Vestigial-like 2 contributes to normal muscle fiber type distribution in mice

Skeletal muscle is composed of heterogeneous populations of myofibers that are classified as slow- and fast-twitch fibers. The muscle fiber-type is regulated in a coordinated fashion by multiple genes, including transcriptional factors and microRNAs (miRNAs). However, players involved in this regulation are not fully elucidated. One of the members of the Vestigial-like factors, Vgll2, is thought to play a pivotal role in TEA domain (TEAD) transcription factor-mediated muscle-specific gene expression because of its restricted expression in skeletal muscles of adult mice. Here, we generated Vgll2 null mice and investigated Vgll2 function in adult skeletal muscles. These mice presented an increased number of fast-twitch type IIb fibers and exhibited a down-regulation of slow type I myosin heavy chain (MyHC) gene, Myh7, which resulted in exercise intolerance. In accordance with the decrease in Myh7, down-regulation of miR-208b, encoded within Myh7 gene and up-regulation of targets of miR-208b, Sox6, Sp3, and Purβ, were observed in Vgll2 deficient mice. Moreover, we detected the physical interaction between Vgll2 and TEAD1/4 in neonatal skeletal muscles. These results suggest that Vgll2 may be both directly and indirectly involved in the programing of slow muscle fibers through the formation of the Vgll2-TEAD complex.

In Drosophila, the nuclear protein, Vestigial, is the essential cofactor of Scalloped, a homolog of TEA domain/ transcription enhancer factor (TEAD/TEF) family of transcription factors and plays a pivotal role in the development and patterning of the wing 22,23 . Vestigial is also a necessary cofactor of Scalloped for muscle differentiation 24 . Mammalian vestigial-like 1-4 (Vgll1-4) proteins possess the TEF-1 interaction domain called the tondu (TDU) motif and contribute to the tissue-specific functions of the TEAD factors [25][26][27] . Of these, Vgll2 (also known as VITO-1) is expressed in the somitic myotome and pharyngeal pouch during mouse development, while it is selectively expressed in skeletal muscle in the adult 25,26 . Previous studies showed that Vgll2 overexpression augments MyoD-mediated myogenic conversion of 10T1/2 cells 26,28 . Conversely, knockdown studies using antisense morpholino showed that inhibition of Vgll2 expression attenuates MyHC expression in C2C12 myoblasts and avian limb muscles 29 . In addition, skeletal muscle specific overexpression of TEAD1 induces fast-to-slow fiber-type transition in adult mice 30 . Thus, previous studies have implicated a role for Vgll2 in in vitro muscle differentiation, but its in vivo function is poorly understood.
In this study, we found that Vgll2-deficient mice exhibited a faster muscle contractile phenotype under basal conditions and significant expression changes of Myh7/miR-208b and its downstream targets of transcriptional repressor proteins for slow-twitch fiber in neonatal skeletal muscles. We further provide evidences that Vgll2 forms a protein complex with TEAD1/4. Our study reveals that Vgll2 has potent activity in normal skeletal muscle fiber distribution in both direct and indirect contribution in vivo.

Results
Preferential Expression of Vgll2 in Slow Muscle. Previous studies suggested a role for Vgll2 in the MyHC expression during myogenic differentiation of C2C12 myoblasts 29 , but the potential function of Vgll2 in adult muscle is unknown. To elucidate the function of Vgll2 in vivo, we first examined whether Vgll2 mRNA displays preferential expression patterns between individual skeletal muscles that differ in fiber type composition in 12-week-old mice. The expression level of Vgll2 mRNA was significantly higher in the Type I and IIa fiber-enriched slow soleus muscle than in the Type IIb fiber-enriched fast gastrocnemius and extensor digitorum longus (EDL) muscles (Fig. 1a). Hence, we speculated that Vgll2 may play a role in the control of the specification of adult muscle fiber-type.

Generation of Vgll2-Deficient Mice.
