Ca2+ Entry Activated byS-Nitrosylation

The coupling between Ca2+ pools and store-operated Ca2+ entry channels (SOCs) remains an unresolved question. Recently, we revealed that Ca2+ entry could be activated in response to S-nitrosylation and that this process was stimulated by Ca2+ pool emptying (Favre, C. J., Ufret-Vincenty, C. A., Stone, M. R., Ma, H-T., and Gill, D. L. (1998) J. Biol. Chem. 273, 30855–30858). In DDT1MF-2 smooth muscle cells and DC-3F fibroblasts, Ca2+ entry activated by the lipophilic NO donor, GEA3162 (5-amino-3-(3,4-dichlorophenyl)1,2,3,4-oxatriazolium), or the alkylator, N-ethylmaleimide, was observed to be strongly activated by transient external Ca2+ removal, closely resembling activation of SOC activity in the same cells. The nonadditivity of SOC and NO donor-activated Ca2+ entry suggested a single entry mechanism. Calyculin A-induced reorganization of the actin cytoskeleton prevented SOC but had no effect on GEA3162-induced Ca2+ entry. However, a single entry mechanism could account for both SOC and NO donor-activated entry if the latter reflected direct modification of the entry channel byS-nitrosylation, bypassing the normal coupling process between channels and pools. Small differences between SOC and GEA3162-activated Ba2+ entry and sensitivity to blockade by La3+ were observed, and in HEK293 cells SOC activity was observed without a response to thiol modification. It is concluded that in some cells, S-nitrosylation modifies an entry mechanism closely related to SOC and/or part of the regulatory machinery for SOC-mediated Ca2+ entry.

kinase receptors, through activation of phospholipase C, generate the second messenger, inositol 1,4,5-trisphosphate. This chemical message diffuses rapidly within the cytosol to interact with inositol 1,4,5-trisphosphate receptors located on the ER, which serve as Ca 2ϩ channels to release luminal stored Ca 2ϩ and generate the initial Ca 2ϩ signal phase (1,3). The resulting depletion of Ca 2ϩ stored within the ER lumen serves as the primary trigger for a message that is returned to the plasma membrane, resulting in the slow activation of "store-operated" Ca 2ϩ entry channels (2, 4 -6). This second Ca 2ϩ entry phase of Ca 2ϩ signals serves to mediate longer term cytosolic Ca 2ϩ elevations and provides a means to replenish intracellular stores (2,4). Whereas receptor-induced generation of inositol 1,4,5-trisphosphate and the function of Ca 2ϩ release channels to mediate the initial Ca 2ϩ -signaling phase is well understood, the mechanism for coupling ER Ca 2ϩ store depletion with Ca 2ϩ entry remains a crucial but unresolved question (4 -6).
Recently, several major channels have been shown to be regulated by thiol nitrosylation, a process becoming recognized as an important nitric oxide (NO)-mediated posttranslational modification affecting control over a diverse array of signaling and regulatory proteins (7)(8)(9)(10)(11)(12). Such S-nitrosylation-mediated effects are direct and independent of activation of guanylyl cyclase, which is a major target for NO and a frequent mediator of the actions of NO (13,14). Studies have revealed that nitrosothiol formation underlies the direct modifying action of NO on a number of important plasma membrane and intracellular channels for Ca 2ϩ and other ions including the N-methyl-Daspartate receptor (8), cyclic nucleotide-gated cation channel (15,16), Ca 2ϩ -activated K ϩ channel (17), L-type Ca 2ϩ channel (18), and the ryanodine receptor Ca 2ϩ release channel (19). For several of these channels, NO donor-induced S-nitrosylation results in channel activation, and this activation is mimicked by alkylation of the same thiol groups (15)(16)(17)(18)(19). Because of the reactivity of thiols toward NO, the sphere of influence of NO can be highly restricted and, rather than diffusion-dependent, NO (or an equivalent of the nitrosonium ion, NO ϩ ) may be donated and exchanged between neighboring protein thiols by local transnitrosation events (7, 9 -11, 15, 16).
