Main

The Rho family of Ras-like GTPases consists of signaling molecules involved in cytoskeleton remodeling, membrane recycling and gene transcription. Rho GTPases play essential roles in controlling cell adhesion, migration, cell proliferation and differentiation1 and particularly in skeletal myogenesis.2, 3, 4, 5, 6, 7, 8, 9, 10 The Rho subfamily contains 21 members clustered in four subgroups that comprise RhoA, RhoB and RhoC; Rnd1, Rnd2 and Rnd3/RhoE; RhoBTB1, RhoBTB2 and RhoBTB3; and Rac1, Rac2, Rac3, Cdc42, TC10, TCL, RhoG, Chp1 and Chp2/Wrch-1. The remaining proteins (RhoD, Rif and RhoH/TTF) delineate distinct stems in the phylogenetic tree. Rho GTPases act as molecular switches that convert extracellular signals into multiple cellular effects by cycling between an inactive GDP-bound and an active GTP-bound state. Their activation is controlled by guanine nucleotide-exchange factors (GEFs), whereas their inactivation is promoted by GTPase-activating proteins (GAPs). The members of the Rnd subfamily (Rnd1, Rnd2 and Rnd3 also called RhoE) and RhoH are an exceptions, as they are always bound to GTP, are not regulated by GEFs and GAPs and have very low, if any, intrinsic GTPase activity.11, 12, 13, 14 Although 21 members are known in mammals, only few of them have been studied extensively. In skeletal muscle cells, the role of RhoA, Rac1 and TC10 has been analyzed. RhoA and Rac1 act both in myogenesis induction and myoblast fusion. RhoA positively regulates MyoD expression and skeletal muscle cell differentiation, as it is required for serum response factor-mediated activation of several muscle-specific gene promoters.2 Conversely, Rac1 inhibits myogenesis induction by preventing myoblast withdrawal from the cell cycle.3, 15 Later on, during the skeletal muscle differentiation program and in contrast to its inhibitory role in myogenesis induction, Rac1 signaling is involved in myoblast fusion in Drosophila.4, 5, 6, 7, 16 Recently, we demonstrated that Rac1 activity is required for myotube formation,17 whereas RhoA activity must be downregulated during this process,8, 9, 18 as that of the Rho-associated kinase (ROCK)I, an effector of RhoA.9, 18 In analyzing the potential mechanisms that might be involved in downregulation of RhoA and ROCKI during myoblast fusion, we focused on RhoE. Indeed, RhoE is ubiquitously expressed and is believed to be a RhoA and ROCKI antagonist.19 RhoE was isolated in a two-hybrid screen that used p190RhoGAP as bait11 and shown to stimulate p190RhoGAP-mediated GAP activity toward RhoA.20 Moreover, RhoE binds and inhibits ROCKI activity.21 Here we show that RhoE is involved in mammalian myoblast fusion in the myogenic cell line C2C12. We demonstrate that RhoE expression increases until the onset of fusion, whereas at the same time the activity of RhoA decreases. Finally, we also show that the knockdown of RhoE by RNA interference impairs downregulation of RhoA and ROCK activities normally observed before myoblast fusion and thus inhibits myotube formation.

Results

Variation of RhoE expression during skeletal muscle cell differentiation

RhoE is ubiquitously expressed and its mRNA has also been detected in skeletal muscle contrary to Rnd1 and Rnd2, the other members of this Rho family subgroup that were not observed in this tissue.13 To analyze their mRNA expression in C2C12 myoblasts, we performed real-time quantitative PCR amplification using specific primer sets.22 RhoE mRNA was found in C2C12 myoblasts and its level was close to that of RhoA and Rac1, which are both involved in the regulation of myogenesis (Figure 1).

Figure 1
figure 1

RhoE expression in C2C12 myoblasts. RhoA, Rac1 and RhoE cDNA levels relative to GAPDH were determined by quantitative RT-PCR amplification. The graph is representative of three independent experiments

Full size image

Rnd proteins are always bound to GTP and variations in their expression might be used to control their activities. We next analyzed RhoE mRNA and protein levels in C2C12 myoblasts induced to differentiate by replacing the growth medium (GM) with the appropriate differentiation medium (DM). At different time points after the shift to DM, cell extracts were prepared to analyze both mRNA and protein levels. The fusion process was concomitantly followed by morphological analysis of myoblast differentiation and appearance of multinucleated myoblasts. We could thus determine that fusion started between 30 and 42 h after DM addition (Figure 2A). As shown in Figure 2B, RhoE mRNA gradually increased until the onset of fusion and then suddenly decreased. These variations in RhoE mRNA levels coincided with those of RhoE protein (Figure 2C). These results show that RhoE is upregulated at the onset of myoblast fusion.

