Developmental Cell

Article

Codependent Activators DirectMyoblast-Specific MyoD TranscriptionPing Hu,1 Kenneth G. Geles,1,6 Ji-Hye Paik,2,3 Ronald A. DePinho,2,3,4,5 and Robert Tjian1,*1Howard Hughes Medical Institute, Department of Molecular and Cell Biology, University of California, Berkeley, Berkeley, CA 94720, USA2Department of Medical Oncology, Dana-Farber Cancer Institute3Department of Medicine4Department of Genetics5Center for Applied Cancer Science, Belfer Foundation Institute for Innovative Cancer Science, Dana-Farber Cancer Institute

Harvard Medical School, Boston, MA 02115, USA6Present address: Wyeth Research, Discovery Oncology, 401 N. Middletown Rd., Pearl River, NY 10965, USA*Correspondence: jmlim@berkeley.edu

DOI 10.1016/j.devcel.2008.08.018

SUMMARY

Although FoxO and Pax proteins represent two impor-tant families of transcription factors in determiningcell fate, they had not been functionally or physicallylinked together in mediating regulation of a commontarget gene during normal cellular transcription pro-grams. Here, we identify MyoD, a key regulator ofmyogenesis, as a direct target of FoxO3 and Pax3/7in myoblasts. Our cell-based assays and in vitro stud-ies reveal a tight codependent partnership betweenFoxO3 and Pax3/7 to coordinately recruit RNA poly-merase II and form a preinitiation complex (PIC) toactivate MyoD transcription in myoblasts. The role ofFoxO3 in regulating muscle differentiation is con-firmed in vivo by observed defects in muscle regener-ation caused by MyoD downregulation in FoxO3 nullmice. These data establish a mutual interdependenceand functional link between two families of transcrip-tion activators serving as potential signaling sensorsand regulators of cell fate commitment in directingtissue specific MyoD transcription.

INTRODUCTION

Embryonic differentiation, postnatal maintenance, and regener-

ation of skeletal muscle in vertebrates are governed by a complex

transcriptional regulatory circuit (Buckingham et al., 2006;

Charge and Rudnicki, 2004). A key player in myogenesis is the

transcription factor MyoD, which is sufficient to transdifferentiate

many types of cells to muscle cells (Choi et al., 1990; Morosetti

et al., 2006; Weintraub et al., 1989, 1991), and is sometimes

referred to as a ‘‘master regulator’’ of skeletal muscle differenti-

ation (Berkes and Tapscott, 2005; Tapscott, 2005). A key chal-

lenge has been to determine what regulates this ‘‘master regula-

tor’’ at the transcriptional level. Distal DNA elements have been

shown to regulate myod transcription during embryonic muscle

differentiation by genetic studies (Asakura et al., 1995; Chen

et al., 2002; Goldhamer et al., 1995; Kucharczuk et al., 1999).

However, the molecular mechanism and key transcription fac-

534 Developmental Cell 15, 534–546, October 14, 2008 ª2008 Else

tors directing the activity of these enhancers remains largely

unknown. Although many transcription factors including SRF,

Sp1, YY1, and p300/CBP have been implicated in affecting

myod transcription (Gauthier-Rouviere et al., 1996; L’Honore

et al., 2003; Roth et al., 2003; Wilson and Rotwein, 2006), none

of these are thought to determine the cell type-specific expres-

sion of MyoD.

FoxOs belong to the large forkhead family of transcription

factors that function to integrate growth signals into diverse tran-

scriptional networks governing a wide range of physiological

processes including proliferation, differentiation, survival, and

metabolism (Accili and Arden, 2004; Carter and Brunet, 2007).

Like all forkhead factors, FoxOs share a conserved DNA binding

domain responsible for recognizing consensus forkhead regula-

tory elements (FRE). FoxOs are also thought to interact with

various activator/repressor partners to regulate their activities

(Gomis et al., 2006). Invertebrates typically encode only one

FoxO, while mammals employ multiple FoxO paralogs: FoxO1,

FoxO3, FoxO4, and FoxO6 (Jacobs et al., 2003; Lam et al.,

2006; van der Heide et al., 2005). These FoxOs all recognize

and bind similar DNA sequences (Furuyama et al., 2000) and

are thus subject to functional redundancy under certain circum-

stances (Paik et al., 2007; Tothova et al., 2007). However, null

mutations in FoxO1, 3, and 4 produce distinct phenotypes in

mice (Castrillon et al., 2003; Hosaka et al., 2004; Jonsson

et al., 2005; Lin et al., 2004), suggesting specific functions of

FoxOs during development. Even for FoxOs involved in the

same cellular process, there seems to be selective utilization of

individual factors under physiological conditions (Paik et al.,

2007). Thus, diversified functions for individual FoxOs and their

mechanisms of specificity remain challenging but important

questions to address. Although some FoxOs have been impli-

cated in muscle differentiation and maintenance (Bois and Gros-

veld, 2003; Hribal et al., 2003; Kitamura et al., 2007; Li et al.,

2007; Liu et al., 2005; Machida et al., 2003; Mammucari et al.,

2007; Sandri et al., 2004), whether this process is regulated by

one particular FoxO or several in combination remains unclear.

Another class of transcription factors, the Pax proteins (paired

and homeodomain containing), is essential for regulating embry-

onic organogenesis and differentiation in metazoans (Lang et al.,

2007). The closely related Pax3 and Pax7 are specifically ex-

pressed in the central nervous system as well as skeletal muscle,

vier Inc.

Developmental Cell

Activators Direct MyoD Transcription

and share some overlapping functions (Buckingham and Relaix,

2007). During embryonic myogenesis, Pax3 and 7 are expressed

exclusively in cells destined to become skeletal muscle (Kassar-

Duchossoy et al., 2005) and have important roles in regulating

expression of myogenic transcription factors (Bajard et al.,

2006; McKinnell et al., 2008). Genetic studies suggest that

Pax3 and Pax7 are potential upstream regulators of myod during

both embryonic and postnatal myogenesis (Bober et al., 1994;

Goulding et al., 1994; Maroto et al., 1997; Relaix et al., 2005;

Seale et al., 2000). However, the direct activation of myod

Figure 1. MyoD Is a Potential FoxO3 Target

in Myoblasts

(A) Schematic representation of FoxO1, 3, and 4.

The conserved forkhead DNA-binding domain is

indicated in yellow. The red bars illustrate the po-

sitions of the shRNA targets. The numbers below

the sequence indicate percentage of identities be-

tween FoxO1, 3, and 4 in CLUSTALW.

(B) Whole-cell lysates of C2C12 cells treated with

FoxO1, 3, or 4 shRNA were analyzed by immuno-

blot with antibodies specific to individual FoxOs.

Immunoblot against GAPDH serves as the loading

control. The percentages of mRNA and proteins

knocked down are indicated at right.

(C) The examples of genes suppressed by FoxO

RNAi. Logarithm of fold reduction in shRNA

treated cells is illustrated in the x axis.

transcription by Pax3 and Pax7 has not

been demonstrated. The importance

of a possible link between FoxOs and

Pax3/7 is underscored by the finding

of a naturally occurring chromosomal

translocation between FoxO1 and Pax3/7

that results in a Pax3/7-FKHR fusion pro-

tein in human alveolar rhabdomyosarco-

mas (Mercado and Barr, 2007). However,

neither a physical nor functional link

between FoxOs and Pax3/Pax7 in non-

transformed cells has been demon-

strated.

Employing a combination of in vitro

biochemistry, molecular genetic loss/

gain of function, and in vivo muscle re-

generation experiments, we have tested

the hypothesis that specific FoxO factors

play a direct role in regulating myogene-

sis. We also identify tissue-specific Pax

activator partners that work in conjunc-

tion with FoxO to serve as key regulators

required to synergistically activate myod

transcription in myoblasts. The identifica-

tion of FoxO3 as a myod transcription ac-

tivator reveals a potential link between

signaling cascades and transcription reg-

ulation of myod in adult muscles. This

connection between FoxO and Pax tran-

scription factors provides muscle cells

with a powerful mechanism to utilize

codependent activators expressed in multiple tissues to direct

cell type-specific transcription.

RESULTS

Expression of MyoD Is Regulated by FoxO3In C2C12 myoblasts, the presence of three FoxO members FoxO1,

FoxO3, and FoxO4 can be detected by RT-PCR and immunoblots.

All three FoxOs share the conserved Forkhead DNA binding do-

main, but each bears different transactivation domains (Figure 1A).

Developmental Cell 15, 534–546, October 14, 2008 ª2008 Elsevier Inc. 535

Developmental Cell

Activators Direct MyoD Transcription

Small hairpin RNA (shRNA) against each specific foxo gene (Fig-

ure 1A, red bars) was incorporated into C2C12 cells to generate

three distinct stable RNA interference (RNAi) lines. Each targeted

FoxO protein was efficiently depleted as determined by immuno-

blots and qRT-PCR (Figure 1B), while expression levels of control

genes (GAPDH) and other nontargeted FoxOs remained largely

unchanged afterFoxOshRNAtreatment (Figure 1B; seeFigureS1

available online).

