Cell surface marker proteins during mouse spermatogenesis: Two-dimensional electrophoretic analysis

Antigens of developing mammalian spermatogenic cells have been identified serologically by a number of workers. Germ cell-specific antigens which first appear during late pachynema of the first meiotic prophase have been described in the mouse (Millette & Bellvé, 1977), the rabbit (O’Rand & Romrell, 1977), and the rat (Tung & Fritz, 1978). Other cell-surface constituents first appear in the membranes of early round spermatids in the guinea-pig and rabbit testis (Tung, Bebe-Han & Evan, 1979; O’Rand & Romrell, 1980). Most of these surface membrane constituents are also present on mature spermatozoa as determined using immunofluorescence or ultrastructural immunohistochemistry. Some surface components of mouse spermatogenic cells, however, are selectively partitioned during late spermiogenesis and are not detected on testicular or epididymal spermatozoa (Millette & Bellvé, 1980). Similar plasma membrane antigens may be partitioned in the rat (Tung & Fritz, 1978).

Send correspondence to: Clarke F. Millette, Ph.D., L.H.R.R.B., Harvard Medical School, 45 Shattuck Street, Boston, MA 02115, U.S.A.

Biochemical characterizations of membrane differentiation antigens on mammalian spermatogenic cells have not been reported previously. Recently, however, procedures for the purification of plasma membranes from isolated highly homogeneous populations of adult mouse spermatogenic cells have been described (Millette, O’Brien & Moulding, 1980), thus facilitating direct biochemical investigations of surface polypeptides specific to differentiating male germ cells. In conjunction with the availability of specific antibodies, including hybridoma products (Bechtol, Brown & Kennett, 1979), immunoprecipitation techniques should soon allow molecular investigations of many spermatogenic cell-surface antigens. Here, the polypeptide composition of plasma membranes from mouse pachytene spermatocytes, round spermatids, and residual bodies has been analysed using 2-dimensional gel electrophoresis. Two plasma membrane proteins have been identified, which are apparently specific to pachytene spermatocytes, and 4 surface constituents have been found to be characteristic of early round spermatids. These polypeptides represent the first membrane components identified as possible markers of individual stages of mammalian spermatogenesis. In addition, short-term in vitro culture of mouse spermatogenic cells indicates that these markers are not artifactually created during the preparation of single-cell suspensions from the seminiferous epithelium.

Adult CD-1 mice aged 60–120 days were obtained from Charles River Breeding Laboratories, Wilmington, Massachusetts. Adult Tac:(SW)fBR mice of the same age were also obtained from Taconic Farms, Inc. (Germantown, New York), and used in some experiments. No differences between membranes isolated from either mouse strain were detected. Seminiferous cell suspensions were prepared according to the procedures of Bellv6 and colleagues (Romrell, Bellv6 & Fawcett, 1976; Bellve et al. 1977 a, b) using sequential incubation in collagenase and trypsin. Purified populations of pachytene spermatocytes, round spermatids, and residual bodies were obtained using unit gravity sedimentation on linear gradients of bovine serum albumin (BSA) in enriched Krebs-Ringer bicarbonate medium (EKRB) as described by Bellv6 et al. (1977b). Most cell separations were conducted using a 2-4% w/v BSA gradient, but some separations were completed using 1–3 % w/v BSA in EKRB. No significant alteration in the relative position or purity of the isolated cell pools resulted from the changed gradient conditions. Cell purity was assayed using Nomarski differential interference microscopy as previously detailed (Romrell et al. 1976). Purities of individual spermatogenic cell types averaged 96 % for pachytene spermatocytes, 95 % for round spermatids, and 91 % for residual bodies. The major contaminants were as follows: pachytene spermatocytes were contaminated with occasional Sertoli cells and binucleated round spermatids, round spermatids were contaminated only with residual bodies, while the fraction of residual bodies co-sedimented with occasional round spermatids (1 %), erythrocytes, and condensing spermatids. Cell viabilities as determined using 0·16 % trypan blue in EKRB were always greater than 95 %. Mixed mouse spermatogenic cell suspensions were obtained by the omission of unit gravity sedimentation following enzymic treatment to remove all interstitial cells and Sertoli cells (Romrell et al. 1976). These suspensions consisted primarily of pachytene spermatocytes, round spermatids, and residual bodies. All cells were washed 3 times in EKRB before the purification of plasma membranes.

