Inositol 1,4,5-trisphosphate and the endoplasmic reticulum Ca2+ cycle of a rat insulinoma cell line.

Regulation of endoplasmic reticulum (ER) Ca2+ cycling by inositol 1,4,5-trisphosphate (IP3) was studied in saponin-permeabilized RINm5F insulinoma cells. Cells were incubated with mitochondrial inhibitors, and medium Ca2+ concentration established by nonmitochondrial pool(s) (presumably the ER) was monitored with a Ca2+ electrode. IP3 degradation accounted for the transience of the Ca2+ response induced by pulse additions of the molecule. To compensate for degradation, IP3 was infused into the medium. This resulted in elevation of [Ca2+] from about 0.2 microM to a new steady state between 0.3 and 1.0 microM, depending on both the rate of IP3 infusion and the ER Ca2+ content. The elevated steady state represented a bidirectional buffering of [Ca2+] by the ER, as slight displacements in [Ca2+], by small aliquots of Ca2+ or the Ca2+ chelator quin 2, resulted in net uptake or efflux of Ca2+ to restore the previous steady state. When IP3 infusion was stopped, [Ca2+] returned to its original low level. Ninety per cent of the Ca2+ accumulated by the ER was released by IP3 when the total Ca2+ content did not exceed 15 nmol/mg of cell protein. Above this high Ca2+ content, Ca2+ was accumulated in an IP3-insensitive, A23187-releasable pool. The maximal amount of Ca2+ that could be released from the ER by IP3 was 13 nmol/mg of cell protein. The data support the concept that in the physiological range of Ca2+ contents, almost all the ER is an IP3-sensitive Ca2+ store that is capable of finely regulating [Ca2+] through independent influx (Ca2+-ATPase) and efflux (IP3-modulated component) pathways of Ca2+ transport. IP3 may continuously modulate Ca2+ cycling across the ER and play an important role in determining the ER Ca2+ content and in regulating cytosolic Ca2+ under both stimulated and possibly basal conditions.

IPS degradation accounted for the transience of the Ca2+ response induced by pulse additions of the molecule. To compensate for degradation, IPS was infused into the medium. This resulted in elevation of [Ca"'] from about 0.2 PM to a new steady state between 0.3 and 1.0 ~L M , depending on both the rate of IP3 infusion and the ER Ca2+ content. The elevated steady state represented a bidirectional buffering of [Ca"'] by the ER, as slight displacements in [Ca"], by small aliquots of Caz+ or the Ca2+ chelator quin 2, resulted in net uptake or efflux of Ca2+ to restore the previous steady state. When IP3 infusion was stopped, [Ca"] returned to its original low level.
Ninety per cent of the Caz+ accumulated by the ER was released by IPS when the total Ca2+ content did not exceed 15 nmollmg of cell protein. Above this high Ca2+ content, Ca2+ was accumulated in an IPS-insensitive, A23187-releasable pool. The maximal amount of Ca2+ that could be released from the ER by IPS was 13 nmol/mg of cell protein.
The data support the concept that in the physiological range of Ca2+ contents, almost all the ER is an IP3sensitive Ca2+ store that is capable of finely regulating [Ca"] through independent influx (Ca2+-ATPase) and efflux (IPS-modulated component) pathways of Ca2+ transport. IPS may continuously modulate Ca2+ cycling across the ER and play an important role in determining the ER Ca2+ content and in regulating cytosolic Ca2+ under both stimulated and possibly basal conditions.
