Abstract

Rat sperm 2B1 antigen (the orthologue of guinea pig sperm PH20) is a plasma membrane-bound glycoprotein that is endoproteolytically cleaved during passage through the epididymis and subsequently migrates from the tail to the acrosomal domain during capacitation. Unlike guinea pig PH20, however, sperm surface 2B1 is insensitive to phosphatidylinositol phospholipase C, nor is it known how endoproteolytic cleavage affects its hyaluronidase activity. In this investigation we have expressed 2B1 cDNA in Chinese hamster ovary cells; we have shown that it contains an internal sequence motif for attachment of a glycosyl phosphatidylinositol (GPI) anchor and that cleavage from a single- into a two-chain molecule causes a significant shift in the optimum pH for hyaluronidase activity. Functionally, these results suggest that 1) 2B1 glycoprotein on rat spermatozoa is attached to the plasma membrane via a GPI anchor and that this is an important factor in its ability to migrate from the tail to the acrosomal domain during capacitation; and 2) endoproteolytic cleavage of 2B1 serves to optimize its hyaluronidase activity immediately before fertilization, thereby facilitating penetration of spermatozoa through the cumulus oophorus.

Introduction

A characteristic feature of many differentiated cells is their ability to sequester newly synthesized proteins and lipids to specific sites on their surface membrane to perform specialized functions. Nonrandom distribution is achieved principally by sorting mechanisms within the trans Golgi network or by generalized membrane insertion followed by repositioning to selected regions on the surface [1, 2]. In addition, a variety of sequence motifs and posttranslational modifications have been shown to facilitate vectoral transport and polarized diffusion. These include glycosylation, addition of lipid anchors, preferential affinity of certain glycoproteins for specific membrane lipids, connection to the submembranous cytoskeleton, and “corralling” between diffusion barriers [3].

Mammalian spermatozoa are highly specialized cells whose DNA is transcriptionally inactive. Unlike somatic cells, therefore, they cannot respond to agonists by transcribing new proteins. At the same time, they have to be sensitive to external signals from their environment and regulate their fertilizing capacity until they are within the immediate vicinity of an egg. One of the mechanisms whereby spermatozoa achieve this is through repositioning certain plasma membrane antigens (especially those crucial for binding to receptors on the egg) from regions where they are inactive to regions where they become active, either because they are modified in the process or because they associate with new membrane components and acquire novel functions. Examples of sperm antigens that change their distribution during posttesticular development are fertilin [4], PH20 [5, 6], PT1 [5], CE9 [7], galactosyltransferase [8], and 2B1 [9]. On guinea pig testicular spermatozoa, fertilin is initially distributed all over the acrosomal plasma membrane, but, after endoproteolytic cleavage in the testis and epididymis and loss of N-terminal peptides, it becomes restricted to the posterior head domain [4]. Ectodomain shedding of an N-terminal peptide from rat sperm CE9 antigen also takes place concomitant with its redistribution from the principal piece of the tail onto the midpiece [7]. Guinea pig PH20 glycoprotein (gpPH20) is unusual in that it migrates from one type of membrane, the postacrosomal plasma membrane, onto a different type of membrane, the inner acrosomal membrane, after exocytosis of the acrosomal vesicle [5]. The rat orthologue of PH20 (original designation, 2B1, retained throughout this paper for the sake of consistency in the literature; [10, 11]) is initially restricted to the sperm tail; but as part of the general membrane changes that accompany capacitation, it is transported across putative diffusion barriers in the neck and equatorial segment regions and accumulates on the acrosomal plasma membrane [9]. Sequence analyses of gpPH20 and its species orthologues have revealed significant homology to bee venom hyaluronidase, and, as a consequence, they have been found to degrade hyaluronic acid [1115]. The final positioning of a membrane-bound hyaluronidase over the rat sperm head suggests that it facilitates penetration through the mass of cumulus cells that surround all mammalian eggs. This hypothesis is supported by the finding that apigenin, an inhibitor of hyaluronidase, blocks penetration of mouse spermatozoa through the cumulus [13]. It is also consistent with an earlier study indicating that immunization of adult guinea pigs with homologous PH20 causes reversible infertility [16].

Notwithstanding these functional implications, the mechanisms regulating polarized migration of gpPH20 and 2B1, as well as the specific effects of endoproteolytic cleavage on their hyaluronidase activity, are still problematic. In both cases, migration is temperature- and Ca2+-dependent but insensitive to metabolic poisons or cytochalasins, properties inconsistent with a role for the cytoskeleton [1719]. Guinea pig PH20 contains a glycosyl phosphatidylinositol (GPI) anchor that is attached to the antigen during synthesis in round spermatids and consequently it has a high diffusion coefficient that is compatible with its ability to migrate from one region of the spermatozoon to another [20]. However, attempts to demonstrate a lipid anchor on 2B1 antigen by phosphatidylinositol phospholipase C (PI-PLC) digestion of whole spermatozoa, or of plasma membrane vesicles derived therefrom, have been unsuccessful [19]. Sequence analysis of 2B1 revealed a hydrophobic domain of approximately 23 amino acid residues toward the C-terminus suggestive of a membrane spanning region [11]. Like gpPH20 [21], 2B1 is endoproteolytically cleaved into a disulfide-bridged heterodimer during sperm passage through the epididymis [19], but it not known whether cleavage is obligatory for its ability to migrate during capacitation or how it affects hyaluronidase activity.

In this investigation, therefore, we have addressed two questions concerning rat sperm 2B1 antigen. First, does 2B1 contain the necessary sequence motif for attachment of a GPI anchor? Second, does endoproteolytic cleavage into a cross-linked heterodimer during epididymal maturation enhance or inhibit 2B1‘s hyaluronidase activity? Briefly, our results show that when Chinese hamster ovary (CHO) cells are transfected with 2B1 cDNA, the expressed protein contains a PI-PLC-sensitive lipid anchor that is required for transport to the plasma membrane. Endoproteolysis of the uncleaved precursor (i.e., testicular) form of sperm 2B1 antigen causes a significant shift in the optimum pH for hyaluronidase activity. These observations emphasize the importance of posttesticular processing of surface membrane antigens for regulating the fertilizing capacity of spermatozoa.

Materials and Methods

Materials

All chemicals and enzymes were of the highest purity available commercially and unless stated otherwise were purchased from Sigma Chemical Co. (London, UK), BDH-Merck (Lutterworth, UK), Pharmacia-LKB (St Albans, UK), or Boehringer Corp. (Lewes, UK). Antibody sources were as follows. A monoclonal antibody (IgG2a subclass) to 2B1 glycoprotein was available from earlier studies [10]. A rabbit polyclonal antibody against calreticulin was obtained from Affinity BioReagents (Golden, CO). Fluorescein isothiocyanate (FITC)-conjugated rabbit anti-mouse (FITC-RAM), tetramethylrhodamine B isothiocyanate (TRITC)-conjugated sheep anti-rabbit (TRITC-SAR), peroxidase-conjugated sheep anti-rabbit (Px-SAR), and peroxidase-conjugated rabbit anti-mouse (PX-RAM) antibodies were purchased from Dako Laboratories (High Wycombe, UK). A rabbit polyclonal antibody specific to trans Golgi network marker antigen 38 (TGN-38) was generously provided by Dr. George Banting, University of Bristol, UK. GMEM culture medium was supplied by Gibco-BRL (Paisley, UK). The mammalian cell expression vector, pEE14, was obtained from Celltech Biologics plc (Slough, UK), and a mutant CHO cell line (G9PLAP.85) deficient in the second step of GPI anchor biosynthesis [22] was a generous gift of Dr. Victoria L. Stevens, Emory University School of Medicine, Atlanta, Georgia.

