Tomato Spotted Wilt Virus NSs Protein Supports Infection and Systemic Movement of a Potyvirus and Is a Symptom Determinant

Plant viruses are inducers and targets of antiviral RNA silencing. To condition susceptibility, most plant viruses encode silencing suppressor proteins that interfere with antiviral RNA silencing. The NSs protein is an RNA silencing suppressor in orthotospoviruses, such as the tomato spotted wilt virus (TSWV). The mechanism of RNA silencing suppression by NSs and its role in virus infection and movement are poorly understood. Here, we cloned and tagged TSWV NSs and expressed it from a GFP-tagged turnip mosaic virus (TuMV-GFP) carrying either a wild-type or suppressor-deficient (AS9) helper component proteinase (HC-Pro). When expressed in cis, NSs restored pathogenicity and promoted systemic infection of suppressor-deficient TuMV-AS9-GFP in Nicotiana benthamiana and Arabidopsis thaliana. Inactivating mutations were introduced in NSs RNA-binding domain one. A genetic analysis with active and suppressor-deficient NSs, in combination with wild-type and mutant plants lacking essential components of the RNA silencing machinery, showed that the NSs insert is stable when expressed from a potyvirus. NSs can functionally replace potyviral HC-Pro, condition virus susceptibility, and promote systemic infection and symptom development by suppressing antiviral RNA silencing through a mechanism that partially overlaps that of potyviral HC-Pro. The results presented provide new insight into the mechanism of silencing suppression by NSs and its effect on virus infection.


Introduction
RNA silencing contributes to the regulation of gene expression and the maintenance of genome integrity in eukaryotes. In plants, nematodes, and insects, gene silencing is additionally an inducible, adaptable, specific, potent, and essential defense system against virus infection (antiviral RNA silencing) [1][2][3][4][5][6][7]. Both endogenous and antiviral RNA silencing initiate with the processing of double-stranded RNA by Dicer-like (DCL) proteins to form small RNAs of 21 to 24 nt that associate with Argonaute (AGO) proteins to guide slicing or translational repression of target RNAs [2,[7][8][9][10][11][12][13]. Sources of viral RNA that might trigger silencing include replication intermediates and self-complementary sequences in viral RNA, yielding primary virus-derived small interfering RNAs (siRNAs) that are necessary but not sufficient to restrict plant virus infection [1]. Restriction of plant virus infection requires silencing amplification by cellular RNA-dependent RNA polymerases (RDR) to establish an antiviral state [3,9,10].
To replicate and establish local and systemic infection, most plant and insect viruses encode silencing suppressor proteins that interfere with both endogenous and antiviral RNA silencing [14]. Mechanisms used by viral suppressors to interfere with RNA silencing include the sequestration of N. benthamiana and A. thaliana plants. In both host species, chimeric TuMV-GFP clones retained the NSs insert in systemically infected tissue.
Silencing suppression-active NSs but not suppressor-deficient mutant NSs restored pathogenicity to suppressor-deficient TuMV-AS9-GFP. Symptoms developed by systemically infected N. benthamiana leaves consisted of a spotted green and yellow mosaic similar to that caused by TSWV. Using A. thaliana plants lacking core components of the RNA silencing machinery [2,3], a genetic analysis showed that NSs is necessary and sufficient to overcome defense mechanisms dependent on DCL, RDR, and AGO proteins, namely antiviral RNA silencing. Collectively, the results described here show that NSs can functionally replace potyviral HC-Pro and support the establishment of local and systemic infection and symptom development by a suppressor-deficient potyvirus. The NSs silencing suppression mechanism overlaps with but is not identical to that of potyviral HC-Pro.

DNA Plasmids
Standard cloning techniques were used to make all constructs used in this study, gateway entry (pENTR) and destination (pMDC32) vectors [42]. Inserts between the 35S promoter and nopaline synthase (NOS) terminator were verified by Sanger sequencing. The sequence of the oligonucleotide primers is in the 5 to 3 orientation.
