Pyridine Nucleotide Metabolites Stimulate Calcium Release from Sea Urchin Egg Microsomes Desensitized to Inositol Trisphosphate*

Inositol trisphosphate (IP3) was previously shown to release Ca2+ from a nonmitochondrial store in sea urchin eggs. In this study, egg homogenates and purified microsomes were monitored with either fura 2 or Ca2+-sensitive minielectrodes to determine whether other stimuli would induce Ca2+ release. Pyridine nucleotides (whose concentrations are known to change at fertil- ization) were found to release nearly as much Ca2+ as did IP3. Average releases/ml of homogenate were: 0.6 p~ IPS, 10.9 nmol of Ca2+; 50 p~ NADP, 7.3 nmol of Ca2+; and 100 p~ NAD, 6.5 nmol of Ca2+ (n = 6). Specificity was demonstrated by screening a series of other phosphorylated metabolites, and none was found to reproducibly release Ca2+. Calcium release induced by IP3 or NADP was immediate, whereas a lag of 1-4 min occurred with NAD. This lag before NAD-induced Ca2+ release led to the discovery that a soluble egg factor (M, > 100,000) converts NAD into a highly active metabolite that releases Ca2+ without a lag. The NAD metabolite (E-NAD) was purified to homogeneity by high pressure liquid chromatography and produced half-maximal Ca2+ release at about 40 nM. Injection of E-NAD into intact eggs produced both an increase in

Inositol trisphosphate (IP3) was previously shown to release Ca2+ from a nonmitochondrial store in sea urchin eggs. In this study, egg homogenates and purified microsomes were monitored with either fura 2 or Ca2+sensitive minielectrodes to determine whether other stimuli would induce Ca2+ release. Pyridine nucleotides (whose concentrations are known to change at fertilization) were found to release nearly as much Ca2+ as did IP3. Average releases/ml of homogenate were: 0.6 p~ IPS, 10.9 nmol of Ca2+; 50 p~ NADP, 7.3 nmol of Ca2+; and 100 p~ NAD, 6.5 nmol of Ca2+ (n = 6 ) .
Specificity was demonstrated by screening a series of other phosphorylated metabolites, and none was found to reproducibly release Ca2+. Calcium release induced by IP3 or NADP was immediate, whereas a lag of 1-4 min occurred with NAD. This lag before NAD-induced Ca2+ release led to the discovery that a soluble egg factor (M, > 100,000) converts NAD into a highly active metabolite that releases Ca2+ without a lag. The NAD metabolite (E-NAD) was purified to homogeneity by high pressure liquid chromatography and produced half-maximal Ca2+ release at about 40 nM. Injection of E-NAD into intact eggs produced both an increase in intracellular Ca2+ (as assayed with indo-1) and a cortical reaction. Following Ca2+ release by each of the active agents (IP3, NAD, and NADP), the homogenates resequestered the released Ca2+ but were desensitized to further addition of the same agent. A series of desensitization experiments showed that homogenates desensitized to any two of these agents still responded to the third, indicating the presence of three independent Ca2+ release mechanisms. This is further supported by experiments using Percoll density gradient centrifugation in which NADP-sensitive microsomes were partially separated from those sensitive to IP3 and NAD.
The two most important events responsible for initiating development in a sea urchin egg after fertilization are the transient increase in intracellular Ca2+ and the subsequent alkalinization of internal pH by 0.5-0.6 pH unit (reviews by . A direct consequence of the Ca2+ transient is the induction of a massive exocytotic reaction leading to the HD17484 and National Science Foundation Grant DCB8602499 (to 11 To whom reprint requests should be addressed, formation of the fertilization envelope. The Ca2+ involved is released from internal stores (1)(2)(3), and several lines of evidence indicate that inositol trisphosphate (IP3)' is the most likely mediator of the Ca2+ release. Immediately after sperm binding, phosphatidylinositol metabolism in the egg is greatly stimulated (4,5), with the phosphatidylinositol content reaching a minimum within 30 s postfertilization (4,5 ) . Concurrently, the phosphorylated intermediates, phosphatidyl 4phosphate and phosphatidyl 4,5-bisphosphate, and the hydrolysis products, diacylglycerol and IP3, were found to increase (4)(5)(6). Evidence suggests that a GTP-binding protein is responsible for activating the lipase activity that leads to the production of IP3 (7). Microinjection of IP3 into a sea urchin egg triggers a transient increase in intracellular Ca" as monitored by fluorescent Ca2+ indicators (8) and also induces the cortical exocytotic reaction (7)(8)(9).
