Roles of Meltrin beta /ADAM19 in the processing of neuregulin.

Meltrin beta/ADAM19 is a member of ADAMs (a disintegrin and metalloproteases), which are a family of membrane-anchored glycoproteins that play important roles in fertilization, myoblast fusion, neurogenesis, and proteolytic processing of several membrane-anchored proteins. The expression pattern of meltrin beta during mouse development coincided well with that of neuregulin-1 (NRG), a member of the epidermal growth factor family. Then we examined whether meltrin beta participates in the proteolytic processing of membrane-anchored NRGs. When NRG-beta1 was expressed in mouse L929 cells, its extracellular domain was constitutively processed and released into the culture medium. This basal processing activity was remarkably potentiated by overexpression of wild-type meltrin beta, which lead to the significant decrease in the cell surface exposure of extracellular domains of NRG-beta1. Furthermore, expression of protease-deficient mutants of meltrin beta exerted dominant negative effects on the basal processing of NRG-beta1. These results indicate that meltrin beta participates in the processing of NRG-beta1. Since meltrin beta affected the processing of NRG-beta4 but not that of NRG-alpha2, meltrin beta was considered to have a preference for beta-type NRGs as substrate. Furthermore, the effects of the secretory pathway inhibitors suggested that meltrin beta participates in the intracellular processing of NRGs rather than the cleavage on the cell surface.

Analysis for transcript expression of meltrin alpha in normal, regenerating, and denervated rat muscle.

Meltrin alpha (a disintegrin and metalloprotease (ADAM) 12) is a recently discovered molecule of the metalloprotease-disintegrin family which has been shown to participate in myotube formation in vitro and in myogenesis in vivo. In this study we investigated meltrin alpha in regenerating rat muscle, which is a condition where satellite cells (SC) contribute to myofiber growth by fusing with one another and with myotubes or muscle fibers. We studied meltrin alpha mRNA expression by RT-PCR and in situ-hybridization in normal adult muscle, in soleus muscle regenerating for 2, 5, or 10 days, and in muscle which had been denervated 1 week, 4 weeks, or 6 months previously. SC do not fuse after denervation. They detach from the principal muscle fiber. Immunohistochemistry using an antibody against M-cadherin was performed in parallel in order to identify SC. Messenger RNA as revealed by RT-PCR was absent in normal adult muscle, but present in regenerating and also in denervated muscle. Meltrin alpha transcript detected by in situ-hybridization was present in regenerating muscle only, not in normal or denervated muscle. It was localized to SC. Taken together, meltrin alpha is absent in normal muscle, and localized to SC in fusing conditions. After denervation, the transcript is upregulated. However, it is so lowly abundant that it fails to be detected by in situ-hybridization. This expression profile suggests a role for meltrin alpha in the fusion of SC with myotubes or muscle fibers, but not in SC adhesion to the adjacent myofiber in normal adult muscle.

Meltrin beta (ADAM19) gene: cloning, mapping, and analysis of the regulatory region.

Meltrin beta (ADAM19) is a member of the metalloprotease-disintegrin family. We report here chromosomal mapping of the mouse and rat meltrin beta genes and cloning and analysis of the mouse upstream regulatory regions. The meltrin beta transcript shows a spatially and temporally restricted expression pattern during morphogenesis, indicating that the actions of this membrane-bound protease are regulated, at least in part, at the transcriptional level. Analysis of the promoter revealed positive and negative regulatory regions upstream of the gene. The former includes a GC-box that appears to be a critical cis-element for activation of the promoter in muscle cells.

Membrane-anchored metalloprotease MDC9 has an alpha-secretase activity responsible for processing the amyloid precursor protein.

MDC9, also known as meltrin gamma, is a membrane-anchored metalloprotease. MDC9 contains several distinct protein domains: a signal sequence followed by a prodomain and a domain showing sequence similarity to snake venom metalloproteases, a disintegrin-like domain, a cysteine-rich region, an epidermal-growth-factor-like repeat, a transmembrane domain and a cytoplasmic domain. Here we demonstrate that MDC9 expressed in COS cells is cleaved between the prodomain and the metalloprotease domain. Further, when MDC9 was co-expressed in COS cells with amyloid precursor protein (APP695) and treated with phorbol ester, APP695 was digested exclusively at the alpha-secretory site in MDC9-expressing cells. When an artificial alpha-secretory site mutant was also co-expressed with MDC9 and treated with phorbol ester, APP secreted by alpha-secretase was not increased in conditional medium. Inhibition of MDC9 by a hydroxamate-based metalloprotease inhibitor, SI-27, enhanced beta-secretase cleavage. These results suggest that MDC9 has an alpha-secretase-like activity and is activated by phorbol ester.

