- Email: [email protected]
A comprehensive study of the relationship between size and protein composition in natural bovine casein micelles
Biochimica et Biophvsica Acta, 789 (1984) 136 143 Elsevier
136
BBA 31988
A C O M P R E H E N S I V E STUDY OF T H E R E L A T I O N S H I P B E T W E E N SIZE AND P R O T E I N C O M P O S I T I O N IN NATURAL BOVINE CASEIN M I C E L L E S W.J. D O N N E L L Y a, G.P. M c N E I L L a W. B U C H H E I M b and T.C.A. M c G A N N a*
" Agricultural Institute, Moorepark Research Centre, Fermoy, Co. Cork (Republic of Ireland) and h Federal Dairy Research Centre, Kiel (F.R.G.) (Received March 12th, 1984)
Key words: Protein compositton," Micelle size," Casein
Casein micelles of bovine skimmed milk were fractionated by permeation chromatography on porous glass (CPG-10, 50 nm followed by CPG-10, 300 nm) at 30°C. Micelles were pooled in eight eluant fractions and their size distribution was determined by electron microscopy. The composition of casein in the eight fractions was determined by quantitative hydroxyapatite chromatography. Micelle size decreased progressively with increasing elution volume, and volume-to-surface average diameter ranged from 154 nm in fraction 1 to 62 nm in fraction 8. Concurrently there was a decrease in relative proportions of a s- and /]-caseins and a large enrichment of K-casein, which changed from 4.1% total casein in fraction 1 to 12.1% total casein in fraction 8. At least half the decrease in a~-casein proportions was attributed to the a~-casein component, but the data also suggested a decline in proportions of as2-casein in the smallest micelle fractions. A plot of K-casein fractional content versus micelle surface-to-volume ratio gave a straight line (correlation coefficient from linear regression 0.98) from which an average K-casein surface coverage of 1.5 m 2 / m g or 47.3 nm2/molecule was obtained. If a constant surface coverage for K-casein is assumed, the parameters of the linear equation predict that micelle voluminosity is inversely related to micelle diameter, being approximately 30% larger in fraction 8 compared to fraction 1.
Introduction
The phosphoprotein (casein) of bovine milk is aggregated into colloidal particles called micelles, ranging in diameter from about 30 nm to greater than 600 nm. Precipitation by calcium, which would otherwise occur at the high calcium concentration of milk, is thereby prevented. The structure of the casein micelle has been the subject of much controversy, but there is now considerable evidence in support of the role of x-casein as colloidal protecting agent through its location at the micelle surface (see, for example, Schmidt [1] for a recent review). While there have been many * Deceased. 0167-4838/84/$03.00 © 1984 Elsevier Science Publishers B.V.
different approaches to the investigation of micelle structure, one of the most informative has been measurement of casein composition in micelle fractions of different size distribution as a means of distinguishing between inner and outer locations for individual caseins. Recently this approach has been refined by developments both in casein compositional analysis and in methods for size-fractionation of milk micelles. In one such study [2] clear differences in casein composition were found between micelle fractions prepared by ultracentrifugation, but no actual size measurements were carried out. A preliminary attempt at developing a precise size-composition relationship was reported by us using permeation chromatography on controlled pore glass for micelle frac-
137 tionation and electron microscopy for size measurement [3]. In order to realise fully the potential of the latter approach, a number of improvements were made in methodology, notably in the procedure for permeation chromatography. These revised techniques have now been applied to an extension of the previous investigations and the present report describes an attempt to define, in comprehensive terms, the relationships between size and casein composition in natural micelles. Materials and M e t h o d s
circumference (C) and the area (A) enclosed by that circumference. For calculation of three-dimensional distributions, this area was assumed to be a circle with d = ~/(4/~r)A. Each particle was then assigned to a size class with a diameter range of 30 nm and relative volume frequencies were calculated as previously described [3,19]. Surface-to-volume ratios (S,,) were calculated from [21] so
4 EC qr 2,4
Micelle fractionation
with volume-to-surface average diameter, d,,, =
Casein micelles were separated into eight size fractions by permeation chromatography on controlled pore glass (CPG-10, 300 nm) using a modification of the procedure described in a previous report [6]. Samples (15 ml) of skim-milk from Friesian herds were chromatographed on a column (90 x 1.6 cm) of CPG-10, 50 nm immediately before fractionation on CPG-10, 300 nm to remove whey protein and non-micellar casein from micellar casein. The micelles were eluted at 30 o C with synthetic milk serum (pH 6.7) containing 0.04% poly(ethylene glycol), at a flow rate of 1.6 ml/min. Column fractions (5 ml) were pooled to give eight bulk fractions of 10 or 15 ml and a portion of each fraction was fixed immediately with glutaraldehyde. Fixation was by dropwise addition of 25% glutaraldehyde with stirring at 20 °C to a final concentration of 1% (v/v). The original milk pH was maintained by simultaneous alkali addition using an automatic titrator. Stirring was continued for 1 h, by which time no further alkali uptake was detectable. Excess glutaraldehyde was removed by overnight dialysis at 4 ° C against water.
b/S~.
Electron microscopy Glutaraldehyde fixed micelle fractions were concentrated 10- to 15-fold in a B15 Minicon concentrator (Amicon Corp., MA, U.S.A.). Freeze fracturing and electron microscopy was carried out at the Federal Dairy Research Centre, Kiel, F.R.G. as described by Schmidt et al. [19]. Micelle diameters were calculated from micrographs with the aid of an Apple microcomputer and graphics tablet. As each particle cross-sectional area was outlined by the graphics tablet, the computer measured the
Duplicate So values were calculated by dividing micrographs randomly into two groups of equal number. Standard deviations were estimated from duplicates. Particles of less than 20 nm diameter were not included in computation of Sv values.
Casein composition Whole casein was precipitated from micelle fractions within 2 h of fractionation by adjusting the pH to 4.6 with 5 M HC1. The precipitates were suspended in water and redissolved at about pH 7.0 by slow addition of 0.1 M N a O H with stirring. The sodium caseinates were stored at - 2 0 ° C as freeze-dried powders. Casein compositional analysis was conducted by fractionation of whole casein on hydroxyapatite followed by quantitative analysis of protein in individual column fractions as previously described [4]. Each sample was analysed in duplicate or triplicate. Proportions of K-casein were corrected for the presence of chymosin-resistant contaminant proteins in the hydroxyapatite column fraction as previously described [5]. One correction factor was obtained for each micelle fraction and for skim-milk. The correction factor decreased in an approximately linear manner as micelle fraction number increased (correlation coefficient for linear regression = 0.91) and in order to minimise error in estimation of true chymosin-sensitive x-casein a new correction factor for each fraction was taken from the regression line.
