Report

The Semaphorin 4D Receptor Plexin-B1 Is a GTPase Activating Protein for R-Ras

See allHide authors and affiliations

Science  06 Aug 2004:
Vol. 305, Issue 5685, pp. 862-865
DOI: 10.1126/science.1097545

Abstract

Plexins are cell surface receptors for semaphorin molecules, and their interaction governs cell adhesion and migration in a variety of tissues. We report that the Semaphorin 4D (Sema4D) receptor Plexin-B1 directly stimulates the intrinsic guanosine triphosphatase (GTPase) activity of R-Ras, a member of the Ras superfamily of small GTP-binding proteins that has been implicated in promoting cell adhesion and neurite outgrowth. This activity required the interaction of Plexin-B1 with Rnd1, a small GTP-binding protein of the Rho family. Down-regulation of R-Ras activity by the Plexin-B1-Rnd1 complex was essential for the Sema4D-induced growth cone collapse in hippocampal neurons. Thus, Plexin-B1 mediates Sema4D-induced repulsive axon guidance signaling by acting as a GTPase activating protein for R-Ras.

Plexins constitute a large family of transmembrane proteins that function as receptors for semaphorins (1). Semaphorins were originally identified as repulsive axonal guidance molecules in the developing nervous system, but have recently been shown to control cell adhesion and migration in other tissues (25). The cytoplasmic domain of plexin family members has been highly conserved during evolution and shares sequence similarity with Ras family–specific GTPase activating proteins (GAPs), especially R-Ras GAP (6, 7). These homologous domains in plexins contain two conserved arginine residues corresponding to those essential for the catalytic activity of common Ras GAPs (8); however, no GAP activity by plexins has been demonstrated so far. R-Ras has been shown to play a key role in cell adhesion by activating integrins and to promote cell migration and neurite outgrowth (912).

Rnd1, a constitutively active Rho family member (13), directly interacts with Plexin-B1, the receptor for Semaphorin 4D (Sema4D) (14). The Rnd1-binding region in Plexin-B1 is located between the two GAP-homologous domains (Fig. 1E). To evaluate whether the interaction of Rnd1 with Plexin-B1 could affect the GAP activity toward R-Ras, we examined the interaction between Plexin-B1 and R-Ras in the presence and the absence of Rnd1. Epitope (Myc)–tagged cytoplasmic domain of Plexin-B1 was expressed in COS-7 cells and used in a pull-down assay with purified glutathione S-transferase (GST)–fused R-Ras preloaded with guanosine 5′-diphosphate (GDP) or guanosine 5′-O-(3′-thiotriphosphate) (GTP-γS). No interaction was detected between R-Ras and the cytoplasmic domain of Plexin-B1 in the absence of Rnd1. In contrast, Plexin-B1 interacted with GTP-bound active R-Ras when it was coexpressed with Rnd1 (Fig. 1A). Rnd1 was also immunoprecipitated with GTP-bound R-Ras, indicating that Plexin-B1 interacted simultaneously with R-Ras and Rnd1. To determine whether the interaction between Plexin-B1 and R-Ras is direct, we purified the recombinant Rnd1, R-Ras, and Myc-tagged cytoplasmic domain of Plexin-B1 from Escherichia coli and used them in immunoprecipitation studies with an antibody against Myc. The cytoplasmic domain of Plexin-B1 interacted with GTP-bound R-Ras only in the presence of Rnd1 (Fig. 1B). A mutant of Plexin-B1 unable to interact with Rnd1 [Plexin-B1-GGA, in which Leu-Val-Pro is mutated to Gly-Gly-Ala at amino acids 1849 to 1851 (14)] did not bind to R-Ras. In addition, mutations in arginine residues in the Ras GAP–homologous domains conserved among Ras GAPs (Plexin-B1-RA, in which Arg is mutated to Ala at amino acids 1677, 1678, and 1984) abolished binding of Plexin-B1 to R-Ras, although Plexin-B1-RA could still bind to Rnd1 (Fig. 1B). Another Ras family member, H-Ras, did not interact with the Plexin-B1-Rnd1 complex (Fig. 1C). The Rnd1-dependent interaction between active R-Ras and Plexin-B1 also occurred in transfected COS-7 cells expressing Myc-tagged full-length Plexin-B1, constitutively active R-Ras (R-Ras-QL), and hemagglutinin (HA) epitope–tagged Rnd1 (Fig. 1D). Neither Plexin-B1-GGA nor Plexin-B1-RA immunoprecipitated R-Ras-QL (Fig. 1D). These results suggest that Plexin-B1 interacts directly and specifically with active R-Ras through its Ras GAP–homologous domains, and that this interaction requires the binding of Rnd1 to Plexin-B1 at the region between the two GAP-homologous domains.

