![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
|
|
|
|
|
||
REVIEW
Department of Molecular Biology, University of Texas Southwestern Medical Center, Dallas, Texas 75390, USA
 |
Abstract |
|---|
 Top Abstract SRF and MADS-box transcription...SRF loss-of-function phenotypesThe myocardin family of...Structure-function studiesOpposing roles of myocardin...Roles of the myocardin...Roles of myocardin and...Loss-of-function phenotypes of...The SRF-myocardin partnership in...Modulation of myocardin activity...Potential roles in human...Other partners of the...Implications and questions for...AcknowledgmentsReferences |
|---|
[Keywords: Myocardin; MRTF; coactivator; smooth muscle; SRF]
). However, the binding of transcription factors to their cognate sites is often insufficient to account for the patterns of expression of their target genes. Indeed, it is not uncommon for a transcription factor to control genes with distinct or even mutually exclusive expression patterns. Many transcription factors are also relatively weak transcriptional activators and function by recruiting coactivators (or corepressors) that do not bind DNA directly but regulate transcription in a DNA sequence-specific manner by associating with DNA-bound factors (Spiegelman and Heinrich 2004). Thus, the association of DNA-binding proteins with accessory factors provides specificity and versatility to gene regulation and expands the regulatory potential of individual cis-regulatory elements.
The regulation of serum response factor (SRF) and its target genes provides a classic example of the diversity of genes controlled by a single DNA-binding protein and the importance of cofactor interactions in the control of gene expression (Shore and Sharrocks 1995). Among the many target genes of SRF are genes involved in cell proliferation and muscle differentiation, which represent opposing programs of gene expression: Muscle genes are repressed by growth factor signals and generally do not become activated until myoblasts exit the cell cycle. The ability of SRF to regulate different sets of downstream effector genes depends on its association with positive and negative cofactors (Treisman 1994; Shore and Sharrocks 1995). SRF thereby serves as a platform to interpret cell identity and signaling by engaging various partners. The recent discovery and mechanistic dissection of the myocardin family of transcriptional coactivators (Wang 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 roles of coactivators in the control of gene expression in general. This article reviews the functions and mechanisms of action of the myocardin family of coactivators and considers their activities in the broader context of signal-dependent and cell-type-specific gene regulation.
|
SRF and MADS-box transcription factors |
|---|
|
TopAbstractSRF and MADS-box transcription...SRF loss-of-function phenotypesThe myocardin family of...Structure-function studiesOpposing roles of myocardin...Roles of the myocardin...Roles of myocardin and...Loss-of-function phenotypes of...The SRF-myocardin partnership in...Modulation of myocardin activity...Potential roles in human...Other partners of the...Implications and questions for...AcknowledgmentsReferences |
|---|
). There are numerous MADS-box proteins in plants, but SRF and the four members of the myocyte enhancer factor-2 (MEF2) family (MEF2A, MEFB, MEFC, and MEFD) are the only MADS-box proteins found in metazoans (Black and Olson 1998). The crystal structures of SRF and MEF2 have revealed commonalities in their modes of DNA binding (Pellegrini et al. 1995; Santelli and Richmond 2000), which are reflected in the similar sequences of their binding sites: CC(A/T)6GG (known as a CArG-box) for SRF and CTA(A/T)4TAG for MEF2.
The conserved N-terminal region of the MADS-box forms an 
-helical structure, which becomes oriented in an antiparallel manner within homodimers to form a bipartite DNA-binding domain. An N-terminal extension of the MADS-box in SRF makes critical base contacts in the minor groove that stabilize DNA binding. A domain C-terminal to the DNA-binding region is oriented away from DNA and is important for homodimerization and interaction with accessory factors. SRF, like other MADS-box transcription factors, interacts with a diverse array of transcriptional regulators to generate tissue-specific and signal-responsive patterns of gene expression (Messenguy and Dubois 2003). The SRF MADS-box induces an unusual degree of bending of the double helix, likely creating a unique shape that facilitates association with other components of the transcriptional regulatory complex (West and Sharrocks 1997).
Approximately 160 genes have been estimated to be direct transcriptional targets of SRF, and about half of these have been experimentally validated (Sun et al. 2006a). The majority of SRF target genes are involved in cell growth, migration, cytoskeletal organization, and myogenesis. The prototypical SRF target gene involved in cell growth, c-fos, is controlled by a single CArG-box, referred to as a serum response element (SRE), that acts in concert with surrounding cis-regulatory elements in the promoter (Norman et al. 1988). The first descriptions of SRF-dependent enhancers of muscle gene expression described duplicated CArG-boxes that functioned cooperatively to activate transcription (Miwa and Kedes 1987; Chow and Schwartz 1990). These initial observations led to the working hypothesis that SRF-dependent genes involved in cell growth are controlled by single CArG-boxes, while SRF-dependent muscle genes are controlled by duplicated CArG-boxes. However, recent genome-wide curation of all SRF target genes and the CArG-boxes that control their expression reveals duplicated CArG-boxes near the start of transcription of most SRF target genes (see Supplemental Material in Sun et al. 2006a). Some CArG-box-dependent muscle genes are expressed in only one muscle cell type (smooth, skeletal, or cardiac), whereas others are expressed in multiple muscle cell types. How such specificity is achieved has not been fully resolved, but likely involves positive and negative modulatory proteins in addition to SRF that act in a gene-specific manner.
