Store-operated Ca2+ Entry Mediated by Orai1 and TRPC1 Participates to Insulin Secretion in Rat β-Cells*

Store-operated Ca2+ channels (SOCs) are voltage-independent Ca2+ channels activated upon depletion of the endoplasmic reticulum Ca2+ stores. Early studies suggest the contribution of such channels to Ca2+ homeostasis in insulin-secreting pancreatic β-cells. However, their composition and contribution to glucose-stimulated insulin secretion (GSIS) remains unclear. In this study, endoplasmic reticulum Ca2+ depletion triggered by acetylcholine (ACh) or thapsigargin stimulated the formation of a ternary complex composed of Orai1, TRPC1, and STIM1, the key proteins involved in the formation of SOCs. Ca2+ imaging further revealed that Orai1 and TRPC1 are required to form functional SOCs and that these channels are activated by STIM1 in response to thapsigargin or ACh. Pharmacological SOCs inhibition or dominant negative blockade of Orai1 or TRPC1 using the specific pore mutants Orai1-E106D and TRPC1-F562A impaired GSIS in rat β-cells and fully blocked the potentiating effect of ACh on secretion. In contrast, pharmacological or dominant negative blockade of TRPC3 had no effect on extracellular Ca2+ entry and GSIS. Finally, we observed that prolonged exposure to supraphysiological glucose concentration impaired SOCs function without altering the expression levels of STIM1, Orai1, and TRPC1. We conclude that Orai1 and TRPC1, which form SOCs regulated by STIM1, play a key role in the effect of ACh on GSIS, a process that may be impaired in type 2 diabetes.

The prevalence of type 2 diabetes is increasing at an alarming rate, because of the combination of aging population, urbanization, physical inactivity, and the increasing prevalence of obesity (1). Type 2 diabetes originates from ␤-cell failure to compensate insulin resistance and secrete the necessary amount of insulin to maintain glucose and lipid homeostasis (2,3). The fine-tuning of insulin secretion in response to glucose and other nutrients relies on a balance of signals controlling dynamic variations in the cytoplasmic Ca 2ϩ concentration ([Ca 2ϩ ] c ) (4). The major pathway for [Ca 2ϩ ] c increase involves metabolism of glucose and other nutrients, which leads to ATP production and closure of the ATP-sensitive potassium channels. This in turn results in membrane depolarization and opening of voltage-dependent Ca 2ϩ channels (VDCCs), 2 leading to a rapid elevation of [Ca 2ϩ ] c , resulting in insulin exocytosis (4). Glucose-stimulated insulin secretion (GSIS) is modulated by several hormones and neurotransmitters, among which acetylcholine (ACh) plays a prominent role (5). ACh binding to M3 muscarinic receptors stimulates the phospholipase C, which generates two second messengers, diacylglycerol (DAG), and inositol 1,4,5-trisphosphate (IP 3 ). DAG plays a key role in the activation of various protein kinase C isoforms, whereas IP 3 binding to its receptor (IP 3 R) promotes the release of Ca 2ϩ from the endoplasmic reticulum (ER) (5).
ER Ca 2ϩ store depletion is known to trigger Ca 2ϩ influx across the plasma membrane (PM) through a family of channels referred to as store-operated Ca 2ϩ channels (SOCs) (6,7). It has been widely accepted that the store-operated Ca 2ϩ entry (SOCE) is a major and ubiquitous Ca 2ϩ influx pathway in nonexcitable cells, necessary for the replenishment of intracellular Ca 2ϩ stores (8). Orai1 was identified as the main protein forming SOCs that conducts the previously called Ca 2ϩ release-activated current (I CRAC ) (9). The activity of Orai1 channels is closely controlled by the ER membrane protein STIM1 (stromal interacting molecule 1), which functions as an ER Ca 2ϩ sensor and translocates upon ER depletion to ER-PM regions of Orai1 clustering (10). Some transient receptor potential canonical (TRPC) channels are responsible for less selective storeoperated currents carried by divalent and monovalent cations. Based on sequence similarity and function, the TRPC channel family is divided into two subgroups, TRPC1/4/5 as SOCs and TRPC3/6/7 as receptor-operated channels (11). Interestingly, TRPC1 has been shown to form STIM1-regulated SOCs together with Orai1 (12,13).