We have developed a cell-free system using egg homogenates and purified microsomes. This system shows ATPdependent Ca2+ sequestration and can release Ca2+ in the presence of IP3 (10). In the current study, we show that a novel metabolite of NAD is generated by a high molecular weight (?100,000) enzyme present in the egg extract and can induce Caz+ release in the cell-free system. The metabolite was purified to apparent homogeneity by HPLC and was found to be active a t a concentration range similar to that of IPS. Microinjection of the purified metabolite into an egg can induce transient Ca2+ release and initiate the cortical exocytotic reaction. Evidence is presented that IP3, the NAD metabolite, and a similar one produced from NADP all act through independent mechanisms in releasing Ca2+, thus suggesting the existence of multiple Ca2+ stores in sea urchin eggs. This interpretation was supported by partial separation of the Ca2+ stores using Percoll density gradient centrifugation.
For some experiments, microsomes were purified by Percoll density gradient centrifugation using a procedure similar to that previously reported (10). Percoll was diluted into 1.3 X concentrated AcIM to produce final concentrations of 25% Percoll and normal strength AcIM. ATP, phosphocreatine, creatine phosphokinase, benzamidine, and EGTA were added to the same concentration as that present in the crude homogenate. Then 3 ml of 25% crude homogenate was layered onto 7 ml of 25% Percoll and centrifuged for 40 min (25,000 X gav, 10 "C). The upper band contained IP3-sensitive microsomes and was removed and utilized for Ca2+ uptake and release assays. This fraction will be referred to as purified microsomes. To prepare microsome-free egg supernatant, the top 2.5 ml of each Percoll gradient was removed and centrifuged an additional 20 min at 100,000 X g. This is referred to as a crude egg supernatant.
Throughout this study, the concentrations of crude homogenates, purified microsomes, or supernatant are expressed as a percentage of the original volume of packed eggs. The crude homogenate was prepared from packed eggs that were diluted (-fold and is called a 25% homogenate. Since the Supernatant from the Percoll gradients is undiluted, it is a 25% supernatant. The volume of purified microsomes recovered from each Percoll gradient was typically 0.75 ml, and since 3 ml of a 25% homogenate was applied, the recovered microsomes would be suspended in the same volume as the original packed egg, thus representing a 100% suspension. For each component, typical concentrations used in this study (based on egg volume) and respective protein concentrations after dilution (mg/ml f S.D.) are: 2.5% crude homogenate (2.84 * 0.48 mg/ml protein, n = 5), 5% purified microsomes (0.30 f 0.06 mg/ml, n = 4), and 25% egg supernatant (12.0 * 3.5 mg/ml, n = 5).

Monitoring ea2+ Fluxes in Homogenates and Microsomes-Ca2+
release and resequestration by egg microsomes were monitored by measuring changes in the Ca2+ concentration of the medium. Both crude homogenates and purified microsomes were diluted into GluIM containing 1 mM ATP, 2.5 mM benzamidine, and 10 p~ EGTA. The medium used for Ca2+ release measurements (GluIM) was different from that used for the homogenization and Percoll separation (AcIM), because microsomes suspended in GluIM responded better to Ca2+ release agents (by releasing about twice as much Ca2+ as when suspended in AcIM), whereas the lower density AcIM was better for the purification procedure, since it allowed twice as much homogenate to be separated on each Percoll gradient. For most experiments, Ca2+ levels in 1-ml aliquots were monitored with the fluorescent Ca2+ indicator, fura 2 (0.5 pM) (11). To minimize the fluorescence interference of reduced pyridine nucleotides, fluorescence was measured at 400-nm excitation (2.5-nm slit) and 540-nm emission (10-nm slit). For each experiment, the relationship between fura 2 response and changes in Ca2+ concentration was determined by adding known amounts of Caz+ to an aliquot of the same homogenate. As will be indicated later, certain key results were verified with Ca2+-sensitive minielectrodes (prepared and used as previously reported, Ref. 10).
Membrane Filtrations and Incubations to Produce Enzyme-actiuated NAD-To remove small molecular weight components present in the egg supernatant, aliquots of supernatant were washed on 100,000 M, cutoff filters (XM-100A membranes, Amicon Corp., Lexington, MA). Typically 5 ml of 25% supernatant was filtered to 0.1-0.2-ml retained volume; 1 ml of GluIM was added and refiltered to 0.1-0.2 ml. The number of washes was repeated 3-4 times and the retentate recovered in 2.5 ml.
To produce enzyme-activated NAD (E-NAD), 1-2.5 mM NAD was added to washed retentate and incubated for 3-7 h at 17 "C. Ca2+ release activity was monitored by adding aliquots of the incubation mixture to 2.5% homogenates, and the incubation was stopped (by placing the mixture on ice) when maximum Ca2+ release activity had developed. The incubation mixture was filtered on a 10,000 M. cutoff filter membrane (Amicon PM-10) to separate the E-NAD from the washed supernatant, and high levels of Ca2+ release activity were recovered in the filtrate.
Alkali Treatment of NAD and NADP-This procedure is based on a method by Lowrey and Passonneau (12) which was reported to alter NAD and NADP but not their reduced forms. Solutions of 50 mM NAD and NADP were prepared in 80 mM K&O3 and 20 mM KHC03, the pH was adjusted to 10.5 with KOH if needed, heated at 60 "C for 6-8 min, placed on ice to stop the reaction, and the pH was readjusted to 8 to prevent further reaction. The resultant products with Ca2+releasing activity are called alkaline-activated NAD (A-NAD) and alkaline-activated NADP (A-NADP).
