Hum. Reprod.

Advance Access published January 25, 2013

Human Reproduction, Vol.0, No.0 pp. 1 14, 2013 doi:10.1093/humrep/des452

ORIGINAL ARTICLE Reproductive biology

Identication of sperm head proteins involved in zona pellucida binding

F.M. Petit 1, C. Serres2, F. Bourgeon 3, C. Pineau 3, and J. Auer 2,*
ne tique mole culaire, Ho pital Antoine Be cle ` re, Clamart 92141, France 2INSERM U1016, De partement de AP-HP, Laboratoire de ge ne tique et De veloppement, Institut Cochin, CNRS UMR8104 and Universite Paris Descartes, Paris 75014, France 3Proteomics Core Ge Facility Biogenouest, Inserm U1085 IRSET, Campus de Beaulieu, Rennes Cedex 35042, France *Correspondence address. E-mail: jana.auer@parisdescartes.fr
1

Downloaded from http://humrep.oxfordjournals.org/ at Health Sciences Library on February 12, 2013

Submitted on September 28, 2012; resubmitted on November 22, 2012; accepted on December 11, 2012

study question: Which human sperm proteins interact with zona pellucida (ZP) glycoproteins, ZPA/2, ZPB/4 and ZPC/3? summary answer: Co-precipitation experiments with recombinant human ZP (rhZP) coated beads demonstrated interactions with
various proteins, including glutathione S-transferase M3 (GSTM) with ZPB/4 and voltage-dependent anion channel 2 (VDAC2) with ZPA/2 and ZPC/3.

what is known already: Regarding sperm ZP binding, several target spot/proteins have been detected in several species, but
not all have been characterized. The limit of these studies was that a mixture of the different ZP glycoproteins was used and did not allow the identication of the specic ZP glycoprotein (ZPA/2, ZPC/3 or ZPB/4) involved in the interaction with the sperm proteins.

study design, size, duration: To identify the human sperm proteins interacting with the oocyte ZP, we combined two
approaches: immunoblot of human spermatozoa targeted by antisperm antibodies (ASAs) from infertile men and far western blot of human sperm proteins overlayd by each of the rhZP proteins.

materials, setting, methods: We used rhZP expressed in Chinese hamster ovary (CHO) cells and ASA eluted from infertile patients undergoing IVF failure. Sperm proteins separated by two-dimensional (2D) electrophoresis recognized by both sperm-eluted ASAs from infertile patients and rhZP were identied by mass spectrometry (MALDI-MS/MS). Some of these proteins were further validated by co-precipitation experiments with rhZP and functional zona binding tests. main results and the role of chance: We identied proteins that are glycolytic enzymes such as pyruvate kinase 3,
enolase 1, glyceraldehyde-3-phosphate dehydrogenase, aldolase A, triosephosphate isomerase, detoxication enzymes such as GSTM or phospholipid hydroperoxide glutathione peroxidase, ion channels such as VDAC2 and structural proteins such as outer dense bre 2. Several of the proteins were localized on the sperm head. However, these proteins have also been described to exert other functions in the agellum. Co-precipitation experiments with rhZP-coated beads conrmed the direct interaction of GSTM with ZP4 and of VDAC2 with ZP2 and ZP3.

limitations, reasons for caution: We used recombinant ZP in place of native ZP. Thus, the post-translational modications of the proteins, such as glycosylations, can be different and can inuence their function. However, CHO cell-expressed rhZP are functional, e.g. can bind human spermatozoa and induce the acrosome reaction. Moreover, the identication of relevant proteins was limited by the need for sufcient amounts of proteins on the preparative 2D-gel to be subsequently analysed in MALDI-TOF MS/MS. wider implications of the ndings: Our results bring new insights on the ability of sperm proteins to exert several functions depending on their sub-cellular localization, either the head or agellum. Their multiple roles suggest that these sperm proteins are multifaceted or moonlighting proteins. study funding/competing interest(s): This work was supported by the grant ReproRio (CNRS, INRA, INSERM and
te dAndrologie de Langue Franc CEA) and the Socie aise.

trial registration number: Not applicable.

Key words: sperm zona pellucida interaction / antisperm antibodies / far WB / moonlighting / proteome

& The Author 2013. Published by Oxford University Press on behalf of the European Society of Human Reproduction and Embryology. All rights reserved. For Permissions, please email: journals.permissions@oup.com

Petit et al.

Introduction
Mammalian fertilization is a complex process involving several molecules distributed on the two gametes, spermatozoon and oocyte. One of these steps is the binding to, and penetration of, the extracellular coat of the oocytes termed the zona pellucida (ZP) by spermatozoa. Immediately after ejaculation, the sperm are not able to recognize and interact with the ZP. The ability of sperm to bind the ZP is achieved only after the last maturation event, called capacitation, which occurs in the female genital tract. This species-specic event stimulates diverse signalling pathways in the spermatozoa leading to the acrosomal exocytosis. During this acrosome reaction, the sperm plasma membrane fuses with the outer acrosomal membrane, causing exposure of the inner acrosome membrane at the surface and release of the acrosomal content. A widely used strategy to identify sperm proteins involved in gamete interaction is based on the inhibition of sperm oocyte binding/fusion by antispermatozoa antibody produced after a mouse/rabbit immunization. This approach requires a time-consuming screening to nd inhibitory antibodies and then the corresponding antigens. Several sperm proteins involved in gamete interaction have been identied by this procedure, such as acrosin, fertilin, Izumo, SPAM1 and ADAM (Topfer-Petersen et al., 1990; McLeskey et al., 1998; Nixon et al., 2007). After validation steps, which often involve the creation of mouse models, usually loss-of-function mouse models for the gene of interest, the protein is either conrmed or ruled out for its role in gamete interaction. Some groups have studied the immunoproteome of human gametes revealed by serum or seminal plasma antisperm antibodies (ASAs) (Shetty et al., 2001, 2008; Bohring and Krause, 2003; Domagala et al., 2007). Stein et al. (2006) combined a proteomic study with subcellular fractionation in order to identify sperm head proteins that mediate the sperm oocyte interaction. Our group has detected several proteins targeted by local ASAs (and not systemic ASAs as generally used) from patients with autoimmune infertility (Auer et al., 1997) and identied and characterized a triosephosphate isomerase (TPI) involved in sperm ZP interaction (Auer et al., 2000, 2004). Another method, also called blot-overlay or far western blot (WB), involves the separation of sperm proteins on one- or two-dimensional (2D) electrophoresis gels, transfer to polyvinylidene diuoride (PVDF) membranes and overlay of the sperm proteins by solubilized ZP glycoproteins. Several target spot/proteins have been detected with this strategy, but not all have been characterized [Shabanowitz and ORand, 1988 (human); Tanii et al., 2001 (mouse); ManaskovaPostlerova et al., 2011 (boar)]. The limit of these studies was that a mixture of the different ZP glycoproteins was used and did not allow the identication of the specic ZP glycoprotein (ZPA/2, ZPC/3 or ZPB/4) involved in the interaction with the sperm proteins. In humans, the difculty in obtaining ZP material in adequate quantity and quality made the above approach less realistic but, in the last 10 years, such a problem has been overcome by the use of human recombinant ZP obtained from diverse cells (Harris et al., 1999; Martic et al., 2004; Chakravarty et al., 2005; Marin-Briggiler et al., 2008; Chirinos et al., 2011). Notably, this approach allows the interactions with the different glycoproteins (ZPA/2, ZPC/3 or ZPB/4) to be distinguished.

Here, we made use of the relatively recently developed proteomic tools and of recombinant human ZP (rhZP) glycoproteins and studied the human sperm receptors for ZP2, ZP3 and ZP4 by direct interaction between rhZP2, rhZP3 or rhZP4 glycoproteins and solubilized sperm membrane proteins using the far WB technique. We compared the results obtained by this approach with those obtained when sperm proteins separated by 2D electrophoresis were recognized by sperm-eluted ASA from infertile patients. The proteins recognized by both ZP glycoproteins and ASA were then identied by mass spectrometry. Some of these proteins were further validated by co-precipitation experiments and functional zona binding tests. With this study, we identied a set of sperm proteins involved in sperm ZP interaction. Several of them are involved in functions other than ZP interaction, which highlights the moonlighting functions of these sperm proteins.

