Selective Gαi Subunits as Novel Direct Activators of Transient Receptor Potential Canonical (TRPC)4 and TRPC5 Channels*

Background: Activation of TRPC4/5 channels is mediated by GPCR activation. Results: TRPC4/5 was activated by the Gαi/o-coupled receptor and the Gαi protein, which interacted directly with each other. Conclusion: Gαi proteins play an essential role as novel activators of TRPC4/5. Significance: Our findings provide new insights into the activation mechanism of inhibitory Gα proteins. The ubiquitous transient receptor potential canonical (TRPC) channels function as non-selective, Ca2+-permeable channels and mediate numerous cellular functions. It is commonly assumed that TRPC channels are activated by stimulation of Gαq-PLC-coupled receptors. However, whether the Gαq-PLC pathway is the main regulator of TRPC4/5 channels and how other Gα proteins may regulate these channels are poorly understood. We previously reported that TRPC4/TRPC5 can be activated by Gαi. In the current work, we found that Gαi subunits, rather than Gαq, are the primary and direct activators of TRPC4 and TRPC5. We report a novel molecular mechanism in which TRPC4 is activated by several Gαi subunits, most prominently by Gαi2, and TRPC5 is activated primarily by Gαi3. Activation of Gαi by the muscarinic M2 receptors or expression of the constitutively active Gαi mutants equally and fully activates the channels. Moreover, both TRPC4 and TRPC5 are activated by direct interaction of their conserved C-terminal SESTD (SEC14-like and spectrin-type domains) with the Gαi subunits. Two amino acids (lysine 715 and arginine 716) of the TRPC4 C terminus were identified by structural modeling as mediating the interaction with Gαi2. These findings indicate an essential role of Gαi proteins as novel activators for TRPC4/5 and reveal the molecular mechanism by which G-proteins activate the channels.

The ubiquitous transient receptor potential canonical (TRPC) channels function as non-selective, Ca 2؉ -permeable channels and mediate numerous cellular functions. It is commonly assumed that TRPC channels are activated by stimulation of G␣ q -PLC-coupled receptors. However, whether the G␣ q -PLC pathway is the main regulator of TRPC4/5 channels and how other G␣ proteins may regulate these channels are poorly understood. We previously reported that TRPC4/TRPC5 can be activated by G␣ i . In the current work, we found that G␣ i subunits, rather than G␣ q , are the primary and direct activators of TRPC4 and TRPC5. We report a novel molecular mechanism in which TRPC4 is activated by several G␣ i subunits, most prominently by G␣ i2 , and TRPC5 is activated primarily by G␣ i3 . Activation of G␣ i by the muscarinic M2 receptors or expression of the constitutively active G␣ i mutants equally and fully activates the channels. Moreover, both TRPC4 and TRPC5 are activated by direct interaction of their conserved C-terminal SESTD (SEC14-like and spectrin-type domains) with the G␣ i subunits. Two amino acids (lysine 715 and arginine 716) of the TRPC4 C terminus were identified by structural modeling as mediating the interaction with G␣ i2 . These findings indicate an essential role of G␣ i proteins as novel activators for TRPC4/5 and reveal the molecular mechanism by which G-proteins activate the channels.
The physiological role of these channels was established recently, demonstrating that TRPC4 and TRPC6 are the molecular candidates for the non-selective cation channels activated by muscarinic receptor stimulation (mI CAT ) in visceral smooth muscle cells. mI CAT mediates the physiological action of acetylcholine in evoking smooth muscle contraction (19). Activation of muscarinic receptors causes the opening of non-selective cationic channels in smooth muscle cells of the gastrointestinal tract (20,21). In these cells, PTX-sensitive G-proteins but not G␤␥ were suggested to mediate channel activation (22,23). At the level of M2 or M3 muscarinic receptors, Sakamoto et al. (24) showed three distinct signaling pathways that activate cationic channels in murine gut smooth muscle cells. The three pathways include the M2, M3, and M2/M3 pathways and were demonstrated using M2 KO, M3 KO, and M2/M3 double KO mice, respectively. In addition, the M2/M3 pathway but not the M2 or M3 pathways involves processes in which Ca 2ϩ has a potentiating effect on channel activation, suggesting that the M3 pathway may facilitate the function of the M2/M3 pathway through inositol 1,4,5-trisphosphate-induced Ca 2ϩ release (24). Similarly, activation of mI CAT requires the simultaneous activation of both the M2 and M3 muscarinic receptors (20), further suggesting involvement of G␣ i in channel activation.
