Radixin Stimulates Rac1 and Ca2+/Calmodulin-dependent Kinase, CaMKII

The ERM (ezrin, radixin, moesin) proteins function as cross-linkers between cell membrane and cytoskeleton by binding to membrane proteins via their N-terminal domain and to F-actin via their C-terminal domain. Previous studies from our laboratory have shown that the α-subunit of heterotrimeric G13 protein induces conformational activation of radixin via interaction with its N-terminal domain (Vaiskunaite, R., Adarichev, V., Furthmayr, H., Kozasa, T., Gudkov, A., and Voyno-Yasenetskaya, T. A. (2000) J. Biol. Chem. 275, 26206–26212). In the present study, we tested whether radixin can regulate Gα13-mediated signaling pathways. We determined the effects of the N-terminal domain (amino acids 1–318) and C-terminal domain (amino acids 319–583) of radixin on serum response element (SRE)-dependent gene transcription initiated by a constitutively activated Gα13Q226L. The N-terminal domain potentiated SRE activation induced by Gα13Q226L; RhoGDI inhibited this effect. Surprisingly, the C-terminal domain also stimulated the SRE-dependent gene transcription. When co-transfected with Gα13Q226L, the C-terminal domain of radixin synergistically stimulated the SRE activation; RhoGDI inhibited this effect. Using in vivo pull-down assays, we have determined that the C-terminal domain of radixin activated Rac1 but not RhoA or Cdc42 proteins. By contrast, Gα13Q226L activated RhoA but not Rac1 or Cdc42. We have also shown that both the C-terminal domain of radixin and Gα13Q226L can stimulate Ca2+/calmodulin-dependent kinase, CaMKII. Activated mutant that mimics the phosphorylated state of radixin (T564E) stimulated Rac1, induced the phosphorylation of CaMKII, and stimulated SRE-dependent gene transcription. Down-regulation of endogenous radixin using small interference RNA inhibited SRE-dependent gene transcription and phosphorylation of CaMKII induced by Gα13Q226L. Overall, our results indicated that radixin via its C-terminal domain mediates SRE-dependent gene transcription through activation of Rac1 and CaMKII. In addition, the radixin-CaMKII signaling pathway is involved in Gα13-mediated SRE-dependent gene transcription, suggesting that radixin could be involved in novel signaling pathway regulated by G13 protein.

tion, regulation of Na ϩ /H ϩ exchanger (NHE1), induction of mitogenesis and apoptosis, and regulation of the extracellular signal-regulated kinase and c-Jun N-terminal kinase pathways (4 -9). Identifying the Rho guanine nucleotide exchange factor, p115RhoGEF, as an effector for G 13 has contributed to the understanding of the some cellular events mediated by G 13 (10,11). For instance, activation of RhoA by G 13 is responsible for actin stress fiber formation, SRE 2 -dependent gene transcription, and NHE1 activity (4,12,13).
However, a number of signaling events regulated by G 13 are Rhoindependent. Thus, G 13 -mediated activation of big mitogen-activated protein kinase (also known as extracellular signal-regulated kinase 5) is Ras-and Rho-independent (14). Similarly, G 13 -mediated chloride conductance (15) and G 13 -mediated activation of protein kinase A (16) are both RhoA-independent. In order to identify novel putative G 13 effectors, our laboratory had used yeast two-hybrid screening and demonstrated that G 13 subunit interacts with cytoskeleton-associated protein radixin (1).
Radixin belongs to the conserved ERM (ezrin, radixin, moesin) protein family. ERM proteins play a role in multiple cell functions, such as cell shape maintenance, formation of microvilli, cell-cell adhesion, cell migration, membrane trafficking, and cell polarity (see reviews in Refs. [17][18][19][20]. ERM proteins consist of the high homology N-terminal FERM (band 4.1, ERM) domain and the C-terminal domain (21). Interaction between the N-terminal FERM domain and the C-terminal domain maintains a dormant inactive state of the ERM proteins. In an active open state, the N-terminal domain binds to the membrane proteins such as CD43, CD44, ICAM1-3, Na ϩ /H ϩ exchanger regulator (NHE3), the cystic fibrosis transmembrane conductance regulator, and the 2-adrenergic receptor, whereas C-terminal domain binds to F-actin. Therefore, ERM proteins were originally regarded as cross-linkers between cytoskeletal actin and the plasma membrane.
