Downloaded from genesdev.cshlp.org on October 6, 2021 - Published by Cold Spring Harbor Laboratory Press REVIEW The myocardin family of transcriptional coactivators: versatile regulators of cell growth, migration, and myogenesis G.C. Teg Pipes, Esther E. Creemers, and Eric N. Olson1 Department of Molecular Biology, University of Texas Southwestern Medical Center, Dallas, Texas 75390, USA The association of transcriptional coactivators with se- gene regulation and expands the regulatory potential of quence-specific DNA-binding proteins provides versatil- individual cis-regulatory elements. ity and specificity to gene regulation and expands the The regulation of serum response factor (SRF) and its regulatory potential of individual cis-regulatory DNA se- target genes provides a classic example of the diversity of quences. Members of the myocardin family of coactiva- genes controlled by a single DNA-binding protein and tors activate genes involved in cell proliferation, migra- the importance of cofactor interactions in the control of tion, and myogenesis by associating with serum re- gene expression (Shore and Sharrocks 1995). Among the sponse factor (SRF). The partnership of myocardin family many target genes of SRF are genes involved in cell pro- members and SRF also controls genes encoding compo- liferation and muscle differentiation, which represent nents of the actin cytoskeleton and confers responsive- opposing programs of gene expression: Muscle genes are ness to extracellular growth signals and intracellular repressed by growth factor signals and generally do not changes in the cytoskeleton, thereby creating a tran- become activated until myoblasts exit the cell cycle. The scriptional–cytoskeletal regulatory circuit. These func- ability of SRF to regulate different sets of downstream tions are reflected in defects in smooth muscle differen- effector genes depends on its association with positive tiation and function in mice with mutations in myocar- and negative cofactors (Treisman 1994; Shore and Shar- din family members. This article reviews the functions rocks 1995). SRF thereby serves as a platform to interpret and mechanisms of action of the myocardin family of cell identity and signaling by engaging various partners. coactivators and the physiological significance of tran- The recent discovery and mechanistic dissection of the scriptional coactivation in the context of signal-depen- myocardin family of transcriptional coactivators (Wang dent and cell-type-specific gene regulation. et al. 2001, 2002), which regulate SRF activity during cell growth, migration, and myogenesis, have provided new insights into the mechanism of action of SRF, as well as The instructions for gene expression are “hard wired” the roles of coactivators in the control of gene expression into DNA sequence in the form of cis-regulatory ele- in general. This article reviews the functions and mecha- ments that bind sequence-specific transcription factors nisms of action of the myocardin family of coactivators and confer specialized patterns of gene expression and considers their activities in the broader context of (Howard and Davidson 2004). However, the binding of signal-dependent and cell-type-specific gene regulation. transcription factors to their cognate sites is often insuf- ficient to account for the patterns of expression of their target genes. Indeed, it is not uncommon for a transcrip- SRF and MADS-box transcription factors tion factor to control genes with distinct or even mutu- ally exclusive expression patterns. Many transcription SRF belongs to the MADS (MCM1, Agamous, Deficiens, factors are also relatively weak transcriptional activators SRF) family of transcription factors, which share homol- and function by recruiting coactivators (or corepressors) ogy in a 57-amino-acid MADS-box that mediates ho- that do not bind DNA directly but regulate transcription modimerization and DNA binding to a dyad symmetri- in a DNA sequence-specific manner by associating with cal A + T-rich DNA consensus sequence (Shore and DNA-bound factors (Spiegelman and Heinrich 2004). Sharrocks 1995). There are numerous MADS-box pro- Thus, the association of DNA-binding proteins with ac- teins in plants, but SRF and the four members of the cessory factors provides specificity and versatility to myocyte enhancer factor-2 (MEF2) family (MEF2A, MEFB, MEFC, and MEFD) are the only MADS-box pro- teins found in metazoans (Black and Olson 1998). The [Keywords: Myocardin; MRTF; coactivator; smooth muscle; SRF] crystal structures of SRF and MEF2 have revealed com- 1Corresponding author. E-MAIL [email protected]; FAX (214) 648-1196. monalities in their modes of DNA binding (Pellegrini et Article is online at http://www.genesdev.org/cgi/doi/10.1101/gad.1428006. al. 