Inositol Trisphosphate Induces Calcium Release from Nonmitochondrial Stores in Sea Urchin Egg Homogenates "

This study presents evidence that inositol trisphosphate (IPS) releases Caz+ from intracellular stores in sea urchin eggs. First, high voltage discharge was used to transiently permeabilize eggs and introduce IP3; the resultant induction of cortical reactions (a well characterized Ca"-dependent event) provided indirect evidence that IPS released Ca2+ from intracellular stores. Next, Ca2+ uptake and release from egg homogenates and homogenate fractions were monitored by both Ca2+ minielectrodes and the fluorescent Ca2+ indicator, quin-2. Both assay methods showed Ca" release upon IP, addition, with a half-maximal response at 50-60 nM IP3 and maximal Ca2+ release at -1 p~ I P S . Homogenates were 300-fold more sensitive to I P 3 than IP,, and Ca2+ release was 95% inhibited by the Ca2+ antagonist TMB-8 (3 mM). Fractionation by density gradient centrifugation showed that activities for Ca2+ sequestration and IP, responsiveness co-purified with endoplasmic reticulum microsomes. Following an initial IP, addition, homogenates were refractory (desensitized) to additional IPS. However, if homogenates were centrifuged and the vesicles resuspended in media lacking IP3, they would respond to added IP3, therefore, showing that desensitization is most likely due to the presence of IP,. This study also shows that the mechanism of IP, action is inherent to the microsomes and ions present in the medium used, with no cytoplasmic factors being required. The stability of this microsome preparation and the purification obtained with density gradient centrifugation make this a promising system with which to further characterize the mechanism of Ips action.

When sea urchin eggs are fertilized, development is activated by the combined effects of a transient increase in intracellular Ca2+ and a long-duration increase in intracellular pH (reviews by Refs. 1-3). The Ca2+ transient is produced by Ca2+ released from intracellular stores (1-3); however, identifying the mechanism by which sperm attachment at the cell surface releases Caz+ from stores within the egg has proved elusive. Only recently has research on phosphatidylinositide metabolism provided a likely mechanism (review by Ref. 4).
Recent studies with sea urchin eggs have reported a 40% increase in phosphatidylinositol 4,5-bisphosphate concentration by 15 s after fertilization of Strongylocentrotus purpuratus eggs (ll), thus linking phosphatidylinositide metabolism with fertilization. Also, IPS injected into Lytechinus pictus eggs induced cortical reactions (12), and since the egg cortical reaction is a Ca2+-mediated event, this injection experiment provides indirect evidence that IPS mediates Caz+ release in these eggs.
We report here the first direct evidence that IPa releases Ca2+ from intracellular stores in sea urchin eggs. Ca2+ uptake by and release from egg homogenates was monitored by both quin-2 and a Ca2+-sensitive minielectrode. Ca2+ uptake was ATP dependent, and IP3 induced the release of Caz+. Percoll density gradients were used to fractionate components of these homogenates and showed co-purification of IP, responsiveness, ATP-dependent. Ca2+ sequestration, and glucose-6phosphatase activity (a marker enzyme for the endoplasmic reticulum); and all three activities were clearly separated from cytochrome 'c oxidase activity (a mitochondrial enzyme). Therefore, IPS induces Ca2+ release from a nonmitochondrial store which is most likely the endoplasmic reticulum.

Animal Maintennnce and Gamete Handling
Gametes from Lytechinuspictus and Strongylocentrotuspurpuratus were collected by intracoelomic injection of 0.5 M KCI. Eggs were spawned directly into reagent grade artificial seawater (ASW, described below), filtered through 95-pm (S. purpuratus) or 125-pm (L. pictus) Nitex screen cloth (Tetko, Inc.), dejellied by brief exposure to pH 5 ASW, and used within 6 h. An aliquot of each egg preparation was tested for fertilizability, and only preparations with 270% of the eggs showing an elevation of fertilization envelopes were used.

13947
IP3 Rekases ea2+ from Sea Urchin Egg ~e m b~a~s mine, 20 mM HEPES, 5 mM NaCl, and 1.0 mM MgC12. The pH was adjusted to 7.2 by the addition of acetic acid.
For certain experiments, EGTA and Ca2+ were added to OCaSW, IM, or egg homogenates to produce media containing a range of free Caz+ concentrations. The resultant free Caz+ levels were computed with the program of Fabiato and Fabiato (15). The stability constants used for EGTA were given by Fabiato and Fabiato (15) and those for gluconate were from Martell and Smith (16). Cap+ and EGTA were added from 0.5 M stocks of CaNaaEGTA and N a G T A (for OCaSW) or CaKzEGTA and yEGTA (for IM) to produce final EGTA con-cen~ations of either 1 mM (OCasW) or 10 mM (IM) and various free Ca2+ concentrations. These stocks were prepared from CaCOs, EGTA, and NaOH or KOH. The concentrations of Na" and K+ were kept constant by the addition of NaCl or KC1 as needed, and free MgZ+ was kept constant by the addition of MgCt,. After adding these reagents, the pH was adjusted to 8.00 t 0.02 for OCaSW or 7.10 -rt 0.02 for IM or egg homogenates. The Ca:EGTA solutions in OCaSW were buffered with 10 mM Tris.
