SEARCH

SEARCH BY CITATION

Keywords:

  • Mohawk;
  • Mkx;
  • Iroquois;
  • Irx;
  • Sin3A;
  • HDAC;
  • Sap18;
  • Tbp;
  • repressor;
  • MRD;
  • myogenesis

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

Mohawk is an atypical homeobox gene expressed in embryonic progenitor cells of skeletal muscle, tendon, and cartilage. We demonstrate that Mohawk functions as a transcriptional repressor capable of blocking the myogenic conversion of 10T1/2 fibroblasts. The repressor activity is located in three small, evolutionarily conserved domains (MRD1–3) in the carboxy-terminal half of the protein. Point mutation analysis revealed six residues in MRD1 are sufficient for repressor function. The carboxy-terminal half of Mohawk is able to recruit components of the Sin3A/HDAC co-repressor complex (Sin3A, Hdac1, and Sap18) and a subset of Polymerase II general transcription factors (Tbp, TFIIA1 and TFIIB). Furthermore, Sap18, a protein that bridges the Sin3A/HDAC complex to DNA-bound transcription factors, is co-immunoprecipitated by MRD1. These data predict that Mohawk can repress transcription through recruitment of the Sin3A/HDAC co-repressor complex, and as a result, repress target genes required for the differentiation of cells to the myogenic lineage. Developmental Dynamics 238:572–580, 2009. © 2009 Wiley-Liss, Inc.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

Mohawk (Mkx) is the sole member of a newly characterized class within the TALE superclass of atypical homeobox genes (PBC, MEIS, TGIF, IRO, and MKX)(Anderson et al.,2006; Murkerjee and Burglin,2007). TALE (three amino acid loop extension) homeobox factors are defined by the inclusion of three additional amino acids in the loop between the first and second helices of their homeodomains. These proteins are essential for a broad set of developmental processes, including cell proliferation, differentiation and positional specification (Selleri et al.,2001; Brendolan et al.,2005; Moens and Selleri,2006; van Tuyl et al.,2006; diIorio et al.,2007). Initial characterization of mouse Mkx revealed a dynamic transcription pattern restricted to progenitors of skeletal muscle, tendon, and cartilage, as well as the sex chords of the male gonad and the ureteric bud tip of the metanephrogenic kidney (Anderson et al.,2006; Liu et al.,2006; Takeuchi and Bruneau,2007). Because Mkx orthologs are highly conserved in both protostomes and deuterostomes, Mkx is believed to play an essential role in regulating decisions regarding differentiation and patterning of these tissues during development. To further our understanding of Mkx, we were interested in determining its function as a transcriptional regulator and its role in tissue differentiation.

Among the TALE genes, Mkx is most closely related to the Iroquois (Irx) class (Anderson et al.,2006; Liu et al.,2006; Takeuchi and Bruneau,2007). Mouse Mkx shares 56% homology with Irx2 over the entire homeodomain (35/63 residues) and 82% (14/17 residues) specifically in Helix III, which physically interacts with DNA. This includes the conservation of an alanine in the DNA recognition domain of Helix III that is believed to participate in defining the target enhancer sequence for Irx proteins. The Irx genes can be distinguished from Mkx by the presence of a highly conserved IRO Box in the carboxy-terminal half of all Irx genes (Burglin,1997). Vertebrate members of the Irx family regulate a diverse set of developmental events, including chamber-specific transcription in the heart, specification and positional identity of the vertebrate neuroectoderm, branching morphogenesis in the lungs, patterning of the nephron, and formation of the zebrafish organizer (Bao et al.,1999; Bruneau et al.,2001; Kudoh and Dawid,2001; Gomez-Skarmeta and Modolell,2002; Matsumoto et al.,2004; Lecaudey et al.,2005; van Tuyl et al.,2006; Reggiani et al.,2007). Studies carried out with chick Irx2 and Drosophila Mirr predict that the Irx family members function to repress transcription (Matsumoto et al.,2004; Bilioni et al.,2005). In addition, transcriptional repression activity by Irx4 has been shown to be essential for proper chamber specification in the chick heart. However, the mechanisms by which the Iroquois genes function are still unknown.

