The reorganization of myofibrils and the re-formation of intercalated discs was examined in neonatal rat cardiac muscle cells during the first 72 h of culture. Rhodamine phalloidin was used to monitor the organizational state of the myofibrils and antibodies to desmoplakin and vinculin were used as markers for the presence of desmosomes and fasciae adherentes, respectively. Tiny punctate desmosomes were observed between muscle cells after 24 h and apparently increased in number and/or size between 24 and 48 h in culture. Fasciae adherentes were not detectable with antibodies to vinculin until after 48 h in culture. Well-defined sarcomeres were restored after 48 h in culture. Once formed the sarcomeric organization of the myofibrils was found to be stable provided they were attached to the sarcolemma via intercalated discs. However, if the myofibrils attached to regions of the membrane that lacked intercalated discs the sarcomeres appeared to break down gradually centripetally. When myofibrils attached to the membrane at the free edges of cells that were not in contact with other muscle cells, the striations stopped abruptly at a considerable distance before the myofibril attached to the membrane. These non-striated terminals elongated between 48 and 72 h and were associated with focal contacts that contained vinculin. Overall the results suggest that cell-cell contact may be critical for the stabilization of normal myofibrillar structure in the heart.

The intercalated disc is a specialized region of the cardiac muscle cell membrane characterized by a grouping of three specialized intercellular junctions, which include the macula adherens, the fascia adherens and the gap junction (McNutt, 1970). The intercalated disc is vital for the integrity and function of the heart. For example, gap junctions mediate the electrical coupling between heart muscle cells (Sheridan & Atkinson, 1985). The fascia adherens and the macula adherens are believed to mediate the adhesions between adjacent muscle cells and also provide attachment sites for cytoskeletal elements to the cell membrane (Severs, 1985). Little is known concerning the molecular constituents of the two adherens junctions or what molecules mediate membrane-cytoskeletal interactions. However, current evidence indicates that the macula adherens and fascia adherens are biochemically and functionally distinct junctions. The macula adherens contains desmoplakin (Mueller & Franke, 1983) in an electron-dense plaque on the cytoplasmic side of the membrane where intermediate filaments appear to attach to the membrane (Arnn & Staehelin, 1981). In contrast, the dense plaque of the fascia adherens contains vinculin and alpha-actinin, which may be involved in the insertion of the thin filaments of the contractile apparatus into the membrane (Tokuyasu et al. 1981).

Enzymic dissociation of neonatal heart tissue not only separates the cells at intercalated discs but also disrupts the sarcomeric structure of the myofibrils, leaving them in a state of disarray (Wollenberger, 1967; Fischmann & Moscona, 1971). Presumably the connections of the myofibrils to the membrane via the fascia adherens are also cleaved. With time in culture intercalated discs re-form and the myofibrils reorganize (Legato, 1972; Perissel et al. 1980). However, the role of various cytoskeletal elements in this reorganization has not been delineated.

In the studies reported here fluorescence microscopy was used to examine the degree of organization of the myofibrils in large numbers of whole cells. The reappearance of the fascia adherens and desmosome components of the intercalated disc was examined using antibodies specific for vinculin and desmoplakin, which served as distinctive molecular markers for the two types of adherens junctions in cardiac muscle (Geiger et al. 1985 ; Cowin et al. 1985). We followed the re-formation of these structures during the first 3 days of culture of neonatal rat heart cells in order to evaluate their potential as an in vitro system in which to investigate the assembly of intercalated discs. The results indicate that re-formation of desmosomes and fasciae adherentes follow different time courses and that the stability of myofibrils in cultured cardiac muscle cells may depend in part on the establishment of cell-cell contacts.

Materials

Pancreatin and Dulbecco’s modified Eagle’s medium (DMEM) were obtained from GIBCO Laboratories (Grand Island, NY). Rhodamine phalloidin was from Molecular Probes, Inc. (Junction City, OR), normal goat serum, goat anti-mouse and goat anti-rabbit secondary antibodies conjugated with fluorescein were obtained from Cooper Biomedical, Inc. (Malvern, PA). Mouse monoclonal antibody to chicken gizzard vinculin (Geiger, 1981) was obtained from Miles Scientific (Naperville, IL). Rabbit antiserum to bovine desmoplakin I was a generous gift from Dr James Arnn, San Diego State University, San Diego, California (Arnn, 1983; Jones & Goldman, 1985).

>Cell culture

Hearts were removed from 3-day-old neonatal rats and prepared for tissue culture according to the method of Decker et al. (1985) with modifications. Heart ventricles were minced and subjected to 30-min cycles of exposure to 0 ·6 mg ml −1 pancreatin. Cells released in the first two cycles were discarded. In order to remove some of the non-muscle cells and enrich cell suspensions for myocytes, the freshly dissociated cells were incubated in tissue culture dishes for 2h at 37°C in DMEM containing 15 % horse serum and 5 % foetal bovine serum in an atmosphere of 95 % air, 5% CO2. The myocyte-enriched cell suspension was seeded into 35 mm tissue culture dishes containing glass coverslips at densities ranging from 2 ×105 to 7 × 105 cells per dish.

