Wild type (WT) N. benthamiana plants, green fluorescent protein (GFP) transgenic (line 16c) N. benthamiana plants and red fluorescent protein (RFP) transgenic N. benthamiana plants were used in the present study. They were grown from seeds in an insect-free growth room at 25°C under a 16 h light and 8 h dark photoperiod. The third and fourth leaves from the top were used for infiltration and inoculation experiments when the plants were about 4 weeks old.

Ten SMoV protein coding regions, including X1, X2, Hel, Pro, Vpg, RdRp encoded by RNA1, MP, CP, Pro2Glu, and P28 encoded by RNA2, were individually amplified by polymerase chain reaction (PCR) from the SMoV Chinese isolate DGHY21 (GenBank accession no. MT070753) according to the method described previously (Fan et al., 2021). Specific primers used in this study were listed in Supplementary Table S1 (containing suitable restriction sites). The PCR products were cloned individually into the pTOPO-Blunt vector (Aidlab, Beijing, China) to generate pTOPO-X1, pTOPO-X2, pTOPO-Hel, pTOPO-Pro, pTOPO-Vpg, pTOPO-RdRp, pTOPO-MP, pTOPO-CP, pTOPO-Pro2Glu, and pTOPO-P28, which were individually digested with specific enzymes for subsequent cloning.

To screen the VSRs, each protein-coding region was subcloned into the pCHF3 vector (Cai et al., 2007) under the control of a 35S promoter and NOS terminator to generate pCHF3-X1, pCHF3-X2, pCHF3-Hel, pCHF3-Pro, pCHF3-Vpg, pCHF3-RdRp, pCHF3-MP, pCHF3-CP, pCHF3-Pro2Glu, and pCHF3-P28. The resulting recombinant pCHF3 constructs were chemically transformed into the Agrobacterium tumefaciens strain EHA105.

For the subcellular localization analysis, the full-length fragments of SMoV Pro2Glu and P28 were inserted into, respectively, the Sma I/Kpn I and Apa I/Kpn I sites of pCAMBIA1300-GFP vector to produce C1300-Pro2Glu and C1300-P28 that contained GFP fusion protein. The resulting recombinant pCAMBIA1300-GFP constructs were chemically transformed into the A. tumefaciens strain GV3101.

For the pathogenicity test, the full-length coding regions of Pro2Glu and P28 were subcloned into the PVX-containing pGR106 vector (a kind gift from David C. Baulcombe, University of Cambridge, Cambridge, United Kingdom) between the Cla I and Sal I restriction sites to produce PVX-Pro2Glu and PVX-P28. The A. tumefaciens strain GV3101 (pJIC SA_Rep) was transformed with the two plasmids.

To generate the Pro2Glu and P28 derivatives, the Glu^{dm329–339aa} mutant with the 329GVSNRKKHRRG339 motif deleted was amplified by PCR from the plasmid pTOPO-Pro2Glu with primers Glu^{dm329–339aa}-F/R. The resulting Glu^{dm329–339aa} was inserted into the BamH I/Pst I sites of pCHF3 or Sma I/Kpn I sites of pCAMBIA1300-GFP to produce pCHF3-Glu^{dm329–339aa} or C1300-Glu^{dm329–339aa}, respectively. The GW motifs of Pro2Glu and P28 were disrupted through W₆A (m4Glu), W₁₁₇A (m1Glu), W₁₇₂A (m2Glu), and W₁₉₁A (m3P28) site-directed mutagenesis using gene synthesis by Sangon Biotech (Shanghai, China) Co., Ltd. The resulting mutants were inserted into the BamH I/Pst I sites of pCHF3 to produce pCHF3-m4Glu, pCHF3-m1Glu, pCHF3-m2Glu, and pCHF3-m3P28.

Agrobacterium-mediated transient gene expression was carried out based on developed protocols (Hamilton et al., 2002). Agrobacterium cultures were pelleted and resuspended to an optical density OD₆₀₀ = 1.0 in a solution containing 10 mM MgCl₂, 10 mM MES (pH 5.8), and 200 μM acetosyringone, and were incubated at room temperature in dark for 3 h before infiltration. For co-infiltration, equal volumes of Agrobacterium cultures harboring 35S-GFP (a 35S promoter-driven construct expressing sense GFP, a kind gift from Prof. Xiuling Yang, Chinese Academy of Agricultural Sciences, Beijing, China) and the tested constructs were mixed, followed by infiltration into fully expanded leaves of 4-week-old 16c or WT N. benthamiana plants. A construct pCHF3-P19 to express tomato bushy stunt virus (TBSV) P19 was used as a positive control, and an empty vector pCHF3 was used as a negative control. GFP fluorescence in infiltrated or systemic leaves was monitored under a handheld long-wavelength UV lamp (Black-Ray Model B-100A, San Gabriel, CA, United States) and was photographed with a Canon EOS-M digital camera featuring a 55 mm yellow filter.