The SRF target genes involved in cell growth can often be distinguished from those involved in myogenesis by the degree to which their CArG-boxes fit the perfect consensus. CArG-boxes in the promoters of several muscle genes deviate from the consensus sequence by one or more residues, resulting in a reduction in SRF-binding affinity (Chang et al. 2001), whereas the CArG-box upstream of c-fos fits the consensus perfectly. The nucleotide substitutions that render muscle gene CArGs nonconsensus are often evolutionarily conserved, and replacement of these CArG-boxes with the perfect consensus CArG-box from the c-fos promoter results in widespread (i.e., non-muscle-specific) expression (Hautmann et al. 1998; Chang et al. 2001). Thus, it appears that cell growth genes are "constitutive" SRF targets, their expression driven by a perfect consensus CArG-box that binds SRF with high affinity, while some myogenic SRF targets have only nonconsensus CArG-boxes that require coactivation, enhancement, or reinforcement of SRF binding to activate their expression in muscle cells.
|
SRF loss-of-function phenotypes |
|---|
|
TopAbstractSRF and MADS-box transcription...SRF loss-of-function phenotypesThe myocardin family of...Structure-function studiesOpposing roles of myocardin...Roles of the myocardin...Roles of myocardin and...Loss-of-function phenotypes of...The SRF-myocardin partnership in...Modulation of myocardin activity...Potential roles in human...Other partners of the...Implications and questions for...AcknowledgmentsReferences |
|---|
; Wei et al. 1998; Kaplan-Albuquerque et al. 2005). A dominant-negative SRF mutant also blocks differentiation of coronary smooth muscle cells (SMCs) in the proepicardial organ of chick embryos (Landerholm et al. 1999) and disrupts skeletal and cardiac muscle differentiation in transgenic mice (Zhang et al. 2001).
A homozygous Srf-null mutation in mice results in lethality at gastrulation (Arsenian et al. 1998). SRF regulates genes involved in cell migration and adhesion, processes required for gastrulation. ES cells lacking SRF also display defects in spreading, adhesion, and migration that correlate with abnormalities in actin stress fibers and loss of expression of genes encoding stress fiber components including vinculin, talin, and a subset of actin isoforms, which are directly regulated by SRF (Schratt et al. 2002). SRF has also been shown to regulate cell survival during development via its control of the anti-apoptotic Bcl-2 gene (Schratt et al. 2004).
The early lethality of Srf-null mice precludes an analysis of the potential roles of SRF in muscle development. However, conditional deletion of Srf from the cardiac muscle lineage results in mid-gestation lethality, accompanied by disruption of sarcomeric structure and abnormalities in muscle gene regulation (Miano et al. 2004; Parlakian et al. 2004; Niu et al. 2005). Smooth muscle deletion of Srf also results in a reduced number of differentiated SMCs near the dorsal aorta, and the few that are detected display cytoskeletal defects, including an absence of thick filaments and bundles of thin filaments (Miano et al. 2004). Deletion of Srf in skeletal muscle results in perinatal lethality from skeletal muscle hypoplasia (Li et al. 2005b).
|
The myocardin family of SRF coactivators |
|---|
|
TopAbstractSRF and MADS-box transcription...SRF loss-of-function phenotypesThe myocardin family of...Structure-function studiesOpposing roles of myocardin...Roles of the myocardin...Roles of myocardin and...Loss-of-function phenotypes of...The SRF-myocardin partnership in...Modulation of myocardin activity...Potential roles in human...Other partners of the...Implications and questions for...AcknowledgmentsReferences |
|---|
). The properties of myocardin highlight one of the difficulties of identifying transcriptional coactivators in general. Myocardin’s role in activating high levels of muscle transcription as an SRF coactivator was not obvious from examination of its predicted protein structure. Subsequent experiments showed that myocardin could stimulate transcription from CArG-dependent muscle enhancers, yet did not bind DNA directly. Instead, myocardin forms a stable ternary complex on CArG-boxes by associating with SRF. In Srf-null cells, myocardin has no detectable transcriptional activity (Wang et al. 2002). Activities of similar coactivator proteins are also critically dependent on the cellular context in which they are assayed: In the absence of a required cofactor, they are inactive.
