HAL Id: hal-00929602https://hal.archives-ouvertes.fr/hal-00929602

Submitted on 1 Jan 1998

HAL is a multi-disciplinary open accessarchive for the deposit and dissemination of sci-entific research documents, whether they are pub-lished or not. The documents may come fromteaching and research institutions in France orabroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire HAL, estdestinée au dépôt et à la diffusion de documentsscientifiques de niveau recherche, publiés ou non,émanant des établissements d’enseignement et derecherche français ou étrangers, des laboratoirespublics ou privés.

Heat stability of reconstituted casein micelle dispersions:changes induced by salt addition

Christelle Le Ray, Jean-Louis Maubois, Frédéric Gaucheron, Gérard Brulé,Paul Pronnier, Fabienne Garnier

To cite this version:Christelle Le Ray, Jean-Louis Maubois, Frédéric Gaucheron, Gérard Brulé, Paul Pronnier, et al.. Heatstability of reconstituted casein micelle dispersions: changes induced by salt addition. Le Lait, INRAEditions, 1998, 78 (4), pp.375-390. �hal-00929602�

Lait (1998) 78, 375-390© Inra/Elsevier, Paris

375

Original article

Heat stability of reconstituted casein micelledispersions: changes induced by salt addition

Christelle Le Raya, Jean-Louis Maubois?", Frédéric Gaucheron",Gérard Brulé", Paul Pronnier", Fabienne Garnier"

a Crealis, ZI Le Teinchurier, rue F-Sauvage, 19100 Brive-la-Gaillarde, Franceb Laboratoire de recherches de technologie laitière, Inra, 65, rue de Saint-Brieuc,

35042 Rennes cedex, France

(Received 6 May 1997; accepted II December 1997)

Abstract - The aim of the present work was to study the heat stability of reconstituted casein micelledispersions (RCMD) after addition of various salt solution: NaCI, CaClz, MgClz, sodium citrate andsodium phosphate. Case in micelle stability to heat treatments (95 "C - 10,20 and 30 min) was eva-luated. Modifications of the pH, of the casein micelle pellet water content and of the minerai and caseindistribution between aqueous and micellar phase were measured. NaCI addition decreased the RCMDpH, increased the amount of diffusible calcium and decreased the amount of supematant casein.CaClz and MgClz addition decreased the RCMD pH and the amount of supematant casein. Sodiumcitrate addition, with a solution at pH 7.4, increased the RCMD pH, drastically increased the amountof diffusible calcium, phosphorus and of supematant casein. It caused the case in micelle's destruc-turation. Sodium phosphate addition, with a solution at pH 7.4, did not modify the RCMD pH andincreased the amount of supernatant casein. Acidification, beyond a pH value of 7.0, induced heataggregation. Calcium and magnesium chloride addition was detrimental to casein micelle heat sta-bility. NaCI, sodium citrate and sodium phosphate addition enhanced reconstituted casein micelle heatstability. The determinant role of the aqueous phase on reconstituted casein micelle physico-che-mical properties was emphasized by this study. © InraJElsevier, Paris

casein micelle / salt / heat stability

Résumé - Stabilité thermique de dispersions de micelles de caséine reconstituées: modificationsproduites par l'addition de sels minéraux. L'objectif du présent travail était d'étudier la stabilitéthermique de dispersions de micelles de caséine reconstituées après l'addition de différentes solutionssalines: NaCI, CaClz, MgClz, citrate de sodium et phosphate de sodium. La stabilité au traitement ther-mique (95 "C, 10-20-30 min) des dispersions des micelles de caséine reconstituées était évaluée. Lesmodifications du pH, de la teneur en eau des culots des dispersions de micelles de caséine et de larépartition des minéraux entre phase aqueuse et micellaire étaient mesurées. L'addition de NaCIdiminuait le pH des dispersions, augmentait la teneur en calcium diffusible et diminuait la teneur en

* Correspondence and reprints. E-mail: maubois@labtechno.roazhon.inra.fr

376 C. Le Rayet al.

caséine des sumageants d'ultracentrifugation. L'addition de CaClz et de MgClz diminuait le pH desdispersions et la teneur en caséine des sumageants d'ultracentrifugation. L'ajout de citrate de sodium,à l'aide d'une solution à pH 7,4, augmentait le pH des dispersions, augmentait fortement les teneursen calcium et phosphore diffusibles et de façon drastique la teneur en caséine des sumageants d'ul-tracentrifugation. Cet ajout entraînait une déstructuration de la micelle. L'addition de phosphate desodium, à l'aide d'une solution à pH 7,4, ne modifiait pas le pH des dispersions et la teneur en caséinedes surnageants d'ultracentrifugation. Une acidification à des valeurs de pH inférieures à 7,0 entraî-nait une agrégation après traitement thermique. L'addition de chlorure de calcium et de magnésiumdiminuait la stabilité thermique des micelles reconstituées. L'addition de NaCI, de citrate de sodiumet de phosphate de sodium améliorait la stabilité thermique des micelles de caséine reconstituées. Danscette étude, nous avons souligné le rôle déterminant de la phase aqueuse sur les propriétés physico-chimiques des dispersions de micelles de caséine reconstituées. © InralElsevier, Paris

micelle de caséine / minéraux / stabilité thermique

1. INTRODUCTION

Incidence of various heat treatments onmilk has been extensively studied. Depen-ding on the heat intensity, various modifi-cations in phosphocalcic equilibria, in wheyprotein solubility as weIl as formation ofcomplexes between x-casein and ~-lacto-globulin, between s-amino groups of pro-teins and lactose and degradation of lactoseoccur [1]. Sorne ofthese changes are rever-sible; others are irreversible and lead to adegradation of the nutritional value and ofthe technological properties of milk.

Most of the published studies on theeffect of heating on casein micellar stabi-litY have been carried out with normal orevaporated milks [11,35,37,46], i.e., withcasein micelles dispersed in a very complexsolution. Sorne works have used solutions ofartificially formed casein micelles as modelsfor studying the effect of individual caseinratio on heat stability [21, 32, 33]. To ourknowledge, no complete study used disper-sions of micellar casein without interactionwith the aqueous phase during heat treat-ment, i.e., in the absence of lactose and wheyproteins.

