Mitochondrial respiration and Ca²⁺ waves are linked during fertilization and meiosis completion

Intracellular Ca²⁺ waves are triggered in somatic cells in response to neuromediators, hormones or growth factors(Berridge, 1997). In a wide variety of egg cells (from plant to human cells), fertilization triggers one to several Ca²⁺ waves. Repetitive Ins(1,4,5)P₃-mediated Ca²⁺ waves lasting over a period of 20 minutes to several hours (called `meiotic Ca²⁺ waves')have been observed in ascidian, mammalian, starfish, molluscan, nemertean and annelid eggs at fertilization (reviewed bySardet et al., 1998;Stricker, 1999;Dumollard et al., 2002). These Ca²⁺ signals are necessary for egg activation and progression through the meiotic cell cycles(Speksnijder et al., 1989;McDougall and Levasseur,1998,McDougall and Levasseur,1998) (reviewed by Nixon et al., 2000). Repetitive Ca²⁺ signals are also thought to be important for early and post-implantation development in mammals(Jones, 1998;Ozil and Huneau, 2001).

Cellular Ca²⁺ homeostasis is the result of Ca²⁺fluxes through the plasma membrane and the intracellular organelles. In every egg studied so far, an extensive and continuous ER network plays a major role in the sperm-triggered Ca²⁺ waves (reviewed bySardet et al., 1998;Dumollard et al., 2002). The network hosts Ca²⁺ channels [Ins(1,4,5)P₃receptors (IP₃R)], which mediate intracellular Ca²⁺release, and Ca²⁺ pumps [sarco-endoplasmic reticulum Ca²⁺ ATPases (SERCAs)]. Little is known about the role of mitochondria in the regulation of sperm-triggered Ca²⁺ signals in eggs. In sea urchins, mitochondria were suggested as a sink for Ca²⁺ released during fertilization(Eisen and Reynolds, 1985;Girard et al., 1991). A role for active mitochondria in the regulation of sperm triggered Ca²⁺oscillations has been suggested in the mouse egg(Liu et al., 2001), but it cannot be inferred from this study what the interplay is between mitochondria and sperm-triggered Ca²⁺ oscillations.

In somatic cells, mitochondria can sequester and release Ca²⁺during Ca²⁺ waves (Boitier et al., 1999; Rizzuto et al.,2000; Hajnoczky et al.,2000). Mitochondrial Ca²⁺ sequestering generally provides a negative feedback on Ins(1,4,5)P₃-induced Ca²⁺ release (IICR) (reviewed byRizzuto et al., 2000;Duchen, 2000). Calcium is also a pivotal `multisite' activator of oxidative phosphorylation in the mitochondria (Hansford, 1994). Calcium activates the dehydrogenases of the Krebs cycle(McCormack et al., 1990) and the electron transport chain (Gunter et al., 1994), and has a direct action on the F0-F1 ATP synthase(Territo et al., 2000). In somatic cells, mitochondrial Ca²⁺ uptake can increase the ratio of NADH/NAD⁺ (i.e. NAD⁺ is reduced to NADH) in the mitochondria (Pralong et al.,1994; Hajnoczky et al.,1995) (reviewed by Duchen,2000) as well as oxygen consumption(Duchen, 2000). Theoretically,mitochondrial ATP production can regulate IICR through ATP^4-, an allosteric regulator of IP₃R opening(Mak et al., 1999;Mak et al., 2001). Therefore active mitochondria may modulate Ca²⁺ signaling at two different levels either via preferential Ca²⁺ fluxes (reviewed byDuchen, 2000;Rizzuto et al., 2000;Hajnoczky et al., 2000) or via local ATP generation (Landolfi et al.,1998; Kennedy et al.,1999).

The first report of an increase in oxygen consumption after fertilization was a study on sea urchins (Warburg,1908), which led to the common belief that fertilization is also accompanied by the metabolic activation of the egg. It is now recognized that respiration in sea urchin eggs and embryos depends solely on electron transport through the mitochondrial respiratory chain, except for a short period after fertilization when secreted ovoperoxidases consume O₂to synthesize H₂O₂(Perry and Epel, 1985;Heinecke and Shapiro, 1992;Yasumasu, 2000). In contrast to sea urchins and molluscs, this non-mitochondrial respiratory burst does not occur in asteroid, echiuroid or ascidian eggs(Schomer and Epel, 1998).

