Calcium Signaling by Cyclic ADP-ribose, NAADP, and Inositol Trisphosphate Are Involved in Distinct Functions in Ascidian Oocytes*

ADP-ribosyl cyclase catalyzes the synthesis of two structurally and functionally different Ca2+releasing molecules, cyclic ADP-ribose (cADPR) from β-NAD and nicotinic acid-adenine dinucleotide phosphate (NAADP) from β-NADP. Their Ca2+-mobilizing effects in ascidian oocytes were characterized in connection with that induced by inositol 1,4,5-trisphosphate (InsP3). Fertilization of the oocyte is accompanied by a decrease in the oocyte Ca2+ current and an increase in membrane capacitance due to the addition of membrane to the cell surface. Both of these electrical changes could be induced by perfusion, through a patch pipette, of nanomolar concentrations of cADPR or its precursor, β-NAD, into unfertilized oocytes. The changes induced by β-NAD showed a distinctive delay consistent with its enzymatic conversion to cADPR. The cADPR-induced changes were inhibited by preloading the oocytes with a Ca2+ chelator, indicating the effects were due to Ca2+ release induced by cADPR. Consistently, ryanodine (at high concentration) or 8-amino-cADPR, a specific antagonist of cADPR, but not heparin, inhibited the cADPR-induced changes. Both inhibitors likewise blocked the membrane insertion that normally occurred at fertilization consistent with it being mediated by a ryanodine receptor. The effects of NAADP were different from those of cADPR. Although NAADP induced a similar decrease in the Ca2+ current, no membrane insertion occurred. Moreover, pretreatment of the oocytes with NAADP inhibited the post-fertilization Ca2+ oscillation while cADPR did not. A similar Ca2+ oscillation could be artificially induced by perfusing into the oocytes a high concentration of InsP3 and NAADP could likewise inhibit such an InsP3-induced oscillation. This work shows that three independent Ca2+ signaling pathways are present in the oocytes and that each is involved in mediating distinct changes associated with fertilization. The results are consistent with a hierarchical organization of Ca2+ stores in the oocyte.

ADP-ribosyl cyclase catalyzes the synthesis of two structurally and functionally different Ca 2؉ releasing molecules, cyclic ADP-ribose (cADPR) from ␤-NAD and nicotinic acid-adenine dinucleotide phosphate (NAADP) from ␤-NADP. Their Ca 2؉ -mobilizing effects in ascidian oocytes were characterized in connection with that induced by inositol 1,4,5-trisphosphate (InsP 3 ). Fertilization of the oocyte is accompanied by a decrease in the oocyte Ca 2؉ current and an increase in membrane capacitance due to the addition of membrane to the cell surface. Both of these electrical changes could be induced by perfusion, through a patch pipette, of nanomolar concentrations of cADPR or its precursor, ␤-NAD, into unfertilized oocytes. The changes induced by ␤-NAD showed a distinctive delay consistent with its enzymatic conversion to cADPR. The cADPR-induced changes were inhibited by preloading the oocytes with a Ca 2؉ chelator, indicating the effects were due to Ca 2؉ release induced by cADPR. Consistently, ryanodine (at high concentration) or 8-amino-cADPR, a specific antagonist of cADPR, but not heparin, inhibited the cADPRinduced changes. Both inhibitors likewise blocked the membrane insertion that normally occurred at fertilization consistent with it being mediated by a ryanodine receptor. The effects of NAADP were different from those of cADPR. Although NAADP induced a similar decrease in the Ca 2؉ current, no membrane insertion occurred. Moreover, pretreatment of the oocytes with NAADP inhibited the post-fertilization Ca 2؉ oscillation while cADPR did not. A similar Ca 2؉ oscillation could be artificially induced by perfusing into the oocytes a high concentration of InsP 3 and NAADP could likewise inhibit such an InsP 3 -induced oscillation. This work shows that three independent Ca 2؉ signaling pathways are present in the oocytes and that each is involved in mediating distinct changes associated with fertilization. The results are consistent with a hierarchical organization of Ca 2؉ stores in the oocyte.
