STUDIES ON MUSCLE DIFFERENTIATION. V. ANTIGENICITIES OF MYOSIN AND OF ACTIN FROM FROG SKELETAL MUSCLES1

Both intact and denatured preparations of myosin and actin from frog skeletal muscles produced in rabbits antisera containing antibodies against authentic myosin and actin, respectively, though being contaminated with antibodies against other proteins. Antigenicity of our frog myosin as revealed in agar diffusion tests was indistinguishable from that of cardiac muscle myosin from the same species. Similarly, skeletal muscle myosins from other amphibians shared to a certain extent immunological characteristics with our frog myosin, but those from avian and mammalian materials did not. Similarity in antigenicity was also demonstrated among our skeletal muscle actin, cardiac muscle actin from the same species and skeletal muscle actin from the other anurans studied. However, skeletal muscle actin from an urodele could not clearly be correlated in its immunological properties with our frog actin, and those from avian and mammalian materials were antigenically different from our frog actin. Thus, the degree of antigenic similarity of these muscle proteins seemed to be correlated with the phylogenic relationship of the animals so far studied. The results also indicated that our antisera could only be applied to immuno-cytological and immuno-embryological studies of myosin and actin when the antisera absorbed with the corresponding antigen preparations were used as negative controls.     

Temporal resolution and sequential expression of muscle-specific genes revealed by in situ hybridization

The expression of muscle-specific mRNAs was analyzed directly within individual cells by in situ hybridization to chicken skeletal myoblasts undergoing differentiation in vitro. The probes detected mRNAs for sarcomeric myosin heavy chain (MHC) or the skeletal, cardiac, and beta isoforms of actin. Precise information as to the expression of these genes in individual cells was obtained and correlated directly with analyses of cell morphology and interactions, cell cycle stage, and immunofluorescence detection of the corresponding proteins. Results demonstrate that mRNAs for the two major muscle-specific proteins, myosin and actin, are not synchronously activated at the time of cell fusion. The mRNA for alpha-cardiac actin (CAct), known to be the predominant embryonic actin isoform in muscle, is expressed prior to cell fusion and prior to the expression of any isoform of muscle MHC mRNA. MHC mRNA accumulates rapidly immediately after fusion, whereas skeletal actin mRNA is expressed only in larger myofibers. Single cells expressing CAct mRNA have a characteristic short bipolar morphology, are in terminal G1, and do not contain detectable levels of the corresponding protein. In a pattern of expression reciprocal to that of CAct mRNA, beta-actin mRNA diminishes to low or undetectable levels in myofibers and in cells of the morphotype which expresses CAct mRNA. Finally, the intracellular distribution of mRNAs for different actin isoforms was compared using nonisotopic detection of isoform-specific oligonucleotide probes. This work illustrates a generally valuable approach to the analysis of cell differentiation and gene expression which directly integrates molecular, morphological, biochemical, and cell cycle information on individual cells.     

Myosin transducer mutations differentially affect motor function, myofibril structure, and the performance of skeletal and cardiac muscles

Striated muscle myosin is a multidomain ATP-dependent molecular motor. Alterations to various domains affect the chemomechanical properties of the motor, and they are associated with skeletal and cardiac myopathies. The myosin transducer domain is located near the nucleotide-binding site. Here, we helped define the role of the transducer by using an integrative approach to study how Drosophila melanogaster transducer mutations D45 and Mhc(5) affect myosin function and skeletal and cardiac muscle structure and performance. We found D45 (A261T) myosin has depressed ATPase activity and in vitro actin motility, whereas Mhc(5) (G200D) myosin has these properties enhanced. Depressed D45 myosin activity protects against age-associated dysfunction in metabolically demanding skeletal muscles. In contrast, enhanced Mhc(5) myosin function allows normal skeletal myofibril assembly, but it induces degradation of the myofibrillar apparatus, probably as a result of contractile disinhibition. Analysis of beating hearts demonstrates depressed motor function evokes a dilatory response, similar to that seen with vertebrate dilated cardiomyopathy myosin mutations, and it disrupts contractile rhythmicity. Enhanced myosin performance generates a phenotype apparently analogous to that of human restrictive cardiomyopathy, possibly indicating myosin-based origins for the disease. The D45 and Mhc(5) mutations illustrate the transducer's role in influencing the chemomechanical properties of myosin and produce unique pathologies in distinct muscles. Our data suggest Drosophila is a valuable system for identifying and modeling mutations analogous to those associated with specific human muscle disorders.     

