CRACM1 Is a Plasma Membrane Protein Essential for Store-Operated Ca²⁺ Entry

Receptor-mediated signaling in nonexcitable cells, immune cells in particular, involves an initial rise in intracellular Ca²⁺ due to release from the intracellular stores. The resulting depletion of the intracellular stores induces Ca²⁺ entry through the plasma membrane through CRAC channels (1–4). This phenomenon is central to many physiological processes such as T cell proliferation, gene transcription, and cytokine release (3, 5–7). Biophysically, CRAC currents have been well characterized (2, 8, 9), but the identity of the CRAC channel itself and the pathway resulting in its activation are still unknown. Recently, STIM1 (for stromal interaction molecule in Drosophila) was identified as an essential component of store-operated calcium entry (10, 11). This protein is located in intracellular compartments that likely represent parts of the endoplasmic reticulum (ER). It has a single transmembrane-spanning domain with a C-terminal Ca²⁺-binding motif that appears to be crucial for its hypothesized function as the ER sensor for luminal Ca²⁺ concentration. When stores become depleted, STIM1 redistributes into distinct structures (punctae) that move toward and accumulate underneath the plasma membrane. Whether or not STIM1 actually incorporates into the plasma membrane is controversial (10, 12, 13). Although STIM1 is required to activate CRAC currents, its presence or even its translocation appears not to be sufficient to mediate CRAC activation, because lymphocytes from patients with severe combined immunodeficiency (SCID) appear to have normal amounts of STIM1 levels and normal function, yet fail to activate CRAC channels (14). This suggests that other molecular components may participate in the store-operated Ca²⁺ entry mechanism.

To identify genes encoding the CRAC channel or other proteins involved in its regulation, we performed a high-throughput, genome-wide RNA interference (RNAi) screen in Drosophila S2R+ cells. The effect of knockdown of each of the ~23,000 genes was tested by fluorescence measurements of intracellular Ca²⁺ concentration in 384-well microplates with an automated fluorometric imaging plate reader (FLIPR, Molecular Devices). Changes in [Ca²⁺]_i were measured in response to the commonly used SERCA [sarcoplasmic and endoplasmic reticulum calcium adenosine triphosphatase (ATPase)] inhibitor thapsigargin, which causes depletion of Ca²⁺ from intracellular stores. An example of responses from this primary screen is illustrated in Fig. 1A, obtained from microplate no. 60. All 63 plates contained wells in which double-stranded RNA (dsRNA) against Rho1 served as negative control and dsRNA against stim1 as positive control. Higher resolution graphs of the real-time [Ca²⁺]_i imaging data are shown in Fig. 1, B and C, from cells treated with dsRNA against Rho1 (mock) and stim1, as well as two genes we later identified as CRAC modulators 1 and 2 (CRACM1 and CRACM2). On the basis of inhibitory efficacy relative to positive and negative controls, we identified ~1500 genes that reduced Ca²⁺ influx to varying degrees (table S1). After eliminating numerous genes based on artifactual fluorescence signals or because they represent known housekeeping genes, cell cycle regulators, and so on, we eventually arrived at 27 candidate genes (table S2) that were subsequently evaluated in a secondary screen using single-cell patch-clamp assays.

From the secondary patch-clamp screen, we identified two novel genes that are essential for CRAC channel function, CRACM1 (encoded by olf186-F in Drosophila and FLJ14466 in human) and CRACM2 (encoded by dpr3 in Drosophila, with no human ortholog). We measured CRAC currents in Drosophila Kc cells after inositol 1,4,5-trisphosphate (IP₃)–mediated depletion of Ca²⁺ from intracellular stores. Both untreated control wild-type cells and cells treated with an irrelevant dsRNA against Rho1 (mock) responded by rapidly activating a Ca²⁺ current with the time course (Fig. 1D) and inwardly rectifying current-voltage (I/V) relation (Fig. 1E) typical of I_CRAC in mammalian (2) and Drosophila (15) cell types. In contrast, CRAC currents were essentially abolished in cells treated with dsRNA for CRACM1 and CRACM2. In some of the experiments on CRACM1, we also applied ionomycin (10 µM) extracellularly on top of the 20 µM IP₃ included in the patch pipette to ensure complete store depletion, but this also failed to induce I_CRAC (fig. S1C). Similarly, CRAC currents were also absent when passive store depletion was induced by the Ca²⁺ chelator BAPTA [(1,2-bis(o-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid)] (fig. S1, A and B).

