characterization of whey protein isolate obtained from milk

Lait (1996) 76, 255-265© Elsevier/INRA

Original article

Characterization of whey protein isolate obtainedfrom milk microfiltration permeate

M Britten1, Y Pouliot2

1Agriculture and Agri-Food Canada, Food Research and Development Centre, 3600, CasavantBlvd West, St Hyacinthe, Ouebec, Canada, J2S 8E3; 2Centre de recherche en sciences et

technologie du lait (STELA) Université Laval, Sainte-Foy, Ouebec, Canada, G1K 7P4

(Received 5 April 1995; accepted 20 July 1995)

Summary - Whey protein was concentrated from milk microfiltration permeate or Cheddar cheesewhey using batch ultrafiltration and diafiltration. Permeate flux declined more rapidly for milk microfil-trate during concentration, but similar flux profiles were observed on both wheys during diafiltration.Protein contents in concentrates from milk microfiltrate and cheese wheys were respectively 95 and74% on a dry matter basis, while their fat contents were respectively 0.2 and 15%. Protein solubilityof the milk microfiltrate isolate was over 95% between pH 3 and 8, while protein solubility of cheesewhey concentrate varied between 75 and 90% over the same pH range. J3-Lactoglobulin in cheesewhey concentrate showed ex1ensive structural alteration as evidenced by the low proportion of SHgroups in native position. J3-Lactoglobulin structural alteration was less severe in protein isolated frommilk microfiltrate, but still significant. Milk microfiltrate protein isolate showed better gelling propertiesthan cheese whey protein concentrate. Cheese whey protein concentrate did not foam at ail, whilestiff and stable foams were produced from milk microfiltrate protein isolate. Gelation and foaming werevery sensitive to the ionic environment and were controlled by a balance between the pH and thecalcium concentration. Reduction of calcium content in milk microfiltrate protein isolate, using salinesolution during diafiltration, had only slight effects of protein functional properties.

milk microfiltration / whey protein 1ultrafiltration / heat set gelation / foaming property

Résumé - Caractérisation d'un isolat protéique de lactosérum obtenu à partir d'un perméat demicrofiltration de lait. Les protéines d'un perméat de microfiltration de lait et d'un lactosérum defromage Cheddar ont été concentrées par ultrafiltration en batch suivie d'une dia filtration. La compa-raison des flux de microfiltration indique un colmatage plus rapide pendant la concentration du micro-filtrat de lait. Les flux de perméation en fin de concentration et pendant la dia filtration étaient toutefoissimilaires. Les teneurs en protéines des concentrés protéiques de micro filtrat de lait et de lactosérumfromager étaient respectivement de 95 et 74%, alors que les concentrations lipidiques atteignaientrespectivement 0,2 et 15,5 %. La solubilité des isolats de microfiltrat de lait était supérieure à 95 %entre pH 3 et 8, alors que la solubilité des concentrés de lactosérum fromager variait de 75 à 90 %sur la même fourchette de pH. La structure native de la J3-lactoglobuline, évaluée par la position dugroupement SH libre, était fortement altérée dans le concentré de lactosérum fromager. L'altérationstructurale de la J3-lactoglobuline était moins importante dans J'isolat de microfiltrat de lait, maistoutefois significative. Les propriétés thermogélifiantes de J'isolat de micro filtrat de lait étaient supé-rieures à celles du concentré de lactosérum fromager. L'isolat de microfiltrat de lait formait des

256 M Britten, Y Pouliot

mousses fermes et stables alors que le concentré de lactosérum fromager refusait de mousser. Lespropriétés gélifiantes et moussantes étaient sensibles à l'environnement ionique et modulées par lacombinaison du pH et de la disponibilité du calcium dans le milieu. La réduction de la teneur en calciumde l'isolat protéique, par l'utilisation d'une solution saline comme solvant de diafiltration, n'a cependanteu qu'une influence mineure sur ses propriétés fonctionnelles.

micro filtration / protéine du lactosérum / ultrafiltration / gélification thermique / propriétémoussante

INTRODUCTION

Recent developments in membrane te-chnology led to commercial uses of cross-flow microfiltration in dairy industry. Micro-filtration is used to remove bacterialcontaminants and suspended materialsfrom milk or whey (Trouvé et al, 1991;Gésan et al, 1993; Mucchetti and Tagliet-ti, 1993). Microfiltration makes possiblealso the selective concentration of nativecasein micelles (Fauquant et al, 1988).This concentrate is heat stable and weilsuited for cheese milk standardizationand enrichment (Pierre et al, 1992). Fur-thermore, casein concentration throughmicrofiltration produces a high value milkmicrofiltrate.

Milk microfiltration permeate is obtainedfrom non-fermented milk, which bringsimportant differences when comparedto cheese whey. Milk microfiltrate is almostfree from bacteria and bacteriophages(Gautier et al, 1994). It requires then verymild heat treatment to ensure salubrity andto be used as in the formulation of culturedproducts. Milk microfiltrate is also free fromdegradation products related to starter cul-ture and milk coagulation enzyme activi-ties. Since milk is not acidified before wheyseparation, the minerai equilibrium is not al-tered and remains similar to that of milk.Compared to cheese whey, lower amounts ofcalcium and phosphorus and almost no or-ganic acids are expected. Finally, milk micro-filtrate is expected to contain only traces offat. Clarification treatments to remove lipidsare not required to produce ingredients withgood functional properties.

