?The?Journal?of?Clinical?Investigation http://www.jci.org Volume 117 Number 9 September 2007
Histone deacetylase degradation and
MEF2 activation promote the formation
of slow-twitch myofibers
Matthew J. Potthoff,1 Hai Wu,1 Michael A. Arnold,1 John M. Shelton,2 Johannes Backs,1
John McAnally,1 James A. Richardson,3 Rhonda Bassel-Duby,1 and Eric N. Olson1
1Department of Molecular Biology, 2Department of Internal Medicine, and
3Department of Pathology, University of Texas Southwestern Medical Center, Dallas, Texas, USA.
Skeletal muscle fibers of adult vertebrates differ markedly with
respect to their contractile and metabolic properties, which reflect
different patterns of gene expression (1). Slow-twitch or type I
myofibers exhibit an oxidative metabolism, are rich in mitochon-
dria, heavily vascularized, and resistant to fatigue. In contrast, fast-
twitch or type II fibers exhibit glycolytic metabolism, are involved
in rapid bursts of contraction, and fatigue rapidly. Myofibers can
be further classified as either type I, IIa, IIx/d, or IIb, depending on
the type of myosin heavy chain (MHC) isoform expressed (2). The
heterogeneity of skeletal myofibers is reflected at the molecular
level in that almost every protein involved in contraction (MHC,
myosin light chain, troponin I, troponin T, troponin C, actinin,
etc.) has at least 2 isoforms expressed discretely in slow (type I)
and fast (type II) fibers (3). In adult animals, specialized myofiber
phenotypes remain plastic and vary in response to contractile load,
hormonal milieu, and systemic diseases (4). Functional overload or
exercise training results in transformation of preexisting fast fibers
to a slow-twitch, oxidative phenotype (5). Conversely, decreased
neuromuscular activity induced by spinal cord injury, limb immo-
bilization, space flight, or blockade of action potential conduction
causes a slow-to-fast myofiber conversion (6).
Functional demands modulate skeletal muscle phenotypes
by activating signaling pathways that modify the gene expres-
sion profile of the myofiber. The signaling pathways involved in
myofiber remodeling are of particular interest because of their
relevance to several human disorders, including muscle dystro-
phy, metabolic disorders, and muscle atrophy (7). Increasing the
abundance of slow, oxidative fibers in the mdx mouse model of
Duchenne muscular dystrophy, for example, reduces the severity
of the dystrophic phenotype (8, 9). Skeletal muscles also play an
important role in whole-body metabolism, such that increasing
the number of type I fibers enhances insulin-mediated glucose
uptake and protects against glucose intolerance (10), which
could have important therapeutic implications for diabetes and
other metabolic diseases.
The myocyte enhancer factor 2 (MEF2) transcription factor, a
key regulator of muscle development, is preferentially activated
in slow, oxidative myofibers (11) and responds to calcium-depen-
dent signaling pathways that promote the transformation of fast,
glycolytic fibers into slow, oxidative fibers (12). The transcrip-
tional activity of MEF2 is repressed by class II histone deacety-
lases (HDACs) (13–15). However, the potential involvement of
MEF2 and class II HDACs in regulating myofiber identity in vivo
has not been explored.
In this study, we show class II HDACs are selectively degraded
by the proteasome in slow, oxidative myofibers, enabling MEF2
to activate the slow myofiber gene program. Consistent with these
conclusions, forced expression of class II HDACs in skeletal mus-
cle or genetic deletion of Mef2c or Mef2d blocks activity-dependent
fast- to slow-fiber transformation whereas expression of a hyper-
active MEF2 protein promotes the slow-fiber phenotype, enhanc-
ing endurance and enabling mice to run almost twice the distance
of WT littermates. These findings provide new insights into the
Nonstandard?abbreviations?used: CaMK, calcium/calmodulin-dependent protein
kinase; DOX, doxycycline; EDL, extensor digitorum longus; GP, gastrocnemius and
plantaris; HDAC, histone deacetylase; MEF2, myocyte enhancer factor 2; MHC, myo-
sin heavy chain; Myo, myogenin promoter/MEF2 enhancer; Myo-Cre, Myo expressing
Cre recombinase; Myo-MEF2C-VP16, Myo expressing a MEF2C-VP16 fusion protein;
Myo-tTA, Myo expressing tTA; NLS, nuclear localization signal; PGC-1α, peroxisome
proliferator-activated receptor–γ coactivator-1α; PLA, plantaris; SkM-KO mice, mice
with a skeletal muscle–specific knockout; tet, tetracycline responsive; tTA, tetracycline
transactivator; WV, white vastus lateralis.
Conflict?of?interest: The authors have declared that no conflict of interest exists.
Citation?for?this?article: J. Clin. Invest. 117:2459–2467 (2007). doi:10.1172/JCI31960.
Related Commentary, page 2388
2460?The?Journal?of?Clinical?Investigation http://www.jci.org Volume 117 Number 9 September 2007
molecular basis of skeletal muscle performance and have impor-
tant implications for possible therapeutic manipulation of muscle
function for amelioration of muscular diseases.
Reduction of class II HDAC proteins in soleus muscle. We speculated that
variations in MEF2 activity patterns seen among different types of
myofibers (12) might arise from differences in the extent of MEF2
repression by class II HDACs. To begin to explore this possibility,
we determined the expression patterns of HDAC proteins in sev-
eral skeletal muscles containing different proportions of fast and
slow myofibers by Western blot analysis (Figure 1A).
Soleus muscle is composed primarily of slow, oxidative fibers
with only a few fast, glycolytic fibers (16). This fiber-type compo-
sition fits the physiological functions of soleus muscle, which is
used almost continuously to maintain posture and resist gravity.
Three other skeletal muscles, plantaris (PLA), extensor digito-
rum longus (EDL), and superficial white vastus lateralis (WV),
contain very few slow fibers (16). As seen in Figure 1A, class I
HDACs (HDAC1, -2, and -3) were expressed at comparable levels
in different muscle groups. In contrast, class II HDACs (HDAC4,
-5, and -7) were expressed preferentially in the fast-fiber–domi-
nant PLA, EDL, and WV muscles, with relatively little expression
in the slow-fiber–enriched soleus (Figure 1A). MEF2 protein
expression levels did not differ among muscle types (Supple-
mental Figure 1; supplemental material available online with
this article; doi:10.1172/JCI31960DS1). The relatively low level
of class II HDAC protein expression in soleus appeared to reflect
a posttranscriptional mechanism since mRNA transcripts encod-
ing class II HDACs were more abundant in soleus than in WV
muscles, as revealed by both RT-PCR and Northern blot analyses
(Figure 1B and Supplemental Figure 2).
