In the majority of the eucaryote, the presence of double-strands RNA induce the inhibition of gene expression base on the complementary of sequence. The best known example is the case of RNA interference in C. elegans which is the first model described, in which the double-strands RNA generate an specific endonuclease who degrade all RNA complementary perfectly to the small RNA guide included in the complex RISC. In addition to this post-transcriptional activity, it has been observed in many eukaryotes the existence of mechanisms related to RNA interference and it inhibit transcription by acting at the chromatin. If these mechanisms have been clearly demonstrated in plants, fungi, there are only several examples of this type of regulation in mammals. Unexpectedly, the targeting the promoter of a gene with small double-stranded RNA can lead to increased expression. This paradoxical response has not been observed so far in mammalian cells, but it raises interest particularly to stimulate the expression of tumor suppressor genes, unfortunely the mechanism is still unknown.My work has focused on studying the induction of expression by small RNAs. They are based first on the development of an experimental approach that allows to monitor the promoter activity of the targeted gene. To do this I used indicator constructions organized around a bidirectional promoter that controls the expression of two fluorescent proteins. When targeting the messenger of one of these proteins, the expression of the other is increased and I was able to show that thisincrease correlates with the amount of RNA messenger polymerase II presented on the bidirectional promoter. Thus, the use of a bidirectional promoter can effectively monitor the level of transcription of the gene targeted by the small RNA. This induction of expression detected in a "contralateral" is not due to an off-target effect of siRNA because it requires the presence of the target sequence on one of the transcripts of the construction indicator. The induction can be observed with many different small RNAs, including the interact as micro RNA. Thus the construction indicator that I developed are biased in an induction response transcriptionally in response to a silencing. The use of a bidirectional promoter is probably the origin of this bias through the possibility of inducing a convergent transcription when the plasmids are circular. In fact, the linearization of the construction indicator removes the induction, at least for the simplest constructions. If the heart of the complex RISC is the protein Ago2, is necessary for the silencing and the induction, I was able to show that in the second case Ago2 was in fact to guide the RISC complex on the transcripts but not to cut it. Indeed, the silencing of proteins TNRC6A and B reduces induction significantly without affecting the silencing if it processe in the siRNA model. Also anchoring the transcript EGFP induces a response similar to the small RNA (silencing and induction). This anchor approach allowed me to identify domaines necessary for silencing and induction and show that they are distinct. This work makes it possible to demonstrate that the transcriptional induction observed in our constructions indicator is due to a activity partner ofArgonaute proteins, the GW182/TNRC6 family. This observation open the way for characterization of the mechanism of this induction by showing that it belongs to a specific activity of the RISC complex.