Recent developments in membranemicrofiltration processing of milk have al-lowed the preparation of micellar phospho-caseinate with no effect on the casein micelle

structure [29, 30]. Micellar phosphocasei-nate powder with a high bacteriological qua-lity can be produced in large quantities bymicrofiltration with a 0.1 um pore sizemembrane [10], carried out in a system usingthe uniform transmembrane concept deve-loped by Sandblom [31], of skim milk pre-viously treated by the Bactocatch processand submitted to diafiltration before spray-drying [34]. This product has been shownto exhibit a micellar-like behaviour in termsof rennet coagulation before and after a100 °C-5 min heat treatment [29], of ~-caseindissociation [30] and of acid gelation [9].

Milk and its proteinaceous derivativesare commonly used in formulated foodscontaining various mineraI salts and otheringredients. Physico-chernical properties ofmilk proteins, as weIl as their heat stability,are generally not predictable because of thecomplexity of the possible interactions be-tween the added proteins and the food com-ponents. Therefore, it appears necessary tostudy the behaviour of model protein solu-tions in progressively more complex sys-tems submitted to heat treatments usuallycarried out by the food industry.

This paper describes the effect of variousenvironments (pH, NaCI, CaCI2, MgCI2,

sodium citrate, sodium phosphate) on phy-sico-chemical properties of reconstitutedcasein micelle dispersion (RCMD) which

Heat stability of casein micelles

is considered to be an interesting model forproteins susceptible to he used per se in rnilkprotein mixture as an ingredient, before andafter three heat treatments at 95 oc (10, 20,and 30 min).

2. MATERIALS AND METHODS

2.1. Preparation of RCMD

The micellar phosphocaseinate powder wasprepared as described by Pierre et al. [29] andSchuck et al. [34]. The chemical composition ofthe micellar phosphocaseinate powder is repor-ted in table J. Powder was reconstituted in ultra-pure water (milliQ) at 50 oc. In one experiment,powder was reconstituted in milk ultrafiltrate at30 "C obtained with a 10 000 Da eut-off mem-brane at 50 oc. The final casein concentrationwas 25 g-L-1 (i.e., 31.1 g-L-lof powder). In ailcases, 0.1 g- L-lof thimerosal (Sigma, SaintLouis, Missouri, USA) was added to the disper-sion to prevent bacterial and fungal growth.

2.2. Acidification of RCMD in water

The RCMO were acidified with 1 mol·L-1HCI to reach the pH values: 6.99, 6.93 and 6.67.Ultrapure water was added to compensate forvolume changes.

2.3. Addition of salts to ReMD in water

The salt solutions used were sodium chloride(NaCI; 4.15 mol-kg:"), calcium chloride (CaCI2;

3.77 mol·kg-I), magnesium chloride (MgCI2;

377

2.12 mol-kg:"), a mixture of disodium and of tri-sodium citrate (fmal pH value: 7.4; 1.25 mol-kg:')and a mixture of sodium dihydrogen orthophos-phate (NaH2P04) and of disodium hydrogenorthophosphate (N~HP04) (final pH value: 7.4;0.45 mol·kg-I). Reagents were of the analyticalgrade. Salt solutions were added to the RCMO inwater, at room temperature. Salt concentrationsadded in final dispersions are reported in table lI.Ouring salt additions, the RCMO were stirredvigorously to ensure rapid and complete mixing.Samples were stirred for one hour at room tem-perature. For sorne experiments carried out at anadjusted pH, HCI 1 N or NaOH 1 N was added toreach the required pH value. pH adjusted sampleswere stirred for one hour at room temperatureand pH was checked and eventually readjustedduring the hour the sam pie was left standing.Ultrapure water was added to compensate forvolume changes induced by salt addition or pHcorrection.

Table J. Chemical composition of the phospho-caseinate powder (g-kg " powder).Tableau J. Composition chimique de la poudrede phosphocaséinate natif (g-kg"" poudre).

Dry matterTotal Protein (N x 6.38)Total Casein (N x 6.38)NCN (N x 6.38)NPN (N x 6.38)LactoseAshes

914.7832.7794.0

38.72.35.4

76.8

NPN: non protein nitrogen. NCN: non casein nitrogen.NPN : azote non protéique. NCN : azote non caséinique.

Table II. Final concentrations of added salts in reconstituted casein micelle-Iike dispersions (RCMO)(mmol-kg"),

Tableau II. Concentrations finales des sels ajoutés dans les dispersions de micelles de caséinesreconstituées (mmol· kg-I).

salt added 1st concentration 2nd concentration 3rd concentration

NaCI 17.4 94.8 177.8CaCI2 10.5 13.7 19MgCI2 2.6 9.3 19.3Sodium citrate 6.0 10.6 17.8Sodium phosphate 1.74 4.19 7.80

378

- -- -- --- --------------------

C. Le Rayet al.

2.4. Heat treatments

Ten g of RCMO in sealed pyrex tubes (exter-nal dimensions: 10 x 1.6 cm; volume = 15 mL)were submerged in an oil-bath, The temperaturewas thermostatically regulated (± 0.1 "C) at therequired assay temperature (95.0 "C). A rockingrate of2 rev-rnirr" was used (RCMO flows fromone end of the tube to the other 4 times-rnirr").The heating times were 10, 20 and 30 min. Thesetimes included a 5-min heating-up period (expe-rimental determination). After heating, RCMOwere immediately cooled to 20 "C and then ana-Iysed. RCMO reconstitutions, heat treatmentsand subsequent analyses were carried out induplicate.

2.5. Physico-chemical analyses

Physico-chemical analyses were carried out at20 oc. Analyses were carried out before and aftersalt additions and heat treatments. Heated andcontrol dispersions were centrifuged at 160g for5 min, in order to remove insoluble aggregates.RCMO samples consisted of these supematants.

pH values were measured with a Portames752 Calimatic pHmeter at 20 "C (Bioblock, 1Il-kirch, France). Total protein contents were deter-mined by Kjeldahl method using a 6.38 con ver-ting factor. The diffusate phases of RCMO wereobtained by ultrafiltration on Centriflo CF 25(eut-off: 25000 Da; Amicon, Epernon, France)after centrifugation at 500 g for 1 h at 20 oc.

Cation concentrations (Na, Ca and Mg) weredetermined by atomic absorption spectrometryas described by Brulé et al. [5] on RCMOsamples and on ultrafiltrates. Phosphorus concen-trations were determined according to the lOFmethod [17] on RCMO samples and on ultra-filtrates. Mineral concentrations in ultrafiltrateswere converted into diffusible mineraI concen-trations in RCMO by multiplying by a 0.96 cor-recting factor as described by Pierre and Brulé[27]; this correction takes into account the exclu-ded volume effect.