Preparation of Phusphoinositols The inositol pol~hosphates used in this study were generously supplied to us by Drs. T. F. Walseth and N. D. Goldberg, University of Minnesota. IPS, IPp, and IPS were prepared by alkaline hydrolysis of their respective phosphatidylinositides using the protocol of Grad0 and Ballou (17). They were then purified by high pressure liquid chromatography on an AG MP-1 anion-exchange resin eluted with trifluoroacetic acid under conditions similar to those used by Axelson et a,?. (18) to separate nucleotides.
Permeabilization of Eggs by High Voltage Discharge (HVD) HVD to determine whether IPS introduced from the extracellular L. pictus eggs were transiently p r e m e a b~~~d by brief pulses of medium would induce CaZ+ release from intracellular stores (as assayed by the induction of cortical reactions). The HVD chamber consisted of two stainless-steel plate electrodes separated by 1 cm in a rectangular chamber of 3-mm depth and 6-mm width. Unless indicated otherwise, each aliquot of eggs was exposed to 10 pulses of 500 V each. The pulses had a half-time of 7-10 ms and a current of 14 A. For each determination, the eggs in 0.4 ml of a 1% egg suspension were washed once with 1 ml of OCaSW, once with 1 mi of OCaSW containiig Ca:EGTA (to produce the desired free Caz+ concentration), and then resuspended in 0.34 ml of the same Ca:EGTA buffer in OCaSW. IPS (7 p~) was added to one-half of this egg suspension, and the other half was used as a control without IPS. Both were subjected to an identical HVD treatment. Nomarski optics (magnification, X 100) was used to assay for the percentage of eggs showing cortical reactions, and 100 eggs were counted for each determination.
Permeabilization was demonstrated in initial experiments using the uptake of an impermeant dye, 6-carboxyfluorescein, as an index. Also, eggs subjected to HVD produced cortical reactions if the free Ca2+ in the medium was =lo-' M, but not if the free Ca2+ was S10" M. The permeabilization was transient, since if the eggs were first permeabilized in M free Ca2+, the subsequent addition of 2 mM Ca2+ 5 min later produced no cortical reactions. However, a higher voltage (1500 V, 10 pulses) did produce long duration permeability in media with 510-M free cas, since no corticd reactions were initiated by the HVD, but subsequent addition of 2 mM Ca2+ after 5 min produced 80-90% cortical reactions.
Egg Homogenate Preparation Unfertilized eggs were dejellied, sequentially washed in OCaSW and IM, and resuspended to 3-25% (v/v) in IM. Unless indicated otherwise, all homogenates contained ATP (0.4 mM) and an ATPregenerating system consisting of phosphocreatine (4 mM) and creatine phosphokinase (2 units/ml) to provide energy for the Ca2+ transport. Protease inhibitors (0.5-1 mM phenylmethylsulfonyl fluoride and 2.5 mM benzamidine) were add&, and the eggs were homogenized with a Dounce-type glass tissue homogenizer with a size "A" glass pestle. The homogenates were centrifuged for 1 min (13,000 X g, 4 "C) in a Microfuge (model 235% Fisher Scientific Co.1 to remove large particles (e.g. plasma membrane fragments, nuclei, and yolk platelets), and the supernatant was saved. For the Percoll gradients, 25% homogenates were utilized; all remaining experiments used 3-9% homogenates. Protein content was assayed with the Bio-Rad protein assay (Bio-Rad); 4% homogenates of L. pictus eggs contained 3.7 mg of protein/ml (S.D. = 0.7, n = 15), and 8% homogenates of S. purpuratus eggs contained 4.6 mg/ml (S.D. = 0.3, n = 2). S. purpuratus homogenates showed 250% decreases in both Caz+pumping activity and responsiveness to IPa by 8 h from egg homogenization, whereas L. pictus homo~nates usually showed no loss during this time. Therefore, L. pictus was used for most experiments reported here.
Monitoring Ca2+ Fluxes Ca2+ uptake by and release from Cap+ stores in egg homogenates was monitored by measuring changes in the Ca2+ concentration of the medium. Two methods were used to measure these Cap+ levels: a Ca'+-sensitive minielectrode and the fluorescent Cas+-sensitive molecule, quin-2. Caz+ Minielectrodes-Ca2+ electrodes were prepared using a protocol similar to that of Prentki (19). Pieces of Teflon tubing (1.5 mm, inner diameter, 4 cm long) were dipped into a solution containing a lO-gl aliquot of the premixed Ca" electrode cocktail (Fluka Chemical Corp.) and 60 pl of 12% (wv) polyvinyl chloride dissolved in tetrahydrofuran. The thin membrane formed at the tip of the tubing was allowed to dry, and the tubing was then Hied with a Ca2+ buffer containing lo-? M free Ca2+ (prepared according to Ref. 20) and stored in this buffer until used (212 h). A silver wire was inserted into the Ca2+ electrode b e e r and connected to a Corning pH/ion meter (model 135). The mini-reference electrode used (model MI-401) was purchased from Microelectrodes, Inc. Assays were conducted with 0.5-or 0.6-ml aliquots of homogenate in a Plexiglas temperaturecontrolled chamber (17 "C) and stirred with a magnetic stirring bar.
