Doench JG, Petersen CP, Sharp PA. (2003) siRNAs can function as miRNAs. Genes Dev 17: 438-42. [Abstract]
MicroRNAs (miRNAs) are a group of small RNAs found in a wide variety of organisms from plants to Drosophila to humans. Single-stranded and approximately 21-25 nt long, most of these non-protein coding RNAs appear to regulate gene expression through a mechanism distinct from siRNA-induced mRNA degradation. lin-4, the first identified miRNA, and let-7, a related miRNA, bind to complementary sequences in the 3' untranslated region (UTR) of target mRNAs, inducing translational repression through an unknown mechanism (1-3). This binding does not induce mRNA degradation, presumably due to the lack of perfect complementarity between the miRNA and binding site. The introduction of perfect complementarity into a let-7 target mRNA induced mRNA degradation, indicating that miRNAs can act as siRNAs and that the two pathways may "overlap". Such overlap suggests that siRNAs may also be able to function as miRNAs. The results of several experiments testing whether siRNAs can function as miRNAs were recently published in Genes & Development by Doench, Petersen, and Sharp. Their results are summarized here.
In their first experiment, Doench and colleagues selected an siRNA known to be effective at silencing CXCR4 in vivo. A CXCR4 binding site was designed and placed in the 3' UTR of a luciferase reporter gene. In one construct, a single CXCR4 binding site was introduced that had perfect siRNA complementarity. In the second construct, 4 CXCR4 binding sites were introduced that contained complementary sequences at their 5' and 3' ends, but not in the middle, producing a "bulge" of non-hybridizing nucleotides. Transfection of the CXCR4 siRNA into HeLa cells and subsequent luciferase assays revealed a greater than 10-fold repression of luciferase expression with both constructs. RT-PCR and Northern analyses demonstrated that suppression had occurred via different mechanisms. With the perfectly complementary construct, luciferase transcript levels were reduced more than 10-fold, in keeping with RNAi-mediated mRNA degradation. With the bulged construct, however, luciferase transcript levels decreased only 1.2-fold, indicating that with "miRNA-like" binding to the target mRNA, luciferase levels were reduced due to translational repression and not reduction of mRNA stability.
Having shown that an siRNA could elicit miRNA-like effects on mRNA translation, the authors were then able to use this system to conduct preliminary tests of miRNA sequence/structure requirements. They found that altering the nucleotide sequence of the "bulge" (ie ACC vs AGG) did not have a significant effect on repression. However, repression was found to increase with increasing number of binding sites placed in the 3' UTR, as well as with increasing siRNA concentration. This was also true of the perfectly complementary binding sites (siRNA mechanism) the only difference being that the perfectly complementary binding sites appeared to function independently of one another, whereas the addition of bulged sites had a greater than additive effect on translational repression.
Endogenous siRNAs have not been identified in mammalian cells, and so, while exogenously derived siRNAs are a powerful tool for modulation of gene expression, a biological function for these small RNAs and associated RNAi machinery in mammalian cells remains to be discovered. miRNAs, on the other hand, are present in a wide variety of organisms and cell-types and, theoretically, could regulate any biological process involving RNA:RNA or RNA:protein interaction. The ubiquitous nature of miRNAs suggests that the day-to-day roles of Dicer and other participants in RNA interference may actually be the processing of miRNAs. The work by Doench and colleagues will undoubtedly precipitate more extensive studies of the mechanism of miRNA-mediated translational repression. But primarily this work suggests that some siRNAs may act as miRNAs to repress translation. Therefore, distinguishing between the two is becoming more difficult and as mentioned, designing gene specific siRNAs may be trickier than previously anticipated. This work should serve to guide researchers toward better siRNA design, since less than perfect complementarity to regions of genes that are not being targeted could theoretically lead to miRNA-like translational repression of genes unrelated to the target.
