Sperm extract injection into ascidian eggs signals Ca2+ release by the same pathway as fertilization

At fertilization, eggs of most if not all species undergo a transient rise in cytosolic free Ca²⁺(see Stricker, 1999). In eggs of ascidians (Speksnijder et al., 1989, 1990a,b; Brownlee and Dale, 1990; Kyozuka et al., 1998) as well as mammals (Miyazaki et al., 1993; Kline et al., 1999), the initial Ca²⁺rise at fertilization begins at the point of sperm-egg fusion, spreads across the egg in the form of a wave, and is followed by additional Ca²⁺transients that also occur in the form of waves. These Ca²⁺rises cause a decrease in cyclin-dependent kinase activity, which stimulates the egg to reenter the meiotic cell cycle (Kline and Kline, 1992; Collas et al., 1995; Sensui and Morisawa, 1996; Albrieux et al., 1997; McDougall and Levasseur, 1998; Levasseur and McDougall, 2000); however, the signalling pathway leading to the initiation of the Ca²⁺rises is only beginning to be understood.

In eggs of echinoderms (Whitaker and Irvine, 1984; Ciapa and Whitaker, 1986; Terasaki and Sardet, 1991; Mohri et al., 1995; Carroll et al., 1997, 1999; Lee and Shen, 1998; Shearer et al., 1999), frogs (Nuccitelli et al., 1993; Runft et al., 1999) and mammals (Miyazaki et al., 1992, 1993), the Ca²⁺rise at fertilization results from the release of Ca²⁺from the endoplasmic reticulum, which is mediated, at least in part, by inositol trisphosphate (IP₃). In ascidian eggs, IP₃ is produced at fertilization (Toratani and Yokosawa, 1995), and introduction of IP₃ causes Ca²⁺release (McDougall and Sardet, 1995; Albrieux et al., 1997; Yoshida et al., 1998), but the question of whether IP₃ is required for Ca²⁺release at fertilization has not been definitively answered (Russo et al., 1996; Yoshida et al., 1998; Wilding et al., 1999).

IP₃ is produced by the phospholipase C (PLC) family of enzymes, which cleave phosphatidylinositol 4,5-bisphosphate to generate IP₃ and diacylglycerol (Singer et al., 1997). The PLC family includes three subgroups: β, γ and δ. PLCγ is activated when its two tandem Src-homology 2 (SH2) domains interact with a specific phosphotyrosine-containing binding site present on an activated tyrosine kinase. This association allows the tyrosine kinase to phosphorylate and activate PLCγ. Recombinant proteins containing the SH2 domains of PLCγ have been used as specific dominant negative inhibitors of PLCγ activation in vitro (Bae et al., 1998) and in vivo (Roche et al., 1996; Carroll et al., 1997, 1999; Wang et al., 1998; Mehlmann et al., 1998; Shearer et al., 1999; Runft et al., 1999). The SH2 domains are believed to block the association of endogenous full-length PLCγ with its activating tyrosine kinase. Injection of PLCγ SH2 domains into echinoderm eggs inhibits the Ca²⁺rise at fertilization, demonstrating that, in starfish and sea urchin eggs, this Ca²⁺rise is initiated by SH2 domain-mediated activation of PLCγ (Carroll et al., 1997, 1999; Shearer et al., 1999). Injection of PLCγ SH2 domains into mouse or frog eggs, however, does not inhibit the Ca²⁺rise at fertilization, indicating that, unlike echinoderms, these vertebrate eggs do not require SH2 domain-mediated activation of PLCγ to initiate the Ca²⁺rise at fertilization (Mehlmann et al., 1998; Runft et al., 1999). Here we investigate whether the Ca²⁺release signalling pathway in ascidians (an evolutionary intermediate deuterostome) resembles that in echinoderms or vertebrates.

In echinoderm fertilization, PLCγ activation (Rongish et al., 1999) relies, directly or indirectly, on the activation of a Src family kinase (Giusti et al., 1999a,b, 2000; Abassi et al., 2000; Kinsey and Shen, 2000), and studies in frog eggs indicate that a Src family kinase may also be required in vertebrate fertilization (Sato et al., 1996, 1998, 1999, 2000; Glahn et al., 1999). Evidence that a Src family kinase plays a role in echinoderm egg activation includes the finding that injection of SH2 domains of Src family kinases into sea urchin and starfish eggs inhibits Ca²⁺release at fertilization (Giusti et al., 1999b; Abassi et al., 2000; Kinsey and Shen, 2000). In this study, we use SH2 domains to investigate the role of a Src family kinase in ascidian fertilization.