Figure 2
figure 2

RhoE expression during C2C12 myoblasts differentiation. C2C12 myoblasts grown in GM or in DM for the indicated times were simultaneously processed for the analysis of the fusion process (A) and for mRNA (B) and protein extracts preparation (C). (A) Phase contrast images of myoblasts and myotubes during the differentiation process. Bar: 20 μm. (B) The graph shows the RhoE cDNA levels relative to the GAPDH level in the same sample at the indicated periods, determined by quantitative RT-PCR amplification. Data are the mean±S.E.M. of five independent experiments. (C) Protein extracts (50 μg/well) from C2C12 myoblasts, collected at the indicated times, were immunoblotted for the assessment of RhoE and α-Tubulin expression. The histogram represents the quantification of RhoE normalized for the amount of α-Tubulin. The histogram shows the quantification of the two bands but similar variations have been observed for each of these bands (data not snown). (D) (a) Serial sections from control (i) and injured TA muscles 4 days after notexin injection (ii) were stained with hematoxylin and eosin. Bar: 100 μm. (b) Expression of M-cadherin and RhoE determined by quantitative RT-PCR amplification after normalization to GAPDH 4 days after notexin or cardiotoxin injection. The graph is representative of three independent experiments

Full size image

To analyze whether RhoE participates in skeletal muscle differentiation in vivo, we examined its expression in a mouse model of muscle regeneration. Skeletal muscle injury was induced by injection of notexin or cardiotoxin in tibialis anterior (TA) muscles. Regeneration was monitored by histological analysis (Figure 2Da) and M-cadherin expression (Figure 2Db) 4 days after injection. RT-PCR amplification showed that the expression of RhoE and M-cadherin were concomitantly upregulated during regeneration.

RhoE is required for myotube formation

To gain further insight into the role of RhoE in myogenesis, we knocked down RhoE expression. We generated stable C2C12 cell lines in which RhoE was silenced by small hairpin RNAs (C2C12 RhoE shRNA). We monitored the efficiency of gene silencing by western blot analysis in five different clones (Figure 3A). In all clones, RhoE protein levels relative to that of Tubulin were decreased by 50–80% (Figure 3B). Parental and C2C12 myoblasts expressing Firefly Luciferase-specific shRNA (C2C12 Luci shRNA) were used as controls. Parental, C2C12 RhoE shRNA and Luci shRNA myoblasts were grown to 80% confluency and shifted to DM for 4 days. Whereas numerous myotubes were observed in the parental and control C2C12 Luci shRNA cells (Figure 3Ca–b), only few myotubes were visible in RhoE shRNA clone C2 (Figure 3Cc). Similar results were obtained with all five RhoE shRNA clones, demonstrating that these effects were not due to clonal variations (data not shown). Quantification of the fusion index clearly illustrates that the knockdown of RhoE impairs myotube formation (Figure 3D). We next examined, by western blot analysis, whether RhoE silencing affected the expression of Myogenin and Troponin T, two muscle-specific proteins (Figure 3E). Their expression was similar in parental, control C2C12 Luci shRNA and C2C12 RhoE shRNA myoblasts, suggesting that induction of these genes does not require RhoE.

Figure 3
figure 3

Inhibition of RhoE expression by RNA interference prevents myotube formation. (A) Luciferase and RhoE shRNAs were delivered in C2C12 myoblasts by retroviral infection. Cells were selected and clones of resistant cells were analyzed for the expression of RhoE and α-Tubulin by immunoblot analysis. (B) The histogram represents the quantification of RhoE normalized for the amount of α-Tubulin in the different clones. Data are the mean±S.E.M. of three independent experiments. The histogram shows the quantification of the two bands but similar variations have been observed for each of these bands (data not snown). (C) May–Grünwald Giesma stain of parental (a), control C2C12 Luci shRNA (b) and C2C12 RhoE shRNA myoblasts (clone B, c) after 4 days in DM. Bar, 50 μm. (D) The histogram represents the fusion index calculated at the indicated times after DM addition for parental, control C2C12 Luci shRNA and C2C12 RhoE shRNA myoblasts (clone B). Data are the mean±S.E.M. of three independent experiments: at least 7000 nuclei were counted per experiment. (E) Cell lysates (20 μg/well) of parental, control C2C12 Luci shRNA and C2C12 RhoE shRNA myoblasts cultured in GM or DM for the indicated periods were assessed by western blot analysis for the expression of Myogenin, Troponin T and α-Tubulin. The histogram represents the quantification of Myogenin (black) and Troponin T (gray) normalized for the amount of α-Tubulin from four independent experiments

Full size image

These data indicate that RhoE is necessary for myotube formation while being dispensable for myogenesis induction.

RhoE controls myoblast elongation and alignment before fusion

We next analyzed RhoE distribution during myogenesis by indirect immunohistochemistry using a specific, affinity-purified antibody against RhoE (Figure 4A). In proliferating C2C12 myoblasts, RhoE had a predominantly perinuclear localization (a) as in fibroblasts.23 After DM addition, RhoE staining increased in myoblasts that were elongating and aligning before fusion (b). Once fusion occurred, it decreased to a level similar to that of proliferating cells (d). Neither the pre-immune serum (data not shown) nor the affinity-purified antibody pre-incubated with the immune peptide (c) revealed such a staining.