Purification of Activated NAD and NADP by HPLC--In order to quantitate the calcium release activity of E-NAD, A-NAD, and A-NADP during purification steps, a unit of Ca2+ release activity was defined as the amount sufficient to produce half-maximal Ca2+ release from 1 ml of a 2.5% crude egg homogenate. Therefore, each fractioncontaining activity was serially diluted, and 2-10 pf was added to 1ml aliquots of crude homogenate to determine the dilution and volume required to produce half-maximal CaZ+ release. From the volume (pl) added to the homogenate, the serial dilution required, and the initial volume of the fraction, the total units of activity in the initial fraction were computed.
All three activators, E-NAD, A-NAD, and A-NADP, were prepared as previously indicated and purified by two sequential separations on 0.46 X 15-cm AG MP-1 anion exchange columns eluted with 1-150 mM trifluoroacetic acid gradients (13). For the first separation, each column was maximally loaded and produced incomplete separation of UV-absorbing peaks (A2M). For each activator, the Ca2+ release activity applied was: E-NAD (1,230 units), A-NAD (1,980 units), and A-NADP (32,000 units). One-ml fractions were collected, vacuum evaporated, resuspended in 0.05-0.1 ml of distilled water, and assayed for Caz+-releasing activity with 2.5% crude homogenates. For each activator, fractions containing activity from the first separation were pooled and rechromatographed, and this time each produced activity associated with a single clearly separated UV peak. Following the two AG MP-1 separations, the percent recoveries of activity and yields (based on initial NAD or NADP) were: E-NAD (110% recovery, 1.8% yield), A-NAD (20% recovery, 0.19% yield), A-NADP (58% recovery, 1.5% yield).
For each Ca2+-releasing factor, 80 units of activity previously purified by the two AG MP-1 separations were analyzed by reverse phase HPLC (14) to determine whether Ca2+-releasing activity copurified with the absorbance at 254 nm. One-ml fractions were collected, dried, and analyzed for Caz+-releasing activity as before.
Monitoring Intracellular Calcium in Intact Eggs-Two sequential iontophoretic injections were used to first load intact eggs with the fluorescent Ca2+ indicator, indo-1 (ll), and then to introduce purified E-NAD or A-NADP. The current parameters used for these injections were 30 pulses/min, 1 s/pulse, and 2-3 nA. Reagent concentrations in the micropipettes were: E-NAD (1.7 units/pl), A-NADP (1.9 units/ pl), indo-1 (10 mM), and NAD (5 mM). Resultant fluorescence changes were monitored at 365-nm excitation and 485-nm emission with a fluorescence microscope equipped with a silicon-intensified target camera (RCA TC1030/H) whose output was digitized. The details of the instrumental setup will be published elsewhere. Autofluorescence from an adjacent uninjected egg was also measured and subtracted from the fluorescence signal of the injected egg.

RESULTS
Calcium Release from Crude Homogenates-In a previous study, we used L. pictus egg homogenates and microsomes as a cell-free system to study Ca2+ uptake and release and found that IP, induced Ca2+ release from a nonmitochondrial store (10). In the present study, the same cell-free system is used, but the Ca2+ movement was monitored with a different fluorescent Ca2+ indicator, fura 2. The fluorescence properties of fura 2 allow the measurement of Ca2+ with little interference from the fluorescence of reduced pyridine nucleotides. At the excitation wavelength used, fura 2 fluorescence decreases in response to increasing Ca2+. Fig. l a shows IP3-induced Ca2+ release from the egg homogenate as indicated by the observed decrease in fura 2 fluorescence.
To determine whether other cell metabolites might stimulate Ca2+ release, a series of phosphorylated molecules was screened. Since the concentrations of pyridine nucleotides are known to change after fertilization of sea urchin eggs (15), NAD and NADP were assayed first. Both were found to release Ca2+ at physiological concentrations ( Fig. 1, b and c).
With six different homogenates, activators were added at concentrations that produced maximal Ca2+ release, and 1-ml aliquots of 2.5% crude homogenate produced Ca2+ releases (k S.D.) of: 10.9 f 3.6 nmol of Ca2+ with 0.6 pM IP3, 6.5 f 1.3 nmol of Ca2+ with 100 p~ NAD, and 7.3 f 2.4 nmol of Ca2+ with 50 p~ NADP. The Ca2+ efflux began immediately upon addition of IP3 and NADP, whereas a delay of 1-2 min occurred before Ca2+ was released by NAD (Fig. IC). Evidence will be presented later that the delay before NAD-induced Ca2+ efflux is due to the conversion of NAD to a more active form.