Electrophoresis Electrophoresis was carried out on polyacrylamide gels essentially by the method of Davies
138
and Law [22]. Gels were analysed on a DCD-16 densitometer (Gelman Instrument Co., Ann Arbor, MI, USA).
firmed by examination of micelle size distributions (Fig. 1), which show a gradual shift in distribution maximum from fraction 1 to fraction 8.
Results
Permeation chromatography Fractionation of skimmilk by permeation chromatography on a dual porous glass column system (CPG-10, 50 nm followed by CPG-10, 300 nm) results in complete separation of micellar from non-micellar casein [24] in contrast to the previously described procedure [6], using CPG-10, 300 nm alone, when partial overlap of micellar and non-micellar peaks occurred. The total micellar protein in the column eluant was pooled into eight fractions of 10-15 ml which were subjected to examination by electron microscopy and to casein compositional analysis by quantitative column chromatography on hydroxyapatite.
Casein composition Table II shows data for casein composition in the eight column fractions and in the original skim-milk. A sharp enrichment of x-casein occurred with decreasing micelle size, compensated by a gradual drop in both a s- and fl-casein. A possible small rise in y-casein with decreasing size is also evident. The extent of K-casein enrichment is emphasised by the change in c~/K ratio from 12.2 in fraction 1 to 3.8 in fraction 8. Agreement between the measured casein composition of the original skim-milk and the composition computed by weighted summation of results for individual fractions (composite skim, Table 1I) was satisfactory and probably well within the limits of accuracy of the measurements.
Micelle size Table I gives details of micelle size data in the eight fractions and in the original skim-milk and Fig. 1 shows corresponding size distributions. A progressive decrease in volume-to-surface average diameter (d,,,) and increase in surface-to-volume ratio (S,,) was observed with increasing elution volume. The measured value for do~ of the original skim-milk (83 nm) can be compared with the range (154-62 nm) in the column fractions. The effectiveness of porous glass fractionation is con-
Micelle dissociation The pattern of behaviour of K- and as-casein across the profile is generally in line with earlier results from this laboratory [6], but the progressive decrease in fl-casein content with decreasing micelle size observed in the present study is opposite to that observed by us previously. Since the chromatographic temperature of the previous report was 20 °C (compared to 30 o C in the present work) it was thought that this may have influenced fl-casein behaviour. In order to check the effect of
TABLE I VOLUME-TO-SURFACE AVERAGE MICELLE DIAMETER (d w) A N D SURFACE-TO-VOLUME RATIO (S,,) OF MICELLE FRACTIONS AFTER CONTROLLED PORE GLASS C H R O M A T O G R A P H Y (CPG-10, 50 NM FOLLOWED BY CPG-10, 300 nm) OF SKIM-MILK Sample CPG-10 fraction 1 2 3 4 5 6 7 8 Skim
Mean elution volume (ml)
Fraction of total micellar casein)
d,., (nm)
S,, ( + S.D.) (nm 1)( × 10 3)
Particles counted
230.0 245.0 257.5 267.5 277.5 287.5 300.0 315.0 -
0.10 0.11 0.13 0.15 0.16 0.15 0.12 0.08 -
154 109 103 91 82 74 71 62 83
39 (0.4) 55 (5.0) 58 (2.0) 66 (1.0) 73 (2.0) 81 (3.0) 84 (3.0) 97 (3.0) 72 (0.1)
630 1 160 1253 1 126 1004 1 531 3 370 3 417 2755
139
(nm) 0 90 180 270 360
temperature on chromatographic behaviour, skim-milk was fractionated at 20 °C and 37 o C on the dual-column system and casein composition was determined in the column fractions. Results for fl-casein proportions at 2 0 ° C and 3 7 ° C are compared with those obtained at 3 0 ° C in Fig. 2. In all cases, change in fl-casein accross the profile was approximately linear with fraction number (correlation coefficients from linear regression analysis were 0.93, 0.98 and 0.98 at 20, 30 and 37 °C, respectively). The slopes of the linear regressions were - 0.57, - 0.99 and - 1.18 at 20, 30 and 37 o C, respectively, showing that the drop in fl-casein proportion accross the profile was much less rapid at 2 0 ° C than at the higher temperatures. Plots of fl + 7x3-casein are also shown in Fig. 2. Inclusion of y-caseins should be appropriate, since they are fragments of fl-casein produced by plasmin-catalysed proteolysis, which might be expected to show analogous associationdissociation behaviour during chromatography and to occupy the same sites in the micelle. In addition, while a plasmin inhibitor was added to skimmilks before chromatography, it is possible that small differences in 3,2.3-casein proportions in individual fractions (Table II) are the result of slight plasmin action during or immediately following fractionation. Once again, the slopes of the regres-
Micelle diometer"
0 90 1BO 270 360
t
,
0.3~"~
,,.
0
,
2
~
0.10~ O. ~
O2
O.
o, -L o
(3
3
o.1
G
o o.-"
4
0.2 0.1t
c
Fig. 1. Three-dimensional micelle size distributions in controlled-pore glass column fractions of skim-milk. Size classes have a diameter range of 30 nm.
GTABLE II
CASEIN COMPOSITION OF SKIM-MILK FRACTIONS OBTAINED BY PERMEATION CHROMATOGRAPHY ON CONTROLLED-PORE GLASS AT 30 o C Coefficients of variation between duplicate or triplicate determinations (%): a s, 2.5; fl, 3.0; x, 2.9; 3'2,3, 16.3; other, not determined; 'other' refers to minor components of casein present in the r-casein fraction from bydroxyapatite, n.d, not determined. Casein composition
Fraction No.
(% total casein)
1
2
3
4
5
6
7
8
% fl
50.1 36.8 4.1 5.3 3.7 42.1
48.8 35.3 5.0 6.7 4.2 42.0
49.1 34.4 6.4 5.3 4.9 39.7
48.5 34.2 7.0 5.5 4.8 39.7
46.6 33.9 8.2 6.2 5.1 40.1
46.2 32.4 9.6 6,3 5.4 38.7
45.6 31.1 11.1 6.6 5.6 37.7
45.7 29.6 12.1 7.1 5.5 36.7
1,2 12.2 2.1
1.2 9.8 2.3
1.2 7.7 2.2
1.2 6.9 2.3
1.2 5.7 1.9
1.2 4.8 2.3
1.2 4.1 2.6
r
Y2,3 Other
/3 + 3'2,3
Skim 46.0 32.4 8.4 7.9 5.4 38.5
Ratios (g/g)
°ts/fl + 3'2.3
%/r
exsl/%2
Composite skim
1.25 3.8 3.0
1.2 5.5 n.d.
47.6 33.6 7.9 6.1 4.9 39.7
140 5O
%
T
0
I
i
I
I
I
i
I
I
1
2
3
4
5
6
•
8
Fraction
No.