The Ca2+ calibration solutions were prepared by adding CaKzEGTA and KrEGTA to IM to give 10 mM EGTA and various free Caz+ concentrations. Benzamidine interfered with the Ca2+ electrode and was omitted from all Cap+ electrode assays.
Quin-8-The fluorescence was measured at 339 nm excitation (2nm slit) and 492 nm emission (10-nm sfit) with a fluorescence spectrophotometer (model 650-105, Perkin-Elmer) used in ratio mode. Samples were stirred with a magnetic stirring bar, the temperature was maintained at 17 "C, and 0.7-ml aliquots of homogenate were used. The Caw concentration of the homogenate was calculated from the relationship, where F is the fluorescence measured during an experiment, F d and Fmm are fluorescence values for minimal and maximal Ca2+ concentrations (measured at the end of each experiment by adding EGTA and then saturating Ca"' , respectively, to each homogenate), and Kd is the dissociation constant for quin-2 (21).
Q~n~~a t i o n of C2+ Released from Vesicles After each experiment, the relationship between Ca*+ electrode or quin-2 response and changes in Ca2+ concentration were determined by adding known amounts of Cap+ to an aliquot of the same homogenate (containing the same reqents). Since the quin-2 response to Caz+ was nonlinear (see i n s e t to Fig. 31, a series of Cap+ additions was always made, with the homogenate starting at the same free Cas+ level as in the experiment being calibrated. This calibration procedure allowed an accurate d e t e r~a t i o n of Cap+ release in the presence of reagents that altered the quin-2 response (3,3',4',5-tetrachlorosalicylanide and TMB-8) and in homogenates containing differing vesicle or protein content (e.g. fractions from Percoll gradients). Also, for quin-2 experiments, the vesicle content and/or quin-2 concentration was adjusted so that Ca2+ release never saturated the quin-2 signal. The Cap+ contamination in IP3 and apyrase preparations were determined by a similar procedure, with each reagent being added to quin-2 in distilled water (pH buffered at 7.1 with 20 mM HEPES), and the resultant response was calibrated with known amounts of Cap+. The respective Cap+ contents were 0.06 nmol of Ca"/nmol of IPa and 0.07 nmol of Ca2+/unit of apyrase, and both Ca2+ contents were insignificant for the experiments reported here.
Vesicle Purification by Percoll Density Gradient Centrifugatbn Percoll was diluted into 1.3 X concentrated IM to produce final concentrations of 25% Percoll and normal strength IM. Since ATP interfered with the assay for glucose-6-phosphatase, one gradient was prepared without ATP or quin-2 and was used for the enzyme assays. A second gradient (Containing 0.5 mM ATP, 4 mM phosphocreatine, 2 units/ml creatine phosphokinase, and 10 pM quin-2) was used for the Ca2+-pumping and IPS responsiveness assays. A third gradient (without ATP and quin-2) was layered with density marker beads designed to calibrate Percoll gradients (Sigma). All three gradients contained 2.5 mM benzamidine.
The egg homogenate (25%) was prepared as previously described and was divided into 2 aliquots, with ATP, phosphocreatine, and creatine phosphokinase added to only 1 aliquot. Then 1 ml of each homogenate (*ATP) was layered onto 9 ml of its respective Percoll gradient (*ATP), and the density beads (suspended in 1 ml of IM) were layered onto the third gradient. The gradients were centrifuged for 50 min (25,000 X gav, 10 "C), photographed, and 1-ml fractions were recovered by upward displacement with 80% sucrose.
Caz+-pumping and IP3-induced Ca2+ release were assayed with quin-2 as previously described. Glucose-6-phosphatase was assayed (as an endoplasmic reticulum marker) by measuring the release of inorganic phosphate (Pi) from glucose 6-phosphate according to the procedure of Morre (22). The reaction was stopped after 2 h (at room temperature) by the addition of trichloroacetic acid, and the Pi was quantitated (23). The medium (lacking vesicles) produced some color in this Pi assay; therefore, blanks were prepared for each fraction by adding an aliquot of the same fraction to incubation medium that already contained trichloroacetic acid and would, therefore, prevent any Pi production by the vesicles. Cytochrome e oxidase was assayed (as a mitochondrial marker) following the procedure of Smith (24). Light scattering at 695 nm was measured (in a spectrophotometer) as an indicator of vesicle content.

