Differential modulation of cell cycle progression distinguishes members of the myogenic regulatory factor family of transcription factors

Authors

  • Kulwant Singh,

    1. Sprott Center for Stem Cell Research, Ottawa Hospital Research Institute, ON, Canada
    Search for more papers by this author
  • F. Jeffrey Dilworth

    Corresponding author
    1. Department of Cellular and Molecular Medicine, University of Ottawa, ON, Canada
    • Sprott Center for Stem Cell Research, Ottawa Hospital Research Institute, ON, Canada
    Search for more papers by this author

Correspondence

F. J. Dilworth, Ottawa Hospital Research Institute, 501 Smyth Rd, Mailbox 511, Ottawa, ON, Canada K1H 8L6

Fax: +1 613 739 6294

Tel: +1 613 737 8899 ext 70339

E-mail: jdilworth@ohri.ca

Abstract

The muscle-specific basic helix–loop–helix proteins MyoD, Myf5, myogenin (Myog) and MRF4 constitute the myogenic regulatory factor (MRF) family of transcription factors that drive muscle gene expression during myogenesis. Having evolved from a single ancestral gene, the spatial and temporal specificity of expression for each family member has been used to define a hierarchical relationship between the four MRFs. Molecular characterization of two of the MRFs (MyoD and Myog) suggests an important distinction between these factors, whereby MyoD establishes an open chromatin structure at muscle-specific genes, whereas Myog drives high levels of transcription of genes within this open chromatin state. Furthermore, recent data have provided an additional distinction between MRF function with respect to cell cycle regulation. Indeed, MyoD has been shown to directly activate genes involved in cell cycle progression, leading to myoblast proliferation. In contrast, Myog has antiproliferative activity through the activation of genes that shut down the cell proliferation machinery, leading to cell cycle exit and myoblast differentiation. Although the transcriptional activities of MyoD and Myog synergize to drive muscle differentiation, it is the expression of Myog that sets in motion a gene expression program that constitutes a ‘point of no return’, leading to cell cycle exit. In this review, we compare and contrast the current literature with respect to MRF function, with a particular emphasis on the differential role of MRFs in modulating the cell cycle.

Abbreviations
bHLH

basic helix–loop–helix

C/H

cysteine/histidine-rich

ChIP

chromatin immunoprecipitation

E

embryonic day

KO

knockout

MAPK

mitogen-activated protein kinase

MRF

myogenic regulatory factor

Myog

myogenin

pRB

retinoblastoma protein

TAD

transactivation domain

Introduction

Developmental gene expression programs are established through the combined activity of tissue-restricted and ubiquitously expressed transcription factors that allow communication between distal and proximal regulatory regions of genes [1, 2]. According to current models, tissue-restricted transcription factors help to establish a remodeled chromatin structure that is permissive to binding of the transcriptional machinery at specific genes within a cell type [3]. The paradigm for such a model has been the transcriptional regulatory network that governs vertebrate skeletal myogenesis. Indeed, expression of the skeletal muscle-specific transcription factor MyoD was long ago shown to be sufficient to mediate lineage reprogramming through activation of the muscle gene expression program in multiple cell types [4]. This seminal finding led to the cloning of three related skeletal muscle-specific transcription factors [5]. Collectively, MyoD, Myf5, myogenin (Myog), and MRF4 constitute a family of transcription factors that have been termed myogenic regulatory factors (MRFs). Characteristic of MRFs, these proteins share a highly conserved variant of the basic helix–loop–helix (bHLH) domain [6] that confers their myogenic potential [7]. Whereas the expression of MRFs is skeletal muscle-specific, their function requires heterodimerization with a member of the ubiquitously expressed E-protein family of bHLH proteins. Upon dimerization, the MRF–E-protein heterodimeric complexes are able to bind the E-box consensus sequence (CANNTG), which is present in the regulatory regions of muscle-specific genes [8]. Although E-boxes are well represented in the genome, with more than 14 million sites, MRFs bind only a small fraction of these sequences [9, 10]. Furthermore, chromatin immunoprecipitation (ChIP) sequencing analysis has shown that MRFs bind both overlapping and distinct loci within the genome [10, 11]. This suggests that the binding of MRFs to their target E-box is limited by chromatin accessibility, but also through properties intrinsic to the individual family members.

