Caged nicotinic acid adenine dinucleotide phosphate. Synthesis and use.

Nicotinic acid adenine dinucleotide phosphate (NAADP) is a metabolite of NADP with Ca2+ mobilizing activity. The Ca2+ release mechanism activated by NAADP as well as the Ca2+ stores that it acts on are different from those activated by either cyclic ADP-ribose or inositol 1,4,5-trisphosphate (IP3) (Lee, H. C., and Aarhus, R. (1995) J. Biol. Chem. 270, 2152-2157). In order to demonstrate unambiguously that NAADP can mobilize Ca2+ stores in live cells, a caged analog was synthesized by reacting NAADP with 1-(2-nitrophenyl)diazoethane. Anion exchange high pressure liquid chromatography (HPLC) was used to purify one particular caged form from the mixture of products. Phosphate analyses following specific enzymatic cleavage indicate that the caging group is on the 2′-phosphate. This is confirmed by 31P NMR spectroscopy, showing that the 2′-phosphate of the caged compound exhibits an altered chemical shift of −2.6 ppm as compared with 2.3 ppm determined for the 2′-phosphate of NAADP. Caged NAADP had no Ca2+ releasing activity at a concentration as high as 1 μM when tested on sea urchin egg microsomes. After photolysis, it released Ca2+, was effective in nanomolar range, and was indistinguishable from authentic NAADP. The regeneration of NAADP after photolysis was also confirmed by HPLC analyses. The analog is particularly susceptible to UV and can be efficiently photolyzed using a spectrofluorimeter. To demonstrate its utility in live cells, caged NAADP was microinjected into sea urchin eggs. Photolysis effectively regenerated NAADP and activated Ca2+ oscillations in the eggs. Removal of external Ca2+ did not prevent the Ca2+ oscillations but only delayed the second Ca2+ peak by about 45 s, indicating that the oscillations are due to release from internal stores and not caused by Ca2+ influx. A mechanism based on sensitization of the Ca2+ release by Ca2+ loading is proposed to account for the Ca2+ oscillation observed.

In addition to inositol trisphosphate, two other independent mechanisms for mobilizing internal Ca 2ϩ stores have been identified in sea urchin eggs. Cyclic ADP-ribose (cADPR) 1 and nicotinic acid adenine dinucleotide phosphate (NAADP) are Ca 2ϩ -mobilizing metabolites derived, respectively, from NAD (1-3) and NADP (4). Two modes of action of cADPR have been documented (reviewed in Ref. 5). In one case it can function as a modulator of the Ca 2ϩ -induced Ca 2ϩ release mechanism and, synergistically with calmodulin, increase the sensitivity of the release mechanism to divalent cations by several orders of magnitude (6 -9). Alternatively, it can also function as a Ca 2ϩ messenger. Nitric oxide, through elevating intracellular cGMP levels, can activate the synthesizing enzyme of cADPR, resulting in an increase in cellular cADPR and mobilization of Ca 2ϩ stores (10,11). Since the Ca 2ϩ releasing activity of cADPR was first described in sea urchin eggs, a variety of mammalian, amphibian, and plant cells have been shown to be responsive to cADPR, indicating its general relevance (1) (reviewed in Ref. 12).
NAADP is not a cyclic molecule; instead, it is formed by replacing the nicotinamide group of NADP with nicotinic acid (4). The Ca 2ϩ release mechanism activated by NAADP has many characteristics of an independent signaling pathway. In sea urchin eggs, NAADP is by far the most effective Ca 2ϩ release agonist and is active at nanomolar concentrations (1,4). Heparin, an antagonist of the IP 3 -receptor, has no effect on the NAADP mechanism (1,4,13), and cell fractionation studies show that the NAADP-sensitive Ca 2ϩ stores can be separated from those responsive to cADPR and IP 3 (4). The NAADP-dependent Ca 2ϩ release is not inhibited by high concentrations of Mg 2ϩ (14), 8-amino-cADPR, an antagonist of the cADPR-receptor (4,15), and does not require calmodulin (4). These properties distinguish it from the cADPR-dependent pathway. NAADP is likely to be operating through a distinct receptor. Specific binding of 32 P-NAADP to sea urchin egg microsomes has been demonstrated, and cADPR has no effect on the binding (16). One novel property of the NAADP mechanism is that, at subthreshold concentrations, NAADP can completely inactivate the release system such that subsequent challenge with a maximal concentration of NAADP is ineffective (16,17). Ligand binding studies show that the self-inactivation occurs at the level of the receptor (16). This novel property is not seen in either cADPR-or IP 3 -dependent Ca 2ϩ release and is likely to be the first description of such a process in receptor-mediated function. Although the structures and functions of cADPR and NAADP are totally distinct, the two Ca 2ϩ agonists can, in fact, be synthesized by the same enzymes (18). Both ADP-ribosyl cyclase and CD38, a lymphocyte antigen which is also a bifunctional enzyme involved in the synthesis of cADPR (reviewed in Ref. 19), can catalyze the exchange of the nicotinamide in NADP with nicotinic acid (18). The base exchange reaction dominates at acidic pH, while at neutral and alkaline pH, both enzymes preferentially cyclize NADP to produce cyclic ADPribose phosphate (18). In this study, we describe the synthesis of caged NAADP and the usefulness of the analog in investigating the Ca 2ϩ release mechanism of NAADP in microsomes as well as in live cells.