Spatially- and temporally-restricted expression of meltrin alpha (ADAM12) and beta (ADAM19) in mouse embryo.

The cloning of the full-length cDNA encoding meltrin beta (ADAM19), one of the metalloprotease-disintegrins expressed in mouse myogenic cells, revealed that the meltrin beta gene encodes a membrane protein closely related to meltrin alpha (ADAM12) which participates in myotube formation in vitro. To delineate the functions of meltrin alpha and beta, we examined the expression patterns of their transcripts during embryogenesis. The meltrin alpha gene is activated in condensed mesenchymal cells that give rise to skeletal muscle, bones and visceral organs. Meltrin beta mRNA, in contrast, is markedly expressed in craniofacial and dorsal root ganglia and ventral horns of the spinal cord, where peripheral neuronal cell lineages differentiate. Heart, skeletal muscle, intestine and lung also express meltrin beta mRNA transiently. Although the meltrin alpha and beta transcripts exhibit distinct expression patterns during embryogenesis, both genes are mainly activated in mesenchymal cells that are derived from both mesoderm and ectoderm.

Bone morphogenetic protein-2 inhibits terminal differentiation of myogenic cells by suppressing the transcriptional activity of MyoD and myogenin.

Bone morphogenetic protein (BMP) is a family of cytokines that induce ectopic bone formation when implanted into muscular tissues. We reported that BMP-2 inhibits the terminal differentiation of C2C12 myoblasts and converts them into osteoblast lineage cells (Katagiri, T., Yamaguchi, A., Komaki, M., Abe, E., Takahashi, N., Ikeda, T., Rosen, V., Wozney, J. M., Fujisawa-Sehara, A., and Suda, T. (1994) J. Cell Biol. 127, 1755-1766). In the present study, we examined the molecular mechanism of the inhibitory effect of BMP-2 on terminal differentiation of myogenic cells. When either MyoD or myogenin cDNA was introduced into C3H10T1/2 (10T1/2) cells with a muscle-specific CAT reporter containing four copies of the right E-box of muscle creatine kinase (MCK) enhancer, the CAT activity was dose-dependently suppressed by BMP-2. Furthermore, BMP-2 inhibited the terminal differentiation of these subclonal 10T1/2 cells that stably expressed MyoD or myogenin into mature myotubes that expressed myosin heavy chain and troponin T. The differentiation of a subclone of the MyoD-transfected NIH3T3 cells into mature muscle cells was also inhibited by BMP-2. BMP-2 induced alkaline phosphatase activity in 10T1/2-derived, but not in NIH3T3-derived MyoD-transfected cells. These cells constitutively expressed exogenous MyoD and myogenin, which were localized exclusively in the nuclei irrespective of the presence and the absence of BMP-2. However, these cells failed to express the mRNAs of endogenous myogenic factors and MCK when cultured with BMP-2. In the electrophoresis mobility shift assay using nuclear extracts of the myogenic cells, MyoD and myogenin bound to the right E-box in the enhancer region of the MCK gene even in the presence of BMP-2. These results suggest that BMP-2 inhibits the terminal differentiation of myogenic cells by suppressing the transcriptional activity of the myogenic factors.

Lysophosphatidic acid and bFGF control different modes in proliferating myoblasts.

