Calcium-induced Acrosomal Exocytosis Requires cAMP Acting through a Protein Kinase A-independent, Epac-mediated Pathway*

Epac, a guanine nucleotide exchange factor for the small GTPase Rap, binds to and is activated by the second messenger cAMP. In sperm, there are a number of signaling pathways required to achieve egg-fertilizing ability that depend upon an intracellular rise of cAMP. Most of these processes were thought to be mediated by cAMP-dependent protein kinases. Here we report a new dependence for the cAMP-induced acrosome reaction involving Epac. The acrosome reaction is a specialized type of regulated exocytosis leading to a massive fusion between the outer acrosomal and the plasma membranes of sperm cells. Ca2+ is the archetypical trigger of regulated exocytosis, and we show here that its effects on acrosomal release are fully mediated by cAMP. Ca2+ failed to trigger acrosomal exocytosis when intracellular cAMP was depleted by an exogenously added phosphodiesterase or when Epac was sequestered by specific blocking antibodies. The nondiscriminating dibutyryl-cAMP and the Epac-selective 8-(p-chlorophenylthio)-2′-O-methyladenosine-3′,5′-cyclic monophosphate analogues triggered the acrosome reaction in the effective absence of extracellular Ca2+. This indicates that cAMP, via Epac activation, has the ability to drive the whole cascade of events necessary to bring exocytosis to completion, including tethering and docking of the acrosome to the plasma membrane, priming of the fusion machinery, mobilization of intravesicular Ca2+, and ultimately, bilayer mixing and fusion. cAMP-elicited exocytosis was sensitive to anti-α-SNAP, anti-NSF, and anti-Rab3A antibodies, to intra-acrosomal Ca2+ chelators, and to botulinum toxins but was resistant to cAMP-dependent protein kinase blockers. These experiments thus identify Epac in human sperm and evince its indispensable role downstream of Ca2+ in exocytosis.

and of amylase from parotid and pancreatic acinar cells (4). In sperm, various activation pathways required to achieve egg-fertilizing ability depend on the intracellular rise of cAMP. These include changes in motility (5), capacitation (6 -9), and the acrosome reaction (AR) 2 (10,11). The AR is a secretory event that is completed by sperm of many species at an early stage of fertilization, when sperm contact glycoproteins of the egg's extracellular matrix, or zona pellucida (ZP) (12). Ca 2ϩ is an essential mediator of the AR. Two phases of ZP-evoked Ca 2ϩ responses have been described, a first, transient phase mediated by voltage-gated channels and a second, sustained phase in which Ca 2ϩ permeates into the sperm cytosol through store depletion-activated channels (SOC) (13). In human sperm, the acrosome behaves as the internal store of releasable Ca 2ϩ (14,15). Ca 2ϩ is released from the acrosome in two phases. The first precedes (and drives) the opening of SOC channels on the plasma membrane, whereas the second follows this opening, taking place later in the fusion cascade, once the biochemical machinery for fusion has been assembled (14,16). In an attempt to refine our understanding of the cAMP-dependent signaling pathways during the AR, we resorted to a plasma membrane streptolysin O (SLO) permeabilization protocol developed in our laboratory (17). Incubating permeabilized sperm with Ca 2ϩ resembles the physiological situation of Ca 2ϩ influx through open SOC channels. Thus, ours constitutes a particularly attractive system to examine relatively late steps of the exocytic cascade, occurring after the sustained Ca 2ϩ influx, while bypassing earlier pathways whose end point is the opening of SOC channels.
In mammalian cells, cAMP is synthesized by a family of nine transmembrane and one soluble adenylyl cyclase (18 -20). The latter, known as sAC, appears to be the predominant form of adenylyl cyclase in sperm. Its direct activation by bicarbonate is thought to be responsible for the cAMP-induced changes in motility and AR mentioned above (21)(22)(23). sAC defines a novel means for generating cAMP, implying that the second messenger can be generated at a distance from the membrane, closer to its required site of action, and circumventing the need for diffusion to reach distant targets. Models whereby cAMP can signal in a complex consisting of both sAC and effectors have been proposed (24). The best characterized cAMP effectors include cAMP-dependent protein kinase (PKA), cyclic nucleotide gated channels, guanine nucleotide exchange factors (25), and the transcription factor cAMP-response element-binding protein (26). For many years, major intracellular effects of cAMP were believed to be mediated by PKA. PKA is composed of two separate polypeptides, the catalytic (C) and regulatory (R) subunits that interact to form an inactive tetrameric holoenzyme (R 2 C 2 ). PKA activation is achieved by binding of four molecules of cAMP to the R subunits, which induces a conformational change in the R subunits and leads to the dissociation of the holoenzyme into two free, catalytically active, C subunits and an R subunit homodimer (27,28). PKA has been described in spermatozoa from several mammalian species. Furthermore, it has been implicated in sperm motility and capacitation through still undefined signaling cascades that culminate in enhanced protein tyrosine phosphorylation. Inhibitor studies have implicated PKA in the ZP-triggered (29) and progesterone-triggered (30) AR of human sperm. It has been proposed that activation by PKA-dependent phosphorylation of Ca 2ϩ channels on the outer acrosomal membrane leads to an increase in cytosolic Ca 2ϩ and, consequently, to the AR (11).
The second well known targets for cAMP are the cyclic nucleotidegated channels. These channels are directly opened by either cAMP of cGMP and are permeable to Ca 2ϩ ions. They form heterooligomeric complexes composed of at least two distinct subunits (␣ and ␤) (31). Both subunits have been found in the flagellum of mammalian sperm and implicated in motility (32).
The most recently described cAMP effector is Epac (exchange protein directly activated by cAMP). Epac is a Rap-specific guanine-nucleotide exchange factor that is activated by the binding of cAMP to a cyclic nucleotide monophosphate-binding domain (33,34). Two isoforms, Epac1 and Epac2, were described in mammalian cells, both containing a regulatory and a catalytic region in the N-and C-terminal portions of the protein, respectively. The regulatory domain contains the cAMP binding site, which autoinhibits the catalytic activity in the absence of cAMP (35). Evidence suggesting a crucial role for Epac in exocytosis first arose from the lack of effect of specific blockers of PKA-and cyclic nucleotide-gated channels on this process (1, 3, 36 -39).
