Cooperative regulation by Rac and Rho of agrin-induced acetylcholine receptor clustering in muscle cells.

A key aspect of neuromuscular synapse formation is the clustering of muscle acetylcholine receptors (AChR) at synaptic sites in response to neurally secreted agrin. Agrin-induced AChR clustering in cultured myotubes proceeds via the initial formation of small microclusters, which then aggregate to form AChR clusters. Here we show that the coupling of agrin signaling to AChR clustering is dependent on the coordinated activities of Rac and Rho GTPases. The addition of agrin induces the sequential activation of Rac and Rho in C2 muscle cells. The activation of Rac is rapid and transient and constitutes a prerequisite for the subsequent activation of Rho. This temporal pattern of agrin-induced Rac and Rho activation reflects their respective roles in AChR cluster formation. Whereas agrin-induced activation of Rac is necessary for the initial phase of AChR cluster formation, which involves the aggregation of diffuse AChR into microclusters, Rho activation is crucial for the subsequent condensation of these microclusters into full-size AChR clusters. Co-expression of constitutively active forms of Rac and Rho is sufficient to induce the formation of mature AChR clusters in the absence of agrin. These results establish that Rac and Rho play distinct but complementary roles in the mechanism of agrin-induced AChR clustering.

During embryonic development, innervation induces the anatomical and biochemical specialization of a defined region of the muscle cell membrane immediately under the motor nerve ending. A prominent aspect of this specialization is the accumulation of high densities of nicotinic acetylcholine receptors (AChR) 1 at these sites (1,2). The aggregation of AChR and other synaptic components is mediated by agrin, a heparansulfated proteoglycan that is synthesized by motor neurons and secreted into the synaptic cleft (3,4). The recruitment of AChR into clusters in postsynaptic membranes ensures high efficiency synaptic transmission at neuromuscular junctions.
Agrin-induced redistribution of surface AChR involves the co-clustering of multiple associated proteins, several of which have been identified to date (2,5). These include the muscle-specific receptor tyrosine kinase (MuSK) (6), the linker protein rapsyn (7), and the scaffolding proteins dystroglycan and utrophin (8). The clustering of MuSK upon its activation by agrin, the formation of AChR complexes with rapsyn, as well as the aggregation of these complexes and their stabilization upon the formation of dystroglycan-utrophin scaffolds appear to be sequential events that are to some extent independently regulated (9 -11). Although the signaling mechanisms that couple agrin activation of MuSK to the clustering of postsynaptic components are incompletely characterized, there is recent evidence for the participation of Src tyrosine kinases (12,13), the Rho GTPases Rac and Cdc42 (14), and Dishevelled, a component of the Wnt signaling pathway (15).
Focal changes in the peripheral actin-based cytoskeleton are thought to underlie the aggregation of AChR at neuromuscular junctions (16 -18). The monomeric G proteins Rac and Rho function to link extracellular signals to dynamic changes in actin cytoskeleton organization leading to the assembly of lamellipodia and actin-myosin filaments, respectively (19 -22). Rac activation induces actin polymerization at the plasma membrane, causing the appearance of lamellipodia with resultant stimulation of cell spreading and motility (23,24). Rho exerts the opposite effect by stimulating actin stress fiber appearance and focal adhesion complex formation to promote cell adhesion and contractility (25,26). As several recent studies have shown, Rac and Rho are mutually inhibitory in several cell types, and the balance between their antagonistic activities is responsible for the dynamic changes in cell morphology, adhesion, and motility (27,28). In other systems Rac serves as an upstream activator of Rho (29).
We have recently shown that agrin triggers the activation of Rac and Cdc42 and that this activation is necessary but not sufficient for formation of full size AChR clusters (14). In this study we present evidence that Rho plays a crucial role in agrin-initiated signaling that is complementary to the contribution of Rac/Cdc42 and that together these Rho family GTPases serve to couple signaling initiated by extracellular agrin to the formation of AChR clusters. diameter glass coverslips in 35-mm culture dishes for microinjection and microscopy experiments, or in 100-mm culture dishes for affinity precipitation assays. The cells were cultured in growth medium consisting of Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum, 5% calf serum, 100 g/ml penicillin, and 100 g/ml streptomycin (Invitrogen) at 37°C with air/5% CO 2 . To stimulate muscle differentiation, the growth medium was replaced with differentiation medium consisting of Dulbecco's modified Eagle's medium containing 2% horse serum (Invitrogen) and 100 g/ml each of penicillin/ streptomycin 2 days after plating. Under these conditions the major proportion of the C2 myoblasts fused to form multinucleated myotubes during the subsequent 2 days. To induce clustering of AChR, the cultures were treated with 5 nM agrin where specified.