Downregulation of class II HDACs in transgenic skeletal muscles trans-
formed toward a slow, oxidative phenotype. Forced expression of consti-
tutively active calcineurin or calcium/calmodulin-dependent pro-
tein kinase IV (CaMKIV) in adult fast, glycolytic fibers of transgenic
mice results in an increase in the number of slow fibers (17, 18).
We used these transgenic mouse models to determine whether
fast- to slow-fiber transformation correlated with a downregula-
tion of class II HDACs, as might be expected if class II HDACs
are involved in the fiber-type switch. The levels of class I HDACs
(HDAC1, -2, and -3) were similar in WV muscles from WT and
transgenic mice whereas the transformation of WV muscles toward
a slow myofiber identity was associated with diminished expres-
sion of class II HDAC (HDAC4, -5, and -7) proteins (Figure 1C),
consistent with the possibility that class II HDACs repress the
expression of slow-fiber genes in fast myofibers.
Class II HDACs redundantly regulate slow, oxidative fiber expres-
sion. To directly examine the potential role of class II HDACs in
regulating fiber-type identity, we analyzed adult skeletal muscles
from mutant mice lacking 1 or more class II HDACs. Hdac5–/– and
Hdac9–/– mice are viable (15, 19), so we were able to analyze the
fiber-type composition of these homozygous mutants with global
Hdac gene deletion in all tissues. However, because Hdac4–/– mice
die at birth from skeletal defects (20) and Hdac7–/– mice die during
embryogenesis from vascular defects (21), we used floxed alleles to
delete these genes specifically in skeletal muscle using a skeletal
muscle–specific Cre recombinase transgene (Myo-Cre) (SkM-KO
mice), thereby avoiding lethality (22). Mice lacking individual class
II HDACs did not display abnormalities in fiber-type switching
or skeletal muscle development (Figure 2A). In contrast, soleus
muscles from Hdac5–/–Hdac9–/– and Hdac4fl/-;Myo-creHdac5–/– dou-
ble-knockout mice showed an increase in the percentage of slow
myofibers from 48% ± 2.6% to 70% ± 3% (P < 0.003) and 63% ± 6.6%
Posttranscriptional downregulation of class II HDACs
in soleus muscle. Soleus (SOL), PLA, EDL, and WV
muscles were dissected from the hind limbs of adult
WT mice (8–10 weeks of age). (A) Protein expres-
sion of HDACs was assayed using antibodies specif-
ic for individual HDAC proteins. α-Tubulin level indi-
cated equal loading. (B) RNA expression of HDACs
in SOL versus WV was analyzed by RT-PCR in the
presence (+) or absence (–) of reverse transcriptase
(RT). Skeletal α-actin primers were used to show
equivalent cDNA input. (C) Immunoblots of HDACs
using WV muscle extracts from WT and 2 transgenic
mouse models (10 weeks old) overexpressing active
calcineurin A (CnA Tg) or CaMKIV (CaMK Tg).
?The?Journal?of?Clinical?Investigation http://www.jci.org Volume 117 Number 9 September 2007
(P < 0.05), respectively (Figure 2, A and B). The fiber-type composi-
tion of the soleus of Hdac5+/–Hdac9–/– mice and Hdac4+/–Hdac5–/–
mice was identical to that of WT mice (data not shown) whereas
Hdac4+/–Hdac5–/–Hdac9+/– mutant mice showed an increase in slow
fibers comparable to that of double-mutant mice (Figure 2A), sug-
gesting that deletion of any combination of 4 alleles of Hdac4, -5,
or -9 results in enhanced slow-fiber gene expression.
Analysis of the expression of transcripts encoding the individual
MHC isoforms revealed an increase in expression of oxidative genes
(MHC type I [MHC-I] and -IIa) in soleus and PLA muscles of Hdac5–/–
Hdac9–/– mutant mice (Figure 2C) compared with Hdac5+/–Hdac9–/– or
Hdac5–/–Hdac9+/– littermates. These results suggested that a reduction
in expression of class II HDAC proteins below a specific threshold
results in an increase in slow and oxidative fibers.
Forced expression of HDAC5 blocks fiber-type switching. Exercise train-
ing transforms preexisting fast fibers to an oxidative phenotype (5).
To determine whether class II HDACs modulated fiber-type switch-
ing in response to exercise, we constructed an inducible skeletal
muscle–specific transgenic system using the myogenin promoter/
MEF2 enhancer (Myo) to drive expression of the tetracycline trans-
activator (tTA) (Myo-tTA), which acts in trans to activate a tetra-
cycline-responsive transgene. In this system, the transgene is not
expressed in the presence of doxycycline (DOX) but is induced when
the drug is removed. To verify the spatial expression of tTA in the
Myo-tTA transgenic mice, we generated transgenic mice harboring
a lacZ reporter gene cloned behind the tetracycline responsive (tet-
responsive) expression cassette (tet-lacZ). When the Myo-tTA trans-
genic mice were crossed to the tet-lacZ reporter mice in the absence
of DOX, lacZ expression was observed specifically in skeletal muscle
without preference for adult slow or fast fibers (data not shown).
The Myo-tTA transgenic mice were bred to responder mice bearing
a tet-responsive transgene encoding a signal-resistant FLAG-tagged
human HDAC5 mutant protein (HDAC5S/A) (23). Expression of the
FLAG-tagged HDAC5 protein was detected by Western blot analysis
Class II HDACs redundantly regulate slow, oxidative fiber expression. (A) Soleus muscles from WT, Hdac5–/– (Hdac5 KO), Hdac9–/– (Hdac9
KO), Hdac4fl/–;Myo-Cre (Hdac4 SkM-KO), Hdac7fl/–;Myo-Cre(Hdac7 SkM-KO), and class II HDAC compound mutant mice were analyzed by
metachromatic ATPase staining. Type I fibers stain dark blue. Type II fibers stain light blue. Original magnification, ×10. Scale bar: 100 μm. (B)
Quantification of fiber-type distribution based on fiber-type analysis in A. (C) Transcripts of MHC isoforms were determined in soleus and PLA
muscles from mice of the indicated genotypes by quantitative real-time PCR. (D) Anti-FLAG M2 antibody on a Western blot analysis of proteins
isolated from GP muscles of 4-week-old Myo-tTA/tet-HDAC5 mice treated with DOX or 10 days after removal of DOX. Tubulin served as a load-
ing control. (E) Metachromatic ATPase staining of GP muscles harvested from sedentary WT, 4-week-exercised control (tet-HDAC5 [no tTA]),
and 4-week-exercised HDAC5 transgenic (Myo-tTA/tet-HDAC5) mice. Original magnification, ×4. Scale bar: 300 μm. Dashed red lines delineate
gastrocnemius (GA) muscle from PLA.
