βγ subunits of heterotrimeric G-proteins contribute to Ca²⁺ release at fertilization in the sea urchin

Fusion of egg and sperm during fertilization causes rapid and widespread changes in the egg. Egg responses, collectively termed `egg activation', may differ between organisms, but exhibit common features such as significant cytoplasmic Ca²⁺ release, increased cytoplasmic pH, cortical granule exocytosis (CGE), initiation of protein synthesis, and DNA replication. Exocytosis of cortical granules is necessary for the formation of a permanent block to polyspermy, and is a consequence of cytoplasmic Ca²⁺ release. In all deuterostome eggs studied to date, Ca²⁺ release from the endoplasmic reticulum (ER) is triggered by inositol (1,4,5)-trisphosphate [Ins(1,4,5)P₃] production (reviewed by Runft et al., 2002), whereas eggs of certain protostome species, such as bivalve mollusks, exhibit influx of external Ca²⁺ through voltage-gated channels at fertilization (Deguchi and Morisawa, 2003). The physiology of egg activation has been extensively documented, although the signaling mechanisms leading to these responses are less well understood.

How is the signal for Ca²⁺ release (Ins(1,4,5)P₃) produced in the deuterostome egg? The phospholipase C (PLC) family of enzymes hydrolyzes a minor membrane phospholipid, phosphatidylinositol 4,5-bisphosphate, to generate Ins(1,4,5)P₃ and diacylglycerol. PLC enzymes are usually cytoplasmic but, upon egg activation, are recruited to the plasma membrane where their substrate resides. The PLC family consists of five subgroups - β, γ, δ, ϵ and ζ - and molecular mechanisms of activation and membrane recruitment differ between the subfamilies (Rhee, 2001; Saunders et al., 2002). PLCβ can be activated by heterotrimeric G-proteins, and PLCγ by tyrosine kinases; less is known about δ, ϵ and ζ isoforms, although the ζ isoform is sperm specific. Signaling mechanisms leading to the production of Ins(1,4,5)P₃ at fertilization are under intense investigation in many deuterostome organisms, and pathways for both G-protein and tyrosine-kinase-mediated Ca²⁺ release have been found in the eggs of both vertebrate and invertebrate organisms (reviewed by Runft et al., 2002).

Evidence for heterotrimeric G-protein signaling at fertilization of invertebrate deuterostomes appears contradictory. Initial experiments in sea urchins showed that GTPγS (causing constitutive G-protein activity) induced CGE, an indicator of cytoplasmic Ca²⁺ release (Turner et al., 1986), whereas GDPβS (blocking G-protein activity) blocked CGE in response to sperm (Turner et al., 1986). These effects appeared to be upstream of Ca²⁺ release as buffering intracellular Ca²⁺ concentration at 0.1 μM prevented GTPγS-mediated activation, whereas injecting Ins(1,4,5)P₃ into GDPβS-injected eggs rescued CGE (Turner et al., 1986). By contrast, inhibition of G-proteins by injection of GDPβS failed to block the raise in Ca²⁺ concentration during fertilization, and even caused Ca²⁺ transients on its own (Crossley et al., 1991). However, it must be kept in mind that experiments utilizing GTPγS or GDPβS do not address the role of specific G-proteins, and will activate or inhibit all GTPases present in the cell (heterotrimeric, as well as small GTPases such as Ras and Rab, and large GTPases such as dynamin). Given the documented feedback and crosstalk of various cell-signaling pathways, these experiments are now hard to interpret.

Other studies utilized more-specific G-protein-targeting reagents to dissect signaling at fertilization in invertebrates. Cholera toxin, which activates Gαs-mediated signaling by ADP-ribosylation of the Gαs protein, induced CGE in sea urchin eggs (Turner et al., 1987). By contrast, pertussis toxin, which ADP-ribosylates and thus inhibits Gαi protein, produced inconsistent effects on CGE, being inhibitory only on occasion (Turner et al., 1987). Activation of the mammalian serotonin receptor expressed exogenously in starfish eggs led to CGE, supporting the presence of a functional G-protein signal transduction pathway that could also be active at fertilization (Shilling et al., 1994). Thus, certain heterotrimeric G-proteins are able to activate echinoderm eggs. However, without particular knowledge of the G-proteins present in the egg and reagents specific to each, it is not clear if heterotrimeric G-proteins actually do function in egg activation.

Because Ca²⁺ release at fertilization is mediated by Ins(1,4,5)P₃ production, more-specific experiments were directed at elucidation of which PLC isoform is active at fertilization. PLCγ (normally activated by tyrosine kinases) can be inhibited by introduction of PLCγ Src-homology 2 (SH2) domains competing away the activator (Src-family kinase) from the PLC. Injection of starfish, sea urchin or ascidian eggs with PLCγ SH2 domains inhibits Ca²⁺ release at fertilization, suggesting that this PLC isoform functions in a Ca²⁺ release pathway (Carroll et al., 1999; Carroll et al., 1997; Runft et al., 2002; Shearer et al., 1999). However, PLCγ SH2 domains do not inhibit Ca²⁺ release initiated by the injection of cholera toxin, suggesting that the mechanism of Ca²⁺ release might be different in this case (Carroll et al., 1999). A candidate Src-family kinase, SFK1, was recently identified in the sea urchin (Giusti et al., 2003).

