The P3N-PIPO Protein Encoded by Wheat Yellow Mosaic Virus Is a Pathogenicity Determinant and Promotes Its Pathogenicity through Interaction with NbRLK6 in Nicotiana benthamiana

Similarly to other potyvirids, the bymovirus wheat yellow mosaic virus (WYMV) encodes a P3N-PIPO protein that is expressed by frameshifting occurring within the open reading frame of the P3 protein. P3N-PIPO is known to be essential for the cell-to-cell movement of several potyviruses, but this has not yet been confirmed for the WYMV. Here, we show that the WYMV P3N-PIPO protein influences disease symptom formation. Infection of Nicotiana benthamiana plants with a potato virus X (PVX)-based vector carrying the WYMV P3N-PIPO gene induced more severe disease symptoms and resulted in higher virus accumulation levels than did infection with PVX lacking the P3N-PIPO gene. N. benthamiana P3N-PIPO-interacting proteins were identified through co-immunoprecipitation (Co-IP) coupled with LC-MS/MS (mass spectrometry), and the interaction between P3N-PIPO and the N. benthamiana receptor-like kinase NbRLK6 was further verified by Co-IP and bimolecular fluorescence complementation (BiFC) of transiently-expressed proteins. Furthermore, our investigation showed that the disease symptom severity and accumulation level of PVX-P3N-PIPO were decreased in N. benthamiana plants when NbRLK6 expression was reduced by tobacco rattle virus-induced gene silencing.


Introduction
Wheat yellow mosaic virus (WYMV), belonging to the genus Bymovirus and being transmitted by the plasmodiophorid Polymyxa graminis, can cause the soilborne viral disease wheat yellow mosaic (WYM) [1]. WYM is one of the most serious diseases affecting wheat production in East Asia [2].
Previous examination of the WYMV genome suggested that it produces ten proteins (P3, 7k, CI, 14k, VPg, NIa, NIb, CP, P1 and P2) derived by proteolytic cleavage of the two polyproteins encoded by the two viral RNAs [1]. In recent years, a small overlapping coding sequence, which is termed PIPO (pretty interesting potyviridae ORF), was discovered by bioinformatics analysis within the genome sequences of viruses in the Potyviridae family [3]. A novel protein named P3N-PIPO, with the PIPO segment being encoded in a different frame within the P3 cistron, is expressed by a viral polymerase slippage mechanism [4][5][6]. In previous studies, P3N-PIPO was shown to be required for efficient cell-to-cell movement of several potyviruses [7][8][9][10][11]. Here we report the results of experiments expressing the WYMV P3-PIPO protein from potato virus X (PVX). This study shows that the WYMV P3NPIPO protein is a pathogenicity determinant, which enhances the pathogenicity of PVX in N. benthamiana. To better understand the mechanism by which WYMV P3N-PIPO influences disease symptom development, we identified host plant proteins that physically interact with P3N-PIPO. Interaction with one of these, NbRLK6, a member of the receptor-like kinase (RLKs) family, was confirmed by co-immunoprecipitation and BiFC assays. Furthermore, our investigation showed that silencing of NbRLK6 expression in N. benthamiana decreased both the severity of PVX disease symptom production and virus accumulation. Previous Co-IP studies of other viruses have not identified RLK6 as a P3-PIPO binding partner. This is the first report of RLK6 as a binding partner of P3-PIPO from any virus.

