Exploring the mechanism of action of the sperm-triggered calcium-wave pacemaker in ascidian zygotes

Sperm trigger a Ca²⁺ increase at fertilization that is responsible for activating virtually every species of egg tested to date (reviewed by Stricker, 1999). In eggs from different organisms, a repetitive Ca²⁺ increase is often required to trigger embryonic development [mammals (Swann and Ozil, 1994); nemerteans (Stricker, 1996); ascidians (McDougall and Sardet, 1995)]. Repetitive Ca²⁺ increases during fertilization usually occur as waves that come from vegetal Ca²⁺-wave pacemakers (reviewed by Dumollard et al., 2002). The first sperm-induced Ca²⁺-wave pacemaker to be described was that of a marine chordate (the ascidian Phallusia mammillata) (Speksnijder, 1992). Similar Ca²⁺-wave pacemakers are now known to exist in the fertilized eggs of nemertean worms (Stricker, 1996) and in mammals (Kline et al., 1999; Deguchi et al., 2000). It is still unknown how these Ca²⁺-wave pacemakers operate and also whether similar mechanisms operate between species.

The large transparent eggs of ascidians still offer one of the clearest examples in the animal kingdom of a Ca²⁺-wave pacemaker (Speksnijder, 1992; McDougall and Sardet, 1995). This pacemaker forms minutes after fertilization because of a sperm-induced Ca²⁺ wave that triggers a cortical contraction, culminating in the formation of a surface protrusion in the vegetal hemisphere termed the contraction pole (CP) (Roegiers et al., 1995). During this cortical contraction, most of the cortical endoplasmic reticulum (cER) is accumulated in a 2-6 μm thick domain located adjacent to the plasma membrane (PM) at the CP (Speksnijder et al., 1993). Once formed, the CP (and, in particular, the domain of cER that accumulates there) is the site of origin of all the repetitive Ca²⁺ waves (McDougall and Sardet, 1995) (Fig. 1). However, in eggs injected with the Ca²⁺-releasing second messenger inositol (1,4,5)-trisphosphate [Ins(1,4,5)P₃], a CP forms but no repetitive Ca²⁺ waves are generated (McDougall and Levasseur, 1998). Clearly, sperm are required to generate the signal that initiates all the repetitive Ca²⁺ waves and the CP acts as a Ca²⁺-wave pacemaker for an as yet unknown reason.

A sperm-borne Ca²⁺ releasing factor probably activates mammalian (Swann, 1990), nemertean (Stricker, 1997), ascidian (Kyozuka et al., 1998) and perhaps most eggs following sperm-egg fusion. The best candidate so far for the role of `sperm factor' is a novel sperm-specific phospolipase C, PLCζ, identified initially in mouse (Saunders et al., 2002) then human (Cox et al., 2002) sperm that triggers Ca²⁺ oscillations when expressed in mouse eggs (Saunders et al., 2002; Cox et al., 2002). However, in invertebrates egg PLCγ is thought to be involved; dual SH2-domain peptides from PLCγ1 inhibit fertilization in starfish (Carroll et al., 1999), sea urchin (Shearer et al., 1999) and ascidian (Runft and Jaffe, 2000) eggs but not in mammalian eggs (Mehlmann et al., 1998). Consistent with a role for PLCs in the ascidian, all of the repetitive Ca²⁺ waves in the ascidian egg following fertilization rely on the Ins(1,4,5)P₃ receptor (InsP₃R) (McDougall and Sardet, 1995; Yoshida et al., 1998).

Two hypotheses have been proposed to explain InsP₃R-based pacemakers. Pacemakers are controlled by either sensitization of InsP₃Rs, as in HeLa cells (Thomas et al., 2000) or accumulation of InsP₃Rs, as in smooth-muscle cells (Hirose and Iino, 1994). However, in fertilized ascidian eggs, the CP is not the most sensitive part of the egg to Ins(1,4,5)P₃ (Dumollard and Sardet, 2001). These data indicate that InsP₃Rs are probably not accumulated or sensitized sufficiently at the CP to affect the Ca²⁺ release properties. Based on these data it has been suggested that the CP acts as a Ca²⁺-wave pacemaker because InsP₃ is preferentially produced in that region of the egg (Dumollard and Sardet, 2001). These data suggest that factors other than the InsP₃R are accumulated at the CP, and that it is these factors that earmark the site as the Ca²⁺-wave pacemaker.