Ca 2ϩ signaling in cells generally involves both its influx from the extracellular medium and its release from intracellular stores. Two families of intracellular Ca 2ϩ release channels have been characterized, namely, the inositol 1,4,5-trisphosphate receptors (InsP 3 R) 1 and the ryanodine receptors (RyR). Unlike the InsP 3 R, the physiological agonist for RyR has not been identified although it is known that its Ca 2ϩ channel property can be modulated by a plant alkaloid, ryanodine. The skeletal muscle RyR isoform is known to be activated directly by conformational coupling with the voltage sensors at the cell surface and does not seem to require an agonist (1,2). On the other hand, other RyR isoforms may require specific ligand for activation, and cyclic ADP-ribose (cADPR) has been proposed to be such an agonist for RyR (3). Cyclic ADP-ribose was discovered during an investigation of Ca 2ϩ signaling mechanisms in sea urchin eggs (4 -6). Subsequent work on a variety of cells indicates cADPR is likely an endogenous regulator of the RyR (7-10). Cyclic ADP-ribose is synthesized from ␤-NAD by the enzyme ADP-ribosyl cyclase, which was first described in sea urchin eggs (4,6) and has since been shown to be ubiquitous (Ref. 11, and reviewed in Ref. 12). A soluble form of the cyclase has been purified from Aplysia ovotestis, and its crystalline structure has been solved recently (13)(14)(15). A membrane-bound homolog of the Aplysia cyclase is CD38, a lymphocyte antigen (16). CD38 is catalytically different from the Aplysia cyclase in that it not only can cyclize ␤-NAD into cADPR but can also catalyze the hydrolysis of cADPR to ADP-ribose (17). Both CD38 and the soluble cyclase can, additionally, catalyze the exchange of the nicotinamide group of ␤-NADP with nicotinic acid to produce nicotinic acid-adenine dinucleotide phosphate (NAADP) (18), a metabolite that can release Ca 2ϩ from intracellular stores different from that mobilized by cADPR (19).
Ca 2ϩ is involved in a complex series of changes accompanying fertilization of ascidian oocytes (20 -23). An immediate decrease in depolarization-activated Ca 2ϩ current is followed by a progressive increase in membrane capacitance (24). The latter is indicative of membrane insertion into the cell surface. Subsequently, a prolonged Ca 2ϩ oscillation occurs, lasting until meiosis is complete. Perfusion of high concentrations of inositol 1,4,5-trisphosphate (InsP 3 ) into unfertilized oocytes through a whole-cell patch pipette can elicit repetitive Ca 2ϩ transients similar to that observed during fertilization (24). However, neither the rapid decrease of Ca 2ϩ current nor membrane insertion is induced by InsP 3 , suggesting other Ca 2ϩ messengers may be involved in mediating these early changes at fertilization (24). In this study, we used the whole-cell patchclamp technique to introduce three independent Ca 2ϩ messengers, cADPR, NAADP, and InsP 3 into the oocytes (25). The resulting Ca 2ϩ mobilization was monitored by two independent methods, by electrical measurements to sense the localized Ca 2ϩ changes near the plasma membrane and by fluorescent indicator to sample the cytoplasmic Ca 2ϩ changes. Results show that all three Ca 2ϩ signaling mechanisms are present in the oocytes and are involved not only in mediating the early electrical changes at the cell surface associated with fertilization but also in modulating the subsequent Ca 2ϩ oscillation.

EXPERIMENTAL PROCEDURES
Ascidian Eggs-As described previously (24), specimens of the hermaphroditic ascidian Phallusia mammillata were collected near Sète on the French Mediterranean coast. Mature oocytes were extracted from the oviduct and kept in artificial sea water (ASW). Sperm was drawn directly from the spermiduct. Chorions and follicle cells surrounding the oocyte were removed either manually, using fine sharpened tungsten needles, or enzymatically (26). Fertilization was induced by inseminating dechorionated eggs with a dilute suspension of sperm.
