DEVELOPMENTAL DYNAMICS 209:233–241 (1997) MRF4 Can Substitute for Myogenin During Early Stages of Myogenesis ZHIMIN ZHU1,2 AND JEFFREY BOONE MILLER1,2* 1Neuromuscular Laboratory, Massachusetts General Hospital, Charlestown, Massachusetts 2Program in Neuroscience, Harvard Medical School, Boston, Massachusetts ABSTRACT MRF4, myogenin, MyoD, and Myf-5 are the four members of the basic helix-loophelix family of muscle-specific regulatory factors (MRFs). We examined whether MRF4 could substitute for myogenin in vivo by determining if the myofiber- and MRF4-deficient phenotype of myogenin (2/2) mice could be rescued by a myogenin promoter-MRF4 transgene. When the transgene was expressed at a physiological level in myogenin-deficient fetuses, we found that expression of the endogenous MRF4 gene was restored to normal levels, whereas MyoD levels were unchanged. Thus, MRF4 can participate in a positive autoregulatory loop and can substitute for myogenin to activate its own promoter. Myogenindeficient fetuses that expressed the transgene also had more myosin, more and larger myofibers, and a more normal ribcage morphology than myogenin-deficient littermates without the transgene. The transgene failed, however, to restore normal numbers of myofibers or viability to myogenin-deficient mice, because the D1.6 kb myogenin promoter fragment was not expressed in most late-forming myofibers. These results demonstrate that MRF4 is able to substitute for myogenin to activate MRF4 expression and promote myofiber formation during the early stages of myogenesis. Dev. Dyn. 209:233–241, 1997. r 1997 Wiley-Liss, Inc. Key words: MRF4; myogenin; myogenesis; regulatory loops; transgenic mouse INTRODUCTION MRF4, myogenin, MyoD, and Myf-5 constitute the basic helix-loop-helix family of muscle regulatory factors (MRFs). During development, each of the MRFs has a unique pattern of expression (Hinterberger et al., 1991; Buckingham, 1992; Smith et al., 1994). MRF4 is the last of the MRFs to be expressed in limb and trunk muscles, appearing about one day later than myogenin and about two days later than Myf-5 and MyoD. From studies of transgenic and knock-out mice, it appears that Myf-5 and MyoD are required for myoblast commitment; myogenin is required for formation of most myotubes; and MRF4 plays a role in myofiber maturation and formation of particular muscles (Braun et al., 1992; Rudnicki et al., 1992, 1993; Hasty et al., 1993; r 1997 WILEY-LISS, INC. Nabeshima et al., 1993; Braun and Arnold, 1995; Patapoutian et al., 1995; Venuti et al., 1995; Zhang et al., 1995; Block et al., 1996). Studies of MRF function in cultured cells and knockout mice suggest that individual MRFs may have unique roles in myogenesis (e.g., Block and Miller, 1992; Rudnicki et al., 1993; Rawls et al., 1995; reviewed Rudnicki and Jaenisch, 1995). Other studies, however, suggest that MRFs may be interchangeable; for instance, each MRF is capable of converting many types of non-myogenic cells into myogenic cells in vitro (Weintraub, 1993) and myogenin can substitute for Myf-5 in vivo (Wang et al., 1996). However, it is not known if other pairs of MRFs are similarly interchangeable in vivo. Previous studies have also shown that individual MRFs regulate each other’s expression, though the particular cross-regulatory pathways that operate in vivo in many cases remain to be determined (Weintraub, 1993). In this study we examine how MRF4 and myogenin participate in cross-regulatory pathways and whether MRF4 can substitute for myogenin during myogenesis in vivo. Based on molecular phylogeny, myogenin is most closely related to MRF4, whereas Myf-5 is most closely related to MyoD, leading to the suggestion that the more closely related pairs of MRFs would be more likely to be interchangeable (Atchley et al., 1994). Consistent with this idea, studies of mice that lack two MRFs suggest that Myf-5 and MyoD can likely substitute for each other, but not for myogenin, though the cellular basis of this complementation is not clear (Rudnicki et al., 1993; Rawls et al., 1995). On the other hand, myogenin, when under control of the Myf-5 promoter, is able to substitute for Myf-5 in vivo (Wang et al., 1996). The cross-regulation and interchangeability of myogenin and MRF4 had not been tested in vivo, though studies of cultured cells suggested that MRF4 can activate the myogenin promoter and that these MRFs may have distinct functional capabilities (Yutzey et al., 1990; Chakraborty et al., 1991; Block and Miller, 1992; Black et al., 1995; Naidu et al., 1995). To find out if MRF cross-regulation and functional interchangeability extends to myogenin and MRF4 in *Correspondence to: Dr. Jeffrey B. Miller, Neuromuscular Laboratory, Massachusetts General Hospital, 149 13th Street, Charlestown, MA 02129. E-mail: Miller@helix.mgh.harvard.edu Received 15 November 1996; Accepted 26 February 1997 234 ZHU AND MILLER vivo, we determined if the abnormal phenotype of myogenin-deficient mice could be rescued by a myogenin promoter-MRF4 transgene. In earlier work (Block et al., 1996), we generated and characterized myo1565MRF4 transgenic mice in which an ,1.6 kb fragment of the myogenin promoter (Cheng et al., 1992) was used to control expression of MRF4 cDNA. This myogenin promoter fragment does not require myogenin for expression (Cheng et al., 1995) and, up until about embryonic day 15 (E15), it appears to be expressed in myogenic cells in the same pattern as the endogenous myogenin gene (Cheng et al., 1992; Block et al., 