Here, we have tested four components of the ascidian Ca²⁺-wave pacemaker to determine which component(s) is required for Ca²⁺-wave pacemaker activity: enhanced Ca²⁺ influx at the CP, phosphatidylinositol (4,5)-bisphosphate [PtdIns(4,5)P₂] accumulation at the CP, cER accumulation and sperm factor enrichment at the CP. We have found that: (1) in the absence of external Ca²⁺ the CP remains the wave Ca²⁺ pacemaker; (2) the accumulation of PtdIns(4,5)P₂ at the CP is not required for Ca²⁺-wave pacemaker activity; (3) in the absence of a cER accumulation the Ca²⁺-wave pacemaker in the ascidian egg correlates with the site of sperm-egg fusion; and finally (4) injection of ascidian sperm extract (aSE) into eggs that were prevented from forming a CP and any ER accumulations still resulted in repetitive Ca²⁺ waves that originated at the cortex, suggesting that the sperm factor might be localized cortically. We therefore propose that the Ca²⁺-releasing factor introduced into the egg either by sperm at fertilization or upon injection aSE becomes enriched at the cortical CP thus making that site act as a Ca²⁺-wave pacemaker.

Ascidiella aspersa were collected from Blyth South Harbour, and kept in holding tanks containing natural sea water (NSW) for 1-2 weeks at 10°C. Eggs and sperm were harvested by first removing the outer tunic of the animal and then puncturing the gonoduct. The eggs were then collected by pipetting and transferred to a Petri dish containing filtered sea water (0.2 μm pore size, Millipore). Next, sperm were collected from the same animal using a pipette and transferred to 1.5 ml plastic tubes and stored at 4°C. To remove the outer chorion of the eggs, the chorionated eggs were treated for 45-60 minutes at 18-19°C with 0.1% trypsin (Sigma), then pipetted gently to shear the digested chorion from the eggs. Following dechorionation, the eggs were transferred to Petri dishes that had been previously coated with gelatin (0.1% gelatin/formaldehyde – the formaldehyde had been previously removed by air drying, followed by washing with tap water then sea water). All glass and plastic ware that came into contact with the eggs had also been coated with gelatin, thus preventing the eggs from sticking to those surfaces.

The plasmid encoding the PH domain (a gift from T. Meyer) was amplified by PCR and subcloned into pRN3-GFP, a specialized vector suitable for in vitro transcription of messenger RNA (see Levasseur and McDougall, 2000). mRNA encoding the PH::GFP construct was synthesized using the mMESSAGE mMACHINE kit (Ambion), resuspended in nuclease-free water and stored in 1 μl aliquots at –20°C until required.

The negative control was a non-PtdIns(4,5)P₂-binding variant of the PH::GFP in which the lysine residues at positions 30 and 32 have both been replaced with asparagine by site-directed mutagenesis [using the GeneEditor™ in vitro Site-Directed Mutagenesis System (Promega)]. This was used to transcribe mRNA in exactly the same way.

The PH::GFP coding region was cloned into the pCal-n vector (Stratagene) and transformed into BL21Gold cells (Stratagene). The PH::GFP chimeric protein was isolated using the pCAL-n Affinity® Protein Expression System (Stratagene) according to the manufacturer's instructions. The elution buffer was exchanged by concentrating the protein using Microcon® 50MWCO filters (Amicon) and dilution with ten volumes of injection buffer (0.5 M KCl, 20 mM PIPES pH 7.2, 100 μM EGTA). These concentration/dilution cycles were repeated twice before adjusting the final concentration to 25 mg ml^–1. Aliquots were stored at –80°C until use.

