Stimulation of Cyclic ADP-ribose Synthesis by Acetylcholine and Its Role in Catecholamine Release in Bovine Adrenal Chromaffin Cells*

Cyclic ADP-ribose (cADPR) is suggested to be a novel messenger of ryanodine receptors in various cellular systems. However, the regulation of its synthesis in response to cell stimulation and its functional roles are still unclear. We examined the physiological relevance of cADPR to the messenger role in stimulation-secretion coupling in cultured bovine adrenal chromaffin cells. Sensitization of Ca2+-induced Ca2+release (CICR) and stimulation of catecholamine release by cADPR in permeabilized cells were demonstrated along with the contribution of CICR to intracellular Ca2+ dynamics and secretory response during stimulation of intact chromaffin cells. ADP-ribosyl cyclase was activated in the membrane preparation from chromaffin cells stimulated with acetylcholine (ACh), excess KCl depolarization, and 8-bromo-cyclic-AMP. ACh-induced activation of ADP-ribosyl cyclase was dependent on the influx of Ca2+ into cells and on the activation of cyclic AMP-dependent protein kinase. These and previous findings that ACh activates adenylate cyclase by Ca2+ influx in chromaffin cells suggested that ACh induces activation of ADP-ribosyl cyclase through Ca2+ influx and cyclic AMP-mediated pathways. These results provide evidence that the synthesis of cADPR is regulated by cell stimulation, and the cADPR/CICR pathway forms a significant signal transduction for secretion.

mobilizing intracellular Ca 2ϩ : Ca 2ϩ -induced Ca 2ϩ release (CICR), mediated by the ryanodine receptors. CICR mediates the amplification and propagation of initial Ca 2ϩ signals, the generation of Ca 2ϩ oscillations, and the propagation of Ca 2ϩ waves in certain types of cells. Depolarizing stimuli can activate release of Ca 2ϩ from ryanodine-sensitive intracellular stores in a number of neuronal cells (1). In cerebellar granule cells, a major component of both K ϩ depolarization-and Nmethyl-D-aspartate-induced elevation of [Ca 2ϩ ]i appears to be due to the release of Ca 2ϩ from intracellular stores, evidence that suggests the importance of intracellular Ca 2ϩ release via CICR (2). Galione et al. (3) recently identified a novel Ca 2ϩ mobilizing agent that is a cyclic metabolite of nicotinamide adenine dinucleotide (NAD ϩ ), cyclic ADP-ribose (cADPR). This agent is as potent as IP 3 in mobilizing Ca 2ϩ in sea urchin eggs and mediates the fertilization-induced Ca 2ϩ wave in these eggs (3,4). The presence of cADPR and the enzyme-catalyzing conversion of NAD ϩ into cADPR and the ability of cADPR to release Ca 2ϩ through an IP 3 -insensitive mechanism have been demonstrated in many tissues (5)(6)(7)(8)(9)(10)(11)(12)(13)(14)(15)(16)(17)(18). cADPR has been shown to activate the cardiac but not the skeletal isoforms of the ryanodine receptor Ca 2ϩ channel and could be a candidate for a nonskeletal type ryanodine receptor endogenous messenger (8).
If cADPR plays a role as a second messenger of CICR, it is necessary to demonstrate that its intracellular levels are under the control of extracellular stimuli. Recently, Takasawa et al. (7) suggested that cADPR is generated in pancreatic islets by glucose stimulation, serving as a second messenger for Ca 2ϩ mobilization and insulin release. However, little evidence is available in this regard for various cell types. In addition, recent studies using type 2 cardiac ryanodine receptors incorporated into planar lipid bilayers have shown the inability of cADPR to cause Ca 2ϩ release in the presence of physiological concentrations of ATP, indicating that cADPR is unlikely to be a second messenger for CICR in vivo (19,20). In addition, Higashida et al. (21) suggest another role of cADPR, that it may mediate the muscarinic acetylcholine (ACh) receptor-induced inhibition of M-type K ϩ currents in NG108 -15. Therefore, the physiological relevance of cADPR remains unclear.
Adrenal medullary chromaffin cells are widely used as a model for the analysis of endocrine and neuronal cell functions. Caffeine is well known for inducing a large increase in [Ca 2ϩ ]i levels through the mobilization of Ca 2ϩ from intracellular Ca 2ϩ stores by stimulating CICR in adrenal chromaffin cells. Spontaneous [Ca 2ϩ ]i fluctuations in rat chromaffin cells are generated by caffeine (22). Thus the presence of a ryanodine-sensitive intracellular Ca 2ϩ store is suggested. Here, we demonstrate the ability of cADPR to cause Ca 2ϩ release and the activation of cADPR synthesis in response to stimuli in bovine adrenal chromaffin cells, the results suggesting that the Ca 2ϩ mobilizing pathway mediated by cADPR may participate in physiological stimulation-induced secretory response of the cells.

