., 2007). The fumonisin working standards were
prepared by diluting a stock with concentration levels of 245 µg/mL, 200 µg/mL,
270 µg/mL of FB1, FB2 and FB3, respectively, with acetonitrile : water (1:1, v/v) to
obtain working standard with 55.13 µg/mL, 25.00 µg/mL and 13.25 µg/mL for
FB1, FB2 and FB3, respectively.
60
3.2.3 Maize Samples

Home-grown maize samples intended for human consumption were collected in
the rural former Transkei area of the Eastern Cape Province following the 2006

harvest and stored at a temperature (4°C) where fumonisins are stable. The well
mixed samples were milled prior to analysis.
3.2.4 Extraction using strong anion exchange (SAX) clean-up
Maize samples were extracted using the method of Sydenham et al. (1996) with
modifications. A milled maize sample (20 g) was extracted by blending in a
homogenizer (Polytron PT 3100, Kinematica, Luzerne, Switzerland) homogenized
for 3 min with methanol : water (3:1, v/v; 10 mL). It was then centrifuged at 4°C
for 10 min at 500 x g. The supernatant was filtered using a MN 617 (185 mm)
filter paper and the pH adjusted to 5.80-6.25 with 1 M NaOH or 1 M HCl. After
centrifugation, an aliquot (10 mL) of the supernatant was cleaned-up using SAX
cartridges (10 mL, 500 mg packing Bond-Elut, Varian, Harbor City, CA, USA),
which were preconditioned with 5 mL methanol followed by another 5 mL
methanol : water (3:1) (flow rate ≤ 2 mL / min, no air was forced through the
column; the column was not allowed to dry through-out the entire clean-up
process). The extracted sample (10 mL) was loaded on the SAX and washed with
5 mL methanol : water (3:1) and 3 mL methanol. After washing with methanol,
the fumonisins were eluted with acetic acid : methanol (1:99, v/v 10 mL) under
gravity. The eluate was evaporated to dryness under nitrogen at < 60°C and
stored at 4°C prior to analysis.
61
3.2.5 Extraction using immunoaffinity columns (IAC) clean-up

Homogeneously mixed samples (20 g) were extracted with 50 mL of the
extraction solvent (acetonitrile + methanol

+ distilled water, 25 + 25 + 50, v + v +
v) by shaking on an orbital shaker for 20 min. The extraction solution was
centrifuged for 10 min at 500 x g and the supernatant was filtered as described
above to avoid the transfer of any solid material. The remaining solid material
was re-extracted with an additional 50 mL of solvent in the same manner as
described above. The filtered supernants was combined and a 10 mL aliquot was
diluted with 40 mL Phosphate buffered saline (PBS) and the solution mixed well.
3.2.6 IAC cleanup
The PBS diluted extract (10 mL) was passed through the FumoniTest column
(Watertown, MA, USA) at a flow rate of 1 to 2 drops per second and the eluate
discarded. The column was washed with 10 mL PBS until air came through the
column and the eluate discarded. The fumonisins were eluated with 2.5 mL of
HPLC grade methanol, at a rate of 1 drop per second. The eluate was evaporated
at 60°C using nitrogen gas and stored at 4°C prior to analysis.
Standards (25 µL) with concentrations of 55.25 µg/mL for FB1, 25.00 µg/mL and
13.25 µg/mL for FB2 and FB3, respectively, were derivatized with 225 µL OPA
reagent and 10 µL injected. The nitrogen dried samples were re-dissolved in 200
62
µL methanol and 25 µL was derivatized with OPA (75 µL), vortexed and 20 µL

injected onto the HPLC exactly 2 min after mixing.

3.3 Chromatography

RP-HPLC was performed on an Agilent Technologies (Wildbronn, Germany) 1260
Infinity pump, Rheodyne 7725i injector and a Phenomenex (Torrance, CA, USA)
Luna C18 5 µm column (150 mm x 4.60 mm). The HPLC instrument was
configured with an Agilent 470 (Waldbronn, Germany) 1100 series diode array
detector (DAD) and an Agilent 1100 series fluorescence detector (FLD) connected
in series. The sequence of the detectors was the DAD first followed by the FLD
(to prevent overpressure, the fluorescence detector should always be the last
module in the flow system). The mobile phase of methanol : 0.1 M sodium
phosphate (77:23, adjusted to pH 3.35 with o-phosphoric acid) was pumped at
an isocratic flow rate of 1 mL/min. Data was collected and analyzed by Agilent
ChemStation software and quantification achieved by comparison of peak areas
with those of authentic fumonisin standards.
3.4 Results and Discussion
3.4.1 Peak resolution
As reliability of data derived from a chromatographic analysis depends on the
effective separation of the analytes of interest from one another and from
additional matrix components, peak resolution was analysed. Complete
resolution of the OPA derivatives was achieved with RP-HPLC C18 column and an
63
isocratic mobile phase of methanol : 0.1 M NaH2PO4 (77:23) in less than 20

minutes. FB-OPA derivatives are generally monitored at excitation wavelengths

335 nm and emission wavelengths 440 nm for fluorescence detection (Shephard,

1998) which are wavelengths used in the

study for FLD detection.
These wavelengths were tested using fumonisin working standards employing
the chromatographic conditions above. Based on the high sensitivity of the DAD
at 335 nm it was selected as wavelength for DAD detection. An iso-absorbance
plot (software programme which displays chromatographic details in 3D
including retention time versus wavelength, from which optimum wavelength
can be selected) was then used to confirm the wavelength selection for DAD.
Injection of fumonisin working standards with concentrations of 55.25 µg/mL for
FB1, 25.00 µg/mL and 13.25 µg/mL for FB2 and FB3 respectively resulted in an
elution order of FB1, FB3 and then FB2. Retention times for FB1, FB2 and FB3 were
4.6, 12.2 and 10.8 min (± 5 %), respectively, for both FLD and DAD (Figures 3.1
and 3.2). Another fumonisin elutes just before FB3 and is only partially separated
from it. This compound has been identified as an epimer of FB 3 (epi-FB3) and is
quantified as part of FB3 (Gelderblom et al., 2007). The two isomers have similar
chemical properties and exhibit similar retention times when analysed on a
reversed-phase HPLC column. To assume the two isomers have the same
chromatographic response factors appears to be reasonable and allows accurate
analysis to be performed.
64
To test for matrix interferences, reagent blanks were analysed and the resultant

chromatogram overlaid with that of fumonisin standards. No interferences were

observed and the background noise of both the blanks and standards were found

to be insignificant for both detectors; therefore no form of baseline correction
was necessary. Satisfactory resolution of the fumonisin analogues was achieved
with analytes identified using their retention times. Chromatographic resolution
of the peaks in the study compared well with previous studies (Shephard et al.,
1990).
65

Figure 3.1 Chromatogram of fumonisin working standard detected by FLD
66

Figure 3.2 Chromatogram of fumonisin working standard detected by DAD
67
3.4.2 Method precision

Precision of the FB-OPA method was determined in terms of intra- and inter-day

analysis. The intra-day precision (daily)
was obtained by injecting three
consecutive working standards and the inter-day precision (day-to-day) was
measured over a period of five days. These parameters were determined to
ensure both the repeatability and reproducibility of the standard preparation
and derivatization is within acceptable variances (i.e. RSD values for intra-day ≤ 5
%; inter-day ≤ 20 %). Based on the results of the intra-day analysis (Table 3.1),
the method’s repeatability was good with RSD for the FLD ≤ 1 % and ≤ 3 % for
DAD, indicating good precision for the OPA method. The method precision gives
an estimation of the variability that can be expected when performing fumonisin
analysis using OPA.
Table 3.1 Intra-day precision of fumonisin working standards (n=3) using
standard peak areas
FLD DAD
FB1 FB2 FB3 FB1 FB2 FB3
Standard 1 35153 13924 11715 16.5 6.7 5.2
Standard 2 35047 13909 11688 16.6 7.0 5.1
Standard 3 35635 14073 11905 17.0 6.6 5.3
Mean 35279 13968 11769 16.7 6.8 5.2
Stdev 313 90 119 0.3 0.2 0.1
RSD (%) 0.9 0.7 1.0 1.6 2.9 1.6
68
The intermediate precision of the method is reflected in the inter-day results

(Table 3.2); the method shows acceptable precision with RSD values ≤ 13 % for

both FLD and DAD which is suitable because inter-day RSD values should be ≤ 20

%. The method is generally reproducible

in terms of standard preparation,
derivatization and injection. These results indicate the OPA method to be
reproducible, precise and repeatable. Based on the intra- and inter-day precision
results, OPA provides adequate precision for fumonisin analysis.
Table 3.2 Inter-day precision of fumonisin working standards using standard
peak areas, (n=18, 3 x std injected/day)
FLD DAD
Day 1 30349 13563 10343 20.0 9.1 7.0
30433 13443 10229 20.2 8.2 5.6
31772 13855 10332 21.4 9.4 7.2
Day 2 32164 13155 10854 20.7 8.8 6.2
32828 13507 10971 19.5 8.5 6.5
31717 13067 10771 19.9 8.3 5.9
Day 3 24749 10134 8248 15.4 7.1 5.4
23923 9811 8044 15.7 8.0 5.5
23915 9814 8590 15.3 6.9 5.2
Day 4 31630 14336 10964 16.5 6.7 5.2
32099 14028 10924 16.6 7.0 5.1
31745 14452 11175 17.0 6.6 5.3
Day 5 35153 13924 11715 16.2 7.1 6.1
35047 13909 11688 16.5 6.9 5.6
35635 14073 11905 16.3 6.4 5.3
Mean 30877 13005 10450 17.8 7.7 5.8
Stdev 3808 1644 1224 2.1 1.0 0.7
RSD (%) 12 13 12 12 13 11
69
3.4.3 Detection limits

The limit of detection (LOD) and quantification (LOQ) were estimated from the

signal-to-noise ratio. The LOQ was obtained at 10:1 signal-to-noise ratio and 3:1

signal-to-noise ratio was used for the LOD. The FB1 analogue was more reliably
detected with lower detection limits (Table 3.3) compared to FB2 and FB3. FLD is
more sensitive than DAD, this is evident in the reduced detector noise (Figures
3.1 and 3.2) and enhanced analyte signal of the FLD (1700 FU) compared to the
DAD (0.8 mAU). The detection limits of the FLD are estimated to be
approximately 20-times more sensitive than DAD when one considers the
amount injected into the column. It is thus reasonable that based on the
detection limits obtained; the method will be able to detect fumonisins at the
concentration levels routinely encountered in contaminated maize with
adequate accuracy and sensitivity.
Table 3.3 Determination of the detection limits using the amount (ng)
injected into the column
Amount (ng) injected into the HPLC column

FB1 FB2 FB3
LOD (s:n=3) FLD 1.0 1.2 1.2
DAD 22 20 20
LOQ (s:n=10) FLD 2.3 2.7 2.7
DAD 37 48 56
70
3.4.4 Recoveries

