Gain-of-function mutant of movement protein allows systemic transport of a defective tobacco mosaic virus

Summary Functional compensation in response to gene dysfunction is a fascinating phenomenon that allows mutated viruses to regain the capabilities of their wild-type parental strains. In this study, we isolated mutants of tobacco mosaic virus capable of CP-independent systemic movement. These gain-of-function mutants lacked the 16 C-terminal amino acids of the movement protein (MP). Whereas this deletion did not affect the cell-to-cell movement of MP, it dramatically enhanced the viral genomic RNA levels and MP accumulation within the infected cells and altered the subcellular localization of MP from exclusively plasmodesmata (PD) to both PD and plasma membrane. The adapted defective virus suppressed the expression of the ethylene pathway and phloem-associated resistance factors in the inoculated leaves. These findings demonstrate the potential for plant viral MPs to gain a new function that allows viral genomes to move systemically in the absence of the natural viral factor that mediates this spread.

The mutation of MP did not change its cell-to-cell moment but subcellular localization The mutation of MP enhanced the viral genomic RNA levels and accumulation of MP TMVDCPmutMP locally suppressed expression of ET pathway and phloemassociated resistance

INTRODUCTION
Functional compensation in response to gene dysfunction is a common phenomenon in many organisms. The fitness losses caused by gene mutations can be buffered or compensated by genetic redundancy in which no or little effect of the mutations is evident due to the same or similar function of one or several other genes (Rutter et al., 2017). This mechanism is rare in RNA viruses, because of their need for compression of genome size leading to virtually no duplicated sequences, fewer control elements, and overlapping reading frames (Krakauer, 2000;Simon-Loriere and Holmes, 2013). Instead, many viruses have evolved the capacity for high mutation rates, leading to numerous variant genomes (Elena et al., 2006); some of these mutations are compensatory, and they often result in the recovery of the wild-type-like phenotypes (Rokyta et al., 2002;Seki and Matano, 2012).
Most plant viruses spread within their hosts by a bimodal process, i.e., by local and systemic movement. Local infection is mediated by cell-to-cell movement, in which the virus moves from the infected to healthy cells through plasmodesmata (PD), the plant intercellular connections (Tabassum and Blilou, 2022). Once the local infection reaches the plant vascular system, systemic movement ensues to spread the viral infection to distant plant tissues (Kappagantu et al., 2020). Tobacco mosaic virus (TMV)-the first virus discovered and, since then, one of the paradigms for plant viruses-has a small, 6.4 kb, positive-sense RNA genome encoding two overlapping replicases, a cell-to-cell movement protein (MP), and a coat protein (CP) (Scholthof et al., 1999). MP is thought to mediate the local spread of the virus whereas CP is absolutely required for the systemic movement (Hipper et al., 2013). Indeed, in absence of CP, the virus can only replicate and spread locally by the cell-to-cell movement mechanism (Hilf and Dawson, 1993;Ryabov et al., 1999;Venturuzzi et al., 2021).

