Food Chemistry 129 (2011) 110–116

Contents lists available at ScienceDirect

Food Chemistry

journal homepage: www.elsevier .com/locate / foodchem

Influence of pH on the dry heat-induced denaturation/aggregation of whey proteins

Muhammad Gulzar, Saïd Bouhallab, Romain Jeantet, Pierre Schuck, Thomas Croguennec ⇑AGROCAMPUS OUEST, UMR 1253, F35000 Rennes, FranceINRA, UMR 1253, F35000 Rennes, France

a r t i c l e i n f o

Article history:Received 14 September 2010Received in revised form 21 February 2011Accepted 14 April 2011Available online 20 April 2011

Keywords:Whey proteinsDry heatingProtein cross-linkingDisulphide bondsProtein hydrolysis

0308-8146/$ - see front matter � 2011 Elsevier Ltd. Adoi:10.1016/j.foodchem.2011.04.037

⇑ Corresponding author at: AGROCAMPUS OUEST,France. Tel.: +33 (0) 2 23 48 59 27.

E-mail address: thomas.croguennec@agrocampus-

a b s t r a c t

The effect of pH on the heat-induced denaturation/aggregation of whey protein isolate (WPI) in the drystate was investigated. WPI powders at different pH values (6.5, 4.5, and 2.5) and controlled water activ-ity (0.23) were dry heated at 100 �C for up to 24 h. Dry heating was accompanied by a loss of soluble pro-teins (native-like b-lactoglobulin and a-lactalbumin) and the concomitant formation of aggregatedstructures that increased in size as the pH increased. The loss of soluble proteins was less when thepH of the WPI was 2.5; in this case only soluble aggregates were observed. At higher pH values (4.5and 6.5), both soluble and insoluble aggregates were formed. The fraction of insoluble aggregatesincreased with increasing pH. Intermolecular disulphide bonds between aggregated proteins predomi-nated at a lower pH (2.5), while covalent cross-links other than disulphide bonds were also formed atpH 4.5 and 6.5. Hence, pH constitutes an attractive tool for controlling the dry heat-induced denatur-ation/aggregation of whey proteins and the types of interactions between them. This may be of greatimportance for whey ingredients having various pH values after processing.

� 2011 Elsevier Ltd. All rights reserved.

1. Introduction

Although whey proteins are widely used as food ingredientsdue to the nutritional and textural properties they add to foodproducts, food technologists have developed processes to extendtheir functionalities. These processes include enzymatic (Kimet al., 2007; Panyam & Kilara, 1996; Rabiey & Britten, 2009), chem-ical (Kidwai, Ansari, & Salahuddin, 1976; Morgan et al., 1999) andphysical modifications (Considine, Patel, Anema, Singh, & Creamer,2007; Gulzar, Croguennec, Jardin, Piot, & Bouhallab, 2009; Patel,Singh, Havea, Considine, & Creamer, 2005). Amongst these pro-cesses, heat treatment in solution under controlled physicochemi-cal conditions was extensively studied (Croguennec, O’Kennedy, &Mehra, 2004; Donato, Schmitt, Bovetto, & Rouvet, 2009; Schmitt,Bovay, Rouvet, Shojaei-Rami, & Kolodziejczyk, 2007) and a correla-tion between the structural modifications of whey proteins and thequality of the final product was established (Alting, Hamer, De Kru-if, Paques, & Visschers, 2003; Alting et al., 2004; Havea, Watkinson,& Kuhn-Sherlock, 2009). In contrast, only a limited number of stud-ies deal with the dry heating of whey proteins (Enomoto et al.,2007, 2009; Ibrahim, Kobayashi, & Kato, 1993; Li, Enomoto, Ohki,Ohtomo, & Aoki, 2005), and it has been shown that the structuralmodifications in proteins during dry heating cannot be extrapo-lated from results obtained in solution (Povey et al., 2009).

ll rights reserved.

UMR 1253, F35000 Rennes,

ouest.fr (T. Croguennec).

Dry heating is known to be an efficient tool to modify the func-tionalities of egg white proteins, such as improvements in gelling,foaming and emulsifying properties (Desfougeres, Lechevalier,Pezennec, Artzner, & Nau, 2008; Kato, Ibrahim, Watanabe, Honma,& Kobayashi, 1989; Matsudomi, Takahashi, & Miyata, 2001; Mine,1997). It has been shown that only minor modifications in the sec-ondary structure of the protein and a slight increase in the acces-sibility of thiol groups and hydrophobic patches, resulting in theformation of soluble aggregates linked with intermolecular disul-phide bonds and also other covalent bonds, may improve the func-tional properties of egg white proteins (Kato, Ibrahim, Watanabe,Honma, & Kobayashi, 1990; Matsudomi et al., 2001; Watanabe,Nakamura, Xu, & Shimoyamada, 2000). Dry heating is usually con-ducted at pH 7–9, which is the natural pH range for egg white pro-teins (Matsudomi et al., 2001; Mine, 1996, 1997). Some resultsindicate that dry heating at acidic pH values can also modify pro-tein functionalities (Desfougeres et al., 2008; Li et al., 2005); how-ever, structural modifications in these conditions wereinadequately described.

Whey protein ingredients are usually obtained from whey withdifferent pH values, mainly acidic or neutral, and pH is known toaffect both the type and kinetics of chemical reactions taking placeduring preparation and subsequent processing such as dry heating(Povey et al., 2009). In this work, we studied the dry heating(100 �C for up to 24 h) of whey proteins under controlled wateractivity (aw 0.23) at three different pH values (2.5, 4.5, and 6.5)in order to better understand the effect of pH on the denatur-ation/aggregation mechanism of whey proteins in the dry state.