Myogenic cells provide excellent in vitro models for studying the cell growth and differentiation. In this study we report that lysophosphatidic acid (LPA), a bioactive phospholipid contained in serum, stimulates the growth and inhibits the differentiation of mouse C2C12 myoblast cells, in a distinct manner from basic fibroblast growth factor (bFGF) whose mitotic and anti-differentiation actions have been well investigated. These actions of LPA were both blocked by pertussis toxin, suggesting the involvement of Gi class of G proteins, whereas bFGF acts through receptor tyrosine kinases. Detailed analysis revealed that LPA and bFGF act differently in regulating the myogenic basic helix-loop-helix (bHLH) proteins, the key players in myogenic differentiation process. LPA stimulates the proliferation of undifferentiated myoblasts allowing the continued expression of MyoD, but in contrast, bFGF does so with the MyoD expression suppressed at the mRNA level. Both compounds maintain the myf-5 expression, and suppress the myogenin expression. In addition, while LPA did not inhibit cell-cell contact-induced differentiation, bFGF strongly inhibited this process. Furthermore, LPA and bFGF act cooperatively in their mitogenic and anti-differentiation abilities. These findings indicate that LPA and bFGF differently stimulate intracellular signaling pathways, resulting in proliferating myoblasts each bearing a distinct expression pattern of myogenic bHLH proteins and distinct differentiation potentials in response to cell-cell contact, and illustrate the biological significance of Gi-mediated and tyrosine kinase-mediated signals.

A metalloprotease-disintegrin participating in myoblast fusion.

Skeletal muscle development involves the formation of multi-nucleated myotubes. This is thought to proceed by the induction of differentiation (acquisition of fusion competence) of myoblast cells, their aggregation, and union of their plasma membranes. Various membrane proteins including N- and M-cadherins, N- and V-CAMs and integrins participate in myotube formation, but the molecular mechanisms of muscle cell fusion are poorly understood. Here we report the identification of three new, myoblast-expressed gene products, meltrin-alpha, beta and gamma, with homology to both viper haemorrhagic factors and fertilin (PH-30), a membrane protein involved in egg-sperm fusion. Meltrin-alpha, a member of the metalloproteinase/disintegrin protein family, appears to be required for myotube formation. Involvement of a fertilin-related protein in myogenesis suggests that there are common mechanisms in gamete and myoblast fusion.

Bone morphogenetic protein-2 converts the differentiation pathway of C2C12 myoblasts into the osteoblast lineage.

The implantation of bone morphogenetic protein (BMP) into muscular tissues induces ectopic bone formation at the site of implantation. To investigate the mechanism underlying this process, we examined whether recombinant bone morphogenetic protein-2 (BMP-2) converts the differentiation pathway of the clonal myoblastic cell line, C2C12, into that of osteoblast lineage. Incubating the cells with 300 ng/ml of BMP-2 for 6 d almost completely inhibited the formation of the multinucleated myotubes expressing troponin T and myosin heavy chain, and induced the appearance of numerous alkaline phosphatase (ALP)-positive cells. BMP-2 dose dependently induced ALP activity, parathyroid hormone (PTH)-dependent 3',5'-cAMP production, and osteocalcin production at concentrations above 100 ng/ml. The concentration of BMP-2 required to induce these osteoblastic phenotypes was the same as that required to almost completely inhibit myotube formation. Incubating primary muscle cells with 300 ng/ml of BMP-2 for 6 d also inhibited myotube formation, whereas induced ALP activity and osteocalcin production. Incubation with 300 ng/ml of BMP-2 suppressed the expression of mRNA for muscle creatine kinase within 6 h, whereas it induced mRNA expression for ALP, PTH/PTH-related protein (PTHrP) receptors, and osteocalcin within 24-48 h. BMP-2 completely inhibited the expression of myogenin mRNA by day 3. By day 3, BMP-2 also inhibited the expression of MyoD mRNA, but it was transiently stimulated 12 h after exposure to BMP-2. Expression of Id-1 mRNA was greatly stimulated by BMP-2. When C2C12 cells pretreated with BMP-2 for 6 d were transferred to a colony assay system in the absence of BMP-2, more than 84% of the colonies generated became troponin T-positive and ALP activity disappeared. TGF-beta 1 also inhibited myotube formation in C2C12 cells, and suppressed the expression of myogenin and MyoD mRNAs without inducing that of Id-1 mRNA. However, no osteoblastic phenotype was induced by TGF-beta 1 in C2C12 cells. TGF-beta 1 potentiated the inhibitory effect of BMP-2 on myotube formation, whereas TGF-beta 1 reduced ALP activity and osteocalcin production induced by BMP-2 in C2C12 cells. These results indicate that BMP-2 specifically converts the differentiation pathway of C2C12 myoblasts into that of osteoblast lineage cells, but that the conversion is not heritable.