Total RNA was extracted from Agrobacterium-infiltrated leaf patches of N. benthamiana line 16c or N. benthamiana WT plants by using an E.Z.N.A.^® Plant RNA Kit (Omega Bio-tek, Norcross, GA, United States) following the manufacturer’s instructions. Then, 2.5 μg of total RNA was separated on 1.5% formaldehyde-agarose gels and was transferred to Hybond N⁺ membranes as instructed (GE Healthcare, Buckinghamshire, United Kingdom). The membrane was hybridized with a digoxigenin-labeled GFP probe, which was made using a PCR DIG probe synthesis kit, and was detected using a detection starter kit II according to the manufacturer’s instructions (Roche Diagnostic, Basel, Switzerland). GelStain (Transgen, Beijing, China) staining was used to visualize the loading controls for the mRNA.

For siRNA blot assay, total RNA were extracted from infiltrated leaves with Trizol reagent (Transgen, Beijing, China) as recommended by the manufacturer. Low-molecular-mass RNAs were enriched from total RNA by adding equal volume of 10 M lithium chloride and incubating 4 h on ice. The RNA was centrifuged at 13,000 rpm for 15 min. The supernatant was precipitated at −20°C for at least 2 h after adding three volumes of ethanol. The low molecular weight (LMW) RNAs were collected by centrifugation for 15 min at 13,000 rpm. The enriched small RNAs were fractionated on a 15% denaturing polyacrylamide (19:1) gels containing 7 M urea in 0.5 × TBE buffer, then transferred to Hybond N⁺ membranes by electroblotting at 20 V for 60 min, and chemically cross-linked via 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide (EDC) as described previously (Pall and Hamilton, 2008). The membrane was hybridized with the digoxigenin-labeled GFP probe which was hydrolyzed into about 50 nt long RNA pieces by treatment with sodium carbonate buffer as described by Dalmay et al. (2000) at 40°C overnight, subsequently, washed at 50°C in 2 × SSC (0.3 M and 0.03 M sodium citrate) and 0.2% sodium dodecyl sulfate.

Total soluble proteins were extracted from Agrobacterium-infiltrated leaf patches of N. benthamiana WT or N. benthamiana line 16c plants by using a Plant Total Protein Lysis Buffer (Sangon Biotech, Shanghai, China). The protein samples were separated by SDS-PAGE. The anti-GFP mouse monoclonal antibody (TransGen Biotech, Beijing, China) was used at 1:5,000 dilution. Western blots were visualized with a goat anti-mouse IgG (H + L) HRP conjugate (TransGen Biotech, Beijing, China) and a chemiluminescence detection system (Tianneng, Shanghai, China). Coomassie blue staining of the large subunit of Rubisco served as loading controls for the Western blot assay.

The Y2H experiment was performed using Super Yeast Transformation Kit (Coolaber, Beijing, China) according to the manufacturers. Plasmids of the bait and prey pairs were co-transformed into yeast cells and cultured at 30°C for 72 h on double dropout (−Leu/−Trp) or quadruple dropout (−Ade/-His/−Leu/−Trp) synthetic dextrose (SD) medium, respectively.

Total RNA was extracted from N. benthamiana systemic leaves infected by PVX, PVX-Pro2Glu and PVX-P28 at 22 dpi. Reverse transcription was conducted according to the published method (Fan et al., 2021). The RT-qPCR reactions in the 20 μl volume containing 1 μl cDNA as template were run using GoTaq qPCR Master Mix (Promega, Madison, United States) according to the instructions supplied by the manufacturer on a MyGo Pro real-time PCR instrument (IT-IS Life Science Ltd., Republic of Ireland). Primers G-35 and G-36 were used for determining PVX genomic RNA accumulation as described elsewhere (Gupta et al., 2018). PP2A was chosen as the internal control for normalization of virus accumulation in N. benthamiana plants. The 2^-ΔΔCT method was used to analyze the RT-qPCR expression data according to Bustin et al. (2009). GraphPad (GraphPad Software Inc., San Diego, CA, USA) was used to analyze the experimental data by one-way ANOVA, and multiple comparisons were done using Tukey’s test (p < 0.05).

