THE ANATOMICAL RECORD 201293-302 (1981) Structure and Function of the Murine Muscle-Tendon Junction JOHN A. TROTTER, KAREN CORBETT, AND BARRY P. AVNER Department of Anatomy (JA.T.,K.C.) and Department of Pharmacology ( B P A J , University of New Mexico School of Medicine, Albuquerque, NM 87131 ABSTRACT The muscle-tendon junctions of the extensor carpi radialis long u s and brevis muscles from adult Balb C BaileyN mice have been examined tensiometrically and ultrastructurally following removal of cellular membrane and soluble cytoplasm by exposure to nonionic detergent. As judged by the ability of the extracted muscle to generate tension upon exposure to ATP and to transmit the generated tension to the tendon, detergent extraction leaves the muscle-tendon junction functionally intact. Electron microscopic analysis of the extracted muscle-tendon junctions reveals that the relationship between the terminal myofilaments and the lamina densa of the basal lamina is retained, despite the extensive extraction of the plasma membrane. Fine filaments (2-7 nm) are seen to connect the lamina densa with an electron-dense intracellular layer into which terminal actin filaments appear to insert. These fine filaments are considered to represent an important component of the structural linkage between myofilaments and connective tissue and hence to be a significant component of the tension transmitting mechanism. Their precise nature is not known, but some part of the filaments must pass through the hydrophobic compartment of the plasma membrane and thus must be a transmembrane component of considerable tensile strength. These studies suggest that detergent-extractable membrane lipids play no significant role in the transmission of tension at the muscle-tendon junction, and that fine filaments, probably protein, are responsible for transmitting tension from myofilaments, through the plasma membrane, to the lamina densa of the basal lamina. Early research concerning the mechanism by which force generated within a skeletal muscle fiber is transmitted to the tendon concentrated on two questions: (a) does force transmission occur only at the ends of the muscle fiber, or does mechanical coupling occur all along the fiber; and (b) does the plasma membrane form a continuous boundary between sarcoplasm and tendon, or are myofilaments continuous with tendon filaments at the myotendinous junction. Modern work supports the first alternative in each case (Bennett, 1955; Gelber et al., 1960). In particular, the electron microscope has revealed that a continuous bilayer membrane forms the external boundary of the cytoplasm of skeletal muscle cells, as it does of all other cells. The presence of a morphologically recognizable membrane implies that a continuous double layer of lipid (including phospholipid, glycolipids, and cholesterol) separates myofilaments from connective tissue components. Lipids, however, are characterized by low shear and tensile strength. Since skeletal muscle fibers can generate forces of approximately lo6 dyn/cm2, and the tensile strength of collagen is greater than lo8 dyn/cm2(Alexander, 1968), it is likely that some nonlipid component of the membrane must transmit force across the lipid domain. The only known nonlipid macromolecular species which spans the thickness of natural membranes is protein. These considerations strongly suggest the hypothesis that transmembrane proteins are essential links between myofilaments and connective tissue filaments. Received February 2, 1981: accepted May 6, 1981 294 J.A. TROTTER, K. CORBETT, AND B.P. AVNER Previous work from this laboratory (Trotter et al., 1978) and others has demonstrated that nonionic detergents can be effectively employed to dissolve many membrane lipids because of their ability to act as lipid solvents by disrupting hydrophobic bonds (Helenius and Simms, 1975). At the same time, nonionic detergents are largely ineffective in disrupting most protein-protein interactions. In particular, the nonionic detergent Triton X-100 has been shown to maintain the contractile apparatus of both striated (Solaro et al., 1971) and smooth muscle (Small, 1977; Gordon, 1978) in a functionally intact state. Similarly, Triton X-100 has no known effects on the protein or polysaccharide components of the connective tissues. In a n effort to test the hypothesis that transmembrane proteins are essential mechanical links between myofilaments and connective tissue filaments, the present studies were undertaken to determine: (a) whether the lipid portion of the plasma membrane of the myotendinous junction can be extracted by Triton X-100; (b) what effect extraction of membrane lipids by Triton X-100 would have on the mechanical properties of the