Neural regulation of mammalian fast and slow muscle myosins: an electrophoretic analysis.

Mammalian nerves to fast and slow muscles have the remarkable property of changing the speed of contraction of muscles following cross-reinnervation. The biochemical basis of speed transformation is the change in myosin in ATPase activity. This paper provides electrophoretic evidence for structural changes in myosin from cross-reinnervated muscles. A method is described for the separation of intact fast and slow muscle myosins by polyacrylamide gel electrophoresis. This method utilizes the fact that ATP and its analogs prevent the formation of myosin polymers in low ionic strength buffers. In this system, normal fast muscle myosin has a higher electrophoretic mobility than slow muscle myosin. Normal rat soleus myosin has a major slow and a minor fast component due to two populations of muscle fibers. The same muscle cross-reinnervated by a fast muscle nerve shows only the fast component, The normal, homogeneous fast extensor digitorum longus muscle has only the electrophoretically fast myosin, but following cross-reinnervation it shows both fast and slow components. These results suggest that mammalian motor nerves can induce or suppress the expression of genes that code for fast and slow skeletal muscle myosins.     

去神经对快,慢肌纤维肌球蛋白ATPase影响的组织化学观察

本文用组织化学方法观察了豚鼠比目鱼肌(SOL)和腓骨第三肌(PT)在去神经后其快、慢纤维肌球蛋白ATPase特性的变化。在正常肌肉中Ⅰ型(慢)纤维和Ⅱ型(快)纤维分别具有酸和碱稳定ATPase活性。慢纤维在去神经后出现了碱稳定ATPase活性,而快纤维则无明显变化。结果表明,只有慢纤维的肌球蛋白ATPase特性才与神经支配有关。     

Diversity in expression of myosin heavy chain isoforms and M-band proteins in rat muscle spindles

The composition of adult rat soleus muscle spindles, with respect to myosin heavy chain isoforms and M-band proteins, was studied by light-microscope immunohistochemistry. Serial sections were labelled with antibodies against slow tonic, slow twitch, fast twitch and neonatal myosin isoforms as well as against myomesin, M-protein and the MM form of creatine kinase. Intrafusal fiber types were distinguished according to the pattern of ATPase activity following acid and alkaline preincubations. Nuclear bag1 fibers were always strongly stained throughout with anti-slow tonic myosin, were positive for anti-slow twitch myosin towards and in the C-region but were unstained with anti-fast twitch and anti-neonatal myosins. The staining of nuclear bag2 fibers was in general highly variable. However, they were most often strongly stained by anti-slow tonic myosin in the A-region and gradually lost this reactivity towards the poles, whereas a positive reaction with anti-slow twitch myosins was found along the whole fiber. Regional staining variability with anti-neonatal and anti-fast myosins was apparent, often with decreasing intensity towards the polar regions. Nuclear chain fibers showed strong transient reactivity with anti-slow tonic myosin in the equatorial region, did not react with anti-slow twitch and were always evenly stained by anti-fast twitch and anti-neonatal myosins. All three intrafusal fiber types were stained with anti-myomesin. Nuclear bag1 fibers lacked staining for M-protein, whereas bag2 fibers displayed intermediate staining, with regional variability, often increasing in reactivity towards the polar regions. Chain fibers were always strongly stained by anti-M-protein. The MM form of creatine kinase was present in all three fiber types, but bag1 fibers were less reactive and clear striations were not observed, in contrast to bag2 and chain fibers. Out of 38 cross sectioned spindles two were found to have an atypical fiber composition (lack of chain fibers) and a rather diverse staining pattern for the different antibodies tested. Taken together, the data show that in adult rat soleus, slow tonic and neonatal myosin heavy chain isoforms are only expressed in the muscle spindle fibers and that each intrafusal fiber type has a unique, although variable, composition of myosin heavy chain isoforms and M-band proteins. We propose that both motor and sensory innervation might be the determining factors regulating the variable expression of myosin heavy chain isoforms and M-band proteins in intrafusal fibers of rat muscle spindles.     

