Regulation of Muscle Differentiation by the MEF2 Family of MADS Box

DEVELOPMENTAL BIOLOGY 172, 2–14 (1995)

MINI-REVIEW View metadata, citation and similar papers at core.ac.uk brought to you by CORE Regulation of Muscle Differentiation by the MEF2 provided by Elsevier - Publisher Connector Family of MADS Box Transcription Factors

Eric N. Olson,1 Michael Perry, and Robert A. Schulz Department of Biochemistry and Molecular Biology, The University of Texas M. D. Anderson Cancer Center, 1515 Holcombe Boulevard, Houston, Texas 77030

INTRODUCTION tube gives rise to muscle precursors that form the myotome, the origin of the vertebral and back muscles. Cells from There has been dramatic progress in recent years toward the ventrolateral edge of the dermamyotome migrate out understanding the molecular mechanisms that regulate to form the muscles of the limbs and body wall. skeletal muscle development. In contrast, relatively little Cardiac muscle arises from cells in the anterior lateral is known of the mechanisms that give rise to cardiac and plate mesoderm, which become committed to a cardiogenic smooth muscle during embryogenesis. These different mus- fate soon after gastrulation. These precardial cells give rise cle cell types express many of the same muscle-specific to a heart tube, which undergoes looping, followed by for- genes. However, they are each unique in several respects, mation of the atria and ventricles. including the spectrum of muscle isoforms expressed, mor- In contrast to skeletal and cardiac muscle, which arise phology, contractile properties, and ability to divide. In prin- from distinct populations of mesodermal precursors, ciple, the expression of muscle genes in skeletal, cardiac, smooth muscle arises throughout the embryo from different and smooth muscle cells could be controlled by a shared populations of mesenchymal cell precursors, as well as from myogenic regulatory program, which is modified within the neural crest (Miano et al., 1994). Among the three major each lineage to confer the unique identities of each muscle muscle cell types, the least is known about the mechanisms cell type. Alternatively, there could be myogenic regulatory that control muscle gene expression in smooth muscle cells. factors unique to each myogenic lineage, which act through different cis-acting DNA sequences to activate muscle structural genes. This review will consider recent evidence Regulation of Muscle Differentiation by Myogenic that suggests the existence of a common myogenic program, bHLH Proteins controlled by the myocyte enhancer factor-2 (MEF2) family of MADS box transcription factors. Much of our knowledge of skeletal muscle development has come from the discovery of the MyoD family of basic helix-loop-helix (bHLH) proteins, which can activate skele- Embryonic Origins of the Three Muscle Cell Types tal muscle gene expression when expressed ectopically in During vertebrate embryogenesis, skeletal, cardiac, and nonmuscle cell types (reviewed in Olson, 1990; Weintraub smooth muscle cells arise from distinct mesodermal precur- et al., 1991; Rudnicki and Jaenisch, 1995). There are four sors in different regions of the embryo (Fig. 1). Skeletal mus- members of this family in vertebrate species, MyoD, Myo- cle is derived from the somites, which form by segmenta- genin, Myf5, and MRF4, which share extensive amino acid tion of the paraxial mesoderm lateral to the neural tube homology within their bHLH regions. During embryogene- (Wachtler and Christ, 1992). The somites appear initially as sis, the myogenic bHLH factors show overlapping, but dis- epithelial spheres, which subsequently become compart- tinct, expression patterns in the skeletal muscle lineage. mentalized to form the dermamyotome and the sclerotome. Myf5 is the first member of the family to be expressed in the The region of the dermamyotome adjacent to the neural mouse, appearing in the dorsomedial region of the rostral uncompartmentalized somites beginning at Day 8.0 post coitum (p.c.). Myogenin is expressed in the myotome about 1 To whom correspondence should be addressed. Fax: (713) 791- a half-day later and MRF4 and MyoD are expressed begin- 9478. ning at Days 9.5 and 10.5 p.c., respectively (reviewed in

0012-1606/95 $12.00 Copyright ᭧ 1995 by Academic Press, Inc. 2 All rights of reproduction in any form reserved.

/ m4070$8017 10-02-95 05:58:56 dba Dev Bio MEF2 Family of MADS Box Transcription Factors 3

The MEF2 Family of MADS Box Transcription Factors

MEF2 was first identified as a DNA binding activity that recognized an A/T-rich element in the muscle creatine kinase (MCK) enhancer that was essential for full enhancer activity (Gossett et al., 1989). Subsequently, MEF2 sites have been found in the promoters and enhancers of the majority of skeletal and cardiac muscle structural genes (Braun et al., 1989; Ianello et al., 1991; Wentworth et al., 1991; Zhu et al., 1991; Nakatsuji et al., 1992; Navankasattu- sas et al., 1992; Muscat et al., 1992; Morisaki and Holmes, 1993; Hidaka et al., 1993; Molkentin and Markham, 1993; Wang et al., 1994; Li and Capetanaki, 1994; Parmacek et al., 1994). The cloning of genes encoding MEF2 factors (also called RSRFs, for related to Serum Response Factors) revealed that these proteins belong to the MADS box family of transcription factors, named for the first four factors in which this domain was identified: MCM1, which regulates mating type-specific genes in yeast, Agamous and Deficiens, which act as homeotic factors that control flower development, and Serum Response Factor, which controls serum- inducible and muscle-specific gene expression (Pollock and Treisman, 1991). Four mef2 genes, referred to as mef2A– mef2D, have been identified in vertebrate species (Pollock and Treisman, 1991; Yu et al., 1992; Martin et al., 1993, FIG. 1. Schematic diagram of vertebrate myogenic lineages. 1994; Breitbart et al., 1993; McDermott et al., 1993; Leifer et al., 1993; Chambers et al., 1994; Wong et al., 1994). There is also a single mef2 gene in Drosophila (Lilly et al., 1994; Nguyen et al., 1994; Taylor et al., 1995) and in Caenorhab- ditis elegans (M. Krause, personal communication). The Buckingham, 1992). The roles of these factors in muscle MEF2 proteins share greater than 80% amino acid homol- development have been confirmed by gene knockout experi- ogy within the 56-amino-acid MADS box at their amino ments, which have shown that MyoD and Myf5 play redun- termini (Fig. 2). Adjacent to the MADS box is a 29-amino- dant roles in the generation of myoblasts, whereas Myo- acid domain, called the MEF2 domain, that is not present genin directs myoblast differentiation (Braun et al., 1992; in other MADS box proteins. Rudnicki et al., 1992, 1993; Hasty et al., 1993; Nabeshima The MADS box mediates DNA binding and dimerization et al., 1993; Olson and Klein, 1994). MRF4 is expressed (reviewed in Shore and Sharrocks, 1995). Consistent with the late in muscle differentiation and may share functions with homology among MADS box proteins, different members of Myogenin (Zhang et al., 1995). Because the myogenic bHLH this family recognize similar A/T-rich DNA sequences. How- proteins are not expressed in cardiac or smooth muscle, ever, the consensus sequence for MEF2 binding, YTA(A/ other regulators must control muscle gene expression in T)4TAR, is distinct from the binding sites of other MADS box these cell types. proteins. The binding sites for MEF2 and other MADS box The myogenic bHLH proteins activate muscle gene ex- proteins exhibit dyad symmetry, allowing each component of pression by binding to the consensus sequence CANNTG the dimeric DNA binding complex to recognize half of the (N, any nucleotide) as heterodimers with ubiquitous bHLH binding site. Mutagenesis of several MADS box proteins in- proteins, known as E-proteins (Murre et al., 1989). This cluding MEF2 has shown that DNA binding requires the 56- DNA sequence, referred to as an E-box, is found in the amino-acid MADS box, in addition to an extension of about control regions of many but not all skeletal muscle genes 30 amino acids on the carboxyl-terminal side of the MADS (Hauschka, 1994). While E-boxes are required for transcrip- box, which is unique to each subclass of MADS box proteins tional activation of many skeletal muscle genes, they are (Pollock and Treisman, 1991; Molkentin and Olson, 1995). not by themselves sufficient and depend on adjacent binding The three-dimensional structure of the MADS box has not sites of other factors to activate skeletal muscle gene tran- yet been elucidated. However, sequence specificity of DNA scription. MEF2 binding sites are often associated with E- binding has been shown to be mediated by basic residues boxes in muscle gene regulatory regions and have been that lie on the same side of a predicted a-helix at the amino- shown to be required for transcriptional activation of those terminal end of the MADS box. Replacing these 28 amino- genes in skeletal muscle cells. terminal residues of the MADS box of SRF with those of

/ m4070$8017 10-02-95 05:58:56 dba Dev Bio 4 Olson, Perry, and Schulz

FIG. 2. Schematic diagrams of the mammalian MEF2 factors. The MADS and MEF2 domains are shown at the amino terminus of each protein. Alternative exons are also indicated. Adapted from Martin et al. (1994).

