Myosin Phosphatase-Rho Interacting Protein

Regulation of vascular smooth muscle cell contractile state is critical for the maintenance of blood vessel tone. Abnormal vascular smooth muscle cell contractility plays an important role in the pathogenesis of hypertension, blood vessel spasm, and atherosclerosis. Myosin phosphatase, the key enzyme controlling myosin light chain dephosphorylation, regulates smooth muscle cell contraction. Vasoconstrictor and vasodilator pathways inhibit and activate myosin phosphatase, respectively. G-protein-coupled receptor agonists can inhibit myosin phosphatase and cause smooth muscle cell contraction by activating RhoA/Rho kinase, whereas NO/cGMP can activate myosin phosphatase and cause smooth muscle cell relaxation by activation of cGMP-dependent protein kinase. We have used yeast two-hybrid screening to identify a 116-kDa human protein that interacts with both myosin phosphatase and RhoA. This myosin phosphatase-RhoA interacting protein, or M-RIP, is highly homologous to murine p116RIP3, is expressed in vascular smooth muscle, and is localized to actin myofilaments. M-RIP binds directly to the myosin binding subunit of myosin phosphatase in vivo in vascular smooth muscle cells by an interaction between coiled-coil and leucine zipper domains in the two proteins. An adjacent domain of M-RIP directly binds RhoA in a nucleotide-independent manner. M-RIP copurifies with RhoA and Rho kinase, colocalizes on actin stress fibers with RhoA and MBS, and is associated with Rho kinase activity in vascular smooth muscle cells. M-RIP can assemble a complex containing both RhoA and MBS, suggesting that M-RIP may play a role in myosin phosphatase regulation by RhoA.

Blood vessel tone is regulated by the contractile state of vascular smooth muscle cells in the blood vessel wall. Diseases characterized by abnormal vascular smooth muscle cell contraction include hypertension, blood vessel spasm, and atherosclerosis (1)(2)(3)(4)(5). Smooth muscle contraction is tightly coupled to myosin light chain phosphorylation (6), which in turn is regulated by the relative activities of myosin light chain kinase and myosin phosphatase. Myosin light chain kinase is activated by intracellular calcium and phosphorylates myosin light chains, leading to cell contraction (7,8). Myosin phosphatase dephosphorylates myosin light chains, leading to smooth muscle cell relaxation (9). Myosin phosphatase activity, once thought to be constitutive, is now known to be highly regulated. Both vasoconstrictor signaling pathways, which lead to inhibition of the phosphatase and cell contraction (reviewed in Ref. 10), and vasodilator signaling pathways, which lead to cell relaxation via activation of myosin phosphatase have been recently defined (11)(12)(13)(14)(15).
Myosin phosphatase is a heterotrimer consisting of a PP1 catalytic subunit, a 130-kDa myosin binding subunit (MBS) 1 and a 20-kDa subunit of unknown function (9, 16 -18). The MBS is a regulatory subunit that targets PP1 to its substrate, myosin light chain (9), and has multiple protein interaction domains, including ankyrin repeats at its amino terminus, and a leucine zipper domain at its carboxyl terminus. MBS binds PP1 and myosin light chain at its amino terminus and the M20 subunit and cGMP-dependent protein kinase 1␣ (cGK) at its carboxyl terminus (Ref. 11, reviewed in Ref. 19). The MBS-cGK interaction is necessary for NO/cGMP-mediated activation of myosin phosphatase (11).
In vascular smooth muscle, G-protein-coupled receptor agonists cause contraction in part by inhibition of myosin phosphatase activity (20). Several downstream signaling pathways that inhibit myosin phosphatase activity have been discovered recently, including RhoA/Rho kinase (21), protein kinase C activation of the inhibitory phosphoprotein CPI-17 (22), and arachidonic acid (23,24). In addition, several kinases copurify with myosin phosphatase, including ZIP-like kinase (25), integrin-linked kinase (26), myotonic dystrophy-related kinase (27), and Raf-1 (28), each of which can phosphorylate MBS and inhibit myosin phosphatase activity. RhoA/Rho kinase has been the most extensively studied myosin phosphatase inhibitor. RhoA binds to a myosin phosphatase complex in vitro, and GTP-bound RhoA, in combination with its downstream effector Rho kinase, inhibits myosin phosphatase activity (21). Specific blockade of Rho kinase has been found to ameliorate hypertension in several rat models (29), as well as to prevent the response to vascular injury and blood vessel spasm in animal models (30 -33). Furthermore, phosphorylation-specific antibodies against inhibitory sites on MBS demonstrate that phosphorylation correlates directly with contractile agonist-mediated myosin phosphatase inhibition (34,35). Recently, a Rho kinase inhibitor has been found to be effective in preventing * This work was supported by Grants HL03987 (to H. K. S.) and HL55309 (to M. E. M.) from the National Institutes of Health and by Grant-in-aid 0050310N (to H. K. S.) from the American Heart Association. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18  coronary artery spasm and treating myocardial ischemia in humans (36,37).