Desensitization Experiments Implicate Three Calcium Release Mechanisms-To determine whether IP3, NAD, and NADP all release Ca2+ via the same mechanism, homogenates were desensitized to IP3 before addition of NAD and NADP. A previous study had shown that homogenates resequester Ca2+ after stimulation by IP3 but are desensitized to subsequent IPS addition (10). Fig. 2 shows that the first addition of IPS to the egg homogenate produced a large Caz+ release followed by resequestration. Subsequent additions induced progressively less Ca2+ release indicating desensitization. The desensitized homogenate, however, could still respond to NAD and release Ca2+ upon its addition. Similar to IP3, NAD also produced desensitization in the homogenate. Finally, NADP stimulated Ca2+ release after desensitization to both IP3 and NAD. The released Ca2+ was again resequestered by the C E 7t homogenate. If the order of reagent addition is changed, each reagent still releases Ca2+, e.g. IP3 stimulated Ca2+ release from homogenates desensitized to NAD and NADP (data not shown).
These results implicate the presence of 3 different Ca2+ release mechanisms in L. pictus eggs. When the experiments depicted in Figs. 1 and 2 were repeated with a Ca2+-sensitive minielectrode, the same results were obtained, therefore, verifying that the observed changes represent real Ca2+ fluxes and not artifacts of the fluorescent Ca2+ assay method. Also, the intracellular Ca2+ blocking agent 8-(diethylamino)-octyl-3,4,5-trimethoxybenzoate (3 mM) produced 290% inhibition of Ca2+ release by all three reagents.
Calcium Release from Purified Microsomes-We had previously shown that Percoll density gradient centrifugation could be used to purify microsomes that released Ca2+ upon stimulation by IP, (10). Using the same protocol but with a lower density medium (AcIM) produced a similar banding pattern on Percoll gradients and allowed a larger volume of crude homogenate to be purified on each gradient.
The upper band (which contained IP3-responsive vesicles) was removed and assayed for response to NAD and NADP. Fig. 3A shows that the purified microsomes responded to both NADP and IP3 but not to NAD. However, if the supernatant from the top of the Percoll gradients was centrifuged at 100,000 x g and added to purified microsomes, responsiveness to NAD was restored (Fig. 3B). A similar time delay between NAD addition and Ca2+ release was observed with both purified microsomes (Fig. 3B) and crude homogenates (Fig. 1). The delay was found to be proportional to the concentration of supernatant added so that Ca2+ release started about twice as fast when the supernatant concentration was doubled (data not shown). After the release, the Ca2+ was again resequestered, and yet the microsomes did not respond to the second addition of NAD. Similar to that observed in the crude homogenate, the desensitized microsomes can still respond to NADP and IPS as shown in Fig. 3B. Comparing Fig. 3, A and B, it can be seen that both the amount and the kinetics of Ca2+ release induced by IP3 and NADP were not affected by the presence of the high speed supernatant.
NAD Is Converted into an Active Form-The characteristics of NAD action are consistent with a soluble enzyme using NAD as a substrate to produce a Ca2+ release activator. To test this hypothesis, NAD (2.5 mM) was incubated with 25% supernatant before being added to a crude homogenate. Fig.  4a shows that such preincubated NAD (N+S in the Fig. 4a) induced Ca2+ release without a lag and at a lower concentration than would induce Ca2+ release without preincubation (Fig. 4b).
Although the preincubated NAD showed no lag in Ca2+releasing activity when added to purified microsomes, supernatant addition was still required for maximal Ca2+ release. Fig. 4c shows that the magnitude of Ca2+ release was proportional to the concentration of supernatant added. In the presence of increasing amounts of high speed supernatant, the same amount of preincubated NAD produced increasing amounts of Ca2+ release. This result implies that an additional factor is necessary for purified microsomes to respond to the Ca"-releasing factor.
The soluble components of the NAD system were fractionated by membrane filtration. Utilizing a series of membrane filters of defined pore sizes, it was determined that the Ca2+releasing factor passed through a 10,000 molecular weight cutoff membrane, whereas both the enzyme and the factor required for responsiveness of purified microsomes were retained by a 100,000 M , cutoff membrane. The combination of these two fractions reconstituted the Ca2+-releasing activity in the purified microsomes with characteristics similar to  homogenates ( a and b) or 5% purified microsomes (c) was monitored with fura 2 as previously described. NAD was preincubated with supernatant by adding 2.5 mM NAD to 25% crude egg supernatant and incubating for 3 h at 17 "C. At the arrows marked N+S, 4 pl of this NAD plus supernatant was added to 1.0 ml of homogenate, thus producing a concentration equivalent to 10 PM original NAD. At the arrow marked NAD, 20 or 100 PM NAD was added, with 20 p~ NAD producing no Ca2+ release and 100 PM producing Ca2+ release after a delay of 1-2 min. Controls of 2.5 mM NAD incubated in AcIM without supernatant and 25% supernatant incubated without NAD produced no Caz+ release activity when added at twice the volume of NAD plus supernatant. In c, preincubated NAD was added to aliquots of 5% purified microsomes previously mixed with 0.25,0.5,1, or 2% supernatant, with increasing supernatant concentration producing increasing Ca2+ release. Controls in which supernatant alone was added produced no Ca2+ release. those described in Fig. 4. Therefore, it appears very likely that the Ca2+-releasing factor itself is a small molecular weight metabolite of NAD, while the converting enzyme and the additional supernatant factor were both high molecular weight proteins.
