p15 encoded by Garlic virus X is a pathogenicity factor and RNA silencing suppressor

Garlic virus X (GarVX) encodes a 15 kDa cysteine-rich protein (CRP). To investigate the function(s) of p15, its subcellular localization, role as a symptom determinant and capacity to act as a viral suppressor of RNA silencing (VSR) were analysed. Results showed that GFP-tagged p15 was distributed in the cytoplasm, nucleus and nucleolus. Expression of p15 from PVX caused additional systemic foliar malformation and led to increased accumulation of PVX, showing that p15 is a virulence factor for reconstructed PVX-p15. Moreover, using a transient agro-infiltration patch assay and a Turnip crinkle virus (TCV) movement complementation assay, it was demonstrated that p15 possesses weak RNA silencing suppressor activity. Removal of an amino acid motif resembling a nuclear localization signal (NLS) prevented p15 from accumulating in the nucleus but did not abolish its silencing suppression activity. This study provides the first insights into the multiple functions of the GarVX p15 protein. Garlic virus X (GarVX), a single-strand positive-sense RNA virus, belongs to the genus Allexivirus of the family Alphaflexiviridae, and encodes six open reading frames (ORFs) sharing a similarity with carlaviruses, potexviruses and foveaviruses in genomic organization [1–4]. ORF1 encodes a 174 kDa protein for viral replication, while ORF2 and ORF3 encode 26 kDa and 11 kDa triple-gene block (TGB) proteins (TGBp1, TGBp2). ORF5 encodes the 26 kDa coat protein (CP), whereas ORF4 and ORF6 encode 32 kDa and 15 kDa proteins of unknown function. Although GarVX and other allexiviruses are present in Allium species globally [1, 3, 5–8], molecular studies of the functions of allexivirus proteins are limited at present. Our previous study showed that GarVX TGBp2 (p11) acts as a virus movement protein (MP) and localizes to plasmodesmata (PD), the cytoplasm and the endoplasmic reticulum (ER) in epidermal cells of Nicotiana benthamiana (N. benthamiana) [9]. Moreover, overexpression of p11 from Potato virus X (PVX) or Tobacco mosaic virus (TMV) induced the unfolded protein response (UPR) and programmed cell death (PCD) via the up-regulation of UPR related genes [10]. ORF6, located at the 3¢-proximal end of the GarVX genome encodes a small, 15 kDa protein (p15) of unknown function, which has features shared with viral cysteine-rich proteins (CRP) [11–17]. There are many positive-strand RNA plant viruses from the genera Hordeivirus, Pecluvirus, Furovirus, Tobravirus, Carlavirus, Pomovirus, Benyvirus and Allexivirus that encode CRP proteins [12, 15, 16, 18–23]. CRP proteins are generally characterized by the presence of multiple cysteine residues in their amino acid sequences, although there is no other significant amino acid sequence similarity between the CRPs of viruses from different genera. For a number of plant viruses their CRPs have been shown to play a role in viral symptom induction and suppression of RNA silencing (VSR) [12, 13, 15, 16, 18–21, 24–26]. Although many studies have described these two properties of plant virus CRPs, the mechanisms of action of these proteins are, for the most part, unknown. Until now the CRPs of allexiviruses have not been studied. Here, we describe Received 30 June 2018; Accepted 16 August 2018 Author affiliations: College of Plant Protection, Shenyang Agriculture University, Shenyang 110161, Liaoning, PR China; Institute of Plant Virology, Ningbo University, Ningbo 315211, PR China; State Key laboratory Breeding Base for Sustainable Control of Pest and Disease, Key Laboratory of Biotechnology in Plant Protection, Zhejiang Academy of Agricultural Sciences, Hangzhou 310021, PR China; College of Plant Protection, Nanjing Agricultural University, Nanjing 210095, PR China; Public Lab, Zhejiang Academy of Agricultural Sciences, Hangzhou 310021, PR China; The James Hutton Institute, Cell and Molecular Sciences Group, Invergowrie, Dundee DD2 5DA, UK. *Correspondence: Yuwen Lu, luyuwen@yeah.net; Fei Yan, fei.yan@mail.zaas.ac.cn

