Molecular Analysis of a Store-operated and 2-Acetyl-sn-glycerol-sensitive Non-selective Cation Channel

We have reported that internal Ca2+ store depletion in HSY cells stimulates a nonselective cation current which is distinct from ICRAC in RBL cells and TRPC1-dependent ISOC in HSG cells (Liu, X., Groschner, K., and Ambudkar, I. S. (2004) J. Membr. Biol. 200, 93–104). Here we have analyzed the molecular composition of this channel. Both thapsigargin (Tg) and 2-acetyl-sn-glycerol (OAG) stimulated similar non-selective cation currents and Ca2+ entry in HSY cells. The effects of Tg and OAG were not additive. HSY cells endogenously expressed TRPC1, TRPC3, and TRPC4 but not TRPC5 or TRPC6. Immunoprecipitation of TRPC1 pulled down TRPC3 but not TRPC4. Conversely, TRPC1 co-immunoprecipitated with TRPC3. Expression of antisense TRPC1 decreased (i) Tg- and OAG-stimulated currents and Ca2+ entry and (ii) the level of endogenous TRPC1 but not TRPC4. Antisense TRPC3 similarly reduced Ca2+ entry and endogenous TRPC3. Yeast two-hybrid analysis revealed an interaction between NTRPC1 and NTRPC3 (CTRPC1-CTRPC3, CTRPC3-CTRPC1, or CTRPC1-NTRPC3 did not interact), which was confirmed by glutathione S-transferase (GST) pull-down assays (GST-NTRPC3 pulled down TRPC1 and vice versa). Expression of NTRPC1 or NTRPC3 induced similar dominant suppression of Tg- and OAG-stimulated Ca2+ entry. NTRPC3 did not alter surface expression of TRPC1 or TRPC3 but disrupted TRPC1-TRPC3 association. In aggregate, our data demonstrate that TRPC1 and TRPC3 co-assemble, via N-terminal interactions, to form a heteromeric store-operated non-selective cation channel in HSY cells. Thus selective association between TRPCs generate distinct store-operated channels. Diversity of store-operated channels might be related to the physiology of the different cell types.

Mammalian canonical TRP 1 channels (TRPCs) have been proposed as Ca 2ϩ -permeable cation channels that are activated in response to stimulation of phosphatidylinositol 4,5-bisphosphate hydrolysis and internal Ca 2ϩ release (1,2). Some TRPC species are capable of forming highly Ca 2ϩ -selective channels, while others form relatively nonselective cation channels that allow permeation of both mono-and divalent cations. In addition, the presently available data suggest that TRPCs can form two types of cation entry channels (3). Store-operated channels (SOCs) that are activated in response to depletion of internal Ca 2ϩ stores. These channels can be activated by agonists, thapsigargin, and by intracellular increase in phosphatidylinositol 1,4,5-trisphosphate. In contrast receptor-operated TRPCs are activated by agonist-stimulated phosphatidylinositol 4,5bisphosphate hydrolysis but not by thapsigargin. The latter TRPCs are also activated by OAG, an analogue of diacylglycerol, the proposed physiological ligand for these channels. TRPC3, TRPC6, and TRPC7 have been reported to form receptor-operated Ca 2ϩ entry channels that can be activated by OAG (4), while TRPC1, TRPC4, and TRPC5 form components of store-operated Ca 2ϩ channels in some cell types (5)(6)(7). The exact mechanisms regulating store-operated and receptor-operated Ca 2ϩ channels or their molecular components have yet been conclusively identified.