Transfection-For experiments utilizing transient transfections, 1 day post-plating C2 myoblast cultures were transfected with the indicated plasmids at a final concentration of 5 g of DNA/ml using either calcium phosphate precipitation or LipofectAMINE reagent (Invitrogen). The transfection medium was replaced with differentiation medium for 3 days prior to addition of 5 nM soluble recombinant neural agrin prepared from COS cells transfected with cDNA encoding the C-terminal fragment of agrin (30).
Rho Pathway Inhibitors-Differentiated C2 cells were treated with either soluble C3 exotransferase from Clostridium botulinum (31)(32)(33) or Y27632 (34,35) at varying time points and were monitored for clustering ability in response to agrin. C3-transferase was applied to C2 myotubes at a concentration of 50 g/ml and preincubated for 2 h at 37°C prior to agrin treatment. The cultures were then incubated for an additional 8 h in the presence of both C3-toxin and agrin. For Y-27632, myotubes were treated with Y-27632 at 20 M concentration, and the cells were incubated at 37°C for 2 h prior to the addition of agrin. 20 M of Y-27632 was then added every 2 h for 8 h to ensure a continuous presence of Y-27632 at a pharmacologically effective concentration. The effect of Y-27632 on agrin-induced AChR clustering was measured by labeling AChR clusters with the AChR ligand tetramethylrhodamineconjugated ␣-bungarotoxin (Molecular Probes).
Microinjection-For experiments involving microinjection, myotube cultures were grown on coverslips and after 3 days in differentiation medium were transferred to Dulbecco's modified Eagle's medium. The injections were performed with sterile borosilicate microinjection capillaries. The capillary was loaded with a solution containing the indicated plasmid or protein in microinjection buffer (50 mM HEPES, pH 7.2, 100 mM KCl, 5 mM NaH 2 PO 4 , pH 7.3, for DNA injections, or 20 mM Tris acetate, pH 7.4, 20 mM NaCl, 1 mM MgCl 2 , 0.1 mM EDTA, 5 mM 2-mercaptoethanol for protein injections) and mounted on the Eppendorf micromanipulator 5171 connected to an Eppendorf microinjector 5242. DNA was at a final concentration of 0.1 g/l. For protein injections the buffer included RhoV14 at 3 g/l and/or RacV12 at 6 g/l. Cytoplasmic injections of plasmid DNA or protein were done at a pressure of 40 -50 hPa for 0.3 s under ambient conditions. Subsequently the cultures were shifted back to differentiation medium and incubated at 37°C for a further 0.5-1 day.
Microscopy-Cells plated on glass coverslips and subsequently either transfected or microinjected were labeled for 1 h with 10 nM tetramethylrhodamine-conjugated ␣-bungarotoxin in Dulbecco's modified Eagle's medium with 1 mg/ml bovine serum albumin for 1 h at 37°C, rinsed with PBS, and fixed in 3.7% formaldehyde with PBS for 30 min to visualize the surface distribution of AChR. After fixation, the cells were permeabilized in 0.2% Triton X-100 with PBS at room temperature for 5 min, blocked with 10 mg/ml bovine serum albumin with PBS for 3 min, and subsequently incubated with anti-T7 antibody (Novagen) for 1 h at 37°C, rinsed with PBS, and stained by incubating with an FITC-conjugated goat anti-mouse antibody (ICN) for 1 h at 37°C. The cultures that were microinjected with protein were co-injected with FITC goat anti-mouse antibody to identify the injected myotubes. The coverslips were mounted on slides using Immu-Mount (Shandon). The images were acquired using a Nikon PCM 2000 laser scanning immunofluorescence confocal microscope and associated imaging system and processed using Adobe Photoshop.