In pancreatic / 3 cells, an elevation in the cytosolic free Ca2+ concentration is thought to play a central role in the induction of insulin release by physiological secretagogues such as glucose or acetylcholine (1,2). Both external Ca2+ and Ca2+ mobilized from intracellular stores appear to contribute to this elevation in cytosolic Ca2+ (1,2). It is presumed, based on extensive evidence in several cell types, that IP3,1 a newly discovered second messenger (3-9), mobilizes internal Ca'+ in insulin-secreting cells in response to the nonfuel secretagogue carbamylcholine (2). Whether IP3 is involved in mediating the elevation in cytosolic Ca2+ in response to fuel secretagogues such as glucose is presently unknown. In a variety of cell types, IP3 has been documented to transiently mobilize Ca'+ from a nonmitochondrial pool (4)(5)(6)(7), identified as the endoplasmic reticulum (7, 10). The reason for the transient nature of the Ca2+ response (Ca'+ release followed by Ca'+ resequestration) is presently unclear, although several possibilities including IP, degradation (5,11), IP3-insensitive compartments (7, 11, E ) , or desensitization (7) have been proposed. Considerable differences in the pattern of IP3-induced Ca2+ responses or IP3 degradation rates have been documented in different cell types or preparations. Thus, permeabilized insulinoma cells (11) are sensitive to a second challenge of IP3, whereas permeabilized human neutrophils (13) and insulinoma (7) or liver (12) microsomes are markedly less responsive to a second pulse addition of the molecule. In addition, microsomal preparations from both insulinoma (7) and liver cells (12) are poorly responsive to IP3 in comparison to permeabilized cells (4, 5 , 11,13). Finally, the rate of Ca'+ reuptake following IP3 addition appears to correlate roughly with IP, degradation in permeabilized hepatocytes (5) and insulinoma cells ( l l ) , whereas in both permeabilized neutrophils' and insulinoma microsomes (7), Ca2+ reuptake occurs rapidly despite very slow degradation of IP3.
This study had three main objectives. First, to understand the reason for the transient nature of the IP3-induced Ca2+ response; second, to quantitate in insulinoma RINm5F cells the IP3-sensitive and -insensitive nonmitochondrial Ca'+ compartments; and third, to evaluate the role of IPS in regulating Ca2+ cycling across the endoplasmic reticulum.

EXPERIMENTAL PROCEDURES
RINm5F cells were maintained in culture as described previously (14,15). Cells (8 X 10') were detached from culture flasks using EDTA and trypsin (15) and were subsequently washed three times in a Ca2'-and MP-free Hanks' basal salt solution (4 "C; pH 7.2) by sedimenting them a t 100 x g for 8 min. The final cell pellet was resuspended in 0.3 ml of the same medium to give a concentration of about 2 X 10' cells/ml and kept on ice until use. Incubations were carried out at 30 "C, pH 7.0, in 0.2 ml of a medium containing KC1 (110 mM), NaCl(10 mM), KH2PO4 (2 mM), MgC1, (1 mM), Hepes (25 mM), oligomycin (1 pg/ml), antimycin A (0.

9185
Inositol Trisphosphate and Cellular Ca2+ Homeostasis and an ATP-regenerating system consisting of phosphocreatine (5 mM) and creatine kinase (20 units/ml). Cells were added at a final concentration of 5 X lo6 cells/ml. Cell permeabilization was initiated by adding saponin at a final concentration of50 pg/ml. This concentration of saponin corresponded to the minimum amount of detergent necessary to optimally permeabilize the cells, i.e. the fastest Ca2+sequestering activity by intracellular structures and the optimal IP3induced Ca2+ response. The preparation and calibration of the Caz+specific minielectrodes have been previously described (16). None of the compounds tested interfered with the Ca2+ electrode. IP3 was infused into the electrode chamber using a 25-pl Hamilton syringe and a small length (15 cm) of polyvinyl chloride tubing (0.5 mm of internal diameter). The rate of IPS infusion was varied as described in the figures using a multispeed transmission infusion pump (Harvard Apparatus Co., Dover, MA, model 600-910/920) connected to the syringe. The traces shown in the figures are taken from representative experiments which have been repeated at least three times. IP3 was produced by the alkaline hydrolysis of ox brain phospha-tidylinositol4,5-bisphosphate and purified by preparative paper chromatography (17).

