A role for Orai in TRPC-mediated Ca²⁺ entry suggests that a TRPC:Orai complex may mediate store and receptor operated Ca²⁺ entry

Abstract

TRPC and Orai proteins have both been proposed to form Ca²⁺-selective, store-operated calcium entry (SOCE) channels that are activated by store-depletion with Ca²⁺ chelators or calcium pump inhibitors. In contrast, only TRPC proteins have been proposed to form nonselective receptor-operated calcium entry (ROCE) cation channels that are activated by Gq/Gi-PLCβ signaling, which is the physiological stimulus for store depletion. We reported previously that a dominant negative Orai1 mutant, R91W, inhibits Ca²⁺ entry through both SOCE and ROCE channels, implicating Orai participation in both channel complexes. However, the argument for Orai participating in ROCE independently of store depletion is tenuous because store depletion is an integral component of the ROCE response, which includes formation of IP3, a store-depleting agent. Here we show that the R91W mutant also blocks diacylglycerol (DAG)-activated Ca²⁺ entry into cells that stably, or transiently, express DAG-responsive TRPC proteins. This strongly suggests that Orai and TRPC proteins form complexes that participate in Ca²⁺ entry with or without activation of store depletion. To integrate these results with recent data linking SOCE with recruitment of Orai and TRPCs to lipid rafts by STIM, we develop the hypothesis that Orai:TRPC complexes recruited to lipid rafts mediate SOCE, whereas the same complexes mediate ROCE when they are outside of lipid rafts. It remains to be determined whether the molecules forming the permeation pathway are the same when Orai:TRPC complexes mediate ROCE or SOCE.

The role of Orai proteins in store operated Ca²⁺ entry has been firmly established (for reviews see refs. 1–3). Recent studies suggest that Orai forms a tetrametic store operated Ca²⁺ channel (4, 5) and this formation may be aided by STIM1, a sensor for endoplasmic reticulum Ca²⁺ content, which facilitates the dimerization of preexisting Orai dimers in the plasma membrane (5). However, there also exists a body of evidence stemming from transient (6) and stable (7) overexpression studies, as well as acute (8), transient (9), and stable (10, 11) suppression studies that strongly implicate participation of TRPC channels in store operated Ca²⁺ entry.

Store operated Ca²⁺ entry (SOCE) is closely related to receptor operated Ca²⁺ entry (ROCE), and it is not known to what extent store operated Ca²⁺ entry contributes to entry of Ca²⁺ after activation of phospholipase C (PLC) in response to exposure of cells or tissues to agents that trigger phosphoinositide hydrolysis with formation of inositol 1,4,5-trisphosphate (IP3) plus diacylglycerol (DAG). From the electrical viewpoint, activation of receptors coupled to intracellular events by the Gq or Gi class of G proteins is followed by activation of a nonselective cation current (cf. 12, 13). This current is transient and inactivates with a time course resembling that reported for activation of a Ca²⁺-release activated Ca²⁺ current (Icrac) (13–15). Ca²⁺ entry during this time increases and continues for as long as agonist is present (16), consistent with a switch from entry carried by nonselective Ca²⁺ permeable cation channels formed by TRPCs to entry mediated by CRAC channels proposed to be formed by TRPCs or Orai proteins, or, as we have proposed, formed by complexes of TRPCs with Orai proteins (17).

Mammalian Orai1 was discovered by 2 approaches: an RNAi suppression screen (15) and the combination of an RNAi suppression screen and positional cloning of a familial mutation responsible for severe combined immunodefficiency (SCID; ref. 18). Not surprisingly, expression of the SCID mutant in normal HEK-293 cells interferes with thapsigargin-stimulated SOCE (19), presumably by interfering with the formation of normal Orai1 multimers. The participation of TRPCs in store operated Ca²⁺ entry involving Orai1 was inferred by us from studies in which expression of exogenous Orai1 increased thapsigargin-stimulated Ca²⁺ entry only in cells stably overexpressing a TRPC (TRPC3 or TRPC6) but not in control cells (17, 19), i.e., by showing an effect of Orai1 that is dependent on the TRPC status of the cells.