ORF6, located at the 3¢-proximal end of the GarVX genome encodes a small, 15 kDa protein (p15) of unknown function, which has features shared with viral cysteine-rich proteins (CRP) [11][12][13][14][15][16][17].There are many positive-strand RNA plant viruses from the genera Hordeivirus, Pecluvirus, Furovirus, Tobravirus, Carlavirus, Pomovirus, Benyvirus and Allexivirus that encode CRP proteins [12,15,16,[18][19][20][21][22][23].CRP proteins are generally characterized by the presence of multiple cysteine residues in their amino acid sequences, although there is no other significant amino acid sequence similarity between the CRPs of viruses from different genera.For a number of plant viruses their CRPs have been shown to play a role in viral symptom induction and suppression of RNA silencing (VSR) [12,13,15,16,[18][19][20][21][24][25][26].Although many studies have described these two properties of plant virus CRPs, the mechanisms of action of these proteins are, for the most part, unknown.Until now the CRPs of allexiviruses have not been studied.Here, we describe experiments done to investigate the functions of the p15 CRP from the Zhejiang isolate of GarVX.
To investigate the sub-cellular localization of GarVX p15 in plants, a binary plasmid was constructed allowing transient expression of p15 with the green fluorescent protein (GFP) fused to its C-terminus (p15-GFP) using a strategy described previously [9].Three days after infiltration into N. benthamiana leaves of agrobacterium cells carrying the p15-GFP construct or an unfused, free GFP control construct, GFP fluorescence was examined by confocal microscopy.In cells expressing p15-GFP, green fluorescence was localized to the cell periphery and the nucleus, including the nucleolus (Fig. 1A, upper panels).The same distribution pattern was observed using an N-terminally GFP-tagged p15 (GFP-p15) (Fig. S1, available in the online version of this article), indicating that the position of the GFP tag did not alter the localization of p15.In contrast, the fluorescence in cells expressing free GFP was cytoplasmic and nuclear (Fig. 1A, bottom panels), but not nucleolar.p15 contains an arginine-rich sequence (RRRAKR) that is a potential nuclear localization signal (NLS) in the middle part of the protein (Fig. 1B, highlighted in red).A similar sequence is also found in the CRP of carlaviruses (Fig. S2).
To verify the role of the NLS in nuclear/nucleolar localization of p15, a NLS deletion mutant (p15m) was constructed in which the entire sequence encoding six amino acid residues was removed (Fig. 1B).The p15m-GFP protein was expressed in plants transformed with a histone H2B-mRFP construct, where the red fluorescence is localized to the plant nucleus.In these experiments, the p15m-GFP protein was localized to the cell periphery in the form of isolated granules but was not present in the nucleus or nucleolus (Fig. 1C).Similar amounts of full-length p15-GFP and p15m-GFP were shown to be expressed in these plants, along with possibly multimeric and truncated forms of p15m-GFP (Fig. S3).
