Crosstalk between microtubules and actin filaments is crucial for various cellular functions, including cell migration, spreading and cytokinesis. The Rac1 GTPase plays a key role in such crosstalk at the leading edge of migrating cells in order to promote lamellipodial formation. However, the mechanism underlying the link between microtubules and Rac1 activation remains unclear. Here, we show that calpain-6 (CAPN6), a non-proteolytic calpain with microtubule-binding and -stabilizing activity, might participate in this crosstalk. Small interfering RNA (siRNA)-induced knockdown of Capn6 in NIH 3T3 cells resulted in Rac1 activation, which promoted cell migration, spreading and lamellipodial protrusion. This increase in Rac1 activity was abolished by knockdown of the Rho guanine nucleotide exchange factor GEF-H1 (officially known as Arhgef2). CAPN6 and GEF-H1 colocalized with microtubules and also interacted with each other through specific domains. Upon knockdown of Capn6, GEF-H1 was shown to translocate from microtubules to the lamellipodial region and to interact with Rac1. By contrast, RhoA activity was decreased upon knockdown of Capn6, although low levels of active RhoA or the presence of RhoA molecules appeared to be required for the Capn6-knockdown-induced Rac1 activation. We suggest that CAPN6 acts as a potential regulator of Rac1 activity, through a mechanism involving interaction with GEF-H1, to control lamellipodial formation and cell motility.
Introduction
Crosstalk between the two major cytoskeletal components, actin filaments and microtubules, is essential for various cellular functions, including cell migration, spreading and cytokinesis. In migrating cells, actin polymerization generates lamellipodial membrane protrusions at the leading edge, and actomyosin contractility in the tail promotes cell-body advancement (Chhabra and Higgs, 2007). In concert with the actin cytoskeleton, the dynamics of microtubules contributes to the establishment of cell polarity and the directional movement of migrating cells. The initial polarization of microtubule assembly is led by actin filaments, and conversely the polarized microtubules contribute to the reorganization of actin filaments (Li and Gundersen, 2008; Rodriguez et al., 2003; Siegrist and Doe, 2007).
Rho family GTPases and their regulatory proteins have been postulated to play central roles in microtubule–actin crosstalk. Microtubule depolymerization, by nocodazole or colchicine, induces Rho activation with an increase in stress fiber formation and cellular contractility (Enomoto, 1996; Ren et al., 1999). This effect is mediated by the Rho guanine-nucleotide-exchange factor GEF-H1 (officially known as ARHGEF2), which is released from microtubules upon depolymerization (Chang et al., 2008; Krendel et al., 2002). By contrast, microtubular growth after washout of nocodazole activates Rac1, leading to actin polymerization in lamellipodial protrusions (Waterman-Storer et al., 1999). Although GEF-H1 has been suggested as a potential mediator of microtubule–actin crosstalk at the leading edge (Siegrist and Doe, 2007; Waterman-Storer et al., 1999), this has not been proven. GEF-H1 was originally reported to be a GEF for both RhoA and Rac1 (Ren et al., 1998), but subsequent reports have not demonstrated a GEF activity for Rac1 (Benais-Pont et al., 2003; Glaven et al., 1999; Krendel et al., 2002; Zenke et al., 2004). Thus, the effect of GEF-H1 on Rac1 activity might depend upon the presence or absence of regulatory factors (Birkenfeld et al., 2008). A recent report has demonstrated that GEF-H1 can promote Rac1 activation in the presence of the p21-activated kinase PAK4 (Callow et al., 2005), indicating that the activation of Rac1 by GEF-H1 might be conditionally regulated.
The calpains are a family of intracellular cysteine proteases whose activity is highly dependent upon Ca2+ ions (Croall and Ersfeld, 2007; Goll et al., 2003; Hanna et al., 2008). Approximately one-half of calpains share a common four-domain structure comprising domains I to IV: domain II is a cysteine protease domain; domain III is related to the C2 domain, a Ca2+- and phospholipid-binding module; and domain IV is characterized by the presence of multiple EF-hand motifs in some members, including the classical calpains (m- and μ-calpains) (Croall and Ersfeld, 2007; Goll et al., 2003). Among the 14 or 15 members of the calpain family in mammals, calpain-6 (CAPN6) is unique in that it lacks the active-site catalytic cysteine residue and is therefore not likely to be a proteolytic enzyme (Dear et al., 1997). In CAPN6, as well as Capn5, the C-terminal structure is defined as a diverged C2-domain (also called domain T) instead of as domain IV, on the basis of similarity to Caenorhabditis elegans TRA-3, a nematode sex determination factor (Barnes and Hodgkin, 1996; Dear et al., 1997; Goll et al., 2003; Mugita et al., 1997). Recently, we have demonstrated that CAPN6 can bind to microtubules, mainly through domain III, and induce microtubule stabilization through non-proteolytic activity (Tonami et al., 2007). Furthermore, inactivation of CAPN6 not only destabilizes microtubules but also promotes formation of lamellipodia and lamellipodial membrane ruffling (Tonami et al., 2007). These findings have led us to study the possibility that CAPN6 participates in microtubule–actin crosstalk in order to contribute to cellular functions. Here, we demonstrate that CAPN6 is a possible mediator of microtubule–actin crosstalk. RNA interference (RNAi)-induced knockdown revealed that CAPN6 suppresses Rac1 activity and lamellipodial formation, in a manner related to changes in microtubule dynamics, through interaction with GEF-H1. Biochemical and immunocytochemical experiments showed that there was a direct association between CAPN6 and GEF-H1. These results might provide a clue to previously unknown regulatory mechanisms that underlie microtubule–actin crosstalk in lamellipodial formation and cell motility.
