Identification of the Potential Virulence Factors and RNA Silencing Suppressors of Mulberry Mosaic Dwarf-Associated Geminivirus

Plant viruses encode virulence factors or RNA silencing suppressors to reprogram plant cellular processes or to fine-tune host RNA silencing-mediated defense responses. In a previous study, Mulberry mosaic dwarf-associated virus (MMDaV), a novel, highly divergent geminivirus, has been identified from a Chinese mulberry tree showing mosaic and dwarfing symptoms, but the functions of its encoded proteins are unknown. In this study, all seven proteins encoded by MMDaV were screened for potential virulence and RNA silencing suppressor activities. We found that V2, RepA, and Rep affect the pathogenicity of a heterologous potato virus X. We showed that V2 could inhibit local RNA silencing and long-distance movement of the RNA silencing signal, but not short-range spread of the green fluorescent protein (GFP) silencing signal in Nicotiana benthamiana 16c plants. In addition, V2 localized to both subnuclear foci and the cytoplasm. Deletion mutagenesis of V2 showed that the basic motif from amino acids 61 to 76 was crucial for V2 to form subnuclear foci and for suppression of RNA silencing. Although the V2 protein encoded by begomoviruses or a curtovirus has been shown to have silencing suppressor activity, this is the first identification of an RNA silencing suppressor from a woody plant-infecting geminivirus.


Introduction
Viruses are intracellular obligate parasites that absolutely depend on the host machinery for their replication and movement. To defend themselves against invading viruses, plants employ several layers of immune responses [1][2][3][4][5]. RNA silencing, a fundamental sequence-specific gene regulatory process, has been demonstrated to be one of the major antiviral defense mechanisms in plants [3,6]. Briefly, viral double-stranded RNA (dsRNA) molecules of different origins, namely highly structured regions of viral single-stranded RNAs, replicative intermediates, or overlapping bidirectional read-through transcripts can trigger RNA silencing, and these are processed into 21-24 nucleotides (nt) small interfering RNAs (siRNAs) by Dicer-like RNases (DCLs). The siRNAs are stabilized through For the PTGS suppression assay, the full-length ORF of V1 was subcloned into the pCHF3 vector [30] between the SmaI and SalI sites to produce pCHF3-V1. The full-length ORFs of V2, V3, RepA, and Rep were individually subcloned into the pCHF3 vector between the BamHI and SalI sites, to yield pCHF3-V2, pCHF3-V3, pCHF3-RepA and pCHF3-Rep. The full-length ORFs of V4 and V5 were individually subcloned into the pCHF3 vector between the KpnI and SalI sites, to obtain pCHF3-V4 and pCHF3-V5. The resulting constructs were individually introduced into the A. tumefaciens strain C58C1 through electroporation.
For subcellular localization analysis, the full-length fragment of MMDaV V2 was inserted into the BamHI and SalI sites of the pCHF3-N-eGFP vector [31] to produce 35S-GFP-V2, which contains a V2 Viruses 2018, 10, 472 4 of 16 N-terminal fusion protein with enhanced green fluorescent protein (eGFP). The resulting plasmid was mobilized into A. tumefaciens strain C58C1 via electroporation.
To generate the V2 mutant dm61-77aa, the 61RRLLRLIRRFSRVKDR76 motif was deleted from the plasmid pGEM-T-V2, using KOD-Plus-Mutagenesis Kit as instructed (Toyobo, Osaka, Japan). The resulting V2 dm61-77aa was inserted into the BamHI and SalI sites of pCHF3 or pCHF3-N-eGFP to produce pCHF3-V2 dm61-77aa and 35S-GFP-V2 dm61-77aa , respectively. All primers used for the generation of the DNA constructs described above are shown in Table 1.