In light of the recently described multiplicity of proteins with which cAMP interacts, functions previously ascribed solely to PKA may need reevaluation. Specifically, we were interested in the possibility of an Epac-mediated, cAMP-dependent signaling pathway in the AR. Here we report a requirement for Epac in Ca 2ϩ -induced acrosomal release in human sperm. Furthermore, we demonstrate that the activation of Epac alone by a specific cAMP analogue is sufficient to achieve maximum exocytosis levels in intact and SLO-permeabilized cells. This exocytosis relies on the bona fide machinery required for fusion in all secretory cells. Epac functions in a relatively early step during the exocytosis cascade, prior to tethering by Rab3, priming by NSF/␣-SNAP, docking by SNAREs, and intra-acrosomal Ca 2ϩ release.
Recombinant Proteins-A plasmid encoding the light chain of botulinum toxin E (His 6 -BoNT/E-LC) in pQE9 was generously provided by Dr. T. Binz (Medizinische Hochschule Hannover, Hannover, Germany). The DNA was transformed into Escherichia coli XL1-Blue (Stratagene, La Jolla, CA), and protein expression was induced overnight at 20°C with 0.2 mM isopropyl-␤-D-thiogalactoside. A plasmid encoding His 6 -NSF in pQE9 (Qiagen) was a kind gift from Dr. S. Whiteheart (University of Kentucky, Lexington, KY). This construct was transformed into E. coli M15pRep4 (Qiagen), and protein expression was induced for 4 h at 30°C with 1 mM isopropyl-␤-D-thiogalactoside. Purification of recombinant proteins was accomplished according to The QIAexpressionist (available on the World Wide Web at www. qiagen.com), except that 0.5 mM ATP, 5 mM MgCl 2 , and 2 mM dithiothreitol were added to all buffers involved in the purification of His 6 -NSF. The expression plasmid pGEX2T containing the cDNA-encoding human Rab3A was generously provided by Drs. M. Colombo and P. Stahl (Washington University, St. Louis, MO). Glutathione S-transferase-Rab3A was expressed in E. coli strain XL1-Blue (Stratagene) and purified on glutathione-Sepharose following the manufacturer's instructions.
SLO Permeabilization and AR Assay-After at least 2 days of abstinence, semen samples were provided by masturbation from healthy volunteer donors who were free from sexually transmitted diseases. Semen was allowed to liquefy for 30 -60 min at 37°C. Highly motile sperm were recovered by swim-up separation for 1 h in gamete preparation medium (GPM; Serono, Aubonne, Switzerland) at 37°C in an atmosphere of 5% CO 2 , 95% air. The composition of GPM is based on Earle's balanced salt solution (0.2 g/liter CaCl 2 , 0.4 g/liter KCl, 0.097 g/liter MgSO 4 , 6.8 g/liter NaCl, 2.2 g/liter NaHCO 3 , 0.14 g/liter NaH 2 PO 4 ⅐H 2 O, 1 g/liter D-glucose, 0.01 g/liter phenol red) supplemented with 0.1 g/liter sodium pyruvate and 1 mg/ml human serum albumin. GPM contains no antibiotics. The pH and osmolality were maintained within the ranges 7.2-7.4 and 278 -288 mosM, respectively. After swim up, sperm concentration was adjusted to 5-10 ϫ 10 6 /ml, and the cells were incubated for at least 2 h under conditions that support capacitation (GPM; 37°C, 5% CO 2 , 95% air). Permeabilization was accomplished as previously described (17). Briefly, washed spermatozoa were resuspended in cold PBS containing 0.4 units/ml SLO for 15 min at 4°C. Cells were washed once with PBS, resuspended in ice-cold sucrose buffer (250 mM sucrose, 0.5 mM EGTA, 20 mM Hepes-K, pH 7) containing 2 mM dithiothreitol, inhibitors were added when indicated, and cells were further incubated for 15 min at 37°C. After the addition of stimulants to the sperm suspensions, incubation proceeded at 37°C for 15 min. For the experiments with the photoinhibitable Ca 2ϩ chelator NP-EGTA-AM, SLO-permeabilized sperm were preloaded with 10 M NP-EGTA-AM before incubating in the presence of inhibitors and stimulants as described, except that all procedures were carried out in the dark (supplemental Fig. S1A). Alternatively, NP-EGTA-AM preloaded sperm were incubated first with 0.5 mM CaCl 2 and second with the inhibitors to test, also in the dark (supplemental Fig. S1B). In all cases, photolysis of the chelator was induced at the end of the second incubation by 2-min exposure to a UV transilluminator FBTIV-614 (Fisher). Incubations proceeded for an additional 5 min at 37°C. 10 l of each reaction mixture were spotted on 8-well slides, air-dried and fixed/permeabilized in ice-cold methanol for 30 s. Acrosomal status was evaluated by staining with FITC-coupled PSA according to Ref. 43. At least 200 cells were scored using a Nikon microscope equipped with epifluorescence optics. Negative (no stimulation) and positive (10 M Ca 2ϩ ) controls were included in all experiments. For each experiment, the data were normalized by subtracting the number of reacted spermatozoa in the negative control from all values and expressing the result as a percentage of the AR observed in the positive control.
SDS-PAGE and Immunoblot Analysis-Sperm were washed in PBS, and proteins were extracted in sample buffer (44), separated on polyacrylamide slab gels according to Laemmli (45), and transferred to 0.22-m nitrocellulose membranes (Schleicher & Schuell). Nonspecific reactivity was blocked by incubation for 1 h at room temperature with 5% nonfat dry milk dissolved in washing buffer (PBS, pH 7.6, 0.1% Tween 20). Blots were incubated with the anti-Epac antibody (0.92 g/ml) preblocked or not with 12 nM (0.02 g/ml) immunogenic peptide for 60 min at room temperature. Horseradish peroxidase-conjugated goat anti-rabbit IgG was used as secondary antibody (0.25 g/ml) with 45-min incubations. Excess first and second antibodies were removed by washing five times for 10 min each in washing buffer. Detection was accomplished with an enhanced chemiluminescence system (ECL; Amersham Biosciences) and subsequent exposure to Eastman Kodak Co. XAR film for 5-30 s.
