- Email: [email protected]
Presence and genetic polymorphism of an epithelial mucin in milk of the goat (Capra hircus)
Comp. Biochem. Physiol. Vol. 103B,No. 1, pp. 261-266, 1992 Printed in Great Britain
0305-0491/92$5.00+ 0.0O © 1992PergamonPress Ltd
PRESENCE AND GENETIC POLYMORPHISM OF AN EPITHELIAL MUCIN IN MILK OF THE GOAT
(CAPRA HIRCUS) WENDY M. CAMPANA,*tRONALDV. JOSEPHSON* and STUARTPATTON:~§ *School of Family Studies and Consumer Sciences, San Diego State University, San Diego, CA 92182, U.S.A.; and J/Department of Neurosciences, Center for Molecular Genetics, 0634J, University of California, San Diego, La Jolla, CA 92093, U.S.A. (Tel: 619-534-6829); (Fax: 619-534-1383) (Received 3 January 1992) Abstract--1. Analysis of individual samples of goat's milk by SDS-PAGE confirmed that they contain a polymorphic, high molecular weight (Mr > 205 kDa) giycoprotein. 2. On SDS-gels, the polymorphism takes the form of two bands of variable mobility which usually stain with equal intensity. This polymorphism resembles that detected in milk mucins of other species and is best explained by an expression of codominant genes containing variable numbers of a tandemly repeated 60-base segment. 3. Analysis of milk fractions provided evidence that the goat mucin is exclusivelya membrane protein, and that it can be purified from other fat globule proteins by gel filtration and peanut lectin affinity chromatography. 4. Among proteins in the goat milk fat globule, the mucin appears to be a strong immunogen but the resulting antibodies applied to Western blots only stained the cow's milk mucin mildly and the guinea pig and human milk mucins not at all. from the goat milk fat globule by SDS-PAGE revealed a high molecular weight component which stained prominently with Schiff's reagents but not at all with Coomassie blue (Patton and Hubert, 1983). These are properties of the milk mucins. Thus we sought to determine whether or not this protein belongs to the family of polymorphic milk mucins detected in other species.
INTRODUCTION Mammals express high molecular weight glycoproteins, now known as epithelial mucins, in their milk (Snow et al., 1977; Shimizu and Yamauchi, 1982; Johnson et al., 1988; Patton et al., 1989; Spicer et al., 1991). These mucins are components of the plasma membrane that envelopes milk fat globules at secretion. Excepting those of the rodents (Spicer et al., 1990, milk mucins of species examined to date exhibit polymorphism on SDS-gels, e.g. human (Patton and Huston,1987; Swallow et al., 1987), chimpanzee (Patton et al., 1989), rhesus monkey (Welsch et al., 1990), horse and cow (Patton et aL, 1989). This polymorphism arises from variable numbers of a tandemly repeated segment of 20 amino acids (Gendler et al., 1987, 1988, 1990; Spicer et al., 1991). Thus, the two alleles inherited for the mucin, one maternal and one paternal, may vary in size with the result that bands of the milk mucin on SDS-gels may differ in mobility. Because of codominant expression of these variable-sized alleles, two bands of equal staining intensity are usually seen for milk samples from individuals. The bovine milk mucin, known as PAS-I because of its staining in SDS-gels with periodic acid-Sehiff's reagents and its high molecular weight (Mather and Keenan, 1975), shows differences in incidence of its polymorphic forms among the major dairy breeds (Patton and Muller, 1992). Worldwide, the goat is also an economically important species for which milk samples from individuals are readily available. Analysis of proteins tPresent address: Department of Dairy and Animal Science, Pennsylvania State University, University Park, PA 16802, U.S.A. §To whom correspondence should be addressed.
MATERIALS AND METHODS
Milk samples Goat milk samples were obtained from two goat farms in the San Diego area. Small quantities (20-40 ml) collected after complete milldngsof individual goats were immediately transported on ice to the laboratory and preserved with sodium azide, final concentration 0.02%. Samples not used were promptly frozen (-70°C) and held for subsequent analysis. Milk samples (2mi) were also obtained from genetically related and unrelated goats of a herd at the International Dairy Research Center, Prairie View, TX. These were shipped on dry ice to our laboratory and held frozen (-70°C) until analysed. Fresh milk samples from individuals of other species (human, cow and guinea pig) were cooled and handled as for goat's milk samples except that storage, when necessary, was at -20°C. The human sample was one selected at random from a large number donated by area mothers producing mature milk. The cow sample was one from those used in a recent study (Patton and Muller, 1992). The guinea pig milk was a gift of Dr Ian Mather of the University of Maryland, College Park, MD. Milk fractionation Milk from an individual goat was placed in 0.4 ml plastic centrifuge tubes (West Coast Scientific, No. 2070) and centrifuged in a swinging bucket-type rotor (International Equipment Co.) for 20 min at 1500g and 20°C. Following the spin, the tubes were cut with a razor blade immediately beneath the fat globule layer. These cream layers were quantitatively transferred to a small test tube and gently
261
262
WENDY M. CAMPANAet al.
