Molecular Mechanism Underlying the Action of Zona-pellucida Glycoproteins on Mouse Sperm

Mammalian oocytes are enveloped by the zona pellucida (ZP), an extracellular matrix of glycoproteins. In sperm, stimulation with ZP proteins evokes a rapid Ca2+ influx via the sperm-specific, pH-sensitive Ca2+ channel CatSper. However, the physiological role and molecular mechanisms underlying ZP-dependent activation of CatSper are unknown. Here, we delineate the sequence of ZP-signaling events in mouse sperm. We show that ZP proteins evoke a rapid intracellular pHi increase that rests predominantly on Na+/H+ exchange by NHA1 and requires cAMP synthesis by the soluble adenylyl cyclase sAC as well as a sufficiently negative membrane potential set by the spem-specific K+ channel Slo3. The alkaline-activated CatSper channel translates the ZP-induced pHi increase into a Ca2+ response. Our findings reveal the molecular components underlying ZP action on mouse sperm, opening up new avenues for understanding the basic principles of sperm function and, thereby, mammalian fertilization.

Here, we study the action of solubilized ZP proteins on mouse sperm. We show that the ZP-induced pH i increase is required to evoke Ca 2+ influx via CatSper. The pH i increase is abolished upon depolarization, which underlies the V m -control of CatSper activation by ZP proteins. Moreover, we show that the ZPinduced pH i increase rests on Na + /H + exchange by NHA1, Alias Slo3 Gene Adcy10 Adenylyl cyclase type 10 Protein Adcy10 Alias sAC

Gene Lrrc52
Leucine-rich repeat-containing protein 52 Protein Lrrc52 but not by sNHE, and requires cAMP synthesis by the soluble adenylyl cyclase sAC. Altogether, our findings answer longstanding questions about the molecular mechanisms underlying ZP action on mouse sperm.