Downloaded from http://humrep.oxfordjournals.org/ at Health Sciences Library on February 12, 2013

Materials and Methods

rhZP glycoproteins and specic antibodies
rhZP produced in Chinese hamster ovary (CHO) cells and their specic anti-sera obtained by rabbit immunization were gifts of Harris et al. (1999). These CHO cells-expressed rhZPs are secreted in the medium and are highly puried.

Sperm membrane fraction

Sperm samples with normal semen parameters (WHO, 2010) were obtained from fertile donors. The motile spermatozoa were selected on Percoll gradients as previously described (McClure et al., 1989). After overnight capacitation in B2 medium (CCD, Paris, France), spermatozoa were washed in 0.05 mmol/l Tris buffer (TB). A pool of capacitated spermatozoa from several donors was constituted for co-precipitation, WB or far WB assays, while individual samples were used for functional tests. For co-precipitation experiments, 1 108 washed spermatozoa were solubilized with 1% NP-40 or 0.1% Triton X-100 detergent in 100 ml of TB supplemented with a protease inhibitors cocktail (Sigma-Aldrich, St Quentin Fallavier, France) for 1 h at 48C. The supernatant was stored at 2 808C until use. For electrophoretic separation, spermatozoa were solubilized either in Laemmli-reducing sample buffer [SDS-polyacrylamide gel electrophoresis (PAGE)] at 958C or in 9 mol/l urea, 2% Triton X-100, 60 mmol/l dithiothreitol, 2% immobilized pH gradient (IPG) buffer (2D) and a protease inhibitors cocktail at 48C. Supernatants were stored at 2 808C until use.

Antisperm antibodies
Sperm samples with high levels of ASA as detected by Immunobeads test (IBT, Sigma-Aldrich) were obtained from infertile men undergoing several unsuccessful in vitro fertilization attempts. Sperm samples from three fertile men with no ASA detectable by IBT were used as negative controls. To obtain ASAs, sperm samples were centrifuged at 600g for 10 min and then sperm pellets were washed twice with phosphate-buffered saline (PBS). The washed pellets containing 5 107 to 1 108 motile spermatozoa were resuspended in 1 2 ml of 100 mmol/l glycine-HCl buffer at a pH of 2.8 under gentle rotation for 15 min at room temperature (RT) and centrifuged for 5 min at 12 000g. The supernatants were neutralized with 3 mol/l Tris, dialysed against PBS overnight at + 48C and ltered through 0.2 mm sterile Acrodisc lters (Gelman Sciences,

Moonlighting sperm ZP binding proteins

3
After the far WB assay as well as after immunodetection, 2D-PVDF membranes were silver stained (Kovarik et al., 1987) for the precise localization of the reactive spots.

France). The titre, immunoglobulin (Ig) class and localization of ASAs were determined by an indirect Immunobeads test (Clarke et al., 1984). All experiments with human samples were conducted in accordance with ethical guidelines and were approved by the ethics evaluation de qualication institutionnelle) of the Institut National committee (comite et de la Recherche Me dicale, INSERM (authorization no. de la Sante 01-013).

In-gel trypsin digestion

Protein spots were excised from 2D gels, and further processed for mass spectrometry thanks to an EttanTM Spot Handling Workstation (GE Healthcare). Briey, before drying, the gel plugs were washed three times in MilliQ water, once in 50% ethanol/50 mmol/l ammonium bicarbonate and once in 75% acetonitrile. Dried plugs were then incubated for 60 min in 20 mmol/l NH4HCO3 supplemented with 8.3 mg/ml se` resquencing grade modied porcine trypsin (Promega, Charbonnie les-Bains, France). Digested peptides were extracted in two successive steps by incubation of gel plugs in 70% acetonitrile and 0.1% triuoroacetic acid. Digested peptides were then dissolved in 0.6 mg/ml a-cyano4-hydroxycinnamic acid in 55% ethanol/27% acetone/0.1% triuoroacetic acid, and further spotted onto a MTP AnchorchipTM MALDI target (384 Scout MTP 600 mm Anchorchip; Bruker Daltonik, GmbH, Bremen, Germany).

Electrophoresis separation
Sperm membrane proteins were further fractionated using 2D gel electrophoresis. Briey isoelectrofocalization was performed on linear immobilized pH 3 10 gradient gel strips of 13 cm, using the Multiphor II system (GE Healthcare, Saclay, France). Strips were cup-loaded at the anodic end with 70 mg of sperm proteins (25 30 106 spz) for analytical gels and isoelectric focalization was performed at 208C for a total of 18.5 kVh. The second dimension was performed in SDS-PAGE on 12% polyacrylamide gels. For preparative gels, the 24 cm 3 10 linear IPG strips were cup-loaded with 90 mg of sperm proteins and focalization was conducted in IPGphor (GE Healthcare) for a total of 71.4 kVh. The second dimension was performed on 12.5% polyacrylamide gels using the DALTSix system (GE Healthcare). At each step, the protein concentration was determined using the bicinchoninic acid protein assay (Sigma-Aldrich). Following migration, 2D gels were silver stained, as previously described (Shevchenko et al., 1996) with minor modications (Com et al., 2003). Gels were scanned with an ImageScanner (GE Healthcare) and then stored at 48C in 1% acetic acid until spot excision. Finally, a pick list was generated using ImageMaster 2D Elite software (GE Healthcare). SDS-PAGE was also performed after co-precipitation of the sperm proteins with rhZP. The protein complexes were separated using a mini PROTEAN II Cell apparatus (Bio-Rad, Marne la Coquette, France) on 12% polyacrylamide gels.

Downloaded from http://humrep.oxfordjournals.org/ at Health Sciences Library on February 12, 2013

Mass spectrometry analysis

Protein identication by mass spectrometry was performed using a MALDI-TOF/TOF mass spectrometer (Ultraex; Bruker Daltonik). Peak lists were generated from MALDI-MS spectra using the FlexAnalysis software (version 3.0; Bruker Daltonik). Following internal calibration with trypsin autodigestion peptides, the monoisotopic masses of tryptic peptides were used to query the NCBInr sequence database (version 20092604, 6833826 sequences), using Mascot server version 2.2 (www. matrixscience.com). Search conditions used were an initial open mass window of 70 ppm for an internal calibration and one missed cleavage allowed. Carbamidomethylation of cysteines was set as xed modications whereas methionine oxidation was set as variable modications. Peptide identications were scored using the probability-based Mowse score (the protein score is 2 108log (P) where P is the probability that the observed match is a random event). In the present experimental conditions, a score . 78 corresponded to a signicant identication (P , 0.05).

Overlay assay (or far WB)

Sperm proteins separated in 1D or 2D electrophoresis were electrotransferred onto PVDF membranes (Hybond-P, GE Healthcare) in renaturing conditions (Dunn, 1986). Prior to incubation, an rhZP solution was preabsorbed with a blank piece of nitrocellulose (Shabanowitz and ORand, 1988). To avoid any non-specic binding, anti-rhZP sera were preabsorbed on human spermatozoa. After incubation in blocking buffer [1% gelatine in TB salin (TBS)] blots were incubated with rhZP (0.75 mg for 1 106 of spz) in TBS modied (TBSm) with 0.05% Tween20, 1 mmol/l Ca2+ and 1 mmol/l Mg2+. Glycoproteins that interacted with sperm proteins were detected with specic anti-rhZP sera (1:4000 in TBSm). After incubation with peroxidaseconjugated anti-rabbit IgG, the binding was detected with the enhanced chemiluminescence (ECL) + detection western blotting system (GE Healthcare).

