Two upstream enhancers collaborate to regulate the spatial patterning and timing of MyoD transcription during mouse development.

MyoD is a member of the basic-helix-loop-helix (bHLH) transcription factor family, which regulates muscle determination and differentiation in vertebrates. While it is now well established that the MyoD gene is regulated by Sonic hedgehog, Wnts, and other signals, it is not known how MyoD transcription is initiated and maintained in response to these signals. We have investigated the cis control of MyoD expression to identify and characterize the DNA targets that mediate MyoD transcription in embryos. By monitoring lacZ reporter gene expression in transgenic mice, we show that regulatory information contained in 24 kb of human MyoD 5' flanking sequence is sufficient to accurately control MyoD expression in embryos. Previous studies have identified two muscle-specific regulatory regions upstream of MyoD, a 4-kb region centered at -20 kb (designated fragment 3) that contains a highly conserved 258-bp core enhancer sequence, and a more proximal enhancer at -5 kb, termed the distal regulatory region (DRR), that heretofore has been identified only in mice. Here, we identify DRR-related sequences in humans and show that DRR function is conserved in humans and mice. In addition, transcriptional activity of MyoD 5' flanking sequences in somites and limb buds is largely a composite of the individual specificities of the two enhancers. Deletion of fragment 3 resulted in dramatic but temporary expression defects in the hypaxial myotome and limb buds, suggesting that this regulatory region is essential for proper temporal and spatial patterning of MyoD expression. These data indicate that regulatory sequences in fragment 3 are important targets of embryonic signaling required for the initiation of MyoD expression.

Transcriptional mechanisms regulating MyoD expression in the mouse.

The basic-loop-helix transcription factors, MyoD and Myf-5, are essential regulators of skeletal muscle lineage determination in vertebrates. Investigations into the signaling and response systems regulating MyoD and Myf-5 gene expression in muscle progenitor cells have led to the identification of molecular signals and DNA regulatory regions that control expression of these muscle regulatory genes. Here we discuss recent advances in these areas, with emphasis on in vivo analyses of MyoD transcription.

Fine-scale transgenic mapping of the MyoD core enhancer: MyoD is regulated by distinct but overlapping mechanisms in myotomal and non-myotomal muscle lineages.

Skeletal muscle lineage determination is regulated by the myogenic regulatory genes, MyoD and Myf-5. Previously, we identified a 258 bp core enhancer element 20 kb 5' of the MyoD gene that regulates MyoD gene activation in mouse embryos. To elucidate the cis control mechanisms that regulate MyoD transcription, we have mutagenized the entire core enhancer using linker-scanner mutagenesis, and have tested the transcriptional activity of enhancer mutants using lacZ reporter gene expression in transgenic mouse embryos. In total, 83 stable transgenic lines representing 17 linker-scanner mutations were analyzed in midgestational mouse embryos. Eight linker-scanner mutations resulted in a partial or complete loss of enhancer activity, demonstrating that MyoD is primarily under positive transcriptional control. Six of these mutations reduced or abolished transgene expression in all skeletal muscle lineages, indicating that activation of MyoD expression in trunk, limb and head musculature is regulated, in part, by shared transcriptional mechanisms. Interestingly, however, two adjacent linker-scanner mutations (LS-14 and LS-15) resulted in a dramatic reduction in transgene expression specifically in myotomes at 11.5 days. At later stages, transgene expression was absent or greatly reduced in myotomally derived muscles including epaxial muscles (deep back muscles) and hypaxial muscles of the body wall (intercostal muscles, abdominal wall musculature). In contrast, head muscles, as well as muscles of the body derived from migrating muscle progenitor cells (e.g. limb, diaphragm), were unaffected by these mutations. In Pax-3-mutant mice, LS-14 and LS-15 transgene expression was eliminated in the body, but was unaffected in the head, yielding an identical expression pattern to the endogenous MyoD gene in mice mutant for both Myf-5 and Pax-3. These data support the hypothesis that LS-14 and LS-15 define the core enhancer targets for Myf-5-dependent activation of MyoD in myotomal muscles.

