Energized mitochondria increase the dynamic range over which inositol 1,4,5-trisphosphate activates store-operated calcium influx

Abstract

In eukaryotic cells, activation of cell surface receptors that couple to the phosphoinositide pathway evokes a biphasic increase in intracellular free Ca²⁺ concentration: an initial transient phase reflecting Ca²⁺ release from intracellular stores, followed by a plateau phase due to Ca²⁺ influx. A major component of this Ca²⁺ influx is store-dependent and often can be measured directly as the Ca²⁺ release-activated Ca²⁺ current (I_CRAC). Under physiological conditions of weak intracellular Ca²⁺ buffering, respiring mitochondria play a central role in store-operated Ca²⁺ influx. They determine whether macroscopic I_CRAC activates or not, to what extent and for how long. Here we describe an additional role for energized mitochondria: they reduce the amount of inositol 1,4,5-trisphosphate (InsP₃) that is required to activate I_CRAC. By increasing the sensitivity of store-operated influx to InsP₃, respiring mitochondria will determine whether modest levels of stimulation are capable of evoking Ca²⁺ entry or not. Mitochondrial Ca²⁺ buffering therefore increases the dynamic range of concentrations over which the InsP₃ is able to function as the physiological messenger that triggers the activation of store-operated Ca²⁺ influx.

Introduction

In non-excitable cells, the second messenger inositol 1,4,5-trisphosphate (InsP₃) evokes Ca²⁺ release from intracellular stores followed by Ca²⁺ influx across the plasma membrane (Putney, 1986; Berridge, 1993). One major route for this Ca²⁺ entry is through store-operated Ca²⁺ channels (SOCs), which are activated by the process of emptying the intracellular Ca²⁺ stores (Parekh and Penner, 1997). Ca²⁺ influx through SOCs is required not only for refilling the intracellular stores but also for regulating a host of physiological processes including secretion, gene transcription and cell proliferation (Parekh and Penner, 1997).

Although several types of SOC have been described, the best characterized to date are the Ca²⁺ release-activated Ca²⁺ (CRAC) SOCs, which give rise to a highly selective whole-cell Ca²⁺ current called I_CRAC (Hoth and Penner, 1992; Parekh and Penner, 1997). I_CRAC can be measured directly using the whole-cell patch-clamp technique. Until recently, I_CRAC was studied routinely in the presence of very high concentrations of Ca²⁺ chelators (mM concentrations of EGTA/BAPTA) in the recording pipette. Such strong Ca²⁺ buffering was used because the current could not be detected with weaker, more physiological levels of Ca²⁺ buffer (0.1 mM EGTA/BAPTA), and this inability to record I_CRAC was attributed to Ca²⁺-dependent inactivation of the underlying CRAC channels. However, we and others have shown that this explanation cannot account for the inability to record I_CRAC in weak Ca²⁺ buffer (Broad et al., 1999; Fierro and Parekh, 2000). We have found that SERCA pumps, which refill the stores, are very powerful in rat basophilic leukaemia-1 (RBL-1) cells (Fierro and Parekh, 1999) and that InsP₃ is unable to deplete stores sufficiently for whole-cell (macroscopic) I_CRAC to develop in weak Ca²⁺ buffer unless the SERCA pumps are inhibited (Fierro and Parekh, 2000; Bakowski and Parekh, 2001). Recently, we have demonstrated that I_CRAC can be activated in weak buffer, even when SERCA pumps are active, provided Ca²⁺ uptake into mitochondria is functional (i.e. when mitochondria are in an energized state in the whole-cell configuration; Gilabert and Parekh, 2000). Furthermore, Ca²⁺-dependent slow inactivation of CRAC channels was reduced when mitochondrial Ca²⁺ uptake was operational. Hence mitochondria are important determinants of store-operated Ca²⁺ influx, determining whether macroscopic I_CRAC activates in weak buffer, to what extent and for how long (Gilabert and Parekh, 2000).

