Caveolin-1 Assembles Type 1 Inositol 1,4,5-Trisphosphate Receptors and Canonical Transient Receptor Potential 3 Channels into a Functional Signaling Complex in Arterial Smooth Muscle Cells*

Physical coupling of sarcoplasmic reticulum (SR) type 1 inositol 1,4,5-trisphosphate receptors (IP3R1) to plasma membrane canonical transient receptor potential 3 (TRPC3) channels activates a cation current (ICat) in arterial smooth muscle cells that induces vasoconstriction. However, structural components that enable IP3R1 and TRPC3 channels to communicate locally are unclear. Caveolae are plasma membrane microdomains that can compartmentalize proteins. Here, we tested the hypothesis that caveolae and specifically caveolin-1 (cav-1), a caveolae scaffolding protein, facilitate functional IP3R1 to TRPC3 coupling in smooth muscle cells of resistance-size cerebral arteries. Methyl-β-cyclodextrin (MβCD), which disassembles caveolae, reduced IP3-induced ICat activation in smooth muscle cells and vasoconstriction in pressurized arteries. Cholesterol replenishment reversed these effects. Cav-1 knockdown using shRNA attenuated IP3-induced vasoconstriction, but did not alter TRPC3 and IP3R1 expression. A synthetic peptide corresponding to the cav-1 scaffolding domain (CSD) sequence (amino acids 82–101) also attenuated IP3-induced ICat activation and vasoconstriction. A cav-1 antibody co-immunoprecipitated cav-1, TRPC3, and IP3R1 from cerebral artery lysate. ImmunoFRET indicated that cav-1, TRPC3 channels and IP3R1 are spatially co-localized in arterial smooth muscle cells. IP3R1 and TRPC3 channel spatial localization was disrupted by MβCD and a CSD peptide. Cholesterol replenishment re-established IP3R1 and TRPC3 channel close spatial proximity. Taken together, these data indicate that in arterial smooth muscle cells, cav-1 co-localizes SR IP3R1 and plasma membrane TRPC3 channels in close spatial proximity thereby enabling IP3-induced physical coupling of these proteins, leading to ICat generation and vasoconstriction.

Several G-protein-coupled receptor agonists activate phospholipase C leading to generation of the second messenger, inositol 1,4,5-trisphosphate (IP 3 ) 2 (1). IP 3 binds to endo/sarcoplasmic reticulum-localized IP 3 receptors leading to Ca 2ϩ release from the endo/sarcoplasmic reticulum (2). In arterial smooth muscle cells, the resultant increase in intracellular Ca 2ϩ concentration ([Ca 2ϩ ] i ) causes vasoconstriction. IP 3 also activates a non-selective cation current (I Cat ), leading to vasoconstriction (3)(4)(5)(6)(7). IP 3 -induced I Cat activation occurs independently of sarcoplasmic reticulum (SR) Ca 2ϩ release and due to a physical interaction between type 1 IP 3 receptors (IP 3 R1) and canonical transient receptor potential (TRPC) 3 channels (4 -6). IP 3 R1 selectively couples to TRPC3 channels due to close spatial proximity of these proteins (4). A wide variety of signal transduction pathways require interacting molecules to form functional macromolecular complexes. These complexes often contain one or more scaffolding proteins that traffic, localize, or compartmentalize signaling molecules (8). However, whether scaffolding proteins are required to assemble IP 3 R1 and TRPC3 channels into a functional macromolecular complex in arterial smooth muscle cells is unclear.
Caveolae are plasma membrane microdomains that act as compartments for several signal transduction molecules, including receptors, ion channels, and scaffolding proteins (9). Caveolae integrity is sustained by cholesterol and caveolin proteins, of which there are three different isoforms (cav-1-3). Caveolin isoforms can exhibit tissue-specific expression (9). Cav-1 is required for caveolae formation in arterial smooth muscle cells, embryonic fibroblasts, lung endothelial, and epithelial cells (10 -13). Cav-1 ablation and caveolae disruption attenuate both pressure-and agonist-induced vasoconstriction, indicating that caveolae play an important role in the regulation of vascular reactivity (11, 14 -17). In arterial smooth muscle cells, peripheral SR membrane can be located nearby plasma membrane caveolae (16). This close spatial proximity may permit or enhance interactions between SRand caveolae-localized plasma membrane proteins. Whether IP 3 R1, TRPC channels, and caveolin proteins physically and functionally interact in arterial smooth muscle cells is unclear.