Solutions-Similar media were described in a previous study (24). The oocytes (ϳ130 m diameter) were filled by diffusion through the whole-cell patch pipette within minutes. The concentrations of the stock solution of cADPR and NAADP were verified by absorbance measurements at 260 nm and by using the published values for extinction coefficients (6,19). Electrophysiological Recordings-Dechorionated eggs in ASW were patch-clamped in the whole-cell configuration using pipettes pulled to resistances of 1.5-3 megohms with a Mecanex BB-CH puller (Geneva, Switzerland). Currents were recorded under voltage-clamp conditions FIG. 1. Calcium release activated by cADPR in the ascidian oocyte. A, perfusion of cADPR (10 nM) into an oocyte induced a rapid decrease in Ca 2ϩ current and an increase in oocyte capacitance. B, these effects were abolished when 5 mM BAPTA was perfused into the oocyte prior to cADPR application, indicating that the observed electrical changes were due to a Ca 2ϩ release activated by cADPR. C, concentration-response curve for the current change induced by cADPR. The Ca 2ϩ current was measured 2 min after the perfusion of cADPR into the oocyte (means Ϯ S.D., 37 oocytes measured). Compounds were perfused into the oocyte through a patch pipette in the whole-cell configuration. The oocyte Ca 2ϩ current was measured subsequently by depolarization pulses between -60 and ϩ30 mV in steps of ϩ10 mV from a holding potential of -80 mV. The peak current intensity measured after a depolarization to -20 mV was used in the plot. The oocytes were incubated in ASW solutions at room temperature. The details of the current measurement were as described previously (24) and are outlined under "Experimental Procedures." with an RK300 amplifier (BioLogic, Claix, France) and pCLAMP software (Axon Instruments, Foster City, CA). Series resistance was compensated electronically (0.5-1.5 megohms). Eggs were successively patched with different pipettes, allowing for changes in pipette solutions as described previously (24,25). Eggs were patch-clamped in the whole-cell configuration and recorded for hours, with the same pipette or with different pipettes, without disturbing the cell integrity (24,25,27,28). Capacitance of the oocyte membrane was evaluated from the steps of current induced under triangle-wave voltage command, as previously explained (29,30).
Calcium Imaging-Most of the procedure was performed as described previously (24). Dechorionated eggs were loaded with 1 mM Fura-2 dextran applied during a 10-min period by diffusion from the patch pipette. Eggs were observed using an epifluorescence Nikon Diaphot 300 microscope through a CF-Fluor ϫ20 objective (numerical aperture 0.75). Fluorescence was collected from the entire cell. Eggs were illuminated at 340 and 380 nm successively using a Lambda 10 optical filter changer (Sutter Instrument Company) and a Technical-Video epifluorescence DX-5 device (Woods Hole, MA). Fluorescence emission was recorded at 510 nm using an Extended Isis CCD camera (Photonic Science, Robertsbridge, UK). Data acquisition and calcium measurement were performed with the Starwise Fluo 220 system (Imstar, Paris). During the acquisition process, at each sampling time, the two source intensity images (excited at 340 and 380 nm) were stored on a hard disk. The calculations of calcium concentration were done according to the formula of Grynkiewicz (31) with predefined values of calibration parameters to provide a pixel by pixel ratiometric image of [Ca 2ϩ ] i as in ref. (24). In the results that are shown below, [Ca 2ϩ ] i were averaged over the apparent whole egg diameter. Data were sampled at 4-s intervals.

RESULTS
cADPR and ␤-NAD Activate Calcium Release and Induce Changes in the Oocyte Calcium Current and Membrane Capacitance-Perfusion of cADPR into an unfertilized ascidian egg through a whole-cell patch pipette elicited two electrophysiological effects, a rapid decrease in the depolarization-activated Ca 2ϩ current and a concomitant but slower increase in membrane capacitance (Fig. 1A). The latter change is indicative of membrane insertion into the cell surface (24). The magnitude of these changes were highly significant, amounting to about 60% decrease in the Ca 2ϩ current and 20% increase in capacitance. In control experiments, perfusion with standard pipette solution without cADPR produced no observable electrical changes (not shown). The measured amplitude of the Ca 2ϩ current remained constant even when the same oocyte was repeatedly patched by different pipettes (24,25,27,28). If the oocyte was preloaded by perfusion with a high concentration of ryanodine (100 M), a RyR blocker, application of cADPR subsequently was ineffective in eliciting any changes (24). In contrast, an inhibitor of the InsP 3 R, heparin, at up to 3 mg/ml, did not prevent the cADPR effects (not shown). These results indicate that cADPR-induced changes may be mediated by a RyRlike Ca 2ϩ release channel. That Ca 2ϩ release indeed was involved was further demonstrated by preloading the oocytes with a Ca 2ϩ chelator, BAPTA. As illustrated in Fig. 1B, subse-FIG. 2. Effects of ␤-NAD on oocyte Ca 2؉ current and membrane capacitance. A, the current and capacitance changes induced by ␤-NAD were qualitatively similar to those triggered by cADPR except with slower kinetics. The initial velocity of the capacitance change shown was about 20 pF/min. It was significantly slower than that activated by cADPR (about 50 pF/min) as shown in Fig. 1. B, the time required to reach the half-maximal current change induced by cADPR and ␤-NAD was compared. There was a time lag of about 120 s for ␤-NAD compared with cADPR at various concentrations, from 1 nM to 10   quent application of cADPR produced no change in either the Ca 2ϩ current or the membrane capacitance. These electrical measurements thus provide a very sensitive assay for monitoring Ca 2ϩ release especially in the localized regions adjacent to the plasma membrane. A concentration-response curve summarizing 37 recordings of the Ca 2ϩ current is shown in Fig. 1C. The half-maximal effective concentration of cADPR was about 0.1 nM, among the lowest values ever reported (32).
