Fig 1.
HCpro, the potyviral suppressor of RNA silencing, is the protein responsible for PVA-induced granules (PGs).
P0YFP was co-expressed with PVAWT (A) and alone (B), and P1YFP (C) and P2YFP (D) were expressed with PVAWT in Nicotiana benthamiana leaves using agroinfiltration and imaged by confocal microscopy three days later. RFP was expressed in A-D to visualize the nucleus. In (A- D) the left panel shows the YFP signal and the right panel an overlay view of YFP, nuclear RFP and the bright field (BF) as Z-stack projections. Arrowheads point out P0YFP-labeled PGs in (A). Scale bar; 10 μm. (E) Frequency of cells/mm2 containing PGs in leaves expressing P0YFP together with PVA proteins P1, HCpro, CI, 6K2, VPg, NIa, NIb or CP. PVAΔGDD was expressed as a positive and GUS as a negative control for PG induction. (F) Frequency of cells/mm2 containing PGs in leaves expressing P0YFP together with GUS, PVAΔGDD or PVAΔHCΔGDD (left panel) or GUS, PVA or PVAΔHC (right panel). (G) Frequency of cells/mm2 showing PGs in leaves expressing P0YFP with PVAWT, PVACPmut, PVAΔGDD or HCpro plotted as a function of time, (H) a western blot analysis of HCpro accumulation in PVAWT, PVACPmut, PVAΔGDD and HCpro samples using anti-HCpro PAb four days post infiltration and (I) virus derived RLUC activities analyzed in the same samples as in (H), again plotted as a function of time (days post infiltration). Modified PVA infectious cDNA constructs used in this study are schematically presented in (S1 Fig). All quantitative data is presented as means and the error bars indicate the standard deviations. (p < 0.01 **, p < 0.05 *).
Fig 2.
HCpro and several host proteins involved in cellular RNA metabolic processes associate with PGs.
HCproRFP was used to induce PGs (A-D). Host proteins fused to fluorescing proteins were co-expressed with HCproRFP as indicated using agroinfiltration of N. benthamiana leaves (using OD600 0.1 for each construct), and their localization was examined by confocal microscopy three days later. HCproRFP co-expressed with P0YFP (A), UBP1YFP (B), DCP1CFP (C) and AGO1CFP (D). The images are Z-stack projections to clearly show the absence or presence of PGs in the cell and signal overlaps were verified from single layer images. Scale bar; 10 μm. (E) The frequency of cells/mm2 showing granule structures labeled by P0YFP, UBP1YFP and AGO1CFP was compared when these host proteins were co-expressed with either HCproRFP or RFP alone (see also S3 Fig). Since DCP1 granules (PBs) are constantly present in all cells regardless of HCpro, these data are not given for DCP1-granules. (F) The frequency of cells/mm2 showing HCpro granules when granule components UBP1, P0, AGO1 and DCP1 were co-expressed using high Agrobacterium concentrations (OD600 0.5). GFP was expressed as control. (p < 0.01 **, p < 0.05 *)
Fig 3.
PGs are RNA granules to which also PVA RNA can localize.
(A) Fractions containing PGs labeled with P0YFP were isolated from PVACPmut infected leaf lysates (+ sample), as described in materials and methods. Lysates from non-infected leaves expressing P0YFP were fractionated similarly (- sample). Three parallel fractions of each sample type were used for fluorescence quantification and relative fluorescence units are given as a mean ± standard deviation. (B) Fractions quantified in (A) were imaged using epifluorescence microscopy to show successful capture of PVA-induced PGs. Scale bar; 500 μm. (C) PG fractions were isolated from mock-infiltrated leaves (-) and PVACPmut RNA-expressing leaves (+) at 3 DAI and subjected to a western blot detection of endogenous HCpro, P0 and AGO1. The asterisks denote the expected position of the corresponding protein. (D) Similar samples enriched for PGs as in (A and B), were incubated with or without RNAse A and imaged using epifluorescence microscopy. Scale bar; 20 μm. (E) RNAse A-mediated release of fluorescence from isolated PGs (D) was quantified by analyzing fluorescence in total, soluble and low-speed pellet fractions after RNAse A treatment compared to control samples (n = 3). (F) PG fractions were prepared from leaves expressing PVA together with either P0YFP (control) or Strep-III-tagged P0 (P0SIII), and subjected to Strep-tag based affinity purification. RNA was isolated from the affinity-purified samples and subjected to reverse transcription (RT+), followed by PCR detection of viral RNA. Total RNA from PVA infected leaves was used as a positive control for RT-PCR. The RT-minus control (RT-) was negative. (G) Bacteriophage λ B-box RNA elements were fused to the 3´ UTR within PVA∆GDD icDNA (PVAB-box; S1 Fig). Binding of λN22RFP to the B-box RNA element enabled visualization of PVAB-box RNA in vivo. PGs were induced either by PVA∆GDD (control) or PVAB-box and visualized by P0YFP (green channel), and λN22RFP was co-expressed to label B-box RNA. The RFP signal was mainly found in nuclei due to the nuclear localization signal present in λN22RFP, but also in the cytoplasm and PGs in the presence of PVAB-box (magenta channel). The images are projections of Z-stacks with a single layer inset from the area indicated with an arrow. Scale bar; 10 μm.
