Accepted Manuscript

Effect of extraction pH on heat-induced aggregation, gelation and microstruc-

ture of protein isolate from quinoa (Chenopodium quinoa Willd)

Geraldine Avila Ruiz, Wukai Xiao, Martinus van Boekel, Marcel Minor, Markus
Stieger

PII: S0308-8146(16)30569-6
DOI: http://dx.doi.org/10.1016/j.foodchem.2016.04.052
Reference: FOCH 19059

To appear in: Food Chemistry

Received Date: 22 June 2015

Revised Date: 4 April 2016
Accepted Date: 15 April 2016

Please cite this article as: Ruiz, G.A., Xiao, W., van Boekel, M., Minor, M., Stieger, M., Effect of extraction pH on
heat-induced aggregation, gelation and microstructure of protein isolate from quinoa (Chenopodium quinoa Willd),
Food Chemistry (2016), doi: http://dx.doi.org/10.1016/j.foodchem.2016.04.052

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Effect of extraction pH on heat-induced aggregation,
gelation and microstructure of protein isolate from
quinoa (Chenopodium quinoa Willd)

Geraldine Avila Ruiz a,*, Wukai Xiao b, Martinus van Boekel b, Marcel Minor c

and Markus Stieger a

a
Food and Biobased Research, Wageningen University and Research Centre, P.O. Box 17,

6700 AA Wageningen, Netherlands

b
Food Quality and Design Group, Wageningen University and Research Centre, P.O. Box

9101, 6700 HB Wageningen, Netherlands

c
Division of Human Nutrition, Wageningen University and Research Centre, P.O. Box 8129,

6700 EV Wageningen, Netherlands

* Corresponding author

E-mail address: geraldine.ruiz@wur.nl (Geraldine Avila Ruiz)

Phone: +31 317 482022

1
Abstract: The aim of this study was to determine the influence of extraction pH on heat-

induced aggregation, gelation and microstructure of suspensions of protein isolates extracted

from quinoa (Chenopodium quinoa Willd). Quinoa seed protein was extracted by alkaline

treatment at various pH values (pH 8 (E8), 9 (E9), 10 (E10) and 11 (E11)), followed by acid

precipitation. The obtained protein isolates were freeze dried. The protein isolates E8 and E9

resulted in a lower protein yield as well as less protein denaturation. These isolates also had a

higher protein purity, more protein bands at higher molecular weights, and a higher protein

solubility in the pH range of 3 to 4.5, compared to the isolates E10 and E11. Heating the 10%

w/w protein isolate suspensions E8 and E9 led to increased aggregation, and semi-solid gels

with a dense microstructure were formed. The isolate suspensions E10 and E11, on the other

hand, aggregated less, did not form self-supporting gels and had loose particle arrangements.

We conclude that extraction pH plays an important role in determining the functionality of

quinoa protein isolates.

Keywords: quinoa, protein, extraction, denaturation, solubility, aggregation, gelation

1. Introduction

Quinoa is an Andean grain that has recently been gaining in popularity around the world.

Quinoa is considered to have a high nutritional value, mainly because of the large amount of

good quality proteins (Abugoch et al., 2009). The total protein content of quinoa (12-23%) is,

on average, higher than that of rice, corn and barley. The amino acid profile of quinoa has

been reported to be better than most cereal and leguminous protein sources. Moreover, quinoa

is gluten-free. Therefore, proteins isolated from quinoa have the potential to be used to enrich

foods and beverages with protein, improving their nutritional value.

2
Quinoa protein isolates (QPI) consist mainly of 11S globulins (37% of total protein) and 2S

albumins (35% of total protein) (Brinegar & Goundan, 1993; Brinegar, Sine, & Nwokocha,

1996). Quinoa’s 11S globulin, also referred to as chenopodin, has a similar structure to

glycinin, the 11S globulin of soy. It is a hexamer consisting of six pairs of acid and basic

polypeptides. The acid and basic polypeptides have molecular weights of 20 to 25 kDa and 30

to 40 kDa, respectively, and are linked to each other by disulphide bonds. Quinoa’s 2S

albumin fraction consists of a heterogeneous population of polypeptides with molecular

weights of 8 to 9 kDa (Abugoch et al., 2009).

QPIs are obtained from quinoa grains by extraction under alkaline conditions, concentration

by acid precipitation and subsequent drying. The potential applications of QPIs in foods and

beverages depend on the functional properties of the QPIs, which are in turn affected by the

protein’s physical, chemical and structural properties (Abugoch, Romero, Tapia, Silva, &

Rivera, 2008; Aora & Alvarado, 2009). These properties are influenced by the extraction

conditions, such as the pH of the aqueous extraction liquid (Martínez & Añón, 1996;

Abugoch et al., 2008; Aora & Alvarado, 2009; Valenzuela, Abugoch, Tapia, & Gamboa,

2013). Obtaining QPIs at an extraction pH of 11 leads to protein denaturation, a higher protein

content, a lower solubility of the QPIs (in a pH range of 4 to 11), and a higher water imbibing

capacity, compared to QPIs extracted at a pH of 9 (Abugoch et al., 2008). Valenzuela et al.

(2013) also found extensive protein denaturation but, in addition to this, they observed

changes in aggregation, dissociation and structure of quinoa protein extracted at a pH higher

than 10. Aora and Alvarado (2009) observed an increasing protein yield as they increased the

extraction pH from 7.5 to 10.5. For amaranth protein isolates, an increase in the extraction

pH resulted in a decreased thermal stability for pH values of 8 and higher, and a decreased

enthalpy of denaturation at pH 11. Furthermore, extraction at pH 8 resulted in albumin-1 and

3
part of the globulins, whereas at a pH higher than 8, albumin-2, glutelin and the remaining

globulins were obtained (Martínez & Añón, 1996).

