As most worm geneticists will tell you, RNA interference (RNAi) is a rapid, easy and specific way in which to inactivate gene function. However, RNAi — which is mediated by double-stranded RNAs that trigger sequence-specific mRNA degradation — is cytotoxic to most mammalian cells, a problem that was overcome last year by using synthetic, short interfering RNAs (siRNAs, see NRG Highlights July 2001). But siRNAs have a shortcoming themselves as a tool for investigating gene function — their effects are transient. Now Thijn Brummelkamp and colleagues have created a new vector, pSUPER, that generates siRNAs in mammalian cells and brings about sustained gene inactivation without cytotoxicity, allowing in vitro loss-of-function phenotypes to be assayed over longer periods of time.

pSUPER has several key features that equip it for the job. It contains an RNA-polymerase-III promoter, a well-defined transcription start site and a termination signal that consists of five Ts. It encodes a small RNA transcript that is cleaved after the second U of the termination signal to generate a transcript that has two overhanging 3′ U nucleotides, as found in synthetic siRNAs. Finally, the gene-specific insert comprises two complementary, target-derived 19-nucleotide (nt) sequences that are separated by a short spacer — an arrangement that is predicted to generate a 19-nt stem–loop structure similar to that of let-7 , the Caenorhabditis elegans gene that controls the timing of developmental events by an RNA-mediated mechanism.

Brummelkamp et al. have used pSUPER to knock down more than ten genes in several mammalian cell lines. Their pSUPER-CDH1 vector worked as well as a synthetic CDH1 siRNA and most effectively when the 19-nt stem structure contained 9 nts of loop sequence. A single-nucleotide mismatch introduced into this sequence resulted in siRNAs that failed to suppress CDH1 expression, illustrating the specificity of siRNA activity. A pSUPER-p53 vector suppressed the endogenous TP53 transcript by ∼90%, and transfected cells failed to undergo p53-mediated G1 arrest following their exposure to ionizing radiation. Moreover, TP53 siRNA was still present in cells two months after transfection, and the cells showed more than a 90% reduction in wild-type p53 levels and no obvious signs of cytotoxicity.

So how can this vector be put to good use? The authors' finding that a single-nucleotide mismatch in the targeting sequence can abrogate siRNA activity points to its potential application in gene therapy. SiRNA-generating vectors could be designed to inactivate disease-associated transcripts that contain point mutations to leave unaffected the expression of the remaining wild-type transcript. These vectors could also be used in high-throughput in vitro screens for loss-of-function phenotypes. How they will be used in vivo, however, remains to be seen.