The most common argument against the involvement of the G-protein-mediated signaling in Ca²⁺ release at fertilization is that PLCγ is responsible for Ins(1,4,5)P₃ production at fertilization. PLCγ activity is typically stimulated by tyrosine kinases and not directly by G-proteins, suggesting that G-proteins do not contribute to the signaling for Ca²⁺ release (Runft et al., 2002). However, recent progress in elucidating signal transduction networks makes this premise worth reevaluating. Studies in mammalian cell cultures suggest that the tyrosine kinase and heterotrimeric G-protein signaling pathways cross-talk considerably. In fact, activation of Src homologs has been reported in many cases to be dependent on activation of heterotrimeric G-proteins (reviewed by Hall et al., 1999; Luttrell and Lefkowitz, 2002). Moreover, the contribution of G-proteins to regulated exocytosis occurs by stimulating Ca²⁺ release from the ER (Zeng et al., 2003) or even downstream of Ca²⁺ release (Blackmer et al., 2001; Kreft et al., 1999). These reports stimulated us to re-investigate the role of G-proteins in signaling at fertilization. We have identified which G-proteins are present in the sea urchin egg, and have used specific reagents to interfere with their function. These results suggest a re-integration of G-proteins into the bigger picture of fertilization signaling.

All chemicals were purchased from Sigma, unless otherwise noted.

Adult Lytechinus variegatus were obtained from Duke University Marine Laboratory (Beaufort, NC) or collected in Tampa, FL, and handled as described (Voronina and Wessel, 2004).

Immunogens for specific G-protein antibodies were designed from the C-terminal peptides of each sea urchin G-protein alpha subunit, and antibodies were generated as described (Voronina and Wessel, 2004). The sequences were obtained from full-length cDNA cloning of each Gα subunit from two species, Strongylocentrotus purpuratus (Sp) and L. variegatus (Lv); the C-terminal peptide sequences for each Gα subunit were identical between the two species. Antibodies to Gαq used for immunoblots were subsequently affinity purified by linking the peptide immunogen to agarose beads (Pierce SulfoLink Kit) with sequential binding and eluting of the antibody as per the manufacturer's protocol. Additional antisera to sea urchin Gαs C-terminal peptide (CRDIIQRMHLRQYEL) were generated in two rabbits by Sigma Genosys. These sera showed results consistent with previously described commercial ones (Voronina and Wessel, 2004).

Protein samples of total egg lysates and cell-surface complex were made as described (Kinsey, 1986) and were subjected to SDS-PAGE followed by immunoblotting as described (Voronina and Wessel, 2004). Cell-surface complex proteins were resolved along with total egg lysate proteins (in a ratio of 1:10). Labeling specificity was ascertained by peptide blocking experiments (Voronina and Wessel, 2004) (data not shown), the diluted antisera were incubated for 1 hour with 300 μM antigenic peptides and then cleared by centrifugation.

Immunofluorescent localization was performed on thick paraffin sections of eggs and embryos that were fixed and processed as previously described (Laidlaw and Wessel, 1994; Voronina and Wessel, 2004). Fertilized eggs were fixed 10 minutes after fertilization; the event of fertilization was confirmed by 100% fertilization envelope (FE) formation in the batch of eggs. For double labeling with hyalin, the sections were incubated with a 2B7 monoclonal antibody against hyalin (1:1) (Wessel et al., 1998) followed by anti-mouse rhodamine-conjugated Fabs (1:20; Jackson Research Laboratories). The G-proteins were then labeled with appropriate primary antibodies as described, followed by FITC-labeled secondary anti-rabbit Fabs (1:20; Jackson Research Laboratories). Signals were visualized and recorded by laser-scanning microscopy with a Zeiss LSM 410 confocal microscope with appropriate filters.

Egg injections were performed as described (http://155.37.3.143/panda/injection/index.html) (Jaffe, 1999; Kiehart, 1982; Voronina et al., 2003). Unfertilized eggs were placed in a Kiehart chamber in artificial seawater (ASW), and injected with appropriate solutions that never exceeded 5% of the cell volume. The injection volume was kept identical between active and control reagents. An oil droplet of dimethylpolysiloxane or a 0.1 mg/ml solution of fluorescent dextran (labeled with rhodamine or Alexa Fluor488) was co-injected into cells as a marker. Both control and test reagents were injected into the same batch of cells, and the cells were then incubated at 22°C. Eggs were fertilized or activated by A23187 addition (10 μg/ml in ASW) or Ins(1,4,5)P₃ injection at least 1 hour post-injection.

