FAK regulates actin polymerization during sperm capacitation via the ERK2/GEF-H1/RhoA signaling pathway

Freshly ejaculated sperm cells are unable to fertilize the ovule; they reach functional maturity after they reside for a period of time within the female reproductive tract, where they undergo a series of physiological and biochemical changes known as capacitation (Austin, 1952; Chang, 1951). This process comprises increases in their metabolism and membrane fluidity, as well as in the concentration of intracellular Ca²⁺, HCO3⁻ and Cl⁻ and cAMP concentration (Nishigaki et al., 2014; Salicioni et al., 2007; Visconti et al., 1995). Other changes associated with capacitation are membrane hyperpolarization (Arnoult et al., 1999), actin polymerization, actin cytoskeleton remodeling (Brener et al., 2003; Cabello-Agüeros et al., 2003) and acquisition of hypermotility (Yanagimachi, 1994). Once capacitation has occurred, spermatozoa undergo a physiologically induced process called acrosome reaction (AR) and become fertilization competent.

The remodeling of the actin cytoskeleton in mammalian spermatozoa during capacitation (Brener et al., 2003; Cabello-Agüeros et al., 2003) is necessary for the AR to progress normally. The presence of the actin-spectrin cytoskeleton on the cortical surface of the plasma membrane prevents the early fusion of the plasma with the outer acrosomal membranes (Breitbart et al., 2005; Brener et al., 2003; Hernández-González et al., 2000; Spungin et al., 1995). The actin cytoskeleton remodeling is also required for activation of phospholipase C, which is involved in the AR (Spungin et al., 1995); the actin cytoskeleton has an important role in hypermotility (Azamar et al., 2007; Finkelstein et al., 2013; Itach et al., 2012) and participates in the fusion of the plasma membrane of the spermatozoa with that of the ovum (Brener et al., 2003; Cabello-Agüeros et al., 2003; Rogers et al., 1989; Sánchez-Gutiérrez et al., 2002; Sosnik et al., 2009).

Focal adhesion kinase (FAK) is a 125 kDa protein without a known receptor tyrosine kinase that plays a central role in integrin signaling as a protein scaffold and in the assembly of focal adhesions (Franchini, 2012). The cytoplasmic domain of β-integrin is known to interact with the FERM domain of FAK and triggers its activation (Cooper et al., 2003; Schaller et al., 1994). This activation comprises autophosphorylation of FAK at residue Tyr397 that plays a scaffolding role for interaction with Src kinase, which phosphorylates FAK at Tyr576 and/or Tyr577 – phosphorylation processes that potentiate FAK kinase activity. Src also phosphorylates FAK at residue Tyr925, a recruitment site for Grb2-SOS (Hall et al., 2011; Lu and Rounds, 2012), which leads to activation of Ras GTPase and the subsequent activation of the Raf1/MEK1/ERK1 and ERK2 pathway (Renshaw et al., 1999). Interestingly, mitogen-activated protein kinases 3 and 1 (MAPK3 and MAPK1, respectively; hereafter referred to as ERK1/2) are also known to regulate actin polymerization in somatic cells through RhoA because they phosphorylate Rho guanine nucleotide exchange factor 2 (ARHGEF2, hereafter referred to as GEF-H1) at Thr678, which enhances the guanine nucleotide exchange activity of GEF-H1 towards RhoA (Fujishiro et al., 2008; Guilluy et al., 2011).

In mammalian spermatozoa, ERK1/2 are activated in response to different factors (de Lamirande and Gagnon, 2002; Nixon et al., 2010). ERK1/2 and the central elements associated with their activation (Shc, Grb2, Ras, Raf1 and MEK) are present in the heads and flagella of different mammalian sperm cells (Almog et al., 2008; Awda and Buhr, 2010; de Lamirande and Gagnon, 2002; Luconi et al., 1998a,b; Nixon et al., 2010). ERK1/2 are involved in different physiological sperm processes – such as capacitation – by regulating phosphorylation of Tyr, motility (de Lamirande and Gagnon, 2002; Luconi et al., 1998a; Luna et al., 2012; Nixon et al., 2010; O'Flaherty et al., 2005, 2006a,b), Ca²⁺ flux (Jaldety and Breitbart, 2015) and activation of phospholipase A2 (Chen et al., 2005).

The presence of FAK in mammalian spermatozoa has recently been reported in a study showing that this kinase might be associated with the regulation of the AR (Roa-Espitia et al., 2016). Specifically, when FAK is inhibited spermatozoa experienced early phosphorylation of Tyr and, consequently, an anticipated AR. Moreover, the authors report FAK to be important for actin polymerization (Roa-Espitia et al., 2016). The inhibition of FAK during capacitation may have different outcomes, but it clearly prevents actin polymerization associated with capacitation. In addition, a signaling cascade including FAK, Ras and ERK1/2 is related to mechanotransduction, proliferation and differentiation (Provenzano et al., 2009; Salasznyk et al., 2007a,b; Ward and Storey, 1984). Therefore, the present study has the aim to show that FAK regulates actin polymerization via the ERK1 and ERK2/GEF-H1/RhoA signaling cascade. Additionally, it is known that mammalian spermatozoa express three different integrins (α5β1, α6β3 and αvβ3), whose ligand is fibronectin. We propose here that this signaling pathway is activated as a result of the interaction between integrins and fibronectin located on the surface of sperm cells. Therefore, we explore upstream and downstream signaling events associated with FAK, and their role in the regulation of actin polymerization during capacitation.

Proteins from whole-sperm extracts were analyzed by western blot using a specific antibody that recognizes ERK1/2 to confirm the expression of the ERK1/2 protein in guinea pig spermatozoa. The antibody detected the two proteins (relative molecular mass of 44 and 42 kDa) in both capacitated and non-capacitated spermatozoa (Fig. 1A). By means of immunocytochemistry, the anti-ERK1/2 antibody detected both proteins in the apical region of the acrosome and along the flagellum in both capacitated and non-capacitated spermatozoa (Fig. 1B).

