Activated Rac1 GTPase Translocates Protein Phosphatase 5 to the Cell Membrane and Stimulates Phosphatase Activity in Vitro*

Physiological studies of ion channel regulation have implicated the Ser/Thr protein phosphatase 5 (PP5) as an effector of Rac1 GTPase signaling, but direct biochemical evidence for PP5 regulation by Rac1 is lacking. In this study we used immunoprecipitation, in vitro binding, cellular fractionation, and immunofluorescence techniques to show that the tetratricopeptide repeat domain of PP5 interacts specifically and directly with active Rac1. Consequently, activation of Rac1 promoted PP5 translocation to the plasma membrane in intact cells and stimulated PP5 phosphatase activity in vitro. In contrast, neither constitutively active RhoA-V14 nor dominant negative Rac1N17, which preferentially binds GDP and retains an inactive conformation, bound PP5 or stimulated its activity. In addition, Rac1N17 and Rac1(PBRM), a mutant lacking the C-terminal polybasic region required for Rac1 association with the membrane, both failed to cause membrane translocation of PP5. Mutation of predicted contact residues in the PP5 tetratricopeptide repeat domain or within Rac1 also disrupted co-immunoprecipitation of Rac1-PP5 complexes and membrane translocation of PP5. Specific binding of PP5 to activated Rac1 provides a direct mechanism by which PP5 can be stimulated and recruited to participate in Rac1-mediated signaling pathways.

Rac1 is a ubiquitously expressed member of the Rho monomeric GTPase family (1) that plays an important role in cytoskeletal reorganization, gene expression, cell morphology, cell motility, and activation of the NADPH oxidase complex (2,3). In many cells inactive GDP-bound Rac1 is complexed with Rho-GDI in the cytoplasm (4). Exogenous stimuli such as growth factors translocate Rac1 to the plasma membrane, where it exchanges GDP for GTP to attain an active state (5,6). Localization of Rac1 to the plasma membrane also requires lipid modification and a C-terminal polybasic region (7)(8)(9).
Although Rac and Rho are structurally closely related monomeric GTPases, they often have opposite physiological effects in the nervous system (10). Rac and Rho also have opposite effects on the regulation of Kv11.1 channels in pituitary cells (11). Because Rho exerts many of its effects through Ser/Thr protein kinases (12), we postulated that Rac might antagonize Rho action on Kv11.1 channels through a protein phosphatase, and we have subsequently implicated the Ser/Thr protein phosphatase, PP5, 2 in Rac1-dependent Kv11.1 stimulation by thyroid hormone (13,14). However, the mechanism by which Rac1 regulates PP5 is unknown, and direct binding of Rac1 to PP5 has not yet been demonstrated.
PP5 is a ubiquitously expressed Ser/Thr protein phosphatase that we purified as a lipid-stimulated phosphatase from mammalian brain (15). Unlike other Ser/Thr protein phosphatases, PP5 contains an N-terminal regulatory domain composed of three TPRs (16 -18) that, together with the C terminus, inhibit basal phosphatase activity (19 -21). The TPR domain mediates in vitro stimulation of PP5 by lipid (19,20). The TPR domain also mediates protein-protein interactions controlling localization and association with substrates (22). In particular we noted that one of the classic Rac1 effectors, NADPH oxidase, interacts with Rac1 through the TPR domain of the p67 subunit (3). We showed that the secondary structures of the two TPR domains from PP5 (23) and p67 (3) have nearly identical three-dimensional structures, and we postulated that Rac1 directly stimulates PP5 activity by disinhibiting the catalytic domain (13). However, Rac1 might also stimulate PP5 activity indirectly by recruiting PP5 to the membrane where its TPR domain would be in close proximity to membrane lipids that may stimulate its activity (15,19).
In this study using recombinant proteins, we demonstrate that active Rac1 specifically and directly binds PP5 and stimulates phosphatase activity in vitro. We also performed cellular fractionation, co-immunoprecipitation and immunofluorescence assays to show that activated Rac1 forms a complex with PP5 in intact cells and promotes PP5 translocation to the plasma membrane. In addition, we confirmed that residue * This work was supported, in whole or in part, by National Institutes of Health Lys 93 within the PP5 TPR domain and Rac1 residue Gly 30 , predicted to be critical contact points (13), are required for Rac-PP5 complex formation in cells. Together, these results indicate that Rac1 regulates PP5 via two mechanisms: direct control of phosphatase activity and regulation of cellular localization.

