Ca²⁺ store dynamics determines the pattern of activation of the store-operated Ca²⁺ current I_CRAC in response to InsP₃ in rat basophilic leukaemia cells

Abstract

1. I_CRAC and intracellular inositol 1,4,5-trisphosphate (InsP₃) concentration is complex. In rat basophilic leukaemia (RBL-1) cells dialysed with high intracellular Ca²⁺ buffer, the relationship is supra-linear with a Hill coefficient of 12 and resembles an apparent ‘all-or-none’ phenomenon. The non-linearity seems to arise from InsP₃ metabolism. However, it is not clear which InsP₃-metabolising pathway engenders the non-linear behaviour nor whether I_CRAC is always activated to its maximal extent by InsP₃.

2. Using the whole-cell patch clamp technique, we dialysed RBL-1 cells with different concentrations of the InsP₃ analogue InsP₃-F. InsP₃-F is broken down by Ins(1,4,5)P₃ 5-phosphatase but is not a substrate for Ins(1,4,5)P₃ 3-kinase. The relationship between InsP₃-F and I_CRAC amplitude was supra-linear and very similar to that with InsP₃ but was distinct from the graded relationship seen with the non-metabolisable analogue Ins2,4,5P₃.

3. In the presence of high intracellular Ca²⁺ buffer, InsP₃-F activated I_CRAC to its maximal extent. With moderate Ca²⁺ buffer, however, sub-maximal I_CRAC could be obtained to a maximal InsP₃-F concentration. Nevertheless, the relationship between the amplitude of I_CRAC and InsP₃-F concentration was still supra-linear.

4. Submaximal I_CRAC in response to InsP₃-F in the presence of moderate Ca²⁺ buffer was due to partial depletion of the stores, because the size of the current could be increased by thapsigargin.

5. The data suggest that first, Ins(1,4,5)P₃ 5-phosphatase is an important factor which contributes to the non-linear relationship between InsP₃ concentration and the amplitude of I_CRAC and second, InsP₃ does not always activate I_CRAC to its maximal extent. At moderate buffer strengths, submaximal I_CRAC is evoked by maximal InsP₃. However, the supra-linear relationship between InsP₃ concentration and amplitude of the current still holds.

In a variety of cell types, stimulation of cell-surface receptors that engage the phosphoinositide pathway evokes a biphasic increase in intracellular Ca²⁺: an initial inositol 1,4,5-trisphosphate (InsP₃)-mediated Ca²⁺ release phase is followed by a smaller but sustained Ca²⁺ influx component (Berridge, 1993; Parekh & Penner, 1997). In most non-excitable cells, emptying of the intracellular Ca²⁺ stores activates a selective Ca²⁺ current called I_CRAC (Ca²⁺ release-activated Ca²⁺ current; Hoth & Penner, 1992; Parekh & Penner, 1997). Ca²⁺ entry through CRAC channels is necessary for prolonging the Ca²⁺ signal following transient Ca²⁺ release, for refilling the stores as well as for a host of key cellular functions including exocytosis, gene transcription and cell proliferation (Parekh & Penner, 1997).

The relationship between the intracellular InsP₃ concentration and activation of the store-operated Ca²⁺ influx can be graded (Jacob, 1990) or supra-linear (Parekh et al. 1997; Hartmann & Verkhratsky, 1998; Liu et al. 1998) in different non-excitable cells. In rat basophilic leukaemia (RBL-1) cells, a model system for studying I_CRAC, the amplitude of the current is steeply related to InsP₃ concentration with a Hill coefficient of 12 (Parekh et al. 1997). This highly non-linear behaviour probably reflects metabolism of InsP₃ because the non-metabolisable InsP₃ analogue Ins2,4,5P₃ generates a graded response (Hill coefficient of 1). Stimulation of cell-surface receptors that engage the phosphoinositide pathway also generates I_CRAC in a non-linear manner, suggesting that this highly supra-linear behaviour might be of physiological relevance (Parekh et al. 1997).