An antibody that specifically recognizes the active, i.e. phosphorylated, form of ERK1/2 (p-ERK1/2) was used to define the activation kinetics of ERK1/2 during capacitation. Results obtained by western blot showed that p-ERK1/2 were almost absent in non-capacitated spermatozoa; however, a significant increase of p-ERK2 levels was observed during the first 2 min of capacitation (at 1.5 min), which peaked at 10 min of capacitation (Fig. 1C,D) and then decreased slowly to levels similar to those of non-capacitated spermatozoa (Fig. 1C). It is important to note that p-ERK1 could only be observed at 10 min of capacitation (Fig. 1C), suggesting that ERK2 is the principal active form in guinea pig spermatozoa. Further analysis of these results showed that the levels of p-ERK at 30 and 60 min of capacitation is significantly higher (P≤0.05) than those shown in non-capacitated spermatozoa (at 0 min). These data suggest that the activity of ERK2 within 30–60 min of capacitation is sufficient to activate its AR-related targets.

The above results were corroborated by immunocytochemistry, which revealed a clear increase in fluorescence at 1.5 min of capacitation and maximum fluorescence at 10 min (Fig. S1A). Additionally, p-ERK2 and ERK1/2 were located in the same sites: the apical region of the acrosome and along the flagellum. These results suggest that ERK2 plays a role in both AR and motility.

We used here the new ERK1/2-specific inhibitor FR180204 (Ohori et al., 2005), to determine in guinea pig spermatozoa the role of ERK2 in sperm capacitation and the AR. For this purpose, spermatozoa were capacitated in the absence or presence of different concentrations of FR180204 (0–15 μM); we found a concentration-dependent reduction of the AR, with maximum reduction at 10 μM (Fig. S1B). Capacitation and AR were assessed by using the chlortetracycline (CTC) fluorescence assay and previously reported patterns therein, i.e. F, B and AR (see Materials and Methods for further details) (Maldonado-García et al., 2017; Roa-Espitia et al., 2016) (Fig. S1C). In respect to non-capacitated spermatozoa after 90 min of capacitation, this assessment showed a significant increase in the number of B patterns (indicating capacitated acrosome-intact cells) and in AR patterns (indicating physiologically capacitated acrosome-reacted cells) (Fig. 2A). When spermatozoa were capacitated in the presence of FR180204 (10 μM), we observed a significant increase in the number of B but not the AR patterns compared with capacitated spermatozoa in the absence of FR180204 (Fig. 2A). Additionally, to determine whether FR180204 also inhibits progesterone-induced AR, spermatozoa were capacitated in the absence or presence of FR180204 (10 μM) for 90 min. Then, the AR was induced by incubating the spermatozoa in the presence of progesterone (10 μM, 20 min). The results showed that the progesterone-induced AR is significantly increased compared with that in spermatozoa not incubated with progesterone (Fig. 2A). Progesterone-induced AR in spermatozoa capacitated in the presence of FR180204 was significantly decreased compared with spermatozoa capacitated in the absence of FR180204 (Fig. 2A). These data suggest that the inhibition of ERK1/2 inhibits spontaneous-AR and physiological-induced AR but not capacitation.

To verify that FR180204 does not inhibit capacitation, we evaluated the distribution of the membrane microdomain marker flotillin-2 (Flot2), since it had been proposed that a hallmark event occurring during capacitation is the redistribution of this marker (Gadella and Boerke, 2016; Maldonado-García et al., 2017; van Gestel et al., 2005). By using immunocytochemistry and a specific antibody that recognizes flotillin-2 (anti-Flot2), we observed in spermatozoa with intact acrosome two types of Flot2 immunolabeling pattern in the sperm head: a dispersed pattern in the acrosome, which is a characteristic of non-capacitated cells (P1) and an apical ridge re-ordered pattern, which is a characteristic of capacitated cells (P2) (Fig. S1D). The spermatozoa capacitated (60 min) in presence or absence of the FR180204 inhibitor showed a significant increase in the number of pattern 2 (P2) of Flot2 compared with those of non-capacitated spermatozoa (Fig. 2B).

We also analyzed the effect of FR180204 on Tyr phosphorylation of proteins (p-Tyr) to corroborate whether ERK1/2 inhibition affects capacitation. The western blot analysis showed a significant increase of p-Tyr at 30 and 60 min of capacitation (Fig. 2C,D). This increase was also observed in spermatozoa capacitated in the presence of FR180204 (10 μM), and was significantly inhibited when spermatozoa were capacitated in the presence of a PKA inhibitor, Rp-cAMPS (Fig. 2C,D). To confirm these results, we inhibited MAP2K1 and MAPK2 (also known as MEK1 and MEK2, respectively; hereafter referred to as MEK1/2) with PD98059, as MEK1/2 activate ERK1/2 through phosphorylation at Tyr204 and Thr202 (Roskoski, 2019). Spermatozoa capacitated in the presence of PD98059 (50 µM) showed a strong decrease in the amount of phosphorylated ERK1/2 (Fig. S1F,G); however, inhibition of MEK1/2 and ERK1/2 only produced a non-significant decrease of p-Tyr phosphorylated proteins (Fig. 2C,D). We also tested spermatozoa capacitated in the presence of a higher (100 µM) concentration of PD98059 to determine whether it was able to significantly inhibit p-Tyr. We found that, a few minutes after the start of incubation, spermatozoa experienced loss of acrosome and immobilization, which indicates that this concentration of PD98059 is very toxic for guinea pig spermatozoa.

Actin cytoskeleton remodeling is one of the main events during capacitation (Brener et al., 2003; Roa-Espitia et al., 2016). Therefore, we wanted to explore whether inhibition of ERK1/2 affects polymerization of actin. We did this by using the ERK1/2-specific inhibitor FR180204 as well as the selective MEK1/2 inhibitor PD98059 that results in indirect inhibition of ERK1/2 (Fig. S1F,G). To accomplish this goal, spermatozoa were stained with phalloidin-FITC, which specifically recognizes F-actin, and the emitted fluorescence was quantified (see Materials and Methods). During capacitation, the amount of F-actin increased steadily until 60 min of incubation; after this time, the amount decreased slightly (Fig. 3A,B). When F-actin was assessed in capacitated spermatozoa in the presence of FR180204 (10 μM) or PD98059 (50 μM), its amount increased steadily until 30 min. However, control levels were not reached because the amount of F-actin was significantly lower at 60 and 90 min (Fig. 3B). These results suggest that ERK1/2 is involved in the modulation of actin polymerization during capacitation.