Figure 4
figure 4

Inhibition of RhoE expression by RNA interference prevents myoblast elongation and alignment. (A) RhoE localization analyzed by indirect immunofluorescence in proliferating C2C12 myoblasts (a) or after 30 (b) or 96 h (d) in DM. Staining obtained with the affinity-purified RhoE antibody pre-incubated with the immune peptide (c). Bar: 10 μm. (B) Phase contrast images of parental (a–c) and C2C12 RhoE shRNA myoblasts (d–f) cultured in GM or DM for the indicated periods. Bar: 10 μm. (C) Scanning electron micrographs of parental and C2C12 RhoE shRNA myoblasts cultured in GM or DM for the indicated periods. Bar: 10 μm. (D) (a) Protein extracts (20 μg/well) from parental and C2C12 RhoE shRNA myoblasts collected at the indicated periods were immunoblotted for M-cadherin and α-Tubulin expression. The histogram represents the quantification of M-cadherin normalized for the amount of α-Tubulin in three independent experiments. (b) M-cadherin localization in parental (i, ii) and C2C12 RhoE shRNA (iii, iv) myoblasts analyzed by indirect immunofluorescence. Shown are deconvolved images of M-cadherin (green) and DNA staining (blue). Bar: 10 μm

Full size image

We then precisely analyzed the morphological modifications occurring during the myogenesis of parental and C2C12 RhoE shRNA myoblasts using phase contrast and scanning electron microscopy (SEM) (Figure 4B and C). During myogenesis, C2C12 myoblasts elongate and align before fusion (Figures 4Ba–c and 5Ca–d).24, 25 In contrast, RhoE knockdown prevented elongation and alignment of myoblasts that remained flat (Figures 4Bd–f and 5Ce–h). Higher magnification analysis by SEM of the dorsal membrane revealed the presence of extensive microvilli and lamellipodia in parental C2C12 at the onset of myoblast fusion (Figure 4Cd), which were absent in C2C12 RhoE shRNA myoblasts (Figure 4Ch).

Figure 5
figure 5

RhoE downregulates RhoA activity. (A) The level of GTP-bound RhoA was measured using GST fused to the Rho-binding domain of the RhoA effector Rhotekin (GST-RBD) in lysates obtained from C2C12 cells in GM or DM from the same kinetic as the one described in Figure 2. RhoA was detected by immunoblotting. Data are representative of three independent experiments. (B) RhoA GTPase activity normalized to the amount of total protein was calculated from panel A and plotted with the RhoE protein levels in cell extracts from the same kinetic (shown in Figure 2B). (C) The level of GTP-bound RhoA was measured in lysates obtained from parental and C2C12 RhoE shRNA myoblasts that were in GM or DM and collected at the indicated times. Data are representative of three independent experiments. (D) Parental and C2C12 RhoE shRNA myoblasts were fixed after 3 days in DM and stained for F-actin (a and c). Overlay of F-actin (red) and DNA staining (blue) are shown in panels b and d. Bar: 10 μm

Full size image

M-cadherin, a member of the Ca2+-dependent cell–cell adhesion molecule family, is involved in myoblast alignment and fusion.8, 26 We thus examined M-cadherin expression and localization in C2C12 RhoE shRNA myoblasts (Figure 4Da). M-cadherin was monitored at different times after DM addition. In C2C12 RhoE shRNA myoblasts cultured in GM, the amount of M-cadherin was similar to that of the parental C2C12 cells. M-cadherin expression decreased in C2C12 RhoE shRNA myoblasts grown in DM compared to parental C2C12 myoblasts (Figure 4Da). In contrast, N-cadherin expression was comparable in C2C12 RhoE shRNA and parental C2C12 myoblasts (data not shown). We then analyzed M-cadherin localization by immunocytochemistry (Figure 4Db). In control elongated and aligned C2C12 myoblasts, M-cadherin accumulated at myoblast-myotube and myotube-myotube contacts (i and ii), but not in C2C12 RhoE shRNA cells (iii and iv).

Altogether, these data show that RhoE accumulates in elongated, aligned myoblasts prior to fusion and that its silencing impairs myoblast elongation and alignment and decreases M-cadherin expression and accumulation at intercellular contacts.