Mitochondrial Ca²⁺ buffering has quite marked effects on Ca²⁺ release to submaximal concentrations of InsP₃. In permeabilized hepatocytes, mitochondrial Ca²⁺ uptake suppresses the positive feedback actions of cytosolic Ca²⁺ on adjacent InsP₃ receptors and this results in less Ca²⁺ release to a submaximal InsP₃ concentration (Hajnoczky et al., 1999). A similar conclusion was reached by Boitier et al. (1999), who found that mitochondrial Ca²⁺ uptake reduced the rate of propagation of Ca²⁺ waves in astrocytes. In Xenopus oocytes, on the other hand, mitochondrial Ca²⁺ uptake increased the frequency of Ca²⁺ waves and this was thought to reflect a reduction in the extent of Ca²⁺-dependent inactivation of InsP₃ receptors (Jouvaille et al., 1995).

Since many Ca²⁺-dependent processes are activated by modest increases in intracellular InsP₃ levels, we have now investigated the effects of mitochondrial Ca²⁺ uptake on the ability of submaximal concentrations of InsP₃ to activate I_CRAC in weak Ca²⁺ buffer. We find that respiring mitochondria enhance the extent of I_CRAC, thus enabling moderate levels of InsP₃ to be more effective in triggering Ca²⁺ influx. In addition, the threshold concentration of InsP₃ that triggers I_CRAC is reduced in the presence of energized mitochondria. Mitochondrial Ca²⁺ uptake therefore increases the dynamic range over which InsP₃ is an effective stimulus for Ca²⁺ entry and may determine the efficacy of relatively weak stimuli in promoting store-dependent Ca²⁺ influx.

Results

A range of InsP₃ concentrations fail to activate I_CRAC consistently in weak Ca²⁺ buffer

When RBL-1 cells are dialysed with a patch pipette containing strong Ca²⁺ buffer (10 mM EGTA) and a maximally effective concentration of InsP₃ (30 µM), virtually all cells respond by generating a relatively large store-operated Ca²⁺ current (Fierro and Parekh, 2000; Gilabert and Parekh, 2000; Bakowski and Parekh, 2001). However, if this experiment is repeated in the presence of more physiological levels of intracellular Ca²⁺ buffering (0.1 mM EGTA), the same concentration of InsP₃ is largely ineffective. The majority of cells (>75%) fail to develop any detectable I_CRAC and, for those that do, the current is substantially smaller than in strong buffer. To examine whether lower, and presumably more physiological, concentrations of InsP₃ were similarly ineffective in weak Ca²⁺ buffer, we dialysed RBL-1 cells with different concentrations of InsP₃ (in 0.1 mM EGTA). In Figure 1A1, the time course of I_CRAC in response to 5 µM InsP₃, a concentration that is just maximal in strong buffer (Parekh et al., 1997), is depicted (open circles in Figure 1A1). Figure 1A2 shows the current–voltage relationship, taken at 100 s. InsP₃ at 5 µM failed to evoke any detectable I_CRAC in this cell. Averaged data from several cells, exposed to different concentrations of InsP₃, are shown in Figure 1B, and the fraction of responding cells is depicted in Figure 1C. In Figure 1B, cells that responded have been pooled together with those that did not. Concentrations of InsP₃ ≤3 µM consistently failed to evoke any discernible current. These concentrations nevertheless do release Ca²⁺ from the stores (Parekh et al., 1997). For higher InsP₃ concentrations, I_CRAC was very small and only ∼25% of the cells responded. For those cells that did respond, I_CRAC was still several fold smaller than the amplitude in strong buffer (–0.51 pA/pF compared with –2.85 pA/pF in strong buffer for 30 µM InsP₃). Hence, over a range of concentrations, InsP₃ is a very weak activator of I_CRAC in the presence of weak intracellular Ca²⁺ buffer.

Fig. 1. Energized mitochondria increase the ability of InsP₃ to activate I_CRAC in weak intracellular Ca²⁺ buffer. (A1) Time course of I_CRAC to 5 µM InsP₃ in the absence (open circles) and presence (filled circles) of mitochondrial cocktail solution. (A2) Current–voltage relationships for the recordings shown in (A1), taken after 100 s. (B) Amplitude of I_CRAC (measured at –80 mV from the voltage ramps and normalized for cell capacitance) is plotted against InsP₃ concentration included in the recording pipette. Open and filled circles correspond to experiments in the absence and presence of the mitochondrial cocktail solution, respectively. (C) The percentage of responding cells is plotted against InsP₃ concentration. Note that, at 3 µM InsP₃, no cell responds in the absence of the mitochondrial cocktail whereas 50% do so when the cocktail is present.

