25

76

Research ArticleReceived: 9 November 2010 Revised: 10 April 2011 Accepted: 18 April 2011 Published online in Wiley Online Library: 8 June 2011

(wileyonlinelibrary.com) DOI 10.1002/jsfa.4478

Ageing-induced solubility loss in milk proteinconcentrate powder: effect of proteinconformational modifications and interactionswith waterEnamul Haque,a Bhesh R Bhandari,a∗ Michael J Gidley,b Hilton C Deetha

and Andrew K Whittakerc

Abstract

BACKGROUND: Protein conformational modifications and water-protein interactions are two major factors believed to induceinstability of protein and eventually affect the solubility of milk protein concentrate (MPC) powder. To test these hypotheses,MPC was stored at different water activities (aw 0.0–0.85) and temperatures (25 and 45 ◦C) for up to 12 weeks. Sampleswere examined periodically to determine solubility, change in protein conformation by Fourier transform infrared (FTIR)spectroscopy and water status (interaction of water with the protein molecule/surface) by measuring the transverse relaxationtime (T2) with proton nuclear magnetic resonance (1H NMR).

RESULTS: The solubility of MPC decreased significantly with ageing and this process was enhanced by increasing water activity(aw) and temperature. Minor changes in protein secondary structure were observed with FTIR which indicated some degree ofunfolding of protein molecules. The NMR T2 results indicated the presence of three distinct populations of water molecules andthe proton signal intensity and T2 values of proton fractions varied with storage condition (humidity) and ageing.

CONCLUSION: Results suggest that protein/protein interactions may be initiated by unfolding of protein molecules thateventually affects solubility.c© 2011 Society of Chemical Industry

Keywords: principal component analysis; proton NMR; milk proteins; water activity; solubility

INTRODUCTIONMilk protein concentrate (MPC) powder is an important dairycommodity produced from skim milk by ultrafiltration/diafiltrationfollowed by spray drying.1 It contains varying amounts of proteinup to 85%. The use of MPC as a food ingredient is increasingworldwide due to its favourable functional properties and highnutritional value. However, MPC gradually loses solubility uponstorage.1 – 3 Generally it is believed that the loss of solubility islinked with conformational modification of protein moleculesduring processing and storage.4 Proteins are stable in their low-energy, native (folded) state; unfolding or denaturation exposeshydrophobic regions of proteins with greater adoption of β-sheet structures, which promote aggregation via protein–proteininteractions, leading to destabilization.5 It is possible to determineunfolding (denaturation) of protein by monitoring changes in theβ-sheet structure. Fourier transform infrared (FTIR) spectroscopy isa well-established method for probing the secondary structure ofproteins.6,7 The amide I (1600–1700 cm−1) band of polypeptidesand proteins is sensitive to changes in secondary structure andanalysis of this band is used to investigate these changes.7,8 In thisstudy, we used FTIR techniques4,9 to probe protein conformationalchanges upon storage of MPC.

Water plays an important role in the physical or texturalcharacteristics as well as the chemical stability of food systems.10

Nuclear magnetic resonance (NMR) offers a variety of methodsfor characterization of protein systems that can be used toinvestigate the nature of water–protein interactions and physicalstates of water over a wide range of timescales.11,12 In thedairy protein field, the technique of NMR relaxometry has beenapplied to the study of rehydration of casein, caseinates andmilk powders,13,14 monitoring hydration and solubility of β-lactoglobulin.15 The relaxation times of water protons are sensitive

∗ Correspondence to: Bhesh R Bhandari, School of Agriculture and Food Sciences,University of Queensland, Brisbane, QLD 4072 Australia.E-mail: b.bhandari@uq.edu.au

a School of Agriculture and Food Sciences, University of Queensland, Brisbane,QLD 4072, Australia

b Centre for Nutrition and Food Sciences, University of Queensland, Brisbane,QLD 4072, Australia

c Centre for Magnetic Resonance and Australian Institute for Bioengineeringand Nanotechnology (AIBN), University of Queensland, Brisbane, QLD4072, Australia