ZP-Induced Ca 2+ -and pH i -Signaling Events in Mouse Sperm
We studied the action of solubilized ZP proteins on mouse sperm. In mice, the ZP consists of three glycoproteins (mZP1-3). In line with previous studies (Avella et al., 2014), staining of isolated oocytes with antibodies against mZP1, mZP2, and mZP3 labeled the ZP surrounding the oocyte ( Figure 1A). The specificity of the anti-ZP antibodies was confirmed by detection of heterologously expressed mZP1, mZP2, and mZP3 in both immunocytochemistry and Western blots (Supplementary Figure S1 and Figure 1C). On Western blots of solubilized ZPs isolated from mouse oocytes, the antibodies detected proteins with apparent molecular weights (M w ) of about 150, 100, and 83 kDa for mZP1, mZP2, and mZP3, respectively ( Figure 1B), similar to what has been shown previously (Bleil and Wassarman, 1980;Wassarman, 1988;Thaler and Cardullo, 1996). Treatment with PNGase F decreased the M w , demonstrating that glycosylation of ZP proteins was preserved during isolation ( Figure 1B). We analyzed the action of ZP proteins on mouse sperm using a stopped-flow apparatus and fluorescent probes for Ca 2+ and pH i . Mixing of capacitated sperm (incubated for 90 min in buffer containing 3 mg/ml BSA and 25 mM HCO − 3 at pH 7.4) with ZP proteins evoked a rapid Ca 2+ increase that reached a plateau within 30-40 s ( Figure 1D); the control signal evoked by mixing with buffer alone was subtracted, setting the control-signal level consistently to F/F 0 (%) = 0. Simultaneous alkalization and depolarization of sperm by mixing with buffer adjusted to pH 8.6 and containing a high K + and a low Na + concentration (dubbed K8.6 buffer) (Babcock and Pfeiffer, 1987) also evoked a rapid Ca 2+ increase that reached a slightly higher [Ca 2+ ] i ( Figure 1D). As a reference for the maximal signal amplitude at saturation of the Ca 2+ indicator, we recorded the Ca 2+ response evoked by the Ca 2+ ionophore ionomycin ( Figure 1D). ZP proteins, K8.6 buffer, and ionomycin also evoked Ca 2+ responses in non-capacitated sperm ( Figure 1E and Supplementary Figures  S2a,b). Yet, relative to the reference signal evoked by ionomycin, the amplitudes of ZP-and K8.6-evoked Ca 2+ response were enhanced upon capacitation ( Figure 1E and Supplementary Figure S2b). To verify that the Ca 2+ responses are carried by CatSper, we measured Ca 2+ responses in wild-type and CatSper1deficient mice. Indeed, the Ca 2+ responses were abolished in sperm from Catsper1 deficient-mice (Catsper1-KO, Figure 1F) (Xia and Ren, 2009).
Next, we studied ZP-induced pH i responses. Mixing of capacitated sperm with ZP proteins evoked a rapid pH i increase that reached a plateau after 20-30 s ( Figure 1G). Mixing of sperm with the weak base NH 4 Cl as a positive FIGURE 1 | Changes in pH i and [Ca 2+ ] i stimulated by solubilized zona pellucida (ZP) glycoproteins in mouse sperm. (A) Mouse oocytes labeled with antibodies directed against mZP1 (purple), mZP2 (cyan), and mZP3 (green); the DNA was labeled using DAPI (blue). (B) Western blot of solubilized ZP glycoproteins before (-) and after (+) PNGase-F treatment. The blots were probed with ZP isoform-specific antibodies. (C) Western blot of heterologously-expressed mouse ZP glycoproteins before (-) and after (+) PNG-F treatment, probed with mZP isoform-specific antibodies; NT: non-transfected cells. (D) Ca 2+ responses in populations of sperm evoked by mixing with 1 ZP/µl, K8.6 buffer, or 2 µM ionomycin; shown are the averages (solid lines, n = 7) and the 95% confidence interval (dashed lines). Shown is the percent change in fluorescence ( F) with respect to the mean of the first three data points recorded immediately after mixing (F 0 ). The control F/F 0 signal observed upon mixing with buffer (control) was subtracted from K8 6-, ZP-, or ionomycin-induced signals, setting the control-signal level to F/F (%) = 0. (E) Relative amplitude of the Ca 2+ responses evoked by mixing with 1 ZP/µl or K8.6 (mean ± SD of the average of the last three data points, n ≥ 7) in non-capacitated (-) and capacitated (+) sperm, normalized to the ionomycin-evoked Ca 2+ responses. (F) Relative amplitude of Ca 2+ responses in wild-type and Catsper1-KO sperm (n = 4). (G) pH i responses evoked by mixing with 1 ZP/µl or 10 mM NH 4 Cl in mouse sperm populations; shown are averages (solid lines, n = 7) and the 95% confidence interval (dashed lines). Shown is the percent change in fluorescence ratio ( R) with respect to the mean of the first three data points recorded immediately after mixing (R 0 ). The control R/R 0 signal observed upon mixing with buffer (control) was subtracted, setting the control-signal level to R/R 0 (%)= 0. (H) Amplitude of pH i responses (mean ± SD of the average of the last three data points, n ≥ 7) evoked by mixing with 1 ZP/µl or 10 mM NH 4 Cl in wild-type and Catsper1-KO sperm (n = 4). (I) pH i responses in sperm bathed in 0 mM Na + buffer (n = 4). (J) Amplitude of pH i responses at 138 or 0 mM extracellular Na + (mean ± SD, n = 4). (K) Amplitude of pH i responses in the absence (-) or presence (+) of 100 µM EIPA (mean ± SD, n = 4). Statistical significance between two groups was determined using two-tailed, unpaired t-test with Welch's correction; p-values are indicated.
control also evoked a rapid pH i increase that reached a slightly higher pH i ( Figure 1G). The ZP-and NH 4 Cl-evoked pH i responses were similar in wild-type and Catsper1-KO sperm ( Figure 1H).
Altogether, these results confirm that ZP action on sperm involves an increase of pH i (Arnoult et al., 1996b) and a Ca 2+ influx via CatSper (Xia and Ren, 2009), and that capacitation enhances the Ca 2+ response (Arnoult et al., 1999). Moreover, because the pH i increase is preserved in Catsper1-KO mice, we conclude that the pH i increase is not evoked by Ca 2+ influx via CatSper. Instead, the ZP-induced pH i increase might underlie the ZP-activation of CatSper.
The ZP-Induced Alkalization Does Not Involve the Na + /H + Exchanger sNHE We aimed to unravel the molecular players underlying the ZP-evoked alkalization. Na + /H + exchange via sNHE has been proposed to take part in the ZP-evoked alkalization (Chavez et al., 2014). In addition to sNHE, two members of the Na + /H + antiporter (NHA) subfamily, NHA1 and NHA2 (encoded by the Slc9b1 and Slc9b2 genes, respectively), have been identified in mouse sperm (Liu et al., 2010;Chen et al., 2016). Thus, NHA1 and NHA2 are also candidates to mediate the ZP-induced alkalization by Na + /H + exchange.
We first probed the role of Na + /H + exchange in the ZP-induced alkalization using Na + substitution and pharmacology. Indeed, the ZP-induced pH i response was abolished by substitution of extracellular Na + by NMDG (N-methyl-D-glucamine) or addition of EIPA, a commonly used non-selective inhibitor of Na + /H + exchangers . The NH 4 Cl-induced pH i response was, however, similar in the absence or presence of Na + or EIPA (Figures 1I-K). These results confirm that the ZP-induced pH i response depends on Na + /H + exchange.
We examined if ZP-induced Na + /H + exchange is mediated by sNHE using Slc9c1 knockout-mice (Slc9c1-KO) (Wang et al., 2003). In Slc9c1-KO sperm, the ZP-induced pH i and Ca 2+ responses were abolished, whereas the pH i and Ca 2+ response evoked by NH 4 Cl and K8.6, respectively, was preserved (Figures 2A,B). However, sNHE interacts with the soluble adenylyl cyclase sAC, encoded by Adcy10 (Wang et al., 2007), which constitutes the principal source of cAMP in mammalian sperm (Esposito et al., 2004;Hess et al., 2005;Xie et al., 2006); Slc9c1-KO sperm lack functional sAC and, therefore, cAMP synthesis (Wang et al., 2007). We wondered whether the failure of ZP proteins to increase [Ca 2+ ] i and pH i in Slc9c1-KO is due to the lack of sNHE, sAC, or both. To test this, we used optogenetics and the membrane-permeable cAMP analog db-cAMP to rescue intracellular cAMP levels. Transgenic expression of the photoactivated adenylyl cyclase bPAC in Slc9c1-KO sperm provides a tool to stimulate cAMP synthesis in a light-dependent manner (Jansen et al., 2015). Light-stimulated cAMP synthesis in Slc9c1-KO/bPAC sperm or incubation of Slc9c1-KO sperm in db-cAMP both restored the ZP-induced pH i and Ca 2+ response (Figures 2C-F). Thus, the Na + /H + exchange stimulated by ZP proteins does not require sNHE, but rather cAMP synthesis by sAC. Using sperm that express the FIGURE 2 | ZP-induced pH i and Ca 2+ responses in sperm lacking sNHE. Intracellular pH (A) and Ca 2+ (B) responses in Slc9c1-KO sperm mean ± 95% confidence interval (n = 4). (C) Representative pH i responses evoked by 1 ZP/µl in wild-type, Slc9c1-KO, and light-stimulated Slc9c1-KO/bPAC sperm. (D) Representative Ca 2+ responses evoked by mixing with 1 ZP/µl in wild-type, Slc9c1-KO, and light-stimulated Slc9c1-KO/bPAC sperm. (E) Amplitudes of pH i and (F) relative amplitudes of Ca 2+ responses in wild-type, Slc9c1-KO, and Slc9c1-KO sperm pre-incubated with 5 mM db-cAMP (mean ± SD, n = 4). (G) Changes in the cerulean/citrine FRET ratio in mlCNBD-FRET sperm evoked by mixing with 50 mM NaHCO 3 or 1 ZP/µl. An increase of the FRET ratio indicates an increase of free intracellular cAMP; mean ± SD, n = 5. Statistical comparison between multiple groups was performed using one-way ANOVA with Dunnett's correction; p-values are indicated.
Frontiers in Cell and Developmental Biology | www.frontiersin.org FRET-based cAMP biosensor mlCNBD-FRET (Mukherjee et al., 2016), we investigated whether ZP proteins enhance sAC activity and, thereby, control intracellular cAMP synthesis. Mixing of mlCNBD-FRET sperm with ZP proteins did not increase cAMP levels, whereas activation of sAC by using 25 mM NaHCO 3 as a control evoked a pronounced cAMP increase ( Figure 2G), demonstrating that ZP proteins do not control cAMP levels in sperm. Altogether, we conclude that the ZP-induced Na + /H + exchange is not mediated by sNHE, but requires cAMP.
Frontiers in Cell and Developmental Biology | www.frontiersin.org flagellum of wild-type, but not of Slc9b1-KO sperm ( Figure 3E) and, as control, HEK293 cells heterologously expressing mouse NHA1 ( Figure 3F). Altogether, these results confirm that NHA1 is expressed in wild-type, but not in Slc9b1-KO mouse sperm (Liu et al., 2010;Chen et al., 2016).
We investigated the action of ZP proteins on Slc9b1-KO sperm. The ZP-induced pH i and Ca 2+ responses were strongly attenuated, whereas the pH i and Ca 2+ responses evoked by NH 4 Cl or K8.6, respectively, were similar in wild-type and Slc9b1-KO (Figures 3G-J). These results suggest that the ZP-induced alkalization depends predominantly on Na + /H + exchange via NHA1. However, genetic ablation of NHA1 reportedly affects sAC expression and, thereby, intracellular cAMP levels (Chen et al., 2016). Thus, the phenotype of Slc9b1-KO sperm might be caused by impaired or abolished cAMP synthesis, resembling the Slc9c1-KO phenotype (Wang et al., 2003). Yet, in contrast to Slc9c1-KO sperm, db-cAMP did not restore the ZP-induced pH i response in Slc9b1-KO sperm ( Figure 3K). These results indicate that lack of NHA1, rather than impaired cAMP synthesis, underlies the attenuated ZPinduced pH i response in Slc9b1-KO sperm. We surmise that the residual Na + /H + exchange in Slc9b1-KO mice is either mediated by NHA2, by another unknown Na + /H + exchanger, or a combination of both.

