In C. elegans hermaphrodites, 131 cells undergo programmed cell death during normal development. We have isolated a new semidominant suppressor of all programmed cell deaths,
nl950. Homozygous
nl950 animals show widespread cell survival and are phenotypically indistinguishable from
ced-3 or
ced-4 mutants. However, whereas the phenotypes of
ced-3 and
ced-4 mutants are recessive, that of
nl950 is semidominant and displays a significant maternal effect. The
nl950 mutation maps to the right arm of LG III, and defines a new gene involved in programmed cell death,
ced-9. Several lines of evidence suggest
ced-9(
nl950) affects all programmed cell deaths. First, the number of extra cells surviving in the anterior pharynx is approximately the same as in
ced-3 and
ced-4 mutants. Second, although many unengulfed cell corpses are seen in ced-l or
ced-5 animals, these corpses are not present in ced-l;
nl950 or
nl950;
ced-5 double mutants. Third,
ced-9(
nl950) suppresses egl-l mutations, which normally cause the HSN neurons to inappropriately undergo programmed cell death. The semidominance of
nl950 suggests that this mutation does not cause a loss of
ced-9 function. We sought loss-of-function alleles of
ced-9 by reverting
nl950. From a screen of 9,000 haploid genomes, we isolated three cisdominant suppressors of
nl950, all tightly linked to
ced-9. One of these suppressors is a deficiency that deletes the entire
ced-9 locus as well as adjacent genes on each side. The other two appear to be bona fide intragenic revertants. The loss of
ced-9 function results in several phenotypes, the two most striking being maternal-effect lethality and zygotic semi-sterility: homozygous
nl950 n2077 animals derived from heterozygotes are viable, but produce only a few eggs, which never hatch. The other revertant,
nl950 n2161, is weaker and somewhat temperature-sensitive. Since
nl950 results in a gain of
ced-9 function and blocks cell death, we reasoned that the loss of
ced-9 function might be lethal because it leads to ectopic cell death. This hypothesis predicts that
ced-3 or
ced-4 mutations could suppress the lethality of
ced-9(lf). Indeed, both
ced-9(lp;
ced-3 and
ced4 ced-9(lp worms generate viable and fertile progny. These experiments demonstrate that the functions of
ced-3 and
ced-4 are essential for the expression of the
ced-9(lf) lethal phenotype. An interesting possibility is that the
ced-9 product prevents
ced-3 and
ced-4 from acting in cells destined to survive, thus protecting these cells from programrned cell death. By contrast, the
ced-9 gene product must be inactivated in cells destined to die, which allows the
ced-3 and
ced-4 activities to kill these cells. We propose that
ced-9(lf) homozygous progeny of
ced-9(lf)/+ heterozygotes do not die because of a matemal contribution of wild-type
ced-9 product to the egg. We are currently cloning
ced-9. RFLP mapping followed by microinjection of cosmids have localized
ced-9 rescuing activity to a 6 kb region covered by the cosmid F46Hl.