How the sperm initiates the signal transduction cascade that leads to the Ca²⁺rise in the egg at fertilization is unknown. One possibility is that influx of extracellular Ca²⁺might initiate this process (see Créton and Jaffe, 1995). This hypothesis, however, is not consistent with observations that Ca²⁺release at fertilization can be initiated in the absence of extracellular Ca²⁺in eggs of mammals (Jones et al., 1998), echinoderms (Schmidt et al., 1982), and ascidians (Speksnijder et al., 1989, 1990a; Sensui and Morisawa, 1996). Other possibilities are that the sperm contacts a receptor on the egg plasma membrane that initiates the signal transduction pathway, or that the fusion of sperm and egg introduces a factor from the sperm into the egg membrane or cytoplasm that induces egg activation. In favor of contact-initiated egg activation, externally applied proteases, lectins and sperm proteins can cause egg activation events in several species: the echiuroid worm Urechis (Jaffe et al., 1979; Gould and Stephano, 1987; Stephano and Gould, 1997), echinoderms (Steinhardt et al., 1971; Carroll and Jaffe, 1995), ascidians (Zalokar, 1980; Speksnijder et al., 1990a; Flannery and Epel, 1998), and amphibians (Shilling et al., 1998; Mizote et al., 1999). In favor of fusion-initiated egg activation, injection of a sperm extract into eggs causes Ca²⁺release and other egg activation events in various species: the nemertean worm Cerebratulus (Stricker, 1997), sea urchins (Dale et al., 1985), ascidians (Dale, 1988; Wilding and Dale, 1998; Kyozuka et al., 1998) and mammals (Swann, 1990; Oda et al., 1999; Perry et al., 2000). The Ca²⁺release induced by sperm extract injection is spatially and temporally like that occurring at fertilization (Swann, 1990; Stricker, 1997; Kyozuka et al., 1998; Wilding and Dale, 1998; Oda et al., 1999; Perry et al., 2000), and in mouse eggs, both fertilization and sperm extract-induced Ca²⁺release are inhibited by an antibody against the IP₃ receptor (Miyazaki et al., 1993; Oda et al., 1999). Here we investigate whether the sperm extract initiates intracellular Ca²⁺release in ascidian eggs by activating the same signal transduction molecules that function at fertilization.

Ciona intestinalis were obtained from the Marine Biological Laboratory (Woods Hole, MA, USA). To obtain gametes, the tunic was removed, the egg or sperm duct was punctured, and the gametes were collected. Eggs were washed twice with natural sea water and observed under a microscope to make sure no sperm were present. Sperm were kept on ice. For fertilization, eggs and sperm were collected from separate animals to facilitate rapid and synchronous fertilization (Rosati and De Santis, 1978; De Santis and Pinto, 1991). Insemination was performed by replacing the solution in the injection chamber (see below) with a suspension of sperm. In most experiments, the suspension of sperm collected from the sperm duct was diluted 1:100 in natural sea water. However, the final sperm concentration was somewhat variable, because the concentration of the suspension collected from the sperm duct was not constant.

Sperm extract was prepared based on methods described by Kyozuka et al. (1998). Sperm were washed 3 times in Ca²⁺-free sea water, and then after centrifugation, the volume of packed sperm was estimated, and the sperm were resuspended in 2.5 μl of extraction buffer (140 mM KCl, 1 mM MgCl₂, 5 mM Hepes, pH 7.0) per μl of packed sperm. Resuspended sperm were then homogenized 50 μl at a time for approx. 10-15 minutes in a microfuge tube on ice using a conical teflon pestle (PGC Scientific, Frederick, MD, USA, #63-5430). The homogenate was observed under a microscope to check that intact sperm or sperm heads were not present. The homogenate was then centrifuged at 4°C at 16,000 g for 10 minutes, or at 100,000 g for 30 minutes. The supernatant was collected as sperm extract, divided into portions, frozen in liquid nitrogen, and stored at −70°C. The protein yield in these extracts was 0.7-2.4 pg per sperm, out of a total protein content of 11 pg per sperm. The total protein content of the sperm was determined after sonication in 1% SDS. Protein concentrations were measured using a BCA assay (Pierce Chemical Co., St Louis, MO, USA) with BSA standards. For the experiments involving SH2 domain injections, the sperm extract was prepared using a 16,000 g centrifugation and the protein concentration was 6 mg/ml.