Identification of the gain-of-function viral mutants with restored systemic movement
To identify an adaptive TMV mutant with capability for systemic movement, we inoculated by agroinfiltration in Nicotiana benthamiana plants with an infectious TMV clone, pTMVDCP G, that lacks CP-and thus is unable to spread systemically-and expresses free GFP to facilitate detection of viral spread. Confirming the inability of TMVDCP G to move systemically, in most of these inoculations (96%), the virus did not spread beyond the inoculated leaves. However, we identified two independently inoculated plants that developed relatively severe systemic symptoms of the viral disease, e.g., leaf curling, shoot stunting, and leaf distortion, in their uninoculated, apical leaves ( Figure 1A). Interestingly, the occurrence of the disease symptoms in the systemic leaves was much more prevalent than the detectible accumulation of the GFP signal in the same leaves ( Figure 1A), suggesting the loss of GFP expression during adaptation. Sequence analyses of the viral genomic region that includes the GFP expression cassette from both systemically moving isolates revealed the complete loss of the GFP coding sequence and of the sequence LIDDDSEATVAESDSF-the 16 C-terminal amino acid residues of MP (16-aa C-terminus) (Figures 1B and S1A). These spontaneous mutants were designated TMVDCPmutMP and TMVDCPmutMP2, both of which lost the 16-aa C-terminus but the TMVDCPmutMP2 also gained 14 new residues from the native 3 0 -untranslated region of the TMV genome ( Figure 1B, asterisk). On the genomic RNA level, the mutMP mutation in TMVDCPmutMP did not interfere with most of the cis-acting elements that remained in the parental strain TMVDCP, i.e., the MP subgenomic promoter, the 3 0 UTR upstream pseudoknot domain, and the 3 0 UTR tRNA-like structure ( Figure S1B) (Grdzelishvili et al., 2000;van Belkum et al., 1985;Zeenko et al., 2002). The only cis-acting element affected by mutMP was the CP subgenomic promoter (Grdzelishvili et al., 2000), which most likely is not biologically relevant for TMVDCP or TMVDCPmutMP because these virus variants have no coding sequences to be transcribed from this promoter ( Figure S1B).
To examine whether these MP mutants indeed represent the causative agents of the systemic symptoms, we reconstructed each of them in a binary vector and evaluated the infectivity of the resulting clones, designated pTMVDCPmutMP and pTMVDCPmutMP2, in N. benthamiana. One week after inoculation, both mutants consistently developed severe systemic symptoms ( Figure 1C). These observations indicate that deletion of the 16-aa C-terminus of MP is sufficient to confer systemic movement ability on the CP-defective TMV, suggesting that the 16-aa C-terminus deletion, termed mutMP, represents a gain of function mutation. Thus, we used the pTMVDCPmutMP mutant for further characterization.
Effects of the mutMP mutation on the cell-to-cell movement and subcellular localization of MP TMV MP itself can move between plant cells without the presence of the viral RNA (Crawford and Zambry ski, 2000). To assess whether the mutMP mutation alters this function, the wild-type MP and mutMP were tagged with CFP and transiently expressed in N. benthamiana leaf epidermis following agroinfiltration. Expression of both MP variants produced a CFP signal in single-cell clusters at 36 h after transfection. Two days after transfection, the cell-to-cell movement of MP was observed as the appearance of R2 cell clusters that accumulated CFP (Figure 2A, left panel). Quantification of the numbers of such clusters did not detect statistically significant differences between the wild-type MP and mutMP in the cell-to-cell movement frequency (Figure 2A, right panel). Thus, mutMP produced no detectable effects on the cellto-cell movement of the MP protein.
We then examined whether mutMP affects the subcellular localization of MP. To this end, we transiently coexpressed CFP-tagged MP and mutMP in N. benthamiana leaf epidermal cells with different fluorescently tagged subcellular localization markers, e.g., PDCB1-mRFP which represents a PD marker (Figures 2B and S2A) or BAM1-mRFP which represents a plasma membrane marker ( Figures 2B and S2B). As expected, MP-CFP exhibited a predominantly punctate appearance diagnostic of PD (Yuan et al., 2018) and colocalized with PDCB1-mRFP but not with upper rows). In contrast, mutMP-CFP was located at both PD and the plasma membrane, colocalizing with their respective marker proteins ( Figures 2B, S2A, and S2B, lower rows). We did not observe colocalization of MP and mutMP with the ER ( Figure S2C), cell wall ( Figure S2D), or nucleocytoplasmic markers ( Figure S2). These observations suggest ll OPEN ACCESS Figure 1. Identification and reconstruction of the TMVDCPmutMP mutants (A) Experimental screening system to identify TMVDCP mutants capable of systemic movement. Two-week-old N. benthamiana plants were agroinfiltrated with pTMVDCP G which expresses GFP, a viral movement marker. Local infection of the viral vector was confirmed at 3 dpi by expression of GFP as detected under UV light. The inoculated plants were monitored until 30 dpi to identify plants that developed systemic viral disease symptoms. RNA was extracted from the tissues from the plant with the most severe symptoms, followed by cDNA synthesis, sequencing, and re-construction of the recovery mutant. (B) Schematic diagrams of the binary vectors pTMV, pTMVDCP, pTMVDCPmutMP, and pTMVDCPmutMP2 with the indicated locations of the Agrobacterium T-DNA left border (LB) and right border (RB) sequences, and the CaMV 35S promoter (p35S) and terminator (35S ter). The genome of TMV contains the indicated untranslated regions (UTRs) and open reading frames of the RNA-dependent RNA polymerase (RdRp), movement protein (MP), and coat protein (CP). The coding sequence for CP is absent from the genome of pTMVDCP. Stop codons that terminate translation of MPs in the three viral mutants, pTMVDCP, pTMVDCPmutMP, and pTMVDCPmutMP2, are indicated. The location and sequence of the 16-aa C-terminal region of MP which is present in pTMVDCP but absent in pTMVDCPmutMP and pTMVDCPmutMP2 is indicated. Asterisk indicates the location of the new 14-aa C-terminal sequence gained by pTMVDCPmutMP2 from the native 3 0 UTR as a result of the corresponding relocation of the stop codon due to the mutMP2 mutation. Effects of the mutMP mutation on the accumulation of the total viral RNA and MP in the inoculated and systemic leaves Does mutMP affect the accumulation of viral RNA and MP in plant tissues? We investigated this question by quantifying the accumulation of the total viral RNA in the leaves inoculated with pTMVDCPmutMP, its parental strain pTMVDCP, and the wild-type virus as a positive control. The total viral RNA comprises the full-length, genomic RNA of the virus as well as its subgenomic RNA transcribed from the subgenomic promoters ( Figure S1A); thus, we performed the qRT-PCR analysis using the primers specific for the MP sequence (Table S1) which detect both the genomic viral RNA and the MP subgenomic RNA species ( Figure S1B). To avoid the interference of local response at R5 days post-inoculation (dpi) with pTMVDCP ( Figure 3A), these experiments were performed at 4 dpi. Figure 3B (left panel) shows that, relative to pTMVDCP, inoculation with pTMVDCPmutMP resulted in ca. 12-fold higher levels of the total viral RNA, comparable to the amounts of the total viral RNA accumulated in the pTMV-inoculated leaves. At the same infection time point in the same leaves, our Western blot analysis revealed substantially higher levels  Table S1). Left panel. Inoculated leaves at 4 dpi. Right panel. Systemic leaves at 14 dpi. Viral RNA accumulation in leaves inoculated with pTMVDCP was set as 1.0. The resulting values were expressed as mean G SE (n = 6). Individual data points are represented by black dots and their numerical values are listed in Table S3. Differences between mean values assessed by the two-tailed t-test are statistically significant for the p-values *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001; p R 0.05 are not statistically significant (ns). iScience Article of MP accumulation in the pTMVDCPmutMP-inoculated leaves as compared both to the pTMVDCP and pTMV inoculations ( Figure 3C, left panel), suggesting the mutMP mutation may increase the stability of MP, known to undergo rapid turnover in the infected cells (Szé csi et al., 1999).
We then considered the effect of the mutMP mutation on the process of systemic infection by analyzing the levels of the total viral RNA and MP levels in the systemic leaves at 14 dpi (Figures S3). Figure 3B (right panel) shows that, as expected, the total viral RNA of pTMVDCP, which does not move systemically, was detected in the systemic leaves only at the background level whereas the RNA of pTMV, also as expected, accumulated to very high levels. The total viral RNA of pTMVDCPmutMP accumulated in the systemic leaves but at levels 3-4 times lower than those of pTMV ( Figure 3B, right panel). This difference may be due to the presence of CP in systemic leaves infected by the wild-type virus which may associate with and protect the viral RNA molecules (Ivanov and Mä kinen, 2012) whereas TMVDCPmutMP does not encode CP. The Western blot analysis of the same leaves showed that infection with pTMVDCPmutMP resulted in a very substantial accumulation of MP which even exceeds that produced by the systemic pTMV infection. Obviously, no MP was observed in the systemic leaves of the plants inoculated with pTMVDCP that does not spread systemically ( Figure 3C, right panel).
Next, we focused our analysis on the genomic viral RNA, without the subgenomic species. To this end, we utilized the primers specific for RdRp (RNA-directed RNA polymerase) (Table S1) which detect only the genomic viral RNA species ( Figure S1B). Similarly, to the accumulation of the total viral RNA (see Figure 3B), the accumulation of the pTMVDCPmutMP genomic RNA in the inoculated leaves was 10-fold higher than that of pTMVDCP and comparable to that of pTMV ( Figure S1C, left panel). Also, in the systemic leaves, the accumulation pattern of the genomic viral RNA mirrored the accumulation of the total viral RNA (compare right panels in Figures S1C and 3B) although we did not observe statistically significant differences between pTMVDCPmutMP and pTMV. Thus, the increase in viral RNA accumulation in the pTMVDCPmutMP-infected local and systemic tissues is general and does not reflect a possible specific increase in the accumulation of the subgenomic MP RNA.
Besides the MP subgenomic promoter, the TMV genome contains the CP subgenomic promoter, although the CP gene itself is absent in pTMVDCPmutMP and in its parental pTMVDCP strain. ( Figure S1B) Thus, to assess the possible effect of the mutMP mutation on the accumulation of transcripts produced from the CP subgenomic promoter, we analyzed the amounts of 3 0 UTR RNA which is located downstream from the CP gene ( Figure S1B) and derives largely from the activity of the CP subgenomic promoter (Grdzelishvili et al., 2000). Figure S1D shows that the inoculated leaves infected with pTMVDCPmutMP accumulated lower amounts of the 3 0 UTR-specific viral RNA than the leaves infected with pTMV or pTMVDCP. The most striking difference was observed in the systemic leaves where the infection with pTMV produced ca. 30-fold more 3 0 UTR-specific viral RNA than the infection with pTMVDCPmutMP ( Figure S1D, right panel). These observations indicate that the mutMP mutation had indeed compromised the activity of the CP subgenomic promoter, with the residual 3 0 UTR-specific transcript most likely generated from the genomic and MP subgenomic promoters.
Finally, our analyses of the viral RNA and the MP protein accumulation in the systemic leaves were confirmed and extended by analyzing the content of the viral RNA-MP complexes. TMV MP is well known to associate with the single-stranded nucleic acids (Brill et al., 2000;Citovsky et al., 1990Citovsky et al., , 1992. Thus, we immunopurified MP from the systemic leaves and analyzed it for the presence of the MP-associated viral RNA relative to the total MP accumulated in the infected cells ( Figure S4A). Figure S4B shows that the pTMVDCPmutMP-infected systemic leaves accumulated ca. 40-fold higher amounts of the viral RNA than the leaves infected by pTMV, consistent with much higher amounts of MP found in these leaves (see Figure 3C, right panel). Potentially, CP of pTMV can encapsidate the viral RNA, thereby reducing its association with MP, whereas this RNA sequestration does not occur with the pTMVDCPmutMP infection where CP does not exist.

