A Muscle-specific Promoter Directs Pitx3 Gene Expression in Skeletal Muscle Cells*

The Pitx homeobox transcription factor genes have been implicated in different developmental processes, including determination of hind limb identity for Pitx1, left-right asymmetry for Pitx2, and eye development and survival of midbrain dopaminergic neurons for Pitx3. Pitx1 and Pitx2 have partly redundant activities in craniofacial development, including in pituitary organogenesis, as indicated by their names. These genes also exhibit redundant activities in the control of hind limb bud growth. Recent studies have shown expression of the three Pitx genes in muscle, with Pitx3 being the most widely expressed in all skeletal muscles. We now report the identification of a muscle-specific promoter within the Pitx3 gene that is situated between the first exon for eye and brain expression and exon 2 that contains the initiator ATG codon. Sequences proximal to this muscle-specific exon 1 are essential and sufficient to confer muscle-specific expression in transgenic mice, they are responsive to myogenic basic helix-loop-helix regulatory factors, and they recruit these factors in vivo. In agreement with exclusive use of the muscle-specific promoter in aphakia mice that are deleted of the brain promoter, the trimethyl-lysine 4 histone H3 promoter signature shifts to this promoter in embryonic day 13 ak limb bud muscle cells. Myogenic basic helix-loop-helix regulatory factor activation of Pitx3 transcription may be part of a positive feedback loop contributing to establishment of the myogenic program.

The Pitx genes play critical roles in early development of several tissues and organs. The first of these homeobox transcription factors, Pitx1 (Ptx1), was discovered for its role in cell-specific transcription of the pituitary Pomc gene (1); on the Pomc gene, it acts as an obligate partner of the highly cell-restricted T-box transcription factor, Tpit (2,3). Beyond their role in pituitary organogenesis and transcription as implied by their names (4,5), the Pitx genes play several roles in early development. Indeed, Pitx1 is first expressed in the lateral plate mesoderm of the posterior half of the embryo, starting as soon as posterior lateral plate mesoderm is formed at gastrulation (6). Throughout later development, Pitx1 remains highly expressed in developing hind limb bud mesenchyme and was shown to determine hind limb identity in mice (7,8), chicks (9,10), and fish (11). In addition, Pitx1 controls the growth ability of hind limb mesenchyme together with Pitx2 (12). Pitx2 is also expressed in the lateral plate mesoderm but preferentially on the left side, where its left-specific expression is controlled by Sonic hedgehog (Shh) and/or nodal (13). The left-specific Pitx2c isoform uses a different promoter than for its bilateral expression in other tissues (14). This alternate promoter is responsible for production of Pitx2a and Pitx2b that arise through differential splicing of exon 2 (4). Left-specific expression is critical for left-right asymmetry and development of the lungs, heart, and stomach (14,15). In the heart, the left-specific isoform of Pitx2, Pitx2c, has activities that are specific or redundant with those of the bilaterally expressed isoforms Pitx2a and Pitx2b (16).
Finally, the third gene of the family, Pitx3, is expressed in the eye (17) and midbrain dopaminergic neurons (18). In agreement with this expression, a natural mouse mutant of the Pitx3 gene promoter, the aphakia (ak) mouse, has lens developmental defect (17), and in humans, Pitx3 mutations have been associated with eye dysmorphogenesis (anterior segment mesenchymal dysgenesis) and with congenital cataracts (19). The ak mouse was also found to exhibit a severe deficit of dopaminergic neurons in the substantia nigra and ventral tegmental area (20 -22) and to have an akinetic phenotype similar to the loss of motor movement presented by patients with Parkinson disease (22). Similar to Parkinson patients, this mouse model exhibits a locomotor response to L-Dopa injection (23).
