Mapping the Interacting Domains of STIM1 and Orai1 in Ca2+ Release-activated Ca2+ Channel Activation*

STIM1 and Orai1 are essential components of Ca2+ release-activated Ca2+ channels (CRACs). After endoplasmic reticulum Ca2+ store depletion, STIM1 in the endoplasmic reticulum aggregates and migrates toward the cell periphery to co-localize with Orai1 on the opposing plasma membrane. Little is known about the roles of different domains of STIM1 and Orai1 in protein clustering, migration, interaction, and, ultimately, opening CRAC channels. Here we demonstrate that the coiled-coil domain in the C terminus of STIM1 is crucial for its aggregation. Amino acids 425–671 of STIM1, which contain a serine-proline-rich region, are important for the correct targeting of the STIM1 cluster to the cell periphery after calcium store depletion. The polycationic region in the C-terminal tail of STIM1 also helps STIM1 targeting but is not essential for CRAC channel activation. The cytoplasmic C terminus but not the N terminus of Orai1 is required for its interaction with STIM1. We further identify a highly conserved region in the N terminus of Orai1 (amino acids 74–90) that is necessary for CRAC channel opening. Finally, we show that the transmembrane domain of Orai1 participates in Orai1-Orai1 interactions.

Ca 2ϩ has been long recognized as an important second messenger in regulating a variety of physiological processes (1). In both excitable and non-excitable cells, it has been found that extracellular Ca 2ϩ enters the cytosol in a typical pathway after depletion of the endoplasmic reticulum (ER) 4 Ca 2ϩ store (2), which is therefore termed either capacitative calcium entry (3) or store-operated Ca 2ϩ (SOC) entry (4). The electrophysiological characteristics of this Ca 2ϩ entry were studied in 1992 and termed Ca 2ϩ release-activated Ca 2ϩ current (I CRAC ) (5). The molecular identity and mechanism of SOC or I CRAC have remained a mystery for a long time. Recently, however, several different groups have identified STIM1 (6,7) and Orai1 (8 -10) to be essential components of the CRAC channel.
STIM1 protein contains multiple discrete regions, including an EF-hand and a sterile ␣-motif domain (SAM) in the N terminus (11,12). In addition to a transmembrane region, STIM1 also contains a coiled-coil domain, a serine-proline-rich region, and a lysine-rich region in the C terminus (13)(14)(15), as shown in Fig. 1. Data clearly indicate that the EF-hand domain in the N terminus of STIM acts as a sensor of Ca 2ϩ concentration in the ER (16,17) and participates in the formation of STIM1 puncta (11). It has been proposed that the ezrin/radixin/moesin (ERM) domain in the C terminus, containing the coiled-coil domain mediates the selective binding of STIM1 to TRPC1 and the cationic lysine-rich region, is essential for the gating of the TRPC1 channel (13). However, the precise roles of these domains of STIM1 in the signaling cascade of I CRAC activation are not known. By truncating different domains from STIM1, we address this question in this study.
It is known that STIM1 and Orai1 functionally interact with each other, because overexpressing both proteins (either human or Caenorhabditis elegans homologues) greatly potentiates the I CRAC current in human embryonic kidney 293 (HEK293) cells (18 -20). We and other groups have found that Ca 2ϩ store depletion induces ER STIM1 aggregation and migration toward the plasma membrane (16,21,22), which lead to co-localized clustering of Orai1 on the surface membrane (21,27), implying a physical interaction between STIM1 and Orai1. Recently, Takahashi et al. (23) showed that the N terminus of murine Orai1 is essential for store-operated Ca 2ϩ entry and is postulated to serve as a binding site for the ERM domain of STIM1. In addition, Orai1 is found to exist both as dimers and oligomers in vivo (24,25). However, the domains of Orai1 that account for the STIM1-Orai1 or Orai1-Orai1 interactions remain unknown. Here we explore this field by co-expressing different truncated forms of Orai1 and STIM1, observing their spatial co-localization, and recording reconstituted SOC Ca 2ϩ entry signal and I CRAC current. Stock solutions of thapsigargin (TG) were prepared in Me 2 SO at a concentration of 2 mM. External solutions were applied using a gravity-driven perfusion system, and the local solution was completely exchanged in less than 2 s. Fura-2/AM was purchased from Invitrogen. Unless otherwise specified all reagents and chemicals were from Sigma-Aldrich.