In order to identify putative target genes for each FoxO protein

in an unbiased manner, we compared the gene expression pro-

files of control RNAi and FoxO1, 3, or 4 RNAi lines by microarray

analysis. RNA prepared from each of the FoxO RNAi lines or

scrambled control RNAi cells were utilized as probe sets for

hybridization. Potential target genes activated by FoxOs, with

more than 2-fold decrease in expression levels were analyzed

(GSE12582). In addition to genes activated by multiple FoxOs,

potential target genes selectively activated by individual FoxOs

were also found. Representative FoxO activated genes are

shown in Figure 1C, and repressed genes are shown in Table

S1. Among these putative target genes, we confirmed previously

identified mammalian FoxO targets such as p21 and insulin

receptor substrate (Bois and Grosveld, 2003; Puig et al., 2003;

Seoane et al., 2004), mammalian homologs of known Drosophila

FoxO targets, such as 4EBP (Marr et al., 2007; Puig et al., 2003),

and putative FoxO targets identified from a computational

analysis of FREs, such as SCAND1 (Xuan and Zhang, 2005). In-

terestingly, a series of genes involved in myogenic differentiation

including myod, myogenin, tropomyosin, and creatine kinase

were effectively downregulated by FoxO RNAi knockdown,

suggesting that FoxOs may play important, but previously unap-

preciated direct roles in muscle differentiation. An important

early marker in myogenic differentiation, MyoD, was specifically

downregulated by FoxO3 depletion (7.9-fold), but not by FoxO1

or FoxO4 RNAi. Among these myogenic regulatory genes, MyoD

is thought to be one of the most upstream regulators in the differ-

entiation cascade, raising the intriguing possibility that this early

marker is a bona fide direct target of FoxO3.

FoxO3 Binds the myod Promoter and ActivatesIts TranscriptionFoxOs are known to be phosphorylated by various kinases and

subsequently translocated into the cytoplasm (Arden, 2006;

Huang andTindall, 2007; Lam etal., 2006). We thereforeexamined

the subcellular localization of FoxO3 in C2C12 myoblasts. Both

immunofluorescent staining of cells and immunoblots of cell ex-

tracts showed predominantly nuclear localization of FoxO3 in

myoblasts (Figure S2). These results suggest that FoxO3 is likely

to be in its active form and regulating transcription of target genes

in vivo.

To confirm the microarray analysis results, we performed

quantitative RT-PCR (qRT-PCR) to measure the relative MyoD

mRNA abundance using U6 small RNA as an internal control in

C2C12 cells treated with either scrambled control or specific

shRNA against each of the three foxo genes. MyoD mRNA levels

were greatly reduced (�85%) in FoxO3 shRNA-treated cells, but

not in scrambled shRNA or FoxO1 shRNA-treated cells, and only

modestly reduced (�30%) in FoxO4 shRNA-treated cells

(Figure 2B). When a CMV-based rescue system was used to ex-

press the coding sequence of a FoxO3 construct that is resistant

536 Developmental Cell 15, 534–546, October 14, 2008 ª2008 Elsev

to shRNA targeted to the 30UTR, nearly normal MyoD mRNA

levels were restored. In contrast, introducing empty vector

or FoxO1 rescue construct into RNAi cells depleted of FoxO3

failed to restore expression of MyoD. Introducing a FoxO4 res-

cue construct to FoxO3 RNAi cells only partially restored

MyoD expression (Figure 2B). These results indicate that the

loss of FoxO3 leads to strong downregulation of MyoD and

that re-expression of FoxO3 and to a lesser extent FoxO4 can

rescue MyoD expression.

We next tested whether overexpression of functional FoxO3

can activate MyoD expression. In order to increase active

FoxO3 concentration in nuclei, we utilized the constitutively ac-

tive FoxO3A3 mutant wherein three Ser/Thr residues have

been changed to Ala thus circumventing the phosphorylation-

dependent cytoplasmic sequestration. When transiently trans-

fected into C2C12 cells, the FoxO3A3 constitutive mutant

strongly stimulated MyoD expression (25-fold), while empty

vector and FoxO1A3 showed no stimulation of MyoD expression

as expected. The FoxO4A3 mutant displayed weak upregulation

of myod transcription (Figure 2A). These gain-of-function studies

taken together with our loss-of-function assays strongly point to

FoxO3 as a primary potent activator of MyoD expression, while

FoxO4 functions partly overlap FoxO3 and modestly activates

myod transcription.

Consistent with this notion, sequence analysis revealed four

potential FRE sites in a 6 kb region upstream of the myod tran-

scription initiation site (Figure 2B), which are between the DRR

and PRR elements important for MyoD regulation during embry-

onic muscle development (Asakura et al., 1995). We next per-

formed ChIP to determine promoter occupancy of FoxOs at

the myod gene using antibodies specifically directed against

each of the three proteins. FoxO3 was efficiently detected at

the myod promoter regions overlapping putative FRE �940

and �1598, but not at �1928 and �2351 (Figure 2B). In marked

contrast to FoxO3, neither FoxO1 nor FoxO4 was detected sig-

nificantly above background at the four potential FREs in C2C12

cells (Figure 2B). As a control, the same FoxO1 and FoxO4 anti-

bodies were used to successfully detect their occupancy at the

p21 promoter (Figure S3A), which is known to be regulated by

these FoxOs (Figure 1C; Seoane et al., 2004). Indeed, it appears

that FoxO3 selectively binds two of the putative FREs at the

myod promoter in myoblasts, suggesting that FoxO3 binding

may correspond to one of the important steps regulating MyoD

expression. RNA polymerase II was also found at the myod

promoter by ChIP (Figure S3B), further confirming the correlation

between FoxO3 promoter occupancy and transcription activa-

tion. Although FoxO4 appeared to have low transcriptional activ-

ity, little, if any, FoxO4 was detected at the myod promoter

by ChIP in myoblasts (Figure 2B). Taken together, these results

suggest that at least FoxO3 is likely part of the active transcrip-

tion machinery directly recruited to the myod promoter in

myoblasts.

Since FoxO3 binds to putative FRE �940 and FRE �1598 of

the myod promoter in C2C12 cells, we next examined whether

FoxO3 is also able to recognize and bind these putative FREs

in vitro by electrophoretic mobility shift assays (EMSA). Purified

recombinant FoxO3 efficiently bound to both the FRE �940

and FRE �1598 probes, but not to mutant probes (Figure 2E),

confirming that FoxO3 can bind to both FRE sequences

ier Inc.

Developmental Cell

Activators Direct MyoD Transcription

specifically in vitro. Consistent with our ChIP results, in luciferase

reporter assays, mutations in the FREs at �940 or �1598 signif-

icantly debilitated myod transcription, while mutations in

FRE�1928 and FRE�2351 had little or no effect on transcription

activity (Figure S4). These results suggest that FRE �940

and �1598 bound by FoxO3 are critical for myod transcription

activation. Taken together, these assays indicate that FoxO3

can specifically bind two FREs in the myod promoter and acti-

vate transcription.

Figure 2. FoxO3 Binds the myod Promoter

and Activates Transcription

(A) qRT-PCR analysis of MyoD mRNA level in cells

treated with shRNA against each FoxO protein as

indicated below each yellow column. The pink col-

umns indicate qRT-PCR analysis of MyoD mRNA

levels in FoxO3 RNAi cells rescued by CMV driven

FoxO constructs as indicated below each column.

The purple columns represent qRT-PCR analysis

of MyoD mRNA levels in cells overexpressing

constitutively active FoxOs as indicated below

each column. The error bars represent values

from standard deviation calculations.

(B) The upper panel is a schematic diagram of the

myod promoter structure. Primers utilized for ChIP

are indicated by orange arrows. ChIP results with

antibodies against each FoxO protein are in the

lower panel.

(C) EMSAs were performed with probe encom-

passing either wild-type or mutant FRE. Recombi-

nant FoxO3 expressed in insect cells binds specif-

ically to the wild-type probe.

FoxO3 and MyoD Can RescueDifferentiation Defectsin FoxO3-Depleted CellsBased on the above results, we proposed

that FoxO3 is responsible for activating

MyoD expression under physiologically

relevant conditions. If this hypothesis

has merit, we might expect that depletion

of FoxO3 will disrupt myotube formation.

