|Home | About | Journals | Submit | Contact Us | Français|
JPY, WZ, MRD and YJC performed the experiments, and all authors participated in the design of the experiments and in writing the manuscript.
Store-operated Ca2+ influx is mediated by store002Doperated Ca2+ channels (SOCs) and is a central component of receptor-evoked Ca2+ signals1. The Orai channels mediate SOCs2–4 and STIM1 is the ER-resident Ca2+ sensor that gates the channels5, 6. How STIM1 gates and regulates the Orai channels is unknown. Here, we report the molecular basis for gating of Orais by STIM1. All Orai channels are fully activated by the conserved STIM1(344–442), which we termed SOAR (the STIM1 Orai Activating Region). SOAR acts in combination with STIM1(450–485) to regulate the strength of interaction with Orai1. Orai1 activated by SOAR recapitulates all the entire kinetic properties of Orai1 activated by STIM1. Mutations of STIM1 within SOAR prevent activation of Orai1 without preventing co-clustering of STIM1 and Orai1 in response to Ca2+ store depletion, indicating that STIM1-Orai1 co-clustering is not sufficient for Orai1 activation. An intact C-terminus α-helicial region of Orai is required for activation by SOAR. Deleting most of Orai1 N terminus impaired Orai1 activation by STIM1, but (Δ1–73)Orai1 interacts with and is fully activated by SOAR. Accordingly, the characteristic inward rectification of Orai is mediated by an interaction between the polybasic STIM1(672–685) and a proline-rich region in the N terminus of Orai1. Hence, the essential properties of Orai1 function can be rationalized by interactions with discrete regions of STIM1.
Ca2+ influx is a critical step of receptor-evoked signaling, and is mediated by plasma membrane channels that are activated in response to release of Ca2+ from the endoplasmic reticulum (ER), and are thus named store-operated Ca2+ influx channels (SOCs)1. Recently, the molecular determinants of SOCs and their gating have been elucidated. STIM1 is a Ca2+ binding protein that functions as the ER Ca2+ sensor that activates SOCs5, 6. STIM1 is a multi-domain protein with N terminal EF hand and SAM domains that reside in the ER lumen and a cytoplasmic region composed of an ERM (Ezrin/radixin/moesin) domain, a S/P-rich domain, and a polybasic lysine- (K) rich domain5, 6. In response to Ca2+ release from the ER, Ca2+ dissociates from the EF hand, and STIM1 clusters next to the plasma membrane to activate the SOCs5–8.
Two channel families have been identified as STIM1-regulated SOCs; the TRPC (TRPCs)9 and Orai channels (Orais)2–4. The TRPCs function as non-selective, Ca2+ permeable channels10, whereas Orai1 mediates the Ca2+ selective Ca2+-release activated (CRAC) current2–4. The STIM1 ERM domain binds TRPCs7, allowing their heteromultimerization11, while STIM1(K684,685K) gats TRPCs by electrostatic interaction with two conserved negative charges within TRPCs TRP Box212.
How STIM1 gates and regulates the Orais is not known. STIM1 is obligatory for the Orais to function as channels4, 13, 14. Gating of the native CRAC current7 and of Orai115, 16 is mediated by the STIM1 C terminus, but the gating does not require the STIM1 S/P and K-domains12. A proline-rich region in the N terminus of Orai1 is suggested to be important for gating of Orai1 by STIM117. However, another study reported that deletion of the first 73 residues of Orai1, which includes the proline-rich region, does not prevent activation of Orai1 by STIM118. To better define gating and regulation of Orai1 by STIM1, we searched for the STIM1 domains that mediate gating and voltage-dependence of Orai1.
CRAC current7 and Orai115, 16 are activated by the STIM1 C terminus, but the S/P and polybasic K- domains are not required for activation of Orai112. We therefore searched for the minimal STIM1 domain that could fully activate Orai1. Drosophila STIM1 terminates at the equivalent human STIM1 residue 48519. Indeed, STIM1(235–485) activates native SOCs and Orai1. Analysis of systematic N and C terminal deletions identified STIM1(344–442) as the minimal sequence that fully activates Orai1. The effects of various STIM1 fragments evaluated in the work (Fig. 1c) on spontaneous Ca2+ influx are summarized in Fig. 1b, and example traces are shown in Fig. 1a. The STIM1 fragments had one of three effects: they activated spontaneous Ca2+ influx; inhibited SOCs; or had no effect. Store depletion by inhibition of SERCA with CPA showed that the STIM1 fragments did not deplete the stores. Since STIM1(344–442) is the minimal domain that activates Ca2+ influx (Fig. 1) and all Orai channels (see below), we termed it SOAR, for the STIM1 Orai Activating Region.
