Overexpressing Sperm Surface f31A-Galactosyltransferase in Transgenic Mice Affects Multiple Aspects of Sperm-Egg Interactions

Sperm surface/31,4-galaetosyltransferase (GalTase) mediates fertilization in mice by binding to specific O-linked oligosaccharide ligands on the egg coat glycoprotein ZP3. Before binding the egg, sperm GalTase is masked by epididymally derived glycosides that are shed from the sperm surface during capacitation. After binding the egg, sperm-bound oligosaecharides on ZP3 induce the acrosome reaction by receptor aggregation, presumably involving Gal'lhse. In this study, we asked how increasing the levels of sperm surface GalTase would affect sperm-egg interactions using transgenic mice that overexpress GalTase under the control of a heterologous promoter. GalTase expression was elevated in many tissues in adult transgenie animals, including testis. Sperm from transgenic males had approximately six times the wild-type level of surface GalTase protein, which was localized appropriately on the sperm head as revealed by indirect immunofluorescence. As expected, sperm from transgenie mice bound more radiolabeled ZP3 than did wild-type sperm. However, sperm from transgenic animals were relatively unable to bind eggs, as compared to sperm from wild-type animals. The mechanistic basis for the reduced egg-binding ability of transgenic sperm was attributed to alterations in two GalTase-dependent events. First, transgenic sperm that overexpress surface GalTase bound more epididymal glycoside substrates than did sperm from wild-type mice, thus masking GalTase and preventing it from interacting with its zona pellucida ligand. Second, those sperm from transgenic mice that were able to bind the zona pellueida were hypersensitive to ZP3, such that they underwent precocious acrosome reactions and bound to eggs more tenuously than did wildtype sperm. These results demonstrate that sperm-egg binding requires an optimal, rather than maximal, level of surface GalTase expression, since increasing this level decreases sperm reproductive efficiency both before and after egg binding. Although sperm GalTase is required for fertilization by serving as a receptor for the egg zona pellucida, excess surface GalTase is counterproductive to successful sperm-egg binding. M AMMALIAN fertilization is initiated when a sperm binds to the zona pellucida matrix surrounding the ~gg. The initial interaction between sperm and egg is a carbohydrate-mediated process in which a receptor on the sperm surface binds to its glycoside ligand on the zona pellucida (Miller and Ax, 1990). In mouse, the zona pellucida ligand has been identified as a class of oligosaccharides bound in an O-glycosidic linkage to the ZP3 polypeptide, one of three glycoproteins that constitute the zona pellucida * These two authors contributed equally to this work. David J. Miller's present address is Department of Animal Sciences, University of Illinois, 132 Animal Sciences Laboratory, 1207 Gregory Drive, Urbana, IL 61801. Address all correspondence to Dr. Barry D. Shur, Dept. of Biochemistry and Molecular Biology, Box 117, The University of Texas M. D. Anderson Cancer Center, 1515 Holcombe Blvd., Houston, TX 77030. Tel.: (713) 792-2555. Fax: (713) 790-0329. (Florman and Wassarman, 1985). Several sperm proteins have been suggested to function as receptors for ZP3 oligosaceharides (Bleil and Wassarman, 1990; Leyton et al., 1992; Miller et al., 1992). To date, only one candidate receptor, /31,4-galactosyltransferase (GalTase) t, satisfies all the criteria required of a sperm receptor for the zona pellueida (Lopez et al., 1985; Shur and Neely, 1988; Miller et al., 1992). GalTase is confined to the sperm plasma membrane overlying the acrosome, as expected for an egg receptor. Affinity-purifiedsperm surface GalTase competitively inhibits sperm-egg binding, as do GalTase-specific perturbants. Sperm GalTase selectively binds oligosaccharides on 1. Abbreviations usedin thispaper: DEPC, pyrocarbonic acid diethyl ester; GalTase, ~l,4-galactosyltransferase; GlcNAc, N-acetylglucosamine; hGH, human growth hormone; MT-I, metallothionein; NGS, normal goat serum; NOG, n-oetylglucopyranoside; nt, nucleotides; PAA, protein-A-agarose; PIC, protease inhibitor cocktail. © The Rockefeller University Press, 0021-9525/94/09/1573/11 $2.00 The Journal