Despite the prominent role of VDCCs in GSIS and [Ca 2ϩ ] c dynamics, various aspects of the intracellular Ca 2ϩ responses of ␤-cells to glucose and secretagogues are still unexplained, suggesting the involvement of additional plasmalemmal Ca 2ϩ channels. Early studies indicate that the emptying of intracellular Ca 2ϩ stores in ␤-cells induces SOCE (14 -18). However the composition of the SOCE-mediated channels and their exact role in GSIS is unclear. A few studies performed in mouse models of insulin-secreting cells suggest that ␤-cells express STIM1 and Orai1 (19,20), as well as several TRPC isoforms (18,21). However, the role of STIM1, Orai1, and TRPCs in insulin secretion remains elusive. The aim of this study was to investigate the composition and potential role of SOCs during GSIS and in particular in response to ACh. We showed that Orai1 and TRPC1 form the channels responsible for SOC-mediated Ca 2ϩ entry in ␤-cells and play a major role, together with the regulating protein STIM1, in the potentiating effect of ACh on GSIS. Finally, we demonstrated that prolonged exposure to supraphysiological glucose concentration impaired SOCs function, suggesting SOCs as therapeutic targets to improve ␤-cell function in type 2 diabetes.
Cell Culture-The rat insulinoma cell line INS-1E (27) (kindly provided by Dr. Pierre Maechler (Centre Médical Universitaire, University of Geneva, Geneva, Switzerland) was maintained as previously described (28,29). Islets of Langerhans from male Wistar rats (Janvier, France) were isolated by collagenase digestion and maintained as previously described (28,29). Islets were dissociated in a 1:1 PBS-Trypsin-EDTA solution at 37°C and mechanically dissociated by pipetting for 2 min. Rat care and euthanasia procedures were approved by the Cantonal Veterinary Office (Service de la Consommation et des Affaires Vétérinaires SCAV-EXPANIM, authorization number 2543).
Cell Transfection-INS1-E cells were transiently transfected with plasmids described above using Lipofectamine 2000 (Life Technologies) as previously described (28,29). Cells were then cultured for a 36-h recovery period before being collected or treated as indicated. Empty pmaxGFP or pcDNA3-YFP vectors were used as control to verify whether transfection itself could affect the Ca 2ϩ response.
Confocal Imaging of YFP-STIM1 Fluorescent Protein-For live cell imaging, INS-1E cells grown on glass-bottomed plates were placed in the on-stage incubator of a Zeiss LSM 710 Quasar Confocal inverted microscope (Zeiss, Germany). Photo excitation was achieved by illumination with the 488-nm line of the argon gas laser, and Z-stack time lapse image acquisition was performed using the accompanying Zen2009 software (Zeiss). Projection of Z-stack images was performed using ImageJ 1.48a software by maximum intensity projection of top to middle stack of the cell of interest.
Insulin Secretion and Human Growth Hormone (hGH) Secretion-Static insulin secretion assays were performed in INS-1E cells in KRBH buffer (130 mM NaCl, 4.8 mM KCl, 0.5 mM NaH 2 PO 4 , 5 mM NaHCO 3 , 2 mM CaCl 2 , 1.2 mM MgCl 2, 10 mM HEPES, 2 mM CaCl 2 ) as previously described (29). Perifusion experiments were performed using chambers developed in Prof. Raddatz's laboratory (30). Groups of 20 -30 islets were perifused at 37°C with KRBH solution (1.6 mM glucose) at a flow rate of 500 l/min for 30 min to establish stable basal insulin secretion. Then the glucose concentration was raised to 16.7 mM glucose, and fractions were collected every minute. Insulin release and content were measured using the rat insulin enzyme immunoassay according to the manufacturer's instructions (Cayman Chemical, Adipogen AG, Liestal, Switzerland).
To specifically assess the exocytosis activity of transfected cells only, we used a hGH secretion reporter assay, as previously described (28). INS-1E cells were transiently co-transfected with a construct encoding the hGH, together with an empty vector (pmaxGFP) as control, or a plasmid encoding mutant version of Orai1 or TRPC1. 48 h later, hGH secretion experiments were performed as in static insulin experiments. The hGH (secreted and content) levels were determined using the hGH ELISA kit (Roche Applied Sciences, Switzerland), and exocytosis activity expressed as the ratio of secreted hGH over hGH content.
