Changes in architecture of the Golgi complex and other subcellular organelles during myogenesis

Myogenesis involves changes in both gene expression and cellular architecture. Little is known of the organization, in muscle in vivo, of the subcellular organelles involved in protein synthesis despite the potential importance of targeted protein synthesis for formation and maintenance of functional domains such as the neuromuscular junction. A panel of antibodies to markers of the ER, the Golgi complex, and the centrosome were used to localize these organelles by immunofluorescence in myoblasts and myotubes of the mouse muscle cell line C2 in vitro, and in intact single muscle fibers from the rat flexor digitorum brevis. Antibodies to the ER stained structures throughout the cytoplasm of both C2 myoblasts and myotubes. In contrast, the spatial relationship between nucleus, centrosome, and Golgi complex was dramatically altered. These changes could also be observed in a low- calcium medium that allowed differentiation while preventing myoblast fusion. Muscle fibers in vivo resembled myotubes except that the ER occupied a smaller volume of cytoplasm and no staining was found for one of the Golgi complex markers, the enzyme alpha-mannosidase II. Electron microscopy, however, clearly showed the presence of stacks of Golgi cisternae in both junctional and extrajunctional regions of muscle fibers. The perinuclear distribution of the Golgi complex was also observed in live muscle fibers stained with a fluorescent lipid. Thus, the distribution of subcellular organelles of the secretory pathway was found to be similar in myotubes and muscle fibers, and all organelles were found in both junctional and extrajunctional areas of muscle.     

Presence of a protein immunologically related to lamin B in the postsynaptic membrane of Torpedo marmorata electrocyte

The Torpedo electrocyte is a flattened syncytium derived from skeletal muscle, characterized by two functionally distinct plasma membrane domains. The electrocyte is filled up with a transversal network of intermediate filaments (IF) of desmin which contact in an end-on fashion both sides of the cell. In this work, we show that polyclonal antibodies specific for lamin B recognizes a component of the plasma membrane of Torpedo electrocyte. This protein which thus shares epitopes with lamin B has a relative molecular mass of 54 kD, an acidic IP of 5.4. It is localized exclusively on the cytoplasmic side of the innervated membrane of the electrocyte at sites of IF-membrane contacts. Since our previous work showed that the noninnervated membrane contains ankyrin (Kordeli, E., J. Cartaud, H. O. Nghiem, L. A. Pradel, C. Dubreuil, D. Paulin, and J.-P. Changeux. 1986. J. Cell Biol. 102:748-761), the present results suggest that desmin filaments may be anchored via the 54-kD protein to the innervated membrane and via ankyrin to the noninnervated membrane. These findings would represent an extension of the model proposed by Georgatos and Blobel (Georgatos, S. D., and G. Blobel. 1987a. J. Cell Biol. 105:105-115) in which type III intermediate size filaments are vectorially inserted to plasma and nuclear membranes by ankyrin and lamin B, respectively.     

Evidence for a polarity in the distribution of proteins from the cytoskeleton in Torpedo marmorata electrocytes

The subcellular distribution of the 43,000-D protein (43 kD or v1) and of some major cytoskeletal proteins was investigated in Torpedo marmorata electrocytes by immunocytochemical methods (immunofluorescence and immunogold at the electron microscope level) on frozen-fixed sections and homogenates of electric tissue. A monoclonal antibody directed against the 43-kD protein (Nghiêm, H. O., J. Cartaud, C. Dubreuil, C. Kordeli, G. Buttin, and J. P. Changeux, 1983, Proc. Natl. Acad. Sci. USA, 80:6403-6407), selectively labeled the postsynaptic membrane on its cytoplasmic face. Staining by anti-actin and anti-desmin antibodies appeared evenly distributed within the cytoplasm: anti-desmin antibodies being associated with the network of intermediate-sized filaments that spans the electrocyte, and anti-actin antibodies making scattered clusters throughout the cytoplasm without preferential labeling of the postsynaptic membrane. On the other hand, a dense coating by anti-actin antibodies became apparent on the postsynaptic membrane in homogenates of electric tissue pointing to the possible artifactual redistribution of a soluble cytoplasmic actin pool. Anti-fodrin and anti-ankyrin antibodies selectively labeled the non-innervated membrane of the cell. F actin was also detected in this membrane. Filamin and vinculin, two actin-binding proteins recently localized at the rat neuromuscular junction (Bloch, R. J., and Z. W. Hall, 1983, J. Cell Biol., 97:217-223), were detected in the electrocyte by the immunoblot technique but not by immunocytochemistry. The data are interpreted in terms of the functional polarity of the electrocyte and of the selective interaction of the cytoskeleton with the innervated and non-innervated domains of the plasma membrane.     