The degree of enrichment for muscle cells was assessed after 24 h in culture for three separate tissue dissociations by periodic acid-Schiff (PAS) staining (Pollinger, 1973). In this procedure muscle cells stain with PAS by virtue of their glycogen content, while non-muscle cells are unstained. Cell suspensions seeded into culture without modification were composed of 54 ·5 ± 3·9% myocytes. After enrichment for muscle cells by differential adhesion, cultures contained 72·8 ± 1·4% myocytes, whereas the other 27·2% were mostly fibroblasts and endothelial cells. However, the distinctive differences between myocytes and non-muscle cells in actin content and organization (see below) suggested that rhodamine phalloidin staining may be a useful alternative to the PAS method for assessing the proportion of muscle cells in primary cultures when the cells are well spread throughout the culture.

Indirect immunofluorescent staining

After 24, 48 and 72 h in culture cells attached to glass coverslips were stained for indirect immunofluorescence. The cells were fixed with a 3·7% formaldehyde in Dulbecco’s phosphate-buffered saline for 10 min at room temperature. The cells were then permeabilized by incubation for 2 min in each of the following: ice-cold 50% acetone-water, acetone at −20°C, 50% acetonewater and finally returned to phosphate-buffered saline. Rhodamine phalloidin and all antibodies were used at a final dilution of 1:20 in phosphate-buffered saline containing 10 mg ml-1 bovine serum albumin. For double-label staining, rhodamine phalloidin and antibodies were each diluted to 1:20 as a single mixture and incubated on coverslips for 30 min at 37°C in a moist chamber. After washing in PBS, cells were incubated for 30 min at 37°C with goat anti-mouse or goat anti-rabbit immunoglobulin G (IgG) secondary antibodies conjugated with fluorescein isothiocyanate diluted 1:20. Finally, the coverslips were washed and mounted on glass slides in Gelvatol containing 100 mg ml−1 of l,4-diazabicyclo-[2.2.2]-octane to retard fading of fluorescein. Replicate coverslips stained with normal rabbit serum, or normal mouse serum in lieu of immune sera gave negative staining reactions except for weak nuclear staining seen with all polyclonal sera used. The nonspecific nuclear fluorescence was exaggerated in photographs where long exposure times were required (e.g. see Fig. 3). Monoclonal antibodies to desmoplakin I and II (Boehringer-Mannhein; Cowin et al. 1985) gave the same results as rabbit polyclonal antibodies to desmoplakin.

The slides were viewed with a Leitz Dialux 20 microscope equipped for epifluorescence. For fluorescein viewing the L2.1 cube was used, which is a very narrow-band blue excitation filter with selective barrier at 525 nm. For rhodamine an N2 narrow-band green excitation filter was used, which excluded FITC excitation. The objective was a 63X/1.40 piano apochromatic phasecontrast lens. Photographs were taken with the Leitz Vario-orthomat camera system using Kodak Tri-X film developed with Diafine (Acufine Inc., Chicago, IL).

Focal contacts were observed with a Zeiss Photo III microscope equipped for interference reflection using a 63X/1.25 Neofluar Antiflex lens. The cells were prepared for immunofluorescence as described above except filiation in formaldehyde was for 30 min. Permeabilization of the cells with acetone or 0·1 % Triton X-100 for 4 min gave similar results. Glass coverslips with adherent cells were mounted on slides by placing them on pieces of no. 1 coverglass and the gap filled with PBS. The edges were sealed with wax to prevent evaporation.

Primary cultures of heart cells from 3-day-old neonatal rats contained at least three different types of cells that exhibited marked differences in their pattern of staining with rhodamine phalloidin (Fig. 1). Muscle cells exhibited a regular cross-banding pattern after staining with phalloidin and in well-developed cells the sarcomeres were silhouetted to reveal Z lines, A bands and I bands (Fig. 1A). In the non-muscle cells the actin was disposed in various arrangements of non-striated actin cables (Fig. 1A,B). In fibroblasts stress fibres were often seen throughout the cytoplasm of the cell (Fig. 1A), while in endothelial cells actin cables were restricted to the cell periphery and were arranged circumferentially (Fig. IB).

Fig. 1.

Contrast between the organization of actin in the three predominant cell types found in cultures of dissociated neonatal rat heart. Staining of 72 h cultures with rhodamine phalloidin. A. A single cardiac muscle cell sitting on a bed of heart fibroblasts. Note that Z lines A and I bands are highlighted in the myofibrils. B. Four presumptive endothelial cells in the centre of the field. The actin cables in non-muscle cells are not striated. Bar, 10 μm.

