Direct interaction of Rnd1 with FRS2 beta regulates Rnd1-induced down-regulation of RhoA activity and is involved in fibroblast growth factor-induced neurite outgrowth in PC12 cells.

The Rho family of small GTPases has been implicated in the reorganization of the actin cytoskeleton and subsequent morphological changes in various cells. Rnd1, a member of this family, has a low intrinsic GTPase activity and exerts antagonistic effects on RhoA signaling. However, how the activity of Rnd1 is regulated has not yet been elucidated. Here we have demonstrated that Rnd1 directly associates with FRS2alpha and FRS2beta, which are docking proteins of fibroblast growth factor (FGF) receptors and play important roles in the intracellular signals induced by FGFs. The interaction of FRS2beta with Rnd1 suppresses the inhibitory effect of Rnd1 on RhoA. Rnd1 binds to the COOH-terminal region of FRS2beta including tyrosine residues essential for the interaction with Shp2. When FGF receptor 1 is activated, it phosphorylates FRS2beta, recruits Shp2, and releases Rnd1 from FRS2beta. The liberated Rnd1 then inhibits RhoA activity. Furthermore, knockdown of Rnd1 by Rnd1-specific short interfering RNAs suppress the FGF-induced neurite outgrowth in PC12 cells. These results suggest that the activity of Rnd1 is regulated by FGF receptor through FRS2beta and that Rnd1 plays an important role in the FGF signaling during neurite outgrowth.

Rif, RhoH/TTF, RhoBTB (1 and 2), and Rnd (1, 2, and 3). Among them, the functions of Rho, Rac, and Cdc42 have been extensively characterized. In neuronal cells, activation of Rac and Cdc42 induces the formation of lamellipodia and filopodia of the growth cone and stimulation of neurite outgrowth. On the other hand, Rho activation induces the inhibition of neuritogenesis and neurite retraction. Recent studies have revealed involvement of Rho family GTPases in the downstream signaling pathways of a variety of neurotrophins and axon guidance molecules to regulate the actin cytoskeleton (2).
Fibroblast growth factors (FGF(s)) 1 constitute a large family of growth factors that influence a wide variety of biological processes such as angiogenesis, embryogenesis, differentiation, and proliferation depending on the cell type (3,4). In the nervous system, they have been shown to stimulate both the differentiation and survival of postmitotic cells as well as being proliferative factors for non-differentiated cells (5). FGFs mediate their pleiotropic responses by binding to and activating a family of receptor tyrosine kinases designated FGF receptors (FGFRs) 1-4 (6). Many of the cellular responses of FGFRs are mediated by the membrane-linked docking proteins, FRS2␣ and FRS2␤ (7)(8)(9)(10)(11). FRS2␣ and FRS2␤ contain myristyl anchors and phosphotyrosine-binding domains in their NH 2 termini and multiple tyrosine phosphorylation sites in their COOH termini that serve as binding sites for the adaptor protein Grb2 and the protein phosphatase Shp2, transducing signals to mitogen-activated protein kinase cascades (8,9,11). Concerning signaling to Rho family GTPases, recent studies indicate that basic FGF induces phosphorylation of the p85 ␤PIX, a guanine nucleotide exchange factor for Rac1/Cdc42 and that phosphorylated p85 ␤PIX mediates Rac1 activation and regulates cytoskeletal reorganization at growth cones (12,13). However, precise mechanisms of FGF signal transduction pathway to Rho family GTPases remain unclear.
The Rnd proteins, Rnd1, Rnd2, and Rnd3/RhoE, comprise a unique branch of Rho family GTPases that lack intrinsic GTPase activity and consequently remain constitutively active (14,15). Prior studies have suggested that Rnd1 has antagonistic effects on RhoA-regulated signaling pathways, and several downstream effectors have been identified, such as Socius and p190RhoGAP (14, 16 -18). However, it has been unclear how the activity of Rnd1 is regulated. Here we show that Rnd1 directly interacts with FRS2␣ and FRS2␤. The Rnd1-induced inactivation of RhoA is regulated by the FGF receptor through FRS2␤, and Rnd1 is involved in the FGF-induced neurite outgrowth. We propose a new role of Rnd1 in the FGF signaling pathway leading to RhoA inactivation and cytoskeletal rearrangements.