Expression of several adhesive macromolecules (N-CAM, L1, J1, NILE, uvomorulin, laminin, fibronectin, and a heparan sulfate proteoglycan) in embryonic, adult, and denervated adult skeletal muscle

Levels of the neural cell adhesion molecule N-CAM in muscle are regulated in parallel with the susceptibility of muscle to innervation: N-CAM is abundant on the surface of early embryonic myotubes, declines in level as development proceeds, reappears when adult muscles are denervated or paralyzed, and is lost after reinnervation (Covault, J., and J. R. Sanes, 1985, Proc. Natl. Acad. Sci. USA, 82:4544-4548). Here we used immunocytochemical methods to compare this pattern of expression with those of several other molecules known to be involved in cellular adhesion. Laminin, fibronectin, and a basal lamina-associated heparan sulfate proteoglycan accumulate on embryonic myotubes after synapse formation, and their levels change little after denervation. L1, J1, nerve growth factor-inducible large external protein, uvomorulin, and a carbohydrate epitope (L2/HNK-1) shared by several adhesion molecules are undetectable on the surface of embryonic, perinatal, adult, or denervated adult muscle fibers. Thus, of the molecules tested, only N-CAM appears on the surface of muscle cells in parallel with the ability of the muscle cell surface to accept synapses. However, four antigens--N-CAM, J1, fibronectin, and a heparan sulfate proteoglycan--accumulate in interstitial spaces near denervated synaptic sites; regenerating axons traverse these spaces as they preferentially reinnervate original synaptic sites. Of particular interest is J1, antibodies to which block adhesion of central neurons to astrocytes (Kruse, J., G. Keihauer, A. Faissner, R. Timpl, and M. Schachner, 1985, Nature (Lond.), 316:146-148). J1 is associated with collagen and other fibrils in muscle and thus may be an extracellular matrix molecule employed in both the central and peripheral nervous systems.     

Role of nerve and muscle factors in the development of rat muscle spindles

The soleus muscles of fetal rats were examined by electron microscopy to determine whether the early differentiation of muscle spindles is dependent upon sensory innervation, motor innervation, or both. Simple unencapsulated afferent-muscle contacts were observed on the primary myotubes at 17 and 18 days of gestation. Spindles, encapsulations of muscle fibers innervated by afferents, could be recognized early on day 18 of gestation. The full complement of spindles in the soleus muscle was present at day 19, in the region of the neuromuscular hilum. More afferents innervated spindles at days 18 and 19 of gestation than at subsequent developmental stages, or in adult rats; hence, competition for available myotubes may exist among afferents early in development. Some of the myotubes that gave rise to the first intrafusal (bag2) fiber had been innervated by skeletomotor (alpha) axons prior to their incorporation into spindles. However, encapsulated intrafusal fibers received no motor innervation until fusimotor (gamma) axons innervated spindles 3 days after the arrival of afferents and formation of spindles, at day 20. The second (bag1) intrafusal fiber was already formed when gamma axons arrived. Thus, the assembly of bag1 and bag2 intrafusal fibers occurs in the presence of sensory but not gamma motor innervation. However, transient innervation of future bag2 fibers by alpha axons suggests that both sensory and alpha motor neurons may influence the initial stages of bag2 fiber assembly. The confinement of nascent spindles to a localized region of the developing muscle and the limited number of spindles in developing muscles in spite of an abundance of afferents raise the possibility that afferents interact with a special population of undifferentiated myotubes to form intrafusal fibers.     

