Kinetic properties of nicotinic acid adenine dinucleotide phosphate-induced Ca2+ release.

Three endogenous molecules have now been shown to release Ca2+ in the sea urchin egg: inositol trisphosphate (InsP3), cyclic adenosine 5′-diphosphate ribose (cADPR), and nicotinic acid adenine dinucleotide phosphate (NAADP), a derivative of NADP. While the mechanism through which the first two molecules are able to release Ca2+ is established and well characterized with InsP3 and cADPR-activating InsP3 and ryanodine receptors, respectively, the newly described NAADP has been shown to release Ca2+ via an entirely different mechanism. The most striking feature of this novel Ca2+ release mechanism is its inactivation, since subthreshold concentrations of NAADP are able to fully and irreversibly desensitize the channel. In the present study we have investigated the fast kinetics of activation and inactivation of NAADP-induced Ca2+ release. NAADP was found to release Ca2+ in a biphasic manner, and such release was preceded by a pronounced latent period, which was inversely dependent on concentration. Moreover, the kinetic features of NAADP-induced Ca2+ release were not altered by pretreatment with low concentrations of NAADP, although the extent of Ca2+ release was greatly affected. Our data suggest that the inactivation of NAADP-induced Ca2+ release is an all-or-none phenomenon, and while some receptors have been fully inactivated, those that remain sensitive to NAADP do so without any change in kinetic features.

Intracellular Ca 2ϩ stores in the sea urchin egg express both inositol trisphosphate (InsP 3 ) 1 -sensitive Ca 2ϩ channels (1) and ryanodine-sensitive Ca 2ϩ channels (RyRs). The latter are activated by a pyridine nucleotide metabolite, cyclic adenosine 5Ј-diphosphate ribose (cADPR) (2). Recently, a third distinct Ca 2ϩ release mechanism has been identified that is potently and selectively activated by a different pyridine nucleotide, nicotinic acid adenine dinucleotide phosphate (NAADP) (3,4). NAADP-induced Ca 2ϩ release can be induced both in the intact egg by microinjection and from microsomes of egg homogenates (4 -6). Ca 2ϩ release induced by NAADP is not blocked by hep-arin, a selective inhibitor of InsP 3 receptors, nor by 8-amino-cADPR, a selective inhibitor of the cADPR-sensitive Ca 2ϩ release (3,4). Moreover, NAADP does not cross-desensitize with either InsP 3 nor cADPR, showing that it induces Ca 2ϩ release via a mechanism independent of InsP 3 Rs and RyRs (3,4,6). This is also supported by the distinct pharmacological properties of the NAADP-sensitive mechanism. For example, NAADP-induced Ca 2ϩ release is blocked by L-type Ca 2ϩ channel blockers such as nifedipine and diltiazem, which have no effect on either of the other two Ca 2ϩ release mechanisms in the sea urchin egg homogenates (6). The NAADP-sensitive Ca 2ϩ store, unlike cADPR-and InsP 3 -sensitive stores, is insensitive to thapsigargin and cyclopiazoic acid, suggesting that the NAADP-sensitive Ca 2ϩ release mechanism resides on a distinct intracellular organelle (7).
A unique feature of the NAADP-induced Ca 2ϩ release is that it can be selectively inactivated by subthreshold concentrations of this molecule, which per se do not cause Ca 2ϩ release (6,8). The mechanism of this unusual form of desensitization is unclear. However, the binding of NAADP cannot be displaced when microsomes are pretreated with [ 32 P]NAADP and then challenged with a high concentration of NAADP after 3 min (8), perhaps indicating that treatment with nonstimulating concentrations of NAADP results in a conformational change of the receptor occluding further association and dissociation of NAADP binding.
Recent reports have indicated that mammalian cells possess the metabolic machinery to regulate intracellular levels of NAADP and therefore possibly utilize it as an intracellular messenger. NAADP can be synthesized in cell extracts from ␤-NADP in the presence of nicotinic acid in various rat tissues, including brain, liver, and spleen (9). Furthermore, it has been shown that ADP-ribosyl cyclase, the enzyme responsible for the cyclization of NAD ϩ to cADPR, and CD38, a lymphocyte differentiation antigen, can also synthesize NAADP (10). The regulation of this class of enzymes is still unclear, but it appears that NAADP is synthesized in the presence of ␤-NADP and nicotinic acid at more acidic pH, while cADPR is produced in the presence of ␤-NAD at more neutral pH values (10).
