Small RNAs such as siRNAs and miRNAs are 20-24 nt non-coding regulatory RNAs that associate with Argonaute (AGO) proteins to form the core of the RNA-induced silencing complex (RISC). Although small RNAs are born double-stranded, only one of the two strands, the guide, is loaded into AGO through a maturation process assisted by the chaperone machinery. Once loaded, mature RISC is guided to complementary mRNAs via base pairing to thereby exert sequence-speci!c repression. Many AGO proteins possess RNaseH endonuclease activity that allows the cleavage (slicing) of mRNAs with nearly perfect complementarity to the small RNA guide. In some cases, secondary siRNAs are produced from the cleavage fragments in a way of ampli!cation loop to further silence the same transcripts, or to produce trans-acting siRNAs (tasiRNAs) that target other mRNAs. While slicing is the prototypical way of regulation for siRNAs and plant miRNAs, endogenous targets in animals rarely present extended base pairing to their cognate miRNAs. In this case silencing occurs via translational repression, often coupled to mRNA decay. This way of regulation is mediated by additional RISC components that associate with AGO via GW (Gly-Trp) motifs that speci!cally bind to hydrophobic cavities exposed on the AGO surface. Plants can also repress miRNA targets by translational inhibition, but the mechanism in this case is totally unknown. Although plant AGOs have hydrophobic pockets homologous to the ones of their animal and yeast counterparts, a decade of thorough quest for GW proteins that might mediate translational repression in plants has been futile. Indeed, very little is still known about how plants combine the two possible ways of miRNA-mediated regulation, how slicing is avoided to allow translational repression to take place, and which are the different outcomes of both pathways from a functional perspective.
In this work, we use the model plant Arabidopsis thaliana to study some of the matters proposed above. Our analyses of AGO1 slicer-de!cient plants suggest that translational repression of miRNA targets is not exerted in planta when slicing is impaired, and mRNA decay is also not detected in these conditions. Additionally, the maturation of siRNA RISC, which is known to be impaired in slicer-de!cient plants, stalls in a loading intermediate that shows trimming of small RNAs by exonucleases while the full duplex remains in association with AGO1. We also describe that slicer-de!cient but not ago1 null mutant plants produce abundant, albeit unphased, tasiRNAs. The discovery suggests that AGO1 is required for the biogenesis of secondary siRNAs, but cleavage of the precursor is only necessary to set the right phase. This discovery changes the actual conception on how production of secondary siRNAs is triggered, not by aberrant features of cleavage fragments as previously thought, but via recruitment of the biogenesis machinery by RISC. Additionally, we generated AGO1 mutants with modi!cations in the hydrophobic pockets designed to either abolish or increase the affinity to GW motifs of binging partners. Surprisingly, plants with a mutation in a highly conserved lysine that conforms the pocket with more affinity for Trp displayed wild type phenotypes. miRNA target regulation occurs normally in these plants, suggesting that the pocket is dispensable for translational repression. We also aimed to increase the affinity of AGO1 for GW proteins by removal of a plant-speci!c endogenous Trp that is located on top of the second pocket. Proteomic analyses of affinity-puri!ed mutant AGO1 provided a list of potential AGO1 partners, with some interesting candidates that have been previously involved in translational regulation. Among them, ECT2 is a YTH domain protein whose homologues in animals and yeast speci!cally bind N6-metyhladenosine (m6A) in mRNA, a post-transcriptional modi!cation essential for cell differentiation and development in many organisms. We validated the association between ECT2 and AGO1 through binding assays, and showed that the GW motifs present in ECT2 might be dispensable for the interaction. Furthermore, we prove that ECT2 is a bona fide m6A-binding protein, and present evidence for the functional redundancy with other ECT family members in Arabidopsis. Finally, we found that the AGO1 tryptophan mutant, which is capable to load small RNAs and cleave RNA targets, is otherwise blocked with the chaperone machinery. This blockage is probably the reason for the developmental defects exhibited by the mutant plants. Based on this result, we propose a novel role for the chaperone machinery in RISC assembly, likely in the rearrangement of the hydrophobic pockets to an active GW-binding conformation.
The understanding of the processes described above has fundamental relevance. In plants as in animals, small RNAs are master regulators of development and stress responses, protect the genome from the deleterious effect of transposons, and provide immunity against viruses. Development of individuals with dysfunctional small RNA machineries is severely impaired, and defects in speci!c miRNAs can cause disease, more speci!cally cancer. This work not only provides relevant data to understand how gene silencing occurs in plants, but also might provide essential information for the same pathway in animals, which is far for being perfectly understood. Indeed, the discoveries made in yeast, plant and bacteria have been fundamental to build the current knowledge on how gene silencing occurs in all organisms.