Since its discovery in Caenorhabditis elegans, RNA interference (RNAi) has revolutionized gene-function analysis. The success of RNAi lies in its simplicity — when introduced into cells of certain organisms, double-stranded (ds) RNAs can cause a sequence-specific mRNA degradation that mimics a loss-of-function phenotype. Elbashir and colleagues now show for the first time that RNAi can work consistently and specifically in mammalian cells in response to short dsRNA molecules (siRNAs). Their study also reveals the existence of two pathways — one general and one sequence specific — that are activated in mammalian cells in response to dsRNA.

For those interested in gene function, RNAi-mediated gene knock-down is a useful and rapid alternative to gene knockouts or conventional antisense technology, and it works efficiently in plants and in many invertebrates. Although RNAi has previously been shown to work in mammalian oocytes and embryos, in mammalian cells dsRNAs longer than 30 nucleotides cause a general, sequence non-specific, mRNA degradation and translational shutdown. As very short dsRNAs mediate specific RNAi effects in Drosophila, Elbashir et al. tested whether such siRNAs might also work in mammalian cells. They synthesized 21-mer dsRNA molecules with sequence homology to three luciferase genes and co-transfected plasmids containing these genes with the corresponding siRNAs into four different types of mammalian cell lines. The authors determined siRNA effectiveness by measuring luciferase luminescence and found that, although the extent of suppression varies — possibly depending on the level of target gene expression — suppression is completely sequence specific. Furthermore, siRNA is effective at significantly lower concentrations than antisense or ribozymes in gene-targeting experiments. Similar results were also obtained for suppressing endogenous genes.

Elbashir et al. investigated the effects of longer, supposedly non-specific, dsRNAs and found that they also act in a sequence-specific manner, but that this is normally masked by the non-specific effects. Their observations indicate that there are two pathways in mammalian cells that can be activated in response to dsRNAs.

RNAi provides unequalled flexibility for rapidly analysing gene function and can be used to determine gene function on a genome-wide scale, as recently shown in C. elegans. The findings of Elbashir et al. have important consequences for those investigating gene function in mammalian cells. But before these applications become practical in mammalian systems, it will be essential to establish easy and efficient ways of delivering siRNA into mammalian cells and rapid phenotypic screens. Questions regarding the biological role of endogenous dsRNAs in gene regulation also remain to be answered.