␤-NAD, the precursor of cADPR, was similarly effective in inducing the two electrical changes as shown in Fig. 2A. However, the kinetics were significantly slower, especially the decrease of the Ca 2ϩ current. The half-time for the current decrease induced by ␤-NAD was 170 s, 3-4 times longer than that induced by cADPR (Fig. 2B). The rate of the capacitance increase induced by ␤-NAD was similarly slower, with an initial rate of about 20 pF/min (cf. Fig. 2A) as compared with about 50 pF/min in the case of cADPR (cf. Fig. 1A). This time lag is consistent with the presence of an ADP-ribosyl cyclase in the oocytes converting ␤-NAD to cADPR, which is then responsible for inducing the observed electrical changes. It would thus be expected that 8-amino-cADPR, a specific antagonist of cADPR (33), should block the effects of ␤-NAD. This was found to be the case as shown in Fig. 3. Perfusion of a high concentration (1 M) of 8-amino-cADPR by itself did not elicit any current or capacitance changes but totally blocked the activating effects of both ␤-NAD and its product, cADPR (Fig. 3, A and B).
Fertilization of ascidian oocytes is accompanied by a large increase in membrane capacitance similar to that induced by cADPR (24,27,30). Evidence suggests that the fertilizationinduced membrane insertion is mediated by Ca 2ϩ released via the RyR (24). As will be shown later (cf. Fig. 9B), this membrane insertion was totally abolished by 8-amino-cADPR.
NAADP Activates Calcium Release and Decreases the Oocyte Calcium Current without Affecting Membrane Capacitance-Results above suggest that an ADP-ribosyl cyclase is present and operative in ascidian oocytes. The cyclase is known to be a multifunctional enzyme capable of catalyzing not only the cyclization of ␤-NAD to produce cADPR but also an exchange reaction to produce yet another Ca 2ϩ release metabolite from ␤-NADP, NAADP (Ref. 18, and reviewed in Ref. 12). The Ca 2ϩ stores sensitive to NAADP in sea urchin eggs homogenates can be separated from those sensitive to cADPR and InsP 3 by density centrifugation (19) and appear to possess a Ca 2ϩ transport system that is insensitive to thapsigargin (34). The effects of NAADP in ascidian oocytes are likewise very different from those of cADPR. Although perfusion of nanomolar concentrations of NAADP into the oocytes induced a decrease in Ca 2ϩ current, the change was much slower than that induced by cADPR (compare Figs. 1A and 4A). The concentration-response curve of NAADP is shown Fig. 4B, summarizing 24 assays with NAADP concentrations ranging from nanomolar to micromolar. The half-maximal effective concentration was about 3 nM, 30-fold higher than that of cADPR (Fig. 1C).
The most dramatic difference between the effects of NAADP and cADPR was, however, in the capacitance measurements. FIG. 4. Effects of NAADP on the oocyte current and capacitance. A, perfusion of NAADP (10 nM pipette concentration) into the oocyte decreased the oocyte current but had no effect on the membrane capacitance. B, the concentration-response curve for the current change induced by NAADP (means Ϯ S.D.). The Ca 2ϩ current was measured 2 min after perfusion of NAADP into the oocyte and was evaluated at the peak intensity of the depolarization-induced current as described in the legend of Fig. 1.