During embryonic development, Mkx transcription is first detected in the dorsomedial and ventrolateral regions of the dermomyotome of maturing somites, which contain highly proliferative myogenic progenitor cells (Anderson et al.,2006). Tight regulation of the switch between proliferation and differentiation of these cells is essential for balancing the maintenance of the myogenic progenitor pool and the expansion of skeletal muscle (Amthor et al.,1999). Differentiation of myogenic progenitors can be measured using the myogenic conversion assay, which measures the ability of the bHLH transcription factor MyoD to convert fibroblast cells to myoblasts. In this assay, we found that Mkx is a strong inhibitor of myogenic differentiation, which was dependent on DNA-binding through the Mkx homeodomain. A biochemical analysis revealed Mkx possesses transcriptional repressor activity, mediated by three conserved domains in the carboxy-terminal region of the protein. This region forms a complex with both the Sin3A/HDAC co-repressor complex and a subset of Polymerase II general transcription factors. Therefore, Mkx can function as a repressor of genes required for the differentiation of tissue progenitors.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

Mkx Is a Transcriptional Repressor

The presence of a homeodomain in Mkx predicts that it regulates transcription, however, neither its native DNA-binding site nor its transcriptional properties are known. Therefore, we used the well-characterized Gal4/UAS transcription assay to determine the DNA binding-dependent transcription activity of Mkx. Full-length Mkx coding sequence was fused to the amino-terminus of the Gal4 DNA-Binding Domain (aa 1-147; Gal4DBD). Because the closely related Irx genes have been reported to act as transcriptional repressors (Matusumoto et al.,2004; Bilioni et al.,2005), the Gal4-Mkx fusion protein was co-transfected into NIH3T3 cells with luciferase reporter 5XUASpGL3Control. This reporter contains the SV40 promoter, the SV40 Enhancer and five copies of the Gal4 UAS, which provides a strong basal luciferase activity and is responsive to Gal4DBD fusion proteins (Fig. 1A). Gal4-Mkx was a potent transcriptional repressor in a dose-dependent manner over a 100-fold range in concentration, as compared with transfection with Gal4DBD alone (Fig. 1B). A 2.6-fold reduction (38.5%) in luciferase activity of the reporter was observed when only 1 ng of Gal4-Mkx encoding plasmid DNA was co-transfected and an 18.2-fold reduction (95.5%) in luciferase activity was observed with 100 ng of Gal4-Mkx. Although the use of the Gal4 DNA-binding domain is artificial, our results demonstrate that Mkx can function as a transcriptional repressor.

thumbnail image

Figure 1. Gal4-Mkx represses transcription in a dose-dependent manner. A: The DNA binding-dependent transcriptional properties of the Gal4-Mkx fusion protein were measured by co-expression in NIH3T3 cells with the Gal4-responsive luciferase reporter, 5XUASpGL3Control. B: Gal4-Mkx is a potent transcriptional repressor at plasmid concentrations ranging from 1 ng to 100 ng. Repression is expressed as the percent luciferase activity in cells transfected with 5XUASpGL3Control and Gal4DBD.

Download figure to PowerPoint

Identification of Three Independent Repressor Domains Within Mkx

To identify domains of Mkx mediating repression activity, we again used the Gal4/UAS assay system. For gross mapping, Gal4DBD fusion proteins were generated containing the Mkx homeodomain [Gal4-Mkx(70-132)], the amino-terminal region [Gal4-Mkx(1-69)] or carboxy-terminal region [Gal4-Mkx(133-354)] to the homeodomain (Fig. 2). Co-transfection of the reporter with the carboxy-terminal fusion protein resulted in an 18.3-fold repression of luciferase activity, relative to the Gal4DBD alone (Fig. 2). Neither the amino-terminal domain nor the homeodomain demonstrated any repressor activity. Therefore, the carboxy-terminal region contains all of the repressor activity.

thumbnail image

Figure 2. Identification of the transcriptional repressor domains in Mkx. A: The Gal4-responsive luciferase reporter, 5XUASpGL3Control, was co-transfected into NIH3T3 cells with expression plasmids for Gal4DBD fusions carrying either full-length Mkx (aa 1-354) or fragments of Mkx as indicated. The homeodomain (green) and Conserved Domains in the carboxy-terminal region (CD-A to E) are labeled. Data are presented for each Gal4DBD fusion as fold repression, which is the inverse of the fold luciferase activity relative to the luciferase activity determined for the Gal4 DNA-binding domain alone.