The delta-crystallin enhancer-binding protein delta EF1 is a repressor of E2-box-mediated gene activation.

The repressor delta EF1 was discovered by its action on the DC5 fragment of the lens-specific delta 1-crystallin enhancer. C-proximal zinc fingers of delta EF1 were found responsible for binding to the DC5 fragment and had specificity to CACCT as revealed by selection of high-affinity binding sequences from a random oligonucleotide pool. CACCT is present not only in DC5 but also in the E2 box (CACCTG) elements which are the binding sites of various basic helix-loop-helix activators and also the target of an unidentified repressor, raising the possibility that delta EF1 accounts for the E2 box repressor activity. delta EF1 competed with E47 for binding to an E2 box sequence in vitro. In lymphoid cells, endogenous delta EF1 activity as a repressor was detectable, and exogenous delta EF1 repressed immunoglobulin kappa enhancer by binding to the kappa E2 site. Moreover, delta EF1 repressed MyoD-dependent activation of the muscle creatine kinase enhancer and MyoD-induced myogenesis of 10T1/2 cells. Thus, delta EF1 counteracts basic helix-loop-helix activators through binding site competition and fulfills the conditions of the E2 box repressor. In embryonic tissues, the most prominent site of delta EF1 expression is the myotome. Myotomal expression as well as the above results argues for a significant contribution of delta EF1 in regulation of embryonic myogenesis through the modulation of the actions of MyoD family proteins.

MyoD and myogenin act on the chicken myosin light-chain 1 gene as distinct transcriptional factors.

Expression of MyoD, myogenin, MRF4, and Myf-5 converts nonmuscle cells to muscle cells. In an attempt to analyze the roles of these factors, we have investigated their effects on transcription driven by the promoter of the chicken myosin alkaline light-chain (MLC1) gene. The activation by CMD1 or c-myogenin (chicken MyoD or myogenin, respectively) was dependent on the existence of a muscle-specific regulatory region located from positions -2096 to -1743. Its distal half, containing a pair of E boxes (CANNTG), had been previously characterized as an enhancer responsive to CMD1 but not to c-myogenin. In this study, we report the identification of another enhancer in the muscle-specific regulatory region which is preferentially responsive to c-myogenin. Deletion and mutation analyses indicated that this enhancer requires a single E box and its flanking sequences. Furthermore, analysis of chimeric proteins of CMD1 and c-myogenin indicated that regions outside the basic helix-loop-helix domain of c-myogenin are involved in the specificity of the enhancer. These results show that CMD1 and c-myogenin act on the MLC1 gene by recognizing different upstream DNA sequences and that direct or indirect interactions between the regions outside the basic helix-loop-helix domain and flanking sequences of E boxes are involved in the target sequence specificity.

Expression of myogenic factors in denervated chicken breast muscle: isolation of the chicken Myf5 gene.

In this study, we have isolated and characterized the chicken Myf5 gene, and cDNA clones encoding chicken MyoD1 and myogenin. The chicken Myf5 and MRF4 genes are tandemly located on a single genomic DNA fragment, and the chicken Myf5 gene is organized into at least three exons. Using genomic and cDNA probes, we further analyzed the mRNA levels of four myogenic factors during chicken breast muscle development. This analysis revealed that myogenin expression is restricted to in ovo stages in breast muscle, and is not detectable in neonatal and adult stages. On the other hand, Myf5 expression is detectable until day 7 post-hatching, and is not found in adult muscle, whereas high levels of MyoD1 and MRF4 are detectable at all stages. To further understand the roles of innervation on muscle maturation, we analyzed the expression of the four myogenic factors in denervated adult breast muscle. We found that MyoD1, myogenin, and MRF4 are induced at high levels in denervated muscle, whereas no change occurs in the level of Myf5. These studies suggest that innervation controls the relative abundance and type of myogenic factors that are expressed in adult muscle, and that when nerve control is removed, the muscle reverts to a neonatal phenotype, with the enhanced expression of three myogenic factors (MyoD1, myogenin, and MRF4).

Upstream region of the myogenin gene confers transcriptional activation in muscle cell lineages during mouse embryogenesis.

