The distribution of adherens junction (AJ) components was investigated in cultured heart myocytes. These cells, derived from either newborn rats or chick embryos, develop elaborate arrays of myofibrils which become extensive and laterally aligned following several days in culture. The Z-disks in these cells, visualized by immunolabeling with antibodies to muscle-specific a- actinin, exhibit a characteristic periodicity of about 2 fan and are in register with those of neighboring myofibrils throughout the sarcoplasm. Vinculin, in these cells, associates with intercellular AJ and cell-matrix adhesions. In addition, this protein is detected in periodic bands located along the lateral cell membranes corresponding to “costamers” previously described by Pardo, J.V., Siliciano, J.D. and Craig, S.W. (Proc. Natn. Acad. Sci. USA, 80,1008). Similarly, N-cadherin, which is predominantly associated with intercellular junctions, is also detected in periodic striations located mainly on the dorsal and lateral cell surfaces. Using computer-aided three-dimensional microscopy confirmed that these vinculin- and N-cadherin-containing structures are located in extrajunctional sites, appar-ently associated with Z-disks of peripheral myofibrils. Based on these findings an alternative pathway is proposed for the assembly of vinculin and N-cadherin, which is not triggered by adhesive interactions with extracellular surfaces but rather by interactions at the membrane-cytoplasm interphase with the periphery of the pre-assembled myofibrills. Moreover, we present evidence that antibodies to N-cadherin, which are capable of blocking AJ formation in culture, have an inhibitory effect also on the development and alignment of myofibrils. We discuss the functional significance of the “costameric” organization of vinculin and N-cadherin and consider its involvement both in the lateral alignment of neighboring muscle cells and in the stabilization of developing myofibrils.

The assembly of the contractile apparatus of cardiac muscle involves several developmental stages including the interaction between neighboring myoblasts via prominent fascia and maculae adherentia, progressive assembly of thin and thick filaments into cytoplasmic arrays and the subsequent appearance of aligned sarcomeric structures with characteristic and uniformly spaced Z-disks and M-lines (Fischman, 1967, 1970). Detailed examination of this process focused much attention on the interaction between myofibrils and the plasma membrane both during myofibrillogenesis and in the mature cells. It had been reported that nascent myofibrils assemble mostly in close proximity to the sarcolemma and then expand until they occupy most of the cell volume (Holtzer et al., 1959; Kelly, 1969; Dlugasz et al., 1984). The peripheral myofibrils appar-ently interact with the membrane via an electron-dense plaque material while the more “medullary” myofibrils are associated with each other and with the sarcoplas-mic reticulum, mostly at the level of the Z-disks (Cheisi et al., 1981). It had further been proposed that peripheral components of the Z-disk, including desmin-containing intermediate filaments and spectrin take part in these intermyofibrillar or myofibril-membrane as-sociations (Granger and Lazarides, 1979; Lazarides, 1980; Goodman et al., 1981; Messina and Lemanski, 1989; Nelson and Lazarides, 1984). Cultured heart muscle cells have become, over the last decade, a most useful system for the study of myofibril assembly at both structural and molecular levels (Claycomb and Palazzo, 1980; Nag and Cheng, 1981). The examination of such cultures by immunofluorescence microscopic labeling revealed the spatial and temporal assembly of the different molecular components of myofibrils into the mature structure. This approach included the localiz-ation of a multitude of proteins such as actin, myosin, α-actinin, tropomyosin, titin, tubulin and vinculin (Nag et al., 1983 ; Wang et al., 1988 ; Sanger et al., 1984 ; Terai et al., 1989; Guo et al., 1986, Lin et al., 1989; Schultheiss et al., 1990).

Especially intriguing was the distribution of vinculin in striated muscle. This 117×103Mr protein was previously shown to be ubiquitously associated with the cytoplasmic faces of all adherens type junctions (AJ) including intercellular junctions and cell-matrix ad-hesions (Geiger et al., 1981, 1985,1987; Burridge, 1986; Burridge et al., 1990). In addition to its prominent association with intercellular junctions and with the termini of myofibrils, vinculin in muscle cells was also detected in periodically spaced “costameric structures” located at the vicinity of Z-disk attachment sites to the plasma membrane (Pardo et al., 1983a and b; Kote-liansky and Gneushev, 1983). Moreover, in view of the apparent involvement of vinculin in the membrane anchorage of actin filaments in cell junctions (Geiger, 1979; Geiger et al., 1981), it was assumed that this protein plays a similar role also in linking the periphery of Z-disks to the sarcolemma (Pardo et al., 1983b; Koteliansky and Gneushev, 1983).

To define the relationships between myofibril as-sembly and membrane anchorage and to determine the involvement of AJ in these processes, we have examined the organization of two AJ components, vinculin and N-cadherin (A-CAM, Hatta et al., 1988; Geiger et al., 1990a), during various phases of cardiac myocyte development in culture. We have found that in addition to its prominent association with intercellular “fascia adherens-Like” junctions (Tokuyasu et al., 1981), vinculin is detected along Z-disks of peripheral myofibrils which are closely associated with the cell membrane. This became evident from the examination of three-dimensional reconstructions of cultured muscle cells, double labeled for vinculin and actin. In addition, we observed a prominent association of N-cadherin with these “costameric” structures along peripheral Z-disks located mostly at the dorsal and lateral cell surfaces. Moreover, addition of N-cadherin antibodies against N-cadherin to cultured heart myocytes resulted in a marked inhibition of myofibrillogenesis. The basis for the localization of vinculin and N-cadherin in these extrajunctional areas and their possible involvement in muscle organization are discussed.

Antibodies

  1. Monoclonal antibodies against A-CAM (N-cadherin) used here included ID 7.2.3 (Volk and Geiger, 1984) or GC-4 (Volk et al., 1990), both are now available from Sigma Immuno-chemicals (St Louis, USA).

  2. Anti-vinculin used here was a monoclonal antibody raised against human vinculin, kindly provided by Dr V. Koteliansky from the Institute of Experimental Cardiology, Moscow, USSR.

