2'-Phospho-cyclic ADP-ribose, a calcium-mobilizing agent derived from NADP.

Cyclic adenosine diphosphoribose (cADPR), a metabolite of NAD, appears to modulate changes in intracellular free Ca2+ levels by activation of ryanodine-sensitive Ca2+ channels. We report here that an ADPR cyclase purified from Aplysia californica readily catalyzes the conversion of NADP to 2'-phospho-cyclic adenosine diphosphoribose (2'-P-cADPR), cyclized at N-1 of the adenine moiety. An enzyme from canine spleen previously shown to contain NAD glycohydrolase, ADPR cyclase, and cADPR hydrolase activities also utilized NADP and 2'-P-cADPR as substrates. The apparent Km value for NADP was 1.6 microM compared with 9.9 microM for NAD, and the Vmax with NADP was twice that with NAD, indicating that 2'-P-cADPR is a likely metabolite in mammalian cells. 2'-P-cADPR was as active as cADPR in eliciting Ca2+ release from rat brain microsomes, but was unable to elicit Ca2+ release following conversion to 2'-P-ADPR by the action of canine spleen NAD glycohydrolase. 2'-P-cADPR and 1-D-myo-inositol 1,4,5-trisphosphate (IP3) appear to act by distinct mechanisms as microsomes desensitized to IP3 still released Ca2+ in response to 2'-P-cADPR and vice versa. Also, inhibition of IP3-induced Ca2+ release by heparin had no effect on release by 2'-P-cADPR. Both 2'-P-cADPR and cADPR appear to act by a similar mechanism based on similar kinetics of Ca2+ release, similar dose-response curves, cross-desensitization, and partial inhibition of release by procaine. The results of this study suggest that 2'-P-cADPR may function as a new component of Ca2+ signaling and a possible link between NADP metabolism and Ca2+ homeostasis.

Rapid changes in the cytosolic levels of free Ca 2ϩ , termed Ca 2ϩ signaling, are involved in the regulation of diverse cellular events including fertilization, muscle contraction, secretion, and proliferation (1,2). Often, Ca 2ϩ signaling is initiated by the release of internal Ca 2ϩ stores into the cytosolic compartment when endogenous second messengers activate membrane Ca 2ϩ channels. Two distinct families of Ca 2ϩ channels have been recognized to play key roles in intracellular Ca 2ϩ mobilization in many cell types. One family consists of channels sensitive to IP 3 , 1 a second messenger generated from cellular phospholip-ids in response to numerous hormones and neurotransmitters (1,3,4). A second family is composed of the ryanodine-sensitive Ca 2ϩ channels (5,6). One possible second messenger of ryanodine-sensitive channels is Ca 2ϩ itself, in a process termed Ca 2ϩ -induced Ca 2ϩ release (6). Recently, cADPR, a naturally occurring metabolite of NAD (7,8), has been implicated as a second messenger of ryanodine-sensitive channels. In sea urchin egg microsomes, cADPR is a potent mediator of Ca 2ϩ release, and its action is cross-potentiated by agents known to affect ryanodine-sensitive Ca 2ϩ channels (9 -13). In mammals, cADPR has been shown to elicit Ca 2ϩ release from microsomes isolated from several different tissues (14 -18) and permeabilized cells (19,20).
While cADPR appears to be a link between NAD metabolism and Ca 2ϩ homeostasis, the metabolic signals that regulate cADPR synthesis and/or degradation are unknown, and our knowledge of cADPR metabolic enzymes is limited. An ADPR cyclase that catalyzes the stoichiometric conversion of NAD to cADPR has been isolated from the marine mollusk Aplysia californica (21,22). In mammals, cADPR appears to be metabolized by multifunctional enzymes that contain ADPR cyclase, cADPR hydrolase, and NAD glycohydrolase activities (23)(24)(25)(26). While it is not yet clear if all mammalian NAD glycohydrolases are involved in cADPR metabolism, many of these enzymes effectively use both NAD and NADP as substrates (27)(28)(29)(30)(31)(32)(33), raising the possibility that they may catalyze the formation of a cyclic nucleotide from NADP. We report here that the Aplysia ADPR cyclase, an enzyme with strong sequence homology to several mammalian NAD glycohydrolases, efficiently catalyzes the conversion of NADP to 2Ј-P-cADPR. In addition, a multifunctional canine spleen NAD glycohydrolase (23) containing ADPR cyclase and cADPR hydrolase activities displays a kinetic preference for NADP over NAD and utilizes 2Ј-P-cADPR as a substrate, indicating that 2Ј-P-cADPR is a likely metabolite in mammalian cells. We also report that 2Ј-P-cADPR is as active as cADPR in eliciting Ca 2ϩ release from rat brain microsomes. Together with recent reports that another possible metabolite of NADP causes Ca 2ϩ mobilization (34,35), the results described here raise the possibility that NADP metabolism may be linked to Ca 2ϩ signaling.