Several studies have also suggested that PTX-sensitive G-proteins play an important role in the activation process of TRPC4 and TRPC5 by GPCR (6,8,25). In our previous study, we showed that G␣ i2 activates TRPC4␤ and that PTX inhibits the activation of TRPC4␤ by stimulation of M2 muscarinic receptors (25). However, the specificity for G␣ i subunits and how the PTX-sensitive G-proteins activate the channels are not yet known; it is assumed that the PTX-sensitive G-proteins activate the channels by an indirect mechanism that involves the generation of second messengers. The well known second messenger of PTX-sensitive G-proteins is cyclic AMP (cAMP), found downstream of adenylate cyclase. PTX-sensitive G-proteins inhibit adenylate cyclase, which in turn decreases cAMP concentration. In our recent study, we showed that cAMP inhibits TRPC4 and TRPC5 currents by activating PKA and phosphorylating TRPC4 and TRPC5 channels (26). Thus, it cannot be the mechanism by which G␣ i activates the channels.
These findings prompted us to ask whether TRPC4 and TRPC5 are activated by other PTX-sensitive G␣ i/o subunits and whether the activation is direct. We were also interested in identifying the TRPC4/5 domain that mediates the interaction with the channels and the activation by G␣ i/o subunits, as well as the roles G␣ q plays in modulating TRPC4/5. In the present study, we focused on the role of G␣ i proteins in regulating TRPC4/5 and report that G␣ i subunits specifically activate TRPC4 and TRPC5 by direct interaction with the channels. Moreover, the regulation is specific to G␣ i subunits. TRPC4 is mainly activated by G␣ i2 , whereas TRPC5 is primarily activated by G␣ i3 . These findings explain how TRPC4 is activated to regulate GI motility. Strategies can now be developed to understand the functional consequences of activation of TRPC4/5 in the central and peripheral nervous systems.

EXPERIMENTAL PROCEDURES
Cell Culture and Transient Transfection, cDNA Clones-Human embryonic kidney (HEK293) cells (ATCC, Manassas, VA) were maintained according to the supplier's recommendations. For transient transfection, cells were seeded in 12-well plates. The following day, 0.5 g/well of pcDNA3 vector containing the cDNA for mouse TRPC4␤ was mixed with 50 -100 ng/well of pEGFP-N1 (Clontech) and transfected using the transfection reagent FuGENE 6 (Roche Molecular Biochemicals), as detailed in the manufacturer's protocol. Human TRPC5-EGFP cDNA and mouse TRPC4␤-EGFP cDNA were also transfected in the same way. Coexpression of TRPC channels with G-proteins or receptors was achieved through a channel to G-protein transfection ratio of 1:1. After 30 -40 h, the cells were trypsinized and transferred to a small recording chamber (RC-11, Warner Instruments) for whole-cell recording. HEK293 cells stably expressing mouse TRPC4␤ were established by G418 selection. The cells were cultured as for the transient transfection, except that the medium was supplemented with G418 (400 g/ml). Human G␣ i1 Q204L , G␣ i2 Q205L , G␣ i3 Q204L , rat G␣ i2 Q205L , and human G␣ q Q209L were cloned into pcDNA3.1ϩ. Human G␤ 1, G␤ 2 , G␤ 3 , G␤ 5 , and bovine G␥ 2 were cloned into pcDNA3.1ϩ (Invitrogen). Human G␣ oA Q205L , M2 receptor (the Missouri S&T cDNA Resource Center), and M3 receptor were cloned into pcDNA3.1ϩ. Human G␤ 1 was used to insert the G␤ 1 W99A and G␤ 1 I80A mutations using the QuikChange site-directed mutagenesis kit (Stratagene).