Despite strong evidence for RhoA/Rho kinase-mediated inhibition of myosin phosphatase, the molecular mechanism for this contractile pathway is not well understood. Activated RhoA and Rho kinase translocate to the cell membrane (38,39), and have also been found colocalized with actin myofilaments (40,41). The mechanism whereby RhoA and Rho kinase are targeted to and inhibit myosin phosphatase and thereby prevent dephosphorylation of myosin light chains in contractile myofilaments remains unclear.
We hypothesized that other signaling proteins regulate RhoA/Rho kinase-mediated inhibition of myosin phosphatase and searched for proteins that interact with both myosin phosphatase and RhoA using yeast two-hybrid screening methods. Here we report the identification and initial characterization of a protein that binds both MBS and RhoA. This myosin phosphatase-RhoA interacting protein (M-RIP) is a potential molecular link between RhoA signaling and myosin phosphatase regulation.

EXPERIMENTAL PROCEDURES
Materials-Vectors pGBT9 and pGAD424, yeast strain Y190, and human aorta Matchmaker cDNA library were from Clontech. Vector pCMV tag was from Stratagene. All enzymes were from New England Biolabs. Y27632 was from Tocris. pGEX vectors were from Pharmacia, pQE vectors from Qiagen, and pCI mammalian expression vector was from Promega. The TA cloning system was from Invitrogen. Antibodies were obtained as follows: anti-MYPT1 from Covance, anti-RhoA, anti-GST, and anti-PP1 from Santa Cruz Biotechnology, M2 antibody from Sigma, anti-cGMP-dependent protein kinase from Stressgen, and anti-ROK␣ from Transduction Labs.
Cell Culture-All cells were maintained in Dulbecco's modified Eagle's medium with 10% fetal bovine serum. COS-7 cells were from the American Type Culture Collection. Human aortic and coronary smooth muscle cells were cultured by the explant method. Ao184 and Co399 cells were immortalized using adenovirus expressing E6 and E7 and selection for G418 resistance. Transfections were performed by electroporation.
Yeast Two-hybrid Screening-Human MBS-Cterm, base pairs 2043-3090 (kind gift of Dr. Masaaki Ito), was cloned into the EcoRI site of yeast two-hybrid vectors pGBT9 and pGAD424, expressing MBS fusion proteins with Gal4 DNA binding and DNA activating domains, respectively. Expression of MBS-Cterm in yeast was confirmed by dimerization when MBS-Cterm-pGBT9 and MBS-Cterm-pGAD424 were coexpressed in yeast (data not shown). Yeast strain Y190 was sequentially transformed with MBS-Cterm-pGBT9 and a human aorta cDNA library (Clontech) using the lithium acetate method. Interactions were identified by growth on selective media, and confirmed by colony lift filter ␤-galactosidase assay. The cDNA library plasmid was extracted from confirmed positive yeast colonies, and sequenced at the Tufts Core Sequencing Facility.
Cloning of Full-length M-RIP-Primers were designed to amplify the 5Ј sequence of M-RIP. The 5Ј primer was based on a human EST that is homologous to the 5Ј sequence of mouse and rat p116RIP3, gb AI 678749 (5Ј-ACCATGTCGGCAGCCAAGGAGAACCCGTGC-3Ј). The 3Ј primer was based on a sequence within yeast two-hybrid clone 6 (5Ј-CCGCTCCTGAGCCAGGGCCTGCTGGATGGG-3Ј). The 5Ј M-RIP sequence was amplified from the human aorta cDNA library. The PCR product was ligated into pCRII using the TA cloning system and sequenced. Full-length M-RIP was generated by overlap extension PCR using 5Ј M-RIP (bp 1-1686) and yeast two-hybrid clone 6 (bp 1599 -3076) alone for the first 5 cycles, then adding M-RIP 5Ј and 3Ј primers (5Ј-ACCATGTCGGCAGCCAAGGAGAACCCGTGC-3Ј and 5Ј-ATT-TCAGGTATCCCACGAGACCTGCTCAAT-3Ј) for 25 cycles. The PCR product was ligated into pCRII using the TA cloning system. Fulllength M-RIP was sequenced in both directions (GenBank TM accession number AY296247).
