Transient receptor potential canonical (TRPC) channels provide cation and Ca2+ entry pathways, which have important regulatory roles in many physio-pathological processes, including muscle dystrophy. However, the mechanisms of activation of these channels remain poorly understood. Using siRNA, we provide the first experimental evidence that TRPC channel 1 (TRPC1), besides acting as a store-operated channel, represents an essential component of stretch-activated channels in C2C12 skeletal myoblasts, as assayed by whole-cell patch-clamp and atomic force microscopic pulling. The channel's activity and stretch-induced Ca2+ influx were modulated by sphingosine 1-phosphate (S1P), a bioactive lipid involved in satellite cell biology and tissue regeneration. We also found that TRPC1 was functionally assembled in lipid rafts, as shown by the fact that cholesterol depletion resulted in the reduction of transmembrane ion current and conductance. Association between TRPC1 and lipid rafts was increased by formation of stress fibres, which was elicited by S1P and abolished by treatment with the actin-disrupting dihydrocytochalasin B, suggesting a role for cytoskeleton in TRPC1 membrane recruitment. Moreover, TRPC1 expression was significantly upregulated during myogenesis, especially in the presence of S1P, implicating a crucial role for TRPC1 in myoblast differentiation. Collectively, these findings may offer new tools for understanding the role of TRPC1 and sphingolipid signalling in skeletal muscle regeneration and provide new therapeutic approaches for skeletal muscle disorders.
- Sphingosine 1-phosphate
- C2C12 myoblasts
- Skeletal myogenesis
- Stretch-activated channels (SACs)
- Store-operated channels (SOCs)
- Lipid microdomains
It is well established that activation of several intercellular signalling pathways and the correct localization of specific intracellular targets are crucially involved in skeletal muscle differentiation (Formigli et al., 2007a; Le Grand and Rudnicki, 2007). It has been recently reported that activation of stretch-activated channels (SACs) may also play a role in muscle biology and myogenesis (Formigli et al., 2007b; Wedhas et al., 2005). SACs are non-selective ion channels that are present in a huge variety of cells and organisms, where they participate in mechanotransduction, converting membrane stretch into electrical Ca2+, K+ and Na+ current across the plasma membrane, and biochemical signals (Ambudkar, 2006; Sachs and Morris, 1998). In particular, SAC expression and activity are prevalent in neonatal skeletal muscle tissue and undifferentiated myoblasts, and decline as myoblasts mature and fuse into myotubes (Formigli et al., 2007b; Haws and Lansman, 1991). Moreover, the inhibition of these channels with several pharmacological compounds, including gadolinium chloride (GdCl3), streptomycin and the spider venom toxin (GsMTx4), has been shown to suppress the expression of the typical myogenic markers and the myoblast-myotube transition, suggesting that SAC activation may lead to increased Ca2+ entries necessary for skeletal muscle differentiation (Formigli et al., 2007b; Wedhas et al., 2005). Notably, there is also evidence supporting a role for SACs in the Ca2+-handling abnormalities observed in muscle diseases such as mdx mouse and human Duchenne dystrophies, stretch-induced muscle damage and cardiac hypertrophy (Allen et al., 2005; Lammerding et al., 2004).
A very important recent development in the understanding of the functional significance of SACs has been the identification of genes encoding members of the transient receptor potential canonical (TRPC) channel family (Clapham, 2003). TRPCs (TRPC1-7) assemble as homo- or hetero-tetramers to form voltage-independent, non-selective cation-permeable pores (Clapham, 2003; Rychkov and Barritt, 2007). Although numerous studies have shown that TRPC1 is a strong candidate for store-operated Ca2+ entry channels (SOCs) in various cell types (Fiorio Pla et al., 2005; Wu et al., 2004), there is also evidence that this channel subunit may function as an SAC. Indeed, TRPC1 is able to translate membrane stretch into cation currents across the plasma membrane (Clark et al., 2008; Maroto et al., 2005; Stiber et al., 2008), and is involved in abnormal SAC activity in mdx mice (Gervasio et al., 2008; Williams and Allen, 2007).
Despite the growing interest in TRPC1 channels, the channel gating, potential regulators and downstream cellular pathways of SAC activation have not been clarified as yet. Among the potential mechanisms, we have demonstrated that cytoskeletal remodelling and stress fibre (SF) formation are able to modulate SAC opening by transmitting tangential forces to the plasma membrane of C2C12 myoblasts (Formigli et al., 2005; Formigli et al., 2007b; Sbrana et al., 2008). In line with this, changes in gating behaviour of SACs observed in dystrophic fibres have been correlated with the irreversible destruction of cortical actin cytoskeleton and with alterations of membrane surface tension and/or composition (Gervasio et al., 2008; Lansman and Franco-Obregón, 2006; Stiber et al., 2008). Sphingolipids are an evolutionarily conserved class of membrane lipids synthesized by all eukaryotic cells. The biological functions of sphingolipids are diverse, encompassing structural roles through their participation in membrane lipid rafts, and signalling role via the involvement of their metabolites in signal transduction pathways. An important sphingolipid metabolite is sphingosine 1-phosphate (S1P), which mainly acts through G-protein-coupled receptors present on mammalian cells, thereby regulating numerous cell functions, including cell proliferation, differentiation and apoptosis (Zeidan and Hannun, 2007). In the last few years, most of the studies of our group have been focused on the role played in skeletal muscle cell biology by S1P, as a modulator of intracellular Ca2+ mobilization, phosphatidic acid synthesis and phospholipid remodelling in skeletal muscle cells (Bencini et al., 2003; Donati et al., 2005; Formigli et al., 2002; Meacci et al., 1999; Meacci et al., 2002a; Meacci et al., 2002b). Recently, we have underscored the physiological relevance of the sphingosine-kinase-S1P axis in C2C12 cell growth arrest and differentiation (Meacci et al., 2008) and identified the essential steps of the pro-myogenic action of S1P in skeletal myoblasts, including formation of SFs, upregulation of connexin 43 protein and its interaction with cortical actin (Formigli et al., 2007b; Squecco et al., 2006).