Plasmid Construction
For the pathogenicity assessment of P3N-PIPO and Co-IP/MS and BiFC assays, the constructs were generated using Gateway technology (Invitrogen Life Technologies, Carlsbad, CA, USA). The full-length sequences of WYMV P3N-PIPO (contain G 2 A 7 motif) and P3N-PIPO-AS (antisense) were amplified with the corresponding primer pairs (Table S1), from the constructed vector PVX-WYMV P3N-PIPO(FS-1)-GFP in our previous study [6]. The NbRLK6 gene was amplified with the corresponding primer pair (Table S1), from N. benthamiana cDNA. The PCR fragments of P3N-PIPO, P3N-PIPO-AS and NbRLK6 were cloned into the pDONR207 entry vector through Gateway technology (Invitrogen Life Technologies, Carlsbad, CA, USA). Then, we subcloned into the destination vectors PVX-attB ( Figure S1), pcL112-nYFP, pcL113-cYFP (these destination vectors were kind gifts from Dr. Stuart MacFarlane) and pBA-Flag-4Myc-DC (kept by Dr. Yanjun Li), via the Gateway ® LR Recombination reaction (Invitrogen Life Technologies, Carlsbad, CA, USA) following the manufacturer's instructions. PVX (pGR106) was a kind gift from David Baulcombe's laboratory, PVX-attB was from Dr. Sean Chapman's laboratory and PVX-P19 was from Dr. Stuart MacFarlane's laboratory.
To silence NbRLK6 expression in N. benthamiana, a 200 bp fragment of the NbRLK6 sequence was amplified and positive-sense-inserted into the pTRV2-LIC vector (kindly provided by Professor Yule Liu from Tsinghua University, Beijing, China) through the LIC method to generate pTRV2-NbRLK6.

Plant Growth and Agrobacterium Infection of Plants
N. benthamiana were grown in a glasshouse under 25 • C-22 • C ± 2 • C (day: night)and 16 h/8 h (light/dark) photoperiod conditions. Agrobacterium cultures containing the experimental plasmids were infiltrated into the leaves of 4-leaf-stage N. benthamiana plants using the previously described method [12].

Co-IP/MS
Agrobacterium cultures carrying the plasmid pBA-Flag-4Myc-P3N-PIPO infiltrated N. benthamiana leaves, and the systemic leaves were collected for the total protein extraction. The P3N-PIPO complexed with interacting plant proteins was isolated following the method described previously [13].
Then, the washed beads carrying P3N-PIPO and co-precipitated host plant proteins were sent for MS analysis using a quadrupole-orbitrap mass spectrometer, Q Exactive HF (Thermo Scientific, Waltham, MA, USA). Finally, data were analyzed according to the previous description [14,15].

BiFC Assay
Plasmids pcL-cYFP-P3N-PIPO and pcL-nYFP-NbRLK6 were separately transformed into A. tumefaciens strain GV3101. The two agrobacterium cultures were mixed at a 1:1 (v/v) ratio and infiltrated into the leaves of N. benthamiana plants. Mixed agrobacterium cultures carrying pcL-cYFP-P3N-PIPO and pcL-nYFP-GUS served as a non-interacting control. At 48 h post-infiltration (hpi), the leaves were collected and examined under a Leica TCS SP8X confocal microscope (Leica Microsystems, Bannockburn, IL, USA).

Virus-Induced Gene Silencing in N. benthamiana
Plasmids pTRV1, pTRV2 and pTRV2-NbRLK6 (Table S2) were individually transformed into A. tumefaciens strain GV3101. Thereafter, the mixed agrobacterium cultures of pTRV1 and TRV:NbRLK6, were infiltrated into the leaves of 3-leaf-stage N. benthamiana plants. The combination of TRV:00 (pTRV1 and pTRV2) was used as a negative control for the VIGS experiments. More details of the VIGS methodology were described previously [17].

Quantitative Reverse Transcription Polymerase Chain Reaction (qRT-PCR)
Total RNA was extracted from leaf tissue using TRIzol reagent (Invitrogen Life Technologies, Carlsbad, CA, USA) following the manufacturer's protocols. The complementary DNA (cDNA) was synthesized using the HiScript III RT SuperMix for qPCR (plus gDNA wiper) (Vazyme, Nanjing, China). qRT PCR was performed on the LightCycler 480 ® Real-Time PCR System (Roche Diagnostics, Basel, Switzerland) using the Universal Blue qPCR SYBR Master Mix (Yeasen, Shanghai, China). The NbUBC (Ubiquitin C) gene was used as an internal reference in the assays. The results were analyzed by the 2-∆∆CT method [18].

