It all started two years ago, say the authors, with the discovery of p53 in Drosophila melanogaster. They reasoned that if Drosophila has p53, then maybe there are parallels between the way that flies and mammals induce p53-dependent cell death. And it seems that they were right, as they now report in Genes and Development.

The authors in question — Arnold Levine and colleagues — started with a p53-dependent pathway that had been observed in flies but not in mammals. Drosophila p53 activates the transcription of a gene called Reaper, the product of which interacts with the inhibitor of apoptosis (IAP) protein DIAP and targets it for ubiquitin-mediated proteolysis. As DIAP's role is to inhibit caspases, its destruction frees up caspases to proceed with an apoptotic response.

So could this happen in mammals? To test this, Levine and co-workers first looked for proteins that might interact with CIAP1 — the human homologue of DIAP1. They generated stable clonal HeLa cell lines that expressed a haemagglutinin-Flag-tagged CIAP1 protein, and then co-purified CIAP1 and its partners from HeLa cell lysates using M2-agarose beads, which recognize the Flag tag. In response to etoposide-induced p53-dependent apoptosis, the authors pulled down two low-molecular-weight proteins. But these weren't CIAP1's binding partners — they were fragments of CIAP1 itself.

Levine and colleagues concluded that CIAP1 is cleaved during p53-dependent apoptosis. When they induced apoptosis in p53-independent manner (using anti-Fas antibodies), the cleavage was no longer seen, suggesting that it requires the p53 pathway. As one characteristic of this pathway is the transcription of p53 target genes, the authors then asked whether de novo protein synthesis is needed for the cleavage. Treatment with cycloheximide blocked the cleavage, consistent with this idea.

Cleavage of CIAP1 was not blocked by caspase inhibitors, which means that the cleavage is independent of caspases and is a cause rather than an effect of the cell-death pathway. So might another type of protease be involved in the cleavage? Serine proteases have previously been implicated in apoptosis, so Levine and co-workers treated HeLa cells with etoposide and a general serine protease inhibitor, and found that both CIAP1 cleavage and apoptosis were blocked. They obtained similar results using primary mouse thymocytes — thymocytes with no p53 genes failed to cleave CIAP and failed to undergo apoptosis.

To identify the protease involved, Levine and colleagues went back to the literature. Previous reports had shown that mammalian IAPs can interact with a serine protease called HTRA2/OMI, so the authors did a northern blot analysis of HTRA2 messenger RNA levels in HeLa cells during treatment with etoposide. They observed a sevenfold increase in HTRA2 mRNA levels; a similar increase was also seen when HeLa cells were transfected with a p53 expression vector.

The authors therefore conclude that a parallel pathway to that first mapped out in flies indeed exists in mammals, with the subtle difference that CIAP1 is destroyed by protease-mediated cleavage rather than being targeted by ubiquitin for destruction.