To determine the influence of Vgll2 deficiency in vivo, we generated Vgll2 knockout mice with targeted replacement of the Vgll2 loci by the lacZ gene in D3 ES cells (Fig. 1b). To construct a targeting vector, we obtained DNA fragments encoding Vgll2 from a mouse EB3 cell genomic DNA library 31,32 . Homologously recombined ES colonies and heterozygous mice were verified by Southern blot analysis (Fig. 1c). After mating heterozygous mice, all genotypes, Vgll2 +/+ , Vgll2 +/− , and Vgll2 −/− were identified (52, 106, and 45, x 2 = 0.68, chi-square test), which segregated according to the laws of Mendelian inheritance, indicating no marked embryonic lethality. Quantitative real-time PCR (qPCR) analysis revealed that Vgll2 transcript was absent in skeletal muscle from Vgll2 −/− mice (Fig. 1a). Moreover, Vgll2 protein was detected only in the nuclear extracts prepared from the Vgll2 +/+ soleus muscle by Western blotting with anti-Vgll2 antibody (see "Materials and Methods"), but not in Vgll2 −/− (Fig. 1d). Compensatory responses in other Vgll family genes such as Vgll3 and Vgll4, which are also expressed in skeletal muscle, were not observed in Vgll2 −/− mice ( Supplementary  Fig. S1).
Vgll2 Deficiency Induces a Slow-to-Fast Transformation of Muscle Fiber-Type. Vgll2 −/− mice displayed no overt abnormalities. The preferential accumulation of Vgll2 mRNA in the slow soleus muscle led us to investigate the fiber type composition in mice lacking Vgll2. To characterize the fiber type in Vgll2 −/− mice, we measured the relative expression levels of several MyHC isoforms by qPCR in neonatal and adult skeletal muscles. In Vgll2 −/− gastrocnemius-plantaris-soleus (GPS) muscle at postnatal day 7 (P7), expression levels of Myh7, encoding slow type I myosin, were significantly reduced by 24% (Fig. 2), whereas the expression of the fast MyHC isoform gene, Myh4, was significantly increased (Fig. 2). These alterations in MyHC expression profiles were also observed in 12-week-old Vgll2 −/− mice. In Vgll2 −/− slow twitch soleus muscle, expression levels of Myh7 were significantly decreased by 47%, whereas fast myosin genes, Myh1 and Myh4 were markedly increased 2.9-and 10.5-fold (Fig. 3a). We did not detect significant changes in the expression levels of all MyHC isoforms in Vgll2 +/− soleus muscle (Supplementary Fig. S2). In addition, in the fast gastrocnemius and EDL muscle groups, expressing low levels of Vgll2 mRNA, Myh7, Myh2, and Myh1 expression levels were significantly reduced, whereas Myh4 expression was sustained (Fig. 3a). Next, we determined the relative content of MyHC isoform type I, IIa (and/ or IIx), and IIb by gel electrophoresis and immunostaining of each skeletal muscle (Fig. 3b,c and Supplementary  Fig. S3). In Vgll2 −/− soleus muscle, a muscle normally rich in type I and IIa fibers, both type I and type IIa fibers were reduced, and an obvious increase in type IIb fibers was detected (Fig. 3b,c and Supplementary Fig. S3). In Vgll2 +/+ soleus muscle, type IIb fibers were very rarely detected (Fig. 3b,c and Supplementary Fig. S3). Moreover, Vgll2 −/− EDL myosin electrophoretogram showed that fast type IIa and IIx myosin were predominantly replaced by the fastest type IIb myosin (Fig. 3b), consistent with the changes in Myh2 and Myh1 expression levels (Fig. 3a). Thus, these observations suggest that a slow-to-fast contractile phenotype transition may occur in Vgll2-deficient mice.