Suppression of the ethylene signaling factors by TMVDCPmutMP
Incompatible interactions between viruses and plants often culminate with a hypersensitive reaction or cell death-like response at the infection loci (Garcia-Ruiz, 2019). Indeed, the local infection of pTMVDCP, but not by pTMVDCPmutMP or pTMV, in N. benthamiana resulted in tissue necrosis that led to partial or complete necrosis and shedding of the inoculated leaves ( Figure 3A)  iScience Article and S3). This reaction most likely represents the antiviral response of the plant, and this response was less efficient against pTMVDCPmutMP and pTMV than against pTMVDCP. That the main functional difference between these viral strains is their capacity or the lack thereof to move systemically suggests that it is the viral factor that allows the systemic movement, i.e., mutMP or CP, that counteracts the resistance. Thus, we examined whether mutMP-which suppressed the necrosis response and helped pTMVDCPmutMP escape the inoculated leaves-can suppress the antiviral signaling of the host. The results of these experiments with pTMVDCPmutMP and pTMV were compared to pTMVDCP, the parental strain of pTMVDCPmutMP, and, therefore, the point of reference for the effects of the mutMP mutation.
For TMV, signaling pathways mediated by salicylic acid (SA), jasmonic acid (JA), and ethylene (ET) are known to regulate the resistance of N. benthamiana to the virus (Zhu et al., 2014(Zhu et al., , 2022) ( Figure 4A). Figures 4B and 4C show that the expression levels of NPR1 and COI1-key genes of the SA and JA pathways, respectively-were not significantly altered in tissues infected by pTMVDCPmutMP relative to pTMVDCP at 4 dpi, i.e., before the onset of necrosis. However, the expression of EIN2, one of the key genes of the ET signaling pathway, was strongly and in a statistically significant fashion suppressed by both pTMVDCPmutMP and pTMV in comparison to pTMVDCP ( Figure 4D).
These effects also were observed for several other ET signaling pathway-related genes ( Figure 4E), i.e., ACO1 and EIN3. Specifically, pTMVDCPmutMP and pTMV exhibited various and statistically significant degrees of suppression of the ACO1 and EIN3 genes relative to pTMVDCP (Figures 4F and 4G). Consistent with their suppressive effects on positive regulators/components of the ET signaling pathway, both pTMVDCPmutMP and pTMV induced the expression of ETR1, a negative regulator of the ET signaling, as compared to pTMVDCP ( Figure 4H). Figure 4 also shows that all tested genes were expressed, to varying degrees, in tissues that were ''mock''-inoculated with the buffer and have not undergone agroinfiltration and have not encountered the virus. Together, these data indicate that pTMVDCPmutMP suppressed several key factors of the ET pathway, but not of the SA or JA pathways, within the inoculated leaves.