pENTR-NIb-NSsHF-CP (TuMV). 6HIS-3xFLAG-tagged NSs (NSs-HF) was introduced in frame between NIb and CP by overlapping PCR. Overlapping fragments A, B, and C were generated as follows: (A) pENTR-NIb-CP as template and primers TuMV-Nib-ENTR and Nib-NSs-R (GTCTGAAT GATCGACTCATAAACACTTGAAGAtgCctggtgataaacacaagcctcagc); (B) pENTR-NSs-6HIS3xFLAG as template and primers Nib-NSs-F (gctgaggcttgtgtttatcaccagGcaTCTTCAAGTGTTTATGAGTCGATCAT TCAGAC) and Flag-CP-R (cctgcatcaagcgtttcacctgcCTG GTGATAGACACACTTGTCATCGTCATCCT TGTAGTCG). (C) pENTR-NIb-CP as template and primers Flag-CP-F (CGACTACAAGGATG ACGATGACAAGTGTGTCTATCACCAGgcaggtgaaacgcttg atgcagg) and pCB302-Ter-Rev. Individual fragments A, B, and C were stitched and amplified with oligos TuMV-Nib-ENTR and pCB302-Ter-Rev and moved into pENTR by TOPO cloning. M was changed to A in NSs to complete an NIa cleavage site at the N terminus of NSs and an extra NIa cleavage site was added between NSs-HF and the coat protein.
Using N. benthamiana plants at the 5 to 6 leaf stage, a standard assay [47] was used to measure silencing suppression of the ssGFP reporter. A. tumefaciens cells carrying the ssGFP (OD600 = 0.25) were infiltrated in combination with NSs (OD600 = 0.5) or controls. Empty pMDC32 and tombusviral p19-HA were used as negative and positive controls, respectively. For each treatment, ten plants were infiltrated at leaves three and four and the experiment was repeated three times. Plants were incubated for four days in a growth chamber at 24 • C with 16 h of light and 8 h of darkness (long day). Photographs of infiltrated leaves were taken under ultraviolet (UV) light at 4 days post infiltration and GFP fluorescence was measured using photographs and the green channel in ImageJ (https: //imagej.net/Welcome) [8,33]. At 4 days post infiltration, infiltrated leaf samples were collected for protein extraction as described [8].
In N. benthamiana and A. thaliana, infection by TuMV and derivatives was launched by agroinfiltration [48]. A. tumefaciens cells carrying TuMV-GFP or derivatives were re-suspended to an OD600 = 0.05. Cells carrying empty pMDC32 were used as the negative control. At the time of infiltration, N. benthamiana plants were at the 5-6 leaf stage and A. thaliana plants were 16 days old. Two N. benthamiana and four A. thaliana leaves were infiltrated per plant, and plants were incubated for 15 days in a growth chamber at 24 • C or 22 • C, respectively, and under long day conditions. At 10 days post inoculation (dpi), leaf samples were collected for protein extraction and analysis. The onset of systemic infection was monitored daily under UV light. The experiment was repeated three times.

Tomato Spotted Wilt Virus Mechanical Inoculation
The Hawaii isolate of (Tomato spotted wilt virus) TSWV was maintained, propagated, and mechanically inoculated as described [8].

Western Blotting
Total protein was extracted as described [8] by grinding leaf samples in 0.5 mL glycine buffer [19] per 1 g of leaves. After a Bradford assay, extracts were normalized to 0.5 mg/mL. For Western blotting, 12.5 µg were used and proteins were separated by gel electrophoresis at 150 V for 60 min and transferred to nitrocellulose membranes. GFP was detected with anti-GFP antibody (Merck Millipore, Darmstadt, Germany) at a 1:4000 dilution. 6HIS3xFlag-tagged NSs was detected with anti-Flag-peroxidase antibody (Sigma-Aldrich, St. Louis, MO, USA) at a 1:10,000 dilution. HA-tagged p19 [15] was detected using anti-HA antibody with peroxidase (Roche, Basel, Switzerland) at a 1:1000 dilution. HSP70 was used as the loading control and detected using primary anti-HSP70 (Merck Millipore, Darmstadt, Germany) at 1:6000 dilution and secondary antibody (goat anti-rabbit immunoglobulin G, NA934-1; GE Healthcare, Little Chalfont, UK) at a 1:10,000 dilution. TuMV CP was detected with anti-CP (PVAS-134 at 1:10,000 dilution) with secondary antibody NA934-1 at a 1:10,000 dilution. Primary and peroxidase-conjugated antibodies were incubated at 4 • C overnight and secondary antibodies for 30 min at room temperature. Chemiluminescence was measured with Clarity Western ECL Substrate and a ChemiDoc ® MP Imaging System (Bio-Rad, Hercules, CA, USA).