The RCMO samples were ultracentrifuged at20 "C for 2 h at 77 000 g in a L8-55 ultracentri-fuge with a 50.2 Ti rotor (Beckman Instrument,Gagny, France). The casein micelle pellet watercontents were determined after a 22-h freezedrying of the RCMO pellet. Each sample wasseparated in 2 tubes. One was used to estimatemicelle pellet water content and the other one toestimate total protein content of the dry pellet.

Casein micelle pellet water contents were expres-sed as g of water per g of total protein.

Supematant protein contents were defined ascasein which did not sediment after a 2-h cen-trifugation at 77 000 g at 20 "C, They were esti-mated by optical density (00) measurements at280 nm (spectrophotometer Beckman, OU 62,Beckman Instrument, Gagny, France) of thesupernatants diluted in 10 mmol-L"! EOTA,pH 10 (adapted from Driehuis and Teemstra [8]).00 of supernatants were calculated by multi-plying the 00 measured by the dilution factor.Reverse-phase high performance Iiquid chro-matography (RP-HPLC) analysis allowed theevaluation of the casein contents in the follo-wing samples: the RCMO samples, a heat-indu-ced precipitate, and ultracentrifugal supematants.Before RP-HPLC analysis, the heat-induced pre-cipitate was redissolved in 7 mol-L"! urea and15 mmol-L:' EOTA. Analyses were performedafter treatment in 10 mmol-L:" OTT (Sigma,Saint-Quentin-en- Yvelines, France) for 1 h at30 oc. The system was composed of a Waters600E multisolvent delivery system, a Waters 486tunable absorbance detector set at 214 nm and aNelson analytical Data System (Perkin Elmer,Saint-Quentin-en-Yvelines, France). The reverse-phase column was a 15 cm Vydac C4, 214 TP54 (Touzart et Matignon, Vitry-sur-Seine,France). The chromatographie conditions werethose of Jaubert and Martin [18]. The percen-tages of solubilized us1- and ~-caseins were esti-mated by ratio of individual casein area in theultracentrifugable supernatants to individualcasein area obtained for the RCMO before ultra-centrifugation.

3. RESULTS AND DISCUSSION

3.1. Characterization of RCMD in water

The pH of RCMD was 7.33. The diffu-sible calcium and phosphorus concentra-tions were about 0.8 and 1.0 mmol-kg ", res-pectively. These results are 10 times lessthan the diffusible calcium and phosphorusconcentrations determined in milk: 10 and12 mmol-L:'. respectively [15]. As shawnby Schuck et al. [34], calcium determina-tian confirms that spray-drying treatmenthad no effect on Ca salt equilibrium sincethe content determined in the microfiltra-tion retentate (data not shawn) and in recons-

Heat stability of casein micelles

tituted RCMD are identical. Increase of pHresulted from the decrease of ionie strength(the diffusible ash content of RCMD was0.2 g-kg") which induces an increase of theapparent pK's of casein carboxyls residuesand the displacement of protons from theaqueous to the micellar phase. This alsoallowed the ionization of groups which pos-sess an apparent pK close to the pH of dis-persion, i.e., His and SerP.

After heat treatment of RCMD, no pHvariation was detected. Such a result isexplained by the lack of milk aqueous phasewhich induces formation of formic acidduring heating of milk with a concomitantrelease of H+ [1].

Casein micelle pellet water content ofRCMD (2.05 g of water per g of total proteinbefore heat treatment) was reduced to 1.78by the treatment of 30 min at 95 oc. Sornedisintegration of casein micelles could haveoccurred during the heating leading to a pre-ferential release of x-casein [36], which isfollowed by a case in micelle reassociationduring cooling [2]. An irreversible change ofcolloidal calcium phosphate (CCP) couldalso occur during heating [28, 45]. The slightmodification of the micelle organization isconfirmed by the observed data (controlsamples in tables III and IV): the applied

379

heat treatment induced an aggregation ofaround 4 % expressed in protein content ofRCMD. The same slight polymerization ofcasein was found by Zin el Din and Aoki[47] when whey protein-free milk withoutlactose was heated at 80 "C for 15 to 75 s.Singh et al. [38] suggested that heat treat-ment of milk appeared to weaken the inter-action forces between casein componentswithin the micelles. In the present study, themicelle pellet water content could have beenreduced by this incomplete restoration ofthe original micellar organization after heattreatment and cooling.

3.2. Acidification of RCMDin the pH range 7.33-6.67

The diffusible calcium concentrations ofthe RCMD adjusted to pH 6.99, 6.93 and6.67 were respectively 1.20, 1.44 and2.23 mrnol-kg ". As expected, the pHdecrease led to calcium solubilization whichis similar to what Famelart et al. [9] obser-ved.

When pH was adjusted to a value lowerthan 7.0, there was an aggregation causedby the applied heat treatment (figure 1). AtpH 6.67, about 40 % of total protein and

Table III. Concentrations of total protein (g-kg") of reconstituted casein micelle-like dispersions(RCMO) as a function of CeCl, concentrations (mmol-kg:") and after heating at 95 "C, 30 min.Before analysis, a centrifugation (160 g, 5 min) was carried out.Tableau III. Concentrations en matière protéique totale (g-kg:") des échantillons de dispersions demicelles de caséines reconstituées en fonction de la concentration en CaCl, (mmol-kg ") et aprèschauffage à 95 "C, 30 min. Les échantillons ont subi une centrifugation (160 g, 5 min) avant analyse.

ceci, (mmol-kg")

heating time (min) 0 10.5 13.7 19.0

0 25.60 25.17 25.16 25.0210 ND 0.85 0.91 0.9320 ND 0.84 0.85 0.8930 24.40 0.84 1.25 0.84

ND: not determined. Total protein = N x 6.38.ND : non déterminé. Matière protéique totale: N x 6.38.

380 C. Le Rayet al.

Table IV. Concentrations of total proteins (g-kg:") of reconstituted casein micelle-like dispersions(RCMO) as a function of MgClz concentrations (rnmol-kg") and after heating at 95 "C, 30 min.Before analysis, a centrifugation (160 g, 5 min) was carried out.Tableau IV. Concentrations en matière protéique totale (g-kg ") des échantillons de dispersions demicelles de caséines reconstituées en fonction de la concentration en MgClz (mmol·kg-1) et après chauf-fage à 95 "C, 30 min. Les échantillons ont subi une centrifugation (160 g, 5 min) avant analyse.

