Everybody has played with
unc-32(
el89) worms, but the molecular identification of this gene has proven elusive. We still are not sure, but have some preliminary evidence that it could correspond to one of the three presently known vacuolar ATPase genes, C05D12.1 on II, F35H10.4 on IV and ZK637.8 on III. In 1988, John Sulston rescued the
el89 allele by transformation with cosmid ZK637. Four more alleles, among which two lethals, were isolated in our lab. The locus became complex (wbgll.5p93). The first injections of subclones of ZK637 pointed to the glutathione reductase as the preferred candidate, but some of the transformants turned out to be gene specific suppressors; in addition, no mutations were found by sequencing the GR gene. More subclones of ZK637 were recently injected at 10 ng/ul, together with pRF4 (100 ng/ul) and/or pPD93-97 (GFP expressed under
myo-3 in the body muscles, at 20 ng/ul). The results are presented in the table below. The data indicate that an 8.2kb fragment (E5-
lin9) is able to rescue at least one Unc and one Let allele. We noticed however, that most arrays rescuing the Uncs are unable to rescue the Lets when crossed in. There may be a certain dosage needed to rescue the lethals, so that negative results in the table may become positive when more arrays are tested (the present number tested is given in parenthesis). Also, even when the lethals are rescued, the adults are usually sterile. The minimal rescuing fragment can encode a vacuolar proton pump homologue. In collaboration with Yuji Kohara, we showed that at least three alternative transcripts of this gene are expressed in the worm: this could account for the complexity of the locus. [THE ORIGINAL DOCUMENT CONTAINS A FIGURE.] Although the
unc-32 locus may well turn out to control the pH of some vesicles, two observations prevent us from being overconfident that we have finally identified the gene: 1- The construct E5-Xk, where the proton pump should be inactivated by a frameshift in the third exon common to the two cDNAs sequenced, rescues the
el89 allele. 2- An identical mutation, G1339A, was found in two independent alleles,
el89 and
fl20, but it lies in the middle of a large intron (1036 or 502 bp long, depending on the alternative splicing).