Diversity in expression of myosin heavy chain isoforms and M-band proteins in rat muscle spindles

Summary The composition of adult rat soleus muscle spindles, with respect to myosin heavy chain isoforms and M-band proteins, was studied by light-microscope immunohistochemistry. Serial sections were labelled with antibodies against slow tonic, slow twitch, fast twitch and neonatal myosin isoforms as well as against myomesin, M-protein and the MM form of creatine kinase. Intrafusal fiber types were distinguished according to the pattern of ATPase activity following acid and alkaline preincubations.Nuclear bag₁ fibers were always strongly stained throughout with anti-slow tonic myosin, were positive for anti-slow twitch myosin towards and in the C-region but were unstained with anti-fast twitch and anti-neonatal myosins. The staining of nuclear bag₂ fibers was in general highly variable. However, they were most often strongly stained by anti-slow tonic myosin in the A-region and gradually lost this reactivity towards the poles, whereas a positive reaction with anti-slow twitch myosins was found along the whole fiber. Regional staining variability with antineonatal and anti-fast myosins was apparent, often with decreasing intensity towards the polar regions. Nuclear chain fibers showed strong transient reactivity with anti-slow tonic myosin in the equatorial region, did not react with anti-slow twitch and were always evenly stained by anti-fast twitch and anti-neonatal myosins. All three intrafusal fiber types were stained with anti-myomesin. Nuclear bag₁ fibers lacked staining for M-protein, whereas bag₂ fibers displayed intermediate staining, with regional variability, often increasing in reactivity towards the polar regions. Chain fibers were always strongly stained by anti-M-protein. The MM form of creatine kinase was present in all three fiber types, but bag₁ fibers were less reactive and clear striations were not observed, in contrast to bag₂ and chain fibers. Out of 38 cross sectioned spindles two were found to have an atypical fiber composition, (lack of chain fibers) and a rather diverse staining pattern for the different antibodies tested.Taken together, the data show that in adult rat solcus, slow tonic and neonatal myosin heavy, chain isoforms are only expressed in the muscle spindle fibers and that each intrafusal fiber type has a unique, although variable, composition of myosin heavy chain isoforms and M-band proteins. We propose that both motor and sensory innervation might be the determining factors regulating the variable expression of myosin heavy chain isoforms and M-band proteins in intrafusal fibers of rat muscle spindles.     

A combined histochemical and immunohistochemical study on the dynamics of fast-to-slow fiber transformation in chronically stimulated rabbit muscle

Summary Chronically stimulated fast-twitch muscles of the rabbit were histochemically and immunohistochemically analyzed in serial cross sections (1) for percentages of fiber types, and (2) for the presence of myosin heavy chain isoforms during fast-to-slow transformation. By four weeks of stimulation the number of type-I fibers had increased more than fourfold, while only about 6% of the original IIB fibers remained. Type-IC and -IIC fibers transiently rose to 20% of the total fiber population. After 16 weeks, the number of type-I fibers had increased to 42%. With prolonged stimulation fewer fibers reacted with antibodies against embryonic and neonatal myosins and more with the antibody against slow myosin. The reaction for embryonic myosin was most often detected in the C fibers (IC, IIC). Immunohistochemical subtypes were observed for each fiber type in the stimulated muscles. The greatest number was seen in type-IIC fibers, which, in addition to their reaction for fast/neonatal and slow myosins, might also react with the antibodies against neonatal/embryonic and embryonic myosins. These findings indicated that the transforming fibers temporarily expressed myosin heavy chain isoforms normally not detectable in adult skeletal muscle. Myotubes reacted strongly with the antibodies against fast/neonatal and embryonic myosins, and some of them also with the antibody against slow myosin. Thus, it appears that under the influence of the low frequency stimulus pattern some of the newly formed myotubes developed into type-I fibers.     

