Activation of M1 muscarinic acetylcholine receptors stimulates the formation of a multiprotein complex centered on TRPC6 channels.

In this study we showed that stimulation of M1 muscarinic acetylcholine receptors (mAChRs) activates endogenous transient receptor potential-canonical, subtype 6 (TRPC6), channels in neuronal PC12D cells. Activation of TRPC6 channels is correlated with the formation of a multiprotein complex containing M1 mAChRs, TRPC6 channels, and protein kinase C (PKC). Formation of the M1 mAChR-TRPC6-PKC complex is transient, with highest levels reached approximately 2 min after stimulation of M1 mAChRs. PKC in the complex phosphorylates TRPC6 on a conserved serine residue in the carboxyl-terminal domain (Ser768 in the TRPC6A isoform and Ser714 in the TRPC6B isoform). The immunophilin FKBP12, the phosphatase calcineurin, and Ca2+-binding protein calmodulin are also recruited to the M1 mAChR-TRPC6-PKC complex following activation of M1 mAChRs and remain stably associated with the TRPC6 channels after M1 mAChRs and PKC have disassociated. Binding of FKBP12, calcineurin, and calmodulin to TRPC6 channels is blocked by the following: 1) inhibition of PKC; 2) mutation of the PKC phosphorylation site (Ser(7168/714)) in the channels; or 3) pretreatment with FK506 or rapamycin, immunosuppressants that directly bind FKBP12. Inhibition of FKBP12 binding blocks the dephosphorylation of TRPC6 channels and the disassociation of M1 mAChRs, without affecting disassociation of PKC. The calcineurin inhibitor cyclosporin A also blocks the dephosphorylation of TRPC6 and prevents the disassociation of M1 mAChRs. Together, these results show that activated TRPC6 channels form the center of a dynamic multiprotein complex that includes PKC and calcineurin, which respectively phosphorylate and dephosphorylate the channels. Phosphorylation of the TRPC6 channels by PKC is required for the binding of FKBP12, which in turn is required for the binding of calcineurin and calmodulin. Subsequent dephosphorylation of the channels by calcineurin is required for the disassociation of M1 mAChRs.

In this study we showed that stimulation of M1 muscarinic acetylcholine receptors (mAChRs) activates endogenous transient receptor potential-canonical, subtype 6 (TRPC6), channels in neuronal PC12D cells. Activation of TRPC6 channels is correlated with the formation of a multiprotein complex containing M1 mAChRs, TRPC6 channels, and protein kinase C (PKC). Formation of the M1 mAChR-TRPC6-PKC complex is transient, with highest levels reached ϳ2 min after stimulation of M1 mAChRs. PKC in the complex phosphorylates TRPC6 on a conserved serine residue in the carboxylterminal domain (Ser 768 in the TRPC6A isoform and Ser 714 in the TRPC6B isoform). The immunophilin FKBP12, the phosphatase calcineurin, and Ca 2؉ -binding protein calmodulin are also recruited to the M1 mAChR-TRPC6-PKC complex following activation of M1 mAChRs and remain stably associated with the TRPC6 channels after M1 mAChRs and PKC have disassociated. Binding of FKBP12, calcineurin, and calmodulin to TRPC6 channels is blocked by the following: 1) inhibition of PKC; 2) mutation of the PKC phosphorylation site (Ser 7168/714 ) in the channels; or 3) pretreatment with FK506 or rapamycin, immunosuppressants that directly bind FKBP12. Inhibition of FKBP12 binding blocks the dephosphorylation of TRPC6 channels and the disassociation of M1 mAChRs, without affecting disassociation of PKC. The calcineurin inhibitor cyclosporin A also blocks the dephosphorylation of TRPC6 and prevents the disassociation of M1 mAChRs. Together, these results show that activated TRPC6 channels form the center of a dynamic multiprotein complex that includes PKC and calcineurin, which respectively phosphorylate and dephosphorylate the channels. Phosphorylation of the TRPC6 channels by PKC is required for the binding of FKBP12, which in turn is required for the binding of calcineurin and calmodulin. Subsequent dephosphorylation of the channels by calcineurin is required for the disassociation of M1 mAChRs.
Stimulation of M1 mAChRs 1 activates phospholipase C-␤ (PLC␤), resulting in increased production of (DAG) and inositol 1,4,5-trisphosphate (IP 3 ) (1). DAG activates protein kinase C (PKC) (2), and IP 3 binds to the IP 3 receptor (IP 3 R) channels in the endoplasmic reticulum (ER), causing the channels to open and release stored Ca 2ϩ (1)(2)(3)(4)(5). Depletion of these stores activates Ca 2ϩ channels in the plasma membrane, increasing Ca 2ϩ levels in the cytoplasm and facilitating the reuptake of Ca 2ϩ into the ER (6,7). Ca 2ϩ channels that open in response to the depletion of ER Ca 2ϩ stores are often referred to as storeoperated channels (SOCs) (8). In addition to SOCs, there are Ca 2ϩ -permeable channels that are activated by G q -coupled receptors independently of the depletion of ER Ca 2ϩ stores. There is currently wide interest in identifying the molecular components of both types of channels and determining the mechanisms by which they are activated and inhibited.
During the last 10 years, cDNAs encoding seven mammalian homologs of Drosophila transient receptor potential and transient receptor potential-like channels have been isolated by molecular cloning (9,10). These channels, designated TRPC (transient receptor potential-canonical), subtypes 1-7 (11), are currently the subject of intense investigation for their potential roles as mammalian SOCs and Ca 2ϩ storeindependent channels (12). Based upon similarities in amino acid sequences, the mammalian TRPCs can be classified into the following four groups: TRPC1, TRPC2, TRPC4/5, and TRPC3/6/7 (13). The TRPC1 subtype is widely expressed in tissues throughout the body, including heart and brain (14). It has been proposed to contribute to store-operated Ca 2ϩ influx (15,16), possibly functioning as heterotetrameric channels containing TRPC4 and/or TRPC5 subunits in some tissues (14,17,18). TRPC2 channels are not expressed in humans but have been shown to be expressed in the vomeronasal organ of rodents, where they function in sex pheromone signaling (19 -21), and in the sperm of mice, where they function in fertilization (22). TRPC4 and -C5 channels are expressed in brain neurons and have variably been described as SOCs or Ca 2ϩ store-independent channels (23). The TRPC3, -C6, and -C7 isoforms form homo-or heterotetrameric (24) channels with each other but not with other TRPC subtypes (18,25). TRPC3/6/7 channels are activated by G q protein-coupled receptors independently of Ca 2ϩ stores, possibly by increases in intramembrane DAG (24,26,27). This activation does not depend upon PKC, and activation of PKC with phorbol esters inhibits the channels (28,29).