P3N-PIPO Enhanced the Pathogenicity of PVX in N. benthamiana
To investigate the properties of the WYMV P3N-PIPO protein, the P3N-PIPO gene was inserted into a PVX-based vector (pGR106), enabling its expression in N. benthamiana plants [19,20]. In addition, either the P19 gene of tomato bushy stunt virus (TBSV), a wellstudied viral pathogenicity determinant, or an antisense version of the WYMV P3-PIPO gene (P3N-PIPO-AS), was cloned into pGR106 as an experimental control. According to our results, the PVX-P3N-PIPO-inoculated plants exhibited irregular leaf mottling at 7dpi, and then severe irregular leaf mottling and leaf deformation at 10 dpi; PVX-P3N-PIPO-ASinoculated plants showed only very mild mosaic symptoms at 10 dpi ( Figure 1a HRP secondary antibody (1:10000, TransGen, Beijing, China). Finally, the blot was treated with the ECL solution (Sangon, Shanghai, China) and examined using a GE Amersham imager 680.

P3N-PIPO Enhanced the Pathogenicity of PVX in N. benthamiana
To investigate the properties of the WYMV P3N-PIPO protein, the P3N-PIPO gene was inserted into a PVX-based vector (pGR106), enabling its expression in N. benthamiana plants [19,20]. In addition, either the P19 gene of tomato bushy stunt virus (TBSV), a wellstudied viral pathogenicity determinant, or an antisense version of the WYMV P3-PIPO gene (P3N-PIPO-AS), was cloned into pGR106 as an experimental control. According to our results, the PVX-P3N-PIPO-inoculated plants exhibited irregular leaf mottling at 7dpi, and then severe irregular leaf mottling and leaf deformation at 10 dpi; PVX-P3N-PIPO-AS-inoculated plants showed only very mild mosaic symptoms at 10 dpi ( Figure 1a   To further investigate the effect of P3N-PIPO expression on PVX infection, the systemically infected leaves from PVX-EV-, PVX-P19-, PVX-P3N-PIPO-AS-and PVX-P3N-PIPOinoculated plants were analyzed by Western blotting and qRT-PCR. The results showed that there were higher levels of PVX coat protein (CP) and viral RNA in the PVX-P3N-PIPOinoculated plants than in the PVX-P3N-PIPO-AS plants (Figure 1b, c). Consequently, it was concluded that the WYMV P3N-PIPO protein could act a determinant of PVX pathogenicity in N. benthamiana plants.

Identification and Analysis of Plant Proteins Interacting with WYMV P3N-PIPO
Investigation of protein-protein interactions is an essential process to understanding the underlying processes taking place during virus-host interactions. To identify potential interacting host factors, the WYMV P3N-PIPO tagged with both FLAG and Myc epitopes was transiently expressed in N. benthamiana leaves by agroinfiltration. Potential complexes of P3N-PIPO and host proteins were initially immunoprecipitated using the FLAG tag, and then, after further washing steps, were re-precipitated using the Myc tag. This approach greatly reduces the co-precipitation of non-specific binding proteins. Subsequently, the remaining captured proteins were identified using LC-MS/MS. This analysis identified 117 potential P3N-PIPO-interacting proteins (File S1). To understand in more detail the putative functional processes associated with the candidate P3N-PIPO-interacting proteins, GO analysis was conducted to provide known information about the biological processes (BP), cellular components (CC) and molecular functions (MF) of the candidate proteins; and KEGG pathway analysis was performed to show other interactions known for the candidate P3N-PIPO-binding proteins.
According to our results, the enriched GO-CC terms were mainly peroxisome, plasma membrane and membrane; the enriched MF terms included transporter activity and catalytic activity; and the BP terms included lipid metabolic process and catabolic process (Figure 2a). The top 10 KEGG pathways enriched by the target genes included amino acid metabolism, lipid metabolism, metabolism of other amino acids and metabolism pathways (Figure 2b). To further investigate the effect of P3N-PIPO expression on PVX infection, the systemically infected leaves from PVX-EV-, PVX-P19-, PVX-P3N-PIPO-AS-and PVX-P3N-PIPO-inoculated plants were analyzed by Western blotting and qRT-PCR. The results showed that there were higher levels of PVX coat protein (CP) and viral RNA in the PVX-P3N-PIPO-inoculated plants than in the PVX-P3N-PIPO-AS plants (Figure 1b, c). Consequently, it was concluded that the WYMV P3N-PIPO protein could act a determinant of PVX pathogenicity in N. benthamiana plants.