CaClz (rnmol-kg ")

Heating time (min) 0 2.6 9.3

0 25.19 25.04 24.9210 NO 18.88 1.1920 NO 18.25 1.1730 23.89 16.61 1.22

19.3

24.721.31.061.06

ND: not determined. Total protein = N x 6.38.ND : non déterminé. Matière protéique totale: N x 6,38.

7

pHFigure 1. Concentrations of total protein (e) and total calcium (+) in RCMO supematants (160 g,5 min) after heat treatment (95 "C, 30 min) as a function of pH.Figure 1. Concentrations en matière protéique totale (e) et en calcium total (+) dans les surna-geants de centrifugation (160 g, 5 min) de phosphocaséinate natif après traitement thermique (95 "C,30 min) en fonction du pH.

calcium were in the sediment obtained bycentrifugation at 160 g-5 min. RP-HPLCanalysis of both the non-heated RCMDsample and sediment samples showed thatthere was no selective precipitation ofcaseins: the proportion of us1- and ~-caseinswere similar (results not shown). Applicationof the same heat treatment to a similar dis-

persion of casein micel1es in milk ultra-fiItrate at pH 6.70 did not induce aggregation(results not shown). Such a difference in thebehaviour of casein miceIIes can be attri-buted to the higher isoelectric pH of preci-pitation obtained at low ionie strength whencompared to the milk ionie strength. Isoio-nic pH of individual caseins are higher than

Heat stability of casein micelles

isoelectric ones at milk ionie strength(1 = 0.08 mol-L -1). When micelles are dis-persed in a medium, they are subject tocharge neutralization at a higher pH whenthe ionie strength of the medium is less than0.08 mol-L -1. Micellar casein might also beregarded as a globulin which was consideredto be a class of protein insoluble in purewater but soluble in dilute saline solutions

7.5

381

[4]. Acidification through gluconodeltalac-tone addition gave similar results for heatprecipitation (results not shown).

3.3. NaCl addition

NaCl addition (0-177.8 rnmol-kg ") tothe RCMD led to an average pH decrease

~

7.3 r\.~ -t------7.4

I TCo

7.2

7.1

7

4.0

,

'"C~

f5~E (5 3.0:J E'(3 E~j 2.0- 0P-.c~ 15 1.0li:: 0.- .co Co

0.0

3

50 100 150 200,NaCI (mmollkg) 1

Figure 2. Physico-chemical characterization of RCMD samples after NaCI addition before heattreatment. (A) pH; (H) diffusible calcium (+) and phosphorus (e) concentrations; (C) micellar pel-let water content.Figure 2. Caractérisation physicochimique des échantillons de dispersions de micelles de caséinesreconstituées après ajout de NaC1, avant traitement thermique. (A) pH ; (B) concentrations en calcium(+) et phosphore (e) diffusibles; (C) teneur en eau des culots micellaires.

382 C. Le Rayet al.

of about 0.1-0.15 unit (figure 2A). H+release is due to the effect of ionie strengthwhich decreases the activity coefficients ofthe diffusible ions and consequently,increases dissociation of the ion pairs. H+release is also due to the exchange betweenNa" and H+ as suggested by Grufferty andFox [14] and Strange et al. [40] becauseapproximately 5 % of the Na" added is dis-placed onto the micellar phase (result notshown).

Diffusible phosphorus levels were notsignificantly modified (figure 2B) but dif-fusible calcium levels were multiplied byfactors of 2.0, 3.5, 4.5 when 17.4, 94.8 and177.8 mmol-kg ", respectively, of sodiumions were added (figure 2B). Our resultsconfirm that NaCI addition induces a pre-ferential solubilization of the calcium bound

to phosphoseryl residues as hypothesizedby Grufferty and Fox [14], Le Graet andBrulé [22] and Famelart et al. [9] owing tothe increase in ionization of these residues.

The casein micelle pellet water contentwas increased by 8, 21, 25 %, respectivelyfor the aforementioned sodium ion addi-tions (figure 2e) as is expected in milk [6, 9,14,35,43]. The amount of supematant pro-tein estimated by Ol) measurement decrea-sed by 25 % when 177.8 mmol-kg"! NaCIwas added (table V). NaCI addition did notsignificantly modify the casein solubilizationin supematants when compared to controlRCMD. The casein micelle pellet watercontent increase could be due to a lowerultracentrifugal effective compression and toa subsequent increase of the amount of pel-leted micelles. The lower ultracentrifugaleffective compression could be explained

Table V. Supematant protein and percentage of supernatant asl- and ~-caseins of reconstitutedcasein micelle-like dispersions (RCMO) containing different salts before heat treatment. Ultracent-rifugation: 77 000 g, 2 h, 20 oc. Control sample: phosphocaseinate suspensed in water without saltaddition. Supematant protein was obtained by optical density at 280 nm in 10 mmol-L:' EDTA ofRCMO supernatants; 00 of supematants were calculated by multiplying the 00 measured by the dilu-tion factor. Percentages of solubilized a 1- and ~-caseins in ultracentrifugal supernatants were esti-mated by ratio of individual casein area ~n ultracentrifugal supematant to the individual case in areain control sam pIe before ultracentrifugation (RP-HPLC analysis).Tableau V. Caséines non ultracentrifugeables et pourcentage de caséines asl et ~ non ultracentri-fugeables dans les dispersions de micelles de caséines reconstituées con~enant différents sels avanttraitement thermique. Ultracentrifugation: 77 000 g, 2 heures, 20 oc. Echantillon témoin: phos-phocaséinate remis en dispersion dans l'eau, sans addition de minéraux. Le taux de caséine nonsédimentable était obtenu par mesure de la DO à 280 nm en présence d'EOTA 10 mrnol-L:' dans lessurnageants de dispersion de phosphocaséinate natif; la DO des surnageants était calculée en mul-tipliant la DO mesurée par le facteur de dilution. Les pourcentages de caséines asl et ~ solubiliséesétaient estimés par le rapport entre la surface obtenue pour une caséine dans le surnageant d'ultra-centrifugation et la surface de la même caséine obtenue dans l'échantillon témoin avant ultracentri-fugation (analyse RP-HPLC).