2462? The?Journal?of?Clinical?Investigation http://www.jci.org Volume 117 Number 9 September 2007
with an anti-FLAG antibody using total protein extracts from gastroc-
nemius and PLA (GP) muscles of 4-week-old tet-HDAC5/Myo-tTA
transgenic mice that had DOX removed 10 days earlier (Figure 2D).
Overexpression of HDAC5 in transgenic mice was confirmed by
probing with an HDAC5 antibody (Figure 2D). FLAG-HDAC5 was
not detected in the presence of DOX (Figure 2D).
To analyze the influence of HDAC5 on fiber-type switching,
DOX was removed from 6-week-old tet-HDAC5/Myo-tTA double-
transgenic mice, and these mice were provided free access to a run-
ning wheel for 4 weeks, a time period shown previously to allow
the transformation of fast, glycolytic fibers to oxidative fibers (12).
Tet-HDAC5/Myo-tTA transgenic mice and control mice ran volun-
tarily at comparable intensities (Supplemental Figure 3, A and B).
Metachromatic ATPase staining of skeletal muscles showed a
pronounced increase in type I and IIa fibers within GP muscles of
exercised tet-HDAC5 mice (without tTA) compared with unexer-
cised mice (Figure 2E). In contrast, GP muscles from exercised tet-
HDAC5/Myo-tTA double-transgenic mice without DOX did not
show an increase in type I and type IIa fibers compared with sed-
entary mice (Figure 2E). Quantification of slow fibers revealed an
approximate 10-fold reduction in the number of slow fibers from
exercised HDAC5 transgenic mice compared with exercised WT
mice (data not shown). Sedentary tet-HDAC5/Myo-tTA double-
transgenic mice with and without DOX displayed normal fiber-
type distributions (data not shown). We conclude that continuous
repression of class II HDAC target genes in adult skeletal muscle
is sufficient to inhibit exercise-induced fiber-type switching.
Requirement of MEF2 for establishing slow, oxidative myofiber identity.
To examine whether class II HDAC regulation of fiber-type switch-
ing occurs through repression of MEF2 activity, we analyzed the
skeletal muscles from individual MEF2 knockout mice. Condi-
tional alleles of Mef2c and Mef2d (24, 25) were deleted specifically
in skeletal muscle using transgenic mice that express Cre recombi-
nase under the control of the myogenin promoter/MEF2 enhancer
(Myo-Cre) (22), which is active in both fast and slow fibers (data
not shown). As shown in Figure 3A, skeletal muscle–specific dele-
tion (SkM-KO) of Mef2c or Mef2d using Myo-Cre resulted in a
reduction in slow fibers within the soleus whereas the abundance
of slow fibers was unaltered in Mef2a–/– mice (Figure 3A) or Mef2c+/–
and Mef2d+/– mice (data not shown).
To further validate the reduction of slow fibers follow-
ing skeletal muscle–specific deletion of Mef2c, we performed
immunohistochemistry for type I fibers using a MHC-I specific
antibody. As shown in Figure 3B, skeletal muscle lacking Mef2c
displayed a loss of type I fibers in the GP and a reduction in num-
ber (and intensity) of type I fibers in the soleus. Moreover, using
glycerol gradient silver staining of MHC isoforms from the Mef2c
SkM-KO soleus muscle, we discovered that these muscles dis-
play a reduction in MHC-I (Figure 3C). Specifically, a decrease in
the percentage of slow myofibers from 48% ± 2.6% to 25% ± 3.6%
(P < 0.002) and 33% ± 11.4% (P < 0.001) was observed in the sole-
us of the Mef2c SkM-KO and Mef2d SkM-KO mice, respectively
(Figure 3D). These findings demonstrate that MEF2C and MEF2D
activate slow-fiber genes and that repression of fiber-type switch-
Requirement of MEF2 for establishing slow, oxidative myofiber identity. Muscles from individual MEF2 knockout mice: Mef2a–/–, Mef2c SkM-KO
(Mef2cfl/-;Myo-Cre), and Mef2d SkM-KO (Mef2dfl/fl;Myo-Cre) skeletal muscle conditional knockout mice were analyzed for fiber-type composi-
tion. (A) Metachromatic ATPase staining of soleus muscle. Type I fibers stain dark blue. Type II fibers stain light blue. Original magnification,
×10. Scale bar: 100 μm. (B) Immunohistochemistry of soleus and GP muscles of Mef2c SkM-KO and WT littermates using an MHC-I specific
antibody. Original magnification, ×2.5. Scale bar: 300 μm. (C) Glycerol gradient silver staining of protein extracts from soleus of WT and Mef2c
SkM-KO mice. MHC-I, -IIa/x, and -IIb isoforms are indicated. (D) Quantification of fiber-type distribution based on metachromatic ATPase
staining of MEF2 knockout mice.
?The?Journal?of?Clinical?Investigation http://www.jci.org Volume 117 Number 9 September 2007
ing by class II HDACs is mediated, at least in part, through their
repressive influence on MEF2C and MEF2D.
Activated MEF2 is sufficient to increase slow-fiber gene expression and
muscle performance. To determine whether MEF2 proteins were not
only necessary, but also sufficient, for properly establishing slow,
oxidative myofiber distribution, we tested whether expression of a
hyperactive MEF2C-VP16 chimera, which is insensitive to HDAC
repression, was sufficient to increase slow, oxidative fiber expression.