CRPs from several viruses have been shown to be symptom determinants, often being involved in modulating virus pathogenicity [18].To examine whether the GarVX p15 protein is a pathogenicity factor, we inserted the p15 gene into an Agrobacterium-delivered infectious clone of PVX [33].In addition, the nuclear localization-deficient mutant p15m gene was also cloned into PVX.As a control in this assay the GFP gene was also inserted into PVX (PVX-GFP).These three PVX constructs were delivered individually into N. benthamiana by agro-infiltration.At 15 days post infiltration (dpi), the upper leaves of plants systemically infected with PVX-p15 showed severe hyperplasia and chlorosis (Fig. 2A, left panels), whereas the symptoms on upper leaves on PVX-p15m-(Fig.2A, middle panels) and PVX-GFP (Fig. 2A, right panels)-infected plants were mild, involving slight leaf crinkling.The p15 gene could be amplified from systemic infected tissue using RT-PCR with p15-specific primers (Table S1), showing that both p15 and p15m were stably maintained in the modified PVX genome at 15 dpi (Fig. 2C).Examination of leaf cross-sections by light microscopy showed that the upper leaves of PVX-p15infected plants had a pronounced disorganization of the palisade mesophyll tissue compared with PVX-p15m-and PVX-GFP-infected leaves (Fig. 2B).Western blotting and Northern blotting were done to examine the accumulation of the PVX CP and viral RNAs in systemically infected leaves of plants infiltrated with PVX-p15, PVX-p15m and PVX-GFP (Fig. 2D, E).The accumulation levels of PVX CP was higher in plants infected with PVX-p15 and PVX-p15m compared to PVX-GFP (Fig. 2D).The viral RNA level was highest in PVX-p15-infected plants, slightly lower in PVX-p15m-infected plants and lowest in PVX-GFP infected plants (Fig. 2E).Although the difference in symptoms caused by PVX-p15m and PVX-GFP was very small, the accumulation of PVX-p15m was obviously increased compared to PVX-GFP (Fig. 2D, E), showing that nuclear localization of p15 is not required for the protein to influence PVX accumulation in systemic leaves.These data suggested that the GarVX p15 protein is a pathogenicity factor that can increase virus accumulation and cause leaf malformation; however, the nuclear localization of p15 is involved in inducing abnormal development of foliage but not in promoting viral infection.
Many plant virus pathogenicity factors [34,35], and several CRPs [12,15,19,20], have been found to interfere with the plant RNA silencing system, which is an evolutionarily conserved, sequence-specific defence mechanism repressing viral infection [36,37].To investigate whether p15 is a viral  S1. (D) Detection of virus CP accumulation in PVX-p15, PVX-p15m and PVX-GFP systemically infected leaves using CP antibody.Lower panel shows Ponceau red staining of leaf proteins as loading control.PVX CP Western blotting was conducted as described before [44].Quantitation calculation of digital images using ImageJ [45], means±SD representing results of three independent replicate experiments.(E) Northern blot analysis of PVX RNA in PVX-p15, PVX-p15m and PVX-GFP systemically infected leaves at 15 dpi.Lower panel shows ethidium bromide staining of rRNA as a loading control.A CP-specific primer pair (Table S1) was used to amplify a CP template for DIG-labelled probe synthesis.Relative RNA analysis was performed according to previous procedures [44].Quantitation calculation of digital images was made from Northern blots using ImageJ; means ±SD represents results of three independent experiments.suppressor of RNA silencing (VSR) we used two different approaches.In the first, we examined whether expression of p15 could interfere with the silencing of a GFP-transgene present in N. benthamiana line 16c plants [38].As described before [19], we induced silencing of the GFP transgene by infiltration of the plants with agrobacteria carrying a second copy of the GFP gene.The plants were simultaneously infiltrated with a second agrobacterium culture expressing either p15, p15m, the well-characterized suppressor p25 from PVX [38], or an empty binary vector.Infiltrated plants were examined at 5 dpi using UV illumination.The patch infiltrated with the GFP binary together with the empty binary vector control had very weak fluorescence, demonstrating active RNA silencing (Fig. 3A, left lower patch).In contrast, the patch infiltrated with PVX p25 and GFP showed brighter green fluorescence at 5 dpi where p25 had prevented silencing from occurring (Fig. 3A, left upper patch).In the patches infiltrated with GFP together with either p15 or p15m plus GFP, an intermediate level of green fluorescence was maintained indicating that both p15 and p15m were able to suppress RNA silencing (Fig. 3A, p15, right upper patch; p15m, right lower patch).Western blotting showed that the GFP protein accumulated to a higher level in p15-or p15m-infiltrated patches than in the empty vector-treated patch but was lower than in the p25-treated patch (Fig. 3B).Northern blot analysis at 5 dpi revealed that expression of GFP mRNA was higher in patches coinfiltrated with GFP and p15 or GFP and p15m than in that co-infiltrated with GFP and empty vector, but was lower than in the GFP and p25-treated patch (Fig. 3C).Further analysis of siRNAs indicated that by 5dpi, GFP siRNA had accumulated to the highest level in the empty vector-treated patch.Both the p15-and p15m-infiltrated patches accumulated less siRNAs than the empty vector patch but more than the p25-treated patch (Fig. 3C).These data all indicated that both p15 and p15m possess RNA silencing suppression ability.