M. Gulzar et al. / Food Chemistry 129 (2011) 110–116 111

The physicochemical parameters of the powders were selected asfollows: An aw of 0.23 corresponds to the aw of whey protein pow-ders produced on an industrial scale; acidic pH values were chosento cover the diversity of commercially produced whey (pH 6.5 and4.5) and extended to pH 2.5 since whey proteins exhibit interestingstructural behaviours when heated in solution at this pH (Bolder,Hendrickx, Sagis, & van der Linden, 2006; Oboroceanu, Wang,Brodkorb, Magner, & Auty, 2010). In addition, at selected pH values(6.5, 4.5, and 2.5), the overall charge of the whey proteins is nega-tive, neutral and positive, respectively.

2. Material and methods

2.1. Materials

The spray-dried WPI (Prolacta, Lactalis Ingredient, Bourgbarré,France) contains 90.1 ± 1.0% proteins (w/w, determined by theKjeldahl method) of which 82% is b-lg and 18% is a-La (determinedby reversed-phase chromatography), 6.7 ± 0.2% moisture (deter-mined by air drying), 0.88 ± 0.08% lactose (determined by the Lac-tose/D-Galactose enzymatic method, Boehringer Mannheim,Darmstadt, Germany), 0.324 ± 0.016% calcium, 0.146 ± 0.012% so-dium (determined by atomic absorption spectroscopy),0.017 ± 0.002% chloride, 0.056 ± 0.005% succinate, 0.020 ± 0.002%sulphate, 0.037 ± 0.002% phosphate, and 0.043 ± 0.006% citrate(determined by ionic chromatography). Approximately 50% of b-Lg and 25% of a-La was lactosylated as assessed by mass spectrom-etry according to the protocol previously reported (Gulzar et al.,2009). Once reconstituted at 10 g l�1, the WPI solution had a pHof 6.5. The protein solubility at pH 7.0 and pH 4.6 (see Section 2.4.2for details) was 97 ± 3% and 93 ± 3% respectively. Glycine was fromAcross Organics (Geel, Belgium); all other chemicals were from Sig-ma–Aldrich (Saint–Quentin–Fallavier, France).

2.2. Preparation of powders

Spray-dried WPI was dissolved in distilled water at a proteinconcentration of 15% and the solution was adjusted to three differ-ent pH values (2.5, 4.5, and 6.5) by using HCl. The solutions werethen lyophilised. The samples containing 10 g of powder werestored for two weeks in a desiccator containing a saturated salt(CH3CO2K) solution in order to maintain a water activity of 0.23.The water activity of the powder was checked using an aw meter(Novasina RTD 200/0 and RTD 33, Pfäffikon, Switzerland).

2.3. Preparation of samples

Powders with three different pH values and an aw of 0.23 wereheated at 100 �C for 0, 8, 16 or 24 h in hermetically-sealed bottles.Subsequently, all the powders were reconstituted at 10 g l�1 in dis-tilled water containing an adequate concentration of NaCl (0.11,0.115, and 0.12 M for dry heated powders at pH 2.5, 4.5, and 6.5,respectively in order to compensate for salts [chloride and sodiumions] introduced during powder preparation and reconstitution atpH 7). All samples reached the same final ionic strength of0.12 M after reconstitution and pH adjustment at 7 by adding1 N NaOH (Solutions 1). Solutions 1 were centrifuged at 10,000gfor 15 min using an Eppendorf 5415C Micro Centrifuge (ScientificSupport, Hayward, California) in order to remove insoluble aggre-gates at pH 7. The supernatants (Solutions 2) contained solubleaggregates, residual native and ‘‘native-like’’ proteins. The pH ofSolutions 1 were lowered to 4.6 using 1 N HCl. Acidified sampleswere centrifuged at 10,000g for 15 min using an Eppendorf5415C Micro Centrifuge (Scientific Support, Hayward, California).Centrifugation resulted in the removal of both soluble and insolu-

ble aggregates at pH 7; subsequently, supernatants at pH 4.6 (Solu-tions 3) with only residual native and ‘‘native-like’’ proteins(soluble proteins at pH 4.6) were recovered. All samples were pre-pared in duplicate.

2.4. Physical analysis

2.4.1. Turbidity measurementProtein samples were diluted 10 times in 0.12 M NaCl solution.

Optical density (OD) at 500 nm was determined before and aftercentrifugation at pH 7 in spectroscopic plastic cuvettes (1 cm pathlength) using a Visible Spectrophotometer S1205 (Unico, France).Turbidity was determined by following the equation:s = (2.303 � OD)/l, where OD is the optical density of samples at500 nm and l is the path length of light in the cuvette.

2.4.2. Protein solubilityThe protein concentration in solutions 1, 2, and 3 was deter-

mined by the Lowry method (Lowry, Rosebrough, Farr, & Randall,1951). Protein solubility at pH 4.6 (Solutions 3) or pH 7 (Solutions2) was expressed as a percentage of the protein recovered in thesupernatants after centrifugation.

2.4.3. Determination of aggregate sizeThe size of the aggregates was determined by dynamic light

scattering using a Zetasizer NanoZS apparatus (Malvern Instru-ment, Worcestershire, UK), which was equipped with a He/Ne laserworking at 633 nm and an attenuator that automatically adjusts thelaser intensity to the specific range for scattered light detection.Protein samples were diluted in a phosphate buffer (0.05 M, pH 7,0.1 M NaCl) and placed in a 10 � 10 mm disposable polystyrene cell(Sarstedt, Germany) equilibrated at 20 �C for measurements.Heated samples were diluted 10 times, while non-heated sampleswere diluted three times to have sufficient signals for measure-ment. The intensity of scattering is detected at 173� (backscatterdetection) to reduce multiple scattering. The hydrodynamic diame-ter of the aggregates was calculated using the Stockes–Einsteinequation, taking the calculated diffusion coefficient from the fit ofthe correlation curve. All the samples were measured in triplicate.

