|Home | About | Journals | Submit | Contact Us | Français|
The triadin isoforms Trisk 95 and Trisk 51 are both components of the skeletal muscle calcium release complex. To investigate the specific role of Trisk 95 and Trisk 51 isoforms in muscle physiology, we overexpressed Trisk 95 or Trisk 51 using adenovirus-mediated gene transfer in skeletal muscle of newborn mice. Overexpression of either Trisk 95 or Trisk 51 alters the muscle fiber morphology, while leaving unchanged the expression of the ryanodine receptor, the dihydropyridine receptor, and calsequestrin. We also observe an aberrant expression of caveolin 3 in both Trisk 95- and Trisk 51-overexpressing skeletal muscles. Using a biochemical approach, we demonstrate that caveolin 3 is associated with the calcium release complex in skeletal muscle. Taking advantage of muscle and non-muscle cell culture models and triadin null mouse skeletal muscle, we further dissect the molecular organization of the caveolin 3-containing calcium release complex. Our data demonstrate that the association of caveolin 3 with the calcium release complex occurs via a direct interaction with the transmembrane domain of the ryanodine receptor. Taken together, these data suggest that caveolin 3-containing membrane domains and the calcium release complex are functionally linked and that Trisk 95 and Trisk 51 are instrumental to the regulation of this interaction, the integrity of which may be crucial for muscle physiology.
In skeletal muscle, the excitation−contraction (EC)1 coupling process takes place at the triads where T-tubules and the reticulum sarcoplasmic terminal cisternae are in close contact. EC coupling requires the expression at the triads of a multimeric calcium release complex (CRC) that includes the T-tubule voltage-dependent calcium channel dihydropyridine receptor (DHPR) and several sarcoplasmic proteins, namely, the calcium release channel ryanodine receptor (RyR) (1,2), the sarcoluminal Ca2+ binding calsequestrin (CSQ), and two of the four triadin isoforms identified to date in skeletal muscle, Trisk 95 (T95) and Trisk 51 (T51) (3,4). In vitro studies had identified a couple of regulatory RyR binding domains on triadin, and these regions are common to both Trisk 95 and Trisk 51 (5,6). Adenoviral-mediated overexpression of either Trisk 95 or Trisk 51 in primary cultures of skeletal muscle further demonstrated that only Trisk 95 plays a role in the regulation of the depolarization-induced calcium release mechanism (4), suggesting a specific role of Trisk 95 in EC coupling. Interestingly, triadin null mice exhibit a structural myopathy (7) with impaired depolarization-induced calcium release (7,8) indicating that the triadins are likely to be involved in the development of human myopathies for which a causative gene has not yet been identified. However, no modification has been yet identified in human triadins, both Trisk 95 and Trisk 51 being expressed in human skeletal muscle (9), and it is currently unknown whether overexpression of Trisk 95 and Trisk 51 would be detrimental for muscle function in vivo.
In this study, we use adenovirus-mediated gene transfer to overexpress Trisk 95 or Trisk 51 in mouse skeletal muscle as an alternative approach to investigating the function of these triadin isoforms. Herein, we show that overexpression of either Trisk 95 or Trisk 51 alters the muscle fiber morphology while leaving the expression of RyR, DHPR, and CSQ unchanged. We also observe that caveolin 3 (Cav-3), an essential structural component of caveolae involved in endocytosis and intracellular trafficking events (10), is aberrantly expressed in both Trisk 95- and Trisk 51-overexpressing skeletal muscles. We further demonstrate that Cav-3 is associated with the CRC via a direct interaction with the RyR, and we propose that the Trisk 95 and Trisk 51 level of expression is critical for the regulation of this interaction.
Wild-type mice (C57BL/6) were bred at the University of Iowa from stocks originally obtained from Jackson Laboratories (Bar Harbor, ME). Triadin null mice have been described previously (7) and were bred and maintained on a C57BL/6 background at Université Joseph Fourier. Animal care and procedures were approved and performed in accordance with the standards set forth by the Institutional Ethics Committee, the National Research Council Guide for the care and use of laboratory animals, the National Institutes of Health, and the Animal Care Use and Review Committee of the University of Iowa.
mAbs against CSQ (clone VIIID12, Affinity BioReagents) and Cav-3 (BD Transduction Laboratories) were used as described in the company datasheet. The sheep anti-DHPR α1 subunit was obtained from Upstate Biotechnology. Polyclonal antibodies against the RyR, the common N-terminal end of tradins, Trisk 95, and Trisk 51 were described previously (2,3,11). Guinea pig anti-Ca2+-ATPase was a gift from A. M. Lompré (12). Crude microsomes were prepared from 1-month-old mouse gastrocnemius muscle as previously described (13).