Characterization of diverse forms of myosin heavy chain expressed in adult human skeletal muscle.

In an attempt to define myosin heavy chain (MHC) gene organization and expression in adult human skeletal muscle, we have isolated and characterized genomic sequences corresponding to different human sarcomeric MHC genes (1). In this report, we present the complete DNA sequence of two different adult human skeletal muscle MHC cDNA clones, one of which encodes the entire light meromyosin (LMM) segment of MHC and represents the longest described MHC cDNA sequence. Additionally, both clones provide new sequence data from a 228 amino acid segment of the MHC tail for which no protein or DNA sequence has been previously available. One clone encodes a "fast" form of skeletal muscle MHC while the other clone most closely resembles a MHC form described in rat cardiac ventricles. We show that the 3' untranslated region of skeletal MHC cDNAs are homologous from widely separated species as are cardiac MHC cDNAs. However, there is no homology between the 3' untranslated region of cardiac and skeletal muscle MHCs. Isotype-specific preservation of MHC 3' untranslated sequences during evolution suggests a functional role for these regions.     

Interaction between troponin and myosin enhances contractile activity of myosin in cardiac muscle

Ca(2+) signaling in striated muscle cells is critically dependent upon thin filament proteins tropomyosin (Tm) and troponin (Tn) to regulate mechanical output. Using in vitro measurements of contractility, we demonstrate that even in the absence of actin and Tm, human cardiac Tn (cTn) enhances heavy meromyosin MgATPase activity by up to 2.5-fold in solution. In addition, cTn without Tm significantly increases, or superactivates sliding speed of filamentous actin (F-actin) in skeletal motility assays by at least 12%, depending upon [cTn]. cTn alone enhances skeletal heavy meromyosin's MgATPase in a concentration-dependent manner and with sub-micromolar affinity. cTn-mediated increases in myosin ATPase may be the cause of superactivation of maximum Ca(2+)-activated regulated thin filament sliding speed in motility assays relative to unregulated skeletal F-actin. To specifically relate this classical superactivation to cardiac muscle, we demonstrate the same response using motility assays where only cardiac proteins were used, where regulated cardiac thin filament sliding speeds with cardiac myosin are >50% faster than unregulated cardiac F-actin. We additionally demonstrate that the COOH-terminal mobile domain of cTnI is not required for this interaction or functional enhancement of myosin activity. Our results provide strong evidence that the interaction between cTn and myosin is responsible for enhancement of cross-bridge kinetics when myosin binds in the vicinity of Tn on thin filaments. These data imply a novel and functionally significant molecular interaction that may provide new insights into Ca(2+) activation in cardiac muscle cells.     

Skeletal muscle myofibrillar protein metabolism in heart failure: relationship to immune activation and functional capacity

Chronic heart failure is characterized by changes in skeletal muscle that contribute to physical disability. Most studies to date have investigated defects in skeletal muscle oxidative capacity. In contrast, less is known about how heart failure affects myofibrillar protein metabolism. Thus we examined the effect of heart failure on skeletal muscle myofibrillar protein metabolism, with a specific emphasis on changes in myosin heavy chain (MHC) protein content, synthesis, and isoform distribution in 10 patients with heart failure (63 +/- 3 yr) and 11 controls (70 +/- 3 yr). In addition, we examined the relationship of MHC protein metabolism to inflammatory markers and physical function. Although MHC and actin protein content did not differ between groups, MHC protein content decreased with increasing disease severity in heart failure patients (r = -0.748, P < 0.02), whereas actin protein content was not related to disease severity. No difference in MHC protein synthesis was found between groups, and MHC protein synthesis rates were not related to disease severity. There were, however, relationships between C-reactive protein and both MHC protein synthesis (r = -0.442, P = 0.05) and the ratio of MHC to mixed muscle protein synthesis (r = -0.493, P < 0.03). Heart failure patients showed reduced relative amounts of MHC I (P < 0.05) and a trend toward increased MHC IIx (P = 0.06). In regression analyses, decreased MHC protein content was related to decreased exercise capacity and muscle strength in heart failure patients. Our results demonstrate that heart failure affects both the quantity and isoform distribution of skeletal muscle MHC protein. The fact that MHC protein content was related to both exercise capacity and muscle strength further suggests that quantitative alterations in MHC protein may have functional significance.     

