The cis‐expression of the coat protein of turnip mosaic virus is essential for viral intercellular movement in plants

Abstract To establish infection, plant viruses are evolutionarily empowered with the ability to spread intercellularly. Potyviruses represent the largest group of known plant‐infecting RNA viruses, including many agriculturally important viruses. To better understand intercellular movement of potyviruses, we used turnip mosaic virus (TuMV) as a model and constructed a double‐fluorescent (green and mCherry) protein‐tagged TuMV infectious clone, which allows distinct observation of primary and secondary infected cells. We conducted a series of deletion and mutation analyses to characterize the role of TuMV coat protein (CP) in viral intercellular movement. TuMV CP has 288 amino acids and is composed of three domains: the N‐terminus (amino acids 1–97), the core (amino acids 98–245), and the C‐terminus (amino acids 246–288). We found that deletion of CP or its segments amino acids 51–199, amino acids 200–283, or amino acids 265–274 abolished the ability of TuMV to spread intercellularly but did not affect virus replication. Interestingly, deletion of amino acids 6–50 in the N‐terminus domain resulted in the formation of aberrant virions but did not significantly compromise TuMV cell‐to‐cell and systemic movement. We identified the charged residues R178 and D222 within the core domain that are essential for virion formation and TuMV local and systemic transport in plants. Moreover, we found that trans‐expression of the wild‐type CP either by TuMV or through genetic transformation‐based stable expression could not rescue the movement defect of CP mutants. Taken together these results suggest that TuMV CP is not essential for viral genome replication but is indispensable for viral intercellular transport where only the cis‐expressed CP is functional.