Given the role of Ca²⁺ waves in somatic cells (reviewed byRizzuto et al., 2000;Duchen, 2000), sperm-triggered Ca²⁺ waves may mediate the metabolic activation of the fertilized egg, but this remains to be demonstrated. Fertilized ascidian eggs provide a particularly favorable model to investigate the relationships between meiotic Ca²⁺ waves and mitochondrial activity for several reasons.

(1) Fertilization does not involve a non-mitochondrial respiratory burst(Schomer and Epel, 1998).

(2) Thousands of rod-shaped mitochondria form a dense 7 μm subcortical layer lining the vegetal and equatorial regions of the egg(Roegiers et al., 1995;Roegiers et al., 1999).

(3) Fertilization triggers two series of very stereotyped Ca²⁺waves which are generated by two distinct Ca²⁺ wave pacemakers (PM1 and PM2) (Sardet et al., 1998;Yoshida et al., 1998;Nixon et al., 2000;Dumollard and Sardet, 2001). Pacemakers PM1 and PM2 are located in the cortex of the egg (pacemaker PM2 is at the vegetal pole of the egg) and they are known to rely mostly if not only on a sperm-triggered Ins(1,4,5)P₃ production(McDougall and Sardet, 1995;Yoshida et al., 1998;Runft and Jaffe, 2000).

(4) A third artificial pacemaker (PM3) can be generated opposite to pacemaker PM2 (Dumollard and Sardet,2001).

In this study, we have recorded two parameters of mitochondrial activity(oxygen consumption and mitochondrial NADH levels) during the period of activity of the Ca²⁺ wave pacemakers PM1, PM2 and PM3 of the ascidian egg. We have also examined the effects of mitochondrial inhibitors on the patterns of Ca²⁺ waves initiated by these three pacemakers. Our studies reveal that mitochondria are stimulated by cytosolic Ca²⁺signals. In turn, the Ca²⁺ wave pacemaker PM2 strictly depends upon Ca²⁺ buffering and ATP production by mitochondria.

Specimens of the ascidian Phallusia mammillata, were collected either from Villefranche sur Mer or Sète (Mediterranean). The ascidiansCiona intestinalis and Ascidiella aspersa were purchased from the Marine Resource Center of the Marine Biological Laboratory in Woods Hole (MA, USA). The animals were kept in seawater at 16-22°C before use. Denuded oocytes were prepared, handled and fertilized as described previously(McDougall and Sardet, 1995;Roegiers et al., 1999).

Unfertilized eggs were introduced into a wedge and injected as described previously (McDougall and Sardet,1995; Dumollard and Sardet,2001). The cytosolic dyes Calcium Green dextran (10 kDa, CG) and Texas Red dextran (10 kDa, TR, Molecular Probes) as well as caged Ins(1,4,5)P₃ (cIP₃, Calbiochem), caged PtdIns(4,5)P₂ (cgPIP₂, Calbiochem) and NPE caged ATP (cATP, Calbiochem) were dissolved in injection buffer consisting of 180 mM KCl, 100 μM EGTA, 30 mM BES pH 7.1. Approximately 1% of the egg volume was injected to give final concentrations of 10-20 μM for CG and TR dextran; 5 μM for cIP₃; 35 μM for cgPIP₂; and 1 mM for cATP.

Injected oocytes were mounted in a perfusion chamber and imaged with a Leica Wild Leitz CLSM or TCS SP2 confocal microscope(McDougall and Sardet, 1995;Dumollard and Sardet, 2001). For UV photorelease of cIP₃, cgPIP₂ or cATP oocytes were co-injected with a mixture of CG and TR and the caged compound(Dumollard and Sardet, 2001). UV flashes were produced on the Leica confocal microscope using epifluorescence illumination (75 W mercury lamp). Some photorelease was also performed on an inverted Zeiss microscope (Axiovert 100 TV) essentially as described previously (Dumollard and Sardet,2001).