Native connectin from porcine cardiac muscle

Native connectin was isolated from porcine cardiac muscle using the method developed for the preparation of native connectin from chicken breast muscle (Kimura et al. (1984) J. Biochem. 96, 1947-1950). It was not necessary to keep cardiac muscle at 0 degrees C before preparation: the proteolysis of alpha-connectin to beta-connectin proceeded during the preparation of myofibrils. Cardiac connectin showed almost the same properties as those of skeletal muscle connectin: mobility in SDS gel electrophoresis, filamentous structure under an electron microscope, circular dichroism spectra, UV absorption spectra, and amino acid composition. Porcine cardiac connectin cross-reacted with antiserum against chicken breast muscle connectin as revealed by an immunoblot method. Immunoelectron microscopical observations revealed an abundance of connectin antigenic sites around the A-I junction area of cardiac myofibrils. Cardiac connectin also interacted with myosin and actin filaments at low ionic strengths to form aggregates. The extent of interaction was somewhat weaker in the case of cardiac connectin than skeletal muscle connectin, regardless of the origin of myosin and actin (porcine cardiac and rabbit skeletal muscles). In conclusion, cardiac connectin is very similar, but not identical to skeletal muscle connectin.     

Kinetic analysis of the slow skeletal myosin MHC-1 isoform from bovine masseter muscle

Several heavy chain isoforms of class II myosins are found in muscle fibres and show a large variety of different mechanical activities. Fast myosins (myosin heavy chain (MHC)-II-2) contract at higher velocities than slow myosins (MHC-II-1, also known as beta-myosin) and it has been well established that ADP binding to actomyosin is much tighter for MHC-II-1 than for MHC-II-2. Recently, we reported several other differences between MHC-II isoforms 1 and 2 of the rabbit. Isoform II-1 unlike II-2 gave biphasic dissociation of actomyosin by ATP, the ATP-cleavage step was significantly slower for MHC-II-1 and the slow isoforms showed the presence of multiple actomyosin-ADP complexes. These results are in contrast to published data on MHC-II-1 from bovine left ventricle muscle, which was more similar to the fast skeletal isoform. Bovine MHC-II-1 is the predominant isoform expressed in both the ventricular myocardium and slow skeletal muscle fibres such as the masseter and is an important source of reference work for cardiac muscle physiology. This work examines and extends the kinetics of bovine MHC-II-1. We confirm the primary findings from the work on rabbit soleus MHC-II-1. Of significance is that we show that the affinity of ADP for bovine masseter myosin in the absence of actin (represented by the dissociation constant K(D)) is weaker than originally described for bovine cardiac myosin and thus the thermodynamic coupling between ADP and actin binding to myosin is much smaller (K(AD)/K(D) approximately 5 instead of K(AD)/K(D) approximately 50). This may indicate a distinct type of mechanochemical coupling for this group of myosin motors. We also find that the ATP-hydrolysis rate is much slower for bovine MHC-II-1 (19 s(-1)) than reported previously (138 s(-1)). We discuss how this work fits into a broader characterisation of myosin motors from across the myosin family.     

Effects of exercise on plasma myosin heavy chain fragments and MRI of skeletal muscle.

The effects of a single series of high-force eccentric contractions involving the quadriceps muscle group (single leg) on plasma concentrations of muscle proteins were examined as a function of time, in the context of measurements of torque production and magnetic resonance imaging (MRI) of the involved muscle groups. Plasma concentrations of slow-twitch skeletal (cardiac beta-type) myosin heavy chain (MHC) fragments, myoglobin, creatine kinase (CK), and cardiac troponin T were measured in blood samples of six healthy male volunteers before and 2 h after 70 eccentric contractions of the quadriceps femoris muscle. Screenings were conducted 1, 2, 3, 6, 9, and 13 days later. To visualize muscle injury, MRI of the loaded and unloaded thighs was performed 3, 6, and 9 days after the eccentric exercise bout. Force generation of the knee extensors was monitored on a dynamometer (Cybex II+) parallel to blood sampling. Exercise resulted in a biphasic myoglobin release profile, delayed CK and MHC peaks. Increased MHC fragment concentrations of slow skeletal muscle myosin occurred in late samples of all participants, which indicated a degradation of slow skeletal muscle myosin. Because cardiac troponin T was within the normal range in all samples, which excluded a protein release from the heart (cardiac beta-type MHC), this finding provides evidence for an injury of slow-twitch skeletal muscle fibers in response to eccentric contractions. Muscle action revealed delayed reversible increases in MRI signal intensities on T2-weighted images of the loaded vastus intermedius and deep parts of the vastus lateralis. We attributed MRI signal changes due to edema in part to slow skeletal muscle fiber injury.(ABSTRACT TRUNCATED AT 250 WORDS)     

Cooperative interactions between the contractile proteins of cardiac and skeletal muscle.