Agrobacterium-Mediated Virus Inoculation and Transient Gene Expression
Agrobacterium-mediated virus inoculation or transient gene expression were carried out as described [32]. Agrobacterium cultures were pelleted and resuspended to an optical density OD 600 = 1.0 in a solution containing 10 mM MgCl 2 , 10 mM MES (pH 5.8), and 100 µM acetosyringone, and were incubated at room temperature for 2-3 h prior to infiltration.
For the recombinant PVX vectors expressing individual ORFs of MMDaV, cultures of A. tumefaciens harboring different constructs were individually infiltrated into N. benthamiana leaves at an OD 600 = 1.0.
To test silencing suppression activity, equal volumes of Agrobacterium cultures harboring 35S-GFP (a 35S promoter-driven construct expressing sense GFP, a kind gift from David C. Baulcombe, University of Cambridge, Cambridge, UK) and the tested constructs were mixed, followed by infiltration into fully expanded leaves of four-week-old 16c or wild type N. benthamiana plants.
Co-infiltration of 35S-GFP with a construct to express tomato bushy stunt virus (TBSV) P19 was used as a positive control, and that with the empty vector of pCHF3 was used as a negative control. For dsRNA-induced PTGS experiments, A. tumefaciens cultures containing the 35S-GFP, 35S-dsGFP (a construct expressing an inverted repeat sequence of GFP) [33], together with either pCHF3-V2, or P19, or the empty pCHF3 vector, were mixed in equal proportions and infiltrated into N. benthamiana 16c plant leaves. GFP fluorescence in infiltrated or systemic leaves was monitored under handheld long-wavelength UV lamp (Black-Ray Model B-100A, San Gabriel, CA, USA) and photographed with a Canon EOS 70D digital camera mounted with a 58 mm yellow filter.

RNA Extraction and Analysis
Total RNA was extracted using TRIzol reagent following the manufacturer's instructions (Invitrogen, Carlsbad, CA, USA). For the N. benthamiana plants inoculated with PVX or recombinant PVX constructs, systemically infected plant leaves were harvested and prepared to analyze MMDaV ORF expression by reverse transcription PCR (RT-PCR). 1 µg of total RNA was reverse transcribed into cDNA using a PrimeScript RT reagent Kit with genomic DNA Eraser (Takara, Dalian, China). Expression of the individual MMDaV ORFs was detected by PCR using primers listed in Table 1. For GFP messenger RNA (mRNA) analysis, 10 µg of total RNA extracted from infiltrated patches of N. benthamiana line 16c plants was separated on 1.2% formaldehyde-agarose gels and transferred to Hybond N+ membranes as instructed (GE Healthcare, Buckinghamshire, UK). The membrane was hybridized with a digoxigenin-labeled GFP probe, which was made using the PCR DIG probe synthesis kit, and was detected using a detection starter kit II according to the manufacturer's instructions (Roche Diagnostic, Basel, Switzerland).

Protein Extraction and Western Blot Analysis
Extraction of total soluble proteins, SDS-PAGE, and Western blot analysis were performed as described [31]. For detection of PVX, proteins were extracted from systemically infected leaves of N. benthamiana plants infected with PVX or PVX recombinant constructs. The anti-CP monoclonal antibody (MAb) raised against PVX (raised in our lab, Institute of Biotechnology, Zhejiang Univiersity, Hangzhou, China) was used at a 1:8000 dilution. For detection of GFP, proteins were extracted from Viruses 2018, 10, 472 5 of 16 infiltrated patches of N. benthamiana line 16c or N. benthamiana plants. The anti-GFP MAb (Roche) was used at a 1:8000 dilution. Western blots were visualised with a secondary peroxidase-conjugated goat antimouse antibody (Cell Signaling Technology, Boston, MA, USA) and a chemiluminescence detection system (Tianneng, Shanghai, China).

Laser-Scanning Confocal Microscopy
Imaging of fluorescent proteins was conducted using a confocal microscopy (LSM880; Carl Zeiss, Jena, Germany) at 36 to 48 h post-infiltration. To stain the nuclei of the leaf epidermal cells, 0.1 µg·mL −1 of 4 -6-diamidino-2-phenylindole (DAPI) was infiltrated into N. benthamiana leaves for 10 to 20 min prior to mounting onto slides as described [34]. For GFP, excitation was set at 488 nm, and emission was set at 500 to 530 nm. For DAPI, excitation was set at 405 nm, and emission was set at 440 to 475 nm.