Fig. 2. Proportional content of fl-casein (open symbols) and fl + y2.3-casein (closed symbols) in fractions of skim-milk obtaincd by permeation chromatography on controlled-pore glass at 20 (O, O), 30 (zx A) and 37(D,I) °C.
sion lines ( - 0.34, - 0 . 8 0 and - 0 . 9 8 at 20, 30 and 37°C), respectively, show a much slower drop in /3 + Y2,3-case'n proportions accross the profile at 20 o C than at the higher temperatures. It therefore appears that the entropy-driven dissociation of these proteins from the micelle is not entirely suppressed at 20 °C, as was previously thought to occur. This partly explains differences in results for fl-casein behaviour in the previous work. A further major factor contributing to the previous results was, undoubtedly, inadequate separation of non-micellar from micellar casein in the single column system for micelle fractionation previously employed, a feature that would tend to elevate fl-casein levels in the smallest-size fractions. The small further increase in slopes of the regression lines at 37 ° C c o m p a r e d to 3 0 ° C (Fig. 2) suggests that even at 3 0 ° C dissociation is not entirely suppressed. However, it is not certain that this small difference is significant in view of the small n u m b e r of data points at 37 ° C, and the constancy of the % / f l + "t2,3 ratio across the profile (Table I|) provides evidenge that little dissociation of f l + y-casein occurred at 3 0 ° C . In contrast, this ratio decreased with increasing elution volume at 20 ° C (data not shown). as, -caseins Separate quantification of %a- and as2-caseins, the components of a~-casein, was not possible with
the chromatographic procedure employed for compositional analysis. However, by means of electrophoresis/densitometry approximate values for O~sl/as2 ratios were obtained for the eight column fractions (Table II). These show a slightly erratic pattern, a fact that can be attributed to the inaccuracies of densitometry. Otherwise the data show a possible increase in the a l / % 2 ratio in fractions 7 and 8, corresponding to a relative decrease in the %z-casein content of the smallest micelles. At most, this increase would account for one-half of the decrease in total a~-casein. N o difference between fractions in relative amounts of individual a~zcasein components (i.e., a~2_6 ) was detectable by densitometry. x-casein heterogeneity Electrophoresis/densitometry was also applied to a study of the x-casein fraction. Fig. 3 shows the electrophoretic patterns of x-casein prepared by hydroxyapatite c h r o m a t o g r a p h y of casein from each of the porous glass column fractions. N o major difference is evident in the glycosylation pattern of this protein accross the profile and densitometry did not reveal any variation in relative proportions of glycosylated and unglycosylated forms.
~,*~:! <<~L ~
~ ~<>~!~!~
~¸ ~ < ~ < > ~ , ; !
~;!~
,~
<~,~,,~,i~:~
~:!~!~!~!~>~'~#~i~
i iii!i!ii!ili!ii!!i!!ili! i~iiii!i~!ii!!!i!!ii~! ii~411i!~i~i~i~~ii!~iiii!i~ii!!ili!i~i~i!i!i~ii!~!!~i~!!]i !i li~i iiiiii!il !!!iiiil~!ii!i~i~i!}~ ¸!~!iii!ilil¸iiiiiiiiii~ii~iiiiiiiiiiii! ii ii!iiiiil Fig. 3. Polyacrylamide gel electrophoresis of x-casein obtained by hydroxyapatite column chromatography of casein from individual controlled-pore glass column fractions of skim-milk. Slots, from left to right, represent fraction 1 to fraction 8, respectively.
141 x-casein
us. S O
A plot of fractional content of x-casein vs. S v for fractions 1 - 8 and skimmed milk approximates to a straight line with a linear regression equation: x-casein = 1.49 S,, -0.024+0.0058
(1)
( C o r r e l a t i o n c o e f f i c i e n t = 0.98)
Discussion
Since the pioneering work of Sullivan et al. [7] which first demonstrated an inverse relationship between micelle size and x-casein content, few data of a comprehensive nature on size-compositional relationships in natural micelles have been published. Possibilities for comparison with the present results are further limited by the fact that some previous studies were based on coldfractionation of micelles when extensive dissociation of caseins, in particular fl-casein, presumably occurred and by the fact that in most studies casein compositional analysis was carried out by techniques which were probably inferior to quantitative column chromatography. Detailed comparisons will therefore be confined to the recent results of Davies and Law [2] who prepared four micelle fractions by ultracentrifugation of skim-milk at 20 o C and examined casein composition by ion-exchange and gel chromatography. The expected enrichment of x-casein with decreasing micelle size was observed by these workers, but the x-casein proportion in the smallest micelle fraction (16.3% of total casein) was much higher than that observed by us. Their x-casein content for skim-milk (11.2%) was also much higher than the value of 8.4% recorded in the present study, reflecting some important difference between groups in methods for r-casein analysis. These workers reported a decrease in proportions of flcasein of the same magnitude as that observed by us, indicating that no appreciable dissociation of fl-casein from the micelle occurred during ultracentrifugation at 20°C. This may reflect the absence of a dilution effect during ultracentrifugation compared to permeation chromatography. Alternatively, any dissociation which occurred during centrifugation was masked by the fact that it resulted in transfer of fl-casein to the supernatant rather than to successive micelle fractions. In both
studies a small drop was found in as2-casein proportions with decreasing micelle size, but the observation by Davies and Law that the proportion of asl-casein does not change with micelle size is not supported by the present results, which indicate a gradual decrease in this component with decreasing size. In neither study was any major difference in glycosylation pattern of r-casein apparent by electrophoretic examination. This is surprising in view of observations that the degree of x-casein glycosylation is a function of micelle size [8,9]. The present results for casein composition differ substantially from those of Yoshikawa et al. [8], who also fractionated skim-milk micelles by permeation chromatography on porous glass at 2 5 ° C and determined casein composition in the micelle fractions by electrophoresis/densitometry. These workers recorded a much steeper drop in a~-casein proportions, and a smaller increase in x-casein proportions, with increasing elution volume than that recorded here. Furthermore, in contrast to the present results a substantial increase in proportions of fl-casein with increasing elution volume was recorded. These results are consistent with the occurrence of considerable dissociation of micellar fl-casein during chromatography. This might have arisen because of the slow flow-rate used by these workers during chromatography which gave residence times that were 10times longer than those in the present study. Dissociation of fl-casein from the micelle is also the most likely explanation for the results of another study [23] which revealed a large increase in proportions of fl-casein with increasing elution volume in micelle fractions obtained by porous glass chromatography of skimmed milk at 4 o C. The clear linear relationship between K-casein fractional content and S o confirms and extends previous conclusions for natural micelles and for laboratory assembled micelles [3]. The results also agree with many general observations in the past of an inverse relationship between x-casein content and micelle size. The main contribution of the present investigation is to quantify that relationship in a manner that would allow more detailed conclusions to be made about micelle structure. The average surface coverage of x-casein (o) in skim-milk may be obtained from the relationship:
142
o = ~-casein
(2)
where V is micelle voluminosity, r-casein is expressed as fractional content of total casein and all r-casein is located at the micelle surface. At an average value of 0.095 for r-casein in bulk milks [5] (giving S,,=0.080 nm -1 from Eqn. 1) and V = 1.8 cm3/g [16], a value of o = 1.5 m2/mg or 47.3 nmZ/molecule is obtained. This value does not taken into consideration the possibility of micelle shrinkage during glutaraldehyde fixation for electron microscopy. Any such shrinkage would mean a corresponding reduction in o for the natural unfixed micelles. No inference can be drawn from these data about the orientation of surface r-casein, i.e. whether it exists as a monolayer occupying parallel to the surface, or whether it protrudes vertically from the surface as postulated by Walstra and co-workers [17,18]. From Eqns. 1 and 2. S:,- V o
= ~s,, + ~
(3)
At constant o, then: v = c + d/S,.
