hMusTRD1 (cid:1) 1 Represses MEF2 Activation of the Troponin I Slow Enhancer*

The novel transcription factor hMusTRD1 (cid:1) 1 (human muscle TFII-I repeat domain-containing protein 1 (cid:1) 1; E., C. S., P., and E. C. (1998) Mol. Cell. Biol. 18, 6641–6652) was identified in a yeast one-hybrid screen as a protein that binds within an upstream enhancer-containing region of the skeletal muscle-specific gene, TNNI1 (human troponin I slow ; hTnI slow ). It has been proposed that hMus - TRD1 (cid:1) 1 may play an important role in fiber-specific muscle gene expression by virtue of its ability to bind to an Inr-like element (nucleotides (cid:2) 977 to (cid:2) 960) within the hTnI slow upstream enhancer-containing region that is necessary for slow fiber-specific expression. In this study we demonstrate that both MEF2C, a known regu-lator of slow fiber-specific genes, and hMusTRD1 (cid:1) 1 regulate hTnI slow through the Inr-like element. Co-transfec- tion assays in C2C12 cells and Cos-7 cells demonstrate that hMusTRD1 (cid:1) 1 represses hTnI slow transcription and prevents MEF2C-mediated

Skeletal muscle formation and maturation requires the coordinated expression of genes resulting in the diversification of skeletal muscle cells into either slow or fast twitch myofibers (2)(3)(4). The establishment of fiber types is due in part to properties intrinsic to myoblasts (5,6) and to the responsiveness of the myofiber to physiological cues such as functional innervation and biomechanical load (7)(8)(9)(10). The molecular mechanisms that dictate these functionally specialized myofiber types at a transcriptional level are less well understood (11)(12)(13). Deletion mapping of fast and slow contractile protein genes has yielded information on potential factors that might be involved in fiberspecific expression (14 -18). Recently, diverse signaling pathways and transcription factors have been described that influence slow myofiber type specification, including the calcium/ calmodulin-dependent phosphatase calcineurin, the Ras mitogen-activated protein kinase pathway, and the transcriptional co-activator peroxisome proliferator-activated receptor-␥ co-activator-1 (PGC-1␣) (1, 19 -30).
Members of the MEF2 gene family are key regulatory factors common to two of these pathways. Paradigms that promote slow fiber specification in mice such as endurance activity and low frequency tonic pacing of the motoneuron result in an increase in MEF2 levels and the induction of a MEF2-responsive transgene (31,32). This occurs concomitant with the induction of slow/oxidative marker genes such as troponin I slow and myoglobin that have been shown to be regulated by MEF2. It is proposed that this regulation occurs through the calcineurin-mediated hypophosphorylation of MEF2 that enhances the function of the MEF2 transcriptional activation domain. Recently, forced expression of PGC-1␣ in fast fiber muscles in mice was shown to be capable of eliciting morphological features and gene expression pattern characteristic of slow fibers (30). It is possible that this regulatory capability involves calcineurin and MEF2 because both factors potentiate the PGC-1␣-mediated activation of the slow fiber-specific promoters of troponin I slow (TnI slow ) and myoglobin.
TnI slow has been used as a model gene to study mechanisms of slow fiber-specific expression. TnI slow is the major isoform of the TnI gene family expressed in all newly formed fibers but is down-regulated in future fast and up-regulated in future slow myofibers during fetal development (33,34). The responsiveness of TnI slow to physiological cues that elicit slow fiber gene expression such as innervation and biomechanical load is well documented (7,9,35,36). The regions of the human and rat TnI slow genes that direct slow fiber-specific expression have been defined using somatic cell injection and transgenic analysis (1,14,16,36). These analyses defined a 157-bp upstream enhancer (USE) 1 region of the human and a similar 128-bp region of the rat TnI slow gene (14,37) that are critical in directing slow myofiber-specific expression. Sequence motif and electrophoretic mobility shift assay (EMSA) analysis of the 157-bp hTnI slow USE identified potential transcription factorbinding sites within this region including a MEF2 consensus element and USE B1 that contains an initiator of transcription or Inr-like element (1). The Inr-like element was shown to be essential for high level, slow fiber-specific expression. In addition, transgenic analysis has demonstrated that the MEF2binding site contributes to expression in the soleus muscle (31).