Consistent with its discovery as a novel cardiac-restricted gene, myocardin expression is largely confined to the cardiovascular system. In vertebrate embryos, the onset of myocardin expression coincides with that of Nkx2-5, the earliest known marker of the cardiac lineage (Wang et al. 2001). Thereafter, myocardin is expressed throughout the heart, as well as in a subset of SMCs within the cardiovascular system and internal organs.
Myocardin shares homology with myocardin-related transcription factor-A (MRTF-A, also called MAL, MKL-1, and BSAC), and MRTF-B (also called MKL-2) (Ma et al. 2001; Mercher et al. 2001; Sasazuki et al. 2002; Wang et al. 2002), which are expressed in a broad range of embryonic and adult tissues: Most cell types tested, including ES cells (Du et al. 2004), express some level of MRTF-A and MRTF-B. During embryogenesis, however, MRTF-A is significantly enriched in mesenchymal cells, muscle cells, and epithelial cells of various organs (Wang et al. 2002; S. Li, D. Wang, and E.N. Olson, unpubl.). Likewise, MRTF-B is most highly expressed in a subset of the branchial arch arteries (Oh et al. 2005) derived from the neural crest and in developing neural structures during embryogenesis.
A single member of the myocardin family exists in fruit flies (Fig. 1) and in several other arthropods, but not in fungi. The fly homolog is most closely related to MRTF-A by primary sequence homology. Thus, the ancestral myocardin gene was likely an ortholog of Mrtf-A that arose during the metazoan transition to multicellularity.
|
|
Structure–function studies |
|---|
|
TopAbstractSRF and MADS-box transcription...SRF loss-of-function phenotypesThe myocardin family of...Structure-function studiesOpposing roles of myocardin...Roles of the myocardin...Roles of myocardin and...Loss-of-function phenotypes of...The SRF-myocardin partnership in...Modulation of myocardin activity...Potential roles in human...Other partners of the...Implications and questions for...AcknowledgmentsReferences |
|---|
). In NIH 3T3 cells, stimuli that promote actin polymerization release MRTFs from G-actin, allowing them to enter into the nucleus. In contrast, myocardin is localized exclusively to the nucleus (Kuwahara et al. 2005), which is likely due to sequence divergence of its RPEL domains and consequent inability to bind actin (Wang et al. 2001; Miralles et al. 2003).
Myocardin family members contain a 35-amino-acid SAP domain, named after SAF-A/B, Acinus, and PIAS, which in other proteins participates in chromosomal dynamics, nuclear breakdown, and apoptotic DNA fragmentation (Aravind and Koonin 2000). The SAP domain is predicted to adopt a helix–linker–helix structure resembling a homeodomain without a DNA recognition motif, and in a few cases, the SAP domain has been shown to serve as a weak DNA-binding module (Gohring et al. 1997; Kipp et al. 2000; Sachdev et al. 2001; Bohm et al. 2005). Deletion mutants of myocardin lacking this region retain the ability to stimulate SRF activity on some promoters, but they are defective on others (Wang et al. 2001), which suggests a possible role for this domain in mediating interactions with other promoter-specific transcription factors.
Association of myocardin and MRTFs with SRF is mediated by a short peptide sequence that includes a basic and glutamine-rich region (Wang et al. 2001, 2002). A coiled-coil motif resembling a leucine zipper mediates homo- and heterodimerization of myocardin and MRTFs and has been proposed to contribute to the cooperativity between CArG-boxes in SRF-dependent muscle genes (Wang et al. 2001, 2002; Miralles et al. 2003; Du et al. 2004). The C-terminal regions of myocardin and MRTFs are somewhat divergent in amino acid sequence, and function as transcription activation domains (TADs). Deletion of these regions generates dominant-negative mutants. The TADs can be replaced by heterologous TADs, such as that from the viral coactivator VP16, and function normally in reporter activation in vitro, indicating that this domain serves a general function in transcriptional activation, but does not contribute to the specificity of these factors for SRF coactivation (Wang et al. 2001).