Indeed, FoxO3 RNAi cells failed to differ-

entiate into myotubes, whereas the

scrambled control and FoxO1 RNAi cells

differentiated normally (Figure 3A). When

a CMV-driven FoxO1, 3, or 4 construct

was introduced into the FoxO3-depleted

cells, only FoxO3 significantly rescued

the differentiation defect (Figure 3B). If

the FoxO3 rescue was achieved in large

measure by restoring MyoD expression,

re-expressing MyoD in these FoxO3

depleted cells should also at least partly

rescue the differentiation defects. As

expected, expression of CMV-driven

MyoD in FoxO3 RNAi cells partially res-

cued myotube formation (Figure 3B).

This partial rescue may be in part due to

inappropriate and unregulated levels of

FoxO3 and MyoD ectopically driven by

the highly active CMV promoter. Taken together, these observa-

tions establish that FoxO3, but not FoxO1 or 4, plays an impor-

tant role in activating myod transcription in cells.

Pax3/7 Binds the myod Promoter and ActivatesTranscriptionAlthough we have identified FoxO3 as an important activator of

myod transcription, FoxO3 is ubiquitously expressed in most

cell types (Anderson et al., 1998; Biggs et al., 2001), while

Developmental Cell 15, 534–546, October 14, 2008 ª2008 Elsevier Inc. 537

Developmental Cell

Activators Direct MyoD Transcription

MyoD expression is limited to muscle cells. This suggests that

tissue specific factors in addition to FoxO3 may be required

to activate myod transcription. Therefore, we performed a series

of chromatographic separations to identify potential partners

that may confer tissue selectivity. C2C12 nuclear extracts

were subjected to ammonium sulfate precipitation followed

by monoQ chromatography and DNA affinity chromatography

with synthetic DNA containing three tandem repeats of

FRE�940 linked to three tandem repeats of FRE�1598. Protein

complexes eluted from the DNA affinity resin were subsequently

analyzed by liquid chromatography tandem mass spectrometry

(LC MS/MS) (Figure 4A). The mass spec results revealed Pax3 or

Pax7 as transcription factors present in the DNA affinity column

eluents in addition to FoxO3. Based on the peptides identified

by LC MS/MS, we could not discriminate Pax3 from the closely

related Pax7, which shares 77% sequence identity to Pax3.

To test whether Pax3/7 has any effect on MyoD expression

in C2C12 cells, we carried out qRT-PCR in cells either lacking

or overexpressing Pax3/7. In cells treated with shRNAs against

Pax3 and Pax7, nearly 95% of the Pax3 and 87% of the

Pax7 mRNA was depleted, indicating an efficient knockdown.

In these same Pax3/7-depleted cells, MyoD mRNA was concom-

itantly downregulated more than 80% (Figure 4B, upper panel).

By contrast, in cells transiently overexpressing Pax3 or Pax7,

production of MyoD was specifically upregulated, while in

both cases FoxO3 mRNA levels remained unaffected (Figure 4B).

Previous mouse genetic studies had implicated Pax3 and

Pax7 as potential upstream activators of MyoD expression

(Bober et al., 1994; Goulding et al., 1994; Seale et al., 2000),

but it had not been shown whether these factors directly regulate

Figure 3. FoxO3 and MyoD Can Rescue

Differentiation Defects in FoxO3-Depleted

C2C12 Cells

(A) Phase contract images of C2C12 cells treated

with control or FoxO3 shRNA before and after

differentiation.

(B) Phase contract images of FoxO3 depleted

C2C12 cells rescued by CMV driven FoxO1, 3, 4,

or MyoD before and after differentiation. The red

arrows indicate myotubes. Bar, 330 mm.

myod transcription. Consistent with

these genetic data, our loss-of-function

and gain-of-function assays targeting

myod transcription suggest that Pax3/7

together with FoxO3 contributes to the

transcription activation of myod in myo-

blasts. Promoter DNA sequence analysis

revealed two potential paired boxes lo-

cated between FRE �940 and FRE

�1598 (Figure 4C). To test for direct

Pax3/7 binding to these sites in C2C12

cells, we performed ChIP experiments

with an antibody recognizing both Pax

proteins. Pax3/7 was significantly en-

riched at the myod promoter region

containing paired box �1502, but not at

the region containing paired box �989

(Figure 4C). The pattern of a Pax recognition site (paired box or

homeobox) lying between two FREs are conserved from rodent

to human (Figure S5). As expected, control (c2) primers mapping

to a region 50 kb upstream of the myod transcription initiation

site did not yield any amplification products above background

(Figure 4C). These data suggest that Pax3/7 binds to paired

box �1502 of the myod promoter in C2C12 cells.

To further examine the binding of Pax3/7 protein to this putative

paired box at �1502, we carried out EMSA. Recombinant Pax3

and Pax7 were found to bind the paired-box probe, but not a mu-

tant probe, although Pax3 binds with significantly higher affinity

than Pax7 at least in vitro (Figure S6A). Consistent with our

ChIP and RNAi results, paired-box mutant promoters displayed

lower activities in luciferase reporter assays (Figure S6B), sug-

gesting that the paired box is required for myod transcription

activation. Together, these cell-based and in vitro studies sug-

gest that Pax3/7 may function as a FoxO3 partner in regulating

myod transcription in myoblasts.

To test for the possible Pax3/7 and FoxO3 co-occupancy at

the endogenous myod promoter, sequential ChIP assays were

performed (Figure 4E). Consistent with the data shown in

Figure 2B, FoxO3 was found at the myod promoter after the first

anti-FoxO3 IP. Intriguingly, Pax3/7 was detected at the same

promoter region containing FRE �1928 and paired box �1502

after the second anti-Pax3/7 IP (Figure 4E), indicating that

FoxO3 and Pax3/7 are able to occupy the myod promoter simul-

taneously. Consistent with these sequential ChIP results, FoxO3

binds Pax3 and Pax7 with higher affinity than FoxO1 and FoxO4

in GST pull-down experiments (Figure S7A), suggesting a spe-

cific protein-protein interaction between FoxO3 and Pax3/7.

538 Developmental Cell 15, 534–546, October 14, 2008 ª2008 Elsevier Inc.

Developmental Cell

Activators Direct MyoD Transcription

FoxO3 and Pax3/7 Activate myod TranscriptionCooperatively and Are Mutually DependentIn order to directly test the functions of FoxO3 and Pax3/7 in

myod transcription, we next set up an in vitro transcription sys-

tem programmed with a DNA template containing the CAT

gene driven by a fragment encompassing �1 to �2000 region

of the myod promoter. Low levels of transcription from this tem-

plate were detected with crude nuclear extracts derived from

proliferating C2C12 cells. We next supplemented these in vitro

transcription reactions with purified recombinant activators. Ad-

dition of recombinant FoxO3 to the in vitro myod transcription

system potently activated transcription (4.7- and 8.7-fold; Fig-

ure 5A). Likewise, Pax3 alone modestly activated myod tran-

scription (2.3- and 4-fold; Figure 5A). Surprisingly, although

Pax7 showed lower affinity for the paired-box DNA than Pax3

in EMSA, it nevertheless showed stronger transcription activa-

tion properties compared to that of Pax3 in the in vitro transcrip-

tion system (6.8- and 14.9-fold; Figure 5A). When combined,

FoxO3 and Pax3 displayed a high degree of synergy for activat-

ing myod transcription in vitro (8- and 21-fold; Figure 5A). Simi-

larly, FoxO3 and Pax7 activated myod transcription coopera-

Figure 4. Pax3/7 Binds the myod Promoter

and Activates Transcription

(A) Purification scheme of FoxO3-associated

activator partners.

(B) The upper panel illustrates qRT-PCR analysis

of Pax3, Pax7, MyoD, and FoxO3 mRNA levels in

C2C12 cells treated with control shRNA (yellow

columns) or shRNA against Pax3 and Pax7 (purple

columns) as indicated below each group of col-

umns. Lower panel illustrates qRT-PCR analysis

of mRNA levels of MyoD and FoxO3 in C2C12 cells

transiently overexpressing vector (green col-

umns), Pax3 (orange columns), or Pax7 (red

columns) as indicated below each group of col-

umns. The error bars represent values from stan-

dard deviation calculations.

(C) The two putative paired-box elements in the

myod promoter are indicated in the upper panel.

Primer pairs utilized in the anti-Pax3/7 ChIP

assays are indicated as yellow arrows. p1 and p2

map paired-box elements as indicated. c1 map

a region 50 kb upstream of the myod transcription

initiation site.

(D) The left panel shows the scheme of the FoxO3-

Pax3/7 sequential ChIP. The results are illustrated

in the right panel.

tively (24-fold; Figure 5A, lanes 10–16).

Consistent with our in vitro transcription

results, FoxO3 and Pax7 display cooper-

ative binding in EMSA on the myod pro-

moter (Figure S7B). Thus, only FoxO3 in

combination with Pax3/7 behave as effi-

cient activators that can synergistically

potentiate myod transcription in vitro.