EXPERIMENTAL PROCEDURES
Plasmids and Antibodies-The expression plasmid pCI-FLAG-PP5(WT) has been described previously (24). The single TPR domain mutant of PP5, pCI-FLAG-PP5E93, and the double TPR domain mutant pCI-FLAG-PP5E93/A126 were generated using polymerase chain reaction. The vector pCI-HA-PP5(WT) was generated by using polymerase chain reaction and pCI-FLAG-PP5(WT) as template. GST-Rac/Rho constructs were provided by John P. O'Bryan (University of Illinois at Chicago) (11). The expression vectors for pcDNA-HA-Rac1N17 (dominant negative Rac1), pcDNA-HA-Rac1(PBRM), and pcDNA-HA-RhoA-V14 (constitutively active) were purchased from The Missouri S & T cDNA Resource Center (Rolla, MO). Expression plasmids pcDNA-HA-Rac1(WT), pcDNA-HA-Rac1L61 (constitutively active), pcDNA-HA-Rac1L61/V30, and pcDNA-HA-Rac1L61/S35 were generated using PCR based site-directed mutagenesis. For creating point mutations in recombinant Rac or Rho proteins, the pcDNA-HA-Rac1 and Rho mutants already generated (as described above) were subcloned into the bacterial expression vector pET21-A. The bacterial expression construct for GST-FLAG-PP5(WT) was described previously (15). The inactive mutant of PP5, GST-FLAG-PP5Q304 (25) was generated using PCR based site-directed mutagenesis. This residue is thought to be critical for catalysis (26), and mutation of His 304 to Ala has previously been used as a catalytically inactive control in cell expression studies (27). Residue numbers refer to the rat form of PP5 and the human forms of Rac1 and RhoA. All newly generated mutants were confirmed by DNA sequence analysis. An expression vector encoding GST-PAK1 (CRIB domain) was a kind gift from John Collard (The Netherlands Cancer Institute, Amsterdam, The Netherlands).
Cell Culture and Transfection-HeLa cells obtained from ATCC were maintained in Dulbecco's modified Eagle's medium supplemented with 10% FBS (Clontech, Mountain View, CA), 1 mM sodium pyruvate, penicillin/streptomycin (100 units/ml, 100 g/ml) (Invitrogen) at 37°C under humidified conditions with 5% CO 2 . The cells were transiently transfected with the indicated expression vectors using Lipofectamine 2000 (Invitrogen) according to the manufacturer's instructions.
Immunoprecipitation-For immunoprecipitation of endogenous PP5 and Rac1, HeLa cells were serum-starved for 20 h and then treated with 20% FBS for 10 min at 37°C (28,29) to activate endogenous Rac1. The cells were then lysed in icecold lysis buffer (50 mM HEPES, pH 7.5, 50 mM NaCl, 5 mM MgCl 2 , 10% glycerol, 1% Triton X-100, protease inhibitor mixture (Roche Applied Science)) and then centrifuged at 10,000 ϫ g for 15 min at 4°C. Endogenous Rac1 complexes were immunoprecipitated from the clarified supernatant using Rac1 monoclonal antibody 23A8 and protein G-agarose (Sigma), whereas anti-mouse IgG antibody (Sigma) was used as a negative control. For overexpression of FLAG-or HA-tagged PP5(WT) only, the cells were transiently transfected for 24 h, and clarified lysates were prepared as described above. The samples were incubated with immobilized FLAG-M2 or HA antibody for 2 h at 4°C. The beads were then washed with lysis buffer, and bound proteins were eluted using FLAG or HA peptide (Sigma) and analyzed by immunoblotting. For cells overexpressing both PP5 and Rac1 or Rho proteins, the cells were co-transfected with FLAG-PP5(WT) or FLAG-PP5-TPR domain mutants (Glu 93 , Glu 93 /Ala 126 ) and either HA-Rac1(WT), HA-Rac1L61, HA-Rac1N17, HA-RhoA-V14, HA-Rac1L61/ V30, HA-Rac1L61/S35, or empty vector. Protein complexes were immunoprecipitated from clarified lysates using immobilized HA antibody and eluted with HA peptide for immunoblot analysis.
Immunoblotting-The proteins were separated by 10% SDS-PAGE and transferred electrophoretically onto nitrocellulose membranes (Bio-Rad). The membranes were blocked in Trisbuffered saline containing 3% nonfat dry milk (Bio-Rad) and then incubated with primary antibodies. Bound primary antibodies were detected using fluorophore-tagged secondary antibodies. Quantification of all immunoblots was done using the Odyssey infrared imaging system (Li-cor Biosciences, Lincoln, NE).