PC12D cells (30) are a rapidly differentiating subline of rat pheochromocytoma-derived PC12 (31). We chose PC12D cells for our studies because they express the M1 subtype of mAChRs, which robustly activate PLC␤ and Ca 2ϩ influx following exposure to the mAChR agonist carbachol (32). We showed previously that Ca 2ϩ influx and PKC activation are the major upstream intracellular events required for the activation of the immediate-early genes zif268 and c-fos in these cells (32). These observations led us to investigate in more detail how M1 mAChRs regulate the influx of extracellular Ca 2ϩ .
Our preliminary studies revealed that stimulation of M1 mAChRs activates Ca 2ϩ influx via both Ca 2ϩ store-dependent and -independent pathways. 2,3 The Ca 2ϩ store-dependent pathway is the major Ca 2ϩ influx pathway activated by carbachol and is also activated by thapsigargin, which empties intracellular Ca 2ϩ stores by inhibiting the sarcoplasmic/endoplasmic reticulum calcium ATPase pump. This influx pathway is mediated by as-yet-unidentified SOCs that are (i) permeable to Mn 2ϩ , (ii) impermeable to Ba 2ϩ , (iii) not inhibited by PKC, (iv) not activated by the DAG analog 1-oleoyl-2-acetyl-sn-glycerol, (v) active for at least 30 min in the continuous presence of carbachol, and (vi) slow to inactivate following inhibition of M1 mAChR with the mAChR antagonist atropine. (The SOCs become inactive as the ER Ca 2ϩ stores refill.) By contrast, the Ca 2ϩ -store independent pathway depends upon channels that are (i) impermeable to Mn 2ϩ , (ii) permeable to Ba 2ϩ , (iii) potently inhibited by PKC, (iv) activated by 1-oleoyl-2-acetyl-sn-glycerol, (v) active for only 2-3 min in the continuous presence of carbachol, and (vi) rapidly inactivated following inhibition of M1 mAChR with atropine. The Ca 2ϩ store-independent channels, but not the SOCs, are blocked in PC12D cells expressing TRPC6 antisense RNA or a TRPC6 amino-terminal peptide, both shown previously (33) to inhibit TRPC6 channels exogenously expressed in COS-7 cells.
The goal of the experiments described in this paper was to elucidate molecular mechanisms by which M1 mAChRs regulate endogenous TRPC6 channels in PC12D cells. These studies show that activation of M1 mAChRs induces the transient formation of a multiprotein complex containing M1 mAChRs, TRPC6, and PKC. The immunophilin FKBP12, the phosphatase calcineurin, and the Ca 2ϩ -binding protein calmodulin are also recruited into this complex but maintain an association with TRPC6 after M1 mAChRs and PKC disassociate. TRPC6 channels in the complex are sequentially phosphorylated by PKC and dephosphorylated by calcineurin during the first 5 min following activation of M1 mAChRs. Phosphorylation of the channels by PKC is required for the recruitment of FKBP12, calcineurin, and calmodulin to the complex. The recruitment of activated calcineurin is required for channel dephosphorylation and the disassociation of M1 mAChRs. These novel findings provide a new framework for understanding the molecular mechanisms by which TRPC6 channels are regulated.
Measurement of Ba 2ϩ Influx-Fifty to 75% confluent PC12D cells were loaded with Fura 2-AM (Calbiochem; 2 M in Krebs-Ringer/ HEPES (KRH) buffer: 6 mM HEPES-NaOH, 125 mM NaCl, 5 mM KCl, 1.2 mM KH 2 PO 4 , 1.2 mM MgSO 4 ⅐7H 2 O, 2 mM CaCl 2 ⅐2H 2 O, 6 mM glucose (pH 7.4)) for 90 min at room temperature in the dark. Cells in individual wells were gently washed four times with nominally Ca 2ϩ -free KRH buffer and immediately used for Ba 2ϩ influx measurements. Changes in intracellular fluorescence were measured in individual cells or small groups of cells using a calcium imaging system comprising a Leica DMIRB inverted microscope equipped with a Lambda G-4 xenon light source (175 watts) and Lambda 10-2 filter wheel (Sutter Instruments, Novato, CA) and Orca-ER CCD camera (Hamamatsu Photonics). Hardware and date collection were performed using OPENLAB 3 imaging software (Improvision, Lexington, MA) running on a Macintosh G4 computer. Fura-2-loaded cells were stimulated with 340 nm light at 0.5-s intervals, and fluorescence was recorded at 510 nm. Data processing and graphing were carried out using Microsoft Excel.
Preparation of siRNAs and Transfection of PC12D Cells-Total cellular RNA from PC12D cells or rat brain was isolated using Trizol reagent (Invitrogen). For each siRNA preparation, 1 g of purified RNA was treated with DNA-Free (Ambion, Austin, TX) to remove genomic DNA prior to use as a template for cDNA synthesis using the reagents in the SuperScript One-step PCR kit (Invitrogen), according to the directions of the manufacturers. RT and PCR amplifications were carried out in the same tubes using a Bio-Rad Thermocycler. PCR primers were designed to amplify segments of the TRPC2 and TRPC6B cDNAs that lack sequence homology to each other and to rat TRPC1, -C3, -C4, -C5, and -C7 cDNAs. A T7 promoter sequence (underlined) 5Ј-TAAT-ACGACTCACTATAGGG-3Ј was included in each of the primers to allow the amplified PCR products to be used as templates for in vitro synthesis of double-stranded (ds) RNA. Primers designed to amplify cDNA segments encoding amino-terminal and carboxyl-terminal regions of TRPC6B were as follows: 5Ј-TAATACGACTCACTATAGGGT-TAAGAAATGGATGTCTGA-3Ј (C6-N forward primer 1); 5Ј-TAATACG-ACTCACTATAGGGTCTTGTTTCGACTCTAAAGA-3Ј (C6-N reverse primer 1); 5Ј-TAATACGACTCACTATAGGGAATGAAAACAGACTGA-CTCA-3Ј (C6-C forward primer 2); and 5Ј-TAATACGACTCACTATAG-GGAAGCTGGATGGTTGAGAATA-3Ј (C6-C reverse primer 2). Primers designed to amplify cDNA segments encoding amino-terminal and carboxyl-terminal regions of TRPC2 were as follows: 5Ј-TAATACGAC-TCACTATAGGGTTCGCCCAACTGGACTGAGA-3Ј (C2-N forward primer 1); 5Ј-TAATACGACTCACTATAGGGTCCAGGGCAATGTACT-GAGG-3Ј (C2-N reverse primer 1); 5Ј-TAATACGACTCACTATAGGGT-CTGGGCAACAGACTGACAG-3Ј (C2-C forward primer 2), and 5Ј-TA-ATACGACTCACTATAGGGCCCTTGGTCTCCAGATCTTC-3Ј (C2-C reverse primer 2). Reaction mixes contained the following: 1ϫ RT-PCR Reaction buffer, 1 g of purified PC12D RNA, 200 nM forward primer, 200 nM reverse primer, 1 l of RT/Platinum Taq Mix, and nuclease-free water added to obtain a final volume of 50 l. The temperature cycling program for the RT and PCR steps was as follows: 1ϫ (45°C for 30 min; 94°C for 2 min), 5ϫ (94°C for 15 s; 45°C for 30 s; 72°C for 60 s), 35ϫ (94°C for 15 s; 50°C for 30 s; 72°C for 60 s), 1ϫ (72°C for 10 min), and 1ϫ (4°C overnight). Following isolation by ethanol precipitation and resuspension in 20 l of nuclease-free water, the PCR products were used to generate siRNAs using reagents in the Dicer siRNA generation kit (Gene Therapy Systems, Inc, San Diego, CA), according to the directions of the manufacturer. Briefly, double-stranded RNAs were synthesized from 1 g of PCRs product using T7 RNA polymerase, treated with DNase to remove the PCR templates, purified by precipitating with LiCl and washing with 70% ethanol, and resuspended in nuclease-free water. siRNAs were generated from the double-stranded RNAs by digestion with Dicer, purified using columns provided in the kit, and quantified by measuring the absorption at 260 nm.