Sample Supernatant % %protein of solubilized of solubilized

(00280 nm) asl-casein ~-casein

Control RCMD 2.63 0.17 9.80NaCI (177.8 mmol-kg ") 1.97 0.06 7.10CaClz (19.0 mmol-kg') 1.36 0.05 2.68MgClz (19.3 mmol-kg ") 1.60 0.18 2.78Sodium citrate (17.8 mmol-kg') 12.49 43.30 67.30Sodium phosphate (6.8 mmol-kg:") 3.00 0.17 9.80

Heat stability of casein micelles 383

by the increase of protein aggregation. Soodet al. [39] found a negative correlation be-tween the voluminosity of casein micellesand their calcium content. Even though wecan not conclude directly to the existenceof an increase in the hydration of the caseinmicelle from our results, we can conclude tothe existence of an increase in casein micellevoluminosity content after sodium chlorideaddition because of this correlation.

RCMD supplemented in NaCl showedno variation of pH, of total protein contentand of mineraI content (results not shown)when the different heat treatments wereapplied, regardless of the initial pH (wheth-er adjusted to 6.8 or not) whereas controlRCMD precipitated (figure 1). Home andDavidson [16] show that NaCl addition tocasein micelles in skim milk stabilize themagainst aggregation by an increase of thehairy layer but with no modification of thecasein hydrodynamic radius. NaCl additionstabilized casein micelles in water againstheat precipitation probably by enhancementof steric repulsions.

3.4. CaCl2 and MgCI2 additions

Cac~andMgC~ (0-19; 0-19.3 mmol-kg',respectively) additions decreased the pH by0.5 and 0.4 pH unit, respectively, for thehighest concentrations added (figures 3Aand 4A, respectively). Although the pro-portion of added calcium ions or sodiumions determined in the diffusible phase wasthe same (90 %), addition of calcium ionsled to a greater pH decrease than additionof sodium ions did (figures 2A and 3A), ascould be expected from the relationship exis-ting between the activity coefficient and theion valence [46]. As also envisaged for NaCladdition, the pH decrease is also due to anH+ release from the micelles following thefixation of the 10 % calcium added to themicelle.

Diffusible calcium concentrations wereincreased by factors of 2.3, 4.3, 6.0 when

2.6,9.3, 19.3 mmol-kg " of Mg2+ wereadded, respectively (figure 4B). Mg2+addi-tion appears to induce a more complex phe-nomenon than Ca2+addition with only a par-tial displacement of bound calcium sincethe total amount of solubilized Ca2+ andMg2+is close to the added Mg2+. CaCl2 andMgCl2 addition decreased the diffusiblephosphorus concentrations only by0.3 mmol-kg:" for both at the highestconcentrations (figure 4B) thus confirmingthat the RCMD is almost entirely colloidal.

The amount of supematant protein decrea-sed by 48 and 39 % when 19 and19.3 mmol-kg"! of CaCl2 and MgCI2, res-pectively, were added (table V). As for pH,Ca2+addition led to a greater decrease in theamount of supematant protein than Mg2+addition did. Solubilized p-casein was redu-ced to 2.68 and 2.78 % when 19 and19.3 mmol-kg " of CaCl2 and MgCl2 wereadded, respectively (table V). The decreasein the amount of supematant protein could bedue to the aggregation of the calcium sensi-tive caseins.

A 19 mmol-kg " calcium ion additiondecreased micellar pellet water content byapproximately 8 % whilst a 19.3 mrnol-kg"!magnesium ion addition did not significantlymodify it at unadjusted pH (figures 3C and4C, respectively). Calcium ions induced ahigher decrease in pH, in the micelle pelletwater content and in the amount of casein insupematants than magnesium ions did. Thiscould be attributed to their different hydra-ted ionie radius (Ca2+ = 0.412 nm, Mg2+ =0.428 nm [26]) and electronegativities(Ca = 1.2; Mg = 1 on the Pauling scale [22]).

After 30 min of heat treatment, less than5 % of the initial total protein content remai-ned dispersed with an addition of about19 mmol-kg " divalent ions (tables 111andIV). Similar experiments carried out at anadjusted pH of 7.4, after divalent ion addi-tion, also induced casein aggregation (resultsnot shown). As observed in milk [37], addi-tion of CaCl2 or MgCl2 to RCMD has a de-stabilizing effect mainly caused by their

384 C. Le Rayet al.

7.5 ,------------------,

=& 7

6.5

"0 ~ 20~ffIll_E 0 15:::l E~.s!3 s 10~ (;:9.<::<Il 0.:::l <Il:t: 0o "li

5

o~=======~====~~ëCIl ~ 3 ~--------------~~~ë .5o CIl0(5

2 0. 2.5~1§o~~CIl ~ 20.2~ III

.!!! ~~ ~ 1.5 '------'------'------'------'-----'~ o 10 15

CaCI2 (mmol/kg)

Figure 3. Physico-chemical characterization of RCMD sampi es after CaCl2 addition before heattreatment. (A) pH; (B) diffusible calcium (+) and phosphorus (e) concentrations; (C) micellar pel-let water content.Figure 3. Caractérisation physicochimique des échantillons de dispersions de micelles de caséinesreconstituées après ajout de CaCI2, avant traitement thermique. (A) pH ; (B) concentrations en calcium(+) et phosphore (e) diffusibles; (C) teneur en eau des culots micellaires.

5

double charge which shields, depending onthe added concentration, the net negativecharge of the casein micelles. Consequently,the zeta potential is reduced [7] as weIl asthe electrostatic repulsions. Jeumink and deKruif [19] point out that calcium chlorideaddition to milk makes casein micellesshrink and as a consequence of the lowerhydration, the van der Waals attractionwould be larger. Hence, interactions bet-ween the micelles increase. The results pre-sented here are consistent with both ana-lyses. Reduction of steric repulsion of casein

c

20

micelles and of electrostatic repulsion, andincrease of van der WaIls attractions by addi-tion of divalent cations (Ca2+, Mg2+) resul-ted in a lowered stability to heat.

