CED-4—The Third Horseman of Apoptosis

activating adaptor proteins have been identified. For The remarkable conservation of physiological cell death example, Mort-1/FADD interacts with certain caspases mechanisms from nematodes to humans has allowed via shared death effector domains (DEDs) and RAIDD/ the genetic pathways of programmed cell death deterCRADD interacts with other caspases via shared CARDs mined in Caenorhabditis elegans to act as a framework to transduce proapoptotic signals, such as those origifor understanding the biology of apoptosis in mammanating from TNF family receptors (Hofmann et al., 1997). lian cells. However, it has been unclear whether the C. In most cells, these cell death signals are poorly blocked elegans cell death gene ced-4 would have a parallel in by BCL-2, whereas BCL-2 is potent at inhibiting cell mammalian cell death. In this issue of Cell, Zou et al. death triggered by other stimuli, such as withdrawal of (1997) report the biochemical identification of a human growth factor, or up-regulation of p53 (Strasser et al., cell death protein whose sequence resembles the nema1995). It may be that the ability of BCL-2 to inhibit tode protein CED-4, providing a spectacular demonstraapoptosis depends upon which adaptor molecule is intion of the combined power of genetic analysis of a volved in caspase activation. simple organism with biochemistry of mammalian cells. Determining precisely how BCL-2-like proteins funcThe Genetic Framework tion has been surprisingly difficult. In various experimenAlthough the first cell death gene to be characterized tal systems, BCL-2 has been found to interact with calat the molecular level was a mammalian gene, bcl-2, cineurin, p53 binding protein, lamin, NIP-1, -2, and -3, genetic analysis of C. elegans mutants had already R-RAS, RAF-1, BAG-1, and galectin-3, as well as intershown that, in the worm, physiological cell death reacting with other BCL-2 family members. Recently it has quired the activity of two ‘‘killer’’ genes, ced-3 and ced-4 been proposed that BCL-2 functions by forming pores (Ellis and Horvitz, 1986). The discovery that human bcl-2 to allow ions or small molecules to cross the outer mitohas functional and structural similarity to the anti–cell chondrial membrane (Reed, 1997). As release of cytodeath gene ced-9 demonstrated that programmed cell chrome c from the intermitochondrial space has been death in the nematode occurred by the same highly associated with apoptosis in a number of systems, and conserved mechanism as apoptosis in mammalian cells as this release is prevented by BCL-2, it has been hy(Vaux et al., 1992; Hengartner and Horvitz, 1994). The pothesized that the key function of BCL-2-like proteins killer gene ced-3 was found to encode a cysteine proteis to somehow retain cytochrome c in the mitochondria ase with aspartic acid specificity (caspase) and was (Kluck et al., 1997; Yang et al., 1997). the archetype of a family of caspases that are the key