MATERIALS AND METHODS
Activity of Aplysia ADPR Cyclase and Canine Spleen NAD Glycohydrolase on NAD and NADP-ADPR cyclase was purified from A. californica as described by Hellmich and Strumwasser (21), and canine spleen NAD glycohydrolase was purified as described by Kim et al. (23). Both enzymes were judged to be homogeneous when analyzed by SDSpolyacrylamide gel electrophoresis. Both enzymes efficiently release nicotinamide from NAD, but the primary product of the Aplysia enzyme is cADPR, while the canine spleen and other mammalian enzymes (24 -26) release cADPR very poorly, and thus, the primary product is ADPR. To compare the relative enzyme activity with NAD and NADP, the enzymes were assayed for release of nicotinamide. Reaction mixtures containing 20 mM HEPES, pH 7.2, 0.1% Triton X-100, and varying concentrations of [ 14 C]NAD or [ 14 C]NADP containing radiolabel in the nicotinamide moiety were incubated with the enzyme in a total volume of 500 l for 30 min at 37°C. Incubation was terminated by addition of 500 l of 0.1% SDS, and samples were applied to 0.5-ml columns of Bio-Rad AG 1-X2 anion-exchange resin. The columns were subsequently washed with 4 ml of deionized water. The amount of [ 14 C]nicotinamide in the flow-through and wash fractions was determined by liquid scintillation counting.
Activity of Canine Spleen NAD Glycohydrolase on 2Ј-P-cADPR-To assay for conversion of 2Ј-P-cADPR to free 2Ј-P-ADPR, reaction mixtures containing varying concentrations of [ 3 H]2Ј-P-cADPR in 20 mM HEPES, pH 7.2, 0.1% Triton X-100, and canine spleen NAD glycohydrolase in a total volume of 50 l were incubated at 37°C for 30 min. The incubation was terminated by addition of 50 l of 0.2 M NaOH followed by 50 l of 0.16 M HCl and 20 mM Tris free base. The [ 3 H]2Ј-P-ADPR produced was converted to adenosine by addition of 10 l of 100 mM MgCl 2 and 20 l containing 0.5 unit of snake venom phosphodiesterase (Worthington) and 0.5 unit of bacterial alkaline phosphatase (Sigma). After incubation at 37°C for 30 min, incubation mixtures were diluted to 1.0 ml with 0.1% SDS and applied to 0.5-ml columns of Bio-Rad AG 1-X2 anion-exchange resin. The columns were subsequently washed with 4 ml of deionized water. The amount of [ 3 H]adenosine in the flow-through and wash fractions was determined by liquid scintillation counting.
Enzymatic Synthesis and Purification of 2Ј-P-cADPR-An immobilized Aplysia ADPR cyclase was prepared by coupling the purified enzyme to an agarose gel. Tresyl chloride-activated agarose (0.5 g; Sigma) was hydrated overnight in 5 ml of 0.2 M sodium bicarbonate, pH 8.2. The purified enzyme (5 mg) was added, and the mixture was incubated with gentle agitation at 4°C for 72 h. This procedure results in an ϳ1.5-ml bed volume. For use, ϳ0.3 ml of gel containing the immobilized enzyme was transferred to a 0.8 ϫ 4-cm Poly-Prep ® chromatography column (Bio-Rad) and washed with 10 ml of 100 mM ammonium formate buffer, pH 9.0 (buffer A). The immobilized enzyme columns were stored in buffer A containing 0.1% (w/v) sodium azide at 4°C until use.