Recently, it was suggested that ERM proteins function as signaling molecules. Interaction of ezrin with the regulatory p85 subunit of phosphatidylinositol 3-kinase is required for phosphatidylinositol 3-kinasedependent cell survival (19). ERM proteins have been shown to interact with and to be involved in Syk-mediated tyrosine kinase signaling (22). Importantly, ERM proteins are involved in the regulation of the Rho pathway. Binding of the guanine nucleotide dissociation inhibitor (RhoGDI) by the N-terminal domain of radixin promotes Rho activation in vitro (2). Interestingly, phosphorylation of ERM at the C terminus by Rho kinase maintains radixin in a relaxed active state (23), suggesting that the ERM proteins can act both upstream and downstream of the Rho signaling pathway. The fact that G␣ 13 interacts with radixin * These studies were supported by National Institutes of Health Grants GM56159, GM65160, and HL06078 and by a grant from the American Heart Association (to T. A. V.-Y.). 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 U.S.C. Section 1734 solely to indicate this fact. 1  and induces the open active state of radixin, resulting in the increased binding for F-actin (1), leads us to propose that radixin may function as a signaling molecule for G␣ 13 . To test this hypothesis, we examined the role of radixin on G␣ 13 -mediated SRE-dependent gene transcription. Extracellular signals such as serum and lysophosphatidic acid cause Rho activation and subsequently actin dynamics change; thereafter, the decrease of free G-actin pool initiates the activation of SRE-dependent gene transcription (24,25). Other members of the small Rho GTPase family, Rac and Cdc42, also stimulate the SRE gene transcription (24). In addition, the increase in intracellular Ca 2ϩ leads to SRE-dependent gene transcription and is mediated by Ca 2ϩ /calmodulin-dependent kinases, CaMKII and CaMKIV (26). Here, we report that radixin via its C-terminal domain mediated SRE-dependent gene transcription through activation of Rac1 and CaMKII. We have also shown that radixin-CaMKII signaling is involved in G␣ 13 protein-mediated SRE-dependent gene transcription.

EXPERIMENTAL PROCEDURES
Materials-Full-length, N-terminal, and C-terminal domains of radixin have been described before (1). pGEX-2T containing rhotekin-Rho binding domain and pGEX4T3 containing p21-binding domain of PAK1 were provided by Dr. A. Schwartz and G. Bokoch (The Scripps Research Institute), respectively. The internal EE-tagged G 13 Q226L, the constitutively activated RhoA and Rac, and the dominant negative RhoA, Rac1, and Cdc42 were purchased from the Guthrie Research Institute (Sayre, PA). The dominant negative CaMKII (K42M) and dominant negative CaMKIV (dCTK75E) were gifts from Dr. M. Rosner (University of Chicago) and Dr. J. Xie (University of Manitoba), respectively. Genistein, piceataneriod, LFM-A13, BAPTA/AM, KN-93, and cyclosporine A were purchased from Calbiochem.
Monoclonal antibodies against RhoA, Rac1, Cdc42, polyclonal antibodies against total and phosphorylated CaMKII, and polyclonal antibody against radixin were from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). Monoclonal antibody against EE epitope was from Covance (Berkeley, CA).
Cell Culture and Transfection-NIH 3T3 fibroblasts were grown in Dulbecco's modified Eagle's medium supplemented with 10% heat-inactivated calf serum, 100 units/ml streptomycin, and 100 units/ml penicillin. Lipofectamine 2000 reagent (Invitrogen) was used for transfection following the manufacturer's instructions.
Radixin Down-regulation-Expression of the endogenous radixin was down-regulated by siRNA using the BD TM Knock-out RNAi Systems (BD Biosciences). The target sequence is the 19 nucleotides corresponding to the coding region between nucleotides 188 and 207 of the mouse radixin. The siRNA sequence contains the target sense sequence, hairpin loop, and target antisense sequence; the same sequence without the target antisense sequence was designed as control siRNA. The primers were chemically synthesized by Sigma-Genosys (The Woodlands, TX). After annealing, they were subcloned into the vector pSIREN-RetroQ (BD Biosciences). The sequences were confirmed by DNA sequencing.
Reporter Gene Assay-SRE-mediated gene expression was determined by the SRE.L reporter system (Stratagene) as described previously (27). Briefly, NIH 3T3 cells at 90% confluence grown on 24-well plates were transfected with the following plasmids (per well): 50 ng of pSRE.L reporter, 50 ng of pCMV-galactosidase (LacZ) (control plasmid for transfection efficiency), and 50 -200 ng of other plasmids indicated in each experiments as described under "Results." The cells were serumstarved overnight before assays. Washed with phosphate-buffered saline buffer, the cells were lysed and assayed following the manufactur-er's instruction for luciferase activity and ␤-galactosidase activity with the Promega assay kit (Promega, Madison, WI). Luciferase activity was normalized to the activity of ␤-galactosidase to correct the difference caused by different transfection efficiency.