Confocal Microscopy
Systemically infected N. benthamiana leaves and controls were collected at 10 dpi for microscopy. Colorimetric images were obtained using a scanner (Epson V600) with visible light. A Nikon TI-2 microscope with a DS-Qi2 camera was used to image whole leaves for GFP (excitation at 470/40 nm and emission at 525/50 nm), pseudocolored green, and chlorophyll autofluorescence pseudocolored red (excitation 620/60 nm and emission at 700/75 nm). A 2× lens was used to image fluorescent data and the resulting images were stitched together with NIS-Elements (Version 5.02, Nikon Instruments, Inc., Tokyo, Japan).

NSs Tagging and Mutational Inactivation
A 6HIS3xFlag tag (HF) was added at the C terminus of TSWV NSs (NSs-HF) and expressed from the 35S promoter in pMDC32 [42] ( Figure 1A). Silencing suppression was tested using a single-stranded GFP transgene in wt N. benthamiana following a standard assay [47]. At 4 days post infiltration, pictures were taken under UV light and GFP fluorescence was estimated using the green channel in ImageJ [8,33]. GFP fluorescence in the presence of NSs-HF was as bright as in the presence of wt NSs. Consistent with previous results for TSWV NSs [34], for both NSs and NSs-HF, GFP fluorescence was 60% ± 10% of that observed in the presence of HA-tagged tombusviral p19 (p19-HA) [15] used as the positive control ( Figure 1B). Thus, the 6HIS3xFlag tag did not affect transgene silencing suppression activity of NSs.
GKT motif and the YL motif are necessary for transgene silencing suppression [44]. To inactivate NSs-HF, S48A, and R51A, mutations were introduced into the RNA-binding domain one [38]. In a second clone, inactivating mutations K182A (NSs-1) and L413A (NSs-2) were introduced in the GKT motif and in the YL motif (NSs-3) [44], respectively ( Figure 1A). Transient assays, as described above, showed that both mutants lost transgene silencing suppression activity ( Figure 1C). Thus, herein these clones are referred to as suppressor-deficient NSs-HF. In contrast, wt NSs-HF suppressed silencing and GFP protein accumulated to 50% ± 12% of that observed in the presence of p19-HA ( Figure 1C). Using an anti-Flag antibody, NSs-HF was detected in infiltrated leaves. In contrast, suppression-deficient NSs-HF mutants were below detection ( Figure 1C). To verify the expression of suppressor deficient NSs-HF clones, their accumulation was determined in the presence and in the absence of p19-HA. Mutant NSs-HF proteins were only detected in the presence of p19-HA and accumulated to only 8% of wt NSs-HF ( Figure 1C). Low accumulation of suppressor-deficient mutant NSs-HF was due to reduced protein stability (see below).  ssGFP was infiltrated alone or in combination with NSs in wt N. benthamiana leaves. Empty vector and HA-tagged protein p19 from tomato bushi stunt virus (p19-HA) were used as negative and positive controls, respectively. Four days after infiltration, GFP fluorescence was visualized and photographs taken under ultraviolet light. The histogram shows average GFP fluorescence ± standard error (18 leaves from three repetitions) relative to leaves infiltrated with p19-HA; (C) Silencing suppression and accumulation of NSs-HF proteins. The experiment was as in (B) plus a control infiltrated with buffer. Representative western blots showing NSs-HF and GFP accumulation. Rubisco was used as loading control. The histogram shows average GFP accumulation ± standard error (four biological replicates) relative to that observed in leaves infiltrated with p19-HA. The blot on the right shows NSs-HF accumulation in the presence (+) and absence (−) of p19-HA. In (B,C) bars with the same letter are not significantly different (Tukey's test with α = 0.05). Scale bars: 1 cm.
In TSWV, the NSs RNA-binding domain one is necessary for transgene silencing suppression, induction of the hypersensitive response [38], and consists of amino acids 45 to 57 [38]. Similarly, the GKT motif and the YL motif are necessary for transgene silencing suppression [44]. To inactivate NSs-HF, S48A, and R51A, mutations were introduced into the RNA-binding domain one [38]. In a second clone, inactivating mutations K182A (NSs-1) and L413A (NSs-2) were introduced in the GKT motif and in the YL motif (NSs-3) [44], respectively ( Figure 1A). Transient assays, as described above, showed that both mutants lost transgene silencing suppression activity ( Figure 1C). Thus, herein these clones are referred to as suppressor-deficient NSs-HF. In contrast, wt NSs-HF suppressed silencing and GFP protein accumulated to 50% ± 12% of that observed in the presence of p19-HA ( Figure 1C). Using an anti-Flag antibody, NSs-HF was detected in infiltrated leaves. In contrast, suppression-deficient NSs-HF mutants were below detection ( Figure 1C). To verify the expression of suppressor deficient NSs-HF clones, their accumulation was determined in the presence and in the absence of p19-HA. Mutant NSs-HF proteins were only detected in the presence of p19-HA and accumulated to only 8% of wt NSs-HF ( Figure 1C). Low accumulation of suppressor-deficient mutant NSs-HF was due to reduced protein stability (see below).