Eggs were introduced into a wedge based on the design of Kiehart (Kiehart, 1982). The microinjection system used has been described previously. Briefly, the injection system consisted of a Narisige three-way hydraulic micromanipulator and compressed air injection system (IM300, Narisige) (McDougall and Levasseur, 1998). Intracellular Ca²⁺ levels were measured using Fura-2-dextran (10 kDa, Molecular Probes), microinjected into eggs to give a final intracellular concentration of 10-20 μM. The mRNA encoding PH::GFP was injected to produce the protein product. Fura-2 and GFP were excited with the appropriate light using a 75 W xenon bulb as the light source. To change the wavelength of excitation appropriate filters (350 nm, 365 nm, 380 nm and 488 nm) housed in a filter wheel (Sutter) were selected using MetaFluor software (Universal Imaging). A triple-pass dichroic filter (Chroma, set 61000) was used to reflect the excited light to the specimen and to pass the appropriate emitted light to a cooled CCD camera (MicroMax, Princeton Instruments). We used an Olympus IX70, fitted with 10×/0.3 NA, 20×/0.75 NA and 60×/1.2 NA water immersion lens for all these experiments.

Ascidian sperm extracts were prepared as previously described (McDougall et al., 2000).

Cytochalasin B (Sigma) was dissolved in distilled water at 1 mg ml^–1 and diluted before use to 2 μg ml^–1. Nocodazole (Sigma) was dissolved in DMSO to 1 mM and diluted prior to use to 2 μM. To visualize the DNA configuration, eggs were bathed in sea water containing 10 μg ml^–1 Hoechst 33342 for 10-20 minutes then excited at 365 nm.

The typical spatiotemporal pattern of Ca²⁺ increase following fertilization of Ascidiella aspersa eggs is shown in Fig.1. Following the initial Ca²⁺ increase (sometimes displaying sinusoidal oscillations, see Fig. 5), a second series of baseline Ca²⁺ spikes was observed (Fig. 1). The first Ca²⁺ wave (Fig. 1B) triggers a cortical contraction of the egg that culminates in formation of a surface protrusion termed the contraction pole (CP) (Roegiers et al., 1995) (Fig. 1C). Following extrusion of the first polar body (pb1), all second-phase Ca²⁺ waves came from the CP and (as previously noted) these waves are faster than the initial wave (McDougall and Sardet, 1995) (Fig. 1D).

To determine whether localized Ca²⁺ influx at the CP was required for the CP to act as a Ca²⁺-wave pacemaker, eggs were fertilized in natural sea water (NSW) and then transferred to nominally Ca²⁺-free sea water supplemented with 2 mM EGTA. In all the eggs examined (n=5) all the repetitive Ca²⁺ waves originated from the CP (the origins of which can be seen in Fig. 2B) even in the absence of external Ca²⁺.

Here we examined whether the specific association of the PH domain of PLC δ-1 with PtdIns(4,5)P₂ could be used to visualize the cellular distribution of PtdIns(4,5)P₂ in an intact ascidian egg during fertilization. To achieve this, a fusion construct between the PH domain of PLC δ-1 and GFP was made (PH::GFP). Either the mRNA encoding this fusion construct or purified PH::GFP protein were microinjected into unfertilized eggs of A. aspersa, where, after sufficient time for expression or translocation, its localization could be observed.

When expressed in the egg, the PH::GFP construct was predominately localized to the CP, where the microvilli are also located (Fig. 3A). Some PH::GFP was still localized on the rest of the PM at this time (Fig. 3C). To examine whether this domain displayed specific binding to PtdIns(4,5)P₂ and could therefore serve as a suitable marker for PtdIns(4,5)P₂ localization, site-directed mutagenesis was performed to replace the basic amino acid in the PtdIns-interacting residues of the PH domain (Lys 30/Asn and Lys 32/Asn). This mutant version of the PH::GFP fusion construct (PH::GFP^**) was uniformly distributed when expressed in the eggs and showed no PM labelling before or during fertilization (Fig. 3B,C, n=5).