EXPERIMENTAL PROCEDURES
Synthesis-NAADP was synthesized by incubating NADP (1 mM) at pH 5.0 with the Aplysia ADP-ribosyl cyclase (25 ng/ml) in the presence of 30 mM nicotinic acid for several hours at 20 -23°C and purified by HPLC using an AG MP-1 column as described previously (18).
The caging reagent 1-(2-nitrophenyl)diazoethane (NPE), originally described by Walker et al. (20) for the synthesis of caged ATP and related nucleotides and more recently used to prepare NPE-cADPR (21), was used to prepare NPE-NAADP. NAADP has three phosphate groups and one carboxylate, all of which are potentially reactive with the caging reagent, provided they are protonated during the caging reaction. The caging reaction was done at two different pH values; first at pH 4.5 and then at pH 1.3. A solution of NAADP (1.0 mg, 1.3 mol) in water (0.5 ml) was prepared at room temperature; the pH was 4.5. A solution of the caging diazoethane reagent (13 mol) in ether (0.5 ml) was added, and the resulting biphasic mixture stirred vigorously for 2 h in darkness, during which the ether layer changed from amber to light yellow. The ether layer was drawn off, and a fresh solution of the caging reagent was added. After stirring for another 3 h, the amber ether layer was drawn off. Thin layer chromatographic analysis indicated complete conversion of starting NAADP (R F 0.13, methanol:choroform:water: acetic acid, 13:10:3.5:0.2, silica gel plates) to a single less polar product (R F 0.87). The aqueous layer was extracted with ether (2 ϫ l.5 ml), then applied directly to a column of Sephadex LH-20 (2 ϫ 6 cm). The product was eluted by gravity with water and lyophilized, yielding 0.5 mg of a pale yellow powder. This caged compound was found to be not very susceptible to photolysis and was not characterized further.
Reducing the pH of the caging reaction to 1.3 resulted in a caged product much more photolabile. NAADP (24 mg, 32 mol) was dissolved in 2.5 ml of water and the solution was adjusted to pH 1.3 by addition of dilute aqueous HCl. The diazoethane reagent (0.2 mmol) in 3 ml ether was added and stirred vigorously for 8 h in darkness as described above. The reagent was replaced twice and stirred each time for 3 h. TLC analysis indicated no remaining NAADP and two major products with R F values of 0.87 and 0.46. The least polar product (R F 0.87) co-eluted on TLC with the higher pH product described above. After the aqueous layer was extracted with ether (2 ϫ 3 ml), the more polar product (R F 0.46) was separated on a column of Sephadex LH-20 (2 ϫ 8 cm). A pale yellow powder of 8.6 mg (32% yield) of the caged product was obtained, which was found to be much more photolabile than the higher pH product described above. This product was further characterized and used in this study. Photolysis with a hand-held UV lamp showed significant conversion to free NAADP, as judged by TLC analysis.
Ca 2ϩ Release in Homogenates and Intact Eggs-Homogenates of sea urchin egg (Strongylocentrotus purpuratus) were prepared as described previously (21). Frozen egg homogenates (25%) were thawed at 17°C for 20 min and diluted to 5% with a medium containing 250 mM N-methylglucamine, 250 mM potassium gluconate, 20 mM Hepes, 1 mM MgCl 2 , 2 units/ml creatine kinase, 8 mM phosphocreatine, 0.5 mM ATP, and 3 M fluo-3, pH 7.2, adjusted with acetic acid. The homogenates were diluted to 2.5% and finally 1.25% with the medium described and were incubated at 17°C for 1 h between dilutions. Ca 2ϩ release was measured spectrofluorimetrically in 1.25% homogenates with an excitation wavelength of 485 nm and emission wavelength of 535 nm. The measurements were done in a cuvette maintained at 17°C, and the homogenates were continuously stirred. The volume of homogenate used was 0.2 ml, and additions were usually made in 2-l volumes.