dispersed in distilled water. The skim milk phases were placed in new 0.4 ml tubes and centrifuged in a Sorval Superspeed RC2-B centrifuge for 2 hr at 50,000g and 4°C. Traces of globules at the top of the tube were removed by cutting the tube with a razor blade directly beneath the surface layer and quantitatively dispersed in distilled water. The infranatant, fluff (membrane) and casein pellet (see Stewart et al., 1972 for description and picture of fractions) were collected separately. The casein pellet was solubilized by sonication in 1% SDS. In preparation for SDS--PAGE, all fractions were adjusted with distilled water to their initial milk volume (1.6 ml). Milk fat globules of the various species were purified from milk samples, dispersed in distilled water, solvent extracted to remove lipids and assayed for protein as previously described (Patton and Huston, 1986). Fractionation o f globule proteins Goat globule protein preparations were separated by molecular weight using Sephadex G-200 (Sigma) in an elution buffer system of 50 mM Tris containing 1% SDS and 0.02% sodium azide, pH 7.2. The lyophilized protein sample was solubilized in 50 mM Tris, pH 7.3, containing 2% SDS and 1% mercaptoethanol (Snow et aL, 1977). The solution was incubated for 2 hr at 37°C and then chromatographed on a Sephadex G-200 column, 2.5 × 53.5 cm. Blue Dextran at a concentration of 2 mg/ml was used to determine void volume. Flow rate was maintained at 6.6 ml/hr. Three-ml fractions were collected and their protein content monitored by absorbance at 280 nm. Fractions containing protein were electrophoresed at room temperature in SDSgels and stained with silver as described in the following section. A suspension of peanut lectin (Arachis hypogaea) beads (Sigma), approximately 1 ml, containing 12 mg of lectin, were exhaustively washed in a buffer system of 0.02M Tris, 0.5M NaCI and 0.05% Tween 20, pH 7.2. Pooled fractions from the Sephadex column containing the high molecular weight glycoprotein, were added to the washed lectin beads and gently mixed on a gyrotory mixer/shaker for 2 hr at ambient temperature. The reacted beads were recovered by centrifuging for 2 min at 1500g, washed with buffer and transferred quantitatively into a filtration tube (BioRad). After extensive washing, this column was eluted with 3.6% lactose solution containing 0.02% sodium azide and 0.05% Tween 20 as the elution buffer. Ten successive 2 ml fractions were collected and then dialyzed, concentrated and stored at 4°C pending analysis by SDS-PAGE. Gel electrophoresis, blotting and staining Protein-containing samples (whole milk, isolated globules, milk fractions and globule protein fractions) were thawed (40°C) when required, denatured and electrophoresed. Addition of 0.05M sodium phosphate buffer (pH 7.0) was omitted for milk samples and milk fractions containing the buffer system of milk. Samples containing 3-20/~g of protein were resolved by SDS-PAGE in a 3% stacking gel and either a 5 or 6% running gel (Laemmli, 1970). Blotting of SDS-gels onto nitrocellulose paper (Schleicher and Schull) was accomplished by the method of Towbin et al. (1979). A pre-stained high molecular weight protein standard (BioRad) was electrophoresed along with samples in gels to be blotted so that blotting efficiency could be evaluated and relative molecular weight of bands on blots observed. Gels were stained for glycoproteins with Schiff's reagents (Fairbanks et al., 1971) and for proteins in general with the silver based procedure of Morrissey (1981). In addition to staining and polymorphic characteristics, mucin bands were identified by comparison to a set of protein standards with Mrs of 29,000 to 205,000 (Sigma). Immunostaining of milk fat globule proteins on nitrocellulose blots was accomplished using rabbit antiserum to goat milk fat globule membrane
(Patton et al., 1980) as the first antibody, goat anti-rabbit IgG conjugated to horseradish peroxidase (Bio Rad) as the second antibody and chloronaphthol as chromogen.
RESULTS
In order for a gel-staining technique to be convenient for detecting milk mucin polymorphism, it should be capable of revealing the mucin in an amount of unfractionated milk that does not overload the gel. Fig. 1 (upper) indicates that goat milk and its fat globule membrane give rise to bands of high molecular weight protein that stain for carbohydrate with PAS reagents. These characteristics coupled with the proximity of the bands and their staining with equal intensity indicate close resemblance to the polymorphic mucin of cow's milk, known as PAS-I. The bands in question stain more strongly with the silver based method, Fig. 1 (lower), the latter being at least five times more sensitive than PAS-staining. Similar results have been reported regarding sensitivity of these methods in detecting bovine PAS-I (Patton and Patton, 1990). F o r the purposes of silver staining, the gel system accommodated samples up to and including 25 #I of milk reasonably well, lanes 7-9. However, the glycoprotein band in lane 10 is poorly resolved and inadequately stained presumably due to sample overload ( ~ 1600/~g of protein). F o r the gel position of the glycoprotein in Fig. 1 with respect to relative molecular weight markers, see Figs 2, 5 and 6. Figure 2 reveals distribution of the high molecular weight glycoprotein, which we henceforth refer to as PAS-I, among goat milk fractions. Position of the PAS-I bands in the figure is defined by brackets. In comparison to the whole milk (lane 2), it is evident
Fig. 1. Portions of a gel from SDS-PAGE analysis comparing periodic acid-Schiffs (upper) and silver staining (lower) for sensitivity to the high molecular weight glycoprotein of goat's milk. Lanes 1 and 6 contain 75 and 15 #g, respectively, of delipidated globule protein from a pooled sample of goat's milk. Lanes 2-5 and 7-10 contain 5, 10, 25 and 50/~1 respectively, of whole goat's milk. The gel contained 6% acrylamide.
Polymorphism of a mucin in goat's milk
263
Fig. 2. SDS--PAGE analysis of goat's milk frations for PAS-I (brackets): lane 1, molecular weight standards (in kDa); lane 2, original whole milk; lane 3, fat globule layer and lane 4, skim milk produced by an initial centrifugation at 1500g. Samples for the remaining lanes were derived by high speed centrifugation (50,000 g for 2 hr) of the skim milk: lane 5, small globule layer: lane 6, infranatant; lane 7, "fluff" fraction; lane 8, casein pellet. The band for xanthine oxidase (I 55 kDa), a principal protein of the milk fat globule, is indicated (arrowhead, here and Fig. 3). All fractions were made to the original starting volume of milk with distilled water, and 15/~1 of each was applied to the gel which contained 6% acrylamide and was stained using the silver method. that the cream layer (lane 3) contains a substantial fraction of the total PAS-I and that perhaps one fourth to one half of the protein occurs in the skim milk (lane 4). Of the fractions derived from the skim milk (lanes 5-8), only the fluff fraction appears to
contain bands of PAS-I and these are comparable in intensity to those of the skim. This is plausible since fluffis known to be rich in membrane, shed from milk fat globules (Wooding, 1971), from plasma membrane of the lactating cell (Plantz et al., 1973) or both.
Fig. 3. SDS-PAGE comparing PAS-I bands (brackets) from milk samples of individual unrelated goats. Whole milk (15 #l) was applied to lanes 1-7. Lane 8 contained 5 #g of delipidated globule protein from a pooled goat's milk sample. The 5%-acrylamide gel was silver stained.
264
WENDVM. CAMPANAet al.
Fig. 4. SDS-PAGE comparing PAS-I bands from milk samples of related goats. Data for three pairs of half-sisters are shown. Samples were 15/~1 of whole goat's milk except for lane R which was 5/tg of delipidated globule protein from a sample of pooled goat's milk. Note the bands held in common (next to small circles) by each half-sister pair. The gel contained 5% acrylamide and was silver stained. Evidence for polymorphism of goat PAS-I can be seen in Fig. 3 which presents an SDS-gel in which the milk proteins of eight unrelated goats, i.e. no sire or dam in common, are resolved. The position of the two PAS-I bands, defined by brackets, vary considerably. Lane 4 exhibits only one band which most likely is two bands superimposed due to homozygosity of the individual for the two alleles. Lanes 2, 5 and 6 contain two bands which have silver stained with somewhat different intensities. This is an exception to the equivalence which has usually, but not always, been observed for other species. Resolution of the proteins in milk samples obtained from genetically related goats is shown in Fig. 4. It can be seen that three sets of half-sisters exhibit one band with mobility in common between half-sister pairs. Thus, it appears that related individuals stand a better chance of exhibiting matching bands of the protein on SDS-gels than do non-relatives.