In the light of all these observations, it appeared to be worth further investigating the structure-function relationships of TRPC1 and its biological relevance in S1P signalling in skeletal myoblasts. To this aim, in the present study, we have demonstrated that TRPC1 acts as a mechano- and Ca2+-sensitive channel and is selectively localized in plasma membrane microdomains by actin remodelling. Notably, the anchorage of TRPC1 channel to lipid rafts represents a crucial mechanism involved in the regulation of channel function. Finally, we provided evidence for the regulation of TRPC1 expression and function by S1P and for its involvement in the accomplishment of skeletal myogenesis. These features make TRPC1 a novel, potential, molecular tool to counteract muscle damage in muscular dystrophies and improve skeletal muscle regeneration.
Expression of TRPC1 in skeletal myoblasts
We used RT-PCR to examine mRNA expression of mammalian TRPC isoforms in C2C12 myoblastic cells. As shown in Fig. 1A, TRPC1 was the predominant isoform in C2C12 myoblasts. Among the other investigated TRPC transcripts, only TRPC3 and TRPC5 could be detected using higher cDNA template amount (not shown), in agreement with previous data (Kunert-Keil et al., 2006; Woo et al., 2008). We then examined TRPC1 protein expression and cellular localization in C2C12 myoblasts by western blot analysis after partitioning proteins in cytosolic, Triton-soluble and Triton-insoluble membrane fractions. As reported in Fig. 1B, a single band corresponding to approximately 80 kDa, the predicted molecular weight of TRPC1, was mainly detected in membrane fractions, and appeared preferentially associated with detergent-insoluble membrane, a fraction largely corresponding to specific plasma membrane microdomains and cytoskeleton. TRPC1 was also detected, although only slightly, in the cytosolic fraction, as also reported for caveolin 1 (Uittenbogaard et al., 1998). To investigate the role of endogenous TRPC1, a mixture of three siRNA duplexes directed against the mouse TRPC1 gene (TRPC1-siRNA; 20 nM) and scrambled siRNA (SCR-siRNA; 20 nM), were transiently transfected into C2C12 cells. Western blots showed that the intensity of the band corresponding to TRPC1 was lower (approximately 65%) in TRPC1-siRNA compared with SCR-siRNA-transfected myoblasts (Fig. 1C). Using confocal microscopy, TRPC1 protein was found to be localized at or near the plasma membrane, consistent with its possible function as an ion channel, and within the cytoplasm, possibly representing a pool of TRPC1 present in the intracellular compartment and/or proteins in the process of being transferred to the apical membrane (Fig. 1D). To confirm the plasma membrane localization of TRPC1, we carried out confocal immunofluorescence on cells previously stained with TRITC-conjugated WGA, a conventional plasma-membrane marker (Sbrana et al., 2008). By superimposing fluorescence and DIC images (Fig. 1D, inset), a strong co-localization of WGA with endogenous TRPC1 was observed along the cell periphery. We also found that the exposure to TRPC1-siRNA dramatically attenuated (65% reduction) the expression level of the channel protein at the plasma membrane and in the intracellular compartments.
TRPC1 channel activity and its regulation by S1P in skeletal myoblasts
To find correlation between TRPC1 protein expression and channel activity, we first examined whether TRPC1 gene silencing abolished stretch-activated ion transients. For this purpose, native, SCR-treated and TRPC1-siRNA-treated cells were loaded with Fluo3-AM, a fluorescent Ca2+ indicator, and stretched with the cantilever of an atomic force microscope. Time-lapse recording showed that, native (not shown) and SCR-siRNA-treated C2C12 cells behaved similarly, showing the characteristic Ca2+ transients after the application of the mechanical stretch (Fig. 2A); the intracellular Ca2+ fluorescence increased close to the tip and then spread to adjacent regions of the cell. Notably, the general amplitude of the fluorescence changes was quite different between these cell populations and TRPC1-siRNA-transfected cells. In fact, after mechanical stretching, the relative fluorescence increase in TRPC1-silenced cells was reduced by approximately 90% (ΔF/Fb=0.44±0.05) compared with controls (ΔF/Fb=1.53±0.10). Moreover, as shown in the same figure, approximately 20% of the TRPC1-silenced cells were totally unresponsive. These results were confirmed by the electrophysiological recordings of SAC-mediated current after inducing mechanical stretch by two microelectrodes (Fig. 2B,C). Indeed, a stretch of 20% in SCR-siRNA cells caused a significant increase in transmembrane current (Im) and conductance Gm/Cm compared with the unstimulated cells at resting length (Fig. 2C, panels a,b; Table 1). By contrast, the mechanical stimulation was ineffective on Im and Gm/Cm variations in cells transfected with TRPC1-siRNA (Fig. 2C, panels c,d; Table 1), confirming the close relationship between TRPC1 expression and SAC amplitude.
We also found that TRPC1 silencing strongly reduced (by approximately 65%) the increase in SAC elicited by stimulation with S1P (1 μM, 30 minutes) (Fig. 2C, panels g-j; Table 1), a bioactive lipid previously shown to activate mechanosensitive channels through cytoskeletal-mediated mechanisms (Formigli et al., 2005). It was worth noting that the current traces corresponded to SAC activity, as no outward or inward currents from voltage-dependent ionic channels were ever observed in C2C12 myoblasts (Fig. 2C). Of interest, the mechanical stimulation caused a stronger increase in Im and Gm/Cm in S1P-stimulated scrambled cells (Table 1) compared with the relative controls (stimulated with S1P at resting length) (Fig. 2C, panels g,h; Table 1), whereas it did not affect the electrophysiological parameters (Table 1), in those treated with TRPC1-siRNA and S1P (Fig. 2C, panels i,j; Table 1).