NSs Supports Systemic Movement of Suppressor-Deficient TuMV-AS9-GFP
When expressed in trans, NSs-HF suppressed local transgene silencing ( Figure 1) and wt NSs suppressed local antiviral RNA silencing, leading to the formation of local infection foci [8]. We hypothesized that NSs supports viral systemic movement. To test this hypothesis, we expressed NSs-HF in cis from TuMV-AS9-GFP, which cannot infect wt N. benthamiana or wt A. thaliana due to an inactivation mutation (AS9) in the silencing suppressor HC-Pro [3]. NSs-HF or NSs-HF-S48A-R51A with inactivating mutations ( Figure 1A) were placed between NIb and the coat protein in a TuMV-AS9-GFP infectious clone ( Figure 2A). Virus infection was launched by agroinfiltration [48] in N. benthamiana and wt A. thaliana. Empty vector and TuMV-AS9-GFP were used as the negative controls while TuMV-GFP [3] was used as a positive control. Virus movement and the establishment of systemic infection was tracked for 15 days based on GFP fluorescence under UV light and confirmed by Western blotting [3]. In both N. benthamiana and wt A. thaliana, NSs-HF restored pathogenicity to suppressor-deficient TuMV-AS9-GFP ( Figures 2B and 3A) and supported systemic virus movement through the entire plant (Table 1). In contrast, chimeric virus expressing suppressor-deficient NSs-HF-S48A-R51A failed to establish local and systemic infection ( Figures 2B and 3A,B and Table 1). NSs-HF-S48A-R51A has inactivating mutations in the RNA-binding domain one [38] (Figure 1). These contrasting differences show that silencing the suppression by NSs-HF was necessary and sufficient to restore establishment of infection and promote systemic movement of suppressor-deficient TuMV-AS9-GFP. This effect requires NSs RNA-binding domain one.     0  0  0  TuMV-AS9-GFP  36  0  0  0  0  TuMV-AS9-GFP-NSs-HF  36  0  0  5  36  TuMV-AS9-GFP-NSs-S48A-R51A-HF  36  0  0  0  0  TuMV-GFP  36  4  23  34   represent average (±standard error) of three repetitions, each consisting of 12 plants per treatment.
(D) Virus accumulation in systemically infected leaves at 10 dpi. Representative western blots showing NSs-HF and coat protein accumulation in the same sample. HSP70 was used as a loading control. The blot was cut in three parts before antibody probing. The histogram shows average coat protein accumulation ± standard error (four biological replicates) relative to TuMV-GFP. Bars with the same letter are not significantly different (Tukey's test with α = 0.05).  In both N. benthamiana and wt A. thaliana, systemic infection of TuMV-AS9-GFP expressing NSs-HF was slower than that observed for TuMV-GFP ( Figure 2B,C and Figure 3C), which could be due to the loss of the NSs insert or to replication or movement defects caused by the presence of NSs-HF in the TuMV genome. To determine the stability of the NSs-HF insert, systemically infected leaf samples were collected from individual N. benthamiana plants at 10 dpi and total protein was extracted for Western blotting. Anti-coat protein (CP) and anti-Flag antibodies were used to probe for both TuMV coat protein and NSs-HF, respectively. In all samples, both coat protein and NSs-HF were detected. Thus, the NSs-HF insert was maintained ( Figure 2D). Coat protein accumulation of TuMV-AS9-GFP expressing NSs-HF was only 20% of that observed for wt TuMV-GFP (Figure 2D), suggesting a defect caused by the NSs-HF insert.
