p116Rip decreases myosin II phosphorylation by activating myosin light chain phosphatase and by inactivating RhoA.

p116Rip was originally found to be a RhoA-binding protein, but its function has been unknown. Here, we clarify the function of p116Rip. Two critical findings were made. First, we found that p116Rip activated the GTPase activity of RhoA in vitro and that p116Rip overexpression in cells consistently diminished the epidermal growth factor-induced increase in GTP-bound RhoA. Second, p116Rip activated the myosin light chain phosphatase (MLCP) activity of the holoenzyme. p116Rip did not activate the catalytic subunit alone, indicating that the activation is due to the binding of p116Rip to the myosin phosphatase targeting subunit MYPT1. Interestingly, the activation of phosphatase was specific to myosin as substrate, and p116Rip directly bound to myosin, thus facilitating myosin/MLCP interaction. The gene silencing of p116Rip consistently and significantly increased myosin phosphorylation as well as stress fiber formation in cells. Based upon these findings, we propose that p116Rip is an important regulatory component that controls the RhoA signaling pathway, thus regulating MLCP activity and myosin phosphorylation in cells.

The actin-myosin system plays a fundamental role in the regulation of cell motility, including cell contractility, migration, division and shape changes. In vertebrates, the phosphorylation of the regulatory light chain RLC20 regulates conventional myosin (myosin II) for both motor activity and filament formation (1)(2)(3). The extent of phosphorylation of myosin II is controlled by the change in the myosin light chain kinase and myosin light chain phosphatase (MLCP) 1 activities. There are two pathways that regulate myosin phosphorylation. One is a Ca 2ϩ -dependent pathway. The increase in cytoplasmic Ca 2ϩ caused by external stimuli activates Ca 2ϩ /calmodulin-dependent myosin light chain kinase, thus increasing myosin phos-phorylation. The other is the Ca 2ϩ -independent pathway. This pathway regulates both myosin phosphorylation and dephosphorylation processes. It has been shown recently that several Ca 2ϩ -independent kinases can phosphorylate myosin in vitro (4 -7). Among them, zipper-interacting protein kinase has been shown to play an important role in phosphorylating myosin in migrating mammalian cells (8). On the other hand, a number of studies have revealed that MLCP is regulated by a RhoA signaling pathway (9,10).
MLCP consists of three subunits: a myosin phosphatase targeting subunit (MYPT1), a 20-kDa small subunit, and a catalytic subunit of the type 1 protein serine/threonine phosphatase family (11)(12)(13). The N-terminal one-third of the large subunit is composed of eight repeat sequences that correspond to the sequence for an ankyrin repeat, a motif that is found in proteins involved in tissue differentiation and cell cycle and cytoskeleton regulation (14). The domain responsible for the binding of MYPT1 to myosin has been studied. Ichikawa et al. (15) showed that the recombinant N-terminal two-thirds of MYPT1 contain the myosin-binding site because phosphorylated heavy meromyosin (HMM) and RLC20 bound to an affinity column made using this recombinant fragment. On the other hand, Johnson et al. (16) reported that the C-terminal 291 residues of the large subunit, but not the N-terminal fragment, bind to myosin. Although the identity of the myosinbinding domain of MLCP remains controversial, it is generally agreed that the large subunit is the myosin targeting subunit.
The interaction of the subunits of MLCP has been studied. Johnson et al. (16) showed that the 20-kDa subunit does not interact with the catalytic subunit, but does interact with the C-terminal 72 residues of MYPT1. Hirano et al. (17) reported that the catalytic subunit binds to MYPT1 at two sites: a relatively strong site in the N-terminal 38 residues and a weaker site in the ankyrin repeat (residues 39 -295). The phosphorylated light chain-binding site is also assigned to the ankyrin repeat. The existence of several isoforms of the type I phosphatase catalytic subunit has been demonstrated (␣ 1 , ␣ 2 , ␥ 1 , ␥ 2 , and ␦) (18 -20), with the catalytic subunit of MLCP identified as a ␦-isoform (11).
It has been postulated that the substrate specificity and regulation of protein phosphatases are governed by their regulatory subunits and that the regulatory subunits act as targeting subunits (21). For MLCP, it is known that its holoenzyme has higher activity than its catalytic subunit (11,12), suggesting that the binding of the regulatory subunits increases MLCP activity. Consistent with this finding, Gong et al. (22) showed that the inhibition of MLCP by a high arachidonic acid concentration is due to the dissociation of the catalytic subunit from the holoenzyme.
A critical finding that shed light on the linkage between RhoA and its downstream cascade was the discovery of Rho-dependent protein kinase (referred to as Rho kinase). Rho kinase was cloned from various tissues (23)(24)(25). The kinase (ϳ160 kDa) is composed of several domains, including a catalytic domain, a coiled-coil domain, a pleckstrin homology domain, and a cysteine-rich zinc finger motif that is homologous to the cysteine-rich domains of protein kinase C (24,25). Interestingly, it was shown that MYPT1 can be phosphorylated by Rho kinase, which results in a decrease in MLCP activity in vitro (26). Rho kinase phosphorylates MYPT1 at two sites in vitro, i.e. Thr 641 and Thr 799 , with Thr 641 responsible for the inhibition of MLCP activity (27). This suggests that activation of the RhoA signaling pathway would phosphorylate MYPT1 by Rho kinase, thus down-regulating MLCP and increasing myosin motor function. It may be asked whether an external stimulus induces the phosphorylation of MYPT1 at Thr 641 . Quite recently, it was shown that during agonist-induced stimulation in smooth muscle, the phosphorylation of MYPT1 at Thr 641 was unchanged, whereas RLC20 phosphorylation was significantly increased (28). This suggests that MYPT1 phosphorylation at Thr 641 is not responsible for the change in myosin phosphorylation in smooth muscle contractile regulation, although it may play a role in other cell motility systems.
