Expression of smooth muscle and nonmuscle myosin heavy chains in cultured vascular smooth muscle cells

We explored the hypothesis that discrepancies in the literature concerning the nature of myosin expression in cultured smooth muscle cells are due to the appearance of a new form of myosin heavy chain (MHC) in vitro. Previously, we used a very porous sodium dodecyl sulfate gel electrophoresis system to detect two MHCs in intact smooth muscles (SM1 and SM2) which differ by less than 2% in molecular weight (Rovner, A. S., Thompson, M. M., and Murphy, R. A. (1986) Am. J. Physiol. 250, C861-C870). Myosin-containing homogenates of rat aorta cells in primary culture were electrophoresed on this gel system, and Western blots were performed using smooth muscle-specific and nonmuscle-specific myosin antibodies. Subconfluent, rapidly proliferating cultures contained a form of heavy chain not found in rat aorta cells in vivo (NM) with electrophoretic mobility and antigenicity identical to the single unique heavy chain seen in nonmuscle cells. Moreover, these cultures expressed almost none of the smooth muscle heavy chains. In contrast, postconfluent growth-arrested cultures expressed increased levels of the two smooth muscle heavy chains, along with large amounts of NM. Analysis of cultures pulsed with [35S] methionine indicated that subconfluent cells were synthesizing almost exclusively NM, whereas postconfluent cells synthesized SM1 and SM2 as well as larger amounts of NM. Similar patterns of MHC content and synthesis were found in subconfluent and postconfluent passaged cells. These results show that cultured vascular smooth muscle cells undergo differential expression of smooth muscle- and nonmuscle-specific MHC forms with changes in their growth state, which appear to parallel changes in expression of the smooth muscle and nonmuscle forms of actin (Owens, G. K., Loeb, A., Gordon, D., and Thompson, M. M. (1986) J. Cell Biol. 102, 343-352). The reappearance of the smooth muscle MHCs in postconfluent cells suggests that density-related growth arrest promotes cytodifferentiation, but the continued expression of the nonmuscle MHC form in these smooth muscle cells indicates that other factors are required to induce the fully differentiated state while in culture.     

Myosin heavy-chain isoforms in adult and developing rabbit vascular smooth muscle

Two monoclonal antibodies specific for smooth muscle myosin (designated SM-E7 and SM-A9) and one monoclonal anti-(human platelet myosin) antibody (designated NM-G2) have been used to study myosin heavy chain composition of smooth muscle cells in adult and in developing rabbit aorta. Sodium dodecyl sulfate/polyacrylamide gel electrophoresis and Western blotting experiments revealed that adult aortic muscle consisted of two myosin heavy chains (MCH) of smooth muscle type, named MHC-1 (205 kDa), and MHC-2 (200 kDa). In the fetal/neonatal stage of development, vascular smooth muscle was found to contain only MHC-1 but not MHC-2. Non-muscle myosin heavy chain, which showed the same electrophoretic mobility as the slower migrating MHC, was expressed in an inverse manner with respect to MHC-2, i.e. it was detectable only in the early stages of development. The distinct pattern of smooth and non-muscle myosin isoform expression during development may be related to the different functional properties of smooth muscle cells during vascular myogenesis.     

The heavy chain of smooth muscle myosin is phosphorylated in aorta cells

The 204-kDa smooth muscle myosin heavy chain (MHC) from rat aorta smooth muscle cells was found to be phosphorylated following isolation of myosin from strips of intact aorta as well as from primary cultures of aorta cells. Two-dimensional maps of the tryptic peptides revealed that the phosphate was confined to only three peptides and gave a similar pattern for the MHC isolated from intact aorta strips and cultured cells. This map was quite different from the phosphopeptide map found for the 196-kDa MHC of nonmuscle myosin isolated from the same cell culture. Smooth muscle MHC purified from primary cell cultures was found to contain approximately 0.7 mol of phosphate/mol of MHC while the nonmuscle MHC contained approximately 0.8 mol of phosphate/mol of MHC. These observations raise the possibility of an additional regulatory mechanism in smooth muscle operating via MHC phosphorylation.     