The sperm extract could be thawed and refrozen up to 3 times without losing activity. However, extract kept at 16-18°C for over an hour showed a detectable loss of activity. The activity of the extract made with the high speed centrifugation was the same as that made with the low speed centrifugation, indicating that the activity is cytosolic rather than an integral membrane component. Heat-inactivated sperm extract was prepared by incubating the extract at 96°C for 10 minutes. This extract was then centrifuged for 2 minutes at 16,000 g and the supernatant was collected and stored at −70°C. Control extract from ascidian ovary was prepared by homogenizing minced ovary as described above for sperm, and collecting the supernatant after centrifugation at 100,000 g for 30 minutes.

Plasmid DNA encoding glutathione-S-transferase (GST)-SH2 domain fusion constructs was obtained from the following sources: bovine GST-PLCγ1 SH2 (N + C) from S. A. Courtneidge (Sugen, Inc., Redwood City, CA, USA; see Carroll et al., 1997), murine GST-SHP2 SH2 (N + C) from T. Pawson (Mount Sinai Hospital, Toronto, Ontario, Canada; see Feng et al., 1993; Carroll et al., 1997), chicken GST-Fyn SH2 from K. Vuori (The Burnham Institute, La Jolla, CA, USA; see Giusti et al., 1999b), and murine GST-Abl SH2 from B. J. Mayer (Harvard Medical School, Boston, MA, USA; see Giusti et al., 1999b). GST fusion proteins were made as previously described (Gish et al., 1995; Carroll et al., 1997), then spin-dialyzed and concentrated in 137 mM NaCl, 2.7 mM KCl, 10 mM phosphate buffer, pH 7.2.

Ciona eggs arrested at metaphase of meiosis I, with chorion and follicle cells intact, were placed in a microinjection chamber between two coverslips separated by two pieces of double stick tape (Kiehart, 1982). All experiments were performed with the eggs in natural sea water. The eggs were microinjected using mercury-filled micropipets, which allows injection of precise picoliter volumes (Hiramoto, 1962). The micropipet tips were broken to a diameter of 2-4 μm and calibrated by expelling a drop of oil in sea water and calculating the volume of the drop based on its diameter. Injected volumes were 1-3% of the egg volume (1800 pl). All injections were performed at 16°C on the stage of an upright microscope with a 20×, 0.5 numerical aperture Plan Neofluar objective (Carl Zeiss, Inc., Thornwood, NY, USA) and a micrometer reticle in the eyepiece.

To prevent the eggs from lysing during the injection, the micropipet was first pushed through the chorion and positioned so that its tip faced the egg equator, and then pushed into the egg until the tip was approximately at the egg center. The injection solution was expelled and the micropipet was then quickly pulled out of the egg to prevent the micropipet from sticking to the chorion and follicle cells. Each micropipet was used only once. Protein and IP₃concentrations given in the text refer to the final concentration in the egg cytoplasm.

Ca²⁺measurements were made using calcium green-1 10-kDa dextran (Molecular Probes, Eugene, OR, USA) at a final concentration of 10 μM in the egg cytoplasm. Fluorescence was detected using a 20×, 0.5 numerical aperture Plan Neofluar objective, 485 nm excitation and 535 nm emission filters, and a photomultiplier tube connected through a current-to-voltage converter to a chart recorder as previously described (Chiba et al., 1990). For imaging calcium green fluorescence, we used a laser scanning confocal microscope (MRC600; BioRad Laboratories, Hercules, CA, USA) with a 20×, 0.5 numerical aperture Plan Neofluar objective. The video output from the confocal microscope was stored on an optical memory disk recorder (see Terasaki et al., 1997; Carroll et al., 1997). Eggs to be used for Ca²⁺measurements were coinjected with calcium green dextran and SH2 domain protein (3% of the egg volume), incubated at 16°C for 1-2 hours, and then inseminated or injected with sperm extract (1% of the egg volume) or IP₃(Calbiochem, La Jolla, CA, USA; 1% of the egg volume). Experiments were done at 16-18°C.