Purified plasma membranes were prepared from isolated populations of pachytene spermatocytes, round spermatids, and residual bodies according to Millette et al. (1980). Briefly, single-cell suspensions of seminiferous cells were suspended in EKRB to yield not more than 1·8 × 10⁸ cells per tube with an optimal number of 1·25 × 10⁸ cells per tube. In experiments involving purified pachytene spermatocytes no more than 6·10⁶ cells per tube were used. Cells were pelleted for 5 min at 200 g using a Beckman TJ-6R centrifuge at 4 °C. All traces of supernatant medium were removed before swelling the cells in hypotonic Tris-buffered saline solution (Brake, Will & Cook, 1978). Tris-saline solution (TBSS) consisted of 0·16M NaCl, 3 HIM MgCl₂, 5 mM KC1 in 10 mM Tris-HCl, pH 7·4 at 4 °C. The hypotonic swelling medium, or homogenization buffer, was TBSS diluted 1/10 with 10 mM Tris-HCl, pH 7·4 at 4 °C. Cell pellets were suspended in exactly o-8 ml homogenization buffer for exactly 5 min at 4 °C. Suspensions were then homogenized using 3 strokes in a small glass Dounce vessel with a Teflon pestle (A. W. Thomas Co., size o, cat. no. 3431-EO4, clearance 0·05×0·10 mm). Immediately after homogenization 0·08 ml 10 x TBSS was added to restore isotonicity for nuclear stabilization. Nuclei and unbroken cells were removed by pelleting for 30 s at 1000 g using a Fisher Model 59 microcentrifuge. Supernatant was removed carefully and transferred to a siliconized glass tube. The pelleted material was resuspended in ·8 ml TBSS (1X) and centrifuged for an additional 10 s at 1000 g as before. The second supernatant was added to the first to yield a final volume of 1 -6 ml consisting of enriched plasma membranes. All supernatants were kept at 4 °C.

Plasma membranes were isolated in high purity from the enriched supernatants by centrifugation on discontinuous sucrose gradients in TBSS. Gradients were loaded according to Monneron & d’Alayer (1978). Exactly 1·5 ml of enriched supernatant was mixed with 1·5 ml of 80% sucrose w/v in TBSS to yield 3·0 ml of 40% w/v sucrose containing membranes. Any remaining supernatant was discarded. All of the 40 % was layered on top of 3·0 ml 45 % sucrose w/v in TBSS in a cellulose nitrate tube (Beckman SW41 rotor). Five millilitres of 30 % sucrose w/v in TBBS were then layered to complete the discontinuous gradient. TBSS alone was used to fill the tube completely. Gradients were centrifuged at 125240 g, (32000 rev/min) for 90 min at 40 °C. Purified plasma membranes were obtained at the 3040 % sucrose interface. This material was pelleted, after dilution in TBSS, by centrifugation at 125 240 g_T for 30 min at 4 °C on a Beckman SW41 rotor. Biochemical and ultrastructural studies have demonstrated that the isolated surface membranes are not significantly contaminated by intracellular cellular organelles or membranes (Millette et al. 1980). The yield of purified membrane obtained from a single preparation of separated pachytene spermatocytes or round spermatids ranged from 150 to 300 µg.