  3. Rabbit polyclonal antibodies against desmin were purchased from Sigma Immunochemicals (St Louis, USA).

  4. Monoclonal antibodies specific for muscle a-actinin were as previously described (Fridlanskaya et al., 1989).

  5. Secondary antibodies used here were DTAF- and rhodamine-lissamine B-modified goat anti-mouse or -rabbit IgG (Jackson ImmunoResearch Labs PA, USA).

Cell culture

Heart muscle cell cultures were prepared from the ventricles of 1- to 3-day-old rats (Wistar strain) or of 7-day-old chicken embryos, using modification of previously described pro-cedures (Nag and Cheng, 1981; Dlugasz et al., 1984). Ventricles were isolated under sterile conditions and washed in Hank’s Ca2+-free solution and again in phosphate-buffered saline (PBS). The tissue was minced and transferred into PBS containing 0.1% trypsin (Difco, USA) for 15–20 minutes at 37°C. After incubation with trypsin solution, the supernatants were discarded and the tissue further processed as follows. For the preparation of rat cardiac myocyte cultures, fresh PBS containing 0.1% trypsin and 0.1% collagenase 1A (Sigma, USA) was added to the tissue for 20 minutes. For chicken embryos, PBS containing 0.1% trypsin and 0.05% collagen-ase was added for 15 minutes. After exposure to the enzyme solutions, cells were collected by centrifugation (3–5 minutes at 1000 rpm in a table top centrifuge) and washed once in Dulbecco”s Modified Eagle”s medium (DMEM) containing 10% fetal calf serum, 2 mM glutamine, 1% antibiotic-antimycotic solution (all purchased from Bio-Lab, Israel). Cells were gently resuspended in new medium and centri-fuged for 2–3 minutes to remove large tissue fragments. Supernatants containing suspended cells were plated on 18 mm coverslips. Fresh enzyme solutions were added to the remaining undissociated tissue and the procedure was re-peated 3–4 times. Cultures were incubated at 37°C in humidified atmosphere of 5% CO2 and 95% air.

Immunofluorescence labeling

For immunofluorescence labeling with antibodies to a- actinin, vinculin, or N-cadherin (A-CAM), cells were per-meabilized at room temperature by 2 minutes exposure to 0.5% Triton X-100 in 50 mM morpholinoethane sulfonate buffer, pH 6.0 and fixed for 30 minutes with 3% paraformal-dehyde. For immunofluorescence labeling with anti-desmin antibodies, cells were fixed and permeabilized for 10 minutes with cold methanol (—20°C) and then for 1 minutes with cold acetone (–20°C). Incubations with antibodies were carried out on Parafilm sheets for 30 minutes at room temperature. Cells were rinsed in PBS after the incubations with the primary and secondary antibodies and mounted in Elvanol.

Immunofluorescence microscopy and three-dimensional reconstructions

Conventional immunofluorescence microscopy and photo-micrography were carried out using Axiophot microscope (Zeiss, Oberkochen, FRG) equipped with filter sets for selective rhodamine/fluorescein fluorescence.

The computerized microscopic system used here was based on the design of Agard and Sedat (Agard et al., 1988, 1989). The system consisted of a Zeiss Axiomat microscope and microvax III workstation which controls image acquisition, light shutters and focus. Low-light-level images were recorded with a cooled, scientific-grade, charge coupled device camera (CCD, Photometries, Arizona, USA, for details see Hiraoka et al., 1987). Images were read into a Mercury Array Processor (Mercury computer systems, Lowel MA. USA) which calculated on-the-fly pixel per pixel correction for illumination and CCD sensitivity, scaled and deconvoluted images essentially in real-time (Chen et al., 1990). Optical sections of double-labeled slides were taken in the fluorescein and rhodamine channels with selective Zeiss filter sets.

The three-dimensional imaging and modeling package (PRISM, Chen et al., 1990, Kam et al., 1991) employs high-resolution (1280×1024, 60 Hz, non-interlaced) video board with 12 Mbytes of memory (Parallax Graphics, Inc. Santa Clara, CA USA), and implements a flexible multiple windowing architecture and interactive display of images series. Versatile color menus allow images of different fluorophore staining of the same sample to be superimposed. A digital movie of sequentially displayed tilt series produce a realistic perception of the three-dimensional distributions of the labeled molecules, and creates a powerful medium for studying complex colocalization of molecules. Pairs of images from such rocking projection series were selected for stereo presentations in this work. The three-dimensional aspect of the data is enhanced by three- to five-fold increase in the separation between sections. Pictures were photographed on Ektachrome 160 or 100/Day light diapositive film using Focus 4700 Imagecorder (Focus Graphics, CA. USA).

Development of myofibrillar structure in cultured cardiac myocytes

In this study, we have examined heart cells from two sources, namely rat and chick, which develop in culture at different rates, and with distinct morphology. Since development of cultured cardiac myocytes was pre-viously described, we will refer here only to major stages in myocyte development. Isolation of cardiac myocytes of 7 day chicken embryos and their transfer to culture resulted, initially, in a dramatic deterioration of sarcomeric structures. Up to the third day after plating only some myofibrils were found in the cytoplasm while from the third day and on, a progressive increase was noted in the number of striated myofibrils, their size and their alignment (Dlugasz et al., 1984; Sanger et al., 1986; see also Fig. 1). At first stress fiber-like structures were frequently noted by phalloidin staining which were decorated by bead-shaped α-actinin-containing structures. These cells were clearly identified as devel-oping muscle cells in as much as they were labeled with the muscle-specific anti-α-actinin antibody and often detected in the same cells as organized myofibrils. The early myofibrils, identified by the typical periodicity of ∽2 μm, were of variable size and usually poorly aligned. Following longer incubation in culture, the stress fiber-like filament arrays became less prominent concomitantly with a marked increase in myofibril alignment and a moderate increase in their number and size, with α-actinin located with their Z-disk.

Fig. 1.