To measure Rho activation, an affinity precipitation method was used (36,37) in which cell lysates prepared with lysis buffer A were incubated with GST fused to the Rho-binding domain from the effector protein Rhotekin (GST-TRBD) bound to glutathione-coupled Sepharose beads for 45 min at 4°C. The beads were washed four times with wash buffer (50 mM Tris, pH 7.2, containing 1% Triton X-100, 150 mM NaCl, 10 mM MgCl 2 , 10 g/ml each of leupeptin and aprotinin, and 0.1 mM phenylmethylsulfonyl fluoride). Bound Rho proteins were eluted with sample buffer (38) and detected by Western blotting using a monoclonal antibody against RhoA (Santa Cruz Biotechnology). The blots were developed using goat anti-mouse antibody coupled to horseradish peroxidase (1:1,000 dilution) and visualized with the ECL detection system (Amersham Biosciences).
In a similar manner Rac activation was measured by affinity precipitation of cellular GTP-bound forms of Rac (39). In this case cell lysates were prepared with lysis buffer B and incubated with GST fused to the Cdc42/Rac (p21)-binding domain of PAK (GST-PBD) bound to glutathione-coupled Sepharose beads for 30 min at 4°C. The fusion protein beads with bound proteins were then washed three times in an excess of lysis buffer, eluted in sample buffer, and then analyzed by Western blotting with a mouse monoclonal antibody against human Rac1 (Transduction Labs) at a 1:1,000 dilution. The blots were developed using sheep anti-mouse coupled to horseradish peroxidase (1:1,000 dilution) and visualized with the ECL detection system (Amersham Biosciences).
For experiments in which Rho and Rac activation was measured in the same cultures, the cells were extracted with lysis buffer B, and the lysates were divided for Rho and Rac determinations as described above. To measure the effect of Rac inactivation on Rho activation, the cells were co-transfected with T7 epitope-tagged wild type Rho (RhoWT) or constitutively activated Rho (RhoV14) and T7 epitope-tagged dominant negative Rac (RacN17). Rho activation was assessed using the TRBD pull-down assay described above, using antibody to the T7 epitope to select for transfected cells.

Inhibitors of the Rho Pathway Impair Agrin-induced AChR
Clustering-We have recently found that constitutively active Rac and Cdc42 mutants cause the formation of AChR microclusters (2-5 m in diameter) but are insufficient to induce full-size AChR clusters (15-20 m in diameter) (14), suggesting that other regulatory components might be required. In view of these findings and the observations implicating the actin cytoskeleton in AChR clustering (3), we have now investigated the potential contribution of Rho. For this purpose we tested the effect of a dominant negative mutant of Rho, RhoN19, on agrin-induced AChR clustering. T7 epitope-tagged RhoN19 plasmid was microinjected into differentiated myotubes, and after overnight incubation in differentiation medium in the absence or presence of agrin, surface AChR on myotubes were visualized by labeling with the AChR ligand tetramethylrhodamine-conjugated ␣-bungarotoxin. The microinjected cells expressing the RhoN19 construct were identified using indirect immunofluorescence with anti-T7 antibody and FITC-tagged second antibody. The dominant negative Rho mutant was found to impair the ability of agrin to cluster AChR (Fig. 1A, panel d) compared with myotubes that were co-injected with vector plus FITC-conjugated antibody to identify the injected cells and treated with agrin under identical conditions (Fig. 1A, panel b). Moreover, as shown in Fig. 1B, the decrease in numbers of agrin-induced full-size AChR clusters is associated with an increase in the number of AChR microclusters present in the myotubes expressing RhoN19.
To verify the contribution of Rho to AChR clustering by pharmacological means, we utilized C3 exotransferase from C. botulinum, an inhibitor of Rho activity (31)(32)(33). When added to differentiated muscle cell cultures prior to agrin treatment, C3 transferase was seen to impair the ability of agrin to induce the formation of AChR clusters (Fig. 2, A, panel b, and B). As in the case of RhoN19 expression, the decrease of AChR clusters was accompanied by an increase in the number of microclusters ( Fig. 2, A, arrowheads, and B). Together these findings suggest that Rho activation is necessary for the induction by agrin of full-size AChR clusters but not for microcluster formation. Moreover, the increased accumulation of microclusters observed upon impairment of Rho is consistent with the possibility that full-size clusters are formed via the aggregation of microclusters and that this step is blocked by Rho inhibitors.