The purpose of this study was to concen-trate protein from milk microfiltration per-meate and cheese whey and to comparetheir functional properties.

MATERIALS AND METHODS

Preparation of protein concentratesfrom milk microfiltrate and cheesewhey

Fresh raw milk was skimmed and microfiltered,with the MFS-7 Alfa-Laval equipment (Alfa-Laval, Lund, Sweden) operated at 50 oC. An ini-tiai filtration using a 1 P 19 Membralox (SCT,Tarbes, France) 1.4 um pore size cartridge(0.85 m length; internai diameter of channels:4 mm; area 0.2 m2) was performed according toTrouvé et al (1991), to reduce bacterial contami-nation. A second filtration using a 0.1 um poresize cartridge was performed according to Pierreet al (1992), to concentrate native phosphoca-seinate. The permeate of this second microfiltra-lion, containing whey proteins, lactose and mine-rais, was used as starting material. A 120 kgmass of milk microfiltrate was batch-concen-trated up to weight concentration factor 16 withthe UFS-1 Alfa-Laval equipment fitted with aPM 10 polysulfone hollow fibre (internaidiameter of fibres: 1.5 mm; area 1.3 m2) ultrafil-tration cartridge (Romicon Corp, Woburn, MA,USA). Temperature during concentration wasmaintained at 50°C. Inlet and oullet pressurewere maintained at 1.8 and 0.6 bar, respectively.Permeate flux was measured and permeatesarnples were withdrawn at weight concentrationfactors (CF) 1, 1.5, 2, 3, 5, 8, 12, and 16. Follow-ing concentration, the retentate (7.5 kg) was dia-filtered by the addition of 37.5 kg of deionizedwater in continuous mode, which correspondedto a final mass dilution rate (a = kg addedwater/kg retentate) of 5. Permeate flux wasmeasured and permeate samples were with-

Milk permeate prote in isolate

drawn at dilution rates (a) 0, 1,2,3,4 and 5. Theretentate was then freeze-dried and stored at-20°C until further analysis. A second concen-trate was produced by the same process but dur-ing diafiltration, sodium chloride solu-tion (0.3% w/w) was used instead of water up toa = 3, in order to further reduce the calcium con-tent of the retentate. Excess sodium wasremoved using deionized water from a = 3 toa = 5. A third concentrate was obtained frompasteurized Cheddar cheese whey (from a localcheese factory). The same process was appliedto cheese whey, with the use of sodium chloridesolution in the early stages of diafiltration. Eachprotein concentrate was produced three timeswith starting mate rial from inde pendent bat-ch es (raw milks or cheese wheys).

Composition analysis

Ultrafiltration permeate samples were analyzedfor total nitrogen (Kjeldahl), ashes (5 h, 550 OC)and calcium (adapted from Pearce, 1977). In ad-dition, freeze-dried protein concentrates wereanalyzed for non-casein nitrogen (NCN) andnon-protein nitrogen (NPN) (Rowland, 1938).The main protein fractions were analyzed byreverse phase HPLC using C4 bonded silica col-umn (Vydac 214TP, The Separations Group,Hesperia, CA, USA) according to Kim etal (1987). Total lipids were determined frommethanol/chloroform (2:1 v/v) extraction accord-ing to Pierre et al (1994). Lactose was spectro-photometrically measured (Dubois et al, 1956)and the minerais (Ca, Mg, Na, K) were detectedusing inductively coupled plasma spectro-metry (ICP spectrometer model 3510, AppliedResearch Laboratories, Sunland, CA, USA).

Retention coefficients

The averaged retention coefficient of solutesduring concentration was calculated fromchanges in retentate composition, according toGlover (1985):

C = Co. CFR

where Co is the initial concentration of solute, C,the solute concentration at concentration factorCF and R, the averaged retention coefficient.The R value was the slope of Log CICo vsLog CF and was obtained through Iinear re-gression analysis.

257

Average retention coefficient during diafiltrationwas also determined from changes in retentatecomposition (Glover, 1985).

Co = C. e (l-R).a (2)

where a is the dilution rate. The R value wasobtained from the slope (1-R) of ln ColC vs a andwas obtained through Iinear regression analysis.