Cell Culture, Transfection, and Fluorescence Imaging-HEK293 cells were cultured as described previously (21) and were transfected with Lipofectamine TM 2000 (Invitrogen) following the manufacturer's instruction. One day prior to the experiment, cells were transferred onto round poly-L-lysinecoated coverslips. Expression levels of pHluorin-tagged STIM1 vectors or EGFP-tagged Orai1 vectors were quantified as the averaged fluorescence intensities obtained from transfected HEK293 cells under wide field illumination, as summarized in supplemental Fig. S1.
For imaging experiments, transfected cells were viewed either under a confocal laser scanning microscope FV500 (Olympus Optical Co., Tokyo, Japan) with 60ϫ oil objective lenses (numerical aperture (NA) ϭ 1.40) or under total internal reflection fluorescence microscope (TIRFM) with 100ϫ oil lenses (NA ϭ 1.45). Detail information regarding the construction and application of confocal and TIRFM are described elsewhere (21). For confocal microscopy experiments, we only showed the middle focal plane throughout this report unless otherwise specified. All images were processed and analyzed using ImageJ (National Institutes of Health).
Single Cell [Ca 2ϩ ] i Measurement-Cells were preincubated with 6 M fura-2/AM at room temperature for 30 min in standard Ringer's solution. [Ca 2ϩ ] i was measured by a dual-wavelength excitation (340/380 nm) photometry system on an inverted microscope (IX71, Olympus, Tokyo, Japan) equipped with polychromatic xenon light source (TILL photonics, Gräfeling, Germany). The emission was collected at 510 Ϯ 5 nm with a photodiode controlled by the TILL photometry system and X-Chart extension of Pulse software (HEKA, Lambrecht, Germany). [Ca 2ϩ ] i was calculated from the background-corrected fluorescence ratio R. Calibration parameters of R min (0.145), R max (4.069), and K eff (3.900 M) were determined previously (28) .
Electrophysiology-Patch clamp experiments were conducted at room temperature using standard whole cell recording configuration. High resolution current recordings were acquired using the EPC-10 (HEKA). After establishment of the whole cell configuration, voltage ramps of 50-ms duration, spanning a range of Ϫ100 to ϩ100 mV, were delivered from a holding potential of 20 mV every 2 s over a period of 300 -500 s. All voltages were corrected for a liquid-junction potential of 10 mV. Currents were filtered at 2.9 kHz and digitized at a rate of 20 kHz. Capacitive currents were determined and corrected before each voltage ramp. The amplitude of the current recorded at Ϫ80 mV was used to monitor the development of I CRAC current. The maximum CRAC currents at Ϫ100 mV were used for statistical analysis. For noise analysis, 300-ms sweeps were acquired at the rate of 20 kHz at a holding potential of Ϫ100 mV, followed by a 50-ms ramp from Ϫ100 to ϩ100 mV. Data were filtered offline at 1 kHz. The mean current and variant were calculated from a 200-ms segment in each 300-ms sweep.
Co-immunoprecipitation and Western Blotting-For transfection-based co-immunoprecipitation assays, HEK293 cells were transfected with the indicated plasmids, lysed in 500 l of lysis buffer (50 mM Tris at pH 8.0, 500 mM NaCl, 0.5% Nonidet P-40, 1 mM dithiothreitol, and protease inhibitor tablets from Sigma-Aldrich), and immunoprecipitated with anti-FLAGagarose beads (Sigma-Aldrich) overnight at 4°C. The beads were washed four times with the lysis buffer and eluted in SDS sample buffer. Proteins in the cell lysates or immunoprecipitated samples were separated by SDS-PAGE, followed by Western blotting with anti-FLAG or anti-GFP (Santa Cruz Biotechnology, Santa Cruz, CA) antibody according to the standard procedures.
Data Analysis-Unless noted otherwise, all current traces were corrected for leak currents. Data analysis was conducted using IGOR Pro 5.01 (Wavemetrics, Portland, OR). Averaged results were presented as the mean value Ϯ S.E. with the number of experiments indicated. Statistical significance was evaluated using Student's t test. Asterisks denote statistical signifi-cance as compared with control, with a p value less than 0.05 (*) and 0.01 (**).