Pulse Additions of IP3 and the Ambient [Ca2+J
Established by Nonmitochondrial Compartments-The purpose of this investigation was to gain insight into the role of IP3 in the regulation of Ca2+ cycling across the endoplasmic reticulum and to quantitate the IP3-sensitive and -insensitive nonmitochondrial Ca2+ pools, using permeabilized insulinoma RINm5F cells as a model. Thus, in all the experiments reported, cells were incubated in the absence of mitochondrial substrates and in the presence of the mitochondrial poisons oligomycin and antimycin A to ensure that only nonmitochondrial Ca2+ transporting activity was studied. It has been shown previously that the nonmitochondrial Ca2+ pool of permeabilized RINm5F insulinoma cells displayed similar characteristics to t.hose observed with endoplasmic reticulumenriched fractions of insulinoma cells (11,18,19). Thus, the ATP-dependent, vanadate-inhibitable, antimycin-and oligoxycin-insensitive nonmitochondrial Ca2+ transporting activity is most probably attributable to the endoplasmic reticulum. It will henceforth be referred to as " E R e a 2 + pool.
Ca2+ sequestration by RINm5F cells incubated in the presence of a low concentration of saponin (for plasma membrane permeabilization) started within 1-2 min after addition of cells to the medium and resulted in a decrease in ambient [Ca'+] (Fig. 1A). Within 15 min, the ER compartment lowered ambient [ea'+] from 1.25 FM to about 0.25 p~ (from -log [ea'+] = 5.9 to -log [Ca"'] = 6.6) (Fig. 1). A pulse addition of IP3 then promoted a rapid and dramatic ea2+ release from the ER. Thus, ambient [Ca'+] increased within 30 s from 0.25 PM to a peak value of 1.0 p~. The released Ca2+ was then slowly resequestered into the store (Fig. lA). The response to subsequent IP3 additions was barely attenuated, and each ea2+ release was followed by ea2+ reuptake into the store. When cells were incubated in the presence of mitochondrial substrates without mitochondrial inhibitors, the Ca2+ response was much less pronounced (about 60% reduced), suggesting that mitochondria dampen the ea2+ response by taking up a fraction of the Ca2+ released from the ER.3 Fig. lA further shows that most of the eaZ+ accumulated by the ER was rapidly released by a maximal concentration of IPS despite the presence of an operative ea2+ uptake system. When successive IP3 additions were made at short time intervals, [Ca"] remained markedly elevated and the small amount of ea2+ taken up between pulses was released by the subsequent IP3 addition (Fig. 1B). This observation did not favor the concept that either desensitization to IP3 or IP3-insensitive compart-M. Prentki, unpublished observation. ments accounted for the transient nature of the IP3-induced ea2+ response. This led us to postulate that if IP3 degradation was responsible for Ca2+ reuptake, ambient [ea'+] might remain elevated provided that IP3 was infused into the medium rather than added in pulses.
IP3 Infusion Studits- Fig. 2 shows that when IP3 was continuously infused into the medium containing the permeabilized cells, Caz+ was progressively released from the ER until a new [Ca"'] steady state was reached. The level of the steady state maintained by the ER was dependent on the rate of IP3 infusion. Thus, instead of eaz+ reuptake following IP3induced ea2+ release, medium [ea2+] remained elevated as long as the infusion continued. This process was fully reversible since, when the infusion of IP3 was stopped, ea2+ resequestration occurred and eventually low medium [Ca'+] levels were restored. In separate experiments, we observed that a high [Ca2+] steady state could be maintained for more than 15 min provided that the medium was continuously infused with IPS (data not shown). Control experiments carried out in the absence of cells indicated that these observations were During the time intervals noted, IP3 was infused into the medium in the amounts indicated (nanomoles of IP3/ml of incubation medium/ min) from a 0.1 mM stock solution of Ips. not attributable to Ca2+ contamination of the compound.