Measurements of Ca 2ϩ Release and [Ca 2ϩ
]i-For measurement of Ca 2ϩ release from digitonin-permeabilized chromaffin cells, cells were washed and suspended in potassium glutamate buffer (145 mM potassium glutamate, 20 mM PIPES, 1 mM EGTA, pH 6.6) containing an ATP generating system (2 mM Mg 2ϩ -ATP, 5 mM creatine phosphate, 40 units/ml creatine phosphokinase) and protease inhibitors (2.5 mM benzamidine, 0.5 mM phenylmethylsulfonyl fluoride, 10 g/ml aprotinin, 10 g/ml leupeptin, and 50 g/ml trypsin inhibitor) and then given permeabilization by incubating cells with digitonin (20 M) for 5 min at 25°C. The cells were washed and resuspended (10 7 cells/ml) in an intracellular medium (KH medium: 140 mM KCl, 10 mM NaCl, 30 mM HEPES, pH 7.0) containing an ATP generating system, protease inhibitors, mitochondrial inhibitors (10 g/ml antimycin A, 10 g/ml oligomycin, and 10 mMNaN 3 ), and 0.025% bovine serum albumin. The permeabilization was checked by measuring the leakage of lactate dehydrogenase. One milliliter of cell suspension was transferred to a fluorescence cuvette and supplemented with fura-2 (1 M). Fluorescence was continuously monitored using a fluorometer at an excitation of 340/380 nm and an emission of 510 nm. Increase in Ca 2ϩ concentration in the medium was calibrated by the addition of known amounts of Ca 2ϩ and expressed as Ca 2ϩ release in the text.
For the measurement of [Ca 2ϩ ]i, cells were incubated at 32°C with 1 M fura-2 AM for 30 min in order to load the dye. Cells were then centrifuged at 15 ϫ g for 10 min and resuspended to yield 3 ϫ 10 6 cells/ml. Fluorescence was measured using a dual-wavelength, fluorescence spectrophotometric mode with an excitation of 340 and 380 nm and an emission of 510 nm. [Ca 2ϩ ]i was calculated from the fluorescent ratio at 340 and 380 nm using the equation of Grynkiewicz et al. (26) and a value of 224 nM for the K d of fura-2. For assaying 45 Ca 2ϩ release, a monolayer culture of chromaffin cells was preincubated in the full culture medium containing 45 CaCl 2 (185 kBq/ml) for 24 -48 h in the CO 2 incubator. Then cells were washed and permeabilized as described above. After a 1-min preincubation in a KH medium containing an ATP generating system, protease inhibitors, mitochondrial inhibitors, and 1 mM EGTA at 32°C, cells were incubated in various conditions. The medium was then immediately separated from the cells, and the 45 Ca 2ϩ released into the medium was quantified by liquid scintillation counting.
Measurement of CA Release-For measurement of CA release from permeabilized chromaffin cells, a monolayer culture of chromaffin cells was permeabilized as described above, preincubated for 1 min in a KH medium containing an ATP generating system, protease inhibitors, and mitochondrial inhibitors with or without 1 mM EGTA, and then incubated for 20 min with cADPR or IP 3 . After the period of incubation, the medium was separated from the cells and used for the CA assay. For this experiment, the KH medium was treated with Calcium Sponge TM S before use to reduce the background of CA released due to contaminated Ca 2ϩ in the medium. Various concentrations of Ca 2ϩ in the medium were made by a Ca 2ϩ -EGTA buffer, and the free Ca 2ϩ concentrations were measured using a selective Ca 2ϩ minielectrode or specific dye indicators, fura-2 and fluo 3. Perchloric acid (5% of final concentration) was added to the incubation medium, which was then centrifuged at 4,500 ϫ g for 15 min. 0.2 ml of the clear supernatant was diluted 20-fold with 3 M acetate buffer to create an appropriate concentration of CA and to adjust the pH to 6.2. Total CA in the medium was determined fluorometrically by the trihydroxyindole method (27), with adrenaline serving as a standard. CA release from intact chromaffin cells was performed as described previously (24,25).