To prepare GST-Rac1 or Rho proteins, Escherichia coli (BL21) cells were transformed with expression plasmids and proteins purified as described by Self and Hall (31) with minor modifications. Bacterial cultures were grown at 37°C to an A 600 of 0.6 and then induced with 30 M ␤-D-1-thiogalactopyranoside for 14 h at room temperature. The cultures were centrifuged at 5000 ϫ g for 10 min at 4°C, and the pellet was resuspended in ice-cold lysis buffer (50 mM Tris, 150 mM NaCl, 5 mM MgCl 2 , 1 mM dithiothreitol, 1 g/ml each of aprotonin, leupeptin, and pepstatin, pH 7.5 at 4°C), sonicated on ice, and centrifuged at 9000 ϫ g for 15 min at 4°C. The supernatant was incubated with glutathione-agarose beads (Sigma) for 1 h at 4°C. Immobilized GST-Rac1 or RhoA proteins were then washed with lysis buffer and stored at 4°C for in vitro binding assays or thrombin-cleaved (31) for use in phosphatase assays. The concentrations of purified recombinant proteins were estimated using the Lowry protein assay (Bio-Rad).
In Vitro Binding Assay-For activation of GST-Rac1 proteins with GTP␥S (Millipore), in vitro loading was performed according to the manufacturer's instructions. In vitro binding of Rac1 to PP5 was performed using GST-Rac1 proteins bound to glutathione-agarose and thrombin-cleaved soluble FLAG-PP5(WT) (0.3 mol of PP5/mol of GST-Rac1 protein). The proteins were mixed in binding buffer (50 mM Tris, pH 7.5, 100 mM NaCl, 5 mM MgCl 2, 0.1% Triton X-100) for 30 min at 4°C. The samples were then briefly centrifuged, the supernatant was discarded, and settled beads were washed in buffer containing 50 mM Tris, pH 7.5, 100 mM NaCl, 5 mM MgCl 2 , and 1% Triton X-100. Bound proteins were boiled in concentrated SDS gel loading buffer and then analyzed by immunoblotting. GST-PAK1 (CRIB domain), treated in parallel with GTP␥S, was used as a negative control.
Activation of Endogenous or Heterologously Expressed Rac1-Cells transfected with FLAG-or HA-tagged PP5(WT) were serum-starved overnight (14 -16 h). To activate endogenous Rac1, the cells were then treated with 20% FBS for 10 min at 37°C (28,29). The cells were then washed with ice-cold PBS and collected for further analysis (immunoprecipitation, fractionation, or immunofluorescence assays). To determine the activation of endogenous Rac1, a GST-PAK binding assay was performed by incubating recombinant GST-PAK1 (CRIB domain) with extracts from untreated or FBS-treated cells, as described previously by Collard and co-workers (30). Samples were boiled in SDS gel loading buffer and analyzed by immunoblotting.
Cellular Fractionation Assays-For studies with heterologously expressed proteins, crude membrane and cytosolic fractions were prepared as described previously for Rac membrane localization by Schwartz and co-workers (32). Briefly, following transfection and treatment the cells were lysed in ice-cold buffer (20 mM Tris-HCl, pH 7.5, 150 mM NaCl, 10 mM MgCl 2 , protease inhibitor mixture). The cells were then sonicated briefly and centrifuged at 1000 ϫ g for 10 min at 4°C. The supernatant was collected and further centrifuged at 100,000 ϫ g for 60 min at 4°C. The supernatant (cytosolic fraction) was saved for analysis. The resulting pellet was washed twice with lysis buffer and then solubilized in lysis buffer containing 1% Triton X-100 for 10 min at 4°C. The sample was then centrifuged at 100,000 ϫ g for 1 h at 4°C. The supernatant represented the solubilized crude membrane fraction. To analyze the effect of Rac1 activation on PP5 localization, HeLa cells transiently transfected with FLAG-PP5(WT) and either HA-Rac1(WT) or HA-Rac1N17 were serum-starved overnight. The cells were then treated with 20% FBS at 37°C for 10 min to activate Rac1 (28, 29) and fractionated as described above.
For fractionation studies measuring endogenous Rac1 or PP5, the cells were serum-starved for 20 h followed by treatment with 20% FBS for 10 min. To extract membrane fractions, the ProteoExtract subcellular proteome extraction kit (Calbiochem) (33) was used as per the manufacturer's instructions. The cytoplasmic fractions, cell lysates, and crude membrane fractions were analyzed by immunoblotting.