InsP₃ still fails to activate I_CRAC consistently when K⁺ is the major intracellular cation

In the preceding experiments, we used Cs⁺ as the dominant cation in our pipette solution. For example, RBL cells express a GTP-dependent K⁺ conductance (McCloskey and Cahalan, 1990), and GTP is a component of the intracellular cocktail used to maintain mitochondria in an energized state (Gilabert and Parekh, 2000; see below). The GTP-dependent K⁺ current is not permeable to Cs⁺, and, therefore, in the presence of Cs⁺, it would not contaminate our recordings. However, under physiological conditions, K⁺ is the major intracellular cation. It is possible that counter-movement of K⁺ is required in order to sustain Ca²⁺ release from the stores, and Cs⁺ might be unable to substitute for K⁺ in this action.

Both InsP₃-gated channels and ryanodine receptors are non-selective cation channels permeable to monovalent and divalent cations, with a higher permeability for divalents (Bezprozvanny and Ehrlich, 1995). The selectivity profile is believed to be very similar for these two Ca²⁺ release channels, which is not unexpected as there is significant homology between the genes encoding these proteins. Although there is little information on the relative permeabilities of Cs⁺ and K⁺ for the InsP₃ receptors, Cs⁺ is quite permeable through ryanodine-sensitive channels. It is more permeable than K⁺ in the channels from skeletal muscle (Smith et al., 1988) and is 0.61 times that of K⁺ in cardiac muscle (Lindsay et al., 1991). Nevertheless, we compared the ability of 30 µM InsP₃ to activate I_CRAC in weak buffer in the presence of either Cs⁺ or K⁺. With Cs⁺, the mean I_CRAC was –0.34 ± 0.04 pA/pF and in K⁺ it was –0.41 ± 0.05 pA/pF (six cells for each condition), and the amplitudes were not significantly different (p >0.3). These results are entirely consistent both with our previous findings in which carbachol, acting via an increase in InsP₃, evoked robust Ca²⁺ release after cells had been dialysed with Cs⁺ (see Figure 3 of Parekh et al., 1997), and with a recent report by Hermosura et al. (2000). These authors showed that the size of the carbachol-evoked transient in RBL cells was very similar in intact cells (where K⁺ is the major intracellular cation) and after extensive dialysis of the cytosol with Cs⁺ (Figure 4 of Hermosura et al., 2000).

Respiring mitochondria potentiate I_CRAC and reduce the threshold for activation by InsP₃

We have found that mitochondrial Ca²⁺ uptake is essential in order for 30 µM InsP₃ to activate I_CRAC in weak Ca²⁺ buffer (Gilabert and Parekh, 2000). Whereas most cells fail to respond to InsP₃ in weak buffer, and for those that do the current is very small (see above), dialysis with a solution designed to maintain mitochondria in an energized state (mitochondrial cocktail solution, see Materials and methods) results in virtually all cells responding to produce a robust I_CRAC. The effects of the cocktail are due entirely to an action on mitochondria because manoeuvres that suppress mitochondrial Ca²⁺ uptake abolish the enhancing effects of the cocktail (Gilabert and Parekh, 2000). To investigate whether mitochondrial Ca²⁺ uptake similarly potentiates I_CRAC to more moderate levels of InsP₃, we dialysed cells with different concentrations of InsP₃ in weak Ca²⁺ buffer but now also in the presence of the mitochondrial cocktail. The results were dramatic. Whereas 5 µM InsP₃ generally failed to activate I_CRAC in weak Ca²⁺ buffer (Figure 1A1, open circles), it evoked a sizeable current in the presence of the cocktail (Figure 1A2, filled circles). The I–V relationship is shown in Figure 1A2. The current is inwardly rectifying, voltage-independent (over 50 ms) and reverses at potentials >50 mV, which are the hallmarks of I_CRAC (Parekh and Penner, 1997). The extent of I_CRAC to different InsP₃ concentrations in the absence (open circles) and presence (filled circles) of the mitochondrial cocktail solution is compared in Figure 1B, where data from responders and non-responders have been pooled together. The percentage of responding cells for each condition is shown in Figure 1C. Kinetic features of the current for the two different conditions are summarized in Figure 2A1 and A2. Several striking differences are apparent in the presence of energized mitochondria. (i) The size of I_CRAC is potentiated over the range of concentrations of InsP₃ that evoke a response (5–30 µM; Figure 1B). (ii) The fraction of cells that respond over this concentration range increases substantially in the presence of respiring mitochondria (Figure 1C). (iii) Whereas 3 µM InsP₃ consistently fails to evoke I_CRAC and is hence a subthreshold concentration in the absence of mitochondrial cocktail, around half of the cells respond to this dose of InsP₃ when mitochondria are energized (Figure 1B and C). (iv) The only kinetic parameter that changes with InsP₃ concentration is the delay before I_CRAC starts to acti vate (Figure 2A1). However, once the current is initiated, then it develops at the same rate (Figure 2A2). (v) If we consider the amplitude of I_CRAC in the presence of cocktail for responding cells only, then this was quite similar for all InsP₃ concentrations that evoked a response [–0.44 ± 0.08 (n = 5), –1.06 ± 0.25 (n = 7), –0.81 ± 0.16 (n = 6) and –1.18 ± 0.16 (n = 21) pA/pF for 3, 5, 10 and 30 µM InsP₃, respectively; the only significant difference was between 3 and 30 µM InsP₃]. Collectively, these results indicate that the threshold concentration of InsP₃ required to evoke I_CRAC is reduced in the presence of energized mitochondria. Mitochondrial Ca²⁺ uptake therefore increases the sensitivity of store-operated Ca²⁺ influx to InsP₃. However, lowering the InsP₃ concentration further (0.5 µM) failed to activate I_CRAC in weak buffer, even in the presence of the mitochondrial cocktail.