of Cell Biology, Volume 126, Number 6, September 1994 1573-1583 1573 on July 7, 2017 jcb.rress.org D ow nladed fom ZP3 that have sperm receptor activity; sperm GalTase does not bind irrelevant zona pellucida oligosaccharides. That the GalTase-binding site on ZP3 is required for its sperm receptor activity is shown by the loss of ZP3's biological activity following the specific removal or blockage of its oligosaccharide substrate for sperm GalTase. These and other observations demonstrate that surface GalTase is required for sperm-egg binding. Whether either of two other sperm proteins suggested to function as ZP3 receptors cooperate with GalTase during fertilization remains unclear. There is evidence that one of these proteins (p56) binds ZP3; however, it behaves as a peripheral membrane protein on the sperm surface (Cheng et al., 1994; Bleil and Wassarman, 1990). Another candidate ZP3 receptor (p95) has recently been identified as sperm hexokinase (Kalab et al., 1994). When sperm are first released from the epididymis, they are unable to bind the zona pellucida. This is due, at least in part, to masking of sperm GalTase by epididymal glycosides. Before reaching the egg, these epididymally derived glycosides are spontaneously shed from the sperm surface, thereby exposing the GalTase-binding site and enabling sperm to bind the zona (Shur and Hall, 1982a). Gal'l'ase substrates purified from epididymal fluids inhibit sperm-egg binding by competing for the GalTase-binding site; irrelevant glycoproteins that are not substrates for sperm GalTase have no effect on sperm-egg binding. When sperm arrive at the egg zona pellucida, sperm surface GalTase mediates fertilization by binding N-acetylglucosamine (GlcNAc)-terminated oligosaccharides on ZP3 (Lopez et al., 1985; Miller et al., 1992). Sperm binding to ZP3 induces the acrosome reaction, which is thought to be initiated by aggregation of sperm receptors by multivalent ZP3 oligosaccharides (Florman et al., 1984; Leyton and Saling, 1989). In this regard, each ZP3 glycoprotein has multiple GalTase-binding oligosaccharides (Miller et al., 1992) and aggregation of surface GalTase using multivalent antibodies is sufficient to initiate the acrosome reaction via a pertussis toxin-sensitive mechanism (Macek et al., 1991; Miller and Shur, unpublished observations). During this process, the plasma membrane overlying the acrosome and the adjacent outer acrosomal membrane are shed, releasing acrosomal proteases and N-acetylglucosaminidase, which enable sperm to penetrate the zona pellucida (Miller et al., 1993a). Egg activation by the fertilizing sperm is also accompanied by release of N-acetylglucosaminidase from cortical granules, which removes the GalTase-binding site from ZP3 oligosaccharides, thus mediating the block to polyspermic binding (Miller et al., 1993b). In light of the central role that sperm GalTase plays in murine fertilization, we explored the effects of overexpressing surface GalTase on the process of sperm-egg binding in transgenic mice. The gene for GalTase encodes two nearly identical proteins: the short GalTase contains 386 amino acids, whereas the long GalTase contains 399 amino acids (Russo et al., 1990). The additional 13 amino acids on the long form are located on the amino-terminal cytoplasmic domain and are required for transporting a portion of GalTase to the cell surface, where it functions as a cell adhesion molecule (Evans et al., 1993; Youakim et al., 1994; Evans and Shur, unpublished observations). Sperm express only the long form of GalTase (Shaper et al., 1990; Pratt and Shur, 1993), and consequently, transgenic animals used in this study were created by injecting a mini-gene specifically encoding long GalTase under the control of a heterologous promoter. We report here, that contrary to our initial expectations, overexpressing surface GalTase on sperm of two independently derived transgenic mouse lines resulted in decreased binding of sperm to eggs, compared to wild-type sperm. This decreased egg-binding ability of transgenic sperm was due to alterations in two GalTase-dependent events during fertilization. First, sperm from transgenic animals bound more epididymal glycosides than did wild-type sperm, effectively masking sperm surface GalTase from its zona pellucida ligand. Second, those sperm from