2.5. Chemical analysis

2.5.1. Gel permeation chromatographySoluble proteins at pH 7 were analysed by High Pressure – Gel

Permeation Chromatography (HP-GPC) using a TSK G3000 SWXL(300 � 7.8 mm i.d.) column (Phenomenex, Le Pecq, France) con-nected to a Waters chromatography system (Milford, USA), con-sisting of a Waters e2695 Separation Module, and a Waters 2489Dual k Absorbance Detector and Empower chromatography appli-cation software to acquire, process and report chromatographicinformation. A 0.05 M phosphate buffer at pH 7 containing 0.1 MNaCl was used to equilibrate the column and to elute the proteinsat a flow rate of 0.8 mL min�1. Proteins were detected at 214 nm.

2.5.2. SDS–PAGE analysisSDS–PAGE was performed under reducing (with DTT) and non-

reducing conditions (without DTT) using a Mini Protean II system(Bio-Rad Laboratories, F Technologies, Dublin, Ireland) as de-scribed by Laemmli (1970) using a 12.5% acrylamide separatinggel and 4% concentration gel. Soluble proteins at pH 7 (Solutions2) were diluted 10-fold with the denaturing buffer (77.975%0.08 M Tris–HCl, pH 6.8; 20% glycerol; 2% SDS; 0.025% bromophe-nol blue). The proteins (10 lg) were loaded into the sample slotsand separated at 75 V for 30 min and at 150 V for 60 min. Gelswere stained with Coomassie Brilliant Blue G250. A low molecularweight marker kit (14.4–94 kg mol�1, Amersham Biosciences,

112 M. Gulzar et al. / Food Chemistry 129 (2011) 110–116

France) was used for molecular weight (MW) calibration. The gelswere scanned by Image Scan II (Amersham, Bioscience). The mono-meric bands of b-lg and a-La were quantified using densitometrysoftware, Image Quant TL 1D (Amersham, Bioscience).

2.5.3. Sulfhydryl quantificationThe quantification of sulphhydryl groups was done according to

the Ellman method (Ellman, 1959). A volume of 100 ll of Solutions2 (soluble proteins at pH 7) was diluted with 900 ll of Tris–glycinebuffer (0.05 M, pH 7) without and with SDS (0.5%), and 25 ll of2,2’-Dinitro-5,50-dithiodibenzoate (DTNB, Merck, Darmstadt, Ger-many) was then added. Total (with SDS) and accessible (withoutSDS) sulphydryl (SH) groups were quantified at 412 nm after180 min of reaction with DTNB using a specific extinction coeffi-cient of 13,600 M�1 cm�1.

3. Results

3.1. Composition of dry heated WPI

The protein composition of WPI powders, dry heated at variouspH values is summarised in Table 1. Non-heated whey protein

Table 1Composition of whey protein powders before and after dry heating treatment at pH 2.5, 4

pH of powders Heating time (hours) Turbidity (500 nm) Insoluble aggreg

2.5 0 0.01 –8 0.01 –

16 0.07 –24 0.14 ± 0.04 –

4.5 0 – –8 0.07 –

16 0.20 ± 0.01 6 ± 224 0.44 ± 0.13 29 ± 2

6.5 0 0.01 –8 0.09 ± 0.01 –

16 0.38 ± 0.04 32 ± 224 0.57 ± 0.02 52

-0,01

0,09

0,19

0,29

0,39

0,49

5 8 11 14 17Time (min)

Abs_

214

nm

Oligomers

8 h

0 h

16 h

24 h

M1

M2

Polymers

B A

-0,01

0,09

0,19

0,29

0,39

0,49

5 8Time

Abs_

214

nm

Fig. 1. HP-GPC profile of dry heated whey proteins (soluble fraction at pH 7) with dry heline), DH-WPI-16 h (green line) and DH-WPI-24 h (blue line). M1 corresponds to a-lactalactoglobulin. (For interpretation of the references to colour in this figure legend, the re

isolate (NH-WPI) samples developed no turbidity after rehydra-tion, regardless of the pH (2.5, 4.5, and 6.5) of the whey proteinsolutions (15% w/w) prior to lyophilisation. They contained noinsoluble aggregates (protein fraction removed by centrifugationat pH 7) irrespective of the pH used for WPI preparation, and con-tained approximately 90% of soluble proteins at pH 4.6, indicatingthat about 10% of proteins exist as soluble aggregates at pH 7, pos-sibly resulting from prior processing.

Almost no change in turbidity was observed for the whey pro-tein isolate (DH-WPI) at pH 2.5 that was dry heated for up to16 h. However, after 24 h of dry heating, the turbidity increasedslightly. The turbidity further increased when the pH of the WPIwas set to 4.5 and 6.5 (for the dry heating treatment).

Dry heating induced a progressive loss of soluble proteins at pH4.6. The kinetics of the loss of soluble proteins at pH 4.6 was almostthe same for the DH-WPI samples at pH 4.5 and 6.5 but differedsignificantly from that of the DH-WPI sample at pH 2.5. Hence,after 24 h of dry heating, the amount of soluble proteins at pH4.6 in DH-WPI (pH 4.5 and 6.5) was 20 ± 1% and 23%, respectivelycompared to 42 ± 1% in DH-WPI (pH 2.5).