Recoveries were determined by spiking maize samples with 400 µL fumonisin

working standards containing 1103, 500 and 270 µg/kg FB1, FB2 and FB3,

respectively. The analysis was repeated

six times for each concentration level
and the recoveries determined using IAC clean-up. Since the OPA method has
been validated both in-house (Shephard et al., 1990) and internationally
(Sydenham et al., 1996), SAX recoveries were confirmed on only two samples
and gave recoveries of up to 75 % as previously obtained by Sydenham et al.,
1996. Consequently, further testing of SAX clean-up recoveries was found to be
unnecessary.
“True analytical blanks” for fumonisin analysis are difficult to find as fumonisins
occur naturally in contaminated maize samples. In this regard, even good
commercial maize can have low levels of fumonisins. Therefore, a sample with
very low fumonisin levels was used as a blank and the fumonisin levels detected
were accounted for in the calculation of the recoveries.
The recoveries measured were lower than expected (Table 3.4) since IAC has
been reported to produce results of between 70 – 95 % as previously reported by
Visconti et al., 2001. Comparison of the FLD and DAD was acceptable for FB1;
however for FB2 the recoveries were lower than expected for DAD. The
recoveries obtained indicate that the method is acceptable for fumonisin analysis
in naturally contaminated maize samples.
71
Table 3.4 Recoveries with IAC clean-up

FLD DAD
(µg/kg) (µg/kg) (µg/kg) (µg/kg) (µg/kg) (µg/kg)

Blank 1 105 20 12 0 0 0
Blank 2 73 18 8 0 0 0
Mean 89 19 10 0 0 0
Sample 1 963 242 154 727 233 104
Sample 2 935 326 212 982 420 272
Sample 3 783 277 180 796 294 173
Sample 4 770 301 229 784 320 190
Sample 5 807 305 287 919 362 252
Sample 6 683 398 218 730 224 97
Mean 824 308 213 823 309 181
Stdev 106 52 46 105 75 73
RSD (%) 13 17 21 13 24 40
Spiked 1103 500 270 1103 500 270
Recovery (%) 67 58 75 67 46 62
3.5 Method application
The derivatization method was successfully applied to 15 maize samples using
SAX and IAC clean-up. Tables 3.5 and 3.6 provide a comprehensive view of the
total fumonisin (µg/kg) levels in the maize samples analyzed. In the study all the
samples analyzed were contaminated with FB1, FB2 and FB3. IAC clean-up results
were not comparable at levels below 500 µg/kg between the two detectors with
DAD detecting only at levels above 140 µg/kg for total fumonisins. SAX clean-up
provided comparable results at all levels analysed. The results obtained were
consistent whether determined by SAX or IAC with both clean-up methods in
72
terms of comparison of the detectors (Figures 3.3 and 4.4) signifying that OPA is

a robust derivatization reagent for fumonisin analysis in terms of clean-up.

Table 3.5 Fumonisin levels (µg/kg)

in naturally contaminated maize samples
cleaned-up with SAX
FLD DAD
Sample FB1 FB2 FB3 Total FB1 FB2 FB3 Total

*
1 70 21 16 106 102 nd nd 102
2 162 60 12 234 253 57 nd 310
3 195 74 18 288 176 81 nd 257
4 229 55 13 296 320 48 nd 368
5 153 56 11 220 239 61 nd 300
6 724 320 87 1132 948 331 nd 1279
7 886 382 107 1375 1025 399 46 1470
8 1829 1076 239 3144 2018 1195 21 3234
9 1431 566 166 2163 1708 517 64 2289
10 1717 614 206 2537 1924 669 212 2805
11 1336 432 131 1900 1362 452 133 1948
12 229 55 13 3120 2053 865 246 3164
13 1083 377 118 1577 1316 410 86 1812
14 3189 2086 812 6088 3417 2144 553 6114
15 2368 1034 337 3740 2551 1280 612 4442
*nd- Not detectable
73
Table 3.6 Fumonisin levels (µg/kg) in naturally contaminated maize

samples cleaned-up with IAC

FLD DAD

*
1 74 22 7 103 nd nd nd nd
2 134 56 48 238 142 nd nd 142
3 181 59 20 260 143 nd nd 143
4 120 29 8 157 nd nd nd nd
5 191 57 16 263 161 nd nd 161
6 496 185 20 701 491 218 nd 709
7 1132 405 86 1623 1199 391 nd 1590
8 2424 1112 193 3730 2524 1174 106 3804
9 1186 390 134 1711 1223 362 nd 1585
10 1593 436 166 2196 1653 403 175 2231
11 1294 306 111 1711 1377 311 nd 1688
12 3313 1087 289 4689 3371 1154 190 4716
13 1124 260 101 1485 1157 289 nd 1445
14 3919 1032 376 5327 4003 951 291 5246
15 2617 904 373 3894 2753 826 176 3756
*
nd- Not detectable
74

Figure 3.3 Comparison of FLD and DAD using SAX celan-up, Total Fumonsins = FB1 + FB2 + FB3
75

Figure 3.4 Comparison of FLD and DAD using IAC celan-up, Total Fumonsins = FB1 + FB2 + FB3
76
3.6 Conclusion

For chromatographic separation of the
fumonisin analogues, OPA was found to
be a very effective derivatization reagent

providing comparable results for both

FLD and DAD. During the study, the detectors were connected in series in order
to allow for comparison of the detectors without any variations produced by
repeat derivatization and injection. Although most sensitive fumonisin analysis
has been done with the use of fluorescence detection, this work indicates that
fumonisins in maize can be comparably determined by UV detection.
Simultaneous detection with FLD and DAD shows the FLD to more sensitive then
DAD. However, the two detectors can be used as alternatives to each other for
maize samples above 500 µg/kg following IAC clean-up compared to SAX clean-
up which provided comparable results at all levels with both detectors. Based on
the results obtained and the Official Analytical Chemists International (AOAC)
adopting the OPA derivatization of fumonisins as the standard HPLC technique
(Wilkes et al., 1998; Sydenham et al., 1996); the method will be used as the
reference method in the other studies in the thesis.
77
References

1. Cawood ME, Gelderblom WCA, Vleggaar R, Behrend Y, Thiel PG and

Marasas WFO (1991) Isolation of fumonisin mycotoxins: A quantitative
approach. Journal of Agriculture and Food Chemistry 39: 1985-196
2. Gelderblom WCA, Sewram V, Shephard GS, Snijman PW, Tenza K, van der
Westhuizen L and Vleggaar R (2007) Structure and natural occurrence of
stereoisomers of the fumonisin B series mycotoxins. Journal of Agricultural
and Food Chemistry 55: 4388-4394
3. Shephard GS, Sydenham EW, Thiel PG and Gelderblom WCA (1990)
Quantitative determination of fumonisins B1 and B2 by high performance
liquid chromatography with fluorescence detection. Journal of liquid
chromatography 13: 2077-2087
5. Sydenham EW, Shephard GS, Thiel PG, Stockenstrom S, Snijman PW and
van Schalkwyk DS (1996) Liquid chromatographic determination of
fumonisins B1, B2 and B3 in corn: AOAC-IUPAC Collaborative Study. Journal
6. Visconti A, Solfrizzo M and De Girolamo A (2001) Determination of
fumonisin B1 and B2 in corn and corn flakes by liquid chromatography with
immunoaffinity columns: collaborative study. Journal of AOAC
78
7. Wilkes JG and Sutherland JB (1998) Sample preparation and high-resolution

separation of mycotoxins possessing carboxyl groups. Journal of


79

CHAPTER 4
Optimization of naphthalene-2,3-dicarboxaldehyde
(NDA) derivatization reagent for fumonisin
derivatization and its applicability to fluorescence
(FLD) and ultraviolet (UV) detection
4.1 Introduction

In chapter 2, reviews of the current methods used for naphthalene-2,3-

dicarboxaldehyde (NDA) derivatization of fumonisins were described. Most

researchers report using similar reaction

conditions viz., sodium borate buffer pH
9, NDA dissolved in acetonitrile or methanol and chromatographic conditions,
with HPLC and LC-MS being the most commonly used techniques (Ware et al.,
1993, Bennett et al., 1994, Silva et al., 2009). The reaction of NDA with the
nucleophilic cyanide anion forms stable and highly fluorescent derivatives (Cho
et al., 2002) and reacts with the primary amine moiety of fumonisin B 1 to form
an N-substituted 1-cynaobenz[f]isoindole derivative (Bennett and Richard, 1994).
Studies reported in this chapter are aimed at methodology development to
optimize both NDA derivatization reaction and instrumentation conditions for
determination of FB1, FB2 and FB3 in maize. Initial conditions used were based on
the method by Scott and Lawrence (1992) with some modifications. Thus the
essential aims were to optimize NDA derivatization for fumonisin determination
in naturally contaminated maize following strong anion exchange (SAX) or
immunoaffinity column (IAC) clean-up, utilizing diode array detection (DAD) as a
practical alternative to fluorescence detection (FLD).
81

4.2.1 Chemicals

All chemicals used were of analytical grade. Methanol, acetone, sodium

hydrogen carbonate (NaHCO3), acetonitrile, o-phosphoric acid (H3PO4),
potassium dihydrogen phosphate (KH2PO4), sodium hydroxide (NaOH), disodium
tetraborate (Na2B4O7.10 H2O), disodium hydrogen phosphate (Na2HPO4.2H2O),
potassium chloride (KCl), sodium dihydrogen phosphate (NaH2PO4), hydrochloric
acid (HCl), sodium chloride (NaCl) were purchased from Merck. Napthalene-2,3-
dicarboxaldehyde (NDA) was purchased from Invitrogen (Molecular Probes). NDA
was prepared by dissolving 4 mg NDA in 8 mL methanol. Phosphate buffered
saline (PBS) was prepared by dissolving 8.0 g sodium chloride, 1.2 g disodium
hydrogen phosphate, 0.2 g potassium dihydrogen phosphate and 0.2 g
potassium chloride in a litre distilled water. The pH was adjusted to 7 with o-
phosphoric acid.
The standards were obtained and prepared as described in Section 3.2.2.
4.2.3 Maize Samples
Home-grown maize samples intended for human consumption were collected in
the rural former Transkei area of the Eastern Cape Province following the 2006
harvest and stored at a temperature (4°C) where fumonisins are stable. The well
mixed samples were milled prior to analysis.
82

Maize samples were extracted using the
method of Sydenham et al. (1996) with
modifications as described in Section 3.2.4.

Extraction of maize samples for IAC clean-up was performed as per
manufacturer’s instructions with minor modifications as described in Section
3.2.5.
4.2.6 IAC cleanup
IAC clean-up was performed as per manufacturer’s instructions with minor
modifications as described in Section 3.2.6.
Standards: Working standard (20 µL) was used, 20 µL of borate buffer (0.1 M)
was added followed by 20 µL potassium cyanide (65.13 mg/100 mL distilled
water) and 40 µL NDA (4 mg/8 mL methanol). The solution was vortexed and
heated at 60°C for 15 minutes, then cooled to 24°C under running tap water.
Mobile phase (100 µL) was then used to dilute the solution and 20 µL injected
into the HPLC system.
Samples: The samples were reconstituted in 200 µL methanol and 100 µL of
borate buffer (0.1 M) was added followed by 100 µL potassium cyanide (65.13
g/100 mL distilled water) and 200 µL NDA (4 mg/8 mL methanol). The solution
83
was vortexed, heated at 60°C for 15 minutes and cooled to 24°C under running

tap water. 500 µL mobile phase was then used to dilute the solution and 20 µL

injected into the HPLC system.