Collection and Extraction of Spermatozoa and Testicular Cells

Spermatozoa were collected from the cauda epididymidis of adult male rats (Porton Wistar strain) by mincing the tissue in 5 ml PBS, pH 7.2, containing 5 mM glucose and 1 mM 4-(2-aminoethyl)-benzenesulphonyl fluoride hydrochloride (AEBSF). Tissue fragments were allowed to settle at room temperature for 5 min and motile spermatozoa in the supernatant recovered by centrifugation at 400 × g for 10 min. Sperm pellets were either resuspended in the appropriate assay buffer for measuring hyaluronidase activity (see later) or extracted for 40 min at 4°C in PBS containing 1% Triton X-100/1 mM AEBSF followed by centrifugation at 12 000 × g for 10 min at 4°C.

For collection of testicular cells, testes from 32-day-old rats were decapsulated and shaken gently in 50 ml serum-free GMEM medium containing 0.25 mg/ml collagenase for 20 min at 22°C. The released seminiferous tubules were washed once in 30 ml GMEM and resuspended in 20 ml GMEM containing 0.25 mg/ml collagenase, 0.03 mg/ml DNase I, and 5 mg/ml BSA; they were then incubated with shaking for 60 min at 22°C. The cell suspension was dispersed by 5 strokes in a loose-fitting Potter homogenizer, and any large tissue fragments remaining were allowed to settle for 5 min. The supernatant, containing a mixture of mainly primary spermatocytes and spermatids, was centrifuged at 400 × g for 5 min. Cells were washed twice more in 10 ml serum-free GMEM and either extracted with 1% Triton X-100/1 mM AEBSF or resuspended in the appropriate assay buffer for PI-PLC as described below.

Testicular spermatozoa were obtained by puncturing the extratesticular rete testis of adult rats whose efferent ductules had been ligated for 18–20 h [19]. Spermatozoa were washed once in PBS containing 5 mM glucose and 1 mM AEBSF, and pellets were either extracted with 1% Triton X-100 or resuspended in the appropriate assay buffer as described below.

Culture of CHO Cells

Both normal and G9PLAP.85 CHO cells were cultured in GMEM medium containing 10% fetal calf serum and 100 IU/ml penicillin on glass coverslips at 37°C in 95% air, 5% CO2. Cell numbers were counted using a hemocytometer, and viability was assessed by trypan blue exclusion.

Construction of pEE14–2B1 Eukaryotic Expression Vector and Transfection of CHO Cells

Specific polymerase chain reaction primers, flanked by EcoRI sites, were made to the extreme ends of rat 2B1 cDNA coding region [11]. Exact sequences of primers were forward primer (5′-GAGATGCAGAATTCGCTAGCTCTCCGTAATGGGATAGTTGC-3′), reverse primer (5′-TCCGAACTGAATTCGCTAGCCTAAGTAGTACTGACTAGCATC-3′). After 10 cycles of polymerase chain reaction (denaturing at 94°C for 1.5 min, annealing at 58°C for 1.5 min, and elongation at 72°C for 2 min), the resulting product (1586 base pairs) was gel purified and cloned into SmaI-cut, dephosphorylated, pUC18 vector DNA (Pharmacia); the sequence was verified on an ABI 377 automated sequencer (Perkin-Elmer, Norwalk, CT). The cDNA insert was then excised with EcoRI restriction endonuclease and recloned in the correct reading frame into the mammalian expression vector pEE14 according to the supplier's instructions (Celltech). The 9.4-kilobase pEE14 vector contains a glutamine synthetase (GS) minigene derived from a hamster GS genomic fragment. This selectable marker has a single intron and GS polyadenylation signals, and expression is driven from an SV40 late promoter. GS provides the only pathway for glutamine synthesis in mammalian cells (using glutamate and ammonia as substrates). Thus, in a glutamine-free medium, GS is an essential enzyme. CHO cells contain endogenous GS enzyme, but methionine sulfoximine (MSX) in excess of 20–25 μM is sufficient not only to inhibit wild-type levels of GS, but also to prevent the growth of the majority of natural variants that arise by amplification of endogenous GS genes. The pEEE14 vector contains a human cytomegalovirus promoter-enhancer (hCMV-MIE) to express the gene of interest at a high level.

Cultured cells were transfected with the pEE14–2B1 construct using DOSPER liposomal transfection reagent (Boehringer) according to the manufacturer's instructions. Forty-eight hours after transfection, selection was initiated by replacement with medium containing 25 μM MSX. Surviving clones were reselected 3–4 wk later with 500 μM MSX. To assay for expression of 2B1 antigen, untransfected CHO cells, transfected CHO cells, and transfected G9PLAP.85 cells were washed in PBS and lysed with 1% SDS for 5 min. Extracts were clarified by centrifugation at 12 000 × g for 5 min and analyzed by nonreducing SDS-PAGE followed by Western blotting onto polyvinylidene fluoride membranes. Blots were probed with 2B1 monoclonal antibody supernatant (or 1:500-diluted 2B1 ascites fluid) followed by 1:10 000-diluted peroxidase rabbit anti-mouse IgG, and bound antibody was detected by chemiluminescence (POD substrate; Boehringer) on Fuji (Tokyo, Japan) x-ray film.

Incubation of Cells with PI-PLC

The standard incubation system consisted of 100 μl of washed cells (∼106/ml) suspended in either serum-free GMEM or 0.264 M sucrose, 10 mM Hepes, pH 7.2, containing 0.5 U PI-PLC (supplied by Sigma or Boehringer or ICN) and incubated at 37°C for 1 h. Samples were centrifuged at 12 000 × g for 10 min; supernatants were removed, and sperm pellets were washed once in 200 μl of GMEM followed by extraction in 1% Triton X-100 containing 1 mM AEBSF for 30 min at 4°C. Supernatants and detergent extracts were analyzed by SDS-PAGE/Western blotting and probed with 2B1 monoclonal antibody/1:10 000-diluted peroxidase-RAM. Control samples contained either no PI-PLC or the complete mixture plus 5 mM ZnCl2 or 100 μM AlCl3 to inhibit the enzyme.

In other experiments, spermatozoa and testicular cells were pretreated with saponin (0.2 mg/ml) or 0.5 M KCl at pH 7.4 or 100 mM NaHCO3 at pH 11.5 for 30 min at 22°C (4°C for saponin) followed by washing in GMEM and incubation with PI-PLC as described above. Supernatants and Triton X-100 extracts were collected and analyzed by SDS-PAGE and Western blotting. CD52 glycoprotein was radiolabeled by the galactose oxidase/sodium [3H]borohydride technique [23], and labeled proteins were separated by SDS-PAGE and detected by fluorography using standard protocols.

Indirect Immunofluorescence and Confocal Microscopy

2B1 glycoprotein was visualized on the surface of spermatozoa and cultured CHO cells using a monoclonal antibody (IgG2a) followed by FITC-RAM IgG as previously described [9]. For detection of intracellular 2B1 antigen, cultured cells were fixed in 100% methanol at −20°C for 5 min followed by 2 washes in PBS at 22°C. Nonspecific binding was blocked with 3% BSA in PBS for 15 min, and cells were incubated with primary and secondary antibodies as described above. Similar protocols were followed for detection of TGN-38 with rabbit polyclonal antibody 44 [24] and calreticulin with rabbit polyclonal antibody PA3–900 [25], followed by incubation in 1:200 TRITC-SAR IgG. Cells were mounted in Mowiol anti-fade solution (Calbiochem-Novabiochem, San Diego, CA) before viewing.