Results
Knockdown of Capn6 promotes cell motility and spreading, with enhancement of lamellipodial membrane ruffling
To investigate the function of CAPN6 in cell motility, we performed Boyden chamber migration assays and scratch-wound-healing assays on NIH 3T3 cells transfected with small interfering RNA (siRNA) targeting Capn6 or with control siRNA. We designed two independent siRNAs for Capn6 (Capn6-1 and Capn6-2), both of which successfully downregulated CAPN6 protein levels (Fig. 1A). In the Boyden chamber assay, knockdown of Capn6 caused a marked enhancement in the three-dimensional migration of the cells in response to platelet-derived growth factor (PDGF)-BB (Fig. 1B). Consistent with this observation are the results of in vitro wound-healing assays, which showed that, upon knockdown of Capn6, cells exhibited extensive lamellipodia formation at the leading edge and a significant increase in the migration distance compared with that of control siRNA-treated cells (Fig. 1C,D). Enhanced cell movement upon knockdown of Capn6 was further confirmed by cell tracking in the random migration assay (Fig. 1E,F). We also examined whether CAPN6 was involved in cell spreading, which shares common mechanisms regulating cytoskeletal organization with those in cell migration (Huveneers and Danen, 2009). For the evaluation of spreading, cells were induced to round up (but not detach) by brief trypsinization and were allowed to re-spread for 2 hours. Knockdown of Capn6 significantly promoted re-spreading of rounded cells (Fig. 1G; supplementary material Fig. S1).
Promotion of cell migration and spreading in cells upon knockdown of Capn6 was accompanied by activation of lamellipodial protrusion and membrane ruffling at the cell edges (Fig. 1H, supplementary material Fig. S2A), as previously described (Tonami et al., 2007). Although these effects are unlikely to be due to an siRNA-induced inhibition of proteolytic activity, because CAPN6 lacks the active-site catalytic cysteine residue, they might be due to derepression of other calpains that are ‘blocked’ by CAPN6. However, the lamellipodial formation induced by knockdown of Capn6 was not affected by treatment with the calpain inhibitors benzyloxycarbonyl-L-leucyl-L-leucinal (ZLLal) or calpeptin (supplementary material Fig. S2B). These findings suggest that CAPN6 is involved in the regulation of the organization of cortical actin and subsequent changes in cell motility and morphology, independent of the proteolytic activity of calpains.
Rac1 activation is involved in the enhancement of cell motility and lamellipodial protrusion observed upon knockdown of Capn6
The enhancement of lamellipodial formation and cell motility described above led us to speculate that Rac1 is activated upon knockdown of Capn6. We therefore estimated Rac1 activity by measuring the levels of active GTP-bound Rac1 using a pull-down assay with PAK1-PBD (the PAK1 p21-binding domain) coated beads. As expected, Rac1 activity was upregulated upon knockdown of Capn6, whereas the total amount of Rac1 was unchanged (Fig. 2A,B). The Capn6-siRNA-induced Rac1 activation was not affected by ZLLal (supplementary material Fig. S2C), as in the case of lamellipodial protrusion. Conversely, overexpression of CAPN6 decreased the amount of Rac1-GTP, although the efficiency was different between the two CAPN6 fusion proteins (GFP–CAPN6 and CAPN6–Myc) (Fig. 2C), indicating that CAPN6 might act as a repressor of Rac1 activity. To test whether increased Rac1 activity was responsible for the effect of Capn6 knockdown, we designed siRNAs targeting Rac1. The efficiency of the Rac1-siRNAs was confirmed by western blotting using a mouse polyclonal anti-Rac1 antibody (Fig. 2D). Co-transfection of Rac1- and Capn6-siRNA significantly suppressed the enhancement of cell motility and lamellipodial protrusions induced by knockdown of CAPN6 (Fig. 2E,F; supplementary material Fig. S3A). We also tested whether the effect of Rac1-siRNA was reproduced upon treatment of cells with the Rac1 inhibitor NSC23766. As expected, NSC23766 effectively suppressed the enhancement of lamellipodia and cell motility caused by Capn6-siRNA (Fig. 2G,H; supplementary material Fig. S3B). These results show that Rac1 activity is involved in the lamellipodial formation and enhanced cell motility that is induced by knockdown of CAPN6.