Suppression of the phloem loading/unloading factors by TMVDCPmutMP
Loading into the phloem and unloading into systemic tissues are the key steps of the systemic spread of the virus after it reaches the vascular system of the inoculated organ. The processes of entry into and egress from the phloem involve several host genes, such as PLM1, GSD1, and cdiGRP ( Figure 5A) (Gui et al., 2014;Ueki and Citovsky, 2002;Yan et al., 2019). We compared the effects of the systemic movementcapable pTMVDCPmutMP and pTMV viruses on the expression of these genes relative to the systemic movement-incapable pTMVDCP virus. PLM1 encodes a sphingolipid biosynthetic enzyme, the absence of which increases the phloem unloading (Yan et al., 2019). Figure 5B shows that pTMVDCPmutMP and pTMV did not significantly alter the expression of PLM1 observed in the absence of the virus, but pTMVDCP strongly activated it. GSD1 encodes a remorin-like protein, the enhanced expression of which impairs transport into the phloem (Gui et al., 2014). Figure 5C shows that pTMVDCPmutMP suppressed the GSD1 expression relative to pTMVDCP whereas pTMV had no statistically significant effect. Finally, cdiGRP codes for a glycine-rich protein that inhibits systemic movement of tobamoviruses (Ueki and Citovsky, 2002), and, again, only pTMVDCPmutMP, but not pTMV, suppressed the expression of cdiGRP with statistical significance relative to pTMVDCP ( Figure 5D).
Callose production by callose synthases and its deposition in the sieve plate is another factor that can interfere with the systemic movement of plant viruses (Wang et al., 2021) ( Figure 5E). Thus, we determined the expression levels of three callose synthase genes, CALS3, CALS7, and CALS8, at 4 dpi in the leaves inoculated with pTMVDCPmutMP, pTMVDCP, or pTMV. These experiments showed that pTMVDCP induced the expression of CALS3 ( Figure 5F) and CALS8 ( Figure 5G) and did not significantly affect the expression of CALS7 ( Figure 5H). Conversely, pTMVDCPmutMP suppressed the expression of CALS3, CALS8, and CALS7 by 3.5-fold, 3.2-fold, and 8.3-fold relative to pTMVDCP, respectively, in a statistically significant fashion ( Figures 5F-5H). pTMV also inhibited the expression of CALS3 and CALS7 (Figures 5F and 5H) but did not affect the expression of CALS8 with statistical significance relative to pTMVDCP ( Figure 5G). Taken together, these results suggest that mutMP may act to suppress numerous and diverse host factors that negatively affect the phloem loading/unloading.   Tables S4 and S5. Differences between mean values assessed by the two-tailed t-test are statistically significant for the p-values *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001; p R 0.05 are not statistically significant (ns). iScience Article potential effects of pTMVDCPmutMP, pTMVDCP, or pTMV inoculation on the expression of ISE1 and ISE2 in the inoculated leaves. Figure S5 shows that both genes were expressed at significantly higher levels in pTMVDCP-inoculated leaves whereas inoculation with pTMVDCPmutMP or pTMV had no statistically significant effects on the expression of ISE1 ( Figure S5A) and ISE2 ( Figure S5B). These results suggest different modes of regulation of the ISE1 and ISE2 expression upon incompatible (pTMVDCP) or compatible (pTMVDCPmutMP and pTMV) interactions between the plant host and the invading viral pathogen. Similarly, Figures 4 and 5 show the basal expression levels for all tested genes in ''mock'' inoculated tissues.  Tables S5-S7. Differences between mean values assessed by the two-tailed t-test are statistically significant for the p values *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001; p R 0.05 are not statistically significant (ns).