Immunofluorescence Microscopy-Cells plated on coverslips were co-transfected with Rac1 and PP5 expression vectors. Twenty-four h post-transfection, the cells were fixed with 3.7% paraformaldehyde for 30 min. The cells were then permeabilized with 0.5% Triton X-100 in PBS for 15 min and incubated in PBS containing 3% bovine serum albumin and 1% goat serum to block nonspecific antibody binding. Then cells were incubated for 1 h at room temperature with HA(Y-11) and PP5 polyclonal antibody to probe for HA-Rac1 and PP5, respectively. Then cells were washed with PBS containing 1% Tween 20 (Sigma) for 15 min and incubated with Alexa Fluor secondary antibodies for 1 h. The coverslips were washed and mounted on glass slides using VECTASHIELD mounting medium containing 4Ј,6-diamidino-2-phenylindole dihydrochloride (Vector Laboratories, Burlingame, CA). Confocal images were acquired using a Radiance 2100 MP Rainbow (Bio-Rad) on a TE2000 inverted microscope (Nikon, Tokyo, Japan) using a 60 ϫ 1.4 NA oil immersion lens. The images were collected sequentially to avoid any possible bleed-through. The Alexa 594 was excited at 543 nm by the green HeNe laser, and the fluorescence emission above 570 nm in wavelength was collected. The Alexa 488 signal was excited using the 488-nm line of the 4-line argon laser, and the emission between 500 nm and 540 nm was collected. Multi-photon excitation for 4Ј,6-diamidino-2-phenylindole dihydrochloride was provided by the Mai Tai laser (Spectra-Physics, Mountain View, CA) at 750 nm, and the emission between 420 and 480 nm was collected. For each area of interest, a z-series was collected and LaserSharp 2000 software was used to create top-down projections and rotating projections.
Fluorescence-based Phosphatase Assays-In vitro phosphatase assays to determine the activation of PP5 by Rac1 were performed using recombinant Rac1 and PP5 proteins. For activation of recombinant Rac1 proteins with GTP␥S (Millipore) or GDP (Sigma) as negative control, loading was performed according to the manufacturer's instructions. The proteins were mixed (0.1 mol of PP5/mol of Rac1 protein) in reaction buffer (50 mM Tris, pH 7.5, 100 mM NaCl, 0.05% bovine serum albumin), and 100 M 6,8-difluoro-4-methylumbelliferyl phosphate (Invitrogen) (34) was added to initiate the reaction. The reactions were carried out at 37°C for 10 min using a fluorometer (FluoroMax 200); the samples were excited at 358 nm, and emission at 450 nm was monitored. As a positive control 250 M arachidonic acid (Calbiochem) was used to activate PP5(WT), whereas inactive PP5Q304, dominant negative Rac1N17, and constitutively active RhoA-V14 were used as negative controls. Similar reactions, as described above, were performed to analyze the effect of certain Rac1 mutants (Leu 61 /Val 30 and Leu 61 /Ser 35 ) on PP5 activity. Arachidonic acid was again used as a positive control in parallel reactions. To analyze the combined effect of Rac1 and arachidonic acid on PP5 activity, 40 M arachidonic acid was used along with constitutively active Rac1L61 (0.1 mol of PP5/mol of Rac1 protein).
Statistical Analyses-The quantified data are presented as the means Ϯ S.E. For statistical comparisons, either one-way analysis of variance followed by Tukey's test for multiple comparisons or Student's t test was used as indicated.

RESULTS
PP5 Interacts Selectively with Active Rac1-In a previous study, our electrophysiological analyses suggested that PP5 binds only active Rac1 (13). With heterologously expressed proteins in pituitary-derived GH 4 C 1 cells, we also demonstrated that constitutively active Rac1 can be co-immunoprecipitated with either PP5 or a chimera containing GFP and the PP5 TPR domain (13). In the present study we extended this observation by demonstrating that endogenous Rac1 and endogenous PP5 can be co-immunoprecipitated from HeLa cells (Fig. 1, A and B). Binding of PP5 requires the active form of Rac1. During serum starvation, Rac1 is in a quiescent GDP-bound state, and treatment with exogenous stimulants such as platelet-derived growth factor or FBS is known to activate Rac1 (5,28,29). Fig. 1C shows that treatment of serum-starved cells with 20% FBS elevated the level of endogenous Rac1 present in HA-PP5(WT) immunoprecipitates. Increased binding of Rac1 in cell extracts to GST-PAK1 (CRIB domain) confirmed that Rac1 was activated by treatment with 20% FBS (Fig. 1D) (28,29).
To further investigate the specificity of PP5 interaction with Rac1, we performed immunoprecipitation studies using HeLa cells transiently overexpressing FLAG-PP5(WT) along with various mutants of Rac1 or the related protein RhoA, which lacks the residues in Rac1 that are proposed to contact PP5 (13) (Fig. 2A). As reported by others (35,36) endogenous and heterologously expressed Rac proteins can sometimes be observed as a pair of closely migrating bands ( Fig. 2A; see also Figs. 4 and 8). Quantitation of immunoprecipitated protein levels indicated a significant increase in PP5 immunoprecipitated with constitutively active Rac1L61 as compared with the amount of PP5 binding to either Rac1(WT), dominant negative Rac1N17, a constitutively active mutant of RhoA, RhoA-V14, or the empty vector negative control (Fig. 2, A and B). Levels of protein expression were similar in all samples (Fig. 2, A and C). This result suggests that PP5 interacts preferentially and selectively with active Rac1.