MEF2A alters the DNA binding specificity of SRF to that of on which factor is bound at the site. The MEF2 site in the MEF2 (Sharrocks et al., 1993). myosin light chain-2 (mlc-2) gene promoter also binds a tissue- MEF2A, MEF2C, and MEF2D bind the same DNA se- restricted zinc finger protein (HF-1b), which appears to play quence, but subtle differences in binding affinity have been an important role in transcriptional activation (Zhu et al., observed among these MEF2 factors, suggesting that their non- 1993), and a serum-inducible cardiac-specific factor, BBF-1 conserved residues may affect their binding properties. DNA (Zhou et al., 1993). sequences surrounding the core consensus sequence also af- MEF2 factors can homo- and heterodimerize, but they do fect DNA binding (Yu et al., 1992). MEF2B fails to bind the not interact with other known MADS box factors (Pollock MEF2 consensus sequence as a homodimer in vivo or in vitro and Treisman, 1991). Dimerization of MEF2 is mediated by a (Pollock and Treisman, 1991; Yu et al., 1992). However, bind- hydrophobic stretch of amino acids toward the C-terminal ing activity is observed with a deletion mutant of MEF2B end of the MADS box, which is predicted to adopt a b-strand containing only the MADS and MEF2 domains, suggesting conformation. The MEF2 domain is also required for efficient that the carboxyl-terminal region of MEF2B inhibits DNA DNA binding, but it does not affect DNA sequence recogni- binding. Intriguingly, MEF2B activates transcription through tion (Molkentin et al., 1995). Most likely, this region confers the MEF2 site in transfected cells, which raises the possibility dimerization specificity and influences the orientation of the that it may have a unique partner in vivo or that it potentiates DNA binding domains of the heterodimeric partners. the activity of other MEF2 factors. The carboxyl termini of MEF2A, MEF2C, and MEF2D have The MEF2 binding site also binds several factors in addition been shown to function as transcription activation domains to MEF2, which in principle can allow for regulation of gene (Wong et al., 1994; Martin et al., 1994; Molkentin et al., 1995). expression through competition for DNA binding. In the core Mutations in this region of the Drosophila MEF2 protein re- of the MCK enhancer, for example, is a low-affinity MEF2 site sult in partial loss-of-function alleles (Ranganayakulu et al., that is essential for enhancer activity in skeletal and cardiac 1995; discussed below). The mef2A, mef2C, and mef2D genes muscle cells. This site also binds the paired-like homeodo- give rise to multiple proteins by alternative splicing within main protein MHox and the POU domain protein Oct-1. Mu- the transcription activation domain, with certain exons being tational analysis of this site has shown that it must be bound muscle-specific and others being ubiquitous. Heterodimeriza- by MEF2 for enhancer activity in muscle cells (Cserjesi et al., tion among different MEF2 proteins and different splice vari- 1994). Since occupancy of this site by these three factors is ants can in principle result in greater than 100 different hetero- mutually exclusive, the activity of the enhancer is dependent dimeric complexes that recognize the same DNA sequence.

/ m4070$8017 10-02-95 05:58:56 dba Dev Bio MEF2 Family of MADS Box Transcription Factors 5

It remains to be determined how the different splice variants lational control might be mediated by the untranslated re- might differ functionally. gions of MEF2 mRNAs because exogenous MEF2 transcripts lacking sequences from the 5؅ and 3؅ untranslated regions are translated in fibroblasts, in which endogenous Regulation of MEF2 Expression MEF2 transcripts are not efficiently translated. The high During mouse embryogenesis, the mef2 genes are ex- degree of sequence conservation of the untranslated regions pressed in precursors of the three myogenic lineages and of MEF2 mRNAs also suggests that they may play a regula- their descendants. mef2C is the first member of the family tory role. to be expressed, appearing in the precardiac mesoderm at Day 7.5 p.c. (Edmondson et al., 1994). Soon thereafter, the A Mutually Reinforcing Network of Myogenic other mef2 genes are expressed in the developing heart and Factors the expression of cardiac muscle structural genes ensues. All of the embryonic expression studies have been per- The cell type distribution of MEF2 DNA binding activity formed with probes that do not distinguish between the has been the subject of much debate. While numerous stud- different exons of the mef2 genes. Expression patterns of ies have documented that MEF2 is highly enriched in differ- muscle-specific and ubiquitous exons of the various mef2 entiated muscle cells (Cserjesi and Olson, 1991; Muscat et genes remain to be determined. The expression of mef2C al., 1992; Hidaka et al., 1993; McDermott et al., 1993; Mol- in the cardiac muscle lineage co-incides with the expression kentin and Markham, 1993), others have reported that it is of the cardiac homeobox gene Nkx-2.5/csx (Lints et al., ubiquitous (Horlick et al., 1990; Pollock and Treisman, 1993; Komuro and Izumo, 1993). A homologue of this gene 1991; Han et al., 1992). The basis for these discrepancies is called tinman has been identified in Drosophila, where it unclear. Consistent with the enrichment of MEF2 activity is required for formation of the dorsal vessel, which is analo- in muscle cells, reporter genes linked to multimerized gous to the heart (Bodmer, 1995). The coexpression of MEF2 MEF2 sites are preferentially expressed in differentiated and this homeodomain protein in the early heart is espe- myocytes (Gossett et al., 1989; Yu et al., 1992). The mecha- cially intriguing in light of the cooperation between MADS nisms that regulate MEF2 expression during myogenesis and homeodomain proteins in other systems (see below) have also been controversial. Gossett et al. (1989) reported and suggests the possibility of direct interactions between that protein synthesis was required for the upregulation of these proteins. MEF2 DNA binding activity that accompanies myoblast In the skeletal muscle lineage, MEF2C is expressed in the differentiation, whereas Buchberger et al. (1994) reported somite myotome a few hours after Myogenin, making it that MEF2 activity was induced in the presence of cyclohex- unlikely that MEF2 is required for the genesis of myoblasts imide. or for the initial activation of myogenin gene expression During differentiation of skeletal myoblasts in vitro, the (Edmondson et al., 1994). As in the cardiac muscle lineage, different MEF2 proteins accumulate sequentially. MEF2D the other mef2 genes are expressed in the skeletal muscle has been reported to be expressed first in myoblasts, but lineage after mef2C. mef2 gene expression is also observed it apparently does not activate muscle target genes until in smooth muscle cells throughout the mouse embryo, myoblasts exit the cell cycle (Breitbart et al., 1993). The where it precedes the expression of muscle structural genes. mechanisms that repress MEF2D activity in myoblasts re- By about Day 14 p.c. of mouse embryogenesis, mef2 tran- main to be determined. Following initiation of the differen- scripts begin to appear in a variety of nonmuscle cell types tiation program by withdrawal of serum, MEF2A protein and by birth, mef2A, mef2B, and mef2D are expressed ubiq- accumulates. MEF2C protein does not appear until late in uitously, except in the brain, where they show highly local- terminally differentiated myotubes (McDermott et al., ized expression patterns (Lyons et al., 1995). mef2C expres- 1993; Martin et al., 1993). These sequential expression pat- sion remains restricted to skeletal muscle, brain, and spleen terns of the MEF2 factors during myogenesis are reminis- in adults. Within the developing brain, expression of the cent of the expression patterns of the myogenic bHLH pro- mef2 genes follows gradients of neuronal differentiation. teins. One of the perplexing aspects of MEF2 regulation is the MEF2 binding activity can be induced in nonmuscle cells disparity between the expression of MEF2 mRNAs and pro- by forced expression of myogenic bHLH factors (Lassar et teins. MEF2A, MEF2B, and MEF2D transcripts are ex- al., 1991; Cserjesi et al., 1994). This upregulation occurs in pressed in a wide range of adult tissues and established cell 10T1/2 fibroblasts in which the complete myogenic pro- lines, but MEF2 protein and DNA binding activity are gram is induced, as well as in CV-1 cells, which are refrac- largely restricted to differentiated muscle cells and neurons. tory to myogenic conversion. These findings led to the ini- The most likely explanation for this disparity is the exis- tial conclusion that MEF2 was present in a regulatory path- tence of a mechanism for translational repression of MEF2 way ‘‘downstream’’ of myogenic bHLH proteins. However, mRNAs in cell types in which the protein is undetected. it was recently reported that forced expression of MEF2A in Indeed, translation control of MEF2A expression has re- fibroblasts can activate expression of Myogenin and MyoD, cently been demonstrated in vascular smooth muscle cells resulting in formation of multinucleate myotubes and mus- (Suzuki et al., 1995). There is evidence to suspect that trans- cle differentiation (Kaushal et al., 1994). The efficiency of

