Casein Micelles as Colloids: Surface Structures and Stabilities D. G. DALGLEISH Department of Food Science, University of Guelph, Guelph, ON, Canada N1G 2W1
ABSTRACT The interactive behavior of bovine casein micelles during processing depends on the structure of the micellar surface. Although some gross features are known, such as that the surface is covered by “hairs” of the macropeptide of k-casein, the details of the surface topology are lacking because of the apparent complexity of the micellar structure or a general lack of understanding of it. It is instructive to attempt to understand the micellar surface by comparing it with some better characterized colloidal systems, such as emulsion droplets or polystyrene latex particles, the surfaces of which are covered by monolayers of adsorbed proteins, either single types or as mixtures. This paper describes such a comparison, combining information from micellar studies and models. The comparison suggests that the surface of the casein micelle is only partly similar to the colloidal systems because of the different structures of the interiors of the particles and the manner in which the surface caseins interact with the interior of the micelle. The k-casein molecules on the micellar surface are probably distributed unevenly and have spaces between them that may permit the passage of other molecules, which might have some important consequences for understanding the behavior of the micelle. ( Key words: casein micelles, structure, emulsions, interfacial structures) INTRODUCTION Casein micelles are considered to be colloidal particles because they consist of aggregated caseins with a stabilizing surface layer; therefore, the micelles may resemble other colloids that are stabilized by proteins, such as emulsions, in their surface structures. A model of the particles has been simplified so that the particles are considered to be solid spheres with a
Received June 23, 1997. Accepted March 16, 1998. 1Present address: CIRDE, 15 Avenue Galile ´ e. 92350 Le PlessisRobinson, France. 1998 J Dairy Sci 81:3013–3018
coating of protein (specifically k-CN) on their surfaces; this coating is considered to provide stability, and the consequences of such a structure have been explored in a recent review (22). Although such a model has the merit of simplicity, it can be difficult to reconcile with the structural complexity of the interior of the casein micelle and even of its surface. The available evidence suggests that the micelle is a more complex entity with properties that reveal a subtlety that cannot be explained adequately by the simple models of colloids. However, the comparison can be used to explain some aspects of micellar behavior. This paper summarizes what is known of the structures of proteins that are adsorbed to the surfaces of colloidal particles and also describes some aspects of the structure of the surface layer in casein micelles. Finally, a synthesis of these two is attempted. COLLOIDAL PARTICLES STABILIZED BY CASEINS First considered are simple casein-stabilized colloids, such as latex particles coated with caseins or the oil droplets in casein-stabilized oil-in-water emulsions. These particles consist essentially of cores that are either solid or immiscible with water to which surfaces the proteins adsorb by hydrophobic interactions with the nonaqueous particles. The adsorbed proteins probably do not penetrate into the nonaqueous phase. Although with a solid latex, penetration is not possible, in principle it is possible for the protein to penetrate partly into the oil phase in an emulsion. The similarity of the dimensions of caseins adsorbed to latex particles ( 5 ) , to triglyceride oil emulsion droplets ( 6 ) , and to hydrocarbon interfaces ( 3 0 ) suggests that the proteins have similar conformations under these different conditions, which in turn suggests that the casein is not significantly submerged in the oil droplet surface but rather lies on the oil-water interface. Caseins adsorbed to the interfaces of these particles prove to have rather extended conformations; even monolayers of the proteins extend from the interface by more than 10 nm (4, 16, 28). Caseins have