The viruses were engineered and produced by the Gene Vector Production Network, at Genethon III (Evry, France). Three type 5 adenoviruses were used in this study, a control virus (AdV-DsRed) with the cDNA of the red fluorescent protein (DsRed), AdV-Trisk 95, an adenovirus with the full-length sequence of rat skeletal muscle T95 (EMBL AJ243304, 687 amino acids), and AdV-Trisk 51, an adenovirus with the full-length sequence of rat skeletal muscle T51 (EMBL AJ243303, 461 amino acids). All the transgenes were under the control of a CMV promoter.
Injections of adenovirus into 3−4-day-old wild-type pups were performed as previously described (14), with some modification: the hamstring, quadriceps, calf, and tibialis anterior muscles of one leg were each injected percutaneously with 1010 pfu diluted in 10 μL of saline; the same muscles of the contralateral leg were each injected with an equal volume of saline. Pups were reintroduced to the mother and kept in quarantine for 5 days. All pups survived after injections. Muscles were harvested 30 days after injection, snap-frozen in liquid nitrogen-cooled isopentane, and stored at −80 °C until they were used.
Rat myogenic L6 cells (clone C5) or COS-7 cells were cultured in Dulbecco’s modified Eagle’s medium supplemented with 10% fetal bovine serum (Invitrogen) and 1% penicillin-streptomycin. L6 cells were infected at a multiplicity of infection (MOI) of 200 with the adenoviruses encoding T95 (AdV-Trisk 95) or encoding T51 (AdV-Trisk 51). The cells were collected 48 h after infection.
The transmembrane domain of human RyR1 (GenBank accession number J05200) corresponding to amino acids 4457−5038 was cloned in pEGFP-C1 (Clontech), to produce a 93 kDa fusion protein corresponding to the GFP fused with the channel part of RyR1. Human caveolin 3 (GenBank accession number AF036365) was cloned in pCDNA3.1 (Invitrogen). COS-7 cells were transfected with the two plasmids (2.5 μg of each plasmid for 8 × 105 cells) using ExGen 500 (Euromedex), and the cells were collected 24 h after transfection.
Five hundred micrograms of microsomes from rat skeletal muscle or 300 μg of cells was solubilized at 2 mg/mL in the presence of 1.6% CHAPS, 0.9 M NaCl, 0.1% phospholipids, 100 μM CaCl2, 50 μM EGTA, 20 mM Pipes (pH 7.1), and protease inhibitors (1 mM diisopropyl fluorophosphates and 100 μM phenylmethanesulfonyl fluoride), and immunoprecipitation was performed with antibodies against rat Trisk 95, rat Trisk 51, or nonimmune serum as described previously (2), using 10 mg of protein A immobilized on Sepharose 4B. All the immunoprecipitated proteins were then analyzed by Western blotting.
The presence of RyR, DHPR, Trisk 95, Trisk 51, calsequestrin, Ca2+-ATPase SERCA, or caveolin 3 in different samples was assayed by Western blotting, using a chemiluminescent reagent (Western lightning Chemiluminescence reagent plus, PerkinElmer Life Science). After electrophoretic separation on a 5−15% acrylamide gel, the proteins were electrotransferred to Immobilon P (Millipore) as previously described (3). The secondary antibodies were labeled with horseradish peroxidase (Jackson ImmunoResearch Laboratories). Quantitative analysis was performed using a Chemidoc XRS and Quantity One (Bio-Rad), and the amount of each protein was normalized to the amount in AdV-DsRed-infected muscles, after loading correction by Coomassie blue staining of myosin, as described previously (15).
Seven micrometer cryosections were prepared and analyzed by immunofluorescence or H&E staining as previously described (16). For immunofluorescent staining, sections were washed with phosphate-buffered saline (PBS), blocked with 3% bovine serum albumin (BSA) in PBS for 30 min at room temperature, and incubated with a primary antibody in a PBS/1% BSA mixture overnight at 4 °C. After incubation for 1 h at room temperature with an Alexa fluor 488-conjugated (Invitrogen) or Cy3-conjugated secondary antibody in a PBS/1% BSA mixture, the samples were mounted with PermaFluor (Beckman Coulter). Sections were observed under an MRC-600 laser scanning confocal microscope (Bio-Rad). Digitized images were captured under identical conditions. Images of H&E staining were photographed using a Leica DM RXA microscope equipped with an Olympus DP70 digital camera.