Immunouorescence staining
For immunostaining on live spermatozoa, a capacitated sperm suspension (5 106 cells/ml) was incubated for 30 min in PBS-5% bovine serum albumin (BSA) to block non-specic staining sites, then with primary antibodies at 1:50 for 1 h at RT (anti-GST, Uptima, Interchim, France; anti-ALDOA, Abnova, Interchim, France) or overnight at 48C [anti-voltagedependent anion channel 2 (VDAC2), Proteintech, Manchster, UK]. After washing, spermatozoa were incubated with the corresponding uorescein isothiocyanate (FITC)-conjugated secondary antibodies or biotinylated secondary antibodies and FITC-conjugated streptavidine (for VDAC2). An immunostaining procedure was also conducted on xed cells. For this, capacitated spermatozoa were incubated for 1 h in 1% paraformaldehyde (PFA) in PBS. To neutralize free reactive aldehyde groups, the cells were then incubated for 30 min in 200 mmol/l glycine in PBS. After washing, spermatozoa were resuspended at a density of 5 106 cells/ ml in PBS-1% BSA and smeared on slides and air dried. They were rst incubated for 30 min in PBS-5% BSA to block non-specic staining sites and then with primary antibodies at a 1:50 for 1 h at RT. The staining with corresponding uorescent secondary antibodies or biotinylated secondary antibodies and FITC-conjugated streptavidine (for VDAC2) was done at RT. An addition of 0.005% saponin to PBS/BSA before and during the staining procedure was needed to facilitate the access of anti-ALDOA and anti-VDAC2 to their sperm targets.

Immunodetection with ASA

Sperm proteins were electrotransferred as described above. The membranes were saturated for 1 h at RT in PBS with 5% low-fat milk powder, and then incubated for 1 h at 378C and overnight at 48C with ASAs obtained from spermatozoa of infertile men complemented with 0.1% Tween20 (PBST) and 1% low-fat milk powder. After washing, the blots were incubated for 1 h at RT with afnity-puried goat anti-human ` gne, France) Ig antibodies conjugated to peroxidase (Biosys, Compie diluted to 1:8000 in PBST with 1% gelatine. The blots were washed twice in PBST and once in PBS. Bound peroxidase was detected by an ECL + western blotting system (GE Healthcare).

4
Rabbit or mouse IgG in place of primary antibodies was included as negative controls. After immunostaining procedures, stained spermatozoa were mounted in glycerol-PBS (Citiuor, London, UK) for observation. The uorescence was examined, using an epiuorescence microscope (Nikon E600, Champigny sur Marne, France) at 630 and 1000 magnications.

Petit et al.

Statistical tests
Students t test and x 2 test were employed as statistical tests and P , 0.05 was considered as signicant.

Results
Detection of sperm proteins interacting with ZP glycoproteins by the far WB method
The sperm proteins targeted by ZP glycoproteins were detected in a far WB assay. This analysis revealed that each ZP glycoprotein interacts with several sperm proteins. About 15 spots or groups of spots were targeted with rhZP2. The molecular weights of these sperm proteins ranged from 18 to 90 kDa. The majority of them were basic with an isoelectric point (pI) between 5.0 and 8.5 (Fig. 1A). Among the basic proteins recognized by rhZP4, three groups of spots had a high intensity. Their molecular weights ranged from 15 to 90 kDa and their pI ranged from 5.5 to 8.0 (Fig. 1B). The signals of the acidic spots with rhZP4 were less intense. Spots targeted with rhZP3 were more numerous, and their pI ranged from 5.2 to 8.2. Their molecular weights ranged from 15 to 75 kDa (Fig. 1C). Control blots which were incubated without rhZP or with the rhZP but without anti-rhZP antibody did not show any spots. For further analysis, we focused on the proteins that were recognized by rhZP in at least two far WB experiments out of the three performed.

Co-precipitation
To obtain beads coated with rhZP, Carboxyl-Adembeads (Ademtech, France) were rstly activated according to the manufacturers instructions. Then, 60 75 mg of rhZP2, rhZP4 and rhZP3 proteins were incubated with 900 mg of magnetic activated beads for 2 h at 408C under agitation. After saturation of the binding sites by 40 mmol/l ethanolamine and three washing steps, 100 ml of sperm membrane protein extract (corresponding to 1 108 cells) was added to the rhZP-coated beads in the presence of a protease inhibitor cocktail. After 2 h of incubation at RT under agitation, the beads were washed three times in 500 ml of lysis buffer, then resuspended in Laemmli reducing buffer and boiled at 958C. Supernatants were then subjected to electrophoresis and WB analysis.

Downloaded from http://humrep.oxfordjournals.org/ at Health Sciences Library on February 12, 2013

Functional tests
Zona binding test
To ensure that the antibody effect on a binding test is not related to their effect on motility, we analysed sperm motility after incubation with antibodies using a Computer-Assisted Sperm movement Analyser (CASA, Hamilton Thorn, USA). We used zona intact unfertilized oocytes recovered after IVF failure. Spermatozoa capacitated overnight (5 105) were added to 500 ml of B2 medium containing zona intact unfertilized oocytes and incubated for 3 h (when primary binding was examined) or 18 h (when zona penetration was examined) at 378C in a 5% CO2/95% air atmosphere in the presence of antibodies. After 3 h of incubation, the oocytes were washed in B2 medium by three to four successive aspirations through a glass pipette (inner diameter of 250 mm) to remove loosely bound spermatozoa. A rst aliquot of washed oocytes was then deposited in a glass depression slide in order to count the spermatozoa bound to each oocyte. The remaining washed oocytes were repeatedly aspirated through a pipette having an inner diameter slightly smaller than the oocytes (125 mm) to detach spermatozoa tightly bound to the ZP. Then, the acrosomal status of the latter spermatozoa was determined. To determine the effect of antibodies on the zona penetration, the oocytes were incubated with spermatozoa for 18 h in the presence of antibodies. After removing adhering spermatozoa from the oocytes, their acrosome reaction rate was determined as described below and oocytes were placed on a slide in a 20 ml drop of 0.2% BSA-PBS to examine the number of spermatozoa that remained rmly bound to or embedded within the ZP or present in the perivitelline space.

Detection of sperm antigens targeted by ASA eluted from sperm of infertile patients
In this study aimed at identifying sperm proteins involved in ZP recognition/binding, we used ASAs directly eluted from spermatozoa provided by infertile patients who underwent IVF failures. In the majority of the cases, a defect of binding and/or penetration of the ZP had been noticed. All the samples of ASAs isolated from spermatozoa of infertile men contained IgA and IgG antibodies. The percentages of the sperm binding immunobeads were 4087% for IgA and 10 95% for IgG and at least 71% of spermatozoa bound to beads in each sample. We analysed the sperm antigens recognized by ASA isolated from the spermatozoa of six individual infertile men and from the pool of ve patients constituted according to their class of ASA (predominantly IgA or IgG). As reported in our earlier study (Auer et al., 1997), we found heterogeneity in the response to ASA targets among individual ASA samples (Fig. 2), as well as among Ig classes. Only a small proportion of sperm proteins resolved in the 2D gels appeared to be targets of ASAs. Thus, 5 30 protein spots were recognized by ASAs with a group of basic proteins of 50 kDa, an area of proteins at around 60 kDa with a rather neutral pI, and an area of more acidic proteins (pI , 5) of 80 kDa. We also noted that the pool of eluted ASA containing IgG antibodies recognized more antigens than those containing predominantly IgA antibodies. No spots appeared in WBs targeted with ASA samples from fertile men.

Acrosome reaction measurement

The acrosomal status of spermatozoa was examined using the modied method of Cross et al. (1986). The suspension of spermatozoa was deposited on slides, air dried and xed in ethanol for 30 min at 48C. Cells were then stained by tetramethyl-rhodamine isothicyanate (TRITC) or FITCPisum sativum agglutinin (PSA) (25 mg/ml in PBS) for 15 min. After washing in distilled water, the slide was mounted with Citiuor and 200 spermatozoa were examined using an epiuorescence microscope. Spermatozoa displaying a uorescent equatorial segment or with no staining of the head were recorded as acrosome-reacted. As only motile sperm were able to bind to the ZP, no viability staining was performed.

