Conformational Activation of Radixin by G13 Protein α Subunit

G13 protein, one of the heterotrimeric guanine nucleotide-binding proteins (G proteins), regulates diverse and complex cellular responses by transducing signals from the cell surface presumably involving more than one pathway. Yeast two-hybrid screening of a mouse brain cDNA library identified radixin, a member of the ERM family of three closely related proteins (ezrin, radixin, and moesin), as a protein that interacted with Gα13. Interaction between radixin and Gα13was confirmed by in vitro binding assay and by co-immunoprecipitation technique. Activated Gα13 induced conformational activation of radixin, as determined by binding of radixin to polymerized F-actin and by immunofluorescence in intact cells. Finally, two dominant negative mutants of radixin inhibited Gα13-induced focus formation of Rat-1 fibroblasts but did not affect Ras-induced focus formation. Our results identifying a new signaling pathway for Gα13 indicate that ERM proteins can be activated by and serve as effectors of heterotrimeric G proteins.

G 13 protein, one of the heterotrimeric guanine nucleotide-binding proteins (G proteins), regulates diverse and complex cellular responses by transducing signals from the cell surface presumably involving more than one pathway. Yeast two-hybrid screening of a mouse brain cDNA library identified radixin, a member of the ERM family of three closely related proteins (ezrin, radixin, and moesin), as a protein that interacted with G␣ 13 . Interaction between radixin and G␣ 13 was confirmed by in vitro binding assay and by co-immunoprecipitation technique. Activated G␣ 13 induced conformational activation of radixin, as determined by binding of radixin to polymerized F-actin and by immunofluorescence in intact cells. Finally, two dominant negative mutants of radixin inhibited G␣ 13 -induced focus formation of Rat-1 fibroblasts but did not affect Ras-induced focus formation. Our results identifying a new signaling pathway for G␣ 13 indicate that ERM proteins can be activated by and serve as effectors of heterotrimeric G proteins.
G 13 protein, one of the heterotrimeric guanine nucleotidebinding proteins (G proteins) 1 regulates diverse and complex cellular responses by transducing signals from the cell surface presumably involving more than one pathway. G␣ 13 regulates Na ϩ /H ϩ exchanger activity (1), regulates the extracellular signal-regulated kinase (2,3) and c-Jun NH 2 -terminal kinase pathways (3,4), participates in embryonic development (5), and promotes assembly of actin stress fibers (6). It also induces mitogenesis and neoplastic transformation (2,7) and apoptosis (8,9). Changes in the organization of the actin cytoskeleton initiated by G␣ 13 are RhoA-dependent (3,6,8), suggesting this small G protein is acting in downstream signaling. Recently, the exchange factor for RhoA, p115RhoGEF, was shown to act as a GTPase-activating protein for G␣ 13 (10). However, only some G 13 -dependent cell responses can be explained by activation of the Rho-dependent pathway. For example, constitutively activated G 13 is a very potent oncogene (2,11), whereas constitutively activated Rho is a weak oncogene (12), which indicates that G 13 may use Rho-independent signaling pathways to transduce mitogenic signal. Similarly, G 13 stimulates activity of the sodium/proton exchanger, NHE1, in both Rhodependent and -independent manner (13).
Here we show that radixin, a member of the ERM family of proteins, interacts with G␣ 13 . This was determined by yeast two-hybrid system, in vitro binding, and co-immunoprecipitation technique. Moreover, activated G␣ 13 induced conformational activation of radixin. Finally, radixin mediated G␣ 13induced neoplastic transformation. Our results identifying a new role for G␣ 13 indicate that ERM proteins can be activated by and serve as effectors of heterotrimeric G proteins.

EXPERIMENTAL PROCEDURES
Yeast Two-hybrid Studies-The yeast two-hybrid MATCHMAKER LexA system (CLONTECH) was used for detecting specific proteinprotein interactions (14). Plasmid pLexA-G␣ 13 , containing the gene encoding G␣ 13 , was constructed by cloning into polylinker of plasmid pLexA (in frame with the LexA coding region) of polymerase chain reaction product synthesized from pcDNA1-G␣ 13 . Mouse brain cDNA library has been screened with pLexA-G␣ 13 . Diploid yeast have been assayed for LEU2 and lacZ reporter gene activity, and G␣ 13 -specific interactor clones have been picked. The cDNAs and deduced sequences of clones testing positive were determined and analyzed.