To investigate the effector pathway by which Rho mediates AChR clustering, we utilized a pharmacological inhibitor of the Rho-associated protein kinase p160 ROCK , Y-27632 (34,35). As shown in Fig. 2 (A, panel c, and B), Y-27632 inhibits the agrin-induced aggregation of AChR into clusters while increasing the number of microclusters (arrowheads) per myotube. These results closely resemble those observed with direct inhibition of Rho, consistent with the possibility that the downstream effects that couple Rho activation to the formation of AChR clusters are mediated through p160 ROCK . The ability to inhibit AChR clustering by blocking the Rho pathway utilizing three separate methods demonstrates that this pathway is necessary for agrin to induce AChR clusters.
Cooperative Induction of AChR Clustering by Constitutively Active Mutants of Rac and Rho-To determine the extent to which Rho activation can mimic the stimulation of AChR clustering by agrin, differentiated myotubes were microinjected with RhoV14 protein, and the effects of this constitutively active Rho mutant on AChR surface distribution were measured after overnight incubation without agrin. As can be seen in Fig. 3A, myotubes that express RhoV14 show altered distribution of AChR (panels a and b) into regions of somewhat elevated AChR density. These areas, although clearly larger than the microclusters induced by the constitutively active mutant of Rac (RacV12) (Fig. 3A, panels c and d), are less dense and more poorly defined than the clusters induced by agrin (Fig. 1A, panel b).
Next, to determine whether these mutant Rho GTPases produce cooperative effects on AChR clustering upon co-expression in muscle cells, we microinjected RacV12 and RhoV14 proteins into the same myotubes and examined the effects on AChR surface distribution after an overnight incubation. We observed that under these conditions, co-expression of RacV12 and RhoV14 significantly increases the amount of full-size clusters above the levels obtained with either alone (Fig. 3A, panels e Similarly, Y-27632, which was added to differentiated muscle cell cultures 2 h prior to agrin treatment and then every 2 h for 8 h, also inhibits AChR cluster formation (panel c) compared with control cells treated with agrin only. The arrows denote full-size clusters; the arrowheads denote microclusters. B, quantitative comparison of the number of aggregations/myotube shows that agrin-induced full-size cluster formation (black bars) is inhibited by more than 40% in C3 transferasetreated cells and by ϳ50% in Y-27632-treated cells compared with control cells. However, the number of agrin-induced microclusters/myotube (gray bars) increases with both C3 and Y-27632 treatment (n ϭ 40; the error bars represent the S.E.). and f). Quantitative comparison of AChR aggregation into microclusters and clusters induced by RacV12 and RhoV14 in these experiments is shown in Fig. 3B. Together these findings indicate that with regard to the contributions of Rac and Rho to agrin-induced AChR clustering, activation of each is necessary, and co-activation of both is sufficient for the induction of fullsize AChR clusters.
Agrin Induces Activation of Rho in Myotubes-To determine whether endogenous Rho in muscle cultures is activated upon agrin treatment, we utilized an affinity precipitation assay based on the selective binding of activated (GTP-bound) Rho to the effector Rhotekin (36,37). As shown in Fig. 4A, agrin activates Rho in C2 myotubes, as monitored by the increased binding of activated Rho to GST-TRBD, a GST fusion protein containing the Rho-binding domain of Rhotekin. In contrast, agrin fails to activate Rho in undifferentiated C2 myoblasts, consistent with the finding that MuSK is selectively expressed in differentiated muscle cells (40).
In view of our findings suggesting that Rac mediates microcluster formation while Rho regulates the subsequent aggregation of microclusters into clusters, it was of interest to compare the time course of activation of Rho with that of Rac upon agrin treatment of C2 myotubes. For this comparison, myotubes were exposed to agrin for various intervals and then extracted, and the lysates were divided into two pools for parallel determina-tions of Rho and Rac activation. As can be seen in Fig. 4B, two differences in the time courses of Rho and Rac are apparent. First, Rac activation occurs sooner following agrin treatment FIG. 3. Constitutively active mutants of Rho (RhoV14) and Rac (RacV12) together induce AChR clustering in the absence of agrin treatment. C2 myotubes were microinjected with RhoV14 protein, RacV12 protein, or both proteins in addition to FITC-conjugated antibody. A, cells injected with RhoV14 exhibit regions of elevated AChR density but few AChR clusters (panels a and b). RacV12-injected cells display only microclusters (panels c and d), whereas myotubes injected with both RhoV14 and RacV12 display full-size clusters (panels e and f). B, quantitative comparison of the microclusters and clusters induced by RacV12 and RhoV14 shows that RacV12 induces a greater than 10-fold increase in microcluster number but has little effect on the number of clusters. Rho V14 plus RacV12 induces a more than 5-fold increase in full-size AChR clusters with no increase in microclusters. In contrast, RhoV14 alone does not induce either microclusters or clusters.