Eggs Permeabilized by HVD
HVD was used to transiently permeabilize eggs and allow IP3 to enter from the extracellular medium. Fig. 1 shows that eggs permeabilized in seawater containing 7 p~ IP, and lo-' to 5 X M free Ca" produced 60-70% cortical reactions, whereas eggs permeabilized in the same media lacking IP, produced 52% cortical reactions. This suggests that during the permeabilization, IP, entry from the medium induced Ca" release from internal stores and triggered the cortical were permeabilized by HVD as described under "Experimental Procedures." In experiments at 10-8-10" M Ca2+, 1 mM EGTA and sufficient Ca2+ was added to produce the indicated free Ca" levels; the experiments at M Ca" were done with ASW lacking EGTA and containing M total Ca' ". Triangles represent experiments containing 7 PM IPa, and circles represent controls without IP3. Eggs showing either partial or complete elevations of the fertilization envelope were scored as having undergone a cortical reaction. Experiments were done in triplicate (S.E. indicated).
reaction. The slight decrease in cortical reactions as free Ca2+ was lowered from 10-7-10-8 is probably due to EGTA buffering the intracellular Ca" released by IPS. Since HVD allows EGTA to enter along with Ips, the increased free EGTA concentration at lo-' M Ca2+ could be sufficient to partially inhibit the IP,-induced Ca2+ increase.
The cortical reaction in most eggs did not propagate and occurred at localized regions corresponding to the cathode and anode sides of each cell. The percentage of cortical reacted eggs decreased as [IPS] in the medium was lowered. Thus, at 5 X M free Ca2+ the percentage dropped from 60% to 55%, and then to 31% as [IP,] was reduced from 7 to 3.5 p~ and then to 1.8 p~, respectively. The size of the surface area covered by the localized cortical reaction also correlated with the [IP,], with >50% of the egg surface covered at 7 PM IP, and progressively smaller areas with lower [IPS]. The high concentrations of both IPS (7 pM) and Ca" ( M in the absence of IP,) required to initiate cortical reactions by this procedure could be explained by the transient nature of the permeabilization.

Ca" Transport and IP3-induced Rebase in Egg Homogenates
Since the previous experiments with intact eggs provide only indirect evidence for IPB-induced Ca2+ release in sea urchin eggs, the system was simplified by using egg homogenates or vesicles isolated from egg homogenates to directly assay Ca2' uptake by and release from intracellular stores. The Ca2+ copcentration of the medium was assayed as an indicator of Ca" uptake or release; thus with this approach, a decreased Ca2+ concentration in the medium corresponds to Ca" uptake by vesicular components in the homogenate, and, conversely, an increased medium Ca" concentration represents an efflux. Fig. 2 shows that the rate of Ca2+ sequestration by an L. pktus egg homogenate (as monitored by quin-2) is greatly increased by the addition of 0.4 mM ATP. Creatine phosphokinase (2 units/ml) and phosphocreatine (4 mM) were also added as an ATP regeneration system to maintain constant ATP levels. The low initial rate of Ca2+ sequestration prior Ca FIG. 2. ATP requirement for Ca" uptake and retention by Caz+ stores. Quin-2 (10 PM) was used to monitor the extravesicular Ca" concentration in 0.9-ml aliquots of a 3% L. pietus egg homogenate. Three experiments utilizing aliquots of the same initial homogenate are shown, and each aliquot was treated identically (-) until the addition of either IPS (---) or apyrase (. . . -). Following apyrase addition, IP3 was added at either 3 min (-. -) or at 25 min (. . . .).At the urmw marked ATP, 0.4 mM ATP plus 2 units/ml creatine phosphokinase and 4 mM phosphocreatine were added. Other arrows represent the addition of 5 nmol of Ca2+ (Ca), 330 nM ( P a ) , or 2 units/ml apyrase (Apyr). The ordinate indicates the approximate free Ca2+ concentrations in the medium and was calibrated as indicated under "Experimental Procedures." The addition of 5 nmol of Ca2+ to 0.9 ml of homogenate produced a free Caz+ change of -200 nM, which is much less than the 5.6 PM change that would occur if no Ca" buffers were present. However, quin-2, gluconate, and egg components all bind Caz' and contribute to the buffered response observed here.

IPS Releases ea2' from Sea
Urchin Egg Membranes to ATP addition was not due to mitochondria since it was not inhibited by the addition of 2.5 p~ 3,3',4',5-tetrachlorosalicylanide (a mitochondria uncoupler) and is probably due to a low endogenous concentration of ATP present in the initial homogenate. After the homogenate had sequestered Ca2+ to an equilibrium level (-30 nM), an aliquot of CaClz (5 nmol) was added to demonstrate the responsiveness of quin-2. After the added Ca2+ had been sequestered, addition of IP3 (330 nM) produced a rapid release of Ca2+ which lasted for about 3-4 min. A comparison of the quin-2 response to the change induced by the Ca" pulse added earlier shows that IP, induced the release of more than 5 nmol of Ca" from the homogenate. The released Ca2+ was then resequestered. The resequestration process depends critically on the availability of ATP, since very little was observed when ATP was removed from the medium by the addition of apyrase, an ATP-hydrolyzing enzyme, as shown in Fig. 2.