Identification of Virulence Factors Encoded by MMDaV
MMDaV has a monopartite genome with seven ORFs (Figure 1). To determine if any of the seven proteins is related to viral pathogenicity, we took advantage of the PVX-based heterologous gene expression system to express the different MMDaV proteins in planta. Leaves of four-week old N. benthamiana plants were inoculated with Agrobacterium cultures containing PVX or recombinant PVX constructs expressing individual MMDaV ORF and were monitored for symptom development. At seven days post infiltration (dpi), non-inoculated systemic leaves exhibited mosaic symptoms characteristic of PVX infection in N. benthamiana plants inoculated with PVX, PVX-V1, PVX-V3, PVX-V4, and PVX-V5, respectively. By 10-12 dpi, N. benthamiana plants inoculated with PVX-V4 and PVX-V5 developed milder mosaic symptoms ( Figure 2A). However, PVX expressing V2 developed necrosis symptoms in inoculated leaves at 5 dpi, followed by downward leaf curling symptoms in emerging leaves at 7 dpi, and death of the apical shoots and eventual plant death at 11 dpi ( Figure 2A). N. benthamiana plants inoculated with PVX-RepA showed necrotic lesions in the inoculated leaves at 4 dpi, followed by downward leaf curling at 6 dpi. Collapse of developing tissue was also observed in emerging leaves inoculated with PVX-RepA at 10 dpi ( Figure 2A). PVX-Rep induced upward leaf curling at 6 dpi, followed by the collapse of developing leaves at 10 dpi ( Figure 2A). peroxidase-conjugated goat antimouse antibody (Cell Signaling Technology, Boston, MA, USA) and a chemiluminescence detection system (Tianneng, Shanghai, China).

Laser-Scanning Confocal Microscopy
Imaging of fluorescent proteins was conducted using a confocal microscopy (LSM880; Carl Zeiss, Jena, Germany) at 36 to 48 h post-infiltration. To stain the nuclei of the leaf epidermal cells, 0.1 μg . mL −1 of 4′-6-diamidino-2-phenylindole (DAPI) was infiltrated into N. benthamiana leaves for 10 to 20 min prior to mounting onto slides as described [34]. For GFP, excitation was set at 488 nm, and emission was set at 500 to 530 nm. For DAPI, excitation was set at 405 nm, and emission was set at 440 to 475 nm.

Identification of Virulence Factors Encoded by MMDaV
MMDaV has a monopartite genome with seven ORFs (Figure 1). To determine if any of the seven proteins is related to viral pathogenicity, we took advantage of the PVX-based heterologous gene expression system to express the different MMDaV proteins in planta. Leaves of four-week old N. benthamiana plants were inoculated with Agrobacterium cultures containing PVX or recombinant PVX constructs expressing individual MMDaV ORF and were monitored for symptom development. At seven days post infiltration (dpi), non-inoculated systemic leaves exhibited mosaic symptoms characteristic of PVX infection in N. benthamiana plants inoculated with PVX, PVX-V1, PVX-V3, PVX-V4, and PVX-V5, respectively. By 10-12 dpi, N. benthamiana plants inoculated with PVX-V4 and PVX-V5 developed milder mosaic symptoms ( Figure 2A). However, PVX expressing V2 developed necrosis symptoms in inoculated leaves at 5 dpi, followed by downward leaf curling symptoms in emerging leaves at 7 dpi, and death of the apical shoots and eventual plant death at 11 dpi ( Figure 2A). N. benthamiana plants inoculated with PVX-RepA showed necrotic lesions in the inoculated leaves at 4 dpi, followed by downward leaf curling at 6 dpi. Collapse of developing tissue was also observed in emerging leaves inoculated with PVX-RepA at 10 dpi ( Figure 2A). PVX-Rep induced upward leaf curling at 6 dpi, followed by the collapse of developing leaves at 10 dpi ( Figure 2A).   To confirm the stability of the inserts, RT-PCR was performed on total RNA extracted from non-inoculated systemic leaves at 7-10 dpi by using specific primers targeting the individual MMDaV ORFs. RT-PCR amplification on each sample obtained expected fragments from N. benthamiana plants infected with recombinant PVX constructs, but not from plants inoculated with PVX, suggesting the maintenance of the MMDaV sequences in the systemically infected leaves. Western blot analysis of the proteins extracted from systemically infected leaves using antibody against PVX CP confirmed PVX infection in all of the inoculated N. benthamiana plants ( Figure 2B). Furthermore, the presence of V2, RepA, and Rep decreased the accumulation of PVX in non-inoculated systemic leaves. Although the expression of V4 and V5 proteins did not enhance the pathogenicity of PVX, their expression also decreased PVX accumulation ( Figure 2B). In summary, the V2, RepA, and Rep proteins of MMDaV enhance the infection severity of a heterologous PVX. To confirm the stability of the inserts, RT-PCR was performed on total RNA extracted from non-inoculated systemic leaves at 7-10 dpi by using specific primers targeting the individual MMDaV ORFs. RT-PCR amplification on each sample obtained expected fragments from N. benthamiana plants infected with recombinant PVX constructs, but not from plants inoculated with PVX, suggesting the maintenance of the MMDaV sequences in the systemically infected leaves. Western blot analysis of the proteins extracted from systemically infected leaves using antibody against PVX CP confirmed PVX infection in all of the inoculated N. benthamiana plants ( Figure 2B). Furthermore, the presence of V2, RepA, and Rep decreased the accumulation of PVX in non-inoculated systemic leaves. Although the expression of V4 and V5 proteins did not enhance the pathogenicity of PVX, their expression also decreased PVX accumulation ( Figure 2B). In summary, the V2, RepA, and Rep proteins of MMDaV enhance the infection severity of a heterologous PVX.