In addition to the data on the half-sisters (Fig. 4), we also determined PAS-I patterns of milk samples from a dam, two of her daughters and a daughter of one of the daughters. It was revealed that a band matching in mobility one of those displayed by the dam was seen for both daughters and the granddaughter (data not shown). In general, the data on genetic relationships as a factor in the observed polymorphism of the goat milk mucin are consistent with those observed for the milk mucins of other species (Swallow et al., 1987; Gendler et al., 1987; Patton and Patton, 1990). The polymorphism apparently is best explained by the operation of codominant genes. As in Fig. 3, there are instances in Fig. 4 of the two PAS-I bands of a sample not staining equally (left of pair 2, and lane R). Moreover, there is evidence of multiple bands in the fight hand member of pair 3. The cause of these phenomena is not known. Gel filtration and peanut lectin affinity chromatography were investigated as means toward isolation and characterization of the goat milk mucin. Two similar experiments were conducted. In the one, 2.9 mg of toal globule protein was applied to the gel column and in the other 3.6 mg. In order to simplify isolation of the mucin, the milk of an individual exhibiting superimposed bands for this protein was used. Representative results from one of these runs and for the lectin affinity purification, which was also repeated, are in Fig. 5. The band patterns of the original globule protein preparation applied to the gel filtration column and the pooled sample of the PASI-rich fractions from the column are shown in lanes 2 and 3, respectively. It is evident that the column product retains a contaminant of M r 155 kDa, presumably xanthine oxidase. Submission of this sample
Fig. 5. SDS-PAGE showing the purification of goat PAS-I by gel filtration on Sephadex G-200 and affinity chromatography employing immobilized peanut lectin: lane 1, molecular weight standards (in kDa); lane 2, 30 #g of the original delipidated goat globule protein preparation; lane 3, 100 t~l of a composite fraction from the Sephadex column; lane 4, 100 #1 of supernatant from reaction of the composite fraction with immobilized peanut lectin; lanes 5 8, 50/~1of fractions 4-7 eluted with lactose from the peanut lectin column. Each fraction originally was 2ml of eluate and was concentrated to 150#1. The SDS-gel contained 6% acrylamide and was silver stained.
Polymorphism of a mucin in goat's milk
265
Fig. 6. Half of a minigel from SDS-PAGE resolving milk fat globule proteins of several species (left) and a Western blot (right) of the other half of the gel bearing the same electrophoresed samples and immunostained with rabbit antiserum to goat globule proteins. Lanes, species and amounts of sample proteins were, respectively: l--goat, 3/~g; 2--cow, 3#g; 3--guinea pig, 3/~g; 4--human, 3/~g; 5--prestained high molecular weight standards (in kDa); 6--goat, 3/zg; 7--cow, 6/~g; 8--guinea pig, 6/~g; and 9--human, 6/~g. The milk mucins (PAS-I) are indicated by arrowheads. The gel contained 6% acrylamide and was silver stained (left half). For immunostaining of blot (right), see text. to the peanut iectin column, removed this component and yielded a single band corresponding to the PAS-I charged to the column (lane 6). In spite of extensive washing of the column, some alien bands, evident in the lectin binding experiment (lanes 4, 7 and 8) apparently arose by lectin dissociation from the beads. The yield of protein in the cleanest fraction (lane 6) was approximately 1 #g. To further explore the relationship between the goat milk mucin and that of the cow, we evaluated the globule proteins of those two species along with those of the human and guinea pig by SDS-PAGE. The resolved proteins were Western blotted and the blots immunostained with antiserum to goat milk fat globule membrane. The several species exhibit prominent mucin bands with differing mobilities in the SDS-gel (arrowheads, Fig. 6, left). The blot results (Fig. 6, right) establish that the goat mucin is a strongly antigenic membrane component (double band at arrowhead, lane 6), and that the bovine globule proteins, including its PAS-I (arrowhead, lane 7), show cross-reactivity, although it is significantly reduced in response compared to that of the goat proteins. There is essentially no cross-reactivity with the guinea pig proteins (lane 8) and only faint non-specific staining in the track of the human globule proteins (lane 9). These results indicate that the goat globule proteins, including its PAS-I, are somewhat related antigenically to those of the cow but little, if at all, to those of the guinea pig or human. With respect to the relatively strong immunostaining by the goat mucin (Fig. 6, lane 6), it has been proposed by Gendler et al. (1988), in the case of the human mucin, that this is due to repetition of epitopes in the tandem repeat region of the molecule. DISCUSSION
Our investigation has established that goat's milk contains a polymorphic mucin resembling those in milk of other species. In comparison to the bovine
mucin, the goat analog may be a considerably larger protein. In the upper region of the gel where the two mucins run (Fig. 6), small variations in mobility translate to large differences in molecular weight. This could result in a greater number of alleles for the goat in comparison to the cow because of the likelihood that the goat gene involves more tandem repeats. In that regard, the variety of band mobilities for goat PAS-I in Fig. 3 seems to exceed that previously observed for the cow (Patton and Patton, 1990; Patton and Muller, 1992). We have also noted that the goat mucin stains more intensely than the cow protein per unit of globule protein on SDS-gels (e.g. Fig. 6, left). This is true with both Schiffs and the silver-based stain. As with milk mucins of the human (Shimizu and Yamauchi, 1982), guinea pig (Johnson et aL, 1988) and cow (Patton et al.. 1989), the goat mucin readily binds to peanut lectin. This indicates the presence of a terminal sequence, fl-D-galactosyl(l-3)-N-acetyl-Dgalactosamine, in the oligosaccharides of the mucins. However, it is a point of interest that the antiserum to goat milk fat globule proteins did not react with such proteins of the human or guinea pig and much less strongly with those of the cow than of the goat (Fig. 6). It has been noted that in immunostaining blots of milk fat globule proteins from 10 species with a monoclonal antibody (HMGF-2) to the human milk mucin MUC-I, only the mucins of the human and chimpanzee responded, the latter weakly. The polyclonal antiserum to the human mucin only stained rhesus mucin in addition to those of the other two primates (S. Patton, unpublished data). Thus there appears to be considerable variation between species in the structure of the milk mucins. One reason for this is suggested by the findings of Spicer et al. (199t) that homology between the human and mouse mucins is low (40%) in the tandem repeat region. Mainly, positions of the serines and threonines were preserved. This led these authors to suggest that oligosaccharides coupled to these amino acids are the important
266
WENDY M. CAMPANAet al.
