RNA decay is an antiviral defense in plants that is counteracted by viral RNA silencing suppressors

Abstract

Exonuclease-mediated RNA decay in plants is known to be involved primarily in endogenous RNA degradation, and several RNA decay components have been suggested to attenuate RNA silencing possibly through competing for RNA substrates. In this paper, we report that overexpression of key cytoplasmic 5'-3' RNA decay pathway gene-encoded proteins (5'RDGs) such as decapping protein 2 (DCP2) and exoribonuclease 4 (XRN4) in Nicotiana benthamiana fails to suppress sense transgene-induced post-transcriptional gene silencing (S-PTGS). On the contrary, knock-down of these 5'RDGs attenuates S-PTGS and supresses the generation of small interfering RNAs (siRNAs). We show that 5'RDGs degrade transgene transcripts via the RNA decay pathway when the S-PTGS pathway is disabled. Thus, RNA silencing and RNA decay degrade exogenous gene transcripts in a hierarchical and coordinated manner. Moreover, we present evidence that infection by turnip mosaic virus (TuMV) activates RNA decay and 5'RDGs also negatively regulate TuMV RNA accumulation. We reveal that RNA silencing and RNA decay can mediate degradation of TuMV RNA in the same way that they target transgene transcripts. Furthermore, we demonstrate that VPg and HC-Pro, the two known viral suppressors of RNA silencing (VSRs) of potyviruses, bind to DCP2 and XRN4, respectively, and the interactions compromise their antiviral function. Taken together, our data highlight the overlapping function of the RNA silencing and RNA decay pathways in plants, as evidenced by their hierarchical and concerted actions against exogenous and viral RNA, and VSRs not only counteract RNA silencing but also subvert RNA decay to promote viral infection.

Fig 1. Molecular characterization of 5’RDGs.

(A) qRT-PCR analysis of NbDCP1, NbDCP2, NbXRN4 and NbPARN expression levels in different tissues of N. benthamiana. Expression was normalized against NbActin transcripts, which serve as an internal standard. Each mean value was derived from three independent experiments (n = 3 samples). Values represent the mean ± standard deviation (SD). (B) Micrographs showing cells from leaves of H2B-RFP transgenic N. benthamiana expressing YFP-NbDCP1, YFP-NbDCP2, YFP-NbXRN4 or YFP-NbPARN. Bars = 50 μm. (C) Western Blot (WB) analysis of total protein extracts from infiltrated leaves as indicated in (B) at 32 hours post infiltration (hpi), antibody against GFP (WB:GFP) was applied. Red asterisks indicate the expected band sizes. Coomassie brilliant blue-stained Rubisco large subunit was used as a loading control.

Fig 2. Interactions and subcellular co-localization of NbDCP1, NbDCP2 and NbXRN4.

(A) BiFC assays between NbDCP1, NbDCP2 and NbXRN4 in H2B-RFP transgenic N. benthamiana leaves at 32 hpi. Yellow fluorescence (green) resulted from the interaction of two tested proteins tagged by the C-terminal half of YFP (C-YFP) or the N-terminal half of YFP (N-YFP). Nuclei of tobacco leaf epidermal cells are indicated by expression of the H2B-RFP transgene (red). P3N-PIPO tagged by C-YFP or N-YFP serves as a negative control. Bars = 50 μm. (B) Co-localization of NbDCP1, NbDCP2 and NbXRN4 in the H2B-RFP transgenic N. benthamiana leaf cells at 32 hpi. The yellow signals result from the overlapping of NbDCP1-CFP (green) with YFP-NbDCP2 (red) or YFP-NbXRN4 (red), or NbDCP2-CFP (green) with YFP-NbXRN4 (red). Insets are the enlarged images of the areas in white boxes in the corresponding panels. H2B-RFP is shown in blue. Bars = 50 μm.

Fig 3. Expression of NbDCP1, NbDCP2, NbXRN4, or NbPARN fails to suppress GFP-induced RNA silencing in N. benthamiana plants.