Distribution of myosin isoenzymes among skeletal muscle fiber types.

Using an immunocytochemical approach, we have demonstrated a preferential distribution of myosin isoenzymes with respect to the pattern of fiber types in skeletal muscles of the rat. In an earlier study, we had shown that fluorescein-labeled antibody against "white" myosin from the chicken pectoralis stained all the white, intermediate and about half the red fibers of the rat diaphragm, a fast-twitch muscle (Gauthier and Lowey, 1977). We have now extended this study to include antibodies prepared against the "head" (S1) and "rod" portions of myosin, as well as the alkali- and 5,5'dithiobis (2-nitrobenzoic acid) (DTNB)-light chains. Antibodies capable of distinguishing between alkali 1 and alkali 2 type myosin were also used to localize these isoenzymes in the same fast muscle. We observed, by both direct and indirect immunofluorescence, that the same fibers which had reacted previously with antibodies against white myosin reacted with antibodies to the proteolytic subfragments and to the low molecular-weight subunits of myosin. These results confirm our earlier conclusion that the myosins of the reactive fibers in rat skeletal muscle are sufficiently similar to share antigenic determinants. The homology, furthermore, is not confined to a limited region of the myosin molecule, but includes the head and rod portions and all classes of light chains. Despite the similarities, some differences exist in the protein compositions of these fibers: antibodies to S1 did not stain the reactive (fast) red fiber as strongly as they did the white and intermediate fibers. Non-uniform staining was also observed with antibodies specific for A2 myosin; the fast red fiber again showed weaker fluorescence than did the other reactive fibers. These results could indicate a variable distribution of myosin isoenzymes according to their alkali-light chain composition among fiber types. Alternatively, there may exist yet another myosin isoenzyme which is localized in the fast red fiber. Those red fibers which did not react with any of the antibodies to pectoralis myosin, did react strongly with an antibody against myosin isolated from the anterior latissimus dorsi (ALD), a slow red muscle of the chicken. The myosin in these fibers (slow red fibers) is, therefore, distinct from the other myosin isoenzymes. In the rat soleus, a slow-twitch muscle, the majority of the fibers reacted only with antibody against ALD myosin. A minority, however, reacted with antiboddies to pectoralis as well as ALD myosin, which indicates that both fast and slow myosin can coexist within the same fiber of a normal adult muscle. These immunocytochemical studies have emphasized that a wide range of isoenzymes may contribute to the characteristic physiological properties of individual fiber types in a mixed muscle.     

Analysis of myosin light and heavy chain types in single human skeletal muscle fibers

In this study, myosin types in human skeletal muscle fibers were investigated with electrophoretic techniques. Single fibers were dissected out of lyophilized surgical biopsies and typed by staining for myofibrillar ATPase after preincubation in acid or alkaline buffers. After 14C-labelling of the fiber proteins in vitro by reductive methylation, the myosin light chain pattern was analysed on two-dimensional gels and the myosin heavy chains were investigated by one-dimensional peptide mapping. Surprisingly, human type I fibers, which contained only the slow heavy chain, were found to contain variable amounts of fast myosin light chains in addition to the two slow light chains LC1s and LC2s. The majority of the type I fibers in normal human muscle showed the pattern LC1s, LC2s and LC1f. Further evidence for the existence in human muscle of a hybrid myosin composed of a slow heavy chain with fast and slow light chains comes from the analysis of purified human myosin in the native state by pyrophosphate gel electrophoresis. With this method, a single band corresponding to slow myosin was obtained; this slow myosin had the light chain composition LC1s, LC2s and LC1f. Type IIA and IIB fibers, on the other hand, revealed identical light chain patterns consisting of only the fast light chains LC1f, LC2f and LC3f but were found to have different myosin havy chains. On the basis of the results presented, we suggest that the histochemical ATPase normally used for fibre typing is determined by the myosin heavy chain type (and not by the light chains). Thus, in normal human muscle a number of 'hybrid' myosins were found to occur, namely two extreme forms of fast myosins which have the same light chains but different heavy chains (IIA and IIB) and a continuum of slow forms consisting of the same heavy chain and slow light chains with a variable fast light chain composition. This is consistent with the different physiological roles these fibers are thought to have in muscle contraction.     

