The initial insight into the genetic basis of apoptosis, or programmed cell death, was gained from ingenious studies of the roundworm Caenorhabditis elegans (for review, see Horvitz 1999). These studies revealed a linear pathway whereby the products of two genes, designated Ced-3 and Ced-4, were necessary and sufficient to trigger the perfectly timed and orchestrated death of 131 preordained cells during development. The relevance of this pathway to higher animals was established by the discovery of apparent mammalian orthologs of these genes and the demonstration that the mammalian Ced-3-related genes encode proteases (designated caspases) whose activities are responsible for the morphological changes characteristic of apoptosis (for review, see Hengartner 2000).

The complexity of the apoptotic program began to increase with the discovery of Bcl-2, a gene whose product causes resistance to apoptosis in lymphocytes (Vaux et al. 1988; McDonnell et al. 1989). Bcl-2 was shown to correct partially the phenotype of a C. elegans mutation in Ced-9, a cell survival gene that functions upstream of Ced-4 and Ced-3 (Vaux et al. 1992). This finding suggested an apparent one-for-one correlation between the C. elegans and mammalian pro-and antiapoptotic pathways. However, this correlation did not explain two observations made in mammalian cells. First, the Bcl-2 protein was found on the membrane of mitochondria, which were not implicated in C. elegans apoptosis; and second, apoptotic changes could be produced in Xenopus laevis oocyte extracts only when a membrane fraction enriched in mitochondria was present (Hockenberry et al. 1990; Newmeyer et al. 1994).