Virus accumulation in systemically infected inflorescence of dcl2-1 dcl3-1 dcl4-2 triple mutants was determined by Western blot in samples collected from individual plants at 15 dpi. Accumulation of TuMV coat protein and NSs-HF in the same sample was determined as described above. Consistent with results obtained in systemically infected N. benthamiana ( Figure 2D), in all systemically infected A. thaliana inflorescence samples, both coat protein and NSs-HF were detected ( Figure 3E). Thus, the NSs-HF insert was maintained after systemic infection of N. benthamiana and A. thaliana. Coat protein accumulated to similar levels for TuMV-AS9-GFP and chimeric viruses expressing NSs-HF or suppressor-deficient NSsHF-S48A-R51A ( Figure 3E). However, NSsHF-S48A-R51A protein accumulated to only 20% of NSs-HF. Potyviruses form a polyprotein that is processed by viral proteases to produce the same number of individual mature proteins [50]. Reduced accumulation in the presence of similar amounts of coat protein made from the polyprotein shows that NSsHF-S48A-R51A protein is less stable than NSs-HF. Accordingly, NSsHF-S48A-R51A is both unstable and inactive in silencing suppression, and the RNA-binding domain one determines both features. Several other NSs mutations that abolished silencing suppression also caused reduced protein accumulation [33]. These observations suggest that NSs silencing suppression activity and stability are linked.
In dcl2-1 dcl3-1 dcl4-2 triple mutant plants, TuMV-GFP and suppressor-deficient TuMV-AS9-GFP established systemic infection of inflorescence at a similar rate ( Figure 3D). dcl2-1 dcl3-1 dcl4-2 triple mutants cannot establish a silencing response to fight virus infection [3,18,49]. Accordingly, in dcl2-1 dcl3-1 dcl4-2 triple mutants, systemic virus movement is independent of silencing suppression. Systemic movement by TuMV-AS9-GFP was similar when expressing active NSs-HF or suppression-deficient NSs-HF-S48A-R51A ( Figure 3D). However, clones carrying an NSs insert established systemic infection slower than those not expressing NSs ( Figure 3D). Thus, the NSs insert caused a slow systemic movement phenotype that was independent of silencing suppression. This could be related to alterations in RNA secondary structures, replication, encapsidation efficiency, polyprotein processing, or a combination. In potyviruses, codon usage bias does not explain translational selection [51] and TSWV infects both N. benthamiana and A. thaliana [52]. Accordingly, changes in codon usage due to the NSs insert are unlikely to affect translation.

NSs Is a Symptom Determinant
N. benthamiana plants systemically infected by chimeric TuMV-AS9-GFP-NSs-HF showed mild symptoms that correlated with a slow establishment of systemic infection ( Figure 2B,C). At 10 dpi, systemically infected leaves showed green and chlorotic spotted mosaic similar to that caused by TSWV [8], and different from TuMV symptoms [3]. These observations suggest that NSs is a symptom determinant. To test this hypothesis, N. benthamiana plants were inoculated the same day with TuMV-AS9-GFP, TuMV-AS9-GFP-NSs-HF, TSWV, or TuMV-GFP. Establishment of systemic infection was monitored for 15 dpi. Symptoms induced by TuMV-AS9-GFP-NSs-HF were similar to those induced by TSWV and less severe than those induced by TuMV-GFP ( Figure 4A). Symptoms induced by TuMV-GFP were more severe than those induced by TSWV. The difference was particularly clear at 7 dpi and faded at 15 dpi ( Figure 4A).

NSs Silencing Suppression Activity Partially Overlaps that of HC-Pro
In TuMV, HC-Pro and VPg are RNA silencing suppressors and determinants of symptom development [2,21]. The results described above showed that TSWV NSs is an RNA silencing suppressor that promotes systemic virus movement and symptom development (Figures 2 and 4). To gain insight into the mechanistic activity of NSs, we did an epistasis analysis by expressing both  In plants inoculated with TuMV-AS9-GFP-NSs-HF, systemically infected leaves showed a mosaic of green and chlorotic spots similar to those caused by TSWV, which were not present in plants inoculated with TuMV-GFP ( Figure 4B). Examination under UV and visible light showed that infection by TuMV-GFP progressed from the base to the top of the leaf and covered most of the leaf surface by 7 dpi (Figures 2B and 4C). In contrast, TuMV-AS9-GFP-NSs-HF infection was localized to spots and did not cover the entire leaf during the 15 days of the experiment (Figures 2B and 4C). Similarly, infection by TSWV induced a mosaic of green and chlorotic spots. Some of the chlorotic spots had localized necrosis that resulted in autofluorescence under UV light ( Figure 4C) that was not due to GFP resulting from infection ( Figure 4D). Confocal microscopy analysis of non-inoculated leaves showed that TuMV-AS9-GFP-NSs-HF caused chlorotic spots that overlapped with GFP fluorescence derived from systemic infection, while normal green areas did not show signs of virus infection ( Figure 4D). Accordingly, in the mosaic of green and chlorotic spots, TuMV-AS9-GFP-NSs-HF infection induced a change in color in an NSs-dependent manner.