Imaging of fluorescent proteins was conducted using a confocal microscopy (LSM880; Carl Zeiss, Jena, Germany) at 60 h post-infiltration. To confirm the nuclei of the leaf epidermal cells, RFP transgenic N. benthamiana leaves were infiltrated with Agrobacterium cultures harboring the C1300-Pro2Glu or C1300-Glu^{dm329–339aa} construct. For GFP, excitation was performed at 488 nm, and emission was recorded at 500 to 530 nm. For RFP, excitation was done at 561 nm, and emission was recorded at 590 to 630 nm.

Agrobacterium cultures harboring the PVX-Pro2Glu and PVX-P28 vectors were diluted to approximately OD₆₀₀ = 0.8 prior to infiltration into leaves of N. benthamiana plants. The infiltrated plants were examined for virus-like symptoms and photographed at various time-points after infiltration.

To identify VSRs of SMoV, a classical GFP-based two-component Agrobacterium infiltration assay was used to screen potential VSRs. Agrobacterium containing 35S-GFP and plasmids expressing individual proteins (X1, X2, Hel, Vpg, Pro, RdRp, MP, CP, Pro2Glu, and P28) of SMoV (Figure 1A) were mixed and co-infiltrated into fully expanded leaves of N. benthamiana line 16c. Three days post infiltration (dpi), GFP fluorescence was evident in all Agrobacterium-infiltrated leaf patches. By 5 dpi, as a consequence of GFP local silencing, GFP fluorescence became weak in leaves co-infiltrated with 35S-GFP and the empty pCHF3 vector and the tested recombinant pCHF3 vectors expressing X1, X2, Hel, Pro, Vpg, RdRp, MP, and CP. However, GFP fluorescence remained strong in leaves co-infiltrated with 35S-GFP and pCHF3-Pro2Glu or pCHF3-P28, resembling the positive control for silencing suppression (pCHF3-P19) (Figure 1B).

The northern blot analysis was carried out using total RNAs (for GFP mRNA) or LMW RNAs (for GFP siRNA) extracted from patches of co-infiltrated leaves. Consistent with the observed GFP fluorescence, GFP mRNA accumulated to higher levels in leaves co-infiltrated with 35S-GFP and pCHF3-P19, pCHF3-Pro2Glu or pCHF3-P28 compared to empty vector (Figure 1C), however, GFP siRNA accumulated at significantly reduced levels in plants expressing P28 and was not detected in P19 and Pro2Glu expressing plants (Figure 1D). Western blot also showed the accumulation of GFP proteins was higher in the leaves co-infiltrated with Pro2Glu + GFP, P28 + GFP or P19 + GFP than in those co-infiltrated with pCHF3 + GFP (Figure 1E). These results indicate that Pro2Glu and P28 encoded by SMoV RNA2 are capable of suppressing local RNA silencing triggered by single-stranded GFP.

To explore whether Pro2Glu and P28 could suppress dsRNA-induced RNA silencing, Agrobacterium suspension containing 35S-GFP, 35S-dsGFP (a construct expressing an inverted repeat sequence of GFP, a kind gift from prof. Xiuling Yang, Chinese Academy of Agricultural Sciences, Beijing, China) and plasmids expressing individual proteins X1, X2, Hel, Vpg, Pro, RdRp, MP, CP, Pro2Glu, and P28 were inoculated into fully expanded leaves of N. benthamiana. At 3 dpi, only the leaf patches co-inoculated by 35S-GFP, 35S-dsGFP and pCHF3-P19 showed strong GFP fluorescence, whereas leaf patches co-inoculated by 35S-GFP, 35S-dsGFP and empty vector, Pro2Glu, P28 or other eight proteins all showed faded GFP fluorescence (Figure 2A). The Northern blot analysis revealed that the leaves co-inoculated with 35S-GFP, 35S-dsGFP and pCHF3-P19 accumulated high levels of GFP mRNA, while GFP siRNA was not detected. Conversely, in other treatments, GFP mRNA remained undetected (Figure 2B) and high levels of GFP siRNA were detected (Figure 2C). Correspondingly, the western blot analysis revealed that the level of GFP protein expression was stronger in leaf patches co-infiltrated by pCHF3-P19 than in other co-infiltrated treatments (Figure 2D). To confirm this observation, we conducted the assay in a leaf of N. benthamiana. At 3 dpi, the same results were observed (Supplementary Figure 1). These data demonstrate that Pro2Glu and P28 failed to suppress dsGFP-induced RNA silencing.

