Lipid Modulation of Calcium Flux through Ca_V2.3 Regulates Acrosome Exocytosis and Fertilization

Summary

Membrane lipid regulation of cell function is poorly understood. In early development, sterol efflux and the ganglioside G_M1 regulate sperm acrosome exocytosis (AE) and fertilization competence through unknown mechanisms. Here, we show that sterol efflux and focal enrichment of G_M1 trigger Ca²⁺ influx necessary for AE through Ca_V2.3, whose activity has been highly controversial in sperm. Sperm lacking Ca_V2.3’s pore-forming α_1E subunit showed altered Ca²⁺ responses, reduced AE, and a strong sub-fertility phenotype. Surprisingly, AE depended on spatio-temporal information encoded by flux through Ca_V2.3—not merely the presence/ amplitude of Ca²⁺ waves. Using both studies in sperm and voltage clamp of Xenopus oocytes, we define a molecular mechanism for G_M1/Ca_V2.3 regulatory interaction, requiring G_M1’s lipid and sugar components and Ca_V2.3’s α_1E and α₂δ subunits. Our results provide mechanistic understanding of membrane lipid regulation of Ca²⁺ flux and therefore Ca²⁺-dependent cellular and developmental processes such as exocytosis and fertilization.

Introduction

Determining the mechanisms by which membrane lipids can change cell function is of rapidly growing interest. Here, we study aspects of the initiating event of developmental biology—fertilization—as a model for lipid regulation of cell function. Although it has long been known that lipids govern the sperm’s ability to undergo acrosome exocytosis (AE) and complete the subsequent steps of fertilization, the molecular mechanisms through which this control is exerted have remained unclear for decades.

Lipids are intimately involved in both the positive and negative regulation of sperm fertilization competence. Removal of sterols from the plasma membrane is strictly required for sperm to become able to fertilize, during a process known as “capacitation” (Travis and Kopf, 2002). Conversely, the seminal plasma protein, SVS2, has been shown to bind the ganglioside G_M1 in murine sperm, and act as a “decapacitation factor,” keeping the sperm quiescent (Kawano et al., 2008). Co-localization of sterols and G_M1 is highly conserved among mammals (Buttke et al., 2006), occurring within dynamic micro-domains segregated to a plasma membrane macro-domain overlying the acrosome (APM), the sperm’s single exocytotic vesicle (Selvaraj et al., 2006, Selvaraj et al., 2009).

In this position, sterols and G_M1 could regulate membrane fusion through several potential mechanisms including modulation of Ca²⁺ flux, which is a known trigger for AE. Interestingly, in several cell types G_M1 has been suggested to influence flux through membrane Ca²⁺ ATPases and exchangers (Ravichandra and Joshi, 1999; Zhao et al., 2004), and voltage-independent gating of Ca_V1.2 channels (Carlson et. al., 1994, Fang et al., 2002), though no molecular mechanisms have been described.

In an attempt to identify Ca²⁺ channels involved in regulating sperm function, recent studies have focused on the CatSper channel complex in both human and mouse sperm. CatSper is pH sensitive and weakly voltage dependent (Ren and Xia, 2010); it mediates progesterone (P4)-induced activation of human sperm (Lishko et al., 2011), and its absence in mouse genetic models results in abnormal motility and infertility (Ren and Xia, 2010). Failure to detect other channel activities in electrophysiological recordings has led some to the highly controversial conclusion that CatSper is the only Ca²⁺ channel in sperm. However, patch-clamp experiments in mouse sperm are typically performed on immature cells that have not completed membrane maturation in either the epididymis or the female tract. Both of these maturational processes involve substantive changes in membrane lipid composition, such as sterol efflux during capacitation. Thus, if sperm Ca²⁺ flux is in some way modulated by membrane lipids, it would likely avoid detection in those highly constrained experimental systems (additional technical limitations to electrophysiological detection of other channels are described in the discussion). Indeed, it has been suggested that other channels could be present, but are undetectable by existing patching methods (Kirichok and Lishko, 2011). This possibility is supported by the findings of a study in which the sperm head itself was patched, and revealed multiple channel activities that were spatially organized (Jimenez-Gonzalez et al., 2007).

Before the current controversy over whether CatSper is responsible for all Ca²⁺ flux in sperm, it was believed that the increase in intracellular Ca²⁺ in the sperm head leading to AE occurs in discrete steps (Florman et al., 2008), where the upstream stimuli include alkalinization and hyperpolarization. CatSper could play this upstream role in this model as it is pH-sensitive and also raises resting Ca²⁺ concentrations upon sterol efflux (Xia and Ren, 2009). Next, an un-identified voltage-gated Ca²⁺ channel (VGCC) would mediate focal, transient Ca²⁺ elevations (Arnoult et al., 1999). Together with the activation of phospholipase Cδ4 (Fukami et al., 2003), these events would trigger a final, sustained elevation in intracellular Ca²⁺ that in mouse sperm is most likely mediated by TRPC2 channels (Jungnickel et al., 2001). As a result, SNARE-mediated fusion of the APM and the outer acrosomal membrane would occur (Yunes et al., 2000).

Inability to characterize the VGCC has certainly contributed to the controversy surrounding models of Ca²⁺ flux and initiation of AE. VGCCs are heteromeric complexes comprised of a pore-forming α₁ subunit and auxiliary α₂δ, β, and γ subunits (Catterall, 2000). Several subtypes of VGCCs have been described in sperm (Darszon et al., 2006), but their in vivo activity has been difficult to characterize. Patch-clamp recordings from developing male germ cells and pharmacological studies suggested the activity of T-type currents (Arnoult et al., 1998). However, mice lacking T-type Ca²⁺ channels Ca_V3.1 (Stamboulian et al., 2004) and Ca_V3.2 (Escoffier et al., 2007) are fertile, and the remaining current in these mice displays characteristics that differ slightly from somatic cell T-type currents (Stamboulian et al., 2004). Additionally, work characterizing depolarization-induced Ca²⁺ influx to simulate zona pellucida (ZP)-induced Ca²⁺ rise in mature mouse sperm proved to be insensitive to blockers of L-, P/Q-, and T-type channels (Wennemuth et al., 2000).