| INTRODUC TI ON
Viruses are obligate intracellular agents that infect all living organisms and exclusively multiply in their host cells. Viral pathogens account for nearly 50% of newly emerging plant diseases and are considered a major constraint to agriculture, threatening global food security (Anderson et al., 2004). The vast majority of known viruses have positive-sense, single-stranded (+ss) RNA genomes. To establish systemic infection, a plant virus, regardless of the viral genome being DNA or RNA, must have the ability to move intercellularly from the primary infected cells to neighbouring cells through plasmodesmata (PD), and further undergo long-distance movement via the phloem and/ or xylem to reach remote sites of an infected plant (Folimonova and Tilsner, 2018;Reagan and Burch-Smith, 2020). PD are specialized intercellular organelles that connect adjacent cells and are gateways to local and systemic infection (Benitez-Alfonso et al., 2010). They allow small molecules to diffuse between cells and regulate the intercellular movement of macromolecules or macromolecular complexes, including virions and viral ribonucleoprotein complexes (vRNPs) (Navarro et al., 2019). A better understanding of how an infecting virus manages to move intercellularly to establish systemic infection may assist in the development of novel strategies for disease control. Previous studies have suggested that viral cell-to-cell movement is achieved through the coordinated action of virus-encoded movement proteins (MPs), viral particles and/or vRNPs, and host factors (Schoelz et al., 2011;Heinlein, 2015;Wang, 2015;Navarro et al., 2019). Typical MPs may be classified into two groups (Navarro et al., 2019). The first group, exemplified by the single dedicated 30 kDa MP of tobamoviruses, increases the size exclusion limit of PD, without affecting PD structure, to allow vRNPs to pass through (Kawakami et al., 2004;Peña and Heinlein, 2012;Liu and Nelson, 2013). The second group of MPs self-interacts to form tubular structures that modify the PD pore by replacing the endoplasmic reticulum (ER)-derived desmotubule and allow the transport of viral particles from the site of virion assembly to neighbouring cells (Ritzenthaler and Hofmann, 2007;Schmitt-Keichinger et al., 2017). Therefore, both MPs and coat proteins (CPs) are required for the viruses with this group of MPs to move intercellularly. MPs of some icosahedral viruses, such as nepo-and comoviruses, belong to this group.
Potyviruses represent the largest group of known plant RNA viruses, including many agriculturally important viruses such as Turnip mosaic virus (TuMV), Plum pox virus (PPV), Soybean mosaic virus (SMV), and Potato virus Y (PVY) (Revers and García, 2015;Wylie et al., 2017;Cui and Wang, 2019;Gibbs et al., 2020). Potyviruses have a positive-sense, single-stranded RNA genome of approximately 10,000 nucleotides that encodes a long open reading frame (ORF) and additional small ORFs resulting from RNA polymerase slippage during viral genome replication (Olspert et al., 2015;Revers and García, 2015;Rodamilans et al., 2015;Hagiwara-Komoda et al., 2016;Cui and Wang, 2019). The polyproteins encoded by these ORFs are processed co-and post-translationally into over 10 mature proteins, including P3N-PIPO and P3N-ALT. Among them, the viral protein P3N-PIPO encoded by a small ORF is a dedicated MP (Cui et al., 2017). P3N-PIPO is a PD-located protein and directs the viral cylindrical inclusion protein (CI) to form conical structures at PD to assist potyviral intercellular movement (Wei et al., 2010). In addition to P3N-PIPO and CI, potyviral CP has also been shown to be essential for potyviral cell-to-cell movement (Dolja et al., 1994(Dolja et al., , 1995Arazi et al., 2001;Kimalov et al., 2004;Seo et al., 2013;. The potyviral CP encompasses a variable N-terminal domain exposed on the virion surface that is susceptible to trypsin treatment, a conserved core domain that interacts with viral RNA and forms the core subunit structure of the virion, and a C-terminal domain that has been implicated in CP-vRNA binding (Zamora et al., 2017;Cuesta et al., 2019;Kezar et al., 2019). Deletion of CP or its important domains or mutation of key charged residues within these domains abolishes potyvirus intercellular movement or systemic infection (Dolja et al., 1994(Dolja et al., , 1995Seo et al., 2013;Kezar et al., 2019). The fact that CPs are required for potyviral intercellular movement raises the possibility that potyviral intercellular spread may occur in the form of virions. To date, several studies have been devoted to better understand the mechanisms underlying the requirement of potyviral CPs in viral movement. Some resulting data are inconsistent among potyviruses. For instance, it has been suggested that the N-terminal region is not required for zucchini yellow mosaic virus (ZYMV) to establish local and systemic infection (Arazi et al., 2001). However, deletion of this region from PVY and tobacco etch virus (TEV) reduces viral local infectivity and abolishes viral systemic infection (Dolja et al., 1994;Kezar et al., 2019). In the case of Wheat streak mosaic virus (WSMV), a member of the genus Tritimovirus in the family Potyviridae, this region is not essential for long-distance movement but is required for efficient cell-to-cell movement . In addition, a few other studies have reported that the potyviral replication complex (VRC) can be found in the extracellular space or move to the adjacent cells (Grangeon et al., 2012(Grangeon et al., , 2013Movahed et al., 2019), implying that potyviruses may move intercellularly in the form of vRNPs.
In this study, we constructed a double-fluorescent protein-tagged TuMV infectious clone that can unambiguously distinguish primary and secondary infected cells to investigate the role of TuMV CP in viral intercellular movement. We confirmed that TuMV CP is not required for viral replication but is indispensable for viral cell-to-cell movement and systemic infection in plants. We further identified the regions and charged amino acids of TuMV CP that are required for TuMV to establish local and systemic infection. We present evidence that CP expressed in trans cannot rescue the movement defect of TuMV CP mutants, suggesting that the cis-expression of CP is essential for TuMV intercellular movement.