Fig. 1.

Contrast between the organization of actin in the three predominant cell types found in cultures of dissociated neonatal rat heart. Staining of 72 h cultures with rhodamine phalloidin. A. A single cardiac muscle cell sitting on a bed of heart fibroblasts. Note that Z lines A and I bands are highlighted in the myofibrils. B. Four presumptive endothelial cells in the centre of the field. The actin cables in non-muscle cells are not striated. Bar, 10 μm.

Cellular organization after 24 h in culture

Rhodamine phalloidin was used to monitor the organizational state of the myofibrils, and antibodies to vinculin and desmoplakin were used as markers to detect the presence of fasciae adherentes and desmosomes, respectively. Twenty-four hours after seeding there was considerable heterogeneity in the degree of organization of the myofibrils in muscle cells (Fig. 2). The most common morphology observed at 24 h was regular arrangements of fluorescent bands or stripes of actin, which probably corresponded to developing I bands (Fig. 2A). The gaps between the bands appeared to be devoid of actin. Other myocytes apparently revealed the first signs of actin bridging the gaps (Fig. 2B). A smaller subset of cells contained variable numbers of striated myofibrils subjacent to the sarcolemma (Fig. 2C). These may represent cells in which the myofibrils were not completely disrupted during dissociation or which had reassembled their myofibrils more rapidly than the majority of the cells. Still other cells contained actin arranged like beads on a string, which looked like myofibrils of very small diameter (Fig. 2D).

Fig. 2.

Heterogeneity in the organization of the myofibrils in muscle cells 24 h after seeding into culture. A-D. Stained with rhodamine phalloidin. A. The staining pattern typical of the majority of the cells. Note that gaps between parallel bands of fluorescence appear to be devoid of actin. B. Sarcomeres appear to be assembling in central regions of the cell. C. Representative of a small subpopulation of cells with well-developed myofibrils adjacent to the sarcolemma, but absent from the deeper cytoplasm. D. Presumptive myocyte with myofibrils of very small diameter following tortuous and branching paths through the cytoplasm. Bar, 10μm.

Fig. 2.

Heterogeneity in the organization of the myofibrils in muscle cells 24 h after seeding into culture. A-D. Stained with rhodamine phalloidin. A. The staining pattern typical of the majority of the cells. Note that gaps between parallel bands of fluorescence appear to be devoid of actin. B. Sarcomeres appear to be assembling in central regions of the cell. C. Representative of a small subpopulation of cells with well-developed myofibrils adjacent to the sarcolemma, but absent from the deeper cytoplasm. D. Presumptive myocyte with myofibrils of very small diameter following tortuous and branching paths through the cytoplasm. Bar, 10μm.

Fig. 3.

Changes in desmosomes and the degree of organization of the myofibrils between 24 and 48 h in culture. A,C. Myocytes stained with rhodamine phalloidin 24 h and 48 h after seeding into culture, respectively. Well-defined sarcomeres can be seen in the myofibrils after 48 h that are apparently lacking after 24 h in culture. Details in the banding pattern of sarcomeres in the myofibrils marked with a bracket (C) are comparable to those in a similar area in brackets in Fig. 6A, shown at higher magnification. B,D. Immunofluorescent staining of the same two fields of cells shown in A and C with antibodies to desmoplakin to visualize desmosomes. B. After 24 h in culture desmosomes are seen as very fine dotted lines of fluorescence at sites of cell-cell contact (arrowheads). D. The staining of desmosomes is markedly more intense in 48h cultures (arrowheads). Tiny spots of fluorescence are also present in the cytoplasm of muscle cells in 24 h and 48 h cultures. Bar, 10μm.

Fig. 3.

Changes in desmosomes and the degree of organization of the myofibrils between 24 and 48 h in culture. A,C. Myocytes stained with rhodamine phalloidin 24 h and 48 h after seeding into culture, respectively. Well-defined sarcomeres can be seen in the myofibrils after 48 h that are apparently lacking after 24 h in culture. Details in the banding pattern of sarcomeres in the myofibrils marked with a bracket (C) are comparable to those in a similar area in brackets in Fig. 6A, shown at higher magnification. B,D. Immunofluorescent staining of the same two fields of cells shown in A and C with antibodies to desmoplakin to visualize desmosomes. B. After 24 h in culture desmosomes are seen as very fine dotted lines of fluorescence at sites of cell-cell contact (arrowheads). D. The staining of desmosomes is markedly more intense in 48h cultures (arrowheads). Tiny spots of fluorescence are also present in the cytoplasm of muscle cells in 24 h and 48 h cultures. Bar, 10μm.

The first elements of the intercalated disc that developed within the first 24 h in culture were desmosomes and gap junctions. Although no probes specific for gap junction proteins were used in this study, nexal junctions between muscle cells had apparently formed since groups of muscle cells beat synchronously (De Mello, 1982). Antibodies to desmoplakin stained very thin continuous or dotted lines of fluorescence at sites of cell-cell contact (Fig. 3B). Vinculin, a component of the fascia adherens, was not detected at sites of cell-cell contact in 24 h cultures.