For preparation of 2Ј-P-cADPR, 1 ml of 10 mM NADP (Calbiochem) in buffer A was passed through a 0.3-ml Aplysia ADPR cyclase column at 4°C. The eluate was collected, and an aliquot was withdrawn for HPLC analysis. HPLC analysis was carried out on a 3.9 ϫ 300-mm Bondapak C 18 column (Waters) with isocratic elution with 100 mM potassium phosphate buffer, pH 6.0, at a flow rate of 1 ml/min. The 2Ј-P-cADPR was purified by anion-exchange chromatography followed by preparative HPLC. In brief, the eluate from the Aplysia ADPR cyclase column was applied to a Poly-Prep ® column containing 0.5 ml of Bio-Rad AG 1-X2 anion-exchange resin previously equilibrated with buffer A. The column was then washed with 20 ml of buffer A followed by 5 ml of deionized water. Elution was performed using 5 ml of 100 mM ammonium formate buffer, pH 4.0. The eluate was applied to a 10 ϫ 270-mm Dynamax ® preparative reversed-phase HPLC column (Rainin Instrument Co. Inc.) with isocratic elution with 0.05% trifluoroacetic acid at a flow rate of 2 ml/min. The 2Ј-P-cADPR peak was collected and lyophilized overnight. To eliminate residual trifluoroacetic acid, the preparation was redissolved in deionized water and subjected to two additional cycles of lyophilization. The final sample was reconstituted in deionized water and stored at Ϫ20°C. The concentration of 2Ј-P-cADPR in stock solutions was determined by absorbance at 254 nm using an extinction coefficient of 14,300 (36).
Characterizations of 2Ј-P-cADPR-For treatment with alkaline phosphatase, 600 nmol of 2Ј-P-cADPR in a total volume of 0.1 ml of 20 mM HEPES, pH 7.2, was incubated with 0.2 unit of bacterial alkaline phosphatase at 37°C for 1 h. Aliquots were removed and analyzed on a 4.6 ϫ 250-mm Partisil-10 strong anion-exchange HPLC column (Whatman) with 50 mM potassium phosphate buffer, pH 6.0, as the mobile phase and a flow rate of 1 ml/min. Treatment with canine spleen NAD glycohydrolase (23) was done by incubating 15 nmol of 2Ј-P-cADPR in 0.05 ml of 20 mM HEPES, pH 7.2, and 0.1% Triton X-100 with 0.5 g of purified enzyme at 37°C for 20 min. HPLC analysis was performed using the Partisil-10 strong anion-exchange HPLC system described above. Treatment with snake venom phosphodiesterase was done by incubating 25 nmol of 2Ј-P-cADPR in 0.1 ml of 20 mM HEPES, pH 7.2, 20 mM potassium phosphate, and 5 mM MgCl 2 with 0.2 unit of snake venom phosphodiesterase at 37°C for 30 min. Aliquots were taken from the reaction mixture and subjected to reversed-phase HPLC with iso-cratic elution with 100 mM potassium phosphate buffer, pH 6.0, at a flow rate of 1 ml/min. 1 H NMR analysis was done using a Varian VXR-400 NMR spectrometer. Three mol of 2Ј-P-cADPR was lyophilized three times in 99.9% D 2 0 prior to NMR analysis: 1  . Assignment of the anomeric ribose protons and the 2Ј-proton was based on spectra of cADPR 2 previously published (36 -38).
Ultraviolet absorption spectra were obtained on a Hitachi U2000 spectrophotometer. The spectral properties of 2Ј-P-cADPR were determined as a function of pH using a number of different buffers at a final concentration of 50 mM, which was also used as a reference. The buffers used at the indicated pH values were sodium acetate, pH 5.0; MES, pH 6.0; HEPES, pH 7.0 and 8.0; TAPS, pH 8.5 and 9.0; and CAPS, pH 10 and 11.
Preparation of Rat Brain Microsomes-Microsomes were prepared fresh on the day of experiment as described previously (14) with minor modifications. Briefly, Wistar rats (180 -200 g) were killed, and the brain tissue was removed and quickly placed in ice-cold buffer B consisting of 250 mM N-methylglucamine, 250 mM potassium gluconate, 20 mM HEPES, 1 mM MgCl 2 , 1 mM dithiothreitol, and 0.1 mg/ml soybean trypsin inhibitor, with pH adjusted to 7.2 with acetic acid. All procedures were performed at 0 -4°C. The tissue was homogenized in a Dounce-type homogenizer with a size A pestle (5-10 strokes) followed by a size B pestle (5-10 strokes). The homogenate was subjected to successive centrifugations at 1000 ϫ g for 5 min, 8000 ϫ g for 10 min, and 100,000 ϫ g for 40 min. The final pellet was resuspended in 5-10 ml of buffer C, which consisted of buffer B supplemented with 1 g/ml oligomycin, 1 mM sodium azide, 1 mM MgATP, 10 mM phosphocreatine, and 10 units/ml creatine phosphokinase. The microsomes were incubated at room temperature for 40 -60 min to allow calcium uptake from the suspension medium. Protein content of the microsomal preparation was estimated by the Bradford method (39).