The competitor peptides for microinjection for Gαq (LQLNLKEYNLV) and Gαi (IKNNLKDCGLF) corresponding to the C-termini of the G-proteins were designed based on the characterized sea urchin sequences (Voronina and Wessel, 2004), synthesized by Sigma Genosys, and resuspended at 10 mM in deionized water. Such reagents were previously demonstrated to inhibit specifically G-protein activity (Gilchrist et al., 1998; Rasenick et al., 1994). Antibodies to be microinjected were purified on protein A-Sepharose (anti-Gαi and anti-Gαq only, since anti-Gαs was already purified on protein A), exchanged into the injection buffer (300 mM glycine, 10 mM Hepes, pH 7.0), and concentrated with Centricon centrifugal filter devices, M_r 30×10³ cut-off size (Millipore). Anti-Gαq Fab fragments were generated by immobilized papain treatment (ImmunoPure Fab Preparation Kit; Pierce), and were subsequently affinity purified as described above for anti-Gαq IgG. For microinjections, Fabs were exchanged into the injection buffer and concentrated as above.

Mastoparan (Mas7; Biomol) was resuspended for microinjection in deionized water at 10 mM and injected to a final concentration in the injected cells of 10 μM. Pertussis toxin (A-subunit; Biomol) was resuspended in deionized water, and activated prior to injection by incubation at 35°C with 50 mM DTT, as previously described (Shilling et al., 1989). The calculated concentration of introduced toxin in the egg cytoplasm was 0.2 μg/ml.

FE formation was scored after challenging injected eggs with sperm. The percentage of eggs exhibiting normal FE was determined for each treatment in two or three independent experiments involving 10-12 cells each. The eggs counted as `abnormal' conformed to either of the following criteria: no FE at all, partial FE (as seen in Fig. 2F), or low FE (with clearance of the envelope from the surface of the cell half or less than that in the non-injected eggs of the same batch). For each egg, the event of fertilization was confirmed by identifying the sperm pronucleus by DNA staining (0.2 μg/ml Hoechst 33258; Molecular Probes).

FM1-43 recordings were performed with the Zeiss LSM 410 confocal microscope. The eggs were first equilibrated in 2 μM FM1-43 solution in ASW for 5 minutes in a Kiehart chamber, and the contents were then replaced with a dilute sperm suspension in ASW/FM1-43 (pre-dilution of sperm in FM1-43 is necessary to prevent exhaustion of the label by binding to sperm). The egg to be recorded was centered, and brought in focus under the 40× objective (C-apochromat 40×, NA 1.2), zoom 1.5×. Time-lapse recording was performed with the following parameters: FITC filter set (excitation: LP 460, emission LP 515; 488 laser line), 1-second scan speed, 2× line averaging, 10-second interval between scans. This way of recording provided the greatest dynamic range of the increase in membrane-associated fluorescence at fertilization in comparison with epifluorescence or fluorimeter recordings (data not shown). The integrated pixel intensity associated with the egg circumference was determined for each frame in Metamorph (Universal Imaging Corporation). The last image before detectable increase in membrane fluorescence was assigned `time 0', and its intensity value was set to 1; the rest of the values were scaled accordingly.

Eggs were injected with a dye mixture containing 4 mM Oregon Green 488 BAPTA-1 (OGB-1) dextran, M_r 10×10³ (contains an average of 1.1 dye molecules per dextran molecule), and 2 mM Texas Red dextran, M_r 10×10³ (Molecular Probes), such that the final concentration of OGB-1 dextran in the cell was 40 μM. Fluorescence was imaged during fertilization on a Zeiss LSM 410 confocal microscope, under 25× objective, NA 0.8, using the dual red/green excitation/recording program. Time-lapse images were collected at 1-second scan speed, 4× line averaging, 10-second interval between scans. The recorded images were quantified in Metamorph, and the ratio of green to red total fluorescence intensity was determined for each time point. The last image before detectable increase in the ratio was assigned `time 0', and its ratio value was set to 1; the rest of the values were scaled accordingly. This way of recording precluded us from effectively detecting the activation potential at sperm-egg fusion, which was determined in a separate set of experiments.

Eggs were injected with OGB-1 dextran (Molecular Probes), such that the final concentration of OGB-1 dextran in the cell was 40 μM. Fluorescence was imaged during fertilization on a Zeiss Axioplan microscope, using a 20× NA 0.5 objective, with a FITC filter set. Time-lapse images were collected at a frequency of 0.86 seconds and were quantified in Metamorph to determine the dynamics of total fluorescence over the egg. Each fluorescence trace was normalized to the fluorescence of the unfertilized egg. The Ca²⁺ wave lag-phase was determined as the time interval between the first increase in fluorescence associated with the activation potential and the first sustainable increase in fluorescence leading to the Ca²⁺ wave (Carroll et al., 1997).