The loss of soluble proteins at pH 4.6 was accompanied by theformation of soluble and/or insoluble aggregates. The amount ofsoluble aggregates in DH-WPI (pH 2.5) increased continuously

.5 and 6.5.

ates at pH 7 (%) Soluble aggregates at pH 7 Soluble protein at pH 4.6

%age Size (nm)

12 ± 3 65 ± 3 88 ± 325 ± 3 91 ± 4 75 ± 342 ± 4 75 ± 7 58 ± 458 ± 1 100 ± 30 42 ± 1

8 ± 1 65 ± 16 92 ± 143 ± 2 54 ± 14 57 ± 267 ± 3 87 ± 4 27 ± 151 ± 3 118 ± 32 20 ± 1

10 ± 2 56 ± 1 90 ± 251 ± 1 94 ± 6 49 ± 139 ± 3 117 ± 13 29 ± 125 220 ± 86 23

C

11 14 17 (min)

-0,01

0,09

0,19

0,29

0,39

0,49

5 8 11 14 17Time (min)

Abs_

214

nm

ating pH 2.5 (A), pH 4.5 (B) and pH 6.5 (C). NH-WPI (black line), DH-WPI-8 h (brownlbumin monomer, while M2 corresponds to monomer and dimer equilibrium of b-ader is referred to the web version of this article.)

M. Gulzar et al. / Food Chemistry 129 (2011) 110–116 113

for up to 24 h of dry heating. In contrast, the denaturation/aggrega-tion of whey proteins for DH-WPI (pH 4.5 and 6.5) was accelerated,resulting in a rapid initial increase in the amount of soluble aggre-gates. After longer heating times, insoluble aggregates instead ofsoluble aggregates formed. A greater amount of insoluble aggre-gates was observed in DH-WPI (pH 6.5). In addition, the resultsindicated that the average size of the aggregates increased withpH during the dry heat treatment.

The insoluble aggregates caused turbidity since all the sampleswere transparent at 500 nm after centrifugation at pH 7 (data notshown). For DH-WPI (pH 2.5), no loss of solubility was seen evenafter 24 h of dry heating. Hence, the slight turbidity observed at500 nm after 24 h of dry heating may be due to insoluble aggre-gates, the concentration of which was below the threshold for pro-tein quantification using the Lowry method. In contrast, asignificant amount of insoluble aggregates was formed after 16 hof dry heating at pH 4.5 and 6.5, reaching 29% and 52% of the totalprotein after 24 h of dry heating at pH 4.5 and 6.5, respectively(Table 1).

Fig. 2. SDS–PAGE profile of dry heated whey proteins (soluble fraction at pH 7) inthe absence (A) and presence (B) of reducing agent DTT. MW, low molecular weightmarkers; M1, a-lactalbumin monomer; M2, b-lactoglobulin monomer; and P,polymers of aggregated proteins.

0

20

40

60

80

100

pH 2.5 pH 4.5 pH 6.5Dry heating pH

% M

onom

eric

β-L

g

Fig. 3. Estimation of monomeric proteins (b-Lg and a-La) for samples dry heated for up tosoluble protein fractions at pH 7.0 (white bar) and pH 4.6 (black bar).

3.2. Characterisation of soluble aggregates

The average hydrodynamic diameter of soluble aggregates (pH7.0) was measured by dynamic light scattering (DLS). Table 1shows that for NH-WPI (pH 2.5, 4.5, and 6.5), the soluble aggre-gates have hydrodynamic diameters of 65 ± 3, 65 ± 16 and56 ± 1 nm, respectively. Dry heating increased the size of theaggregates, which increased further at higher pH values. The aver-age hydrodynamic diameters of the soluble aggregates in DH-WPI(pH 2.5, 4.5, and 6.5) after 24 h of dry heating were 100 ± 30,118 ± 32, and 221 ± 86 nm, respectively.

Fig. 1 shows the size exclusion chromatography profiles for NH-and DH-WPI samples. The three NH-WPI profiles were very simi-lar; they showed two major peaks eluting at 12.2 and 12.8 min,corresponding to the b-lactoglobulin monomer–dimer equilibriumand the a-lactalbumin monomer, respectively. Some oligomericforms were eluted between 10 and 11.5 min, while a chromato-graphic peak eluted at 7 min (excluded volume) indicated the pres-ence of polymers. The presence of oligomers and polymers in thecommercial WPI powders was in agreement with the quantifica-tion of soluble aggregates determined by the Lowry method. TheDH-WPI (pH 2.5) chromatographic profiles exhibited a significantdecrease in native b-lactoglobulin and a-lactalbumin with increas-ing the dry heating time (Fig. 1A). At the same time, the proportionof larger aggregates increased. The decrease in native b-lactoglob-ulin and a-lactalbumin was faster when the pH (of the powders)was higher than 2.5, which was in agreement with the solubilitymeasurements at pH 4.6. In addition, the polymer fraction at pH4.5 even just after 16 h of dry heating was much greater as com-pared to the polymer fraction observed after 24 h of dry heatingat low pH (2.5). However, after 24 h of dry heating at pH 4.5, thepolymer fraction was decreased, which indicated that the aggre-gates were centrifuged and confirmed the results presented in Ta-ble 1. This aggregation process was further accelerated for DH-WPI(pH 6.5).

SDS–PAGE analysis in the absence and presence of the reducingagent dithiothreitol (DTT) was used to determine the type ofintermolecular interactions in the formed aggregates. In the ab-sence of DTT (Fig. 2), the NH-WPI (pH 2.5, 4.5, and 6.5) samplesshowed two major bands corresponding to monomers of a-lactal-bumin and b-lactoglobulin. Small quantities of oligomers andpolymers were also observed, confirming the results of size exclu-sion chromatography. By increasing the dry heating time, theintensity of the bands corresponding to monomers of a-lactalbu-min and b-lactoglobulin decreased. The bands on top of the sepa-rating gel and in the wells, corresponding to large aggregates,

0

20

40

60

80

100

pH 2.5 pH 4.5 pH 6.5Dry heating pH

% M

onom

eric

α-L

a

24 h by SDS-SCAN from SDS–PAGE gels. Ratios were calculated based on recovered

0

10

20

30

0 8 16 24Heating Time (hours)

nm S

H /

mg

prot

ein

0

10

20

30

0 8 16 24Heating Time (hours)

nm S

H /

mg

prot

ein

0

10

20

30

0 8 16 24Heating Time (hours)

nm S

H /

mg

prot

ein

C

B

A

Fig. 4. Sulfhydryl groups quantification in the WPI dry heated at pH 2.5 (A), pH 4.5(B) and pH 6.5 (C) using the Ellman method. The full line represents accessible freesulphydryls in the absence of SDS, while the dashed line represents total freesulphydryls. The sulphydryls were corrected by taking into account the solubleamount of proteins at pH 7.