Myogenic determination occurs independently in somites and limb buds.

Gene targeting has indicated that the bHLH transcription factors Myf-5 and MyoD are required for myogenic determination because skeletal myoblasts and myofibers are entirely ablated in mouse embryos lacking both Myf-5 and MyoD. Entrance into the skeletal myogenic program during development occurs following the independent transcriptional induction of either Myf-5 or MyoD. To identify sequences required for the de novo induction of MyoD transcription during development, we investigated the expression patterns of MyoD-lacZ transgenes in embryos deficient in both Myf-5 and MyoD. We observed that a 258-bp fragment containing the core of the -20-kb MyoD enhancer activated expression in newly formed somites and limb buds in compound mutant embryos lacking both Myf-5 and MyoD. Importantly, Myf-5- and MyoD-deficient presumptive muscle precursor cells expressing beta-galactosidase were observed to assume nonmuscle fates primarily as precartilage primordia in the trunk and the limbs, suggesting that these cells were multipotential. Therefore, cells are recruited into the MyoD-dependent myogenic lineage through activation of the -20-kb MyoD enhancer and this occurs independently in somites and limb buds.

MyoD and Myf-5 define the specification of musculature of distinct embryonic origin.

Mounting evidence supports the notion that Myf-5 and MyoD play unique roles in the development of epaxial (originating in the dorso-medial half of the somite, e.g. back muscles) and hypaxial (originating in the ventro-lateral half of the somite, e.g. limb and body wall muscles) musculature. To further understand how Myf-5 and MyoD genes cooperate during skeletal muscle specification, we examined and compared the expression pattern of MyoD-lacZ (258/2.5lacZ and MD6.0-lacZ) transgenes in wild-type, Myf-5, and MyoD mutant embryos. We found that the delayed onset of muscle differentiation in the branchial arches, tongue, limbs, and diaphragm of MyoD-/- embryos was a consequence of a reduced ability of myogenic precursor cells to progress through their normal developmental program and not because of a defect in migration of muscle progenitor cells into these regions. We also found that myogenic precursor cells for back, intercostal, and abdominal wall musculature in Myf-54-/- embryos failed to undergo normal translocation or differentiation. By contrast, the myogenic precursors of intercostal and abdominal wall musculature in MyoD-/- embryos underwent normal translocation but failed to undergo timely differentiation. In conclusion, these observations strongly support the hypothesis that Myf-5 plays a unique role in the development of muscles arising after translocation of epithelial dermamyotome cells along the medial edge of the somite to the subjacent myotome (e.g., back or epaxial muscle) and that MyoD plays a unique role in the development of muscles arising from migratory precursor cells (e.g., limb and branchial arch muscles, tongue, and diaphragm). In addition, the expression pattern of MyoD-lacZ transgenes in the intercostal and abdominal wall muscles of Myf-5-/- and MyoD-/- embryos suggests that appropriate development of these muscles is dependent on both genes and, therefore, these muscles have a dual embryonic origin (epaxial and hypaxial).

Regulated demethylation of the myoD distal enhancer during skeletal myogenesis.

myoD is one of a family of four related basic helix-loop-helix transcription factors involved in the specification and differentiation of skeletal muscle. We previously identified a 258-bp distal enhancer that is sufficient for embryonic activation of myoD and is highly conserved between humans and mice. In this paper, we show using a modified bisulfite deamination/PCR amplification method that the distal myoD enhancer is completely unmethylated at all the CpG sites tested in myogenic cells and a subpopulation of somite cells. Conversely, the distal enhancer in nonmuscle cells and tissues is methylated to an average level of > 50% and we find no chromosomes in these tissues with a completely unmethylated enhancer. We present evidence that demethylation of the distal enhancer in somites of mouse embryos precedes myoD transcription, suggesting that demethylation of the distal enhancer is an active, regulated process that is essential for myoD activation. We also show by analysis of transgenic mice carrying a human distal enhancer/reporter construct in which the three enhancer CpG sites have been mutated that methylation of the distal enhancer is not required to prevent precocious or ectopic embryonic myoD expression. We propose that a subset of somite cells demethylate the distal enhancer in response to specific developmental signals, thus making the enhancer accessible and able to respond to subsequent signals to activate the myoD gene.