Free SH groups determination

Total free SH groups were determined spectro-photometrically from reaction with DTNB. Ac-cording to Shimada and Chetlel (1989), the SHgroup in the native state of ~-Iactoglobulin (posi-tion 119/121) reacts more slowly with DTNB inpresence of SDS than other SH groups resultingfrom SH/S-S interchange reactions. The kineticdata of the reaction (reacting SH groups vs time)were then fitted to a double exponential equa-tion. The first exponential represented the con-tribution from the slow reacting SH groups andthe second, the contribution from the fast reac-ting SH groups:

SH, = SHs. (1-e-·0034') + SHF. (1_e-·04612') (3)

(1)

where SH, is the concentration of SH group re-acting with DTNB at time t, SHs, the concentra-tion of slow reacting SH groups in the proteinsample and, SHF, the concentration of fast reac-ting SH groups in the protein sam pie. Rate con-stant of the first exponential (-0.0034) was ob-tained from kinetic data of whey protein isolatedfrom acidified fresh raw milk and fitted to a singleexponential modal. ~-Lactoglobulin in fresh rawmilk was assumed to contain exclusively slowreacting SH groups. The rate constant for thesecond exponential (-0.0461) was obtainedfrom kinetic data of denatured whey protein iso-late (1% solution heated at 90 "C for 20 min),fitted to a double exponential with -0.0034 usedas the rate constant of the first exponential. SHgroup concentrations are reported as urnol per9 of ~-Iactoglobulin (according to the proteincomposition (table 1), the theoretical contributionof BSA to the free SH groups content of the iso-lates was less th an 1% and was neglected incalculations).

Solubility of protein concentra tesProtein solubility was determined from whey pro-tein concentrates suspended in deionized water.


Table 1. Composition of whey protein concentrateson dry matter basis (g kg-I).Composition des concentrés protéiques surbase sèche (g kg-I).

Concentra te from MM MM-LC' CW

Protein 9588" 9628 779NCN (x 6.38) 8958 8908 660j3-lactoglobulin 7948 7888 591a-Iactalbumin 718 78a 548

BSA 308 248 148

NPN (x 6.38) 98 88 38

Lipids 28 2a 155

Lactose 13.4a 14.2a 22.2

Ash 18a 20a 23a

Ca 3.3 1.0a 1.1a

Mg 0.36 0.09a 0.108

Na 0.7a 4.4a 3.0a

K 1.5 0.4a 0.2a

P 0.58 0.3a 0.58

"Low calcium (diafiltered with NaCI solution).Faible teneur en calcium (diafiltré avec une so-lution de NaCI)."Means in a same row, followed by the sameletter are not significantly different (a = 0.05).Les moyennes d'une même ligne suivies de lamême lettre ne sont pas significativement dif-férentes (a = 0,05).

The effect of pH and added calcium on the solu-bility were tested through a 11 x 4 completelyrandomized design. The pH of the solution wasadjusted (between 3 and 8 with 1 mol L-1 HCIor 1 mol L-1 NaOH) and CaCI2 was added at fourdifferentlevels (0, 5,10 and 15 mmol L:'). Finalprotein concentration was adjusted to 10 9 L-l.Protein dispersions were centrifuged at 20 000 gfor 15 min. Solubility was obtained from the ab-sorbance ratio between the supernatant and thedispersion before centrifugation. Absorbancewas measured at 280 nm on a sam pie aliquotdiluted 1: 10 (v/v) in dissociating bull-er (50 mmol L-l EDTA; 8 mol L-1 urea; pH 10).

Gelling properties of proteinconcentrates

Whey protein concentrates were dispersed indeionized water. The effect of pH (4.0,6.0 and 8.0)

and added calcium (0, 5, 10 and 15 mmol L") ongelling properties were tested through a split plotfactorial design, with the whey concentrates andthe pH in the main plot and the level of addedcalcium in the subplot. Final protein concentra-tion was adjusted to 75 9 L-1. A 2-ml sam pie ofprotein solution was placed in a 11-mm internaidiameter test tube and heated at 90 "C for30 min. Gels were cooled in an ice bath for 5 minand allowed to stand 2 h at 22 ± 1 "C before ana-Iysis. Strain-deformation curves were monitoredat the same temperature using a texture ana-Iyzer (model TA-XT2, Texture TechnologiesCorp, Scarsdale, NY, USA). The probe was a fiatcylinder (6-mm diameter) which penetrated thegel at a speed of 0.8 mm S-1. Maximal deforma-tion was set to 15 mm. Gel strength was associ-ated with the force required to fracture thegel (maximum on strain-deformation curve).Broken gels were then centrifuged (2500 rpm,15 min), and syneresis index was calculatedfrom the volume proportion of free serum. Ali gelswere prepared and analyzed in triplicates.

Foaming properties of proteinconcentrates

Whey protein concentrates were dispersed indeionized water. The effect of pH (4.0, 6.0 and 8.0)and added calcium (0, 5,10 and 15 mmoL L-l) onfoaming properties were tested through a splitplot factorial design, with the whey concentratesand the pH in the main plot and the level of addedcalcium in the subplot. Final protein concentra-tion was adjusted to 50 9 L-1. A 100-mLsamplewas whipped at 22 "C ± 1 "C with an electricalmixer (model Mixmaster, Sunbeam Corporation,Toronto, Canada) operating at maximum speedfor 5 min. Foam expansion factor was calculatedfrom the density ratio between the solution andthe foam. Foam firmness was measured with aBrookfield viscometer (Model DVII) mounted ona Helipath support and fitted with a 'T'typespindle. Spindle rotational speed was set to12 rpm, while vertical displacement was fixed at2.5 cm rnirr", Apparent viscosity was measuredalter t-orn penetration. Foam drainage stabilitywas defined as the time required to drain half thefoam weight. For that purpose, a 50-mL fun-nel ('combitip' from Brinkmann, Rexdale, Ca-nada) was filled with freshly prepared foam andplaced over a balance. The balance was con-nected to a computer which collected weightdata of drained Iiquid as a function of time.