NADH fluoresces at 470 nm with peak absorption at 356 nm(Masters and Chance, 1993). A 470±20 nm bandpass filter (excitation: bandpass 360±10 nm,Chroma Opticals) was used to collect images of intracellular NADH concentration ([NADH]). The mitochondria-rich vegetal subcortex containing mostly mitochondria (Roegiers et al.,1995; Roegiers et al.,1999) (Fig. 5A,B)displayed a higher fluorescence (Fig. 2A). For each NADH image, the one pixel line running in the center of the entire mitochondria-rich domain (lines ab and a′b′ inFig. 2A) was extracted by image analysis using Visilog software [details about the segmentation method can be obtained by contacting C. C.(cibert@ijm.jussieu.fr)]. To prevent motion artifacts, each extracted line of a NADH image (ab and a′b′ in Fig. 2A)was divided by the same line extracted in an essentially NADH-insensitive reference image [exc, bandpass 380± 10 nm; em, 470nm (Chroma Opticals)]acquired quasi simultaneously with the NADH image. These divided extracted lines (which are ratiometric measurements of [NADH]_mito) were stacked together to display the variations of [NADH]_mito with time(running from left to right in Fig. 2A `time image').

Oxygen fluxes in a single egg were measured using the oxygen-sensitive self-referencing vibrating probe in the NVPF facility at the Marine Biological Laboratory in Woods Hole (MA, USA) (Land et al., 1999). The vibrating electrode was positioned within 5μm of a denuded egg adhering to a coverslip in a petri dish(Fig. 1B). For fertilization experiments, activated sperm were added to the dish, while the egg was observed using time-lapse recording. The presence of Ca²⁺ waves was monitored by recording the wave of contraction that accompanies each Ca²⁺ wave (Roegiers et al.,1999). When indicated, FCCP was added to give a final concentration of 1 μM. In some experiments cIP₃ was photoreleased by UV illumination and [Ca²⁺]_c was monitored by recording the Calcium Green fluorescence using a Zeiss Atto Arc imaging system while simultaneously recording the oxygen fluxes with the oxygen probe.

The eggs were held under the microscope in a perfusion chamber and either FCCP (1 μM in sea water, Sigma), cyanide anion (CN^-; 2 mM in sea water, Sigma) or oligomycin (100 μM in sea water, Sigma) perfused. Washing of the inhibitors was carried out by performing several perfusions of sea water. The perfusions lasted less than 1 minute. During that period no recording could be made.

Fertilization in ascidians elicits two stereotyped series of Ca²⁺ waves (Fig. 1A)initiated by two distinct Ca²⁺ wave pacemakers: PM1 and PM2(Dumollard and Sardet, 2001)(examples of such repetitive Ca²⁺ waves can be seen on our web site athttp://biodev.obs-vlfr.fr/biomarcell/ascidies/calcium.html). Calcium wave pacemaker PM1 starts at the point of sperm entry emitting first a sustained fertilization Ca²⁺ wave (lasting 3 to 4 minutes) followed by two to four smaller Ca²⁺ waves and the emission of the first polar body (pb1, Fig. 1A PM1). After a pause lasting 2 to 4 minutes, a second series of Ca²⁺ waves(for a period of 15-20 minutes) initiates in a vegetal protrusion called the contraction pole. The Ca²⁺ wave pacemaker PM2 produces these waves until the end of meiosis II (Fig. 1A PM2). Pacemaker PM2 is situated in a zone of cortical ER accumulation sandwiched between the plasma membrane and the 7 μm mitochondria-rich layer (see Fig. 5A). All meiotic Ca²⁺ waves traverse this layer of densely packed mitochondria.

To measure the oxygen fluxes around a single ascidian egg, we used an oxygen-sensitive, self-referencing microelectrode(Fig. 1B). This technique has been successfully used to measure local oxygen consumption in living cells and embryos (Land et al., 1999;Trimarchi et al., 2000). We found that the basal oxygen consumption in the vegetal pole region of the unfertilized ascidian egg was 38±10% higher than the oxygen consumption measured near the animal pole region of the same egg (data not shown,n=7). This is consistent with the presence of the bulk of mitochondria in a subcortical location in the vegetal hemisphere of the egg(see Fig. 5A).

Maximum oxygen consumption was measured by uncoupling mitochondria with the permeant protonophore FCCP that induces mitochondria to consume oxygen but also to hydrolyze ATP (see Fig. 3). Upon addition of 1 μM FCCP, the oxygen consumption of the egg increased more than six times in the example shown, reflecting the maximum rate of oxygen reduction by mitochondria(Fig. 1D). The graph shown inFig. 1B represents a measurement made with the oxygen electrode positioned near the vegetal pole of the egg. Although there was some variability due to interspecimen differences or alternative positioning of the electrode, a general pattern of oxygen consumption was observed in the eight eggs examined (five Phallusiaand three Ciona) (Fig. 1B). With the basal level of oxygen consumption before fertilization assigned a value of 100%, we observed that, on average, oxygen consumption rose during the activity of Ca²⁺ wave pacemaker PM1 to 167±18%, and then decreased during extrusion of polar body 1 to 140±15%. Oxygen consumption rose again to reach a maximum value during the period of Ca²⁺ wave pacemaker PM2 activity (173±14%). After the cessation of the meiotic Ca²⁺ waves, the oxygen consumption then stabilized to 126±11%(Fig. 1B).