Indeed, skeletal muscle expression of MEF2C-VP16 (Myo-MEF2C-
VP16) was sufficient to increase the number of slow fibers in the PLA,
which is normally composed primarily of fast fibers (Figure 4B and
data not shown). This increase in slow-fiber abundance was remark-
ably similar to the fiber-type increase observed in WT mice after exer-
cise training (compare to Figure 2F). Analysis of muscle fiber mark-
ers and mitochondrial proteins by Western blot analysis revealed an
increase in the slow-fiber–specific contractile protein, Troponin I,
and the type I fiber oxidative proteins myoglobin and cytochrome c
(Figure 4A) in Myo-MEF2C-VP16 transgenic mice, confirming the
results of metachromatic ATPase staining. In addition, an increase in
expression of other metabolic genes and important metabolic tran-
scription factors, such as peroxisome proliferator-activated recep-
tor–γ coactivator-1α (PGC-1α), was observed in Myo-MEF2C-VP16
transgenic skeletal muscles (Supplemental Figure 4). These findings
demonstrate that MEF2 is sufficient to drive fast- to slow-fiber trans-
formation, mimicking the effect of exercise in vivo.
To examine the functional consequences of the increase in
slow fibers and oxidative capacity of Myo-MEF2C-VP16 trans-
genic muscles, we measured the endurance of these mice by
forced treadmill exercise on a 10% incline. As shown in Figure 4,
C and D, Myo-MEF2C-VP16 transgenic mice displayed a 75%
increase in running time and a 94% increase in distance, respec-
tively, compared with WT mice. Thus, activation of MEF2 is suf-
ficient to enhance skeletal muscle oxidative capacity and mito-
chondrial content, thereby diminishing muscle fatigability and
Class II HDACs are ubiquitinated and degraded by the proteasome.
To begin to understand the mechanistic basis for the lack of
accumulation of class II HDAC proteins in slow skeletal muscle
fibers, we examined the half-life of HDAC5 in vitro using a stable
C2C12 muscle cell line that constitutively expressed FLAG-tagged
HDAC5. Inhibition of protein synthesis with cycloheximide for 4
hours resulted in a precipitous decrease in the level of HDAC5 pro-
tein in skeletal myocytes (Figure 5A). In contrast, α-tubulin was
stable over this time period. The proteasome inhibitor MG132
blocked the degradation of HDAC5 (Figure 5B), suggesting that
HDAC5 is degraded by the proteasome pathway.
Since ubiquitination is a prerequisite for degradation by the
proteasome (26), we examined whether HDAC5 was ubiquitinat-
ed. Indeed, ubiquitinated HDAC5 was readily detectable when
HA-tagged ubiquitin was expressed in the HDAC5-expressing cell
line in the presence of MG132 (Figure 5C). The signal-resistant
HDAC5S/A mutant (13), lacking the regulatory phosphorylation
sites (serines 259 and 498), was ubiquitinated to the same extent as
the WT HDAC5 protein (Figure 5C), indicating that phosphoryla-
tion of HDAC5 on these residues is not a prerequisite for ubiqui-
tination. FLAG-HDAC4, -7, and MITR (a splice variant of HDAC9)
were also ubiquitinated in C2C12 cells (data not shown).
To determine whether ubiquitination of HDAC5 occurs in
the nucleus or cytoplasm, we compared the ubiquitination
of WT HDAC5, which is expressed primarily in the nucleus
(Figure 5D), and an HDAC5 mutant with a mutation in the nucle-
ar localization signal, HDAC5(K270R) (Figure 5D). As shown in
Figure 5E, the K270R mutant was not ubiquitinated, whereas
fusion of an SV40–nuclear localization signal (SV40-NLS) to the
HDAC5(K270R) protein restored nuclear localization and ubiqui-
tination. Together these results demonstrate that HDAC5 ubiqui-
tination occurs in the nucleus.
Blockade to HDAC degradation in vivo by MG132. To determine
whether class II HDACs are degraded via the proteasome pathway
in slow, oxidative myofibers in vivo, 8-week-old, WT C57BL/6 male
mice were injected with either DMSO or MG132 and expression
of HDAC4 and -5 was examined in the soleus, GP, tibialis anterior,
and EDL skeletal muscles after 6 hours, a time period shown pre-
viously to provide proteasome inhibition in vivo (27). Treatment
with MG132 in vivo increased the level of HDAC4 and HDAC5
Activated MEF2 is sufficient to increase slow-fiber expression. (A)
Western blot analysis of Myo-MEF2C-VP16 transgene expression
using an anti-VP16 antibody. Expression of the slow-fiber–specific
troponin I and oxidative markers myoglobin and cytochrome c in pro-
tein extract of GP muscles of Myo-MEF2C-VP16 transgenic mice. (B)
Metachromatic ATPase staining of gastrocnemius and PLA muscles
of WT and Myo-MEF2C-VP16 transgenic mice. Original magnifica-
tion, ×4. Scale bar: 300 μm. Dashed red lines delineate gastrocnemius
muscle from PLA. (C and D) Exercise endurance and muscle perfor-
mance showing total time running (min; C) and total distance run (m; D)
of Myo-MEF2-VP16 transgenic muscles were analyzed by forced
treadmill exercise. Eight-week-old Myo-MEF2C-VP16 transgenic and
WT male mice with similar body weights were subjected to forced
treadmill exercise (n = 5 for each group) on a 10% incline.
2464?The?Journal?of?Clinical?Investigation http://www.jci.org Volume 117 Number 9 September 2007
Ubiquitination and degradation of class II HDACs in vitro and in vivo. (A) C2C12 cells stably expressing FLAG-tagged HDAC5 (C2C12-HDAC5)
were treated with cycloheximide (CHX, 25 μM) for 0, 1, 2, or 4 hours before cells were lysed and FLAG-HDAC5 expression was measured by
Western blot analysis using anti-FLAG M2 antibody. Tubulin immunoblot showed equivalent loading of each lane. (B) C2C12-HDAC5 cells were
treated with cycloheximide and MG132, singly or in combination for 4 hours, and FLAG-HDAC5 expression was analyzed. (C) C2C12-HDAC5
and C2C12 cells stably expressing CaMK-resistant HDAC5 (S259/498A) were transfected with or without HA-tagged ubiquitin (HA-Ub) and
treated with or without MG-132 (25 μM) for 4 hours; the ubiquitination status of WT and mutant HDAC5 was analyzed. FLAG expression in inputs
shows equal loading. (D) Subcellular localization of Myc-HDAC5 (WT), Myc-HDAC5(K270R), or Myc-SV40 NLS-HDAC5(K270R) in C2C12 cells.