Yeast Two-hybrid Screening-A rat brain cDNA library fused to the GAL4 activation domain of the pACT2 vector (Clontech) was screened using pGBKT7/Rnd1 S229 as bait in the yeast strain AH109 according to the manufacturer's instructions. Interaction between the bait and library proteins activates transcription of the reporter genes HIS3, Ade, and lacZ. From 1.3 ϫ 10 7 transformants, 647 colonies grew on selective medium lacking histidine and adenine and were also positive for ␤-galactosidase activity. One of these, clone 2-107 was found to encode the carboxyl-terminal 247 amino acids of FRS2␤.
Cell Culture and Transfection-293T cells were cultured in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum, 4 mM glutamine, 100 units/ml penicillin, and 0.2 mg/ml streptomycin under humidified conditions in 95% air and 5% CO 2 at 37°C. PC12 cells were grown in Dulbecco's modified Eagle's medium containing 10% horse serum and 5% fetal bovine serum. Transient transfections were carried out with Lipofectamine PLUS (Invitrogen) for 293T cells or Lipofectamine 2000 (Invitrogen) for PC12 cells, according to the manufacturer's instructions.
Immunoblotting-Proteins were separated by 10 or 12.5% SDS-PAGE and were electrophoretically transferred onto a polyvinylidene difluoride membrane (Millipore Corp.). The membrane was blocked with 3% low fat milk in Tris-buffered saline (10 mM Tris-HCl, pH 7.5, 150 mM NaCl) and then incubated with primary antibodies. The primary antibodies were detected with horseradish peroxidase-conjugated secondary antibodies and the ECL detection kit (Amersham Biosciences).
In Vitro Binding Assays-All glutathione S-transferase (GST)-fused proteins were purified from Escherichia coli as described previously (17). Protein concentration was determined by comparing with bovine serum albumin standards after SDS-PAGE and by staining with Coomassie Brilliant Blue.
For pull-down assays, 293T cells transfected with HA-tagged Rho GTPases were rinsed once with phosphate-buffered saline and lysed with the ice-cold cell lysis buffer (20 mM Tris-HCl, pH 7.5, 2 mM MgCl 2 , 1 mM dithiothreitol, 0.2% Triton X-100, 1 mM phenylmethylsulfonyl fluoride, 10 g/ml aprotinin, and 10 g/ml leupeptin). Cell lysates were then centrifuged for 10 min at 16,000 ϫ g at 4°C. The supernatants were incubated for 10 min at 4°C with 10 g of GST fusion proteins and subsequently incubated with glutathione-Sepharose beads for 2 h at 4°C. After the beads were washed with the ice-cold cell lysis buffer, the bound proteins were eluted in Laemmli sample buffer and analyzed by SDS-PAGE and immunoblotting with anti-HA antibody.
To examine direct interaction between the FRS2␤ and Rnd1, dot blot assay was performed as described previously (21). Five g of GST and GST-fused FRS2␤ (amino acids 247-493) were spotted onto a nitrocellulose membrane and allowed to dry for 1 h at room temperature. The membrane was blocked with buffer A (50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 10 mM MgCl 2 , and 1 mM dithiothreitol) containing 5% low fat milk for 1 h at 4°C. The membrane was then incubated for 1 h at 4°C in buffer A containing 20 g of GST-fused Rnd1 or Rac1 preloaded with GTP␥S. The membrane was washed with buffer A and then incubated . Then they were immobilized on glutathione-Sepharose beads, and bound proteins and lysate input (Lysate, lane 1) were analyzed by immunoblotting with anti-HA antibody. B, GST or GST-fused FRS2␤ (amino acids 247-493) was spotted onto a nitrocellulose membrane. The membrane was incubated with GSTfused Rnd1 preloaded with GTP␥S. GST-Rac1 was used as a negative control. Associated GST-Rnd1 and GST-Rac1 were detected with anti-Rnd1 or anti-Rac1 antibody, respectively. C, 293T cells were transfected with expression vectors encoding HA-tagged Rnd1, RhoA V14 , or Cdc42 V12 together with Myc-tagged FRS2␣ or FRS2␤. The cell lysates were immunoprecipitated (IP) with anti-Myc antibody. The immunoprecipitates and total lysates were analyzed by immunoblotting with anti-HA or anti-Myc antibody.