Identification of Suppressors of Local RNA Silencing
Discovery of MMDaV by small RNA-based deep sequencing [25] suggested the pivotal role of RNA silencing in targeting MMDaV in mulberry trees. To test whether any of the MMDaV ORFs is able to counteract the host RNA silencing response, a classical GFP-based two-component agroinfiltration assay was used to screen potential VSRs. In this system, sequences encoding the seven ORFs were individually cloned into a binary vector (pCHF3) under the control of a 35S promoter. A mixture of Agrobacterium containing 35S-GFP and plasmids expressing individual ORFs of MMDaV was co-infiltrated into fully expanded leaves of N. benthamiana GFP line 16c plants. As expected, strong GFP fluorescence was evident in all leaf patches agroinfiltrated with a mixture of 35S-GFP, either with pCHF3 empty vector or pCHF3 vectors expressing individual MMDaV ORFs, at 2 to 3 dpi. At 5 dpi, as a consequence of GFP local silencing, GFP fluorescence was completely lost in the patches agroinfiltrated with 35S-GFP, plus the empty pCHF3 vector. As previously described, agroinfiltration of 35S-GFP and TBSV P19, a positive control for silencing suppression, elicited a strong GFP fluorescence in the infiltrated area at 5 dpi. Examination of the GFP fluorescence in leaf patches co-infiltrated with Agrobacterium mixtures containing 35S-GFP and the tested MMDaV ORFs showed that only fluorescence in the case of V2 was comparable with that of P19 at 5 dpi ( Figure 3A), which could last even at 8 dpi ( Figure S1). None of the other ORFs evaluated had a detectable effect on GFP expression ( Figure 3A), nor on the prevention of GFP silencing by 8 dpi ( Figure S1). Interestingly, agroinfiltration of 35S-GFP with the RepA or Rep ORF of MMDaV caused necrotic lesions in infiltrated leaf patches ( Figure 3A and Figure S1), suggesting the role of RepA and Rep in pathogenicity that may be independent of silencing suppressor activity. As the severe necrosis of the patches caused by infiltration of 35S-GFP plus RepA did not allow for the extraction of RNA of good quality, Northern blot analysis was carried out using total RNAs extracted from leaf patches agroinfiltrated with 35S-GFP plus the empty vector, P19, or the other six ORFs (V1, V2, V3, V4, V5, and Rep). Consistent with the observed GFP fluorescence, GFP mRNA accumulated to higher levels in leaves expressing V2 and P19 than in those transformed with the empty vector or the other ORFs ( Figure 3B). Protein gel blots showed that GFP protein was low in leaves co-infiltrated with 35S-GFP together with the empty vector or the other five ORFs. In contrast, higher GFP protein levels were detected in leaves co-infiltrated with 35S-GFP and P19, or with 35S-GFP and V2 ( Figure 3B). Similar transient assays were carried out in wild-type N. benthamiana plants by infiltration with Agrobacterium carrying 35S-GFP and empty vector, or 35S-GFP and V2 or 35S-GFP and P19. As expected, GFP fluorescence faded at 3 dpi in the infiltrated area in control conditions. However, tissues infiltrated with 35S-GFP and V2 or with 35S-GFP and P19 showed strongly increased GFP fluorescence ( Figure 3C). Concurrently, levels of GFP protein determined by Western blot analysis correlated with the visualized GFP fluorescence ( Figure 3D). These results suggest that MMDaV V2 is a potent suppressor of local RNA silencing triggered by sense GFP.
Since dsRNA is considered to be a strong inducer of RNA silencing [35], we examined whether MMDaV V2 was able to suppress dsRNA-induced RNA silencing in N. benthamiana 16c plants. GFP fluorescence was only observed in co-infiltrations that contained 35S-GFP, 35S-dsGFP, and P19 at 3 dpi, but not in leaf patches infiltrated with 35S-GFP, 35S-dsGFP, and empty vector or in those infiltrated with 35S-GFP, 35S-dsGFP, and MMDaV V2 ( Figure S2), suggesting that MMDaV V2 does not suppress dsRNA-induced RNA silencing.