(preserved) functional aspects of the mucins. It appears that variations in structure of the mucin peptide backbones and their genes will provide interesting clues to genentic relationships of species. The human milk mucin and truncated forms of it can be detected in the circulation of some breast cancer patients by means of antibodies (Burchell et al., 1984). These observations coupled with evidence that the mucin occurs in human skim milk (Shimizu et al., 1986) suggest that a secreted (soluble) as well as a membrane bound form of the mucin may exist. While it is quite conceivable that tumor genetics may produce aberrant mucin forms, our data indicate only one form of the goat milk mucin. Distribution of the mucin in goat's milk is explained by its being an integral membrane protein. It was detected in fat globules where it is a component of the membrane and in the membrane-rich (fluff) fraction of skim milk. The bands in the SDS-gel (Fig. 2) showed no evidence that there were differences in the mucin at the two sites, i.e. similar mobilities and the proportions of the individual bands in each pair seem equal as judged by staining intensity. The human milk mucin carries variations in blood group antigens in its oligosaccharides (Dion et al., 1990) as well as polymorphism in its peptide chain. This makes these proteins unusual genetic markers which may have useful predictive value regarding yield and properties of milk. Acknowledgements--We wish to thank John S. O'Brien for
use of laboratory facilities; Adela Mora, Lila Sturgess and Fran Brown for samples of goat's milk; and Jeffrey Hubert for his interest in this work. This research was supported by contracts with the National Institute of Child Health and Human Development (NO1-HD6-2920) and the National Dairy Research and Promotion Board. REFERENCES
Burchell J., Wang D. and Taylor-Papadimitriou J. (1984) Detection of the tumor-associated antigens recognized by the monoclonal antibodies HMGF-I and 2 in serum from patients with breast cancer. Int. J. Cancer 34, 763-768. Di0n A. S., Williams J., Herlyn M. and Major P. (1990) Human milk fat globule membrane glycoproteins express blood group-related determinants primarily on mucin-like epithelial membrane antigens and gp 70. Biochem. Int. 22, 295-302. Fairbanks G., Steck T. L. and Wallach D. F. H. (1971) Electrophoretic analysis of the major polypeptides of the human erthrocyte membrane. Biochemistry 10, 2606-2610. Gendler S. J., Burchell J. M., Duhig T., Lamport D., White R., Peaker M. and Taylor-Papadamitriou J. (1987) Cloning of partial cDNA encoding differentiation and tumor-associated mucin glycoproteins expressed by human mammary epithelium. Proc. Natn. Acad. Sci. USA 84, 6060-6064. Gendler S. J., Lancaster C. A., Taylor-Papadimitriou J., Duhig T., Peat N., Burchell J., Pemberton I., Lalani E.-N. and Wilson D. (1990) Molecular cloning and expression of human tumor-associated polymorphic epithelial mucin. J. biol. Chem. 265, 15286-15293. Gendler S. J., Taylor-Papadamitriou J., Duhig T., Rothbard J. and Burchell J. (1988) A highly immunogenic region of a human polymorphic epithelial mucin expressed by carcinomas is made up by tandem repeats. J. biol. Chem. 263, 12820-12823. Johnson V. G., Greenwalt D. E., Madara P. J. and Mather I. H. (1988) Purification and characterization of
a differentiation-specific sialoglycoprotein of lactatingguinea-pig mammary tissue. Biochem J. 251, 507-514. Laemmli U. K. (1970) Cleavage of structural proteins during the assembly of the head of the bacteriophage T4. Nature 227, 680-685. Mather I. H. and Keenan T. W. (1975) Studies on the structure of milk fat globule membrane. J. Membrane Biol. 21, 65-85. Morrissey J. H. (1981) Silver stain for proteins in polyacrylamide gels: a modified procedure with enhanced uniform activity. Anal. Biochem. 117, 307-310. Patton S., Bogus E. R., Stemberger B. H. and Trams E. G. (1980) Antiserum to the milk fat globule membrane: preparation and capacity to suppress milk secretion. Biochim. Biophys. Acta 597, 216-233. Patton S. and Hubert J. (1983) Binding of concanavalin A to milk fat globules and release of the lectin-membrane complex by Triton X-100 J. Dariy Sci. 66, 2312-2319. Patton S. and Huston G. E. (1986) A method for isolation of milk fat globules. Lipids 21, 170-174. Patton S. and Huston G. E. (1987) Difference between individuals in high-Mr glycoproteins from human mammary epithelia. FEBS Lett. 216, 151-154. Patton S., Huston G. E., Jenness R. and Vaucher Y. (1989) Differences between individuals in high-molecular weight glyeoproteins from mammary epithelia of several species. Biochim. Biophys. Acta 980, 333 338. Patton S. and Muller L. D. (1992) Genetic polymorphism of the epithelial mucin, PAS-I, in milk samples from the major dairy breeds. J. Dairy Sci. 75, 863-867. Patton S. and Patton R. S. (1990) Genetic polymorphism of PAS-I, the mucin-like glycoprotein of bovine milk fat globule membrane. J. Dariy Sci. 73, 3567-3574. Plantz P. E., Patton S. and Keenan T. W. (1983) Further evidence of plasma membrane material in skim milk. J. Dairy Sci. 56, 978-983. Shimizu M. and Yamauchi K. (1982) Isolation and characterization of mucin-like glycoprotein in human milk fat globule membrane J. Biochem. 91, 515-524. Shimizu M., Yamauchi K., Miyauchi Y., Sakurai T., Tokugawa K. and Mcllhinney R. A. J. (1986) High-Mr glycoprotein profiles in human milk serum and fat globule membrane. Biochemical J. 233, 725--730. Snow L. D., Colton D. G. and Carraway K. L. (1977) Purification and properties of the major sialoglycoprotein of the milk fat globule membrane. Arch. Biochem. Biophys. 179, 690-697. Spicer A. P., Parry G., Patton S. and Gendler S. (1991) Molecular cloning and analysis of the mouse homologue of the tumor-associated mucin, MUC1, reveals conservation of potential O-glycosylation sites, transmembrane and cytoplasmic domains and loss of minisatellite-like polymorphism. J. Biol. Chem. 266, 15099-15109. Stewart P. S., Puppione D. L. and Patton S. (1972) The presence of microvilli and other membrane fragments in the non-fat phase of bovine milk. Z. Zellforsch. 123, 161-167. Swallow D. M., Gendler S., Griffiths B., Corney G., TaylorPapadimitriou J. and Bramwell M. E. (1987) The human tumor-associated epithelial mucins are coded by an expressed hypervariable gene locus PUM. Nature 328, 82-84. Towbin H., Staehelin T. and Gordon J. (1979) Electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose sheets: procedure and some applications. Proc. Natn. Acad. Sci. USA 76, 4350-4354. Welsch U., Schumacher U., Buchheim W., Schinko I.. Jenness R. and Patton S. (1990) Histochemical and biochemical observations on milk-fat-globule membranes from several mammalian species. Acta Histochem. 40, $59-64. Wooding F. B. P. (1971) The mechanism of secretion of the milk fat globule. J. Cell Sci. 9, 805-821.