GEF-H1 mediates Rac1 activation induced by knockdown of Capn6
To explore the link between CAPN6 and Rac1, we focused on GEF-H1, which is known to be associated with microtubules, as is CAPN6. Signals for endogenous CAPN6 and GEF-H1 were partially overlapped within the cytoplasm and were observed in a reticular pattern (Fig. 3A). The specificity of staining for CAPN6 and GEF-H1 and their colocalization to microtubules was confirmed by immunostaining after siRNA knockdown, together with western blotting for GEF-H1 (supplementary material Figs S4, S5). To test whether GEF-H1 could activate Rac1, we transfected a GEF-H1–His6 expression vector into control cells and cells subjected to siRNA knockdown of Capn6. GEF-H1 overexpression increased Rac1 activity both in the control cells and in cells with Capn6 knockdown (Fig. 3B). When each of the two siRNAs targeting GEF-H1 (Gef-h1-1 and Gef-h1-2) was introduced into Capn6-knockdown cells, it significantly suppressed the Rac1 activation induced by Capn6 siRNA (Fig. 3C,D). Furthermore, knockdown of GEF-H1 plus Capn6 suppressed the enhancement of cell migration and lamellipodial protrusions (Fig. 3E,F; supplementary material Fig. S6). GEF-H1 knockdown appeared to decrease slightly the amount of GTP-Rac1 under serum-free conditions (Fig. 3C,D), but this effect was marginal. GEF-H1 knockdown did not affect Rac1 activity in PDGF-stimulated cells (data not shown), indicating that GEF-H1 is not involved in PDGF-induced Rac1 activation. These results suggest that GEF-H1 is involved in the Rac1 activation induced by knockdown of CAPN6.
CAPN6 promotes microtubular association of GEF-H1 and prevents it from activating Rac1-mediated microtubule–actin crosstalk
To examine the effect of Capn6 knockdown on the intracellular behavior of GEF-H1, we stained control and Capn6-knockdown cells for endogenous GEF-H1 and compared its localization with the distribution of microtubules and actin filaments. In control siRNA-treated cells, GEF-H1 was largely colocalized to microtubules (Fig. 4A,B; supplementary material Fig. S7). However, upon knockdown of Capn6, GEF-H1 appeared to re-distribute from microtubules into the lamellipodial region (Fig. 4C,D; supplementary material Fig. S7), where it colocalized with peripheral actin filaments (Fig. 4E–H). By contrast, such translocation of GEF-H1 was hardly ever observed in lamellipodia induced by PDGF stimulation (supplementary material Fig. S8).
To confirm further the behavior of GEF-H1, we overexpressed GEF-H1–His6 in control cells and in cells subjected to Capn6 knockdown. GEF-H1–His6 also changed its distribution pattern, re-distributing from microtubules into the lamellipodial region upon Capn6 knockdown (Fig. 4I–L), which led to the colocalization of GEF-H1 with lamellipodial actin filaments (Fig. 4M–P). These observations indicate that GEF-H1 is readily mobilized from the microtubule network to peripheral actin filaments in the absence of CAPN6.
We investigated further whether CAPN6 regulates the interaction between GEF-H1 and microtubules. Cell lysates from control or Capn6-siRNA-transfected NIH 3T3 cells were subjected to a microtubule co-sedimentation assay. Co-sedimentation of both GEF-H1 and CAPN6 in the microtubule-containing pellet was intensified in the presence of paclitaxel, a microtubule-stabilizing agent (Fig. 5A). In Capn6-siRNA-transfected cells, the relative amount of GEF-H1 associated with microtubules was significantly decreased (Fig. 5A,B). In this assay, the total amount of microtubules sedimented (stabilized microtubules) was the same in control and Capn6-siRNA-treated cell lysates when in the presence of paclitaxel (Fig. 5A), indicating that the amount of GEF-H1 binding to microtubules was decreased in the absence of CAPN6. Furthermore, GEF-H1 was co-precipitated with the Rac1-GTP complex from cell lysates from Capn6-knockdown cells in a Rac1-activity assay (Fig. 5C). These results suggest that CAPN6 promotes GEF-H1–microtubule association and prevents GEF-H1 from activating Rac1-mediated microtubule–actin crosstalk.