Quantitative and qualitative heterogeneity in smooth muscle myosin heavy chains

Previous studies have shown that smooth muscle myosin consists of two heavy chains (MHCs) of unequal molecular weight; however, it is not clear whether there are intermuscle, inter- and intraspecies differences in the MHCs. The purpose of these experiments was to quantitatively and qualitatively compare MHCs in different smooth muscles. Extracts of bovine aorta (BAo), dog saphenous vein (dSV) and femoral artery (dFA), and rat aorta (rAo), femoral artery (rFA), carotid artery (rCA), ileum (rGI) and uterus (rUt) were electrophoresed on 5% polyacrylamide-1% SDS gels. All tissues exhibited two MHCs with molecular weights of 207,000 (MHC1) and 204,000 (MHC2) daltons. In all cases the proportion of total MHC made up by MHC1 was greater than that by MHC2. Based on their relative proportions (MHC1:MHC2), the tissues fell into one of three groups: (1) 55:45 - rAo, rCA, dFA; (2) 60:40 - dSV, BAo, rGI; and (3) 65:35 - rUt, rFA. Group 1 differed significantly from group 3 in the proportion of each MHC. One dimensional peptide maps indicated that BAo, dSV and dFA were similar while subtle differences existed between rUt and rAo. Differences between rUt and rAo were also observed in their cross-reactivity to a monoclonal antibody to smooth muscle MHC, confirming the differences seen on peptide maps. These results indicate that there are intertissue and inter- and intraspecies differences in smooth muscle MHCs. The significance of these differences to muscle function remains to be determined.     

Myosin heavy-chain isoforms in human smooth muscle

The myosin heavy-chain composition of human smooth muscle has been investigated by sodium dodecyl sulfate/polyacrylamide gel electrophoresis, enzyme immunoassay, and enzyme-immunoblotting procedures. A polyclonal and a monoclonal antibody specific for smooth muscle myosin heavy chains were used in this study. The two antibodies were unreactive with sarcomeric myosin heavy chains and with platelet myosin heavy chain on enzyme immunoassay and immunoblots, and stained smooth muscle cells but not non-muscle cells in cryosections and cultures processed for indirect immunofluorescence. Two myosin heavy-chain isoforms, designated MHC-1 and MHC-2 (205 kDa and 200 kDa, respectively) were reactive with both antibodies on immunoblots of pyrophosphate extracts from different smooth muscles (arteries, veins, intestinal wall, myometrium) electrophoresed in 4% polyacrylamide gels. In the pulmonary artery, a third myosin heavy-chain isoform (MHC-3, 190 kDa) electrophoretically and antigenically distinguishable from human platelet myosin heavy chain, was specifically recognized by the monoclonal antibody. Analysis of muscle samples, directly solubilized in a sodium dodecyl sulfate solution, and degradation experiments performed on pyrophosphate extracts ruled out the possibility that MHC-3 is a proteolytic artefact. Polypeptides of identical electrophoretic mobility were also present in the other smooth muscle preparations, but were unreactive with this antibody. The presence of three myosin heavy-chain isoforms in the pulmonary artery may be related to the unique physiological properties displayed by the smooth muscle of this artery.     

Smooth muscle myosin is composed of homodimeric heavy chains.

Vertebrate smooth muscle myosin heavy chains (MHCs) exist as two isoforms with molecular masses of 204 and 200 kDa (MHC204 and MHC200) that are generated from a single gene by alternative splicing of mRNA (Nagai, R., Kuro-o, M., Babij, P., and Periasamy, M. (1989) J. Biol. Chem. 264, 9734-9737). A dimer of two MHCs associated with two pairs of myosin light chains forms a functional myosin molecule. To investigate the isoform composition of the MHCs in native myosin, antibodies specific for MHC204 were generated and used to immunoprecipitate purified bovine aortic smooth muscle myosin from a solution containing equal amounts of each isoform. MHC204 quantitatively removed from this mixture was completely free of MHC200. Immunoprecipitation of the supernatant with an antiserum that recognizes both isoforms equally well revealed that only MHC200 remained. We conclude that only homodimers of MHC204 and MHC200 exist under these conditions. A method is described for the purification of enzymatically active MHC204 and myosin on a protein G-agarose high performance liquid chromatography column containing immobilized MHC204 antibodies. We show, using an in vitro motility assay, that the movement of actin filaments by myosin containing 204-kDa heavy chains (0.435 +/- 0.115 microns/s) was not significantly different from that of myosin containing 200-kDa heavy chains (0.361 +/- 0.078 microns/s) or from myosin containing equal amounts of each heavy chain isoform (0.347 +/- 0.082 microns/s).     