To assess cell biological correlates of STIM1 activation of Orai1, we monitored the distribution of EGFP-STIM1 fragments and mCherry-Orai1 in HEK293 cells (Fig. 1e and Fig. S1). In all cases, STIM1 constructs that activate Orai1 co-localized with Orai1 at the plasma membrane. Several constructs, such as SOAR, redistributed from the cytosol to the plasma membrane when co-expressed with Orai1, independent of the native STIM1. Other constructs associated with the plasma membrane when expressed alone, but knock-down of native Orai1 prevented their plasma membrane enrichment (STIM1(344–460) in Fig. 1e). N and C terminal truncations of SOAR that failed to activate Orai also failed to co-localize with Orai1 (Figs. 1e and S1b). Depletion of internal Ca2+ stores had no further effect on their co-localization with Orai1 (not shown). This is consistent with a recent report showing that the STIM1 C terminus does not cluster with Orai120. These data indicate a precise correlation between the ability of STIM fragments to localize to the plasma membrane with Orai, and to activate the channel.
Fig. S2a shows that SOAR is conserved among vertebrate species ranging from human to Zebrafish and in invertebrates, but with lower sequence identity for C. elegans. Attempts to purify SOAR expressed in bacteria were not successful. However, we expressed and purified STIM1(336–485). The CD spectra of STIM1(336–485) (Fig. S2b) indicates a 45% α-helical structure, which compares with the predicted 62%. Both, gel filtration (red trace) and light scattering (blue trace) analyses indicate that STIM1(336–485) exists as a dimer in solution (Fig. S2c). Interestingly, recent work suggests that Orai1 functions as a tetramer21, 22 and assembles with two STIM1 molecules to form the functional channel21. The dimer formation of STIM1(336–485) is consistent with these findings.
The effect of SOAR on Orai1 current was measured at a 1SOAR:1Orai1 expression ratio, which is optimal for maximal activation of Orai1 (Fig. S3). Fig. 2 shows that SOAR and the cytosolic STIM1 fragments that co-localize with Orai1 fully activate Orai1, and SOAR activates Orai1 independent the native STIM1 (Fig. 2d). Current measurement starting within 10 sec of break-in shows that about 50% of the Orai1 current induced by the STIM1 fragments is spontaneously active (Fig. 2e), with further activation likely due to chelation of cytoplasmic Ca2+ by the slowly diffusing BAPTA. Importantly, all kinetic parameters of the Orai1 current activated by SOAR are identical to the parameters of Orai1 current activated by STIM1 (Figs. 2a, 2b, 2f), and STIM1 and SOAR similarly activated the Na+ current of the pore mutant Orai1(E106D) (Figs. 2g).
Disruption of the Orai1 C-terminus α-helix by the Orai1(L273S) mutation inhibits interaction with and activation of Orai1 by STIM115. Similarly, SOAR does not co-localize with Orai1(L273S) at the plasma membrane (Fig. 3a), and STIM1 and SOAR do not activate Orai1(L273S) (Fig. 3b). Hence, activation of Orai1 requires interaction of SOAR with the Orai1 C terminus.
The Orai1 N terminus encompasses aa 1–86. Deletion of aa 1–73 reduces activation of Orai1 by STIM118. Fig. 3c–e shows that STIM1 poorly activates (Δ1–73)Orai1. Most strikingly, SOAR fully activates (Δ1–73)Orai1, and (Δ1–73)Orai1 activated by SOAR has the same properties as Orai1 activated by SOAR or STIM1. Fig. S4d shows that (Δ1–73)Orai1 recruits SOAR to the plasma membrane and, in response to store depletion, co-clusters with STIM1 at the plasma membrane. SOAR and STIM1(344–460) co-IP equally well with Orai1 and with (Δ1–73)Orai1, but not with Orai1(L273S) (Fig. S4e). The specificity of the co-IP is confirmed by the lack of co-IP of STIM1(344–425) and Orai1 or (Δ1–73)Orai1. SOAR and STIM1(344–460) did not co-IP with an Orai1 mutant lacking the C-terminus, however, Orai1(ΔC) was mostly retained in the ER which makes this construct less informative (not shown). Hence, SOAR binding and activation of Orai1 requires the C-terminus but not the N-terminus of Orai1.