Changes in cellular organization between 24 h and 48 h

There were marked changes in the staining of desmosomes and in the degree of organization of the myofibrils between 24 and 48 h in culture. Well-defined sarcomeres were not present in the majority of cells at 24h (Figs 2A, 3A). During the next day in culture, other elements of the sarcomere were inserted in the gaps between the bands of fluorescence as evidenced by the presence of A and I bands visualized with phalloidin staining in 48 h cultures (Fig. 3C). The banding pattern of the sarcomeres in the cells shown in Fig. 3C (bracket) were comparable to those shown in Fig. 6A (bracket) when viewed at higher magnification. Clearly a further reorganization of the contractile proteins had taken place between 24 and 48 h in culture.

Changes in the staining of desmosomes with antibodies to desmoplakin were also dramatic between 24 and 48h in culture. As shown in Fig. 3D, the staining of desmosomes had become markedly more intense by 48 h in culture. This might suggest an increase in the number and/or surface area of the junctions. In addition, spots of fluorescence were seen in the cytoplasm of the cells (Fig. 3D). The spots, also present in 24 h cultures, could represent desmosomes that had been internalized after tissue dissociation (Overton, 1968) or precursors for new desmosomes en route to the cell surface (Jones & Goldman, 1985) or both. No further change in the staining of desmosomes was detected between 48 and 72 h in culture. Desmoplakin was detected only in muscle cells; non-muscle cells were not stained by this antibody.

While desmosomes were first detected after 24 h in culture, fasciae adherentes were not detected by immunofluorescent staining with antibodies to vinculin until after 48 h in culture (Fig. 4A,B). Sites of cell—cell contact were labelled with very fine lines of fluorescence that were difficult to capture on film. Furthermore, there was no change in the intensity of staining up to 72 h in culture.

Fig. 4.

The distribution of vinculin in 48 h cultures of cardiac myocytes, A. Phasecontrast view of a group of three muscle cells. B. The same cells after immunofluorescent staining with monoclonal antibodies to vinculin. Lines of fluorescence at sites of cell-cell contact mark the presence of fasciae adherentes junctions (arrows). C,D. Double-label staining of a muscle cell with rhodamine phalloidin (C) and antibodies to vinculin (D). Note that the terminals of myofibrils are not striated and that they are associated with plaque-like deposits of vinculin where they attach to the sarcolemma (arrowheads). Bar, 10 μm.

Fig. 4.

The distribution of vinculin in 48 h cultures of cardiac myocytes, A. Phasecontrast view of a group of three muscle cells. B. The same cells after immunofluorescent staining with monoclonal antibodies to vinculin. Lines of fluorescence at sites of cell-cell contact mark the presence of fasciae adherentes junctions (arrows). C,D. Double-label staining of a muscle cell with rhodamine phalloidin (C) and antibodies to vinculin (D). Note that the terminals of myofibrils are not striated and that they are associated with plaque-like deposits of vinculin where they attach to the sarcolemma (arrowheads). Bar, 10 μm.

When myofibrils terminated in the free edges of cells that were not in contact with other muscle cells the striations of the myofibrils stopped abruptly a considerable distance before the myofibril attached to the sarcolemma (Fig. 4C). The non-striated terminals appeared continuous with striated portions of the myofibril deeper in the cytoplasm. Furthermore, elongated membrane plaques that stained with antibodies to vinculin were associated with the non-striated terminals of myofibrils but not with striated regions of myofibrils (Fig. 4D). The vinculin plaques corresponded exactly to the location and extent of non-striated terminations of myofibrils where they attached to the membrane at the free edges of muscle cells (Fig. 4C,D). Interference reflection combined with double-label immunofluorescent staining showed a clear association between non-striated myofibril terminals, vinculin plaques and areas of focal contact with the substratum (Fig. 5). The focal contacts in the myocytes were very similar to those seen in fibroblasts (Izzard & Lochner, 1976).

Fig. 5.

Correspondence between non-striated myofibril terminals, vinculin plaques and focal contacts. A,C. 48 h and 72 h cardiac myocytes, respectively, stained with rhodamine phalloidin. B,D. Same cells stained for immunofluorescence with antibodies to vinculin. Insets show the interference reflection images of selected regions of the two cells. Dark areas in interference reflection mark the location of focal contacts of the cells with the substratum. Arrowheads indicate some of the sites of correspondence between the three cell structures. All panels are at the same magnification. Bar, 10 μm.

Fig. 5.