It has been reported that Pax3-FKHR,

but not Pax3, strongly induced many

myogenic genes including myod upon

transfection into NIH 3T3 cells (Khan

et al., 1999). Recent studies in mouse ES cells also suggest

that Pax3 is not sufficient for full MyoD activation. As expected,

when transfected into NIH 3T3 cells or D3 ES cells, the combina-

tion of Pax3/7 and FoxO3 activated myod transcription robustly,

while Pax3 or FoxO3 alone barely activated myod transcription

(Figure S8). These results are consistent with the observation

from our in vitro studies suggesting cooperative activation of

myod transcription by Pax3/7 and FoxO3.

We next used a double-template transcription system to fur-

ther confirm the functions of FoxO3 and Pax3/7. In these reac-

tions, both wild-type and mutant myod promoters are present

in the same transcription reaction. First, we examined transcrip-

tion from the myod promoter containing a mutant paired box. As

expected, supplementing the system with recombinant Pax3/7

failed to activate transcription from the paired-box mutant tem-

plate. Surprisingly, addition of FoxO3 was also insufficient to

activate transcription from the paired-box mutant promoter.

However, transcription directed by the wild-type myod promoter

in the same reaction was not affected (Figure 5B). These obser-

vations suggest that Pax proteins are important, probably

obligate, partners for FoxO3 to direct myod transcriptional

Developmental Cell 15, 534–546, October 14, 2008 ª2008 Elsevier Inc. 539

Developmental Cell

Activators Direct MyoD Transcription

Figure 5. FoxO3 and Pax3/7 Activate myod Transcription Cooperatively and Mutually Depend on Each Other

(A) In vitro transcription from myod promoter driven CAT template was performed with nuclear extracts from proliferating C2C12 myoblasts supplemented with

recombinant FoxO3, Pax7, or Pax3. The amount of proteins utilized in each reaction is indicated above each lane. In order to see cooperative activation by FoxO3

and Pax7, less protein was used as indicated in lanes 11–16. The fold of transcription activation by the activators is illustrated under each lane.

(B) Codependence between FoxO3 and Pax3/7 to activate myod transcription. Templates containing both wild-type and mutant myod promoter were supple-

mented with proliferating C2C12 myoblast nuclear extracts. The transcription product from the wild-type promoter is 15 nt longer than the one from the mutant

promoter. The transcripts from the mutant template are indicated in the upper panel. The transcripts from the wild-type promoter are indicated in the lower panel.

Lanes 1–7 indicate in vitro transcription directed by paired-box mutant promoter. Lanes 8–14 indicate in vitro transcription directed by myod promoter with both

FRE �940 and �1598 mutated.

(C) qPCR analysis of enrichment of the myod or b-actin promoter regions in anti-Pax3/7 or RNA polymerase II phospho-Ser5 ChIP as indicated below each group

of columns in cells treated with FoxO3 or control shRNAi as indicated by blue or yellow columns, respectively. The error bars represent values from standard

deviation calculations.

(D) qPCR analysis of enrichment of the myod or b-actin promoter in anti-FoxO3 or RNA polymerase II phospho-Ser5 ChIP as indicated below each group of

columns in cells treated with Pax3/7 or control shRNAi as indicated by purple or yellow columns, respectively. The error bars represent values from standard

deviation calculations.

activation. We also tested transcription from a myod promoter

containing mutant FREs in the double template system. As ex-

pected, FoxO3 failed to activate transcription from this mutant

FRE template. Interestingly, both Pax3 and Pax7 also failed to

activate transcription from the mutant FRE promoter. FoxO3

and Pax3/7 dependent transcription from the wild-type promoter

in the same reaction was not affected (Figure 5B). These in vitro

studies taken en toto suggest that FoxO3 is necessary for Pax3/7

dependent myod transcription activation.

Having established that FoxO3 and Pax3/7 can activate tran-

scription of myod in vitro, we next investigated potential inter-

540 Developmental Cell 15, 534–546, October 14, 2008 ª2008 Else

actions between FoxO3 and Pax3/7 in C2C12 cells. First, we

asked what happens to Pax3/7 binding at the myod promoter

in the absence of FoxO3. Indeed, ChIP experiments performed

in FoxO3-depleted cells revealed that neither Pax3/7 nor active

RNA polymerase II can be detected at the myod promoter in

the absence of FoxO3. As a control, we confirmed that active

RNA polymerase II is detected at the non-FoxO3-regulated

b-actin promoter (Figure 5C). Next, we carried out the converse

experiments using C2C12 cells depleted of both Pax3 and

Pax7. In this case, the amount of FoxO3 detected at FRE

�1598 and �940 was reduced to about 50% compared to

vier Inc.

Developmental Cell

Activators Direct MyoD Transcription

that in control RNAi cells. Curiously, we also detected some

FoxO4 in addition to FoxO3 at FRE �1598 in the absence of

Pax3/7, while FRE �940 was only recognized by FoxO3. Impor-

tantly, little, if any, RNA polymerase II was recruited to the

myod promoter, whereas its recruitment to the b-actin gene

was not affected (Figure 5D). Consistent with our in vitro tran-

scription results (Figure 5B), these ChIP experiments suggest

a mutual codependence between FoxO3 and Pax3/7 for estab-

lishing an active PIC. Apparently, only when both activators are

present is RNA polymerase II recruited and myod transcription

activated.

FoxO3 and Pax3/7 Both Bind the myod Promoterin Satellite CellsThe observations described thus far established FoxO3 and

Pax3/7 as codependent activators for myod transcription in

C2C12 cells. To determine whether this situation also occurs in

primary cells, we next carried out experiments using satellite

Figure 6. FoxO3 and Pax3/7 Bind the myod

Promoter in Satellite Cells

(A) Phase contract images of undifferentiated and

differentiated satellite cells. Bar, 80 mm.

(B) Immunofluorescent staining images of FoxO3

and Pax7 in undifferentiated satellite cells and

fibroblasts isolated from the same mice. Bar,

10 mm.

(C) ChIP performed in satellite cells and primary fi-

broblasts isolated from the same mice. F1 primers

(Figure 2B) were used to map FRE�940. F2 primers

(Figure 2B) were used to map FRE �1598. p1

primers (Figure 4C) were used to map the paired

box �1502. The PCR reactions are performed

with 32P-labeled primers.

cells (muscle stem cells), which are

responsible for postnatal growth and

muscle regeneration, isolated from new-

born mouse skeletal muscle (Collins,

2006; Conboy et al., 2005; Wagers and

Conboy, 2005), wherein MyoD is highly

expressed to confirm the presence of

FoxO3 and Pax3/7 at the myod promoter.

In vitro differentiation experiments con-

firmed that over 90% of the cells isolated

were indeed satellite cells (Figure 6A).

FoxO3 and Pax3/7 were both easily de-

tected in the nuclei of these cells. In

marked contrast, FoxO3 was diffusely

distributed in both nuclei and cytoplasm

of primary fibroblasts isolated from the

same mice, while Pax3/7 was not de-

tected in these nonmuscle cells (Fig-

ure 6B). ChIP experiments confirmed

the presence of FoxO3 at FRE �1598

and �940, while Pax3/7 occupied the

paired box at the myod promoter in satel-

lite cells, but not in primary fibroblasts

(Figure 6C). These experiments establish

that FoxO3 and Pax3/7 can specifically

target the myod promoter in primary cells and suggest that

FoxO3 and Pax3/7 likely operate together to activate myod

transcription in vivo.

FoxO3 Null Mice Display Muscle Regeneration DefectsThe results above suggest that FoxO3 and Pax3/7 directly acti-

vate myod transcription in both C2C12 and primary cells. If this

hypothesis has merit, we might expect that both Pax3/7 and

FoxO3 knockout mice will have lower MyoD expression levels

and display some degree of muscle defects. Consistent with

our hypothesis, satellite cells from Pax7 null mice showed

reduced differentiation potential (Seale et al., 2000) while Pax3

mutant mice lacked skeletal muscle in limbs (Bober et al.,

1994; Goulding et al., 1994). However, it was not known whether

FoxO3 knockout mice display any muscle defects. We anticipate

that if FoxO3�/� animals show any muscle defects, they might

resemble MyoD�/� mice wherein knockout mice do not display

any dramatic muscle phenotypes largely due to compensatory

Developmental Cell 15, 534–546, October 14, 2008 ª2008 Elsevier Inc. 541

Developmental Cell

Activators Direct MyoD Transcription

effects by Myf5 (Rudnicki et al., 1993). However, satellite cells

isolated from these knockout mice showed distinctly reduced

potential to form myotubes, thus revealing a muscle regenera-

tion defect after injury (Cornelison et al., 2000; Sabourin et al.,

1999; Yablonka-Reuveni et al., 1999).