To the best of our knowledge, the heat-induced aggregation, gelation and microstructure of

QPIs have not yet been investigated. Only the cold-induced aggregation and gelation

properties (at pH 8.5 and 10.5) have been described (Mäkinen, Zannini, & Arendt, 2015). For

potential commercial applications of QPIs in foods and beverages, it is important to explore

the functional properties of QPIs that have not been further processed, both during and after

thermal treatment, as this simulates the processing that food products containing QPIs would

undergo.

These studies of functional properties were all carried out on bitter quinoa varieties. Sweet

quinoa varieties are saponin-free (<0.11%), and thus need to be processed less after

harvesting, which facilitates large-scale production (Wright, Pike, Fairbanks, & Huber, 2002;

Arendt & Zannini, 2013). Wageningen University and Research Centre in the Netherlands has

been developing sweet quinoa varieties suitable to be grown on a commercial scale in

northwest Europe (Limburg & Masterbroek, 1997; Mastebroek, van Loo, & Dolstra, 2002).

The functional properties of sweet quinoa varieties have not yet been studied. The post-

harvest removal of saponins from traditional bitter quinoa varieties has been demonstrated to

increase the protein efficiency ratio, but to decrease the nitrogen solubility, emulsifying and

foaming properties (Chauhan, Cui, & Eskin, 1999a; Chauhan, Eskin, & Mills, 1999b;

Lindeboom, 2005). Therefore, it is important to verify the influence of the inherent absence of

saponins on the functional properties, as well as on the underlying physico-chemical

properties and protein content, of protein isolates from sweet quinoa.

The aim of this study was to determine the effect of extraction pH on both the previously

studied QPI properties (protein purity, protein yield, molecular weight distribution, thermal

properties, solubility) and on the not-yet-studied heat-induced properties (aggregation,

4
gelation, microstructure) of suspensions of QPIs obtained from sweet quinoa. We used the

sweet quinoa variety Atlas, which is based on breeding lines designed and tested by

Mastebroek et al. (2002), and which shows a good agronomic performance.

2. Material and methods

2.1. Materials

Quinoa seeds (Chenopodium quinoa Willd) of the sweet variety Atlas were supplied by the

Agricultural Research Institute (INIA) in Santiago, Chile. Petroleum ether (boiling range 40-

60°C) was used (Sigma-Aldrich Chemie GmbH, Schnelldorf, Germany).

2.2. Preparation of quinoa protein isolates

Quinoa protein isolates were prepared using a modified method previously described

(Abugoch et al., 2008). Quinoa seeds were ground with a Fritsch Mill Pulverisette 14 (Idar-

Oberstein, Germany) using a speed of 7000 rpm, and sieved through a 200 µm sieve, to

produce flour. The flour was defatted in a soxhlet extractor for 24 hours, using petroleum

ether and 17% w/w flour (Pelgrom, Wang, Boom, & Schutyser, 2015). After defatting, the

petroleum ether was removed by evaporation. The defatted flour was suspended in deionized

water (10% w/w), and the pH adjusted to 8, 9, 10 and 11 by the addition of 2 N NaOH. These

suspensions were stirred for 4 hours at room temperature and stored at 4°C for 16 hours to

maximize protein solubilization. Then the suspensions were centrifuged at 10°C for 30 min at

6000g. The pH of the supernatants was adjusted to pH 4.5 using 2N HCl, and the supernatants

were centrifuged for 30 min at 13000g and 10°C. The precipitated pellets were re-suspended

in deionized water (5% w/w). To rinse remaining salts the suspensions were centrifuged for

30 min at 13000g and 10°C, re-suspended in deionized water (5% w/w) and neutralized by the

5
addition of 2 N NaOH. The suspensions were frozen by dipping them into liquid nitrogen, and

were subsequently freeze-dried for 72 h (Chris Epsilon 2-6D Freeze Dryer, Osterode am Harz,

Germany). Finally, the dried protein isolates were ground with a kitchen blender for 1 min to

turn them into powder.

2.3. Determination of protein yield and purity

Amounts of 8 to 15 mg QPI were weighed in tin cups and dried overnight at 60°C. The

nitrogen content was determined using the Dumas methodology by sample combustion in a

Dumas Flash EA 1112, Series NC analyzer (Wigan, UK), and converted to a crude protein

percentage using a protein factor of 5.85 (Becker et al., 1981; Castellani, Martínez, & Añón,

1998; Abugoch et al., 2008). Measurements were performed in duplicate for isolates obtained

in duplicate from two separate extractions. The protein yield was calculated as follows:

     % ×   

Protein yield (%) =      % ×   
× 100

     % ×   

Protein purity (%) =   
× 100

The protein loss was calculated as follows:

                 

Protein loss (%) =      % ×   
× 100

2.4. Determination of molecular weight distribution

Sodium dodecyl sulphate–polyacrylamide gel electrophoresis (SDS–PAGE) was used to

determine the molecular weight distribution of the quinoa protein isolate fractions, using a

method previously described (Yi et al., 2013). Polyacrylamide gel electrophoresis (PAGE)

was performed using a NuPAGE Electrophoresis System (Invitrogen Corp., Carlsbad, CA).

First, the protein suspensions (1% w/w) were prepared in deionized water (pH 6.5±0.1) and

centrifuged for 1 min at 13000g. Then the supernatants were diluted with 1 x NuPAGE® LDS

6
Sample Buffer and deionized water, before applying the samples to the gel. NuPAGE®

Novex® Bis-Tris Gels (1–200 kDa), containing 12% acrylamide (4% acrylamide stacking

gel), were used. The molecular weight markers were from NuPAGE® Novex® (Mark 12™

Unstained Standard, 2.5–200 kDa). The protein bands produced by the electrophoresis were

stained with Simply Blue™ SafeStain.