Two questions arise from this. First, which of the InsP₃-metabolising enzymes accounts for the non-linear relationship between InsP₃ concentration and activation of I_CRAC? Second, is I_CRAC invariably activated to its maximal extent by InsP₃ (apparent ‘all-or-none’ activation) or can the current be activated submaximally under certain conditions? Whilst graded I_CRAC has been reported for both a non-metabolisable InsP₃ analogue (Parekh et al. 1997) and following passive depletion of stores (Fierro & Parekh, 1999), these conditions are non-physiological. It is not known whether InsP₃ itself activates the current to a submaximal extent. We have designed experiments to address these issues. We have first investigated the biochemical mechanism which engenders supra-linear activation and then examined whether the size of I_CRAC is always maximal following a sufficiently high increase in intracellular InsP₃. We find that the relationship between InsP₃ concentration and the size of I_CRAC is set largely by cytoplasmic Ins(1,4,5)P₃ 5-phosphatase and is complex. I_CRAC can be activated maximally or submaximally by a given InsP₃ concentration, and this is determined, at least in part, by Ca²⁺ reuptake into the stores, which operates efficiently at low to moderate concentrations of intracellular Ca²⁺ buffer. Even when I_CRAC is submaximal, the steep relationship between InsP₃ concentration and the amplitude of the current still holds.

METHODS

Rat basophilic leukaemia (RBL-1) cells, which were obtained from the American Type Culture Collection, were cultured as described previously (Parekh et al. 1997; Fierro & Parekh, 1999).

Patch clamp experiments were conducted in the tight-seal whole-cell configuration at room temperature (18–25°C) as described previously (Hamill et al. 1981; Parekh et al. 1997; Fierro & Parekh, 1999). Sylgard-coated, fire-polished pipettes had DC resistances of 2.5-3.7 MΩ when filled with standard internal solution that contained (mM): caesium glutamate, 145; NaCl, 8; MgCl₂, 1; Hepes, 10; Mg-ATP, 2; pH 7.2 with CsOH. The Ca²⁺ chelators EGTA (Sigma) or BAPTA tetracaesium salt (Molecular Probes) were added to this solution at the specified concentrations (described in the text). 3-Deoxy-3-fluoro-myo-inositol 4,5-trisphosphate (InsP₃-F; added to the pipette solution) was obtained from Calbiochem. Thapsigargin was from Alomone Laboratories. All other chemicals were purchased from Sigma. A correction of +10 mV was applied for the subsequent liquid junction potential that arose from this glutamate-based internal solution. Extracellular solution contained (mM): NaCl, 145; KCl, 2.8; CaCl₂, 10; MgCl₂, 2; CsCl, 10; glucose, 10; Hepes, 10; pH 7.4 (NaOH). I_CRAC was measured by applying voltage ramps (-100 to +100 mV in 50 ms) at 0.5 Hz from the holding potential of 0 mV as described previously (Parekh et al. 1997). Currents were filtered using an 8-pole Bessel filter at 2.5 kHz and digitised at 100 μs. Currents were normalised 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 two to four ramp currents, and then subtracting this from all subsequent currents. Data are presented as means ±s.e.m., and statistical evaluation was carried out using Student's unpaired t test.

RESULTS

InsP₃ can be metabolised in mammalian cells through two pathways: an Ins(1,4,5)P₃ 3-kinase converts InsP₃ to InsP₄ and Ins(1,4,5)P₃ 5-phosphatase degrades InsP₃ to InsP₂ (Shears, 1992). The InsP₃ analogue InsP₃-F is resistant to Ins(1,4,5)P₃ 3-kinase activity (Kozikowski et al. 1990; Sarrany et al. 1992) but is readily susceptible to breakdown by the Ins(1,4,5)P₃ 5-phosphatase (Sarrany et al. 1992). InsP₃-F is a full agonist on the InsP₃-gated Ca²⁺ release channels, although it has a slightly lower affinity (Sarrany et al. 1992). We therefore used InsP₃-F to dissect out the metabolic pathway that underlay the supra-linear activation of I_CRAC.