Finally, in all the experimental conditions, the Im/Vm relationship evaluated by ramp voltage-clamp protocol was linear over the membrane potential examined and the reversal potentials (Vrev) ranged from -7 to 0 mV in SCR- and TRPC1-siRNA cells, and from 0 to 7 mV in S1P-stimulated cells (Fig. 2C, panels k,l). These Vrev values were compatible with the solutions used for evaluating the involvement of cationic channels, thus excluding the contribution of currents through voltage-dependent anionic SAC (Cl-selective) and Ca2+-selective ion channels (Formigli et al., 2005; Formigli et al., 2007b).
In view of the previous data underscoring the ability of S1P to promote Ca2+ release from the intracellular stores (Meacci et al., 2002b; Formigli et al., 2002) and to induce inositol-3-phosphate and diacylglycerol generation leading to store-operated Ca2+ entry (Blom et al., 2005; Mehta at al., 2005) in several cell types, we asked whether treatment with the bioactive lipid could also affect SOC-dependent transmembrane currents in C2C12 cells. As shown in Table 2, the treatment with 50 μM GdCl3 or 100 μM 2-APB, an inositol-3-phosphate receptor-dependent SOC inhibitor (Lievremont et al., 2005), reduced Gm/Cm by approximately 27% and 40% in unstimulated cells, respectively, and by approximately 60% and 65% in S1P-stimulated cells. Moreover, the ion current was completely abolished when the stimulated cells were silenced for TRPC1 expression.
All these data taken together lead us to suggest that TRPC1-mediated SAC and SOC currents in C2C12 myoblasts were modulated by the sphingoid molecule.
Functional localization of TRPC1 in lipid microdomains in skeletal myoblasts
As TRPC-based channels may be associated with highly structured cholesterol-enriched lipid-microdomains (Lockwich et al., 2000; Murata et al., 2007; Kannan et al., 2007), we next searched for TRPC1 localization in lipid microdomains and the possible functional repercussion on the channel activity in C2C12 myoblasts. To this purpose, we used centrifugation to equilibrium on discontinuous sucrose density gradients to isolate lipid rafts, which were then analyzed by western blotting for the presence of the typical marker protein caveolin 1 (Cav1) (Fig. 3A). Because lipid microdomains are also characterized by the enrichment in cholesterol and sphingolipids, in some experiments C2C12 cells were labelled with [3H]sphingosine and sucrose density gradient fractions analysed, after lipid extraction, for the presence of [3H]sphingolipids (Fig. 3B). As expected, raft-containing detergent-resistant membranes (DRM), corresponding to fractions 2 to 5, were enriched in lipid raft-protein marker (Fig. 3A) and in [3H]sphingolipids (Fig. 3B), whereas high-density membranes (HDM), corresponding to fractions 10-11 and 12 (detergent-insoluble pellet likely corresponding to cytoskeleton), contained calnexin, which reportedly is not found in DRM (Fig. 3A). As shown in the same figure, TRPC1 co-purified with DRM fractions and physically interacted with Cav1 (Fig. 3C). In particular, TRPC1 was detected not only in DRM but also in HDM fractions as two distinct molecular weight bands: one at approximately 80 kDa and the other at approximately 90 kDa. These findings, demonstrating the existence of post-translationally modified TRPC1 forms in caveolar lipid rafts and in detergent-insoluble fractions, are in agreement with other studies performed in different cell lines (Lockwich et al., 2000; Yuan et al., 2003). Co-localization of TRPC1 with lipid microdomains was also confirmed by confocal immunofluorescence (Fig. 3D, panels a-c). In particular, C2C12 cells were exposed to the fluorescent Alexa 488-cholera toxin subunit B (CT-B, 1 μg/ml), used as a lipid raft marker (Nagy et al., 2002), and immunostained for TRPC1 expression. In agreement with the above reported data, confocal immunofluorescence revealed that TRPC1 and raft-specific signals overlapped in native myoblasts (overlap coefficient 0.48±0.01), consistent with TRPC1 localization to myoblast lipid microdomains. This assumption was confirmed by the analyses of the fluorescence intensity profiles along lines traced in Fig. 3D, showing a significant degree of co-localization of TRPC1 with CT-B, especially along the plasma membrane.
We also showed that exposure of C2C12 cells to methyl-β-cyclodextrin (MβCD, 2 mM), which selectively depletes cholesterol from the cell membranes, resulted in alterations in both TRPC1 cellular localization and SAC sensitivity (Fig. 3D-F). Indeed, MβCD treatment reduced co-localization with CT-B (overlap coefficient 0.1±0.01) (Fig. 3D, panels d-f) and caused a redistribution of TRPC1 from DRM to HDM (fraction 11, Fig. 3E). Notably, the treatment with MβCD did not modify actin assembly in these cells (Fig. 3D, panels d,f). Patch clamp recordings showed that S1P-induced Gm/Cm increase was markedly inhibited (approximately by 40%) after MβCD treatment compared with untreated cells (Fig. 3F), suggesting that the correct localization of a pool of TRPC1 channels in cholesterol-enriched lipid microdomains was crucial for the channel function in C2C12 myoblasts.