(A) Schematic representation of the RNA fragment derived from expression vectors GFP, GF and dsGF. (B, C) Pictures of representative agroinfiltrated leaves taken at 5 dpi under UV light. Leaf patches were agroinfiltrated with three vectors including 35S-GFP, 35S-GF (or 35S-dsGF) and one of the following vectors: an empty vector (Vec), Myc-tagged-NbDCP1, NbDCP2, NbXRN4, NbPARN, and TBSV p19 (B). Similar results were obtained from three independent experiments. (D, E) Analyses of relative accumulations of GFP mRNAs by specific qRT-PCR in the infiltrated leaves shown in (B, C) at 5 dpi. NbActin serves as an internal standard. Each mean value was calculated based on three independent experiments (n = 3 samples). Values represent the mean ± SD. Double asterisks indicate a highly significant difference compared to 35S-GFP+35S-GF+Vec (D) or 35S-GFP+35S-dsGF+Vec (E) (P < 0.01, Student’s t test). (F, G) Accumulations of GFP protein, Myc-tagged-NbDCP1, NbDCP2, NbXRN4, NbPARN protein, GFP mRNAs and siRNAs in the infiltrated leaves shown in (B, C) at 5 dpi. Protein levels were analyzed by immunoblot analysis using antibodies against GFP (WB:GFP) or Myc (WB:Myc). Coomassie brilliant blue (CBB) staining of the large subunit of Rubisco, ethidium bromide staining of rRNA, and U6 serve as a loading control for immunoblot, mRNA blot and siRNA blot, respectively. The values of GFP siRNAs/U6 were quantified by ImageJ software and then normalized against the mean value corresponding to the Vec treatment, which was set to 1.00.

Fig 4. Silencing of NbDCP1, NbDCP2, NbXRN4, or NbPARN suppresses S-PTGS, but not IR-PTGS in N. benthamiana plants.

(A, D) Visualization of green fluorescence of representative agroinfiltrated leaves. Leaf patches were agroinfiltrated with three expression vectors including 35S-GFP, 35S-GF (or 35S-dsGF) and one of the following vectors: an empty vector (Vec), dsNbDCP1, dsNbDCP2, dsNbXRN4, dsNbPARN, P19, Myc-tagged-NbDCP1, NbDCP2, NbXRN4, and NbPARN. Pictures were taken at 5 dpi under UV light. Similar results were obtained from three independent experiments. (B, E) Relative accumulation of GFP mRNAs in agroinfiltrated patches of the infiltrated leaves indicated in (A) and (D), respectively. Total RNA was isolated at 5 dpi and GFP mRNA was analyzed by qRT-PCR. NbActin serves as an internal standard. Each mean value was derived from three independent experiments (n = 3 samples). Values represent the mean ± SD. Double asterisks indicate a highly significant difference compared to 35S-GFP+35S-GF+Vec (B) or 35S-GFP+35S-dsGF+Vec (E) (P < 0.01, Student’s t test). (C, F) Accumulation of GFP protein, GFP mRNA and siRNAs in the infiltrated leaves shown in (A) and (D), respectively. Samples were collected at 5 dpi. CBB staining of the large subunit of Rubisco, ethidium bromide staining of rRNA, and U6 serve as a loading control for immunoblot, mRNA blot and siRNA blot, respectively. The values of GFP siRNAs/U6 were quantified by ImageJ software and then were normalized against the mean value corresponding to the Vec treatment, which was set to 1.00.

Fig 5. Expression of NbDCP1, NbDCP2, NbXRN4, or NbPARN suppresses GFP expression in RDR6-deficient N. benthamiana plants.

(A) Visualization of green fluorescence of representative agroinfiltrated leaves. Pictures of representative agroinfiltrated leaves were taken at 5 dpi under UV light. Wild type (Wt) N. benthamiana or dsRDR6 transgenic N. benthamiana (dsRDR6) plants were agroinfiltrated with three expression vectors including 35S-GFP, 35S-GF and one of the following vectors: an empty vector (Vec) as a control, Myc-tagged-NbDCP1, NbDCP2, NbXRN4, and NbPARN. Similar results were obtained from three independent experiments. (B) Accumulation of GFP protein, GFP mRNA and siRNAs in the infiltrated leaves shown in (A) at 5 dpi. CBB staining of the large subunit of Rubisco, ethidium bromide staining of rRNA, and U6 serve as a loading control for immunoblot, mRNA blot and siRNA blot, respectively. (C) Relative accumulation of GFP mRNAs analyzed by specific qRT-PCR in the infiltrated leaves shown in (A) at 5 dpi. NbActin serves as an internal standard. Each mean value was calculated based on three independent experiments (n = 3 samples). Values represent the mean ± SD. Double asterisks indicate a highly significant difference compared to 35S-GFP+35S-GF+Vec/dsRDR6 (P < 0.01, Student’s t test). The values of GFP siRNAs/U6 were quantified by ImageJ software and then were normalized against the mean value corresponding to the Vec treatment in Wt N. benthamiana plants, which was set to 1.00.