Plasmid Construction-Rat TRPC6B cDNA lacking a stop codon was obtained by PCR amplification from pEF-BOS-TRPC6B (33) using the forward primer 5Ј-ATGAGCCGGGGTAATGAAAAG-3Ј and reverse primer 5Ј-TCTGCGGCTTTCCTCTTGTTT-3Ј. The PCR product was subcloned into pCRII-TOPO vector (Invitrogen). The TRPC6B insert was excised by EcoRI digestion and cloned in EcoRI-digested pCMV-Tag5A (Stratagene), which encodes a "Myc" tag downstream from the EcoRI restriction site. Digestion with HindIII was used to screen for correct insertions of the insert, yielding the plasmid pCMV-TRPC6Bmyc. The codon for Ser 714 was converted to GCT (encoding alanine) using the "QuikChange TM site-directed mutagenesis kit" (Stratagene) with the forward primer 5Ј-CCCTTCAATCTTGTGCCAGCTC-CAAAGTGCCTGCTTTATCT-3Ј and reverse primer 5Ј-GAGATAAAG-CAGGGACTTTGGAGCTGGCACAAGATTGAAGGGG-3Ј. Briefly, 50 ng of pCMV-TRPC6B-myc was used as template for PCR amplification with 125 ng (each) forward and reverse primers and 2.5 units of Pfu Turbo DNA polymerase. The PCR conditions were as follows: 1ϫ (95°C for 30 s) and 16ϫ (95°C for 30 s, 55°C for 1 min, and 68°C for 8 min). The reaction mix was cooled on ice for 2 min and treated with 10 units of DpnI at 37°C for 1 h. One l of DpnI-treated DNA was introduced into transformation-competent DH5-alpha cells (Invitrogen) and kanamycin-resistant colonies isolated by culturing on agar plates containing 50 g/ml kanamycin. Plasmids recovered from the bacteria were screened for the presence of the mutation by DNA sequencing. DNA sequencing of the entire TRPC6B-myc coding segment of pEF-CMV-TRPC6B-S714A confirmed that the sequence was identical to the wildtype TRPC6B isoform, except for the Ser 714 codon (AGT), which was changed to a codon for alanine (GCT).
Transfection of PC12D Cells-siRNAs were transfected into PC12D cells using Oligofectamine Transfection Reagent (Invitrogen), as recommended by the manufacturer. Briefly, cells cultivated to ϳ50% confluency in Cellstar 6-well tissue culture plates (Greiner Bio-One, Inc., Longwood, FL) were transfected with 0.5 g of siRNA, 50 ng of pEGFP (enhanced green fluorescent protein expression vector, pEGFP-N2; Clontech). Two or 3 days following transfections, the cells were loaded with Fura-2 as described above. Ba 2ϩ influx measurements were carried out on EGFP-expressing cells identified visually using a fluorescein fluorescence filter cube.
Western Blotting-Proteins were resolved by SDS-PAGE (10% polyacrylamide resolving gel, 5% stacking gel) and were electrophoretically transferred to nitrocellulose membranes (Hybond ECL nitrocellulose, 0.2 m, Amersham Biosciences). After the transfer, membranes were blocked for 2 h at room temperature with TNT buffer (20 mM Tris (pH 7.5), 137 mM NaCl, and 0.1% Tween 20) containing 5% powdered skim milk. The membranes were then exposed to anti-M1 mAChR (1:1000), anti-TRPC6 (1:2000), anti-PKC␣ (1:2000), anti-phospho-Ser PKC substrate (1:2000), anti-FKBP12 (1:2000), anti-calcineurin (1:1000), or anti-calmodulin (1:500) antibodies at the indicated dilutions in blocking buffer overnight at 4°C. The membranes were washed three times with TNT buffer and incubated in TNT buffer containing anti-rabbit IgG antibodies or anti-goat IgG antibody cross-linked with horseradish peroxidase (Jackson ImmunoResearch; 1:5000 dilution) for 1 h at room temperature. Membranes were washed three times, and proteins were visualized by enhanced chemiluminescence (ECL kit, Pierce) according to the manufacturer's instructions. Fig. 1 shows that addition of carbachol to Fura-2-loaded PC12D cells in nominally Ca 2ϩ -free KRH buffer causes a transient increase in intracellular fluorescence, correlating with the release of Ca 2ϩ from the ER and its rapid expulsion from the cells. Subsequent addition of 500 M BaCl 2 to the medium results in an increase in Fura-2 fluorescence due to the influx of Ba 2ϩ . Ba 2ϩ is not a substrate for plasma membrane Ca 2ϩ pumps and is therefore not expelled from the cells (34,35). Thus, Ba 2ϩ -dependent increases in Fura-2 fluorescence reflect Ba 2ϩ influx and accumulation only and not the balance of influx and efflux (as is the case for Ca 2ϩ ). The release of Ca 2ϩ from internal stores and Ba 2ϩ influx are both mediated by M1 mAChRs, because both are blocked in cells pretreated with atropine.

Activation of M1 mAChRs Activates Endogenous TRPC6 Channels in PC12D Cells-
Based upon reverse transcriptase (RT)-PCR and Western blot analyses, we previously determined that PC12D cells express four TRPC channel subtypes: TRPC1, TRPC3, TRPC6, and TRPC7. 3 The TRPC6 channel expressed in PC12D cells is the TRPC6B isoform, 3 which lacks 54 amino acid residues present in the amino terminus of the TRPC6A isoform (33). Carbachol-stimulated Ba 2ϩ influx in PC12D cells is blocked by exogenous expression of anti-TRPC6 mRNA or the amino-terminal domain of TRPC6B, 3 which may act in a dominantnegative manner to block TRPC6 channel assembly (36,37). Fig. 2A shows that carbachol-stimulated Ba 2ϩ influx, but not release Ca 2ϩ from ER stores, is blocked in PC12D cells transfected with siRNA specific for rat TRPC6 mRNA. By contrast, transfection of PC12D cells with anti-TRPC2 siRNA does not block carbachol-stimulated Ba 2ϩ influx. Fig. 2B shows that PC12D cells transfected with anti-TRPC6 siRNA are selectively depleted in TRPC6 protein, with no decrease in levels of TRPC3 or TRPC7 protein. By contrast, TRPC6 protein is not depleted in cells transfected with anti-TRPC2 siRNA.