Figure 1 summarises the effects of dialysing cells with different concentrations of InsP₃-F. The pipette solution was strongly buffered with BAPTA and Ca²⁺ was clamped close to 120 nM in order to prevent spontaneous depletion of stores and hence activation of I_CRAC. We constructed a dose-response curve for each coverslip used and Fig. 1a summarises a typical experimental protocol in which cells were dialysed with different concentrations of InsP₃-F. The time courses of the currents (obtained from voltage ramps spanning −100 to +100 mV in 50 ms, top right panel of Fig. 1a) are shown on the left and the I–V relationships on the right. The amplitude of the current to 30 μM InsP₃-F had a peak value of −3.08 ± 0.21 pA pF⁻¹ (n= 18, Fig. 1B). This was not significantly different to that seen in the presence of 30 μM InsP₃+ 10 mM BAPTA (-3.05 ± 0.38 pA pF⁻¹, n= 7) or 30 μM InsP₃-F + 10 mM EGTA + 2 μM thapsigargin (-2.88 ± 0.26 pA pF⁻¹, n= 5; we have observed that I_CRAC in BAPTA tends to be slightly larger than that in EGTA but this is not significant; Fierro & Parekh, 1999), and indicates that 30 μM InsP₃-F activated I_CRAC to its maximal extent. The relationship between InsP₃-F concentration and the amplitude of I_CRAC was clearly supra-linear with a Hill coefficient of 17 (Fig. 1B, ^). This was strikingly similar to that seen with InsP₃ (Fig. 1B, ▵; Hill coefficient of 13 obtained with EGTA as the chelator; Parekh et al. 1997) but markedly different from that seen with the non-metabolisable analogue Ins2,4,5P₃ (Fig. 1B, •) (Parekh et al. 1997). The lowest concentration of InsP₃-F that activated maximal I_CRAC (7.5 μM) did so after a delay that was significantly longer than that seen with higher concentrations (P < 0.05, Fig. 1C). However, once the current was activated, both the time constant of activation and time-to-peak were the same irrespective of the concentration used (P > 0.1, Fig. 1C). The supra-linear relationship between InsP₃ concentration and activation of I_CRAC therefore seems to reflect the activity of cytoplasmic Ins(1,4,5)P₃ 5-phosphatase. This family of enzymes has a high capacity for InsP₃ and K_D values in the micromolar range (Shears, 1992). Therefore they can effectively suppress modest increases in intracellular InsP₃ levels, preventing macroscopic activation of I_CRAC. We predict that the relationship between the size of I_CRAC and the InsP₃-F concentration would be less steep in the presence of Ins(1,4,5)P₃ 5-phosphatase block. Specific blockers of this class of enzyme are lacking and agents that inhibit the phosphatase are plagued with additional effects (for example 2,3-diphosphoglycerate (DPG), which is often used to inhibit the Ins(1,4,5)P₃ 5-phosphatase (Streb et al. 1985), can also interfere with the Ins(1,4,5)P₃ 3-kinase (Guillemette et al. 1990), inhibit the InsP₃ receptor directly (Guillemette et al. 1990), and affect other signalling pathways like phospholipase D (Kanaho et al. 1993). Nevertheless, we tested the effects of DPG and disulfiram (Fowler et al. 1993), another putative Ins(1,4,5)P₃ 5-phosphatase inhibitor, on the InsP₃-F-I_CRAC relationship. However, we found that disulfiram (20 μM, added to the bathing medium) seemed to block CRAC channels directly, because the current could not be activated even in the presence of a maximal concentration of InsP₃ (30 μM in 10 mM EGTA; 4/4 cells). With 1 mM DPG in the pipette, three of seven cells responded to InsP₃ and 10 mM EGTA after a sizeable delay (the other cells behaved normally), consistent with reports that DPG inhibits the InsP₃ receptor (Guillemette et al. 1990). We were therefore forced to abandon these drugs.

Figure 1. Highly supra-linear activation of I_CRAC by InsP₃-F in the presence of high exogenous Ca²⁺ buffer.