Fluorescence was observed with a Zeiss Axiophot photomicroscope (Carl Zeiss, Thornwood, NY) or scanned with a Leica TCS-NT confocal laser scanning microscope equipped with a Kr/Ar laser (488-, 568-, and 647-nm lines) attached to a Leica DM RBE upright epifluorescence microscope (Leica, Solms, Germany). All images were processed with Leica software for 2D image analysis.

Fluorescence-Activated Cell Sorter (FACS) Analysis of CHO Cells Stained with 2B1 Monoclonal Antibody

Release of 2B1 glycoprotein from the surface of CHO cells after PI-PLC treatment was verified by analysis on a FACS Excalibor cell sorter. Control and PI-PLC-digested cells were stained with 2B1 monoclonal antibody/FITC-RAM and sorted at a rate of approximately 5000 cells/sec at a wavelength of 488 nm with a 530 band pass filter and setting of 400 mW. Debris and cell dimers were gated out using forward and side light scatter.

Assay of Hyaluronidase Activity

Hyaluronidase activity in suspensions of whole cells or Triton X-100 extracts was measured by a microplate turbidimetric assay as described previously [26] using hyaluronic acid as substrate. Activity is expressed in Units using Sigma hyaluronidase (H-3506) as reference standard or as relative activity under defined conditions. Blanks contained all the reagents except enzyme sample. For determining pH optima of hyaluronidases from different sources, a series of 10 mM sodium phosphate buffers were adjusted to different pHs with 1 mM HCl or 1 mM NaOH at 22°C. Final pHs were checked with a pH electrode in scaled-up samples containing all of the reagents. Inhibitors were dissolved in dimethyl sulfoxide (apigenin) or ethanol (kaempferol, quercetin) or buffer (heparin, IgGs) and were added to assay mixtures to the final concentrations as indicated in Results. Total protein was measured by the dye binding procedure [27] using BSA as standard or by the BCA assay (Pierce, Rockford, IL) when interfering levels of detergents were present.

Results

Attempts to Release 2B1 Glycoprotein from Spermatozoa with PI-PLC

Since previous attempts to demonstrate a GPI anchor on 2B1 glycoprotein on cauda spermatozoa had been unsuccessful [11], we next investigated whether 1) the antigen on immature germ cells (round spermatids and spermatozoa collected from the testis and caput epididymidis) would be more susceptible to PI-PLC or 2) preexposure of cauda spermatozoa to mild dissociating agents would improve accessibility of the enzymes to a GPI anchor. The rationale behind these experiments was that the plasma membranes of round spermatids and spermatozoa in the epididymis are significantly different in composition from those on cauda spermatozoa [28]. On spermatids and immature spermatozoa, therefore, or on cauda spermatozoa from which loosely bound extrinsic membrane proteins had been extracted, the putative GPI anchor on 2B1 antigen may become more susceptible to cleavage by enzymes.

The results of these experiments are shown in Fig. 1. As reported previously, 2B1 antigen could not be solubilized from washed cauda spermatozoa by incubation with 5 U/ml of PI-PLC. Preexposure of spermatozoa to 0.5 M KCl or pH 11.5 or pH 4.5 or saponin did not improve the situation, as all of the immunoreactive 2B1 antigen was recovered in the pellet. PI-PLC treatment also failed to solubilize 2B1 antigen from round spermatids or testicular spermatozoa or spermatozoa from the caput epididymidis (results not shown).

Fig. 1

A) Western blot of supernatant (S) and pellet (P) fractions from cauda spermatozoa that had been either left untreated, incubated with PI-PLC or preincubated in pH 4.5, pH 11.5, 0.5 M KCl, or saponin followed by PI-PLC treatment. Arrow indicates position of 2B1 glycoprotein. B) Fluorograph of supernatant and pellet fractions from cauda spermatozoa prelabeled with galactose oxidase/sodium [3H]borohydride followed by PI-PLC digestion. The major labeled glycoprotein is CD52 which contains a GPI anchor and is susceptible to PI-PLC

Fig. 1

A) Western blot of supernatant (S) and pellet (P) fractions from cauda spermatozoa that had been either left untreated, incubated with PI-PLC or preincubated in pH 4.5, pH 11.5, 0.5 M KCl, or saponin followed by PI-PLC treatment. Arrow indicates position of 2B1 glycoprotein. B) Fluorograph of supernatant and pellet fractions from cauda spermatozoa prelabeled with galactose oxidase/sodium [3H]borohydride followed by PI-PLC digestion. The major labeled glycoprotein is CD52 which contains a GPI anchor and is susceptible to PI-PLC

As positive controls for the above experiments, CD52 glycoprotein on cauda spermatozoa was radiolabeled with galactose oxidase/sodium [3H]borohydride as previously described [23] and digested with PI-PLC. CD52 is a major GPI-linked glycoprotein on rat spermatozoa [23, 29] that is known to be susceptible to PI-PLC. As shown in Fig. 1, ∼50% of the glycoprotein appeared in a soluble supernatant fraction, confirming the efficacy of the enzyme treatment and incubation conditions.

It has to be concluded from these experiments, therefore, that 2B1 antigen on rat spermatozoa either does not have a classic GPI anchor or that if such an anchor is present, it is highly resistant to hydrolysis by PI-PLC.

Expression of Recombinant 2B1 Glycoprotein in CHO Cells

In view of the foregoing results, we next investigated whether 2B1 glycoprotein contained internal sequence information for attachment of a GPI anchor. For this purpose, CHO cells (which are very efficient in synthesis of GPI) were transfected with pEE14 containing 2B1 cDNA, and transformed cells were selected in MSX medium (see Materials and Methods). Controls consisted of 1) CHO cells that were not transfected, 2) transfected G9PLAP.85 mutant CHO cells that are unable to complete the GPI attachment step, and 3) transfected CHO cells grown in the presence of 10 mM d-mannosamine, a known inhibitor of glycan extension during GPI synthesis [30]. Expressed 2B1 antigen was detected in cell lysates by Western blotting and on whole cells by fluorescence staining with 2B1 monoclonal antibody/FITC-RAM. The endoplasmic reticulum (ER) and Golgi were colocalized with anti-calreticulin/TRITC-SAR and anti-TGN-38/TRITC-SAR, respectively.

Nontransfected CHO cells did not stain with 2B1/FITC-RAM (results not shown). In contrast, transfected cells bound 2B1/FITC-RAM strongly over their entire plasma membrane with irregularly shaped aggregates appearing in the cytoplasm, particularly in a small region adjacent to the nucleus (Fig. 2a). The cytoplasm also stained strongly with anti-calreticulin/TRITC-SAR (Fig. 2b) for ER. When the foregoing confocal images were superimposed, it became apparent that the 2B1-positive aggregates were localized within the Golgi and ER (Fig. 2c), suggesting that they represented newly synthesized 2B1 glycoprotein en route to the surface. The densely stained area adjacent to the nucleus stained positively with TGN-38/TRITC-SAM antibody, identifying it as the trans Golgi network in which many glycoproteins are posttranslationally processed during protein trafficking (Fig. 2, j–l). Transfected mutant G9PLAP.85 CHO cells also expressed 2B1, but in this case the glycoprotein accumulated in large amounts within the ER and Golgi, as shown by colocalization with calreticulin and TGN-38 antibodies, respectively, and could not be detected on the surface membrane (Fig. 2, d–f; compare with Fig. 2, a–c). A similar distribution was observed after the addition of the inhibitor 10 mM d-mannosamine to the culture medium (Fig. 2, g–i). Retention within the ER and trans Golgi network is characteristic of glycoproteins deficient in GPI anchors [31, 32].