Further analysis of these results showed that the F-actin levels reached in the presence of inhibitors (FR180204 or PD98059) at 60 and 90 min of capacitation is ∼50% less compared with those in spermatozoa capacitated in the absence of the inhibitors. This amount of F-actin is significantly higher (P≤0.01) than that in non-capacitated spermatozoa (at 0 min), which suggests that this amount of F-actin is sufficient for capacitation to be carried out successfully but is insufficient to accomplish the AR.

We have recently reported that the focal adhesion kinase (FAK) of 125 kDa modulates actin polymerization during capacitation and that inhibition of FAK by PF573228 inhibits this process (Roa-Espitia et al., 2016). We now suggest that the signaling pathway FAK/Grb2/SOS/Ras/MEK regulates activation of ERK1/2. To corroborate this hypothesis, we first analyzed the effects of FAK inhibition on ERK2 activation. When ERK2 reached maximum activation at 10 min, spermatozoa were capacitated for 10 min in the absence and presence of PF573228 (5 μM) or the absence and presence of FAK inhibitor 14 (5 μM). Immunocytochemistry with anti-p-ERK1/2 antibody showed fluorescent areas within the apical region of the acrosome and the flagellum of capacitated spermatozoa (Fig. S1E); sperm cells capacitated in the presence of PF573228 showed less fluorescence (Fig. S1E). To confirm these results, p-ERK1/2 levels were analyzed by western blotting of complete protein extracts. p-ERK2 was detected in extracts of sperm capacitated for 10 min, but was only detected at low levels in extracts of non-capacitated sperm cells and extracts of sperm cells capacitated in the presence of PF573228 or FAK inhibitor 14 (5 μM) (Fig. 4A,B).

It has been postulated that phosphorylation of FAK at Tyr925 is essential for its association with Grb2 and to modulate the activation of ERK1/2 (Hall et al., 2011). Therefore, the phosphorylation status of FAK was analyzed by western blotting using antibody against FAK phosphorylated at Tyr925 (p-FAK-Tyr925). Basal levels of p-FAK-Tyr925 were found in non-capacitated spermatozoa (Fig. 4C, top panel); levels of p-FAK-Tyr925 increased significantly in the first minutes of capacitation and within 30 min, decreasing thereafter (Fig. 4C,D). These results suggest that, during capacitation, FAK is associated with Grb2 and modulates activation of ERK1/2.

Grb2 is a scaffold protein that enables the activation of ERK1/2 via the canonical signaling pathway; therefore, the interaction of the Grb2 protein with FAK was determined in order to confirm participation of FAK in the activation of ERK2. FAK was immunoprecipitated, and the western blot analysis revealed that Grb2 coimmunoprecipitated together with FAK (Fig. 4E). A densitometric analysis showed that significantly more Grb2 coimmunoprecipitates with FAK in capacitated spermatozoa compared with non-capacitated spermatozoa; such increase was inhibited by PF573228 (Fig. 4E,F). Additionally, reciprocal assays were carried out and similar results were obtained; significantly more FAK coimmunoprecipitated together with Grb2 in capacitated spermatozoa compared with non-capacitated spermatozoa or capacitated in the presence of PF573228 (Fig. 4G,H). As a control, immunoprecipitation was performed in the presence of a non-specific IgG. Neither FAK nor Grb2 was coimmunoprecipitated. These results confirmed that, during capacitation, FAK is involved in the activation of ERK2.

Confocal microscopy showed that both FAK and Grb2 are located in the acrosomal region, and the flagella of non-capacitated and capacitated spermatozoa (Fig. 4I). FAK and Grb2 strongly colocalized throughout the flagellum but did not colocalize in the acrosomal region (Fig. 4I). When spermatozoa were capacitated for 10 min, FAK and Grb2 were found to be strongly colocalized in the apical region of the acrosome and throughout the flagellum (Fig. 4I). However, colocalization was lost in the apical region of the acrosome when spermatozoa were capacitated in the presence of PF573228 (Fig. 4I).

In somatic cells, ERK1/2 modulates actin polymerization through RhoA, enhancing the activity of GEF-H1 through phosphorylation at Thr678 (Fujishiro et al., 2008; Waheed et al., 2013). Thus, to determine whether FAK and ERK1/2 regulate actin polymerization in spermatozoa, we searched for evidence of ERK1/2-mediated GEF-1 phosphorylation. To this end, we used western blot analysis to detect GEF-H1 protein expression; a protein band with the molecular mass of 100 kDa was observed in capacitated and non-capacitated spermatozoa (Fig. 5A). Immunocytochemical analysis showed GEF-H1 in the acrosome and along the flagellum of non-capacitated and capacitated spermatozoa (Fig. 5B).

Commercial antibodies to detect phosphorylation of GEF-H1 at Thr678 are currently unavailable. Therefore, GEF-H1 was isolated by immunoprecipitation and phosphorylation of Thr was analyzed using an antibody against phosphorylated Thr residues (anti-p-Thr). The results showed low Thr phosphorylation of GEF-H1 in non-capacitated spermatozoa (Fig. 5C). When a similar assay was performed on capacitated spermatozoa, the amount of phosphorylated Thr residues increased significantly, whereas significantly less phosphorylation of Thr residues was detected in capacitated spermatozoa in the presence FAK inhibitor (PF573228) or ERK1/2 inhibitor (FR180204), as evidenced by densitometric analysis (Fig. 5C,D). This result suggests that FAK and ERK1/2 are involved in a phosphorylation process that enhances GEF-H1 activity.

To confirm whether GEF-H1 is related to RhoA, GEF-H1 was isolated by immunoprecipitation. We then determined whether RhoA coimmunoprecipitated with GEF-H1. Results showed that RhoA coimmunoprecipitated with GEF-H1 to a lesser extent in non-capacitated spermatozoa (Fig. 5E). The amount of RhoA coimmunoprecipitated with GEF-H1 increased significantly in capacitated spermatozoa (Fig. 5E,F). However, when spermatozoa were capacitated in the presence of PF573228 or FR180204, the amount of RhoA coimmunoprecipitated with GEF-H1 was significantly lower compared with spermatozoa capacitated in the absence of inhibitors, as evidenced by densitometric analysis (Fig. 5F). These results suggests that GEF-H1 is involved in the activation of RhoA.