Using the yeast one-hybrid assay, a novel 944-aa DNA-binding protein, MusTRD1, was isolated that binds to the Inr-like element within the USE B1 region (1). Mutation of the Inr-like core sequence (GATTAA) within the USE-B1 region (nucleotides Ϫ977 to Ϫ960) to USE-B1b (GATatc) abrogated both slow fiber-specific expression from the USE and DNA binding by MusTRD1. MusTRD1 contains five repeat domains that share ϳ70% homology to the six repeat domains of the Inr-like binding protein TFII-I (38). Two MusTRD1 homologues, GTF3 and CREAM (consisting of 959 aa), from humans and three mouse orthologues, BEN 1, 2, and 3, have been isolated (39 -43). We have discovered multiple MusTRD isoforms in mice that arise from alternative splicing of a single gene, Gtf2ird1, on mouse chromosome 5 (44). In keeping with the nomenclature that we have devised for the mouse isoforms, we have changed the name of human MusTRD1 to hMusTRD1␣1 (human muscle TFII-I repeat domain-containing protein 1␣1). The name of the isoform is based on the peptide sequence. The first number indicates the unique combination of C-terminal repeat domains present. Isoforms contain 1 of 2 possible C-terminal exons that are designated "␣" (exon 31) or "␤" (exon 30). The final number indicates which of the seven possible specific combinations of unique C-terminal exons is present.
In this study, we show that hMusTRD1␣1 acts as a repressor of hTnI slow transcription and that repression occurs through the hTnI slow USE. We show that MEF2C (myocyte enhancer factor 2C) transcriptional activation via the USE is dependent upon an intact Inr-like element and is blocked by hMus-TRD1␣1. hMusTRD1␣1 binds to the Inr-like element via two DNA-binding domains (DBDs) located between aa 351 and 458 and aa 544 and 944, respectively. Transient transfection analysis of a deletion series of hMusTRD1␣1 revealed that repression can occur in the absence of DNA binding. hMusT1␣1 is a potent repressor by itself, and GST pull-down analysis revealed that it can interact directly with nuclear receptor co-repressor (NCoR). This report demonstrates that hMusTRD1␣1 interacts directly with MEF2C and can prevent MEF2C from binding to its cognate binding site. The data support a role for hMus-TRD1␣1 as a repressor of slow fiber-specific genes by preventing MEF2-mediated transcriptional activation.
Reverse Transcription-PCR-Total RNA was extracted from Cos-7 cells, undifferentiated C2C12 myoblasts, and C2C12 cultures maintained 3 days in differentiation medium using Trizol® (Invitrogen) and reverse transcribed with Impromptu® (Promega). PCR was conducted using the primers: MusTRDF, GAGCTACAGTCAGACTTCCTCAG, and MusTRDR, TCTCTGACTTGTCATGGACGATG, which were designed to regions conserved between mouse and human.
Transactivation Assays-C2C12 cells (5 ϫ 10 4 cells/well) were cultured overnight in low glucose Dulbecco's modified Eagle's medium (Invitrogen) with 20% fetal bovine serum and 0.5% chick embryo extract. Transfections were performed using LipofectAMINE 2000™ (Invitrogen) with 250 ng of pTnI slow USEB1tkluc or pTnI slow USEB1btkluc and 100-ng quantities of the pcDNA3.1hMusTRD1␣1-truncation series, pcDNAMEF2C or combinations of the above. After transfection the cells were allowed to differentiate for 36 h in Dulbecco's modified Eagle's medium with 2% horse serum before harvesting.
Luciferase assays were performed according to the supplied protocol (Promega) using a Top Count Microplate scintillation and luminescence counter (Packard). The activities were normalized to protein concentration and expressed as ratios of empty vector activity. The experiments were performed three times, and a representative result is shown.