Short-term in vitro cell culture was conducted with mixed populations of adult CD-1 mouse seminiferous cells or with purified populations of pachytene spermatocytes and round spermatids. Cells (2 × 10⁷ per aliquot) were cultured in 5 ml Minimal Eagle’s Medium (MEM) containing 10% foetal calf serum and added L-glutamine (4 mM). Cultures were kept at 33 °C with a 5 % CO, in air atmosphere. In radiolabelling experiments, cells were incubated with 50×Ci/ml [‘H]leucine (> 110Ci/mmol; New England Nuclear). At specified times, cells were removed from culture, pelleted, and washed 3 times in 5-ml aliquots of cold phosphate-buffered saline (PBS; 8·00 g NaCl, 1·15 g Na₂HPO₄, 0·20 g KH₂PO₂, 0·20 g KC1 per 1. H₂O, pH 7·4). It should be noted that adult mouse spermatogenic cells do not adhere to culture dishes in vitro. As a result, no enzymes or chelating agents were required for cell harvesting. Pellets were suspended in 1 ml 5 % TCA and kept at 4 °C overnight before washing with cold 5% TCA and dissolution in 1 ml 0·1 N NaOH. Aliquots were assayed for TCA-insoluble counts using liquid scintillation spectrometry. All cultures were set up in triplicate. In some experiments, pelleted samples were resuspended directly in sample buffer for polyacrylamide gel electrophoresis.

Plasma membranes were isolated from mixed populations of adult mouse seminiferous cells after 8 h in vitro culture as already described. No alteration of the membrane isolation procedure was required. Viability assays conducted using 0·16 % trypan blue in MEM indicated that all cell populations were >89% viable at the end of the culture period. In addition, haemocytometer counts revealed no significant loss of cells during the short-term in vitro incubations.

Polyacrylamide gel electrophoresis in one dimension was conducted according to Laemmli (1970) using 20% glycerol and 40 mM dithiothreitol in the sample buffer. The concentration of the running gel was 12% acrylamide. Two-dimensional gel electrophoresis was conducted according to O’Farrell (1975) using a linear 7–12 % acrylamide gradient for the second dimension. Gels were fixed and stained with Coomassie brilliant blue R (Sigma Chemical Co., 0·1 % in 45 % methanol, 10% acetic acid). Gels were destained in 7·5 % acetic acid in 5 % methanol and photographed using Ektapan film. For autoradiography gels were dehydrated in dimethylsulphoxide (DMSO), equilibrated with 2,5-diphenyloxazole in DMSO, rehydrated and dried under vacuum according to Bonner & Laskey (1974). Kodak X-Omat R film was prefogged according to Laskey & Mills (1975) and exposed to gels for 5 days. Equal amounts of protein were loaded on each gel track. Protein was determined using the fluram method (Nakamura & Pisano, 1976).

Highly-purified populations of mouse pachytene primary spermatocytes, round spermatids, and residual bodies were prepared according to the procedures of Bellve et al. (1977 a, b). All preparations were carefully screened microscopically to minimize contamination by somatic cells or by unwanted germ cells. Purified plasma membranes were prepared using the method of Millette et al. (1980). Pelleted membranes were suspended and solubilized directly in sample buffer and analysed by 2-dimensional polyacrylamide gel electrophoresis as described by O’Farrell (1975), except that a linear acrylamide gradient from 7·5 to 12% was used in the second dimension. Over no distinct spots were detected reproducibly in each preparation of plasma membranes isolated from separated adult mouse spermatogenic cells. The silver nitrate staining method of Merril, Switzer & Van Keuren (1979) and Switzer, Merril & Shifrin (1979) was also employed in these studies. This procedure is of high sensitivity, but it was difficult to control the staining intensity in our experiments. No significant differences were detected in comparisons of gels stained with either silver nitrate or Coomassie brilliant blue. The positions, shapes, and relative intensities of all analysed polypeptides were constant and independent of the particular experimental preparation or the exact amount of protein loaded into each gel. Although computer-assisted analysis was not conducted in this study, little difficulty was encountered in recognizing polypeptides unique to particular gels.

As might be expected, the majority of proteins detected in this study were found in all 3 preparations: pachytene spermatocyte membranes (Fig. 1), round spermatid membranes (Fig. 2), and residual body membranes (Fig. 3). Major surface components as judged by the intensity of Coomassie brilliant blue staining were predominantly of greater than 45 K molecular weight, in good agreement with the 1-dimensional electrophoretic results of Millette et al. (1980). Most of the major membrane constituents exhibited pls ranging from 5·5 to 6·5, although reproducible staining patterns were also obtained for materials staining as elongated smears with pl values approximating 7·1 to 7·4. Greater than 95 % of all the proteins detected on any one gel could also be detected on any other gel at an appropriate sample load.