In the present study we have investigated the transient kinetics of the NAADP-induced Ca 2ϩ release mechanism in the sea urchin egg homogenates and compared it with the transient kinetics of the InsP 3 -sensitive and cADPR-sensitive Ca 2ϩ release. Such approach can lead to further understanding of the similarities and differences between these release mechanisms and to provide information on the activation and inactivation mechanisms of NAADP-induced Ca 2ϩ release. NAADPinduced Ca 2ϩ release is composed of a fast and a slow phase, which can be described by two rate constants. Similar biexponential equations can also fit InsP 3 -induced and cADPR-induced Ca 2ϩ release. NAADP triggers Ca 2ϩ release after a pro-nounced latency inversely dependent on the concentration of NAADP, while no latency is present before InsP 3 -and cADPRinduced Ca 2ϩ release. Desensitization by subthreshold concentrations of NAADP does not affect either of the rate constants, but affects the magnitude of release. This finding suggests that NAADP desensitization proceeds in an all-or-none manner, with inactivation affecting a subpopulation of receptors while others are unaffected. Ca 2ϩ Release Assays-Homogenates (2.5%) of unfertilized Lytechinus pictus eggs were prepared as described previously (11), and Ca 2ϩ loading was achieved by incubation at room temperature for 3 h in an intracellular medium (IM) consisting of potassium gluconate, 250 mM; N-methylglucamine, 250 mM; Hepes, 20 mM (pH 7.2); MgCl 2 , 1 mM; ATP, 0.5 mM; phosphocreatine, 10 mM; creatine phosphokinase, 10 units/ml; oligomycin, 1 mg/ml; antimycin, 1 mg/ml; sodium azide, 1 mM; fluo-3, 3 mM. Free Ca 2ϩ concentration was measured by monitoring fluorescence intensity at excitation and emission wavelengths of 490 and 535 nm, respectively. Fluorimetry was performed at 17°C using 500 l of homogenate in a Perkin-Elmer LS-50B fluorimeter. Additions were made in 5-l volume, and all chemicals were added in IM containing 10 M EGTA. Basal concentrations of Ca 2ϩ were typically between 100 and 150 nM. Sequestered Ca 2ϩ was determined by monitoring decrease in fluo-3 fluorescence during microsomal loading and by measuring Ca 2ϩ release in response to ionomycin (5 M) and was constant between experiments.

Collection of Sea Urchin Eggs-Eggs
Preparation of Purified Microsomes and Supernatant Fractions-Microsomes were purified from 50% egg homogenates by the Percoll density centrifugation method described previously (11). Briefly, a fractionating buffer was prepared by diluting Percoll stock to 25% in a modified intracellular medium (333 mM N-methylglucamine, 333 mM potassium acetate, 27 mM Hepes, 1.3 mM MgCl 2 , pH titrated to 7.2 with acetic acid). This was then supplemented with 0.5 mM ATP, 2 units/ml creatine phosphokinase, 4 mM phosphocreatine, 10 M EGTA. 1 ml of 50% egg homogenate was layered on 9 ml of this solution and centrifuged at 27,000 ϫ g for 30 min at 15°C. The top fraction (1 ml) was collected and represented the supernatant fraction. The Ca 2ϩ -storing, cADPR-and InsP 3 -sensitive microsomes formed a distinct tight band half way down the tube. 1 ml of this was collected using a disposable syringe to puncture the vessel wall. The fractions were aliquoted and stored at Ϫ70°C until use.
Stopped-flow Measurements-Rapid kinetics of Ca 2ϩ release by InsP 3 , cADPR, and NAADP were carried out on 2.5% homogenate prepared as described above. Stopped-flow measurements were performed as described elsewhere (12)(13)(14). Ca 2ϩ -loaded homogenate was introduced in a 2.5-ml syringe of a stopped-flow fluorimeter (Applied Photophysics, model SX17 MV), while a 250-l syringe was filled with either IM, InsP 3 , or NAADP (diluted in IM) at a concentration 10 times the concentration desired in the mixing chamber, as the mixing ratios of the two syringes were 10:1. Temperature of homogenate in either the syringes or mixing compartment was maintained at 17°C by a circulating water bath. Fluorescence changes of fluo-3 were monitored with an exciting the sample at 490 nm and measuring the emission above 515 nm. Fluo-3 fluorescence was captured over 55 s in a split-time base mode, with 200 recordings taken in the first 5 s and 200 recordings taken in the remaining 50 s. Each experiment represents the average of at least six acquisitions.