To confirm that FAK and ERK1/2 participate in the activation of RhoA, the active form of RhoA (RhoA-GTP) was isolated by pull-down assay. Results showed a significantly higher amount of RhoA-GTP in capacitated spermatozoa compared with non-capacitated spermatozoa (Fig. 5G,I). During capacitation, levels of RhoA-GTP increased in the first 5 min minutes and peaked at 90 min (Fig. 5G,H). As control, spermatozoa were capacitated for 90 min in the presence of the RhoA-specific toxin Clostridium botulinum C3 (Evans et al., 2004). The amount of recovered RhoA-GTP was significantly less than that recovered from spermatozoa capacitated for 90 min in the absence of C3 (Fig. 5G). Furthermore, significantly lesser amount of RhoA-GTP was recovered in spermatozoa capacitated for 60 min in the presence of PF573228, PD98059 or FR180204 compared with that recovered from spermatozoa capacitated for the same time without any inhibitors, as evidenced by densitometric analysis (Fig. 5I,J).

RhoA has been shown to modulate actin polymerization during capacitation (Brener et al., 2003; Fiedler et al., 2008; Romarowski et al., 2015). Therefore, we estimated the amount of F-actin on the basis of levels of FITC-phalloidin fluorescence. When spermatozoa were capacitated in the presence of the RhoA inhibitor C3, we observed F-actin levels that were less than those of spermatozoa capacitated in the absence of C3 (Fig. 5K). These results confirm the participation of FAK and ERK1/2 in one of the pathways that lead to the activation of RhoA, culminating in polymerization of the actin cytoskeleton within the acrosome during capacitation.

The classic FAK activation pathway is known to involve association of FAK with β1 or β3 chains of integrin; when bound to an extracellular matrix protein, such as fibronectin; these chains then undergo a conformational change that enables FAK autophosphorylation at Tyr397 (Cooper et al., 2003; Schaller et al., 1994). Recently, our group has shown that FAK is associated with β1-integrin in guinea pig spermatozoa (Roa-Espitia et al., 2016). Moreover, fibronectin has been found in the extracellular surface of human, equine, porcine and bovine spermatozoa (Ekhlasi-Hundrieser et al., 2005; Fusi and Bronson, 1992; Glander et al., 1987; Koehler et al., 1980; Miranda and Tezon, 1992; Thys et al., 2009; Vuento et al., 1984). On the basis of these data, we propose that FAK is activated through the interaction of spermatozoa heads during capacitation, when sperm rosettes are formed. Such interaction would take place through integrins binding to fibronectin present on sperm cell surfaces. To corroborate this hypothesis, we established whether guinea pig spermatozoa express fibronectin on their surface. An anti-fibronectin antibody showed fibronectin located on the surface of the acrosome and the middle piece of guinea pig spermatozoa (Fig. 6A). Western blot analysis determined that the protein recognized by the antibody against fibronectin had a molecular mass of ∼250 kDa, which corresponds to fibronectin, as shown by fibronectin purified from human blood plasma used as a control (Fig. 6B).

To determine whether rosette formation depended on interaction between fibronectin and integrin, sperm cells were capacitated in the absence or presence of tripeptide RGD (100 μM) or anti-integrin β1 antibody (1:25), and the number of rosettes per microscopic field was counted (Fig. S2A,B). The average number of rosettes per field after 60 min of capacitation was significantly lower in the case of spermatozoa capacitated in the presence of a peptide that inhibits the binding of integrins to fibronectin, the tripeptide RDG, and spermatozoa capacitated in the presence of anti-integrin β1 antibody (Fig. 6C). Moreover, capacitation in the presence of tripeptide RGD showed a significant tendency to increase the AR (Fig. 6D); a similar effect was observed when spermatozoa were capacitated in the presence of fibronectin (Fig. 6E). The levels of p-ERK2 (Tyr204) were measured by western blotting to determine whether the interaction between RGD and integrin had any effects on the activation of ERK1/2 (Fig. 6F). Results showed a significant increase of p-ERK2 in spermatozoa capacitated in the presence of tripeptide RGD compared to those capacitated in the absence of the tripeptide (Fig. 6F,G). Taken together, these results suggest that the fibronectin/integrin pathway is associated with the activation of ERK2.

Actin cytoskeleton polymerization and remodeling are two of the main processes that take place during capacitation, and their inhibition during this sperm process has important consequences on fertilization (Brener et al., 2003; Castellani-Ceresa et al., 1993; Rogers et al., 1989). However, little is known about how actin polymerization is regulated during capacitation. Our study is the first describing a signaling pathway that regulates actin cytoskeleton polymerization during the capacitation of guinea pig spermatozoa, with FAK and ERK1/2 having preponderant roles. To the best of our knowledge, no other study has described this pathway for any cell type.

Our results show that, as in other mammalian spermatozoa, guinea pig spermatozoa expressed ERK1 and ERK2, with ERK2 being predominant (Fig. 1A). Interestingly, when activation of ERK1/2 was analyzed on the basis of phosphorylation of Tyr204, only activation of ERK2 could be observed (Fig. 1B). Similar results have been reported for human and mouse spermatozoa. This suggests that, in sperm cells, ERK2 is the main isoform playing a role in processes associated with ERK1/2, such as capacitation, AR and motility (Almog et al., 2008; Awda and Buhr, 2010; Chiu et al., 2010; de Lamirande and Gagnon, 2002; du Plessis et al., 2001; Jaldety and Breitbart, 2015; Luconi et al., 1998a; Luna et al., 2012; Luo et al., 2015; Rahamim Ben-Navi et al., 2016). This hypothesis is supported by the fact that ERK1-knockout mice are viable and fertile (Pages et al., 1999), whereas ERK2-knockout mice result in embryonic lethality (Yao et al., 2003). Additionally, we saw that a very high activation of ERK1/2 between 5 and 20 min of capacitation (Fig. 1C,D). This high activity might be required for the activation of different target proteins, such as RhoA, since the increase in ERK1/2 activity coincides with that in RhoA activity (Fig. 5H). Although ERK1/2 activity declines after 20 min of capacitation, it is higher than the activity exhibited by non-capacitated sperm cells, suggesting that this activity is sufficient to activate its target proteins required for AR.

