In this study, we have described complex effects of 2-APB on the amplitude and kinetic behaviour of ICRAC. These include an increase of current amplitude (potentiation) and speeding of inactivation by low concentrations of 2-APB (< 5 μm), as well as inhibition of the current and removal of fast inactivation by higher concentrations (≥ 5 μm). Importantly, none of these effects appeared to occur through the IP3 receptor, as previously postulated by the conformational coupling hypothesis for store-operated Ca2+ entry. In the sections that follow, we discuss some possible underlying mechanisms and implications of these phenomena.
2-APB exerts complex effects on the amplitude and kinetics of ICRAC
We were surprised to find that 2-APB had much more complex effects on ICRAC than only the slow inhibition reported previously (Braun et al. 2001). Low doses (1–5 μm) reversibly enhanced the amplitude of ICRAC by up to fivefold and slightly increased the rate of fast Ca2+-dependent inactivation. At higher doses (≥ 5 μm) of the drug, the slow inhibition of ICRAC also became apparent, along with the disappearance of fast inactivation. All of these effects were seen in Jurkat, RBL and DT40 cells; thus, they appear to be independent of the cell type (T cells, mast cells and B cells) and species (human, rat and chicken), suggesting that they may apply to CRAC channels in general. Thus, the reason why potentiation was not observed in a previous study of RBL cells (Braun et al. 2001) probably stems not from the cell type but from a combination of other factors. The amplitude of potentiation was less pronounced in RBL cells than in Jurkat or DT40 cells (compare Fig. 8C with Fig. 8A and Fig. 3A), and this smaller effect may have been exacerbated in other studies by the use of high concentrations of 2-APB (100 μm), which produce a predominant inhibitory effect, and by pre-application of 2-APB rather than acute application after ICRAC had been activated.
The degree to which these 2-APB effects are specific for CRAC channels is mostly unknown. The inhibitory effects of 2-APB seem to be relatively non-specific, as they apply to IP3 receptors (Sugawara et al. 1997), endogenous store-operated channels in HEK 293 cells (Ma et al. 2000) and the mitochondrial Ca2+ release machinery (Fig. 1B) in addition to ICRAC. However, to our knowledge, 2-APB is the first example of a pharmacological agent that can enhance the activation of ICRAC beyond the level achieved by complete store depletion. Because Ca2+ entry through CRAC channels is a key requisite step for the activation of T cells by antigen, potentiation of ICRAC may be useful as a novel therapeutic approach to boost activation of immune cells in immunodeficient individuals. In addition, potentiation by 2-APB may prove to be a useful diagnostic tool for testing putative CRAC channel genes and probing the physiological functions of CRAC channels in vivo.
Possible mechanisms for the potentiation of ICRAC by 2-APB
In resting cells in which Ca2+ stores were full and which displayed no Ca2+ influx through CRAC channels, low doses of 2-APB were ineffective in raising [Ca2+]i or activating ICRAC. Thus, rather than activating ICRAC by itself, 2-APB appears to boost the activity of CRAC channels once they are activated by store depletion (Fig. 3E). This potentiation does not result from an increased emptying of Ca2+ stores, since 2-APB increases ICRAC severalfold even after treatment with TG, which is known to deplete the ionomycin-releasable stores in Jurkat cells by > 90 % (see Fig. 1A of Hoth et al. 1997).
Although 2-APB does reduce the level of Ca2+-dependent fast inactivation at doses as low as 5 μm, this effect is not responsible for potentiation of the current. Conditions that limit or prevent fast inactivation, including holding potentials more positive than +30 mV or the use of Ba2+ as the charge carrier, did not reduce the amount of potentiation by 2-APB (Fig. 7). In principle, 2-APB could act by interfering with other known forms of ICRAC inactivation, such as slow Ca2+-dependent inactivation (Zweifach & Lewis, 1995a) or kinase-dependent inactivation (Parekh & Penner, 1995). However, this seems unlikely, as both processes would be minimized by our recording conditions (10 mm EGTA and 0 ATP in the recording pipette).