| TuMV CP is indispensable for viral cell-to-cell and systemic movement but not required for viral replication
To unambiguously determine viral cell-to-cell movement, we constructed a double-fluorescent (green and mCherry) protein-tagged TuMV infectious clone, which is designated pCBTuMV-GFP//mCherry and serves as a wild-type (WT) virus for this study (Figure 1a). In this vector, two gene expression cassettes are placed within the transfer DNA (T-DNA) borders: one for transcription of mRNA coding for mCherry fused with a luminal endoplasmic reticulum (ER) retention signal (mCherry-HDEL) and the other for the TuMV genome tagged by green fluorescent protein (GFP). On agroinfiltration of this clone into Nicotiana benthamiana leaf cells, both GFP and mCherry fluorescent proteins are expected to be expressed in the primary infected cells, leading to the emission of green and red fluorescence signals. The secondary infected cells as a result of viral intercellular movement would emit green fluorescence only as the recombinant TuMV genome contains the GFP sequence. After confirming that this WT clone allows for the differential visualization of primary and secondary infected cells (see below), we used it as a parental plasmid and constructed two additional clones. The first, ∆GDD, has a deletion in the coding sequence for the glycine-aspartic acid-aspartic acid (GDD) motif that is the active site of the RNA-dependent RNA polymerase (also NIb) (Shen et al., 2020). The second, ∆CP, was created by the deletion of almost the entire coding region for the CP except two short stretches essential for the cleavage sites ( Figure 1b). ∆GDD is a replication-defective mutant and serves as a control (Deng et al., 2015) ( Figure 1c (Figure 1f), suggesting no viral intercellular movement occurred for these two mutants. Potyviral cell-to-cell movement requires active genome replication so that a small percentage of viruses encoding P3N-PIPO, a dedicated movement protein, is generated (Cui et al., 2017).
As the ∆GDD mutant lost cell-to-cell movement ability, the expression of double fluorescent proteins in agroinfiltrated cells with ∆GDD could be due to the 35S promoter activity.
To determine if deletion of CP compromises viral replication leading to the inability to move between cells, we isolated mesophyll protoplasts from 4-week-old healthy N. benthamiana seedlings and conducted a protoplast transfection assay with ∆CP. Quantitative RT-PCR (RT-qPCR) analyses revealed that the level of either viral plus-strand or negative-strand RNA in ∆CPtransfected protoplasts 48 hrs post-transfection (hpt) did not significantly differ from that in WT-transfected protoplasts but was significantly higher than that in the protoplasts transfected with ∆GDD ( Figure 1e). Taken together these data suggest that TuMV CP is essential for viral intercellular and systemic movement but is not required for viral replication.

| Identification of CP segments required for viral cell-to-cell movement
The TuMV CP comprises 288 amino acid residues with a molecular mass of approximately 33 kDa. Based on the recently released atomic model (PDB: 6T34) (Cuesta et al., 2019), TuMV CP is divided into the N-terminus (N, amino acids 1-97), core (amino acids 98-245), and C-terminus (C, amino acids 246-288) domains ( Figure 1b). To identify the CP segments that are essential for TuMV cell-to-cell movement, we created four partial CP deletion mutants using the parental WT virus (pCBTuMV-GFP//mCherry) ( Figure 1b and Table 2 all CP mutants and controls at 14 dpi, and then analysed by RT-PCR with TuMV-specific primers. Consistently, viral RNA was detected from the Δ6-50 sample but not from the Δ51-199, Δ200-283, or Δ265-274 samples ( Figure 1d). Next, we conducted a protoplast transfection assay to check the replication capacity of these deletion mutants in the N. benthamiana protoplasts. At 48 hpt, the viral RNA levels of CP deletion mutants were comparable to that of the WT (Figure 1e). These data further confirm that CP is dispensable for TuMV replication.
We then examined the cell-to-cell movement ability of the mutants. The N. benthamiana leaves agroinfiltrated with each of these mutants were subjected to confocal microscopy at 4 dpi. In leaf tissues agroinfiltrated with the mutant ∆6-50, we found isolated individual cells emitting double fluorescent signals and also large foci of green fluorescence, similar to what was observed for the WT (Figure 1f). However, in the agroinfiltrated leaf regions with Δ51-199, Δ200-283, or Δ265-274, we observed isolated cells highlighted by double fluorescence but did not detect any cells emitting green fluorescence only ( Figure 1f). These data suggest that similar to ∆CP, mutants Δ51-199, Δ200-283, and Δ265-274 lose the ability to move from the primary infected cells to neighbouring cells.