Differentiation of fiber types in wing muscles during embryonic development: effect of neural tube removal

The embryonic precursors of the avian slow (type I and III) and fast (type II) fibers can be distinguished from each other early in muscle formation (stage 28, V. Hamburger and H. L. Hamilton, J. Morphol, 88, 49-92, 1951) on the basis of the differential sensitivity of their myosin ATPases. To test the neural dependence of fiber type differentiation, the source of motor innervation was eliminated by excision of the brachial neural tube at stages 16-18 before muscles are innervated. Removal of the brachial neural tube did not affect the number of primary myotubes in a sample muscle of the forelimb (ulnimetacarpalis dorsalis, UMD) up until stage 36. Myosin ATPase staining at a variety of pHs revealed the typical patterns of fiber types in muscles of neural-tube free embryos in stages 35-37. These muscles included the anterior latissimus dorsi, brachialis, and UMD which showed presumptive type III staining (type IIIEMB), the pronator superficialis and flexor carpi ulnaris which showed embryonic type II staining (type IIEMB), and the triceps brachii muscles which showed characteristic arrangements of both type IEMB and type IIEMB fibers. The normal patterns of type IEMB and type IIEMB myotubes were also seen in muscles containing a heterogeneous mixture of fiber types such as the biceps brachii, extensor metacarpi radialis, and adductor indicis muscles, although the intensity of acid-stable ATPase staining of the type IEMB myotubes in these muscles was lower than in innervated muscles. It is concluded that the earliest differentiation of muscle fiber types is independent of the nervous system.     

The development of the pattern of innervation in chicken hindlimb muscles: evidence for specification of nerve-muscle connections

The pattern of innervation in 13 chicken hindlimb muscles was studied at various stages of development in order to examine the mechanisms which regulate its formation. The pattern of innervation was visualized by examining the distribution of fiber types within each muscle. It was found that the fiber type which a myotube acquired was influenced by both its time of formation and its position within a muscle. The earliest generation of myotubes (primary) had a marked tendency to become type I fibers, whereas, in contrast, the later generation of myotubes (secondary) tended to differentiate into type II fibers. There were regions of muscle, however, in which primary myotubes differentiated into type II fibers and other regions in which secondary myotubes acquired type I characteristics. During the development of some muscles the pattern of fiber types changed as a result of either a selective loss of type I fibers or, in other cases, a rearrangement of some of the initial neuromuscular contacts. These observations are consistent with the pattern of innervation of a muscle being established as a result of differential projection patterns of fast and slow motoneurons and the existence of some type of chemoaffinity where particular myotubes are preferentially innervated by particular motoneurons.     

Extent of multiple innervation of purkinje cells by climbing fibers in the olivocerebellar system of weaver,reeler, and staggerer mutant mice

Multiple innervation of cerebellar Purkinje cells (PCs) by climbing fibers (CFs) has been described recently in adult weaver, reeler, and staggerer mutant mice, instead of the monoinnervation found in normal adults. In the present study, the extent of this multiple innervation was estimated by two methods, using both evoked and spontaneous activity of the olivocerebellar system. Concordant values were obtained: the mean number of CF collaterals per PC was between 3.5 and 4 in weaver and staggerer and close to 3.2 for the multiply innervated PCs of reeler mice. These values are of the same order of magnitude as those for the transient multiple innervation in developing rats (Mariani and Changeux, 1981a, b).     

Dynamic retention of TGN membrane proteins in Saccharomyces cerevisiae

In a secretory pathway organelle like the Golgi complex, resident proteins are retained in the face of substantial protein flux to subsequent destinations. Recently, molecular genetic strategies have been used to study membrane protein retention in a compartment of the yeast Saccharomyces cerevisiae that is analogous to the trans Golgi network (TGN) of mammalian cells. These studies have defined retention signals containing aromatic amino acids in the TGN proteins' cytoplasmic domains. The identification of mutants that fail to retain TGN proteins has offered the first glimpse into the components involved in retention. The phenotypes of these mutants suggest that retention involves retrieval of TGN proteins from an endosomal compartment.     

Nonacceptance of innervation by innervated neonatal rat muscle

Denervated adult muscle accepts innervation and has high levels of extrajunctional acetylcholine (ACh) receptor, compared to innervated adult muscle. If the high receptor density or any externally oriented part of the receptor molecule permitted or triggered functional synaptogenesis, then innervated neonatal muscle, with its known high extrajunctional sensitivity, should also accept extra synapses from implanted motor nerves. This prediction was tested by implanting the common peroneal nerve into innervated lateral gastrocnemius muscle in 25 neonatal rats and studying the innervation achieved 1–8 weeks later. With one exception, zero or negligible twitch tensions were obtained when the implanted nerve was stimulated. Intracellular recording in two cases showed no evidence of subthresholdevoked potentials in surface muscle fibers. In contrast, when the original motor nerve was cut at the time of common peroneal nerve implantation, reinnervation occurred as soon as 4 days later, and substantial indirect twitches (most observed qualitatively) were invariably found 6–7 days after operation. Four to eight weeks after nerve implantation into denervated muscle, substantial twitch tensions were obtained upon stimulation of the implanted nerve. α-Bungarotoxin binding to extrajunctional ACh receptors per unit surface area was similar in innervated neonatal and denervated adult muscle. Therefore, nonacceptance of additional functional innervation in neonatal muscle implies that a high average density of extrajunctional ACh receptor is not sufficient to permit or trigger functional neuromuscular junction formation.     

