In addition to morphological criteria, biochemical analyses on the ZP was performed to categorise the oocytes as "fertilised" or "unfertilised". As proteolysis of ZP2 is a known biochemical modification of the ZP occurring after the fertilisation, we determined by western blotting the cleavage status of ZP2 in ZP isolated from the unfertilised and fertilised oocytes used in our work (Figure 1). In unfertilised embryos, the anti-rhZP2 antibody detected the ZP2 as a major band at 100 kDa (lane 1 in Figure 1A and 1B), which disappeared almost totally in fertilised oocytes and was essentially replaced by two bands of 70 kDa and 30 kDa (lane 2 in Figure 1A and 1B). This result indicates that our hZP preparations were suitable to represent a fertilised and unfertilised human oocyte/egg material.

After 3 h hours of gamete co-incubation, the zonae pellucidae of all inseminated oocytes, whether fertilised or not, were able to bind spermatozoa (Figure 2). The number of spermatozoa bound by intact ZP varied from 12 to 76 for unfertilised and from 6 to 65 for fertilised oocytes. However, the mean number of spermatozoa bound per fertilised oocyte was significantly lower compared to unfertilised oocyte (p < 0.05, Table 1).

The mean number of embedded spermatozoa was significantly higher after co-incubation of spermatozoa with unfertilised intact hZP compared to fertilised hZP (p < 0.001; Table 4). The penetration rate (proportion of penetrated hZP) was significantly higher for unfertilised than for fertilised oocytes (p < 0.05, Table 4) in each of the 3 experiments. The overall penetration rate for unfertilised and fertilised oocytes was 57% and 19%, respectively.

For practical and ethical reasons, only unfertilised or fertilised oocytes which are usually discarded in an IVF program can be used for experiments. Consequently, the oocytes and zygotes have to be cultured for 48 h before deciding to use them for research. Could this condition be responsible for uncertain or dubious physiological relevance of measured ZP bioactivity? It has been shown previously that unfertilised oocytes collected after IVF failure could be used to study sperm – ZP interaction 17. In our study, unfertilised oocytes were mature but no pronucleus was observed 18 hours post insemination and no cleavage occurred within 48 hours. Furthermore, we controlled that there was no ZP2 proteolysis in these unfertilised oocytes. On the contrary, fertilised oocytes defined by the presence of at least 3 PN at 18 hours after fertilisation or by the presence of 2 PN but abnormal cleavage after 48 hours had a ZP2 cleaved in two bands of 70 and 30 kDa, denoting a limited proteolysis likely due to cortical granule exocytosis 18. Therefore, we are confident that the material used in our study was relevant to unfertilised oocytes or fertilised zygotes to analyse hZP bioactivity.

In mice, spermatozoa do not bind to fertilised oocytes and the ZP3 protein purified from two-cell embryos cannot induce the AR in mouse spermatozoa 4. Surprisingly, we observed that zona pellucida of human fertilised oocytes, retain its capacity to bind spermatozoa, albeit about half as efficiently as zona from unfertilised oocytes and was able to induce an intra-spermatic calcium influx of the same amplitude as hZP from unfertilised oocytes. The spermatozoa bound to fertilised oocytes were found to be both acrosome-reacted and acrosome-intact spermatozoa. A recent study has reported that acrosome-reacted spermatozoa have a drastically reduced or zero ability to directly bind ZP 19. Therefore, it may be hypothesised that acrosome-reacted spermatozoa observed on fertilised ZP were acrosome intact before attachment to ZP, and underwent the AR upon contact with hZP. Our data suggest that (1) hZP of fertilised oocytes is able to bind sperm and to induce AR (also confirmed by test using solubilised ZP of fertilised oocytes) and (2) acrosome-reacted spermatozoa remain attached to fertilised ZP despite the cleavage of ZP2. Thus the human glycoproteins involved in sperm binding and AR induction – presumably ZP4 (ZPB) and ZP3 (ZPC) 1213 – seem to remain functional after fertilisation. This is the original point of our study that contrasts with observations in mice showing that fertilisation modifies the ZP, preventing sperm binding and AR induction. The ZP of mouse embryos was clearly shown to be unable to induce the AR in mouse spermatozoa, in experiments using purified ZP3 from the embryo as an inducer 3, but post-fertilisation modifications of mouse ZP3 have not been demonstrated directly. Thus, the mouse gamete interaction does not seem to model the situation in human in a reliable manner. Could the additional glycoprotein, ZP4, present in human oocyte but not in mice account for this difference? Does ZP4 work individually or in a heterocomplex with ZP3, as described in boars 10?