Western Blotting and Co-immunoprecipitation-Transfected cells were collected and lysed using 300 l of binding buffer (50 mM HEPES, pH 7.4, 120 mM NaCl, 2 mM EDTA, 2 mM MgCl 2 , complete protease inhibitor mixture tablet, phosphatase inhibitor mixture tablet (Roche Applied Science), and 0.5% Triton X-100). The lysates were sonicated, and any insoluble material was removed by centrifugation at 13,300 ϫ g for 10 min. For co-immunoprecipitation of TRPC4␤-GFP and TRPC5-GFP with G␣ i2 and G␣ i3 , anti-GFP antibody (1 g, Invitrogen, A11122) was added to 100 l cell extract and incubated for 12 h at 4°C. Then, 50 l of a 1:1 slurry of protein G-Sepharose 4B beads was added to the antibody-extract mix and incubated for 12 h at 4°C. Beads were washed three times with binding buffer; proteins were released from the beads with 50 l of 2ϫ SDS-loading buffer and analyzed with 10% or 8% SDS-PAGE. G␣ i2 and G␣ i3 were co-precipitated with GFP antibody and probed by mouse monoclonal anti-G␣ i2 antibody (2 g, Santa Cruz Biotechnology, sc-13534) and mouse monoclonal anti-EE antibody for G␣ i3 (2 g, Covance, MMS-115P). The mouse monoclonal anti-G␣ i2 antibody was used for a reciprocal co-immunoprecipitation with GFP antibody in a sequential experiment. Co-immunoprecipitation of TRPC4␤-GFP with G␣ q was achieved using the same procedures. G␣ q was probed by mouse monoclonal anti-G␣ q antibody (Santa Cruz Biotechnology, sc-136181). Rat brain from day 15 was homogenized on ice using a Dounce homogenizer. The homogenate buffer had the same composition as the binding buffer. Homogenates were centrifuged at 13,000 rpm for 30 min at 4°C. Supernatants were recentrifuged at 13,000 rpm at 4°C for 20 min. Supernatants were pre-cleared with protein G-Sepharose beads for 1 h at 4°C and centrifuged at 2000 rpm for 2 min at 4°C. Fifty microliters of a 1:1 slurry of protein G-Sepharose beads was added to the rabbit polyclonal antibody (anti-G␣ i2 , Santa Cruz Biotechnology, sc-7276 and anti-G␣ i3 , Santa Cruz Biotechnology, sc-262) extract mix and incubated for 12 h at 4°C. The omission of primary antibody was used as a control. Beads were washed three times with binding buffer. Immunoprecipitated proteins were probed with anti-TRPC4 (NeuroMab, 75-119) and anti-TRPC5 (NeuroMab, 75-104) on an 8% SDS-PAGE gel. Anti-TRPC4 and Anti-TRPC5 were used for reciprocal pulldowns. Controls were omission of the primary antibody and substitution of normal anti-mouse IgG with non-immune serum (Santa Cruz Biotechnology, sc-2025). Rabbit polyclonal anti-G␣ i2 (2 g, Santa Cruz Biotechnology) and rabbit polyclonal anti-G␣ i3 (2 g, Santa Cruz Biotechnology) antibodies were used to probe co-immunoprecipitation samples on a 10% SDS-PAGE gel.
Surface Biotinylation-Cells were washed with and suspended in PBS. Suspended cells were incubated in 0.5 mg/ml sulfo-NHS-LC-biotin (Pierce) in PBS for 30 min on ice. Free biotin was quenched by the addition of 100 mM glycine in PBS. Lysates were prepared in lysis buffer by being passed 7-10 times through a 26-gauge needle after sonication. Lysates were centrifuged at 13,300 ϫ g for 10 min at 4°C, and protein concentrations of the supernatants were determined. Forty microliters of a 50% slurry of avidin beads (Pierce) was added to cell lysates equivalent to 400 g of protein. After incubation for 1 h at room temperature, beads were washed three times with 0.5% Triton X-100 in PBS, and proteins were extracted in sample buffer. Collected proteins were then analyzed by 8% SDS-PAGE gel and probed by anti-GFP antibody (Invitrogen, A11122).
GST Pulldown Assays-The C-terminal domain of TRPC4-(621-890) was cloned into BamHI-SalI restriction sites of pGEX4T-1 (Amersham Biosciences Pharmacia) by PCR. The GST fusion constructs were expressed and purified from Escherichia coli (BL21(DE3)). Briefly, E. coli were grown in liquid cultures containing 0.1 mM isopropyl 1-thio-␤-D-galactopyranoside with vigorous agitation for 18 h at 20°C to an A 600 of 0.6. Channel protein was purified from the soluble extract using glutathione-agarose beads (Amersham Biosciences). GST fusion proteins appeared to be sensitive to degradation and carefully utilized in subsequent binding assays conducted within 24 h following purification. Histidine-tagged G␣ i2 Q205L protein was expressed in E. coli (BL21(DE3)) from the pET15b plasmid containing full-length human G␣ i2 Q205L cDNA. Histidine-tagged forms of G␣ i2 Q205L were purified using immobilized Ni 2ϩ -nitrilotriacetic acid affinity chromatography. Binding between G␣ i2 Q205L and the GST fusion C-terminal domain of TRPC4 was allowed to occur for 1 h at room temperature on a plate rotator. Each reaction sample was subsequently centrifuged at 500 ϫ g for 5 min. After three washes with 500 l of PBS with 0.1% Triton X-100, the GST protein-G-protein complexes were eluted with 15 l of 2ϫ SDS sample buffer, and the entire sample was run on a 10% polyacrylamide-SDS gel. Mouse monoclonal anti-G␣ i2 (Santa Cruz Biotechnology) and anti-GST antibodies (Santa Cruz Biotechnology, sc-138) were used for immunoblot analyses. Unless otherwise stated, all pulldown assays were repeated three times for each condition.