Activation of M1 mAChRs Stimulates the Formation of a Complex between the Receptors and TRPC6 -To investigate molecular mechanisms underlying the regulation of TRPC6 channels, we explored the possibility that activation of M1 mAChRs induces the formation of a complex that includes the receptors and channels. For this purpose, we examined whether exposure to carbachol allows the M1 mAChRs and TRPC6 to be coimmunoprecipitated. Fig. 3A (top set) shows that stimulation of M1 mAChRs with carbachol for 2 min causes TRPC6 channels to coimmunoprecipitate with M1 mAChRs, and that this coimmunoprecipitation is blocked by pretreatment of the cells with atropine. Fig. 3A (bottom set) shows, conversely, that exposure to carbachol causes the M1 mAChRs to coimmunoprecipitate with TRPC6 channels. Again, coimmunoprecipitation is blocked by atropine. Taken together, these results indicate that activation of M1 mAChRs stimulates the formation of a complex with TRPC6. It is not clear whether this complex depends upon direct interactions between the receptor and channel or requires intervening proteins.
M1 mAChR-TRPC6 Complex Formation Is Transient and Reversible-Our previous studies suggested that TRPC6 channels are transiently activated by carbachol, activating and inactivating within about 2-3 min. 3 We therefore examined the time course of the M1 mAChR-TRPC6 complex formation for 3 min following exposure to carbachol. Fig. 3B shows that M1 mAChR-TRPC6 complex formation begins within 20 s after activation of M1 mAChRs. The formation of this complex is reversible, because addition of atropine after 40 s of carbachol exposure not only blocks further increases in coimmunoprecipitations but also reduces previously attained levels.
PKC Is Recruited into the M1 mAChR-TRPC6 Complex-DAG released following activation of M1 mAChRs would be expected to activate PKC. Because DAG is presumably generated in the proximity of M1 mAChRs, we investigated whether PKC is also recruited into the M1 mAChR-TRPC6 complex. Fig. 4A shows that PKC coimmunoprecipitates with TRPC6 channels (top set) and M1 mAChRs (bottom set) following exposure to carbachol for 2 min. Fig. 4B shows that atropine blocks the formation of the M1 mAChR-PKC complex and causes the previously formed complex to disassociate. Identical results were obtained when anti-TRPC6 or anti-PKC antibodies was used for the initial immunoprecipitation and coimmunoprecipitating proteins examined by Western blotting. 4 TRPC6 Channels Are Phosphorylated by PKC-Analysis of the rat TRPC6A and -B amino acid sequences using Prosite (us.expasy.org/tools/scanprosite) (38)  and Western blot analysis of the immunoprecipitates were performed as described under "Experimental Procedures." In the top set, M1 mAChRs that coimmunoprecipitated with TRPC6 channels were detected by probing the transfer membranes with anti-M1 mAChR antibodies (top row). Immunoprecipitation of TRPC6 channels was confirmed by stripping the membranes and reprobing them with anti-TRPC6 antibodies (bottom row). In the bottom set, TRPC6 channels that coimmunoprecipitated with M1 mAChRs were detected by probing the membranes with anti-TRPC6 antibodies (top row). Membranes were then stripped and reprobed with anti-M1 mAChR antibodies (bottom row). The results shown are representative of two independent experiments. Identical results were obtained when PC12D cells in KRH buffer containing 2 mM CaCl 2 or in DMEM were exposed to carbachol plus/minus atropine. 4 B, PC12D cells in DMEM were treated with 500 M carbachol for the indicated times. At 40 s the cells were exposed to vehicle (upper set) or 10 M atropine (bottom set). Cell extracts were prepared in lysis buffer and subjected to immunoprecipitation using anti-M1 mAChR antibodies prior to Western analysis using anti-TRPC6 antibodies (top row in each set). The membranes were then stripped and reprobed with anti-M1 mAChR antibodies (bottom row in each set). The last band in each row is the indicated protein detected in PC12D cell extracts without immunoprecipitation. The results shown (A and B) are representative of two independent experiments. TRPC6A channel isoform (Ser 14 , Thr 629 , Ser 768 , Ser 835 , Ser 892 , and Ser 928 ) and five homologous sites in the TRPC6B channel isoform (Thr 575 , Ser 714 , Ser 781 , Ser 838 , and Ser 874 ). To obtain direct evidence for phosphorylation of TRPC6 by PKC, we examined the ability of antibodies specific for the phosphoserine in PKC substrates to bind TRPC6 channels before and after activation of PKC with phorbol ester. As shown in Fig. 5A (top), anti-phospho-Ser PKC substrate antibodies bind TRPC6 channels immunoprecipitated from cells exposed to PMA. Binding is not observed for TRPC6 channels immunoprecipitated from nonstimulated cells or cells pretreated with GF109203X, a potent inhibitor of the ␣, ␤I, ␤II, and ␥ subtypes of PKC (39). As shown in Fig. 5A (bottom), anti-phospho-Ser PKC substrate antibodies immunoprecipitate TRPC6 channels only from cells exposed to PMA and not from nonstimulated cells or cells pretreated with GF109203X. As shown in Fig. 5B, TRPC6 channels are also phosphorylated following exposure to carbachol. This phosphorylation is transient, with peak levels reached within 1.5-2 min. As shown in Fig. 5C, pretreating the cells with GF10203X blocks the binding of the anti-phospho-Ser PKC substrate antibodies to the channels. TRPC6 channel phosphorylation is also blocked in cells pretreated with atropine, 4 indicating that channel phosphorylation depends upon activation of M1 mAChRs. The time course of the phosphorylation parallels the time course of the formation and dissociation of the M1 mAChR-TRPC6-PKC complex.
PKC Phosphorylates a Conserved Serine Residue Located Near the TRPC6 Channel Carboxyl Terminus-A recent study by Trebak et al. (29) showed that TRPC3 channels are phosphorylated by PKC on Ser 712 in the carboxyl-terminal domain. TRPC6A and TRPC6B channels each contain a homologous PKC consensus phosphorylation site, Ser 768 in TRPC6A and Ser 714 in TRPC6B.