CAPN6 physically interacts with GEF-H1
Next, we overexpressed Myc-tagged CAPN6 and His6-tagged GEF-H1 in NIH 3T3 cells to test whether CAPN6 might physically interact with GEF-H1 in vivo. GEF-H1–His6 was detected in the immunoprecipitate from the anti-Myc antibody (Fig. 6A). We further examined whether endogenous CAPN6 and GEF-H1 associated with each other. As expected, GEF-H1 was co-precipitated with CAPN6 (i.e. it was detected with immunoprecipitates from the anti-CAPN6 antibody) (Fig. 6B). This association between CAPN6 and GEF-H1 in vivo was not affected by treatment with nocodazole (supplementary material Fig. S9), suggesting that this association is independent of microtubules. To examine whether CAPN6 and GEF-H1 could directly interact with each other and, if so, which domain was involved, we performed a GST pull-down assay using various GST–CAPN6 mutants coexpressed with GEF-H1–His6protein in NIH 3T3 cells (Fig. 6C). GEF-H1–His6 was pulled down by GST-fused full-length CAPN6 and the CAPN6 domains II, III and T (Fig. 6C). We also found that the precipitates from the GST-fused full-length CAPN6 and the CAPN6 domains III and T contained β-tubulin (Fig. 6C), indicating that β-tubulin might mediate the pull-down of GEF-H1–His6. This possibility is supported by our previous finding that CAPN6 binds to microtubules through domain III and domain T (Tonami et al., 2007). By contrast, β-tubulin was not detected in the precipitate of GST-fused CAPN6 domain II (Fig. 6C). The same result was reproduced by using in-vitro-translated GEF-H1–His6 protein (data not shown). To determine which domain(s) of GEF-H1 were responsible for the interaction with CAPN6 domain II, we performed GST pull-down assays using GST-fused CAPN6 domain II and in-vitro-translated GEF-H1 derivatives (Fig. 6D). This mapped the association between GEF-H1 and CAPN6 domain II to the zinc-finger-containing N-terminal region (amino acids 1–240) of GEF-H1 (Fig. 6D). These results suggest that CAPN6 and GEF-H1 directly interact, most probably through domain II and the zinc-finger-containing N-terminal region, respectively.
Knockdown of Capn6 suppresses RhoA activity
It is well known that GEF-H1 is a RhoA GEF and increases RhoA activity when it is released from microtubules (Birkenfeld et al., 2008; Chang et al., 2008; Krendel et al., 2002). Therefore, we examined whether RhoA activity might be increased upon knockdown of CAPN6. Unexpectedly, RhoA activity was largely decreased in cells upon Capn6 knockdown (Fig. 7A). This effect was also seen upon treatment with NSC23766 (Fig. 7B), indicating that the decrease in RhoA activity was not due to Rac1 activation. It has been suggested recently that RhoA inactivation might increase Rac1 activity because RhoA suppresses Rac1 through a Rho-associated protein kinase (ROCK)-dependent mechanism (Narumiya et al., 2009). To test this possibility, we compared the contribution of GEF-H1 in Rac1 activation between cells upon Capn6 knockdown and cells treated with Y-27632, a ROCK inhibitor. We found that the Y-27632-induced lamellipodial formation was not suppressed by Gef-h1-siRNA (supplementary material Fig. S10). Lamellipodial translocation of GEF-H1 was not observed in Y-27632-treated cells (supplementary material Fig. S8), indicating that GEF-H1-dependent Rac1 activation upon Capn6 knockdown is not likely to be due to ROCK inactivation. Interestingly, Capn6-knockdown-induced Rac1 activation was not observed upon co-transfection of RhoA siRNA (Fig. 7C), but was still apparent upon treatment of cells with Y-27632 (supplementary material Fig. S11A). The lamellipodial formation induced by Capn6 siRNA was also insensitive to Y-27632 (supplementary material Fig. S11B). Taken together, these results suggest that knockdown of CAPN6 causes suppression of RhoA activity independently of Rac1 activity, but that a ROCK-independent activity of RhoA or the presence of RhoA molecules, even if they are not activated, is necessary for GEF-H1-mediated Rac1 activation.
Capn6 expression is suppressed by serum
To investigate the physiological relevance of the GEF-H1-mediated Rac1 activity induced by the knockdown of CAPN6, we searched for factors that could affect CAPN6 expression. Notably, we found that Capn6 mRNA and CAPN6 protein levels were downregulated by serum (Fig. 8A,B). To confirm the suppressive effect of serum on Capn6 expression, we stimulated NIH 3T3 cells with fetal calf serum (FCS) for the indicated times after a 12-hour starvation. We found that Capn6 mRNA levels were decreased by the addition of 10% FCS, in a time-dependent manner (Fig. 8C), and that this was accompanied by a decrease in CAPN6 protein levels (Fig. 8D). These results suggest that CAPN6–GEF-H1–Rac1 signaling might be regulated by serum-derived factor(s).
Discussion
In the present study, we characterize CAPN6, a non-proteolytic calpain that is associated with microtubules, as a regulator of Rac1-mediated lamellipodial formation and cell motility. siRNA-mediated Capn6 knockdown resulted in Rac1 activation and subsequent enhancement of lamellipodial formation and cell motility. This Capn6-knockdown-induced Rac1 activation was dependent on GEF-H1. Biochemical evidence indicated that there was a direct interaction between CAPN6 and GEF-H1 through specific domains. In the absence of CAPN6 activity, GEF-H1 appears to be dissociated from the microtubules and translocated to the cortical actin network, where lamellipodial formation is upregulated. CAPN6 expression was downregulated by serum, which activates Rac1 and cell motility, suggesting that the CAPN6–GEF-H1–Rac1 regulatory pathway might contribute to the serum-dependent control of cell motility.