Calcium Release Assays-Extramicrosomal Ca 2ϩ measurements were conducted in a Hitachi F-2000 spectrofluorometer using the fluorescent dye Fura-2 (Sigma) with excitation and emission wavelengths at 340 and 505 nm, respectively. The assay medium consisted of 0.5 mg/ml microsomal protein in 2 ml of buffer C. All measurements were made in the presence of 1 M Fura-2, and the mixture was continuously stirred at 30°C. Once a constant base-line level of fluorescence was established, 1-10 l of the test compound was added, and changes in Fura-2 fluorescence were monitored. At the end of each experiment, Ca 2ϩ signals were calibrated by adding 1 mM EGTA followed by 2.5 mM CaCl 2 to obtain F min and F max , respectively. The following equation was used to calculate the ambient free Ca 2ϩ concentration: (40). All solutions of 2Ј-P-cADPR, cADPR, and IP 3 were tested for the presence of Ca 2ϩ contamination by adding the compound to the reaction mixture in the absence of microsomes. In these experiments, no Ca 2ϩ signal was observed for the samples, indicating that they were virtually free of contaminating Ca 2ϩ .

Action of A. californica ADPR Cyclase on NADP-The A.
californica ADPR cyclase is distinct from other enzymes known to catalyze the conversion of NAD to cADPR in that cADPR is the primary product of the Aplysia enzyme (22), while the other enzymes generate primarily ADPR (23)(24)(25)(26). When tested with NADP as a substrate, the Aplysia enzyme readily catalyzed release of nicotinamide. The apparent K m values determined from Lineweaver-Burk plots were 190 M for NAD and 160 M for NADP, and the V max values were very similar for the two substrates.
To identify the product generated from NADP, an immobilized enzyme preparation was prepared by coupling the purified enzyme to agarose. This preparation has been previously used to convert micromole amounts of NAD to cADPR by a single passage at 4°C through a 0.3-ml column of the immobilized enzyme. 3 The result obtained when a 1-ml solution of 10 mM NADP was passed through a column of immobilized enzyme is shown in Fig. 1B. Passage through the column resulted in the disappearance of ϳ90% of NADP and the appearance of nicotinamide and unidentified material eluting at ϳ4 min. The material eluting at 4 min was purified by anion-exchange chromatography followed by preparative reversed-phase HPLC. The final preparation showed a single peak on reversed-phase HPLC (Fig. 1C). The NADP preparation used contained a small amount of contaminating material eluting at ϳ6 min (Fig. 1A). A recent report has shown that a commercial preparation of NADP (Sigma) contained contaminating NAADP, which is active in mobilizing Ca 2ϩ from sea urchin egg microsomes (35). Therefore, the possibility that the contaminant in our NADP preparation (Calbiochem) was NAADP was examined. When the NADP preparation was subjected to anion-exchange HPLC, the contaminating material eluted at 7 min compared with an elution time of 60 min for NAADP, demonstrating that the material did not correspond to NAADP.
Characterization of Material Produced from NADP-The material derived from passage of NADP through the Aplysia ADPR cyclase column was characterized by both enzymatic and spectral methods. When subjected to anion-exchange HPLC ( Fig. 2A), Ͼ95% of the UV-absorbing material eluted at 21 min, with a trace amount of material eluting at 8 min. The elution time of 21 min did not correspond to the elution times of cADPR (5 min), ADPR (7 min), and 2Ј-P-ADPR (43 min). Incubation with alkaline phosphatase resulted in a near quantitative conversion of the material eluting at 21 min to material that coeluted with cADPR, while the material that eluted at 8 min was unaffected by the phosphatase treatment (Fig. 2B). Incubation with purified canine spleen NAD glycohydrolase, an enzyme previously shown to catalyze the hydrolysis of cADPR to ADPR (23), resulted in a near quantitative conversion to material that coeluted with 2Ј-P-ADPR (data not shown). Incubation with snake venom phosphodiesterase, which does not act on cADPR (23), did not affect the chromatographic properties of the material, while in a positive control, ADPR was completely converted to material that comigrated with AMP (data not shown). In total, the enzymatic characterizations indicated that the material was a cyclic form of 2Ј-P-ADPR.