(4)
At constant V, then: o
1
f + e/S,,
(5)
where a - f are constants. Since d and e are negative, Eqn. 4 predicts an increase in micelle voluminosity with decreasing size, while Eqn. 5 predicts a decrease in r-casein surface coverage with decreasing micelle size. If neither V nor o is constant, then these parameters vary inversely with micelle size, but as a working hypothesis the situation described by Eqn. 4 is favoured for the following reasons: (a) higher voluminosity of small micelles ~a consistent with experimental data for artificial micelles [11] and with the observation that hydration and translucence of slowly sedimenting natural micelles is increased [20]; and (b) a constant surface coverage of K-casein is consistent with the existence of a dynamic equilibrium between surface submicelles and supernatant as advocated by Slattery [12]. Waugh and Noble [11]
found a gradual change in micelle voluminosity by ultracentrifugation of calcium aggregated suspensions of a~ and r-casein with different relative proportions. When the fractional content of Kcasein increased from 0.048 to 0.11, voluminosity increased by about 40%. A similar change in Kcasein proportion occurs from fraction 1 to fraction 8 of the present study, and from Eqns. 1 and 2 the corresponding change in voluminosity is calculated to be an increase of 34%, which is in excellent agreement with the experimental data for artificial micelles. The value calculated above for o is not invariant but applies to the equilibrium position in a skim-milk of defined x-casein and colloidal calcium phosphate content. A new value for a would be reached if the equilibrium is disturbed. Examples are the redistribution of micelle size which occurs during storage of unfixed micelle fractions after permeation chromatography (unpublished observation) or which occurs after partial removal of colloidal calcium phosphate from skim-milk [13]. The relationship between variable o in these systems and such micelle properties as rennetability would be worthy of investigation. The above computations are based on the assumption that the x-casein content of the micelle interior is negligible. While a low value for internal K-casein is indicated by the data of Eqn. 1, extrapolation to S,, = 0 is not possible because of the influence of variable voluminosity on the parameters of the r / S , , equation. However, direct evidence from electron microscopy of micelles containing gold-labelled caseins [14] places r-casein almost exclusively on the micelle surface. The fact that the a J f l + y ratio is independent of micelle size indicates that these components are distributed in an identical way throughout the micelle. Yet the ability of much of the fl-casein to migrate in and out of the micelle without apparently affecting micelle structure [15] appears to indicate a non-essential structural role for a major fraction of this protein. We postulate that in vivo formation of micelles is predominantly the result of deposition of amorphous calcium phosphate on c~s-casein or %/r-casein complexes and that flcasein associates with the primary aggregate only partly through calcium phosphate linkage and mainly through hydrophobic bonding to a~-casein. No satisfactory analogy has been found in bio-
143
logical systems for the process of colloidal stabilisation of milk phosphoproteins. Indeed, the word micelle itself is, by conventional meaning, a misnomer for the casein aggregates and might be more properly applied to describe casein submicelles. The manner in which these submicelles are further aggregated into polymers whose size distribution is controlled by the presence of a surface-active protein component may yet prove unique to milk.
Acknowledgement G.P.M. was a graduate student of the Biochemistry Department, University College, Galway, during this study.
References 1 Schmidt, D.G. (1982) in Developments in Dairy Chemistry, Vol. 1 (Fox, P.F., ed.), pp. 61-86, Applied Science Publishers Ltd., Barking 2 Davies, D.T. and Law, A.J.R. (1983) J. Dairy Res. 50, 67-75 3 McGann, T.C.A., Donnelly, W.J., Kearney, R.D. and Buchheim, W. (1980) Biochim. Biophys. Acta 630, 261-270 4 Barry, J.G. and Donnelly, W.J. (1981) J. Dairy Res. 48, 437-446 5 Barry, J.G. and Donnelly, W.J. (1980) J. Dairy Res. 47, 71-82 6 McGann, T.C.A., Kearney, R.D. and Donnelly, W.J. (1979) J. Dairy Res. 46, 307-311
7 Sullivan, R.A., Fitzpatrick, M.M. and Stanton, E.K. (1959) Nature 183, 616-617 8 Yoshikawa, M., Takeuchi, M., Sasaki, R. and Chiba, H. (1982) Agric. Biol. Chem. 46, 1043-1048 9 Slattery, C.W. (1978) Biochemistry 17, 1100-1104 10 Schmidt, D.G., Koops, J. and Westerbeek, D. (1977) Neth. Milk Dairy J. 31, 328-341 11 Waugh, D.F. and Noble, R.W. (1965) J. Am. Chem. Soc. 87, 2246-2257 12 Slattery, C.W. (1977) Biophys. Chem. 6, 59-64 13 McGann, T.C.A. and Pyne, G.T. (1960) J. Dairy Res. 27, 403-417 14 Schmidt, D.G. and Both, P. (1982) Milchwissenschaft 37, 336-337 15 Downey, W.K. (1973) Neth. Milk Dairy J. 27, 218-219 16 Buchheim, W. and Prokopek, D. (1976) Milchwissenschaft 31,462-465 17 Walstra, P. (1979) J. Dairy Res. 46, 317-323 18 Walstra, P., Bloomfield, V.A., Wei, G.J. and Jenness, R. (1981) Biochim. Biophys. Acta 669, 258-259 19 Schmidt, D.G., Walstra, P. and Buchheim, W. (1973) Neth. Milk Dairy J. 27, 128-142 20 Morr, C.V. (1973) J. Dairy Sci. 56, 544-552 21 Weibel, E.R. (1973) in Principals and Techniques of Electron Microscopy, Vol. 3 (Hayat, M.A., ed.), pp. 239-310, Von Nostrand-Reinhold, New York 22 Davies, D.T. and Law, A.J.R. (1977) J. Dairy Res. 47, 83-90 23 Ekstrand, B. and Larsson-Raznikiewicz, M. (1978) Biochim. Biophys. Acta 536, 1-9 24 McNeill, G.P., Donnelly, W.J. and McGann, T.C.A. (1981) Research Report, Food Science and Technology, pp. 14-15, An Foras Taluntais, Dublin
136
BBA 31988
A C O M P R E H E N S I V E STUDY OF T H E R E L A T I O N S H I P B E T W E E N SIZE AND P R O T E I N C O M P O S I T I O N IN NATURAL BOVINE CASEIN M I C E L L E S W.J. D O N N E L L Y a, G.P. M c N E I L L a W. B U C H H E I M b and T.C.A. M c G A N N a*
" Agricultural Institute, Moorepark Research Centre, Fermoy, Co. Cork (Republic of Ireland) and h Federal Dairy Research Centre, Kiel (F.R.G.) (Received March 12th, 1984)
Key words: Protein compositton," Micelle size," Casein
Casein micelles of bovine skimmed milk were fractionated by permeation chromatography on porous glass (CPG-10, 50 nm followed by CPG-10, 300 nm) at 30°C. Micelles were pooled in eight eluant fractions and their size distribution was determined by electron microscopy. The composition of casein in the eight fractions was determined by quantitative hydroxyapatite chromatography. Micelle size decreased progressively with increasing elution volume, and volume-to-surface average diameter ranged from 154 nm in fraction 1 to 62 nm in fraction 8. Concurrently there was a decrease in relative proportions of a s- and /]-caseins and a large enrichment of K-casein, which changed from 4.1% total casein in fraction 1 to 12.1% total casein in fraction 8. At least half the decrease in a~-casein proportions was attributed to the a~-casein component, but the data also suggested a decline in proportions of as2-casein in the smallest micelle fractions. A plot of K-casein fractional content versus micelle surface-to-volume ratio gave a straight line (correlation coefficient from linear regression 0.98) from which an average K-casein surface coverage of 1.5 m 2 / m g or 47.3 nm2/molecule was obtained. If a constant surface coverage for K-casein is assumed, the parameters of the linear equation predict that micelle voluminosity is inversely related to micelle diameter, being approximately 30% larger in fraction 8 compared to fraction 1.