Fig. 2. Effects of mitochondrial Ca²⁺ buffering on the kinetics of I_CRAC under conditions where SERCA pumps are active. The delay before I_CRAC activates (A1) and the time to peak (A2) are plotted against InsP₃ concentration for cells with energized mitochondria. The time to peak was corrected for the delay. (B1) The time course of I_CRAC is shown for a cell dialysed with 30 µM InsP₃ in weak Ca²⁺ buffer in the absence (open circles) and presence (filled circles) of mitochondrial cocktail. Note that the amplitude of the current was similar for the two cells, but I_CRAC declined much more quickly when mitochondria were not energized. (B2) The extent of decline of I_CRAC is compared for cells dialysed in the absence (open bar) and presence (filled bar) of mitochondrial cocktail. (B3) The extent of decline of I_CRAC in the presence of cocktail is plotted against the different InsP₃ concentrations. *p <0.05 and ***p <0.001.

The relationship between InsP₃ concentration and extent of I_CRAC is highly non-linear in weak Ca²⁺ buffer

We had reported previously that the relationship between InsP₃ concentration and the amplitude of I_CRAC was highly non-linear in strong Ca²⁺ buffer, with a Hill coefficient of 12 (Parekh et al., 1997; Glitsch and Parekh, 2000). This supralinear relationship still held in the presence of moderate Ca²⁺ buffering (Glitsch and Parekh, 2000). However, it was not clear whether the steep relationship was also valid in the presence of weak Ca²⁺ buffer because the current could not be measured consistently under those conditions. Since we can record robust I_CRAC in weak Ca²⁺ buffer provided mitochondria are energized, we have been able to address this important issue directly. As shown in Figure 1B, concentrations of InsP₃ <3 µM fail to evoke any detectable I_CRAC, whereas 5 µM InsP₃ generates maximal current. Fitting this dose–response curve with a modified Hill equation yielded a Hill coefficient of 17. Because the relationship is so steep, the Hill coefficient (derived from the fit) is only an approximation. However, the key point is that I_CRAC is related supralinearly to InsP₃ concentration in the presence of physiological levels of intracellular Ca²⁺ buffering and energized mitochondria. Once the threshold concentration of InsP₃ is exceeded, then only small further increases in InsP₃ will result in maximal activation of Ca²⁺ influx.