DNA Constructs-Full-length M-RIP with an amino-terminal FLAG tag was constructed as follows. Amino-terminal M-RIP was amplified from full-length M-RIP using primers that incorporated a 5Ј FLAG sequence (5Ј-GATGAATTCCGACCATGGACTACAAGGACGACGATG-ACAAGTCGGCAGCCAAGGAGAACCCGTGCAGG-3Ј and 5Ј-CCGCT-CCTGAGCCAGGGCCTGCTGGATGGG-3Ј) and the PCR product that included the amino-terminal half of M-RIP was cloned into mammalian expression vector pCI. All but the FLAG-tagged amino terminus of M-RIP was replaced by adding the AgeI/NotI fragment from wild-type M-RIP to generate full-length FLAG-tagged M-RIP. M-RIP coiled-coil domains were amplified from full-length M-RIP using the following primers: CC1, 5Ј-TGCGGTCGACCTCGCACGTGGCCTGCAGCACGT-AGCC-3Ј and 5Ј-TGCGGTCGACCTCCCGGCCCAGGGCCACCCTCAG-CTG-3Ј; CC2, 5Ј-CCGGAATTCCGAGGGTTTGCAGCAATGGAAGAAA-CG-3Ј and 5Ј-TGCGGTCGACAGTCAGCAGCGTCCGCAACCGTGTGA-T-3Ј; CC3, 5Ј-CCGGAATTCGCCTATGAACTAGAGGTCTTATTGCG-G-3Ј and 5Ј-TGCGGTCGACGGGGACTTCTCCCCCAGTGCTTCCGTT-3Ј. PCR products were cloned into pGEX and pQE vectors for expression as GST and polyhistidine fusion proteins, respectively. M-RIP-(545-823) and M-RIP-(545-878) were amplified from full-length M-RIP by PCR using the following primers: M-RIP-(545-823), 5Ј-CGCGGATCC-GCTGAGTTCCGTCCCATCCAGCAG and 3Ј-GGCGAATTCCCGAGA-GGACCACCAGTTCCCGC. M-RIP-(545-878) was amplified using the 5Ј M-RIP-(545-823) primer and the 3Ј CC2 primer. Both PCR products were cloned into pGEX and pCMV-Tag2B (Clontech) vectors for expression of GST fusion proteins in bacteria and FLAG-tagged proteins in mammalian cells, respectively. Wild-type FLAG-RhoA cDNA and GST-Rhotekin RBD were kind gifts of Dr. Naoki Mochizuki. RhoA was cloned into pQE for expression of polyhistidine-tagged RhoA in bacteria. Rhotekin RBD includes amino acids 1-99 of mouse Rhotekin (42), amplified by PCR from a mouse cDNA library, and ligated into pGEX4T-3. GST-MBS LZ (amino acids 847-1030) was made as described (11). GST-MBS LZ mutant was made as described (43). GFP-MBS was prepared by amplifying full-length human MBS from the Clontech human aorta library, followed by ligation into pEGFP. GFP-MBS LZ mutant was prepared by replacing the COOH-terminal domain of GFP-MBS with the COOH-terminal domain of the GST-MBS LZ mutant. All DNA constructs were fully sequenced in both directions.
Production of Anti-M-RIP Antisera-cDNA encoding the 5Ј 480 bp of M-RIP was amplified from full-length M-RIP using PCR primers 5Ј-ATCGAATTCATGTCGGCAGCCAAGGAGAAC-3Ј and 5Ј-CGATCCTC-GAGCTTCTTCTGATTCTGCTTGTT-3Ј. The PCR product was ligated into pGEX and sequenced. The GST fusion protein of the NH 2 -terminal M-RIP (RIP-(1-160)) was expressed in and purified from bacteria as described (11). GST-RIP-(1-160) was eluted from glutathione-agarose beads and specific rabbit antiserum was produced at Alpha Diagnostic International, San Antonio TX. Crude antisera were purified on protein A beads as described (44).