A histidine-tagged phosducin construct (S to A mutant at multiple positions) was generously provided by B. Willardson (Thulin et al., 2001). All sites of inhibitory phosphorylation were removed from this expressed protein, making it a particularly potent Gβγ scavenger. A histidine-tagged βARKct expression construct was generously provided by H. Hamm (Blackmer et al., 2001). The 6-histidine fusion expression constructs were transformed into BL-21 strain of Escherichia coli, and induced to express the fusion proteins with 1 mM IPTG for 3 hours at 37°C. The cells were harvested, resuspended in the native lysis buffer (50 mM NaH₂PO₄, 300 mM NaCl, 10 mM imidazole, pH 8.0), and lysed by rapid freeze-thaw followed by sonication. Insoluble fractions were removed by centrifugation at 13,000 g for 30 minutes. The supernatants containing soluble fusion proteins were bound to Ni-NTA resin (ProBond; Invitrogen), and washed with 20 mM imidazol-containing wash buffer. The bound proteins were eluted with increasing stepwise concentrations of imidazole (50 mM, 100 mM and 250 mM), and the majority of protein eluted with the 100 mM imidazole fraction. For microinjection, the proteins were then exchanged into injection buffer (300 mM glycine, 10 mM Hepes, pH 7.0), and concentrated with Microcon centrifugal filter devices, M_r 10×10³ cut-off size (Millipore).

We have cloned four Gα subunits - Gαi, Gαq, Gαs and Gα12 - from two species of sea urchin, produced and characterized antibodies against these proteins, and found them expressed in sea urchin oocytes and eggs (Voronina and Wessel, 2004). Here, we document the localization of Gα proteins with antibodies directed against subunit-specific C-terminal peptides of sea urchin Gα subunits (Fig. 1A-D) double labeled with an antibody to the cortical granule content protein hyalin (Wessel et al., 1998). Three of the G-proteins, Gαi, Gαq and Gαs, are associated with the cell cortex in eggs, whereas Gα12 is distributed throughout the cytoplasm.

To test biochemically if Gαi, Gαq and Gαs proteins are enriched in the egg cortex, we performed fractionation of eggs, followed by western blotting of the isolated cell-surface complex (CSC, containing plasma membrane with docked cortical granules), in comparison with the total egg lysate (Fig. 1E). The quality of fractionation was assessed by blotting for the known cortical granule components enriched in the CSC (MGB) (Haley and Wessel, 2004) and for yolk protein absent from the CSC (YP30) (Wessel et al., 2000). Gαi, Gαs and Gαq are all detected in CSC, and are considerably enriched in this fraction compared with whole egg lysate (note that egg samples are loaded at 10× the mass of the CSC samples). In agreement with the immunolocalization data, Gα12 is not enriched in the CSC. Documentation of anti-Gα subunit immunospecificity includes competitive ablation of the signal (both by western blotting and immunolocalization) by the immunogenic peptides (Voronina and Wessel, 2004) (data not shown). Higher molecular weight bands detected by anti-Gα antisera represent nonspecific binding of the antibodies to the abundant content proteins of cortical granules, and are not ablated by peptide competition. These epitiopes are not accessible to antibodies subsequently microinjected into the cytoplasm. The presence of Gαi and Gαs at the cell surface of mature sea urchin eggs is consistent with a previous report detecting ADP-ribosylated substrates of pertussis toxin (Gαi) and cholera toxin (Gαs) in CSC (Turner et al., 1987). Gαq and Gαs integrate into the plasma membrane upon exocytosis of the cortical granules at fertilization, whereas Gαi moves to endocytic vesicles (data not shown). We pursued a functional analysis of the G-proteins localized to the cell surface (Gαs, Gαq and Gαi), since they are poised to receive signal input at fertilization.

To test G-protein function at fertilization, we injected commonly accepted and widely used specific inhibitory reagents. These include antibodies and peptides that prevent interaction of a particular Gα subunit with its cognate receptor (Gilchrist et al., 1998; Manning, 1999; Rasenick et al., 1994), or pertussis toxin, which inhibits Gαi-mediated signaling by irreversible ADP-ribosylation of the protein (Manning, 1999). These types of reagents have shown great efficacy in a diverse array of cells. Furthermore, they offer a significant advantage of affecting a specific target.

We used FE formation as an initial screen for G-protein function, since CGE is a consequence of Ca²⁺ signaling and formation of FE reports CGE. For these experiments, we determined the fraction of the egg population that exhibited normal FE elevation upon fertilization following treatment with the G-protein inhibitors (Fig. 2). Inhibition of Gαs by injection of anti-Gαs antibodies interfered most with CGE (90% of eggs with aberrant CGE), closely followed by inhibition of Gαq (73% of eggs with aberrant CGE), whereas interfering with Gαi had the least effect (25% of eggs with aberrant CGE; Fig. 2A-C). These effects are specific since denatured antibodies did not have any effect on fertilization, and nonspecific IgGs from preimmune antisera did not alter CGE either (Fig. 2A-C). For microinjections, we chose 100 μg/ml final concentration of antibodies in the cytoplasm, which is at the lower range of published effective inhibitory concentrations (Moore et al., 1994; Runft et al., 1999; Williams et al., 1998). The effectiveness of anti-Gαq and anti-Gαs at these low concentrations gives us greater confidence in the specificity of these reagents.

