Amplification of CRAC current by STIM1 and CRACM1 (Orai1)

Abstract

Depletion of intracellular calcium stores activates store-operated calcium entry across the plasma membrane in many cells. STIM1, the putative calcium sensor in the endoplasmic reticulum, and the calcium release-activated calcium (CRAC) modulator CRACM1 (also known as Orai1) in the plasma membrane have recently been shown to be essential for controlling the store-operated CRAC current (I_CRAC)^1–4. However, individual overexpression of either protein fails to significantly amplify I_CRAC. Here, we show that STIM1 and CRACM1 interact functionally. Overexpression of both proteins greatly potentiates I_CRAC, suggesting that STIM1 and CRACM1 mutually limit store-operated currents and that CRACM1 may be the long-sought CRAC channel.

Receptor-mediated release of Ca²⁺ from intracellular stores induces Ca²⁺ entry through calcium release-activated calcium (CRAC) channels^5–7. Previous studies have identified STIM1 as the potential sensor for endoplasmic reticulum luminal Ca²⁺ concentration^1,8,9. When Ca²⁺ is depleted from intracellular stores, STIM1 translocates to vesicular structures (punctae) underneath the plasma membrane, where it is hypothesized to activate CRAC channels residing in the plasma membrane. A second protein, CRACM1, has recently been identified as essential for activating CRAC channels^3,4. This protein contains four transmembrane domains, is located in the plasma membrane and, therefore, may represent the CRAC channel itself, a subunit of the channel, or a regulatory molecule that couples to the channel. When overexpressed individually, neither STIM1 nor CRACM1 can significantly potentiate I_CRAC^1–4.

To address the potential interaction of STIM1 and CRACM1, both proteins were overexpressed individually, or in combination, in HEK293 and Jurkat T cells and the CRAC currents were measured in response to Ca²⁺ store depletion by 20 µM inositol 1,4,5-trisphosphate (Ins(1,4,5)P₃). Both cell types normally exhibit native CRAC currents of approximately 0.5 pA pF⁻¹ and approximately 3 pA pF⁻¹, respectively⁴ (Fig. 1a, c), with typical inwardly rectifying current-voltage (I/V) relationships^5,10,11 (Fig. 1b, d). Consistent with previous work^1,9, overexpression of STIM1 alone caused a small-to-modest increase in I_CRAC in HEK293 and Jurkat cells (Fig. 1a, c). CRACM1 overexpression alone did not affect the CRAC currents induced by store depletion in HEK293 cells (Fig. 1a, b) and caused a small reduction in I_CRAC in Jurkat cells (Fig. 1c, d). Unless simply due to a general effect of transfection or variability of I_CRAC across preparations, this reduction may be due to some kind of dominant–negative effect. Taken together, the available data on CRACM1 and STIM1 suggest that the individually expressed proteins, although essential for I_CRAC manifestation, cannot significantly amplify the current. This would indicate that these proteins are either not sufficient to generate large CRAC currents or that they are stoichiometrically linked and limit each others’ ability to generate CRAC currents above normal. Therefore, we co-overexpressed both proteins in HEK293 cells (see Supplementary Information, Fig. S1) and assessed store-operated currents by patch clamp.

Figure 1.