To determine whether the carbachol-stimulated phosphorylation of TRPC6 channels in PC12D cells takes place on these conserved serines, we tested the effects of replacing the Ser 768 in TRPC6A and Ser 714 in TRPC6B with alanine. Fig. 6A shows that TRPC6B-myc and TRPC6 S714A-myc channels exogenously expressed in PC12D cells can be immunoprecipitated using anti-TRPC6 or anti-Myc antibodies. Fig. 6B shows that exposure to PMA for 2 min results in phosphorylation of Myctagged wild-type TRPC6B channels and that this phosphorylation is blocked by pretreatment with GF109203X. By contrast, Myc-tagged TRPC6B-S714A channels are not phosphorylated following exposure to PMA. Immunoprecipitation with antiphospho-Ser PKC antibodies showed that endogenous TRPC6 channels are phosphorylated in TRPC6B-myc and TRPC6B-S714A-myc transfected cells following exposure to PMA and that this phosphorylation is blocked by pretreatment with GF109203X. Fig. 6C shows that similar results are obtained following exposure of TRPC6B-myc and TRPC6B-S714A-myc transfected cells to carbachol. Analysis of TRPC6A-S768A-myc yielded similar results. 4 Nearly identical results were also obtained with TRPC6A and -B channels containing glycine in place of Ser 768/714 . 4 Taken together, these data show that PKC phosphorylates TRPC6B channels on Ser 768/714 . Because the anti-phospho-Ser PKC substrate antibodies do not recognize phosphothreonine residues, we could not determine whether PKC phosphorylates Thr 629/575 . Our results show, however, that Ser 768/714 is the major site for serine phosphorylation by PKC in the TRPC6A and -B isoforms.
Activation of M1 mAChRs Recruits Immunophilin FKBP12, Calcineurin, and Calmodulin to the M1 mAChR-TRPC6-PKC Complex-A recent study by Sinkins et al. (40) showed that TRPC6 exogenously expressed in Sf9 insect cells or endogenously expressed in rat brain coimmunoprecipitates with FKBP12 (FK506-binding protein, 12 kDa), an "immunophilin" that is best known as the molecular target of the immunosuppressants FK506 and rapamycin (41).
The cellular functions of FKBP12 are poorly understood, but several studies provide evidence for a role in regulating Ca 2ϩ release from the sarcoplasmic reticulum and endoplasmic reticulum. FKBP12 binds to type 1 ryanodine receptor (RyR1) calcium channels in sarcoplasmic reticulum of striated muscle (42)(43)(44) and to the IP 3 R Ca 2ϩ channels in the endoplasmic reticulum (45)(46)(47), and has been proposed to regulate Ca 2ϩ release through those channels (42)(43)(44)(45)(46)(47). FKBP12.6, an isoform of FKBP12, binds to the type 2 ryanodine receptor (RyR2) of cardiac smooth muscle and regulates its channel properties (48). FKBP12-binding sites in these three channels have been identified, and each contains a conserved leucyl-prolyl (LP) or valyl-prolyl (VP) dipeptide.
Mutagenesis experiments by Sinkins et al. (40) showed that FKBP12 also binds to the carboxyl-terminal domain of TRPC6A at the amino acid sequence 759 LPVPFNLVP 767 . The LP dipeptide and the Pro at position ϩ9 in this sequence are conserved in TRPC1-C7 and in the TRPC6B isoform. The VP dipeptide is conserved in TRPC3, -C6, and -C7, which binds FKBP12, whereas a isoleucyl-propyl (IP) dipeptide is found at the homologous positions in TRPC1, -C4, and -C5, which bind to a distinct immunophilin, FKBP52 (40). Binding between TRPC6 and FKBP12 is disrupted by low concentrations of FK506 (0.5-1 M) (40).
To determine whether FKBP12 also associates with TRPC6 in PC12D cells, we performed reciprocal immunoprecipitation assays using antibodies specific for both proteins. The Western blots depicted in Fig. 7, A and B (1st rows), show that coimmunoprecipitation of FKBP12 with TRPC6 is not observed in unstimulated cells or cells pretreated with atropine but is robustly detected 2 and 10 min following activation of M1 mAChRs with carbachol. Fig. 7, C and D (1st rows), shows that FKBP12 coimmunoprecipitates with M1 mAChRs at 2 min, but not at 10 min, after stimulation of mAChRs with carbachol.
The molecular structure of part of FK506 closely resembles the LP dipeptide. Based upon this resemblance and the fact that the FKBP12-FK506 complex binds calcineurin, Cameron et al. (47) proposed that the FKBP12-LP complex in the IP 3 R functions as a binding site for calcineurin. Unlike the FKBP12-FK506 complex, which potently inhibits calcineurin, FKPB12 bound to an LP-containing segment of the IP 3 R was proposed to serve as anchor for calcineurin without inhibiting its phosphatase activity. In fact, calcineurin bound to the FKPB12-IP 3 R complex was shown to dephosphorylate PKA-phosphorylated MAP-2 and PKC-phosphorylated IP 3 R (46).
Our findings that PKC phosphorylates TRPC6 following M1 mAChR activation combined with the discovery by Sinkins et al. (40) that TRPC6 binds FKBP12 suggested that calcineurin might also be targeted to and dephosphorylate TRPC6 channels. To test this idea, we investigated whether calcineurin is recruited into the M1 mAChR-TRPC6-PKC complex following exposure of the cells to carbachol. Because calcineurin is activated by the binding of the Ca 2ϩ -bound form of calmodulin, we also investigated whether calmodulin is recruited to the complex. As shown in Fig. 7, A and B (2nd and 3rd rows), no coimmunoprecipitation of calcineurin or calmodulin with TRPC6 is observed in unstimulated cells or cells pretreated with atropine, but robust coimmunoprecipitation is observed 2 and 10 min after activation of M1 mAChRs with carbachol in the absence of atropine. Fig. 7, C and D (2nd and 3rd rows), shows that calcineurin and calmodulin coimmunoprecipitate with M1 mAChRs at 2 min, but not 10 min, after exposure to carbachol.
To determine whether the association of FKBP12, calcineurin, and calmodulin with M1 mAChRs depends upon TRPC6 channels, we examined the coimmunoprecipitation of these proteins with M1 mAChRs in cells transfected with anti-TRPC6 siRNA. As shown in Fig. 8, A and B (1st to 4th rows), depletion of TRPC6 channels prevents carbachol-stimulated association of FKBP12, calcineurin, and calmodulin with M1 mAChRs. Surprisingly, association between PKC and M1 mAChRs was also blocked in anti-TRPC6 siRNA-treated cells (Fig. 8, A and B, 5th rows). Fig. 8C shows that TRPC6 protein is not detected in cells pretreated with anti-TRPC6 siRNA. Taken together, these data suggest that TRPC6 channels form the center of the carbachol-stimulated protein complex.