Infection by TuMV-AS9-GFP-NSs-HF was distributed in spots and GFP fluorescence was more intense in the mesophyll than in the veins. In contrast, in plants inoculated with TuMV-GFP, non-inoculated leaves were uniformly covered by virus infection from the bottom to the tip and GFP fluorescence was more intense in the veins than in the mesophyll ( Figure 4D). Thus, individually, NSs and HC-Pro determined the pattern of systemic infection and spatial virus distribution in the leaves.
In A. thaliana, symptoms caused by TuMV-GFP were more severe than those observed for TuMV-AS9-GFP-NSs-HF ( Figure 3B). Symptoms caused by TuMV-AS9-GFP-NSs-HF were more severe than those caused by TuMV-AS9-GFP expressing suppressor-deficient NSs-HF ( Figure 3B). Thus, in A. thaliana, as in N. benthamiana, NSs promoted systemic infection and symptom development. Both effects are dependent on NSs silencing suppression.
These results show that silencing suppression by NSs restored pathogenicity to suppressor-deficient TuMV-AS9-GFP, promoting local, systemic infection, and symptom development following a pattern similar to that of TSWV.

NSs Silencing Suppression Activity Partially Overlaps that of HC-Pro
In TuMV, HC-Pro and VPg are RNA silencing suppressors and determinants of symptom development [2,21]. The results described above showed that TSWV NSs is an RNA silencing suppressor that promotes systemic virus movement and symptom development (Figures 2 and 4). To gain insight into the mechanistic activity of NSs, we did an epistasis analysis by expressing both HC-Pro and NSs-HF from the same viral construct. If HC-Pro and NSs have similar mechanistic activity and one suppressor is dominant, their combined effect will be similar to that of a single infection. In the absence of dominance, suppressors could compete for a limiting substrate, interfering with each other, and resulting in mild symptoms. In contrast, if HC-Pro and NSs have different mechanistic activities, their cumulative effect would be synergistic and would cause more severe symptoms than that of a single infection. To test this model, N. benthamiana ( Figure 5A) and A. thaliana ( Figure 6A) plants were inoculated with chimeric TuMV-GFP expressing NSs-HF or suppressor-deficient NSsHF-S48A-R51A (Figure 2A). Establishment of systemic infection was monitored for 15 dpi and virus accumulation in systemically infected tissue was determined at 10 dpi. TuMV-GFP expressing NSs-HF or suppressor-deficient NSsHF-S48A-R51A established systemic infection at a similar rate and slower than parental TuMV-GFP in both N. benthamiana ( Figure 5B,C) and A. thaliana (Figure 6) plants. Additionally, symptoms were less severe in plants infected with chimeric TuMV expressing NSs than those caused by parental TuMV-GFP (Figures 5A and 6A). In systemically infected N. benthamiana leaves, TuMV-GFP expressing active or suppressor-deficient NSs-HF accumulated to 50% of parental TuMV-GFP ( Figure 5D). These differences might be related to the slow movement phenotype caused by NSs when expressed from the TuMV genome ( Figure 3D).   Figure 4C). This is in contrast with the spotted distribution in leaves systemically infected by TuMV-AS9-GFP expressing NSs-HF (Figures 2A and 4C). Accordingly, even in the presence of NSs, spatial distribution of TuMV-GFP in systemically infected tissue was determined by HC-Pro.

Discussion
Transgene silencing suppression activity has been demonstrated for NSs of several orthotospoviruses, including TSWV [33,34], impatiens necrotic spot virus [54], groundnut ringspot virus [33], tomato yellow ring virus [33], capsicum chlorosis virus [55], and groundnut bud necrosis virus [56]. NSs blocks the spread of the transgene silencing signal [33], prevents local antiviral RNA silencing [8,55], is required for persistent infection of and transmission by thrips [36], is required for systemic movement in plants [36], and in mixed infections of iris yellow spot virus and TSWV result in synergism, leading to enhanced virus accumulation and movement in Datura stramonium [57]. However, the mechanism of RNA silencing suppression by orthotospoviral NSs and its role in virus infection have not been determined.