MMDaV V2 Does Not Suppress Cell-to-Cell Movement of Silencing Signal
In plants, after silencing is initiated in a single cell, the silencing signal can spread locally from cell to cell. To investigate whether the MMDaV V2 protein that was identified as a local silencing suppressor can inhibit the cell-to-cell spread of RNA silencing, we monitored GFP fluorescence at the edge of the infiltrated patches in N. benthamiana 16c plants. In this system, movement of the silencing signal from the agroinfiltrated area causes a strong reduction of GFP transgene expression in a zone of 10 to 15 adjacent cells, which can be visualized under UV light as a characteristic red ring around the infiltration area [36]. As expected, when the leaves of N. benthamiana 16c plants were co-infiltrated with 35S-GFP and the empty pCHF3 vector, a significant decrease of GFP expression and an obvious red ring was observed in the cells surrounding the agroinfiltrated patches at 6 dpi ( Figure 4A). In contrast, no red ring was developed around the infiltrated zone expressing 35S-GFP and P19, as previously described ( Figure 4A). Significantly, a visible red ring was observed around the infiltrated area expressing 35S-GFP and MMDaV V2 ( Figure 4A). These results suggest that MMDaV V2 does not suppress the short range (cell-to-cell) spread of RNA silencing.

MMDaV V2 Does Not Suppress Cell-to-Cell Movement of Silencing Signal
In plants, after silencing is initiated in a single cell, the silencing signal can spread locally from cell to cell. To investigate whether the MMDaV V2 protein that was identified as a local silencing suppressor can inhibit the cell-to-cell spread of RNA silencing, we monitored GFP fluorescence at the edge of the infiltrated patches in N. benthamiana 16c plants. In this system, movement of the ( Figure 4A). In contrast, no red ring was developed around the infiltrated zone expressing 35S-GFP and P19, as previously described ( Figure 4A). Significantly, a visible red ring was observed around the infiltrated area expressing 35S-GFP and MMDaV V2 ( Figure 4A). These results suggest that MMDaV V2 does not suppress the short range (cell-to-cell) spread of RNA silencing.