Monoclonal antibody assessment of tissue- and species-specific myosin light chain kinase isozymes

Monoclonal antibodies raised against chicken gizzard smooth muscle myosin light chain kinase were used for immunological and structural studies of this enzyme. Epitope mapping of trypsin-digested chicken gizzard enzyme showed that MM-1, 2, 3, 4, 5, 6, and 7 bind to 65 kDa (trypsin-digested) and 60 kDa (chymotrypsin-digested) fragments which contain the catalytic domain of the kinase. Kinetic analysis demonstrated that MM-7 inhibited kinase activity competitively with respect to ATP and noncompetitively with respect to myosin light chain, thereby indicating that MM-7 binds at or near the ATP binding site of the enzyme. Immunoblot analysis revealed that all these antibodies (MM-1 to 12) reacted with the enzyme (130 kDa) from intestinal and vascular smooth muscles, whereas 5 (MM-1, 3, 4, 6, and 9) or 3 (MM-1, 3, and 4) of 12 antibodies did not cross-react with chicken cardiac muscle or with blood platelet myosin light chain kinase (130 kDa), respectively. None of these antibodies showed cross-reactivity against skeletal muscle myosin light chain kinase. As for mammalian species, MM-11 and 12 reacted with myosin light chain kinase of vascular smooth muscle (140 kDa) and MM-11 cross-reacted with the enzyme (140 kDa) from cardiac muscle of rat and rabbit. These data suggest the existence of at least 4 subspecies of myosin light chain kinase in chicken tissues and the heterogeneity of tissue- and species-specific isozyme forms.     

Evidence for inserted sequences in the head region of nonmuscle myosin specific to the nervous system. Cloning of the cDNA encoding the myosin heavy chain-B isoform of vertebrate nonmuscle myosin.

The complete amino acid sequence of a vertebrate nonmuscle myosin heavy chain-B isoform (MHC-B, 1976 amino acids, 229 kDa) has been deduced by using cDNA clones from chicken brain libraries. The chicken nonmuscle MHC-B shows overall similarity in primary structure to other MHCs in the areas contributing to the ATP-binding site and actin-binding site. Similar to other nonsarcomeric MHC IIs, there is a short uncoiled tail sequence at the carboxyl terminus of the molecule. It is in the uncoiled tail sequence that the greatest number of differences in amino acids sequence between MHC-A and B were found, which allowed generation of isoform-specific antibodies. These antibodies were used to determine the relative content of MHC-A and MHC-B in various tissues. During the cloning of the cDNA encoding chicken brain MHC-B, we found a 63-nucleotide insertion encoding 21 amino acids located in the head region of the MHC near to the actin-binding site and a 30 nucleotide insertion encoding 10 amino acids near to the ATP-binding site. Analysis using S-1 nuclease showed that both inserts are expressed in a tissue-dependent manner; mRNA containing the inserts is present in tissues of the nervous system, but is absent from other non-muscle cells, which contain the noninserted isoform of MHC-B. Similar inserts were found in corresponding positions in human cerebellar mRNA. Antibodies raised against a peptide synthesized based on the 21 amino acid insert found in chickens recognize a MHC isoform in the same tissues that are enriched for the mRNA. These insertions appear to be a mechanism for generating additional MHC-B isoforms specific to the nervous system.     

Embryonic chicken gizzard: smooth muscle and non-muscle myosin isoforms

Antibodies to smooth muscle and non-muscle myosin allow the development of smooth muscle and its capillary system in the embryonic chicken gizzard to be followed by immunofluorescent techniques. Although smooth muscle development proceeds in a serosal to luminal direction, angiogenetic cell clusters develop independently at the luminal side close to the epithelial layer, and the presumptive capillaries invade the developing muscle in a luminal to serosal direction. The smooth muscle and non-muscle myosin heavy chains in this avian system cannot be separated by SDS polyacrylamide gel electrophoresis and do not show isoform specificity in immunoblotting, unlike the system found in mammals. Only two myosin heavy chains with M_r of 200 and 196 kDa were separable and considerable immunological cross-reactivity was found between the denatured myosin isoform heavy chains.     