myotendinous junction; and (c) what are the ultrastructural characteristics of the detergent-extracted myotendinous junction. MATERIALS AND METHODS Detergent extraction of muscle-tendon preparations Extensor carpi radialis longus and brevis muscles were dissected from adult Balb C BaileylJ mice under sodium pentothal-induced anesthesia. The exposed tissues were maintained in a moist condition by irrigation with physiologic saline (see below). Prior to resection, the muscles were tied distally by their tendons and proximally by their muscle bellies to 0.5-mm diameter tungsten wire, using 000 silk sutures. The muscles were then severed at their origins and insertions. The muscles tied to tungsten wires were extracted for 48 h r in ice-cold extraction solution. For the physiological experiments, the extraction solution contained EGTA. For morphological examination the extraction solution contained EGTA, CaCI,, or neither. (The exact compositions of the extraction solutions and the composition of abbreviated substances are given below.) They were then rinsed in a n identical solution without PMSF or Triton X-100. In some experiments, both sutures were tied to muscle (i.e., the distal suture was tied proximal to the myotendinousjunction), or both sutures were tied to the same tendon (i.e., the proximal suture was tied distal to the myotendinous junction). In this series of experiments 41 muscles were extracted and analyzed as described below. Determination of isometric tension generated by extracted muscles Extracted and rinsed muscles were suspended by the sutures in a tissue bath containing 10 ml of equilibration solution (see below) at 30°C, with one suture fixed to the bottom of the chamber and the other connected to a Grass model FT.03 force displacement transducer. The initial tension on each muscle was adjusted to 0.5g and maintained at that level during the equilibration period of 20-30 min. Nitrogen gas was bubbled through the solution throughout the experiment. Contraction was initiated by the addition of 0.1 M ATP and 1 M CaC1, to the final concentrations indicated in the “contracting solution” below. Mechanical responses were recorded isometrically on a Grass model 7 polygraph. Fixation a n d microscopy Fresh and extracted muscles were fixed by immersion in 2.4% glutaraldehyde, 0.1 M sodium phosphate buffer, pH 7.0, at room temperature for 18 hr. They were then rinsed in 0.1 M phosphate buffer, and postfixed in icecold 1%Os04, 0.1 M phosphate buffer, pH 6.0, for 60 min. After rinsing in water for 30 min, the muscles were immersed in 0.5% uranyl acetate in water for 60 min, and were then dehydrated in a series of increasing concentrations of ethanol. The ethanol was replaced by propylene oxide, and the muscles were finally embedded in Epon. Thin ( 0 . 5 ~ sections ) were stained with Richardson’s (1960) stain and were photographed in bright field illumination on a Zeiss photomicroscope. Ultrathin sections (0.06~)were stained with uranyl acetate and lead citrate and were subsequently examined with a n Hitachi model HU-11-C electron microscope. Solutions used The physiologic saline contained: 125 mM NaCl; 2.7 mM KC1; 25 mM Tris-HC1, pH 7.5; 1.8 mM CaCl,; 0.5 mM MgCl,; 0.12 mM glucose. MUSCLE-TENDON JUNCTION 295 to generate tension when provided with MgATP (Fig. 3). Extraction also leaves functionally intact the mechanical junction between the contractile apparatus and the tendon, as is indicated by the ability of the tendon to transmit tension from the extracted muscle fiber to the force transducer. That this force is generated totally by the muscle is shown by the complete lack of response of the tendon to ATP. Electron microscopy of the extracted muscles reveals that the plasma membrane (as well as the membranous organelles of the sarcoplasm) are removed by the detergent (Figs. 4, 5), whereas the basal lamina retains its structure and its spatial relationship to the cell periphery. The thin (actin) filaments end in a dense layer which is composed of dense globules in a less dense background (Figs. 4, 5). A slight indication of a similar organization can also Materials be seen in unextracted fibers (Fig. 2). This All chemicals were reagent quality. Di- dense layer is connected to the lamina densa thiothreitol, phenylmethylsulfonyl fluoride, of the basal lamina by fine filaments which PIPES, Tris, ATP, imidazole and Triton X-100 traverse the lamina lucida and the space forwere purchased from the Sigma Chemical merly occupied by the lipoprotein bilayer of the plasma membrane. The filaments do not stain Company. well with uranyl acetate and lead citrate; nor RESULTS is this staining enhanced by using fixation proMurine extensor carpi radialis longus and tocols employing alcian