Indirect Immunofluorescence-Cells maintained in GPM for 3 h under capacitating conditions were washed twice with PBS and allowed to air-dry on polylysine-coated, 9-mm round coverslips before fixing/ permeabilizing in 2% paraformaldehyde, 0.1% Triton X-100 in PBS for 10 min at room temperature. Sperm were incubated in PBS containing 50 mM glycine for at least 30 min at room temperature and washed twice with PBS containing 0.4% polyvinylpyrrolidone (PVP; average M r ϭ 40,000; ICN) (PBS/PVP). Nonspecific staining was blocked by incubation in 2% bovine serum albumin in PBS/PVP for 1 h at room temperature. Anti-Epac antibodies were diluted in blocking solution (7.6 g/ml), pretreated or not with blocking peptide (0.1 M, 0.16 g/ml), added to the coverslips, and incubated for 20 h at 4°C in a moisturized chamber. After washing three times with PBS/PVP, TRITC-goat antirabbit IgG (15 g/ml in 0.5% horse serum in PBS/PVP) was added and incubated for 1 h at room temperature. After washing as before, samples were mounted with Gelvatol and stored at 4°C in the dark. Slides were examined with an Eclipse TE3000 Nikon microscope equipped with a Plan Apo 60ϫ/1.40 oil objective and a Hamamatsu Orca 100 camera (Hamamatsu Corp., Bridgewater, NJ) operated with MetaMorph 6.1 software (Universal Imaging, Downingtown, PA). Background was subtracted, and brightness/contrast were adjusted to render an all-or-nothing labeling pattern using Jasc Paint Shop Pro 6.02 (Jasc Software; available on the World Wide Web at www.corel.com).
Protein Determination-Protein concentrations were determined by the Bio-Rad Protein assay in 96-well microplates. Bovine serum albumin was used as a standard, and the results were quantified on a Bio-Rad 3550 Microplate Reader.

Reagents that Mimic cAMP Trigger AR in SLO-permeabilized
Sperm-If the opening of Ca 2ϩ channels elicited by PKA (11) was the only cAMP-dependent signaling pathway during the AR, it would be expected that reagents that augment the intracellular concentration of cAMP would not have an effect in cells where the plasma membrane is permeabilized by SLO, bypassing any requirement for ion channel opening. However, when SLO-permeabilized human sperm were incubated with 1 mM Bt 2 cAMP they underwent acrosomal exocytosis of a magnitude similar to that achieved by 10 M Ca 2ϩ (Fig. 1A). Interestingly, the cAMP endogenously synthesized by sperm must be enough to cause a maximal response, given that 100 M IBMX (a PDE inhibitor) elicited the same level of exocytosis as either Ca 2ϩ or Bt 2 cAMP (Fig. 1A). These data suggest that an alternative mechanism, not related to PKAinduced opening of Ca 2ϩ channels, operates during the cAMP-elicited AR. This cAMP-dependent step must be located downstream of the influx of Ca 2ϩ into the cytosol through SOC channels.
cAMP-induced AR Does Not Require Extracellular Ca 2ϩ -The addition of micromolar concentrations (10 -500 M) of Ca 2ϩ triggers acrosomal exocytosis in permeabilized spermatozoa. The same effect is observed without Ca 2ϩ addition when Bt 2 cAMP is used as inducer (Fig.  1A). Under these conditions, free Ca 2ϩ concentration in the reaction mixture, which contains 0.5 mM EGTA, is on the order of 10 Ϫ7 M (14). When 5 mM EGTA is added to the system, free Ca 2ϩ concentration drops to less than 10 nM. Even at this very low Ca 2ϩ concentration, sperm underwent exocytosis in response to Bt 2 cAMP (Fig. 1B). These results show that the AR elicited by cAMP proceeds without an influx of Ca 2ϩ into the cytosol from the extracellular milieu, lending support to the hypothesis that a mechanism other than the opening of Ca 2ϩ channels by PKA mediates the exocytosis brought about by cAMP.
Ca 2ϩ Induces AR in Permeabilized Sperm through a cAMP-mediated Pathway-Next, we asked if cAMP and Ca 2ϩ initiate independent cascades or achieve acrosomal release through shared signaling pathways. Hence, we investigated whether Ca 2ϩ -triggered exocytosis requires cyclic nucleotides. To this end, we destroyed endogenous sperm cAMP by loading permeabilized cells with 2 g/ml catalytically active, cAMPspecific PDE before inducing the AR. This pretreatment caused an 80% reduction of the exocytotic response compared with untreated samples (Fig. 1C), suggesting that Ca 2ϩ requires cAMP to stimulate exocytosis. In other words, Ca 2ϩ and cAMP do not operate independently to activate the AR cascade.
AR in Permeabilized Sperm Is Insensitive to PKA Inhibitors-To eliminate the possibility that cAMP is acting through PKA to modulate the AR, we selected drugs that inhibit the activity of the catalytic subunit of PKA through different mechanisms of action. Preincubation with 10 M H89 did not affect the AR triggered by 1 mM Bt 2 cAMP ( Fig. 2A, gray bars) or 100 M IBMX ( Fig. 2A, dotted bars). Likewise, preincubation with 15 g/ml unmyristoylated peptide inhibitor of PKA (PKI; fragment 6 -22 amide) (46) did not influence the AR triggered by Bt 2 cAMP ( Fig. 2A). These data indicate that challenging with cAMP elicits exocytosis in permeabilized human sperm through a PKA-independent mechanism.
Given that Ca 2ϩ signals through cAMP (Fig. 1C) and that the latter accomplishes exocytosis through a mechanism that does not rely on PKA ( Fig. 2A), it would be expected that the Ca 2ϩ -induced AR would also be independent of this kinase. In fact, when we preincubated permeabilized sperm with 10 M H89, 15 g/ml PKI (Fig. 2B), or 500 M Rp-cAMPS (Fig. 3B, see below) but challenged with Ca 2ϩ instead of Bt 2 cAMP or IBMX, all of the inhibitors had no or very little effect on exocytosis. These data indicate that, as is the case with cAMP, Ca 2ϩ does not utilize a PKA-dependent pathway to mediate acrosomal exocytosis in permeabilized sperm.