Phylogenetic analysis indicates that the four vertebrate MRF genes have evolved from a single ancestral MRF gene as a result of gene duplication events and subsequent divergent mutations [12]. Indeed, many invertebrates, such as Caenorhabditis elegans, Drosophila, sea urchins, and acidians, continue to develop their musculature in the presence of a single MRF gene [13-15]. However, the more complex musculature of vertebrates has forced the evolution of four MRFs to regulate the complex gene expression program in myogenesis. To establish this complex regulation of gene expression, MRFs have retained high conservation of their DNA-binding domain (bHLH), while the transactivation domains of each transcription factor have diverges. This has allowed MyoD, Myf5, Myog and MRF4 to retain a certain degree of functional overlap while acquiring the unique properties required for development of the mammalian musculature. This has been clearly demonstrated in genetic models, where knock-in studies have shown that MRFs display a certain degree of functional redundancy, but are not completely interchangeable [16, 17]. Although elegant experiments involving swapping between MyoD and Myog have provided some insights into the roles of specific protein domains in mediating differential transactivation by these transcription factors [18, 19], most studies examining the mechanisms of transactivation have focused on the activities of individual MRFs. Thus, we currently lack a good understanding of the specific functional roles played by individual MRFs in the processes of muscle development, repair, and homeostasis. Here, we review the current literature concerning MRF functions that supports the notion that MRFs play differential roles in mediating cell cycle progression.

Understanding the role of MRFs through mouse genetics

Studies over the last few decades have made significant advances in demarcating the functions of individual MRFs in myogenesis. Importantly, these studies have shown that these muscle-specific transcription factors have evolved a hierarchical relationship that retains a certain degree of functional redundancy, while highlighting the fact that each MRF also possess unique properties that subtly affect the myogenic process. Mouse genetic studies have played a key role in identifying the hierarchical relationship between MRFs, whereby MyoD and Myf5 are generally viewed as factors involved in the determination of myogenic cells, whereas Myog and MRF4 are more closely associated with terminal differentiation and homeostasis of myofibers [5]. The definition of MyoD and Myf5 as determination factors comes from genetic studies in mice, where the double knockout (KO) of these two MRFs resulted in postnatal lethality, owing to a complete absence of skeletal myoblasts or myofibers [20-22]. The absence of an overt phenotype in mice lacking either MyoD or Myf5 provides strong evidence for functional overlap between these two determination factors [20-22]. However, further exploration of myogenesis in the absence of MyoD or Myf5 has revealed subtle difference between these two transcription factors. Developmental studies of the single KO mice showed delayed myogenesis in the epaxial myotome in the absence of Myf5, whereas myogenesis in the hypaxial myotome was delayed in mice lacking MyoD [23]. Further evidence for MyoD and Myf5 determining distinct muscle populations in the myotome comes from studies showing that ablation of Myf5-expressing cells failed to prevent MyoD-dependent skeletal muscle differentiation [24, 25]. This suggests that MyoD and Myf5 are responsible for determining distinct myogenic populations in the myotome that have the potential to compensate for each other in development.

Mice that lack Myog continue to specify the muscle lineage through the formation of myoblasts. However, these mice show perinatal lethality, because of severe disruption of myoblast differentiation and muscle fiber formation, leading to the idea that Myog is a differentiation factor in the myogenic process [26, 27]. This functional distinction between Myog and its related family members MyoD and Myf5 is further highlighted by studies showing that the Myog/MyoD or Myog/Myf5 double KO mice specify the muscle lineage but do not form muscle fibers – a phenotype similar to that of the Myog KO mouse [28]. These findings demonstrate that the function of Myog does not overlap with those of MyoD and Myf5, and that Myog acts downstream of MyoD and Myf5 in skeletal muscle development as a differentiation factor.

Studies have demonstrated that MRF4 can act as both a determination and a differentiation factor [29]. Similar to the effect of KO of the determination factors MyoD and Myf5, MRF4 null mice do not show any overt muscle phenotype [30]. This would suggest that the determination role of MRF4 is functionally redundant with MyoD and Myf5, whereas its differentiation role is redundant with that of Myog. Indeed, MRF4 KO mice did show strong upregulation of Myog expression, suggesting that MRF4 probably functions to downregulate Myog expression in the mature myofiber [30]. Interestingly, MRF4/MyoD KO mice died at birth, showing a phenotype highly similar to that of the Myog null mutant mice [31]. In MRF4/MyoD double KO mice, Myog was expressed, but this expression was insufficient to support normal myogenesis in vivo. This suggests that MRF4 and MyoD play a redundant role in mediating skeletal muscle differentiation during development. Taken together, these genetic studies establish the need for a minimum of two MRFs to generate functional muscle during development: one to determine the muscle lineage, and a second to permit terminal differentiation.

