To find genes required for differentiation of non-neuronal ectodermal cells, we have looked for mutations displaying similar defects as mutations in
lin-26, a gene that has been shown to specify and/or maintain the fates of hypodermal and glial-like cells. The
che-14(
e1960) mutation had been shown to display ultrastructural defects in glial-like cells (Perkins et al., Dev. Biol. 56: 110-156, 1986) similar to those observed in
lin-26(
n156) mutants by electron microscopy (Labouesse et al., Development 122: 2579-2588, 1996). This similarity prompted us to investigate in more details the function of
che-14. Using a dye-filling test as an assay, we isolated one new
che-14 mutation,
mc16, in a non-complementation screen against
e1960 after examining 10000 genomes. We found that
mc16, as
e1960, affects chemosensory organs by closing the amphid and phasmid channels and that both induce a partial larval lethality probably due to a hypodermal defect. We have cloned
che-14 by complementation rescue of the dye filling defect due to
e1960. A 9 kb EcoRI subclone gives perfect rescue and contains only one ORF. The
che-14 gene is predicted to encode a protein with several transmembrane domains that displays 45% similarity/20% identity with PATCHED, the HEDGEHOG receptor. We identified the two known
che-14 mutations, showing that we have indeed cloned
che-14, as G->A transitions that affect splice donor sites located towards the 3' end of the gene. Their nature and positions raise the possibility that
e1960 and
mc16 are partial loss-of-function mutations. To determine the null phenotype, we are screening a library of TMP/UV mutagenized animals using primers within
che-14. To look at the expression pattern of
che-14, we constructed a
che-14::gfp fusion. This fusion is expressed in (probably all) non-neuronal ectodermal cells which is consistent with the Che-14 phenotype. The subcellular localisation is membranous, consistent with the predicted transmembrane structure. Based on the CHE-14 sequence, computer programs predict several topologies and even different numbers of transmembrane domains (between 9 and 11). To determine the CHE-14 topology we are using a strategy first inspired from the work on SEL-12 (Li X. & Greenwald I., Neuron, 17: 1015-1021, 1996). After inserting the GFP before and after a predicted transmembrane domain, we have observed that one fusion is fluorescent but not the other, although both are expressed as shown by staining of the fusion protein with antibodies against GFP. We suppose that the GFP can only fluoresce if intracellular and we will further use this strategy to determine the topology of CHE-14.