Immunofluorescent localization of actin (A,C) and vinculin (B,D) in chicken cardiomyocytes cultured for 3 days (A,B) or 6 days (C, D). Note the presence of vinculin both in focal contacts and in an apparent association with some of the Z-disks. The matching paired-arrows point to regions along the same myofibril that exhibit highly variable levels of vinculin staining. The bar indicates 10 μm.

Fig. 1.

Immunofluorescent localization of actin (A,C) and vinculin (B,D) in chicken cardiomyocytes cultured for 3 days (A,B) or 6 days (C, D). Note the presence of vinculin both in focal contacts and in an apparent association with some of the Z-disks. The matching paired-arrows point to regions along the same myofibril that exhibit highly variable levels of vinculin staining. The bar indicates 10 μm.

In rat cardiomyocyte cultures, discrete sarcomeric structures were noted already one day following plating, though the cells were usually small and the myofibrils short and poorly aligned (see Fig. 2A). Upon longer incubation in culture, the cells underwent considerable spreading and differentiation showing conspicuous increase in the content and alignment of myofibrils (Fig. 2C). At the termini, myofibrils were often associated with stress-fiber-like actin-containing structures displaying beaded α-actinin distribution.

Fig. 2.

Immunofluorescent localization of actin (A,C), vinculin (B,D,F) and desmin (E) in rat cardiomyocytes cultured for 3 (A,B,E,F) or 6 days (C,D). Notice the prominent association of vinculin labeling with focal contacts, intercellular AJ and Z-disks at both early and advanced stages of development and the more extensive labeling for desmin along the latter two areas (the matched arrows in E and F point to desmin-containing Z-disks which appear to be devoid of vinculin). The bar indicates 10 μm.

Fig. 2.

Immunofluorescent localization of actin (A,C), vinculin (B,D,F) and desmin (E) in rat cardiomyocytes cultured for 3 (A,B,E,F) or 6 days (C,D). Notice the prominent association of vinculin labeling with focal contacts, intercellular AJ and Z-disks at both early and advanced stages of development and the more extensive labeling for desmin along the latter two areas (the matched arrows in E and F point to desmin-containing Z-disks which appear to be devoid of vinculin). The bar indicates 10 μm.

These filament bundles terminated at the ventral cell surface, in an apparent association with focal contacts (see also Guo et al., 1986 and Fig. 2).

Organization of vinculin in cultured cardiac myocytes

At early stages of myofibrillogenesis in the chick (3 days in culture, Fig. 1A,B), close association between actin-containing fibers and vinculin was noted. Unlike non-muscle mesenchymal cells, in which vinculin was associated predominantly, with focal contacts located at the termini of stress fibers (the arrowheads in Fig. 1A,B point to focal contact-associated vinculin in a fibroblast present in the culture) prominent beaded vinculin staining was detected along the developing myofibrils with an apparent average periodicity of about 1 μm (Fig. 1A,B, arrows). Upon longer incubation (6 days), vinculin staining was detected both at the termini of myofibrils and along their length where it was associ-ated with many (but apparently not all) Z-disks (Fig. 1C,D). This was verified by careful alignment of the actin- and vinculin-labeling (see for example matched arrow in Fig. 1C,D). It is noteworthy that vinculin antibodies did not react with other structures period-ically located along the myofibrils.

The apparent association of vinculin with Z-disk area was even more conspicuous in rat cardiomyocyte cultures. The presence of vinculin in both substratum adhesions and Z-disks was already clear following 3-day in culture (Fig. 2A,B) and became more extensive after incubation for additional 3 days (Fig. 2C,D). Often, when groups of cells were examined, additional labeling was noticed along the intercellular junctions formed between adjacent cardiac cells (two muscle cells or one muscle cell and one fibroblast) adopting a typical “intercalated disc-like” organization (Fig. 2F). Ad-ditional analysis indicated that another component of the Z-disk periphery, namely desmin was already associated with Z-disks following 3-days in culture. Comparison of desmin staining to that of α-actinin and vinculin suggested that a-actinin is associated with Z-disks earlier than the other two proteins and that desmin staining along Z-disks appears earlier and is more extensive than that of vinculin (Fig. 2E,F). It was however clear that, whenever present, the labeling of desmin and vinculin was overlapping.

To explore the possibility that vinculin is selectively associated only with Z-disks of “peripheral” myofibrils (i.e those located immediately subjacent to the plasma membrane), we have determined the relative distri-butions of vinculin and actin in the same developing muscle cells by examination of optical sections pro-duced by computer-aided microscopy. We have selected to examine cultured rat cardiomyocytes after 3 days in culture. At that stage, prominent association of vinculin with the Z-disk area was already detected, yet the cells were moderately spread and thus allowed for distinction between myofibrils located at different focal levels. As evident from the optical sections of the double-labeled cell (Fig. 3), most of the ventral actin-containing myofibrils were associated with periodic vinculin striations (Fig. 3A,B) while in sections located at the more dorsal levels, most myofibrils (except for the most peripheral ones) were cleared of vinculin (Fig. 3C,D).

Fig. 3.

Optical sections of 3-day cultured rat cardiomyocytes, double-labeled for vinculin (A,C) and actin (B,D). The focal planes selected are ventral (A,B), and close to the dorsal surface (C,D). Notice that the majority of actin-containing ventral myofibrils are associated with vinculin at their termini and along their sarcomere, while closer to the dorsal surface only cortical myofibrils are positive for vinculin. Bar indicates 10 μm.

Fig. 3.

Optical sections of 3-day cultured rat cardiomyocytes, double-labeled for vinculin (A,C) and actin (B,D). The focal planes selected are ventral (A,B), and close to the dorsal surface (C,D). Notice that the majority of actin-containing ventral myofibrils are associated with vinculin at their termini and along their sarcomere, while closer to the dorsal surface only cortical myofibrils are positive for vinculin. Bar indicates 10 μm.

