In C. elegans, null alleles of the a-integrin (
pat-2) and b-integrin (
pat-3) genes result in lethality. Mutant embryos are Paralyzed and Arrest at the Two-fold stage of embryogenesis. This Pat phenotype is attributed to failure of the myofilament lattice in body-wall muscle cells to organize and become contractile. Therefore, mutational and functional characterization of sarcomere assembly provides a model for analysis of integrin-mediated signaling pathways.
pat-4 was identified in a screen for genes required for this assembly process. Transformation rescue of
pat-4 reveals it is the homologue of the vertebrate gene integrin-linked kinase (ILK) (Hannigan, 1996). PAT-4 and ILK are 58% identical and contain four ankyrin repeats at the N-terminus, a PH-domain, and a C-terminal kinase domain. In vitro studies show the kinase domain of ILK binds b-integrin (
pat-3), and the ankyrin repeats bind PINCH, which is homologous to
unc-97 (Tu, 1999). Interestingly,
unc-97(RNAi) was recently shown to generate the Pat phenotype (Hobert, 1999). Overexpression of ILK in tissue culture cells promotes focal adhesion and fibronectin assembly, mediates Wnt signaling, and regulates the activities of GSK-3 and AKT/PKB (Wu, 1998). We constructed a
pat-4::GFP fusion gene which completely rescues the mutant phenotype. The fusion gene product is expressed in body-wall muscle cells where it co-localizes subcellularly with both PAT-3 and UNC-97 at M-line and dense body attachment sites. In addition,
pat-4::GFP is expressed in the touch neurons PLM and ALM , the vulva, uterine, and anal depressor muscle cells, the spermatheca, and the distal tip cells.
pat-4::GFP is also expressed transiently in the pharynx during embryogenesis. Ectopic expression of a full length mouse ILK::GFP fusion gene, which is 99% identical to human ILK, also localizes to dense body and M-line attachment sites. We were able to rescue
pat-4 mutant animals when ectopic expression of
pat-4 was restricted to body-wall muscle cells using the
myo-3 promoter. These rescued animals grow slowly, have a reduced brood size, and are dumpy with a truncated tail. We are currently trying to determine if these mosaic animals have a neuronal phenotype due to a loss of
pat-4 expression in the touch neurons. Current models of myofilament assembly suggest perlecan (UNC-52) is secreted by body-wall muscle cells on their basal surface, and integrin heterodimers (PAT-2/PAT-3) polarize to this surface and initiate myofilament assembly. Subsequently, the dense body protein vinculin (DEB-1) polarizes to these sites and M-line and dense bodies assemble. In
pat-4 mutants, both UNC-52 and PAT-3 are able to localize properly, but DEB-1 and the MH-42 antigen (an M-line component) are disorganized. This suggests
pat-4 activity is required for the transduction of downstream integrin-mediated signaling events required for assembly. In order to determine if PAT-3 and PAT-4 physically interact in vivo we have recently started an experiment which will measure fluorescent resonance energy transfer (FRET) between PAT-3::CFP and PAT-4::YFP fusion proteins using fluorescent lifetime imaging microscopy (FLIM). So far, we have been able to make lifetime measurements of PAT-4::GFP and are currently generating the strains required for the FRET analysis.