(E) The ubiquitination status of cytoplasmic HDAC5 [Myc-HDAC5(K270R)] or nuclear HDAC5 [Myc-SV40 NLS-HDAC5(K270R)] was analyzed.
(F and G) WT C57BL/6 males (8 weeks old) were IP injected with DMSO or MG132 for 6 hours. Protein was isolated from SOL, GP, tibialis
anterior (TA), and EDL muscles and analyzed for expression of (F) HDAC4 or (G) HDAC5. Tubulin shows equal loading. (H) Treatment with
MG132 decreases MEF2 activation. Six hours after DesMEF-lacZ mice were injected with DMSO or MG132, mice were run for approximately
3 hours using forced treadmill exercise. Skeletal muscles were then isolated from DMSO- and MG132-treated DesMEF mice and analyzed for
lacZ expression. LacZ expression was reduced in MG132-treated muscles (where class II HDAC expression was increased).
?The?Journal?of?Clinical?Investigation http://www.jci.org Volume 117 Number 9 September 2007
protein expression in the soleus to that of the EDL (Figure 5, F
and G, respectively), consistent with the in vitro results and dem-
onstrating that class II HDAC proteins are specifically degraded by
the proteasome in slow and oxidative myofibers.
Finally, since treatment with MG132 results in an increase in
class II HDAC protein, we examined its effect on MEF2 activ-
ity in vivo using a transgene reporter line (DesMEF lacZ) that
expresses lacZ under control of 3 tandem MEF2 sites (28). Prior
studies showed that the expression of lacZ in these mice provides
a faithful measure of MEF2 activity (12). Sedentary MEF2 report-
er mice (DesMEF lacZ) do not express lacZ in skeletal muscle.
However, lacZ expression is observed when DesMEF mice are
exercised (12). Therefore, we injected DesMEF lacZ transgenic
mice with either DMSO or MG132 and after 6 hours ran the mice
for approximately 3 hours using forced treadmill exercise. Skel-
etal muscles were then isolated from DMSO- and MG132-treated
DesMEF lacZ mice and analyzed for lacZ expression. Soleus mus-
cle from exercised, DMSO-treated DesMEF lacZ mice expressed
lacZ, while lacZ expression in exercised, MG132-treated DesMEF
lacZ muscles was significantly reduced (Figure 5H). These results
demonstrate that inhibition of the proteasome in vivo, which
prevents HDAC degradation, reduces the activation of MEF2 and
slow, oxidative fiber gene expression.
The results of this study show that slow and oxidative myofiber
identity and muscle performance are governed by the balance
between positive and negative signaling by MEF2 and class II
HDACs, respectively. Degradation of class II HDAC proteins allows
sustained activation of MEF2, which promotes the establishment
of slow and oxidative myofibers and, strikingly, enhances muscle
endurance and fatigue resistance (Figure 6).
The contractile properties of skeletal myofibers reflect a combina-
tion of developmental and extrinsic inputs. During embryogenesis,
fast- and slow-twitch fibers are patterned by specific developmental
cues (3). After birth, the pattern of motor innervation plays a key
role in influencing muscle fiber type. Phasic motor neuron firing
at high frequency (100–150 Hz) promotes the formation of fast
fibers, which display brief, high-amplitude calcium transients and
low ambient calcium levels (<50 nM) whereas tonic motor neuron
stimulation (10–20 Hz) favors the formation of slow fibers, which
maintain higher intracellular calcium levels (100–300 nM) (29).
Calcineurin and CaMK have been implicated in the transduction of
calcium-dependent signals that upregulate the expression of oxida-
tive, slow-fiber–specific genes in skeletal muscle (17, 18, 30). How-
ever, the precise mechanisms whereby these signaling pathways
modulate the slow-fiber phenotype have not been defined (30).
Our results show that skeletal muscle of transgenic mice that
has undergone a fast-to-slow myofiber transformation in response
to activated calcineurin and CaMK displays a reduction in abun-
dance of class II HDAC proteins, suggesting that these calcium-
dependent signaling pathways act, at least in part, by enhanc-
ing degradation of class II HDAC proteins. The calcineurin and
CaMK signaling pathways have also been shown to promote the
phosphorylation of class II HDACs on a series of conserved ser-
ine residues, which mediate signal-dependent nuclear export and
derepression of MEF2 target genes (13, 14). Thus, it seems likely
that some combination of the 2 mechanisms — proteolysis and
regulated nuclear export — regulates class II HDACs and thereby
MEF2 activity and myofiber identity.
Nuclear factor of activated T cells (NFAT) transcription factors,
which serve as transcriptional mediators of calcineurin signal-
ing, have also been implicated in the control of slow-fiber gene
expression (30, 31), as has the nuclear receptor PPARδ (32) and
PGC-1α (33). Recently, PGC-1β was implicated in regulating type IIx/d
fiber formation (34). MEF2 interacts with NFAT (35) and PGC-1α
(36) and also regulates PGC-1α expression (23). Indeed, the
MEF2-VP16 superactivator upregulated PGC-1α but not PGC-1β
expression in skeletal muscle (Supplemental Figure 4). Thus,
MEF2 serves as a nodal point for the control of multiple down-
stream transcriptional regulators of the slow-fiber phenotype and
can potentially confer calcium sensitivity to other factors via its
signal-dependent interaction with class II HDACs.
Our results show that ubiquitination of class II HDACs occurs in
the nucleus. Nuclear ubiquitination of transcriptional activators has
also been described as a mechanism to regulate the extent and dura-
tion of activation of transcriptional activators (37). The ubiquitina-
tion of class II HDACs provides another mechanism by which tran-
scriptional activators may become activated. Class II HDACs and
MEF2 are also sumoylated, which enhances the repressive activity of
class II HDACs (38, 39). Whether sumoylation might be regulated
during myofiber specification in vivo has not been addressed.
The mechanism that directs ubiquitination and degradation of
class II HDACs in response to calcium signaling remains to be deter-
mined. It is tempting to speculate that phosphorylation of HDACs
serves as a signal for ubiquitination by recruiting specific E3 ligases.
Identification of the E3 ligase(s) for class II HDACs and the signals
regulating their degradation are currently under investigation.