Electrophoretic Mobility Shift Analysis-hMusTRD1␣1 350-, 458-, 564-, 689-, 786-, 944-, and 500-aa proteins and MEF2C protein were produced by in vitro coupled transcription-translation reactions using TNT® rabbit reticulocyte lysate (Promega). The oligonucleotides used include trimerized B1 or B1b elements, B1(3X) and B1b(3X) (described elsewhere), MEF2 (nucleotides Ϫ908 to Ϫ891 from the USE; CATC-CTAAAAATACCCGG), and mutMEF2 (the same region with four mutated residues in the core MEF2 consensus sequence (CATCCTAAtgc-gACCCGG) and were labeled using Klenow to fill 3Ј overhangs in the presence of [␣-32 P]dCTP and purified on 15% acrylamide gels. DNA binding reactions containing 40 l of 10 mM Hepes (pH 7.9), 1 mM dithiothreitol, 0.2 g/l poly(dI-dC), 10% glycerol, 0.1-0.5 ng of labeled oligonucleotides, and in vitro translated proteins were incubated for 20 min at room temperature. In assays involving antibodies or competitors, the sheep anti-hMusTRD1␣1 1-20aa polyclonal antibody, preimmune serum, 2, 4, and 6 l of in vitro translated hMusTRD1␣1 944aa , and 4 l of rabbit reticulolysate as a control were preincubated for 20 min at room temperature prior to the addition of DNA. 1 ng of annealed, unlabeled B1b(3X) oligonucleotide fragment was included in antibody shift assays to reduce nonspecific binding of proteins present in reticulocyte lysates. DNA-protein complexes were electrophoresed through a 4% (w/v) nondenaturing polyacrylamide gel in 0.5ϫ TBE and visualized by autoradiography. The experiments were performed three times, and a representative result is shown.
Immunoprecipitation-Cos-7 cells transfected with pCMXNCoR and pcDNA3.1hMusTRD1␣1 350 myc/his were lysed using Nonidet P-40 lysis buffer (20 mM Tris, pH 7.5, 1 mM EDTA, 1 mM EGTA, 0.1 mM phenylmethylsulfonyl fluoride, 150 mM NaCl, 1% Nonidet P-40 containing Complete TM (Roche Applied Science) protease inhibitor mixture). 100 g of lysate was mixed with an anti-NCoR Rb88 polyclonal antibody directed against the C-terminal 2239 -2453 aa region and protein A/G agarose (Santa Cruz) and incubated at 4°C for 30 min. The complexes bound to the agarose beads were washed in 1ϫ lysis buffer, then boiled in SDS loading buffer, and subjected to 10% SDS-PAGE before transfer onto polyvinylidene difluoride HybondP membrane (Amersham Biosciences). The filter was incubated with the 9E10 anti-c-myc monoclonal antibody (Santa Cruz), and binding of the secondary antibody was detected using ECL Super Signal TM (Pierce).
GST Pull-down Assays-Expression of GST-NCoR 1649 -2453 , GST-hMusTRD1␣1 944 , and GST alone in the BL21(DE3)pLysS Gold Escherichia coli strain (containing the pACYC-T7-Trx plasmid) was confirmed by SDS-PAGE. GST pull-down assays were performed by co-incubation of GST or GST fusion proteins with in vitro translated 35 S-labeled hMusTRD1␣1 350aa or MEF2C initially for 30 min at 30°C and for a further 40 min at room temperature in PPI buffer (20 mM Hepes, pH 7.9, 200 mM KCl, 1 mM EDTA, 4 mM MgCl 2 , 1 mM dithiothreitol, 0.1% Nonidet P-40, and 10% glycerol). Bound proteins were resolved by 10% SDS-PAGE. The gels were dried and exposed to a phosphorimaging screen, and the bands were quantified using a Molecular Dynamics Storm 860 reader and ImageQuant software (Molecular Dynamics).