Early reports on expression profiles of Pitx genes had noted expression of Pitx1 in a small subset of muscles (6), of Pitx2 in early myoblasts (24), and of Pitx3 in extraocular muscles, tongue and head muscles (19). 5 Muscle expression of these genes was not studied systematically until recently, as we conducted a detailed analysis of their expression patterns in early precursors and in developing muscle (25). It thus appears that Pitx3 is expressed in all skeletal muscles. Since Pitx3 is the only member of the family to be expressed in all muscles, we have undertaken to define the mechanism of its expression and its role in muscle development. Since a natural mouse mutant of the Pitx3 gene already exists, the ak mouse, we assessed muscle expression of Pitx3 in ak mice and found it to be normal. The present report delineates a novel muscle-specific promoter within intron 1 of the Pitx3 gene that is not affected by the ak mutation. Further, we defined the muscle-specific Pitx3 promoter in transgenic mice experiments and showed that it is adequate to target gene expression in early skeletal muscle, thus mimicking expression of the endogenous gene. In normal muscle, both brain and muscle initiation sites appear to be used, but only muscle-specific exon 1m is transcribed in ak mice, in agreement with a shift of the trimethyl-lysine 4 histone H3 (H3K4me3) promoter signature in ak muscle cells. Finally, we show that Pitx3 expression is stimulated by myogenic bHLH 6 regulatory factors (MRFs) in a positive feedback regulatory mechanism that serves to establish the myogenic program of differentiation.

EXPERIMENTAL PROCEDURES
Whole Mount in Situ Hybridization-The protocol used (available on the World Wide Web) is from the laboratory of Dr. Janet Rossant. A rat Pitx3 cDNA EcoRI-BamHI fragment encompassing the entire translated sequence was used as probe.
RNA Analyses-Total RNA was prepared from adult mouse muscle, spleen, or e14.5 embryos with TRIzol (Invitrogen) according to the manufacturer's directions. Purification of mRNA was performed using the Qiagen (Valencia, CA) Oligotex Maxi Kit following the supplier's protocol.
Transgene Constructs-Mouse Pitx3 gene sequences were cloned from a genomic 129/sv DASH2 phage library (1). A genomic 4-kb Pitx3 promoter fragment was inserted upstream of LacZ (from ␣-GSU-LacZ) (28) to create plasmid JA1639. The addition of the 3Ј two-thirds of Pitx3 intron 1 yielded JA1642. JA1673 was obtained by deleting the 2.2-kb putative alternative promoter region in JA1642. For TG1704 and TG1707 transgenes, the Pitx3 4-kb promoter in TG1639 was replaced by an intronic fragment corresponding to the conserved promoter region (corresponding to Ϫ1775/ϩ12 of the muscle promoter) with or without a less conserved 1.5-kb upstream region, respectively. Insertion of the 3-kb 5Ј part of intron 1 into TG1704 yielded TG1706. All constructs were checked by restriction and sequencing; maps and sequences are available upon request.
Transgenic Mice-All DNA fragments for microinjection were produced by SacII-SalI digestion and purified on agarose gel. Transgenic mice were produced as previously described (29). For transient transgenics, foster mice were sacrificed at 12.5 days postcoitum, taking the day of injection as 0.5 days postcoitum.
Cell Culture and Transfection-L6 and C2C12 cells were cultured in Dulbecco's modified Eagle's medium containing 20% fetal bovine serum and transfected in Dulbecco's modified Eagle's medium without serum for 4 h with Lipofectamine according to the manufacturer's directions. Usually, for 35-mm wells, 1 g of ␤-galactosidase reporter was used per well. For differentiation, cells were grown to confluence, and then medium was changed to Dulbecco's modified Eagle's medium containing 2% horse serum, and cells were maintained in culture for 24 or 48 h. In experiments to assess responsiveness to MyoD, expression vectors for both MyoD and E47 (same amount for each) were co-transfected with reporter, and reporter activity was measured 24 h posttransfection in proliferation conditions. ␤-Galactosidase activity was assayed in lysates with a luminometer using the Galacto-Light system (TROPIX) as described by the manufacturer or by cytochemistry using X-gal as a substrate in formaldehyde-fixed cells to correlate activity with differentiated morphology.
Software-Sequence alignments were performed using ClustalW (available on the World Wide Web), and graphic genomic alignments were performed with LAGAN and VISTA (both available on the World Wide Web). The promoter prediction software is also available online.

Pitx3 Is Expressed in Developing Muscles of Aphakia Mice-In
order to establish expression of Pitx3 in developing muscles, we used whole mount in situ hybridization with mouse embryos at day 12.5 (e12.5) of development. This analysis revealed expression of Pitx3 in myotomes and developing muscle masses of the limbs (Fig. 1, A and B). Muscle masses of both fore limb and hind limb buds express Pitx3, and expression is also observed in forming abdominal muscle masses. It thus appears that Pitx3 is very widely expressed in most developing skeletal muscles. We provided a detailed investigation of this expression profile elsewhere (25).
