Localization of the transmembrane proteoglycan syndecan-4 and its regulatory kinases in costameres of rat cardiomyocytes: A deconvolution microscopic study



Syndecan-4 (syn-4), a transmembrane heparan sulfate-containing proteoglycan, is unique among the four members of the syndecan family in its specific cellular localization to complex cytoskeletal adhesion sites, i.e., focal adhesions. During early phenotypic redifferentiation of neonatal cardiomyocytes in culture, immunolocalization reveals syn-4 to be heavily concentrated in the perinuclear endoplasmic reticulum-Golgi region, with little found at the peripheral regions. Subsequently, syn-4 becomes localized to a cytoskeletal adhesion complex unique to striated muscle, the costamere. Soon after redifferentiation of myofibrils in cultured neonatal cardiomyocytes, syn-4 is present only in costameres, not in focal adhesions. In cultured adult cardiomyocytes, it is present in both costameres and focal adhesions—the latter in two distinct regions of the spread cardiomyocytes, reflecting localization with two types of actin-containing filaments. The fact that syn-4 is observed early in the costameric regions, as opposed to later in the focal adhesions, suggests that it may play an initial role in early adhesion/signal transduction mechanisms in close proximity to the contractile apparatus, as well as in transmission of contractile force to the collagenous extracellular matrix (ECM) which surrounds the cardiac myofibers in situ. With respect to possible regulatory mechanisms of syn-4, we localized syn-4 with both the ϵ isoform of protein kinase C and the tyrosine kinase pp60csrc in costameric regions. These findings suggest that syn-4 may not only play a role in cellular adhesion and contractile force transmission, it may also, through ser, thr, and tyr phosphorylation, be part of an interactive signal transduction mechanism in myocardial functioning via these adhesive cytoskeletal complexes. Anat Rec 268:38–46, 2002. © 2002 Wiley-Liss, Inc.

Over the past decade and a half, it has come to light that the cytoskeleton is involved in complicated signal transduction mechanisms mediated by specialized cytoskeletal adhesion complexes, in addition to its role in various mechanisms of physical linkage of cells to the extracellular matrix (ECM) fibrous proteins. The components and their regulatory pathways are avidly being studied by numerous investigators using a wide range of techniques.

The mammalian syndecans are a highly conserved, four-member family (syn-1 to -4, named in the order of their cloning) of transmembrane proteoglycans which share several common structural features (Fig. 1) : a short (28–34 aa) cytoplasmic COOH-terminal cytoplasmic domain, containing two highly conserved domains (C1 and C2) and an intervening variable region (V); a highly conserved transmembrane domain (25 aa); and a variable-length ectodomain to which the glycosaminoglycans heparan sulfate and chondroitin sulfate are attached (Bernfield et al., 1992; David, 1993; Carey, 1997) (see also Fig. 1 and an excellent diagram in Rapraeger, 2000).

Figure 1.

Diagrammatic representation of the proteoglycan syn-4 spanning a plasma membrane. The outermost of the two external domains represents sites for attachment of heparin sulfate (HS) and chondroitin sulfate (CS) chains (dark squiggles and bent line) as well as cell binding domains. The following region is the highly conserved transmembrane domain. The inner three domains (only the middle segment is variable among the syndecans) constitute various ser, thr, or tyr phosphorylation sites as well as cytoskeletal protein binding sites (postsynaptic density 95, disk large, zona occludens-1 (PDZ)) as well as oligomerization and phosphotidylinositol (4,5) bisphosphate binding sites.

It is through the posttranslational modifications to the ectodomain region (Fig. 1) that the syndecans are able to interact with a number of macromolecules, such as growth factors (Elenius et al., 1990; Salmivirta et al., 1992; Salmivirta and Jalkanen, 1995; Gallagher, 1995); other cells' receptors; and ECM components, such as fibronectin. With respect to the present studies, it was recently demonstrated that syn-4, in addition to syn-1 and -2 bind to fibrillar Type 1 collagen (Liu et al., 1998), the major collagen comprising the extensive intercellular complexes in the myocardium. McFall and Rapraeger (1997) reported that in addition to the well-documented binding of syndecan's extracellular heparan sulfate domains to extracellular proteins, certain cell surface proteins of other cells may interact directly with the core protein of syn-4. A role for syn-1 and -4 in cell-to-cell adhesion has been suggested (Stanley et al., 1995; see also Fig. 2 in Rapraeger, 2000).

