- Email: [email protected]
whey protein solution blends
Food Research International, Vol. 30, No. 5, pp. 327±334, 1997 # 1998 Canadian Institute of Food Science and Technology Published by Elsevier Science Ltd Printed in Great Britain PII: S0963-9969(97)00056-2 0963-9969/98 $19.00+0.00
Thermal stability of skim milk/whey protein solution blends W. Rattrayy & P. Jelen* Department of Agricultural, Food and Nutritional Science, University of Alberta, Edmonton, Alberta, Canada, T6G 2P5 Dried protein fractions consisting predominantly of -lactalbumin, -lactoglobulin or a whey protein concentrate (WPC) were dissolved in skim milk ultra®ltration permeate at 3.44% (w/w) protein. Each solution, denoted as -FS, -FS or WPCS, was blended with skim milk at ratios of 100/0±60/40 (g skim milk gÿ1 whey protein solution), to increase the ratio of whey to casein protein without changing the concentration of total protein. The blends were adjusted to pH values in the range 6.6±7.1 and heat stability determined by measuring the heat coagulation time at 140 C. At initial pH 6.6±6.7, blends of skim milk and -FS were slightly more heat stable than normal skim milk; gel electrophoresis indicated that thermal association between -la and casein may have augmented heat stability. From pH 6.8±7.1, the blends were considerably less heat stable than normal skim milk, possibly due to the formation of soluble complexes between -la and -casein and accompanying heat-sensitive, -casein-depleted casein micelles. In most cases, enrichment of skim milk with -lg, by blending with -FS or WPCS, caused a drastic reduction in heat stability; this was attributed to the low heat stability of -lg. An exception to this trend was the high heat stability of a 70/30 blend of skim milk and -FS or skim milk and WPCS; electrophoresis indicated that the heat stable entity was a complex between -lg and casein. # 1998 Canadian Institute of Food Science and Technology. Published by Elsevier Science Ltd Keywords: heat stability, whey protein, -lactalbumin, -lactoglobulin, skim milk/whey protein solution blends.
INTRODUCTION
addition of whey protein concentrate (WPC) to milk allows for the production of yoghurt with better texture and stability (Jelen et al., 1987; Abd El Salam et al., 1991), sour milk with a reduced viscosity (Buchheim et al., 1986; Jelen et al., 1987) or the manufacture of texturized dairy products (de Wit, 1989). Increasing the whey protein content of milk and dairy products could be justi®ed also from a nutritional viewpoint. In feeding trials with rats, McDonough et al. (1976) demonstrated that dried WPC had a greater protein eciency ratio than skim milk powder. Recently, McIntosh et al. (1995) and Wong and Watson (1995) showed that, at least in the case of laboratory animals, whey protein appeared to possess immunomodulatory properties, confering increased resistance to the growth of tumours. Because almost all dairy processes require heat treatments, information on the heat stability of milk with a modi®ed protein component is important. The heat
Modi®cation of the protein fraction of bovine milk presents new opportunities for the dairy industry. The term `protein adjustment' implies a relatively drastic alteration of the protein fraction of milk, which may or may not be accompanied by an altered ratio of whey protein to casein, and is usually carried out to impart technological bene®ts (Rattray and Jelen, 1996). Yoghurt made from ultra®ltered milk exhibits improved texture and stability (Puhan, 1992; Savello and Dargan, 1995), while ultra®ltration (UF) of milk for cheese making increases the yield of cheese (Bech, 1993). The y
In memoriam: This paper is dedicated to the memory of its principal author, William Rattray, Ph.D., who perished in a mountaineering accident on Mont Blanc in August 1997. *To whom correspondence should be addressed. Fax: 001 403 492 4265; e-mail: [email protected] 327
328
W. Rattray, P. Jelen
stability of milk which had its protein content adjusted by enrichment with -lactalbumin ( -la) was investigated by Peter et al. (1996); direct addition of -la to skim milk caused a modest decline in heat stability when measured at the unadjusted pH (6.7) of milk. Patocka et al. (1993) ultra®ltered dierent types of whey to obtain UF retentate with a protein content similar to that of skim milk. The UF retentate was mixed with skim milk to obtain dierent ratios of whey protein to casein while maintaining constant total protein. When the UF retentate was derived from sweet or ultracentrifugal whey, blends of skim milk and UF retentate had heat stabilities, measured at 90 C and pH 6.5±6.7, in excess of 30 min. In contrast, mixtures of acid whey UF retentate and skim milk, adjusted to pH 6.5±6.7, were heat labile at a ratio >30/70, due to the high calcium content of the acid whey UF retentate, resulting in calcium-modulated precipitation of whey proteins. Similar trends were found by Abd-El-Salam et al. (1991), who reported that a mixture of liquid WPC and bualo milk, at a ratio >30/70, coagulated rapidly at 80 C. The scope of these studies was limited, as heat stability was measured at a limited range of pH and temperature; and no attempt was made to determine the relative importance of speci®c milk proteins on heat stability. The aim of this study was to determine the heat stabilities of blends of skim milk and whey protein solutions; as described above, such information may be of increased relevance to the dairy industry in light of the desirable technological and nutritional properties of protein-adjusted milk. Dried industrially produced fractions of -la ( -F); -lg ( -F); or a WPC were dissolved in skim milk UF permeate and the solutions ( -FS, -FS or WPCS) blended with skim milk, to increase the ratio of whey protein to casein without changing the total protein. The pH of the blends was adjusted to values in the range 6.4±7.1 and heat stability assessed by measuring the heat coagulation time (HCT) at 140 C. MATERIALS AND METHODS Materials Fractions of -la, -lg or a WPC used in this study were all produced on an industrial scale. The -F and -F were provided by Protose Separations Inc., Canada; the manufacturing procedure was not disclosed. The WPC was `Alacen' brand No. 132, obtained from New Zealand Milk Products, Inc., USA; details of product manufacture were not provided. The -F, -F or the WPC were dissolved in skim milk UF permeate (SMP) at 3.44% (w/w) protein and each solution blended with skim milk (3.44%, w/w, protein) at ratios of 100/0±40/ 60 (g skim milk gÿ1 whey protein solution), so that the
ratio of whey protein to casein was increased but total protein was constant. To produce SMP, pasteurized (72 C for 15 s), bulk skim milk, acquired from a local dairy was ultra®ltered at 20 C and 400 kPa, for 6 h, using a pilot-scale Carbosep UF unit (SFEC Carbosep, Bollene, France), which had a tubular arrangement of zirconium oxide membranes (cut o 20 000 Da). In some experiments, the heat stabilities of solutions of -F, -F or the WPC were determined in the absence of casein, for which purpose the protein preparation was dissolved in SMP at 1.7% (w/w) protein. Analyses All raw materials were assayed for total solids, protein and ash by routine analytical methods (AOAC, 1990). The heterogeneity of the protein in the -F, -F or WPC was determined by sodium dodecyl sulphate polyacrylamide gel electrophoresis (SDS±PAGE). To this end, the protein preparations were dissolved in SMP and then mixed with SDS reducing buer (an aqueous solution of 9% SDS, 15% mercaptoethanol, 30% glycerol, 0.01% bromothymol blue, 150 mM Tris HCl at pH 6.8) so that a ®nal protein concentration of 2 mg mlÿ1 was obtained. A 12.5% polyacrylamide gel was then injected with 5 l of the protein solution and the proteins separated by application of an electric potential of 75 V for 20 min, followed by 150 V for 1 h. After separation the gel was stained with Coomassie Blue for 12 h and then destained with a water-methanol-acetic acid solution. Species of protein were readily identi®ed on the basis that, in the presence of SDS, electrophoretic motility would be proportional to molecular weight. Heat stability Prior to measuring the heat stabilities of solutions of whey protein preparations or blends of whey protein solutions and skim milk, the pH was adjusted, at 20 C, to values in the range 6.4±7.1, using 0.5 N lactic acid or 0.5 N NaOH. The samples were stored overnight at 4 C, equilibrated to 20 C, and the pH remeasured and readjusted if necessary. Immediately after the ®nal pH adjustment, the heat stability was measured as the heat coagulation time (HCT) at temperatures of 96±140 C by the method of Davies and White (1966). This standard heat-coagulation test uses a temperature of 140 C as otherwise the normally high heat stability of milk would necessitate extraordinarily long assay times; in the present conditions the use of the 140 C was also advantageous because the ability of a protein-adjusted milk product to withstand prolonged heating at this temperature implies that it would be stable in almost all industrial heating conditions such as UHT processing or in-can sterilization. All HCT values were determined in triplicate or quadruplicate, from separately prepared
Thermal stability of skim milk/whey protein solution blends protein solutions, and mean values calculated from all data. The behaviour of speci®c proteins during heating of the skim milk/whey protein solution blends was investigated further using SDS±PAGE. An aliquot (10 ml) of required sample was placed in a glass tube (i.d. 12 mm), the tube sealed and immersed in a 120 C oil bath for 5 min; this relatively mild heat treatment was used in an attempt to simulate conditions during the early stages of the heat coagulation process and heat-aggregate the most labile proteins. The sample was cooled quickly to 20 C and centrifuged at 100,000 g and 20 C for 1 h, using a Beckman Ultracentrifuge (model L8-70M, Beckman Instruments Inc., Palo Alto, CA). The purpose of centrifugation was to sediment the most heat sensitive whey proteins, which would be expected to heat-aggregate independently and/or complexed to casein micelles. A portion (10 0.5 mg) of ultracentrifugal (UC) pellet was dissolved in 1 ml of SDS reducing buer, and 5 l of this solution applied to a 12.5% SDS gel and the proteins separated and identi®ed using the same procedure as described above. RESULTS AND DISCUSSION Composition of raw materials Compositional data con®rmed that the SMP was a dilute solution of lactose and minerals (Table 1). The lactose content was similar to that of skim milk, but the mineral content was much lower, because, on a mass basis, about 2/3 of the total minerals, especially calcium and phosphorus, in milk are associated with casein and thus would be retained during UF (Green et al., 1984). The concentration of nitrogen in the SMP was low and probably consisted of non-protein nitrogen, namely, low molecular weight peptides, urea and other nitrogenous compounds, capable of permeating through UF membranes. As expected, the -F and -F were high in protein and low in lactose and ash. Electrophoretograms (results not shown) revealed that -la or -lg were the most abundant proteins in the -F or -F, respectively; the -F was more heterogeneous, containing also noticeable quantities of -lg and bovine serum albumin (BSA). The WPC had a relatively high amount of non-protein Table 1. Average (n = 2) composition (%, w/w) of materials used to prepare protein-adjusted milk Product Skim milk permeate -Fraction -Fraction Whey protein concentrate a
% Protein = n 6.38.