We recently utilized a combination of membrane-permeant NO donors and alkylators to probe the role of S-nitrosylation in the process of Ca 2ϩ entry and its relationship to Ca 2ϩ pool depletion (20). A novel class of lipophilic NO donors, including the oxatriazole-5-imine derivative, GEA3162, activated Ca 2ϩ entry independent of the well defined NO target, guanylyl cyclase. Strikingly similar Ca 2ϩ entry induced by cell permeant alkylators indicated that this Ca 2ϩ entry process was activated through thiol modification. Significantly, Ca 2ϩ entry activated by NO donors or alkylators was stimulated by Ca 2ϩ pool depletion, which increased the rate and size of the Ca 2ϩ response and the sensitivity to thiol modifiers. These results led us to postulate that S-nitrosylation may underlie activation of an important store-operated Ca 2ϩ entry mechanism. Here we have examined the relationship between store-operated Ca 2ϩ entry occurring independently of S-nitrosylation and Ca 2ϩ entry activated in response to S-nitrosylation.
Measurement of Intracellular Calcium-Cells grown on coverslips for 1 day were transferred to Hepes-buffered Krebs medium (107 mM NaCl, 6 mM KCl, 1.2 mM MgSO 4 , 1 mM CaCl 2 , 1.2 mM KH 2 PO 4 , 11.5 mM glucose, 0.1% bovine serum albumin, 20 mM Hepes-KOH, pH 7.4) and loaded with fura-2/AM (2 M) for 10 min at 20°C. Cells were washed, and the dye was allowed to deesterify for a minimum of 15 min at 20°C. Approximately 95% of the dye was confined to the cytoplasm as determined by the signal remaining after saponin permeabilization (25,26). Fluorescence emission at 505 nm was monitored with excitation at 340 and 380 nm; Ca 2ϩ measurements are shown as 340/380-nm ratios obtained from a groups of 10 -12 cells. Details of these Ca 2ϩ measurements were recently described for DDT 1 MF-2 (27) and DC-3F cells (24). Resting Ca 2ϩ levels in DDT 1 MF-2 cells were approximately 60 -90 nM and 25-50 nM in DC-3F cells; maximal activation by GEA3162 resulted in up to 600 nM Ca 2ϩ . All measurements shown are representative of a minimum of three, and in most cases, a much larger number of independent experiments.

RESULTS AND DISCUSSION
Our previous results revealed that NO donors including nitroprusside, nitrite, and the lipophilic donor, GEA3162, were effective in directly inducing Ca 2ϩ entry (20). A highly similar entry of Ca 2ϩ was induced with alkylators including NEM and 4-vinylpyridine, indicating that the activation of Ca 2ϩ entry resulted from thiol modification, either nitrosylation or alkylation (20). The Ca 2ϩ entry observed with NO donors or with alkylators, in both cases, was substantially enhanced by emptying Ca 2ϩ pools before administration of the activator. Pool emptying increased three parameters of thiol modifier-induced Ca 2ϩ entry: the time-dependence of entry, the size of the Ca 2ϩ entry response, and the sensitivity to thiol modifier (20). The most prominent of these effects was the time dependence. Thus, in normal cells with filled pools, there was a pronounced lag in the Ca 2ϩ entry response to thiol modifiers of at least 1 min; after pool emptying, the Ca 2ϩ entry response was extremely rapid, suggesting that pool emptying had allowed the Ca 2ϩ entry channel to alter its configuration to expose a thiol group that was important in modifying channel activity (20). The question of whether this putative entry channel was indeed the store-operated Ca 2ϩ channel was important to address.