Precedence from studies of STIM1 activation of TRPC1 indicates that STIM1 binding and gating is mediated by distinct regions of the molecule12. We searched for mutants of STIM1 that might distinguish these properties for Orai. Residues 344–350 of SOAR are essential for activation of Orai1 (Fig. 1b), and Figure 4 shows that STIM1(LQ347/348AA) and SOAR(LQ347/348AA) do not activate Orai1. SOAR(LQ347/348AA) does not co-localize or co-IP with Orai1 (Fig. 4c and 4f). By contract, when expressed alone, STIM1(LQ347/348AA) is found in tubular-like structures in untreated cells and it clusters into puncta in response to store depletion (Fig. 4d). Moreover, when co-expressed with Orai1, STIM1(LQ347/348AA) co-clusters with Orai1 in cells with depleted stores (arrows in Fig. 4e), and co-IPs with Orai1 (Fig. 4f). Importantly, STIM1(LQ347/348AA) is functional towards TRPC1 (Fig. 4h), suggesting that the overall structure of STIM1(LQ347/348AA) is not disrupted.
Together, the results in Fig. 4 indicate that SOAR is essential for activation of Orai1 by STIM1. Furthermore, the interaction and co-clustering of STIM1 with Orai1 is not sufficient for activation of Orai1 by STIM1. This implies that SOAR within STIM1 actively gates Orai1 to open the channel.
The differential action of SOAR versus STIM1 in activating (Δ1–73)Orai1 (Fig 3) suggests that another domain in STIM1 normally interacts with the N terminus of Orai1 and this is required for the action of SOAR when it is part of STIM1. To examine this hypothesis, we noted that the N-terminus of Orai contains prolines grouped at positions 3, 5; 7–9; 14; 17, 18; 39, 40 and 43–47. Mutation of prolines 14; 17, 18; 43, 44; and 45–47 to alanines had no effect on activation of Orai1 by STIM1 or SOAR. Mutation of prolines 7–9 resulted in a mutant with low channel activity with STIM1 and SOAR (not shown), independent of surface expression (Fig. S6a). However, Orai1(P3,5A) and Orai1(P39,40A) behave similar to (Δ1–73)Orai1 with respect to activation by STIM1 and SOAR (Fig. 5a– 5c). The I/V plots for activation of Orai1(P3,5A) and Orai1(P39,40A) by STIM1 are flattened at negative membrane potentials, with the channel becoming both, inwardly and outwardly rectifying. This is illustrated by calculating the % increase in current on changing the membrane potential from −70 to −100 mV (Fig. 5f), which shows about 25% increase for Orai1 and only 5–7% increase for Orai1(P3,5A) and Orai1(P39,40A). The pairs of prolines (3,5) and (39,40) are required to observe the change in the Orai1 I/V, since the single mutants Orai1(P3A), Orai1(P39A) and Orai1(P40A) behave like wild-type Orai1 when activated by STIM1 (Fig. 5d) or by SOAR (not shown).
The reduced activation of Orai1(P3,5A) and Orai1(P39,40A) by STIM1 is not due to reduced surface expression (Fig. S6a). To the extent tested, the properties of the Orai1(P3,5A) and Orai1(P39,40A) activated by STIM1(344–460) are similar to those of Orai1 activated by STIM1 (Figs. 5b and 5d and Fig. S5d). The pore mutation E106D in the context of Orai1(P3,5A) and Orai1(P39,40A) activated by STIM1 or STIM1(344–460) resulted in a large monovalent current, but with a small current with STIM1 (Figs. S5a–S5c).
The preceding analysis supports a model in which the Orai1 N-terminus interacts with a region of STIM1 that normally inhibits the action of SOAR. We examined a series of mutants and pinpointed the polybasic K-domain. Fig. 5g and 5h and Fig. S6b show that STIM1(ΔK) and point mutations of lysines in the K-domain rescue the reduced activation of Orai1(P3,5A) by STIM1. The same results were obtained with all the STIM1 K-domain mutants with Orai1(P39,40A) (not shown). However, the mutant in which all seven K-domain lysines were substituted by glutamates did not activate Orai1(P3,5) (Fig. 5h). Hence, intact K-domain is required for its interaction with the Orai1 mutants. These observations indicate that the positive charges within the K-domain regulate the ability of SOAR in the context of wild-type STIM1 to activate Orai1. Consistent with this model, deletion of the K domain strongly enhances STIM1 binding to Orai1 (Fig. S5e).
The present study reveals the molecular mechanism by which STIM1 gates the Orai channels. SOAR within STIM1 is required and sufficient for full activation of all Orai channels by interacting with the C terminus of Orai1. The STIM1 C- terminus multimerizes two Orai1 dimers to form the active channel20. STIM1(336–485) exists as a dimer in solution (Fig. S2c). A potential mechanism for gating of the Orais by STIM1 is that the Orai C termini prevent multimerization and assembly of functional channels. Binding of STIM1 molecules through their SOAR domains to the C termini of Orais can serve to deflect the C termini, thereby allowing assembly and opening of the channels. Hence, SOAR within STIM1 actively gates the Orai channels. Thus, clustering of Orai1 is neither sufficient nor necessary for activation of the channel. This is also supported by the function of STIM1(LQ/AA).