* This work was supported, in whole or in part, by National Institutes of Health NHLBI Grants R01 HL67061 and HL094378 (to J. H. J.), and K01 HL096411 (to A. A.). 1 To whom correspondence should be addressed: 894 Union Ave., Memphis TN 38163. Tel.: 901-448-1208; Fax: 901-448-7126; E-mail: jjaggar@uthsc. edu. 2 The abbreviations used are: IP 3  In the present study, we examined the significance of caveolae in mediating IP 3 -induced I Cat activation in cerebral artery smooth muscle cells and vasoconstriction in pressurized cerebral arteries. We show that caveolae disassembly using cyclodextrin, cav-1 knockdown using RNAi, or a cav-1 scaffolding domain (CSD) peptide that competes with endogenous cav-1 for interacting protein binding, inhibit IP 3 -induced I Cat activation in smooth muscle cells and vasoconstriction. Data indicate that IP 3 R1 and TRPC3 physically interact with cav-1, and that cav-1 and caveolae maintain plasma membrane TRPC3 and SR IP 3 R1 in a functional signaling complex in cerebral artery smooth muscle cells, thereby permitting IP 3 -induced vasoconstriction.

Tissue Preparation and Smooth Muscle Cell
Isolation-Animal protocols used were reviewed and approved by the Animal Care and Use Committee at the University of Tennessee Health Science Center. Male Sprague-Dawley rats (6 -8 weeks) were euthanized by intraperitoneal injection of sodium pentobarbital (150 mg/kg). The brain was removed and resistance-size cerebral arteries (posterior cerebral, cerebellar, and middle cerebral, Ͻ200 m) were harvested and maintained in ice-cold (4°C), oxygenated (21% O 2 -5% CO 2 -74% N 2 ) physiological saline solution (PSS) containing (in mM): 119 NaCl, 4.7 KCl, 24 NaHCO 3 , 1.2 KH 2 PO 4 , 1.6 CaCl 2 , 1.2 MgSO 4 , and 11 glucose (pH 7.4). Individual smooth muscle cells were dissociated from cerebral arteries using a HEPESbuffered solution containing (in mM): 55 NaCl, 80 sodium glutamate, 5.6 KCl, 2 MgCl 2 , 10 HEPES, and 10 glucose (pH 7.3), as previously described (18) Briefly, cerebral arteries were placed into isolation solution containing 0.7 mg/ml papain, 1 mg/ml dithioerythreitol, and 1 mg/ml bovine serum albumin (BSA) for 12 min at 37°C. Arteries were then immediately transferred to isolation solution containing 1 mg/ml collagenase F and H (2:1), 100 M CaCl 2 , and 1 mg/ml BSA for 10 min at 37°C. Arteries were subsequently washed in isolation solution and triturated using a fire-polished glass pipette to yield single smooth muscle cells. Cells were maintained at 4°C, and used for experimentation within 8 h.
Pressurized Artery Diameter Measurement-Arterial diameter was measured using pressurized artery myography, as previously described (19). Briefly, an arterial segment 1-2 mm in length was cannulated at each end in a pressurized artery chamber (Living Systems Instrumentation; Burlington, VT).
The chamber was continuously perfused with PSS equilibrated with a mixture of 21% O 2 , 5% CO 2 , 74% N 2 , and maintained at 37°C. Arterial endothelium was denuded by introducing air bubbles into the artery lumen for 1 min followed by a wash with PSS. Intravascular pressure was changed using an attached reservoir and monitored using a pressure transducer. Pressurized arteries were observed with a charge-coupled device camera attached to an inverted microscope (Nikon TE 200). Arterial diameter was measured at 1 Hz using the automatic edge-detection function of IonWizard software (Ionoptix; Milton, MA). Pharmacological agents were applied via chamber perfusion after arteries had developed steady-state spontaneous myogenic tone. Myogenic tone (%) was calculated as (1 Ϫ active diameter/passive diameter) ϫ 100.