Use of type-specific antimyosins to demonstrate the transformation of individual fibers in chronically stimulated rabbit fast muscles

Continuous stimulation of a rabbit fast muscle at 10 Hz changes its physiological and biochemical parameters to those of a slow muscle. These transformations include the replacement of myosin of one type by myosin of another type. Two hypotheses could explain the cellular basis of these changes. First, if fibers were permanently programmed to be fast or slow, but not both, a change from one muscle type to another would involve atrophy of one fiber type accompanied by de novo appearance of the other type. Alternatively, preexisting muscle fibers could be changing from the expression of one set of genes to the expression of another. Fluorescein-labeled antibodies against fast (AF) and slow (AS) muscle myosins of rabbits have been prepared by procedures originally applied to chicken muscle. In the unstimulated fast peroneus longus muscle, most fibers stained only with AF; a small percentage stained only with AS; and no fibers stained with both antibodies. In stimulated muscles, most fibers stained with both AF and AS; with increasing time of stimulation, there was a progressive decrease in staining intensity with AF and a progressive increase in staining intensity with AS within the same fibers. These results are consistent with a theory that individual preexisting muscle fibers can actually switch from the synthesis of fast myosin to the synthesis of slow myosin.     

Myosin types in cultured muscle cells

Fluorescent antibodies against fast skeletal, slow skeletal, and ventricular myosins were applied to muscle cultures from embryonic pectoralis and ventricular myocadium of the chicken. A number of spindle-shaped mononucleated cells, presumably myoblasts, and all myotubes present in skeletal muscle cultures were labeled by all three antimyosin antisera. In contrast, in cultures from ventricular myocardium all muscle cells were labeled by anti-ventricular myosin, whereas only part of them were stained by anti-slow skeletal myosin and rare cells reacted with anti-fast skeletal myosin. The findings indicate that myosin(s) present in cultured embryonic skeletal muscle cells contains antigenic determinants similar to those present in adult fast skeletal, slow skeletal, and ventricular myosins.     

Heterogeneity of fast-oxidative muscle fibers of chicken demonstrated by anti-myosin monoclonal antibodies

Summary The fiber type composition of two fast muscles of the chicken, namely, adductor superficialis (AS) and pectoralis major (PM) was examined by the histochemical myosin ATPase staining and immunochemical techniques using monoclonal antibodies (McAbs). Two new McAbs produced against the myosin of the anterior latissimus dorsi (ALD) muscle of the chicken and named ALD-122 and ALD-83 were characterized to be specific for myosin heavy chain (MHC) and for myosin light chain-1 respectively. They were used in conjunction with previously reported McAbs specific for slow MHC (ALD-47), fast MHC (MF-14) and fast light chain-2 (MF-5). By the histochemical ATPase test most muscle fibers of AS and PM muscles reacted as IIA and IIB respectively. By immunofluorescent staining with the anti-MHC McAbs, ALD-122, and MF-14, the fibers of AS, muscle showed remarkable heterogeneity whereas PM muscle fibers reacted, uniformly. Differences in the myosin light chain composition of AS and PM muscles were also found by SDS-gel electrophoresis and immunoblot analysis with the anti-light chain McAb, ALD-83. The study clearly indicated that the histochemically homogenous (type IIA) AS muscle is composed of several subpopulations of fibers which differ in their myosin composition and that this heterogeneity of the muscle is not simply due to presence of variable amounts of slow myosin in its fibers.     

Comparative studies on 5,5'-dithiobis-(2-nitrobenzioc acid)-treated myosins.