Moisture Proteina Lactose Ash 94.8 4.1 5.2 4.2
0.31 89.9 89.0 78.8
4.4 1.0 1.0 11.8
0.36 2.95 2.81 3.89
329
constituents and its protein component was heterogeneous. When the -F, -F or WPC were dissolved in SMP at 3.44% protein, the ®nal pH values were in the range 6.6±6.8 and changes in the pH of skim milk (6.7) were very small ( 0.05 pH units) upon admixture of any whey protein solution at any ratio. Heat stability of solutions of -fraction, -fraction or WPC To permit a better understanding of the thermal behaviour of proteins in the blends of skim milk with whey protein solution, the heat stability of individual whey protein preparations in SMP was measured at 96± 140 C. The -FS was extremely heat stable (Fig. 1), coagulating only at high temperatures and low pH values (140 C and < pH 6.7; 120 C and < pH 6.5). This was in contrast to the behaviour of the -FS or WPCS, which coagulated quickly at 120 C and almost instantaneously at 140 C; these solutions were resistant to coagulation only at a relatively low temperature (96 C) and high pH values (6.7±6.9). The heat stability results indicated that the thermal behaviour of the -FS was consistent with the known resistance of -la to heat coagulation (Larsen and Rolleri, 1955; Matsudomi et al., 1992), while that of the -F or WPC was dominated by -lg, documented as being sensitive to heat precipitation (de Wit, 1981; Matsudomi et al., 1992). Near neutral pH, -lg denatures at 75±80 C, while -la, in the presence of calcium, denatures at 60±65 C (Ruegg et al., 1977; de Wit et al., 1983; Bernal and Jelen,
Fig. 1. Mean (n = 3) heat coagulation time (HCT) of pHadjusted whey protein solutions, made by dissolving -fraction, -fraction or whey protein concentrate (WPC) in skim milk ultra®ltration permeate to 1.7% (w/w) protein; -fraction solution at 120 (*) or 140 C (*); -fraction solution at 96 (&) or 120 C (&); WPC solution at 96 (~) or 120 C (~). Lines exiting the graph indicate heat stabilities >50 min. At 140 C and all pH values, -fraction and WPC solutions had a heat stability of 0 min; at 96 C and all pH values, -fraction solution had a heat stability >50 min.
330
W. Rattray, P. Jelen
1984; Paulsson et al., 1985; Paulsson and Dejmek, 1990; Xiong et al., 1993; Hollar et al., 1995; Anema and McKenna, 1996); in the present study both -la and -lg would be expected to exhibit unfolded conformations at the high temperatures of 96±140 C. Anema and McKenna (1996) observed that at temperatures > 85 C for -la or >100 C for -lg, the proteins underwent aggregation reactions, evident by a reduction in entropy. Thus, in the heating conditions used in the present study, it can be concluded that in ionic environment of SMP, unfolded -lg was prone to rapid and extensive intermolecular interaction, leading to the formation of visible protein clots, whereas aggregation between unfolded -la molecules was much more limited such that visible protein clots did not form. The latter observation is consistent with the work of Chaplin and Lyster (1986), who showed that prolonged heating (100 C for up to 30 min) of a-la, in 0.1 M phosphate buer at pH 7.0, led to irreversible denaturation and polymerization, but the extent of polymerization was limited, allowing -la to remain soluble. The limited propensity of -la for thermal association was also demonstrated by Paulsson et al. (1986), who could not gel a 20% solution of -la, upon heating at 90 C and pH 6.6, in contrast to -lg which gelled at 5% protein. Hines and Foegeding (1993) observed a similar trend; when a 7% -la solution, at pH 7, in the presence of 100 mM NaCl, was heated at 80 C for 3 h, protein aggregation was extremely slow resulting in the formation of a very weak gel. The resistance of -la to thermal coagulation can sometimes be attributed to its calcium-assisted renaturation on cooling (Ruegg et al., 1977; Bernal and Jelen, 1984; Patocka et al., 1987), but continuous heating at 140 C would not allow renaturation. Vanderheeren and Hanssens (1994) showed that binding of the hydrophobic probe, bis-ANS (1,10 -bi [4anilo] naphthalene-5,50 -disulphonate), to apo- -la at pH 7.5 was maximal at 25 C; two strong binding sites were identi®ed. At 25 C, the protein probably existed in the molten globule state, corresponding to a loosening of tertiary structure with preservation of secondary structures, which allowed bis-ANS to more easily penetrate the hydrophobic interior of the molecule. At 80 C, apo- -la exhibited an unfolded conformation and the accessibility of hydrophobic groups to bis-ANS increased further; however, the loss of tertiary structure caused dissipation of hydrophobic clusters and only one weak binding site for bis-ANS was identi®ed. Similar trends occurred when -la was heated in the presence of calcium, although the eects took place at higher temperatures, because of stabilization of protein conformation by calcium. It has not been proven whether the reduced hydrophobic interaction between -la and bisANS at high temperature could be extrapolated to imply that relatively weak hydrophobic interactions would occur between unfolded proteins, but if this was
the case, it would, at least in part, explain the high resistance of -la to thermal precipitation. It is also possible that the limited ability of -la for self-association, compared to -lg, was related to a lower rate of sulphydryl group activity in -la. Considering that -lg contains a free sulphydryl group whereas -la does not, thiol group interactions between denatured -la molecules can be expected to be less prevalent than in the case of -lg. Calvo et al. (1993) showed that when -la and casein micelles were heated together at 90 C for 24 min, -la did not aggregate with itself or with casein micelles. However, the presence of -lg promoted aggregation of -la. These authors proposed that -lg could act as a catalyst, via its free sulphydryl group, which interacted with the sulphydryl bridges of -la, thus allowing -la to aggregate with itself, -lg or -casein. Similarly, Matsudomi et al. (1992) showed that a synergistic interaction between -la and -lg, in the formation of heat-induced gels, was due predominantly to thiol-disulphide interchange reactions between the two protein types during heating. At increased pH, it is likely that -la and -lg were denatured to a greater extent, as repulsive intramolecular electrostatic forces would be expected to increase. However, the increased heat stabilities of the -FS, -FS or WPCS, with increasing pH, indicate that intermolecular association between denatured protein molecules was also inhibited by increased electrostatic repulsion. Possibly, longer heating times were required to accentuate hydrophobic and thiol group interactions involved in protein aggregation or more extensive unfolding of protein was required to initiate association. The heat stability of the WPCS was lower than that of the -FS, probably due to its higher content of minerals; small (0±200 mM) changes in the concentration of calcium have a great eect on the thermal aggregation of -lg, due to the ability of calcium to form cross-bridges between proteins (Xiong et al., 1993). In succeeding experiments, the contrasting heat stabilities between -FS and -FS or WPCS were evidenced by the drastically dierent eects of these solutions on the heat stability of skim milk. Heat stability of blends of skim milk and -fraction solution In general, at 140 C and initial pH 6.6±6.7, mixtures of skim milk and -FS were more heat stable than skim milk alone, but considerably less heat stable when the initial pH was in the range 6.8±7.1 (Fig. 2). An exception to this trend was the very low heat stability, at pH 6.6, of a 40/60 combination of skim milk and -FS; this was probably due to the presence of the heat-labile -lg in the -FS. Skim milk itself showed a type A HCT/ pH pro®le, thought to be due to a pH-dependent interaction between serum proteins and -casein (Kudo, 1980; Singh and Fox, 1985, 1986, 1987a,b).
Thermal stability of skim milk/whey protein solution blends
Fig. 2. Mean (n = 3) heat coagulation time (HCT) of pHadjusted blends of skim milk and -fraction solution. The ratio of skim milk to -fraction solution was 100/0(- - -), 95/5 (*), 90/10 (*), 70/30 (&), 50/50 (&) or 40/60 (~).