A common concern in using whole antibodies in inhibition studies is that they are multivalent, and could lead to artifactual inhibition of signaling by crosslinking the antigens, as opposed to Fabs that only block the accessibility of the antigen. Therefore, we generated monovalent Fab fragments of the anti-Gαq antibodies, and affinity purified these Fab fragments. Injection of 4 μg/ml of affinity-purified anti-Gαq Fabs prevented FE formation in 80% of treated eggs on average, similar to the effect produced by introduction of whole anti-Gαq IgGs (Fig. 2B).

As an independent test of Gα function at fertilization, we also interfered with G-protein signaling by injecting alternative reagents. A C-terminal competitor peptide for Gαq recapitulated the antibody-induced phenotype inhibiting CGE in 55% of cases on average. Consistent with the phenotypes produced by anti-Gαi antibodies, inhibition of Gαi by microinjection of competitor peptide or pertussis toxin seldom interfered with fertilization (83% and 70% normal fertilization on average in the case of peptide and pertussis toxin respectively).

Is the activation of G-proteins sufficient to cause CGE? Earlier reports indicate that activation of Gαs signaling by treatment with cholera toxin induces CGE by inducing Ca²⁺ release (Turner et al., 1987), apparently utilizing a PLCγ-independent signaling pathway (Carroll et al., 1999). In our hands, artificial activation of Gαi-mediated signaling by injection of mastoparan failed to induce CGE. Furthermore, mastoparan did not interfere with CGE upon insemination (Fig. 2C). Thus, our overall results suggest that Gαs and Gαq play the dominant role in signal generation for egg activation, making them the focus of our subsequent studies.

To relate the qualitative morphological observation of interference with FE formation to the quantitative changes in CGE, we adapted an assay quantifying CGE at fertilization (Carroll and Jaffe, 1995; Terasaki, 1995). We used FM1-43, a membrane impermeable lipophilic dye that fluoresces only upon insertion into a hydrophobic environment. Sea urchin eggs bathed in FM1-43 solution accumulate a saturable, stable and quantitative amount of dye in their plasma membrane. Upon egg activation, a significant amount of membrane is added to the surface of the cell resulting from CGE. This new plasma membrane quickly accumulates additional FM1-43 from the media, and the fluorescence of the egg increases proportionally to the amount of additional new membrane. We have analyzed these membrane dynamics by FM1-43 incorporation at fertilization by quantifying fluorescence intensities of the recorded images, and found that a significant and reproducible 7-9-fold increase in fluorescence would occur at the cell surface during fertilization in the time span of 1 minute (Fig. 3A). This was clearly distinguishable from FM1-43 fluorescence of unfertilized eggs during the same time span.

We used this approach to quantify CGE in cells injected with the subunit-specific inhibitors of G-protein signaling described above. This technique allows one both to quantify the amounts of exocytosis, and to record the dynamics of secretion (Fig. 3B-D). The results indicate that inhibition of Gαs or Gαq signaling individually caused approximately 50% decrease in the amount of membrane added to the cell surface, consistent with the observed effect of respective inhibitors on FE formation (Fig. 3C,D). In turn, inhibition of Gαi signaling by both competitor peptide (Fig. 3B) and pertussis toxin treatment (data not shown) slowed down membrane addition compared with the control (Fig. 3B), but the Gαi-inhibited cells still reached control levels by 5 minutes post-fertilization. However, cells challenged by Gαs and Gαq inhibitors accumulated significantly less FM1-43 label than controls, even after 5 minutes (P<0.02 by Student's t-test; the groups of values are also significantly distinct by ANOVA test, P<0.01). These data allow us to propose an explanation for the observed low capability of Gαi inhibitors to interfere with FE formation (see also Turner et al., 1987): these treatments do not prevent CGE completely, but slow its dynamics, therefore, in a certain proportion of eggs the exocytosed contents of cortical granules never reach a critical concentration to form an envelope. Our results are consistent with the model that G-protein subtypes mediate distinct signaling events at fertilization.

The observed effects of G-protein inhibitors on CGE could be due to a direct effect of the G-proteins on the secretion machinery, or to an indirect effect on Ca²⁺ dynamics. To distinguish between these possibilities, we measured Ca²⁺ levels in the cells injected with inhibitors of G-protein signaling. For this purpose, we used recordings of eggs injected with Oregon Green 488 BAPTA-1 (OGB-1) dextran, which is a fluorescent Ca²⁺ sensor. The OGB-1 dextran is co-injected into the cell with Texas Red (TR) dextran, which is insensitive to Ca²⁺ levels, and fluorescence of the cell in both channels is recorded, such that a ratio of OGB-1 fluorescence to TR fluorescence can be determined. These ratiometric readings allow the precise quantification of Ca²⁺-dependent fluorescence as they normalize for potential variation in injected dye quantities, or photobleaching during fluorescence recordings (Stricker and Whitaker, 1999). The recording parameters in these experiments required 10-second intervals between frames, precluding us from reliably detecting the activating potential at sperm-egg fusion and thus measuring the latent period between fertilization and cytoplasmic Ca²⁺ transient (see, however, Fig. 6 below).