RhoE controls the downregulation of RhoA activity at the onset of myoblast fusion

RhoE antagonizes RhoA activity,12, 20 which is highly regulated during myogenesis. Indeed, RhoA is activated and required for myogenesis induction, but its activity must be downregulated to allow myotube formation.2, 8, 9, 27 First, we measured RhoA activity in C2C12 myoblasts from the same kinetic as the one described in Figure 2. Figure 5 shows the crude data (A) and RhoA activity normalized to the amount of total protein (B). As previously reported, we observed an increase in RhoA-GTP level at the beginning of myogenesis. Then, RhoA-GTP content decreased and reached a minimum after 2 days in DM (i.e., at the onset of fusion) (Figure 5B). As the samples used for the experiments described in Figures 2 and 5 were similar, we reported the amount of RhoE protein onto the same graph (Figure 5B) and showed that the maximum RhoE expression correlated to the lowest RhoA activity. We next measured RhoA activity in C2C12 RhoE shRNA myoblasts. In control C2C12 cells, we observed a decrease in RhoA activity 30 h after DM addition, that is, at the onset of fusion. Conversely, in C2C12 RhoE shRNA cells, we could not detect a decrease in RhoA activity at 30 h after DM addition (Figure 5C) or at later time points. This suggests that RhoE-dependent pathways are not involved in the control of RhoA activation during myogenesis induction and that they are implicated in RhoA activity's downregulation at the onset of myoblast fusion. To further demonstrate that RhoA activity was maintained in C2C12 RhoE shRNA myoblasts, we used the organization of the F-actin cytoskeleton as a functional readout (Figure 5D). After 3 days in DM, parental myoblasts showed a decrease in the number of stress fibers (Figure 5Da and b), particularly in myotubes, as previously reported.25 In contrast, C2C12 RhoE shRNA myoblasts still presented with a higher amount of stress fibers, a hallmark of RhoA activation (Figure 5Dc and d).

To confirm that myoblast fusion inhibition after RhoE silencing was dependent on continuous RhoA activation, we treated C2C12 RhoE shRNA myoblasts with the cell-permeable exoenzyme C3, which inactivates RhoA.28 As RhoA activity is required for myogenesis induction,2, 3 C3 transferase was added 6 h after induction of differentiation. As shown in Figure 6A, RhoA inhibition partially rescued inhibition of myoblast fusion caused by RhoE silencing. Quantification of the fusion index 3 days after DM addition revealed a four-fold increase in myoblast fusion in C2C12 RhoE shRNA myoblasts treated with C3 transferase, as compared to the untreated ones (Figure 6B). These data show that RhoE is responsible for the downregulation of RhoA, a crucial step in myoblast fusion.

Figure 6
figure 6

RhoA inhibition partially rescues myoblast fusion. Control C2C12 Luci and RhoE shRNA myoblasts were induced to differentiate by replacing the GM with the appropriate DM. Six hours later, C3 transferase was added, if appropriate. DM with or without C3 transferase was replaced every 24 h. (A) Images of control C2C12 Luci shRNA (a) and C2C12 RhoE shRNA myoblasts (b) after 4 days in DM or C2C12 RhoE shRNA myoblasts after 4 days in DM containing C3 transferase (c). Bar, 50 μm. (B) The histogram represents the fusion index calculated 96 h after DM addition for control C2C12 Luci and RhoE shRNA C2C12 myoblasts treated, or not, with C3 transferase. Data are the mean±S.E.M. of three independent experiments, where at least 3000 nuclei/experiment were counted

Full size image

RhoE activates p190RhoGAP at the onset of myoblast fusion

p190RhoGAP is a RhoE effector involved in RhoE-induced inactivation of RhoA in fibroblasts.20 We thus tested whether RhoE can also interact with p190RhoGAP in C2C12 myoblasts. Immunoprecipitation using an anti-RhoE antibody was performed in extracts from C2C12 myoblasts in GM or after the indicated times in DM (Figure 7A). Endogenous p190RhoGAP co-immunoprecipitated with RhoE (Figure 8Aa), and this association increased at the onset of myoblast fusion (Figure 7Ab).

Figure 7
figure 7

RhoE binds to p190RhoGAP and regulates its activity at the onset of the fusion process. (A) (a) Cell lysates of C2C12 myoblasts cultured in GM or at the indicated times after DM addition were immunoprecipitated using an anti-RhoE antibody and immunoblotted for the presence of p190RhoGAP and RhoE. −, without anti-RhoE antibody; +, with anti-RhoE antibody. (b) The histogram shows the quantification of p190RhoGAP associated with RhoE. Data are the mean±S.E.M. of three independent experiments. (B) Control and RhoE shRNA C2C12 myoblasts cultured in GM or at different times after DM were lysed and subjected to pulldown analysis with GST-RhoAL63. The precipitated material and total lysates were analyzed for p190RhoGAP levels by immunoblotting. Three independent experiments were analyzed by densitometry. The histogram represents the GAP activity normalized to the total protein content

