Report

The CRAC Channel Activator STIM1 Binds and Inhibits L-Type Voltage-Gated Calcium Channels

See allHide authors and affiliations

Science  01 Oct 2010:
Vol. 330, Issue 6000, pp. 101-105
DOI: 10.1126/science.1191027

Abstract

Voltage- and store-operated calcium (Ca2+) channels are the major routes of Ca2+ entry in mammalian cells, but little is known about how cells coordinate the activity of these channels to generate coherent calcium signals. We found that STIM1 (stromal interaction molecule 1), the main activator of store-operated Ca2+ channels, directly suppresses depolarization-induced opening of the voltage-gated Ca2+ channel CaV1.2. STIM1 binds to the C terminus of CaV1.2 through its Ca2+ release–activated Ca2+ activation domain, acutely inhibits gating, and causes long-term internalization of the channel from the membrane. This establishes a previously unknown function for STIM1 and provides a molecular mechanism to explain the reciprocal regulation of these two channels in cells.

Excitable and nonexcitable cells are distinguished by their ability to increase their concentration of intracellular calcium ([Ca2+]i) in response to membrane depolarization (1). Excitable cells such as neurons and myocytes have voltage-gated Ca2+ channels (VGCCs) that are activated by depolarization and are essential for synaptic vesicle release, contraction, and electrical excitability (24). In contrast, nonexcitable cells such as lymphocytes and mast cells lack voltage-gated Ca2+ influx but have Ca2+ release–activated Ca2+ (CRAC) channels (57). These are activated by receptors that deplete the internal Ca2+ stores and are important for regulating gene expression and controlling cell proliferation and differentiation (5). Excitable cells express store-operated Ca2+ channel proteins, but these contribute little to Ca2+ influx (8, 9), whereas nonexcitable cells express VGCC proteins but lack voltage-gated Ca2+ currents (10, 11). The underlying mechanisms that account for the reciprocal regulation of these Ca2+ influx pathways are not understood.

VGCCs are composed of an α1 subunit that contains the pore and voltage sensor of the channel, and β and α2δ subunits that modulate trafficking and gating (12, 13). The α1 subunit has four repeats of six transmembrane domains and cytoplasmic N and C termini (14). L-type Ca2+ channels (LTCs) have a large single-channel conductance, are sensitive to dihydropyridine blockers, and are encoded by the CaV1 family of α1 subunits (15). CaV1.2 channels are the most abundant LTCs in the heart and brain and are essential for cardiac contraction and for neuronal function (16).

CRAC channels are composed of STIM (1719) and Orai proteins (2022). STIM1 is a single-pass endoplasmic reticulum protein with an intraluminal EF hand and cytoplasmic coiled-coil and lysine-rich domains (1719). Upon depletion of the Ca2+ stores, STIM1 forms oligomers that translocate to endoplasmic reticulum–plasma membrane junctions and bind to Orai channels at the plasma membrane (17, 23). STIM1 binds to Orai via a cytoplasmic region called the CRAC activation domain (CAD) that is both necessary and sufficient for channel opening (24, 25).

Rat cortical neurons express LTCs that generate a [Ca2+]i rise after membrane depolarization (Fig. 1A). These cells show little store-operated Ca2+ influx, as treatment of the cells with 1 μM thapsigargin (TG) to deplete the internal stores does not cause a [Ca2+]i rise (Fig. 1A). To test whether depletion of stores affects voltage-gated Ca2+ channels, we treated cells with 1 μM TG and then stimulated the cells with a depolarizing pulse of KCl (Fig. 1, B and C). Treatment with TG led to a 21% decrease in the initial slope of the [Ca2+]i rise, which suggests that depletion of the internal stores inhibited the conductance by VGCCs. Because the slope of the [Ca2+]i rise reflects sources of Ca2+ other than plasma membrane Ca2+ channels, we used whole-cell patch clamping to measure LTC activity directly. We transfected human embryonic kidney (HEK) 293 cells with CaV1.2, α2δ, and β1b subunits and used whole-cell patch clamping to measure the CaV1.2 currents. Treatment of these cells with TG led to a 15% decrease in the amplitude of the CaV1.2 currents over a period of 200 s, consistent with the idea that depletion of stores inhibits CaV1.2 channels (Fig. 1D). Depletion of the stores did not alter the current-voltage (I-V) relationship or the inactivation of the channels. To test whether inhibition of CaV1.2 was reversible, we treated cells with the reversible sarco-endoplasmic reticulum calcium adenosine triphosphatase (SERCA) inhibitor cyclopiazonic acid (CPA) (Fig. 1E). Treatment of cells expressing CaV1.2 with CPA led to a 10% inhibition of CaV1.2 currents that was reversed after removal of CPA, which suggests that inhibition of CaV1.2 is not specific to TG and can be reversed by refilling of stores (Fig. 1, D and E).