with 3% low fat milk in Tris-buffered saline containing anti-Rnd1 or anti-Rac1 antibody. These antibodies were detected by using horseradish peroxidase-conjugated secondary antibody and the ECL detection kit.
Immunofluorescence Microscopy-PC12 cells were seeded onto round 13-mm glass coverslips coated with poly-D-lysine, and then they were transfected with expression vectors encoding GFP and HA-RhoA V14 . At 5 h after transfection, the medium was changed with Dulbecco's modified Eagle's medium containing 100 ng/ml acidic FGF (aFGF) (Sigma) and 5 g/ml heparin. At 48 h after treating with aFGF, cells on coverslips were fixed with 4% paraformaldehyde in phosphate-buffered saline for 15 min. Cells on coverslips were mounted in 90% glycerol containing 0.1% p-phenylenediamine dihydrochloride in phosphatebuffered saline. Alternatively, PC12 cells were transiently transfected with expression vectors encoding GFP and Rnd1-specific siRNA or GFP and Myc-tagged FRS2␤. At 48 h after transfection, they were stimu-lated with 100 ng/ml aFGF and 5 g/ml heparin. PC12 cells possessing one or more neurites longer than the diameter of the cell body were scored as positive. We scored more than 50 cells in each experiment.
Measurement of RhoA Activity-Measurement of RhoA activity was performed as described previously (22). Briefly, 293T cells were transfected with an expression vector encoding HA-tagged wild-type RhoA together with Myc-tagged FRS2␤ and GFP-tagged Rnd1. At 36 h after transfection, the cells were washed with ice-cold Tris-buffered saline, and lysed with the cell lysis buffer (50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 30 mM MgCl 2 , 0.5% Triton X-100, 0.5% sodium deoxycholate, 10% glycerol, 1 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride, 10 g/ml aprotinin, and 10 g/ml leupeptin) containing 20 g of GST-fused Rho-binding domain of mouse Rhotekin. The cell lysates were centrifuged for 5 min at 16,000 ϫ g at 4°C, and the supernatants were incubated with glutathione-Sepharose beads for 1 h at 4°C. The beads were washed with the lysis buffer, and bound proteins were analyzed by SDS-PAGE and immunoblotting.

Rnd1
Directly Interacts with FRS2␣ and FRS2␤-In an attempt to identify binding proteins of Rnd1, we employed a two-hybrid screening of a rat brain cDNA library with Rnd1 S229 , a mutant of Rnd1 lacking the carboxyl-terminal CAAX motif, as bait. Approximately 1.3 ϫ 10 7 clones were screened, and several positive clones were isolated. Sequence analyses revealed that one of them encoded the COOH-terminal region of rat FRS2␤. FRS2␣ and FRS2␤ function as mediators of signaling by FGFRs (7)(8)(9)(10)(11). To determine whether Rnd1 interacts with FRS2␣ and FRS2␤ in vitro, HA-tagged Rnd1 was expressed in 293T cells and pull-down assays were performed by using the purified GST-fused full-length of FRS2␣ and FRS2␤. As shown in Fig. 1A, GST-FRS2␣ and FRS2␤ interacted with Rnd1. Rnd2 also interacted with FRS2␣ and FRS2␤, but Rnd3 and constitutively active RhoA (RhoA V14 ), Rac1 (Rac1 V12 ), and Cdc42 (Cdc42 V12 ) did not, indicating that FRS2 proteins specifically interact with Rnd1 and Rnd2. A dot blot assay with GTP␥S-loaded purified GST-Rnd1 as a probe showed that the GTP-bound Rnd1 directly interacted with FRS2␤ (Fig. 1B).