/ m4070$8017 10-02-95 05:58:56 dba Dev Bio 6 Olson, Perry, and Schulz

FIG. 3. Myogenic bHLH proteins and MEF2 factors in the skeletal muscle lineage. Early mesodermal regulators induce the expression of myogenic bHLH factors during myoblast determination. Myogenic bHLH factors autoregulate their expression and induce the expression of MEF2, which binds the promoters of several myogenic bHLH genes, amplifying and maintaining their transcription. MEF2 and myogenic bHLH factors collaborate to induce muscle structural genes during differentiation. While MEF2 can be induced by myogenic bHLH factors in tissue culture, this has not yet been demonstrated in vivo. Other regulators might also initially induce MEF2 expression in skeletal muscle cells independent of myogenic bHLH factors. This schematic model is highly simplified and does not take into account differences in expression of individual myogenic bHLH and MEF2 factors during embryogenesis. myogenic conversion by MEF2A was similar to MyoD. promoter, for example, contains a MEF2 site that is required These results suggest that skeletal muscle determination for high level transcription in cultured muscle cells (Ed- and differentiation are not controlled by a linear genetic mondson et al., 1992; Buchberger et al., 1994). This site is pathway, but rather by two classes of regulators that regu- also required for expression of a myogenin–lacZ transgene late each others’ expression in a mutually reinforcing regu- in the limb buds and somites of transgenic mice (Cheng et latory network (Fig. 3). It should be pointed out that other al., 1993; Yee and Rigby, 1993). In the absence of this site, investigators have been unable to demonstrate stable con- transgene expression in the limb buds and the dorsal regions version of fibroblasts to differentiated muscle cells by forced of the somites is lost. However, myogenin transcription in expression of MEF2 (J. Martin, T. Firulli, and E. Olson, un- the ventral regions of the somites does not require the MEF2 published). The basis for these differences is unclear. site. These results suggest that Myogenin is expressed in The ability of MEF2 factors to regulate muscle gene ex- two distinct populations of muscle cells; one of which repression has also been examined in the Xenopus animal quires MEF2 and the other of which is independent of MEF2, cap system. Animal pole cells from blastula stage embryos for expression of myogenin. The existence of two popula- normally differentiate as ectoderm and neural tissues. Spec- tions of somitic muscle cell precursors has also been dem- ification of the developmental fate of these cells can be onstrated by somite transplantation experiments (Ordahl changed by exposure to growth factors that induce the dif- and Le Douarin, 1992). It is possible that the myogenin– ferentiation of mesodermal derivatives, such as muscle. Al- lacZ transgene containing a mutated MEF2 site is able to though forced expression of MEF2D in animal caps was distinguish between these two populations. Because MEF2 unable to induce expression of endogenous myogenic bHLH gene expression is initiated after myogenin expression in genes, activation of the endogenous cardiac-specific gene the somites and limb buds, it is likely that MEF2 partici- mlc-2 was detected in isolated animal pole explants dis- pates in an indirect autoregulatory loop to amplify and sected from early blastula stage embryos that were injected maintain myogenin gene expression, rather than initially as fertilized eggs with synthetic MEF2D transcripts (Cham- activating myogenin expression (Fig. 3). bers et al., 1994). This result is consistent with the expres- MEF2 has also been shown to regulate expression of the sion of MEF2D in cardiac muscle cell precursors and sup- Xenopus MyoDa gene. In this case, a consensus MEF2 bind- ports a role for MEF2 in the differentiation of cardiac muscle ing site overlaps precisely with the XMyoDa TATA box, cell lineages. Expression of the mlc-2 gene was activated the binding site for the multisubunit transcription factor by ectopic expression of MEF2D but not by expression of TFIID (Leibham et al., 1994). Binding of both factors to the MEF2A, suggesting that functional differences exist be- specialized TATA motif is required since transactivation is tween members of the MEF2 protein family. abolished by promoter mutations that selectively prevent binding of either factor. Activation of the XMyoDa pro- MEF2 Factors Regulate Some Myogenic bHLH moter by MEF2 requires only the MADS/MEF2 domains; Genes this activation is independent of a region toward the C- Several recent studies have revealed a role for MEF2 in terminus of MEF2 which is required to activate transcrip- the regulation of the myogenic bHLH genes. The myogenin tion of promoters with separate binding sites for MEF2 and

/ m4070$8017 10-02-95 05:58:56 dba Dev Bio MEF2 Family of MADS Box Transcription Factors 7

TFIID (Wong et al., 1994). Presumably, different types of analysis of MEF2 has been facilitated by the mutagenesis interactions occur between MEF2 and the transcriptional and phenotypic analysis of the single mef2 gene of Drosoph- machinery depending on whether MEF2 is bound to the ila, called D-mef2. The MADS and MEF2 domains of the TATA box or elsewhere. D-MEF2 protein share greater than 85% amino acid identity Since the DNA binding component of TFIID (TBP) inter- with the corresponding regions of the mammalian MEF2 acts with the minor groove and MEF2 with the major groove factors. Outside of these conserved motifs, D-MEF2 di- of DNA, both factors could, in principle, occupy the same verges significantly from the vertebrate proteins, but like site simultaneously. However, it has not yet been possible those factors it is rich in glutamine, serine, threonine, and to demonstrate simultaneous binding of recombinant TBP proline in its carboxyl-terminal region. D-MEF2 binds the and MEF2 to the XMyoDa TATA box in vitro. This might same DNA sequence as its vertebrate homologues and can reflect a requirement for additional components (TAFs?) activate transcription in Drosophila and mammalian cells necessary for a stable complex or it could indicate that bind- through the MEF2 binding site (Lilly et al., 1994; Nguyen ing of these factors to the XMyoDa TATA motif is mutually et al., 1994). Within the conserved region of the gene, the exclusive. In the latter case, prior binding of MEF2 might positions of D-mef2 introns map to the same codons as prevent inactivation of the promoter by chromatin assem- in the mammalian mef2 genes. Thus, this structural and bly or other inhibitory events. Because activation of XMyoD functional conservation suggests that the Drosophila and expression precedes that of MEF2 in early Xenopus em- mammalian mef2 genes evolved from a common ancestral bryos, MEF2 appears to function downstream of XMyoD in gene more than 600 million years ago. the myogenic pathway. The binding of MEF2 to the XMy- D-mef2 is expressed during embryogenesis in a profile oDa TATA box might constitute a simple and direct mecha- that is strikingly similar to the MEF2 expression patterns nism for stabilizing and amplifying XMyoDa expression in observed in early mouse development. Gene transcripts are differentiating muscle cells. MyoD genes from other species first detected in cells of the ventral furrow in the late blasto- appear to be regulated by different mechanisms than Xmy- derm/early gastrula and are restricted to the mesodermal oDa, since the mouse and chicken myoD genes do not con- cell layer during germ band extension (Lilly et al., 1994; tain MEF2 sites in their proximal promoters (Tapscott et Nguyen et al., 1994; Taylor et al., 1995). During the reorga- al., 1992; Dechesne et al., 1994). nization of the mesoderm during stage 10 and thereafter Binding of MEF2 to the TATA boxes of muscle-specific (Bate, 1993), there is a dynamic pattern of D-mef2 expres- genes may explain the observation that certain muscle sion in the external and internal mesoderm cell layers and genes require a specific TATA box for expression. Overlap- in the precursors of the heart. During germ band retraction ping binding sites for MEF2 and TBP are present in the and dorsal closure of the embryo, the muscles of the body promoters of the mouse and rat MRF4 genes (Naidu et al., wall, gut, and heart are formed. D-mef2 transcripts are pres- 1995; Black et al., 1995). Since MEF2 can also transactivate ent in all of these muscle types during their differentiation the MRF4 promoter upon binding to the TATA box, this into the final muscle structures. Immunolocalization exper- architecture appears to be biologically meaningful rather iments using a D-MEF2 antibody shows the expression of than fortuitous. The presence of overlapping binding sites a nuclear protein that faithfully follows the accumulation for MEF2 and TBP is not restricted to the promoters of of D-mef2 RNA in the mesoderm, muscle cell lineages, and myogenic regulatory genes. The myoglobin gene is effi- differentiated muscles (Fig. 4) (Lilly et al., 1995; Bour et al., ciently expressed in muscle cells only with its native TATA 1995). box, but not when that sequence is replaced with the TATA There is a lack of D-mef2 expression in twist mutant box of the SV40 promoter (Wefald et al., 1990). The myoglo- embryos and expression is severely reduced in snail mutant bin TATA box has been recently shown to bind MEF2 embryos, implicating these two transcription factors as (Grayson et al., 1995). Analogous situations exist with the probable regulators of early D-mef2 expression (Lilly et al., chick b-globin gene promoter, which is activated by binding 1994; Nguyen et al., 1994). D-mef2 appears to be a direct of the erythroid-specific protein cGATA-1 to the TATA mo- target of the Twist protein as ectopic expression of a twist tif (Fong and Emerson, 1992) and the pituitary specific regu- cDNA in the epidermis under the control of a heat shock latory gene PIT-1/GHF-1 TATA box, which binds a pitu- promoter results in the expression of D-mef2 RNA in the itary-specific factor (McCormick et al., 1990). These and same cells (Taylor et al., 1995). Intriguingly, D-mef2 is coex- other examples are illustrative of a general control mecha- pressed with tinman in the ventral furrow, undifferentiated nism involving the direct interaction of regulatory factors mesoderm, and cardiac muscle lineage in the Drosophila with specialized TATA boxes. embryo, just as the mammalian homologues of these genes are coexpressed in the precardiac mesoderm in the mouse embryo. However, the early mesodermal expression of D- Genetic Analysis of MEF2 Function in Drosophila mef2 is independent of tinman (Lilly et al., 1994; Nguyen Given the overlapping expression patterns and possible et al., 1994). In the future, it will be especially interesting redundancy of the vertebrate MEF2 genes, it may be difficult to determine the mechanisms that regulate D-mef2 expres- to address their functions by gene targeting without simul- sion in the mesoderm and three myogenic lineages. Prelimi- taneously inactivating multiple loci. However, the genetic nary analyses of D-mef2 regulatory sequences in Drosophila