The nature of thermal interactions between serum proteins and casein was investigated further by SDSPAGE of the UC pellets of heated (120 C 5 minÿ1) skim milk/ -FS blends. At pH 6.6, the increased ratio of -FS to skim milk led to greater amounts of -la, -lg and BSA in the UC pellets (Fig. 3, columns 1±3). The results re¯ected the composition of the -F which consisted primarily of -la but contained also smaller amounts of -lg and BSA and implied that increasing the amount of these proteins in the blends enhanced their thermal association with casein, especially for -la, the most plentiful protein. The whey proteins must have been complexed to the casein in order to be sedimented at 100,000 g for 1 h, because, unlike casein, their sedimentation coecients would be too low to permit
Fig. 3. Electrophoretic pro®le of ultracentrifugal pellets obtained from blends of skim milk and -fraction solution which had been pH adjusted, heat at 120 C for 10 min and cooled at 20 C. At pH>6.6, the ratio of skim milk to -fraction solution was (1) 100/0, (2) 60/40 or (3) 40/60; at pH 6.8 it was (4) 100/0 or (5) 50/50; at pH 7.1 it was (6) 100/0 or (7) 50/50.
331
exclusive sedimentation. Therefore, in skim milk, -la co-aggregated with the casein micelles, as opposed to a much more limited tendency for self-aggregation when dissolved in SMP (Fig. 1). Law et al. (1994) also showed a complexation reaction between casein and whey proteins; heating milk at natural pH and 80±140 C for 5 min caused extensive complexation of -la, -lg and bovine serum albumin with casein, allowing these proteins to co-precipitate with casein when the cooled milk was adjusted to pH 4.6. As described above, the work of Calvo et al. (1993) may indicate that the natural presence of -lg in skim milk is necessary for -la to react with casein micelles. However, the heating conditions used in the present study (140 C) were much more severe than that used by the above authors (90 C), obtruding a direct comparison between the two studies. At initial pH 6.6 or 6.7, the high heat stability of the blends of skim milk and -FS may have been due to increased complexation of -la onto casein micelles, increasing micellar charge and hydration and thus heat stability, as proposed by Fox and Hearn (1978). The presence of a small amount of -lg in the -F may have caused the decline in heat stability at pH 6.6, at the lowest ratio of skim milk to -FS (40/60), consistent with the increased concentration of -lg in the UC pellet of the heated mixture (Fig. 3, column 3). This seems likely considering that at 140 C and pH 6.6 the -FS had a HCT of 11 min, in contrast to the -FS which coagulated almost instantly (Fig. 1). Furthermore, skim milk with low levels of added -FS was also very unstable due to independent coagulation of -lg and/or -lg complexed to casein, as will be discussed below. An eect similar to the present results was reported by Peter et al. (1996) who observed that the addition of increasing amounts of a dried fraction of -la to milk, at pH 6.7, caused a gradual decline in heat stability. At initial pH 6.8 or 7.1, UC pellets from heated blends of skim milk and -FS did not contain increased quantities of any whey protein (Fig. 3, columns 4±7), indicating that most of the added whey protein component remained soluble during heating and the low heat stabilities were caused by coagulation of casein alone; yet the sensitivity of casein to coagulation was increased in the presence of extra whey proteins. The reduced heat stability of the caseins in the 50/50 blends is also evidenced by the increased amounts of casein in the UC pellets (Fig. 3, columns 5 and 7) despite the fact that the blends contained lower amounts of casein than in skim milk. The low heat stability of type A milk near pH 6.9 may be due to the formation of soluble whey protein/casein complexes and concomitant heat labile -caseindepleted micelles (Kudo, 1980; Singh and Fox, 1985, 1986, 1987a,b); possibly, in the skim milk/ -FS blends, the rate of this destabilizing interaction would be increased because of an increased concentration of -la and/or -lg.
332
W. Rattray, P. Jelen
Heat stability of blends of skim milk and -fraction solution Skim milk in combination with -FS had a substantially dierent heat stability pattern, compared to skim milk mixed with -FS; this was expected in view of the starkly dierent thermal properties of the -FS and -FS (Fig. 1). In most cases, blends of skim milk and FS showed a greatly reduced heat stability, compared to normal skim milk (Fig. 4). For all combinations of skim milk and -FS, the zone of minimal heat stability deepened and was extended to more alkaline pH values. At initial pH 6.7, low levels of added -FS caused a modest increase in heat stability, but when the ratio of -FS to skim milk was >10/90, heat stability declined markedly. At a 70/30 ratio of skim milk to -FS and at pH 6.8, the HCT increased; it appeared that the entire HCT-pH curve was shifted to more alkaline pH values. At all pH values, 60/40 mixtures were extremely unstable; a maximum heat stability of only about 5 min was recorded when the initial pH of the blends was 6.8. Gel electrophoresis of the UC pellets of heated (120 C for 5 min) 50/50 blends revealed that large quantities of -lg, as well as casein, were present (Fig. 5), suggesting that the heat sensitivity of the mixtures was caused by coagulation of -lg alone and/or of -lg which had heat-aggregated with casein micelles. Unlike -la, -lg was unstable in SMP at 120 and 140 C. Accordingly, the low heat stability of the skim milk / FS combinations at 140 C could have resulted from exclusive coagulation of -lg or -lg/casein micelle complexes or both, all of which would sediment by ultracentrifugation. Compared to mixtures of skim milk with -FS at pH 6.6 (Fig. 2, columns 1±3), skim milk with -FS showed much greater quantities of -lg in the UC pellets, attesting that denatured -lg was much
more prone than denatured -la to intermolecular interaction with itself and/or casein. This concurs with Mottar et al. (1989), who reported that a more severe heating of milk was required to induce -la association onto casein micelles, as compared to -lg. In the absence of casein, at pH 6.8, the -FS was heat sensitive (Fig. 1), while unaltered skim milk also displayed a relatively low heat stability at pH 6.8. This suggests that in the 70/30 blends, at pH 6.8, the heat stable entity must have been a -lg/casein micelle complex, which the SDS±PAGE results also indicate (Fig. 5, column 5). The result indicates that a shift of the heat stability/pH curve to more alkaline pH values occurred. When normal skim milk is heated at pH 6.7, -lg heatprecipitates onto the surface of casein micelles to increase their heat stability (Kudo, 1980; Singh and Fox, 1985, 1986, 1987a,b); the peak heat stability of the 70/30 mixture at pH 6.8 indicates that in the presence of extra -lg, a favourable degree of interaction between -lg and casein occurred at a higher pH value. The slightly improved heat stability, at pH 6.7, of 95/5 or 90/ 10 combinations, also may have been due to thermal complexation between -lg and casein micelles. At > pH 6.9, all blends of skim milk and -FS had low heat stabilities, possibly because of a more rapid formation of soluble -casein/ -lg complexes and hence heat labile, -casein depleted micelles and/or the sole aggregation of -lg. Patocka et al. (1993) showed that combinations of skim milk and acid whey UF retentate, at pH 6.5±6.7, coagulated at 90 C; the high calcium content of the UF retentate promoted formation of casein/whey protein co-precipitates. The same authors found that mixtures of skim milk and sweet or UC whey UF retentate were heat stable at 90 C and pH 6.5±6.7, attributed to lower levels of calcium in the UF retentate. The blends of skim
Fig. 4. Mean (n = 3) heat coagulation time (HCT) of pHadjusted blends of skim milk and -fraction solution. The ratio of skim milk to -fraction solution was 100/0 (- - -), 95/5 (*), 90/10 (*), 70/30 (&) or 60/40 (&).
Fig. 5. Electrophoretic pro®le or ultracentrifugal pellets obtained from blends of skim milk and -fraction solution, which had been pH adjusted, heated at 120 C for 10 min and cooled to 20 C. At pH 6.7, the ratio of skim milk to -fraction solution was (1) 100/0 or (2) 50/50; at pH 6.8 it was (3) 100/0, (4) 90/10, (5) 70/30 or (6) 50/50; at pH 7.1 it was (7) 100/0 or (8) 50/50.