A critical question is how the RhoA signaling pathway regulates MLCP activity, thus myosin phosphorylation. To address this question, we studied the MYPT1-interacting proteins using the yeast two-hybrid system. One of the positive clones is a p116 Rip , previously found to be a RhoA-interacting protein (29). It was reported that overexpression of p116 Rip attenuates lysophosphatidic acid-induced cell shape change, suggesting that p116 Rip might inhibit RhoA function. The mechanism by which p116 Rip attenuates the lysophosphatidic acid-induced cell shape change is unclear. Quite recently, Surks et al. (30) reported that p116 Rip interacts with MYPT1. To date, the function of p116 Rip has not been understood. Here, we found that p116 Rip activates MLCP activity. The activation is achieved by p116 Rip facilitation of the binding of MYPT1 to myosin. Furthermore, we found that p116 Rip has RhoGAP activity, thus inactivating RhoA activity. The gene silencing of p116 Rip consistently and significantly increased myosin phosphorylation in cells. Based upon these findings, we propose that p116 Rip is an important regulatory component that controls the RhoA signaling pathway, thus regulating MLCP activity and myosin phosphorylation in cells.

EXPERIMENTAL PROCEDURES
Yeast Two-hybrid Analysis-Yeast two-hybrid screening was performed using the Matchmaker Two-Hybrid System 3 (Clontech) according to the manufacturer's instructions. Yeast AH109 strains were sequentially transformed with the pGBKT7 vector, coding for full-length rat MYPT1 with a leucine zipper fused in-frame to the DNA-binding domain of Gal4, and then with a human aorta cDNA library constructed in the pACT2 vector, coding for the activation domain of Gal4. Initial transformants were selected as being positive on synthetic complete plates lacking tryptophan/leucine/histidine/adenine and containing 15-Bromo-4-chloro-3-indolyl-␣-galactopyranoside (X-␣-gal) on the surface of the plates.
Vector and cDNA Constructs-The coding sequence for the Rho-binding domain of rhotekin (amino acids 7-89) was amplified from a mouse trachea RNA preparation by reverse transcription-PCR and subcloned into the pGEX4T2 vector. The cDNAs were generated from total RNA of human trachea by reverse transcription-PCR and used as template to generate a full-length p116 Rip cDNA (DDBJ/GenBank TM /EBI accession number AB189741). Amplified PCR products were subcloned into pFBHT containing a FLAG tag at the N terminus or the pEGFP vector. The entire coding region of rat MYPT1 cDNA was subcloned into pFBHTc at the SalI site. The cDNAs of RhoA and Rho kinase were gifts of Dr. Alan Hall (University College of London, London, United Kingdom) and Dr. T. Leung (University of Singapore), respectively.
Purification of Proteins-Smooth muscle myosin and myosin light chain kinase were prepared from frozen turkey gizzards as described previously (32,33). Xenopus oocyte calmodulin was expressed in Escherichia coli and purified as described previously (34,35). Recombinant MLC20 was prepared as described previously (34,35). HMM was prepared by ␣-chymotryptic hydrolysis of gizzard myosin (36). Recombinant MLCP holoenzyme was purified by co-infecting Sf9 cells with viruses expressing rat His-MYPT1 with a leucine zipper, rat PP1␦, and rat M21 as described previously (37). The expressed MLCP holoenzyme was purified by Ni 2ϩ -agarose affinity chromatography (QIAGEN Inc.) according to the manufacturer's protocol. Recombinant p116 Rip , MYPT1, RhoA, Rho kinase, and PP1␦ were also purified following the same procedure.
In Vitro Binding Assay-Ten g of FLAG-tagged p116 Rip was mixed with 10 g of His-MYPT1 and/or 10 g of His-RhoA in 30 mM Tris-Cl (pH 7.5), 50 mM NaCl, and 0.5% Nonidet P-40 for 1 h at 4°C. Ten g of FLAG-tagged p116 Rip was also mixed with 20 g of Rho kinase under the same conditions. Anti-FLAG antibody-agarose was added to the mixture, and the mixture was incubated for another 30 min. Anti-FLAG antibody-agarose was washed three times with the same buffer, and the bound proteins were eluted with FLAG peptide (Sigma) in buffer containing 30 mM Tris-Cl (pH 7.5) and 150 mM NaCl. The bound proteins were analyzed by Western blotting or Coomassie Blue staining. For the binding of p116 Rip to myosin, 10 g of FLAG-tagged p116 Rip and 17.5 g of unphosphorylated or phosphorylated myosin were mixed in buffer containing 30 mM Tris-Cl (pH 7.5), 50 mM NaCl, and 0.5% Nonidet P-40 for 1 h at 4°C. The bound proteins were determined as described above.
Immunoprecipitation-COS-7 cells attached to the culture dish were washed three times with cold phosphate-buffered saline and scraped into lysis buffer containing 50 mM Tris-Cl (pH 7.5), 150 mM NaCl, 1% Triton X-100, 1 mM phenylmethylsulfonyl fluoride, and 10 g/ml leupeptin. The cells were lysed through a 26-gauge needle. The samples were centrifuged at 14,000 ϫ g for 5 min at 4°C. Supernatants were incubated with protein A-Sepharose for 1 h at 4°C, and the resin was centrifuged to remove nonspecific binding proteins. The supernatants were incubated overnight with control IgG, anti-MYPT1 antibody, or anti-p116 Rip antibody at 4°C and then incubated with protein A-Sepharose for 1 h at 4°C. The resin was precipitated, and the bound proteins were resolved by SDS-PAGE.