To measure the impact of the rise in cytosolic Ca²⁺concentration ([Ca²⁺]_c) on oxygen consumption, we artificially raised [Ca²⁺]_c by photoreleasing injected cIP₃ with a global UV flash(Dumollard and Sardet, 2001). Increased oxygen consumption was associated with each Ca²⁺transient caused by the photorelease of cIP₃(Fig. 1C). There was no measurable delay between the onset of the Ca²⁺ transient and the onset of the oxygen consumption transient measured with the vibrating electrode. The decay of Ca²⁺ signal preceded the decrease in oxygen consumption (Fig. 1C).

Calcium can directly or indirectly activate several mitochondrial dehydrogenases of the Krebs cycle, thereby reducing the NAD⁺/NADH pool (i.e. increasing the amount of reduced NADH and decreasing the oxidized NAD⁺, Fig. 5D)(Duchen, 2000;Rizzuto et al., 2000). As NADH is endogenously fluorescent (Masters and Chance, 1993), we imaged NADH in a single ascidian egg(Fig. 2).

The vegetal subcortex of the egg contains mostly mitochondria(Fig. 5A,B)(Roegiers et al., 1995;Roegiers et al., 1999); we therefore chose to measure the variations of [NADH] in this mitochondria-rich domain to monitor the variations of [NADH]_mito. This domain is easily distinguished as a cortical crescent displaying higher NADH fluorescence (Fig. 2A, see Materials and Methods). As with oxygen consumption(Fig. 1B), variations of[NADH]_mito measured in the middle of the mitochondria-rich region(corresponding to the vegetal contraction pole)(Roegiers et al., 1995) showed two main increases during the period of operation of the Ca²⁺ wave pacemakers PM1 and PM2 (Fig. 2B). Blocking the reduction of oxygen by the electron-transport chain with CN^- led to an increase in [NADH]_mito(Fig. 2C). By contrast,perfusion of the uncoupler FCCP, caused a decrease in [NADH]_mito in accordance with its presumed effect on the mitochondria (see Figs3,5). These observations imply that there is a strong positive correlation between meiotic Ca²⁺waves and the activation of mitochondrial respiration.

In order to examine the role of active mitochondria on the pattern,frequency and amplitude of Ca²⁺ waves in the ascidian egg, we used an array of mitochondrial inhibitors, which inhibit Ca²⁺ influx into mitochondria. Energized mitochondria are characterized by an electrochemical potential that is mainly powered by a pH gradient generated by the electron transport chain (Fig. 3E). This electrochemical gradient (around -150 mV) is the driving force for Ca²⁺ influx into the mitochondria via a Ca²⁺uniporter (Gunter et al.,1994; Duchen,2000). Calcium influx can be abolished by simply collapsing the pH gradient across the inner mitochondrial membrane (with CN^- or FCCP see Fig. 3G,H). Perfusion of unfertilized eggs with FCCP or CN^- for 10-20 minutes caused only a slight elevation of [Ca²⁺]_c(Fig. 3A, 0.03±0.02%,n=6, Fig. 3D). By contrast, when mitochondrial inhibitors were perfused after eggs have been activated by cIP₃ injection, they caused a long-lasting increase in the resting [Ca²⁺]_c(Fig. 3B). Upon washout of CN^-, [Ca²⁺]_c returned rapidly to its basal level (Fig. 3B). The FCCP-induced increase of resting [Ca²⁺]_c was faster(0.25±0.05%, n=7, Fig. 3D) than the CN^--induced Ca²⁺ increase(0.10±0.04%, n=9). Interestingly, when Ca²⁺transients were generated by cIP₃ photolysis, the continuous presence of CN^- or FCCP (for more than 10 minutes in the example shown in Fig. 3B) apparently did not affect the rate of the Ins(1,4,5)P₃-mediated Ca²⁺ rise or its decay (Fig. 3B is representative of 10 experiments).

