Effect of Adenophostin A on Ca2+ Entry and Calcium Release-activated Calcium Current (I crac) in Rat Basophilic Leukemia Cells*

In most non-excitable cells, calcium influx is signaled by depletion of intracellular calcium stores, a process known as capacitative calcium entry. Adenophostin A, a potent activator of the inositol 1,4,5-trisphosphate receptor, has been reported to activate Ca2+ entry in Xenopus oocytes to a greater extent than expected on the basis of its ability to release calcium stores. In this study, we compared the abilities of adenophostin A and inositol 2,4,5-trisphosphate ((2,4,5)IP3) to release Ca2+ from intracellular stores, to activate Ca2+ entry, and to activate calcium release-activated calcium current (I crac) in rat basophilic leukemia cells. Under conditions of low intracellular Ca2+ buffering (0.1 mm BAPTA), adenophostin A-induced Ca2+ release and activation of I crac could be monitored simultaneously. However, other reagents that would be expected to deplete Ca2+ stores ((2,4,5)IP3, 3-fluoro-inositol 1,4,5-trisphosphate, thapsigargin, and ionomycin) were unable to activate I crac under this low Ca2+buffering condition. Adenophostin A activatedI crac after a significant delay, longer than the delay for Ca2+ release. Thus, adenophostin A activatesI crac as a consequence of release of intracellular Ca2+, rather than directly acting on store-operated channels. The unique ability of adenophostin A to activate I crac under conditions of low intracellular Ca2+ buffering suggests an additional site of action, perhaps in preventing or reducing rapid Ca2+-dependent inactivation of store-operated Ca2+ channels.

Stimulation of G-protein-coupled receptors and tyrosine kinase receptors activates phospholipase C, which generates inositol 1,4,5-trisphosphate ((1,4,5)IP 3 ) 1 and diacylglycerol. (1,4,5)IP 3 binds to its receptor on the membrane of the endoplasmic reticulum and releases Ca 2ϩ stored therein. The depletion of intracellular Ca 2ϩ stores then signals the opening of plasma membrane Ca 2ϩ channels, a process known as capacitative calcium entry (1,2). The signaling mechanism for the activation of capacitative calcium entry is not understood. A direct coupling model suggests that store depletion involves a functional coupling between (1,4,5)IP 3 receptors on the endo-plasmic reticulum membrane and Ca 2ϩ entry channels on the plasma membrane (3,4). Other models suggest that store depletion generates one or more second messengers which in turn activate Ca 2ϩ entry channels (5).
Adenophostin A, a compound isolated from the culture broth of the Penicillium brevicompactum, is the most potent known agonist for the (1,4,5)IP 3 receptor. Its affinity for the (1,4,5)IP 3 receptor is about 100-fold greater than that of (1,4,5)IP 3 (6). Recently, two reports (7,8) have shown that injection of adenophostin A into Xenopus oocytes produced a greater activation of Ca 2ϩ entry (based on the activity of Ca 2ϩ -dependent Cl Ϫ channels) than would be expected solely as a result of release of intracellular Ca 2ϩ stores, that is when compared with (1,4,5)IP 3 . With low concentrations of adenophostin A, DeLisle et al. (7) reported a stimulation of Ca 2ϩ -dependent Cl Ϫ current without a detectable release of intracellular Ca 2ϩ . This suggests that in addition to its action as an agonist for the (1,4,5)IP 3 receptor, adenophostin A either acts downstream of (1,4,5)IP 3 in the pathway signaling capacitative calcium entry (perhaps on the channels themselves) or preferentially releases Ca 2ϩ from stores coupled to the activation of Ca 2ϩ entry. In this study, we have examined the effects of adenophostin A on Ca 2ϩ entry in a mammalian cell line, rat basophilic leukemia cells (RBL-1). Adenophostin A and (2,4,5)IP 3 , both non-metabolizable analogs of (1,4,5)IP 3 , were compared with regard to their abilities to release Ca 2ϩ from intracellular stores, to activate Ca 2ϩ entry, and to activate the well characterized calcium release-activated calcium current (I crac ) (9). While we find no evidence for a direct activation of calcium entry or of I crac by adenophostin A, we have discovered an unusual and unique ability of adenophostin A to support I crac under conditions of limited intracellular calcium buffering. Adenophostin A will be a useful tool for further investigations into the signaling mechanisms for capacitative calcium entry.