Individual and combined overexpression of STIM1 and CRACM1. (a) Normalized average time course of Ins(1,4,5)P₃-induced (20 µM) I_CRAC in HEK293 cells. Currents of individual cells were measured at −80 mV, normalized by their respective cell size, averaged and plotted versus time (± s.e.m.). Cytosolic calcium was clamped to near zero with 10 mM BAPTA. Traces represent native I_CRAC in wild-type cells (WT, black circles: n = 10), cells transfected with CRACM1 + GFP (red circles; n = 28) or STIM1 + GFP expressing cells (green circles; n = 13). (b) Average current–voltage (I/V) relationships of I_CRAC extracted from representative HEK293 cells at 60 s, representing leak-subtracted currents evoked by 50 ms voltage ramps from −100 to +100 mV, normalized to cell size (pF). Traces represent native I_CRAC in wild-type cells (n = 6), cells transfected with CRACM1 + GFP (n = 13) and STIM1 + GFP expressing cells (n = 5). (c) Normalized average time course of Ins(1,4,5)P₃-induced (20 µM) I_CRAC in Jurkat cells. Currents were analysed as in a (n = 21 for control; n = 11 for CRACM1; n = 12 for STIM1). (d) Averaged I/V traces of I_CRAC extracted from representative Jurkat cells at 60 s. Analysis as in b (n = 19 for wild type; n = 7 for CRACM1; n = 12 for STIM1). (e) Normalized average time course of I_CRAC in HEK293 or Jurkat cells expressing STIM1 + CRACM1. Analysis as in a (n = 14 for HEK293; n = 17 for Jurkat cells). The time course of I_CRAC in wild-type Jurkat cells is included for comparison (same data as in c). (f) Average current–voltage (I/V) data traces of I_CRAC extracted from representative HEK293 (red) or Jurkat cells (green) expressing STIM1 + CRACM1. Analysis as in b (n = 14 for HEK293; n = 17 for Jurkat cells). The Jurkat wild-type data trace is plotted for comparison (same data as in d).

The co-overexpression of STIM1 and CRACM1, in both HEK293 and Jurkat cells, is sufficient to generate enormous membrane currents of approximately 30 pA pF⁻¹ on store depletion by Ins(1,4,5)P₃ (Fig. 1e). These currents are significantly larger than the corresponding native currents evoked by the same experimental protocol and amount to an approximately 60-fold increase in HEK293 cells and approximately tenfold in Jurkat cells. The currents exhibit a similar time course of activation (Fig. 1e) and the same inwardly rectifying I/V relationship (Fig. 1f), as the well-characterized CRAC current^5,10. It should be noted that the average current presented in the graph represents an underestimate, as it excludes cells in which I_CRAC increased to well above 50–100 pA pF⁻¹. In these cells, the massive Ca²⁺ influx presumably saturated the intracellular Ca²⁺ chelator BAPTA (1,20bis(o-aminophenoxy) ethane-N,N,N',N'-tetraacetic acid), causing the current to inactivate^5,10 (see Supplementary Information, Fig. S2).

To further characterize these currents, some specific properties that are considered the hallmarks of native CRAC currents were assessed. We established that the current could be activated by other stimuli that cause store depletion. Passive store depletion by 10 mM BAPTA, in the absence of Ins(1,4,5)P₃, also recruited CRAC-like currents (Fig. 2a). These developed after a characteristic delay that is likely to represent the time required for Ca²⁺ to leak out of the stores⁵. The cells were next perfused with a solution in which the Ca²⁺ concentration was buffered to approximately 150 nM to avoid store depletion and this did prevent current activation (Fig. 2b). Subsequent application of ionomycin to release Ca²⁺ from intracellular stores, however, rapidly activated large currents that exhibited the same rectifying I/V as I_CRAC (Fig. 2c).

Figure 2.