3.5. Sodium citrate addition

Citrate addition (0-17.8 mmol-kg ") tothe RCMD increased the pH up to 8.08(figure 5A), despite a citrate solution pHvalue of7.4. By chelating the colloidal cal-cium, citrate increased the second apparentpK of the phosphoseryl residues and allowed

Heat stability of case in micelles

~g~7.1 L:::::------ ..w 7 1

20)0-",-a::::~ ~ 15-a..sË E 10::J .~

:2 8l<U Co 0><Il <U:0 E'ln -0:J C~ <Uo

o

•

3,--------------=,,----,

51.5 +-----+----+----+-----+-------1

o 20

385

10 15MgCI2 (mmol/kg)

Figure 4. Physico-chemical characterization of RCMD samples after MgClz addition before heattreatment. (A) pH; (B) diffusible calcium (+), phosphorus (e) and magnesium (.Â.) concentrations;(C) micellar pellet water content.Figure 4. Caractérisation physicochimique des échantillons de dispersion de micelles de caséinesreconstituées après ajout de MgClz, avant traitement thermique. (A) pH ; (B) concentrations en cal-cium (+), phosphore (e) et magnésium (.Â.) diffusibles; (C) teneur en eau des culots micellaires.

groups which possess pKs close to the pH ofthe dispersion to bind protons; both pheno-mena resulted in the RCMD pH increase.Moreover, because of the low bufferingcapacity of the dispersing phase, the conse-quence of this phenomenon on the disper-sion pH is higher than in milk serum (datanot shown). Calcium and magnesium chlo-ride addition act conversely by inducingfixation of calcium and magnesium onto themicellar phase.

Citrate ions increased the diffusible cal-cium ion concentrations by 4.1, 6.3 and

8.8 times for addition of 6.0, 10.6 and17.8 mrnol-kg ", respectively (figure SB).Diffusible phosphorus concentrations werecorrespondingly increased by factors of 1.3,1.7 and 2.4 (figure SB). The binding of cal-cium and phosphorus to micelles due to thepH increase was more than compensated bytheir solubilization caused by citrate addi-tion. The ratio of solubilized calcium to solu-bilized phosphorus was 2.4 mol-mol"! when17.8 rnmol-kg " of citrate were added, avalue far from the classical one of 1.5 foundin casein micelles [1] which could be explai-

386 C. Le Rayet al.

8.5 r--------------r':""T'--,

8.0Ia.

7.5

7.01Il:J(;s:a.1Ilo

s:a.'O~ffi~EO:J E'i3 E~~al:0'iii:èi5

8

6

4

2

o o 5 1510citrate (mmol/kg)

Figure 5. Physico-chemical characterization of RCMD samples after sodium citrate addition beforeheat treatment. (A) pH; (B) diffusible calcium (+) and phosphorus (e) concentrations; (C) micellarpellet water content.Figure 5. Caractérisation physicochimique des échantillons de dispersions de micelles de caséinesreconstituées après ajout de citrate de sodium avant traitement thermique. (A) pH ; (B) concentrationsen calcium (+) et phosphore (e) diffusibles; (C) teneur en eau des culots micellaires.

ned by a sequestrating effect of the addedcitrate on both, calcium involved in CCPand bound to phosphoseryl residues.

The amount of supematant protein increa-sed by a factor of 4.7 when 17.8 mmol-kg "sodium citrate was added to RCMD (table \1).Solubilized usl-casein represented 43.30 %of the casein total area when 17.8 mmol.kg:'sodium citrate was added (table V). Solu-bilized ~-casein represented 67.30 % of thiscasein total area (table \1). Citrate additioninduced a release of large amounts of caseinfrom the micellar phase and a solubilization

of us1-casein because of the disorganizationof CCP. Lin et al. [23] and Griffin et al. [13]reported the dissociation of casein micellesafter EDT A addition to milk through dia-lysis with a non-discriminate solubilizationof the four caseins. The dissociation ofreconstituted casein micelles which occur-red in the present study, after sodium citratedirect addition, happened with a highersolubilization of the most phosphorylatedcasein quantified: the us1-casein. The dis-crepancy between our results and the afo-rementioned results can be explained by the

Heat stability of casein micelles 387

mode of addition: direct addition as opposedto dialysis addition.

Citrate addition also drastically increa-sed the casein micelle pellet water contentfrom 2.05 to 8.41 g of water-g'" of total pro-tein for a 7.8 mmol-kg " addition (figure5C). This increase in the amount of thecasein micelle pellet water content can beexplained by the reduction of the micellarsize induced by sodium citrate addition(results not shown). As mentioned forsodium chloride, Sood et al. [39] found anegative correlation between the volumi-nosity of casein micelles and their calciumcontent. We can indirectly conclude that anincrease of casein micelle voluminosity aftercitrate addition occurs because of this cor-relation.

Our results are in agreement with thoseobtained with milk. Citrate ion addition tomilk brings about substantial disintegrationof the micelles by chelation of calcium ionsthereby causing a shift in the distributionfrom the colloidal to the diffusible phase[25]. As a result, a number of interrelatedchemieal and physical properties of the milkare altered: a) reduced sedimentation of themicelles by ultracentrifugation [25], b)increased viscosity [45], c) reduced turbi-dit Y [20, 25] and lightness decrease [20],and d) increased numbers of small particlesand residual open micelle 'skeleton' [45].

No modification of the pH, of the mine-rai distribution between aqueous and micel-lar phase and of protein content was obser-ved after heat treatrnents (results not shown).A heat treatment under the same experi-mental conditions but at an adjusted pH of6.7 was carried out and showed no caseinaggregation (results not shown). Sodiumcitrate protected casein micelles against heataggregation as it improves milk heat stability[24] probably by solubilizing micellar salts.

3.6. Sodium phosphate addition

Phosphate ion addition to the RCMD(0-6.8 mmol-kg'") induced no modification

of the pH values (figure 6A). The pH of thesodium phosphate solution was close to theRCMD control pH. The diffusible calciumconcentrations were not modified (figure6B). It suggests that the added inorganicphosphate ions did not displace calcium ionsfrom the mieellar to the diffusible phase ascitrate did. The calcium affinity is higherfor phosphoseryl residues and for inorga-nie phosphate contained in CCP than for theinorganic phosphate added to RCMD. Thisaffinity difference can be explained by therelatively high electronegativity of the phos-phoseryl clusters and/or by their localconformation [12]. Solubilized phosphorosrepresented 56, 63 and 71 % of the phos-phoros concentration added (figure 6B). Soa part of the phosphate added was bound tocasein micelles. The potential sites of inter-action for the negatively charged inorganicphosphate could be case in micellar calciumand magnesium sites but also basic aminoacid residues (arginyl, Iysyl residues andfree œ-amino terminal) [42]. Calcium phos-phate solubility in RCMD diffusible phaseis lower than in milk diffusible phase.According to Walstra and Jenness [46], theapparent solubility product of CaHP04 inmilk diffusible phase is 6 times higher thanin an infinitely diluted solution. In our case,less phosphorus and calcium ions werefound in the diffusible phase because of thelow ionie strength of the medium. The CCPsolubilization is limited by its solubility pro-duct in the RCMD aqueous phase.