molecular masses around 19,000 to 23,000 Da, and globular proteins of this mass (e.g., b-LG) have dimensions considerably smaller than the measured dimensions of the observed caseins. This extension of the casein molecules demonstrates the nonglobular natures of the caseins and that the molecules consist of hydrophobic and hydrophilic domains (29), which are close to the interface and projecting into solution, respectively ( 3 ) . The extension of the caseins away from the surface also depends on their charge because reduction of the charge causes the thickness of the adsorbed layer to decrease ( 4 ) . This adsorbed structure, in which the hydrophobic moiety interacts with the interface and the hydrophilic part protrudes into solution, has been modeled for b-CN and observed in other research (32). For b-CN adsorbed to oil droplets or to latex particles, the protruding part of the molecule consists of the N-terminal region ( 3 ) , where most of the protein charge (from the phosphoserine residues) is to be found, and the majority of the mass of the protein is close to the interface. The other caseins are harder to model in this way because they do not present an obvious distinction between the hydrophilic and hydrophobic regions in the primary structures. However, measurement has shown that adsorbed caseins, of which k-CN is of particular relevance to this discussion, also form extended adsorbed layers; their conformation depends partly on the extent of disulfide-linked polymers present in the k-CN (5, 28). The conformations of the adsorbed caseins also depend on the extent to which the interface is covered by the protein. If other surfactants are absent, the molecules of the protein can spread to cover the interface; thus, at an overall surface coverage of 1 mg·m–2, casein molecules appear to be stretched to their maximum extent, and the thickness of the adsorbed layer is less (16). Conversely, the presence of excess casein increases the monolayer coverage to a maximum value of 3 mg·m–2, the parts of the molecules in contact with the interface adopt a more compact conformation, and the hydrophilic moieties protrude further from the interface. Therefore, the addition of casein to the low protein emulsion causes increased adsorption and conformational change (16). These observations indicate that the caseins are relatively mobile on an oil-water interface; at least, conformation of the caseins might change, and it might also be possible to move them around (to accommodate added molecules). The saturating surface coverage is important because it permits the maximum dimensions of closely packed protein molecules to be calculated. For a molecular mass of 23,000 Da, a coverage of 3 mg·m–2 corresponds to an area per molecule of 13 Journal of Dairy Science Vol. 81, No. 11, 1998
nm2 so that the centers of the molecules must be 3 to 4 nm distant from one another in their most closely packed and adsorbed state. The area per molecule increases to about 40 nm2 for a surface coverage of 1 mg·m–2. However, the volume occupied by the protein molecules ( ∼35 nm3) is constant in the two cases. The presence of caseins adsorbed to oil-water or solid-water interfaces stabilizes the particles; in many cases, hydrolysis of the proteins by trypsin does not destabilize an emulsion made using sodium caseinate ( 2 ) because the adsorbed as-CN and b-CN are attacked first, either because they protrude more from the interface or because of greater affinity between the b-CN and as-CN and trypsin (13). This hydrolysis may even increase the stability of the particles to the presence of calcium ions because the adsorbed kCN remains intact until the calcium-sensitive caseins have been destroyed, then is itself attacked ( 1 ) . kCasein apparently can exhibit stabilizing properties when remote from its natural place in the casein micelles, but whether k-CN is a major cause of the stability of the original caseinate-stabilized emulsions has not been demonstrated. Both as-CN and b-CN are capable of the efficient stabilization of emulsions alone or in mixtures lacking k-CN. Of the individual caseins, k-CN is the poorest emulsifier, as estimated by the sizes of the droplets in emulsions formed using this protein (14). The emulsifying capacity is improved when the protein is treated with mercaptoethanol, thus breaking intermolecular disulfide bonds and possibly breaking up hydrophobic clusters. However, the protein has been shown to adsorb to the surfaces of polystyrene latex particles and to give varying layer thicknesses, depending on the variant of the casein that is used and the extent of glycosylation, although, in all cases, substantial layers with thicknesses in the range 10 to 15 nm are formed (28). When treated with rennet, the layers of adsorbed k-CN are diminished because of the loss of the macropeptide, and the thickness of this layer appears to be of the order of 10 nm, which is similar to the dimensions of the other adsorbed caseins. Having summarized some of the important aspects of caseins adsorbed to interfaces, we may now try to contrast these observations with the behavior of true casein micelles. PROPERTIES OF CASEIN MICELLES Regardless of the detailed internal structure of the casein micelle, it is more homogeneous in composition than that of an emulsion droplet or a casein-coated latex particle. That is, casein micelles contain pro-
SYMPOSIUM: CASEIN MICELLE STRUCTURE