The averaged traces were then analyzed using nonlinear regression analysis programs supplied by Applied photophysics and Biosoft. The progress of Ca 2ϩ release from at least three different homogenate preparations used in this study were shown to be biphasic and could be best fitted to a biexponential profile using the following equation, fluo-3 signal in these experiments was never saturating, these units are related to Ca 2ϩ changes. In experiments where NAADP was used, Ca 2ϩ release was preceded by a latency or lag phase. This latency was quantified and then subtracted prior to further analysis of the Ca 2ϩ release process. In a control experiment the effect of the Ca 2ϩ pumps responsible for the re-uptake of Ca 2ϩ in these preparations were also quantified, and the rate constants of the uptake were essentially negligible compared with Ca 2ϩ release by either InsP 3 , cADPR, or NAADP. In fact, the rate constant for Ca 2ϩ was 0.0008 Ϯ 0.000004 s Ϫ1 , while the rate constants observed for Ca 2ϩ release were at least 30 times faster.
Materials-NAADP was either purchased from Research Biochemicals International (St. Albans, United Kingdom (UK)) or was a kind gift of Prof. T. F. Walseth (University of Minnesota), fluo-3 was from Molecular Probes (Cambridge, UK). All other chemicals were from Sigma (Poole, UK).

RESULTS AND DISCUSSION
NAADP released Ca 2ϩ dose-dependently with an EC 50 of 25 Ϯ 5 nM (Fig. 1, A and B). Furthermore, NAADP, 3 nM, which per se induced only a negligible release, was able to desensitize the mechanism to a subsequent addition of a maximal concentration of NAADP (200 nM) ( Fig. 1C; see also Refs. 6 and 8). Performing experiments at 7°C did not alter significantly the extent of Ca 2ϩ release nor the extent of inactivation (Fig. 1).
If quantal release, first described for the InsP 3 -induced Ca 2ϩ release, is defined by the observation that submaximal concentrations of mobilizing agent can rapidly release Ca 2ϩ from the Ca 2ϩ stores without affecting their ability to respond to further maximal concentrations of the same agent (15-17) (incremen-tal detection), NAADP does not release Ca 2ϩ in a quantal manner. It therefore appears either quantal Ca 2ϩ release, which has been also described for the ryanodine receptors (18), is not central in the operation of all intracellular Ca 2ϩ channels or that incremental detection is not necessarily associated with quantal release.
To investigate the fast kinetics of the Ca 2ϩ fluxes activated by NAADP, InsP 3 , or cADPR, we have employed stopped-flow analysis, with a temporal resolution of 25 ms. In the present experiments Ca 2ϩ pump inhibitors have not been used for two reasons: (i) the rate of Ca 2ϩ uptake in our preparations was at least 30 times slower than the slower Ca 2ϩ release parameter examined and therefore negligible (see also "Materials and Methods"); (ii) NAADP-induced Ca 2ϩ release is known to be affected differently from cADPR and InsP 3 by Ca 2ϩ uptake inhibitors (e.g. it is insensitive to thapsigargin and cyclopiazoic acid) (7). When NAADP, InsP 3 , or cADPR were added to the homogenate, it resulted in an increased fluorescence signal representing Ca 2ϩ release (Figs. 2-4). Supramaximal concentrations of NAADP (1 M) (6) released Ca 2ϩ with a t1 ⁄2 of 11.2 Ϯ 1.0 s. Supramaximal concentrations of cADPR (3 M) (6) released Ca 2ϩ with a t1 ⁄2 of 6.5 Ϯ 0.9 s. In comparison, addition of maximal InsP 3 concentrations (10 M) (6) resulted in a more rapid Ca 2ϩ release, with a t1 ⁄2 of 2.7 Ϯ 0.8 s. InsP 3 -induced Ca 2ϩ release appears slower to that observed in cerebellar microsomes, where maximal InsP 3 concentrations release Ca 2ϩ with a t1 ⁄2 of less than a second (14). This difference may well reflect the relative abundance of InsP 3 -sensitive channels in the two systems.