A, patch clamp recordings investigating the effects of dialysing cells from the same coverslip with different concentrations of InsP₃-F are shown exactly as the experiment was carried out. The left-hand graphs depict the time course of I_CRAC, measured from voltage ramps at −80 mV. The right-hand graphs show the corresponding I–V relationships, taken at the time indicated by the arrows. B shows the overall InsP₃-F concentration-I_CRAC amplitude curve. Data were fitted using the analysis program IGOR (Wavemetrics, OR, USA). Included in the graph are the corresponding relationships for InsP₃ (the physiological isomer) and Ins2,4,5P₃ (non-metabolisable analogue). For the InsP₃ curve we have included previously published data (Parekh et al. 1997) together with new experiments. The Hill coefficient (Hn) increased from 12 to 13. Our previous Ins2,4,5P₃ data have been recently reproduced by Huang & Putney, 1998. Fitting their data yielded a Hill coefficient in the range of 2, reasonably similar to our value of 1. C shows the delay (left) and both the time-to-peak and time constant (τ) of the exponential fits (right). Two cells which gave clear currents to 5.25 μM InsP₃-F have been excluded from the left graph (but not the right graph) because the delays (100 and 81 s) would mask the significant difference in delay between 7.5 μM and the higher concentrations, by increasing the range of the y-axis. There were no significant differences between time-to-peak or time constant of activation. For each graph, points represent means ±s.e.m. of 8–18 cells, except for 5.25 μM InsP₃-F in C (n= 2). Cells were dialysed with standard internal solution supplemented with 12.5 mM BAPTA and 6.0 mM CaCl₂. Free Ca²⁺ was calculated to be 122 nM and the free BAPTA concentration 6.25 mM. Dialysis with this solution without InsP₃ failed to activate I_CRAC over a 400 s period (8 cells).

Taken together, these results indicate that I_CRAC activates in a supra-linear manner to InsP₃ in the presence of high intracellular Ca²⁺ buffer. Furthermore, the range of InsP₃ concentrations over which I_CRAC activates is so narrow (2-fold) under these conditions that it approximates to an apparent ‘all-or-none’ process.

We then examined whether first, InsP₃ supra-linearly activated I_CRAC under more physiological conditions (i.e. weak intracellular Ca²⁺ buffering) and second, if this was always to its maximal extent. Unfortunately, the current cannot be measured when cells are dialysed with InsP₃ in low Ca²⁺ buffer (Fig. 2a, InsP₃-F + 0.1 mM EGTA; see also Huang et al. 1998; Fierro & Parekh, 2000). We therefore determined the lowest concentrations of buffer that would still enable us to reliably measure macroscopic I_CRAC and, having found this, we then examined the relationship between InsP₃-F concentration and the size of I_CRAC. Cells were dialysed with 30 μM InsP₃-F and different concentrations of EGTA (range, 0.1-10 mM). For 30 μM InsP₃-F, concentrations of EGTA of less than 0.35 mM generally failed to support any macroscopic activation of I_CRAC (Fig. 2a). Close to maximum current was seen with around 1 mM EGTA (-2.33 ± 0.19 compared with −2.64 ± 0.31 pA pF⁻¹ for 10 mM EGTA). However, in the presence of 0.6 mM EGTA, we obtained I_CRAC that clearly had a submaximal amplitude (-1.39 ± 0.21 pA pF⁻¹, n= 8; Fig. 2a). The relationship between exogenous buffer concentration and I_CRAC amplitude in Fig. 2a was supra-linear with a Hill coefficient of 3.5. Figure 2B shows an experiment with a threshold concentration of 3 μM InsP₃-F instead of 30 μM (30 cells; 0.1-1 mM EGTA). For comparison, the corresponding data with 30 μM InsP₃-F were also included. For EGTA concentrations of 0.6 mM or less, no macroscopic I_CRAC was detected. In the presence of 1 mM EGTA and 3 μM InsP₃-F, a small current was occasionally seen (6 of 13 cells) but its properties were no different to those of the current evoked by 1 mM EGTA alone (50 % of cells responded to 1 mM EGTA alone compared with 46 % for 1 mM EGTA + 3 μM InsP₃-F; their time-to-peak values were 120.1 ± 20.4 and 122.67 ± 44.52 s, respectively, and the overall extent of I_CRAC was similar; P > 0.1; Fig. 2B; see also Fierro & Parekh, 1999).