Cytosolic Ca 2ϩ Monitoring-In most experiments, loading of the cells was performed at 25°C in the dark in KRBH supplemented with 2 mM CaCl 2 , 20 mM glucose, and 3 M Fura-2/AM plus 10% (w/v) Pluronic F-127 (Life Technologies) for 40 min. Following loading, cells were washed and observed under a Zeiss inverted microscope (ϫ40 oil immersion fluorescence objective). Fura-2-AM was excited at 340/380 nm with a Visichrome holographic monochromator, and emission fluorescence was monitored at 510 nm using a Hamamatsu Orca ER coded CCD camera. Images were treated with MetaFluor to evaluate the ratio of fluorescence emitted at 340 and 380 nm (Cellular Imaging Facility, University of Lausanne, Lausanne, Switzerland). To study the thapsigargin or acetylcholine-activated Ca 2ϩ entry, loaded cells were washed in Ca 2ϩ -free KRBH containing 0.1 mM EGTA and depleted with 5 M thapsigargin (Tg) or 100 M ACh in the presence of the ATP-sensitive potassium channel opener diazoxide (D; 250 M) and of the VDCCs blocker verapamil (V; 10 M). Subsequently, 2 mM Ca 2ϩ (in the presence of diazoxide ϩ verapamil (DϩV)) was added to the medium, and the peak amplitude of the fluorometric signal was measured as the ⌬F (340/380 nm ratio). ER Ca 2ϩ depletion was calculated as the area under the curve based on the ⌬F normalized to basal line over time in 0 mM Ca 2ϩ .
PCR Amplification-Total RNA from INS-1E cells or freshly isolated rat islets were extracted using Tripure isolation reagent (Roche Diagnostics) and reverse-transcribed using the ImProm-2 reverse transcription System (Promega AG) as previously described (28,29). PCR was performed using Titanium Taq PCR kit (Takara Bio Europe/Clontech). Primer sequences for classic PCR are provided in Table 1.
Co-immunoprecipitation and Duolink Proximity Ligation Assay-After lysis, 2 g of antibody was added to a volume of lysates containing 300 g of proteins diluted with 600 l of NET solubilization buffer (50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 5 mM EDTA, 0.05% Nonidet P-40 (v/v)) and incubated at 4°C overnight with constant mixing. Then the protein-antibody complex was incubated for 1 h at 4°C with constant mixing with 40 l of protein A/G magnetic beads (Millipore). The immune complexes were collected with a magnetic stand and washed three times in NET, and samples were subjected to SDS-PAGE. The negative control were performed using lysates with beads (ctrl(Ϫ)) without any antibody. For technical reasons, it was not possible to perform immunoblotting of Orai1 because its molecular weight is approximately the same as the IgG light chain.
The Duolink in situ proximity ligation assay (Olink Bioscience, Uppsala, Sweden) is a sensitive method to detect proteinprotein interaction in intact tissue using oligonucleotide-conjugated secondary antibodies that, when in close proximity, allow an in situ polymerization reaction (31). INS-1E cells were fixed for 5 min in Ϫ20°C acetone and processed according to manufacturer's instructions and as previously described (32). The following antibodies were used to test protein-protein interaction: mouse anti-TRPC1 (1:500; Santa Cruz Biotechnology), polyclonal goat anti-Orai1 (1:500; sc-74778; Santa Cruz Biotechnology), and rabbit anti-STIM1 (1:1000; Sigma-Aldrich). The proximity ligation assay fluorescent signal was quantified using the ImageJ software (National Institutes of Health, Bethesda, MD) as follows: 30 images from 4 distinct experiments were converted to a 32-bit format, and the signal to noise ratio was determined by applying the Yen thresholding method. A binary image was then created, and the number of pixels of the duolink signal was measured. The data were normalized to the number of cells (nuclei imaged by DAPI staining) in each image.