Spatial and temporal control of MRF expression

The four members of the MRF family can act as master regulators of skeletal myogenesis, whereby their exogenous expression can hijack the inherent gene expression program of a nonmuscle cell and drive it towards a myogenic fate [4, 32-34]. Therefore, the location, timing and expression levels of the MRFs during embryonic development are tightly regulated to ensure the accurate progression of the developmental process. In the murine model, precursor cells for myogenesis originate in the segmentally arranged nascent somites that flank both side of the neural tube and notochord. As embryonic development progresses (Fig. 1), a portion of the somites transforms into a transient structure called the dermomyotome [35]. Within these structures, expression of MRFs is first detected at around embryonic day (E)8.0, as sonic hedgehog signaling from the notochord and floor plate induces epaxial expression of Myf5 in the dorsal lips of the dermomyotome, committing these cells to become the epaxial myotome [36, 37]. At E10.5, Wnt (from the dorsal ectoderm) and Bmp4 (from the lateral mesoderm) signaling establishes MyoD expression in the hypaxial dermomyotome, causing these cells to establish the hypaxial myotome [38, 39]. Transcripts for Myog and MRF4 are first detected at E8.5 and E9.0, respectively, and their expression is evenly distributed throughout the myotome [39-42]. Consistent with a dual role in both determination and differentiation of the muscle lineage, MRF4 transcripts show biphasic expression, whereby they are downregulated by E11.5, but reappear at E16.0 in differentiated muscle fibers [40]. These MRFs execute the myogenic differentiation program, via expression of downstream targets, that results in the development of trunk muscle from the epaxial myotome, whereas limbs, diaphragm and body wall muscle develop from the hypaxial myotome.

Figure 1.

A schematic representation of embryonic and postnatal skeletal myogenesis. During vertebrate myogenesis, somites differentiate and subdivide to give rise to the dermomyotome (DM) and sclerotome (SCL) in response to signals from the neural tube (NT) and notochord (N). Cells from the dorsomedial lip (DML) migrate under the DM to form the epaxial myotome, which is the main source of the deep back musculature. Similarly, cells originating in the ventrolateral lip (VLL) migrate under the DM to generate the hypaxial myotome, which gives rise to the lateral trunk musculature. A proportion of cells from the VLL undergoing epithelial–mesenchymal transition delaminate and migrate to the region of presumptive limb muscle development (migrating limb precursors). The paired homeobox gene Pax3 has been demonstrated to be a key regulator of embryonic skeletal myogenesis, and sufficient to activate the expression of Myf5 and MyoD and initiate the myogenic program that leads to the development of myoblasts. Although myogenic precursor cells express both Pax3 and Pax7, Pax7-expressing cells are more prominent in the central region of the DM, and this is the population that gives rise to the major pool of Pax7+ satellite cells (SC) in the body. SCs are mitotically quiescent in adult muscle, and are located underneath the basal lamina. In response to muscle injury, SCs are activated and give rise to Myf5/MyoD cells, which return to the quiescent stage, and a Myf5+/MyoD+ myoblast population. Myoblasts generated from both embryonic and postnatal myogenesis undergo extensive proliferation, which leads to the generation of Myog+ myocytes in response to appropriate differentiation cues. Finally, these myocytes fuse to form myotubes and subsequently myofibers, which continue to express the terminal differentiation marker MRF4, structural and metabolic genes such as the myosin heavy chain (MHC) and the muscle creatinine kinase (MCK) genes.

