|Home | About | Journals | Submit | Contact Us | Français|
Cadherins are homophilic cell–cell adhesion molecules implicated in many fundamental processes, such as morphogenesis, cell growth, and differentiation. They accumulate at cell–cell contact sites and assemble into large macromolecular complexes named adherens junctions (AJs). Cadherin targeting and function are regulated by various cellular processes, many players of which remain to be uncovered. Here we identify the small GTPase Rab35 as a new regulator of cadherin trafficking and stabilization at cell–cell contacts in C2C12 myoblasts and HeLa cells. We find that Rab35 accumulates at cell–cell contacts in a cadherin-dependent manner. Knockdown of Rab35 or expression of a dominant-negative form of Rab35 impaired N- and M-cadherin recruitment to cell–cell contacts, their stabilization at the plasma membrane, and association with p120 catenin and led to their accumulation in transferrin-, clathrin-, and AP-2–positive intracellular vesicles. We also find that Rab35 function is required for PIP5KIγ accumulation at cell–cell contacts and phosphatidyl inositol 4,5-bisphosphate production, which is involved in cadherin stabilization at contact sites. Finally, we show that Rab35 regulates myoblast fusion, a major cellular process under the control of cadherin-dependent signaling. Taken together, these results reveal that Rab35 regulates cadherin-dependent AJ formation and myoblast fusion.
Cadherins are highly conserved transmembrane receptors that mediate calcium-dependent cell–cell adhesion and form adherens junctions. They play essential roles during embryonic development by regulating cell differentiation, growth, and migration and in the maintenance of tissue architecture in adult life (Takeichi, 1995 ; Halbleib and Nelson, 2006 ; Harris and Tepass, 2011 ). Perturbation of cadherin function is associated with cancer cell invasion and metastasis (Christofori, 2003 ). Cadherins mediate homotypic cell–cell adhesion through their extracellular domain (Troyanovsky, 2005 ), whereas their cytoplasmic domains interact with a range of proteins that link cadherins to the cytoskeleton and to cell signaling pathways (Kemler, 1993 ; Perez-Moreno et al., 2003 ).
Formation of cell–cell contacts is a multistep process that includes cadherin association with catenins, delivery of cadherin–catenin complexes to the plasma membrane (PM), lateral diffusion in the PM toward cell–cell contact sites, cadherin oligomerization, homotypic interactions, and association of cadherin complexes with their intracellular partners and the actin cytoskeleton. Moreover, cadherin-based cell–cell contacts are dynamic adhesive structures, and the trafficking and turnover of cadherins to and from the PM play an important role in this dynamic behavior (Yap et al., 2007 ; Schill and Anderson, 2009 ; Baum and Georgiou, 2011 ). Much attention has been focused on the role of the Rho and Arf families of small GTPases in cadherin-dependent adhesion (Braga et al., 2000 ; Fukata and Kaibuchi, 2001 ; Palacios et al., 2001 ; Lozano et al., 2003 ), whereas the implication of the Rab family of small GTPases that includes >60 proteins is much less known. Rab GTPases define specific trafficking routes within the secretory and endocytic pathways by controlling several transport steps, such as vesicle formation, motility, docking, and fusion (Zerial and McBride, 2001 ; Stenmark, 2009 ). In particular, Rab11 has been involved in E-cadherin recycling and apical membrane formation in mammals and Drosophila (Desclozeaux et al., 2008 ; Roeth et al., 2009 ), Rab5 and Rab7 in lysosomal targeting of E-cadherin in Src-induced epithelial-to-mesenchymal transition (Palacios et al., 2005 ), and Rab5 and Rab11 in N-cadherin trafficking during neuronal migration (Kawauchi et al., 2010 ). To determine whether some Rab family members might play a role in cadherin-dependent adhesion, we analyzed their localization and found that Rab35 strongly accumulated at cell–cell contacts. Rab35 is ubiquitously expressed and localizes at the PM and in endocytic compartments and controls a fast endocytic recycling pathway (Kouranti et al., 2006 ; Patino-Lopez et al., 2008 ). Rab35 has been also involved in cytokinesis, phagocytosis, and neurite outgrowth (Kouranti et al., 2006 ; Chevallier et al., 2009 ; Dambournet et al., 2011 ; Egami et al., 2011 ; Kobayashi and Fukuda, 2012 ). Moreover, several types of cargo, such as T-cell and major histocompatibility class I (MHCI) receptors, KCa2.3 Ca2+-activated K+ channels, and the oocyte receptor in Caenorhabditis elegans, require Rab35 for their recycling (Patino-Lopez et al., 2008 ; Allaire et al., 2010 ; Gao et al., 2010 ). In addition to its role in the regulation of membrane trafficking, Rab35 modulates actin organization directly through its effectors, by controlling Rac1 and Cdc42 localization at the PM, or through Arf6 (Zhang et al., 2009 ; Shim et al., 2010 ; Dambournet et al., 2011 ; Egami et al., 2011 ; Kobayashi and Fukuda, 2012 ).