Genetic Ablation of NHA1 Affects Sperm Motility
Considering that the ZP-induced alkalization is predominantly mediated by NHA1, we analyzed the phenotype of Slc9b1-KO mice in more detail. In line with previous results, the fertility of Slc9b1-KO males was severely impaired (Chen et al., 2016): only 2 out of 21 matings (Slc9b1-KO males with wildtype females) produced offspring (Supplementary Table S2). The testis and epididymis weight and average sperm count were similar in wild-type and Slc9b1-KO mice ( Supplementary  Table S2), yet Slc9b1-KO sperm largely failed to fertilize oocytes in vitro (Supplementary Table S2), which might be attributed to the severely impaired ZP-induced Ca 2+ and pH i signaling. In previous studies, the number of sperm cells that were motile was significantly reduced in Slc9b1-KO mice (Chen et al., 2016). We observed a similar phenotype: only 55 ± 5% of Slc9b1-KO sperm showed progressive motility compared to 86 ± 5% of wild-type sperm. The decrease in motility has been attributed to a reduced sAC protein expression level (Chen et al., 2016). Thus, we investigated sAC function in wild-type versus Slc9b1-KO sperm. To this end, we studied the flagellar beat frequency (Hansen et al., 2018), which is controlled by sAC: activation of sAC by NaHCO 3 rapidly increases intracellular cAMP levels and the flagellar beat frequency. In sperm that lack sAC, cAMP synthesis and the action of NaHCO 3 is abolished, rendering the sperm immotile (Wennemuth et al., 2003;Esposito et al., 2004;Hess et al., 2005;Xie et al., 2006;Carlson et al., 2007). To investigate whether sAC dysfunction underlies the defect in sperm motility in Slc9b1-KO sperm, we compared the basal flagellar beat frequency, determined at about 60 µm distance from the center of the sperm head, between wild-type and Slc9b1-KO sperm. Under basal conditions, the beat frequency was similar (WT: 11 ± 3 and KO: 13 ± 3 Hz, n ≥ 13), and stimulation with 25 mM NaHCO 3 increased the frequency to a similar extent (WT: 20 ± 6% and KO: 22 ± 9%, n ≥ 13). Thus, the lack of NHA1 does not impair the sAC-control of flagellar beat frequency, suggesting that cAMP synthesis is not impaired.
When analyzing the flagellar beat in detail, we noticed that the beat frequency along the flagellum was not uniform in Slc9b1-KO (Figures 4A,B and Supplementary Movies S1-S2). At ≤ 80 µm distance from the sperm head, the flagellar beat frequency was similar in Slc9b1-KO and wild-type sperm, whereas at > 80 µm, the frequency was considerably faster in Slc9b1-KO sperm ( Figure 4B). Strikingly, Slc9b1-KO sperm displayed a stiff midpiece, which prevented to reliably determine the beat frequency at the first 20 µm of the flagellum. To describe this defect in quantitative terms, we compared the amplitudes of the curvature angle along the flagellum as a measure for the beat amplitude ( Figure 4C) (Hansen et al., 2018). In the midpiece, the amplitude was lower in Slc9b1-KO sperm compared to wild-type sperm, reflecting the restricted movement ( Figure 4C). Farther along the flagellum, the amplitude was similar between wild-type and Slc9b1-KO sperm ( Figure 4C).
A stiff midpiece has also been observed in sperm lacking calcineurin or the CatSper-channel subunit CatSper ζ (Miyata et al., 2015;Chung et al., 2017). The CatSper-channel complex forms Ca 2+ signaling domains along the flagellum that are organized in four longitudinal columns (Chung et al., 2014). Loss of the channel complex disrupts this organization (Chung et al., 2014(Chung et al., , 2017. To examine whether a defect in the Ca 2+ signaling domains might underlie the stiff midpiece in Slc9b1-KO sperm, we used super-resolution microscopy (3D-STORM). In wild-type and Slc9b1-KO sperm, an anti-CatSper1 antibody labeled four distinct columns aligned longitudinally along the flagellum, as previously described for wild-type sperm using the same antibody (Chung et al., 2014), demonstrating that the Ca 2+ -signaling domains are preserved in Slc9b1-KO sperm ( Figure 4D). Furthermore, the density distribution of CatSper1 labeling along the flagellum was not different between wild-type and Slc9b1-KO sperm (Supplementary Figure S3). Thus, the molecular mechanism underlying the motility defect of Slc9b1-KO sperm remains unclear, but we propose that a combination of defective ZP signaling and altered flagellar beat pattern in Slc9b1-KO sperm underlies the sub-fertility of male Slc9b1-KO mice.