Distribution of N-cadherin (A-CAM) in developing chicken cardiomyocytes

Immunolabeling of cultured chick myocytes with anti-N-cadherin antibodies yielded intense staining of intercellular contacts, in accordance with our previous results (Volk and Geiger, 1984). This junctional staining pattern was detected already one day after plating (Fig. 4A) and persisted in the same culture for 7-10 days. However, it was noted that, in addition to the labeling at cell-cell contacts, periodically spaced spots and bands were also present, located along the ventral, dorsal or lateral surfaces of the cells. This was first noted at 3 days after plating (Fig. 4B) and became more prominent by 6 days (Fig. 4C).

Fig. 4.

Immunofluorescent localization of N-cadherin in chicken cardiomyocytes after 1 (A), 3 (B) and 6 days (C) in culture. Initially, labeling is restricted to cell-cell contacts whereas at more advanced stages punctate staining, corresponding to sarcomeric periodicity, is evident (see areas indicated by arrows in B and C). Bar represents 10 μm.

Fig. 4.

Immunofluorescent localization of N-cadherin in chicken cardiomyocytes after 1 (A), 3 (B) and 6 days (C) in culture. Initially, labeling is restricted to cell-cell contacts whereas at more advanced stages punctate staining, corresponding to sarcomeric periodicity, is evident (see areas indicated by arrows in B and C). Bar represents 10 μm.

To obtain a more accurate view of the spatial relationship between N-cadherin and actin, we have examined cells displaying different stages of myo-fibrillar organization using the computer-aided three-dimensional microscopy. Fig. 5 shows cultured heart muscle cells which were fixed before extensive spread-ing had occurred. In these cells, some actin bundles and primordial myofibrils are detected throughout the cytoplasm. The N-cadherin labeling appears in two major forms in these cells: long, nearly continuous-line running along the cell-cell junction and spots, scattered over the cell periphery. Examination of the overlaps between cadherin and actin labeling (represented by overlapping blue and yellow areas which appear white in these micrographs) indicated that most of these cadherin-containing spots were, in fact, associated with actin. On the other hand, the myofibrils which were not located at the cell periphery were apparently devoid of N-cadherin labeling.

Fig. 5.

Computer-generated stereoscopic view of 2-day chick cadiomyocytes, double-labelled for actin (blue) and N-cadherin (yellow). Notice that N-cadhenn is predominantly associated with actin in the intercellular junction with only sporadic patches along the cell surface.

Fig. 5.

Computer-generated stereoscopic view of 2-day chick cadiomyocytes, double-labelled for actin (blue) and N-cadherin (yellow). Notice that N-cadhenn is predominantly associated with actin in the intercellular junction with only sporadic patches along the cell surface.

As pointed out above, at more advanced stages of development, N-cadherin-labeled striations become apparent. The relationship of these structures to the myofibrillar network was determined by computerized optical sectioning analysis as shown in Fig. 6, as well as by careful examination of flatly spread cells in which the images were essentially two dimensional (Fig. 7). The three-dimensional reconstitution of the N-cadherin/ac-tin double-labeled cells is shown in the bottom panel (Fig. 6G,H). Stereoscopic examination showed that cadherin labeling (which appears yellow or white -when superimposed on the “blue” actin) is mostly restricted to the periphery of the cells and is present in the form of long streaks located at cell-cell contact regions, scattered spots and uniformly spaced striations with ∽2 μmperiodicity. The latter were often located on the dorsal cell surface with no apparent relationship to intercellular junctions. Since the reconstituted images of whole cells seem too complex for detailed spatial analyses, we have selected a narrower range of optical sections located near either the ventral or the dorsal focal planes of the cells. Examination of the superimposed images (Fig. 6B,E) displaying N-cad-herin (yellow) (Fig. 6A,D) and actin (blue) (Fig. 6C,F) indicates that the N-cadherin labeling close to the ventral area was mostly restricted to cell-cell contacts while in dorsal areas these striations were apparently associated with the control region of the I-band along the sarcomeric actin. This notion was further supported by two lines of evidence. Linear computer scanning along the myofibrillar axis confirmed that the periodici-ties of actin and N-cadherin were 2.1 ± 0.2 and 1.93 ± 0.2 μm, respectively. There was however large differ-ence in the width of the labeled bands. The average of full width of actin bands at half peak intensity was 1.15 ± 0.2 μm while the average width of cadherin bands was only 0.47 ± 0.2 μm. High resolution examination of cultured chick cells labeled for N-cadherin and either α- actinin or actin confirmed that the cadherin labeling was indeed associated with the Z-disc area (Fig. 7).

Fig. 6.

Computer-generated optical sections (A-F) and stereoscopic new (G,H) of differentiated chick cadiomyocytes, double-labelled for actin (blue) and N-cadherin (yellow). The free dorsal cell surface is depicted in A-C, showing overlapping stnanons (white in panel B) containing both proteins along peripheral myofibrils. At the ventral focal plane (D-F), close to the substratum, N-cadherin is apparently absent from the myofibrils and detected mostly at cell borders. The three-dimensional relationships between the two proteins are displayed in the stereopair (G,H). Distribution of N-cadherin in chick cardiac muscle indicates periodic association of the label with the periphery of Z-disks (I). Occasionally, when two adjacent cells retracted during processing, it became evident that the N-cadherin patches in neighboring cells are in register (J).

Fig. 6.

Computer-generated optical sections (A-F) and stereoscopic new (G,H) of differentiated chick cadiomyocytes, double-labelled for actin (blue) and N-cadherin (yellow). The free dorsal cell surface is depicted in A-C, showing overlapping stnanons (white in panel B) containing both proteins along peripheral myofibrils. At the ventral focal plane (D-F), close to the substratum, N-cadherin is apparently absent from the myofibrils and detected mostly at cell borders. The three-dimensional relationships between the two proteins are displayed in the stereopair (G,H). Distribution of N-cadherin in chick cardiac muscle indicates periodic association of the label with the periphery of Z-disks (I). Occasionally, when two adjacent cells retracted during processing, it became evident that the N-cadherin patches in neighboring cells are in register (J).

Fig. 7.