blue, ruthenium red, brevis muscles were chosen for this study be- cetylpyridinium chloride, or tannic acid. Meascause (a) they are fusiform muscles with dis- urements of the filament diameters are actinct myotendinous junctions and long ten- cordingly uncertain, but are estimated to fall in the range of 2 to 7 nm. The filaments cross dons; (b) they are small enough to permit convenient fixation, embedment, and section- the lamina lucida and the extracted membrane ing through the entire thickness of the mus- space without interruption. Their orientation cles; and (c) they are large enough to produce is more or less perpendicular to the long axis of the actin filaments, but deviations from the measurable tension using techniques readily available. A light micrograph of an extracted perpendicular are frequent. In most cases, the dimensions of the lamina muscle (Fig. 1)shows that the muscle fiber and the connective tissue mutually interdigitate at lucida and lamina densa are not appreciably the myotendinous junction. Electron micro- altered by detergent extraction in the presence of EGTA. However, in some instances the lamgraphs (Fig. 2) of the myotendinous junction ina lucida is greatly widened (Fig. 6), and in of an unextracted (control) muscle show that the junction is characterized by the subplas- these instances the filaments which connect malemmal density into which thin (actin) fil- the lamina densa to the actin-filament-density aments appear to insert, a continous plasma are greatly lengthened. In extreme cases, the membrane, and a basal lamina composed of a muscle fiber separates completely from the lamina densa, which remains tightly adherent 30-nm thick lamina densa and a 20-nm thick lamina lucida. The lamina lucida appears to to the connective tissue matrix (Fig. 7). The be traversed by fine filaments connecting the widening of the lamina lucia and the complete plasma membrane to the lamina densa. Ex- separation of myofiber from lamina densa octernal to the lamina densa is a filamentous cur to a lesser extent in a modified extraction solution, from which the EGTA is omitted. matrix. Extraction of the muscles with Triton X-100 However, if the “extraction solution” is releaves the contractile apparatus intact, as is placed by the “calcium extraction solution,’’ shown by the ability of the extracted muscles which contains 1.5 mM CaCl,, the myofibers The Extraction solution contained: 1% v/v Triton X-100;2 mM MgC12; 2 mM EGTA; 10 mM PIPES. NaOH, pH 6.0; 50 mM KC1; 1 mM DTT; 1 mM PMSF. The calcium extraction solution contained: 1% Triton X-100;2 mM MgC1,; 1.5 mM CaC1,; 10mM PIPES, pH 6.0; 50 mM KCl; 1 mM DTT; 1 mM PMSF. The equilibration solution contained: 10mM Imidazole-HC1, pH 7.0; 2 mM MgC1,; 50 mM KC1; 1 mM DTT. The contracting solution contained: 10 mM Imidazole-HC1, pH 7.0; 2 mM MgC1,; 0.1 mM CaC1,; 1.5 mM ATP; 50 mM KC1; 1 mM DTT. Abbreviations used: DTT; Dithiothreitol; PIPES; piperazine-N,N’-bis (2-ethanesulfonic acid); EGTA; ethylene glycol-bis-(B-aminoethy1 ether) N,N‘-tetraacetic acid; PMSF; phenylmethylsulfonyl fluoride. 296 J.A. TROTTER, K. CORBETT, AND B.P. AVNER 297 MUSCLE-TENDON JUNCTION do not separate from the lamina densa, nor does the lamina lucida show any indication of widening. In regions other than the myotendinous junction, the periphery of detergent-extracted myofibers is characterized by a distinct space separating myofibrils from basal lamina (Fig. 8). The width of this space is quite variable (up to several microns) and contains a variable amount of electron-dense material in the form of globules (-- 20 nm in diameter), flocculent material, and filaments (6-13 nm). Similar regions of unextracted fibers show that this space is occupied by mitochondria and other membranous structures, and by ribosomes in a dense, vaguely filamentous, matrix (Fig. 9). The dense material juxtaposed to the plasmalemma in the unextracted cell (Fig. 9)has apparent connections to the basal lamina in the extracted cells (Fig. 8). The dense material is sometimes seen to form transverse plates in register with the Z lines of the most peripheral myofibrils (not illustrated). DISCUSSION The results presented here establish that exposure of skeletal muscle to Triton X-100for 48 hr results in the complete elimination of morphologically identifiable membranes, both plasma membrane and internal membranes. Since no chemical studies have been performed, it is impossible to specify the extent to which various membrane components have been extracted. It seems prudent, therefore, to consider the effect of Triton X-100 on muscle to be the destruction of the integrity of membranes, rather than the complete extraction of membrane components. Indeed, the ability of extracted muscles