Post-acquisition line-scan analysis was performed to compare image intensity in fertilized eggs expressing PH::GFP or PH::GFP^**. For each egg, a line intensity profile was produced (schematically represented on the eggs, Fig. 3C, insets). Eggs expressing PH::GFP revealed a large increase in fluorescence at the CP region shortly after fertilization (Fig. 3B,C). By contrast, however, the PH::GFP^** was uniformly distributed both before and during fertilization (Fig. 3C). Next, we wanted to determine whether the accumulation of PtdIns(4,5)P₂ played a role, if any, during the generation of the periodic Ca²⁺ waves from that site.

To examine whether this accumulation of PtdIns(4,5)P₂ at the CP was due to an increased density of microvilli at the CP or to some other localization mechanism, and to determine the role this accumulation of microvilli and PtdIns(4,5)P₂ played during the generation of Ca²⁺ waves, fertilized eggs were treated with the fungal alkaloid cytochalasin B to disrupt the actin cytoskeleton and so disrupt microvilli. Eggs were previously injected with Fura-2-dextran (10-20 μM) and then injected with mRNA encoding PH::GFP and fertilized. Following fertilization and the accumulation of PtdIns(4,5)P₂ at the CP, eggs were treated with 2 μg ml^–1 cytochalasin B. Within 1 minute of adding cytochalasin B, the eggs rounded up, the microvilli disappeared and the PH::GFP became more uniform around the PM (Fig. 4, compare II and V). The accumulated PtdIns(4,5)P₂ at the CP therefore reflected an accumulation of microvilli. In addition, the Ca²⁺ waves still continued even though the PtdIns(4,5)P₂ and microvilli were confirmed to be absent.

To determine precisely the location of the Ca²⁺-wave pacemaker in the absence of a CP, eggs were labelled with Nile Blue beads to mark the location of CP before addition of cytochalasin B (Fig. 4I, IV, VII). The pb1 also acts as an indicator of the position of the CP because it often lies opposite to the CP (Fig. 4I). Following the addition of cytochalasin B to fertilized eggs, the CP was no longer present and the second phase of Ca²⁺ waves still originated from a vegetal site (Fig. 4IV) close to where the tuft of microvilli and the CP previously resided (Fig. 4I). In addition, the site of the Ca²⁺-wave pacemaker did not colocalize with the position of the sperm aster, as shown by the clear separation between the position of the sperm aster (Fig. 4VII) and the Ca²⁺-wave pacemaker (Fig. 4VI). From these data, we concluded that neither the CP protrusion nor the accumulation of PtdIns(4,5)P₂ was required for the functioning of the Ca²⁺-wave pacemaker because all the second-phase Ca²⁺ waves still originated from the CP protrusion area, even when the CP and localized PtdIns(4,5)P₂ were no longer present.

During contraction of the actomyosin basket, the cER becomes concentrated rapidly at the CP (Speksnijder et al., 1993) and it has been suggested that this cER domain plays a role in the determination of the CP as a Ca²⁺-wave pacemaker (Speksnijder et al., 1993). Before formation of the CP and its associated Ca²⁺-wave pacemaker activity, ER first accumulates around the growing sperm aster and it has been shown that this accumulation of ER correlates with the location of the mobile Ca²⁺-wave pacemaker (Dumollard and Sardet, 2001), which is equivalent to the first phase of sinusoidal Ca²⁺ waves that are sometimes seen during fertilization in A. aspersa (compare Figs 1 and 5). However, recent evidence casts doubt on ER accumulation having a role in Ca²⁺-wave pacemakers, because these ER accumulations are not the part of the egg that is most sensitive to InsP₃ (Dumollard and Sardet, 2001). By preventing ER accumulation at the CP (absence of the cortical contraction served as a positive control for cytochalasin B) and/or at the sperm aster (absence of sperm aster and pb1 extrusion served as positive controls for nocodazole), eggs were fertilized to determine whether the frequency of Ca²⁺ oscillations would be reduced by the absence of an accumulation of ER at the CP and/or sperm aster (as would be expected if these sites act as pacemakers because of their accumulation of ER). Some eggs were previously injected with oil-saturated dialcylcarbocyanine (DiI) to label the ER, thus confirming that ER accumulations were abolished when the cytoskeleton was disrupted (data not shown).