ZP-Induced Signaling Requires a Sufficiently Negative Membrane Potential
Although it is unknown how ZP proteins activate Na + /H + exchange via NHA1, the membrane potential seems to be an important factor (Zeng et al., 1995;Arnoult et al., 1999;De La Vega-Beltran et al., 2012;Chavez et al., 2014). Thus, we tested the role of the V m in ZP signaling. In wild-type sperm, that were depolarized by incubation in high extracellular potassium ([K + ] o = 138 mM), the ZP-induced pH i and Ca 2+ responses were abolished, whereas the pH i and Ca 2+ responses evoked by NH 4 Cl or K8.6, respectively, were preserved (Figures 5A-D). This suggests that the ZP-induced pH i increase requires a more negative V m . Slo3 and its auxiliary subunit Lrrc52 form the principal K + channel in mouse (Santi et al., 2010;Yang et al., 2011;Zeng et al., 2011Zeng et al., , 2015 and human sperm (Brenker et al., 2014). We studied the action of ZP proteins on Kcnu1and Lrrc52-KO sperm, which both feature a depolarized V m (Santi et al., 2010;Yang et al., 2011;Zeng et al., 2011Zeng et al., , 2015. In Kcnu1-KO and Lrrc52-KO sperm, the ZP-evoked pH i and Ca 2+ responses were abolished, whereas the pH i and Ca 2+ responses evoked by NH 4 Cl and K8.6, respectively, were preserved (Figures 5E,F,I-L). Remarkably, hyperpolarizing Kcnu1-and Lrrc52-KO sperm using the K + ionophore valinomycin (Santi et al., 2010;Chavez et al., 2014) restored the ZP-induced pH i and Ca 2+ responses (Figures 5G-I,L), supporting the notion that ZP signaling requires a sufficiently negative V m . Altogether, these results corroborate that a negative V m , set by Slo3, enables the ZP-induced pH i increase and, thereby, ZP-induced Ca 2+ influx via CatSper. However, the mechanism underlying the V m -control of NHA1 remains to be elucidated.