Full size image
Figure 8
figure 8

ROCK inhibition partially rescues myoblast fusion in RhoE deficient cells. (A) Cell lysates of C2C12 myoblasts cultured in GM or at different times after DM addition were immunoprecipitated using an anti-RhoE antibody and immunoblotted to assess the presence of ROCK1 and RhoE. −, without anti-RhoE antibody; +, with anti-RhoE antibody. (B) Parental and RhoE shRNA expressing C2C12 myoblasts cultured in GM were co-transfected with GFP-tagged RhoA and myc-tagged ROCK1 expressing plasmids. Cell lysates were immunoprecipitated using a mouse monoclonal anti-myc antibody and immunoblotted to assess the presence of ROCK1 (WB myc) and RhoA (WB RhoA). −, without anti-myc antibody; +, with anti-myc antibody. (C) Phospho-MYPT-1 and α-Tubulin levels were analyzed in total lysates by immunoblotting from control and C2C12 RhoE shRNA myoblasts cultured in GM or at different times after DM addition. The histogram shows the quantification of MYPT-1 phosphorylation normalized to the total protein content. Data are the mean±S.E.M. of three independent experiments. (D) Images of C2C12 RhoE shRNA myoblasts after 4 days in DM (a) or DM containing Y-27632 (b). Bar, 50 μm. (E) The histogram represents the fusion index calculated at the indicated times after DM addition for control C2C12 Luci and RhoE shRNA myoblasts treated or not with Y-27632. Data are the mean±S.E.M. of three independent experiments, where at least 3000 nuclei/experiments were counted

Full size image

We then analyzed the GAP activity of p190RhoGAP at different time points during myoblast differentiation in parental and C2C12 RhoE shRNA myoblasts (Figure 7B). Activated p190RhoGAP was pulled down using GST-RhoAL63, which is locked in its GTP-bound state.29 Pulleddown samples were then immunoblotted with an anti-p190RhoGAP antibody. The activity of p190RhoGAP gradually increased in C2C12 myoblasts after DM addition. In contrast, in C2C12 RhoE shRNA myoblasts, we did not observe any increase in the activity of p190RhoGAP. These data support the notion that p190RhoGAP interaction with RhoE allows the increase of its activity during myoblast differentiation.

RhoE effect on myoblast fusion is also dependent on ROCK

RhoE binds to the RhoA effector ROCKI and inhibits its activity.21 Moreover, inhibition of ROCK activity is required for myoblast fusion.9, 18 We thus analyzed whether ROCK was involved in the inhibition of myoblasts fusion induced by RhoE silencing. We first tested whether ROCKI was associated with RhoE in C2C12 myoblasts using immunoprecipitation (Figure 8A). ROCKI co-immunoprecipitated with RhoE in C2C12 myoblasts in GM or after the indicated times in DM. We also analyzed whether RhoE impacts on the binding of RhoA to ROCK1. We compared the association of RhoA to ROCK1 in parental and C2C12 RhoE shRNA expressing C2C12 myoblasts. We found that the level of RhoA bound to ROCK increases upon RhoE silencing (Figure 8B). We next analyzed ROCK activity by assessing the phosphorylation of a ROCK substrate, the myosin binding subunit of the myosin light chain phosphatase MYPT-130 (Figure 8C). As previously reported, we observed a decrease in ROCK activity during myogenesis in control cells, but not in C2C12 RhoE shRNA myoblasts induced to differentiate.

To gain insight into the role of ROCK in RhoE-induced myoblast fusion, we tested the effect of Y-27632, a pharmacological ROCK inhibitor. C2C12 RhoE shRNA myoblasts were grown in DM in the presence of Y-27632. Inhibition of ROCK partially rescued myoblast fusion in C2C12 RhoE shRNA cells (Figure 8D), as confirmed by quantification of the fusion index (Figure 8E). These data argue for a participation of RhoE in ROCK inhibition required for myoblast fusion.

Discussion

Unraveling the molecular mechanisms that govern myoblast fusion is important for the understanding of muscle diseases where fusion is affected or for the repair process that involves adult satellite–cell fusion. Among the signaling pathways involved in myoblast fusion, the decrease in RhoA/ROCK activities appears to be crucial.8, 9, 18 We have searched for upstream regulators of RhoA/ROCK signaling during myoblast fusion and demonstrate here that RhoE plays an important role in their regulation and in skeletal muscle cell differentiation.

RhoE is required for myoblast fusion

Our results prove that RhoE is required for myoblast fusion in C2C12 myoblasts, although it is dispensable for myogenesis induction. Indeed, RhoE knockdown impairs myotube formation without altering the level of muscle-specific gene products. We report modifications in RhoE expression during myogenesis consistent with a role during myoblast fusion, as its expression level in C2C12 myoblasts increases until the onset of myoblast fusion before resuming its basal level once fusion is well under way. Its implication in muscle differentiation is reinforced by the results we obtained in an in vivo model for muscle regeneration. Moreover, the data of RhoE localization during myogenesis combined with the analysis of the phenotype of C2C12 myoblast, in which RhoE was silenced, suggest that RhoE participates in myoblast elongation and alignment before fusion. The variations in RhoE expression during myogenesis are consistent with the way this molecule is regulated. Indeed, RhoE binds only weakly to GDP, has very low, if any, intrinsic GTPase activity and therefore exists predominantly in a GTP-bound state. RhoE is not regulated by GEFs or GAPs but rather by its expression level and post-translational modifications such as phosphorylation.21, 31 Further studies will be required to understand how RhoE expression is controlled during myogenesis.