The transmembrane portion of the syndecans is highly conserved among the members of the syndecan family (Fig. 1). The three cytoplasmic domains are very active, with phosphorylation target sites for both protein kinase C (ser and thr) and tyrosine kinases (tyr), and binding sites for certain cytoskeletal proteins (Fig. 1). Of particular interest is the fact that the cytoplasmic domain of syn-1 and -4 may associate with actin-containing microfilaments (Carey et al., 1994a,b, 1996). This suggests that, like two other bifunctional transmembrane proteins, the large family of heterodimeric integrins and the proteoglycan dystroglycan, the extensive intracellular actin microfilament cytoskeleton may be stabilized or organized through syn-4's interaction with ECM components (Carey, 1997; Saoncella et al., 1999). The fact that most cell types express several syndecans (Kato et al., 1994), usually in different cellular locations, suggests that these molecules may serve different roles (Kim et al., 1994; Carey, 1997).

Syn-4 (also called ryudocan (Kojima et al., 1992) or amphiglycan (David et al., 1992)) is unique among other syndecans in that it is specifically localized in actin cytoskeletal-attaching focal adhesions (Woods and Couchman, 1994), whereas syn-1 presents a diffuse immunostaining pattern (Carey et al., 1996). Syn-2 also is absent from focal adhesions. The concept of differential expression of separate syndecans in the same tissue is supported by previous findings (Lian et al., 1998) that syn-4 is found in myocardial cells, whereas syn-1 is localized to endothelial cells of blood vessels. Although earlier immunological and molecular studies failed to detect syn-1 in heart (Hayashi et al., 1987), other studies determined that syn-2 and -4 are present in intact myocardium (Kim et al., 1994).

Earlier studies linking protein kinase C-α (PKCα) to fibroblast focal adhesion formation (Woods and Couchman, 1992) have been further expanded by studies indicating that localization of syn-4 in focal adhesions is dependent on PKC (Baciu and Goetinck, 1995). Recent findings (Oh et al., 1997a) that PKCα binds to and is activated by a cytoplasmic 28 amino acid domain of syn-4 (Horowitz and Simons, 1998) demonstrate that this class of proteoglycans can function in transmembrane signaling, perhaps through phosphorylation/dephosphorylation pathways.

In addition to the multicomponent focal adhesions (Burridge et al., 1988; Jockusch et al., 1995), cardiac and skeletal muscles possess unique adhesive, regularly spaced, striated cytoskeletal complexes that anchor contractile myofibers at precise intervals to the overlying sarcolemma and, ultimately, to the ECM. These striations, termed costameres (L.costa, rib) (Pardo et al., 1983), contain many of the same proteins (integrin, talin, focal adhesion kinase, paxillin, etc.) as found in focal adhesions (Wu et al., 1999), suggesting that similar regulatory and signal transduction mechanisms may operate at both adhesion sites (Sharp et al., 1997). In cardiac muscle, costameric structures are thought to be the sites for transmission of generated systolic force to the ECM components (Danowski et al., 1992; Sharp et al., 1997), requiring contractile activity to maintain structural/functional organization (Sharp et al., 1997). It is of interest that actin-myosin contractility in other cellular systems appears to also be necessary for the assembly of focal adhesion complexes, demonstrating yet another similarity in function between focal adhesion and costameric adhesion systems (Sastry and Burridge, 2000).

The nonreceptor proto-oncogene tyrosine kinase pp60c-src was first demonstrated in focal adhesions in Rous sarcoma virus (now referred to as pp60v-src) transformed kidney and fibroblast cell lines in the early 1980s (Rohschneider, 1980; Nigg et al., 1982). Even at this early date, the studies suggested that pp60c-src was somehow associated with the intracellular cytoskeleton. The suggestion by Ott and Rapraeger (1998) that the tyrosine kinase pp60csrc, the cellular homologue of the viral transforming oncogene, may be a syn-4 regulator led us to study the localization of this kinase in cardiomyocytes. Our findings are not only the first report of syn-4 as a part of the cardiomyocyte cytoskeletal adhesion mechanisms (focal adhesions as well as costameres), but also suggest that syn-4 and two kinases, the ϵ isoform of PKC and pp60csrc, may constitute part of a regulatory pathway at the unique cellular junction, the costamere.

Given the potential myocardial functional significance of syn-4 as a “double-headed” anchor site joining extracellular matrices to intracellular contractile proteins, as well as a component in a transmembrane signal transduction mechanism/putative regulatory pathway, we undertook studies to ascertain its presence at the myofibrillar–sarcolemmal–extracellular force transmission site (i.e., the costamere) in neonatal and adult cultured cardiomyocytes.


Neonatal Cardiomyocyte Isolation and Culture

Cardiomyocytes isolated from 3- to 4-day-old Sprague-Dawley rat pup ventricles and maintained in DMEM-based culture medium (VanWinkle et al., 1994, 1995, 1996) were plated on the multicomponent, biosynthetic ECM Cardiogel, as recently described (VanWinkle et al., 1996). All tissue culture reagents were purchased from Gibco (Grand Isle, IA). On the matrix Cardiogel, the neonatal cells became redifferentiated and exhibited spontaneous contractile behavior by 3 days after plating and were used for the immunological studies described herein within 4 days.