Download figure to PowerPoint

Potential repression domains within the carboxy-terminal region were identified based on amino acid conservation among vertebrate orthologs of Mkx. A multiple sequence alignment of Mkx orthologs from Mus musculus (GenBank Accession # NM_177595), Homo sapiens (GenBank accession no. NM_173576), and the two orthologs from Danio rerio, (MkxA, GenBank accession no. EU280785; MkxB, GenBank accession no. EU280786), revealed five highly conserved domains (CD-A through CD-E) in the carboxy-terminal region (sequence alignment and CDs are identified in Supp. Fig. S1, which is available online). Recently, CD-A, -B, and -D were also reported to be conserved among protostomes (Murkerjee and Burglin,2007). Of interest, CD-E is found within 25 residues at the carboxy-terminus that are not present in protostomes.

Fragments containing individual CDs (CD-A [aa 133-174], CD-B [aa 207-241], CD-C [aa 247-283], CD-D [aa 301-322] and CD-E [aa 327-354]) were subcloned in-frame with the Gal4DBD and tested for their ability to repress the 5XUASpGL3Control reporter. CD-A, CD-A with the adjacent homeodomain, and CD-C were unable to repress luciferase transcription (Fig. 2). In contrast, CD-B was a potent repressor and resulted in a 21.5-fold repression in luciferase activity, relative to the Gal4DBD alone. Among protostomes and deuterostomes, only the central 13 aa of CD-B (M. musculus KYKSSLLNRYLND; aa 281-230) are highly conserved (Murkerjee and Burglin,2007). When fused to the Gal4DBD these residues were able to repress transcription of the reporter 30-fold. We renamed this sequence Mohawk Repressor Domain 1 (MRD1). A carboxy-terminal fragment containing both CD-D and CD-E [Gal4-Mkx(284-354)] led to an 11.9-fold reduction in luciferase activity. Individual analysis of fragments containing either CD-D or CD-E revealed a 6.0- and 5.0-fold reduction, respectively (Fig. 2). We renamed these regions MRD2 and MRD3, respectively.

To determine the necessity of MRD1, 2, and 3 for the repression activity of Mkx, individual and combinatorial deletions of these domains were made in the context of the full-length protein. Deletion of MRD1 led to a reduction in repressor activity, (6.3-fold vs. 30.0-fold), while deletion of all three domains [Gal4-Mkx(ΔMRD1-3)] resulted in a complete loss of repressor function (Fig. 2). MRDs do not share any significant sequence homology with known functional domains, based on a TBLASTN search of the available sequenced genomes (data not shown). Therefore, MRDs 1–3 represent novel repression domains that can function independently to repress transcription and are necessary for the repressor function of Mkx.

Determination of Residues Required for MRD1 Function

A comparison of MRD1 sequence from mouse to the distantly related Drosophila melanogaster revealed that only 5 of 13 amino acids are identical, while 5 are highly similar (Fig. 3A). As a fusion to the Gal4DBD domain, Drosphila MRD1 was able to repress transcription of the 5XUASpgl3Control reporter in mouse NIH3T3 cells, albeit less strongly than mouse MRD1 (Fig. 3A). This suggests that distantly related protostome orthologs may share a conserved mechanism of MRD1-mediated repression.

thumbnail image

Figure 3. Residues required for MRD1 repressor function. A: The Gal4-responsive luciferase reporter, 5XUASpGL3Control, was co-transfected into NIH3T3 cells with expression plasmids for Gal4DBD fusions carrying mouse MRD1, Drosophila MRD1, or mouse MRD1 alanine substitution mutants. Data are presented for each Gal4DBD fusion as fold repression, which is the inverse of the fold luciferase activity relative to the luciferase activity determined for the Gal4 DNA-binding domain alone. B,C: I-TASSER program predicted models of MRD1 (B) and the artificial MRD1 (C) secondary structures with the side chains.

Download figure to PowerPoint

To determine the specific amino acids required for MRD1 function, alanine substitution mutants were created as fusions to the Gal4DBD and tested for their ability to repress the Gal4 re-sponsive 5XUASpgl3Control reporter. Alanine substitution of residues K218, Y219, K220, L223, R226, and Y227 greatly reduced MRD1 repressor function, whereas substitution of L224 or D230 had little effect (Fig. 3A). We next tested the highly substituted artificial MRD1 (aMRD1 sequence: KYKAALAARYAAA), which only contained those residues found to be critical for repressor function. aMRD1 was able to repress transcription at a level equivalent to MRD1.

