Search

Cell adhesion to the extracellular matrix is mediated by elaborate networks of multiprotein complexes consisting of adhesion receptors, cytoskeletal components, signaling molecules, and diverse adaptor proteins. To explore how specific molecular pathways function in the assembly of focal adhesions (FAs), we performed a high-throughput, high-resolution, microscopy-based screen. We used small interfering RNAs (siRNAs) to target human kinases, phosphatases, and migration- and adhesion-related genes. Multiparametric image analysis of control and of siRNA-treated cells revealed major correlations between distinct morphological FA features. Clustering analysis identified different gene families whose perturbation induced similar effects, some of which uncoupled the interfeature correlations. Based on these findings, we propose a model for the molecular hierarchy of FA formation, and tested its validity by dynamic analysis of FA formation and turnover. This study provides a comprehensive information resource on the molecular regulation of multiple cell adhesion features, and sheds light on signaling mechanisms regulating the formation of integrin adhesions.

Focal adhesions (FAs) regulate cell migration. Vinculin, with its many potential binding partners, can interconnect signals in FAs. Despite the well-characterized structure of vinculin, the molecular mechanisms underlying its action have remained unclear. Here, using vinculin mutants, we separate the vinculin head and tail regions into distinct functional domains. We show that the vinculin head regulates integrin dynamics and clustering and the tail regulates the link to the mechanotransduction force machinery. The expression of vinculin constructs with unmasked binding sites in the head and tail regions induces dramatic FA growth, which is mediated by their direct interaction with talin. This interaction leads to clustering of activated integrin and an increase in integrin residency time in FAs. Surprisingly, paxillin recruitment, induced by active vinculin constructs, occurs independently of its potential binding site in the vinculin tail. The vinculin tail, however, is responsible for the functional link of FAs to the actin cytoskeleton. We propose a new model that explains how vinculin orchestrates FAs.

The transition of cell-matrix adhesions from the initial punctate focal complexes into the mature elongated form, known as focal contacts, requires GTPase Rho activity. In particular, activation of myosin II-driven contractility by a Rho target known as Rho-associated kinase (ROCK) was shown to be essential for focal contact formation. To dissect the mechanism of Rho-dependent induction of focal contacts and to elucidate the role of cell contractility, we applied mechanical force to vinculin-containing dot-like adhesions at the cell edge using a micropipette. Local centripetal pulling led to local assembly and elongation of these structures and to their development into streak-like focal contacts, as revealed by the dynamics of green fluorescent protein-tagged vinculin or paxillin and interference reflection microscopy. Inhibition of Rho activity by C3 transferase suppressed this force-induced focal contact formation. However, constitutively active mutants of another Rho target, the formin homology protein mDia1 (Watanabe, N., T. Kato, A. Fujita, T. Ishizaki, and S. Narumiya. 1999. Nat. Cell Biol. 1:136-143), were sufficient to restore force-induced focal contact formation in C3 transferase-treated cells. Force-induced formation of the focal contacts still occurred in cells subjected to myosin II and ROCK inhibition. Thus, as long as mDia1 is active, external tension force bypasses the requirement for ROCK-mediated myosin II contractility in the induction of focal contacts. Our experiments show that integrin-containing focal complexes behave as individual mechanosensors exhibiting directional assembly in response to local force.

We have used the selective farnesylation inhibitor HR12 [cysteine-N(methyl)valine-N(cyclohexyl) glycine-methionine-O-methyl-ester] to study the role of oncogenic Ras in cytoskeletal reorganization in Ha-ras(V12)-transformed Rat1 cells (Rat1/ras). Application of HR12 resulted in complete restoration of the cytoskeleton and associated cell adhesions disrupted by oncogenic Ras. This included an increase in the number and size of focal adhesions, accompanied by massive stress fiber formation and enhanced tyrosine phosphorylation. Furthermore, HR12 induced assembly of adherens junctions and dramatically elevated the level of the junctional components, cadherin and beta-catenin. HR12 was unable to restore the nontransformed phenotype in cells expressing farnesylation-independent, myristylated Ras. Examination of the main Ras-regulated signaling pathways revealed that HR12 induced a dose- and time-dependent decline in Erk1&2 activation (t(1/2) approximately 6 h), which correlated with the accumulation of nonfarnesylated oncogenic-Ras. Inhibition of the Mek/Erk pathway in Rat1/ras cells, using the Mek inhibitor, PD98059, resulted in complete cytoskeletal recovery, indistinguishable from that induced by HR12. Moreover, a constitutively active Mek mimicked the effect of ras transformation in Rat1 cells, and prevented HR12-induced cytoskeletal effects in Rat1/ras cells. No such effects were observed after treatment of Rat1/ras cells with the phosphatidylinositol 3-kinase inhibitor LY294002. These findings establish the Mek/Erk pathway as the dominant pathway involved in conferring the cytoskeletal and junctional manifestations of the Ras-induced transformed phenotype.