CAPN6 is a unique member of the calpain family – one of the crucial catalytic cysteine residues is replaced with lysine in humans and mice, whereas this residue is conserved in all other mammalian calpain members (Dear et al., 1997). Indeed, the present study and our previous results (Tonami et al., 2007) demonstrate that the effects of Capn6 knockdown and overexpression are not affected by calpain inhibitors, indicating that the microtubule association and/or stabilization, and the GEF-H1-mediated Rac1 activation seen upon Capn6 knockdown, are independent of the proteolytic activity. Recently, non-proteolytic functions of other mammalian calpains have been suggested following gene-targeting studies in mice (Hata et al., 2010; Ojima et al., 2010). Although calpains have proteolytic activities, they might also function as structural regulators and chaperones for other molecules. These findings might shed light on the concealed side of non-proteolytic calpain functions.
Biochemical analysis and immunostaining indicated that there was a direct interaction between CAPN6 and GEF-H1. The GST pull-down assay suggests that it was domain II of CAPN6, a cysteine protease domain in other calpain members, and the zinc-finger-containing N-terminal region of GEF-H1, that are most likely to mediate this interaction. We have previously reported that CAPN6 stabilizes microtubules through interactions mediated by domain III and, to a lesser extent, domain T (Tonami et al., 2007). However, the microtubule binding of GEF-H1 involves the N-terminal region, as well as the C-terminal and Dbl-homology domains (Krendel et al., 2002; Zenke et al., 2004). Mutations in the N-terminal region of GEF-H1 abolish its binding to microtubules and, thereby, increase its enzymatic activity (Krendel et al., 2002; Zenke et al., 2004). Thus, CAPN6 might reinforce the association between GEF-H1 and microtubules, and negatively regulate GEF-H1 activity by interacting with this region, although it remains to be resolved whether microtubule binding is necessary for the inhibition of GEF-H1 by CAPN6. The zinc-finger-containing N-terminal region of GEF-H1 contains a domain similar to the C1 diacylglycerol-binding domain of the atypical protein kinase C (aPKC) family (Birkenfeld et al., 2008). Interestingly, CAPN6 domains III and T have a similarity to the C2 domain, a Ca2+- and phospholipid-binding module (Croall and Ersfeld, 2007; Goll et al., 2003). The C1 and C2 domains are known to interact within several PKC isozymes in order to determine selective lipid binding and membrane interactions (Colon-Gonzalez and Kazanietz, 2006). Thus, the interaction between CAPN6 and GEF-H1 might involve the C1 and C2 domains from each other assembling to create a functional unit.
The absence of CAPN6 is likely to cause increased microtubule instability and release of GEF-H1. Previous studies have shown that GEF-H1 is released and activated upon microtubule depolymerization (Birukova et al., 2006; Chang et al., 2008; Krendel et al., 2002). Therefore, a destabilization of microtubules caused by inactivation of CAPN6 might be responsible for GEF-H1-dependent Rac1 activation. Furthermore, the microtubule co-sedimendation assay in the presence of paclitaxel demonstrated that the affinity of GEF-H1 for stabilized microbtubules was decreased upon knockdown of Capn6, suggesting that CAPN6 potentiates the GEF-H1–microtubule association. However, nocodazole-induced microtubule destabilization did not recapitulate the effect of CAPN6 knockdown on Rac1 activation and lamellipodial formation (K.T., unpublished data). Thus, destabilization of microtubules is insufficient for GEF-H1-dependent Rac1 activation and lamellipodial formation. Inactivation or disappearance of CAPN6 and/or subsequent cytoskeletal changes, such as enhanced microtubule turnover, are required for translocation of GEF-H1 to the leading edge and Rac1 activation. CAPN6 might prevent GEF-H1 from translocating to the lamellipodia, to activate Rac1, even in a microtubule-free state. We present a model of CAPN6-regulated microtubule–actin crosstalk mediated by GEF-H1 in Fig. 9.
Although GEF-H1 released from microtubules has been shown to activate RhoA in some studies (Birukova et al., 2006; Chang et al., 2008; Krendel et al., 2002), RhoA activity was downregulated in cells upon Capn6 knockdown. In this situation, GEF-H1 released from microtubules might fail to access and/or activate RhoA and would then interact with Rac1 and activate it. Although the mechanism of Capn6-knockdown-induced RhoA inactivation remains unknown, it might explain why GEF-H1 activates Rac1 instead of RhoA upon Capn6 knockdown.