114 M. Gulzar et al. / Food Chemistry 129 (2011) 110–116

intensified. Intermediates were not observed in significant quanti-ties. Increasing the pH for dry heating accelerated the conversion ofmonomeric proteins (a-lactalbumin and b-lactoglobulin) into lar-ger aggregates, in agreement with size exclusion chromatographyand solubility measurements.

The estimation of monomeric proteins (a-lactalbumin and b-lactoglobulin) from SDS PAGE (with no reducing agent) in samplessoluble at pH 7 and pH 4.6 showed that the quantity of monomericproteins at pH 7 was only slightly greater than that at pH 4.6. Thisindicates that aggregates are mainly covalent. Samples containedonly a small amount of non-native monomers that precipitatedat pH 4.6 as already shown for b-lactoglobulin (Croguennec, Bou-hallab, Molle, O’Kennedy, & Mehra, 2003; Surroca, Haverkamp, &Heck, 2002). The proportion of soluble monomers at pH 7, whichprecipitated at pH 4.6, was greater for a-lactalbumin, especiallyat pH 2.5 (Fig. 3).

To further investigate the type of covalent interactions, thesamples soluble at pH 7 were analysed by SDS–PAGE in the pres-ence of the reducing agent, DTT (Fig. 2). Under these conditions,the polymer bands observed in the absence of DTT (for NH-WPIsamples at pH 2.5, 4.5, and 6.5) were no longer detected in thepresence of DTT, indicating intermolecular disulphide bonds be-tween these proteins. The polymer bands observed in DH-WPI(pH 2.5) in the absence of DTT disappeared in the presence ofDTT (in profit of monomeric proteins), showing that the proteinsin these aggregates were joined together by disulphide bonds.However, the recovery of the proteins in the monomeric bandwas incomplete, and new bands above (16,200 Da) and below(12,100 Da,) the monomeric band of a-lactalbumin were observed.For DH-WPI (pH 4.5 and 6.5), a slight increase in the intensity ofthe monomer bands was observed, corresponding to the decreasein the intensity of the polymer band. This indicated that the pro-teins were held together by disulphide bonds. However, covalentbonds other than disulphide bonds also participated in proteincross-linking, as significant amounts of polymers were still concen-trated on top of the separating gel.

Intermolecular disulphide bonds were formed from either inter-molecular sulphydryl/disulphide interchange or oxidation of sul-phydryl groups between two molecules. Compared to sulphydryl/disulphide interchange, sulphydryl group oxidation led to a reduc-tion in total free sulphydryls. The latter were quantified usingDTNB in the presence of SDS. For NH-WPI (pH 2.5, 4.5 and 6.5), to-tal free sulphydryls were 27.3 ± 0.1, 28.4 ± 0.4, and 30.0 ± 0.3 nmo-les of SH mg�1 of protein, respectively (Fig. 4). Regardless of the pH,dry heating induced a significant decrease in total sulphydryls forall samples (20.4 ± 0.2, 20.8 ± 0.7, and 23.5 ± 0.1 nmoles of SHmg�1 of protein after 24 h of dry heating at pH 2.5, 4.5, and 6.5,respectively). This suggests that both sulphydryl/disulphide inter-change reactions and oxidation of sulphydryl groups occurred.Intermolecular disulphide bonds were formed in the powders dur-ing dry heating and did not result from reactions during resolubi-lisation of the powders at pH 7.0 since the direct dispersion of thedry heated powders in the SDS PAGE denaturing buffer (absence ofDTT) did not modify the electrophoretic bands (results not shown).At the same time, exposed sulphydryl groups (accessible to DTNBin the absence of SDS) increased. In fact, for NH-WPI (pH 2.5, 4.5,and 6.5), exposed sulphydryls were 5.3 ± 0.7, 4.1 ± 1.6, and5.5 ± 0.5 nmoles of SH mg�1 of protein, respectively, representingbetween 14% and 19% of exposed sulphydryl groups of whey pro-teins. For DH-WPI, a two-fold increase in exposed sulphydrylswas observed after 24 h of heating (Fig. 4).

4. Discussion

A considerable amount of research has been conducted on thedenaturation/aggregation of whey proteins in aqueous solution,but very little is known about the dry heating of whey proteins,especially the influence of the pH of the concentrate prior to dry-ing. Due to the reduced water content of powders, proteins aremore stable with regard to heat denaturation. The denaturationtemperature of the proteins in the WPI powder, adjusted to a wateractivity (aw) of 0.23, is 163 �C (Zhou & Labuza, 2007); a value sim-ilar to this was determined in the present study (data not shown).This temperature is significantly higher than the denaturation tem-perature for the same proteins in solution. b-Lg, the major wheyprotein that is largely responsible for the structural and functionalproperties of whey proteins in solution, has a denaturation tem-perature of around 80 �C, which can vary slightly depending onthe pH of the solution (Verheul, Roefs, & de Kruif, 1998). Decreas-ing the water content causes an increase in the glass transitiontemperature of powders. Compared to solutions, the molecular