Co-immunoprecipitation and Western Immunoblotting-Arterial lysate (yielding ϳ1.5 mg total protein) was harvested from cerebral arteries using ice-cold non-denaturing lysis buffer (Thermo Scientific, Rockford, IL). Because of the small size (Ͻ200 m diameter) of the cerebral arteries used in this study, arteries pooled from ϳ15 rats were required for each co-IP experiment. Co-immunoprecipitation (co-IP) was done using the Thermo Scientific Pierce co-IP kit following the manufacturer's protocol. Briefly, the cav-1 antibody (Abcam) was first immobilized for 2 h using AminoLink Plus coupling resin. The resin was then washed and incubated with arterial lysate overnight. After incubation, the resin was again washed and protein eluted using elution buffer. A negative control that was provided with the IP kit to assess nonspecific binding received the same treatment as the co-IP samples, including the cav-1 antibody. In this control, the coupling resin is not amine-reactive preventing covalent immobilization of the primary antibody onto the resin. Samples were analyzed by Western blotting using rabbit polyclonal anti-cav-1 (Abcam Inc., Cambridge, MA), mouse monoclonal anti-IP 3 R1 (UC Davis/NINDS/NIMH NeuroMab Facility), rabbit polyclonal anti-TRPC3 (Alomone Labs Ltd. Jerusalem, Israel), and horseradish peroxidase-conjugated secondary antibodies.
Western blotting was performed as previously described (4 -6). Briefly, rat cerebral artery lysate protein concentrations were determined spectrophotometrically. Proteins were separated by 7.5% gradient SDS-polyacrylamide gel electrophoresis and transferred onto nitrocellulose membranes using a Mini Trans Blot Cell (Bio-Rad). Membranes were then incu-bated with respective antibodies and developed using enhanced chemiluminescence (Thermo Scientific).
Chemical Loading-Caveolin-1 scaffolding domain peptide (CSD) was inserted into intact cerebral arteries using the chemical loading technique modified from a previously described method (14,21). Briefly, cerebral arteries were first placed into a Ca 2ϩ -free (EGTA, 10 mM), high K ϩ (120 mM) solution supplemented with ATP (5 mM) at 4°C for ϳ30 min. Arteries were then incubated in the presence of CSD in Mg 2ϩ (2 mM)-and ATP (5 mM)-containing Ca 2ϩ -free solution at 4°C for ϳ2 h. Arteries were further placed in 10 mM Mg 2ϩand ATP (5 mM)-containing Ca 2ϩ -free solution at 4°C for ϳ30 min, after which the solution was changed to one containing Na ϩ (140 mM) and K ϩ (5 mM) and maintained at room temperature for 30 min. Ca 2ϩ was then added incrementally to the solution to reach a concentration of 1.8 mM. Arteries were then placed into DMEM-F12 supplemented with 1% penicillin and streptomycin overnight (37°C, 5% CO 2 ) prior to use.
Immunofluorescence Resonance Energy Transfer (Immuno-FRET) and Confocal Imaging-ImmunoFRET was performed using a modified version of our previously described method (4). Briefly, paraformaldehyde-fixed cerebral artery smooth muscle cells were permeabilized with 0.1% Triton X-100 for 1 min at room temperature. Following a 1 h of incubation in PBS containing 5% bovine serum albumin (BSA), smooth muscle cells were treated overnight at 4°C with mouse monoclonal anti-IP 3 R1 (UC Davis/NINDS/NIMH NeuroMab Facility), sheep polyclonal anti-TRPC3 (Abcam), rabbit polyclonal anti-cav-1 (Abcam), or anti-IP 3 R1 plus anti-TRPC3 or anti-TRPC3 plus anti-cav-1 at a dilution of 1:100 each in PBS containing 5% BSA. After a wash and block with PBS containing 5% BSA, cells were incubated for 1 h at 37°C with secondary antibodies: Alexa 546-conjugated donkey anti-mouse for IP 3 R1 (1:100 dilution; Invitrogen, Carlsbad, CA) and Alexa 488-conjugated donkey anti-sheep for TRPC3 (1:100 dilution; Invitrogen). For FRET measurements using TRPC3 and cav-1 primary antibodies, Alexa 546-conjugated donkey anti-sheep and Alexa 488-conjugated goat anti-rabbit secondary antibodies were used, respectively. Following wash and mount, fluorescence images were acquired using a Zeiss LSM Pascal laser-scanning confocal microscope. Alexa 488 and Alexa 546 were excited at 488 and 543 nm and emission collected at 505-530 and Ͼ560 nm, respectively. Images were acquired using a z-resolution of ϳ0.8 m. Negative controls prepared by omitting primary antibodies did not exhibit fluorescence. Images were background-subtracted and normalized FRET (N-FRET) was calculated on a pixel-by-pixel basis for the entire image and in regions of interest (within the boundaries of the cell) using the Xia method (22) and Zeiss LSM FRET Macro tool version 2.5.