1996). We have now examined how the MRF4-deficient and myofiber-defective phenotype of myogenin-deficient mice (Hasty et al., 1993) is altered by expression of the myo1565-MRF4 transgene. The results show that MRF4 can substitute for myogenin to activate its own promoter and promote myofiber formation during early stages of myogenesis in vivo. RESULTS To analyze MRF cross-regulation and to determine if MRF4 can substitute for myogenin, we used crossbreeding to introduce a myogenin promoter-MRF4 transgene (myo1565-MRF4, locus 43; Block et al., 1996) into myogenin (2/2) mice. After analyzing 114 threeweek-old progeny resulting from these crosses, we found no surviving myogenin (2/2) mice, either with or without the transgene. All pups that were both myogenin (2/2) and transgene-positive died within a few minutes of delivery, as do myogenin-deficient neonates without the transgene (Hasty et al., 1993; Nabeshima et al., 1993). If the myo1565-MRF4 transgene had restored viability to myogenin (2/2) mice, we would have expected ,16 of the 114 survivors to be myogenin (2/2) and transgene-positive. Though they died at birth, myogenin (2/2) pups that carried the transgene appeared to be bigger and better formed, and to have larger mechanically-evoked limb movements than myogenin (2/2) littermates that did not carry the transgene (not shown). The failure of the myo1565-MRF4 transgene to prevent death was not due to a general lack of transgene expression, because MRF4 mRNA and protein were expressed in myogenin-deficient mice (Figs. 1 and 2). The myo1565-MRF4 mRNA, which is larger than the endogenous MRF4 mRNA, was expressed at least from embryonic day 17 (E17) to birth in myogenin (2/2) mice (Fig. 1 and not shown). As noted (Hasty et al., 1993), the amount of endogenous MRF4 mRNA is much lower in myogenin (2/2) than myogenin-positive fetuses (Fig. 1). However, when the myo1565-MRF4 transgene (locus 43) was expressed in myogenin (2/2) fetuses, the amount of endogenous MRF4 mRNA was increased to a level similar to that seen in wild-type littermates (Fig. 1). The combined amount of endogenous and locus 43 transgene MRF4 mRNAs in E18 myogenin-deficient, transgene-positive mice was at least as great as the amount of endogenous MRF4 and myogenin mRNAs in Fig. 1. The myo1565-MRF4 transgene was expressed and increased endogenous MRF4 expression in myogenin-deficient fetuses. Total RNA was prepared from limb muscles of E18–19 littermates with the indicated combinations of myo1565-MRF4 (locus 43) and myogenin expression. Samples (10 µg) were separated by electrophoresis and hybridized with probes for MRF4 or myogenin. Probes were of equal specific activity and autoradiograms were exposed for equal times (1 day). All samples were analyzed at the same time; some lanes were juxtaposed for presentation. The transgene was expressed at about equal levels in the presence or absence of myogenin and produced a MRF4 mRNA that was larger than that produced from the endogenous gene. MRF4 mRNA was barely detectable in myogenin (2/2) fetuses, but was abundant when myo1565MRF4 was expressed. myogenin-positive littermates that have normal muscle phenotypes (Fig. 1). In contrast, the transgene did not affect Myf-5 or MyoD expression, because, with or without the transgene, Myf-5 mRNA was not detectable and MyoD mRNA was expressed at the same level in myogenin-deficient and -positive fetuses (not shown). Consistent with their low level of MRF4 mRNA, myogenin (2/2) fetuses without the transgene showed only weak MRF4 immunostaining in a few muscles and no staining in most muscles (Fig. 2A, D). In contrast, MRF4 protein was easily detectable in myogenin (2/2) mice that carried the transgene (Fig. 2B, E). In these fetuses, MRF4 was expressed in every myofiber that formed, and strong staining for MRF4 was found in every muscle examined including the diaphragm, intercostal, thoracic wall, pelvic, brachial, forelimb, hindlimb, head, and neck muscles. In myogenin-positive controls without the transgene, MRF4 staining varied in intensity among muscles, with moderate staining in jaw muscles (Fig. 2C), weak staining in intercostals, and no staining in the diaphragm (Fig. 2F). From these initial protein and mRNA analyses, therefore, it appeared that MRF4 was expressed from the transgene at a physiological level in myogenin (2/2) fetuses, and that this MRF4 expression activated expression of the endogenous MRF4 gene, without affecting expression of the endogenous Myf-5 and MyoD genes. We also examined whether the transgene affected expression of pre- and post-synaptic components of the neuromuscular junction. By RT-PCR, mRNA for the alpha subunit of AChR, which is the AChR subunit that binds alpha-Bungarotoxin, was detectable in myogenindeficient and wild-type mice, with or without the transgene (not shown). Furthermore, by staining with alphaBungarotoxin, we found that myofibers in both myogenin-deficient and myogenin-positive muscles contained single AChR clusters typical of normal neuromuscular junctions, and this normal clustering was not affected by the transgene (Fig. 2G–I). Also, we examined myogenin-deficient and normal muscles with anti- MRF4 IN MYOGENIN (2/2) MICE 235 Fig. 2. MRF4 and acetylcholine receptor expression in fetuses with different combinations of myogenin and myo1565-MRF4 expression. Sections were made of E18–19 fetuses that were myogenin-deficient (A, D, G); myogenin-deficient and myo1565-MRF4-positive (B, E, H), and myogenin-positive (C, F, I). Sections