Lytechinus pictus eggs were used for the microinjection experiments. The procedures for microinjection by pressure were as described previously (21). Ca 2ϩ changes in the injected eggs were measured using fluo-3. Samples were dissolved in the injection buffer containing 0.5 M KCl, 0.1 mM EGTA, 10 mM Hepes, pH 6.7. The injected volume was about 1% of the egg. Eggs were attached to the bottom of a protamine sulfate-coated culture dish and incubated with artificial seawater (ASW) containing 460 mM NaCl, 27 mM MgCl 2 , 28 mM MgSO 4 , 10 mM CaCl 2 , 10 mM KCl, 2.5 mM NaHCO 3 , pH 8.0. In some experiments, ASW was changed to Ca 2ϩ -free seawater (0CaSW) after microinjection using gravity perfusion. 0CaSW had the same composition as ASW except without CaCl 2 , and the concentration of NaCl was increased to 470 mM. In most experiments, the 0CaSW was supplemented with 0.2-1 mM EGTA. All media were kept at 17°C.
Photolysis-Activation of caged cADPR in egg homogenates was achieved in a Hitachi spectrofluorimeter (S-2000) by alternating the excitation wavelength every 2 s between 350 nm for photolysis and 485 nm for monitoring fluo-3 fluorescence.
In some experiments, photolysis with UV light (10 15 quanta/s) for 1 min was done in a Rayonet photochemical reactor (Southern New England Ultraviolet Co.) at 0 -4°C.
The UV photolysis and fluorescence measurement of individual eggs were done using the InCa 2ϩ imaging system (Intracellular Imaging Inc., Cincinnati, OH). Excitation was provided by a 300-Watt xenon lamp equipped with filters for 340-and 485-nm light. During photolysis, the excitation light was alternated between the uncaging (340 nm) and the monitoring (485 nm) wavelengths, both of which were reflected by a BCECF Sp dichroic filter toward the objective. The fluo-3 fluorescence was selected by a long pass filter with a 500-nm cutoff and monitored by a CCD camera. Fluo-3 fluorescence was measured every 4 s. Photolysis was performed for 3.5 s between measurements. As the xenon lamp aged, the intensity of UV light decreased. In some experiments, to compensate for the diminished UV intensity, the 340-nm filter was removed during photolysis.
HPLC Analyses-HPLC separation was done with columns packed with AG MP-1 resin (Bio-Rad) and eluted with a nonlinear gradient of trifluoroacetic acid similar to that described previously (15). The final purification of the caged NAADP was achieved using a 0.5 ϫ 5-cm Mono Q column (Pharmacia Biotech Inc.). The product was eluted using a gradient of water (solvent A) and 1 M triethylamine bicarbonate (solvent B, pH 8.8): 0 -12 min, 0% B; linearly increased to 20% B from 12 to 16 min, linearly increased to 30% B from 16 to 36 min, linearly increased to 100% from 36 to 37 min and held at 100% B for 3 min before returning to 0% B.