One candidate not ruled out by previous studies is VGCC Ca_V2.3. Despite considerable investigation in multiple cell types, its physiological role remains the least understood of all VGCCs (Weiergraber et al., 2006). These channels introduce large inward Ca²⁺ currents (Bourinet et al., 1996) and are responsible for the majority of the “residual”, or R-type current (Randall and Tsien, 1995). R-type channels have been defined based on their resistance to inhibitors of other high-voltage activated (L-, N-, and P/Q-type) Ca²⁺ channels and by their sensitivity to the spider venom peptide SNX-482 (Bourinet et al., 2001). The Ca_V2.3 VGCC is widely distributed, being found in the central nervous system, endocrine (Pereverzev et al., 2005), cardiovascular (Weiergraber et al., 2005), reproductive (Sakata et al., 2002) and gastrointestinal systems (Grabsch et al., 1999). Several studies have shown their modulation by membrane binding partners, including SNARE proteins (Wiser et al., 2002, Cohen and Atlas, 2004), and confirmed their involvement in exocytosis in various cell types (Wang et al., 1999; Albillos et al., 2000; Vajna et al., 2001). Although involvement of Ca_V2.3 channels in sperm function has been suggested (Westenbroek and Babcock, 1999; Wennemuth et al., 2000; Sakata et al., 2001, 2002), no conclusive evidence has been provided. Interestingly, sperm from Ca_V2.3 null mice have been reported to display aberrant motility and to have different responses in Ca²⁺ uptake in the sperm head, though direct involvement in AE or fertilization apparently was not investigated.

We now describe a mechanism through which membrane lipids can modulate an R-type VGCC. Cell biological, pharmacological, and genetic evidence confirm that focal clustering of the extracellular sugars of G_M1, or loss of sterols from the membrane, help stimulate AE through the activation of Ca_V2.3 channels. Notably, mice lacking the α_1E subunit of this channel have marked subfertility in natural matings and severe defects in AE and in vitro fertility. In addition, our comparisons between sperm from Ca_V2.3 null and wild type mice show the exquisite spatial and temporal control of Ca²⁺ flux needed to achieve AE. Remarkably, the presence or amplitude of a wave is not sufficient to induce exocytosis, a point of considerable potential importance for the multiple cell types whose functions rely on regulated exocytosis. Lastly, using a reductionist, heterologous expression system and voltage-clamping, we show that G_M1’s ability to exert these effects is not restricted to sperm and is mechanistically dependent upon the presence of the α₂δ subunit. Together, our data led us to create a molecular model for G_M1/lipid regulation of Ca_V2.3 function in sperm, thereby controlling AE and fertilization.

Results

Focal enrichment of G_M1 induces AE in sperm incubated under capacitating conditions

Our prior characterization of the conservation of G_M1 segregation to the APM, as well as the temporary binding of SVS2 to G_M1 to maintain functional quiescence (Kawano et al., 2008), led us to hypothesize that this lipid plays an important role in fertilization. Upon adding the pentavalent B subunit of cholera toxin (CTB) to live, capacitated sperm, we noted that a sub-population exhibited clearing of labeled CTB signal over the apical acrosome (AA), potentially consistent with AE (Selvaraj et al., 2007). To investigate whether this clearing involved loss of the membrane alone or was also associated with a loss of acrosome matrix as occurs during physiologic AE, we assessed acrosome status with Coomassie staining after treatment with CTB or commonly employed agonists for AE (P4 or solubilized ZP). Addition of Ca²⁺ ionophore (A23187) provided a positive control, and sperm incubated under capacitating conditions alone gave a baseline for spontaneous AE or loss of acrosome matrix as a result of experimental handling. Exposure to CTB induced AE in sperm incubated in capacitating conditions at levels statistically similar to sperm incubated with P4 or ZP (Figure 1B), and which were significantly higher than seen in the control capacitated sperm.

Figure 1. CTB and G_M1 induce acrosome exocytosis (AE).

A) Representative Coomassie staining patterns of sperm having undergone AE (left) or acrosome intact (right). Arrows point to absence or presence of staining of acrosome matrix proteins. B) Box whisker plots of percentage of capacitated (C) wild-type sperm undergoing AE in response to Ca²⁺ ionophore (A23187), progesterone (P4), solubilized zona pellucida (ZP), or CTB (n=4 trials, ≥100 sperm counted for each condition). Sperm were incubated under capacitating conditions for 50 minutes and treated with P4 (3 µg/ml), ZP (2 ZP/µl), or CTB (1.5 µg/ml) to induce AE for 10 minutes prior to Coomassie assessment. As shown, CTB induced a significant increase in the percent of AE (p<0.0001) that was statistically similar to AE% induced by P4 or ZP. C) Box whisker plots of percentage of capacitated sperm (C) undergoing AE in response to 25µM of exogenous G_M1, asialo-G_M1 (Asialo), ceramide (Cer), or solvent control (DMSO, data not shown) (n=4). Exogenous G_M1 induced a significant increase in AE% (p=0.005) compared to capacitated sperm (C), Asialo, Cer or a solvent control (DMSO, data not shown). D) Box whisker plots of the percentage of sperm undergoing AE in response to CTB in the presence of nickel (Ni). Ni significantly inhibited CTB-induced AE at 30 and 100μM (p=0.02).