J Sci Food Agric 2011; 91: 2576–2581 www.soci.org c© 2011 Society of Chemical Industry

25

77

Ageing-induced solubility loss in milk protein concentrate powder www.soci.org

Table 1. Saturated salt solutions used to equilibrate MPC85 atdifferent values of relative humidity

Saturated salt % Equilibrium relative humidity/(aw)

Phosphorus pentoxide 0%/(0.00)

Potassium acetate 22%/(0.22)

Potassium carbonate 43%/(0.43)

Potassium iodide 69%/(0.69)

Sodium chloride 75%/(0.75)

Potassium chloride 84%/(0.84)

indicators of the local chemical and physical environment inheterogeneous materials. This is a result of the dependenceof both the longitudinal (T1) and particularly the transverse(T2) relaxation times on the strength of residual dipole–dipolecouplings with near-neighbour protons. These couplings aremoderated by molecular motion and exchange with protonson protein molecules. Thus it is often possible to identify watermolecules in a range of distinct environments from the distributionof (T2) relaxation times of the water protons. The aim of this studywas to investigate the changes in water–protein interactions andto relate this to protein conformational modifications and changesin solubility upon ageing of MPC.

MATERIALS AND METHODSMilk protein concentrate powder was supplied by MurrayGoulburn Co-operative Co. Ltd (Brunswick, Victoria, Australia).Analysis of the powder was carried out by the manufacturer usingthe method described in Australian standard 2300 (StandardsAustralia 1995). This showed the composition on a dry weightbasis was proteins (82.4% w/w), lactose (4% w/w), fat (1.6% w/w),and ash (7.3% w/w). The moisture content (dry basis) of MPCpowder was determined by AOAC Official Method 927.05 (AOACInternational 2000) and found as 5.5% w/w in our laboratory. MPCpowder samples were kept at−4 ◦C (as powder solubility remainedunchanged for at least 1 year at or below this temperature)prior to use for experimentation. MPC powder samples wereequilibrated to relative humidity range 20–85% by keeping themin vacuum-sealed desiccators containing saturated salt solutions(Table 1). Samples were kept in this condition at room temperature(24 ± 1 ◦C) for the experimentation period. In the case of high-temperature storage studies, samples were equilibrated to wateractivity (aw) = 0 for one week at 24 ◦C prior to storing them invacuum-sealed desiccators at equilibrium relative humidity (ERH)0% and 74% in an incubator at 45 ± 1 ◦C for up to 5 weeks.Solubility was examined after 3, 4 and 5 weeks of storage. aw ofthe samples were measured using an AquaLab 3TE water activitymeter (Decagon Devices Inc., Pullman, WA, USA).

Solubility testsSolubility of MPC85 was determined using a modification ofthe procedure described by Anema et al.3 Duplicate 5% MPCsolutions were prepared from stored samples using MilliQ water.The solutions were continuously stirred using a mechanical stirrer(rotor speed 400 rpm) at room temperature for 30 min. Samplesolutions (40 mL) were transferred to 50 mL centrifugation tubesand centrifuged at 1000 × g for 10 min at 20 ◦C. The supernatantwas then filtered under vacuum (GF/A microfiber filter paper,

1.6 µm pore size, Whatman) and duplicate of 5 g aliquots of thefiltrate weighed into pre-weighed aluminium dishes containing∼20 g acid-washed sand and dried at 105 ◦C for 24 h. The sampledishes were cooled to room temperature in a desiccator containingdry silica gel prior to weighing. The solubility of the MPC samplewas calculated using the following equation:

%Solubility = Solids in the supernatant

Solids in the solution× 100

FTIR analysisStorage-induced protein conformational changes were measuredusing an FTIR spectrometer (Spectrum 100, PerkinElmer, Waltham,MA, USA) with an attenuated total reflection (ATR) cell (UniversalATR). About 2–3 mg MPC powder (stored under different relativehumidity and temperature conditions) was placed on the ATRcrystal and the absorbance intensity was measured at 4 cm−1

resolution in the wavenumber range from 4000 to 400 cm−1. AllFTIR analyses were carried out in triplicate. FTIR was conducted at25 ◦C temperature.