Another STIM1 domain, the polybasic K-rich domain, regulates Orai1 and is separate from opening of Orai1 by SOAR. The K-domain regulates the activity of Orai1 at negative membrane potential by communicating with the Orai1 N terminal proline-rich domain. It appears that the K-domain retards interaction of STIM1 with Orai1. In one model, the K-domain interacts with a region of Orai1 or STIM1 to prevent SOAR binding and this is reversed when the K-domain interacts with Orai1 N terminus.
SOAR and the K-domain appear to have different roles in gating the TRPC and Orai channels. SOAR within the STIM1 ERM domain participates in binding of STIM1 to TRPCs, but is not sufficient to activate TRPC1 (Fig. S6c). The STIM1 K-domain is essential for opening of TRPCs12, but inhibits Orai1. Under physiological conditions, the balance of association of the K-domain and SOAR with the two channel types will determine the extent of regulation of Orai1 and TRPC channels by STIM1 and Ca2+ influx.
The TRPC1, STIM1 (wild-type, ΔK, All K/E, 681X), and Orai1 clones were described previously 7, 11. The human Orai2 and Orai3 clones were obtained from Open Biosystems (Orai2 clone #: BC069270 and Orai3 clone #: BC015555). Orai2 and Orai3 were cloned into mCherry-Red-p3XFLAG vector12 using NotI(5’) and SalI(3’). The mCherry-Red is downstream and next to the 3X FLAG and was cloned into the HindIII site of the multi-cloning region of the p3XFLAG vector. SOAR and the various STIM1 fragments were generated by PCR and cloned into the pEGFP-C1 (Clontech) vector using EcoRI (5’ and 3’). Hence, all STIM1 fragments are tagged with EGFP at their N terminus. All point mutations on STIM1 and Orai1 were generated using the site-directed mutagenesis kit (Stratagene). The antibodies used were monoclonal anti-STIM1 (BD Biosciences), polyclonal anti-tubulin (Cell Signaling Technology), monoclonal HRP-conjugated anti-GFP (Santa Cruz Biotech) and monoclonal anti-FLAG and HRP-conjugated anti-FLAG (Sigma-Aldrich). Anti-FLAG antibodies were used for co-IP, while HRP-conjugated anti-GFP and anti-FLAG antibodies were used for Western blotting. The siRNA sequence used to knockdown human Orai1 is: 5’-GCCAUAAGACUGACCGACAGUUCCA-3’. A 6-hour siRNA transfection of HEK293 cells was done using Lipofectamine 2000 (day 1). The amount of siRNA used was 0.8 µg per 12-well with HEK cells at 80–90% confluency. After transferring the cells containing siRNA to 35mm dish on day 2, they were transfected with plasmid for 6 hours on day 3. The total amount of cDNA used per 35 mm dish was 0.5 µg/ml. Localization assays using confocal microscopy was done, current was measured, or cells were harvested and extracted for co-IP analysis the following day (day 4). Thus, the total time for cells exposed to siRNA was 72 hours.
Transfected cells were harvested and lysed using 500 µL of binding buffer: 1× PBS buffer containing 1 mM NaVO3, 10 mM NaPyrophosphate, 50 mM NaF [pH 7.4], and 1% Triton X-100. The cell extracts were sonicated, and insoluble material was spun down at 30,000 × g for 20 min. For the co-IP experiments, 1 µg of myc antibody was added to 100 µL of cell extract and incubated for 1 hr at 4 C. Then, 50 µL of 1:1 slurry of protein G sepharose 4B beads were added to the antibody-extract mix and incubated for an additional hr at 4 C. Beads were washed 3 × 10 min with binding buffer, proteins were released from the beads with 50 µL of SDS-loading buffer. 25 µL was loaded onto 8% tris-glycine SDS-PAGE gels. Gels were transferred onto PVDF membrane, and Western blot analysis was done.
Transfected cells were washed once with 1X PBS on ice. 0.5 mg/mL of EZ-Link Sulfo-NHS-SS-Biotin (Pierce) was added to the cells for 30 min on ice. Afterwards, the biotin was quenched with 50 mM glycine on ice for 10–15 min. The cells were then processed as described above to make cell extract. 50 µL of 1:1 slurry of immobilized avidin beads(Pierce) were added to 100 µL of cell extract and incubated for 2 hrs at 4 C. Beads were washed 3 × 10 min with binding buffer, proteins were released from the beads with 50 µL of SDS-loading buffer. 25 µL was loaded onto 8% tris-glycine SDS-PAGE gels. Gels were transferred onto PVDF membrane, and Western blot analysis was done.