Thermal stability of skim milk/whey protein solution blends milk and -FS, used in the current study, would be expected to be similar in composition to blends of skim milk and sweet or UC whey used by Patocka et al. (1993). The results imply that, without elevated calcium levels, the presence of extra -lg was also detrimental to the heat stability of skim milk, the eect occurring over a range of pH values, albeit at a much higher temperature (140 C). Heat stability of blends of skim milk and WPC solution Combinations of skim milk and WPCS had HCT-pH pro®les (Fig. 6) broadly similar to skim milk/ -FS blends, indicating that heat stability was in¯uenced mainly by an increased concentration of -lg, the most abundant protein in the WPC. In most cases, blends of WPCS and skim milk were much less heat stable than the skim milk alone, a conspicuous exception being a pronounced increase at pH 6.8 and 70/30 ratio of skim milk to WPCS. The results indicated that the HCT/pH pro®le was shifted to more alkaline pH values, as also occurred with a 70/30 ratio of skim milk to -FS. Gel electrophoresis of UC pellets obtained from heated samples (results not shown) gave patterns very similar to those observed for the combinations of skim milk and -FS (Fig. 5); this implies that the heat stability pattern was again due to the presence of extra -lg and the heat instability may have been caused by aggregation of -lg alone and/or -lg in association with casein. Similarly, the SDS± PAGE results indicated that the high thermal stability of a 70/30 blend of skim milk and WPCS at pH 6.8 could have been due to the formation of a complex between -lg and -casein. Studies have shown that whey protein products rich in -lg have similar thermal properties to more puri®ed
333
-lg preparations (de Wit, 1981; Hines and Foegeding, 1993), so it should not be surprising that the overall heat stability and gel electrophoretic patterns of skim milk combined with -FS or WPCS were similar. However, the actual HCT values of skim milk/WPC solution combinations were higher, suggesting an in¯uence of other proteins and/or non-protein constituents in the WPC, the nature of which was not established. CONCLUSIONS The thermal behaviour of industrial whey protein preparations was dominated by their most abundant protein constituent; -la in the -F; and -lg in the -F or the WPC. Con®rming the well known dierences in the heat stabilities of -la and -lg, the -FS was resistant to coagulation in SMP at temperatures up to 140 C, whereas -lg quickly formed visible protein clots at 140 C. This was probably due to a greater degree of thermal unfolding for -lg as compared to -la, implying that the overall heat stability of -la, in terms of conformational changes and aggregation, is greater than that of -lg. The heat stabilities of blends of skim milk and whey protein solutions were dependent on the type of whey protein preparation used and the initial pH. At the unaltered pH (6.7) of the mixtures, -la had a bene®cial eect on heat stability up to at least a 40/60 ratio of skim milk to -FS, whereas -lg caused a small increase in heat stability at < 90/10 ratio but was deleterious at higher levels of addition. In general, at pH values > 6.7, extra -la or especially -lg caused the heat stability to decline; pronounced exceptions were the high heat stabilities at pH 6.8 of 70/30 combinations of skim milk and -FS or WPCS. REFERENCES
Fig. 6. Mean (n = 3) heat coagulation time (HCT) of pHadjusted blends of skim milk and whey protein concentrate solution. The ratio of skim milk to whey protein concentrate solution was 100/0 (- - -), 95/5 (*), 90/10 (*), 70/30 (&) or 60/40 (&).
Abd El Salam, M. H., El-Shibny, S., Mahfouz, M. B., El-Dein, H. F., El-Atriby, H. M. and Antila, V. (1991) Preparation of whey protein concentrate from salted whey and its use in yogurt. J. Dairy Res. 58, 503±510. Anema, S. G. and McKenna, A. B. (1996) Reaction kinetics of thermal denaturation of whey proteins in heated reconstituted whole milk. J. Agric. Food Chem. 44, 422±428. AOAC (1990) Ocial Methods of Analysis, 13th edn. Association of Ocial Analytical Chemists, Arlington, VA, pp. 807±808, 834±835. Bech, A. (1993) Characterising ripening in ultra®ltered cheese. Int. Dairy J. 3, 329±342. Bernal, V. and Jelen, P. (1984) Eect of calcium binding on thermal denaturation of bovine -lactalbumin. J. Dairy Sci. 67, 2452±2454. Buchheim, V. W., Peters, K. H. and Kaufmann, W. (1986) Technological, physico-chemical and sensory properties of milk and sour milk with a modi®ed casein/whey protein ratio. Milchwissenschaft 41, 139±141.
334
W. Rattray, P. Jelen
Calvo, M. M., Leaver, J. and Banks, J. M. (1993) In¯uence of other whey proteins on the heat-induced aggregation of -lactalbumin. Int. Dairy J. 3, 719±727. Chaplin, L. C. and Lyster, R. L. J. (1986) Irreversible heat denaturation of bovine -lactalbumin. J. Dairy Res. 53, 249±258. Davies, D. T. and White, J. C. D. (1966) The stability of milk protein to heat. 1. Subjective measurement of heat stability of milk. J. Dairy Res. 33, 67±81. de Wit, J. N. (1981) Structure and functional behaviour of whey proteins. Neth. Milk Dairy J. 35, 47±64. de Wit, J. N. (1989). The use of whey protein products. In Developments in Dairy Chemistry. 4. Functional Milk Proteins. P. F. Fox, ed. pp. 323±346. Applied Science, London. de Wit, J. N., Klarenbeek, G. and Hontelez-Backx, E. (1983) Evaluation of functional properties of whey protein concentrates and whey protein isolates. I. Isolation and characterization. Neth. Milk Dairy J. 37, 37±49. Fox, P. F. and Hearn, C. M. (1978) Heat stability of milk: in¯uence of denaturable proteins and detergents on pH sensitivity. J. Dairy Res. 45, 159±172. Green, M. L., Scott, J., Anderson, M., Grin, M. A. C. and Glover, F. A. (1984) Chemical characterization of milk concentrated by ultra®ltration. J. Dairy Res. 51, 267±273. Hines, M. E. and Foegeding, E. A. (1993) Interactions of -lactalbumin and bovine serum albumin with -lactoglobulin in thermally induced gelation. J. Agric. Food Chem. 41, 341±346. Hollar, C. M., Parris, N., Hsieh, A. and Cockley, K. D. (1995) Factors aecting the denaturation and aggregation of whey proteins in whey protein concentrate mixtures. J. Dairy Sci. 78, 260±267. Jelen, P., Buchheim, W. and Peters, K. H. (1987) Heat stability and use of milk with modi®ed casein/whey protein content. Milchwissenschaft 42, 418±421. Kudo, S. (1980) The heat stability of milk: formation of soluble proteins and protein-depleted micelles at elevated temperatures. N. Z. J. Dairy Sci. Technol. 15, 255±263. Larsen, B. L. and Rolleri, G. D. (1955) Heat denaturation of the speci®c serum proteins in milk. J. Dairy Sci. 38, 351±360. Law, A. J. R., Horne, D. S. and Leaver, J. (1994) Heatinduced changes in the whey proteins and caseins. Milchwissenschaft 49, 125±129. Matsudomi, N., Oshita, T., Sasaki, E. and Kobayash, K. (1992) Enhanced heat-induced gelation of -lactoglobulin by -lactalbumin. Bioscience, Biotechnology and Biochemistry 56, 1697±1700. McIntosh, G. H., Regester, G. O., LeLeu, R. K., Royle, P. J. and Smithers, G. W. (1995) Dairy products protect against dimethylhydrazine-induced intestinal cancers in rats. J. Nutr. 125, 809±816. Mottar, J., Bassier, A., Joniau, M. and Baert, J. (1989) Eects of heat-induced association of whey proteins and casein micelles on yogurt texture. J. Dairy Sci. 72, 2247±2256. McDonough, F. E., Alford, J. A. and Womack, M. (1976) Whey protein concentrate as a milk extender. J. Dairy Sci. 59, 34±38.
Patocka, J., Drathen, M. and Jelen, P. (1987) Heat stability of isolated whey protein fractions in highly acidic conditions. Milchwissenschaft 42, 700±703. Patocka, G., Jelen, P. and Kalab, M. (1993) Thermostability of skimmilk with modi®ed casein/whey protein content. Int. Dairy J. 3, 35±48. Paulsson, M. and Dejmek, P. (1990) Thermal denaturation of whey proteins in mixtures with casein studied by dierential scanning calorimetry. J. Dairy Sci. 73, 590±600. Paulsson, M., Hegg, P. O. and Castberg, H. B. (1985) Thermal stability of whey proteins studied by dierential scanning calorimetry. Thermochim. Acta 95, 435±440. Paulsson, M., Hegg, P. and Castberg, H. B. (1986) Heatinduced gelation of individual whey proteins. A dynamic rheological study. J. Food Sci. 52, 87±90. Peter, S., Rattray, W. and Jelen, P. (1996) Heat stability and sensory quality of protein-standardized 2% fat milk. Milchwissenschaft 51, 611±616. Puhan, Z. (1992) Standardization of milk protein by membrane processes for product manufacture. In New Applications of Membrane Processes, pp. 23±32. IDF special issue No. 9201, Brussels, Belgium. Rattray, W. and Jelen, P. (1996) Protein standardization of milk and dairy products. Trends Food Sci. Technol. 7, 227± 234. Ruegg, M., Morr, U. and Blanc, B. (1977) A calorimetric study of the thermal denaturation of whey proteins in simulated milk ultra®ltrate. J. Dairy Res. 44, 509±520. Savello, P. A. and Dargan, R. A. (1995) Improved yoghurt physical properties using ultra®ltration and very-high temperature heating. Milchwissenschaft 50, 86±90. Singh, H. and Fox, P. F. (1985) Heat stability of milk: pH-dependent dissociation of micellar -casein on heating milk at ultra high temperature. J. Dairy Res. 52, 529± 538. Singh, H. and Fox, P. F. (1986) Heat stability of milk: further studies on the pH-dependent dissociation of micellar -casein. J. Dairy Res. 53, 237±248. Singh, H. and Fox, P. F. (1987a) Heat stability of milk: role of -lactoglobulin in the pH-dependent dissociation of micellar -casein. J. Dairy Res., 54, 509±521. Singh, H. and Fox, P. F. (1987b) Heat stability of milk: in¯uence of colloidal and soluble salts and protein modi®cation on the pH-dependent dissociation of micellar -casein. J. Dairy Res. 54, 523±534. Xiong, Y. L., Dawson, K. A. and Wan, L. (1993) Thermal aggregation of -lactoglobulin: eect of pH, ionic environment, and thiol reagent. J. Dairy Sci. 76, 70±77. Vanderheeren, G. and Hanssens, I. (1994) Thermal unfolding of bovine -lactalbumin. Comparison of circular dichroism with hydrophobicity measurements. J. Biol. Chem. 269, 7090±7094. Wong, C. W. and Watson, D. L. (1995) Immunomodulatory eects of dietary whey proteins in mice. J. Dairy Res. 62, 359±368.