The FCCP- or CN^--induced Ca²⁺ increase described above could be inhibited by preincubating the egg in seawater containing 100μM oligomycin (which blocks the F0-F1 ATP synthase) for 20 minutes(Fig. 3C,D). This strongly suggests that the FCCP/CN^--induced Ca²⁺ increase was due to ATP depletion after the reversal of the mitochondrial ATP synthase when mitochondria were depolarized. The ATP depletion provoked by uncoupled mitochondria may well be responsible for an increased Ca²⁺ leak into the cytosol. Under these new net Ca²⁺ fluxes, the resting[Ca²⁺]_c equilibrates to a new, slightly higher level(Fig. 3B,Fig. 4C,E). The resting[Ca²⁺]_c did not increase further after 3 minutes of mitochondrial inhibitor perfusion, indicating that some Ca²⁺extrusion from the cytosol could still operate (see Discussion).

Fig. 4A,B,D illustrates the differential sensitivity of the two physiological pacemakers (PM1 and PM2) and of the artificial pacemaker PM3 (Dumollard and Sardet, 2001) to FCCP perfusion. Calcium wave pacemaker PM1 triggered by sperm entry was not affected by the presence of FCCP before or during fertilization (Fig. 4A,n=3), while the artificial pacemaker PM3 created by global photorelease of cgPIP₂ in the unfertilized egg was only slightly affected (Fig. 4D). We observed a gradual decrease of peak [Ca²⁺]_c and a gradual increase of inter-spike [Ca²⁺]_c about 4 minutes after the perfusion of FCCP (n=4, Fig. 4D). In contrast, the continuous presence of FCCP inhibited the pacemaker PM2 (n=3, Fig. 4A), which stopped emitting Ca²⁺ waves within seconds after the perfusion of FCCP (n=4,Fig. 4B). In addition,perfusing FCCP after the meiotic Ca²⁺ waves had ceased induced a transient elevation of [Ca²⁺]_c superimposed to the slow and sustained Ca²⁺ rise due to mitochondrial ATP hydrolysis(Fig. 4C, n=3;Fig. 4E, n=4). Importantly, this transient rise in [Ca²⁺]_c was localized to the vegetal hemisphere of the egg where the subcortical mitochondria-rich layer is located (data not shown). This FCCP-releasable Ca²⁺ pool was not present in the unfertilized egg(Fig. 3A) and it is different from the FCCP- or CN^--induced Ca²⁺ leak observed in eggs activated by Ins(1,4,5)P₃ or fertilized eggs. The most likely interpretation is that this Ca²⁺ pool resides in mitochondria that have sequestered and accumulated Ca²⁺ during the passage of the multiple Ca²⁺ waves triggered by PM1 and PM2.

CN^- had a similar effect to FCCP on the sperm-triggered pacemakers PM1 and PM2: pacemaker PM2 was inhibited, but PM1 was not perturbed(Fig. 4E, n=4;Fig. 4F, n=4). As with FCCP, PM2 activity recovered rapidly and completely upon washout of CN^- (Fig. 4E,F). The similar effects of the two different mitochondrial inhibitors on the different pacemakers demonstrate that active mitochondria are necessary for the maintenance of Ca²⁺ wave pacemaker PM2 but that they only play minor roles in sustaining the activities of pacemakers PM1 or PM3.

In order to examine the role played by the decreased level of ATP in the inhibition of pacemaker PM2, we used UV photolysis of NPE-caged ATP(He et al., 1998). Fertilized eggs were first exposed to CN^- and then flashed with UV light to increase [ATP]_c globally. Increasing [ATP]_c in fertilized eggs perfused with CN^- partially restored Ca²⁺ wave pacemaker PM2 activity(Fig. 4F). Global photorelease of ATP caused Ca²⁺ waves to initiate in the contraction pole where PM2 is located (data not shown). However recovery of PM2 activity was only partial compared with the rapid and complete recovery observed after washout of CN^- (Fig. 4F). These observations suggest that mitochondrial ATP hydrolysis and the resultant decrease in [ATP]_c accounts in part for the inhibition of the Ca²⁺ wave pacemaker PM2 by mitochondrial inhibitors.