We therefore set out to characterize the differentiation poten-

tial of satellite cells isolated from FoxO3 null mice (Castrillon

et al., 2003). Based on immunostaining of satellite cell markers

c-Met and Pax7, over 85% of the cells isolated from FoxO3

+/+, +/�, and�/�mice were satellite cells (Figure S9). The mor-

phologies of the satellite cells from all three sources appeared to

be similar in proliferation medium, suggesting self-renewal of

these cells remained largely unaffected. After differentiation,

myosin expression and multinucleate myotubes were easily de-

tected in wild-type cells. By contrast, only a small percentage of

FoxO3 heterozygous or homozygous satellite cells expressed

myosin and these cells formed much smaller myotubes (mea-

sured by the number of nuclei per cell) upon differentiation

(Figure 7A). These observations suggest a reduced differentia-

tion potential of satellite cells isolated from FoxO3 deficient

mice. We next examined the expression level of MyoD and

myogenin in these cells. RNA isolated from similar numbers of

FoxO3 +/+, +/�, or �/� cells were subjected to qRT-PCR anal-

ysis. Compared to wild-type cells, MyoD expression levels were

significantly reduced in both heterozygous (27% of wild-type)

and homozygous (13.6% of wild-type) KO satellite cells, while

the transcript levels of pax7, pax3, foxo1, and foxo4 were largely

unaffected in all littermates (Figure 7B).

To further investigate the contribution of FoxO3 in regulating

MyoD in vivo, muscle injury and regeneration studies were car-

ried out by intramuscular injection of cardiotoxin (CTX). Two

slightly different methods of CTX injection were utilized in paral-

lel. One group of mice was injected with 25 ml of 42 mM CTX, the

other group with 100 ml of 10 mM CTX (Yan et al., 2003). Hind leg

skeletal muscle sections from each time point taken during the

muscle regeneration process in both groups were stained with

anti-laminin and DAPI. In order to obtain a representative regen-

eration profile for wild-type mice, the numbers of cells containing

center located nuclei were counted and plotted (Figure S10). At

day 5 after injection with 25 ml of CTX and at day 7 after injection

of the larger volume, the numbers of cells with centrally located

nuclei decreased dramatically compared to earlier time points

(Figure S10). This suggests that under our experimental condi-

tions, wild-type mice largely recovered from wounding at day

5�7 in the case of 25 ml injections or day 7�11 in the case of

100 ml injections. Next, we examined the regeneration status of

FoxO3�/� mice injected with CTX at day 5, 7 and day 12, 16.

In sharp contrast to the wild-type, FoxO3�/�mice showed signif-

icantly higher percentages of cells containing center located

nuclei (Figures 7C and 7D), characteristic of incomplete muscle

cell regeneration. These injury and reparation experiments reveal

a distinct delay in skeletal muscle regeneration in FoxO3�/�mice

compared to wild-type animals and are consistent with our pro-

posal that FoxO3 plays an important role in vivo in skeletal mus-

cle differentiation, most likely via activating myod transcription.

Taken together, these observations suggest that FoxO3 null

mice more or less phenocopy MyoD null mice with respect to

muscle defects and support the notion that FoxO3 is an impor-

tant activator required for MyoD expression in vivo.

542 Developmental Cell 15, 534–546, October 14, 2008 ª2008 Else

DISCUSSION

Although FoxOs have previously been implicated functioning in

muscle differentiation (Bois and Grosveld, 2003; Hribal et al.,

2003; Kitamura et al., 2007; Machida et al., 2003; Sandri et al.,

2004), skeletal muscle-specific genes directly targeted by FoxOs

had not been identified. Our findings indicate that FoxO3 (but not

FoxO1 or FoxO4) binds a subset of FREs in the myod promoter to

work in concert with Pax3/7 in regulating cell type-specific tran-

scription activation in myoblasts. The contribution of FoxO3 in

directing myod transcription activation in vivo was further con-

firmed by the observed muscle regeneration defects in FoxO3

null mice.

Identification of FoxO3 as an important myod transcription ac-

tivator may provide a handle to explore the potential signaling

pathways governing muscle regeneration. Interestingly, FoxO1

was found to negatively regulate MyoD expression indirectly

through the Delta-Notch pathway (Holterman et al., 2007; Kita-

mura et al., 2007). The potential repression of MyoD by FoxO1

together with our results of direct activation of MyoD by FoxO3

suggests an intriguing mechanism to fine tune MyoD expression.

Muscle differentiation may therefore utilize selected FoxOs in

partnership with Pax3/7 to integrate inputs from multiple signal-

ing pathways.

The apparent Kd of FoxO3 and Pax3/7 binding individually to

the DNA elements in the myod promoter is on the order of

10�7M in vitro (P.H., unpublished data), which is rather modest

compared to a typical DNA binding protein, such as GAL4 (ap-

parent Kd, �10�11M) (Kamachi et al., 2000; Parthun and Jaehn-

ing, 1990). This is consistent with our finding that neither FoxO3

nor Pax3/7 alone binds promoter DNA efficiently to form a stable

DNA-activator complex capable of recruiting a functional PIC. It

appears that to efficiently assemble an active PIC via FoxO3 and

Pax3/7 at the myod promoter both protein-DNA and protein-

protein interactions mediated by this hitherto unknown activator

partnership must take place to trigger transcription activation

synergistically. This may therefore represent a useful and effi-

cient combinatorial mechanism to direct cell type-specific tran-

scription while utilizing two activators shared by many cell types.

This codependence is reminiscent of the mechanism utilized by

Sox2 and Pax6 to drive lens-specific transcription of d-crystallin

gene (Kondoh et al., 2004; Lang et al., 2007; Lefebvre et al.,

2007).

Although under normal conditions FoxO3 is predominantly de-

tected occupying the FRE at �1598 of the myod promoter, curi-

ously, we found that some FoxO4 can be detected at this site in

myoblasts when Pax3/7 is depleted. It is known that Pax3/7 is

not expressed in myotubes, but MyoD expression persists in

myotubes. It will be interesting to survey the identity of FoxOs

binding to the myod promoter in myotubes versus myoblasts

and explore additional mechanisms involved in cell type-specific

transcription activation during later stages of myogenesis. In-

triguingly, we now know there is a switching of the core transcrip-

tion machinery from the canonical holo-TFIID to a TRF3/TAF3

complex during myoblasts differentiation to myotubes (Deato

and Tjian, 2007). While FoxO3 and Pax3/7 appear to work in

concert with TFIID at the MyoD promoter in myoblasts, it will be

important to identify the key activator(s) that function together

with the TRF3/TAF3 complex in myotubes.

vier Inc.

Developmental Cell

Activators Direct MyoD Transcription

Figure 7. FoxO3 Null Mice Display Muscle

Regeneration Defects

(A) Immunofluorescent staining images of myosin

in differentiated satellite cells isolated from +/+,

+/�, and �/�mice. The average number of nuclei

per cell is indicated next to the images. Bar,

330 mm.

(B) qRT-PCR analysis of MyoD, Pax3, Pax7,

FoxO1, and FoxO4 mRNA levels in satellite

cells isolated from +/+, +/�, �/� mice. The error

bars represent values from standard deviation

calculations.

(C) Immunohistology staining images of laminin

and DAPI with frozen skeletal muscle sections

from +/+ and �/� mice injected with CTX or

PBS. CTX was injected intramuscularly in the right

hind leg. PBS was injected intramuscularly in the

left hind leg of the same mouse. Laminin staining

was indicated by red color. Green color was artifi-

cially applied to DAPI staining. Representative

staining images of 25 ml CTX injection after 5 days

are illustrated in panels (a) and (b). Representative

staining images of 100 ml CTX injection after

12 days are illustrated in panels (c) and (d). Repre-

sentative staining images of PBS injection are

illustrated in panels (e) and (f). Bar, 30 mm.

(D) Percentages of cells containing center located

nuclei in skeletal muscles sections from wild-type

or FoxO3 null mice. Days of recovery after injury

and volumes of injections are indicated below

each bar.

Developmental Cell 15, 534–546, October 14, 2008 ª2008 Elsevier Inc. 543

Developmental Cell

Activators Direct MyoD Transcription

EXPERIMENTAL PROCEDURES

Constructs and Antibodies

The mouse myod promoter was PCR from genomic DNA isolated from C2C12

cells. Plasmids containing FoxO1 and FoxO3 were kind gifts from Dr. D. Accili

(Columbia University, New York, NY). Plasmid containing FoxO4 was a kind

gift from Dr. O. Puig (University of California, Berkeley, Berkeley, CA). Pax3

and Pax7 genes were cloned from cDNA pool made from C2C12 cells. DNA

encoding shRNA was synthesized and cloned into pSM vector (Paddison

et al., 2004).