(Received 7 February 1997; accepted 24 August 1997)
Thermal stability of skim milk/whey protein solution blends W. Rattrayy & P. Jelen* Department of Agricultural, Food and Nutritional Science, University of Alberta, Edmonton, Alberta, Canada, T6G 2P5 Dried protein fractions consisting predominantly of -lactalbumin, -lactoglobulin or a whey protein concentrate (WPC) were dissolved in skim milk ultra®ltration permeate at 3.44% (w/w) protein. Each solution, denoted as -FS, -FS or WPCS, was blended with skim milk at ratios of 100/0±60/40 (g skim milk gÿ1 whey protein solution), to increase the ratio of whey to casein protein without changing the concentration of total protein. The blends were adjusted to pH values in the range 6.6±7.1 and heat stability determined by measuring the heat coagulation time at 140 C. At initial pH 6.6±6.7, blends of skim milk and -FS were slightly more heat stable than normal skim milk; gel electrophoresis indicated that thermal association between -la and casein may have augmented heat stability. From pH 6.8±7.1, the blends were considerably less heat stable than normal skim milk, possibly due to the formation of soluble complexes between -la and -casein and accompanying heat-sensitive, -casein-depleted casein micelles. In most cases, enrichment of skim milk with -lg, by blending with -FS or WPCS, caused a drastic reduction in heat stability; this was attributed to the low heat stability of -lg. An exception to this trend was the high heat stability of a 70/30 blend of skim milk and -FS or skim milk and WPCS; electrophoresis indicated that the heat stable entity was a complex between -lg and casein. # 1998 Canadian Institute of Food Science and Technology. Published by Elsevier Science Ltd Keywords: heat stability, whey protein, -lactalbumin, -lactoglobulin, skim milk/whey protein solution blends.
INTRODUCTION
addition of whey protein concentrate (WPC) to milk allows for the production of yoghurt with better texture and stability (Jelen et al., 1987; Abd El Salam et al., 1991), sour milk with a reduced viscosity (Buchheim et al., 1986; Jelen et al., 1987) or the manufacture of texturized dairy products (de Wit, 1989). Increasing the whey protein content of milk and dairy products could be justi®ed also from a nutritional viewpoint. In feeding trials with rats, McDonough et al. (1976) demonstrated that dried WPC had a greater protein eciency ratio than skim milk powder. Recently, McIntosh et al. (1995) and Wong and Watson (1995) showed that, at least in the case of laboratory animals, whey protein appeared to possess immunomodulatory properties, confering increased resistance to the growth of tumours. Because almost all dairy processes require heat treatments, information on the heat stability of milk with a modi®ed protein component is important. The heat
Modi®cation of the protein fraction of bovine milk presents new opportunities for the dairy industry. The term `protein adjustment' implies a relatively drastic alteration of the protein fraction of milk, which may or may not be accompanied by an altered ratio of whey protein to casein, and is usually carried out to impart technological bene®ts (Rattray and Jelen, 1996). Yoghurt made from ultra®ltered milk exhibits improved texture and stability (Puhan, 1992; Savello and Dargan, 1995), while ultra®ltration (UF) of milk for cheese making increases the yield of cheese (Bech, 1993). The y
In memoriam: This paper is dedicated to the memory of its principal author, William Rattray, Ph.D., who perished in a mountaineering accident on Mont Blanc in August 1997. *To whom correspondence should be addressed. Fax: 001 403 492 4265; e-mail: [email protected] 327
328
W. Rattray, P. Jelen
stability of milk which had its protein content adjusted by enrichment with -lactalbumin ( -la) was investigated by Peter et al. (1996); direct addition of -la to skim milk caused a modest decline in heat stability when measured at the unadjusted pH (6.7) of milk. Patocka et al. (1993) ultra®ltered dierent types of whey to obtain UF retentate with a protein content similar to that of skim milk. The UF retentate was mixed with skim milk to obtain dierent ratios of whey protein to casein while maintaining constant total protein. When the UF retentate was derived from sweet or ultracentrifugal whey, blends of skim milk and UF retentate had heat stabilities, measured at 90 C and pH 6.5±6.7, in excess of 30 min. In contrast, mixtures of acid whey UF retentate and skim milk, adjusted to pH 6.5±6.7, were heat labile at a ratio >30/70, due to the high calcium content of the acid whey UF retentate, resulting in calcium-modulated precipitation of whey proteins. Similar trends were found by Abd-El-Salam et al. (1991), who reported that a mixture of liquid WPC and bualo milk, at a ratio >30/70, coagulated rapidly at 80 C. The scope of these studies was limited, as heat stability was measured at a limited range of pH and temperature; and no attempt was made to determine the relative importance of speci®c milk proteins on heat stability. The aim of this study was to determine the heat stabilities of blends of skim milk and whey protein solutions; as described above, such information may be of increased relevance to the dairy industry in light of the desirable technological and nutritional properties of protein-adjusted milk. Dried industrially produced fractions of -la ( -F); -lg ( -F); or a WPC were dissolved in skim milk UF permeate and the solutions ( -FS, -FS or WPCS) blended with skim milk, to increase the ratio of whey protein to casein without changing the total protein. The pH of the blends was adjusted to values in the range 6.4±7.1 and heat stability assessed by measuring the heat coagulation time (HCT) at 140 C. MATERIALS AND METHODS Materials Fractions of -la, -lg or a WPC used in this study were all produced on an industrial scale. The -F and -F were provided by Protose Separations Inc., Canada; the manufacturing procedure was not disclosed. The WPC was `Alacen' brand No. 132, obtained from New Zealand Milk Products, Inc., USA; details of product manufacture were not provided. The -F, -F or the WPC were dissolved in skim milk UF permeate (SMP) at 3.44% (w/w) protein and each solution blended with skim milk (3.44%, w/w, protein) at ratios of 100/0±40/ 60 (g skim milk gÿ1 whey protein solution), so that the
ratio of whey protein to casein was increased but total protein was constant. To produce SMP, pasteurized (72 C for 15 s), bulk skim milk, acquired from a local dairy was ultra®ltered at 20 C and 400 kPa, for 6 h, using a pilot-scale Carbosep UF unit (SFEC Carbosep, Bollene, France), which had a tubular arrangement of zirconium oxide membranes (cut o 20 000 Da). In some experiments, the heat stabilities of solutions of -F, -F or the WPC were determined in the absence of casein, for which purpose the protein preparation was dissolved in SMP at 1.7% (w/w) protein. Analyses All raw materials were assayed for total solids, protein and ash by routine analytical methods (AOAC, 1990). The heterogeneity of the protein in the -F, -F or WPC was determined by sodium dodecyl sulphate polyacrylamide gel electrophoresis (SDS±PAGE). To this end, the protein preparations were dissolved in SMP and then mixed with SDS reducing buer (an aqueous solution of 9% SDS, 15% mercaptoethanol, 30% glycerol, 0.01% bromothymol blue, 150 mM Tris HCl at pH 6.8) so that a ®nal protein concentration of 2 mg mlÿ1 was obtained. A 12.5% polyacrylamide gel was then injected with 5 l of the protein solution and the proteins separated by application of an electric potential of 75 V for 20 min, followed by 150 V for 1 h. After separation the gel was stained with Coomassie Blue for 12 h and then destained with a water-methanol-acetic acid solution. Species of protein were readily identi®ed on the basis that, in the presence of SDS, electrophoretic motility would be proportional to molecular weight. Heat stability Prior to measuring the heat stabilities of solutions of whey protein preparations or blends of whey protein solutions and skim milk, the pH was adjusted, at 20 C, to values in the range 6.4±7.1, using 0.5 N lactic acid or 0.5 N NaOH. The samples were stored overnight at 4 C, equilibrated to 20 C, and the pH remeasured and readjusted if necessary. Immediately after the ®nal pH adjustment, the heat stability was measured as the heat coagulation time (HCT) at temperatures of 96±140 C by the method of Davies and White (1966). This standard heat-coagulation test uses a temperature of 140 C as otherwise the normally high heat stability of milk would necessitate extraordinarily long assay times; in the present conditions the use of the 140 C was also advantageous because the ability of a protein-adjusted milk product to withstand prolonged heating at this temperature implies that it would be stable in almost all industrial heating conditions such as UHT processing or in-can sterilization. All HCT values were determined in triplicate or quadruplicate, from separately prepared
Thermal stability of skim milk/whey protein solution blends protein solutions, and mean values calculated from all data. The behaviour of speci®c proteins during heating of the skim milk/whey protein solution blends was investigated further using SDS±PAGE. An aliquot (10 ml) of required sample was placed in a glass tube (i.d. 12 mm), the tube sealed and immersed in a 120 C oil bath for 5 min; this relatively mild heat treatment was used in an attempt to simulate conditions during the early stages of the heat coagulation process and heat-aggregate the most labile proteins. The sample was cooled quickly to 20 C and centrifuged at 100,000 g and 20 C for 1 h, using a Beckman Ultracentrifuge (model L8-70M, Beckman Instruments Inc., Palo Alto, CA). The purpose of centrifugation was to sediment the most heat sensitive whey proteins, which would be expected to heat-aggregate independently and/or complexed to casein micelles. A portion (10 0.5 mg) of ultracentrifugal (UC) pellet was dissolved in 1 ml of SDS reducing buer, and 5 l of this solution applied to a 12.5% SDS gel and the proteins separated and identi®ed using the same procedure as described above. RESULTS AND DISCUSSION Composition of raw materials Compositional data con®rmed that the SMP was a dilute solution of lactose and minerals (Table 1). The lactose content was similar to that of skim milk, but the mineral content was much lower, because, on a mass basis, about 2/3 of the total minerals, especially calcium and phosphorus, in milk are associated with casein and thus would be retained during UF (Green et al., 1984). The concentration of nitrogen in the SMP was low and probably consisted of non-protein nitrogen, namely, low molecular weight peptides, urea and other nitrogenous compounds, capable of permeating through UF membranes. As expected, the -F and -F were high in protein and low in lactose and ash. Electrophoretograms (results not shown) revealed that -la or -lg were the most abundant proteins in the -F or -F, respectively; the -F was more heterogeneous, containing also noticeable quantities of -lg and bovine serum albumin (BSA). The WPC had a relatively high amount of non-protein Table 1. Average (n = 2) composition (%, w/w) of materials used to prepare protein-adjusted milk Product Skim milk permeate -Fraction -Fraction Whey protein concentrate a
% Protein = n 6.38.