Store-operated Ca 2ϩ entry channels display a further important characteristic. In many cells, the entry of Ca 2ϩ , activated after pool depletion, becomes deactivated with time, and transient removal and readdition of extracellular Ca 2ϩ is a well described means for reactivating the entry mechanism (24, 28 -31). The effect is frequently referred to as the Ca 2ϩ "overshoot" response, since it results in a transiently high reactivation of Ca 2ϩ entry, which then deactivates once again with time (24,28). The results in Fig. 1 reveal that transient removal of external Ca 2ϩ has a dramatic enhancing effect on the operation of Ca 2ϩ entry activated by GEA3162. Untreated DDT 1 MF-2 cells exposed briefly to nominally Ca 2ϩ -free medium then re-turned to medium containing normal Ca 2ϩ showed no change in cytosolic Ca 2ϩ (Fig. 1A). The addition of GEA3162 at 25 M, a submaximal concentration under normal conditions (20), in the continued presence of external Ca 2ϩ induced a modest rise of Ca 2ϩ after a lag of approximately 2 min. If external Ca 2ϩ was removed before the addition of 25 M GEA3162 (Fig. 1B), no significant change in Ca 2ϩ occurred for several minutes, confirming the lack of any effect of the NO donor on release of internal Ca 2ϩ . However, upon readdition of external Ca 2ϩ , a rapid entry of Ca 2ϩ occurred, resulting in a considerably larger peak of Ca 2ϩ (Fig. 1B) than that observed in the continuous presence of external Ca 2ϩ (Fig. 1A). Thus the removal and readdition of Ca 2ϩ considerably enhanced the effectiveness of GEA3162. Control experiments revealed that prolonged (10 min) removal of external Ca 2ϩ did not cause any release of Ca 2ϩ from pools and that following such prolonged external Ca 2ϩ removal, no entry of Ca 2ϩ was observed upon the readdition of Ca 2ϩ in the absence of GEA3162. The stimulatory effect of transient Ca 2ϩ removal on the action of GEA3162 was further characterized as shown in Fig. 2. In this experiment external Ca 2ϩ was transiently removed after the addition of different GEA3162 concentrations. After adding GEA3162 at 1 M, a 3-min period of external Ca 2ϩ removal resulted in only a very slight entry of Ca 2ϩ ( Fig. 2A). However, after the addition of 10 M GEA3162, the transient removal of external Ca 2ϩ triggered a much more significant and rapid increase in Ca 2ϩ following Ca 2ϩ readdition (Fig. 2B). Under normal conditions of external Ca 2ϩ , GEA3162 at 10 M was below its effective threshold and induced almost no Ca 2ϩ entry (20). Therefore, the entry of Ca 2ϩ observed after the brief removal of external Ca 2ϩ represents a real potentiation of the effect of GEA3162. The resultant increase in Ca 2ϩ was rapid but transient and began to decline after 1 min. Removal of Ca 2ϩ prevented any further entry of Ca 2ϩ , and the Ca 2ϩ level fell rapidly. After a further 3 min, readdition of Ca 2ϩ resulted again in a rapid and transient entry of Ca 2ϩ . The entry of Ca 2ϩ could be repeatedly reactivated by transient removal of Ca 2ϩ (Fig. 2B). Although the peak size of the response to 10 M GEA3162 after the initial transient Ca 2ϩ depletion was smaller, the peaks following subsequent brief periods of Ca 2ϩ removal were larger and ap- This pattern of deactivation and reactivation by transient removal of Ca 2ϩ is highly similar to the operation of storeoperated Ca 2ϩ entry channels. As shown in Fig. 3A, after thapsigargin-induced pool emptying in the absence of external Ca 2ϩ , readdition of Ca 2ϩ caused a large increase in cytosolic Ca 2ϩ , reflecting a high level of store-operated Ca 2ϩ entry. The entry of Ca 2ϩ rapidly deactivated with time, and subsequent removal of external Ca 2ϩ prevented any further Ca 2ϩ entry. Upon the readdition of external Ca 2ϩ , maximal store-operated Ca 2ϩ entry was restored. This overshoot response pattern classically reflects the operation of store-operated Ca 2ϩ entry and is believed to represent the function of Ca 2ϩ -binding sites, which negatively control store-operated Ca 2ϩ entry (24,28,29,31). According to such a model, as Ca 2ϩ increases in the cytosol, binding of Ca 2ϩ to such regulatory sites inhibits entry; transient external Ca 2ϩ removal prevents Ca 2ϩ entry, allowing cytosolic Ca 2ϩ to fall rapidly as Ca 2ϩ is pumped out of the cell. As a result, Ca 2ϩ dissociates from the regulatory site, permitting the channel to become fully reactivated; upon readdition of Ca 2ϩ , a high level of Ca 2ϩ entry is again observed. As shown in Fig. 3A, this process could be repeated many times. However, it is important to reiterate that the entry observed was completely dependent on pool depletion. Thus, transient removal of Ca 2ϩ at the beginning of the trace before pools were emptied induced no entry of Ca 2ϩ . Indeed, in experiments with normal, pool-filled cells, repeated transient removal and readdition of Ca 2ϩ over a period of 30 min induced no change in cytosolic Ca 2ϩ (not shown). The means of activation, the appearance, and the size of the overshoot responses after pool emptying were all remarkably similar to those described above, activated in response to the NO donor. Yet in the case of the NO donor, pools were not emptied. This point is reinforced from the data shown in Fig. 3B. Thus addition of 15 M GEA3162 before thapsigargin had no effect on the size of the Ca 2ϩ pool released by subsequent addition of thapsigargin. Moreover, in this experiment the size of overshoots induced after application of both GEA3162 and thapsigargin was not measurably different from that induced by each agent alone. Also, addition of 15 M GEA3162 to cells after pool depletion with thapsigargin resulted in little significant change in the size of overshoots induced by repeated transient Ca 2ϩ removal (Fig. 3A). Thus, thapsigargin and NO donor induced similar shaped and sized overshoots that did not appear to be additive. This suggested they were activating either the same or a closely coupled entry mechanism.