4.2.8 Recoveries
The maize samples were spiked with fumonisin working standards (40 µL)
directly onto the dry milled maize samples. Since maize without fumonisin was
not available, the unspiked maize samples were analyzed for fumonisins and
these unspiked levels were taken into account for the calculation of the
recoveries.
4.3 Chromatography
RP-HPLC was performed on an Agilent Technologies (Wildbronn, Germany) 1260
Infinity pump, Rheodyne 7725i injector and a Phenomenex (Torrance, CA, USA)
Luna C18 5 µm column (150 mm x 4.60 mm) which was configured as described
in Section 3.3. The mobile phase was prepared by combining methanol (780 mL) :
0.1 M sodium phosphate (NaH2PO4) (220 mL), and the pH of the mixture was
adjusted to pH 3.35 with o-phosphoric acid. The mobile phase was filtered using
0.45 µm X 47mm filter paper with vacuum and pumped at 1 mL/min flow rate.
Data was collected and analyzed by Agilent ChemStation software and
quantification was achieved by comparison of peak areas with those of authentic
fumonisin standards.
84


Figures 4.1 and 4.2 are chromatograms obtained from a 10 µL injection of a

fumonisin working standard derivatized with NDA at levels of 55.13 µg/mL, 25.00
µg/mL and 13.25 µg/mL for FB1, FB2 and FB3 respectively using FLD and DAD
detection. Analytical separation of the FB-NDA derivatives was performed using
isocratic elution with methanol : NaH2PO4 (78:22) mobile phase at 1 mL/min flow
rate. The NDA derivatives were separated with retention times for FB1, FB2 and
FB3 at 6.31, 14.87 and 11.87 (±5 %) min respectively. An examination of the FLD
and DAD chromatograms revealed that a number of other peaks (labelled A and
B) in addition to the analyte peaks were present.
A reagent blank was prepared to test the interference of these peaks; the
resultant chromatogram obtained from the reagent blank was overlaid with that
of the standard chromatogram. All the peaks except for analyte peaks were
present in the reagent blank chromatogram. This led to the conclusion that the
peaks are from the reagents and do not interfere with the quantification or
resolution of the analytes. Furthermore, Lillard et al., 1998 also observed peak A
as a reagent peak during separation of amine NDA derivatives and suggested
that it was formed as a by product of the benzoin condensation (Roach et al.,
1987) or other side products (Kwakman et al., 1990) of NDA upon exposure to
aqueous buffer conditions. Since the peak is present in completely independent
experiments it was ruled out as a laboratory contaminant.
85

Figure 4.1 Chromatogram of fumonisins working standard detected by FLD

86

Figure 4.2 Chromatogram of fumonisin working standard detected by DAD
87

Reproducibility was determined by
measuring the intra- and inter-day
repeatability of the working standards.

The intra-day was measured by injecting
three working standards in one day and the inter-day was measured over a
period of five days. The excellent intra-day (Table 4.1) results obtained for FLD
and DAD demonstrates the precision of the derivatization method. Comparison
of the results across the two detectors suggests that the FLD has better
repeatabilities with RSD values ≤ 2 %. Inter-day (Table 4.2) results of both
detectors are higher than their corresponding intra-day results. However the
inter-day results are still acceptable with both detectors reporting RSD values ≤ 8
% at n = 15 which is below the maximum standard value of 20 %.
Table 4.1 Intra-day precision of fumonisin working standards (n=3) using
reported as peak areas
FLD DAD
Standard 1 99819 55625 34884 456 247 151
Standard 2 99512 55075 35001 453 237 154
Standard 3 101148 57079 35005 441 241 159
Mean 100160 55926 34964 450 242 154
Stdev 870 1035 69 8 5 4
RSD (%) 0.9 1.9 0.2 1.8 2.1 2.7
88
The precision of the method is acceptable and demonstrates both the

reproducibility and repeatability of the derivatization method to be satisfactory.

Furthermore, it allows for the use of standard peak areas for quantitation during

method optimization. The precision

results are in accordance with the
performance characteristics of FB1 and FB2 as regulated by the Commission
Directive of the European Commission (EC, 2005).
Table 4.2 Inter-day precision of fumonisin working standards (n=15)
reported as peak areas
FLD DAD
Day 1 114205 53905 32857 482 253 133
110123 52786 33105 486 253 140
108985 52055 32996 510 263 140
Day 2 129982 61197 34623 495 239 140
126829 60671 34852 481 231 140
128419 60910 35430 479 228 138
Day 3 127202 61118 31727 545 281 160
122601 58140 29928 547 285 160
132047 63281 35671 555 278 165
Day 4 124543 63407 30718 505 278 136
130530 66972 31294 528 287 142
126289 64486 32361 550 267 155
Day 5 109950 61791 33012 524 273 145
107691 59853 32758 545 297 140
110155 61122 31310 523 277 139
Mean 120637 60113 32843 517 266 145
Stdev 9211 4269 1717 28 21 10
RSD (%) 8 7 5 5 8 8
89
4.4.3 Detection Limits

The limit of detection (LOD) was calculated as the amount of analyte injected

resulting in peak heights of three times the maximum noise height whereas the

limit of quantification (LOQ) was calculated

as the amount of analyte injected
giving a peak height ten times the maximum noise peak height. The NDA
detection limits indicate that the FLD is 10-times more sensitive than that of the
DAD with fumonisin standards (Table 4.3) and 100-times more sensitive with
naturally contaminated samples (Table 4.4). Although FB1 has the lowest LOD, it
may be stated that the LODs of the FB2 and FB3 analogues are also satisfactory.
Table 4.3 Amount (ng) injected into HPLC Column for standards
FB1 FB2 FB3
LOD (s:n=3) FLD 0.11 0.50 0.27

DAD 13.78 12.5 6.63
LOQ (s:n=10) FLD 11.03 25.00 13.25
DAD 55.13 25.00 27.00
Table 4.4 Limits of detection (LOD) and quantification (LOQ) expressed as levels
in sample (µg/kg) following IAC clean-up
FB1 FB2 FB3
LOD (s:n=3) FLD 0.004 0.03 0.08
DAD 0.3 170 180
LOQ (s:n=10) FLD 0.03 0.1 0.4
DAD 3 300 350
90
4.4.4 Stability of NDA derivatives

Stability tests were performed by treating both standards and samples to the

same environmental conditions over a period of three days. Initial experiments

showed a monotonic decrease in response

(approximately 10 %) over an 8 hour
period when derivatized samples were left at room temperature. This was
overcome by storing both standards and samples at - 22°C after derivatization.
Improvement in both the repeatability and precision of the method (Table 4.4)
was observed as the derivatives under these conditions remained stable after six
consecutive injections (~ 120 min) of the same standard.
NDA stability was further evaluated by storing derivatized standards and
derivatized maize extracts at -22°C over three consecutive days. The FB-NDA
responses (Figure 4.3 A) were stable for 24 hours; after which a decrease in
detection (approximately 10 %) was observed in the FLD response. In contrast,
the DAD response (Figure 4.3 B) on day 2 and 3 apparently increased 10 % over
day 1. Previous reports described increases in FB-NDA response after 24 hours
(Bennett et al., 1994, Lino et al., 2006). These results suggest that NDA
derivatives are suitable for auto-injection or over-night analysis with working
standards being injected between samples to allow for better quantification and
to accommodate any derivative instability.
91
Table 4.5 Stability of FB-NDA standard after six consecutive injections (~ 120 min)

FLD DAD

Injection 1 63445 37292 9452 256 145 36
Injection 2 70098 41288 10680
289 171 45
Injection 3 67623 39935 10385 276 157 50
Injection 4 66881 39354 10274 293 160 44
Injection 5 63322 36994 9738 255 149 37
Injection 6 64251 37651 9952 262 161 41
Mean 65937 38752 10080 272 157 42
Stdev 2718 1710 451 17 9 5
RSD (%) 4 4 4 6 6 12
Figure 4.3 Stability of FB-NDA derivatives (µg/kg) over period of three days
(72 hours). Results calculated from mean (n=6) ± standard
deviation. (A) Stability of FLD, (B) stability of DAD
92
4.5 Robustness

4.5.1 Optimization of mobile phase

Different organic solvents used as HPLC mobile phase components were

examined for their suitability to provide the shortest run time without
compromising on the resolution of the closely eluting peaks (FB2 and FB3). Figure
4.4 shows the effect change in mobile phase composition has on the retention
times of the analytes. From the graphic representation, 77 % and 80 % methanol
composition causes interferences between FB1 with peaks A and B, and FB3 with
peak B. Selection of mobile phase was consequently influenced by the separation
of peaks A and B from analyte peaks.
Figure 4.4 Effect of methanol concentration on retention time and

interference with reagent peaks
93
Most studies on NDA use acetonitrile as an organic solvent in their mobile phase

because it provides significantly lower retention times and more sensitive results

(Lino et al., 2007). Acetonitrile was found to be the best solvent since it provided

an excellent baseline resolution coupled

with short analysis time.
The results suggested that the elution order of the FB-NDA derivatives is
dependent on the type of mobile phase used as the relative retention of the
analytes were altered with the use of acetonitrile. Methanol : 0.1 M NaH 2PO4
(78:22) mobile phase provides an elution order of FB1, FB3 and FB2. However in
acetonitrile : water : acetic acid (65:35:1) as eluent, the elution order of FB 2 with
FB3 and FB3 with its isomer epi-FB3 was interchanged, a phenomenon which has
not previously been reported. Finally, further method optimization allowed for
the development of a methanol mixture as the HPLC mobile phase component of
choice since it not only yielded comparable results to acetonitrile but is cheaper
than acetonitrile which was unavailable during the period of the study.
From the different proportions of methanol and 0.1 M NaH 2PO4, the best
separation of all the fumonisin B analogues within the shortest analysis time was
obtained with mobile phase methanol : 0.1 M NaH2PO4 (78:22, pH 3.35). This was
consequently selected as the mobile phase of choice for the chromatographic
separation of NDA derivatives.
94
4.5.2 Optimization of wavelength

FB-NDA derivatives are generally monitored
at excitation wavelengths of 420 and
246 nm and emission wavelengths of 500 and 418 nm (Bennett et al., 1994; De

Montigny et al., 1987). These wavelengths were tested using fumonisin working
standards. Based on the high sensitivity of the FLD at excitation 420 nm and
emission 500 nm; these were selected as optimum wavelengths. The DAD
absorption wavelengths (248 nm, 250 nm, 252 nm, 256 nm,) were examined and
252 nm in our hands provided the best sensitivity (Figure 4.5). An iso-absorbance
plot (software programme which displays chromatographic details in 3D
including retention time versus wavelength, from which optimum wavelength
can be selected) was then used to confirm the wavelength selection for DAD.
Figure 4.5 Optimum wavelength selections for UV absorbance relative to

peak area, results calculated from mean (n=3)
95
4.5.3 Reaction time

Figures 4.6 A and B demonstrate the effect heating time has on standard peak
areas. The results are based on five standards

heated for 15 and 30 minutes at
60°C. Both the FLD and DAD showed a decrease in standard area for FB 1 and FB2
when the reaction time was increased from 15 minutes to 30 minutes. FB 3 was
not affected by change in the reaction time as it remained constant during
heating. Since FB1 and FB2 produced higher standard area when heated for 15
minutes and the fumonisins B analogues are analysed simultaneously, the
reaction time for the derivatization of the fumonisins with NDA was selected to
be 15 minutes.
Figure 4.6 Effect of reaction time on peak areas. Results reported as mean
(n=5) ± standard deviation. (A) Comparison of FLD, (B) Comparison
of DAD
96
4.5.4 Effect of reaction temperature