The core genetic components and pathway of plant antiviral RNA silencing have been determined using DNA and positive-strand RNA viruses and experimental model plants such as N. benthamiana and A. thaliana [3,18,45,49,[58][59][60][61]. Less is known about negative-strand and ambisense RNA viruses, mainly due to the lack of genetically tractable hosts and, in the case of orthotospoviruses, due to the lack of infectious clones. To gain insight into the mechanism of RNA silencing suppression by NSs and its effect on virus infection, we used suppressor-deficient TuMV-AS9-GFP, N. benthamiana, and A. thaliana to create a genetically tractable system. Potato virus X and tobacco mosaic virus have been used as vectors to test viral silencing suppressors [62,63]. These systems, and the TuMV vector described here, might cause unexpected effects. The genetic determinants of antiviral RNA silencing against TuMV have been determined and several A. thaliana mutants have been characterized [2,3], which allow for the identification of unexpected effects that are independent from silencing suppression. Indeed, a genetic analysis ( Figure 3D) showed that the TuMV vector described here has a slow systemic movement phenotype caused by the NSs insert and that is independent from silencing suppression.
A 6HIS-3xFlag (HF) tag was added at the C terminus of TSWV NSs ( Figure 1) and a tagged NSs-HF was placed between NIb and the coat protein in a TuMV-AS9-GFP infectious clone (Figure 2A). Collectively, these results show that symptoms caused by TuMV-GFP and the spatial distribution of virus infection are similar when expressing HC-Pro in combination with active or suppressor-deficient NSs, suggesting that the mechanisms used to suppress antiviral silencing by HC-Pro and NSs are similar and that HC-Pro is dominant over NSs.

Discussion
Transgene silencing suppression activity has been demonstrated for NSs of several orthotospoviruses, including TSWV [33,34], impatiens necrotic spot virus [54], groundnut ringspot virus [33], tomato yellow ring virus [33], capsicum chlorosis virus [55], and groundnut bud necrosis virus [56]. NSs blocks the spread of the transgene silencing signal [33], prevents local antiviral RNA silencing [8,55], is required for persistent infection of and transmission by thrips [36], is required for systemic movement in plants [36], and in mixed infections of iris yellow spot virus and TSWV result in synergism, leading to enhanced virus accumulation and movement in Datura stramonium [57]. However, the mechanism of RNA silencing suppression by orthotospoviral NSs and its role in virus infection have not been determined.
The core genetic components and pathway of plant antiviral RNA silencing have been determined using DNA and positive-strand RNA viruses and experimental model plants such as N. benthamiana and A. thaliana [3,18,45,49,[58][59][60][61]. Less is known about negative-strand and ambisense RNA viruses, mainly due to the lack of genetically tractable hosts and, in the case of orthotospoviruses, due to the lack of infectious clones. To gain insight into the mechanism of RNA silencing suppression by NSs and its effect on virus infection, we used suppressor-deficient TuMV-AS9-GFP, N. benthamiana, and A. thaliana to create a genetically tractable system. Potato virus X and tobacco mosaic virus have been used as vectors to test viral silencing suppressors [62,63]. These systems, and the TuMV vector described here, might cause unexpected effects. The genetic determinants of antiviral RNA silencing against TuMV have been determined and several A. thaliana mutants have been characterized [2,3], which allow for the identification of unexpected effects that are independent from silencing suppression. Indeed, a genetic analysis ( Figure 3D) showed that the TuMV vector described here has a slow systemic movement phenotype caused by the NSs insert and that is independent from silencing suppression.
A 6HIS-3xFlag (HF) tag was added at the C terminus of TSWV NSs ( Figure 1) and a tagged NSs-HF was placed between NIb and the coat protein in a TuMV-AS9-GFP infectious clone (Figure 2A). The NSs insert was maintained in both N. benthamiana and A. thaliana during a single infection cycle initiated by agroinfiltration (Figures 2D and 3E). Because both host species are amenable to agroinfiltration, genetic stability of the NSs insert after several passages by mechanical inoculation is not necessary and was not determined.
Although the chimeric virus moved at a lower rate ( Figure 2C, Figure 3C, Figure 5C, and Figure 6B), NSs-HF supported the establishment of infection and the systemic movement of suppressor-deficient TuMV-AS9-GFP in N. benthamiana and A. thaliana ( Figures 2B and 3A). In contrast, NSs-HF with inactivating mutations in the RNA-binding domain one failed to promote systemic infection in N. benthamiana and wt A. thaliana ( Figures 2B and 3A). The lack of infection was due to the lack of silencing suppression because the virus harboring NSs-HF with inactivating mutations systemically infected dcl2-1 dcl3-1 dcl4-2 A. thaliana triple mutants compromised in antiviral RNA silencing ( Figure 3A).