Here we show that Rab35 is recruited to cell–cell contacts in a cadherin-dependent manner. Rab35 knockdown dramatically affects N-, M-, and E-cadherin recruitment to cell–cell contacts and the PM and leads to accumulation of cadherins in intracellular vesicles in both myoblasts and HeLa cells. Absence of Rab35 activity decreases the accumulation of phosphatidyl inositol 4,5-bisphosphate (PI(4,5)P2) and PIP5KIγ at cell–cell contacts, a change that also participates in the loss of cadherins at these sites. We thus identify Rab35 as a new regulator of adherens junction (AJ) formation.
To investigate the possible involvement of Rab family members in cadherin-dependent adhesion, we expressed wild-type Rab4, Rab5, Rab7, Rab8, Rab11, and Rab35 fused to green fluorescent protein (GFP) in C2C12 mouse myoblasts and HeLa cells and then monitored their localization and that of N- and M-cadherin. In both cell lines, only Rab35 accumulated at cell–cell contact sites, where it colocalized with N- and M-cadherin (Figure 1, A and B, for myoblasts; Supplemental Figure S1, A and B, for HeLa cells).
Moreover, cadherins triggered Rab35 recruitment to cell–cell contact sites. Indeed, in mouse L cells, which do not express endogenous cadherins, Rab35 did not accumulate at cell contacts. Conversely, upon expression of exogenous N-, M-, or E-cadherin, Rab35 was recruited to cell contacts, where it colocalized with the expressed cadherin (Figure 1C). This is specific for Rab35, because none of the other tested Rab family members (Rab4, Rab5, Rab7, and Rab11) was recruited to cell–cell contacts in a cadherin-dependent manner (Supplemental Figure S1C).
Finally, in immunoprecipitation experiments using anti–N- or -M-cadherin antibodies and whole-cell extracts of C2C12 myoblasts and HeLa cells that express wild-type Rab35 (Rab35WT) fused to GFP, Rab35 was immunoprecipitated together with endogenous N-cadherin (Figure 1D, a and c) or M-cadherin (Figure 1Db), as revealed by Western blot analysis. Similarly, endogenous Rab35 was immunoprecipitated together with M-cadherin in C2C12 myoblasts (Supplemental Figure S1D). Rab11 was not found in the M-cadherin complex, whereas the Rab35 effector fascin was.
These findings show that Rab35 is complexed with N- and M-cadherin and colocalizes with these cadherins at cell contact sites. They also indicate that Rab35 recruitment at cell–cell contacts is induced by cadherins.
To get insights into Rab35 function at cell–cell contact sites, we generated by retroviral infection stable C2C12 cell lines in which Rab35 expression was inactivated by RNA interference. Two Rab35 short hairpin RNAs (shRNAs; Rab35 shRNA1 and Rab35 shRNA2) were used to generate a pool (Figure 2A) and various cell clones that were selected using a resistance marker (Supplemental Figure S2, A and B). As controls, parental and C2C12 myoblasts expressing a luciferase shRNA (Luci shRNA) were used (Fortier et al., 2008 ). We first investigated whether Rab35 inactivation influenced N- and M-cadherin localization. In parental and Luci shRNA C2C12 myoblasts, N- and M-cadherin accumulated at cell–cell contacts (Figure 2B, a and b and g and h, respectively; arrows). In contrast, their accumulation at cell–cell contacts was perturbed in Rab35 shRNA C2C12 myoblasts (Figure 2B, c–f and i–l, arrowheads), whereas their intracellular localization increased (Figure 3). We confirmed this result by expressing a dominant-negative mutant of Rab35 (Rab35S22N) in C2C12 myoblasts (Supplemental Figure S2C). Again, N- and M-cadherin localization at cell–cell contacts was disrupted in 60% of C2C12 and HeLa cells expressing Rab35S22N (Supplemental Figures S2C and S3 and Supplemental Videos S1–S4). Moreover, cadherins accumulated in intracellular vacuoles (asterisks in Supplemental Figure S2C and Supplemental Figures S4 and S5) that were induced by Rab35S22N expression, as previously shown for other cargoes (Kouranti et al., 2006 ; Patino-Lopez et al., 2008 ; Chesneau et al., 2012 ). Thus cadherin localization at cell–cell contacts was perturbed in Rab35-knockdown cells and upon expression of Rab35S22N.