Introduction
The phosphoprotein (casein) of bovine milk is aggregated into colloidal particles called micelles, ranging in diameter from about 30 nm to greater than 600 nm. Precipitation by calcium, which would otherwise occur at the high calcium concentration of milk, is thereby prevented. The structure of the casein micelle has been the subject of much controversy, but there is now considerable evidence in support of the role of x-casein as colloidal protecting agent through its location at the micelle surface (see, for example, Schmidt [1] for a recent review). While there have been many * Deceased. 0167-4838/84/$03.00 © 1984 Elsevier Science Publishers B.V.
different approaches to the investigation of micelle structure, one of the most informative has been measurement of casein composition in micelle fractions of different size distribution as a means of distinguishing between inner and outer locations for individual caseins. Recently this approach has been refined by developments both in casein compositional analysis and in methods for size-fractionation of milk micelles. In one such study [2] clear differences in casein composition were found between micelle fractions prepared by ultracentrifugation, but no actual size measurements were carried out. A preliminary attempt at developing a precise size-composition relationship was reported by us using permeation chromatography on controlled pore glass for micelle frac-
137 tionation and electron microscopy for size measurement [3]. In order to realise fully the potential of the latter approach, a number of improvements were made in methodology, notably in the procedure for permeation chromatography. These revised techniques have now been applied to an extension of the previous investigations and the present report describes an attempt to define, in comprehensive terms, the relationships between size and casein composition in natural micelles. Materials and M e t h o d s
circumference (C) and the area (A) enclosed by that circumference. For calculation of three-dimensional distributions, this area was assumed to be a circle with d = ~/(4/~r)A. Each particle was then assigned to a size class with a diameter range of 30 nm and relative volume frequencies were calculated as previously described [3,19]. Surface-to-volume ratios (S,,) were calculated from [21] so
4 EC qr 2,4
Micelle fractionation
with volume-to-surface average diameter, d,,, =
Casein micelles were separated into eight size fractions by permeation chromatography on controlled pore glass (CPG-10, 300 nm) using a modification of the procedure described in a previous report [6]. Samples (15 ml) of skim-milk from Friesian herds were chromatographed on a column (90 x 1.6 cm) of CPG-10, 50 nm immediately before fractionation on CPG-10, 300 nm to remove whey protein and non-micellar casein from micellar casein. The micelles were eluted at 30 o C with synthetic milk serum (pH 6.7) containing 0.04% poly(ethylene glycol), at a flow rate of 1.6 ml/min. Column fractions (5 ml) were pooled to give eight bulk fractions of 10 or 15 ml and a portion of each fraction was fixed immediately with glutaraldehyde. Fixation was by dropwise addition of 25% glutaraldehyde with stirring at 20 °C to a final concentration of 1% (v/v). The original milk pH was maintained by simultaneous alkali addition using an automatic titrator. Stirring was continued for 1 h, by which time no further alkali uptake was detectable. Excess glutaraldehyde was removed by overnight dialysis at 4 ° C against water.
b/S~.
Electron microscopy Glutaraldehyde fixed micelle fractions were concentrated 10- to 15-fold in a B15 Minicon concentrator (Amicon Corp., MA, U.S.A.). Freeze fracturing and electron microscopy was carried out at the Federal Dairy Research Centre, Kiel, F.R.G. as described by Schmidt et al. [19]. Micelle diameters were calculated from micrographs with the aid of an Apple microcomputer and graphics tablet. As each particle cross-sectional area was outlined by the graphics tablet, the computer measured the
Duplicate So values were calculated by dividing micrographs randomly into two groups of equal number. Standard deviations were estimated from duplicates. Particles of less than 20 nm diameter were not included in computation of Sv values.
Casein composition Whole casein was precipitated from micelle fractions within 2 h of fractionation by adjusting the pH to 4.6 with 5 M HC1. The precipitates were suspended in water and redissolved at about pH 7.0 by slow addition of 0.1 M N a O H with stirring. The sodium caseinates were stored at - 2 0 ° C as freeze-dried powders. Casein compositional analysis was conducted by fractionation of whole casein on hydroxyapatite followed by quantitative analysis of protein in individual column fractions as previously described [4]. Each sample was analysed in duplicate or triplicate. Proportions of K-casein were corrected for the presence of chymosin-resistant contaminant proteins in the hydroxyapatite column fraction as previously described [5]. One correction factor was obtained for each micelle fraction and for skim-milk. The correction factor decreased in an approximately linear manner as micelle fraction number increased (correlation coefficient for linear regression = 0.91) and in order to minimise error in estimation of true chymosin-sensitive x-casein a new correction factor for each fraction was taken from the regression line.
Electrophoresis Electrophoresis was carried out on polyacrylamide gels essentially by the method of Davies
138
and Law [22]. Gels were analysed on a DCD-16 densitometer (Gelman Instrument Co., Ann Arbor, MI, USA).
firmed by examination of micelle size distributions (Fig. 1), which show a gradual shift in distribution maximum from fraction 1 to fraction 8.