PP5 Directly Binds Active Rac1-We next investigated whether the interaction between PP5 and Rac1 is direct by performing in vitro binding assays using recombinant Rac1 and FLAG-PP5 proteins. GST-Rac1 fusion proteins were treated with or without GTP␥S, followed by incubation with recombinant PP5. Binding of PP5 to untreated Rac1(WT) was similar to that with an unrelated GST-fusion protein, GST-PAK1 (CRIB domain), but was increased by treatment of Rac1(WT) with GTP␥S (Fig. 3). This was not a nonspecific effect of GTP␥S, because GST-PAK1 (CRIB domain) was also treated with GTP␥S. Constitutively active Rac1L61 showed a modest increase in PP5 binding relative to untreated Rac1(WT), whereas PP5 bound to GTP␥S-treated Rac1L61 was similar to that of GTP␥S-treated Rac1(WT). The increase in binding seen after GTP␥S treatment of Rac1L61 is likely due to the fact that constitutively active Rac1 expressed in bacteria is not fully loaded with GTP (37), and GTP␥S treatment saturated the remaining empty nucleotide-binding sites. These results indicate that PP5 interacts directly with Rac1 and further support the conclusion that PP5 binds to Rac1 in its active form.
PP5-Rac1 Interaction Is Mediated via Key Residues in the PP5 TPR Domain and in Rac1-Based on the similar folding pattern of the TPR domains of PP5 and the NADPH oxidase p67 subunit that binds Rac1, we had predicted that Lys 93 and Lys 126 within the TPR domain of PP5 would contact Rac1 (13). Fig. 4A shows that mutation of PP5 Lys 93 to Glu or combined mutation of PP5 Lys 93 to Glu and Lys 126 to Ala decreased co-immunoprecipitation with constitutively active Rac1L61. Likewise, based on the crystal structure of active Rac1 and the TPR domain of p67 (3), Gly 30 within Rac1 should contact the TPR domain of PP5. When Gly 30 was mutated to Val in Rac1L61, co-immunoprecipitation of Rac1L61/V30 with PP5(WT) was decreased to the same level as co-immunoprecipitation with the TPR mutant PP5E93, used as a negative control. Residue Thr 35 is a key residue in the effector domain of Rac1 required for binding every downstream target for Rac1 regulation described thus far (7). Mutation of Rac1 Thr 35 to Ser also dramatically decreased its co-immunoprecipitation with PP5(WT), similar to results seen with the TPR mutant PP5E93.

JOURNAL OF BIOLOGICAL CHEMISTRY 3875
shows that the expression levels of all mutant proteins were similar to that of PP5(WT) or active Rac1L61. These results are consistent with our model predicting that PP5 binds active Rac1 through its TPR domain in a manner similar to the interaction of Rac1 and the p67 subunit of NADPH oxidase (13).
Active Rac1 Translocates PP5 to the Plasma Membrane-Active Rac1 is known to localize at the plasma membrane (38), whereas PP5 is reported to be in the cytoplasm and nucleus (27,39,40). To examine whether PP5 localizes to the plasma membrane as a function of Rac1 activation, we performed immunofluorescence assays in HeLa cells with transiently transfected FLAG-tagged PP5 and HA-tagged Rac1 proteins. The cells were immunostained, and confocal images were captured as described under "Experimental Procedures." We observed that a portion of PP5 co-localized at the plasma membrane in the presence of constitutively active Rac1L61 (Fig. 5A), whereas dominant negative Rac1N17 failed to translocate PP5 to the plasma membrane (Fig. 5D). Rac1L61/S35 translocated to the membrane and caused similar morphological changes as Rac161L but did not promote PP5 localization to the membrane. This is consistent with the observation that RacL61/S35 did not co-immunoprecipitate PP5 and shows that localization of PP5 at the membrane in Fig. 5A is not simply caused by Rac-induced morphological changes near the membrane. Similar results were observed with RacL61/V30 (supplemental Fig. S1B). Fig. 5G shows that in the majority of cells co-expressing Rac1L61 and PP5, a portion of PP5 was translocated to the cell membrane. To determine whether activation of Rac1(WT) promotes PP5 translocation to the membrane, HeLa cells expressing FLAG-PP5(WT) along with HA-Rac1(WT) or the polybasic region mutant HA-Rac1(PBRM) were serum-starved and then treated with or without 20% FBS. Fig. 5 (B, C, and F) shows that FBS stimulation increased the number of cells with membraneassociated PP5 when HA-Rac1(WT) was co-expressed, compared with HA-Rac1(PBRM) (Fig. 5F and supplemental Fig. S1A), the polybasic region mutant that lacks the C-terminal residues essential for Rac1 association with membrane fractions (8,9). This indicates that FBSinduced translocation of PP5 to the cell membrane was specific for activation of Rac1. In the absence of heterologously expressed active Rac1, PP5 was localized diffusely throughout the cytoplasmic compartment, whereas treatment with 20% FBS caused a modest amount of PP5 translocation to the membrane (supplemental Fig. S1C). This is consistent with the observation that 20% serum treatment activates endogenous Rac1 (Fig. 1D). Fig. 5G shows that in the majority of cells co-expressing Rac1L61 and PP5, a portion of PP5 was translocated to the cell membrane, compared with cells expressing Rac1N17, Rac1L61/S35, or Rac1L61/V30. The proportion of cells with membrane-localized PP5 appeared lower in the case of co-expression with Rac1N17, compared with cells with Rac1L61/S35 or Rac1L61/V30. This is likely due to the fact that Rac1N17 is a dominant negative mutant preventing any activation of endogenous Rac1.
