Community Research and Development Information Service - CORDIS

Final Report Summary - NUAGO1 (Exploring Nuclear Localization and Functions of Arabidopsis AGO1)

Plants have evolved a large variety of pathways over the DICER–AGO consortium, which most likely underpins important aspects of their phenotypic plasticity. AGO proteins are the main RNA silencing effectors across kingdoms. The Arabidopsis genome encodes 10 AGO genes, defining three major phylogenetic clades: (i) AGO1, -5, and -10; (ii) AGO2, -3, and -7; and (iii) AGO4, -6, -8, and -9. Canonical eukaryotic AGOs contain four main domains: a variable N-terminal domain and the more highly conserved PAZ, MID, and PIWI domains. AGOs fold into a bilobal structure displaying a central groove for sRNA binding. Whereas AGO4, -6, and -9 associate mostly with 24-nt siRNA mediating transcriptional gene silencing (TGS), AGO1, -2, -5, -7 and -10 bind 21–22-nt molecules acting in posttranscriptional gene silencing (PTGS). Arabidopsis genome also encodes four DCL genes, which all harbor a helicase ATP-binding domain, a helicase C-terminal domain, a PAZ domain, two RNaseIII domains and one or two C-terminal dsRNA binding domains (dsRBD). DCL1 binds to imperfect stem-loops contained in primary miRNA transcripts to release ~21-nt miRNA/miRNA* duplexes, while DCL2, -3 and -4 process long, near-perfect dsRNA into populations of 22-, 24- and 21-nt siRNAs, respectively. DCL3 is the main DCL required for RdDM. DCL4 and its surrogate, DCL2, produce siRNAs derived from viruses and from long inverted-repeats (IRs). DCL4 is also required for the biogenesis of so called trans-acting siRNA (tasiRNAs) involved in plant development as well as a small number of evolutionary young miRNA, including miR822.
During the current project we focus in two main goals that involve the both key proteins from sRNA silencing pathway: AGOs and DCLs proteins. Concretely: (i) analyze AGOs subcellular localization, in particular focusing in the putative nuclear localization and function of AGO1, and (ii) identify for the first time in a direct way, specific dsRNA substrates of DCLs proteins.
(i) Regarding the first objective: Initially, we expressed all full-length AGOs sequences (with exception of AGO8) fused to fluorescence reporter (mCherry or GFP) under their respective native regulatory sequences in Arabidopsis WT plants to analyze its respective subcellular localization. Results illustrated a wide diverse subcellular localization of the different AGO in Arabidopsis. This fact prompted us to compare and analyze in detail AGOs sequences. Surprisingly, beside the cytoplasmic localization of AGO1, the amino acid analysis exposed the presence of a weak non-canonical nuclear localization signal (NLS) and highly conserved canonical nuclear exportation signal (NES). Both signal showed functional performance validated by transient experiment in Nicotiana. To validate NES function in Arabidopsis, pAGO1:GFP-AGO1 was transformed into heterozygous ago1-3+/- mutants plants. Transgenic plants were selected and genotyped to get transgenic homozygous ago1-3-/- pAGO1:GFP-AGO1 plants. Complementation was confirmed by phenotype and sRNA levels analysis. Following experiment showed for the first time that AGO1 shuttles between a steady cytoplasmic and shortly transitory nuclear subcellular localization.
Thorough investigation of AGO1 nuclear function(s) entails, we implemented a commonly used nucleo-cytoplasmic fractionation technique based on Percoll gradients. Even though after fractionation the main control for proteins (H3 and UGPase) and RNA (U6 and tRNA) showed and expected behavior we could detect two main problems (vias of contamination): (a) we found a important presence of nuclear soluble proteins in cytoplasm fraction and (b) significant contamination of different debris in the nuclear fraction. At the same time, we set up a new protocol to purify nuclei using flow cytometry sorting. This approach allowed us to isolate extremely pure nuclei reduced highly the debris contamination in the nuclear fraction. Then, we performed nuclear fractionation from WT and deficient in RNA silencing proteins plants. After nuclear sorting we preformed total and nuclear RNA extraction and we conducted comparative quantitative RT-PCR analyses to analyze miRNAs, miRNAs precursors and miRNAs target in both fractions. Analysis of miRNAs precursors showed an expected accumulation in the nuclear fractions of hyl1, hasty and control plants. Analysis of mature miRNAs accumulations showed that miRNAs tested are mostly accumulated in cytoplasmic fraction. However, we couldn’t see an accumulation of any miRNAs in nuclear fraction of hasty mutants. Therefore should be reconsidered the proposed function of HASTY on miRNA exportation. Finally, by qPCR we measured miRNA targets levels in hyl1, hasty and control plants. We detected a significant (and expected) enrichment of miRNA target levels in hyl1 in comparison with control plants in cytoplasmic fraction. However, in nuclear fractions, we couldn’t detect any significant differences between hyl1, hasty and control plants. This result suggests that, at least the miRNAs target tested would be regulated by a “cytoplasmic PTGS”.
(ii) Although DCLs play a central role in plant sRNA biogenesis, their dsRNA substrates have never been directly isolated. Instead, these substrates are usually indirectly inferred by mapping sRNA sequencing data to the genomic loci that presumably produce the dsRNA precursors. Recently, Dicer immuno-precipitation coupled to cross-linking was used to identify direct dsRNA substrates in Human and C. elegans. In order to improve this approach we have produced catalytically inactive versions of DCL1, 2, 3 and 4 using point mutations introduced in the RNase III domains of the proteins to obtain an objective and unbiased view of the potential DCL substrates at a genome-wide scale.
To obtain a global understanding of DCL functions in small RNA biology, full-length genomic WT sequences of DCL1, DCL2, DCL3 and DCL4 and the catalytically inactive versions were cloned under their endogenous promoters and were fused to either FLAG or hemagglutinin (HA) epitopes on the 3’ end. These construct were used to generate stable dicing-proficient complemented dcl mutant plants (those transformed with DCLs WT versions) and dicing-deficient dcls transformed plants (those transformed with DCLs WT versions). RNA-immunoprecipitation (RIP) procedures for DCL1 have been set up efficiently in the lab. High-throughput RNA sequencing analysis of DCL1-MUT/dcl1-7 RIP allowed the discovery of novel groups of DCL1 substrates in addition to the near-entirety of all know miRNAs and IR precursors described in the literature. These groups include protein-coding gene mRNAs, tRNA, transposable elements (TE) RNA and RNA from intergenic regions. Computational analysis showed that 75% of theses regions were associated with sRNA showing imperfect stem-loop secondary structures of miRNA-like precursors. These novel groups of DCL1 substrates as well as the new putative miRNAs, their mRNA targets and possible biological implications are currently being characterized in details. The remaining 25% of structured RNA bound to DCL1 that did not produce sRNAs was dominated by fold-back structures found within mRNA of protein-coding genes with biases in 5’- and 3’-UTRs. Currently investigating if, and to what extent, these regions undergo single cuts (i.e. not generating sRNA) by DCL1 as a novel means of post-transcriptional regulation of gene expression not involving the downstream steps of RNA silencing.

Related information


Tel.: +41446339360
Follow us on: RSS Facebook Twitter YouTube Managed by the EU Publications Office Top