0305-0491/92$5.00+ 0.0O © 1992PergamonPress Ltd
PRESENCE AND GENETIC POLYMORPHISM OF AN EPITHELIAL MUCIN IN MILK OF THE GOAT
(CAPRA HIRCUS) WENDY M. CAMPANA,*tRONALDV. JOSEPHSON* and STUARTPATTON:~§ *School of Family Studies and Consumer Sciences, San Diego State University, San Diego, CA 92182, U.S.A.; and J/Department of Neurosciences, Center for Molecular Genetics, 0634J, University of California, San Diego, La Jolla, CA 92093, U.S.A. (Tel: 619-534-6829); (Fax: 619-534-1383) (Received 3 January 1992) Abstract--1. Analysis of individual samples of goat's milk by SDS-PAGE confirmed that they contain a polymorphic, high molecular weight (Mr > 205 kDa) giycoprotein. 2. On SDS-gels, the polymorphism takes the form of two bands of variable mobility which usually stain with equal intensity. This polymorphism resembles that detected in milk mucins of other species and is best explained by an expression of codominant genes containing variable numbers of a tandemly repeated 60-base segment. 3. Analysis of milk fractions provided evidence that the goat mucin is exclusivelya membrane protein, and that it can be purified from other fat globule proteins by gel filtration and peanut lectin affinity chromatography. 4. Among proteins in the goat milk fat globule, the mucin appears to be a strong immunogen but the resulting antibodies applied to Western blots only stained the cow's milk mucin mildly and the guinea pig and human milk mucins not at all. from the goat milk fat globule by SDS-PAGE revealed a high molecular weight component which stained prominently with Schiff's reagents but not at all with Coomassie blue (Patton and Hubert, 1983). These are properties of the milk mucins. Thus we sought to determine whether or not this protein belongs to the family of polymorphic milk mucins detected in other species.
INTRODUCTION Mammals express high molecular weight glycoproteins, now known as epithelial mucins, in their milk (Snow et al., 1977; Shimizu and Yamauchi, 1982; Johnson et al., 1988; Patton et al., 1989; Spicer et al., 1991). These mucins are components of the plasma membrane that envelopes milk fat globules at secretion. Excepting those of the rodents (Spicer et al., 1990, milk mucins of species examined to date exhibit polymorphism on SDS-gels, e.g. human (Patton and Huston,1987; Swallow et al., 1987), chimpanzee (Patton et al., 1989), rhesus monkey (Welsch et al., 1990), horse and cow (Patton et aL, 1989). This polymorphism arises from variable numbers of a tandemly repeated segment of 20 amino acids (Gendler et al., 1987, 1988, 1990; Spicer et al., 1991). Thus, the two alleles inherited for the mucin, one maternal and one paternal, may vary in size with the result that bands of the milk mucin on SDS-gels may differ in mobility. Because of codominant expression of these variable-sized alleles, two bands of equal staining intensity are usually seen for milk samples from individuals. The bovine milk mucin, known as PAS-I because of its staining in SDS-gels with periodic acid-Sehiff's reagents and its high molecular weight (Mather and Keenan, 1975), shows differences in incidence of its polymorphic forms among the major dairy breeds (Patton and Muller, 1992). Worldwide, the goat is also an economically important species for which milk samples from individuals are readily available. Analysis of proteins tPresent address: Department of Dairy and Animal Science, Pennsylvania State University, University Park, PA 16802, U.S.A. §To whom correspondence should be addressed.
MATERIALS AND METHODS
Milk samples Goat milk samples were obtained from two goat farms in the San Diego area. Small quantities (20-40 ml) collected after complete milldngsof individual goats were immediately transported on ice to the laboratory and preserved with sodium azide, final concentration 0.02%. Samples not used were promptly frozen (-70°C) and held for subsequent analysis. Milk samples (2mi) were also obtained from genetically related and unrelated goats of a herd at the International Dairy Research Center, Prairie View, TX. These were shipped on dry ice to our laboratory and held frozen (-70°C) until analysed. Fresh milk samples from individuals of other species (human, cow and guinea pig) were cooled and handled as for goat's milk samples except that storage, when necessary, was at -20°C. The human sample was one selected at random from a large number donated by area mothers producing mature milk. The cow sample was one from those used in a recent study (Patton and Muller, 1992). The guinea pig milk was a gift of Dr Ian Mather of the University of Maryland, College Park, MD. Milk fractionation Milk from an individual goat was placed in 0.4 ml plastic centrifuge tubes (West Coast Scientific, No. 2070) and centrifuged in a swinging bucket-type rotor (International Equipment Co.) for 20 min at 1500g and 20°C. Following the spin, the tubes were cut with a razor blade immediately beneath the fat globule layer. These cream layers were quantitatively transferred to a small test tube and gently
261
262
WENDY M. CAMPANAet al.