PP5 Is Present in Membrane Fractions of Cells with Activated Rac1-As a second approach to determine whether PP5 is translocated to the membrane as a function of Rac1 activation, we asked whether PP5, like Rac1, is elevated in membrane fractions following FBS stimulation. Fig. 6 shows that treatment of HeLa cells with 20% FBS significantly increased To determine whether PP5 translocation to the membrane fraction specifically requires Rac1 activation, we examined the effect of FBS stimulation in cells overexpressing FLAG-PP5 together with either HA-tagged Rac1(WT) or the dominant negative mutant Rac1N17. Compared with cells with WT-Rac1 treated with FBS, lower levels of PP5 were present in the membrane fractions of cells expressing dominant negative Rac1N17 (Fig. 7A). As expected, membrane association of Rac1(WT) in cells treated with FBS was higher than in serum-starved cells, whereas dominant negative Rac1N17 was undetectable in the membrane fractions, even with FBS treatment. The absence of Rac1 or PP5 in membrane fractions from cells expressing dominant negative Rac1N17 was not simply due to differences in protein expression, as shown by the levels of FLAG-PP5 or HA-Rac1 detected in cell lysates (Fig. 7B). The presence of a small fraction of Rac1(WT) in the membrane fraction of unstimulated cells may represent modest activation by residual growth factors or other stimulants that would not be expected to affect Rac1N17 (43)(44)(45).
The level of PP5 in the membrane fraction was significantly increased in cells expressing constitutively active Rac1L61 compared with that in cells expressing dominant negative Rac1N17, unstimulated Rac1(WT), or empty vector (Fig. 7, C  and F). In cells expressing constitutively active RhoA-V14, the level of membrane-associated PP5 was comparable with control (Fig. 7F), demonstrating that the response was specific for Rac1. Fig. 7 (D and G) shows that similar levels of PP5 and Rac1 or RhoA proteins were expressed in all cell samples. We detected the cytosolic marker protein glyceraldehyde-3-phosphate dehydrogenase (41,42) in the cytoplasmic but not the membrane fraction (Fig. 7E), indicating that our fractionation protocol was efficient, and the presence of PP5 in the membrane fraction was not simply due to nonspecific trapping.
Consistent with results of co-immunoprecipitation studies (Fig. 4), the mutation of TPR residues proposed to bind Rac1 reduced the PP5 levels in the membrane fraction from cells expressing active Rac1L61. Fig. 8A shows that membrane localization of PP5E93 or PP5E93/A126 was greatly reduced compared with PP5(WT) in cells expressing Rac1L61. Likewise, when constitutively active Rac1L61 contained mutations of residues critical for binding PP5 (Rac1L61/V30) or binding effectors (Rac1L61/S35), membrane localization of PP5(WT) was dramatically reduced compared with Rac1L61 (Fig. 8C). Fig. 8  (B and D) shows that equal levels of protein were expressed in all cell samples.