MEF2C Activates and hMusTRD1␣1
Represses Transcription through the hTnI slow USE-The transcriptional regulation of the hTnI slow USE via MEF2C and hMusTRD1␣1 was investigated initially in C2C12 muscle cultures. Expression vectors encoding hMusTRD1␣1 or MEF2C were co-transfected into C2C12 cells with luciferase reporter gene constructs driven by the hTnI slow USE (Fig. 1). MEF2C elicited an approximate 2-fold increase in hTnI slow USE transcriptional activity in differentiated C2C12 cultures (Fig. 1B, left panel). In contrast, hMusTRD1␣1 repressed the basal activity mediated by the hTnI slow USE by ϳ2-fold. To assess the transcriptional effects of hMusTRD1␣1 on MEF2C-mediated transactivation, hMus-TRD1␣1 and MEF2C expression constructs were co-transfected together with the hTnI slow USE. hMusTRD1␣1 repressed the MEF2C-mediated transcriptional activation of the hTnI slow USE by ϳ3-fold. The specific involvement of the Inr-like element in this regulation was examined using USE-B1b that contains a mutation in this element that prevents hMus-TRD1␣1 binding (1) (Fig. 1A). There is a 36-fold reduction in activity upon mutation of the GATTAA core sequence demonstrating that basal transcriptional activity is mediated by the B1 region (Fig. 1B). Furthermore, mutation of this core sequence resulted in a 68-fold reduction of MEF2C-mediated induction of the hTnI slow USE, demonstrating that this site is necessary for MEF2C-mediated transcriptional activation. These results also suggest that an unidentified factor binding FIG. 1. hMusTRD1␣1 represses MEF2C activation through the B1 region in the hTnIslow USE. Transient co-transfections of C2C12 cells were performed using expression constructs for hMusTRD1␣1 (1␣1 944aa ) or MEF2C and (A) reporter constructs containing the intact hTnI slow USE or mutated USE B1b site linked to luciferase (B, C2C12). Transient co-transfections of Cos-7 cells were performed using the expression construct for hMusTRD1␣1 (1␣1 944aa ) and (A) reporter constructs containing the trimerized B1 or B1b regions linked to luciferase (B, Cos-7). Activation or repression is expressed as fold induction of basal (empty) expression vector activity that was set at 1. The columns represent the mean values of triplicates; the bars indicate S.E. C, reverse transcription-PCR was performed on total RNA isolated from undifferentiated C2C12 myoblasts (Mb), differentiated C2C12 myotubes (Mt), Cos-7 cells and embryonic day 13.5 mouse hindlimbs (ED 13.5). PCR was performed on 5 ng of pcDNA3.1hMusTRD1␣1 944aa myc/his (pcDNA) as a positive control.
hMusTRD1␣1 Represses MEF2 Transcriptional Activation to B1 is needed for basal activity and activation mediated by MEF2C and that hMusTRD1␣1 competes with the same DNA site to mediate repression.
hMusTRD1␣1-mediated repression of the USE could be due to an intrinsic property of the protein. Alternatively, it could repress by modifying other myogenic factors present in C2C12 cells, or it could recruit other MusTRD isoforms because these cells express a number of MusTRD isoforms, including the mouse orthologue of hMusTRD1␣1 (44) (Fig. 1C). To discriminate these possibilities, expression studies were conducted in Cos-7 cells that do not express myogenic regulatory factors and express negligible amounts of MusTRD transcripts. In addition, the B1 region of the USE was used so that the Inr-like element was present but not the MEF2-binding site. The B1 region was trimerized and linked to a heterologous promoter driving luciferase to achieve a sufficient level of expression in Cos-7 cells (Fig. 1A). hMusTRD1␣1 was co-transfected together with a construct bearing either the B1 or B1b region. hMusTRD1␣1 repressed basal transcriptional activity by ϳ3-fold (Fig. 1B). Mutation of the core GATTAA sequence in the B1b version resulted in a 4-fold loss of basal transcriptional activity in Cos-7 cells. These data demonstrate that hMusTRD1␣1-mediated repression can occur in the absence of MEF2C and its binding site.
hMusTRD1␣1 Contains Two DNA-binding Domains-To locate the region(s) of hMusTRD1␣1 responsible for DNA binding, we examined the functional capabilities of a deletion series of hMusTRD1␣1 cDNAs. As a first step in this process, we demonstrated that in vitro translated full-length hMusTRD1␣1 (hMusTRD1␣1 944aa ) binds to the Inr-like element within the B1 region of the hTnI slow USE ( Fig. 2A). Oligonucleotides containing either a trimerized B1 or B1b region were incubated with in vitro translated hMusTRD1␣1 944aa protein. A protein-DNA complex formed when the B1 region was used as a probe but did not form with the mutated oligonucleotide. The presence of hMusTRD1␣1 in the slowest migrating complex was demonstrated using an ␣-MusTRD1␣1 1-20aa antibody directed against the first 20 aa of hMusTRD1␣1. The specificity of the antibody was shown by Western immunoblotting (Fig. 2B). A supershift of the upper complex occurred in the presence of the antibody, demonstrating that it contains hMusTRD1␣1 binding to the B1 region.
