We have been studying
oz36, an apparent recessive gain-of-function allele of the
let-42 gene that displays a variable tumorous germline phenotype. In some cases,
oz36 hermaphrodites have late-onset tumorous germ lines. That is, the distal-to-proximal polarity normally found within each gonad arm is established correctly, but with time the distal mitotic zone expands and in extreme cases fills the arm entirely. In other cases, the germline polarity does not appear to be established appropriately. Rather, with the exception of a few sperm (usually < 15), germ cells at the proximal end of the gonad arm continue to proliferate. DAPI staining of gonads dissected from such animals reveals that they contain both the normal distal-to-proximal gonad polarity (mitotic germ cells followed by meiotic germ cells followed by oocytes) as well as a second, reversed, polarity (oocytes are next followed by meiotic and then mitotic germ cells). Which, if either, of the germline phenotypes arises depends on three variables: 1) temperature (cs), 2) maternal genotype (+/oz36 vs
oz36/oz36 ), and 3) smg mutant background (enhancement of the tumorous phenotype in a smg mutant background is consistent with
oz36 acting as a gain-of-function). At 20 in a smg(+) background, greater than 90% of the homozygous
oz36 self-progeny of
oz36/+ hermaphrodites are phenotypically wild type; the remainder have a late-onset tumorous phenotype. In contrast, 96% of the
smg-1;
oz36 self-progeny at 15 display proximal proliferation while 4% are late-onset tumorous. Conditions intermediate to those just described result in most animals showing a late-onset tumorous phenotype. Thus, proximal proliferation appears to be a more severe defect than the late-onset tumorous phenotype.
oz36 males can also be tumorous. Two other genes that can mutate to a tumorous germline phenotype,
gld-1 and
glp-1, have been described. In
gld-1(null) mutants, female, but not male, germ cells fail to progress beyond the pachytene stage of meiotic prophase I and instead re-enter the mitotic cell cycle and form tumors (1). In contrast,
glp-1(gf) female and male germ cells either never leave the mitotic cell cycle (thus forming tumors) or the gonads are late-onset tumorous (2).
oz36 germline tumors appear most similar to
glp-1(gf) tumors. Therefore, we have begun to examine whether
oz36 interacts with the
glp-1 signaling pathway. Strong
glp-1(lf) alleles, including
q175, the putative null allele (3), are epistatic to
oz36. However, both the germline (25) and embryonic (20) defects of
glp-1(
q231ts), a partial loss-of-function allele (3), are suppressed by
oz36. Moreover,
g38, a second (loss-of-function?) allele of
let-42 (4), suppresses the
q231 embryonic defect at 20 but not the germline defect at 25. At 15,
q231oz36 double mutants have phenotypically wild type, not tumorous, germ lines. Taken together, these data suggest that
oz36 requires
glp-1 activity for tumor formation. In addition,
let-42 may negatively regulate
glp-1 .
oz36 and
g38 also display a maternal effect sterile phenotype at 25. A 3kb genomic fragment containing a single predicted gene (ZK1128.2) (5) was found to be sufficient for rescue of the ts MES phenotype. Sequence analysis of PCR products generated from a C. elegans cDNA library revealed that the exons had been predicted correctly and that
let-42 is spliced to SL1.
let-42 is predicted to encode a 495 a.a. protein. The only significant homology within the computer database is to a human aorta EST. Sequence analysis of
oz36 and
g38 is almost complete. To date, a single point mutation has been found within each allele.
g38 has a missense mutation while
oz36 contains a nonsense codon. Both mutations are approximately halfway through the
let-42 coding region. 1) Francis et al., 1995. Genetics 139: 579-606. 2) Berry et al., in preparation. 3) Austin and Kimble, 1987. Cell 51: 589-599. 4) Cassada et al. 1981. Devel. Biol. 84: 193-205. 5) Wilson et al., 1994. Nature 368: 32-38.