Fig 6. TuMV infection upregulates the RNA decay pathway in N. benthamiana.

(A, B) The expression levels of NbDCP1, NbDCP2, NbXRN4 and NbPARN were analyzed by qRT-PCR in mock (infiltration buffer) or TuMV-infiltrated N. benthamiana leaves at 3 dpi (A) or upper new leaves at 10 dpi (B). NbActin was used as an internal standard. Each mean value was calculated based on three independent biological repeats (n = 3 samples). Values represent the mean ± SD. Double asterisks indicate a highly significant difference compared to mock at 3 dpi (A) or 10 dpi (B) (P < 0.01, Student’s t test). (C) Confocal microscopy analysis of cells co-expressing NbDCP1-CFP (green) and 6K2-YFP (panel I, red), or 6K2-NIa-VPg-YFP (panel II, red), or NIb-YFP (pane III, red) at 32 hpi. Bars, 25 μm. (D) Confocal microscopy of cells co-expressing NbDCP1-YFP (green) and TuMV-6K2-mCherry (panel I and panel II, red) or TuMV-CFP-NIb (panel III, red) at 72 hpi. The enlarged image of the area in the white box in panel I is shown in panel II. Bars = 50 μm.

Fig 7. Expression of NbDCP1, NbDCP2, NbXRN4, or NbPARN reduces TuMV RNA accumulation.

(A, B) GFP fluorescence in Wt or RDR6-deficient (dsRDR6) N. benthamiana leaves co-infiltrated with TuMV-GFP and one of the following vectors: Vec, NbDCP1, NbDCP2, NbXRN4 or NbPARN at 3 dpi. (C) qRT-PCR analyses of TuMV RNA levels. RNA was extracted from the infiltrated patches shown in (A) at 3 dpi. Each value was normalized against NbActin transcripts in the same sample. Error bars represent SD (n = 3 independent biological repeats). Double asterisks indicate a highly significant difference compared to the treatment of Vec (P < 0.01, Student’s t test). (D, E) GFP fluorescence in Wt or RDR6-deficient (dsRDR6) N. benthamiana leaves co-infiltrated with TuMV-GFP-ΔGDD (a replication-defective mutant) and one of the following vectors: Vec, NbDCP1, NbDCP2, NbXRN4 or NbPARN at 3 dpi. (F) Accumulation of TuMV siRNAs in the infiltrated patches shown in (A, B, D, E) at 3 dpi. Northern blotting was performed using DIG-labeled DNA probes complementary to the TuMV genome. U6 serves as a loading control for siRNA blot, respectively. The values of GFP siRNAs/U6 were quantified by ImageJ software and then were normalized against the mean value corresponding to the Vec treatment, which was set to 1.00.

Fig 8. Knock-down of NbDCP1, NbDCP2, NbXRN4 or NbPARN facilitates TuMV infection.