Enzymatic Cleavage and Phosphate Measurements-Caged NAADP, NAADP, and photolyzed caged NAADP, all at 28 M, were incubated with 2 units/ml alkaline phosphatase (Sigma), or 3.5 units/ml nucleotide pyrophosphatase (Sigma), or both enzymes together, in the presence of 10 mM MgCl 2 and 80 mM triethylamine bicarbonate, pH 8.8, for 20 min at 37°C. The total volume of the reaction mixture was 10 l. The phosphate released by the enzymes was measured by adding 0.1 ml of the Malachite Green reagent (22) and 10 l of 34% sodium citrate. Absorbance at 660 nm was measured and compared with sodium phosphate standards. 31 P NMR Analyses-Spectra were collected at an observation frequency of 161.9 MHz using a Varian UNITY plus 400 FT-NMR spectrometer which was equipped with a computer switchable Nalorac 4N400 -5ϩ 5 mm 4-nucleus probe ( 1 H, 19 F, 13 C, 31 P). All samples were dissolved in a D 2 O solution buffered at pH 7.5 with 10 mM Hepes. All spectra were acquired at an ambient room temperature of 22-23°C. A phosphoric acid (H 3 PO 4 ) solution in methanol in a 2-mm coaxial tube was used as external 31 P NMR standard for zero chemical shift calibration for all spectra. Negative chemical shifts are upfield of (lower frequency than) H 3 PO 4 . The acquisition conditions used for all the spectra include spectral width of 48543 Hz; acquisition time of 0.338 s; number of acquisition data points of 32768; pulsed flip angles ranging from 20 to 50°, depending on the experiment (PW 90 was about 50 s); relaxation delay of 3-4 s, depending on the experiment; and number of transients ranging from 400 to 30,000, depending on the experiment. All 31 P NMR spectra were proton-decoupled, with the decoupler gated on only during the acquisition period of the free induction decay. Data processing included zero-filling the free induction decays to 65536 data points and using an exponential weighting apodization function with a line broadening of 2.5 Hz to improve the signal-to-noise ratio in the transformed spectrum.

RESULTS
Synthesis and Purification-We have previously synthesized caged cADPR that is particularly photolabile. Effective photolysis can be accomplished with standard spectrofluorimeters or epifluorescence setups, and no specialized equipment is required (21). NAADP represented a special challenge since it has three phosphate groups as well as a carboxyl group which are all reactive toward the NPE reagent (20). Since it is known that the NPE reagent reacts mainly with protonated groups, the caging reaction was performed at either pH 4.5 or 1.3 so that different ionizable groups could be sampled. The caged product obtained at the higher pH value was found to be difficult to photolyze. Even after prolonged exposure to a handheld UV lamp, no detectable NAADP was produced as judged by TLC analysis. In sea urchin egg homogenate, the caged material was inactive as a Ca 2ϩ mobilizer, but prolonged photolysis only slowly and weakly generated free NAADP (data not shown). The product was not characterized further, but the poor photolysis properties are characteristic of carboxylates caged as NPE esters (23).
We next reduced the pH of the reaction to 1.3, a condition at which all four groups would be reactive. We did not have prior knowledge of which phosphate group would be the most appropriate for caging and which specific groups should be protected from being caged. Therefore, the strategy was not to focus on a particular group, but instead, to separate the mixture of products by chromotography and analyze each fraction using sea urchin egg homogenates as a bioassay for Ca 2ϩ release (1). The product was expected not to release Ca 2ϩ on its own but to be easily photolyzed using a spectrofluorimeter and to regenerate the Ca 2ϩ release activity of NAADP.
TLC analyses of the caged products produced at pH 1.3 showed two main spots with R F values of 0.87 and 0.46, respectively. The more polar products (R F 0.46) were separated using a Sephadex column (see "Experimental Procedures"). Fig. 1 (top) shows a HPLC chromatograph of the caged products analyzed on an AG MP-1 column. It was found to contain several components. All three peaks were collected and tested for Ca 2ϩ release activity before and after photolysis. Only the smallest peak, indicated with an asterisk, satisfied our criteria. The product was purified one more time on an AG MP-1 column followed by twice on a Mono-Q column. Fig. 1 (bottom) shows the purified product eluted as a single peak on the second Mono-Q column. Starting from 5 mg of the product mixture, about 50 g of purified caged NAADP was obtained.
Structural Determination-To determine which of the phosphates is caged, we used nucleotide pyrophosphatase to cleave the pyrophosphate linkage of the molecule and alkaline phosphatase to release the phosphate groups as inorganic phosphate. Alkaline phosphatase should also cleave the 2Ј-phosphate group if it is not caged. The P i released was measured using the Malachite Green method (22). Fig. 2 shows that treatment with alkaline phosphatase released about 1 mol of P i per mol of NAADP (open bars), which corresponds to the 2Јphosphate of the molecule. Nucleotide pyrophosphatase cleaved the molecule but did not release any P i by itself. After nucleotide pyrophosphatase treatment, all the phosphate groups of the cleaved molecule should be susceptible to alkaline phosphatase. Indeed, treatment with both enyzmes (MIX) produced about 3 mol of P i /mol of NAADP, as expected. In contrast to NAADP, treatment of caged NAADP (black bars) with alkaline phosphatase produced no P i , indicating that the 2Ј-phosphate is caged. Treatment with the combined enzymes released the two phosphates forming the pyrophosphate linkage. The pattern of P i released from the photolyzed caged NAADP (gray bars) following the enzyme treatments is the same as that of NAADP. These results are consistent with the caged group being attached to the 2Ј-phosphate of NAADP.