Eggs treated with nocodazole alone displayed both series of Ca²⁺ oscillations (Fig. 5Ai, n=8). The microtubule-based sperm aster did not form and the pbs were not extruded. These eggs displayed the actomyosin-dependent cortical contraction, and the second series of Ca²⁺ waves displayed a frequency not significantly different from control fertilizations (Table 1).

Eggs treated with cytochalasin B alone also displayed both phases of Ca²⁺ oscillations (Fig. 5Aii, n=9). The actomyosin-dependent cortical contraction and pb extrusion were both inhibited. However, the formation of the microtubule-based sperm aster was unaffected. These eggs displayed a frequency of second-phase Ca²⁺ oscillations not significantly different from control fertilizations (Table 1).

When eggs were treated with both nocodazole and cytochalasin B to prevent both forms of ER accumulation, the fertilized eggs still displayed two series of Ca²⁺ waves that all originated from the point of sperm fusion (Fig. 5B, n=18). To determine more precisely the site of sperm-egg fusion, eggs were bathed in the vital DNA dye Hoechst 33342 and a mixture of cytochalasin B (2 μg ml^–1) and nocodazole (2 μM) then injected with Ca²⁺-Green-1-dextran (10-20 μM) and fertilized. Hoechst 33342 staining revealed that the position of the sperm nucleus corresponded to the initiation site of both phases of Ca²⁺ waves when both ER accumulation at the CP and sperm aster formation were prevented (Fig. 5B). Finally, the frequency of Ca²⁺ spikes during the second phase in eggs treated with the cytoskeleton inhibitors was similar to those in untreated eggs (controls) (Table 1). These data indicate that the accumulation of cER at the CP (or around the sperm nucleus) is not required for the generation of the second series of Ca²⁺ waves and has no effect on the frequency of the second phase of Ca²⁺ waves. We did find, however, that the rate of Ca²⁺-wave propagation was slower in eggs treated with cytochalasin B (data not shown).

It has been shown in mouse and ascidian eggs that the cortex is the site of the initial Ca²⁺ release when sperm extract (SE) is injected centrally [mouse (Oda et al., 1999); ascidian (Kyozuka et al., 1998)] and that the sensitivity to SE and InsP₃ is higher in the cortex of mouse oocytes (Oda et al., 1999). In the ascidian, the CP is not the region of the egg that is most sensitive to InsP₃ (Dumollard and Sardet, 2001), even though all of the second-phase Ca²⁺ waves originate there when aSE is injected centrally (McDougall et al., 2000). These data suggest that aSE accumulates near or on the PM at the CP following the cortical contraction.

We were therefore interested in determining whether the cortex plays a role in anchoring the factor present in aSE near the PM. Eggs were injected with Ca²⁺-Green-1-dextran (10-20 μM) and soaked in cytochalasin B (2 μg ml^–1) and nocodazole (2 μM) for approximately 30 minutes. After this, aSE was injected into the centre of the egg (0.1-0.5% cytoplasmic volume, ∼3-20 mg ml^–1 injected by pipette) (Fig. 6). After a lag time of approximately 2 minutes, the initial Ca²⁺ transient originated from a site on the cortex instead of the deeper cytoplasmic injection site (Fig. 6). All the Ca²⁺ waves originated from the near the PM (Fig. 6, 6/6 eggs).