Statistical Analysis-The data are presented as means Ϯ S.E. Comparisons were performed by two-tailed unpaired Student's t test and between at least three groups with one-way or twoway analysis of variance completed by Fisher's least significant difference post hoc test for multiple comparisons. A p value of Յ0.05 was considered statistically significant.

Inhibition of SOCs Leads to Impaired GSIS in isolated rat islets and INS-1E cells-
To determine whether SOCs channels contribute to insulin secretion, 20 -30 rat islets were placed in an islet perifusion chamber infused with KRBH supplemented with low (1.6 mM) followed by high (16.7 mM) glucose concentration, together or not with common SOCs inhibitors, SKF-96365 (SKF) or BTP2 (33). SKF and BTP2 impeded the first peak (5-12 min) and nearly fully blocked the second phase (13-30 min) of GSIS (Fig. 1, A and B), thereby reducing total insulin release by about 70 and 60%, respectively (Fig. 1, A and B). Further static insulin secretion experiments performed in INS-1E cells confirmed that SKF and BTP2 partially inhibited insulin secretion induced by 16.7 mM glucose (Fig. 1C). ACh or the SERCA pump inhibitor Tg led to ER Ca 2ϩ depletion and activation of SOCs, and potentiated GSIS in ␤-cells (4,15,34). SKF or BTP2 fully blocked the potentiating effect of 100 M ACh or 1 M Tg in INS-1E cells (Fig. 1C). These data suggest that SOCs participate to GSIS, especially in potentiating mechanisms linked to ER Ca 2ϩ release. Of note, SKF and BTP2 had no impact on insulin content in INS-1E cells (data not shown).
Acetylcholine and Thapsigargin Trigger SOCE in INS-1E and Dispersed Rat Islet Cells-To further investigate the role of SOCs in GSIS, we first studied the effect of ACh and Tg on [Ca 2ϩ ] c at stimulating 20 mM glucose concentration in Fura-2loaded ␤-cells. SOCs were activated by ER Ca 2ϩ store depletion using 2 M Tg or 100 M ACh in the absence of external Ca 2ϩ and then exposed to 2 mM Ca 2ϩ to elicit SOCE in INS-1E cells (Fig. 2, A and B) or dispersed rat islet cells (Fig. 2, C and D) in high glucose conditions. To selectively study SOCE without interference from VDCCs, the cells were hyperpolarized with the ATP-sensitive potassium channel opener diazoxide (D) combined with the VDCCs blocker verapamil (V) (15,35). As shown in Fig. 2, store depletion in Ca 2ϩ -free solution with Tg or ACh induced large Ca 2ϩ release from the ER in INS-1E cells and in dispersed rat islet cells. When extracellular Ca 2ϩ was reintroduced, both Tg-and ACh-mediated depletion stimulated SOCE in INS-1E cells (Fig. 2, A and B) and dispersed rat islet cells (Fig. 2, C and D). In these conditions, i.e. when VDCCs were blocked, SOCs inhibition by SKF or BTP2 completely abolished Tg and ACh-mediated Ca 2ϩ entry (Fig. 2, A-D). In contrast, the SOCs inhibitors BTP2 or SKF had no significant effect on Ca 2ϩ entry stimulated by acute exposure to high glucose in nondepleted INS-1E cells (Fig. 2E). As expected, this Ca 2ϩ entry was reduced by 60% in the presence of DϩV, confirming that high glucose-stimulated Ca 2ϩ entry is essentially due to VDCCs activity (Fig. 2E). This last result also confirmed that the SOCs inhibitors did not affect VDCCs activity.