Protein Purification, in Vitro Binding Assay, Immunoprecipitation, and Western Blotting-G-protein subunits were purified from recombinant baculovirus-infected Sf9 cells as described (10). To construct GST fusion proteins, DNA sequences corresponding to the indicated sequences of radixin complementary DNA were amplified by polymerase chain reaction and subcloned into vector pGEX-2T (Amersham Pharmacia Biotech). Each construct was confirmed by DNA sequencing. GST fusion proteins were expressed in DH5a cells and purified on glutathione-agarose beads. In vitro binding assay was performed as described (16). HA-tagged radixin and wild type or constitutively active G␣ 13 were transiently expressed in COS-7 cells. Cells were lysed in 20 mM Tris (pH 7.5), 1 mM dithiothreitol, 100 mM NaCl, 1 mM EGTA, 5 mM MgCl 2 , and 1% Triton X-100. Where indicated, AMF (50 M AlCl 3 , 10 mM MgCl 2 , and 5 mM NaF) was included. Lysates were normalized for protein concentration, and proteins were immunoprecipitated with anti-HA 12CA5 antibody and protein A-agarose for 16 h at 4°C. Immunoprecipitates were washed, separated by 8% SDS-polyacrylamide gel electrophoresis (SDS-PAGE), and immunoblotted with 12CA5 antibody (BabCo) or G␣ 13 antibody (Santa Cruz Biotechnology, Inc.).
In Vitro and in Vivo F-actin Co-sedimentation Assay-Actin-binding protein kit (Cytoskeleton, Inc.) was used according to manufacturer's instruction to determine the interaction of recombinant radixin with F-actin. To isolate total membrane and cytosol fractions, NIH3T3 cell lines stably expressing vector only or G␣ 13 Q226L were scraped and 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. resuspended in homogenization buffer containing 20 mM Tris-HCl (pH 8.0), 1 mM EDTA, 1 mM dithiothreitol, and protease inhibitors. Cells were homogenized using a 27-gauge needle, and nuclei were removed by centrifugation at 800 ϫ g for 10 min. Membrane and cytosol fractions were obtained by centrifugation for 1 h at 4°C at 150,000 ϫ g in Beckman Instruments ultracentrifuge. Membrane fraction was resuspended in homogenization buffer at 2 mg/ml protein concentration. Protein concentrations of cytosol fractions obtained from vector only and G␣ 13 Q226L-expressing cells were equilibrated and formation of F-actin was initiated by a 40-fold polymerization buffer containing 200 mM MgCl 2 , 4 M KCl, 100 mM ATP. The samples were incubated for 1 h at 37°C, and F-actin was pelleted by ultracentrifugation for 1 h at 4°C at 150,000 ϫ g. The F-actin-containing pellets were rinsed with homogenization buffer and resuspended in 200 l of the same buffer. Membrane, cytosol, and F-actin fractions were separated by 8% SDS-PAGE and immunoblotted with radixin and G␣ 13 antibodies (Santa Cruz Biotechnology, Inc.) and actin antibody (gift of Mark Rasenick).
Immunofluorescent and Confocal Microscopy-Cell were grown on gelatin-coated coverslips, serum-starved for 24 h, washed with phosphate-buffered saline, fixed with 4% paraformaldehyde, permeabilized with 0.1% Triton X-100, and blocked with 2% bovine serum albumin for 30 min. Thereafter, cells were incubated for 1 h at room temperature with 15 g/ml G␣ 13 or radixin (Santa Cruz Biotechnology, Inc.) polyclonal rabbit or goat antibodies, correspondingly. Following three washes in phosphate-buffered saline, fluorescein isothiocyanate-conjugated or tetramethyl rhodamine isothiocyanate-conjugated species-appropriate secondary antibodies (10 g/ml, Pierce) were added for an additional 30 min. Co-localization studies of G␣ 13 and radixin in NIH3T3 cells were performed using dual-wavelength laser scanning confocal microscopy with a Zeiss LSM 510 equipped with 40ϫ waterimmersion objective.
Focus Formation Assay-Focus formation assay was performed as described (2). Rat-1 cells (60-mm dishes in triplicate) were transfected with indicated cDNA constructs and cultured for 18 days in the presence of 5% calf serum. Foci were stained with Giemsa and counted in three independent experiments.