Our results show that adult skeletal muscle phenotypes are
dictated by the extent of repression of MEF2 by class II HDACs
and that the fast-fiber phenotype results from the absence of
MEF2 activity. The fact that 4 Hdac alleles needed to be deleted
A model for the control of slow and oxidative fibers by MEF2 and class II
HDACs. Motor nerve activity regulates MEF2 activity and myofiber iden-
tity through ubiquitination (Ub) and degradation of class II HDACs.
2466?The?Journal?of?Clinical?Investigation http://www.jci.org Volume 117 Number 9 September 2007
to observe an increase in slow-fiber gene expression suggests
that there is substantial functional redundancy among different
HDACs with respect to repression of the slow-fiber gene program.
Forced expression of a signal-resistant mutant of HDAC5 is suf-
ficient to suppress the slow-fiber phenotype, whereas expression
of a hyperactive MEF2-VP16 chimera is sufficient to override the
repressive influence of endogenous class II HDACs and drive the
slow-fiber phenotype. These findings suggest that the fast-fiber
phenotype represents a “default” gene program resulting from
the absence of MEF2 activity.
Therapeutic potential. The ability of activated MEF2 to enhance
oxidative capacity and endurance of skeletal muscle suggests
opportunities for therapeutically enhancing muscle performance
by stimulating MEF2 activity. Increasing the number of slow fibers
in skeletal muscle via MEF2 also represents a potential method for
treating metabolic and muscular diseases (8, 9). One could imag-
ine augmenting MEF2 activity by interfering with the repressive
activity of class II HDACs by modulating the signaling pathways
that control HDAC phosphorylation, subcellular localization, or
degradation. In this regard, HDAC inhibitors have recently been
shown to suppress muscle pathology associated with muscular
dystrophy (40, 41) in mice.
In addition to regulating skeletal muscle gene expression and
function, MEF2 signaling has been shown to drive pathological
cardiac growth and remodeling, which result from signal-depen-
dent phosphorylation and nuclear export of class II HDACs in
cardiac myocytes (19). These adverse consequences of MEF2
activation pose interesting challenges to the goal of enhancing
MEF2 activity in skeletal muscle while avoiding possible cardio-
toxicity of such strategies and, conversely, to pharmacologically
preventing cardiac dysfunction without diminishing skeletal
Plasmid constructs, tissue culture, and cell transfection. The expression vector
encoding HA-ubiquitin was described previously (42). Myc-NLS-HDAC5
was a kind gift from Tim McKinsey. The K270R point mutation was gener-
ated by site-directed mutagenesis (Stratagene). Myo-tTA and Myo-MEF2C-
VP16 transgenic constructs were generated by cloning the tTA cassette
from the pUDH 15-1 vector (43), fusing a VP16 activation domain in frame
to the C terminus of full-length MEF2C, respectively, and cloning them
into the HindIII site of a pBSIISK(+) vector containing a hGHpolyA tail.
The myogenin promoter/MEF2C enhancer (22) was then cloned upstream
of the cassettes into KpnI/XhoI sites.
C2C12 myoblasts were grown in DMEM supplemented with fetal bovine
serum and antibiotics (100 U/ml penicillin and 100 μg/ml streptomycin).
For transient transfection assays, cells were plated and transfected 12
hours later using FuGENE (Roche Applied Science) following the man-
ufacturer’s instructions. Cells were harvested 24-48 hours after transfec-
tion. C2C12 stable lines expressing FLAG-tagged HDAC5 or FLAG-tagged
HDAC5S/A were established by G418 selection of C2C12 clones transfected
with plasmids encoding WT HDAC5 or a CaMK-resistant HDAC5 mutant,
in which Ser259 and Ser498 were mutated into alanines (13). Cells were
treated with cycloheximide (Sigma-Aldrich) and/or MG132 (Calbiochem;
EMD Biosciences) at indicated concentrations.
Generation of transgenic mice and knockout mice. Transgenic mice that express
constitutively active forms of calcineurin or CaMKIV under the control of a
muscle-specific enhancer from the muscle creatine kinase (MCK) gene are
described elsewhere (17, 18). Tet-HDAC5S/A transgenic mice are described
elsewhere (23). Myo-tTA, tet-lacZ, and Myo-MEF2C-VP16 transgenic mice
were generated by injecting linearized constructs into the pronuclei of fer-
tilized oocytes as previously described (44).
Hdac5–/– (15), Hdac9–/– (19), and Hdac7 conditional knockout mice had
been generated previously (21). Hdac4 conditional mice were generated by
flanking exon 6 with loxP sites, which results in an out-of-frame mutata-
tion in the Hdac4 allele. Mef2a–/– and Mef2c and Mef2d conditional knockout
mice have been described previously (24, 45). Correct gene targeting was
confirmed by Southern blot analysis, genomic sequencing, and RT-PCR.
Immunoprecipitation and immunoblotting. Immunoprecipitations were
performed as previously described (20). Antibodies against HA (1:1,000;
Sigma-Aldrich), FLAG M2 (1:4,000; Sigma-Aldrich), Myc (1:1,000; Santa
Cruz Biotechnology Inc.), α-tubulin (1:5,000; Sigma-Aldrich), HDAC1,
-2, and -3 (1:1,000 for all; Sigma-Aldrich), HDAC4 (1:500; Santa Cruz
Biotechnology Inc.), HDAC5 (1:1,000; Millipore), HDAC7 (1:1,000; Cell
Signaling Technology), MEF2A (1:1,000; Millipore), MEF2C (1:1,000;
Santa Cruz Biotechnology Inc.), MEF2D (1:2,500; BD Biosciences),
troponin I slow (1:2,500; Santa Cruz Biotechnology Inc.), cytochrome c
(1:2,500; BD Biosciences — Pharmingen), and myoglobin (1:3,000; Dako)
were used for immunoblot analyses. For analyzing endogenous class II
HDAC proteins, the ECL Advance Western Blotting Detection Kit
(Amersham Biosciences) was used.
In vivo pharmacological studies. Myo-tTA and tet-HDAC5S/A transgenic
mice were bred while receiving DOX (200 μg/ml) in water as previously
described (23). Myo-tTA\tet-HDAC5S/A transgenic mice were maintained
on DOX as needed. Voluntary wheel-running experiments were per-
formed and measured as previously described (12). Animals were allowed
to acclimate to running cages for 4 days prior to running recordings.