Double immunofluorescent localization of N-cadherin (A,C) and either o’-actin (B) or actin (D). The arrows point to identical location. Notice that N-cadherin is associated with sharply delineated bands, superimposed on <r-actinin striations, and the central region of the I-bands. Bar represents 10 μm.

Fig. 7.

Double immunofluorescent localization of N-cadherin (A,C) and either o’-actin (B) or actin (D). The arrows point to identical location. Notice that N-cadherin is associated with sharply delineated bands, superimposed on <r-actinin striations, and the central region of the I-bands. Bar represents 10 μm.

Periodic organization of N-cadherin at or near Z-disk-membrane contacts is not exclusively present in cultured heart cells. In fact, in the first report on A-CAM distribution, we have noted that, in heart tissue, this protein was associated not only with intercalated disks but also with lateral membranes, where it often displayed a typical sarcomeric periodicity and was apparently in register with peripheral Z-disks (Volk and Geiger, 1984, see Fig. 5A). An example of frozen sections of adult chicken heart following immunofluor-escent labeling with anti-A-CAM/N-cadherin anti-bodies is shown in Fig. 61,J, indicating that this protein is indeed present along Z-disks, supporting the obser-vations made with cultured myocytes.

Effect of anti-N-cadherin on myofibrillogenesis in culture

In view of the apparent association of N-cadherin with the periphery of Z-disks, we have explored the possibility that this protein is involved in myofibrillo-genesis. For that purpose chicken cardiac myocytes were plated on glass coverslips in the presence of purified IgG fraction of monoclonal anti-A-CAM antibody (clone GC4), which was previously found to be an effective inhibitor of adherens junction formation in cultured chick lens epithelial cells (Volk and Geiger, unpublished results). As a control, we have used class-matched monoclonal antibodies directed against IgE. In both cases, the IgG fraction was isolated by DEAE-cellulose chromatography and used at final concen-trations of 0.2-0.4 mg/ml. The cells were incubated in culture for 2 days, then fixed and immunolabeled for μ- actinin to examine the extent of myofibrillogenesis.

As shown in Fig. 8A, incubation with the control antibodies, even at a relatively high concentration, did not impair the development of elaborate myofibrillar arrays. Anti-A-CAM, on the other hand, strongly inhibited the development of a-actinin containing myofibrils. This was manifested by poor alignment of the filament bundles and the presence of “beaded” rather than sarcomeric structures. This effect appeared to depend on the concentration of inhibitory antibodies (compare Fig. 8B,C) and was readily reversible: replacement of anti-A-CAM with regular medium resulted in the re-acquisition of typical myofibrillar organization within 24 hours (Fig. 8D).

Fig. 8.

Effect of anti-N-cadherin antibodies on myofibril development. Chicken cardiomyocytes were plated in culture in the presence of 0.4 mg/ml of an irrelevant mouse IgG (A), or anti-N-cadherin (0.2 mg/ml-B or 0.4 mg/ml-C). Following two days in culture, cells were fixed and labeled for a-actinin. Notice that in the control cells a-actinin was present both along the Z-disks and intercellular junctions (double-headed arrow in A while myofibril assembly was markedly inhibited by the treatment with anti-N-cadherin as apparent from the reduction in the number and size of Z-disks and the presence of a-actinin-containing “beaded” filaments (arrow in B). This inhibition was reversible since fully developed Z-disks appeared within one day after removal of the inhibitory antibodies (D). The bar indicates 10 μm.

Fig. 8.

Effect of anti-N-cadherin antibodies on myofibril development. Chicken cardiomyocytes were plated in culture in the presence of 0.4 mg/ml of an irrelevant mouse IgG (A), or anti-N-cadherin (0.2 mg/ml-B or 0.4 mg/ml-C). Following two days in culture, cells were fixed and labeled for a-actinin. Notice that in the control cells a-actinin was present both along the Z-disks and intercellular junctions (double-headed arrow in A while myofibril assembly was markedly inhibited by the treatment with anti-N-cadherin as apparent from the reduction in the number and size of Z-disks and the presence of a-actinin-containing “beaded” filaments (arrow in B). This inhibition was reversible since fully developed Z-disks appeared within one day after removal of the inhibitory antibodies (D). The bar indicates 10 μm.

In this study, we have examined the subcellular distribution of the AJ components vinculin and N-cadherin in cultured cardiomyocytes at different stages of myofibrillogenesis. The observations described here indicated that these molecules are present in two distinct cellular sites: the first are intercellular AJ and the other are membrane regions associated with cortical myofibrils.

The presence of AJ in heart muscle has been amply documented in recent years. It had been shown that vinculin, α-actinin and A-CAM (N-cadherin) are enriched along the fascia adherens in cardiomyocytes both in culture and in situ (Tokuyasu et al., 1981; Koteliansky and Gneushev, 1983; Volk and Geiger, 1984; Geiger et al., 1990a). The assembly of these and other components into a mature AJ has not been well characterized at the molecular level. Yet it had been widely accepted that this process is initiated by immobilization of the relevant receptors (cadherins in cell-cell AJ) followed by assembly of the submembrane plaque and cytoskeletal proteins (Geiger et al., 1985, 1987, 1990b). This was mainly based on the observation that AJ-specific molecules are rarely detected in extrajunctional sites and that cell-cell and cell-matrix adhesions appear to act as nucleation sites for the organization of the microfilament system. The only observation that was apparently inconsistent with this view is the finding of organized vinculin in extrajunctio-nal “rib shaped” sites in muscle cells known as “costameres” (Pardo et al., 1983a,b, see also Fig 5b in Schultheiss et al., 1990). The mechanism of costamer formation and their physiological roles are, however, still obscure. In the present study, we have extended the characterization of costamers and showed that not only cytoplasmic components of AJ such as vinculin but also the respective membrane receptor, namely N-cadherin, becomes associated with myofibrillar structures in differentiating cardiomyocytes. This organized N-cad-herin, displayed as regularly spaced dots or bands, had a rather uniform periodicity of about 2 μm, which was essentially identical to that found in muscle sarcomeres. Furthermore, it was noted that N-cadherin was colocal-ized with only a limited fraction of the myofibrils. It was therefore important to determine what are the special features of that subpopulation of myofibrills which colocalized with N-cadherin and to which substructures along the sarcomere was the cadherin apparently bound.