The three-dimensional structure of MRD1 and aMRD1 were predicted using the web-based I-TASSER program, which is a hierarchical protein structure modeling approach based on secondary-structure enhanced profile–profile threading and the interactive implementation of the threading assembly refinement program (Wu et al.,2007; Zhang,2007,2008). Based on this program, both MRD1 and aMRD1 are predicted to form short alpha helices (Fig. 3B,C). From the models, it is clear that the predicted critical residues for MRD1 function are present on the same side of the alpha helix (Fig. 3B). To further demonstrate the importance of secondary structure for MRD1 function, we substituted a proline residue in place of asparagine at position 225. The proline substitution mutant, which is predicted to break the alpha helix secondary structure, resulted in a complete loss of repressor function (Fig. 3A). Therefore, secondary structure appears to be critical for MRD1 function.

Mkx Associates With Components of the Sin3A/HDAC Co-repressor Complex

Chromatin remodeling by the recruitment of Histone Deacetylases (HDACs) to a gene locus is a common mechanism for transcriptional repression (Burke and Baniahmad,2000). HDACs are targeted to specific gene loci complexed with other proteins, notably as part of the Sin3A/HDAC or NuRD co-repressor complexes (Ayer,1999). We tested the ability of components of the Sin3A/HDAC complex to form stable interactions with Mkx using co-immunoprecipitation. Myc-epitope tagged Mkx (MT-Mkx) and HA-epitope tagged members of the Sin3A/HDAC complex (Hdac1, Sin3A, and Sap18) were co-expressed in NIH3T3 cells. MT-Mkx was immunoprecipitated from cell lysates using an anti-myc antibody and co-precipitation of members of the Sin3A/HDAC co-repressor complex was determined by Western Blot using an anti-HA antibody. MT-Mkx, but not the myc-tag only control, was able to co-precipitate both Hdac1 and Sin3A (Fig. 4A,B). The fragment of Mkx carboxy-terminal to the homeodomain [MT-Mkx(133-354)] also bound Sin3A. In contrast, neither the homeodomain [MT-Mkx(70-132)] nor the region amino-terminal to the homeodomain [MT-Mkx(1-69)] was able to bind HA-Sin3A. This is consistent with our previous finding that the repressor function of Mkx is in the carboxy-terminal domain of the protein. However, the 13 aa MRD1 was not sufficient to co-precipitate Sin3A (Fig. 4B).

thumbnail image

Figure 4. The carboxy-terminal region of Mkx forms a stable complex with the Sin3A/HDAC co-repressor complex. The ability of Mkx to form a stable complex with Sin3A, Hdac1, and Sap18 was examined by co-immunoprecipitation/western blot. Myc-tag Mkx-specific fusion proteins were co-expressed in NIH3T3 cells with HA-tag Hdac1, Sin3A or Sap18. Whole cell lysates were immunoprecipitated with an anti-myc antibody and western blots were probed with an anti-HA antibody. A: HA-tag Hdac1 was co-immunoprecipitated by myc-tag Mkx and not by the myc-tag only control. B: HA-tag Sin3A was co-immunoprecipitated by full-length Mkx and the carboxy-terminal domain (aa 133-354), but not by the amino-terminal region (aa 1-69), the homeodomain (aa 70-132) or by MRD1. HA-tag Sap18 bound to Mkx in a similar manner as Sin3A, but additionally to the homeodomain and MRD1.

Download figure to PowerPoint

Sin3A-associated protein, 18 kDa (Sap18), has been shown to mediate the interaction between Sin3A and DNA-bound transcription factors (Zhang et al.,1997; Sheeba et al.,2007). Using the same co-immunoprecipitation assay, Sap18 was found to bind to full-length Mkx, the Mkx homeodomain, the region carboxy-terminal to the homeodomain, but not the region amino-terminal to the homeodomain (Fig. 4B). Of interest, MRD1 was sufficient to co-immunoprecipitate Sap18 as well. These results indicate that Mkx recruits the Sin3A/HDAC co-repressor complex.