Fig. 2

Indirect immunofluorescence detection of 2B1 glycoprotein, calreticulin and TGN-38 in CHO cells as detected by confocal laser microscopy. a) Normal CHO cells transfected with 2B1 cDNA and stained with 2B1 monoclonal antibody/FITC-RAM. Note strong fluorescence on the surface membrane. b) Same cells as in a double-stained with anti-calreticulin/TRITC-SAR. Fluorescence is only observed intracellularly. c) Superimposition of images shown in a and b. d) G9PLAP–85 cells transfected with 2B1 cDNA and stained with 2B1 monoclonal antibody/FITC-RAM. Note strong fluorescence in the cytoplasm with the nucleus appearing unstained. e) Same cells as in d double-stained with anti-calreticulin/TRITC-SAR. f) Superimposition of images in d and e. g) Normal CHO cells cultured in the presence of 10 mM d-mannosamine followed by 2B1/FITC-RAM. Note cytoplasmic fluorescence. h) Same cells as in g stained with calreticulin/TRITC-SAR. i) Superimposition of images in h and g. j) Normal CHO cells transfected with 2B1 cDNA and stained with 2B1/FITC-RAM. k) Same cells as in j doubled stained with TGN-38/TRITC-SAR. Note the strong reaction in the Golgi region. l) Superimposition of images in j and k. The Golgi and surface membrane are stained red and green, respectively. ×630

Fig. 2

Indirect immunofluorescence detection of 2B1 glycoprotein, calreticulin and TGN-38 in CHO cells as detected by confocal laser microscopy. a) Normal CHO cells transfected with 2B1 cDNA and stained with 2B1 monoclonal antibody/FITC-RAM. Note strong fluorescence on the surface membrane. b) Same cells as in a double-stained with anti-calreticulin/TRITC-SAR. Fluorescence is only observed intracellularly. c) Superimposition of images shown in a and b. d) G9PLAP–85 cells transfected with 2B1 cDNA and stained with 2B1 monoclonal antibody/FITC-RAM. Note strong fluorescence in the cytoplasm with the nucleus appearing unstained. e) Same cells as in d double-stained with anti-calreticulin/TRITC-SAR. f) Superimposition of images in d and e. g) Normal CHO cells cultured in the presence of 10 mM d-mannosamine followed by 2B1/FITC-RAM. Note cytoplasmic fluorescence. h) Same cells as in g stained with calreticulin/TRITC-SAR. i) Superimposition of images in h and g. j) Normal CHO cells transfected with 2B1 cDNA and stained with 2B1/FITC-RAM. k) Same cells as in j doubled stained with TGN-38/TRITC-SAR. Note the strong reaction in the Golgi region. l) Superimposition of images in j and k. The Golgi and surface membrane are stained red and green, respectively. ×630

When expressed 2B1 glycoprotein was extracted from transfected CHO cells and analyzed by SDS-PAGE/Western blotting, it migrated with a molecular mass of ∼60 kDa under nonreducing conditions (Fig. 3A). This compares favorably with the estimated size of molecular mass 58 kDa for native testicular sperm 2B1 [19]. 2B1 expressed in G9PLAP85 cells was slightly smaller in size (molecular mass ∼59 kDa) and less heterogenous than that in normal CHO cells, presumably because of incomplete processing. More significantly, when normal CHO cells expressing surface 2B1 were incubated with PI-PLC under conditions identical to those described for spermatozoa in Fig. 1, > 80% of the antigen was solubilized and recovered in supernatant fractions (Fig. 3B). The addition of divalent cations such as Zn2+ and Al3+, which are known to inhibit PI-PLC activity, either prevented, or substantially inhibited, solubilization of 2B1 glycoprotein, thus confirming the specificity of the enzyme. Ca2+ ions, however, did not have an inhibitory effect.

Fig. 3

A) Western blot of Triton X-100 extracts of 2B1 transfected and untransfected CHO cells (normal and mutant G9PLAP.85 cell lines). B) Western blot of supernatant (S) and pellet (P) fractions from transfected normal CHO cells expressing 2B1 glycoprotein on their surface membrane following treatment with PI-PLC in the presence and absence of Ca2+, Zn2+, and Al3+. Arrows indicate position of 2B1 glycoprotein

Fig. 3

A) Western blot of Triton X-100 extracts of 2B1 transfected and untransfected CHO cells (normal and mutant G9PLAP.85 cell lines). B) Western blot of supernatant (S) and pellet (P) fractions from transfected normal CHO cells expressing 2B1 glycoprotein on their surface membrane following treatment with PI-PLC in the presence and absence of Ca2+, Zn2+, and Al3+. Arrows indicate position of 2B1 glycoprotein

Confirmation that 2B1 glycoprotein was released from the surface of transfected CHO cells by PI-PLC was obtained by FACS analysis. PI-PLC-Treated cells showed a clear shift in peak fluorescence intensity relative to non-PI-PLC-treated cells (Fig. 4).

Fig. 4

FACS analysis of fluorescence intensity on 2B1 transfected CHO cells before (solid line) and after (dashed line) treatment with PI-PLC. Note the shift to the left following PI-PLC indicating a decrease in surface staining intensity (Fl1-H)

Fig. 4

FACS analysis of fluorescence intensity on 2B1 transfected CHO cells before (solid line) and after (dashed line) treatment with PI-PLC. Note the shift to the left following PI-PLC indicating a decrease in surface staining intensity (Fl1-H)

Taken together, these results indicate that 2B1 glycoprotein contains the correct internal sequence motif for attachment of a GPI anchor and suggests that this is likely to be its means of attachment to the sperm plasma membrane despite its resistance to PI-PLC.

Effects of Endoproteolytic Cleavage on Hyaluronidase Activity of 2B1 Glycoprotein

During passage of spermatozoa through the epididymis, 2B1 glycoprotein undergoes endoproteolytic cleavage between residues Arg312 and Ser313 to produce a 2-chain molecule held together by one or more disulfide bridges. This processing can be mimicked in vitro by controlled incubation of immature caput spermatozoa with pancreatic trypsin [19]. However, it is not known how internal cleavage and the presumed conformational changes that ensue affect the hyaluronidase activity of 2B1 glycoprotein [11]. To investigate this problem, equal numbers of washed testicular and cauda spermatozoa were incubated with hyaluronic acid at pH 6.6 and 7.1. Since previous work has shown that ∼84% of total hyaluronidase activity in homogenates of rat spermatozoa is associated with the membrane fraction [11], it is reasonable to presume that this represents 2B1 glycoprotein. Results showed that testicular spermatozoa contained 31% (at pH 6.6) and 37% (at pH 7.1) of the activity of equal numbers of cauda spermatozoa (means of 3 experiments). When activity was expressed relative to micrograms of total protein, the respective values were 41% at pH 6.6 and 44% at pH 7.1. Thus, although endoproteolytic cleavage of 2B1 glycoprotein is not obligatory for its hyaluronidase activity, transformation from a single- to a 2-chain molecule clearly has an enhancing effect.

To gain further insights into this phenomenon, we next investigated the relative activity of cleaved and uncleaved 2B1 glycoprotein over a range of pHs. A pH profile of hyaluronidase activity in Triton X-100 extracts of cauda spermatozoa revealed a broad range of activity between 5.5 and 8.5 with an optimum at pH 6.6 (Fig. 5A). A similar broad pH range was observed for Triton X-100 extracts of testicular sperm except that the optimum was at pH 7.1. Treatment of testicular sperm with pancreatic trypsin before Triton X-100 extraction, however, lowered the pH optimum to 6.6; i.e., it behaved like the cauda form of 2B1. Thus, internal cleavage of 2B1 glycoprotein clearly affected its enzymic properties.