In the light of recent observations that actin-binding proteins modulate membrane raft dynamics (Viola and Gupta, 2007), we next investigated whether cytoskeletal reorganization affected TRPC1 microdomain membrane localization. At first, we provided evidence for a physical association of TRPC1 with cytoskeletal proteins by co-immunoprecipitation experiments. In fact, as shown in Fig. 4A, endogenous TRPC1 co-immunoprecipitated with cortactin, an actin-binding protein expressed underneath the plasma membrane (Buday and Downward, 2007). Of note, this association was dependent on actin remodelling, as TRPC1-cortactin association was increased or decreased in response to S1P-induced SF formation and treatment with the actin-disrupting factor, dihydrocytochalasin B (DHCB, 1 μg/ml), respectively. Notably, S1P and DHCB treatment did not significantly affect TRPC1 expression (Fig. 4A). We then investigated the influence of actin cytoskeleton status on lipid raft localization of TRPC1 by confocal immunofluorescence. The treatment with DHCB significantly altered the association of TRPC1 to these membrane domains (overlap coefficient 0.33±0.02), provoking a dramatic dispersion of the immunostaining throughout the cells (Fig. 4B, panels a-c). Moreover, S1P-induced SF formation increased TRPC1 localization to membrane lipid rafts (overlap coefficient 0.7±0.025) (Fig. 5B, panels d-f). These latter data, in conjunction with the findings that cholesterol depletion impaired TRPC1 channel activation, without affecting actin filament assembly (Fig. 3D, panel e), supported the idea that cytoskeletal remodelling promoted by S1P and lipid raft integrity were both crucial for TRPC1 function in C2C12 myoblasts.
Role of TRPC1 expression on skeletal myoblast differentiation and its modulation by S1P
Since previous reports of our group and others (Formigli et al., 2007b; Wedhas et al., 2005) have demonstrated that the pharmacological inhibition of SACs blocks skeletal myogenesis and that S1P exerts its pro-myogenic action through SAC activation, we finally examined the relevance of TRPC1 in skeletal muscle differentiation and the ability of the bioactive lipid to modulate the channel expression. To this purpose, native, SCR- and TRPC1-siRNA-treated C2C12 cells were cultured in differentiation medium (DM) and assayed for the accomplishment of skeletal muscle differentiation in the presence or absence of S1P. The results of this analysis, shown in Fig. 5, demonstrated that, in both unstimulated and S1P-stimulated cells, transfection with TRPC1-siRNA significantly reduced the expression of the myogenic markers myogenin and α-sarcomeric actin (Fig. 5A,B), and the tendency of these cells to fuse into multinucleated myotubes, compared with SCR-siRNA-transfected cells (Fig. 5B). Moreover, the expression of TRPC1 was tightly regulated during myogenesis (Fig. 6). In fact, there was an increased expression of TRPC1 in C2C12 myoblasts after 24 hours, followed by a progressive decrease along with differentiation (72 hours) in agreement with the reported changes in SAC density and activity during skeletal myogenesis (Formigli et al., 2007b). Interestingly, TRPC1 expression was affected by both positive and negative regulators of muscle cell differentiation; indeed, the channel expression was markedly upregulated in response to S1P (Fig. 5A; Fig. 7B,C) and downregulated after the addition to the cell medium of 1 ng/ml transforming growth factor β (TGFβ) (Fig. 7), a well-known protein interfering with muscle development and regeneration (Olson et al., 1986).
Finally, to evaluate the contribution of TRPC1/SAC- and SOC-mediated currents in sustaining C2C12 cell differentiation, patch-clamp analysis was performed in differentiating myoblasts. It was found that TRPC1 silencing significantly reduced transmembrane currents evoked by mechanical stretching (Fig. 8A), whereas the treatment with the SOC-inhibitor, 2-APB, did not substantially modify Gm/Cm after 24 hours incubation in DM (Fig. 8B). All these data lead us to conclude that TRPC1 could play a preferential role as an SAC, rather than an SOC, channel in the regulation of the early phases of myoblast differentiation. This assumption was further confirmed by data obtained in C2C12 cells induced to differentiate by S1P (Fig. 8A,B). Collectively, all these findings provided the first experimental evidence that TRPC1 plays a crucial role in C2C12 cell differentiation and that mechanotransduction could represent a critical mechanism of the pro-myogenic action of S1P in skeletal muscle cells.
TRPC proteins constitute a family of highly conserved cation-permeable channels with multiple physiological roles. Although the numerous mechanisms of activation, including receptor-operated, ligand-induced, intracellular store depletion-mediated, and membrane-stretch activation are known, the exact physiological functions of TRPC-containing channels are currently under debate (Dietrich et al., 2007; Gottlieb et al., 2007; Maroto et al., 2005). In the present study, we have investigated the functional expression of TRPC1, the best candidate to be a stretch-activated channel of the canonical TRP family (Allen et al., 2005; Nilius et al., 2007). TRPC1 was detected at relatively high levels by RT-PCR, and represented the predominant channel isoform in C2C12 skeletal myoblasts, leading us to better correlate its expression with its physiological function. TRPC1 appeared to be implicated in SAC mechanisms in these cells, on the basis that cation currents promoted by mechanical stretches (using patch microelectrodes and AFM tether-pulling) were strongly attenuated by treatment with siRNA specific to TRPC1, thus adding additional evidence for a role of TRPC1 as an SAC (Maroto et al., 2005; Stiber et al., 2008). These data are in contrast with those recently published showing that the amplitude of mechanosensitive currents is not significantly altered by both TRPC1 overexpression (Gottlieb et al., 2007) and ablation (Dietrich et al., 2007). However, it is likely that the ectopic expression of tagged-TRP proteins or the compensation by other isoforms in TRPC1 knockout mice, may affect the proper protein function (Salgado et al., 2008). Although the data reported here characterize TRPC1 as a component of SACs, a role for this channel as an SOC cannot be ruled out, especially taking into consideration the well-documented activation of TRPC1 by Ca2+ store depletion (Fiorio Pla et al., 2005; Wu et al., 2004) and the emerging assumption that SACs and SOCs may belong to the same population or share common constituents, including TRPC1 (Ducret et al., 2006). In line with this, we have shown here that the chemical stimulation of C2C12 cells with S1P, a bioactive lipid known to mediate intracellular Ca2+ store release (Formigli et al., 2002; Meacci et al., 2002b; Spiegel and Milstein, 2003), was able to increase TRPC1-mediated transmembrane ion current, and that this increase is partially prevented by 2-APB, a specific inhibitor of SOCs (Lievremont et al., 2005).