Interrogations of Functions of Different Domains in STIM1-
To study the function of different domains of STIM1 in Ca 2ϩ store depletion-induced STIM1 aggregation and interaction with Orai1, we constructed different truncated forms of STIM1 and co-expressed them with wild-type (WT) Orai1. As shown in Fig. 1A, STIM1-⌬N was devoid of the EF hand and SAM domains in the N terminus but retained the signal peptide; STIM1-⌬C1 was lacking the entire cytosolic C terminus, whereas STIM1-⌬C2 lacked the serine-proline-rich region and the polycationic residue region. The STIM1-⌬C3 mutant was truncated such that it lacked most of the polycationic residues in the C terminus.
In accord with our previous data (21), WT pHluorin-STIM1 displayed diffused ER distribution in the resting state; it migrated toward cell periphery and co-localized with Orai1-mOrange on the plasma membrane after store depletion (supplemental Fig. S2). Deletion of the EF-hand and SAM domain in the N terminus of STIM1 led to a spontaneously targeting of STIM1 to the plasma membrane that was independent of store depletion, as shown in the confocal images obtained from the same cell before and after TG (1 M) pretreatment for at least 5 min ( Fig. 2A). Moreover, the fluorescence intensity profile of STIM1-⌬N and Orai1-mOrange along the line that was imposed across the cell overlapped, indicative of spatial colocalization of these proteins. Consistent with the optical experiment, cells co-expressing WT Orai1 and STIM1-⌬N exhibited a elevated basal [Ca 2ϩ ] i (supplemental Fig. S3A) and a spontaneously active inward current. The channel was blocked by 10 M La 3ϩ (supplemental Fig. S3C) and had a current volt-age relationship similar to native CRAC channel (supplemental Fig.  S3D), indicating that it represented a bona fide CRAC channel. Expressing the pHluorin-tagged ⌬C1 mutant in HEK293 cells resulted in a diffuse ER distribution, which failed to aggregate into clusters upon ER Ca 2ϩ store depletion (Fig.  2B). In contrast, STIM1-⌬C2 clusters were found in non-stimulated HEK293 cells, but many were not targeted to the cell periphery, as manifested by the fact that fluorescence intensity profile of STIM1-⌬C2 peaked in the cell interior in Fig. 2C. After store depletion, STIM1-⌬C2 clusters did not migrate to the surface membrane, neither did they co-localize with the WT Orai1 on the surface membrane. In parallel, SOC Ca 2ϩ influx was inhibited in cells co-expressing Orai1-mOrange with either STIM1 mutant (supplemental Fig. S3, A and  B), and no CRAC current was detected in these cells, either (supplemental Fig. S8).
On the other hand, STIM1-⌬C3 exhibited normal ER distribution in rest condition; upon store depletion, they aggregated and migrated to the cell periphery and co-localized with Orai1-mOrange on the surface membrane, as shown by the overlapped fluorescence intensity profiles of the STIM1-⌬C3 and Orai1-mOrange along the lines that ran across the cell (Fig. 2D). Moreover, the STIM1-⌬C3 mutant was fully capable of reconstituting SOC influx when co-expressed with Oria1 in HEK293 cells (supplemental Fig. S3, A and B). Co-expressed STIM1-⌬C3 with Orai1 also produced large CRAC current (Fig. 3A). The I-V relationship of I CRAC current was not altered by expression of the STIM1-⌬C3 mutant (Fig. 3B), and the averaged current density was also similar in cells co-expressing STIM1-⌬C3 and Orai1-mOrange (30 Ϯ 6 pA/picofarad, n ϭ 8), as compared with cells co-expressing pHluorin-STIM1 and Orai1-mOrange (34 Ϯ 5 pA/picofarad, n ϭ 8). However, intracellular perfusion of BAPTA induced slower initiation of I CRAC current in cells co-expressing STIM1-⌬C3 and Orai1 (delay ϭ 62 Ϯ 11 s), as compared with cells co-expressing WT STIM1 and Orai1 (delay ϭ 104 Ϯ 23 s, p Ͻ 0.05) (Fig. 3D). This effect was not due to a difference in the exogenous fusion protein expressed, because overexpressed STIM1-⌬C3 and WT STIM1 were at a similar level (supplemental Fig. S1A). Taken together, these experiments indicate that different regions in STIM1 play different roles in clustering and migration of STIM and its functional interaction with Orai1 upon store depletion.