were stained either for MRF4 using a horseradish peroxidase system so that MRF4-positive nuclei appear dark (A–F) or for acetylcholine receptor with a fluorescent system so that receptor clusters appear white (G–I). MRF4 was barely detectable in diaphragm and jaw muscles of myogenin-deficient fetuses that did not express the transgene (A, D); whereas it was expressed at high levels in these muscles of myogenin-deficient littermates that expressed the transgene (B, E). In myogenin-positive fetuses, MRF4 was expressed at high levels in jaw muscles, but not at all in the diaphragm (C, F). Acetylcholine receptor clusters were found in littermates with all combinations of myogenin and transgene expression (G–I). Scale bar 5 40 µm for A–F and 60 µm for G–I. bodies to the synaptic vesicle protein, SV2 (Buckley and Kelly, 1985). SV2 staining was concentrated near AChR clusters in both myogenin-deficient and myogeninpositive muscles, with or without the transgene (not shown). Though expression of myo1565-MRF4 did not completely restore myogenin-deficient mice to wild-type, the transgene did produce a partial rescue of the myogenin-deficient phenotype. First, limbs of E18–19 myogenin-deficient fetuses that expressed the transgene contained several times more myosin heavy chain (MHC) protein than the limbs of myogenin-deficient littermates without the transgene (Fig. 3A). However, even this increased amount of MHC was much less than found in limbs of myogenin-positive fetuses, which contained high levels of MHC with or without the transgene (Fig. 3A). Transgene expression in myogenin-deficient mice also improved myofiber morphology. In the absence of the transgene, the MHC-expressing cells in E18–19 myogenin (2/2) fetuses have a very wide range of morphologies. Some muscles of myogenin (2/2) fetuses, e.g., the diaphragm (Fig. 4D, G) and body wall muscles, contained MHC-expressing cells that were mostly small, mononucleated, and not striated, whereas other muscles, e.g., the intercostal (Fig. 4A) and most limb muscles, contained a mixture of similarly small myofibers and more normal-sized myofibers that were multinucleated and striated. In contrast, when the transgene was expressed in myogenin (2/2) fetuses, such very small MHC-expressing cells were no longer present in the diaphragm (Fig. 4E, H), body wall, or intercostal (Fig. 4B) muscles; rather these muscles contained multinucleated, striated, and apparently normal diameter myofi- 236 ZHU AND MILLER Fig. 3. Effects of myo1565-MRF4 expression on the myosin heavy chain content of myogenin-deficient muscles. A: Expression of myo1565MRF4 increased the total amount of myosin in myogenin-deficient limb muscles. Crude extracts were prepared from limbs of E18–19 littermates that had the indicated combinations of myo1565-MRF4 (locus 43) and myogenin expression. Equal amounts (10 µg) of total protein were analyzed in each lane by SDS-PAGE and immunoblotting with mAb F59, which reacts with all isoforms of mouse MHCs. MHC was more abundant in the limbs of myogenin-deficient fetuses that expressed the transgene than in myogenin-deficient littermates that did not express the transgene, though it was still less abundant than in myogenin-expressing limbs. B: With or without the transgene, myogenin-deficient muscles had more slow MHC and less perinatal MHC than myogenin-expressing muscles. MHC was purified from the crude extracts described in A, and equal amounts (0.5 µg) of total MHC were analyzed in each lane by SDS-PAGE and immunoblotting. As indicated, all MHC isoforms were detected and equal sample loading was confirmed with mAb F59; the slow MHC isoform was detected with mAb BA-F8; and the perinatal MHC isoform was detected with mAb BF-34. All samples were analyzed on the same gel and immunoblot; some lanes were juxtaposed for presentation. bers, which were similar to myofibers in the same muscles of myogenin-positive fetuses (Fig. 4C, F, I). In addition, very large diameter myofibers were found in myogenin-deficient fetuses that expressed myo1565-MRF4 but not in myogenin-deficient littermates without the transgene. We identified and measured the diameters of the largest myofibers in sections of E18 littermates. The mean diameter 6 SD was 12.8 6 2.2 µm for myogenin (2/2) myofibers (n 5 112), 18.2 6 3.5 µm (n 5 123) for myofibers that were myogenin-deficient but expressed myo1565-MRF4, and 16.1 6 3.4 µm (n 5 129) for myogenin-positive myofibers. The largest fibers in myogenin-deficient muscles were significantly smaller (P , 0.01) than those in either myogenin-positive or myogenin-deficient, transgene-positive muscles. Though expression of the transgene improved myofiber morphology, it did not restore normal numbers of myofibers. Most muscles of myogenin-positive fetuses, e.g., the diaphragm and intercostals (Fig. 4C, F), contained many more myofibers than did the same muscles in myogenin (2/2) littermates that expressed myo1565MRF4 (Fig. 4B, E). Thus, when compared to myogeninpositive muscles, the myofibers in many myogenin (2/2), transgene-negative muscles were fewer in number and poorly formed; whereas myofibers in the same myogenin (2/2), transgene-positive muscles, though still few in number, were well formed. Expression of the myo1565-MRF4 transgene also improved, albeit incompletely, the abnormal ribcages of myogenin-deficient neonates (Fig. 5). As noted by Hasty et al. (1993), the ribs of myogenin-deficient fetuses are short, nearly straight, and connect perpendicularly to the sternum