Reduced Exercise Endurance in Vgll2 KO Mice. Changes in contractile properties of skeletal muscle have a considerable impact on physical activity. We therefore assessed the endurance capacity and grip strength of mice under involuntary conditions. Muscle endurance in Vgll2 +/+ and Vgll2 −/− mice were evaluated by treadmill running to exhaustion. After acclimatization, mice were allowed to run on the treadmill set to a 10% incline at increasing speed until the mice were unable to remain on the treadmill despite prodding. Vgll2 −/− mice were able to run for a significantly shorter time than Vgll2 +/+ mice did. Vgll2 +/+ mice ran for an average of 21.6 min, while Vgll2 −/− mice ran for 14.2 min (Fig. 4a). Correspondingly, the running distance was significantly lower for Vgll2 −/− mice compared to that ran by Vgll2 +/+ mice. The distance covered by Vgll2 −/− mice was 200.3 meters on average, whereas Vgll2 +/+ mice ran 374.7 meters on average (Fig. 4b). Grip strength was not reduced in Vgll2 −/− mice (Fig. 4c). These results indicate that, in skeletal muscles, Vgll2 contributes greatly to exercise endurance. Direct Binding of Vgll2 to TEF/TEAD in the Skeletal Muscle. Vgll2 possesses an evolutionarily conserved TDU motif, which binds to TEAD transcription factors. Previous studies suggested a role for TEF/TEAD family member, TEAD1 and TEAD4 in the regulation of slow muscle gene expression 30,[33][34][35][36] . Vgll2 deficient mice showed fiber type composition changes in skeletal muscle and reduced endurance. Our findings suggest that Vgll2 may be required for TEAD-mediated muscle fiber-type regulation, as one of the cofactors. To find Figure 1. Expression patterns of Vgll2 mRNA expression in several muscles, and generation of Vgll2-deficient mice. (a) Vgll2 mRNA levels were measured by qPCR in the soleus (SOL), gastrocnemius (GAS), and extensor digitorum longus (EDL) muscles from 12-week-old Vgll2 +/+ and Vgll2 −/− mice (n = 6). *P < 0.05 vs. Vgll2 +/+ soleus muscles. Data are presented as mean ± SEM. For comparison, the expression level in Vgll2 +/+ soleus was arbitrarily set at 1. (b) Schematic diagram showing the strategy used to generate the Vgll2 null allele. The top row depicts the wild-type Vgll2 allele, which consists of three exons (solid boxes). The initiation Met codon for Vgll2 exists in exon 1. The second row depicts the targeting construct. The third row depicts the mutated Vgll2 allele. The lacZ expression cassette with a SV40 polyadenylation signal, which was followed by the loxP-flanked puromycin resistant gene expression cassette (Puro) in the reverse orientation, was fused to the initiation codon of Vgll2. The diphtheria toxin A expression cassette (DT-A) was included at the 5′-end of the targeting vector. The external probe used for Southern blot analysis lies outside the 5′-homologous arm (gray box). RI, EcoRI; RV, EcoRV. (c) Southern blot analysis for genotyping wild-type (Vgll2 +/+ ), heterozygous Vgll2-deficient (Vgll2 +/− ), and homozygous Vgll2-deficient (Vgll2 −/− ) animals. Digestion with EcoRI of Vgll2 +/+ and Vgll2 −/− alleles gave rise to fragments of 2.2 and 4.6 kb, respectively. An uncropped original image is shown in Supplementary information. (d) Vgll2 protein levels were assessed by Western blotting using nuclear protein extracts obtained from the soleus muscle of 12-week-old Vgll2 +/+ and Vgll2 −/− mice. Histone H2B served as a loading control. Analyses were performed on four mice per genotype. Data from two mice per genotype are shown. Uncropped original images are shown in Supplementary information. the physical interaction between Vgll2 and TEAD1 in skeletal muscles, we conducted co-immunoprecipitation experiments using extracts from neonatal (P7) GPS muscles. As expected, endogenous TEAD1 and TEAD4 proteins were efficiently co-immunoprecipitated by the anti-Vgll2 antibody only in Vgll2 +/+ extracts, but not in Vgll2 KO extracts (Fig. 5a). In addition, reciprocal co-immunoprecipitation with an anti-TEAD1 antibody precipitated endogenous Vgll2 only in Vgll2 +/+ extracts (Fig. 5b). Similarly, co-immunoprecipitation with an anti-TEAD4 antibody precipitated endogenous Vgll2 only in Vgll2 +/+ extracts (Fig. 5c). Together, our findings raise the possibility that Vgll2 plays a cooperative role in TEAD1/4-associated transcription in skeletal muscle through protein binding.