MMDaV V2 Inhibits Systemic Silencing of GFP
In plants, the silencing signal can also move systemically through the phloem. To find out the effect of MMDaV V2 on long-range movement of RNA silencing, GFP fluorescence was monitored in the newly emerging leaves of N. benthamiana 16c plants infiltrated with 35S-GFP and the empty pCHF3 vector, or 35S-GFP and V2, or 35S-GFP and P19. At 20 dpi, red fluorescence was observed in upper uninfiltrated leaves of 35S-GFP and empty vector-infiltrated plants (16 out of 24 plants   Figure 4B). In contrast, upper leaves of N. benthamiana 16c plants infiltrated with 35S-GFP plus MMDaV V2 or P19, maintained GFP fluorescence. As shown in Figure 4B, the efficiency of MMDaV V2 to inhibit systemic silencing was 87.5% (21 of 24 plants showed inhibition of systemic silencing), indicating that MMDaV V2 was able to inhibit long-distance spread of RNA silencing.

The Basic Motif Is Required for MMDaV V2 Subnuclear Foci Localization and PTGS Suppression
To explore which region(s) within MMDaV V2 is crucial for suppressing RNA silencing, we analyzed the MMDaV V2 sequence, and found that the RRLLRLIRRFSRVKDR motif between amino acid 61 and 76 defined a bipartite basic nuclear localization signal (NLS). Previous studies have indicated that the NLS of several VSRs is essential for the RNA silencing suppressor activity [31,[37][38][39], prompting us to determine whether this motif is crucial for V2 to suppress PTGS. At a first step, we determined the subcellular localization of V2 via transient expression of eGFP-fused constructs into N. benthamiana leaves. An examination of leaves expressing eGFP, as a control, by laser scanning confocal microscopy at 36 h postinfiltration showed that both eGFP and eGFP-V2 were localized in the nucleus and in the cytoplasm. A close-up of the images showed that expression of GFP-V2 localized to the subnuclear foci as indicated by DAPI staining ( Figure 5A). Interestingly, the GFP-V2 fusion protein formed one or more discrete fluorescent foci inside the nucleus ( Figure 5A). To determine whether the basic motif of MMDaV V2 is required for its VSR activity, V2 dm61-77aa was cloned into pCHF3 to evaluate its ability to suppress RNA silencing. Leaves monitored under UV light showed that co-expression of 35S-GFP and empty vector or 35S-GFP and V2 dm61-77aa led to a loss of GFP fluorescence at 3 dpi ( Figure 6A). This observation was further confirmed by examining the relative GFP protein level extracted from the corresponding leaf patches (Figure 6B), suggesting To investigate whether the predicted NLS plays a role in protein localization, the 61RRLLRLIRRFSRVKDR76 motif was deleted from V2 and a cassette to express the V2 mutant variant, 35S-GFP-V2 dm61-77aa , was generated. In contrast to the wild-type V2, the fluorescence of 35S-GFP-V2 dm61-77aa was generally distributed throughout the nucleus and the cytoplasm. Despite the observation that deletion of the putative NLS from MMDaV V2 did not prevent localization of the protein in the host nucleus, 35S-GFP-V2 dm61-77aa appeared not to form discrete foci in the plant nuclear ( Figure 5A). To ensure that the inability of the 35S-GFP-V2 dm61-77aa to form the subnuclear foci was not due to destabilization or impaired production of the mutant V2 protein, Western blot analysis was carried out to detect the eGFP fusion proteins using anti-GFP antibody. Detection of the expected protein sizes revealed that the deletion of the 61RRLLRLIRRFSRVKDR76 motif does not destabilize or impair the production of the V2 protein ( Figure 5B). Together, these results suggest that the basic motif of MMDaV V2 contains signals conditioning formation of the subnuclear foci in plant cells.
To determine whether the basic motif of MMDaV V2 is required for its VSR activity, V2 dm61-77aa was cloned into pCHF3 to evaluate its ability to suppress RNA silencing. Leaves monitored under UV light showed that co-expression of 35S-GFP and empty vector or 35S-GFP and V2 dm61-77aa led to a loss of GFP fluorescence at 3 dpi ( Figure 6A). This observation was further confirmed by examining the relative GFP protein level extracted from the corresponding leaf patches ( Figure 6B), suggesting that the 61RRLLRLIRRFSRVKDR76 motif is required for MMDaV V2 to suppress RNA silencing. Taken together, these findings indicate that the 61RRLLRLIRRFSRVKDR76 motif was vital for V2 subnuclear foci localization and RNA silencing suppression.