ACh binding to M3 muscarinic receptors stimulates the PLC␤, which generates DAG and IP 3 . DAG plays a key role in the activation of TRPC3/6/7 channels referred to as receptor-

STIM1/Orai1/TRPC1 and GSIS
operated channels, whereas IP 3 binding to its receptor (IP 3 R) promotes the release of Ca 2ϩ from ER activating SOCs (5). Inhibition of IP 3 R using 20 M xestospongin C (36) reduced AChinduced ER Ca 2ϩ depletion by 70% and decreased the AChinduced SOCE by 60% (Fig. 2F), suggesting that ACh-mediated extracellular Ca 2ϩ influx required IP 3 R-mediated Ca 2ϩ release in INS-1E cells. STIM1, Orai1, and TRPC1 Form a Molecular Complex upon ER Ca 2ϩ Depletion in ␤-Cells-TRPC1, STIM1, and Orai1 were expressed in INS-1E cells and freshly isolated rat islets at the mRNA and protein levels (Fig. 3A). In response to ER Ca 2ϩ store depletion, STIM1 is known to aggregate and translocate to ER-PM region, where it interacts with and activates Orai1-TRPC1 (10,37). To study the intracellular localization of STIM1 in INS-1E cells, we overexpressed a fluorescent YFP-STIM1 protein. As expected, YFP-STIM1 displayed a reticular localization in control condition (11 mM glucose, t0) and rapidly aggregated to discrete regions close to the PM in cells exposed to 1 M Tg or 100 M ACh (Fig. 3B). We also found that TRPC1 co-immunoprecipitated with Orai1 and STIM1 and that STIM1 co-immunoprecipitated with Orai1 and TRPC1, demonstrating the existence of a ternary complex between TRPC1, STIM1, and Orai1 in untreated (data not shown) and in Tg-treated INS-1E cells (Fig. 3C). Proximity ligation assays (31) further demonstrated an increased interaction between Orai1, TRPC1, and STIM1 in cells treated with ACh or Tg (Fig. 3D) compared with untreated cells, indicating that ER Ca 2ϩ depletion promotes the formation of this ternary complex. Of note, as compared with negative controls (ctrl(Ϫ)) where only one antibody against Orai1, TRPC1, or STIM1 was used, INS-1E cells displayed little or no STIM1, Orai1, and TRPC1 interactions in normal conditions (ctrl) (Fig. 3D).
STIM1, Orai1, and TRPC1 Are Critical in SOCE in INS-1E Cells-To assess the specific and respective role of Orai1, TRPC1, and STIM1 in SOCE, native or dominant negative constructs of Orai1, TRPC1, and STIM1 were expressed in INS-1E cells 36 h before performing the same SOCE protocol as in Fig.  2 with (Fig. 4, B-D) or without DϩV (Fig. 4, A and E). We favored a dominant negative strategy to suppress the function (via point mutation in the ionic pore) of a target channel without affecting its endogenous expression, which might induce compensatory regulation of other channels. SOCE was analyzed in Ca 2ϩ imaging experiments in transfected cells only identified by YFP or GFP tag. Of note, GFP/YFP-transfected cells displayed similar SOCE compared with untransfected cells (data not shown). YFP-Orai1 overexpression had no significant effect on SOCE as compared with pmaxGFP-transfected cells (GFP) (Fig. 4A). In contrast, SOCE was increased by 40% in YFP-STIM1-positive cells, and the co-overexpression of YFP-STIM1 and YFP-Orai1 tripled the SOCE amplitude (Fig. 4A), demonstrating a crucial role of the STIM1-Orai1 duet in SOCE. TRPC1 overexpression alone increased SOCE by 40%, and the co-overexpression of YFP-STIM1 and TRPC1 further increased SOCE by 30% compared with TRPC1 alone (Fig. 4A). STIM1 has been proposed to gate TRPC1 through its K-rich domain (38). The overexpression of STIM1 deleted in its K-rich domain (YFP-STIM1-⌬K) alone or the co-overexpression of YFP-STIM1-⌬K with TRPC1 did not increase SOCE (Fig. 4A), suggesting a regulatory role of STIM1 in TRPC1-mediated SOCE.
Overexpression of Orai1-E106D, a dominant negative pore mutant for Orai1 (9), reduced Tg or ACh-induced SOCE to the same extend as BTP2 or SKF, suggesting that Orai1 is the predominant protein forming SOCs in ␤-cells (Fig. 4, B and C). Overexpression of TRPC1-F562A, a dominant negative pore mutant for TRPC1 (25), also reduced by 50% the SOCE triggered by Tg or ACh (Fig. 4, B and C). The combined inactivation of TRPC1 and Orai1 channels had no additive effect (Fig. 4,  B and C), suggesting that TRPC1 and Orai1 are mutually dependent to mediate SOCE. We also observed that TRPC3 is  DECEMBER 18, 2015 • VOLUME 290 • NUMBER 51 expressed in rat ␤-cells (Fig. 4D). TRPC3 blockade using the dominant negative fragment of TRPC3 NTRPC3-YFP (26) had no effect on Tg-stimulated Ca 2ϩ influx (Fig. 4E). These Ca 2ϩ imaging data show that Orai1 and TRPC1, but not TRPC3, mediate SOCE, and that STIM1 constitutes the main trigger to activate Orai1/TRPC1 in INS-1E cells.