DISCUSSION
The function of mammalian sperm is controlled by external cues that engage various signaling molecules. How these molecules are integrated into signaling pathways is not well-understood. Here, we show that the synthesis of cAMP and a sufficiently negative membrane potential prime mouse sperm to transduce binding of ZP proteins into rapid H + and Ca 2+ signaling events (Figure 6). This ZP-induced Ca 2+ increase might be involved in the control of swimming behavior and acrosomal exocytosis.
The CatSper channel is the principal pathway for Ca 2+ entry into mammalian sperm (Quill et al., 2001;Ren et al., 2001;Kirichok et al., 2006;Lishko and Kirichok, 2010). We propose that the action of ZP proteins on CatSper is indirect, mediated by a pH i -signaling pathway (Figure 6): any experimental condition that abolishes the ZP-induced pH i increase abolishes the Ca 2+ influx, and vice versa, conditions that restore the pH i response also restore the Ca 2+ influx. This result is consistent with earlier reports, demonstrating a ZP-evoked increase of pH i and [Ca 2+ ] i in mouse sperm (Bailey and Storey, 1994;Murase and Roldan, 1996). In patch-clamp experiments, ZP proteins did not enhance monovalent CatSper currents recorded from sperm isolated from the corpus of the ductus epididymis (Xia and Ren, 2009). Yet, sperm from this region of the epididymis are unable to undergo capacitation (Yanagimachi, 1994), which potentiates the Ca 2+ response (see Figure 1D). Thus, the use of non-capacitated sperm might have hampered the detection of a direct ZP protein action on CatSper. Nevertheless, these results support our model that in mouse sperm, ZP proteins activate CatSper via intracellular alkalization (Figure 6). The ZP-induced pH i increase rests on FIGURE 6 | Model of the ZP-signaling pathway in mouse sperm. ZP glycoproteins bind to a yet unknown receptor (recognition), which initiates a signaling cascade. This cascade requires a hyperpolarized membrane potential (V m ), set by the Slo3/Lrrc52-channel complex, and cAMP synthesis by sAC (priming). Thereby, a pH i response, resting on Na + /H + exchange by NHA1, and probably NHA2, is enabled that promotes Ca 2+ influx through CatSper (signaling). Shaded objects indicate that the signaling mechanism and function of the molecules has not been established yet.
Na + /H + exchange that is predominantly mediated by NHA1 (Figure 6). The residual pH i response in Slc9b1-KO might be carried by NHA2; future work needs to address this question using Scl9b1/2 knockout mice. The pH i response is only observed at the negative V m set by Slo3 (Santi et al., 2010;Zeng et al., 2011Zeng et al., , 2015. It remains to be elucidated how the V m dependence is integrated into the ZP-signaling pathway. In mouse sperm, K + currents and, thus, V m are strongly controlled by pH i (Navarro et al., 2007;Zeng et al., 2011Zeng et al., , 2013. The control of the ZP-induced pH i response by V m , the control of V m by Ca 2+ and pH i , and the interplay of CatSper and Slo3 during ZP signaling deserve further studies. We identified cAMP as a key player in the ZP-signaling pathway in mouse sperm: basal cAMP synthesis by sAC is required for the pH i response and ensuing CatSper activation. sNHE, which might form a physical or functional complex with sAC, is however not responsible for the ZP-induced alkalization. It has been proposed that NHA1 and NHA2 control sAC expression and, thereby, cAMP synthesis and motility (Chen et al., 2016). We also find Slc9b1-KO males are infertile and feature a significantly reduced number of motile sperm. However, we show that sAC function is unaffected in NHA1-KO sperm as the control of the flagellar beat frequency by HCO − 3 remains unchanged. Thus, the pathomechanism underlying the reduced number of motile sperm in Slc9b1-KO males remains to be elucidated. Yet, the loss of NHA1 alters the flagellar beat pattern along the flagellum with the midpiece being stiff and the rest of the flagellum being more flexible. Although the ZPinduced alkalization via NHA1 activates CatSper, the exchanger is not required for the organization of CatSper in quadrilateral columns. Future studies are required to elucidate whether NHA1 is part of these columns to form nanodomains of pH and Ca 2+ signaling. Whether the loss of NHA1 affects other downstream processes that control the flagellar beat pattern or if the axonemal structure/dynein function is altered is not known and warrants further studies. Sperm must undergo the acrosome reaction to penetrate through the ZP. Across species, binding of sperm to the native ZP, to isolated ZPs, as well as to solubilized ZP proteins evokes the acrosome reaction in sperm (Florman and Storey, 1982;Bleil and Wassarman, 1983;Cherr et al., 1986;O'Rand and Fisher, 1987;Cross et al., 1988;Florman and First, 1988;Uto et al., 1988;Arnoult et al., 1996a;Schroer et al., 2000;Tollner et al., 2003;Gupta et al., 2012). This suggested that in vivo, sperm undergo the acrosome reaction primarily upon binding to the ZP (Hirohashi and Yanagimachi, 2018). Recent studies utilizied fluorescentlylabeled sperm from transgenic mouse models to determine the acrosomal status and acrosome reaction in live mouse sperm upon fertilization in vitro and ex vivo within the oviduct. These experiments revealed that at least in mice, most sperm undergo the acrosomal exocytosis during their journey across the oviduct or within the cumulus cell-matrix rather than upon binding to the zona pellucida (Jin et al., 2011;Hino et al., 2016;La Spina et al., 2016;Muro et al., 2016). Moreover, acrosome-reacted mouse sperm seem to reach the surface of the oocyte more readily than acrosome-intact sperm (Hildebrand et al., 2010), and mouse sperm can bind to the ZP without undergoing the acrosome reaction (Baibakov et al., 2007). Thus, the significance of the acrosomal exocytosis and signaling events evoked by binding of sperm to the ZP for fertilization in mice or other mammalian species still remain to be elucidated.
In summary, our study provides new insights and at the same time, raises new questions about the action of ZP proteins on mammalian sperm. The identification of the ZP receptor(s) and binding site(s) on sperm is certainly essential to delineate the whole ZP-signaling pathway. However, this question has remained unanswered since the initial characterization of ZPsperm interaction (Bleil and Wassarman, 1980). Our results present NHA1, sAC, and Slo3/Lrrc52 as new players in the sperm ZP-signaling pathway. This might be the basis for future approaches, unraveling the molecular mechanisms underlying infertility and the design of new contraceptives.