The new Caz+ steady state during IP3 infusion presumably reflected a balance between Ca2+ influx (via the CaZ+-ATPase) and a separate IP3-sensitive Ca2+ efflux component (13) (Fig. 3A). This indicates that Ca2+ cycling across the ER occurred at steady state and that the steady state reflects the balance between the influx and efflux components of Ca2+ transport. It should be noted that the first quin 2 addition lowered [Ca'+] less than subsequent additions (Fig. 3). This observation was made consistently and was attributed to some nonspecific quin 2-binding sites in the preparation. It was of interest to determine whether bidirectional buffering of ambient [Ca'+] occurs without IPS infusion. In order to obtain a Ca2+ steady state in the absence of IP3 infusion and in the range of Ca2+ concentrations (above 0.2 PM Ca2+) where the electrode is optimally sensitive, cells were incubated in the absence of an ATP-regenerating system (Fig. 3B). Indeed, it has been shown previously that in the presence of an ATPregenerating system, endoplasmic reticulum vesicles are capable of decreasing ambient [Ca"] to lower values than in the absence of ATP-regenerating system, due to the fact that ADP markedly influences the extramicrosomal Ca2+ steady state (18). As shown in Fig. 3B, the ER was also capable of similar bidirectional buffering of [Ca"] without IPS infusion, although it occurred a t lower [Ca'+] levels. This suggests that IPS may not be absolutely necessary for Ca2+ cycling to occur across the ER and that other possible route(s) of Ca2+ efflux (such as pump reversal (18,20)) may operate. It is also possible that the bidirectional buffering of ambient [Ca'+] observed in the absence of IP3 infusion may occur due to some IPS production by the cells during incubation. Although the ER buffered ambient [Ca"] in a bidirectional manner, the previous [Ca'+] steady state was not "exactly" restored after the small additions of Ca2+ or quin 2. Ambient [Ca"] remained slightly elevated after a Ca2+ addition and slightly decreased after quin 2 addition (Fig. 3). Thus, we investigated whether 6.6 1 IP, infusion (0.12nmol/ml/min 1 6.7l the Ca2+ steady state obtained during IP3 infusion depended on the ER Ca2+ content. As shown in Fig. 4, both the Ca2+ level reached before and during IP, infusion depended markedly on ER Ca2+ content. This was demonstrated by allowing cells to accumulate Ca2+ from various starting ambient Ca2+ concentrations and finding that the higher the Caz+ content of the ER, the higher the Ips-induced Ca2+ steady state (Fig. 4).
IP3-sensitive and -insensitive Nonmitochondrial Compartments of RINm5F Cells-We carried out experiments to better characterize and quantify the nonmitochondrial pools of RINm5F cells. Cells were allowed to accumulate Ca2+ from the medium until a lower ambient [Ca"] was reached (Fig.  5A). At this point vanadate (1 mM) was added. Vanadate was used a t a maximal concentration to completely block ATPdependent nonmitochondrial Ca2+ influx (not shown). This resulted in a progressive increase in [Ca"] until a steady state was obtained. A further addition of vanadate had no effect, indicating that Ca2+ contamination did not account for the response. Subsequent addition of IP3 rapidly released the remaining Ca2+ previously accumulated by the ER. As seen in Fig. 5A, vanadate released a small fraction of the Ca2+ that had been accumulated until a new steady state was reached, but was not capable of releasing all the accumulated Ca2+. One interpretation of this effect of vanadate is that only a small fraction of Ca2+ can be released from the store through indicates that the ER of permeabilized cells is not "leaky" to Ca2+.