As described above and earlier (20), thiol modification by either nitrosylation or alkylation activated a very similar entry of Ca 2ϩ . We therefore examined whether the stimulatory action of transient Ca 2ϩ removal also activated Ca 2ϩ entry induced by alkylators. Experiments utilized the DC-3F fibroblast cell line in which responses to NO donors and alkylators were similar to DDT 1 MF-2 cells. As shown in Fig. 4A, the addition of the alkylator, NEM, at 10 M induced only a slight increase in cytosolic Ca 2ϩ (Fig. 4A). However, if extracellular Ca 2ϩ was transiently removed for just a short (2 min) period, a substantial entry of Ca 2ϩ immediately followed the readdition of Ca 2ϩ (Fig. 4B). As with DDT 1 MF-2 cells, transient Ca 2ϩ removal without alkylator or NO donor present had no effect on cytosolic Ca 2ϩ in DC-3F cells (24). The shape and time dependence of the transient Ca 2ϩ removal-induced entry response seen after NEM treatment (Fig. 4B) and after pool emptying (24) were impressively similar in the DC-3F cells. NEM-induced entry of Ca 2ϩ into DDT 1 MF-2 cells was similarly potentiated by transient removal of Ca 2ϩ (data not shown).
Taken together, the above results and those published previously (20) revealed that Ca 2ϩ entry activated by S-nitrosyla- tion was stimulated by both Ca 2ϩ pool depletion and transient external Ca 2ϩ removal, the two primary conditions for activating store-operated Ca 2ϩ entry channels. From this it could be concluded that S-nitrosylation was affecting a process closely linked with store-operated Ca 2ϩ entry. However, other approaches to determining the relationship between the Ca 2ϩ entry mechanisms have revealed some interesting differences. In recent work, we determined that the operation of storeoperated Ca 2ϩ entry appears to involve trafficking of the ER toward the plasma membrane (32). One of the approaches to this work was to utilize the phosphatase inhibitor, calyculin A, which in two distinct cell types induced a profound redistribution of actin resulting in the formation of a tight ring of cortical actin filaments subjacent to the plasma membrane. This cortical actin appeared to act as a physical barrier to prevent close interaction between the ER and plasma membrane (32). Under this condition, the activation of store-operated Ca 2ϩ entry by pool emptying with thapsigargin was blocked in both cell types. We therefore compared the action of calyculin A on Ca 2ϩ entry activated by thapsigargin-induced Ca 2ϩ pool depletion with its effects on Ca 2ϩ entry activated by NO donor.