Fig. 1.

Interaction between Plexin-B1 and R-Ras requires Rnd1. (A) Lysates from transfected COS-7 cells expressing the Myc-tagged cytoplasmic domain of Plexin-B1 (Myc-PlexB1-Cyt) and HA-tagged Rnd1 were used in a pull-down assay with GST-R-Ras preloaded with GDP or GTP-γS. Bound proteins and total cell lysates (Lysates) were analyzed by immunoblotting with antibodies against Myc and HA. (B and C) Recombinant Myc-PlexB1-Cyt, Rnd1, R-Ras, and H-Ras were purified from E. coli. Myc-PlexB1-Cyt and GTP-γS–loaded Rnd1 were incubated with R-Ras (B) or H-Ras (C) preloaded with GDP or GTP-γS, and then immunoprecipitated with an antibodyagainst Myc. Bound and total proteins were analyzed by immunoblotting with antibodies against R-Ras, Rnd1, and Myc. (D) Lysates from transfected COS-7 cells expressing the indicated proteins were immunoprecipitated with an antibody against Myc, and bound proteins were analyzed by immunoblotting with antibodies against R-Ras, HA, and Myc. (E) Plexin-B1 constructs. The Rnd1-binding region is indicated. Letters indicate specific amino acid residues within domains (A, Ala; F, Phe; G, Gly; L, Lys; P, Pro; R, Arg; V, Val). Numbers indicate amino acid position within the sequence.

To determine whether Plexin-B1 promotes the intrinsic GTPase activity of R-Ras in vitro, we incubated purified R-Ras preloaded with [γ-32P]GTP with the membrane fraction from transfected COS-7 cells expressing full-length Plexin-B1 and Rnd1. Expression of Plexin-B1 or Rnd1 alone had little effect on the intrinsic GTPase activity of R-Ras in the presence and the absence of Sema4D. In contrast, stimulation of cells with Sema4D increased hydrolysis of bound GTP when R-Ras was incubated with the membrane fraction from COS-7 cells expressing Plexin-B1 and Rnd1 (Fig. 2A). Expression of Rnd1 and Plexin-B1-GGA or Plexin-B1-RA did not stimulate the GTPase activity of R-Ras in the presence and the absence of Sema4D (Fig. 2A). Expression of Plexin-B1 constructs in the membrane fractions was similar, as verified by immunoblot analysis (15).

Fig. 2.

Plexin-B1 stimulates the intrinsic GTPase activityof R-Ras in a ligand-dependent manner. (A) Purified R-Ras preloaded with [γ-32P]GTP was incubated with the membrane fraction from COS-7 cells transfected with the indicated plasmids in the presence (solid lines) or the absence (dashed lines) of Sema4D at 25°C for the indicated times. The radioactivity bound to R-Ras was determined by a nitrocellulose filtration assay. Results are the means ± SEM of three independent experiments. (B) COS-7 cells transfected with expression plasmids encoding Myc-tagged Plexin-B1 (Myc-PlexB1), GFP-tagged Rnd1, and HA-tagged R-Ras were stimulated with Sema4D for the indicated times. The cell lysates were incubated with GST-fused Ras binding domain of c-Raf-1 (GST-RBD). Bound R-Ras and total cell lysates were analyzed by immunoblotting with antibodies against HA (R-Ras), Myc (Plexin-B1), and GFP (Rnd1). (C) COS-7 cells transfected with the indicated plasmids were stimulated with Sema4D for 5 min. The cell lysates were incubated with GST-RBD, and bound R-Ras and total cell lysates were analyzed by immunoblotting. Relative R-Ras activity was determined by the amount of R-Ras bound to GST-RBD normalized to the amount of R-Ras in cell lysates analyzed by NIH Image software (28). Results are the means ± SEM of three or five independent experiments.