Milk permeate protein isolate

Statistical analyses

Analysis of variance was used to determine if thefactors and their interactions had a significanteffect on the measured properties (SAS InstituteInc, Cary, NC, USA). Statistical analyses wereperformed at an a = 0.05. Composition andphysicochemical properties of whey protein con-centrates were compared through multiple com-parisons of least square means at controlleda level of 0.05. Contrast analyses were used tocompare the gelling and foaming behaviours ofwhey protein concentrates. Error bars on graphsrepresent the standard error obtained from thestatistical models.

RESULTS AND DISCUSSION

Concentration of whey protein

Cheese whey and milk microfiltrate werebatch-concentrated up to CF 16. As aconsequence of membrane fouling, thepermeate flux permeate rapidly decreasedwith increasing CF (fig 1a). In the earlystages of concentration, flux declined fas-ter with milk microfiltrate than with cheesewhey (P = 0.0001). At CF greater than 10,permeate fluxes values were similar andaveraged 18 L h-1 m-2• Considering the ge-nerally accepted relationship between ul-trafiltration flux decline and residual lipidsin whey (de Wit and de Boer, 1975; Merinand Gordin, 1983; Piot et al, 1984; Daufinet al, 1992), lower flux observed duringconcentration of milk microfiltrate was notexpected. However, Iipid content is not theonly composition factor controlling permeateflux. Calcium phosphate precipitation duringconcentration of dairy fluids is responsible forsevere flux decline (Merin and Cheryan,1980; Glover, 1985). The pH of milk microfil-trate was slightly higher than that of cheesewhey (6.6 vs 6.2), which could have promo-ted minerai precipitation. In order to reduceminerai fouling, it would be appropriate toslightly reduce the pH of milk microfiltrate orto operate the ultrafiltration unit at a tempera-ture lowerthan the temperature used duringmilk microfiltration.

259

",--------,-,

20

1°0 5 la 15

cooceneetcn fader (Cr)o 1 2 3 4 5

dilutiOn rate (Cl)

Fig 1. Permeate flux during (a), batch-concen-tration (e, milk microfiltrate; ., cheese whey)and (b), diafiltration of whey.Flux de perméation pendant (a), la concentra-tion en batch (e, microfiltrat de lait;., lactosé-rum de fromage) et (b), la dia filtration du lactosé-rum.

The average retention coefficients (R) fornitrogen compounds during concentrationof milk microfiltrate and cheese whey werenot significantly different (P = 0.4221) andaveraged 0.83 (data not shown). Calciumretenti on during concentration was slightlyhigher for cheese whey (R = 0.18) than formil k microfiltrate (R = 0.16) (P = 0.0401).This difference was attributed to the portion ofnon-permeable calcium which increased withthermal treatments (pasteurization) of cheesemilk and whey (Brulé and Fauquant, 1981).

Diafiltration of whey protein

After concentration (CF = 16), cheesewhey (CW) and milk microfiltrate (MM)were diafiltered in continuo us mode up toa = 5. The permeate flux was not affectedeither by the type of whey or the diafiltrationsolvent (water vs NaCI solution)(P = 0.3658). However, the permeate fluxshowed variations associated with dilutionrate (a) (P=0.0213) (fig 1b). From a=Otoa = 3, the flux gradually increased up to120% of its initial value. Flux increase wasattributed to the viscosity decrease of UF-permeate resulting from the elimination oflactose (Peri et al, 1973). Further diafiltra-


~ -1s

dilutionrate(u)

Fig 2. Change in calcium concentration du ringdiafiltration of whey concentrate .• , MM-concen-trate diafiltered with deionized water; 0, MM-concentrate diafiltered with NaCI solution up toa = 3, followed with deionized water; e, CW-concentrate diafiltered with NaCI solution up toa = 3, followed with deionized water.Évolution de la concentration en calcium pen-dant la dia filtration du concentré de lactosérum.., concentré de microfiltrat de lait dia filtré avecl'eau déionisée; 0, concentré de microfiltrat delait dia filtré avec une solution de NaCI jusqu'àa = 3, puis avec l'eau déionisée ; e, concentréde lactosérum de fromage diafiltré avec unesolution de Na CI jusqu'à a = 3, puis avec l'eaudéionisée.

tion reversed the trend, with permeate fluxalmost back to its initial value at a = 5, sug-gesting further fouling of the UF membrane.

Average retention for nitrogen com-pounds was not affected either by the typeof whey or the diafiltration sol-vent (P = 0.6241) with a coefficient of 0.95.Calcium retenti on coefficient increased du-ring diafiltration as evidenced by the non-li-near relationship between ln (Co/C) vs acurves (fig 2). This deviation from li-nearity was attributed to the portion ofnon-permeable calcium. Non-permea-ble calcium, expressed on a protein ba-sis averaged 3.4 mg g-' in milk microfil-trate protein isolate. It was reduced by70% when sodium chloride was used inthe early stages of diafiltration. Use ofsodium chloride during diafiltration of cheesewhey led to a similar result.