It has been proposed that the seven TRPC proteins (TRPC1-7) can assemble to form homomers or heteromers (8). Heteromeric interactions between TRPCs can potentially generate a wide variety of different channels and there are convincing data to substantiate these suggestions. Heterologous expression of different combinations of TRPC channels generate distinct channels (9,10). Furthermore, fluorescence resonance energy transfer measurements have confirmed interactions between heterologously expressed TRPCs (11). Immunoprecipitation experiments have demonstrated association of both endogenous and exogenously expressed TRPCs (9,10,12,13). In aggregate, these previous studies suggest that heteromeric interactions occur between members of two groups of TRPCs: TRPC1/TRPC4/TRPC5 and TRPC3/TRPC6/TRPC7. Interestingly, these two groups also represent receptor-operated and store-operated channels, respectively. The latter group is also suggested to be activated by OAG (4). Exceptions to these findings have also been reported. Montell and co-workers (14) reported that exogenously expressed TRPC1 and TPRC3 coassemble to form a heteromeric complex. Novel heteromeric associations between endogenous TRPCs were reported by * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. Strubing et al. (10) in embryonic brain. Lintschinger et al. (15) reported that exogenously expressed TRPC1 and TRPC3 share the property of activation by diacylglycerols and co-assemble to form channels regulated by phospholipase C and Ca 2ϩ but not by Tg. However, there is little information regarding the contribution of endogenous TRPCs, and specifically of heteromeric TRPC channels, to store-operated Ca 2ϩ entry seen in different cell types. Furthermore, the physiological implication of such diverse SOCs is not yet understood. Downstream events regulated by Ca 2ϩ entry likely determine the type of channel that is formed in different cells (1,2,16).
We have reported earlier that the salivary gland cells lines HSG (human submandibular gland ductal cells) and HSY (human parotid gland ductal cells) display distinct carbachol-stimulated Ca 2ϩ oscillations (17). Furthermore, internal Ca 2ϩ store depletion activates a non-selective cation current in HSY cells, which is distinct from the relatively Ca 2ϩ -selective cation current (I SOC ) activated in HSG cells. pCa/pNa was 40 and 4.6 for HSG and HSY cells, respectively. Evidence for anomalous mole fraction behavior of Ca 2ϩ /Na ϩ permeation was obtained with HSG cells but not HSY cells (18). Importantly, our studies have suggested that TRPC1, but not TRPC3, is an integral component of the store-operated channel in HSG cells (7,19). In this study we have analyzed the molecular components of the nonselective, store-operated channel in HSY cells and present data to demonstrate that it is formed by selective heteromeric interactions between endogenously expressed TRPC1 and TRPC3.

EXPERIMENTAL PROCEDURES
Cell Culture and Transfection-HSY cells were cultured in Earle's minimal essential medium supplemented with 10% fetal calf serum, 2 mM glutamine, 1% penicillin/streptomycin at 37°C in 5% CO 2 . For transfection, 1 g of DNA was combined with 2.4 g of Lipofectamine 2000 (Invitrogen) in serum-free medium and incubated at room temperature for 20 min. The mixture was then added to the cells. 5 h after the start of transfection, the medium was replaced with fresh medium containing serum and antibiotics. For functional studies, cells were plated onto glass coverslips or MatTek tissue culture dishes. Cells were transiently transfected with 5 g of plasmids encoding antisense TRPC1, antisense TRPC3, NTRPC1, or NTRPC3 and 1 g GFP-encoding plasmid. GFP-positive cells were selected for the experiment 24 h after transfection.
Patch clamp experiments were performed in the tight-seal whole cell configuration at room temperature (22-25°C) using an Axopatch 200B amplifier (Axon Instruments, Inc.). Development of the current was assessed by measuring the current amplitudes at a potential of Ϫ80 mV, taken from high resolution currents in response to voltage ramps ranging from Ϫ90 to 90 mV over a period of 1 s imposed every 4 s (holding potential was 0 mV) and digitized at a rate of 1 KHz. Liquidjunction potentials were less than 8 mV and were not corrected. Capacitative currents and series resistance were determined and minimized. For analysis, current recorded during the first ramp was used for leak subtraction of the subsequent current records. Tg and OAG were purchased from Calbiochem.
[Ca 2ϩ ] i Measurements-Fura2 fluorescence was measured in single HSY cells cultured for 24 h in glass bottom Matek tissue culture dishes. Cells were loaded with Fura2 by incubation with 1 M Fura2-AM at 37°C for 1 h. Fluorescence was recorded using a Till Photonics-Polychrome 1V spectrofluorimeter and Metafluor imaging software (Universal Imaging Corp.) with excitation at 340 and 380 nm and emission at 510 nm. Cells were bathed in external medium (similar to that used for electrophysiology experiments), and where required Ca 2ϩ was omitted from the assay medium (Ca 2ϩ -free medium). All conditions and additions are indicated in the text. All experiments were repeated using at least three different passages of cells. Student's t test was used to statistically evaluate the data.