In general, potentiation of ICRAC could occur through increases in the number of activatable channels (N), the single-channel current (i) or the open probability of the channels (Po). Without single-channel recordings, it is difficult to discriminate between these possible mechanisms, but we can make some educated guesses by exploiting the changes in fast inactivation that accompany ICRAC potentiation. Previous work has shown that ICRAC inactivation is influenced by the local [Ca2+]i around individual CRAC channels (Zweifach & Lewis, 1995b). 2-APB at low concentrations does not induce inactivation by itself, since Ba2+ currents do not inactivate even in the presence of the drug (Fig. 7C); instead, a reasonable conclusion is that it increases the speed of fast Ca2+-dependent inactivation, possibly by producing an increase in the local [Ca2+]i around CRAC channels. Fast inactivation is strongly dependent on the amplitude of i but not on N (Zweifach & Lewis, 1995b). Thus, if potentiation were due to an increase in i, then one would expect a significant increase in the extent of inactivation, particularly under conditions where the current increases significantly (e.g. by a factor > 2; see Fig. 5). This was not observed, suggesting that i is probably not affected by 2-APB. In addition, because CRAC channels inactivate independently (Zweifach & Lewis, 1995b), an increase in N would not be expected to affect inactivation. It might be argued that overlap of Ca2+ microdomains between adjacent CRAC channels might increase with N, undermining channel independence and increasing inactivation. To test this possibility, we reduced the current with SKF 96365, reasoning that this would reduce the overlap of microdomains and therefore reduce the inactivation if overlap existed. We found that, in the presence of a potentiating dose of 2-APB, a half-blocking dose of SKF 96365 did not affect the rate of inactivation (data not shown), making the overlap of domains unlikely.
If 2-APB does not influence N or i, the most likely explanation is that 2-APB potentiates ICRAC by increasing Po. The resulting increase in accumulation of intracellular Ca2+ at the inactivation sites would be expected to enhance inactivation, though perhaps to a smaller degree than would be expected from a comparable increase in i. A mechanism of this sort has been proposed to underlie the speeding of inactivation of voltage-gated L-type Ca2+ channels by Bay K 8644, which increases the Po of these channels (Noceti et al. 1998). Further experiments on ICRAC at the single-channel level will be needed to resolve this issue.
Inhibition of ICRAC and the removal of inactivation by 2-APB
At concentrations ≥ 5 μm, 2-APB slowly inhibited ICRAC and removed fast inactivation. The reversal of these effects after 2-APB was washed out of the bath was even slower and was often incomplete. In this sense, the inhibition by 2-APB is similar to that by imidazole compounds like SKF 96365 or econazole (Franzius et al. 1994; Christian et al. 1996). It is not known whether 2-APB (or SKF 96365) acts by blocking the channel pore directly, or by altering channel gating. Our data suggest that 2-APB does not act through a simple 1:1 interaction with the CRAC channel pore. The on-rate for inhibition, calculated from the time constant of ICRAC inhibition in Fig. 6B, and K1/2 for current inhibition (10 μm, Fig. 4B) was 1/500 μm−1 s−1. Thus, a simple 1:1 binding interaction would predict an off-rate of ∼1/50 s−1, or a time constant of recovery of ∼50 s. This is much faster than the observed recovery time (22 % recovery at 160 s; estimated recovery time constant of > 600 s). The discrepancy cannot be explained by slow washout of the drug, as the potentiation effect of 2-APB reversed rapidly after drug removal (Fig. 3A). Instead, these results suggest that the mechanism of inhibition is more complex than simple pore blockade, and that the recovery from inhibition is not limited by the dissociation rate of 2-APB from its target.
2-APB inhibited both the amplitude of ICRAC and the extent of fast inactivation with similar time courses, raising the intriguing possibility that these two effects may be causally linked. It is possible, for example, that 2-APB at high concentrations interrupts the coupling between the CRAC channel and a component that confers fast inactivation, and the activity of the CRAC channel declines because this component is necessary for maintaining the channel in an activatable state. Such a mechanism might also explain the kinetic discrepancy discussed above, if recovery of the inactivation component were much slower than the unbinding of 2-APB. Further experiments on the molecular underpinnings of fast inactivation may give clues as to the mechanism of ICRAC inhibition by 2-APB.