Further information was obtained from 1 H NMR spectroscopy (Fig. 3). The 1 H NMR spectrum was very similar to that of cADPR (36 -38). The spectrum also indicated that the material was primarily a single compound. A unique similarity to the spectrum of cADPR was the presence of two sets of chemical shifts between 6.0 and 6.2 ppm, which correspond to the anomeric protons of both ribose moieties. The chemical shifts at this frequency indicate that both ribose anomeric carbon atoms are bound to nitrogen, i.e. that the nucleotide is cyclic. A second unique similarity was an isolated signal with a chemical shift between 5.4 and 5.5 ppm. In cADPR, this signal is due to the ribose 2Ј-proton, which is shifted far downfield relative to the other ribose protons (36 -38). This signal appears as a well resolved triplet, while a broad singlet was seen for the cyclic form of 2Ј-P-ADPR. This difference can be attributed to the presence of a phosphate esterified to the 2Ј-carbon, with a resulting phosphorus-proton coupling.
The position of cyclization was studied by obtaining UV absorption spectra as a function of pH. Fig. 4 shows spectra FIG. 2. Treatment of material produced from NADP with alkaline phosphatase (phosphomonoesterase). Material derived from NADP by passage through an ADPR cyclase column was purified and incubated in the absence (A) or presence (B) of bacterial alkaline phosphatase. As a control, an equal amount of the alkaline phosphatase was incubated alone (C). The reaction mixtures were analyzed by anionexchange HPLC.

FIG. 1. Utilization of NADP by
Aplysia ADPR cyclase. A sample of NADP was passed through a column of immobilized A. californica ADPR cyclase, and aliquots of the reaction mixture were analyzed by reversed-phase HPLC as described under "Materials and Methods." The HPLC chromatograms of the NADP solution before and after passage through the ADPR cyclase column are shown in A and B, respectively. The material eluting at 4 min was subsequently purified as described under "Materials and Methods," and an aliquot of the purified material was analyzed by reversedphase HPLC as shown in C.
obtained at pH values of 5.0, 9.0, and 11.0. With increasing pH, the compound displayed a hyperchromic effect at 260 nm, a more pronounced shoulder at 267 nm, and a markedly increased absorbance in the region at 280 -310 nm. These pH-dependent spectral changes are unique to cADPR and other adenine nucleotides substituted at N-1 of the adenine ring (41), indicating that N-1 was the position of cyclization in 2Ј-P-cADPR. The spectral changes in the region at 280 -310 nm are due to dissociation of an adenine ring proton (41). Fig. 4 (inset) shows a plot of the absorbance at 300 nm as a function of pH. From these data, a pK a of ϳ9.0 was determined for 2Ј-P-cADPR. The corresponding pK a value for cADPR is ϳ8.2 (41). The higher pK a value of 9.0 can be attributed to the presence of the additional phosphate group present in the molecule. Taken together, the data from both enzymatic and spectral characterizations demonstrate that the compound generated from NADP by the action of immobilized A. californica ADPR cyclase is 2Ј-P-cADPR, cyclized at N-1 of the adenine moiety (Fig. 5).

Utilization of NADP and 2Ј-P-cADPR by Canine Spleen NAD
Glycohydrolase-Our laboratory has described previously the isolation and characterization from canine spleen of a multifunctional enzyme that contains NAD glycohydrolase, ADPR cyclase, and cADPR hydrolase activities (23). This enzyme uses either NAD or cADPR as substrate. Similar enzymes have been described in other mammalian cells (24 -26), indicating that multifunctional NAD glycohydrolases may be the rule for cADPR metabolism in mammalian cells. To address the question as to whether 2Ј-P-cADPR was a likely metabolite in mammalian cells, we determined if the canine spleen enzyme would utilize NADP and 2Ј-P-cADPR as substrates. Fig. 6A shows that the enzyme utilized NADP as a substrate. From three separate experiments, an apparent K m value for NADP of 1.6 Ϯ 0.2 M was determined. The corresponding K m value for NAD was determined with the same preparation of enzyme, and the value obtained was not significantly different from the value of 9.9 M that we had previously reported (23). The relative V max values for NADP and NAD were also determined, 2Ј-Phospho-cADP-ribose and the V max for NADP was ϳ200% of that observed for NAD. Also, competition experiments with NAD and NADP indicated that NAD was a competitive inhibitor of NADP and vice versa (data not shown). Accordingly, the specificity constant value k cat /K m for NADP is 12 times greater than the value for NAD, indicating that the NADP is a more preferred substrate than NAD. The canine spleen and other mammalian enzymes release cADPR very poorly; thus, the primary product released from the enzyme when NAD is used as a substrate is ADPR rather than cADPR (23)(24)(25)(26). Likewise, the primary product released from the canine spleen enzyme when NADP was used as a substrate was 2Ј-P-ADPR rather than 2Ј-P-cADPR (data not shown). Fig. 6B shows that 2Ј-P-cADPR was also a substrate for the enzyme. From three separate experiments, the apparent K m value for 2Ј-P-cADPR was 140 Ϯ 14 M. The corresponding K m value for cADPR was also determined and was similar to the value of 46 M reported previously (23). The V max for 2Ј-P-cADPR was ϳ10% of that observed for cADPR. In total, our results show that the canine spleen NAD glycohydrolase functions more effectively with NADP than with NAD as a substrate, but less effectively with 2Ј-P-cADPR than with cADPR. These data indicate that the conversion of NADP to 2Ј-P-cADPR is likely to occur in vivo.