Assay of ADP-ribosyl Cyclase-Cultured chromaffin cells were incubated in various conditions and then washed rapidly twice with 10 ml of chilled buffer containing 0.34 M glucose, 1 mM MgCl 2 , 10 mM 2-mercaptoethanol, 0.1 mM phenylmethylsulfonyl fluoride, 10 g/ml leupeptin, 10 g/ml aprotinin, 50 g/ml soybean trypsin inhibitor, 20 mM HEPES, pH 7.2, and immersed in liquid N 2 . Cells were homogenized in the buffer using Polytron for 20 s followed by six strokes with a Thomas homogenizer equipped with Teflon pestle. The homogenate was centrifuged at 100,000 ϫ g for 30 min at 0°C. The resulting pellet was suspended in the same buffer as a membrane fraction. The homogenate and membrane fraction were then used for assays of enzyme activity and protein concentration.
ADP-ribosyl cyclase activity was determined by measuring the production of cADPR from [ 32 P]NAD ϩ as substrate. An assay mixture containing 10 l of enzyme preparation, 40 l of 250 M [ 32 P]NAD ϩ (37 kBq), 20 mM HEPES, pH 7.2, and 0.1% Triton X-100 was incubated at 37°C for 10 min. The resulting [ 32 P]cADPR was separated from substrate and metabolites by selective hydrolysis with snake venom phosphodiesterase 1 followed by purification with dihydroxyboronyl Bio-Rex 70 (DHB Bio-Rex 70) column chromatography according to the reported procedure (28). In brief, the reaction mixture was treated with snake venom phosphodiesterase 1 (0.3 unit), diluted with ammonium formate buffer, pH 9.0, and applied to a DHB Bio-Rex 70 column. After washing the column with the above buffer, [ 32 P]cADPR was eluted with 5 ml of deionized water and quantified by liquid scintillation counting to estimate cyclase activity. Eighty ml of eluant from the DHB Bio-Rex 70 column was neutralized, freeze-dried, and then dissolved with 200 l of deionized water. Further analysis of the sample by HPLC was performed using a TSKgel QAE-2SW column (0.46 ϫ 25 cm) and TSKgel ODS-80Ts column (0.46 ϫ 15 cm) connected in tandem. Elution was performed with 0.25 M ammonium formate buffer, pH 4.0, at a flow rate of 0.5 ml/min and monitored by the absorbance at 260 nm. One-ml fractions were assayed for determination of Ca 2ϩ release activity as described above and for radioactivity.
The activity of ADP-ribosyl cyclase was also determined by measuring cyclic GDP-ribose (cGDPR) fluorometrically using NGD ϩ as a substrate (29). cGDPR is resistant to hydrolysis so this procedure was demonstrated to be highly sensitive and convenient for ADP-ribosyl cyclase-like enzymes such as CD38 (29). Reaction mixtures containing 60 M NGD ϩ , 20 mM Tris, pH 7.4, and homogenate and membranes of chromaffin cells (0.8 mg of protein/ml) were maintained at 37°C for 1 h to continuously monitor the fluorescence of cGDPR on a Hitachi 850 spectrophotometer at excitation/emission wavelengths of 300/410 nm. The fluorescence intensity was calibrated and converted to molar concentration by use of authentic cGDPR. NGD ϩ was HPLC-purified according to the procedure of Graeff et al. (29).
Assay of cADPR Hydrolase Activity-The hydrolase activity was determined using 200 M [ 3 H]cADPR (7.4 kBq) as a substrate as reported (28). ADPR, the product of cADPR was first converted to AMP by treatment with phosphodiesterase, and the AMP was separated on the DHB Bio-Rex 70 column. The flow-through and wash fractions (4 ml) were combined, and [ 3 H]AMP was quantitated by liquid scintillation counting. Statistical analyses were performed with the Student's t test.

Characterization of cADPR-induced Ca 2ϩ
Release from Digitonin-permeabilized Chromaffin Cells-cADPR stimulated Ca 2ϩ release in a concentration-dependent manner from permeabilized chromaffin cells with larger maximal response and lower affinity than that induced by IP 3 (Fig. 1A). In the presence of 1 M IP 3 , in which IP 3 caused the maximal release, additional 1 M IP 3 caused no further increase in Ca 2ϩ release, whereas cADPR brought about further increase in the release. As cADPR may induce Ca 2ϩ release by different mechanisms than that induced by IP 3 , we first examined the pharmacological characteristics of cADPR-induced Ca 2ϩ release in chromaffin cells.