Since the ak mouse was shown to be a promoter mutant of the Pitx3 locus (30, 31), we wanted to verify a putative defect of Pitx3 expression in muscle. Surprisingly, we found that muscle expression of Pitx3 is intact in ak embryos and that in this tissue, it appears undistinguishable from wild-type littermates (Fig. 1, C and D). In order to verify that Pitx3 protein indeed has normal expression in ak muscles and further that it is expressed in adult muscle, we performed Western blot analyses of hind limb muscles using an antibody specific for Pitx3. This analysis revealed a similar band in both wild-type and ak mice at about 38 kDa (Fig. 1E).
A Muscle-specific Pitx3 Exon 1-Since the ak deletion removes the brain and eye (lens) exon 1 and the upstream region of the previously described Pitx3 gene (31), we reasoned that a different exon 1 and promoter might be active in muscles and that this putative muscle exon 1 could be spliced onto a common exon 2. The quest for an alternate initiation site for muscle Pitx3 expression was facilitated by the search of genome data bases for expressed sequence tags containing Pitx3 sequences. Indeed, a BLAST search on the UCSC genome server identified a rat expressed sequence tag (BF543032) matching with exons 2 and 3 of Pitx3 but containing different exon 1 sequences (thereafter called exon 1m for "muscle") by comparison with the already known exon 1 (hereafter called exon 1l,b for lens, brain expression). The structure of the Pitx3 mouse locus is schematically represented in Fig. 2A, showing the position of both exon 1l,b and the novel putative exon 1m; the figure also shows the position of the two genomic deletions identified in ak mice (30,31). It is interesting to note that putative exon 1m lies within a region of similar sequence conservation as those of exons 1l,b, 2, 3, and 4 (Fig. 2B).
In order to pinpoint a putative transcription start site for exon 1m, the BDGP neural network promoter prediction software (32) was used to predict putative initiation site(s) as indicated in Fig. 3A; this prediction was made with a high score (0.83, where the maximum is 1). This putative start site lies within a region of high sequence conservation between mouse, rat, and human sequences (Fig. 3A). In order to validate this prediction, a series of RT-PCRs were carried out using e14.5 embryo mRNA and different primers. A common 3Ј primer located in exon 2 was used in all cases, and different 5Ј primers (Fig. 3A) located downstream of putative initiation (primer 5, 4, 3, 2, or 1) or just upstream of this predicted start site (primer 6, 7, or 8) were used. No RT-PCR products could be amplified using adult spleen RNA as control (data not shown), and a PCR product of the expected size was obtained with primer 5 located between putative nucleotides Ϫ6 and ϩ18 (Fig. 3B). In contrast, no PCR product could be obtained with primer 6, which lies at positions Ϫ45 to Ϫ26 bp relative to the putative start site (Fig. 3B). In order to verify that primer 6 was active in PCR, control reactions were conducted using genomic DNA substrate; a fragment of the expected size was obtained (Fig. 3C). DNA sequencing of the RT-PCR products confirmed that this fragment contains the expected exon 1m, exon 2, and exon 3 sequences, as predicted for the proposed muscle mRNA shown in Fig. 2A. Similar results were obtained with other reverse primers. It is noteworthy that the initiator ATG codon is found in exon 2 of the Pitx3 gene; it is therefore predicted that the Pitx3 protein should be the same in muscle as in eye and midbrain. These data suggest that muscle expression of Pitx3 is driven by a promoter different from the one used in eye lens and brain, but this activity may still use the exon 1l,b initiation site. Indeed, using RT-PCR with a forward primer in exon 1l,b and the exon 2 reverse primer, it was found that the exon 1l,b is also used in muscle together with exon 1m (Fig. 3B). In order to assess the relative use of each first exon in muscle, quantitative RT-PCR was performed (Fig. 3D). This analysis revealed about 8 times more exon 1l,b-containing mRNA compared with exon 1m. In agreement with the deletion of exon 1l,b in ak mice, no PCR product corresponding to exon 1l,b sequences was detected in ak muscle, but the relative abundance of exon 1m in these tissues was increased about 2.8-fold compared with wildtype muscle (Fig. 3D).