Adult Cardiomyocyte Isolation and Culture

Cardiomyocytes were obtained from mature (200 g), heparinized male Sprague-Dawley rats by aseptic retrograde Langendorff perfusion of left ventricles with Joklik-modified minimal essential medium (MEM) containing 0.1% collagenase (#CLS2;Worthington, Freehold, NJ) and 0.1% trypsin for 30–45 min followed by mincing of the ventricles in fresh medium and shaking on a reciprocating bath at 37°C for 10 min. Following centrifugation at 50 ×g in a sterile 15-ml conical tube, cells were resuspended in Joklik's medium containing 4% bovine serum albumin (BSA), followed by centrifugation and resuspension in Joklik's medium containing 2% BSA. Calcium concentration was incrementally increased from 0 to 1 mM by addition of 20-μl aliquots of 200-mM CaCl2 stock. Following centrifugation, cardiomyocytes were resuspended in Joklik's medium containing 0.1% trypsin, 2% BSA, and 1 mM CaC12, shaken at 37°C for 10 min and recentrifuged. The pellet was washed twice in the above buffer minus trypsin, and the cardiomyocytes were resuspended in Dulbecco's MEM containing 1%:5% penicillin:streptomycin, 10% fetal bovine serum, 10 μg/ml insulin, and 3 μg/m1 cytosine arabinoside (AraC). This procedure routinely produces cardiomyocyte populations containing 80–85% noncontracting elongate cells. Cardiomyocytes were plated on laminin- or Cardiogel-coated glass coverslips and maintained at 37°C, 5%CO2. Cardiomyocyte culture medium was replaced with fresh AraC-containing medium every other day for 7–10 days. Cardiomyocytes were examined 1 week and 2 weeks after plating by the immunofluorescent microscopic techniques described below.

Syn-4 Monoclonal Antibody

The monoclonal antibody to syn-4 (#150.9) was generated against the synthetic peptide ESIRETEVIDPQDLLE(C) representing the N-terminal amino acids of human and rat syn-4 conjugated to keyhole limpet hemocynanin (KLH) (Longley et al., 1999).

Confocal Scanning Laser Immunofluorescence Microscopy (CSLIM) and Deconvolution Microscopy (DM)

For localization of syn-4, cardiomyocytes were rinsed in PBS and fixed in cold (–20°C) 20% methanol for 10 min, followed by extensive washing in PBS. Nonspecific binding was blocked by preincubation with 10% goat serum (Sigma Chemical Co., St. Louis, MO) in PBS. Mouse IgG (cell supernatant #150.9) to syn-4 was incubated with coverslips for 45 min at 37°C. Following washing in PBS, coverslips were incubated with a 1:500 dilution of Texas Red-conjugated goat anti-mouse IgG (Molecular Probes, Eugene, OR) in PBS containing 0.05% Tween 20 and 10% goat serum. Following washing in PBS, coverslips were mounted using the polyvinyl Elvanol (a generous gift from DuPont Chemicals, Wilmington, DE). While this fixation protocol was optimal for syn-4 localization, attempts at subsequent colocalization staining with polyclonal and monoclonal antibodies to a variety of focal adhesion/costameric cytoskeletal components (tensin, paxillin, and talin, among others) which we routinely employ for colocalization at these adhesion sites (VanWinkle et al., 1994, 1995, 1996) were generally not successful. The one exception we found was a monoclonal antibody to vinculin (Vin-11-5; ICN Immunobiologicals, Costa Mesa, CA). This latter antibody was useful after the cold methanol fixation protocol, hence our use of the focal adhesion and costameric vinculin component when searching for colocalizion with syn-4. Immunolabeling specificity of the syn-4 monoclonal antibody was assessed by incubation (30 min at 37°C) of the monoclonal (#150.9), with the synthetic peptide coupled to KLH prior to incubation of the mixture with cardiomyocytes as detailed above for immunolocalization. This epitope blocking resulted in the abolition of immunolocalization of the syndecan monoclonal to cardiomyocyte cytoskeletal adhesion sites (costameres and focal adhesions) (data not shown).

For visualization of PKCϵ, coverslips were washed in PBS followed by fixation in 3.7% formaldehyde (Tousimis, Rockville, MD)/PBS for 5 min followed by permeabilization in 0.5% Triton X-100 for 2 min. Antibody to a synthetic decapeptide of PKCϵ (aa728-737) was obtained from Research and Diagnostic Antibodies (Benicia, CA) and used at a 1:100 dilution in PBS for 45 min at 37°C, followed by visualization with secondary Texas Red-conjugated antibody as described above. When desired, nuclei were visualized with 4′,6′-diamidino-2-phenylindole (DAPI) (Molecular Probes) at 0.1 μg/ml prior to mounting in Elvanol. Coverslips were examined using CSLIM (Molecular Dynamics Inc., Sunnyvale, CA) in addition to an applied precision delta vision wide field DM system (API, Issaquah, WA). Images represent single deconvolved optical sections or stacks of sections (each section is 200 nm thick). Three images (Figs. 3a and b, and 4) were obtained as a three-dimensional volume rendering using Imaris software (Bitplane, Zurich, Switzerland). This latter procedure employs stacks of serial optical sections in 2D, which when processed contain sufficient depth cues, shading, and highlighting to allow the perception of the third dimension (Messerli et al., 1993).