The calcium activation of the ATPase (ATP phosphohydrolase, EC 3.6.1.3) activity of cardiac actomyosin reconstituted from bovine cardiac myosin and a complex of actin-tropomyosin-troponin extracted from bovine cardiac muscle at 37 degrees C was studied and compared with similar proteins from rabbit fast skeletal muscle. The proteins of the actin complex were identified by polyacrylamide gel electrophoresis in sodium dodecyl sulfate. Half-maximal activation of the cardiac actomyosin was seen at a calcium concentration of 1.2 +/- 0.002 (S.E. of mean) muM. A hybridized reconstituted actomyosin made with cardiac myosin and the actin-tropomyosin-troponin complex extracted from rabbit skeletal muscle was also activated by calcium but the half-maximal value was shifted to 0.65 +/- 0.02 (S.E. of mean) muM Ca2+. Homologous rabbit skeletal actomyosin showed half-maximal activation at 0.90 +/- 0.01 (S.E. of mean) muM Ca2+ and the value for a hybridized actomyosin made with rabbit skeletal myosin and the actin-complex from cardiac muscle was found at 1.4 +/- 0.03 (S.E. of mean) muM Ca2+ concentration. Kinetic analysis of the Ca2+ activated ATPase activity of reconstituted bovine cardiac actomyosin indicated some degree of cooperativity with respect to calcium. Double reciprocal plots of reconstituted actomyosins made with bovine cardiac actin complex were curvilinear and significantly different than those of reconstituted actomyosins made with the rabbit fast skeletal actin complex. The Ca2+-dependent cooperativity was of a mixed type as determined from Hill plots for homologous reconstituted bovine cardiac and rabbit fast skeletal actomyosin. The results show that cooperative interactions in reconstituted actomyosins were greater when the actin-tropomyosin-troponin complex was derived from cardiac than skeletal muscle.     

Some motile properties of fast characean myosin

We improved a motility assay system by using an affinity-purified antibody against the C-terminal globular domain of characean myosin. This improvement allowed us to study the sensitivity to ionic strength or the processivity of characean myosin. The sliding velocity of actin filaments on a characean myosin-coated surface was unaffected by ionic strength. This property is unlike that of skeletal or smooth muscle myosin and suggests that the binding manner of characean myosin to actin is different from that in other muscle myosins. The sliding velocity decreased when the MgADP concentration was raised. The extent of inhibition by MgADP on the motile activity of characean myosin was almost the same as in skeletal muscle or cardiac myosin. The number of sliding filaments on the characean myosin-coated surface decreased drastically with a decrease in the motor density. The motor density required to produce a successful movement of actin filament was about 200 molecules/microm(2). These results suggest that the characean myosin is not a processive motor protein.     

Chaperone-mediated reversible inhibition of the sarcomeric myosin power stroke

Molecular chaperones are required for successful folding and assembly of sarcomeric myosin in skeletal and cardiac muscle. Here, we show that the chaperone UNC-45B inhibits the actin translocation function of myosin. Further, we show that Hsp90, another chaperone involved in sarcomere development, allows the myosin to resume actin translocation. These previously unknown activities may play a key role in sarcomere development, preventing untimely myosin powerstrokes from disrupting the precise alignment of the sarcomere until it has formed completely.     