Myocardin recruits chromatin remodeling enzymes to SRF target genes. Association with the histone acetyltransferase p300 enhances and interaction with class II histone deacetylases represses expression of SRF target genes (Cao et al. 2005). In support of a role for myocardin in altering chromatin structure, myocardin–SRF complexes have been shown to associate with a specific variant of histone H3 on SMC gene loci in vivo (McDonald et al. 2006). Myocardin also appears to enhance SRF binding to nonconsensus CArG-boxes associated with smooth muscle SRF gene targets, whereas the perfect consensus CArG-boxes found upstream of the immediate early gene c-fos bind SRF with the same affinity in the presence and absence of myocardin (Hendrix et al. 2005). The mechanism whereby myocardin enhances SRF DNA binding has not been defined. Perhaps the interaction of the two proteins stabilizes SRF in a conformation that facilitates association with DNA, or with DNA in a particular promoter context.
|
Opposing roles of myocardin and Ets factors in the control of SRF |
|---|
|
TopAbstractSRF and MADS-box transcription...SRF loss-of-function phenotypesThe myocardin family of...Structure-function studiesOpposing roles of myocardin...Roles of the myocardin...Roles of myocardin and...Loss-of-function phenotypes of...The SRF-myocardin partnership in...Modulation of myocardin activity...Potential roles in human...Other partners of the...Implications and questions for...AcknowledgmentsReferences |
|---|
).
|
, 1998; Sharrocks et al. 1997). The SRF-binding regions of myocardin and MRTFs resemble the predicted secondary structure of the B-box, although they lack direct amino acid homology with this region of Elk-1. Deletion of this region of myocardin abolishes its ability to associate with SRF and activate SRF-dependent genes, and replacement of this domain with the Elk-1 B-box restores these functions (Wang et al. 2004). Myocardin and Elk-1 compete for interaction with a common docking site in the MADS-box of SRF (Miralles et al. 2003; Wang et al. 2004). Their mutually exclusive association with this site creates a binary switch in which growth signals can modulate the phenotype of SMCs. When SMCs are stimulated with PDGF, Elk-1 becomes phosphorylated by the MAP kinase signaling pathway, facilitating its association with SRF and favoring the displacement of myocardin (Wang et al. 2004). Although Elk-1 is a coactivator of SRF, its activity is substantially weaker than that of myocardin, such that the displacement of myocardin from SRF by Elk-1 results in an overall decrease in the expression of smooth muscle genes. Consistent with this model, lowering endogenous levels of Elk-1 in SMCs results in an increase in expression of a subset of smooth muscle genes, likely via derepression of SRF–myocardin complexes on the promoters of these genes (Zhou et al. 2005). |
Roles of the myocardin family in Rho and cytoskeletal signaling |
|---|
|
TopAbstractSRF and MADS-box transcription...SRF loss-of-function phenotypesThe myocardin family of...Structure-function studiesOpposing roles of myocardin...Roles of the myocardin...Roles of myocardin and...Loss-of-function phenotypes of...The SRF-myocardin partnership in...Modulation of myocardin activity...Potential roles in human...Other partners of the...Implications and questions for...AcknowledgmentsReferences |
|---|
). Members of the Rho family, including RhoA, Cdc42, and Rac1, use multiple pathways to increase the F-actin:G-actin ratio in different cell types (Geneste et al. 2002). In fibroblasts, multiple downstream effectors of activated RhoA promote F-actin accumulation: mDia1 by promoting assembly of F-actin (Copeland and Treisman 2002), and ROCK by stabilizing F-actin (Sotiropoulos et al. 1999). Calcium influx through voltage gated calcium channels during SMC contraction stimulates myocardin expression via a Rho/ROCK-dependent pathway (Wamhoff et al. 2004), and myocardin siRNA suppresses expression of contractile protein gene targets of SRF and myocardin, demonstrating that myocardin expression is regulated by, and part of the response to, changes in actin dynamics associated with smooth muscle physiology.
The actin-sensitivity of SRF has been proposed to be due to sequestration of MRTF-A in the cytoplasm by direct binding to G-actin via the N-terminal RPEL motifs (Miralles et al. 2003; Posern et al. 2004). Actin polymerization in response to Rho signaling results in activation of SRF as a consequence of the translocation of MRTF-A/MAL to the nucleus (Fig. 3; Miralles et al. 2003; Posern et al. 2004). However, significant areas of uncertainty remain regarding the actin-sensitivity of MRTF nuclear localization. For example, mutants of RhoA have been isolated that stimulate SRF activity but do not promote stress fiber formation (Sahai et al. 1998; Cen et al. 2004), indicating that, among the many pathways regulated by RhoA, there may be an alternate effector of SRF transcriptional activity. It remains to be determined whether these SRF-activating mutants of RhoA affect nucleocytoplasmic partitioning of the MRTFs. The mechanisms that regulate nuclear import of MRTF-A also remain to be defined. How is MRTF-A dislodged from actin monomers (Miralles et al. 2003; Posern et al. 2004)? Does actin polymerization mask the domain of actin that binds MRTF-A, or is there an actin-polymerizing factor that binds G-actin with a higher avidity than does MRTF-A?
|
). STARS promotes nuclear localization of MRTF-A and MRTF-B, suggesting that it may compete with their RPEL motifs for association with actin and thereby provide a mechanism for signaling from the contractile apparatus to the nucleus (Kuwahara et al. 2005). STARS may also drive MRTFs into the nucleus indirectly, by stabilizing F-actin and decreasing the cytoplasmic F:G actin ratio. STARS is up-regulated during normal heart development and pathological cardiac hypertrophy (G.C.T. Pipes, K. Kuwahara, and E.N. Olson, in prep.), settings in which subsets of the muscle actin genes (all of which are CArG-dependent) are activated and G-actin monomers are incorporated into an expanding sarcomere. Some mechanism must exist to shield SRF activity from feedback inhibition by binding of actin monomers to MRTFs during periods of sarcomeric growth. STARS or some other muscle actin-binding protein is likely to serve this function.