We have examined how the sperm-induced Ca²⁺-wave pacemaker in the ascidian egg functions and conclude that the pacemaker probably contains an accumulation of the sperm factor that generates all the repetitive Ca²⁺ waves. In addition, we find that the accumulated domain of cER, Ca²⁺ influx (following pb1) and PtdIns(4,5)P₂ enrichment at the CP are not required for Ca²⁺-wave pacemaker activity. Finally, we demonstrate that injecting aSE into eggs treated with microfilament- and microtubule-disrupting drugs triggers repetitive Ca²⁺ waves that originate near the PM, and also that all the repetitive Ca²⁺ waves originate at the site of sperm-egg fusion in fertilized eggs prevented from forming a CP.

There are two phases of Ca²⁺ waves in the ascidian. The first is between fertilization and pb1 extrusion (∼0-7 minutes). We discounted Ca²⁺ influx as having a role in setting the placement of the Ca²⁺-wave pacemaker. However, because external Ca²⁺ was present during sperm-egg fusion, it is possible that Ca²⁺ influx during that time somehow sets the site of the Ca²⁺-wave pacemaker. The regulation and pattern of Ca²⁺ oscillations are the same as normal fertilization following injection of aSE (Kyozuka et al., 1998; Runft and Jaffe, 2000; McDougall et al., 2000) where no sperm PM is present. It therefore seems unlikely that external Ca²⁺ influx through sperm-borne Ca²⁺ channels results in the establishment of the Ca²⁺-wave pacemaker because the pacemaker is also established following injection of aSE. The simplest interpretation of our data is that a factor(s) introduced into the egg with the sperm stays near the cortex. Following introduction into the egg, this factor stimulates local production of InsP₃, which in turn spreads across the egg, causing the first Ca²⁺ increase. This Ca²⁺ increase in turn stimulates a cortical contraction, dragging cortical components towards the vegetal hemisphere of the egg (Roegiers et al., 1995). We hypothesize that the sperm factor is also dragged cortically. It is interesting that, before the stable Ca²⁺-wave pacemaker forms at the CP, a mobile Ca²⁺-wave pacemaker is present (McDougall and Sardet, 1995). This mobile Ca²⁺-wave pacemaker correlates with an accumulation of ER around the sperm aster (Dumollard and Sardet, 2001). Here, we define the accumulation or aggregation of ER as the assembly of ER microdomains (∼2 μm diameter) into larger structures (several μm diameter). We hypothesize that the mobile Ca²⁺-wave pacemaker represents the movement of the ascidian sperm factor from the site of sperm-egg fusion towards the vegetal hemisphere of the egg.

The second phase is between pb1 extrusion and pb2 extrusion (∼7-25 minutes after fertilization). Following formation of the CP, all Ca²⁺ waves come from that site. We hypothesize that the CP contains an accumulation of the sperm factor (perhaps not all but an enrichment of it) for the following reasons. When the cortical contraction is abolished, all second-phase Ca²⁺ waves come from the site of sperm-egg fusion. This shows that the factor probably remains near the site of sperm-egg fusion, even though the sperm factor might be a soluble protein. Also, when the CP protrusion was removed pharmacologically, all the Ca²⁺ waves still came from the site previously occupied by the CP rather than the site of sperm-egg fusion. We do not know whether and, if so, how, the sperm factor is anchored to or near the PM but it might be through the association with membrane lipid rafts. It has been shown that lipid rafts function as a signalling scaffold where several cell signalling molecules, including PLCs and Src kinases accumulate (reviewed in Zajchowski and Robbins, 2002) and recent reports have shown that the disruption of these domains perturbs the fertilization signalling pathway in sea urchin and Xenopus eggs (Belton et al., 2001; Sato et al., 2002).