Whole-cell Patch Clamp Experiment-The whole-cell configuration was used to measure TRPC channel current in HEK cells as described previously (13,19,23,27). Cells were transferred to a small chamber on the stage of an inverted microscope (TE2000S, Nikon Japan), and attached to coverslips in the small chamber for 10 min prior to patch recording. Currents were recorded using an Axopatch 200B patch clamp amplifier (Axon Instrument). Bath solutions were constantly perfused with a physiological salt solution at a rate of 1-2 ml/min. Glass microelectrodes with 2-4 megohm resistance were used to obtain gigaohm seals. After establishing the whole-cell configuration, the external solution was changed from Normal Tyrode to Cs ϩ -rich external solution. The current was recorded in 500-ms duration RAMPs from ϩ100 to Ϫ100 mV and from a holding membrane potential of Ϫ60 mV. pCLAMP software (version 10.2) and Digidata 1440A (Axon Instruments) were used for data acquisition and application of command pulses. Data were filtered at 5 kHz and displayed on a computer monitor. Data were analyzed using pCLAMP (version 10.2) and Origin software (Microcal origin, version 7.5).
Solutions and Drugs-For all TRPC channel recordings, physiological salt solution containing 135 mM NaCl, 5 mM KCl, 2 mM CaCl 2 , 1 mM MgCl 2 , 10 mM glucose, and 10 mM HEPES. The pH was adjusted to 7.4 using NaOH. Cs ϩ -rich external solution was prepared by replacing NaCl and KCl with equimolar CsCl. The pipette solution contained 140 mM CsCl, 10 mM HEPES, 0.2 mM Tris-GTP, 0.5 mM EGTA, and 3 mM Mg-ATP. The pH was adjusted to 7.3 with CsOH. Pertussis toxin was purchased from Calbiochem (La Jolla, CA), and carbachol, HEPES, and GTP␥S were purchased from Sigma.
Modeling of Interaction Between G␣ i2 and TRPC4-The homology model of G␣ i2 structure was constructed using MODELLER 9v8 (28) based on two G␣ structures that have high sequence homology with G␣ i2 (Protein Data Bank codes 2ODE and 1GP2). Due to the absence of a known protein structure homologous to TRPC4, the structure of TRPC4 C terminus was modeled using I-TASSER, which combines the methods of threading, ab initio modeling, and structural refinement (29,30). The docking of the TRPC4 C terminus on the G␣ i2 was initiated because the putative binding site of TRPC4 (amino acids 701-720) contains positive charges conserved among TRPC channels. An ionic interaction was assumed between the sites of the mTRPC4 and G␣ i2 . There are nine sites where the G␣ i family has negative charges, but the G␣ q/11 family does not. Seven crystal structures of protein complexes between G␣ and other proteins (Protein Data Bank codes 2RGN, 2ODE, 3CX6, 2G83, 1FQJ, 2BCJ, and 1GP2) were inspected visually, and it was found that only specific areas of G␣ participated in interactions with other proteins. Therefore, manual docking was carried out between G␣ i2 and the TRPC4 C terminus, emphasizing the close interaction between the two conserved positive charges of TRPC4, Arg-711 and Lys-715, and the conserved positive charges of G␣ i2 , Asp-252.
Statistics-All data are expressed as means Ϯ S.E. Statistical significance was determined using paired or unpaired Student's t tests. p values of Ͻ 0.05 were considered statistically significant. The number of cell recordings is represented by n.

RESULTS
Expression of TRPC4 alone results in a minimal spontaneous current (2.1 Ϯ 1.1 pA/pF; see Fig. 2B, mock, open column, n ϭ 5), compared with TRPC5, which showed a significant basal current (Fig. 2C, 36.6 Ϯ 9.2 pA/pF, n ϭ 14; see also Ref. 26). Although HEK cells endogenously express muscarinic receptors, most likely the M3 subtype (31), the endogenous receptors do not activate TRPC4 (Fig. 1B, mock) (32). In the case of TRPC5, the endogenous muscarinic receptor elicited only small transient TRPC5 activation. Therefore, we analyzed the TRPC4 and TRPC5 currents induced by heterologously expressed muscarinic receptors or by intracellular infusion of GTP␥S through the patch pipette. The TRPC maximal inward currents (in Cs ϩ -rich solution) at negative membrane potentials (Ϫ60 mV) are represented as a current density (pA/pF). In all cases, maximal peak inward currents were obtained by subtracting the current recorded in bath Cs ϩ .