At the moment, it remains unclear whether the Rac1 activation by GEF-H1 is a direct effect or whether it is mediated by RhoA. Contrary to the downregulation of RhoA activity, Rac1 activation in cells upon Capn6 knockdown was abolished by RhoA knockdown, indicating that this Rac1 activation appears to be dependent on RhoA. Recently, Nalbant and colleagues have reported that siRNA-induced GEF-H1 depletion causes localized RhoA inactivation at the leading edge, resulting in decreased directional migration in HeLa cells (Nalbant et al., 2009). Other reports have demonstrated that RhoA is activated prior to Rac1 and Cdc42, and functions upstream of Rac1 and Cdc42 activation at the leading edge (El-Sibai et al., 2008; Machacek et al., 2009). Although the RhoA–ROCK pathway suppresses Rac1 activity, RhoA can activate Rac1 through mammalian diaphanous homolog 1 (DIA1, also known as DIAP1), a formin family member that catalyzes actin nucleation and polymerization (Watanabe et al., 1997), when the level of RhoA activity is low (Narumiya et al., 2009; Tsuji et al., 2002). Indeed, DIA1, but not ROCK, is associated with membrane ruffles in nascent lamellipodia (Kurokawa and Matsuda, 2005) and it stabilizes microtubules specifically in leading edge adhesions (Palazzo et al., 2001). Together with these findings, our present data suggest that, in the absence of CAPN6, GEF-H1 could activate Rac1, possibly through a RhoA-DIA1-dependent mechanism at the leading edge, in a spatiotemporally controlled manner in certain types of cells. Further experiments are required to determine how RhoA activity is suppressed upon CAPN6 inactivation, whether GEF-H1 can directly catalyze nucleotide exchange on Rac1 and how GEF-H1-induced RhoA and Rac1 activation is interrelated.
Here, we propose a function for non-proteolytic calpains in microtubule–actin crosstalk which is mediated through Rho family GTPase regulation. CAPN6 shares common structural characteristics with other calpain family members, such as Capn5 and C. elegans TRA-3 (Barnes and Hodgkin, 1996; Dear et al., 1997; Mugita et al., 1997), constituting an evolutionarily conserved subfamily. Although CAPN5 and TRA-3 are proteolytic calpains, it might be interesting to examine whether this non-proteolytic activity is also shared by these calpains and whether it is conserved across species.
Materials and Methods
Reagents
NSC23766, Y-27632 and calpeptin were purchased from Calbiochem. ZLLal was from Peptide Institute (Osaka, Japan). PDGF-BB, paclitaxel and nocodazole were from Sigma.
Plasmids
Wild-type, green fluorescent protein (GFP)-fused and glutathione transferase (GST)-fused CAPN6 expression vectors have been previously described (Tonami et al., 2007). For the expression of His6- or Myc-tagged proteins, the open reading frames of murine Capn6 and GEF-H1 were subcloned, in frame, into the pcDNA3.1/V5-His TOPO (Invitrogen) or Myc-containing pcDNA3 vector (Invitrogen). All of the constructs were verified by sequencing.
RNAi
All the stealth small interfering RNA (siRNA) duplexes were synthesized by Invitrogen. Two Capn6 siRNAs (Capn6-1 and Capn6-2) were designed to target nucleotides 1518–1542 and 1935–1959 of the mouse Capn6 mRNA sequence (GenBank accession no. NM_007603), respectively. Rac1 siRNA, targeting nucleotides 631–655 (GenBank accession no. NM_009007), and RhoA siRNA, targeting nucleotides 574–598 (GenBank accession no. NM_016802), were obtained from Invitrogen Stealth Select RNAi. For GEF-H1 knockdown, two siRNAs (Gef-h1-1 and Gef-h1-2) were designed to target nucleotides 1085–1109 and 3410–3434 of the mouse GEF-H1 mRNA sequence (GenBank accession no. NM_008487), respectively. As a negative control, siRNA with a scrambled sequence was prepared for each specific siRNA (Control-1 to Control-4). The sequences of siRNAs are shown in supplementary material Table S1.
Cell culture and transfection
NIH 3T3 cells were cultured in Dulbecco's modified Eagle's medium (DMEM) containing 10% fetal calf serum (FCS) and antibiotics at 37°C under 5% CO2. For transfection, cells were grown to 50–90% confluence and were treated with a mixture of plasmid DNA and Lipofectamine LTX (Invitrogen). The siRNAs were transfected into NIH 3T3 cells using Lipofectamine RNAiMAX (Invitrogen) according to the manufacturer's protocol. After 4 hours of incubation, the cells were re-fed with medium containing FCS and were allowed to recover for 18–48 hours.
Cell migration assays
Cell migration was evaluated by the use of three methods: a three-dimensional Boyden chamber assay, a two-dimensional scratch-wound-healing assay and a random migration assay. For the Boyden chamber assay, trypsinized cells were suspended in DMEM plus 0.1% BSA and were then added to the top chambers to give a final concentration of 1.5×105 cells per well. The bottom chambers were filled with DMEM plus 0.1% BSA containing 100 ng/ml PDGF-BB. The polycarbonate membrane filters (8-μm-pore size) (Transwell Permeable Supports) were precoated with 5 μg of type IV collagen per filter. The migration was evaluated after a 4-hour incubation at 37°C. For the scratch wound assay, cells transfected with siRNA were starved in DMEM plus 0.5% FCS for 24 hours before the assay. The confluent cell monolayer was wounded with a plastic cell scraper. The remaining cells were washed twice with culture medium to remove cell debris and incubated at 37°C. Cell migration was evaluated at the wounded front at 6 hours after scratching. For the random migration assay, cells transfected with siRNA were cultured in DMEM plus 1% FCS for 16 hours before the assay. The cells were tracked for 4 hours at 5-minute intervals on an Olympus LCV100 microscope and were analyzed with MetaMorph software (Universal Imaging, Molecular Devices).