Fig. 5

Activity of hyaluronidase from rat spermatozoa (A and C), CHO cells (B), bull testis (D), and guinea pig spermatozoa (E) as a function of pH. See text for details of preparation of the extracts

Fig. 5

Activity of hyaluronidase from rat spermatozoa (A and C), CHO cells (B), bull testis (D), and guinea pig spermatozoa (E) as a function of pH. See text for details of preparation of the extracts

Recombinant 2B1 glycoprotein expressed in CHO cells also showed hyaluronidase activity, but in this case the pH profile was different from that observed for either rat testicular or cauda spermatozoa. Instead of a clearly definable peak of activity, a wide plateau was observed between pH 3.0 and 6.5 with a rapid decrease toward pH 8.5 (Fig. 5B).

Most (∼85%) of rat sperm hyaluronidase is membrane bound, with the remainder released by sonication into a soluble supernatant fraction [11]. To investigate whether there were any significant differences between these two forms of hyaluronidase (from cauda sperm), activity profiles were measured between pH 3 and 8. Optimum peaks of activity were observed at pH 6.7 for the membrane-bound enzyme and from pH 6.0 to 6.7 for the soluble form (Fig. 5C).

For comparison, pH activity profiles were investigated for soluble and membrane-bound forms of hyaluronidase released from guinea pig spermatozoa (cauda) and from commercially available bull testicular hyaluronidase (Sigma). Bull testicular hyaluronidase showed maximal activity at pH 3.9 that fell to ∼50% of this level between pH 5.0 and 8.0 (Fig. 5D). The soluble form of guinea pig sperm hyaluronidase also reached peak activity below pH 4.0 (maximum at pH 3.7) with a gradual decrease to a slightly lower level between pH 4.5 and 7.5 (Fig. 5E). By contrast, the membrane-bound form was considerably less active below pH 4.0, only reaching a plateau similar to the soluble form above pH 5.0 (Fig. 5E).

Hyaluronidase activity (measured at pH 6.6) in Triton X-100 extracts of rat cauda spermatozoa was inhibited by between 75% and 85% in the presence of 250 μM heparin or apigenin, or kaempferol or quercetin (Table 1). High levels of 2B1 IgG (in the form of 1:100-diluted ascites fluid) inhibited activity by ∼61% whereas control ascites fluid, containing a monoclonal antibody (2F7) that does not bind to rat spermatozoa, inhibited it by only ∼25%.

Table 1

Effects of various inhibitors, antibodies, detergents and solvents on the hyaluronidase activity of 2B1 glycoprotein.*

Compound % Inhibition 
1. Triton X-100 (1%) 
2. Triton X-100 (1%) + heparin (0.25 mg/ml) 33 
3. Triton X-100 (1%) + heparin (1.0 mg/ml) 94 
4. Triton X-100 (1%) + dimethylsulphoxide (1%) 
5. Triton X-100 (1%) + apigenin (100 μM) 30 
6. Triton X-100 (1%) + apigenin (250 μM) 88 
7. Triton X-100 (1%) + ethanol (1%) 
8. Triton X-100 (1%) + ethanol (1%) + kaempferol (100 μM) 20 
9. Triton X-100 (1%) + ethanol (1%) + kaempferol (250 μM) 78 
10. Triton X-100 (1%) + ethanol (1%) + quercetin (100 μM) 49 
11. Triton X-100 (1%) + ethanol (1%) + quercetin (250 μM) 76 
12. Triton X-100 (1%) + ethanol (1%) + 2B1 ascites (5 μl) 36 
13. Triton X-100 (1%) + ethanol (1%) + 2B1 ascites (20 μl) 61 
14. Triton X-100 (1%) + ethanol (1%) + 2F7 ascites (20 μl) 25 
Compound % Inhibition 
1. Triton X-100 (1%) 
2. Triton X-100 (1%) + heparin (0.25 mg/ml) 33 
3. Triton X-100 (1%) + heparin (1.0 mg/ml) 94 
4. Triton X-100 (1%) + dimethylsulphoxide (1%) 
5. Triton X-100 (1%) + apigenin (100 μM) 30 
6. Triton X-100 (1%) + apigenin (250 μM) 88 
7. Triton X-100 (1%) + ethanol (1%) 
8. Triton X-100 (1%) + ethanol (1%) + kaempferol (100 μM) 20 
9. Triton X-100 (1%) + ethanol (1%) + kaempferol (250 μM) 78 
10. Triton X-100 (1%) + ethanol (1%) + quercetin (100 μM) 49 
11. Triton X-100 (1%) + ethanol (1%) + quercetin (250 μM) 76 
12. Triton X-100 (1%) + ethanol (1%) + 2B1 ascites (5 μl) 36 
13. Triton X-100 (1%) + ethanol (1%) + 2B1 ascites (20 μl) 61 
14. Triton X-100 (1%) + ethanol (1%) + 2F7 ascites (20 μl) 25 
*

Values shown are selected points from titration curves (5 measurements) to illustrate the range of inhibition by the various compounds; each experiment was performed at least twice.

Discussion

This work has shown that rat sperm 2B1 glycoprotein (the orthologue of gpPH20) contains a C-terminal sequence motif for attachment of a GPI anchor. Together with its ability to bind Triton X-114 [11], the evidence suggests that despite the failure of PI-PLC to release the antigen from spermatozoa, 2B1 is attached to the plasma membrane via a lipid anchor. Both immature (testicular) and mature (epididymal) forms of 2B1 show hyaluronidase activity, although endoproteolytic processing into a disulfide-linked dimer (such as occurs during sperm maturation in the epididymis) causes a significant reduction in the pH for optimal enzyme activity.

Criteria for Determining That a Membrane Protein Contains a GPI Anchor

GPI is a widespread linkage for attaching glycoproteins to cell membranes, and more than 50 examples are now known in eukaryotic systems (reviewed in [33]). To demonstrate the presence of such an anchor conclusively, four properties of the glycoprotein have to be investigated. One must examine, first, the ability of the glycoprotein to bind detergents such as Triton X-114; second, release of the antigen from the membrane with PI-PLC or phosphatidylinositol phospholipase D (PI-PLD) or nitrous acid; and third, the presence of a C-terminal motif that satisfies the “w, w + 2 rule” [34] for predicting a GPI attachment site. Fourth, failing any of the above, the glycoprotein should be expressed and correctly localized in a heterologous system known to be efficient in synthesizing and attaching GPI moieties.

Previous experiments on 2B1 glycoprotein demonstrated its tenacity for the sperm plasma membrane, which, together with its partitioning into Triton X-114, suggested that it was an integral membrane protein [11, 19]. Unlike guinea pig and mouse PH20, however, rat 2B1 could not be solubilized from whole spermatozoa or isolated plasma membranes derived therefrom with PI-PLC or PI-PLD [11]. Attempts to cleave the glucosamine-glycan bond with nitrous acid were also inconclusive because of the lability of the epitope recognized by the monoclonal antibody to acidic pH (unpublished results). Resistance of GPI-anchored membrane proteins to PI-PLC (or PI-PLD) is not, in fact, unusual [35]. It has been attributed to various factors such as an additional fatty acyl chain on the inositol ring [36] and/or the presence of N-linked oligosaccharide chains on an asparagine residue adjacent to the GPI attachment site [37]. Since these are covalent modifications made to a protein during its biosynthesis, they offer an explanation why the extraction of loosely bound proteins from the surface membrane with high salt solutions, pH 11.0 buffers, saponin, trypsin, etc., is unable to confer PI-PLC sensitivity on 2B1; in short, resistance to cleavage is an inherent property of the glycoprotein.