-X-X-

Dry heating at

pH 4.5 & 6.5

-X-X

-

Native whey proteins

Insoluble aggregates

Soluble aggregates

PeptidesSoluble

aggregates

-S-S-

-S-S-

-S-S--S-S-

Non-native monomers

Non-native monomers

-S-S-

-S-S

-

-S-S

--X

-X-

Dryheatin

g

at pH2.5

Native monomers

Native monomers

-X-X-

Dry heating at

pH 4.5 & 6.5

-X-X

-

Native whey proteins

Insoluble aggregates

Soluble aggregates

PeptidesSoluble

aggregates

-S-S-

-S-S-

-S-S--S-S-

Non-native monomers

Non-native monomers

-S-S-

-S-S

-

-S-S

--X

-X-

Dryheatin

g

at pH2.5

Native monomers

Native monomers

Fig. 5. Proposed mechanism of how dry heating modulates the denaturation/aggregation mechanism of whey proteins. –S–S–, disulphide bond; –X–X– (present in solubleaggregates at pH 6.5), covalent bonds other than disulphide bonds.

M. Gulzar et al. / Food Chemistry 129 (2011) 110–116 115

motion in powders is retarded, especially at temperatures close tothe glass transition (Zhou & Labuza, 2007). Hence, the chemicalreactions involved in protein aggregation are slowed down as aconsequence of the substantial decrease in molecular motion.

The temperature selected for dry heating in this study was be-low the denaturation temperature of the proteins in the whey pro-tein powders. However, denaturation/aggregation of wheyproteins occurred in agreement with previous studies (Poveyet al., 2009; Zhou & Labuza, 2007). In addition to the effect of aw

in the powders and temperature for dry heat treatment (Zhou & La-buza, 2007), the rate of denaturation/aggregation of whey proteinswas pH-dependent also. The rate of denaturation/aggregation ofwhey proteins was faster and the hydrodynamic size of the aggre-gates increased with increasing pH. Lower quantities of solubleproteins at pH 4.6 and higher quantities of insoluble aggregateswere observed after dry heating whey proteins at pH 4.5 and 6.5compared to at pH 2.5. Working in a different pH range (neutralto alkaline), Mine (1996) suggested that increasing the pH fordry heating increased the rate of denaturation/aggregation of eggproteins and the molecular weight of the aggregates because pro-teins had a higher propensity for polymerisation through sulphyd-ryl/disulphide interchange reactions.

When heated above its denaturation temperature, protein isreadily unfolded, and hydrophobic patches and sulphydryl groupsinitially buried inside the protein are exposed, which may furtheradd to aggregation of proteins. Even if the denaturation tempera-ture of the proteins in the powders was higher than the dry heatingtemperature (100 �C), an increase in the accessibility of sulphydrylgroups initially buried inside the protein structures was observed.At elevated temperatures, chemical reactions such as hydrolysis,dehydration, deamidation, as well as intramolecular and intermo-lecular cross-links (Gerrard, 2002) involving functional groupsclose to each other may catalyse protein denaturation and proteinaggregation. Hydrolysis of whey proteins was clearly observed forDH-WPI at pH 2.5 and could be a preliminary step for proteinaggregation in powder, as suggested when heating proteins insolution at acidic pH (Akkermans et al., 2008). Compared to proteinunfolding in solution, the increase in exposed sulphydryl groups ismuch more limited in powders, regardless of the pH for dry heattreatment. The sulphydryl accessibility induced during dry heatingdetermined in this study is half that reported for WPI heated inaqueous solution at 68.5 �C for up to 2 h (Alting et al., 2003). Pro-tein unfolding is probably reduced due to lower protein mobilityin low water content systems.

Regardless of the pH for dry heating, intermolecular disulphidebonds linked whey proteins together within the aggregates. How-ever, significant amounts of intermolecular cross-links other thandisulphide bonds were also present within the aggregates formed

at pH 4.5 and 6.5. Intermolecular cross-links other than disulphidebonds were observed by other researchers in dry heated egg white(Kato et al., 1989; Watanabe, Xu, & Shimoyamada, 1999) and wheyproteins (Enomoto et al., 2009; Li et al., 2005). However, theirstructure remains unclear. Intermolecular isopeptide and lysino-alanine cross-links were suggested. These cross-links are ratherlimited in protein solutions where intermolecular disulphidebonds prevailed (Schokker, Singh, Pinder, & Creamer, 2000). Asfor disulphide bond formation, the rate of these reactions increaseswith increasing pH (Gerrard, 2002). Intermolecular cross-linksother than disulphide bonds were not observed at pH 2.5, indicat-ing that the rate of formation of these cross-links was even lowerthan for disulphide bond. Under the same pH condition in solution(pH 2.5), intermolecular disulphide bonds between whey proteinswere hindered, leading mainly to non-covalent aggregates due to alower reactivity of the sulphydryl groups (Schokker et al., 2000).However, the high concentration of sulphydryl groups in wheyprotein powders may compensate for the low reactivity of sul-phydryl groups at acidic pH, as indicated by Swaisgood (2005). Inaddition, when proteins are in powder form, sulphydryl groupsare also able to be oxidised more readily than in solution. Sincea-La has no free sulphydryl within its structure, it is unable to par-ticipate in sulphydryl oxidation. After destabilisation by dry heat-ing, a-La could be left as non-native monomers at pH 2.5, wheremainly disulphide bonds participated in protein cross-linking, asindicated in Fig. 3.

At fixed experimental conditions, pH clearly modulates thechemical reactions as well as the aggregation kinetics of WPI,resulting in products of different composition. Fig. 5 gives a sum-mary of the main findings of this study: whey protein denatur-ation/aggregation by dry heating at pH 2.5 results in productscontaining mainly disulphide-linked soluble aggregates, native-like proteins and peptides. In contrast, dry heating at pH 6.5 and4.5 favored larger insoluble and soluble aggregates containing pro-teins linked together by disulphide bonds as well as by intermolec-ular cross-links other than disulphide bonds, and a small amountof native-like proteins.