In our study, the biochemical analysis of the hZP collected after fertilisation confirmed that ZP2 was cleaved, demonstrating that sperm binding and the AR can occur whether ZP2 is cleaved or uncleaved. Using a model of chimeric human-mouse oocytes obtained by transgenesis, Rankin et al. 20 suggested that sperm binding is a function of the supramolecular structure of the ZP, modulated by the cleavage status of ZP2, with sperm binding occurring when ZP2 is uncleaved but not when ZP2 is cleaved. Our observation of sperm binding to the zona pellucida when human fertilised oocytes with cleaved ZP2 were re-inseminated suggests that Rankin's model, although appropriate for oocytes with three zona glycoproteins as in mouse or chimeric human-mouse ZP, would not be applicable to species with four zona proteins, as in humans.

Intact hZP from fertilised and unfertilised human oocytes are equally efficient in inducing the AR, with similar rates as reported previously for unfertilised oocytes 21. By contrast, solubilised ZP-induced AR rates measured in the present study were two times higher (ΔAR = 40%) than those previously reported (ΔAR = 20%) 2223. This difference can arise from different experimental conditions (time of capacitation, incubation medium...) or from the probes used to reveal the AR. We used the anti-CD46 antibody coupled to the flow cytometry to measure solubilised hZP-induced AR and not the PSA lectin labelling widely used in the other studies. From our knowledge, this is the first time that the flow cytometry, an objective method which allows counting on several thousand cells, was used to measure ZP-induced AR. As PSA lectin is not the most suitable probe for AR measurement in flow cytometry 24, we have chosen anti-CD46 antibody that gives for acrosome-reacted cells a high fluorescent peak well separated from the low fluorescent peak of intact cells 25. By contrast, the fluorescent peaks of intact and acrosome-reacted spermatozoa are more or less overlapped when PSA lectin staining is used 26. In our study, the use of two different probes to mark the AR induced by either solubilised (anti-CD46) or intact (PSA lectin) ZP, does not make it possible to compare the results. Indeed, it is known that different methods of AR staining never give the same results when the probes target different sites in the sperm cells, measuring different steps in the AR kinetics 27. Our objective was to compare the AR inducing ability of ZP of fertilised and unfertilised oocytes within a same experiment.

In our experiments, more than 80% of fertilised oocytes co-incubated with capacitated spermatozoa had no spermatozoa within their ZP. This observation shows that human fertilised oocytes cannot be more penetrated by spermatozoa and is consistent with the murine model as demonstrated in previous studies 428. This inhibition of sperm penetration could be temporally correlated with the proteolytic cleavage of ZP2, demonstrated after fertilisation in mice 29 and in humans [our study].

In summary, the events of ZP sperm binding/AR induction and zona penetration can be dissociated in humans. The main step of polyspermy blocking would be the inhibition of the acrosome-reacted sperm penetration through the zona after fertilisation. Some post-fertilisation architectural alterations could provoke a hardening of the zona which make it impossible to be lysed by the acrosomal enzymes.

Among the different functions of ZP analysed in this work, we have brought particular attention to the ZP-induced calcium influx in spermatozoa. Until now, this function of ZP had not been studied in humans, except with recombinant ZP3 as the inducer 56. Our study brings detailed information on intraspermatic calcium increase induced by solubilised hZP. When disaggregated by acid treatment, it retained its ability to induce an intracellular increase in calcium concentration, as previously reported for AR induction 303132.