DOX was removed from the mice for the days of acclimation so that the
animals could begin expressing the transgene. After 4 days, wheel-run-
ning activity was measured continuously for 4 weeks.
Forced treadmill exercise experiments for analyzing Myo-MEF2C-VP16
transgenic mice were performed as follows. Prior to exercise, mice were
accustomed to the treadmill (Columbus Instruments) with a 5-minute run
at 7 m/min once per day for 2 days. The exercise test was performed on a 10%
incline for 10 m/min for the first 60 minutes, followed by 1 m/min incre-
ment increases for 3- to 15-minute intervals, then 45 minutes at 13 m/min,
followed by 1 m/min increment increases at 15-minute intervals until
exhaustion. Forced exercise ended when mice were unable to avoid repeated
electric shocks. MEF2 reporter mice (DesMEF lacZ) were run for approxi-
mately 3 hours at 9 m/min.
In vivo proteasome inhibition experiments were performed by
intraperitoneally delivering DMSO or MG132 as previously described for
6 hours (27), except 30 μmol/kg body weight MG132 was used for injec-
tions. All experiments involving animals were reviewed and approved by
the Institutional Animal Care and Research Advisory Committee of the
University of Texas Southwestern Medical Center.
RNA isolation and analysis. Total RNA was prepared from mouse tissues
using TRIzol (Invitrogen) following the manufacturer’s instructions.
Of total RNA, 1.5 μg was converted to cDNA using oligo dT primer and
SuperScript II Reverse Transcriptase (Invitrogen). For PCR reactions, 2%
of the cDNA pool was amplified. PCR cycles were optimized for each set
of primers. Sequences for HDAC PCR primers have been described previ-
ously (46). Quantitative real-time PCR was performed for indicated MHC
isoforms using SYBR Green (Applied Biosystems). Northern blots were
performed with 20 μg of total RNA in each lane and probed in ULTRAhyb
(Ambion) with labeled HDAC4, -5, or β-actin cDNA.
Fiber-type and immunohistological analysis. Soleus and GP muscles were
harvested from mice and flash frozen in embedding medium or fixed
in 4% paraformaldehyde as previously described (47). Fiber-type analysis
using metachromatic ATPase staining (48) and glycerol gradient silver
?The?Journal?of?Clinical?Investigation http://www.jci.org Volume 117 Number 9 September 2007
staining were performed as previously described (47). Paraffin sections
were stained with a MHC-I antibody, followed by treatment with DAB
Statistics. Data are presented as mean ± SEM. Differences between groups
were tested for statistical significance using the unpaired 2-tailed Student’s
t test. Values of P < 0.05 were considered significant.
We thank Andrew Williams, Yuri Kim, Dillon Phan, Bryan
Young, and Cheryl Nolen for technical assistance and Alisha
Tizenor for assistance with graphics. This work was supported
by grants from the NIH, the D.W. Reynolds Clinical Cardiovas-
cular Research Center, the Texas Advanced Technology Program,
the Muscular Dystrophy Association (to E.N. Olson), and the
Robert A. Welch Foundation.
Received for publication February 27, 2007, and accepted in revised
form May 29, 2007.
Address correspondence to: Eric N. Olson, Department of Molec-
ular Biology, University of Texas Southwestern Medical Center,
5323 Harry Hines Boulevard, Dallas, Texas 75390-9148, USA.
Phone: (214) 648-1187; Fax: (214) 648-1196; E-mail: eric.olson@
1. Williams, R.S., and Neufer, P.D. 1996. Regulation
of gene expression in skeletal muscle by contractile
activity. In The handbook of physiology. L.B. Rowell
and J.T. Shepard, editors. American Physiology
Society. Bethesda, Maryland, USA. Oxford Univer-
sity Press. New York, New York, USA. 1124–1150.
2. Pette, D., and Staron, R.S. 2000. Myosin isoforms,
muscle fiber types, and transitions. Microsc. Res.
3. Schiaffino, S., and Reggiani, C. 1996. Molecular
diversity of myofibrillar proteins: gene regula-
tion and functional significance. Physiol. Rev.
4. Baldwin, K.M., and Haddad, F. 2001. Effects of dif-
ferent activity and inactivity paradigms on myosin
heavy chain gene expression in striated muscle.
J. Appl. Physiol. 90:345–357.
5. Sugiura, T., Miyata, H., Kawai, Y., Matoba, H., and
Murakami, N. 1993. Changes in myosin heavy
chain isoform expression of overloaded rat skeletal
muscles. Int. J. Biochem. 25:1609–1613.
6. Talmadge, R.J. 2000. Myosin heavy chain isoform
expression following reduced neuromuscular activ-
ity: potential regulatory mechanisms. Muscle Nerve.
7. Bassel-Duby, R., and Olson, E.N. 2006. Signaling
pathways in skeletal muscle remodeling. Annu. Rev.
8. Chakkalakal, J.V., et al. 2004. Stimulation of calci-
neurin signaling attenuates the dystrophic pathol-
ogy in mdx mice. Hum. Mol. Genet. 13:379–388.
9. Stupka, N., et al. 2006. Activated calcineurin
ameliorates contraction-induced injury to skel-
etal muscles of mdx dystrophic mice. J. Physiol.
10. Ryder, J.W., Bassel-Duby, R., Olson, E.N., and Zierath,
J.R. 2003. Skeletal muscle reprogramming by activa-
tion of calcineurin improves insulin action on meta-
bolic pathways. J. Biol. Chem. 278:44298–44304.
11. Wu, H., et al. 2000. MEF2 responds to multiple
calcium-regulated signals in the control of skeletal
muscle fiber type. EMBO J. 19:1963–1973.
12. Wu, H., et al. 2001. Activation of MEF2 by muscle
activity is mediated through a calcineurin-depen-
dent pathway. EMBO J. 20:6414–6423.
13. McKinsey, T.A., Zhang, C.L., Lu, J., and Olson, E.N.
2000. Signal-dependent nuclear export of a histone
deacetylase regulates muscle differentiation.
14. McKinsey, T.A., Zhang, C.L., and Olson, E.N. 2002.
MEF2: a calcium-dependent regulator of cell divi-
sion, differentiation and death. Trends Biochem. Sci.
15. Chang, S., et al. 2004. Histone deacetylases 5 and
9 govern responsiveness of the heart to a subset
of stress signals and play redundant roles in heart
development. Mol. Cell. Biol. 24:8467–8476.