Quantitative Analysis of the Binding of p116 Rip to MYPT1 or HMM-A fixed amount of FLAG-tagged p116 Rip  The proteins bound to p116 Rip were analyzed by Western blotting (for MYPT1) or Coomassie Blue staining (for HMM). To determine the amount of MYPT1 bound to p116 Rip , various amounts of purified MYPT1 were also analyzed by Western blotting to serve as standards. The amount of MYPT1 or HMM bound to p116 Rip was determined by densitometry and fit to a single rectangular hyperbola using Prism Version 4.0 software.

Measurement of Protein Expression Levels in Cells-
The homogenates of COS-7 cells were prepared as described above. Protein concentration was measured using a DC protein assay kit (Bio-Rad). The standard curves were obtained with 2.5-10 ng of purified p116 Rip and 15-25 ng of purified MYPT1, and the data was fit by linear regression. Approximately 40 g of the total COS-7 cell lysates, which fell on the linear portion of the standard curves of p116 Rip and MYPT1, was subjected to SDS-PAGE, followed by Western blotting.
Protein Phosphatase and Kinase Assay-After incubation of smooth muscle myosin with 1 M microcystin LR for 10 min at room temperature, myosin (5 mg/ml) was phosphorylated with 10 g/ml myosin light chain kinase and 10 g/ml calmodulin in buffer containing 0.2 mM [␥-32 P]ATP, 150 mM KCl, 1 mM MgCl 2 , 0.2 mM CaCl 2 , 1 mM dithiothreitol (DTT), 30 mM Tris-HCl (pH 7.5), and 0.1 M microcystin LR at 25°C for 30 min. The phosphorylated myosin was precipitated by adding 8 volumes of salt-free buffer containing 15 mM MgCl 2 , 1 mM DTT, and 30 mM Tris-HCl (pH 7.5). After centrifugation at 14,000 ϫ g for 5 min, the precipitated myosin was dissolved with high salt buffer (0.4 M KCl, 1 mM DTT, and 30 mM Tris-HCl (pH 7.5). The precipitation-solubilization cycle was repeated again. The phosphorylated myosin was finally dissolved in high salt buffer at a concentration of 25 M. MLC20 was also phosphorylated as described previously (34). The phosphorylated MLC20 was precipitated with 10% trichloroacetic acid. The precipitates were washed three times with water and dissolved in 50 mM NaCl, 30 mM Tris-HCl (pH 8.0), 1 mM EGTA, and 1 mM DTT. Protein phosphatase assays were carried out at 25°C using [␥-32 P]ATP-labeled myosin or MLC20 (final concentration of 3 or 0.75 M, respectively) as substrate in the presence of 100 mM NaCl, 30 mM Tris-HCl (pH 7.5), 1 mM EDTA, 0.2 mg/ml bovine serum albumin, and 1 nM MLCP or 20 nM PP1␦. MLCP at 10 nM was used when phospho-MLC20 was used as substrate. The reactions were terminated by the addition of 10% trichloroacetic acid. After sedimentation of the proteins at 5000 ϫ g for 5 min, Cerenkov counting was performed to determine the radioactivity of the supernatant containing the liberated 32 P. A protein kinase assay was performed as described previously (28).
GTP Hydrolysis of RhoA-RhoA (3 M) was first loaded with 10 M [␥-32 P]GTP in buffer containing 20 mM Tris-HCl (pH 7.5), 50 mM NaCl, and 1 mM EDTA at 30°C for 10 min. GTP hydrolysis was initiated by incubating the [␥-32 P]GTP-bound RhoA with or without p116 Rip in the above buffer containing 5 mM MgCl 2 , 0.1 mM GTP, and 1 mg/ml bovine serum albumin at 30°C. The reaction was terminated at 10 min by adding 20 volumes of ice-cold dilution buffer containing 20 mM Tris-HCl (pH 7.5), 50 mM NaCl, and 5 mM MgCl 2 . The diluted samples were immediately filtered through 25-mm BA85 nitrocellulose filters (Schleicher & Schü ll) and washed with 5 ml of ice-cold dilution buffer. The filters were assessed for radioactivity using a Beckman scintillation counter.
Rhotekin Binding Assay-HeLa cells were mock-transfected or transfected with a p116 Rip -expressing vector for 24 h and serum-starved for 48 h. Approximately 70% of the total cells were mock-or p116 Riptransfected. The transfected cells were treated with 10 ng/ml epidermal growth factor. RhoA activity was assessed at the indicated times using the Rho-binding domain (RBD) of rhotekin as described (38). Briefly, cells were washed with ice-cold phosphate-buffered saline and lysed in 50 mM Tris (pH 7.5), 1% Triton X-100, 0.1% SDS, 500 mM NaCl, 10 mM MgCl 2 , 10 g/ml leupeptin, and 1 mM phenylmethylsulfonyl fluoride. Cell lysates were clarified by centrifugation at 14,000 ϫ g for 5 min at 4°C, and supernatants were incubated with GSH-RBD beads (8 g) at 4°C for 45 min. The beads were washed three times with Tris buffer containing 1% Triton X-100, 150 mM NaCl, 10 mM MgCl 2 , 10 g/ml leupeptin, and 0.1 mM phenylmethylsulfonyl fluoride. Bound RhoA proteins were detected by Western blot analysis.