A Muscle Cell-specific Pitx3 Gene Promoter-In order to test whether the putative Pitx3 promoters are active in vivo and show the expected specificity, we tested the two predicted promoter sequences of the Pitx3 locus in transgenic mice. The promoter fragments were fused to ␤-galactosidase (lacZ) coding sequences, and founder e12.5 transgenic embryos were assessed for transgene expression using LacZ activity as reporter. A 4-kb promoter fragment for exon 1l,b that is expressed in eye lens and midbrain (mb) was found to be largely inactive in transgenic embryos (Fig. 4, construct F). This transgene did not exhibit any expression in lens or midbrain and exhibited only weak expression in the myotome region of only one embryo. In contrast, four of seven embryos had faint ectopic expression at unrelated sites. Stable mouse transgenic lines had been established with the same transgene previously, and they showed similar weak and inconsistent patterns of expression.  NOVEMBER 9, 2007 • VOLUME 282 • NUMBER 45

JOURNAL OF BIOLOGICAL CHEMISTRY 33195
Various putative promoter fragments for exon 1m were also assessed in transgenic mice, and the most consistent and strongest expression was observed with the shortest transgenic promoter fragment (construct A). This 1.6-kb promoter fragment contains two regions of high sequence conservation: a promoter-proximal region of about 350 bp and another upstream region of about 300 bp at Ϫ1450 bp (Fig. 2B). This transgene directed very strong expression in myotomes, limb, and facial muscles with very infrequent ectopic expression (Fig. 4, construct A); LacZ activity was never detected in midbrain or eye lens. The addition of more upstream sequences (constructs B and C) did not further enhance muscle expression of transgenes; in fact, expression was weaker ( Fig. 4) with particular loss of expression at the level of craniofacial muscles, of the epaxial domain of the myotome, and of the most caudal somites (beyond the hind limb level). In fact, transgene expression in face muscles was only consistently observed for construct A; nine of 11 founders had facial expression, compared with three of nine for construct B and four of 17 for construct C. A short region of unconserved upstream sequences (absent in construct C) could not be tested in transgenic mice, because the presence of these sequences rendered plasmids very unstable. Since construct F conferred expression at low frequency (one of seven) in putative myogenic cells, we wanted to test whether these sequences may act in concert with the muscle-specific promoter and whether the musclespecific sequences could act as enhancers. A transgene containing both promoters (construct D) exhibits similar muscle specificity as construct B or C; it is also less active than construct A. In agreement with the hypothesis that the highly conserved sequences around exon 1m (Fig. 2B) that are active in construct A are important for muscle-specific activity, it was observed that deletion of those sequences completely abolished transgenic promoter activity (Fig. 4, construct E). Collectively, these data clearly identify a muscle-specific promoter/enhancer upstream of exon 1m of the Pitx3 gene and indicate that this promoter is necessary and sufficient for muscle-specific expression of Pitx3. The maintenance of Pitx3 expression in ak muscles is completely consistent with the existence of separate regulatory sequences for muscle-specific expression of the gene.
The Pitx3 Muscle-specific Promoter Is Active in Myoblasts and in Differentiating Muscle Cells-The early pattern of muscle expression for transgene A (Fig. 5, A and B) is extremely similar to the pattern of Pitx3 mRNA expression (Fig. 1). It includes forming skeletal muscle masses of limbs, as revealed by analysis of tissue sections from X-gal-stained embryos (Fig. 5, C and D). In order to ascertain the muscle specificity of this LacZ activity, we performed co-labeling by immunohistochemistry with Pitx3 or MyoD. This demonstrated co-expression of MyoD with LacZ in developing FL muscle masses (Fig.  5, E and F) and in the myotome (Fig.  5, G and H). A similar distribution was observed for Pitx3 and LacZ (Fig. 5, I and J). Only a subset of MyoD-positive cells stain for LacZ in both myotome and limb muscles; in more caudal myotomes (Fig. 5I), LacZ-positive cells are also less abundant than Pitx3-positive cells (Fig. 5J), suggesting a delay between endogenous Pitx3 and transgene A LacZ activity. These data clearly indicate that the muscle-specific exon 1m promoter of the Pitx3 gene is only active in muscle cells.