Of importance, we did not detect any significant difference in the total expression levels of N- and M-cadherin in cells expressing Rab35 shRNA1 and 2 in comparison to parental and control Luci shRNA C2C12 myoblasts (Figure 2C). To further quantify N- and M-cadherin levels at the cell surface of Luci shRNA and Rab35 shRNA C2C12 myoblasts, we used a cell surface biotinylation assay (Figure 2D). Quantification of the results showed that Rab35 knockdown decreased the amount of both N- and M-cadherin at the PM. Because cadherin stability at cell–cell contacts is dependent on the binding of p120 catenin (Davis et al., 2003 ; Xiao et al., 2003 ), we then analyzed whether Rab35 silencing decreased p120 catenin association with N- and M-cadherin. Although the level of p120 catenin was not modified in Rab35 shRNA C2C12 myoblasts (Figure 2Ea), p120 catenin association with N-cadherin (Figure 2Eb) and M-cadherin (Figure 2Ec) was significantly decreased. In contrast, the association of β-catenin with M-cadherin was not modified in Rab35 shRNA C2C12 myoblasts.
Moreover, the reduced accumulation of cadherins at the PM of Rab35 shRNA C2C12 myoblasts was accompanied by increased N- and M-cadherin accumulation in transferrin-labeled recycling compartments (Figure 3A). Similar results were obtained upon Rab35S22N expression (Supplemental Figure S5). Accumulation of N- and M-cadherin in transferrin-labeled recycling compartments was even higher when cell–cell contact disruption induced by Rab35S22N was enhanced by ethylene glycol tetraacetic acid (EGTA) treatment (Supplemental Figure S5). As previously described (Rahajeng et al., 2012 ), functional inhibition of Rab35 led to accumulation of transferrin in the perinuclear region (Figure 3A, d and j). These N- or M-cadherin–containing endocytic vesicles were positive for AP-2, clathrin, and to a lesser extent Rab5 (Figure 3B) but not for other endocytic markers, such as EEA1, Rab4, and Rab11 (unpublished data). Accumulation of N- or M-cadherin in AP-2–, clathrin-, and Rab5-positive vesicles was higher in C2C12 myoblasts that express Rab35S22N than in cells expressing Rab35WT (Figure 3B).
Taken together, these data show that Rab35 silencing does not affect the total expression level of N- and M-cadherin but perturbs their localization at the PM and their stabilization at cell–cell contacts.
We previously demonstrated PI(4,5)P2 accumulation and colocalization with N- and M-cadherin at cell contacts in C2C12 myoblasts (Taulet et al., 2009 ; Bach et al., 2010 ). Given that Rab35 regulates the localization of PI(4,5)P2 at the PM (Kouranti et al., 2006 ), we then assessed whether Rab35 knockdown affected its accumulation and colocalization with N-cadherin in C2C12 myoblasts. To this aim, we transfected parental and Rab35 shRNA C2C12 myoblasts with a construct that expresses the PH domain of PLCδ fused to GFP (PH-PLCδ-GFP) that specifically binds to PI(4,5)P2 (Varnai and Balla, 1998 ). PI(4,5)P2 accumulated at the PM and at cell–cell contacts in control cells (Figure 4Aa, arrow) but not in Rab35 shRNA C2C12 myoblasts (Figure 4Ac, arrowhead). To confirm that Rab35 regulates PI(4,5)P2 accumulation at cell contacts, we analyzed the localization of PH-PLCδ-GFP during cell–cell contact reformation upon Ca2+ recovery after chelation with EGTA. Rab35 knockdown affected PI(4,5)P2 accumulation at cell–cell contacts also after Ca2+ recovery (Figure 4A, compare b and d). Specifically, whereas 70% of cell–cell contacts showed PI(4,5)P2 enrichment in control cells, this dropped to 30% in Rab35 shRNA C2C12 myoblasts. Expression of the Rab35 dominant-negative mutant Rab35S22N in C2C12 myoblasts also reduced PI(4,5)P2 accumulation at cell–cell contacts (Figure 4B). In contrast, no perturbation of PI(4,5)P2 localization at cell–cell contacts was observed after expression of Rab35WT or Rab35Q67L, a constitutively active form of Rab35. Moreover, both Rab35 isoforms colocalized with PI(4,5)P2 at cell–cell contacts (Figure 4B). We also analyzed PI(4,5)P2 levels at cell–cell contacts using an antibody that specifically detects this phosphoinositide. To perform a more accurate quantification of PI(4,5)P2 at cell–cell contacts, we expressed red fluorescent protein (RFP) fused to a CAAX sequence that addresses RFP to the PM to normalize the signal given by the anti-PI(4,5)P2 antibody. Line-scan analysis confirmed the decrease of PI(4,5)P2 expression at cell–cell contacts in Rab35-depleted cells (representative line scan shown in Figure 4C).