Muscarinic Receptor Stimulation Activates TRPC4 and TRPC5 through Endogenous G␣ i/o -First, we tested the effects of the expressed G␣ i -coupled M2 and the G␣ q -coupled M3 muscarinic receptors on the activation of TRPC4 or TRPC5 by endogenous G-proteins in HEK cells (Fig. 1). Activation of the M2 receptors with carbachol elicited 2-3-fold higher TRPC4 and TRPC5 currents than activation of the M3 receptors. The TRPC4 and TRPC5 currents showed a typical doubly rectifying current-voltage relationship (Fig. 1, A and E). The stimulation of the M2 receptor increased both TRPC4 and TRPC5 currents in a dose-dependent manner (Fig. 1, B and F, open column, 1, 10, and 100 M; n ϭ 5, n ϭ 4, n ϭ 12, and n ϭ 4, n ϭ 4, n ϭ 10), whereas the stimulation of the M3 receptor increased the TRPC4 current dose-dependently and the TRPC5 current dose-independently at the concentration range 1-100 M carbachol (Fig. 1, B and F, closed column, 1, 10, and 100 M; n ϭ 5, n ϭ 8, n ϭ 16, and n ϭ 5, n ϭ 7, n ϭ 11). Treatment with PTX markedly inhibited the TRPC4 and TRPC5 currents activated by M2 receptor stimulation and GTP␥S ( Fig. 1, C, D, G, and H; mock/PTX; n ϭ 8/n ϭ 8, n ϭ 5/n ϭ 5, n ϭ 7/n ϭ 4, and n ϭ 6/n ϭ 3). Thus, the M2 receptor-G␣ i/o pathway was more effective than the M3 receptor-G␣ q pathway in channel activation by engaging the endogenous PTX-sensitive G␣ proteins.
Specific G␣ Isoforms Increase TPRC4 and TRPC5 Activity-To determine which G␣ isoform is involved in the activation of TRPC4 and TRPC5, we used constitutively active forms of the G␣ i subunits (G␣ QL mutants). Intracellular application of GTP␥S through the patch pipette increased the TRPC4 current to 45.0 Ϯ 7.2 pA/pF (Fig. 2B, mock, closed column, n ϭ 13). The TRPC4 channel was activated to a different extent by all constitutively active G␣ i/o subunits, even in the absence of GTP␥S.
Constitutively active G␣ i3 and G␣ o mimicked the activation of TRPC4 by GTP␥S (open/closed column, n ϭ 10/n ϭ 8 and n ϭ 9/n ϭ 11). Constitutively active G␣ i1 activated the TRPC4 channel, whereas application of GTP␥S significantly inhibited the TRPC4 current (open/closed column, n ϭ 13/n ϭ 11). Constitutively active G␣ i2 (G␣ i2 Q205L ) was the most effective activator among the G␣ i subunits tested (Fig. 2B, open columns, and supplemental Fig. S1A, n ϭ 6). Notably, the application of GTP␥S had no further effect on the TRPC4 current, indicating that G␣ i2 fully activates TRPC4 (closed column, n ϭ 8). Furthermore, stimulation of the M2 receptors (Fig. 1) and G␣ i2 (Fig. 2) activated TRPC4 to the same extent. Of particular significance, constitutively active G␣ q was unable to activate TRPC4 (Fig. 2B, open/closed column, n ϭ 4/n ϭ 3). Moreover, constitutively active G␣ q inhibited the stimulatory effect of GTP␥S. This was addressed further below.
G␤␥ Isoforms Are Not Required for TPRC4 and TRPC5 Activation-Although G␣ i/o subunits are key activators of TRPC4 or TRPC5, G␤␥ subunits may also be involved in the activation of the channels, as is the case with GIRK channels, or in the regulation of the channels by altering the availability of activated G␣ i . To address these questions, we tested the effects of various G␤␥ combinations on TRPC4 and TRPC5 activity. Figs. 3, A-C show that none of the G␤␥ combinations tested activated the channels or reduced the activation by GTP␥S. The exception is G␤ 2 ␥ 2 , which slightly inhibited the activation of TRPC5, likely by sequestering some of the G␣ i3 even in the presence of GTP␥S. Moreover, even the free form G␤ 1 I80A mutant (18, 20, 33, 34) did not activate TRPC4 or TRPC5 ( Fig.  3B; open/closed column, n ϭ 4/n ϭ 5, and Fig. 3C; open/closed column; n ϭ 3/n ϭ 14). These results indicate that the PTXsensitive G␣ i2/3 subunits are the activators of TRPC4 and TRPC5.