Discussion
With limited coding capacity, plant viruses of the family Geminiviridae rely extensively on the host machinery for their infection cycles, such as replication and cell-to-cell and systemic movement. Geminiviruses therefore redirect and reprogram multiple plant processes to achieve a successful

Discussion
With limited coding capacity, plant viruses of the family Geminiviridae rely extensively on the host machinery for their infection cycles, such as replication and cell-to-cell and systemic movement. Geminiviruses therefore redirect and reprogram multiple plant processes to achieve a successful infection [27,40]. To combat a geminivirus infection, plants have employed sophisticated defense systems. The RNA silencing machinery, an antiviral defense conserved in plants and other eukaryotic organisms, is implicated as the major effective mean to overcome geminivirus infection [28]. As a counterdefense, geminiviruses evolve to encode silencing suppressors that target different steps of the RNA silencing pathway [27,28,41]. In this study, we screened for the potential virulence factors and RNA silencing suppressors encoded by MMDaV. We report that the V2, RepA, and Rep proteins of MMDaV affect the pathogenicity of PVX. We present evidence that MMDaV V2 is a strong suppressor of PTGS, and that it localizes to subnuclear foci and the cytoplasm. Finally, we show that the motif from amino acids 61 to 76 is crucial for MMDaV V2 to form subnuclear foci in plant cells and to suppress PTGS.
V2 from several Old World begomoviruses has been identified to have an RNA silencing suppressor activity [42][43][44][45][46][47][48]. The V2 protein of tomato yellow leaf curl virus (TYLCV) is proposed to suppress PTGS by direct interaction with suppressor of gene silencing 3 (SGS3) [47], or through competition with SGS3 for dsRNA substrates [49]. The V2 protein encoded by tomato yellow leaf curl China virus was shown to bind siRNAs [42]. In a recent study, V2 from a curtovirus is reported to act as a PTGS suppressor, possibly by hindrance of the RNA-dependent RNA polymerase 6 (RDR6) function [50]. Despite similarity in genome location, the length and amino acid sequence of MMDaV V2 are highly diverse from its begomovirus and curtovirus counterpart. In this study, it is clear from the co-infiltration leaf patch assays performed on N. benthamiana 16c or wild-type N. benthamiana plants that MMDaV V2 suppresses RNA silencing triggered by sense GFP RNA. In spite of the fact that MMDaV V2 could not prevent cell-to-cell spread of the silencing signal, MMDaV V2 could interfere with systemic spread of the silencing signal. As MMDaV V2 does not inhibit local silencing induced by dsRNA, it probably targets the upstream steps of dsRNA production. Sense GFP-triggered silencing requires the conversion to dsRNA by the action of RDR6, in concert with other co-factors, such as SGS3 [51][52][53]. It is not known whether MMDaV V2 suppresses PTGS by impairing the RDR6/SGS3 pathway. Failing in detecting an interaction between MMDaV V2 and SGS3 suggests that there may be mechanistic differences between MMDaV V2 and TYLCV V2 (Xiuling Yang, Dongxue Wang, Xueping Zhou. Unpublished. Institute of Plant Protection, Chinese Academy of Agricultural Sciences, Beijing, China. 2018). The mode of action and molecular targets of this divergent V2 protein remain to be elucidated.
It is interesting to note that, different from the V2 protein of Old World begomoviruses and beet curly top virus, the GFP-tagged MMDaV V2 protein localized both to subnuclear foci and the nucleus. In silico prediction of MMDaV V2 domains showed that it contains a bipartite NLS (RRLLRLIRRFSRVKDR) from amino acids 61 to 76. Despite the fact that deletion of the putative NLS from MMDaV V2 did not prevent translocation of the protein into the plant nucleus, mutagenesis experiment demonstrated that the predicted basic motif of MMDaV V2 is indispensable not only for subnuclear foci localization but also for silencing suppression. In previous studies, nuclear import was found mandatory for several VSRs, such as βC1 of tomato yellow leaf curl China betasatellite and tomato leaf curl China betasatellite, P37 of Pelargonium line pattern virus, and P6 of cauliflower mosaic