The cytoplasmic domain of the cellular adhesion molecule integrin assembles a plaque of adaptor and signaling proteins that regulates many fundamental cellular functions including adhesion, growth, differentiation, and apoptosis. This protein complex has been most extensively analyzed in the focal adhesions, attachments that are formed by vertebrate cells in culture. C. elegans body wall muscle cell dense bodies and M-lines are analogs of focal adhesions that are amenable to genetic analysis. We are contributing to a comprehensive genetic dissection of these structures by studying a subset of the pat genes ( P aralyzed, A rrested elongation at T wo-fold) that are required for their normal assembly. A partial list of these genes includes
pat-2 and
pat-3, which are integrin subunit genes (Williams and Waterston, unpubl.; Gettner et al. , 1995, JCB 129:1127),
unc-112 , which encodes a novel dense body/M-line component (Rogalski et al ., 2000, JCB 150:235),
pat-4 , which encodes integrin-linked kinase (see abstract by Mackinnon et al. , this meeting), and
pat-6 , which we focus on here. We molecularly isolated the
pat-6 gene and have found that it encodes the worm homologue of actopaxin, a human actin binding protein discovered very recently in a yeast two-hybrid screen for proteins that bind paxillin, a focal adhesion component (Nikolopoulos and Turner, 2000, JCB 151:1435). Although a recent database search failed to identify any other proteins homologous to the N-terminal half of PAT-6, the C-terminal half includes two calponin homology (CH) domains, a feature of many actin-binding proteins. PAT-6 is located in body wall muscle dense bodies and M-lines. It also colocalizes with integrin at several other sites including the attachments of single sarcomere muscles, within the spermatheca, and within touch neurons. We are using the collection of pat mutants in combination with antisera and/or functional GFP-tagged fusion proteins to investigate the role of
pat-6 in dense body and M-line assembly. To date, we have found that
pat-6 is not required for the normal assembly of UNC-52/Perlecan into the adjacent basal lamina, for the initial polarization of integrin to the basal membrane of muscle cells, nor for the entry of integrin into nascent dense bodies and M-lines. In contrast,
pat-6 is required for maturation of dense bodies and M-lines, including the proper assembly of at least one membrane-distal protein, UNC-89, and for the arrangement of these nascent attachments into a well-ordered striated array. In reciprocal experiments we have addressed the requirements for proper PAT-6 assembly. As predicted, PAT-6 requires UNC-52/perlecan and integrin to assemble normally at the sarcolemma. We were surprised to find that PAT-6 is not detectable in body wall muscle cells when PAT-4 has been removed by mutation. Although this result is consistent a number of interesting possibilities, we currently favor the hypothesis that PAT-4 and PAT-6 interact, perhaps directly, and that this interaction stabilizes the PAT-6 protein. Consistent with this idea, we have found through yeast two-hybrid experiments that PAT-4 binds directly to PAT-6. The kinase domain of PAT-4 and C-terminal CH domain of PAT-6 are both necessary and sufficient for binding. We failed to detect interactions between PAT-6 and any of several other dense body proteins in similar two-hybrid experiments. In our initial structure/function analysis of PAT-6, we have assayed ability mutant PAT-6 lacking either the N-terminal or the C-terminal CH domain to 1) localize properly in animals that are also expressing wild-type PAT-6, and 2) to rescue
pat-6 null animals. Our results to date show that N-terminal CH domain is not required for localization, but is required for rescue. In contrast, the C-terminal CH domain is required both functions. Current experiments are addressing the function of the N-terminal half of PAT-6.