TRPC4 Is Inhibited by Increased G␣ q Activity-Activation of the G␣ q -PLC pathway has been shown to modestly activate TRPC4 and TRPC5 (Fig. 1) by an unknown mechanism (1, 7, 8). Hence, the inhibition of TRPC4/5 by the constitutively active G␣ q (Fig. 2) was completely unexpected. These results imply that intense overstimulation of a G␣ q -activated pathway inhibits TRPC4 and TRPC5. We considered several potential mechanisms, including increased or decreased cytoplasmic Ca 2ϩ , interference of channel interaction with G␣ i , reduced surface expression of TRPC4, modified cellular PIP 2 , and channel phosphorylation by PKC, known to induce desensitization of TRPC5 with phosphorylation of residue Thr-972 at the C terminus (35).
Loading the cells with BAPTA-AM recovered only 15 Ϯ 5.6% (mock, 56.7 Ϯ 11.2 pA/pF (n ϭ 6); recovery by BAPTA, 9.8 Ϯ 3.1 pA/pF; n ϭ 6, BAPTA with GTP␥S; n ϭ 12, G␣ q QL; n ϭ 3, Fig. 4A), and inhibition of PKC with Goe6976 recovered only FIGURE 2. Effect of G␣ isoforms on TPRC4 and TRPC5 activity. A, representative I-V relationships of TRPC4 and TRPC5 show the effect of constitutively active G␣ QL mutants on the electrophysiological properties of the TRPC4 and TRPC5 channels. B, summary of TRPC4 current density activation by G␣ subunits and/or by GTP␥S. Note the variable effects of the G␣ mutants. All G␣ mutants activate TRPC4 channels without an activator (e.g. GTP␥S or carbachol (CCh)). G␣ i2 provided the most effective activation of TRPC4, whereas G␣ q inhibited TRPC4. Current density is represented by maximal current peaks (subtracted Cs ϩ basal current) at Ϫ60 mV in Cs ϩ solution and is indicated by means Ϯ S.E. Statistical significance was denoted by an asterisk (open column) and number sign (closed column) at p Ͻ 0.05. C, summary of TRPC5 activation by G␣ subtypes showing that the most effective activator is G␣ i3 and that most G␣ mutants inhibit the TRPC5 channel, with G␣ q inhibiting TRPC5. Current density was obtained by the methods described above. Statistical significance was denoted by an asterisk (open column) and number sign (closed column) at p Ͻ 0.05. n. s., not significant.
These findings rule out the effects of changes in cytosolic Ca 2ϩ , PKC, and altered interaction with G␣ i in channel inhibition by G␣ q as the major inhibitors of the current induced by activated G␣ q . Moreover, they point to an indirect effect of G␣ q on channel function.

FIGURE 4. Inhibition of TRPC4 by G␣ q is rescued by PIP 2 . A-C show current densities in HEK293 cells stably expressing mTRPC4␤ and its inhibition by G␣ q .
Loading the cells with BAPTA-AM and inhibition of PKC with Goe6976 did not effectively reverse channel inhibition by G␣ q . High intracellular Ca 2ϩ (5 M) reversed channel inhibition by G␣ q . D, TRPC4 and C-terminal truncation TRPC4 mutants did not co-IP with G␣ q . E, intracellular application of diC8-PIP 2 (50 M) almost recovered the inhibition of TRPC4 by G␣ q . All current densities represent subtracted maximal current peaks at Ϫ60 mV in Cs ϩ solution and are indicated by means Ϯ S.E. Statistical significance was denoted by an asterisk (p Ͻ 0.05). All whole-cell currents were recorded in the intracellular application of GTP␥S. F, surface expression of TRPC4 and G␣ q co-expressed TRPC4 in HEK cells. Surface expression of TRPC4 was not altered by G␣ q QL as determined by co-expression and surface biotinylation. Immunoblots of surface and total were detected by anti-GFP antibody (upper panel). Expression of endogenous G␣ q and transfected G␣ q QL were detected by G␣q antibody (bottom panel). IB, immunoblot. N. B., non buffered. 51.83 Ϯ 12.83% of the PIP 2 control (open/closed column, mock, PIP 2 alone, G␣ q Q209L , and PIP 2 with G␣ q Q209L ; n ϭ 18/26, n ϭ 4/15, n ϭ 3/18, and n ϭ 3/14, Fig. 4E). PIP 2 was reported to inhibit TRPC4␣ but not the TRPC4␤ isoform (8), the isoform used in the present work. In addition, we reported that PIP 2 slows TRPC5 desensitization (7). Thus, the combined results in Fig. 4 indicate that PIP 2 and perhaps increased cytoplasmic Ca 2ϩ are required for TRPC4 and TRPC5 activation.