Casein micelle pellet water contentincreased by 8, 16 and 34 % for addition of1.74,4.19 and 7.8 mmol-kg ", respectively(figure 6C). As explained for sodium chlo-ride and sodium citrate addition, we canconclude that phosphate ion addition increa-sed casein micelle voluminosity. Phosphateaddition increased the amount of supematantprotein by 14 % when compared to controlRCMD (table V). It did not modify the per-centage of solubilized us1- and ~-caseins(table V). Phosphate addition causes a non-discriminate solubilization of the casein stu-died.

388 C. Le Rayet al.

7.4

7.5 ,------------------,

7.3Ic.

7.2

7.1

7

7..1-----'--------'------------'

QL--. L- __ ---..I --l ...-J

ë ~3.Q ,..-----------------2 <:52u 22 ~2.5~~--~~~22.Q~ '"'" ~'IDe>u ~~ 1.5 L--. '--- '-- __ ---' _

Q 4phosphate (mmol/kg)

Figure 6. Physico-chemical characterization of RCMD sampI es after sodium phosphate additionbefore heat treatment. (A) pH; (B) diffusible calcium (+) and phosphorus (e) concentrations; (C)micellar pellet water content.Figure 6. Caractérisation physicochimique des échantillons de dispersions de micelles de caséinesreconstituées après ajout de phosphate de sodium, avant traitement thermique. (A) pH ; (B) concen-trations en calcium (+) et phosphore (e) diffusibles; (C) teneur en eau des culots micellaires.

2

After heat treatments, no modification ofpH, mineraI repartition between aqueousand micellar phase and of protein contentwas measured (same concentrations added,same heat treatment; results not shown). Anexperiment carried out at an adjusted pH of6.8 did not induce the heat aggregation ofcasein micelles. Such results agree with theimprovement of the milk heat stability byaddition of sodium phosphate which wasshown by Sweetsur and Muir [41]. Sodium

6 8

phosphate addition protect the RCMDagainst heat instability probably by solubi-Iizing the micellar mineraIs.

4. CONCLUSION

The dispersing aqueous phase which SUf-

rounds casein micelles strongly influencesthe major physico-chemical properties forthe casein micelle dispersions (pH, micelle

Heat stability of casein micelles

pellet water content, mineraI and individualcasein distribution between micellar andaqueous phase). Sodium chloride, sodiumcitrate and sodium phosphate addition byreducing the amount of micellar mineraI andincreasing stabilizing forces (steric repul-sions, etc.) protect RCMD against heat insta-bility. Calcium and magnesium chloride byincreasing the amount of micellar mineraIsand the van der Waals attraction and redu-cing steric and electrostatic repulsions indu-ced heat instability.

Sodium chloride was also found to sta-bilize faba bean protein (vicilin and legu-min) against thermal denaturation whilst cal-cium and magnesium chloride destabilizedthe protein isolate [3]. The denaturation tem-perature of whey protein concentrate deter-mined by differencial scanning calorimetrywas reduced by calcium and magnesiumchloride addition but increased in the pre-sence of sodium chloride on the alkaline sideof the isoelectric zone [44]. Chloride saltsseem to have similar effects on caseinmicelles dispersions, on whey protein concen-trate and on faba bean protein isolate.

ACKNOWLEDGMENTS

We wish to thankM.H. Famelart, P. Schuck,M. Piot, Y. Le Graet, F. Michel for their expertcontribution to this work. This work was sup-ported by Danone and by the National Associa-tion for Technical Research.

REFERENCES

[11 Alais C.; Science du lait: principes des tech-niques laitières, 4e éd., SEPAIC, Paris, 1984.

[21 Aoki T., Suzuki H., Imamura T., Formation ofcasein in whey protein-free milk heated at hightemperature, Milchwissenschaft 29 (1974)589-594.

[3] Amtfield S.D., Murray E.D., Ismond M.A.H.,Effect of salts on the thermal stability of sto-rage proteins from faba bean, J. Food Sci. 51(1986) 371-377.

[4] Brailsford Robertson T., The physical chemistryof the proleins, Longmans, Green and Co., NewYork, 1918.

389

[5] Brulé G., Maubois J.L., Fauquant J., Étude de lateneur en éléments minéraux de produits obtenuslors de l'ultrafiltration de lait sur membrane,Lait 54 (1974) 600-614.

[6] Creamer L.K., Water absorption by rennetedcasein micelles, Milchwissenschaft 40 (1985)589-591.

[71 Dalgleish D.G. Measurement of electrophore-tic mobilities and zeta-potentials of particlesfrom milk using laser Doppler electrophoresis,J. Dairy Res. 51 (1984) 425-438.

[8] Driehuis F., Teemstra EJ.M., Method for tur-bidimetric estimation of lactic acid in fat-contai-ning milk bacteria, Neth. Milk Dairy J. 46 (1992)209-215.

[9] Famelart M.H., Lepesant F., Gaucheron F., LeGraet Y., Schuck P., pH-induced physicoche-mical modifications of native phosphocaseinatedispersions: influence of aqueous phase, Lait76 (1996) 445-460.

[10] Fauquant J., Maubois J.L., Pierre A., Microfil-tration du lait sur membrane minérale, Tech.Lait. 1038 (1988) 21-23.

[II] Fox P.F., Heat-induced changes in milk prece-ding coagulation, J. Dairy Sei, 64 (1981)2127-2137.

[12] Gaucheron F., Le Graet Y., Sinbandhit S.,Guénot P., Brulé G., Binding of calcium to ~-casein in the presence of inorganic phosphate,Sci. Aliments 15 (1995) 481-489.

[13] Griffin A., Lyster R.L.J., Priee J.C., The dissa-gregation of calciurn-depleted casein micelles,Eur. J. Biochem. 174 (1988) 339-343.