Data processingSpectra were baseline corrected and normalized, and secondderivatives of the amide I and II (1700–1500 cm−1) bands weredetermined to indicate the position of individual component peakswithin the amide envelope. The amide I region (1700–1600 cm−1)of the spectra was deconvolved using PeakFit 4.11, SeaSolveSoftware Inc., San Jose, CA, USA. The deconvoluted spectrumwas fitted with Gaussian band shapes by an iterative curve-fitting procedure (second-derivative deconvolution, baselinelinearization, automatic smoothing). The resultant peaks wereassigned to different protein secondary structures (α-helical, β-sheet and turn). Peak assignment of deconvolved amide I bands(1700–1600 cm−1) (Fig. 1) was done using the results of Farrellet al.16 and Byler and Susi17 as a primary guide. Peaks between1640 and 1620 cm−1 are assigned as β-sheet, 1650 ± 2 cm−1 asα-helix and between 1700 and 1660 cm−1 as β-turns. The peak at∼1648 cm−1 was considered to represent a loop or helix and thepeak at ∼1658 cm−1 a large loop and assigned as a α-helix.

Statistical significance of the change in secondary structuralcomponents between different sample groups, i.e. aged and non-aged, stored at high and low aw (three replicates for each sample)was determined by calculating one-way analysis of variance usingthe software Minitab 15 (Minitab Inc., State College, PA, USA).

NMR analysisA spectrometer (Bruker MSL-300) with resonance frequency300.13 MHz was used for proton transverse relaxation time T2

measurement. T2 relaxation curves were obtained using the single-point Carr–Purcell–Meilboom–Gill (CPMG) pulse sequence.18 T2

relaxation times were measured using CPMG echo trains with pulsespacing TE ranging from 10 to 100 ms and with a maximum of512 pulses. The T2 distributions were obtained from the decayof the signal intensity plotted against time using the opensource software CONTIN, which is based on a Laplace transformalgorithm.19 The NMR experiments were carried out at 25 ◦C.

RESULTS AND DISCUSSIONSolubilityThe initial solubility of fresh MPC powder was 70 ± 2.7% andmoisture content (mc) on a dry weight basis was 5.4%. Solubility

J Sci Food Agric 2011; 91: 2576–2581 c© 2011 Society of Chemical Industry wileyonlinelibrary.com/jsfa

25

78

www.soci.org E Haque et al.

600110016002100260031003600

Wave number cm-1

Amide IAmide II

(a)

(b)

Figure 1. Representative (a) FTIR whole spectra and (b) deconvolution ofamide I band of unequlibrated MPC powder.

declined (Table 2) even during storage at room temperature foronly 4 weeks (65% for samples stored at ERH 23%, mc 5.5% and30% for sample stored at ERH 85%, mc 17%). After 11 weeksof storage, solubility was negligible for samples stored at highERH (>50%). It should be noted that the test used measuressolubility after a fixed time of 30 min; increased solubility occursafter longer times of stirring in water.2 In the case of storage athigh temperature (45 ◦C) and high humidity conditions (ERH =74%), the solubility decreased rapidly from around 70% initiallyto 24.2 ± 0.3% and 14.3 ± 2.9% for samples stored for 3 weeksat ERH 0% and 74%, respectively. After 4 weeks of storage, thesolubility was negligible (Table 2). This observation is consistentwith previous work on the effect of temperature and ageing timeon the solubility of high-protein powder.3 The study by Anemaet al. indicated that MPC solubility decreased with increasingageing time and temperature and suggested that casein was theinsoluble fraction of MPC powder and cross-linking of the proteinat the surface could be the cause of insolubility indicated bya gel electrophoresis study.3 Havea et al. showed that insolublematerial obtained during MPC rehydration consisted of largeparticles (∼100 µm) where the casein micelles were fused togetherby some form of protein–protein interactions predominantlyhydrophobic in nature.1 In a recent study, Mimouni et al. showedthat the low solubility of MPC powder (the same one as used inthe present study) is a consequence of slow dissolution kineticsrather than presence of a large amount of insoluble matter,at least for the samples of less than 6 months old. They alsoshowed that the rehydration process consists of two overlappingsteps, namely disruption of agglomerated particles and releaseof material from the powder particles to the surrounding phase.Both occur simultaneously and the latter process is suggested tobe the rate-limiting step.2 These studies, along with our currentresults, suggest that ageing induces physicochemical changes in