Preparation of Fusion Proteins-GST and His 6 fusion proteins were grown in bacteria overnight at 37°C in LB with 150 g/ml ampicillin. The culture was diluted 10-fold with LB/Amp, and incubated an additional 1 h. Isopropyl-1-thio-␤-D-galactopyranoside was added to a final concentration of 0.1 mM for GST fusion proteins and 1 mM for His 6 fusion proteins, and the cells were incubated 4 h, pelleted, and frozen. For GST fusion proteins, the cell pellet was thawed and resuspended in 35 ml of 20 mM Tris, pH 8, 100 mM NaCl, 1.5 mM EDTA, 0.1% Sarkosyl, 0.25 mg/ml lysozyme, 2 mM PMSF, 10 mM benzamidine, 20 mM dithiothreitol, and 0.01 g/ml each of aprotinin, leupeptin, and pepstatin A. The cell lysate was incubated on ice for 15 min, then EDTA and Sarkosyl were added to final concentrations of 5 mM and 1.4%, respectively. The lysate was sonicated, centrifuged, and to the supernatant was added 20 ml of 10% Triton X-100 and glutathione-agarose beads. This was incubated for 2 h at 4°C, following which the beads were washed with cold PBS. GST fusion proteins were either stored at 4°C bound to glutathione-agarose beads, or eluted from beads in 50 mM Tris, pH 8.0, and 15 mM reduced glutathione, snap frozen, and stored at Ϫ80°C. For His 6 fusion proteins, the bacterial pellet was resuspended in lysis buffer consisting of 50 mM NaH 2 PO 4 , pH 8.0, 300 mM NaCl, 10 mM imidazole, 0.01 g/ml each of aprotinin, leupeptin, and pepstatin A, 2 mM PMSF. Lysozyme was added to a final concentration of 1 mg/ml and incubated for 30 min on ice. The lysate was then sonicated and centrifuged at 12,500 rpm for 20 min. The supernatant was incubated with 0.5 ml of Ni-NTA beads (Qiagen) for 2 h, then washed with lysis buffer. His 6 fusion proteins were then stored at 4°C bound to Ni-NTA beads, or eluted from the beads in 50 mM Tris, pH 8.0, 150 mM NaCl, 5 mM MgCl 2 , 1 mM dithiothreitol, and 250 mM imidazole, snap frozen, and stored at Ϫ80°C.
Purification and Loading of His 6 RhoA-The procedure was based on that of Diekmann and Hall (45). Briefly, His 6 RhoA was purified as described above for His 6 fusion proteins. Eluted His 6 RhoA was dialyzed in 10 mM Tris, pH 7.5, 150 mM NaCl, 2 mM MgCl 2 , and 0.1 mM dithiothreitol, then snap frozen. Two g of purified His 6 RhoA was loaded in 50 mM Tris, pH 7.5, 5 mM EDTA, and 0.5 mg/ml bovine serum albumin with either 200 M GTP␥S or GDP␤S for 15 min at room temperature, followed by the addition of MgCl 2 to a final concentration of 60 mM.
Solubility Assay-Subconfluent cultured human aortic smooth muscle cells were rinsed twice with cold PBS and lysed in 50 mM Tris, pH 7.5, 1 mM EDTA, 0.5% Nonidet P-40, 2 mM PMSF, 0.01 g/ml each of aprotinin, leupeptin, and pepstatin A, and 250 or 350 mM NaCl. Lysates were incubated 1 h at room temperature, then centrifuged at 14,000 rpm for 20 min. The supernatant fraction was precleared with protein A beads, then used for immunoprecipitation with polyclonal anti-M-RIP antibody as described below. An equivalent percentage of each fraction was analyzed by anti-M-RIP immunoblot. Co-immunoprecipitation Assays-Confluent cultured human aortic smooth muscle cells were rinsed in PBS, then lysed in buffer A (50 mM Tris, pH 7.6, 7 mM MgCl 2 , 2 mM EDTA, 2 mg/ml n-dodecyl-B-maltoside, 0.4 mg/ml cholesteryl hemisuccinate, 0.6 M NaCl, 10 mM sodium molybdate, 2 mM PMSF, and 0.01 g/ml each of aprotinin, leupeptin, and pepstatin A) for MBS immunoprecipitations and in buffer B (40 mM Tris, pH 7.5, 0.275 M NaCl, 4 mM EDTA, 2% Triton X-100, 20% glycerol, 50 mM ␤-glycerol phosphate, 2 mM PMSF, and 0.01 g/ml each of aprotinin, leupeptin, and pepstatin A) for M-RIP and cGMP-dependent protein kinase immunoprecipitations. Cell lysates were incubated 1 h at room temperature, then centrifuged at 14,000 rpm for 20 min at 4°C. The supernatant was precleared with protein A beads and incubated overnight with polyclonal anti-M-RIP, anti-MBS, or anti-cGMPdependent protein kinase 1. Protein A beads were added, and the lysates were incubated 2 h. The beads were washed three times with buffer C (50 mM Tris, pH 7.6, 7 mM MgCl 2 , 2 mM EDTA), then proteins were eluted in SDS sample buffer and analyzed by SDS-PAGE and immunoblotting as above. For co-immunoprecipitation of MBS with FLAG-M-RIP domains, COS-1 cells were transfected by electroporation, and lysates were prepared with buffer A as above after 48 h. Immunoprecipitations were performed as above except that M2 antibody was used for immunoprecipitation, and protein G beads were used to collect antigen-antibody complexes.