Truncated versions of hMusTRD1␣1 were generated by progressive C-terminal deletion of putative regulatory regions as well as an N-terminally deleted peptide lacking repeat domains 1 and 2 (⌬N444aa) (Fig. 3A). PCR products were subcloned into the pcDNA3.1myc/his expression vector for in vitro translation of proteins that were 350, 458, 564, 689, 786, and 500 aa in length (Fig. 3B). These proteins and plasmids were used in subsequent EMSA and transfection assays, respectively (Figs. 3C and 4). hMusTRD1␣1 350aa , containing repeat domain 1 only, was unable to bind the B1 region (Fig. 3C). In contrast, hMusTRD1␣1 458aa clearly binds DNA, demonstrating that a DNA-binding domain (DBD1) is located in the N terminus between aa 351 and 458. The presence of a second DNA-binding domain (DBD2) between aa 544 and 944 was demonstrated by the binding of hMusTRD1␣1N⌬ 444aa , which lacks DBD1. hMusTRD1␣1 786aa binds to the B1 region most avidly as indicated by a band of greater intensity in comparison with the other deletions. This suggests that hMusTRD1␣1 786aa may contain both DBDs and that DBD2 may exist in aa 544 -786. The binding activities of hMusTRD1␣1 458aa , hMusTRD1␣1 564aa , hMusTRD1␣1 786aa , and hMusTRD1␣1N⌬ 444aa were lost upon mutation of the B1 region, further indicating that the GATTAA site is the core area for interaction for both DBD1 and DBD2. These data demonstrate that hMusTRD1␣1 contains two DBDs.
hMusTRD1␣1 Can Repress in the Absence of DNA Binding-To determine the functional significance of the two DBDs in hMusTRD1␣1-mediated repression, the hMusTRD1␣1 Cterminal truncation and N-terminal deleted constructs were co-transfected into Cos-7 cells along with a luciferase construct driven by the trimerized B1 region. All constructs repressed activity (Fig. 4). These data demonstrate that hMusTRD1␣1mediated repression is a dominant feature of this protein and that repression capability resides in at least one (aa 1-350) and possibly more regions of the protein (Fig. 3A). In addition, because hMusTRD1␣1 350aa does not contain a DBD but can repress, repression capability does not rely on interaction of the protein with its cognate DNA-binding site.
Mammalian one-hybrid assays in Cos-7 cells were used to confirm that hMusTRD1␣1 can repress in the absence of binding to its cognate DNA-binding site. Fusion proteins containing the DBD of GAL4 linked to hMusTRD1␣1 944aa were constitutively expressed in Cos-7 cells in combination with a luciferase reporter gene construct driven by a trimerized GAL4 DNA-binding site (Fig. 5A). Reporter gene activity was determined and expressed as fold repression over basal GAL4 DBD activity (Fig. 5B). hMusTRD1␣1 944aa was able to repress the basal activity by ϳ3.5fold by utilizing the GAL4 DBD. The repressive capability of hMusTRD1␣1 944aa was tested against the repression domains present in the N-terminal portion of the potent co-repressor NCoR (45). Repression was similar with both factors. Taken together, these results validate the co-transfection results shown in Fig. 4 and demonstrate that hMusTRD1␣1 944aa can repress hMusTRD1␣1 Represses MEF2 Transcriptional Activation without binding to its cognate DNA-binding site and therefore must contain intrinsic repression capabilities. In addition, these results show that the repressive activity of hMusTRD1␣1 is comparable with a known repressor molecule, NCoR, that functions without directly binding DNA.
hMusTRD1␣1 and NCoR Can Physically Interact in Vivo and in Vitro-The ability of hMusTRD1␣1 to interact directly with the common co-repressor NCoR was determined using co-immunoprecipitation assays. hMusTRD1␣1 350aa was used because it is the smallest peptide that can repress in the absence of DNA binding. An antibody that recognizes the Cterminal region of NCoR was used to successfully co-immunoprecipitate a protein complex containing full-length NCoR and hMusTRD1␣1 350aa from Cos-7 cells (Fig. 6A). GST pull-down assays were used to demonstrate in vitro interactions between 35 S-labeled hMusTRD1␣1 350aa and the region of NCoR that contains the nuclear receptor interaction domains responsible for interaction with other proteins, GST-NCoR 1649 -2453 (45) (Fig. 6B). These protein-protein interaction data confirmed that hMusTRD1␣1 can interact with NCoR via a mechanism that is independent of DNA binding. Taken together, these findings suggest that NCoR and hMusTRD1␣1 could cooperate to mediate the transcriptional repression of the hTnI slow gene.