The vesicles also require ATP to retain their sequestered Ca2+ since removal of ATP by apyrase produced a slow but sustained release of Ca". Addition of IP, at various times after the apyrase produced the typically fast release of Ca" so that the sum of the Ca" released by apyrase followed by IP3 is nearly the same as that released in the control experiment by IP3 alone. The respective amounts of Ca2+ released (per mg of protein) were: IP3 alone, -4.1 nmol; apyrase followed by IP, at 3 min, -4.1 nmol; and apyrase followed by IP3 at 23 min, -3.6 nmol. The Ca2+ release was quantitated by adding known amounts of Ca2+ to aliquots of homogenate +apyrase as described under "Experimental Procedures." This experiment indicates that the Ca" release by apyrase is probably from the same store as that released by IP3 and also demonstrates that the Ca2+-release mechanism activated by IP, does not require ATP.
The ATP requirement for retention of sequestered Ca2+ indicates that these vesicles are inherently leaky to CaZ+, and a dynamic relationship between Ca" leakage and sequestration maintains the Ca2+ gradient between the vesicles and extravesicular medium. This suggests the possibility that IPS could release Ca" by inhibiting the Ca2+ pump, although the slow kinetics of Ca" release by apyrase argues against this mechanism.

Steady State Concentration to Which Homogenates Pumped Ca2+
Ca" Electrode-The Ca2+ electrodes were calibrated with Ca2+-EGTA-buffered IM in the presence of homogenates and indicated that the minimum detection limit of the Ca2' electrode is about 100 nM. In four experiments with two preparations of 4% L. pictus homogenates, the steady state free Ca2+ concentration to which the homogenates pumped was found to be below the detection limit of the Ca2+ electrode, thus indicating the vesicular components in these homogenates can pump Ca" to 5100 nM free Ca2+.
Quin-2-Quin-2 has a lower detection limit for Ca" than does the Ca2+ electrodes, and since it also has a faster response time, quin-2 was used for the remaining experiments. The effective K d for quin-2 was determined to be 170 nM in IM at pH 7.1. This Kd was the same in the presence of homogenates and was unaffected by benzamidine (2.5 mM), phenylmethylsulfonyl fluoride (0.5 mM), ATP (0.4-0.5 mM), phosphocreatine (4 mM), or creatine phosphokinase (2 units/ml), which are the concentrations of these reagents routinely added to L. pictus homogenates. This & is somewhat higher than the value of 115 nM reported for quin-2 in a medium designed for mammalian cells (21) and probably reflects the higher ionic strength of the IM used here. The steady state Ca2+ concentration to which 4% L. pictus homogenates pumped was found to range from 20-30 nM in fresh homogenates to 40-50 nM in homogenates assayed 6-24 h later. This free Ca2+ is consistent with the 1100 nM determined by Ca" electrodes.

Characterization of Ca2+ Uptake and Release
The dose-response relationship for IPS is shown in Fig. 3. The Ca2+ released from a 4% L. pictus egg homogenate was maximal at 1 pM and half-maximal between 50 and 60 nM. At saturating doses of IP3, 11 nmol of Ca" was released from 0.7 ml of the 4% homogenate (2.9 nmol of Ca2+/mg of protein). The half-maximal dose is 40-fold less IP3 than the 2 p~ that produced a half-maximal response with HVD-permeabilized eggs (Fig. 1); however, as indicated previously, the difference is probably due to the transient nature of the permeabilization produced by HVD using the parameters selected.
The effects of homogenate pH on Ca" pumping and IP3induced Ca" release were investigated next. A L. pictus egg homogenate (6%) was allowed to pump Ca2+ to a steady state level at pH 7.1, and then the pH of the medium was adjusted to various pH values. The steady state Ca2+ level showed little change (<lo nM) after the pH adjustments. The rate of Ca2+pumping activity at the different pH values was determined by adding 1.25 nmol of Ca2+ and measuring the proportion of this Ca2+ that had been sequestered by 2 min. Finally, after the homogenate had sequestered this added Ca", IP3 (290 nM) was added, and the Ca2+ released was quantified as previously described. The results showed little difference in either Ca2+ pumping or IP, response between pH 7.1 and 7.5; however, both responses are reduced by about 50% at pH 6.7. The respective Ca2+-pumping rates at pH 6.7, 7.1, and 7. Since fertilization increases the intracellular pH of L. pictus eggs from about 6.8 to 7. 3 (25, 26), these results indicate that both Ca2+ pumping and IP, responsiveness may be greater in fertilized eggs.
The Ca2+ release mechanism is highly specific for IP3. Thus when IP, and IPI (5-20 p~ each) were assayed for induction of Ca" release from L. pictus homogenates, 16 pM IP2 produced the same Ca" release as did 60 nM IP3, and 20 pM IP, produced no Ca2+ release. Therefore, the specificity for inducing Caz+ release is 300-fold greater for IP3 than for IPz and >300-fold greater than for IP1. Also, when 43 nM IPS was added 5 min after 20 p~ IP, or 5 p~ IP2, no inhibition of Ca2+ release was observed, therefore, showing that neither reagent inhibits IP3 action.