Nomenclature
For most of the proteins, we have used the alias in the text. The official nomenclature is indicated in Table 1.

Sperm Preparation
Mouse sperm were isolated as previously described (Raju et al., 2015). In brief, sperm were isolated by incision of the cauda epididymis followed by a swim-out in modified TYH medium (in mM: 135 NaCl, 4.8 KCl, 2 CaCl 2 , 1.2 KH 2 PO 4 , 1 MgSO 4 , 5.6 glucose, 0.5 sodium pyruvate, 10 lactic acid, 10 HEPES, pH 7.4 adjusted at 37 • C with NaOH). After 15-30 min swim-out at 37 • C, sperm were collected and counted. For capacitation, sperm were incubated for 90 min in TYH containing 3 mg/ml BSA and 25 mM NaHCO 3 substituting 25 mM of the NaCl; the pH was adjusted to 7.4. Measurements under depolarized membrane potential were performed in TYH with 135 mM KCl and 5 mM NaCl at pH 7.4. For experiments in the absence of extracellular Na + , Na + was substituted by N-methyl-Dglucamine (M2004, Sigma-Aldrich) and the pH was adjusted with HCl. Valinomycin and db-cAMP were added after swim-out and were present throughout the experiment. Prm1-bPAC/Slc9c1-KO sperm were isolated under dim red light. Light-activation of Prm1-bPAC/Slc9c1-KO sperm was performed in a custommade rack equipped with blue LEDs during sperm capacitation. Experiments were performed with capacitated sperm unless otherwise indicated. The cAMP analogs (db-cAMP) were already added during the capacitation phase for 90 min.

Isolation of Mouse Zona pellucidae
For ZP isolation, wild-type female mice were super-ovulated by intraperitoneal injection of 10 I.U. hCG (human Chorionic Gonadotropin; ProSpec, Rehovot, Israel) 3 days before the experiment. 14 h before oocyte isolation, mice were injected with 10 I.U. PMSG (Pregnant Mare's Serum Gonadotropin; ProSpec). Mice were killed by cervical dislocation and oviducts were dissected. Cumulus-enclosed oocytes were prepared from the oviducts in TYH buffer containing 300 µg/ml hyaluronidase (Sigma). After 15 min, cumulus-free oocytes were transferred into fresh buffer and washed twice. Zonae pellucidae and oocytes were separated by shear forces generated by expulsion from 50 nm pasteur pipettes. Zona pellucidae were counted, transferred into fresh buffer, diluted to a concentration of 1 ZP per ul, and solubilized by incubation at 75 • C for 15 min (Thaler and Cardullo, 1996). Animal experiments were performed in accordance with the relevant guidelines and regulations and approved by the local authorities (LANUV) AZ84-02.05.40.13.127.