RhoE decreases RhoA and ROCK activities

RhoA GTPase activity is tightly controlled throughout myogenesis. At the beginning of myogenesis, RhoA is activated through an N-cadherin-mediated pathway10, 27 and is then downregulated at the onset of myoblast fusion.8, 18 We report here that RhoE is required for the decrease of RhoA activity during myoblast fusion. This is the first example of a regulatory pathway that negatively controls RhoA activity to allow fusion to occur. In fibroblasts, RhoE interacts with p190RhoGAP in vitro and increases its GAP activity toward GTP-bound RhoA.20 We demonstrate here, on endogenous proteins, that the highest association between RhoE and p190RhoGAP occurs at the onset of myoblast fusion. Interestingly, C2C12 myoblasts in which RhoE was silenced show defects both in p190RhoGAP activation and RhoA inhibition at the onset of myoblast fusion. Finally, myoblast fusion could be partially rescued in C2C12 RhoE shRNA by inactivating RhoA with C3 exoenzyme. In conclusion, these data explain how RhoE, through the binding and activation of its effector p190RhoGAP, antagonizes RhoA activity and participates in the induction of myoblast fusion. RhoE also inhibits RhoA signaling by binding to the RhoA-activated serine/threonine kinase ROCK, whose activity must be decreased during myogenic differentiation to allow myoblast fusion.9 Addition of the ROCK inhibitor Y-27632 partially rescued myoblast fusion in C2C12 RhoE shRNA myoblasts as efficiently as C3 the exoenzyme. Moreover, Y-27632 did not strengthen the effect of C3 exoenzyme alone (data not shown), suggesting that RhoA acts through ROCK to repress myoblast fusion. Nevertheless, as the recovery of myoblast fusion with these drugs was only partial, other yet unknown RhoE effectors might participate in regulating myoblast fusion. For instance, the RhoE effector Socius is involved in RhoE-induced disassembly of actin stress fibers,32 and Rnd2 and RhoE interact with the RhoGAP MgcRacGAP.33 Nevertheless, both Socius and MgcRacGAP are predominantly expressed in the testis, which emphasized the need to identify other skeletal muscle-specific RhoE effectors.

Spatial RhoE targeting

The mechanisms responsible for RhoE activation and function in myoblast fusion are most probably linked to its increased expression, which allows its accumulation above a threshold level, thus favoring its binding to p190RhoGAP and ROCK1 that leads to the inhibition of the RhoA/ROCK pathway. Moreover, RhoE recruitment to specific plasma membrane sites might facilitate its association with p190RhoGAP and ROCK1. Indeed, in fibroblasts and epithelial cells, RhoE colocalizes with cadherins in adherens junctions.34 Although M-cadherin is involved in myoblast fusion,8 we could not detect any association of RhoE with either M-cadherin or N-cadherin in C2C12 myoblasts (data not shown). Also β1-integrin, another transmembrane protein involved in myoblast fusion, was not associated with RhoE (data not shown).35 In neurons, RhoE and Rnd1 interact with plexins, which are semaphorin receptors involved in axon guidance through regulation of p190RhoGAP-dependent RhoA downregulation.28, 36 Semaphorin 4C participates in myoblast fusion,37 and therefore it would be important to test whether RhoE associates with this kind of complex in myoblasts.

RhoE/RhoA/ROCKI signaling and myoblast fusion

Here we demonstrate for the first time that RhoE participates in skeletal muscle differentiation through the downregulation of RhoA/ROCK signaling, an event required for myoblast fusion.8, 9 ROCK phosphorylates both the myosin binding subunit of myosin light chain phosphatase and the myosin regulatory light chain, which induces the actin–myosin 2 interaction and promotes actin cytoskeleton assembly and cell contractility.30, 38 When RhoA and ROCK activities are decreased, Myosin 2 activity is also reduced. In C2C12 myoblasts in which RhoE was silenced, stress fibers are not reduced and myoblast fusion is impaired. Interestingly, the loss of stress fibers occurs in pre-fusing myoblasts.25 Nevertheless, we could only partially rescue myoblast fusion with RhoA or ROCK inhibitors, and synergy was not obtained by the addition of both (data not shown). Although we cannot rule out a possible toxicity of these drugs, this suggests the existence of other RhoE-dependent pathways controlling myoblast fusion. Indeed, we found that the inhibitor of myosin light chain kinase (MLCK) ML7 slightly increases the effect of C3 exoenzyme on the recovery of myoblast fusion in RhoE-deficient myoblasts (data not shown). Myosin 2 inhibition with blebbistatin in C2C12 myoblasts in which RhoE was silenced also resulted in rescue of myoblast fusion (data not shown), supporting the idea that a decrease in acto-myosin contraction is required for myoblast fusion. These data suggest a synergy between a RhoA/ROCK and a MLCK pathway downstream of RhoE, both converging on the reduction of myosin activity. Upstream regulation of MLCK can occur through Rac1 and its effector Pak, which phosphorylates and inactivates MLCK. Moreover, there is a functional antagonistic balance between the Rho and Rac subgroups of the Rho family, and we have previously reported such a crosstalk in skeletal muscle cells.8, 17 Indeed, Rac1 signaling is required for myoblast fusion, both in Drosophila and in mammalian myoblasts.4, 5, 6, 7, 17 Interestingly, our detailed examination of the membrane morphological modifications by SEM and pulldown assay (data not shown) indicate that Rac1 activity is not increased in C2C12 myoblasts in which RhoE was knocked down. Thus RhoE-dependent reduction of RhoA activity might also contribute to Rac1 activation. Our results establish that RhoE-dependent downregulation of RhoA/ROCK signaling also contributes to correct functioning of M-cadherin, a protein required for myoblast fusion, which is regulated by RhoA GTPase.8 Indeed, we show that M-cadherin is downregulated and mislocalized in myoblast in which RhoE was silenced and that its expression is partially rescued by the addition of RhoA or ROCK inhibitors (data not shown).