Myogenic bHLH Regulatory Factors Activate the Pitx3 Muscle Promoter-In order to further define the Pitx3 muscle promoter, we used transfection of promoter constructs into muscle cell lines. Initially, we tested the activity of the Ϫ1597 bp Pitx3 promoter (construct A) fused to LacZ in proliferating C2C12 cells, and we observed very low ␤-galactosidase activity (Fig. 6A). However, significant activity was observed with the same reporter in C2C12 cells following 24 h in differentiation conditions (33). In agreement with this promoter/reporter activation, endogenous Pitx3 mRNA levels were found to increase in similar differentiation relative to proliferation conditions (Fig. 6B). We then assessed the effect of MyoD on the Ϫ1597 bp Pitx3 promoter using L6 cells, another muscle-derived cell line; similar results were obtained in C2C12 cells (data not shown). This experiment clearly indicated that MyoD can activate the Pitx3 muscle promoter (Fig. 6C). We then localized MRF-responsive sequences using deletions of the Ϫ1597 bp reporter plasmid. Thus, deletion to Ϫ590 bp (Del B) completely prevented responsiveness to MyoD (Fig. 6C). In contrast, deletion to Ϫ1388 bp (Del A) did not affect the responsiveness to MyoD, suggesting that active sequences are present between Ϫ1388 and Ϫ590 bp (Fig. 6C). Interestingly, numerous muscle-specific E-boxes that are the targets of MRFs are present in this interval, and the recruitment of MyoD to the Pitx3 locus was recently shown using a ChIP-on-chip strategy (34) in agreement with the present data showing activation of the Pitx3 promoter. . In vivo activity of the Pitx3 muscle-specific promoter. Various transgenes (A-F) were constructed using putative promoter sequences for eye/brain or muscle expression of Pitx3, as shown in the diagrams, and fused with the ␤-galactosidase (LacZ) reporter gene. For each transgene, e12.5 transgenic embryos were stained for ␤-galactosidase activity. A representative embryo for each transgene is shown on the right of the table giving the number of embryos showing expression in each of muscle, lens, midbrain (mb), or ectopic sites. In each case, the total number of embryos examined is indicated.
The differentiation of C2C12 cells is accompanied by up-regulation of MyoD activity, and in agreement with the deletion analyses (Fig. 6C), we found that the different deletion mutants of the Pitx3 promoter exhibit similar activity when assessed upon differentiation of C2C12 cells, with the Ϫ590 bp promoter (Del B) showing no activity (Fig. 6D). Collectively, these data are in agreement with the expression of Pitx3 in differentiating muscle cells (25) and with a scheme in which Pitx3 expression is subject to positive feedback up-regulation by MyoD during muscle cell differentiation.
Chromatin Signature of Muscle-specific Promoter-The action of MyoD on the Pitx3 muscle regulatory sequences appeared to depend principally on sequences between Ϫ1388 and Ϫ590; this 702-bp region contains multiple putative MRF binding sites, and we developed a PCR strategy to investigate MRF recruitment to this region by ChIP (Fig. 7A). We found significant recruitment of both MyoD and myogenin to this region (Fig. 7B, primers C) compared with the exon 1l,b regulatory region (primers A) or with exon 4 of the gene (primers E).
It appears that both promoters 1l,b and 1m are active in muscle (Fig. 3D), although only promoter 1m was active in transgenics (Fig. 4). In order to validate this at the chromatin level, we used ChIP against H3K4me3 that has been reported to be a signature for active promoters (35,36). This analysis (Fig. 7C) revealed a higher level of H3K4me3 at the exon 1l,b promoter (primers B) than at exon 1m (primers D), but the level of H3K4me3 at exon 1m was significantly increased in ak mice and was absent from the exon 1l,b region when probed with primers B (Fig. 7A) that are beyond the deletion end point in the ak genome. Although not different in WT and ak limb buds, the levels of H3K4me3 at exon 4 are unusually high for transcribed FIGURE 5. Muscle specificity of transgenic Pitx3 promoter activity. A transgene A e12.5 transgenic embryo is shown after whole mount in situ hybridization (A and B) for comparison with whole mount in situ hybridization for endogenous Pitx3 (Fig. 1). Expression is observed in elongating myotomes (B) as well as in developing limb bud muscle masses. The embryo was sectioned for co-staining, as indicated.  sequences; it is possible that this region may contain another promoter, possibly for antisense transcripts (37). In agreement with the levels of H3K4me3 and with relative promoter usage in muscle (Fig. 3D), more RNA polymerase II was found to be present at exon 1l,b than exon 1m in muscle (Fig. 7D), but this ratio was reversed in ak muscles. Collectively, these data indicate that both promoters 1l,b and 1m are used in muscle cells but that the muscle regulatory sequences are primarily situated upstream of exon 1m.