The calcium influx induced by solubilised hZP displayed qualitative and quantitative aspects similar to those reported for other species. Using spectrofluorimetry, we found that the calcium response to hZP may be biphasic or monophasic, depending on the sample. Previous studies using spectrofluorimetry and a similar probe showed that the Ca²⁺ response to solubilised mouse ZP 33 or recombinant hZP3 56 was biphasic, with a first transient peak and a second sustained phase. However, in nine of the 21 cases analysed in our study, the calcium response appeared to be monophasic, with only sustained increase in calcium concentration. Either the transient calcium peak did not occur or it developed so rapidly that it was not recorded in our conditions. If the initial peak lasted 200 ms as in mice 16, the peak development was probably too rapid to be detected, and this could explain the previous reports of a monophasic increase in calcium concentration in response to ZP in cattle 34 and in mice 1435. However, Shirakawa and Miyazaki 36 detected no transient calcium peak with recordings taken at 0.3-second intervals in the hamster sperm head after the addition of solubilised ZP. These results and ours suggest that a sustained calcium influx, keeping [Ca²⁺]_i high for a sufficiently long period, may be sufficient to trigger the AR. The importance of the sustained phase in AR induction is also illustrated by the similar AR rates induced by hZP in sperm samples with monophasic and biphasic calcium responses.

The calcium response to solubilised human ZP is sensitive to pimozide (>80% inhibition), suggesting that VOCC is involved in this ionic event. T-type VOCC activity, involving the Cay3.2 channel in particular, has been associated with AR in humans 3738 and it has been shown that ZP-induced calcium entry and AR depend on T-type VOCC in mice 39.

We adopted the nomenclature recently proposed by Conner et al. 40, which follows the proposals of the NCBI committee. The four ZP genes are named ZP1, ZP2, ZP3 and ZP4, and encode proteins ZP1 to ZP4, respectively. The ZP2, ZP3 and ZP4 genes correspond to the ZPA, ZPC and ZPB genes, respectively, in Harris' nomenclature 41.

BWW medium contained 95 mM NaCl, 4.8 mM KCl, 1.7 mM CaCl₂, 1.2 mM KH₂PO₄, 1.2 mM MgSO₄, 5.6 mM glucose, 0.25 mM sodium pyruvate, 20 mM HEPES -free acid, 25 mM NaHCO₃, 20 mM lactic acid, pH adjusted to 7.5 with NaOH. The osmolarity of this solution was 300–310 mosmol/l. In some experiments, a modified BWW medium devoid of Ca²⁺ was used. This medium was prepared by omitting CaCl₂ and adding 500 μM EGTA (Ca²⁺ free BWW).

B2 medium was obtained from CCD (Paris, France) and Dulbecco's phosphate-buffered saline (PBS) was obtained from Sigma (Sigma-Aldrich, Saint-Quentin Fallavier, France).

Pimozide dissolved in absolute ethanol, Fura-2 AM dissolved in DMSO, sodium pyruvate, fluorescein isothiocyanate conjugated-goat anti-mouse antibody (FITC-GAM Ig), bovine serum albumin (BSA; fraction V) and HEPES were purchased from Sigma. Rhodamine-labelled Pisum sativum agglutinin (TRITC-PSA) was obtained from Vector (AbCys S.A., Paris, France). Anti-CD46 monoclonal antibodies were purchased from Immunotech (Marseille, France). All other chemicals (salts for buffers) were purchased from Merck/VWR (VWR, Fontenay-sous-Bois, France). Antibodies against human recombinant ZP2, produced in CHO cells, were a gift from J Harris (Zonagen, Inc., Woodlands, Texas, USA).

The selected and capacitated motile spermatozoa (1 × 10⁷/ml) were incubated for 45 minutes at 37°C with 2 μM Fura-2 AM in 1% BSA in BWW, in a 5% CO₂/95% air atmosphere. They were then washed by dilution (1/3-v/v) in BWW and centrifuged (600 g, 10 min). The pellets were suspended at a density of 1.5 × 10⁷cells/ml in BWW and kept at room temperature in the dark for a further 30 minutes before use. Aliquots (13 μl) of sperm suspension were placed in a glass microcuvette at 37°C and 4 μl of solubilised hZP or 5 mM NaH₂PO₄ buffered in 2 × BWW (control) were added. We used an hZP concentration of 2 hZP/μl, which lies within the range of concentrations used in most studies 3032. Fluorescence was recorded with a PTI M-2001 spectrofluorimeter (Kontron Instruments, Saint-Quentin-en-Yvelines, France) at an emission wavelength of 505 nm, using dual excitation at 340 and 380 nm (5 nm bandpass). Recordings were taken over 15 minutes and the sperm suspension was then collected and incubated for a further 15 min at 37°C in a 5% CO₂/95% air atmosphere to allow completion of the AR.