The Ca2+ blocking agent TMB-8 was found to reversibly block the Ca2+ release induced by IP3. The TMB-8 experiments were monitored with quin-2, and addition of 3 mM TMB-8 to 4% L. pictus homogenates inhibited 95% of the Ca2+ release induced by 290 nM Ips, with half-maximal inhibition being observed at about 1 mM TMB-8. The reversibility of the TMB-8 inhibition was tested by a dilution experiment. Ca2+ release (induced by 290 nM IPS) from a 9% homogenate was totally inhibited by 3 mM TMB-8; however, following a 3-fold dilution into IM lacking TMB-8 (thus producing a final TMB-8 concentration of 1 mM), the IP3 addition produced 0.91 f 0.04 nmol of Ca2+/mg of protein. This was nearly the same as the 1.09 k 0.04 nmol of Ca2+ released from controls diluted from a 9% homogenate not previously exposed to TMB-8 and to which 1 mM TMB-8 was added. Other controls assayed in the absence of TMB-8 released 3.3 0.1 nmol of Ca2+. Experiments were conducted in triplicate, standard deviations are indicated, and TMB-8 was added at least 4 min before IP3 responsiveness was assayed.

Homogenates Are Refractory to a Second Dose of IP,
Following the addition of a saturating dose of IP, (430 nM), egg homogenates did not respond to a second aliquot of IP3 (Fig. 4A). Also, sequential additions of a submaximal dose of IP3 (70 nM) produced decreasing reponses with each additional dose (Fig. 4B). However, if these homogenates were allowed to incubate for 3.2-3.5 h before additional aliquots of IPS were added, the homogenate exposed to submaximal IP3 had completely recovered its responsiveness, and the homogenate initially exposed to higher IP3 had recovered about half of its initial responsiveness. The time course of recovery was determined next. A 6% L. pictus homogenate was treated with 290 nM IP3, and the recovery was monitored by assaying the responsiveness of 0.7-ml aliquots to 290 nM IP3 at l-h intervals. The assay times and Ca2+ released per mg of protein were: initial, -2.7 nmol of Ca2+; 1 h, -0.03 nmol; 2 h, -0.90 nmol; 3 h, -2.1 nmol; and 4 h, -2.7 nmol.
The recovery of responsiveness may be due to the removal of IP3 from the medium (e.g. hydrolysis by endogenous IP3 phosphatase) and subsequent refilling of responsive vesicles 2.51-A 8min with Ca2+. If this is the case, one would expect that washing with IM containing no IP3 should rejuvenate the vesicles. A 4% L. pictus egg homogenate was first exposed to 400 nM IP, and divided into three portions. One portion was centrifuged (180,000 X g, 20 min, 10 "C) to pellet the vesicles and then resuspended into fresh IM containing no IP,. Addition of 290 nM IP3 induced 6.77 nmol of Ca2+ release/mg of protein (S.D. = 0.23, n = 3). On the other hand, an identical dose of IP, added to the control portion (which had not been washed) produced only 0.08 & 0.06 nmol/mg ( n = 3) of Ca2+ release. Similar nonresponsiveness (0.08 & 0.01 nmol of Ca2+/mg) was found for vesicles pelleted but resuspended with the original medium containing IP3 (instead of the fresh IM without IP3). A more accurate comparison is obtained if the response of washed vesicles (6.77 nmol of Ca2+/mg of protein) is corrected for the loss of soluble proteins (57% of the total protein); such a correction yields 2.9 nmol of Ca2+/mg of protein and does not alter the qualitative results.
These results are consistent with the continuous presence of a high concentration of IP, in the medium being the cause of the refractory response. This centrifugation and resuspension experiment also shows that soluble factors from the egg do not mediate the action of IP3, and the apyrase experiment (Fig. 2) provided evidence against a requirement for ATP. Therefore, the mechanism of IPS action must utilize only components inherent to the vesicles and ions present in the medium used.