Whereas in vivo studies have provided important spatial information about the expression of MRFs, the hierarchical relationship regulating the expression of the different family members has been defined through ex vivo studies using satellite cells. Satellite cells (SCs) are resident muscle stem cells that are mitotically quiescent in healthy adult muscle and become activated upon damage to the muscle fiber [43]. Overall, these studies have shown that the progression of activated satellite cells towards the myogenic lineage is mainly controlled by expression of Myf5 and MyoD [44]. Recent work suggests that Myf5 sits at the top of the hierarchy, as it was shown that the majority of quiescent satellite cells transcribe the Myf5 gene and are poised to enter the myogenic program. To maintain these poised cells in the quiescent state, mRNAs encoding Myf5 remain untranslated, owing to miR-31-dependent sequestration of the transcripts in mRNP granules. Upon satellite cell activation, mRNP granule dissociation results in the release of sequestered transcripts and rapid translation of the Myf5 mRNAs [45]. Soon afterwards, MyoD expression is initiated through the activity of Pax3/7, Six1/4, and FoxO3, to permit expansion of a cell population that is committed to the myogenic lineage [46, 47]. In the proliferating myoblasts, the MyoD gene is transcribed through a TFIID-dependent mechanism, whereby a moderate level of expression is ensured through spatial localization near the periphery of the nucleus [48, 49]. As myoblasts initiate differentiation towards myotubes, the MyoD gene moves towards the lumen of the nucleus, where a TAF3/TRF3-dependent transcriptional mechanism results in a high level of MyoD expression [48, 49]. Under these conditions, MyoD induces Myog expression, which results in downregulation of Myf5 expression [50]. This switch in expression from Myf5 to Myog coincides with cell cycle exit and a commitment to differentiate [2, 51]. The combined activity of MyoD and Myog leads to the expression of the MRF4 gene and other late muscle differentiation genes to permit the formation of multinucleated fibers. In mature muscle fibers, expression of MyoD and Myog is then downregulated, whereas MRF4 continues to be expressed at high levels to act as the predominant MRF in adult muscle [41].

The sequential expression of MRFs during muscle differentiation, combined with their functional redundancy, would suggest that the timing of MRF expression could be responsible for the differential roles of family members in muscle formation. Supporting such a possibility, insertion of the Myog cDNA into the Myf5 locus in Myf5/MyoD double KO mice allowed Myog to rescue the absence of skeletal muscle [16]. The formation of healthy myofibers in these knock-in mice demonstrated that Myog can act to specify the muscle lineage in the proper context. However, it is noted that these mice died at birth, owing to reduced skeletal muscle formation. Thus, whereas Myog was able to specify the muscle lineage, it appears that Myog may not have allowed for the expansion of the muscle progenitor population to generate sufficient cells to establish a complete musculature. Therefore, these findings demonstrate that there are functional differences inherent to MRF family members beyond their specific spatial and temporal expression patterns during development.

Divergence within structural domains of MRFs

Although all four MRFs are able to induce lineage reprogramming of fibroblasts towards a skeletal muscle lineage, studies in cultured cells have demonstrated differential efficiency between MRFs in activating the expression of silent muscle genes [18, 52]. To explain this divergent ability to activate gene expression within chromatin, differences in protein structure have been explored. Examination of the protein sequences shows that MRFs vary in size from 224 amino acids (Myog) to 319 amino acids (MyoD). Structurally, the MRFs are highly similar, in that they possess a conserved bHLH domain for DNA binding that is flanked by less conserved N-terminal and C-terminal domains that mediate transcriptional activation (Fig. 2). However, the divergent amino acid sequences in the transactivation domains (TADs) of MRFs raises the possibility that they may be responsible for the inherent differences between family members. This hypothesis was tested by Ishibashi et al. [53], who interchanged the TADs between MyoD and Myf5. These experiments demonstrated that the N-terminal and C-terminal domains of MyoD and Myf5 were interchangeable for the activation of gene expression involved in myoblast proliferation. However, in conditions of differentiation, the N-terminal and C-terminal TADs of MyoD act in a synergistic manner to induce the expression of muscle differentiation genes, and this synergy could not be achieved in combination with either of the Myf5 TADs [53].

Figure 2.

A schematic diagram of the structure of MyoD as a representative of the MRF family. The amino acids encompassed by the different domains are indicated. Various known post-translational modifications (PTMs) are presented in table below, and their conservation is determined on the basis of multiple sequence alignment.