To determine the effect of reaction temperature,
experiments were performed
at two different temperatures (24°C

and 60°C) using fumonisin working

standards. Temperature estimations were selected from literature (Lamba et al.,
2008; Scott and Lawrence 1992) with all experiments performed for 15 minutes.
The FB-NDA derivatization reaction was found to be temperature dependent
(Figures 4.7 and 4.8) as the responses increased with increase in temperature.
Performing the experiments at room temperature (24°C) resulted in half the
reaction efficiency compared to 60°C for both the FLD and DAD. Due to increase
in standard peak area when the reaction was heated for 60°C and the improved
RSD values, 60°C was selected as reaction temperature for NDA derivatization.
Figure 4.7 Effect of reaction temperature on peak areas. Results reported as

mean (n=5) ± standard deviation for FLD
97

Figure 4.8 Effect of reaction temperature on peak areas. Results reported as

mean (n=5) ± standard deviation for DAD
4.5.5 Optimization of buffer concentration
Optimization of buffer concentration was done by performing experiments at
three different buffer concentrations (0.05 M, 0.08 M and 0.1 M, all adjusted to
pH 9.5). Based on the results of five standard injections at each concentration
level, the use of different buffer concentrations affected the fluorescence
response of the FB-NDA derivatives. Although no major difference in HPLC
responses were observed between 0.05 M and 0.08 M buffers, a steep increase
(approximately 10 % from the others) in standard area was observed with 0.1 M
(Figure 4.9) with RSD values ≤ 3 % for all fumonisin analogues. The DAD response
was not affected by changes in buffer concentration (Figure 4.10). Since the FLD
and DAD are run simultaneously, it is convenient to use the same buffer for all
their preparations. Given that 0.1 M provided optimum results for FLD and buffer
98
concentration change not influencing DAD response, 0.1 M was selected as the

buffer concentration for fumonisin derivatization.

Figure 4.9 Effect of buffer concentration on peak areas for mean (n=5)
± standard deviation for FLD
Figure 4.10 Effect of buffer concentration on peak areas for mean (n=5)
± standard deviation for DAD

99
4.5.6 Effect of extractants

To improve the extraction efficiencies,
two different extraction solvents were
tested for their efficiency to extract fumonisins from maize. Approaches in
literature for improving the extraction of fumonisins from maize include the use
of EDTA as an extraction solvent (Sydenham et al., 1995, Dombrink-Kurtman et
al., 1999, Scott et al., 1996) whereas methanol : water (3:1) has been reported to
provide increased extraction efficiencies compared to other solvent mixtures
(Shephard, 1998).
Methanol : 0.1 M EDTA (3:1, v/v) was tested as an extraction solvent for its
efficiency for fumonisin extraction compared to the widely used methanol :
water (3:1, v/v). Table 4.5 provides detailed results on the extraction
experiments using both solvents. Results of the extraction with methanol : 0.1 M
EDTA (3:1, v/v) as a solvent provided similar results to the methanol : water (3:1,
v/v) extraction solvent. Agreement between extraction solvents was excellent
with the highest variation noted with methanol : water (3:1, v/v) as it produced
RSD values varying from 7 – 10 % compared to the methanol : 0.1 M EDTA (3:1,
v/v) with RSD values between 2 – 10 %. Use of methanol : 0.1 M EDTA (3:1, v/v)
caused some difficulties as the 0.1 M EDTA precipitated out of solution when it
was added to the methanol.
Due to the precipitation of EDTA and it not providing any improvement in results
when compared to methanol : water (3:1, v/v), the latter was chosen as the
100
extraction solvent for the study as used previously for fumonisin extraction

(Shephard et al., 1990).

101

Table 4.6 Solvent extraction efficiency for NDA derivatization
of maize samples (20 g / 100 mL) following SAX clean-up

FLD DAD
Solvent Extraction FB1 ( µg/g) FB2 ( µg/g) FB3 ( µg/g) FB1 ( µg/g) FB2 ( µg/g) FB3 ( µg/g)
MeOH : H2O
(3:1, v/v)
1 1197 547 191 974 534 197
2 1150 530 234 924 516 242
3 1120 526 206 886 534 237
4 1149 549 208 974 563 233
5 1144 512 224 842 511 236
6 980 437 210 841 421 252
Mean 1123 517 212 907 513 233
Stdev 74 41 15 61 49 19
RSD (%) 7 8 7 7 10 8
MeOH : 0.1 M EDTA
(3:1, v/v)
1 1240 551 188 1216 564 211
2 1321 571 201 1288 610 247
3 1261 544 201 1245 582 218
4 1314 560 213 1289 608 249
5 1315 557 216 1171 612 267
6 1285 567 215 1102 610 263
Mean 1290 558 205 1218 598 242
Stdev 33 10 11 73 20 23
RSD (%) 3 2 5 6 3 10
102
4.6 Recoveries

The accuracy of the optimized method was determined by measuring the

percent recoveries of the method. This was achieved by spiking maize samples

with 1103 µg/kg, 500 µg/kg and 270 µg/kg of FB1, FB2 and FB3 respectively. In
order to optimize recoveries two different extraction solvents (methanol : water
3:1 and methanol : 0.1M EDTA 3:1) were investigated, but little difference was
seen between the two and hence methanol : water (3:1) as previously used for
maize samples analysed by SAX clean-up was selected (Sydenham et al., 1996).
The accuracy and repeatability of the method is generally within acceptable
limits for both FLD and DAD following SAX clean-up (Table 4.6). In addition to SAX
clean-up, a similar recovery experiment was performed using IAC clean-up. A
good basis of comparison between the FLD and DAD methods was thus achieved,
even though a decrease in NDA-FB1 recoveries was observed with DAD following
IAC clean-up (Table 4.7).
Initial recovery experiments were done by spiking directly on the eluate of the
SAX cartridge to test the accuracy of the clean-up and derivatization method. The
results obtained from those experiments were comparable to those of spiking
directly on the maize sample. This signifies that little fumonisin is lost in the
extraction process, clean-up stage and derivatization; indicating the accuracy and
effectiveness of both sample preparation and derivatization for fumonisin
analyses when using NDA derivatization. Recovery results of both SAX and IAC
clean-up are acceptable according to the values established by the European
103
Commission which recommends recoveries of between 60 – 120 % for individual

FB methods (EC, 2005).

Table 4.7 Fumonisin recoveries (%)

from maize samples cleaned-up with SAX

FLD DAD
(ng/g) (ng/g) (ng/g) (ng/g) (ng/g) (ng/g)
Blank 1 295 110 21 175 126 62
Blank 2 341 126 36 203 149 91
Mean 318 118 28 189 137 76
Sample 1 1197 547 191 974 534 197
Sample 2 1150 530 234 924 516 242
Sample 3 1120 526 206 886 534 237
Sample 4 1149 549 208 974 563 233
Sample 5 1144 512 224 842 511 236
Sample 6 980 437 210 841 421 252
Mean 1123 517 212 907 513 233
Stdev 74 41 15 61 49 19
RSD (%) 7 8 7 7 10 8
Spike 1103 500 270 1103 500 270
Recoveries (%) 73 80 68 65 75 58
Table 4.8 Fumonisin recoveries (%) from maize samples cleaned-up with IAC
FLD DAD
Blank 1 113 67 12 406 703 nd
Blank 2 156 91 18 143 657 173
Mean 135 79 15 275 680 87
Sample 1 835 415 177 834 912 214
Sample 2 843 403 266 812 910 252
Sample 3 803 942 144 716 1227 311
Sample 4 793 749 161 646 1173 242
Sample 5 789 402 135 684 881 258
Sample 6 823 395 244 1100 844 245
Mean 814 551 188 799 991 254
Stdev 23 236 54 165 164 32
RSD (%) 3 43 29 21 17 13
spiked 1103 500 270 1103 500 270
Recovery (%) 62 94 64 48 62 62
104
4.7 Method application

To demonstrate the applicability of the
optimized method, analysis was applied
to 15 maize samples collected from the

Eastern Cape Province former Transkei
samples was routinely carried out using

area, South Africa. The analysis of the
SAX and IAC clean-up with simultaneous detection with HPLC-FLD and DAD. The
same samples were used to allow for comparison between the two clean-up
methods. Fumonisins were detected in all the samples with contamination levels
varying from 93 µg/kg to 4120 µg/kg for total fumonisins.
Comparison of FLD and DAD utilizing SAX clean-up was good with all samples
comparable at all levels (Table 4.8; Figure 4.12). The IAC clean-up however
showed poor comparison between the detectors as the results produced low
levels at high fumonisin contamination and high levels at low fumonisin
contamination with FLD, a trend that was reported by Chu et al., 1994 (Table 4.9,
Figure 4.13). Nonetheless, above 1800 µg/kg, the FLD and DAD could compare
after IAC clean-up.
105
Table 4.9 Comparison of HPLC-FLD and DAD following SAX clean-up (µg/kg)

FLD DAD


1 90 17 1 108 120 29 *nd 149
2 166 55 10 231 184 64 25 273
3 229 88 18 336 242 99 38 380
4 271 76 20 367 165 116 47 328
5 172 54 15 240 179 84 25 288
6 733 250 48 1030 663 245 35 943
7 442 176 20 638 370 184 nd 554
8 1712 832 227 2771 1522 809 197 2528
9 507 182 20 709 480 229 nd 709
10 713 248 23 983 663 265 41 969
11 1171 345 58 1575 997 326 129 1452
12 2065 747 201 3014 1875 688 241 2804
13 523 150 14 686 498 161 41 699
14 1754 453 284 2492 1598 465 338 2402
15 318 108 136 562 301 115 111 528

*nd- Not detectable
106
Table 4.10 Comparison of HPLCFLD and DAD following IAC clean-up (µg/kg)

FLD
DAD
1 59 22 12 93 325 442 *nd 766
2 145 58 16 219 211 347 137 695
3 156 59 10 225 343 387 nd 730
4 148 43 11 202 194 310 nd 504
5 173 53 21 248 403 383 nd 786
6 503 257 25 785 627 802 nd 1429
7 942 327 47 1316 1043 778 151 1972
8 1961 928 127 3016 1944 1277 140 3361
9 1370 370 80 1820 1467 737 243 2447
10 1518 428 168 2115 1631 804 144 2579
11 1133 297 59 1490 1204 690 183 2077
12 2410 825 108 3343 2506 1171 153 3830
13 1015 251 66 1332 1073 627 nd 1700
14 2713 673 278 3664 2797 1073 249 4120
15 1971 676 227 2874 2043 968 205 3215

*nd- Not detectable
107

Figure 4.11 Comparison of FLD and DAD following SAX clean-up, Total Fumonisins = FB1 + FB2 + FB3
108

Figure 4.12 Comparison of FLD and DAD following IAC clean-up, Total Fumonisins = FB1 + FB2 + FB3
109
4.8 Conclusion