Orai1 Interacts with STIM1 at Its Cytoplasmic C Terminus-Orai1 is reported to functionally and physically interact with STIM1 and Orai1 (24,25,29). To locate sites in Orai1 that were responsible for this interaction, we truncated Orai1-EGFP of cytoplasmic C terminus (Orai1-⌬C), N terminus (Orai1-⌬N), and both cytoplasmic C and N terminus (Orai1-D⌬). We also made a cytoplasmic N terminus of Orai1 fused to the trans-membrane domain of Syx1A tagged by pHluorin (Orai1-Nt-TM) (Fig. 1B). As a con-trol, we have previously shown that overexpressed STIM1-mOrange in HEK293 cells aggregated into clusters that co-localized with Orai1-EGFP on the surface membrane after store depletion (21,27). In contrast, as observed using TIRFM in Fig. 4A, Orai1-⌬C on the surface membrane failed to aggregate and colocalize with overexpressed STIM1-mOrange clusters after Ca 2ϩ store depletion by applying another sarcoplasmic-endoplasmic reticulum calcium ATPase pump blocker tert-butylhydroquinone (100 M). Moreover, neither significant storeoperated Ca 2ϩ entry (supplemental Fig. S4, A and B) nor CRAC current (supplemental Fig. S9) was observed in cells co-expressing Orai1-⌬C and STIM1-EGFP. In contrast, truncation of the cytoplasmic N terminus of Orai1 did not affect its cellular distribution, because Orai1-⌬N exhibited normal periphery distribution in rest conditions and co-localized with STIM1 clusters upon store depletion (Fig. 4B). Orai1-D⌬, where the double cytoplasmic region was removed, behaved similar to Orai1-⌬C when overexpressed with STIM1 (data not shown), implying that the cytoplasmic C terminus of Orai1 is crucial for its aggregation and co-localization with STIM1. This is further supported by the fact that overexpressed chimeric protein Orai1-Nt-TM, containing the cytoplasmic N terminus of Orai1, was targeted to the plasma membrane but failed to co-localize with STIM1 clusters after store depletion (Fig. 4C).
Although Orai1-⌬N responded to store depletion with the same spatial rearrangement as the WT Orai1, it failed to reconstitute either SOC Ca 2ϩ entry (supplemental Fig. S4, A  and B) or CRAC current (supplemental Fig. S9) when co-expressed with STIM1, highlighting the essential role of cytoplasmic N terminus in CRAC channel opening.
Amino Acids 74 -90 in the Cytoplasmic N Terminus of Orai1 Are Required for CRAC Channel Activation-By comparing the amino acid sequence of the human, mouse, Drosophila, and C. elegans Orai families (supplemental Fig. S5), we found that the  (STIM1-⌬N, A) or whole cytoplasmic C terminus (STIM1-⌬C1, B), before and after TG stimulation (1 M) in the same HEK293 cells. Images were captured at the middle plane using confocal microscopy. STIM1-⌬N was constitutively targeted to the plasma membrane, while STIM1-⌬C1 was unable to migrate to the cell periphery whether ER store was depleted or not. The right panels show spatial profiles of fluorescence intensity of STIM1 mutants (green) and Orai1-mOrange (red) along the lines imposed on the images. The x axis displays the distance relative to the start point of the line, and the y axis displays the fluorescence intensity (FI). C, deletion of the serine-proline-rich region and the polycationic residues region (STIM1-⌬C2) led to spontaneous STIM1 aggregations independent of Ca 2ϩ store depletion but was deficient in targeting to the plasma membrane. D, STIM1-⌬C3 exhibited normal ER distribution at rest and co-localized with Orai1-mOrange at the cell periphery after store depletion. Each set of images represents five similar experiments. OCTOBER 5, 2007 • VOLUME 282 • NUMBER 40 amino acids from 74 to 90 in the cytoplasmic N terminus of Orai1 were highly conserved among different species. Therefore, we constructed a truncation sparing the conserved cytoplasmic N terminus (Orai1-⌬N73) (Fig. 1B). The amplitude of store-operated Ca 2ϩ entry was increased by 2-fold in HEK293 cells co-expressed with Orai1-⌬N73 and STIM1-mOrange (1.1 Ϯ 0.12 M, n ϭ 20), compared with non-transfected cells (0.36 Ϯ 0.05 M, n ϭ 27, p Ͻ 0.001), as shown in Fig. 5 (A and B).