To examine whether Vgll2 deficiency alters the expression levels of other TEAD cofactors such as Yap and Taz 27, 37, 38 , we quantified these genes by qPCR. However, we did not detect significant changes in the expression levels of these genes ( Supplementary Fig. S4).  and Myh4 (IIb) were measured by qPCR in the soleus, gastrocnemius, and EDL muscles from 12-week-old Vgll2 +/+ and Vgll2 −/− mice (n = 6). For comparison, the expression level of these genes in Vgll2 +/+ mice was arbitrarily set at 1. Data are presented as mean ± SEM. *P < 0.05 vs. Vgll2 +/+ in each muscle. (b) High resolution gel electrophoresis for MyHC isoform separation of protein extracts isolated from the soleus and EDL muscles of 12-week-old Vgll2 +/+ and Vgll2 −/− mice. Equal amounts (50 ng) of total protein were separated on an 8% acrylamide gel containing glycerol. The gel was then stained with the silver stain method. Uncropped original images are shown in Supplementary information. (c) Representative images of immunostained soleus muscles isolated from 12-week-old Vgll2 +/+ and Vgll2 −/− mice (n = 3). Scale bar: 200 μm.

Downregulation of miR-208b Expression in Vgll2 Deficient Skeletal Muscle and Increase in its
Target Proteins. We next sought to investigate the molecular mechanisms underlying slow-to-fast fiber-type transition in Vgll2 −/− mice. Previous studies showed that two miRNAs, miR-208b and miR-499, serve as fiber-type modulators by inhibiting the activity of transcriptional repressors of genes encoding slow-twitch contractile proteins such as Sox6, Sp3, and Purβ 21, 39-42 . Although these miRNAs share a comparable seed sequence and target the same mRNAs, they are differentially regulated: miR-208b is encoded within the Myh7 gene, and miR-499 is encoded within Myh7b gene 43 . We measured the relative expression levels of these miRNAs in neonatal skeletal muscles by qPCR. At P7, consistent with gene expression results of their host MyHC mRNAs (Fig. 2), the expression level of miR-208b was significantly reduced in Vgll2 −/− GPS muscles by 39%, whereas miR-499 expression was slightly increased (Fig. 6).
In order to further validate the reduction in miR-208b in Vgll2 −/− skeletal muscles, we compared protein levels of its downstream targets, Sox6, Sp3, and Purβ by Western blotting analysis. In Vgll2 −/− GPS muscles at P7, Sox6 and Sp3 protein levels were markedly increased, while Purβ protein level was moderately increased (Fig. 7a). Collectively, our findings suggest that Vgll2 directly and indirectly participates in the specification of mature skeletal muscle fiber characteristics.

Discussion
In the present study, we provide the first in vivo evidence that the transcription cofactor, Vgll2, plays a role in muscle fiber-type distribution and physical performance. The Absence of Vgll2 results in shifts in fiber type of skeletal muscle toward a faster phenotype and decreases in slow isoform MyHC gene and its intronic miRNA, Figure 4. Analysis of muscle functions in Vgll2 null mice. (a,b) Physical endurance was measured by using a treadmill running test. After acclimation, mice ran on a 10% slope with a protocol using increasing speed until mice were exhausted. Exhaustion was defined as the inability of the mice to remain on the treadmill despite physical prodding. Running time (a) and speed were measured, and the distance (b) was calculated. (c) Skeletal muscle strength was assessed by using the grip strength test. Grip strength was measured in each mouse 20 times and the highest value of each experiment was used. Male mice (11-13-week-old) were used in all experiments (n = 8). Data are presented as mean ± SEM. *P < 0.05 vs. Vgll2 +/+ mice. Figure 5. Physical interaction analysis between endogenous Vgll2 and TEAD1 in skeletal muscles at the neonatal stage. (a) Immunoprecipitation was conducted with the anti-Vgll2 antibody followed by Western blot analysis to detect TEAD1, TEAD4 and Vgll2 in GPS muscle extracts from Vgll2 +/+ and Vgll2 −/− mice at P7. (b) Immunoprecipitation was conducted with an anti-TEAD1 antibody followed by Western blot analysis to detect Vgll2 and TEAD1 in P7 GPS muscle extracts. (c) Immunoprecipitation was conducted with an anti-TEAD4 antibody followed by Western blot analysis to detect Vgll2 and TEAD4 in P7 GPS muscle extracts. Control samples were GPS muscle lysates incubated without anti-Vgll2 or -TEAD1 antibody (no-Ab). Uncropped original images are shown in Supplementary information. Myh7/miR-208b, and deregulation of miR-208b targets. As physical interaction between Vgll2 and TEAD1/4 were detected in neonatal muscles, we propose a possible mechanism involved in the programing of slow muscle fibers through the formations of Vgll2-TEAD1/4 complex (Fig. 7b).