2.5. Solubility measurements

The solubility of the QPIs was determined using a modified method previously described

(Abugoch et al., 2008). The QPIs were suspended in deionized water (1% w/w) and stirred for

1 h at room temperature (pH 6.5±0.1). The suspensions were mixed with an Ultra Turrax for 3

min at 4000 rpm, and homogenized (Labho Scope Homogenizer, Delta Instruments, Drachten,

Netherlands) at 150 bar for 10 runs. The homogenized suspensions were adjusted to a pH

range from 3 to 9, and centrifuged for 30 min at 8500g and 10°C. The protein purity of the

supernatants was determined as described in section 2.3., using 200 µl of sample.

Measurements were performed in duplicate for isolates obtained in duplicate. The solubility at

each pH was calculated as follows:

        % ×      

Solubility (%) = × 100
     % ×   

We define solubility as the percentage protein remaining in solution (protein solubility) after

centrifuging the protein suspension for 30 min at 8500g and 10°C, using a Centrifuge 5430 R

(Eppendorf AG, Hamburg, Germany), assuming that not all protein is molecularly dissolved

but in suspension. To obtain the mass of the supernatant, the supernatant was weighed.

7
 2.6. Particle size determination

The protein suspensions (1% w/w) were prepared in the same way as for the solubility

analysis, for a pH range of 3 to 9. Instead of centrifuging, the suspensions were filtered with a

0.45 µm diameter filter. The particle size of the filtrates was quantified with a Malvern

Zetasizer Nano (Malvern Instruments, Worcestershire, UK), using a modified method

previously published (Zhou, Qi, Neil Lewis, & Carpenter, 2015). The z-averaged

hydrodynamic diameter (z-average) in nm was recorded. Data were collected at 20°C using a

material refractive index of 1.45, a dispersant refractive index of 1.330 and a measurement

angle of 173° (backscatter). For each sample, three measurements were performed.

Measurements were performed in duplicate for isolates obtained in duplicate.

2.7. Determination of thermal properties

The thermal properties of the QPIs were assessed by Differential Scanning Calorimetry

(DSC), using a modified method previously described (Abugoch et al., 2008). Hermetically

sealed aluminum pans were filled with 25-50 mg of 20% w/w suspensions of isolates,

dispersed in deionized water. The DSC samples were heated from 20 to 140°C at a rate of

10°C/min, using a PerkinElmer Diamond series differential scanning calorimeter, equipped

with an intracooler 2P. A double, empty pan was used as reference. The denaturation

parameters were calculated using Pyris Software (Version 11, PerkinElmer), with the

denaturation temperature (Td) value corresponding to the maximum transition peak, and the

transition enthalpy (∆H) calculated from the area below the transition peaks. Measurements

were performed in duplicate for isolates obtained in duplicate.

8
 2.8. Effect of heating on particle size and gelation properties

For the particle size measurements, 1% w/w suspensions were prepared as described in

Section 2.6. The suspensions were filtered through a 0.2 µm-diameter filter. The

measurements were made at temperatures from 20 to 90°C, at intervals of 10°C, with an

equilibration time of 5 min after each heating step. To avoid evaporation, the samples were

covered with a thin layer of paraffin oil and sealed with a plastic stopper.

For the gelation measurements, a modified previously described method was used (Yi et al.,

2013). The protein suspensions (10% w/w) were prepared in deionized water and stirred for 1

h at room temperature (pH 6.5±0.1). Oscillatory strain tests were performed using a stress-

controlled rheometer (Physica MCR 300, Anton Paar, Graz, Austria) equipped with stainless

steel and titanium concentric cylinder geometry (CC-10, diameter inner cylinder: 10.00 mm;

diameter cup: 10.845 mm). To prevent evaporation, samples were covered with a thin layer of

paraffin oil. The samples were heated from 20 to 90°C at a heating rate of 1°C/min, kept at

90°C for 5 min, and cooled to 20°C at a rate of 3°C/min. During the temperature ramp, the

storage modulus G’ and loss modulus G’’ were determined, by applying a strain amplitude of

1% at a frequency of 0.1 Hz. The temperature at which G’ started to increase considerably

and became greater than the background noise was designated as the gelation temperature

(Renkema, Knabben, & van Vliet, 2001). Measurements were performed in duplicate for

isolates obtained in duplicate.

2.9. Determination of microstructure of heat treated quinoa protein isolates

The microstructure of the heat-treated QPIs was analyzed using a modified method previously

described (Auty, Twomey, Guinee, & Mulvihill, 2001). Suspensions of the isolates were

prepared in the same way as for the gelation measurements, except that rhodamine B was

9
added to the suspensions before heat treatment. After performing the oscillatory strain tests,

the micrographs of the heat-treated suspensions E8, E9, E10 and E11 were obtained using a

Confocal Scanning Light Microscope (Zeiss LSM510, Jena, Germany), with an excitation

wavelength of 488 nm, emission channel 1 of ≥635 nm (red), emission channel 3 of 545-635

nm (green) and an emission channel 2 of 505-545 nm (cyan). The resolution of the obtained

micrographs was 250 x 250 μm.

2.10. Statistical data analysis

Statistical data analysis was performed using SPSS (V19, SPSS Inc., Chicago, IL, USA).

One-way ANOVA followed by least-squares difference posthoc testing (LSD) were

performed to identify significant differences between mean values. A significance level of p

< 0.05 was chosen.