hMusTRD1␣1 Abrogates MEF2C Binding to the hTnI slow USE MEF2 Site through Direct Interaction-We have shown that hMusTRD1␣1 is capable of repressing MEF2C-mediated activation of the USE. The binding of MEF2C to its recognition site within the USE in the presence of hMusTRD1␣1 was examined by EMSA in an effort to understand the mechanism of this repression. Oligonucleotides corresponding to the functional MEF2 binding site within the USE (Ϫ908 to Ϫ891) bound efficiently to in vitro translated MEF2C. However, oligonucleotides containing the same sequence except for four altered bases in the core recognition sequence (mutMEF2) FIG. 3. hMusTRD1␣1 contains two DNA-binding domains. A, hMusTRD1␣1 contains five TFII-I-like repeated domain structures of 79 -93aa (RD). C-and N-terminally truncated versions of hMusTRD1␣1 were generated by progressive removal of each repeat domain starting at the carboxyl end of the molecule, in addition to an N-terminally deleted hMusTRD1␣1 lacking repeat domains 1 and 2 (⌬N444aa). B, in vitro translation products of C-and N-terminally deleted versions of hMusTRD1␣1 (black arrowheads). C, gel shift experiments were performed with in vitro translated hMusTRD1␣1 proteins that were incubated with a trimerized B1 (B1(3X)) or mutated trimerized B1b (B1b(3X)) region from the hTnI slow USE. DNA:hMusTRD1␣1-specific complexes are shown (arrowheads), and nonspecific (NS) complexes are also indicated (*). The presence of DBD1 and/or DBD2 in the peptide is indicated.
hMusTRD1␣1 Represses MEF2 Transcriptional Activation failed to bind, thus demonstrating the specificity of the interaction (Fig. 7A). The addition of increasing amounts of hMus-TRD1␣1 blocked the interaction between MEF2C protein and the oligonucleotide containing its binding site in a dose-dependent manner (Fig. 7A).
This finding suggests that hMusTRD1␣1-mediated repression could occur through a direct interaction between hMus-TRD1␣1 and MEF2C. To test this further, GST pull-down assays were performed using GST-hMusTRD1␣1 fusion pro-tein and 35 S-labeled in vitro translated MEF2C. A band corresponding to MEF2C is visible in the GST pull-down eluates, and quantitation using phosphorimaging analysis demonstrated a significant increase in binding efficiency of GST-hMusTRD1␣1 over the GST control (Fig. 7B). This shows that MEF2C and hMusTRD1␣1 are capable of direct interaction.

DISCUSSION
These data address key aspects of the transcriptional regulation of hTnI slow and potentially of many slow fiber-specific genes. The Inr-like element within the hTnI slow USE is an essential regulatory element involved in both transcriptional activation and repression. We demonstrate that mutation of the Inr-like element prevents MEF2C-mediated activation via the USE. This key finding shows that USE function relies on cooperation between proteins binding at the Inr-like site and MEF2 proteins. The data are consistent with a model whereby MEF2-mediated transcriptional activation occurs through the functional MEF2binding site of the USE and requires occupancy of the Inr-like element by an unknown activating factor(s) (Fig. 8A). hMus-TRD1␣1 is found to repress activity of the USE, and this may be achieved through interaction with the Inr-like element or in a DNA-independent manner (Fig. 8B). hMusTRD1␣1 can bind to the Inr-like element by virtue of two sequence-specific DNAbinding domains and therefore may compete with the activating factor for occupancy of the binding site. Alternatively, hMus-TRD1␣1 may utilize its ability to interact directly with NCoR and

hMusTRD1␣1 Represses MEF2 Transcriptional Activation
MEF2C. It could prevent the proper interaction of MEF2C with the enhancing factor either by sequestering MEF2C in conjunction with NCoR or by occupying the Inr-like element (Fig. 8B).