Trisk 95 and Trisk 51, the two major skeletal muscle triadin isoforms, were overexpressed in vivo using injection of adenovirus into newborn mice. As a control, the muscles were injected with an adenovirus encoding DsRed. To determine the effect of Trisk 95 or Trisk 51 overexpression on the expression of their calcium release complex’s partners, immunoblot analysis of crude muscle extracts was conducted 30 days after gene transfer (Figure (Figure1A,B).1A,B). Protein expression levels were normalized to the amounts measured in control muscles infected with AdV-DsRed (Figure (Figure1A,1A, lane 3). Both Trisk 95 and Trisk 51 were overexpressed by 60% (Figure (Figure1A,1A, lanes 1 and 2, and Figure Figure1B),1B), but the expression levels of the RyR, the α1 subunit of DHPR, CSQ, and SERCA were not changed (Figure (Figure1B).1B). This experiment was conducted on muscles from different animals, with similar results (no modification of the expression level of the proteins assayed, except for that of the overexpressed triadin). Nevertheless, the level of overexpression depends on the infection level, which is not identical among all the animals. Using Western blot on muscle homogenate, the infected fibers are mixed with noninfected fibers, which constitute the majority of the fibers. To discriminate between infected and noninfected fibers and to evaluate more precisely the modification induced on the other proteins, we chose to analyze the expression of the CRC components in Trisk 95- and Trisk 51-overexpressing muscles by immunofluorescence. While the stainings for RyR, DHPR, and CSQ were unchanged in Trisk 95-infected fibers, we could observe a reduction in the level of Trisk 51 expression in Trisk 95-infected fibers (Figure (Figure1C),1C), an effect that was not observed in a Western blot. Likewise, only the level of Trisk 95 expression was reduced in Trisk 51-infected fibers (Figure (Figure11D).
Histological analysis revealed no sign of muscle damage in control skeletal muscles injected with either saline (data not shown) or AdV-DsRed (Figure (Figure2A,2A, panel a). However, Trisk 95- and Trisk 51-infected skeletal muscles were characterized by the presence of fibrotic tissue, atrophic fibers, and regenerating fibers (Figure (Figure2A,2A, panels b and c). The extent of muscle damage was greater in T51-infected muscles than in T95-infected muscle, as assessed by the increased level of fibrosis and the presence of necrotic fibers. To specifically characterize Trisk 95- and Trisk 51-infected fibers, muscle sections were analyzed after immunostaining with caveolin 3 (Cav-3) as a plasma membrane marker (Figure (Figure2B).2B). In contrast to muscles infected with AdV-DsRed (Figure (Figure2B,2B, panels a−c and g−i), membrane staining revealed that the size of fibers infected by AdV-Trisk 95 and AdV-Trisk 51 was abnormal. Trisk 95-infected fibers were smaller than noninfected fibers (Figure (Figure2B,2B, panels d−f), and Trisk 51-infected fibers had a round shape (Figure (Figure2B,2B, panels j−l) and lost the contact between them. In contrast with the control AdV-DsRed-infected fibers (Figure (Figure2B,2B, insets), the expression pattern of Cav-3 was altered in fibers overexpressing Trisk 95 and Trisk 51. In these latter fibers, Cav-3 was shown to accumulate intracellularly and to colocalize with the triadins. As Cav-3 is expressed during the differentiation of the satellite cells in myotubes (17), it could be hypothesized that the aberrant staining of Cav-3 specifically occurs in regenerating fibers. However, staining of nuclei indicated that none of the cells displaying an intracellular staining of Cav-3 were regenerating cells (data not shown). Immunoblot analysis showed a similar amount of Cav-3 in crude extracts of infected muscle as compared to control samples from DsRed-infected muscles (Figure (Figure2C,D),2C,D), suggesting most probably a relocalization of Cav-3 upon triadin overexpression.
In vivo modification of triadin expression results in modification of Cav-3, indicating a functional link between the two proteins. We studied their possible association by immunoprecipitation in rat skeletal muscle (Figure (Figure3A).3A). Using isoform specific antibodies, both Trisk 95 and Trisk 51 co-immunoprecipitated Cav-3 (Figure (Figure3A,3A, lanes 2 and 3). Trisk 95 and Trisk 51 also co-immunoprecipitated each other and the RyR, which indicates that Cav-3 is associated with the CRC.