| Deletion of the segment amino acids 6-50 of the N domain results in the production of aberrant viral particles
Because the N-terminal region of some potyviral CPs such as PPV is known to be involved in the CP-CP interaction, which is important for virion assembly (Zilian and Maiss, 2011), we determined whether ∆6-50 can form regular virions in plant cells. Crude virus extracts were prepared from N. benthamiana leaves systemically infected by WT or to TuMV CP detected a protein of slightly larger than 33 kDa from the WT sample and a protein of about 28 kDa from the ∆6-50 sample, consistent with the predicted molecular mass for WT CP and ∆6-50, respectively ( Figure 2a). The crude virus extracts were also subjected to negative staining and transmission electron microscopy (TEM) analysis.
Numerous typical flexuous rod-shaped viral particles of 600-900 nm in length were easily observed in the WT sample ( Figure 2b). In addition to the typical TuMV particles as shown in the WT sample, a small percentage of particles in the ∆6-50 sample had remarkably elongated length F I G U R E 2 Effects of deletion of a coat protein (CP) segment (amino acids 6-50) on virion assembly and analysis of the self-interactions of different CP regions. (a) Immunoblotting analysis of total proteins extracted from TuMV wild-type (WT) or Δ6-50-infected plants at 12 days postinoculation (dpi). Total protein extracts were probed with TuMV CP antibody. (b) Transmission electron microscopy (TEM) analysis of virion assembly of Δ6-50 mutant. Crude virion preparations were obtained from the symptomatic tissue of WT or Δ6-50-infected plants at 12 dpi. The preparation was subjected to negative staining and TEM. (c) Analysis of the self-interactions of different CP regions. Top panel: schematic representation of truncated portions for protein-protein interaction assays. Middle panel: yeast two-hybrid (Y2H) analysis in yeast. Yeast competent cells co-transformed with bait and prey plasmids were plated on double dropout (DDO) medium lacking tryptophan and leucine to test for double transformation, and on quadruple dropout (QDO) medium lacking tryptophan, leucine, histidine, and adenine for protein-protein interaction. Yeast co-transformed with AD-CP and BD-CP serves as a positive control. Yeast co-transformed with AD-CP and empty BD plasmid serves as a negative control. Bottom panel: bimolecular fluorescence complementation analysis of CP-truncated mutants in Nicotiana benthamiana plants. Combination of CP-YN/CP-YC serves as a positive control and CP-YN/YC as the negative control. Experiments were repeated three times. Bars = 40 µm ( Figure 2b). To quantify the distribution of aberrant particles in the ∆N samples, we measured 120 particles (randomly selected) and found 5% of particles showed the elongation phenotype with an average length of 1,500 nm. No such elongated particles were observed in the WT sample. These findings suggest that though the N-terminal segment amino acids 6-50 of the N domain is probably involved in TuMV virion assembly, deletion of this region is not sufficient to completely disrupt the formation of morphologically typical viral particles.
Because self-interaction is required for virion assembly, we generated four truncated CP mutants ( Figure 2c) and examined their ability to self-interact using both yeast two-hybrid (Y2H) and bimolecular fluorescence complementation (BiFC) assays. Y2H results revealed that CP200-288 did not self-interact, whereas CPwt, CP1-50, CP1-199, and CP51-288 did ( Figure 2c). Consistent results were obtained with the BiFC assay in planta ( Figure 2c).These data suggest that the C domain is unlikely to be involved in virion assembly through self-interaction whereas other regions probably are.

| Identification of residues essential for TuMV infection
Given the fact that the CP fragment encompassing the C-terminal segment of the core domain and the entire C domain lacks self-interaction ability ( Figure 2c) but is essential for viral intercellular moment  (Figure 3d,e). Obviously, the group III mutants lost the ability to move systemically.