Site-directed mutagenesis and deletion of three phosphorylation sites of calsequestrin of skeletal muscle sarcoplasmic reticulum. Effects on intracellular targeting

Calsequestrin (CS) is segregated to the junctional sarcoplasmic reticulum (jSR) of skeletal muscle fibers and is responsible for intraluminal Ca(2+) binding. A chimeric CS-HA1, obtained by adding the nine-amino-acid viral epitope hemagglutinin (HA1) to the carboxy-terminal of CS and shown to be correctly segregated to skeletal muscle jSR in vivo (A. Nori, K. A. Nadalini, A. Martini, R. Rizzuto, A. Villa, and P. Volpe, 1997, Am. J. Physiol. 272, C1420-C1428), is mutagenized in order to identify domains of CS involved in targeting. Since a putative targeting mechanism of CS implies phosphorylation-dependent steps in the endoplasmic reticulum (ER) and/or Golgi complex, five CS-HA1 mutants disrupting the three phosphorylation sites of CS (Thr(189), Thr(229), and Thr(353)) were engineered by either site-directed mutagenesis or deletion: CS-HA1DeltaP1 (Thr(189) --> Ile); CS-HA1DeltaP2 (Thr(229) --> Asn); CS-HA1DeltaP1,2; in which Thr(189) and Thr(229) were changed to Ile and Asn, respectively; and CS-HA1Delta14(COOH) and CS-HA1Delta49 (COOH), in which 14 residues (Glu(354)-Asp(367)) and 49 residues (Asp(319)-Asp(367)), respectively, were deleted at the carboxy-terminal. Mutant cDNAs were transiently transfected in either HeLa cells, cultured myoblasts of rat skeletal muscle, or regenerating soleus muscle fibers of adult rats. Each CS-HA1 mutant was identified by Western blot as a single polypeptide of the predicted molecular weight. The intracellular localization of CS-HA1 mutants was studied by immunofluorescence using specific antibodies against either CS or HA1. CS-HA1 mutants colocalized with ER markers, e.g., calreticulin, and partially overlapped with Golgi complex markers, e.g., alpha-mannosidase II, in HeLa cells and myotubes. CS-HA1 mutants were expressed and retained in ER and ER/SR of HeLa cells and myotubes, respectively, and correctly segregated to jSR of regenerating soleus muscle fibers. Thus, the targeting mechanism of CS in vivo is not affected by phosphorylation(s); i.e., sorting and segregation of CS appear to be independent of posttranslational phosphorylation(s).     

Spatial and temporal expression of acetylcholine receptor RNAs in innervated and denervated rat soleus muscle

In adult vertebrate skeletal muscle acetylcholine receptors are localized to the neuromuscular junction. Upon denervation, this distribution changes, with new receptors appearing in extrajunctional regions of the muscle fiber. The location of acetylcholine receptors in innervated or denervated muscle may result, in part, from the distribution of their RNAs. This was tested by assaying for receptor RNAs in junctional and extrajunctional regions of innervated and denervated rat soleus muscle using in situ hybridization and RNAase protection assays. These experiments showed alpha, beta, and delta subunit RNAs concentrated beneath the endplates of innervated muscle fibers. Following denervation, there was an unequal distribution of receptor RNAs along the muscle fiber, with highest levels occurring in extrajunctional regions near the endplate. These data are consistent with a nonuniform pattern of gene expression in adult skeletal muscle fibers.     