FIG. 4. Agrin-induced activation of Rho in differentiated muscle cells is dependent on Rac activation.
Agrin-induced Rho activation in nontransfected C2 myoblasts and myotubes was measured by the increase of endogenous Rho bound to a Rhotekin domain (TRBD) that specifically binds the GTP-bound form of Rho. Agrin treatment causes increased association of Rho with GST-TRBD in myotubes but not in myoblasts (A). The time course for Rho activation in response to agrin shows the onset of activation by 15 min after treatment and a sustained activation of Rho for at least 90 min (B) compared with Rac, which is activated by 5 min but inactive by 60 min in the same cells. C, Rho activation by agrin is eliminated in myotubes expressing the dominant negative Rac mutant, RacN17. D, RacN17 does not impair the expression of a constitutively active Rho mutant (RhoV14), demonstrating that the inhibitory effect of RacN17 on agrin-induced Rho activation (C) does not reflect cellular toxicity. than the activation of Rho; a significant increase in Rac activity is evident within 5 min, whereas Rho activation is not appreciable until 15 min after agrin addition. An additional difference is that the activation of Rac is clearly transient, as shown by the return of Rac activity to basal levels by 60 min after agrin exposure. In contrast, Rho activation reaches maximum by 15 min and has not diminished by 90 min after the onset of agrin treatment.
The finding that Rac activation occurs earlier than Rho activation in response to agrin led us to test the possibility that Rho activation might be dependent on the prior activation of Rac. We measured the agrin-induced activation of Rho in muscle cells expressing dominant negative Rac (RacN17) to determine whether this mutant would block agrin-induced Rho activation. As shown in Fig. 4C, RacN17 prevents the activation of Rho by agrin. To control for the possibility that the absence of agrin-induced Rho activation reflected muscle cell toxicity associated with RacN17 expression, we measured the effects of RacN17 expression on the ability of muscle cells to express constitutively activated Rho (RhoV14), using the TRBD pull down assay. As shown in Fig. 4D, constitutive Rho activation was unaffected by RacN17 expression. These findings support the notion that the pathway to Rho activation by agrin passes through Rac. DISCUSSION A crucial feature of synaptogenesis is the establishment at postsynaptic membranes of a well delineated surface region that displays sharply elevated sensitivity to neurotransmitter. In the case of the neuromuscular synapses, this is accomplished by the clustering of AChR initiated by spatial cues originating at the motor nerve ending and acting on the adjacent patch of muscle membrane. The postsynaptic sensor component of this focal signaling mechanism includes the receptor tyrosine kinase, MuSK, which activates a signaling pathway to couple the binding of neurally derived agrin to the assembly of subsynaptic complexes and to the aggregation of diffusely distributed surface AChR into high density clusters. Our recent findings that activation of Rac/Cdc42 is essential for AChR clustering and is sufficient for mediating agrin-induced AChR microclusters but not full-size clusters (14) suggests that additional regulatory steps are necessary to bring about the aggregation of microclusters to form full-size clusters. This has led us to examine the contribution of Rho, a monomeric G protein that plays a critical role in the formation of adhesion plaques, well characterized surface structures to which AChR cluster regions bear a resemblance, to agrin-induced AChR clustering.
In this study we provide evidence that Rho, acting in conjunction with Rac, plays a critical role in coupling agrin signaling to AChR cluster formation. We show that in differentiated myotubes Rho undergoes activation upon agrin treatment. Inhibition of the Rho pathway impairs agrin-induced AChR clustering. Conversely, stimulation of the Rho pathway with the constitutively active Rho mutant, RhoV14, stimulates the formation of regions of elevated AChR density on the myotube surface, but these areas are more poorly defined with lower staining intensity than the full-size clusters induced by agrin. However, co-expression of RacV12 with RhoV14 results in the induction of full-size high density AChR clusters. These findings lead to the conclusion that Rac and Rho are each necessary but individually insufficient for agrin-induced AChR cluster formation, whereas co-activation of Rac and Rho can serve to mediate AChR clustering, mimicking the effect of agrin.