Phosphorylation of the ryanodine receptor by calmodulin-dependent protein kinase has been reported for sarcoplasmic reticulum from cardiac muscle (30,31), where calmodulin inhibits the channel activity (31). Phosphorylation of ryanodine receptor and IP 3 receptor by cAMP-dependent protein kinase (A-PK) also has been reported, although the functional role of the phosphorylation in the regulation of Ca 2ϩ release is not clear (30 -35). In the present study, 8-bromo-cyclic-AMP (8-Br-cAMP) potentiated both cADPR-and IP 3 -induced Ca 2ϩ release; calmodulin potentiated cADPR-induced release but did not affect IP 3 -induced release (Fig. 1B). The catalytic subunit of A-PK (A-PK C.S.) potentiated cADPR-and IP 3 -induced Ca 2ϩ release, and the potentiating effects of an A-PK C.S. were blocked by the A-PK inhibitor ( Table I). 8-Bromo-cyclic-GMP (8-Br-cGMP) was without effect (data not shown). A calmodulin antagonist, W-7, reduced cADPR-induced release and blocked the potentiations by calmodulin of cADPR-and caffeine-induced Ca 2ϩ release. Neither calmodulin nor W-7 affected IP 3induced Ca 2ϩ release.
Heparin, CsCl, and tetraethylammonium, known inhibitors of IP 3 -induced Ca 2ϩ release, all inhibited IP 3 -induced Ca 2ϩ release in this experiment also but had no effect on cADPRinduced Ca 2ϩ release. A classical inhibitor of ryanodine receptors, benzocaine, and a novel inhibitor, imperatoxin inhibitor (IpTxi), which has been purified from scorpion venom and shown to specifically block ryanodine receptors of skeletal and cardiac muscle (36), both, inhibited cADPR-but not IP 3 -induced release. Caffeine potentiated the effect of cADPR, and cADPR greatly potentiated caffeine-induced release. Such cross-potentiation was not observed between IP 3 and cADPR or caffeine.
Thapsigargin, shown to specifically block the endoplasmic reticulum but not muscle sarcoplasmic reticulum Ca 2ϩ -ATPase, inhibits Ca 2ϩ uptake into endoplasmic reticulum and then empties Ca 2ϩ in the stores (37). In cells pretreated with 20 nM thapsigargin, IP 3 failed to increase Ca 2ϩ release. cADPRand caffeine-induced Ca 2ϩ release were only blocked by higher concentrations of thapsigargin (Fig. 2). The concentration-response relationship of thapsigargin-induced inhibition shows that Ca 2ϩ release in response to IP 3 is more sensitive to the inhibition than to cADPR. Taken together, these results may suggest that cADPR and caffeine act on the same Ca 2ϩ release mechanism from the same stores, which are different from IP 3 -sensitive stores, and support the hypothesis that cADPR may work through a Ca 2ϩ release system similar to the CICR system in the sarcoplasmic reticulum.
A number of studies on cADPR-induced Ca 2ϩ release from sea urchin eggs have shown that the actions of ryanodine and caffeine are similar to those of cADPR (4,38). Therefore, the ability of cADPR to sensitize CICR was examined by measuring Ca 2ϩ -induced 45 Ca 2ϩ release. Fig. 3 shows the effects of cADPR and caffeine on 45 Ca 2ϩ release from digitonin-permeabilized cells that had been prelabeled with 45 CaCl 2 . cADPR stimulated 45 Ca 2ϩ release, with the peak effect occurring 10 s after the addition of cADPR. The effect of cADPR was transient in comparison with the long lasting release of 45 Ca 2ϩ induced by caffeine. cADPR and caffeine markedly enhanced 45 Ca 2ϩ release induced by low concentrations of Ca 2ϩ but did not modify the maximal Ca 2ϩ -induced 45 Ca 2ϩ release.
cADPR-induced CA Release-Increases in the concentration of Ca 2ϩ in the medium from 0.02 M to a peak effect of 3 M stimulated release of CA in digitonin-permeabilized chromaffin cells (Fig. 4A). The Ca 2ϩ -induced CA release was not inhibited by IpTxi. In this cell preparation, cADPR (1 M) caused an increase in CA release, and the effect of cADPR was blocked by EGTA and IpTxi (Fig. 4B). These results show the ability of cADPR to induce CA release via Ca 2ϩ release through ryanodine receptor Ca 2ϩ channel.