Fura-2 Loading-The attached cells were mounted in a Teflon chamber and incubated with 1 M fura-2/AM (Molecular Probes) for 25 min at room temperature. The cells were then washed and bathed in a HEPES-buffered physiological saline solution (in mM: 150 NaCl, 4.7 KCl, 1.8 CaCl 2 , 1.13 MgCl 2 , 10 glucose, and 10 HEPES, pH 7.2; HPSS) at room temperature for at least 10 min before Ca 2ϩ measurements were made.
Fluorescence Measurements-The fluorescence of the fura-2-loaded cells was monitored with a photomultiplier-based system, mounted on a Nikon Diaphot microscope equipped with a Nikon 40ϫ (1.3 N.A.) Neofluor objective. The fluorescence light source was provided by a Deltascan D101 (Photon Technology International Ltd.), equipped with a light path chopper and dual excitation monochromators. The light * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Electrophysiology-The normal external saline was HPSS described above. Nominally Ca 2ϩ -free saline was the same except no CaCl 2 was added. The bath volume (0.5 ml) was rapidly exchanged with a gravity perfusion system. The labels in the figures indicate the exact times when new bath solution was introduced, without any correction for the dead time required for a new solution to reach the cell. For fluorescence experiments, the patch pipette (2-4 MS, Corning glass, 7052) contained (in mM) 150 KCl, 10 NaCl, 2 MgCl 2 , 10 HEPES, 0.1 EGTA, 0.05 fura-2 free acid, 1 MgATP, pH 7.2. For measurement of I crac , the pipette solution was (in mM) 140 Cs-Asp, 2 MgCl 2 , 10 HEPES, 10 BAPTA-Cs 4 , 1 MgATP (free Ca 2ϩ 100 nM, pH 7.2). In some I crac experiments 0.1 mM BAPTA or EGTA was used instead of 10 mM BAPTA. The bath solution contained (in mM) 140 NaCl, 4.7 KCl, 10 CsCl, 10 CaCl 2 (or 10 MgCl 2 for Ca 2ϩ -free solution), 1.13 MgCl 2 , 10 glucose, and 10 HEPES, pH 7.2.
Whole cell voltage clamp was carried out as described previously (12,13). The holding potential was 30 mV, and Ca 2ϩ entry was measured by stepping to Ϫ60 mV. I crac was measured at a potential of Ϫ100 mV from voltage ramps between Ϫ100 and ϩ60 mV over a period of 160 ms. All voltages were corrected for a 10 mV liquid-junction potential. Data acquisition and analysis were performed with Axopatch-1C amplifier and PCLAMP 6.1 software (Axon Instruments, Burlingame, CA). Currents were filtered at 1 kHz and digitized at 300-fs intervals. Fig. 1 illustrates the protocol used to assess Ca 2ϩ entry with a series of concentrations of (2,4,5)IP 3 and adenophostin A. Cells patch-clamped in the whole cell configuration were held at 30 mV to prevent Ca 2ϩ entry. The pipettes contained differing concentrations of (2,4,5)IP 3 or adenophostin A which in-duced release of intracellular Ca 2ϩ . For (2,4,5)IP 3 , this release was generally accompanied by the appearance of [Ca 2ϩ ] i oscillations. This was true for adenophostin A as well, but only with the lower concentrations. 2 Where indicated, the holding potential was changed to Ϫ60 mV to provide a controlled driving force for Ca 2ϩ entry, and a subsequent rise in [Ca 2ϩ ] i occurred.