PKA inhibitors have been reported to block exocytosis, raising the possibility that the inhibitors were ineffective in our hands. When capacitated intact human sperm were stimulated with 15 M progesterone, AR ensued (Fig. 2C, open bars). Exocytosis was abrogated by preincubation with 10 M H89 (Fig. 2C), in agreement with previous observations of PKA dependence of the AR in intact sperm and eliminating the possibility of drug inactivity. Taken together, our data unveil a new, cAMP-dependent, PKA-independent signaling pathway during acrosomal exocytosis.
Selective Epac Activation Induces Acrosomal Exocytosis-Our findings that the AR in permeabilized sperm requires cAMP but is independent of PKA suggest that another cAMP target(s) might be mediating the effects. Epac is a plausible candidate, and we therefore turned to investigate its role in exocytosis using several different approaches. Bt 2 cAMP has been linked to Epac-mediated responses in whole cells (47,48) but also to those of PKA; therefore, we resorted to the recently designed 8-pCPT-2Me-cAMP Epac-specific cAMP analogue. When nonpermeabilized sperm were exposed to 50 M 8-pCPT-2Me-cAMP, they achieved an exocytotic response similar to that brought about by FIGURE 1. Ca 2؉ and reagents that augment cAMP trigger AR in SLO-permeabilized sperm through a common signaling pathway. A, permeabilized spermatozoa were incubated with 0.5 mM CaCl 2 (10 M free Ca 2ϩ , estimated by MAXCHELATOR, a series of programs for determining the free metal concentration in the presence of chelators; available on the World Wide Web at www.stanford.edu/ϳcpatton/maxc.html) (black bar), 1 mM Bt 2 cAMP (gray bar), or 100 M IBMX (dotted bar) for 15 min at 37°C. B, permeabilized spermatozoa were incubated with or without 5 mM EGTA for 15 min at 37°C. Acrosomal exocytosis was then initiated by adding 0.5 mM CaCl 2 (10 M free Ca 2ϩ ) (black bar) or 1 mM Bt 2 cAMP (gray bars) and a further 15 min of incubation at 37°C. C, 2 g/ml PDE4D was introduced into permeabilized cells and incubated for 15 min at 37°C prior to stimulation with 0.5 mM CaCl 2 (10 M free Ca 2ϩ ) for an additional 15 min. Sperm were fixed, acrosomal exocytosis was evaluated by FITC-PSA binding, and data were normalized (mean Ϯ S.E.) as described under "Experimental Procedures." Statistical analysis is provided in Table S1.  Table S2. the physiological agonist progesterone (Fig. 2C). A cGMP analogue with an identical pCPT substitution in position 8 did not trigger exocytosis when the same concentration was added to sperm (Fig. 2C). These data suggest the presence of Epac in human sperm and its functional role in the AR. In contrast to progesterone, 8-pCPT-2Me-cAMP-elicited AR was insensitive to 10 M H89 (Fig. 2C, dashed bars), pointing to the PKA independence of the Epac-related signaling pathway.
The addition of increasing concentrations of 8-pCPT-2Me-cAMP to SLO-permeabilized sperm induced the AR in a dose-response fashion (Fig. 3A, gray circles). 50 M 8-pCPT-2Me-cAMP achieved the maximum exocytotic response, of a magnitude comparable with that of 10 M free Ca 2ϩ (Fig. 3A, open square). As was the case in nonpermeabilized sperm, the effect was insensitive to 10 M H89 (Fig. 3A, black  circles). We further investigated this matter by applying a drug with a different mechanism of action. Rp-cAMPS competitively inhibits PKA by binding, but not dissociating, the holoenzyme. When added to permeabilized human sperm, 500 M Rp-cAMPS did not prevent the AR triggered by calcium (Fig. 3B, black bars) or 8-pCPT-2Me-cAMP (Fig.  3B, dashed bars). The addition of Rp-cAMPS alone had no effect (data not shown). These data suggest that Epac activation is sufficient to elicit exocytosis in permeabilized sperm, and it does so in a PKA-independent manner. They also strengthen the view that calcium does not rely on PKA activity to accomplish exocytosis.
As expected, 50 M 8-pCPT-2Me-cAMP triggered exocytosis even in the virtual absence of extracellular Ca 2ϩ (Fig. 3B), as shown for Bt 2 cAMP (Fig. 1B). These data indicate that extracellular Ca 2ϩ is not required for 8-pCPT-2Me-cAMP-triggered AR, supporting the notion that Epac governs signaling pathways unrelated to the opening of Ca 2ϩ channels.
Presence and Localization of Epac in Human Sperm-Responsiveness to 8-pCPT-2Me-cAMP is an accepted marker of the presence of Epac in cells. Thus, the inducibility of the AR by this analogue strongly implies that Epac is not only present but also functionally important for human sperm exocytosis. Specific antibodies were used to detect the presence and localization of Epac in human sperm by Western blot and indirect immunofluorescence. As shown on Fig. 3C (anti-Epac lane), the anti-Epac antibodies recognized a single protein band in sperm extracts. This signal was specific, since detection was abolished by preincubation of the probe with the synthetic peptide (Fig. 3C, anti-Epac* lane). The apparent molecular mass of sperm Epac is ϳ100 kDa, corresponding better with the expected molecular mass of human Epac1 (881 amino acids) than that of human Epac2 (1011 amino acids). Indirect immunofluorescence was used to localize Epac on fixed, permeabilized human sperm. The antibody reacted with the sperm head in the acrosomal region (Fig. 3D, top). This pattern was specific, since it was not observed when the antibody had been previously blocked with the synthetic pep-  Table S3. C, whole sperm proteins were extracted in Laemmli sample buffer and analyzed by Western blot with an anti-Epac antibody preblocked (Epac* lane) or not (Epac lane) with 0.019 g/ml (12 nM) Epac peptide. Protein from 7 ϫ 10 6 cells were loaded. M r standards (ϫ 10 3 ) are indicated on the left. D, top, human sperm were fixed, permeabilized, and immunostained with antibodies to Epac (7.6 g/ml) followed with a TRITC-labeled goat anti-rabbit antibody. Bottom, the primary antibody was preblocked by incubation with 0.16 g/ml Epac peptide. Shown are epifluorescence micrographs of typically stained cells (bar, 6 m). tide (Fig. 3D, bottom panels). Taken together, our data indicate that Epac is present in human sperm, localizes to the acrosomal region, and exhibits a positive role in the AR that is independent of PKA and the opening of Ca 2ϩ channels.