These data show that Rab35 is required for PI(4,5)P2 accumulation at cell–cell contacts.
PI(4,5)P2, N- and M-cadherin (Taulet et al., 2009 ; Bach et al., 2010 ), and Rab35 (this study) accumulate at cell contacts in C2C12 myoblasts. Moreover, recruitment and activation of PIP5Kγ at sites of N-cadherin ligation, resulting in PI(4,5)P2 production, have been reported (El Sayegh et al., 2007 ). We thus asked whether PI(4,5)P2 was required for N- and M-cadherin accumulation at cell–cell contacts. To decrease PI(4,5)P2 level in cultured myoblasts, cells were treated with agents that reduce (calcimycin, LiCl) or mask (neomycin) PI(4,5)P2 (Griffin and Hawthorne, 1978 ; Hallcher and Sherman, 1980 ; Gabev et al., 1989 ; Laux et al., 2000 ). Whereas in control cells, N- and M-cadherin accumulated at cell–cell contacts (Figure 5A, a and e, arrows), treatments that perturbed PI(4,5)P2 disrupted cadherin accumulation at cell–cell contacts (Figure 5A, b–d and f–h, arrowheads). We then analyzed the effect of PI(4,5)P2 reduction on N- and M-cadherin localization by using stable C2C12 cell lines in which PIP5KIγ expression was silenced by RNA interference (El Sayegh et al., 2007 ; Ling et al., 2007 ; Bach et al., 2010 ). Moreover, PI(4,5)P2 expression at cell–cell contacts was significantly lower in PIP5KIγ shRNA myoblasts than in control cells (Figure 5B). Both N- and M-cadherin accumulation at cell–cell contacts was perturbed in PIP5KIγ shRNA myoblasts (Figure 5C, d–f) in comparison to control Luci shRNA C2C12 cells (Figure 5C, a–c). Similarly, Rab35 was not recruited to cell–cell contacts in PIP5KIγ shRNA myoblasts compared with control Luci shRNA C2C12 myoblasts (Supplemental Figure S6). This is in agreement with our data showing that cadherins recruit Rab35 to cell–cell contact sites (Figure 1C). Moreover, whereas the level of p120 catenin was not modified in PIP5KIγ shRNA C2C12 myoblasts (Figure 5Da), p120 catenin association with N-cadherin (Figure 5Db) and M-cadherin (Figure 5Dc) was decreased. Conversely, β-catenin binding was not affected. The weak association of N- and M-cadherin with p120 catenin when PI(4,5)P2 is reduced reflects an unstable cadherin complex at cell junctions.
Finally, we analyzed PIP5KIγ localization in Rab35-knockdown cells. PIP5KIγ accumulated at cell–cell contacts in control cells, where it colocalized with N- and M-cadherin (Figure 6A, a–c and g–i, arrows). In contrast, in Rab35 shRNA C2C12 myoblasts, PIP5KIγ was no longer accumulated at cell–cell contacts, whereas its punctuated intracellular localization increased (Figure 6A, d–f and j–l). To analyze in more detail PIP5KIγ distribution, we cotransfected C2C12 myoblasts with hemagglutinin (HA)-tagged PIP5KIγ and either RFP-Rab35WT or RFP-Rab35S22N. PIP5KIγ was mostly found at the PM and accumulated at cell–cell contacts when coexpressed with Rab35WT (Figure 6B, a–c). Conversely, it colocalized with Rab35S22N in clusters located outside cell–cell contacts (Figure 6B, d–f). Moreover, N-cadherin was also found in these clusters with PIP5KIγ and Rab35S22N (Figure 6C). Taken together, these data show that 1) Rab35 is required for PIP5KIγ accumulation at cell–cell contacts and 2) PI(4,5)P2 generated at cell–cell contacts is required for cadherin-dependent intercellular junction formation.