The differences in associations of TRPC6 channels and M1 mAChRs with FKBP12, calcineurin, and calmodulin at 2 and 10 min after exposure to carbachol (Figs. 7 and 8) led us to examine the time course of these associations in more detail. Fig. 9A (top set of blots) shows the time course of association of FKBP12, calcineurin, calmodulin, M1 mAChRs, and PKC with TRPC6 channels during the first 10 min after activation of M1 mAChRs. Although it is not possible to directly compare levels of each protein in the complex (because of likely differences in the affinity between the antibodies used in the Western blot analysis), it is possible to compare times where maximal binding of each protein occurs. Thus, maximal association of M1 mAChRs and PKC with TRPC6 channels occurs ϳ2 min after the addition of carbachol, a time when levels of coimmunoprecipitating FKBP12, calcineurin, and calmodulin are still increasing. M1 mAChR and PKC levels are significantly reduced or absent at 5 min and undetectable at 10 min, at which time the binding of FKBP12, calmodulin, and calcineurin is still robust. These results suggest that M1 mAChRs and PKC associate with TRPC6 channels at the same time or slightly earlier than FKBP12, calcineurin, and calmodulin and that the later proteins remain in the TRPC6 complex after M1 mAChRs and PKC have disappeared.
As predicted from the results of Sinkins et al. (40), pretreatment of PC12D cells with FK506 blocks the association of FKBP12 with TRPC6 (Fig. 9A, bottom set of blots). FK506 also blocks the association of calcineurin and calmodulin with TRPC6, consistent with the proposed role for the FKBP12-TRPC6 complex as an anchor for calcineurin/calmodulin. By contrast, FK506 does not disrupt the interaction between TRPC6 and PKC. Most significantly, pretreatment with FK506 allows the coimmunoprecipitation of M1 mAChRs and TRPC6 at late (5 and 10 min) time points, suggesting that the binding of FKBP12 and/or calcineurin is required for the release of M1 mAChRs from the complex.
Nearly identical results were obtained in experiments examining the coimmunoprecipitation of the FKBP12, calcineurin, and calmodulin with M1 mAChRs (Fig. 9B), PKC (Fig. 9C), or calcineurin (Fig. 9D). In the last case, FK506 was found to block the interaction of calcineurin with TRPC6 and PKC but not with calmodulin, consistent with the ability of the FKBP12-FK506 complex to bind to and inhibit calcineurin/calmodulin. Calcineurin Dephosphorylates TRPC6 Channels-As mentioned above, calcineurin that is presumably anchored to FKBP12 on the IP 3 R was found to dephosphorylate IP 3 Rs previously phosphorylated by PKC (46,47). If calcineurin recruited to the M1 mAChR-TRPC6-PKC complex plays a role in the dephosphorylation of TRPC6 by PKC, pretreatment with calcineurin inhibitors would be expected to prolong TRPC6 phosphorylation. The experiments depicted in Fig. 10 show that this is the case. Pretreatment of PC12D cells with FK506, which binds FKBP12 to form a potent calcineurin inhibitor, prevents the dephosphorylation of TRPC6 channels, which usually occurs 3-10 min after activation of M1 mAChR with carbachol (Fig. 10, A and B). Similarly, pretreatment with cyclosporin A, which binds ubiquitously expressed cyclophilin to form a potent calcineurin inhibitor (49,50), also blocks the dephosphorylation of PKC-phosphorylated TRPC6 (Fig. 10C). By contrast to FK506, the immunosuppressant rapamycin binds to FKBP12, but the resulting complex does not inhibit calcineurin (51,52). Pretreatment of the cells with rapamycin, however, is also effective in preventing the dephosphorylation of PKC-phosphorylated TRPC6 (Fig. 10D). This result suggests that efficient dephosphorylation of the TRPC6 channels requires not only the activation of calcineurin but also the targeting of activated calcineurin to FKBP12 in the M1 mAChR-TRPC6-PKC complex.
To obtain additional insights into the actions of cyclosporin A and rapamycin, we examined the effects of these inhibitors on the formation of the multiprotein complexes. The results depicted in Fig. 11A show that pretreatment of the cells with cyclosporin A prevents the recruitment of calcineurin and calmodulin to the M1 mAChR-TRPC6-PKC complex without interfering with the recruitment of FKBP12 (Fig. 11A, 1st to 3rd rows). As observed for FK506, cyclosporin A also blocks the dissociation of M1 mAChRs and TRPC6 (Fig. 11A, 4th row), without changing the transient association between PKC and TRPC6 (Fig. 11A, 5th row).
Rapamycin is similar to FK506 in that it blocks the recruitment of FKBP12, calcineurin, and calmodulin (Fig. 10B, 1st to  3rd rows). It also blocks the dissociation of M1 mAChR and TRPC6 without affecting the transient association of PKC with TRPC6 (Fig. 11B, 4th and 5th rows). Taken together, these results suggest that recruitment of activated calcineurin to the M1 mAChR-PKC complex is required for the dephosphorylation of PKC-phosphorylated TRPC6 and release of M1 mAChRs from the complex.

PKC Phosphorylation of TRPC6 Channels Is Required for
Binding of FKBP12, Calcineurin, and Calmodulin- Fig. 12A  (1st to 3rd rows) shows that inhibiting PKC with GF109203X blocks the coimmunoprecipitation of FKBP12, calcineurin, and calmodulin with the TRPC6 channels. Similar to FK506, GF109203X has no effect on coimmunoprecipitation of M1 mAChRs and TRPC6 channels at 2 min but inhibits the disassociation of the M1 mAChRs and the channels normally seen at 10 min (Fig. 12A, 4th row). GF109203X has no affect on the association of TRPC6 channels and PKC (Fig. 12A, 5th row). Fig. 12, B and C, shows that coimmunoprecipitation of FKBP12, calcineurin, and calmodulin with TRPC6B channels is blocked when Ser 714 is replaced with alanine. TRPC6B-S714A-myc channels also maintain an association with M1 mAChRs at 10 min after addition of carbachol, even in the absence of pretreatment with GF109203X (compare 4th rows in Fig. 12, B and C). Nearly identical results to those shown in Fig. 12C were obtained with TRPC6A and -B channels containing glycine in place of Ser 768/714 . 4 These observations, together with the results described above, imply that disassociation of M1 mAChRs from TRPC6 channels requires three sequential events as follows: 1) phosphorylation of TRPC6 channels on Ser 768/714 by PKC; 2) binding of FKBP12 to a site immediately adjacent to Ser 768/714 ; and 3) dephosphorylation of Ser 768/714 by calcineurin targeted to the TRPC6/FKBP12 anchor. DISCUSSION In this paper we describe a multiprotein complex containing M1 mAChRs, TRPC6 channels, PKC, immunophilin FKBP12, calcineurin, and calmodulin that forms following the activation of M1 mAChRs in PC12D cells. To our knowledge, this is the first study to show that TRPC6 channels physically associate with M1 mAChRs and the first to show that the channels are phosphorylated by PKC and dephosphorylated by calcineurin. Our working model for the regulation of TRPC6 channels is depicted in Fig. 13.
Activation of M1 mAChRs in PC12D cells stimulates a rapid and robust influx of extracellular Ba 2ϩ (Fig. 1), which we have shown here (Fig. 2) and elsewhere 3 to be mediated by TRPC6 channels. Stimulation of M1 mAChRs in these cells activates PLC␤ 3 resulting in increased levels of the second messengers IP 3 (32) and DAG. Because TRPC6 channels are activated by analogs of DAG (24,26,33), it is likely that DAG participates in the activation of the TRPC6 channels following stimulation of M1 mAChRs in PC12D cells. The observation that Ba 2ϩ accumulation reaches a plateau within 2-3 min after the addition of carbachol (Fig. 1) suggests that the channels are active for only 2-3 min following stimulation of M1 mAChRs.