Close examination and comparison of plasma membranes from the individual classes of adult mouse spermatogenic cells did reveal a limited number of consistent differences. Two proteins, labelled Pa and Pb in Fig. 1, were found only in surface membranes of adult mouse pachytene spermatocytes. Pa has an estimated molecular weight of just under 90 K and a pl of just over 5·6. Protein Pb has an estimated molecular weight of 56·5 K and a pl of 6·0. No significant amount of either polypeptide was detected in membranes isolated from purified round spermatids or from purified residual bodies.

In comparison, 2-dimensional gels of purified plasma membranes from mouse round spermatids revealed 4 major components not detected in either pachytene spermatocyte membranes or residual body membranes. These components have been labelled RSa, RSb, RSc, and RSd (Fig. 2). Proteins RSa and RSb have a similar pl of 5·9 and molecular weights of between 90 K and 95 K. Quantitatively, RSa predominates as judged by the intensity of Coomassie brilliant blue staining. RSa and RSb are clearly 2 different proteins. Protein RSc has a pl of 5·5 and a molecular weight of approximately 88 K. RSc is easily distinguished from component Pa seen on gels of pachytene spermatocyte membranes. A difference of over 0·1 pl unit separates the position of these constituents. The polypeptide labelled RSd has a molecular weight of 58 K and is present as a characteristic elongated smear ranging from pl 6·0 to 6·3. The slightly higher molecular weight and the elongation of RSd distinguish this component from polypeptide Pb detected in pachytene spermatocyte surface membranes.

Finally, electrophoretic analysis of residual body membranes indicates great similarity with the membranes of both pachytene spermatocytes and round spermatids. No reproducible differences unique to residual body preparations were detected. Very little material corresponding in position to Pa, RSa, RSb, RSc, or RSd was detected in any analysis of residual body membranes. The positions of RSa and RSd have been labelled in Fig. 3 to indicate the maximum amount of these 2 constituents seen in residual bodies. This small amount of material may be accounted for by slight contaminations of residual bodies by other germ cells. No material corresponding to Pb was ever detected in residual body membranes.

These data, therefore, suggest that proteins RSa, RSb, RSc, and RSd are plasma membrane markers which first appear during the early stages of spermiogenesis in the mouse. These components are the first such surface proteins to be described biochemically using electrophoretic techniques. Proteins Pa and Pb, furthermore, represent the first identified spermatogenic cell-surface constituents which disappear from the plasma membrane between pachynema and the generation of haploid spermatids. It is possible that the 6 proteins described here may represent differentiation markers specific to individual classes of mouse spermatogenic cells with Pa and Pb specific to pachytene spermatocytes and the RS components specific to developing spermatids. The absence of these components from residual body membranes supports this hypothesis, but a definitive demonstration must await the analysis of plasma membranes purified from both earlier germ cells and mature mouse spermatozoa. Neither of these materials is yet available in high purity suitable for detailed biochemical study.

Due to the unavoidable exposure of mouse spermatogenic cells to both collagenase and trypsin during their isolation, it was important to ensure that Pa, Pb, and the RS proteins were not created as artifacts of enzymic treatment. Accordingly, experiments were conducted to establish satisfactory short-term in vitro culture conditions for separated mouse pachytene spermatocytes and round spermatids. After isolation using unit gravity sedimentation (Bellv et al. 1977 a, b), cells were cultured in Minimal Eagle’s Medium containing foetal calf serum at 33 °C as described earlier. Cell counts taken with a haemacytometer revealed no loss of cells up to 48 h of in vitro culture, after which time cell numbers decreased rapidly in all cultured cell populations. Viability assays made with trypan blue also showed greater than 90% viability for all cell populations up to 48 h. No morphological changes were detected in mixed germ cell populations or in purified pachytene spermatocytes or round spermatids when observed using Nomarski differential interference or phase contrast optics. Ultrastructural analysis has not yet been completed.