[14] Grufferty M.B., Fox P.F., Effect of added NaCIon physicochemical properties of milk, Ir. J.Sci. Technol. 9 (1985) 1-9.

[15] Holt C., The milk salts: their secretion, concen-trations and physical chemistry, in: Fox P.F.(Ed.), Developments in dairy chemistry -3- Lac-tose and minor constituents, Elsevier, London,1985, pp. 143-181.

[16] Home D.S., Davidson C.M., The effect of envi-ronmental conditions on the steric stabilizationof casein micelles, Colloid Polymer Sei. 264(1986) 727-734.

[17] IDF (International Dairy Federation), Determi-nation of total phosphorus (photometrie method),Standard 33C, International Dairy Federation,Brussels, 1987.

[18] Jaubert A., Martin P., Reverse-phase HPLC ana-Iysis of goat caseins. Identification of lX,j andlX,2 genetic variants, Lait 72 (1992) 235-247.

[19] Jeumink TJ.M., de Kruif K.G., Calcium concen-tration in milk in relation to heat stability andfouling, Neth. Milk Dairy J. 49 (1995) 151-165.

[20] Johnston D.E., Murphy RJ., Effects of sornecalcium-chelating agents on the physical pro-perties of acid-set milk gels, J. Dairy Res. 59(1992) 197-208.

390 C. Le Rayet al.

[21] Kudo S., Influence of lactose and urea on the tation par atomisation de phosphocaséinate natifheat stability of artificial milk systems, NZ J. obtenu par microfiltration sur membrane, Lait 74Dairy Sei, Technol. 15 (1980) 197-200. (1994) 375-388.

[22] Le Graet Y., Brulé G., Les équilibres minéraux [35] Singh H., Effects of high temperatures on caseindu lait: influence du pH et de la force ionique, micelles, NZ J. Dairy Sci. Technol. 23 (1988)Lait 73 (1993) 51-60. 257-273.

[23] Lin S.H.C., Leong S.L., Dewan R.K., Bloom- [36] Singh H., Fox P.F., Heat stability of milk:field V.A., Morr C.V., Effect of calcium ion on influence of colloidal and soluble salts and pro-the structure of native bovine casein micelles, tein modification on the pH-dependent disso-Biochemistry II (1972) 1818-1821. dation of micellar x-casein, J. Dairy Res. 54

[24] Mc Crae c.H., Muir D.D., Heat stability in milk, (1987) 523-534.in: Heat induced changes in milk, IDF special [37] Singh H., Creamer L.K., Heat stability of milk,issue 9501, 1995, pp. 206-230. in: Fox P.F. (Ed.), Advanced dairy chemistry -

[25] Morr C.V., Sorne effects of pyrophosphate and 1- Proteins, Elsevier, London, 1992, pp.citrate ions upon the colloidal caseinate-phos- 621-656.phate micelles and ultrafiltrate of raw and hea- [38] Singh H., Roberts M.S., Munro P.A., Tet Teoted skimmilk, 1. Dairy Sci. 50 (1967) 1038-1044. C, Acid-induced dissociation of casein micelles

[26] Nightingale Jr E.R., Phenomenological theory of in milk: effects of heat treatrnent, 1. Dairy Scî. 79ion solvation, effective radii of hydrated ions, (1996) 1340-1346.J. Phys. Chem. 63 (1959) 1381-1387. [39] Sood S.M., Gaind D.K., Dewan R.K., Correlation

[27] Pierre A., Brulé G., Mineral and protein equili- between micelle solvation and calcium content,bria between the colloidal and soluble phases NZ J. Dairy Sci. Technol. 14 (1979) 32-34.of milk at low temperatures, J. Dairy Res. 48 [40] Strange E.D., van Hekken DL, Holsinger V.H.,(1981) 417-428. Effect of sodium chloride on the solubility of

[28] Pierre A., Brulé G., Fauquant J., Étude de la caseins, J. Dairy Sci. 77 (1994) 1216-1222.mobilité du calcium dans le lait à J'aide du cal- [41] Sweetsur A.W.M., Muir D.D., The use of per-cium 45, Lait 63 (1983) 473-489. mitted additives and heat treatment to optimize

[29] Pierre A., Fauquant J., Le Graet Y., Piot M., the heat stability of skim milk and concentra-Maubois J .L., Préparation de phosphocaséinate ted skim milk, J. Soc. Dairy Technol. 33 (1980)natif par micro filtration sur membrane, Lait 72 101-105.(1992) 461-474. [42] ter Horst M.G., A new theory on the caseinate

[30] Pouliot M., Pouliot Y., Britten M., Maubois 1.L., complex in milk, Neth. Milk Dairy J. 17 (1963)Fauquant J., Study of the dissociation of ~-casein 185-192.from native phosphocaseinate, Lait 74 (1994) [43] van Hooydonk A.C.M., Hagedoom H.G., Boer-325-332 rigter I.J., The effect of various cations on the

[31] Sandblom (Alfa Laval), Filtering process, Swe- renneting ofmilk, Neth. Milk Dairy J. 40 (1986)dish Brevet, 74/16257,1974. 369-390.

[32] Schmidt D.G., Koops J., Properties of artificial [44] Varunsatian S., Watanabe K., Hayakawa S.,casein micelles. 2. Stability towards ethanol, Nakamura R., Effects of Ca2+, Mg2+ and Na"dialysis, pressure and heat in relation to casein on heat aggregation of whey protein concen-composition, Neth. Milk Dairy J. 31 (1977) trates, J. Food Sei, 48 (1983) 42-46.342-357. [45] Visser J., Minihan A., Smits P., Tjan S.B.,

[33] Schmidt D.G., Both P., Koops J., Properties of Heertje L, Effects of pH and temperature on theartificial casein micelles. 3. Relationship be- milk salt system, Neth. Milk Dairy J. 40 (1986)tween salt composition, size and stability 351-368.towards ethanol, dialysis and heat, Neth. Milk [46] Walstra P., Jenness R., Dairy chemistry andDairy J. 33 (1979) 40-48. physics, John Wiley and Sons, New York, 1984.

[34] Schuck P., Piot M., Mejean S., Le Graet Y., [47] Zin el Din M., Aoki T., Polymerization of caseinFauquant J., Brulé G., Maubois J.L., Deshydra- on heating milk, Int. Dairy 1. 3 (1993) 581-588.