In a review of the colloidal properties of casein micelles, Payens' concluded that a component due to steric stabilization was necessary to explain micellar stability. Prior to this, Holt' postulated the existence of a hydrophilic "hairy" coat to the micelle to explain its inherent stability despite low {-potential values. Walstra? demonstrated that the high voluminosity of the casein micelle could only be explained by assuming the existence of such a hairy surface layer. He further proposed that the hairs would consist predominantly of the C-terminal portion of K-casein. The action of chymosin would then be to remove the hairs, destablizing the micelle and simultaneously decreasing its hydrodynamic radius. The latter effect was confirmed by Walstra et a1.,4 who used photon-correlation-spectrosco- py (PCS) measurements to show that the micellar radius initially decreased by approximately 5 nm following the addition of chymosin before aggregation finally ensued.
Coagulation of casein micelles in milk may also be induced by the addition of ethanoL5s6 Though recent studies of the ethanol stability of milk have concentrated on effects due to milk s a l t ~, ~, s it is well known that in an unfavorable solvent environment, proteins and other polymers undergo a conformational transition to a collapsed form, accompanied by a dramatic reduction in the effective volume of the molecule. Solvent conditions which lead to the collapse are often similar to those leading to aggregati~n.~ The action of the ethanol in micellar aggregation may therefore be to collapse the "hairs," thus removing the steric stabilizing component from the system. T o test this hypothesis, PCS was used to measure the micellar hydrodynamic radius in the presence of ethanol up to those levels producing aggregation.
Fresh milk samples were obtained from the Hannah Research Institute farm bulk tank following the morning milking. Skim milk was prepared from this by centrifugation, as described by Horne and Parker.'O The contaminating fat was further depleted by successive filtration through glass-fiber filters (Whatman GF/A) and cellulose nitrate membrane, 0.8-pm pore (Millipore). Defined fractions of casein micelles were prepared by successive centrifugation steps, as detailed by Dalgleish et al.I1 For PCS measurements, the skim milk was diluted into a 20 mM imidazole/HCl buffer at pH 7.0, which contained 5 mM Ca2+ as CaC12 and 50 mM NaC1. Separate studies have shown that casein micelles remain stable in this calcium buffer system with no change in radius or molecular weight (turbidity) for greater than 3 h. Alcohol-containing buffers were prepared by mixing an equal volume of aqueous ethanol solution with a buffer solution whose components were a t twice the above values. The rennet used was a commercial powder preparation (Kasel-