Labelling experiments using [³H]leucine were conducted to assay protein synthesis by mouse spermatogenic cells in vitro. These results are shown in Fig. 4. TCA-precipitable label was readily obtained in these cultures with pachytene spermatocytes consistently yielding higher levels of labelling per cell than round spermatids. Pachytene spermatocytes are, of course, larger than round spermatids. Compared with the haploid spermatids, pachytene spermatocytes contain 3·6 times the total cellular protein as determined using the Lowry assay (Lowry, Rosebrough, Farr & Randall, 1951). The total number of TCA-insoluble counts for both pachytene spermatocytes and round spermatids plateaued between 18 and 24 h of in vitro incubation. These data were highly reproducible, with absolute values for [³H]leucine incorporated on a per cell or per protein basis rarely differing by more than 10% between experiments. Further characterization of the metabolic activity of isolated mouse spermatogenic cells in vitro is needed, but these data indicate that the germ cells do synthesize protein in culture and might be expected to repair any cell membrane polypeptides removed or cleaved proteolytically during the initial stages of cell separation. Similar culture studies using highly-purified populations of rat round spermatids have been reported previously by Nakamura, Romrell & Hall (1978) and Nakamura & Hall (1977, 1978).

To assay for the variety of cellular proteins synthesized in vitro, populations of isolated pachytene spermatocytes and round spermatids were cultured separately in [³H]leucine. At various times, cells were removed from culture, washed, and analysed using 1-dimensional polyacrylamide gel electrophoresis. The results of a typical experiment are shown in Fig. 5. The pattern of total cellular proteins obtained with Coomassie blue staining is identical to that found previously using whole cell homogenates immediately after cell separation, without in vitro culture (Millette et al. 1980). No significant differences were detected as a result of culture in either spermatogenic cell type.

Fluorographic analysis of the 1-dimensional gels (Fig. 5) reveals that both pachytene spermatocytes and round spermatids synthesize numerous polypeptides during short-term in vitro culture, with molecular weights ranging from over 100 K to approximately 10 K. The exact patterns of proteins synthesized under the conditions used are difficult to analyse satisfactorily using 1-dimensional electrophoresis due to the many polypeptides showing incorporation of [³H]leucine, but the results shown in Fig. 5 indicate that isolated mouse spermatogenic cells in vitro readily synthesize proteins. Recent investigations using [³H]mannose have established, in addition, that plasma membrane constituents including glycosylated proteins are synthesized in vitro (Millette and Zak, unpublished results).

Having established suitable short-term in vitro culture conditions for single-cell suspensions from the mouse seminiferous epithelium, experiments were conducted to assess directly the effects of enzymic cell separation on germ-cell surface polypeptides. Mixed populations of adult mouse spermatogenic cells were prepared using collagenase and trypsin according to Bellve et al. (1977 a, b). Plasma membranes were prepared as described (Millette et al. 1980) from some cell preparations immediately after washing away exogenous protease. Other cell preparations were cultured in vitro under conditions already detailed for 8 h before harvest and the preparation of purified plasma membranes. Both sets of purified surface membranes were then compared using 2-dimensional polyacrylamide gel electrophoresis (Fig. 6).

The results of this experiment indicate that only a very limited number of surface polypeptides detected using Coomassie brilliant blue are affected by collagenase and trypsin. Immediately after cell separation, electrophoretic analysis reveals 2 components, labelled e1 and e2 (Fig. 6A), which are more prominent before than after in vitro culture. Polypeptide e1 has a molecular weight of approximately 85 K and a pl of 5·6. Polypeptide e2 has a molecular weight of 60 K and a pl of close to 5·4. A total of 110 to 120 different proteins can be detected electrophoretically in membranes prepared immediately following cell separation. Of these, only components e1 and e2 are not also detected following 8 h in vitro culture. It is likely, therefore, that e1 and e2 represent the products of partial enzymic digestion during the cell separation procedure. It should also be noted, however, that there is no evidence of increased numbers of low-molecular-weight constituents in membrane preparations from uncultured spermatogenic cells. This implies that any adverse effects of collagenase and trypsin are highly selective.