5. Conclusions

Dry heating slows down the process of denaturation/aggrega-tion of whey proteins compared to heating in solution. Restrictingprotein mobility by reducing water activity is one way of control-ling the rate of denaturation/aggregation of whey proteins. How-ever, some chemical reactions occurred, which were notobserved to the same extent in solution. The pH for dry heatinghas a significant influence on the rate of denaturation/aggregation

116 M. Gulzar et al. / Food Chemistry 129 (2011) 110–116

of whey proteins and the nature of the intermolecular cross-linksformed during aggregation as well as the size and solubility ofthe aggregates formed (Fig. 5). At pH 2.5, where soluble, smallaggregates were formed, protein–protein cross-links involvedmainly disulphide bonds. Moreover, under this condition, peptidesresulting from whey protein hydrolysis were observed. At higherpH values (4.5 and 6.5) disulphide bonds and other intermolecularcross-links were observed within whey protein aggregates. Theoccurrence of such reactions increased with increasing pH, result-ing in larger aggregates, some of which were insoluble.

Because of its prominent effect on the chemical reactions, thedenaturation/aggregation kinetics and the type of aggregatesformed, the pH for dry heating clearly constitutes a powerful toolto diversify and optimise the composition and consequently theproperties of whey protein based products. Given their composi-tional difference, the functional and nutritional properties of dryheated samples deserve consideration. For instance, dry heatingat acidic pH is well suited for applications where insoluble aggre-gates are not allowed.

Acknowledgements

The authors are grateful to the Higher Education Commission ofPakistan and INRA, France for financial support of this researchwork. We would also like to thank Julien Jardin for the mass spec-trometry analysis and Florence Rousseau for her help in DLSexperiments.

References

Akkermans, C., Venema, P., van der Goot, A. J., Gruppen, H., Bakx, E. J., Boom, R. M.,et al. (2008). Peptides are building blocks of heat-induced fibrillar proteinaggregates of +p

p� lactoglobulin formed at pH 2. Biomacromolecules, 9,

1474–1479.Alting, A. C., Hamer, R. J., De Kruif, C. G., Paques, M., & Visschers, R. W. (2003).

Number of thiol groups rather than the size of the aggregates determines thehardness of cold set whey protein gels. Food Hydrocolloids, 17, 469–479.

Alting, A. C., Weijers, M., De Hoog, E. H. A., van de Pijpekamp, A. M., Stuart, M. A. C.,Hamer, R. J., et al. (2004). Acid-induced cold gelation of globular proteins:Effects of protein aggregate characteristics and disulfide bonding on rheologicalproperties. Journal of Agricultural and Food Chemistry, 52, 623–631.

Bolder, S. G., Hendrickx, H., Sagis, L. M. C., & van der Linden, E. (2006). Fibrilassemblies in aqueous whey protein mixtures. Journal of Agricultural and FoodChemistry, 54, 4229–4234.

Considine, T., Patel, H. A., Anema, S. G., Singh, H., & Creamer, L. K. (2007).Interactions of milk proteins during heat and high hydrostatic pressuretreatments – A review. Innovative Food Science & Emerging Technologies, 8, 1–23.

Croguennec, T., Bouhallab, S., Molle, D., O’Kennedy, B. T., & Mehra, R. (2003). Stablemonomerie intermediate with exposed Cys-119 is formed during heatdenaturation of beta-lactoglobulin. Biochemical and Biophysical ResearchCommunications, 301, 465–471.

Croguennec, T., O’Kennedy, B. T., & Mehra, R. (2004). Heat-induced denaturation/aggregation of beta-lactoglobulin A and B: Kinetics of the first intermediatesformed. International Dairy Journal, 14, 399–409.

Desfougeres, Y., Lechevalier, V., Pezennec, S., Artzner, F., & Nau, F. (2008). Dry-heating makes hen egg white lysozyme an efficient foaming agent and enablesits bulk aggregation. Journal of Agricultural and Food Chemistry, 56, 5120–5128.

Donato, L., Schmitt, C., Bovetto, L., & Rouvet, M. (2009). Mechanism of formation ofstable heat-induced beta-lactoglobulin microgels. International Dairy Journal,19, 295–306.

Ellman, G. L. (1959). Tissue sulfhydryl groups. Archives of Biochemistry andBiophysics, 82, 70–77.

Enomoto, H., Hayashi, Y., Li, C. P., Ohki, S., Ohtomo, H., Shiokawa, M., et al. (2009).Glycation and phosphorylation of alpha-lactalbumin by dry heating: Effect onprotein structure and physiological functions. Journal of Dairy Science, 92,3057–3068.

Enomoto, H., Li, C. P., Morizane, K., Ibrahim, H. R., Sugimoto, Y., Ohki, S., et al. (2007).Glycation and phosphorylation of beta-lactoglobulin by dry-heating: Effect onprotein structure and some properties. Journal of Agricultural and FoodChemistry, 55, 2392–2398.

Gerrard, J. A. (2002). Protein–protein crosslinking in food: Methods, consequences,applications. Trends in Food Science & Technology, 13, 391–399.

Gulzar, M., Croguennec, T., Jardin, J., Piot, M., & Bouhallab, S. (2009). Coppermodulates the heat-induced sulfhydryl/disulfide interchange reactions of beta-lactoglobulin. Food Chemistry, 116, 884–891.

Havea, P., Watkinson, P., & Kuhn-Sherlock, B. (2009). Heat-induced whey proteingels: Protein–protein interactions and functional properties. Journal ofAgricultural and Food Chemistry, 57, 1506–1512.

