The p122 subunit of Tobacco Mosaic Virus replicase is a potent silencing suppressor and compromises both small interfering RNA- and microRNA-mediated pathways

Abstract

One of the functions of RNA silencing in plants is to defend against molecular parasites, such as viruses, retrotransposons, and transgenes. Plant viruses are inducers, as well as targets, of RNA silencing-based antiviral defense. Replication intermediates or folded viral RNAs activate RNA silencing, generating small interfering RNAs (siRNAs), which are the key players in the antiviral response. Viruses are able to counteract RNA silencing by expressing silencing-suppressor proteins. It has been shown that many of the identified silencing-suppressor proteins bind long double-stranded RNA or siRNAs and thereby prevent assembly of the silencing effector complexes. In this study, we show that the 122-kDa replicase subunit (p122) of crucifer-infecting Tobacco mosaic virus (cr-TMV) is a potent silencing-suppressor protein. We found that the p122 protein preferentially binds to double-stranded 21-nucleotide (nt) siRNA and microRNA (miRNA) intermediates with 2-nt 3' overhangs inhibiting the incorporation of siRNA and miRNA into silencing-related complexes (e.g., RNA-induced silencing complex [RISC]) both in vitro and in planta but cannot interfere with previously programmed RISCs. In addition, our results also suggest that the virus infection and/or sequestration of the siRNA and miRNA molecules by p122 enhances miRNA accumulation despite preventing its methylation. However, the p122 silencing suppressor does not prevent the methylation of certain miRNAs in hst-15 mutants, in which the nuclear export of miRNAs is compromised.

FIG. 1.

Effects of cr-TMV infection and expression of the p122 silencing suppressor. (A and B) Infection of A. thaliana with cr-TMV leads to a strong phenotype after 2 weeks (A) compared to uninfected plants (B). (C and D) Reversion of silenced GFP in cr-TMV-infected transgenic Amp × GFP (C) and Amp (D) plants at 14 days p.i. GFP fluorescence was assessed in the plants under UV light using a dissecting microscope, and the photographs were taken 14 days after inoculation. (E) Accumulation of virus-specific siRNAs during cr-TMV infection. Total RNA was extracted from cr-TMV-infected N. bethamiana plants at 4 days p.i. and A. thaliana plants at 14 days p.i. and separated on denaturing agarose (for viral RNAs) and in 15% denaturing polyacrylamide gel (for siRNAs). (F) Suppression of RNA silencing by p122. N. benthamiana GFP16C leaves were agroinfiltrated with 35S-GFP, 35S-GF-IR, 35S-HcPro, and 35S-p122, as indicated. The GFP fluorescence was monitored under UV light, and the photographs were taken 3 days after agroinfiltration. (G) The p122 silencing suppressor inhibits GFP RNA degradation but does not impair primary siRNA production. Leaves of the N. benthamiana GFP16C line were infiltrated with 35S-GFP (lane 1); 35S-GFP and 35S-GF-IR (lane 2); 35S-GFP and HC-Pro (lane 3); 35S-GFP and p122 (lane 4); 35S-GFP, 35S-GF-IR, and HcPro (lane 5); 35S-GFP, 35S-GF-IR, and p122 (lane 6); and 35S-GFP, 35S-GF-IR, and sigma 3 (lane 7). The RNA samples extracted 60 h after infiltration were subjected to Northern analysis using appropriate probes to detect GFP mRNA and GF-IR RNA and GFP-, GF-, and P-specific siRNAs.

FIG. 2.

p122 protein inhibits siRNA-guided target cleavage. p122 inhibited target cleavage in vitro in the direct-competition assay (lanes 3 to 7) but did not interfere with preprogrammed RISC activity (indirect-competition assay, lanes 8 to 12). In the direct-competition assay, Drosophila embryo extract, target RNA (0.5 nM), labeled siRNA (5 nM), and either empty-vector-infiltrated (lane 1) or p122-infiltrated N. benthamiana plant extracts were used at different dilutions, and all components were added simultaneously. In the indirect-competition assay, siRNAs were incubated with the Drosophila embryo extract for 20 min, and then target RNA and p122-infiltrated N. benthamiana plant extracts were added in different dilutions. The effect of p122 on RISC-mediated cleavage was monitored by the detection of cleavage products. Lane M shows RNA size markers; size is indicated in nucleotides.

FIG. 3.