Activation of ADP-ribosyl Cyclase and cADPR Hydrolase-To ascertain whether cADPR was a physiological messenger of CICR, the presence of synthesis and degradation systems for cADPR and their regulation by physiological cell stimulation was examined. Fig. 5 shows the presence of ADPribosyl cyclase activity in adrenal chromaffin cells. The product formed by the incubation of [ 32 P]NAD ϩ with a homogenate of adrenal medulla was separated from the substrate and other metabolites by a selective hydrolysis with snake venom phosphodiesterase and then purified with cation-exchange chromatography, as reported previously (28). The incubation product was used for an assay of Ca 2ϩ release from digitonin-permeabilized chromaffin cells. The product released Ca 2ϩ in the IpTxi-sensitive manner (Fig. 5A). Identification of cADPR related to Ca 2ϩ -releasing activity in the product was further carried out using HPLC (Fig. 5, B, a). Elution profiles revealed a single peak at the same retention time as authentic cADPR with regard to both radioactivity and Ca 2ϩ -releasing activity. This procedure made it possible to assay ADP-ribosyl cyclase easily and accurately. Most cyclase activity was found in the centrifuged particulate fractions but was not detectable in the 120,000 ϫ g supernatant (data not shown). In the homogenate prepared from the cells treated with ACh and excess KCl, ADP-ribosyl cyclase activity increased shortly after the additions (Fig. 5C). ACh-induced activation of the cyclase was further confirmed by HPLC. Both radioactivity (Fig. 5, B, b) and Ca 2ϩ -releasing activity (Fig. 5, B, c) were facilitated at the same retention time as authentic cADPR upon ACh-treated cells.
The presence of an enzyme that can degrade cADPR was noted during measurements of ADP-ribosyl cyclase activity in various tissue extracts. It has been reported that cADPR is degraded immediately by hydrolase and that the hydrolase activity, like ADP-ribosyl cyclase, is widely distributed (38), so we examined cADPR hydrolase activity in chromaffin cells. In the homogenate prepared from the cells treated with ACh (30 M), cADPR hydrolase activity increased shortly after the addition of ACh (Fig. 5D). The extracts from both stimulated and nonstimulated cells were examined for their ability to promote Ca 2ϩ release from digitonin-permeabilized cells. However no release of Ca 2ϩ was detected in this assay system (data not shown). Thus the amount of cADPR might be less than that required for detectable release of Ca 2ϩ .
Detecting ADP-ribosyl cyclase activity was further performed by fluorometric assay of accumulated cGDPR from NGD ϩ as substrate instead of NAD ϩ , which is resistant to hydrolysis (29). As shown in Fig. 6A, incubation of 60 M NGD ϩ with homogenate prepared from chromaffin cells resulted in a progressive increase in the fluorescence, which can be calibrated using cGDPR as a standard. Identification of cGDPR as being responsible for the changes in fluorescence of the NGD ϩ products was further performed using HPLC. The fluorescent  products had the same retention time as those obtained with authentic cGDPR (data not shown).
Mechanism of Receptor-mediated Activation of ADP-ribosyl Cyclase-The involvement of cAMP and calmodulin in the activation of ADP-ribosyl cyclase by ACh in intact cells was examined, since both agents enhanced cADPR-induced Ca 2ϩ release in permeabilized cells (Fig. 1, Table I). In addition to the possibility of interaction of these agents with ryanodine receptor as suggested (32), another possible mechanism was an increase in cellular cADPR level by these agents. The cGDPR fluorescence actually increased linearly during the incubation of the homogenate preparation with 8-Br-cAMP in the presence of an ATP generating system. The rate of cGDPR synthesis increased from 77.0 Ϯ 4.3 to 202.4 Ϯ 11.1 pmol/mg of protein/ min with the addition of 8-Br-cAMP, and the effect was blocked by a A-PK blocker, R p -cAMP-S (Fig. 6A). Calmodulin was without effect (data not shown). In contrast to the observation shown in Fig. 6A, the membrane preparations from the AChtreated cells revealed a rapid increase in fluorescence initially in the assay where ATP generating system was absent (Fig.  6B). Similar activation by forskolin and 8-Br-cAMP was observed (Fig. 6B). The activation of the enzyme by ACh was rapid, starting within 10 s and reaching a maximum at 1 min after the addition of ACh (Fig. 6C).