Myoblast fusion is a multi-step process in which actin cytoskeleton dynamics plays a key role. We thus propose that RhoE is an actor of myoblast fusion through the regulation of the organization of the actin cytoskeleton and M-cadherin function, which in turn controls myoblast alignment and elongation.

Materials and Methods

Cell culture

C2C12 mouse myoblasts were cultured as described.27 To induce differentiation, GM was replaced by DM consisting of DMEM/Ham's F-12 supplemented with 2% FCS. Stable cell lines derived from C2C12 myoblasts were cultured under the same conditions. Cell permeable C3 transferase (0.4 μg/ml, Cytoskeleton) or the ROCK inhibitor Y-27632 (10 μM, Calbiochem) were added 6 h after the addition of DM. Drug-containing DM was refreshed every 24 h.

Quantification of mRNA by real-time quantitative PCR amplification

Total RNAs from myoblasts in culture were prepared using the Rneasy minikit from QIAGEN. Total RNAs from TA muscles were extracted following the acid/guanidinium thiocyanate/phenol/chloroform method using RNable (Eurobio, France) according to the manufacturer's instructions. cDNAs were synthetized from 5 mg of total RNA using superscript II Reverse Transcriptase (Invitrogen) and random hexamer primers. The levels of the various cDNAs were determined by quantitative PCR amplification using the SYBR green I with a Light Cycler (Roche Diagnostics, Sommerville, NJ, USA). All primers used were previously described.22, 39 GAPDH mRNA was amplified as an internal control. The specificity of the primers was checked by DNA sequencing.

Gel electrophoresis and immunoblotting

Cell extracts were prepared as previously described.8 Protein concentration was determined with a BCA protein assay kit (Pierce Chemical, Rockford, IL, USA). Protein extracts (20 μg) were resolved on polyacrylamide gels (12%) and transferred onto Immobilon-P membranes. Membranes were incubated with monoclonal antibodies against RhoE (1:1000; Upstate), troponin T (1:1000; Sigma-Aldrich), myogenin (1:500; BD Biosciences PharMingen, San Diego, CA, USA), M-cadherin (1:200, NanoTools, Munich, Germany) and α-tubulin (1:100). Membranes were processed as described previously.8

Small hairpin RNA retroviral vector construction and establishment of stable cell lines

To suppress endogenous RhoE expression, annealed double strand oligonucleotides GATCCGTCAGAACGTGAAATGCAAAGTTCAAGAGACTTGCATTTCACGTTCTGACTTTTTTACGCGTG(top) and AATTCACGCGTAAAAAAGTCAGAACGTGAAATGCAAGTCTCTTGA ACTTTGCATTTCACGTTCTGACG (bottom) were inserted into the RNAi-Ready pSIREN-RetroQ-ZsGreen vector (Clontech, BD Biosciences). Bold letters correspond to oligonucleotides 291–310 of the mouse and human RhoE cDNA sequence (NM_028810). The pSIREN-RetroQ-ZsGreen vector, which targets the firefly Luciferase, was used as a control.

Phoenix cells expressing RhoE shRNA or Luciferase shRNA were grown to collect retrovirus-containing cell-free supernatants. Infection was performed on exponentially growing C2C12 myoblasts (5 × 105 cells per 60-mm dish) by using 5 ml of viral supernatant. Positive GFP cells were selected by FACS and cloned by limited dilution. RhoE expression was assayed on five random clones for RhoE shRNA and on a pool of Luciferase shRNA expressing C2C12.

Fusion index determination

Cells growing onto 100-mm dishes were fixed in 3.7% formaldehyde in phosphate-buffered saline (PBS) followed by a 3-min permeabilization in 0.1% Triton X-100 in PBS and saturated in PBS containing 0.1% bovine serum albumin. Nuclei were stained with Hoechst stain (0.1 μg/ml, Sigma-Aldrich) and washed with PBS. Fusion index, that is the number of nuclei in multinucleated myotubes divided by the total number of nuclei, was determined by the MRI Cell Image Analyzer software.8 At least 7000 nuclei in each experiment were counted. Only cells with three nuclei or more were considered to be myotubes.