The distal human myoD enhancer sequences direct unique muscle-specific patterns of lacZ expression during mouse development.

Transgenic mice carrying the bacterial lacZ reporter gene under the control of the regulatory elements of the human myoD gene have been produced. The developmental expression of the myoD reporter transgene in somites, limb buds, visceral arches, and cephalocervical regions was studied in transgenic embryos by beta-gal staining. In somites, the spatiotemporal pattern of transgene expression was different from other muscle-specific regulatory and structural genes and revealed that myoD-expressing cells arise in distinct patterns in somites that are dependent on position along the anterior-posterior (AP) body axis (occipital and cervical vs thoracic and more posterior myotomes). Transgene expression did not follow a strict anterior to posterior sequence of activation and therefore was not strictly correlated with somite developmental age. Moreover, the pattern of transgene expression along the dorsal-ventral myotomal axis was dependent on somite position along the anterior-posterior axis. While myoD expression is first detected after the myotome is well-formed, transgene expression in the dorsal and ventral medial lips of the dermatome suggests a function for myoD in the expansion of the myotome. Whole-mount in situ hybridization confirmed that these unique patterns of transgene expression in somites, as well as expression in limb buds, visceral arches, and other myogenic centers, are concordant with the distribution of endogenous myoD transcripts. These results shed new light on the developmental differences between myotomes at different positions along the AP and DV axis and demonstrate a unique axial pattern of somitic myoD expression, suggesting a specific role of myoD in myotome lineage determination and differentiation.

MSX1 inhibits myoD expression in fibroblast x 10T1/2 cell hybrids.

Transfer of human chromosome 11, which contains the myoD locus, from primary fibroblasts into 10T1/2 cells results in activation of myoD. In contrast, hybrids that retain human chromosome 11 and additional human chromosomes fail to activate myoD. We show that human chromosome 4 inhibits myoD activation. myoD enhancer/promoter reporter constructs show that repression is at the transcriptional level. Chromosome fragment-containing hybrids localize the repressing activity to the region of 4p that contains the homeobox gene MSX1. MSX1 is expressed in primary human fibroblasts and in 10T1/2 cells containing human chromosome 4, while parental 10T1/2 cells do not express Msx1. Forced expression of Msx1 represses myoD enhancer activity. Msx1 protein binds to the myoD enhancer and likely represses myoD transcription directly. Antisense MSX1 relieves repression mediated by chromosome 4. We conclude that MSX1 inhibits transcription of myoD and that myoD is a target for homeobox gene regulation.

Embryonic activation of the myoD gene is regulated by a highly conserved distal control element.

MyoD belongs to a small family of basic helix-loop-helix transcription factors implicated in skeletal muscle lineage determination and differentiation. Previously, we identified a transcriptional enhancer that regulates the embryonic expression of the human myoD gene. This enhancer had been localized to a 4 kb fragment located 18 to 22 kb upstream of the myoD transcriptional start site. We now present a molecular characterization of this enhancer. Transgenic and transfection analyses localize the myoD enhancer to a core sequence of 258 bp. In transgenic mice, this enhancer directs expression of a lacZ reporter gene to skeletal muscle compartments in a spatiotemporal pattern indistinguishable from the normal myoD expression domain, and distinct from expression patterns reported for the other myogenic factors. In contrast to the myoD promoter, the myoD enhancer shows striking conservation between humans and mice both in its sequence and its distal position. Furthermore, a myoD enhancer/heterologous promoter construct exhibits muscle-specific expression in transgenic mice, demonstrating that the myoD promoter is dispensable for myoD activation. With the exception of E-boxes, the myoD enhancer has no apparent sequence similarity with regulatory regions of other characterized muscle-specific structural or regulatory genes. Mutation of these E-boxes, however, does not affect the pattern of lacZ transgene expression, suggesting that myoD activation in the embryo is E-box-independent. DNase I protection assays reveal multiple nuclear protein binding sites in the core enhancer, although none are strictly muscle-specific. Interestingly, extracts from myoblasts and 10T1/2 fibroblasts yield identical protection profiles, indicating a similar complement of enhancer-binding factors in muscle and this non-muscle cell type. However, a clear difference exists between myoblasts and 10T1/2 cells (and other non-muscle cell types) in the chromatin structure of the chromosomal myoD core enhancer, suggesting that the myoD enhancer is repressed by epigenetic mechanisms in 10T1/2 cells. These data indicate that myoD activation is regulated at multiple levels by mechanisms that are distinct from those controlling other characterized muscle-specific genes.