Myogenin, one of the MyoD-related factors containing basic-helix-loop-helix motifs, is transcriptionally activated in skeletal muscle lineages in somites and limb buds during embryogenesis. In an attempt to understand regulatory mechanisms which govern transcriptional activation of the myogenin gene, transgenic mice bearing the lacZ gene driven by the upstream region of the myogenin gene were generated. Stereoscopic visualization of LacZ-positive cells of these transgenic mouse embryos revealed that the upstream region of the myogenin gene conferred its transcriptional activation in cells of the skeletal muscle lineages in somites, limb buds, and visceral arches. Moreover, transient LacZ expression in newly formed somites in addition to its strong activation in myotomal regions of mature somites with a rostro-caudal gradient raised the possibility that myogenin is transcriptionally activated in immature somites before myotome formation.

Differential trans-activation of muscle-specific regulatory elements including the mysosin light chain box by chicken MyoD, myogenin, and MRF4.

We have isolated cDNAs encoding a chicken homologue of MRF4 (cMRF4) in addition to chicken MyoD (CMD1) and myogenin (c-myogenin) described previously. In an attempt to understand the roles that cMRF4, CMD1, and c-myogenin play in chicken myogenesis, the effects of these factors on muscle-specific cis-elements identified in regulatory regions of myosin alkali light chain (MLC) genes were examined. The promoter analysis of some of MLC genes has revealed two sorts of muscle-specific positive regulatory elements to date, an enhancer located upstream of the adult type LC1 gene and a cis-element, termed an MLC box, conserved among promoters of various MLC genes. The LC1 enhancer was exclusively trans-activated by CMD1. Although c-myogenin also activated transcription driven by the LC1 promoter, it was suggested that c-myogenin requires a cis-element(s) other than the CMD1-responsive enhancer. Chicken MRF4 could not trans-activate any of the constructs containing the LC1 promoter. In contrast, the promoter of the embryonic L23 gene was trans-activated by all of the three factors. From deletion and mutation analysis, the MLC box was shown to be involved in their positive regulation. These results extend previous observations that individual myogenic regulatory factors exhibit different capabilities in transcriptional activation of muscle-specific genes by acting distinctively upon their regulatory elements.

Localization of mRNAs coding for CMD1, myogenin and the alpha-subunit of the acetylcholine receptor during skeletal muscle development in the chicken.

Myogenin and CMD1, the chicken homologue of MyoD, transactivate the promoter of the alpha-subunit of the acetylcholine receptor (AChR) in chicken fibroblasts. The expression of these three genes was followed by in situ hybridization. In two-day-old embryos the CMD1 gene is expressed shortly before the AChR alpha-subunit and the myogenin genes. At day 19 extrajunctional AChR mRNA clusters have disappeared and myogenin mRNAs are no longer detected in PLD muscle. Moreover, both myogenin and CMD1 mRNA levels increase after muscle denervation in chicks. These data are compatible with a role for myogenic factors in the induction and maintenance of extra-junctional expression of the AChR genes during early muscle development. Using digoxygenin labelled RNA probes, we also show that the mRNAs for the AChR alpha-subunit display a punctated, probably perinuclear distribution, whereas mRNAs for myogenic genes accumulate in the sarcoplasm around subsets of nuclei in the muscle fiber.

Positive and negative gene regulation in muscle.

By the analysis of cis and trans-acting element involved in transcriptional regulation of chicken myosin alkali light chain genes, we have identified the MLC box, muscle specific enhancer element and negative regulatory element. The MLC box is an essential element for the expression of MLC genes located at approximately 100 bp upstream from mRNA start sites. The core sequence of MLC box is similar to the consensus of actin gene CArG box and SRE of c-fos oncogene. In vitro DNA-protein binding assay has revealed that the MLC box, CArG box and SRE might bind to a common or a similar protein complex. CMD1, cMyogenin, and cMRF4 transactivate the promorter with an intact MLC box, but not the promoter lacking MLC box, indicating that the MLC box itself is transactivated by the myogenic regulatory factors. This transactivation must have been due to the indirect effect of the myogenic regulatory factors, because chicken myogenic factors do not bind to the MLC box. A cis element identified at about 150 bp upstream from the cap site of cardiac MLC gene suppresses the cardiac MLC gene expression in skeletal muscle cells but not in cardiac muscle cells. The protein(s) bound to NRE might be identical with one of proteins bound to SRE. NRE may block the function of MLC box and resultantly inhibits the expression of cardiac MLC1 gene in skeletal muscle cells. Skeletal muscle enhancer at -2 kb of skeletal MLC1f gene is composed of two subelements P and D, cooperative action between them is required for sufficient enhancer activity. CMD1 and myogenin bind to the enhancer sequences of skeletal MLC1 gene and MCK gene and transactivate these genes preferentially in skeletal muscle cells. In addition to the CMD1 responsible enhancer, another cis-element is required for transactivation of the MLC1f gene by cMyogenin. An E-box adjacent to MLC box may co-work with the enhancer to increase the expression of MLC1f gene. Muscle specific and developmentally regulated expression of MLC gene family is regulated by the combination of these cis and trans-acting elements.