Earlier studies have suggested that microclusters and clusters represent distinct stages of neural factor-induced AChR aggregation, with microclusters serving as the precursors of full-size clusters in cultured muscle cells (41)(42)(43)(44), as well as in developing neuromuscular synapses (2). The dispersal of clusters in response to pharmacological agents (protein kinase C activators and phosphoprotein phosphatase inhibitors) similarly proceeds via microclusters as intermediates (43,45). Under the experimental conditions used in the present experiments, exposure of C2 myotubes to agrin results in the appearance of numerous microclusters (2-5 m diameter) within 2 h and the subsequent accumulation of full-size AChR clusters (15-20 m diameter) starting after a further 1-2 h. Several of our findings suggest that Rac and Rho act at distinct points but serve cooperative roles in the pathway that couples agrin signaling to AChR cluster formation. Whereas dominant negative Rac (RacN17) abolishes all agrin-induced AChR aggregation, inhibition of Rho-mediated signaling in agrintreated myotubes is seen to selectively impair AChR aggregation into full-size AChR clusters but does not inhibit the formation of AChR microclusters. On the contrary the number of microclusters increases when Rho or its effector pathway is blocked. Constitutively active Rac induces AChR aggregation exclusively into microclusters but requires the co-activation of Rho to produce clusters. These results support the notion that agrin-induced AChR microclusters are the precursors of fullsize clusters. Moreover, Rho activation appears to be unnecessary for agrin-induced microcluster formation but crucial for the subsequent conversion of microclusters into full-size clusters.
What is the relationship between Rac and Rho in the signaling pathway that couples agrin stimulation to surface AChR aggregation? Our current finding that upon agrin addition to muscle cells Rac activation precedes and overlaps with Rho activation is consistent with the observed sequence of events in which Rac-mediated formation of microclusters is followed by the Rho directed condensation of microclusters into full-size clusters. Additionally, we find that agrin fails to cause Rho activation in muscle cells expressing dominant negative Rac, suggesting that in these cells Rac activation is necessary for the agrin-induced Rho stimulation that directs cluster formation. A hypothetical scheme that illustrates how the Rho GTPases cooperate to induce high AChR site density membrane domains in response to agrin signaling is presented in Fig. 5.
The multiprotein cytoskeletal complexes that assemble at postsynaptic membrane sites in the neuromuscular junction and in the AChR cluster regions in cultured muscle cells closely resemble the focal adhesions involved in cell-substrate adhesion and signaling (46 -48). Rho GTPases serve crucial roles in the formation of focal complexes and adhesions (25). Rac and Cdc42 regulate the formation of small focal complexes, whereas Rho is required for the formation of focal adhesions by the aggregation of focal complexes (26,28,49). Moreover, similar conditions produce dispersal of both AChR clusters and focal adhesions. The dispersal of AChR clusters into microclusters induced by the protein kinase C activator 12-O-tetradecanoylphorbol-13-acetate in cultured myotubes (43) may be related to actomyosin inactivation in these cultures (50) and resembles the 12-O-tetradecanoylphorbol-13-acetate-triggered, Rho inactivation-dependent disassembly of focal adhesions seen in Madin-Darby canine kidney cells (51,52). Our present results showing that the block in AChR cluster formation by inhibition of Rho activity is associated with an increase in number of AChR microclusters parallel the recent observation that the Rho inactivation-induced block of focal adhesions is accompanied by the formation of new focal complexes (28).
A comparison of the time courses for the activation of Rac and Rho in response to agrin provides a potential insight into the regulatory mechanisms that coordinate their activities. Rac is activated shortly after agrin treatment, whereas Rho activity remains at basal levels for a longer duration. The selective early activation of Rac would allow an interval for Rac-induced formation of AChR microclusters. Upon the subsequent activation of Rho, these microclusters would then undergo aggregation into clusters through Rho-mediated cytoskeletal dynamics. It is noteworthy that the prolonged activation of Rho by agrin is accompanied by the inactivation of Rac and is concomitant with the accumulation of full-size AChR clusters on the surface of myotubes. In conclusion, these findings show that Rac and Rho play sequential and cooperative roles in the pathway coupling neural agrin signaling to AChR clustering in myotubes. In more general terms our results support the possibility that cellular events responsible for the formation of synapses including axon guidance and postsynaptic membrane assembly utilize similar signaling pathways in which Rho family GTPases play homologous roles.