Heterologous Expression of ZP Glycoproteins
The cDNA sequence of mZP1, mZP2, and mZP3 was amplified via PCR. A hexa-histidine tag was inserted upstream of the conserved furin cleavage site, an AgeI restriction site was added to the 5 end, and a Kpnl restriction site to the 3 end by nested PCR. The PCR product was cloned into a pHLsec vector (kindly provided by Prof. Yvonne Jones, Wellcome Trust Center for Human Genetics, University of Oxford, United Kingdom) using AgeI and Kpnl. pHLsec-mZP1, pHLsec-mZP2 and pHLsec-mZP3 were transiently transfected in HEK293T cells (ATCC-CRL-3216) using polyethyleneimine (Life Technologies, Carlsbad, United States).

STORM Imaging and Analysis of Sperm Flagellar Proteins
STORM imaging experiments were performed in an imaging buffer (50 mM Tris, pH 8, 10 mM NaCl) with an oxygen scavenging system (0.5 mg/mL glucose oxidase, 40 µg/ml catalase, 10% glucose, and 10 mM 2-aminoethanethiol). 10.000-60.000 frames were acquired per data set using 647 nm excitation at 100 mW. A 405 nm laser was used to maintain an adequate number of localizations per frame. A cylindrical lens was introduced in the detection path for astigmatism 3D STORM acquisition. Perfect focus system from Nikon was used to minimize axial drift and a vibration isolation table was used to minimize lateral drift. STORM movies were analyzed as described previously using the Nikon software package based on a technology developed by Dr. Xiaowei Zhuang (Huang et al., 2008). Briefly, fluorescence peaks corresponding to individual molecules were identified in each frame and fit using least-squares or maximum-likelihood estimation with a twodimensional Gaussian to determine the (x,y) position of each molecule. For 3D imaging, the ellipticity of the Gaussian fit was used to assign a z coordinate. Drift correction was applied using cross-correlation.
STORM images were rendered with each localization plotted as a Gaussian. Images were filtered to reject molecules with low photon number (below 500 photons). Molecules with aspect ratio higher than 1.5 for 2D and 2.5 for 3D datasets were rejected. Moreover, molecules that appear for > 10 consecutive frames were rejected. Non-specifically bound antibodies can give background in the STORM images, which appears as scattered localizations with low local densities. This background noise was removed by a local density filter. Low-density localizations were filtered out by removing a molecule if it was surrounded by fewer than 10 localizations in the 80 × 80 nm region.

Beta-Galactosidase Staining of Testis Sections
Testis of adult male mice were isolated, punctured twice with a cannula and incubated overnight at RT in 4% PFA. After a single washing step in PBS for 10 min, testis were transferred into a 10% sucrose solution for 1 h and subsequently incubated overnight in 30% sucrose. On the next day, testis were embedded in Tissue TEK (Sakura Finetek) and stored at −80 • C. The testis was sectioned in 16 µm thick cross-sections using a 2800 Frigocut-E cryostat (Reichert-Jung, Nußloch)

Measurement of Changes in Intracellular Ca 2+ and pH in Mouse Sperm
Changes in [Ca 2+ ] i and pH i in mouse sperm were recorded in a rapid-mixing device in the stopped-flow mode (SFM400; Bio-Logic, Claix, France) after loading with the fluorescent Ca 2+ indicator Cal-520-AM (AAT Bioquest, Sunnyvale, United States) or the fluorescent pH indicator BCECF-AM (Thermo Fisher), respectively. Changes in [Ca 2+ ] i were measured as previously described (Strünker et al., 2011) with minor modifications. In brief, sperm were loaded with Cal-520-AM (5 µM) in the presence of Pluronic F-127 (0.02% v/v) at 37 • C for 45 min. After incubation, excess dye was removed by three centrifugation steps (700 g, 7 min, RT). The pellet was resuspended in buffer and equilibrated for 5 min at 37 • C. The sperm suspension (5 × 10 6 sperm/ml) was rapidly mixed 1:1 (v/v, 100:100 µl) with the respective stimulant [ZP, K8.6, 2 µM ionomycin (Tocris)] at a flow rate of 0.5 ml/s. Fluorescence was excited by a SpectraX Light Engine (Lumencor, Beaverton, United States), whose intensity was modulated with a frequency of 10 kHz. The excitation light was passed through a BrightLine 475/28 nm filter (Semrock, Rochester, United States) onto the cuvette. Emission light was passed through a BrightLine 536/40 filter (Semrock) and recorded by photomultiplier modules (H10723-20; Hamamatsu Photonics). The signal was amplified and filtered through a lockin amplifier (7230 DualPhase; Ametek, Paoli, United States). Data acquisition was performed with a data acquisition pad (PCI-6221; National Instruments, Austin, United States) and Bio-Kine software v. 4.49 (Bio-Logic). Ca 2+ signal traces are depicted as the percent change in fluorescence ( F) with respect to the mean of the first three data points recorded immediately after mixing (F 0 ). Mean ± 95% CI are shown to visualize the true range of the data. The control F/F 0 signal observed upon mixing with buffer (control) was subtracted from K8 6-, ZP-, or ionomycin-induced signals, setting the control-signal level to F/F 0 (%) = 0. The K8.6 solution (Babcock and Pfeiffer, 1987) contained (in mM: 4.8 NaCl, 138 KCl, 2 CaCl 2 , 1.2 KH 2 PO 4 , 1 MgSO 4 , 5.6 glucose, 0.5 sodium pyruvate, 10 lactic acid, 10 HEPES, pH 8.6, adjusted with KOH) to depolarize the V m and simultaneously increase pH i to activate CatSper. Bar graphs show the maximal amplitude of the ZP-or K8.6-evoked Ca 2+ response (average of last three data points), normalized to the respective ionomycin-evoked Ca 2+ response (average of last three data points) (relative signal amplitude). To measure pH changes in the stopped-flow mode, sperm were loaded with BCECF-AM (10 µM) at 37 • C for 15 min. The pellet was resuspended in TYH and equilibrated for 5 min at 37 • C. The excitation light was passed through a BrightLine 452/45 nm filter (Semrock) onto the cuvette. Emission light was passed in parallel through a BrightLine 494/20 filter and a BrightLine 540/10 filter (Semrock). pH signals are depicted as the percent change in fluorescence ratio ( R) of 494 nm/540 nm with respect to the mean of the first three data points recorded immediately after mixing (R 0 ) when a stable fluorescence signal was observed. The R/R 0 signal evoked by mixing with buffer (control) was subtracted from ZP-or NH 4 Cl-induced signals. Bar graphs show the maximal amplitude of the ZP-or NH 4 Cl-evoked pH response (average of last three data points).