STIM1/Orai1/TRPC1 and GSIS
Functional Orai1 and TRPC1 Channels Are Required to Maintain INS-1E Cell Secretory Function-To further investigate the role of Orai1 and TRPC1 in ␤-cell function, we studied the effect of the dominant negative mutants Orai1-E106D or TRPC1-F562A on GSIS in INS-1E cells. To avoid confounding factors caused by transfection efficiency, the exocytosis activity was assessed in transfected cells only, using a reporter system with a plasmid encoding the hGH, which is secreted in the same granules as insulin (28). Orai1-E106D or TRPC1-F562A expression reduced glucose-induced hGH exocytosis by 50 and 30%, respectively, and fully blocked the potentiating effect of 100 M ACh or 1 M Tg (Fig. 5). Of note, selective inhibition of TRPC3 using Pyr3 at 10 M (39) affected neither basal nor stimulated insulin secretion (data not shown). These results strengthen the idea of the specific nature of the store-dependent Ca 2ϩ entry carried by TRPC1 and Orai1 in regulating GSIS and more, particularly in the potentiating effect of ACh.
Chronic High Glucose Leads to ER Ca 2ϩ Depletion and Impaired Tg-induced SOCE in INS-1E Cells-Several lines of evidence underscore a role of chronic hyperglycemia referred to as "glucotoxicity" in dysfunctional insulin secretion and ␤-cell failure in type 2 diabetes (40). INS-1E cells were incubated in the presence of 30 mM glucose for 72 h (G30), a condition known to result in glucose desensitization and impaired GSIS (41,42). To indirectly measure ER Ca 2ϩ content, INS-1E cells were treated with 5 M thapsigargin in the absence of extracellular Ca 2ϩ . The resulting [Ca 2ϩ ] c rise directly depends on the Ca 2ϩ concentration in this organelle (43). This first phase was followed by a second "descending" phase where [Ca 2ϩ ] c returned to basal level (Fig. 6A), which directly reflects the Ca 2ϩ buffering/export capacities of the cells (44,45). As compared with cells maintained at control 11 mM glucose concentration (G11), G30 cells displayed decreased ER Ca 2ϩ release (without changes in the kinetics of the Ca 2ϩ release), suggesting reduced ER Ca 2ϩ stores (Fig. 6A). In addition, the Tg-induced SOCE were decreased by 70% in G30 condition (Fig. 6B). Time course analysis revealed no effect of high glucose concentration on the protein levels of STIM1 (0.98 Ϯ 0.02 at 72 h), Orai1 (0.92 Ϯ 0.02 at 72 h), and TRPC1 (0.89 Ϯ 0.06 at 72 h) (Fig. 6C), suggesting an effect of prolonged high glucose on SOC activity.

Discussion
GSIS relies mostly on Ca 2ϩ entry through VDCCs. However, intracellular sequestration and release of Ca 2ϩ from ER via pathways such as IP 3 R (4, 46) and ryanodine receptors (47)(48)(49) also have important roles in glucose-induced [Ca 2ϩ ] i regulation and insulin secretion. Hence, in mouse ␤-cells, the production of IP 3 induced by ACh triggers a rapid depletion of ER Ca 2ϩ stores, which activates SOCs (14,16,17) and contributes to the amplifying pathway in GSIS (4). It was previously described that in excitable pancreatic ␤-cells, the modest SOCE modulates the membrane potential (15,18) and is significant for the [Ca 2ϩ ] c regulation, suggesting a potential role of such Ca 2ϩ entry in insulin secretion in pancreatic ␤-cells. Currently, the nature of the channels mediating SOCE and their role in ␤-cell physiology is unclear. Here, we demonstrated the crucial role of SOCs in rat ␤-cell function and identified Orai1 and TRPC1 as the main pore-forming subunits of SOCs regulated by the Ca 2ϩ sensor protein STIM1.