Interacting Domains of STIM1 and Orai1
Intracellular perfusion of high concentrations of BAPTA induced pronounced whole cell current in cells co-expressing Orai1-⌬N73 and STIM1-mOrange (Fig. 5C), which displayed similar CRAC channel characteristics such as an inwardly rectifying current-voltage relationship (Fig. 5D), an increase in divalent-free current in Ca 2ϩ -free bath solution, and first potentiation then inhibition of the current by 50 M 2-APB (supplemental Fig. S6). However, co-expressing Orai1-⌬N73 and STIM1-mOrange in HEK293 cells only reconstituted ϳ50% of the SOC Ca 2ϩ entry, and ϳ36% of I CRAC current compared with cells co-expressing STIM1-mOrange and Orai1-EGFP (Fig. 5, C and E). Such reduction of SOC Ca 2ϩ entry and CRAC current is not due to a decreased expression of Orai1-⌬N73 (supplemental Fig. S1B), and Orai1-⌬N73 exhibited normal plasma membrane localization similar to WT Orai1, as examined using TIRFM (data not shown). To further explore the underlying mechanism, we analyzed the variance and averages of whole cell current to infer characteristics of a single CRAC channel. In cells overexpressing STIM1 and Orai1, the variance of CRAC current versus its amplitude could be best fitted with a line with slope of 3.0 Ϯ 0.2 fA (Fig. 5G), which represented single CRAC channel current multiplied by the result of 1 subtracted by the channel opening probability (30). This value was not different in cells co-transfected with Orai1-⌬N73 and STIM1 (3.0 Ϯ 0.15 fA, Fig. 5I), indicating that single channel conductance might remain unaltered after truncating part of the N terminus of Orai1.
Functional Interaction between Different Orai1-Endogenous storeoperated Ca 2ϩ entry was severely inhibited in HEK293 cells overexpressing WT Orai1 (70 Ϯ 18 nM, n ϭ 22) compared with non-transfected control cells (360 Ϯ 53 nM, n ϭ 24) (Fig. 6A), presumably reflecting a direct interaction of exogenous Orai1 with endogenous STIM1 or Orai1 that led to a functionally inactivated CRAC channel. Therefore, we probed possible interactions base on this criteria. Transfecting HEK293 cells with Orai1-⌬N was equally effective as WT Orai1 in reducing the amplitude of SOC Ca 2ϩ entry (88 Ϯ 21 nM, n ϭ 21, Fig. 6B). Although the inhibition was less pronounced compared with previous results, overexpressing Orai1-⌬C or Orai1-D⌬ also significantly reduced the amplitude of endogenous SOC Ca 2ϩ entry (218 Ϯ 26 nM, n ϭ 24 and 118 Ϯ 20 nM, n ϭ 24 respectively) (Fig. 6B). It has been recently suggested that the N terminus of mouse Orai1 acts as a dominant negative mutant that inhibits SOC entry (23). Consistent with this idea, SOC Ca 2ϩ influx in cells expressing Orai1-Nt-TM was also significantly suppressed (160 Ϯ 25 nM, n ϭ 21).
Because Orai1-D⌬ was incapable of forming clusters co-localized with STIM1, we wondered if Orai1-D⌬ inhibited SOC influx through interacting with endogenous Orai1. To test that possibility, 3ϫFLAG-tagged WT Orai1 was co-expressed with either EGFP-tagged Orai1-D⌬ or EGFP-tagged transmembrane domain of Syx1A (EGFP-TM). The two forms of Orai1 interacted with one another, as shown by immunoblotting Orai1-D⌬-EGFP, which was co-immunoprecipitated with Orai1-3ϫFLAG (Fig. 6C). As a control, anti-FLAG beads failed to co-immunoprecipitate EGFP in control cells transfected with 3ϫFLAG-tagged Orai1 and EGFP-TM. These results indi-cate that the transmembrane domain of Orai1 is involved in the Orai1-Orai1 interaction.

