A defining characteristic of apoptosis is chromosomal DNA

apoptosis is chromosomal DNA degradation by deoxyribonuclease (DNase)-mediated fragmentation. In mammals, these DNases include DNA fragmentation factor 40 kDa (DFF40) and its chaperone and inhibitor, DFF45. During apoptosis, caspases cleave DFF45, releasing DFF40 to cleave DNA. In Caenorhabditis elegans, however, no homologues of DFF40 and DFF45 have been identified, and it is unclear how apoptotic DNA degradation is initiated. Nakagawa et al. now reveal that in C. elegans, this process is mediated by the RNase Dicer 1 (DCR-1). The nuclease CPS-6 (CED-3 protease suppressor 6) is involved in apoptotic DNA degradation, and CPS-6-mutant worms accumulate TUNEL-stained nuclei owing to a defect in resolving 3′ hydroxyl DNA breaks (which are labelled by TUNEL). The authors carried out RNA interference (RNAi)-mediated knockdown of nucleases in CPS-6mutant worms and found that DCR-1 knockdown reduced the number of TUNEL-stained cells. DCR-1deletion mutants showed a strong reduction of TUNEL-stained cells and contained fewer apoptotic cells, suggesting that DCR-1 is required to generate TUNEL-reactive DNA breaks in apoptotic cells and promote apoptosis upstream of CPS-6 and other nucleases. DCR-1 processes doublestranded RNA (dsRNA) to generate small RNAs that are involved in gene silencing. Worms defective for other RNAi components do not show the same apoptotic phenotypes as DCR-1 mutants. So how does DCR-1 mediate the processing of DNA in apoptosis? Because apoptotic DNA degradation occurs downstream of caspases, the authors investigated whether DCR-1 is a direct substrate of the caspase cell death protein 3 (CED-3; a caspase 8 homologue) and found that it is. Interestingly, cleavage of DCR-1 by CED-3 occurs in one of the two RNaseIII domains in DCR-1 (which are required for binding and cleaving dsRNA), destroying the ability of DCR-1 to process dsRNA. The CED-3-mediated cleavage of DCR-1 generates a carboxyterminal product that retains the other RNaseIII domain. Unlike full length DCR-1, truncated DCR-1 can generate 3′ hydroxyl DNA breaks, suggesting that DCR-1 undergoes a protease-mediated conversion from an RNase to a DNase. Finally, as truncated DCR-1 can rescue the cell death defect of DCR-1 mutants, but expression of a dcr-1 transgene lacking a CED-3 cleavage site cannot, truncated DCR-1 is sufficient to mediate the pro-apoptotic function of DCR-1. DCR-1-deletion mutants also exhibit aberrant vulva anatomy owing to defective microRNA production, which is rescued by expression of DCR-1 that cannot undergo CED-3-mediated cleavage, but not by expression of truncated DCR-1. DCR-1 therefore has independent functions in RNAi and apoptotic DNA fragmentation. Thus, unexpectedly, in addition to processing small RNAs, DCR-1 can fragment DNA by switching from being an RNase to a DNase. It will be interesting to see whether this conversion occurs in the regulation of other nucleases and in mammals. Debbie Walker