Figure 2. Supra-linear activation of I_CRAC in the presence of moderate Ca²⁺ buffering.

A, relationship between intra-pipette EGTA concentration and the amplitude of I_CRAC (evoked by dialysis with 30 μM InsP₃-F). Note that I_CRAC was clearly submaximal in the presence of 0.6 mM EGTA. Each point is the mean ±s.e.m. of at least 5 cells. B, graph showing the relationship between EGTA concentration and size of I_CRAC when the threshold dose of 3 μM InsP₃-F (IP₃-F) was used. Data for 30 μM have been included for comparison. The EGTA concentration ranged from 0.1 to 1.0 mM (each point is the mean ±s.e.m. of at least 5 cells and error bars are within the circles). The open triangle that overlaps with the filled circle at 1 mM is the size of I_CRAC seen following dialysis with 1 mM EGTA and no InsP₃-F. C, summary of pooled results when cells were dialysed with different concentrations of InsP₃-F in the presence of 1 mM EGTA. Each point is the mean ±s.e.m. of at least 5 cells. Depending on how we treated the response to 3 μM InsP₃-F, two different fits were obtained. If we included the size of I_CRAC with 1 mM EGTA alone (-0.3 ± 0.11 pA pF⁻¹), and then subtracted this value from the amplitudes seen when different concentrations of InsP₃-F were included, we obtained the dashed fit (raw data). However, if we arbitrarily set the amplitude of I_CRAC to zero for 3 μM InsP₃-F (because there was no significant difference between 1 mM EGTA alone and 3 μM InsP₃-F + 1 mM EGTA), and subtracted the value for 3 μM InsP₃-F + 1 mM EGTA from the currents seen with higher concentrations of InsP₃-F, then we obtained the continuous fit (corrected data). The dashed fit is an underestimate, whereas the continuous one is an overestimate, of the true Hill coefficient (Hn). D, graph summarising the InsP₃-F concentration-I_CRAC relationship in the presence of 0.6 mM EGTA. Whereas the majority of cells failed to generate a current to 3 μM InsP₃-F + 0.6 mM EGTA or 0.6 mM EGTA alone, two cells from one preparation did generate a clear I_CRAC to 3 μM InsP₃-F (-1.5 and −1.2 pA pF⁻¹). In this same preparation, 0.6 mM EGTA also produced a current rapidly (-1.0 pA pF⁻¹). Because this was so uncharacteristic, we did not include any cells from this preparation.

We then determined the relationship between InsP₃-F concentration and activation of I_CRAC in the presence of moderate intracellular Ca²⁺ buffering (1 and 0.6 mM EGTA). Results are summarised in Fig. 2C and D. The relationship between InsP₃-F concentration and the amplitude of I_CRAC was clearly supra-linear for both cases (Hill coefficient in the range 3.5-28 and apparent K_D of 4.1 μM for 1 mM EGTA (see legend to Fig. 2); corresponding values for 0.6 mM EGTA were 16 and 3.4 μM, respectively). Therefore even in the presence of quite moderate amounts of Ca²⁺ buffer and when the current is submaximal (as in 0.6 mM EGTA), I_CRAC is still supra-linearly related to the cytoplasmic InsP₃ concentration.