Identication of sperm proteins targeted by both eluted ASA and ZP glycoproteins

We limited our identication work to the spots targeted by both eluted ASAs and rhZP. For a precise localization of the targeted

Moonlighting sperm ZP binding proteins

5
ImageMaster 2D Platinum software able to take into account the differences between the gels. Only spots that were well visible on the preparative silver-stained gel were excised, digested and analysed by mass spectrometry. Among the spots selected for mass spectrometry analysis, 14 were identied with a high score and high coverage (between 22 and 75%) and these corresponded to 9 different proteins (Table I). Five glycolytic enzymes represented the most important group [pyruvate kinase 3 (PK3), enolase 1 (ENO1), glyceraldehyde-3phosphate dehydrogenase (GAPDH), aldolase A (ALDOA) and TPI]; two proteins were known to be involved in detoxication processes, glutathione S-transferase M3 (GSTM) and phospholipid hydroperoxide glutathione peroxidase (PHGPx); one protein was an ionic channel, VDAC2 and another protein, outer dense bre 2 (ODF2), was a cytoskeletal protein.
Downloaded from http://humrep.oxfordjournals.org/ at Health Sciences Library on February 12, 2013

Co-precipitation of sperm proteins with recombinant ZP glycoproteins

To further assess the ability of sperm proteins to interact specically with rhZP, we carried out co-precipitation experiments using sperm extract incubated with rhZP-coated magnetic beads. We focused this analysis on ALDOA, TPI, GSTM and VDAC2, four of the sperm proteins shown to interact with one or two ZP glycoproteins in the far WB/ASA assay and for which an antibody was commercially available. After co-incubation of sperm extract with rhZP-coated beads, we precipitated TPI, GSTM and VDAC2 with rhZP-coated beads. The characteristic 36 kDa band of sperm TPI (Auer et al., 2004) was detected in 1% NP40 extract and found to interact with rhZP3 and rhZP4 to a larger extent than with rhZP2 (data not shown). GSTM was detected as a 26 kDa band in 0.1% TX100 as well as in 1% NP40 sperm extract and preferentially co-precipitated with rhZP4 when compared with rhZP2 and rhZP3 (Fig. 4A). VDAC2 which runs as a 3334 kDa protein in 1% NP40 sperm extract was precipitated with rhZP2 or rhZP3 but not with rhZP4-coated beads (Fig. 5A). Similar results were obtained for these three proteins in two independent experiments. For ALDOA, a 44 kDa band (consistent with its theoretical molecular weight) was detected in 1% NP40 sperm extracts but not after precipitation with rhZP2, rhZP4 or rhZP3 (Fig. 6A).

Localization of sperm proteins interacting with ZP

Figure 1 Sperm proteins targeted by rhZP in far western blots (WB) assays. 2D WB of human sperm proteins was probed with solubilized rhZP2 (A), rhZP4 (B) and rhZP3 (C) and revealed with corresponding anti-rhZP2, anti-rhZP4 or anti-rhZP3 antibodies as described in the Materials and Methods section. Molecular identities of some proteins identied by mass spectrometry are indicated on the blots. The spots with low intensity are shown on the right of the blots with increased contrast. The pI gradient (310) used for IEF is reported at the top of each blot and the molecular mass (Mr) standards in kDa used for the second dimension are indicated on the left of each blot. In human spermatozoa, double staining of capacitated spermatozoa with anti-GSTM and PSA lectin revealed GSTM immunoreactive sites in the head region overlying the acrosome of intact (PSA positive) spermatozoa (Fig. 4B a and a ). The GSTM staining was not observed on acrosome-reacted (PSA negative) spermatozoa. No GSTM staining was observed in the control incubation with rabbit IgG instead of anti-GSTM antibody (Fig. 4B b), nor in sperm treated with 0.05% Triton X100. On PFA-xed sperm cells, VDAC2 was detected in the region overlying the acrosome of intact cells and at the agellum level (Fig. 5B a and c). We failed to detect any staining with anti-VDAC2 antibody on live spermatozoa. However, anti-VDAC2 immunodetection performed on spermatozoa kept in suspension and treated with saponin worked. Under such conditions, VDAC2 staining was observed in the post-

spots on 2D gels, exposed lm and corresponding silver-stained blots were scanned and overlapped. The resulting spots were then positioned on a scanned preparative gel (Fig. 3) and analysed with the

Petit et al.

Downloaded from http://humrep.oxfordjournals.org/ at Health Sciences Library on February 12, 2013

Figure 2 Sperm proteins targeted by antisperm auto-antibodies. 2D WBs of human sperm proteins revealed by ASAs eluted from spermatozoa of three infertile men (A C) illustrate the heterogeneity of the immune response. Molecular identities of some antigens identied by mass spectrometry are indicated on the blots. The pI gradient (3 10) used for IEF is reported at the top of each blot and the molecular mass (Mr) standards in kDa used for the second dimension are indicated on the left of each blot.

Figure 3 Preparative silver-stained 2D gel of human sperm proteins. Preparative 2D gel of human sperm proteins solubilized and separated in 24 cm 3 10 linear immobilized pH gradient (IPG) strips and silver stained. Spots revealed in far WBs and on ASA immunoblots were compared and then localized on this gel using the Image Master software. They were excised and processed for identication by MALDI-TOF-TOF analysis. The positions of these spots are indicated by crosses and the names are noted. The pI gradient used for IEF is reported at the top of gel and the molecular mass (Mr) standards in kDa used for the second dimension are indicated on the left of the gel.

acrosomal region of the majority of the cells, either as a double uorescent band or as a calyx at the head base (Fig. 5B e). No staining was visible in control experiments (Fig. 5B b, d and f).

No anti-ALDOA staining could be detected in live sperm cells. An additional treatment with 0.005% saponin was needed to reveal an immunostaining in the sperm head. This latter was conned to the

Moonlighting sperm ZP binding proteins

Table I ZP and ASA binding human sperm proteins identied by Mass ngerprinting.
Function Protein name Species Accession numbera Mr expb pI exp Matched peptides/ totalc 10/24 22/51 5/9 14/29 10/25 7/20 10/35 10/28 16/42 % coverage Mascot score Target

............................................................................................................................................................................................
Glycolytic enzyme PK3 isoform 2 ENO1 GAPDH ALDOA TPI 1 Detoxication GSTM PHGPx 4 Ion transport Cytoskeleton VDAC ODF of sperm tails Homo sapiens Homo sapiens Homo sapiens Homo sapiens Homo sapiens Homo sapiens Homo sapiens Homo sapiens Homo sapiens NP_872270 AAH04458 NP_055179 NP_000025 AAH17917 NP_000840 P36969 AAB59457 NP_702915 66 47 47 42 33 27 17 34 47 7.3 6.8 7.4 8.6 5.2 4.9 8.5 6.7 5.5 24 53 19 59 51 38 41 47 22 102 200 133 189 142 83 74 126 100 ZP2-ZP3-ZP4 ZP2-ZP3-ZP4 ZP2-ZP3-ZP4 ZP4 ZP3-ZP4 ZP3-ZP4 ZP2-ZP3-ZP4 ZP2-ZP3 ZP3-ZP4 ASA ASA ASA ASA ASA ASA ASA ASA ASA

Downloaded from http://humrep.oxfordjournals.org/ at Health Sciences Library on February 12, 2013

PI, Isoelectric point. a National Center for Biotechnology Information (NCBI) sequence identication number. b Experimental relative mass in kDa. c The number of matched peptides versus the total number of peptides.

Figure 4 Validation of GSTM as sperm ZP binding protein. (A) Sperm proteins extracted in 1% NP40 were precipitated using magnetic beads
coated with rhZP glycoproteins. The WB of sperm proteins co-precipitated with rhZP2 (A), rhZP4 (B) or rhZP3 (C) and of sperm extract (E) were revealed with an anti-GSTM antibody. (B) Localization of GSTM on human spermatozoa. Double staining of capacitated spermatozoa using an anti-GSTM and a secondary antibody conjugated to FITC (a) followed by FITC-Pisum sativum agglutinin (PSA) lectin conjugated to tetramethylrhodamine isothicyanate (TRITC) to reveal the acrosomal content (a ). GSTM staining is visible in the acrosomal region of intact spermatozoa (PSA positive). The arrows indicate two acrosome-reacted spermatozoa (PSA negative) without GST labelling. Rabbit IgG was used in place of anti GSTM antibody in the control staining experiment (b).

equatorial band and at the base of the head of acrosome intact spermatozoa (Fig. 6B a and c). Anti-ALDOA antibody also stained the agellum. Of note, ethanol or 0.05% Triton X100 treatments eliminated

the sperm head staining but preserved the immunoreactivity of the agellar principal piece (Fig. 6B d). Control staining using mouse IgG in place of anti-ALDOA is shown in Fig. 6B b.

Petit et al.