Addition of carbachol to the cells stimulates the formation of a complex between M1 mAChRs and TRPC6 channels with a time course that parallels the activation of Ba 2ϩ influx (Fig. 3). The formation of this complex depends upon the continuous stimulation of M1 mAChRs and is reversible, because addition of atropine rapidly blocks the formation of new complexes and disrupts previously formed complexes (Fig. 3). Atropine also rapidly blocks carbachol-stimulated Ba 2ϩ influx. 3 Activation of M1 mAChRs recruits PKC to the M1 mAChR-TRPC6 complex with a time course identical to that for M1 mAChRs and TRPC6 channels (Fig. 4). This was somewhat surprising because activation of PKC is correlated with inactivation of TRPC6 channels (28,29,33). To reconcile these two observations, we hypothesize that DAG produced by stimulation of M1 mAChRs activates TRPC6 channels more rapidly than DAG-activated PKC can inhibit the channels. A similar bimodal regulation by DAG has been proposed recently to explain the transient activation of exogenously expressed TRPC3 Disassociation of PKC from TRPC6 channels is not dependent upon the dephosphorylation of the TRPC6 channels. FK506 blocks the binding of FKBP12 to the TRPC6 channels and inhibits calcineurin. Rapamycin blocks the binding of FKBP12 to TRPC6 channels, without inhibiting calcineurin. Cyclosporin A inhibits calcineurin, without blocking the binding of FKBP12 to TRPC6 channels. channels following stimulation of mAChRs in HEK293 cells (28).
By using antibodies that specifically recognize phosphoserine residues in PKC substrates, we demonstrated that TRPC6 channels are phosphorylated following exposure to phorbol ester or carbachol (Fig. 5). Phosphorylation of TRPC6 channels is detected as early as 20 s after activation of M1 mAChRs, peaks at about 2 min, and then declines to undetectable levels by 5 min. These observations are consistent with a recent report showing that PKC phosphorylates TRPC3 channels on a conserved serine residue, Ser 712 , in the carboxylterminal domain (29). Substitution of alanine for Ser 714 in the TRPC6B isoform blocked PKC phosphorylation (Fig. 6), demonstrating that the analogous serine residues are phosphorylated in TRPC3 and TRPC6B channels.
Phosphorylation of TRPC6 channels by PKC is correlated with channel inhibition. Evidence for this includes the following observations: 1) TRPC6 channels are inhibited in PC12D cells exposed to 100 nM PMA; and 2) this inhibition is blocked by pretreating the cells with 2 M GF109203X. 3 Also, Myctagged TRPC6B-S714A channels exogenously expressed in COS-7 cells are activated normally following stimulation of M1 mAChRs but are not inhibited in cells pretreated with 100 nM PMA. 4 By contrast, Myc-tagged TRPC6B wild-type channels are completely inhibited by pretreatment with PMA. 4 In addition to inhibition of the channels by PKC, there must also be PKC-independent mechanisms by which TRPC6 channels are inactivated. Evidence for this includes the observation that pretreating the cells with GF109302X does not increase the rate or extent of carbachol-stimulated Ba 2ϩ influx. 4 Instead, similar or slightly reduced rates of Ba 2ϩ influx are observed in GF109203X-pretreated cells. As discussed below, a possible PKC-independent mechanism for the inactivation of TRPC6 channels is Ca 2ϩ -dependent binding of calmodulin to the carboxyl terminus, as proposed by Zhu and co-workers (54,55).
The recent report by Sinkins et al. (40) showing that TRPC6 channels bind FKBP12 led us to investigate whether this immunophilin is a component of the M1 mAChR-TRPC6-PKC complex. Figs. 7-9 show that FKBP12 does not associate with TRPC6 channels prior to activation of M1 mAChRs but rapidly binds to the M1 mAChR-TRPC6-PKC complex following exposure to carbachol. FKBP12 enters the complex at the same time or slightly later than M1 mAChRs and PKC and remains after M1 mAChRs and PKC have disassociated. In addition to FKBP12, calcineurin and calmodulin are also recruited into the M1 mAChR-TRPC6-PKC complex with a similar time course (Figs. 7-9). Formation of a complex between these proteins and M1 mAChRs is dependent upon TRPC6, because no association is observed in TRPC6-depleted cells (Fig. 8). As predicted by the model of Cameron and co-workers (47), the association of calcineurin and calmodulin with TRPC6 channels is dependent upon the binding of FKBP12, because their binding is blocked by FK506 (Fig. 9).
Our findings differ from those of Sinkins et al. (40) in that we observe coimmunoprecipitation of FKBP12 and TRPC6 channels only after activation of M1 mAChRs, whereas they found FKBP12 and the channels to be constitutively associated. This difference may reflect differences in our experimental systems. Whereas we studied endogenous TRPC6 and FKBP12 in PC12D cells, Sinkins et al. (40) examined exogenously expressed TRPC6 channels and FKBP12 in Sf9 cells. It is possible that high levels of exogenously expressed proteins favor association in the absence of TRPC6 channel activation. Sinkins et al. (40) also showed that FKBP12 and TRPC6 channels coimmunoprecipitate from rat brain lysates. It is possible that the immunoprecipitated FKBP12 from brain was bound to a subset of activated TRPC6 channels. A third possibility is that factors controlling the binding of FKBP12 to TRPC6 channels are different in different types of cells.
As mentioned above, the observation that FK506 blocks the binding of FKBP12 to the TRPC6 channels is reminiscent of the FK506-sensitive binding of FKBP12 to IP 3 R1-3 (45)(46)(47). Based upon the resemblance between FKBP12/FK506 and FKBP12 bound to a conserved LP dipeptide in the IP 3 Rs, Cameron et al. (47) proposed that the binding of FKBP12 creates a docking site for calcineurin on the IP 3 Rs and that calcineurin anchored at this site dephosphorylates IP 3 Rs that had been phosphorylated by PKC. This model is controversial, however, and alternative schemes have been proposed (56). Because our data seemed to fit the model of Cameron et al. (47), we decided to investigate whether calcineurin targeted to TRPC6/FKBP12 can dephosphorylate TRPC6 channels that had been phosphorylated by PKC.