RESULTS
The [Ca 2ϩ ] i rise at Ϫ60 mV with the protocol in Fig. 1 may give a reasonable estimate of Ca 2ϩ entry at different concentrations of (2,4,5)IP 3 or adenophostin, but the initial [Ca 2ϩ ] i spike is not likely a reliable indicator of the magnitude of release. As in previous studies (14,15), we utilized the ionomycin residual technique, illustrated in Fig. 2. In this experiment, a cell incubated in nominally Ca 2ϩ -free medium (ϩ 1 mM EGTA to prevent Ca 2ϩ entry due to ionomycin) was activated by 250 nM adenophostin A in the pipette. Subsequent addition of a high concentration (10 M) of ionomycin induced a small release of Ca 2ϩ compared with a control cell, indicative of significant depletion of intracellular Ca 2ϩ stores by adenophostin A. Fig. 3 (top) summarizes the data for Ca 2ϩ release by adenophostin A and (2,4,5)IP 3 . It was not possible to obtain graded release of Ca 2ϩ with adenophostin A. 0.25 M adenophostin A depleted Ca 2ϩ to the same extent as the highest concentration tested, 10 M, and lower concentrations (0.1 and 0.05 M) failed to consistently activate release and entry. However, as reported previously (15), one concentration of (2,4,5)IP 3 , 10 M, induced a consistent and partial depletion of Ca 2ϩ stores. Fig. 3 (bottom) shows data for Ca 2ϩ entry in response to adenophostin A and (2,4,5)IP 3 within the concentration range of reproducible Ca 2ϩ release. At the highest concentrations tested, both adenophostin A (10 M) and (2,4,5)IP 3 (100 M) released Ca 2ϩ and activated Ca 2ϩ entry to a similar extent. At 10 M (2,4,5)IP 3 , release of Ca 2ϩ was incomplete, and likewise, entry was lower than at other concentrations. Interestingly, at the intermediate concentrations of (2,4,5)IP 3 , even though (2,4,5)IP 3 appeared to release Ca 2ϩ maximally, the Ca 2ϩ entry response was significantly less that that seen with adenophostin A.
The level of Ca 2ϩ entry in Figs. 1 and 3 is determined not only by Ca 2ϩ entry but also the extrusion across the plasma membrane and uptake into the endoplasmic reticulum. Therefore, we decided to examine the effect of adenophostin A and (2,4,5)IP 3 on Ca 2ϩ entry by using techniques that avoid these complicating factors. In the experiments shown in Fig. 4, prior to changing the holding potential to Ϫ60 mV, the external medium was changed from one containing 1.8 mM Ca 2ϩ to one containing 10 mM Ba 2ϩ . Ba 2ϩ is able to permeate the channels opened by depleting intracellular stores but is not a substrate  for plasma membrane or endoplasmic reticulum Ca 2ϩ pumps (16 -18). The summarized data in Fig. 4 shows that Ba 2ϩ entry is not significantly different when activated by 25 M (2,4,5)IP 3 or 0.25 M adenophostin A.
Another technique for assessing Ca 2ϩ entry that avoids effects of Ca 2ϩ pumps is to measure the current associated with entry directly, the calcium release-activated calcium current, or I crac (19). Fig. 5 shows that adenophostin A activates an inward current in RBL-1 cells and that this current has the following properties indicating that it is I crac : (i) the current is strongly inwardly rectifying with a magnitude and current-voltage relationship similar to that previously described for I crac by Hoth and Penner (19); (ii) the current is observed with strong intracellular calcium buffering (10 mM BAPTA) but is lost when external calcium is removed (Fig. 5). Consistent with the result from the Ba 2ϩ experiments, the amount of I crac activated with adenophostin A was the same as that with (2,4,5)IP 3 at all concentrations except for 10 M (2,4,5)IP 3 , which only partially depleted the intracellular Ca 2ϩ stores (Fig. 6). Therefore, the results suggest that adenophostin A and (2,4,5)IP 3 activate Ca 2ϩ entry similarly. Because of the different results from those obtained in the Ca 2ϩ entry experiments, in some experiments we followed the exact same protocol as used in Fig. 1, that is holding the cell at 30 mV and measuring I crac with 1.8 mM external Ca 2ϩ at 300 s after the establishment of whole cell configuration. Again, there was no difference in the amount of I crac activated (1.84 Ϯ 0.28 pA/pF, n ϭ 7, for 0.25 M adenophostin A versus 1.82 Ϯ 0.11 pA/pF, n ϭ 6, for 25 M (2,4,5)IP 3 ). The potential meaning of the disparate findings with measurement of net [Ca 2ϩ ] i changes compared with Ba 2ϩ entry or I crac will be addressed under "Discussion." The results to this point reveal no clear distinctions