Ca 2ϩ Mobilizing Activity of 2Ј-P-cADPR-The Ca 2ϩ mobilizing activity of 2Ј-P-cADPR was studied using rat brain microsomes. Fig. 7A shows that addition of 10 M 2Ј-P-cADPR resulted in Ca 2ϩ release from the microsomes. The pattern of Ca 2ϩ release appeared to be biphasic, with an initial rapid release followed by a sustained but slower rate of release. A very similar pattern of Ca 2ϩ release was observed by addition of 10 M cADPR (Fig. 7C). This similarity raised a concern that the Ca 2ϩ release observed following addition of 2Ј-P-cADPR was the result of conversion of 2Ј-P-cADPR to cADPR by the rat brain microsomes, with the observed Ca 2ϩ release being due to cADPR rather than to 2Ј-P-cADPR. To examine this possibility, samples were removed from incubation mixtures such as those shown in Fig. 7 and subjected to anion-exchange HPLC analyses. There was no detectable decrease in 2Ј-P-cADPR, even after a prolonged incubation, indicating that little or no conversion of 2Ј-P-cADPR to cADPR occurred under these conditions (data not shown).
The Ca 2ϩ release elicited by both 2Ј-P-cADPR and cADPR was dose-dependent (Fig. 8). EC 50 values of 1.3 M for 2Ј-P-cADPR and 1.5 M for cADPR were determined. The Ca 2ϩ release was specific for the cyclic nucleotides as addition of 50 M 2Ј-P-ADPR (Fig. 7B) and ADPR (Fig. 7D) did not result in 2Ј-Phospho-cADP-ribose detectable Ca 2ϩ release. Also shown is Ca 2ϩ release following addition of 10 M IP 3 (Fig. 7E). In contrast to the patterns observed with 2Ј-P-cADPR and cADPR, the IP 3 response showed a rapid phase of Ca 2ϩ release followed by rapid reuptake of Ca 2ϩ into the microsomes.
Current evidence indicates that cADPR and IP 3 elicit Ca 2ϩ release through distinct mechanisms (10). To determine if the mechanism of Ca 2ϩ release induced by 2Ј-P-cADPR was also distinct from the IP 3 system, cross-desensitization and inhibition studies were done. Fig. 9A shows that a second addition of IP 3 resulted in Ca 2ϩ release that was only 4% of the first addition, consistent with desensitization of the IP 3 -sensitive release mechanism. However, a subsequent addition of 2Ј-P-cADPR elicited approximately the same amount of Ca 2ϩ release observed without prior additions of IP 3 . Fig. 9B shows that a second addition of 2Ј-P-cADPR resulted in Ca 2ϩ release that was 29% of the first addition, indicating a partial desensitization of the 2Ј-P-cADPR release mechanism. In this case, a subsequent addition of IP 3 still resulted in Ca 2ϩ release that was approximately the same as observed without prior additions of 2Ј-P-cADPR (Fig. 9B). The relationship between 2Ј-P-cADPR and IP 3 was examined further by the use of heparin, a potent inhibitor of IP 3 -sensitive Ca 2ϩ channels (1). Pretreatment with 600 g/ml heparin completely abolished the action of IP 3 , but the Ca 2ϩ release activity of 2Ј-P-cADPR was unaffected (data not shown). In total, these data provide evidence that the mechanism of 2Ј-P-cADPR-induced Ca 2ϩ release is distinct from the IP 3 pathway.