to transmit force to their tendons argues forcefully that a t least one component of the plasmalemma is not extracted by Triton X-100: Fig. 1. Light micrograph of the muscle-tendon junction of a fiber (F) of the extensor carpi radialis longus muscle. T, tendon; N, nucleus of fibrocyte. A region comparable but not identical to that enclosed in the dotted square is shown in Figure 2. x 1,600. Fig. 2. Electron micrograph of a portion of the muscle-tendon junction region of an unextracted (control) muscle. The lamina lucida (L) is populated by numerous fine filaments, many of which appear to connect the plasma membrane (open arrowhead) with the lamina densa (D). The cytoplasmic surface of the plasma membrane is characterized by a dense layer (with faintly visible globules) (arrowhead) into which the thin filaments appear to insert. x 97,000. namely, that element (or elements) which mechanically couples the contractile proteins to the extracellular tensile structures. The fact that Triton X-100is known to be very effective as a solvent of membrane lipids, and very ineffective as a disrupter of protein-protein interactions (Helenius and Simons, 1975), strongly suggests that proteins which span the hydrophobic domain of the muscle membrane are the elements which form this mechanical link. The number of putative polypeptides forming this linking structure is not determinable from these studies. There may be a single transmembrane polypeptide, analogous to Band 3 in the erythrocyte membrane (Lux, 1979),which binds to contractile proteins on the cytoplasmic side of the membrane and to extracellular macromolecules on the other. Or there may be several polypeptides which interact within the plane of the membrane. Further studies will be required to explore these possibilities. The ultrastructural analysis of detergentextracted muscle-tendon junctions tends to corroborate this view. Although a morphol- 2 MIN T t t 3 A G Fig. 3. Force generated by detergent-extracted muscles plotted versus time. In the upper tracing (M) the sutures were both tied to muscle; therefore, this graph displays the force generated by the myofibrils. In the middle tracing (MT), one suture was tied to the tendon; this graph thus displays the force transmitted from myofibrils to tendon. In the bottom tracing (T),both sutures were tied to tendon. At point “ A , ATP was added; a t point “ G , glutaraldehyde was added. The drop in tension upon addition of glutaraldehyde is inconsistent from experiment to experiment and therefore cannot be interpreted. 298 J.A. TROTTER, K. CORBETT, AND B.P. AVNER Figs. 4 and 5. Electron micrographs of detergent-extracted muscle-tendon junctions. The myofilaments of the muscle fibers (F) are well preserved, a s is the basal lamina. No cell membrane is seen. The lamina ludica (L) is crossed by many fine filaments that run between the peripheral density of the sarcoplasm and the lamina densa (D). Fine arrows indicate regions in which the fine filaments are especially well seen; thick arrows indicate regions in which thicker filaments are observed. The sarcoplasm has a globular appearance, especially in the peripheral regions. x 97,000. 299 MUSCLE-TENDON JUNCTION Fig. 6. In the presence of EGTA, the lamina densa of some detergent-extracted muscle-tendon junctions is partially separated from the sarcoplasm by a greatly widened lamina lucida (L). Fine filaments are seen crossing the lam- Fig. 7. This muscletendon junction, extracted in Triton X-100 in the presence of EGTA, has separated almost completely. The lamina densa (D) has remained adherent to the connective tissue, in which collagen (C) fibrils are seen. The :-"i . , ~ , i ~av nnn rln-mn nlnh..la, -ot..m n C thn t n r m k - 1 rnrnnn1o.m ;o .smll 300 J.A. TROTTER, K. CORBETT, AND B.P. AVNER Fig. 8. A nonterminal portion of n detergent-extracted muscle fiber shows a gap between the myofibrils (F) and the lamina densa (D). This gap is occupied by varying amounts of sarcoplasmic material composed of granular and filamentaus components. x 40.000. Fig. 9. A nonterminal region of an unextracted muscle shows that the space between the myofibrils and the plasma membrane is occupied by membranes, mitochondria (M), and ribosomes. X 40,000. MUSCLE-TENDON JUNCTION ogically recognizable plasmalemma is missing, structural continuity between intracytoplasmic electron-dense material and the lamina densa is maintained by a population of fine filaments. Filaments which extend between the lamina densa and the plasmalemma at the myotendinous junction have been previously described (Hanak and Bock, 1971; Korneliussen, 1973; Ajiri, 1978). However, the present work extends previous observations in two significant ways: (1)It demonstrates that the filaments