between the actions of (2,4,5)IP 3 and adenophostin A. Thus, the ability of adenophostin A to activate Ca 2ϩ entry and I crac may result from its ability to release intracellular Ca 2ϩ stores, rather than from a direct effect on calcium entry channels. To investigate further this issue, we examined the temporal relationship between the activation of I crac and Ca 2ϩ release, as we did in an earlier study for (2,4,5)IP 3 (15). We measured the latency for I crac activation with adenophostin A and compared that with the latency for Ca 2ϩ release from intracellular stores. We found that adenophostin A activated I crac with a latency correlating to, but longer than, the latency for Ca 2ϩ release (Fig. 7). A similar pattern for Ca 2ϩ release and I crac activation was found for (2,4,5)IP 3 (15). Thus, adenophostin A appears to activate I crac only after Ca 2ϩ release from intracellular Ca 2ϩ stores is well under way, suggesting that adenophostin A activates I crac as a result of the release of intracellular Ca 2ϩ , rather than through a direct action on store-operated Ca 2ϩ channels or some other site downstream of Ca 2ϩ release.
One complication in the interpretation of the kinetic data in Fig. 7 is that the measurement of Ca 2ϩ release and I crac are carried out under different experimental conditions. In the fura-2 experiments to measure Ca 2ϩ release, the pipette solution contains a modest amount of Ca 2ϩ buffer (0.1 mM EGTA) so that net changes in cytoplasmic Ca 2ϩ can be observed. On the other hand, measurements of I crac in this study (as well as in all other published studies) were carried out with a much higher concentration of Ca 2ϩ buffer in the pipette (10 mM BAPTA or EGTA) to minimize Ca 2ϩ -dependent inactivation of I crac . Thus, we attempted to design a protocol that would permit measurement of I crac and Ca 2ϩ release (with fura-2) simulta- The protocol was similar to that for Fig. 1, except that at 300 s, the medium was changed from one containing 1. neously on a single RBL-1 cell. The cells were loaded with fura-2 (see "Experimental Procedures"), and cells were patchclamped with pipettes containing the Cs ϩ -containing I crac intracellular solution with 50 M fura-2 but with BAPTA reduced to 0.1 mM. To minimize Ca 2ϩ entry and Ca 2ϩ -induced inactivation, we held the cell at ϩ30 mV. I crac was assayed by a short ramp (from Ϫ100 to 60 mV for 160 ms) every 1-5 s. With this protocol, we expect Ca 2ϩ entry to be greatly reduced during the inter-ramp interval and to be mainly limited to the short duration of the ramp. As shown in Figs. 8 and 9, this procedure resulted in the development of a significant inward current when 0.5 M adenophostin A was included in the pipette. This current had an I-V relationship expected of I crac (Fig. 8); with these ionic conditions, no other known current could produce such an I-V relationship. The current was dependent on extracellular Ca 2ϩ but did not depend on the level of [Ca 2ϩ ] i , consistent with Ca 2ϩ acting as a charge carrier rather than as an activator of the current. Thus, the current has all of the properties expected of the depletion-activated current, I crac , seen with stronger intracellular Ca 2ϩ buffers. Surprisingly, (2,4,5)IP 3 (two batches from two different sources) completely failed to activate any detectable I crac (Table I) with this protocol even though Ca 2ϩ is almost depleted to the same extent as with adenophostin A (Fig. 3). Furthermore, three other store-depletion reagents (3-F-IP 3 , thapsigargin, and ionomycin) similarly failed to activate I crac with the low Ca 2ϩ buffer protocol (Table  I). 3-F-IP 3 is another nonhydrolyzable (1,4,5)IP 3 analog. The K d for 3-F-IP 3 for the (1,4,5)IP 3 receptor is comparable to that of (1,4,5)IP 3 (13 nM compared with 6 nM for (1,4,5)IP 3 by the same assay) but still higher than that for adenophostin A. The average peak Ca 2ϩ release induced by 3-F-IP 3 was 256.8 Ϯ 23.5 nM (n ϭ 9 at 50 M) and 330.0 Ϯ 1.0 nM (n ϭ 2 at 100 M), respectively, similar to that observed for (2,4,5)IP 3 and adenophostin A ( Fig. 1 and not shown). Finally, with the high Ca 2ϩ buffer protocol, all of the above store-depleting reagents activated I crac to a similar extent with almost 100% success rate (not shown), suggesting that the difference between adenophostin A and other store-depletion reagents is only seen with the more physiological low Ca 2ϩ buffering protocol.