Fluorescein isothiocyanate (FITC)-labeled antennapedia peptide was inserted into smooth muscle cells of intact cerebral arteries using chemical loading procedure. FITC fluorescence was background-subtracted. Control arteries underwent the same chemical loading and culture protocol, but were not exposed to FITC-labeled antennapedia. FITC fluorescence was imaged using a Zeiss LSM5 Pascal laser-scan-ning confocal microscope. FITC was excited at 488 nm and emission collected at 505-530 nm.
Statistical Analysis-Data are expressed as means Ϯ S.E. of the mean. Statistical significance was calculated by using Student's t-tests for paired or unpaired data and analysis of variance followed by Student-Newman-Keuls post-hoc test for multiple data sets. p Ͻ 0.05 was considered significant.

Caveolae Are Required for IP 3 -induced I Cat Activation in Arterial Smooth
Muscle Cells and Vasoconstriction-To examine the significance of caveolae for IP 3 -induced I Cat activation, freshly-isolated cerebral artery smooth muscle cells were exposed to methyl ␤-cyclodextrin (M␤CD; 30 min), a cholesterol-depleting agent that disassembles caveolae (15,23). M␤CD treatment reduced mean smooth muscle cell capacitance (control versus M␤CD-treated (pF): 12.1 Ϯ 1.0, n ϭ 6 versus 7.4 Ϯ 1.2, n ϭ 7; p Ͻ 0.05), consistent with a decrease in cell surface area. M␤CD treatment also reduced IP 3 -induced I Cat density in arterial smooth muscle cells by ϳ89% (Fig. 1, A and B). To study the importance of cholesterol in  FEBRUARY 11, 2011 • VOLUME 286 • NUMBER 6 JOURNAL OF BIOLOGICAL CHEMISTRY 4343 mediating this effect, cholesterol-loaded M␤CD (Chol/ M␤CD; ϳ1 h) was used to replenish cholesterol in M␤CDtreated arterial smooth muscle cells (15,23,24). Cholesterol replenishment reversed the M␤CD-induced decrease in cell capacitance (to 10.7 Ϯ 0.9 pF, n ϭ 6, p Ͼ 0.05 when compared with control). Chol/M␤CD also reversed the inhibitory effect of M␤CD on IP 3 -induced I Cat activation (Fig. 1, A and B).

Caveolin-1 Assembles IP 3 R1-TRPC3 Signaling Complex
The functional significance of caveolae for IP 3 -induced vasoconstriction was investigated by measuring diameter regulation of pressurized (60 mmHg), endothelium-denuded cerebral arteries. Control arteries developed ϳ25% mean myogenic tone (Fig. 2, A and B). M␤CD treatment (2 h) reduced myogenic tone to ϳ7%, or by ϳ72% (Fig. 2, A and B). M␤CD treatment also reduced vasoconstriction induced by Bt-IP 3 , a membrane-permeant IP 3 analog, by ϳ79% (Fig. 2, C  and D). Chol/M␤CD treatment (ϳ2 h) restored both myogenic tone and IP 3 -induced vasoconstriction in arteries treated with M␤CD (Fig. 2, A-D). These data suggest that intact caveolae are necessary for IP 3 -induced I Cat activation in arterial smooth muscle cells and myogenic tone and IP 3 -induced vasoconstriction in pressurized arteries.