teins throughout their structures that are quite similar in nature to the proteins on the surface. The surface caseins in the micelle are almost certainly not adsorbed in the same way as the proteins are adsorbed to hydrophobic interfaces. Hydrophobic interactions very likely play a role in defining the structure of the micelle, but the surface proteins cannot be washed off the micelles by small molecule surfactants as they are from emulsions (12). Other types of bonds link the surface of the micelle to the proteins in the interior. Moreover, because interactions occur between the caseins and micellar calcium phosphate (19), any surface caseins that interact in this way likely differ in conformation from those physically adsorbed to interfaces. Thus, because a clear definition of the boundary between the interior of the micelle and the surrounding serum is not possible, the proteins on or near the surface of the micelle might not behave similarly to the same proteins in the adsorbed state. For example, b-CN on an oil-water interface is adsorbed by means of its C-terminal region (3, 29, 32). However, in the micelle, the protein could be attached through its N-terminal region, where serine phosphates are present and capable of interacting with micellar calcium phosphate (23). In which case, any part of this casein that protrudes into solution is likely to be derived from the C-terminal end of the protein; because the C-terminal is highly hydrophobic, direct contact with the aqueous phase may be unlikely. Such behavior is likely to be less for k-CN because each molecule contains only one serine phosphate residue and will react only weakly with the calcium phosphate. The caseins in the micelle are known to possess some mobility; for example, cooling causes b-CN to be liberated from the micelles into the serum. When the milk is again warmed, the b-CN returns to the micelle but the brief liberation of the b-CN shows that some weak (hydrophobic) bonds exist in the structure. It is not known whether the b-CN comes from the interior of the micelle, but if the model shown by Holt and Horne ( 2 2 ) is correct, diffusion from the interior of the micelles to the serum is likely to be hindered somewhat, and so the dissociating b-CN may come from rather close to the micellar surface. Similarly, when the cooled micelles are warmed and the b-CN reassociates with them, it seems unlikely that the bCN will be able to penetrate deeply into the micelle. Thus, during reassociation, the dissociated b-CN may alter the structure of the surface of the casein micelle. Further evidence of protein mobility has come from studies of the nuclear magnetic resonance spectra of the micelles; these spectra show that, in the native
micelles at room temperature (25°C), the only mobile material appears to be the macropeptide of k-CN (18, 36). However, as the temperature is increased, there is evidence for a general loosening of the structure of the micelle and a greater mobility of individual proteins (31, 35). Of the different caseins, k-CN is the least likely to bind to calcium phosphate but does not dissociate readily from the micelle when the calcium phosphate is dissolved ( 2 1 ) or from the micelle during cooling or when surfactants are added. Thus, the links between the k-casein and the other caseins cannot be simply ascribed to hydrophobic bonding. Therefore, in casein micelles, the k-CN does not act simply as a hydrophobic surface-active material that emulsifies the rest of the caseins, although some aspects of its behavior may suggest this relationship. It is generally agreed that the surface of the micelle is covered by a diffuse layer made up of the macropeptide of k-CN. The existence of this hairy layer was demonstrated by the change in hydrodynamic diameter measured by dynamic light scattering during the renneting of milk (39). This and subsequent experiments (17, 20, 24) performed on highly diluted micelle systems suggest that the thickness of the macropeptide layer of the k-CN is about 5 nm. However, a measurement by a different light-scattering technique in concentrated milk has given a higher value of about 10 nm, which is closer to the value found in emulsions (26). The lower value, however, is in agreement with the thickness of casein layers spread on an interface to the extent of 1 mg·m–2. A calculation of the effect of solvent drainage through the hairy layer and the comparison with experimental results has also suggested that the layer might be as much as 12 nm thick and that relatively few of the k-CN molecules on the surface actually participated in the formation of the layer (20). In addition to being removed by the action of rennet, the hairs of the k-CN appear to be flattened by the action of ethanol, presumably because ethanol is a poor solvent, causing the macropeptide layer to collapse (24). The change in diameter caused by ethanol is also consistent with a layer about 10 nm thick, although this thickness seems to be dependent upon the solution in which the micelles are suspended and might arise from conformational changes in more than k-CN (25). Decrease of pH also diminishes the thickness of the macropeptide layer, whether measured directly by the change in diameter as the pH is altered ( 2 5 ) or by the decrease in the micellar diameter caused by renneting at different pH values (M. Morisot, C. Tranchant, and D. G. Dalgleish, 1997, unpublished observations). Journal of Dairy Science Vol. 81, No. 11, 1998