Downloaded from http://humrep.oxfordjournals.org/ at Health Sciences Library on February 12, 2013

Figure 5 Validation of VDAC2 as sperm ZP binding protein. (A) Sperm proteins extracted in 1% NP40 were co-precipitated using magnetic beads coated with rhZP glycoproteins. The WB of sperm extract (E) and sperm proteins co-precipitated with rhZP2 (A), rhZP4 (B) or rhZP3 (C) were revealed with anti-VDAC2 antibody. (B) VDAC2 localization on PFA-xed (a d) or live (e and f) human spermatozoa permeabilized by 0.005% saponin. Spermatozoa were PFA-xed before (a and b) or after (c and d) spreading on slides. Spermatozoa were labelled with anti-VDAC2 and FITC secondary staining (a, c and e) and TRITC-PSA lectin (a ). The arrow indicates the VDAC2 staining over the acrosomal area of intact spermatozoa (PSA positive). Rabbit IgG was used in place of anti-VDAC2 in the control staining experiments (b, d and f).

Functional tests
We next tested the role of ALDOA and VDAC2 in sperm binding to zona intact oocytes. We thus performed gamete interaction tests in the presence of anti-ALDOA or anti-VDAC2 antibody (Tables II V, respectively). In four independent experiments, the number of sperm bound to the ZP was signicantly decreased in the presence of anti-ALDOA when compared with incubation with control IgG (Table II). During the binding tests, we veried that anti-ALDOA did not immobilize the spermatozoa. In one experiment measuring sperm motility by CASA, a slight decrease in the percentage of motile spermatozoa was noted with anti-ALDOA (62%) when

compared with IgG (81%). We examined the acrosomal status of those spermatozoa bound to the ZP after 3 h of interaction, in the presence of anti-ALDOA or control IgG. In three separate experiments, no difference in the rate of acrosome reaction induced by ZP was observed between the two groups (data not shown). Incubation with anti-VDAC2 antibody during the in vitro gamete interaction led to a reduction by half of the number of spermatozoa bound to or penetrating the ZP (Tables III and IV). Exposure to antiVDAC2 during the binding tests did not signicantly alter the sperm motility. Indeed, the percentage of motile spermatozoa measured by CASA after 3 h of contact with anti-VDAC2 was 60.0 + 6.24%

Moonlighting sperm ZP binding proteins

Downloaded from http://humrep.oxfordjournals.org/ at Health Sciences Library on February 12, 2013

Figure 6 Aldolase localization on human spermatozoa. (A) Sperm proteins extracted in 1% NP40 were co-precipitated using magnetic beads coated with rhZP glycoproteins. The WB of sperm proteins co-precipitated with rhZP2 (A), rhZP4 (B) or rhZP3 (C) and of sperm extract (E) were revealed with anti-ALDOA antibody. (B) Live spermatozoa were permeabilized by 0.005% saponin (a c) or ethanol (d) treatment before immunostaining. They were stained with anti-ALDOA and FITC-conjugated secondary antibody (a and d) or double stained with anti-ALDOA and FITC secondary staining followed by TRITC-PSA to label the acrosomal content (c). Mouse IgG replaced anti-ALDOA in the control experiment (b).

Table II Sperm binding to zona-intact unfertilized human oocytes in the presence of anti-ALDOA antibody.
Experiment 1 2 3 4 Control 98.20 + 24.54 (5)a 62.71 + 13.49 (7) 17.71 + 1.78 (7) 97.00 + 15.31 (8) Anti-ALDOA 44.50 + 21.38 (4)* 20.62 + 3.82 (13)*** 9.50 + 4.04 (8)* 48.17 + 12.47 (6)**

Table III Sperm binding onto ZP of intact unfertilized human oocytes in the presence of anti-VDCA2 antibody.
Experiment 1 2 3 Control 30.00 + 6.73 (8)a 72.71 + 12.04 (7) 32.14 + 4.06 (14) Anti-VDAC2 16.43 + 4.84 (7) 40.50 + 10.99 (8) 13.00 + 4.86 (9)*

........................................................................................

........................................................................................

Results are expressed as the number of spermatozoa per oocyte (mean + SEM). a The number of oocytes is given in parentheses. Signicantly different from control oocytes (Students t test, *P , 0.05, **P , 0.02 and ***P , 0.0001).

Results are expressed as the number of spermatozoa per oocyte (mean + SEM). a The number of oocytes is given in parentheses. *Signicantly different from control oocytes (Students t test, P , 0.01).

Discussion
New strategy for identication of ZP receptor proteins
In the present study, we aimed to identify human sperm proteins that interact with ZP glycoproteins using a double approach: a far WB

compared with 70.0 + 8.54% in the presence of IgG control (P 0.09; n 3). The induction of the acrosome reaction in the spermatozoa bound to ZP was unaffected by the presence of anti-VDAC2 (Table V).

10
assay testing for direct interaction between sperm proteins and ZP glycoproteins and an indirect approach using ASAs directly eluted from spermatozoa of infertile patients with failure of conventional IVF. With this strategy, we increased the probability of identifying sperm proteins actually involved in gamete interaction. These two approaches have already been independently employed to identify sperm proteins presumed to play a role in gamete interaction. Indeed, several studies have been carried out to identify sperm antigens that react with ASAs, which are responsible for 215% of IVF failure. They therefore constitute an appropriate tool for such a research. In the most recent studies using 2D gel electrophoresis coupled to MALDI-TOF-MS/MS, diverse proteins such as HSP70, Caspase3, LDHC4, VDAC2, ODF1 and GSTMu5 have been identied as the targets of serum or seminal plasma ASAs (Shetty et al., 1999; Bohring et al., 2001; Bohring and Krause, 2003; Chiu et al., 2004; Paradowska et al., 2006; Domagala et al., 2007). Although the humoral response has been described to interfere with fertility, local ASAs would be a better tool for identifying sperm membrane antigens actually involved in IVF failure. For this reason, we rather used ASAs bound to and eluted from spermatozoa of infertile men to identify relevant sperm-specic antigens. We conrmed the previously observed heterogeneity of the immune response of each infertile individual in intensity and in the number and range of protein spots that were revealed. Despite this heterogeneity, we noted that some antigens were common to different individuals distributed over three to four areas of proteins recognized by a majority of samples. Only a few studies report the use of a far WB assay, a method routinely used to identify interacting proteins (Edmondson and Roth, 2001; Machida and Mayer, 2009), with the aim of characterizing human sperm receptors for the ZP. Using radio-labelled solubilized

Petit et al.

Table IV Sperm penetration of ZP of intact unfertilized human oocytes in the presence of anti-VDCA2 antibody.
Experiment 1 2 3 Control 10.30 + 2.47 (10)a 35.18 + 10.03 (11) 6.17 + 2.19 (12) Anti-VDAC2 5.33 + 1.84 (9) 27.18 + 5.13 (11) 1.79 + 0.87 (14)*

........................................................................................

Results are expressed as the number of spermatozoa per oocyte (mean + SEM). a The number of oocytes is given in parentheses. *Signicantly different from control oocytes (Students t test, P , 0.01).

ZP, ORand et al. (1985) and Shabanowitz and ORand (1988) detected several sperm receptors at 16, 18, 19, 35 and 60 kDa without further characterization. More recently in the porcine species, multiple sperm plasma membrane proteins interacting with native ZP fragments were identied, including spermadhesin AQN-3, SED-1 (also known as P47), fertilin-beta and peroxiredoxin (van Gestel et al., 2007). In humans, fucosyltransferase has been proposed to be a sperm receptor for the intact and solubilized ZP (Chiu et al., 2007). In the present study, we adapted the far WB technique to optimize the protein protein or protein carbohydrate interaction between sperm proteins and human rhZP. Protein conformation is known to play a role in protein protein interaction. Since sperm proteins were denatured for electrophoresis separation, we used renaturing transfer conditions, which allow the proteins to refold partially into their native structure. Under such conditions, we observed different target proteins distributed between 18 and 100 kDa with a pI ranging from 5 to 9 depending on the recombinant ZP used. This relatively large pattern in molecular weight corresponds quite well to the proteins already identied as ZP binding proteins in humans by ORand et al. (1985) and Shabanowitz and ORand (1988). Several spots of the same protein were also revealed, undoubtedly corresponding to post-translational modications (PTMs), such as phosphorylation and/or glycosylation. These two PTMs have been described to play an essential role in different steps of fertilization (i.e. capacitation, binding) via modication of protein function and/or of the antigenicity of spermatozoa . From this point of view, it is noticeable that in our far WB assays we used rhZP(s) expressed in CHO cells in place of native solubilized ZP. These recombinant proteins expressed in the CHO cell undergo PTMs that can be different from the native protein produced in human oocytes. This may affect the protein function since the sugar moieties of ZP glycoproteins play a critical role in the gamete interaction. However, it has been reported that CHO-expressed rhZPC can bind human spermatozoa and induce the acrosome reaction (Bray et al., 2002; Marin-Briggiler et al., 2008). So we believe that certain sugars (if not the totality) of these CHO-expressed rhZP are similar to those of the native ZP. Other rhZP(s) able to interact with spermatozoa have been produced in Escherichia coli and baculovirus expression systems (Chakravarty et al., 2005). It remains to be seen up to which point the results of the studies can be inuenced by the system of expression of proteins. Five of the sperm proteins identied by our double approach belong to the category of glycolytic enzymes involved in energy production

Downloaded from http://humrep.oxfordjournals.org/ at Health Sciences Library on February 12, 2013

Table V Induction of acrosome reaction by ZP in the presence of anti-VDCA2 antibody.