To determine whether Pro2Glu and P28 could inhibit activation of systemic host RNA silencing based defense responses, we agroinfiltrated two fully expanded leaves of N. benthamiana 16c line with 35S-GFP and empty vector, pCHF3-Pro2Glu, pCHF3-P28 or pCHF3-P19. GFP fluorescence was monitored in the newly emerging leaves of infiltrated plants. The plants with major and minor veins of upper young leaves turning red were assumed to be silenced systemically. At 28 dpi, 95% of plants co-infiltrated with 35S-GFP and empty vector showed systemic silencing, but only 32.4 and 35.3% of plants co-infiltrated with, respectively, 35S-GFP plus pCHF3-Pro2Glu or pCHF3-P28 exhibited systemic silencing (Figure 3A and Supplementary Table 2). The GFP mRNA and siRNA level of plants upper young leaves were analyzed by Northern blot. The results revealed high accumulation of GFP mRNA and negligible accumulation of GFP siRNA in patches co-infiltrated with 35S-GFP plus pCHF3-Pro2Glu, pCHF3-P28 or pCHF3-P19 (Figures 3B,C). These data demonstrate that Pro2Glu and P28 effectively inhibit activation of systemic host RNA silencing-based defense responses.

Some VSRs, such as apple chlorotic leaf spot virus (ACLSV) P50, suppress systemic host RNA silencing without interfering with local silencing. To find out whether SMoV also encodes other suppressor that is unable to suppress local RNA silencing but inhibits systemic RNA silencing, agroinfiltration containing 35S-GFP and the other recombinant plasmids (pCHF3-X1, -X2, -Hel, -Vpg, -Pro, -RdRp, -MP or -CP) into fully expanded leaves of N. benthamiana 16c line was done. At 28 dpi, red fluorescence was observed in newly emerging leaves in every treatment with 35S-GFP and pCHF3-X1, -X2, -Hel, -Vpg, -Pro, or –MP. More than 90% of infiltrated plants showed red fluorescence in the other two treatments with 35S-GFP and pCHF3-RdRp or pCHF3-CP (Figure 3A and Supplementary Table 2). The GFP mRNA level of plants upper young leaves was analyzed by Northern blot. The results were consistent with the observed GFP fluorescence (Figure 3B). Taken together, these data indicate that X1, X2, Hel, Vpg, Pro, RdRp, MP or CP of SMoV do not efficiently inhibit systemic silencing of GFP.

To determine the subcellular localizations of Pro2Glu and P28, we generated GFP-fused constructs C1300-Pro2Glu and C1300-P28, and individually infiltrated them into N. benthamiana leaves. At 60 h post-infiltration, the control leaves expressing GFP showed that GFP protein was localized in the nucleus and cytoplasm. Similarly, Pro2Glu and P28 proteins were observed in the nucleus and cytoplasm (Supplementary Figure 2).

Some studies reported that nuclear localization signal (NLS) is required for VSR suppressor activity (Wang et al., 2004; González et al., 2010; Yang et al., 2018). To explore which regions within Pro2Glu and P28 are crucial for their suppressor activity, we analyzed the SMoV Pro2Glu and P28 sequences using cNLS Mapper (Kosugi et al., 2009), and found that the GVSNRKKHRRG motif between amino acids 329 and 339 might be the NLS of the Pro2Glu protein, but the available NLS motif was not predicted in the P28 sequence. Hence, we deleted the 329GVSNRKKHRRG339 motif from Pro2Glu and constructed a cassette (Figure 4A; C1300-Glu^{dm329–339aa}) to clarify the subcellular localization of the Pro2Glu mutant variant. Parallel with the wild-type Pro2Glu, the fluorescence of C1300-Glu^{dm329–339aa} was generally distributed in the nucleus and cytoplasm (Figure 4B). We further confirmed the nuclear localization of Pro2Glu and Glu^{dm329–339aa} in RFP transgenic N. benthamiana plants (Figure 4B), demonstrating that the region covering aa 329–339 is not required for nuclear localization of Pro2Glu protein.