/ m4070$8017 10-02-95 05:58:56 dba Dev Bio 8 Olson, Perry, and Schulz

FIG. 4. D-MEF2 expression in a stage 14 Drosophila embryo. A stage 14 Drosophila embryo was stained for D-MEF2 protein expression, ﬁleted along the ventral side, and ﬂattened beneath a cover slip. Expression of D-MEF2 can be seen in pharyngeal (pm) and somatic muscle (sm) cells, as well as in the forming dorsal vessel (dv).

germline transformant lines have identified separable en- tors tinman and bagpipe, which are required for the forma- hancer elements that properly drive the expression of a lacZ tion of the heart and visceral muscle (Azpiazu and Frasch, reporter gene in cells of the ventral furrow, mesoderm, and 1993; Bodmer, 1993), are expressed correctly in mutant em- somatic, visceral, and cardiac muscle lineages (Lilly et al., bryos. The formation of the somatic musculature in Dro- 1995; C. Chromey, G. Ranganayakulu, B. Zhao, E. Olson, sophila is believed to occur by the fusion of founder cells and R. Schulz, unpublished). with fusion-competent myoblasts (Rushton et al., 1995). To address the possible functions of D-mef2 in mesoderm Markers for the founder cells, such as the MyoD homolog differentiation and muscle development, P element inser- nautilus (Michelson et al., 1990; Paterson et al., 1991) and tional and ethyl methane sulfonate (EMS) chemical muta- the homeobox genes S59 (Dohrman et al., 1990) and apter- genesis screens have been carried out to recover lethal mu- ous (Bourgouin et al., 1992), are expressed in their normal tations in the gene. Embryos homozygous for a P-element- pattern in D-mef2 mutant embryos. Thus, the phenotype induced 25-kb chromosome deletion that removes essential of the D-mef2 deficiency embryos suggests that the gene D-mef2 regulatory sequences express the D-MEF2 protein acts at a relatively late stage within the different myogenic at very low levels and fail to form normal differentiated lineages to control cell differentiation (Fig. 5). muscles (Lilly et al., 1995). The use of molecular markers A more detailed understanding of the role of D-mef2 in for the precursors of the somatic, visceral, and cardiac mus- muscle cell differentiation was obtained through the analy- cles showed that muscle precursor cells were specified and sis of mutants containing point mutations in the gene. The positioned normally. For example, the mesodermal regula- phenotypic analysis of embryos homozygous for EMS-in-

/ m4070$8017 10-02-95 05:58:56 dba Dev Bio MEF2 Family of MADS Box Transcription Factors 9

FIG. 5. Myogenic lineages in Drosophila. Schematic representation of the three myogenic lineages from Drosophila and the genes that are expressed in those lineages. D-mef2 and tinman are coexpressed in the uncommitted mesoderm; both genes are dependent on twist and snail for expression. The prospective mesoderm gives rise to the somatic, visceral, and cardiac muscle lineages, all of which express D-mef2. tinman is expressed initially in the cardiac and visceral muscle lineages, but ultimately becomes restricted to the dorsal vessel. In D-mef2 mutant embryos, tinman is expressed normally. Formation of somatic muscles is believed to occur by fusion of founder cells with fusion-competent myoblasts. Both of these cell types express D-MEF2. In D-mef2 mutant embryos, nautilus and the homeobox genes S59 and apterous are expressed normally in founder cells, but there is no fusion. Activation of muscle structural genes in all three lineages is dependent on D-MEF2.

duced null (Bour et al., 1995) or severe loss-of-function (Ran- MEF2 for their differentiation. Alternatively, D-MEF2 may ganayakulu et al., 1995) alleles of D-mef2 revealed the par- interact with different sets of cofactors in different somatic tial differentiation of somatic muscle precursors. However, muscle precursors and mutations in the D-MEF2 protein an absence of myoblast cell fusion and muscle fiber forma- could selectively disrupt interactions with some cofactors tion was observed, and a dramatic decrease in the myoblast and not with others, resulting in the absence of specific population followed due to programmed cell death. Such a muscles. Whereas weak alleles of D-mef2 result in selective phenotype is consistent with the ability of the Drosophila ablation of certain somatic muscle fibers, but not others, embryo to eliminate cells that have failed to complete their even relatively weak alleles that have only minor effects normal differentiation program (Abrams et al., 1993). In- on the somatic musculature completely eliminate certain triguingly, apoptosis is restricted to the somatic muscle aspects of the differentiation of cardiac and visceral muscle population and is not observed in cardiac or visceral muscle cells. This suggests that these muscle cell types are more precursors. sensitive to MEF2 activity than somatic muscle cells. One Further studies on embryos expressing hypomorphic D- combination of D-mef2 alleles allows for the occasional mef2 alleles has revealed novel functions of the gene that survival of transheterozygous mutant adults. These adults could not be resolved with complete loss-of-function alleles are flightless and exhibit severe defects in the patterning (Ranganayakulu et al., 1995). In particular, certain somatic and organization of the indirect flight muscles present in muscles are selectively lost in embryos with partial D- the thorax. Therefore, D-MEF2 is required for both the lar- MEF2 activity arising from truncated D-MEF2 proteins, in- val and adult myogenic programs. D-MEF2 is expressed in dicating that D-mef2 is required for both the formation and adepithelial cells which are the precursors of the adult tho- patterning of body wall muscle in the Drosophila embryo. racic muscles. Given the simultaneous expression of Twist This could be explained if the myoblasts that normally form (Bate et al., 1991) and D-MEF2 in these cells, and the recent the muscles that are absent require a higher threshold of D- demonstration of the cooperativity of the myogenic bHLH