Correspondence between non-striated myofibril terminals, vinculin plaques and focal contacts. A,C. 48 h and 72 h cardiac myocytes, respectively, stained with rhodamine phalloidin. B,D. Same cells stained for immunofluorescence with antibodies to vinculin. Insets show the interference reflection images of selected regions of the two cells. Dark areas in interference reflection mark the location of focal contacts of the cells with the substratum. Arrowheads indicate some of the sites of correspondence between the three cell structures. All panels are at the same magnification. Bar, 10 μm.

In non-muscle cells vinculin was localized in tiny focal adhesion plaques at sites of cell-to-substratum contacts (Geiger, 1979). In endothelial cells vinculin plaques were located along the cell border while in fibroblastic cells the plaques were often scattered all over the ventral cell surface in association with focal contacts seen by interference reflection (unpublished observations).

Cellular organization after 72 h in culture

After 72 h, greater than 90% of the cells contained what appeared to be mature parallel myofibrils in register with one another (Fig. 6A). Thick transverse bands of actin marked the location of sites of cell-cell contact. The bands corresponded with the presence of intercalated discs as manifested by immunofluorescent staining with antibodies to desmoplakin (Fig. 7) and vinculin (not shown). The actin bands at sites of cell-cell contact were occasionally seen after 48 h in culture, but they were present in large numbers after 72 h in culture.

Fig. 6.

Myocytes 72 h after seeding into culture. A. Rhodamine phalloidin staining of muscle cells showing intercalated discs marked by thick transverse bands of actin (arrows). Note that the myofibrils are striated continuously where they attach to the membrane at intercalated discs (bracket). B. Myocytes stained with rhodamine phalloidin showing very’ long non-striated terminals of myofibrils at free edges of the cells. Centrally, the myofibrils have retained their sarcomeric organization. C. The same field doublelabel-stained with antibodies to vinculin showing continued association of plaques with myofibril terminals. Bar, 10 μm.

Fig. 6.

Myocytes 72 h after seeding into culture. A. Rhodamine phalloidin staining of muscle cells showing intercalated discs marked by thick transverse bands of actin (arrows). Note that the myofibrils are striated continuously where they attach to the membrane at intercalated discs (bracket). B. Myocytes stained with rhodamine phalloidin showing very’ long non-striated terminals of myofibrils at free edges of the cells. Centrally, the myofibrils have retained their sarcomeric organization. C. The same field doublelabel-stained with antibodies to vinculin showing continued association of plaques with myofibril terminals. Bar, 10 μm.

Fig. 7.

Double-label immunofluorescent staining of cardiac muscle cells after 72 h in culture. A. Rhodamine phalloidin. B. The same field stained with antibodies to desmoplakin. The thick transverse bands of actin seen in A (arrows) correspond to the location of desmosomes between cells seen in B. Bar, 10 μm.

Fig. 7.

Double-label immunofluorescent staining of cardiac muscle cells after 72 h in culture. A. Rhodamine phalloidin. B. The same field stained with antibodies to desmoplakin. The thick transverse bands of actin seen in A (arrows) correspond to the location of desmosomes between cells seen in B. Bar, 10 μm.

When the bands at sites of cell-cell contact were present, the myofibrils of each cell were striated continuously right up to the intercalated disc on both sides of the junctions (Figs 6A, 7A). At the opposite end of the cells where the same myofibrils terminated at the free edge of the cell, they were not striated (Fig. 7A). In addition the length of the non-striated terminal region of myofibrils was of the order of two to three times longer at 72 h than those seen at 48 h (cf. Figs 4C, 6B). In 48 h cultures, some of the myofibrils were striated all the way to the end (Fig. 3C), while others had non-striated terminals up to 10–12μm in length. After 72 h in culture, non-striated terminals approximately 20 μm in length were common and they ranged up to 30 Jim long (Fig. 6B). These values were based on measurements of terminals whose boundaries were clearly defined. Accurate measurements of the remaining terminals could not be made because they overlapped and crossed over one another, making it difficult to determine precisely where they ended. Nevertheless, examination of hundreds of cells from several different preparations indicated that the increase in the length of the non-striated terminals between 48 and 72 h was quite obvious even without extensive quantitative measurements. They continued to be associated with membrane deposits of vinculin along much of their length (Fig. 6C).

When neonatal rat heart was dissociated into single cells with proteolytic enzymes, intercellular junctions were cleaved and the sarcomeric organization of the myofibrils was disrupted. During the first two days in culture, the cells gradually reassembled their myofibrils, aligned them into parallel arrays and regenerated intercalated discs.