Myosins prepared from chicken and rabbit fast and slow muscles were treated with 5,5'-dithiobis-(2-nitrobenzoic acid) (Nbs2). About half of the thiol groups of the fast muslce myosins reacted with Nbs 2, but in slow muscle myosins, only about 10-20% of the thiol groups reacted. This treatment removed 50-60% of the L2 components, Nbs2 light chain, from fast muscle myosins, but did not result in specific dissociation of the light chains in slow myscle myosins. The treatment sometimes released L4 component from chicken muscle myosins instead of L2 component. The changes of myosin ATPase [EC 3.6.1.3] activities caused by this treatment did not correlate with the release of Nbs2 light chain, but were dependent upon the species, chicken or rabbit.     

Curare-induced transformation of myosin pattern in developing skeletal muscle fibers

The effects of neuromuscular block on the pattern of distribution of myosin isozymes in developing skeletal muscle fibers was examined by immunocytochemistry. The homogeneous population of fibers in the anterior latissimus dorsi (ALD) of the 18-day chick embryo was converted by curare to a mosaic of at least two categories of fibers. Normally all fibers in this slow muscle reacted with antibodies against slow myosin (anti-ALD). They also reacted with an antibody specific for the alkali 1 light chain (anti-delta 1) but not the alkali 2 light chain (anti-delta 2) of fast myosin. After treatment with curare, which inhibits neuronal cell death and increases the number of axonal endings, ALD muscle fibers continued to react with anti-delta 1, but many now reacted with anti-delta 2 as well. The same fibers failed to react with anti-ALD. From this it can be concluded that the myosin in this population was converted to a type not normally present. The changes, therefore, are not merely a result of the preferential loss of a slow type of fiber, nor are they a result of delayed maturation. In contrast, curare had no apparent effect on the fast posterior latissimus dorsi (PLD). As in the normal muscle at 18 days, all fibers reacted strongly with anti-delta 1 and to variable degrees with anti-delta 2, and very few fibers reacted with anti-ALD. Our observations suggest that the dual response to antibodies against fast and slow myosin during development is not a necessary consequence of multiple axon terminals. We present evidence that curare induces the expression of a different myosin in the embryonic ALD, and we suggest that the selective transformation of the fiber population may be a manifestation of a change in composition of the motoneuron pool.     

Morphological and biochemical correlates of skeletal muscle contractility in the cat. II. Physiological and biochemical studies.

Isometric twitch characteristics and biochemical parameters of isolated myosin and sarcoplasmic reticulum have been compared in three cat hind limb muscles. The fast twitch caudofemoralis and the slow twitch soleus are almost pure muscles as judged from histochemical studies. Isolated myosin from the caudofemoralis is not only 2- to 3-fold higher in its ATPase activities than that of the soleus, but also in non-dissociated forms has greater electrophoretic mobility than the soleus myosin. Purified myosins from fast muscles as well as soleus exhibited three light chains upon electrophoresis. However, the intact non-solubilized myosins differed in electrophoretic mobilities. The sarcoplasmic reticulum fraction isolated from caudfemoralis exhibits faster rates of Ca++ binding and uptake than soleus, and when fit to a two component model, the caudofemoralis SR exhibits a higher amount of a fast binding site than does soleus SR, features reflected in differences in the relaxation time of the two muscles. In contrast, the fast twitch tibialis anterior has been shown to be a gradient of fiber types and its isometric twitch may be separated by selective nerve stimulation, into a fast and a slow twitch component. Our findings that myosin fractions, as well as sarcoplasmic reticulum fractions isolated from these two components differ with respect to their biochemical characteristics add support to the possibility of a dual function in this muscle.     