Preparation and Transfection of Small Interfering RNA (siRNA)-The selected sequences were submitted to a BLAST search to ensure that only the p116 Rip gene was targeted. p116 Rip siRNA1 (AATGGCAGCGACGGT-TCTT, corresponding to bp 203-221 of p116 Rip ) and p116 Rip siRNA2 (AATTGACTTGTCCGCATGT, corresponding to bp 1290 -1308 of p116 Rip ) were used. A pair of 67-nucleotide complementary oligonucleotides were synthesized separately with the addition of a BamHI site to the 5Ј-end and an EcoRI site to the 3Ј-end. The annealed 67-bp cDNA fragment with p116 Rip siRNA was cloned into the pSIREN-DNR-DsRed vector (BD Knockout RNAi Systems, Clontech), which has two promoter regions for producing siRNA and DsRed, respectively. Negative control oligonucleotides were obtained from the RNAi-ready pSIREN-DNR-DsRed express kit (Clontech). Transfection of pSIREN vectors into HeLa cells was performed using FuGENE 6 (Roche Applied Science).
Immunofluorescence Staining and Image Processing-Two different anti-rabbit polyclonal antibodies were used for the independent localization of p116 Rip and MYPT1 as described previously (39). Briefly, COS-7 cells were incubated with anti-rabbit MYPT1 polyclonal antibody, followed by Cy5-conjugated anti-rabbit IgG and unlabeled anti-rabbit IgG to block anti-rabbit MYPT1 antibody. This blocking step assured that all free sites for anti-rabbit IgG interaction were covered before the second series of polyclonal and secondary labeled antibodies. After this blocking step, cells were incubated sequentially with anti-rabbit p116 Rip polyclonal antibody and fluorescein isothiocyanate-conjugated anti-rabbit IgG. As a control, fluorescein isothiocyanate labeling was observed after the deletion of the second polyclonal antibody, providing a measure of the efficacy of the blocking step. Actin filaments were stained with phalloidin-conjugated Alexa Fluor 546 during the incubation with the second secondary antibody. Immunocytochemistry was performed as described previously (8,31,40). Fluorescence images were viewed using a DM IRB laser scanning confocal microscope (Leica) controlled by TCS SPII systems (Leica). All images were taken with the same laser output to directly compare the fluorescence signal intensities. Images were processed using Adobe® Photoshop® Version 7.0 software.

RESULTS
Binding of p116 Rip to MYPT1-The interaction between p116 Rip and MYPT1 was found by yeast two-hybrid screening (Fig. 1A). The entire coding region of rat MYPT1 was used as bait, and the positive clones were subjected to X-␣-gal assay to eliminate false positive clones. Two of the positive clones, TM3 and TM15, encoded bp 1840 -3111 and 1459 -3111 of human p116 Rip , respectively (Fig. 1A). The binding between p116 Rip , RhoA, and MYPT1 was studied. GFP-tagged p116 Rip expressed in COS-7 cells was immunoprecipitated with anti-GFP antibodies. The immunoprecipitated complex was analyzed by SDS-PAGE, followed by Western blot analysis (Fig. 1B). MYPT1 was co-immunoprecipitated with GFP-p116 Rip using anti-GFP antibodies, but not control IgG (Fig. 1B, left panels). Co-immunoprecipitation of p116 Rip with GFP-MYPT1 using anti-GFP antibodies was also observed with the cell lysates expressing GFP-tagged MYPT1 (Fig. 1B, right panels). These results support the interaction between p116 Rip and MYPT1. To further address this issue, we examined the binding of endogenous p116 Rip and MYPT1.
The COS-7 cell lysates were immunoprecipitated with anti-p116 Rip antibody (Fig. 1C, left panels). Both MYPT1 and RhoA were co-immunoprecipitated. On the other hand, when anti-MYPT1 antibodies were used for the immunoprecipitation, both p116 Rip and RhoA were co-immunoprecipitated (Fig. 1C, right panels). To determine whether co-immunoprecipitation of the three proteins resulted from the direct binding of these proteins, we expressed full-length MYPT1 and p116 Rip using the baculovirus expression system, and the isolated proteins were subjected to a binding assay. Full-length p116 Rip flanked with a FLAG tag at the N-terminal end was incubated with MYPT1, and anti-FLAG antibody-agarose was added to the mixture. After the resin was washed exhaustively to eliminate the unbound proteins, FLAG-p116 Rip was eluted with FLAG peptide. MYPT1 co-eluted with FLAG-p116 Rip , indicating the direct binding between p116 Rip and MYPT1 (Fig. 1D). There was no nonspecific binding of MYPT1 to anti-FLAG antibodyagarose (Fig. 1D). Fig. 1E also shows the direct binding between p116 Rip and RhoA. We examined RhoA(Q63L), an active form, and RhoA(T19N), an inactive form, for binding to p116 Rip . Both the active and inactive forms of RhoA were added to FLAG-p116 Rip with anti-FLAG antibody-agarose and eluted with FLAG peptide as described above. A similar amount of RhoA was eluted, indicating that p116 Rip binds to RhoA irrespective of the activity of RhoA. The binding of wild-type RhoA to p116 Rip was also observed (data not shown). To determine whether RhoA and MYPT1 bind together to p116 Rip without interfering with each other, both were added to FLAG-p116 Rip /anti-FLAG antibodyagarose and eluted with FLAG peptide. Both MYPT1 and RhoA were eluted from the resin, and the amount of bound MYPT1 and RhoA was similar to that observed for MYPT1 or RhoA alone (Fig. 1F). These results suggest that p116 Rip directly binds to MYPT1 and RhoA and that the binding of one does not interfere with the binding of the