To determine whether the observed effect of CTB was due to focal enrichment from CTB’s ability to bind up to 5 molecules of G_M1, as opposed to non-specific membrane perturbations, AE was measured following addition of 25µM exogenous G_M1. This concentration of G_M1 has been shown to stimulate G_M1–induced signaling events in somatic cells without affecting overall membrane integrity or cholesterol efflux (Yatomi et al., 1996). As seen with the addition of CTB, exogenous G_M1 significantly increased the incidence of AE in capacitated sperm (Figure 1C), but neither G_M1 nor CTB affected the viability or AE of non-capacitated sperm (data not shown; n=4).

Glycosphingolipids like G_M1 are complex molecules capable of interacting via their sugar and/or lipid components. Our first investigation into the molecular mechanism of G_M1’s effects revealed that neither ceramide nor asialo-G_M1 (lacking the extracellular sialic acid), were able to induce a significant elevation in AE as compared to G_M1 (Figure 1C). Similarly, addition of sialic acid alone did not induce AE (data not shown). These results suggest that G_M1 exerted its effects in sperm by means of the extracellular sugars, particularly the sialic acid, but only if specifically oriented by the ceramide tail in the outer leaflet.

To identify whether the effects of G_M1 on AE were being mediated by a regulation of Ca²⁺ flux, we tested the effect of co-incubation with nickel, a non-specific inhibitor of Ca²⁺ conducting channels. CTB-induced AE was inhibited by nickel in a concentration dependent manner (Figure 1D). Investigating further, we found that extracellular Ca²⁺ was essential for evoked membrane exocytosis, and that inhibition of Ca²⁺ influx (including block of SOCC) with either Cd²⁺ (100µM) or La³⁺ (10µM) prevented CTB-induced AE (Figure 2A). The effect of 30μM Ni²⁺ suggested that T- and L-type Ca²⁺ channels were not involved in G_M1-mediated AE. In support of this, nifedipine, which has reported selectivity for L-type channels, failed to inhibit CTB-induced AE even at concentrations up to 100μM. Kurtoxin type KL1, which has been shown to inhibit LVA Ca²⁺ channels and specifically T-type Ca_V3.1 and Ca_V3.2 in sperm (Lopez-Gonzalez et al., 2003), had no effect, suggesting that T-type channels were not involved. However, in agreement with our Ni²⁺ and Cd²⁺ data, blocking Ca_V2.3 Ca²⁺ channels with SNX-482 prevented the CTB-induced AE (Figure 2A; it also inhibited G_M1-induced AE [data not shown]). Together, these pharmacological data suggested that both Ca_V2.3 channels and SOCC were necessary for G_M1-mediated AE. No further increase in the occurrence of AE was seen with the combination of both P4 and CTB or P4 and G_M1, suggesting that these combinations were acting on the same physiologically relevant population of cells (data not shown).

Figure 2. Pharmacologic and genetic data suggesting the importance of α_1E in AE.

A) The effect of various pharmacological inhibitors on AE in capacitated sperm (C) in response to CTB. CTB-induced AE required extracellular Ca²⁺ as shown by the lack of response in Ca²⁺-depleted media (de-Ca). CTB-induced AE was inhibited by nonspecific Ca²⁺ influx inhibitors La³⁺ (La) or Cd²⁺ (Cd), but not L-type channel inhibitor nifedipine (Nif) or T-type channel inhibitor kurtoxin KL1 (Kur). The Ca_V2.3 inhibitor SNX-482 (SNX, 500nM) inhibited CTB-induced AE (n=4 trials, p<0.005). B) Indirect immunofluorescence images of wild-type (upper) or an α_1E homozygous null mouse sperm labeled with antibodies against the α_1E subunit of the Ca_V2.3. A high magnification view of representative single sperm is shown (n=4 separate experiments with both wild-type and α_1E-null sperm, with >100 sperm assessed for each). C) Percentages of AE in α_1E null sperm (KO) and wildtype sperm (WT) in response to Ca²⁺ ionophore A23187, P4, CTB, G_M1, and ZP (n=4). See also Figure S1.

Genetic investigation of Ca_V2.3 involvement in AE

We next took a genetic approach to determine whether Ca_V2.3 plays a role in AE, to complement the pharmacological data. Sperm from mice lacking the α_1E subunit of Ca_V2.3 have previously been described as having normal intracellular Ca²⁺ concentrations, but altered Ca²⁺ dynamics (Sakata et al., 2002). These sperm had a straightened flagellar wave-form and delayed and lower Ca²⁺ rises in response to mannosylated-BSA, an agent that can induce AE in human sperm (Amin et al., 1996). However, a direct role for this channel in AE or other steps of fertilization has not been described. We therefore utilized a mouse model null for the α_1E subunit (Figure S1) to investigate this further.

Consistent with a previous report (Wennemuth et al., 2000), α_1E was expressed in the AA of mature sperm, but was absent in sperm from the null mice (Figure 2B). During in vitro assays of AE, sperm from α_1E-null mice showed severe defects, failing to undergo AE in response to P4, CTB, G_M1, or ZP (Figure 2C). Occurrence of AE in response to Ca²⁺ ionophore was normal in these sperm, suggesting that membrane fusion machinery was intact and functional in these sperm (Figure 2C). Defects in fertility were also noted in natural matings, with α_1E-null mice having significantly smaller litters than age and season-matched wild type mice of the background strain (null: 4 pups ± 2.1 S.D. [n=13] versus WT: 7.1 pups ± 1.6 S.D. [n=11], p=0.001). Matings of null males with wild type females and vice versa demonstrated that the subfertility was solely due to a male factor (null male/WT female: 3.5 pups ± 3.5 S.D. [n=4] versus WT male/null female: 7.25 pups ± 1.7 S.D. [n=8], p=0.00002). Careful investigation using in vitro fertilization (IVF) assays revealed that the presence or absence of the cumulus cells did not significantly impact the relative fertility of the sperm from the null mice (Table 1). Importantly, removal of the ZP from the oocytes resulted in complete rescue of the phenotype (Table 1), providing compelling demonstration that the defect was early in sperm-egg interactions, consistent with a physiologic failure in AE and/or penetration of the ZP.