Previous research suggests that ERK1/2 have an important role in processes, such as capacitation, AR and motility. This suggestion is based on the use of MAPK pathway inhibitors that specifically inhibit MEK1/2, as MEK1/2 directly activate ERK1/2 (Almog et al., 2008; Awda and Buhr, 2010; Chiu et al., 2010; de Lamirande and Gagnon, 2002; du Plessis et al., 2001; Jaldety and Breitbart, 2015; Kim et al., 2015; Luconi et al., 1998a,b; Luna et al., 2012; Luo et al., 2015; Rahamim Ben-Navi et al., 2016). In our study here, we used the new inhibitor FR180204 that acts directly on the kinase activity of ERK1/2 (Ohori et al., 2005). Moreover, unlike previous reports, our results of CTC, membrane microdomain redistribution and p-Tyr analyses showed that inhibition of ERK1/2 allows capacitation in guinea pig spermatozoa, but not the AR (Fig. 2). Similar results have been reported for mouse spermatozoa by using matrine alkaloid, which inhibits activation of ERK1/2 (Luo et al., 2015). In that study, the AR, induced by the Ca²⁺ ionophore A23187, was inhibited but capacitation was not (Luo et al., 2015). Therefore, ERK1/2 is likely to have a preponderant role in the AR – rather than in capacitation.

However, ERK1/2 proteins have been shown to be involved in the increase of phosphorylation at Tyr residues that occur during capacitation because Tyr phosphorylation was found to decrease in the presence of ERK pathway inhibitors (i.e. CGP78850, a competitor of Grb2–SH2 interactions; the Ras inhibitor FTI-277; the Raf inhibitor ZM336372; the MEK inhibitor U126) (O'Flaherty et al., 2006a,b). However, inhibition of ERK1/2 through the specific inhibitor FR180204 was not observed to affect phosphorylation at Tyr residues (Fig. 2C). This might be due to the fact that the use of inhibitors upstream of ERK1/2 affect other signaling pathways involved in the activation of tyrosine kinase proteins (O'Flaherty et al., 2006b). Furthermore, inhibition of MEK1/2 and ERK1/2 through PD98059 during capacitation did not result in significant changes of Tyr phosphorylation. We, therefore, suggest that the MEK1 and MEK2/ERK1 and ERK2 signaling pathway does not participate in the phosphorylation of Tyr residues that occurs during capacitation – unlike in other mammalian species in which this pathway has been related to Tyr phosphorylation (Awda and Buhr, 2010; de Lamirande and Gagnon, 2002).

Our data corroborate those reported by Jaldety and Breitbart (2015) concerning the early activation of ERK2, namely 1.5–30 min after capacitation. This is enough time for its effects to be observed – even at the end of capacitation period (60–90 min) – when a greater increase in actin polymerization was observed (30–60 min, Fig. 3B) and some of the ERK1/2 target proteins, whose participation in the AR is essential, such as phospholipase A2 (Chen et al., 2005) or Ca²⁺ channels (Jaldety and Breitbart, 2015), were activated. Activation of ERK2 seems to depend on factors intrinsic to the capacitation medium or the sperm cell. In a recent paper, we proposed that FAK has a role in the capacitation and AR processes, specifically in actin polymerization related to capacitation (Roa-Espitia et al., 2016). FAK inhibition prevents its interaction with Grb2, and the activation of ERK2 (Fig. 4). Grb2 is a scaffolding protein that triggers the canonical ERK1/2 activation pathway (Lu and Rounds, 2012; Renshaw et al., 1999). We propose that FAK is responsible for regulating the activation of ERK2, probably via the fibronectin/integrin/FAK pathway, which occurs during the formation of aggregates (spermatozoa attaching to one another by their heads, Fig. S2A,B). This mechanism is characteristic of sperm capacitation in guinea pig and other species (Fisher and Hoekstra, 2010; Monclus and Fornes, 2016; Moore et al., 2002). Our hypothesis is based on: (1) Expression of fibronectin on the surface of guinea pig spermatozoa, similar to other mammalian spermatozoa (Ekhlasi-Hundrieser et al., 2005; Fusi and Bronson, 1992; Glander et al., 1987; Koehler et al., 1980; Miranda and Tezon, 1992; Thys et al., 2009; Vuento et al., 1984). (2) Expression of integrin α5β1 and αvβ3 in human, mouse and bovine spermatozoa (Barraud-Lange et al., 2007; Boissonnas et al., 2010; Thys et al., 2009). Integrins that recognize fibronectin as a ligand (Barczyk et al., 2010) and, through its β1 and β3 chains, can interact with FAK and trigger its activation (Cooper et al., 2003; Schaller et al., 1994; Switala-Jelen et al., 2004). (3) β1-Integrin and FAK, in addition to other proteins, form a focal adhesion complex in guinea pig, mouse and human spermatozoa (Roa-Espitia et al., 2016). (4) The RDG peptide and anti-integrin β1 antibody prevent aggregation of spermatozoa, which occurs during capacitation; the presence fibronectin or RDG peptide increases the AR (Fig. 6) (Diaz et al., 2007). (5) The RGD tripeptide increases the activity of ERK2 (Fig. 6F) and, (6) prevents the aggregation of spermatozoa (Fig. 6C). This increase in ERK2 activity could be related to the increase in AR. However, the importance of the association between fibronectin and spermatozoa during capacitation is not limited to the activation of ERK1/2 and actin polymerization. It has also been shown to be involved in activation of the cAMP/PKA pathway and Tyr phosphorylation of proteins (Diaz et al., 2007; Martínez-León et al., 2015). Therefore, the interaction of spermatozoa during capacitation through the fibronectin/integrin system is crucial for sperm capacitation.