The structural assignment is further confirmed by 31 P NMR analyses. Fig. 3 compares the 31 P NMR spectra of two standards, NAADP and nicotinic acid adenine dinucleotide (NAAD), with that of the purified caged NAADP. The chemical shift was FIG. 1. Purification of caged NAADP by HPLC. Sequential chromatographic steps on anion exchange columns were used to purify a particular form of caged NAADP from a mixture of products generated by the caging reaction. The first step used an AG MP-1 column (1st AG MP-1). The fraction most easily photolyzed is indicated by the asterisk. The number of moles of P i released per mol of nucleotide was determined. Alkaline phosphatase was used to cleave specifically the 2Ј-phosphate. Nucleotide pyrophosphatase cleaves the pyrophosphate bond and allows the phosphate groups to be released by alkaline phosphatase. calibrated with an external phosphoric acid sample that had its chemical shift set to zero. The spectrum of NAADP shows two peaks at 2.3 ppm and Ϫ12.2 ppm. The negative parts/million peak represents the diphosphate of NAADP since its integrated area is about twice (2.09) that of the other peak. The two phosphorus atoms of the diphosphate are chemically equivalent and both are expected to have the same chemical shift. This assignment is supported by the spectrum of NAAD. The compound lacks the 2Ј-phosphate but contains the same diphosphate, and its spectrum shows only one peak at Ϫ12.2 ppm, a chemical shift value very similar to the diphosphate peak of NAADP. The spectrum of the caged NAADP also shows two peaks. The diphosphate peak has a chemical shift of Ϫ12.4, very similar to the diphosphate peaks of both NAADP and NAAD. The graphical integration shown above the spectrum indicates the area of the diphosphate peak is 1.78 times that of the 2Ј-phosphate peak. The chemical shift of the 2Ј-phosphate of the caged compound is changed to Ϫ2.6 ppm from 2.3 ppm of the 2Ј-phosphate of NAADP. This dramatic shift of the 2Јphosphate is consistent with the caging group being attached there. A noticeable feature of the spectrum of caged NAADP is that the peaks are broader than that of NAADP. The reason for the broadening is not known but could be due to reduced mobility of the caged compound in solution. The hydrophobicity of the cage group could promote intermolecular association especially at the relatively high concentration (Ϸ4 mg/ml) needed for obtaining the 31 P NMR spectrum. In any case, results from specific enzymatic cleavage and 31 P NMR analyses are consistent with the caging group being on the 2Ј-phosphate. The structure of caged NAADP is shown in Fig. 4.
Photolysis- Fig. 5B shows that addition of caged NAADP from 90 to 900 nM to egg homogenates produced no Ca 2ϩ release. The small jumps of fluo-3 fluorescence at high concen-trations of caged NAADP were due to addition artifacts (e.g. small Ca 2ϩ contamination in the samples), since they were present even when the release mechanism was totally inactivated (tracing labeled 8Ј in Fig. 5A). A novel property of the NAADP-sensitive Ca 2ϩ release is that the release mechanism can be totally inactivated by pretreatment with a subthreshold concentration of NAADP as low as 1 nM (16,17). The multistep procedure described above for the purification of caged NAADP was designed to ensure that the contaminating NAADP is below the self-inactivating levels. Fig. 5 also shows the results of testing the inactivating effect of caged NAADP. Egg homogenates were pretreated with 1 nM NAADP and subsequently challenged with a maximal concentration (40 nM) (Fig. 5A). After 2 min of pretreatment, the response to 40 nM NAADP was substantially reduced (tracing labeled 2Ј) as compared to without pretreatment (tracing labeled 0Ј) and was totally eliminated after 8 min of pretreatment (tracing labeled 8Ј). Pretreatment of the homogenates with 90 nM caged NAADP for 2 min produced very little inactivation (Fig. 5B). At 540 nM caged NAADP, the extent of inactivation following the pretreatment was similar to that effected by 1 nM of NAADP (comparing Fig.  5A, 2Ј, with Fig. 5B, 540 nM). This extent of inactivation by the caged compound can be accounted for if the sample is contaminated with about 0.1-0.2% NAADP.