The observation that IPS releases Ca2+ after total blockade of Ca2+ influx with vanadate suggests, as observed in permeabilized human neutrophils (13), that IP3 acts by stimulating a n independent efflux component of the ER and not by inhibiting the influx component of Ca2+ transport. The presence of an additional nonmitochondrial compartment was detected since the Ca2+ ionophore A23187 released further Ca2+ from the cells (Fig. 5A). Fig. 5B shows that the Ca2+ that could be released by vanadate alone (Fig. 5A) was also IP3releasable. Indeed, the final amount of Ca2+ rapidly released by IPS (in Fig. 5B) was the same as that obtained in Fig. 5A when IP3 was added after a plateau Ca2+ level had already been obtained with vanadate. Thus, in the experiments reported in Figs. 6 and 7, the Ips-releasable Ca2+ is the amount of Ca2+ rapidly released from the cells after the simultaneous addition of vanadate plus IP3. Vanadate was added t o eliminate the problem of Ca2+ resequestration by the ER during IPS-induced Ca2+ release.
The amount of Ca2+ released by IP3 was measured as a function of ATP-dependent Ca2+ accumulation by the ER (Fig. 6). It was observed that the IP3-releasable Ca2+ was linearly related to the amount of Ca2+ taken up by the ER. Fig. 6 further shows that about 90% of the accumulated Ca2+ was Ips-releasable when cells were allowed to accumulate relatively small amounts of Ca2+ (less than 15 nmol/mg of cell protein). In another series of experiments, the amount of Ca2+ released by IP3 or by the ionophore A23187 (after IP3) was evaluated as a function of total releasable Ca2+ (IPS + A23187) (Fig. 7). We observed that the IP3-releasable Ca2+ was linearly related to the increase in total cell Ca" content until the total releasable Ca2+ was about 16 nmol/mg of cell protein. Above this point, the additional Ca2+ taken up was accumulated in an IP3-insensitive, A23187-sensitive compartment (Fig. 7). Thus, the maximal amount of Ca2+ that could and the Cas+ accumulated into the nonmitochondrial compartment. Cells ( 1 0 ' ) were incubated as described in the legend to Fig. 1 and with an experimental design similar to that shown in Fig.  5A. The points shown in the figure were obtained from separate traces with three different cell preparations. In order to have various amounts of Ca2+ accumulated in the nonmitochondrial pool, cells were incubated for various times before the addition of vanadate plus IPJ and at various starting ambient Ca2+ concentrations (from 1 to 6 p~ Ca"). The ATP-dependent Ca2+ uptake was calculated from the lowering of ambient [CaZ+] following the addition of ATP to the medium (not including the small immediate chelation of medium Ca2+ by ATP) until vanadate (1 mM) and Ips (IO p~) were simultaneously added to the medium (see also Fig. 5A). The Ips-releasable Ca2+ was calculated from the increase in ambient [Ca2+] following the simultaneous addition of vanadate plus IPS until a plateau was reached after about 3 min (see also Fig. 5B). Both uptake and release of Ca2+ were calculated using separate Ca2+ calibration curves similar to the one shown in Fig. 5C. The amount of CaZ+ that was released by IP3 plus vanadate before cells were allowed to take up Ca2+ was 4.1 k 0.4 nmol of Ca2+/mg of cell protein (mean k S.E. of three separate experiments). This "endogenous" IPrreleasable Ca" was subtracted from all the experimental values to give the points shown in the figure. ment(s) remained constant at about 3 nmol/mg of cell protein over quite a wide range of Ca2+ contents. The A23187-releasable compartment(s) increased only after the total releasable Ca2+ exceeded about 16 nmol/mg of cell protein (Fig. 7).