and vertebrae, whereas ribs in myogeninpositive fetuses are longer, markedly curved, and connect to the sternum and vertebrae at acute angles (Fig. 5). We consistently (n 5 17) observed that when the transgene was expressed in myogenin-deficient fetuses, the ribs were longer and markedly more curved than in myogenin-deficient littermates without the transgene, though not as long or curved as in myogenin-positive littermates (Fig. 5). The ribs in transgene-expressing mice also attached to the sternum and vertebrae at an acute angle, that was clearly different from the nearly perpendicular attachment in myogenin-deficient mice without the transgene, though not as acute as in myogenin-positive littermates (Fig. 5). Further analyses suggested that the transgene specifically failed to improve development of later forming, slow MHC-negative fibers in myogenin (2/2) mice. First, limb muscles from myogenin-positive mice contained a smaller proportion of the slow MHC isoform and more of the perinatal isoform than did limb muscles from myogenin (2/2) littermates, with or without the transgene (Fig. 3B). In muscles of all genotypes, the Adult Fast Type IIA, IIB, or IIX isoforms were not detectable, whereas the amount of the embryonic MHC isoform paralleled the amount of total MHC (not shown). Furthermore, muscles in myogenin-positive fetuses contained a smaller percentage of slow MHC-expressing myofibers than those in myogenin (2/2), transgenepositive fetuses (Fig. 4, compare panels F and I to panels E and H). In the E19 diaphragm, only 10.2% (51 out of 500) of the myofibers expressed slow MHC in myogenin-positive fetuses, whereas 53% (187 out of 353) expressed slow MHC in myogenin (2/2), transgenepositive littermates (P , 0.0001). Slow MHC expression was also found in diaphragms of myogenindeficient littermates without the transgene (Fig. 4G), but these fibers were poorly formed making it difficult to determine the percentage that expressed slow MHC. Slow MHC-expressing myofibers are among the first to be formed during myogenesis, whereas later forming myofibers generally express perinatal but not slow MHC (Condon et al., 1990). Thus, myogenin (2/2) fetuses, even with the transgene, appear defective in later stages of myogenesis. Lack of expression of the myogenin promoter fragment in late stages of myogenesis prevented full rescue of the myogenin-deficient phenotype by the myo1565MRF4 transgene (Fig. 6). In E18–19 myogenin-positive fetuses that carried a myo1565-LacZ transgene (Cheng et al., 1992), we found that LacZ was expressed in only a subset of the myofibers and muscles (Fig. 6A–D). (This expression was not fiber type-specific, because many mixed fiber type muscles such as the diaphragm showed no staining.) In contrast, we consistently found 237 MRF4 IN MYOGENIN (2/2) MICE Fig. 4. The myo1565-MRF4 transgene improved myofiber morphology in myogenin-deficient muscles. Sections were prepared of intercostal (A–C) and diaphragm (D–I) muscles of E19 littermates with the indicated combinations of myo1565-MRF4 (locus 43) and myogenin expression. Sections were analyzed either by immunostaining for all MHC with only mAb F59 (A–C) or by double immunostaining for both all MHC with mAb F59 (D–F) and slow MHC with mAb BA-F8 (G–I). Myogenin-deficient muscles that did not express the transgene (A, D) contained large numbers of very small MHC-expressing cells; in contrast, such small myocytes were nearly absent from myogenin-deficient muscles that expressed the transgene (B, E). Diaphragms of myogenin-deficient, transgene-expressing fetuses (E, H) contained a larger percentage of slow MHC-expressing myofibers, though fewer myofibers in total, than the diaphragms of myogenin-positive littermates (F, I). Scale bar 5 45 µm. the myogenin protein throughout myo1565-LacZ fetuses, even in muscles that failed to express LacZ (Fig. 6E). Near the time of birth, therefore, the myo1565 promoter fragment was expressed in many fewer myogenic cells than the endogenous myogenin gene, whereas at earlier stages of myogenesis (,E15) expression of this promoter fragment more closely matches that of the endogenous gene in myogenic cells (Cheng et al., 1992; Block et al., 1996). Finally, we examined how the quantity, as well as the location, of transgene expression affected the results. The analyses described above focused on mice that were heterozygous for locus 43 of the myo1565-MRF4 transgene. We also produced myogenin-deficient fetuses that were homozygous for locus 43 and found that they contained about twice as much of both the transgene and the endogenous MRF4 mRNAs as heterozygotes (not shown). Despite the increased MRF4 expression, homozygotes did not show any improvements in phenotype beyond those seen in heterozygotes. Furthermore, we found that two additional transgene loci, 42-2 and 42-5, were expressed at ,10% of the level of locus 43 (not shown). The low levels of expression from these loci were sufficient to produce small increases in the amount of endogenous MRF4 mRNA in myogenin-deficient fetuses (not shown), but were insufficient to produce the improvements in diaphragm myofibers or the ribcage that were produced by the much more highly expressed locus 43 (not shown). DISCUSSION Myogenin-deficient fetuses that expressed the myogenin promoter-MRF4 transgene, myo1565-MRF4, had more myosin, larger myofibers, increased expression of the endogenous MRF4 gene, and improved ribcage morphology compared to their myogenin-deficient littermates that did not