Table I.

Average success of in vitro fertilization for homozygous and heterozygous matings.

Oocyte Status		Sperm Genotype	% Fertilized (n)	SEM
Zona intact	Cumulus intact	WT	69% (134)	7.3
	Cumulus intact	KO	26% (121)	9.2
Zona intact	cumulus free	WT	65% (106)	6.4
	cumulus free	KO	19% (104)	8.4
Zona free		WT	68% (91)	3.3
		KO	78% (148)	7.8

Single cell Ca²⁺ imaging

Our observation of significant defects in AE in the α_1E-null sperm led us to hypothesize that a functional interaction between G_M1 and α_1E might be involved in Ca²⁺ mobilization required for membrane fusion and AE. Thus, we sought to investigate the early steps in sperm Ca²⁺ response to addition of exogenous G_M1 or CTB and in response to sterol efflux, a physiological requirement for sperm to become able to fertilize. For these experiments we utilized high speed Ca²⁺ imaging of single sperm with a controlled delivery system for local administration of G_M1, CTB, or 2-hydroxypropyl-β-cyclodextrin (2-OHCD). To stimulate Ca²⁺ mobilization, a short (10 sec) puff of stimulant was applied to individual sperm at 37°C, while recording the changes in Fluo4 fluorescence intensity at image rates of up to 50 frames per second.

Using G_M1 or CTB, we observed two major types of spatially and temporally resolved Ca²⁺ mobilization dynamics in individual sperm. As illustrated in Figure 3A, Ca²⁺ “transients” typically initiated within seconds after the addition of G_M1 or CTB. In some cells, Ca²⁺ transients were followed by a Ca²⁺ rise that most often initiated from the connecting piece/midpiece (Ho and Suarez, 2001), and propagated towards the AA; we refer to these sustained elevations as “waves”.

Figure 3. Single cell imaging reveals altered CTB- or G_M1-mediated Ca²⁺ transients in α_1E-null sperm.

A) Fluorescence changes as measured in a CD1 sperm loaded with Fluo4-AM, representing the changes in Ca²⁺ over time in the region of interest (ROI) indicated by the yellow circle on left panel. B) Distribution of Ca²⁺ transient events over sperm heads (each green dot represents the location of a single transient in a different sperm, compiled here into single images), following CTB addition to CD1 or KO sperm. C) Summary of Ca²⁺ transient events that occurred in CD1 or α_1E KO sperm upon various experimental conditions. D) Statistical analysis of transient lag time, as measured from the beginning of stimulation to the onset of the Ca²⁺ signal. E) Statistical analysis of integrated Ca²⁺ influx for the duration of the transient rise phase. Sperm from 3 or more mice were imaged for each experimental condition (n>25 for each condition tested; p value indications are as follows: 0.05<**>0.01, ***<0.005). See also Figure S2.

In terms of transient and wave characteristics, we found no difference between sperm from strain-matched controls and CD1 mice, which are a standard outbred model for sperm cell physiology studies (data not shown). We observed G_M1- and CTB-induced Ca²⁺ transients in ~50% of both α_1E-null or control CD1 sperm (Figure 3C). However, in sperm from control mice the Ca²⁺ transients typically localized along the AA whereas no consistent spatial pattern was observed for transients in the α_1E-null sperm, with the majority occurring in the equatorial segment (Figure 3B). Notably, in CD-1 sperm the spatial distribution of Ca²⁺ transients corresponded precisely with the localization pattern of the α_1E subunit as determined by indirect immunofluorescent labeling (Figure 2B). Addition of G_M1 in the presence of 10µM La³⁺ significantly reduced Ca²⁺ transient occurrence, consistent with an influx of extracellular Ca²⁺ rather than release from internal stores. Interestingly, upon addition of exogenous G_M1 to CD1 sperm incubated under non-capacitating conditions (NON-CAP), we observed high variability in the occurrence of transients. Identification of Ca²⁺ transients in non-capacitated sperm suggests that signaling processes dependent upon exposure to stimuli for capacitation are not required for activation of the involved Ca²⁺ channel when G_M1 is enriched in the membrane by exogenous addition.

Although Ca²⁺ transients occurred in equivalently sized populations of both α_1E-null sperm and wild type CD1 sperm, there were several dynamic differences in these responses in addition to having different spatial localizations. Transient Ca²⁺ responses in control sperm initiated 8.9±1.5 seconds following addition of CTB, whereas these transients appeared significantly earlier in response to G_M1 (2.2±0.4 seconds; Figure 3D). Comparable differences in transient lag time were observed in the α_1E null sperm in response to G_M1 or CTB. This difference would be consistent with slower G_M1 clustering kinetics that depended upon multivalent binding with CTB versus a more rapid process of G_M1 membrane integration.

Strikingly, there was a ~3 fold decrease in total Ca²⁺ influx in the α_1E-null sperm during the transient response upon addition of either G_M1 or CTB (Figure 3E). Also of note, mild depolarization of the sperm membrane by addition of high K⁺ (60mM) failed to stimulate Ca²⁺ transients (Figure S2A), further suggesting that in sperm, Ca_V2.3 voltage dependent activation is strongly modulated by its lipid environment.

Although capable of responding to exogenous G_M1 (Figure 3C), sperm incubated under non-capacitating conditions demonstrated a reduced Ca²⁺ influx in response to G_M1 similar to that seen in the null sperm (Figure 3E). This observation suggested that an α_1E-activating/potentiating process during capacitation is required for increasing open probability and/or conductance. In support of this, puffing 2-OHCD, which stimulates capacitation by mediating sterol efflux, onto non-capacitated sperm resulted in Ca²⁺ transients with similar characteristics to the addition of G_M1 or CTB (Figure S2A and B), suggesting a common mechanism to channel modulation by sterol efflux and focal ganglioside enrichment. Indeed, sterol efflux would be predicted to be the most physiologically relevant means of modulating G_M1 dynamics within the plasma membrane. Our observations are consistent with physiological regulation of sperm function through sterol efflux and G_M1, and are explored further below.