We have also shown that TRPC1 was predominantly localized in lipid microdomains in C2C12 myoblasts, as previously reported in other cell types (Murata et al., 2007; Kannan et al., 2007). Current evidence suggests that lipid microdomains, dynamic plasma membrane structures enriched in cholesterol and sphingolipids, concentrate and segregate several membrane signalling proteins and serve as scaffolds for activation of Ca2+-dependent mechanisms (Ambudkar, 2006). In particular, we have found that TRPC1 protein physically interacted with caveolin 1, and co-localized with lipid-raft-specific fluorescent marker, providing the first evidence for the existence of a morpho-functional interaction between TRPC1 protein and lipid microdomains in skeletal myoblasts. As cholesterol depletion by MβCD caused dispersion of Cav1 without affecting actin cytoskeletal assembly, contrary to previous studies (Morachevskaya et al., 2007), it is likely that lipid raft integrity may account for the correct localization of TRPC1 and TRPC1/SAC sensitivity in skeletal myoblasts. This assumption fits well with the recent observations that lipid rafts determine clustering of known regulators of TRPC1, such as Stim1, in HSG and HEK293 cells (Pani et al., 2008). From our findings, it may be assumed that destruction of lipid microdomains may either change the intrinsic properties of the channel and/or affect its specific interaction with other proteins in the lipid microenvironment or with proteins linked to the actin cytoskeleton (Levitan and Gooch, 2007). This is mostly on the basis of the following observations: (1) TRPC1 physically interacted with the actin-binding protein cortactin, as also previously suggested for TRPC3 (Lockwich et al., 2001); (2) this physical association was enhanced by S1P and prevented by the F-actin disrupter agent, DHCB; (3) actin cytoskeletal remodelling by DHCB and S1P correlated with the regulation of TRPC1/raft co-localization. These data are consistent with the idea that actin cytoskeleton may anchor TRPC1 to lipid microdomains not only for structural purpose, but also for channel function regulation. In line with this assumption, recent observations have indicated that altered TRPC-protein (dystrophin, caveolin 3) association and the loss of Homer1 scaffolding of TRP channels contribute to the abnormal SAC activity observed in dystrophic myofibres (Gervasio et al., 2008; Stiber et al., 2008; Vandebrouck et al., 2007). Of interest, very recently the expression of other scaffold proteins required for store-operated Ca2+ entry, Stim1 and Orai1, has also been demonstrated in skeletal myotubes (Lyfenko and Dirksen, 2008).
It is thought that Ca2+ entry via TRPC channels (Ambudkar, 2006) regulates wide-ranging biological functions, including regulation of smooth muscle cell proliferation, endothelin-evoked arterial contraction, neuronal differentiation and cardiac hypertrophy (Rychkov and Barritt, 2007). On this basis, and in consideration of our previous findings that Ca2+ influx is a prerequisite for S1P-induced myoblast differentiation (Squecco et al., 2006), and the evidence reported here that TRPC1 localized especially after S1P stimulation into lipid microdomains, where key Ca2+ signalling proteins are concentrated (Murata et al., 2007), we have next explored the potential involvement of TRPC1 in skeletal myogenesis. Of interest, the silencing of TRPC1 expression dramatically hampered the ability of S1P to promote the expression of muscle differentiation markers and myoblast-myotube transition, in agreement with the results of our group and others showing that the pharmacological inhibition of SACs prevents skeletal myogenesis (Formigli et al., 2007b; Wedhas et al., 2005). Interestingly, we have also shown that TRPC1 expression was tightly regulated during myogenesis and modulated by known factors involved in skeletal myogenesis: the pro-myogenic agent, S1P (Meacci et al., 2008), and a potent inhibitor of skeletal muscle specific gene expression, TGFβ (Olson et al., 1986). Moreover, the weak ability of 2-APB to affect S1P-induced activation of trans-membrane currents in 24-hours differentiating cells, points to a predominant role of TRPC1 in sustaining Ca2+ influx through SACs during myoblast differentiation promoted by the sphingolipid. It is also worth noting that our findings do not exclude a contribution of Ca2+ entry via SOCs in the regulation of the early response to S1P, suggesting the possibility that Ca2+ influx through SOC and SAC/TRPC1 channels may activate different signalling pathways that culminate in a range of temporally diverse responses.
In conclusion, in the present study we have shown for the first time that TRPC1 represents an essential constituent of SACs in C2C12 myoblasts and that its expression is modulated in skeletal myogenesis. Although TRPC1 is a protein that may form homomeric and/or heteromeric channels (Clapham, 2003), the fact that its downregulation by specific siRNA leads to functional effects (reduced trans-membrane currents) indicates that this isoform, either alone or combination with others, has a crucial role in determining SAC function and the related Ca2+-mediated signals. It is probable that the channel localization in lipid rafts, and its interaction with caveolin 1 and the underlying cytoskeleton, may represent a structural platform necessary for SAC/TRPC1 channel activation. Of note, the upregulation of TRPC1 and its recruitment into membrane microdomains by S1P represent novel mechanisms by which the bioactive lipid exerts its pro-myogenic action on these cells, linking SAC-opening to gene expression. These findings are particularly important in view of the role played by sphingolipid signalling in satellite cell biology and in tissue regeneration, and offer new possible targets for therapeutic intervention in those pathological conditions in which alterations of cation concentration appear to be crucially involved.