After 8 h of in vitro culture, purified plasma membranes from adult mouse seminiferous cells no longer contain large amounts of constituents e1 and e2 (Fig. 6 B). Instead, 5 polypeptides not seen in membranes prepared immediately after cell separations are evident. These components have been indicated in Fig. 6 B. Protein ci has an estimated molecular weight of 98 K and a pl of 5·6. Protein C2 has a similar pl, but is of lower molecular weight, approximately 90 K. This constituent (c2) is similar in position to Pa found in surface membranes of purified pachytene spermatocytes and may be coincident with component Pa. Constituent C3 has a molecular weight of 61 K and a pl of 5·9–6·0, C4 has a molecular weight of 52 K and a pl of 5·8, and C5 has a molecular weight of 48 K and a pl of 5·6. These 5 constituents (C1–C5) represent the only new membrane constituents detected after the in vitro culture of mouse spermatogenic cells. Presumably, they represent surface proteins degraded enzymically during the cell separation procedure and replaced biosynthetically during short-term in vitro culture.

The similarity of 2-dimensional electrophoretic patterns detected before and after in vitro cell culture is striking. Results shown in Fig. 6 indicate that the enzymic cell separation procedures affect only 7 components out of a total of 110–120 polypeptides as assayed here. Only one of the constituents specific to particular stages of spermatogenesis (Pa) is apparently damaged proteolytically in this experiment (Fig. 6). Furthermore, as shown in Fig. 1 component Pa is detected consistently as a major component even in plasma membranes from cells assayed immediately following enzymic treatment. Therefore, the cell differentiation marker proteins [Pa, Pb, Rs(a–d)] do not represent enzyme degradation products.

The close correlation of polypeptide compositions seen before and after short-term in vitro culture suggests that the plasma membranes of isolated adult mouse spermatogenic cells are highly representative of their in vivo constitution. It is not yet evident to what extent spermatogenic cells in vitro mimic their physiological biosynthetic capability, but it is apparent that only 4% (5 of 120) of the proteins analysed here are degraded during cell separation to the extent that they cannot be identified without biosynthetic recovery procedures. Ninety-four per cent (113 of 120) of polypeptides from these cells seem unaltered electrophoretically by collagenase and trypsin. The gel system used in these studies is capable of resolving clearly singlecharge differences between proteins (Steinberg, O’Farrell, Friedrich & Coffino, 1977). Plasma membranes from freshly isolated mouse spermatogenic cells seem well-suited, therefore, for intensive biochemical analysis of the majority of cell surface polypeptides.

Cell-surface components specific to developing mammalian spermatogenic cells have been described serologically in a variety of species. Millette & Bellve (1977) first demonstrated in the mouse that new membrane constituents appear during late pachynema and that these constituents are maintained as surface components of the mature spermatozoon. Membrane antigens showing a similar temporal expression during spermatogenesis were also demonstrated initially by O’Rand & Romrell (1977) in the rabbit. Soon after, Tung & Fritz (1978) reported the presence of specific plasma membrane antigens during later spermatogenesis in the rat. Tung et al. (1979) reported that new surface proteins appear during early spermiogenesis in the guinea pig. Romrell & O’Rand (1978) have also conducted a study of the mobility of membrane antigens during rabbit spermatogenesis, while O’Rand & Romrell (1980) recently described 2 sub-classes of autoantigens in the same species. One class appears as already described during pachynema, but other components first appear on differentiating spermatids. Finally, Millette & Bellve (1980) demonstrated that still different cell-surface molecules appear throughout spermatogenesis in the mouse, but are not present on testicular or epididymal spermatozoa. These antigens seem to be partitioned selectively to the residual body. Similar components may exist in the rat (Tung & Fritz, 1978). Serological studies of plasma membrane antigens and their temporal expression during mammalian spermatogenesis have been reviewed by Millette (1979).