Within the TADs of MyoD, two structural domains have been identified that allow MyoD to activate the expression of silent genes during myogenesis [18]. These domains are termed the cysteine/histidine-rich (C/H) domain and helix 3 domain, and they lie within the N-terminal and C-terminal TADs respectively. Interestingly, swapping of the C/H and helix 3 domains of MyoD into Myog allows Myog to efficiently activate the expression of silent muscle genes [18]. Characterization of the C/H and helix 3 domains showed that this structure is important for targeting MyoD to E-boxes within repressive chromatin through an interaction with the homeodomain protein Pbx1 [19]. Once tethered to the chromatin by Pbx1, MyoD can then recruit the SWI/SNF ATP-dependent chromatin remodeling complex [54] and the p300 histone acetyltransferase [55, 56] to muscle-specific genes for the establishment of an open chromatin structure. Importantly, the C/H and helix 3 domains of MyoD and Myf5 are highly conserved, whereas the domains diverge in the determination factors Myog and MRF4 [18]. Genome-wide binding analyses have demonstrated that MyoD and Myf5 share a significant proportion of their targets in proliferating myoblasts [10]. This would suggest that a shared ability to remodel chromatin might be at the heart of the redundant ability of MyoD and Myf5 to efficiently specify the muscle lineage. Importantly, the C/H and helix 3 domains of Myog cannot mediate remodeling of chromatin within the promoter of muscle genes [18, 52]. Instead, Myog is a strong activator of transcription at loci with an open chromatin structure previously established by MyoD [57]. That being said, Myog can function in concert with transcription factors such as Mef2D to recruit chromatin remodeling machinery at muscle genes to permit transcriptional activation [58]. This observation would explain how Myog could act as a specification factor when expressed from the Myf5 locus in the MyoD/Myf5 double KO background [16]. Thus, MyoD and Myf5 possess an activity that allows for the opening of chromatin that is not shared by Myog.

Domain swapping studies have also revealed functional differences between MyoD and MRF4, whereby the N-terminal domain of MRF4 can act as either a transcriptional activator or as a repressor, depending on the promoter context [59]. Among the target genes repressed by MRF4 is the cardiac α-actin gene, which is activated by MyoD expression [59]. Direct competition studies on the cardiac α-actin promoter showed that the repressive property of MRF4 predominates over MyoD-mediated transactivation, suggesting that the relative levels of different MRFs may modulate the transcriptional output of specific muscle genes [59]. Interestingly, MRF4 activity is modulated by the p38 mitogen-activated protein kinase (MAPK) signaling pathway during the course of myoblast differentiation [60]. Phosphorylation of MRF4 by p38 MAPK at Ser31 and Ser42 within the N-terminal transactivation domain inhibits its function, permitting activation of the cardiac α-actin gerne while blocking CKm gene expression [60]. Thus, the N-terminal TAD of MRF4 cooperates with p38 MAPK to selectively modulate the expression of the late myogenic transcriptional program. Taken together, these results suggest that divergent TADs within the MRF family play a role in modulating their targeting to specific genomic loci.

MRFs and their transcriptional targets

The binding of several MRFs has been explored in cultured cell systems with high-throughput technologies such as ChIP microarray and ChIP sequencing [9, 10, 57, 61]. Among these, genome-wide binding of MyoD has been the most thoroughly explored. The consensus among these different studies is that MyoD binds a large number of genes in muscle cells while modifying gene expression at only a fraction of its targets [9, 10, 57, 61]. However, an important finding of the study by Cao et al. was the observation that binding of MyoD within the genome correlated well with opening of the chromatin structure through acetylation of histones [9]. This opening of chromatin has recently been shown to establish enhancer elements that can control the expression of both coding and noncoding RNAs [62]. This would suggest that a major role of MyoD in the myogenic process is the opening of chromatin at specific loci to establish the muscle-specific chromatin state. Consistent with this idea, studies in the Tapscott laboratory demonstrated that MyoD-dependent activation of late muscle genes required the activity of Myog [57]. On the basis of these findings, it was proposed that the transcriptional activation domain of Myog drives high-level expression at genes that have an open chromatin structure established by MyoD [57]. Nevertheless, MyoD can also directly modulate the activation of gene expression. During differentiation, MyoD-dependent activation of immediate-early genes occurs in the absence of Myog [57]. Similarly, MyoD has been shown to directly regulate the expression of a subset of genes in proliferating myoblasts through direct binding to the promoter [9, 10]. Gene ontology analysis of these different target genes revealed that MyoD displayed enhanced binding in myotubes at genes involved in muscle development, whereas genes involved in the regulation of cell cycle were preferentially bound by MyoD in proliferating myoblasts [9, 61]. Thus, MyoD binding facilitates the expression of genes that promote proliferation or differentiation of myoblasts, depending on the cellular environment.

Genome-wide occupancy of Myf5 has been reported recently in proliferating myoblasts [10]. Comparative analysis of MyoD and Myf5 binding sites showed a highly significant overlap of ~ 30% [10]. This result is consistent with the notion that MyoD and Myf5 share a role in defining myoblast identity. Unfortunately, the study did not report the ontology of Myf5 target genes, and thus it remains to be seen whether Myf5 modulates myoblast proliferation through direct upregulation of genes involved in cell cycle regulation.