other, indicating that RhoA and MYPT1 bind to distinct regions in p116 Rip . This premise was also supported by the results shown in Fig. 1C, where immunoprecipitation of the endogenous proteins by anti-p116 Rip antibody yielded both MYPT1 and RhoA. It should be noted that the binding of RhoA to p116 Rip was weaker than that of MYPT1 to p116 Rip and that although the majority of the added MYPT1 co-eluted from the resin with p116 Rip , only part of the added RhoA was recovered with p116 Rip . Fig. 1G shows the binding of MYPT1 to p116 Rip as a function of MYPT1 concentration. The extent of binding showed a saturation curve, and a maximum binding stoichiometry of 0.83 mol of . C, co-immunoprecipitation of endogenous p116 Rip , MYPT1, and RhoA. Untransfected COS-7 cell lysates were immunoprecipitated with control IgG, anti-p116 Rip antibody, or anti-MYPT1 antibody. D and E, interaction between purified p116 Rip and MYPT1 and between p116 Rip and RhoA, respectively, in vitro. FLAG-tagged p116 Rip and MYPT1 or RhoA were mixed as described under "Experimental Procedures." Proteins bound to anti-FLAG antibody-agarose were eluted with FLAG peptide and analyzed by Western blotting. F, interaction between purified p116 Rip , RhoA, and MYPT1. FLAG-p116 Rip , RhoA, and MYPT1 were mixed, and the proteins bound to anti-FLAG antibody-agarose were eluted with FLAG peptide and analyzed as described under "Experimental Procedures." G, binding of p116 Rip to MYPT1. A fixed amount of FLAG-tagged p116 Rip was mixed with various amounts of MYPT1 and subjected to pull-down assay, followed by Western blot analysis. The amount of MYPT1 bound to p116 Rip was determined by densitometric analysis as described under "Experimental Procedures." H, protein expression of p116 Rip Fig. 1H shows the expression levels of p116 Rip and MYPT1 in COS-7 cells. Total cell homogenates were subjected to SDS-PAGE, followed by Western blotting, and the signals were normalized to estimate the expression levels of these proteins using the isolated p116 Rip and MYPT1 of known concentration as standards. Based upon the results, we estimated the molar ratio of the expression level of p116 Rip to that of MYPT1 to be 1:2. It was reported previously (30) that p116 Rip binds to the leucine zipper motif of MYPT1, suggesting that it binds to the isoform with the leucine zipper. It should be noted that although p116 Rip interacted with RhoA, it did not directly bind to Rho kinase (Fig. 1I). Fig. 1J shows the localization of endogenous p116 Rip and MYPT1 in COS-7 cells, and p116 Rip localization at stress fibers agrees with previous reports (30,41). Consistent with the in vitro data shown above, MYPT1 and p116 Rip co-localized well. The results support that the binding of p116 Rip to MYPT1 takes place in cells.
Effect of p116 Rip on MLCP Activity-Because MYPT1 is a regulatory subunit of MLCP, we examined whether the binding of p116 Rip to MYPT1 affects phosphatase activity. As shown in Fig. 2A, p116 Rip significantly activated the myosin phosphatase activity of the MLCP holoenzyme in a concentration-de-

FIG. 2. Effect of p116 Rip on MLCP or PP1␦ activity and interaction between p116 Rip and myosin.
A, activation of phosphatase activity by p116 Rip . Various concentrations of p116 Rip were preincubated with MLCP or PP1␦ for 5 min at 25°C, and the phosphatase reaction was started by the addition of [␥-32 P]ATP-labeled myosin. Values are means Ϯ S.E. of three independent experiments and are expressed as 100% of phosphatase activity in the absence of p116 Rip . B, effect of p116 Rip on the phosphorylation of MYPT1 at Thr 641 by Rho kinase. MYPT1 was preincubated with or without p116 Rip for 5 min at 25°C and phosphorylated by Rho kinase for the indicated times. Western blotting using phosphorylation site-specific antibodies (anti-phospho-Thr 641 MYPT1 antibody (Ab.)) determined the phosphorylation of MYPT1 at Thr 641 . C, effect of p116 Rip on binding between MYPT1 and PP1␦. The MLCP holoenzyme (70 nM) was incubated with or without 500 nM p116 Rip for 1 h at 4°C. MLCP were immunoprecipitated (I.P.) with anti-MYPT1 antibody. Immunoprecipitates were analyzed by Western blotting using anti-p116 Rip , anti-MYPT1, and anti-PP1␦ antibodies. D, p116 Rip facilitates myosin (but not MLC20) dephosphorylation. Phosphatase activity was determined as described for A. Values are means Ϯ S.E. of three independent experiments. E, binding of p116 Rip to myosin. Smooth muscle myosin was phosphorylated by myosin light chain kinase using [␥-32 P]ATP as described previously (34). Unphosphorylated or phosphorylated myosin was incubated with or without FLAG-tagged p116 Rip as described under "Experimental Procedures." Proteins bound to anti-FLAG antibody-agarose were eluted with FLAG peptide and subjected to SDS-PAGE, followed by autoradiography. MHC, myosin heavy chain. F, binding of p116 Rip to HMM. A fixed amount of FLAG-tagged p116 Rip was mixed with various amounts of HMM and subjected to pull-down assay, followed by Coomassie Blue staining. The amount of HMM bound to p116 Rip was determined by densitometric analysis as described under "Experimental Procedures." pendent manner. On the other hand, p116 Rip did not affect the phosphatase activity of the PP1␦ catalytic subunit (Fig. 2A). These results suggest that the binding of p116 Rip to MYPT1 alters the interaction between MYPT1 and myosin, thus activating phosphatase activity.