DISCUSSION
The results demonstrate that the transient nature of the IP3-induced Ca2+ response that has been observed in a variety of cell types (4-8) is due to degradation of IP, in RINm5F insulinoma cells, not to Ca2+ reaccumulation by an IP3-insensitive nonmitochondrial compartment or desensitization to the molecule. We ruled out desensitization by showing that high Ca2+ levels could be maintained by the ER if several pulse additions of IP, were made in rapid succession. Furthermore, if the molecule was infused, a Ca2+ steady state was maintained by the ER and Ca2+ reuptake occurred only when infusion was stopped. The possibility that Ca2+ reuptake following IP3 addition was due to an IP3-insensitive ER compartment was discarded by showing that at relatively low Ca2+ contents, almost all (about 90%) of the ER pool of RINm5F cells was IP3-releasable. It should be added that the conclusion that the IPS-induced Ca2+ response is transient due to IP3 degradation is in accord with previous studies carried out in insulinoma RINm5F cells (11) and hepatocytes (5) showing that [32P]IP3 degradation correlated approximately with the time course of Ca2+ reuptake. Thus, assuming a similar mode of action of IPS in other cell types or preparations, it is likely that the lack of a "second IP3-induced Ca2+ response" observed in permeabilized human neutrophils (13) or rat insulinoma (7) and liver (12) microsomes is due to an IP,-insensitive nonmitochondrial compartment present in these preparations.
It has been suggested that within the intact cell only a part of the ER is sensitive to IP3 (8) and that the molecule releases only a small fraction of the ER Ca2+ content from a specialized region of the ER (21). The view that the ER possesses IP3sensitive and -insensitive regions which are noncommunicating Ca2+ compartments or that only a small fraction of the ER Ca2+ content is IP3-sensitive does not seem to apply to insulinoma cells. Indeed, three separate arguments contradict this view. First, a pulse addition of a maximal concentration of IP3 released almost all the Ca2+ accumulated by the ER, even in the presence of an operative Ca2+ influx component (Fig. 1). It is of interest to note that a similar experiment performed on permeabilized neutrophils (13) or hepatocytes (21) released only 25-50% of the Ca2+ accumulated into the nonmitochondrial compartment. Second, the maintenance of a Ca2+ steady state by the ER during IPS infusion (Figs. 2-4) suggests that it is unlikely that an IPS-insensitive ER compartment is of major importance to Ca2+ buffering in RINm5F cells. Indeed, were a nonmitochondrial IPS-insensitive compartment important, the Ca2+ released by the IP,-sensitive pool could be resequestered by the IP3-insensitive ER, and consequently, the Caz+ response would be transient even during IPS infusion. Third, we have shown (Figs. 6 and 7) that at reasonable Ca2+ contents (less than 15 nmol of Ca2+/mg cell protein) almost all the Ca2+ accumulated by the ER was released by IP3; for comparison, most cells have a total Ca2+ content of about 10 nmol/mg of protein (22). When the total nonmitochondrial cellular Ca2+ content exceeded 16 nmol/mg of cell protein, Ca2+ was accumulated exclusively in an IPSinsensitive compartment. The nature of this IP3-insensitive nonmitochondrial compartment which sequesters Ca2+ at high Ca2+ loadings is presently unknown. It is possible that at high and unphysiological Ca2+ contents, an IPS-insensitive ER is formed due to a deleterious action of such large amounts of Ca2+. Thus, the ER might be fragmented into IP3-sensitive and -insensitive regions. Alternatively, at high Ca2+ concentrations and following saturation of the ER, Ca2+ might be accumulated in other structures (such as Golgi elements, lysosomes, coated vesicles, etc.). A third possibility is that the releasing mechanism induced by IP3 is inhibited at high Ca2+ concentrations (23). Further studies are required to evaluate these possibilities. It can be supposed that the IP3-insensitive, A23187-releasable pool(s) which remain constant (at about 3 nmol Ca2+/mg protein) over quite a wide range of Ca2+ contents consists of nonmitochondrial compartments such as secretory granules, lysosomes, and coated vesicles. Indeed, insulin secretory granules are known to contain large amounts of A23187-releasable Ca2+ (18), despite the fact that they do not appear to play a role in short-term regulation of cytosolic Ca2+ (18). Hence, we favor the concept that in intact insulinoma cells containing reasonable amounts of Ca2+, the ER is primarily a single Caz+ compartment with all regions communicating in such a way that virtually all the ER Caz+ content is IPS-releasable. This does not exclude the possibility that some regions of the ER contain more IP, "receptors" than others or even that some regions are devoid of IPS receptors. Indeed, such possibilities could explain the heterogeneity of microsomal vesicles generated by the fragmentation of the ER when cells are homogenized (8). The view that the ER is a single IP3-sensitive Ca2+ compartment is compatible with morphological observations carried out in a variety of tissues showing that the ER membrane is continuous throughout the cell and that it encloses a single lumen (24, 25).