RNA silencing against TuMV is dependent on DCL4 and DCL2 [3]. Accordingly, the gain in pathogenicity in suppression-deficient TuMV-AS9-GFP was dependent on NSs silencing suppression activity ( Figure 2B) to overcome the antiviral defense mediated by DCL4 and DCL2. DCL4 and DCL2 process ds viral RNA to form small virus-derived siRNAs [3,49] that associate with AGO proteins to target single-strand viral RNA [2,11,13,64]. TSWV-infected plants and thrips [65] accumulate TSWV-derived siRNAs [66][67][68]. Accordingly, NSs does not prevent the biogenesis of virus-derived siRNAs. However, TSWV isolates harboring a suppressor-defective NSs protein accumulated virus-derived siRNAs to higher levels than isolates harboring a functional NSs, and the virus-derived siRNAs had a different genomic distribution [68]. These observations suggest that NSs acts downstream of virus-derived formation by DCL proteins, possibly at the association with AGO proteins, RNA silencing amplification, or both. Systemic infection of plants by RNA viruses is restricted in a spatial manner by AGO and RDR proteins [2,3,11]. AGO2 and AGO5 restrict infection in leaves, while AGO1 and AGO10 protect inflorescence from systemic infection by TuMV [2]. Similarly, the redundant activity of RDR1 and RDR6 protect inflorescence from systemic infection by suppressor-deficient TuMV-AS9-GFP [3]. Accordingly, infection of rdr1-1, rdr6-15, and ago2-1 single-mutant plants by suppressor-deficient TuMV-AS9-GFP is restricted to leaves while inflorescences are not infected [2]. However, TuMV-AS9-GFP expressing NSs-HF established systemic infection of inflorescence in wt A. thaliana plants and in rdr1-1, rdr6-15, and ago2-1 single-mutant plants (Tables 1 and 2). Thus, NSs interferes with the antiviral activity of AGO and RDR proteins.
In addition to HC-Pro, in TuMV, VPg is a suppressor of RNA silencing that causes degradation of SGS3 and RD6, which are involved in RNA silencing amplification [21]. In TuMV-AS9-GFP, VPg is not enough to promote infection of N. benthamiana or wt A. thaliana [3]. When expressed in cis from TuMV-AS9-GFP, NSs restored pathogenicity and systemic movement both in N. benthamiana and in A. thaliana (Figures 2 and 3). Accordingly, NSs can functionally replace potyviral HC-Pro.
In plants infected by TuMV-AS9-GFP expressing NSs, virus special distribution co-localized with chlorotic spots typical of TSWV infection symptoms (Figure 4), instead of those induced by TuMV (Figure 4), suggesting that NSs is a symptom determinant. Consistent with these observations, transgenic plants expressing groundnut bud necrosis virus (GBNV) NSs showed symptoms similar to those infected by GBNV [56]. Symptoms in plants infected by TuMV-GFP expressing NSs resembled symptoms induced by TuMV ( Figure 5A), and systemic infection followed a pattern similar to that of TuMV-GFP ( Figure 5B). Accordingly, HC-Pro is dominant over NSs.
Potyviral HC-Pro suppresses antiviral RNA silencing by sequestering small RNAs and preventing their association with AGO proteins [2,17], while CMV 2a targets the initiation step [9]. Consistent with this model, potyviral HC-Pro reactivated expression of a GUS transgene [69], and transgenic GFP silenced plant regained GFP expression upon infection with TSWV or PVY while plants infected with CMV did not regain GFP expression [35]. NSs of TSWV, INSV, GRSV, and TYRV bind small interfering RNAs and miRNA/miRNA* duplexes in vitro [54]. Similarly, potyviral HC-Pro binds several miRNAs and miRNAs* [2,[15][16][17]. These observations suggest that NSs suppresses RNA silencing by a mechanism that includes binding of virus-derived siRNAs and host siRNAs. This model is consistent with the presence of three RNA-binding domains in NSs [38].

Conclusions
A genetically tractable system was developed to determine the role of TSWV NSs in silencing suppression, pathogenicity, and virus systemic movement. This system is based on a potyvirus infectious clone, N. benthamiana, and A. thaliana, and is amenable to modifications to study any protein of interest. Using this system, a genetic analysis showed that NSs can functionally replace potyviral HC-Pro, and is genetically stable when expressed in cis from TuMV. NSs suppresses an antiviral defense mechanism that is dependent on DCL, RDR, and AGO proteins to condition virus susceptibility, promote virus infection, systemic movement, and symptom development.