We designed experiments to understand why I_CRAC was submaximal in the presence of 0.6 mM EGTA, despite dialysis with a maximal concentration of InsP₃-F. We reasoned that stores might refill partially in the presence of such a concentration of Ca²⁺ buffer and that this would reduce the overall extent of I_CRAC activation. If this was the case, then one would predict that the size of I_CRAC should increase in the presence of the sarco-endoplasmic reticulum Ca²⁺-ATPase (SERCA) blocker thapsigargin (Thastrup et al. 1989), since the latter should prevent any Ca²⁺ reuptake and hence refilling of the stores. The results of Fig. 3a and B show that this was indeed the case. Inclusion of thapsigargin in the pipette solution (with 30 μM InsP₃-F and 0.6 mM EGTA) had quite dramatic effects on the properties of I_CRAC. The delay before the current activated was substantially reduced (3-fold, P < 0.01; Fig. 3a and B) and, strikingly, the size of the current increased significantly (P < 0.01, Fig. 3a and B). Hence SERCA-mediated reuptake can slow the onset of I_CRAC as well as reduce the overall extent of activation. However, the rate of development of I_CRAC was the same for both conditions (Fig. 3B). The size of I_CRAC in 30 μM InsP₃-F + 0.6 mM EGTA + thapsigargin was similar to that seen in the presence of either InsP₃+ 10 mM EGTA or InsP₃+ 10 mM EGTA + thapsigargin and is therefore the maximal amplitude that can be recorded. Inhibition of SERCA pumps can therefore increase the size of I_CRAC evoked by a maximal concentration of InsP₃-F in moderate (0.6 mM) but not high (10 mM) EGTA, consistent with enhanced Ca²⁺-dependent SERCA reuptake in the presence of moderate Ca²⁺ buffer (Mogami et al. 1998). An alternative explanation for the actions of thapsigargin is that the subsequent elevation of intracellular Ca²⁺ inhibits the activity of the Ins(1,4,5)P₃ 5-phosphatase so that InsP₃ concentrations in the cytosol are increased and hence I_CRAC becomes larger. However, the Ins(1,4,5)P₃ 5-phosphatase is only weakly sensitive to Ca²⁺ changes and the most parsimonious explanation of the thapsigargin results is that it reflects reduced Ca²⁺ reuptake.

Figure 3. Graded I_CRAC can be obtained in the presence of moderate Ca²⁺ buffer because of SERCA uptake and inhibition of the InsP₃ receptor.

A shows recordings for a cell dialysed with 30 μM InsP₃-F and 0.6 mM EGTA (•) and for a cell dialysed with this solution but supplemented with 2 μM thapsigargin (thap), a SERCA blocker. Note the dramatic increase in the size of I_CRAC following inhibition of the pumps. B, summary of amplitudes, delays and times to peak for experiments as in A. C, cells dialysed with InsP₃ and 1 mg ml⁻¹ heparin (an inhibitor of the InsP₃ receptor) can produce a wide range of different I_CRAC amplitudes, in spite of the presence of a maximal InsP₃ concentration. This experiment will depend on the relative speed of entry of InsP₃ and heparin into the cytosol (proportional to series resistance), subsequent diffusion in the cytosol and access to the InsP₃ receptors. Cells were dialysed with a solution in which Ca²⁺ was buffered at 225 nM (3 mM Ca-EGTA and 2 mM EGTA). The decay in the current reflects store refilling because it could be reduced by inclusion of thapsigargin in the pipette (5 cells; data not shown). Dialysis with heparin fails to alter the activation of I_CRAC evoked by ionomycin (Parekh & Penner, 1995) or passive depletion (Fierro & Parekh, 1999), indicating that it is not interfering directly with the activation mechanism itself nor with the CRAC channels.

