A sperm membrane protein that binds in a species-specific manner to the egg extracellular matrix is homologous to von Willebrand factor.

We have purified a sperm membrane protein, designated zonadhesin, that binds in a species-specific manner to the extracellular matrix (zona pellucida) of the egg, and cloned its cDNA. The cDNA encodes a novel protein with a single transmembrane segment separating a 36 amino acid, highly basic intracellular C terminus from a 2418-amino acid extracellular region. The extracellular sequence specifies a mosaic protein comprising a unique N-terminal domain, a mucin-like domain, and five tandem domains proximal to the membrane that are homologous to prepro von Willebrand factor. The N-terminal and mucin-like domains were absent from zonadhesin that bound to the egg extracellular matrix, suggesting that processing occurs during sperm maturation and/or capacitation. By Northern blotting and in situ hybridization, zonadhesin mRNA was detected only within the testis, where it was expressed primarily in haploid spermatids. The unique domain structure of zonadhesin suggests multiple functions, one of which is to mediate sperm adhesion to the zona pellucida.

Fertilization in both invertebrates and vertebrates is species-specific, and events such as motility activation, gamete adhesion, and induction of the acrosome reaction show absolute or relative species specificity (1)(2)(3)(4)(5). Thus, sperm membrane proteins that interact with the egg in a species-specific manner are likely to be primary gamete recognition and/or signaling components. Efforts to reach a molecular understanding of mammalian fertilization have been hampered by the apparent asynchrony of the sperm population; at any given time only a fraction of spermatozoa possess the ability to fertilize the egg. To overcome this problem we used the porcine egg extracellular matrix (zona pellucida) as an affinity medium to isolate proteins from large quantities of pig sperm membranes (6). One of these proteins bound in a species-specific manner to the egg extracellular matrix and migrated in SDS-PAGE at M r 150,000 under nonreducing conditions and as M r 105,000 and 45,000 subunits (p105 and p45) after disulfide bond reduction (6). We have now purified this sperm membrane protein (named zonadhesin), obtained partial amino acid sequence, and isolated its cDNA. Zonadhesin is expressed by the haploid spermatid and is homologous to both von Willebrand factor and mucins.
RNA Isolation-Total RNA was isolated by homogenizing tissues in guanidinium thiocyanate and N-lauroyl sarcosine, extracting with acidic phenol/CHCl 3 , and precipitating with isopropyl alcohol (8). Poly(A) ϩ RNA was isolated from total RNA by oligo(dT)-cellulose chromatography (Fast Track Kit, Invitrogen). To obtain mRNAs with intact 5Ј ends for 5Ј-RACE, RNAs larger than 5 kb in were purified from total RNA by gel filtration on a 0.5 ϫ 6-cm column of Sephacryl S-1000 in 10 mM Tris-HCl pH 7.5, 200 mM NaCl, 10 mM vanadyl ribonucleoside complex.
cDNA Cloning-PCR (denatured at 94°C for 1 min, annealed at 50°C for 2 min, extended at 72°C for 3 min; 40 cycles) with pig testis cDNA as template (53 ng/100 l reaction) and degenerate oligonucleotide primers (sense: GAATTCGAATTCGA(A/G)GGICA(A/G)CCIC-CIGCITT(C/T)TA(C/T)(C/T)T; antisense: GGATCCGGATCCCAIGCIG-GIGC(C/T)TG(A/G)AAIGCIGC(C/T)TG; 0.5 M each) based on the sequences of two tryptic peptides (EGQPPAFYL and AQEQCQAAFQA-PAWANCAT, respectively), produced a 500-bp product that was cloned and sequenced (9) to confirm that it encoded part of p105. The cloned insert was then labeled with [␣-32 P]dCTP and used to screen oligo(dT)and random hexamer-primed pig testis cDNA libraries constructed in ZAP II (10). Stringency washes were 3 ϫ 15 min at 65°C in 0.1 ϫ SSC, 0.2% SDS, 0.1% Na 4 P 2 O 7 . Phage clones were rescued as pBluescript plasmids, and inserts were then sequenced with or without prior subcloning into M13 filamentous phages. Approximately 1.3 kb of 5Ј end cDNA was cloned by 5Ј-RACE (5Ј Amplifinder Kit, Clontech) using nested antisense primers (reverse transcription primer: TGGAAGGGT-GGGTTCTTTAAAGAC; downstream PCR primer: GACAGAGAAT-TCAGAAGGTCCTCGACAGGT; upstream PCR primer: GACGAAT-TCAAATATCAGCTGCATGGGCTG). PCR for 5Ј-RACE (denatured at 94°C for 0 s, annealed at 55-65°C for 0 s, extended at 72°C for 30 s) was done using an air-driven thermal cycler (Idaho Technologies). Primary amplification products (35 cycles with the downstream PCR primer) were excised from low melting point agarose gels, and 0.25% of each was used as template for re-amplification for 30 -35 cycles with the upstream PCR primer. RACE products were cloned in pBluescript and sequenced by double-stranded cycle sequencing using an Applied Biosystems automated sequencer. Both strands of cloned cDNAs were sequenced at least once, and the composite sequence was assembled and analyzed using DNASTAR software. The presence of a polyadenylation signal and a poly(A) tail in the 60-base 3Ј-untranslated sequence and of in-frame stop codons in the 297 base 5Ј-untranslated sequence confirmed that the cloned cDNAs spanned the entire zonadhesin mRNA.