It is known that Rho kinase phosphorylates MYPT1 and that the phosphorylation of MYPT1 at Thr 641 inhibits MLCP activity (27). We examined whether the binding of p116 Rip to MYPT1 affects the phosphorylation of MYPT1 by Rho kinase. MYPT1 was phosphorylated by Rho kinase in the presence or absence of p116 Rip , and the phosphorylation of Thr 641 was monitored using phosphorylation site-specific antibodies (28). As shown in Fig. 2B, the phosphorylation of MYPT1 at Thr 641 was not affected by p116 Rip binding.
Because it is known that MLCP activity is affected by the binding of MYPT1 to the catalytic subunit, we examined the effect of p116 Rip on the association of MYPT1 with the catalytic subunit. p116 Rip was added to the MLCP holoenzyme and immunoprecipitated with anti-MYPT1 antibodies. All of the catalytic subunit was co-immunoprecipitated regardless of the presence of p116 Rip (Fig. 2C), suggesting that p116 Rip does not interfere with the association of MYPT1 and PP1␦.
It is critical to determine whether the activation of the MLCP-induced dephosphorylation of myosin by p116 Rip is due to the activation of the enzyme itself or to the change in the interaction between myosin and MYPT1. To address this question, we examined the effect of p116 Rip on the MLCP-induced dephosphorylation of isolated MLC20. Quite interestingly, the dephosphorylation rate of MLC20 was not significantly increased in the presence of p116 Rip (Fig. 2D), suggested that the increase in the MLCP-induced dephosphorylation of myosin by p116 Rip resulted from the p116 Rip -induced changes in the interaction between MLCP and myosin. Therefore, we examined whether p116 Rip binds directly to myosin or strengthens the binding between MYPT1 and myosin. FLAG-tagged p116 Rip was mixed with myosin, and the complex was precipitated with anti-FLAG antibody-agarose. As shown in Fig. 2E, myosin was pulled down by p116 Rip . On the other hand, myosin was not precipitated by the resin without p116 Rip . Fig. 2F shows the binding of HMM to p116 Rip as a function of HMM concentration. We used HMM to estimate the binding stoichiometry; we did not use whole myosin because myosin forms filaments under physiological ionic conditions, and the binding of one molecule of myosin in the filament to p116 Rip coprecipitates a number of myosin molecules with the p116 Rip -bound resin. This is avoided using the soluble fragment of myosin, HMM. The extent of binding showed a saturation curve, and a maximum binding stoichiometry of 1.1 mol of head HMM/mol of p116 Rip with an K d of 1.0 ϫ 10 Ϫ6 M was obtained. It is known that the phosphorylation of myosin at MLC20 Ser 19 changes the conformation of myosin (42,43). Therefore, we examined the effect of myosin phosphorylation on the binding of p116 Rip to myosin. There was no significant change in binding upon MLC20 phosphorylation (Fig. 2E). It should be noted that myosin could be precipitated by high speed centrifugation, but not by low speed centrifugation (100 ϫ g) used for the experiment. These results clearly demonstrate that p116 Rip directly binds to myosin.
Effect of p116 Rip on RhoA GTPase Activity-We next examined whether p116 Rip influences RhoA function. It has been shown previously that RhoA activity depends upon the nucleotide to which it is bound (44). GTP-RhoA and GDP-RhoA are the active and inactive forms, respectively. The regulatory proteins (GTPase-activating protein and GDP-GTP exchange factor) control the relative presence of the two forms. The GTPaseactivating protein favors the production of the GDP-bound form, whereas the GDP-GTP exchange factor induces the ex-change of the bound GDP with GTP to facilitate the production of the GTP-bound form (44). We first determined the effect of p116 Rip on the rate of GTP hydrolysis by measuring the decrease in the GTP-bound RhoA concentration. As shown in Fig.  3A, p116 Rip significantly decreased the amount of GTP-bound RhoA in a p116 Rip concentration-dependent manner. This result suggests that p116 Rip enhances the GTPase activity of RhoA, i.e. GTPase-activating protein activity. We then examined the RhoGEF activity of p116 Rip . RhoA was incubated with GDP, and then [␥-35 S]GTP␥S was added in the presence or absence of p116 Rip . The production of GTP-RhoA was monitored as described under "Experimental Procedures." p116 Rip had no influence on the exchange of GTP for GDP (data not shown).
p116 Rip Inhibits RhoA Activity in Cells-As p116 Rip increased the GTPase activity of RhoA, we anticipated that p116 Rip would inactivate RhoA by destabilizing the GTP-bound form of RhoA. To investigate whether p116 Rip affects RhoA activity in cells, we examined the effect of p116 Rip on RhoA activity by measuring the amount of GTP-RhoA in a rhotekin binding assay. HeLa cells were transfected with GFP-p116 Rip or GFP alone. The protein expression of GFP-p116 Rip was confirmed by Western blot analysis (Fig. 3B). p116 Rip overexpression did not affect the RhoA expression level (Fig. 3B). After serum starvation, the cells were challenged by epidermal growth factor for stimulation, and total cell lysates were obtained at various times after the stimulation. Beads coated with the RBD of rhotekin were added to the lysates for pulldown assay. As shown in Fig. 3C, the RBD-binding form of RhoA in the GFP-transfected cell lysates increased with time after stimulation. On the other hand, the RBD-binding form of RhoA was markedly diminished in cells transfected with p116 Rip . These results are consistent with the finding that p116 Rip activated RhoA GTPase activity, suggesting that p116 Rip inhibits RhoA activity in cells.