To address the former issue, we have used the computerized three-dimensional microscopic examin-ation of cells, double labeled for both actin and N-cadherin. Careful plane-by-plane examination of these specimens, along with the phase-contrast or Nomanski images, revealed the relative topologies of the two molecule. Their spatial interrelationships indicated that only myofibrils located at the cell periphery exhibit such association with cadherin and that the transverse N-cadherin bands were largely in register with Z-disks. Electron microscopic analyses are now in progress to assess these relationships at a molecular resolution. It is noteworthy that while this co-distribution was studied here primarily in differentiating cultured cells examin-ation of cadherin distribution in intact chicken cardiac muscle using either A-CAM-specific or pan-cadherin antibodies indicated that prominent labeling was located at the periphery of Z-disks (Fig. 8, see also Fig 5a in Volk and Geiger, 1984, and Figs 4 and 5 in Geiger et al., 1990).

The molecular mechanism leading to this extrajunc-tional organization of N-cadherin is not clear. However, it is conceivable that the molecular interactions that link together the various constituents in intercellular AJ are also responsible for these extrajunctional assemblies.

Biochemical studies characterizing these interactions are mostly available for the cell-matrix AJ where it was shown that integrin can bind talin (Horwitz et al., 1986), or a-actinin talin may bind to vinculin (Burridge and Mangeat, 1984) and vinculin to α-actinin (Otto, 1983; Wilkins et al., 1983; Belkin and Koteliansky, 1987; Wachsstock et al., 1987) and paxillin (Turner et al., 1990) which is believed to be regulatory molecule, a-actinin has been known, for a long time, to bind and cross-link actin (Maruyama and Ebashi, 1965) and was recently shown to directly bind to integrin (Otey et al., 1990) and to another AJ constituent, zyxin (Crawford and Beckerle, 1991). It is likely that a similar chain of interactions exists also in cell-cell AJ (see, for Dis-cussion: Geiger and Ginsberg, 1991). These binding data left, however, unresolved the intriguing question: how is the directionally of AJ assembly determined? In the common molecular models for AJ formation, it was proposed that surface interactions, mediated by cadher-ins (in cell-cell AJ), or integrins (in cell-matrix AJ), lead to the sequential local accretion of the cytoplasmic components of the junctional plaque (such as vinculin) and subsequently to the assembly of actin and associ-ated proteins at the junctional site. The present results, on the other hand, raise the possibility that the organization of AJ constituents may be a bidirectional process, namely triggered not only by proper nucleation sites provided by extracellular multivalent ligands (a contact with an external surface), but also by pre-existing cytoskeletal structures. The presence of such “reverse assembly” is suggested in the present studies by the progressive organization of N-cadherin in non-junctional regions where no external organizer is detected. It is thus conceivable that the localization of both vinculin and N-cadherin in such sites, in close proximity to Z-disks of peripheral myofibrils is driven by a chain of interactions directed from pre-assembled intracellular structures towards the periphery.

Does the regular organization of N-cadherin along the lateral borders of cardiac myocyte have a functional significance? One tempting speculation is that the presence of “contact receptors”, which are believed to be capable of mediating homophilic cell-cell adhesions, may contribute to the proper register of laterally neighboring muscle cells. Such paracellular alignment, effectively bridging between the peripheries of Z-disk in neighboring cells, may play an important role in efficient and coordinated contractility. Furthermore, the inhibition experiments shown in Fig. 8 support the notion that surface-associated N-cadherin may indeed play an important role not only in the establishment of intercellular contacts but also in myofibrillar organization.

This study was supported by grants from the Ch. Revson Foundation and the DKFZ-NCRD Israeli-German program. B.G. holds the E. Neter Chair in Cell and Tumor Biology. Z.K. is the Israel Pollak Professor of Biophysics.