In Situ Hybridization-Paraffin-embedded sections of pig testis (Novagen) were de-waxed and hybridized with digoxigenin-labeled RNA probes, washed at high stringency, and digoxigenin was then detected by incubation with alkaline phosphatase-conjugated anti-digoxigenin followed by color development with 5-bromo-4-chloro-3-indoylphosphate and nitro blue tetrazolium (12, 13). Sense and antisense probes derived from the 1566-bp BamHI-SalI 3Ј end fragment of the zonadhesin cDNA were synthesized with T3 or T7 polymerases using templates of appropriately linearized pBluescript plasmids containing the 1566-bp insert.

RESULTS AND DISCUSSION
Zonadhesin was purified to apparent homogeneity based on its zona pellucida binding activity (Fig. 1). Approximately 30 g of p105 and 10 -15 g of p45 were obtained from 880 mg of sperm membrane protein, using 44 mg of native, particulate zona pellucida as the affinity matrix. Amino acid sequences of eight p105 and five p45 tryptic peptides (Table I) were not present in existing protein sequence data bases, suggesting that zonadhesin was a novel protein.
Degenerate oligonucleotide primers based on the sequences of two p105 tryptic peptides were used to clone cDNAs encompassing the zonadhesin coding sequence by a combination of PCR, cDNA library screening, and 5Ј-RACE. The 7785-base composite sequence of the cDNAs contains a single major open reading frame, and the 2476-amino acid sequence deduced from it includes the sequences of the eight p105 and five p45 peptides in its C-terminal two-thirds (Table I and Fig. 2).
A 29-amino acid putative signal peptide (14) is present at the N terminus of the deduced sequence (Fig. 2a). The sequence also predicts a 2418-amino acid extracellular region, a single membrane-spanning segment, and a 36-amino acid intracellular C terminus containing numerous basic residues. Cleavage of the putative signal peptide at serine 29 would produce a 2447-amino acid mature polypeptide chain with a calculated molecular molecular mass of 267,000 Da. This predicted mo-  (21). Note the characteristic alignment of nearly all of the cysteine residues and presence of the CGLCG motif in D1 and D2. c, C-terminal sequence. The transmembrane segment is underlined. In b, alignment was by the clustal method using the PAM 250 alignment matrix. Pairwise alignment parameters were: window size ϭ 5, gap penalty ϭ 3. Multiple alignment parameters were: gap penalty ϭ 10, gap length penalty ϭ 10.  (15); the changes that occur in either environment could involve proteolysis of zonadhesin. A requirement for sperm surface proteolytic activity to facilitate sperm adhesion to the zona pellucida has been reported (16). A region of highly repetitive sequence is present between amino acids 300 and 700 of zonadhesin; it comprises 53 imperfect repeats of the consensus sequence PTE(K/R)(P/T)T(V/I). Repetitive sequences rich in proline and threonine are characteristic of mucins, which are O-glycosylated on numerous serines and threonines and have extended structures owing to their high proline content (17). These structural properties reflect the functions of mucins in regulating cellular interactions; the large, carbohydrate-rich domains extend beyond most other cell surface glycoproteins and thereby either inhibit or promote cell adhesion (17,18). The mucin-like domain of zonadhesin could function similarly during sperm migration through the male or female reproductive tracts, inhibiting inappropriate trapping of spermatozoa or promoting adhesion to the oviductal isthmus, which serves as a sperm reservoir in several species (19,20).
