During the development of a C. elegans hermaphrodite 131 cells from many different tissues undergo programmed cell death. Junying Yuan used mosaic analysis to show that these deaths are probably suicides ( WBG 9(1):58). How, then, do cells decide to die? To understand this we are studying genes that affect this decision in four cells in the pharynx: the sisters of NSML and NSMR, and of I2L and I2R. In wild- type animals these four cells all die, but in
ces-1(
n703sd) and
ces-2(
n732ts) worms they can survive (WBG 9 (2): 61 and CSH Abstracts 1987 pg .149; ces stands for cell death specification). We have shown that
ces-1(
n703) and
ces-2(
n732) can eliminate four dead cell bodies normally found in the pharynxes of
ced-1; which suggests that they can prevent four specific cell deaths. Furthermore, John Sulston directly observed that the NSM and I2 sisters in
ces-1 (
n703) animals do not die ( personal communication) . Though mutations in
ces-1 and
ces-2 can prevent these four cell deaths, these mutations do not appear to affect any.other dying cells. In
ces-2(
n732ts) and
ces-2(
n732)/Df worms at 25 C the NSM sisters survive about half of the time, but the I2 sisters only rarely. Furthermore, in both
ces-2(
n732)/+ and Df/+ worms the NSM sisters survive a few percent of the time, although they always die in the wild type. Since in both experiments
ces-2(
n732) at 25 C behaves like a deficiency,
n732 probably reduces
ces-2 gene activity. Because
n703 is a semidominant allele of
ces-1, and so might increase gene activity, we measured the effects of different doses of this gene on cell survival. Deficiencies in this region do not show the semidominant phenotype: in Df/+ animals all four cells die just as in the wild type . Furthermore, in
ces-1 (
n703)/Df worms the I2 sisters survive much less frequently than in
ces-1 (
n703)/+ animals . This suggests that the allele
ces-1 (
n703) does have increased gene activity. To explore this hypothesis, we reverted the dominant phenotype of
ces-1 (
n703) to find alleles of this gene with reduced or no activity. From a screen of about 10,000 haploid genomes we found two revertants. Both are closely linked to
ces-1(
n703) and both eliminate the effect of
n703 in cis, so we believe they are probably intragenic mutations. Because in gene dosage studies both revertant alleles (
ces-1 (
n703 n1406) and
ces-1 (
n703 n1434) ) show much less activity than a + allele, but more activity than deficiencies, they probably reduce gene function significantly. (They may even eliminate gene function, but if so, the deficiencies we used must have a weak influence on cell survival beyond the fact that they do not contain
ces-1.) In both revertant strains the NSM and I2 sisters die as in the wild type, and no other cells appear to be affected. We believe these revertants restore normal programmed cell death, because in
ces-1 (
n703 n1406);
ced-3 and
ces-1 (
n703 n1434);
ced-3 strains the NSM and I2 sisters survive, just as in other
ced-3 strains. To study the interactions of
ces-1 and
ces-2, we built the
ces-1(
n703 n1406)
ces-2(
n732) double mutant. Preliminary results show that in this strain the NSM and I2 sisters die as in the wild type, indicating that
ces-1 gene product is required if mutations in
ces-2 are to prevent these cell deaths. One simple model that explains all of these relationships is that the genes involved in the NSM and I2 sister deaths act in this order.
ces-2 -->
ces-1 -->
ced-3,
ced-2 - -> nuc-