As predicted by the model by Cameron et al., dephosphorylation of TRPC6 channels, which is usually observed 3-10 min after activation of M1 mAChRs, is blocked by the calcineurin inhibitors FK506 and cyclosporin A (Fig. 10). Interpretation of the effects of FK506 is complicated, however, by the fact that FK506 inhibits both the phosphatase activity of calcineurin and the binding of FKBP12 to TRPC6 channels (Fig. 9). By contrast, the calcineurin inhibitor cyclosporin A does not block the binding of FKBP12 to the channels (Fig. 11A). The fact that cyclosporin A blocks the dephosphorylation of PKC-phosphorylated TRPC6 channels therefore provides strong evidence for the involvement of calcineurin. Significantly, dephosphorylation of TRPC6 channels is also blocked by the rapamycin, which blocks the binding of FKBP12 to TRPC6 channels ( Fig.  11B) but does not inhibit calcineurin (51,52). The observation that rapamycin blocks the dephosphorylation of TRPC6 channels suggests that the dephosphorylation of TRPC6 channels requires not only calcineurin activation but also the targeting of the activated calcineurin to the channels.
Because calcineurin is activated by binding the Ca 2ϩ -bound form of calmodulin, we investigated whether calmodulin is recruited to the TRPC6-centered protein complex. As expected, calmodulin enters the complex with the same time course as calcineurin (Fig. 9, A-D). Binding of both proteins to TRPC6 channels is also blocked by FK506 (Fig. 9), rapamycin (Fig.  11B), GF109203X (Fig. 12A), and the S714A mutation (Fig. 12, B and C), each of which blocks the binding of FKBP12. Calmodulin coimmunoprecipitates with FKBP12 and calcineurin in the presence of FK506, consistent with the binding of FKBP12/FK506 to the activated (i.e. calmodulin-bound) form of calcineurin (Fig. 9D). Taken together, these results show that calmodulin and calcineurin bind to TRPC6 in a correlated manner, and the association of both proteins with TRPC6 is dependent on the binding of FKBP12.
By using in vitro binding assays and patch clamp recordings, Zhu and co-workers (54,55) showed that calmodulin can regulate TRPC3 and TRPC4 channel activity by directly binding to a specific site, the CIRB (calmodulin/IP 3 receptor binding) domain, located in the carboxyl-terminal domain of each channel. The site of calmodulin binding in the CIRB domain overlaps with the binding site for IP 3 R, leading the authors to propose that calmodulin modulates binding of IP 3 R to TRPC channels. Specifically, Zhu and co-workers (54,55) hypothesize that activation of TRPC channels involves the displacement of loosely bound ("tethered") calmodulin by an amino-terminal segment of the IP 3 R. Influx of Ca 2ϩ then allows calmodulin to bind tightly to the TRPC channels, preventing further activation by IP 3 R.
As described above, we did not observe the coimmunoprecipitation of calmodulin with TRPC6 prior to activation of M1 mAChRs, and binding of calmodulin after activation was dependent upon the presence of FKBP12 in the complex. Our immunoprecipitation data, however, are not necessarily in conflict with the model of Zhu and co-workers (54,55). Weakly bound calmodulin may have been displaced from the TRPC6 under our immunoprecipitation conditions, and tightly bound calmodulin may be difficult to detect using our antibodies. Thus, although our results do not provide direct support for the direct binding of calmodulin to TRPC6, they also do not exclude this possibility. Because the FKBP12-binding site, 759 LPVPFNLVP 767 (40), does not overlap with the site where calmodulin is predicted to bind, 843 HLNSFSNPPRQ-YQKIMKRLIKRYVLQAQID 872 (54), TRPC6 channels could accommodate the binding of calmodulin at both sites. As mentioned above, the Ca 2ϩ -dependent binding of calmodulin to the carboxyl-terminal segment of TRPC6 is an attractive mechanism for PKC-independent inactivation of the channel.
The observation that inhibition of PKC or substitution of alanine for Ser 714 in TRPC6B blocks the binding of FKBP12, calcineurin, and calmodulin to the channels (Fig. 12, A-C) suggests that phosphorylation of this conserved serine, which is located immediately adjacent to the FKBP12-binding site, is required for the binding of FKBP12. We are currently attempting to directly test the effects of PKC phosphorylation of Ser 768/714 on FKBP12 binding to TRPC6 channels using purified proteins.
The observation that dissociation of M1 mAChRs from TRPC6 channels is inhibited by blocking PKC phosphorylation of the channels with GF109203X or by substitution of alanine for Ser 768/714 (Fig. 12, A-C) is probably related to the requirement for phosphorylation of Ser 768/714 for the binding of FKBP12, i.e. phosphorylation of the channels by PKC is required for the binding of FKBP12, and the binding of FKBP12 is required for targeting calcineurin to the complex. As described for calcineurin-mediated dephosphorylation of the TRPC6 channels, release of M1 mAChRs from the complex requires both the activation of calcineurin by calmodulin and the targeting of the activated calcineurin to the complex. The requirement for activated calcineurin is shown by the fact that the calcineurin inhibitor cyclosporin A blocks the disassociation of M1 mAChRs from TRPC6 channels without blocking the binding of FKBP12 (Fig. 11A). The requirement for targeting activated calcineurin to the complex is shown by the fact that rapamycin blocks the disassociation of M1 mAChRs (Fig. 11B). We are currently attempting to determine whether dephosphorylation of TRPC6 channels is required for the dissociation of M1 mAChRs.
Taken together, the results obtained in this study are consistent with the following sequence of events (Fig. 13). 1) Activation of M1 mAChRs stimulates the formation of a multiprotein complex containing the M1 mAChRs, TRPC6 channels, and PKC. 2) PKC phosphorylates the channels on a conserved serine located adjacent to the FKBP12-binding site. 3) FKBP12 binds to the phosphorylated channels. 4) Activated calcineurin/ calmodulin binds to the FKBP12/TRPC6 anchor, either before or after calcineurin dephosphorylates the channels. 5) Dephosphorylation of the channels (or perhaps another protein in the complex) by calcineurin triggers the release of M1 mAChRs from the TRPC6 channels.
This study provides several new insights into the molecular rearrangements and covalent modifications undergone by TRPC6 channels following their activation. Additional studies will be required to determine whether the multiprotein complex detected by coimmunoprecipitation results from direct interactions between the various protein components or depends upon indirect interactions through an as-yet-unidentified molecular scaffold. The Drosophila transient receptor potential channels in the rhabdomere of the photoreceptor cells of the eye are known to be linked to PLC and PKC via the molecular scaffold INAD (57)(58)(59). TRPC1 channels have recently been shown to function within a molecular complex containing PLC␤, G␣ q/11 , IP 3 R, calmodulin, caveolin, and the molecular adapter homer (60 -63). TRPC4 and TRPC5 channels have been shown to bind the molecular scaffold Na ϩ /H ϩ exchanger regulatory factor (53), which may be considered one of the functional equivalents of INAD in mammals. Na ϩ /H ϩ exchanger regulatory factor does not bind TRPC6 channels, however, and alternative candidate scaffolding proteins have not yet been described.
Many more experiments will be required to understand the molecular mechanisms by which TRPC6 channels are regulated and the physiological significance of this regulation. The present study shows, however, that a satisfactory explanation for channel activation and inhibition must take into account the roles of multiprotein complex formation and channel phosphorylation and dephosphorylation.