Fusion Protein Interaction Assays-For fusion protein interactions with proteins from cell lysates, confluent cells were rinsed twice with ice-cold PBS, lysed in buffer A, incubated for 1 h at room temperature, then centrifuged at 14,000 rpm for 20 min at 4°C. The supernatant was mixed with GST fusion proteins prebound to glutathione-agarose beads. After overnight incubation at 4°C, the beads were washed with buffer C, eluted with SDS sample buffer, and bound proteins were analyzed by immunoblotting. For fusion protein interactions with purified proteins, fusion proteins immobilized on beads were incubated in buffer A for 1 h with purified soluble protein previously eluted from beads. After incubation, the beads were washed three times with buffer C, bound proteins were eluted in SDS sample buffer and analyzed by immunoblotting with anti-GST antibodies.
Stress Fiber Preparation-This procedure was adapted from Katoh et al. (40). Human aortic smooth muscle cells were grown on two 100-mm dishes to near confluence, washed with cold PBS, then extracted with 10 ml of triethanolamine extraction buffer (2.5 mM triethanolamine, 1 g/ml leupeptin and pepstatin A, 20 g/ml aprotinin) for 30 min, shaking, with replacement of extraction buffer every 2-3 min. Remaining cell components were then further extracted using 10 ml of Triton buffer (0.5% Triton X-100, 1 g/ml leupeptin and pepstatin A, 20 g/ml aprotinin in PBS) for Triton extractions or 10 ml of glycerol buffer (50% glycerol, 1 g/ml each of leupeptin and pepstatin A, 20 g/ml aprotinin in PBS) for glycerol extractions for 5 min while shaking, with replacement of extraction buffer twice. Triton or glycerol was then removed by washing with 10 ml of aprotinin/PBS (20 g/ml aprotinin, 1 g/ml leupeptin and pepstatin A in PBS) for 10 min while shaking with one replacement of wash buffer. Remaining insoluble material was then scraped in aprotinin/PBS, and homogenized with a Z-shaped 21-gauge needle. The insoluble debris was pelleted at 1,000 ϫ g for 5 min, and stress fibers were isolated by centrifugation of the supernatant at 100,000 ϫ g for 1 h. The stress fiber pellet was boiled in 0.1 ml of protein sample buffer and subjected to SDS-PAGE and immunoblotting with the indicated antibodies. For immunostaining of purified stress fibers, human coronary artery smooth muscle cells were grown on coverslips and glycerol-extracted as described above. Stress fibers were then fixed and immunostained as described below. All antibody dilutions were 1:100.
Kinase Assay-Subconfluent human aortic smooth muscle cells were lysed and M-RIP was immunoprecipitated as described for the solubility assay above, with 350 mM NaCl. The M-RIP and nonimmune IPs were washed with Rho kinase assay buffer (adapted from Feng et al. Immunofluorescence Staining-Human coronary artery smooth muscle cells were plated on coverslips in Dulbecco's modified Eagle's medium with 10% fetal bovine serum. The coverslips were washed with PBS twice, then fixed in 3.7% paraformaldehyde. The cells were permeabilized with 0.3% Triton X-100 and 10% donkey serum, then Miscellaneous-Amino acid alignments were performed using the BLAST 2 sequences program from the National Center for Biotechnology Information. Protein structure prediction was made using the Simple Modular Architecture Research Tool, 2 phosphorylation sites were predicted using Omiga (Oxford Molecular) and nuclear localization signals were predicted using PredictNLS Online. 3

RESULTS
Identification and Cloning of M-RIP-To identify proteins involved in regulation of myosin phosphatase, base pairs 2043-3090 of human MBS (MBS-Cterm) were used as bait in a yeast two-hybrid screen of a human aorta library. Two clones, 6 and 11, encoded 3Ј regions of a cDNA with high homology to a murine RhoA-interacting protein, p116 RIP3 (47). When retransformed into yeast, both clones interacted with MBS-Cterm but not with Gal4 DNA-binding domain alone (data not shown). Clone 6 was homologous to bp 1617-3072 and clone 11 was homologous to bp 2109 -3072 of murine p116 RIP3 .