In vitro Fertilization
Superovulation in females was induced as described above. HTF medium (EmbryoMax Human Tubal Fluid; Merck Millipore) was mixed 1:1 with mineral oil (Sigma-Aldrich) and equilibrated overnight at 37 • C. Sperm were capacitated for 90 min in TYH medium supplemented as indicated above. On the day of preparation, 100 µl drops of HTF were covered with the medium/oil mixture and 10 5 sperm were added to each drop. Cumulus-enclosed oocytes were prepared from the oviducts of superovulated females and added to the drops. After 4 h at 37 • C and 5% CO 2 , oocytes were transferred to fresh HTF. The number of 2-cell stages was evaluated after 24 h.

Sperm Motility Analysis
Freely beating sperm were observed in shallow perfusion chambers with 200 µm depth, which allowed to exchange the buffer during recordings. Sperm were tethered to the glass surface by lowering the BSA to 0.3 mg/ml. An inverted dark-field video microscope (IX71; Olympus) with a 10 x objective (UPlanFL, NA 0.4; Olympus) and an additional 1.6 x magnification lens (16x final amplification) was combined with a high-speed camera (Dimax; PCO). Dark-field videos were recorded with a frame rate of 200 Hz. The temperature was 37 • C (Incubator; Life Imaging Services). The flagellar beat was analyzed using SpermQ (Hansen et al., 2018). SpermQ outputs the parameter curvature angle as a measure for flagellar bending. The curvature angle at a given point on the flagellum was determined by the angle between the tangential vector at the given point and the tangential vector at the point 10 µm proximal on the flagellum. The beat frequency at a given point on the flagellum was determined by the highest peak in the frequency spectrum obtained by Fast-Fourier-Transformation of the time-course of the parameter curvature angle at the given flagellar point. The amplitude of the curvature angle for a given point was determined as the absolute difference between the median of the five highest and the median of the five lowest curvature angle values in the entire time-course at the given point.

Identification of NHA1 by Mass Spectrometry
Sperm were isolated from wild-type C57Bl/6 mice and subjected to mass spectrometry as described previously (Raju et al., 2015).

Statistical Analysis
Statistical analyses for graphs shown in figures has been performed in GraphPad Prism. Statistical significance between two groups was determined using two-tailed, unpaired t-test with Welch's correction, statistical significance between multiple groups was determined using one-way ANOVA with Dunnett's correction. The respective details are indicated in the figure legends.

DATA AVAILABILITY STATEMENT
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

AUTHOR CONTRIBUTIONS
MB established and performed the pH i and Ca 2+ fluorimetry, prepared the native ZPs, and generated and analyzed the NHA1-KO. CS also performed pH i and Ca 2+ fluorimetry. HH performed STORM imaging. JJ and JH analyzed sperm motility.
CT performed the mass spectrometry. PN performed genotyping of knockout mice. HF provided protocols for ZP isolation and analysis. HF, UK, CB, LJ, and LH analyzed and/or interpreted data, participated in drafting the manuscript, and revised the manuscript critically for important intellectual content. TS and DW conceived the project, designed and coordinated the experiments, analyzed and/or interpreted data, and wrote the manuscript. All authors contributed to the article and approved the submitted version.