3. Results and discussion

3.1. Protein yield and purity

The protein yield significantly increased as the extraction pH increased (F(3,4)=205.5; p <

0.001), from 36.3 % (g protein/100 g flour) for E8, to 52.0 % for E11 (Figure 1). This

suggests that the solubility of the proteins increased in more extreme alkaline conditions

(Lestari, Mulder, & Sanders, 2010; Valenzuela et al., 2013). At a more alkaline pH, proteins

are increasingly negatively charged due to ionization of the carboxyl groups and

deprotonation of the amine groups. As a result, electrostatic repulsion between the negatively

charged proteins is enhanced. This increases protein-water interactions and thereby protein

solubility.

10
The protein yield range is in agreement with a previous study on bitter quinoa (Chenopodium

quinoa Willd) from which a protein yield of 47 % at extraction pH 8 and 0.5 N NaCl was

estimated (Brinegar & Goundan, 1993). The protein yields of the present study are slightly

lower than the ones calculated based on the data of Aora & Alvarado (2009). A very recent

study reported a maximum protein yield of 76.3% at extraction pH 11 and 0.1 N NaCl

(Guerreo-Ochoa, Pedreschi, & Chirinos, 2015). This is a very similar maximum yield to that

obtained in the present study (at pH 11 yield is 74.3%). For other protein sources, such as

paprika and soybean, Guerreo-Ochoa et al. (2015) reported maximum protein yields of 12.2%

and 33.0% for extraction pH 9. The protein yield of QPI E9 calculated in the same way was

63.1%. An increase in protein yield with increasing extraction pH was also found by Aora and

Alvarado (2009) for quinoa protein, and by Martínez and Añón (1996) for amaranth protein.

The vast majority of protein was lost during the alkalinization and precipitation steps (Figure

1). In the alkalinization step, the protein loss decreased with increasing extraction pH. The

protein yield increased with extraction pH by about the same ratio as the protein loss in the

alkalinization step decreased. This indicates that more protein was solubilized from the grain

matrix and ended up in the final protein concentrate.

Protein purity of the QPIs significantly decreased with increasing extraction pH (F(3,4)=9.9; p

< 0.05) from 88% for E8, to 82% for E11. The decrease in purity may be caused by an

increase in the amount of non-protein components co-precipitating with the protein isolates at

pH values higher than 9, as theorized by Lestari et al. (2010).

The purity of the saponin-free QPIs obtained in our study was higher than the values

previously reported in literature (52 to 85%), even with some studies that used protein factor

of 6.25, as compared to the protein factor of 5.85 used in the present study (Chauhan et al.,

1999a; Aluko & Monu, 2003; Lindeboom, 2005; Abugoch et al., 2008; Aora & Alvarado,

11
2009). The higher protein purity in the present study might be due to a longer alkalinization

time (16 h) than in most other studies (8-120 min), which allowed more protein to be

solubilized from the grain.

The protein purity decreased slightly with increasing extraction pH, in contrast to results

shown in the literature (Martínez & Añón, 1996; Abugoch et al., 2008; Aora & Alvarado,

2009).

3.2. Molecular weight distribution

The SDS-PAGE analysis (Figure 2) showed numerous bands of varying intensity in the

protein isolates E8, E9, E10 and E11. There are bands at 6kDa, 33kDa, 38 kDa, and 50kDa.

For E8, E9 and E10, the most intense bands were found at 50 kDa. These bands could

correspond to 11S globulin (Brinegar & Goundan, 1993; Abugoch et al., 2009). The bands for

E11 were more diffuse and, at lower molecular weights, were more pronounced than the

bands of the protein extracts obtained at lower pH. For E8, the high molecular weight

fractions dominated the 6kDa fractions. As the extraction pH increased, the protein fractions

of lower molecular weight became more prominent, and for E11 they dominated the 50kDa

fractions. The SDS profiles indicated that globulin and other high molecular weight protein

fractions could be obtained at extraction pH values ranging from 8 to 10. At extraction pH 11

these fractions might have been hydrolyzed into fractions with lower molecular weights, as

well as associated through increased hydrophobic interactions and intermolecular disulfide

bonds into insoluble aggregates that were removed by centrifugation before performing the

electrophoresis (Mauri & Anon, 2008; Lestari et al., 2010; Valenzuela et al., 2013). This

would explain the fainter bands at high and intermediate molecular weights for a higher

extraction pH. At higher pH values, proteins of lower molecular weight might be more

successfully extracted, similar to albumin-2 (Martínez & Añón, 1996).

12
For all SDS gels, a considerable amount of protein remained at the top of all lanes that did not

penetrate the gel. As 1-200kDa gel was used, this means that a considerable amount of

proteins with molecular weights higher than 200 kDa was present in the isolates and in the

defatted flour.

The SDS profiles of the QPIs were similar to profiles published by Abugoch et al. (2008) and

Valenzuela et al. (2013).

The profile of the defatted flour showed even more bands than the protein isolates, however,

their intensities were more evenly distributed, probably as a result of the much lower protein

concentration. Some of the flour’s protein fractions (66-116 kDa, 26-30kDa) were not visible

(or were hardly visible) in the isolates, while other fractions (50 kDa, 38 kDa and 33 kDa)

were much more prominent in the isolates (E8, E9 and E10) than in the flour. The comparison

of the isolates with the defatted flour shows that the alkaline extraction generated a different

protein composition to the one originally present in the quinoa grain.