The concept of SOCE, proposed in 1986 (50), constitutes the predominant pathway of Ca 2ϩ entry in nonexcitable cells. The  DECEMBER 18, 2015 • VOLUME 290 • NUMBER 51 original role of SOCE was to maintain ER Ca 2ϩ levels. Although still under debate, there is a plethora of evidence that Orai1 forms the Ca 2ϩ selective pore-forming unit of Ca 2ϩ releaseactivated Ca 2ϩ channels and that TRPCs, especially TRPC1 channels, form nonselective cation channels gated by store depletion. STIM1 serves as a Ca 2ϩ sensor in the ER, which clusters proximally to the PM to activate Orai1 and/or TRPC channels when ER Ca 2ϩ stores are depleted (11-13, 25, 51).

STIM1/Orai1/TRPC1 and GSIS
Here, we report that STIM1, TRPC1, and Orai1 mRNAs and proteins are present in rat ␤-cells. Overexpression of YFPtagged STIM1 in rat ␤-cells confirms its translocation and aggregation to discrete regions close to the PM in response to ER Ca 2ϩ store depletion by Tg or ACh. We further show that ER Ca 2ϩ depletion stimulates the formation of a ternary complex composed of Orai1, TRPC1, and STIM1. This is in line with findings in mouse ␤-cells that STIM1 accumulates to subplasmalemmal region to co-cluster with Orai1 in response to ER depletion (19,20). However, here we describe for the first   time in ␤-cells that TRPC1 is an integral part of the STIM1/ Orai1 complex. Overexpression experiments in Ca 2ϩ imaging further confirm that both Orai1 and TRPC1 require STIM1 to form functional SOCs. Of note, the sole overexpression of Orai1 did not significantly increase SOCE, whereas the co-overexpression of YFP-STIM1 and YFP-Orai1 had a synergetic effect on SOCE. It is well described that the STIM1 to Orai1 ratio highly influences their activity. Some reports proposed that a ratio of 2/1 STIM1/Orai1 is optimal for maximal SOCE activity (52,53). Thus, Orai1 overexpression alone may "dilute" the STIM1 binding to individual Orai1 channels, which could explain our data. In contrast, TRPC1 overexpression, alone or in combination with STIM1, increased SOCE, suggesting TRPC1 as an enhancer of SOCE, rather than the main SOCforming partner. Specific blockade of Orai1 using the dominant negative Orai1 (Orai1-E106D) pore mutant abolished Tg-induced SOCE, whereas TRPC1 blockade using the TRPC1-F562A pore mutant only partially blocked Tg-induced SOCE, supporting the major role of Orai1. Moreover, concomitant inactivation of both channels had no further effect on SOCE, which is in accordance with our results and in line with previous reports (54,55), supporting that TRPC1 and Orai1 are mutually dependent to form functional SOCs.
Although STIM1 interacts with and gates Orai1 via a cytosolic SOAR domain (amino acids 344 -442) (12,56), gating of TRPC1 by STIM1 following store depletion is achieved by a lysine-rich region in the STIM1 C terminus ( 684 KK 685 ) (38,57). The overexpression of truncated STIM1 lacking the K-rich domain alone is not able to increase SOCE as the native STIM1. In addition, the co-overexpression of STIM1-⌬K with TRPC1 totally prevented the TRPC1-mediated SOCE. These data suggest a crucial role for STIM1 in store-dependent activation of TRPC1 channels in rat ␤-cells. Altogether these data suggest that Orai1/TRPC1/STIM1 may be key players to regulate the insulin secretion in ␤-cells.