Mechanisms Associated with Interaction between G␣ i2 with C Terminus of TRPC4-Together, the results in Figs. 1-4 indicate that activation of G␣ i subunits by GPCRs is the primary mechanism for activating TRPC4 and TRPC5. This raised the question of whether activation of the channels requires direct interaction with the G␣ i subunits, as was shown for other channels regulated by G␣ (36) and G␤␥ (33,37) subunits. To address this question, we identified the TRPC4 and TRPC5 domain that might interact with the G␣ i subunits. To characterize the association between TRPC4␤ with G␣ i2 in vivo, HEK cells were transfected with TRPC4␤-GFP and G␣ i2 , and their association was analyzed by co-immunoprecipitation. Immunoprecipitation of G␣ i2 pulled down TRPC4␤-GFP (Fig. 5A, upper panel). Likewise, G␣ i2 was present in TRPC4␤-GFP immunoprecipitates (Fig. 5A, lower panel). Similar co-immunoprecipitation occurred between TRPC5 and G␣ i3 (supplemental Fig. S2A). Pulldown assays were utilized to examine the binding of purified G␣ i2 Q205L to GST fusion protein containing the C-terminal domain of TRPC4. G␣ i2 Q205L bound to the C-terminal domain of TRPC4 (Fig. 5A, right panel).

DISCUSSION
We report here that the primary mechanism for the activation of TRPC4 and TRPC5 in gastric smooth muscle in vivo is through activation of GPCR. TRPC4 and TRPC5 activation appears to require both M2-G i/o and M3-G q/11 muscarinic receptors (20). The mechanism involves specific activation of the channels by G␣ i subunits. Receptors coupled to G␣ i2 primarily activate TRPC4 and receptors coupled to G␣ i3 primarily activate TRPC5 to mediate the receptor-stimulated monovalent cation current and the Ca 2ϩ influx.
Several studies have reported modulation of TRPC4 and TRPC5 activity by GTP␥S-activated PTX-sensitive G-proteins and G i/o -coupled receptors (6,8,25). Activation of TRPC4 by muscarinic receptor stimulation of GI smooth muscle was inhibited by PTX and was shown to be critical in evoking and regulating GI tract motility (21). Our results provide a molecular mechanism to explain the activation of TRPC4 by the PTXsensitive G-protein signaling. A unique feature of the mI CAT is that it requires simultaneous activation of both the M2 and M3 muscarinic receptors. Conversely, cationic channel activation in murine gut smooth muscle cells involves three separate pathways (24). 1) The M3-G␣ q -PLC␤ system, which transiently activates the 70-pS and 120-pS cationic channels concurrently with inositol 1,4,5-trisphosphate-induced Ca 2ϩ release. 2) The M2 pathway, which transduces signals from M2 receptors via G␣ o to the 70-pS cationic channel and shifts the transient activation toward a longer open mode. 3) The M2/M3 pathway, which transmits M2 signals via G␣ o and M3 signals via a G␣ qindependent PLC, to the 70-pS cationic channel, resulting in a much longer open mode. The latter pathway does not work well when either the M2 or the M3 receptors are lacking or when either G␣ o or PLC are inactivated. In addition, the M2/M3 pathway, but not the M2 or M3 pathway, involves processes in which Ca 2ϩ has a potentiating effect on channel activation, suggesting that the M3 pathway may facilitate the function of the M2/M3 pathway through inositol 1,4,5-trisphosphate-induced Ca 2ϩ release (38). We argue that if the 70-pS channel is mediated by TRPC4, then the pathways involving the M2 receptors use G␣ i2 to directly activate the channel.
In addition to TRPC4/5, activation by the G␣ i/o subunits have been reported to regulate several other TRP channels. For example, in G o -coupled mGluR6, G␣ o closes a downstream nonselective cation channel in ON bipolar cells that is mediated by TRPM1-L (39). Pheromone sensing in the vomeronasal organ is mediated by V1R-G i and V2R-G o complexes that activate TRPC2 (40). Whether activation of these channels is by direct interaction with G␣ i/o subunits as shown here for TRPC4/TRPC5 remains to be determined.