3.3. Thermal properties of quinoa protein isolates

A single endotherm peak at around 97°C (denaturation temperature Td) was observed for E8,

E9 and E10, but not for E11 (Figure 3). This is in agreement with Abugoch et al. (2008), who

reported a single endotherm at 98°C for extraction pH 9, and no endotherm at extraction pH

11 for QPIs. Another study, analyzing protein isolates from amaranth, also observed

endotherms from 94 to 100°C for extraction pH 9 to 11 (Martínez & Añón, 1996). A single

peak generally suggests that the protein isolates consisted either of one protein species, or of

several species with similar thermostability. The SDS-PAGE results showed that globulin

appeared to be the most prominent protein species in isolates E8, E9 and E10. Furthermore,

isolated globulin from amaranth has been found to have a major endotherm at 97°C (Martínez

& Añón, 1996). Globulins from other plant sources have also been shown to have a Td in this

temperature range (soybean Td = 92°C, broadbean Td = 94°C, sunflower Td = 95°C) (Ma,

13
Harwalkar, & John, 1991). Therefore, it is very likely that the endotherm peak from the

present QPIs can be attributed to globulin. The high Td of quinoa globulin shows that the

protein is stable up to 97°C. This is the result of numerous remaining hydrophobic

interactions and disulfide bonds that connect globulin’s acidic and basic subunits to each

other. (Kinsella, Damodaran, German, & Wilcke, 1985; Martínez & Añón, 1996; Abugoch et

al., 2008).

There is no obvious relationship between the denaturation temperature and the extraction pH

of the QPIs (Figure 3). For amaranth protein, Martínez & Añón (1996) observed only a slight

overall decrease of Td (by 2-3°C) from extraction pH 8 to 11. Other studies, of suspensions

from amaranth and sunflower protein, reported a much sharper decrease in Td (by 10-20°C)

from pH 8 to 11 (Castellani et al., 1998; Molina, Petruccelli, & Añón, 2004). It seems that the

extraction pH has much less effect on Td than the pH of the protein suspension.

The denaturation enthalpies (∆H) for the isolates ranged from 0 to 10.2 J/g protein (Figure 3).

For E9 the denaturation enthalpy was 7.2 J/g, which is lower than the value that Abugoch et

al. (2008) reported for extraction pH 9 (12.4 J/g). For extraction pH 11, no endotherm could

be observed in the present study, which is in agreement with Abugoch et al. (2008). The lower

denaturation enthalpies compared to the literature might be due to the longer alkalinization

step used in our study (16 h in the present study compared to 30 min in the study of Abugoch

et al. (2008)), which led to more protein denaturation. The denaturation enthalpy is known to

be correlated with the content of ordered secondary structure of a protein (Wang,

Hettiarachchy, Qi, Burks, & Siebenmorgen, 1999). Alkaline treatment with subsequent acid

treatment decreases the extent of ordered secondary structure of proteins through disruption of

hydrogen bonds and hydrophobic interactions, to the point of irreversible changes in

conformation, leading to a more denatured state of the proteins (Martínez & Añón, 1996).

14
The denaturation enthalpy significantly decreased with increasing extraction pH (F(3,4)=47.8;

p < 0.001). A higher extraction pH leads to more protein denaturation, which reduces the

amount of heat necessary to denature the remaining native protein structure. The thermogram

of E11 indicates that the proteins were already denatured, as no endotherm could be detected.

The decrease in denaturation enthalpy with an increase in extraction pH is in agreement with

studies on quinoa, amaranth and sunflower protein (Martínez & Añón, 1996; Castellani et al.,

1998; Molina et al., 2004; Abugoch et al., 2008; ).

3.4. Solubility and particle size of QPIs

The solubility curves of the protein isolates E8, E9, E10 and E11 in aqueous solution, over a

pH range of 3 to 9, have an inverse bell shape (Figure 4). The solubility values for all isolates

ranged from 20 to 60% at pH 3, and from 35 to 73% at pH 7 to 9. The lowest protein

solubility, of around 5%, was observed at pH 4 to 6. The low solubility plateau can be

attributed to globulins, as they have been found to have the lowest solubility at pH 4 to 6

(Brinegar & Goundan, 1993; Martínez & Añón, 1996; González-Pérez & Vereijken, 2007;

Nishinari, Fang, Guo, & Phillips, 2014). Isolate E8 had the highest solubility at pH 3 and 4,

while E9 had the highest solubility at pH 7 and 8 compared to the other isolates (F(2,3)=27.0;

p < 0.05) with the exception of E10. The low solubility plateau was at a higher pH value for

E8 than for the other isolates. From soybean it is known that the association of the basic

subunit with the acidic subunit of the 11S soy protein tends to increase solubility of the basic

subunit (Kinsella et al., 1985). SDS-PAGE showed the highest amount of protein fractions

corresponding to intact 11S globulin for E8 and E9, which might explain the higher solubility

of E8 and E9 at many pH values, compared with E10 and E11. It is known that solubility

decreases with molecular weight and increases with surface polarity (Kinsella et al., 1985).

Therefore, the lower solubility of E10 and E11 may have resulted from their low molecular

15
weight protein fractions, and the greater degree of denaturation of proteins in general, leading

to the exposure of hydrophobic groups and thus decreased surface polarity. The consequence

would be increased protein aggregation, via hydrophobic interactions, to big, insoluble

particles, which is in line with the fainter bands of E10 and E11 for higher molecular weights

on the SDS-PAGE gel.

The solubility profiles are consistent with those of QPIs reported previously by Chauhan et al.

(1999a), Aluko & Monu (2003), Mäkinen et al. (2015), Aora & Alvarado (2009), and in

contrast to the solubility profiles reported by Abugoch et al. (2008), where solubility

increased with pH continuously from pH 3 to 11. The trend of a higher solubility at lower

extraction pH is in agreement with Abugoch et al. (2008), who observed a significantly higher

solubility for extraction pH 9 than for pH 11.