express the transgene. On the other hand, the transgene failed to restore normal numbers of myofibers or viability to myogenin-deficient mice, apparently because the ,1.6 kb myogenin promoter fragment was not expressed in most late-forming myofibers. However, based on the improved phenotype of 238 ZHU AND MILLER Fig. 5. Expression of myo1565-MRF4 improved ribcage morphology in myogenin-deficient mice. Alcian Blue was used to stain cartilage in skeletons from E19 littermates with the indicated combinations of myo1565-MRF4 (locus 43) and myogenin expression. Ribcages of myogenin-deficient neonates that expressed the transgene (middle) were larger in diameter, had longer ribs, and showed more normal angles of rib attachment than ribcages of myogenin-deficient littermates that did not express the transgene (left). Missing limbs were used for additional biochemical or histological assays. Fig. 6. The myo1565-LacZ transgene is not expressed in the same pattern as the myogenin protein in myogenin-positive fetuses. Sections of E18–19 myo1565-LacZ fetuses were stained with X-Gal so that LacZpositive cells appears dark (A–D) or by immunofluorescence with an anti-myogenin mAb so that myogenin-expressing nuclei appear light (E). Phase contrast (A) and bright-field (B) views show a back muscle in which a subset of the myofibers expressed LacZ. The additional phase contrast (C) and bright-field (D) views show a different shoulder muscle in which none of the myofibers expressed LacZ. Though the muscle shown in C and D failed to express the myogenin promoter-LacZ transgene, staining of an adjacent section showed that many myonuclei did contain the myogenin protein (E). Scale bar 5 30 µm. MRF4 IN MYOGENIN (2/2) MICE myogenin-deficient fetuses that expressed the transgene, we conclude that MRF4 is able, at least in early stages of myogenesis, to functionally substitute for myogenin. MRF4 proved able to substitute for myogenin to increase expression of the endogenous MRF4 gene. There is a very low level of MRF4 mRNA in the muscles of myogenin (2/2) fetuses (Hasty et al., 1993), whereas myogenin-positive fetuses have high levels of MRF4 mRNA. Expression of the myo1565-MRF4 transgene in myogenin-deficient fetuses led to a large increase in MRF4 mRNA produced from the endogenous MRF4 gene, but did not affect expression of either MyoD or Myf-5. Thus, in the presence of the transgene, MRF4 substituted for myogenin to activate, perhaps indirectly, its own promoter, at least in those fibers that formed in myogenin-deficient, myo1565-MRF4-positive fetuses. Such positive autoregulation, if found in all normal myofibers, could maintain the high level of MRF4 found in adult myofibers that ordinarily have little or no myogenin (Musaro et al., 1995). MRF4 was also able to substitute for myogenin in the development of early-forming myofibers. In the absence of myogenin, the myosin-expressing cells in some fetal muscles, notably the diaphragm, were mononucleate and not striated (Fig. 5, cf. Hasty et al., 1993; Nabeshima et al., 1993); whereas upon MRF4 expression from myo1565-MRF4 the same muscles contained a population of large diameter, striated, and multinucleated myofibers even in the absence of myogenin. These ‘‘rescued’’ myofibers were predominantly slow MHCexpressing and thus likely to have been primary fibers that formed early in development (Condon et al., 1990). We cannot be sure that MRF4 restored every aspect of muscle gene expression to these early myogenindeficient fibers; nonetheless, it is clear that MRF4 is able to substitute for myogenin to markedly improve early myogenesis. Alternative approaches are needed to determine whether MRF4 is similarly able to substitute for myogenin in fetal myogenesis. Though expression of the myo1565 fragment appears to closely match that of the endogenous myogenin gene in myogenic cells until E15 (Cheng et al., 1992, 1995; Block et al., 1996), we found in older fetuses that the myo1565 promoter fragment was expressed in many fewer muscles and myofibers than the myogenin protein. Thus, introduction of the myo1565-MRF4 transgene into myogenin-deficient fetuses did not test whether MRF4 could substitute for myogenin in the predominantly fast myofibers that form during later stages of development. Perhaps the myo1565 fragment lacks regulatory elements necessary for later expression. Alternatively, this fragment may be more sensitive than the endogenous promoter to repression by MRF4, which becomes abundant and appears to repress myogenin expression in neonates (Zhang et al., 1995). It might be possible, by using knock-in technology (Wang et al., 1996) or a longer myogenin promoter fragment (Buonanno et al., 1993; Yee and Rigby, 1993), to determine if MRF4 substitutes 239 for myogenin in later stages of myogenesis, though even these approaches may fail if MRF4 represses the myogenin promoter in late fetal cells. Neuromuscular junctions appeared to form in myogenin-deficient mice. We found by RT-PCR and staining with alpha-Bungarotoxin that the alpha subunit of the AChR was expressed in myofibers and assembled into normal looking clusters even in the absence of myogenin. In contrast, an initial analysis by a less sensitive technique, northern blotting, suggested that the mRNA for alpha-AChR was not expressed in myogenindeficient mice (Hasty et al., 1993), leading to the suggestion that myogenin-deficient mice die due to lack of neuromuscular junctions (Venuti et al., 1995). However, we observed that myogenin-deficient