IP3 receptors are not needed for ICRAC activation or inactivation
The mechanism underlying the activation of store-operated Ca2+ channels has been an area of active debate and, despite considerable efforts, no consistent picture has emerged. The leading candidate hypotheses include activation by a diffusible factor released from depleted stores, fusion of store-operated channel-laden vesicles with the plasma membrane, and direct activation of store-operated channels by conformational coupling to IP3 receptors in the ER membrane (for reviews, see Putney & McKay, 1999; Prakriya & Lewis, 2001a). The conformational coupling hypothesis has received much recent support from evidence that IP3 receptors bind directly to some members of the TRP family, thought by some to be the molecular correlates of CRAC channels (Kiselyov et al. 1998, 1999a; Boulay et al. 1999). Moreover, inhibitors of IP3 receptors, such as 2-APB and heparin, also inhibit store-operated Ca2+ entry and the currents mediated by the putative store-operated channel TRP3 (Kiselyov et al. 1998; Ma et al. 2000). It has been suggested that fast inactivation of CRAC channels may arise through the inhibitory effects of Ca2+ on IP3 receptors to which they are coupled (Berridge, 1995).
Our data are inconsistent with a necessary role for IP3 receptors in the activation and inactivation of ICRAC. First, we found that heparin failed to affect activation or the maintenance of ICRAC in store-depleted cells, even at a concentration that completely prevented the ability of IP3 to activate ICRAC. A similar result was described recently in RBL cells (Broad et al. 2001). In contrast, currents mediated by recombinant TRP3 channels in excised patches from HEK 293 cells or endogenous ‘CRAC-like’ channels from A431 cells require addition of IP3 and are blocked potently by heparin (Kiselyov et al. 1998, 1999b). Thus, the ability to activate ICRAC through passive depletion, coupled with the inability of heparin to prevent this, suggests that CRAC channels differ from these other channels in not requiring even background levels of IP3 for activation. Second, the activation and properties of ICRAC were apparently normal in DT40 cells in which all three subtypes of IP3 receptor had been knocked out by homologous recombination. These results confirm and extend earlier work by Sugawara et al. (1997) showing that TG was able to induce Ca2+ entry in these cells. Thus, our results show that CRAC channels do not require IP3 receptors or IP3 for their activation. However, these results do not rule out conformational coupling involving other proteins.
2-APB affects CRAC channel gating independently of IP3 receptors
In the mutant DT40 cells lacking all three subtypes of IP3 receptors, 2-APB elicited the same combination of ICRAC potentiation, inhibition and changes in fast inactivation as in wild-type cells. Thus, the actions of 2-APB on ICRAC are probably unrelated to any effect on IP3 receptors. These results therefore show that inhibition of Ca2+ entry by 2-APB cannot be used to test the conformational coupling model for ICRAC activation.
If IP3 receptors are not involved, what then is the target for 2-APB's effects on CRAC channel activity? One possible site of action we considered was mitochondria, as 2-APB inhibited Ca2+ release from these organelles (Fig. 1B). In Jurkat T cells, uptake and release of Ca2+ by mitochondria is required to enable the cells to maintain a high rate of Ca2+ entry through CRAC channels (Hoth et al. 1997). To test the possibility that the effects of 2-APB on ICRAC are an indirect consequence of modulating mitochondrial function, we treated cells with 1 μm antimycin A1 + 1 μm oligomycin to inhibit mitochondrial Ca2+ uptake. The inhibitors did not alter the ability of 2-APB to potentiate and inhibit ICRAC (data not shown). Therefore, the effects of 2-APB on ICRAC appear to be distinct from its effects on mitochondrial Ca2+ transport.