The role of mitochondrial Ca²⁺ influx on the activity of the Ca²⁺ wave pacemaker PM2 was then investigated by perfusing CN^- on fertilized eggs in the presence of oligomycin. When incubated for 20 minutes in the presence of oligomycin, eggs were fertilized normally and the Ca²⁺ wave pacemakers PM1 and PM2 were not profoundly affected (Fig. 4G,n=3). CN^- perfusion during the period of Ca²⁺wave pacemaker PM2 activity in these eggs still drastically inhibited PM2(Fig. 4G) even though no ATP hydrolysis by mitochondria occurs under these conditions (see drawings inFig. 3H,I). Therefore, the rapid inhibition of pacemaker PM2 is probably due to a rapid mitochondrial depolarization that blocks mitochondrial Ca²⁺ influx. Cytosolic Ca²⁺ buffering by the mitochondria may thus be crucial for Ca²⁺ wave pacemaker PM2 activity, particularly in the presence of oligomycin a situation during which no mitochondrial ATP synthesis can occur(see Discussion). Finally we photoreleased NPE caged ATP during PM2 activity,in order to characterize a potential regulation of PM2 activity by intracellular ATP during the normal operation of PM2 (i.e. without any mitochondrial inhibitors, Fig. 4H, n=4). The most striking effect of adding exogenous ATP during Ca²⁺ wave pacemaker PM2 activity was to decrease the period between successive waves elicited by pacemaker PM2 (bars inFig. 4H show this decrease is of 21% in this experiment).

Together, these observations suggest that complex functional interactions between mitochondria and surrounding Ca²⁺-release sites on the ER network occur during the operation of Ca²⁺ wave pacemaker PM2 and that they involve both local Ca²⁺ fluxes and mitochondrial ATP production.

This study provides the first experimental evidence that both meiotic Ca²⁺ waves and Ins(1,4,5)P₃-mediated Ca²⁺ signals can stimulate mitochondrial respiration in an egg. Therefore, meiotic Ca²⁺ waves not only control egg activation and the completion of the meiotic cell cycles (Spesknijder et al., 1989;McDougall and Sardet, 1995;Yoshida et al., 1998;Levasseur and McDougall, 1998), but they also stimulate the energetic metabolism of the zygote.

Our experiments with mitochondrial inhibitors and caged ATP also show that the stable vegetal Ca²⁺ wave pacemaker PM2, which operates in ascidian zygotes during completion of meiosis II, strictly depends on mitochondrial activity in contrast to pacemakers PM1 and PM3. Because Ca²⁺ wave pacemaker PM2 is located in a cortical domain of ER accumulation sandwiched between the plasma membrane and the subcortical mitochondria-rich domain (McDougall and Sardet, 1995; Roegiers et al.,1999; Dumollard and Sardet,2001), we discuss our results in the light of possible functional relationships between ER and mitochondria.

Using two non-invasive recording techniques, we have been able to measure oxygen consumption and mitochondrial NADH levels of individual ascidian eggs. Basal oxygen consumption of the egg was highest in the vegetal pole region where the bulk of subcortical mitochondria are located(Roegiers et al., 1995;Roegiers et al., 1999;Dumollard and Sardet, 2001),suggesting that the oxygen consumption we measured was mainly due to mitochondrial respiration. The increases in both O₂ consumption and[NADH]_mito observed at fertilization indicate that mitochondrial activity is stimulated by sperm entry.

Ca²⁺ appears to be the stimulus for increased mitochondrial respiration as both O₂ consumption and [NADH]_mitoincrease during the activity of the Ca²⁺ wave pacemakers PM1 and PM2. Most remarkably, Ca²⁺ itself is sufficient to produce the increase in O₂ consumption as demonstrated by the transient burst of O₂ consumption in response to Ins(1,4,5)P₃. These data suggest that the repetitive Ca²⁺ waves emitted by the two pacemakers (PM1 and PM2) are responsible for the sperm-triggered activation of the energetic metabolism in the egg.

The activation of mitochondrial metabolism by the sperm-triggered Ca²⁺ waves suggests that, in these eggs, mitochondria take up cytosolic Ca²⁺ during the passage of a Ca²⁺ wave. The appearance of a FCCP-releasable Ca²⁺ pool after the activity of the pacemakers PM1 and PM2 further supports such hypothesis. Indeed after meiotic Ca²⁺ oscillations have occurred, FCCP perfusion induced a Ca²⁺ transient in the vegetal region of the egg (superimposed on the slow and sustained increase in [Ca²⁺]_c induced by FCCP or CN^- perfusion in artificially activated or fertilized eggs)that most probably reflected release of Ca²⁺ from the mitochondria upon mitochondrial depolarization.