STRUCTURE OF THE MICELLAR SURFACE Studies of the composition of micelles of different sizes have demonstrated that most, if not all, of the kCN is to be found on the surface of the micelle (9, 15). However, other caseins must also be present on the micellar surface, and their function in this position has not been established. As has been described, the maximum surface coverage of emulsions (i.e., closely packed caseins) is about 3 mg·m–2, and the surface that is available to a protein molecule is about 13 nm2. However, from the measured sizes of casein micelles [mean diameters of about 200 nm ( 8 ) ] and the content of k-CN in whole casein [about 12% (11)], we can make an approximate calculation that the available k-CN can cover the surfaces of the micelles to the extent of only 1 mg·m–2 at maximum, so about one-third of the surface is covered by this protein. This value may indeed be an overestimate, because assuming smaller mean micelle sizes or smaller contents of k-CN in the calculation leads to an even smaller value for surface coverage. A similar value can be calculated by estimating what proportion of the total casein molecules in casein micelles is present on the micellar surface by calculation from the dimensions of the individual caseins and of the casein micelles. A slightly smaller result is implied by the results of Donnelly et al. (15), who suggested an area of 47 nm2 per molecule of k-CN, which corresponds to a surface coverage of about 0.8 mg·m–2. Thus, most calculations agree that about one-third of the micellar surface is covered by k-CN, and the remainder must be occupied by other proteins; a similar value was calculated from measurements of the protein compositions of micelles of different sizes ( 9 ) . Which of the other caseins shares the surface with k-CN is less clear. Compositional analysis of micelles separated by centrifugation (9, 10) suggests that asCN are both on the surface and in the interior of the micelle and that b-CN mainly occupies interior positions. However, compositional studies of micelles separated by chromatography on controlled pore glass suggest that both as-CN and b-CN may share the surface with k-CN (15), and measurements of the rates of hydrolysis of different proteins in micelles treated with trypsin ( 1 3 ) suggest that b-CN is the protein nearest to the surface. So far, this discrepancy is unresolved. However, whatever casein shares the micellar surface with k-CN, it likely will possess a structure different from that adopted when the same proteins are adsorbed, because of the interactions between caseins and calcium phosphate within the micelle. Thus, the peptides that are protruding from the micellar surface and creating a hairy layer ( 2 2 ) probably are from k-CN alone, as has been shown in Journal of Dairy Science Vol. 81, No. 11, 1998
renneting experiments. These observations also show that the k-CN attaches to the micelle by means of its para-k-CN moiety, which is similar to its adsorbed state in colloids. However, because only approximately one molecule in three on the micellar surface is k-CN, the hairs may be well spaced, at about one kcasein molecule per 40 to 50 nm2 of micellar surface, compared with a packed monolayer of casein of 3 mg·m–2, in which the area occupied by a single protein is in the region of 13 nm2. Conceivably, other caseins also help to provide the hairy surface, but this scenario is unlikely because no mobile portions of the other caseins (as in the macropeptide of k-CN) have been detected by nuclear magnetic resonance experiments of unheated casein micelles. It is also likely that the distribution of k-CN in the micellar surface layer is not homogeneous because kCN isolated from micelles exists in a mixture of disulfide-linked polymeric forms of different degrees of polymerization (34), and this oligomerization is also present in the protein in the micelles themselves (38). Thus, the micellar surface is likely to contain small islands of k-CN that are surrounded by other caseins. Possibly, these caseins might represent the surface heterogeneities that are observed by electron microscopy (37), but at present no evidence supports this notion. From a distance (i.e., from the perspective of an approaching micelle) this heterogeneity of the surface may be too small to be detectable, but it may