Evidence that 2Ј-P-cADPR and cADPR were acting via a similar mechanism was obtained by examination of the effects of 2Ј-P-cADPR on cADPR-induced Ca 2ϩ release and vice versa. In the experiment shown in Fig. 10A, a second addition of 20 M 2Ј-P-cADPR resulted in Ca 2ϩ release that was 24% of the first addition. The subsequent addition of cADPR resulted in Ca 2ϩ release that was ϳ32% of a control without 2Ј-P-cADPR pretreatment, indicating that the 2Ј-P-cADPR release mechanism was cross-desensitized to that of cADPR, even though the release mechanism was only partially desensitized. A similar result was observed when successive additions of cADPR were followed by addition of 2Ј-P-cADPR (Fig. 10B).
Partial desensitization of Ca 2ϩ release was also observed following addition of subsaturating amounts of 2Ј-P-cADPR. For example, when the microsomes were treated with successive additions of 1 M 2Ј-P-cADPR, the response to the second, third, and fourth additions was reduced to 25-35% of the initial addition (data not shown). In the same experiment, when 1 M cADPR was added following the four additions of 2Ј-P-cADPR, Ca 2ϩ release was ϳ25% of a control without 2Ј-P-cADPR pretreatment, again indicating cross-desensitization between the 2Ј-P-cADPR and cADPR mechanisms.
Evidence that both 2Ј-P-cADPR and cADPR were operating through a ryanodine-sensitive Ca 2ϩ channel was obtained by examining the effect of procaine, a partial antagonist of ryan- FIG. 8. Ca 2؉ release from rat brain microsomes elicited by 2-P-cADPR and cADPR. Isolated rat brain microsomes were treated with different concentrations of either 2Ј-P-cADPR or cADPR, and Ca 2ϩ release was monitored fluorometrically using the Ca 2ϩ indicator Fura-2 as described under "Materials and Methods." Measurements were made at 100 s following addition of cADPR or 2Ј-P-cADPR. Approximately 30 nM Ca 2ϩ was released by saturating concentrations of 2Ј-P-cADPR and cADPR, and this value was normalized to 100%. All other measurements are expressed relative to the maximal value.
FIG. 9. Interaction of IP 3 -and 2-P-cADPR-induced Ca 2؉ release from rat brain microsomes. Rat brain microsomes were treated with two successive additions of IP 3 followed by 2Ј-P-cADPR (A) and conversely, two successive additions of 2Ј-P-cADPR followed by IP 3 (B). The final concentrations of each Ca 2ϩ -mobilizing agent added are indicated.

DISCUSSION
Studies with cADPR have provided a link between NAD metabolism and regulation of Ca 2ϩ signaling (10). The results described here, together with recent reports that another possible NADP metabolite, NAADP, causes Ca 2ϩ release in sea urchin egg microsomes (34,35), suggest a possible link between NADP metabolism and Ca 2ϩ signaling. As shown here, the Aplysia ADPR cyclase utilizes NAD and NADP with very similar efficiency to generate cyclic nucleotides (Fig. 1). In mammalian cells, the only enzymes that have been demonstrated to metabolize cADPR are NAD glycohydrolases that catalyze both the synthesis of cADPR from NAD and the hydrolysis of cADPR to ADPR (23)(24)(25)(26). Our results with the canine spleen enzyme using NADP as a substrate indicate that the conversion of NADP to 2Ј-P-cADPR in vivo is likely. The K m value of the enzyme for NADP is well below estimated cellular concentrations of NADP (42), and the enzyme displays a kinetic preference for NADP over NAD as reflected by a significantly higher ratio of k cat /K m for NADP compared with NAD. Evidence that the multifunctional canine spleen enzyme uses the same mechanism (23) with NADP as with NAD is supported by the observation that NAD and NADP are competitive substrates and that the enzyme also uses both cADPR and 2Ј-P-cADPR as substrates. While it remains to be determined if NADP is converted to 2Ј-P-cADPR in mammalian cells, the efficiency with which the canine spleen enzyme uses NADP warrants a search for 2Ј-P-cADPR in vivo.
The maintenance of Ca 2ϩ homeostasis requires the precise regulation of cytosolic free Ca 2ϩ levels. These levels are in turn controlled by membrane Ca 2ϩ channels through which Ca 2ϩ moves between different intracellular compartments and between intracellular and extracellular compartments. The elevation of cytosolic free Ca 2ϩ levels is often initiated by activation of members of one or both of two different families of intracellular Ca 2ϩ channels, the IP 3 -sensitive channels (1,3,4) and the ryanodine-sensitive channels (5,6). Studies of sea urchin egg microsomes indicate that cADPR modulates ryanodine-sensitive Ca 2ϩ channels (9 -13). Although less well characterized, cADPR also appears to act on ryanodine-sensitive Ca 2ϩ channels in mammalian cell microsomes (14,15). The results presented here indicate that 2Ј-P-cADPR, similar to cADPR, causes Ca 2ϩ release by a mechanism distinct from IP 3 .