We have previously shown that ACh increased adenylate cyclase activity and accumulation of cAMP in chromaffin cells resulting from an influx of Ca 2ϩ (24,39). We therefore examined whether the activation of ADP-ribosyl cyclase by ACh is primarily mediated by Ca 2ϩ influx. The action of ACh but not forskolin was abolished by the treatment of cells with Ca 2ϩ -free medium and an inhibitor of the voltage-operated Ca 2ϩ channel (VOC), diltiazem (Fig. 6D). The treatments with A-PK blockers such as the structurally unrelated compounds H-89 and R p -cAMP-S abolished the effect of both ACh and forskolin and also 8-Br-cAMP. These results strongly suggested that ADP-ribosyl cyclase was activated by ACh via Ca 2ϩ influx and the resulting cAMP-mediated mechanism. It has been reported that cGMP can stimulate cADPR production in sea urchin eggs (40) and in PC12 cells (12). A similar mechanism responsible for cGMP does not seem to operate in chromaffin cells because sodium nitroprusside or 8-Br-cGMP did not affect the ADP-ribosyl cyclase activity in the same experimental protocol. 2 Role of CICR in Secretory Response-Whether CICR participates in Ca 2ϩ transience and CA release in chromaffin cells was examined using IpTxi. In digitonin-permeabilized chromaffin cells, IpTxi inhibited Ca 2ϩ release in response to cADPR, caffeine, and ryanodine but not to IP 3 (Table I). Pretreatment of intact chromaffin cells with IpTxi reduced AChinduced [Ca 2ϩ ]i increases (Fig. 7A).
As the increase in [Ca 2ϩ ]i after stimulation of nicotinic ACh receptors in adrenal chromaffin cells is believed due to Ca 2ϩ influx through its channel and VOC, the time required for the 2 K. Morita, S. Kitayama, and T. Dohi, unpublished observations.  (30 M) or from the sample without enzyme preparation was further analyzed by HPLC using a TSKgel QAE-2SW column and TSKgel ODS-80Ts column connected in tandem. Elution was performed with 0.25 M ammonium formate buffer, pH 4.0, and monitored by the absorbance at 260 nm as indicated by line with the positions of authentic compounds marked at the top (a). The elution from the HPLC column was collected in 1-ml fractions for quantitating the radioactivity (b) and for confirming the Ca 2ϩ -release activity (c). C, time courses of the activation of ADP-ribosyl cyclase in the cells treated with ACh (30 M) and excess KCl (30 mM). After incubation with or without the indicated stimulants, cells were washed and homogenized with glucose-HEPES buffer containing protease inhibitors, then processed to the subsequent steps as described in (A) and quantified by liquid scintillation counting. D, cADPR hydrolase activity was determined as described under "Experimental Procedures." Values are the mean Ϯ S.E. of three or four experiments assayed in duplicate. reduction of [Ca 2ϩ ]i to half the peak high of [Ca 2ϩ ]i (t1 ⁄2 ) was plotted as a function of [Ca 2ϩ ]i obtained with various treatments. A linear correlation was found between t1 ⁄2 and [Ca 2ϩ ]i in each condition (Fig. 7B). In the presence of diltiazem, the correlation between the peak high of the ACh-induced [Ca 2ϩ ]i increase and t1 ⁄2 was very similar to that in the absence of diltiazem. However, t1 ⁄2 at the various concentrations of ACh in the presence of IpTxi was shortened, and the slope became steep. Similarly, the correlation between the peak high of excess KCl-induced [Ca 2ϩ ]i increase and t1 ⁄2 was not modified by diltiazem but was shifted to the left by IpTxi. Thus, it is suggested that the attenuation of rise in [Ca 2ϩ ]i is hastened by IpTxi through blockade of CICR. In other words, CICR contributes to a maintenance of the sustained [Ca 2ϩ ]i rise during cell stimulation.
To further determine the involvement of CICR in CA release, we examined the effect of IpTxi on CA release that was induced by ACh in intact chromaffin cells. ACh at concentrations of 3-100 M induced CA release with maximal effect at 30 M.