The z-averaged particle size for the QPIs varied from 50 to 3761 nm over a pH range of 3 to 9

(Figure 4). The z-averaged particle sizes above 450 nm in the pH range from 4.5 to 6 may be

explained by the occasional passage of particles larger than 450 nm through the filter (pore

size 450 nm), due to slightly more pressure applied to the syringe to filter the protein

suspensions, as a result of a higher resistance. This particularly occurred at pH values where

solubility was the lowest, and thus more big particles were present in the protein suspensions.

This hypothesis about the correlation of the high z-averaged particle sizes with the low

solubility plateau in the pH range of 4.5 to 6 is further confirmed by the observation that the

largest particle size of E8 shifted to a higher pH in the same way as its corresponding low

solubility plateau.

16
 3.5. Effect of heating on particle size and gelation behavior of quinoa protein isolates

The z-averaged particle size of the QPI suspensions at pH 6 remained constant up to 50°C

(Figure 5). From 60°C onwards, the z-averaged particle size was significantly higher for E8,

and especially E9, compared with E10 and E11 (F(3,8)=919.0; p < 0.001).

This suggests that heating induced protein aggregation at temperatures of 50°C and higher, for

particles (smaller than 450 nm) extracted at a low pH, while it induced less or no aggregation

for QPI particles extracted at a higher pH. It seems that the more denatured proteins resulting

from extraction at a higher pH (E10 and E11) could not undergo further association and

aggregation at higher temperatures, while the less denatured proteins resulting from extraction

at a lower pH (E8 and E9) still had the functional capacity to do so. The aggregation of E8

and E9 may be the consequence of increased disulfide bond formation. In line with this,

Mäkinen et al. (2015) reported significant reductions of the free and total SH group content of

QPI suspensions that had undergone heat-treatment at pH 8.5. We could also infer that an

extraction pH of 9 caused the highest degree of aggregation from 70°C upwards.

The G’ moduli of the isolate suspensions during heating and subsequent cooling are shown in

Figure 5. The G’ values increased considerably for E8 and E9 at around 70°C, while for E10

and E11 the G’ value increased only during the cooling phase. The gelation temperature of E8

and E9 (around 70°C) is similar to that of amaranth and pea protein isolates (O'Kane, Happe,

Vereijken, Gruppen, & van Boekel, 2004; Shevkani, Singh, Rana, & Kaur, 2014).

The G’ values at the end of the cooling phase for E8 (5000 Pa) and E9 (3300 Pa) were similar

to, or higher than, those for amaranth protein (up to 3800 Pa, 10% protein suspension,

extraction pH 9), pea protein (2000 Pa, 7.5 and 9.9% protein suspension, extraction pH 8) and

sunflower protein (500 Pa, 10% protein suspension, extraction pH 9) reported previously for

similar heating profiles (O'Kane et al., 2004; González-Pérez & Vereijken, 2007; Shevkani et

17
al., 2014). This suggests that stronger gels can be formed from quinoa protein than from other

plant proteins at comparable protein concentrations.

The G’ values of E10 and E11 showed that quinoa protein extracted under strongly alkaline

conditions did not gel during heating, and only formed a soft gel during cooling. A reason for

this seems to be the higher extent of protein denaturation, which may have led to flocculation

and sedimentation of larger particles, and to reduced aggregation of smaller particles,

resulting in a weaker tendency to form a network. The higher G’ values of E8 and E9

compared to E10 and E11 seemed to result from higher initial solubility and particle sizes,

favoring the interaction and aggregation of proteins during heating. When comparing these

results to the DSC results, we observed a difference between gelation temperature (around

70°C) and denaturation temperature (around 97°C) for E8 and E9. This difference may be

explained by an initial hydration and swelling of the proteins from 60°C to 70°C (as indicated

by an increasing particle size), leading to more protein-protein interactions. At about 75°C,

DSC thermograms showed the beginning of a heat flow decline with isolates E8 and E9,

indicating the start of a phase transition (protein unfolding). The sequence and overlap of the

two events could be responsible for an exponential rise of the degree of network formation.

3.6. Microstructure

The microstructure of the heat-treated QPIs differed considerably for E8 and E9, compared

with E10 and E11 (Figure 6). The heat treated suspensions of isolates E8 and E9 revealed

irregular particles of 15-30 µm embedded in a dense protein matrix, with larger pores

deprived in protein. By contrast, the suspensions of isolate E10 and E11 revealed many

particles of a smaller size (5-15 µm) and rounder shape, which seemed more loosely arranged

in a matrix, with a few small pores deprived in protein.

This suggests that the QPIs obtained at low extraction pH (E8 and E9) formed denser quinoa

protein networks during the heat treatment, via particle association, yielding an agglomerated

18
network structure. This structure seems to be responsible for the high G’ values. At a high

extraction pH, small particles do not seem to interact with each other, while big particles may

have flocculated into the background plane, giving the whole a more continuous structure. As

a result, this loose and inhomogeneous mass may explain the low G’ final values.

Mäkinen et al. (2015) observed a more irregular, aggregated gel structure, with larger pores,

for cold-induced QPI gels previously heat-treated at pH 8.5 compared to pH 10.5. This

morphology is similar to what CSLM pictures show in the present study at similar pH values,

but then of protein extraction instead of heat treatment post-extraction. The heat-treated QPI

suspensions from both studies differ, however, in their gelation behaviour, which reveals the

impact of varying the pH at different steps of QPI production and processing on a functional

level.