of the lamina lucida are linked to cytoplasmic structures, either by directly passing through the plasma membrane or by being directly connected with transmembrane elements; and (2) it suggests that these filaments actually carry tension, since both they and the ability to transmit tension to tendons are preserved by Triton X-100 extraction of the muscles. The smallest filaments observed in the extracted lamina lucida are about 2 nm, and the largest are about 7 nm. There may be several populations of filaments present in these regions; or, alternatively, the filaments may have a tendency to aggregate, either naturally or as a result of experimental intervention. The filaments may also possess, or they may be part of a filament complex that possesses, considerable extensibility after a lengthy (48 hr) exposure to EGTA. Stretched and broken filaments are observed when the lamina lucida is moderately widened (up to 60 nm). That is, a filament length in excess of 60 nm has not been observed. When the lamina lucida exceeds 60 nm, as in the case of complete muscle-tendon separation, the filaments are absent, presumably because they have broken. Inclusion of CaC12 during extraction of the muscles prevents this separation of basal lamina from sarcoplasm. Although it is impossible to identify the sites which are stabilized by Ca2+,it is reasonable to think that they are normally external to the outer leaflet of the plasma membrane, since this locale is exposed to the extracellular concentration of Ca2+(1.5 mM). It is probable, in any case, that a Ca2+sensitive site is involved in the linking of the lamina densa to myofilaments, but not in the junction between lamina densa and extracellular matrix. The Ca2+-sensitivesite might be within the lamina densa. This notion is strengthened by our recent observation that in cultured myotubes the linkages between myofibrils and the substratum are insensitive to the concentration of Ca2+(unpublished observations). 301 “Microfibrils”(Hanack and Bock, 1971; Ajiri et al., 19781,“thread-like or spine-like profiles” (Korneliussen, 1973), and “filamentous structures” (Nakao, 1976) in the lamina lucida of the myotendinous junction; and, “pillars or partitions” or “strands” have been described in the corresponding layers of desmosomes (Kelly, 1966). The character and function of these structures has remained speculative, however. The present study, by showing a correlation between the presence of the fine filaments and of a functionally intact myotendinous junction, strengthens the idea that they serve as mechanical linkers in the muscle-tendon junction. About the physical and chemical character of the filaments, we have little information. It is worth noting that the protein microfibrils associated with elastic filaments are approximately 8 nm in diameter, and may be composed of finer filaments, approximately 1.5 nm (Gotte and Serafini-Fracassini, 1963) to 2.5 nm (Serafini-Fracassini, 1978). The triple helix of collagen also has a diameter of approximately 1.4 nm and associates to form fibrils of 8 nm and greater (Eyre, 1979), which are probably composed of multiple repeats of 8-nm units (Parry and Craig, 1979). Since the lamina densa of basal laminae is composed in large part of type IV collagen (Kefalides, 1980) and is morphologically composed of randomly arranged fibrils in a granular matrix, the possibility exists that the connecting fibrils described herein are in part composed of basal lamina collagen. Carrying this speculation further, one could suggest that the transmembrane components which are bound intracellularly to actin (either directly or indirectly) are bound extracellularly to collagen microfibrils in the lamina lucida. These microfibrils might then coalesce to form fibrils that are in turn cross-linked to the collagen fibrils of the lamina densa. The crosslinked random filament structure of the lamina densa would be expected to provide the system with a certain amount of rubbery elasticity (Alexander, 1968). Again, we have no evidence at this time that the filaments of the lamina lucida are collagen or elastin. However, the extensibility of the filaments in the lamina lucida suggests that they might be an elastic element in series with myofilaments and tendon, and thereby serve as a damper during contraction. A similar suggestion has been advanced by Hanak and Bock (1971). Concerning the transmembrane component of the linking system of the myotendinous junction, we also have little information. 