For F-actin localization (Fig. 2b), Bodipy-phallacidin (Molecular Probes) was employed as described previously (VanWinkle et al., 1995).

Figure 2.

Complete view of a stack of 200-nm optical sections across an entire freshly isolated adult cardiomyocyte, showing a striated pattern of sarcolemmal syn-4 (red indicates syn-4 alone; yellow indicates colocalization of actin (green) and syndecan (red)). To the right is a small section taken from a whole cell and magnified, showing syn-4 in very regular 2 μm (red) patterns adjacent to colocalized syn-4 and actin (yellow).

Construction and Application of the pp60csrc-EGFP Expression Vector

Human pp60csrc cDNA (kindly provided by Gary Gallick, M.D. Anderson Cancer Center, Houston, TX) was amplified from the expression vector pcDNA3-pp60csrc using the T7 primer and the Nhe-pp60csrc 3 (AGATGCTAGCGCGAGGTTCTCCCCGGGCTGG) (which introduces an NH3 site at the 3′ end of the pp60csrc coding sequence) and cut with Hind III and Nhe I. Enhanced green fluorescent protein (EGFP) was isolated from the vector pEGFP-N (Clontech, Palo Alto, CA) after cutting with Nhe I and Xho I. Plasmid pcDNA3-pp60c-src was cut with Hind III and Xho and the vector free of insert was isolated. A three-way ligation was set up between the EGFP fragment, pcDNA3, and the amplified pp60c-src to obtain pcDNA3 pp60c-src-EGFP which was used for subsequent transfection of neonatal cardiomyocytes on glass coverslips. LipoTAXI (Stratagene, La Jolla, CA) in serum-antibiotic-free DMEM was mixed with pcDNA pp60csrc-EGFP and incubated for 30 min. The construct mixture was then added to the coverslips, which were then incubated for 6 hr at 37°C. The coverslips were then washed in complete medium for 48 hr and fixed in paraformaldehyde as described above.

Animal Care

The use of animals in this study followed approved guidelines for animal welfare. This project was approved by the University of Texas Houston Medical School Center for Laboratory Animal Medicine and Care (NIH Welfare Assurance #A3413-01; USDA 74-R-068; WBV HSC-AWC-98-093).


Syn-4 Immunolocalization in Cardiomyocytes

As revealed by immunofluorescence microscopy (both by CSLIM and DM) of freshly isolated adult cardiomyocytes, the external surfaces of the elongate freshly isolated adult cardiomyocytes are replete with regularly spaced striations strongly stained for syn-4 (left panel of Fig. 2). The rectangle on the left in Figure 2 shows the repeat colocalization at higher magnification. Myofibrillar actin and syn-4 colocalized (green = actin; red = syndecan; yellow = colocalized syndecan:actin) at each Z-band region with 2 μm periodicity (Fig. 2). Freshly isolated neonatal cardiomyocytes are not suitable for this type of study, as they immediately round up during the isolation procedure.

Neonatal cardiomyocytes were placed in tissue culture on the fibroblast-derived, ECM Cardiogel 1–2 days after plating (at which time striated myofibrils and costameres are not well differentiated, nor is spontaneous contractility observed (VanWinkle et al., 1994). Syn-4 is observed in large aggregates in the perinuclear region (Fig. 3a). This location and general morphology coincide well with the perinuclear location of the rough endoplasmic reticulum-Golgi network in cultured cardiomyocytes we previously observed using fluorescent ceramide probes (Pagano et al., 1991) (data not shown), and suggests extensive new synthesis of syn-4 during early redifferentiation of the cardiomyocyte. It should be noted that at this early stage, syndecan is not yet localized to the sarcolemma (not shown in figure). Primitive spotty focal adhesions hold the spherical neonatal cell down as it is spreading (there is no sarcolemmal syn-4 yet), and only after the establishment of striated myofibrils and their attendant costameric complexes is syn-4 found at the sarcolemma.