In most cases, sperm showing the transient increase also demonstrated a secondary Ca²⁺ mobilization typically propagating from the connecting piece towards the AA as waves, and resulting in a sustained elevation in cytosolic Ca²⁺ levels (Figure 4A). We observed these Ca²⁺ waves in over 70% of both the CD1 or α_1E-null sperm (Figure 4B). Interestingly, 40% of all control sperm underwent AE following the CTB-mediated Ca²⁺ wave, whereas only 5% of the α_1E null sperm demonstrated subsequent AE (as determined by the loss of fluorescence in the AA, Figure S3 and Movie S1). These results suggest that in response to clustering of G_M1 with CTB, cytosolic Ca²⁺ elevation by itself is required but is not sufficient for induction of AE. In addition, although α_1E null mice demonstrated high Ca²⁺ wave occurrence, they had significantly decreased AE percentage in response to ZP addition (Figure S2A), consistent with a central role for Ca_V2.3-mediated transients to facilitate AE.

Figure 4. α_1E-null sperm have severe defects in ability to undergo AE and altered Ca²⁺ wave dynamics.

A) Time line analysis of a CD1 sperm loaded with Fluo4-AM, representing the changes in Ca²⁺ over time as they occur along the line indicated in the left panel. Brighter colors correspond to higher Ca²⁺ concentrations. In the time line panel, white arrows indicate a Ca²⁺ transient (left arrow) followed by a Ca²⁺ wave (right arrow). B) Summary of CTB-induced Ca²⁺ wave occurrence in CD1 or α_1E KO sperm incubated under capacitating conditions (5.5mM glucose and 2mM 2-OHCD), with relative occurrence of AE (cells that demonstrated AE as percentage of cells with Ca²⁺ waves; yellow area). C) Statistical analysis of wave rise time, calculated from a single exponent fit to individual traces. D) Average Ca²⁺ elevation at wave peak (relative to baseline Ca²⁺). E) Integrated Ca²⁺ influx during the rise phase of Ca²⁺ waves. Data are presented as AVG±SEM (n>25). Sperm from 3 or more mice were imaged for each experimental condition. p values are 0.05>**>0.01. See also Figure S3 and Movie S1.

Both CD1 and α_1E null sperm demonstrated comparable lag time in Ca²⁺ wave initiation upon CTB addition (89.1±10.2 and 89.9±10.0 sec, respectively). However, the rise time of Ca²⁺ waves was significantly shorter in the α_1E null sperm (502±77 msec) compared to CD1 sperm (842±108 msec; p<0.05 Figure 4C). Although the relative fluorescence increase was slightly but significantly higher in the α_1E null sperm (p=0.05, Figure 4D), the waves in these cells diminished faster resulting in a 2-fold decrease in the integrated Ca²⁺ influx compared to CD1 sperm upon CTB addition (Figure 4E). Finally, we found comparable wave velocities in CD1 and α_1E null sperm (5.18±0.56 and 4.5±0.58 µm/sec respectively), consistent with an intact Ca²⁺-induced Ca²⁺ release mechanism that supports wave propagation independently of Ca_V2.3 activation. Together, our data suggest that the spatial and temporal information contained within the transients either influences the resulting wave and/or locally changes the subsequent wave’s ability to facilitate/promote AE. The spatial and temporal relationship between the transient and wave are therefore critical in determining whether the flux will result in AE.

The α₂δ₁ subunit is involved in G_M1-mediated sperm Ca²⁺ flux

In light of the partitioning of α₂δ into membrane rafts (e.g. Robinson et al., 2011), and because both G_M1 and α₂δ are similar in having highly charged extracellular sugars, we hypothesized that this subunit might play a significant role in channel modulation. First, we confirmed the expression of α₂δ_1/2/4 in developing male germ cells using RT-PCR (data not shown). Next, we localized α₂δ isoforms expressed in sperm, finding both α₂δ₁ and α₂δ₄ in the AA (Figure 5A–B), with higher correspondence of α₂δ₁ with α_1E. To verify the specificity of α₂δ₁ localization, we performed a peptide block of the antibody by pre-adsorbing it with the peptide against which it was generated (Figure S4). The localization patterns of the α₂δ subunits were only marginally different in the α_1E null sperm (Figure 5B), suggesting coupling to other α₁ subunits reported in mouse sperm (Westenbroek and Babcock, 1999). To investigate a possible functional role of α₂δ₁ in G_M1 regulation of sperm Ca²⁺ flux, we utilized the α₂δ_1/2-specific inhibitor, gabapentin (GBP, Taylor, 2009). A 20–40 minute exposure to GBP (100µM) induced a significant reduction in Ca²⁺ transient occurrence and AE (Figure 5C–D), which taken together with our localization results, suggest that α₂δ₁ subunits are likely involved in regulating Ca²⁺ transients and demonstrate again the strong functional association between Ca²⁺ transients and AE.

Figure 5. The α₂δ₁ subunit is involved in G_M1-mediated sperm Ca²⁺ flux.