Based on the chromatographic resolution of the fumonisin analogues in the

naturally contaminated samples, the selected optimized conditions are suitable
for NDA derivatization and the detection

of fumonisins in maize. The reaction of
fumonisins with NDA yielded FB-NDA derivatives which were found to be stable,
sensitive and selective for both FLD and DAD methodologies.
The derivatization using NDA is fast with total retention time less than 15
minutes. The analytical procedure supports using methanol : water (3:1) as
extraction medium and purification using either SAX or IAC allows for
simultaneous detection and quantification of FB1, FB2 and FB3. Aside from the
difference in limits of detection, a comparative study of FLD and DAD for the
analysis of fumonisins in maize demonstrated that the response achieved by
both detectors is sensitive enough for the analysis of fumonisin in naturally
contaminated samples. Both detectors would be appropriate for quantification
purposes with the highest sensitivity achieved by FLD.
Safety
The cyanide anion is a highly lethal and toxic reagent and consequently it is
necessary to observe and apply very strict and appropriate treatment conditions
for the safe disposal of all waste materials containing the cyanide anion.
110
References

1. Bennett GA and Richards JL (1994) Liquid chromatographic method for

analysis of naphthalene dicarboxaldehyde

derivative of fumonisins. Journal

2. Cho YH, Yoo HS, Min JK, Lee EY, Hong SP, Chung YB and Lee YM (2002)
Comparative study of naphthalene-2,3-dicarboxaldehyde and o-
phthalaldehyde fluorogenic reagents for chromatographic detection of
sphingoid bases. Journal of Chromatography A 977: 69-76
3. Chu FS and Li GY (1994) Simultaneous occurrence of fumonisin B1 and
other mycotoxins in moldy corn collected from the People’s Republic of
China in regions with high incidences of esophageal cancer. Applied and
4. Dombrink-Kurtzman MA and Dvorak TJ (1999) Fumonisin content in masa
and tortillas from Mexico. Journal of Agricultural and Food Chemistry 47:
622-627
5. European Commission Regulation (EC) 2005/38/EC of 6 June 2005.Official
Journal of the European Union L 143/18
6. Kwakman PJM, Koelewijn H, Kool I and Brinkman UATh and DeJong GJ
(1990) Naphthalene- and anthracene-2,3-dialdehyde as pre-column
labelling reagents for primary amines using reversed- and normal-phase
111
liquid chromatography with peroxyoxalate chemiluminescence detection.

Journal of Chromatography 511: 155-166

7. Lamba S, Pandit A, Sanghi SK, Gowri

VS, Tiwari A, Baderia VK, Singh DK and
Nigam P (2008) Determination of aliphatic amines by high performance
liquid chromatography- amperometric detection after derivatization with
naphthalene-2,3-dicarboxaldehyde. Analytica Chimica Acta 614: 190-195
8. Lillard SJ, Chiu DT, Scheller RH, Zare RN, Rodriguez-Cruz SE, Williams ER,
Orwar O, Sandberg M and Lundqvist JA (1998) Separation and
characteristics of amines from individual atria gland vesicles of Aplysia
californica. Analytica Chemistry 70: 3517-3524
9. Lino CM, Silva LJ, Pena ALS and Silveira MI (2006) Determination of
fumonisins B1 and B2 in Portuguese maize and maize-based samples by
HPLC with fluorescence detection. Analytical Bioanalytical Chemistry 384:
1214-1220
10. Lino CM, Silva LJG, Pena ALS, Fernandez M and Manes J (2007) Occurrence
of fumonisins B1 and B2 in broa typical Portuguese maize bread.
International Journal of Food Microbiology 118: 79-82
11. De Montigny PD, Stobaugh JF, Givens RS, Carlson RG, Srinivasachar K,
Sternson LA and Higuchi T (1987) Naphtyhalene-2,3-dicaboaxaldehyde/
cyanide ion: A rationally designed fluorogenic reagent for primary amines.
Analytical Chemistry 59: 1096-1101
112
12. Roach MC and Harmony MD (1987) Determination of amino acid at

subfemtomole levels by high performance liquid chromatography with

laser-induced fluorescence detection. Analytical Chemistry 59: 411-415

13.
Scott PM and Lawrence GA (1992) Liquid chromatographic determination
14. Scott PM and Lawrence GA (1996) Determination of hydrolysed fumonisin
B1 in alkali-processed corn foods. Food Additives and Contaminants 13:
823-832
15. Shephard GS, Sydenham EW, Thiel PG and Gelderblom WCA (1990)
Quantitative determination of fumonisins B1 and B2 by high performance
liquid chromatography with fluorescence detection. Journal of liquid
chromatography 13: 2077-2087
17. Silva L, Fernandez-Frazon M,Font G, Pena A, Silveira I, Lino C and Manes J
(2009) Analysis of fumonisin in corn-based food by liquid chromatography
with fluorescence and mass spectrometry. Food Chemistry 112: 1031-1037
18. Sydenham EW, Stockenstrom S, Thiel PG, Shephard GS, Koch KR and
Marasas WFO (1995) Potential of alkaline hydrolysis for the removal of
113
fumonisins from contaminated corn. Journal of Agricultural and Food

Chemistry 43: 1198-1201

19. Sydenham EW, Shephard GS, Thiel

PG, Stockenstrom S, Snijman PW and

van Schalkwyk DS (1996) Liquid chromatographic determination of
fumonisins B1, B2 and B3 in corn: AOAC-IUPAC Collaborative Study. Journal
20. Ware GM, Francis O, Kaun SS, Umrigar P, Carmen A, Carter L and Bennett
GA (1993) Determination of fumonisin B1 in corn by high performance
liquid chromatography with fluorescence detection. Analytical letter 26:
1751-1760
114

CHAPTER 5
An evaluation of dansyl chloride (DnS-Cl) for fumonisin
derivatization analysed by HPLC with fluorescence (FLD)
and ultraviolet (UV) detection

5.1 Introduction

Dansyl chloride (DnS-Cl) reacts with both primary and secondary amino groups to

provide stable derivatives. DnS-Cl derivatives are known to combine the unique

feature of being both fluorescent and detectable in the UV region (Loukou et al.,
2003). The only HPLC determination of fumonisin (Dasko et al., 2006) utilizing
DnS-Cl was FB1 in beer. In this chapter the extent to which DnS-Cl can be used for
the derivatization of fumonisins in naturally contaminated maize samples was
investigated.
Initial conditions used were based on the method by Dasko et al., (2006) with
modifications. The method was optimized to obtain optimum conditions
following strong anion exchange (SAX) and immunoaffinity (IAC) clean-up, pre-
column derivatization and HPLC detection with diode array (DAD) and
fluorescence detection (FLD). Dansyl chloride reacts with amines by nucleophilic
substitution and forms fluorescent dansyl derivatives (Legua et al., 1999) and
produced both sensitive and reproducible results with the analysis of biogenic
amines (Mo Dugo et al., 2006).
5.2.1 Chemicals
All chemicals used were of analytical grade. Sodium carbonate (Na2CO3), sodium
chloride (NaCl), o-phosphoric acid (H3PO4) and dansyl chloride (DnS-Cl) were
purchased from Merck. DnS-Cl derivatizing solution was prepared by dissolving
116
100 mg DnS-Cl in 10 mL acetone. Phosphate buffered saline (PBS) was prepared

as described in Section 3.2.1.


The standards were obtained and prepared as described in Section 3.2.2.
5.2.3 Maize Samples
Home-grown maize samples were collected as described in Section 3.2.3.
3.2.4.
3.2.5.
Standards: Working standard (20 µL) was added to DnS-Cl (20 µL, 100 mg/10 mL)
followed by 20 µL NaHCO3 (2 M, saturated). The solution was vortexed and
heated at 40°C for 10 minutes then cooled to 24°C under running tap water.
H3PO4 (1 M, 20 µL) was added and 20 µL injected into the HPLC system.
117
Samples: Samples were reconstituted in 100 µL methanol and derivatized by

adding 200 µL DnS-Cl (100 mg/10 mL) followed by 200 µL NaHCO3 (2 M,

saturated), heated at 40°C for 10 minutes on a heating mantle, cooled to 24 °C

and 200 µL H3PO4 (1 M) added and 20 µL injected into the HPLC system.
5.3 Chromatography
Reversed–phase high-performance liquid chromatography (RP-HPLC) was
performed on an Agilent Technologies (Wildbronn, Germany) and configured as
described in Section 3.3. The mobile phase was prepared by combining methanol
(740 mL) : 0.1 M sodium phosphate (NaH2PO4) (260 mL), and the pH of the
mixture was adjusted to pH 3.35 with o-phosphoric acid. The mobile phase was
filtered using 0.45 µm X 47mm filter paper with vacuum and pumped at 1
mL/min flow rate. Data were collected and analyzed by Agilent ChemStation
software and quantification achieved by comparison of peak areas with those of
authentic fumonisin standards.
A typical fumonisin B chromatogram was obtained when fumonisin working
standard (containing 55.13 µg/kg of FB1, 25.00 µg/kg and 13.25 µg/kg of FB2 and
FB3, respectively) was derivatized with DnS-Cl (Figures 5.1 and 5.2).
Chromatographic resolution of the dansyl derivatives was obtained with an
isocratic elution using MeOH : 0.1 M NaH2PO4 (74:26) mobile phase with well
118
resolved analyte peaks in less than 15 minutes and an elution order of FB1, FB3

and FB2. Retention times of 4.98, 13.01 and 11.83 min (± 5 %) for FB1, FB2 and

FB3, respectively, for both FLD and DAD were attained.

The chromatogram of fumonisin derivatives showed additional peaks as was also
seen for NDA. Interference of the peaks were examined by overlaying the
chromatograms of reagent blank with working standards, the peaks (A and B) did
not interfere with those of the FB analogues. Removal of the peaks proved
difficult, with one of the additional peaks probably dansyl dimethylamine, the
most abundant product of dansyl reactions (Seiler et al., 1978). Since they did
not interfere with both the resolution and quantification of the analyte peaks, no
further attempts were made to remove them and they were classified as reagent
peaks.
119

Figure 5.1 Chromatogram of combined fumonisins working standard detected by FLD
120

Figure 5.2 Chromatogram of combined fumonisins working standard detected by DAD
121

The repeatability and reproducibility of the method was determined by intra-
and inter-day precision and reported as

standard peak area and RSD in tables 5.1

and 5.2, respectively. The inter-day precision was obtained by injecting three to
four working standards a day. Results obtained by FLD were more reproducible
(RSD ≤ 1 %) compared to results obtained with DAD (RSD ≤ 6 %) (Table 5.1).
Table 5.1 Intra-day precision using peak areas of working standards (n=3)
FLD DAD
Standard 1 56903 32119 26325 77.3 39.8 31.7
Standard 2 56436 32527 26516 74.2 42.0 31.7
Standard 3 57731 32754 26671 81.3 44.5 35.0
Mean 57023 32467 26504 78 42 33
Stdev 656 322 173 4 2 2
RSD (%) 1.2 0.9 0.7 5 6 6
Inter-day precision was obtained by analyzing three standards per day over a
period of four days. Table 5.2 shows inter-day results for FLD and DAD. FLD
performed better (RSD ≤ 12 %) compared to DAD (RSD ≤ 16 %). The results meet
122
the performance criteria for EU regulatory purposes (EC, 2005), which state that

the repeatability RSD must be ≤ 20 % for fumonisin levels ≤ 100 µg/kg.