Sperm entry was detected by injecting eggs with 4’6-diamidino-2-phenylindole, dihydrochloride (DAPI) (Molecular Probes). The DAPI was dissolved in distilled water at 1 mg/ml, and injected to obtain a final concentration of 10 μg/ml. Fluorescence was detected using a 20×, 0.5 numerical aperture Plan Neofluar objective, and 330 nm excitation and 400 nm long-pass emission filters, and photographed using TMAX 400 film (Eastman Kodak Co., Rochester, NY, USA).

The Ca²⁺rise at fertilization in Ciona eggs consists of several components. The initial response that is seen in records of total calcium green fluorescence is a Ca²⁺action potential (Figs 1A, 2A, 4A, asterisks), resulting from Ca²⁺entry through voltage-gated channels in the egg plasma membrane (Goudeau et al., 1992). The action potential is triggered by the depolarization that occurs at the time of sperm-egg fusion (McCulloh and Chambers, 1992), and serves to amplify the depolarization and establish an electrical block to polyspermy (Goudeau and Goudeau, 1993; Goudeau et al., 1994). It is followed by a much larger Ca²⁺rise (Figs 1A, 2A, 4A), which crosses the egg in the form of a wave (Fig. 5A) and is due to release of Ca²⁺from intracellular stores (Speksnijder et al., 1989, 1990a,b; Brownlee and Dale, 1990; Yoshida et al., 1998; Kyozuka et al., 1998). Ca²⁺remains high for several minutes and then begins to oscillate. We used the action potential as a marker of when fertilization occurred and measured the timing of intracellular Ca²⁺release with respect to this.

To investigate the role of PLCγ at fertilization in ascidian eggs, Ciona eggs were coinjected with calcium green dextran and PLCγ SH2 domains, or SH2 domains from a control protein, the phosphatase SHP2. These eggs were then inseminated while total calcium green fluorescence was monitored using a photomultiplier. All eggs injected with the control SH2 domains at a final concentration of 1 mg/ml in the egg cytoplasm showed Ca²⁺rises similar to those that normally occur at fertilization, with the release of intracellular Ca²⁺starting at an average of 18 seconds after the action potential rise (Fig. 1A, Table 1). In contrast, 5 out of 7 eggs injected with 1 mg/ml (20 μM) PLCγ SH2 domains showed an action potential but no subsequent Ca²⁺release (Fig. 1B). The other 2 eggs injected with 1 mg/ml PLCγ SH2 domains did release Ca²⁺at fertilization, but with a delay of several minutes after the action potential (Fig. 1C, Table 1). In these eggs, the peak amplitude of the Ca²⁺increase, and its rate of rise, were the same as in control eggs, although the duration of the initial Ca²⁺elevation was shorter (Table 1). The three subsequent Ca²⁺oscillations appeared to be normal, but recordings were not continued beyond this point. All eggs injected with 0.1 mg/ml (2 μM) PLCγ SH2 domains showed a Ca²⁺increase of normal amplitude, rate of rise and duration, but the delay between the action potential and Ca²⁺release was about 1 minute longer than in controls (Table 1). These data indicate that, as in echinoderm eggs, SH2 domain-mediated activation of PLCγ is needed to initiate the release of Ca²⁺from intracellular stores at fertilization in Ciona eggs.

To examine if SH2 domain-mediated activity of a Src family tyrosine kinase is needed for release of intracellular Ca²⁺at fertilization in ascidian eggs, Ciona eggs were coinjected with calcium green dextran and SH2 domains from the Src family kinase Fyn, or control SH2 domains from the non-Src family kinase Abl. The SH2 domain of Fyn acts as a specific dominant negative inhibitor of the activation of Src family kinases (Roche et al., 1995; Giusti et al., 1999b; Abassi et al., 2000; Kinsey and Shen, 2000). The activation of several members of the Src kinase family is probably inhibited by the Fyn SH2 domain, since the SH2 domains of the mammalian Src family kinases that have been studied have similar phosphopeptide binding specificities (Songyang et al., 1995). Which particular Src family kinases are present in ascidian eggs is unknown.