A sequence analysis of 2B1 revealed hydrophobic domains at both the N- and C-termini [11], the latter being a characteristic feature of nascent forms of GPI-containing proteins. During processing in the ER, the hydrophobic C-terminal peptide is removed to reveal a specific (previously internal) residue for GPI attachment. From an analysis of over 20 authenticated GPI-containing proteins, Kodukula et al. [33, 34] have compiled a w, w + 2 rule to predict putative GPI attachment motifs with an accuracy of 80%, a figure similar to that for identifying the cleavage site of N-terminal signal peptides [38]. The rule predicts that the attachment site w can be one of several residues with small side chains (Ala, Asn, Asp, Cys, Gly, or Ser), w + 1 may be any amino acid except Pro or Trp, and w + 2 is usually Gly or Ala. This is succeeded by a stretch of 8–12 hydrophilic amino acids and subsequently by a 15- to 30-residue hydrophobic peptide. Analysis of the C-terminal sequence of 2B1 reveals w, w + 1, and w + 2 sites at Asn437, Ser438, and Ala439, respectively, followed by 7 hydrophilic residues (Phe440 to Lys446) and 31 hydrophobic residues (Gly447 to Pro477). The w, w + 2 index is 0.8 (cf. a value of 1.0, meaning most probable). Thus, theoretically a sequence motif for attachment of a GPI anchor is present in rat 2B1. A similar motif can be identified near the C-terminus of guinea pig and mouse PH20 [11].

The veracity of this site was shown by expressing rat recombinant 2B1 in CHO cells. The glycoprotein was transported via classic pathways involving the ER to the plasma membrane from which it was readily released by hydrolysis with exogenous PI-PLC. The evidence, therefore, satisfies 3 of the 4 criteria referred to above and is overwhelmingly in favor of 2B1 attachment to the sperm plasma membrane via a GPI anchor. The GPI anchor also helps explain its susceptibility to antibody-induced patching [10] and its ability to migrate from the tail onto the acrosomal domain during capacitation [9], as proteins with GPI anchors have higher diffusion coefficients than typical type I transmembrane proteins [21]. It is of interest that the other major GPI-anchored glycoprotein on rat spermatozoa is CD52, which is readily released by PI-PLC [23, 29].

Functionality of Rat 2B1 Glycoprotein as a Membrane Bound Hyaluronidase

The discovery that gpPH20 and its species orthologues in the rat, mouse, monkey, and human show hyaluronidase activity is consistent with a putative role in facilitating sperm penetration through the mass of cumulus oophorus cells that surround mammalian eggs [13]. This hypothesis is enforced by the appropriate location of the antigen over the acrosomal plasma membrane (mouse, monkey, human, and capacitated rat spermatozoa; [15, 19, 39]) and by the ability of hyaluronidase inhibitors, such as apigenin, to block sperm penetration through the cumulus [13, 39]. The presence of a membrane-bound form of hyaluronidase, however, presents something of a conundrum, as the enzyme has long been regarded as a soluble intra-acrosomal protein whose release correlates closely with damage to the plasma membrane and acrosome. In fact, there are considerable species differences. In bull spermatozoa, > 97% of hyaluronidase activity is intracellular and appears in the cytosol following sonication, whereas in rat, human, and monkey spermatozoa the converse is true: > 85% is recoverable in the membrane pellet with the remaining 15% in the cytosol [11]. Guinea pig spermatozoa are closer to bull spermatozoa, with 78% soluble and 22% membrane bound. Ultrastructural localization with gold-labeled antibodies to PH20 has revealed that a proportion of the bound form is also found on the inner acrosomal membrane in human and monkey spermatozoa [15]. Similar studies on rat 2B1 have detected the antigen only on the surface membrane and not in the acrosomal contents [9], which correlates closely with the data mentioned above on the partitioning behavior of hyaluronidase activity.

In a detailed study of gpPH20, Hunnicutt et al. [25] concluded that 1) the primary structures of the soluble and bound forms are very similar, if not identical, and that the soluble form is derived from the bound form during the acrosome reaction by removal of the GPI anchor, thereby implying the presence of an intra-acrosomal PI-PLC. PI-PLC activity has been described in bull and goat spermatozoa [40, 41], but its intracellular location is not clear from these studies. 2) Membrane-bound PH20 undergoes endoproteolytic cleavage into a disulfide-linked dimer, whereas soluble PH20 remains uncleaved despite containing the protease-sensitive site between residues Arg311 and Ser312. 3) Membrane-bound and soluble PH20 have different pH activity curves. An unresolved question with gpPH20, however, is whether endoproteolytic cleavage has any effect on its hyaluronidase activity as a function of pH, as uncleaved membrane-bound PH20 cannot be recovered from acrosome-intact spermatozoa. Rat 2B1 overcomes this problem as it is predominantly membrane-bound and is normally cleaved during sperm maturation in the epididymis. Our results show that testicular (uncleaved) 2B1 has a pH optimum at 7.1 whereas cauda (cleaved) 2B1 has maximum activity at pH 6.6. Trypsinization of testicular 2B1 under conditions that hydrolyze the Arg312-Ser313 site causes a shift in peak activity from pH 7.1 to pH 6.6. Thus, the structural changes that must ensue after cleavage of 2B1 clearly have an affect on its activity as a hyaluronidase. Like soluble PH20 from guinea pig and human spermatozoa, the small amount of soluble 2B1 has a more acidic pH profile than membrane-bound 2B1 (pH 6.2 vs. pH 6.6, respectively). It would seem, therefore, that removal of the GPI anchor and endoproteolytic cleavage combine to produce a soluble enzyme with a lower pH optimum than that of its bound counterpart. In this respect it would be of interest to investigate the small amount of bound hyaluronidase activity in bull spermatozoa, as the soluble form has a very acidic pH optimum of 3.6.

The physiological significance of these different pH activity profiles of soluble and bound sperm hyaluronidase will become apparent only when more information is available about the microenvironment within the mass of cumulus oophorus cells and on the surface of the zona pellucida. Traces of hyaluronic acid are present in the zona [42], and it has been speculated that sperm hyaluronidase causes a localized dissolution or softening as a prelude to sperm binding to zona glycoproteins [25]. Hunnicutt et al. [25] have further suggested on the basis of differential antibody inhibition experiments that gpPH20 is bifunctional, i.e., that it behaves first of all as a hyaluronidase and then functions as a secondary binding molecule to prevent release of acrosome-reacted spermatozoa from the zona surface. The epitopes recognized by the inhibiting monoclonal antibodies to gpPH20 have not been identified; hence it is not known whether hyaluronidase activity and zona binding activity reside in the same or different domains of the protein. To some extent, data on rat 2B1 support this concept of separate functional domains. 2B1 monoclonal antibody has been shown to inhibit fertilization in vitro [43] but has only a weak inhibitory effect on hyaluronidase activity. This can be explained on the basis that 2B1 monoclonal antibody binds to an epitope near the N-terminus [11] whereas site-directed mutagenesis experiments on human hyaluronidase indicate that the crucial residues for enzyme activity lie within the middle third of the sequence [44]. Thus, a monoclonal antibody to an N-terminal epitope would not be expected to block enzyme activity unless it caused steric hindrance or conformational changes as a consequence of binding.