Table 5.2 Inter-day precision using

peak areas of working standards injected
(n=12)
FLD DAD
Day 1 56903 32119 26325 77.3 39.8 31.7
56436 32527 26516 74.2 42.0 31.7
57731 32754 26671 81.3 44.5 35.0
Day 2 60929 28972 20177 104.0 60.3 33.5
60737 33117 21764 106.7 57.9 36.4
62663 32104 21315 111.0 56.1 34.5
Day 3 62206 33186 26181 109.5 52.6 39.0
67248 34381 27639 122.5 51.0 42.6
63089 35943 28057 115.3 58.6 43.2
Day 4 55809 32432 22626 96.1 59.9 49.1
53593 31948 21326 89.9 51.9 34.4
51576 30950 20733 87.3 53.7 34.2
Mean 59077 32536 24111 97.9 52.4 37.1
Stdev 4495 1699 3011 15.9 7.0 5.3
RSD (%) 8 5 12 16 13 14
123
5.4.3 Detection Limits

The detection limits reached with FLD were generally better than those obtained
with DAD (Table 5.3). The LOD values were higher than expected in comparison

to other studies (Loukou et al., 2003) with FB1 and FB3 determined with
comparable sensitivity for both detectors. The values obtained are generally
satisfactory for the analysis to be performed using the method. These results
indicate that DnS-Cl derivatization reagent provides adequate sensitivity for
fumonisin analysis.
Table 5.3 Limits of detection (LOD) and quantitation (LOQ) in terms of
amount injected (ng) onto HPLC column
Amount (ng) injected into HPLC column
FB1 FB2 FB3
LOD (s:n=3) FLD 4.3 3.9 2.1
DAD 17.2 15.6 16.6
LOQ (s:n=10) FLD 8.6 6.9 8.3
DAD 34.5 62.5 33.1
124
5.5 Robustness

5.5.1 Selection of derivatization solvent

Standard solutions of DnS-Cl are prepared

by dissolving the pure compound in
either acetone (Dasko et al., 2006) or acetonitrile (Heimbecher et al., 1997).
These solvents were tested in this study for their suitability as reaction solvents
for the preparation of the dansyl reagent. The reactions of fumonisins with DnS-
Cl prepared in acetonitrile were problematic, producing little to no
chromatographic peaks suggesting either an incomplete or no reaction between
DnS-Cl (prepared in acetonitrile) with the fumonisins. Kang et al., 2006 reported
that an increase in DnS-Cl volume in solution can directly increase the intensity
of the derivative; therefore the volume of DnS-Cl in solution was increased from
100 µL to 300 µL. This improved the reaction, but produced poor
chromatographic resolution, reproducibilities and formed unstable derivatives.
Analysis of six standards derivatized with DnS-Cl prepared in acetonitrile were
scattered with no repeatability. Under the conditions employed, acetonitrile was
found to be an inappropriate solvent for DnS-Cl preparation with very low FLD
and DAD intensities.
As acetonitrile was shown to be unsuitable, acetone as a reagent solvent was
tested. The reaction of fumonisins with DnS-Cl (prepared in acetone) occurred
spontaneously with satisfactory chromatographic resolution and reproducible
peaks. DnS-Cl prepared in acetone reacted more intensely with the fumonisins
125
compared to acetonitrile (Figure 5.3). Consequently, DnS-Cl reagent for further

use in the study was prepared in acetone.

Figure 5.3 Comparison of acetonitrile and acetone as DnS-Cl reagent

solvents. Results reported as mean (n=6) ± standard deviation. (A)
Comparison of FLD, (B) Comparison of DAD
5.5.2 Wavelength Selection
Wavelength selection was determined by analyzing working standards at various
wavelengths reported in literature (Molins-Legua et al., 1998; Dasko et al., 1996)
for their suitability to provide optimum sensitivity and intensity. Optimum results
were obtained at excitation wavelength 247 nm and emission wavelength 510
nm for FLD (Table 5.4). For DAD, 230 nm provided highest sensitivity (Table 5.5).
Iso-absorbance plots (software programme which displays chromatographic
details in 3D including retention time versus wavelength, from which optimum
126
wavelength can be selected) were used to confirm the optimum wavelength for

DAD.

Table 5.4 Effect of wavelength on

working standard peak areas for FLD
Ex 252 Ex 247 Ex 360 Ex 365

Em510 Em 510 Em 510 Em 510
FB1 38797 50132 6024 47196
FB2 22190 28519 3627 2737
FB3 15521 20486 2579 23299
Table 5.5 Effect of wavelength on working standard peak areas for DAD
230 nm 254 nm 280 nm 335 nm
FB1 76 67 6 25
FB2 40 35 ND 14
FB3 29 24 ND 9
5.5.3 Reaction temperature selection
Temperature experiments were performed at 24°C, 40°C and 60°C for 15
minutes as these are the mostly used temperatures in literature (Smith et al.,
1985; Dasko et al., 1996). An increase in standard peak area was observed when
the reaction was heated compared to when it was performed at room
temperature (24°C). Heating the derivatives improved both the reaction yield
and RSD values from 15 % to ≤ 2 % (Figure 5.4). Further increase in temperature
to 60°C showed an insignificant increase in response for the FB 2 and FB3 of the
FLD. In contrast, the DAD showed a decrease in response between 40°C to 60°C
127
except for FB3; which showed a slight increase in response (Figure 5.5). Since, the

FLD results showed increased standard area for FB1 and DAD showing increased

results for FB1 and FB2 when experiments were performed at 40°C, this was

selected as the optimum reaction temperature.

Figure 5.4 Effect of reaction temperature on peak areas of FLD, results

calculated on mean (n=5) ± standard deviation
Figure 5.5 Effect of reaction temperature on peak areas of DAD, results

calculated on mean (n=5) ± standard deviation
128
5.5.4 Reaction time selection

Different reaction times (10 min, 15 min
and 30 min) were tested, based on the
peak areas no major difference in HPLC

responses were observed between 15
min and 30 min. Optimum reaction time was obtained at 10 minutes for FLD
(Figure 5.6) with a slightly lower response for DAD at 10 min compared to other
reaction times (Figure 5.7). To save on analysis time, 10 minutes was selected as
the suitable reaction time for DnS-Cl derivatization of the FB analogues.
Figure 5.6 Effect of reaction time at 40 °C on peak areas for FLD. Results
reported as mean (n=5) ± standard deviation
129

Figure 5.7 Effect of reaction time at 40 °C on peak areas for DAD. Results
reported as mean (n=5) ± standard deviation
5.6 Recoveries
Accuracy of the optimized method was examined by analyzing recoveries of
fumonisins. Initial recoveries were determined by spiking directly on the dry
milled maize at levels of 1103 µg/kg, 500 µg/kg and 270 µg/kg for FB1, FB2 and
FB3, respectively, and performing clean-up with IAC. The recoveries obtained
with FLD were poor ≤ 30 % (Table 5.6). Due to poor chromatography and the
presence of larger FB2 peaks then FB1, the recoveries for DAD for this experiment
could not be calculated. To improve on the results obtained, the derivatization
procedure was tested for its ability to derivatize fumonisins in maize samples;
this was done by spiking into the 10 mL samples eluated from the SAX prior to
the drying step. Recovery results were again poor, with FLD achieving ± ≤ 40 %
for the three FB analogues. Apparent recoveries as determined by DAD were
130
incongruous (10 % for FB1, 80 % for FB2 and 77 % for FB3), as compared to the

FLD.

Table 5.6 Determining the recovery

of the maize extraction using IAC
FLD
FB1 (ng/g) FB2 (ng/g) FB3 (ng/g)
Blank 1 57 0 0
Blank 2 31 18 9
Mean 44 9 4
Sample 1 359 166 83
Sample 2 316 156 84
Sample 3 343 154 83
Sample 4 346 156 78
Sample 5 328 149 72
Sample 6 279 148 69
Mean 328 155 78
Stdev 29 7 7
RSD (%) 9 4 8
Spike 1103 500 270
Recoveries (%) 26 29 27
Table 5.7 Determination of the recovery of the derivatization procedure
FLD DAD
Blank 1 101 31 6 421 742 0
Blank 2 108 36 10 406 831 0
Mean 104 33 8 413 787 0
Sample 1 648 228 123 692 1433 222
Sample 2 628 235 125 415 1146 264
Sample 3 483 176 96 658 1150 181
Sample 4 570 216 107 597 1158 164
Sample 5 563 183 97 507 1040 199
Sample 6 533 176 97 344 1194 214
Mean 571 202 108 535 1187 207
Stdev 61 27 13 138 131 35
RSD (%) 11 13 12 26 11 17
Spike 1103 500 270 1103 500 270
Recoveries (%) 42 34 37 11 80 77
131
An evaluation of the peak purity for UV absorption was done on FB 1. The results

obtained from the peak purity test showed that all calculations were within the

calculated threshold (Figure 5.8) and no impurities are present under the analyte

peaks. The result thus far suggests that the current derivatization conditions
used are not suitable for derivatization of naturally contaminated samples.
Figure 5.8 Purity peak check results
Figure 5.9 below illustrates the type of chromatogram obtained with DnS-Cl
derivatives detected by DAD. Due to poor resolution of FB 1 peak from the
132
additional peaks in the chromatogram (which could possibly be due to matrix

interference) and FB2 providing higher results compared to FB1, DAD analysis of

DnS-Cl derivatives was abandoned and the study was continued with only FLD.

Figure 5.9 Chromatogram of naturally contaminated maize sample with DAD
5.7 Optimization of recoveries using naturally contaminated samples
The fluorescent products obtained by the reaction of fumonisins with DnS-Cl
produced low recoveries. To improve on the recoveries obtained, optimization of
the reaction was performed by spiking directly on the cleaned-up eluate to allow
for both the recoveries and the derivatization process to be examined. The
parameters analysed included temperature, time, volume of DnS-Cl and pH as
they are known to affect the yield of DnS-Cl derivatives (Kang et al., 2006).
133
General conditions used:

For optimization of the samples, the
following conditions were used unless
otherwise stated as per optimization parameter

(i.e. during sample optimization,

except for the parameter tested, everything was kept constant):
Samples re-dissolved in 100 µL methanol, 200 µL DnS-Cl and 200 µL NaHCO3
added, heated for 10 minutes at 40°C, cooled to 24°C under tap water and 200
µL H3PO4 added and 20 µL injected into HPLC.
5.7.1 Effect of reaction temperature and time
Since both reaction temperature and time can affect the rate of derivative
formation and hence HPLC response, optimization of these parameters was
undertaken first.
Temperature experiments were performed first; all experiments were performed
for 10 minutes at two temperatures (40°C and 60°C). The maximum intensity was
obtained at 60°C (Table 5.8). Reaction time experiments were therefore
performed at 60°C. Slight differences in HPLC responses were observed between
10 min and 30 min because 10 min provided shorter reaction time (Table 5.9) it
was selected as reaction time. Therefore, heating at 60°C for 10 min was selected
as optimum temperature and time for sample derivatization. However
optimization of the reaction temperature and time did not improve recoveries.
134
Table 5.8 Effect of reaction temperature on maize derivatized for 15
minutes with DnS-Cl
40°C
60°C
(ng/g)
(ng/g) (ng/g) (ng/g) (ng/g) (ng/g)
Blank 1 7 0 0 15 7 0
Blank 2 8 0 0 11 4 0
Mean 8 0 0 13 5 0
Sample 1 147 48 24 218 64 41
Sample 2 142 40 22 225 70 40
Sample 3 144 37 20 132 39 22
Sample 4 126 36 20 257 77 43
Sample 5 140 40 21 234 70 40
Sample 6 154 46 24 244 73 45
Mean 142 41 22 218 65 38
Stdev 9 5 2 45 14 8
RSD (%) 6 12 8 20 21 21
Spike 1103 500 270 1103 500 270
Recoveries (%) 12 8 8 19 12 14
Table 5.9 Effect of reaction time on maize derivatized at 60°C with DnS-Cl
10 min 15 min 30 min