Analysis of the genome-wide binding of Myog that was performed within the ENCODE project has not yet been published [63]. Thus, our understanding of Myog binding is shaped mostly by ChIP array studies performed to identify associations with defined promoter regions [57, 61]. These studies revealed that Myog is bound to 75% of the promoters that are targeted by MyoD in differentiating myotubes [61]. This finding is consistent with the view that MyoD establishes an open chromatin structure that, in turn, permits binding of Myog to establish a high level of transcriptional activation [57]. Examination of the genes co-bound by MyoD and Myog identified genes involved in muscle development [57, 61]. Interestingly, microarray studies examining the role of Myog in differentiation recently identified genes involved in cell cycle progression as key transcriptional targets that are downregulated by Myog during differentiation [51]. This suggests that Myog is an important modulator of cell cycle exit during differentiation. Thus, in contrast to MyoD, which promotes proliferation in growing myoblasts, Myog attenuates the expression of genes that mediate cell cycle progression.

MRFs as modulators of the cell cycle

Studies from cultured cell systems support the notion that MyoD and Myog have opposing roles in modulating the cell cycle. In fact, both MyoD and Myf5 have been shown to promote expansion of the muscle progenitor population [64-66]. In contrast, Myog appears to possess intrinsic activity that is required to mediate cell cycle exit [51]. Indeed, it was recently demonstrated that ectopic expression of Myog in proliferating myoblasts leads to exit from the cell cycle [51]. This finding provides an explanation for the phenotype observed in the studies where Myog is expressed from the Myf5 locus in the Myf5/MyoD double KO mouse [16]. These mice had specified the muscle lineage and established healthy muscle fibers. However, the mice died at birth, owing to insufficient musculature. Thus, expression of Myog from the Myf5 locus is sufficient to determine the muscle lineage, although it probably caused precocious exit from the cell cycle, leading to impaired expansion of the progenitor pool, resulting in less muscle. Furthermore, this finding might explain why stable expression of Myog in primary myoblasts has never been reported. Taken together, these findings support the notion that MyoD and Myf5 promote expansion of muscle progenitor pools, whereas Myog induces cell cycle exit.

Characterization of satellite cells from MyoD KO [65] and Myf5 KO [66] mice showed that they both had proliferation defects. This suggests that the two determination MRF factors play nonredundant roles in the cell cycle. Further evidence for this notion come from studies showing the distinct and contrasting expression patterns of MyoD and Myf5 during the different phases of the cell cycle [67]. Myf5 protein levels peak in G0, decrease during G1, and then rise again at the end of G1 where they remain stable through mitosis (Fig. 3). In contrast, MyoD has been shown to block the G1/S transition [68]. Thus, MyoD protein levels peak in mid-G1, are reduced to their minimum level on G1/S transition, and are reaugmented from S to M [67]. These changes in MyoD and Myf5 protein levels during the cell cycle are modulated through post-translational modifications that signal their degradation. Although not all of the mechanisms regulating MRF stability have been established, several studies have highlighted modifications that lead to degradation of MyoD or Myf5. In particular, the Myf5 protein level is modulated by its phosphorylation and subsequent degradation at mitosis [69, 70]. In the case of MyoD, degradation at late G1 is mediated by the ubiquitin proteasome system, which is triggered by cyclin E/CDK2-dependent phosphorylation of MyoD on Ser200 [71, 72]. The transcriptional activity of MRFs can also be modulated during the cell cycle, as it has been shown that MyoD is phosphorylated at Ser5 and Ser200 by cyclin B/Cdc2 in mitosis to inhibit its DNA-binding activity and transcriptional activation ability [73]. Thus, the activities of Myf5 and MyoD are dynamically regulated throughout the cell cycle via modulation of transcriptional activity and protein abundance.

Figure 3.