Here we expand on these studies by providing further connections between Orai1 and TRPCs. We took advantage of the fact that a low concentration of GdCl₃ (1 μM) blocks SOCE, but not ROCE mediated by TRPC channels, to show that Orai1, expressed at the level that enhances SOCE, leads to appearance of Gd³⁺-resistant ROCE and that the SCID mutant of Orai1, Orai1[R91W], not only inhibits SOCE and ROCE in HEK293 cells but also Ca²⁺ entry elicited by activation of TRPC3 with 1-oleoyl-2-acetyl-sn-glycerol (OAG). OAG-induced Ca²⁺ entry is specific for the TRPC3 subfamily of TRPCs (TRPC3, TRPC6 and TRPC7) (13, 20). We propose a model that incorporates our data and recent data from the literature indicating that SOCE appears to occur through channels located in lipid rafts whereas ROCE may occur outside of the lipid rafts.

Results

In previous studies we had found that expression of low levels of Orai1 [corresponding to a transfection with 50–60 ng of input expression plasmid in our transfection protocol (17)] enhances SOCE in HEK-293 cells expressing TRPC3 or TRPC6 in stable form. We expanded our studies to 2 additional TRPCs,TRPC1 and TRPC7, by testing the effect of Orai1 in cells stably expressing them. We demonstrate that Orai1 enhanced SOCE in both cell lines, with the exception that cells expressing TRPC7 responded to 5-fold lower inputs of Orai1. As a result, at plasmid levels that are optimal for TRPC3 expressing cells, the TRPC7 cells were either unaffected or slightly inhibited by Orai1 (Figs. 1 and 2).

Fig. 1.

Enhancement of SOCE in HEK293 cells stably expressing TRPC1 (T1–8 cells). Cells were cotransfected in 60 mm dishes with empty pcDNA3 (Clontech) or 60 ng of pOrai1 (pcDNA3 carrying the cDNA of wild-type Orai1 under the control of the CMV promoter) and peYFP as described (17). After 24 h, cells were replated onto coverslips for an additional 24 h. SOCE was monitored in cells loaded with Fura2 by dual wavelength ratiometric fluorescence video microscopy as described (44, 17).

Fig. 2.

Effect of varying the input pOrai1 on SOCE in HEK293 cells stably expressing TRPC7 (T7–2 cells).

We reported further that mutant Orai1 not only inhibited SOCE but also ROCE (19). These findings strongly suggested that the physiological roles of Orai may not be restricted to its participation in SOCE. However, because ROCE includes an aspect of SOCE triggered by IP3-induced store depletion, we sought another method to stimulate TRPC function that would not include a store depletion phase. One such form of Ca²⁺ entry mediated by TRPC is the DAG-stimulated Ca²⁺ entry seen in cell lines expressing one of the members of the TRPC3 subfamily of TRPCs. We found that the SCID mutant inhibits the OAG-activated Ca²⁺ entry. This is shown for TRPC3 expressed stably (Fig. 3A) or transiently (Fig. 3B) and was seen also in TRPC6 expressing cells (data not shown). Wild-type Orai1 did not affect Ca²⁺ entry stimulated by OAG (Fig. 3A).

Fig. 3.

(A) Orai1[R91W] inhibits Ca²⁺ entry triggered by OAG (100 μg/ml) in HEK cells that stably express TRPC3 (clone T3H1), and (B) in HEK-293 cells that transiently express TRPC3. (A) T3H1 cells were transfected with 0.1 μg peYFP (Control, deep blue), 0.1 μg peYFP plus 60 ng of pOrai1[R91W] (R91W, bluish-green), or with 0.1 μg peYFP plus 60 ng pOrai1 (wild-type Orai1, red) and analyzed for their response to OAG. (B) HEK-293 cells were cotransfected with 1.0 μg pTRPC3 and 0.1 μg peYFP (TRPC3, deep blue) or with 1.0 μg pTRPC3, 60 ng pOrai1[R91W] and 0.1 μg peYFP (TRPC3 + W91R, blue-green) and analyzed for their response to OAG. Transfections were in 60 mm dishes as described under Methods. The OAG responses were tested 48 h after transfection. The OAG activation protocol was as described (17). Similar results were obtained with HEK-293 cells stably expressing TRPC6 (data not shown).