nevertheless affect the behavior of the micelle. Such a description of the surface suggests a possible explanation of a number of reactions by the casein micelles. First, the change in diameter caused by removing the k-CN macropeptides from micelles by the action of rennet [about 5 nm (39)] is smaller than the corresponding change in latex particles coated with k-CN. This difference can be explained if the molecules of k-CN are sufficiently separated on the micellar surface that the hairy layer is hydrodynamically draining. A calculation based on this premise ( 2 0 ) has indeed shown that the hairy layer may be as much as 12 nm thick, in agreement with the results from the simple colloids (6, 28). In addition, this dimension can be used to predict the onset of rennetinduced coagulation of casein micelles based on a simple geometric model ( 7 ) . A second use of the distributed k-CN model is to help explain how other protein molecules may readily penetrate the surface. For example, it has been a problem to explain how enzymes (e.g., rennet) can penetrate the hairy layer to find the rennet-sensitive Phe-Met bond. Although this bond may be in a configuration to be attacked by chymosin (33), it is hard to explain how the rennetsensitive bond can be on the outermost part of the hairy layer and yet liberate a long peptide, changing
SYMPOSIUM: CASEIN MICELLE STRUCTURE
the hydrodynamic diameter of the micelle, when it is cleaved. Similarly, trypsin or other enzymes seem to be capable of penetrating within the surface layer to attack the other caseins (13). Another specific point of attack of the casein micelles is the formation of disulfide bonds between micellar k-CN and b-LG from the serum when milk is heated (27). This reaction involves the cysteine residues in the para-k-CN part of the molecule, which must certainly be inside the hairy layer of glycomacropeptide. However, if the hairy layer is diffuse or has gaps in it because of the distribution of the k-CN on the micellar surface, the diffusion of the incoming whey proteins may be relatively unhindered. A final benefit of this model is that the liberation of b-CN from the micelle during cooling may be understood because the hairy layer of k-CN may be sufficiently diffuse to permit the passage of the individual molecules of b-CN. CONCLUSIONS In summary, a comparison of adsorbed k-CN on colloidal surfaces and micellar k-CN suggests that the micellar surface must be only partially covered by kCN and that the k-CN is heterogeneously distributed on the surface. This surface coverage provides steric stabilization against the approach of large particles such as other micelles, but the small-scale heterogeneities and the gaps between the k-CN molecules provide relatively easy access for molecules of the dimensions of individual proteins and smaller. REFERENCES 1 Agboola, S. O., and D. G. Dalgleish. 1996. Enzymatic hydrolysis of milk proteins for emulsion formation. 1: Kinetics of protein breakdown and storage stability of emulsions. J. Agric. Food Chem. 44:3631–3636. 2 Agboola, S. O., and D. G. Dalgleish. 1996. Enzymatic hydrolysis of milk proteins for emulsion formation. 2: Effects of calcium, pH and ethanol on the stability of emulsions. J. Agric. Food Chem. 44:3637–3642. 3 Atkinson, P. J., E. Dickinson, D. S. Horne, and R. M. Richardson. 1995. Neutron reflectivity of adsorbed b-casein and blactoglobulin at the air-water interface. J. Chem. Soc. Faraday Trans. 91:2847–2854. 4 Brooksbank, D. V., C. M. Davidson, D. S. Horne, and J. Leaver. 1993. Influence of electrostatic interactions on b-casein layers adsorbed on polystyrene latices. J. Chem. Soc. Faraday Trans. 89:3419–3425. 5 Dalgleish, D. G. 1990. The conformations of proteins on solid/ water interfaces—caseins and phosvitin on polystyrene latices. Colloids Surf. 46:141–155. 6 Dalgleish, D. G. 1993. The sizes and conformations of the proteins in adsorbed layers of individual caseins on latices and in oil-in-water emulsions. Colloids Surf. B 1:1–8. 7 Dalgleish, D. G., and C. Holt. 1988. A geometrical model to describe the initial aggregation of partly renneted casein micelles. J. Colloid Interface Sci. 123:80–84. 8 Dalgleish, D. G., and D. S. Horne. 1985. A photon correlation spectroscopy study of size distributions of casein micelle suspensions. Eur. Biophys. J. 11:249–258.