Contrary to that observed with cADPR, NAADP never elicited any capacitance changes as shown in Figs. 4A and 5A. This is further emphasized in Fig. 5A. Of the 19 oocytes perfused with NAADP, none of them showed any increase in membrane capacitance. In contrast, all 31 oocytes responded to cADPR with a capacitance increase averaging to about 40 pF after 5 min. As a control, ADP-ribose (ADPR), the hydrolysis product of cADPR, was also tested, and it too did not induce any capacitance changes (Fig. 5A).
Of the 19 oocytes perfused with NAADP and not showing a capacitance change, all responded with an average of about 60% decrease in Ca 2ϩ current after 2 min of perfusion, a magnitude similar to that induced by cADPR (Fig. 5B). The current decrease induced by either NAADP or cADPR could be blocked by preloading the oocytes with BAPTA, indicating in both cases that the effect was due to Ca 2ϩ release activated by the agonists. There was, however, a major difference between these agonists with respect to the action of the antagonist, 8-amino-cADPR, which inhibited only the current decrease induced by cADPR but not that induced by NAADP (Fig. 5B). This latter result indicates that the Ca 2ϩ release mechanism activated by NAADP was distinct from that activated by cADPR, consistent with that reported in sea urchin eggs (19). This difference was also supported by an experiment in which application of a high concentration of ryanodine to oocytes (100 M) blocked the cADPR-sensitive decrease in Ca 2ϩ current (24) but, similar to 8-amino-cADPR, did not inhibit the NAADP-induced changes (data not shown).
It has previously been reported that ADPR can interfere with channels and Ca 2ϩ events in ascidian oocytes (35). This is confirmed in Fig. 5B. The effect of ADPR on the current, however, was not mediated by Ca 2ϩ release since preloading the oocytes with BAPTA, which effectively eliminated the actions of both cADPR and NAADP, did not inhibit the effect of ADPR. The effect of ADPR appeared to be specific to that molecule since other similar compounds such as 1 mM ADP (not shown) or 8-amino-cADPR (Fig. 3A) did not cause a decrease in the current. An unexpected feature of the ADPR effect was its sensitivity to inhibition by 8-amino-cADPR (Fig. 5B). It is possible that the action of ADPR is related to its ability to covalently react with amino groups of proteins (36). Irrespective of the exact mechanism, it is clear that ADPR did not mobilize Ca 2ϩ since its action was not blocked by BAPTA (Fig. 5B) and thus was not investigated further.
The NAADP Signaling Is Independent of RyR, but Is Related to InsP 3 R-The effects of NAADP and cADPR on the current are independent (Fig. 6A). As usual, perfusion of NAADP into the oocyte decreased its Ca 2ϩ current. Subsequent perfusion of cADPR into the same oocyte induced a further decrease in its Ca 2ϩ current and an increase in membrane capacitance. As described above, both cADPR-induced changes can be attributed to the involvement of RyR. The magnitude of the subsequent current change was typical of that induced by cADPR alone, indicating that cADPR and NAADP act independently and that their effects on the current were essentially additive.
The inactivating effect of NAADP on the oocyte Ca 2ϩ current is, in some respects, similar to that observed with InsP 3 (24). Fig. 6B shows that perfusion with InsP 3 induced a slow decrease in the current very much like that seen with NAADP (Fig. 4A). Also, neither NAADP nor InsP 3 altered the membrane capacitance. However, the receptor for NAADP is distinct from that of InsP 3 since neither heparin, an InsP 3 receptor antagonist, nor pretreatment of the oocytes with InsP 3 inhibited the action of NAADP (Fig. 6B, right panel). Nevertheless, the InsP 3 and the NAADP-sensitive Ca 2ϩ stores appear to be able to interact functionally. Thus, as shown in Fig.  6C, pretreatment of the oocytes with NAADP could render InsP 3 incapable of causing a decrease in the Ca 2ϩ current. Indeed, pretreatment with NAADP (n ϭ 4, Fig. 6C, right panel) was found to be as effective as heparin in blocking the subsequent action of InsP 3 . It thus appears that the order of addition is important. Prior activation by NAADP inhibits InsP 3 but the converse is not true; no effect was seen on the action of NAADP by pretreatment with InsP 3 . These intriguing and apparently paradoxical observations were investigated further.
