We have been analyzing the ability of constructs containing fusions of the
ced-3 and
ced-4 cDNAs to different promoters to kill cells in which these promoters are active. We have used promoters for the
mec-7 gene (expressed in the touch cells and a few other cells; A. Fire, M. Hamelin, and J. Culotti, personal communication), the
unc-30 gene (expressed in the VD and DD neurons; Y. Jin, personal communication), and the heat-shock promotor (expressed throughout the animal; A. Fire, P. Candido, personal communication). We initially examined integrated lines containing both a
mec-7 -
ced-3 and a
mec-7 -
ced-4 construct in
ced-9 (
n2812);
ced-3 (
n717)animals to avoid any potential inhibitory effects of
ced-9 .We found that the touch cells were missing in these animals. By observing the cell lineages of QL and its progeny we found that PVM and occasionally SDQL undergo a cell death process that is morphologically indistinguishable from normal programmed cell deaths. The
mec-7 promoter we have been using is active in both PVM and SDQL. These results support the idea suggested by mosaic analysis of
ced-3 and
ced-4 (Yuan and Horvitz, Developmental Biology, 138, 33-41 1992) that these genes act cell-autonomously to kill cells, since we never observed any dying cells near PVM or SDQL. These results also suggest that the machinery necessary for engulfment of cell corpses is either constitutively present or is induced by the activities of
ced-3 and
ced-4 ,since engulfment of corpses was generally normal. In one case we directly observed the hypoderm engulfing a dying PVM. To elucidate interactions among
ced-3 ,
ced-4 ,and
ced-9 ,we have begun to analyze the effects of the constructs described above in various genetic backgrounds. So far we have found that introducing these constructs into animals of different genetic backgrounds has different effects. Killing of target cells is more efficient in the absence of endogenous
ced-9 ,since
ced-3 fusions kill more efficiently in
ced-9 (
n2812);
ced-3 (
n717)animals than in
ced-3 (
n717)animals, and
ced-4 fusions kill more efficiently in
ced-4 (
n1162)
ced-9 (
n2812)animals than in
ced-4 (
n1162)animals. We also found that the
mec-7 -
ced-3 fusions kill less efficiently in the
ced-4 (
n1162)
ced-9 (
n2812)background than in the
ced-9 (
n2812);
ced-3 (
n717)background and that
mec-7 -
ced-4 fusions kill less efficiently in the
ced-9 (
n2812);
ced-3 (
n717)background than in the
ced-4 (
n1162)
ced-9 (
n2812)background. These results suggest that
ced-3 function is augmented by
ced-4 function and vice versa and also suggest that
ced-3 and
ced-4 are normally expressed in the touch cells even though these cells normally do not die. Finally, we would like to suggest that
ced-3 and
ced-4 can be used to genetically ablate cells in which a given promoter is active, and in the near future we hope to make available convenient vectors that can be used with any promoter.