dispersed in distilled water. The skim milk phases were placed in new 0.4 ml tubes and centrifuged in a Sorval Superspeed RC2-B centrifuge for 2 hr at 50,000g and 4°C. Traces of globules at the top of the tube were removed by cutting the tube with a razor blade directly beneath the surface layer and quantitatively dispersed in distilled water. The infranatant, fluff (membrane) and casein pellet (see Stewart et al., 1972 for description and picture of fractions) were collected separately. The casein pellet was solubilized by sonication in 1% SDS. In preparation for SDS--PAGE, all fractions were adjusted with distilled water to their initial milk volume (1.6 ml). Milk fat globules of the various species were purified from milk samples, dispersed in distilled water, solvent extracted to remove lipids and assayed for protein as previously described (Patton and Huston, 1986). Fractionation o f globule proteins Goat globule protein preparations were separated by molecular weight using Sephadex G-200 (Sigma) in an elution buffer system of 50 mM Tris containing 1% SDS and 0.02% sodium azide, pH 7.2. The lyophilized protein sample was solubilized in 50 mM Tris, pH 7.3, containing 2% SDS and 1% mercaptoethanol (Snow et aL, 1977). The solution was incubated for 2 hr at 37°C and then chromatographed on a Sephadex G-200 column, 2.5 × 53.5 cm. Blue Dextran at a concentration of 2 mg/ml was used to determine void volume. Flow rate was maintained at 6.6 ml/hr. Three-ml fractions were collected and their protein content monitored by absorbance at 280 nm. Fractions containing protein were electrophoresed at room temperature in SDSgels and stained with silver as described in the following section. A suspension of peanut lectin (Arachis hypogaea) beads (Sigma), approximately 1 ml, containing 12 mg of lectin, were exhaustively washed in a buffer system of 0.02M Tris, 0.5M NaCI and 0.05% Tween 20, pH 7.2. Pooled fractions from the Sephadex column containing the high molecular weight glycoprotein, were added to the washed lectin beads and gently mixed on a gyrotory mixer/shaker for 2 hr at ambient temperature. The reacted beads were recovered by centrifuging for 2 min at 1500g, washed with buffer and transferred quantitatively into a filtration tube (BioRad). After extensive washing, this column was eluted with 3.6% lactose solution containing 0.02% sodium azide and 0.05% Tween 20 as the elution buffer. Ten successive 2 ml fractions were collected and then dialyzed, concentrated and stored at 4°C pending analysis by SDS-PAGE. Gel electrophoresis, blotting and staining Protein-containing samples (whole milk, isolated globules, milk fractions and globule protein fractions) were thawed (40°C) when required, denatured and electrophoresed. Addition of 0.05M sodium phosphate buffer (pH 7.0) was omitted for milk samples and milk fractions containing the buffer system of milk. Samples containing 3-20/~g of protein were resolved by SDS-PAGE in a 3% stacking gel and either a 5 or 6% running gel (Laemmli, 1970). Blotting of SDS-gels onto nitrocellulose paper (Schleicher and Schull) was accomplished by the method of Towbin et al. (1979). A pre-stained high molecular weight protein standard (BioRad) was electrophoresed along with samples in gels to be blotted so that blotting efficiency could be evaluated and relative molecular weight of bands on blots observed. Gels were stained for glycoproteins with Schiff's reagents (Fairbanks et al., 1971) and for proteins in general with the silver based procedure of Morrissey (1981). In addition to staining and polymorphic characteristics, mucin bands were identified by comparison to a set of protein standards with Mrs of 29,000 to 205,000 (Sigma). Immunostaining of milk fat globule proteins on nitrocellulose blots was accomplished using rabbit antiserum to goat milk fat globule membrane
(Patton et al., 1980) as the first antibody, goat anti-rabbit IgG conjugated to horseradish peroxidase (Bio Rad) as the second antibody and chloronaphthol as chromogen.
RESULTS
In order for a gel-staining technique to be convenient for detecting milk mucin polymorphism, it should be capable of revealing the mucin in an amount of unfractionated milk that does not overload the gel. Fig. 1 (upper) indicates that goat milk and its fat globule membrane give rise to bands of high molecular weight protein that stain for carbohydrate with PAS reagents. These characteristics coupled with the proximity of the bands and their staining with equal intensity indicate close resemblance to the polymorphic mucin of cow's milk, known as PAS-I. The bands in question stain more strongly with the silver based method, Fig. 1 (lower), the latter being at least five times more sensitive than PAS-staining. Similar results have been reported regarding sensitivity of these methods in detecting bovine PAS-I (Patton and Patton, 1990). F o r the purposes of silver staining, the gel system accommodated samples up to and including 25 #I of milk reasonably well, lanes 7-9. However, the glycoprotein band in lane 10 is poorly resolved and inadequately stained presumably due to sample overload ( ~ 1600/~g of protein). F o r the gel position of the glycoprotein in Fig. 1 with respect to relative molecular weight markers, see Figs 2, 5 and 6. Figure 2 reveals distribution of the high molecular weight glycoprotein, which we henceforth refer to as PAS-I, among goat milk fractions. Position of the PAS-I bands in the figure is defined by brackets. In comparison to the whole milk (lane 2), it is evident
Fig. 1. Portions of a gel from SDS-PAGE analysis comparing periodic acid-Schiffs (upper) and silver staining (lower) for sensitivity to the high molecular weight glycoprotein of goat's milk. Lanes 1 and 6 contain 75 and 15 #g, respectively, of delipidated globule protein from a pooled sample of goat's milk. Lanes 2-5 and 7-10 contain 5, 10, 25 and 50/~1 respectively, of whole goat's milk. The gel contained 6% acrylamide.
Polymorphism of a mucin in goat's milk
263
Fig. 2. SDS--PAGE analysis of goat's milk frations for PAS-I (brackets): lane 1, molecular weight standards (in kDa); lane 2, original whole milk; lane 3, fat globule layer and lane 4, skim milk produced by an initial centrifugation at 1500g. Samples for the remaining lanes were derived by high speed centrifugation (50,000 g for 2 hr) of the skim milk: lane 5, small globule layer: lane 6, infranatant; lane 7, "fluff" fraction; lane 8, casein pellet. The band for xanthine oxidase (I 55 kDa), a principal protein of the milk fat globule, is indicated (arrowhead, here and Fig. 3). All fractions were made to the original starting volume of milk with distilled water, and 15/~1 of each was applied to the gel which contained 6% acrylamide and was stained using the silver method. that the cream layer (lane 3) contains a substantial fraction of the total PAS-I and that perhaps one fourth to one half of the protein occurs in the skim milk (lane 4). Of the fractions derived from the skim milk (lanes 5-8), only the fluff fraction appears to
contain bands of PAS-I and these are comparable in intensity to those of the skim. This is plausible since fluffis known to be rich in membrane, shed from milk fat globules (Wooding, 1971), from plasma membrane of the lactating cell (Plantz et al., 1973) or both.
Fig. 3. SDS-PAGE comparing PAS-I bands (brackets) from milk samples of individual unrelated goats. Whole milk (15 #l) was applied to lanes 1-7. Lane 8 contained 5 #g of delipidated globule protein from a pooled goat's milk sample. The 5%-acrylamide gel was silver stained.
264
WENDVM. CAMPANAet al.
Fig. 4. SDS-PAGE comparing PAS-I bands from milk samples of related goats. Data for three pairs of half-sisters are shown. Samples were 15/~1 of whole goat's milk except for lane R which was 5/tg of delipidated globule protein from a sample of pooled goat's milk. Note the bands held in common (next to small circles) by each half-sister pair. The gel contained 5% acrylamide and was silver stained. Evidence for polymorphism of goat PAS-I can be seen in Fig. 3 which presents an SDS-gel in which the milk proteins of eight unrelated goats, i.e. no sire or dam in common, are resolved. The position of the two PAS-I bands, defined by brackets, vary considerably. Lane 4 exhibits only one band which most likely is two bands superimposed due to homozygosity of the individual for the two alleles. Lanes 2, 5 and 6 contain two bands which have silver stained with somewhat different intensities. This is an exception to the equivalence which has usually, but not always, been observed for other species. Resolution of the proteins in milk samples obtained from genetically related goats is shown in Fig. 4. It can be seen that three sets of half-sisters exhibit one band with mobility in common between half-sister pairs. Thus, it appears that related individuals stand a better chance of exhibiting matching bands of the protein on SDS-gels than do non-relatives.