Materials and Methods
Cell cultures and treatments
Murine C2C12 skeletal myoblasts obtained from American Type Culture Collection (ATCC, Manassas, VA), were routinely grown in 100 mm dishes in Dulbecco's modified Eagle's medium (DMEM) and cultured as previously reported (Formigli et al., 2005; Meacci et al., 1999). In some experiments the cells were treated for 30 minutes with the following reagents: S1P, 1 μM, Calbiochem, La Jolla, CA); methyl-β-cyclodextrin (MβCD, 2 mM, Sigma, Milan, Italy) to deplete cell membrane cholesterol; and dihydrocytochalasin B (DHCB 1 μg/ml, Sigma) to inhibit actin polymerization.
For myogenic differentiation studies, cells were induced to differentiate by incubation in differentiation medium (DM), containing 2% horse serum (HS, Sigma) in the presence or absence of S1P and transforming growth factor beta (TGFβ, 1 ng/ml, Sigma).
Reverse transcription and cDNA amplification
Total cellular RNA was extracted from control C2C12 cells, using TRIREAGENT (Sigma) according to the manufacturer's protocol and quantified spectrophotometrically at 260 nm. Then, 1 μg of total RNA was used for reverse transcription (RT) using Superscript II reverse transcriptase (Invitrogen, Carlsbad, CA). Different amounts of cDNA were used for PCR in the presence of the mouse-specific primers for TRPC isoforms, designed in the coding region as reported (Kunert-Keil et al., 2006). β-Actin was amplified using specific primers (forward: TCATGTTTGAGACCTTCAACACCC; reverse: GATGGAATTGAATGTAGTTTC) and used as an internal control to normalize relative levels of gene expression. PCR products were separated by electrophoresis onto 1.2% agarose gel, and exact size was evaluated by comparison with PCR Low Ladder (Sigma).
Cell lysates and subcellular fractionation
Lysate preparation and subcellular fractionation were performed as previously described (Meacci et al., 2000a; Meacci et al., 2000b) with minor modifications. C2C12 myoblasts or differentiating cells were lysed in homogenization buffer containing 0.2 M Tris (pH 7.4), 1 mM EDTA, 15 mM NaF, 1 mM dithiothreitol and protease inhibitors. Membrane/cytosol fraction preparation was done by centrifuging in an SW50.1 rotor for 1 hour at 200,000 g at 4°C. Detergent-insoluble membranes were obtained by resuspending the total membrane fraction in homogenization buffer containing 1% Triton-X100 and centrifuging as above. Protein concentration was measured by Bradford microassay (Bio-Rad, Hercules, CA).
Silencing of TRPC1 by siRNA
To inhibit the expression of TRPC1, short interference RNA duplexes (siRNA) (Santa Cruz Biotechnology, Santa Cruz, CA) corresponding to three distinct regions of the DNA sequence of mouse TRPC1 gene (NM_011643): GAGAAAUGCUGUUACC - AUA, CCAGAAUAUUCAACAACGA and GAACUUAAGUCGUCUGAAA) were used in combination. A non-specific scrambled (SCR) siRNA (Santa Cruz Biotechnology) was used as control. C2C12 myoblasts at 80% confluence were transfected using Lipofectamine 2000 reagent (1 mg/ml; Invitrogen) with the mixed combination of TRPC1-siRNA duplexes or with SCR-siRNA (20 nM) as reported (Meacci et al., 2008). After 5 hours, transfected cells were shifted in fresh medium for additional 48 hours, and then used. The specific knock-down of TRPC1 was evaluated by western blotting and confocal immunofluorescence. The efficiency of transfection was estimated to be approximately 70%.
Western blot analysis and immunoprecipitation
Immunoblotting was performed as previously reported (Meacci et al., 2000a; Meacci et al., 2008). Total cell lysates, bound fractions of precipitates, lysates obtained from the subcellular fractionation procedures, or samples from sucrose density gradient purification (see Lipid raft analysis) were resuspended in Laemmli's sample buffer and subjected to sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE). Proteins were transferred on PVDF membranes (Hybond-P, Amersham Pharmacia Biotech, Uppsala, Sweden). To immunodetect endogenous TRPC1, specific rabbit polyclonal anti-TRPC1 antibodies (Santa Cruz Biotechnology) were used. Specific monoclonal antibodies against endogenous caveolin 1 (Transduction Laboratories, Lexington, KY), myogenin (F5D, Sigma), α-sarcomeric actin (Dako, Carpenteria, CA), calnexin (Sigma), cortactin (Sigma) and β-actin (Santa Cruz Biotechnology) were used. Hybridization with primary antibodies was followed by incubation with peroxidase-conjugated goat anti-mouse or anti-rabbit IgG1 (Santa Cruz Biotechnology). Proteins were detected by enhanced chemiluminescence (ECL; Amersham Pharmacia Biotech). Immunoprecipitation was performed as previously reported (Meacci et al., 2000a; Squecco et al., 2006). Briefly, cell lysates and a pool of DRM fractions, obtained from the sucrose density gradient centrifugation of the lysate of approximately 6×106 cells, were resuspended in lysis buffer and, after pre-clearing with Protein A-Sepharose (Amersham Pharmacia Biotech), 1-4 μg of primary antibody were added. Antibody-protein complexes were collected with Protein A-Sepharose. The immunocomplex was washed and dissolved in 4× Laemmli, proteins separated by SDS-PAGE and immunoblotted.