/ m4070$8017 10-02-95 05:58:56 dba Dev Bio 10 Olson, Perry, and Schulz and MEF2 proteins in regulating skeletal muscle gene ex- tion with other regulatory factors. SRF, for example, inter- pression (Kaushal et al., 1994; Naidu et al., 1995), Twist acts with a group of ETS-domain proteins, referred to as and D-MEF2 may interact in the control of gene expression ternary complex factors (TCFs), which bind a site adjacent in adult muscle differentiation. to the serum response element in the c-fos promoter (Treis- Information is emerging on potential targets of D-MEF2 man, 1994). This interaction is dependent on a region adja- transcriptional activity in the different muscle cell lineages. cent to the dimerization domain of SRF. DNA binding by In the cardiac lineage, the dorsal vessel is formed, yet myo- SRF has also been shown to be potentiated by interaction sin protein is not detected (Lilly et al., 1995; Bour et al., with the homeodomain protein Phox, also called MHox. 1995; Ranganayakulu et al., 1995). The individual mhc, Enhanced DNA binding by SRF in the presence of Phox is mlc-alk, and mlc2 subunit genes are not expressed in the believed to occur by acceleration of the on-rate of DNA dorsal vessel of mutant embryos and thus serve as likely binding (Grueneberg et al., 1992). It has not been possible targets of D-MEF2 (Ranganayakulu et al., 1995). Both ge- to demonstrate direct physical interaction between SRF and netic and molecular studies implicate the aPS2 integrin sub- Phox, so it is unclear precisely how this effect is achieved. unit gene as a target of D-MEF2 in visceral muscle. Null MCM1 has been shown to regulate cell-type-specific gene mutations in the inflated locus (Brabant and Brower, 1993; expression in the budding yeast Saccharomyces cerevisiae Brown, 1994), which encodes this integrin subunit, result through interaction with accessory proteins (Herskowitz, in an identical midgut phenotype to that observed in D- 1989). The identities of the two haploid cell types in yeast, mef2 mutant embryos. This common phenotype correlates a and a, require the activation and repression of distinct with the lack of aPS2 gene expression and muscle-specific sets of genes. MCM1 controls cell-type-specific genes in enhancer function in the absence of D-mef2 function (Ran- yeast by acting in conjunction with two coregulators, a1 ganayakulu et al., 1995). Given the presence of a high-affin- and a2, which are expressed only in a cells. In a cells, ity D-MEF2 site in the aPS2 enhancer, it is probable that this MCM1 inhibits the expression of a-specific genes by binding essential integrin subunit gene is a direct target of D-MEF2 their promoters in collaboration with the transcriptional transcriptional activation in visceral muscle. repressor a2. a-Specific genes are activated in the same cells Future challenges in the study of the D-mef2 gene will by MCM1 in collaboration with the homeodomain protein include the analysis of its complex regulation, more detailed a1. Neither the a2 repressor nor the a1 activator are present studies on its function in the larval and adult myogenic in a cells; a-specific genes are therefore not expressed, even programs, and the identification of D-MEF2 target genes. though MCM1 is present. Thus, the response of a gene to The use of Drosophila genetics will be essential for pre- MCM1 is dictated not simply by the presence of an MCM1 cisely determining the position and function of this MEF2 binding site, but also by the sites surrounding that site, as family member in the genetic hierarchy controlling meso- well as the affinity of MCM1 for that site. derm differentiation and muscle development. Our current MCM1 also mediates the actions of pheromones in yeast, knowledge of the genes and events involved in this complex which act through the cell surface receptors STE2 and STE3. process is summarized in Fig. 5. Transcription activation of pheromone-responsive genes is controlled by MCM1 and the transcriptional regulator STE12, which form a complex on the promoters of these Regulation of Cell-Specific Gene Expression by genes (Dolan and Fields, 1991). Activation of pheromone- MADS Box Proteins responsive genes leads to cell cycle arrest and morphological An important unanswered question is how MEF2 can reg- changes prior to cell fusion. Expression of the SW15 gene, ulate muscle gene expression in multiple myogenic lin- which encodes a cell-type-specific transcription factor re- eages, which express overlapping but distinct subsets of quired for mating type switching, is also controlled by muscle-specific genes. The mlc-2A gene, for example, is MCM1 (Lydall et al., 1991). Transcriptional activation of controlled by MEF2 and is expressed in ventricular cardiac SWI5 by MCM1 is mediated by the cooperative interaction myocytes, but not in fast skeletal muscle (Lee et al., 1992). of MCM1 with another regulatory protein SFF to form a Similarly, the myogenin gene is expressed in skeletal, but transcriptional complex on the SWI5 gene control region. not in cardiac or smooth muscle cells (Edmondson and The functions of MADS box proteins in regulating cell-type- Olson, 1989). Thus, the ability of MEF2 to activate its target specific gene expression have also been well characterized genes must be influenced by a cell’s identity or develop- in plants, in which combinations of MADS box proteins mental history. act within a genetic network that establishes floral organ In thinking about how MEF2 might regulate muscle gene identity (Weigel and Meyerowitz, 1994). expression in multiple muscle cell types, it is useful to Given the propensity of MADS box proteins to act combi- consider the mechanisms whereby other MADS box pro- natorially with other classes of transactivators, it is tempt- teins regulate programs of cell-specific gene expression. ing to speculate that MEF2 factors may also interact with There are many examples in which MADS box proteins act other regulators to control muscle gene expression. Indeed, at the endpoints of signal transduction pathways to regulate MEF2 has been shown to bind DNA cooperatively with inducible genes and to confer cell identity. In most cases, myogenic bHLH proteins, resulting in synergistic activation MADS box proteins exert these activities through coopera- of muscle-specific transcription (Funk and Wright, 1992).

/ m4070$8017 10-02-95 05:58:56 dba Dev Bio MEF2 Family of MADS Box Transcription Factors 11

The possibility that myogenic bHLH proteins influence the pression of the c-jun gene in HeLa cells (Han et al., 1992; activity of MEF2 is suggested by the finding that MEF2A Han and Prywes, 1995). The possibility that MEF2 might physically interacts with Myogenin and MyoD (Kaushal et confer serum inducibility to certain genes is particularly al., 1994). This interaction is mediated by the MADS box interesting in light of the role of SRF in the control of mus- of MEF2A and the basic region and first helix of MyoD. cle and serum-inducible gene expression (Treisman, 1990). MyoD interacts specifically with the MADS box of MEF2 Perhaps these different MADS domain proteins rely on sim- and not with that of SRF. Activation of muscle-specific tran- ilar mechanisms to control these seemingly disparate forms scription by MyoD is dependent on two residues (alanine- of gene expression. threonine) in the center of the basic region and a third resi- due (lysine) in the junction between the basic region and Future Questions helix-1. Replacement of these residues with the residues at the corresponding positions in E12 abolishes the ability to In summary, there is convincing evidence in support of activate muscle transcription without affecting DNA bind- a role for MEF2 factors in the control of muscle gene expres- ing. Conversely, if these residues are introduced into E12, sion in multiple muscle cell types. However, many im- they confer the ability to induce myogenesis (Davis and portant questions remain to be answered. In particular, if Weintraub, 1992). MEF2 fails to interact with E12 or with MEF2 controls differentiation of diverse muscle cell types as MyoD mutants lacking the three residues critical for myo- the genetic studies in Drosophila indicate, what determines genesis, but it can interact with an E12 mutant containing whether a cell is skeletal, cardiac, or smooth muscle? We the three myogenic residues (Kaushal et al., 1994). Thus, the favor a model in which MEF2 provides a function that is interaction between MEF2 and bHLH proteins correlates essential for myogenesis in general and in which the addi- precisely with myogenic activity. Since myogenic bHLH tional specificity that is unique to each myogenic lineage proteins are not expressed in cardiac or smooth muscle, it arises from combinatorial interactions between MEF2 and will be interesting to determine the identities of the factors other regulators that are restricted to each myogenic lin- with which MEF2 interacts in those cell types to activate eage. In skeletal muscle cells, it is likely that MEF2 cooper- muscle gene expression. ates with myogenic bHLH proteins to establish the skeletal Other factors have also been shown to mediate MEF2- muscle phenotype. Whether similar bHLH proteins are ex- dependent activation of muscle gene expression. A DNA pressed in cardiac and smooth muscle cells remains to be binding activity referred to as MAF1, for example, binds a determined, but it is likely that MEF2 has a conserved, MEF3 motif in the aldolase A gene promoter and is required common function in the differentiation of these different for activation of the gene by MEF2 (Hidaka et al., 1993). muscle cell lineages. Given the many parallels between The MEF3 motif is also present in the promoters of the muscle and neural differentiation (Jan and Jan, 1993), it will cardiac troponin C and myogenin genes. Activation of mlc- also be interesting to determine whether MEF2 collaborates 2 transcription in cardiac myocytes also requires synergism with neurogenic bHLH factors such as NeuroD (Lee et al., between MEF2 and a ubiquitous factor, HF-1a, which binds 1995) to regulate neural differentiation. an adjacent site in the promoter (Navankasattusas et al., MEF2 expression in skeletal muscle cells is also induced 1992). by myogenic bHLH proteins. It will be of interest to determine the mechanisms that lead to MEF2 expression in other muscle cell types and to determine to what extent the func- Functions for MEF2 Factors in Addition to Muscle tions and position of MEF2 in the regulatory pathways lead- Gene Regulation ing to muscle formation have been conserved from Dro- In addition to their expression in myogenic lineages, sophila to mammals. members of the MEF2 family exhibit highly specific expression patterns in the developing brain and appear to demar- ACKNOWLEDGMENTS cate gradients of neuronal maturation. MEF2 gene expression is especially pronounced in the cerebellum, cerebral We are grateful to B. Lilly for the photograph of the D-MEF2- cortex, and hippocampus (Lyons et al., 1995). The functions stained Drosophila embryo. We thank K. Tucker for editorial assis- of MEF2 in neural cells remain to be determined, and no tance. The authors are supported by grants from the National Insti- target genes for MEF2 regulation have been identified in tutes of Health (E.N.O. and M.P.), the Muscular Dystrophy Associa- neural cells thus far. The use of Drosophila genetics may tion (E.N.O., M.P., and R.A.S.), the National Science Foundation allow for the elucidation of the function of MEF2 in neu- (R.A.S.), and the Human Frontier Science Program (E.N.O. and ronal cell differentiation. D-MEF2 is expressed in specific R.A.S.). neurons of the central nervous system (G. Ranganayakulu and R. Schulz, unpublished) and genetic approaches exist REFERENCES to generate mosaic animals that have wild-type gene function in muscle cells, yet are mutant for D-mef2 function in Abrams, J. M., White, K., Fessler, L. I., and Steller, H. (1993). Pro- neuronal cells. grammed cell death during Drosophila embryogenesis. Develop- MEF2 has also been implicated in serum-inducible ex- ment 117, 29 –43.