Identification of the serine residue phosphorylated by protein kinase C in vertebrate nonmuscle myosin heavy chains

Two-dimensional mapping of the tryptic phosphopeptides generated following in vitro protein kinase C phosphorylation of the myosin heavy chain isolated from human platelets and chicken intestinal epithelial cells shows a single radioactive peptide. These peptides were found to comigrate, suggesting that they were identical, and amino acid sequence analysis of the human platelet tryptic peptide yielded the sequence -Glu-Val-Ser-Ser(PO4)-Leu-Lys-. Inspection of the amino acid sequence for the chicken intestinal epithelial cell myosin heavy chain (196 kDa) derived from cDNA cloning showed that this peptide was identical with a tryptic peptide present near the carboxyl terminal of the predicted alpha-helix of the myosin rod. Although other vertebrate nonmuscle myosin heavy chains retain neighboring amino acid sequences as well as the serine residue phosphorylated by protein kinase C, this residue is notably absent in all vertebrate smooth muscle myosin heavy chains (both 204 and 200 kDa) sequenced to date.     

Cloning of the cDNA encoding a neuronal myosin heavy chain from mammalian brain and its differential expression within the central nervous system.

The complete amino acid sequence of a neuronal myosin heavy chain (MHC) from mammalian brain (1999 amino acids, 230 kDa) has been deduced by sequencing cDNA clones isolated from a rat brain cDNA library. The library was screened using an affinity-purified polyclonal antibody that had been raised against myosin purified from a neuronally-derived cell line (Neuro-2A). Restriction digests of genomic DNA from Neuro-2A cells and rat brain are consistent with an identity of the sequenced isoform from these two sources. RNA blot analysis demonstrates this myosin to exhibit differential expression within the cerebral cortex and spinal cord. No expression was observed in liver, kidney, heart, spleen or skeletal muscle, or even within other regions of the brain. The sequence of this neuronal MHC is compared with those of other non-muscle MHCs, to which it shows an overall similarity of structure, especially with respect to conserved regions within the head (ATP binding site, actin binding site, reactive thiols) and the presence of an alpha-helical coiled-coil tail that can be arranged as 28-residue repeating units plus four skip residues. A unique non-helical tailpiece composed of 72 amino acid residues marks the C-terminus of this neuronal myosin isoform.     

Production of specific antibodies to contractile proteins and their use in immunofluorescence microscopy. III. Antiobody against human uterine smooth muscle myosin.

The preparation of highly purified myosin from surgical specimen of human uterine muscle is described. Antibodies were raised in rabbits against this immunogen. In immunodiffusion, they react with uterine and chicken gizzard muscle myosin, no reaction is observed between uterine myosin and the anti-chicken-gizzard- myosin. In immunofluorescence, anti-uterine-myosin stains smooth muscle in the contractile and "modulated" state and non-muscle cells such as fibroblasts, platelets and endothelium of various species. Thus, these antibodies contrast anti-gizzard-myosin, which has previously been shown to be specific for contractile state muscle cells. We therefore conclude that the uterine myosin preparation consists of two immunogens, the one being associated with cell contractility and the other, termed cytoplasmic myosin, with motility and mitosis. The latter is indistinguishable from the myosin present in non-muscle cells and can be absorbed specifically with actomyosin from blood platelets.     

Alternative splicing and cycling kinetics of myosin change during hypertrophy of human smooth muscle cells

We investigated in vivo expression of myosin heavy chain (MHC) isoforms, 17 kDa myosin light chain (MLC₁₇), and phosphorylation of the 20 kDa MLC (MLC₂₀) as well as mechanical performance of chemically skinned fibers of normal and hypertrophied smooth muscle (SM) of human myometrium. According to their immunological reactivity, we identified three MHC isoenzymes in the human myometrium: two SM-MHC (SM1 with 204 kDa and SM2 with 200 kDa), and one non-muscle specific MHC (NM with 196 kDa). No cross-reactivity was detected with an antibody raised against a peptide corresponding to a seven amino acid insert at the 25K/50K junction of the myosin head (a-25K/50K) in both normal and hypertrophied myometrium. In contrast, SM-MHC of human myomatous tissue strongly reacted with a-25K/50K. Expression of SM1/SM2/NM (%) in normal myometrium was 31.7/34.7/33.6 and 35.1/40.9/24 in hypertrophied myometrium. The increased SM2 and decreased NM expression in the hypertrophied state was statistically significant (P < 0.05). MHC isoform distribution in myomatous tissue was similar to normal myometrium (35.3/35.3/29.4). In vivo expression of MLC_17a increased from 25.5% in normal to 44.2% in hypertrophied (P < 0.001) myometrium. Phosphorylation levels of MLC₂₀ upon maximal Ca²⁰-calmodulin activation of skinned myometrial fibers were the same in normal and hypertrophied myometrial fibers. Maximal force of isometric contraction of skinned fibers (pCa 4.5, slack-length) was 2.85 mN/mm² and 5.6 mN/mm² in the normal and hypertrophied state, respectively (P < 0.001). Apparent maximal shortening velocity (Vmax_app, extrapolated from the force-velocity relation) of myometrium rose from 0.13 muscle length s ¹ (ML/s) in normal to 0.24 ML/s in hypertrophied fibers (P < 0.001). J. Cell. Biochem, 64:171–181. © 1997 Wiley-Liss, Inc.     