CSD Peptide Attenuates IP 3 -induced I Cat Activation and Vasoconstriction-The cav-1 N terminus contains a CSD that interacts with and regulates a wide variety of signaling proteins containing a CSD-interacting motif (25,26). CSD-interacting motifs are characterized as ⌽X⌽XXXX⌽ and ⌽XXXX⌽XX⌽, where ⌽ is a tryptophan, phenylalanine, or  tyrosine, and X is any amino acid (26). As shown in Fig. 5A, the N and C termini of IP 3 R1 and the C terminus of TRPC3 channels contain CSD-binding motifs, suggesting that IP 3 R1 and TRPC3 channels may both interact with cav-1.
To study the physiological function of the CSD in IP 3 -induced I Cat activation and vasoconstriction, we used a synthetic peptide corresponding to the cav-1 CSD sequence (amino acids 82-101, DGIWKASFTTFTVTKYWFYR). This CSD peptide competes with endogenous cav-1 for interacting protein binding. CSD peptide was conjugated to an antennapedia (ANT) internalization sequence (ANT-CSD, amino acids 43-58, RQIKIWFQNRRMKWKK) to improve membrane permeability. An ANT peptide was used as an experimental control. To examine membrane permeability of the peptide, a FITC-labeled ANT peptide (ANT-FITC) was used. Confocal imaging of arteries exposed to ANT-FITC indicated successful introduction into smooth muscle cells (Fig. 5B).
Cerebral Artery cav-1 Immunoprecipitates with Both TRPC3 and IP 3 R1-To test the hypothesis that cav-1 exists in a macromolecular protein complex with IP 3 R1 and TRPC3 channels, co-IP was performed. A cav-1 antibody co-immunoprecipitated cav-1, IP 3 R1, and TRPC3 from cerebral artery lysate (Fig. 7). These data indicate that cav-1, IP 3 R1, and TRPC3 channels are located within the same macromolecular complex in cerebral arteries.

Cav-1 and TRPC3 Are Spatially Localized in Arterial Smooth
Muscle Cells-Spatial localization of TRPC3 channels to cav-1 in arterial smooth muscle cells was studied using immunoFRET. Fig. 8A indicates that cav-1 and TRPC3 are co-localized in the arterial smooth muscle cell plasma membrane. Alexa 488-and Alexa 546-conjugated secondary antibodies bound to cav-1 and TRPC3 channel primary antibodies, respectively, generated mean N-FRET of 23.0 Ϯ 1.2% (n ϭ 16). These data indicate that cav-1 and TRPC3 are spatially localized in the cerebral artery smooth muscle cell plasma membrane.

IP 3 R1 and TRPC3 Are in Located in Close Spatial Proximity and Separated by Caveolae Disruption or CSD Peptide in Arterial Smooth Muscle
Cells-ImmunoFRET was also used to measure the spatial proximity of IP 3 R1 and TRPC3 channels   FEBRUARY 11, 2011 • VOLUME 286 • NUMBER 6 in arterial smooth muscle cells. Alexa 488-and Alexa 546labeled secondary antibodies bound to TRPC3 channel and IP 3 R1 primary antibodies, respectively, generated mean N-FRET of ϳ19% (Fig. 8, B and C). Fig. 8B illustrates that im-munoFRET was localized both to the plasma membrane and intracellularly. M␤CD treatment and ANT-CSD peptide reduced mean N-FRET to ϳ12 and 10%, respectively (Fig. 8C). The M␤CD-induced reduction in N-FRET was reversed by subsequent exposure of M␤CD-treated cells to Chol/M␤CD (Fig. 8C). These data indicate that caveolae and the cav-1 CSD maintain IP 3 R1 and TRPC3 in close spatial proximity in arterial smooth muscle cells.

DISCUSSION
Data in this study indicate that cav-1 maintains IP 3 R1 and TRPC3 channels in a signaling complex that enables IP 3 -induced I Cat activation and vasoconstriction in cerebral artery smooth muscle cells. We show that: 1) caveolae disassembly reduces myogenic tone and attenuates IP 3 -induced I Cat activation and vasoconstriction, 2) cav-1 knockdown does not alter IP 3 R1 and TRPC3 channel expression but reduces myogenic tone and IP 3 -induced vasoconstriction, 3) the N and C termini of IP 3 R1 and the C terminus of TRPC3 contain CSDinteracting motifs, 4) a CSD peptide inhibits IP 3 -induced I Cat activation and vasoconstriction, 5) cav-1, IP 3 R1 and TRPC3 are located within the same macromolecular complex in cerebral arteries, and 6) the close spatial proximity of IP 3 R1 and  . Cav-1 antibody co-immunoprecipitates cerebral artery cav-1 (ϳ22 kDa), IP 3 R1 (ϳ270 kDa), and TRPC3 (ϳ90 kDa). Lysate supernatant (ϳ40 g of protein) was used as the input control. Negative control (-ve ctrl.) received the same concentration of cav-1 antibody except that the coupling resin was replaced with control agarose resin that is not aminereactive.