Results
Permeation chromatography Fractionation of skimmilk by permeation chromatography on a dual porous glass column system (CPG-10, 50 nm followed by CPG-10, 300 nm) results in complete separation of micellar from non-micellar casein [24] in contrast to the previously described procedure [6], using CPG-10, 300 nm alone, when partial overlap of micellar and non-micellar peaks occurred. The total micellar protein in the column eluant was pooled into eight fractions of 10-15 ml which were subjected to examination by electron microscopy and to casein compositional analysis by quantitative column chromatography on hydroxyapatite.
Casein composition Table II shows data for casein composition in the eight column fractions and in the original skim-milk. A sharp enrichment of x-casein occurred with decreasing micelle size, compensated by a gradual drop in both a s- and fl-casein. A possible small rise in y-casein with decreasing size is also evident. The extent of K-casein enrichment is emphasised by the change in c~/K ratio from 12.2 in fraction 1 to 3.8 in fraction 8. Agreement between the measured casein composition of the original skim-milk and the composition computed by weighted summation of results for individual fractions (composite skim, Table 1I) was satisfactory and probably well within the limits of accuracy of the measurements.
Micelle size Table I gives details of micelle size data in the eight fractions and in the original skim-milk and Fig. 1 shows corresponding size distributions. A progressive decrease in volume-to-surface average diameter (d,,,) and increase in surface-to-volume ratio (S,,) was observed with increasing elution volume. The measured value for do~ of the original skim-milk (83 nm) can be compared with the range (154-62 nm) in the column fractions. The effectiveness of porous glass fractionation is con-
Micelle dissociation The pattern of behaviour of K- and as-casein across the profile is generally in line with earlier results from this laboratory [6], but the progressive decrease in fl-casein content with decreasing micelle size observed in the present study is opposite to that observed by us previously. Since the chromatographic temperature of the previous report was 20 °C (compared to 30 o C in the present work) it was thought that this may have influenced fl-casein behaviour. In order to check the effect of
TABLE I VOLUME-TO-SURFACE AVERAGE MICELLE DIAMETER (d w) A N D SURFACE-TO-VOLUME RATIO (S,,) OF MICELLE FRACTIONS AFTER CONTROLLED PORE GLASS C H R O M A T O G R A P H Y (CPG-10, 50 NM FOLLOWED BY CPG-10, 300 nm) OF SKIM-MILK Sample CPG-10 fraction 1 2 3 4 5 6 7 8 Skim
Mean elution volume (ml)
Fraction of total micellar casein)
d,., (nm)
S,, ( + S.D.) (nm 1)( × 10 3)
Particles counted
230.0 245.0 257.5 267.5 277.5 287.5 300.0 315.0 -
0.10 0.11 0.13 0.15 0.16 0.15 0.12 0.08 -
154 109 103 91 82 74 71 62 83
39 (0.4) 55 (5.0) 58 (2.0) 66 (1.0) 73 (2.0) 81 (3.0) 84 (3.0) 97 (3.0) 72 (0.1)
630 1 160 1253 1 126 1004 1 531 3 370 3 417 2755
139
(nm) 0 90 180 270 360
temperature on chromatographic behaviour, skim-milk was fractionated at 20 °C and 37 o C on the dual-column system and casein composition was determined in the column fractions. Results for fl-casein proportions at 2 0 ° C and 3 7 ° C are compared with those obtained at 3 0 ° C in Fig. 2. In all cases, change in fl-casein accross the profile was approximately linear with fraction number (correlation coefficients from linear regression analysis were 0.93, 0.98 and 0.98 at 20, 30 and 37 °C, respectively). The slopes of the linear regressions were - 0.57, - 0.99 and - 1.18 at 20, 30 and 37 o C, respectively, showing that the drop in fl-casein proportion accross the profile was much less rapid at 2 0 ° C than at the higher temperatures. Plots of fl + 7x3-casein are also shown in Fig. 2. Inclusion of y-caseins should be appropriate, since they are fragments of fl-casein produced by plasmin-catalysed proteolysis, which might be expected to show analogous associationdissociation behaviour during chromatography and to occupy the same sites in the micelle. In addition, while a plasmin inhibitor was added to skimmilks before chromatography, it is possible that small differences in 3,2.3-casein proportions in individual fractions (Table II) are the result of slight plasmin action during or immediately following fractionation. Once again, the slopes of the regres-
Micelle diometer"
0 90 1BO 270 360
t
,
0.3~"~
,,.
0
,
2
~
0.10~ O. ~
O2
O.
o, -L o
(3
3
o.1
G
o o.-"
4
0.2 0.1t
c
Fig. 1. Three-dimensional micelle size distributions in controlled-pore glass column fractions of skim-milk. Size classes have a diameter range of 30 nm.
GTABLE II
CASEIN COMPOSITION OF SKIM-MILK FRACTIONS OBTAINED BY PERMEATION CHROMATOGRAPHY ON CONTROLLED-PORE GLASS AT 30 o C Coefficients of variation between duplicate or triplicate determinations (%): a s, 2.5; fl, 3.0; x, 2.9; 3'2,3, 16.3; other, not determined; 'other' refers to minor components of casein present in the r-casein fraction from bydroxyapatite, n.d, not determined. Casein composition
Fraction No.
(% total casein)
1
2
3
4
5
6
7
8
% fl
50.1 36.8 4.1 5.3 3.7 42.1
48.8 35.3 5.0 6.7 4.2 42.0
49.1 34.4 6.4 5.3 4.9 39.7
48.5 34.2 7.0 5.5 4.8 39.7
46.6 33.9 8.2 6.2 5.1 40.1
46.2 32.4 9.6 6,3 5.4 38.7
45.6 31.1 11.1 6.6 5.6 37.7
45.7 29.6 12.1 7.1 5.5 36.7
1,2 12.2 2.1
1.2 9.8 2.3
1.2 7.7 2.2
1.2 6.9 2.3
1.2 5.7 1.9
1.2 4.8 2.3
1.2 4.1 2.6
r
Y2,3 Other
/3 + 3'2,3
Skim 46.0 32.4 8.4 7.9 5.4 38.5
Ratios (g/g)
°ts/fl + 3'2.3
%/r
exsl/%2
Composite skim
1.25 3.8 3.0
1.2 5.5 n.d.
47.6 33.6 7.9 6.1 4.9 39.7
140 5O
%
T
0
I
i
I
I
I
i
I
I
1
2
3
4
5
6
•
8
Fraction
No.