The smooth muscle 132 kDa cyclic GMP-dependent protein kinase substrate is not myosin light chain kinase or caldesmon.

Atrial natriuretic peptide (ANP) stimulates the phosphorylation of three cyclic GMP-dependent protein kinase substrate proteins of 225, 132, and 11 kDa (P225, P132 and P11 respectively) in the particulate fraction of cultured rat aortic smooth muscle cells [Sarcevic, Brookes, Martin, Kemp & Robinson (1989) J. Biol. Chem. 264, 20648-20654]. Vrolix, Raeymaekers, Wuytack, Hofmann & Casteels [(1988) Biochem. J. 255, 855-863] have reported the presence of a 130 kDa cyclic GMP-dependent protein kinase substrate protein in the membrane fraction of pig aorta or stomach, and suggested that it may be myosin light chain kinase (MLCK). The aim of the present study was to determine whether P132 from rat aorta was MLCK or caldesmon. Although P132 co-migrates with purified chicken gizzard MLCK on SDS/polyacrylamide gels, it is distinct from rat aortic MLCK. Partially purified MLCK from rat aorta migrated as a 145 kDa protein on SDS/polyacrylamide gels. Immunoblotting the partially purified rat aortic MLCK with antibody to bovine tracheal MLCK identified rat aortic MLCK (145 kDa) and a corresponding 145 kDa protein in the particulate fraction of cultured rat aortic smooth muscle cells, but did not detect the 132 kDa protein. Phosphopeptide maps of purified rat aortic MLCK prepared by digestion with Staphylococcus aureus V8 protease were distinct from those of P132. P132 was not caldesmon, since antibodies to caldesmon cross-reacted with 136 and 76 kDa proteins in the particulate fraction of rat aortic cells, but not with P132. Furthermore, caldesmon was partially extracted from the particulate into the soluble fraction by heating at 90 degrees C, whereas P132 was not. These results demonstrate that the ANP-responsive cyclic GMP-dependent protein kinase substrate of 132 kDa from rat aortic smooth muscle cells is not MLCK or caldesmon.     

Developmental changes in expression of contractile and cytoskeletal proteins in human aortic smooth muscle

To describe phenotypic changes of human aortic smooth muscle cells (SMCs), proportion of smooth muscle and nonmuscle variants of actin, myosin heavy chains (MHCs), vinculin, and caldesmon, during prenatal and several months of postnatal development was determined. In aortic SMCs from 9-10-week-old fetus, both nonmuscle and smooth muscle-specific variants of all four proteins were present, however, the nonmuscle forms were more abundant. During development, a shift towards the expression of muscle-specific variants was observed, although the time course of changes in protein variant content was not similar for all the proteins studied. By the 24th week of gestation, fractional content of alpha-smooth muscle actin and smooth muscle MHCs was rather close to that in the mature SMCs, and comprised approximately 80 and 90%, respectively, of the levels characteristic of SMCs from adult aortic media. On the contrary, fractional ratio of meta-vinculin and 150-kDa caldesmon was still rather low in the aorta from the 24-week-old fetus, did not increase in a 2-month-old child aorta, and did not reach the level characteristic of mature SMCs even in the 6-month-old child aorta. Thus changes in alpha-smooth muscle actin and smooth muscle MHC fractional content occur mainly during the prenatal period of development, before the 24th week of gestation; while meta-vinculin and the 150-kDa caldesmon proportion increases mainly in the postnatal period, during several months after birth. In the "Discussion," phenotypes of SMCs from developing aorta were compared to those from different layers of the adult aortic wall.     