Moisture Proteina Lactose Ash 94.8 4.1 5.2 4.2
0.31 89.9 89.0 78.8
4.4 1.0 1.0 11.8
0.36 2.95 2.81 3.89
329
constituents and its protein component was heterogeneous. When the -F, -F or WPC were dissolved in SMP at 3.44% protein, the ®nal pH values were in the range 6.6±6.8 and changes in the pH of skim milk (6.7) were very small ( 0.05 pH units) upon admixture of any whey protein solution at any ratio. Heat stability of solutions of -fraction, -fraction or WPC To permit a better understanding of the thermal behaviour of proteins in the blends of skim milk with whey protein solution, the heat stability of individual whey protein preparations in SMP was measured at 96± 140 C. The -FS was extremely heat stable (Fig. 1), coagulating only at high temperatures and low pH values (140 C and < pH 6.7; 120 C and < pH 6.5). This was in contrast to the behaviour of the -FS or WPCS, which coagulated quickly at 120 C and almost instantaneously at 140 C; these solutions were resistant to coagulation only at a relatively low temperature (96 C) and high pH values (6.7±6.9). The heat stability results indicated that the thermal behaviour of the -FS was consistent with the known resistance of -la to heat coagulation (Larsen and Rolleri, 1955; Matsudomi et al., 1992), while that of the -F or WPC was dominated by -lg, documented as being sensitive to heat precipitation (de Wit, 1981; Matsudomi et al., 1992). Near neutral pH, -lg denatures at 75±80 C, while -la, in the presence of calcium, denatures at 60±65 C (Ruegg et al., 1977; de Wit et al., 1983; Bernal and Jelen,
Fig. 1. Mean (n = 3) heat coagulation time (HCT) of pHadjusted whey protein solutions, made by dissolving -fraction, -fraction or whey protein concentrate (WPC) in skim milk ultra®ltration permeate to 1.7% (w/w) protein; -fraction solution at 120 (*) or 140 C (*); -fraction solution at 96 (&) or 120 C (&); WPC solution at 96 (~) or 120 C (~). Lines exiting the graph indicate heat stabilities >50 min. At 140 C and all pH values, -fraction and WPC solutions had a heat stability of 0 min; at 96 C and all pH values, -fraction solution had a heat stability >50 min.
330
W. Rattray, P. Jelen
1984; Paulsson et al., 1985; Paulsson and Dejmek, 1990; Xiong et al., 1993; Hollar et al., 1995; Anema and McKenna, 1996); in the present study both -la and -lg would be expected to exhibit unfolded conformations at the high temperatures of 96±140 C. Anema and McKenna (1996) observed that at temperatures > 85 C for -la or >100 C for -lg, the proteins underwent aggregation reactions, evident by a reduction in entropy. Thus, in the heating conditions used in the present study, it can be concluded that in ionic environment of SMP, unfolded -lg was prone to rapid and extensive intermolecular interaction, leading to the formation of visible protein clots, whereas aggregation between unfolded -la molecules was much more limited such that visible protein clots did not form. The latter observation is consistent with the work of Chaplin and Lyster (1986), who showed that prolonged heating (100 C for up to 30 min) of a-la, in 0.1 M phosphate buer at pH 7.0, led to irreversible denaturation and polymerization, but the extent of polymerization was limited, allowing -la to remain soluble. The limited propensity of -la for thermal association was also demonstrated by Paulsson et al. (1986), who could not gel a 20% solution of -la, upon heating at 90 C and pH 6.6, in contrast to -lg which gelled at 5% protein. Hines and Foegeding (1993) observed a similar trend; when a 7% -la solution, at pH 7, in the presence of 100 mM NaCl, was heated at 80 C for 3 h, protein aggregation was extremely slow resulting in the formation of a very weak gel. The resistance of -la to thermal coagulation can sometimes be attributed to its calcium-assisted renaturation on cooling (Ruegg et al., 1977; Bernal and Jelen, 1984; Patocka et al., 1987), but continuous heating at 140 C would not allow renaturation. Vanderheeren and Hanssens (1994) showed that binding of the hydrophobic probe, bis-ANS (1,10 -bi [4anilo] naphthalene-5,50 -disulphonate), to apo- -la at pH 7.5 was maximal at 25 C; two strong binding sites were identi®ed. At 25 C, the protein probably existed in the molten globule state, corresponding to a loosening of tertiary structure with preservation of secondary structures, which allowed bis-ANS to more easily penetrate the hydrophobic interior of the molecule. At 80 C, apo- -la exhibited an unfolded conformation and the accessibility of hydrophobic groups to bis-ANS increased further; however, the loss of tertiary structure caused dissipation of hydrophobic clusters and only one weak binding site for bis-ANS was identi®ed. Similar trends occurred when -la was heated in the presence of calcium, although the eects took place at higher temperatures, because of stabilization of protein conformation by calcium. It has not been proven whether the reduced hydrophobic interaction between -la and bisANS at high temperature could be extrapolated to imply that relatively weak hydrophobic interactions would occur between unfolded proteins, but if this was
the case, it would, at least in part, explain the high resistance of -la to thermal precipitation. It is also possible that the limited ability of -la for self-association, compared to -lg, was related to a lower rate of sulphydryl group activity in -la. Considering that -lg contains a free sulphydryl group whereas -la does not, thiol group interactions between denatured -la molecules can be expected to be less prevalent than in the case of -lg. Calvo et al. (1993) showed that when -la and casein micelles were heated together at 90 C for 24 min, -la did not aggregate with itself or with casein micelles. However, the presence of -lg promoted aggregation of -la. These authors proposed that -lg could act as a catalyst, via its free sulphydryl group, which interacted with the sulphydryl bridges of -la, thus allowing -la to aggregate with itself, -lg or -casein. Similarly, Matsudomi et al. (1992) showed that a synergistic interaction between -la and -lg, in the formation of heat-induced gels, was due predominantly to thiol-disulphide interchange reactions between the two protein types during heating. At increased pH, it is likely that -la and -lg were denatured to a greater extent, as repulsive intramolecular electrostatic forces would be expected to increase. However, the increased heat stabilities of the -FS, -FS or WPCS, with increasing pH, indicate that intermolecular association between denatured protein molecules was also inhibited by increased electrostatic repulsion. Possibly, longer heating times were required to accentuate hydrophobic and thiol group interactions involved in protein aggregation or more extensive unfolding of protein was required to initiate association. The heat stability of the WPCS was lower than that of the -FS, probably due to its higher content of minerals; small (0±200 mM) changes in the concentration of calcium have a great eect on the thermal aggregation of -lg, due to the ability of calcium to form cross-bridges between proteins (Xiong et al., 1993). In succeeding experiments, the contrasting heat stabilities between -FS and -FS or WPCS were evidenced by the drastically dierent eects of these solutions on the heat stability of skim milk. Heat stability of blends of skim milk and -fraction solution In general, at 140 C and initial pH 6.6±6.7, mixtures of skim milk and -FS were more heat stable than skim milk alone, but considerably less heat stable when the initial pH was in the range 6.8±7.1 (Fig. 2). An exception to this trend was the very low heat stability, at pH 6.6, of a 40/60 combination of skim milk and -FS; this was probably due to the presence of the heat-labile -lg in the -FS. Skim milk itself showed a type A HCT/ pH pro®le, thought to be due to a pH-dependent interaction between serum proteins and -casein (Kudo, 1980; Singh and Fox, 1985, 1986, 1987a,b).
Thermal stability of skim milk/whey protein solution blends
Fig. 2. Mean (n = 3) heat coagulation time (HCT) of pHadjusted blends of skim milk and -fraction solution. The ratio of skim milk to -fraction solution was 100/0(- - -), 95/5 (*), 90/10 (*), 70/30 (&), 50/50 (&) or 40/60 (~).