Experiment 3h

..........................................................
Anti-VDAC2 53.00* 72.80 67.33

18 h

..........................................................
Anti-VDAC2 86.00 62.60 62.04

Control 1 2 3 78.00 71.50 73.47

Control 91.00 70.22 58.67

.............................................................................................................................................................................................

Results are expressed as percentage of acrosome-reacted spermatozoa. *Signicantly different from control oocytes (x 2 test, P , 0.02).

Moonlighting sperm ZP binding proteins

11
remains to be addressed whether these glycolytic enzymes truly exert their catalytic activity when localized in the sperm head. Almost of all these glycolytic proteins are described as multifunctional with non-enzymatic moonlighting properties (Kim and Dang, 2005; Sriram et al., 2005) and may therefore have a different role according to their localization. Here, we propose a new role for these sperm glycolytic enzymes: they would act as potential sites of recognition and/or binding for ZP glycoproteins when located in the sperm head. We showed that ALDOA interacts with a ZP glycoprotein (ZP4) directly, in a far WB assay, and indirectly, in the functional ZP binding test. Its localization on the equatorial segment and at the basis of the head of acrosome-intact spermatozoa is also in agreement with its interaction with ZP4, a glycoprotein involved in the rst steps of the gamete interaction, before the acrosome reaction. It is known that sperm receptors for ZP act as lectin-like proteins that specically recognize sugar chains on ZP (Wassarman, 1995). We propose that glycolytic enzymes, which have oligosaccharides as substrates of their catalytic activity in the agellum, bind to the sugar moieties of ZP glycoproteins without exerting their catalytic activity, when located on the sperm head. Recently, the concept of multirecognition molecules assembled in a functional complex located in lipid rafts of the sperm membrane has emerged (Nixon et al., 2009) and these would interact with the carbohydrate segment of ZP glycoproteins. Multimolecular complexes involving glycolytic enzymes have already been described (Campanella et al., 2005). In spermatozoa, three testis-specic isoenzymes, HK1-S, muscle-type PFK and GSTM5 form a molecular complex associated with the mouse brous sheath (Nakamura et al., 2010). The existence of similar complexes with glycolytic enzymes remains to be demonstrated in the human sperm head.

(PK3, GAPDH, ENO1, ALDOA, TPI). The VDAC2 can also be included in this group. Indeed, VDAC2 transports adenine nucleotides, Ca2+ and other metabolites at the level of the outer mitochondrial membrane and therefore participates in energy production. Other proteins identied in the present study include proteins involved in detoxication processes (GSTM3 and PHGPx) and a structural protein belonging to the agellar cytoskeleton (ODF2). At rst sight, these results may seem surprising, as these proteins are primarily described to be located in the sperm tail. However, the literature reports, and our immunouorescence experiments revealed that they are also localized on the sperm head (on the plasma membrane or acrosomal membrane, insofar as soft permeabilization is sometimes necessary to visualize them). Additionally, co-precipitation experiments conrmed that some of them actually interact with ZP proteins. This is the case for GSTM with rhZP4, VDAC2 with rhZP2 and rhZP3, and TPI with rhZP3 and rhZP4. We thus demonstrate here that proteins with several functions according to their localization exist in the spermatozoa. The phenomenon by which a protein can perform more than one function has already been described and the protein is then termed moonlighting (Sirover, 1999; Kim and Dang, 2005; Sriram et al., 2005). Moreover, it can be noted that several of these proteins have germ cell-specic isoforms, differing from isoforms expressed in somatic cells (GAPDH, ENO1, ALDOA, TPI, etc.) (Edwards and Grootegoed, 1983; Welch et al., 1992; Russell and Kim, 1996; Vemuganti et al., 2007). This could also explain why the function of these proteins may be different in the gametes compared with somatic cells.

Downloaded from http://humrep.oxfordjournals.org/ at Health Sciences Library on February 12, 2013

Glycolytic enzymes identied as ZP binding proteins

With regard to the group of glycolytic enzymes identied in this work, it is worth noting that the majority have a double distribution, mitochondrial and extra-mitochondrial, in sperm cells. Recent studies reported the presence of PK3, GAPDH or TPI in the acrosome, in addition to the expected agellar localization. In the present work, we found ALDOA in the agellum and in the head of human spermatozoa, precisely at the equatorial band level of intact cells. The fact that the detection of ALDOA required permeabilization by saponin suggests that ALDOA is located at the internal side of the plasma membrane or at the acrosomal membrane. The observation of ALDOA in both the head and the agellum of human spermatozoa is an original result. The loss of the ALDOA staining in the sperm head and at the midpiece following ethanol treatment despite its preservation in the principal piece seems to indicate that ALDOA is not similarly organized in the head and in the principal piece where it is anchored to the brous sheath. Our group has already shown that TPI which is targeted by sperm autoantibodies is localized at the acrosomal level in human spermatozoa (Auer et al., 2004). Two other enzymes of the glycolytic pathway, hexokinase 1 and 6-phosphofructokinase (PFK), are also present in the acrosomal area of boar sperm (Kamp et al., 2007), increasing to six the number of glycolytic enzymes localized in the sperm head. Up to now, the function of these glycolytic enzymes in the human sperm head had not been demonstrated. Feiden et al. (2008) described GAPDH on the acrosome of boar spermatozoa and concluded that GAPDH is implicated in local ATP supply which drives ion pumps involved in the initiation of the acrosome reaction. But it

Detoxication enzyme GSTM and ZP binding

We have identied GSTM as a ligand for human ZP4 and ZP3. This result conrms the receptor activity of GSTM, which has already been reported in goat spermatozoa (Gopalakrishnan et al., 1998). In the present study, GSTM was found at the surface of the human sperm head in a region overlying the acrosome of intact spermatozoa and was not maintained after the acrosome reaction. This suggests that the GSTM-ZP4/ZP3 interaction occurs in the rst steps of gamete recognition. In the goat, GST was characterized as a protein of the sperm surface and was shown to bind specically to ZP during the rst phase of sperm oocyte interactions (Hemachand et al., 2002). Using the far WB method, the authors identied GST as a ligand for a ZP3-like protein. Our study agrees and reinforces the idea that GSTM is a sperm molecule for ZP3/ZP4 recognition during initial binding. This sperm-specic role of GST adds the detoxication function of GSTM which, by eliminating reactive oxygen species via glutathione, prevents lipid membrane peroxidation, a process highly damaging to sperm membrane integrity (Hemachand and Shaha, 2003). Thus, in spermatozoa, GSTM would also have a double role of cell protection and oocyte recognition.