All control eggs injected with 1 mg/ml Abl SH2 domains showed Ca²⁺rises similar to those that normally occur at fertilization (Fig. 2A, Table 1), while 5 out of 8 eggs injected with 1 mg/ml (25 μM) Fyn SH2 domains showed no Ca²⁺release following the action potential (Fig. 2B, Table 1). 3 of the 8 eggs injected with Fyn SH2 domains did release intracellular Ca²⁺at fertilization, but the time from the action potential to intracellular Ca²⁺release was increased by several minutes (Fig. 2C, Table 1). In these eggs, the peak amplitude of the Ca²⁺increase, and its rate of rise, were the same as in control eggs; the duration of the initial Ca²⁺elevation was shorter, but the subsequent three oscillations appeared to be normal (Fig. 2C, Table 1). These results indicate that SH2 domain-mediated activity of a Src family tyrosine kinase is required for initiating intracellular Ca²⁺release at fertilization in ascidian eggs. The action potential component of the Ca²⁺rise still occurs in eggs injected with PLCγ or with Fyn SH2 domains, indicating that this Ca²⁺rise does not depend on PLCγ or Src family kinases.

To confirm that the inhibition of Ca²⁺release at fertilization by PLCγ or Fyn SH2 domains is upstream of IP₃ production, Ciona eggs were injected with 1 mg/ml PLCγ SH2 or Fyn SH2 domains followed by 50 nM IP₃ (close to the minimum amount of IP₃ needed to initiate Ca²⁺release). Control eggs injected with 50 nM IP₃, without SH2 domains, showed a Ca²⁺rise immediately after injection (within the several seconds required to open the photomultiplier shutter), which was followed by 2-4 oscillations (n=3). A longer series of Ca²⁺transients like that at fertilization is seen only with constant perfusion of IP₃ (Albrieux et al., 1997). In response to injection of 50 nM IP₃, all eggs that had been preinjected with PLCγ SH2 domains (n=4) or Fyn SH2 domains (n=4) showed an immediate Ca²⁺rise followed by 2-4 oscillations, as in controls (data not shown).

To confirm that sperm entry occurred in the PLCγ and Fyn SH2 domain injected eggs, we used the DNA stain DAPI. The DAPI was injected, rather than applied externally, in order to stain only internalized sperm. The DAPI injection was made 2-3 minutes after insemination, and the egg was observed 1-7 minutes after the injection, before the dye began to leak out of the egg (Kaji et al., 2000). A calcium green recording was made in each of these eggs, to be sure that no Ca²⁺release had occurred at the time sperm nuclei were first observed in the cytoplasm. 1-3 nuclei were detected in the cytoplasm of eggs injected with 1 mg/ml of the SH2 domains of PLCγ (n=4) or Fyn (n=2) (Fig. 3). Due to the optical density of the egg cytoplasm, it is possible that additional sperm entries occurred but were not detected. The occurence of polyspermy in the SH2 domain-injected eggs suggests that Ca²⁺release may be necessary for the establishment of a polyspermy block.

Studies of the ascidian species Ciona savignyi (Kyozuka et al., 1998) and Ciona intestinalis (Wilding and Dale, 1998) have demonstrated that injecting ascidian sperm extract into these eggs stimulates a series of Ca²⁺rises similar to that at fertilization. We confirmed these findings using Ciona intestinalis; eggs injected with sperm extract produced a series of Ca²⁺rises resembling that at fertilization, except that no action potential was seen (Fig. 4A,B, Tables 1 and 2). As reported for C. savignyi (Kyozuka et al., 1998), the sperm extract initiated a Ca²⁺rise in C. intestinalis that originated at the egg surface and propagated across the egg in the form of a wave like that at fertilization (Fig. 5). Control eggs injected with heat-inactivated sperm extract (n=5) did not show Ca²⁺rises, indicating that the activating factor is most likely a protein (Fig. 4C).

To quantitate the amount of sperm extract required to release Ca²⁺, various amounts of sperm extract were injected into eggs. All eggs injected with 35-135 pg of sperm extract protein showed Ca²⁺rises, starting 84±149 seconds after injection (mean ± s.d., n=18). With a 2-12 pg injection, Ca²⁺release did not occur (n=4). A single Ciona intestinalis sperm contains approx. 11 pg of total protein. Therefore, 12-35 pg of sperm extract corresponds to the amount of protein present in approx. 1-3 sperm, indicating that, under our experimental conditions, microinjection of between 1 and 3 sperm equivalents of extract protein is needed to cause Ca²⁺release. Since some Ca²⁺releasing activity may be lost in the process of preparing the sperm extract (see Materials and Methods), these observations are consistent with the idea that the protein present in a single sperm is sufficient to cause Ca²⁺release. Furthermore, microinjection of the sperm extract into the center of the egg cytoplasm differs from the process of sperm-egg fusion, which introduces sperm proteins in a concentrated form near the egg plasma membrane.