Composition of whey protein iso/ates

Compositions of MM-isolates and CW-concentrate are presented in table 1. Pro-tein concentration of MM-isolates avera-ged 96% on a dry matter basis. TheCW-concentrate showed lower proteincontent (77.9%) due to the significant a-mount of lipids concentrated with proteinsduring ultrafiltration. In MM-isolates, thepH 4.6-insoluble fraction averaged 7.0% oftotal protein. Pierre et al (1992) found a si-milar proportion in milk microfiltration per-meate (6.4%) and associated it to the pre-sence of caseins. The pH 4.6-insolubleprotein fraction in CW-concentrate was hi-gher (15.3%). Denaturation of whey pro-teins, induced by pasteurization steps du-ring cheese making would explain thedifference. The proportion of the variouswhey proteins was altered by the concen-tration process. The weight ratio betweenp-Iactoglobulin and a-Iactalbumin, which issi i 9 ht 1 Y hi 9 he r th a n 4: 1 in mil k s e-rum (Marshall, 1982), increased to at least10:1 in ultrafiltration concentrate. Polyme-ric membranes, such as those used in thepresent study are generally characterizedby a diffuse eut-off and allow slight proteinpermeation according to molecularsize (Cheryan, 1986). A lower extent of pro-teolysis in milk microfiltrate compared tocheese whey would explain the lower NPNcontent in MM-isolates. Lipid content of MM-isolates was much lowerthan in CW-concen-trate. The minerai composition of MM-isolatewas modified by the use of sodium chlorideinstead of water in the early stages of diafil-tration: Ca, Mg and K concentrations werereduced by about 70%, while sodium increa-sed from 0.7 9 kg-1 to 4.4 9 kg-'.

Protein denaturation

The kinetics of the reaction between freeSH groups and DTNB reagent was used todetermine the extent of SH/S-S inter-change of p-Iactoglobulin (table Il). The


Table II. SH groups characteristics in wheyprotein concentrates (urnol g-1 ~-Iactoglobulin).Caractéristiques des groupements SH dans lesconcentrés protéiques Mmol g-I ~-Iactoglobuline).

SH;nterchanged SHnative SHtot

MM-isolate

MM-isolate LC" 10.1a 28.7a 38.8a

CW-concentrate 23.3a 6.6 29.9

'Means in a same row, lollowed by the same leller arenot signilicantly different (a = 0.05). **Low calcium (dia-liltered with NaCI solution).*Les moyennes d'une même ligne suivies de la mêmelettre ne sont pas significativement dif-férentes (a = 0,05). **Faible teneur en calcium (dia-filtré avec une solution de NaCI).

slow-reacting SH group content of p-Iacto-globulin in MM-isolates was close to29 urnol s' and represented 75% of totalSH groups. The concentration process in-duced significant structure alterationthrough SH/S-S interchange reactions re-sulting in 25% decrease of native p-Iacto-globulin content. In the present study, wheywas batch-concentrated. lt was then sub-mitted to pumping and was maintained at50 "C for a long period of time. Such condi-tions have been shown to induce partialprotein unfolding (Harris et al, 1989). In or-der to preserve the native structure of P-lactoglobulin, it would be appropriate to usea system configuration with shorter resi-dence time and reduce filtration tempera-ture. p-Lactoglobulin from CW-concentrateshowed a lower level of slow reacting SHgroups which represented 22% of total SHgroups. This proportion of SH group in na-tive position is close to the expected equi-librium value (20%) assuming an extensiveand random interchange reaction.

According to its molecular mass (18 362Da; Swaisgood, 1982) the total free SHgroup content of native p-Iactoglobulin is

261

54.51lmol g-1. A lower free SH groupcontent reflects the formation of intermole-cular disulphide bonds or other oxidationproducts (Donovan and Mulvihill, 1987). Intable Il, it can be seen that the total SHgroup content of p-Iactoglobulin in MM-iso-lates averaged 38.5 urnol g-1, correspon-ding to 29.4% of the original SH being oxi-dized. Batch ultrafiltration, performed at50 "C, was likely to induce significant al-teration of whey protein structure. A lowerSH content of p-Iactoglobulin was foundin CW-concentrate, where 45.1 % of theoriginal content of SH groups had beenoxidized.

So/ubility of whey protein iso/ates

Solubility of protein isolates was measuredat various pHs. Solubility was not influen-ced by the addition of calcium (P = 0.2601)and results presented in figure 3 are theaverages of solubility measured at the fourcalcium levels. The solubility of MM-iso-lates averaged 97% at pH values below4.5 and higher than 5.5. A slight decreasewas observed at the isoelectric point of P-lactoglobulin (IP = 5.3). The CW-concen-trate showed lower solubility than MM-iso-

100 ,-----------------,

95

80

75 L.L __ -'--_----:- __ :-_---;---_----;:---'

pH

Fig 3. Solubility of whey protein concentrates.., MM-Isolate; 0, low calcium MM-isolate;e, CW-concentrate.Solubilité des concentrés protéiques de lactosé-rum .• , isolat de microfiltrat de lait; 0, isolat demicrofiltrat de lait faible en calcium; e, concentréprotéique de lactosérum de fromage.


lates at any pH tested. Furtherrnore, solu-bility in the low pH portion of the profile didnot improve compared to the solubility atthe isoelectric point of ~-Iactoglobulin. Thesolubility profile suggests that proteins incheese whey suffered from severe treat-ment probably associated with the cheesemaking practices or whey pasteurizationconditions.