May–Grünwald Giemsa staining

Cells were washed twice with PBS, incubated with May–Grünwald diluted in PBS for 5 min and next with Giemsa diluted in PBS for 10 min.

Muscle regeneration model

Mice (male C57BL/6J, 6 week old) were anesthetized by intraperitoneal injection of pentobarbital. Twenty microliters of 1 μM notexin (Latoxan, Valence, France) or 20 μl of 10 μM cardiotoxin (Latoxan, Valence, France) were injected into the TA muscles. Muscles were collected at various times after injection. Contralateral uninjected TA muscles were used as control. Muscles were used for hematoxylin/eosin staining and to isolate total RNA for RT-PCR.

Immunocytochemistry

Cells were fixed in 4% paraformaldehyde in PBS followed by a 2-min permeabilization in 0.1% Triton X-100 (in PBS). RhoE was stained using an affinity-purified antibody raised against an N-terminal peptide of RhoE (SQKLSSKSIMDPNQNVK) as previously described,21 and produced by Eurogentec (Seraing, Belgium). Control experiments were performed using the preimmune serum and affinity-purified anti-RhoE blocked with saturating levels of its specific peptide antigen. M-cadherin was stained using a polyclonal antibody as described.8 Anti-RhoE and anti-M-cadherin were revealed using Alexa Fluor 488-conjugated goat anti-rabbit antibody (Molecular Probes, Invitrogen). F-actin expression was analyzed using TRITC-conjugated phalloidin (Sigma-Aldrich) and nuclei were stained with Hoechst (0.1 μg/ml, Sigma-Aldrich). Images were captured with a MicroMax 1300 CCD camera (RS-Princeton Instruments, Treuton, NJ, USA) driven by the MetaMorph (v.7, Universal Imaging Corp, Westchester, PA, USA) Software. Stacks of images were captured with a piezo stepper (E662, Physik Instruments) with a Z step of 0.2 μm. Stacks were then restored with Huygens deconvolution software (Scientific Volume Imaging) and the restored images were viewed in 3D with MetaMorph. Images were processed using Adobe Photoshop and Adobe Illustrator.

Scanning electron microscopy

Parental and C2C12 RhoE shRNA myoblasts were grown on Thermanox™ coverslips (Nunc) either in GM or in DM. After washing in PBS, myoblasts were fixed in PBS containing 3% glutaraldehyde for 2 h and post-fixed with 1% OsO4 in distilled water for 1 h. After alcohol dehydration, myoblasts were dried in hexamethyldisilazane (Acros Organics, NJ, USA) for 2 min, coated with Gold-palladium and observed using a Hitachi S4000 scanning microscope at 10 kV. For each condition, at least 100 cells were examined.

RhoA GTPase activity assay

Proliferating or differentiating C2C12 myoblasts were lysed and processed to measure the total and GTP-bound RhoA levels as described previously8 or by using the luminometry based G-LISA™ Rho activation assay (Cytoskeleton Inc.).

Immunoprecipitation

C2C12 cells were lysed for 10 min in ice-cold extraction protein buffer.8 Protein extracts (1 mg) were immunoprecipitated using a mouse monoclonal anti-RhoE antibody (1/1000 dilution; Upstate). Negative controls were performed using IgG from murine serum (Sigma-Aldrich). Half of each immunoprecipitation was separated on an 8% polyacrylamide gel and then transferred onto nitrocellulose. p190RhoGAP and ROCKI detection were performed using monoclonal antibody (1/250; BD Transduction Laboratories). The other half of each immunoprecipitation was separated on a 15% polyacrylamide gel, transferred on Immobilon-P membranes and immunoblotted with a polyclonal anti-RhoE antibody.

Three 100-mm dishes were co-transfected with 3 μg GFP-tagged Rho and myc-tagged ROCK1 plasmids using JetPEI. Twenty-four hours after the transfection, cells were lysed and protein extracts (2 mg) were immunoprecipitated using a mouse monoclonal anti-myc antibody from ascites (1/1000 dilution). Negative controls were performed using IgG from murine serum (Sigma-Aldrich). Eighty percent of each immunoprecipitation was separated on an 8% polyacrylamide gel and then transferred onto nitrocellulose. ROCKI detection was performed using monoclonal anti-myc antibody from ascites (1/10 000 dilution).

Twenty percent of each immunoprecipitation was separated on a 15% polyacrylamide gel, transferred on Immobilon-P membranes and immunoblotted with a monoclonal anti-RhoA antibody (1/300 dilution, Santa Cruz, Tebu).

p190 RhoGAP activity assay

RhoA L63GST (a generous gift from K Burridge) was prepared as described.29 Cells were lysed at different times after DM addition and processed as described.28

ROCK activity assay

Cells extracts were lysed as described.40 Protein extracts (60 μg) were resolved on polyacrylamide gels (10%) and transferred onto Immobilon-P membranes. Membranes were probed with phospho-Mypt1 polyclonal antibody (1/1000) (Upstate) and α-tubulin (1:100).