A) Representative images of localization of α₂δ subunits. B) Quantification of α₂δ localization patterns for wild-type (wt) verus α_1E null sperm (data collected from 4 wt and 2 α_1E null mice with >110 sperm for each). Different regions of the sperm head are abbreviated as indicated in the inset: AA-apical acrosome, PA-post acrosome, ES-equatorial segment, ND-not detected. C) Ca²⁺ transient occurrence in CD1 sperm stimulated with G_M1 (blue bar, n=42 from 3 mice) or with G_M1 in the presence of 100µM GBP (white bar, n=32 from 3 mice). D) G_M1-induced Ca²⁺ wave occurrence with or without GBP (100µM), with relative occurrence of AE (cells that demonstrated AE as percentage of cells with Ca²⁺ waves denoted by yellow shading). E) Six days after cRNA injection with α_1E (2 ng/oocyte), β_2A (5 ng/oocyte), and/or α₂δ (5 ng/oocyte), Xenopus oocytes were two electrode voltage clamped. Ca²⁺ currents were elicited from a holding potential of –80 mV by voltage steps applied at 5-sec intervals to test potentials between –35 to +45 mV in response to 160 ms pulse duration. G/V curves of α_1E and β_2A (open circles) or Ca_V2.3 (closed circles). F) Effect of G_M1 on G/V curves (Ca_V2.3 red-filled circles, α_1E + β_2A open circles). See also Figure S4.

To investigate the molecular mechanism by which Ca_V2.3 is modulated by G_M1, we utilized the Xenopus oocyte heterologous expression system, injecting various combinations of cRNA of α_1E, α₂δ₁ and β₂A into stage V–VI Xenopus oocytes (Note that both β1 and β2 have been reported in the AA (Serrano et al., 1999)). Six days later we utilized the two electrode voltage clamp technique to measure Ca²⁺ currents elicited by various test potentials. As expected (Qin et al., 1998), co-expression of the α₂δ₁ and α_1E subunits in the oocytes resulted in a moderate reduction in voltage sensitivity (as demonstrated by the right shift of the G/V curve) of the channel conductance (Figure 5E). However, the G/V relationship of Ca_V2.3 clearly showed a significant left shift in the presence of G_M1 (Figure 5F). The substantial left shift in α_1E voltage dependence in the presence of α₂δ₁ strongly suggests that this subunit is required for mediating the activation of the functional channel complex by its lipid environment, specifically by G_M1.

Because these experiments were conducted under voltage clamp conditions, the data collected were insensitive to any additional effects G_M1 might be exerting on membrane potential. Therefore, we investigated sperm membrane potential and found that G_M1 indeed caused significant depolarizations. When data from single sperm incubated with 3mM 2-OHCD and then exposed to 25µM G_M1 were analyzed as a single population, the G_M1 induced average potential changes of +23.1mV±15 mV. However, we observed a larger depolarization (+30.3mV±15 mV) in the capacitated subpopulation (as indicated by a more hyperpolarized membrane potential of <−40mV (De La Vega-Beltran et al., 2012)) versus those being less hyperpolarized (>−40mV and considered non-capacitated, +18.9mV±14 mV, p=0.03 for comparison of randomly chosen paired samples, or p=0.02 if analyzed as unpaired samples; Figure S4). Together, these data show that not only did G_M1 increase voltage sensitivity, making it more likely that the channel would open under more negative potentials, but it also shifted the membrane to a more depolarized state, and this change was enhanced in cells that had undergone sterol efflux.

Discussion

It is now recognized that the exquisite regulation of sperm function by lipids makes these cells useful models for fundamental studies of how lipids regulate cell function (Levental et al., 2011). As in sperm, sterols and G_M1 play key regulatory roles in many physiological and pathological processes. Though examples include some of our most pressing medical problems ranging from diabetes to Alzheimer’s disease to atherosclerosis, our understanding of the underlying molecular mechanisms through which lipids act in these processes is at a nascent stage. It is in the context of this larger challenge of cell biological understanding that we have explored how lipids control sperm function in fertilization, the initiating process of developmental biology.

Positive regulation of sperm function by lipids has been known for decades in the form of a strict requirement for sterol efflux. Conversely, SVS2 provides negative regulation, binding G_M1 and retarding the capacitation process. Our investigations revealed that these lipids modulated Ca_V2.3 activity, which mediates a transient rise in Ca²⁺ that spatiotemporally encodes information required for subsequent AE. Mechanistically, we show that this functional regulation depended both on sugar and lipid components of G_M1 and multiple subunits of Ca_V2.3.

The identity of sperm Ca²⁺ channels involved in fertilization has been controversial for some time, though the nature of that controversy has recently changed. High expression of Ca_V2.3 message in developing male germ cells (Lievano et al., 1996), was at odds with a lack of electrophysiological activity in those immature male germ cells (Arnoult et al., 1998; Sakata et al., 2001). Similarly, reports from Ca²⁺ imaging and pharmacologic studies in mature sperm that suggested possible Ca_V2.3 activity (Sakata et al., 2002; Wennemuth et al., 2000), stood in contrast with recent patch clamping experiments that show only CatSper activity (Lishko et al., 2011; Ren and Xia, 2010).

Several methodological differences can account for discrepancies between data obtained by patch clamping and our data (and those of others). First, reports that only CatSper (Kirichok and Lishko, 2011) or CatSper and Ksper activity (Zeng et al., 2013) are detectable note that channels gated or sensitized by other factors (e.g. changes in lipid composition such as sterol efflux or alteration of membrane rafts) would be missed. Second, sperm have scant cytoplasmic space and the membrane is tightly affixed to the underlying cytoskeleton (Noiles et al., 1997). Therefore, patching of mouse sperm is typically performed on residual blebs of cytoplasm lost from normal sperm, and these cells are also often swollen by incubation in hypotonic solutions. Channel activities dependent upon membrane organization or interactions with intracellular partners would likely be perturbed. Third, data presented here were collected at 37°C, whereas patch clamping was performed at room temperature (22–25°C), and phase transitions occur in sperm membranes between these temperatures (Holt and North, 1984; Wolf, 1990). This is potentially an important point for membrane lipid regulation of the channel, and one that should be investigated further using biophysical approaches. Fourth, our data were collected directly from the heads of motile, live sperm as opposed to cytoplasmic or proximal droplets from immotile, immature sperm. Lastly, the signal from a localized transient might be diminished in a whole cell patch at the cytoplasmic droplet (approximately 35 µm away). This distance could impart “space clamp” issues that arise from the fact that current signals are summed from areas with heterogeneous membrane potentials due to non-linear clamping of membrane domains with varying conductance (Bar-Yehuda and Korngreen, 2008). This problem is exacerbated in cells with elongated morphologies, such as sperm, in which multiple membrane micro- and macro-domains and diffusion barriers would have to be crossed for a transient in the apical acrosome to be detected in a cytoplasmic droplet at the flagellar annulus. Our finding of capacitation-dependent, lipid regulation of Ca_V2.3 provides a clear reason why Ca_V2.3 activity largely escaped detection previously, potentially resolving both sets of discrepancies in the literature.