Cloning, Expression, and Purification of GST Constructs-The N and C terminus of TRPC1 and TRPC3 were cloned into pGEX5.1 (Amersham Biosciences) vector using PCR-based strategy and expressed as GST fusion proteins. 500 ml of Escherichia coli (BL-21) expressing various GST constructs (log phase culture, OD 0.5-0.6) were induced with 0.2 mM isopropyl ␤-D-thiogalactopyranoside for 3 more h, and GST fusion proteins were purified as described (20). Briefly cells were lysed using a French press at 1200 p.s.i. Cell debris was removed by centrifugation at 10,000 ϫ g for 10 min. The supernatant was loaded on a glutathione S-transferase-agarose affinity column, washed, and eluted using 10 mM reduced glutathione.
Cell Surface Biotinylation-Control and NTRPC3-expressing cells were washed with ice-cold phosphate-buffered saline and incubated for 20 min with 0.5 mg/ml Sulfo-NHS-Biotin (Pierce) on ice. Cells were then washed with buffer containing 0.1 M glycine and solubilized with 2 ml of immunoprecipitation buffer. Biotinylated proteins were pulled down with NeutrAvidin-linked beads (Pierce). Bound fraction was washed and released with SDS-PAGE sample buffer and analyzed by Western blotting using anti-TRPC3 or anti-TRPC1 antibodies as described above.

Thapsigargin and OAG Stimulate Similar Currents in HSY
Cells-In whole cell patch clamp experiments perfusion of HSY cells with 1 M Tg in the external medium increased membrane currents (Fig. 1A). The Tg-induced cation currents inactivated relatively fast. Consistent with our previous report (18) the I-V relationship was linear with a reversal potential of ϩ5 mV. Perfusion of cells with OAG also generated a current that had similar characteristics (pattern of I-V curve and E rev , see Fig.  1B) as that induced by Tg. Furthermore, when cells were perfused with OAG ϩ Tg-containing medium the current generated was similar in characteristics and magnitude to that seen with Tg alone (Fig. 1C). Maximum current densities (picoampere/picofarad) measured in the three conditions were Tg, 6.3 Ϯ 0.7 (n ϭ 5); OAG, 4.5 Ϯ 0.5 (n ϭ 4); OAG ϩ Tg, 6.4 Ϯ 0.9 (n ϭ 6). Thus, the effects of OAG and Tg on the current were not additive. In addition, both Tg-and OAG-stimulated currents were completely blocked by 1 M Gd 2ϩ (Fig. 1, D and E). The patch clamp data were substantiated with Fura2 measurements. 1 M Tg rapidly increased intracellular Ca 2ϩ concentration in cells incubated in Ca 2ϩ -free external medium and addition of 1 mM Ca 2ϩ to the bath induced a rapid significant increase in [Ca 2ϩ ] i (Fig. 1F, note that basal Ca 2ϩ entry is minimal in these cells). Peak Fura2 fluorescence (340/380 ratio) was 1.2 Ϯ 0.3 (n ϭ 118). OAG alone did not induce increase in [Ca 2ϩ ] i until external Ca 2ϩ was replenished (Fig. 1G). Peak 340/380 fluorescence ratio was 1.0 Ϯ 0.3 (n ϭ 121). These data demonstrate that Tg induces internal Ca 2ϩ release and Ca 2ϩ entry, while OAG induces only Ca 2ϩ entry. This is consistent with previous reports showing that OAG directly activates some TRPC channels such as TRPC3, -6, and -7 (4, 22-24). Importantly, and consistent with the current measurements, addition of Tg ϩ OAG induced internal Ca 2ϩ release and Ca 2ϩ entry that were similar to that induced by Tg alone (Fig. 1H). Peak 340/380 fluorescence ratio was 1.2 Ϯ 0.4 (n ϭ 135). In aggregate, the data in Fig. 1 suggest that OAG and Tg stimulate the same Ca 2ϩ entry mechanism in HSY cells.