Effects of innervation on the distribution of acetylcholine receptors in regenerating skeletal muscles of adult chickens

In order to determine the roles of nerves in the formation of clusters of acetylcholine receptors (AChRs) during synaptogenesis, we examined the distribution of AChRs in denervated, nerve-transplanted (neurotized) muscles and in regenerated skeletal muscles of adult chickens by fluorescence microscopy using curaremimetic toxins. In the denervated muscles, many extrajunctional clusters developed at the periphery of some of the muscle nuclei of a single muscle fiber and continued to be present for up to 3 months. The AChR accumulations originally present at the neuromuscular junctions disappeared within 3 weeks. In the neurotized muscles, line-shaped AChR clusters developed at 4 days after transection of the original nerve, but no change in the distribution of AChRs had occurred even at 2 months after implantation of the foreign nerve. The line-shaped AChR clusters were found to be newly formed junctional clusters as they were associated with nerve terminals of similar shape and size. Some of both the line-shaped and extrajunctional clusters were formed at least partly by the redistribution of preexisting AChRs. Finally, based on the above observations, the regenerating muscle fibers in normal muscles and in denervated muscles were examined: The extrajunctional clusters appeared in both kinds of muscles at 2 weeks after injury. Afterward, during the innervation process, the line-shaped AChR clusters developed while the extrajunctional clusters disappeared in the innervated muscles. In contrast with this, in the absence of innervation, only the extrajunctional clusters continued to be present for up to 3 months. These results demonstrate clearly that the nerve not only induces the formation of junctional clusters at the contact site, but also prevents the formation of clusters at the extrajunctional region during synaptogenesis.     

Role of lipid rafts in agrin-elicited acetylcholine receptor clustering

Emerging concepts of membrane organization point to the compartmentalization of the plasma membrane into distinct lipid microdomains. This lateral segregation within cellular membranes is based on cholesterol-sphingolipid-enriched microdomains or lipid rafts which can move laterally and assemble into large-scale domains to create plasma membrane specialized cellular structures at specific cell locations. Such domains are likely involved in the genesis of the postsynaptic specialization at the neuromuscular junction, which requires the accumulation of acetylcholine receptors (AChRs), through activation of the muscle specific kinase MuSK by the neurotropic factor agrin and the reorganization of the actin cytoskeleton. We used C2C12 myotubes as a model system to investigate whether agrin-elicited AChR clustering correlated with lipid rafts. In a previous study, using two-photon Laurdan confocal imaging, we showed that agrin-induced AChR clusters corresponded to condensed membrane domains: the biophysical hallmark of lipid rafts [F. Stetzkowski-Marden, K. Gaus, M. Recouvreur, A. Cartaud, J. Cartaud, Agrin elicits membrane condensation at sites of acetylcholine receptor clusters in C2C12 myotubes, J. Lipid Res. 47 (2006) 2121-2133]. We further demonstrated that formation and stability of AChR clusters depend on cholesterol. We also reported that three different extraction procedures (Triton X-100, pH 11 or isotonic Ca++, Mg++ buffer) generated detergent resistant membranes (DRMs) with similar cholesterol/GM1 ganglioside content, which are enriched in several signalling postsynaptic components, notably AChR, the agrin receptor MuSK, rapsyn and syntrophin. Upon agrin engagement, actin and actin-nucleation factors such as Arp2/3 and N-WASP were transiently recovered within raft fractions suggesting that the activation by agrin can trigger actin polymerization. Taken together, the present data suggest that AChR clustering at the neuromuscular junction relies upon a mechanism of raft coalescence driven by agrin-elicited actin polymerization.     

Acetylcholine receptor degradation in adult rat diaphragms in organ culture and the effect of anti-acetylcholine receptor antibodies.

Acetylcholine receptor located at the neuromuscular synapse of normal innervated adult muscle fibers is extremely stable metabolically. We have studied the kinetics of receptor degradation in both normal innervated and denervated rat diaphragms in organ culture. These studies show that degradation of receptor-bound 125I-alpha-bungarotoxin is a valid measure of junctional receptor degradation. Degradation of junctional receptor is similar or identical to degradation of extrajunctional receptor in many ways: 1) both require energy, 2) both are inhibited by specific lysosomal protease inhibitors, 3) both are inhibited by treatment with colchicine, and 4) both are stimulated by treatment with anti-acetylcholine receptor antibodies. The one important distinction between degradation of junctional and extrajunctional receptor is a 10-fold difference in rate constant for the process.     