Kinetics of decay of I_CRAC in the presence of energized mitochondria

Mitochondrial Ca²⁺ uptake reduces both the rate and extent of Ca²⁺-dependent slow inactivation of CRAC channels in RBL cells (Gilabert and Parekh, 2000) and Jurkat T lymphocytes (Hoth et al., 2000). In order to isolate this inactivation mechanism, Ca²⁺-dependent store refilling by SERCA pumps had to be suppressed. Because SERCA pumps are very powerful in RBL cells, we wanted to see whether mitochondria could prolong the duration of I_CRAC even when these pumps were active. To this end, we compared the time course of I_CRAC following activation by InsP₃ in weak Ca²⁺ buffer in the absence and presence of the mitochondrial cocktail. Most cells failed to generate I_CRAC to InsP₃ in the absence of cocktail, whereas for the minority that did so the current was transient. Figure 2B1 shows a typical response to InsP₃ in the absence of cocktail (open circles, control) and the extent of decline of the current is summarized in the histogram of Figure 2B2. I_CRAC declined almost completely within 200 s. However, in the presence of cocktail, the current decayed much more slowly such that the amplitude of I_CRAC at times >100 s was significantly larger than the case when mitochondria were not maintained in an energized state (Figure 2B1 and B2). Therefore, under conditions where SERCA pump activity is maintained, mitochondrial Ca²⁺ buffering is still an important factor that prolongs the time course of I_CRAC. There was some variability in the extent of decay for I_CRAC in the presence of energized mitochondria between different cells (see also Gilabert and Parekh, 2000), but overall the decay was not significantly different between the various InsP₃ concentrations (Figure 2B3).

The facilitatory effects of mitochondria are suppressed by moderate concentrations of Ca²⁺chelator

Thin-section electron microscopic studies on RBL-2H3 cells have revealed that a small fraction of the mitochondrial surface is closely apposed to the endoplasmic reticulum (ER) membrane (Hajnoczky et al., 2000). This suggests that Ca²⁺ released by InsP₃ receptors on the ER might need to diffuse, at least over short distances, in the cytosol before they are taken up by mitochondria. To see whether a slow Ca²⁺ chelator could intercept the diffusing Ca²⁺ ions and what concentration of this chelator suppressed the facilitatory effects of mitochondrial Ca²⁺ uptake, we constructed dose–response curves to EGTA in the absence and presence of the mitochondrial cocktail solution. Results are summarized in Figure 3A. For control (non-cocktail-treated) cells, 30 µM InsP₃ failed to evoke any I_CRAC in 0.1 mM EGTA. EGTA at 0.35 mM was at around the threshold concentration for detection (3/9 cells failed to respond). In the presence of 0.6 mM EGTA, all cells responded and I_CRAC was ∼50% of the maximum, the maximum being obtained in the presence of 10 mM EGTA. These results are very similar to our previous findings (Glitsch and Parekh, 2000). The presence of the cocktail significantly potentiated I_CRAC in the presence of weak to moderate concentrations of EGTA (<0.6 mM EGTA; Figure 3A) and virtually all cells responded under these conditions (90% in 0.1 mM and 100% in 0.35 mM EGTA). In 0.6 mM EGTA, the current was enhanced slightly by cocktail but this was not significant (Figure 3A). No facilitatory effects of the cocktail were seen in 10 mM EGTA. These results are in good agreement with those reported by Csordas et al. (1999) who found that 0.6 mM EGTA suppressed the increase in intramitochondrial Ca²⁺ following InsP₃-mediated Ca²⁺ release in permeabilized RBL-2H3 cells. Because slow buffers such as EGTA are unable to reduce Ca²⁺ levels at distances <20 nm from open Ca²⁺-permeable channels such as InsP₃-gated channels (Neher, 1998), our results indicate that a molecular distance significantly larger than 20 nm separates the Ca²⁺ release and uptake sites on the ER and mitochondria, respectively. This is entirely consistent with the findings of Csordas et al. (1999), who calculated an average distance of 100 nm between InsP₃ receptors and mitochondrial Ca²⁺ uptake sites in permeabilized RBL-2H3 cells.

Fig. 3. Moderate concentrations of the slow Ca²⁺ chelator EGTA suppress the potentiating effects of mitochondrial cocktail on I_CRAC. (A) Amplitude of I_CRAC is plotted against EGTA concentration in the absence (open circles) and presence (filled circles) of mitochondrial cocktail. I_CRAC was potentiated by the cocktail for EGTA concentrations <0.6 mM. (B) In the presence of respiring mitochondria, 30 µM InsP₃ and 0.1 mM EGTA, inhibition of SERCA pumps by thapsigargin (2 µM) results in a further increase in the amplitude of I_CRAC. (C) Energized mitochondria potentiate the size of I_CRAC when Jurkat T cells are dialysed with InsP₃ (30 µM) and 0.1 mM EGTA. *p <0.05 and ***p <0.001.