In 24 h cultures, the predominant pattern observed after staining actin in muscle cells with rhodamine phalloidin was parallel bands of fluorescence spaced approximately 2μm apart. This is similar to the distance between Z lines in mature myofibrils. The gaps between the bands appeared to be devoid of actin. This pattern suggested that the bands represented Z discs in the process of developing into I bands in the early stages of the reassembly of sarcomeres. This interpretation is supported by electron-microscopic observations of others. For example, Cedergren & Harary (1964) saw Z material with clusters of thin filaments emanating from them in opposite directions in cultured neonatal rat heart cells. Similar structures were seen in vivo by Chacko (1973) in cardiac muscle cells in 10-day rat embryos. Such structures could give rise to the staining pattern observed with rhodamine phalloidin in the light microscope. Furthermore, double-label immunofluorescent staining of cultured embryonic chick heart muscle with rhodamine phalloidin and antibodies to alpha-actinin indicate that both label Z lines (unpublished observations). In addition, the former stains actin in the I bands of myofibrils.

Intercalated discs were assembled during the first 48 h in culture. Gap junctions presumably had formed during the first 24 h in culture, since groups of muscle cells were seen to beat synchronously. Desmosomes were present at 24h, but fasciae adherentes could not be detected by immunofluorescent staining with antibodies to vinculin until 48 h in culture. Electron microscopic observations of cultures similar to ours by Perissel et al. (1980) indicate that cell junctions morphologically consistent with fasciae adherentes had formed between muscle cells after 24 h in culture. Furthermore, they found that developing desmosomes were often quite small compared to fasciae adherentes in 24 h cultures. Our own electron-microscopic observations corroborate their findings; all three types of junction were present in 24h cultures (unpublished observations). These results combined with the fact that vinculin was not detectable by immunofluorescent staining until after 48 h in culture suggests that the fasciae adherentes present in 24 h cells were not biochemically complete. The disparity between our immunofluorescence results and electron-microscopic data raises two important considerations in studies of the genesis of specialized cell junctions. First, although cell junctions may appear to be mature morphologically, in fact they may not have completed their development at the molecular level. Second, more molecular markers which distinguish between junctions with similar morphologies are needed in order to identify unambiguously particular junctions (Geiger et al. 1983), especially in developing systems.

The wide transverse bands of actin stained with rhodamine phalloidin corresponded to intercalated discs at sites of contact between muscle cells. Thus, rhodamine phalloidin appears to be a useful and convenient probe with which to follow the organizational state of the myofibrils of cultured rat cardiac cells and it may also provide a marker for the presence of intercalated discs.

Reassembly of sarcomeres took place during the first 48 h in culture as manifested by the gradual delineation of sarcomeres in the myofibrils. Beginning at 48 h, the terminals of some of the myofibrils were not striated at the free edges of cells that were not in contact with other muscle cells. The non-striated terminals elongated between 48 and 72 h strongly suggesting that a gradual remodelling or disassembly of sarcomeres was occurring centripedally (cf. Figs 4C, 6B). At the same time, the sarcomeric organization of the myofibrils remained stable at sites of cell-cell contact where intercalated discs had formed.

Non-striated terminal regions of myofibrils were observed in two previous studies of cultured cells. One was an electron-microscopic study of amphibian myotome (Peng et al. 1981) and the other was an immunofluorescence study of embryonic chick heart cells (Dlugosz et al. 1984). In both studies the non-striated terminals were interpreted to be regions of myofibril assembly. However, the possibility that the terminals represented instead sites of myofibril breakdown was not ruled out. The distribution of vinculin was not examined in these experiments and no changes in the length of the non-striated terminals were noted. It is possible that similar terminals in rat cells may provide a scaffold for the assembly of myofibrils under appropriate conditions. However, in our cultures of neonatal cells the thick, non-striated cables of actin appeared to elongate with time in culture, suggesting that they were being left in the wake of sarcomeric disassembly along myofibrils. Thus, sarcomeric assembly may be a reversible process, which somehow involves actin cables continuous with the terminals of myofibrils.

In vivo the muscle cells in the heart make contact with other muscle cells via intercalated discs and the myofibrils are striated all the way up to their terminals, where they attach to the membrane at fasciae adherentes junctions (McNutt, 1970). In the intact heart the distribution of vinculin is restricted to the fascia adherens (Tokuyasu et al. 1981) and rib-like deposits on the cytoplasmic side of the sarco-lemma in association with Z discs (Pardo et al. 1983). Precisely what role vinculin plays in these membrane-cytoskeletal interactions is not known since recent data indicate that vinculin does not bind actin (Wilkins & Lin, 1986). Therefore, it may be only indirectly involved in the attachment of thin filaments to the cell membrane.