We studied the human ortholog of CRACM1, a 37.7-kD protein encoded by gene FLJ14466, to confirm that its function is conserved across species and that it is involved in store-operated Ca²⁺ entry. We used small interfering RNA (siRNA)–mediated silencing of human CRACM1 in human embryonic kidney cells (HEK293) and human T cells (Jurkat). The selective knockdown of CRACM1 message was confirmed by semiquantitative reverse transcription polymerase chain reaction (RT-PCR) analysis (Fig. 2A). Two different CRACM1-specific siRNA sequences caused a 60 to 70% inhibition of calcium influx in response to thapsigargin-induced store depletion in HEK293 cells (Fig. 2B). Patch-clamp recordings obtained from siRNA-treated cells, responding to intracellular IP₃ perfusion, demonstrated a nearly complete inhibition of CRAC currents (Fig. 2, D and E). In Jurkat cells, siRNA-mediated inhibition of Ca²⁺ influx was close to 20% (Fig. 2C) and not as dramatic as in the HEK293 cells. However, I_CRAC in Jurkat cells was effectively reduced by both siRNA sequences (Fig. 2, F and G). The differences in the efficacy of suppressing the changes in [Ca²⁺]_i in HEK293 and Jurkat cells (see Fig. 2, B and C) are likely due to the different magnitudes of I_CRAC in these two cell types. CRAC current densities in HEK293 cells (~0.5 pA/pF) are much smaller than typically seen in Jurkat cells (~2.5 pA/pF), and a further inhibition may explain the more dramatic reduction in the Ca²⁺ signal than that observed in Jurkat cells (note that the remaining CRAC current densities in siRNA-treated Jurkat cells, while strongly reduced, are ~0.4 pA/pF and comparable to the normal CRAC current densities of untreated HEK293 cells). Taken together, these data indicate that CRACM1 is also a key modulator of store-operated CRAC currents in human cells.

Given that the knockdown of CRACM1 inhibited CRAC activation, we wanted to know whether overexpression would enhance Ca²⁺ influx and CRAC current densities. HEK293, Jurkat, and RBL-2H3 cells were infected with a Myc-tagged CRACM1 and green fluorescent protein (GFP) retrovirus, and overexpression of the protein was confirmed in HEK293 cells by immunoprecipitation followed by Western blotting (Fig. 3A). However, we did not detect any increase in CRAC current amplitudes above control levels in either HEK293 (Fig. 3B) or Jurkat cells (fig. S2A) and only a slight increase in RBL cells (fig. S2B). These data suggest that CRACM1, although necessary for CRAC activation, does not in and of itself generate significantly larger CRAC currents.

An important question is whether CRACM1 localizes to the ER (as does STIM1) or to the plasma membrane. To address this question, we tagged CRACM1 on either end (Myc–C terminus and flag–N terminus) and transfected the constructs into HEK293 cells. After 24 hours, immunofluorescence confocal analysis revealed no staining in intact cells expressing either construct, which suggested that both tags are intracellular. After permeabilizing the cells, both constructs were detected by the fluorescent antibody and showed predominant peripheral staining of the plasma membrane (Fig. 3, C and D). These data fit well with the hydropathy profile of CRACM1, which predicts a topology of four transmembrane domains, with both ends facing the cytosol (fig. S2C).

In summary, our results demonstrate that the protein CRACM1 is essential for store-operated Ca²⁺ influx via CRAC channels. Although the overexpression of CRACM1 does not alter the magnitude of CRAC currents, the plasma membrane localization of this protein and the presence of multiple transmembrane domains point toward a direct role for CRACM1 in store-operated calcium influx. A number of possible functions can be envisioned for CRACM1. First, CRACM1 could function as the CRAC channel itself. In this scenario, the unaltered CRAC currents in CRACM1 overexpressing cells might be due to a limiting factor upstream of CRAC channel activation (e.g., STIM1). Second, CRACM1 could be a subunit of a multimeric channel complex, in which case the other subunit(s) could become the limiting factor(s) during overexpression. Finally, CRACM1 might function as a plasma membrane acceptor or docking protein, possibly for STIM1 or some other as-yet-unidentified component of the signaling machinery that ultimately leads to CRAC channel activation and store-operated Ca²⁺ entry.