Rabbit polyclonal antibodies were affinity purified with antigen. Pax3 and

Pax7 monoclonal antibody (for immunofluorescent staining) were from The

Developmental Studies Hybridoma Bank. Anti-MyoD was from BD biosci-

ences. Anti-Myosin was from Upstate Biotechnology. Pax polyclonal anti-

bodies (for ChIP), a-GAPDH, a-c-met, and a-RNA polymerase II Ser5, a-lam-

inin was from Abcam. FITC-anti-CD34 is from Beckman Coulter. a-TBP was

from Biodesign International.

Cell Culture, Transfection, RNAi, Differentiation, and Muscle Injury

C2C12 cells (ATCC) were maintained in DMEM (Sigma) supplemented with

10% fetal bovine serum (FBS) (Sigma) at 37�C with 5% CO2. C2C12 cells

were transfected with lipofectamin 2000 (Invitrogen). Satellite cells were iso-

lated from mice legs as described (Carlson and Conboy, 2007). In brief,

mice younger than 1 month were sacrificed, and muscle tissues were dis-

sected out. The dissected muscle tissues were digested with collagenase

(Roche) and dispase (Roche), triturated followed by multiple sedimentations,

washes, and filtration. They were grown in F10 nutrition mix (Invitrogen) sup-

plemented with 15% FBS, 10 ng/ml bFGF (Invitrogen), 1000 units of LIF (Milli-

pore) at 37�C, 5% CO2. The cells were passed for less than 20 passages. Both

C2C12 cells and satellite cells were differentiated in DMEM containing 2%

horse serum (Invitrogen). C2C12 cells were grown to 100% confluence before

differentiation.

For skeletal muscle regeneration experiments, 25 ml of 42 mM or 100 ml of

10 mM CTX dissolved in PBS was injected intramuscularly to the right hind

leg of 6-week-old +/+ or �/� mice. As control, PBS was injected to the left

leg. The mice were sacrificed at 1, 2, 3, 4, 5, 6, 7, 11, 12, 14, and 16 days,

and frozen sections were made from skeletal muscles.

shRNA sequences against FoxOs were obtained from http://codex.cshl.

edu/scripts/main.pl. Control RNAi construct was obtained from Open Biosys-

tems. After transfection, the cells were selected and maintained in complete

DMEM medium containing 2 mg/ml puromycin. siRNA against Pax3 and

Pax7 were obtained from Dharmacon and transfected into C2C12 cells using

Oligofectamine (Invitrogen). Sequences were listed in Supplemental Experi-

mental Procedures.

DNA Microarrays

Microarray analysis was performed utilizing Affymetrix murine 430 2.0 chips.

The data were analyzed by GCOS software (Affymetrix). All chips have compa-

rable background values. The change p value is set at higher than 0.999. Other

values were listed in the Supplemental Experimental Procedures. Promoters of

the potential target genes were searched for FREs [(G/A/C)T(C/A)AA(T/C)A(A/

C)] at http://rna.berkeley.edu/�siwu/dnascanner.htm.

qRT-PCR

RNA was isolated from cells with RNeasy mini kit (QIAGEN) followed by

reverse transcription using M-LV RT (Ambion). cDNA was used as template

for qPCR with SYBR green master mix (Bio-Rad), followed by analysis with

DDCt method from at least four replicas. The error bars represent values

from standard deviation calculations (see Supplemental Experimental Proce-

dures for formula).

Immunofluorescent and Immunohistology Staining

Cells were fixed with 4% formaldehyde and permeablized in 0.1% Triton

X-100, then blocked with 0.5% goat serum and incubated with primary

antibody. Alexa 594- or Alexa 488-conjugated secondary antibodies (Molecu-

lar Probes) were used. The DNA contents of cells were stained with 300 nM

DAPI (Molecular Probes). Actin was stained with Alexa 488-conjugated

phalloidin (Molecular Probes). Frozen sections were made by cryostat

544 Developmental Cell 15, 534–546, October 14, 2008 ª2008 Else

(Brighter Instrument). After fixed in acetone for 10min at 4�C, the sections

were stained. Samples were visualized by 510 Meta Confocal Microscope

(Zeiss) or AxioImager M1 fluorescence microscope (Zeiss).

ChIP

Proliferating C2C12, or satellite cells, or C2C12 cells depleted FoxO3/Pax3/7

were crosslinked and processed as described in (Deato and Tjian, 2007)

except that wash buffer 2 contains 1 M LiCl.

EMSA

FoxO3 EMSA was performed as described in Puig and Tjian (2005). Pax3/7

EMSA was performed in the same buffer as FoxO3 in addition of 500 ng calf

thymus DNA (GE Healthcare) and 2 mM spermidine (Sigma) and ran on 5%

polyacrylamide gel at room temperature. Probe sequences were listed in

Supplemental Experimental Procedures.

In Vitro Transcription

Proliferating C2C12 nuclear extracts were used to transcribe from myod

promoter driven CAT in vitro. Transcription analysis was performed by primer

extension as described (Puig et al., 2003).

Chromatography

Proliferating C2C12 nuclear extracts were subject to 30%�40% ammonium

sulfate precipitation, followed by Poros-HQ 20 column (Applied Biosystems)

and eluted by linear KCl gradient in Buffer D (25 mM HEPES [pH 7.9],

0.2 mM EDTA, 2 mM MgCl2, 10% glycerol, 1 mM DTT, and KCl). The fractions

containing FoxO3 were pooled and applied to DNA affinity column. The DNA

affinity chromatography was performed as described (Kadonaga and Tjian,

1986). DNA bait was oligomers of the basic oligo unit containing three repeats

of FRE �940 followed by three FRE �1598 ranging from 50 to 70 mers.

ACCESSION NUMBERS

The complete DNA microarray data were deposited in the Gene Expression

Omnibus (GEO) at NCBI (http://www.ncbi.nlm.nih.gov/geo/) with series acces-

sion number GSE12582.

SUPPLEMENTAL DATA

The Supplemental Data include ten figures, one table, and Supplemental Ex-

perimental Procedures and can be found with this article online at http://

www.developmentalcell.com/cgi/content/full/15/4/534/DC1/.

ACKNOWLEDGMENTS

We thank M. Haggart for technical assistance, A. Fisher for assistance in large

scale cell culture, D. Schichnes for help on microscope usage, and D. King and

L. Kohlstaedt for help on peptide synthesis and mass spec. We thank I. Con-

boy and M. Carlson for help on satellite cell isolation, M. Marr for help on in vitro

transcription assays, and B. Gan for help on isolation of primary myoblasts

from knockout mice. We also thank M. Deato, K. Wright, Y. Fong, W. Liu, R.

Colemen, F. Herrera, B. Guglielmi, E. Olson, M.E. Buckingham, and M. Rud-

nicki for critical reading of the manuscript. We thank members of the Tjian lab-

oratory for valuable suggestions and discussions. J.-H.P. is Damon Runyon

Fellows supported by the Damon Runyon Cancer Research Foundation.

R.T. is an investigator of the Howard Hughes Medical Institute and Director

of the Li Ka-Shing Center for Biomedical and Health Sciences.

Received: January 25, 2008

Revised: July 21, 2008

Accepted: August 29, 2008

Published: October 13, 2008

REFERENCES

Accili, D., and Arden, K.C. (2004). FoxOs at the crossroads of cellular metab-

olism, differentiation and transformation. Cell 117, 421–426.

vier Inc.

Developmental Cell

Activators Direct MyoD Transcription

Anderson, M.J., Viars, C.S., Czekay, S., Cavenee, W.K., and Arden, K.C.

(1998). Cloning and characterization of three human forkhead genes that com-

prise an FKHR-like gene subfamily. Genomics 47, 187–199.

Arden, K.C. (2006). Multiple roles of FOXO transcription factors in mammalian

cells point to multiple roles in cancer. Exp. Gerontol. 41, 709–717.

Asakura, A., Lyons, G.E., and Tapscott, S.J. (1995). The regulation of MyoD

gene expression: conserved elements mediate expression in embryonic axial

muscle. Dev. Biol. 171, 386–398.

Bajard, L., Relaix, F., Lagha, M., Rocancourt, D., Daubas, P., and Buckingham,

M.E. (2006). A novel genetic hierarchy functions during hypaxial myogenesis:

Pax3 directly activates Myf5 in muscle progenitor cells in the limb. Genes Dev.

20, 2450–2464.

Berkes, C.A., and Tapscott, S.J. (2005). MyoD and the transcriptional control

of myogenesis. Semin. Cell Dev. Biol. 16, 585–595.

Biggs, W.H., 3rd, Cavenee, W.K., and Arden, K.C. (2001). Identification and

characterization of members of the FKHR (FOX O) subclass of winged-helix

transcription factors in the mouse. Mamm. Genome 12, 416–425.

Bober, E., Franz, T., Arnold, H.H., Gruss, P., and Tremblay, P. (1994). Pax-3 is

required for the development of limb muscles: a possible role for the migration

of dermomyotomal muscle progenitor cells. Development 120, 603–612.