Our finding that elevated Ca2+ steady states can be maintained by the ER when IP, is constantly infused into the permeabilized cells is of particular interest. We demonstrated that bidirectional buffering of ambient [CaZ+] occurred at steady state, since the ER tended to restore the previous Ca2+ level when small additions of either Ca2+ or the Ca2+ chelator quin 2 were added to the system. However, this Ca2+ steady state established by the ER is not strictly a "set point" as it is for the "extramitochondrial Ca2+ steady state" (16,26,27).
Thus, unlike the mitochondria which are known to maintain an extramitochondrial Ca2+ steady state that is independent of the Ca2+ load over quite a wide range (16,26,27), the ''extra ER Ca2+ steady state" was markedly dependent on the Ca'+ load. The results suggest that when a constant Ip3 level is achieved during IP, infusion, the ER maintains a Ca" steady state at which Ca' ' influx via the Ca"-ATPase equals Ca2+ efflux through an independent IP3-regulated component. Extended to the intact cell, such Ca'+ cycling through independent influx and efflux pathways would allow very precise regulation of cytosolic [Ca"] by the ER as well as a means of finely regulating the ER Ca'+ content. At least two different ways of controlling Ca2+ cycling across the ER can occur. First and most important is regulation of the efflux component by the cytosolic IPS concentration. Second, is the modulation of the influx component by the cytosolic phosphorylation potential ([ATP]/[ADP][Pi]) (2,18), which is the driving force of the Ca2+-ATPase. The second mode of control is likely to be of less importance in most cells since the phosphorylation potential of cells is in general remarkably stable. Under pathophysiological conditions where the phosphorylation potential is lowered and consequently Ca'+ influx is decreased, this could be a means by which Ca'+ escapes the ER. In pancreatic islets, however, such modulation of the influx component might be of physiological significance (18). Indeed, it can be hypothesized that when @ cell glucose phosphorylation by glucokinase (28) is acutely stimulated by a stepwise elevation in glucose concentration, the pancreatic islet phosphorylation potential might be transiently lowered, leading to a transient mobilization of Ca2+ from the ER. This would be a very simple mechanism for linking metabolic changes induced by glucose to the early Ca'+ mobilization (2). However, this does not preclude the possibility that fuel secretagogues also mobilize Ca2+ from the ER through enhanced IP, production.
The exact concentrations of IP, that occur in cells under basal or stimulated conditions are a t present unknown. Rough estimates based on isotopic methods give values in the micromolar range (8,21,29) which are also the levels active in permeabilized cells (8). It should be added that in some tissues both the cytosolic Ca'+ concentration (21,30,31) and the IP3 level (3, 21, 29, 32) remain significantly elevated over basal values as long as a high concentration of the agonist is present. Thus, it can be hypothesized that IP3 may not be a molecule that only transiently stimulates Ca2+ efflux from the ER. Our IP3 infusion experiments suggest that the IP3-sensitive independent efflux component of the ER could be continuously regulated by this molecule. Consequently, the size of the ER Ca'+ pool is likely to be continuously determined by the concentration of IP, in the cytosol. Studies carried out in several tissues have suggeted that the ER, together with the plasma membrane, may be involved in establishing the cytosolic free Ca" concentration (18, 21, 22, 33). Hence, when extended to the in vivo situation, our data suggest that IPS by finely modulating the rate of Ca'+ cycling across the ER may play a major role in regulating the level of cytosolic Ca'+ under both basal and stimulated conditions.