Purification of Enzyme-activated NAD-The behavior of these factors on membrane filters was used to prepare an extract highly enriched for the NAD-derived Ca2+-releasing factor. The procedure is described under "Experimental Procedures," and in summary consists of first washing the supernatant on the 100,000 molecular weight cutoff filter to both remove small molecular weight cytoplasmic components and concentrate the enzyme, and then incubating with NAD until maximum Ca2+-releasing activity was developed. Finally the large molecular weight components were removed by filtration on the 10,000 molecular weight cutoff membrane, and the resultant Ca2+-releasing activity was recovered in the eluent.
This enzyme-activated NAD (E-NAD) was further purified by anion exchange HPLC using a procedure known to provide good separation of nucleotides (13). Fig. 5A shows that the initial preparation contained several peaks that absorb at 254 nm, with Ca2+-releasing activity correlating with a small UV peak that eluted at 24 min. In the preparation analyzed here, 2.5 pmol of NAD was incubated with washed supernatant to produce 1230 units of Ca" release activity; then the large molecular weight components were removed by filtration on a 10,000 M, cutoff membrane filter (as described under "Experimental Procedures"). UV absorbance at 254 nm (A,) and Ca" release activity (percent recovered activity) are plotted. One-ml fractions were collected, vacuum evaporated, reconstituted in 0.1 ml of distilled water, and assayed for Ca2+-releasing activity. A, initial separation on an anion exchange (AG MP-1) column using the protocol described for the first separation under "Experimental Procedures." One hundred forty-five units were applied and 155 units recovered. Due to overloading, the retention times for various UV peaks in this chromatogram may not be accurate. B, reverse-phase HPLC analysis of activity purified by two sequential separations on AG MP-1 columns (as described under "Experimental Procedures"). Eighty units were applied and 67 units recovered.
Upscaling the procedure to preparative range using a sample 8 . 5~ larger than that shown in Fig. 5A'resulted in incomplete separation of UV-absorbing peaks. All the fractions containing Ca2+-releasing activity were, therefore, combined and rechromatographed on a second AG MP-1 column. The second chromatogram showed clearly separated UV peaks with the Ca2+ release activity correlating with a single peak that was eluted at 17.8 min (data not shown). The slight difference in elution times for E-NAD between Fig. 5A (24 min) and the second preparative run (17.8 min) is due to either different AG MP-1 columns being used or else the column in Fig. 5A was overloaded.
Next, an aliquot of E-NAD purified by the two sequential separations was further analyzed by reverse phase HPLC (14). A single UV peak with a retention time of 9.4 min copurified with the Ca2+-releasing activity (Fig. 5B). Since anion exchange and reverse phase HPLC separate according to different principles, these results, therefore, provide strong evidence that E-NAD is purified to homogeneity and contains a moiety that absorbs at 254 nm.
In a first attempt to characterize E-NAD, UV spectra were compared for E-NAD, NAD, adenosine 5'-diphosphoribose (ADP-ribose), adenosine 5'-monophosphate (AMP), and pnicotinamide mononucleotide (NMN). The spectra for E-NAD, NAD, ADP-ribose, and AMP were nearly identical with each showing a peak at 257-260 nm and a minimum at 227-233 nm. NMN, on the other hand, showed a significantly different spectrum with a small peak at 262 nm and a minimum at 245 nm. These results indicate that the adenine group is retained in E-NAD but provide no information concerning other possible modifications of the NAD structure.
The absorbance at 258 nm was then used to estimate the active concentration of E-NAD. Assuming the extinction coefficient at 258 nm is the same for both NAD and E-NAD, it could be determined that 80 units/ml of E-NAD would correspond to 2.8 ~L M NAD. One unit/ml was then calculated to be 35 nM; and from the definition of a unit of activity, 35 nM is the approximate concentration that produces one-half maximal Ca2+ release. Since the extinction coefficient of E-NAD is unlikely to be too much higher than NAD, this estimate, although quite rough, does put the active concentration of E-NAD in the range of that of IP,.