FB1 FB2 FB3 FB1 FB2 FB3 FB1 FB2 FB3
(ng/g) (ng/g) (ng/g) (ng/g) (ng/g) (ng/g) (ng/g) (ng/g) (ng/g)
Blank 1 7 0 0 15 7 0 12 8 0
Blank 2 13 9 0 11 4 0 15 9 0
Mean 10 5 0 13 5 0 14 8 0
Sample 1 435 189 101 218 64 41 418 153 82
Sample 2 423 198 101 225 70 40 443 182 92
Sample 3 457 192 96 132 39 22 431 176 88
Sample 4 449 187 94 257 77 43 407 158 83
Sample 5 410 170 88 234 70 40 409 173 88
Sample 6 424 191 98 244 73 45 447 180 94
Mean 433 188 96 218 65 38 426 170 88
Stdev 18 10 5 45 14 8 17 12 5
RSD (%) 4 5 5 20 21 21 4 7 5
Spike 1103 500 270 1103 500 270 1103 500 270
Recoveries (%) 38 37 36 19 12 14 37 32 33
135
5.7.2 Effect of DnS-Cl volume

Given that excess DnS-Cl was already used in the study and that it can either
compensate for side reactions (Seiler, 1971; Heimbecher et al., 1997) or increase
the response on undesirable reagent compounds (Molions-Legua et al., 1998),
the effect of DnS-Cl volume was examined.
DnS-Cl volume in the derivatization reagent was investigated at 100 µL, 200 µL
and 300 µL. A sharp increase in response was obtained from 100 µL to 200 µL.
However, a further increase in volume to 300 µL resulted in a decrease in
response (Figure 5.10). This may be due to the decomposition of the derivatives
from the excess DnS-Cl (Tapuhi et al., 1981). Since 200 µL provided optimum
results it was selected for sample derivatization. Calculation of recoveries from
the results showed no improvement in recoveries.
Figure 5.10 Effect of reagent volume (DnS-Cl) on reaction yield, results

reported on mean (n=6) ± standard deviation
136
5.7.3 Effect of pH

Na2CO3 was selected as an alternative to NaHCO3 because it could provide both

increased pH, which has been reported to increase the yield of DnS-Cl derivatives

(Heimbecher et al., 1997) and provide

the basic condition necessary for
derivatization. Experiments to test the effect of buffer change and pH were
carried out simultaneously (i.e. buffer change from NaHCO3 to Na2CO3 provided a
pH change from 8.25 to pH 11.05). The chromatographic resolution of fumonisin
working standards with the use of Na2CO3 was baseline with increased response
for all 3 analogues. However, when the samples were derivatized poor resolution
of the peaks was obtained with two chromatographic peaks formed at similar
retention times to FB1 (Figure 5.11). Although identification of the FB1 peak
could be done by the use of retention times, integrating the double peak would
be an inaccurate quantification of FB1.
Figure 5.11 Naturally contaminated maize sample with DAD
137
The source of the double peak was further investigated by preparing working

standards using Na2CO3. However, the original FB1 responses could not be

reproduced since its area increased disproportionally to that of FB2 and FB3,

suggesting some interference beneath the peak (Figure 5.12) e.g. non-separation
from reagent peak A and B (Figures 5.1 and 5.2).
Troubleshooting (i.e. detector settings, cleaning injector port and use of minimal
injection volume (10 µL)) of the chromatographic system used indicated that it
had no influence on the FB1 standard area and on the split sample peak. It was
thus concluded that DnS-Cl is an inappropriate derivatization reagent for the
analysis of fumonisins in maize.
Figure 5.12 Fumonisin working standard derivatized with DnS-Cl and Na2CO3
as buffer
138
5.8 Conclusion

The effectiveness of pre-column derivatization of fumonisins with DnS-Cl, HPLC

separation and subsequent application was evaluated. Different experimental

conditions in order to improve method

performance for fumonisin analysis were
used with initial optimization experiments with working standards providing
reproducible, repeatable and precise results. Application of the optimized
method to naturally contaminated maize samples produced unreliable results for
DAD, resulting in the discontinuation of the DAD analysis. Although baseline
resolution of the peaks with maize samples was obtained with FLD, the poor
recoveries could not be improved even after investing the influence of DnS-Cl
concentration, pH, buffer, derivatization time and temperature.
Generally, it seems that DnS-Cl derivatives are less desirable due to the presence
of analytical interferences and suspiciously higher results of FB2 compared to FB1.
Although all analytes could be identified with FLD with the use of retention
times; the recovery results obtained suggests it to be unsuitable for the analysis
of naturally contaminated maize samples due to either no reaction or incomplete
reaction, and the formation of secondary peaks.
139
References

derivatization agents in HPLC-fluorescence

detection analysis of
2. European Commission Regulation (EC) 2005/38/EC of 6 June 2005, Official
Journal of the European Union L 143/18
3. Heimbecher S, Lee YC, Tabibi SE and Yalkowsky SH (1997) Derivatization
and high-performance liquid chromatographic analysis of
pentaazapentacosane pentahydrochloride. Journal of Chromatography B
691: 173-178
neuroactive amino acids of biological samples. Clinica Chimica Acta 366:
352-356
chemical ionization mass spectrometry. Journal of Chromatography A 996:
103-113
140
6. Mo Dugo G, Vilasi F, La Torre GL and Pellicano TM (2006) Reverse phase

HPLC/DAD determination of biogenic amines as dansyl derivatives in

experimental red wines. Food Chemistry 95: 672-676

7. Molins-Legua C, Campins-Falco P and Sevillano-Cabeza A (1998) Automated
pre-column derivatization of amines in biological samples with dansyl
chloride and with or without post-column chemilumiescence formation by
using TCPO-H2O2. The Analyst 123: 28712876
8. Seiler N (1971) Identification and quantitation of amines by thin-layer
chromatograph. Journal of Chromatography A 63: 97-112
9. Seiler N and Knodgen B (1978) Determination of di- and polyamines by high
performance liquid chromatographic separation of their dimethylamino-
naphthalene-1-sulfonyl derivatives. Journal of Chromatography 145: 29-39
10. Smith MA and Davies PJ (1985) Separation and quantitation of polyamines
in plant tissue by high performance liquid chromatography of their dansyl
derivatives. Plant Physiology 78: 89-91
11. Sydenham EW, Shephard GS, Thiel PG, Stockenstrom S, Snijman PW and
Van Schalkwyk DS (1996) Liquid chromatographic determination of
fumonisins B1, B2 and B3 in corn: AOAC-IUPAC collaborative study. Journal
141
12. Tapuhi Y, Schimdt D, Lindner W and Karger BL (1981) Dansylation of amino

acids for high-performance liquid chromatography analysis. Analytical

Biochemistry 115: 123-129

142

CHAPTER 6
Comparisons of methods, General Discussion,

Recommendations and Conclusion
6.1 Comparison of OPA and NDA

Validation of the methods for recovery and repeatability were applied to 15

maize samples and are reported in Chapters 3 and 4. The same samples were

used for both derivatization reagents. From a single extraction, 2 aliquots of the
centrifuged extract were independently cleaned-up on SAX, with one eluate
being derivatized with OPA and the other with NDA. This allowed accurate
comparison between the OPA (FLD and DAD) with NDA (FLD and DAD) by
importantly, avoiding variation in the extraction step (Table 6.1).
Although comparison between the detectors was good following SAX clean-up,
the results from the two derivatives were comparable only up to 1000 µg/kg. It
was found that with concentrations above 1000 µg/kg, the comparison was
frequently poor, with NDA being lower than 50% of the OPA (Figure 6.1). This
would suggest that certain of the home-grown maize samples studied contained
inhibitors to the complete NDA derivatization reaction of their SAX extracts. Due
to these poor comparisons, the alternate IAC clean-up method was investigated,
since it produces a cleaner extract for derivatization.
In a similar manner to the SAX experiment, single samples were extracted and
duplicate clean-ups performed, one for OPA and one for NDA derivatization. For
OPA, a good comparison was obtained between FLD and DAD, whereas for NDA,
all results were higher with DAD, especially at levels below 1000 µg/kg. The
reason for this trend was not apparent (Figure 6.2). Comparison of OPA and NDA
derivatives showed much improved results over the previous comparison of the
144
derivatives after SAX clean-up. In comparing the derivatives using the DAD

response, again comparisons above 1000 µg/kg were superior to those below

this level.

The OPA results were consistent whether determined either after SAX or IAC
clean-up, showing that OPA is a robust derivatization reagent for fumonisin
analysis (Figure 6.3). For NDA derivatization (Figure 6.4), IAC clean-up produced
much cleaner extracts, which resulted in improved comparison with OPA (Table
6.1). Recent applications of NDA fumonisin analysis with FLD used IAC clean-up
(Lino et al., 2006; Lino et al., 2007; Silva et al., 2009). An older method using C18
RP-SPE of fumonisins from mouldy maize reported that compared to OPA, NDA
gave higher values at lower contamination levels and lower values at higher
contamination levels (Chu et al., 1994).
145
Table 6.1 Total fumonisin levels (FB1+FB2+FB3; µg/kg) in naturally contaminated
maize cleaned-up with SAX and IAC, derivatized with OPA or NDA

FLD DAD

SAX IAC SAX IAC
Sample OPA NDA OPA NDA OPA NDA OPA NDA
*
1 106 108 103 93 102 149 ND 766
2 234 231 238 219 310 273 142 695
3 288 336 260 225 257 380 143 730
4 296 367 157 202 320 328 ND 504
5 220 240 263 248 300 288 161 786
6 1132 1030 701 785 1279 943 709 1429
7 1375 638 1623 1316 1470 554 1590 1972
8 3144 2771 3730 3016 3234 2528 3804 3361
9 2163 709 1711 1820 2289 709 1585 2447
10 2537 983 2196 2115 2805 969 2231 2579
11 1900 1575 1711 1490 1948 1452 1688 2077
12 3120 3014 4689 3343 3164 2804 4716 3830
13 1577 686 1485 1332 1812 699 1445 1700
14 6088 2492 5327 3664 6114 2402 5246 4120
15 3740 562 3894 2874 4442 528 3756 3215
146

Figure 6.1 Comparison of OPA with NDA following SAX clean-up for FLD and DAD,
Figure 6.2 Comparison of OPA and NDA following IAC clean-up for FLD and DAD,
147

Figure 6.3 Comparison of IAC and SAX with OPA derivatization for FLD and DAD,
Figure 6.4 Comparison of IAC and SAX with NDA derivatization for FLD and DAD,
148
6.2 General Discussion

Chromatographic methods used to analyse fumonisins in maize were

investigated with regard to: (1) derivatising
reagent (OPA, NDA and DnS-Cl) (2)
clean-up method (SAX and IAC) and (3) detection (FLD and DAD). Table 6.2
provides a summary of the optimal conditions used for OPA, NDA and DnS-Cl
derivatization of fumonisins in maize.
Table 6.2 Chromatographic parameters for determination of fumonisins as