beta-Catenin and plakoglobin are homologous proteins that function in cell adhesion by linking cadherins to the cytoskeleton and in signaling by transactivation together with lymphoid-enhancing binding/T cell (LEF/TCF) transcription factors. Here we compared the nuclear translocation and transactivation abilities of beta-catenin and plakoglobin in mammalian cells. Overexpression of each of the two proteins in MDCK cells resulted in nuclear translocation and formation of nuclear aggregates. The beta-catenin-containing nuclear structures also contained LEF-1 and vinculin, while plakoglobin was inefficient in recruiting these molecules, suggesting that its interaction with LEF-1 and vinculin is significantly weaker. Moreover, transfection of LEF-1 translocated endogenous beta-catenin, but not plakoglobin to the nucleus. Chimeras consisting of Gal4 DNA-binding domain and the transactivation domains of either plakoglobin or beta-catenin were equally potent in transactivating a Gal4-responsive reporter, whereas activation of LEF-1- responsive transcription was significantly higher with beta-catenin. Overexpression of wild-type plakoglobin or mutant beta-catenin lacking the transactivation domain induced accumulation of the endogenous beta-catenin in the nucleus and LEF-1-responsive transactivation. It is further shown that the constitutive beta-catenin-dependent transactivation in SW480 colon carcinoma cells and its nuclear localization can be inhibited by overexpressing N-cadherin or alpha-catenin. The results indicate that (a) plakoglobin and beta-catenin differ in their nuclear translocation and complexing with LEF-1 and vinculin; (b) LEF-1-dependent transactivation is preferentially driven by beta-catenin; and (c) the cytoplasmic partners of beta-catenin, cadherin and alpha-catenin, can sequester it to the cytoplasm and inhibit its transcriptional activity.

Intercellular adherens junctions between cultured lens epithelial cells are highly Ca2+-dependent and are readily dissociated upon chelation of extracellular Ca2+ ions. Addition of Ca2+ to EGTA-treated cells results in the recovery of cell-cell junctions including the reorganization of adherens junction-specific cell adhesion molecule (A-CAM), vinculin, and actin (Volk, T., and B. Geiger, 1986, J. Cell Biol., 103:000-000). Incubation of cells during the recovery phase with Fab' fragments of anti-A-CAM specifically inhibited the re-formation of cell-cell adherens junctions. This inhibition was accompanied by remarkable changes in microfilament organization manifested by an apparent deterioration of stress fibers and the appearance of fragmented actin bundles throughout the cytoplasm. Incubation of EGTA-dissociated cells with intact divalent anti-A-CAM antibodies in normal medium had no apparent inhibitory effect on junction formation and did not affect the assembly of actin microfilament bundles. Moreover, adherens junctions formed in the presence of the divalent antibodies became essentially Ca2+-independent, suggesting that cell-cell adhesion between them was primarily mediated by the antibodies. These studies suggest that A-CAM participates in intercellular adhesion in adherens-type junctions and point to its involvement in microfilament bundle assembly.

Specific binding (conjugation) of cytotoxic T lymphocytes (CTL) to target cells (TC) is the first step in a multistage process ultimately resulting in dissolution of the TC and recycling of the CTL. We examined the position of the microtubule organizing center (MTOC) of immune CTL bound to specific TC. Immunofluorescence labeling of freshly prepared CTL-TC conjugates with tubulin antibodies indicated that the MTOC in essentially all conjugated CTL but not in the conjugated TC were oriented towards the intercellular contact site. This finding was corroborated by electron microscopy examination of CTL-TC conjugates fixed either immediately after conjugation or during the lytic process. Antibody-induced caps of membrane antigens of CTL such as H-2 and Thy 1, did not show a similar relationship to the MTOC. Incubation of CTL-TC conjugates, 10-15 min at room temperature, resulted in an apparent deterioration of the microtubular system of conjugated CTL. It is proposed that the CTL plasma membrane proximal to the MTOC is particularly active in forming stable intercellular contacts, resulting in CTL-TC conjugation, and that subsequent modulation of the microtubular system of the CTL may be related to the cytolytic response and to detachment of the effector cell.