Epac Mediates Ca 2ϩ and cAMP-triggered AR-We have shown that the activation of Epac by exogenously added cAMP suffices to elicit the AR in human sperm. We have also shown that Ca 2ϩ requires cAMP to achieve exocytosis. Next, we asked whether Ca 2ϩ relies on Epac to fulfill its role as AR inducer. When added to SLO-permeabilized sperm, 1 g/ml anti-Epac antibody, but not a nonimmune rabbit IgG, effectively abrogated Ca 2ϩ -triggered AR (Fig. 4A). The effect of the anti-Epac polyclonal was abolished by preincubation with the synthetic peptide against which the anti-Epac antibody was raised (Fig. 4A), suggesting that the inhibitory effect of the anti-Epac was specific and due to binding to endogenous Epac. As expected, the AR elicited by both Bt 2 cAMP (Fig. 4B, gray bars) and 8-pCPT-2Me-cAMP (Fig. 4B, dashed bars) was also blocked by the anti-Epac antibody. These results reveal that Epac is indispensable for human sperm acrosomal exocytosis and suggest that cAMP plays its role in Ca 2ϩ -triggered AR through the activation of Epac.
cAMP-induced, Epac-mediated AR Requires a Functional Fusion Machinery-Results from our laboratory show that the AR in permeabilized sperm proceeds through a unidirectional sequence of events put in motion by Ca 2ϩ stimulation (16). The next series of experiments were designed to characterize the cAMP-induced AR of human sperm, defining the place of the activation of Epac in the sequence of events leading to exocytosis.
First, we set out to determine whether the tethering of the acrosome to the plasma membrane elicited by Rab3A takes place before or after Epac activation. Specific antibodies against Rab3A (200 g/ml) were introduced into SLO-permeabilized human sperm before challenging with 10 M free Ca 2ϩ (Fig. 5A, black bars), 1 mM Bt 2 cAMP (Fig. 5A, gray bars), or 50 M 8-pCPT-2Me-cAMP (Fig. 5A, dashed bars). Pretreatment with the anti-Rab3A antibodies inhibited Ca 2ϩ -, Bt 2 cAMP-, and 8-pCPT-2Me-cAMP-triggered AR by 98, 87, and 98% respectively. To test the specificity of this effect, the active sites of the antibodies were blocked by preincubation with recombinant Rab3A before the addition to permeabilized sperm. Inhibition was no longer observed (Fig. 5A, anti-Rab3A* 3 calcium), suggesting that their inhibitory effect was due to binding to endogenous Rab3A. These data indicate that Epac-mediated exocytosis requires Rab3A. Second, we resorted to specific anti-␣-  Table S4.  Table S5. SNAP (50 g/ml) and anti-NSF (whole serum diluted 1:300) antibodies that prevent the AR by sequestering the endogenous proteins and therefore the priming of the fusion machinery. Ca 2ϩ (Fig. 5B, black bars), Bt 2 cAMP (Fig. 5B, gray bars), and 8-pCPT-2Me-cAMP (Fig. 5B, dashed bars) failed to trigger exocytosis in sperm pretreated with these antibodies. Once again, the effect of the antibodies was specific, since they no longer inhibited exocytosis when preincubated with excess recombinant NSF (Fig. 5B, anti-NSF* 3 calcium) or ␣-SNAP (42). BoNT/E is a potent inhibitor of exocytosis due to its highly specific proteolytic cleavage of the SNARE protein SNAP-25. When added to SLO-permeabilized sperm, 300 nM BoNT/E caused a marked inhibition of the Bt 2 cAMP-dependent (Fig. 5B, gray bars) and 8-pCPT-2Me-cAMPdependent (Fig. 5B, dashed bars) AR. Since both cAMP analogues effectively induced the AR in the absence of extracellular Ca 2ϩ (Figs. 1B and 3B), these results suggest that the analogues themselves, through Epac activation, are directly or indirectly able to bring about a correct priming by NSF/␣-SNAP and the SNARE-dependent docking of the acrosome to the plasma membrane. Third, we investigated the requirement of Ca 2ϩ mobilized from the acrosome. To this end, we chelated Ca 2ϩ in the lumen of the acrosome with 10 M BAPTA-AM, a permeant-chelating agent that accumulates in membranebound compartments. This treatment blocked the AR when both Bt 2 cAMP (Fig. 5C, gray bars) and 8-pCPT-2Me-cAMP (Fig. 5C, dashed bars) were used as inducers. Similar results were obtained when the Ca 2ϩ -ATPase pump, responsible for taking up Ca 2ϩ into the acrosome, was inhibited with 10 M CPA (Fig. 5C). We then asked whether cAMP-induced AR requires the release of Ca 2ϩ from the acrosomal store. Loading with 100 M 2-APB, an IP 3 -sensitive Ca 2ϩ channel blocker, precluded exocytosis elicited by Bt 2 cAMP (Fig. 5C, gray bars) and 8-pCPT-2Me-cAMP (Fig. 5C, dashed bars). These data show that cAMP-triggered, Epac-mediated exocytosis depends on the efflux of intra-acrosomal Ca 2ϩ .