Fig. 2. Proportional content of fl-casein (open symbols) and fl + y2.3-casein (closed symbols) in fractions of skim-milk obtaincd by permeation chromatography on controlled-pore glass at 20 (O, O), 30 (zx A) and 37(D,I) °C.
sion lines ( - 0.34, - 0 . 8 0 and - 0 . 9 8 at 20, 30 and 37°C), respectively, show a much slower drop in /3 + Y2,3-case'n proportions accross the profile at 20 o C than at the higher temperatures. It therefore appears that the entropy-driven dissociation of these proteins from the micelle is not entirely suppressed at 20 °C, as was previously thought to occur. This partly explains differences in results for fl-casein behaviour in the previous work. A further major factor contributing to the previous results was, undoubtedly, inadequate separation of non-micellar from micellar casein in the single column system for micelle fractionation previously employed, a feature that would tend to elevate fl-casein levels in the smallest-size fractions. The small further increase in slopes of the regression lines at 37 ° C c o m p a r e d to 3 0 ° C (Fig. 2) suggests that even at 3 0 ° C dissociation is not entirely suppressed. However, it is not certain that this small difference is significant in view of the small n u m b e r of data points at 37 ° C, and the constancy of the % / f l + "t2,3 ratio across the profile (Table I|) provides evidenge that little dissociation of f l + y-casein occurred at 3 0 ° C . In contrast, this ratio decreased with increasing elution volume at 20 ° C (data not shown). as, -caseins Separate quantification of %a- and as2-caseins, the components of a~-casein, was not possible with
the chromatographic procedure employed for compositional analysis. However, by means of electrophoresis/densitometry approximate values for O~sl/as2 ratios were obtained for the eight column fractions (Table II). These show a slightly erratic pattern, a fact that can be attributed to the inaccuracies of densitometry. Otherwise the data show a possible increase in the a l / % 2 ratio in fractions 7 and 8, corresponding to a relative decrease in the %z-casein content of the smallest micelles. At most, this increase would account for one-half of the decrease in total a~-casein. N o difference between fractions in relative amounts of individual a~zcasein components (i.e., a~2_6 ) was detectable by densitometry. x-casein heterogeneity Electrophoresis/densitometry was also applied to a study of the x-casein fraction. Fig. 3 shows the electrophoretic patterns of x-casein prepared by hydroxyapatite c h r o m a t o g r a p h y of casein from each of the porous glass column fractions. N o major difference is evident in the glycosylation pattern of this protein accross the profile and densitometry did not reveal any variation in relative proportions of glycosylated and unglycosylated forms.
~,*~:! <<~L ~
~ ~<>~!~!~
~¸ ~ < ~ < > ~ , ; !
~;!~
,~
<~,~,,~,i~:~
~:!~!~!~!~>~'~#~i~
i iii!i!ii!ili!ii!!i!!ili! i~iiii!i~!ii!!!i!!ii~! ii~411i!~i~i~i~~ii!~iiii!i~ii!!ili!i~i~i!i!i~ii!~!!~i~!!]i !i li~i iiiiii!il !!!iiiil~!ii!i~i~i!}~ ¸!~!iii!ilil¸iiiiiiiiii~ii~iiiiiiiiiiii! ii ii!iiiiil Fig. 3. Polyacrylamide gel electrophoresis of x-casein obtained by hydroxyapatite column chromatography of casein from individual controlled-pore glass column fractions of skim-milk. Slots, from left to right, represent fraction 1 to fraction 8, respectively.
141 x-casein
us. S O
A plot of fractional content of x-casein vs. S v for fractions 1 - 8 and skimmed milk approximates to a straight line with a linear regression equation: x-casein = 1.49 S,, -0.024+0.0058
(1)
( C o r r e l a t i o n c o e f f i c i e n t = 0.98)
Discussion
Since the pioneering work of Sullivan et al. [7] which first demonstrated an inverse relationship between micelle size and x-casein content, few data of a comprehensive nature on size-compositional relationships in natural micelles have been published. Possibilities for comparison with the present results are further limited by the fact that some previous studies were based on coldfractionation of micelles when extensive dissociation of caseins, in particular fl-casein, presumably occurred and by the fact that in most studies casein compositional analysis was carried out by techniques which were probably inferior to quantitative column chromatography. Detailed comparisons will therefore be confined to the recent results of Davies and Law [2] who prepared four micelle fractions by ultracentrifugation of skim-milk at 20 o C and examined casein composition by ion-exchange and gel chromatography. The expected enrichment of x-casein with decreasing micelle size was observed by these workers, but the x-casein proportion in the smallest micelle fraction (16.3% of total casein) was much higher than that observed by us. Their x-casein content for skim-milk (11.2%) was also much higher than the value of 8.4% recorded in the present study, reflecting some important difference between groups in methods for r-casein analysis. These workers reported a decrease in proportions of flcasein of the same magnitude as that observed by us, indicating that no appreciable dissociation of fl-casein from the micelle occurred during ultracentrifugation at 20°C. This may reflect the absence of a dilution effect during ultracentrifugation compared to permeation chromatography. Alternatively, any dissociation which occurred during centrifugation was masked by the fact that it resulted in transfer of fl-casein to the supernatant rather than to successive micelle fractions. In both
studies a small drop was found in as2-casein proportions with decreasing micelle size, but the observation by Davies and Law that the proportion of asl-casein does not change with micelle size is not supported by the present results, which indicate a gradual decrease in this component with decreasing size. In neither study was any major difference in glycosylation pattern of r-casein apparent by electrophoretic examination. This is surprising in view of observations that the degree of x-casein glycosylation is a function of micelle size [8,9]. The present results for casein composition differ substantially from those of Yoshikawa et al. [8], who also fractionated skim-milk micelles by permeation chromatography on porous glass at 2 5 ° C and determined casein composition in the micelle fractions by electrophoresis/densitometry. These workers recorded a much steeper drop in a~-casein proportions, and a smaller increase in x-casein proportions, with increasing elution volume than that recorded here. Furthermore, in contrast to the present results a substantial increase in proportions of fl-casein with increasing elution volume was recorded. These results are consistent with the occurrence of considerable dissociation of micellar fl-casein during chromatography. This might have arisen because of the slow flow-rate used by these workers during chromatography which gave residence times that were 10times longer than those in the present study. Dissociation of fl-casein from the micelle is also the most likely explanation for the results of another study [23] which revealed a large increase in proportions of fl-casein with increasing elution volume in micelle fractions obtained by porous glass chromatography of skimmed milk at 4 o C. The clear linear relationship between K-casein fractional content and S o confirms and extends previous conclusions for natural micelles and for laboratory assembled micelles [3]. The results also agree with many general observations in the past of an inverse relationship between x-casein content and micelle size. The main contribution of the present investigation is to quantify that relationship in a manner that would allow more detailed conclusions to be made about micelle structure. The average surface coverage of x-casein (o) in skim-milk may be obtained from the relationship:
142
o = ~-casein
(2)
where V is micelle voluminosity, r-casein is expressed as fractional content of total casein and all r-casein is located at the micelle surface. At an average value of 0.095 for r-casein in bulk milks [5] (giving S,,=0.080 nm -1 from Eqn. 1) and V = 1.8 cm3/g [16], a value of o = 1.5 m2/mg or 47.3 nmZ/molecule is obtained. This value does not taken into consideration the possibility of micelle shrinkage during glutaraldehyde fixation for electron microscopy. Any such shrinkage would mean a corresponding reduction in o for the natural unfixed micelles. No inference can be drawn from these data about the orientation of surface r-casein, i.e. whether it exists as a monolayer occupying parallel to the surface, or whether it protrudes vertically from the surface as postulated by Walstra and co-workers [17,18]. From Eqns. 1 and 2. S:,- V o
= ~s,, + ~
(3)
At constant o, then: v = c + d/S,.