From the experiment depicted in Fig. 9, some interesting observations about the nature of the Ca 2ϩ -mediated inactivation of I crac can be made. In this particular experiment, the initial sampling of I crac occurred every second. This resulted in an elevated [Ca 2ϩ ] i which did not return to base line, suggesting that with this schedule, sufficient Ca 2ϩ enters during the ramp (presumably at a very high rate due to diminished inactivation) to maintain the average [Ca 2ϩ ] i at a level similar to that seen with sustained activation at a physiological membrane potential (as in Fig. 1). Yet, this sustained level of around 250 nM [Ca 2ϩ ] i is obviously insufficient to induce inactivation because a significant inward current develops. This is consistent with the prior suggestion (20, 21) that Ca 2ϩ -dependent inactivation of I crac results from the action of Ca 2ϩ at a site very close to the mouth of the channel and is insensitive to the [Ca 2ϩ ] i in the bulk of the cytoplasm (22). Note that when the inter-ramp interval was increased to 5 s, the average [Ca 2ϩ ] i declined substantially, whereas I crac changed little. Removal and restoration of external Ca 2ϩ reversed and reinstated I crac with minimal effects on [Ca 2ϩ ] i . Finally, changing the holding potential to Ϫ60 mV caused I crac to disappear completely, because sustained entry of Ca 2ϩ through the channel causes inactivation to a level below that which can be detected. How-  6. I crac is activated to a similar extent with adenophostin  A or (2,4,5)IP 3 . Cells were held at 30 mV. I crac was determined every 5 s after the establishment of the whole cell configuration with a voltage ramp from Ϫ100 to 60 mV. I crac was measured as the average current (from 10 to 12 ms after the initiation of ramp). Non-I crac current or leak current was taken as the current in nominally Ca 2ϩ -free solution or as the current before I crac was activated, and this current was subtracted. Current density was then determined by normalizing I crac amplitude to cell capacitance which is proportional to cell surface area. The external [Ca 2ϩ ] was 10 mM. The data for I crac with (2,4,5)IP 3 are the same as those reported previously (15) and were obtained as a part of the same series of experiments as for adenophostin A. Means Ϯ S.E. from 4 to 17 independent determinations. ever, the steady-state [Ca 2ϩ ] i level at Ϫ60 mV was not substantially different from that in the initial phase of the experiment, when the inter-ramp interval was 1 s but I crac was well developed. Again, this is consistent with the conclusion that it is the Ca 2ϩ concentration in the microenvironment of the Ca 2ϩ channel rather than the global, cytoplasmic [Ca 2ϩ ] i which is responsible for inactivation of the Ca 2ϩ entry channels. It is presumed that when Ca 2ϩ entry occurs physiologically, i.e. at Ϫ60 mV, the observed sustained entry of Ca 2ϩ is due to I crac , but the level of the current is below that which can be detected with current technology. However, this has not been unequivocally established by either the current studies or any previous work on the electrophysiology of store-operated Ca 2ϩ channels.
The low Ca 2ϩ -buffering protocol with adenophostin A permits simultaneous measurement of I crac activation and Ca 2ϩ release, and thus the latencies for these two events can be directly compared. Fig. 10 illustrates a typical finding with adenophostin A; there was a clear delay for the activation of I crac after Ca 2ϩ release was initiated. Fig. 11 summarizes a number of experiments using this technique to simultaneously examine Ca 2ϩ release and I crac activation. Two concentrations of adenophostin A, 0.5 and 2.0 M, were employed, and as was seen in the earlier experiments, at both concentrations there was a significant delay between the initiation of release and the initiation of I crac . DISCUSSION Previous reports suggest that adenophostin A can activate Ca 2ϩ influx either in the absence of Ca 2ϩ release (7) or with minimal Ca 2ϩ release (8). These investigations based their conclusions on the activities of Ca 2ϩ -dependent Cl Ϫ channels, an indirect reporter of [Ca 2ϩ ] i changes. In this study, we ex-

FIG. 7. Ca 2؉ release and I crac activation with adenophostin A.
