In vitro dicing: RNAi technology for high throughput loss-of-function screening

RNA interference (RNAi) was first discovered in C. elegans and in a very short time has been elucidated as a well-conserved, enzyme-mediated cellular process in which small non-coding RNAs regulate gene expression. One such RNA, the small interfering RNA (siRNA), can silence gene expression by triggering the destruction of an mRNA. Pools of siRNAs are produced when Dicer, an RNase III family enzyme, cleaves or "dices" long double stranded RNA (dsRNA) into small double-stranded fragments, each ∼21 nucleotides in length. These siRNA duplexes are contained within a multicomponent protein complex, which unwinds the duplex, mediates binding of the antisense strand to the complementary target mRNA (the sense strand) and contains nuclease activities that degrade the target mRNA. The specificity and simplicity of siRNA-mediated gene silencing encouraged scientists to exploit the cellular RNAi machinery especially in vertebrate model systems where traditionally loss-of-function studies have been tedious and difficult. However, not just any siRNA will trigger specific gene silencing rather many different siRNAs must be designed and tested in order to find an efficacious one. Thus, a simple method to produce a large number of different efficacious siRNAs is needed especially if siRNAs are to be utilized for large-scale reverse genetic screening. Nature has already devised a method; in invertebrates large dsRNAs are diced into pools of siRNAs and these pools are very efficacious probably because the mRNA is targeted by multiple siRNAs. Pools of siRNAs could ultimately be produced in vertebrate cells by simply introducing a large dsRNA, but in most somatic cells dsRNAs larger than 30 base pairs are toxic. We therefore decided to make pools of siRNAs in vitro by treating a large dsRNA with a recombinant human Dicer. This approach, "in vitro dicing", is indeed very efficacious. With equivalent potency nearly every pool of diced siRNAs (d-siRNAs) elicits specific gene silencing in mammalian cells and d-siRNA-mediated gene suppression results in phenotypes that accurately predict gene function. Furthermore in vitro dicing is amenable to high throughput format. Taken together these findings suggest that in vitro dicing is ideal for large-scale loss-of-function reverse genetic analysis in vertebrate somatic cells.