The activation of apoptotic cascades triggers a series of events, one of which is the condensation and fragmentation of chromosomal DNA. In mammals, at least three proteins have been implicated in this process — DFF40/CAD, the apoptosis-inducing factor (AIF) and a mitochondrial endonuclease termed endonuclease G (EndoG). How do these proteins work, and are their pro-apoptotic functions conserved in other species? These are the questions tackled by Ding Xue and colleagues, who report their results in Science.

The authors started by cloning wah-1, which is the Caenorhabditis elegans homologue of AIF. They then examined the effects of inhibiting the expression of wah-1 using RNA-mediated interference (RNAi) and found that, in wah-1(RNAi)-treated worms, the appearance of embryonic cell corpses during normal development was delayed. This phenotype is similar to that observed in cps-6 mutant worms (CPS-6 is the worm orthologue of mammalian EndoG), and so indicates a role for WAH-1 in the progression of apoptosis. However, a cps-6 mutation did not enhance the wah-1(RNAi) phenotype, indicating that WAH-1 and CPS-6 might function in the same pathway.

Xue and co-workers next used TUNEL assays to show that, like AIF, WAH-1 induces chromosome fragmentation. And, using a WAH-1–green fluorescent protein (GFP) fusion protein, they showed that WAH-1 localizes to the mitochondria. So how is WAH-1 activated? The release of apoptogenic factors from mitochondria in response to pro-apoptotic stimuli is triggered by so-called 'BH3-domain-only' proteins. In worms, the most upstream cell-death activator is also a BH3-domain protein, EGL-1. The authors asked whether global expression of egl-1 could release WAH-1 from the mitochondria, and they found that it could — the punctate mitochondrial staining pattern of WAH-1–GFP became more uniformly spread throughout the cell, and more GFP fluorescence was seen in nuclei, indicating the release of this protein from the mitochondria to mediate chromosome fragmentation.

These results indicate that the mitochondrial cell-death pathway is indeed conserved in worms, but what about the mechanism of AIF-mediated DNA degradation? Hints that WAH-1 and CPS-6 act in the same pathway led Xue and colleagues to examine whether these two proteins interact directly. GST-fusion-protein pulldown experiments confirmed that they interact in vitro, and the authors further showed that they act synergistically to induce cell killing. They monitored this in vitro by looking at the cleavage of DNA at various concentrations, and found that cleavage occurred at lower concentrations of protein when the two proteins were incubated together than when either protein was used alone. In vivo, the ectopic expression of wah-1 or cps-6 alone in touch-receptor neurons led to the death of 5–8% or 15–23% of the neurons, respectively. When cps-6 and wah-1 were expressed together, however, 53–75% of the neurons were killed.

These results, say the authors, indicate that “EndoG and AIF ... define a single conserved DNA-degradation pathway initiated from the mitochondria”. However, there are some differences between the worm and human pathways. For example, whereas AIF acts in a caspase-independent manner, WAH-1 requires functional CED-3 (the worm caspase homologue) for full activity. Moreover, unlike AIF, WAH-1 does not bind FAD and does not seem to have an oxidoreductase activity. These data fit in well, then, but the puzzle is still far from solved.