SERCA pumps can compete with mitochondria for removal of cytosolic Ca²⁺

Inspection of Figures 1B and 3A reveals that, in energized mitochondria and weak Ca²⁺ buffer, I_CRAC is generally around –1 pA/pF. In strong buffer, the current is almost three times larger. This difference in current amplitudes could arise from some vestigial Ca²⁺-dependent inactivation of the CRAC channels in weak buffer such that the current size is reduced. Alternatively, it could reflect some Ca²⁺-dependent store refilling, implying that not all of the Ca²⁺ released by InsP₃ is taken up by mitochondria but that some of this Ca²⁺ is resequestrated into the stores. This latter scenario would constitute a form of physiological antagonism between two major Ca²⁺ clearance mechanisms in RBL cells, with mitochondria enhancing depletion of InsP₃-sensitive Ca²⁺ stores and hence activation of I_CRAC, and SERCA pumps promoting store refilling and therefore a reduction in the extent of activation of I_CRAC. To distinguish between these possibilities, we compared the size of I_CRAC between cells dialysed with InsP₃ + weak Ca²⁺ buffer + mitochondrial cocktail in the absence and presence of thapsigargin. Results are summarized in the histogram of Figure 3B. Inclusion of thapsigargin resulted in an almost 3-fold increase in the size of the current. The amplitude of I_CRAC now was similar to that seen in strong buffer (InsP₃ + 10 mM EGTA; Figure 3A). Hence, SERCA pumps can still resequestrate sufficient Ca²⁺ even in the presence of energized mitochondria such that I_CRAC cannot activate to its maximum extent. This would be consistent with the notion that mitochondria and SERCA pumps compete for Ca²⁺ and that this impacts upon the extent of store depletion and subsequent activation of I_CRAC.

Energized mitochondria potentiate I_CRAC in weak Ca²⁺ buffer in Jurkat T lymphocytes

In Jurkat T lymphocytes, like RBL-1 cells, we have found that macroscopic I_CRAC can be activated only weakly following dialysis with InsP₃ and 0.1 mM EGTA (Fierro et al., 2000) We therefore investigated whether energized mitochondria could enhance the size of the current, as is the case in RBL cells. As shown in Figure 3C, the amplitude of I_CRAC was significantly larger in the presence of the mitochondrial cocktail. These results complement the reports by Hoth et al. (1997, 2000), who found that mitochondrial Ca²⁺ buffering reduced Ca²⁺-dependent slow inactivation of CRAC channels in the Jurkat cell line. Like RBL cells, energized mitochondria in Jurkat T lymphocytes seem to increase the size of I_CRAC following dialysis with InsP₃ in weak Ca²⁺ buffer.

Discussion

Our new findings demonstrate that energized mitochondria determine whether macroscopic I_CRAC activates over a range of stimulus intensities. By reducing the threshold concentration of InsP₃ required to evoke I_CRAC, mitochondria increase the dynamic range over which this ubiquitous second messenger is able to control, via store depletion, the extent of Ca²⁺ influx. A cartoon summarizing this effect is shown in Figure 4 (see legend for explanation). The increased sensitivity to lower levels of InsP₃ may be an important factor that helps determine whether weak/moderate stimulation of cell surface receptors can promote store-operated Ca²⁺ entry. It is intriguing to consider that regulation of mitochondrial Ca²⁺ uptake and/or spatial distribution of mitochondria relative to the ER might be a novel mechanism to dictate whether weak/moderate receptor stimulation evokes Ca²⁺ influx or not.