To determine the region of FRS2␣ and FRS2␤ required for the interaction with Rnd1, various truncated forms of FRS2␣ and FRS2␤ were purified as GST fusion proteins, and their bindings to HA-tagged Rnd1 expressed in 293T cells were examined by a pull-down assay (Fig. 2). HA-tagged Rnd1 was precipitated by the GST-fused COOH-terminal region of FRS2␣ (FRS2␣-CT, amino acids 411-508) or FRS2␤ (FRS2␤-CT, amino acids 393-493) and the wild-type proteins (FRS2␣-WT and FRS2␤-WT) or the fragment corresponding to the original yeast clone (FRS2␤-CT*, amino acids 247-493), whereas it was not precipitated by the GST-fused FRS2␤ lacking CT region (FRS2␤-⌬CT, amino acids 2-392) or GST alone (Fig. 2, B and C). These results indicate that the binding site of Rnd1 is located at the COOH-terminal region of FRS2␣ and FRS2␤.
Tyrosine Phosphorylation of FRS2␤ Releases Rnd1 from FRS2␤ and Recruits Shp2-Upon FGF stimulation, FRS2␣ and FRS2␤ are phosphorylated by activated FGFRs and recruit Grb2 and Shp2. FRS2␤-CT contains two tyrosine phosphorylation sites, Tyr-418 and Tyr-456, and their phosphorylations are essential for the interaction with Shp2 (9, 11). Therefore, we next examined whether phosphorylation of FRS2 affects its interaction with Rnd1. 293T cells were cotransfected with Myctagged FRS2␤, HA-tagged Rnd1, and FGFR1, and the lysates were then immunoprecipitated with anti-Myc antibody. FRS2␤ was tyrosine phosphorylated when it was coexpressed with FGFR1 and bound to Shp2. In contrast, Rnd1 was dissociated from the phosphorylated FRS2␤ (Fig. 3A). The dissociation of Rnd1 from FRS2␤ was not observed when they were coexpressed with kinase-inactive FGFR1 (FGFR1-KD). We found that FRS2␣ also had Rnd1-releasing activity, although it was weak compared with that of FRS2␤ (data not shown). We generated FRS2␤ containing Y418F and Y456F substitutions in the Shp2-binding region (FRS2␤-2F) to examine whether tyrosine phosphorylation in this region is required for the dissociation of Rnd1 from FRS2␤. As observed in previous studies (9, 11), FRS2␤-2F was phosphorylated by FGFR1, probably at tyrosine residues in the Grb2-binding region. However, FRS2␤-2F did not interact with Shp2 in the presence of FGFR1. On the other hand, Rnd1 interacted with FRS2␤-2F in the absence and presence of FGFR1 (Fig. 3A). We further examined whether one or both phosphorylation sites of FRS2␤ are essential for the loss of Rnd1 binding by using each of single point mutants of FRS2␤ (FRS2␤-Phe-418 and FRS2␤-Phe-456). Both FRS2␤-Phe-418 and FRS2␤-Phe-456 interacted with Rnd1 in the absence of FGFR1 and partially released Rnd1 in the presence of FGFR1 (Fig. 3B). Taken together, these results suggest that the interaction of Rnd1 with FRS2␤ is regulated by the FGFR-induced phosphorylation of FRS2␤ and that both Tyr-418 and Tyr-456 are important for the regulation of Rnd1 binding.