Identification and Analysis of Plant Proteins Interacting with WYMV P3N-PIPO
Investigation of protein-protein interactions is an essential process to understanding the underlying processes taking place during virus-host interactions. To identify potential interacting host factors, the WYMV P3N-PIPO tagged with both FLAG and Myc epitopes was transiently expressed in N. benthamiana leaves by agroinfiltration. Potential complexes of P3N-PIPO and host proteins were initially immunoprecipitated using the FLAG tag, and then, after further washing steps, were re-precipitated using the Myc tag. This approach greatly reduces the co-precipitation of non-specific binding proteins. Subsequently, the remaining captured proteins were identified using LC-MS/MS. This analysis identified 117 potential P3N-PIPO-interacting proteins (File S1). To understand in more detail the putative functional processes associated with the candidate P3N-PIPO-interacting proteins, GO analysis was conducted to provide known information about the biological processes (BP), cellular components (CC) and molecular functions (MF) of the candidate proteins; and KEGG pathway analysis was performed to show other interactions known for the candidate P3N-PIPO-binding proteins.
According to our results, the enriched GO-CC terms were mainly peroxisome, plasma membrane and membrane; the enriched MF terms included transporter activity and catalytic activity; and the BP terms included lipid metabolic process and catabolic process (Figure 2a). The top 10 KEGG pathways enriched by the target genes included amino acid metabolism, lipid metabolism, metabolism of other amino acids and metabolism pathways (Figure 2b).