These findings led us to generate the following model (Figure 6) of membrane lipid regulation of sperm function: 1) Under normal physiological conditions, G_M1 and the α_1E subunit co-localize to the fusogenic AA. 2) Binding of G_M1 to components of seminal plasma such as SVS2 provides negative regulation, preventing the interaction of G_M1’s extracellular sugars with the α₂δ subunit. 3) High concentrations of membrane sterols reduce membrane fluidity and lateral diffusion of G_M1, further decreasing the probability for stimulatory interactions. 4) During capacitation, SVS2 is lost from the plasma membrane. 5) Sterol efflux causes changes in membrane fluidity and dynamics promoting or altering interactions between G_M1 and the Ca_V2.3 α₂δ₁ subunit, in addition to G_M1’s local depolarizing effect on the membrane, leading to a focal Ca²⁺ transient. 6) That transient primes the sperm for a subsequent wave, originating from the direction of the posterior ring/connecting piece/flagellum, to facilitate AE.

Figure 6. Model for the regulation of sperm Ca²⁺ flux by the membrane lipid environment.

This model builds upon reports that sterol efflux can activate CatSper via alkalinization and hyperpolarization, and reports that SVS2 inhibits capacitation by binding G_M1. Data presented in the current paper demonstrate that G_M1 functionally regulates the activity of the Ca_V2.3 channel and that this is dependent upon the α₂δ subunit. Transient Ca²⁺ rise through the α_1E pore-forming subunit spatially and/or temporally encodes information that allows a subsequent Ca²⁺ wave to result in acrosome exocytosis. Future experiments will be needed to determine how the transient interacts with the wave and/or modulates the local membrane environment to enable fusion and exocytosis. In the figure, voltage dependent inactivation is abbreviated as “VDI.”.

In support of this last step of the model, our data showed that although alternative, compensatory G_M1-Ca²⁺ mobilization pathways were present in sperm lacking Ca_V2.3, those transients were mis-localized and did not enable subsequent waves to culminate in AE. One possible explanation for the importance of the spatiotemporal properties of the transient is that it is involved in the assembly or priming of a functional excitosome complex consisting of SNAREs, synaptotagmin and Ca_V2.3 (Cohen and Atlas, 2004) which is required to enable acrosome and plasma membrane fusion in AE.

Our model could include alternative cellular events that would also increase G_M1-Ca_V2.3 interactions. For example, point fusion events between the plasma membrane and acrosomal membrane are believed to occur late in capacitation and early in AE (Jin et al., 2011; Kim and Gerton, 2003). These fusions might result in lipid transfer between these membranes, providing additional G_M1 in the APM or changing lipid dynamics within the plane of the plasma membrane. Note that the multiple possible factors promoting G_M1-Ca_V2.3 interactions (SVS2 removal, sterol efflux, membrane fusions leading to more G_M1 on the surface) are not mutually exclusive.

Through one or more of these factors, clustering of G_M1 might promote voltagegated Ca²⁺ current in several ways: 1) modification of upstream or secondary signaling molecules (Maehashi et al., 2003); 2) direct interaction with channel subunits (Zhang et al., 2005); 3) induction of a local change in membrane potential by accumulation of negative surface charges (Piret et al., 2005); 4) local increase in driving force of current flow through open channels by concentrating cations, including Ca²⁺, in the extracellular space immediately adjacent to the plasma membrane due to the negative charge provided by G_M1’s sialic acid residue; or 5) alteration of the electric field sensed by the channel’s gating elements induced by the sialic acid (Bennett et al., 1997).

One aspect of VGCC regulation that is in accordance with our data has been well established; namely, the role of the highly glycosylated, extracellular auxiliary α₂δ subunit in increasing channel conductivity (see Davies et al., 2007 for review). Segregation of the α₂δ subunit to membrane rafts has been hypothesized to prevent full activation of the channel until raft-association of the pore-forming subunit occurs (Davies et al., 2006). In contrast to some reports on its chronic effects (e.g. Hendrich et al., 2008), but in agreement with others (e.g. Peng et al., 2011), our data revealed an acute inhibition of Ca²⁺ transients and AE by GBP, which together with our voltage clamp data support a hypothesis in which the G_M1-Ca_V2.3 regulatory interaction depends in part upon the α₂δ₁ subunit. Like the α₂δ subunit itself, G_M1 has both extracellular sugars and a domain within the membrane, and both G_M1 and sterols are focally organized in dynamic raft microdomains in sperm (Selvaraj et al., 2009; Asano et al., 2009). Our finding that depolarization alone was not sufficient to activate the Ca_V2.3 channel-mediated transients (Figure S2), suggests that the α₂δ might be segregated from the α_1E subunit in the sperm head, leading to a significant reduction of the voltage sensitivity of the channel as evidenced by our voltage clamp data. In our growing appreciation of the functional effects of lipids, it is clear that not only are the species of lipid important, but also the organization and dynamics of those lipids in the local membrane microenvironment.

Our studies have answered some long-standing questions in development, such as how sterol efflux or the ganglioside G_M1 can control whether a sperm can fertilize. We revealed surprising spatiotemporal complexities in how Ca²⁺ transients or waves can or can’t lead to exocytosis. Mechanistically, we defined requirements for the sugar and lipid components of G_M1 to regulate the activity of the Ca_V2.3 channel in sperm, and using a reductionist, heterologous expression system, we showed that this interaction requires the α₂δ₁ subunit of the channel. These improvements in understanding the molecular nature of how lipids can alter cell function might lead to new pharmacological targets or manipulations not only in sperm, but also in the many other cell types in which lipids regulate physiological and pathological processes.