PEVK domain of titin: an entropic spring with actin-binding properties

The PEVK domain of the giant muscle protein titin is a proline-rich sequence with unknown secondary/tertiary structure. Here we compared the force-extension behavior of cloned cardiac PEVK titin measured by single-molecule atomic force spectroscopy with the extensibility of the PEVK domain measured in intact cardiac muscle sarcomeres. The analysis revealed that cardiac PEVK titin acts as an entropic spring with the properties of a random coil exhibiting mechanical conformations of different flexibility. Since in situ, titin is in close proximity to the thin filaments, we also studied whether the PEVK domain of cardiac or skeletal titin may interact with actin filaments. Interaction was indeed found in the in vitro motility assay, in which recombinant PEVK titin constructs slowed down the sliding velocity of actin filaments over myosin. Skeletal PEVK titin affected the actin sliding to a lesser degree than cardiac PEVK titin. The cardiac PEVK effect was partially suppressed by physiological Ca(2+) concentrations, whereas the skeletal PEVK effect was independent of [Ca(2+)]. Cosedimentation assays confirmed the Ca(2+)-modulated actin-binding propensity of cardiac PEVK titin, but did not detect interaction between actin and skeletal PEVK titin. In myofibrils, the relatively weak actin-PEVK interaction gives rise to a viscous force component opposing filament sliding. Thus, the PEVK domain contributes not only to the extensibility of the sarcomere, but also affects contractile properties.     

Mapping of O‐linked β‐N‐acetylglucosamine modification sites in key contractile proteins of rat skeletal muscle

O‐linked β‐N‐acetylglucosamine (O‐GlcNAc) is a widespread modification of serine/threonine residues of nucleocytoplasmic proteins. Recently, several key contractile proteins in rat skeletal muscle (i.e., myosin heavy and light chains and actin) were identified as O‐GlcNAc modified. Moreover, it was demonstrated that O‐GlcNAc moieties involved in contractile protein interactions could modulate Ca²⁺ activation parameters of contraction. In order to better understand how O‐GlcNAc can modulate the contractile activity of muscle fibers, we decided to identify the sites of O‐GlcNAc modification in purified contractile protein homogenates. Using an MS‐based method that relies on mild β‐elimination followed by Michael addition of DTT (BEMAD), we determined the localization of one O‐GlcNAc site in the subdomain four of actin and four O‐GlcNAc sites in the light meromyosin region of myosin heavy chains (MHC). According to previous reports concerning the role of these regions, our data suggest that O‐GlcNAc sites might modulate the actin–tropomyosin interaction, and be involved in MHC polymerization or interactions between MHC and other contractile proteins. Thus, the results suggest that this PTM might be involved in protein–protein interactions but could also modulate the contractile properties of skeletal muscle.     

Smooth muscle and skeletal muscle myosins produce similar unitary forces and displacements in the laser trap.

Purified smooth muscle myosin in the in vitro motility assay propels actin filaments at 1/10 the velocity, yet produces 3-4 times more force than skeletal muscle myosin. At the level of a single myosin molecule, these differences in force and actin filament velocity may be reflected in the size and duration of single motion and force-generating events, or in the kinetics of the cross-bridge cycle. Specifically, an increase in either unitary force or duty cycle may explain the enhanced force-generating capacity of smooth muscle myosin. Similarly, an increase in attached time or decrease in unitary displacement may explain the reduced actin filament velocity of smooth muscle myosin. To discriminate between these possibilities, we used a laser trap to measure unitary forces and displacements from single smooth and skeletal muscle myosin molecules. We analyzed our data using mean-variance analysis, which does not rely on scoring individual events by eye, and emphasizes periods in the data with constant properties. Both myosins demonstrated multiple but similar event populations with discrete peaks at approximately +11 and -11 nm in displacement, and 1.5 and 3.5 pN in force. Mean attached times for smooth muscle myosin were longer than for skeletal-muscle myosin. These results explain much of the difference in actin filament velocity between these myosins, and suggest that an increased duty cycle is responsible for the enhanced force-generating capacity of smooth over skeletal-muscle myosin.     

Fast skeletal muscle isoforms exhibit the highest incorporation level into myofibrils and stress fibers among members of myosin alkali light chain isoform family