A human aorta library was next probed for the 5Ј sequence of human p116 RIP3 . PCR using the human aorta library as template yielded a 1,686-bp product that was highly homologous to the 5Ј sequence of murine p116 RIP3 . The full-length human clone was constructed by overlap extension PCR using yeast two-hybrid clone 6 and the 5Ј 1,686-bp sequence (Fig. 1A). The full-length human sequence is 85% identical at the nucleotide level and 90% identical at the amino acid level to murine and rat p116 RIP3 . The RhoA-binding domain of this human clone is 88% identical at the amino acid level to the similar domain on murine p116 RIP3 . This human cDNA is hereafter called M-RIP.
Analysis of the M-RIP cDNA predicts a protein of 1,024 amino acids with multiple protein interaction domains, including a pair of pleckstrin homology domains flanking two polyproline motifs, and three carboxyl-terminal coiled-coil domains (Fig. 1B). Sites for myristylation as well as for phosphorylation by protein kinase C, cyclic nucleotide-dependent protein kinases, and tyrosine kinases are also present in the M-RIP protein.

Detection and Localization of M-RIP in Vascular Smooth
Muscle Cells-Specific M-RIP antisera were raised and tested first by immunoblotting of lysates from COS-7 cells transfected with full-length M-RIP cDNA. Anti-M-RIP recognized a specific band of 125 kDa in untransfected COS-7 cells that was augmented by overexpression of full-length M-RIP ( Fig. 2A). A parallel immunoblot of the same lysates in which anti-M-RIP was preabsorbed with immunogen failed to identify any M-RIP band (data not shown). Anti-M-RIP was next used to probe lysates from two different human arterial smooth muscle cell lines. Anti-M-RIP identified a specific 125-kDa band in these lysates, supporting that M-RIP is expressed in human vascular smooth muscle cells (Fig. 2B). M-RIP from vascular smooth muscle cells was completely insoluble at 250 mM NaCl, but could be solubilized and immunoprecipitated under high ionic strength conditions (Fig. 2C).
Cultured human coronary artery smooth muscle cells were immunostained with anti-M-RIP (Fig. 2D, left panel) and with phalloidin (Fig. 2D, middle panel) to label actin fibers. M-RIP localized primarily in a filamentous pattern in the cytoplasm, similar to the distribution of actin filaments. Overlay of the two images revealed that M-RIP colocalized with actin myofila-ments (Fig. 2D, right panel). In control experiments, M-RIP antibody preincubated with M-RIP immunogen failed to label VSMCs (data not shown). Despite the presence of a putative nuclear localization signal between residues 151 and 156, M-RIP was not detected in the nucleus of vascular smooth muscle cells (data not shown).
In Vivo Interaction between M-RIP and Myosin Phosphatase-To explore whether the interaction between M-RIP and MBS detected in yeast occurs in intact vascular smooth muscle cells, immunoprecipitation experiments were performed with human aortic smooth muscle cell lysates. Immunoprecipitation of either M-RIP or MBS led to recovery of both proteins (Fig.  3A), supporting an in vivo interaction between M-RIP and MBS in vascular smooth muscle cells. Furthermore, M-RIP immunoprecipitation also led to recovery of PP1, and cGMP-dependent protein kinase (Fig. 3, B and C), supporting that M-RIP is associated with the myosin phosphatase complex in vivo. In GST pull-down interaction studies, binding between M-RIP and either PP1 or cGMP-dependent protein kinase could not be detected, suggesting that M-RIP interacts with these proteins indirectly via their known interactions with MBS (data not shown).