In vitro RISC formation is inhibited by p122 protein. In the direct-competition (comp.) assay, Drosophila embryo extract (extr.), labeled siRNA, and pBIN empty-vector-infiltrated (lane 11) or p122-infiltrated (lanes 5 to 10) plant extract dilutions were added at the same time. In the indirect-competition reactions, embryo extracts were preincubated with labeled siRNA prior to the addition of the pBIN-infiltrated (lane 18) or p122-infiltrated (lanes 12 to 17) plant extracts. Control reactions were done with labeled siRNA only (lane 1), embryo extract and labeled siRNA (lane 2), and siRNA and pBIN-p122-infiltrated or empty-vector-infiltrated plant extract (lanes 3 and 4, respectively). The different forms of siRNA containing silencing-related complexes were separated in 3.9% native acrylamide gels. The positions of RISC, RISC loading complex (RLC), siRNA-DICER2-R2D2, and p122-siRNAs are indicated.

FIG. 4.

p122 protein does not inhibit preassembled siRISC or miRISC activity in vivo. (A) GFP-Cym or GFP-PolV sensor constructs were infiltrated into Cym19stop-infected recovered leaves. GFP mRNA and protein were analyzed 3 days p.i. in Northern blot and Western immunoblot assays, respectively. (B) Preprogrammed miRISC activity is not affected by the presence of p122. The sensor construct GFP-171.1, bearing a perfect target site for miR171, is cleaved irrespective of the presence of the p122 protein. Noncleavable GFP-171.2 sensor, which bears a mutation in the target site, was used as a control. HC-Pro coinfiltration was also used as a control. For Western blot analysis, anti-GFP and anti-His antibodies were applied.

FIG. 5.

Affinities of p122 protein for different RNA duplexes. Twenty-one-nucleotide bona fide siRNA and 19-nt blunt-ended RNA duplexes (A) or 21-nt siRNA and miR171 duplexes (B) were incubated with a dilution series of p122-infiltrated plant extracts and loaded on a 5% native 0.5× Tris-borate-EDTA gel. (C) Determination of relative binding affinities of p122 extract for different RNA molecules. (D) Relative affinities of p122 protein for miRNA duplexes miR171a, miR171b, and miR171c compared with siR171. (E) Binding affinities for 21-nt siRNA and for 49-nt dsRNA using pBIN-p122-infiltrated or cr-TMV-infected plant extract. The complexes formed ran at the same mobility. Control reactions are shown without protein extract or with empty-vector-infiltrated plant extract. (F) Plant extracts infected with mutant virus not coding for p122 do not show any binding.

FIG. 6.

Accumulation and 3′-terminal methylation of miRNAs and siRNAs upon cr-TMV infection. (A) Elevated accumulation of the different miRNAs and miRNA*s have been observed in cr-TMV-infected plants. Total RNAs were extracted from cr-TMV-infected or uninfected wt or mutant A. thaliana plants as indicated. For hybridization, labeled locked nucleic acid oligonucleotides complementary to the indicated miRNA and miRNA* were used. For virus-derived siRNAs and for U6 as a loading control, labeled complementary transcripts were applied as probes. For the identification of the methylation statuses of miRNAs and siRNAs, total RNAs from infected and uninfected plants were subjected (+) or not (−) to the β-elimination reactions. (B) Northern analysis of AGO1 mRNA from mock- and virus-infected A. thaliana wt and mutant plants. Mouse total RNA was used as a control.

FIG. 7.

Effect of p122 protein on the methylation of GFP-derived siRNA and overexpressed miRNA 171c. Total RNA extracts were subjected (+) or not (−) to β-elimination reactions, and the blots were hybridized as for Fig. 6. (A) GFP-IR-derived siRNAs are fully methylated in the absence of suppressor proteins. HC-Pro completely inhibited the methylation of 21-nt-long siRNAs and partially inhibited the methylation of 22-nt siRNAs (lane 6). p122 protein partially interfered with the methylation of 21- and 22-nt-long siRNA species (lane 8). Neither suppressor blocked the methylation of 24-nt siRNA molecules (lanes 6 and 8). Nonmethylated synthetic RNA oligonucleotides were used as positive controls for β-elimination reactions (lanes 9 and 10). (B) Overexpressed miR171c and miR171c* are fully methylated in the absence of silencing suppressors (line 4). HC-Pro and p122 proteins partially inhibited the methylation of miR171c (lanes 6 and 8) but not that of miR171c*. Positive controls for the β-elimination test are shown in lanes 10 and 12.