To assess the GAP activity of Plexin-B1 in cells, we determined the presence of GTP-bound R-Ras in transfected COS-7 cells expressing Plexin-B1, Rnd1, and R-Ras by using a pull-down assay with a fusion protein containing GST and the Ras-binding domain of c-Raf-1, which selectively isolates active R-Ras (16). After stimulation of cells with Sema4D, GTP-bound R-Ras rapidly decreased to a minimum within 5 min (Fig. 2B). However, addition of Sema4D had no effect on GTP-bound R-Ras in the absence of Rnd1 or when Plexin-B1-GGA or Plexin-B1-RA was expressed (Fig. 2C). These data indicate that Plexin-B1 acts as a GAP toward R-Ras through the Ras GAP–homologous domains, and that the interaction of Rnd1 with Plexin-B1 and the ligand stimulation are essential for Plexin-B1 GAP activity in vitro and in cells.

It has been reported that Plexin-B1 binds to Rho-specific guanine nucleotide exchange factors PDZ-RhoGEF and LARG through the receptor's C-terminal PDZ domain–binding motif, thus activating RhoA (1721). However, mutations in the Ras GAP–homologous domains in Plexin-B1 had no effect on the ability of Plexin-B1 to activate RhoA (fig. S1). Furthermore, deletion of the PDZ domain–binding motif or expression of the dominant-negative form of PDZ-RhoGEF, both of which completely inhibited RhoA activation by Plexin-B1, had no effect on the R-Ras GAP activity (figs. S1 and S2). These results indicate that Plexin-B1 induces down-regulation of R-Ras activity separately from its activation of RhoA.

To address the biological role of the R-Ras GAP activity manifested by Plexin-B1, we used differentiated PC12 cells that express endogenous Plexin-B1 and exhibit neurite retraction in response to Sema4D (17). Stimulation of nerve growth factor (NGF)–differentiated PC12 cells with Sema4D decreased GTP-bound R-Ras in the cells (Fig. 3A). To determine whether the down-regulation of R-Ras activity is required for the Sema4D-induced neurite retraction, we used NGF to differentiate transfected PC12 cells expressing constitutively active R-Ras (R-Ras-QL), and then examined cell morphology after stimulation of the cells with Sema4D. Tips of neurites were detected by also expressing green fluorescent protein (GFP). PC12 cells expressing GFP alone retracted their neurites within 3 hours after stimulation with Sema4D. In contrast, expression of R-Ras-QL completely blocked the Sema4D-induced neurite retraction (Fig. 3B). To confirm that the Sema4D-induced neurite retraction depends on the GAP activity of Plexin-B1, we stimulated transfected PC12 cells expressing Plexin-B1-GGA or Plexin-B1-RA with Sema4D and compared their morphologies with that of cells expressing wild-type (WT) Plexin-B1. Similar to untransfected cells, differentiated PC12 cells expressing WT Plexin-B1 retracted their neurites in response to Sema4D. Expression of Plexin-B1-GGA or Plexin-B1-RA blocked the Sema4D-induced neurite retraction (Fig. 3C). In addition, expression of the myristoylated GAP domain of R-Ras GAP, which exhibits a specific GAP activity toward R-Ras (22), or knockdown of R-Ras expression by the use of an R-Ras–specific short interfering RNA (siRNA) expression vector, mimicked the Sema4D-induced morphological change (Fig. 3D and fig. S3). We also observed the involvement of Rnd1 in the Sema4D-induced neurite retraction with a Rnd1-specific siRNA expression vector that effectively reduced the amount of Rnd1 protein (fig. S4). These results indicate that the GAP activity exhibited by the Plexin-B1-Rnd1 complex toward R-Ras is essential for the Sema4D-induced neurite retraction. However, inhibition of RhoA signaling by expression of a mutant Plexin-B1 lacking the PDZ domain– binding motif also suppressed the Sema4D-induced neurite retraction (15), as reported previously (17). Thus, the Sema4D-induced neurite retraction requires both the R-Ras GAP activity and the PDZ-RhoGEF–mediated RhoA activation by Plexin-B1.

Fig. 3.