1995; Santelli and Richmond 2000), which are re- GENES & DEVELOPMENT 20:1545–1556 © 2006 by Cold Spring Harbor Laboratory Press ISSN 0890-9369/06; www.genesdev.org 1545 Downloaded from genesdev.cshlp.org on October 6, 2021 - Published by Cold Spring Harbor Laboratory Press Pipes et al. flected in the similar sequences of their binding sites: consensus CArG-box from the c-fos promoter results in CC(A/T)6GG (known as a CArG-box) for SRF and widespread (i.e., non-muscle-specific) expression (Haut- CTA(A/T)4TAG for MEF2. mann et al. 1998; Chang et al. 2001). Thus, it appears The conserved N-terminal region of the MADS-box that cell growth genes are “constitutive” SRF targets, forms an ␣-helical structure, which becomes oriented in their expression driven by a perfect consensus CArG-box an antiparallel manner within homodimers to form a that binds SRF with high affinity, while some myogenic bipartite DNA-binding domain. An N-terminal exten- SRF targets have only nonconsensus CArG-boxes that sion of the MADS-box in SRF makes critical base con- require coactivation, enhancement, or reinforcement of tacts in the minor groove that stabilize DNA binding. A SRF binding to activate their expression in muscle cells. domain C-terminal to the DNA-binding region is ori- ented away from DNA and is important for homodimer- ization and interaction with accessory factors. SRF, like SRF loss-of-function phenotypes other MADS-box transcription factors, interacts with a Experiments in cultured cells using dominant-negative diverse array of transcriptional regulators to generate tis- SRF mutants or siRNA to down-regulate SRF expression sue-specific and signal-responsive patterns of gene ex- have revealed essential roles for SRF in serum-dependent pression (Messenguy and Dubois 2003). The SRF MADS- cell growth and skeletal muscle differentiation (Soulez et box induces an unusual degree of bending of the double al. 1996; Wei et al. 1998; Kaplan-Albuquerque et al. helix, likely creating a unique shape that facilitates as- 2005). A dominant-negative SRF mutant also blocks dif- sociation with other components of the transcriptional ferentiation of coronary smooth muscle cells (SMCs) in regulatory complex (West and Sharrocks 1997). the proepicardial organ of chick embryos (Landerholm et Approximately 160 genes have been estimated to be al. 1999) and disrupts skeletal and cardiac muscle differ- direct transcriptional targets of SRF, and about half of entiation in transgenic mice (Zhang et al. 2001). these have been experimentally validated (Sun et al. A homozygous Srf-null mutation in mice results in 2006a). The majority of SRF target genes are involved in lethality at gastrulation (Arsenian et al. 1998). SRF regu- cell growth, migration, cytoskeletal organization, and lates genes involved in cell migration and adhesion, pro- myogenesis. The prototypical SRF target gene involved cesses required for gastrulation. ES cells lacking SRF also in cell growth, c-fos, is controlled by a single CArG-box, display defects in spreading, adhesion, and migration referred to as a serum response element (SRE), that acts that correlate with abnormalities in actin stress fibers in concert with surrounding cis-regulatory elements in and loss of expression of genes encoding stress fiber com- the promoter (Norman et al. 1988). The first descriptions ponents including vinculin, talin, and a subset of actin of SRF-dependent enhancers of muscle gene expression isoforms, which are directly regulated by SRF (Schratt et described duplicated CArG-boxes that functioned coop- al. 2002). SRF has also been shown to regulate cell sur- eratively to activate transcription (Miwa and Kedes vival during development via its control of the anti-ap- 1987; Chow and Schwartz 1990). These initial observa- optotic Bcl-2 gene (Schratt et al. 2004). tions led to the working hypothesis that SRF-dependent The early lethality of Srf-null mice precludes an analy- genes involved in cell growth are controlled by single sis of the potential roles of SRF in muscle development. CArG-boxes, while SRF-dependent muscle genes are However, conditional deletion of Srf from the cardiac controlled by duplicated CArG-boxes. However, recent muscle lineage results in mid-gestation lethality, accom- genome-wide curation of all SRF target genes and the panied by disruption of sarcomeric structure and abnor- CArG-boxes that control their expression reveals dupli- malities in muscle gene regulation (Miano et al. 2004; cated CArG-boxes near the start of transcription of most Parlakian et al. 2004; Niu et al. 2005). Smooth muscle SRF target genes (see Supplemental Material in Sun et al. deletion of Srf also results in a reduced number of differ- 2006a). Some CArG-box-dependent muscle genes are