DISCUSSION
It is firmly established that STIM1 forms aggregates after ER Ca 2ϩ store depletion (6,7,21), which depends on the EF-hand and SAM domains in the N terminus of STIM1 (11,16). In the present study, we further confirm this theory, because deletion of EF-hand and SAM domains leads to spontaneous plasma membrane targeting of STIM1 to the cell periphery and constitutive opening of the CRAC channel ( Fig. 2 and supplemental  Fig. S3). We also found that STIM1-⌬C1 was unable to form aggregates before or after store depletion, whereas STIM1-⌬C2 spontaneously aggregated into clusters in resting condition. The difference is likely to be the presence of the coiled-coil region, which is present in STIM1-⌬C2 but absent in the ⌬C1 mutant, therefore strongly supporting an essential role for the coiled-coil domain in STIM1 aggregation (31). Huang et al. (13) reported that the ERM domain of STIM1 is essential for its binding with the TRPC1 channel, therefore suggesting a similar interaction between STIM1 and Orai1. In our study, Ca 2ϩ store depletion failed to induce STIM1-⌬C2 migration and Orai1 aggregation in cells co-expressing these two vectors, and this phenomenon was rescued by replacing STIM1-⌬C2 with STIM1-⌬C3. Because STIM1-⌬C3 vector contains a serineproline-rich region and part of the ERM domain compared with the ⌬C2 mutant (13), we further delineate these regions as possible STIM1-Orai1 interaction sites.
Human STIM1 has a polycationic tail, which is found to be essential for the gating of the TRPC1 channel (13). Many membrane-residing small GTPase proteins share a similar structure and use this polycationic tail to guide their targeting to the phosphatidylinositol 4,5-bisphosphate-enriched region in the plasma membrane (32). Therefore, it is tempting to speculate that this polycationic region determines the targeting of STIM1 to the cell surface. In agreement with this hypothesis, both TGinduced migration of STIM1-⌬C3 puncta to the cell periphery (data not shown) and development of I CRAC current induced by intracellular perfusion of BAPTA were slowed down, implying a reduced efficiency in the trafficking of STIM1 to the cell periphery. However, the amplitude of averaged I CRAC current reconstituted by STIM1-⌬C3 and Orai1 was not significantly different from that reconstituted by STIM1 and Orai1, indicating that the tail region is not required for CRAC activation. It is possible that the presence of the single basic amino acids in the tail region (amino acid 671) in our constructs, which was absent in previous studies (13), provides enough guiding signal for STIM1. However, Yeung et al. (33) showed that reducing the cationic residues in the K-ras tail down to two residues released it from plasma membrane subdomains, and at least four cationic residues in the tail were required for obvious binding of the protein to the cell periphery. Moreover, this polycationic region is not conserved in C. elegans or Drosophila (34). However, co-expressing STIM1 and Orai1 from C. elegans reconstitutes large I CRAC current in HEK293 cells (20). Therefore, we favor the concept that the highly conserved serine-proline-rich region serves as a ubiquitous targeting signal of STIM1, and the FIGURE 4. The cytoplasmic C terminus of Orai1 is essential for its co-localization with STIM1. TIRFM was used to examine distributions of EGFPtagged Orai1 mutants and mOrange-tagged STIM1 mutants before or after store depletion. A, Orai1-⌬C was distributed throughout the surface membrane at rest (upper left panel). Depletion of the ER Ca 2ϩ store induced STIM1-mOrange aggregation into puncta underneath the plasma membrane, whereas Orai1-⌬C failed to co-localize with STIM1. As shown in the merged image (right), there is little overlap of Orai1 and STIM1 either before or after store depletion. B, Orai1-⌬N and STIM1-mOrange exhibited normal distribution at rest. After store depletion, Orai1-⌬N (left) and STIM1 (right) aggregated into co-localized clusters. C, Orai1-Nt-TM distributed homogenously throughout the surface membrane before and after store depletion and was incapable of forming clusters that co-localized with STIM1-mOrange. Each set of images represents at least five similar experiments. OCTOBER 5, 2007 • VOLUME 282 • NUMBER 40 polycationic tail is a vertebrate adaptation that helps STIM1 migration but not required for CRAC activation.