(A) GFP fluorescence in N. benthamiana leaves co-infiltrated with TuMV-GFP and one of the following expression vectors: an empty vector (Vec) as a control, NbDCP1, dsNbDCP1, dsNbDCP2, dsNbXRN4 and dsNbPARN at 4 dpi. (B) Relative TuMV RNA levels determined by qRT-PCR. RNA was extracted from the infiltrated patches shown in (A) at 4 dpi. Each value was normalized against NbActin transcripts in the same sample. Error bars represent SD (n = 3 independent biological repeats). *, P < 0.05; **, P < 0.01 (Student’s t test). (C) GFP fluorescence in systemic leaves of plants pre-treated with TRV-GUS, TRV-NbDCP1, TRV-NbDCP2, TRV-NbXRN4, or TRV-NbPARN and then infected by TuMV-GFP was photographed under UV light at 6 dpi. (D) Relative TuMV RNA levels determined by qRT-PCR. RNA was extracted from plants in (C) at 14 dpi. *, P < 0.05; **, P < 0.01 (Student’s t test). (E) Accumulation of GFP protein and TuMV siRNAs in the systemic leaves of plants in (C) at 14 dpi. CBB staining of the large subunit of Rubisco and U6 serve as a loading control for immunoblot, mRNA blot and siRNA blot, respectively. The values of GFP siRNAs/U6 were quantified by ImageJ software and then were normalized against the mean value corresponding to the TRV-GUS treatment, which was set to 1.00.

Fig 9. Yeast two-hybrid (Y2H) assays between the four 5’RDGs and 11 TuMV viral proteins.

(A, B) Summary of Y2H assay results. Y2H Gold yeast strains co-transformed with the indicated plasmids were plated on synthetic dextrose (SD)/-Trp, -Leu, -His, -Ade medium to identify protein interactions at 3 days after transformation. Proteins were fused to either the Gal4 DNA binding (BD) or activation (AD) domain. No means no positive interaction between two tested proteins, Yes means positive interaction between two tested proteins, and ‘-’ means that NbDCP1 has self-activation activity when fused to BD vector, indicating a unauthentic interaction. (C, D) Y2H assays for NbDCP2 and VPg (C), and NbXRN4 and HC-Pro (D). Y2H Gold yeast strains co-transformed with the indicated plasmids were subjected to 10-fold serial dilutions and plated on SD/-Trp, -Leu, -His, -Ade medium to identify protein interactions at 3 days after transformation. Cells co-transformed with AD-T7-T+BD-T7-53 serve as positive controls; cells co-transformed with AD-NbDCP2 or AD-NbXRN4 and the empty BD, or the empty AD and BD-NbDCP2 or BD -NbXRN4 are negative controls.

Fig 10. VPg negatively regulates the formation of cytoplasmic NbDCP1/NbDCP2 granules.

(A) BiFC assays for possible protein-protein interactions in planta. NbDCP1, NbDCP2 and VPg were fused with YN or YC. The fused proteins were transiently expressed in H2B-RFP transgenic N. benthamiana leaves. Confocal microscopy was carried out at 32 hpi. Yellow fluorescence (green) was observed in the leaf cells co-expressing NbDCP1 + NbDCP2 or NbDCP2 + VPg, but not in the cells co-expressing NbDCP1 + VPg, or NbDCP1 + NbDCP2 in the presence of VPg. Nuclei of tobacco leaf epidermal cells are indicated by expression of the H2B-RFP transgene (red). Bars = 25 μm. (B) Co-localization of VPg + NbDCP1, VPg + NbDCP2, NbDCP1 + NbDCP2 in the presence of an empty vector (+Vec) or VPg (+VPg) in wild type N. benthamiana leaf cells. Confocal microscopy was carried out at 32 hpi. Bars = 25 μm. (C) The average number of NbDCP1/NbDCP2 co-localization granules (per 10 cells) when they were co-infiltrated with Vec (+Vec) or VPg (+VPg). Independent infiltration experiments were repeated three times and 30 cells in total were used to quantify. Values represent the mean number of the NbDCP1/NbDCP2 granules (per 10 cells) ± SD. Student’s t test was performed to compare differences, and double asterisks indicate a highly significant difference (P < 0.01). (D) Immunoblotting analyses of NbDCP2-CFP at 2 dpi. Total proteins or proteins isolated from the cytoplasm or nuclei were probed with GFP antibodies, and total proteins were also incubated with Myc antibodies to detect Myc-VPg. Equal loading for the nuclear and total protein samples was monitored by probing with Histone H2B antibodies and CBB staining, respectively.

Fig 11. HC-Pro interacts with XRN4 and inhibits XRN4 activity.