Inhibitory Effect of NAADP on Calcium Oscillation-It was found that the action of InsP 3 was complex (24). As shown in Fig. 7A, it was capable of inducing not only a single event of Ca 2ϩ release but prolonged Ca 2ϩ oscillation lasting as long as InsP 3 was applied through the patch pipette. Simultaneous perfusion of both NAADP and InsP 3 did not block the first Ca 2ϩ release induced by InsP 3 but effectively and reproducibly eliminated the subsequent Ca 2ϩ oscillation (n ϭ 11), as illustrated in Fig. 7B. All the InsP 3 -induced Ca 2ϩ oscillations, including the first one, could be inhibited if the oocytes were first preincubated for minutes with NAADP before InsP 3 application (Fig.  7C). This inhibitory effect is reversible, since subsequent simultaneous application of InsP 3 and NAADP 30 min later induces a pattern of Ca 2ϩ signal identical to that depicted in Fig. 7B (not shown). It thus appears that, when InsP 3 and NAADP were simultaneously applied, there was insufficient time for NAADP to effect total inhibition of the InsP 3 -induced oscillation (Fig. 7B). The results obtained with NAADP preincubation (Fig. 7C) are consistent with those shown in Fig. 6C.
The InsP 3 -induced Ca 2ϩ oscillation was very similar to that which occurs after fertilization (24). As shown in Fig. 8A, fertilization was accompanied by an initial Ca 2ϩ transient lasting more than 5 min, which was then followed by several smaller transients with a periodicity of about 2-3 min and each lasting for about 1 min. Pretreatment of the oocyte with NAADP for several minutes reproducibly eliminated these oscillations (Fig. 8B). The first Ca 2ϩ transient was not inhibited by NAADP but did appear to be significantly shortened by the NAADP treatment (n ϭ 3). This first Ca 2ϩ transient has previously been shown to be contributed by both Ca 2ϩ release through InsP 3 R and RyR as well as by Ca 2ϩ influx (24). Although NAADP did not inhibit the first Ca 2ϩ transient after fertilization, it was effective in blocking the post-fertilization Ca 2ϩ oscillation, which is likely to be mediated mainly by InsP 3 . The inhibitory effect of NAADP on the Ca 2ϩ oscillation was specific since neither pretreatment with cADPR nor 8-amino-cADPR was capable of blocking the post-fertilization Ca 2ϩ oscillation (data not shown). On the other hand, the precursor of NAADP, ␤-NADP, had the same effects, in all respects, as NAADP itself (not shown). This is not surprising since ␤-NADP could have been enzymatically converted to NAADP. Also, commercial ␤-NADP preparations are known to be contaminated with significant amounts of NAADP. Indeed, it was the contamination that had led to the discovery of the Ca 2ϩ releasing effect of NAADP (19).
Further evidence that the inhibitory effect of NAADP was restricted only to the Ca 2ϩ oscillation is shown in Fig. 9A. Pretreatment of oocytes with NAADP did not abolish the fer-FIG. 6. The interrelationship of the effects induced by NAADP, cADPR, and InsP 3 . A, the effects of NAADP and cADPR are independent. An oocyte was first perfused with 10 nM NAADP. After the decrease in the oocyte current stabilized, 10 nM cADPR was perfused into the same oocyte through a different patch pipette. B, an oocyte was first perfused with InsP 3 . After the decrease in the oocyte current stabilized, NAADP was perfused into the same oocyte through a different patch pipette. Further decrease in oocyte current was recorded. The panel on the right shows the mean of the current decrease induced by NAADP following the pretreatment of either standard pipette solution (Standard) (n ϭ 19, mean Ϯ S.D.), InsP 3 (n ϭ 2), or 3 mg/ml heparin (n ϭ 2). C, conditions are similar to panel B except NAADP was applied first, followed by InsP 3 . The panel on the right shows the mean of the current decrease induced by 100 nM InsP 3 following the pretreatment of either buffer (Standard) (n ϭ 13), NAADP (n ϭ 4), or 3 mg/ml heparin (n ϭ 6). Values are mean Ϯ S.D. The results of this figure are obtained with concentrations of cADPR or NAADP ranging from nanomolar to micromolar. The peak Ca 2ϩ current decrease was evaluated 2 min after InsP 3 application. tilization-associated membrane insertion, as indicated by a normal increase in membrane capacitance, nor did it totally inhibit the first Ca 2ϩ transient, which is consistent with that shown in Fig. 8B. In contrast, 8-amino-cADPR totally eliminated the capacitance increase as shown in Fig. 9B, whereas the Ca 2ϩ oscillation pattern following the first transient was not affected by this cADPR antagonist (not shown).