Although several in vitro studies provided evidence that Vgll2 promotes skeletal muscle differentiation in cultured muscle cells 26,28 , its in vivo role in this capacity has not been investigated in skeletal muscle tissues. Herein, we show that Vgll2 is involved in the establishment and maintenance of adult skeletal muscle phenotypes. At P7, before the establishment of adult muscle fiber phenotype, gene expression analysis by qPCR revealed that Vgll2 deficient skeletal muscle presented a faster MyHC profile. Furthermore, this faster MyHC profile was maintained in adult mice. Both transcript and protein levels of MyHCIIa in adult Vgll2 −/− mice were dramatically decreased in slow and fast twitch muscles when compared to that in WT mice. Similarly, reduced expression of MyHCI was observed at both the transcript and protein levels. The changes in fast isoforms of MyHC expression were prominent in the slow soleus muscle. Transcript and protein levels of MyHCIIx and IIb were dramatically increased. On the other hand, MyHCIIx expression was significantly decreased while that of MyHCIIb was sustained in fast muscle groups such as the gastrocnemius and EDL. Taken together, our experimental results show that Vgll2 may be required not only for the establishment, but also for the maintenance of the adult slow muscle phenotype. A conditional knockout approach may serve to clear Vgll2 function in skeletal muscle development and exclude the possibility of the loss of Vgll2 during development that might affect muscle fiber type distribution in adult mice. Muscle fiber composition is an important determinant of muscle endurance in animals, along with the capacity for mitochondrial fuel oxidation and ATP synthesis [44][45][46] . Consequently, slow-to-fast fiber-type transformation often leads to a reduction in exercise endurance [47][48][49] . In the treadmill running experiments, the running distance and time were significantly decreased in Vgll2 −/− mice, indicating that genetic deletion of Vgll2 impaired the endurance capacity of the mice. On the other hand, there was no difference in grip strength between Vgll2 −/− and control animals, suggesting that Vgll2 activity is dispensable for the maintenance of muscle integrity. Thus, slow-to-fast fiber type transformation of skeletal muscle in Vgll2 −/− mice has distinct effects on physical activity.
Previous reports showed that Vgll2 can activate TEAD family transcription factors by physical interaction in in vitro assays 26 . We verified that Vgll2 forms a complex with TEAD1/4 in neonatal muscle. TEAD1/4 regulates the expression of muscle specific genes. Overexpression of TEAD1 induces a transition toward a slow muscle contractile phenotype and satellite cell hyperplasia 30,50 . These previous data indicate that TEAD1 are sufficient factors in both muscle specific gene expression and regeneration. However, they require additional factors or coactivators for gene expression 51 . It is noteworthy that TEAD1/4 expression is not restricted to specific tissues. Hence, skeletal muscle-specific functions are conferred by muscle-specific cofactors. In recent studies, several cofactors for TEAD family members have been identified, including YAP/TAZ, p160 family of nuclear receptor coactivators, and other Vgll members 26,27,37,38,[52][53][54] . Vgll2 is exclusively expressed in skeletal muscles in adult animals 25,26 . In addition, we now show that Vgll2 is more highly expressed in slow muscle groups. Thus, our data suggest the possibility that Vgll2 contribute to TEAD-dependent regulation of muscle-specific genes, especially in slow-twitch muscle.