[Ca2+]i was measured about 24 hrs post transfection using a PTI image acquisition system. [Ca2+]i was measured by loading the cells with Fura2 and recording Fura2 fluorescence at excitation wavelengths of 340 and 380 nm and collecting the light emitted at wavelength above 500 nm. [Ca2+]i is expressed as the 340/380 ratio.
After HEK cells were transfected with the chosen EGFP-STIM1 fragment +/− mCherry Red-Orai1, the cells were fixed with 4% paraformaldehyde and mounted onto microscope slides. Images were collected under 400X magnification using the LaserSharp 2000 (Bio-Rad) software and a Bio-Rad confocal microscope.
The STIM1(336–485) sequence was cloned into the pET 28 vector and expressed as an N-terminal Hisx6-fusion protein in BL21 Codon Plus cells (Stratagene) following the manufacture’s guidelines. The fusion protein was purified using TALON metal affinity resin (Clontech) following the manufacturer’s protocol, and the Hisx6 tag was removed using Thrombin (Roche). The STIM1(336–485) fragment was concentrated using Amicon Ultra 10K concentrators (Millipore) and run over a Superdex 200 gel filtration column (GE Healthcare). Light scattering measurements of the STIM1(336–485) fragment were collected with a miniDAWN TREOS (Wyatt Technology) while eluting off of the Superdex 200 gel filtration column in 20 mM Tris pH 8.0, 100 mM NaCl. ASTRA V software (Wyatt Technlogy) was used to analyze the data. The STIM1(336–485) CD spectrum was collected on a Jasco J-810 spectropolarimeter in 10 mM HEPES pH 7.4, 140 mM KCl. The helical content of STIM1(336–485) was calculated using the Greenfield and Fasman method fh = (([Θ]222–[Θ0]222)/[Θ100]222) where [Θ]222 is the observed molar elipticity at 222 nm, [Θ0]222 is the mean molar elipticity of a model peptide with no alpha helix and/[Θ100]222 which is completely helical 24. The 0% and 100% helix estimates used were 2,000 and 30,000 deg·cm2/dmol 25, 26.
The whole cell configuration was used to measure the Orai1 CRAC current in HEK cells co-transfected with Orai1, the Orai1 mutants and with STIM1, SOAR or the STIM1 fragments as detailed before 7, 11, 12. The standard pipette solution contained (in mM): 140 Cs aspartate, 6 MgCl2, 10 BAPTA, and 10 Hepes (pH 7.2 with CsOH). The standard bath solution contained (in mM): 130 NaCl, 5 KCl, 10 CaCl2, 1 MgCl2, and 10 Hepes (pH 7.4 with NaOH). The divalent-free (DVF) solution contained (in mM): 150 NaCl, 10 EDTA, and 10 Hepes (pH 7.4 with NaOH). The current was recorded by 400 ms rapid alterations of membrane potential (RAMPs) from −100 to +100 mV from a holding potential of 0 mV. The current recorded at −100 mV was used to plot the time course of current development and to calculate current density in pA/pF. The averages of multiple experiments are given as mean ±s.e.m of the number of experiments performed.
TRPC1 current was measured in transiently transfected HEK cells by whole current recording, as described previously 11, 12, 27. The pipette solution contained (in mM) 140 CsCl, 2 MgCl2, 1 ATP, 5 EGTA, 1.5 CaCl2 (free Ca2+ 70 nM) and 10 HEPES at pH 7.2 with CsOH. The bath solution contained (in mM) 140 NaCl or 140 NMDG-Cl, 5 KCl, 0.5 EGTA and 10 HEPES at pH 7.4 with NaOH or NMDG-OH−). Cells were transfected with TRPC1 and empty vector or TRPC1 and SOAR. RAMPs of −100 to +100 mV were used to record the TRPC1current. The cells were also transfected with the M3 receptor so that the current can be consistently activated by stimulating the cells with 100 µM carbachol. The current recorded at −100 mV was used to calculate current density as pA/pF.
This work was supported in part by a grant from the National Institutes of Health Grants DE12309 and DK38938 and the Ruth S. Harrell Professorship in Medical Research to S. M. and by the National Institute on Drug Abuse (NIDA; DA00266, DA10309) and the National Institute of Mental Health (NIMH; MH068830) to P. F. W.