In Vivo Rho GTPase Activation Assay-Activation of Rho proteins in vivo was determined by using pull-down assays as described by us earlier (28). These assays involve the use of glutathione S-transferase (GST) fusion proteins containing the GTP-dependent binding domains from effectors that bind the various Rho GTPases. Rhotekin interacts with GTP-bound RhoA but not with Rac1 or Cdc42. Conversely, the PAK serine/threonine kinase interacts with activated Rac1 and Cdc42 but not with RhoA. pGEX expression vectors encoding GST fusion proteins that contain the isolated GTP-dependent binding domains of the Rac1 and Cdc42 effector PAK1 (amino acids 70 -132 of PAK1; PAK PBD) or the RhoA effector rhotekin (amino acids 7-89 of rhotekin; rhotekin RBD) were used for the bacterial expression of GST fusion proteins. Confluent NIH 3T3 cells (100-mm dishes) were transfected with the indicated cDNAs for 48 h. Cells were serum-starved 24 h prior the experiment. Cells were then quickly washed with ice-cold Tris-buffered saline and lysed in lysis buffer (50 mM Tris, pH 7.4, 1% Triton X-100, 0.5% sodium deoxycholate, 0.1% SDS, 500 mM NaCl, 10 mM MgCl 2 , 10 g/ml each of aprotinin and leupeptin, and 1 mM phenylmethylsulfonyl fluoride. Cell lysates were clarified by centrifugation at 14,000 ϫ g at 4°C for 2 min, and equal volumes of cell lysates were incubated with GST-RBD rhotekin beads or GST-RBD PAK beads (15 g) at 4°C for 1 h. The beads were washed three times with wash buffer (50 mM Tris, pH 7.4, 1% Triton X-100, 150 mM NaCl, 10 mM MgCl 2 , 10 g/ml each of aprotinin and leupeptin, and 0.1 mM phenylmethylsulfonyl fluoride), and bound Rho, Rac1, or Cdc42 were eluted by boiling in Laemmli sample buffer. Samples eluted from the beads and the total cell lysates were then separated on 12.5% SDS-polyacrylamide gel electrophoresis, transferred to nitrocellulose, and analyzed by Western blotting using appropriate antibodies.
Western Blotting for Detection of Phosphorylation of CaMKII-Cells were washed with ice-cold phosphate-buffered saline, lysed in ice-cold radioimmune precipitation buffer (50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 1% Nonidet P-40, 0.5% sodium deoxycholate, 0.1% SDS, 10 l/ml protease inhibitor mixture), and sonicated on ice. The insoluble material was removed from the lysates by centrifugation at 14,000 ϫ g for 10 min. Protein concentration of the total cell lysates was determined using a Bio-Rad DC protein assay kit (Bio-Rad). 50 g of proteins were subjected to SDS-PAGE, transferred to a polyvinylidene difluoride membrane, and analyzed by immunoblotting with the appropriate antibodies.
Statistical Analysis-For statistical analysis, Student's t test was used to compare data between two groups. Values are expressed as mean Ϯ S.D. of three independent experiments. p Ͻ 0.05 was considered statistically significant.

Activation of SRE-dependent Gene Transcription by the C-terminal
Domain of Radixin-Rho proteins, including Rho, Rac, and Cdc42, are regulated by the guanine nucleotide dissociation inhibitors (GDIs), which inhibit the dissociation of the nucleotide bound to these proteins. The dissociation of GDI from the protein is a prerequisite for membrane association and activation of these Rho proteins by guanine nucleotide exchange factors (29). Based on the evidence that G␣ 13 interacts with and activates radixin (1) and that RhoGDI interacts with N-terminal domain of radixin, thereby displacing RhoGDI from Rho proteins in in vitro studies (2), it was reasonable to expect that domain(s) of radixin may regulate G␣ 13 -mediated activation of RhoA and subsequent cellular signaling.
We tested how expression of the N-terminal (amino acids 1-318), or C-terminal (amino acids 319 -583) domains of radixin affected the G␣ 13 -mediated SRE-dependent gene transcription. As shown in Fig. 1A, RhoGDI inhibited SRE activation induced by G␣ 13 Q226L expression, supporting the notion that Rho proteins mediate SRE activation induced by G␣ 13 Q226L. The N-terminal domain of radixin did not exhibit any basal activity compared with the empty vector. However, the N-terminal domain of radixin potentiated SRE activation induced by G␣ 13 Q226L. RhoGDI inhibited SRE activation induced by the combination of G␣ 13 Q226L and the N-terminal domain of radixin (Fig. 1A).
Interestingly, the C-terminal domain of radixin itself induced a 2-fold increase of the SRE-dependent gene transcription that was inhibited by RhoGDI (Fig. 1A). Surprisingly, when co-transfected with G␣ 13 Q226L, the C-terminal domain radixin synergistically stimulated the SRE activation (Fig. 1A). This activation was also inhibited by RhoGDI, suggesting that Rho proteins are involved in this signaling pathway.
To rule out the possibility that observed synergistic stimulation of SRE was due to changes in expression of transfected proteins, we analyzed whether the expression of G␣ 13 Q226L with internal EE epitope was affected by the C-terminal domain of radixin. Detection of G␣ 13 Q226L-EE with EE antibody showed that the C-terminal domain of radixin did not change the expression of G 13 Q226L-EE (Fig. 1B), suggesting that observed stimulation of SRE activity was not due to increased expression of G␣ 13 . G␣ 13 Q226L-EE showed the similar to G␣ 13 Q226L stimulation of SRE-dependent gene transcription (data not shown). Reprobing of the same blots with G␣ 13 antibody did not reveal a detectable difference of G␣ 13 expression in cells transfected with G␣ 13 Q226L-EE or vector alone, suggesting that overexpression of the transfected construct was minimal.