/ m4070$8017 10-02-95 05:58:56 dba Dev Bio 12 Olson, Perry, and Schulz

Azpiazu, N., and Frasch, M. (1993). tinman and bagpipe: Two ho- specific enhancer binding factor MEF-2 independently of other meo box genes that determine cell fates in the dorsal mesoderm muscle-specific gene products. Mol. Cell. Biol. 11, 4854–4862. of Drosophila. Genes Dev. 7, 1325–1340. Davis, R. L., and Weintraub, H. (1992). Acquisition of myogenic Bate, M. (1993). ‘‘The Development of Drosophila melanogaster’’ specificity by replacement of three amino residues from MyoD (M. Bate and A. Martinez Arias, Eds.). Cold Spring Harbor Labora- into E12. Science 256, 1027–1030. tory Press, Cold Spring Harbor, NY. Dechesne, C. A., Wei, Q., Eldridge, J., Gannoun-Zaki, L., Millas- Bate, M., Rushton, E., and Currie, D. A. (1991). Cells with persistent seau, P., Bougueleret, L., Caterina, D., and Paterson, B. (1994). twist expression are the embryonic precursors of adult muscles E-box and MEF2-independent muscle-specific gene expression, in Drosophila. Development 113, 79–89. positive autoregulation, and cross-activation of the chicken Black, B. L., Martin, J. F., and Olson, E. N. (1995). The mouse MRF4 myoD1 (CMD1) promoter reveal an indirect regulatory pathway. promoter is trans-activated directly and indirectly by muscle- Mol. Cell. Biol. 14, 5474–5486. specific transcription factors. J. Biol. Chem. 270, 2889–2892. Dohrman, C., Azpiazu, N., and Frasch, M. (1990). A new Drosophila Bodmer, R. (1993). The gene tinman is required for specification of homeobox gene is expressed in mesodermal precursor cells of the heart and visceral muscles in Drosophila. Development 118, distinct muscles during embryogenesis. Genes Dev. 4, 2098– 719–729. 2111. Bodmer, R. (1995). Heart development in Drosophila and its rela- Dolan, J. W., and Fields, S. (1991). Cell-type specific transcription tionship to vertebrates. Trends Cardiovasc. Med. 5, 21–28. in yeast. Biochim. Biophys. Acta 1088, 155–169. Bour, B. A., O’Brien, M. A., Lockwood, W. L., Goldstein, E. S., Edmondson, D. G., Cheng, T.-C., Cserjesi, P., Chakraborty, T., and Bodmer, R., Taghert, P. H., Abmayr, S. M., and Nguyen, H. T. Olson, E. N. (1992). Analysis of the myogenin promoter reveals (1995). Drosophila MEF2, a transcription factor that is essential an indirect pathway for positive autoregulation mediated by the for myogenesis. Genes Dev. 9, 730–741. muscle-specific enhancer factor MEF-2. Mol. Cell. Biol. 12, 3665– Bourgouin, C., Lundgren, S. E., and Thomas, J. B. (1992). apterous 3677. is a Drosophila LIM domain gene required for the development Edmondson, D. G., Lyons, G. E., Martin, J. F., and Olson, E. N. of a subset of embryonic muscles. Neuron 9, 549–561. (1994). MEF2 gene expression marks the cardiac and skeletal Brabant, M. C., and Brower, D. L. (1993). PS2 integrin requirements muscle lineages during mouse embryogenesis. Development 120, in Drosophila embryo and wing morphogenesis. Dev. Biol. 157, 1251–1263. 49–59. Edmondson, D. G., and Olson, E. N. (1989). A gene with homology Braun, T., Rudnicki, M. A., Arnold, H. H., and Jaenisch, R. (1992). to the myc similarity region of MyoD1 is expressed during myo- Targeted inactivation of the muscle regulatory gene Myf-5 results genesis and is sufficient to activate the muscle differentiation in abnormal rib development and perinatal death. Cell 71, 369– program. Genes Dev. 3, 628 –640. 382. Fong, T. C., and Emerson, B. M. (1992). The erythroid-specific pro- Braun, T., Tannich, E., Buschhausen-Denker, G., and Arnold, tein cGATA-1 mediates distal enhancer activity through a spe- H. H. (1989). Promoter upstream elements of the chicken cardiac cialized b-globin TATA box. Genes Dev. 6, 521–532. myosin light chain 2-A gene interact with trans-acting regulatory Funk, W. D., and Wright, W. E. (1992). Cyclic amplification and factors for muscle-specfiic transcription. Mol. Cell. Biol. 9, 2513– selection of targets for multicomponent complexes: Myogenin 2525. interacts with factors recognizing binding sites for basic helix- Breitbart, R. E., Liang, C.-S., Smoot, L. B., Laheru, D. A., Mahdavi, loop-helix, nuclear factor 1, myocyte-specific enhancer-binding V., and Nadal-Ginard, B. (1993). A fourth human MEF2 transcrip- factor 2, and COMP1 factor. Proc. Natl. Acad. Sci. USA 89, 9484– tion factor, hMEF2D, is an early marker of the embryonic lineage. 9488. Development 118, 1095–1106. Gossett, L. A., Kelvin, D. J., Sternberg, E. A., and Olson, E. N.

Brown, N. H. (1994). Null mutations in the aPS2 and bPS integrin (1989). A new myocyte-specific enhancer-binding factor that rec- subunit genes have distinct phenotypes. Development 120, ognizes a conserved element associated with multiple muscle- 1221–1231. specific genes. Mol. Cell. Biol. 9, 5022–5033. Buchberger, A., Ragge, K., and Arnold, H. H. (1994). The myogenin Grayson, J., Williams, R. S., Yu, Y.-T., and Bassel-Duby, R. (1995). gene is activated during myocyte differentiation by pre-existing, Synergistic interactions between heterologous upstream activa- not newly synthesized transcription factor MEF-2. J. Biol. Chem. tion elements and specific TATA sequences in a muscle-specific 269, 17289–17296. promoter. Mol. Cell. Biol. 15, 1870–1878. Buckingham, M. (1992). Making muscle in mammals. Trends Grueneberg, D. A., Natesan, S., Alexandre, C., and Gilman, Genet. 8, 144–149. M. Z. (1992). Human and Drosophila homeodomain proteins that Chambers, A. E., Logan, M., Kotecha, S., Towers, N., Sparrow, D., enhance the DNA-binding activity of serum response factor. Sci- and Mohun, T. J. (1994). The RSRF/MEF2 protein SL1 regulates ence 257, 1089–1095. cardiac muscle-specific transcription of a myosin light-chain Han, T.-H., Lamph, W. W., and Prywes, R. (1992). Mapping of epi- gene in Xenopus embryos. Genes. Dev. 8, 1324–1334. dermal growth factor-, serum-, and phorbol ester-responsive se- Cheng, T.-C., Wallace, M., Merlie, J. P., and Olson, E. N. (1993). quence elements in the c-jun promoter. Mol. Cell. Biol. 12, 4472– Separable regulatory elements govern myogenin transcription in 4477. embryonic somites and limb buds. Science 261, 215–218. Han, T.-H., and Prywes, R. (1995). Regulatory role of MEF2D in Cserjesi, P., Lilly, B., Hinkley, C., Perry, M., and Olson, E. N. (1994). serum induction of the c-jun promoter. Mol. Cell. Biol. 15, 2907– Homeodomain protein MHox and MADS protein myocyte en- 2915. hancer-binding factor-2 converge on a common element in the Hasty, P., Bradley, A., Morris, J. H., Edmondson, D. G., Venuti, J., muscle creatine kinase enhancer. J. Biol. Chem. 269, 16740– Olson, E. N., and Klein, W. H. (1993). Muscle deficiency and 16745. neonatal death in mice with a targeted mutation in the myogenin Cserjesi, P., and Olson, E. N. (1991). Myogenin induces muscle- gene. Nature 364, 501–506.