Co-expression of CRACM1 and STIM1 produces a current with the characteristic features of I_CRAC. (a) Normalized average time course of I_CRAC in STIM1 + CRACM1 expressing HEK293 cells (black circles; n = 3). Currents of individual cells were measured at −80 mV, normalized by cell size (pF), averaged and plotted versus time (± s.e.m.). Passive store-depletion was induced by clamping cytosolic calcium to near zero using 10 mM BAPTA. For comparison, the solid line reproduces the Ins(1,4,5)P₃-induced data shown in Fig. 1e. (b) Time course of a representative HEK293 cell expressing STIM1 + CRACM1 (n = 3), where I_CRAC was induced by application of 2 µM ionomycin for 3 s, as indicated by the arrow. Intracellular calcium was clamped to 150 nM with 10 mM BAPTA and 4 mM CaCl₂ to prevent passive store-depletion before treatment with ionomycin. (c) I/V data trace of I_CRAC extracted from the same cell as in b at 110 s. The voltage protocol was as in Fig. 1b. (d) Representative time course of I_CRAC evoked in a HEK293 cell expressing STIM1 + CRACM1 by 20 µM Ins(1,4,5)P₃. Extracellular calcium was removed as indicated by the black bar (average inhibition at the end of application = 94% ± 1%, n = 15). (e) Representative HEK293 cell expressing STIM1 + CRACM1 where 10 mM extracellular Ca²⁺ was replaced with equimolar Ba²⁺ or Sr²⁺ during the time indicated by the bar (average inhibition by Ba²⁺ = 95% ± 1%, n = 14; and by Sr²⁺ = 81% ± 2%, n = 16). (f) Representative HEK293 cell expressing STIM1 + CRACM1 superfused with a divalent-free (DVF) external solution (n = 3). (g) Application of 50 µM 2-APB initially facilitates and then inhibits Ins(1,4,5)P₃-induced I_CRAC in a representative HEK293 cell expressing STIM1 + CRACM1 (n = 5).

The current produced by STIM1–CRACM1 overexpression was as specific for Ca²⁺ ions as I_CRAC^5,10. Removal of extracellular Ca²⁺ (nominally Ca²⁺-free), while retaining normal levels of Mg²⁺, inhibited the inward currents evoked by store depletion (Fig. 2d). Another specific characteristic of I_CRAC is that it can carry Ba²⁺ and Sr²⁺ currents, albeit at smaller levels than Ca²⁺ (refs ^5,11,12). Ion substitution experiments in which 10 mM extracellular Ca²⁺ was replaced by equimolar amounts of Ba²⁺ and Sr²⁺ resulted in a strong inhibition of the current and very limited steady-state permeation of these ions compared with Ca²⁺ (Fig. 2e). I_CRAC can also carry monovalent cations, such as Na⁺, when removing all divalent ions from the extracellular solution and this transiently increases inward currents¹⁰. The large currents in STIM1–CRACM1 overexpressing cells showed the same behaviour (Fig. 2f). Whether or not the current exhibited the pharmacological profile of I_CRAC, which is known to be enhanced by low concentrations of 2-aminoethoxydiphenylborate (2-APB) and inhibited at higher doses^13,14, was also assessed. Treatment with 50 µM 2-APB, after an initial increase as 2-APB concentration built up, completely blocked the current. Although these properties of the large CRAC currents are generally in excellent agreement with those of native I_CRAC, it is noteworthy that Sr²⁺ and Ba²⁺ currents seem smaller than those described for native I_CRAC (refs ^5,11,12). This may reflect a genuine, small Ba²⁺- and Sr²⁺-permeability of the CRAC channels in STIM1–CRACM1 overexpressing cells, and/or compromised channel function, as a consequence of the removal of extracellular Ca²⁺ (ref. ¹⁵). This difference in divalent permeation may also hint at the possibility of species differences, or that additional proteins — possibly other members of the STIM or CRACM families — may participate in shaping native CRAC currents.

In summary, our data establish that the co-overexpression of STIM1 and CRACM1 greatly amplifies store-operated currents and that these currents possess most of the defining characteristics of I_CRAC. This suggests that STIM1 and CRACM1 are entirely sufficient to control the magnitude of the CRAC current. As individual overexpression of either protein fails to augment I_CRAC, we conclude that they mutually represent limiting factors for CRAC current manifestation, although in Jurkat cells there may be a surplus of CRACM1 compared with STIM1, as STIM1 overexpression can enhance I_CRAC approximately twofold (Fig. 1c). Although more complex interpretations are conceivable, the most parsimonious interpretation of the fact that additional CRACM1 overexpression can amplify I_CRAC 10–60-fold (and in some cells well above 100-fold) would be that CRACM1 itself constitutes the CRAC channel.