DISCUSSION
The present report defines a skeletal muscle-specific transcription unit of the Pitx3 gene. This gene was previously known to be expressed in the eye lens and in a subset of dopaminergic neurons of the midbrain through the activity of a common promoter and exon 1. The present report defines a new promoter located between the eye/midbrain promoter and the ATG-containing exon 2. The muscle-specific Pitx3 promoter appears to be entirely contained within a 1.6-kb DNA fragment that mimics in transgenic mice the expression of endogenous Pitx3 in muscles. The muscle-specific Pitx3 promoter can be activated by the myogenic bHLH regulatory factor, MyoD. The action of MRFs on the Pitx3 promoter is consistent with the expression pattern of Pitx3 during muscle development.
A Muscle-specific Promoter and Exon of the Pitx3 Locus-The skeletal muscle-specific Pitx3 transcription unit defined in the present work includes a new exon 1 (exon 1m) and important regulatory sequences located within 1659 bp upstream (Figs. 4 -7). Transgenic mice analyses of this muscle-specific promoter suggest that it is entirely sufficient for muscle expression (Fig. 4). The activity of the exon 1m promoter in transgenic mice is entirely consistent with the expression in developing myotomes of a GFP knock-in allele of the Pitx3 gene (38). The muscle-specific activity of exon 1m regulatory sequences in transgenic mice contrasts with its partial usage in normal muscle. Indeed, all criteria (mRNA abundance, RNA Pol II recruitment, and H3K4me3 enrichment) indicate that exon 1l,b is used preferentially even under the influence of muscle regulatory sequences. It is only upon deletion of exon 1l,b in ak mice that this initiation is no longer used in favor of exon 1m (Fig.  3D) and that H3K4me3 enrichment at exon 1l,b is lost (Fig. 7C). Exon 1l,b appears to have strong transcription initiation capacity; it is used in the brain and eye, but regulatory sequences directing this activity have not yet been identified and are not within the upstream 4 kb (Fig. 4, transgene F). The muscle regulatory sequences (between Ϫ1388 and Ϫ590 bp) that are 6 kb downstream from exon 1l,b appear to preferentially use this initiation relative to exon 1m initiation that is closer. The mechanism for promoter selection in such case remains poorly understood.
Role of Pitx Transcription Factors in the Program of Muscle Gene Expression-The onset of Pitx3 expression with differentiation of Pitx2-positive progenitors (25) is consistent with the activity of the Pitx3 exon 1m shown in transgenic mice (Figs. 4 and 5). In addition, the maintenance of Pitx2 expression in Pitx3 Ϫ/Ϫ muscles suggests that at least one Pitx factor must be present for maintenance of the myogenic programs (25) and that Pitx2 may contribute in part to activation of Pitx3 expression. The earliest Pitx gene expressed during muscle cell differentiation is Pitx2 with expression in proliferating muscle pro- genitors (24,25,39). This early expression is progressively replaced by Pitx3 in differentiated and postmitotic muscle cells (25). Pitx3 expression persists in adult skeletal muscles (Fig. 1E).
Whereas MRFs expression precedes that of Pitx2 in myotomes, it is the reverse in developing limb bud muscle masses, with Pitx3 expression following Pitx2, together with MyoD and cell differentiation (25). In this context, the present demonstration of an activation of the Pitx3 exon 1m promoter by MyoD is likely to represent one mechanism responsible for activation of the Pitx3 muscle-specific promoter. In addition, MRF activation of the Pitx3 promoter may also contribute to a positive feedback loop by which MRFs and Pitx3 reinforce the myogenic program. This positive autoregulatory loop could operate both in myotome-derived and limb muscles. Direct synergism between Pitx3 and MRFs may also serve a maintenance and amplification function for muscle-specific gene transcription. Indeed, this is consistent with the action of Pitx3 on genes encoding muscle-specific proteins, including regulatory factors, such as MyoD and myogenin, and structural proteins like troponins. 7 These genes are also targets of MRFs and synergistic activation of transcription is exerted on them by Pitx3 and MRFs. Taken collectively, the data suggest that Pitx factors play essential roles in muscles to maintain the myogenic program of gene expression. The identification of muscle-specific regulatory sequences (Fig. 4) within the Pitx3 locus as reported here is consistent with such a role in muscle function.