To analyze the functional consequence of Rab35 knockdown, and because N- and M-cadherins are two major regulators of myogenesis induction and myoblast fusion (Charrasse et al., 2002 , 2006 ), we asked whether Rab35 knockdown impaired myogenesis. First, C2C12 cells were grown to 80% confluency in growth medium (GM) and then shifted to differentiation medium (DM) for 4 d. Rab35 expression level increased during the differentiation process (Figure 7A). Then parental, Luci shRNA, and Rab35 shRNA C2C12 myoblasts were induced to differentiate, and the expression of myogenin, troponin T, and myosin heavy chain (three myogenic markers) and myotube formation were analyzed. We observed a little decrease in the expression of these myogenic markers in Rab35 shRNA myoblasts (Figure 7B). These results were confirmed using four different Rab35 shRNA clones (Rab35 shRNA1 A1 and A7; Rab35 shRNA2 A1 and A10, unpublished data). Analysis of myogenin and troponin T expression by immunofluorescence after 2 d in DM also confirmed these results (unpublished data). In addition, time-lapse imaging of parental and Rab35 shRNA myoblasts revealed that Rab35 silencing perturbed myoblast fusion (Supplemental Video S5; left, parental C2C12 cells; right, Rab35 shRNA2 myoblasts). A significant reduction in the number of myotubes and in the fusion index was observed in Rab35 shRNA myoblasts after 4 d in DM (Figure 7C, d–g) in comparison to controls (Figure 7C, a–c). Comparable results were obtained with the different Rab35 shRNA clones and pooled cells. Moreover, these findings were confirmed using C2C12 myoblasts that express the dominant-negative mutant Rab35S22N-GFP after selection by fluorescence-activated cell sorting (FACS) or with a resistance marker (G418; (Figure 7C, h and i). These data show that Rab35 is required for myogenesis and more specifically for myotube formation and fusion.
AJs are very dynamic structures, and their continuous remodeling is vital for the maintenance of tissue integrity, morphogenetic movements, delamination, and epithelial-to-mesenchymal transition (Harris and Tepass, 2011 ). Here we report that the Rab35 GTPase is a new regulator of cadherin-dependent AJ formation through regulation of cadherin trafficking and stabilization at cell–cell contacts.
Endocytosed cadherins can undergo multiple fates, including degradation or recycling back to the PM. In Rab35-depleted cells, cadherin localization at the cell surface was reduced (Figure 3D), whereas their total level was not decreased, suggesting that Rab35 regulates the recycling of internalized cadherins. Previous studies linked Rab35 to the recycling pathways from endosomes to the PM of cargoes that are internalized in a clathrin-dependent or clathrin-independent way, such as transferrin or MHCI (Kouranti et al., 2006 ; Patino-Lopez et al., 2008 ; Allaire et al., 2010 ). This suggests that clathrin-dependent and clathrin-independent cargoes could use common machineries for recycling back to the cell surface from recycling endosomes. This is consistent with the fact that cadherins undergo endocytic processing through several internalization routes from the PM (Delva and Kowalczyk, 2009 ; Schill and Anderson, 2009 ; Harris and Tepass, 2011 ). Rab35 does not generally control the recycling of all transmembrane proteins but allows selective recycling of some cargoes from early endosomes. Indeed, it was previously demonstrated that Rab35 controls recycling of MHCI (Allaire et al., 2010 ; Chesneau et al., 2012 ) but not of β1-integrin from early endosomes (Allaire et al., 2010 ). Until now, only few cargoes that require Rab35 for their recycling had been identified, and here we demonstrate that cadherins are among them.
