Dicer-like (DCL) enzymes play a pivotal role in RNA silencing in plants, processing the long double-stranded RNA (dsRNA) that triggers silencing into the primary short interfering RNAs (siRNAs) that mediate it. The siRNA population can be augmented and silencing amplified via transitivity, an RNA-dependent RNA polymerase (RDR)-dependent pathway that uses the target RNA as substrate to generate secondary siRNAs. Here we report that Arabidopsis DCL2–but not DCL4-is required for transitivity in cell-autonomous, post-transcriptional silencing of transgenes. An insertion mutation in DCL2 blocked sense transgene-induced silencing and eliminated accumulation of the associated RDR-dependent siRNAs. In hairpin transgene-induced silencing, the dcl2 mutation likewise eliminated accumulation of secondary siRNAs and blocked transitive silencing, but did not block silencing mediated by primary siRNAs. Strikingly, in all cases, the dcl2 mutation eliminated accumulation of all secondary siRNAs, including those generated by other DCL enzymes. In contrast, mutations in DCL4 promoted a dramatic shift to transitive silencing in the case of the hairpin transgene and enhanced silencing induced by the sense transgene. Suppression of hairpin and sense transgene silencing by the P1/HC-Pro and P38 viral suppressors was associated with elimination of secondary siRNA accumulation, but the suppressors did not block processing of the stem of the hairpin transcript into primary siRNAs. Thus, these viral suppressors resemble the dcl2 mutation in their effects on siRNA biogenesis. We conclude that DCL2 plays an essential, as opposed to redundant, role in transitive silencing of transgenes and may play a more important role in silencing of viruses than currently thought.

One recently identified mechanism that regulates mRNA abundance is RNA silencing, and pioneering work in Arabidopsis thaliana and other genetic model organisms helped define this process. RNA silencing pathways are triggered by either self-complementary fold-back structures or the production of double-stranded RNA (dsRNA) that gives rise to small RNAs (smRNAs) known as microRNAs (miRNAs) or small-interfering RNAs (siRNAs). These smRNAs direct sequence-specific regulation of various gene transcripts, repetitive sequences, viruses, and mobile elements via RNA cleavage, translational inhibition, or transcriptional silencing through DNA methylation and heterochromatin formation. Early genetic screens in Arabidopsis were instrumental in uncovering numerous proteins required for these important regulatory pathways. Among the factors identified by these studies were RNA-dependent RNA polymerases (RDRs), which are proteins that synthesize siRNA-producing dsRNA molecules using a single-stranded RNA (ssRNA) molecule as a template. Recently, a growing body of evidence has implicated RDR-dependent RNA silencing in many different aspects of plant biology ranging from reproductive development to pathogen resistance. Here, we focus on the specific functions of the six Arabidopsis RDRs in RNA silencing, their ssRNA substrates and resulting RDR-dependent smRNAs, and the numerous biological functions of these proteins in plant development and stress responses.

No abstract

RNA silencing is a highly conserved pathway in the network of interconnected defense responses that are activated during viral infection. As a counter-defense, many plant viruses encode proteins that block silencing, often also interfering with endogenous small RNA pathways. However, the mechanism of action of viral suppressors is not well understood and the role of host factors in the process is just beginning to emerge. Here we report that the ethylene-inducible transcription factor RAV2 is required for suppression of RNA silencing by two unrelated plant viral proteins, potyvirus HC-Pro and carmovirus P38. Using a hairpin transgene silencing system, we find that both viral suppressors require RAV2 to block the activity of primary siRNAs, whereas suppression of transitive silencing is RAV2-independent. RAV2 is also required for many HC-Pro-mediated morphological anomalies in transgenic plants, but not for the associated defects in the microRNA pathway. Whole genome tiling microarray experiments demonstrate that expression of genes known to be required for silencing is unchanged in HC-Pro plants, whereas a striking number of genes involved in other biotic and abiotic stress responses are induced, many in a RAV2-dependent manner. Among the genes that require RAV2 for induction by HC-Pro are FRY1 and CML38, genes implicated as endogenous suppressors of silencing. These findings raise the intriguing possibility that HC-Pro-suppression of silencing is not caused by decreased expression of genes that are required for silencing, but instead, by induction of stress and defense responses, some components of which interfere with antiviral silencing. Furthermore, the observation that two unrelated viral suppressors require the activity of the same factor to block silencing suggests that RAV2 represents a control point that can be readily subverted by viruses to block antiviral silencing.

Many biotechnological applications require high-level expression of transgenes in plants. One strategy to achieve this goal was the production of potato virus X (PVX) "amplicon" lines: transgenic lines that encode a replicating RNA virus vector carrying a gene of interest. The idea was that transcription of the amplicon transgene would initiate viral RNA replication and gene expression, resulting in very high levels of the gene product of interest. This approach failed, however, because every amplicon transgene, in both tobacco and Arabidopsis thaliana, was subject to post-transcriptional gene silencing (PTGS). In PTGS, the transgene is transcribed but the transcripts fail to accumulate as a result of sequence-specific targeting and destruction. Even though the amplicon locus is silenced, the level of beta-glucuronidase (GUS) activity in a PVX/GUS line is similar to that in some transgenic lines expressing GUS from a conventional (not silenced) GUS locus. This result suggested that the very high levels of expression originally envisioned for amplicons could be achieved if PTGS could be overcome and if the resulting plants did not suffer from severe viral disease. Here we report that high-level transgene expression can be achieved by pairing the amplicon approach with the use of a viral suppressor of PTGS, tobacco etch virus (TEV) helper component proteinase (HC-Pro). Leaves of mature tobacco plants co-expressing HC-Pro and a PVX/GUS amplicon accumulate GUS to approximately 3% of total protein. Moreover, high-level expression occurs without viral symptoms and, when HC-Pro is expressed from a mutant transgene, without detrimental developmental phenotypes.