We focused on the phenotypes of Gαs- and Gαq-inhibited eggs since they appear most relevant to egg activation. Control eggs injected with the Ca²⁺-imaging mix exhibit the following characteristics: ∼1.6 times increase in relative fluorescent intensity in the time span of 40-60 seconds, and a slower reuptake of Ca²⁺ during the following 4-5 minutes. By contrast, the eggs injected with 100 μg/ml of anti-Gαs or with 50 μM of Gαq competitor peptide showed a significantly reduced Ca²⁺ release, exhibiting a decreased rate of Ca²⁺ release and reaching a peak of approximately 60% and 68% of the control eggs respectively (Fig. 4, Table 1). The experimental eggs also at times displayed abnormal propagation of the Ca²⁺ wave as compared with the control (Fig. 4D).

As a separate test of whether the observed defects of G-protein inhibition are upstream or downstream of the Ca²⁺ release, we asked if the effects of G-protein inhibition could be rescued by providing the eggs with Ca²⁺. Such treatment would rescue exocytosis in the case of G-proteins being upstream of Ca²⁺ release, but not in the case of G-proteins having a direct influence on the vesicle fusion machinery. Injection of 28 nM Ins(1,4,5)P₃ (Turner et al., 1986) into the sea urchin eggs pre-injected with anti-Gαq Fabs was sufficient to induce CGE in 100% of cases (n=9; Fig. 2B). This argues that the effect of G-protein inhibition is upstream of Ca²⁺ release.

Inhibition of the signaling through two different Gα subunits, Gαs and Gαq, produces the same result of interference with CGE. This phenomenon could be explained by both Gαs and Gαq acting on a common downstream effector. An alternative hypothesis is that activation of members of different heterotrimeric G-protein complexes generates one common consequence of liberating Gβγ subunits, which are the signaling mediators (Fig. 5B). Suppression of signaling through one particular Gα would decrease the overall abundance of released Gβγ subunits, and result in the decrease or suppression of downstream signaling.

To test the hypothesis that Gβγ complexes mediate G-protein-dependent signaling at fertilization, we overexpressed, purified and introduced Gβγ-sequestering proteins into sea urchin eggs and then challenged the eggs with sperm. The Gβγ inhibitors we used are the C-terminal domain of β-adrenergic receptor kinase, βARK1, or GRK2, (Blackmer et al., 2001), and phosducin (Thulin et al., 2001). Both these reagents serve as scavengers and bind to free Gβγ subunits, thereby preventing their interaction with normal cellular partners in a dominant fashion. The phosducin construct we used contains multiple substitutions of alanine for serine, rendering it insensitive to regulatory phosphorylation that could otherwise reduce its affinity for Gβγ complexes (Thulin et al., 2001). Introduction of either of these proteins into sea urchin eggs interfered with FE formation upon sperm addition (Fig. 5A). We tested several concentrations of injected phosducin, finding that the effects are concentration dependent (Fig. 5C); the difference between the effects of the 15 μg/ml and 113 μg/ml phosducin is statistically significant (P<0.001) by the Wilcoxin Rank Sum test. Phosducin injected at 45 μg/ml was the minimal concentration causing complete inhibition of CGE in at least some eggs. However, even as little as 15 μg/ml phosducin in the egg resulted in interference with CGE as determined by formation of low or incomplete FEs (Fig. 5C). The βARKct-mediated effect was also dependent on the amount of protein introduced into the egg (data not shown). These effects were specific, as they were abolished by denaturing phosducin or βARK prior to injection into the eggs (Fig. 5A). Recording of the dynamics of cortical granule secretion by FM1-43 in these injected eggs confirmed that exocytosis is strongly reduced (Fig. 5D). We conclude from these data that the Gβγ subunits are the primary mediators of the Gαs and Gαq signaling at fertilization.

Injection of phosducin or βARKct also significantly (P<0.01) decreased Ca²⁺ release at fertilization in the treated eggs (46% and 69% of control on average; Fig. 6A). Phosducin appeared to be a more efficient inhibitor than βARKct: it was inhibitory at lower concentrations, and the effects were more pronounced (Fig. 5D, Fig. 6A). These data complement the assays of CGE by FE formation and support the hypothesis of differential sensitivity of cortical granule subpopulations to available free Ca²⁺ (Blank et al., 1998; Matese and McClay, 1998; Terasaki, 1995). Variability in the threshold Ca²⁺ concentration that activates the fusion machinery of cortical granules is one possible explanation for the observed phenomenon.