The nature of thermal interactions between serum proteins and casein was investigated further by SDSPAGE of the UC pellets of heated (120 C 5 minÿ1) skim milk/ -FS blends. At pH 6.6, the increased ratio of -FS to skim milk led to greater amounts of -la, -lg and BSA in the UC pellets (Fig. 3, columns 1±3). The results re¯ected the composition of the -F which consisted primarily of -la but contained also smaller amounts of -lg and BSA and implied that increasing the amount of these proteins in the blends enhanced their thermal association with casein, especially for -la, the most plentiful protein. The whey proteins must have been complexed to the casein in order to be sedimented at 100,000 g for 1 h, because, unlike casein, their sedimentation coecients would be too low to permit
Fig. 3. Electrophoretic pro®le of ultracentrifugal pellets obtained from blends of skim milk and -fraction solution which had been pH adjusted, heat at 120 C for 10 min and cooled at 20 C. At pH>6.6, the ratio of skim milk to -fraction solution was (1) 100/0, (2) 60/40 or (3) 40/60; at pH 6.8 it was (4) 100/0 or (5) 50/50; at pH 7.1 it was (6) 100/0 or (7) 50/50.
331
exclusive sedimentation. Therefore, in skim milk, -la co-aggregated with the casein micelles, as opposed to a much more limited tendency for self-aggregation when dissolved in SMP (Fig. 1). Law et al. (1994) also showed a complexation reaction between casein and whey proteins; heating milk at natural pH and 80±140 C for 5 min caused extensive complexation of -la, -lg and bovine serum albumin with casein, allowing these proteins to co-precipitate with casein when the cooled milk was adjusted to pH 4.6. As described above, the work of Calvo et al. (1993) may indicate that the natural presence of -lg in skim milk is necessary for -la to react with casein micelles. However, the heating conditions used in the present study (140 C) were much more severe than that used by the above authors (90 C), obtruding a direct comparison between the two studies. At initial pH 6.6 or 6.7, the high heat stability of the blends of skim milk and -FS may have been due to increased complexation of -la onto casein micelles, increasing micellar charge and hydration and thus heat stability, as proposed by Fox and Hearn (1978). The presence of a small amount of -lg in the -F may have caused the decline in heat stability at pH 6.6, at the lowest ratio of skim milk to -FS (40/60), consistent with the increased concentration of -lg in the UC pellet of the heated mixture (Fig. 3, column 3). This seems likely considering that at 140 C and pH 6.6 the -FS had a HCT of 11 min, in contrast to the -FS which coagulated almost instantly (Fig. 1). Furthermore, skim milk with low levels of added -FS was also very unstable due to independent coagulation of -lg and/or -lg complexed to casein, as will be discussed below. An eect similar to the present results was reported by Peter et al. (1996) who observed that the addition of increasing amounts of a dried fraction of -la to milk, at pH 6.7, caused a gradual decline in heat stability. At initial pH 6.8 or 7.1, UC pellets from heated blends of skim milk and -FS did not contain increased quantities of any whey protein (Fig. 3, columns 4±7), indicating that most of the added whey protein component remained soluble during heating and the low heat stabilities were caused by coagulation of casein alone; yet the sensitivity of casein to coagulation was increased in the presence of extra whey proteins. The reduced heat stability of the caseins in the 50/50 blends is also evidenced by the increased amounts of casein in the UC pellets (Fig. 3, columns 5 and 7) despite the fact that the blends contained lower amounts of casein than in skim milk. The low heat stability of type A milk near pH 6.9 may be due to the formation of soluble whey protein/casein complexes and concomitant heat labile -caseindepleted micelles (Kudo, 1980; Singh and Fox, 1985, 1986, 1987a,b); possibly, in the skim milk/ -FS blends, the rate of this destabilizing interaction would be increased because of an increased concentration of -la and/or -lg.
332
W. Rattray, P. Jelen
Heat stability of blends of skim milk and -fraction solution Skim milk in combination with -FS had a substantially dierent heat stability pattern, compared to skim milk mixed with -FS; this was expected in view of the starkly dierent thermal properties of the -FS and -FS (Fig. 1). In most cases, blends of skim milk and FS showed a greatly reduced heat stability, compared to normal skim milk (Fig. 4). For all combinations of skim milk and -FS, the zone of minimal heat stability deepened and was extended to more alkaline pH values. At initial pH 6.7, low levels of added -FS caused a modest increase in heat stability, but when the ratio of -FS to skim milk was >10/90, heat stability declined markedly. At a 70/30 ratio of skim milk to -FS and at pH 6.8, the HCT increased; it appeared that the entire HCT-pH curve was shifted to more alkaline pH values. At all pH values, 60/40 mixtures were extremely unstable; a maximum heat stability of only about 5 min was recorded when the initial pH of the blends was 6.8. Gel electrophoresis of the UC pellets of heated (120 C for 5 min) 50/50 blends revealed that large quantities of -lg, as well as casein, were present (Fig. 5), suggesting that the heat sensitivity of the mixtures was caused by coagulation of -lg alone and/or of -lg which had heat-aggregated with casein micelles. Unlike -la, -lg was unstable in SMP at 120 and 140 C. Accordingly, the low heat stability of the skim milk / FS combinations at 140 C could have resulted from exclusive coagulation of -lg or -lg/casein micelle complexes or both, all of which would sediment by ultracentrifugation. Compared to mixtures of skim milk with -FS at pH 6.6 (Fig. 2, columns 1±3), skim milk with -FS showed much greater quantities of -lg in the UC pellets, attesting that denatured -lg was much
more prone than denatured -la to intermolecular interaction with itself and/or casein. This concurs with Mottar et al. (1989), who reported that a more severe heating of milk was required to induce -la association onto casein micelles, as compared to -lg. In the absence of casein, at pH 6.8, the -FS was heat sensitive (Fig. 1), while unaltered skim milk also displayed a relatively low heat stability at pH 6.8. This suggests that in the 70/30 blends, at pH 6.8, the heat stable entity must have been a -lg/casein micelle complex, which the SDS±PAGE results also indicate (Fig. 5, column 5). The result indicates that a shift of the heat stability/pH curve to more alkaline pH values occurred. When normal skim milk is heated at pH 6.7, -lg heatprecipitates onto the surface of casein micelles to increase their heat stability (Kudo, 1980; Singh and Fox, 1985, 1986, 1987a,b); the peak heat stability of the 70/30 mixture at pH 6.8 indicates that in the presence of extra -lg, a favourable degree of interaction between -lg and casein occurred at a higher pH value. The slightly improved heat stability, at pH 6.7, of 95/5 or 90/ 10 combinations, also may have been due to thermal complexation between -lg and casein micelles. At > pH 6.9, all blends of skim milk and -FS had low heat stabilities, possibly because of a more rapid formation of soluble -casein/ -lg complexes and hence heat labile, -casein depleted micelles and/or the sole aggregation of -lg. Patocka et al. (1993) showed that combinations of skim milk and acid whey UF retentate, at pH 6.5±6.7, coagulated at 90 C; the high calcium content of the UF retentate promoted formation of casein/whey protein co-precipitates. The same authors found that mixtures of skim milk and sweet or UC whey UF retentate were heat stable at 90 C and pH 6.5±6.7, attributed to lower levels of calcium in the UF retentate. The blends of skim
Fig. 4. Mean (n = 3) heat coagulation time (HCT) of pHadjusted blends of skim milk and -fraction solution. The ratio of skim milk to -fraction solution was 100/0 (- - -), 95/5 (*), 90/10 (*), 70/30 (&) or 60/40 (&).
Fig. 5. Electrophoretic pro®le or ultracentrifugal pellets obtained from blends of skim milk and -fraction solution, which had been pH adjusted, heated at 120 C for 10 min and cooled to 20 C. At pH 6.7, the ratio of skim milk to -fraction solution was (1) 100/0 or (2) 50/50; at pH 6.8 it was (3) 100/0, (4) 90/10, (5) 70/30 or (6) 50/50; at pH 7.1 it was (7) 100/0 or (8) 50/50.
Thermal stability of skim milk/whey protein solution blends milk and -FS, used in the current study, would be expected to be similar in composition to blends of skim milk and sweet or UC whey used by Patocka et al. (1993). The results imply that, without elevated calcium levels, the presence of extra -lg was also detrimental to the heat stability of skim milk, the eect occurring over a range of pH values, albeit at a much higher temperature (140 C). Heat stability of blends of skim milk and WPC solution Combinations of skim milk and WPCS had HCT-pH pro®les (Fig. 6) broadly similar to skim milk/ -FS blends, indicating that heat stability was in¯uenced mainly by an increased concentration of -lg, the most abundant protein in the WPC. In most cases, blends of WPCS and skim milk were much less heat stable than the skim milk alone, a conspicuous exception being a pronounced increase at pH 6.8 and 70/30 ratio of skim milk to WPCS. The results indicated that the HCT/pH pro®le was shifted to more alkaline pH values, as also occurred with a 70/30 ratio of skim milk to -FS. Gel electrophoresis of UC pellets obtained from heated samples (results not shown) gave patterns very similar to those observed for the combinations of skim milk and -FS (Fig. 5); this implies that the heat stability pattern was again due to the presence of extra -lg and the heat instability may have been caused by aggregation of -lg alone and/or -lg in association with casein. Similarly, the SDS± PAGE results indicated that the high thermal stability of a 70/30 blend of skim milk and WPCS at pH 6.8 could have been due to the formation of a complex between -lg and -casein. Studies have shown that whey protein products rich in -lg have similar thermal properties to more puri®ed
333
-lg preparations (de Wit, 1981; Hines and Foegeding, 1993), so it should not be surprising that the overall heat stability and gel electrophoretic patterns of skim milk combined with -FS or WPCS were similar. However, the actual HCT values of skim milk/WPC solution combinations were higher, suggesting an in¯uence of other proteins and/or non-protein constituents in the WPC, the nature of which was not established. CONCLUSIONS The thermal behaviour of industrial whey protein preparations was dominated by their most abundant protein constituent; -la in the -F; and -lg in the -F or the WPC. Con®rming the well known dierences in the heat stabilities of -la and -lg, the -FS was resistant to coagulation in SMP at temperatures up to 140 C, whereas -lg quickly formed visible protein clots at 140 C. This was probably due to a greater degree of thermal unfolding for -lg as compared to -la, implying that the overall heat stability of -la, in terms of conformational changes and aggregation, is greater than that of -lg. The heat stabilities of blends of skim milk and whey protein solutions were dependent on the type of whey protein preparation used and the initial pH. At the unaltered pH (6.7) of the mixtures, -la had a bene®cial eect on heat stability up to at least a 40/60 ratio of skim milk to -FS, whereas -lg caused a small increase in heat stability at < 90/10 ratio but was deleterious at higher levels of addition. In general, at pH values > 6.7, extra -la or especially -lg caused the heat stability to decline; pronounced exceptions were the high heat stabilities at pH 6.8 of 70/30 combinations of skim milk and -FS or WPCS. REFERENCES
Fig. 6. Mean (n = 3) heat coagulation time (HCT) of pHadjusted blends of skim milk and whey protein concentrate solution. The ratio of skim milk to whey protein concentrate solution was 100/0 (- - -), 95/5 (*), 90/10 (*), 70/30 (&) or 60/40 (&).