Fig. 1

Regulation of calcium influx through CaV1.2 by STIM1. (A) Fura-2 Ca2+ imaging of primary cortical neurons stimulated with KCl or with thapsigargin (TG; n = 17). Extracellular [Ca2+] is indicated. (B and C) Depolarization-induced increase in [Ca2+]i in cortical neurons before and after treatment with TG (B) and measurements of the rate of increase in [Ca2+]i (n = 18, *P < 0.05) (C). (D and E) Time course of Ca2+ currents measured in HEK293 cells expressing CaV1.2, before or after treatment with 1 μM TG or 5 μM CPA. (F and G) [Ca2+]i measurements of cortical neurons transfected with a control (n = 16) or a STIM1 plasmid (n = 18). Rate of change in [Ca2+]i after depolarization for control and wild-type cells is shown in (G) (*P < 0.05). (H) Current traces (above) and peak I-V relationships measured in HEK293 cells expressing CaV1.2 in the presence or absence of STIM1 with 5 mM Ba2+ as a charge carrier. (I) Normalized current amplitudes recorded in response to +10 mV depolarizing pulses.

To determine whether STIM1 has a role in suppressing CaV1.2 currents after store depletion, we introduced STIM1 into cortical neurons and measured KCl-stimulated [Ca2+]i elevations. Expression of STIM1 for 9 to 12 hours caused a decrease in the increase in [Ca2+]i (Fig. 1F) that was also observed in Neuro2A neuroblastoma cells (fig. S1). To measure the effects of STIM1 on CaV1.2 currents directly, we measured Ba2+ (Fig. 1H) and Ca2+ (fig. S2) currents in HEK293 cells expressing CaV1.2 and STIM1. Expression of STIM1 eliminated CaV1.2 currents in about 40% of cells and reduced Ba2+ and Ca2+ currents in the rest, but had no effect on the I-V relationship of the channels (Fig. 1H). Thus, expression of STIM1 inhibits activation of both endogenous and heterologously expressed CaV1.2 channels.

In contrast to neurons, T lymphocytes show no voltage-dependent [Ca2+]i elevations (10, 11) but have a large store-operated Ca2+ influx (Fig. 2A) pathway and high levels of STIM1. To test whether endogenous STIM1 can suppress CaV1.2 channels, we introduced yellow fluorescent protein (YFP)–labeled CaV1.2, along with β1b and α2δ1 subunits, into Jurkat T cells and measured the depolarization-induced Ca2+ influx (Fig. 2A). Although the YFP-labeled channel was clearly expressed, we did not detect any depolarization-induced increase in [Ca2+]i, consistent with the notion that STIM1 suppresses CaV1.2 (Fig. 2A). To determine whether STIM1 is necessary to suppress CaV1.2 activity in T cells, we used three lines of Jurkat T cells that lack store-operated Ca2+ influx (26). Sequencing the STIM1 and Orai genes in these cells revealed that three of them—CJ1, M101, and M108—contain either nonsense or missense mutations in the stim1 gene (fig. S3 and table S1) and have significantly reduced amounts of STIM1 (Fig. 2B). We introduced CaV1.2, β1b, and α2δ1 into all three cell lines and measured the depolarization-induced increase in [Ca2+]i. In contrast to wild-type Jurkat T cells, the three mutant cell lines showed voltage-activated increases in [Ca2+]i, suggesting that loss of STIM1 allowed CaV1.2 expression in these cells (Fig. 2, C and D, and fig. S4). To confirm that this was due to loss of STIM1, we expressed STIM1 in the M101 cell line along with CaV1.2 (Fig. 2E). This restored store-operated Ca2+ entry in M101 cells and also prevented activation of CaV1.2 channels (Fig. 2, F and G). As a further test of the idea that STIM1 inhibits CaV1.2 channels in T cells, we designed a short hairpin RNA (shRNA) against STIM1. The STIM1 shRNA efficiently eliminated the expression of STIM1 protein relative to a scrambled shRNA control (Fig. 2H) and allowed the functional expression of CaV1.2 channels, indicating that STIM1 suppresses VGCC function in Jurkat cells (Fig. 2, H and I).