transgenic mice that were able to bind the zona were hypersensitive to ZP3, such that they underwent acrosome reactions earlier than did wild-type sperm and bound to eggs more tenuously. Thus, simply increasing the level of surface GalTase on sperm does not lead to a subsequent increase in sperm binding, but rather, leads to decreased binding, illustrating that successful fertilization requires an optimal, rather than a maximal, level of sperm receptors for the egg. Materials and Methods Construction of EV-142 Expression Plasmids Containing Long GalTase The EV-142 expression vector contains the mouse metallothionein (M/:l) promoter fused to a fragment of the human growth hormone gene (hGH) containing the polyadenylation signal (Low et al., 1985). Construction of the plasmid containing long GaiTase cDNA was performed as described previously (Evans et al., 1993), with the following modification. The 4.7-kb fragment containing MT-1, PDLGT eDNA, and the hGH 3' untranslated sequences was excised from the vector by EcoRI digestion, and ligated into EcoRI-cut Bluescript (Stratagene Inc., La Jolla, CA). Intron 4 from the GalTase genomic DNA was inserted into the eDNA as follows: a 5-kb fragment was amplified from a partial genomic GalTase clone, using primers flanking a unique StuI restriction site within exon 3 and a unique BsmI site within exon 5. The polymerase chain reaction product was isolated, and digested with BsmI and NdeI and the appropriate 2-kb fragment containing intron 4 was ligated into the vector containing the GalTase eDNA that was previously digested with NdeI and BsmI. The resulting cDNA/genomic hybrid was sequenced across the ligation junctions and into the intron. Production and Identification of Transgenic Mice The EV-142 plasmid containing the long GaiTase cDNA/genomic hybrid insert was digested with EcoRI to release the insert, which was then purified. Transgenic mice were created by microinjecting the linearized insert into the pronuclei of mouse zygotes. The zygotes were isolated from B6D2F~ females mated to B6D2Fl males and were introduced into the ampullae of CD1 foster mothers. Offspring were genotyped by Southern blotting of DNA isolated from tails, using a 600-bp fragment of the hGH sequences as probe. Positive founder mice were mated with wild-type littermates to establish the Ft generation. Subsequently, genotyping was done by slot blot analysis of tail DNA, using hGHas probe and a 1-kb fragment of the major urinary protein (MUP) gene (Derman, 1981) as an internal standard for quantitation of the DNA. GalTase Enzyme Assays Adult tissues. To assay for inducibility of the transgene, 25 mM ZnSO4 was added to the drinking water 4 d before sacrifice. Tissues were isolated from adult male hemizygous transgenic or wild-type mice between 2 and 4 mo of age. Tissue was homogenized on ice in 1 ml of medium B (140 mM NaCI, 4 mM KC1, 20 mM Hepes; pH 7.2) containing freshly added protease inhibitor cocktail (PIC; 2/~g/ml antipain, 0.1% aprotinin, 10 ~g/ml benzamide, 1/~g/mi chymostatin, 1 ~tg/ml leupeptin, 1 t~g/ml pepstatin), with a Polytron (Brinkman Instrs., Inc., Westbury, NY) for 6 s at a setting of 8. Homogenates were cleared of debris by centrifugation at 13,000 g for 5 rain at 4°C. The supernatant was removed to a fresh tube, and n-octylglucoThe Journal of Cell Biology, Volume 126, 1994 1574 on July 7, 2017 jcb.rress.org D ow nladed fom pyranoside (NOG) was added to a final concentration of 30 mM. The sampies were incubated on ice for 2 h with frequent trimration. Insoluble material was removed by centrifugation for 5 min at 4"C. Protein concentrations were determined (Bradford, 1976). Enzyme assays wore conducted for 1 h at 37"C in a total volume of 50 ttl, and contained 10 pg protein, 100 t~M uridine 5tdiphosphate [3H]galactose (UDp[3I-I]Gai; 220 dpm/pmol; Dupont-New England Nuclear, Wilmington, DE), 1 mM Y-AMP, 10 mM MnC12, and 30 mM GlcNAc, in medium B/PIC. The reaction was stopped by addition of 10 #1 ice-cold 0.2 M EDTA-Tris-HC1, pH 7.2.50 ~tl was subjected to high-voltage borate electrophoresis to separate the 3H-labeled galactosylated product from unused UDp[3H]Gal and its breakdown products. The radiolabeled product remaining at the origin was quantitated by liquid scintillation spectroscopy. Sperm. Cauda epididymal sperm were collected into 2 ml