Interacting Domains of STIM1 and Orai1
We also looked at functions of different domains of Orai1 in CRAC activation. It has recently been reported that the N terminus of murine Orai1 is essential for the store-operated Ca 2ϩ entry (23). In agreement with that report, we also found that truncation of the cytoplasmic N terminus of Orai1 failed to reconstitute SOC Ca 2ϩ entry when co-expressed with STIM1. Because there are five positively charged arginine residues in the N terminus of Orai1, it was postulated previously that these residues might bind to the ERM domain of STIM1 (23). However, truncation of the whole cytoplasmic N terminus did not affect store depletion-induced aggregation of Orai1 and its colocalization with STIM1, whereas deletion of the cytoplasmic C terminus diminished such responses. Conversely, overexpressing either Orai1-⌬N or Orai1-⌬C in HEK293 cells inhibited endogenous SOC Ca 2ϩ entry. This inhibition was reduced in cells co-expressing Orai1-⌬N and STIM1 but was not altered in cells co-expressing Orai1-⌬C and STIM1 (supplemental Fig.  S7). Moreover, the plasma membrane-tethered N terminus of Orai1 did no co-localize with STIM1 clusters formed after store depletion (Fig. 4C). These three lines of evidence suggest that it is the cytoplasmic C terminus and not the N terminus of Orai1 that is crucial for the STIM1-Orai1 interaction. Rather, we believe that the N terminus of Orai1 is essential for the opening of CRAC channel. We identified a highly conserved region (amino acids 74 -90) in the N terminus that is absolutely required for I CRAC activation. Because this region is conserved in human Orai1, Orai2, and Orai3, it may well explain the phenomenon wherein different Orai genes generate or augment store-operated Ca 2ϩ entry and reconstitute I CRAC current (25,(35)(36)(37). The reconstitute whole cell I CRAC current was smaller in cells co-expressing STIM1 and Orai1-⌬N73, compared with cells co-transfected with STIM1 and Orai1. Because the unitary single CRAC channel current revealed by the variance analysis of mean whole cell CRAC current was not altered (Fig. 5), the reduction in current could be due to a decreased open probability of single CRAC channel or an increase in the number of silent CRAC channels on the surface membrane (30). Whether any one of these hypotheses is true requires further investigation.
Orai1 forms dimers or oligomers in the non-stimulated resting state and store-depleted state (24,25). Functionally overexpressing Orai1 in HEK293 and HeLa cells is found to inhibit endogenous SOC Ca 2ϩ entry (18,35), implying that overexpressed Orai1 may interact with endogenous Orai1 and become insensitive to the store-depletion signal. In the present study, TG-induced endogenous SOC Ca 2ϩ entry was significantly inhibited in cells overexpressing Orai1-⌬C or Orai1-D⌬. Because Orai1-⌬C and -D⌬ are deficient in their interaction with STIM1, the inhibition could only be due to interaction of endogenous Orai1 with these truncations. Moreover, coimmunoprecipitation experiments directly showed that Orai1-D⌬ physically interacted with full-length Orai1. On the other hand, endogenous SOC Ca 2ϩ entry was inhibited by overexpressing Orai1-Nt-TM (Fig. 6, A and B), implying that the cytoplasmic N terminus may also participate in Orai1 oligomerization. Taking these findings together, we conclude that the transmembrane domain of Orai1 participates in Orai1 dimerization and oligomerization but does not exclude roles of other domains in the same processes.
In conclusion, after Ca 2ϩ store depletion, STIM1 aggregation requires the coiled-coil domain in the C terminus of STIM1. STIM1 migrates toward the surface membrane under guidance of both the serine-proline-rich region and the polycationic region, interacts with the cytoplasmic C terminus of Orai1, and leads to clustering of the latter on the opposing plasma membrane. Although the N terminus of the Orai1 is not required for its interaction with STIM1, amino acids 74 -90 are essential for the opening of the CRAC channel. Finally, we propose that the transmembrane domain of Orai1 participates in Orai1-Orai1 interactions. . Orai1 interact with each other using the transmembrane domain. A, averaged SOC influx in HEK293 cells expressing WT Orai1-EGFP or Orai1 mutants after depletion of intracellular Ca 2ϩ store by TG. Non-transfected HEK293 cells were used as controls. Numbers of cells tested for each experiment are labeled on the figure. B, histograms of averaged peak Ca 2ϩ elevation of HEK293 cells in the indicated categories. * and **, statistical significance as compared with control. C, co-immunoprecipitation analysis of interaction between 3ϫFLAG-tagged WT Orai1 and EGFP-tagged Orai1-D⌬. EGFP-tagged transmembrane domain of Syx1A was used as a control. HEK293 cells were co-transfected with plasmids at a ratio of 1:1 encoding Orai1-3ϫFLAG and EGFP-TM or Orai1-3ϫFLAG and Orai1-D⌬-EGFP. Western blots of 3ϫFLAG-tagged wild-type Orai1 protein in whole cell lysate (lower) and EGFP-tagged Orai1-D⌬ (upper) using the indicated antibodies. OCTOBER 5, 2007 • VOLUME 282 • NUMBER 40 JOURNAL OF BIOLOGICAL CHEMISTRY 29455