In addition to the data on the half-sisters (Fig. 4), we also determined PAS-I patterns of milk samples from a dam, two of her daughters and a daughter of one of the daughters. It was revealed that a band matching in mobility one of those displayed by the dam was seen for both daughters and the granddaughter (data not shown). In general, the data on genetic relationships as a factor in the observed polymorphism of the goat milk mucin are consistent with those observed for the milk mucins of other species (Swallow et al., 1987; Gendler et al., 1987; Patton and Patton, 1990). The polymorphism apparently is best explained by the operation of codominant genes. As in Fig. 3, there are instances in Fig. 4 of the two PAS-I bands of a sample not staining equally (left of pair 2, and lane R). Moreover, there is evidence of multiple bands in the fight hand member of pair 3. The cause of these phenomena is not known. Gel filtration and peanut lectin affinity chromatography were investigated as means toward isolation and characterization of the goat milk mucin. Two similar experiments were conducted. In the one, 2.9 mg of toal globule protein was applied to the gel column and in the other 3.6 mg. In order to simplify isolation of the mucin, the milk of an individual exhibiting superimposed bands for this protein was used. Representative results from one of these runs and for the lectin affinity purification, which was also repeated, are in Fig. 5. The band patterns of the original globule protein preparation applied to the gel filtration column and the pooled sample of the PASI-rich fractions from the column are shown in lanes 2 and 3, respectively. It is evident that the column product retains a contaminant of M r 155 kDa, presumably xanthine oxidase. Submission of this sample
Fig. 5. SDS-PAGE showing the purification of goat PAS-I by gel filtration on Sephadex G-200 and affinity chromatography employing immobilized peanut lectin: lane 1, molecular weight standards (in kDa); lane 2, 30 #g of the original delipidated goat globule protein preparation; lane 3, 100 t~l of a composite fraction from the Sephadex column; lane 4, 100 #1 of supernatant from reaction of the composite fraction with immobilized peanut lectin; lanes 5 8, 50/~1of fractions 4-7 eluted with lactose from the peanut lectin column. Each fraction originally was 2ml of eluate and was concentrated to 150#1. The SDS-gel contained 6% acrylamide and was silver stained.
Polymorphism of a mucin in goat's milk
265
Fig. 6. Half of a minigel from SDS-PAGE resolving milk fat globule proteins of several species (left) and a Western blot (right) of the other half of the gel bearing the same electrophoresed samples and immunostained with rabbit antiserum to goat globule proteins. Lanes, species and amounts of sample proteins were, respectively: l--goat, 3/~g; 2--cow, 3#g; 3--guinea pig, 3/~g; 4--human, 3/~g; 5--prestained high molecular weight standards (in kDa); 6--goat, 3/zg; 7--cow, 6/~g; 8--guinea pig, 6/~g; and 9--human, 6/~g. The milk mucins (PAS-I) are indicated by arrowheads. The gel contained 6% acrylamide and was silver stained (left half). For immunostaining of blot (right), see text. to the peanut iectin column, removed this component and yielded a single band corresponding to the PAS-I charged to the column (lane 6). In spite of extensive washing of the column, some alien bands, evident in the lectin binding experiment (lanes 4, 7 and 8) apparently arose by lectin dissociation from the beads. The yield of protein in the cleanest fraction (lane 6) was approximately 1 #g. To further explore the relationship between the goat milk mucin and that of the cow, we evaluated the globule proteins of those two species along with those of the human and guinea pig by SDS-PAGE. The resolved proteins were Western blotted and the blots immunostained with antiserum to goat milk fat globule membrane. The several species exhibit prominent mucin bands with differing mobilities in the SDS-gel (arrowheads, Fig. 6, left). The blot results (Fig. 6, right) establish that the goat mucin is a strongly antigenic membrane component (double band at arrowhead, lane 6), and that the bovine globule proteins, including its PAS-I (arrowhead, lane 7), show cross-reactivity, although it is significantly reduced in response compared to that of the goat proteins. There is essentially no cross-reactivity with the guinea pig proteins (lane 8) and only faint non-specific staining in the track of the human globule proteins (lane 9). These results indicate that the goat globule proteins, including its PAS-I, are somewhat related antigenically to those of the cow but little, if at all, to those of the guinea pig or human. With respect to the relatively strong immunostaining by the goat mucin (Fig. 6, lane 6), it has been proposed by Gendler et al. (1988), in the case of the human mucin, that this is due to repetition of epitopes in the tandem repeat region of the molecule. DISCUSSION
Our investigation has established that goat's milk contains a polymorphic mucin resembling those in milk of other species. In comparison to the bovine
mucin, the goat analog may be a considerably larger protein. In the upper region of the gel where the two mucins run (Fig. 6), small variations in mobility translate to large differences in molecular weight. This could result in a greater number of alleles for the goat in comparison to the cow because of the likelihood that the goat gene involves more tandem repeats. In that regard, the variety of band mobilities for goat PAS-I in Fig. 3 seems to exceed that previously observed for the cow (Patton and Patton, 1990; Patton and Muller, 1992). We have also noted that the goat mucin stains more intensely than the cow protein per unit of globule protein on SDS-gels (e.g. Fig. 6, left). This is true with both Schiffs and the silver-based stain. As with milk mucins of the human (Shimizu and Yamauchi, 1982), guinea pig (Johnson et aL, 1988) and cow (Patton et al.. 1989), the goat mucin readily binds to peanut lectin. This indicates the presence of a terminal sequence, fl-D-galactosyl(l-3)-N-acetyl-Dgalactosamine, in the oligosaccharides of the mucins. However, it is a point of interest that the antiserum to goat milk fat globule proteins did not react with such proteins of the human or guinea pig and much less strongly with those of the cow than of the goat (Fig. 6). It has been noted that in immunostaining blots of milk fat globule proteins from 10 species with a monoclonal antibody (HMGF-2) to the human milk mucin MUC-I, only the mucins of the human and chimpanzee responded, the latter weakly. The polyclonal antiserum to the human mucin only stained rhesus mucin in addition to those of the other two primates (S. Patton, unpublished data). Thus there appears to be considerable variation between species in the structure of the milk mucins. One reason for this is suggested by the findings of Spicer et al. (199t) that homology between the human and mouse mucins is low (40%) in the tandem repeat region. Mainly, positions of the serines and threonines were preserved. This led these authors to suggest that oligosaccharides coupled to these amino acids are the important
266
WENDY M. CAMPANAet al.