Mkx Interacts With a Subset of Polymerase II General Transcription Factors

Components of the RNA Polymerase II general transcription factors (GTFs) are common targets of transcriptional repressors, and represent an HDAC-independent mechanism of repression. This occurs through direct binding or interaction with co-repressor complexes, which has been noted for the Sin3A/HDAC complex (Heinzel et al.,1997; Laherty et al.,1997; Nagy et al.,1997). To determine whether Mkx interacts with GTFs, co-immunoprecipitations were performed with MT-Mkx and HA-epitope tagged Tbp, Tbp-like 1 (Tbpl1) and GTFs that had been previously shown to interact with co-repressors, including TFIIA1 and TFIIB, (Burke and Baniahmad,2000). Full-length Mkx was able to co-immunoprecipitate TFIIA1, TFIIB, and Tbp (Fig. 5A). Of interest, Mkx did not co-precipitate Tbpl1, a gene that is 73% similar to Tbp and that can also bind TATA box elements and initiate transcription (Fig. 5A). We further show that Tbp is specifically co-immunoprecipitated by the region carboxy-terminal to the Mkx homeo-domain, but by neither the region amino-terminal to the homeodomain nor the homeodomain itself (Fig. 5B). This demonstrates that Mkx can interact with a specific subset of RNA Polymerase II GTFs, which potentially can contribute to a second mechanism of transcriptional repression.

thumbnail image

Figure 5. Mkx interacts with Tbp, TFIIA1, and TFIIB. The ability of Mkx to form a stable complex with members of the RNA Polymerase II general transcription factors was examined by co-immunoprecipitation/Western blot. A: The myc-tag Mkx fusion proteins were co-expressed in NIH3T3 cells with HA-tag versions of Tbp, Tbpl1, TFIIA1 and TFIIB. Whole cell lysates were immunoprecipitated with an anti-myc antibody and subsequently probed with an anti-HA antibody. Tbp, TFIIA1 and TFIIB all interacted with full-length myc-tag Mkx. B: HA-tag Tbp was co-immunoprecipitated by full-length Mkx and the carboxy-terminal domain (aa 133-354), but not by the amino-terminal region (aa 1-69) or the homeodomain (aa 70-132).

Download figure to PowerPoint

Mkx Inhibits the Differentiation of MyoD-Initiated Myoblasts

It has been previously reported that Mkx is expressed in the myogenic progenitor cells in the dermomyotome compartment of mouse somites (Anderson et al.,2006). Initiation of Mkx transcription appears in progressively posterior somites at a rate similar to the appearance of myoblasts in the myotome of these somites. This suggests that Mkx may play a role in regulating the onset of muscle differentiation. To test this, we examined the impact of Mkx on the conversion of 10T1/2 mouse fibroblasts to myoblasts by MyoD. Ectopic expression of MyoD in 10T1/2 cells induces the complete myogenic program, characterized by an exit from the cell cycle and the formation of multinucleate myotubes that express skeletal muscle-specific structural proteins (Tapscott et al.,1988).

MyoD, driven by the EMSV promoter (Davis et al.,1987), was expressed in 10T1/2 cells for 36 hours and then the cells were allowed to differentiate for 5 days under low mitogen conditions. Differentiation to the muscle lineage was assessed by immunolabeling for the skeletal muscle-specific protein myosin heavy chain (MHC; Fig. 6A,B). Co-expression of MyoD with full-length Mkx reduced the number of MHC-positive cells and appeared to reduce the number of multinucleated myotubes compared with MyoD alone. Quantifying the number of MHC-positive cells was used to assess the impact of Mkx on myoblast differentiation. MyoD expression alone was used as a reference of 100% differentiation (Fig. 6B). Co-expression of Mkx with MyoD resulted in a 48.2% reduction in the number of MHC-positive myotubes (Fig. 6B). The inhibition of myogenesis by Mkx was disrupted (14.0% vs. 48.2%), when Helix III, the DNA-binding recognition helix of the homeodomain (aa 114-130), was deleted (Fig. 6B). This demonstrates that Mkx inhibits myogenesis in a DNA-dependent manner. Given that Mkx is expressed in myogenic progenitors in the mouse embryo, this argues that Mkx plays a role in restricting the differentiation of progenitors to the myogenic pathway.