Lack of CP drives the evolution of the TMV genome and reveals the potential new function of MP
The members of the genus Tobamovirus known today have co-evolved with their hosts, and, through natural selective pressure, the viral proteins have gained and perfected the specific biological functions essential for their fitness in specific hosts (Gibbs, 1999). On the other hand, like most RNA viruses, the TMV replication is error prone which can lead to diverse variations in protein functions (Elena et al., 2006). Deletion of CP from the viral genome creates a novel and powerful selective pressure for the evolution of the defective TMV genome which has lost the capacity for systemic movement but remained able to replicate and accumulate potential gain-of-function mutations. Taking advantage of this unique experimental system, we demonstrated that the CP-deficient virus can evolve and recover the ability for systemic infection by a gain-of-function modification of its MP, mutMP, which was achieved through the loss of its 16 C-terminal amino acids. Normally, MP is not a direct participant in the systemic movement of the virus per se (Hilf and Dawson, 1993). What is the molecular and functional basis for this newly acquired systemic movement capacity of this gain-of-function MP mutant?
The 16-aa C-terminal domain of MP may serve as a negative regulator for the constitutive presence of MP in the host cells. Indeed, while the TMV RNA-dependent RNA polymerase (RdRp) and CP proteins are expressed throughout the course of infection, the presence of MP is only transient (Szé csi et al., 1999;Watanabe et al., 1984) most likely due to its destruction by the 26S proteasome (Reichel and Beachy, 2000).
Removal of this C-terminal domain substantially increases the accumulation of the resulting mutMP protein in both inoculated and systemic leaves, suggesting that mutMP may, at least in part, escape proteasomal degradation and that the deleted MP domain may contain post-translational modifications that signal degradation. Indeed, a deep-learning-based motif prediction (Wang et al., 2020) indicated that the mutMP mutation compromises a ubiquitination site at the amino acid residue K250 and abolishes four phosphorylation sites at the residues S258, T261, S265, and S267 ( Figure S6). Interestingly, phosphorylation is known to regulate the ubiquitin-mediated degradation of viral (Hé ricourt et al., 2000) and host proteins (Liu et al., 2009). Therefore, the complete lack of the ubiquitination and phosphorylation sites in the ''new'' C-terminus of mutMP may stabilize mutMP by rendering it less susceptible to the ubiquitin/proteasome system (UPS) of the cell.
In addition, phosphorylation of the MP residues S258, T261, and S265 was demonstrated directly and suggested to represent a mechanism to sequester and functionally inactivate MP (Citovsky et al., 1993); this notion is consistent with the observations that the C-terminal phosphorylation sites are dispensable for the cell-to-cell movement of MP in N. benthamiana [not shown and (Trutnyeva et al., 2005)]. Thus, deletion of the C-terminal domain may in fact further activate the protein and contribute to the altered function of mutMP.
The requirement for CP for their systemic movement has become an exclusive rule for most plant viruses (Hipper et al., 2013). A unique exception is the members of the genus Umbravirus, which naturally lack a CPencoding gene but are systemically infectious in the form of RNP complexes (Ryabov et al., 2001;Taliansky et al., 2003). Similarly, viroids represent such an exception for subviral agents which often exist in association with helper viruses, e.g., alpha-satellites can replicate their own genomes but depend on their helper begomoviruses for systemic infection (Badar et al., 2021); and tombusvirus-like associated RNAs are capable of autonomous replication but also depend on a virus of the genus Polerovirus as a helper for systemic movement and aphid transmission (Passmore et al., 1993). In our experimental system of the CP-deficient TMV genome incapable of systemic transport, we identified a subviral agent capable of replication and CP-independent systemic movement.