RESULTS AND DISCUSSION
G␣ 13 Interacts with Radixin in the Yeast Two-hybrid Screening-To further delineate components of signaling pathways that are important for apparently diverse cellular functions, we used yeast two-hybrid screening of a mouse brain cDNA library (total 4.3 ϫ 10 6 cDNAs) to search for proteins interacting with G␣ 13 . Using full-length mouse G␣ 13 as "bait," the screening yielded five positive clones, one of which was a fragment (171-583 amino acids) of radixin. To characterize this interaction of G␣ 13 with radixin further, pairs of hybrid plasmids containing full-length and domain fragments of radixin were transformed together with G␣ 13 into the yeast reporter strain EGY48 and co-transformants were selected on synthetic drop-out media lacking leucine at 30°C for 4 days. Growth on selective media and ␤-galactosidase activity indicated that G␣ 13 interacted only with the full-length protein and the amino-terminal (amino acid residues 1-318), but not carboxyl-terminal domain fragment of radixin (Fig. 1A). This places the site for interaction to the membrane-binding region of radixin.
G␣ 13 Interacts with Radixin in an Activation-dependent Manner-To confirm that amino-terminal domain of radixin interacted directly with G␣ 13 , we examine their binding using purified components in vitro. Purified G␣ 13 bound to purified GST-radixin-FL (full-length) and to GST-radixin-N (amino-terminal domain), whereas G␣ 13 binding to GST-radixin-C (carboxyl-terminal domain) was very limited (Fig. 1B); there was no detectable binding of G␣ 13 to control GST. Importantly, G␣ 13 bound to GST-radixin-FL and GST-radixin-N only in the presence of AlF 4 Ϫ , an activator of the G␣ subunit that promotes a conformation similar to that of the transition state for GTP hydrolysis (17), suggesting a potential role of radixin as a G␣ subunit effector protein. AlF 4 Ϫ -dependent interaction was also detected between GST-radixin-FL and G␣ 12 , but not G␣ i or G␣ q (data not shown).
We next demonstrated binding of G␣ 13 with radixin by coimmunoprecipitation technique (Fig. 1, C and D). Wild type, GDP-bound G␣ 13 or a constitutively activated GTP-bound mutant form of G␣ 13 (G␣ 13 Q226L) (1) were co-expressed with HA-tagged radixin in COS-7 cells (Fig. 1C). Antibodies against GST fusion proteins were pull-down by centrifugation with GST-agarose beads, and bound G␣ 13 was analyzed by SDS-PAGE and G␣ 13 antibodies. C, expression of HA-radixin and wild type G␣ 13 (G␣13wt) or constitutively activated G␣ 13 Q226L (G␣13QL) in COS-7 cells. Cell lysates were prepared in the absence or presence of AMF; equal amounts of protein from each sample were separated by SDS-PAGE and immunoblotted with anti-HA (12CA5) or G␣ 13 antibodies as indicated. D, co-immunoprecipitation of radixin and G␣ 13 with anti-HA epitope antibody. Cell lysates were prepared as described under "Experimental Procedures," and proteins were immunoprecipitated with anti-HA epitope and immunoblotted with anti-HA or G␣ 13 antibodies.
HA epitope co-immunoprecipitated appreciable amounts of G␣ 13 Q226L, and only trace amounts of wild type GDP-bound G␣ 13 (Fig. 1D). However, when AlF 4 Ϫ was added, a 2-3-fold increase in the amount of immunoprecipitated wild type G␣ 13 was detected (Fig. 1D). We did not detect interaction of endogenous G␣ 13 with HA-radixin (Fig. 1D), probably due to low abundance of G␣ 13 in most cell types. Taken together, these data indicate that G␣ 13 binds radixin in an activation-dependent manner.
We examined the subcellular distribution of G␣ 13 and radixin in NIH3T3 cells stably expressing G␣ 13 Q226L using laser scanning confocal microscopy. Optical sectioning (0.5-m-thick confocal sections) of antibody-labeled cells showed that there was a striking overlap of G␣ 13 and radixin (Fig. 2). Both proteins were co-localized at the plasma membrane, at the nuclear membrane, and in the form of small dots. This co-distribution is consistent with our observation that G␣ 13 and radixin form a complex.