A schematic overview of crosstalk between MRFs and cell cycle regulation. MyoD and Myf5 show distinct and contrasting expression patterns during the different phases of the cell cycle. The Myf5 protein level peaks in G0, decreases during G1, and then rise again at the end of G1 where they remain stable through mitosis. In contrast, the MyoD protein level peaks in mid-G1, decreases to its minimum level in the G1/S transition, and is reaugmented from S to M. The levels of both proteins during the different phases of the cell cycle are regulated by phosphorylation-dependent degradation via the 26S proteasome. In proliferating myoblasts, MyoD initiates the expression of two genes, CDC6 and MCM2, which are primarily involved in making chromatin operational for DNA replication and progression of cells through S-phase. In response to appropriate differentiation signals, MyoD induces the expression of Myog, and establishes a transcriptionally permissive chromatin structure at the p21cip, p57kip and pRB genes. Myog and MyoD then synergize to upregulate the expression of these key regulators of cell cycle exit – p21cip and p57kip repress the activity of Cdks and cyclins, whereas pRB targets E2F family members, which are major regulators of the expression of Cdks and cyclins. Myog also interfers with cell cycle progression by upregulating the expression of miR-20a and LATS2. miR-20a is a microRNA that is well characterized for its ability to downregulate the transcription factors E2F1 and E2F3 through targeting the 3′-UTRs of their mRNAs. LATS2 is a protein kinase that has been implicated in targeting of the transcriptional repressor complex DREAM to E2F target genes to block cell cycle progression.

Although distinctions in the expression patterns have been observed, the relative importance of MyoD and Myf5 to cell cycle progression has not been established. Genome-wide binding studies have previously established that MyoD binds to the transcriptional regulatory region of genes with critical roles in the cell cycle [9, 57, 61]. Characterization of specific promoters in growth-stimulated quiescent myoblasts showed that MyoD directly activates expression of the CDC6 and MCM2 genes, which are involved in preparing chromatin for DNA replication, and consequently progression of cell through S-phase [64]. We note that, whereas MyoD shows greater efficiency in establishing transcription at two genes that modulate the cell cycle, Myf5 can also perform this function, showing some degree of functional redundancy [64]. Future mechanistic studies based on the plethora of genome-wide data should prove highly informative with respect to the transcriptional activation of specific genes by Myf5 and MyoD in proliferating myoblasts.

An interesting distinction between MyoD and Myf5 is their ability to modulate cell cycle progression in response to DNA damage [74]. The cellular response to DNA damage is to block cell cycle progression at specific checkpoints, which allows DNA repair to prevent the propagation of genomic mutations in daughter cells [75, 76]. In the case of skeletal myogenesis, MyoD is a genuine target for the differentiation checkpoint. In response to genotoxic stress, the transcriptional activity of MyoD is repressed to prevent myotube formation in cells that have arrested the cell cycle because of DNA damage. This transcriptional inhibition requires phosphorylation of MyoD at Tyr30 by c-Abl tyrosine kinase [77]. Interestingly, the consensus site for Abl-mediated tyrosine phosphorylation is absent in Myf5, making this protein insensitive to the DNA damage response. The differential susceptibility of MyoD and Myf5 to c-Abl-mediated repression is not surprising, as MyoD levels peak in mid-G1, where myoblasts exit the cell cycle, whereas Myf5 levels are minimal at this stage of the cell cycle. Interestingly, the c-Abl consensus binding site (YDDP) is also absent in Myog and MRF4, suggesting that this role in facilitating DNA repair is unique to MyoD [74, 77].

For successful execution of the myogenic differentiation program, myoblasts must exit the cell cycle. An extensive literature has supported a role for MyoD in the induction of cell cycle exit during terminal differentiation (reviewed in [78]). Indeed, it has been shown that MyoD−/− myoblasts fail to exit the cell cycle, as the transcriptional regulator nuclear factor-κB (NF-κB) maintains a nuclear localization to activate the transcription of key cell cycle regulators [79]. However, recent studies examining bromodeoxyuridine staining in differentiating myoblasts have demonstrated that MyoD expression in promyogenic conditions is not sufficient for cell cycle exit [51]. Instead, it appears that the activation of Myog expression by MyoD constitutes a ‘point of no return’, where Myog initiates a gene expression program that commits the differentiating myoblast to exit the cell cycle [51]. Consistent with the existing literature, Myog-induced cell cycle exit appears to occur through the activation of CDKN1a (p21cip), as well as pathways that inhibit the expression of E2F target genes – cell cycle targets that have previously been attributed to MyoD [45, 78, 80-82]. Indeed, exogenous expression of Myog in proliferating myoblasts leads to upregulation of p21cip, which triggers cell cycle exit [51]. This appears to be a direct modulation of p21cip expression, as a binding site for Myog has been identified at the CDKN1a promoter (unpublished observation based on the ENCODE data from CalTech, K. Singh and F. J. Dilworth). In the case of E2F target genes, Myog takes a multipronged approach to ensure efficient silencing of these loci. In a first mechanism, Myog upregulates transcription of the miR17-92 cluster of microRNAs [51]. Among these microRNAs, miR-20a has been shown to target the mRNA of E2F1, E2F2 and E2F3 to inhibit their translation [83-85]. Therefore Myog can indirectly downregulate the cellular protein levels of E2F family members. A second mechanism used by Myog to inhibit E2F target gene expression is upregulation of Lats2 kinase [51], which has been shown to inhibit transactivation by E2F proteins. Lats2 was identified in a small hairpin RNA screen of factors that function with retinoblastoma protein (pRB) to mediate cell cycle exit [86]. There, it was shown that Lats2, a component of the Hippo pathway [87], is required to initiate a phosphorylation cascade that leads to pRB-dependent binding of the repressive complex DREAM to E2F target genes [86]. This finding is particularly interesting, as it has been shown that downregulation of pRB and Arf allows Myog-expressing myocytes to re-enter the cell cycle [88]. Furthermore, MyoD-induced myogenesis leads to upregulation of pRB to mediate cell cycle exit [89] through the establishment of repressive histone methylation at cell cycle genes [90]. Although we have not examined whether Myog can directly lead to upregulation of pRB and ARF, we note that MyoD and Myog both bind to the promoter and an intronic enhancer of the Rb1 gene (unpublished observation based on the ENCODE data from Caltech). On the basis of this observation, we suggest that MyoD may create an open chromatin structure at the Rb1 gene that facilitates strong transactivation by Myog. However, this possibility still needs to be formally tested. Nevertheless, it appears that MyoD and Myog act in a coordinated manner to ensure cell cycle exit through the activation of key genes involved in suppressing E2F activity. The mechanism by which NF-κB activity is inhibited during cell cycle exit, and whether Myog participates in this process, remain to be determined.