To further analyze the SRE-dependent gene transcription regulated by the C-terminal domain of radixin, we evaluated how different amounts of C-terminal domain of radixin cDNA affected G␣ 13 Q226Ldependent SRE activation (Fig. 1C). Data showed that the C-terminal domain of radixin stimulated SRE-dependent gene transcription in a dose-dependent manner both alone and in the presence of G␣ 13 Q226L (Fig. 1C).
Involvement of Rac1 in SRE Activation Induced by the C-terminal Domain of Radixin-To evaluate the effect of radixin on activation of individual Rho GTPases, we used the Rho-binding domain of the RhoA effector, rhotekin, to affinity-precipitate active RhoA, as a direct readout for RhoA activation. We have also used the Rac1-and Cdc42-binding domains of Rac1 and Cdc42 effector PAK to affinity-precipitate active Rac1 and Cdc42 as a direct readout for Rac1 and Cdc42 activation.
NIH 3T3 cells were transfected with either the C-terminal domain of radixin or G␣ 13 Q226L. Data showed that G␣ 13 Q226L induced a 5-6fold increase in RhoA activity ( Fig. 2A). Importantly, G␣ 13 Q226L did not activate Rac1 or Cdc42 ( Fig. 2A). The C-terminal domain of radixin expressed alone or in the presence of G␣ 13 Q226L did not affect RhoA activity ( Fig. 2A; data not shown). In addition, wild-type and N-terminal domain of radixin did not affect the activity of Rho proteins (data not shown).
By contrast, the C-terminal domain of radixin activated Rac1 but not RhoA and Cdc42 ( Fig. 2A). The C-terminal domain of radixin expressed in the presence of G␣13Q226L did not further enhanced Rac1 activity (data not shown). Equal protein expression of the C-terminal domain of radixin or G␣ 13 Q226L was controlled by Western blotting (data not shown), which confirmed that the different effects of these proteins on activation of Rho GTPases was not due to difference in the amount of expressed protein. These data provided the direct evidence that in mammalian cells, G␣ 13 Q226L stimulated RhoA, whereas the C-terminal domain of radixin stimulated Rac1 proteins.
To further support the observation that the C-terminal domain of radixin stimulated Rac1 protein, we have used dominant negative mutants of these GTPases. Data showed that dominant negative mutants of Rac1 (T17N) but not RhoA (T19N) or Cdc42 (T17N) inhibited SRE activation induced by the C-terminal domain of radixin (Fig.  2B). Thus, this result further corroborated the finding that C-terminal A, effect of G␣ 13 Q226L on SRE activity induced by N-terminal and C-terminal domains of radixin. NIH 3T3 cells were co-transfected with 50 ng of SRE.L luciferase reporter, 50 ng of ␤-galactosidase, 50 ng of RhoGDI, 50 ng of G␣ 13 Q226L, 100 ng of N-terminal domain, and the C-terminal domain of radixin as indicated. Empty vector pCDNA3.1 was used to make the final DNA amount equal in all transfections. The cells were incubated in serum-free medium for 18 h and then lysed for luciferase and ␤-galactosidase assays. Luciferase activity was normalized to ␤-galactosidase activity and was expressed as -fold stimulation from the basal level. Data represent mean Ϯ S.D. from three experiment performed in triplicate. *, p Ͻ 0.05, significant difference. B, expression of G␣ 13 Q226L in transfected cells. NIH 3T3 cells were transfected with G␣ 13 Q226L with internal EE tag, G␣ 13 Q226L-EE. The cell lysates were subjected to 10% SDS-PAGE. Expression of G␣13Q226L-EE was detected with a monoclonal antibody against EE tag. The total levels of G␣ 13 were detected with a polyclonal antibody against G␣ 13 . C, dose-dependent stimulation of SRE transcription by C-terminal domain of radixin. NIH 3T3 cells were co-transfected with 50 ng of SRE.L luciferase reporter, 50 ng of pCMV-␤gal, and the indicated amounts of the C-terminal domain of radixin in the absence or presence of 50 ng of G␣ 13 Q226L. The luciferase activity was determined as described above. Data represent mean Ϯ S.D. from one of at least three experiment performed in triplicate.
domain of radixin regulated activity of Rac1 but not RhoA or Cdc42 proteins.
In addition, the data showed that both C3 toxin and dominant negative Rac1(T17N) inhibited G␣ 13 -induced SRE activation (Fig. 2C). The different sensitivities of reporter assay and pull-down assay may have contributed to the differences in the data.