Table 2. Solubility of MPC 85 powder after 4 and 11 weeks of storageat room temperature at different aw values and storage at 45 ◦C and ataw 0.0 and 0.75 for 3–4 weeks

aw

Storage temp.(◦C)

Solubility %(3 weeks)

Solubility %(4 weeks)

Solubility %(11 weeks)

0.23 24 ± 1 65 ± 0.8 35 ± 14

0.45 24 ± 1 48 ± 2.2 16.6 ± 0.2

0.69 24 ± 1 42 ± 0.04 10.8 ± 2.7

0.85 24 ± 1 29.6 ± 0.01 7.5 ± 10.6

0.00 45 ± 1 24.2 ± 0.33 14.1 ± 7.3

0.74 45 ± 1 14.3 ± 2.9 6.82 ± 0.2

15

20

25

30

35

40

45

50

55

0 0.2 0.4 0.6 0.8 1

Water activity, aw

% R

elat

ive

pro

po

rtio

n

α-helix

β-sheet

β-turn

Figure 2. Relative fractions of different structural components after roomtemperature (24 ± 1 ◦C) storage of MPC 85 for 3 and 10 weeks at differentaw. Solid and dashed lines represent corresponding structural componentsafter 3 and 10 weeks.

MPC powder particles/molecules which lead to a rapid declinein solubility. High temperature and aw greatly enhance thisprocess.

Protein conformational modificationsThe FTIR whole spectra and deconvolution of amide I band ofMPC powder is shown in Fig. 1. FTIR analysis of the amide Iband showed minor changes in secondary structural componentswith time of storage and/or water activity conditions (Fig. 2) atmoderate (24 ± 1 ◦C) temperature storage. The relative fraction ofα-helix was slightly higher in samples stored for 3 weeks than thatof samples stored for 10 weeks for all levels of water activity (Fig. 2).Consequently, the relative fractions of β-sheet components werea little higher after 10 weeks than after 3 weeks for all levelsof water activity (Fig. 2). The content of β-turn showed a slightincrease after 10 weeks compared to that after 3 weeks for loweraw samples but decreased for higher aw (>0.23) samples. Thusthere was a decrease in helical components and a corresponding

wileyonlinelibrary.com/jsfa c© 2011 Society of Chemical Industry J Sci Food Agric 2011; 91: 2576–2581

25

79

Ageing-induced solubility loss in milk protein concentrate powder www.soci.org

increase in β-components (specifically β-sheet) with ageing ofMPC. These transformations of helical and loop components(folded structure) to sheet and turn components (unfolding)indicate changes in the secondary structural conformation ofthe proteins during storage. However, these changes in structuralcomponents at different water activities were minor (±1–9%).Overall, the effect of storage temperature and aw on the relativefractions of different secondary structural components covereda restricted range (α-helix 26–36%, β-sheet 42–49% and β-turn20–29%) (Fig. 2) consistent with relatively minor average changesin protein conformation. Statistical analysis did not show anysignificant difference between different sample groups. In allcases P-values greater than 0.05 were obtained which indicatethat the limited change in protein conformation was statisticallynot significant.

The solubility of MPC powders declined rapidly with time ofstorage and increasing water activity. High temperature enhancedthis process (Table 2). It is generally believed that unfolding of thesecondary structure of a protein leads to alteration of its functionalproperties such as solubility. It was expected that such a changewould be detected in the FTIR analysis but this was not apparentin these results. The reason for this could be that the proteinconformational change is partial (or not extensive), but that aminor average change could lead to a major change in solubilityby influencing the protein–protein interactions, which seems tobe responsible for the slow dissolution kinetics of MPC.20 This isfurther discussed below.