Downloaded
In a second approach to further examine the RNA silencing suppression ability of p15, a Turnip crinkle virus (TCV) movement complementation assay was done [31,32,39].This assay uses a TCV construct (Ti-plasmid version vector, PZP-TCV-sGFP), which harbours the GFP gene in place of the virus CP gene.The TCV CP is a strong suppressor of silencing and is required for cell-to-cell movement of TCV.The GFP derivative of TCV is, by itself, unable to move between cells and is confined to the initial site of infection.In these experiments plants were infiltrated with an extremely diluted (OD 600 =0.001) culture of Agrobacterium culture harbouring PZP-TCV-sGFP, which is sufficient to express TCV-sGFP in widely separated single cells throughout the leaf.The TCV-GFP culture was mixed with a second Agrobacterium culture (OD 600 =0.5) expressing either p15 or p15m.Agrobacterium cultures carrying an empty vector or expressing the Turnip mosaic virus (TuMV) HC-Pro silencing suppressor protein were used as negative or positive controls, respectively.As expected, using the empty vector (negative control) the GFP fluorescence produced during TCV replication was confined to small foci of one to two cells at 3 dpi (Fig. 3D, left bottom panel), confirming that TCV-sGFP movement was restricted in the absence of a suppressor protein.In contrast, TuMV HC-Pro was able to complement the cell-to-cell movement of the CP (VSR)deficient TCV-sGFP, producing infection foci of four to 30 cells (Fig. 3D, left upper panel).At the same time, the movement of TCV-sGFP was monitored when either p15 or p15m was transiently expressed with the virus.Both p15 and p15m supported the movement of TCV-sGFP, producing foci of three to 15 cells for p15 (Fig. 3D, right upper panel), and foci of three to nine cells for p15m (Fig. 3D, right bottom panel), respectively.Examination of the numbers of fluorescent foci consisting of one, two, three or more than four cells confirmed that p15 and p15m were able to complement TCV-sGFP movement but less effectively than TuMV HC-Pro (Table 1).Collectively, the observation of fluorescent foci and numerical data demonstrate that the p15 protein can function as a weak VSR, and that removal of the NLS did not completely abolish its silencing suppression activity.
In conclusion, we have shown that the p15 encoded by GarVX ORF6 is localized not only to the cytoplasm, but also to the nucleus and nucleolus of plant cells, and we have identified a NLS sequence motif within the protein that directs nuclear and nucleolar localization.When expressed by PVX, p15 enhances virus accumulation and increases virus symptom severity.We also showed that p15 has weak RNA silencing suppression activity in the leaf infiltration patch assay and PZP-TCV-sGFP complementation experiment.Lastly, we showed that the presence of p15 in the nucleus is not required for it to affect virus accumulation or to suppress RNA silencing; however, the induction of foliar malformation is suggested as linking to p15 nuclear localization.
There are several potential mechanisms by which plant virus silencing suppressor proteins in general, and CRPs in particular, could and do have their effects [40].One mechanism is the physical interference with RNA silencing components.For example, the tomato bushy stunt virus (TBSV) p19 protein and the cucumber mosaic virus (CMV) 2b protein bind to siRNAs, and the CMV 2b and TCV P38 bind to Argonaute (Ago) proteins.For CRPs, the tobacco rattle virus (TRV) 16K protein has been shown not to bind siRNAs but to interact with Ago4 [41].The influence of silencing suppressor proteins on symptom development can also result from physical interaction between suppressor proteins and host components.Once again the CMV 2b protein provides an example, where it interacts with plant catalase (CAT) 3 resulting in H 2 O 2 accumulation and induction of tissue necrosis [42].For CRPs, the p12 protein of chrysanthemum virus B causes plant leaf malformations by binding to the DNA promoter of the upp-L gene and upregulating its expression [32].The UppL protein is itself a transcription factor that activates plant cell proliferation.To attempt to understand the mechanism of action of GVX p15, we did experiments to examine whether p15 could bind to siRNAs (21nt single-or double-strand RNA) in vitro, but we did not detect this activity for p15 (Fig. S4).