Actin polymerization during capacitation is important for mammalian spermatozoa to acquire their fertilizing ability, such that they can undergo normal AR and are able to fuse with the ovules (Brener et al., 2003; Cabello-Agüeros et al., 2003; Castellani-Ceresa et al., 1993; Rogers et al., 1989; Spungin et al., 1995). Inhibition of ERK1/2 or MEK1/2 kinases allows only ∼50% of F-actin to be polymerized (Fig. 2A,B). For controls, spermatozoa were capacitated in the absence of ERK1/2 and MEK1/2 inhibitors. This decreased amount of F-actin might be sufficient for capacitation to take place but is not enough for the AR to occur, suggesting that complete actin polymerization is necessary for both processes to succeed. These data also suggest that the fibronectin/integrin/FAK/MEK1 and MEK2/ERK2/GEF-H1/RhoA signaling pathway is only one of those involved in actin polymerization during capacitation. Mammalian spermatozoa express at least three different Rho proteins, i.e. Cdc42, RhoA and Rac1 (Delgado-Buenrostro et al., 2016; Ducummon and Berger, 2006), that assemble and organize the cytoskeleton in different structures (Hall, 1998). It is possible that each of the Rho proteins polymerizes actin in specific regions of the spermatozoon; Rac1, for example, only polymerizes actin in the apical region of the acrosome (Ramírez-Ramírez et al., 2020). We, therefore, suggest that the signaling pathway presented here directs actin polymerization in the head region and, mainly, in the acrosome (Fig. 3A).

The FAK/ERK1 and ERK2 pathway has also been associated with actin cytoskeleton remodeling in processes, such as migration and differentiation (Hyväri et al., 2018; Zhang et al., 2014). As already discussed, FAK is relevant for activation of ERK1/2 and inhibition of these kinases prevents capacitation-related actin polymerization (Fig. 3); this revealed the function of FAK as an ERK1/2 effector, connecting it to actin polymerization. Fujishiro et al. proposed that ERK1/2 can activate the GEF-H1 of RhoA (Fujishiro et al., 2008). However, our study is the first report of GEF-H1 expression in sperm cells. This GEF is activated during capacitation – as shown by the increase of GEF-H1 phosphorylated at Thr that is inhibited by PF573228 and FR180204 (Fig. 5). We also observed that the interaction between GEF-H1 and RhoA in capacitated spermatozoa is blocked by inhibiting FAK or ERK1/2 (Fig. 5E,F). Together, these data suggest that FAK and ERK1/2 are regulating RhoA through GEF-H1, which leads to polymerization of the actin cytoskeleton. The two signaling pathways FAK/ERK1 and ERK2, and ERK1 and ERK2/GEF-H1/RhoA are related to cytoskeleton remodeling, which has been separately reported (Fujishiro et al., 2008; Hyväri et al., 2018; Zhang et al., 2014). In our paper here, we report for the first time that these two pathways are connected and that they might have a relevant role in sperm physiology.

The interaction of spermatozoa could be a form of interaction in those that form rosettes during capacitation, where both fibronectin associated with the sperm surface and integrins have an important role. Different evidence supports the importance of fibronectin in this interaction process: species in which spermatozoa do not form rosettes also contain fibronectin associated with their surface (Ekhlasi-Hundrieser et al., 2005; Fusi and Bronson, 1992; Leahy et al., 2011; Thys et al., 2009; Vilagran et al., 2015), which is released during the capacitation process (Hernández-Silva et al., 2020). This released fibronectin might have an autocrine or paracrine role, either in vivo or in vitro, when associated with the integrins of another spermatozoon. Interestingly, the amount of fibronectin within the oviductal liquid increases during the fertile stage of the estrous cycle (Osycka-Salut et al., 2017), and this would have an impact on sperm physiology. Therefore, although spermatozoa do not form rosettes, fibronectin might have an essential role in the activation of various signaling pathways (Martínez-León et al., 2019; Martínez-León et al., 2015).

Therefore, we propose that fibronectin/integrin/FAK/MEK1 and MEK2/ERK2/GEF-H1/RhoA is one of the signaling pathways that regulate actin cytoskeleton polymerization during capacitation in guinea pig spermatozoa. No other study has described this pathway in sperm cells or any other cell type. Interestingly, at least in vitro, this signaling pathway depends on sperm–sperm interaction: the integrins of one of sperm cell associate with the fibronectin found on the surface of another sperm cell (Fig. 7), which allows activation of this signaling pathway leading to actin polymerization and remodeling of the cytoskeleton during capacitation.

PKA inhibitor Rp-cAMPS triethylammonium salt (#A165) and reagents were purchased from Sigma-Aldrich, except protease inhibitor (Complete™ cocktail tablets), which was obtained from Roche Diagnostics and Molecular Biochemicals (Mannheim, Germany). Nitrocellulose membranes, acrylamide, N,N′-methylenebisacrylamide and sodium dodecyl sulfate (SDS) were obtained from Bio-Rad Laboratories (Hercules, CA). C3 Rho inhibitor (CT04) was purchased from Cytoskeleton Inc (Denver, CO). Fibronectin human plasma (#341635) and Enhanced Chemiluminescence (ECL) reagent were purchased from Amersham (Buckinghamshire, UK) or Millipore (Temecula, CA). FAK inhibitor PF573228 (#3239) was obtained from Tocris Bioscience (Bristol, BS11, UK). ERK1/2 inhibitor FR180204 (#sc-203945), MEK1/2 inhibitor PD98059 (#sc-3535A) and FAK inhibitor 14 (#sc-203950) were obtained from Santa Cruz Biotechnology Inc. (Santa Cruz, CA).

Anti-ERK1/2 (#M5670; 1:5000) was obtained from Sigma-Aldrich. Anti-p-Tyr-100 (#9411S; 1:2000) was purchased from Cell Signaling Technology. Anti-p-Thr (#139200; 1:1000) was obtained from Thermo Fisher Scientific (Waltham, MA). Anti-FAK (#sc-551; 1:250), anti-p- FAK-Y925 (#sc-11766; 1:250), anti-Grb2 (#sc-255; 1:250), anti-Flot2 (#sc-28320; 1:50) and anti-p-ERK1/2 (#sc7383; 1:250) were purchased from Santa Cruz Biotechnology Inc. (Santa Cruz, CA). Anti-GEF H1 (#GTX125893; 1:500) was obtained from Gene Tex. Anti-RhoA (#ab68826; 1:500) was obtained from Abcam (Cambridge, UK). Secondary antibodies labeled with horseradish peroxidase (HRP), TRITC or FITC were obtained from Jackson Immunoresearch Laboratories Inc. (West Grove, PA). The anti-utrophin 71 antibody (#Up71; 1:1000) was kindly donated by Dr Dominique Mornet, from INSERM U592 (Paris, France).