These results show that three independent Ca 2ϩ signaling mechanisms are present and operative in ascidian oocytes. Each mechanism is activated by a different Ca 2ϩ messenger, cADPR, NAADP and InsP 3 , and each appears to be involved in mediating specific changes occurring during fertilization and early development of the oocyte. DISCUSSION It is generally believed that cells possess multiple types of Ca 2ϩ stores (reviewed in Ref. 37). The recent discoveries of cADPR and NAADP (4 -6, 19) in addition to InsP 3 provide credence to such a belief. Indeed, it has been shown that the NAADP-sensitive stores can be physically separated from those sensitive to the other two Ca 2ϩ agonists (32,38). Of the three, the InsP 3 -dependent mechanism is ubiquitous in cells (39). The cADPR-dependent mechanism is also quite widespread, being present in a variety of cells from plant to mammalian tissues (32). In contrast, the distribution of the NAADP-mechanism in cells is just beginning to be explored. This study is of notable importance since it is the first to demonstrate that the mechanism is present in a cell other than sea urchin eggs. Although both sea urchins and ascidians are marine animals, they diverged in evolution hundreds of millions of years ago. The fact that NAADP-dependent Ca 2ϩ signaling is present in these two widely different species suggests that the mechanism may prove to be as widespread as those mediated by cADPR or InsP 3 . In this study, pharmacological evidence is obtained, indicating all three mechanisms may be mediated by its own independent receptor, consistent with that shown in sea urchin eggs (32). Thus, 8-amino-cADPR inhibited the effect of cADPR while heparin blocked the effect of InsP 3 , and neither antagonist had any effect on the action of NAADP (Figs. 5 and 6).
The intriguing question of why cells possess three different Ca 2ϩ signaling mechanisms was explored in this study. Our results show that each of the three mechanisms has its own special function. Evidence is presented indicating the cADPR mechanism is involved in mediating the insertion of membranes associated with fertilization. Thus, neither NAADP nor InsP 3 can induce an increase in the membrane capacitance (Figs. 4 and 6). Perfusion of cADPR into oocytes mimicked the capacitance increase seen at fertilization while preapplication of 8-amino-cADPR blocked the increase naturally occurring ( Figs. 1 and 9). It is likely that cADPR is also responsible for mediating the decrease in membrane Ca 2ϩ current during fertilization. Although all three Ca 2ϩ agonists can induce a decrease in membrane Ca 2ϩ current in the oocyte, only cADPR can do so with fast enough kinetics (Figs. 1, 4, and 6) as compared with those of fertilization (24). The fast changes induced by cADPR or fertilization are consistent with the cortical localization of the Ca 2ϩ stores (25).
That this effect of cADPR on the current is due to Ca 2ϩ release activated by the agonist is shown by the fact that preloading the oocytes with BAPTA totally inhibited the change (Fig. 5B). However, measurements of the cytosolic Ca 2ϩ using Fura-2 dextran detected no Ca 2ϩ changes induced by either cADPR (24) or NAADP (Fig. 7C). This could be due to the relatively low temporal and spatial resolution of our classical fura-2 measurement. Future experiments using other probes, in conjunction with confocal microscopy may allow direct measurement of the local Ca 2ϩ in the cortical region. In any case, the experiments using BAPTA presently serve to validate the current measurement as a method for monitoring localized Ca 2ϩ release close to the plasma membrane. This technique has been widely used and is generally accepted (24,25,35). The use of the Ca 2ϩ chelator to distinguish between Ca 2ϩ -dependent and -independent effects is important since, in principle, factors other than Ca 2ϩ may be able to induce a similar change in Ca 2ϩ current. This is the case for ADPR, whose inactivating effect on the Ca 2ϩ current is not inhibitable by BAPTA, indicating it is not mediated by Ca 2ϩ mobilization (Fig. 5B). The ADPR-effect is likely to be due to a direct interaction of the metabolite with ion channels. Indeed, it has been reported that ADPR can directly activate a K ϩ -channel in arterial smooth muscles (40).