A chicken embryo-specific myosin light chain (L23) appears to be identical with one of brain myosin light chains

A novel embryo-specific myosin light chain of 23 kDa molecular weight (L23) was found previously in embryonic chicken skeletal, cardiac, and smooth muscles (Takano-Ohmuro et al. (1985) J. Cell Biol. 100, 2025-2030). When we examined myosin in embryonic and adult brain by two-dimensional electrophoresis, 23 kDa myosin light chain present in brain (Burridge & Bray (1975) J. Mol. Biol. 99, 1-14) comigrated with L23. Two monoclonal antibodies, EL-64 and MT-185d, were applied to clarify the identity of the brain 23 kDa myosin light chain and the chicken embryonic muscle L23. The two antibodies recognize different antigenic determinants in the L23 molecule; the former antibody is specific for L23, whereas the latter recognizes the sequence common to fast skeletal muscle myosin light chains 1 and 3, and also L23. The immunoblots combined with two-dimensional gel electrophoresis showed that both EL-64 and MT-185d can bind to the brain 23 kDa myosin light chain as well as the chicken embryonic muscle L23. These results indicate that chicken brain and chicken embryonic muscles contain a common myosin light chain of 23 kDa molecular weight.     

Myosin isoforms and cell heterogeneity in vascular smooth muscle. I. Developing and adult bovine aorta

Monoclonal anti-smooth muscle (SM-E7, SM-F11, and BF-48) and anti-nonmuscle (NM-A9 and NM-G2) myosin antibodies, Western blotting, and immunocytochemical procedures were used to study myosin isoform composition and distribution in the smooth muscle (SM) cells of bovine aorta differentiating in vivo and in vitro. Two myosin heavy chain (MHC) isoforms were identified by SM-E7 in adult aorta: SM-MHC-1 (Mr = 205 kDa) and SM-MHC-2 (Mr = 200 kDa), respectively. When tested with the SM-F11 antibody, SM-MHC-2 isoform showed distinct antigenic properties compared to SM-MHC-1. Two bands of 205 and 200 kDa were also present in the aortic SM tissue from 3-month-old fetus and were equally recognized by the BF-48 antibody. The 200-kDa SM myosin isoform was labeled by SM-F11 but not by SM-E7, thus indicating the existence of a fetal-specific SM-MHC-2 isoform. At the cellular level, both developing and adult bovine aortic tissues showed the existence of distinct patterns of myosin isoform expression. Three or even more aortic cell populations are differently distributed in areas which appear as (1) a network of interconnecting sheet-like or compact tissue (early fetus) and (2) enriched of collagenous-elastic or muscular tissue (adult animal). In addition, the SM-MHC-2 isoform of the fetal type appears to be uniquely distributed in cultured SM cells grown in vitro from adult bovine aortic explants. Our data indicate that in bovine aorta (1) MHC isoform expression is developmentally regulated and (2) the distribution of myosin isoforms is heterogenous both among and within aortic cells. These findings may be related to the distinct physiological properties displayed by SM during vascular myogenesis.     

Immunohistochemical studies on the distribution of cellular myosin II isoforms in brain and aorta.

The distribution of nonmuscle myosin isoforms in brain and aorta was studied by using polyclonal antibodies against two synthetic peptides selected from a region near the carboxyl terminus of bovine brain (peptide IIB) and human macrophage (peptide IIA) myosin. Immunoblots of brain homogenates and purified myosin showed two major bands stained by anti-peptide IIB (MIIB1 and MIIB2) and a minor band stained by anti-peptide IIA (MIIA2). Polyclonal anti-human platelet myosin antibodies did not react with MIIB isoforms. In cryosections from bovine, rat, and mouse brains, anti-peptide IIB stained most neuronal cells. In bovine cryosections, glial staining was also observed. In contrast, anti-peptide IIA and anti-platelet myosin antibodies primarily stained blood vessels. In bovine aorta, the anti-peptide antibodies recognized four bands, MIIB3, MIIB4, MIIA1, and MIIA2. Only MIIA2 was recognized by anti-human platelet myosin antibodies. In bovine aorta cryosections, anti-peptide IIB stained smooth muscle cells in tunica intima and tunica media but did not stain endothelial cells. Anti-peptide IIA stained smooth muscle cells in the tunica media, and endothelial cells of vaso vasorum but not of aorta. Only polyclonal anti-platelet myosin antibodies stained the endothelial cells of aorta tunica intima. These results indicate that multiple isoforms of cellular myosins exist in mammals, that these isoforms are expressed in a cell specific manner, and that the major myosin isoforms isolated from whole brain originate from neurons and, at least in bovine brain, from glia, but not from blood vessels.