The effect of G-protein signaling of the Ca²⁺ release could be in the initiation and/or the amplification step of the pathway. To test each of these possibilities, we measured the lag phase of the Ca²⁺ transient (the delay between gamete fusion and the initiation of Ca²⁺ release), using high-speed single-channel time-lapse recordings of OGB-1 dextran fluorescence in the eggs at fertilization. This assay is less quantitative than ratiometric analysis but provides better time resolution. The Ca²⁺ action potential functions as a fast electrical block to polyspermy and occurs simultaneously with sperm-egg fusion (McCulloh and Chambers, 1992). The action potential is detectable by this assay and serves as an indicator of gamete fusion. We find that inhibition of Gβγ signaling by injection of 80 μg/ml phosducin significantly (P<0.02) increases the delay between gamete fusion and the initiation of the Ca²⁺ wave to an average of 31 seconds (Fig. 6B), compared with either 7 seconds in the untreated eggs or 17 seconds in the eggs injected with denatured protein. We conclude that Gβγ signaling is a necessary signaling pathway contributing to the initiation and amplification of the cytoplasmic Ca²⁺ transient.

Additionally, we tested whether CGE could be rescued by supplementing cells treated with Gβγ scavengers with Ca²⁺. Eggs were injected with 130 μg/ml of phosducin and then exposed to the Ca²⁺ ionophore A23187 in artificial seawater (normally containing 10 mM Ca²⁺). In these experiments, 100% of phosducin-injected eggs (n=4) were able to exocytose cortical granules and form full FEs as well as the control cells (data not shown). As a more rigorous and physiologically relevant test, we asked if CGE in phosducin-injected eggs could be rescued by Ins(1,4,5)P₃. Providing phosducin-injected eggs with Ins(1,4,5)P₃ (28 nM) also induced CGE as assessed by normal FE formation in 100% of cases (n=5; Fig. 6C). The Ins(1,4,5)P₃ added (28 nM) is less than that normally experienced by a sea urchin egg at fertilization (67 nM increase by 30 seconds post-fertilization) (Lee and Shen, 1998). The ability of low Ins(1,4,5)P₃ levels to cause CGE suggests that the effect of phosducin is a result of interfering with a signal transduction pathway leading to Ca²⁺ release, and not due to the interference with the cortical granule fusion machinery.

We further tested this conclusion by assessing the response of phosducin-treated eggs to suboptimal concentrations of Ins(1,4,5)P₃. Although sub-physiological doses of Ins(1,4,5)P₃ can trigger CGE, sub-threshold levels of Ins(1,4,5)P₃ would produce a condition sensitized to the effects downstream of Ins(1,4,5)P₃ production. This allowed us to ask rigorously whether injected phosducin has effects downstream of Ins(1,4,5)P₃ production. We determined the minimal concentration of Ins(1,4,5)P₃ that was sufficient to induce FE formation when injected into the center of the cell (indicating near-normal Ca²⁺ transient and CGE) in 80% of non-treated eggs within 1 minute of injection (10 nM). This amount of Ins(1,4,5)P₃ was then supplied to phosducin or control treated eggs to determine if CGE would still occur (Fig. 6D). Phosducin-injected eggs were indistinguishable from controls injected with the denatured protein by either the delay of response, or the percentage of eggs able to initiate CGE within 1 minute of injection. Thus, we conclude that the major contribution of the Gβγ signaling at fertilization is upstream of the Ca²⁺ release and does not influence the exocytosis machinery directly or Ins(1,4,5)P₃ responsiveness.

Study of sea urchin eggs has led to the identification of key elements of the Ca²⁺ transient at fertilization. Recently, this research has determined several molecular participants in this process: SFK1 kinase activates PLCγ, which produces Ins(1,4,5)P₃ (reviewed by Runft et al., 2002). However, the upstream signaling events of egg activation are not clear, and the cellular factors impinging on the activity of SFK1 and PLCγ in the egg are not well understood (Sato et al., 2004). The participation of heterotrimeric G-proteins in egg activation has been controversial, largely because of a lack of specific G-protein reagents. The results presented here help resolve this controversy by identifying which Gα subunits are present in the egg, and by using inhibitory reagents specific for identified Gα subunits to address their specific function, if any, in fertilization.

We bridge several levels of effector activity manifest by the egg during its activation in this study, and find that signaling mediated by two of the heterotrimeric G-proteins in the egg, Gαs and Gαq, is necessary for normal CGE during sea urchin fertilization. The release of free Gβγ subunits produced by G-protein activation contributes to the timing and amplification of the cytoplasmic Ca²⁺ transient at sea urchin fertilization. We propose integration of Gβγ-mediated signaling into the current model of egg activation, potentially through contribution to activation of Src homolog (Giusti et al., 2003). Such activation of Src through Gβγ released by Gαi and Gαq activation has been previously reported in mammalian tissue culture cells (Canet-Aviles et al., 2002; Igishi and Gutkind, 1998; Luttrell et al., 1997). Our model re-integrates a G-protein contribution with Src signaling in echinoderm fertilization instead of positioning them as contradictory and mutually exclusive mechanisms. Another mechanism allowed by our evidence is the recently proposed gating of the Ins(1,4,5)P₃ receptor Ca²⁺ channels directly by the free Gβγ subunits (Zeng et al., 2003).