G␣ 13 Induces Conformational Activation of Radixin-Radixin and other members of ERM family of membrane-actinlinking proteins exist in structurally different forms and conformational activation is required to expose otherwise masked interaction sites (15, 18 -20). The concept that ERM proteins are regulated by conformational changes has gained much support in last years (21). The conformational activation of ERM proteins results in unmasking of a high affinity binding site for F-actin in the COOH-terminal domain of ERM (20,22,23).
Therefore, to test whether interaction between G␣ 13 and radixin results in activation of radixin, we first examined the ability of radixin to bind F-actin, as a way of assessment of radixin's conformational activation. Unmasking of radixin's actin-binding site was determined using in vitro co-sedimentation assay (20, 24, 25) (Fig. 3A). Purified actin was polymerized in the presence of divalent cations and ATP, and after addition of purified radixin and AlF 4 Ϫ -treated G␣ 13 , F-actin, and any bound proteins were separated by ultracentrifugation. Western blotting showed that radixin alone did not co-sediment with F-actin. However, in the presence of activated G␣ 13 , about 50% of radixin was associated with F-actin (Fig. 3A), whereas addition of GDP-bound G␣ 13 did not significantly increase the association of radixin with F-actin (Fig. 3A). Additionally, G␣ 13 was equally distributed between soluble and pellet fractions in the presence of radixin but did not co-sediment with F-actin in the absence of radixin (Fig. 3B), suggesting that interaction with radixin may have caused the co-sedimentation of G␣ 13 with F-actin.
To corroborate these results, we determined whether G␣ 13 promoted binding of radixin to endogenous F-actin. We hypothesized that if G␣ 13 promotes conformational activation of radixin, then in cells expressing G␣ 13 Q226L larger fraction of endogenous radixin will be associated with F-actin. To test this hypothesis, the cytosol fractions of NIH3T3 cells expressing vector only or G␣ 13 Q226L were prepared by ultracentrifugation in low ionic strength buffer in the absence of divalent cations (see "Experimental Procedures"); under these conditions F-actin spontaneously depolymerizes (24). Actin polymerization was then induced in the cytosol fraction (25), and F-actin was separated from non-sedimentable forms of actin by ultracentrifugation. Western blotting detected radixin in comparable amounts in both control and G␣ 13 Q226L-expressing cells (Figs. 3C and 4D). After polymerization of actin and ultracentrifugation, the pellet obtained from control cells contained only trace amounts of radixin and G␣ 13 (Fig. 3C). Remarkably, the pellet from G␣ 13 Q226L-expressing cells contained considerable amounts of both radixin and G␣ 13 , suggest- ing that the two proteins co-sedimented with F-actin (Fig. 3C). To test for specificity, we probed cytosol and membrane fractions with antibodies against primarily membranebound G␣ q and cytosolic mitogen-activated protein kinase/ extracellular signal-regulated kinase kinase. Both proteins could not be detected in the F-actin pellet. 2 We finally examined whether activated G␣ 13 Q226L induced conformational activation of radixin in vivo. Another manifes-tation of the conformationally active state of radixin in cells is its redistribution to newly formed apical microvilli in the cells (26). Whether or not G␣ 13 is associated with a conformationally active state of radixin was tested using immunofluorescence microscopy of NIH3T3 fibroblast cell lines stably expressing either vector alone or the constitutively active G␣ 13 Q226L subunit. In these cells, G␣ 13 Q226L induced assembly of actin stress fibers (data not shown) and caused the appearance of numerous short apical membrane microvilli (Fig. 4, A and B); however, radixin did not localize along stress fibers but instead showed diffused staining (Fig. 4, A and B). Importantly, the Ϫ -G␣ 13 . Purified actin was polymerized as described under "Experimental Procedures" and incubated with purified radixin and G␣ 13 . Western blotting of soluble (S) and pellet (P) fractions after ultracentrifugation determined association of radixin with F-actin in the presence of G␣ 13 . B, G␣ 13 does not interact with F-actin in the absence of radixin. C, interaction of endogenous radixin with F-actin in G␣ 13 Q226L-expressing cell lysates. G␣ 13 increases co-sedimentation of radixin with F-actin. Total cytosol fractions were prepared as described under "Experimental Procedures." Protein concentrations of cytosol fractions obtained from vector only and G␣ 13 Q226L-expressing cells were adjusted to similar amounts, formation of F-actin was initiated by addition of 40ϫ polymerization buffer, and F-actin fractions were obtained as described under "Experimental Procedures." Equal amounts of proteins from original cytosol and F-actin fractions were separated by SDS-PAGE and immunoblotted with radixin, G␣ 13 , or actin antibodies.

