The dimerization of RIG-I CTD reported previously may simply reflect the 59 triphosphorylated bivalency of the dsRNA ligand used. Surprisingly, RIG-I dimerization in the presence of the 62-mer 59ppp-dsRNA could not be observed in cellula. This could be explained by an unbalanced molar ratio of RIG-I protein to 59ppp-dsRNA in the intracellular milieu, a competition with other 59ppp-RNA binding proteins and/or the highly dynamic interaction of RIG-I with 59ppp-dsRNA despite a Kd in the 160 pM range. The incubation of very stable 59ppp- panhandle RNA with dsRNA of variable length with cellular extracts from RIG-I transfected cells allows the observation of RIG-I oligomerization, at least if the dsRNA exceeds 46 bp in length. According to the proposed model, one molecule of RIG-I would bind the RNA 59ppp end and enter the RNA using ATP hydrolysis. Several RIGI molecules would enter an RNA this way and form a RNA mediated oligomer. Contrary to the cooperative association of MDA5 along RNA, RIG-I molecules do not self-oligomerize to form a long filament but multiple proteins can bind to the same RNA, forming a RNA-poly-RIG-I scaffold that falls apart if the long RNA is cleaved by RNAse treatment. In vivo, RIG-I oligomerization was reported once by pull down assay of Flag- and Myc-tagged RIG-I. However, the lack of clear differences between the data obtained in infected and non-infected cells, questions whether any RNAinduced RIG-I oligomerization had really occurred. In addition, multiple combinations of RIG-I and RIG-I domains and PR-171 molecular weight subdomains such as between RIG-I and CARDs, RIG-I and RIG-I-D-CARDs, CTD and CARDs, CTD and helicase, CTD and were also reported. While one cannot exclude that some of the reported interactions could reflect cis-interactions between RIG-I domains bridged or not by viral RNA, the other interactions would suggest multiple oligomerization sites within RIG-I. However, none of them are supported by available RIG-I crystal structures. In contrast, in our work, we did not observe selfassembly of RIG-I upon recognition of synthetic or viral RNA by co-immunoprecipitation assay or using the more sensitive PCA assay. Furthermore, RIG-I dimerization hardly occurred even after being grafted with the gcn4 dimerization signal. We strongly favour that a monomeric RIG-I-RNA complex is the minimal functional signal transduction unit in full agreement with biochemically defined monomeric RIG-I-RNA complexes that are able to activate the IFN response. Thus, so far there is no convincing evidence that, upon RNA recognition, RIG-I could self-oligomerize, and the model of RIG-I oligomerization for enabling signal transduction is inconsistent with all cell biological, biochemical and structural biological studies that have endeavoured to quantitatively assess the stoichiometry of RIG-I in its activated state. Rather, a single dsRNA can bind several RIG-I molecules and this can occur or not during viral infection . Further down the signalling cascade, tandem CARDs of RIG-I associate with free K63 polyubiquitin in a helical tetramer complex that becomes engaged in a complex interaction with membrane anchored MAVS. This scaffold would associate multiple RNA-RIG-I signal units to several MAVS molecules.