9 Dalgleish, D. G., D. S. Horne, and A.J.R. Law. 1989. Sizerelated differences in bovine casein micelles. Biochim. Biophys. Acta 991:383–387. 10 Davies, D. T., and A.J.R. Law. 1983. Variation in the protein composition of bovine casein micelles and serum casein in relation to micellar size and milk temperature. J. Dairy Res. 50: 67–75. 11 Davies, D. T., and A.J.R. Law. 1987. Quantitative fractionation of casein mixtures by fast protein liquid chromatography. J. Dairy Res. 54:369–376. 12 de Feijter, J. A., J. Benjamins, and M. Tamboer. 1987. Adsorption displacement of proteins by surfactants in oil-in-water emulsions. Colloids Surf. 27:243–266. 13 Diaz, O., A. M. Gouldsworthy, and J. Leaver. 1996. Identification of peptides from casein micelles by limited trypsinolysis. J. Agric. Food Chem. 44:2517–2522. 14 Dickinson, E., R. H. Whyman, and D. G. Dalgleish. 1987. Colloidal properties of model oil-in-water food emulsions stabilized separately by as1-casein, b-casein and k-casein. Pages 40–51 in Food Emulsions and Foams. E. Dickinson, ed. R. Soc. Chem., London, England. 15 Donnelly, W. J., G. P. McNeill, W. Buchheim, and T.C.A. McGann. 1984. A comprehensive study of the relationship between size and protein composition in natural bovine casein micelles. Biochim. Biophys. Acta 789:136–143. 16 Fang, Y., and D. G. Dalgleish. 1993. Dimensions of the adsorbed layers in oil in water emulsions stabilized by caseins. J. Colloid Interface Sci. 156:329–334. 17 Griffin, M.C.A. 1987. Structural studies of casein micelles using photon correlation spectroscopy. J. Colloid Interface Sci. 115: 499–506. 18 Griffin, M.C.A., and G.C.K. Roberts. 1985. A 1H-NMR study of casein micelles. Biochem. J. 228:273–276. 19 Holt, C. 1992. Structure and stability of bovine casein micelles. Adv. Prot. Chem. 43:63–151. 20 Holt, C., and D. G. Dalgleish. 1986. Electrophoretic and hydrodynamic properties of bovine casein micelles interpreted in terms of particles with an outer hairy layer. J. Colloid Interface Sci. 114:513–524. 21 Holt, C., D. T. Davies, and A.J.R. Law. 1986. Effects of colloidal calcium phosphate content and free calcium ion concentration in the milk serum on the dissociation of bovine casein micelles. J. Dairy Res. 53:557–572. 22 Holt, C., and D. S. Horne. 1996. The hairy casein micelle: evolution of the concept and its implications for dairy technology. Neth. Milk Dairy J. 50:85–111. 23 Holt, C., and M.J.J.M. van Kemenade. 1989. The interaction of phosphoproteins with calcium phosphate. Pages 175–213 in Calcified Tissue. D.W.L. Hukins, ed. Macmillan, London, England. 24 Horne, D. S. 1984. Steric effects in the coagulation of casein micelles by ethanol. Biopolymers 23:989–993. 25 Horne, D. S., and C. M. Davison. 1986. The effect of environmental conditions on the steric stabilization of casein micelles. Colloid Polym. Sci. 264:727–734. 26 Horne, D. S., and C. M. Davidson. 1993. Direct observation of decrease in size of casein micelles during the initial stage of renneting of skim milk. Int. Dairy J. 3:61–71. 27 Jang, H. D., and H. E. Swaisgood. 1990. Disulfide bond formation between thermally denatured b-lactoglobulin and k-casein in casein micelles. J. Dairy Sci. 73:900–904. 28 Leaver, J., D. Brooksbank, and D. S. Horne. 1994. Influence of glycosylation and surface charge on the binding of k-casein to latex particles. J. Colloid Interface Sci. 162:463–469. 29 Leaver, J., and D. G. Dalgleish. 1990. Topography of b-casein adsorbed at an oil/water interface as determined from the kinetics of trypsin-catalysed hydrolysis. Biochim. Biophys. Acta 1041:217–222. 30 Leaver, J., and D. G. Dalgleish. 1992. Variations in the binding of b-casein to oil-water interfaces detected by trypsin-catalysed hydrolysis. J. Colloid Interface Sci. 149:49–55. Journal of Dairy Science Vol. 81, No. 11, 1998
31 Leaver, J., and G. Thomson. 1993. Influence of heating and cooling on the trypsinolysis of caseins in bovine milk. Milchwissenschaft 48:378–381. 32 Leermakers, F.A.M., P. J. Atkinson, E. Dickinson, and D. S. Horne. 1996. Self-consistent-field modeling of adsorbed bcasein: effects of pH and ionic strength on surface coverage and density profile. J. Colloid Interface Sci. 178:681–693. 33 Plowman, J. E., and L. K. Creamer. 1995. Restrained molecular dynamics study of the interaction between bovine k-casein peptide 98–111 and bovine chymosin and porcine pepsin. J. Dairy Res. 62:451–467. 34 Rasmussen, L. K., P. Hojrup, and T. E. Petersen. 1992. The multimeric structure and disulfide-bonding pattern of bovine kcasein. Eur. J. Biochem. 207:215–222.
Journal of Dairy Science Vol. 81, No. 11, 1998
35 Rollema, H. S., and J. A. Brinkhuis. 1989. A 1H-NMR study of bovine casein micelles; influence of pH, temperature and calcium ions on micellar structure. J. Dairy Res. 56:417–425. 36 Rollema, H. S., J. A. Brinkhuis, and H. J. Vreeman. 1988. 1H-NMR studies of bovine k-casein and casein micelles. Neth. Milk Dairy J. 42:233–248. 37 Schmidt, D. G. 1982. Electron microscopy of milk and milk products: problems and possibilities. Food Microstruct. 1: 151–165. 38 Talbot, B., and D. F. Waugh. 1970. Micelle-forming characteristics of monomeric and covalent polymeric k-caseins. Biochemistry 9:2807–2813. 39 Walstra, P., V. A. Bloomfield, G. J. Wei, and R. Jenness. 1981. Effect of chymosin action on the hydrodynamic diameter of casein micelles. Biochim. Biophys. Acta 669:258–259.