EGTA-induced depletion of Ca2+ ions from the culture medium of Madin-Darby bovine kidney epithelial cells results in rapid splitting of adherens-type junctions and the detachment of the vinculin- and actin-containing filament bundle from the cytoplasmic faces of the plasma membrane of the zonula adhaerens. This process was monitored by phase-contrast microscopy, combined with electron microscopy and immunofluorescent localization of the two proteins. It is shown that shortly after extracellular free Ca2+ concentration is lowered to the micromolar range, the actin-containing, junction-associated belt of microfilaments, together with the vinculin-rich junctional plaque material, is irreversibly detached as one structural unit from the plasma membrane, contracts, and is displaced towards the perinuclear cytoplasm where it gradually disintegrates. Other actin- and vinculin-containing structures present in the same cells, notably the focal contacts at the substratum, are not similarly affected by the Ca2+ depletion and retain both the adhesion to the external surface and the association with the plaque and microfilament components. Electron microscopic examination has shown that the membrane domain of the zonulae adhaerentes, unlike that of desmosomes, is not endocytosed after Ca2+ removal and that the displaced actin- and vinculin-containing plaque and filament belt are not associated with a particular membrane. It is further shown that upon restoration of normal Ca2+ levels in the culture medium, new intercellular contacts are established gradually by accretion of both vinculin and actin into new belt-like plaque- and microfilament-containing structures.

Actin, keratin, vinculin and desmoplakin organization were studied in primary mouse keratinocytes before and during Ca2+-induced cell contact formation. Double-label fluorescence shows that in cells cultured in low Ca2+ medium, keratin-containing intermediate filament bundles (IFB) and desmoplakin-containing spots are both concentrated towards the cell center in a region bounded by a series of concentric microfilament bundles (MFB). Within 5-30 min after raising Ca2+ levels, a discontinuous actin/vinculin-rich, submembranous zone of fluorescence appears at cell-cell interfaces. This zone is usually associated with short, perpendicular MFB, which become wider and longer with time. Later, IFB and the desmoplakin spots are seen aligned along the perpendicular MFB as they become redistributed to cell-cell interfaces where desmosomes form. Ultrastructural analysis confirms that before the Ca2+ switch, IFB and desmosomal components are found predominantly within the perimeter defined by the outermost of the concentric MFB. Individual IF often splay out, becoming interwoven into these MFB in the region of cell-substrate contact. In the first 30 min after the Ca2+ switch, areas of submembranous dense material (identified as adherens junctions), which are associated with the perpendicular MFB, can be seen at newly formed cell-cell contact sites. By 1-2 h, IFB-desmosomal component complexes are aligned with the perpendicular MFB as the complexes become redistributed to cell-cell interfaces. Cytochalasin D treatment causes the redistribution of actin into numerous patches; keratin-containing IFB undergo a concomitant redistribution, forming foci that coincide with the actin-containing aggregates. These results are consistent with an IF-MF association before and during desmosome formation in the primary mouse epidermal keratinocyte culture system, and with the temporal and spatial coordination of desmosome and adherens junction formation.

The recently described adherens junction-specific 135-kD protein (Volk, T., and B. Geiger, 1984, EMBO (Eur. Mol. Biol. Organ.) J., 3:2249-2260) was localized along cardiac muscle intercalated discs by immunogold labeling of ultrathin frozen sections. Analysis of this labeling indicated that the 135-kD protein, adherens junction-specific cell adhesion molecule (A-CAM), is tightly associated with the plasma membrane unlike vinculin labeling, which was present along the membrane-bound plaques of the fascia adherens. In cultured chick lens cells, A-CAM was associated with Ca2+-dependent junctions that were cleaved upon a decrease of extracellular Ca2+ concentrations to less than or equal to 0.5 mM. In the chelator-separated junction, A-CAM became exposed to exogenously added antibodies or to proteolytic enzymes. Upon addition of trypsin to EGTA-treated cells, A-CAM was cleaved into three major cell-bound antigenic peptides with apparent molecular masses of 78, 60, and 46 kD, suggesting that the extracellular domain of A-CAM has a size greater than or equal to kD. Incubation of electrophoretic gels with 125I-concanavalin A (Con A) indicated that one of the major Con A-binding proteins in chicken lens membranes is a integral of 135-kD glycoprotein that was partially purified on Con A-Sepharose column and identified as A-CAM by immunoblotting. Detergent partitioning assay using Triton X-114 biphasic system was carried out to determine whether A-CAM displays properties of an integral membrane protein. This assay indicated that the intact A-CAM molecule was recovered in the buffer phase but its cell-associated tryptic peptides, which presumably lost a great part of the A-CAM extracellular extension, readily partitioned into the detergent phase. The results obtained in this and in the following paper (Volk, T., and B. Geiger, 1986, J. Cell Biol., 103:1451-1464) strongly suggest that A-CAM is a Ca2+-dependent adherens junction-specific membrane glycoprotein that is involved in intercellular adhesion in these sites.