| Compromised viral infectivity of CP point mutants is attributed to deficient cell-to-cell movement
To exclude the possibility that viral replication is affected in the CP point mutants, we conducted a protoplast transfection assay. The plus-strand viral RNA level was quantified by RT-qPCR at 48 hpt. All CP point mutants replicated well as their viral RNA levels showed no significant difference with that of the WT virus ( Figure 4a).
We then checked the ability of these mutants to move from the primary infected cells to the neighbouring cells by confocal microscopy. Different from the group I and II mutants, the group III mutants failed to move between cells and showed the phenotype of abolished cell-to-cell movement as none of cells emitting green fluorescence only could be found (Figure 4d). This explains why the group III virus completely lost the ability to establish systemic infection.

| Effects of CP point mutations on virus assembly
As potyviral cell-to-cell movement is often correlated with virion assembly (Dolja et al., 1994(Dolja et al., , 1995Seo et al., 2013), we examined if the Because the group III mutants lost the ability to move between cells, we transfected R178A, D222A, and WT into N. benthamiana mesophyll protoplasts. The transfected protoplasts were subjected to immunoblotting and TEM analyses. As expected, the CP protein and typical virions were detected from the sample prepared from the protoplasts transfected with WT (Figure 5b,c). However, no virions were found from the protoplasts transfected with R178A and D222A ( Figure 5b) and the CP level of these two mutants was remarkably low compared to that of WT (Figure 5c). These data suggest that the cell-to-cell movement-defective mutants R178A and D222A cannot form regular viral particles and their CPs are probably not stable.

| CP stability is essential for cell-tocell movement
The fact that the group III mutants R178A and D222A retain replication competency comparable toWT (Figure 4a) but accumulate very low levels of CPs prompted us to hypothesize that the movementdefective phenotype of the group II and III mutants is associated with the instability of the CP protein. To test this, we conducted a transient expression assay to detect CP accumulation levels. The CP coding sequence of WT and that of the movement-defective mutants R178A, D222A, E268A, R269A, and ER were cloned into the Gateway binary vector pEarleyGate100, which does not contain a fluorescent protein tag, and the resulting expression vectors were agroinfiltrated into N. benthamiana leaves for transient expression.
Total proteins were extracted from the agroinfiltrated leaf areas 48 hr postinfection (hpi) for detection of CP by immunoblotting. In comparison with the CP of WT, the CPs of all movement-defective mutants accumulated at much lower levels (Figure 6a), suggesting a possible link between CP stability and viral intercellular movement.
In plants, most proteins are degraded by the ubiquitin/26S proteasome system, which represents the major protein degradation pathway. To test if this pathway is involved in depleting CP with point mutations, N. benthamiana leaves transiently expressing the CPs of movement-defective mutants were treated with MG132 (carbobenzoxy-Leu-Leu-leucinal), a proteasome inhibitor that is known to block the proteolytic activity of the 26S proteasome complex.
Proteins were then isolated, followed by immunoblotting analyses.
MG132 treatment effectively increased the CP accumulation levels of R178A, D222A, E268A, R269A, and ER ( Figure 6b). These data suggest that the proteasome-ubiquitin pathway contributes to the CP degradation, and residues R178, D222, E268, and R269 are crucial for CP stability in plants.

| CP expressed in cis not in trans functions in viral cell-to-cell and systemic movement
To test whether WT CP can rescue the movement-defect of the mutants ∆CP, R178A, and D222A, we generated transgenic CP   To further examine if this also holds true in the context of viral infection, we checked whether the movement-defective phenotype of the TuMV CP mutants could be rescued by a wild-type TuMV.
A wild-type TuMV without any fluorescent protein tag (TuMV-WT) was agroinfiltrated into N. benthamiana leaves together with one of the three movement-defective mutants ΔCP, R178A, or D222A.
Confocal microscopy analysis revealed that TuMV-WT failed to restore the movement defect of the mutants in the inoculated leaves ( Figure 8). In these plants, the movement-defective mutants could not be detected in the upper newly emerged leaves (Figure 8). Taken together these data suggest that the TuMV CP functions in cis for viral intercellular movement.