Finally, we reported also that expression of high levels of Orai1 and STIM1 cause the appearance of Gd³⁺-resistant ROCE (19). This phenomenon was explored further. We found that, even though our initial experiments indicated that both STIM1 and Orai1 were needed at high concentrations such as those achieved with 1 μg of each of the input expression plasmids, Orai1 can elicit Gd³⁺-resistant ROCE at low levels of expression in the absence of STIM1 (Fig. 4Right) and that Gd³⁺-resistant ROCE can be generated in HEK293 cells by transfecting as little as 25–50 ng of the Orai1 expression plasmid (Fig. 5). This is the same amount as was used to obtain TRPC-dependent enhancements of SOCE in cells expressing exogenous TRPCs in stable form. High amounts of the Orai1 expressing plasmid are inhibitory even in the presence of Gd³⁺ (bottom trace of Fig. 5).

Fig. 4.

Wild-type (wt) Orai1 causes the appearance of Gd³⁺-resistant ROCE in HEK-293 cells and Orai1[R91W] inhibits ROCE both in the absence (Left) and the presence (Right) of 1 μM Gd³⁺. Note: no exogenous TRPC was present.

Fig. 5.

Induction of Gd³⁺-resistant ROCE by varying the amount of input pOrai1 shown in ng per 60 mm dish. All transfections also contained constant amounts of plasmids directing the expression of exogenous V1a vasopressin receptor and peYFP. At the concentrations tested the Gd-resistant ROCE was maximal at 50 ng pOrai1 (trace #5). Ten-fold higher amounts of plasmid were inhibitory (trace #6). Trace #1, (ROCE in the absence of Gd³⁺) was obtained from cells transfected only with peYFP and pV1aR. Note: that no exogenous TRPC was present.

In other experiments we found by real-time qPCR that coexpression of high levels of Orai1 and STIM1, which enhanced SOCE by more that 10-fold (19), did not significantly affect the endogenous levels of mRNAs coding for expression of endogenous TRPC1, TRPC3, TRPC4, or TRPC6 mRNA, which are the TRPC channels expressed in our HEK-293 cells (data not shown).

Discussion

Fig. 6 summarizes our hypothesis that the location of a TRPC:Orai complex determines whether it functions as a SOCE or a ROCE channel. TRPC:Orai channels activated outside the confines of lipid rafts are assumed to mediate ROCE, whereas channels located in lipid rafts mediate SOCE. The conversion from ROCE to SOCE and translocation from the nonraft to the raft domains may be mediated through STIM binding and dimerization of the preexisting Orai dimers, which might be the predominant form of TRPC:Orai channels that mediate ROCE.

Fig. 6.

Model of TRPC (purple):Orai (red) complexes operating as ROCE channels outside of lipid rafts and as SOCE (CRAC) channels within the confines of lipid rafts to which the complex is recruited by the multifunctional C terminus of activated STIM1 (blue; ref. 26). STIM1 is shown interacting by coiled-coiled domain interaction with the C terminus of Orai1 (yellow; ref. 27) and by ionic interaction of its C-terminal KK dipetide with the DD dipeptide of TRPC (light blue; ref. 31). The IP3 receptor (cyan) is shown in its IP3-activated form that confers to it the ability to interact with a region of TRPCs located in their C-termini (pink; ref. 34).

Evidence has accumulated to indicate that the channel complexes responsible for SOCE and ROCE do not assemble in the same membrane compartments. Specifically, store depletion causes (i) STIM1 to aggregate into submembranous puncta (21, 22) and (ii) the translocation of Orai1 (23, 24) and TRPCs (25, 26) to these puncta. As a consequence, STIM1 and Orai1, as well as STIM1 and the TRPC molecules, come into close enough proximity to appear as colocalized when viewed through a confocal microscope to allow for Foerster resonance energy transfer (FRET) from STIM1-CFP to Orai1-YFP (23, 27) and from STIM1-CFP to TRPC1-YFP (26) and to allow for 3-way coimmunoprecipitation from normal human platelet lysates of STIM1 with Orai1 and TRPC1, of Orai1 with STIM1 and TRPC1, and of TRPC1 with STIM1 and Orai1 (28) or simple coimmunoprecipitation of TRPC1 with STIM1 and Orai1 as seen in human submaxillary gland cells (29).