302 J.A. TROTTER, K. CORBETT, AND B.P. AVNER Freeze-fracture studies on microvilli have shown that 8-nm filaments which bind to intracellular actin filaments penetrate the lipid bilayer (McNutt, 1978).In desmosomesas well, traversing filaments have been described which seem to bind to 10-nm filaments intracellularly and to pierce or traverse the membrane (Hull and Staehelin, 1979; Kelly and Kuda, 1980).Also, we have recently described transmembrane linking elements which apparently bind the actin of macrophages to extracellular substrata (Trotter, 1981). The nature of these structures is uncertain; in particular, the number of molecules involved is unknown. The filaments may be homo- or heteropolymers of polypeptides or other macromolecules. In the myotendinous junction as well, there is no evidence that a monomolecular filament extends from actin filaments to the lamina densa. Rather, it seems more likely that at least three distinct proteins are involved in the transmission of force: (1)A protein that binds actin and transmits tension to the “membrane”; vinculin (Block and Geiger, 1980) is a possible candidate for this protein. (2) A protein that spans the hydrophobic portion of the membrane. (3) A protein that transmits tension from the “membrane” to the lamina densa. At the moment, in terms of chemistry, we are unable to do more than suggest possible candidates for these proteins. Nonetheless, the fact that we can morphologically identify the tension-transmitting elements that are of physiological significance encourages the belief that a molecular descrption of this region is possible. ACKNOWLEDMENTS The authors express appreciation to Ms. Suzanne Newel1 for her help in preparation of the manuscript, and to Ms. Judi DeLongo for her help in preparing thin and ultrathin sections. This work was supported by a grant from the Muscular Dystrophy Association. LITERATURE CITED Ajiri, T., T. Kimura, R. Ito, and S. Inokuchi 11978) Microfibrils in the myotendinous junction. Acta Anat., 102:433-t39. Alexander, R.M. (1968)Animal Mechanics. University of Washington Press, Seattle. Bennett, H.S. (1955)Modem Concepts of the structure of striated muscle. Amer. J. Phys. Med., 34:46-67. Bloch, R.J., and B. Geiger (1980)The localization of acetylcholine receptor clusters in areas of cell-substrate contact in cultures of rat myotubes. Cell, 21:25-35. Eyre, D.R. (1980)Collagen: Molecular diversity in the body’s protein scaffold. Science, 207:1315-1322. Gelber, D., O.H. Moore, and H. Ruska (1960)Observations of the myo-tendon junction in mammalian skeletal muscle. Zeit. fur Zellforsch., 52:396-400. Gordon, A.R. (1978) Contraction of detergent-treated smooth muscle. Proc. Nat. Acad. Sci., 75:35273530. Gotte, L., and A. Serafini-Fracassini (1963)Electron microscopic observations on the structure of elastin. J. Atheroscler. Res., 3:247-251. Hanak, H., and P. Bock (1971)Die Feinstruktur der muskelsehnenverbindung von Skelett-und Henmuskel. J. U1trastruct. Res., 36:68-85. Helenius, A,, and K. Simons (1975)Solubilization of membranes in detergents. Biochim. Biophys. Acta, 415:29-79. Hull, B.E., and L.A. Staehelin (1979)The terminal web. J. Cell Biol., 81:67-82. Kefalides, N.A. (1980)Chemistry of basement membranes: structure and synthesis. Advan. Microcir., 9:295322. Kelly, D.E. (1966)Fine structure of desmosomes, hemidesmosomes and an adepidermal globular layer in developing newt epidermis. J. Cell Biol., 28.51-72. Kelly, D.E., and A. Kuda (1980)Desmosomal fine structure revealed by dark-shadow printing and complementary freeze-fracture techniques. Anat. Rec., 196:95a. Komeliussen, H. (1973)Ultrastructure of myotendinous junctions in Myxine and rat. Z.Anat. Entwick1.-Gesch., 142:91-101. Lux, S.E. (1979)Dissecting the red cell membrane skeleton. Nature (London),281:426429. McNutt, N.S. (1978)A thin-section and freeze-fracture study of microfilament-membrane attachments in choroid plexus and intestinal microvilli. J. Cell Biol., 79:774-787. Nakao, T. (1976)Some observations on the fine structure of the myotendinous junction in myotomal muscles of the tadpole tail. Cell Tiss. Res., 166:241-254. Richardson, K.G., L.Jarrett, and E.H. Finke (1960)Embedding in epoxy resins for ultrathin sectioning in electron microscopy. Stain Technol., 35:313-323. Serafini-Fracassini,A,, J.M. Field, and J. Hinnie (1978)The primary filament of bovine elastin. J. Ultrastruct. Res., 65~190-193. Small, J.V. (1977)Studies on isolated smooth muscle cells: The contractile apparatus. J. Cell Sci., 24:327-349. Solaro, R.J., D.C. Pang, and F.N. Briggs (1971)The purification of cardiac myofibrils with Triton X-100.Biochim. Biophys. Acta, 245:259-262. Trotter, J.A., B.A. Foerder, and J.M. Keller (1978)Intracellular fibers in cultured cells: Analysis by scanning and transmission electron microscopy and by SDS-polyacrylamide gel electrophoresis. J. Cell Sci., 31:369-392. Trotter, J.A. (1981)The organization of actin in spreading macrophages. Exp. Cell Res., 132:235-248.