(A) BiFC assays to confirm the HC-Pro and NbXRN4 interaction in H2B-RFP N. benthamiana leaves at 32 hpi. HC-Pro and NbXRN4 were fused with YN and YC. Yellow fluorescence (green) was observed in the cells co-expressing YN-HC-Pro+YC-NbXRN4 or YC-HC-Pro+YN-NbXRN4. Co-expression of YN-HC-Pro+YC-NbDCP1 or YC-HC-Pro+YN-NbDCP1 did not result in any detectable fluorescence in H2B-RFP N. benthamiana leaves at 32 hpi, which revealed that HC-Pro did not interact with NbDCP1. Bars = 25 μm. (B) Co-localization analysis of HC-Pro-CFP and YFP-NbXRN4 in N. benthamiana leaves at 32 hpi. Panel I: HC-Pro-CFP was expressed alone, Panel II: YFP-NbXRN4 was expressed alone, Panel III: HC-Pro-CFP and YFP-NbXRN4 were expressed together. Bars = 25 μm. (C) Y2H assays for AtXRN4 and HC-Pro. Y2H Gold yeast strains co-transformed with the indicated plasmids were subjected to 10-fold serial dilutions and plated on SD/-Trp, -Leu, -His, -Ade medium to identify protein interactions at 3 days after transformation. Cells co-transformed with AD-T7-T+BD-T7-53 serve as positive controls; cells co-transformed with AD-HC-Pro and the empty BD, or the empty AD and BD-AtXRN4 are negative controls. (D) BiFC assays revealed that HC-Pro interacted with AtXRN4 in H2B-RFP N. benthamiana leaves at 32 hpi. P3N-PIPO and AtXRN4 served as a negative control for the protein-protein interaction assay. Bars = 25μm. (E) Phenotypes of Col-0, xrn4 mutant, and transgenic Arabidopsis plants transformed with 35S:Myc-AtXRN4, 35S:HC-Pro-CFP-1, 35S:HC-Pro-CFP-2 and 35S:HC-Pro-CFP-1/35S:Myc-AtXRN4 at 20 days after sowing. 35S:HC-Pro-CFP-1/35S:Myc-AtXRN4 plants were obtained by genetic crosses between 35S:HC-Pro-CFP-1 and 35S:Myc-AtXRN4 Arabidopsis plants. T2 generation plants were used in the above experiments. The confirmation of xrn4 mutant, transgenic plants expressing 35S:Myc-AtXRN4, 35S:HC-Pro-CFP-1, 35S:HC-Pro-CFP-2 and 35S:HC-Pro-CFP-1/35S:Myc-AtXRN4 was shown in S7 Fig and S10 Fig. (F, G, H) qRT-PCR analysis of AtEBF1, AtRAP2.4, AtNMT expression in Col-0, xrn4 mutant, and transgenic plants carrying 35S:Myc-AtXRN4, 35S:HC-Pro-CFP-1 and 35S:HC-Pro-CFP-1/35S:Myc-AtXRN4 at 20 days after sowing (F), in Col-0 and xrn4 mutant leaves infiltrated by empty vector (Vec) and HC-Pro at 3 dpi (G), and in Col-0 newly emerged leaves of mock- or TuMV-infected plants 12 dpi (H). AtActinII gene was used as an internal control. The data were analyzed using Student’s t test and asterisks denote significant differences between treatments (* P <0.05, ** P <0.01).

Fig 12. Model for counter-defense by VSRs against cytoplasmic RNA decay- and RNA silencing-mediated plant defenses.

After transcription, mRNAs are deadenylated, decapped and degraded by XRN4-mediated 5’–3’ decay or exosome-mediated 3’–5’ degradation. Deadenylated, decapped, or XRN4-partially degraded RNAs facilitate RDR6 to transform ssRNA into dsRNA to degrade the left RNAs by PTGS pathway. Potyviruses encode at least two VSRs: HC-Pro and VPg. HC-Pro interacts with XRN4 to inhibit its slicing activity and VPg disrupts the interaction between NbDCP1 and NbDCP2 by targeting NbDCP2 to the nucleus. As a result, RNA decay-mediated plant defense is compromised. In addition, VPg interacts with SGS3 and mediates its degradation via ubiquitination and autophagy pathways to block RDR6-mediated anti-viral response [36].