cAMP/Epac Govern an Early
Step during the Ca 2ϩ -triggered Exocytotic Cascade-To determine the site of action of Epac when Ca 2ϩ initiates the AR, we resorted to a reversible blocker of exocytosis that prevents the AR by sequestering intra-acrosomal Ca 2ϩ (14). The reversible blocker of choice was the photolabile Ca 2ϩ chelator NP-EGTA-AM. UV photolysis of NP-EGTA releases the caged Ca 2ϩ rapidly and with high photochemical yield (49). In our SLO-permeabilized human sperm model, the membrane-permeable compound NP-EGTA-AM crosses the plasma and outer acrosomal membranes, accumulates inside the acrosome, and thus precludes the availability of intra-acrosomal Ca 2ϩ . UV photolysis of NP-EGTA-AM rapidly replenishes the acrosomal Ca 2ϩ pool, resuming exocytosis (see schematic in Fig. 6A and supplemental Fig. S1). In combination with AR inhibitors, NP-EGTA-AM serves to place the requirement for fusion-related factors before or after the intra-acrosomal Ca 2ϩ -sensitive step. Briefly, NP-EGTA-AM allows an AR inducer to prepare the fusion machinery up to the point when intra-acrosomal Ca 2ϩ is required. Inhibitors are then added, and the tubes are illuminated. Resistance to inhibitors, reflected in unaffected exocytosis, implies that their targets are required upstream of intra-acrosomal Ca 2ϩ efflux. Sensitivity to inhibitors, revealed by blocked exocytosis, means their targets are located after the intraacrosomal Ca 2ϩ -sensitive step (see schematic in Fig. 6A and supplemental Fig. S1). The AR is always prevented when the inhibitors are added prior to the inducer and maintained throughout the experiment. We asked whether the cAMP/Epac-dependent step takes place before or after intra-acrosomal Ca 2ϩ release by loading permeabilized sperm with NP-EGTA-AM and hydrolyzing cAMP to 5Ј-AMP with 2 g/ml PDE (Fig. 6B, gray bars) or blocking Epac function with an antibody (Fig.   6C, gray bars). Both reagents inhibited exocytosis when added before, but not after, challenging with Ca 2ϩ . These results indicate that cAMP/ Epac are necessary early, before Ca 2ϩ is released from the acrosome, in the fusion cascade. This is in complete agreement with data attained when cAMP was used to elicit the AR (Fig. 5C) and suggests a role for cAMP/Epac prior to intra-acrosomal Ca 2ϩ efflux.

DISCUSSION
Elevated cAMP concentrations serve as a switch to activate most if not all of the signaling pathways during sperm maturation, rendering the spermatozoa competent for fertilization, and also the AR. Thus, reagents that increase intracellular cAMP levels, such as the membranepermeable analogues Bt 2 cAMP and 8-Br-cAMP, xanthine PDE inhibitors (to prevent cAMP hydrolysis), and the adenylyl cyclase stimulator forskolin, induce exocytosis in mammalian sperm (50). One of the mechanisms proposed to explain these observations involves PKA activation by cAMP, leading to the opening of Ca 2ϩ channels (11). We now know that it is most likely that the cAMP-mediated signaling mechanism is much more complex than was believed earlier, and many cAMPmediated effects that were previously ascribed to PKA are in fact transduced by Epac. Therefore, it is necessary to reformulate concepts of cAMP-mediated signaling to include the contribution of Epac. To this end, we have used human sperm permeabilized with the bacterial toxin SLO. Selective permeabilization of the plasma membrane was achieved by SLO binding to cholesterol at 4°C (the toxin is inactive at this temperature (51)) and removing the excess toxin by washing and centrifugation. The fraction of SLO that remained bound to the plasma membrane was activated with dithiothreitol, and pore formation was initiated by incubation at 37°C. This procedure not only allows access to intracellular compartments but also preserves the normal reducing cytosolic environment without causing any noticeable functional or structural damage. Furthermore, it achieves regulated acrosomal release and has been used to demonstrate the roles of Rab3A (17), NSF (52), synaptotagmin VI (53), the SNARE complex (16,54), calmodulin (55), protein-tyrosine kinases and phosphatases (56), and ␣-SNAP (42) in the Ca 2ϩ -dependent AR of human sperm. Furthermore, both intact and permeabilized sperm respond to challenge with cAMP analogues (Figs. 2C and 3, A and B) and BoNTs (16,54) in a similar fashion.
In this particular study, permeabilization bestows the additional advantage of uncovering cAMP-dependent, post-SOC channel opening pathways, otherwise masked by the PKA dependence of pre-SOC channel opening in intact sperm (57). Here, we show that reagents that augment or mimic cAMP triggered the AR in a PKA-independent fashion (Figs. 2 and 3). We make such a strong assertion based on data gathered after selecting a variety of inhibitors with different mechanisms of action. Thus, H89 is a competitive inhibitor of ATP (58), whereas PKI competes with protein/peptide substrates (59). Both of these compounds block PKA activity through its catalytic subunit. In contrast, Rp-cAMPS competes with cAMP for its binding sites on the regulatory subunit but, unlike cAMP, is unable to dissociate the holoenzyme (60). In intact sperm, PKA has been claimed to exhibit a positive role in the AR by indirectly opening SOC channels on the plasma membrane (57) and thus allowing a sustained influx of Ca 2ϩ into the cytosol, which ultimately leads to exocytosis (61). Consistent with these observations, PKA is required in intact (Fig. 2C) but not in permeabilized (Figs. 2, A and B, and 3, A and B) sperm. When SLO-permeabilized sperm are suspended in a solution containing Ca 2ϩ , the situation resembles that of open SOC channels. Therefore, we interpret any modulation of exocytosis taking place in SLO-treated cells as indicative of a post-SOC opening step in intact sperm. Given that Epac is required in per- FIGURE 6. cAMP and Epac are required for Ca 2؉ -triggered human sperm AR before intra-acrosomal Ca 2؉ efflux. A, schematic representation of the use of NP-EGTA-AM in permeabilized sperm. Shown is the sequence of the addition of reagents. NP, NP-EGTA-AM; h, photolysis of NP-EGTA-AM by UV illumination and release of intra-acrosomal Ca 2ϩ . B and C, permeabilized spermatozoa were loaded with 10 M NP-EGTA-AM (NP) for 15 min at 37°C to chelate intra-acrosomal Ca 2ϩ . Acrosomal exocytosis was then initiated by adding 0.5 mM CaCl 2 (10 M free Ca 2ϩ ) (calcium). After a further 15 min at 37°C to allow exocytosis to proceed up to the intra-acrosomal Ca 2ϩ -sensitive step, sperm were treated for 15 min at 37°C with 2 g/ml PDE4D or 1 g/ml anti-Epac antibody. All of these procedures were carried out in the dark. UV flash photolysis of the chelator was induced at the end of the incubation period (h), and the samples were incubated for 5 min to promote exocytosis (NP 3 calcium 3 inhibitor 3 h). Sperm were then fixed, and AR was measured as described under "Experimental Procedures." Several controls were included (black bars): background AR in the absence of any stimulation (control); AR stimulated by 10 M free Ca 2ϩ (calcium), inhibitory effect of NP-EGTA-AM in the dark (NP 3 calcium 3 dark) and the recovery upon illumination (NP 3 calcium 3 h); and inhibitory effect of the inhibitors when present throughout the experiment (NP 3 inhibitor 3 calcium 3 h). The data were normalized as described under "Experimental Procedures" (mean Ϯ S.E.). Statistical analysis is provided in the corresponding Table S6. meabilized sperm, we deduce that this protein displays its positive role in exocytosis downstream of SOC opening. Because PKA catalyzes a step prior to the activation of these channels, we would like to suggest that PKA and Epac act in a sequential manner to achieve sperm exocytosis. The fact that the activity of these two cAMP targets is required for exocytosis is not a sperm oddity. For instance, cAMP stimulates exocytosis of two different vesicle pools in melanotrophs, one through the PKA-dependent modulation of Ca 2ϩ channels and the other through the Epac-dependent modulation of the secretory machinery (62). In the calyx of Held, cAMP increases the Ca 2ϩ affinity for secretion (37) and enhances vesicle priming (63) and synaptic potentiation (64) through Epac activation. Epac also regulates transmitter release at the crayfish neuromuscular junction (36) and progesterone secretion by human granulosa cells (48).