Five homologous domains, in tandem, follow the mucin-like domain (Fig. 2b). Sequence comparisons (PIR data base) revealed that these domains are homologous to the D-domains of prepro-von Willebrand factor (vWF) (21). Conservation of nearly all of the cysteines within the five domains is readily apparent, and several other amino acids are also conserved at various positions in the sequences. The D0 domain of zonadhesin is about one-fourth the length of the D1-D4 domains, and it truncates at the same point in its sequence as a partial Ddomain present in prepro-vWF. The sequence CGLCG of prepro-vWF D1-D3 domains is conserved in the zonadhesin D1 and D2 domains. The vicinal cysteines in this sequence may mediate covalent oligomerization of prepro-vWF monomers (22). Four vWF-like D-domains were recently identified in the human intestinal mucin MUC2 (23), which also exists as disulfide-bonded oligomers and possesses the CGLCG motif in its D1 and D3 domains. We previously observed high M r sperm proteins (SDS-PAGE, nonreducing conditions) that bound to the porcine zona pellucida and appeared to be covalent oligomers of zonadhesin (6). Thus, the CGLCG motif in the D1 and D2 domains of zonadhesin may mediate covalent oligomerization, which in turn could be important in sperm adhesion to the zona pellucida because of the increased binding avidity of multivalent interactions. The zona pellucida is itself a repetitive structure, comprising strands of noncovalently associated multimers of ZP2-ZP3 heterodimers (24), that could interact productively with an oligomeric receptor. Oligomerization could also promote membrane protein aggregation that may be important for induction of the sperm acrosome reaction (25,26).
Northern blots of poly(A) ϩ RNAs from several pig tissues detected expression of a 7.5-8 kb zonadhesin mRNA in testis, but not in heart, liver, lung, spleen, brain, kidney, or epididymis (Fig. 3). Because testis is a heterogeneous tissue containing several different cell types, we localized zonadhesin expression . Note that the field comprises primarily seminiferous tubules, with three blood vessels and some connective tissue. b, hybridization of an antisense RNA probe to a parallel section (same field as a), viewed with dark field illumination. c, hybridization of a sense RNA probe to a parallel section (same field, illumination, and exposure as b). Comparing a-c, specific hybridization of the antisense RNA probe is confined to the seminiferous tubules (positions of two blood vessels are marked for orientation). Only a subset of seminiferous tubules express high levels of the zonadhesin mRNA. d, phase contrast view (no counterstain) of two adjacent seminiferous tubules (boxed region of b). Note the difference in the level of zonadhesin expression between the two tubules. Expression in the strongly stained tubule is high in haploid spermatids. The clear zones appearing as halos surrounding the large round nuclei of diploid spermatocytes (arrowheads) indicate an absence of expression in the cytoplasm of these cells. more precisely by in situ hybridization (Fig. 4). Some but not all of the seminiferous tubules in testicular sections hybridized strongly with the zonadhesin probe, and this hybridization was restricted to the germinal epithelia of the tubules. The apparent expression of zonadhesin mRNA by only a fraction of the tubules in a given tissue section probably reflects the asynchrony of spermiogenesis among tubules. Within strongly hybridizing tubules, expression was detected only in the haploid spermatids. Spermatid-specific expression of zonadhesin supports the hypothesis that this protein mediates sperm-specific function(s).
The domain structures of zonadhesin, prepro-vWF, and MUC2 are illustrated schematically in Fig. 5. Each is a mosaic protein (27) and among them zonadhesin is smallest and is unique in having a putative transmembrane segment and its D-domains in tandem. The mucin-like domain of zonadhesin is also a substantially smaller fraction of the total protein than is the mucin part of MUC2, which is a prototype for secreted mucins.
In addition to their functions in oligomerization, vWF Ddomains bind heparin (28). This property is of potential interest, because heparin and/or other glycosaminoglycans promote capacitation and the acrosome reaction of bovine and hamster spermatozoa in vitro (29); however, the molecular target(s) for these agents have not been identified. It is possible that the D-domains of zonadhesin mediate these heparin/glycosaminoglycan effects. Like heparin, zona pellucida glycoproteins contain sulfated carbohydrate (30); this further suggests that zonadhesin's binding activity may derive in part from interaction of its D-domains with sulfated carbohydrates in the egg extracellular matrix.
Zonadhesin differs from other potential gamete recognition proteins of spermatozoa for which cDNAs have been cloned (31)(32)(33)(34). The substrate specificity of sperm surface galactosyltransferase, a M r 60,000 plasma membrane-associated variant of the Golgi enzyme, is consistent with its hypothesized function in gamete adhesion (35). Monoclonal and polyclonal antibodies to PH-20, a M r 56,000 -64,000 sperm surface hyaluronidase, inhibit adhesion of guinea pig spermatozoa to the zona pellucida (32,36). Recently, two different proteins potentially involved in mouse (33) and human (34) sperm-egg adhesion (sp56 and ZRK, respectively) were also described. Bookbinder et al. (33) reported a correlation between detectability of sp56 in spermatozoa from various species and ability of the cells to adhere to mouse eggs; species specificity has not been addressed for the other molecules. Thus, zonadhesin is presently the only protein known to bind in a species-specific manner to the egg extracellular matrix.