Although mitochondria were suggested to be a sink for the Ca²⁺released during the fertilization wave in sea urchin eggs(Eisen and Reynolds, 1985;Girard et al., 1991), no link to mitochondrial respiration could be established because of the concomitant non-mitochondrial respiratory burst occurring in these eggs at fertilization(Schomer and Epel, 1998;Yasumasu, 2000). In a very recent study, mitochondrial Ca²⁺ has been suggested to play a role in the control of mitochondrial respiration in sea urchin eggs(Fujiwara et al., 2001). However this study predicts that a high [Ca²⁺]_mito in the unfertilized sea urchin egg would inhibit mitochondrial respiration(Fujiwara et al., 2001), which is in complete contradiction with the previous studies on somatic cells, on sea urchins (Eisen and Reynolds,1985; Girard et al.,1991) and with our findings in ascidian eggs.

The presence in ascidian eggs of three Ca²⁺ wave pacemakers with distinct locations and Ca²⁺ wave patterns provides an ideal model system with which to study the influence of mitochondrial activity on the regulation of intracellular Ca²⁺ waves. Perfusions of mitochondrial inhibitors during the period of activity of the different pacemakers revealed that, whereas Ca²⁺ wave pacemaker PM2 was immediately and completely blocked upon FCCP or CN^- perfusion, the pacemakers PM1 and PM3 were only slightly altered after lengthy perfusions of these mitochondrial inhibitors. Several observations suggest that the rapid inhibition of the Ca²⁺ wave pacemaker PM2 is not caused by emptying Ca²⁺ stores due to a global depletion in ATP levels. First, the activated egg could still respond to Ins(1,4,5)P₃ after more than 20 minutes of FCCP perfusion and the fertilized egg could respond to Ins(1,4,5)P₃ during PM2 inhibition by FCCP or CN^- (R. D., unpublished). Second, the resting[Ca²⁺]_c remained relatively low for long periods of time during metabolic inhibition of the pacemaker PM2, implying that ATP-dependent Ca²⁺ extrusion from the cytosol was still operating, with a part of it probably refilling the ER Ca²⁺ stores. Therefore, mechanisms other than global ATP depletion must account for the inhibition of Ca²⁺ wave pacemaker PM2 by mitochondrial inhibitors.

In most eggs studied so far, Ins(1,4,5)P₃-induced Ca²⁺ release from the ER Ca²⁺ stores supports the repetitive Ca²⁺ waves observed at fertilization. Specifically in ascidians, it is known that injecting heparin(McDougall and Sardet, 1995)or the function blocking antibody specific to the IP₃R1 (18A10)(Yoshida et al., 1998) both inhibit the pacemaker PM2. Moreover, blocking specifically PLCγ activity inhibits both pacemakers (Runft and Jaffe,2000) further illustrating that the pacemakers PM1 and PM2 rely mostly if not only on a sperm-triggered Ins(1,4,5)P₃production. Given the central role of Ins(1,4,5)P₃mediated Ca²⁺ release in PM2 activity, IP₃Rs may be a key target for the regulation by mitochondria. There are at least two mechanisms by which mitochondria can regulate the opening of the IP₃R channels: first, by locally buffering[Ca²⁺]_c; second, by controlling the local concentration of ATP (as MgATP^2- and ATP^4- ions).

In recent years it has been established that mitochondria sense the high[Ca²⁺]_c in the vicinity of the IP₃Rs and take up significant amount of Ca²⁺ during the passage of Ca²⁺waves (Hajnoczky et al., 2000;Duchen, 2000,Rizzuto et al., 2000). In such`Ca²⁺ microdomains' created between clusters of IP₃Rs and juxtaposed mitochondria (Fig. 5D), the local Ca²⁺ buffering provided by mitochondria can promote IP₃R opening or closure depending on a bell shape sensitivity to Ca²⁺ at a given[Ins(1,4,5)P₃]_c (i.e. Ca²⁺ above a low threshold level promotes opening, while higher Ca²⁺ levels inhibit IP₃R opening). At low Ins(1,4,5)P₃levels, moderate Ca²⁺ levels inhibit IP₃R opening, while under a higher concentration of Ins(1,4,5)P₃, higher levels of Ca²⁺ are required to inhibit IP₃R opening(Bootman and Lipp, 1999;Mak et al., 1999;Mak et al., 2001).