As shown in Fig. 5, we were able to directly observe the two means of Ca 2ϩ entry activation in a single trace. In normal DDT 1 MF-2 cells, typical activation of Ca 2ϩ entry via pool depletion and application of GEA3162 is shown in Fig. 5A. After the addition of 2 M thapsigargin, rapid release of Ca 2ϩ from pools was observed. The slower secondary peak was due to entry of Ca 2ϩ dependent on pool depletion; this became deactivated with time, and cytosolic Ca 2ϩ decreased to a reduced level. The basal Ca 2ϩ level reached was slightly higher than normal resting Ca 2ϩ , reflecting a small level of residual Ca 2ϩ entry. After removal of external Ca 2ϩ , this low level of entry was abolished, and upon subsequent readdition of Ca 2ϩ , the typical large overshoot of Ca 2ϩ entry was observed, consistent with that described above. After the store-operated entry had once again deactivated, the addition of a 100 M GEA3162 clearly activated a large increase in Ca 2ϩ entry, and this entry again deactivated with time. In similar experiments, lower GEA3162 concentrations also induced entry of Ca 2ϩ , although the entry was less prolonged (not shown). Note that in these experiments, deactivation of store-operated Ca 2ϩ entry occurred for a longer period of time compared with that in Fig. 3, and there was no further removal of external Ca 2ϩ . Thus, if GEA3162 was activating the same Ca 2ϩ entry pathway as pool emptying, then this result would suggest that GEA3162 was able to reverse the deactivation process occurring as a result of Ca 2ϩ inhibition.
Clearly, the picture was very different in the presence of 100 nM calyculin A. The experiment shown in Fig. 5B included a 10-min pretreatment with the phosphatase inhibitor that was sufficient to rearrange cortical actin into a tight band closely associated with the plasma membrane (32) and prevent storeoperated Ca 2ϩ entry. As seen in this experiment, whereas the size of the release peak of Ca 2ϩ in response to thapsigargin was unchanged, almost no Ca 2ϩ entry followed, and hence the second peak was eliminated. A very slight level of Ca 2ϩ entry appeared to remain, as seen by the removal of external Ca 2ϩ ; however, subsequent readdition of Ca 2ϩ caused only a small reactivation of Ca 2ϩ entry (that is, almost no overshoot), conclusively demonstrating that activation of store-operated entry had been blocked. Whereas the overshoot response was blocked, application of 100 M GEA3162 gave an immediate and large activation of Ca 2ϩ entry. The effects of applying lower GEA3162 concentrations was also unaltered by calyculin A treatment (not shown); thus, although the duration and peak sizes of the responses were smaller than with 100 M GEA3162, they were not different to the responses observed without calyculin A. Therefore, on the basis of sensitivity to modification by calyculin A, it might be concluded that the entry activated by pool emptying was quite distinct from that activated by S-nitrosylation.
However, this interpretation is not necessarily valid. Thus, we concluded from our recent work that the coupling mechanism between pools and store-operated Ca 2ϩ entry channels is through a trafficking event involving physical movement of some component of the ER toward the plasma membrane (32). In this model, the channels may be preexisting within the plasma membrane, but activation may be elicited by close approach of the ER membrane. Indeed, recent evidence suggests that for one type of Ca 2ϩ entry channel, the TRP3 channel, activation may occur via a reversible interaction between the ER-located inositol 1,4,5-trisphosphate receptor and the TRP3 channel itself or a closely related component within the Each of the agents was maintained in medium after addition throughout successive changes of medium with or without Ca 2ϩ . plasma membrane (33). Although there is some uncertainty about whether pool emptying is necessary for activation of this particular channel (33,34), there is certainly precedent for believing that the activation of store-operated Ca 2ϩ entry channels may involve interactions between the ER and plasma membrane (32). The action of calyculin A on preventing storeoperated Ca 2ϩ entry is believed to result from physical interruption of the coupling process that occurs between the ER and plasma membrane as a result of reorganization of F-actin into a tight cortical layer beneath the plasma membrane (32). Evidence for this action of calyculin A was based on the morphological redistribution of actin observed. In addition, the inhibitory action of calyculin A on store-operated Ca 2ϩ entry by cytochalasin D could be reversed by cytochalasin D. Thus, depolymerization of actin with cytochalasin D after calyculin A caused the cortical actin barrier to be removed and the coupling between ER and the plasma membrane to be reestablished, permitting activation of store-operated Ca 2ϩ entry (32). Since this action of calyculin A is not considered to be a direct effect on the channel itself but rather a prevention of the interaction with ER, it appeared to us that the lack of effect of calyculin A on GEA3162-induced Ca 2ϩ entry might reflect a direct action of the NO donor on the entry channel, perhaps circumventing or even mimicking the activation that results from interaction with ER.