VDAC2 as a site of ZP binding

Interestingly, we found by the two approaches (ASA WB and rhZP far WB) that VDAC2 interacts with ZP2 and ZP3 glycoproteins. The facts

12
that the presence of anti-VDAC2 antibodies decreases its ability to bind to ZP and that VDAC2 protein is localized on the sperm head corroborate the role of VDAC2 as a sperm ZP ligand. More precisely, VDAC2 was found on the acrosome of xed human sperm, in addition to its localization on the agellum, as reported in bovine spermatozoa (Hinsch et al., 2001, 2004; Triphan et al., 2008). On live permeabilized sperm cells, we observed VDAC2 staining at the base of the head (calyx), similarly to that found by Liu et al. (2009). As VDAC2 was only detected after a permeabilization treatment by saponin during the immunostaining procedure, we deduced that the epitopes recognized by commercial antibody were on the internal side of the plasma membrane or on the outer acrosomal membrane. Thus, in our conditions, we could not differentiate between the plasma membrane and the outer acrosomal membrane for the localization of VDAC2, as Triphan et al. (2008) found in their study. According to the literature, VDAC2 is a voltage-dependent, and not an ion-specic, channel commonly located on the outer mitochondrial membrane (topical review: Shoshan-Barmatz and Israelson, 2005). However, it also exhibits other extra-mitochondrial localizations as in the plasma membrane of lymphocytes (De Pinto et al., 2010) or the endoplasmic/sarcoplasmic reticulum membrane of skeletal muscle (Shoshan-Barmatz and Israelson, 2005). Recently, it has been reported that VDAC2 could play a role in the rapid transfer of the calcium released from the endoplasmic reticulum through a ryanodine receptor (RyR) and IP3 receptor (IP3R) to the outer mitochondrial membrane (Csordas and Hajnoczky, 2009). The presence of the same receptors RyR and IP3R in the acrosomal membrane suggests that VDAC2 could be localized at the level of this membrane in sperm cells where it could be involved in ATP and calcium transport between the acrosome and the cytoplasm (Triphan et al., 2008). A scaffolding function of VDAC2 has been evoked by De Pinto et al. (2010) suggesting that VDAC2 has the potential to assemble with other proteins to form a complex. Indeed, the association of VDAC2 with cytosolic enzymes including HK-1, ALDOA and GAPDH has been reported in the skeletal muscle (Shoshan-Barmatz and Israelson, 2005) and, in this complex, HK1 could modulate VDAC channel activity (Azoulay-Zohar et al., 2004). With this in view, we can hypothesize, in the context of the sperm head, that the glycolytic enzymes localized on the acrosome could contribute to VDAC biological function. Whether the interaction ZP-VDAC2, reported in this study, is able to modulate the permeability of VDAC2 and, as a consequence, regulate the calcium signal remains to be demonstrated.

Petit et al.

Acknowledgements
The authors thank Drs Marta De Almeida and Luc Camoin for their advice and scientic expertise and Drs Julie Coquet and Sandrine Barbaux for their critical reading of the manuscript. The authors also thank Prof. Pierre Jouannet (Service de Biologie de la Reproduc pital Cochin, Paris) for his kind collaboration notably in protion, Ho viding human gametes of infertile patients and donors.

Authors roles
F.M.P., C.S. and J.A. contributed to the study design, execution, analysis of results and drafting of the manuscript. F.B. and C.P. contributed to the proteome analysis and a critical reading of the manuscript.
Downloaded from http://humrep.oxfordjournals.org/ at Health Sciences Library on February 12, 2013

Funding
This work was supported by the grant ReproRio (CNRS, INRA, te dAndrologie de Langue Franc INSERM and CEA) and the Socie aise.

Conict of interest
None declared.

References
Auer J, Senechal H, De Almeida M. Sperm-associated and circulating IgA and IgG classes of antibodies recognise different antigens on the human sperm plasma membrane. J Reprod Immunol 1997;34:121 136. Auer J, Senechal H, Desvaux FX, Albert M, De Almeida M. Isolation and characterisation of two sperm membrane proteins recognised by sperm-associated antibodies in infertile men. Mol Reprod Dev 2000; 57:393 405. Auer J, Camoin L, Courtot AM, Hotellier F, De Almeida M. Evidence that P36, a human sperm acrosomal antigen involved in the fertilization process is triosephosphate isomerase. Mol Reprod Dev 2004; 68:515 523. Azoulay-Zohar H, Israelson A, Abu-Hamad S, Shoshan-Barmatz V. In self-defence: hexokinase promotes voltage-dependent anion channel closure and prevents mitochondria-mediated apoptotic cell death. Biochem J 2004;377:347 355. Bohring C, Krause W. Immune infertility: towards a better understanding of sperm (auto)-immunity. The value of proteomic analysis. Hum Reprod 2003;18:915 924. Bohring C, Krause E, Habermann B, Krause W. Isolation and identication of sperm membrane antigens recognized by antisperm antibodies, and their possible role in immunological infertility disease. Mol Hum Reprod 2001;7:113 118. Bray C, Son JH, Kumar P, Harris JD, Meizel S. A role for the human sperm glycine receptor/Cl(-) channel in the acrosome reaction initiated by recombinant ZP3. Biol Reprod 2002;66:91 97. Campanella ME, Chu H, Low PS. Assembly and regulation of a glycolytic enzyme complex on the human erythrocyte membrane. Proc Natl Acad Sci USA 2005;102:2402 2407. Chakravarty S, Suraj K, Gupta SK. Baculovirus-expressed recombinant human zona pellucida glycoprotein-B induces acrosomal exocytosis in capacitated spermatozoa in addition to zona pellucida glycoprotein-C. Mol Hum Reprod 2005;11:365 372.

Conclusion
Our work led to the identication of human sperm proteins that are recognized both by ASA eluted from the spermatozoa of patients who have had IVF failures and by rhZP glycoproteins. This double approach allows us to reliably assign to these proteins a role in gamete interaction and specically in binding/recognition of the ZP. These proteins were already known for exerting functions at the agellar level such as glycolysis (ALDOA, TPI, ENO1) or ion transport (VDAC2). Here, we propose an additional function for these proteins (zona binding) when they are located in the sperm head. Their multiple roles depending on their sub-cellular localization suggest that these sperm proteins are multifaceted or moonlighting proteins.

Moonlighting sperm ZP binding proteins

13
Kovarik A, Hlubinova K, Vrbenska A, Prachar J. An improved colloidal silver staining method of protein blots on nitrocellulose membranes. Folia Biol (Praha) 1987;33:253 257. Liu B, Wang Z, Zhang W, Wang X. Expression and localization of voltage-dependent anion channels (VDAC) in human spermatozoa. Biochem Biophys Res Commun 2009;378:366 370. Machida K, Mayer BJ. Detection of protein-protein interactions by far-western blotting. Methods Mol Biol 2009;536:313 329. Manaskova-Postlerova P, Davidova N, Jonakova V. Biochemical and binding characteristics of boar epididymal uid proteins. J Chromatogr B Analyt Technol Biomed Life Sci 2011;879:100 106. Marin-Briggiler CI, Gonzalez-Echeverria MF, Harris JD, Vazquez-Levin MH. Recombinant human zona pellucida protein C produced in Chinese hamster ovary cells binds to human spermatozoa and inhibits sperm-zona pellucida interaction. Fertil Steril 2008;90:879 882. Martic M, Moses EK, Adams TE, Liu DY, Gook DA, Garrett C, Dunlop ME, Baker GH. Recombinant human zona pellucida proteins ZP1, ZP2 and ZP3 co-expressed in a human cell line. Asian J Androl 2004;6:3 13. McClure RD, Nunes L, Tom R. Semen manipulation: improved sperm recovery and function with a two-layer Percoll gradient. Fertil Steril 1989;51:874 877. McLeskey SB, Dowds C, Carballada R, White RR, Saling PM. Molecules involved in mammalian sperm-egg interaction. Int Rev Cytol 1998; 177:57 113. Nakamura N, Mori C, Eddy EM. Molecular complex of three testis-specic isozymes associated with the mouse sperm brous sheath: hexokinase 1, phosphofructokinase M, and glutathione S-transferase mu class 5. Biol Reprod 2010;82:504 515. Nixon B, Aitken RJ, McLaughlin EA. New insights into the molecular mechanisms of sperm-egg interaction. Cell Mol Life Sci 2007; 64:1805 1823. Nixon B, Bielanowicz A, McLaughlin EA, Tanphaichitr N, Ensslin MA, Aitken RJ. Composition and signicance of detergent resistant membranes in mouse spermatozoa. J Cell Physiol 2009;218:122 134. ORand MG, Matthews JE, Welch JE, Fisher SJ. Identication of zona binding proteins of rabbit, pig, human, and mouse spermatozoa on nitrocellulose blots. J Exp Zool 1985;235:423 428. Paradowska A, Bohring C, Krause E, Krause W. Identication of evolutionary conserved mouse sperm surface antigens by human antisperm antibodies (ASA) from infertile patients. Am J Reprod Immunol 2006;55:321 330. Russell DL, Kim KH. Expression of triosephosphate isomerase transcripts in rat testis: evidence for retinol regulation and a novel germ cell transcript. Biol Reprod 1996;55:11 18. Shabanowitz RB, ORand MG. Molecular changes in the human zona pellucida associated with fertilization and human sperm-zona interactions. Ann N Y Acad Sci 1988;541:621 632. Shetty J, Naaby-Hansen S, Shibahara H, Bronson R, Flickinger CJ, Herr JC. Human sperm proteome: immunodominant sperm surface antigens identied with sera from infertile men and women. Biol Reprod 1999; 61:61 69. Shetty J, Diekman AB, Jayes FC, Sherman NE, Naaby-Hansen S, Flickinger CJ, Herr JC. Differential extraction and enrichment of human sperm surface proteins in a proteome: identication of immunocontraceptive candidates. Electrophoresis 2001;22:3053 3066. Shetty J, Bronson RA, Herr JC. Human sperm protein encyclopedia and alloantigen index: mining novel allo-antigens using sera from ASApositive infertile patients and vasectomized men. J Reprod Immunol 2008; 77:23 31. Shevchenko A, Wilm M, Vorm O, Mann M. Mass spectrometric sequencing of proteins silver-stained polyacrylamide gels. Anal Chem 1996;68:850 858.