The channels assembled in this manner are not only aggregates of the required proteins but concentrate in domains rich in cholesterol and sphingomyelins referred to as lipid rafts (24, 26, 30). Participation of a TRPC in thapsigargin-activated, store-operated currents (SOC) had also been inferred from TRPC1 knockdown experiments in human submaxillary gland cells (29) and from TRPC1 knockout experiments in mouse salivary glands (11) in which SOCE was reduced.

Initial agonist induced ROCE, which by definition does not involve STIM1, finds TRPCs (26, 30) and Orai proteins (24) diffusely distributed in the nonlipid raft domains of the plasma membrane from where they are recruited to lipid rafts by activated STIM1. The slow development of Icrac is consistent with the need for redistribution of plasma membrane proteins under the influence of activated STIM1 and assembly of the proteins forming the influx channel.

At the molecular level, the C terminus of STIM1 has been shown to interact with the C terminus of Orai1 through their respective coiled coil domains (27) and with the C terminus of TRPC1 through the interaction of the C-terminal KK dipeptide of STIM1 and a DD dipeptide of TRPC1 located very close to the TRPC TRP box (31).

Assembly of the Orai:TRPC complex in lipid rafts appears to be an obligatory event for the development of a store operated current SOC, as disruption of lipid rafts prevents the development of both the STIM1-to-TRPC1 FRET signal and the store depletion-activated current (26).

Activation of ROCE is rapid with time lags in the order of a few seconds at most. It presumably arises from entry occurring outside of lipid rafts, although there is no a priori reason that activation could not also occur in the environment of lipid rafts. We presume that ROCE is, at least initially, carried by TRPC channels activated secondarily to PLC activation. Despite intensive studies, a unifying mechanism by which PLC activation leads to TRPC activation has not yet been found. Some TRPCs can be activated by DAG formed by the Gq/Gi-activated PLCβ [these include TRPC2 (32), TRPC3 (13), TRPC6 (13), TRPC7 (20) and possibly also Drosophila TRP channels (33)]. Other TRPCs, e.g., TRPC4, TRPC5, and TRPC1, do not appear to be activated by a DAG. All TRPCs have the potential to interact with an N-terminal region of the IP3 receptor (34, 35), which, as shown for TRPC3, has the ability to displace inhibitory Ca²⁺/calmodulin (36). Yet, this is still not likely to be a common mechanism of activation of all TRPCs, including Drosophila TRPs, as Drosophila photoreceptor cells lacking the IP3 receptor are activated by light (37). We hypothesize that rapid activation of TRPC channels may arise therefore by a combination of DAG-mediated activation and a protein::protein interaction between the TRP channel molecules and PLCβ. In support of the latter, TRPC4 has been coimmunoprecipitated with PLCβ from mouse brain lysates (38), and TRPC1 and PLCβ have been coimmunoprecipitated with caveolar complexes from salivary gland cell lysates (39).

One of our current, rather unexpected findings is the interference by Orai1[R91W] with activation of TRPC3 and TRPC6 by OAG. This finding adds Orai1 to TRPCs in the picture of ROCE. One would not expect Orai1[R91W] to interfere in a dominant negative way with OAG activation if a multimeric Orai were not part of the channel formed by TRPCs. We propose that Orai:TRPC complexes operating in lipid rafts are inhibited by Gd³⁺, and those operating outside of lipid rafts, if formed by Orai1:TRPC3/6/7, show resistance to the lanthanide.

Neither expression of low levels of Orai1 nor that of STIM1 in HEK293 cells (19) affect their SOCE. Expression of high levels of Orai inhibits SOCE. Coexpression of high levels of Orai1 and STIM1 relieves the inhibition by excess Orai1 and increases SOCE at least 10-fold (19, 40–43). Orai and STIM1 levels are therefore balanced in HEK-293 cells. As a result, appearance of Gd³⁺-resistant channels upon expression of low levels of Orai1 requires association to preexisting “unpaired” TRPC channels.