The Activity of Rnd1 Is Regulated by Its Interaction with FRS2␤-Rnd1 has much lower affinity to GDP than GTP and no detectable intrinsic GTPase activity (14), and no GTPaseactivating protein or inhibitory binding protein specific for Rnd1 has been identified so far. Thus, it remains unknown how the cellular activities of Rnd1 are regulated. We next examined the possibility that FRS2␤ functions as an inhibitory binding protein for Rnd1 and that the activity of Rnd1 is regulated by FGFR-induced phosphorylation of FRS2␤. Several lines of evidence indicate that Rnd1 counteracts the biological functions of RhoA (14, 16 -18). Therefore, we measured the level of active GTP-bound RhoA in cells by using GST-fused Rho-binding domain of Rhotekin to precipitate GTP-bound RhoA from the cell lysates (23). HA-tagged wild-type RhoA was cotransfected with GFP-tagged Rnd1 in the absence and presence of Myctagged FRS2␤ and FGFR1 in 293T cells, and precipitated GTPbound RhoA was detected with anti-HA antibody. Expression of Rnd1 dramatically decreased RhoA activity in the cells. However, the down-regulation of RhoA activity by Rnd1 was significantly attenuated in the cells coexpressing Rnd1 and FRS2␤ (Fig. 4). On the other hand, expression of FGFR1 restored the Rnd1-induced inactivation of RhoA, whereas the expression of FGFR1-KD did not. Furthermore, coexpression of Rnd1 with FRS2␤-2F suppressed the Rnd1-induced down-reg- These results indicate that the Rnd1-induced down-regulation of RhoA activity is regulated by the interaction of Rnd1 with FRS2␤.
Knock-down of Rnd1 Suppresses the FGF-induced Neurite Outgrowth in PC12 Cells-FGFs play important roles in neurogenesis, axon growth, and differentiation (24), and previous reports showed that FGFs stimulate neurite extension in PC12 cells (8,25). Rho family GTPases regulate neuronal morphology through reorganizing the actin cytoskeleton (2,26,27). Among them, activation of RhoA has been known to stabilize cortical actin and prevents neurite outgrowth, whereas inactivation of RhoA promotes neurite outgrowth in PC12 cells (2,26). In addition, ectopic expression of Rnd1 in PC12 cells stimulates neurite extension (28). These observations led us to examine whether Rnd1 is involved in the FGF-induced neurite outgrowth in PC12 cells by inhibiting RhoA activity. Indeed, stimulation of PC12 cells with aFGF decreased GTP-bound RhoA in the cells (Fig. 5A). To determine whether the downregulation of RhoA activity is required for the FGF-induced neurite outgrowth, we transfected HA-tagged RhoA V14 and examined cell morphology after stimulation of the cells with aFGF. PC12 cells expressing GFP alone extend their neurites after stimulation with aFGF. In contrast, expression of HAtagged RhoA V14 completely blocked the aFGF-induced neurite extension (Fig. 5B). We next examined the involvement of Rnd1 in the FGF-induced neurite extension by expressing vectors encoding Rnd1-specific siRNAs. Three siRNAs targeting different coding regions of the rat Rnd1 cDNA were designed (Rnd1 siRNA-A, -B, and -C), and two of them (Rnd1 siRNA-A and -B) significantly reduced exogenously expressed Myc-tagged rat Rnd1 (Fig. 5C). However, they had no effect on the Rnd2 expression (data not shown). Similar to untransfected cells, PC12 cells coexpressing GFP and control siRNA or Rnd1 siRNA-C that had little effect on Rnd1 expression caused neurite outgrowth in response to aFGF. However, coexpression of GFP and Rnd1 siRNA-A or -B significantly blocked the aFGFinduced neurite outgrowth (Fig. 5D). We also examined whether expression of FRS2␤-2F has the same effect as the Rnd1 siRNA in decreasing neurite outgrowth in PC12 cells, because Rnd1 was not dissociated from FRS2␤-2F (Fig. 3A). However, we could not detect PC12 cells expressing the FRS2␤-2F mutant, although wild-type FRS2␤ was expressed in the same transfection