It is important to note that the mechanisms described above for Myog-mediated cell cycle exit are all indirect, acting by controlling the upregulation of target genes that are directly involved in blocking cell cycle progression. This distinction is important, as it explains observations made by the Walsh group showing that a small percentage of differentiating myoblasts staining positive for Myog can proceed through a round of cell cycle to incorporate bromodeoxyuridine [91]. Thus, we propose that expression of Myog during myogenic differentiation represents a ‘point of no return’, where the MRF initiates a gene expression program that commits the myoblast to withdraw from the cell cycle. In this model, cells that had committed to division prior to achieving sufficient expression of Myog target genes would complete a final round of replication before permanently exiting the cell cycle.

The molecular characterization of cultured cell systems has thus provided some mechanistic insights into the differential roles of MRFs in cell cycle progression as defined through ex vivo cellular studies of myoblasts from KO mice. Combining these findings, we propose the following generalization: the MRFs that play a role in determination (MyoD and Myf5) facilitate cell cycle progression, whereas the MRFs that mediate differentiation (Myog and MRF4) induce cell cycle exit. Confirmation of this generalization will require studies to examine the role of MRF4 in the cell cycle.

Conclusions and perspective

The functional redundancy within the MRF family is evident from genetic studies in mice showing that the development of the murine skeletal musculature can be directed through the activities of as few as two MRFs – a determination MRF and a differentiation MRF. Although it has been difficult to unravel the subtle differences in function between MRFs, high-throughput technologies that permit us to pursue unique target genes as well as changes in gene expression and chromatin structure at a global level have begun to provide novel insights into the importance of the multiple MRFs in mediating myogenesis. These studies have identified a role for the determination MRFs in specifying the muscle lineage through the opening of chromatin structure at genes involved in muscle development, but also in expanding the muscle progenitor population. In addition, they have identified a role for the differentiation MRF, Myog, in activating high-level expression of muscle genes that lie in an open chromatin structure while also activating signaling cascades to ensure cell cycle withdrawal. Further examination of the plethora of information present in the high-throughput datasets is certain to provide us with further insights into the subtle difference between the various MRFs. Unfortunately, our knowledge of MRF4 function in myogenesis remains poorly characterized, owing to the absence of good cellular models for study of its function ex vivo. The current development of more sensitive techniques to exploit ChIP sequencing and RNA sequencing technologies will hopefully soon allow for the isolation of MRF4-expressing cells from the developing mouse, so that we can explore the distinct functional features of this MRF, which displays both determination and differentiation characteristics.

Acknowledgements

We would like to thank M. Brand for critically reading the manuscript. Work in the Dilworth laboratory on MRF function is supported by a grant from the Canadian Institutes of Health Research (MOP-77778). F. J. Dilworth is the Canada Research Chair in Epigenetic Regulation of Transcription.

Ancillary