IpTxi reduced CA release induced by each concentration of ACh (Fig. 8A). Caffeine and ryanodine markedly potentiated AChinduced CA release, and IpTxi blocked the potentiation by these compounds (Fig. 8B). DISCUSSION A growing body of evidence supports the idea that cADPR might be an endogenous caffeine-like substance that regulates ryanodine receptor of type 2 in cardiac muscle and type 3 in other types of cells but not type 1 in skeletal muscle (38). The ability of cADPR to sensitize these receptors to the stimulatory effect of Ca 2ϩ provides a possible mechanism for intracellular Ca 2ϩ mobilization and thus for the generation of both calcium oscillation and calcium waves. Through video-imaging analysis of fura-2-loaded adrenal chromaffin cells, it has been shown that the [Ca 2ϩ ]i increase brought about by Ca 2ϩ influx through plasma membrane by nicotinic agonists or high K ϩ was initially restricted to a subplasmalemmal region. This restricted increase was then followed by more widespread elevation of [Ca 2ϩ ]i throughout the cytoplasm (41). Thus, it was hypothesized in chromaffin cells that the second phase of the Ca 2ϩ increase was due to the release of Ca 2ϩ from internal stores by a CICR-dependent mechanism.
In this study, we demonstrated that 1) cADPR mediates the sensitization of CICR, 2) cADPR metabolism is under the control of cell activation, and 3) cADPR/CICR pathway forms a positive feedback loop in secretory response in adrenal chromaffin cells.
cADPR-induced Ca 2ϩ Release-Our findings of the crosspotentiation of the effect of cADPR with caffeine or ryanodine but not with IP 3 and of the specific antagonism of the effect of cADPR by a ryanodine receptor antagonist suggest that cADPR releases Ca 2ϩ via a different mechanism from that of IP 3 . Chromaffin cells, therefore, seem to possess two functionally distinct Ca 2ϩ stores sensitive to either IP 3 or cADPR and caffeine. The different sensitivity to thapsigargin-induced inhibition of IP 3 -and of caffeine-or cADPR-induced Ca 2ϩ release in chromaffin cells agrees well with the conclusion that the target pools for cADPR and IP 3 are distributed differently within cells.
Phosphorylation of ryanodine receptors and of IP 3 receptors by calmodulin and by A-PK have been shown in various tissues, but the regulation of Ca 2ϩ release through phosphorylation was unclear. In the present study, the potency of cADPR but not of IP 3 in releasing Ca 2ϩ was enhanced by calmodulin in chromaffin cells. Both cADPR-and IP 3 -induced Ca 2ϩ release were enhanced by A-PK. Lee et al. (42,43) and Tanaka and Tashjian (44) suggest that cADPR requires the accessory or intermediate proteins to activate ryanodine receptors and actually identify two cADPR-binding proteins, 140 kDa and 100 kDa, in egg microsomes (45). Although it is not evident whether cADPR directly interacts with ryanodine receptor proteins or indirectly through its additional target factor(s) in adrenal chromaffin cells, Ca 2ϩ release mediated by a cADPR-dependent process seems to receive positive modulations by A-PK and/or calmodulin in the cells.
Regulation of cADPR Synthesis-The membrane-bound protein that possesses ADP-ribosyl cyclase activities has been purified from lymphocyte, a 40-kDa protein called CD38, a lymphocyte surface antigen, and also from the spleen, a 39-kDa protein. Moreover, these proteins have cADPR hydrolase activities (28,46). The amino acid sequence of the Aplysia ADPribosyl cyclase has been determined (47) and found to be homologous to CD38. Takasawa et al. (48) and Kato et al. (49) stress the importance of CD38 as a regulator of the level of cADPR in pancreatic ␤ cells, where ATP generated during glucose metabolism in islets inhibits the cADPR hydrolase activity of CD38, thereby increasing the accumulation of cADPR. However, the physiological role of expressing an enzyme bifunctionally metabolizing cADPR on the surface of lymphocytes has not yet been elucidated. Although various tissues have been found to possess the activity of ADP-ribosyl cyclase and hydrolase, especially the brain (5,29,50), little is known about the control of ADP-ribosyl cyclase. To our knowledge, the present results are the first demonstration that ADP-ribosyl cyclase is activated in response to ACh, a physiological stimulation in adrenal chromaffin cells using NAD ϩ and NGD ϩ as substrate. ACh stimulated not only ADP-ribosyl cyclase activities but also hydrolysis of cADPR. Therefore, it is obvious that the increase in ADP-ribosyl cyclase activities induced by ACh is not due to an inhibition of cADPR hydrolysis. Recently, increased cADPR formation has been described in intestinal longitudinal muscle upon cholecystokinin administration, although the process leading to ADP-ribosyl cyclase activation was not investigated (16).