16. Burkholder, T.J., Fingado, B., Baron, S., and Lieber,
R.L. 1994. Relationship between muscle fiber types
and sizes and muscle architectural properties in the
mouse hindlimb. J. Morphol. 221:177–190.
17. Naya, F.J., et al. 2000. Stimulation of slow skeletal
muscle fiber gene expression by calcineurin in vivo.
J. Biol. Chem. 275:4545–4548.
18. Wu, H., et al. 2002. Regulation of mitochondrial
biogenesis in skeletal muscle by CaMK. Science.
19. Zhang, C.L., et al. 2002. Class II histone deacety-
lases act as signal-responsive repressors of cardiac
hypertrophy. Cell. 110:479–488.
20. Vega, R.B., et al. 2004. Histone deacetylase 4 controls
chondrocyte hypertrophy during skeletogenesis.
21. Chang, S., et al. 2006. Histone deacetylase 7 main-
tains vascular integrity by repressing matrix metal-
loproteinase 10. Cell. 126:321–334.
22. Li, S., et al. 2005. Requirement for serum response
factor for skeletal muscle growth and maturation
revealed by tissue-specific gene deletion in mice.
Proc. Natl. Acad. Sci. U. S. A. 102:1082–1087.
23. Czubryt, M.P., McAnally, J., Fishman, G.I., and
Olson, E.N. 2003. Regulation of peroxisome pro-
liferator-activated receptor gamma coactivator 1
alpha (PGC-1 alpha) and mitochondrial function
by MEF2 and HDAC5. Proc. Natl. Acad. Sci. U. S. A.
24. Arnold, M.A., et al. 2007. MEF2C transcription fac-
tor controls chondrocyte hypertrophy and bone
development. Dev. Cell. 12:377–389.
25. Haberland, M., et al. 2006. Regulation of HDAC9
gene expression by MEF2 establishes a negative
feedback loop in the transcriptional circuitry of
muscle differentiation. Mol. Cell. Biol. 27:518–525.
26. Pickart, C.M. 2001. Mechanisms underlying ubiqui-
tination. Annu. Rev. Biochem. 70:503–533.
27. Luker, G.D., Pica, C.M., Song, J., Luker, K.E., and
Piwnica-Worms, D. 2003. Imaging 26S proteasome
activity and inhibition in living mice. Nat. Med.
28. Naya, F.J., Wu, C., Richardson, J.A., Overbeck, P.,
and Olson, E.N. 1999. Transcriptional activity of
MEF2 during mouse embryogenesis monitored
with a MEF2-dependent transgene. Development.
29. Olson, E.N., and Williams, R.S. 2000. Calcineurin sig-
naling and muscle remodeling. Cell. 101:689–692.
30. Chin, E.R., et al. 1998. A calcineurin-dependent
transcriptional pathway controls skeletal muscle
fiber type. Genes Dev. 12:2499–2509.
31. Delling, U., et al. 2000. A calcineurin-NFATc3-
dependent pathway regulates skeletal muscle differ-
entiation and slow myosin heavy-chain expression.
Mol. Cell. Biol. 20:6600–6611.
32. Wang, Y.X., et al. 2004. Regulation of muscle fiber
type and running endurance by PPARdelta. PLoS
33. Lin, J., et al. 2002. Transcriptional co-activator
PGC-1 alpha drives the formation of slow-twitch
muscle fibres. Nature. 418:797–801.
34. Arany, Z., et al. 2007. The transcriptional coactivator
PGC-1beta drives the formation of oxidative type
IIX fibers in skeletal muscle. Cell Metab. 5:35–46.
35. Blaeser, F., Ho, N., Prywes, R., and Chatila, T.A.
2000. Ca(2+)-dependent gene expression medi-
ated by MEF2 transcription factors. J. Biol. Chem.
36. Moore, M.L., Park, E.A., and McMillin, J.B. 2003.
Upstream stimulatory factor represses the induc-
tion of carnitine palmitoyltransferase-Ibeta expres-
sion by PGC-1. J. Biol. Chem. 278:17263–17268.
37. Kodadek, T., Sikder, D., and Nalley, K. 2006. Keep-
ing transcriptional activators under control. Cell.
38. Zhao, X., Sternsdorf, T., Bolger, T.A., Evans, R.M.,
and Yao, T.P. 2005. Regulation of MEF2 by histone
deacetylase 4- and SIRT1 deacetylase-mediated
lysine modifications. Mol. Cell. Biol. 25:8456–8464.
39. Gregoire, S., and Yang, X.J. 2005. Association
with class IIa histone deacetylases upregulates the
sumoylation of MEF2 transcription factors. Mol.
Cell. Biol. 25:2273–2287.
40. Minetti, G.C., et al. 2006. Functional and mor-
phological recovery of dystrophic muscles in
mice treated with deacetylase inhibitors. Nat. Med.
41. Avila, A.M., et al. 2007. Trichostatin A increases
SMN expression and survival in a mouse model of
spinal muscular atrophy. J. Clin. Invest. 117:659–671.
42. Hakak, Y., and Martin, G.S. 1999. Ubiquitin-
dependent degradation of active Src. Curr. Biol.
43. Resnitzky, D., Gossen, M., Bujard, H., and Reed, S.I.
1994. Acceleration of the G1/S phase transition by
expression of cyclins D1 and E with an inducible
system. Mol. Cell. Biol. 14:1669–1679.
44. Cheng, T.C., Wallace, M.C., Merlie, J.P., and Olson,
E.N. 1993. Separable regulatory elements governing
myogenin transcription in mouse embryogenesis.
45. Naya, F.J., et al. 2002. Mitochondrial deficiency and
cardiac sudden death in mice lacking the MEF2A
transcription factor. Nat. Med. 8:1303–1309.
46. Wu, H., and Olson, E.N. 2002. Activation of the
MEF2 transcription factor in skeletal muscles
from myotonic mice. J. Clin. Invest. 109:1327–1333.
47. Oh, M., et al. 2005. Calcineurin is necessary for the
maintenance but not embryonic development of
slow muscle fibers. Mol. Cell. Biol. 25:6629–6638.
48. Ogilvie, R.W., and Feeback, D.L. 1990. A metachro-
matic dye-ATPase method for the simultaneous
identification of skeletal muscle fiber types I, IIA,
IIB and IIC. Stain Technol. 65:231–241.