(preserved) functional aspects of the mucins. It appears that variations in structure of the mucin peptide backbones and their genes will provide interesting clues to genentic relationships of species. The human milk mucin and truncated forms of it can be detected in the circulation of some breast cancer patients by means of antibodies (Burchell et al., 1984). These observations coupled with evidence that the mucin occurs in human skim milk (Shimizu et al., 1986) suggest that a secreted (soluble) as well as a membrane bound form of the mucin may exist. While it is quite conceivable that tumor genetics may produce aberrant mucin forms, our data indicate only one form of the goat milk mucin. Distribution of the mucin in goat's milk is explained by its being an integral membrane protein. It was detected in fat globules where it is a component of the membrane and in the membrane-rich (fluff) fraction of skim milk. The bands in the SDS-gel (Fig. 2) showed no evidence that there were differences in the mucin at the two sites, i.e. similar mobilities and the proportions of the individual bands in each pair seem equal as judged by staining intensity. The human milk mucin carries variations in blood group antigens in its oligosaccharides (Dion et al., 1990) as well as polymorphism in its peptide chain. This makes these proteins unusual genetic markers which may have useful predictive value regarding yield and properties of milk. Acknowledgements--We wish to thank John S. O'Brien for
use of laboratory facilities; Adela Mora, Lila Sturgess and Fran Brown for samples of goat's milk; and Jeffrey Hubert for his interest in this work. This research was supported by contracts with the National Institute of Child Health and Human Development (NO1-HD6-2920) and the National Dairy Research and Promotion Board. REFERENCES
Burchell J., Wang D. and Taylor-Papadimitriou J. (1984) Detection of the tumor-associated antigens recognized by the monoclonal antibodies HMGF-I and 2 in serum from patients with breast cancer. Int. J. Cancer 34, 763-768. Di0n A. S., Williams J., Herlyn M. and Major P. (1990) Human milk fat globule membrane glycoproteins express blood group-related determinants primarily on mucin-like epithelial membrane antigens and gp 70. Biochem. Int. 22, 295-302. Fairbanks G., Steck T. L. and Wallach D. F. H. (1971) Electrophoretic analysis of the major polypeptides of the human erthrocyte membrane. Biochemistry 10, 2606-2610. Gendler S. J., Burchell J. M., Duhig T., Lamport D., White R., Peaker M. and Taylor-Papadamitriou J. (1987) Cloning of partial cDNA encoding differentiation and tumor-associated mucin glycoproteins expressed by human mammary epithelium. Proc. Natn. Acad. Sci. USA 84, 6060-6064. Gendler S. J., Lancaster C. A., Taylor-Papadimitriou J., Duhig T., Peat N., Burchell J., Pemberton I., Lalani E.-N. and Wilson D. (1990) Molecular cloning and expression of human tumor-associated polymorphic epithelial mucin. J. biol. Chem. 265, 15286-15293. Gendler S. J., Taylor-Papadamitriou J., Duhig T., Rothbard J. and Burchell J. (1988) A highly immunogenic region of a human polymorphic epithelial mucin expressed by carcinomas is made up by tandem repeats. J. biol. Chem. 263, 12820-12823. Johnson V. G., Greenwalt D. E., Madara P. J. and Mather I. H. (1988) Purification and characterization of
a differentiation-specific sialoglycoprotein of lactatingguinea-pig mammary tissue. Biochem J. 251, 507-514. Laemmli U. K. (1970) Cleavage of structural proteins during the assembly of the head of the bacteriophage T4. Nature 227, 680-685. Mather I. H. and Keenan T. W. (1975) Studies on the structure of milk fat globule membrane. J. Membrane Biol. 21, 65-85. Morrissey J. H. (1981) Silver stain for proteins in polyacrylamide gels: a modified procedure with enhanced uniform activity. Anal. Biochem. 117, 307-310. Patton S., Bogus E. R., Stemberger B. H. and Trams E. G. (1980) Antiserum to the milk fat globule membrane: preparation and capacity to suppress milk secretion. Biochim. Biophys. Acta 597, 216-233. Patton S. and Hubert J. (1983) Binding of concanavalin A to milk fat globules and release of the lectin-membrane complex by Triton X-100 J. Dariy Sci. 66, 2312-2319. Patton S. and Huston G. E. (1986) A method for isolation of milk fat globules. Lipids 21, 170-174. Patton S. and Huston G. E. (1987) Difference between individuals in high-Mr glycoproteins from human mammary epithelia. FEBS Lett. 216, 151-154. Patton S., Huston G. E., Jenness R. and Vaucher Y. (1989) Differences between individuals in high-molecular weight glyeoproteins from mammary epithelia of several species. Biochim. Biophys. Acta 980, 333 338. Patton S. and Muller L. D. (1992) Genetic polymorphism of the epithelial mucin, PAS-I, in milk samples from the major dairy breeds. J. Dairy Sci. 75, 863-867. Patton S. and Patton R. S. (1990) Genetic polymorphism of PAS-I, the mucin-like glycoprotein of bovine milk fat globule membrane. J. Dariy Sci. 73, 3567-3574. Plantz P. E., Patton S. and Keenan T. W. (1983) Further evidence of plasma membrane material in skim milk. J. Dairy Sci. 56, 978-983. Shimizu M. and Yamauchi K. (1982) Isolation and characterization of mucin-like glycoprotein in human milk fat globule membrane J. Biochem. 91, 515-524. Shimizu M., Yamauchi K., Miyauchi Y., Sakurai T., Tokugawa K. and Mcllhinney R. A. J. (1986) High-Mr glycoprotein profiles in human milk serum and fat globule membrane. Biochemical J. 233, 725--730. Snow L. D., Colton D. G. and Carraway K. L. (1977) Purification and properties of the major sialoglycoprotein of the milk fat globule membrane. Arch. Biochem. Biophys. 179, 690-697. Spicer A. P., Parry G., Patton S. and Gendler S. (1991) Molecular cloning and analysis of the mouse homologue of the tumor-associated mucin, MUC1, reveals conservation of potential O-glycosylation sites, transmembrane and cytoplasmic domains and loss of minisatellite-like polymorphism. J. Biol. Chem. 266, 15099-15109. Stewart P. S., Puppione D. L. and Patton S. (1972) The presence of microvilli and other membrane fragments in the non-fat phase of bovine milk. Z. Zellforsch. 123, 161-167. Swallow D. M., Gendler S., Griffiths B., Corney G., TaylorPapadimitriou J. and Bramwell M. E. (1987) The human tumor-associated epithelial mucins are coded by an expressed hypervariable gene locus PUM. Nature 328, 82-84. Towbin H., Staehelin T. and Gordon J. (1979) Electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose sheets: procedure and some applications. Proc. Natn. Acad. Sci. USA 76, 4350-4354. Welsch U., Schumacher U., Buchheim W., Schinko I.. Jenness R. and Patton S. (1990) Histochemical and biochemical observations on milk-fat-globule membranes from several mammalian species. Acta Histochem. 40, $59-64. Wooding F. B. P. (1971) The mechanism of secretion of the milk fat globule. J. Cell Sci. 9, 805-821.