Gene Silencing of p116 Rip Increases MLC20 Phosphorylation-Because p116 Rip affected both MLCP and RhoA activities, we anticipated that p116 Rip would be involved in the regulation of myosin II phosphorylation in cells. To address this issue, we studied the effect of the elimination of p116 Rip on myosin phosphorylation in cells. We attempted to eliminate the expression of p116 Rip by transfecting the p116 Rip -specific siRNA into cells. Two sequences were chosen for the production of siRNA, siRNA1 and siRNA2. HeLa cells were transfected with the pSIREN-DNR-DsRed vector containing either p116 Rip siRNA sequence or control sequence (see "Experimental Procedures"). The transfected cells could be identified by the expression of DsRed. The transfection efficiency was ϳ80%. The cells were harvested and subjected to Western blot analysis to evaluate the efficiency of gene silencing. Both siRNA1 and siRNA2 efficiently eliminated p116 Rip expression, whereas the p116 Rip siRNAs did not affect the expression of MYPT1 or MLC20 (Fig.  4A). An important finding is that p116 Rip siRNA significantly increased the phosphorylation of MLC20 as revealed by Western blotting using the antibodies specific to phospho-Ser 19 of MLC20 and to diphosphorylated MLC20 (8) as probes (Fig. 4A). The extent of phosphorylation of MLC20 in siRNA-treated cells was estimated using the phosphorylation site-specific antibody as a probe. Total cell homogenates were phosphorylated by exogenous myosin light chain kinase/calmodulin to completely phosphorylate MLC20, and this was used as a reference for 100% phosphorylation of MLC20. As shown in Fig.  4B, MLC20 phosphorylation was increased from nearly 5% to ϳ25% of total myosin after siRNA treatment. Because the transfection efficiency was ϳ80%, it was calculated that nearly 30% of the myosin in the siRNA-transfected cells was phosphorylated.
The increase in MLC20 phosphorylation upon elimination of p116 Rip was also shown by immunostaining of the cells with the phosphorylation site-specific antibodies to MLC20. As shown in Fig. 4C, the elimination of p116 Rip by siRNA markedly increased MLC20 phosphorylation in both the cell cortical region and stress fibers. On the other hand, the signal intensity and localization probed by antibodies recognizing MLC20 regardless of its phosphorylation state were the same in the siRNA-transfected and control cells. Fig. 5 shows the effect of p116 Rip elimination by siRNA on stress fiber formation. For the cells treated with two different siRNAs for p116 Rip , stress fiber formation was significantly increased compared with the control cells. Note that the non-siRNA-transfected cells on the same coverslip did not show any increase in stress fiber formation. The number of cells showing stress fibers was significantly increased upon siRNA transfection (Fig. 5B). The fraction of untransfected cells showing stress fibers was the same as that of control siRNA-transfected cells (Fig. 5B). It should be noted that although ϳ20% of the control cells showed clear stress fibers, the number of stress fibers was clearly lower than in p116 Rip siRNA-transfected cells. It is known that RhoA controls stress fiber formation (45,46). It is also known that an increase in MLC20 phosphorylation stabilizes stress fibers (47,48). Therefore, these results are consistent with the finding that p116 Rip has RhoGAP activity and MLCPactivating function. DISCUSSION We have found that p116 Rip binds to both RhoA and MYPT1. Quite recently, Surks et al. (30) demonstrated an interaction between p116 Rip and MYPT1 using yeast two-hybrid screening. However, the function of p116 Rip has not been clarified. Our results are consistent with this recent report and further show that p116 Rip can bind to both RhoA and MYPT1 simultaneously, thus forming a ternary complex. It was shown previously that the RhoA-binding region and MYPT1-binding region  19 in HeLa cells was immunostained with anti-diphosphorylated MLC20 antibody and visualized by fluorescein isothiocyanateconjugated anti-rabbit IgG (green). DsRed-conjugated siRNA (red) was visualized with fluorescence signals. MLC20 was immunostained with anti-MLC20 antibodies recognizing MLC20 regardless of its phosphorylation state, followed by Cy5-conjugated anti-mouse IgM (blue). Scale bar ϭ 10 m. reside in the first coiled-coil region and the second/third coiledcoil region of p116 Rip , respectively (30). Our results are consistent with this finding and suggest that MYPT1 and RhoA do not interfere with each other in their binding to p116 Rip .
The critical finding of this study is that p116 Rip increases RhoA GTPase and MLCP activities. It is anticipated that the activation of RhoA GTPase activity increases GDP-RhoA, an inactive form, thus decreasing the GTP-bound form of RhoA. In support of this finding, the rhotekin binding assay revealed that the expression of p116 Rip inhibits RhoA activation by epidermal growth factor in cells. This result suggests that p116 Rip down-regulates RhoA activity in cells, thus forming the inactive form of RhoA. This is consistent with the RhoGAP activity of p116 Rip . We also observed that the elimination of p116 Rip by p116 Rip -specific siRNA resulted in an increase in stress fiber formation. Because it is known that the activation of RhoA increases stress fiber formation (45,46), the elimination of p116 Rip with RhoGAP activity increases GTP-bound RhoA, thus increasing stress fiber formation.