Water–protein interactionsIn all cases, the T2 (proton transverse relaxation time, millisecondtimescale) distribution profile could be fitted by three populationsof relaxation times (Fig. 3(a)). In reality, this is likely to be anapproximation of a broader distribution of exponential functions,but provides a minimum number of relaxation time ranges thatcan be interpreted in terms of water–protein interactions. Theshort T2 component (∼0.01–2 ms) is assigned to water protons inclose proximity to the protein surface that exchange rapidly andmay have low mobility. The longest T2 component (∼5–28 ms) isassociated with the fraction of water that is furthest from a proteinsurface and which may have higher mobility. The intermediateT2 component (∼1.0–7.4 ms) relates to an intermediate form ofwater that is moderately mobile. Generally, peak widths indicateuniformity of the samples, with broader peaks corresponding toless uniform samples.21,22 The peak of the short T2 componentwas narrow (Fig. 3(a)), indicating that this protein-associatedwater fraction was relatively uniform. Peak width (distributionof relaxation time) became progressively wider for the other twowater components (Fig. 3(a)), suggesting that there was greatervariation in the physical environments of the protons and thatthese fractions were less uniform. After 3 weeks of storage at roomtemperature (24±1 ◦C) and at different aw there was a decrease inT2 with increasing water content up to aw 0.45 (mc 8.2%), and thena further increase upon increase of aw (mc) was observed (Fig. 3(b)).The overall T2 values increased with added water, as expected foran increased average distance to protein surfaces. This indicatesthat the exchange (of water) became more important as water isadded to the low-moisture-content samples. The proton fraction(calculated by integrating the peak area for each proton pool) withshort T2 values increased progressively with increasing MC andconsequently proton fractions of intermediate- and long-T2 poolsdecreased (Fig. 4). This indicates increased interaction of waterwith protein sites, resulting in reduced mobilization of water.

0 5 10 15 20 25 30

T2 (ms)

Inte

nsi

ty

(a)

0

1

2

3

4

5

0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6

T2 (ms)

Inte

nsi

ty (

a.u

.)

aw 0.0aw 0.23aw 0.45aw 0.69aw 0.85

(b)

Short

Intermediate

Long

Figure 3. (a) Example of distribution of proton transverse relaxation timesof MPC powder and (b) T2 of short proton pool of MPC samplesstored at room temperature (24 ± 1 ◦C) and at different aw for3 weeks.

0

20

40

60

80

100

0 0.23 0.45 0.69 0.85aw

Fra

ctio

ns

(%)

short

intermediate

long

Figure 4. T2 proton fractions of MPC stored at room temperature(24 ± 1 ◦C) for (a) 3 weeks at different aw. Proton fractions weredetermined by calculating the corresponding peak area from linear T2distribution plot.

J Sci Food Agric 2011; 91: 2576–2581 c© 2011 Society of Chemical Industry wileyonlinelibrary.com/jsfa

25

80

www.soci.org E Haque et al.

After 11 weeks of storage at room temperature, the T2

distribution pattern was similar to that after 3 weeks. That is,a decrease in short T2 with increase of aw up to aw 0.45 andsubsequent increase with further increase of aw was observed(figure not included). The difference in short T2 with change in aw

was less pronounced than after 3 weeks of storage. This change inshort T2 may suggest that proton spins close to the protein surfacewere experiencing a similar physical environment, indicating asimilar average arrangement of protein molecules at the differentstorage times. Another noticeable change was that the intensityof the signals from intermediate- and long-T2 fractions decreasedmore for samples stored at high aw (≥0.45, mc 8.2%) than forsamples stored at lower aw (figure not included). On the otherhand, for samples kept at lower aw (<0.45), proton fractions withshort T2 values decreased and those of intermediate and longvalues increased after 11 weeks storage compared to that after3 weeks of storage. This suggests that at higher aw (mc) watermolecules from more mobile phases (intermediate and long pool)migrated to close to the protein surface and became less mobile,with the distinction between short and intermediate T2 fractionsdiminishing. However, at lower aw, water molecules of the short-T2