Fig 4.
HCpro mutants lacking RNA silencing suppression and eIF4E binding are deficient in PG induction.
P0YFP was co-expressed with HCproRFP (A), 4Ebd-HCproRFP (B) or sd-HCproRFP (C) in N. benthamiana leaves using agroinfiltration. P0YFP and HCproRFP signals were detected by confocal microscopy at 3 DAI and presented as Z-stack projections covering an area of multiple epidermal cells to clearly convey the frequency of PGs. Scale bar; 50 μm. (D) In order to achieve comparable levels of native HCproRFP and the mutants, 4Ebd-HCproRFP and sd-HCproRFP expression was increased by increasing the Agrobacterium concentration used in infiltrations (indicated by the OD600 values). P0YFP was co-expressed with these. Protein accumulation was followed by western blot analysis using anti-RFP for HCpro and anti-GFP for P0 detection. (E) The frequency of cells/mm2 showing PGs was calculated using fluorescence microscopy in parallel with analyzing the protein levels in (D). (F) Frequency of cells/mm2 showing P0-labeled granules during co-expressed of P0YFP with plant viral VSRs HCpro, P19, 2b and P25. (G) Frequency of cells/mm2 showing YFP-tagged TuMV HCpro or HCpro AS9 in PG-like structures. The PG frequencies are presented as means and the error bars indicate the standard deviations. (p < 0.01 **, p < 0.05 *).
Fig 5.
Viral protein VPg together with HCpro and other PG components regulates viral translation.
P0YFP and PVACPmut were expressed to induce PGs in N. benthamiana leaves, and the effect of co-expressed RFP (A) or VPgRFP (B) on P0YFP-labeled PGs was examined by confocal microscopy three days later. RFP fluorescence was used to verify VPg and control expressions. Scale bar; 100 μm. (C—D) Frequency of cells/mm2 containing P0YFP-labeled PGs in leaf tissues during expression of either RFP or VPgRFP. PVACPmut served as an inducer of the PGs in (C) and HCpro in (D). (E—F) PVAΔGDD RLUC activities and RNA levels were determined during co-expression of the vRNA with GUS, HCpro, VPg, VPg and HCpro in N. benthamiana leaves at 3 DAI. (G) PVAΔHCΔGDD RLUC activity levels were determined during co-expression of vRNA with GUS, VPg, VPg and either HCpro, 4Ebd-HCpro or sd-HCpro in N. benthamiana leaves at 3 DAI. (H) The capacity of VPg to elevate PVAΔGDD translation was analyzed during RNA hairpin (hp)-induced silencing of UBP1 and VCS in N. benthamiana leaves. Empty non-recombined silencing vector (hp-) and phytoene desaturase (PDS) hp-constructs were used as controls. VPg or GUS was expressed with PVAΔGDD in the different silencing backgrounds and RLUC activity was determined as a reporter of viral protein expression at 5 DAI. (I) The capacity of UBP1 and UBP1rrm to affect VPg promoted PVAΔGDD RLUC translation was analyzed. Here, GUS or VPg were co-expressed with PVAΔGDD during ectopic expression of GUS, UBP1, or UBP1rrm mutant lacking RNA-binding domains. All quantitative data is presented as mean and the error bar indicates the standard deviation. (J) The effect of UBP1 and VCS silencing and the effect of UBP1rrm overexpression on PVA-induced PG formation was analyzed by determining the frequency of cells/mm2 showing PGs in N. benthamiana leaves at 3 DAI. (p < 0.01 **, p < 0.05 *).