This appears to be the case since freshly purified samples of the caged compound at 200 nM exhibited essentially no inactivation as shown in Fig. 6. Even at 900 nM of caged NAADP, the inactivation due to pretreatment was only about 30 -40%. Results in Fig. 2 indicate NAADP is very sensitive to alkaline phosphatase while the caged compound is not. Treatment of the 900 nM sample with alkaline phosphatase (ϩAPase), indeed, essentially removed all the inactivation. Results described in Figs. 5 and 6 thus show that caged NAADP is biologically inactive. It does not release Ca 2ϩ nor does it induce inactivation. Fig. 7 shows that photolysis of caged NAADP regenerates NAADP. The samples, before and after photolysis, were analyzed by HPLC. The retention time of the photolyzed product was the same as that of NAADP, which was shifted as compared with caged NAADP. The photolyzed product was effec- tive in releasing Ca 2ϩ as shown in Fig. 8 and, in fact, its concentration dependence was indistinguishable from that of NAADP. Also shown in Fig. 8 is that caged NAADP, before photolysis, had no Ca 2ϩ releasing activity at concentrations as high as 1 M. Further evidence that the photolyzed product is NAADP is provided by its desensitization of the NAADP-dependent Ca 2ϩ release in egg homogenates. The results are shown in the inset of Fig. 8. At 28 nM, the photolyzed product induced rapid Ca 2ϩ release. The egg homogenate became totally desensitized such that subsequent addition of 80 nM NAADP, after the Ca 2ϩ was resequestered, did not produce any release. Homogenates desensitized to prior exposure to 80 nM NAADP also did not respond to the photolyzed product. This cross-desensitization indicates the photolyzed product is indeed, NAADP.
Ca 2ϩ Release Activity- Fig. 9 shows that NAADP can be regenerated from caged NAADP using a spectrofluorimeter. The excitation wavelength was alternated between 350 nm for photolysis and 485 nm for monitoring the fluorescence of the Ca 2ϩ indicator, fluo-3. Addition of caged NAADP to the egg homogenates with the alternating UV excitation turned on resulted in Ca 2ϩ release after a brief delay (Fig. 9A). The delay was more prominent at the lower concentrations. Comparison of the Ca 2ϩ release activity with that induced by NAADP itself shows that about 1% of the caged NAADP added was photolyzed. This low efficiency is due to the relative weak UV excitation light of the spectrofluorimeter. Fig. 9B compares the concentration-response of NAADP and caged NAADP with or without UV photolysis. Because NAADP is effective in releasing Ca 2ϩ at nanomolar concentrations, the low efficiency of photolysis by the spectrofluorimeter does not hamper its use in Photolysis was induced by UV exposure in a photochemical reactor for 1 min at 0 -4°C. The concentration-response of the photolyzed caged NAADP (filled squares) was indistingushable from that of NAADP (opened squares), and caged NAADP (filled triangles) was inactive. The inset shows the cross-desensitization between NAADP and the photolyzed product, indicating the latter was, in fact, NAADP. this setting. A common problem in measuring Ca 2ϩ release from a suspension of permeabilized cells or cell-free assays, such as that shown in Fig. 9, is distinguishing Ca 2ϩ release from Ca 2ϩ contamination in the samples. The caged analog should be useful since it is not biologically active until photolysis, which can be accomplished conveniently by simply alternating the excitation wavelength in a spectrofluorimeter.