thumbnail image

Figure 6. Mkx inhibits the differentiation of MyoD-converted myoblasts. A: Mouse 10T1/2 fibroblast cells were transfected with expression vectors carrying MyoD or MyoD and Mkx and challenged to differentiate into myotubes by culturing in low mitogen containing conditions. B: The degree of differentiation was assessed by counting the number of myotubes, as identified by immunostaining for an antibody against skeletal muscle-specific myosin heavy chain (MHC). The myogenic activity of Mkx or Mkx(ΔH3), a mutation that deletes the DNA-recognition helix of the homeodomain, is expressed as the percentage of MHC-positive myotubes, compared with the total number of myotubes formed as a result of MyoD transfection alone, which was given the arbitrary value of 100%. The average of three independent experiments reveal that Mkx is able to inhibit muscle differentiation, while Mkx(ΔH3) relieved most, but not all of the inhibition. As well, Mkx, Mkx(ΔH3) and the myc-tag empty vector were alone not sufficient to induce muscle differentiation.

Download figure to PowerPoint

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

Mkx is a recently identified homeobox gene transcribed in the discrete embryonic progenitor cells of the musculoskeletal system, including skeletal muscle, tendon, and cartilage (Anderson et al.,2006). Mkx has the potential to play a critical role in the genetic pathways that lead to the development of these tissues. In this regard, the expression pattern of Mkx overlaps with Pax3 and paraxis in skeletal muscle progenitors, scleraxis in tendon progenitors and Sox9 in cartilage progenitors of the axial skeleton (Anderson et al.,2006). Here, we describe the results of our analysis of the biochemical function of Mkx in cell culture. Mkx can function as a transcriptional repressor and contains three small repressor domains. Co-IP experiments predict that Mkx represses transcription through either HDAC-dependent chromatin remodeling or association with the general transcription machinery. Finally, a myogenic conversion of 10T1/2 cells demonstrated that expression of Mkx was able to block MyoD-directed muscle differentiation. Overall, this study identifies Mkx as potent regulator of skeletal muscle differentiation that functions through three domains to direct transcriptional repression.

The TALE superclass of homeodomain-containing proteins have been shown to be important regulators of cell proliferation, differentiation, and patterning of a diverse set of tissues during the development of both protostomes and deuterostomes. Several members of the TALE superclass have been reported to act as transcriptional repressors, including genes in the closely related Irx, Pbc, and TGIF classes (Asahara et al.,1999; Sharma and Sun,2001; Wotton et al.,2001; Matsumoto et al.,2004). We found that Mkx can also act as a transcriptional repressor. This activity was mapped to three short domains that were able to function independently when fused to a heterologous DNA-binding domain. The short MRD's likely function as protein–protein interaction domains with co-repressors. This is highlighted by the fact that residues critical for MRD1 function all occur on the same side of an alpha-helix, creating a potential protein interaction surface. Short alpha-helices have been shown to participate in protein–protein interactions for several transcription factors, including BCL2 and MDM2 (Arkin and Wells,2004). In the case of MRD1, we were able to identify Sap18 as a binding partner. Short modular repression motifs have been described for other transcriptional repressors. Most notable are the 55 aa en D domain of the engrailed protein, the 57 aa C2D2 domain of eve, and the 13 aa Sin3A Interacting Domain (SID) of the Mad family (Jaynes and O'Farrell,1991; Han and Manley,1993a,b; Eilers et al.,1999). The MRD sequences of Mkx do not share significant homology with these motifs and therefore represent novel repression domains. It is interesting to note that MRD3 only appears in vertebrate orthologs of this gene. Considering the ability of MRD3 to independently repress transcription, this gives vertebrate Mkx orthologs an additional regulatory function.