Nevertheless, our work confirms that p15, currently the only allexivirus CRP to be studied in detail, functions as a pathogenicity factor (at least in the context of a PVX infection) and RNA silencing suppressor, although further studies are required to fully understand the mechanism of silencing suppression by p15.*In each assay, Agrobacterium strain GV3101 harbouring PZP-TCV-sGFP (OD 600 =0.001) mixed with (1 : 1) Agrobacterium strain GV3101 harbouring each of the constructs (#) individually (OD 600 =0.5) for co-expression in N. benthamiana leaves.

Fig. 2 .
Fig. 2. Heterologous expression of p15 promotes PVX infection and induces severe symptom production.(A) (a) Foliar malformation on N. benthamiana plants infected with PVX-expressing p15 (PVX-p15); mild symptoms on N. benthamiana plants infected with (b) PVX-p15m and (c) PVX-GFP.Photographs taken at 15 dpi.(B) Cross-section and light microscopy of N. benthamiana leaves infected with PVX-p15, PVX-p15m and PVX-GFP at 15 dpi.Each construct was examined in four leaves of independent plants.Sample preparation was performed following a previous description [43].Scale bar, 50 µm.(C) Stability of p15 and p15m PVX constructs in upper systemic leaves confirmed by RT-PCR.Primer pair is listed in TableS1.(D) Detection of virus CP accumulation in PVX-p15, PVX-p15m and PVX-GFP systemically infected leaves using CP antibody.Lower panel shows Ponceau red staining of leaf proteins as loading control.PVX CP Western blotting was conducted as described before[44].Quantitation calculation of digital images using ImageJ[45], means±SD representing results of three independent replicate experiments.(E) Northern blot analysis of PVX RNA in PVX-p15, PVX-p15m and PVX-GFP systemically infected leaves at 15 dpi.Lower panel shows ethidium bromide staining of rRNA as a loading control.A CP-specific primer pair (TableS1) was used to amplify a CP template for DIG-labelled probe synthesis.Relative RNA analysis was performed according to previous procedures[44].Quantitation calculation of digital images was made from Northern blots using ImageJ; means ±SD represents results of three independent experiments.

Fig. 3 .
Fig. 3. Characterization of the GarVX p15 protein as a weak suppressor of RNA silencing.(A) GFP image of agro-infiltrated patches of N. benthamiana line 16 c under UV at 5 dpi.The constructs used for agro-infiltration are stated above each patch and the density of each infiltrated agrobacterium culture was OD 600 =0.5.(B) Western blot detection of the accumulation of the GFP protein extracted from the agro-infiltrated patches.Lower panel shows Ponceau red staining of leaf proteins as loading control.Quantitation calculation of digital images of Western blots using ImageJ; means±SD represents results of three independent experiments.(C) Upper: accumulation of GFP mRNA in co-agro-infiltrated patches determined by Northern blot.Empty vector and p25 were used as negative and positive control, respectively.rRNA stained by ethidium bromide as the loading control.Lower: GFP siRNA detection in co-agro-infiltrated patches by Northern blot.Empty vector and p25 were used as negative and positive control, respectively.rRNA stained by ethidium bromide as the loading control.(D) Confocal imaging of GFP fluorescence in infection foci of TCV-sGFP.Co-infiltration of PZP-TCV-sGFP and HCpro (left upper panel); PZP-TCV-sGFP and p15 (right upper panel); PZP-TCV-sGFP and empty vector (left bottom panel); PZP-TCV-sGFP and p15m (right bottom panel).Scale bar, 75 µm.

Table 1 .
Complementation of TCV-sGFP movement in infection foci as observed by the spread of GFP fluorescence from the initial infected cell