Top, a single RBL-1 cell, loaded with fura-2, was held at 30 mV under conventional whole cell configuration. At time 0, the whole cell configuration was established. After a short latency, adenophostin A (2 M in the pipette) caused a transient Ca 2ϩ release from intracellular Ca 2ϩ stores. The arrows indicate the intervals for estimation of latency for release and time to peak release. Middle, adenophostin A (2 M in the pipette) activated I crac with a latency somewhat longer than that for Ca 2ϩ release. The current was measured at the potential of Ϫ100 mV from voltage ramps from Ϫ100 to 60 mV and plotted versus time. The nonspecific current (the current before the induction of I crac or the current remaining when external Ca 2ϩ was replaced) was subtracted.
The arrows indicate the intervals for estimation of latency of I crac and development time for I crac . Bottom, cumulative data for the four intervals illustrated above determined with four different pipette concentrations of adenophostin A. Each point represents average measurements from 8 or more cells. The latency for Ca 2ϩ release and I crac decreases with increasing concentration of adenophostin A. However, a significant discrepancy exists between the time for store depletion and activation of I crac at all concentrations. The development time for I crac was relatively constant .   FIG. 8. Adenophostin-induced current responses in a strongly  buffered cell (10 mM BAPTA) and a weakly buffered cell (0.1 mM  BAPTA). The gray solid line represents data from a highly buffered cell and the black dotted line from a weakly buffered cell. The current response to the ramp protocol shown (Ϫ100 mV to 60 mV) reveals identical current-voltage relations under the two experimental conditions. Currents under the low buffer condition were generally smaller than under high buffer conditions (about 2/3), however. In this example the low buffer trace has been scaled by a factor of 1.49 to illustrate the similar shapes of the two curves. Membrane potential was held at 30 mV and I crac assessed every second with ramps from Ϫ100 to 60 mV lasting 160 ms. Where indicated, the interval between ramps was increased from 1 to 5 s. Also where indicated, external Ca 2ϩ (which was 10 mM) was removed and restored. Finally, the holding potential was changed from 30 to Ϫ60 mV. amined the effects of adenophostin A and (2,4,5)IP 3 on Ca 2ϩ entry assessed indirectly by measurement of net [Ca 2ϩ ] i changes and directly by measuring the store-operated Ca 2ϩ current (I crac ). For the most part, our results suggest that adenophostin A acts in a manner similar to that of (2,4,5)IP 3 in activating Ca 2ϩ entry, through release of intracellular Ca 2ϩ , rather than as a direct activator of store-operated Ca 2ϩ channels or through some other mechanism independent of intracellular Ca 2ϩ store depletion. We conclude this because with adenophostin A (i) we never observed activation of entry in the absence of Ca 2ϩ release, and (ii) there was a substantial delay between the activation of release and entry, similar to that observed previously for (2,4,5)IP 3 .
Had we only utilized [Ca 2ϩ ] i changes as an indicator of entry, we might have been misled into concluding that adenophostin was more efficient at inducing entry than (2,4,5)IP 3 . With intermediate concentrations of (2,4,5)IP 3 , net Ca 2ϩ entry was not maximal, despite apparent complete depletion of intracellular Ca 2ϩ stores (Fig. 3). There are two possible explanations for this phenomenon: either (i) (2,4,5)IP 3 (and not adenophostin A) activates a pathway for Ca 2ϩ removal, or (ii) despite the ability of (2,4,5)IP 3 to deplete stores completely in the absence of Ca 2ϩ entry (i.e. with the protocol shown in Fig. 2), in the presence of Ca 2ϩ entry some refilling of the stores occurs. We consider the latter alternative the more likely because previous studies have shown that (2,4,5)IP 3 acts as a partial agonist compared with (1,4,5)IP 3 when rate of release rather than extent of release is measured (23). Thus, in the absence of Ca 2ϩ entry, (2,4,5)IP 3 would be able to deplete completely the stores through submaximal activation of the (1,4,5)IP 3 receptor, but when entry elevates [Ca 2ϩ ] i , some refilling would occur. Presumably adenophostin A is a full agonist for the receptor producing a maximal permeability increase, and little or no refilling would occur.