Fig 6.
Viral suppressors of RNA silencing HCpro and 2b support VPg-mediated viral translation.
Various VSRs were expressed with PVA RNA either alone or together with VPg in N. benthamiana leaves, and their effect on RLUC activity derived from PVAΔGDD (A) and PVAΔHCΔGDD (B) was measured at 3 DAI. In (C), the FLUC activities derived from a non-viral control FLUC mRNA co-infiltrated with the PVA constructs was determined from the same samples as in (A and B). FLUC activities were not affected by VSRs and were in average decreased in the presence of VPg showing that RLUC changes in (A and B) were virus-specific. (D) PVAΔHC was inoculated using a low Agrobacterium density (OD600 0.0005) together with indicated viral suppressors of silencing. Potentially successful infections were allowed to spread from the initially transformed cells and viral RLUC activities were determined 6 DAI. Luciferase activities are presented as means and the error bars indicate the standard deviation. (p < 0.01 **, p < 0.05 *).
Fig 7.
PGs are important for PVA infection and they can associate with viral replication complexes.
(A-B) PG-associated proteins UBP1, VCS, P0 and eIF4E/(iso)4E were silenced in N. benthamiana leaves via hairpin (hp)-constructs. Empty hairpin construct (hp-) and hpPDS were used as controls and hpRLUC to block infection via vRNA silencing. To initiate PVA infection from individual cells, RLUC-tagged PVAWT was inoculated using a low Agrobacterium density (OD600 0.0005). Infection was then allowed to spread from the initially infected cells in the silencing backgrounds. Six days later, viral RLUC activities (A) and the coat protein levels (B) were determined and they are presented as means with error bars indicating the standard deviation. (C—D) PG-associated proteins UBP1 and VCS were silenced in N. benthamiana using TRV-mediated VIGS followed by mechanical inoculation of PVA and quantification of PVA in systemically infected leaves using ELISA (C) and qPCR (D) at 7 DAI. (E) Plants were inoculated with PVA tagged with an additional copy of 6K2CFP and infection was let to spread into systemic, non-inoculated leaves. P0YFP, VCSYFP and UBP1YFP were subsequently expressed in the systemically infected leaves using Agrobacterium infiltration and imaged together with 6K2-labeled VRCs. The overlay shows PG-markers and VRCs as separate structures adjacent to each other in the presented single-layer images. Scale bar; 3 μm. (F—G) N. benthamiana plants were inoculated with PVA-HCRFP-6KY icDNA tagged simultaneously with HCproRFP and an additional copy of 6K2YFP (see schematic presentation of the construct in S1 Fig) using Agrobacterium infiltration (F) and virion inoculation (G). PGs were visualized via HCproRFP and VRCs via 6K2YFP in upper, non-inoculated leaves by confocal microscopy upon development of systemic infection, and presented as single layer images. Similar to (E) PGs were detected in the vicinity of VRCs as well as separated from them. Scale bar; 10 μm (upper) and 5 μm (lower). (p < 0.01 **, p < 0.05 *).
Fig 8.
A model for PGs in PVA infection.
Viral RNAs engage in different processes after replication including translation. This involves recruitment of the cytoplasmic translational machinery, which potentially exposes viral RNAs to hostile cytoplasmic conditions. RNA silencing-associated structures, potyvirus-induced granules and polysomes may overlap to a yet unknown extent, but are illustrated here as alternative destinations for replicated viral RNA. HCpro, UBP1, VCS, P0 and AGO1 are redistributed to PGs together with translationally inactive viral RNA when active viral translation is not supported. PG assembly and translational repression are promoted by UBP1 whereas translational activation of viral RNA occurs via an cooperative action orchestrated by VPg together with PG components including P0, VCS, HCpro and eIF(iso)4E. VPg disrupts the formation of PGs, underscoring that PG formation and viral translation are interrelated processes in PVA RNA gene expression. We propose that PGs play an important role in protecting viral RNA from antiviral silencing and thereby necessary to achieve optimal virus accumulation in plants where RNA silencing is active.