Down-regulation of R-Ras activity by Plexin-B1 is essential for the Sema4D-induced neurite retraction in PC12 cells. (A) PC12 cells were differentiated with NGF for 36 hours and then stimulated with mouse Sema4D for the indicated times. GTP-bound R-Ras isolated with GST-RBD was detected with an antibody against R-Ras. (B and C) PC12 cells were transfected with expression plasmids encoding R-RasQL (B) or various constructs of Plexin-B1 (C) together with an expression plasmid encoding GFP. They were differentiated with NGF for 36 hours and then treated with mouse (B) or human (C) Sema4D for 3 hours. The transfected cells are shown by the fluorescence of GFP. Cells with neurites were defined as cells that possessed at least one neurite longer than the twice the diameter of the cell body, and were scored as a percentage of the total number of transfected cells. Results are the means ± SEM of three independent experiments in which 50 cells were counted. (D) Differentiated PC12 cells were microinjected with an expression plasmid encoding GFP alone or together with a plasmid encoding myristoylated GAP domain of R-Ras GAP (Myr-R-Ras GAP) and cultured for 3 hours. Bar, 20 μm.

Sema4D induces growth cone collapse in primary hippocampal neurons from rat embryos (18). Plexin-B1 and Rnd1 are endogenously expressed in cultured hippocampal neurons (18, 23), and endogenous R-Ras protein was also detected by immunoblot analysis (15). To determine whether the GAP activity of Plexin-B1 toward R-Ras is involved in Sema4D-induced growth cone collapse in hippocampal neurons, we reduced expression of Rnd1 in rat hippocampal neurons by RNA interference (fig. S5). Similar to untransfected neurons, transfected neurons expressing GFP and control siRNA caused growth cone collapse in response to Sema4D. However, expression of GFP and the Rnd1 siRNA suppressed the Sema4D-induced growth cone collapse (Fig. 4A). Expression of R-Ras-QL also blocked Sema4D-induced growth cone collapse (Fig. 4B). In addition, expression of R-Ras siRNA promoted growth cone collapse (fig. S6). These results suggest that inactivation of R-Ras by the Plexin-B1-Rnd1 complex is necessary for the Sema4D-induced growth cone collapse in hippocampal neurons.

Fig. 4.

Down-regulation of R-Ras activity by Plexin-B1 is required for the Sema4D-induced growth cone collapse in rat hippocampal neurons. (A) Hippocampal neurons transfected with control or the Rnd1-specific siRNA expression plasmid together with a plasmid encoding GFP were stimulated with Sema4D for 1 hour. (B) Hippocampal neurons transfected with expression plasmids encoding GFP or GFP-tagged R-RasQL were stimulated with Sema4D for 1 hour. Growth cones of the transfected cells are shown by the fluorescence of GFP. A growth cone was defined by the presence of lamellipodia and filopodia visualized by the staining of filamentous actin (F-actin) with Alexa 594-conjugated phalloidin. Growth cone–positive cells were scored as a percentage of the total number of transfected cells, and results are the means ± SEM of three independent experiments in which 30 cells were counted. Bar, 10 μm.

Rnd1 also binds to Plexin-A1 and is required for the Sema3A-Plexin-A1–mediated repulsion (6, 24). In addition, mutations in the GAP homologous domains in Plexin-A1 suppressed the Plexin-A1–mediated repulsion (24). To determine whether the down-regulation of R-Ras activity was also involved in Plexin-A–mediated growth cone collapse, we expressed GFP-tagged R-Ras-QL in rat hippocampal neurons. R-Ras-QL suppressed the Sema3A-induced growth cone collapse (fig. S7), indicating that the down-regulation of R-Ras activity is also required for the Sema3A/Plexin-A–mediated growth cone collapse.

Direct regulation of R-Ras activity by Plexin-B1 acting as a GAP serves as a mechanism for transmembrane receptor–mediated signal transduction. The down-regulation of R-Ras activity is also involved in the Sema3A-induced repulsion. Activation of R-Ras promotes cell adhesion and controls cell migration and neurite outgrowth by activating integrins (912). It has been recently reported that Sema3A and Sema4D inhibited integrin-mediated cell adhesion and migration (2527). Therefore, the down-regulation of R-Ras activity by plexins may suppress integrin activation and thereby reduce cell adhesiveness, leading to growth cone collapse and neurite retraction. Considering that the GAP-homologous domains are well conserved among different plexin subfamilies, direct regulation of R-Ras activity by plexin is likely to be a major signaling pathway for cellular functions mediated by semaphorins.

Supporting Online Materials

www.sciencemag.org/cgi/content/full/305/5685/862/DC1

Materials and Methods

Figs. S1 to S7

References

References and Notes

View Abstract

Navigate This Article