Ca2+ release by E-NAD and IPS was then compared, with each being used at concentrations that produce greater than 90% maximal Ca2+ release. In two experiments with 2.5% crude homogenates, IP3 (600 nM) released 14 and 8.6 nmol of Ca2+/ml of homogenate, and E-NAD ( 5 units/ml) released 11.5 and 6.3 nmol of Ca2'/ml. Therefore, E-NAD released an average of 78% as much Ca2+ as did IPS. Desensitization experiments showed that E-NAD desensitized the component of the Ca2+ release activity which was sensitive to NAD but did not desensitize microsomes to either IPS or NADP. Calcium Release Induced by Injection of Activated NAD into Intact Eggs-The egg was first loaded with the fluorescent Ca2+ indicator, indo-1; then the fluorescence was monitored as E-NAD was iontophoretically injected. Fig. 6 shows that each injection was followed by a transient decrease in fluorescence, thus demonstrating Ca2+ release followed by resequestration. After the first injection, the egg was observed with phase contrast optics and had undergone a complete cortical reaction. If more E-NAD was injected 5.5 min after the first injection, additional Ca" was released. This shows that any desensitization produced by E-NAD is reversed by about 5 min and implies that the egg possesses a system for removing Three more eggs were injected with indo-1 followed by E-E-NAD. Ca*+ release (decreases in indo-1 fluorescence) was already detected after the first current pulse. Complete cortical reaction was observed at the time period labeled CR. In the second injection of E-NAD, a total of 18 current pulses was used. Change in fluorescence was detected after the first three pulses.
NAD, and all showed free Ca2+ increases followed by cortical reactions. In three control experiments, the same protocol was used, only the pipettes were filled with 5 mM NAD instead of E-NAD. All three eggs showed much less increases in Ca2+ (6-7-fold smaller fluorescence decreases than with E-NAD) and no cortical reactions. Two of these control eggs were subsequently fertilized by adding sperm, and both showed large Ca2+ increases followed by cortical reactions, indicating they were not damaged by the microinjection.
NAD Analogs and Alkaline-activated NAD-Several commercially available analogs for NAD were screened for Ca2+releasing activity upon addition to 2.5% crude homogenates. Only NADP released Ca2+ at less than 100 p~; however, NADP cannot be E-NAD since NADP does not desensitize homogenates to E-NAD and vice versa. Molecules found to release Ca2+ at 100-500 PM include NADH, NADPH, adenosine 5'-diphosphoribose (ADP-ribose), and nicotinic acid adenine dinucleotide. All four showed no lag before Ca2+ release, and all four desensitized homogenates to E-NAD; therefore, they are most likely acting as analogs to E-NAD. Little or no Ca2+ release was induced by 200-500 PM nicotinamide, nicotinamide mononucleotide, or &-NAD. Also none of these desensitized homogenates to E-NAD.
A comparison of the structures of active NAD analogs indicates that activation of NAD may involve neutralization of the positive charge on the nicotinamide moiety. Lowry and Passonneau (12) reported an alkali treatment that alters NAD but not NADH, with a major initial reaction being a hydrolysis that opens the nicotinamide ring and neutralizes the positive charge originally associated with the ring (16).
Alkali treatment of NAD (using the procedure described under "Experimental Procedures") produced a substance with Ca2+-releasing activity very similar to that of E-NAD. The alkali product produced Ca2+ release without a lag, desensitized homogenates only to NAD, and required the equivalent of 50 p~ original NAD for maximal Ca2+ release. This alkaline-activated NAD is called A-NAD.
A-NAD was then purified by anion exchange HPLC and analyzed by reverse phase HPLC using the same procedures as for E-NAD. Respective elution times for the UV peaks associated with Ca2+-releasing activity for E-NAD and A-NAD were identical in both the anion exchange system (17.8 min) and the reverse phase system (9.4) min). The UV spec-trum of the HPLC-purified A-NAD was also found to be the same as E-NAD and NAD.
The active concentration of A-NAD was determined from the absorbance at 258 nm, assuming the same extinction coefficient as NAD, and 48 nM was found to induce halfmaximal Ca2+ release from 2.5% crude homogenates. This is close to the 35 nM active concentration determined for E-NAD. Furthermore, A-NAD also desensitized microsomes only to E-NAD and NAD but not to IP, or NADP. All together, these results indicate that E-NAD and A-NAD are probably the same molecule.
Alkaline-activated NADP-Both the enzyme and the alkaline treatments as developed for NAD were also tested on NADP to determine if NADP could also be converted to a more active form. Incubation of NADP plus supernatant using the same conditions as used for the production of E-NAD did not produce a convincing increase in Ca2+-releasing activity; however, alkaline treatment of NADP produced about a 30fold enhancement. Before the alkaline treatment, the concentration of NADP required to produce half-maximal and maximal Ca2+ release was found to be about 75 and 400 p~, respectively. After alkaline treatment, these values were reduced to about 2.5 and 10 p~, respectively.
Alkaline-treated NADP (A-NADP) was then purified by anion exchange HPLC as was used for E-NAD and A-NAD. On the anion exchange column, activity eluted with a UV peak at 30 min as compared to the elution time of 17.8 min for E-NAD. These elution times clearly show that A-NADP is different from E-NAD. Also, the purified A-NADP induced Ca2+ release from microsomes desensitized to both IP3 and E-NAD as shown in Fig. 7.