Potential of 2B1/PH20 Glycoprotein as a Contraceptive Immunogen

Inhibition of fertilization in vitro by monoclonal antibodies to various sperm antigens is well documented [45], although there is no consensus as to the relative importance of these antigens or the way in which they relate to one another (if at all). The report [16] that reversible infertility could be achieved in adult male and female guinea pigs following immunization with purified homologous PH20 prompted considerable interest in this antigen as a contraceptive immunogen. Unfortunately, the efficacy of PH20 has not been confirmed in other species. Immunization of female rabbits with homologous recombinant PH20 had no significant effect on their fertility despite high titers of circulating antibodies [46]. We have also investigated the effects of immunizing male and female rats with bacterially expressed fragments of 2B1 representing nucleotides 436–879, 898–1375, and 1376–1863 [11]. Despite high titers of anti-2B1 antibodies in the blood of all immunized animals (> 1:12 500 dilution required to achieve preimmune levels in ELISA assays), there was no significant effect on their fertility (as judged by the number of implantation sites) following natural matings (unpublished results). These findings cast doubt on the claims for PH20 as a contraceptive immunogen, with the proviso that in the guinea pig, native sperm PH20 was used, whereas all the recent studies have relied on recombinant PH20. It is possible that specific epitopes on the native protein, not found on bacterially expressed PH20, are crucial for providing the correct immunological response that causes infertility. A further complication is that male guinea pigs immunized with gpPH20 readily develop autoimmune orchitis leading to azoospermia in many animals [47]. This was not the case in our studies in the rat, as there were no overt signs of testicular damage or loss of spermatozoa from the cauda epididymidis. The potential of PH20 as a contraceptive immunogen, therefore, remains controversial and may be species specific.

In conclusion, the likelihood that 2B1 antigen on rat spermatozoa contains a GPI anchor helps to explain its ability to migrate from the tail domain to the acrosomal domain during capacitation while remaining within the plane of the plasma membrane [9]. The mechanisms that enable it to move against a large concentration gradient and across putative intramembranous barriers, however, are major questions that remain to be answered [48]. Nonetheless, the ability of spermatozoa to process and reposition an antigen that has clear functional relevance (to help spermatozoa penetrate through the cumulus oophorus) illustrates their ability to self-regulate their fertilizing capacity and respond to external signals without transcribing new proteins.

Acknowledgments

We thank Dr. George Banting (Bristol University, UK) for the gift of TGN-38 antibody, Dr. Victoria Stevens (Emory University, Atlanta, GA) for generously providing the G9PLAP.85 cell line, Mr. Nigel Miller (BI) for the FACS analysis, and Mrs. Alicia Ma (BI) for able technical assistance.

References

1.
Eaton
S
,
Simons
K
.
Apical, basal and lateral cues for epithelial polarization.
Cell
 
1995
;
82
:
5
8
.
2.
Ikonen
E
,
Simons
K
.
Protein and lipid sorting from the trans-Golgi network to the plasma membrane of polarized cells.
Semin Cell Dev Biol
 
1998
;
9
:
503
509
.
3.
Rodriguez-Boulan
E
,
Nelson
JW
.
Morphogenesis of the polarized epithelial cell phenotype.
Science
 
1989
;
245
:
718
725
.
4.
Blobel
CP
,
Myles
DG
,
Primakoff
P
,
White
JH
.
Proteolytic processing of a protein involved in sperm-egg fusion correlates with acquisition of fertilization competence.
J Cell Biol
 
1990
;
111
:
69
78
.
5.
Myles
DG
,
Primakoff
P
.
Localized surface antigens of guinea-pig sperm migrate to new regions prior to fertilization.
J Cell Biol
 
1984
;
99
:
1634
1641
.
6.
Phelps
BM
,
Myles
DG
.
The guinea pig sperm plasma membrane protein, PH-20, reaches the surface via two transport pathways and become localized to a domain after an initial uniform distribution.
Dev Biol
 
1987
;
123
:
63
72
.
7.
Petruszak
JAH
,
Nehume
CL
,
Bartles
JR
.
Endoproteolytic cleavage in the extracellular domain of the integral plasma membrane protein CE9 precedes its redistribution from the posterior to the anterior tail of the rat spermatozoan during epididymal maturation.
J Cell Biol
 
1991
;
115
:
917
927
.
8.
Scully
NF
,
Shafer
JH
,
Shur
BD
.
Spatial and temporal expression of cell surface galactosyltransferase during mouse spermatogenesis and epididymal maturation.
Dev Biol
 
1987
;
124
:
111
124
.
9.
Jones
R
,
Shalgi
Hoyland J
,
Phillips
DM
.
Topographical rearrangement of a plasma membrane antigen during capacitation of rat spermatozoa in vitro.
Dev Biol
 
1990
;
139
:
349
362
.
10.
Gaunt
SJ
,
Brown
CR
,
Jones
R
.
Identification of mobile and fixed antigens on the plasma membrane of rat spermatozoa using monoclonal antibodies.
Exp Cell Res
 
1983
;
144
:
275
284
.
11.
Hou
S-T
,
Ma
A
,
Jones
R
,
Hall
L
.
Molecular cloning and characterization of rat sperm surface antigen 2B1 a glycoprotein implicated in sperm-zona binding.
Mol Reprod Dev
 
1996
;
45
:
193
203
.
12.
Gmachl
M
,
Kreil
G
.
Bee venom hyaluronidase is homologous to a membrane protein of mammalian sperm.
Proc Natl Acad Sci USA
 
1993
;
90
:
3569
3573
.
13.
Lin
Y
,
Mahan
K
,
Lathrop
WF
,
Myles
DG
,
Primakoff
P
.
A hyaluronidase activity of the sperm plasma membrane protein PH-20 enables sperm to penetrate the cumulus cell layer surrounding the egg.
J Cell Biol
 
1994
;
125
:
1157
1163
.
14.
Thaler
CD
,
Cardullo
RA
.
Biochemical characterization of a glycosylphosphatidylinositol-linked hyaluronidase on mouse sperm.
Biochemistry
 
1995
;
34
:
7788
7795
.
15.
Cherr
GN
,
Meyers
SA
,
Yudin
AI
,
VandeVoort
CA
,
Myles
DG
,
Primakoff
P
,
Overstreet
JW
.
The PH-20 protein in Cynomolgus macaque spermatozoa: identification of two different forms exhibiting hyaluronidase activity.
Dev Biol
 
1996
;
175
:
142
153
.
16.
Primakoff
P
,
Lathrop
W
,
Woolman
L
,
Cowan
AE
,
Myles
DG
.
Fully effective contraception in male and female guinea pigs immunized with the sperm protein PH-20.
Nature
 
1988
;
335
:
543
546
.
17.
Primakoff
P
,
Hyatt
H
,
Myles
DG
.
A role for the migrating sperm surface antigen PH-20 in guinea pig sperm binding to the egg zona pellucida.
J Cell Biol
 
1985
;
101
:
2239
2244
.
18.
Phelps
BM
,
Primakoff
P
,
Koppel
DE
,
Low
MG
,
Myles
DG
.
Restricted lateral diffusion of PH-20, a PI anchored sperm membrane protein.
Science
 
1988
;
240
:
1780
1782
.
19.
Jones
R
,
Ma
A
,
Hou
S-T
,
Shalgi
R
,
Hall
L
.
Testicular biosynthesis and epididymal endoproteolytic processing of rat sperm surface antigen 2B1.
J Cell Sci
 
1996
;
109
:
2561
2570
.
20.
Cowan
AE
,
Myles
DG
,
Koppel
DE
.
Migration of the guinea pig sperm protein PH-20 from one localized surface domain to another does not occur by a simple diffusion-trapping mechanism.
Dev Biol
 