To investigate if an extract made from a tissue other than sperm could initiate Ca²⁺rises in the ascidian egg, we prepared an extract from a Ciona ovary. Of three eggs injected with the ovary extract (216 pg protein), none showed a Ca²⁺rise during a 15 minute recording period. This result indicates that the activating factor present in sperm is not present in ovary.

To test if the sperm extract could initiate Ca²⁺rises in the egg when applied externally, 70 pl of sperm extract (98 pg protein) was injected into the approx. 700 pl space between the chorion (external egg envelope) and the egg plasma membrane. Of three eggs tested, none showed a Ca²⁺rise during a 15 minute recording period. Injection of fluorescent 10 kDa dextran into the perichorion space showed that fluorescence was retained in this space for at least 30 minutes, indicating that sperm extract probably remained in this space during the recording period. These results show that the sperm extract is only effective when injected into the egg cytoplasm, and does not initiate Ca²⁺rises in the egg by an external action on the egg surface.

If a soluble sperm factor initiates Ca²⁺release at fertilization, then sperm extract injection should cause Ca²⁺release by the same pathway as operates at fertilization. To determine if sperm extract-induced Ca²⁺release requires SH2-mediated activation of PLCγ, Ciona eggs were coinjected with calcium green dextran and 1 mg/ml PLCγ SH2 domains or control SHP2 SH2 domains. After 1-2 hours, these eggs were injected with sperm extract (108 pg protein) and calcium green fluorescence was monitored. While 9 out of 10 eggs preinjected with control SH2 domains showed a normal series of Ca²⁺rises in response to sperm extract injection, none of the eggs preinjected with PLCγ SH2 domains showed any Ca²⁺release in response to sperm extract injection (Fig. 6A,B, Table 2). These results indicate that, like Ca²⁺release at fertilization, sperm extract-induced Ca²⁺release requires SH2 domain-mediated activation of PLCγ.

We next investigated if, as at fertilization, sperm extract-induced Ca²⁺release requires SH2 domain-mediated activity of a Src family kinase. Ciona eggs were coinjected with calcium green dextran and with 1 mg/ml Fyn SH2 domains or control Abl SH2 domains and then, 1-2 hours later, injected with sperm extract (108 pg protein). Although 6 out of 7 eggs preinjected with control SH2 domains showed a normal series of Ca²⁺rises in response to sperm extract injection, only 1 of 6 eggs preinjected with Fyn SH2 domains showed Ca²⁺release in response to sperm extract injection (Fig. 7A,B, Table 2), and in the one egg that released intracellular Ca²⁺, the release was delayed (Table 2). These results indicate that, like Ca²⁺release at fertilization, sperm extract-induced Ca²⁺release also requires SH2 domain-mediated activity of a Src family kinase.

While a Ca²⁺rise appears to be an essentially universal signal for restarting the cell cycle in eggs at fertilization, the pathway leading to this Ca²⁺rise varies among different species (see Introduction). Along the evolutionary branch that includes echinoderms, ascidians and vertebrates, the Ca²⁺rise results from Ca²⁺release from the endoplasmic reticulum, and this process requires IP₃ (see Table 3). In echinoderms and ascidians, IP₃ production results from an SH2 domain-mediated activation of a Src like kinase and PLCγ, whereas in vertebrates (frog and mouse), a different and as yet unidentified pathway leads to IP₃ production at fertilization (see Table 3).

In this study of ascidian eggs, we demonstrate that PLCγ is required for the release of intracellular Ca²⁺at fertilization, by showing that injection of eggs with the SH2 domains of PLCγ, which act as specific dominant negative inhibitors of PLCγ activation, inhibits Ca²⁺release at fertilization. Likewise, we demonstrate that a Src-like kinase is required in this pathway, by showing that injection of the SH2 domain of the Src family kinase Fyn inhibits Ca²⁺release at fertilization. That tyrosine kinase activity is required for ascidian egg activation is also indicated by studies showing that the tyrosine kinase inhibitor erbstatin inhibits the surface contraction that occurs at about 5 minutes after insemination (Ueki and Yokosawa, 1997).