Cell spreading assay
At 48 hours after siRNA transfection, cells were incubated in trypsin-EDTA, until rounding but not detachment was observed, then the trypsin was carefully aspirated and fresh medium with 10% serum was added to stop the reaction. Cells were visualized by phase-contrast microscopy and scored for the percentage of spread cells at the indicated times.
Immunofluorescence microscopy
Cells were washed with preincubated general tubulin buffer (GTB) {80 mM PIPES [piperazine-N, N′-bis (2-ethanesulfonic acid)] pH 7, 1 mM MgCl2 and 1 mM EGTA} at 37°C and fixed with either an ice-cold methanol and acetone mixture (1:1; for 10 minutes) or 4% paraformaldehyde in GTB for 15 minutes at room temperature. Paraformaldehyde-fixed samples were permeabilized with 0.2% Triton X-100 in GTB. Fixed cells were then washed, with GTB at room temperature, three times. After being blocked with 5% skimmed milk powder or 3% goat serum in GTB, the cells were incubated with primary antibodies against the following proteins or tags: GFP (rabbit polyclonal; MBL), His6 (mouse monoclonal; Invitrogen), α-tubulin (mouse monoclonal, Sigma), β-tubulin (rabbit polyclonal; Sigma), GEF-H1 (mouse monoclonal; Abcam), anti-CAPN6 domain II (rabbit polyclonal Abcam) and the CAPN6 C-terminus (rabbit polyclonal; Transgenic). Then, the cells were washed, and stained with FITC- or rhodamine-conjugated donkey anti-(mouse IgG) or -(rabbit IgG) (Jackson Immunoresearch), Alexa-Fluor-488- or -555-conjugated goat anti-(mouse IgG) or -(rabbit IgG) (Molecular Probes) or DyLight-405-conjugated goat anti-(mouse IgG) (Thermo Scientific) secondary antibodies. For actin staining, 2.5 units/ml of rhodamine- or FITC-labeled phalloidin (Invitrogen) was added to the reaction buffer containing FITC-conjugated secondary antibody. The cells were viewed using a fluorescence microscope (Nikon TE300), a confocal microscope (Nikon D-ECLIPSE C1) or a LSM510 META laser-scanning confocal microscope (Carl Zeiss).
Immunoprecipitation
Cells were lysed with 0.1% NP-40 in GTB and were subjected to immunoprecipitation with the indicated antibodies using standard procedures. Myc-tagged CAPN6 was precipitated with an anti-Myc antibody (Upstate) prebound to agarose beads. For immunoprecipitation of endogenous CAPN6, the rabbit polyclonal anti-CAPN6 antibodies (N- and C-terminal specific; Transgenic) were used. Normal rabbit IgGs (Santa Cruz Biotechnology) served as negative controls.
Western blotting
For whole-cell lysate preparation, cells were solubilized in PBS containing 1% Triton X-100, 0.1% sodium deoxycholate, 0.02% SDS, 1 mM phenylmethylsulfonyl fluoride, 0.5 mM vanadate and protease inhibitor cocktail (Sigma). Cell lysates and immunoprecipitates were separated by SDS-PAGE (7.5% or 15% gels), electrotransferred onto a polyvinylidene difluoride membrane and then subjected to immunoblotting with primary antibodies against the following proteins or tags: CAPN6 (rabbit polyclonal; Transgenic), GFP (rabbit polyclonal; MBL), GEF-H1 (rabbit monoclonal; Cell Signaling Technology), GEF-H1 (mouse monoclonal; Abcam), His6 tag (mouse monoclonal; Invitrogen), RhoA (mouse polyclonal; Santa Cruz Biotechnology), Rac1 (mouse polyclonal; Upstate), β-actin (mouse monoclonal; Sigma) and anti-tubulin or anti-(acetylated tubulin) antibodies (mouse monoclonals; Sigma). The membranes were then washed with 0.1% Tween 20 in Tris-buffered saline pH 7.6, and incubated with peroxidase-conjugated anti-(rabbit IgG) or -(mouse IgG) antibody (DAKO). The signals were detected using the enhanced chemiluminescence detection system (Amersham Bioscience) or the POD Immunostain set (Wako). The signal intensity was quantified with ImageJ 1.43 (NIH).