Identification of DNA-binding protein(s) in the developing heart.

An antiserum (anti-H2) directed at the second helix of the helix-loop-helix (HLH) protein MyoD1 reacts with a protein expressed during avian cardiac myocyte differentiation. Indirect immunohistochemical whole mount staining with anti-H2 detected a protein expressed in stage 11 hearts, but not in hearts of older embryos. At the cellular level, this staining is confined to the nucleus of cardiac cells suggesting that these proteins may have DNA-binding abilities. Several proteins were immunoprecipitated by anti-H2 from stage 11 heart tissue. Protein extracts from similarly staged hearts, when incubated with the muscle-specific enhancer sequence of muscle creatinine kinase (MCK), gave a stage-specific band shift in electromobility shift assays (EMSA), and these protein-DNA complexes were recognized and supershifted by anti-H2. Incubation with a MCK sequence containing a mutated E box did not produce a shift. The specific shift was present as early as stage 6, remained through stage 13, and disappeared by stage 17. These data suggest the presence of at least one protein that is transiently expressed in the differentiating cardiac myocyte, that is immunochemically reactive with an antiserum raised against the second helix of MyoD1, and that binds to a muscle-specific DNA enhancer sequence.

Regulatory elements that control the lineage-specific expression of myoD.

The molecular basis of skeletal muscle lineage determination was investigated by analyzing DNA control elements that regulate the myogenic determination gene myoD. A distal enhancer was identified that positively regulates expression of the human myoD gene. The myoD enhancer and promoter were active in myogenic and several nonmyogenic cell lines. In transgenic mouse embryos, however, the myoD enhancer and promoter together directed expression of a lacZ transgene specifically to the skeletal muscle lineage. These data suggest that during development myoD is regulated by mechanisms that restrict accessibility of myoD control elements to positive trans-acting factors.

Ganglia implantation as a means of supplying neurotrophic stimulation to the newt regeneration blastema: cell-cycle effects in innervated and denervated limbs.

Regulation of blastema cell proliferation during amphibian limb regeneration is poorly understood. One unexplained phenomenon is the relatively low level of active cell cycling in the adult newt blastema compared to that of larval axolotls. In the present study, we used ganglia implantation as a means of "superinnervating" normally innervated adult newt blastemas to test whether blastema cell subpopulations are responsive to nerve augmentation. The effectiveness of implanted ganglia to provide neurotrophic stimulation was demonstrated in denervated blastemas. Blastemas implanted with 2 dorsal root ganglia and simultaneously denervated 14 days after amputation exhibited control levels of cell cycle activity 6 days later, as measured by 3H-thymidine pulse labeling. Denervated blastemas that were sham-operated or implanted with pituitary glands exhibited cell-cycle declines similar to those of denervated blastemas without implanted ganglia. Thus, 2 implanted ganglia provide neurotrophic stimulation equivalent to that of the normal nerve supply. Dorsal root ganglia implanted into normally innervated blastemas, which should effectively double neurotrophic activity to the blastema, had no effect on cell-cycle activity, innervated blastemas implanted with ganglia for 6 days exhibited pulse labeling indices similar to those of normally innervated blastemas. These data indicate that neurotrophic stimulation is not normally limiting in innervated limbs, and that some other factor, whether extrinsic or intrinsic to blastema cells, accounts for the relatively low level of active cell cycling in the adult newt blastema.

An extracellular matrix molecule of newt and axolotl regenerating limb blastemas and embryonic limb buds: immunological relationship of MT1 antigen with tenascin.