Abd El Salam, M. H., El-Shibny, S., Mahfouz, M. B., El-Dein, H. F., El-Atriby, H. M. and Antila, V. (1991) Preparation of whey protein concentrate from salted whey and its use in yogurt. J. Dairy Res. 58, 503±510. Anema, S. G. and McKenna, A. B. (1996) Reaction kinetics of thermal denaturation of whey proteins in heated reconstituted whole milk. J. Agric. Food Chem. 44, 422±428. AOAC (1990) Ocial Methods of Analysis, 13th edn. Association of Ocial Analytical Chemists, Arlington, VA, pp. 807±808, 834±835. Bech, A. (1993) Characterising ripening in ultra®ltered cheese. Int. Dairy J. 3, 329±342. Bernal, V. and Jelen, P. (1984) Eect of calcium binding on thermal denaturation of bovine -lactalbumin. J. Dairy Sci. 67, 2452±2454. Buchheim, V. W., Peters, K. H. and Kaufmann, W. (1986) Technological, physico-chemical and sensory properties of milk and sour milk with a modi®ed casein/whey protein ratio. Milchwissenschaft 41, 139±141.
334
W. Rattray, P. Jelen
Calvo, M. M., Leaver, J. and Banks, J. M. (1993) In¯uence of other whey proteins on the heat-induced aggregation of -lactalbumin. Int. Dairy J. 3, 719±727. Chaplin, L. C. and Lyster, R. L. J. (1986) Irreversible heat denaturation of bovine -lactalbumin. J. Dairy Res. 53, 249±258. Davies, D. T. and White, J. C. D. (1966) The stability of milk protein to heat. 1. Subjective measurement of heat stability of milk. J. Dairy Res. 33, 67±81. de Wit, J. N. (1981) Structure and functional behaviour of whey proteins. Neth. Milk Dairy J. 35, 47±64. de Wit, J. N. (1989). The use of whey protein products. In Developments in Dairy Chemistry. 4. Functional Milk Proteins. P. F. Fox, ed. pp. 323±346. Applied Science, London. de Wit, J. N., Klarenbeek, G. and Hontelez-Backx, E. (1983) Evaluation of functional properties of whey protein concentrates and whey protein isolates. I. Isolation and characterization. Neth. Milk Dairy J. 37, 37±49. Fox, P. F. and Hearn, C. M. (1978) Heat stability of milk: in¯uence of denaturable proteins and detergents on pH sensitivity. J. Dairy Res. 45, 159±172. Green, M. L., Scott, J., Anderson, M., Grin, M. A. C. and Glover, F. A. (1984) Chemical characterization of milk concentrated by ultra®ltration. J. Dairy Res. 51, 267±273. Hines, M. E. and Foegeding, E. A. (1993) Interactions of -lactalbumin and bovine serum albumin with -lactoglobulin in thermally induced gelation. J. Agric. Food Chem. 41, 341±346. Hollar, C. M., Parris, N., Hsieh, A. and Cockley, K. D. (1995) Factors aecting the denaturation and aggregation of whey proteins in whey protein concentrate mixtures. J. Dairy Sci. 78, 260±267. Jelen, P., Buchheim, W. and Peters, K. H. (1987) Heat stability and use of milk with modi®ed casein/whey protein content. Milchwissenschaft 42, 418±421. Kudo, S. (1980) The heat stability of milk: formation of soluble proteins and protein-depleted micelles at elevated temperatures. N. Z. J. Dairy Sci. Technol. 15, 255±263. Larsen, B. L. and Rolleri, G. D. (1955) Heat denaturation of the speci®c serum proteins in milk. J. Dairy Sci. 38, 351±360. Law, A. J. R., Horne, D. S. and Leaver, J. (1994) Heatinduced changes in the whey proteins and caseins. Milchwissenschaft 49, 125±129. Matsudomi, N., Oshita, T., Sasaki, E. and Kobayash, K. (1992) Enhanced heat-induced gelation of -lactoglobulin by -lactalbumin. Bioscience, Biotechnology and Biochemistry 56, 1697±1700. McIntosh, G. H., Regester, G. O., LeLeu, R. K., Royle, P. J. and Smithers, G. W. (1995) Dairy products protect against dimethylhydrazine-induced intestinal cancers in rats. J. Nutr. 125, 809±816. Mottar, J., Bassier, A., Joniau, M. and Baert, J. (1989) Eects of heat-induced association of whey proteins and casein micelles on yogurt texture. J. Dairy Sci. 72, 2247±2256. McDonough, F. E., Alford, J. A. and Womack, M. (1976) Whey protein concentrate as a milk extender. J. Dairy Sci. 59, 34±38.
Patocka, J., Drathen, M. and Jelen, P. (1987) Heat stability of isolated whey protein fractions in highly acidic conditions. Milchwissenschaft 42, 700±703. Patocka, G., Jelen, P. and Kalab, M. (1993) Thermostability of skimmilk with modi®ed casein/whey protein content. Int. Dairy J. 3, 35±48. Paulsson, M. and Dejmek, P. (1990) Thermal denaturation of whey proteins in mixtures with casein studied by dierential scanning calorimetry. J. Dairy Sci. 73, 590±600. Paulsson, M., Hegg, P. O. and Castberg, H. B. (1985) Thermal stability of whey proteins studied by dierential scanning calorimetry. Thermochim. Acta 95, 435±440. Paulsson, M., Hegg, P. and Castberg, H. B. (1986) Heatinduced gelation of individual whey proteins. A dynamic rheological study. J. Food Sci. 52, 87±90. Peter, S., Rattray, W. and Jelen, P. (1996) Heat stability and sensory quality of protein-standardized 2% fat milk. Milchwissenschaft 51, 611±616. Puhan, Z. (1992) Standardization of milk protein by membrane processes for product manufacture. In New Applications of Membrane Processes, pp. 23±32. IDF special issue No. 9201, Brussels, Belgium. Rattray, W. and Jelen, P. (1996) Protein standardization of milk and dairy products. Trends Food Sci. Technol. 7, 227± 234. Ruegg, M., Morr, U. and Blanc, B. (1977) A calorimetric study of the thermal denaturation of whey proteins in simulated milk ultra®ltrate. J. Dairy Res. 44, 509±520. Savello, P. A. and Dargan, R. A. (1995) Improved yoghurt physical properties using ultra®ltration and very-high temperature heating. Milchwissenschaft 50, 86±90. Singh, H. and Fox, P. F. (1985) Heat stability of milk: pH-dependent dissociation of micellar -casein on heating milk at ultra high temperature. J. Dairy Res. 52, 529± 538. Singh, H. and Fox, P. F. (1986) Heat stability of milk: further studies on the pH-dependent dissociation of micellar -casein. J. Dairy Res. 53, 237±248. Singh, H. and Fox, P. F. (1987a) Heat stability of milk: role of -lactoglobulin in the pH-dependent dissociation of micellar -casein. J. Dairy Res., 54, 509±521. Singh, H. and Fox, P. F. (1987b) Heat stability of milk: in¯uence of colloidal and soluble salts and protein modi®cation on the pH-dependent dissociation of micellar -casein. J. Dairy Res. 54, 523±534. Xiong, Y. L., Dawson, K. A. and Wan, L. (1993) Thermal aggregation of -lactoglobulin: eect of pH, ionic environment, and thiol reagent. J. Dairy Sci. 76, 70±77. Vanderheeren, G. and Hanssens, I. (1994) Thermal unfolding of bovine -lactalbumin. Comparison of circular dichroism with hydrophobicity measurements. J. Biol. Chem. 269, 7090±7094. Wong, C. W. and Watson, D. L. (1995) Immunomodulatory eects of dietary whey proteins in mice. J. Dairy Res. 62, 359±368.
(Received 7 February 1997; accepted 24 August 1997)