Caveolin-1 Assembles IP 3 R1-TRPC3 Signaling Complex
TRPC3 is disrupted by caveolae disassembly and a CSD peptide. In summary, data indicate that that cav-1 and caveolae spatially localize IP 3 R1 and TRPC3 channels in cerebral artery smooth muscle cells, enabling IP 3 -induced physical coupling, leading to I Cat generation and vasoconstriction.
Three IP 3 R isoforms (1-3) have been cloned (27,28). IP 3 R1 is the primary molecular and functional isoform expressed in arterial smooth muscle cells (6). In cerebral artery smooth muscle cells, plasma membrane TRPC3 channels are located in close spatial proximity to IP 3 R1 (4). IP 3 activates plasma membrane TRPC3 channels by inducing a direct interaction between the IP 3 R1 N terminus and TRPC3 channel calmodulin and IP 3 R binding (CIRB) domains (4). IP 3 -induced TRPC3 channel activation causes Na ϩ influx, resulting in membrane depolarization, voltage-dependent Ca 2ϩ channel activation, an intracellular Ca 2ϩ ([Ca 2ϩ ] i ) elevation, and vasoconstriction. Therefore, IP 3 R1 and TRPC3 channels directly couple in arterial smooth muscle cells (4,5).
Several TRP channels, including TRPC3, localize within caveolae in mammalian cells (29). Such localization may enhance TRP channel interaction with other cell signaling molecules (29). However, whether caveolae localize TRPC3 channels nearby other regulatory proteins in arterial smooth muscle cells was unclear. Our data indicate that cav-1 and TRPC3 are spatially localized in the arterial smooth muscle cell plasma membrane, suggesting that TRPC3 may be localized in caveolae. In rat tail arteries and pulmonary artery smooth muscle cells, cholesterol depletion by M␤CD reduced caveolae density and altered plasma membrane morphology, and these effects were reversed by cholesterol replenishment (15,23). To examine a role for caveolae and cav-1 in enabling functional IP 3 R1 to TRPC3 channel physical coupling in arterial smooth muscle cells, we used M␤CD at concentrations previously utilized to effectively disrupt plasma membrane caveolae (15,23,30,31). M␤CD reduced smooth muscle cell membrane capacitance and this was reversed by subsequent cholesterol treatment. Reduced cell capacitance has also been reported in arterial smooth muscle cells lacking cav-1, and in M␤CD-treated uterine and urinary bladder smooth muscle cells (13,32,33). These studies and ours indicate that acute and chronic loss of caveolae reduces smooth muscle cell surface area. In another previous study, M␤CD-mediated caveolae disruption did not alter arterial smooth muscle cell membrane capacitance (34). Reasons for these contradictory observations are unclear. M␤CD modulates depolarization of skinned skeletal muscle fibers (35). However, M␤CD treatment did not alter membrane potential or basal intracellular Ca 2ϩ concentration in cerebral artery smooth muscle cells (34). These studies suggest that the effect of M␤CD on membrane potential may be cell type-dependent. Here, M␤CD and cav-1 knockdown attenuated myogenic constriction in cerebral arteries, consistent with observations in cerebral and mesenteric arteries from cav-1 knock-out mice and M␤CDtreated rat mesenteric arteries (16,17,30). It is unclear why M␤CD and cav-1 knockdown reduced myogenic tone, whereas the CSD peptide did not alter myogenic tone. One explanation may be that cav-1 regulates myogenic tone via a CSD-independent mechanism. Alternatively, the CSD peptide used may more effectively block physical coupling of IP 3 R1 to TRPC3 channels than it inhibits CSD-dependent mechanisms that lead to cerebral artery myogenic constriction.