Agard
,
D. A.
,
Hlraoka
,
Y.
and
Sedat
,
J. W.
(
1988
).
Three-dimensional light microscopy of diploid Drosophila chromosomes
.
Cell Motility and the Cytoskeleton
10
,
18
27
.
Agard
,
D. A.
,
Hlraoka
,
Y.
,
Shaw
,
P.
and
Sedat
,
J. W.
(
1989
).
Fluorescence microscopy in three dimensions
.
Meth, in Cell Biol
. Vol.
30
, Ch.13, pp.
353
377
.
Belkin
,
M.
and
Koteliansky
,
V. E.
(
1987
).
Interaction of iodinated vinculin, metavinculin and a-actinin with cytoskeletal proteins
.
FEBS Lett
.
220
,
291
294
.
Burridge
,
K.
(
1986
).
Substrate adhesions in normal and transformed fibroblasts: Organization and regulation of cytoskeletal, membrane and extracellular matrix components at focal contacts
.
Cancer Rev
.
4
,
18
78
.
Burridge
,
K.
and
Mangeat
,
P.
(
1984
).
An interaction between vinculin and talin
.
Nature
308
,
744
746
.
Burridge
,
K.
,
Nuckolls
,
G.
,
Otey
,
C.
,
Pavalko
,
F.
,
Simon
,
K.
and
Turner
,
C.
(
1990
).
Actin-membrane interaction in focal adhesion
.
Cell Diff. Dev
.
32
, (in press).
Chen
,
H.
,
Sedat
,
J. W.
and
Agard
,
D. A.
(
1990
).
Manipulations, display, and analysis of three-dimensional biological images
.
Handbook of Biological Confocal Microscopy
, (ed. J.B. Pawley) Ch.13 pp.141-150.
New York
:
Plenum Press
.
Chiesi
,
M.
,
Ho
,
M. M.
,
Inesi
,
G.
,
Somlyo
,
A. V.
and
Somlyo
,
P.
(
1981
).
Primary role of sarcoplasmic reticulum in phasic contractile activation of cardiac myocytes with shunted myolemma
.
J. Cell Biol
.
91
,
728
742
.
Claycoinb
,
W. C.
and
Palazzo
,
M. C.
(
1980
).
Culture of the terminally differentiated adult cardiac muscle cell: a light and scanning electron microscope study
.
Devi. Biol
.
80
,
466
482
.
Crawford
,
A. W.
and
Beckerle
,
M. C.
(
1991
).
Purification and characterization of zyxin, an 82,000 dalton component of adherens junctions
.
J. Biol. Chem
.
266
,
5847
5853
.
Dlugasz
,
A. A.
,
Antin
,
P. B.
,
Nachimias
,
V. T.
and
Holtzer
,
H.
(
1984
).
The relationship between stress fiber-like structures and nascent myofibrils in cultured cardiac myocytes
.
J. Cell Biol. 99,2268-2278
.
Fischman
,
D. A.
(
1967
).
An electron microscope study of myofibril formation in embryonic chick skeletal muscle
.
J. Cell Biol
.
32
,
557
575
.
Fischman
,
D. A.
(
1970
).
The synthesis and assembly of myofibrils in embryonic muscle
. In
Current Topics in Developmental Biology
, Vol.
5
(ed.
A.A.
Moscona
).
New York
:
Academic Press
.
Fridlanskaya
,
L.L.
,
Goncharova
,
E. I.
,
Borisov
,
A. B.
,
Krylova
,
T. A.
and
Pinaev
,
G. F.
(
1989
).
The monoclonal antibody to the muscle a-actinin as a marker for the skeletal and cardiac myogenesis
.
Tsitologia
31
,
1234
1237
.
Geiger
,
B.
(
1979
).
A 130K protein from chicken gizzard; its localization at the termini of microfilament bundles in cultured chicken cells
.
Cell
18
,
193
205
.
Geiger
,
B.
,
Dutton
,
A.
,
Tokuyasu
,
K. T.
and
Singer
,
S. J.
(
1981
).
Immunoelectron microscope studies of membrane-microfilament interactions: Distributions of O’-actinin, tropomyosin, and vinculin in interstitial epithelial brush border and chicken gizzard smooth muscle cells
.
J. Cell Biol
.
91
,
614
628
.
Geiger
,
B.
and
Ginsberg
,
D.
(
1991
).
The cytoplasmic domain of adherens-type junctions
.
Cell Motil. Cytoskel., in press
.
Geiger
,
B.
,
Ginsberg
,
D.
,
Salomon
,
D.
and
Volberg
,
T.
(
1990a
).
The molecular basis for the assembly and modulation of adherence-type junctions
.
Cell Diff. and Dev
.
32
,
343
354
.
Geiger
,
B.
,
Volberg
,
T.
,
Ginsberg
,
D.
,
Bitzur
,
S.
,
Sabanay
,
I.
and
Hynes
,
O.
(
1990b
).
Broad spectrum pan-cadherin antibodies, reactive with the C-terminal 24 amino acid residues of N-cadherin
.
J. Cell Sci
.
97
,
607
614
.
Geiger
,
B.
,
Volk
,
T.
and
Volberg
,
T.
(
1985
).
Molecular heterogeneity of adherens junctions
.
J. Cell Biol
.
101
,
1523
1531
.
Geiger
,
B.
,
Volk
,
T.
,
Volberg
,
T.
and
Bendori
,
R.
(
1987
).
Molecular interactions in adherens-type junctions
.
J. Cell Sci
.
Suppl. 8
,
251
272
.
Goodman
,
S. R.
,
Zagon
,
I. S.
and
Kulikowski
,
R. R.
(
1981
).
Identification of a spectrin-like protein in non-erythroid cells
.
Proc, natn. Acad. Sci. USA
.
78
,
7570
7574
.
Granger
,
B. L.
and
Lazarides
,
E.
(
1979
).
Desmin and vimentin coexist at the periphery of the myofibril Z-disk
.
Cell 18,1053-1063
.
Guo
,
J.-I.
,
Jacobson
,
S. L.
and
Brown
,
D. L.
(
1986
).
Rearrangement of tubulin, actin and myosin in cultured ventricular cardiomyocytes of the adult rat
.
Cell Motil. Cytoskel
.
6
,
291
304
.
Hatta
,
K.
,
Nose
,
A.
,
Nagafuchi
,
A.
and
Takeichi
,
M.
(
1988
).
Cloning and expression of cDNA encoding a neural calcium-dependent cell adhesion molecule: its identity in the cadherin gene family
.
J. Cell Biol
.
106
,
873
881
.
Hlraoka
,
Y.
,
Sedat
,
J. W.
and
Agard
,
D. A.
(
1987
).
The use of charged-coupled device for quantitative optical microscopy of biological structures
.
Science
238
,
36
41
.
Holtzer
,
H.
,
Abott
,
J.
and
Cawanaugh
,
M. W.
(
1959
).
Some properties of embryonic cardiac myoblasts
.
Expl. Cell Res