Bois, P.R., and Grosveld, G.C. (2003). FKHR (FOXO1a) is required for myotube

fusion of primary mouse myoblasts. EMBO J. 22, 1147–1157.

Buckingham, M., and Relaix, F. (2007). The role of Pax genes in the develop-

ment of tissues and organs: Pax3 and Pax7 regulate muscle progenitor cell

functions. Annu. Rev. Cell Dev. Biol. 23, 645–673.

Buckingham, M., Bajard, L., Daubas, P., Esner, M., Lagha, M., Relaix, F., and

Rocancourt, D. (2006). Myogenic progenitor cells in the mouse embryo are

marked by the expression of Pax3/7 genes that regulate their survival and

myogenic potential. Anat. Embryol. (Berl.) 211 (Suppl 1), 51–56.

Carlson, M.E., and Conboy, I.M. (2007). Loss of stem cell regenerative capac-

ity within aged niches. Aging Cell 6, 371–382.

Carter, M.E., and Brunet, A. (2007). FOXO transcription factors. Curr. Biol. 17,

R113–R114.

Castrillon, D.H., Miao, L., Kollipara, R., Horner, J.W., and DePinho, R.A. (2003).

Suppression of ovarian follicle activation in mice by the transcription factor

Foxo3a. Science 301, 215–218.

Charge, S.B., and Rudnicki, M.A. (2004). Cellular and molecular regulation of

muscle regeneration. Physiol. Rev. 84, 209–238.

Chen, J.C., Ramachandran, R., and Goldhamer, D.J. (2002). Essential and

redundant functions of the MyoD distal regulatory region revealed by targeted

mutagenesis. Dev. Biol. 245, 213–223.

Choi, J., Costa, M.L., Mermelstein, C.S., Chagas, C., Holtzer, S., and Holtzer,

H. (1990). MyoD converts primary dermal fibroblasts, chondroblasts, smooth

muscle, and retinal pigmented epithelial cells into striated mononucleated

myoblasts and multinucleated myotubes. Proc. Natl. Acad. Sci. USA 87,

7988–7992.

Collins, C.A. (2006). Satellite cell self-renewal. Curr. Opin. Pharmacol. 6, 301–

306.

Conboy, I.M., Conboy, M.J., Wagers, A.J., Girma, E.R., Weissman, I.L., and

Rando, T.A. (2005). Rejuvenation of aged progenitor cells by exposure to

a young systemic environment. Nature 433, 760–764.

Cornelison, D.D., Olwin, B.B., Rudnicki, M.A., and Wold, B.J. (2000).

MyoD(�/�) satellite cells in single-fiber culture are differentiation defective

and MRF4 deficient. Dev. Biol. 224, 122–137.

Deato, M.D., and Tjian, R. (2007). Switching of the core transcription machin-

ery during myogenesis. Genes Dev. 21, 2137–2149.

Furuyama, T., Nakazawa, T., Nakano, I., and Mori, N. (2000). Identification of

the differential distribution patterns of mRNAs and consensus binding

sequences for mouse DAF-16 homologues. Biochem. J. 349, 629–634.

Gauthier-Rouviere, C., Vandromme, M., Tuil, D., Lautredou, N., Morris, M.,

Soulez, M., Kahn, A., Fernandez, A., and Lamb, N. (1996). Expression and

activity of serum response factor is required for expression of the muscle-

Develop

determining factor MyoD in both dividing and differentiating mouse C2C12

myoblasts. Mol. Biol. Cell 7, 719–729.

Goldhamer, D.J., Brunk, B.P., Faerman, A., King, A., Shani, M., and Emerson,

C.P., Jr. (1995). Embryonic activation of the myoD gene is regulated by a highly

conserved distal control element. Development 121, 637–649.

Gomis, R.R., Alarcon, C., He, W., Wang, Q., Seoane, J., Lash, A., and Mas-

sague, J. (2006). A FoxO-Smad synexpression group in human keratinocytes.

Proc. Natl. Acad. Sci. USA 103, 12747–12752.

Goulding, M., Lumsden, A., and Paquette, A.J. (1994). Regulation of Pax-3

expression in the dermomyotome and its role in muscle development. Devel-

opment 120, 957–971.

Holterman, C.E., Le Grand, F., Kuang, S., Seale, P., and Rudnicki, M.A. (2007).

Megf10 regulates the progression of the satellite cell myogenic program.

J. Cell Biol. 179, 911–922.

Hosaka, T., Biggs, W.H., 3rd, Tieu, D., Boyer, A.D., Varki, N.M., Cavenee, W.K.,

and Arden, K.C. (2004). Disruption of forkhead transcription factor (FOXO)

family members in mice reveals their functional diversification. Proc. Natl.

Acad. Sci. USA 101, 2975–2980.

Hribal, M.L., Nakae, J., Kitamura, T., Shutter, J.R., and Accili, D. (2003).

Regulation of insulin-like growth factor-dependent myoblast differentiation

by Foxo forkhead transcription factors. J. Cell Biol. 162, 535–541.

Huang, H., and Tindall, D.J. (2007). Dynamic FoxO transcription factors. J. Cell

Sci. 120, 2479–2487.

Jacobs, F.M., van der Heide, L.P., Wijchers, P.J., Burbach, J.P., Hoekman,

M.F., and Smidt, M.P. (2003). FoxO6, a novel member of the FoxO class of

transcription factors with distinct shuttling dynamics. J. Biol. Chem. 278,

35959–35967.

Jonsson, H., Allen, P., and Peng, S.L. (2005). Inflammatory arthritis requires

Foxo3a to prevent Fas ligand-induced neutrophil apoptosis. Nat. Med. 11,

666–671.

Kadonaga, J.T., and Tjian, R. (1986). Affinity purification of sequence-specific

DNA binding proteins. Proc. Natl. Acad. Sci. USA 83, 5889–5893.

Kamachi, Y., Uchikawa, M., and Kondoh, H. (2000). Pairing SOX off: with part-

ners in the regulation of embryonic development. Trends Genet. 16, 182–187.

Kassar-Duchossoy, L., Giacone, E., Gayraud-Morel, B., Jory, A., Gomes, D.,

and Tajbakhsh, S. (2005). Pax3/Pax7 mark a novel population of primitive myo-

genic cells during development. Genes Dev. 19, 1426–1431.

Khan, J., Bittner, M.L., Saal, L.H., Teichmann, U., Azorsa, D.O., Gooden, G.C.,

Pavan, W.J., Trent, J.M., and Meltzer, P.S. (1999). cDNA microarrays detect

activation of a myogenic transcription program by the PAX3-FKHR fusion

oncogene. Proc. Natl. Acad. Sci. USA 96, 13264–13269.

Kitamura, T., Kitamura, Y.I., Funahashi, Y., Shawber, C.J., Castrillon, D.H.,

Kollipara, R., DePinho, R.A., Kitajewski, J., and Accili, D. (2007). A Foxo/Notch

pathway controls myogenic differentiation and fiber type specification. J. Clin.

Invest. 117, 2477–2485.

Kondoh, H., Uchikawa, M., and Kamachi, Y. (2004). Interplay of Pax6 and

SOX2 in lens development as a paradigm of genetic switch mechanisms for

cell differentiation. Int. J. Dev. Biol. 48, 819–827.

Kucharczuk, K.L., Love, C.M., Dougherty, N.M., and Goldhamer, D.J. (1999).

Fine-scale transgenic mapping of the MyoD core enhancer: MyoD is regulated

by distinct but overlapping mechanisms in myotomal and non-myotomal mus-

cle lineages. Development 126, 1957–1965.

L’Honore, A., Lamb, N.J., Vandromme, M., Turowski, P., Carnac, G., and Fer-

nandez, A. (2003). MyoD distal regulatory region contains an SRF binding

CArG element required for MyoD expression in skeletal myoblasts and during

muscle regeneration. Mol. Biol. Cell 14, 2151–2162.

Lam, E.W., Francis, R.E., and Petkovic, M. (2006). FOXO transcription factors:

key regulators of cell fate. Biochem. Soc. Trans. 34, 722–726.

Lang, D., Powell, S.K., Plummer, R.S., Young, K.P., and Ruggeri, B.A. (2007).

PAX genes: roles in development, pathophysiology, and cancer. Biochem.

Pharmacol. 73, 1–14.

Lefebvre, V., Dumitriu, B., Penzo-Mendez, A., Han, Y., and Pallavi, B. (2007).

Control of cell fate and differentiation by Sry-related high-mobility-group box

(Sox) transcription factors. Int. J. Biochem. Cell Biol. 39, 2195–2214.

mental Cell 15, 534–546, October 14, 2008 ª2008 Elsevier Inc. 545

Developmental Cell

Activators Direct MyoD Transcription

Li, H., Liang, J., Castrillon, D.H., DePinho, R.A., Olson, E.N., and Liu, Z.P.