pool seem to move to outer (intermediate and long) fractions.Recent studies on MPC solubility by scanning electron mi-

croscopy (SEM) suggest that during storage increased interactionbetween and within casein micelles results in compaction of mi-celles and formation of a closely packed monolayer skin of caseinmicelles that seem to be responsible for the slow dissolution kinet-ics of MPC.20 Light scattering studies2 show that surface wettabilityis not rate determining, but that casein micelle solubilization is.Further evidence for the aggregation of casein micelles as thecause of the reduced rate of solubilization comes from studiesof the milk components solubilized during rehydration.23 Thisshowed that, apart from milk components associated with caseinmicelles (casein, calcium, etc.), all components were solubilizedrapidly, e.g. whey proteins, lactose, sodium. Our studies showed agradual decline of solubility with ageing under more adverse con-ditions such as high ERH (or aw) and temperature. Also studies onprotein conformational change indicate some limited unfolding ofprotein (transformation of helical components to β-components)upon ageing. This protein aggregation was aggravated by adversestorage conditions such as high water activity and had a nega-tive impact on solubility. The water–protein interaction behaviourobserved by proton transverse relaxation time suggests that asproteins unfold water molecules can interact with a greater pro-portion of the surface of the protein. The traditional view is thatcharged groups on proteins alter the normal structure of the waterin their vicinity, and water becomes oriented in layers around thesegroups. As aw is increased (0.2–0.7), the nonionic groups and theaccessible peptide bonds are thought to become progressivelysaturated.24 Based on our observations, it could be suggested thatlimited unfolding of protein (specifically casein) molecules maybe related to protein aggregation and may initiate casein–caseininteraction within a micelle. Adverse storage conditions aggravatethis transformation process. It is unlikely that protein unfolding issolely responsible for the whole transformation process but it maybe the initiation step. The fact that only a limited change in proteinconformation was observed is consistent with the limited nature ofthe overall micelle aggregation, as evidenced by the kinetic natureof the solubilization change (i.e. after sufficient rehydration time,even long-term stored samples are fully solubilized) and the likelylocalization of micelle aggregation effects to the surface of pow-der particles. Although we propose that protein conformational

changes are associated with the reduced rate of solubility of storedpowders, more studies are required to understand the molecularmechanism of subsequent steps.

CONCLUSIONIntegrating the results obtained from characterization of MPCduring storage using FTIR for protein conformational modificationsand NMR relaxometry for water–protein interactions suggeststhat there is a link between minor protein unfolding/refolding andgradual loss of solubility upon ageing of MPC. Storage at higherwater activity resulted in more water molecules in close proximityto the protein surface and also slightly higher protein unfolding,which could eventually lead to loss of solubility.

REFERENCES1 Havea P, Protein interactions in milk protein concentrate powders. Int

Dairy J 16:415–422 (2006).2 Mimouni A, Deeth HC, Whittaker AK, Gidley MJ and Bhandari BR,

Rehydration process of milk protein concentrate powder monitoredby static light scattering. Food Hydrocolloids 23:1958–1965 (2009).

3 Anema SG, Pinder DN, Hunter RJ and Hemar Y, Effects of storagetemperature on the solubility of milk protein concentrate (MPC85).Food Hydrocolloids 20:386–393 (2006).

4 Kher A, Udabage P, McKinnon I, McNaughton D and Augustin MA,FTIR investigation of spray-dried milk protein concentrate powders.Vibrat Spectrosc 44:375–381 (2007).

5 Arrondo JLR and Goni FM, Structure and dynamics of membraneproteins as studied by infrared spectroscopy. Prog Biophys Mol Biol72:367–405 (1999).

6 Kumosinski TF and Farrell HM, Determination of the global secondarystructure of proteins by Fourier-Transform Infrared (FTIR) spec-troscopy. Trends Food Sci Technol 4:169–175 (1993).