Fig. 4. Cartoon summary of the role of mitochondria in activation of I_CRAC for moderate increases in InsP₃. (A) The resting situation, where most CRAC channels are closed. The stores are sufficiently full that I_CRAC is deactivated and Ca²⁺ that leaks out of the stores is taken back up by the SERCA pumps. (B) Moderate increases in InsP₃ evoke substantial Ca²⁺ release from the stores but sufficient Ca²⁺ is resequestrated by the pumps so that the threshold for macroscopic activation of I_CRAC is not reached. The current therefore does not activate to detectable levels. (C) In the presence of energized mitochondria, these moderate concentrations of InsP₃ are now able to activate macroscopic I_CRAC. Mitochondria take up some of the Ca²⁺ that has been released by InsP₃. This will result in less Ca²⁺ being available for the SERCA pumps, resulting in a reduction in net re-uptake. In addition, mitochondrial Ca²⁺ buffering may reduce the extent of Ca²⁺-dependent inactivation of the InsP₃ receptors, thereby enabling greater Ca²⁺ release and hence store depletion. Mitochondrial Ca²⁺ uptake will also reduce the rate and extent of Ca²⁺-dependent slow inactivation, and this will increase the size of I_CRAC as well as prolong its duration (Gilabert and Parekh, 2000).

Because relatively high concentrations of InsP₃ or receptor engagement are required to activate I_CRAC, it has been proposed that additional Ca²⁺ entry pathways may be involved during weaker levels of stimulation (Shuttleworth, 1999). Our findings, demonstrating that energized mitochondrial Ca²⁺ uptake can reduce the amount of InsP₃ that is necessary to evoke I_CRAC, suggest instead that I_CRAC can be activated by lower InsP₃ concentrations when care is taken to maintain mitochondria in an energized state. It is likely that addition to our pipette solution of further components that support respiring mitochondria may result in greater Ca²⁺ uptake by this organelle and hence a further reduction in the InsP₃ concentration required to activate I_CRAC.

In RBL-1 cells, like certain other non-excitable cells, low concentrations of InsP₃ can trigger Ca²⁺ release and reduce the amount of Ca²⁺ within the stores without evoking any Ca²⁺ influx (Parekh et al., 1997; Hartmann and Verkhratsky, 1998; Liu et al., 1998). Growing evidence from RBL-1 cells points towards a specialized subcompartment of the ER that is involved specifically in the activation of I_CRAC (Parekh et al., 1997; Broad et al., 1999; Krause et al., 1999). This store seems harder to deplete since somewhat higher concentrations of InsP₃ are required to activate Ca²⁺ influx than Ca²⁺ release. It has been suggested that low levels of InsP₃ fail to access these stores because the InsP₃ 5-phosphatase breaks down these concentrations very efficiently (Hermosura et al., 2000). Only high concentrations of InsP₃, which are well above the K_M of the phosphatase, can mobilize the stores and thus activate I_CRAC. An alternative explanation is that low concentrations of InsP₃ do access these specialized stores but that the SERCA pumps are so active that they prevent the Ca²⁺ released by these low InsP₃ concentrations from depleting stores sufficiently for I_CRAC to activate. Energized mitochondria reduce the threshold concentration of InsP₃ that is required to evoke I_CRAC. Because the activity of the 5-phosphatase is not thought to be Ca²⁺ dependent (Shears, 1992) and presumably, therefore, would not be affected by mitochondrial Ca²⁺ uptake, this indicates that low concentrations of InsP₃ (3 µM) probably do in fact mobilize these specialized stores and hence that metabolism of InsP₃ is unlikely to be the main factor preventing these InsP₃ concentrations from evoking the current. Instead, enhanced mitochondrial Ca²⁺ uptake would facilitate store depletion by reducing both possible Ca²⁺-dependent inactivation of InsP₃ receptors and Ca²⁺ uptake via SERCA pumps.

Our results also suggest that mitochondrial Ca²⁺ uptake and Ca²⁺ ATPases of the ER, two major Ca²⁺ removal mechanisms in these and other cells (Herrington et al., 1996; Tinel et al., 1999), can functionally antagonize one another. In the absence of energized mitochondria, I_CRAC is generally not activated by InsP₃ in weak Ca²⁺ buffer unless SERCA pumps are blocked (Fierro and Parekh, 2000; Gilabert and Parekh, 2000). This indicates that a sizeable fraction of the Ca²⁺ released by InsP₃ is resequestrated into the stores such that the intraluminal Ca²⁺ content does not fall sufficiently for macroscopic I_CRAC to activate. On the other hand, in the presence of energized mitochondria and active SERCA pumps, enough Ca²⁺ is taken up by the mitochondria (and hence away from the pumps) so that stores are depleted to an extent that macroscopic I_CRAC activates, albeit to a submaximal level (Gilabert and Parekh, 2000). This dynamic interplay between SERCA pumps and mitochondrial Ca²⁺ uptake sites might require close apposition between the two Ca²⁺ removal mechanisms so that effective competition can take place. Recent morphological evidence strongly suggests that this is the case in RBL cells. Csordas and Hajnoczky (2001) have found that almost every mitochondrion has a region that is positioned very close to the SERCA pumps of the ER. It is likely therefore that at these mitochondria–ER junctions, competition between the two organelles for removing Ca²⁺ would be particularly strong.