Characterization of Binding between M-RIP and Myosin
Phosphatase-The M-RIP-MBS interaction was explored further using GST and His fusion protein interaction studies. GST fusions of various M-RIP domains were incubated with vascular smooth muscle cell lysates and recovered MBS was detected by immunoblotting. The individual coiled-coil domains of M-RIP were tested for binding to MBS (Fig. 1C). The second coiled-coil domain (CC2) of M-RIP bound MBS strongly, whereas the CC1 and CC3 domains had little or no interaction with MBS (Fig. 4A). The CC2 domain includes the carboxylterminal end of the putative Rho-binding domain for murine p116 RIP3 (47) (Fig. 4B). These data indicate that MBS binds a domain of M-RIP distinct from and 3Ј to the Rho-binding domain (Fig. 4C).
The carboxyl terminus of MBS contains a leucine zipper domain that mediates binding to other coiled-coil domains, including the leucine/isoleucine zipper domain of cGMP-dependent protein kinase 1␣ (11,12,43). To determine whether this domain of MBS mediates binding to the CC2 of M-RIP, wild-type COOH-terminal MBS (MBS LZ ) and mutant COOHterminal MBS in which all leucines in the leucine zipper domain were mutated to alanine (MBS LZ mutant), were expressed as GST fusion proteins and tested for binding to full-length M-RIP. Wild-type GST-MBS LZ interacted with M-RIP, whereas GST-MBS LZ mutant bound no appreciable M-RIP from the same lysates (Fig. 4D). Full-length GFP-MBS or full-length GFP-MBS LZ mutant also was studied in COS-1 cells known to express endogenous MBS. 4  GST-M-RIP-(545-878) bound HisRhoA, without preference for GDP or GTP binding state of the low molecular weight G-protein (Fig. 5B).
M-RIP Assembles a Complex including the Myosin Binding Subunit and RhoA-The data above support that RhoA binds an M-RIP domain that includes amino acids 545-823 (Fig. 5B), whereas MBS binds a domain that includes amino acids 823-878 (Fig. 4, B and C). An M-RIP construct containing both domains was expressed in COS-1 cells and used to test the hypothesis that this region of M-RIP assembles a ternary complex of M-RIP with MBS and RhoA (Fig. 5C)   for the binding of RhoA to M-RIP (Fig. 5C).
M-RIP Is Associated with RhoA/Rho Kinase in Vivo-RhoA and Rho kinase have been shown to colocalize with actin stress fibers, and are present in stress fiber preparations made by glycerol extraction, but are lost from stress fibers extracted with Triton X-100 (40,41). When Triton-and glycerol-extracted stress fibers from vascular smooth muscle cells were examined by immunoblot, both contained actin, MBS, and M-RIP, whereas glycerol-extracted stress fibers, but not Triton-extracted stress fibers, also contained RhoA and Rho kinase (Fig.  6A). Immunostaining of glycerol-extracted vascular smooth muscle cells revealed colocalization of M-RIP, MBS, and RhoA with actin stress fibers (Fig. 6B). Because of the requirement of detergent and high ionic strength conditions for M-RIP solubilization, RhoA could not be detected in the M-RIP immunopellet by immunoblot. However, the M-RIP immunopellet contained kinase activity (Fig. 6C) that was inhibited 45% (p ϭ 0.005, n ϭ 3) by the Rho kinase inhibitor Y27632 (Fig. 6D) supporting an association between M-RIP and RhoA/Rho kinase in vivo.

DISCUSSION
Myosin phosphatase activity is regulated by both contractile agonists and nitrovasodilators. RhoA and its downstream effector Rho kinase mediate contractile agonist-induced myosin phosphatase inhibition, but the mechanisms whereby RhoA interacts with myosin phosphatase remain unclear. Activated RhoA and Rho kinase translocate to the cell membrane, and a subpopulation of RhoA/Rho kinase has been identified on actin stress fibers (40,41). The mechanism whereby RhoA/Rho kinase localize to actin stress fibers is unknown.
The protein described here, M-RIP, is expressed in human vascular myocytes, and is bound to myosin phosphatase in these cells. Immunostaining additionally shows that M-RIP colocalizes with actin myofilaments, consistent with a recent report showing p116 RIP3 binding to actin (49). M-RIP is thus localized to the contractile filament where myosin light chain phosphorylation regulates the contractile state, suggesting a role for M-RIP in myosin phosphatase regulation. The amino terminus of M-RIP contains adjacent pleckstrin homology domains and polyproline motifs, a structural combination also found on Bruton's tyrosine kinase where it mediates binding to both actin and G␣ 12 (50,51). This region of p116 RIP3 has recently been shown to mediate actin binding and actin bundling activity in vitro (49). Whereas p116 RIP3 has been noted to be present in the cell nucleus (47), we could not detect endogenous M-RIP in the nucleus of vascular smooth muscle cells, implying either a difference in localization between the murine and human homologs, or cell type-specific nuclear localization.