7 Barth A, Infrared spectroscopy of proteins. Biochim Biophys Acta1767:1073–1101 (2007).

8 Kumosinski TF and Unruh JJ, Quantitation of the global secondarystructure of globular proteins by FTIR spectroscopy: comparisonwith X-ray crystallographic structure. Talanta 43:199–219 (1996).

9 Oliver CM, Kher A, McNaughton D and Augustin MA, Use of FTIR andmass spectrometry for characterization of glycated caseins. J DairyRes 76:105–110 (2009).

10 Bell LN and Labuza TP, Moisture Sorption: Practical Aspects of IsothermMeasurement and Use (2nd edn). American Association of CerealChemists, St Paul, MN (2000).

11 Krushelnitsky A and Reichert D, Solid-state NMR and protein dynamics.Prog Nucl Magn Reson Spectrosc 47:1–25 (2005).

12 Belloque J and Ramos M, Application of NMR spectroscopy to milkand dairy products. Trends Food Sci Technol 10:313–320 (1999).

13 Davenel A, Schuck P, Mariette F and Brule G, NMR relaxometry as anon-invasive tool to characterize milk powders. Lait 82:465–473(2002).

14 Schuck P, Davenel A, Mariette F, Briard V, Mejean S and Piot M,Rehydration of casein powders: effects of added mineral salts andsalt addition methods on water transfer. Int Dairy J 12:51–57 (2002).

15 Granizo DP, Reuhs BL, Stroshine R and Mauer LJ, Evaluating thesolubility of powdered food ingredients using dynamic nuclearmagnetic resonance (NMR) relaxometry. LWT Food Sci Technol40:36–42 (2007).

16 Farrell HM, Wickham ED, Unruh JJ, Qi PX and Hoagland PD, Secondarystructural studies of bovine caseins: temperature dependence ofbeta-casein structure as analyzed by circular dichroism and FTIRspectroscopy and correlation with micellization. Food Hydrocolloids15:341–354 (2001).

17 Byler DM and Susi H, Examination of the secondary structure ofproteins by deconvolved FTIR spectra. Biopolymers 25:469–487(1986).

18 Meiboom S and Gill D, Modified spin-echo method for measuringnuclear relaxation times. Rev Sci Instrum 29:688–691 (1958).

19 Provencher SW, CONTIN: a general purpose constrained regularizationprogram for inverting noisy linear algebraic and integral equations.Comput Phys Commun 27:229–242 (1982).

wileyonlinelibrary.com/jsfa c© 2011 Society of Chemical Industry J Sci Food Agric 2011; 91: 2576–2581

25

81

Ageing-induced solubility loss in milk protein concentrate powder www.soci.org

20 Mimouni A, Deeth HC, Whittaker AK, Gidley MJ and Bhandari BR,Investigation of the microstructure of milk protein concentratepowders during rehydration: alterations during storage. J Dairy Sci93:463–472 (2010).

21 Noronha N, Duggan E, Ziegler GR, O’Riordan ED and O’Sullivan M,Investigation of imitation cheese matrix development using lightmicroscopy and NMR relaxometry. Int Dairy J 18:641–648 (2008).

22 Kuo MI, Gunasekaran S, Johnson M and Chen C, Nuclear magneticresonance study of water mobility in pasta filata and non-pastafilata Mozzarella. J Dairy Sci 84:1950–1958 (2001).

23 Mimouni A, Deeth HC, Whittaker AK, Gidley MJ and Bhandari BR,Rehydration of high-protein-containing dairy powder: slow- andfast-dissolving components and storage effects. Dairy Sci Technol90:335–344 (2010).

24 Kinsella JE and Fox PF, Water sorption by proteins: milk and wheyproteins. CRC Crit Rev Food Sci Nutr 24:91–139 (1986).

J Sci Food Agric 2011; 91: 2576–2581 c© 2011 Society of Chemical Industry wileyonlinelibrary.com/jsfa