To characterize the compartments in which cadherins are located after Rab35 depletion, various GFP-tagged Rab family members were expressed, such as Rab4, 5, 7, 8, and 11. With the exception of Rab5, none of the tested Rab members localized in the compartment in which cadherins accumulated. We observed an enlargement of Rab5-positive early endosomes, as described (Allaire et al., 2010 ). A clear colocalization upon Rab35 depletion between cadherins and transferrin was detected (Figure 3D). This is consistent with studies linking Rab35 to the recycling of transferrin back to the PM (Kouranti et al., 2006 ; Patino-Lopez et al., 2008 ; Chesneau et al., 2012 ). Finally, we demonstrate that cadherins and clathrin heavy chain or AP-2 colocalize in cells that express dominant-negative Rab35. This is not surprising because Rab35 and connecdenn, a GEF for Rab35, are both expressed in clathrin-coated vesicles (Allaire et al., 2010 ).
Besides its role in cadherin localization at the PM, Rab35 could also affect cadherin stabilization at cell–cell contacts through a decrease of PI(4,5)P2 production at cell–cell contact sites. We confirm here, as previously shown in epithelial cells (Ling et al., 2007 ), that the production of PI(4,5)P2 at cell–cell contact sites is important for cadherin stability. Recruitment and activation of PIP5KIγ at sites of N- and E-cadherin ligation result in PI(4,5)P2 production (Akiyama et al., 2005 ; El Sayegh et al., 2007 ; Ling et al., 2007 ). PIP5KIγ interacts directly with the cytoplasmic tail of cadherins and is required for E-cadherin–mediated AJ assembly. Moreover, expression of a kinase-inactive form of PIP5KIγ showed that PI(4,5)P2 generation is required for AJ assembly (Ling et al., 2007 ) and local actin assembly at AJs (El Sayegh et al., 2007 ). When Rab35 is inhibited, PIP5KIγ is not accumulated at cell–cell contacts and is sequestered in membrane vesicles with cadherins and Rab35. The colocalization of PIP5KIγ with cadherins in these structures is not surprising, because cadherin and PIP5KIγ are associated (El Sayegh et al., 2007 ) and are relocated to similar intracellular areas when cell–cell contacts are disrupted upon reduction of calcium ion concentration (Ling et al., 2007 ). As a consequence, PI(4,5)P2 production at cell–cell contacts is impaired. These results argue for a role of Rab35 in the control of cadherins and PIP5KIγ localization at cell–cell contacts. In stable C2C12 cell lines in which PIP5KIγ expression was silenced by RNA interference we observed 1) a decrease in the amount of PI(4,5)P2 at cell–cell contacts (Figure 5B) and 2) perturbation of N- and M-cadherin accumulation at cell–cell contacts (Figure 5C). The perturbation of N- and M-cadherin accumulation at cell–cell contacts in PIP5KIγ shRNA myoblasts is always less pronounced than in Rab35 shRNA myoblasts. This indicates that inhibition of PIP5Kγ affects only PI(4,5)P2 production at cell–cell contacts, which then perturbs cadherin stabilization. Conversely, inhibition of Rab35 affects both cadherin and PIP5K delivery to the plasma membrane and thus has a more pronounced effect on cadherin accumulation at cell–cell contacts. Indeed, in addition to the inhibition of PIP5K and PI(4,5)P2 production at cell–cell contacts, which impairs cadherin stabilization, Rab35 inhibition also blocks cadherin delivery into vesicles/clusters outside cell–cell contacts.
Functionally, we found that Rab35 is required for myoblast fusion to occur and observed that Rab35 expression level increased during myoblast fusion. Until now the only Rab involved in myoblast fusion was Rab11 in Drosophila (Bhuin and Roy, 2009 ). The defect in myoblast fusion we observed in C2C12 myoblasts in which Rab35 was silenced could be explained by the decrease of cadherin accumulation and PI(4,5)P2 production at cell–cell contacts. Indeed, myoblast fusion is characterized by PI(4,5)P2 accumulation at cell–cell contact sites (Bach et al., 2010 ). PI(4,5)P2 decrease at contact sites might thus impair the recruitment at fusion sites of PH domain–containing proteins involved in the regulation of myoblast fusion, such as Trio, Rac1, and Arf6, through the reorganization of the actin cytoskeleton and of PM dynamics (Macia et al., 2008 ). Moreover, Rab35 could participate in the control of myoblast fusion through the regulation of EHD1, one of the four carboxyl-terminal Eps15 homology domain-containing (EHD) proteins. Indeed, EHD proteins have a role in endocytic regulation and have been linked to several Rab family members through their association with mutual effectors (Naslavsky and Caplan, 2011 ). Of interest, Rab35 controls early endosomal recruitment of EHD1 (Allaire et al., 2010 ), a protein that was recently shown to be involved in myoblast fusion (Posey et al., 2011 ).