Involvement of CaMKII in SRE Activation Induced by C-terminal Domain of Radixin and G␣ 13 Q226L-Partial inhibition by dominant negative Rac1 of the SRE activation induced by the C-terminal domain of radixin suggested that an additional pathway involved in SRE activation might exist. Since some tyrosine kinases and two isoforms of Ca 2ϩ / calmodulin-dependent kinase, CaMKII and CaMKIV, have been shown to stimulate SRE gene transcription (26, 30 -35), we determined how inhibitors of tyrosine kinases and inhibitors of CaMK affect the SRE activation induced by the C-terminal domain of radixin. At concentrations chosen for the activity against the respective target enzymes, gen-eral tyrosine kinase inhibitor genistein (50 M) and selective Syk tyrosine kinase inhibitor piceataneriod (10 M) did not affect the transcriptional activity (data not shown), suggesting that tyrosine kinases are probably not involved in the SRE activation induced by the C-terminal domain of radixin.
To test the involvement of the Ca 2ϩ /calmodulin-dependent kinase pathway, we examined how intracellular calcium chelator (BAPTA/ AM), calmodulin kinase inhibitor (KN-93), and calcineurin inhibitor (cyclosporine) affect SRE activation induced by the C-terminal domain of radixin. Our results showed that BAPTA/AM and KN-93 inhibited SRE activation (Fig. 3A), suggesting that the Ca 2ϩ /calmodulin-dependent kinase pathway may be involved in the regulation of SRE gene transcription induced by the C-terminal domain of radixin.
Because both CaMKII and CaMKIV stimulate SRE gene transcription and KN-93 inhibits the activity of both kinases, next we examined whether the dominant negative mutants of CaMKII and CaMIV can affect the SRE activation induced by the C-terminal domain of radixin. As shown in Fig. 3B, the dominant negative mutant of CaMKII but not of CaMKIV inhibited the SRE activation induced by the C-terminal domain of radixin.
Because dominant negative Rac1 induced partial inhibition of SRE activity (Fig. 2B), we tested whether simultaneous inhibition of Rac1 and CaMKII inputs will further decrease the SRE activity induced by the C-terminal domain of radixin. Data showed that both dominant negative Rac1 and kinase-dead CaMKII partially inhibited SRE activity induced by C-terminal domain of radixin (Fig. 3C). Simultaneous inhibition of Rac1 and CaMKII by dominant negative constructs further inhibited SRE activity (Fig. 3C), suggesting that both inputs contributed to the SRE activity induced by the C-terminal domain of radixin.
We tested the involvement of the Ca 2ϩ /calmodulin-dependent kinase pathway in Ga 13 -mediated signaling. Our results showed that BAPTA/AM and KN-93 inhibited SRE activation (Fig. 4A), suggesting that Ca 2ϩ /calmodulin dependent kinase pathway may be involved in the regulation of SRE gene transcription induced by the G␣ 13 protein. Interestingly, dominant negative mutants of CaMKII and CaMKIV inhibited the SRE activation induced by G␣ 13 Q226L (Fig. 4A), suggesting that both CaMKII and CaMKIV are involved in G␣ 13 -mediated SRE-dependent gene transcription.
The phosphorylation state of CaMKII was analyzed using specific antibody targeted to the autophosphorylation site of CaMKII at residue of Thr 286 . We observed significant increase of the endogenous CaMKII phosphorylation (Fig. 4, B and C) in cells transfected with the C-terminal domain of radixin, whereas total CaMKII expression remained constant. Wild-type radixin or its N-terminal domain did not induce CaMKII phosphorylation (data not shown). Importantly, G␣ 13 Q226L also induced CaMKII phosphorylation (Fig. 4, B and C). Co-transfection of C-terminal domain of radixin together with G␣ 13 Q226L resulted in a further increase of CaMKII phosphorylation. These data provided evidence that both the C-terminal domain of radixin and G␣ 13 Q226L stimulated CaMKII.