(2007). FoxO4 regulates tumor necrosis factor alpha-directed smooth muscle

cell migration by activating matrix metalloproteinase 9 gene transcription. Mol.

Cell. Biol. 27, 2676–2686.

Lin, L., Hron, J.D., and Peng, S.L. (2004). Regulation of NF-kappaB, Th activa-

tion, and autoinflammation by the forkhead transcription factor Foxo3a. Immu-

nity 21, 203–213.

Liu, Z.P., Wang, Z., Yanagisawa, H., and Olson, E.N. (2005). Phenotypic mod-

ulation of smooth muscle cells through interaction of Foxo4 and myocardin.

Dev. Cell 9, 261–270.

Machida, S., Spangenburg, E.E., and Booth, F.W. (2003). Forkhead transcrip-

tion factor FoxO1 transduces insulin-like growth factor’s signal to p27Kip1 in

primary skeletal muscle satellite cells. J. Cell. Physiol. 196, 523–531.

Mammucari, C., Milan, G., Romanello, V., Masiero, E., Rudolf, R., Del Piccolo,

P., Burden, S.J., Di Lisi, R., Sandri, C., Zhao, J., et al. (2007). FoxO3 controls

autophagy in skeletal muscle in vivo. Cell Metab. 6, 458–471.

Maroto, M., Reshef, R., Munsterberg, A.E., Koester, S., Goulding, M., and

Lassar, A.B. (1997). Ectopic Pax-3 activates MyoD and Myf-5 expression in

embryonic mesoderm and neural tissue. Cell 89, 139–148.

Marr, M.T., 2nd, D’Alessio, J.A., Puig, O., and Tjian, R. (2007). IRES-mediated

functional coupling of transcription and translation amplifies insulin receptor

feedback. Genes Dev. 21, 175–183.

McKinnell, I.W., Ishibashi, J., Le Grand, F., Punch, V.G., Addicks, G.C., Green-

blatt, J.F., Dilworth, F.J., and Rudnicki, M.A. (2008). Pax7 activates myogenic

genes by recruitment of a histone methyltransferase complex. Nat. Cell Biol.

10, 77–84. Published online December 9, 2007. 10.1038/ncb1671.

Mercado, G.E., and Barr, F.G. (2007). Fusions involving PAX and FOX genes in

the molecular pathogenesis of alveolar rhabdomyosarcoma: recent advances.

Curr. Mol. Med. 7, 47–61.

Morosetti, R., Mirabella, M., Gliubizzi, C., Broccolini, A., De Angelis, L., Taglia-

fico, E., Sampaolesi, M., Gidaro, T., Papacci, M., Roncaglia, E., et al. (2006).

MyoD expression restores defective myogenic differentiation of human meso-

angioblasts from inclusion-body myositis muscle. Proc. Natl. Acad. Sci. USA

103, 16995–17000.

Paddison, P.J., Cleary, M., Silva, J.M., Chang, K., Sheth, N., Sachidanandam,

R., and Hannon, G.J. (2004). Cloning of short hairpin RNAs for gene knock-

down in mammalian cells. Nat. Methods 1, 163–167.

Paik, J.H., Kollipara, R., Chu, G., Ji, H., Xiao, Y., Ding, Z., Miao, L., Tothova, Z.,

Horner, J.W., Carrasco, D.R., et al. (2007). FoxOs are lineage-restricted redun-

dant tumor suppressors and regulate endothelial cell homeostasis. Cell 128,

309–323.

Parthun, M.R., and Jaehning, J.A. (1990). Purification and characterization of

the yeast transcriptional activator GAL4. J. Biol. Chem. 265, 209–213.

Puig, O., and Tjian, R. (2005). Transcriptional feedback control of insulin recep-

tor by dFOXO/FOXO1. Genes Dev. 19, 2435–2446.

Puig, O., Marr, M.T., Ruhf, M.L., and Tjian, R. (2003). Control of cell number by

Drosophila FOXO: downstream and feedback regulation of the insulin receptor

pathway. Genes Dev. 17, 2006–2020.

Relaix, F., Rocancourt, D., Mansouri, A., and Buckingham, M. (2005). A Pax3/

Pax7-dependent population of skeletal muscle progenitor cells. Nature 435,

948–953.

546 Developmental Cell 15, 534–546, October 14, 2008 ª2008 Else

Roth, J.F., Shikama, N., Henzen, C., Desbaillets, I., Lutz, W., Marino, S.,

Wittwer, J., Schorle, H., Gassmann, M., and Eckner, R. (2003). Differential

role of p300 and CBP acetyltransferase during myogenesis: p300 acts

upstream of MyoD and Myf5. EMBO J. 22, 5186–5196.

Rudnicki, M.A., Schnegelsberg, P.N., Stead, R.H., Braun, T., Arnold, H.H., and

Jaenisch, R. (1993). MyoD or Myf-5 is required for the formation of skeletal

muscle. Cell 75, 1351–1359.

Sabourin, L.A., Girgis-Gabardo, A., Seale, P., Asakura, A., and Rudnicki, M.A.

(1999). Reduced differentiation potential of primary MyoD�/� myogenic cells

derived from adult skeletal muscle. J. Cell Biol. 144, 631–643.

Sandri, M., Sandri, C., Gilbert, A., Skurk, C., Calabria, E., Picard, A., Walsh, K.,

Schiaffino, S., Lecker, S.H., and Goldberg, A.L. (2004). Foxo transcription fac-

tors induce the atrophy-related ubiquitin ligase atrogin-1 and cause skeletal

muscle atrophy. Cell 117, 399–412.

Seale, P., Sabourin, L.A., Girgis-Gabardo, A., Mansouri, A., Gruss, P., and

Rudnicki, M.A. (2000). Pax7 is required for the specification of myogenic

satellite cells. Cell 102, 777–786.

Seoane, J., Le, H.V., Shen, L., Anderson, S.A., and Massague, J. (2004). Inte-

gration of Smad and forkhead pathways in the control of neuroepithelial and

glioblastoma cell proliferation. Cell 117, 211–223.

Tapscott, S.J. (2005). The circuitry of a master switch: Myod and the regulation

of skeletal muscle gene transcription. Development 132, 2685–2695.

Tothova, Z., Kollipara, R., Huntly, B.J., Lee, B.H., Castrillon, D.H., Cullen, D.E.,

McDowell, E.P., Lazo-Kallanian, S., Williams, I.R., Sears, C., et al. (2007).

FoxOs are critical mediators of hematopoietic stem cell resistance to physio-

logic oxidative stress. Cell 128, 325–339.

van der Heide, L.P., Jacobs, F.M., Burbach, J.P., Hoekman, M.F., and Smidt,

M.P. (2005). FoxO6 transcriptional activity is regulated by Thr26 and Ser184,

independent of nucleo-cytoplasmic shuttling. Biochem. J. 391, 623–629.

Wagers, A.J., and Conboy, I.M. (2005). Cellular and molecular signatures of

muscle regeneration: current concepts and controversies in adult myogenesis.

Cell 122, 659–667.

Weintraub, H., Tapscott, S.J., Davis, R.L., Thayer, M.J., Adam, M.A., Lassar,

A.B., and Miller, A.D. (1989). Activation of muscle-specific genes in pigment,

nerve, fat, liver, and fibroblast cell lines by forced expression of MyoD. Proc.

Natl. Acad. Sci. USA 86, 5434–5438.

Weintraub, H., Dwarki, V.J., Verma, I., Davis, R., Hollenberg, S., Snider, L.,

Lassar, A., and Tapscott, S.J. (1991). Muscle-specific transcriptional activa-

tion by MyoD. Genes Dev. 5, 1377–1386.

Wilson, E.M., and Rotwein, P. (2006). Control of MyoD function during initiation

of muscle differentiation by an autocrine signaling pathway activated by

insulin-like growth factor-II. J. Biol. Chem. 281, 29962–29971.

Xuan, Z., and Zhang, M.Q. (2005). From worm to human: bioinformatics

approaches to identify FOXO target genes. Mech. Ageing Dev. 126, 209–215.

Yablonka-Reuveni, Z., Rudnicki, M.A., Rivera, A.J., Primig, M., Anderson, J.E.,

and Natanson, P. (1999). The transition from proliferation to differentiation is

delayed in satellite cells from mice lacking MyoD. Dev. Biol. 210, 440–455.

Yan, Z., Choi, S., Liu, X., Zhang, M., Schageman, J.J., Lee, S.Y., Hart, R., Lin,

L., Thurmond, F.A., and Williams, R.S. (2003). Highly coordinated gene regula-

tion in mouse skeletal muscle regeneration. J. Biol. Chem. 278, 8826–8836.

vier Inc.