The specific action of PLCγ and Fyn SH2 domains was demonstrated by the lack of effect of control SH2 domains from other proteins on Ca²⁺release, by the ability of IP₃ to bypass the PLCγ and Fyn SH2 domain inhibition, and by the lack of inhibition of sperm entry by PLCγ and Fyn SH2 domains. Additional specificity controls have been done in echinoderm eggs, where it was found that injecting SH2 domains of several other kinases and phosphatases, or point-mutated forms of PLCγSH2 and SrcSH2, had no effect on Ca²⁺release at fertilization (Carroll et al., 1997, 1999; Shearer et al., 1999; Giusti et al., 1999b; Abassi et al., 2000; Kinsey and Shen, 2000). Specificity was further established by the findings in echinoderm eggs that PLCγ SH2 domains do not inhibit Ca²⁺release in response to PLCβ stimulation, cholera toxin, cGMP or cADP ribose (Carroll et al., 1997, 1999), and that Fyn SH2 domains have only a small effect on Ca²⁺release in response to cGMP (Kinsey and Shen, 2000).

The concentration dependence of the inhibitory effects of SH2 domains is in the low micromolar range, and shows some variability among different animal species. In sea urchin eggs, 1-2 μM of the mammalian PLCγ SH2 domains is sufficient to completely inhibit Ca²⁺release at fertilization (Carroll et al., 1999; Shearer et al., 1999), and a significant delay is seen at 0.1-0.3 μM (Shearer et al., 1999). In contrast, in starfish and ascidian eggs, 20 μM of the PLCγ SH2 domains is required to inhibit Ca²⁺release completely, and 2 μM results in a delay but not complete inhibition (Carroll et al., 1997; present results). This difference in sensitivity between species is not understood, but could be related to differences in the amino acid sequences of the endogenous PLCγ proteins in these organisms (see Shearer et al., 1999). Fyn SH2 domains completely inhibit or substantially delay Ca²⁺release at fertilization in ascidian, sea urchin and starfish eggs at 25 μM (Giusti et al., 1999b; Abassi et al., 2000; Kinsey and Shen, 2000; present results). Although lower concentrations have not been examined in ascidian eggs, 2.5 μM Fyn SH2 domains significantly delay Ca²⁺release at fertilization in sea urchin and starfish eggs (Giusti et al., 1999b; Abassi et al., 2000).

Our finding that PLCγ SH2 domains can completely inhibit Ca²⁺release at fertilization indicates that IP₃ is required to initiate Ca²⁺release at fertilization in ascidians, and that the other known Ca²⁺release channel in the endoplasmic reticulum, the ryanodine receptor, does not initiate this Ca²⁺release. This is consistent with previous studies showing that ruthenium red, an inhibitor of the ryanodine receptor, does not inhibit Ca²⁺release at fertilization in ascidians (Wilding and Dale, 1998; Yoshida et al., 1998).

Our results support the hypothesis that in ascidian eggs, a Src family kinase, directly or through intermediate molecules, phosphorylates and activates PLCγ. The Src family kinases are among the tyrosine kinases known to participate in the activation of PLCγ in other cells (Arkinstall et al., 1995; Melford et al., 1997; Clements and Koretsky, 1999; Schlesinger et al., 1999), and studies in echinoderms indicate that activation of a Src family kinase leads to activation of PLCγ at fertilization. Recombinant SH2 domains of PLCγ associate with a starfish egg Src family kinase in a fertilization-dependent manner (Giusti et al., 1999a), and this association occurs by 15 seconds post-insemination. In addition, PLCγ from fertilized sea urchin eggs can form a stable complex with recombinant domains of Fyn (Kinsey and Shen, 2000). Injection of starfish eggs with PLCγ SH2 domains delays Ca²⁺release in response to injection of recombinant Src protein, indicating that Src acts upstream of PLCγ (Giusti et al., 2000). As in most species, the release of Ca²⁺at fertilization of ascidian eggs is a regenerative process that proceeds across the egg in a wave; Ca²⁺then remains high for several minutes before beginning to oscillate (see Introduction). Our results with ascidian eggs indicate that the SH2 domains of PLCγ and the Src family kinase function primarily in the initiation of Ca²⁺release, and have only a minor effect on the subsequent pattern of Ca²⁺elevation. This conclusion is based on the observation that in the few cases where fertilization caused a Ca²⁺response in the presence of PLCγ or Fyn SH2 domains, its initiation was delayed but its peak amplitude and rate of rise were unaffected. The only effect of the PLCγ and Fyn SH2 domains that we observed on the Ca²⁺response was that the duration of the initial Ca²⁺plateau was somewhat shorter. As in many other cells, such as HeLa cells responding to histamine (Bootman et al., 1997), amplification of Ca²⁺release in the ascidian egg probably occurs when Ca²⁺and/or IP₃ at the sperm interaction site rises to a critical threshold, leading to an almost ‘all-or-none’ response. In echinoderm eggs, the initiation and amplification steps in Ca²⁺release appear to be somewhat more interdependent. In these eggs, the injection of PLCγ or Src family kinase SH2 domains both delays and reduces the amplitude of the Ca²⁺release (Carroll et al., 1997, 1999; Shearer et al., 1999; Giusti et al., 1999b; Abassi et al., 2000; Kinsey and Shen, 2000). Whether an initiating signal results in a graded or all-or-none response may depend on the balance between positive feedback components that raise Ca²⁺, such as Ca²⁺activation of phospholipase C and the IP₃ receptor channel, and other factors that lower Ca²⁺, such as degradation of IP₃, Ca²⁺pump activity, and the inhibition of some IP₃ receptor isoforms by high Ca²⁺(see Miyazaki et al., 1993; Bootman et al., 1997; Hagar et al., 1998).