We demonstrate for the first time that caveolae disassembly and a CSD peptide reduces IP 3 -induced I Cat activation in smooth muscle cells and vasoconstriction in pressurized arteries. Co-IP data demonstrated that cav-1, IP 3 R1, and TRPC3 channels are contained within the same macromolecular complex in cerebral arteries. We also show that both IP 3 R1 and TRPC3 channels contain CSD binding motifs. The CSD peptide did not alter smooth muscle cell capacitance, suggesting that caveolae integrity and thus, cell surface area remained unchanged. However, the CSD peptide attenuated IP 3 -induced I Cat activation in smooth muscle cells and vasoconstriction in pressurized arteries. We tested the hypothesis that caveolae/cav-1 spatially localize plasma membrane TRPC3 channels nearby SR membrane IP 3 R1 channels, thereby permitting physical coupling of these proteins. The Förster distance between Alexa 488 and 546 is 6.4 nm (Invitrogen). Given that secondary antibodies bound to IP 3 R1 (Alexa 546) and TRPC3 (Alexa 488) generated significant N-FRET, these data indicate that IP 3 R1 and TRPC3 channels are located in very close spatial proximity. Our data indicate that FRET was localized both at the plasma membrane and within the cytosol. Intracellular protein complexes containing both IP 3 Rs and TRPC3 have previously been reported in HEK-293 cells (36). We have also previously shown that a proportion of TRPC3 protein is intracellular in rat cerebral arteries and that TRPC3 immunofluorescence exhibits both plasma membrane and intracellular localization in arterial smooth muscle cells (4). Conceivably, intracellular TRPC3 channels may be located on organelle membranes that are nearby SR IP 3 R1 channels or stored within the SR prior to trafficking and in close proximity to SR IP 3 R1. Caveolae disruption and the CSD peptide both reduced the IP 3 R1/TRPC3 N-FRET signal, indicating that these proteins became spatially separated. These data indicate that cav-1, IP 3 R1 and TRPC3 channels form a macromolecular complex. Additional data indicate that this complex enables physically coupling of SR IP 3 R1 to plasma membrane TRPC3 channels thereby allowing direct communication between intracellular and plasma membrane proteins. M␤CD and CSD peptide reduced IP 3 -induced I Cat and M␤CD, CSD peptide, and cav-1 knockdown attenuated IP 3induced vasoconstriction. Therefore, cav-1 and caveolae enable IP 3 R1 to physically couple to TRPC3 channels in smooth muscle cells to control arterial diameter.
Recombinant cav-1 interacts with both endogenous IP 3 R3 and recombinant TRPC1 in HEK-293 cells (37). Expression of a CSD-deleted cav-1 mutant in HEK-293 cells and a human dermal microvascular endothelial cell line increased storeoperated Ca 2ϩ entry (SOCE) and attenuated a thapsigargininduced IP 3 R3-TRPC1 interaction in HEK-293 cells (37). These studies suggested that through its scaffolding domain, cav-1 regulates an IP 3 R3-TRPC1 complex and SOCE in cultured HEK-293 and endothelial cells (37,38). Store-operated cation currents have been described in a variety of vascular smooth muscle cells, including aorta, mesenteric artery, and portal vein (39). In contrast, SR Ca 2ϩ depletion does not acti- FEBRUARY 11, 2011 • VOLUME 286 • NUMBER 6 vate a I Cat or alter IP 3 -induced I Cat activation in cerebral artery smooth muscle cells or modify IP 3 -and agonist-induced cerebral artery constriction at low intravascular pressure (5,6). SR Ca 2ϩ depletion also did not induce contraction in cerebral arteries at low or no pressure (5,40,41). Collectively, these studies indicate that SOCE does not appear to exist in contractile cerebral artery smooth muscle cells.

Caveolin-1 Assembles IP 3 R1-TRPC3 Signaling Complex
In conclusion, we demonstrate that cav-1 maintains IP 3 R1 and TRPC3 channels in a functional macromolecular complex in cerebral artery smooth muscle cells. Cav-1 enables IP 3 to stimulate physical coupling between IP 3 R1 and TRPC3 channels, leading to I Cat activation and vasoconstriction.