FIG. 4. G␣ 13 induces conformational activation of radixin in vivo.
A, NIH3T3 cells expressing vector only. B, NIH3T3 cells expressing G␣ 13 Q226L. Staining of endogenous radixin showing appearance of numerous microspikes and microvilli and apparent increased intensity of radixin staining in B. Exposure time for both images was 500 ms, and the median brightness distribution was 20 Ϯ 9.5 and 143 Ϯ 21, for A and B, respectively. Median brightness distributions for 3 coverslips subjected to the same treatment were determined and averaged as described under "Experimental Procedures." C, expression of RhoN19 did not change distribution and increased intensity of radixin staining (compare with B). D, G␣ 13 does not induce up-regulation of radixin synthesis or redistribution of radixin between membrane and cytosol fractions. A membrane and cytosol fraction from equal amount of cells expressing vector-only or G␣ 13 Q226L were obtained as described under "Experimental Procedures"; proteins were separated by SDS-PAGE and immunoblotted with radixin antibody. apparent intensity of radixin staining in cells expressing G␣ 13 Q226L was dramatically increased (Fig. 4, A versus B), although the protein concentration of radixin remained unchanged in both cells lines as determined by Western blotting (Figs. 3C and 4D). This indicates that the increase in intensity of the immunofluorescence staining was not due to up-regulation of radixin expression. Similarly, radixin distribution between cytosol and membrane fractions did not differ in G␣ 13 Q226L-expressing cells when compared with vector-onlyexpressing cells (Fig. 4D). Therefore, the change in immunofluorescence staining was not due to recruitment of radixin from the cytosol to the membrane, but was the result of an increase in epitope availability for antibody due to a change in the conformational state of radixin (21). The G␣ 13 Q226L-induced unmasking of radixin epitope for antibodies was not dependent on RhoA, because dominant negative RhoA mutant did not affect immunofluorescence of radixin (Fig. 4, A-C), while inhibited G␣ 13 -induced actin polymerization, presumably mediated by RhoA (data not shown).
Because G␣ 13 functionally interacts with thrombin receptors (5), we next examined the intracellular distribution of radixin in serum-starved human microvascular endothelial cells in response to thrombin. The apparent distribution of endogenous radixin changed, most notably due to the increase in staining intensity and the appearance of intensely stained membrane protrusions (Fig. 5, A and B). Western blotting analysis did not detect changes in cytosol and membrane contents of radixin before and after the stimulation of the endothelial cells with thrombin (Fig. 5C), supporting the notion that G␣ 13 -dependent increase of apparent intensity and distribution of radixin staining resulted from a change in the conformational state of radixin. However, as thrombin receptors are coupled to multiple G proteins, we are currently investigating the possible involvement of individual G proteins in a thrombin-induced redistribution of radixin. Taken together, these data indicate that G␣ 13 directly induces conformational activation of radixin.
As ERM proteins are the substrates of serine/threonine and tyrosine kinases, phosphorylation may positively regulate their activities by stabilizing ERMs in conformationally "open" state (20,21). In Swiss 3T3 cells, mutationally activated Rho induces phosphorylation of ERM proteins (26). However, Rho kinase FIG. 6. Radixin inhibits focus formation induced by G␣ 13 in Rat-1 cells. Rat-1 cells were transfected with indicated cDNA constructs (amount of cDNA indicated in Table I). Eighteen days later foci were stained and counted. Experiment was performed three times with similar results. does not phosphorylate full-length radixin in a conformationally "inactive" state (20). Radixin can serve as a Rho kinase substrate mostly in conformationally "active" state in vitro; full-length radixin is a poor substrate for Rho kinase, whereas COOH-terminal domain of radixin can be readily phosphorylated by Rho kinase (20). In addition, recently it was shown that phosphatidyl inositol bisphosphate is involved in the RhoA-dependent activation of ERM proteins (27). Our data suggest that G␣ 13 -induced conformational changes of radixin may provide its availability as a substrate for Rho kinase, which also can be indirectly activated by G␣ 13 (30). This potentially suggests that G␣ 13 may recruit two independent signaling pathways resulting in "stabilized" activation radixin. Although this hypothesis is currently under investigation, here we demonstrate a novel mechanism of activation of radixin via direct interaction with G␣ 13 .