Figure 3.

a: Three-dimensional, volume-rendered, deconvolved image of a stack (20 200-nm-thick optical sections) of syn-4 localization adjacent to the nucleus in neonatal rat cardiomyocytes 2 days after plating. The image represents typical perinuclear localization of syn-4 in amorphous complexes (red) similar to rough endoplasmic reticulum-Golgi surrounding the nucleus (blue). The white bar in this image represents 15 μm. In the subsequent panels, the bar represents 7.5 μm. b: Three-dimensional, volume-rendered, deconvolved fluorescent image of a stack of optical sections of a neonatal cardiomyocyte 3 days after plating on Cardiogel ECM. Syn-4 (green-yellow) is localized only in striated costameric regions in these cells (arrowheads in rectangular region taken from larger photomicrograph); there is no localization of syn-4 in focal adhesion regions (large arrow) (compare with focal adhesion localization in part d, above right). Colocalization of vinculin (red) with syndecan indicates that the striations are costameres. Insert at higher magnification shows yellow-red costameric colocalization of syn-4; vinculin (arrows) is meant to isolate a small region in order to visualize the costameres, rather than for high magnification as in Figure 2 insert. Bar at right is equal to 7.5 μm. c: Cluster of neonatal cardiomyocytes following 4-day plating on laminin-coated coverslips. In contrast to cardiomyocytes maintained on the multiple-component biomatrix Cardiogel (above), the single matrix component laminin-supported cells do not exhibit striated costameric localization of syn-4. Although syn-4 is present in these cells, it appears as wispy sarcolemmal filaments, mostly in perinuclear regions near the ventral surface of the cells (arrows). Empty, circular black “holes” represent nuclei that were not stained in this figure. d: Adult cardiomyocytes 4 days after plating on Cardiogel exhibit strong syn-4 localization in striated regions (arrowheads) as well as to two regions of focal adhesions: one terminates the striated myofibrils, the other terminates the stress fibers near the periphery of the myocytes (arrows). e: Adult cardiomyocytes as above, but stained for localization of the epsilon isoform of protein kinase C. Note that PKC-ϵ localizes in both the striated costameric regions (arrowheads) and in the focal adhesions terminating the myofibrils (small arrows). Note especially the lack of this isoform staining in the peripheral focal adhesions. Nuclei are stained with DAPI (blue). f: Localization of syn-4 in cardiac fibroblast focal adhesions (arrowheads). Note that focal adhesions may occur anywhere on the ventral surface, not only at terminal peripheral “points.”

After 3–5 days in culture, neonatal cardiomyocyte myofibrils redifferentiate, exhibiting myofibrillar striations as well as spontaneous contractility. It should be noted that while approximately half of adult cardiomyocytes do not lose their original myofibrils, and remain elongated, virtually all neonatal cardiomyocytes become spherical when isolated and only regain myofibrillar arrangement after 3–5 days in culture. Syn-4 is localized exclusively in costameric striations and not in focal adhesions, and exhibits a 2-μm (Z-band to Z-band sarcomeric distance) periodicity at the periphery of the cardiomyocytes adjacent to the ECM (Fig. 3b). It is important to note that in most cardiomyocytes, syn-4 is localized only to the focal adhesions and not in the costameric striations that are observed later. These striations are identified as costameres by the colocalization of the costameric/focal adhesion component vinculin (Fig. 3b). Thus, despite the fact that integrin-based, vinculin-containing focal adhesions are well established in cultured neonatal cardiomyocytes at this time (VanWinkle et al., 1994, 1995), syn-4 is not present at the termini of contractile fibers at this early stage (Fig. 3b).

However, as cells progress in dedifferentiation (4–5 days), syn-4 does become localized to the focal adhesions as well. It should be noted also that by this time spontaneous contractile activity of the spread cardiomyocytes has become much more robust than in the earlier stages of dedifferentiation.

Our cardiac fibroblast-derived ECM material, Cardiogel, contains collagen Types I and III, laminin, fibronectin, proteoglycans, and other ECM components (VanWinkle et al., 1996). Since syn-4 in fibroblast focal adhesions colocalizes with fibronectin in those regions (Baciu and Goetinck, 1995), we cultured cardiomyocytes on laminin matrix alone to test cardiomyocyte syn-4 specificity for singular ECM components. Neonatal cardiomyocytes adhere to and spread, and develop costameres, focal adhesions, and organized myofibrils on laminin (VanWinkle et al., 1996). Despite the presence of vinculin-containing costameres, however, syn-4 is not observed in these adhesive sites (Fig. 3c) in cardiomyocytes maintained on laminin alone. While there is diffuse, wispy labeling across the cells, this may reflect the presence of fibronectin, a component of serum to which syn-4 may be responding. Conversely, fibroblasts plated on laminin in the absence of protein synthesis are able to establish focal adhesions (Woods and Couchman, unpublished findings).