/ m4070$8017 10-02-95 05:58:56 dba Dev Bio MEF2 Family of MADS Box Transcription Factors 13

Hauschka, S. D. (1994). The embryonic origin of muscle. In ‘‘Myol- (1993). Nkx-2.5: A novel murine homeobox gene expressed in ogy, Basic and Clinical’’ (A. G. Engel and C. Franzini-Armstrong, early heart progenitor cells and their myogenic descendants. De- Eds.), 2nd ed., Vol. 1, pp. 3 –73. McGraw-Hill, New York. velopment 119, 419–431. Herskowitz, I. (1989). A regulatory hierarchy for cell specialization Lydall, D., Ammerer, G., and Nasmyth, K. (1991). A new role for in yeast. Nature 342, 749–757. MCM1 in yeast: Cell cycle regulation of SW15 transcription. Hidaka, K., Yamamoto, I., Arai, Y., and Mukai, T. (1993). The MEF- Genes Dev. 5, 2405–2419. 3 motif is required for MEF-2-mediated skeletal muscle-specific Lyons, G. E., Micales, B. K., Schwarz, J., Martin, J. F., and Olson, induction of the rat aldolase A gene. Mol. Cell. Biol. 13, 6469– E. N. (1995). Expression of mef2 genes in the mouse central ner- 6478. vous system suggests a role in neuronal maturation. J. Neurosci. Horlick, R. A., Hobson, G. M., Patterson, J. H., Mitchell, M. T., 15, 5727–5738. and Benfield, P. A. (1990). Brain and muscle creatine kinase genes Martin, J. F., Miano, J. M., Hustad, C. M., Copeland, N. G., Jenkins, contain common TA-rich recognition protein-binding regulatory N. A., and Olson, E. N. (1994). A Mef2 gene that generates a elements. Mol. Cell. Biol. 10, 4826–4836. muscle-specific isoform via alternative mRNA splicing. Mol. Ianello, R. C., Mar, J. H., and Ordahl, C. P. (1991). Characterization Cell. Biol. 14, 1647–1656. of a promoter element required for transcription in myocardial Martin, J. F., Schwarz, J. J., and Olson, E. N. (1993). Myocyte en- cells. J. Biol. Chem. 266, 3309–3316. hancer factor (MEF) 2C: A tissue-restricted member of the MEF- Jan, Y. N., and Jan, L. Y. (1993). HLH proteins, fly neurogenesis, 2 family of transcription factors. Proc. Natl. Acad. Sci. USA 90, and vertebrate myogenesis. Cell 75, 827 –830. 5282–5286. Kaushal, S., Schneider, J. W., Nadal-Ginard, B., and Mahdavi, V. McCormick, A., Brady, H., Fukushima, J., and Karin, M. (1990). (1994). Activation of the myogenic lineage by MEF2A, a factor The pituitary-specific regulatory gene GHF-1 contains a minimal that induces and cooperates with MyoD. Science 266, 1236– cell type-specific promoter centered around its TATA box. Genes 1240. Dev. 5, 1490–1503. Komuro, I., and Izumo, S. (1993). Csx: A murine homeobox-con- McDermott, J. C., Cardoso, M. C., Yu, Y., Andres, V., Leifer, D., taining gene specifically expressed in the developing heart. Proc. Krainc, D., Lipton, S. A., and Nadal-Ginard, B. (1993). hMEF2C Natl. Acad. Sci. USA 90, 8145–8149. gene encodes skeletal muscle- and brain-specific transcription Lassar, A. B., Davis, R. L., Wright, W. E., Kadesch, T., Murre, C., factors. Mol. Cell. Biol. 13, 2564–2577. Voronova, A., Baltimore, D., and Weintraub, H. (1991). Func- Miano, J. M., Cserjesi, P., Ligon, K. L., Periasamy, M., and Olson, tional activity of myogenic HLH proteins requires hetero-oligo- E. N. (1994). Smooth muscle myosin heavy chain exclusively merization with E12/E47-like proteins in vivo. Cell 66, 305 –315. marks the smooth muscle lineage during mouse embryogenesis. Lee, J. E., Hollenberg, S. M., Snider, L., Turner, D. L., Lipnick, N., Circ. Res. 75, 803–812. and Weintraub, H. (1995). Conversion of Xenopus ectoderm into Michelson, A. M., Abmayr, S. M., Bate, M., Martinez Arias, A., neurons by NeuroD, a basic helix-loop-helix protein. Science 268, and Maniatis, T. (1990). Expression of a MyoD family member 836–844. prefigures muscle pattern in Drosophila embryos. Genes Dev. 4, Lee, K. J., Ross, R. S., Rockman, H. A., Harris, A. N., O’Brien, 2086–2097. T. X., van Bilson, M., Shubeita, H. E., Kandolf, R., Brem, G., Molkentin, J. D., and Markham, B. E. (1993). Myocyte-specific en- Price, J., Evans, S. M., Zhu, H., Franz, W.-M., and Chien, K. R. hancer-binding factor (MEF-2) regulates alpha-cardiac myosin (1992). Myosin light chain-2 luciferase transgenic mice reveal heavy chain gene expression in vitro and in vivo. J. Biol. Chem. distinct regulatory programs for cardiac and skeletal muscle-spe- 268, 19512–19520. cific expression of a single contractile protein gene. J. Biol. Chem. Molkentin, J. D., Black, B., Martin, J. F., and Olson, E. N. (1995). 267, 15875–15885. Submitted. Leibham, D., Wong, M.-W., Cheng, T.-C., Schroeder, S., Weil, Morisaki, T., and Holmes, E. W. (1993). Functionally distinct ele- P. A., Olson, E. N., and Perry, M. (1994). Binding of TFIID and ments are required for expression of the AMPD1 genes in myo- MEF2 to the TATA element activates transcription of the Xeno- cytes. Mol. Cell. Biol. 13, 5854–5860. pus MyoDa promoter. Mol. Cell. Biol. 14, 686–699. Murre, C., McCaw, P. S., Vaessin, H., Caudy, M., Jan, L. Y., Jan, Leifer, D., Krainc, D., Yu, Y.-T., McDermott, J. C., Breitbart, R. E., Y. N., Cabrera, C. V., Buskin, J. N., Hauschka, S. D., Lassar, Heng, J., Neve, R. L., Kosofsky, B., Nadal-Ginard, B., and Lipton, A. B., Weintraub, H., and Baltimore, D. (1989). Interactions be- S. A. (1993). MEF2C, a MADS/MEF2-family transcription factor tween heterologous helix-loop-helix proteins generate complexes expressed in a laminar distribution in cerebral cortex. Proc. Natl. that bind specifically to a common DNA sequence. Cell 58, 537– Acad. Sci. USA 90, 1546–1550. 544. Li, H., and Capetanaki, Y. (1994). An E box in the desmin promoter Muscat, G. E. O., Perry, S., Prentice, H., and Kedes, L. (1992). The cooperates with the E box and MEF-2 sites of a distal enhancer human skeletal alpha-actin gene is regulated by a muscle-specific to direct muscle-specific transcription. EMBO J. 13, 3580–3589. enhancer that binds three nuclear factors. Gene Expression 2, Lilly, B., Galewsky, S., Firulli, A. B., Schulz, R. A., and Olson, 111–126. E. N. (1994). D-MEF2: A MADS box transcription factor ex- Nabeshima, Y., Hanaoka, K., Hayasaka, M., Esumi, E., Li, S., Non- pressed in differentiating mesoderm and muscle cell lineages dur- aka, I., and Nabeshima, Y. (1993). Myogenin gene disruption re- ing Drosophila embryogenesis. Proc. Natl. Acad. Sci. USA 91, sults in perinatal lethality because of severe muscle defect. Na- 5662–5666. ture 364, 532–535. Lilly, B., Zhao, B., Ranganayakulu, G., Paterson, B. M., Schulz, Naidu, P. S., Ludolph, D. C., To, R. Q., Hinterberger, T. J., and R. A., and Olson, E. N. (1995). Requirement of MADS domain Konieczny, S. F. (1995). Myogenin and MEF2 function synergisti- transcription factor D-MEF2 for muscle formation in Drosophila. cally to activate the MRF4 promoter during myogenesis. Mol. Science 267, 688–693. Cell. Biol. 15, 2707–2718. Lints, T. J., Parsons, L. M., Hartley, L., Lyons, I., and Harvey, R. P. Nakatsuji, Y., Hidaka, K., Tsujino, S., Yamamoto, Y., Mukai, T.,