3.7. Conclusion

We conclude that the extraction pH affected the previously studied properties of QPIs (purity,

yield, molecular weight distribution, denaturation and solubility) in a similar way to literature

findings. The variation of heat-induced properties (aggregation, gelation and microstructure)

with extraction pH, which had not previously been studied, revealed new insights into these

properties. QPIs obtained from extraction at pH values below 9 could be used to prepare

semi-solid gelled foods. QPIs obtained from extraction at pH values higher than 10 lost the

capacity to form a strong gelled network upon heating. These QPIs could be used for

beverages or other liquid food applications. Future research could focus on finding such

applications for QPIs, but also on maximizing protein yield and purity, while minimizing

protein loss.

19
 4. Acknowledgments

The authors gratefully acknowledge the financial support of the “IPOP Customized Nutrition”

program of Wageningen University and Research Centre.

5. Abbreviations used

E8, protein isolated at pH 8; E9, protein isolated at pH 9; E10, protein isolated at pH 10; E11,

protein isolated at pH 11; QPI, quinoa protein isolate

6. References

Abugoch, L., Castro, E., Tapia, C., Añón, M. C., Gajardo, P., & Villarroel, A. (2009).

Stability of quinoa flour proteins (Chenopodium quinoa Willd.) during storage.

International Journal of Food Science and Technology, 44(10), 2013-2020.

Abugoch, L. E., Romero, N., Tapia, C. A., Silva, J., & Rivera, M. (2008). Study of some

physicochemical and functional properties of quinoa (Chenopodium quinoa willd)

protein isolates. Journal of Agricultural and Food Chemistry, 56(12), 4745-4750.

Aluko, R. E., & Monu, E. (2003). Functional and bioactive properties of quinoa seed protein

hydrolysates. Journal of Food Science, 68(4), 1254-1258.

Aora, J. C. C., & Alvarado, K. (2009). Aislados proteínicos de granos altoandinos

chenopodiaceas; quinua "chenopodium quinoa" – cañahua "chenopodium pallidicaule"

por precipitación isoeléctrica. Revista Boliviana de Química, 26(1), 12-20.

Arendt, E. K., & Zannini, E. (2013). Quinoa. 409-438.

20
Auty, M. A., Twomey, M., Guinee, T. P., & Mulvihill, D. M. (2001). Development and

application of confocal scanning laser microscopy methods for studying the

distribution of fat and protein in selected dairy products. J Dairy Res, 68(3), 417-427.

Becker, R., Wheeler, E. L., Lorenz, K., Stafford, A. E., Grosjean, O. K., Betschart, A. A., et

al. (1981). A Compositional Study of Amaranth Grain. Journal of Food Science, 46(4),

1175-1180.

Brinegar, C., & Goundan, S. (1993). Isolation and characterization of chenopodin, the 118

seed storage protein of Quinoa (Chenopodium quinoa). Journal of Agricultural and

Food Chemistry, 41(2), 182-185.

Brinegar, C., Sine, B., & Nwokocha, L. (1996). High-Cysteine 2S Seed Storage Proteins from

Quinoa (Chenopodium quinoa). Journal of Agricultural and Food Chemistry, 44(7),

1621-1623.

Castellani, O. F., Martínez, E. N., & Añón, M. C. (1998). Structural Modifications of an

Amaranth Globulin Induced by pH and NaCl. Journal of Agricultural and Food

Chemistry, 46(12), 4846-4853.

Chauhan, G. S., Cui, W., & Eskin, N. A. M. (1999a). Effect of saponin on the surface

properties of quinoa proteins. International Journal of Food Properties, 2(1), 13-22.

Chauhan, G. S., Eskin, N. A. M., & Mills, P. A. (1999b). Effect of saponin extraction on the

nutritional quality of quinoa (Chenopodium quinoa Willd.) proteins. J. Food Sci.

Tech., 36(2), 123-126.

González-Pérez, S., & Vereijken, J. M. (2007). Sunflower proteins: overview of their

physicochemical, structural and functional properties. J Sci Food Agric, 87(12), 2173-

2191.

21
Guerreo-Ochoa, M. R., Pedreschi, R., & Chirinos, R. (2015). Optimised methodology for the

extraction of protein from quinoa (Chenopodium quinoa Willd.). International Journal

of Food Science & Technology, n/a-n/a.

Kinsella, J. E., Damodaran, S., German, B., & Wilcke, A. M. A. L. (1985). V -

Physicochemical and Functional Properties of Oilseed Proteins with Emphasis on Soy

Proteins New Protein Foods (pp. 107-179): Academic Press.

Lestari, D., Mulder, W., & Sanders, J. (2010). Improving Jatropha curcas seed protein

recovery by using counter current multistage extraction. Biochemical Engineering

Journal, 50(1-2), 16-23.

Limburg, H., & Masterbroek, H. D. (1997). Breeding high yielding lines of Chenopodium

quinoa Willd. with saponin free seed. In O. Stølen, K. Bruhn, K. Pithan & J. Hill

(Eds.), Small Grain Cereals and Pseudo-Cereals (pp. 103–114). Copenhagen,

Denmark: Proc COST 814 workshop, 22–24 February 1996.

Lindeboom, N. (2005). Studies on the characterization, biosynthesis and isolation of starch

and protein from quinoa (Chenopodium quinoa Willd.). University of Saskatchewan,

Canada.

Ma, C. Y., Harwalkar, V. R., & John, E. K. (1991). Thermal Analysis of Food Proteins

Advances in Food and Nutrition Research (Vol. Volume 35, pp. 317-366): Academic

Press.

Mäkinen, O., Zannini, E., & Arendt, E. (2015). Modifying the Cold Gelation Properties of

Quinoa Protein Isolate: Influence of Heat-Denaturation pH in the Alkaline Range.

Plant Foods for Human Nutrition, 1-7.