Morphological analyses and immunofluorescence
C2C12 cells grown on glass coverslips were fixed in 0.5% buffered paraformaldehyde (PFA) and immunodetected as previously described (Formigli et al., 2005) using rabbit polyclonal anti-TRPC1 (Santa Cruz Biotechnology) and mouse monoclonal anti-myogenin (Sigma). The immunoreactions were revealed by incubation of the cells with goat anti-rabbit Alexa Fluor 488- or Alexa Fluor 568-conjugated IgG or with goat anti-mouse Alexa 488-conjugated IgG (Molecular Probes Inc., Eugene, OR). Negative controls were carried out by replacing the primary antibodies with non-immune rabbit serum; cross-reactivity of the secondary antibodies was tested in control experiments in which primary antibodies were omitted. Counterstaining was performed with TRITC- or Alexa Fluor 647-labelled phalloidin (Molecular Probes) to reveal F-actin. In some experiments, living cells were first incubated with TRITC-conjugated wheat germ agglutinin (TRITC-WGA, 1:25, Molecular Probes) for 10 minutes at room temperature to label plasma membrane and then fixed and immunostained for the expression TRPC1.
The coverslips containing the immunolabelled cells were then viewed under a confocal Leica TCS SP5 microscope (Leica Microsystems, Mannheim, Germany) equipped with a HeNe/Ar laser source for fluorescence measurements and with differential interference contrast (DIC) optics. The observations were performed using a Leica Plan Apo 63×/1.43NA oil immersion objective. Series of optical sections (1024 × 1024 pixels each; pixel size 204.3 nm) 0.4 μm in thickness were taken through the depth of the cells at intervals of 0.4 μm. Images were then projected onto a single `extended focus' image. When needed, a single optical fluorescent section and DIC images were merged to view the precise distribution of the immunostaining. Densitometric analysis of the intensity of TRPC1 fluorescence was performed on digitized images of C2C12 cells using ImageJ software (NIH). At least 50 different cells were analysed in each experiment (three preparations/experiment).
Myotube formation was also observed under an inverted phase contrast microscope Nikon Diaphot 300 (Nikon).
Measurements of stretch-induced Ca2+ influx by atomic force microscopy
C2C12 cells cultured on glass coverslips were loaded with the fluorescent Ca2+ dye Fluo3-AM (1 μM, Molecular Probes) for 20 minutes at 37°C. To visualize stretch-activated Ca2+ transients, the cells were mechanically stimulated using the cantilever of an atomic force microscope (PicoSPM, Molecular Imaging, Phoenix, AZ), as previously described (Charras and Horton, 2002). Observations were performed during the mechanical stimulations to ensure that the cells were not visibly damaged as well as after stimulation to visualize intracellular Ca2+ transients. The intracellular Ca2+ levels were expressed as relative fluorescence [ΔF/Fb: ratio of fluorescence difference, peak-basal (Fp-Fb), to basal (Fb) values], as previously reported (Viciencio et al., 2006).
The electrophysiological behaviour of C2C12 cells was examined by the whole cell patch-clamp technique in voltage-clamp conditions, as previously described (Formigli et al., 2005). During the experiments, the cells were superfused with a physiological bath solution containing 150 mM NaCl, 5 mM KCl, 2.5 mM CaCl2, 1 mM MgCl2, 10 mM D-glucose and 10 mM HEPES. The patch pipettes were filled with a solution containing 150 mM CsBr, 5 mM MgCl2, 10 mM EGTA and 10 mM HEPES, which was filtered through 0.22 μm pores. pH was titrated to 7.4 with NaOH and to 7.2 with TEA-OH for bath and pipette solution, respectively. Pulse protocol of stimulation consisted of 100 millisecond step pulses ranging from -80 to 0 mV, applied from a holding potential (HP) of -60 mV every 10 seconds in 10 mV steps. By using this electrophysiological protocol, the current-voltage relation was linear in all experimental conditions, confirming our previous reports showing that voltage-dependent ionic channels are not expressed in C2C12 myoblasts (Formigli et al., 2005; Formigli et al., 2007b). Moreover, to better evaluate the reversal potential we used ramp voltage-clamp pulses applied from -120 to +40 mV at a rate of 100 mV/second, HP -60 mV. Electrode capacitance was compensated before disrupting the patch. Access resistance (Ra) was not compensated for monitoring membrane area. The area beneath the capacitative transient and the time constant of the transient's decay (τ) were used to calculate the cell linear capacitance (Cm) and Ra from τ=RaCm. The measurement of membrane resistance (Rm) and capacitance (Cm) were corrected as reported previously (Formigli et al., 2005). The SAC and/or SOC current (Im) was determined by subtracting from the total current (Im*) the leak current (Im,leak). This latter was evaluated as the residual current recorded 3 minutes after adding in the bath solution at the end of any experimental session gadolinium chloride (GdCl3; 50 μM; Sigma), a SAC/SOC blocker (Ducret et al., 2006) (Fig. 2C, panels a-c). The SAC and/or SOC conductance (Gm) given in the results for any experimental conditions was evaluated from Im. In order to reveal the presence of SACs, the cells were stretched using two patch microelectrodes, as previously reported (Zhang et al., 2000). Briefly, one microelectrode was positioned by a hydraulic micromanipulator at the centre of the cell and used for electrophysiological measurements, while the other one was sealed 10-20 μm far, and moved by another hydraulic micromanipulator to stretch the cell. The stretch extent was 20% and its effect on transmembrane currents reported as the percentage of change in distance between the two microelectrodes (L) relative to the original value=(ΔL/Loriginal) 100. The SAC sensitivity evaluated by deviation of Gm was expressed as the change relative to the original length: G*m/Cm=[(Gm/Cm)stretched-(Gm/Cm)original]/(Gm/Cm)original=Δ(Gm/Cm)/(Gm/Cm)original.