An association between the DHPR and Cav-3 was proposed in skeletal muscle (18) and shown in cardiac muscle (19). Thus, it is possible that the molecular complex between the triadins and Cav-3 requires a direct interaction of Cav-3 with the DHPR. To test the direct interaction between the DHPR and Cav-3, we analyzed the co-immunoprecipitation in rat skeletal muscle of Cav-3 with DHPR or the triadins (Figure (Figure3A).3A). The DHPR efficiently co-immunoprecipitated Cav-3 and the RyR (Figure (Figure3A,3A, lane 5), but not the triadins. Conversely, both Trisk 95 and Trisk 51 co-immunoprecipitated Cav-3 but not the DHPR (Figure (Figure3A,3A, lanes 2 and 3), suggesting that the association of triadin and Cav-3 is not DHPR-dependent.
To test whether Cav-3 and the triadins are able to associate directly, we studied their association in the L6 rat muscle cell line that does not express the known CRC proteins but expresses Cav-3. L6 cells were infected with AdV-Trisk 95 or AdV-Trisk 51 to induce the expression of each triadin isoform (Figure (Figure3B,3B, lanes 2−4). The absence of the RyR and the presence of Cav-3 were confirmed in these cells (Figure (Figure3B,3B, lanes 2−4). Direct association of Cav-3 with either triadin was studied by co-immunoprecipitation with the anti-T95 and anti-T51 antibodies. Although Trisk 95 and Trisk 51 were efficiently immunoprecipitated, Cav-3 was co-immunoprecipitated with none (Figure (Figure3B,3B, lanes 5−7), indicating that Cav-3 is not engaged in a direct interaction with any of the triadins. As both triadins are linked in skeletal muscle, as demonstrated by their co-immunoprecipitation (Figure (Figure3A),3A), it is possible that the simultaneous presence of Trisk 95 and Trisk 51 is required for association with Cav-3. To test this hypothesis, L6 cells were co-infected with Adv-Trisk 95 and AdV-Trisk 51. However, Cav-3 was not co-immunoprecipitated with either anti-triadin isoform antibody (data not shown), suggesting that the presence of the two isoforms is not sufficient for the association with Cav-3
An association between RyR2 and Cav-3 was shown in cardiac muscle (20). To test whether Cav-3 directly interacts with RyR1 in skeletal muscle, we took advantage of the triadin null mouse model (7) to analyze the co-immunoprecipitation of Cav-3 with RyR1 in the absence of triadin. As shown in Figure Figure4A,4A, Cav-3 was efficiently co-immunoprecipitated with the RyR in both WT (Figure (Figure4A,4A, lane 1) and triadin null skeletal muscle microsomes (Figure (Figure4A,4A, lane 3), demonstrating that triadin was not necessary for this association and that the RyR−Cav-3 interaction could be direct. To confirm the direct interaction of the RyR and Cav-3, immunoprecipitation experiments were performed in a non-muscle cell line. COS-7 cells were cotransfected either with Cav-3 and GFP-RyR transmembrane domain (last 582 amino acids fused in frame with GFP on its N-terminal part) or with Cav-3 and GFP alone. The anti-GFP antibody efficiently co-immunoprecipitated the RyR and Cav-3 (Figure (Figure4B,4B, lane 6) but failed to co-immunoprecipitate Cav-3 in the presence of GFP alone (Figure (Figure4B,4B, lane 5), confirming the direct interaction between Cav-3 and the RyR1 transmembrane domain.
In vivo overexpression of the triadin isoform Trisk 95 or Trisk 51 was induced by adenoviral-mediated gene transfer in the hindlimb muscles of newborn mice. Overexpression of one triadin isoform induced a reduction in the expression level of the second one. The different triadin isoforms are issued of the alternative splicing of the same gene (9), and even if the mechanisms driving the specific expression of each isoform are unknown, it is clear that the two triadins Trisk 95 and Trisk 51 not only are always associated together, as seen in the co-immunoprecipitation experiments, but also mutually regulate their expression, to keep the total amount of triadin in the skeletal muscle almost constant. The expression of the main components of the calcium release complex, i.e., RyR, DHPR, the Ca2+-ATPase SERCA, and CSQ, was not affected. However, an interesting finding of our study was the alteration of Cav-3 expression in triadin-infected fibers. Using a biochemical approach, we demonstrated the existence of a molecular complex, including the triadins and Cav-3. Using differential immunoprecipitations in WT and triadin null skeletal muscle, as well as in muscle and non-muscle cells, we further demonstrated that Cav-3 was associated with the CRC via a direct interaction with the transmembrane domain of RyR1. Our data also suggest that two different populations of CRC exist in the skeletal muscle triads, one involving a DHPR−RyR−Cav-3 complex and one involving a Trisk 95−Trisk 51−RyR−Cav-3 complex, as shown in Figure Figure3.3. This could be related to the early electronic microscopy study demonstrating that only half of the RyR was coupled to the DHPR (1), and one could imagine that the DHPR and triadin are mutually excluding each other in the CRC, the RyR uncoupled to DHPR interacting with triadin.