The activation of ADP-ribosyl cyclase by ACh appears to be Ca 2ϩ -dependent, because activation was blocked by omitting extracellular Ca 2ϩ and also by an inhibitor of VOC. Therefore, it seems likely that the stimulation of ACh receptors does not directly activate the cyclase but regulates enzyme activity via the increase of [Ca 2ϩ ]i. 8-Br-cAMP and forskolin mimic the ACh-induced activation of ADP-ribosyl cyclase. Moreover, blockers of A-PK blocked the activation of the cyclase activity induced by ACh, whereas forskolin-induced activation of the cyclase was not inhibited by the omission of external Ca 2ϩ and the VOC inhibitor. We have previously demonstrated that ACh increases cAMP levels in the chromaffin cells by the activation of adenylate cyclase resulting from an increased influx of Ca 2ϩ (39). Taken together, it was concluded that ACh elevates the Ca 2ϩ influx through the plasma membrane in chromaffin cells, resulting in an activation of adenylate cyclase activity, and consequently, that the increased cAMP initiates the activation of ADP-ribosyl cyclase activity via a A-PK-dependent mechanism.
Activation of the cyclase induced by 8-Br-cAMP was characterized as an increase in initial velocity of the enzyme activity in intact cells, whereas a time-dependent progressive increase was observed in the assay in which 8-Br-cAMP was directly added in the homogenate containing the ATP generating system. Therefore, these findings suggest that the continuous phosphorylation of ADP-ribosyl cyclase or its related protein(s) by A-PK may be required for the maintenance of the cyclase activity. The deduced amino acid sequences of CD38 from human, mouse, and rat and of Aplysia ADP-ribosyl cyclase have several consensus phosphorylation sites for A-PK (38,51).
Physiological Role of CICR in Stimulation-Secretion Coupling-cADPR-induced Ca 2ϩ release in the nervous system has been demonstrated in brain microsomes (8 -10). Its involvement in depolarization-induced increase in [Ca 2ϩ ]i was also shown in bullfrog sympathetic ganglion cells (14) and in cultured Purkinje neurons (11), although the physiological significance has not been elucidated.
Our present findings suggest that the ADP-ribosyl cyclase/ cADPR pathway contributes to the spread of Ca 2ϩ to intracellular regions, which in turn regulate CA release response in chromaffin cells. This possibility was further supported by the specific inhibition by IpTxi of cADPR-induced Ca 2ϩ release from permeabilized chromaffin cells and of the stimulationevoked [Ca 2ϩ ]i rise and CA release in intact cells. These effects of IpTxi seem unlikely to be due to the inhibition of Ca 2ϩ influx through the plasma membrane because IpTxi did not affect the stimulation-evoked 45 Ca 2ϩ influx (data not shown).
ACh causes biphasic [Ca 2ϩ ]i rise, an initial transient rise followed by sustained rise, in bovine adrenal chromaffin cells. Both phases are largely dependent on the presence of extracellular Ca 2ϩ (52). Thus it is thought that Ca 2ϩ influx through the plasma membrane is the main source for the [Ca 2ϩ ]i rise caused by ACh in the cells. An analysis of the relationship between the peak height of [Ca 2ϩ ]i and the time required for its half decay in the presence or absence of diltiazem and IpTxi revealed a clear difference in the mode of blockade of [Ca 2ϩ ]i transient by these agents. Namely, diltiazem equally reduced the peak and sustained phases of ACh-induced [Ca 2ϩ ]i rise, but IpTxi specifically reduced the sustained rise in [Ca 2ϩ ]i. These results may reflect the differences in the mechanism of action between an inhibitor of VOC, diltiazem, and an inhibitor of CICR, IpTxi, suggesting that transient rise of [Ca 2ϩ ]i is mainly due to a Ca 2ϩ influx from the extracellular space and that the sustained rise of [Ca 2ϩ ]i is constituted by a Ca 2ϩ influx and a concurrently occurring Ca 2ϩ release from intracellular stores by CICR. Conclusions-It has been suggested that Ca 2ϩ entry and the sustained rise of [Ca 2ϩ ]i are essential for the activation of exocytosis in bovine chromaffin cells (52). Therefore, the present results suggest that the CICR pathway, which is activated through Ca 2ϩ influx, contributes to the time-dependent rise of [Ca 2ϩ ]i and to the maximal exocytotic response in chromaffin cells.
The evidence that cADPR is synthesized and degraded quickly in response to agonists, that cADPR stimulates the release of Ca 2ϩ by a ryanodine-receptor-sensitive pathway, and that the physiological stimulation of [Ca 2ϩ ]i transient and secretory response were blocked by an inhibitor of CICR satisfies criteria for cADPR as a second messenger of CICR in adrenal chromaffin cells.