In culture, vinculin was found not only in intercalated discs but also in extra-junctional sites in association with actin in non-striated terminals of myofibrils where they attach to the cell membrane at the free edges of muscle cells. When myocytes were unable to form junctional contacts with other muscle cells, they adapted by attaching the distal ends of myofibrils to membrane plaques that contained vinculin. The fact that the non-striated terminals of myofibrils were associated with focal contacts that contained vinculin suggests a structural and functional homology with focal adhesion contacts described in non-muscle cells such as fibroblasts. Actin cables in fibroblasts attach to the cell membrane at focal contact sites that contain vinculin (Geiger, 1979), alpha-actinin (Lazarides & Burridge, 1975), talin (Burridge & Connell, 1983), and a number of other proteins (Mangeat & Burridge, 1984). The plaques present in cultured rat heart myocytes have not been analysed for the presence of talin, and the immunofluorescent localization of alpha-actinin was inconclusive due to the poor cross-reactivity with rat tissue of antibodies to chicken alpha-actinin. It is noteworthy that vinculin and alpha-actinin have been shown to be present in the fascia adherens in intercalated discs in vivo, but talin is not a constituent of this structure (Geiger et al. 1985). Thus, the homology between the fascia adherens and focal adhesion contacts appears to be incomplete.

Although the membrane plaques and fascia adherens junctions appear to be structurally and functionally related membrane specializations, myofibrils in cultured cells that attached to the membrane via the fascia adherens appeared to be stable, while those that attached to the membrane at extrajunctional sites were not, as evidenced by their lack of sarcomeric organization. This suggests that fundamental differences exist between the two types of myofibrillar attachment sites, which may relate to changes in the membrane associated with the establishment of cell contacts. Apparently, the plaques lack structures or constituents present at cell-cell contact sites that are necessary for the maintenance or stabilization of the sarcomeric organization of myofibrils in rat cardiac muscle. Cell contacts do not appear to be necessary for the initial reassembly of myofibrils, since isolated single cells contain them. However, our results do suggest that cell-cell contact may have a critical role in stabilizing normal myofibrillar structure in cultured cardiac muscle cells. Therefore, it will be important to explore further structural and functional relationships between the two membrane specializations by extending this work to include ultrastructural and molecular analyses.

If our proposal that cell contact is important for the stability of myofibrils proves true, then it has major implications concerning the mechanisms that may underlie diseases affecting cardiac muscle, such as muscular dystrophy (Jasmin & Eu, 1979) and hypertrophic obstructive cardiomyopathy (Ferrans et al. 1972), where a breakdown or disorganization of the myofibrils is associated with detachment of the intercalated discs or altered cell-cell interactions.

This project was supported in part by BRSG S07RR 05370-24 awarded by the Biomedical Research Support Grant program, Division of Research Resources, the National Institutes of Health and by the Chicago Heart Association. The authors thank Dr Robert S. Decker for reading the manuscript and for the enthusiasm he expressed during the course of this work. We also thank Dr James Arnn for generously providing us with antibodies to desmoplakin.