In contrast, in cultured adult cardiomyocytes, syn-4 localizes in repeat striations reminiscent of striated “costameres” containing vinculin (Fig. 3d) and other cytoskeletal proteins (VanWinkle et al., 1994, 1995, 1996), as well as in focal adhesions (Fig. 3d). As shown above, syn-4 is not localized to focal adhesions in neonatal cells, even though other focal adhesion components are present. The lack of syn-4 in early functional neonatal cardiomyocyte focal adhesions is in contrast to its prominent focal adhesion location in fibroblasts (Woods and Couchman, 1994) (Fig. 3f).

Focal adhesions in fibroblasts have been shown to activate as well as bind to PKCα (Woods and Couchman, 1992), a step which appears to also involve recruitment of the actin cytoskeletal regulator phosphatidyl 4,5-bisphosphate (Oh et al., 1998). This regulatory step appears to be phosphorylation of ser183, which immediately precedes the binding and activation of PKC (Horowitz and Simons, 1998). However, unlike fibroblasts, it is the ϵ isoform—not the α isoform—that is present in cardiomyocyte focal adhesions and costameric striations (Distanik et al., 1994) (data not shown) (Fig. 3e).

Many cultured cardiomyocytes exhibit two distinct locations of focal adhesions: the first, more cortically located foci serve as the physical anchoring termini for the striated myofibrils containing the sarcomeric actin isoform; the other focal adhesions are more peripheral and anchor nonstriated actin filaments comprised of the smooth muscle actin isoform (Eppenberger-Eberhardt et al., 1990; Messerli et al., 1993) (Snuggs and VanWinkle, unpublished findings). Syn-4 is present in both classes of focal adhesions (Fig. 3e). It is obvious from various micrographs in Figure 2, as well in Figures 2 and 3, that both cytoskeletal adhesion complexes, focal adhesions, and costameres are extensive, occupying a large area at the cardiomyocyte–ECM interface.

pp60csrc Localization in Cardiomyocytes

Following transfection of the EGFP-pp60csrc, the probe is found heavily localized in two regions of the cardiomyocyte: the broad contractile filament-terminating focal adhesions, and the striated costameric adhesion sites (Fig. 4). Cells transfected with the EGFP probe minus the pp60csrc exhibited only nonspecific background cytoplasmic fluorescence (not shown). As judged by the amount of probe signal for incorporated cytoskeletal adhesion pp60csrc, this tyrosine kinase is plentiful in cultured cardiomyocytes. The photomicrograph in Figure 4 is the first published demonstration of newly synthesized pp60csrc by direct (green fluorescence protein) protocols, and shows a deconvolution image of the pp60csrc localized with syn-4, PKCϵ, and vinculin, as well as numerous other focal adhesions and costameric constituents.

Figure 4.

Deconvolved optical image stack of a large cultured neonatal cardiomyocyte illustrating pp60csrc-GFP (green) localization in focal adhesions (large arrows) and striated costameres in peripheral as well as centrally located myofibrils (arrowheads). 3D imaging rendered by Imaris software (Bitplane, Zurich). Bar = 7.5 μm.


Syn-4 Localization

Our studies indicate that the transmembrane proteoglycan, syn-4, as well as two of its possible phosphorylation regulators (PKC ϵ (ser, thr) and pp60c-src (tyr)) are components of both focal adhesions and costameric complexes in mammalian cardiomyocytes. These multicomponent cytoskeletal specializations serve as cellular anchors through which the cardiomyocytes not only maintain their attachment to the ECM but also transmit their contractile force (Danowski et al., 1992; Sharp et al., 1997). The relative amount and size of cardiomyocyte focal adhesions and costameres in this study is not unexpected given the considerable spontaneous contractile activity of the cells directed against the underlying ECM, which requires extensive and strong anchorage complexes. Indeed, the contractile force generated by single isolated cardiomyocytes (freshly isolated, noncultured) has been shown to be considerable: 22.3 N/mm2 (Palmer et al., 1996) for “nonskinned cardiomyocytes” (sarcolemma intact), and 0.57 mg for skinned cardiomyocytes (8–13 μm wide, 35–60 μm long) (Fabiato and Fabiato, 1978).

It has been suggested that the syndecan extracellular heparan sulfate–chondroitin sulfate chains may serve at least three functions: 1) participation in cell-to-matrix adhesion in concert with the membrane spanning and cytoskeletal anchoring proteins, the integrins; 2) concentration of growth factors near their respective receptors; and 3) cell-to-cell adhesion (Iba et al., 2000) and possible gene regulation through nuclear targeting of calcium/calmodulin-dependent serine kinase (CASK) (Rapraeger, 2000).

A previous study (Asundi et al., 1997) demonstrated immunolocalization of another proteoglycan, glypican, in the sarcolemma at the lateral (extracellular contact) as well as at the intercalated disk (cell-to-cell contact). Overall, glypican sarcolemmal staining is in contrast to that of the syn-4 reported here, which occurs only at sarcolemma-ECM adhesion sites (focal adhesions and costameres) and not at the intercalated disk cell-cell adhesion regions. Syn-1 has been shown to colocalize with actin microfilaments in other cell types (Rapraeger et al., 1986; Carey et al., 1994a, b). Recent studies have identified an actin filament binding domain in the central region of the syn-4 cytoplasmic domain (Saoncella et al., 1999).