condition. This may be because overexpression of FRS2␤-2F mutant was toxic to PC12 cells (data not shown). These results indicate that Rnd1 is involved in the FGF-induced neurite outgrowth in PC12 cells. DISCUSSION Most of Rho family GTPases cycle between inactive GDPbound and active GTP-bound forms. However, Rnd1 has been shown to be constitutively active, because it has substitutions at amino acid residues known to be important for GTP hydrolysis (14). Although several downstream proteins that interact with Rnd1 have been identified, it has not yet to be elucidated how its function is regulated. Our findings demonstrate an FRS2␤ phosphorylation-dependent mechanism for regulating Rnd1 function. We showed that FRS2␤ directly associates with Rnd1 and suppresses the activity of Rnd1. The binding region of Rnd1 is located at the COOH-terminal region of FRS2␤ including Tyr-418 and Tyr-456, which are essential for the interaction with the SH2 domain of Shp2. Phosphorylation of FRS2␤ by FGFR1 at Tyr-418 and Tyr-456 recruits Shp2 and instead releases Rnd1 from FRS2␤, and then dissociated Rnd1 inhibits RhoA. These results suggest that the activity of Rnd1 is regulated by FGFR through FRS2␤. A previous report showed that expression of Rnd3, another Rnd subfamily mem-ber that is also known to be a constitutively active GTPase, can be induced by expressing activated Raf in Madin-Darby canine kidney cells, but the expression of Rnd1 is not affected (29). Similarly, stimulation of fibroblasts with platelet-derived growth factor promotes the synthesis of Rnd3 protein (30). Our results showed that Rnd3 does not interact with FRS2␣ and FRS2␤. In this respect, these constitutively active GTPases may be differently regulated; the activity of Rnd1 is controlled by the interaction with inhibitory binding proteins such as FRS2␤, whereas Rnd3 is regulated at the level of transcription or translation.
Rho family GTPases are critical regulators of neurite outgrowth (2,26,27). Among them, Rac and Cdc42 activation are involved in promoting neurite outgrowth, whereas RhoA inhibits the outgrowth, and down-regulation of RhoA activity is required for the neurite outgrowth (2,26,27). It seems likely that Rho GTPases are also involved in FGF-induced neurite outgrowth. Indeed, a recent study showed that basic FGF induces Rac1 activation through p85 ␤PIX, which in turn regulates cytoskeletal reorganization at growth cones (12,13). We show here that FGFR-induced phosphorylation of FRS2␤ triggers the Rnd1-induced down-regulation of RhoA activity. The down-regulation of endogenous RhoA activity was observed in PC12 cells when they were stimulated with FGF. In addition, overexpression of constitutively active RhoA, or knock-down of endogenous Rnd1 by Rnd1-specific siRNA inhibited FGF-induced neurite outgrowth in PC12 cells. These observations suggest that down-regulation of RhoA is one of important signaling pathways in FGF-induced neurite outgrowth, and that Rnd1 is involved in this effect.
It has been reported that FGF-2 (basic FGF) increases functional excitatory synapses on cultured hippocampal neurons (31). Recently, we have shown that Rnd1 is highly expressed in hippocampus during the stage of synapse formation and plays a role in spine formation (20). In contrast, active Rho has been shown to strongly induce a loss of mature-shaped spines, whereas inhibition of RhoA activity resulted in elongated spine necks (32,33). Considering our present results and these observations, FGFs may enhance synapse formation by Rnd1induced inactivation of RhoA.
In conclusion, we identified FRS2␤ as an upstream regulator of Rnd1 and demonstrated that Rnd1 plays an important role for FGF-induced neurite outgrowth in PC12 cells. This study not only links FRS2␤ to Rnd1 but also contributes to the understanding of control mechanism for constitutively active GTPases.