Microsomes desensitized to further addition of IP 3 could still release Ca 2ϩ in response to 2Ј-P-cADPR and vice versa (Fig. 9). Also, levels of heparin that completely blocked release by IP 3 did not affect Ca 2ϩ release by 2Ј-P-cADPR.
Our results also suggest that 2Ј-P-cADPR elicits Ca 2ϩ release by a mechanism similar to that of cADPR based on the similar kinetics of Ca 2ϩ release elicited by the two nucleotides (Fig. 7), the similar dose-response curves (Fig. 8), cross-desensitization (Fig. 10), and the partial inhibition of Ca 2ϩ release by procaine (Fig. 11). The possibility that 2Ј-P-cADPR was being converted to cADPR by the brain microsomes and that the observed Ca 2ϩ release was due to cADPR can be ruled out by the observations that there was no detectable conversion of 2Ј-P-cADPR to cADPR under the conditions used and that very similar dose-response curves of cADPR and 2Ј-P-cADPR were observed. The concentrations of 2Ј-P-cADPR and cADPR that gave half-maximal Ca 2ϩ release in our study, 1.3-1.5 M (Fig.   FIG. 11. Effect of procaine on Ca 2؉ mobilizing activity of cADPR and 2-P-cADPR. The microsomes were pretreated with 5 mM procaine for 1 h. Either cADPR or 2Ј-P-cADPR was then added, and the Ca 2ϩ release was monitored fluorometrically. The results are expressed as percentage of the controls in which procaine was omitted from the incubation mixture. FIG. 10. Cross-desensitization of Ca 2؉ release by 2-P-cADPR and cADPR. The effects of two successive additions of saturating concentrations of 2Ј-P-cADPR on subsequent Ca 2ϩ release by cADPR and vice versa are shown in A and B, respectively. The final concentration (20 M) of the nucleotides represents the concentration that induced maximal Ca 2ϩ release from the rat brain microsomes (see Fig. 8).
2Ј-Phospho-cADP-ribose 8), are somewhat higher than those reported by other workers who observed maximal release at 0.25 M cADPR for rat brain microsomes (14). Whether this reflects differences in the amount of Ca 2ϩ loaded into the microsomes or some other difference will require further study.
It is interesting that the kinetics of Ca 2ϩ release of 2Ј-P-cADPR and cADPR differ from those of IP 3 in the brain microsomes. Addition of IP 3 to the microsomes resulted in a rapid release followed by a rapid re-uptake of Ca 2ϩ , while 2Ј-P-cADPR and cADPR caused an initial rapid release followed by a slower but prolonged increase in Ca 2ϩ (Fig. 7). Previous studies also have observed a prolonged Ca 2ϩ release by cADPR in rat pituitary cells (20) and rat brain microsomes (14). The difference in kinetics suggests that the mode of Ca 2ϩ channel activation caused by cADPR and 2Ј-P-cADPR may be fundamentally different from that of IP 3 . The rate of Ca 2ϩ release during the prolonged phase that occurs following 2Ј-P-cADPR and cADPR is probably underestimated as it presumably reflects the actual rate of Ca 2ϩ release minus the rate of Ca 2ϩ re-uptake.
Even if cADPR and 2Ј-P-cADPR act on the same Ca 2ϩ channels, the linkage of both NAD and NADP metabolism to Ca 2ϩ signaling raises interesting metabolic possibilities as the NAD(H) and NADP(H) pools are functionally distinct. The NAD(H) pool is maintained in a highly oxidized state (42) as NAD serves as a hydride acceptor in multiple catabolic reactions and as a source of ADP-ribose for cellular ADP-ribose transfer enzymes (43). In contrast, the NADP(H) pool is normally maintained in a highly reduced state to provide NADPH as a source of reducing equivalents for anabolic pathways (42). The highly reduced state of the NADP(H) pool under normal metabolic conditions makes it a potentially sensitive indicator of metabolic conditions that cause oxidative stress. In that vein, it is of interest that the studies of Richter and Kass (44) have closely linked agents that oxidize the NADP(H) pool to the activation of a mitochondrial NAD glycohydrolase and to rapid Ca 2ϩ efflux from mitochondria.