As it has been reported that S1P activates inositol-3-phosphate- and SOC activity because of the decrease of the Ca2+ concentration in the endoplasmic reticulum lumen (Meacci et al., 2002b; Mehta et al., 2005), some experiments were performed with cells pre-treated for 30 minutes with GdCl3 (50 μM) or with the inositol-3-phosphate receptor blocker, 2-aminoethoxydiphenil borate (2-APB, 100 μM, Sigma, stock solution in methanol, 1000×) 3 minutes before the electrophysiological recording. For the latter experiment current records were made in the presence of 2-APB added in the bath solution 3 minutes before the electrophysiological recording.
The Cm value was considered as an index of the cell surface area assuming that membrane-specific capacitance is constant at 1 μF/cm2. To allow comparison of test current recorded from different cells, the membrane conductance was normalized to Cm, Gm/Cm (in nS/pF).
Lipid raft analysis
Biochemical isolation of lipid rafts
Preparation of detergent-resistant membrane microdomains (DRM) was done essentially as described in Meacci et al. (Meacci et al., 2000b), by lysing the cells in 25 mM MES (pH 6.5), 150 mM NaCl, 1% Triton X-100, 1 mM Na3VO4, 1 mM β-glycerolphosphate, 1 mM phenylmethylsulfonyl fluoride (PMSF), 10 μg/ml aprotinin and protease inhibitors (Roche, Indianapolis, IN), and 40 strokes of a Dounce homogenizer (Wheaton, Milleville, NJ). Lysates were mixed with an equal volume of 85% sucrose, overlaid with 3 ml of 30% sucrose and 1 ml of 5% sucrose. Fractions (0.5 ml each) were collected after centrifugation using an SW41 rotor for 16 hours at 200,000 g at 4°C. The DRM corresponded to the interphase between the 30% and 5% sucrose layers, while the fractions 10-12 corresponded to high density membrane (HDM); in particular fraction 12 represented detergent-insoluble pellet. Membrane/cytosolic and lipid raft fraction were diluted with 4× Laemmli's sample buffer, and separated by SDS-PAGE and analysed by immunoblotting. Sphingolipid content in lipid rafts was determined by labelling overnight the cells with 0.1 μCi of [3H]sphingosine essentially as previously reported (Meacci et al., 2000b). Detergent-resistant membrane microdomains were prepared from labelled cells as described above and lipid extracted by each gradient fractions and separated by TLC as previously described (De Palma et al., 2006; Meacci et al., 2008).
Lipid raft staining
Lipid rafts were labelled by the fluorescent-conjugate of cholera toxin subunit B (CT-B), using the Alexa Fluor 488 Lipid Raft Labeling Kit (Invitrogen). In particular, living native and treated C2C12 cells were incubated with 1 μg/ml of Alexa Fluor 488-conjugated CT-B for 10 minutes at 4°C. After washing with chilled PBS, the cells were fixed in buffered PFA for 10 minutes at room temperature and processed to reveal TRPC1 by confocal immunofluorescence. Quantitative assessment of co-localization between TRPC1 and CT-B fluorescence signals was performed by calculating the overlap coefficient (ranging from 0, minimum co-localization degree, to 1, maximum co-localization degree), using the Leica Application Suite Software. At least 50 different cells were analysed in each experiment (three preparations per experiment). Co-localization of TRPC1 (red) with lipid raft microdomains (CT-B, green) was also analysed by superimposing lines in predefined positions of the confocal image (along the plasma membrane and inside the cell) and calculating the number of co-localized red peaks divided by the total number of red peaks for each line, using Image J software (NIH). Five cells from three different experiments were analysed.
In immunoblot experiments, densitometric analysis of the bands was performed using Imaging and Analysis Software by Bio-Rad (Quantity-One). Band intensity of western blot analysis was reported as relative percentage (mean ± s.e.m.), obtained by calculating the specific protein/β-actin or caveolin-1 band intensity ratios and normalizing to control, set as 100. Statistical significance was determined by Student's t-test, with a value of P<0.05 considered significant. In the immunofluorescence experiments, data were evaluated as a mean ± s.e.m. of the indicated experiments. Statistical analysis of all these parameters was performed using one-way ANOVA or by Student's t-test; a P<0.05 was considered significant. Calculations were performed using GraPhPad Prism software program (GraPhPad, San Diego, CA). In electrophysiological experiments, statistical analysis of differences between the experimental groups was performed by one-way ANOVA and Newman-Keuls post-test (P<0.05 was considered significant). Calculations were made with Graph Pad Prism statistical program. pClamp9 (Axon Instruments, CA), SigmaPlot and SigmaStat (Jandel Scientific, San Rafael, CA) were used for mathematical and statistical analysis.
The authors are grateful to Daniele Nosi (Dept Anatomy, Histology, Forensic Medicine, University of Florence, Italy) for his valuable contribution in confocal and AFM image acquisition and processing. This work was supported by grants from Fondazione Cassa di Risparmio di Pistoia e Pescia to E.M., Fondazione Banche di Pistoia e Vignole to E.M., Ente Cassa di Risparmio di Firenze to L.F., S.Z.-O., University of Florence (ex. 60%) to E.M., L.F., F.F. and S.Z.-O.
- Accepted January 12, 2009.
- © The Company of Biologists Limited 2009