Gelling properties

Thermal gels were produced from wheyprotein solutions. The effects of pH adjus-tment and calcium addition on gel proper-ties were tested. The gel strength was si-gnificantly affected by both variables andthe type of whey protein concen-trate (P = 0.0001) (fig 4). Cheese wheyconcentrate produced very soft gels withmaximum gel strength values around 50 9under optimum conditions (pH 4.0). Pre-sence of fat in the concentrate (Morr et al,1993) and evidence of severe protein de-naturation (Mangino et al, 1987) could ex-plain the poor gelling performance.

350,-----.r------,r---~

300

250

100

o 0

Fig 4. Effect of pH and added calcium on wheyprotein gel strength. a, pH 4; b, pH 6; c, pH 8;., MM-isolate; 0, low calcium MM-isolate;., CW-concentrate.Effet du pH et de l'ajout de calcium sur la fermetédes gels protéiques. a, pH 4 ; b, pH 6; c, pH 8 ;., Isolat de microfiltrat de lait; 0, isolat demicrofiltrat de lait faible en calcium ; ., con-centré protéique de lactosérum de fromage.

Protein isolate from milk microfiltrate gel-led at the three pHs tested. From contrastanalysis, it was shown that the use of so-dium chloride during the early stages of dia-filtration did not influence the gelstrength (P = 0.8524). lt seems that cal-cium complexes did not affect gelation.However, the addition of ionic calcium in-creased the gel strength and this effect wasmore important as the pH increased. Cal-cium ions promoted protein-protein interac-tions through cross-Iinking negatively char-ged residues. Increasing pH increased theproportion of residues involved in calciumbridges. lt has been shown that thestrength of thermal gels from clarified wheyprotein concentrates increased with pH in-creasing from 5 to 9 (Gault et al, 1990). Atlow pH, the gel strength was essentially ex-plained by non-specifie interactions be-tween protein molecules, while calciumbridges and disulphide bond formation ex-plained gel strength at higher pH (Schmidtet al, 1979). A combination of high pH andhigh ionic calcium content produced thestrongest gels.

Whey protein gels were centrifuged andsyneresis was monitored (fig 5). Ali the va-riables studied (pH, calcium, proteinsource) and their statistical interactionshad significant effects on gel synere-sis (P = 0.0001). As a general trend, gelsfrom MM-isolates showed lower levels ofsyneresis than gels from CW-concentrate.Again, the use of sodium chloride duringthe early steps of diafiltration did not affectsyneresis (P = 0.1500). Minimum synere-sis was observed on gels obtained atpH 4.0. The aggregated protein matrix for-med at this pH was soft (fig 4) and highlyhydrated. Addition of calcium for gels pro-duced at pH 8.0 decreased syneresis. Un-der alkaline conditions, calcium ions contri-buted to the formation of strong gelmatrix (fig 4) which resisted deformationupon centrifugation. Under these condi-tions, low syneresis was related to the gelmatrix mechanical properties.


a b c

0

0

H0

~0

~°r+ I-ido ±

100

90

80

30

10 15 0 5 10 15 0addecIcalcîum(mmolr')

10 15

Fig 5. Effect of pH and added calcium on wheyprotein gel syneresis. a, pH 4; b, pH 6; c, pH 8;., MM-isolate; 0, low calcium MM-isolate; .,CW-concentrate.Effet du pH et de l'ajout de calcium sur la synérèsedes gels protéiques. a, pH 4 ; b, pH 6 ; c, pH 8 ;., isolat de microfiltrat de lait; 0, isolat demicrofiltrat de lait faible en calcium; ., concen-tré protéique de lactosérum de fromage.

Foaming properties

Foaming properties of whey protein solu-tions were determined at various pHs anddifferent levels of added calcium. Cheesewhey protein concentrate did not foam atail over the entire range of conditions tes-ted. The lipid content of CW-concen-trate (table 1) was responsible for defi-cient foaming behaviour (Maubois, 1988;Joseph and Mangino, 1988). Protein iso-lates from milk microfiltrate showed muchbetter foaming properties (fig 6). Foamexpansion was slightly lower for low cal-cium isolate (use of sodium chloride inthe early steps of diafiltration)(P = 0.0069), and the difference remai-ned constant over the range of conditionstested. The addition of calcium to the pro-tein solution had no significanteffect onfoam expansion (P = 0.2354). lt sug-gests that transport and spreading ofprotein molecules at the air-water inter-face during the foaming process are notinfluenced by calcium concentration.However, foam expansion of MM-isolates