Activated Rac1 Stimulates PP5 Phosphatase Activity-To examine the effect of Rac1 on PP5 activity, we performed in vitro phosphatase assays using recombinant proteins with 6,8difluoro-4-methylumbelliferyl phosphate as substrate (34). When incubated with Rac1(WT), dominant negative Rac1N17, or constitutively active RhoA-V14, phosphatase activity was unchanged from the control activity of PP5 alone (Fig. 9A). In contrast, phosphatase activity was increased ϳ2-3-fold by constitutively active Rac1L61 or by Rac1(WT) pretreated with GTP␥S. Treatment of Rac1(WT) with GDP and treatment of Rac1N17 with either GTP␥S or GDP had no effect on the phos-

Regulation of Protein Phosphatase 5 by Rac1
FEBRUARY 5, 2010 • VOLUME 285 • NUMBER 6 phatase activity of PP5, showing that increased PP5 activity required the active form of Rac1 (data not shown). Catalytically inactive PP5Q304 was not stimulated by constitutively active Rac1L61, Rac1(WT), or arachidonic acid (supplemental Fig.   S2B), demonstrating that the increase in activity shown in Fig.  9A was due to PP5(WT). As a positive control, assays performed in parallel showed that under the assay conditions used, PP5(WT) was activated ϳ10 -15-fold by arachidonic acid, similar to previous reports (supplemental Fig. S2A) (15,20). The magnitude of Rac1 stimulation of PP5 is similar to that reported with G␣ 12 and G␣ 13 (46). Mutation of Gly 30 or Thr 35 within Rac1L61 prevented stimulation of PP5 (Fig. 9B). Thus, consistent with our co-immunoprecipitation results (Fig. 4) and immunofluorescence studies (Fig. 5, E and G, and supplemental Fig. S1B), mutation of residues 30 or 35 appears to disrupt functional interaction of Rac1L61 with PP5. When using a submaximal concentration of arachidonic acid (40 M) that modestly stimulates PP5, the combined effects of Rac1L61 and arachidonic acid on PP5 activity were synergistic (Fig. 9C).

DISCUSSION
Our study shows that active Rac1 binds and translocates PP5 to the plasma membrane and that Rac1 directly stimulates PP5 activity in vitro. This observation is consistent with our studies demonstrating that PP5 and Rac1 drive thyroid hormone modulation of Kv11.1 ion channels in GH 4 C 1 cells (13) and that Thr 895 is the phosphorylation site on the Kv11.1 protein for PP5-mediated regulation (14). Thus, by recruiting PP5 to the  membrane, activated Rac1 could promote dephosphorylation of Kv11.1 channel proteins and other targets in or near the membrane by three synergistic mechanisms: activating PP5 through direct binding, increasing the proximity of PP5 to its target substrate, and simultaneously increasing its proximity to membrane lipids that may further stimulate its activity. Active Rac1 (GTP-bound or constitutively active) is known to activate the NADPH oxidase in neutrophils by binding the TPR domain of the NADPH p67 subunit at the plasma membrane (2,3,47). The biochemical data reported here demonstrate that Rac1 has a similar effect on PP5. This conclusion is also supported by our previous physiological studies of Rac1-dependent Kv11.1 stimulation, in which the TPR domain of PP5 functioned as a dominant negative inhibitor (13). Because the secondary structures  (41,42) to determine the efficiency of cellular fractionation. F, quantitation of FLAG-PP5(WT) present in membrane fractions as a function of HA-Rac1 or HA-RhoA proteins. PP5 levels in each sample were normalized to cadherin in membrane fractions. G, quantitation of expression levels of FLAG-PP5(WT), HA-Rac1, and HA-RhoA proteins in cell lysates, normalized to cadherin. The results represent the means Ϯ S.E. (n ϭ 5 or 6). The statistical analysis was carried out using one-way analysis of variance followed by Tukey's test. *, p Ͻ 0.001. The membrane fractions and cell lysates were subjected to SDS-PAGE and Western blot analysis using FLAG and HA antibodies to detect FLAG-PP5 and HA-Rac1 proteins, respectively. Cadherin was used as the loading control. A, membrane fractions from cells co-transfected with constitutively active Rac1L61 and either FLAG-PP5(WT) or FLAG-PP5 TPR domain mutants (Glu 93 and Glu 93 /Ala 126 ). B, levels of FLAG-PP5 and HA-Rac1 proteins in cell lysates. C, membrane fractions from cells co-transfected with FLAG-PP5(WT) or TPR domain mutants (a single point mutant (E93) or a double point mutant (E93/ A126)) and either Rac1L61, Rac1L61/V30 predicted not to interact with PP5 (13), or Rac1L61/S35, a Rac1 effector domain mutant (7). D, levels of FLAG-PP5 and HA-Rac1 proteins in cell lysates. The dashed lines indicate that a band was spliced out of the blot figure. Similar results were obtained in three independent experiments.
of the TPR domains of PP5 and p67 are nearly identical, we were able to predict two essential contact points between Rac1 and PP5 (13). Mutation of those two residues in the isolated TPR domain of PP5 prevented it from inhibiting channel activation by thyroid hormone and Rac1 (13). In contrast, overexpression of the full-length PP5 protein with the same two mutations in the TPR domain blocked thyroid hormone modulation of Kv11.1 channels (13). Our study confirms that mutation of Lys 93 within the TPR domain of PP5 decreases binding to Rac1L61-and Rac1-induced translocation to the membrane, as compared with wild type PP5. The findings that Rac1 binds both p67 (3) and PP5 via a TPR domain raise the possibility that other TPR-containing proteins may also be targets of Rac1.