The Sin3A/HDAC core complex has been well studied for its ability to direct gene-specific transcriptional repression. In addition to the acetyltransferase activity of HDACs, the core complex serves as a scaffold for bringing other enzymes that modify the nucleosome and general transcription machinery (Yang et al.,2002,2003). Neither Sin3A nor HDACs possess intrinsic DNA-binding activity, indicating that interaction with DNA-binding transcription factors is necessary to localize the repressor complex. To date, several transcription factors, including TALE superclass homeodomain proteins Pbx1 and TGIF, have been shown to bind Sin3A (Sharma and Sun,2001; Wotton et al.,2001; Silverstein and Ekwall,2005). Here, we demonstrate that Mkx also binds the Sin3A/HDAC complex. Sin3A, Hdac1 and Sap18 all interacted with the carboxy-terminal region of Mkx, which acts as a functional repressor in vivo. Interestingly, we were able to show that Sap18 can interact with MRD1 alone. Sap18 has been reported to play a role in stabilizing the Sin3A/HDAC complex and interacts with Sin3A (Zhang et al.,1997). Localization of Sap18 through a heterologous DNA-binding domain to a reporter gene results in transcriptional repression (Zhang et al.,1997). Thus, Sap18 recruitment can explain the repression observed by MRD1 and may also be important for bridging Mkx to the larger Sin3A/HDAC complex. The Mkx homeodomain was also able to co-immunoprecipitate Sap18, which was surprising, considering that this domain has no repressor activity. For future work, it will be interesting to determine the targets of MRD2 and MRD3 and determine whether they help stabilize the Sin3A/HDAC complex interaction of Mkx or function through novel mechanisms.

The ability to co-immunoprecipitate TFIIA1, TFIIB, and Tbp raises the possibility that Mkx is able to repress transcription through an HDAC-independent mechanism. General transcription factors are essential for the formation of the RNA Polymerase II initiation complex at the transcriptional start site. Binding to these proteins can lead to a disruption of the initiation complex, resulting in gene silencing (Heinzel et al.,1997; Laherty et al.,1997; Nagy et al.,1997). Our analysis has demonstrated that the large carboxy-terminal region of Mkx can bind Tbp, which is the same region that also bound the Sin3A/HDAC complex. Since Tbp and TFIIB binding to Sin3A has been reported (Burke and Baniahmad,2000), it remains a possibility that Mkx mediated-repression occurs through one large complex.

Development of skeletal muscle is dependent on the balance between the self-renewal of myogenic progenitor cells associated with the periphery of the muscle and their differentiation into myofibers. Cells at the dorsomedial and ventrolateral lips (DML and VLL, respectively) of the dermomyotome are the source of the self-renewing premyogenic cells that contribute to the expanding myotome, the anlage for skeletal muscle (Christ and Ordahl,1995; Cossu et al.,1996; Denetclaw et al.,1997; Venters and Ordahl,2002). Specification of these cells by the bHLH muscle regulator factors MyoD, Myf5, Mrf4, and myogenin, is balanced by muscle antagonists such as Notch, Pax3, Pax7, and follistatin, which promote proliferation and block the transcription of muscle-specific genes and muscle agonists such as myostatin. We became interested in the role Mkx plays in skeletal muscle differentiation based on its dynamic expression in the myogenic progenitors during development. Mkx is transcribed specifically in the regions of the DML and VLL, indicating a role in controlling differentiation of the premyogenic cells (Anderson et al.,2006). To test this hypothesis, we used the classic myogenic conversion cell-culture assay. We found that Mkx inhibited the differentiation of MyoD-induced myoblasts in a DNA-binding dependent manner. While the genetic pathway through which this occurred is unclear at this time, Mkx may function upstream of the myogenic factors and repress their transcription. This is also supported by the observation that Mkx is not transcribed in the myotome compartment of the somite, thus indicating that down-regulation of Mkx transcription may be a necessary step in differentiation of the myogenic lineage.

EXPERIMENTAL PROCEDURES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

Plasmids

Full-length gene-coding sequences and subfragments were cloned into the CS2MT (6 × amino-terminal Myc epitopes), CS2HA (2 × HA amino-terminal epitopes)(Rupp et al.,1994) or CS2G4 (Gal4DBD 1-147) expression plasmids using polymerase chain reaction (PCR) amplification of mouse embryonic cDNA. Restriction sites were included during PCR amplification to facilitate cloning and stop codons were introduced where appropriate. CS2G4 was constructed by subcloning the HindIII/EcoRI fragment from pSG424 (Sadowski et al.,1988) into the HindIII/EcoRI sites of CS2MT, which removed the 6 x Myc epitopes. Five copies of the Gal4 UAS were PCR amplified from Gal4E1b5CAT (Wilson-Rawls et al.,2004) and subcloned into the KpnI/BglII sites of pGL3Promoter and pGL3Control (Promega, Madison, WI), to create 5XUASpGL3Promoter and -Control, respectively. All clones were sequenced for verification.