The fact that IBMX brings about exocytosis means that the amount of cAMP synthesized by sperm would be high enough to support secretion, were it not hydrolyzed by endogenous PDEs. Ca 2ϩ -induced AR in permeabilized human sperm was mediated by cAMP but was independent of PKA (Figs. 2B and 3B). Instead, it involved Epac. It is currently unknown whether Ca 2ϩ relies on a threshold of cAMP concentrations normally present in unstimulated cells or if it somehow increases cAMP levels by altering the balance between its synthesis by cyclases and degradation by PDEs. Interestingly, Ca 2ϩ stimulation of sAC activity has been well documented (65,66). The new cAMP/Epac-related signaling pathway we have characterized also acts in nonpermeabilized cells (Fig.  2C). However, deciphering it in nonpermeabilized cells remains difficult, because we lack Epac-related AR blockers that permeate through the plasma membrane. For instance, we used antibodies directed against Epac as specific inhibitors (Fig. 4) based on the hypothesis that the binding of an immunoglobulin on or close to an active site of a protein should alter its activity and/or its interaction with other proteins located upstream or downstream the signaling cascade, or, as the case may be here, with cAMP. This antibody was raised against a peptide mapping to within the single cAMP binding domain in Epac1 and cAMP-B in Epac2, corresponding to the highly exposed ␤-sheets 7 and 8 in the latter crystal structure (41). The region bears no similarity with the cAMP binding domains of human PKA RI␣, PKA RII␣, PKA RI␤, or PKA RII␤ (34) or with any other mammalian protein in public data bases, allowing us to regard our antibody as highly specific for Epac recognition. Two commercially available anti-Epac antibodies have been applied in a similar experimental strategy to inhibit the activity of H,K-ATPase in SLO-permeabilized rat kidney cells (67). In sperm, Epac exhibits acrosomal localization (Fig. 3D), consistent with its proposed role in the exocytosis of this granule. A single band corresponding to the molecular mass of Epac1 was detected by Western blot in whole sperm extracts (Fig. 3C). We would like to point out, however, that given that the peptide to which our antibody was raised is conserved in both Epac1 and Epac2, we cannot rule out the possibility that sperm contains an isoform of Epac2 of smaller molecular mass than that described in somatic cells. To the best of our knowledge, ours is the first report on the presence and function of Epac in mammalian sperm. Indirect immunofluorescence experiments had previously suggested a subplasmalemmal localization for Epac in mouse spermatocytes and spermatids (68). Functional roles were not explored in that article. It is unknown at this point whether the apparent discrepancy between the localization of Epac reported by Berruti and us is due to the use of gametes from different species (mouse versus human), different maturation stage (spermatogenic versus mature sperm), or different antibodies (commercial versus custom made).
The first line of evidence for the requirement of Epac in the AR comes from the use of specific antibodies (see above). As shown in Fig. 4, loading of SLO-permeabilized sperm with anti-Epac antibodies reduced the exocytotic response of sperm to Ca 2ϩ and cAMP analogues. We and others have previously shown that the acrosome behaves as a Ca 2ϩ -storing organelle (14,15). Release of intravesicular Ca 2ϩ takes place after Rab3-elicited tethering and SNARE-mediated docking of the acrosome to the plasma membrane and is necessary for human sperm AR. By using a photosensitive Ca 2ϩ chelator in combination with anti-Epac antibodies and recombinant PDE, we have been able to show that cAMP/Epac are required in sperm exocytosis before the release of Ca 2ϩ from the acrosome (Fig. 6). The AR is not a wholesale, instantaneous release of components from a fluid-filled, membrane-bound sack, but rather the acrosomal contents consist of soluble proteins and an insoluble acrosomal matrix, which is the last to be released after membrane fusion. It is this matrix we detect with PSA. We claim that anti-Epac antibodies actually block membrane fusion and not simply acrosomal matrix dispersal, given that exocytosis proceeds in their presence when added after releasing intra-acrosomal Ca 2ϩ with NP-EGTA-AM (Fig. 6).