Radixin Inhibited G␣ 13 -but Not Ras-dependent Cell Transformation-To further investigate the physiological relevance of the interaction between G␣ 13 and radixin, we studied the modulation of G␣ 13 -dependent cellular responses by radixin. The recognized cellular effect of G␣ 13 is neoplastic transformation, which occurs by yet unknown mechanism (2,11). Therefore, we next examined the effect of radixin on cellular transformation. Activated G␣ 13 induced focus formation when transfected into Rat-1 cells (Fig. 6, Table I), which is consistent with previously published data (2,11). This cellular transformation was not Ras-or Rho-dependent, because it could not be inhibited by dominant negative Ras (RasN17) or dominant negative Rho (RhoN19) (Fig. 6, Table I). Under the same conditions, RasN17 was functionally active since it was able to inhibit Ras-induced focus formation (Fig. 6, Table I). Similarly, the functional activity of RhoN19 was also confirmed by its ability to inhibit Ras-induced focus formation (Fig. 6, Table I).
Overexpression of full-length radixin or its deletion mutants had no effect on basal rate of Rat1 transformation (Fig. 6, Table I). However, overexpression of both COOH-and NH 2terminal domains of radixin dramatically inhibited G␣ 13 -induced focus formation by 62% and 76%, respectively (Fig. 6, Table I). The COOH-and NH 2 -terminal domains were shown to exert dominant negative effect on radixin in different cellu-lar systems (18), presumably by competing with radixin's Facting binding site and membrane binding site, respectively. Thus, our data suggest that endogenous radixin mediates G␣ 13 -induced transformation, and its binding to both F-actin and to the membrane are necessary for this function. Importantly, the deletion mutants of radixin did not affect Rasinduced focus formation (Fig. 6, Table I). This is consistent with our data showing that G␣ 13 -induced transformation is not Rasdependent (Fig. 4, Table I), and this also indicates that Rasinduced transformation does not require radixin, but recruits a distinct signaling pathway. Overexpression of full-length radixin did not additionally enhance G␣ 13 -induced transformation, but rather slightly reduced the effects of both G␣ 13 Q226L and Ras, the significance of which is questionable. This suggests that the amounts of endogenous radixin in Rat1 cells are sufficient for mediating the effect of G␣ 13 Q226L.
The mechanism by which radixin mediates G␣ 13 -dependent focus formation is not yet understood. Recently ezrin, a protein closely related to radixin, has been shown to interact with p85 subunit of phosphatidylinositol (PI) 3-kinase and promote cell survival (28). Our preliminary data show that PI 3-kinase is also involved in G␣ 13 function, although the connection between PI 3-kinase and radixin is yet to be examined. Furthermore, the role of cortical cytoskeleton, which is obviously regulated by radixin at focal adhesions and other cell surface structures (21), could be important in G␣ 13 -induced transformation. Thus, the relative importance of one or the other pathways and the precise sequence of molecular events under different conditions of cell activation remain to be established.
In conclusion, we have shown that G␣ 13 directly interacts with radixin, a member of a distinct family of ERM proteins (Fig. 7). Interaction of G␣ 13 with radixin occurs in a G␣ 13activation-dependent manner and results in conformational activation of radixin. This represents a novel signaling pathway induced by G␣ 13 . Functionally, radixin mediates the G␣ 13dependent neoplastic transformation. To our knowledge, this is the first evidence that an ERM proteins can be activated by and serve as direct effectors of a heterotrimeric G protein. Finally, because ERM proteins and proteins containing homology to membrane-binding domain of ERM proteins (29), such as the brain tumor suppresser neurofibromatosis 2 gene product mer-  13 . Activated G␣ 13 binds to radixin and induces conformational activation of radixin. Radixin mediates G␣ 13 -dependent neoplastic transformation. G␣ 13 contributes to Rho activation via interaction with p115RhoGEF (10). Direct interactions are shown in solid lines. Indirect effects are in dashed lines. Potential cross-talk between radixin and RhoA pathways, which are currently under investigation, are in dotted lines.
lin, cell-cell-contact protein talin, cytoplasmic tyrosine phosphatase, regulate a variety of signal transduction pathways, the G␣ 13 -ERM protein link should provide a new means to extend G-protein signaling to a broad range of physiological processes.