How do our findings on specific G-protein involvement in the signaling at echinoderm fertilization fit with previous results? Earlier reports showing a variable effect of pertussis toxin on FE formation in sea urchin can now be explained by slower kinetics of exocytosis resulting from Gαi inhibition (Turner et al., 1987). A recent report suggests that mastoparan, a G-protein-activating peptide, induces Ca²⁺-independent CGE in sea urchin eggs, whereas the inactive mastoparan analog, MAS-17, is without effect (Lopez-Godinez et al., 2003). However, we did not detect induction of CGE by microinjecting mastoparan. The eggs in the Lopez-Godinez et al. study were bathed in solutions with a high concentration of mastoparan, taking advantage of the potential ability of the peptide to cross the plasma membrane. However, it is unclear how much mastoparan actually enters the cell. The concentration we tested in our study is 10 μM, which is ten times lower than the smallest effective concentration in the study by Lopez-Godinez et al. (Lopez-Godinez et al., 2003); in our hands, mastoparan concentrations higher than 25 μM were toxic upon injection. Furthermore, mastoparan is known to produce nonspecific effects at high concentrations, which include plasma membrane pore formation (Suh et al., 1996), or activation of phospholipase D2 (Chahdi et al., 2003). A commonly accepted control for the specificity of mastoparan is the ability to inhibit its effects by Gαi-inhibitory reagents: competitor peptides, inhibitory antibodies or pertussis toxin (Kreft et al., 1999). The pertussis toxin treatment described by Lopez-Godinez et al. (Lopez-Godinez et al., 2003) failed to suppress mastoparan effects; therefore, the functional relevance of this report to physiological G-protein signaling is unclear.

In our hands, inhibiting G-protein signaling (Gαs, Gαq or Gβγ) in the egg never extinguished the Ca²⁺ transient completely, but it was significantly attenuated. We cannot presently distinguish between two possible explanations for this observation: (1) our reagents not being 100% effective in inhibiting G-protein signaling by competing with endogenous protein interaction partners, and/or (2) G-proteins function in the modulation - either by feedback or cross-talk - of the SFK-PLCγ-Ins(1,4,5)P₃ pathway required for Ca²⁺ release. In either scenario, the G-proteins are clearly required for timely, full-scale Ca²⁺ release at fertilization. The complexity of Ca²⁺-generating signaling at fertilization has emerged in recent years, implicating contribution of many signaling pathways to the eventual release of cytoplasmic Ca²⁺ (reviewed by Galione and Churchill, 2002). In addition to the dominant role of Ins(1,4,5)P₃, cyclic ADP-ribose (cADPr) is involved in generation of prolonged Ca²⁺ release from the ER (Leckie et al., 2003), and nicotinic acid adenine dinucleotide phosphate (NAADP) is hypothesized to act as a local trigger for Ca²⁺-induced Ca²⁺ release and to contribute to the generation of the Ca²⁺ activation potential (Churchill et al., 2003; Galione and Churchill, 2002). Production of cADPr at fertilization occurs in response to nitric oxide and cGMP, and heterotrimeric G-protein activity is linked to the cADPr pathway in certain mammalian tissues (Higashida et al., 2001).

Signaling mechanisms in vertebrate eggs (mouse, fish and Xenopus) at fertilization appear distinct from that in invertebrates; neither PLCγ SH2 domains, nor G-protein inhibitors such as pertussis toxin, anti-Gαq antibody, or sequestration of Gβγ subunits are inhibitory to the Ca²⁺ release (reviewed by Runft et al., 2002; Sato et al., 2004). Nevertheless, signaling through cytoplasmic tyrosine kinases still appears important, as tyrosine kinase inhibitors suppress Ca²⁺ release in Xenopus eggs (Sato et al., 2000), and SH2 domains of Fyn suppress Ca²⁺ transients in zebrafish eggs (Kinsey et al., 2003). The contribution of sperm factor (a sperm cytoplasmic protein inducing egg activation when introduced into the cytoplasm) is well documented for mammalian egg activation. Recent evidence supports a novel sperm PLCζ as a candidate for this activity, which produces a Ca²⁺ rise, enhanced by Ca²⁺-dependent autoamplification of sperm PLC activity (Carroll, 2001; Cox et al., 2002; Saunders et al., 2002; Swann et al., 2001). Other factors that might activate PLC in vertebrate eggs are Ca²⁺ introduced by sperm, increase in the amount of the lipid PLC substrate phosphatidylinositol (4,5)bisphosphate, or improved accessibility of this substrate to PLC (reviewed by Mehlmann et al., 1998). Clearly, identification of signaling pathways, participants and interactions at work in different fertilization mechanisms will enable deeper understanding of evolved traits in this fundamental process.

We thank past and present members of the Providence Institute of Molecular Oogenesis for support and thoughtful discussions, particularly M. Leguia for data on G-protein detergent extractions. We are grateful to R. Créton for providing the protocol and assistance for live-cell Ca²⁺ recordings. We appreciate the willingness of H. Hamm and B. Willardson to share βARK and phosducin expression constructs, especially the latter's suggestion to use multiply substituted phosducin. We would also like to thank L. Jaffe for critical review of this manuscript while in preparation. This research was supported by grants from the NIH and NSF.