1991
;
144
:
189
198
.
21.
Phelps
BM
,
Koppel
DE
,
Primakoff
P
,
Myles
DG
.
Evidence that proteolysis of the surface is an initial step in the mechanism of formation of sperm cell surface domains.
J Cell Biol
 
1990
;
111
:
1839
1847
.
22.
Stevens
VL
,
Zhang
H
,
Harreman
M
.
Isolation and characterization of a Chinese hamster ovary (CHO) mutant defective in the second step of glycosylphosphatidylinositol biosynthesis.
Biochem J
 
1996
;
313
:
253
258
.
23.
Moore
A
,
White
TW
,
Ensrud
KM
,
Hamilton
DW
.
The major maturation glycoprotein found on rat cauda epididymal sperm surface is linked to the membrane via phosphatidylinositol.
Biochem Biophys Res Commun
 
1989
;
160
:
460
468
.
24.
Wilde
A
,
Reaves
B
,
Banting
G
.
Epitope mapping of 2 isoforms of a trans-Golgi network specific integral membrane protein TGN38/41.
FEBS Lett
 
1992
;
313
:
235
238
.
25.
Denning
GM
,
Leidal
KG
,
Holst
VA
,
Iyer
SS
,
Pearson
DW
,
Clark
JR
,
Nauseef
WM
,
Clark
RA
.
Calreticulin biosynthesis and processing in human myeloid cells: demonstration of signal peptide cleavage and N-glycosylation.
Blood
 
1997
;
90
:
372
381
.
26.
Hunnicutt
GR
,
Primakoff
P
,
Myles
DG
.
Sperm surface protein PH-20 is bifunctional: one activity is a hyaluronidase and a second, distinct activity is required in secondary sperm-zona binding.
Biol Reprod
 
1996
;
55
:
80
86
.
27.
Bradford
MM
.
A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principal of protein dye binding.
Anal Biochem
 
1979
;
72
:
248
254
.
28.
Jones
R
.
Plasma membrane structure and remodelling during sperm maturation in the epididymis.
J Reprod Fertil Suppl
 
1998
;
53
:
73
84
.
29.
Hale
G
,
Rye
PD
,
Warford
A
,
Lauder
I
,
Britobabapulle
A
.
The glycophosphatidylinositol-anchored lymphocyte antigen CDW52 is associated with the epididymal maturation of human spermatozoa.
J Reprod Immunol
 
1993
;
23
:
189
205
.
30.
Lisanti
MP
,
Rodriguez-Boulan
E
.
Glycophospholipid membrane anchoring provides clues to the mechanism of protein sorting in polarized epithelial cells.
TIBS
 
1989
;
15
:
113
118
.
31.
Delahurity
MD
,
Stafford
FJ
,
Yuan
LC
,
Shay
D
,
Bonifacino
JS
.
Uncleaved signals for glycosylphosphatidylinositol anchoring cause retention of precursor proteins in the endoplasmic reticulum.
J Biol Chem
 
1993
;
268
:
12017
12027
.
32.
Field
MC
,
Moran
P
,
Li
WL
,
Keller
GA
,
Caras
IW
.
Retention and degradation of proteins containing an uncleaved glycosylphosphatidylinositol signal.
J Biol Chem
 
1994
;
269
:
10830
10837
.
33.
Udenfriend
S
,
Kodukula
K
.
How glycosylphosphatidylinositol-anchored membrane proteins are made.
Annu Rev Biochem
 
1995
;
64
:
563
591
.
34.
Kodukula
K
,
Gerber
LD
,
Aruthauer
R
,
Briuk
L
,
Udenfriend
S
.
Biosynthesis of glycosylphosphatidylinositol (GPI)-anchored membrane proteins in intact cells. Specific amino acid requirements adjacent to the site of cleavage and GPI attachment.
J Cell Biol
 
1993
;
120
:
657
664
.
35.
Huang
KS
,
Li
S
,
Low
MG
.
Glycosylphosphatidylinositol-specific phospholipase D.
Methods Enzymol
 
1991
;
197
:
567
575
.
36.
Roberts
WL
,
Myhar
JJ
,
Kuksis
A
,
Low
MG
,
Rosenberry
TL
.
Lipid analysis of the glycoinositol phospholipid membrane anchor of human-erythrocyte acetylcholinesterase. Palmitoylation of inositol results in resistance to phosphatidylinositol-specific phospholipase C.
J Biol Chem
 
1988
;
263
:
18766
18775
.
37.
Guthor
MLS
,
Dealmeida
MLC
,
Rosenberry
TL
,
Ferguson
MAJ
.
The detection of phospholipase-resistant and phospholipase-sensitive glycosyl-phosphatidylinositol membrane anchors by western blotting.
Anal Biochem
 
1994
;
219
:
249
255
.
38.
von Heijne
G
.
A new method for predicting signal sequence cleavage sites.
Nucleic Acids Res
 
1986
;
14
:
4683
4691
.
39.
Sabeur
K
,
Cherr
GN
,
Yudin
AI
,
Primakoff
P
,
Li
M-W
,
Overstreet
JW
.
The PH-20 protein in human spermatozoa.
J Androl
 
1997
;
18
:
151
158
.
40.
Vanha-Perttula
T
,
Kasurinen
J
.
Purification and characterization of phosphatidylinositol-specific phospholipase-C from bovine spermatozoa.
Int J Biochem
 
1989
;
21
:
997
1007
.
41.
Bansal
P
,
Atreja
SK
.
Phosphoinositide specific phospholipase-C activity of goat spermatozoa in transit from the caput to the cauda epididymis.
Ind J Biochem Biophys
 
1991
;
28
:
307
311
.
42.
Talbot
P
.
Hyaluronidase dissolves a component in the hamster zona pellucida.
J Exp Zool
 
1984
;
229
:
309
316
.
43.
Shalgi
S
,
Matilyahu
A
,
Gaunt
SJ
,
Jones
R
.
Identification of antigens on rat spermatozoa with a potential role in fertilization.
Mol Reprod Dev
 
1990
;
25
:
286
296
.
44.
Arming
S
,
Strobl
B
,
Wechselberger
C
,
Kreil
G
.
In vitro mutagenesis of PH-20 hyaluronidase from human sperm.
Eur J Biochem
 
1997
;
247
:
810
814
.
45.
McLeskey
SB
,
Dowds
C
,
Carballada
R
,
White
RR
,
Saling
PM
.
Molecules involved in mammalian sperm-egg interaction.
Int Rev Cytol
 
1998
;
177
:
57
113
.
46.
Holland
MK
,
Andrews
J
,
Clarke
H
,
Walton
C
,
Hinds
LA
.
Selection of antigens for use in a virus-vectored immunocontraceptive vaccine: PH-20 as a case study.
Reprod Fertil Dev
 
1997
;
9
:
117
124
.
47.
Tung
KSK
,
Primakoff
P
,
Woolman-Gamer
L
,
Myles
DG
.
Mechanism of infertility in male guinea pigs immunized with sperm PH-20.
Biol Reprod
 
1997
;
56
:
1133
1141
.
48.
Cowan
AE
,
Nakhimovsky
L
,
Myles
DG
,
Koppel
DE
.
Barriers to diffusion of plasma membrane proteins form early during guinea pig spermiogenesis.
Biophys J
 
1997
;
73
:
507
516
.

Author notes

1
We are grateful to the BBSRC for financial support and for award of a postgraduate CASE studentship to one of us (G.J.S.).