Despite similarities in the actions of adenophostin A and (2,4,5)IP 3 , one particularly interesting and potentially significant difference was noted. Only adenophostin A was capable of consistently activating I crac in cells patch-clamped with low Ca 2ϩ buffer solutions, i.e. when [Ca 2ϩ ] i was permitted to fluctuate. (2,4,5)IP 3 failed to support I crac under this condition, even when employed at a concentration of 100 M; at this concentration its Ca 2ϩ release and Ca 2ϩ entry activation appear maximal and similar to adenophostin A (Table I and Fig.  3). It has not previously been possible to observe I crac in RBL cells without the presence of high concentrations of intracellular buffers. In the present study, even with adenophostin A I crac could only be observed when entry was largely prevented by the use of a positive holding potential. We assume that during calcium entry under physiological conditions, i.e. when an increase in the fura-2 signal is observed, this entry is mediated at least in part by I crac at a level below that which can be detected   Fig. 9, that is when Ca 2ϩ entry proceeds continuously through the channels, neither agent produces detectable I crac because Ca 2ϩ -dependent inactivation makes the current too small to detect. Note that Ca 2ϩ -dependent inactivation is assumed to underlie the failure of (2,4,5)IP 3 , thapsigargin, and ionomycin to activate detectable I crac with low Ca 2ϩ buffers, but this is not proven. The Ca 2ϩ -dependent inactivation that occurs close to the mouth of the channel (21) should be considerably reduced by holding at ϩ30 mV. However, there are other slower mechanisms of inactivation (24) that may be more sensitive to the global [Ca 2ϩ ] in the cytoplasm. One possibility is that adenophostin A prevents or attenuates only the rapid inactivation that occurs in or near the cytoplasmic opening of the store-operated calcium channel. Regardless of the explanation, the ability to measure physiologically relevant changes in [Ca 2ϩ ] i resulting from Ca 2ϩ release simultaneously with measurement of I crac will be a useful tool for future research into the quantitative and temporal relationships between Ca 2ϩ discharge and I crac activation. One example in the present study is the clear demonstration of a delay between the release of Ca 2ϩ and the initiation of I crac . This delay supports the idea of a diffusible signal responsible for the activation of the capacitative calcium entry channels, rather than a direct physical interaction between the endoplasmic reticulum and the plasma membrane (4). Such an observation has been made previously with (2,4,5)IP 3 , but it was necessary to compare cells activated under very different conditions (release with low buffering; I crac with high buffering) (15) weakening the interpretation of these experiments considerably.
Unlike the previous work with Xenopus oocytes (7, 8), we found no evidence for a dissociation between the ability of adenophostin A to release Ca 2ϩ and its ability to activate entry. The substantial delay between release of Ca 2ϩ and activation of I crac argues against a direct activation of the channels by adenophostin A. The findings in oocytes may be related to the same property of adenophostin A which in the current study results in the consistent activation of I crac under low Ca 2ϩbuffering conditions. Alternatively, the studies in oocytes utilized Ca 2ϩ -dependent chloride currents for assessment of [Ca 2ϩ ] i changes, and it is possible that they may sometimes underestimate the extent of Ca 2ϩ release, particularly if it occurs at sites which are not in the proximity of the Ca 2ϩsensitive chloride channels. Regardless of the explanation, it is clear from the earlier work in oocytes and from the present work in RBL-1 cells that adenophostin A may have interesting actions in addition to its ability to activate (1,4,5)IP 3 receptors with high affinity. It may thus prove a useful tool for unraveling the molecular pathways that regulate capacitative calcium entry.