An important and consistent feature of the NAADP-induced Ca 2ϩ release was that low concentrations of NAADP did not release Ca 2ϩ instantly, but rather a pronounced latency period was observed prior to release (Figs. 2 and 5). This latent period had a duration of up to several seconds and was found to be inversely dependent on the concentration of NAADP used (Fig.  5). However, at higher NAADP concentrations (3 and 10 M), Ca 2ϩ release was not preceded by a detectable latency (Figs. 2 and 5). In contrast, no lag phase was observed with any concentration of InsP 3 (Fig. 3) This is in agreement with most reports on the fast kinetics of the InsP 3 receptor (14) and consistent with direct gating of InsP 3 Rs by InsP 3 , although a lag phase has been reported for InsP 3 -induced Ca 2ϩ release in rat basophilic leukemia cells (13). cADPR was also not preceded by a lag phase. To investigate whether this lag phase was due to the conversion of NAADP to an active metabolite, stopped-flow experiments were performed at 7°C. If an enzymatic conversion was responsible for the latency, then lowering the temperature should prolong the lag phase. The lag phases at both 7 and 17°C did not markedly differ at all concentrations tested (Fig. 4). Moreover, NAADP was able to release Ca 2ϩ to the same extent in both conditions (Fig. 1). It therefore seems unlikely that an enzymatic conversion is responsible for the observed latency and that Ca 2ϩ pumps affect release in this system.
The marked lag phase observed could be dependent on the requirement of a modulatory protein to bind the NAADP receptor. To investigate this possibility, fluorimetric determinations of NAADP-induced Ca 2ϩ release under various conditions were performed. In Percoll-purified microsomes, which should lack soluble cytosolic proteins, NAADP still released Ca 2ϩ , although to a minor extent in respect to the 2.5% homogenate. Unlike cADPR-sensitive Ca 2ϩ stores, NAADP-sensitive Ca 2ϩ stores are not all contained in this microsomal fraction (4), and the minor extent of release observed is most likely due to this observation. Addition of calmodulin (1 M) did not affect NAADP-induced Ca 2ϩ release in microsomes, in contrast to the potentiating effect that it has on cADPR-induced Ca 2ϩ release (19). Addition of FK506 (10 M) and rapamycin (10 M), which are known to affect Ca 2ϩ -releasing mechanisms by inhibiting modulatory proteins associated with RyRs and InsP 3 Rs (20), did not affect NAADP-induced Ca 2ϩ release significantly. Moreover, the reconstitution of the preparation with 2.5% supernatant, only slightly enhanced the total amount of Ca 2ϩ released by NAADP (120 Ϯ 12% compared with controls). Ca 2ϩ release by NAADP is also unaffected by preincubation of 2.5% homogenate with cholesteryl hemisuccinate (5 mM), an agent which rigidifies membranes and slows down protein movement in the membrane (21,22). Taken together, these observations suggest that modulatory proteins are unlikely to be involved in the opening of the channel, although the possibility of a closely associated uncharacterized modulatory protein cannot presently be ruled out. The observed lag phase could also be due to the presence of one or more slow temperature-insensitive lim-iting steps prior to channel opening.