A stable cellular marker for the analysis of mouse chimeras: the bacterial chloramphenicol acetyltransferase gene driven by the human elongation factor 1 alpha promoter.

We have developed a method of marking of mouse cells by means of transfection of a foreign gene. The transgene chosen here was the plasmid pEF321CAT which contains the bacterial chloramphenicol acetyl transferase (CAT) gene linked to the promoter region of the human polypeptide chain elongation factor 1 alpha (hEF1 alpha) gene. Evaluation of the plasmid pEF321CAT as a cellular marker for mouse cells involved intensive examination of a transgenic mouse carrying pEF321CAT. The CAT gene was expressed in all tissues examined, demonstrating that the hEF1 alpha promoter was active in a wide range of mouse cells. The plasmid itself did not exert any harmful effect on the normal development of mice, and the CAT activity was immunohistologically detectable on sectioned tissues by the use of anti-CAT serum. When the plasmid was transferred into embryonal carcinoma (EC) cells and embryonic stem (ES) cells, the CAT gene was also found to be expressed constantly irrespective of their differentiation. These results demonstrated that the plasmid pEF321CAT can be used as a reliable and feasible cellular marker that would distinguish unequivocally the cells of each of genotype in chimeric tissues.

Myogenin contains two domains conserved among myogenic factors.

Previously, three structurally related proteins, MyoD, myogenin, and Myf5, have been identified, and each of them was found to convert C3H10T1/2 fibroblasts to myoblasts when their respective cDNAs were expressed under the control of a viral promoter. Here, we describe the cloning and DNA sequencing of myogenin cDNAs from chicken and mouse. They encode polypeptides highly homologous to each other, but the polypeptide sequences we have obtained are not homologous in the carboxyl-terminal 70 amino acids with those previously reported for mouse, rat, and human because of a single base deletion in the previously reported cDNAs. Determination of genomic sequence coding for myogenin revealed that the mouse myogenin cDNA presented here corresponds to a correct transcript of the gene, and the nucleotide sequence of the previously reported cDNA is incorrect. A comparison of chicken and mouse myogenins with other myogenic regulatory factors, MyoD and Myf-5, identified a domain with an interesting feature located in the carboxyl terminus of these proteins in addition to the myc homology domain previously reported.

Regulation of the chicken embryonic myosin light-chain (L23) gene: existence of a common regulatory element shared by myosin alkali light-chain genes.

The transcriptional regulation of the chicken myosin alkali light-chain (MLC) L23 gene was analyzed. Two different types of cis-regulatory regions were identified: one was a silencerlike region located between 3.7 and 2.7 kilobases upstream of the mRNA initiation site, and the other was essential for the expression of L23 in skeletal muscle cells and was located between 106 and 91 base pairs upstream of the cap site. This 16-base-pair cis-acting element was designated as the MLC box since it is well conserved in various muscle-specific MLC promoter regions. The activity of the MLC box showed tissue specificity. To analyze the relationship between the nucleotide sequence and the activity of the MLC box precisely, mutation analysis was performed. The 16-base-pair sequence was indispensable for the active transcription of L23 gene, and the MLC box could function in either orientation. The inverted sequence of the MLC box was similar to the sequence of the alpha-actin CArG box. By using a gel mobility retardation assay, the nuclear protein(s) that binds to both MLC box and CArG box was detected with nuclear extract prepared from chicken embryonic breast muscle. These observations imply that a common factor regulates the coordinate expression of these contractile proteins in muscle differentiation.