Male Dunkin-Hartley guinea pigs (Cavia porcellus) weighing 800–1000 g were obtained from the bioterium at Cinvestav-IPN. All animal experiments and handling procedures were approved by the Internal Committee for Laboratory Animal Care and Use of the CINVESTAV-IPN (CICUAL No. 0122-14) following the American Veterinary Medical Association guidelines. Every effort was made to minimize potential animal pain, stress or distress.

Capacitation was performed as described by Rogers and Yanagimachi (1975). Spermatozoa were obtained from the ductus deferens and cauda epididymis of guinea pigs and then washed in phosphate-buffered saline (PBS) pH 7.4 (140 mM NaCl, 2.7 mM KCl, 1.5 mM KH₂PO₄, 8.1 mM Na₂HPO4). Sperm cells (35×10⁶ cell ml⁻¹) were capacitated by incubation at 37°C in minimal culture medium (MCM-PL) pH 7.8, containing: 105.9 mM NaCl, 1.71 mM CaCl₂, 25.07 mM NaHCO₃, 0.25 mM sodium pyruvate, 20 mM DL-lactic acid 85%. Capacitation was conducted for the desired time. Spermatozoa were also capacitated in the presence of different drugs (PF573228, FAK inhibitor 14, PD98059, FR180204, C3 or Rp-cAMPS), which were added at the beginning of capacitation. Any change is indicated in the text.

Non-capacitated and capacitated spermatozoa were fixed using 1.5% (v/v) formaldehyde and 0.1% (v/v) glutaraldehyde in PBS. After 1 h, spermatozoa were collected by centrifugation (600 g for 3 min). Pelleted spermatozoa were incubated in 50 mM NH₄Cl for 10 min, rinsed twice in PBS and once in bi-distilled water. Microscope slides were prepared using this suspension, air-dried at room temperature overnight and stored at 4°C. Sperm cells were permeabilized in PBS with 0.1% Triton for 20 min at room temperature and washed in PBS. Antibodies against ERK1/2 (1:100), p-ERK1/2 (1:50), FAK (1:50), Grb2 (1:50), Flot2 (1:50) and GEF H1 (1:50) were diluted in PBS with 1% bovine serum albumin (BSA) (blocking solution) and incubated on slides overnight at room temperature. Slides were incubated for 2 h at 37°C with the appropriate TRITC- or FITC-conjugated secondary antibody diluted in blocking solution. Samples were mounted in Gelvatol, covered with glass coverslips, adequately sealed, and stored at −20°C until observations. Stained cells were imaged using a confocal laser scanning microscope (Leica TCS SP8) and analyzed by LAS AF Lite imaging software (Ver. 2.6.3).

The presence of fibronectin on sperm cell surfaces was determined as mentioned above, except spermatozoa that were not permeabilized after fixation, which were treated directly with anti-fibronectin antibody (1:100) in blocking solution.

To define the specificity of antibodies, two controls were performed: (1) Sperm was incubated only with the secondary antibody; (2) non-specific antibody was used as the primary antibody. These control procedures did either not show a fluorescence signal at all or only a very weak one. Results are shown in Fig. S2C.

Non-capacitated spermatozoa and spermatozoa capacitated at different times were centrifuged (600 g) for 3 min and suspended in RIPA buffer (25 mM TRIS HCl pH 7.6, 150 mM NaCl, 1% NP-40, 1% sodium deoxycholate, 0.1% SDS), supplemented with protease inhibitors [5 mg ml⁻¹ soybean trypsin inhibitor, 100 mg ml⁻¹ benzamidine, 30 mg ml⁻¹ pepstatin, 30 mg ml⁻¹ leupeptin, 30 mg ml⁻¹ aprotinin, 1 mM PMSF diluted in dimethyl sulfoxide, 20 mg ml⁻¹ iodoacetamide, 1 mM sodium orthovanadate, 10 mM sodium fluoride, 10% glycerol, 2.5% complete mini protease inhibitor cocktail (1 tablet diluted in 1 ml H₂O)]. Samples were then incubated for 20 min on ice and centrifuged (5000 g) for 30 min at 4°C. Supernatants were collected and protein concentration was determined (Bradford, 1976). These samples were boiled for 5 min in sample buffer at pH 7.0 (Laemmli, 1970), separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) in 10% polyacrylamide non-gradient gels, transferred to nitrocellulose membranes and blocked using 5% skim milk in PBS with 1% Triton pH 7.5. Membranes were incubated overnight at 4°C with primary antibodies: anti- ERK1/2 (1:5000), anti-p-ERK1/2 (1:250), anti-FAK (1:250), anti-p-FAK (1:250), anti- GEF H1 (1:500) and anti-Grb2 (1:250), antibodies were diluted in PBS with 1% Triton and 3% bovine serum albumin (BSA); membranes were subsequently incubated with their respective secondary antibody: HRP-conjugated anti-rabbit (1:10,000), HRP-conjugated anti-mouse (1:5000), or HRP-conjugated anti-goat (1:5000). The nitrocellulose membranes were bathed with Amersham ECL and exposed to MXB film inside an autoradiography cassette or with the Odyssey Fc Imaging System. The densitometric analysis of the images was performed using the ImageJ 1.8.0_112 program. After photoactive protein detection, membranes were stripped and immunoblotted again. Up71 was used to normalize for protein loading.