Martínez, E. N., & Añón, M. C. (1996). Composition and Structural Characterization of

Amaranth Protein Isolates. An Electrophoretic and Calorimetric Study. Journal of

Agricultural and Food Chemistry, 44(9), 2523-2530.

22
Mastebroek, H. D., van Loo, E. N., & Dolstra, O. (2002). Combining ability for seed yield

traits of Chenopodium quinoa breeding lines. Euphytica, 125(3), 427-432.

Mauri, A. N., & Anon, M. C. (2008). Mechanical and Physical Properties of Soy Protein

Films with pH-Modified Microstructures. Food Science and Technology International,

14(2), 119-125.

Molina, M. I., Petruccelli, S., & Añón, M. C. (2004). Effect of pH and Ionic Strength

Modifications on Thermal Denaturation of the 11S Globulin of Sunflower (Helianthus

annuus). Journal of Agricultural and Food Chemistry, 52(19), 6023-6029.

Nishinari, K., Fang, Y., Guo, S., & Phillips, G. O. (2014). Soy proteins: A review on

composition, aggregation and emulsification. Food Hydrocolloids, 39, 301-318.

O'Kane, F. E., Happe, R. P., Vereijken, J. M., Gruppen, H., & van Boekel, M. A. J. S. (2004).

Heat-Induced Gelation of Pea Legumin: Comparison with Soybean Glycinin. Journal

of Agricultural and Food Chemistry, 52(16), 5071-5078.

Pelgrom, P. J. M., Wang, J., Boom, R. M., & Schutyser, M. A. I. (2015). Pre- and post-

treatment enhance the protein enrichment from milling and air classification of

legumes. Journal of Food Engineering, 155(0), 53-61.

Renkema, J. M. S., Knabben, J. H. M., & van Vliet, T. (2001). Gel formation by β-

conglycinin and glycinin and their mixtures. Food Hydrocolloids, 15(4–6), 407-414.

Shevkani, K., Singh, N., Rana, J. C., & Kaur, A. (2014). Relationship between

physicochemical and functional properties of amaranth (Amaranthus

hypochondriacus) protein isolates. International Journal of Food Science &

Technology, 49(2), 541-550.

Valenzuela, C., Abugoch, L., Tapia, C., & Gamboa, A. (2013). Effect of alkaline extraction

on the structure of the protein of quinoa (Chenopodium quinoa Willd.) and its

23
 influence on film formation. International Journal of Food Science & Technology,

48(4), 843-849.

Wang, M., Hettiarachchy, N. S., Qi, M., Burks, W., & Siebenmorgen, T. (1999). Preparation

and Functional Properties of Rice Bran Protein Isolate. Journal of Agricultural and

Food Chemistry, 47(2), 411-416.

Wright, K. H., Pike, O. A., Fairbanks, D. J., & Huber, C. S. (2002). Composition of Atriplex

hortensis, Sweet and Bitter Chenopodium quinoa Seeds. Journal of Food Science,

67(4), 1383-1385.

Yi, L., Lakemond, C. M., Sagis, L. M., Eisner-Schadler, V., van Huis, A., & van Boekel, M.

A. (2013). Extraction and characterisation of protein fractions from five insect species.

Food Chem, 141(4), 3341-3348.

Zhou, C., Qi, W., Neil Lewis, E., & Carpenter, J. F. (2015). Concomitant Raman spectroscopy

and dynamic light scattering for characterization of therapeutic proteins at high

concentrations. Analytical Biochemistry, 472(0), 7-20.

Figure 1. (A) Protein yield and protein purity on dry matter basis of the quinoa protein
isolates E8, E9, E10 and E11. (B) Protein loss expressed as amount of protein lost relative to
total protein in flour determined as protein content in the pellet of the alkaline suspension
(alkalinization), in the supernatant of the precipitated protein (precipitation) and in the
supernatant of the rinsed protein (rinsing) of the QPIs E8, E9, E10 and E11. Error bars
represent the standard deviation. Error bars represent the standard deviation.

Figure 2. SDS-PAGE profile of the QPIs E8, E9, E10, E11 and defatted quinoa flour. Lane
M: molecular weight marker; lane FL: defatted quinoa flour.

Figure 3. (A) DSC thermograms of QPIs E8, E9, E10 and E11. (B) Enthalpy (∆H) and
denaturation temperature (Td) of the QPIs. Error bars represent standard deviation.

24
Figure 4. Solubility (A) and z-averaged particle size (B) of the QPIs E8, E9, E10 and E11 in
suspension at pH values ranging from 3 to 9.

Figure 5. (A) Z-averaged particle size of the QPIs E8, E9, E10 and E11 in suspension at pH 6
as a function of temperature. (B) Storage modulus G’ of the QPIs E8, E9, E10 and E11 in
suspension (10% w/w) at pH 6.5 as a function of time. Heating and cooling temperature is
plotted as a secondary axis.

Figure 6. 10% w/w suspensions of the QPIs after heat treatment. Size of pictures is 250 x 250
µm. In green the protein phase is shown.

Conflict of interest

The authors declare no competing financial interest.

TOC graphic

pH 8-11

Quinoa Protein
grain isolate

25
Figure 1
Geraldine Avila Ruiz

26
Figure 2

Geraldine Avila Ruiz

27
Figure 3

Geraldine Avila Ruiz

28
Figure 4

Geraldine Avila Ruiz

29
Figure 5

Geraldine Avila Ruiz

30
Figure 6

Geraldine Avila Ruiz

Color online only

31
- As extraction pH increased, protein yield increased but protein purity decreased.

- As extraction pH increased, protein denaturation increased.

- As extraction pH increased, solubility decreased from pH 3 to 4.5.

- Extraction pH 8 and 9 allowed the formation of semi-solid gels upon heating.

32