of dmKRBT (120 mM NaCI, 2 mM KC1, 2 mM CaCI2, 10 mM NaI-ICO3, 1.2 mM MgSO4.7H20, 5.6 mM glucose, 1.1 mM sodium pyruvate, 25 mM TAPSO [3-{N-tr/smethylamino}-2-hydroxy propane sulfonlc acid], 18.5 mM sucrose, 6 rng/mi BSA, pH 7.3), filtered through a nytex filter after 15 min, and capacitated at 37"C for 1 h (Neill and Olds-Clarke, 1988). Sperm were then centrifuged three times in 5 ml of medium B/PIC at 800 g for 5 min to remove epididymal soluble GalTase. Sperm wore counted on a hemacytometer, and adjusted to give 1 x 105 sperm/35 ~tl. 35 /~1 of sperm was assayed for Garlhse activity as above, except that the cocktail included UDP[3H]Gal at 440 dpm/pmol. Only cell surface activity is assayed under these conditions, since sperm are unusual in that all of their GalTase is localized to the plasma membrane; there is no intracellular pool of biosynthetic GalTase in mature spermatozoa (Lopez and Shur, 1987; Shur and Ncely, 1988). An aliquot of the last wash was also assayed for activity to insure that all soluble epididymal GalTase had been removed from the washed sperm pellet. The assays were performed three times, each in duplicate. RNase Protection Assays RNA was isolated from testes of adult male hemizygous transgenic or wildtype mice, using the guanidinium method (Chomczynski and Sacchi, 1987); was resuspended in 100 /zl of pyrocarbonic acid diethyl ester (DEPC) dH20, and was quantitated by measuring absorbance at OI)26o. The probe for RNase protection assays was generated using a fragment of GalTase from -239 to +28% inserted into the pGEM 3Z vector (Promega Corp., Madison, WI), and linearized with HindHI. The 32p-labeled riboprobe ([32p]CTE 800 Ci/mmol; Amersham Corp., Arlington Heights, IL) was prepared by transcription from the T7 promoter, using the Ambion Maxiscript kit (Ambion, Inc., Austin, TX) according to the manufacturer's instructions and gel purified. 10 t~g of RNA was hybridized with 50,000 cpm of probe overnight at 58"C. RNase digestion, precipitation, and recovery of protected riboprobe was done using the Ambion RNase Protection kit, according to the manufacturer's instructions, and electrophoresed on a 5 % acrylamide/8 M urea gel. Control samples included probe combined with yeast RNA, either with or without RNase digestion. Immunoprecipitation of GalTase Activity Surface GaiTase activity was immunoprecipitated from sperm using an antiserum (anti-peptide 1) generated to the 13 amino acids unique to the cytoplasmic domain of long GalTase (Youakim et al., 1994). Briefly, sperm were isolated from the cauda epididymis as described above and washed three times in medium B/PIC. The washed cells were resuspended in medium B/PIC and NOG was added to a final concentration of 30 mM. The cells were solubilized at 4°C for 4 h with frequent trituration and were centrifuged at 13,000 g for 5 rain. The supernatant was assayed for protein concentration (Bradford, 1976). Concurrently, 50 ftl of protein-A-agarose (PAA; Pierce Chem. Co., Rockford, IL) that had been washed overnight in medium B/PIC/NOG with 5% BSA was incubated for 4 h at 4°C on a rotator with either 15/~1 antipeptide 1 serum or preimmune serum. The PAA-antibody complex was washed three times with medium B/PIC/NOG/BSA, after which 100 ttg of lysate was added. The complex was incubated overnight on a rocking platform. In some instances, peptide was added to the lysate (60 ttg/ml). The PAA-antibody complex was centrifuged and the GalTase activity remaining in the supernatant was assayed as described above. Western Immunoblotting Sperm were recovered, washed, and solubilized as described above for immunoprecipitation, except that BSA was omitted from all buffers. 250 #g of protein was electrophoresed on a 10% nonreducing polyacrylamide gel. Proteins were transferred to nitrocellulose by electrophoretic transfer for 4 h at room temperature. The filters wore blocked overnight at 4°C in PTB (PBS, pH 7.2, containing 0.02% Tween 20 and 5% BSA). The filters were incubated with primary antibodies (anti-peptide 1 immune and preimmune serum) diluted 1:100 in PTB, and washed three times (20 min/wash) with FIB. nSI-Labeled goat anti-rabbit IgG (ICN, Irvine, CA) in PTB was added to filters for 3 h at 40C, and filters were washed as before. The filters were exposed to x-ray film for 14 h to visualize irnmunoreactive proteins.