The UV spectrum of A-NADP was the same as for NADP and NAD; therefore, the adenine group was retained. Again, assuming the extinction coefficient of A-NADP is the same as NAD, its half-maximal active concentration can be estimated to be 38 nM, very similar to that of E-NAD. The maximal amount of Caz+ release in crude homogenates induced by A-NADP was on average 67% of that released by IPS and 85% of that released by E-NAD (n = 2). Also, all three reagents showed similar kinetics of Ca2+ release, with release beginning immediately and being maximal by 2 min.
Purified A-NADP was injected into intact eggs, and free Ca2+ was monitored using the same protocol as for E-NAD. A-NADP produced both transient Ca2+ increases (similar to that shown in Fig. 6) and cortical reactions in all four eggs injected, therefore, showing that intact eggs respond to A-NADP.  0.47 f 0.21 99.5 f 0.21 "Homogenates were separated on Percoll gradients as described under "Experimental Procedures." Results from three homogenate preparations are averaged, and percent activity (+ SD) is reported to normalize differences in absolute values.
'Cytochrome c oxidase was assayed following the procedure of Smith (29).
mogenates, the response to NADP always decreased more than the response to either IP3 or NAD (e.g. Figs. 3 and 7). One possible explanation would be if part of the NADPresponsive microsomes was lost when the upper band was removed from the Percoll gradients. To test this hypothesis, upper and lower bands (the only two bands present on those Percoll gradients) were recovered separately and assayed for response to IP,, E-NAD, and A-NADP. Table I shows that the upper band contained over 90% of the responsiveness to IP, and E-NAD. On the other hand, the A-NADP sensitivity distributed mainly in the lower band with only about 23% of the activity in the upper band. These results clearly show that the responsiveness to A-NADP resides in a different Ca2+ store.
Table I also compares the distribution of a mitochondrial marker enzyme, cytochrome c oxidase, with that of the Ca2+releasing activities; and the former was found almost exclusively (99.5%) in the lower band. This indicates that the IP, and the E-NAD-sensitive Ca2+ stores are nonmitochondrial. The situation is less clear for the A-NADP-sensitive Ca2+ store. Although the larger portion of the activity was in the lower mitochondrial band, the distribution pattern between the two bands does not suggest that it copurified with the mitochondria. Thus, from the three gradients summarized in Table I, the upper band contained, on the average, 23% of the A-NADP sensitivity but less than 0.5% of the mitochondrial activity. This disparity of distribution pattern was even more striking in one of the three gradients, showing 42% of A-NADP activity in the upper band but with only 0.3% of the cytochrome c oxidase activity. These results, therefore, suggest that the A-NADP-sensitive Ca2+ store is also nonmitochondrial and with a buoyant density slightly higher than the two other nonmitochondrial Ca2+ stores found in the upper band.

DISCUSSION
The discovery that IP3 can induce Caz+ release directly from internal stores attracts a great deal of interest focusing on the role of IP, as a second messenger for intracellular Ca2+ changes. In most of the systems studied so far, IP, production is stimulated as a result of the interaction between external stimuli and cell surface receptors (reviews by Refs. 17 and 18). Similarly, interaction between sperm and sperm receptors on the surface of sea urchin eggs activates a GTP-binding protein which then leads to the production of IP, and Ca2+ release from internal stores (7). However, not all intracellular increases are mediated by surface receptor activation. For example, in the sea urchin egg, about 15 min after the first Ca2+ changes at fertilization, the intracellular Ca2+ increases transiently again (19). This occurs in the absence of external stimulus and is correlated with the pronuclear migration (19). Thereafter, multiple Ca2+ transients can be detected correlating with various mitotic events (19).
We are interested in knowing if all of these Ca2+ transients are triggered by IPS and, if not, whether other activators can be identified. We had previously developed a cell-free system for studying the IPS-sensitive Ca2+ release mechanism (10). This system is ideally suited to screening a large number of potential Ca2+ release activators. Pyridine nucleotides were tested first because it is known that shortly after fertilization, there is a large and rapid conversion of NAD to NADP which is followed by the reduction of NADP to NADPH (15). These changes in pyridine nucleotides have been postulated to have important consequences for the later activation of the nucleus (20). In this study, we showed that both NAD and NADP at physiological concentrations of 50-100 p~ were able to induce Ca2+ release in the cell-free system. On the average, the amount of Ca2+ released is 60-70% of that released by IP3. Unlike that of IP3 which releases Ca2+ immediately upon addition, the kinetics of Ca2+ release induced by NAD shows a lag period. This suggested that NAD needed to be converted to an active form before it could release Ca2+. This was found to be the case, since preincubation of NAD with a high speed supernatant of the egg homogenate produced an active form of NAD (E-NAD) which was purified to apparent homogeneity by HPLC. Microinjection of purified E-NAD into sea urchin eggs induced Ca2+ release and triggered cortical exocytosis, thus indicating it is active within a living cell.