263

" ba c

,,~ ~ ; i rIi t :§: :l: -,

10 ~

2

I-r ~ ~ l l l l l l l lI~ ~ .L

0 10 15 0 5 10 15 0added calcium (mmol r~

10 15

Fig 6. Effect of pH and added calcium on wheyprotein foam expansion factor. a, pH 4; b, pH 6;c, pH 8; ., MM-isolate; 0, low calcium MM-iso-late; ., CW-concentrate.Effet du pH et de l'ajout de calcium sur le facteurd'expansion des mousses protéiques. a, pH 4 ;b, pH 6; c, pH 8; ., isolat de microfi/trat de lait ;0, isolat de microfiltrat de lait faible en calcium;., concentré protéique de lactosérum de fromage.

was influenced by the pH of the solution. AspH increased, the foam expansion factordecreased. Alkaline conditions increasedthe net negative charge of protein mole-cules. Transport to the air-water interfacewas then slowed down by increased elec-trostatic repulsion between proteins in so-lution and proteins already adsorbed at theinterface (Kinsella and Whitehead, 1989).Increasedcharge of protein molecules alsopromoted intra-molecular repulsions andspreading at the interface. These two fac-tors were associated with the formation ofthin and unstable protein films, susceptibleto collapse during whipping.

Addition of calcium, and pH adjustmenthad various effects on foam firm-ness (fig 7). Increasing both calciumconcentration and pH increased the firm-ness of foams from MM-isolates. Theseconditions promoted the formation of cal-cium bridges leading to a more rigid mem-brane at the air-water interface. It was no-ticed that the use of sodium chloride duringdiafiltration (Iow calcium isolate) reduced


10 15 0 5 10 15 0added calcium (1'T'fflCII rI

10 15

Fig 7. Effect 01 pH and added calcium on wheyprotein loam lirmness. a, pH 4 ; b, pH 6; c, pH 8;., MM-isolate; 0, low calcium MM-isolate;e, CW-concentrate.Effet du pH et de l'ajout de calcium sur la fermetédes mousses protéiques. a, pH 4 ; b, pH 6 ; c,pH 8;., isolat de microfiltrat de lait; 0, isolatde microfiltrat de lait faible en calcium; e, con-centré protéique de lactosérum de fromage.

the firmness of foams produced at pH 8.0.Higher level of monovalent cations in theMM-isolate diafiltered with saline solu-tion (table 1)could have masked sorne re-sidues which otherwise would have partici-pated in calcium bridges. From theseresults, it seems that foam firmness is con-trolled by the combined effect of pH, mono-valent cations and available ionic calciumon the electrostatic repulsions and calciumbridge formation between protein mole-cules at the air-water interface.

Foam stability results reflected also thecombined effect of the variables understudy on foaming properties (fig 8). At lowpH, foam was stable and not affected bythe addition of calcium. As pH increased,foam stability tended to decrease, butwas improved by the addition of calcium.There was no significant effect asso-ciated with the use of sodium chloride du-ring the early stage of diafiltration(P = 0.4804).

6b c

:t+~~ t.

e

4

2

0T T T IT T T T TIl

15 1 15o 5 10 15 0added calcium (mmol r')

Fig 8. Effect 01 pH and added calcium on wheyprotein loam stability. a, pH 4; b, pH 6; c, pH 8;., MM-isolate; 0, low calcium MM-isolate;e, CW-concentrate.Effet du pH et de l'ajout de calcium sur la stabilitédes mousses protéiques. a, pH 4 ; b, pH 6 ; c,pH 8 ;., isolat de microfiltrat de lait; 0, isolatde microfiltrat de lait faible en calcium; e, con-centré protéique de lactosérum de fromage.

CONCLUSION

Concentration of proteins from milk micro-filtrate led to a high value protein isolatewith improved solubility, gelling and foa-ming properties (compared to cheesewhey). However, it has been shown that thestructure of ~-Iactoglobulin had been alte-red by the concentration conditions used inthe present study. To further improve thequality of the isolate, the filtration condi-tions should be revised in order to reduceprotein structure alteration.

The functional properties of the isolatewere strongly dependent on the ionic envi-ronment (pH, calcium content). In order todevelop the desired properties, the qualityof the ionie environment is as important asthe quality of the protein ingredient. The ba-lance between pH and calcium contentshould be controlled to ensure optimal be-haviour of whey protein isolates in complexfood systems.


ACKNOWLEDGMENTS

The authors wish ta thank S Bastrash, HJ Girouxand N Raymond for technical assistance.

REFERENCES

Brulé G, Fauquant J (1981) Mineral balance in skim-milkand milk retentate: effect of physicochemical char-acteristics of the aqueous phase. J Dairy Res 48,91-97

Cheryan M (1986) Ultrafiltration Handbook. Technomicpublishing co, Landcaster, USA

Daufin G, Michel F, Merin U (1992) Study of defattedwhey protein concentrates (WPC) withdrawn froman industrial plant. Lait72, 185-199

de Wit JN, de Boer R (1975) Ultrafiltration of cheesewhey and some functional properties of the resultingwhey protein concentrate. Neth Milk Dairy J 29,198-211

Donovan M, Mulvihill DM (1987) Thermal denaturationand aggregation of whey proteins. Ir J Food Sci Te-ehno/11,87-100