Several well-characterized extracellular matrix (ECM) components have been localized to the amphibian limb regenerate, but the identification and characterization of novel ECM molecules have received little attention. Here we describe, using mAb MT1 and immunocytochemistry, an ECM molecule expressed during limb regeneration and limb development. In limb stumps, mAb MT1 reactivity was restricted to tendons, myotendinous junctions, granules in the basal layers of epidermis, periosteum (newts) and perichondrium (axolotls). In regenerating limbs, reactivity in the distal limb stump was first detected 5 days and 1 day after amputation of newt and axolotl limbs, respectively. In both species, mAb MT1 recognized what appeared to be an abundant blastema matrix antigen, localized in both thin and thick cords between and sometimes closely associated with blastema cells. Reactivity was generally uniform throughout the blastema except for a particularly thick layer that was present immediately beneath the wound epithelium. During redifferentiation stages, mAb MT1 reactivity persisted among blastema cells and redifferentiating cartilage but was lost proximally in areas of muscle and connective tissue differentiation. During the entire period of embryonic limb development, mAb MT1 reactivity was seen in the ECM of the mesenchyme and in a layer beneath the limb bud ectoderm, similar to its distribution during regeneration. Considerable mAb MT1 reactivity was also associated with the developing somites. The reactivity of mAb MT1 in blastema and limb bud was similar if not identical to that of a polyclonal Ab against tenascin (pAbTN), a large, extracellular matrix glycoprotein implicated in growth control, inductive interactions, and other developmental events. This pAbTN effectively competed against mAb MT1 binding on blastema sections. In immunoblots, both mAb MT1 and pAbTN recognized a very high molecular weight (approximately Mr 1000 x 10(3)) protein in blastema extracts of both newts and axolotls. mAb MT1 immunoprecipitated a protein of Mr 1000K size which reacted to both mAb MT1 and pAbTN in immunoblots. These data show that tenascin is in the matrix of the urodele blastema and limb bud, and suggest that mAb MT1 identifies urodele tenascin.

A developmentally regulated wound epithelial antigen of the newt limb regenerate is also present in a variety of secretory/transport cell types.

The role of the wound epithelium in amphibian limb regeneration is not understood. We showed previously that monoclonal antibody (mAb) WE3 stains the wound epithelium but not skin epidermis, suggesting that the WE3 antigen may be a marker for, or be important in, the function of the wound epithelium. In the present study, we conducted an extensive immunohistochemical survey of adult newt tissues to define the distribution of the WE3 antigen. The results show that the antigen is most commonly found in tissues specialized in macromolecular secretion and/or ion transport. Since the enzyme, carbonic anhydrase, serves as a useful marker for a variety of specialized transporting cell types, we examined whether this enzyme was present in WE3-reactive cells. Of the tissues examined, a striking degree of colocalization of carbonic anhydrase and the WE3 antigen was observed, further strengthening the view that the WE3 antigen is an important constituent of specialized transporting cells. A preliminary biochemical characterization suggests that the antigen is probably a glycoprotein, which elutes during gel filtration as a species of over 660 kDa. Possible implications for the function of the wound epithelium are discussed.

Cell cycle controls and the role of nerves and the regenerate epithelium in urodele forelimb regeneration: possible modifications of basic concepts.

Data from pulse and continuous labeling with [3H]thymidine and from studies with monoclonal antibody WE3 have led to the modification of existing models and established concepts pertinent to understanding limb regeneration. Not all cells of the adult newt blastema are randomly distributed and actively progressing through the cell cycle. Instead, many cells are in a position that we have designated transient quiescence (TQ) and are not actively cycling. We postulate that cells regularly leave the TQ population and enter the actively cycling population and vice versa. The size of the TQ population may be at least partly determined by the quantity of limb innervation. Larval Ambystoma may have only a small or nonexisting TQ, thus accounting for their rapid rate of regeneration. Examination of reactivity of monoclonal antibody WE3 suggests that the early wound epithelium, which is derived from skin epidermis, is later replaced by cells from skin glands concomitant with blastema formation. WE3 provides a useful tool to further investigate the regenerate epithelium.