.
16
,
595
601
.
Horwitz
,
A.
,
Duggan
,
K.
,
Buck
,
C.
,
Beckerle
,
M. C.
and
Burridge
,
K.
(
1986
).
Interactions of plasma membrane fibronectin receptor with talin - a transmembrane linkage
.
Nature
320
,
531
533
.
Kam
,
Z.
,
Chen
,
H.
,
Sedat
,
J. W.
and
Agard
,
D. A.
(
1991
).
Analysis of three-dimensional image data: Display and feature tracking
. In
Electron Tomography
(ed. J. Frank) New York: Plenum (
In
Press
).
Kelly
,
D. E.
(
1969
).
Myofibrillogenesis and Z-band differentiation
.
Anat. Rec
.
163
,
403
426
.
Kotehansky
,
V. E.
and
Gneushev
,
G. N.
(
1983
).
Vinculin localization in cardiac muscle
.
FEBS Lett
.
159
,
158
160
.
Lazarides
,
E.
(
1980
).
Intermediate filaments as mechanical integrators of cellular space
.
Nature
283
,
249
256
.
Lin
,
Z.
,
Holtzer
,
S.
,
Schulthelss
,
T.
,
Murray
,
J.
Masaki
,
T.
,
Fischman
,
D. A.
and
Holtzer
,
H.
(
1989
).
Polygons and adhesion plaques and the disassembly and assembly of myofibrils in cardiac myocytes
.
J. Cell Biol
.
108
,
2355
2367
.
Maruyama
,
K.
and
Ebashi
,
S.
(
1965
).
o’-actimn, a new structural protein from striated muscle II. Action of actin
.
J. Biochem
.
58
,
13
19
.
Messina
,
D. A.
and
Leman Ski
,
L. F.
(
1989
).
Immunocytochemical studies of spectrin in hamster cardiac tissue
.
Cell Motil. Cytoskel
.
12
,
139
149
.
Nag
,
A. C.
and
Cheng
,
M.
(
1981
).
Adult mammalian cardiac muscle cells in culture
.
Tissue Cell
.
13
,
515
523
.
Nag
,
A. C.
,
Cheng
,
M.
,
Fischman
,
D. A.
and
Zak
,
R.
(
1983
).
Long-term cell culture of adult mammalian cardiac myocytes: electron microscopic and immunofluorescent analyses of myofibrillar structure
.
J. Molec. Cell. Cardiol
.
15
,
301
317
.
Nelson
,
W. J.
and
Lazarides
,
E.
(
1984
).
Assembly and establishment of membrane-cytoskeleton domains during differentiation. Spectrin as a model system
. In
Cell Membranes: Methods and Reviews
. Vol.
2
(eds.
L.
Glaser
,
W.A.
Frazer
,
E.L.
Elson
) pp
219
246
.
New York
:
Plenum Publishing Corp
.
Otey
,
C. A.
,
Pavalco
,
F. M.
and
Burridge
,
K.
(
1990
).
An interaction between o’-actinin and the /JI integrin subunit in vitro
.
J. Cell Biol. Ill, 721-729
.
Otto
,
J. J.
(
1983
).
Detection of vinculin-binding proteins with an 125I-vinculin gel overlay technique
.
J. Cell Biol
.
97
,
1283
1287
.
Pardo
,
J. V.
,
Silidano
,
J. D.
and
Craig
,
S. W.
(
1983a
).
Vinculin is a component of an extensive network of myofibril-sarcolemma attachment regions in cardiac muscle fibers
.
J. Cell Biol
.
97
,
1081
1088
.
Pardo
,
J. V.
,
Sillciano
,
J. D.
and
Craig
,
S. W.
(
1983b
).
A vinculin-containing cortical lattice in skeletal muscle: Transverse lattice elements (“costameres”) mark sites of attachment between myofibrils and sarcolemma
.
Proc Natn. Acad. Sci. USA
80
,
1008
1012
.
Sanger
,
J. M.
,
Mittal
,
B.
,
Pochapin
,
M.
and
Sanger
,
J. W.
(
1986
).
Observations of microfilament bundles in living cells microinjected with fluorescently labelled contractile proteins
.
J. Cell Sci. Suppl
.
5
,
17
44
.
Sanger
,
J. W.
,
Mittal
,
B.
and
Sanger
,
J. M.
(
1984
).
Formation of myofibrils in spreading chick cardiac myocytes
.
Cell Motil
.
4
,
405
416
.
Schulthelss
,
T.
,
Zhongxiang
,
L.
,
Mei-Hua
,
L.
,
Murray
,
J.
,
Fischman
,
D. A.
,
Weber
,
K.
,
Masaki
,
T.
,
Michihiro
,
I.
and
Holtzer
,
H.
(
1990
).
Differential distribution of subsets of myofibrillar proteins in cardiac nonstriated and striated myofibrils
.
J. Cell Biol
.
110
.
1159
1172
.
Teral
,
M.
Komiyama
,
M.
and
Shimada
,
Y.
(
1989
).
Myofibril assembly is linked with vinculin, a--actinin, and cell-substrate contacts in embryonic cardiac myocytes in vitro
.
Cell Motil. Cytoskel
.
12
,
185
194
.
Tokúyasu
,
K. T.
,
Dutton
,
A. H.
,
Geiger
,
B.
and
Singer
,
S. J.
(
1981
).
Ultrastructure of chicken cardiac muscle as studied by double immunolabelling in electron microscopy
.
Proc. natn. Acad. Sci. USA
78
,
7619
7623
.
Turner
,
C. E.
,
Glenney
Jr,
J. R.
and
Burridge
,
K.
(
1990
).
Paxillin: A new vinculin-binding protein present in focal adhesions
.
J. Cell Biol. Ill, 1059-1068
.
Volk
,
T.
and
Geiger
,
B.
(
1984
).
A 135-kd membrane protein of intercellular adherens junctions
.
EMBO J
.
3
,
2249
2260
.
Volk
,
T.
,
Volberg
,
T.
,
Sabanay
,
I.
and
Geiger
,
B.
(
1990
).
Cleavage of A-CAM by endogenous proteinases in cultured lens cells and in developing chick embryos
.
Devi Biol
.
139
,
314
326
.
Wachsstock
,
D. H.
,
Wilkins
,
J. A.
and
Lin
,
S.
(
1987
).
Specific interaction of vinculin with oactinin
.
Biochem. Biophys. Res. Commun
.
146
,
554
560
.
Wang
,
S.-M.
,
Greaser
,
M. L.
,
Schultz
,
E.
,
Bulinski
,
J. C.
,
Lin
,
J. J.-C.
and
Lessard
,
J. L.
(
1988
).
Studies on cardiac myofibrillogenesis with antibodies to titin, actin, tropomyosin, and myosin
.
J. Cell Biol
.
107
,
1075
1083
.
Wilkins
,
J. A.
,
Chen
,
K. Y.
and
Lin
,
S.
(
1983
).
Detection of high molecular weight vinculin binding proteins in muscle and nonmuscle tissues with an electroblot-overlay technique
.
Biochem. Biophys. Res. Commun
.
116
,
1026
1032
.