Although these investigations have established the existence of multiple surface proteins specific to differentiating male germ cells, 2 major unanswered questions remain. First, most studies to date have been conducted serologically with little or no attempt at the biochemical characterization of particular membrane molecules. O’Rand & Porter (1979) have conducted preliminary studies of a sperm membrane sialoglycoprotein, but its exact relation to other immunologically-described constituents is not yet clear. Silver, Artzt & Bennett (1979) have analysed both total cellular extracts and enriched membrane preparations from mixed populations of mouse spermatogenic cells using 2-dimensional gel electrophoresis after in vitro labelling with [³⁵S]methionine. These investigators have identified a p63/6·9 polypeptide which is encoded for by a gene of the T locus. Second, all constituents identified at present exist on multiple spermatogenic cell types including pachytene spermatocytes, round spermatids, residual bodies, and spermatozoa isolated from various positions along the male reproductive tract. True differentiation antigens specific to individual stages of spermatogenesis have not previously been demonstrated.

The application of hybridoma technology promises to facilitate analysis of these problems. Monoclonal antibodieg have in fact been produced against mixed suspensions of mouse spermatogenic cells (Bechtol et al. 1979) and against mature guinea-pig spermatozoa (Myles, Primakoff and Bellvd, in preparation). Membrane constituents specific to individual classes of spermatogenic cells have not yet, however, been described using this approach. Proteins Pa and Pb of mouse pachytene spermatocytes, as well as the RS components of round spermatids, therefore, are the first presumptive differentiation antigens described for mammalian spermatogenesis. In addition, the results described here provide the first detailed electrophoretic analyses of purified plasma membranes isolated from highly homogeneous populations of mammalian spermatogenic cell populations.

Presently, plasma membrane preparations from mouse spermatozoa are not yet available in sufficiently high purity to allow direct biochemical analysis of polypeptide compositions. The absence of the RS proteins on residual bodies suggests, but does not prove, that these pioteins are indeed differentiation markers. It should now be possible, however, to prepare monoclonal antibodies directed specifically to any of the 4 RS proteins. Such reagents could then be used to assay mature spermatozoa using a variety of immunological techniques to test for the expression of RS(a-d). Likewise, purified plasma membranes from early mouse spermatogenic cells at present are available in enriched, but not purified, preparations (O’Brien & Millette, unpublished results). Further purification of this material is required in order to determine the stage of spermatogenesis where components Pa and Pb are first expressed. Other workers have established that pachytene spermatocytes initiate the expression not only of multiple cell-surface polypeptides, but also of specific intracellular enzymes such as LDH-C₄ (Meistrich, Trostle, Frapart & Erickson, 1977) and cytochrome C₄ (Wheat, Hintz, Goldberg & Margoliash, 1977). It seems possible, therefore, that constituents Pa and Pb may indeed represent differentiation antigens for mouse pachytene spermatocytes.

As indicated by in vitro culture experiments, components Pa, Pb, and RS(a–d) are not artifacts resultant from the enzymic cleavage of plasma membrane proteins during cell separation. Under conditions where spermatogenic cells synthesize both intracellular and cell surface polypeptides, less than 4 % of the assayed membrane proteins are altered as assayed electrophoretically after 8 h of incubation. Although the half-lives of spermatogenic cell membrane components are not known, investigations of the rate at which cultured cells replace lost surface proteins suggest that 8 h is a sufficient culture period. Turner, Strominger & Sanderson (1972), for example, assayed the ability of lymphocytes in vitro to replace HLA-2 antigens after proteolysis by papain and found complete regeneration of the histocompatibility determinants within 6 h. Similarly, Cook, Will, Proctor & Brake (1976) noted that Na⁺, K⁺-ATPase is replaced on HeLa cell membranes within 3–6 h. Further efforts must be conducted to correlate the in vitro biosynthetic capability of isolated mouse spermatogenic cells with their in vivo activity; but Nakamura & Hall (1977, 1978, 1980) and Nakamura et al. (1978), using culture conditions similar to those described here, have already demonstrated the feasibility of this experimental system. This suggests strongly that direct biochemical characterization of plasma membranes from separated classes of mouse spermatogenic cells, in conjunction with the production of specific antibodies, will facilitate greatly our knowledge of molecular structure and function of male germ cell surfaces.

We are grateful for the participation of Mr Eric Rosenthal in the in vitro cell culture studies and for the excellent assistance of Mr Steven Borack, photographer, in the preparation of this manuscript. This research was supported by Grant 11267 from the National Institute of Child Health and Human Development.