The caged analog is also useful in single cell measurements. We have previously shown that photolyzing caged NAADP loaded into sea urchin eggs can induce Ca 2ϩ oscillations (16). In some cases, these Ca 2ϩ oscillations persist for more than 30 min. The possibility that Ca 2ϩ influx may be involved in generating these oscillations was investigated by removing external Ca 2ϩ . We focused on the first two Ca 2ϩ oscillations that occur within 6 -7 min after photolysis. Fig. 10 shows that removal of external Ca 2ϩ does not prevent the Ca 2ϩ change induced by photolysis nor does it inhibit the subsequent Ca 2ϩ oscillation that occurs spontaneously. The main effect of removal of external Ca 2ϩ is a delay of the occurrence of the second Ca 2ϩ peak. This delay separates the second peak farther from the first peak and makes it appear more prominent in the case of 0CaSW. Table I summarizes the results from 16 eggs in 0CaSW and 17 eggs in ASW. Both the magnitude of the two Ca 2ϩ peaks and the time of the first Ca 2ϩ peak are independent of external Ca 2ϩ . These results show that the internal stores are the main source of the Ca 2ϩ changes induced by photolyzing caged NAADP as well as the subsequent Ca 2ϩ oscillation. It should be noted that in another five eggs in 0CaSW, photolysis induced no change in internal Ca 2ϩ . It is likely that these eggs had suffered damage during microinjection. The leakage of EGTA into the eggs could have buffered the Ca 2ϩ changes. In ASW the second peak occurred at 115.7 Ϯ 6.6 s after the start of photolysis. In 0CaSW, the second peak occurred at 161.5 Ϯ 12.6 s, a 45-s delay. The exact mechanism of how removal external Ca 2ϩ can delay the Ca 2ϩ oscillation remains to be elucidated. One possibility is proposed in the "Discussion." DISCUSSION As described in the introduction, the Ca 2ϩ release mechanism activated by NAADP has many properties of a signaling pathway. In this study, we describe two other properties of NAADP that strengthen its signaling role. First, it is highly effective in live cells, which can be unambiguously demonstrated using the caged analog. Indeed, loading of eggs with about 200 nM caged NAADP was more than sufficient (Table I), a concentration which is 10-fold lower than that required for caged cADPR (21). Removal of external Ca 2ϩ did not inhibit the Ca 2ϩ change induced by photolyzing caged NAADP in live eggs, indicating the source of Ca 2ϩ is from internal stores.
Second, NAADP can be degraded effectively by phosphatases, such as nucleotide pyrophosphatase and alkaline phospha- tase. An effective enzymatic system for removal is a hallmark of a signaling molecule, whose action needs to be terminated once its function is completed. The degradation pathway is particularly important for NAADP because of its potent selfinactivation property. The general presence of phosphatases in cells should ensure that NAADP is degraded.
Photolysis of caged NAADP in live eggs produces not only a single Ca 2ϩ transient but also spontaneous Ca 2ϩ oscillations that last for more than 30 min (16). The oscillation is not abolished by removal of external Ca 2ϩ , indicating that the Ca 2ϩ also comes from internal stores (Fig. 10). The exact mechanism is not known. One possibility is that the Ca 2ϩ released from the NAADP stores is sequestered by cADPR-sensitive stores, overloading the latter and triggering spontaneous release. It has previously been shown that Ca 2ϩ overloading not only can trigger spontaneous release in egg homogenates, but also can sensitize the stores to cADPR by 50-fold or more (24). The sensitization could conceivably enable the basal concentration of cADPR endogenously present in eggs to activate further release, generating the second Ca 2ϩ peak seen in Fig. 10. If external Ca 2ϩ is present, influx could hasten the overloading and/or sensitization of the cADPR stores and speed up the occurrence of the second Ca 2ϩ peak as shown in Fig. 10. The overloading mechanism could also account for the ineffectiveness of either cADPR or IP 3 in triggering Ca 2ϩ oscillation, since both agonists release from the same stores (16,25), and the majority of the Ca 2ϩ released by either of them would likely be sequestered by the NAADP-sensitive stores, which may not have the same type of Ca 2ϩ release mechanism that is sensitive to intravesicular Ca 2ϩ . Whether this overloading mechanism is in fact operative in eggs remains to be established, but it does provide rationalization of how the interplay of various Ca 2ϩ stores can result in generation of Ca 2ϩ oscillation.
The synthesis of caged NAADP described in this study produces only modest yield. This is not a major problem for biological use because of the incredible potency of NAADP in inducing Ca 2ϩ release. The main reason for adopting the nonselective strategy we have used is that there was no a priori knowledge of which of the four acidic groups should be caged. The results described in this study establish that the caging group attached to the 2Ј-phosphate is the most appropriate for photolysis. Improvement on the yield can now be rationally designed by protecting the other groups before the caging reaction. I Calcium oscillations induced by photolyzing caged NAADP Eggs were loaded with caged NAADP and 30 -190 M fluo-3 and incubated either in ASW or 0CaSW. Photolysis was achieved by exposure to UV for 17-25 s. Maximal fluorescence increase (F max ) was normalized to the initial fluo-3 fluorescence before photolysis (F 0 ). T 1 and T 2 , respectively, denote the time of the first and second Ca 2ϩ peak from the start of photolysis. All values are mean ϮS.E. The number of eggs used for measurements was 17 for ASW and 16 for 0CaSW.