Adaptive virulence conferred by mutMP
Hypersensitive reaction (HR) is a typical response upon recognition by the host resistance (R) proteins of their corresponding viral factor, e.g., HR in response to the N protein against the helicase domain of the TMV replicase protein (Tran et al., 2014). When the R proteins are absent or insufficiently induced, systemic necrosis appears at later stages of infection (Abebe et al., 2021;Roshan et al., 2018). N. benthamiana does not carry the N gene but still develops a TMV-induced necrosis (Guo et al., 2015) which likely inhibits the virus multiplication (Komatsu et al., 2010). We observed that the local necrosis response in N. benthamiana against TMV is more severe in the absence of CP but is suppressed by the recovery mutant ll OPEN ACCESS iScience 25, 105486, December 22, 2022 iScience Article pTMVDCPmutMP or the presence of CP. Thus, besides gaining the function of CP in systemic movement per se, mutMP also evolved to exhibit a CP-like virulence function in mitigating the host immune response.
Interactions between plants and viruses usually result in the accumulation of SA, JA, or ET (Carr et al., 2010).
In N-mediated resistance against TMV, ET is highly accumulated and accelerates the HR (Knoester et al., 2001;Ohtsubo et al., 1999). The involvement of an ET-induced transcription factor in the resistance of N. benthamiana against TMV (Zhu et al., 2022) suggests that the ET signaling pathway also functions in the absence of the N gene. Expression profiling of signaling-related genes demonstrated that both mutMP and CP may act as counter-defense factors to suppress the components of the ET signaling pathway, but not of the JA or SA pathways, in local leaves, i.e., to downregulate the ET signaling positive regulators ACO1, EIN2, and EIN3, and to upregulate the negative regulator ETR1. These activities of mutMP could contribute to the delayed local necrosis and maintain conditions conducive to virus replication, as indicated by the higher local TMV RNA accumulation and successful systemic infection of pTMVDCPmutMP.
The host plant can further restrict the cell-to-cell movement of the virus from the infected into the neighboring uninfected cells during the viral approach to the leaf vein, representing yet another defense layer that prevents the viral systemic infection (Nyalugwe et al., 2016). Our analyses indicate differential transcriptional reprogramming of phloem-associated factors by the viruses with different abilities to move systemically. For example, pTMVDCP-that fails to move systemically-induces the expression of PLM1, a callose-independent negative regulator of phloem transport, and of callose synthase genes CALS3 and CALS8 whose overexpression can block the phloem transport (Vaté n et al., 2011;Yan et al., 2019). In contrast, pTMVDCPmutMP and pTMV-that spread systemically-suppress the induction of these regulators and strongly downregulates CALS7, another callose synthase gene responsible for callose deposition in the phloem (Xie et al., 2011). The effects of mutMP and CP on the host phloem loading/unloading factors do not always parallel each other; for instance, the expression of a cell wall-associated and cadmium-induced glycine-rich protein cdiGRP was dramatically suppressed by pTMVDCPmutMP but not by pTMV. Taken together, our data suggest that the adaptation of a CP-deficient TMV virus in the N. benthamiana host creates a multifunctional mutMP that retains its cell-to-cell movement function and gains, at least in part, the function of CP to promote systemic movement and suppress host immunity.
In summary, we propose a model for the TMV-host plant interactions that occur when the systemic movement capacity is lost in pTMVDCP and regained in pTMVDCPmutMP. The loss of CP restricts the ability of TMV to move systemically and induces the ET signaling and local necrosis. The lack of CP in pTMVDCP also increases the expression of phloem-associated resistance factors that regulate the processes of phloem loading and unloading. In contrast, the gain-of-function mutMP mutant is systemically movable and increases the viral RNA accumulation. The mutMP mutant virus accumulates at a higher level and gains a CP-like virulence suppressing the host defenses, e.g., local necrosis, ET pathway, and phloem-associated resistance factors.