Purification of Vesicles by Percoll Density Gradient
Centrifugation Percoll density gradient centrifugation of egg homogenates produced two major bands of vesicles as is shown in Fig. 5A. The large pellets at the bottom of each tube (Fig. 5A)  bands of density beads are also indicated by arrows, and their respective densities were 1.017, 1.033, and 1.048. B, vesicle and protein distribution in the gradients f ATP. Protein content (0, expressed as mg/ml) is plotted for the gradient without ATP, and vesicles content determined by light scattering (AG95m) is plotted for both gradients with (A) and without (0) ATP. C, distribution of activities for Ca2+ sequestration (A) and IP3-induced Ca2+ release (0). 200-pl aliquots of each fraction were diluted to 700 pl of IM (containing the same concentrations of ATP, phosphocreatine, creatine phosphokinase, and quin-2 as in the gradient), and each diluted fraction was incubated at least lh h at 17 "C to allow those with Ca2+-sequestering activity to pump Ca2+ to an equilibrium level. Then, 1.25 nmol of Ca2+ (5 pl of 0.25 mM CaC12) was added, and the rate of sequestration was measured (and plotted as nmol of Ca2+/4 min). Next 290 nM IPS was added, and the Ca2+ release was monitored and calibrated by adding known amounts of Ca2+ to a second aliquot of the same diluted fraction (using the protocol illustrated by the inset to Fig. 3). The inset to this figure shows Ca2+ sequestration and IP3 response assays for fraction 5. D, distribution of enzyme activities. Fractions were assayed for glucose-6-phosphatase (0) and cytochrome c oxidase (0) using the protocol described under "Experimental Procedures." found to consist of sedimented Percoll particles and contained no vesicles. When the gradients containing homogenates were compared with the calibrating gradient containing density beads, the densities of the major two vesicle bands were determined to be about 1.023 and 1.040 g/ml, respectively. The upper band in the gradient containing ATP was much sharper than the corresponding band without ATP. Fractions collected from both gradients were analyzed for turbidity at content was also measured for fractions from the gradient without ATP, and three peaks were found (Fig. 5B). Two of these peaks correspond to the vesicle bands. The main protein peak was at the top of the gradient (fractions 1 and 2 ) and most likely represents soluble protein in the homogenate. Fig. 5C shows that the activities for Ca2+ sequestration and IPS-induced Ca2+ release co-purify in a region occupied by the upper vesicle band in the gradient containing ATP. It is likely 695 nm, which was used as a measure of the vesicle content, that the diffused pattern of the upper band in the gradient and the results are shown in Fig. 5B. In the absence of ATP, without ATP is due to the loss of Ca2+ from the vesicles. As vesicles in the upper band were divided equally between was shown earlier, the IPS-responsive vesicles require ATP to fractions 4 and 5, while nearly all of the vesicles were concenretain their Ca2+ (Fig. 2). The Ca2+ loss in the absence of ATP trated in the denser fraction 5 when ATP was present. Protein would, therefore, reduce the vesicle density and result in spreading of the vesicles to a lower density region of the gradient.
Marker enzyme distributions were analyzed only in the gradient without ATP, since ATP interferes with one of the enzyme assays (glucose-6-phosphatase). Fig. 5 0 shows that activity for cytochrome c oxidase (a marker enzyme for mitochondria) was localized in the higher density band, which was completely devoid of ATP-dependent Ca2+ transport and Ips-induced Ca2+ release activity. On the other hand, the activity for glucose-6-phosphatase (a marker enzyme for the endoplasmic reticulum) showed two peaks in the gradient, with the major peak of activity coinciding with the IP3induced Ca2+ release activity. The complete separation of the mitochondrial marker enzyme activity from the IPS-induced Ca2+ release activity provides strong evidence that the latter is nonmitochondrial. The co-purification with glucose-6-phosphatase suggests it is most likely a component of the endoplasmic reticulum network.
When the recovered activities for Ca2+ sequestration and IP3 responsiveness (Fig. 5C) were compared to the total applied activities (assayed as a 4% homogenate but extrapolated to 1 ml of a 25% homogenate), it. was determined that 87 and 65% of the respective applied activities were recovered. Also, Fraction 5 contained 62 and 73% of the recovered sequestration and release activities but only 30% of the vesicles (as assayed by Aegbnrn); therefore, this purification method enriches the active vesicles with little loss of applied activity.

DISCUSSION
This study presents evidence that IPS releases Ca" from intracellular stores in sea urchin eggs. First, HVD was used to transiently permeabilize eggs and introduce 1P3, with the resultant induction of cortical reactions (a well-characterized Ca2+-dependent event, Refs. 1-3) being assayed as an indicator of intracellular Ca" release. Both our HVD results (Fig. 1) and the microinjection study of Whitaker and Irvine (12) show that IP3 introduction into eggs induces cortical reactions, thus providing similar indirect evidence that IP, releases Ca2+ from internal stores in sea urchin eggs.
To directly assay Ca2+ uptake by and release from intracellular stores, egg homogenates and microsomal fractions were utilized. The extravesicular Ca" concentration was monitored by either a Ca" minielectrode or. quin-2, and both assay methods showed increases in medium Ca" levels upon addition of IP3. The observed changes represent Ca2+ release from stores in the homogenate and are not due to Ca2+ contamination in the IP3 (which was 0.06 nmol of Ca2+/nmol of IP,). In a typical experiment such as the response to 30Q nM IP, shown in Fig. 3, <1% of the Ca" change is due to Ca" introduced with IP3. Also no change was observed when IP, was added to vesicles refractory to a second IP, dosage (Fig.  4A), and addition of IP, at concentrations above saturating levels (>1 p~) produced no additional Ca2+ change (Fig. 3).
The response of L. pietus homogenates to IP, was halfmaximal at 50-60 nM and maximal at 1 p M (Fig. 3). At 30 nM IP3, 20% maximal Ca2+ release occurred; this IP, concentration is close to the 10 nM IPS estimated to be the minimum concentration that would induce cortical reactions when injected into eggs (12). When compared to other cell types, the sea urchin egg is more sensitive to IP, then are most mammalian cells, which show half-maximal responses to IP, in the range of 0.1-2 pM (5-10). An exception is a hamster insulinoma cell line reported to have a half-maximal response at 25 nM IP, when assayed as saponin-permeabilized cells; however, the half-maximal response was at 200 nM IP, when assayed as partially purified microsomes (8). Since this result clearly shows that the sensitivity to IP, can be greatly altered by cell disruption, the physiologically active concentration of IP, in intact cells may be lower than the values reported for most permeabilized cells and cell homogenates.