Fig. 2

Suppression of CaV1.2 function in T lymphocytes by STIM1. (A) Fura-2 measurements of [Ca2+]i in Jurkat T cells expressing CaV1.2 or a control plasmid and treated with KCl or TG (n = 44). (B) Western blot of cell extracts of wild-type (WT), CJ1, M101, and M108 Jurkat T cells. (C) Ca2+ imaging of M101 mutant cells with and without CaV1.2 (n = 30). (D) Measurement of peak [Ca2+]i in WT or M101 Jurkat T cells treated with TG or KCl in the presence or absence of CaV1.2. (E) Overexpression of STIM1 in M101 Jurkat T cells rescues increase in [Ca2+]i induced by TG (n = 13). (F) Expression of STIM1 along with CaV1.2 reduces the KCl-induced [Ca2+]i elevation in the M101 mutants (control, n = 15; STIM1/CaV1.2, n = 20). (G) Quantification of peak depolarization-induced [Ca2+]i in M101 cells expressing STIM1 and CaV1.2 as indicated. (H) Western blot of extracts from cells expressing shRNA-STIM1 or shRNA-scrambled. (I) [Ca2+]i measurements in WT Jurkat T cells expressing CaV1.2 in the presence or absence of the shRNA STIM1.

The observation that STIM1 suppresses CaV1.2 currents led us to test whether the two proteins interact in cells. We labeled CaV1.2 and STIM1 with FLAG and Myc epitope tags, respectively, and then expressed the proteins in HEK293T cells and measured their interaction by coimmunoprecipitation. Immunoprecipitation of FLAG-tagged CaV1.2 resulted in coimmunoprecipitation of Myc-tagged STIM1 only in cells expressing both proteins (Fig. 3A). The reverse experiment of immunoprecipitating STIM1 and blotting for CaV1.2 confirmed that the two proteins form a complex when they are expressed in cells (Fig. 3B). To determine whether the two proteins also interact when expressed at endogenous levels, we performed coimmunoprecipitation of CaV1.2 and STIM1 from SH-SY5Y human neuroblastoma cells that express both proteins. STIM1 was associated with immunoprecipitated CaV1.2, indicating that the interaction also occurs under endogenous conditions (Fig. 3C).

Fig. 3

Direct interaction of STIM1 with CaV1.2. (A and B) Western blots of cell lysates (left) and immunoprecipitates (right) from HEK293T cells expressing full-length CaV1.2, STIM1, or both. (C) Western blot of cell lysates and immunoprecipitates from SH-SY5Y cells incubated with antibodies to CaV1.2 and probed with antibodies to STIM1. (D) Fura-2 Ca2+ measurements of Neuro2A cells expressing CaV1.2 alone, CaV1.2 with STIM1 lacking the CAD domain (STIM1-∆CAD), or CaV1.2 with the CAD peptide. (E and F) Western blots of lysates and immunoprecipitates of HEK293T cells expressing full-length CaV1.2 (E) or C terminus of CaV1.2 (F) along with YFP-CAD. (G) Mapping of the domains of the C terminus of CaV1.2 that bind to STIM1. (H) Immunoprecipitation of the CaV1.2 CT fragment c and STIM1. (I) Ca2+ imaging of Neuro2A cells expressing CaV1.2 lacking the CT (CaV1.2∆CT) with and without full-length STIM1 or the CAD domain. (J) Ca2+ imaging of Jurkat T cells expressing CaV1.2 alone or along with CaV1.2-CT fragment c peptide (n = 10).

The CAD of STIM1 mediates the binding and activation of Orai1 channels. We therefore investigated whether this domain is necessary for inhibiting CaV1.2 (24). We introduced STIM1 lacking the CAD (STIM1∆CAD) or the CAD peptide alone into cells expressing CaV1.2. STIM1∆CAD did not suppress depolarization-induced [Ca2+]i increase (Fig. 3D), whereas CAD alone caused 50% suppression of the [Ca2+]i increase, indicating that this domain is necessary and sufficient for inhibition of CaV1.2 (Fig. 3D). To determine whether the CAD binds to CaV1.2, we introduced YFP-tagged CAD into HEK293T cells expressing CaV1.2. Immunoprecipitation of the CAD peptide resulted in coimmunoprecipitation of the channel, indicating that the CAD can bind to CaV1.2 independently of the rest of STIM1 (Fig. 3E).

To find which domains of CaV1.2 bind to the STIM1 CAD, we overexpressed the CAD along with the intracellular loops of CaV1.2 in HEK293T cells and investigated their binding by coimmunoprecipitation. Immunoprecipitation of the CAD peptide resulted in strong coimmunoprecipitation of the C terminus of CaV1.2 (Fig. 3F). We subdivided the C-terminal domain into five peptides and tested which domain coimmunoprecipitated with the STIM1 CAD. A region between amino acids 1809 and 1908 of the C terminus of CaV1.2 interacted with the CAD of STIM1 (Fig. 3G). This same domain also interacted with full-length STIM1 protein, indicating that this region can interact with the CAD in its endogenous context (Fig. 3H).