Spermatozoa suspension (20×10⁶) was recuperated, centrifuged, washed three times with PBS supplemented with phosphatase inhibitors [complete mini protease inhibitor cocktail (1 tablet diluted in 50 ml PBS), 1.0 mM sodium orthovanadate and 0.02 mM genistein], after resuspended in sample buffer (Laemmli, 1970) without mercaptoethanol and boiled for 5 min as previously described (Visconti et al., 1995). After centrifugation at 5000 g for 20 min at 4°C, pellets were discarded, supernatants added to 5% 2-mercaptoethanol and stored at −70°C. Each sample was again boiled for 5 min in sample buffer, resolved using 10% SDS-PAGE, and transferred onto a nitrocellulose membrane (Towbin et al., 1979). Nitrocellulose membranes were blocked using PBS containing 0.1% Tween-20 and 5% skimmed milk. Anti-p-Tyr antibody was diluted (1:2000) in blocking solution, added to the nitrocellulose membranes, and incubated overnight at 4°C. The membranes were then washed 5× for 7 min and incubated with HRP secondary anti-mouse (1:5000). Immunoreactive proteins were detected by chemiluminescence using Amersham ECL and exposed the Odyssey Fc Imaging System. Densitometric analysis of the ImageJ 1.8.0_112 program was used, considering the sum of the densitometry measurements of all the bands that appear for each lane. Up71 was used to normalize the gel loading.

This procedure, first described by Ward and Storey (1984), was modified for guinea pig spermatozoa according to Maldonado-García et al. (2017). At the time of the assay, 45 μl non-capacitated and capacitated (90 min) spermatozoa suspension in absentia or presence of FR180204 inhibitor were mixed with 45 μl chlortetracycline (CTC) solution (750 mM CTC in 130 mM NaCl, 5 mM cysteine, 20 mM TRIS pH 7.8) and incubated for 20 s in a water bath at 37°C. Immediately after incubation, the CTC-sperm suspensions were fixed with 0.75 μl glutaraldehyde (12.5%) in 0.5 mM TRIS pH 7.4 at room temperature to achieve a final concentration of 0.1% glutaraldehyde. A 10 μl sample of the CTC-sperm suspension was smeared onto a glass slide, put onto a coverslip, adequately sealed, and stored for 48 h at 4°C to clear strong background fluorescence. The CTC solution was kept in a light-shielded container at 4°C at all times. To physiologically-induced AR, under capacitating conditions in the absence or presence of the FR180204 inhibitor, the spermatozoa were incubated with Progesterone (4-Pregnene-3,20-dione, 10 μM) for 20 min before performing the CTC assay, which was performed as mentioned above. All fluorescence images were obtained by using an epifluorescence microscope (Olympus BX5) and were registered in Nis-Element 3.1. Five-hundred spermatozoa per sample were classified as expressing one of three CTC staining patterns. These are: F pattern, i.e. faint fluorescence in the acrosome region, characteristic of non-capacitated acrosome-intact cells; B pattern, i.e. bright fluorescence in the acrosomal region and one band along the equatorial segment that is characteristic of capacitated acrosome-intact cells; AR pattern, i.e. fluorescence in the post-acrosomal region and equatorial segment, characteristic of physiologically capacitated acrosome-reacted cells (CTC excitation at 330–380 nm, emission at 420 nm).

Non-capacitated and capacitated spermatozoa at different times in the absentia or the presence of PD98059 inhibitor or FR180204 inhibitor were fixed on slides as described above. Sperm smears were incubated for 1 h at 37°C with FITC-phalloidin (1:25) diluted in PBS and washed five times in PBS, and added to coverslips in Gelvatol. Fluorescence images were obtained using a confocal laser scanning microscope (Leica TCS SP8) and the fluorescence intensity of the acrosomal region of 500 spermatozoa was analyzed for each condition, using the NIS-Elements 3.1 imaging software. The fluorescence intensity of the different capacitation times (N) was normalized with respect to that exhibited by non-capacitated spermatozoa (N0). The relation N/N0 shows the changes that F-actin experiences during capacitation.

Anti-GEF H1, anti-Grb2 or anti-FAK antibodies were bound to protein A/G-agarose for immunoprecipitation assays per manufacturer's instructions (Pierce^® Crosslink Immunoprecipitation Kit). Antibody (20 µg) was coupled to 20 µl of the resin, incubated for 1 h at room temperature, and then washed to eliminate excess antibody. Crosslinking was achieved by incubating the bound antibody in the presence of disuccinimidyl suberate at room temperature for 60 min and washing to eliminate excess non-crosslinked antibody. Sperm protein extracts (500 µg) from spermatozoa capacitated, spermatozoa non-capacitated, and spermatozoa capacitated in the presence of inhibitors were incubated with constant agitation for 12 h at 4°C. Proteins not bound to the A/G-agarose antibody were recovered by centrifugation at 4000 g after three washes. Bound proteins were eluted, recovered by centrifugation at 4000 g and boiled in Laemmli sample buffer (Bio-Rad Laboratories, Hercules, CA) for 5 min. Finally, these proteins were separated by SDS-PAGE, transferred to nitrocellulose membranes and subjected to immunoblot analysis.

RhoA activity was determined by pull-down assay using the RhoA Pull-down Activation Assay Biochem Kit (Cytoskeleton). Proteins (300 µg) from non-capacitated spermatozoa, capacitated spermatozoa at different times, or spermatozoa capacitated in the presence of inhibitors (PF573228, FR180204, PD98059 or C3) were incubated with 50 µg agarose-conjugated Rhotekin-RBD at 4°C for 1 h. Proteins not bound to the Rhotekin-RBD were recovered by centrifugation at 5000 g at 4°C for 3 min after two washes. Proteins precipitated with Rhotekin GTP-RhoA were boiled in Laemmli sample buffer to enable the release of active RhoA, and then separated by SDS-PAGE, transferred to nitrocellulose membranes, and subjected to immunoblot analysis.

Experiments were replicated no less than three times. All data are presented as mean±s.d. Statistical significance was analyzed using t-test or ANOVA for comparisons between two groups or multiple comparisons, respectively. SigmaPlot version 11.0 was used for the analysis, and P≤0.05 was considered statistically significant.

We wish to thank Jaime Escobar, Chief Manager at Unidad de Microscopia Confocal (Cell Biology Department, CINVESTAV-IPN) for providing access to confocal facilities.

The peer review history is available online at https://jcs.biologists.org/lookup/doi/10.1242/jcs.239186.reviewer-comments.pdf