Experimental Procedures

Reagents and Animals

All reagents were purchased from Sigma (St. Louis, MO), unless otherwise noted. Monoclonal antibody against mouse α_1E, α₂δ subunits, and blocking peptide were purchased from Santa Cruz (Santa Cruz, CA) or AbGent (AbGent, SD). Goat anti-mouse serum secondary antibodies (Alexa-488 or Alexa-555 conjugated) were from Invitrogen (Carlsbad, CA). Male CD-1 mice were purchased from Charles River Laboratories (Kingston, NY). Male and female B6129SF/J mice were purchased from Jackson laboratories (Bar Harbor, ME). Sperm from young males (3–6 months) were used for all assays of function, but retired breeder CD-1 mice were used for preliminary studies and for some localization studies. All animal procedures were performed under the guidelines of the Institutional Animal Care and Use Committee at Cornell University.

Sperm Collection and Handling

For murine sperm, a modified Whitten’s medium (MW; 22mM HEPES, 1.2mM MgCl₂, 100mM NaCl, 4.7mM KCl, 1mM pyruvic acid, 4.8mM lactic acid hemi-Ca²⁺ salt, pH 7.35; (Travis et al., 2004)) was used for all incubations. Glucose (5.5mM), NaHCO₃ (10mM), and 2-hydroxypropyl-β-cyclodextrin (2-OHCD; 3mM) were supplemented as needed. Cauda epididymal murine sperm collection was described previously (Travis et al., 2001). All steps of collection and washing were performed at 37°C using MW medium, using methods to minimize membrane damage. After the initial washes but prior to experimental incubations, motility assessment was carried out, and samples showing <60% motility were not used.

Sperm Capacitation and Induction of AE

2×10⁶ sperm were incubated in 300µl of MW with glucose (base media) as non-capacitating media or MW base media supplemented with both 10mM NaHCO₃ and 1mM 2-OHCD as capacitating conditions (pH=7.35). The dead spaces of tubes used for all incubations were filled with nitrogen to avoid the generation of bicarbonate anions in the aqueous media. Ca²⁺ channel inhibitors were added 10 minutes prior to the addition of AE agonists. Progesterone (3μg/ml final concentration, in DMSO) was used as a positive control to induce AE in capacitated murine sperm after 50 minutes of incubation. CTB (1.5μM final concentration, in PBS) was added where indicated to assess its affect on AE. The exogenous lipids ceramide, asialo-G_M1, and G_M1 (in DMSO; 25mM stock concentration) were also added (25μM) where indicated. Sperm were then processed for Coomassie assessment of AE as described previously (Visconti et al., 1999). Groups were compared using a Kruskal-Wallis/Wilcoxon test in Kaleidagraph and SAS v9.3.

Single Cell Ca²⁺ Imaging

Sperm from single mice were incubated with glucose (5.5mM) and fluo4-AM (0.5µM) for 1h at 37°C, with 2-OHCD (2mM) included where indicated. Following capacitation and dye loading, sperm were plated on poly-D-lysine-coated 35mm coverslip dishes (MatTek Corp. Ashland, MA), and 2ml of warm MW base media supplemented with 20mM CaCl₂ were added. The dish with the sperm was then mounted on a Zeiss 510 microscope with a heated 100X oil immersion objective (37°C). Fluo4 was excited using the 488 nm line of a krypton/argon laser and viewed with a 505–550 nm BP filter. Selected cells were approached with a ~5 μm-diameter pulled-glass capillary, positioned within 100 μm from the cell and prefilled with stimulating solution. The stimulating solutions were made at 5X the normal concentration (e.g. G_M1 or A-G_M1=125µM, CTB=25µM), to compensate for the diffusion distance between the capillary and the cell. Sperm cells were imaged at 5–50 Hz while applying a puff of 10s/5psi from the pipette, controlled by a Picospritzer III (FMI Medical Instruments). Unless otherwise stated, all data are presented as mean±SEM. For each of the experiments, the number of cells analyzed (n) is presented in the text. Data were processed and plotted using Origin 8 (OriginLab) and Excel (Microsoft). Statistical comparisons were performed using either a Student’s t-test or a Wilcoxon test with indications as follows: *>0.05, **>0.01, ***<0.005.

Expression of Ca_V2.3 and Two-Electrode Voltage Clamp in Xenopus Oocytes

Human α_1E Ca_V2.3 (#L27745) and rat β_2A (#M80545) were kindly donated by N. Qin and L. Birnbaumer (University of North Carolina, Chapel Hill, NC). Rabbit skeletal α₂δ₁ (#M86621) was from A. Schwartz (University of Ohio, Athens, OH). Expression of channel subunits and two-electrode voltage clamp were preformed as previously described (Cohen and Atlas, 2004). Bath solutions contained 5 mM Ca(OH)₂, 50 mM N-methyl-D-glucamine, 1 mM KOH, 40 mM tetraethyl ammonium, 5 mM HEPES, titrated to pH 7.5 with methanesulfonic acid CH₃SO₃HE. Experiments were carried out 3 times, with n>8 for each group in every experiment. G_M1 was added to the bath solution prior to current recording. Peak current and G/Gmax values were analyzed by Clampfit 9.

Highlights.

• Ca_V2.3 mediates spatiotemporal regulation of acrosome exocytosis (AE)

• G_M1 regulates sperm Ca²⁺ influx, AE & fertilization via Ca_V2.3’s α_1E & α₂δ₁ subunits

• Mechanism of α_1E regulation involves G_M1’s lipid & sugar moieties & α₂δ₁