Rac1 and Rho activity assays
Cells were cultured in serum-free DMEM, except during experiments using Y-27632, when the DMEM was supplemented with 1% FCS. Rac1-GTP and Rho-GTP were quantified with the Rac/Cdc42 assay (PAK1-PBD, agarose conjugate) and Rho activation assay reagents (Rhotekin-RBD, agarose conjugate) (Upstate), respectively. Briefly, the cells were harvested in lysis buffer [25 mM HEPES pH 7.5, 150 mM NaCl, 1% Igepal CA-630, 10 mM MgCl2, 1 mM EDTA, 10% glycerol, 1 mM phenylmethylsulfonyl fluoride, 0.5 mM vanadate and protease inhibitor cocktail (Sigma)] and then cleared lysates were incubated with the Rac/Cdc42 assay reagent or the Rho activation assay reagent for 1 hour at 4°C. Beads were washed three times with lysis buffer, and the proteins bound to the beads were separated by SDS-PAGE (15% gels) and analyzed by immunoblotting with polyclonal antibodies against Rac1 or RhoA.
GST pull-down assay
For the cell lysate preparation, NIH 3T3 cells were solubilized in pull-down buffer (20 mM Tris-HCl pH 7.6, 1% Triton-X-100, 0.25% sodium deoxycholate and 0.25 M NaCl) containing 1 mM phenylmethylsulfonyl fluoride, 0.5 mM vanadate and protease inhibitor cocktail (Sigma). Unbroken cells and cellular debris were removed by centrifugation at 20,000 g at 4°C for 15 minutes. Then, GST-fusion proteins, bound to the beads, were mixed with the lysates and incubated at 4°C for 6 hours. The beads were washed three times with pull-down buffer containing protease inhibitors, and the bound proteins were eluted by adding 2.5× sample buffer and boiling for 5 minutes. These samples were subjected to SDS-PAGE (7.5% gels) and proteins were detected by western blotting with antibodies against His6 or GEF-H1. The amounts of GST-fusion proteins were grossly estimated by Ponceau staining.
Microtubule co-sedimentation assay
Cells were lysed, with 1% NP-40 in GTB containing protease inhibitor cocktail (Sigma), by passing them through a 30-gauge syringe needle. After incubation on ice for 30 minutes, to depolymerize the microtubules, the lysate was centrifuged at 20,000 g for 80 minutes at 4°C to remove cellular debris. The supernatant was diluted with GTB and 0.1% NP-40, and then divided into two tubes. 20 μM Paclitaxel (Taxol) or vehicle [dimethyl sulfoxide (DMSO)] was added to each sample. After an incubation for 30 minutes at 37°C, the reaction mixture was centrifuged at 20,000 g for 40 minutes at room temperature. The resultant pellets were resuspended in lysis buffer, separated by SDS-PAGE (7.5% gels) and subjected to western blotting analysis with whole lysates or supernatants. The relative amounts of microtubule-associated GEF-H1 were estimated by the comparison of the signal intensity of GEF-H1 in pellets divided by that of total α-tubulin in whole lysates (inputs).
RT-PCR
Total RNA was extracted from NIH 3T3 cells with the use of TRIzol reagent (Invitrogen), and samples (4 μg) were then reverse-transcribed with the use of the first-strand cDNA synthesis kit (GE Healthcare) with oligo(dT) primers. The resultant cDNAs were amplified with Ex Taq polymerase (Takara) in a thermocycler. The primer sequences were 5′-GAATTCATGGGTCCTCCTCTGAAGCT-3′ (sense) and 5′-GAATTCGAGCTCAGTGAGATCATCGC-3′ (antisense) for the mouse Capn6 mRNA, and 5′-GGTGTGAACCACGAGAAATAT-3′ (sense) and 5′-AGATCCACGACGGACACATT-3′ (antisense) for mouse Gapdh mRNA. Thermal cycling was performed for 18 to 23 cycles, to maintain PCR conditions within the linear range of amplification before saturation was reached. Each cycle consisted of 30 seconds of denaturation at 94°C, 30 seconds of annealing at 58°C and 2.5 minutes (for Capn6) or 30 seconds (for Gapdh) of extension at 72°C.
Statistical analysis
Data are expressed as means±s.e.m. Comparisons between two groups were performed using Student's t-tests, whereas multiple comparisons between more than two groups were analyzed by one-way ANOVA and post hoc tests. A Mann–Whitney U-test was used to analyze cell tracking in the random migration assay. Values of P<0.05 were considered statistically significant differences.
Acknowledgements
We are grateful to all the members of the Department of Physiological Chemistry and Metabolism, The University of Tokyo and the Calpain Project (Rinshoken, Tokyo, Japan) for valuable support and discussions. This work was supported in part by the Global COE Program (Integrative Life Science based on the Study of Biosignaling Mechanisms) of the Ministry of Education, Culture, Sports, Science and Technology (MEXT), Japan, and by grants-in-aid for scientific research from MEXT (grant no. 18076007 to H.S.), grants-in-aid for scientific research from the Japan Society for the Promotion of Science (grant nos 22790272 to K.T., 20590275 to Y.K., 20370055 to H.S. and 21390238 to H.K.) and grants-in-aid for scientific research from the Ministry of Health, Labour and Welfare of Japan (grant no. 09156294 to H.K.).