In conclusion, we report here that Rab35 controls cadherin-mediated AJ formation and myoblast fusion. Given that Rab35 regulates cadherin function, it will now be interesting to investigate whether dysfunction of Rab35 might be associated with cancer development or with skeletal muscle pathologies such as muscular dystrophies.
HeLa cells were cultured in DMEM supplemented with 10% fetal bovine serum. Culture and differentiation of C2C12 mouse myoblasts were done as described (Charrasse et al., 2006 ). PIP5KIγ shRNA C2C12myoblasts were described previously (Bach et al., 2010 ).
EGTA was used at 2 mM for 1 h, and then cells were rinsed in GM for 1 h. Calcimycin (50 μM; Sigma-Aldrich, St. Louis, MO), LiCl (1 mM; Sigma-Aldrich), and neomycin (1 mM; Sigma-Aldrich) were added to GM for 16 h.
shRNA constructs were made using the retroviral vector pSIREN-RetroQ according to the manufacturer’s protocol (BD Biosciences, San Diego, CA). To suppress endogenous Rab35 expression, the annealed double-strand oligonucleotides GATCCGgacaacttggcgaaacagcTTCAAGAGAgctgtttcgccaagttgtcCTTTTTTACGCGTG (top) and AATTCACGCGTAAAAAAGgacaacttggcgaaacagcTCTCTTGAAgctgtttcgccaagttgtcCG (bottom) were inserted into the RNAi-Ready pSIREN-RetroQ vector (Clontech, Mountain View, CA) to produce Rab35 shRNA1. Bold letters correspond to oligonucleotides 680–688 of the mouse Rab35 cDNA sequence (NM-198163). For Rab35 shRNA2, the sequence in bold letters of the top strand of the shRNA1 was replaced by GATCCGccttctcaggcagctacat-TTCAAGAGAatgtagctgcctgagaaggCTTTTTTACGCGTG (oligonucleotides 255–273 of the Rab35 sequence). Luciferase shRNA was used as control (Fortier et al., 2008 ). Retrovirus production in Phoenix cells, infection, and selection were performed as described (Fortier et al., 2008 ). Cells were grown continuously in 1 μg/ml puromycin. Different clones were isolated by limited dilution. Rab35 silencing was assessed in 11 random clones and a pool of Rab35 shRNA1 pSIREN-RetroQ C2C12 myoblasts and in 10 random clones and a pool of Rab35 shRNA2 pSIREN-RetroQ C2C12 myoblasts. All presented experiments were performed with at least the pool and two random clones in triplicate.
Cells cultured in 100-mm dishes were rinsed in cold phosphate-buffered saline (PBS) and lysed in 10 mM piperazine-N,N′-bis(2-ethanesulfonic acid), pH 7.0, 100 mM NaCl, 300 mM sucrose, 3 mM MgCl2, 0.5% NP40, 1 mM EDTA, 1 mM orthovanadate, and protease inhibitor cocktail (Sigma-Aldrich). Then 30- to 50-μg protein extracts were resolved on polyacrylamide gels (8, 12, and 15%) and transferred onto Immobilon-FL membranes. Membranes were incubated with monoclonal antibodies against N-cadherin (1:1000), β-catenin (1:500), p120 catenin (1:1000), and Rab11 (1:1000; BD Transduction Laboratories, Lexington, KY), troponin T (1:500), α-tubulin (1:500), and myosin (1:2000; Sigma-Aldrich), fascin (1:500; Abcam, Cambridge, MA), and M-cadherin (1:200; NanoTools, Munich, Germany) or with polyclonal antibodies against Rab35 (1:800; gift from A. Echard), myogenin (Santa Cruz Biotechnology, Santa Cruz, CA), GFP (1:800; Rockland Immunochemicals, Gilbertsville, PA), and actin (1:1000; Sigma-Aldrich). After washing, membranes were incubated with IRDyeR 680 or 800 secondary antibodies (LI-COR Biosciences, Lincoln, NE). Detection and analysis were performed using the Odyssey Infrared imaging system (LI-COR Biosciences). Quantification analysis of Western blots was performed by densitometry using Odyssey, version 3.0, and ImageJ (National Institutes of Health, Bethesda, MD).
Cell lysates were obtained as described. Monoclonal anti–N- or M- cadherin antibodies or irrelevant monoclonal antibody (1 μg) were incubated with protein G (Dynabeads; Invitrogen, Carlsbad, CA) at room temperature for 30 min. After washing, 1 mg of protein extract was added at room temperature for 1 h.