E-NAD was found to be of small molecular weight, since it passed through a 10,000 molecular weight cutoff filter. The UV absorption spectrum of E-NAD was identical to that of NAD and ADP-ribose suggesting that the modification was likely to be on the nicotinamide and not on the adenine group. An alkaline treatment (known to attack the nicotinamide group) was tested on NAD and was found to be effective in generating the active form of NAD. The active substance (A-NAD) was shown to be identical to E-NAD in both the UV absorption characteristics and also retention times on anion exchange and reverse phase HPLC. This strongly indicates that both A-NAD and E-NAD are the same substance. A similar alkaline treatment was also able to activate NADP.
The active substance (A-NADP) has an identical UV absorption spectrum as E-NAD and A-NAD but is retained longer on anion exchange HPLC. This is likely due to the additional phosphate group and indicates that A-NADP is structurally different from E-NAD and A-NAD.
The active concentrations of E-NAD, A-NAD, and A-NADP that produced half-maximal Ca2+ release were all estimated to be 30-50 nM. This estimate was based on the assumption that they all have the same extinction coefficient as NAD. This assumption should be quite reasonable since the extinction coefficient of NAD is quite insensitive to modifications on the nicotinamide group. Thus, a survey of published extinction coefficients for NAD, NADP, NADH, NADPH, ADP-ribose, ADP, adenosine, and adenine shows that none of them differ by more than 25%. The uncertainty in the estimated half-maximal concentrations is, therefore, also very likely not to exceed 25%, and this puts them in the same active range as IP3. Screening of common derivatives of NAD showed that none of them can induce Ca2+ release at this concentration range. This suggests that E-NAD is likely to be a novel derivative of NAD specifically serving as a Ca2+ release activator.
The enzyme in the supernatant of the egg extract which was responsible for converting NAD to its active form was found to be of high molecular weight, since it was retained by a 100,000 molecular weight cutoff filter. Also found in the retentate is another factor which was required for E-NAD to release Ca2+ from egg microsomes. The presence of these factors in the high speed supernatant of the egg extract suggests both are cytoplasmic components, although the possibility that they were solubilized during homogenization cannot be excluded. In any case, it is likely that regulatory mechanisms exist in vivo for controlling either the enzyme activity or the availability of the enzyme itself. Thus, the NAD system, as characterized in this study, possesses at least three potential controlling sites: first, the availability of the substrate NAD; second, the regulation of the converting enzyme; and third, the availability of the high molecular weight factor. Regulation at one or more of these sites should provide a tight control of the Ca2+ release activity of E-NAD. Furthermore, the egg also possesses a very active removal system for E-NAD. This can be inferred from the microinjection experiment shown in Fig. 6. The egg took only about 5 min to recover from the first injection of E-NAD and released Ca2+ again in response to a second injection. In comparison, a similar microinjection experiment using IP3 showed a recovery period of about 15 min (8). Therefore, the NAD system appears to have all the necessary characteristics of a signaling system.
Recent studies have suggested that GTP (21) and arachidonic acid (22) can also induce Ca2+ release from internal stores independently of IP3. We have tested both in our system and found that GTP could induce Ca2+ release only at a much higher concentration (2-4 mM) than reported to be effective (3-5 /IM) in cultured neuronal cells (21). The effect was also not very reproducible and was not studied further. Arachidonic acid at 20 pM (twice the effective concentration in pancreatic islets (22)) did not release Ca2+ in our system and neither did it potentiate nor inhibit Ca2+ release induced by IP3 or NAD. However, it (20 p~) did specifically block 90-98% of the Ca2+ release induced by NADP and A-NADP with half-maximal inhibition occurring at 5 p~.
Pyridine nucleotides have also been implicated in the organic peroxide-induced Ca2+ release from mitochondria. It was proposed that NAD-or NADP-dependent ADP-ribosylation of a mitochondrial protein mediates hydroperoxideinduced Ca2+ release (23). This mechanism is unlikely to explain the pyridine nucleotide-induced stimulation of Ca2+ release in sea urchin egg homogenates since: 1) Ca2+ release does not correlate with the presence of mitochondria (Table  I); 2) the addition of an organic peroxide is not needed for pyridine nucleotide-induced Ca2+ release; and 3) ADP-ribosylation is probably not involved, since NAD is converted into a soluble metabolite (E-NAD) that is neither nicotinamide nor ADP-ribose.
Another important conclusion of this study is that sea urchin eggs possess multiple systems for releasing Ca2+ from internal stores. Two lines of evidence support this conclusion. First, repeated additions of either IP3, E-NAD, or A-NADP can induce desensitization. Microsomes desensitized to any two of the activators will release Ca2+ in response to the third (e.g. Figs. 2, 3, and 71, indicating all three activators act independently. Second, Percoll density gradient centrifugation allowed a partial separation of the A-NADP-sensitive Ca2+ store from those sensitive to IP, and E-NAD (Table I). All three stores appear to be nonmitochondrial since their