Another important finding of this study is that the binding of p116 Rip to MYPT1 activates MLCP activity. The elimination of p116 Rip by the p116 Rip -specific siRNA consistently increased MLC20 phosphorylation. This result supports the notion that the p116 Rip -induced activation of MLCP takes place in vivo. The activation of the myosin light chain dephosphorylation activity is due to the binding of p116 Rip to MYPT1 since p116 Rip did not activate the PP1␦ catalytic subunit. Quite interestingly, p116 Rip increased MLCP-catalyzed myosin dephosphorylation, but not isolated MLC20 dephosphorylation. Consistent with this finding, we have found that p116 Rip directly binds to myosin. Taking these results into account, it is thought that p116 Rip interacts with both myosin and MYPT1 simultaneously, facilitating MLCP to access the phosphorylation site of myosin at MLC20 and enhancing the dephosphorylation of myosin. It is known that MYPT1 or the MLCP holoenzyme binds to myosin with moderate binding activity. On the other hand, it is also known that MLCP activity is tightly associated with myosin under physiological ionic conditions and is barely removed from myosin preparations (49). Theoretically, MLCP should be easily removed from myosin if the binding of MLCP to myosin is moderate. Our findings suggest that p116 Rip enhances the binding of MLCP to myosin, and it is plausible that the tight association of MLCP activity in myosin preparations may be due to the presence of p116 Rip , although the amount of p116 Rip and MLCP present in myosin preparations is negligible. It was reported previously that p116 Rip can bind to actin with a dissociation constant of 0.5 M (41). p116 Rip consistently showed a discrete localization at the stress fibers and in the cell  cortical region, where actin and myosin are the major components. Therefore, p116 Rip functions as a scaffolding protein that holds MLCP at the actomyosin structure in cells, thus promoting the efficiency and specificity of MLCP against myosin dephosphorylation.
Although p116 Rip was originally found to be a RhoA-interacting protein, we found that the binding of p116 Rip to RhoA is rather weak. This is consistent with the finding that RhoA and p116 Rip do not show significant co-localization in cells. It is known that the active form of RhoA (i.e. GTP-RhoA) translocates from the cytosol to the membrane (25,50). A question is how p116 Rip localized in the actomyosin-based cytoskeleton such as stress fibers interacts with GTP-RhoA to induce GTP hydrolysis, resulting in inactive or inactivated RhoA. Since the overexpression of p116 Rip in HeLa cells resulted in the inactivation of RhoA, the RhoGAP function of p116 Rip should operate in the cellular environment. Although GTP-RhoA translocates to the membrane, it is reasonable to think that GTP-RhoA is in the equilibrium between the membrane-bound and the cytosolic molecules. Cytosolic GTP-RhoA encounters p116 Rip and is converted to GDP-RhoA, thus reducing GTP-RhoA on the membrane in the equilibrium between the membrane and cytosolic molecules. The net result is a decrease in the membrane-bound active RhoA molecules.
Previously, Surks et al. (30) reported that immunoprecipitates from cell lysates using anti-p116 Rip antibodies contained protein kinase activity that was inhibited by Y27632 and claimed that p116 Rip bound to Rho kinase, although the presence of Rho kinase in the immunoprecipitates was not directly shown. In the present study, we found that p116 Rip does not bind directly to Rho kinase. Therefore, it is likely that if Rho kinase was in the immunoprecipitates of p116 Rip , the kinase was co-immunoprecipitated with p116 Rip via RhoA bound to p116 Rip .
It is known that the RhoA pathway plays an important role in the regulation of myosin phosphorylation (9,10). The activation of RhoA activates Rho kinase, which results in the inhibition of MLCP activity, thus increasing myosin phosphorylation (26). Therefore, it is anticipated that p116 Rip facilitates the decrease in myosin phosphorylation by diminishing RhoA-dependent MLCP inhibition.
Our findings provide the following scenario for the function of p116 Rip . p116 Rip binds to GTP-RhoA to accelerate GTP hydrolysis, thus forming GDP-RhoA. The GDP-RhoA produced associates with the guanine nucleotide dissociation inhibitor, resulting in the dissociation of p116 Rip from the RhoA-RhoGDI complex. The dissociated p116 Rip becomes available for binding to GTP-RhoA to initiate another cycle.
Based upon our findings, we propose a model explaining the function of p116 Rip as follows (Fig. 6). p116 Rip co-localizes with actomyosin structures such as stress fibers and cortical actin in cells, where p116 Rip forms a complex with MYPT1 and myosin. p116 Rip holds MLCP with myosin, thus facilitating efficient and specific dephosphorylation of myosin in cells. As a result, myosin phosphorylation at this cytoskeletal structure decreases. On the other hand, p116 Rip facilitates the GTP hydrolysis of RhoA to produce GDP-RhoA. GDP-RhoA then forms a complex with RhoGDI in the cytosol. p116 Rip shifts the equilibrium of GTP-RhoA and GDP-RhoA toward the GDP-bound inactive form. Because Rho kinase phosphorylates MYPT1, which inactivates MLCP activity (26), it is anticipated that the inactivation of RhoA by p116 Rip also contributes to the increase in MLCP activity, thus reducing myosin phosphorylation. Therefore, it is postulated that p116 Rip contributes to the decrease in myosin phosphorylation via activation of the myosin dephosphorylation activity of MLCP and the inactivation of the RhoA pathway.