Chirinos M, Carino C, Gonzalez-Gonzalez ME, Arreola E, Reveles R, Larrea F. Characterization of human sperm binding to homologous recombinant zona pellucida proteins. Reprod Sci 2011;18:876 885. Chiu WW, Erikson EK, Sole CA, Shelling AN, Chamley LW. SPRASA, a novel sperm protein involved in immune-mediated infertility. Hum Reprod 2004;19:243 249. Chiu PC, Chung MK, Koistinen R, Koistinen H, Seppala M, Ho PC, Ng EH, Lee KF, Yeung WS. Glycodelin-A interacts with fucosyltransferase on human sperm plasma membrane to inhibit spermatozoa-zona pellucida binding. J Cell Sci 2007;120:33 44. Clarke GN, Stojanoff A, Cauchi MN, McBain JC, Speirs AL, Johnston WI. Detection of antispermatozoal antibodies of IgA class in cervical mucus. Am J Reprod Immunol 1984;5:61 65. Com E, Evrard B, Roepstorff P, Aubry F, Pineau C. New insights into the rat spermatogonial proteome: identication of 156 additional proteins. Mol Cell Proteomics 2003;2:248 261. Cross NL, Morales P, Overstreet JW, Hanson FW. Two simple methods for detecting acrosome-reacted human sperm. Gamete Res 1986;15:213226. Csordas G, Hajnoczky G. SR/ER-mitochondrial local communication: calcium and ROS. Biochim Biophys Acta 2009;1787:1352 1362. De Pinto V, Messina A, Lane DJ, Lawen A. Voltage-dependent anion-selective channel (VDAC) in the plasma membrane. FEBS Lett 2010;584:1793 1799. Domagala A, Pulido S, Kurpisz M, Herr JC. Application of proteomic methods for identication of sperm immunogenic antigens. Mol Hum Reprod 2007;13:437 444. Dunn SD. Effects of the modication of transfer buffer composition and the renaturation of proteins in gels on the recognition of proteins on Western blots by monoclonal antibodies. Anal Biochem 1986; 157:144 153. Edmondson DG, Roth SY. Identication of protein interactions by far Western analysis. Curr Protoc Mol Biol 2001;Chapter 20:Unit 20 26. Edwards YH, Grootegoed JA. A sperm-specic enolase. J Reprod Fertil 1983;68:305 310. Feiden S, Wolfrum U, Wegener G, Kamp G. Expression and compartmentalisation of the glycolytic enzymes GAPDH and pyruvate kinase in boar spermatogenesis. Reprod Fertil Dev 2008;20:713 723. Gopalakrishnan B, Aravinda S, Pawshe CH, Totey SM, Nagpal S, Salunke DM, Shaha C. Studies on glutathione S-transferases important for sperm function: evidence of catalytic activity-independent functions. Biochem J 1998;329:231 241. Harris JD, Seid CA, Fontenot GK, Liu HF. Expression and purication of recombinant human zona pellucida proteins. Protein Expr Purif 1999; 16:298 307. Hemachand T, Shaha C. Functional role of sperm surface glutathione S-transferases and extracellular glutathione in the haploid spermatozoa under oxidative stress. FEBS Lett 2003;538:14 18. Hemachand T, Gopalakrishnan B, Salunke DM, Totey SM, Shaha C. Sperm plasma-membrane-associated glutathione S-transferases as gamete recognition molecules. J Cell Sci 2002;115:2053 2065. Hinsch KD, Asmarinah, Hinsch E, Konrad L. VDAC2 (porin-2) expression pattern and localization in the bovine testis. Biochim Biophys Acta 2001; 1518:329 333. Hinsch KD, De Pinto V, Aires VA, Schneider X, Messina A, Hinsch E. Voltage-dependent anion-selective channels VDAC2 and VDAC3 are abundant proteins in bovine outer dense bers, a cytoskeletal component of the sperm agellum. J Biol Chem 2004;279:15281 15288. Kamp G, Schmidt H, Stypa H, Feiden S, Mahling C, Wegener G. Regulatory properties of 6-phosphofructokinase and control of glycolysis in boar spermatozoa. Reproduction 2007;133:29 40. Kim JW, Dang CV. Multifaceted roles of glycolytic enzymes. Trends Biochem Sci 2005;30:142 150.

Downloaded from http://humrep.oxfordjournals.org/ at Health Sciences Library on February 12, 2013

14
Shoshan-Barmatz V, Israelson A. The voltage-dependent anion channel in endoplasmic/sarcoplasmic reticulum: characterization, modulation and possible function. J Membr Biol 2005;204:57 66. Sirover MA. New insights into an old protein: the functional diversity of mammalian glyceraldehyde-3-phosphate dehydrogenase. Biochim Biophys Acta 1999;1432:159 184. Sriram G, Martinez JA, McCabe ER, Liao JC, Dipple KM. Single-gene disorders: what role could moonlighting enzymes play? Am J Hum Genet 2005;76:911 924. Stein KK, Go JC, Lane WS, Primakoff P, Myles DG. Proteomic analysis of sperm regions that mediate sperm-egg interactions. Proteomics 2006; 6:3533 3543. Tanii I, Oh-oka T, Yoshinaga K, Toshimori K. A mouse acrosomal cortical matrix protein, MC41, has ZP2-binding activity and forms a complex with a 75-kDa serine protease. Dev Biol 2001;238:332 341. Topfer-Petersen E, Cechova D, Henschen A, Steinberger M, Friess AE, Zucker A. Cell biology of acrosomal proteins. Andrologia 1990; 22(Suppl 1):110 121.

Petit et al.

Triphan X, Menzel VA, Petrunkina AM, Cassara MC, Wemheuer W, Hinsch KD, Hinsch E. Localisation and function of voltage-dependent anion channels (VDAC) in bovine spermatozoa. Pugers Arch 2008; 455:677 686. van Gestel RA, Brewis IA, Ashton PR, Brouwers JF, Gadella BM. Multiple proteins present in puried porcine sperm apical plasma membranes interact with the zona pellucida of the oocyte. Mol Hum Reprod 2007; 13:445 454. Vemuganti SA, Bell TA, Scarlett CO, Parker CE, de Villena FP, OBrien DA. Three male germline-specic aldolase A isozymes are generated by alternative splicing and retrotransposition. Dev Biol 2007;309:1831. Wassarman PM. Mammalian fertilization: egg and sperm (glyco)proteins that support gamete adhesion. Am J Reprod Immunol 1995;33:253 258. Welch JE, Schatte EC, OBrien DA, Eddy EM. Expression of a glyceraldehyde 3-phosphate dehydrogenase gene specic to mouse spermatogenic cells. Biol Reprod 1992;46:869 878. WHO. WHO Laboratory Manual for Examination and Processing of Human Semen, 5th edn. Geneva: Cambridge University Press, 2010.

Downloaded from http://humrep.oxfordjournals.org/ at Health Sciences Library on February 12, 2013