Triadin (Trisk 95 or Trisk 51) overexpression results in an increased level of intracellular Cav-3 staining, which is indicative of an in vivo functional link between the two proteins. As we cannot demonstrate Cav-3 overexpression equivalent to triadin overexpression, this increased level of intracellular Cav-3 labeling is most probably due to Cav-3 relocalization. Our finding that Cav-3 is functionally linked to the proteins of the CRC raises the question of the physiological relevance of the association between the caveolae and the SR. Cav-3 was proposed to play a role in calcium homeostasis (21−23), and previous studies showed that the IP3R and SERCA, both involved in calcium homeostasis (24), were detected within the caveolae of several different cell types, including muscle cells (21,22). More recently, Li et al. (24) showed the presence of microdomains enriched in cholesterol, sphingolipids, and Cav-3 in the SR membrane. Assuming that the caveolae−SR contacts are partly responsible for the control of the cytoplasmic free Ca2+ concentrations in very precisely defined spaces (25), the CRC involving triadins could be located at those sites. Previous studies showed that caveolae are sites of extracellular Ca2+ entry (26) and suggested that caveolae are involved in refilling depleted intracellular calcium stores, a mechanism called store-operated calcium entry (SOCE). Caveolae could then act as intermediates between the SR and the plasma membrane. We have previously shown that Trisk 95 can modulate the mechanism coupling the depletion of intracellular Ca2+ stores to extracellular Ca2+ entry. The overexpression of Trisk 95 decreases the level of this Ca2+ entry in myotubes (27). It would therefore be possible that the overexpression of Trisk 95 disturbs the SOCE by its effect on Cav-3. The relocalization of Cav-3 from the plasma membrane to intracellular compartments following the overexpression of Trisk 95 would directly disturb the Ca2+ entry by caveolae and would therefore reduce the level of store-operated Ca2+ entry.
Both Cav-3 null mice and transgenic mice overexpressing Cav-3 display a dystrophic phenotype (17). A likely explanation for the Cav-3 null mice phenotype is the alteration of the intracellular trafficking and targeting of membrane proteins and protein complexes, such as dysferlin (2,28) and the dystrophin glycoprotein complex (17). Noteworthy is the fact that Cav-3 null skeletal muscle also exhibits abnormal DHPR and RyR staining, suggesting that Cav-3 is involved in triad formation during muscle development (29,30). Trisk 95- and Trisk 51-overexpressing muscles exhibited obvious signs of pathology that included central nucleated fibers and heterogeneous fiber sizes that resemble those of dystrophic muscles. Given that Trisk 95 and Trisk 51 seem to mutually regulate their expression level and given that triadin null mice exhibit muscle weakness (7), it is tempting to speculate that a precise expression level of the triadins is crucial for normal muscle function (31); an excess of triadin could be as deleterious as a lack of triadin on the muscle physiology. However, while Trisk 95 overexpression was shown to block EC coupling in skeletal muscle (4), Trisk 51 overexpression has no effect on EC coupling (4). Therefore, it is unlikely that the muscle pathology observed in triadin-overexpressing muscle results entirely from EC coupling dysfunction. In this context, one could speculate that Trisk 95 or Trisk 51 overexpression perturbs the traffic of proteins forming the Ca2+ release machinery by Cav-3 trapping, thus resulting in the abnormal muscle phenotype observed here. A mechanism of Cav-3 trapping has been proposed for the mutant form of Cav-3, P104L, responsible for some cases of myopathy (LGMD1C) (32,33); therefore, this mechanism has already been shown to be responsible for a pathology. Although no disease-causing mutation has been identified yet in the triadin gene, it is possible that modification of the triadin expression level underlies a new pathophysiological mechanism for human myopathies, possibly involving Cav-3 alteration.
We thank Genethon (Evry, France) for the production of the adenoviruses used in this study.
†This work was supported by grants from Association Française contre les Myopathies (AFM), from Groupement d'Intêret Scientifique (GIS)-Maladies Rares, from Agence Nationale de la Recherche (ANR-Maladies Rares), from Société Française de Myologie (SFM), and from Fondation Ducoin.