Limitations of the study
This study focuses on the CP-independent systemic transport of defective TMV mutants infecting N. benthamiana. The spectrum of this ability should be confirmed with other defective tobamoviruses or in additional natural hosts of TMV. Also, only some of the most relevant resistance signaling pathways are characterized in this study. Therefore, a high-throughput transcriptomic or proteomic analysis is necessary for the full understanding of the systemic changes of host signaling pathways suppressed by the pTMVDCPmutMP gain-of-function mutant.

STAR+METHODS
Detailed methods are provided in the online version of this paper and include the following:

ACKNOWLEDGMENTS
The work in the V.C. laboratory was supported by grants from NIH, NSF, NSF/NIFA, and BARD to V.C.

AUTHOR CONTRIBUTIONS
P-T.T. and M-S.V.P. conducted the experiments and analyzed the experimental data. P-T.T. and V.C. wrote, reviewed, and edited the manuscript.

DECLARATION OF INTERESTS
The authors declare no competing interests.

Lead contact
Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Phu-Tri Tran (phutri.tran@stonybrook.edu).

Materials availability
Plasmids generated in this study are available from the lead contact.

Data and code availability
Individual data points of quantitative graphs are available in Tables S2-S10.
This paper does not report original code.
Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request (phutri.tran@stonybrook.edu).