Other GTPases have also been reported to recruit PP5 to membranes. Yamaguchi et al. (46) reported that constitutively active forms of the heterotrimeric G protein ␣ subunits, G␣ 12 and G␣ 13 , recruit PP5 to the plasma membrane in COS-7 cells. They also showed that active G␣ 12 and G␣ 13 subunits interact specifically with the TPR domain of PP5 and modestly stimulate PP5 activity in vitro (46). However, the residues in the switch 1 domain of Rac1 that we demonstrated to be essential for Rac1 signaling through PP5 (13) are not conserved in G␣ 12 and G␣ 13 subunits, and activated G␣ 13 has the opposite effect on Kv11.1 activity in GH 4 C 1 cells as activated Rac1 (11). We have observed that activated Rac1 can also produce a modest 2-3-fold increase in PP5 phosphatase activity in vitro, which is comparable with the stimulation by G␣ 12 and G␣ 13 reported by Yamaguchi et al. (46) but much less than the effect of exogenous lipids (15). It is possible that recruitment of PP5 in close proximity to a substrate like the Kv11.1 channel is sufficient to promote its dephosphorylation at physiologically significant rates. Alternatively, recruitment of PP5 to the membrane in close proximity to lipids or another protein partner at the membrane surface may further increase PP5 activity. The synergistic stimulation of PP5 by Rac1 at a submaximal concentration of arachidonic acid suggests that Rac together with lipid binding may indeed be more potent than Rac binding alone and that recruitment to the membrane may potentially serve two roles, enhancing Rac stimulation and increasing local concentrations of a physiological substrate. It will be important in future studies to examine these possibilities in extracts from stimulated cells.
The residue proposed to interact with PP5, Rac1(G30), is conserved in all three forms of Rac (13). Co-immunoprecipitation or membrane translocation of PP5 by active Rac was disrupted by mutation of Gly 30 and also by a mutation of Thr 35 . Mutation of residue Thr 35 in Rac is known to disrupt binding FIGURE 9. Effect of Rac1 on the phosphatase activity of PP5. In vitro fluorescence-based phosphatase assays were performed using recombinant PP5 and Rac1 or RhoA proteins as described under "Experimental Procedures." A, the activation of PP5 by Rac1 or RhoA proteins is represented as fluorescence intensity in arbitrary units/min/g PP5 on the y axis. Control represents the basal phosphatase activity of PP5(WT) in the absence of Rac1 or RhoA proteins. The results represent the means Ϯ S.E. (n ϭ 3). The statistical analysis was carried out using one-way analysis of variance followed by Tukey's test. *, p Ͻ 0.05. B, the effect of Rac1 mutants (Rac1L61/V30 predicted not to interact with PP5 (13) or Rac1L61/S35, a Rac1 effector domain mutant (7)) on PP5(WT) activity is represented as fluorescence intensity in arbitrary units/min/g PP5 on the y axis. C, the effect of Rac1L61, arachidonic acid (40 M), and Rac1L61 ϩ arachidonic acid (40 M) on PP5(WT) activity is represented as fluorescence intensity in arbitrary units/min/g PP5 on the y axis. All of the recombinant proteins used in phosphatase assays were thrombin-cleaved: Rac1QL, Rac1L61; Rac1WT, Rac1 wild type; Rac1WT(G), Rac1 wild type loaded with GTP␥S; RhoA-V14, RhoA-V14; PP5(WT), PP5 wild type.
with the p67 subunit of NADPH oxidase complex and its subsequent activation (47). Residue Thr 35 is also conserved in all forms of Rac and is required for interaction of Rac with all effectors known to date (7). Thus, PP5 is a potential effector for Rac1, 2, and 3. It will be important in the future to determine whether PP5 participates in other Rac-mediated responses. For example, Rac1 signaling is implicated in several disease states, including breast cancer (48,49); however, the signal transduction pathways mediating Rac1 responses are not completely understood. Although little is known about the roles and regulation of PP5 in breast epithelial cells, this phosphatase has been reported to be elevated in breast cancer and is also implicated in the promotion of cell proliferation, primarily through inhibition of pathways initiating cell cycle arrest and apoptosis (50 -52).
Finally, Kv11.1 potassium channels are essential for the rhythmic excitability of endocrine cells and cardiac myocytes (53,54), and polymorphisms in the human ERG1 gene that encodes Kv11.1 channels are associated with cardiac arrhythmias (54,55). For these reasons it will be important to examine the involvement of Rac and PP5 in the regulation of Kv11.1 channels in these tissues.