The regulation of actin gene expression by SRF and the sensitivity of MRTFs to the state of actin polymerization creates a regulatory loop whereby changes in cell shape that influence cytoskeletal structure can promote actin synthesis (Fig. 3). This transcriptional–cytoskeletal regulatory circuit appears to be quite ancient, as studies in multiple lower metazoans show that inactivation of SRF causes cytoskeletal defects (Guillemin et al. 1996; Escalante et al. 2004; Sun et al. 2006a). SRF, the multiple CArG-dependent actin genes, their protein products, and the MRTFs thus constitute a novel mechanism that maintains cell shape, extrusive activity, and/or contractile potential in homeostatic balance with SRF transcriptional output. Nucleation of stress fibers (via Rho) or expression of F-actin-binding proteins (such as STARS during sarcomeric expansion) tips this balance toward stimulation of SRF transcriptional activity. Inhibition of SRF activity by expression of a naturally occurring splice variant of SRF (Davis et al. 2002; Chang et al. 2003) or actin depolymerization tips the balance toward loss of contractility and sarcomere regression.
|
Roles of myocardin and MRTFs in myogenesis |
|---|
|
TopAbstractSRF and MADS-box transcription...SRF loss-of-function phenotypesThe myocardin family of...Structure-function studiesOpposing roles of myocardin...Roles of the myocardin...Roles of myocardin and...Loss-of-function phenotypes of...The SRF-myocardin partnership in...Modulation of myocardin activity...Potential roles in human...Other partners of the...Implications and questions for...AcknowledgmentsReferences |
|---|
; Small et al. 2005). Conversely, ectopic expression of myocardin in Xenopus embryos results in activation of cardiac gene expression throughout the embryo; cardiac gene expression in response to myocardin is particularly robust in spinal cord neurons for reasons that are unclear. Similarly, injection of myocardin mRNA into Xenopus animal cap explants results in expression of numerous, but not all, myocardial genes (Small et al. 2005). Forced expression of myocardin in fibroblasts or stem cells is sufficient to activate the expression of smooth muscle genes, with only a subset of cardiac genes being expressed (Wang et al. 2003; Pipes et al. 2005; van Tuyn et al. 2005). This promyogenic activity requires the SRF-binding region and TAD of myocardin (Wang et al. 2003).
MRTF-A and MRTF-B can also activate smooth muscle gene expression when overexpressed in transfected fibroblasts (Wang et al. 2002; Du et al. 2004; Selvaraj and Prywes 2004). Why, then, don’t MRTF-A and MRTF-B activate smooth muscle genes in the many nonmuscle cell types in which they are expressed? Nonmuscle tissues may express negative regulators of MRTF activity or lack essential cofactors or signals required for establishment of an SRF–MRTF complex on smooth muscle promoters. Perhaps there is a threshold level of expression of MRTFs that is required for smooth muscle gene activation, but not achieved by the endogenous levels of expression of MRTFs. It is also possible that the state of chromatin renders muscle target genes of SRF–MRTF inaccessible in nonmuscle cells.
MKL-1/MRTF-A has been implicated in skeletal muscle differentiation in vitro. Si-RNA for MRTF-A blocks fusion of skeletal myoblasts into multinucleated myotubes and prevents activation of SRF-dependent muscle genes (Selvaraj and Prywes 2003). Similarly, transgenic expression of dominant-negative MRTF-A in skeletal muscle in vivo results in abnormally thin muscle fibers with significant fibrosis, consistent with MRTF activity being required for muscle growth. This form of myopathy resembles that in mice lacking Srf in skeletal muscle (Li et al. 2005b). Expression levels of muscle genes (both CArG-dependent and -independent) are vastly decreased in these mice.