Furthermore, Glu^{dm329–339aa} was cloned into pCHF3 to evaluate its ability to suppress RNA silencing. Fully expanded leaves of N. benthamiana 16c line were co-infiltrated with 35S-GFP and pCHF3-P19, pCHF3 empty vector, pCHF3-Glu^dm329-339aa or pCHF3-Pro2Glu, and monitored under UV light. At 4 dpi, strong GFP fluorescence was observed in the leaf patches co-infiltrated with 35S-GFP and pCHF3-Pro2Glu or pCHF3-P19. However, GFP fluorescence was weak in the leaf patches co-infiltrated with 35S-GFP and pCHF3-Glu^dm329-339aa or empty vector (Figure 4C). This observation was confirmed by the Northern blot analysis that examined the relative GFP mRNA level extracted from the corresponding leaf patches (Figure 4D). To determine whether the 329GVSNRKKHRRG339 motif deletion destabilized Pro2Glu protein. Western blot analysis was performed to detect the GFP fusion proteins using anti-GFP antibody. Detection of the expected protein sizes revealed that the deletion of the 329GVSNRKKHRRG339 motif significantly decreased the level of the Pro2Glu protein (Figure 4E). Taken together, these results demonstrate that the 329GVSNRKKHRRG339 motif from Pro2Glu is not necessary for Pro2Glu nucleus localization. Since the 11 amino acids deletion makes the Pro2Glu unstable, the role of this motif in RNA silencing remains inconclusive.

The GW motif has been considered to be involved in direct interaction with AGO proteins to inhibit AGO-mediated gene silencing (El-Shami et al., 2007; Lian et al., 2009; Gupta et al., 2019). We analyzed the Pro2Glu and P28 sequences, and identified three conserved GW motifs in Pro2Glu of all available SMoV isolates (Supplementary Figures 3A–C) and a conserved GW motif in P28 of Chinese SMoV isolates (Supplementary Figure 3D). These GW motifs were disrupted through WA site-directed mutagenesis (Figures 5A,B). The resulting mutants were confirmed by Sanger sequencing (Supplementary Figure 4). Agroinfiltration of 35S-GFP along with pCHF3-m1Glu or pCHF3-m4Glu into leaves of N. benthamiana 16c line showed weak green fluorescence under UV light at 4 dpi compared to leaves co-infiltrated with wild-type pCHF3-Pro2Glu. Concurrently, leaves agroinfiltrated with pCHF3-m2Glu or pCHF3-m3P28 showed green fluorescence similar to wild-type pCHF3-Pro2Glu or pCHF3-P28, respectively (Figure 5C).

GFP expression was tested by Northern blot hybridization of total RNA extracted from infiltrated regions at 4 dpi. The results showed that GFP mRNA accumulated at lower levels in pCHF3-m1Glu- or pCHF3-m4Glu-infiltrated leaf patches compared with wild-type pCHF3-Pro2Glu. For pCHF3-m2Glu- or pCHF3-m3P28-infiltrated regions, GFP mRNA accumulation was equivalent to wild-type pCHF3-Pro2Glu or pCHF3-P28, respectively (Figure 5D). These data indicate that two GW (G₅W₆ and G₁₁₆W₁₁₇) motifs of Pro2Glu are indispensable to suppression of local RNA silencing.

Most VSRs have been suggested to enhance synergistically the severity of infection when they are expressed from a heterologous virus-based expression vector (Cao et al., 2005; Cañizares et al., 2008; Tatineni et al., 2012; Samuel et al., 2016). To test the hypothesis that Pro2Glu and P28 could enhance pathogenicity of a heterologous virus, we took advantage of the PVX-based heterologous gene expression system to express the Pro2Glu and P28 proteins in WT N. benthamiana plants. Agrobacterium cultures containing pGR106 empty vector (PVX) or recombinant PVX constructs expressing individual Pro2Glu and P28 were infiltrated in the third and fourth leaves from top of 4-week-old N. benthamiana plants. The inoculated plants were monitored for symptom development. At 10 dpi, non-inoculated systemic leaves exhibited severe mosaic symptoms in N. benthamiana plants inoculated with PVX and PVX-P28 (Supplementary Figure 5). By 22 dpi, the upper leaves inoculated with PVX-P28 appeared crinkled and had necrotic lesions (Figure 6A). However, the plants inoculated with PVX-Pro2Glu showed mild mosaic symptom at 10 dpi (Supplementary Figure 5). By 22 dpi, mild mosaic symptoms in the plants inoculated with PVX-Pro2Glu were similar to the control (PVX-inoculated) plants (Figure 6A). These observations indicate P28 is a putative symptom determinant, and Pro2Glu does not aggravate the symptoms of PVX.

To determine the genomic RNA copies of PVX, RT-qPCR was performed using total RNA extracted from systemically infected leaves. The result showed that PVX-Pro2Glu and PVX-P28 accumulated with, respectively, 47- and 18-fold increases compared to wild-type PVX at 22 dpi (Figure 6B), indicating that Pro2Glu and P28 significantly enhanced heterologous virus PVX accumulation in N. benthamiana plants.