2-APB was more effective at modulating ICRAC when applied extracellularly than when it was applied intracellularly through the recording pipette. A similar result was reported for RBL cells by Braun et al. (2001), who concluded that the drug's site of action was extracellular. The interpretation of this experiment is complicated by several factors. First, 2-APB may exist in multiple forms, including a protonated and unprotonated monomer, as well as a dimer (van Rossum et al. 2000), but the relative efficacy of these various compounds in affecting Ca2+ entry in general is unknown. Second, the ability of lipophilic compounds to interact with intracellular sites when applied intracellularly via a patch pipette is a function of their membrane permeability. Thus, when applied through the recording pipette, the concentration of a hydrophobic form of 2-APB near the plasma membrane may be much lower than expected, due to the fact that the rate-limiting diffusion step for such a hydrophobic compound may be across the tip of the pipette rather than across the plasma membrane. We reasoned that increasing the protonation of the drug should reduce permeability to some degree, and this should reduce the efficacy of extracellular drug if its site of action is intracellular. This approach has been applied previously in studies of membrane-permeant K+ channel blockers. For example, methadone is known to block K+ channels at an intracellular site, yet it is effective only when applied from outside the cell. Acidification of the extracellular medium was shown to inhibit the blocking action of methadone by reducing the concentration of the uncharged, lipophilic species (Horrigan & Gilly, 1996). In our experiments, reducing extracellular pH from 7.4 to 6.4 failed to decrease either potentiation or inhibition by 2-APB, implying that lowering the concentration of the unprotonated (and presumably membrane-permeant) form of 2-APB had no effect. The relative amounts of the protonated and unprotonated forms of 2-APB will depend on its pKa, which is unknown, but is likely to be determined by its ethylamine group (pKa= 10.8; Weast & Lide, 1989). Based on this pKa value, one would predict that only 0.04 % of the 2-APB will be unprotonated at pH 7.4, but that this fraction will decrease by a factor of 10 at pH 6.4. Thus, the lack of an effect from lowering the pH (Fig. 10C) is consistent with an extracellular but not an intracellular site of action. However, given the various complications outlined above, a direct examination of this issue using membrane-impermeant variants of 2-APB will be needed to identify firmly the active species of 2-APB and its site of action.
The opposing effects of 2-APB on ICRAC are reminiscent of the effects of the plant alkaloid ryanodine on the sarcoplasmic reticulum Ca2+ release channel in cardiac muscle. At submicromolar (1–50 nm) concentrations, ryanodine stabilizes the Ca2+ release channel in a subconductance state with a dramatically higher Po, whereas at micromolar (10–1000 μm) concentrations, ryanodine closes the channel (Buck et al. 1992; Humerickhouse et al. 1993). These effects are manifested as increases in Ca2+ flux and muscle contraction by dilute ryanodine and inhibition of Ca2+ flux and muscle contraction by concentrated ryanodine. Binding studies indicate that the Ca2+ release channel has a high-affinity site, thought to mediate the stimulatory effects of ryanodine, and one or more low-affinity sites that mediate channel closure (Sutko et al. 1997). Thus, by analogy, it is possible that 2-APB has two separate binding sites, a high-affinity site that mediates ICRAC potentiation and a low-affinity site that produces inhibition. Alternatively, different species of 2-APB may be responsible for the different effects: monomers, which are more prevalent at low concentrations, may mediate potentiation, whilst dimers, which are favoured at higher concentrations, may be responsible for inhibition. Further studies will be needed to resolve these issues.
In summary, we found that 2-APB elicits complex effects on ICRAC. Low concentrations of 2-APB reversibly enhanced ICRAC, whereas high concentrations also inhibited ICRAC. 2-APB also altered the gating of ICRAC by removing fast Ca2+-dependent inactivation. It was surprising to find that so many different properties of CRAC channels were affected, and an important question is whether the various effects of 2-APB are linked in some way or whether they are truly independent. Given that 2-APB affects the function of so many Ca2+ signalling proteins (IP3 receptors, CRAC channels, TRP channels and the mitochondrial Ca2+ release mechanism), the drug clearly has limited utility as a specific inhibitor of store-operated Ca2+ entry. However, some effects of 2-APB (e.g. potentiation and removal of inactivation) may be sufficiently unique that identification of the parts of the molecule that are responsible for these effects and the proteins with which they interact could provide new clues to the mechanisms underlying activation, deactivation and inactivation of CRAC channels.