Myocardin is constitutively nuclear and specifically expressed in cardiac and smooth muscle myocytes, while the MRTFs shuttle between the nucleus and the cytoplasm in fibroblasts. In striated muscle, however, the activity of STARS protein is predicted to drive MRTFs into the nucleus and effectively make them constitutively nuclear (although this has not yet been shown in vivo). Thus, in contrast to other cell types, muscle cells have multiple mechanisms driving high levels of SRF transcriptional activity via myocardin family members, particularly during sarcomeric expansion, myocyte growth, and hypertrophy.
|
Loss-of-function phenotypes of myocardin family genes |
|---|
|
TopAbstractSRF and MADS-box transcription...SRF loss-of-function phenotypesThe myocardin family of...Structure-function studiesOpposing roles of myocardin...Roles of the myocardin...Roles of myocardin and...Loss-of-function phenotypes of...The SRF-myocardin partnership in...Modulation of myocardin activity...Potential roles in human...Other partners of the...Implications and questions for...AcknowledgmentsReferences |
|---|
). The lack of smooth muscle gene expression in null embryos is, surprisingly, not due to a cell-autonomous defect in SMC differentiation in the embryo proper, as the creation of chimeric mice derived from wild-type and myocardin–/– ES cells revealed that myocardin-null SMCs can differentiate and be incorporated into vascular tissues with normal morphology (Pipes et al. 2005). Recent studies suggest that the lethality of myocardin-null embryos reflects an essential role of myocardin in a population of extraembryonic SMCs (S. Li, D.Z. Wang, and E.N. Olson, unpubl.). Proof of this hypothesis awaits tissue-specific deletion of myocardin from embryonic and extraembryonic SMCs.
The potential role of myocardin in heart development has been clouded by the lack of a cardiac phenotype in myocardin-null embryos. How can the lack of a cardiac phenotype of myocardin-null mice be explained in light of the ability of myocardin siRNA or dominant-negative mutants to disrupt heart development in Xenopus embryos? It seems likely that MRTFs or other cardiac transcription factors can compensate for the loss of myocardin function in the early heart. The generation of compound mutant mice will be required to address this possibility. With respect to myocardin’s apparent ability to induce cardiac gene expression in Xenopus embryos (Small et al. 2005) but not in cultured mammalian cells (Wang et al. 2003), the environment of the embryo may be permissive for the actions of myocardin due to the presence of cofactors or signaling molecules that are absent from cultured cells. Alternatively, or in addition, negative regulators may interfere with the cardiogenic activity of myocardin in vitro.
Mice lacking MRTF-A are viable and display an intriguing defect specific to lactating females in which myoepithelial cells, which provide the contractility required for secretion of milk from the mammary gland, fail to differentiate and undergo apoptosis (Li et al. 2006; Sun et al. 2006b). As a result, offspring of MRTF-Anull mothers fail to thrive. MRTF-B is also expressed in myoepithelial tissue, but despite its high degree of similarity to MRTF-A, is unable to replace the function of MRTF-A, suggesting there are significant differences in how these factors are regulated, or in their interactions with other transcription factors. Alternatively, it may be that MRTF-A and MRTF-B are wholly redundant, and deleting MRTF-A lowers the amount of transcriptional activation of SRF target genes below a required threshold.
Mice lacking MRTF-B display defects in differentiation of SMCs derived from the neural crest, although two different targeted alleles of the gene result in phenotypes of differing severity. A lacZ enhancer trap allele results in vascular defects and perinatal lethality with incomplete penetrance (Li et al. 2005a), whereas a deletion mutation introduced into the gene by homologous recombination results in lethality at E13.5 with complete penetrance (Oh et al. 2005). The -smooth muscle actin gene, which is SRF-dependent, is expressed normally in the heart and aorta of MRTF-B-null mice, but is down-regulated in the neural-crest-derived SMCs, which display the mutant phenotype. Thus, in a subset of smooth muscle tissue (i.e., neural crest-derived SMCs), the developmental roles of MRTF-B appear to have diverged from the roles of myocardin and MRTF-A. An interesting question for future investigation is why MRTF-A cannot substitute for MRTF-B in neural crest cells, and why MRTF-B cannot substitute for MRTF-A in differentiating myoepithelial cells.
|
The SRF–myocardin partnership in fruit flies |
|---|
|
TopAbstractSRF and MADS-box transcription...SRF loss-of-function phenotypesThe myocardin family of...Structure-function studiesOpposing roles of myocardin...Roles of the myocardin...Roles of myocardin and...Loss-of-function phenotypes of...The SRF-myocardin partnership in...Modulation of myocardin activity...Potential roles in human...Other partners of the...Implications and questions for...AcknowledgmentsReferences |
|---|
; Guillemin et al. 1996). There is a single myocardin family member in Drosophila, referred to as MAL-D or DMRTF, which, like its mammalian counterparts, functions as a strong SRF coactivator (Han et al. 2004; Somogyi and Rorth 2004). Determination of the function of DMRTF during embryogenesis has been complicated by a strong maternal contribution to the embryo. Nevertheless, loss of DMRTF function results in abnormalities in tracheal branching similar to, but less severe than, those in blistered mutant embryos. The defects in tracheal development in DMRTF RNAi-injected embryos are similar to those of embryos lacking Rac1 and Rac2 (Chihara et al. 2003), suggesting that DMRTF, like its mammalian counterparts, responds to Rho family members.