Our findings indicate that regulation of TRPC4 and TRPC5 by G-proteins is more complex than assumed previously. The newly discovered mechanism for activation of TRPC4 and TRPC5 suggests that Ca 2ϩ influx through these channels can be activated by several mediators depending on receptor stimulation. As reported before and shown in Fig. 1, TRPC4 and TRPC5 can clearly be activated by G␣ q -coupled receptors and mediated by several of the mechanisms reported before. However, activation by G␣ q -coupled receptors appears to be modest (Fig. 1). More significant activation of TRPC4/5 is elicited by stimulation of G␣ i -coupled receptors that is mediated by direct activation of the channels by G␣ i subunits. The G␣ i binding domain in TRPC4 and TRPC5 shares binding motifs with other regulatory molecules (CIRB, SESTD1) in the C-terminal region of TRPC4 and TRPC5. Modeling and mutation analyses indicate that the 711 RNLVKR 716 region is important for the interaction between G␣ i2 and TRPC4 (supplemental Fig. S6). By contrast with GIRK channels, G␤␥ subunits do not appear to be involved in the direct activation of TRPC4 or TRPC5 by G␣ i . It FIGURE 5. Interaction of G␣ i2 with the C terminus of TRPC4. A, TRPC4 co-immunoprecipitates with G␣ i2 . HEK cells were transfected with empty vector, G␣ i2 alone, and G␣ i2 with TRPC4-GFP and were used to test reciprocal co-IP of TRPC4 and G␣ i2 . TRPC4 C-terminal GST fusion protein was used to probe direct binding between G␣ i2 and TRPC4. Binding of the TRPC4 C-terminal GST fusion protein with recombinant human G␣i2 Q205L protein was shown by G␣ i2 -antibody in vitro binding assay. B, a schematic of GFP-fused TRPC4. C, summary of the effects of G␣ i2 on TRPC4 C-terminal truncation mutants in the presence and absence of GTP␥S stimulation. Current densities are represented by subtracted maximal maximal current peaks at Ϫ60 mV in Cs ϩ solution and are indicated by means Ϯ S.E. D, schematic of GFP-fused TRPC4 deletion and truncation mutants used (upper panel). Wild-type TRPC4 and mutants were probed using the GFP antibody in immunoblotting (bottom). E, summary of the effects of G␣ i2 on current by TRPC4 deletion and truncation mutants in the presence and absence of GTP␥S stimulation. Current density was obtained by the methods described above. F, interaction between G␣ i2 , TRPC4, and mutants was tested by co-IP, and the interaction site was mapped to the 700ϳ728 region (the SESTD1 domain of TRPC4␤). G, a schematic of GFP-fused C-terminal fragments of TRPC4 (upper panel) and their co-IP with G␣ i2 . H, the association between G␣ i with TRPC4/5 in vivo. G␣ i2 and G␣ i3 were immunoprecipitated from rat brain extract and were probed for TRPC4 and TRPC5 to show co-IP in vivo (upper panel). TRPC4 and TRPC5 were co-immunoprecipitated reciprocally. Lanes of IgG and control (Ϫ) did not show G␣ i binding. Input was indicated as 10% input of brain extract. IB, immunoblot. FIGURE 6. The interaction modeling of the TRPC4 C terminus (amino acids 701-720) with G␣ i2 . A, a model of association between G␣ i2 and TRPC4. An ionic interaction was assumed between the two proteins. Amino acids from 671 to 758 of TRPC4 are shown. B, summary of effects of G␣ i2 Q205L on current by mTRPC4 and mTRPC4 mutants in the presence and absence of GTP␥S stimulation. mTRPC4 mutants were substituted at CIRB residues (R711K, R711A, N712R, K715A, and R716A). All current densities represent maximal current peaks (subtracted Cs ϩ basal current) at Ϫ60 mV in Cs ϩ solution and are indicated by means Ϯ S.E. Statistical significance was denoted by an asterisk (p Ͻ 0.05).
is likely that the major role of G␤␥ is regulating channel function by sequestering G␣ i/o subunits in the resting state. The mode of TRPC4/TRPC5 regulation by G␣ i/o subunits reported here likely constitutes a major pathway in smooth muscle. It will be interesting to explore the full potential of this form of regulation in other cell types, particularly when the regulation by G␣ i/o -coupled receptors involves changes in cellular Ca 2ϩ . Examples include acetylcholine-induced vasoregulation, lung microvascular permeability (41,42) and the increase of 5-hydroxytryptamine 2 receptor-coupled GABA release in thalamic interneurons (43). In summary, the present findings expand knowledge regarding signal transduction by cytoplasmic Ca 2ϩ beyond the G␤␥ to the G␣ subunits of G␣ i -coupled receptors. Moreover, our findings point to a signaling function that is activated rather than inhibited by G␣ i proteins and add an important function to the repertoire of functions activated by G␣ i subunits.