| The cis-expression of CP is essential for TuMV translocation in plants
In this report, we present evidence that deletion of TuMV CP or important segments does not affect viral replication but abolishes viral cell-to-cell and long-distance movement in plants (Figure 1). We show that mutations of some charged residues conserved in these segments among potyviruses also result in a similar movement-defective phenotype (Figure 3). Overall, these data are consistent with recent and previous findings for other potyviruses such as TEV, PVY, and SMV that potyviral CP is required in virus translocation or systemic infection in plants (Seo et al., 2013;Kezar et al., 2019).
To examine if CP provided in trans could rescue the movement defect of TuMV CP deletion or point-mutation mutants, we generated transgenic Arabidopsis plants overexpressing TuMV CP and then inoculated them with the movement-defective TuMV CP mutants. We confirmed CP expression in these plants (Figure 7), and observed VLPs in these transgenic plants, in agreement with recently published data (Cuesta et al., 2019). To our surprise, we found that overexpression of CP failed to rescue these movement-defective mutants (Figure 7). We further revealed that CP provided by a wild-type TuMV did not recover the movement defect of all the three CP mutants tested (Figure 8).
Altogether these data clearly demonstrate that for TuMV intercellular movements, only the cis-expressed CP is functional. Inconsistent with our results, an earlier study with a recombinant TEV tagged by the reporter protein β-glucuronidase (TEV-GUS) showed that overexpression of TEV CP in transgenic Nicotiana tabacum plants partially restores the movement defect of four CP mutants: an N-terminal CP deletion mutant, two single point-mutation mutants of the charged residues in the core domain, and a double-mutation mutant of these two residues (Dolja et al., 1994

| The role of the N-terminal domain of TuMV CP in viral translocation and virion assembly
The N-terminal domain of potyviral CPs exposed on the surface of virus particles is highly variable among potyviruses and contains the DAG motif that is required for aphid transmission (Allison et al., 1985;Shukla et al., 1988;Dolja et al., 1991;López-Moya et al., 1999;Nigam et al., 2019). In this study, we found that the N-terminal segment amino acids 6-50 of the N-terminal domain of TuMV CP was dispensable for viral  (Kezar et al., 2019). Different from these, the N-terminal segment amino acids 6 to 27 of WSMV CP is not essential for long-distance movement but is required for efficient cell-to-cell movement . It seems that the requirement of the variable

| The role of the core domain of TuMV CP in viral intercellular movement and virion assembly
In this study, we revealed that alanine substitution of charged residues R178 and D222 in the core domain abolished viral cell-to-cell movement in N. benthamiana plants ( Figure 4d) and no virions were detected in N. benthamiana protoplasts transfected with these two point mutants (Figure 5b), suggesting that Arg at 178 and Asp at 222 are critical for TuMV virion assembly and intercellular movement.
Similar results have been observed in the corresponding analogous point mutants of a few other potyvirid species, including TEV (Dolja et al., 1994) and WSMV . It is tempting to speculate that there is a correlation between virion assembly and potyviral intercellular movement. This suggestion contrasts against the assumption made by Grangeon and colleagues that TuMV moves as VRCs for cell-to-cell movement, based on their observation that the motile 6K2-containing vesicles (VRCs) enable vRNA transport to PD by trafficking along ER/microfilaments and pass through the PD to the neighbouring cells (Grangeon et al., 2012(Grangeon et al., , 2013. However, our data cannot exclude the possibility that CP expressed in cis also supports TuMV and related viruses to move intercellularly as RNPs. It has been suggested that positively charged residue R178 may interact with negatively charged residues D222 via a salt bridge, which may be critical for protein stability, virion assembly, and viral cell-to-cell movement (Dolja et al., 1991). To test this hypothesis, we created a TuMV CP double mutant (DR) in which Arg178 and Asp222 were switched. Similar to the single mutants R178D and D222R, the double mutant DR failed to move in N. benthamiana plants (Figure 4d). As structural changes might occur after switching the two residues, our data could not completely reject the salt bridge hypothesis.