Radixin Mutant Mimicking Phosphorylated State Activates Rac1 and Induces CaMKII Phosphorylation and SRE-dependent Gene
Transcription-Activation of radixin requires relief of the intramolecular association, and this is believed to involve phosphorylation of threonine 564 (17). To determine whether the C-terminal radixin accurately reflects an activated state of radixin, we made the mutant T564E radixin, which mimics the phosphorylated state of radixin (17). Data showed that radixin (T564E) induced activation of Rac1 (Fig. 5A), phosphorylation of CaMKII (Fig. 5B), and SRE activation (Fig. 5C). Together, these A, NIH 3T3 cells were transfected with 5 g of G␣ 13 Q226L, C-terminal domain of radixin, or Cdc42V12 as indicated. In vivo Rho GTPase activation assay was performed as described under "Experimental Procedures." Shown are Western blots from a representative experiment that demonstrates the activation of RhoA in response to G␣ 13 Q226L but not to the C-terminal domain of radixin. By contrast, Rac1 is activated in response to the C-terminal domain of radixin but not to G␣ 13 Q226L. Cdc42 activity was not affected by either G␣ 13 Q226L or the C-terminal domain of radixin. Three additional experiments gave similar results. B, C-terminal domain of radixin activates SRE via Rac1 but not RhoA or Cdc42 proteins. NIH 3T3 cells seeded onto 24-well plates were transfected with 50 ng of pSRE.L, 50 ng of pCMV-␤gal, 100 ng of C-terminal domain of radixin, and 200 ng of dominant negative RhoA, Rac1, and Cdc42 as indicated. The luciferase activity was determined as described in Fig. 1. The presented data are the mean Ϯ S.D. from three independent experiments. *, p Ͻ 0.05, significant difference. C, inhibition of G␣ 13 Q226Lmediated SRE-dependent gene transcription by dominant negative Rac1 and C3 toxin. NIH 3T3 cells seeded onto 24-well plates were transfected with 50 ng of pSRE.L, 50 ng of pCMV-␤gal, 50 ng of G␣ 13 Q226L, and 100 ng of dominant negative Rac1 or 100 ng of C3 toxin plasmid. The luciferase activity was determined as described in the legend to Fig. 1. The presented data are the mean Ϯ S.D. from three independent experiments. *, p Ͻ 0.05, significant difference. results suggested that the C-terminal domain of radixin can reflect the activated state of radixin.
Involvement of Radixin in G␣ 13 -mediated SRE Gene Transcription-To test the involvement of radixin in G␣ 13 -mediated signaling, we used siRNA technology to reduce the expression of endogenous radixin in NIH 3T3 cells. Cells were transfected with constructs containing either radixin or control siRNA, and 24 h later cell lysates were analyzed using Western blotting with antibodies against radixin, ezrin, or moesin. Densitometry analysis showed that expression of the endogenous radixin was down-regulated by ϳ50% upon transfection of the NIH 3T3 cells with vector containing radixin siRNA (Fig. 6A). The transfection of the control siRNA  vector did not affect the expression of the endogenous radixin. However, expression of endogenous ezrin or moesin was not affected by either control or radixin siRNAs (Fig. 6A), suggesting that siRNA-induced downregulation of radixin was specific.
Next, we tested whether down-regulation of the endogenous radixin can affect G␣ 13 -mediated activation of SRE. Data showed that in the cells expressing vector containing radixin siRNA, SRE activity induced by G␣ 13 Q226L was reduced by ϳ50% (Fig. 6B). In the cells expressing control siRNA, G␣ 13 -induced SRE activation was not affected. Importantly, SRE activity induced by a downstream effector of G␣ 13 , p115RhoGEF, was not affected by down-regulation of radixin (Fig. 6B), indicating that the reduction of G␣ 13 Q226L-mediated SRE activation by radixin siRNA was specific. Furthermore, radixin siRNA but not control siRNA also reduced the phosphorylation of CaMKII by G␣ 13 Q226L (Fig. 6C). Taken together, these data suggested that radixin-CaMKII signaling is involved in the G␣ 13 -induced SRE gene transcription.

DISCUSSION
In the present study, we have demonstrated for the first time that radixin can activate Rac1 and induce phosphorylation of CaMKII. Activation of these proteins resulted in the stimulation of SRE-dependent gene transcription. We have also demonstrated that G␣ 13 can induce CaMKII phosphorylation. In addition, reducing endogenous radixin with siRNA demonstrated that the radixin-CaMKII signaling pathway is involved in G␣ 13 -mediated SRE-dependent gene transcription. The partial inhibition by dominant negative Rac1(T17N) suggests the possibility that the radixin-Rac1 signaling pathway is also involved in G␣ 13mediated SRE-dependent gene transcription. Taken together, our studies suggest that radixin could be involved in a novel signaling pathway FIGURE 5. Radixin (T564E) activates Rac1, induces the phosphorylation of CaMKII, and stimulates SRE-dependent gene transcription. A, NIH 3T3 cells transfected with 5 g of vector (Vect) or wild-type (WT) or T564E radixin as indicated. The pull-down assay was used to detected the activation of Rac1 as described under "Experimental Procedures." B, NIH 3T3 cells transfected with 5 g of vector or wild-type or T564E radixin as indicated. The phosphorylated and total CaMKIIs were detected as described in Fig. 4. C, NIH 3T3 cells were co-transfected with 50 ng of SRE.L luciferase reporter, 50 ng of ␤-galactosidase (LacZ), and 100 ng of empty vector pCDNA3 or of wild-type or T564E radixin, as indicated. The luciferase activity was determined as described in the legend to Fig. 1. Data represent mean Ϯ S.D. from three experiment performed in triplicates. *, p Ͻ 0.05, significant difference compared with the cells transfected with empty vector.

FIGURE 6. Radixin is involved in G␣ 13 -mediated SRE-dependent gene transcription.