Syn-4 and Phosphorylation

Given that certain structural-functional aspects of focal adhesions (and, by extension, costameres) are regulated by phosphorylation/phosphatase mechanisms, it is not surprising that we were able to localize kinases specific for both ser-thr (PKC ϵ) and tyr (pp60csrc). Focal adhesions and costameres contain at least three major tyrosine kinases (focal adhesion kinase (FAK), pp60csrc, and calcium-dependent tyrosine kinase (CADTK)) in addition to their many cytoskeletal substrates (β-integrin, tensin, paxillin, vinculin, cortactin, etc.) (Jockusch et al., 1995; Burridge and Chrzanowska-Wodnicka, 1996) (Van Winkle and Snuggs, unpublished data). In the Cl and C2 cytoplasmic domains, all of the syndecans contain two tyrosine residues whose surroundings render them as consensus tyrosine kinase substrates, (C1 {DEGSY} and C2 (constituting the extreme carboxy terminus of the molecule {EFYA} (Ott and Rapraeger, 1998). However, phosphorylation of key ser-thr residues also plays an important regulatory role in these cytoskeletal/ECM adhesion sites (Burridge and Chrzanowska-Wodnicka, 1996). Recent studies regarding syn-4 phosphorylation have shown that Ser183 located in the cytoplasmic tail of syn-4 is regulated by a novel PKC, although the functional role of such phosphorylation remains unknown at this time (Horowitz and Simons, 1998). Of possible significance regarding the syn-4 localization in the myofibrillar binding costamere is that the carboxy terminus contains a PDZ binding site that binds several cytoskeletal proteins, syntenin, CASK, synectin, and synbindin (Simons and Horowitz, 2001). This suggests the possibility of considerable syn-4–cytoskeletal interaction.

Of at least 10 PKC isoforms, only the non-Ca2+-dependent ϵ isoform is localized to costameric/Z-band regions (Fig. 3e) (Distanik et al., 1994) and to focal adhesions in heart, in distinct contrast to specific localization of PKCα to focal adhesions in other cell types (Woods and Couchman, 1992, 1994). Interestingly, PKCϵ is concentrated in those more cortically-located focal adhesions which terminate the striated myofibrils, but not at the more peripheral focal adhesions. We have observed (data not shown) that the striated actin terminates at the cortical focal adhesions, with the nonstriated actin portion continuing with its terminus connecting to the more peripheral focal adhesions. This terminal, nonstriated area stains intensely for the smooth muscle actin isoform. In addition, PKCϵ is not observed at the intercalated disk junctions between adjacent cardiomyocytes (Fig. 3e). The apparent disparity between localization of the PKCα in fibroblast lines and the exclusive localization of PKCϵ in cardiac muscle adhesion complexes may suggest a separate role for these two PKC isoforms in different tissue types. The α isoform is a predominant subtype in adult heart (Bogoyevitch et al., 1993; Puceat et al., 1994; Rybin and Steinberg, 1994), and in other tissues it is involved in numerous cellular functions, from phosphoinositide metabolism (Huwiler et al., 1991), gene expression (Kang et al., 1996; Li et al., 1996) to cell adhesion (Chun et al., 1996). Its role in the cardiac myofibrillar/cytoskeletal system is not fully understood, although it has been shown that cardiac PKC-α activity is increased by increased systolic stretch (Puceat et al., 1994). PKC-α is translocated to the membrane fraction in response to agonists such as phenylephrine, carbachol, endothelin, and phorbol myristate (Bogoyevitch et al., 1993), and to arachidonic acid (Huang et al., 1997). Since PKC ϵ localizes in costameres or focal adhesions in cardiomyocytes, it is tempting to equate positional colocalization of PKCϵ and syn-4 with a functional interaction such as that shown for focal adhesion syn-4 and PKC-α in fibroblasts (Oh et al., 1997b). Final resolution of a possible functional interaction between these two macromolecules would help to clarify the role of PKC-ϵ in heart. Given the constantly emerging role of cytoskeletal tyrosine and serine/threonine phosphorylation and regulation of cell adhesion to ECM components (Sharp et al., 1997), further study of the cytoskeletal association of syn-4 and neighboring protein kinases with the cardiomyocyte cytoskeleton is warranted.