The presence of N- and M-cadherin at the cell surface was analyzed as described previously (Charrasse et al., 2006 ). Quantification of at least three independent experiments was performed using the Odyssey system.
Cells were transfected with plasmids encoding HA-, GFP-, or RFP-tagged Rab35WT, Rab35Q67L, or Rab35S22N (from Genaro Patino-Lopez, National Institutes of Health/National Cancer Institute) or N-cadherin, M-cadherin, PH-PLCγ, PIP5KIγ, CAAX, Rab4WT, Rab5WT, Rab7WT, Rab8WT, or Rab11WT using JetPEI (Polyplus-transfection SA, Illkirch, France) and then fixed in 3.2% paraformaldehyde in PBS for 15 min. After 5-min permeabilization with 0.1% Triton X-100 in PBS, cells were incubated with mouse monoclonal anti-HA (1:2000), monoclonal anti–N-cadherin (1:200), anti–E-cadherin (1:1000), anti–clathrin heavy chain (1:100) and anti–adaptin-α (AP-2, 1:100; all from BD Transduction Laboratories) or affinity-purified anti–M-cadherin and N-cadherin (1:50) antibodies. Antibodies were revealed with Alexa Fluor 546–conjugated, Alexa Fluor 488–conjugated, or Alexa Fluor 633–conjugated goat anti–mouse or rabbit antibodies (Molecular Probes, Eugene, OR).
For transferrin uptake, cells were incubated in DMEM at 37°C for 30 min. Tetramethyl-rhodamine–labeled transferrin (20 μg/ml in medium containing 1% BSA; Molecular Probes) was internalized at 37°C for 45 min to reach the turnover steady state, and then cells were rinsed twice before fixation and cadherin expression analysis by immunofluorescence. For PI(4,5)P2 detection, an anti-PI(4,5)P2 monoclonal antibody (2C11; Santa Cruz Biotechnology) was used according to the protocol described in Hammond et al. (2009 ).
Images were taken either with a Zeiss LSM Meta 510 confocal microscope or with a MetaMorph-driven (Molecular Devices, Sunnyvale, CA) wide-field Axioimager Z2 fluorescence microscope (Zeiss, Jena, Germany) with a PL APO 63× objective (numerical aperture 1.32; Leica, Melville, NY) and a CoolSNAP HQ camera (Photometrics, Woburn, MA). Images were processed using Photoshop and Illustrator (Adobe, San Jose, CA).
Sixteen-bit images were captured and epifluorescence images were first restored with Huygens (Scientific Volume Imaging, Hilversum, Netherlands; Ponti et al., 2007 ). Huygens is an iterative program that can reassign light, after encoding as gray levels, to its sources in the stack with a very high probability using a point spread function. This process removes the fuzziness contained in the stack while keeping the three-dimensional information. In the present study the maximum likelihood estimation algorithm was used throughout. Restored stacks were then further processed with Imaris (Bitplane, Zurich, Switzerland). Colocalization was analyzed with the Imaris Colocalization module.
Parental and Rab35 shRNA C2C12 myoblasts were grown to confluence before analysis by time-lapse microscopy. Alternatively, parental and Rab35 shRNA C2C12 myoblasts were transfected with N-cadherin–GFP/RFP, M-cadherin–GFP, Rab35WT-RFP, or Rab35S22N-RFP. Time-lapse confocal microscopy was performed using a spinning Nipkow disk. Images were saved as .tif files and further compiled into QuickTime movies (Apple, Cupertino, CA) using the MetaMorph program.
We are grateful to Genaro Patino-Lopez (National Institutes of Health) for providing various plasmids encoding the RFP- and GFP-tagged Rab35 constructs. We thank the Montpellier RIO Imaging Facility. This work was supported by the Ligue Nationale contre le Cancer (Equipe labellisée), the Institut National du Cancer, and the Association Française contre les Myopathies. C.G.R. was supported by the Institut National de la Santé et de la Recherche Médicale. A.E. was supported by grants from the Institut Pasteur, the Center National de la Recherche Scientifique, and the Schlumberger Foundation for Education and Research.
This article was published online ahead of print in MBoC in Press (http://www.molbiolcell.org/cgi/doi/10.1091/mbc.E12-02-0167) on November 28, 2012.