virus (CaMV) and strawberry vein banding virus, to suppress RNA silencing [31,[37][38][39]54]. Nuclear import of CaMV P6 is conducted via two importin-α-dependent NLSs, and it is required for CaMV infection and suppression of the nuclear RNA silencing factor dsRNA-binding protein 4 (DRB4) [54]. In the case of CMV 2b, the NLS sequence in the Fny2b is coincident with the domain required for sRNA binding [55]. Although earlier studies showed that nuclear enrichment of 2b was required for RNA silencing suppression, latter studies found that small RNA binding activity is required for VSR activity, while nuclear localization is dispensable [55]. The nuclear-cytoplasmic partitioning of the 2b protein allows CMV to regulate the balance between virus accumulation and suppression of RNA silencing [56]. At this moment, we do not know how the subnuclear foci localization of MMDaV V2 influences its VSR activity. Nonetheless, our results demonstrate that silencing suppression is one of several possible nuclear function of MMDaV V2.
VSRs generally have parallel functions. Besides working as silencing suppressors, they may also fulfill other non-silencing suppression tasks during infection. Therefore, VSRs can often differentially affect pathogenicity in heterologous systems. In agreement with previous studies [42,44], expression of MMDaV V2 by PVX induced leaf curling and a HR-like cell death symptoms that were distinct from the symptoms induced by PVX infection, suggesting that the MMDaV V2 protein might play some role in pathogenicity or virulence. Although the heterologous expression of MMDaV V2 led to an enhancement of symptoms leading to plant death at 11 dpi, decreased PVX accumulation was observed in non-inoculated N. benthamiana leaves when compared to those inoculated with PVX. It is generally thought that due to the suppression of antiviral defenses by VSRs, chimerical PVXs carrying heterologous VSRs accumulate at higher levels when compared with the wild-type PVX. However, previous studies indicated that the combination of PVX with some VSRs like P19 and HC-Pro that are capable of inducing a systemic necrosis response in N. benthamiana does not relate to an increase in the PVX genomic RNA levels, but enhances and/or stabilizes PVX subgenomic RNAs [57]. In the case of the heterologous expression of tomato chlorosis virus p22 suppressor, no obvious differences in PVX accumulation were observed in the viral accumulation of PVX-p22 and PVX [58]. Since the enhanced pathogenicity associated with recombinant PVX-V2 does not seem to link to a more efficient viral accumulation, other mechanisms must be involved. It is also remarkable that the areas infiltrated with 35S-GFP plus RepA or 35S-GFP plus Rep became necrotic at three to four dpi in the co-infiltration leaf patch assays, suggesting that RepA and Rep may be associated with other plant defenses.
The disease denoted mulberry mosaic dwarf (MMD) has seriously affected Morus alba in China for a long time, and it has a close association with MMDaV. Our findings present the first identification of a VSR from a woody plant-infecting geminivirus, providing a starting point for understanding the mechanisms that are involved in MMDaV-induced MMD.
Supplementary Materials: The following are available online at http://www.mdpi.com/1999-4915/10/9/472/s1. Figure S1. The effect of MMDaV ORFs on the prevention of GFP silencing. Nicotiana benthamiana 16c plants were infiltrated with a mixture of Agrobacterium cultures containing 35S-GFP and pCHF3 vectors expressing individual ORFs of MMDaV, respectively. The constructs used for infiltration are indicated. Photographs were taken under UV light with a yellow filter-mounted Canon camera or under natural light at 8 dpi. Note that co-infiltration of 35S-GFP with RepA or Rep caused necrotic lesions in infiltrated areas at 8 dpi. Figure S2. MMDaV V2 does not suppress dsRNA-induced RNA silencing. N. benthamiana 16c plants were infiltrated with a mixture of Agrobacterium cultures containing 35S-GFP, 35S-dsGFP and MMDaV V2. Co-infiltration of N. benthamiana 16c plants with 35S-GFP, 35S-dsGFP and the empty pCHF3 vector or P19 served as negative or positive controls, respectively. Photographs were taken under UV light at 3 dpi.