F I G U R E 8 Trans-complementation assay in
We speculated that the charged residues R178 and D222 might be important for proper folding on translation to form the appropriate functional three-dimensional structure and thus mutation of these two residues might affect CP stability. Indeed, immunoblotting analyses revealed that the CP accumulation levels of the mutants R178A and D222A in N. benthamiana leaves were remarkably reduced compared to that of WT CP (Figure 6a). Treatment with the proteasome inhibitor MG132 could partially inhibit R178 and D222 CP degradation ( Figure 6b), suggesting that these charged residues at least partially contribute to CP stability, which unavoidably affects CP functionality.
In a recent study, Gallo and colleagues revealed that the formation of stable PPV virions requires CP and other viral factors such as HC-Pro and a replication-proficient RNA (Gallo et al., 2018). Their data suggest a functional link between RNA replication and virion assembly.

Recent near-atomic structural analyses of PVY and TuMV virions and
VLPs strongly suggest that CP-RNA interaction is crucial for the helical configuration and stability of the virion (Cuesta et al., 2019;Kezar et al., 2019). Therefore, it is likely that TuMV virion assembly requires CP expression and viral replication in a tightly coordinated manner.

| The role of the C domain of TuMV CP in viral cell-to-cell movement
The   (Dolja et al., 1995;Seo et al., 2013;Kezar et al., 2019). Therefore, the C domain of potyviral CPs is crucial for viral cell-to-cell and long-distance movement.
How does the C domain function to support potyviral intercellular movement? Previous studies suggested that the C-terminus of potyviral CP is exposed on the virion surface and possesses TA B L E 2 Summary of the phenotypes of coat protein (CP) mutants in Nicotiana benthamiana a disordered short segment at the very end (Shukla et al., 1988;Zamora et al., 2017). Opposed to this, a recent study has determined the near-atomic structure of PVY virions and revealed that the C domain of PVY CP is completely buried in the lumen of the viral filament, forming a compact conical structure (Kezar et al., 2019). The C-terminal region of SMV CP was shown to be involved in CP intersubunit interactions, implying a possible role of this domain in virion assembly (Kang et al., 2006;Seo et al., 2013). However, TEV and PVY could form virions without the C-terminal domain (Dolja et al., 1995;Kezar et al., 2019), indicating that this domain is not necessary for virion assembly. In this study, we found that the C-terminal region (including the C domain) of TuMV CP did not self-interact (Figure 2c).
It is possible that the C domain of TuMV CP is required for CP-vRNA binding, as suggested in two independent cryoelectron microscopy structure studies of WMV and PVY virions (Zamora et al., 2017;Kezar et al., 2019). Alternatively, the C domain of TuMV CP may serve as a site for the CP to interact with other key players (such as host factors) to regulate TuMV cell-to-cell and systemic movement.
It is of great interest to isolate such factors and characterize their functional roles in viral intercellular movement.

| Plant materials, genetic transformation, and protoplast work
N. benthamiana and Arabidopsis thaliana ecotype Col-0 plants used for agroinfiltration were prepared as described previously . In brief, plants were grown in pots with  (Yoo et al., 2007;Wu et al., 2009). Transfection assay was carried out essentially as described previously (Deng et al., 2015;Wu et al., 2020).

| Construction of double fluorescent proteintagged TuMV infectious clone pCBTuMV-GFP// mCherry and CP mutants
The TuMV infectious clone pCamibaTunos/GFP (pCamTuMV-GFP) and the mini-binary vector pCB301 were reported previously (Xiang et al., 1999;Cotton et al., 2009). The GFP-tagged TuMV infectious clone pCBTuMV-GFP was constructed by double digestion of pCam-biaTunos/GFP with XmaI and ApaI, and ligation of the TuMV-GFP fragment into the corresponding sites of pCB301. Using a PPV infectious clone tagged with GFP and mCherry-HDEL (Cui and Wang, 2016) as a template, we amplified the 35S-mCherry-HDEL-Nos fragment and introduced an ApaI digestion site to this fragment by PCR.
The amplified fragment was digested with ApaI and inserted into the