At moderate InsP₃ concentrations (nM to low µM range), Ca²⁺ inactivation of InsP₃ receptors becomes more prominent (Mak et al., 1998). Through their effects on buffering cytosolic Ca²⁺, mitochondria might reduce this inactivation process and thereby promote further Ca²⁺ release from the stores. If the pumps are now inhibited, then the amplitude of I_CRAC increases to the maximal extent. Hence, in respiring mitochondria, some of the released Ca²⁺ is still taken back up into the InsP₃-sensitive stores. Coordinated regulation of the Ca²⁺ transport capacities of these two organelles would therefore have quite marked effects on store depletion and subsequent Ca²⁺ influx.

Mitochondrial Ca²⁺ buffering prolongs the time course of I_CRAC even when the powerful SERCA pumps are active (Figure 2B), and therefore it is an important factor that helps determine the extent of Ca²⁺ influx under conditions where other Ca²⁺ removal mechanisms are still operational. Under physiological conditions, therefore, mitochondria are powerful intracellular Ca²⁺ buffering organelles that help prolong the duration of Ca²⁺ influx.

Our results demonstrate that mitochondria are key orchestrators of store-operated Ca²⁺ entry in RBL cells. Mitochondria are involved in three crucial aspects of Ca²⁺ influx: (i) they determine whether macroscopic I_CRAC activates or not (Gilabert and Parekh, 2000); (ii) they help set the time course of I_CRAC following its activation; and (iii) they reduce the levels of intracellular InsP₃ required to activate the current. By sensitizing cells to lower InsP₃, mitochondria may determine whether relatively weak stimuli are capable of evoking Ca²⁺ influx or not.

Materials and methods

Cell culture

RBL-1 cells and Jurkat T lymphocytes, which were bought from Cell Bank at the Sir William Dunn School of Pathology, Oxford University, were cultured as previously described (Fierro and Parekh, 2000; Fierro et al., 2000).

Electrophysiology

Patch-clamp experiments were conducted in the tight-seal whole-cell configuration at room temperature (20–23°C) as previously described (Parekh et al., 1997; Fierro and Parekh, 1999). Sylgard-coated, fire-polished pipettes had DC resistances of 2.5–4 MΩ when filled with standard internal solution that contained 145 mM Cs glutamate, 8 mM NaCl, 1 mM MgCl₂, 0.1 mM EGTA, 2 mM Mg-ATP and 10 mM HEPES pH 7.2 with CsOH. A correction of +10 mV was applied for the subsequent liquid junction potential that arose from this glutamate-based internal solution. In some experiments, Cs⁺ was replaced with K⁺ (see text). Mitochondrial cocktail contained 2 mM pyruvic acid, 2 mM malic acid, 1 mM NaH₂PO₄, 0.5 mM cAMP, 0.5 mM GTP and 0.5 mM MgCl₂. Extracellular solution contained 145 mM NaCl, 2.8 mM KCl, 10 mM CaCl₂, 2 mM MgCl₂, 10 mM CsCl, 10 mM glucose and 10 mM HEPES pH 7.4 with NaOH. I_CRAC was measured by applying voltage ramps (–100 to +100 mV in 50 ms) at 0.5 Hz from a holding potential of 0 mV as previously described (Parekh et al., 1997). Currents were filtered using an eight-pole Bessel filter at 2.5 kHz and digitized at 100 µs. Currents were normalized by dividing the amplitudes (measured from the voltage ramps at –80 mV) by the cell capacitance. Capacitative currents were compensated before each ramp by using the automatic compensation of the EPC 9-2 amplifier. All leak currents were subtracted by averaging the first few ramp currents (usually two), and then subtracting this from all subsequent currents.

Data are presented as the mean ± SEM, and statistical analysis was carried out using both Student’s t and Mann–Whitney non-parametric tests. Thapsigargin was purchased from Alomone Laboratories. All other chemicals were from Sigma.