Because inhibition of SERCA-mediated Ca²⁺ uptake into the stores increases the size of I_CRAC in moderate Ca²⁺ buffer, we examined whether the converse was also true: namely, if inhibition of Ca²⁺ efflux from the stores, as I_CRAC developed, could turn the current off. We reasoned that, in the absence of further InsP₃-mediated Ca²⁺ release, SERCA pumps should rapidly refill the stores, thereby deactivating the current. The experiment of Fig. 3C was designed to test this. We dialysed cells with 30 μM InsP₃ and heparin (1 mg ml⁻¹), a competitive antagonist of the InsP₃ receptor. Because of its large size, heparin diffuses into the cytosol relatively slowly and hence I_CRAC would initially start to activate before appreciable amounts of intracellular heparin would have built up. However, as intracellular heparin levels rise, it should displace InsP₃ from its receptor, thereby causing I_CRAC to switch off as SERCA pumps refill the stores. Results are summarised in Fig. 3C. We obtained a wide dispersion of I_CRAC amplitudes, despite dialysis with a maximally effective InsP₃ concentration. Some cells generated a large I_CRAC whereas others produced rather small currents. Such pronounced variation was not seen in the absence of heparin. Hence a range of I_CRAC amplitudes can be obtained to a fixed InsP₃ concentration when Ca²⁺ release is compromised. The results also indicate that both activation and maintainance of I_CRAC in moderate Ca²⁺ buffering requires the InsP₃ receptor to be in an open conducting state in order to compensate SERCA reuptake. Although our results suggest that the InsP₃ receptor indirectly controls I_CRAC through the Ca²⁺ content of the stores, it is also conceivable that a more direct regulation involving a physical interaction between the CRAC channels and InsP₃ receptors also contributes (Kiselyov et al. 1998). The results of Fig. 3 demonstrate that activation of I_CRAC by InsP₃ is not a true ‘all-or-none’ phenomenon because, once the current was activated, it did not invariably reach its maximal extent.

DISCUSSION

Our results provide new insights into the activation of I_CRAC. The second messenger InsP₃ does not invariably activate I_CRAC to its maximal extent, because submaximal currents can be observed at moderate levels of intracellular Ca²⁺ buffering. This seems to reflect a dynamic interplay between Ca²⁺ release and thapsigargin-sensitive Ca²⁺ reuptake at the level of the InsP₃-sensitive Ca²⁺ stores. It is interesting that, in moderate buffer, SERCA-mediated reuptake prevents InsP₃ from depleting the stores to the extent that I_CRAC activates maximally. It seems unlikely that primary active transport (Ca²⁺ uptake) can match the efflux of Ca²⁺ through the large conductance InsP₃-gated channels. As we have recently suggested (Fierro & Parekh, 2000), the InsP₃ receptors on the stores may inactivate, at least partially, hence reducing the extent of Ca²⁺ movement out of the stores. Furthermore, the fall in intraluminal as well as the increase in cytosolic Ca²⁺ may both further enhance SERCA-mediated reuptake (Mogami et al. 1998).

The finding that I_CRAC activates only partially in response to InsP₃ in moderate buffer and that this is due to SERCA activity indicates that regulation of Ca²⁺ uptake may represent a powerful means for grading the extent of activation of I_CRAC in the presence of the weak to moderate levels of intracellular Ca²⁺ buffering seen in intact cells.

Although I_CRAC can be submaximal in the presence of moderate intracellular Ca²⁺ buffer, the highly supra-linear relationship between InsP₃ concentration and activation of the current still holds. However, we cannot be sure that the supra-linear behaviour is valid under conditions of weak buffering, since we and others consistently fail to record activation of the current by InsP₃ under these conditions.

The supra-linear relationship between InsP₃ concentration and activation of I_CRAC seen in moderate to high intracellular Ca²⁺ buffer probably arises from the cytoplasmic activity of Ins(1,4,5)P₃ 5-phosphatase. Although this conclusion is drawn from experiments with a phosphatase-resistant InsP₃ analogue and the standard pharmacological tools we have employed are not refined enough to selectively interfere with the enzyme, nevertheless this opens up the possibility of new experiments using molecular genetics to test this hypothesis in the future. Although the human Ins(1,4,5)P₃ 5-phosphatase has been cloned, the corresponding enzyme in the rat has not. Nevertheless, we have tried to reduce expression of the rat inositol 5-phosphatase using antisense oligonucleotides based on the human sequence (supplied by Biognostik, Goettingen, Germany), but without success. Cloning of the rat enzyme should enable us to design more rational antisense oligonucleotides. Interestingly, in certain human malignant haemopoetic cells, the activity of this enzyme is significantly reduced compared with that in normal cells (Mengubas et al. 1994). Because Ca²⁺ influx through CRAC channels has been linked to cell growth and proliferation (Berridge, 1995), it is tempting to speculate that the abnormal cell growth in these forms of human leukaemia reflects, at least in part, the non-linear relationship between InsP₃ concentration and the size of I_CRAC.