Therefore, we sought to determine other parameters defining operation of the entry mechanisms to examine any differences between store-operated and GEA3162-activated entry. Our attention turned toward examination of cation sensitivity and specificity of the entry processes. We investigated the passage of different divalent alkaline-earth cations and of the blocking action of La 3ϩ . The passage of Sr 2ϩ and Ba 2ϩ ions could be assessed directly by fura-2 ratio-fluorimetry. Experiments revealed that both store-operated entry and entry activated by GEA3162 allowed passage of Sr 2ϩ ions (Fig. 6). The addition of 25 M GEA3162 to DDT 1 MF-2 cells in medium containing 1 mM Sr 2ϩ in place of 1 mM Ca 2ϩ activated entry of Sr 2ϩ with kinetics similar to that of Ca 2ϩ (Fig. 6A). When using cells in which pools had been pool-depleted in the absence of external Ca 2ϩ (Fig. 6B), the addition of Sr 2ϩ caused a rapid entry of Sr 2ϩ through store-operated Ca 2ϩ entry channels. The kinetics of deactivation of Sr 2ϩ entry were not dissimilar from those for Ca 2ϩ entry (compare with Fig. 5A), although there did not appear to be a residual of Sr 2ϩ entry as appeared for Ca 2ϩ . This might suggest that a second entry channel was activated by store depletion. Application of 25 M GEA3162 also clearly induced Sr 2ϩ entry (Fig. 6B), and clearly, the emptying of pools activated both the rate and extent of entry, consistent with the results on Ca 2ϩ entry (20). The slightly slower deactivation of Sr 2ϩ entry may result from less efficient pumping of Sr 2ϩ out of the cell and/or a difference in the relative ability of Sr 2ϩ to effect deactivation of entry channels. We did not observe any significant difference in the Sr 2ϩ concentration dependence of entry activated by pool emptying as opposed to GEA3162 (not shown). However, this result was significantly different from results obtained with Ba 2ϩ (Fig. 7). In this experiment, the concentration dependence of externally applied Ba 2ϩ entry activated by pool emptying was compared with that for Ba 2ϩ entry activated by GEA3162. In cells in which pools were emptied, removal followed by the readdition of Ca 2ϩ caused the familiar overshoot of Ca 2ϩ entry (Fig. 7). Further removal of Ca 2ϩ followed by the addition instead of Ba 2ϩ resulted in Ba 2ϩ entry, which, as for Sr 2ϩ , was also detectable by fura-2 ratiofluorimetry. The entry appeared long-lasting, since Ba 2ϩ is a poor Ca 2ϩ pump substrate (31) and, hence, was not removed from cells by the plasma membrane pump even if the entry mechanism did become activated (compare with Sr 2ϩ and Ca 2ϩ in Figs. 5A and 6B, respectively). Under these conditions, only very slow entry of Ba 2ϩ was observed at 0.1 mM Ba 2ϩ , although entry was larger and more rapid with higher levels of Ba 2ϩ (Fig. 7A). If the experiment was undertaken with pool-filled cells in the absence of Ca 2ϩ , the addition of 25 M GEA3162 activated entry when Ba 2ϩ was subsequently added (Fig. 7B). However, the Ba 2ϩ dependence of entry appeared measurably different. Now, even 0.02 mM Ba 2ϩ resulted in significant entry. From analyses from several experiments using ranges of Ba 2ϩ , it appeared that the rate of store-operated Ba 2ϩ entry was half-maximal at approximately 1 mM Ba 2ϩ , whereas the rate of GEA3162-dependent Ba 2ϩ entry was activated halfmaximally with approximately 0.1 mM Ba 2ϩ .