and wildtype myofibers, with or without the transgene, had single large AChR clusters. Such clusters indicate successful innervation, because never-innervated myofibers have multiple small AChR clusters rather than single large clusters (Gautam et al., 1996). The presence of a properly localized presynaptic protein, SV2, and motor neuron axons (Brennan et al., 1996) also suggests that neuromuscular junctions form in myogenin-deficient mice. Whether these synapses are fully functional remains to be determined, but, if so, it is likely that myogenin-deficient neonates die because they have insufficient muscle mass, specifically in the diaphragm, to breathe successfully. To what extent are the MRFs functionally interchangeable? In culture, each of the four MRFs is able to convert non-myogenic C3H10T1/2 cells into myogenic cells and each is able to trans-activate particular muscle-specific promoters (Weintraub, 1993), suggesting that MRFs are functionally interchangeable. On the other hand, some promoters are trans-activated by only a subset of the MRFs (e.g., Yutzey et al., 1990) and analyses of MRF-deficient mice show that each MRF plays a unique role in normal myogenesis (Rudnicki and Jaenisch, 1995), suggesting that MRFs may not be interchangeable. Recently, Wang et al. (1996) showed by knock-in that myogenin can substitute for Myf-5 in vivo; and we show here that MRF4 can, as so far tested, substitute for myogenin in vivo. Thus, in at least two cases, one MRF can substitute for another. Further work in vivo is needed to determine whether these MRF pairs are functionally interchangeable in the reverse order, i.e., Myf-5 for myogenin or myogenin for MRF4. It is sometimes argued that natural selection must eliminate genes that encode interchangeable proteins (Hochgeschwender and Brennan, 1994). Knock-out experiments, however, suggest that some genes in mice are dispensable, though such conclusions must be tempered by the lack of natural selection in a laboratory and the limited number of assays performed (Colucci-Guyon et al., 1994; Galou et al., 1996). On the other hand, knock-in experiments with engrailed and MRF genes (Hanks et al., 1995; Wang et al., 1995), as well as our work, show clearly that some functions can be carried out by more than one transcription factor. Further work will show whether these factors are 240 ZHU AND MILLER completely interchangeable or have a subset of unique functions. Perhaps the coding regions of potentially interchangeable proteins can remain in a genome as long as their regulatory regions produce distinct expression patterns. EXPERIMENTAL PROCEDURES Mice Transgenic mice carrying a myogenin promoter-MRF4 fusion gene—termed myo1565-MRF4—were generated previously (Block et al., 1996). The myo1565-MRF4 fusion gene includes the rat MRF4 cDNA (Rhodes and Konieczny, 1989) under control of nucleotides 21565 to 118, relative to the transcription start site, of the mouse myogenin promoter region (Cheng et al., 1992). Adult myogenin (1/2) heterozygotes (mygml/1; Hasty et al., 1993) as well as adult myo1565-LacZ mice (Cheng et al., 1992) were a gift of E.N. Olson. We crossed myo1565-MRF4 heterozygotes (Block et al., 1996) with myogenin (1/2) heterozygotes to generate mice that were both myo1565-MRF4 (1/2) and myogenin (1/2). These double heterozygotes were then crossed with myogenin (1/2) mice. Among the progeny of these crosses were myogenin (2/2) homozygotes, of which approximately half also carried the myo1565MRF4 transgene. DNA and RNA Analyses A PCR assay identified mice that carried myo1565MRF4 (Block et al., 1996), and a Southern blot assay identified myogenin (2/2), (1/2), and (1/1) mice (Hasty et al., 1993). To identify endogenous and transgene MRF4 mRNAs, 10 µg samples of total RNA isolated (Chomczynski and Sacchi, 1987) from skinned and deboned hindlimb muscles of E17 to P1 mice were separated by electrophoresis, transferred to nylon, and hybridized with a 32P-labeled MRF4 probe (Block et al., 1996). Probes for myogenin, MyoD, and Myf-5 mRNAs were also used (Block et al., 1996; Dominov and Miller, 1996). RT-PCR was used to detect mRNA for the alpha subunit of the acetylcholine receptor (Wheatley et al., 1992). Histology and Antibodies Skeletons of fetal and newborn mice were prepared and stained with Alcian Blue to detect cartilage (Zhang et al., 1995). For immunohistology, all muscle isoforms of myosin heavy chain (MHC) were detected with mAb F59 (Miller et al., 1985; Miller and Stockdale, 1986a,b; Miller et al., 1989; Miller, 1990; Smith and Miller, 1992); the slow MHC isoform was detected with mAb BA-F8 and the perinatal MHC isoform was detected with mAb BF-34 (Gorza, 1990); and the Adult Fast Type IIA, IIB, and IIX, as well as the embryonic, MHC isoforms were detected with specific mAbs (Gorza, 1990). Skeletal muscle isoforms of myosin light chains (MLC) were detected with mAbs F310 and T14 (Crow et al., 1983). Myogenin protein was detected with mAb F5D (Wright et al., 1991). The MyoD, Myf-5, and MRF4 proteins were detected with specific polyclonal antisera (Smith et al., 1993, 1994). The synaptic vesicle-specific SV2 protein was detected with mAb 10H, a gift of K.M. Buckley (Buckley and Kelly, 1985). 