Expression and Heteromeric Interaction between Endogenous TRPC1 and TRPC3 in HSY Cells-To identify the channels activated by Tg and OAG, we examined the endogenous expression of TRPC proteins in HSY cells. TRPC1, TRPC3, and TRPC4 were detected by Western blotting, while TRPC5 and TRPC6 were not detected ( Fig. 2A, proteins detected using the respective antibodies are indicated below each blot; TRPC2, which is a pseudogene in human cells, and TRPC7 were not examined). Lane 1 in each blot was loaded with a sample containing crude membrane preparations from cells expressing exogenous TRPCs, as indicated (the expressed proteins were further confirmed using the anti-epitope tag antibodies; data not shown). Lane 2 was loaded with a sample containing crude membrane preparation from HSY cells and demonstrates the presence of endogenous TRPCs. Thus, HSY cells appear to express three different TRPC proteins, TRPC1, TRPC3, and TRPC4. Although distinct store-operated currents have been detected in different cell types, there is little information regarding the molecular composition of the endogenous channels mediating these currents or the role of endogenous TRPC proteins in the generation of these channels.
Both exogenously expressed and endogenous TRPCs have been shown to form homo-and heteromultimers. Furthermore, heterologous expression of TRPCs has demonstrated that different combinations of TRPCs generate channels with distinct characteristics. These heteromeric TRPC complexes have been shown to form both store-operated and store-independent cation channels (9,10,15,19,20,(23)(24)(25). Thus we examined possible heteromerization between the endogenously present TRPCs in HSY cells. When anti-TRPC3 was used to pull down TRPC3, TRPC1 was co-immunoprecipitated (Fig. 2B). Reverse IP using anti-TRPC1 pulled down TRPC1 and TRPC3. Furthermore, IP of TRPC1 did not pull down TRPC4 (Fig. 2B). These data suggest a selective interaction between TRPC1 and TRPC3 in HSY cells.
Antisense TRPC1 and Antisense TRPC3 Decrease Store-operated Ca 2ϩ Entry in HSY Cells-To examine the contribution of TRPC1 and TRPC3 to the OAG and Tg-stimulated currents described above, cells were transiently transfected with TRPC1-as (C1-as) or TRPC3-as (C3-as), and expression of TRPC1 and TRPC3 was assessed. TRPC3-as induced a decrease in endogenous TRPC3 but did not affect expression of TRPC1 or TRPC4 (Fig. 2C). Similarly, TRPC1-as induced a  Fig. 1 for details) and I-V plots are shown in each case. Average data comparing the current densities with that in control cells (see Fig. 1) are shown in C. TRPC1-as and TRPC3-as induce a similar decrease in Tg-, OAG-, and Tg ϩ OAGstimulated currents. Values marked are significantly different (p Ͻ 0.05; n Ն 7) from unmarked values in the same group but not from each other. decrease in TRPC1 but not TRPC3 or TRPC4. These data demonstrate the specificity of the antisense treatments.
Consistent with the decrease in TRPC1 expression, TRPC1-as induced a significant reduction in both TG and OAG stimulated membrane currents (Fig. 3, A and C). The current seen with OAG, Tg, or OAG ϩ Tg were all similarly decreased. Importantly, the remaining currents displayed similar I-V relationships as seen in control HSY cells. Similar effects on Tgand OAG-stimulated currents were seen with TRPC3-astreated cells (Fig. 3, B and C). The average current densities, relative to that in control cells, are shown in Fig. 3C. Thus, antisense-induced decrease in specific TRPC proteins does not alter channel characteristics indicating that both TRPC1 and TRPC3 contribute to the OAG-and Tg-stimulated currents. This also suggests that the remaining TRPCs do not reassemble to generate channels with different compositions.
The effect of antisense treatment on the store-operated current was further confirmed by Fura2 measurements. Expression of TRPC1-as or TRPC3-as induced similar decreases in Tg-, OAG-, and Tg ϩ OAG-stimulated Ca 2ϩ entry (Fig. 4, B and  C, respectively). Fig. 4A shows the responses in control cells which were transfected with vector ϩ GFP. Thus, both TRPC1 and TRPC3 significantly contribute to Tg and OAG-stimulated Ca 2ϩ entry. Together with the data showing their association (Fig. 2), these data strongly suggest that both TRPC1 and TRPC3 co-assemble to generate a heteromeric store-operated Ca 2ϩ -permeable cation channel in HSY cells.