The next problem is how the factor reaches to the CP, because the sperm nucleus clearly borders the CP but is not near its centre (Fig. 4VII). One possibility is that it is accumulated in the CP owing to the cortical contraction; if this is the case, it probably remains near the PM. For example, beads attached to the surface of the egg, near the site of sperm-egg fusion, move further into the CP than the underlying sperm nucleus (Roegiers et al., 1995). This is probably because the sperm nucleus lies on the cytoplasmic side of the actomyosin basket and therefore experiences more viscous drag than PM proteins tethered to the actomyosin basket. Finally, the frequency of Ca²⁺ oscillations is not significantly affected by preventing formation of the CP. This shows that the local accumulation of cER, PtdIns(4,5)P₂ and other factors, such as egg PLCs, at the CP are not required, and that sufficient cER, PtdIns(4,5)P₂, PLCs and other components of the signalling cascade can be found around the egg periphery. The CP is therefore the Ca²⁺-wave pacemaker by default because by preventing its formation, Ca²⁺ waves are still generated at the correct frequency, albeit from the site of sperm-egg fusion. Cytochalasin B treatment does appear to reduce the frequency of Ca²⁺ oscillations in another species of ascidian (Dumollard and Sardet, 2001). However, because the object of that work was to show that three Ca²⁺-wave pacemakers can co-exist in fertilized ascidian zygotes, no statistical analysis was provided concerning the frequency of oscillations. It is therefore impossible to know unequivocally whether cytochalasin treatment leads to a significant reduction in oscillation frequency (Dumollard and Sardet, 2001). It is clear from our present and previous experiments that the range of oscillation frequency is large (McDougall et al., 2000). One of the last aspects of the vegetal Ca²⁺-wave pacemaker that has not been tested is whether it provides a cue to polarity and thus whether it has significance for later developmental events.

There is evidence that, in the mammalian egg, the sperm-specific PLCζ can induce Ca²⁺ oscillations that mimic fertilization (Saunders et al., 2002). In other organisms, such as echinoderms, it has often been reported that PLCγ and Src-like tyrosine kinases are involved in generating the Ca²⁺ increase at fertilization (Carroll et al., 1999; Shearer et al., 1999; Giusti et al., 2003). In ascidians, there is some evidence that PLCγ and Src-like tyrosine kinases are likewise involved in triggering the first Ca²⁺ increase at fertilization (Runft and Jaffe, 2000). However, PLCζ has not yet been identified in a non-mammalian system.

In invertebrates, both PLCγ and forms of Src have been proposed to be involved in the mechanism of egg activation. These proteins are accumulated to the PM when activated (reviewed in Rebecchi and Pentyala, 2000; Featherstone, 1997), so the ascidian sperm factor might be tethered to the membrane by one of these proteins. Alternatively, the ascidian sperm factor might be a form of PLCζ that remains to be identified. As for localisation near the PM, it is not known yet how PLCζ is localized in mammalian eggs following expression of PLCζ copyRNA or fertilization. One final point to make about localization is the nuclear localisation of mammalian sperm factor following the formation of pronuclei (Marangos et al., 2003). If the ascidian sperm factor were located in a pronucleus following fertilization, the male pronucleus would make the ideal choice because it is closest to the CP. However, because the Ca²⁺ oscillations stop before the pronuclei form (McDougall and Levasseur, 1998), it is clear that the Ca²⁺ waves generated by the ascidian sperm can be interrupted in a way that does not require pronucleus formation. The Ca²⁺ oscillations in the ascidian egg correlate well with the activity of cyclin-dependent kinase 1 (Cdk1) activity, pause when Cdk1 activity falls and finally stop when Cdk1 activity falls a second time, a few minutes before formation of the pronuclei (McDougall and Levasseur, 1998). Indeed, Ca²⁺ oscillations can be maintained indefinitely by preventing Cdk1 inactivation (Levasseur and McDougall, 2000).

Finally, we wondered whether ER integrity might play a role in the pattern of Ca²⁺ oscillations observed, but found that the ER remains continuous in both control and cytochalasin-treated eggs during the period that the Ca²⁺ oscillations normally pause (data not shown). To resolve the localisation and regulation of the sperm factor fully will clearly require its identification.

We are grateful to T. Meyer for the gift of the PH::GFP construct. We thank Patrick Chang for critically reading the manuscript. This work was supported by the Wellcome Trust.