P3N-PIPO Interacts with NbRLK6
Among the 117 potential P3N-PIPO-interacting proteins identified in our screen, we selected a putative receptor-like kinase NbRLK6 for further study, as RLK proteins are known to play a prominent role in various plant-pathogen interactions [21]. Biomolecular fluorescence complementation (BiFC) assays were performed to verify the interaction of WYMV P3N-PIPO and NbRLK6 in plants. Thus, using agroinfiltration, we transiently co-expressed cYFP-P3N-PIPO and nYFP-NbRLK6 in N. benthamiana leaves. This resulted in the production of a YFP fluorescence signal in the cell cytoplasm at 48 h post-infiltration, suggesting physical interaction between these two proteins. However, no YFP fluorescence was observed in leaves co-infiltrated with cYFP-P3N-PIPO and the non-interacting protein nYFP-GUS, ruling out the possibility of YFP auto-fluorescence from the cYFP-P3N-PIPO protein (Figure 3a).
Among the 117 potential P3N-PIPO-interacting proteins identified in our screen, we selected a putative receptor-like kinase NbRLK6 for further study, as RLK proteins are known to play a prominent role in various plant-pathogen interactions [21]. Biomolecular fluorescence complementation (BiFC) assays were performed to verify the interaction of WYMV P3N-PIPO and NbRLK6 in plants. Thus, using agroinfiltration, we transiently coexpressed cYFP-P3N-PIPO and nYFP-NbRLK6 in N. benthamiana leaves. This resulted in the production of a YFP fluorescence signal in the cell cytoplasm at 48 h post-infiltration suggesting physical interaction between these two proteins. However, no YFP fluorescence was observed in leaves co-infiltrated with cYFP-P3N-PIPO and the non-interacting protein nYFP-GUS, ruling out the possibility of YFP auto-fluorescence from the cYFP-P3N-PIPO protein (Figure 3a).
As a further test for interaction, WYMV P3N-IPO tagged with Myc and NbRLK6 tagged with FLAG were co-expressed by agroinfiltration in N. benthamiana leaves. As a non-interaction control, WYMV P3N-PIPO-Myc was co-expressed with GFP-FLAG. Leaf extracts were precipitated using a bead-immobilized anti-FLAG antibody, and the collected proteins were examined by Western blotting. In these experiments, P3N-PIPO was co-precipitated in combination with NbRLK6-FLAG, but not with GFP-FLAG (Figure 3b)   Figure 3. Verification of the interaction between the WYMV P3N-PIPO protein and NbRLK6 in agro infiltrated N. benthamiana. (a) Detection of the interaction between P3N-PIPO and NbRLK6 by bimolecular fluorescence complementation test. Fluorescence, bright field (DIC) and merged images of the lower epidermises of leaves infiltrated with cYFP-and nYFP-fusion constructs were taken using a laser scanning confocal microscope. Bar = 50 μm. (b) Detection of the interaction between P3N-PIPO and NbRLK6 by Co-IP assay. Protein extracts were precipitated using an anti-FLAG antibody. Input refers to the total protein extract. Flag-IP refers to immunoprecipitated proteins, which were detected using either anti-FLAG or anti-Myc antibodies. "+" and "−" represent the presence or absence of specific constructs in the agroinfiltration mixture. Fluorescence, bright field (DIC) and merged images of the lower epidermises of leaves infiltrated with cYFP-and nYFP-fusion constructs were taken using a laser scanning confocal microscope. Bar = 50 µm. (b) Detection of the interaction between P3N-PIPO and NbRLK6 by Co-IP assay. Protein extracts were precipitated using an anti-FLAG antibody. Input refers to the total protein extract. Flag-IP refers to immunoprecipitated proteins, which were detected using either anti-FLAG or anti-Myc antibodies. "+" and "−" represent the presence or absence of specific constructs in the agroinfiltration mixture.
As a further test for interaction, WYMV P3N-IPO tagged with Myc and NbRLK6 tagged with FLAG were co-expressed by agroinfiltration in N. benthamiana leaves. As a noninteraction control, WYMV P3N-PIPO-Myc was co-expressed with GFP-FLAG. Leaf extracts were precipitated using a bead-immobilized anti-FLAG antibody, and the collected proteins were examined by Western blotting. In these experiments, P3N-PIPO was co-precipitated in combination with NbRLK6-FLAG, but not with GFP-FLAG (Figure 3b).
To study whether NbRLK6 and P3N-PIPO co-localized in plants, leaves were infiltrated with constructs expressing NbRLK6 fused to the green fluorescent protein (NbRLK6-GFP) and P3N-PIPO fused to the red fluorescent protein (P3N-PIPO-mRFP). When expressed alone, P3N-PIPO (mRFP) was associated with the nucleus and cytomembrane, and NbRLK6 (GFP) aggregated to form vesicles (Figure 4a). While the relatively weak fluorescence precluded visualization of P3N-PIPO-RFP at the plasmodesmata (PD), fusion of P3N-PIPO to GFP, as expected, clearly showed localization of P3N-PIPO with PD ( Figure S2). The co-expression led to the redistribution of P3N-PIPO to the vesicles of NbRLK6 (Figure 4b), which further verified the interaction between P3N-PIPO and NbRLK6, as identified using Co-IP (Figure 3). trated with constructs expressing NbRLK6 fused to the green fluorescent protein (NbRLK6-GFP) and P3N-PIPO fused to the red fluorescent protein (P3N-PIPO-mRFP). When expressed alone, P3N-PIPO (mRFP) was associated with the nucleus and cytomembrane, and NbRLK6 (GFP) aggregated to form vesicles (Figure 4a). While the relatively weak fluorescence precluded visualization of P3N-PIPO-RFP at the plasmodesmata (PD), fusion of P3N-PIPO to GFP, as expected, clearly showed localization of P3N-PIPO with PD ( Figure S2). The co-expression led to the redistribution of P3N-PIPO to the vesicles of NbRLK6 (Figure 4b), which further verified the interaction between P3N-PIPO and NbRLK6, as identified using Co-IP (Figure 3). .

Silencing of NbRLK6 Expression in N. benthamiana Decreases PVX Accumulation
To determine the role of NbRLK6 in virus infection, NbRLK6 expression was silenced in N. benthamiana using the TRV-based VIGS vector and inoculated with empty vector (TRV:00) as a control. By 7 dpi, the TRV:00 and TRV:NbRLK6-inoculated plants showed