In eggs of various species, injection of a sperm extract causes intracellular Ca²⁺release like that at fertilization (see Introduction). Our present results show that for ascidians, the signalling pathway initiated by sperm extract injection requires the same components, a Src family kinase and PLCγ, that are required at fertilization. These findings support the hypothesis that during fertilization, intracellular Ca²⁺release is initiated by a factor from the sperm that enters the egg as a consequence of sperm-egg fusion.

The component in the ascidian sperm extract that stimulates egg activation is heat labile, and is between 30 and 100 kDa in molecular mass (Kyozuka et al., 1998). Sperm extract activity is not affected by high speed centrifugation (Wilding and Dale 1998; present results) or by the presence of protease inhibitors in the extraction buffer (Wilding and Dale, 1998), indicating that the factor is soluble and is probably not a protease. The activating factor found in sperm is not present in ovary (present results).

In mammals, several specific sperm proteins have been investigated as possible candidates for the egg activating factor, but none has been definitively established (Sette et al., 1997; Wolosker et al., 1998; Swann and Parrington, 1999; Perry et al., 2000). Our findings indicate that, in ascidians, the activating factor in the sperm extract may be a regulator, directly or indirectly, of a Src family kinase in the egg. Src family kinases can be regulated by kinases or phosphatases that affect the phosphorylation state of two regulatory tyrosines and also by molecules that bind to their SH2 and SH3 domains (Erpel and Courtneidge, 1995; Boerner et al., 1996; Brown and Cooper, 1996; Xu et al., 1999; Thomas, 1999). Our evidence that the SH2 domain of a Src family kinase is required for Ca²⁺release at fertilization suggests that the Src activator that functions at fertilization may act by way of the Src SH2 domain. Examples of such regulators in other signalling systems include the focal adhesion kinase FAK (Thomas et al., 1998; Schaller et al., 1999), the platelet-derived growth factor receptor (see Erpel and Courtneidge, 1995), and the ‘immune receptor tyrosine activation motifs’ of antigen receptors (Johnson et al., 1995); these regulatory proteins bind to the SH2 domains of Src family kinases, causing them to adopt an active conformation.

In summary, our results support the hypothesis that activation of ascidian eggs at fertilization is initiated when sperm-egg fusion introduces a soluble factor from the sperm into the egg cytosol, which directly or indirectly activates a Src family kinase. This Src family kinase then directly or indirectly activates PLCγ, which produces the IP₃ that stimulates Ca²⁺release from the endoplasmic reticulum.

We thank S. A. Courtneidge, T. Pawson, K. Vuori, and B. J. Mayer for providing DNA. We also thank D. J. Carroll and A. F. Giusti for their help and encouragement, M. Terasaki for assistance with the imaging experiments, K. R. Foltz and L. M. Mehlmann for useful discussions, S. Davis for advice on working with ascidians, C. S. Ramarao and D. Serwanski for assistance in making fusion proteins, and S. Miyazaki for advice on the use of DAPI injection to detect sperm entry. Supported by a grant from the National Institutes of Health to L. A. J.