A, down-regulation of radixin using siRNA in NIH 3T3 cells. NIH 3T3 cells grown on 60-mm dishes were transfected with 5 g of cDNAs of control and radixin siRNA. 24 h later, 50 g of total cell lysates from each sample was subjected to 4 -20% gradient SDS-PAGE and probed with antibodies against radixin, ezrin, or moesin as indicated. B, radixin siRNA inhibits G␣ 13 Q266L-induced but not p115RhoGEF-induced activation of SRE. NIH 3T3 cells were co-transfected with 50 ng of SRE.L luciferase reporter, 50 ng of ␤-galactosidase, 50 ng of G␣ 13 Q226L, 50 ng of p115RhoGEF, and 200 ng of radixin siRNA or control siRNA as indicated. 24 h later, cells were incubated in serum-free medium for an additional 12 h. Luciferase activity was measured in the cell extracts, normalized to ␤-galactosidase activity, and expressed as a -fold stimulation from the basal level. Data represent mean Ϯ S.D. from three experiments performed in triplicate. *, p Ͻ 0.05, significant difference. C, phosphorylation of CaMKII induced by G␣ 13 Q226L is inhibited by radixin siRNA. NIH 3T3 cells grown on 60-mm dishes were transfected with 1 g of G␣ 13 Q226L or empty vector pCDNA3 and 5 g of radixin siRNA or control siRNA, as indicated. 24 h later, 50 g of total cell lysates from each sample was subjected to 8% SDS-PAGE. Phosphorylated and total CaMKIIs were detected with appropriate antibodies. regulated by G␣ 13 protein. However, we cannot exclude the possibility that G␣ 13 also activates CaMKII in a radixin-independent manner (Fig. 7).
Activation of Rac1 by C-terminal Domain of Radixin-In contrast to the N-terminal domain, the C-terminal domain of radixin alone induced a 2-fold increase in SRE activity (Fig. 1). We explored the possible mechanism of this activation using pull-down assays for activation of Rho proteins. The results indicated that the C-terminal domain of radixin activated Rac1, but not RhoA or Cdc42 (Fig. 2). Because Rac1 can also stimulate SRE-dependent gene transcription, these data suggested that Rac1 might mediate the stimulation of SRE induced by the C-terminal domain of radixin. Inhibition of C-terminal domain-induced SRE activity by Rac1 (T17N) but not by RhoA (T19N) or Cdc42 (T17N) confirmed that the C-terminal domain of radixin stimulated SRE via Rac1 (Fig. 2B). Together, these data indicated that Rac1 is at least in part responsible for the stimulation of gene transcription by the C-terminal domain of radixin.
Activation of Rac1 by the C-terminal domain of radixin was consistent with the other studies. Overexpression of ezrin T567D, in which an aspartic acid mimics the constitutive threonine phosphorylation of the C terminus, resulted in increased membrane ruffling and lamellipodia (36). In addition, it was recently reported that ezrin T567D induced Rac1 activation in Madin-Darby canine kidney cells (37). Therefore, it seems that activated ERM proteins in an open relaxed state may activate Rac1. Our results indicated that the C-terminal domain of radixin mediated Rac1 activation.
Furthermore, ezrin was shown to be cleaved by calpain following the stimulation with phorbol 12-myristate 13-acetate (38). The physiological significance of this proteolytic cleavage is currently unclear, although it might be a mechanism of releasing N-and C-terminal domains.
How the C terminus of radixin activates Rac1 remains unknown. It is possible that the C-terminal domain of radixin may regulate Rac-specific guanine nucleotide exchange factors (GEFs). One candidate is Tiam, a specific GEF for Rac. Interestingly, CaMKII was shown to activate Tiam (39,40). Since the C-terminal domain of radixin also induced the autophosphorylation of CaMKII (Fig. 6), it is possible that CaMKII and Tiam are involved in the activation of Rac1. This hypothesis is currently being tested in the laboratory.
Activation of CaMKII by C-terminal Domain of Radixin-Because dominant negative Rac1 induced partial inhibition of SRE activity (Fig.  2B), we tested whether other signaling pathways are involved in the regulation of SRE activity induced by the C-terminal domain of radixin. We determined that intracellular calcium chelator BAPTA/AM and CaMK inhibitor KN-93 inhibited SRE activation induced by the C-terminal domain of radixin (Fig. 3). Furthermore, we demonstrated that the dominant negative mutant of CaMKII but not CaMKIV inhibited the SRE activity induced by the C-terminal domain of radixin (Fig. 3). Using the specific antibody against the autophosphorylation site at Thr 286 of CaMKII, we determined that both the C-terminal domain of radixin and G␣ 13 Q226L increased the autophosphorylation of CaMKII (Fig. 4). These results provided strong evidence indicating the involvement of CaMKII in SRE activity induced by the C-terminal domain of radixin.