In contrast to cADPR and NAADP, perfusion with InsP 3 induces a cytoplasmic Ca 2ϩ oscillation in the oocytes similar to that observed after fertilization. Preloading the oocytes with heparin blocks the oscillation (41). Together, these results indicate that post-fertilization Ca 2ϩ oscillation is mediated by InsP 3 . The exact mechanism of how these Ca 2ϩ oscillations are generated is not known, but our results point to the critical involvement of the NAADP-sensitive Ca 2ϩ stores. It has previously been shown in sea urchin eggs that NAADP itself is an inactivator of the NAADP-dependent Ca 2ϩ release mechanism and can totally desensitize the release mechanism even at non-activating concentrations (38,42). In the ascidian oocytes, our results show that pretreatment with NAADP, which presumably would discharge the stores and inactivate the release mechanism, can effectively inhibit the Ca 2ϩ oscillation. It is possible that the oscillation requires the functional interaction between the InsP 3 -and the NAADP-sensitive stores. Inactivating the NAADP-mechanism by the pretreatment could disrupt the critical interaction and thus block the oscillation, even if the InsP 3 -sensitive stores are fully functional. Indeed, it has previously been proposed that the interaction between the NAADP-and the cADPR/InsP 3 -sensitive Ca 2ϩ stores may be responsible for mediating the Ca 2ϩ oscillation seen in sea urchin eggs (32,42). Irrespective of the exact mechanism, the inhibition by pretreatment with NAADP does suggest the involvement of NAADP, together with InsP 3 , in mediating the post-fertilization Ca 2ϩ oscillation. Moreover, the known pH regulation of the NAADP production could allow the natural alkalinization of the cytoplasmic pH that occurs after fertilization to influence the pattern of post-fertilization Ca 2ϩ oscillation (43).
The results presented in this study are consistent with a model where the three independent Ca 2ϩ stores are arranged in a hierarchical manner. It is proposed that the cADPR-sensitive stores are localized next to the plasma membrane. Their mobilization by cADPR would rapidly raise the cortical Ca 2ϩ concentration, resulting in activation of a fast decrease in oocyte current and membrane insertion. Since the action of cADPR can be blocked by ryanodine (24), it is suggested that a RyR-like release channel is involved. Such a channel has indeed been immunolocalized to the cortical region of the oocytes (25). The next level of the hierarchical organization is suggested to be the NAADP-sensitive stores. The farther distance of these stores from the plasma membrane accounts for the NAADP-induced current decrease being slower than that of cADPR. The attenuation of the Ca 2ϩ release from these stores by local buffering could also contribute to the slow change and could account for the ineffectiveness of NAADP to activate the capacitance increase. The cortical region where the cADPRand NAADP-sensitive stores are localized may represent a diffusion barrier for either large molecules, such as Fura-2 dextran, or for Ca 2ϩ ions. The latter could be due to rapid resequestration of Ca 2ϩ and/or buffering by Ca 2ϩ binding proteins. Thus the cortical Ca 2ϩ changes can be detected readily by electrical measurements but not by the cytoplasmic Fura-2 dextran. In contrast, the Ca 2ϩ release induced by InsP 3 can be detected by the cytoplasmic probe, indicating that the stores are distributed in the cytoplasm. It is proposed that the InsP 3sensitive stores interact specifically with the NAADP-sensitive stores to generate the observed Ca 2ϩ oscillation. The exact nature of the interaction remains to be determined. One possibility is that the Ca 2ϩ released from one type of store is sequestered by the other, resulting in overloading and activating spontaneous release. A similar proposal has been advanced to account for the effect of NAADP on Ca 2ϩ oscillations seen in sea urchin eggs (32,42). Self-desensitization of the NAADP mechanism by pretreatment with the agonist could disrupt the crucial interaction between the Ca 2ϩ stores and thus terminate the oscillation. This hierarchical model was designed to account for results presented in this study and will serve as our working model for future investigations of the Ca 2ϩ signaling pathways in the oocyte.