blistered also plays an important role in Drosophila wing development by promoting the formation of intervein tissue and suppressing vein formation. Expression of a dominant-negative DMRTF mutant in wing imaginal discs reduces wing interveins, whereas overexpression of DMRTF promotes the formation of excess intervein tissue, defects reminiscent of blistered loss- and gain-of-function phenotypes, respectively. Expression of dominant-negative DMRTF also disrupts dorsal migration of mesodermal cells, a process required for regional specification of cardiac, somatic, and visceral muscles (Han et al. 2004).
Migration of border cells, a specialized group of follicle cells, into germline tissue during oogenesis requires directed movement and extension of filopodial processes, which is accompanied by nuclear accumulation of DMRTF (Somogyi and Rorth 2004). In DMRTF mutant flies, border cells are unable to migrate effectively due to a defective actin cytoskeleton.
The target genes of SRF–DMRTF that control cell migration and tracheal branching in Drosophila remain to be defined. However, the striking similarities between the defects resulting from loss of function of SRF and myocardin family members in vertebrates and Drosophila suggest that these proteins represent an ancient and evolutionarily conserved system for coupling changes in cell shape and extracellular signaling with cell migration during development.
|
Modulation of myocardin activity during pathological growth and remodeling of muscle cells |
|---|
|
TopAbstractSRF and MADS-box transcription...SRF loss-of-function phenotypesThe myocardin family of...Structure-function studiesOpposing roles of myocardin...Roles of the myocardin...Roles of myocardin and...Loss-of-function phenotypes of...The SRF-myocardin partnership in...Modulation of myocardin activity...Potential roles in human...Other partners of the...Implications and questions for...AcknowledgmentsReferences |
|---|
; Xing et al. 2006), a sensitive marker of cardiac myocyte hypertrophy controlled by SRF. Myocardin transcriptional activity is also negatively regulated via phosphorylation of myocardin by glycogen synthase kinase-3
, a known suppressor of hypertrophic signaling (Badorff et al. 2005). Furthermore, myocardin transcript levels are up-regulated in failing heart (Torrado et al. 2003), consistent with the up-regulation of CArG-dependent muscle genes during pathological remodeling of the heart. Given that heart failure and cardiac hypertrophy are accompanied by increases in CArG-dependent muscle gene expression (with some exceptions), the increase in myocardin expression may represent a maladaptive response to cardiac injury, suitable for future therapeutic intervention.
Pathological remodeling of the vessel wall during atherosclerosis and restenosis involves a switch in SMC phenotype, from a differentiated, contractile to a proliferative, "synthetic" state (Owens et al. 2004). Signals that perturb myocardin activity will, in turn, block transcription of SMC contractile genes, and may promote phenotypic switching (Fig. 4). The forkhead transcription factor Foxo4 is up-regulated in vascular SMCs following injury and interacts with myocardin with consequent inhibition of transcriptional activity. The association of Foxo4 and myocardin appears to play a key role in modulation of smooth muscle growth and differentiation in this process, as Foxo4 and myocardin can both be isolated from the chromatin of smooth muscle genes under proliferating conditions (Liu et al. 2005). The Kruppel transcription factor KLF4 is also up-regulated during phenotypic dedifferentiation of SMCs and blocks activation of SMC contractile genes by decreasing myocardin expression (Liu et al. 2005) and interfering with SRF binding to CArG-boxes in intact chromatin, as well as recruiting repressive chromatin remodeling activities to SMC promoters (McDonald et al. 2006). The in vivo relevance of KLF4-mediated repression of myocardin expression is difficult to reconcile with recent reports of vast increases in myocardin immunoreactivity in synthetic SMCs found following vascular injury (Doi et al. 2005), although other reports describe a decrease in myocardin transcript levels under similar conditions (Hendrix et al. 2005).