Silencing of NbRLK6 Expression in N. benthamiana Decreases PVX Accumulation
To determine the role of NbRLK6 in virus infection, NbRLK6 expression was silenced in N. benthamiana using the TRV-based VIGS vector and inoculated with empty vector (TRV:00) as a control. By 7 dpi, the TRV:00 and TRV:NbRLK6-inoculated plants showed no obvious symptoms ( Figure 5a); however, the qRT-PCR assay suggested that the mean transcription level of NbRLK6 in TRV:NbRLK6-inoculated plants was reduced by approximately 80% compared with that in TRV:00-inoculated control plants (Figure 5b). Thereafter, the effects of NbRLK6 on PVX-P3N-PIPO accumulation and symptom induction were investigated. By 14 dpi, the TRV:00 control plants inoculated with PVX-P3N-PIPO exhibited severe leaf curling and mosaic symptoms, whereas only mild leaf curling symptoms were observed on the NbRLK6-silenced plants after inoculation with PVX-P3N-PIPO (Figure 5a). qRT-PCR and Western blotting results showed that, compared with control plants, the accumulation level of PVX was significantly reduced in the NbRLK6-silenced plants (Figure 5c,d). Taken together, these results indicate that NbRLK6 plays important roles in virus accumulation and the development of disease symptoms in PVX-infected plants.
no obvious symptoms ( Figure 5a); however, the qRT-PCR assay suggested that the mean transcription level of NbRLK6 in TRV:NbRLK6-inoculated plants was reduced by approximately 80% compared with that in TRV:00-inoculated control plants (Figure 5b). Thereafter, the effects of NbRLK6 on PVX-P3N-PIPO accumulation and symptom induction were investigated. By 14 dpi, the TRV:00 control plants inoculated with PVX-P3N-PIPO exhibited severe leaf curling and mosaic symptoms, whereas only mild leaf curling symptoms were observed on the NbRLK6-silenced plants after inoculation with PVX-P3N-PIPO (Figure 5a). qRT-PCR and Western blotting results showed that, compared with control plants, the accumulation level of PVX was significantly reduced in the NbRLK6-silenced plants (Figure 5c,d). Taken together, these results indicate that NbRLK6 plays important roles in virus accumulation and the development of disease symptoms in PVXinfected plants.

Discussion
In plants, RLKs play a key role in regulating plant development and plant responses to environmental stimuli [22][23][24]. Many studies have revealed that RLKs are involved in plant defense against bacteria, fungi, oomycetes, and even insects [25].

Discussion
In plants, RLKs play a key role in regulating plant development and plant responses to environmental stimuli [22][23][24]. Many studies have revealed that RLKs are involved in plant defense against bacteria, fungi, oomycetes, and even insects [25].
A typical RLK contains an extracellular domain, a single-pass transmembrane domain, and a cytoplasmic kinase domain [26]. Usually, the extracellular domain of RLKs was involved in perceiving the invasion of pathogens. As viruses are intracellular pathogens, the involvement of RLKs in plant-virus interactions seems counterintuitive. However, more and more evidence shows that RLKs affect plant susceptibility to viruses, and in some cases, interact with viral proteins [27][28][29][30][31][32][33][34]. The nuclear shuttle protein (NSP) of bipartite geminiviruses acts as a virulence factor, which also binds to Arabidopsis RLKs, and thus inhibits their function in anti-viral defense [28,[35][36][37]. Arabidopsis NSP-interacting kinases 1, 2, and 3 (AtNIK1-3) were the first RLKs identified to interact with viral protein [35]. The C4 protein of geminiviruses has been the most extensively studied in terms of the interaction between RLKs and viral proteins. Typically, the C4/AC4 protein of geminiviruses has been revealed to interact with RLKs in the CLAVATA 1 (CLV1) clade [25,38].
Numerous studies have indicated that P3N-PIPO is the movement protein of potyviruses [7][8][9][10][11]. In addition, P3N-PIPO determines the virulence of clover yellow vein virus (ClYVV) in both resistant and susceptible peas [39]. However, the mechanism underlying this phenomenon remains unresolved. Here, our study revealed that WYMV P3N-PIPO is a determinant of pathogenicity, which enhances the symptom severity and accumulation level of PVX in N. benthamiana. We discovered that NbRLK6, a member of the plant RLK family, interacts with WYMV P3N-PIPO, and confirmed the interaction by BiFC and Co-IP assays. Furthermore, reducing NbRLK6 expression in N. benthamiana by TRV-VIGS decreased the accumulation of the chimeric virus PVX-P3N-PIPO and alleviated the viral disease symptoms.
Intriguingly, over the past few years, most studies on the interaction between RLKs and viral proteins were focused on those of geminiviruses, which have DNA genomes and nuclear replication cycles [25]. Our study focused on the RNA virus PVX and suggests that more work should be performed on viruses from different families in order to more systematically study the interaction between RLKs and viral proteins.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.