Ibrahim, H. R., Kobayashi, K., & Kato, A. (1993). Improvement of the surfacefunctional-properties of beta-lactoglobulin and alpha-lactalbumin by heating ina dry state. Bioscience, Biotechnology, and Biochemistry, 57, 1549–1552.

Kato, A., Ibrahim, H. R., Watanabe, H., Honma, K., & Kobayashi, K. (1989). Newapproach to improve the gelling and surface functional-properties of dried egg-white by heating in dry state. Journal of Agricultural and Food Chemistry, 37,433–437.

Kato, A., Ibrahim, H. R., Watanabe, H., Honma, K., & Kobayashi, K. (1990). Structuraland gelling properties of dry-heating egg-white proteins. Journal of Agriculturaland Food Chemistry, 38, 32–37.

Kidwai, S. A., Ansari, A. A., & Salahuddin, A. (1976). Effect of succinylation (3-carboxypropionylation) on conformation and immunological activity ofovalbumin. Biochemical Journal, 155, 171.

Kim, S. B., Seo, I. S., Khan, M. A., Ki, K. S., Lee, W. S., Lee, H. J., et al. (2007). Enzymatichydrolysis of heated whey: Iron-binding ability of peptides and antigenicprotein fractions. Journal of Dairy Science, 90, 4033–4042.

Laemmli, U. K. (1970). Cleavage of structural proteins during the assembly of thehead of bacteriophage T4. Nature, 227, 680–685.

Li, C. P., Enomoto, H., Ohki, S., Ohtomo, H., & Aoki, T. (2005). Improvement offunctional properties of whey protein isolate through glycation andphosphorylation by dry heating. Journal of Dairy Science, 88, 4137–4145.

Lowry, O. H., Rosebrough, N. J., Farr, A. L., & Randall, R. J. (1951). Proteinmeasurement with the Folin phenol reagent. Journal of Biological Chemistry,193, 265–275.

Matsudomi, N., Takahashi, H., & Miyata, T. (2001). Some structural properties ofovalbumin heated at 80 degrees C in the dry state. Food Research International,34, 229–235.

Mine, Y. (1996). Effect of pH during the dry heating on the gelling properties of eggwhite proteins. Food Research International, 29, 155–161.

Mine, Y. (1997). Effect of dry heat and mild alkaline treatment on functionalproperties of egg white proteins. Journal of Agricultural and Food Chemistry, 45,2924–2928.

Morgan, F., Venien, A., Bouhallab, S., Molle, D., Leonil, J., Peltre, G., et al. (1999).Modification of bovine beta-lactoglobulin by glycation in a powdered state or inan aqueous solution: Immunochemical characterization. Journal of Agriculturaland Food Chemistry, 47, 4543–4548.

Oboroceanu, D., Wang, L. Z., Brodkorb, A., Magner, E., & Auty, M. A. E. (2010).Characterization of beta-lactoglobulin fibrillar assembly using atomic forcemicroscopy, polyacrylamide gel electrophoresis, and in situ fourier transforminfrared spectroscopy. Journal of Agricultural and Food Chemistry, 58, 3667–3673.

Panyam, D., & Kilara, A. (1996). Enhancing the functionality of food proteins byenzymatic modification. Trends in Food Science & Technology, 7, 120–125.

Patel, H. A., Singh, H., Havea, P., Considine, T., & Creamer, L. K. (2005). Pressure-induced unfolding and aggregation of the proteins in whey protein concentratesolutions. Journal of Agricultural and Food Chemistry, 53, 9590–9601.

Povey, J. F., Perez-Moral, N., Noel, T. R., Parker, R., Howard, M. J., & Smales, C. M.(2009). Investigating variables and mechanisms that influence protein integrityin low water content amorphous carbohydrate matrices. Biotechnology Progress,25, 1217–1227.

Rabiey, L., & Britten, M. (2009). Effect of whey protein enzymatic hydrolysis on therheological properties of acid-induced gels. Food Hydrocolloids, 23, 2302–2308.

Schmitt, C., Bovay, C., Rouvet, M., Shojaei-Rami, S., & Kolodziejczyk, E. (2007). Wheyprotein soluble aggregates from heating with NaCl: Physicochemical,interfacial, and foaming properties. Langmuir, 23, 4155–4166.

Schokker, E. P., Singh, H., Pinder, D. N., & Creamer, L. K. (2000). Heat-inducedaggregation of beta-lactoglobulin AB at pH 2.5 as influenced by ionic strengthand protein concentration. International Dairy Journal, 10, 233–240.

Surroca, Y., Haverkamp, J., & Heck, A. J. R. (2002). Towards the understanding ofmolecular mechanisms in the early stages of heat-induced aggregation of beta-lactoglobulin AB. Journal of Chromatography. A, 970, 275–285.

Swaisgood, H. E. (2005). The importance of disulfide bridging. BiotechnologyAdvances, 23, 71–73.

Verheul, M., Roefs, S. P. F. M., & de Kruif, K. G. (1998). Kinetics of heat-inducedaggregation of beta-lactoglobulin. Journal of Agricultural and Food Chemistry, 46,896–903.

Watanabe, K., Nakamura, Y., Xu, J. Q., & Shimoyamada, M. (2000). Inhibition againstheat coagulation of ovotransferrin by ovalbumin dry-heated at 120 degrees C.Journal of Agricultural and Food Chemistry, 48, 3965–3972.

Watanabe, K., Xu, J. Q., & Shimoyamada, M. (1999). Inhibiting effects of egg whitedry-heated at 120 degrees C on heat aggregation and coagulation of egg whiteand characteristics of dry-heated egg white. Journal of Agricultural and FoodChemistry, 47, 4083–4088.

Zhou, P., & Labuza, T. P. (2007). Effect of water content on glass transition andprotein aggregation of whey protein powders during short-term storage. FoodBiophysics, 2, 108–116.