L. pictus egg homogenates showed a much greater sensitivity to IPS than to its analogs IPz and IP1. 300-fold more IP2 (55 nM IP, versus 16 pM IP2) was required to induce halfmaximal Ca2+ release, and 20 p~ IP, induced no Ca2+ release. Similar specificities for IP, over IP, and IP, have been reported for injected L. pictus eggs (12) and for permeabilized mammalian cells and cell homogenates (6, 7, 10).
IP3-induced Ca" release was inhibited by TMB-8, a Ca2+ antagonist reported to block Ca2+ release from the sarcoplasmic reticulum of both smooth and striated muscle cells (27). Inhibition was 95% at 3 mM TMB-8 and was reversed by dilution into media lacking TMB-8. In intact L. pictus eggs, TMB-8 was reported to inhibit both the cortical reaction and the release of intracellular Ca2+ at fertilization (28). It thus appears that TMBS is either blocking the Ca2+-release mechanism or is interfering with the interactions between IP, and its receptor.
Ca2+ sequestration by egg homogenates was ATP dependent (Fig. 2), and the steady state level to which egg homogenates pumped Ca2+ was assayed by both Ca" electrodes and quin-2. L. pictus homogenates pumped Ca2+ to the detection limit of the Ca2+ electrodes (-100 nM), and when assayed with quin-2, were found to pump Ca2+ to 20-30 nM in freshly prepared homogenates. This is consistent with the 4 0 0 nM free intracellular Ca2+ reported for intact unfertilized Arbacia punctulata eggs assayed with the Ca2+-sensitive photo protein aequorin (29) but considerably below the 144 nM free Caz+ reported for intact unfertilized L. pictus eggs assayed with fura 2 (30). The reason for this difference is not known.
ATP-dependent Ca2+ sequestration in sea urchin eggs and embryos has also been demonstrated in studies utilizing 45Ca2+ (14, 31, 32). Silver et al. (31) demonstrated Ca2+ uptake by vesicles associated with the mitotic apparatus of fertilized eggs. Suprynowicz et al. (14) demonstrated Ca2+ uptake by fertilized eggs permeabilized by HVD and by homogenates of both fertilized and unfertilized eggs (32); in each case the Ca2+-sequestering threshold was estimated to be about 100 nM. The homogenates were also reported to sequester Ca2+ 5 times faster at pH 7.4 than 6.8, which is similar to the 3.3fold increase we measured between pH 6.7 and 7.5.
When egg homogenates were fractionated by Percoll density gradient centrifugation, the activities for Ca2+ sequestration, IP3-induced Ca2+ release, and glucose-6-phosphatase (an enzyme associated with the endoplasmic reticulum) all copurified (Fig. 5) and showed clear separation from cytochrome c oxidase (an enzyme asgociated with the mitochondria). Therefore, the Ca2+ store that is responsive to IP, is most likely a component of the endoplasmic reticulum network. This result is supported by a study with A. punctulata eggs loaded with aequorin and whose organelles were stratified by centrifugation. When fertilized, these eggs showed Ca" to be released from a region enriched in endoplasmic reticulum but depleted for mitochondria (29). Studies with several mammalian cell types have also shown that IP3-responsive vesicles co-purify with endoplasmic reticulum-associated enzymes (5, 7,8,9).
These purification studies and the experiment where L. pictus microsomes were centrifuged and resuspended also show that no cytoplasmic factors are required for IP, action. The mechanism of IPS action must utilize only components inherent to the vesicles and ions present in the medium used.
When L. pictus egg homogenates were exposed to saturating doses of IP3 (e.g. 430 nM in Fig. 4), Ca" was released for 1-2 min, whereupon resequestration began. However, the homogenates were refractory (desensitized) to subsequent IP, addition, even after the Ca" concentration had returned to its initial level. Similar desensitization to IP3 has been reported for three insulinoma cell lines (6)(7)(8). When the desensitized egg homogenates were centrifuged and washed with fresh medium lacking IP3, the desensitization was reversed, indicating that the continuous presence of IP3 in the medium may be the cause. It is possible that there is more than one type of Caz+-sequestering vesicle in the homogenate. In the presence of IP,, the IP3-responsive vesicles would not resequester Ca2+, and the released Ca2+ would be transported into another type of vesicle which is not responsive to IP,. Reversal of desensitization would, therefore, require the removal of IP, and subsequent redistribution of Ca" back into responsive vesicles. Such a model would explain both the time-dependent recovery we observed and the centrifugation and washing experiment (see "Results"). This study provides the first direct evidence that IP, induces Ca2+ release from intracellular stores in sea urchin eggs and, therefore, supports the hypothesis that IPS mediates the sperm-induced Ca2+ increase that activates development. Determining the mechanism of IP, action is, therefore, critical to understanding how these cells respond to sperm. This system should be very useful for future studies into the mechanism of IP3 action, since the IP, responsiveness of microsomes is very stable, and we have demonstrated in this study the feasibility of purifying the active components using Percoll density gradient centrifugation.