Arnn
,
J.
(
1983
).
Dissertation, University of Colorado, Boulder, CO
.
Arnn
,
J.
&
Staehelin
,
L. A.
(
1981
).
The structure and function of spot desmosomes
.
Dermatology
20
,
330
339
.
Burridge
,
K.
&
Connell
,
L.
(
1983
).
A new protein of adhesion plaques and ruffling membranes
.
J. Cell Biol
.
97
,
359
367
.
Cedergren
,
B.
&
Harary
,
I.
(
1984
).
In vitro studies on single beating rat heart cells. VI. Electron microscopic studies of single cells
.
J. Ultrastruct. Res
.
11
,
429
442
.
Chacko
,
K. J.
(
1973
).
Observations on the ultrastructure of developing myocardium of rat embryos
.
J. Morph
.
150
,
681
710
.
Cowin
,
P.
,
Kapprell
,
H.
&
Franke
,
W. W.
(
1985
).
The complement of desmosomal plaque proteins in different cell types
.
J. Cell Biol
.
101
,
1442
1454
.
Decker
,
R. S.
,
Decker
,
M. L.
,
Thomas
,
V.
&
Fuseler
,
J. W.
(
1985
).
Responses of cultured cardiac myocytes to lysosomotropic compounds and methylated amino acids
.
J. Cell Sei
.
74
,
119
135
.
De Mello
,
W. C.
(
1982
).
Intercellular communication in cardiac muscle
.
Circulation Res
.
51
,
1
9
.
Dlugosz
,
A. A.
,
Antin
,
P. B.
,
Nachmias
,
V. T.
&
Holtzer
,
H.
(
1984
).
The relationship between stress fiber-like structures and nascent myofibrils in cultured cardiac myocytes. J’
.
Cell Biol
.
99
,
2268
2278
.
Ferrans
,
V. J.
,
Morrow
,
A. G.
&
Roberts
,
W. G.
(
1972
).
Myocardial ultrastructure in idiopathic hypertrophic subaortic stenosis. A study of operatively excised left ventricular outflow tract muscle in 14 patients
.
Circulation
45
,
769
792
.
Fischmann
,
D. A.
&
Moscona
,
A. A.
(
1971
).
Reconstruction of heart tissue from suspensions of embryonic myocardial cells: Ultrastructural studies on dispersed and reaggregated cells
. In
Cardiac Hypertrophy
(ed.
R.
Alpert
), pp.
125
139
.
New York
:
Academic Press
.
Geiger
,
B.
(
1979
).
A 130kd protein from chicken gizzard: its localization at the termini of microfilament bundles in cultured chicken cells
.
Cell
18
,
193
205
.
Geiger
,
B.
(
1981
).
Involvement of vinculin in contact-induced cytoskeletal interactions
.
Cold Spring Harbor Symp. quant. Biol
.
46
,
671
682
.
Geiger
,
B.
,
Schmid
,
E.
&
Franke
,
W. W.
(
1983
).
Spatial distribution of proteins specific for desmosomes and adherens junctions in epithelial cells demonstrated by double immunofluorescence microscopy
.
Differentiation
23
,
189
205
.
Geiger
,
B.
,
Volk
,
T.
&
Volberg
,
T.
(
1985
).
Molecular heterogeneity of adherens junctions
.
J. CellBiol
.
101
,
1523
1531
.
Izzard
,
C. S.
&
Lochner
,
L. R.
(
1976
).
Cell-to-substrate contacts in living fibroblasts. An interference reflexion study with an evaluation of the technique
.
J. Cell Sci
.
27
,
129
.
Jasmin
,
G.
&
Eu
,
H. Y.
(
1979
).
Cardiomyopathy of hamster dystrophy. Ann
.
N.Y. Acad. Sci
.
317
,
46
58
.
Jones
,
J. C. R.
&
Goldman
,
R. D.
(
1985
).
Intermediate filafnents and the initiation of desmosome assembly. J’
.
CellBiol
.
101
,
506
517
.
Lazarides
,
E.
&
Burridge
,
K.
(
1975
).
Alpha-actinin: immunofluorescent localization of a muscle structural protein in non-muscle cells
.
Cell
6
,
289
298
.
Legato
,
M. J.
(
1972
).
Ultrastructural characteristics of the rat ventricular cell grown in tissue culture with special reference to sarcomerogenesis.J
.
molec. cell. Cardiol
.
4
,
209
317
.
Mangeat
,
P.
&
Burridge
,
K.
(
1984
).
Actin-membrane interaction in fibroblasts: what proteins are involved in this interaction?
.
Cell Biol
.
99
,
95s
103s
.
Mcnutt
,
N. S.
(
1970
).
Ultrastructure of intercellular junctions in adult and developing cardiac muscle. Am
.
J. Cardiol
.
25
,
169
183
.
Mueller
,
H.
&
Franke
,
W. W.
(
1983
).
Biochemical and immunological characterization of desmoplakins I and II, the major polypeptides of the desmosomal plaque
.
J. molec. Biol
.
163
,
647
671
.
Overton
,
J.
(
1968
).
The fate of desmosomes in trypsinized tissue
.
J. exp. Zool
.
168
,
203
214
.
Pardo
,
J. V.
,
Siliciano
,
J. D.
&
Craig
,
S.
(
1983
).
Vinculin is a component of an extensive network of myofibril-sarcolemma attachment regions in cardiac muscle fibres
.
J. Cell Biol
.
97
,
1081
1088
.
Peng
,
H. B.
,
Wolosewik
,
J. J.
&
Cheng
,
P.
(
1981
).
The development of myofibrils in cultured muscle cells: a whole mount and thin-section electron-microscopic study
.
Devi Biol
.
88
,
121
136
.
Perjssel
,
B.
,
Charbonne
,
F.
,
Moalic
,
J. M.
&
Malet
,
P.
(
1980
).
Initial stages of trypsinized cell culture of cardiac myoblasts: Ultrastructural data
.
J. molec. cell. Cardiol
.
12
,
63
75
.
Pollinger
,
I.
(
1973
).
Identification of cardiac myocytes in vivo and in vitro by the presence of glycogen and myofibrils
.
Expl Cell Res
.
76
,
243
.
Severs
,
N. J.
(
1985
).
Intercellular junctions of the cardiac intercalated disc
.
Adv. Myocard
.
5
,
223
242
.
Sheridan
,
J. D.
&
Atkinson
,
M. M.
(
1985
).
Physiological roles of permeable junctions. A
.
Rev. Physiol
.
47
,
337
353
.
Tokuyasu
,
K. T.
,
Dutton
,
A. H.
,
Geiger
,
B.
&
Singer
,
S. J.
(
1981
).
Ultrastructure of chicken cardiac muscle as studied by double immunolabeling in electron microscopy
.
Proc. natn. Acad. Sci. U.SA
.
78
,
7614
7627
.
Wilkins
,
J. A.
&
Lin
,
S.
(
1986
).
A re-examination of the interaction of vinculin with actin
.
J. Cell Biol
.
102
,
1085
1092
.
Wollenberger
,
A.
(
1967
).
Rhythmic and arhythmic contractile activity of single myocardial cells cultured in vitro
.
Circulation Res
.
15
,
Suppl. 2
,
184
.