In plants, small interfering RNAs (siRNAs) are a quintessential class of RNA interference (RNAi)-inducing molecules produced by the endonucleolytic cleavage of double stranded RNAs (dsRNAs). In order to ensure robust RNAi, the siRNAs are amplified through a positive feedback mechanism called transitivity. Transitivity relies on RNA-DIRECTED-RNA POLYMERASE 6 (RDR6)-mediated dsRNA synthesis using siRNA-targeted RNA. This secondary dsRNA is subsequently cleaved into secondary, mainly phased, siRNAs (phasiRNAs) by DICER-LIKE (DCL) endonucleases. As primary siRNAs, secondary siRNAs are also loaded into ARGONAUTE proteins (AGOs) to form an RNA-induced silencing complex (RISC) reinforcing cleavage of the target RNA. Although the molecular players underlying transitivity are well established, the mode of action of transitivity remains elusive. In this study, we investigated the influence of primary target sites on transgene silencing and transitivity using the GFP-expressing Nicotiana benthamiana 16C line, high pressure spraying protocol (HPSP), and synthetic 22-nucleotide (nt) long siRNAs. We found that the siRNA targeting the 3 prime end of the GFP transgene was less efficient in inducing silencing when compared to the siRNAs targeting the 5 prime end and middle region of the GFP. Moreover, sRNA sequencing of locally silenced leaves showed that the amount but not the profile of secondary RNAs are shaped by the occupancy of the primary siRNA triggers on the target RNA. Our findings suggest that RDR6-mediated dsRNA synthesis is not primed by primary siRNAs and that dsRNA synthesis appears to be generally initiated at the 3 prime end of the target RNA.

Combinations of epigenetic modifications H3K4me3 and H3K27me3 implicate bistable feature which alternates between on and off state allowing rapid transcriptional changes upon external stimuli. Target of Rapamycin (TOR) functions as a central sensory hub to link a wide range of external stimuli to gene expression. However, the mechanisms underlying stimulus-specific transcriptional reprogramming by TOR remains elusive. Our in silico analysis in Arabidopsis demonstrates that TOR-repressed genes are associated with either bistable or silent chromatin domains. Both domains regulated by TOR signaling pathway are associated with high level of H3K27me3 deposited by CURLY LEAF (CLF) in specific context with LIKE HETEROCHROMATIN PROTEIN1 (LHP1). Chromatin remodeler SWI2/SNF2 ATPase BRAHMA (BRM) activates TOR-repressed genes only at bistable chromatin domains to rapidly induce biotic stress responses. Here we demonstrated both in silico and in vivo that TOR represses transcriptional stress responses through global maintenance of H3K27me3.