In this study, SOCs were activated by ER Ca 2ϩ store depletion using either the SERCA pump inhibitor thapsigargin or through ACh-mediated IP 3 generation. We observed that ACh elicits robust IP 3 -induced Ca 2ϩ release and triggers similar SOCE as thapsigargin in rat ␤-cells. However blockade of TRPC1 and Orai1 did not fully abolished ACh-induced SOCE, suggesting the involvement of other TRP-dependent Ca 2ϩ channels. Besides SOCE, ACh may also induce receptor-operated Ca 2ϩ entry through activation, by PKC-and/or DAG, of nonselective cation channels formed by other members of the TRPC family (58). This hypothesis is supported by studies showing that ACh triggers both SOC-dependent and SOC-independent inward currents in ␤-cells (14,15,17). We observed that TRPC3, which forms receptor-operated channels regulated by the DAG/PKC pathway (59,60), is not involved in Tg-mediated Ca 2ϩ entry and insulin secretion, supporting the hypothesis that TRPC1 is the main SOC-forming TRPC isoform (54,55,58). We also show that ACh-mediated Ca 2ϩ entry requires IP 3 R-mediated Ca 2ϩ release, suggesting that ACh-induced receptor-operated Ca 2ϩ entry is minimal in ␤-cells. Altogether these data demonstrate that ACh triggers the opening of SOCs made of Orai1 and TRPC1 and regulated by the translocation of STIM1. We further demonstrated that block-ade of Orai1 and TRPC1 via pharmacological inhibition or channel-dead mutant strategies impaired GSIS and fully blocked the potentiating effect of ACh and Tg on insulin secretion in INS-1E cells. SOCs inhibition altered mostly the second phase of insulin release, which is in line with the hypothesis that SOCs are mainly involved in the fine tuning of [Ca 2ϩ ] c during the second phase where ER Ca 2ϩ stores play a prominent role (61). Thus, we propose that SOCs contribute to the cholinergic amplifying pathway of insulin secretion (14,16,17), either through the SOC primordial role in refilling ER Ca 2ϩ reservoirs or because this Ca 2ϩ influx directly shapes [Ca 2ϩ ] c oscillations (7) and pulsatile insulin secretion during the second phase. Our data also suggest a role for SOCs during the first phase of insulin release, when Ca 2ϩ release from the ER is considered to play a minor role (62,63). Of interest, recent reports indicate that STIM1 and Orai1 suppress L-and T-type Ca 2ϩ channel activity in various cell types (64 -66). We also previously showed that TRPC channels inhibit the L-type Ca 2ϩ channel activity in the developing heart (30). In contrast, TRPC3 and TRPC6 channels have been shown, in vascular smooth muscle cells, to trigger sufficient depolarization to open VDCCs (67)(68)(69). Therefore, the Orai1/TRPC1/STIM1 complex may also contribute to insulin secretion via effects on VDCCs activity, which would have an effect on both phases of insulin release. The functional interaction between SOCs and VDCCs in ␤-cells deserves further investigation.
Other TRP channels have been shown to contribute to Ca 2ϩ oscillations and GSIS, including Ca 2ϩ -permeable TRPM2 channel (70), monovalent cations TRPM4 (71) and TRPM5 channels (72), and nonselective cation TRPV4 channels (73). Our results and these previous reports ascertain the co-existence of both voltage-dependent and voltage-independent Ca 2ϩ influx in ␤-cells.
SOCE is a ubiquitous and major mechanism for Ca 2ϩ influx in mammalian cells. As a result, SOCs contribute to the function of all systems and organs and have physiological as well as pathophysiological relevance (7). In particular, accumulating evidence indicates altered SOCs function and/or expression associated with diabetic complications, including diabetes nephropathy, retinopathy, and peripheral neuropathies (74). It is well established that chronic exposure of isolated ␤-cells to hyperglycemic conditions results in glucose desensitization and impaired GSIS (41,42). Although more studies are required to carefully assess the impact of glucose on ER Ca 2ϩ levels, our data demonstrate for the first time that prolonged exposure to supraphysiological glucose concentration results in reduced ER Ca 2ϩ release and decreased SOCE in INS-1E cells. Interestingly, this dysfunction seems independent of STIM1, Orai1, and TRPC1 expression, suggesting that hyperglycemia directly affects the activity of SOCs. SOCE being a key mechanism to replenish the ER Ca 2ϩ reservoirs upon depletion, SOC dysfunction may also result in ER stress, which plays a key role in ␤-cell dysfunction and death (75). Further investigations are required to decipher the pathophysiological mechanisms by which SOCs are regulated in diabetes and to evaluate the therapeutic potential of the ternary TRPC1/Orai1/STIM1 complex.