5-bromo-4-chloro-3indolyl-b-D-galactoside (X-gal) was used as substrate to locate b-Galactosidase activity in sections of myo1565LacZ fetuses (Cheng et al., 1995). Tissue sections were prepared, incubated with primary antibodies, and stained with appropriate fluorescein-, Texas Red-, Cy3-, or horseradish peroxidase-conjugated secondary antibodies (Kachinsky et al., 1994; Smith et al., 1993, 1994; Block et al., 1996). Sections from unfixed or paraformaldehyde-fixed tissues were used to detect MHC isoforms, whereas only paraformaldehyde-fixed tissues were used to detect other antigens and for multiple staining protocols (Miller et al., 1985; Smith et al., 1994). As controls, primary antibodies were omitted from selected sections on a slide. To identify acetylcholine receptor (AChR) clusters, sections were incubated sequentially with biotinylated alpha-Bungarotoxin (Molecular Probes, Eugene OR) and Cy3-conjugated streptavidin (Jackson Immunoresearch, Malvern, PA), with each reagent used at 1 µg/ml for 1 hr at room temperature. As a control, staining was abolished when unlabeled alpha-Bungarotoxin was added at 50 µg/ml. To measure myofiber diameters, sections made through corresponding regions of littermates were examined and the largest myofibers were identified in matched muscles in each section. The diameters of these large myofibers were measured, mean diameters 6 S.D. were calculated, and mean differences were examined for statistical significance by ANOVA and t-test. Differences in fiber type proportions were examined for significance using Fisher’s exact test and Chi square with Yates correction using the InStat computer program (v. 1.12, GraphPad Software, San Diego, CA). ACKNOWLEDGMENTS We thank Kathleen Buckley, Rick Boyce, Janice Dominov, S.F. Konieczny, En Li, Jennifer Moss, Craig Neville, Eric Olson, Ken Rosen, Nadia Rosenthal, Frank Stockdale, and W.E. Wright for antibodies, reagents and/or advice. Z.Z. was supported by a NIH postdoctoral fellowship. This work was supported by grants from the NIH, USDA, and Muscular Dystrophy Association, and an Established Investigatorship from the American Heart Association to J.B.M. REFERENCES Atchley, W.R., Fitch, W.M., and Bronner-Fraser, M. (1994) Molecular evolution of the MyoD family of transcription factors. Proc. Natl. Acad. Sci. USA 91:11522–11526. Black, B.L., Martin, J.F., and Olson, E.N. (1995) The mouse MRF4 promoter is trans-activated directly and indirectly by musclespecific transcription factors. J. Biol. Chem. 270:2889–2892. Block, N.E., and Miller, J.B. (1992) Expression of MRF4, a myogenic helix-loop-helix protein, produces multiple changes in the myogenic program of BC3H-1 cells. Mol. Cell. Biol. 12:2484–2492. Block, N.E., Zhu, Z., Kachinsky, A.M., Dominov, J.A., and Miller, J.B. (1996) Acceleration of somitic myogenesis in embryos of myogenin promoter-MRF4 transgenic mice. Dev. Dyn. 207:382–394. Buonanno, A., Edmondson, D.G., and Hayes, W.P. (1993) Upstream sequences of the myogenin gene convey responsiveness to skeletal muscle denervation in transgenic mice. Nuc. Acids Res. 21:5684–5693. MRF4 IN MYOGENIN (2/2) MICE Braun, T., Rudnicki, M.A., Arnold, H.-H., and Jaenisch, R. (1992) Targeted inactivation of the muscle regulatory gene Myf-5 results in abnormal rib development and perinatal death. Cell 71:3369–3382. Braun, T., and Arnold, H.-H. (1995) Inactivation of Myf-6 and Myf-5 leads to alterations of skeletal muscle development. EMBO J. 14:1176–1186. Brennan, T.J., Olson, E.N., Klein, W.H., and Winslow, J.W. (1996) Extensive motor neuron survival in the absence of secondary skeletal muscle fiber formation. J. Neurosci. Res. 45:57–68. Buckingham, M. (1992) Making muscle in mammals. Trends Gen. 8:144–149. Buckley, K.M., and Kelly, R.B. (1985) Identification of a transmembrane glycoprotein specific for secretory vesicles of neural and endocrine cells. J. Cell Biol. 100:1284–1294. Chakraborty, T., and Olson, E.N. (1991) Domains outside of the DNA-binding domain impart target gene specificity to myogenin and MRF4. Mol. Cell. Biol. 11:6103–6108. Cheng, T.-C., Hanley, T.A., Mudd, J., Merlie, J.P., and Olson, E.N. (1992) Mapping of myogenin transcription during embryogenesis using transgenes linked to myogenin control region. J. Cell Biol. 119:1649–1656. Cheng, T.C., Tseng, B.S., Merlie, J.P., Klein, W.H., and Olson, E.N. (1995) Activation of the myogenin promoter during mouse embryogenesis in the absence of positive autoregulation. Proc. Natl. Acad. Sci. USA 92:561–565. Chomczynski, P., and Sacchi, N. (1987) A single-step method of RNA isolation by acid guanidinium thiocyanate-phenol-chloroform extraction. Anal. Biochem. 162:156–159. Colucci-Guyon, E., Portier, M.M., Dunia, I., Paulin, D., Pournin, S., and Babinet, C. (1994) Mice lacking vimentin develop and reproduce without an obvious phenotype. Cell 79:679–694. Condon, K., Silberstein, L., Blau, H.M., and Thompson, W.J. (1990) Development of muscle fiber types in the prenatal rat hindlimb. Dev. Biol. 138:256–274. Crow, M.T., Olson, P.S., and Stockdale, F.E. (1983) Myosin light chain expression during avian muscle development. J. Cell Biol. 96:736–744. Dominov, J.A., and Miller, J.B. (1996) POU homedomain genes and myogenesis. Dev. Genet. 19:108–118. Galou, M., Colucci-Guyon, E., Ensergueix, D., Ridet, J.L., Ribotta, M.G.Y., Privat, A., Babinet, C., and Dupouey, P. (1996) Disrupted glial fibrillary acidic protein network in astrocytes from vimentin knockout mice. J. Cell Biol. 133:853–863. Gautam, M., Noakes, P.G., Moscoso, L., Rupp, R., Scheller, R.H., Merlie, J.P., and Sanes, J.R. (1996) Defective neuromuscular synaptogenesis in agrin-deficient mutant mice. Cell 85:525–535. Gorza, L. (1990) Identification of a novel type 2 fiber population in mammalian skeletal muscle by combined used of histochemical ATPase and anti-myosin monoclonal antibodies. J. Histochem. Cytochem. 