Our data suggest that inhibiting mitochondrial Ca²⁺ accumulation during PM2 activity using CN^-, while blocking mitochondrial ATP hydrolysis with oligomycin rapidly stops the ability of the pacemaker PM2 to generate Ca²⁺ waves. This finding is consistent with the hypothesis that at the low Ins(1,4,5)P₃ levels driving the pacemaker PM2 (Dumollard and Sardet,2001; Dumollard et al.,2002), Ca²⁺ buffering by mitochondria is necessary to keep the local [Ca²⁺]_c in the activating part of the bell shape curve. The reason why PM1 is insensitive to inhibition of mitochondrial Ca²⁺ accumulation is that it is driven by a larger increase in Ins(1,4,5)P₃ levels induced at the activation wave (Dumollard and Sardet,2001; Dumollard et al.,2002). Under these conditions of high Ins(1,4,5)P₃ levels, the [Ca²⁺]_cmust reach higher levels to inhibit IP₃Rs opening and Ca²⁺ buffering by mitochondria would not be so crucial.

A role for local control of ATP is indicated by the finding that photoreleasing ATP^4- in eggs in which PM2 was inhibited by CN^- perfusion caused a reinitiation of Ca²⁺ waves from the contraction pole. Mitochondrial ATP production in an `ATP microdomain' (as shown in Fig. 5D) can modulate intracellular Ca²⁺ release(Yang and Steele, 2000). ATP^4- is an allosteric regulator of IP₃R1 opening(Mak et al., 1999;Mak et al., 2001) the predominant IP₃R in the egg cell(Fissore et al., 1999; Brind et al., 2001) and also the most sensitive to ATP(Miyakawa et al., 1999;Maes et al., 2000). Electrophysiological studies have revealed that ATP^4- (but not the MgATP^2- complex) sensitizes the IP₃R to activation by Ca²⁺ (Mak et al.,1999; Mak et al.,2001). Moreover when[Ins(1,4,5)P₃]_c is low, ATP^4- can also suppress Ca²⁺-dependent inhibition of the channel opening(Mak et al., 2001). During mitochondrial inhibition of the pacemaker PM2, some of the photoreleased ATP^4- will directly bind to the IP₃Rs, thereby suppressing Ca²⁺-dependent inhibition of the channels and rendering the vegetal contraction pole region excitable for the initiation of global Ca²⁺ waves. In addition, the remaining ATP^4- ions will bind Mg²⁺ ions and form MgATP^2- complexes that can be used by SERCAs to refill the ER Ca²⁺ stores(Fig. 5D). Such a model does explain why, while PM2 is completely blocked by mitochondrial inhibition, the egg can still respond to the rather high Ins(1,4,5)P₃increases achieved by uncaging cIP₃ (between 100 and 500 nM after each UV flash) (Dumollard and Sardet,2001) and PM1 (which functions under high Ins(1,4,5)P₃ levels) is not altered by mitochondrial inhibition. This model can also account for the increase in frequency of Ca²⁺ oscillations generated by PM2 after the photorelease of ATP^4-. As the threshold [Ca²⁺]_c for triggering a wave can be lowered by increased ATP^4-(Mak et al., 1999;Mak et al., 2001) this threshold can be reached more quickly after a wave.

Taken together, our observations and the above considerations indicate that functional interactions between ER and mitochondria occur during sperm-triggered Ca²⁺ waves via two mechanisms: local Ca²⁺ fluxes between the two organelles and mitochondrial production of ATP. While local Ca²⁺ buffering in the `Ca<sup>2+</sup>microdomain' may suppress the Ca<sup>2+</sup> inhibition of the IP<sub>3</sub>R, local ATP production in the `ATP microdomain' would sensitize and potentiate IICR (Fig. 5D). Even though this work raises the possibility that both modes of communication between ER and mitochondria are at work during the generation of sperm-triggered Ca²⁺ waves, only precise measurement of[ATP^4-]_c, [MgATP^2-]_c,[Ca²⁺]_c in the `Ca²⁺ and ATP microdomain' in the living zygote will allow us to define the complex roles mitochondria play in the regulation of the meiotic Ca²⁺ wave pacemakers.

We thank John Carroll for critical reading of the manuscript, and Michael Duchen for helpful suggestions and advice. This work was supported by AFM and ARC grants to C. S.