OPA, NDA and DnS-Cl derivatives
OPA NDA DnS-Cl
Column Luna 5 µm, 75 mm Luna 5 µm, 75 mm Luna 5 µm, 75 mm

x 4.6 mm x 4.6 mm x 4.6 mm
Mobile phase MeOH:0.1 NaHPO4 MeOH:0.1 NaHPO4 MeOH:0.1

(77:23) (78:22) NaHPO4 (76:24)
Flow rate 1 mL/ min 1 mL/ min 1 mL/ min
Injection volume Standards – 10 µL Standards – 10 µL Standards – 10 µL
Samples – 20 µL Samples – 20 µL Samples – 20 µL
Buffer (s) 1M Na2B4O7.10H2O 1M Na2B4O7.10H2O 2M NaHCO3
2M Na2CO3
Temperature (s) 24 °C 60°C 60°C
Reaction time 2 minutes 15 minutes 10 minutes
DAD wavelength (s) 335 nm 252 nm 230 nm
FLD wavelength (s) Excitation 335 nm Excitation 420 nm Excitation 247 nm
Emission 440 nm Emission 500 nm Emission 510 nm
149
Internal validation of the methods was performed using characteristics such as

precision, accuracy, specificity and application to naturally contaminated maize

samples. The optimized analytical methods proved to be both selective and

sensitive. This is evident in the ability

of the methods to produce accurate
measures of the analytes in the presence of unknown components and any other
products that may be expected to be present in the sample matrix and with
comparable fumonisins levels in a wide range of fumonisin contamination. This
study emphasized that the use of appropriate fluorometric derivatization
procedures is of considerable importance for accurate determination of
fumonisins. The robustness of the procedure allowed us to identify some critical
steps in the methods for fumonisin analysis in maize. In particular, the extraction
of fumonisins from sample matrix, clean-up and derivatization reagent which
were all demonstrated to be critical.
The OPA method was found to be sensitive, reliable and reproducible for
fumonisin analysis in maize. Although unstable, it has been found to be a suitable
derivatization reagent (Shephard et al., 1996). Baseline chromatographic
separation was obtained with retention times of the standard peak allowing
identification of the analyte peaks. One of the objectives of this study was to
determine the extent to which DAD could be used as an alternative to FLD for
OPA derivatization. This was achieved by obtaining comparable results between
the detectors, although FLD proved to be more sensitive with lower detection
limits and significantly higher peak areas. Good comparison between these two
detectors was demonstrated with SAX and IAC clean-up.

150
The NDA method was optimized and validated. NDA provided comparable results

to those of OPA and essentially confirming a previous study by Bennett and

Richard 1994. It proved to be a suitable alternative for OPA, especially following

IAC clean-up, and for laboratories requiring

overnight analysis of fumonisins
which, due to their high stability, provides an added advantage over OPA. NDA
proved to be sensitive as well as sufficiently selective in the application of
fumonisin analysis.
Dansyl derivatives are mainly used for amino acid (De Jong et al., 1982; Kang et
al., 2006) and biological amine analysis (Loukou et al., 2003; Proestos et al.,
2008). Only one study has utilized DnS-Cl for fumonisin derivatization in beer
samples and was reported to form stable complexes with satisfactory
chromatographic separations (Dasko et al., 2006). Scott and Lawrence, 1992
derivatized fumonisins in maize and found it to form good derivatives. However
it also produced analytical interferences with the maize matrix.
As a result of the investigations performed in this study, it was found that DnS-Cl
could on the one hand form stable, sensitive and highly fluorescent derivatives
with fumonisin standards. However, on the other hand, due to the low
recoveries from maize it could not be used for the analysis of naturally
contaminated maize samples as previously reported by Scott and Lawrence
1992. This showed that the maize matrix caused analytical interferences with
derivatives thereby suppressing the reaction between the fumonisin amino
groups with the dansyl chloride reagent. Interferences with the maize were
151
mainly observed with DAD; therefore the DAD investigations were abandoned.

However, DnS-Cl has some potential to be utilized with FLD.

6.3 Recommendations

Although cyanide has been shown to be a suitable nucleophile for NDA
derivatization, it is a highly toxic substance and hence an alternative would
be much safer for the environment. Therefore, future research should
focus on finding other suitable nucleophiles for NDA derivatization.
DnS-Cl is known to be a non-specific reagent as it reacts with amino groups
of many compounds as well as hydroxyl groups of phenols and some
alcohols (Smith et al., 1985). Reaction of fumonisins with DnS-Cl occurs at
the more nucleophilic amino functional group, which is where fumonisin
derivatization often occurs. Further studies should be conducted to
address the issues surrounding the apparent matrix interference of DnS-Cl
derivatives on fumonisin and investigate the mechanism between DnS-Cl
with fumonisins as they are structurally suited for DnS-Cl reaction.
6.4 Conclusion
This study uniquely investigated three derivatization reagents systematically for
fumonisin analysis in South African home-grown maize intended for human
consumption with concurrent FLD and DAD detection and with two clean-up
methods (SAX and IAC).
152
This study shows that although FLD is more sensitive, UV detection can be used

as a reliable alternative for fumonisin analysis of OPA derivatives. OPA and NDA

have proven to be excellent fluorogenic reagents for accurate determination of

fumonisins in naturally contaminated

maize samples and can be used as
alternatives to each other employing both SAX and IAC clean-up with either FLD
or DAD. Even though DnS-Cl derivatization could not be applied to maize
samples, the study can be the basis for investigating the use of DnS-Cl in other
related matrixes. In conclusion this study has shown that UV detection can be
utilized as an alternative to FLD for fumonisin analysis in naturally contaminated
maize irrespective of the clean-up method or the derivatization agent.
153
References

1. Bennett GA and Richards JL (1994)

Liquid chromatographic method for

analysis of naphthalene dicarboxaldehyde derivative of fumonisins. Journal

2. Chu FS and Li GY (1994) Simultaneous occurrence of fumonisin B1 and
other mycotoxins in moldy corn collected from the People’s Republic of
China regions with high incidences of esophageal cancer. Applied and
derivatization agents in HPLC-fluorescence detection analysis of
4. De Jong C and Hughes GJ (1982) Amino acid analysis by high performance
liquid chromatography: An evaluation of the usefulness of pre-column DnS
derivatization. Journal of Chromatography 241: 345-359
neuroactive amino acids of biological samples. Clinica Chimica Acta 366 :
352-356
6. Lino CM, Silva LJ, Pena ALS and Silveira MI (2006) Determination of
fumonisins B1 and B2 in Portuguese maize and maize-based samples by
154
HPLC with fluorescence detection. Analytical Bioanalytical Chemistry 384:

1214-1220

7. Lino CM, Silva LJG, Pena ALS, Fernandez M and Manes J (2007) Occurrence

of fumonisins B1 and B2 in broa typical Portuguese maize bread.
International Journal of Food Microbiology 118: 79-82
chemical ionization mass spectrometry. Journal of Chromatography A 996:
103-113
9. Proestos C, Loukatos P and Komaitis M (2008) Determination of biogenic
amines in wines by HPLC with precolumn dansylation and fluoremetric
determination. Analytical, Nutritional and Clinical Methods 106: 1218-1224
of fumonisins with 4-fluoro-7-nitrobenzofurazan. Journal of Association
and Official Analytical Chemistry International 75: 829-834
155
12. Silva L, Fernandez-Frazon M,Font G, Pena A, Silveira I, Lino C and Manes J

(2009) Analysis of fumonisin in corn based food by liquid chromatography

with fluorescence and mass spectrometry. Food Chemistry 112: 1031-1037

13. Smith MA and Davies PJ (1985) Separation and quantitation of polyamines
in plant tissue by high performance liquid chromatography of their dansyl
derivatives. Plant Physiology 78: 89-91
156

Dr. G.S Shephard

Uploaded by

Copyright:

Available Formats

Dr. G.S Shephard

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

Dr. G.S Shephard

Uploaded by

Copyright:

Available Formats

Determination of fumonisins in maize by High Performance Liquid

A thesis submitted in partial fulfilment of the requirements for the degree

Magister Scientiae in the Department of Chemistry, University of the

Supervisors: Professor I.R Green

Dr. G.S Shephard

Co-supervisor: Dr. L. van der Westhuizen

phthaldialdehyde, naphthalene-2,3-dicarboxaldehyde and dansyl chloride

MSc Chemistry Thesis, Department of Chemistry, University of the Western Cape

Keywords: Fumonisins; fluorescence detector; ultraviolet detector; o-

phthaldialdehyde; naphthalene-2,3-dicarboxaldehyde; dansyl chloride; high-

performance liquid chromatography; strong anion extraction; immunoaffinity;

diode array detector.

Fumonisins, carcinogenic mycotoxins produced by various Fusarium

The most abundant naturally occurring fumonisins analogues in maize are

derivatization reagent for fumonisins as they lack useful chromophores or

fluorophores. Alternative derivatization reagents, naphthalene-2,3-

dicarboxaldehyde (NDA) and dansyl chloride (DnS-Cl), were investigated in this

fluorescence detectors to determine UV detection as an alternative following

12.5 ng and 6.6 ng, respectively. Subsequently naturally contaminated maize

respectively, with DAD. Subsequently the NDA derivatization was compared to

the OPA derivatization, as well as an alternative sample clean-up with

immunoaffinity column (IAC) by analyzing naturally contaminated maize samples

comparable results irrespective of the clean-up method or the derivatization

agent. The investigation of DnS-Cl as a derivatization reagent resulted in LOD for

derivatives were formed with fumonisin working standards, matrix interferences

for fumonisin analysis in naturally contaminated maize. In conclusion this study

has shown that UV detection can be utilized as an alternative to FLD for

fumonisin analysis in naturally contaminated maize irrespective of the clean-up

method or the derivatization agent.

phthaldialdehyde, naphthalene-2,3-dicarboxaldehyde and dansyl chloride

quoted have been indicated and acknowledged by complete references.

Full name............................................................. Date..................................................

To my supervisors Dr Gordon Shephard

advice and valuable contributions to the study are much appreciated.

To my fiancé, Thami Tsolekile, thank you for your patience, constant

Unit, MRC thank you for your support.

The National Research Foundation for funding the study.

Table of Contents vii

Contributions of thesis xvi

CHAPTER 3: Ultraviolet (UV) detection of fumonisin B analogues as

5.1 Introduction 116

CHAPTER 6: Comparisons of methods, General Discussion,

Table 3.3 Determination of the detection limits using the amount 70

Table 3.4 Recoveries with IAC clean-up 72

Table 3.5 Fumonisin levels (µg/kg) in naturally contaminated 73

Table 3.6 Fumonisin levels (µg/kg) in naturally contaminated 74

Table 4.1 Intra-day precision of fumonisin working standards 88

Table 4.2 Inter-day precision of fumonisin working standards 89

Table 4.3 Amount (ng) injected into HPLC Column 90

Table 4.4 Limits of detection (LOD) and quantification (LOQ) 90

Table 4.5 Stability of FB-NDA standard after six consecutive 92

Table 4.6 Solvent extraction efficiency for NDA derivatization of 102

Table 4.10 Comparison of HPLCFLD

Table 5.2 Inter-day precision using peak areas of working 123

Table 5.3 Limits of detection (LOD) and quantitation (LOQ) in 124

Table 5.4 Effect of wavelength on working standard peak areas 127

Table 5.5 Effect of wavelength on working standard peak areas 127

Table 5.7 Determination of the recovery of the derivatization 131

Table 5.8 Effect of reaction temperature on maize derivatized for 135