Purification and characterization of myosins from human and rabbit skeletal muscles by using specific monoclonal antibodies

By using immunoaffinity column chromatography slow (I) and fast (IIA, IIB) myosins were isolated from human (vastus lateralis) and rabbit (tibialis anterior, psoas and conoidal bundle) skeletal muscles. The peptide pattern revealed that slow (I) and fast (IIA, IIB) myosin heavy chains are quite distinct, as are those from pure slow (conoidal bundle) and fast (psoas) rabbit skeletal muscles. Unlike Billeter et al. (1981) the authors observed that fast human myosins were always associated with a small amount of slow myosin light chains. The fast myosins (IIA, IIB) from rabbit tibialis anterior muscle did not appear very distinct and contained only fast myosin light chains. These myosins were different from the IIB myosin from the psoas muscle. Ten per cent of the fibres revealed histochemically as fast IIA also reacted with an anti-slow myosin antibody. The classical histochemical techniques appear inadequate to demonstrate the existing differences among fibre types, but the monoclonal antibodies hold promise.     

Chronic denervation of rat hemidiaphragm: maintenance of fiber heterogeneity with associated increasing uniformity of myosin isoforms

During several months of denervation, rat mixed muscles lose slow myosin, though with variability among animals. Immunocytochemical studies showed that all the denervated fibers of the hemidiaphragm reacted with anti-fast myosin, while many reacted with anti-slow myosin as well. This has left open the question as to whether multiple forms of myosin co-exist within individual fibers or a unique, possibly embryonic, myosin is present, which shares epitopes with fast and slow myosins. Furthermore, one can ask if the reappearance of embryonic myosin in chronically denervated muscle is related both to its re-expression in the pre-existing fibers and to cell regeneration. To answer these questions we studied the myosin heavy chains from individual fibers of the denervated hemidiaphragm by SDS PAGE and morphologically searched for regenerative events in the long term denervated muscle. 3 mo after denervation the severely atrophic fibers of the hemidiaphragm showed either fast or a mixture of fast and slow myosin heavy chains. Structural analysis of proteins sequentially extracted from muscle cryostat sections showed that slow myosin was still present 16 mo after denervation, in spite of the loss of the selective distribution of fast and slow features. Therefore muscle fibers can express adult fast myosin not only when denervated during their differentiation but also after the slow program has been expressed for a long time. Light and electron microscopy showed that the long-term denervated muscle maintained a steady-state atrophy for the rat's life span. Some of the morphological features indicate that aneural regeneration events continuously occur and significantly contribute to the increasing uniformity of the myosin gene expression in long-term denervated diaphragm.     

Expression of myosin heavy chain isoforms in developing rat muscle spindles

Summary The development of muscle spindles, with respect to the expression of myosin heavy chain isoforms was studied in rat hind limbs from 17 days of gestation up to seven days after birth. Serial cross-sections were labelled with antibodies against slow tonic, slow twitch and neonatal isomyosins, myomesin, laminin and neurofilament protein.At 17–18 days of gestation, a small population of primary myotubes expressing slow tonic myosin were identified as the earliest spindle primordia. These myotubes also expressed slow twitch and, to a lesser extent, neonatal myosin. At 19–20 days of gestation a second myotube became apparent; this staining strongly with anti-neonatal myosin. A day later this secondary myotube acquired reactivity to anti-slow tonic and anti-slow twitch myosins. By birth, a third myotube was present; this staining strongly with anti-neonatal myosin but otherwise unreactive with the other antibodies against myosin heavy chains. Three days after birth a fourth myotube, with identical reactivity to the third one, became apparent. Regional variation in the expression of isomyosins, which was present since birth in the two nuclear bag fibers was further enhanced: the nuclear bag₂ staining strongly with anti-slow tonic and antineonatal in the equatorial region and with decreasing intensity towards the poles, whilst with anti-slow twitch the stainability was low in the equatorial and high in the polar region. The nuclear bag₁ fiber showed a homogeneous staining: high with anti-slow tonic, moderate with anti-neonatal, and displayed stainability to antislow twitch myosin in the polar regions only. No regional variation was found along the chain fiber/myotube. At seven days after birth, the pattern of reactivity was similar to that found in the adult spindles, except for the bag₁ fiber which still expressed neonatal myosin.We show that slow tonic myosin is expressed from early development and it is a reliable marker of developing bag fibers. We suggest that muscle spindles are formed from special cell lineages of which the primary generation myotubes expressing slow tonic myosin represent the primordium of muscle spindles.     