We calibrated the fluorescent signal by carrying out parallel experiments in which sperm were permeabilised by incubation with 0.05% Triton X-100, to which 10 mM EGTA (pH 9.5) was subsequently added, for determination of the maximal and minimal fluorescence signals, respectively. [Ca²⁺]_i was calculated from the 340/380 nm excitation ratio, as described by Grynkiewicz et al. 44 and reported by Yang et al. 45.

Capacitated spermatozoa suspended in B2 medium were labelled with fluorescein isothiocyanate (FITC), as previously described 46, to distinguish them from spermatozoa eventually remaining bound to oocytes of IVF. Briefly, 100 μl of 1% FITC in 0.1 M KOH were diluted in 5 ml of PBS supplemented with 1% glucose. Capacitated spermatozoa were diluted 1/2 or 1/4 in the FITC solution and incubated for 10 min at 37°C then passed through a Percoll gradient in order to select only motile stained spermatozoa. We incubated 5 × 10⁵ FITC-labelled capacitated spermatozoa with 500 μl of B2 medium containing 7 to 23 unfertilised or fertilised oocytes with an intact zona pellucida for 3 h (primary binding) or 18 h (secondary binding/penetration) at 37°C in an atmosphere containing 5% CO₂/95% air.

For primary binding experiments, oocytes were washed in B2 medium by repeated aspiration in a glass pipette (internal diameter 250 μm), to dislodge loosely bound spermatozoa. The first aliquot of washed oocytes was placed in a glass depression slide and the number of FITC-labelled spermatozoa bound to each oocyte was determined with a Nikon E600 epifluorescence microscope (Nikon, St Quentin-en-Yvelines, France). The second aliquot of washed oocytes was treated as described below to determine the acrosomal status of bound spermatozoa.

After 18 hours of incubation time, the oocytes were manipulated according to the procedure reported by Liu and Baker 47 to assess the sperm-zona pellucida penetration. Briefly, the oocytes were washed to remove loosely bound spermatozoa and repetitively aspirated in and out of a pipette with an internal diameter of 125 μm to dislodge spermatozoa bound to the ZP, which were then examined for AR. These oocytes, cleared of spermatozoa on the surface of the ZP, were placed on a slide in a 20 μl drop of 0.2% BSA in PBS and the number of FITC-labelled spermatozoa remaining embedded within the zona pellucida or present in the perivitelline space were counted under the epifluorescence microscope at × 400 magnification.

Isolated zonae pellucidae were solubilised (v/v) in reducing sample buffer (20% glycerol, 4% SDS, 0.125 M Tris-HCl pH 6.5, 10% β-ME), heated for 5 min at 96°C and centrifuged for 5 min at 12000 g. The supernatant (equivalent of 3 solubilised ZP) was subjected to electrophoresis at 150 V for 45 min. One dimensional SDS-PAGE was performed with a 7% acrylamide gel, using a Mini PROTEAN II Cell Apparatus (Bio-Rad Laboratories, Marne-la-Coquette, France).

The separated proteins were electrotransferred onto PVDF membranes (GE Healthcare Bio-Sciences, France) for ZP2 detection. The blots were blocked by incubation with Tris buffer containing 1% (w/v) gelatin and 0.1% (v/v) Tween-20 and were then incubated with anti-rhZP2 antibodies diluted 1:2000 in the same buffer. Antibody binding was detected by incubation with horseradish peroxidase-conjugated anti-IgG antibody (dilution 1:30000), by enhanced chemiluminescence (ECL+; GE Healthcare Bio-Sciences).

The data are expressed as means ± SEM and were compared using Student's t test or the Mann and Whitney test. For the AR and penetration rates, chi² tests were used for comparisons. Values of p < 0.05 were considered statistically significant.