|
|
Potential roles in human disease |
|---|
|
TopAbstractSRF and MADS-box transcription...SRF loss-of-function phenotypesThe myocardin family of...Structure-function studiesOpposing roles of myocardin...Roles of the myocardin...Roles of myocardin and...Loss-of-function phenotypes of...The SRF-myocardin partnership in...Modulation of myocardin activity...Potential roles in human...Other partners of the...Implications and questions for...AcknowledgmentsReferences |
|---|
, 2002 Dastugue et al. 2002; Hrusak and Porwit-MacDonald 2002; Ballerini et al. 2003; Duchayne et al. 2003; Gilliland et al. 2004; Hsiao et al. 2005). The resulting RBM15–MRTF-A fusion protein contains almost the entire OTT protein sequence fused near the N terminus of MRTF-A, after the first predicted RPEL domain. All of the domains of MRTF-A required for activation of SRF target genes are thus present in this fusion protein. The mutant protein causes uncontrolled cell proliferation through an unknown mechanism. The oncogenic fusion protein displays enhanced transcriptional activity toward the growth-factor-inducible c-fos and Egr1 promoters, but its ability to activate smooth muscle genes is unaltered (Cen et al. 2003). How the RBM15–MRTF-A fusion promotes tumorigenesis remains to be determined, but given the role of SRF/MRTFs in regulating cytoskeletal assembly during cell migration, it will be of interest to determine whether this protein partnership plays a role in invasive cell migration during cancer metastasis. |
Other partners of the myocardin family |
|---|
|
TopAbstractSRF and MADS-box transcription...SRF loss-of-function phenotypesThe myocardin family of...Structure-function studiesOpposing roles of myocardin...Roles of the myocardin...Roles of myocardin and...Loss-of-function phenotypes of...The SRF-myocardin partnership in...Modulation of myocardin activity...Potential roles in human...Other partners of the...Implications and questions for...AcknowledgmentsReferences |
|---|
effector Smad3, which interacts with myocardin and stimulates its activity in a CArG-independent manner (Qiu et al. 2005). Mutation of the CArG-boxes, in the SM22 promoter, the prototypical smooth muscle target gene of SRF, reduced but did not ablate synergistic activation of the promoter by myocardin and Smad3. In contrast, Smad1 interacts with myocardin, resulting in transcriptional synergy in a CArG-dependent manner (Callis et al. 2005). While these findings suggest that myocardin can coactivate transcription through a DNA-binding partner other than SRF, it should be kept in mind that Smad3 also binds SRF, raising questions as to whether this form of coactivation is truly SRF-independent (Qiu et al. 2005). Furthermore, given the stability of the complex formed by SRF and myocardin on CArG-containing DNA, it is unclear how this alternative interaction with myocardin could form in the presence of SRF in vivo.
The transcriptional activity of myocardin is subject to repression by a variety of corepressor proteins. The hairy-related transcription factor (HRT)-2, which is expressed in developing vasculature (Nakagawa et al. 1999), associates with myocardin and represses its activity through a mechanism yet to be defined (Proweller et al. 2005). Similarly, GATA factors are capable of repressing myocardin’s coactivation of SRF by competing for a common docking site on SRF, although this effect is only observed on a subset of smooth muscle enhancer elements. On the CArG-dependent enhancer of the smooth muscle myosin heavy chain gene, GATA-6 can synergize with myocardin and SRF to coactivate high levels of transcription, highlighting the context-dependent activity of myocardin (Yin and Herring 2005). Similarly, GATA-4 binds directly to myocardin and can synergize with myocardin to activate transcription from the cardiac-specific Nkx2-5 enhancer in a CArG-box-dependent manner, yet it prevents activation of the cardiac ANF promoter by myocardin (Oh et al. 2004).
|
Implications and questions for the future |
|---|
|
TopAbstractSRF and MADS-box transcription...SRF loss-of-function phenotypesThe myocardin family of...Structure-function studiesOpposing roles of myocardin...Roles of the myocardin...Roles of myocardin and...Loss-of-function phenotypes of...The SRF-myocardin partnership in...Modulation of myocardin activity...Potential roles in human...Other partners of the...Implications and questions for...AcknowledgmentsReferences |
|---|
)? Gene regulation by the recruitment of a TAD to a sequence-specific DNA-binding protein expands the regulatory potential of the DNA-bound transcription factor by allowing for the exchange of multiple positive and negative cofactors. This form of transcriptional regulation is suited for the rapid activation or repression of sets of genes that are unlinked and cannot be controlled by a shared enhancer, such as muscle genes. This mechanism is also especially amenable to mediating rapid changes in gene expression in response to extracellular signals, a central function of SRF and MEF2. The evolution of SRF and myocardin proteins as separate components of a transcriptional activator, rather than being contained in a single polypeptide, is also reflective of, and likely responsible for, the plasticity of the smooth muscle phenotype. In this regard, it is instructive to compare and contrast the activities of myocardin, as a regulator of smooth muscle differentiation, with MyoD, a master regulator of skeletal muscle development. Both proteins are capable of activating downstream muscle genes associated with the smooth and skeletal muscle phenotypes, respectively. However, there are key differences in the biology of smooth and skeletal muscle cells that reflect, at least in part, the properties of myocardin-SRF and MyoD.
In contrast to SMCs, which can switch their phenotypes between proliferative and differentiated states in response to extracellul