The COOH-terminal domain of M-RIP interacts with both MBS and RhoA. The RhoA-binding domain of M-RIP overlaps the amino-terminal 95 amino acids of CC2, which raised the possibility that M-RIP interacts with MBS indirectly via RhoA. However, we have found that both RhoA and MBS bind M-RIP directly to separate adjacent domains (amino acids 545-823 for RhoA and 823-878 for MBS). Our data support a model where M-RIP binding brings MBS and RhoA into proximity (Fig. 7).
The COOH-terminal domain of MBS contains a leucine zipper domain that mediates binding to a leucine/isoleucine zipper in cGMP-dependent protein kinase 1␣ (43). The present study shows that the MBS leucine zipper domain also binds the COOH-terminal 55 amino acids of the M-RIP CC2 domain. Thus the same domain of MBS binds to both proteins. Our data support that MBS can bind M-RIP and cGMP-dependent protein kinase 1␣ simultaneously because M-RIP can interact with cGMP-dependent protein kinase 1␣, as shown by coimmunoprecipitation, without evidence of direct binding, suggesting that the interaction occurs indirectly with MBS as the intermediary protein. It will be of interest in future studies to explore the mechanism of this multimeric interaction. p116 RIP3 was initially described as a RhoA-binding protein (45), although a subsequent study was unable to confirm binding to RhoA (49). In contrast, using both purified protein interaction studies and cell lysates, we find that M-RIP binds RhoA directly and independently of nucleotide binding state. The high ionic strength and detergents required to solubilize and immunoprecipitate M-RIP would be expected to disrupt binding to RhoA and Rho kinase (40). However, using a glycerol bound to glutathione-agarose beads. Bound His 6 RhoA was detected using anti-RhoA antibodies. Amounts of input GST fusion proteins were equivalent as detected by Ponceau stain (data not shown). C, COS-1 cells transfected with FLAG-RhoA were lysed in buffer A and incubated with GST, GST-Rhotekin RBD , GST-M-RIP-(545-823), and GST-M-RIP-(545-878) bound to glutathione-agarose beads. Aliquots of the bound protein were separated on 12.5 and 7.5% protein gels and immunoblots for FLAG (M2) and MBS were performed, respectively. 5% of input lysate is shown on the left. extraction method shown to preserve RhoA and Rho kinase localization to stress fibers (40) we found that M-RIP copurified with RhoA and Rho kinase. Immunostaining of glycerol-extracted coronary myocytes confirmed that stress fiber architecture and colocalization of M-RIP, MBS, and RhoA with these structures was preserved. Furthermore, despite the presence of detergent and high ionic strength, Rho kinase activity could be detected in the M-RIP immunopellet because of the high sensitivity of autoradiography. Coupled with the in vitro direct binding studies, these data strongly support that M-RIP interacts with RhoA in vivo.
The M-RIP interaction occurs without preference for the nucleotide binding state of RhoA, whereas Kimura et al. (21) found that only GTP-RhoA interacted with myosin phosphatase. If M-RIP links RhoA to myosin phosphatase, it is unclear how GTP dependence of the RhoA-M-RIP-myosin phosphatase interaction occurs. Further study is required to determine whether a complex interaction between M-RIP, RhoA, and MBS establishes GTP dependence of the interaction, or whether a Rho kinase phosphorylation event stabilizes the complex.
Both our study of M-RIP and a recent report of p116 RIP3 indicate that these proteins colocalize with actin filaments. Interestingly, Mulder et al. (49) show that p116 RIP is colocalized with both stress fibers and lamellepodia, has actin bundling properties in vitro, and causes actin filament disassembly when overexpressed. Studies are currently underway to determine whether M-RIP has a direct effect on myosin phosphatase activity, and whether it modulates the smooth muscle contractile state by regulating actin myofilament structure and composition.
In summary, M-RIP, the human homolog of p116 RIP3 , binds myosin phosphatase and RhoA in vascular smooth muscle cells. Our binding studies support a model where M-RIP brings MBS and RhoA into proximity (Fig. 7). The binding to myosin phosphatase and RhoA and localization to actin myofilaments suggest that M-RIP may target RhoA to the myosin phosphatase complex to regulate the myosin phosphorylation state.