TRPC1-TRPC3 Interact via Their N-terminal Domains-To further examine the heteromerization of TRPC1-TRPC3, yeast two-hybrid analysis was used to map out the possible regions involved in their interaction. An interaction was detected between NTRPC1 and NTRPC3 (Fig. 5A). CTRPC3-NTRPC1, NTRPC3-CTRPC1, and CTRPC1-CTRPC3 did not interact. Furthermore, this interaction was mapped to amino acids 1-73 in TRPC1 (see Supplemental Fig. 1). NTRPC1-NTRPC3 interaction was confirmed using GST fusion protein pull-down assays (Fig. 5B). GST-NTRPC3, but not GST-CTRPC3, pulled down full-length HA-TRPC1 from lysates of HEK 293 cells stably expressing HA-TRPC1 (no proteins were detected in samples from control non-transfected cells, first panel). Conversely, GST-NTRPC1 and GST-NTRPC3, but not GST-CTRPC1 or GST-CTRPC3, pulled down HA-TRPC3 from lysates of HEK 293 cells stably expressing HA-TRPC3. These data demonstrate that (i) TRPC3 monomers can interact to form homomultimers, and (ii) that TRPC3 and TRPC1 can interact to form TRPC1-TRPC3 heteromultimers. Additionally, we found that the NTRPC1-TRPC3 and NTRPC3-TRPC1 interactions were relatively stable and did not dissociate even in the presence of 0.5% SDS (data not shown). These results are consistent with reports showing that Drosophila TRP and TRPL co-assembly is mediated through their N terminus (14). Homomultimeric TRPC3 has not been assessed in this study, and its role, if any, in HSY cell Ca 2ϩ entry is presently not known.
NTRPC1 and NTRPC3 Suppress Store-operated Channel Activation in HSY Cells-The data discussed above suggest that the N-terminal domains of TRPC1 and TRPC3 are involved in assembly of the TRPC1-TRPC3 heteromeric channel in HSY cells. This was further examined by expressing either NTRPC1 or NTRPC3 in these cells. Significantly, both NTRPC1 and NTRPC3 induced similar dominant supression of Tg-and OAGstimulated currents in these cells. Furthermore, the characteristics of the remaining currents were similar to that seen in control cells (Fig. 5, C and D). Importantly, expression of or NTRPC3 in HSY cells (i) did not alter the expression of TRPC1 or TRPC3 and (ii) disrupted TRPC1-TRPC3 association (Fig.  5E). Immunoprecipitation of TRPC1 from lysates of cells expressing NTRPC3 pulled down less TRPC3 than from lysates of control cells. However, the amount of TRPC1 pulled down by anti-TRPC1 antibody was similar in control and NTRPC3expressing cells, and similar levels of TRPC1 and TRPC3 were detected in the input samples from control and NTRPC3 cells. In addition, Fig. 5F shows that NTRPC3 does not affect the FIG. 5. TRPC1 and TRPC3 interact via N-terminal regions to generate store-operated cation channel in HSY cells. A, yeast two-hybrid interactions between NTRPC1 and NTRPC3 (interaction was confirmed by ␤-galactosidase assays; data not shown). B, GST pull-down assays. GST proteins are indicated on top of the blots. Lysates of control non-transfected cells (panel 1) and cells expressing either HA-TRPC1 (panel 2) or HA-TRPC3 (panel 3) were used (indicated below the blot). IB antibody was anti-HA for all the blots. GST was used as a control (shown in panel 3, first lane). C, Tg-and OAG-activated membrane currents (trace shows currents measured at Ϫ80 mV, see the legend to Fig. 1 for details) and respective I-V relationships in control and NTRPC3-and NTRPC1-expressing cells. D, average current densities obtained in each case (n ϭ at least 6 cells in each experiment). Significant difference from respective controls is indicated by **; marked values are not different from each other. E, effect of NTRPC3 expression on co-IP of TRPC1 and TRPC3. TRPC1 was immunoprecipitated from lysates of control cells (Con) and cells transfected with NTRPC3 (NTRPC3) using anti-TRPC1 antibody (upper panels). Lower panels show input protein. Blots were probed for TRPC3 or TRPC1 as indicated (IB: TRPC3 and IB TRPC1, respectively). F, surface expression of TRPC1 and TRPC3 in control and NTRPC3-expressing cells. Cells were biotinylated as described under "Experimental Procedures." Upper blots show levels of TRPC1 (left) or TRPC3 (right) in the biotinylated fraction pulled down from cell lysates using avidin-linked beads (IP-avidin; IB was done using anti-TRPC1 or anti-TRPC3 antibodies, respectively). Lower blots show the level of proteins in the input (cell lysate). IP, immunoprecipitated; IB, immunoblotted. plasma membrane expression of either endogenous TRPC1 or TRPC3 in HSY cells. The upper blots show the level of the proteins in the biotinylated fraction (pull-down with avidin) obtained from cells expressing NTRPC3 and control cells (blots on the left show TRPC1, and blots on the right show TRPC3; input protein is shown in the lower blots). These data provide strong evidence that the store-operated channel in HSY cells is assembled via N-terminal interactions between TRPC1 and TRPC3. Consistent with these findings, it has been reported previously that expression of NTRPC3 suppresses a Tg-stimulated TRPC3-dependent conductance in human umbilical vein endothelial cells (26).