If interaction between CaV1.2 and STIM1 is required to inhibit channel conductance, then a channel lacking the C-terminal domain should not be suppressed by STIM1. We expressed CaV1.2 lacking the C terminus (CaV1.2∆CT) along with STIM1 in mouse Neuro2A cells. Introduction of the CaV1.2∆CT led to a depolarization-induced [Ca2+]i rise in these cells that was unaffected by coexpression of STIM1 or the STIM1 CAD (Fig. 3I). Thus, binding of STIM1 to the CaV1.2 C terminus is required for suppression of channel function. To test this further, we expressed the CaV1.2 C terminus in Jurkat T cells to determine whether this domain would bind to endogenous STIM1 and prevent it from suppressing CaV1.2 channels. In contrast to cells expressing CaV1.2 alone, which showed no increase in [Ca2+]i after depolarization, cells expressing the channels and the peptide showed a large increase in [Ca2+]i (Fig. 3J), indicating that the C terminus of CaV1.2 is critical for the ability of STIM1 to suppress the activity of CaV1.2 channels.

We next investigated whether STIM1 and CaV1.2 colocalize in cells after depletion of stores. We introduced hemagglutinin (HA)–tagged CaV1.2 and FLAG-tagged STIM1 into HEK293 cells and used quantitative immunocytochemistry to examine their colocalization. We observed significant colocalization of the two proteins in resting cells and a slight increase after treatment with 1 μM TG (Fig. 4A). Because CaV1.2 and STIM1 interacted significantly before store depletion and seemed to colocalize in intracellular vesicles, we examined whether STIM1 altered the surface expression of CaV1.2 channels. We transfected hippocampal neurons with CaV1.2 containing an intracellular YFP and an extracellular HA tag (27) and measured the surface fraction of CaV1.2 by staining unpermeabilized cells with fluorescent antibodies to HA (Fig. 4, B to D). In the absence of STIM1, a large fraction of the CaV1.2 channels were expressed on the cell surface, but introduction of STIM1 led to a 73% decline in cell surface CaV1.2 channels. To determine whether STIM1 increases the internalization of CaV1.2, we expressed dominant negative dynamin-1 (DN-dyn1), a potent inhibitor of internalization, in cells expressing CaV1.2 and STIM1 (Fig. 4, B to D). DN-dyn1 prevented the loss of CaV1.2 cell surface expression in cells expressing STIM1, providing evidence that STIM1 causes internalization of the channels.

Fig. 4

Regulation of surface expression of CaV1.2 by STIM1. (A) Hippocampal neurons expressing HA-CaV1.2 and FLAG-STIM1 were fixed and stained with antibodies to HA or FLAG. Cells are shown before and after treatment with 1 μM TG. (B) YFP and HA staining of hippocampal neurons expressing CaV1.2 containing an N-terminal YFP and an extracellular HA tag along with STIM1, or STIM1 and dominant negative dynamin 1 (DN-Dyn1). (C) Measurement of the fraction of CaV1.2 channels on the cell surface from the experiment shown in (B) (n > 50; mean ± SEM). (D) Histogram of CaV1.2 cell surface expression.

Our studies support a model in which STIM1 is recruited to CaV1.2 channels after depletion of stores. It acutely reduces CaV1.2 currents by about 15% and chronically triggers CaV1.2 internalization, leading to complete loss of functional CaV1.2 channels. STIM1 therefore acts as switch that promotes Ca2+ entry through Orai channels and inhibits Ca2+ entry through CaV1.2. This is likely to be important for selecting among different Ca2+-activated signaling cascades, as Orai and CaV1.2 activate different signaling pathways. In addition, regulation of CaV1.2 by STIM1 could be important for the inhibition of CaV1.2 that occurs after activation of PLC-coupled receptors such as the muscarinic acetylcholine receptor and the D2 dopamine receptor (28, 29). Together these findings provide a previously unknown set of functions for STIM1 in mammalian cells.

Supporting Online Material

www.sciencemag.org/cgi/content/full/330/6000/101/DC1

Materials and Methods

Figs. S1 to S4

Table S1

References and Notes

  1. Supported by NIH grants DP1OD003889 and R21MH087898, the Simons Fund for Autism Research, the California Institute for Regenerative Medicine grant TG 01159. We thank A. Budzillo for help with experiments, and R. Lewis and A. Nigh for feedback on the manuscript.
View Abstract

Navigate This Article