Myosin types in human skeletal muscle fibers

By combining enzyme histochemistry for fiber typing with immunohistochemistry for slow and fast myosin a correlation between fiber type and myosin type was sought in human skeletal muscle. Fiber typing was done by staining for myofibrillar ATPases after preincubation at discriminating pH values. Myosin types were discriminated using type specific anti-rabbit myosin antibodies shown to cross-react with human myosin and were visualized by a protein A-peroxidase method. Type I fibers were shown to contain slow myosin only, type IIA and IIB fibers fast myosin only, and type IIC fibers both myosins in various proportions. When muscle biopsies from well-trained athletes were investigated essentially the same staining pattern was observed. However, rarely occurring type I fibers with high glycolytic activity were detected containing additional small amounts of fast myosin and occasional type IIA fibers had small amounts of slow myosin. Based on the observation of various fiber types in which slow and fast myosin coexist we propose a dynamic continuum of fibers encompassing all fiber types.     

Myosin types in human skeletal muscle fibers

Summary By combining enzyme histochemistry for fiber typing with immunohistochemistry for slow and fast myosin a correlation between fiber type and myosin type was sought in human skeletal muscle. Fiber typing was done by staining for myofibrillar ATPases after preincubation at discriminating pH values. Myosin types were discriminated using type specific anti-rabbit myosin antibodies shown to cross-react with human myosin and were visualized by a protein A-peroxidase method. Type I fibers were shown to contain slow myosin only, type IIA and IIB fibers fast myosin only, and type IIC fibers both myosins in various proportions. When muscle biopsies from well-trained athletes were investigated essentially the same staining pattern was observed. However, rarely occurring type I fibers with high glycolytic activity were detected containing additional small amounts of fast myosin and occasional type IIA fibers had small amounts of slow myosin. Based on the observation of various fiber types in which slow and fast myosin coexist we propose a dynamic continuum of fibers encompassing all fiber types.     

Expression of myosin heavy chain isoforms in developing rat muscle spindles

The development of muscle spindles, with respect to the expression of myosin heavy chain isoforms was studied in rat hind limbs from 17 days of gestation up to seven days after birth. Serial cross-sections were labelled with antibodies against slow tonic, slow twitch and neonatal isomyosins, myomesin, laminin and neurofilament protein. At 17-18 days of gestation, a small population of primary myotubes expressing slow tonic myosin were identified as the earliest spindle primordia. These myotubes also expressed slow twitch and, to a lesser extent, neonatal myosin. At 19-20 days of gestation a second myotube became apparent; this staining strongly with anti-neonatal myosin. A day later this secondary myotube acquired reactivity to anti-slow tonic and anti-slow twitch myosins. By birth, a third myotube was present; this staining strongly with anti-neonatal myosin but otherwise unreactive with the other antibodies against myosin heavy chains. Three days after birth a fourth myotube, with identical reactivity to the third one, became apparent. Regional variation in the expression of isomyosins, which was present since birth in the two nuclear bag fibers was further enhanced: the nuclear bag staining strongly with anti-slow tonic and antineonatal in the equatorial region and with decreasing intensity towards the poles, whilst with anti-slow twitch the stainability was low in the equatorial and high in the polar region. The nuclear bag fiber showed a homogeneous staining: high with anti-slow tonic, moderate with anti-neonatal, and displayed stainability to anti-slow twitch myosin in the polar regions only. No regional variation was found along the chain fiber/myotube.(ABSTRACT TRUNCATED AT 250 WORDS)