38:257–265. Hanks, M., Wurst, W., Anson-Cartwright, L., Auerbach, A.B., and Joyner, A.L. (1995) Rescue of the En-1 mutant phenotype by replacement of En-1 with En-2. Science 269:679–682. Hasty, P., Bradley, A., Morris, J.H., Edmondson, D.G., Venuti, J.M., Olson, E.N., and Klein, W.H. (1993) Muscle deficiency and neonatal death in mice with a targeted mutation in the myogenin gene. Nature 364:501–506. Hinterberger, T., Sassoon, D.A., Rhodes, S.J., and Konieczny, S.F. (1991) Expression of the muscle regulatory factor MRF4 during somite and skeletal myofiber development. Dev. Biol. 147:144–156. Hochgeschwender, U., and Brennan, M.B. (1994) Redundant genes? Nature Genet. 8:219–220. Kachinsky, A.M., Dominov, J.A., and Miller, J.B. (1994) Myogenesis and the intermediate filament protein, nestin. Dev. Biol. 165:216–228. Miller, J.B. (1990) Myogenic programs of mouse muscle cell lines: Expression of myosin heavy chain isoforms, MyoD1, and myogenin. J. Cell Biol. 111:1149–1159. Miller, J.B., Crow, M.T., and Stockdale, F.E. (1985) Slow and fast myosin heavy chain expression defines three types of myotubes in early muscle cell cultures. J. Cell Biol. 101:1643–1650. Miller, J.B., and Stockdale, F.E. (1986) Developmental origins of skeletal muscle fibers: Clonal analysis of myogenic cell lineages based on fast and slow myosin heavy chain expression. Proc. Nat. Acad. Sci. USA 83:3860–3864. 241 Miller, J.B., and Stockdale, F.E. (1986) Developmental regulation of the multiple myogenic cell lineages of the avian embryo. J. Cell Biol. 103:2197–2208. Miller, J.B., Teal, S.B., and Stockdale, F.E. (1989) Evolutionarily conserved sequences specific for striated muscle myosin heavy chain isoforms: Epitope mapping by cDNA expression. J. Biol. Chem. 264:13122–13130. Musaro, A., Deangelis, M.G.C., Germani, A., Ciccarelli, C., Molinaro, M., and Zani, B.M. (1995) Enhanced expression of myogenic regulatory genes in aging skeletal muscle. Exp. Cell Res. 221:241–248. Nabeshima, Y., Hanaoka, K., Hayasaka, M., Esumi, E., Li, S., Nonaka, I., and Nabeshima, Y. (1993) Myogenin gene disruption results in perinatal lethality because of severe muscle defect. Nature 364:532– 535. Naidu, P.S., Ludolph, D.C., To, R.Q., Hinterberger, T.J., and Konieczny, S.F. (1995) Myogenin and MEF2 function synergistically to activate the MRF4 promoter during myogenesis. Mol. Cell. Biol. 15:2707–2718. Patapoutian, A., Yoon, J.K., Miner, J.H., Wang, S., Stark, K., and Wold, B. (1995) Disruption of the mouse MRF4 gene identifies multiple waves of myogenesis in the myotome. Development 121:3347–3358. Rawls, A., Morris, J.H., Rudnicki, M., Braun, T., Arnold, H.-H., Klein, W.H., and Olson, E.N. (1995) Myogenin’s functions do not overlap with those of MyoD or Myf-5 during mouse embryogenesis. Dev. Biol. 172:37–50. Rhodes, S.J., and Konieczny, S.F. (1989) Identification of MRF4: A new member of the muscle regulatory factor gene family. Genes Dev. 3:2050–2061. Rudnicki, M.A., Braun, T., Hinuma, S., and Jaenisch, R. (1992) Inactivation of MyoD in mice leads to upregulation of the myogenic HLH gene Myf-5 and results in apparently normal muscle development. Cell 71:383–390. Rudnicki, M.A., Schnegelsberg, P., Stead, R.H., Braun, T., Arnold, H.-H., and Jaenisch, R. (1993) MyoD or Myf-5 is required for the formation of skeletal muscle. Cell 75:1351–1360. Rudnicki, M.A., and Jaenisch, R. (1995) The MyoD family of transcription factors and skeletal myogenesis. Bioessays 17:203–209. Smith, T.H., and Miller, J.B. (1992) Distinct myogenic programs of embryonic and fetal mouse muscle cells: Expression of the perinatal myosin heavy chain isoform in vitro. Dev. Biol. 149:16–26. Smith, T.H., Block, N.E., Rhodes, S.J., Konieczny, S.F., and Miller, J.B. (1993) A unique pattern of expression of the four muscle regulatory factor proteins distinguishes somitic from embryonic, fetal, and newborn mouse myogenic cells. Development 117:1125–1133. Smith, T.H., Kachinsky, A.M., and Miller, J.B. (1994) Somite subdomains, muscle cell origins, and the four muscle regulatory factor proteins. J. Cell Biol. 127:95–105. Venuti, J., Morris, J., Vivian, J.L., Olson, E.N., and Klein, W.H. (1995) Myogenin is required for late but not early myogenesis during mouse development. J. Cell Biol. 128:563–576. Wang, Y., Schnegelsberg, P.N.J., Dausman, J., and Jaenisch, R. (1996) Functional redundancy of the muscle-specific transcription factors Myf5 and myogenin. Nature 379:823–825. Weintraub, H. (1993) The MyoD family and myogenesis: Redundancy, networks, and thresholds. Cell 75:1241–1244. Wheatley, L.M., Urso, D., Tumas, K., Maltzman, J., Loh, E., and Levinson, A.I. (1992) Molecular evidence for the expression of nicotinic acetylcholine receptor alpha-chain in mouse thymus. J. Immunol. 148:3105–3109. Wright, W.E., Binder, M., and Funk, W. (1991) Cyclic amplification and target selection (CASTING) for the myogenin consensus binding site. Mol. Cell. Biol. 11:4104–4111. Yee, S.-P., and Rigby, P.W.J. (1993) The regulation of myogenin gene expression during embryonic development of the mouse. Genes Dev. 7:1277–1289. Yutzey, K.E., Rhodes, S., and Konieczny, S. (1990) Differential trans activation associated with the muscle regulatory factors MyoD1, myogenin, and MRF4. Mol. Cell. Biol. 10:3934–3944. Zhang, W., Behringer, R.R., and Olson, E.N. (1995) Inactivation of the myogenic bHLH gene MRF4 results in up-regulation of myogenin and rib anomalies. Genes Dev. 9:1388–1399.