Isoproteins of myosin alkali light chain (LC) were co-expressed in cultured chicken cardiomyocytes and fibroblasts and their incorporation levels into myofibrils and stress fibers were compared among members of the LC isoform family. In order to distinguish each isoform from the other, cDNAs of LC isoforms were tagged with different epitopes. Expressed LCs were detected with antibodies to the tags and their distribution was analyzed by confocal microscopy. In cardiomyocytes, the incorporation level of LC into myofibrils was shown to increase in the order from nonmuscle isoform (LC3nm), to slow skeletal muscle isoform (LC1sa), to slow skeletal/ventricular muscle isoform (LC1sb), and to fast skeletal muscle isoforms (LC1f and LC3f). Thus, the hierarchal order of the LC affinity for the cardiac myosin heavy chain (MHC) is identical to that obtained in the rat (Komiyama et al., 1996. J. Cell Sci., 109: 2089-2099), suggesting that this order may be common for taxonomic animal classes. In fibroblasts, the affinity of LC for the nonmuscle MHC in stress fibers was found to increase in the order from LC3nm, to LC1sb, to LC1sa, and to LC1f and LC3f. This order for the nonmuscle MHC is partly different from that for the cardiac MHC. This indicates that the order of the affinity of LC isoproteins for MHC varies depending on the MHC isoform. Further, for both the cardiac and nonmuscle MHCs, the fast skeletal muscle LCs exhibited the highest affinity. This suggests that the fast skeletal muscle LCs may be evolved isoforms possessing the ability to associate tightly with a variety of MHC isoforms.     

Synthetic Medium for Cellulolytic Anaerobe,Ruminococcus albus

Some physico-chemical and heat-induced gelling properties of myosin heavy chains (MHCs) from rabbit skeletal muscle were studied. MHCs were found to be almost devoid of ATPase activity, possibly because of the absence of myosin light chains. MHCs formed precipitates below and ionic strength of 0.3, and above this ionic strength MHCs became soluble. On heating to 65°C at low concentrations of the protein, MHCs lost their solubility and formed aggregates even at high ionic conditions. Partially irreversible change in the conformation of MHC (from helical to random coil) also occurred in heated MHCs.

The heat-induced gel strength of MHCs was found to be almost equal to that of intact myosin molecules with identical protein concentrations. This suggests that the gelling potential of myosin is solely confined in MHC. Although myosin light chains do not seem to contribute to the gelling power of myosin, possibly they provide some stability to the gel if the pH is varied above the optimum value, i.e., 6.0. The addition of actin to a MHC system weakened the gel strength of the latter proportionally to the quantity of added actin.     

Contents of myofibrillar proteins in cardiac, skeletal, and smooth muscles

The in situ contents of myosin, actin, alpha-actinin, tropomyosin, troponin, desmin were estimated in dog cardiac, rabbit skeletal, and chicken smooth muscles. Whole muscle tissues were dissolved with 8 M guanidine hydrochloride and subjected to two-dimensional gel electrophoresis, which is a nonequilibrium pH gradient electrophoresis (Murakami, U. & Uchida, K. (1984) J. Biochem. 95, 1577-1584) with some modification. The amount of protein in a spot on a slab gel was determined by quantification of the extracted dye. Dye binding capacity of individual myofibrillar proteins was determined by using the purified protein. Myosin contents were 82 +/- 7 pmol/mg wet weight in cardiac muscle, 105 +/- 10 pmol/mg wet weight in skeletal muscle, and 45 +/- 4 pmol/mg wet weight in smooth muscle. Actin contents were 339 +/- 15 pmol/mg wet weight in cardiac muscle, 625 +/- 27 pmol/mg wet weight in skeletal muscle, and 742 +/- 13 pmol/mg wet weight in smooth muscle. The subunit stoichiometry of myosin in the three types of muscles was two heavy chains and four light chains, and there was one light chain 2 for every heavy chain. The molar ratio of actin to tropomyosin was 7/1 in the three types of muscles. Striking differences were seen in the molar ratio of myosin to actin: 1.0/4.1 in cardiac muscle, 1.0/6.0 in skeletal muscle, and 1.0/16.5 in smooth muscle.     

Actin and myosin expression during development of cardiac muscle from cultured embryonal carcinoma cells

P19 embryonal carcinoma cells are multipotential stem cells that differentiate into striated muscle as well as some other cell types when aggregated and exposed to dimethyl sulfoxide (DMSO). Immunofluorescence experiments using monospecific antibodies indicated that the majority of muscle cells were mononucleate and contained four myosin isoforms normally found in cardiac muscle; atrial and ventricular myosin heavy chains, ventricular myosin light chain 1, and atrial myosin light chain 2. Northern blot analysis of RNA isolated from differentiating cultures indicated that cardiac actin and skeletal actin mRNAs were expressed at similar levels and with identical kinetics during the differentiation of P19-derived myocytes. These results demonstrate that most of the P19-derived myocytes are of the cardiac type and suggest that they closely resemble the cells of the early embryonic myocardium.