Above, we have identified the molecular components of a store-operated, non-selective cation channel in HSY cells. We show that this channel is activated by thapsigargin and more uniquely also by OAG. The effect of OAG and Tg are nonadditive demonstrating that both activate the same channel. Importantly, our data show that endogenous TRPC1 and TRPC3 co-assemble heteromerically to form this store-operated cation channel in HSY cells and that channel assembly involves interaction between their N-terminal domains. This channel is distinct from the TRPC1-dependent SOCs in HSG cells (7,19,20) and Ca 2ϩ release-activated channel in RBL cells. The molecular components of the latter, relatively well studied channel, are presently unknown (27,28). Notably, assembly of TRPC1 and TRPC3 in HSY cells generates a channel that is relatively non-selective for Ca 2ϩ when compared with the channels in HSG and RBL cells (pCa/pNa are 4, 40, and Ͼ400 in HSY, HSG, and RBL cells, respectively). This channel also does not exhibit anomalous mole fraction behavior of Ca 2ϩ / Na ϩ permeation, unlike the channels in RBL and HSG cells. Typically, TRPC3 has been suggested to interact with TRPC6/ TRPC7 and form non-store-operated Ca 2ϩ entry channels (4). In fact, in our previous studies we have shown that TRPC3 is not involved in SOC activity in HSG cells (7). However, nontypical interactions of TRPC3 with TRPC5 and TRPC1 have been reported in embryonic brain (10), and store-dependent regulation of heterologously expressed TRPC3 has also been reported (4). It is important to note that the characteristics of the current we have measured in HSY cells are distinct from that generated due to expression of TRPC1 ϩ TRPC3 ϩ TRPC5 in this previous study (10). Consistent with this we show that TRPC5 is not expressed in HSY cells (see Fig. 2). Furthermore, TRPC1, TRPC3, and TRPC4 are expressed in HSY cells, but only TRPC1 and TRPC3 are co-immunoprecipitated. Thus, we can suggest that TRPC1 and TRPC3 selectively interact to form a heteromultimeric store-operated cation channels in HSY cells, although the involvement of TRPC7 cannot be presently excluded. Association of endogenous TRPC1-TRPC3 in Ca 2ϩ entry has been suggested in two other studies. Non-storeoperated TRPC1 ϩ TRPC3 channels, sensitive to both agonist and OAG, were described by Sydorenko et al. (29). In a recent paper Wu et al. (30) have shown that store-operated calcium entry is up-regulated in differentiating H19-7 cells and can be correlated with expression of TRPC1 and TRPC3.
In conclusion, we report here the presence of a Tg-and OAG-sensitive, store-operated non-selective cation channel in HSY cells. We demonstrate that TRPC1 and TRPC3 co-assemble, via N-terminal domain interactions, to form this novel channel. Thus, selective assembly of endogenous TRPC proteins in cells can generate distinct cation channels that are activated by agonist stimulation of the cells, internal Ca 2ϩ store-depletion, or both. Further studies are required to understand the physiological relevance of such distinct store-operated calcium channels in different cell types.