Both CaMKII and CaMKIV have been shown to stimulate SRE activity (26). The mechanism for CaMKIV-mediated SRE-dependent gene transcription had been defined (35). Thus, CaMKIV phosphorylates transcriptional repressor histone deacetylase and dissociates it from serum response factor, removing this repression molecule from the transcription complex. Recently, CaMKII was shown to use the similar mechanism to stimulate the myocyte enhancer factor-2 transcription factor in neuron cells (41). It is conceivable that CaMK II might use a similar mechanism to activate SRE-dependent gene transcription.
G␣ 13 /Radixin Signaling Pathway-Radixin is involved in the signaling pathways regulated by of RhoA, phosphatidylinositol 3-kinase, and tyrosine kinases (17). We have previously reported that activated G␣ 13 , via the N-terminal domain, interacts with radixin and increases radixin binding to F-actin. In the present studies, we have shown that both N-terminal and C-terminal domains were involved in the G␣ 13dependent SRE activation (Fig. 1A). Down-regulation of endogenous radixin resulted in the inhibition of SRE activation induced by G␣ 13 Q226L (Fig. 6B). Although G␣ 13 Q226L did not activate Rac1, we showed that the radixin-CaMKII signaling is involved in G␣ 13 -stimulated SRE-dependent gene transcription. The lines of evidence include that (i) BAPTA/AM, KN-93, and dominant negative CaMKII inhibited SRE activation by both the C-terminal domain of radixin and G␣ 13 Q226L (Figs. 3 and 4), (ii) both the C-terminal domain of radixin and G␣ 13 Q226L induced the phosphorylation of CaMKII (Fig. 4), and (iii) down-regulation of radixin using siRNA reduced the phosphorylation of CaMKII induced by G␣ 13 Q226L (Fig. 6).
According to the proposed model, G␣ 13 should be able to activate Rac1 (Fig. 7). However, we did not observe that in the pull-down assay ( Fig. 2A). On the another hand, the dominant negative mutant of Rac1(T17N) partially inhibited the G␣ 13 Q226L-mediated SRE gene transcription in the reporter assay (Fig. 2C). This discrepancy could be due to several possibilities. (i) Rac1 activation by G␣ 13 Q226L is cell condition-dependent. Thus, it was reported that stably expressed G␣ 13 Q226L induced Rac1 activation in NIH 3T3 cells (42), suggesting that G␣ 13 Q226L may activate Rac1 under certain cellular conditions. (ii) The small inhibition of dominant negative mutant Rac1 also suggests that G␣ 13 may activate Rac1, but the activation may be not strong enough to detect it using a pull-down assay. One major reason for small inhibition is that G␣ 13 Q226L activates RhoA, which in turn dominantly stimulates SRE. Therefore, G␣ 13 Q226L may stimulate SRE via three pathways: 1) RhoA; 2) radixin-CaMKII; 3) radixin-Rac1. Considering the three parallel pathways and dominance of RhoA, the radixin-Rac1stimulated SRE should be relatively small. It should be pointed out that inhibition was significant. (iii) The pull-down assay to assess the direct activation of small G proteins is capable of detecting femtomolar amounts of the activated enzyme. By contrast, the quantitative reporter assay that uses firefly luciferase allows detection of subattomole amounts of enzyme (43). In addition, one of the dominant signalings of G␣ 13 Q226L is the activation of RhoA. It is known that activation of RhoA could inhibit activation of Rac1 (44). The degree of RhoA activation by G␣ 13 Q226L in different cell conditions may affect Rac1 activation by G␣ 13 Q226L in a dynamic manner. In other words, the Rac1 activation by G 13 Q226L could be transient, and this would be more difficult to detect by the pull-down assay. By contrast, interference with the function of endogenous Rac1 by the dominant negative mutant could be relatively stable after the mutant is expressed. (iv) While this paper was in revision, a study was published describing that G␣ 13 mediates reactive oxygen species production by Rac activation (45). Direct activation of Rac1 by G␣ 13 was not shown in this paper. However, the authors determined that an inhibitor of G␣ 12/13 , the G␣ 12/13 -specific regulator of G protein signaling domain of p115RhoGEF, inhibited angiotensin-induced activation of Rac1. (v) It is also possible that G␣ 13 Q226L does not activate Rac1 when transiently transfected in NIH 3T3 cells. If this were the case, G␣ 13 would not use radixin in the signaling. Instead, G␣ 13 Q226L can activate the downstream CaMKII in a radixin-independent manner. This means that G␣ 13 Q226L and radixin would independently regulate the phosphorylation and activity of CaMKII and then the downstream SRE gene transcription. Our data do not exclude this possibility. Taken together, it should be pointed out that the direct signaling from G␣ 13 to radixin is likely, but further evidence is needed (Fig. 7), and further studies are also needed to distinguish among the above possibilities.
In summary, our studies indicate that radixin stimulates SREdependent gene transcription. This activity requires activation of Rac1 and CaMKII by radixin via its C-terminal domain. Our results also suggest that the radixin-CaMKII signaling is involved in G␣ 13 -mediated SRE-dependent gene transcription.