Dubois M, Gilles K, Hamilton JK, Rebers PA, SmithF (1956) Colorimetrie method for determination ofsugars and related substances. Anal Chem 28, 350-356

Fauquant J, Maubois JL, Pierre A (1988) Microfiltrationdu lait sur membrane minérale. Tech Lait Market1028,21-31

Gault P, Mahaut M, Korolczuk J (1990) Caractéristiquesrhéologiques et gélification thermique du concentréprotéique de lactosérum. Lait 70, 217-232

Gautier M, Rouault A, Méjean S, Fauquant J, MauboisJL (1994) Partition of Laetoeoceus laetis bacterio-phage during the concentration of micellar casein bytangential 0.1 urn pore size microfiltration. Lait 74,419-423

Gésan G, Merin U, Daufin G, Maugas JJ (1993) Perfor-mance of an industrial cross-flow rnicrofiltrationplant for clarifying rennet whey. Neth Milk Dairy J 47,121-135

Glover FA (1985) Ultrafiltration and reverse osmosis forthe dairy industry. The National Institute for Re-search in Dairying, Technical Bulletin 5, Reading, UK

Harris JL, Pecar MA, Pearce RJ (1989) Effect of theprocessing equipment on protein functionality in theconcentration of cheese whey by ultrafiltration. AustJ Dairy Teehno/44, 78-81

Joseph MSB, Mangino ME (1988) The effect of milk fatglobule membrane protein on the foaming and gelationproperties of beta-Iactoglobulin solutions and wheyprotein concentrates. Aust J Dairy Techno/43, 9-11

Kim VA, Chism GW, Mangino ME (1987) Determinationof the beta-Iactoglobulin, alpha-Iactalbumin and bo-vine serum albumin of whey protein concentratesand their relationship to protein functionality. J FoodSei 52, 124-127

Kinsella JE, Whitehead DM (1989) Proteins in whey:chemical, physical, and functional properties. AdvFood Nutr Res 33, 343-348

265

Mangino ME, Kim JH, Dunkerley JA, Zadow JG (1987)Factors important to the gelation of whey proteinconcentrates. Food Hydroeo/loids 1,277-282

Marshall KR (1982) Industrial isolation of milk proteins:whey proteins. In: Developments in dairy cnemis-try. 1. Proteins (Fox PF, ed) Applied Science Publish-ers, London

Maubois JL (1988) Whey, its biotechnological significa-tion. In: Proe [jh Int Bioteehnol Symp, VolumeII. (Durand G, Bobichon L, Florent J, eds) SociétéFrançaise de Microbiologie

Merin U, Gordin, S (1983) Microfiltration of sweetcheese whey. NZJ Dairy Sei Teehno/18, 153-160

Merin U, Cheryan M (1980) Factors affecting the me-chanism of flux decline during ultrafiltration of cot-tage cheese whey. J Food Proeess Preserv4, 183-198

Morr CV, Ha YW (1993) Whey protein concentrates andisolates: Processing and functional properties. CritRev Food Sei Nutr33, 431--476

Mucchetti G, Taglietti P (1993) Microbiological stabiliza-tion of whey by cross flow microfiltration. Lait 73,79-84

Pearce KN (1977) The complexometric determination ofcalcium in dairy products. NZJ Dairy Sei Teehno/12,113-115

Peri C, Pompei C, Rossi F (1973) Process optimizationin skim milk protein recovery and purification by ul-trafiltration. J Food Sei 38, 135-140

Pierre A, Fauquant J, Le Graet V, Piot M, MauboisJL (1992) Préparation de phosphocaséinate natifpar microfiltration sur membrane. Lait 72,461--474

Pierre A, Le Graet V, Daufin G, Michel F, Gésan G (1994)Whey microfiltration performance: influence of pro-tein concentration by ultrafiltration and of physico-chemical pretreatment. Lait 74, 65-77

Piot M, Maubois JL, Schaegis P, Veyre R, LuccioniA (1984) Microfiltration en flux tangentiel des lacto-sérums de fromagerie. Lait 64, 102-120

SAS (1985) User's guide: Statisties, Version 5 Edition,SAS Inst Inc, Cary, NC, USA

Rowland SJ (1938) The determination of the nitrogendistribution in milk. J Dairy Res 9, 42--46

Schmidt RH, IIlingworth BL, Deng JC, Cornell JA (1979)Multiple regression and response surface analysisof the effects of calcium chio ride and cysteine onheat-induced whey protein gelation. J Agrie FoodChem 27, 529-532

Shimada K, Chellel JC (1989) Sulfhydryl group/disul-phide bond interchange reactions during heat-indu-ced gelation of whey protein isolate. J Agrie FoodChem37,161-168

Swaisgood HE (1982) Chemistry of milk protein. In: De-velopment in dairy ehemistry-1. Proteins. (Fox PF,ed) Elsevier Applied Science Publishers, London

Trouvé E, Maubois JL, Piot M, Madec MN, Fauquant J,Rouault A, Tabard J Brinkmann G (1991) Rétentionde différentes espèces microbiennes lors de l'épu-ration du lait par microfiltration en flux tangentiel. Lait71,1-13

characterization of whey protein isolate obtained from milk

Documents