The other finding reported here is the appearance of Gd³⁺-resistant ROCE upon expression of Orai1. Coexpression of Orai1 and STIM1 leads to greatly enhanced SOCE and Icrac. Thus, expression of Orai1 in a cell expressing superstoichiometric levels of a TRPC from the TRPC3 subfamily could lead to the accumulation of Gd³⁺-resistant TRPC:Orai1 complexes as we found (Figs. 4 and 5), which we presume to occur outside of lipid rafts (Fig. 6). Recruitment to lipid rafts depends on STIM1 (26). Thus, under condition of ROCE, and without an overexpression of STIM1, the assembly of the CRAC conformation of the complex is not favored.

Fig. 7 presents a pictorial view of the sequence of events that we propose lead from a resting cell to a stimulated cell, first allowing Ca²⁺to enter through the ROCE configuration of the TRPC:Orai channel followed by reconfiguration into a CRAC channel within the confines of lipid rafts. Although the data is not shown, the initial ROCE channel may be activated not only outside of lipid rafts but also within their confines.

Fig. 7.

Model of the events triggered by phospholipase Cβ activated by a Gq/Gi-coupled GPCR (receptor signaling shown by red arrow). Nonactivated TRPC channels (1) are depicted as TRPC:Orai complexes based on our finding that spontaneous activities of TRPC3 channels (17) and TRPC6 channels (C. Erxleben and D.L. Armstrong, personal communication) are reduced by expression of low levels of Orai1. The receptor-activated TRPC (2) is shown in association with Orai based on the finding reported here that a dominant negative variant of Orai (Orai1[R91W]) inhibits ROCE as well as DAG activated Ca²⁺ entry. Activated STIM1 (3) is shown as a dimer with a dual TRPC and Orai interacting C terminus (dark green) to reflect multimerizaton and activation of its TRPC and Orai interacting ability induced by loss of Ca²⁺ caused by active store depletion (4) triggered by IP3 formed by PLC activity or passive store depletion without activation of PLC (light blue arrows). The activated CRAC channel (5) is shown in a lipid raft to which it is directed by activated STIM1 molecules. (Modified from figure 6 in ref. 19).

Further studies are required to describe more accurately the molecular makeup of the channels mediating ROCE, including answering the question as to which of the molecules carries the ionic currents. In the most formal analysis, there is the same evidence available for TRPCs forming ion channels as for Orai proteins forming ion channels.

Methods

Cell and Molecular Biology.

For transfections, cDNAs coding for eYFP (Clontech), V1a vasopressin receptor (V1aR), myc-tagged Orai and mutants, and STIM1 were placed into pcDNA3 (Invitrogen). For expression of proteins in HEK and HEK derived cells, cells were grown in 60 mm dishes to ca. 80% confluence and were transfected (Lipofectamine method, Invitrogen) with a mixture of plasmids that included plasmids coding for eYFP (0.1–0.5 μg), the indicated plasmids coding for wild-type or mutant Orai1, TRPC3 tagged with the HA epitope at the C terminus, or combinations thereof and varying amounts of pcDNA3 without insert (empty vector) to give 5.0 μg total plasmid DNA. For ROCE measurements, the V1aR cDNA in pcDNA3 (pV1aR, 1.5 μg) was included in the transfection mixture. For measurement of [Ca²⁺]_i, the cells were replated 24 h after the transfection onto polylysine coated coverslips and allowed to recover for another 24 h before preparing them for SOCE or ROCE measurements. Loading with Fura2 was done by incubation with 2 μg Fura2 a.m. in Hepes buffered saline for 30 min at 37 °C. Changes in [Ca²⁺]_i were monitored by dual excitation fluorescence ratiometric video microscopy as described in refs. 44 and 17.

Quantitative reverse transcription-PCR (qPCR) was performed on total RNA from control mock transfected HEK-293 cells and Orai1 plus STIM1 transfected HEK-293 cells, isolated 48 h after transfection, with reagents and gene specific primers (TRPC1, cat # PPH15081A; TRPC3, cat #PPH12808E; TRPC4, cat #PPH15312A; TRPC5, cat #PPH14971A; TRPC6, cat# PPH13135A; TRPC7, cat#PPH15892A; GAPDH, PPH00150E) purchased from SuperArray Bioscience Corporation following the vendor's instructions. Thermal cycling was performed in a BioRad iCycler using the cycling protocol suggested by the supplier of the reagents.

Other materials and all other methods were as in ref. 17.