The results with Ba 2ϩ indicated a significant difference in the apparent selectivity for passage of cations activated by store emptying as opposed to GEA3162. In other studies on Ca 2ϩ entry, criteria for defining differences between putative entry channels have rested on the effectiveness of La 3ϩ in blocking passage of Ca 2ϩ ions (35). Although we noted that 10 M La 3ϩ was sufficient to block both pool emptying as well as NO donor-activated entry, a more careful analysis of the La 3ϩ dependence of such blockade again revealed a significant difference between the two modes of activation of entry (Fig. 8). The experimental approach was similar to that used in Fig. 7, except varying concentrations of La 3ϩ were added either after pool emptying and subsequent Ca 2ϩ removal (Fig. 8A) or following 25 M GEA3162 addition in the absence of Ca 2ϩ (Fig.  8B). The Ca 2ϩ entry activated by subsequent Ca 2ϩ addition to the pool-emptied cells was completely blocked by 10 M La 3ϩ . Indeed, 1 M La 3ϩ induced a large decrease in the rate of Ca 2ϩ entry, and even 0.1 M La 3ϩ had a significant inhibitory action (Fig. 8A). From other experiments, induction of a half-maximal rate of Ca 2ϩ entry was observed with approximately 0.5 M La 3ϩ . After GEA-dependent activation of entry by Ca 2ϩ readdition (Fig. 8B), whereas Ca 2ϩ entry was blocked by 10 M La 3ϩ , the sensitivity to La 3ϩ appeared measurably different from the store-operated entry. Thus, 1 M La 3ϩ had only a small effect, and the entry of Ca 2ϩ was blocked half-maximally at approximately 5 M La 3ϩ .
Overall, our conclusions concerning the nature and relationship of the two modes of activating Ca 2ϩ entry present an interesting picture. The S-nitrosylation-activated entry is stimulated by both of the two conditions that define the operation of store-activated Ca 2ϩ entry, Ca 2ϩ pool emptying and transient external Ca 2ϩ removal. Moreover, when both mechanisms are simultaneously activated by external Ca 2ϩ addition, their effects can appear nonadditive. Whereas the effects of cytoskeletal reorganization by calyculin A indicate that the mode of activation of the entry mechanisms may be quite distinct, we could reconcile such a difference by considering a possible direct action of S-nitrosylation on the channel, which might circumvent the process of pool-emptying. If this were the case, then the action of pool emptying might help to change the conformation of the channel to increase the availability of a reactive thiol toward nitrosylation or alkylation. In addition, transient Ca 2ϩ removal may also cause reconfiguration of the channel into a more susceptible conformation. However, careful analysis of divalent cation selectivity and La 3ϩ blockade does reveal a significant difference between the two modes of Ca 2ϩ entry. A difference in ion conductivity is difficult to reconcile with the premise that both activities result from activation of a single channel, even though recent evidence for the "slip-mode" operation of Na ϩ channels (36) provides some precedent for such changes. At this stage, we might conclude that the two entry mechanisms are distinct but related. This conclusion may derive strength from other observations. Thus, we have observed that in human embryonic kidney HEK293 cells, in which store-operated Ca 2ϩ entry can be activated, the entry of Ca 2ϩ is not stimulated by GEA3162 or alkylators. 2 This may underscore an impression that has been previously suggested (2,4,6) that store-operated Ca 2ϩ entry differs significantly between cells and may represent a family of distinct channel proteins and/or different association with regulatory proteins. Interest in the TRP family of channels as potential members of the store-operated family has revealed at least six different gene products that vary significantly with respect to their ion selectivity and, more importantly, their possible relationship to pool emptying (33,34,37,38). We utilized the HEK293 cell lines stably transfected by Zhu et al. (35) to express the TRP-3 channel protein. 2 In these cells as well as the parent and control-transfected cells, the lack of action of either GEA3162 or alkylators indicated that the TRP-3 channel was not a likely target for activation of Ca 2ϩ entry by S-nitrosylation. These results notwithstanding, it is interesting to consider that very significant differences in both the conductance properties and the ability to couple to store depletion were noted by Xu et al. (39) for operation of Drosophila-derived transient receptor potential (TRP) and TRP-like (TRPL) proteins cotransfected into mammalian cells. Thus, these experiments suggested that the coassembly of channel monomers into multimeric channels could confer quite different properties related to the relative makeup of the distinct subunits within the channel assemblies. Thus, it is possible that differences in physiological operation (for example, pool coupling and desensitization by Ca 2ϩ ), susceptibility to S-nitrosylation, as well as ion selectivity among different cells may all be related to the relative expression of an extended family of related but distinct channel proteins.