Ott and Rapraeger (1998) first observed that two specific tyrosine kinase inhibitors, genistein and herbimycin, inhibit syndecan phosphorylation, apparently at highly conserved residues in the cytoplasmic regions of syn-4. The means by which syn-4 is uniquely localized to adhesion sites, such as focal adhesions and costameres, is not understood, although most reports point to PKC. Although it cannot be determined from our studies, plentiful localization of syn-4 in both of the cardiomyocyte adhesion systems may indicate a phenomenon recognized in other syndecan-based systems: oligomerization of syndecan through the transmembrane domain interactions (Carey, 1997; Ott and Rapraeger, 1998; Zimmermann and David, 1999). It may be possible that the unconserved or variable region (V) in the cytoplasmic domain (which has been assigned as a PKC-interactive domain in syn-4 located between two highly conserved regions, Cl and C2) may represent part of a mechanism by which syn-4 is directed to and maintained in specific cytoskeletal adhesion regions. That integrins and the core protein of syn-4 apparently interact in a cooperative manner to form cytoskeletal adhesion–signaling complexes has been suggested for some time (Woods et al., 1986; Echtermeyer et al., 1999). (See Simons and Horowitz (2001) and Woods and Couchman (2001) for recent reviews on syn-4 and cell signaling.)

Both adult and neonatal cardiomyocyte cultures are contaminated by 5–10% fibroblasts, which usually disappear in long-term culture due to the constant presence of Ara C. As shown in other fibroblast lines (Woods and Couchman, 1994; Baciu and Goetinck, 1995), syn-4, in contrast to other members of the syndecan family, localizes exclusively at the focal adhesions in this cell type (Fig. 3f). PKC-ϵ was not localized in the terminal focal adhesions in cardiac fibroblasts (data not shown) but, in agreement with previous observations (Goodnight et al., 1995), was seen as diffuse perinuclear staining.

Potential Roles of Syn-4 in Cardiomyocytes

In situ, each cardiomyocyte comprising the myocardium is surrounded by an extensive collagenous ECM surrounding each myofiber as a thick fibrous layer (see especially Fig. 3 in Icardo and Colvee (1998)) that provides cellular anchoring against which the cardiomyocyte force developed during systole is transmitted (Winegrad and Robinson, 1978). The cellular role(s) of the syndecans include at least three signaling cascades: 1) cadhedrin/catenin-mediated cell-to-cell adhesion, which also may include a gene regulation cascade through syndecan's interaction with CASK; 2) heparan sulfate-mediated growth factor signaling; and 3) integrin-coordinated cell-ECM adhesion (Saoncella et al., 1999; Couchman and Woods, 1999; Rapraeger, 2000). The implication of pp60csrc in signaling cascades coupled with syndecans has been suggested in a variety of tissues (Hsveh et al., 1998; Kinnunen et al., 1998; Ott and Rapraeger, 1998; Yamashita et al., 1999). The role of syn-4 at the costamere remains conjectural, although several possibilities exist: 1) Its persistence, indeed increased organization, at these regions in older, well-differentiated contractile cells suggests a role in mechanical adhesion of the myofibrils to ECM components since it is at the costameres that developed contractile tension is transmitted to the surrounding matrix (Danowski et al., 1992; Sharp et al., 1997). 2) Although the fixation protocol employed for the syn-4 antibody employed in this study is detrimental to the localization of most other antibodies, it is of interest that antibodies to the bFGF receptor localize in costameres in cardiomyocytes (Kardami et al., 1991), suggesting a possible role for syn-4 as a regulator of low-affinity bFGF co-receptor interactions in these cells. 3) Finally, the possibility that syn-4 at this adhesion/mechanical transduction site in cardiomyocytes may serve as an activation or signaling mechanism involving PKC-ϵ warrants further study. The addition of syn-4 to the list of syndecans that bind intracellular actin (Saoncella et al., 1999) suggests a physical intra-/extracellular linkage via syn-4. Longley et al. (1999) recently found that deletion of the syn-4 cytoplasmic domain results in considerable changes not only in the functional ability of cellular migration, but in the actual shape of the cell itself.

Ott and Rapraeger (1998) suggested that the syndecan tyrosylphosphorylation is tightly controlled by a constitutively active kinase, and that this kinase might be pp60csrc. Of particular interest is the fact that the tyrosine kinase pp60csrc is so prominent in the costameres, as well as in the two adhesive sites in the focal adhesions (Fig. 2). While this finding does not preclude the presence and effects of other tyrosine kinases, namely focal adhesion kinase (FAK) or other members of the src family, it does suggest that src may play an important role in the signal transduction through costameres.

In summary, in addition to its recognized presence in focal adhesions in other cell types, mainly fibroblast cell lines, we have localized the transmembrane proteoglycan syn-4 to the cytoskeletal adhesion complex unique to striated muscle, the costamere. Given the colocalization of two important kinases to both adhesion structures in cardiomyocytes, it is possible that both sites serve at least two functional roles: as a means of transmission of contractile force to the surrounding ECM, and as a factor in a yet to be determined regulatory signaling pathway.