A growing number of signaling pathways have been found to converge on the MEF2 proteins, thereby regulating their essential role in transcriptional control of muscle-specific genes (reviewed in Ref. 47). MEF2C and MEF2A are substrates for p38 mitogen-activated protein kinase (48,49), and MEF2C is a substrate for BMK1/ERK5 (50). Other pathways involve calcium-sensitive proteins including calcineurin, which activates MEF2 by direct dephosphorylation (32) and Ca 2ϩ -calmodulindependent protein kinase, which activates MEF2 by alleviating the repression imposed by the histone deacetylases (HDACs) (51)(52)(53). Signaling via Ca 2ϩ -dependent pathways has recently been proposed as a potential means of fiber-type adaptation because sustained patterns of nerve stimulation elevate intracellular calcium levels (20), and recent experiments in mice support this hypothesis (31,54). Additionally, the transcriptional co-activator PGC-1␣ has been found to elicit slow fiberspecific gene expression, as well as mitochondrial biogenesis, through interaction with MEF2 (30). Therefore, MEF2 is a key factor involved in translating endurance activity and electrical activity-mediated changes in intracellular Ca 2ϩ levels within muscle fibers into muscle gene transcriptional activity. It is clear that by disrupting the transcriptional activation capacity of MEF2C in the USE of hTnI slow , as demonstrated in this study, hMusTRD1␣1 acts at a nodal point in the regulation pathway of slow fiber-specific genes.
hMusTRD1␣1 repression can occur through at least two distinct mechanisms. hMusTRD1␣1 can interact directly with the B1 region via two DBDs. Truncated versions of hMusTRD1␣1 containing either or both DBDs bind specifically to the B1 region of the USE through the core binding motif GATTAA defined by O'Mahoney et al. (1). In all instances, mutation of this sequence to GATatc prevented binding. Furthermore, the A, transcriptional transactivation through the hTnIslow USE is achieved when the B1 region is occupied by an enhancing factor ("X") interacting with MEF2C bound to the functional MEF2C-binding site. B, repression is achieved either by the sequestering of MEF2C by an NCoR:hMusTRD1␣1 (1␣1) complex or occupation of the B1 region by hMusTRD1␣1 that prevents the binding of the enhancing factor. hMusTRD1␣1 Represses MEF2 Transcriptional Activation ␣-hMusTRD1␣1 1-20aa antibody was capable of directly blocking hMusTRD1␣1 DNA binding activity, further verifying the specificity of this DNA-protein complex. The DBDs in hMusTRD1␣1 are located within aa 351-458 and 544 -944. Calvo et al. (42) recently reported the isolation of a series of truncated GTF3 proteins encoded by the mouse orthologue of the hMusTRD1␣1encoding gene. The GTF3 proteins were isolated by their ability to bind to the rat TnI slow slow upstream regulatory element, which acts similarly to the human USE. On the basis of the truncations, it was concluded that the DBD was located in the C-terminal region, which would be consistent with DBD2 demonstrated here.
The association of hMusTRD1␣1 with the nuclear receptor co-repressor N-CoR demonstrates a second mode of repression. NCoR functions as a co-repressor not only for nuclear hormone receptors but also for multiple classes of transcription factors (55,56). The multiple, N-terminal repression domains mediate interactions via mSin3 with large complexes containing class I HDACs (57) or, by direct association, with class II HDACs (58), thereby modifying chromatin structure through histone hypoacetylation. Co-immunoprecipitation studies recently revealed an association between hMusTRD1␣1 and the class I histone deacetylase HDAC3 (59). It is possible that this interaction is mediated by an NCoR-dependent mechanism. In support of this claim, Wen et al. (59) found that in addition to HDAC3, eight other proteins including NCoR and TFII-I, a structurally related to hMusTRD1␣1 protein, specifically coeluted with histone deacetylase activity through an anti-HDAC3 immune affinity column.
Recent reports have demonstrated that the mouse MusTRD homologues GTF3 and BEN also act as transcriptional repressors in other experimental systems (40,42). BEN was shown to repress TFII-I-mediated activation of the c-fos promoter despite its apparent inability to bind either to the c-fos promoter or to TFII-I (43). It was concluded that the repression occurs through competition for nuclear shuttling components and transcriptional co-factors. On the basis of current knowledge, it would seem that the MusTRDs are complex multifaceted proteins distinctly different from the basic helix-loop-helix transcription factor family. The discovery of multiple splice isoforms of MusTRD, which modify regions of the C-terminal half (44), may be viewed as an additional complexity. On the other hand, the functional consequences of these modifications may provide an entry point for understanding the diverse mechanisms by which the MusTRDs regulate transcription.