Development of the neuromuscular junction in the chick embryo: The number,distribution, and stability of acetylcholine receptors

The number, distribution, and stability of skeletal muscle acetylcholine receptors during development of the neuromuscular junction in the chick embryo were studied. The distribution and turnover of receptors labeled with ¹²⁵I-labeled α-bungarotoxin were determined by quantitative autoradiography on individual teased muscle fibers. Each posterior latissimus dorsi muscle fiber, which in the adult is singly innervated, has a high density of acetylcholine receptors at a single spot from embryonic Day 10 through hatching. The spots stain more intensely than elsewhere for acetylcholinesterase and are assumed to be end plates. The receptors at these spots are presumed to be junctional receptors. The junctional receptor density remains constant at 10⁴/μm² from embryonic Day 14 through adult life, although the area of the junction increases 40-fold. In contrast, the extrajunctional receptor density drops precipitously from 250/μm² on Day 14 to only 10/μm² on Day 19. This decrease in extrajunctional receptor density can be prevented by chronic paralysis with curare. The rate of autoradiographic grain loss from junctional and extrajunctional regions after a pulse injection of ¹²⁵I-labeled α-bungarotoxin indicates that both classes of embryonic receptors turn over at the same rate ($\text{t}{}^{\mathrm{<mtext>1</mtext><mtext>2</mtext>}}\text{? 30}\text{hr}$).     

Expression of muscle-gene-specific isozymes of phosphorylase and creatine kinase in innervated cultured human muscle

Isozymes of creatine kinase and glycogen phosphorylase are excellent markers of skeletal muscle maturation. In adult innervated muscle only the muscle-gene-specific isozymes are present, whereas aneurally cultured human muscle has predominantly the fetal pattern of isozymes. We have studied the isozyme pattern of human muscle cultured in monolayer and innervated by rat embryo spinal cord explants for 20-42 d. In this culture system, large groups of innervated muscle fibers close to the ventral part of the spinal cord explant continuously contracted. The contractions were reversibly blocked by 1 mM d-tubocurarine. In those innervated fibers, the total activity and the muscle-gene-specific isozymes of both enzymes increased significantly. The amount of muscle-gene-specific isozymes directly correlated with the duration of innervation. Control noninnervated muscle fibers from the same dishes as the innervated fibers remained biochemically immature. This study demonstrated that de novo innervation of human muscle cultured in monolayer exerts a time-related maturational influence that is not mediated by a diffusable neural factor.     

Myomesin and M protein: Differential expression in embryonic fibers during pectoral muscle development

By applying immunocytochemistry using monoclonal antibodies, we found that the myofibrillar M band of both presumptive type-I and -II fibers in the pectoralis major muscle of chickens contains two high-molecular-weight proteins, i.e., myomesin (Mr, 185,000) and M protein (Mr, 165,000), early in embryonic development (7 days in ovo), even though adult type-I fibers lack M protein. The developmental expression of M protein is unusual in that, from 10 to 14 days in ovo, it is gradually suppressed not only in presumptive type-I fibers but also in presumptive type-II fibers formed from primary-generation myotubes. This latter suppression is transient, as M protein is expressed in all adult type-II fibers derived from both the primary- and second-generation myotubes. Myomesin, on the other hand, is continuously expressed in all myotubes throughout development. This finding shows that myomesin and M protein expression is regulated independently in different myotube populations, and that the suppression of M protein in primary-generation myotubes accounts for the delayed accumulation of M protein during development, as previously revealed by biochemical analysis. Presumptive type-I fibers, which form in the deep portion of the muscle, become concentrated in a narrow band known as the red strip.     

Distribution of fiber types in embryonic chick limb muscles innervated by foreign motoneurons

The differentiation of distinct myotube fiber types in chick limb muscle development is coincident with innervation. The role of motoneurons in influencing fiber type differentiation was analyzed by causing chick hind limb muscles to be innervated by inappropriate motoneurons and then examining experimental muscles for changes in the distribution of myosin ATPase fiber types. Motoneuron innervation of limb muscles was altered by performing either limb shifts, limb reversals, or large spinal cord reversals on early neural tube or limb bud stage chick embryos. The distribution of fiber types was then analyzed in muscles from stage 36 (E10) to stage 45 (E20) embryos after processing hind limb sections for myosin ATPase histochemistry. In the majority of experimental muscles examined (267/312), the distribution of myosin ATPase fiber types was unaltered. In the remaining experimental muscles (14%), alterations in the distribution of myosin ATPase fiber types occurred, indicating that in some cases, foreign innervation may alter the developmental program of differentiating myotubes. The results suggest that myotubes differentiate myosin ATPase staining characteristics according to an intrinsic program and that these differentiating myotubes are selectively innervated by motoneurons of the appropriate type under most conditions including normal development. Under exceptional circumstances of motoneuron-muscle fiber type mismatch, embryonic motoneurons can alter fiber type expression.