Transient Transfections and Luciferase Reporter Assays

NIH3T3 mouse fibroblast cells were seeded at 4 × 104 cells/well in complete media (DMEM supplemented with 10% Newborn Calf Serum) in 24-well tissue-culture dishes. Each well was transfected with a total of 400 ng of plasmid DNA using 1 μl of Lipofectamine (Invitrogen, Carlsbad, CA) and 4 μl of PLUS reagent (Invitrogen), according to manufacture's protocol. Transfected cells were lysed 24 hr after transfection in 100 μl/well Luciferase Cell Culture Lysis Buffer (Promega) and subjected to a single freeze–thaw cycle at −80°C. Luciferase activity was measured for each well by reacting 20 μl of cell lysate with 100 μl of Luciferase Assay Buffer (Promega) in white 96-well plates, using an FLx800 microplate reader (BioTek Instruments, Inc., Winooski, VT). Sample variables were performed in triplicate per experiment and each experiment was repeated at least three times. Data are presented from a single representative experiment.

Myogenic Conversion Assay in C3H10T1/2 Cells

C3H10T1/2 mouse fibroblast cells were seeded at 1.9 × 105 cells in DMEM supplemented with 10% fetal bovine serum in 35-mm tissue culture dishes coated with 0.1% gelatin. Upon plating, the cells were immediately transfected according to manufacture's protocol with 6 μl of Fugene6 (Roche Applied Science, Indianapolis, IN) and 2 μg of plasmid DNA and incubated for 36 hr. The culture medium was removed and the cells were incubated in low mitogen containing media (DMEM supplemented with 2% horse serum). Myogenic conversion was scored after 5 days by the expression of skeletal specific myosin, which was immunolabeled using the Anti-Myosin MY32 antibody (Sigma-Aldrich, St. Louis, MO), as described in (Wilson-Rawls et al.,1999).

Co-immunoprecipitations and Western Blots

NIH3T3 cells were transfected as stated above, with the exception that cells were transfected in 60-mm plates with 2 μg of plasmid DNA. According to each experiment, Myc-epitope tag fusion proteins were co-expressed with target hemagglutinin (HA)-epitope tag fusion proteins. Immunoprecipitations were carried out using mouse monoclonal Anti-Myc antibody (Invitrogen) as described in Wilson-Rawls et al. (1999). Initially, the expression of all epitope tagged proteins was confirmed by Western blot using antibodies specific. For each pulldown experiment, Western blots were performed to ensure equivalent total input protein.

For Western blotting, proteins were transferred to an Immobilon-P PVDF membrane (Millipore, Billerica, MA), dried in 100% methanol and used for western blotting as previously described in (Wilson-Rawls et al.,1999), with the following exceptions. For Co-IP experiments, the primary antibody was mouse Anti-HA (Invitrogen), diluted 1:500 in 1% Carnation nonfat dry milk in Tris Buffered Saline containing 0.1% Tween-20 (MTBST). The secondary antibody was alkaline phosphate conjugated Anti-Mouse IgG (Invitrogen), diluted 1:10,000 in MTBST. For Gal4/UAS experiments, mouse monoclonal RK5C1 antibody (Santa Cruz Biotechnology, Santa Cruz, CA) was used as a primary antibody to confirm the expression of Gal4DBD fusion proteins (Western blots not shown). The secondary antibody was the same as for the Co-IP experiments. Membranes were imaged on a Molecular Dynamics STORM chemifluorescent imager using ECF substrate (GE Healthcare Bio-Sciences Corp., Piscataway, NJ).

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

We thank Dr. Anthony Firulli for providing the CS2MT plasmid, Dr. Steve Tapscott for providing the CS2HA plasmid, Dr. Scott Bingham for assistance in DNA sequencing, and Dr. James Elser. We also thank Dr. Kenro Kusumi, Dr. Lei Lei, Dr. Brian Verrelli, and Megan Rowton for their valuable discussion.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

Additional Supporting Information may be found in the online version of this article.

FilenameFormatSizeDescription
DVDY_21873_sm_SuppFigS1.tif6261KSupp. Fig. S1. A multiple-sequence alignment using ClustalW was performed using a select group of vertebrate Mkx orthologs. This clearly revealed five highly conserved domains (CDs) within the carboxy-terminal region of Mkx. The homeodomain is shaded gray and the amino acids comprising each CD are identified by a black bar above the alignment.

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.