We made substantial progress in all the three main lines of research that we set out to pursue.
(1) Installment of positive feedback loops to produce amplified siRNA populations
1.1 A new interaction site in AGO1 with relevance for siRNA amplification
We screened mutations in the AGO1 protein to find regions specifically required for siRNA amplification. All mutants were tested for endogenous miRNA activity and for siRNA amplification defects. These efforts identified an uncharacterized protein-protein interaction site in AGO1 with a clear requirement for some, but not all, types of siRNA amplification, and no requirement for miRNA activity. A combination of approaches led to identification of the protein whose binding to the interaction site is necessary for siRNA amplification.
1.2 Forward genetics and a second interaction site in AGO1 relevant for siRNA amplification
We carried out a full genomic screen instead aimed at identifying genes required specifically for AGO1-mediated siRNA amplification. We recovered a point mutant that identifies a new site in AGO1 itself implicated in siRNA amplification. Follow-up work focused on how it is related to the other site described above, in particular whether there is redundancy between them. This turned out to be difficult, because the most obvious double mutants had unexpected defects related to chaperone binding, forcing us to go through multiple rounds of mutant design. This very promising work could not be fully completed during the PATHORISC project, but we remain hopeful that our results can form the basis for further investigation.
1.3 siRNA amplification and the discovery of an intracellular double-stranded RNA sensing pathway
During our efforts to define molecularly why targeting RISC to endogenous mRNAs, contrary to viral ones, generally does not lead to siRNA amplification, we ran into an unexpected result: We found that simultaneous inactivation of two systems that preclude endogenous miRNAs from triggering siRNA amplification also leads to very strong autoimmune phenotypes. We investigated this finding further and found clear evidence that the double-stranded RNA (dsRNA) generated by the siRNA amplification machinery activates an intracellular dsRNA sensing pathway that turns on innate immune responses via receptors that we identified. We also showed that these intracellular dsRNA-dependent immune responses depend on the ribonuclease DICER-LIKE2 (DCL2), and that they are key for basal antiviral responses mediated by DCL2. These are extremely important results, because they change our understanding of how basal antiviral resistance in plants works and introduce intracellular dsRNA sensing in plants as new field of research.
(2) RISC re-programming with a focus on degradation of free ARGONAUTE
2.1 Insights into specific recognition and degradation of free ARGONAUTE
We previously identified a defined a 14-amino acid segment in AGO1 (the N-coil) as particularly important for a interactions with series of interactors implicated in regulated proteolysis. In this project, we used biochemical and biophysical approaches to show (a) that the interactions with the 14-amino acid segment are direct and involve specific amino acid side chains required for interaction, (b) crucially, that the 14-amino acid segment is a structural switch that is only accessible for interaction in the RNA-free state of ARGONAUTE, (c) that mutations in the 14-aa segment significantly stabilize the RNA-free form of ARGONAUTE in intact plants.
We also set up kinetic assays combined with mathematical modeling that allowed us to estimate to that at least 50% of all synthesized ARGONAUTE protein is degraded prior to RNA association. This result immediately suggests high biological importance. It also has implications for RISC reprogramming, because it means that the system actually displays the dynamics that would be required for massive amounts of newly synthesized small RNA to enter into mature RISC complexes. This work is described in two papers: one is published in Biochemical Journal, getting the journal cover (see image), the other as a pre-print on BiorXiv close to journal publication.
(3) N6-methyladenosine in mRNA and use of RISC target sites
Our initial hypothesis that N6-methyladenosine in mRNA of miRNA targets could dictate facultative use of miRNA binding sites could not be confirmed by our work. Nonetheless, the results obtained by our work allowed us to show that a small clade of N6-methyladenosine-binding proteins have important roles in plant organogenesis. We defined a high-confidence set of mRNAs bound by N6-methyladenosine-binding proteins, and have done progress on defining the subset that is important for growth stimulation and on what the binding of N6-methyladenosine-binding proteins to these mRNAs actually does. Interestingly, mRNAs encoding AGO proteins are prominent targets of N6-methyladenosine-binding proteins, leading us to rethink what the relationship between N6-methyladenosine and AGO proteins actually is. While we did not have time within PATHORISC to investigate these revised hypotheses, they form the basis for efforts to secure further funding.