NAADP-induced Ca 2ϩ release was best fitted to a biexponential function (see Fig. 6 for an example) as defined by the  17°C (filled circles) and 7°C (open squares). Values from one representative experiment (n ϭ 8) out of three performed. equation given above as were InsP 3 -and cADPR-induced Ca 2ϩ release. A monoexponential process proved to be completely unsatisfactory in all three cases cases. On the other hand, the Ca 2ϩ release process that followed addition of low concentrations of cADPR (below 50 nM) could not be fitted to any standard mathematical fit routinely employed. The pattern of Ca 2ϩ release induced by low concentrations of cADPR appeared sigmoidal, with the rate of release slowly increasing throughout the first few seconds of release. This can be explained by the regulatory role that cADPR plays on Ca 2ϩ -induced Ca 2ϩ release (2). It is possible that at low concentrations of cADPR, the early events of Ca 2ϩ release potentiate further Ca 2ϩ release. Rate constants for both the fast and slow phases of Ca 2ϩ release induced by NAADP, InsP 3 , and cADPR were comparable in value at saturating concentrations of agonists used (for 10 M NAADP the fast rate constant was 0.42 s Ϫ1 and the slow rate constant was 0.04 s Ϫ1 ; for 10 M InsP 3 the fast and slow rate constants were 0.40 and 0.08 s Ϫ1 , respectively, and for 5 M cADPR the fast and slow rate constants were 0.45 and 0.05 s Ϫ1 , respectively; Figs. 2-4). The InsP 3 -and cADPR-induced Ca 2ϩ release rate constants increased with concentration reaching a maximum at around 1 M (Figs. 3 and 4). The increases in the rate constants appear to closely follow the trend in increase in amplitude. Unlike in cerebellar microsomes (14), the slow rate constant for InsP 3 was responsible for most Ca 2ϩ release. This was also observed for NAADP-induced Ca 2ϩ release (Fig. 2). However, for NAADP-induced Ca 2ϩ release, the slow rate constant was concentration-insensitive, although the extent of Ca 2ϩ release by this phase increased steadily with increasing concentrations of NAADP. This finding is difficult to rationalize. The most likely explanation is that a fine balance exists between the activation and inactivation of this receptor. Since the EC 50 for inactivation is ϳ100fold lower than for activation, at low concentrations a high proportion of receptors will be desensitized without being activated. Increasing concentrations of NAADP will increase the probability that the receptors will activate before inactivating and increase the amplitude of release. As shown in Fig. 7 (for text see below), partial inactivation modifies the amplitude but not the kinetic parameters and therefore this model is plausible. In contrast, the fast rate constant increased with increasing concentrations of NAADP, but the extent of Ca 2ϩ release appeared to plateau at 100 nM and contributed no more than 25% of the overall release. Experiments conducted at 7°C did not substantially alter NAADP-induced Ca 2ϩ release, although the slow rate constant decreased by approximately 30% (data not shown). The presence of two kinetic components can suggest that NAADP releases Ca 2ϩ from two populations of Ca 2ϩ stores with different Ca 2ϩ -accumulating capacities, which would account for the two phases. Another hypothesis is that in our homogenate preparation the majority of vescicles sensitive to NAADP have a low proportion of NAADP channels, which account for the slow phase of Ca 2ϩ release, but a few vescicles contain a high proportion of NAADP receptors. With increasing concentrations of NAADP a higher proportion of receptors on this latter population of vescicles is activated and therefore Ca 2ϩ fluxes are faster.
The cooperativity of NAADP-induced Ca 2ϩ release was assessed for the fast rate constants and found to have a Hill coefficient of 1.1, suggesting that either only one binding site is involved in Ca 2ϩ release, of that if multiple sites are present they act independently of each other in influencing Ca 2ϩ channel activity. In comparison, it has been demonstrated previously that cADPR-induced Ca 2ϩ release in sea urchin eggs is a cooperative process with a Hill coefficient of approximately 1.8 (23).
We then examined the kinetic properties of NAADP-induced Ca 2ϩ release after partial desensitization with lower concentrations of NAADP. When this was performed, the magnitude of Ca 2ϩ release by maximal concentrations of NAADP was reduced in a concentration-dependent manner (Fig. 7). Interestingly, both the fast and slow rate constants were unaffected by the pretreatment with low desensitizing concentrations of NAADP (Fig. 7). One possibility is that inactivation occurs in an all-or-none manner, with submaximal inactivating concentrations of NAADP causing the full inactivation of a proportion of receptors. Increasing the magnitude of the pretreatment, this proportion of inactivated receptors would progressively increase. In such a way, while a proportion of the receptors has been shifted to the desensitized form by the pretreatment, the active ones preserved the kinetic characteristics expressed by the control.
In conclusion, NAADP-induced Ca 2ϩ release, although distinct, shows some common kinetic features with the established Ca 2ϩ -releasing molecule InsP 3 . Both release Ca 2ϩ in a biphasic manner and with similar rate constants. In both cases the slow phase of release is responsible for the majority of Ca 2ϩ release. NAADP, unlike InsP 3 , releases Ca 2ϩ after a latency. Furthermore, NAADP-induced Ca 2ϩ release appears not to be cooperative. Moreover, NAADP inactivation is most likely an intrinsic all-or-none property of the receptor. In fact, partial inactivation does not alter the kinetic properties of the release, although it alters the amplitude, which suggests that the number of receptors responding has diminished but the active receptors have not been altered by pretreatment.