| Cell-to-cell movement of TuMV CP mutants
Agrobacterial cells harbouring TuMV CP mutants were inoculated into 3-to 4-week-old N. benthamiana plants at OD 600 = 0.0001. The cell-to-cell movement ability of mutants was determined by confocal microscopy by monitoring the green fluorescent foci on the agroinfiltrated N. benthamiana leaf using a TCS SP2 confocal laser scanning microscope (Leica) starting at 2 dpi.
For trans-complementation assay in transgenic Arabidopsis lines, 3-week-old Arabidopsis plants expressing TuMV CP were agroinfiltrated with movement-defective TuMV mutants at an OD 600 value of 0.003 and cell-to-cell movement was monitored on the agroinfiltrated leaf area using a TCS SP2 confocal laser scanning microscope starting at 4 dpi and continually monitored each day until 10 dpi.
Long-distance movement was evaluated by visual observation for symptoms, green fluorescence under UV light, and RT-PCR. Three independent experiments were performed and each trial included at least five individual plants.
For trans-complementation assay with WT TuMV, two agrobacterial cultures harbouring TuMV-WT (tag-free) and one of the three mutants ΔCP, R178A, or D222A were mixed in three different ratios, 100:1, 10:1, and 1:1 (vol/vol), and the mixed cultures were infiltrated into levels of 3-to 4-week-old N. benthamiana plants. Coinfiltration of TuMV-WT and pCBTuMV-GFP//mCherry-WT served as the control. Cell-to-cell and long-distance movement of the mutants were evaluated as described above.

| Crude virion preparation and electron microscopy
To obtain the crude virion preparation from N. benthamiana leaf tissues, 0.1 g of leaves with symptoms were ground in a mortar and pestle with 150 µl of potassium phosphate buffer (pH 7.0) containing 0.1% β-mercaptoethanol. Following the centrifugation at 12,000 × g for 5 min at 4°C, the supernatant was filtered through a 40 µm Nylon Cell Strainer (Corning Inc.). Fifty microlitres of chloroform were then added and vortexed for 1 min, the homogenate was then centrifuged at 12,000 × g for 10 min at 4°C. The aqueous phase was collected and subjected to centrifugation at 12,000 × g for 30 min at 4°C. The resulting aqueous phase was used for negative staining. For crude virion preparation of N. benthamiana mesophyll protoplasts, protoplasts were harvest at 72 hpt by centrifugation at 50 × g for 2 min. Fifty microlitres of potassium phosphate buffer (pH 7.0) were added to the pellet, followed by the procedure presented above. Formvar/carbon-coated EM grids (Electron Microscopy Sciences) were incubated with the crude virion preparation for 2 min and stained with 2% phosphotungstic acid (PTA), pH 7.0, for 2 min. Grids were allowed to dry prior examination on a transmission electron microscope (JEM-1200EXII, JEOL Ltd) operated at 80 kV.

| RNA extraction and RT-qPCR
Total RNA was extracted from leaf tissues or protoplasts using the Plant Total RNA Mini Kit (Geneaid) as instructed. For firststrand cDNA synthesis, 1 μg of RNA was pretreated by DNase I (Invitrogen) at 37°C for 20 min. The reverse transcription reaction was performed using a SuperScript III First-Strand Synthesis System (Thermo Scientific) following the manufacturer's instructions. RT-qPCR analyses of TuMV RNA levels were performed as described previously (Wu et al., 2020).

| Y2H and BiFC assays
Y2H and BiFC assay were essentially as described .

| Protein work
Immunoblotting and relative quantification of proteins were performed as described .

| Chemical treatment
Chemical treatment with the proteasome inhibitor MG132 was performed essentially as described . In brief, 100 μM MG132 (Sigma-Aldrich) was infiltrated into N. benthamiana leaves for 12 hr before harvesting.

ACK N OWLED G M ENTS
We are indebted to Professor Jean-François Laliberté (Institut

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.