The second line of evidence comes from the use of the Epac-selective cAMP analogue 8-pCPT-2Me-cAMP. This compound is a very efficient stimulant of Epac in vitro and in vivo, with half-maximal activation achieved at concentrations 15 times lower than those of cAMP itself. Most importantly, this analogue is unable to activate PKA either in vitro or in vivo (69, 70). 8-pCPT-2Me-cAMP triggered the AR in permeabi- lized (Fig. 3, A and B) and intact (Fig. 2C) human sperm in concentrations similar to those used in a whole range of somatic cells. In both cases, exocytosis was resistant to H89 (Figs. 2C and 3A), in agreement with the notions that 8-pCPT-2Me-cAMP does not signal through PKA and that Epac activation is sufficient to accomplish the AR. Rp-cAMPS exhibited a similar behavior (Fig. 3B). Whereas there is wide agreement on the antagonistic role of this analogue on PKA activation, its effects toward Epac are controversial. This analogue binds Epac with low affinity (70). Rehmann et al. (41) have recently referred to their own unpublished observation that Rp-cAMPS does not activate Epac; rather, it inhibits cAMP-induced Epac activation (71). Other laboratories have reported a nonantagonistic role of Rp-cAMPS toward Epac (72), and yet other studies appear to indicate that Rp-cAMPS could activate Epacrelated pathways (3,62,67,73). To help resolve this controversy, we would like to contribute the information that, at least in human sperm, Epac is not activated by Rp-cAMPS. Nor does this analogue antagonize the stimulatory role of cAMP on exocytosis.
8-pCPT-2Me-cAMP and Bt 2 cAMP were presumably acting through Epac activation and not through an influx of Ca 2ϩ into the cytosol, since they elicited acrosomal exocytosis in the virtual absence of Ca 2ϩ in the medium (Figs. 1B and 3B). Under standard conditions, influx of Ca 2ϩ into the cell leads to the activation of Rab3A, which in turn mediates tethering of the acrosome to the plasma membrane. Next, priming by NSF/␣-SNAP takes place, followed by SNARE protein assembly in trans complexes and therefore SNARE-dependent docking of the acrosome. The docking machinery contains or interacts with the Ca 2ϩ sensor synaptotagmin. Upon binding Ca 2ϩ mobilized from the acrosome through IP 3 -sensitive channels, synaptotagmin undergoes a conformational change that ultimately (and indirectly) promotes fusion. Both cAMP analogues used throughout this study required Rab3 (Fig. 5A), ␣-SNAP/ NSF, SNAREs (Fig. 5B), and an efflux of intra-acrosomal Ca 2ϩ (Fig. 5C) to bring about the AR. A model depicting our current thinking on the mechanisms underlying the AR is shown in Fig. 7, where, for simplicity, only SNAREs are drawn. Initially, SNAREs are locked in inactive cis complexes on plasma and outer acrosomal membranes. Upon Ca 2ϩ entrance into the cytoplasm, sAC is stimulated, leading to an increase in cAMP levels. This cAMP binds to and activates Epac. Later on, Rab3A is activated, triggering the tethering of the acrosome to the plasma membrane. Next, ␣-SNAP/NSF disassemble cis SNARE complexes on both membranes. Monomeric SNAREs are free to assemble in loose trans complexes, causing the irreversible docking of the acrosome to the plasma membrane. At this point, Ca 2ϩ is released from inside the acrosome through IP 3 -sensitive Ca 2ϩ channels to trigger the final steps of membrane fusion, which require SNAREs (presumably in tight trans complexes) and synaptotagmin.
In many cell types, an increase in intracellular cAMP concentration regulates Ca 2ϩ -triggered exocytosis (39,62,74). Unlike the situation in human sperm (this work), in most of them, an elevation of cAMP alone in the absence of a Ca 2ϩ rise is not sufficient to trigger exocytosis (75). Nevertheless, a limited range of cells use cAMP as a major trigger for exocytosis. The cAMP-dependent pathways coexist with Ca 2ϩ -dependent ones for exocytosis in these cells, and as we report here, it is likely that they use a common final SNARE-dependent mechanism. For instance, cAMP-triggered exocytosis in the parotid gland, one of the best studied, requires VAMP2. The involvement of other SNAREs or other components of the general fusion machinery in cAMP-triggered exocytosis had not, until now, been investigated (75), with the exception of the pathway cAMP-Epac2-Rim-Rab3-calcium sensor (1,38,76,77). The exact mechanisms by which cAMP/Epac assemble the whole fusion machinery in preparation for exocytosis are currently under investiga-tion in our laboratory. Although this particular aspect has not been specifically addressed yet, other laboratories have suggested various avenues for Epac signaling in a number of biological phenomena. Thus, Epac1 has been shown to utilize pathways involving direct proteinprotein interactions with Rap1 (3,69,70,78) and Epac2 with Rim2 (38), Piccolo (76), and the sulfonylurea receptor 1, a subunit of ATP-sensitive K ϩ channels (77). Activation of Epac1 (79) and Epac2 (2) mobilizes Ca 2ϩ from stores regulated by ryanodine receptors. Furthermore, activation of Epac1 mobilizes Ca 2ϩ through IP 3 -sensitive channels via Rap2B and phospholipase C-related pathways (80). Recently, a Rapindependent cascade was described for Epac1 activation of JNK, involving its Ras-exchanger motif (REM) (rather than its guanine nucleotide exchange factor) domain (81).
The existence of Epac immediately raises many questions regarding the mechanism of cAMP-mediated signaling. Since both PKA and Epac are expressed in sperm, an increase in intracellular cAMP levels will lead to the activation of both enzymes and perhaps other potential cAMP target(s) as well. It is conceivable that whereas PKA acts through a discrete set of signaling pathways, Epac may enlist distinct routes, and the net cellular effects of cAMP are dictated by the sum of these events. Both synergistic and antagonistic effects of PKA and Epac have been claimed to transduce cAMP signaling, depending on the cell type (47,67,70,82,83). It would appear that ours is the first system where PKA and Epac contribute to achieve a biological response, the AR, in a sequential manner, so much so that direct Epac activation seemingly bypasses the need for PKA-mediated protein phosphorylation. In our working model, binding of progesterone (and perhaps also ZP) to a sperm plasma membrane receptor results in PKA activation, presumably through accumulation of cAMP. PKA-mediated protein phosphorylation would contribute to the opening of Ca 2ϩ channels, leading to the influx of this ion into the cytosol. This Ca 2ϩ sets into motion all of the exocytotic machinery. Our findings suggest that it does so through cAMP-mediated Epac activation.