Many striated muscle-specific genes contain muscle-specific cytidine-adenosine-thymidine (MCAT) elements in the control region [55][56][57] . MyHCI is also a well-established MCAT-element containing gene. Several previous studies described that the Myh7 promoter was activated by TEAD proteins through the direct binding of TEAD proteins on MCAT sequences [58][59][60] . Together with our results, these data suggest that MyHCI might be one of the direct targets of the Vgll2-TEAD1/4 complex. Interestingly, miR-208b, which works as a slow type regulator in a miRNA-mediated transcriptional regulatory network, was also decreased in the GPS muscles of Vgll2 −/− neonates in parallel with Myh7 expression. miR-208b represses negative regulators of slow-twitch muscle genes such as Sox6, Sp3, and Purβ 21 . In our study, the expression of Sox6, Sp3, and Purβ proteins was increased in Vgll2 deficient GPS muscles. As miR-208b uses the Myh7 promoter 21 , the Vgll2-TEAD1 complex may also regulate miR-208b expression through Myh7 promoter regulation.
In our gene expression analysis, MyHCIIa was remarkably decreased in Vgll2-deficient muscles. Impairment of exercise tolerance in Vgll2-deficient mice could be induced by downregulation of not only MyHCI, but also MyHCIIa because type IIa fibers are mitochondria rich and are resistant to fatigue. Thus, MyHCIIa might be a target of Vgll2. Although several studies described the regulatory mechanisms of MyHCIIa expression 8, 61, 62 , MCAT and its related elements have not been identified in the MyHCIIa regulatory region. Additional studies are required to understand how Vgll2-associated transcription pathway participates in this regulation. The expression of Myh1, encoding MyHCIIx, is also significantly decreased in Vgll2 deficient GPS muscles at P7. Given that Myh1 is one of the Sox6 target genes 41 , it raises the possibility that the decrease in Myh1 expression is associated with an increase in Sox6 protein. Thus, both direct and indirect involvements of Vgll2 are proposed for muscle fibers regulation.
This study sheds light on the role for Vgll2 in slow-twitch myofiber gene expression in vivo. After determination by basal regulation, muscle fiber type is modified in response to external stimuli like muscle usage. Previous studies implicated TEAD1 in the exercise-induced fiber type remodeling 30 . Additional experiments are necessary to elucidate the participation of Vgll2 in the regulation of exercise-induced fast-to-slow fiber type transition. Furthermore, previous studies pointed out the possibility that Vgll2 can interact with another transcription factor, MEF2C 26 . Further analyses are warranted to fully investigate Vgll2 interaction factors and its downstream target genes.
In summary, this study provides the first in vivo evidence that Vgll2 participates in muscle fiber-type specification and in the maintenance of normal muscle fiber composition at the adult stage under basal conditions. Moreover, our findings suggested the direct and indirect involvement of the Vgll2-TEAD1 complex in the regulatory program for slow skeletal muscle gene expression.

Methods
Antibody Production. Peptides corresponding to residues 293 to 308 (QSLGLSVDSGKRRREC) of mouse Vgll2 were synthesized as an immunogen for the production of polyclonal antibodies in rabbits using a customized service (MBL, Nagoya, Japan). After verification of Vgll2 specificity by ELISA and Western blotting, the antiserum was affinity-purified against the peptide.
Construction of the Targeting Vector. The targeting construct (Fig. 1b)

contained 1.3 kb and 7.4 kb of
Vgll2 homologous sequences on the 5′ and 3′ regions of the lacZ expression cassette with a SV40 polyadenylation signal, which followed the Vgll2 initiation codon. The loxP-flanked puromycin-resistance gene expression cassette (Puro) driven by a phosphoglycerate kinase promoter was inserted in the reverse orientation after the lacZ expression cassette. The diphtheria toxin A expression cassette (DT-A) was included at the 5′-end of the targeting vector for negative selection.