(4)
At constant V, then: o
1
f + e/S,,
(5)
where a - f are constants. Since d and e are negative, Eqn. 4 predicts an increase in micelle voluminosity with decreasing size, while Eqn. 5 predicts a decrease in r-casein surface coverage with decreasing micelle size. If neither V nor o is constant, then these parameters vary inversely with micelle size, but as a working hypothesis the situation described by Eqn. 4 is favoured for the following reasons: (a) higher voluminosity of small micelles ~a consistent with experimental data for artificial micelles [11] and with the observation that hydration and translucence of slowly sedimenting natural micelles is increased [20]; and (b) a constant surface coverage of K-casein is consistent with the existence of a dynamic equilibrium between surface submicelles and supernatant as advocated by Slattery [12]. Waugh and Noble [11]
found a gradual change in micelle voluminosity by ultracentrifugation of calcium aggregated suspensions of a~ and r-casein with different relative proportions. When the fractional content of Kcasein increased from 0.048 to 0.11, voluminosity increased by about 40%. A similar change in Kcasein proportion occurs from fraction 1 to fraction 8 of the present study, and from Eqns. 1 and 2 the corresponding change in voluminosity is calculated to be an increase of 34%, which is in excellent agreement with the experimental data for artificial micelles. The value calculated above for o is not invariant but applies to the equilibrium position in a skim-milk of defined x-casein and colloidal calcium phosphate content. A new value for a would be reached if the equilibrium is disturbed. Examples are the redistribution of micelle size which occurs during storage of unfixed micelle fractions after permeation chromatography (unpublished observation) or which occurs after partial removal of colloidal calcium phosphate from skim-milk [13]. The relationship between variable o in these systems and such micelle properties as rennetability would be worthy of investigation. The above computations are based on the assumption that the x-casein content of the micelle interior is negligible. While a low value for internal K-casein is indicated by the data of Eqn. 1, extrapolation to S,, = 0 is not possible because of the influence of variable voluminosity on the parameters of the r / S , , equation. However, direct evidence from electron microscopy of micelles containing gold-labelled caseins [14] places r-casein almost exclusively on the micelle surface. The fact that the a J f l + y ratio is independent of micelle size indicates that these components are distributed in an identical way throughout the micelle. Yet the ability of much of the fl-casein to migrate in and out of the micelle without apparently affecting micelle structure [15] appears to indicate a non-essential structural role for a major fraction of this protein. We postulate that in vivo formation of micelles is predominantly the result of deposition of amorphous calcium phosphate on c~s-casein or %/r-casein complexes and that flcasein associates with the primary aggregate only partly through calcium phosphate linkage and mainly through hydrophobic bonding to a~-casein. No satisfactory analogy has been found in bio-
143
logical systems for the process of colloidal stabilisation of milk phosphoproteins. Indeed, the word micelle itself is, by conventional meaning, a misnomer for the casein aggregates and might be more properly applied to describe casein submicelles. The manner in which these submicelles are further aggregated into polymers whose size distribution is controlled by the presence of a surface-active protein component may yet prove unique to milk.
Acknowledgement G.P.M. was a graduate student of the Biochemistry Department, University College, Galway, during this study.
References 1 Schmidt, D.G. (1982) in Developments in Dairy Chemistry, Vol. 1 (Fox, P.F., ed.), pp. 61-86, Applied Science Publishers Ltd., Barking 2 Davies, D.T. and Law, A.J.R. (1983) J. Dairy Res. 50, 67-75 3 McGann, T.C.A., Donnelly, W.J., Kearney, R.D. and Buchheim, W. (1980) Biochim. Biophys. Acta 630, 261-270 4 Barry, J.G. and Donnelly, W.J. (1981) J. Dairy Res. 48, 437-446 5 Barry, J.G. and Donnelly, W.J. (1980) J. Dairy Res. 47, 71-82 6 McGann, T.C.A., Kearney, R.D. and Donnelly, W.J. (1979) J. Dairy Res. 46, 307-311
7 Sullivan, R.A., Fitzpatrick, M.M. and Stanton, E.K. (1959) Nature 183, 616-617 8 Yoshikawa, M., Takeuchi, M., Sasaki, R. and Chiba, H. (1982) Agric. Biol. Chem. 46, 1043-1048 9 Slattery, C.W. (1978) Biochemistry 17, 1100-1104 10 Schmidt, D.G., Koops, J. and Westerbeek, D. (1977) Neth. Milk Dairy J. 31, 328-341 11 Waugh, D.F. and Noble, R.W. (1965) J. Am. Chem. Soc. 87, 2246-2257 12 Slattery, C.W. (1977) Biophys. Chem. 6, 59-64 13 McGann, T.C.A. and Pyne, G.T. (1960) J. Dairy Res. 27, 403-417 14 Schmidt, D.G. and Both, P. (1982) Milchwissenschaft 37, 336-337 15 Downey, W.K. (1973) Neth. Milk Dairy J. 27, 218-219 16 Buchheim, W. and Prokopek, D. (1976) Milchwissenschaft 31,462-465 17 Walstra, P. (1979) J. Dairy Res. 46, 317-323 18 Walstra, P., Bloomfield, V.A., Wei, G.J. and Jenness, R. (1981) Biochim. Biophys. Acta 669, 258-259 19 Schmidt, D.G., Walstra, P. and Buchheim, W. (1973) Neth. Milk Dairy J. 27, 128-142 20 Morr, C.V. (1973) J. Dairy Sci. 56, 544-552 21 Weibel, E.R. (1973) in Principals and Techniques of Electron Microscopy, Vol. 3 (Hayat, M.A., ed.), pp. 239-310, Von Nostrand-Reinhold, New York 22 Davies, D.T. and Law, A.J.R. (1977) J. Dairy Res. 47, 83-90 23 Ekstrand, B. and Larsson-Raznikiewicz, M. (1978) Biochim. Biophys. Acta 536, 1-9 24 McNeill, G.P., Donnelly, W.J. and McGann, T.C.A. (1981) Research Report, Food Science and Technology, pp. 14-15, An Foras Taluntais, Dublin