/ m4070$8017 10-02-95 05:58:56 dba Dev Bio 14 Olson, Perry, and Schulz

Yanagihara, T., Kishimoto, T., and Sakoda, S. (1992). A single Tapscott, S. J., Lassar, A. B., and Weintraub, H. (1992). A novel MEF-2 site is a major positive regulatory element required for myoblast enhancer element mediates MyoD transcription. Mol. transcription of the muscle-specific subunit of the human phos- Cell. Biol. 12, 4994–5003. phoglycerate mutase gene in skeletal and cardiac muscle cells. Taylor, M. V., Beatty, K. E., Hunter, H. K., and Baylies, M. K. (1995). Mol. Cell. Biol. 12, 4384–4390. Drosophila MEF2 is regulated by twist and is expressed in both Navankasattusas, S., Zhu, H., Garcia, A. V., Evans, S. M., and the primordia and differentiated cells of the embryonic somatic, Chien, K. R. (1992). A ubiquitous factor (HF-1a) and a distinct visceral and heart musculature. Mech. Dev. 50, 29 –41. muscle factor (HF-1b/MEF-2) form an E-box-independent path- Treisman, R. (1990). The serum response element. Trends Bio- way for cardiac muscle gene expression. Mol. Cell. Biol. 12, chem. Sci. 17, 423–426. 1469–1479. Treisman, R. (1994). Ternary complex factors: Growth factor regu- Nguyen, H. T., Bodmer, R., Abmayr, S. M., McDermott, J. C., and lated transcriptional activators. Curr. Opin. Genet. Dev. 4, 96– Spoerel, N. A. (1994). D-mef2: A Drosophila mesoderm-specific 101. MADS box-containing gene with a biphasic expression profile Wachtler, F., and Christ, B. (1992). The basic embryology of skeletal during embryogenesis. Proc. Natl. Acad. Sci. USA 91, 7520– muscle formation in vertebrates: The avian model. Dev. Biol. 3, 7524. 217–227. Olson, E. N. (1990). MyoD family, a paradigm for development? Wang, G., Yeh, H.-I., and Lin, J. J.-C. (1994). Characterization of Genes Dev. 4, 1454–1461. cis-regulating elements and trans-activating factors of the rat Olson, E. N., and Klein, W. H. (1994). bHLH factors in muscle cardiac troponin T gene. J. Biol. Chem. 269, 30595–30603. development: Dead lines and commitments, what to leave in and Wefald, F. C., Devlin, B. H., and Williams, R. S. (1990). Functional what to leave out. Genes Dev. 8, 1–8. heterogeneity of mammalian TATA-box sequences revealed by Ordahl, C. P., and Le Douarin, N. M. (1992). Two myogenic lineages interaction with a cell-specific enhancer. Nature 344, 260 –262. within the developing somite. Development 114, 339 –353. Weigel, D., and Meyerowitz, E. M. (1994). The ABCs of floral home- Parmacek, M. S., Ip, H., Jung, F., Shen, T., Martin, J. F., Vora, otic genes. Cell 78, 203–209. A. J., Olson, E. N., and Leiden, J. M. (1994). A novel myogenic Weintraub, H., Davis, R., Tapscott, S., Thayer, M., Krause, M., regulatory circuit controls slow/cardiac troponin C gene tran- Benezra, R., Blackwell, T. K., Turner, D., Rupp, R., Hollenberg, scription in skeletal muscle. Mol. Cell. Biol. 14, 1870–1885. S., Zhuang, Y., and Lassar, A. (1991). The myoD gene family: Paterson, B. M., Walldorf, U., Eldridge, J., Dubendorfer, A., Frasch, Nodal point during specification of the muscle cell lineage. Sci- M., and Gehring, W. J. (1991). The Drosophila homologue of ence 251, 761–766. vertebrate myogenic-determination genes encodes a transiently Wentworth, B. M., Donoghue, M., Engert, J. C., Berglund, E. B., and expressed protein marking primary myogenic cells. Proc. Natl. Rosenthal, N. (1991). Paired MyoD binding sites regulate myosin Acad. Sci. USA 88, 3782–3786. light chain gene expression. Proc. Natl. Acad. Sci. USA 88, 1242– Pollock, R., and Treisman, R. (1991). Human SRF-related proteins: 1246. DNA-binding properties and potential regulatory targets. Genes Wong, M.-W., Pisegna, M., Lu, M.-F., Leibham, D., and Perry, M. Dev. 5, 2327–2341. (1994). Activation of Xenopus MyoD transcription by members Ranganayakulu, G., Zhao, B., Dokidis, A., Molkentin, J. D., Olson, of the MEF2 protein family. Dev. Biol. 166, 683–695. E. N., and Schulz, R. A. (1995). A series of mutations in the Yee, S. P., and Rigby, P. W. (1993). The regulation of myogenin D-MEF2 transcription factor reveal multiple functions in larval gene expression during embryonic development of the mouse. and adult myogenesis in Drosophila. Dev. Biol. 171, 169 –181. Genes Dev. 7, 1277–1289. Rudnicki, M. A., Braun, T., Hinuma, S., and Jaenisch, R. (1992). Yu, Y. T., Breitbart, R. E., Smooth, L. B., Lee, Y., Mahdavi, V., Inactivation of MyoD in mice leads to up-regulation of the myo- and Nadal-Ginard, B. (1992). Human myocyte-specific enhancer genic HLH gene Myf-5 and results in apparently normal muscle factor 2 comprises a group of tissue-restricted MADS box tran- development. Cell 71, 383–390. scription factors. Genes Dev. 6, 1783–1798. Rudnicki, M. A., and Jaenisch, R. (1995). The MyoD family of tran- Zhang, W., Behringer, R. R., and Olson, E. N. (1995). Inactivation scription factors and skeletal myogenesis. BioEssays 17, 203– of the myogenic bHLH gene MRF4 results in upregulation of 209. myogenin and rib anomalies. Genes Dev. 9, 1388–1399. Rudnicki, M. A., Schnegelsberg, P. N. J., Stead, R. H., Braun, T., Zhou, M. D., Goswami, S. K., Martin, M. E., and Siddiqui, M. A. Arnold, H. H., and Jaenisch, R. (1993). MyoD or Myf-5 is required (1993). A new serum-responsive, cardiac tissue-specific transcrip- for the formation of skeletal muscle. Cell 75, 1351–1359. tion factor that recognizes the MEF2-site in the myosin light Rushton, E., Drysdale, R., Abmayr, S., Michaelson, A., and Bate, chain-2 promoter. Mol. Cell. Biol. 13, 1222–1234. M. (1995). Mutations in a novel gene, myoblast city, provide Zhu, H., Garcia, A. V., Ross, R. R., Evans, S. M., and Chien, K. R. evidence in support of the founder cell hypothesis for Drosophila (1991). A conserved 28-base-pair element (HF-1) in the rat cardiac muscle development. Development 121, 1979–1988. myosin light-chain-2 gene confers cardiac-specific and a-adrener- Sharrocks, A. D., von Hesler, F., and Shaw, P. E. (1993). The identi- gic-inducible expression in cultured neonatal rat myocardial fication of elements determining the different DNA binding spec- cells. Mol. Cell. Biol. 11, 2273–2281. ificities of the MADS box protein p67SRF and RSRFC4. Nucleic Zhu, H., Nguyen, V. T., Brown, A. B., Pourhosseini, A., Garcia, Acids Res. 21, 215–221. A. V., van Bilson, M., and Chien, K. R. (1993). A novel, tissue- Shore, P., and Sharrocks, A. D. (1995). The MADS-box family of restricted zinc finger protein (HF-1b) binds to the cardiac regula- transcription factors. Eur. J. Biochem. 229, 1–13. tory element (HF-1b/MEF-2) in the rat myosin light-chain 2 gene. Suzuki, E., Guo, K., Kolman, M., Yu, Y.-T., and Walsh, K. (1995). Mol. Cell. Biol. 13, 4432–4444. Serum induction of MEF2/RSRF expression in vascular myocytes is mediated at the level of translation. Mol. Cell. Biol. 15, 3415– Received for publication May 24, 1995 3423. Accepted July 18, 1995

/ m4070$8017 10-02-95 05:58:56 dba Dev Bio