The dy e- f illing ( dyf ) defective gene, dyf-8 , was first isolated in a screen for animals that fail to take up a lipophillic dye into amphid and phasmid neurons (Starich et al , 1995, Genetics 139: 171-188 ). We cloned this gene using two novel strategies: SNP mapping (Wicks et al. in press) and Mos transposon-insertion mutagenesis. The original m509 allele of dyf-8 was isolated from a mutator strain, and therefore possibly arose as a consequence of a Tc1-insertion event. A transposon-insertion display (Wicks et al , 2000. Dev. Biol, 221: 295-307 ) was used to identify all inserts that co-segregated with the mutant phenotype. We identified genomic flanks for six Tc1 inserts near the centre of LGX, ranging in position from --3.42 mu to +1.16 mu. We then crossed dyf-8 into a CB4856 background. Bulked segregant analysis was used to verify the approximate map position and identify snip-SNPs that flank the mutation. Recombination events between these flanking markers were mapped with a high density SNP map, and the recombinants phenotyped. We were able to map the mutant phenotype with the progeny from a single cross to a 40 kb interval that included one of the six known transposon inserts in the dyf-8 mutant. This insert was in sixth exon of the predicted ORF C43C3.3. In a parallel series of experiments, we used the Mos1 transposon from Drosophila mauritiana to identify mutants ( Bessereau,J.-L. et al . 2000, European WM abstract) defective for the detection of high osmolarity (Osm). We determined whether the mutants were defective for the uptake of a lipophillic dye into amphid and phasmid neurons. Seven of the ten mutants were Dyf. Of these, three contained transposon insertions suggesting that the defects were possibly caused by the insertion event. We used inverse single worm PCR (ISWP) to clone genomic sequences flanking the transposons. One of the mutants (ox195) contained an insert in the tenth exon of C43C3.3. m509 animals, transgenic for the C43C3 cosmid, were completely rescued with respect to the Dyf phenotype. C43C3.3 is predicted to encode a peptide with three suggestive features: First, the N-terminus consists of a signal sequence, suggesting that the peptide is secreted. Second, a single predicted transmembrane domain is found near the C-terminus. In between, on what would be the extracellular domain of this membrane-bound protein, is a Zona Pellucida domain, which has been shown in other proteins to mediate some forms of protein-protein interactions on the extracellular surface of cells. Our working hypothesis is that dyf-8 directly, or indirectly, mediates the stabilisation of the cilia/support cell interactions. We are testing this hypothesis through further genetic and biochemical analysis of the gene.