Generation of Vgll2-Deficient Mice. The D3 mouse embryonic stem cells 63 were electroporated with the targeting vector and selected in medium containing puromycin. Targeted clones were identified by Southern blotting using the DIG High prime DNA Labeling and Detection Starter Kit II (Roche, Basel, Switzerland) with 5′-external probes. Chimeric mice were obtained by injection of positive ES cells into blastocysts and crossed with wild-type C57BL/6 J mice (Japan SLC, Hamamatsu, Japan) to test for germline transmission of the disrupted Vgll2 allele. The genotypes of offspring were examined by Southern blotting analysis. To establish a mouse line carrying the disrupted Vgll2 allele, heterozygous mice were backcrossed with C57BL/6 J mice. The mouse experiments were approved by the Animal Care and Use Committee of the National Cerebral and Cardiovascular Center in Japan, and were performed in accordance with the institutional and national guidelines and regulations.

High Resolution Gel Electrophoresis. Protein extraction and high resolution gel electrophoresis for
MyHC isoform separation were performed as described previously 64 . Briefly, muscle tissues were isolated from both legs per mouse was homogenized in buffer (0.1 M Tris-HCl, pH 8.0, 10% SDS, 40 mM DTT, 5 mM EDTA) supplemented with protease inhibitor cocktail (Complete Ultra Mini Protease inhibitor cocktail tablets used as instructed; Roche) using a BioMasher II homogenizer (Nippi), and 50 ng of protein sample was separated on 8% acrylamide gel containing glycerol by 140 V for 22 h at 4 °C. The gel was then stained with Silver Stain II Kit (Wako) according to the manufacturer's instructions.
Immunofluorescence. Immunohistochemistry techniques were used for fiber type determination.
Soleus muscles were harvested and frozen in liquid nitrogen cooled isopentane. Frozen sections (7 μm) were cut in a cryostat on microscope slides. Slides were allowed to warm up to room temperature, fixed in 4% paraformaldehyde-PBS for 20 min at 4 °C, and permeabilized with 0.1% Triton X-100-PBS for 10 min at 4 °C, and immersed in 0.01 M citrate buffer (pH 3.0) at 37 °C for 30 min for antigen retrieval. Slides were then blocked with 10% normal goat serum (Nichirei, Tokyo, Japan) for 1 h at room temperature, and incubated at 4 °C for overnight with the following primary antibody cocktail: MyHCI antibody (BA-F8), MyHCIIa antibody (SC-71), and MyHCIIb antibody (BF-F3) were mixed in equal volume. Slides were then incubated for 1 h at room temperature with the following secondary antibody cocktail: Alexa Fluor 488 goat anti-mouse IgG2b (Invitrogen, A-21141, 1:200), Alexa Fluor 647 goat anti-mouse IgG1 (Invitrogen, A-21240, 1:200), and Alexa Fluor 568 anti-mouse IgM (Invitrogen, A-21043, 1:200). Images were captured under the FLUOVIEW FV10i (Olympus, Tokyo, Japan), then merged and pseudocolored in FV10-ASW (Olympus). All primary antibodies against MyHC isoforms were purchased from DSMZ (Braunschweig, Germany). Collected cell culture supernatants were used as primary antibody solutions.
Treadmill Exercise. The muscle endurance test was performed as described previously 65 . In brief, we used a MK-680S treadmill (Muromachi Kikai Co., Ltd., Tokyo, Japan). For 3 days, animals were acclimated to treadmill running for 5 min at a speed of 10 m/min on a 0% grade. After acclimation, animals ran on a treadmill with a 10% uphill grade starting at a speed of 10 m/min for 5 min. Every subsequent 2 min, the speed was increased by 2 m/min until the mice were exhausted. Exhaustion was defined as the inability of the animal to remain on the treadmill despite mechanical prodding. Running time and speed were measured, and the distance was calculated.
Grip Strength. Grip strength was measured as described previously 65 . In brief, we use a MK-380M grip strength meter (Muromachi Kikai Co., Ltd.). The grip strength of each individual mouse was measured 10 times, the same measurements were repeated on the next day, and the highest value of each experiment was used.

Statistical analyses.
Quantitative analyses were performed on at least 3 independent biological samples.
Data are expressed as the mean ± SEM. Data were analyzed by using the Student's t-test or one-way ANOVA with Dunnett post hoc test. P < 0.05 was considered as significant.