Plant growth
Nicotiana benthamiana plants were grown on soil in an environment-controlled chamber at 23 C under a 16-h light (100 mmol photons mÀ2 sÀ1)/8-h dark cycle. One week after sowing, the seedlings were transplanted into 10 3 10 3 15 cm pots filled with vermiculite. Three-week-old N. benthamiana plants were used for virus inoculation and agrobacterium-mediated transient expression.

Bacterial growth
For the Agrobacterium-mediated transient expression, the cells of Agrobacterium tumefaciens strain EHA105 harboring the appropriate plasmids were individually added to 1 mL of LB broth containing appropriate antibiotics for vector selection. After the broth cultures had grown for 24 h at 28 C and 200 rpm, 0.1 mL of each was transferred to 4.9 mL of LB broth containing the same antibiotics and 200 mM acetosyringone. The 5-mL cultures were grown under the same conditions for 16 h before the Agrobacterium cells were collected for inoculation into plants. iScience Article 5.7, 200 mM acetosyringone) to OD 600 of 0.1, 0.2, or 0.001 for virus inoculation, subcellular localization, and movement assays, respectively. For subcellular localization, the suspensions of bacteria with the tested constructs were mixed at a 1:1:1 vol/vol ratio with the suspensions of bacteria harboring reference constructs that express free YFP (a nucleocytoplasmic marker), PDCB1-mRFP (a plasmodesmal marker), or BAM1-mRFP (a plasma membrane marker). These cell mixtures were infiltrated in two abaxial sides of fully expanded leaves of three-week-old N. benthamiana.
To monitor the expression of pTMVDCP G, the inoculated plants were periodically imaged using a digital camera with a UV filter under the 377 nm UV light in a dark room. Subcellular localization of MP-CFP was recorded at 2 dpi under a laser scanning confocal microscope (LSM 900, Zeiss) with a 403 objective lens and CFP-, YFP-, and mRFP-specific filters. The cell-to-cell movement of MP-CFP was scored at 2 dpi as multi-cell clusters by counting them under a confocal microscope with a 103 objective lens and a CFP-specific filter.

Quantitative RT-PCR (qRT-PCR)
To quantify viral RNA accumulations and transcriptional expression of the host genes, total RNA from 50 mg of the green leaf tissue around the inoculation site, or from an uninoculated, systemic leaf if so indicated, was extracted by the TRIzol reagent and utilized as a reverse transcription template to synthesize cDNA using the RevertAid Revert Transcription Kit and Hexa-random primers. Quantitative PCR (qPCR) was performed as described (Tran and Citovsky, 2021), using a QuantStudioä 3 real-time PCR system (Applied Biosystems #A28567) and the PowerUp SYBR Green Master Mix (Applied Biosystems #A25741) with the cycling regimen recommended by the manufacturer and gene-specific primers listed in Table S1. Fold change in gene expression is normalized to an internal control gene (Livak and Schmittgen, 2001) for which we utilized the N. benthamiana F-BOX gene (Liu et al., 2012). Fold change for each condition was calculated by the delta-delta Ct (cycle threshold, i.e., the number of PCR cycles required for the signal to become detectable above the background) method as described (Livak and Schmittgen, 2001;Tran et al., 2018). The resulting fold change was expressed relative to that in the pTMVDCP-inoculated plants, which was set to 1.0; pTMVDCP is the parental strain of pTMVDCPmutMP, which represents the reference point for the pTMVdelCPmutMP movement and serves as a control for possible effects of agroinoculation.

Western blotting
To detect the accumulation of the MP or mutMP proteins, total proteins from 100 mg of green leaf tissue around the inoculation site, or from an uninoculated, systemic leaf if so indicated, were extracted by grinding and heating in 0.5 mL of 1X sample buffer (50 mM Tris-Cl pH 6.8, 2% SDS, 0.1% bromophenol blue, 10% glycerol, 100 mM DTT) for 10 min at 95 C, followed by centrifugation at 11,000xg for 1 min. The supernatant (15 mL) was resolved by SDS polyacrylamide gel electrophoresis (PAGE) on a 10% gel and electro-blotted onto PVDF membranes (Immobilon #IPVH00010). The $55-kDa large unit of Rubisco was used as the loading control and visualized on the blots by the Ponceau S (Sigma-Aldrich #P7170) staining. The blots were then blocked by casein (1%, pH 8 in the TBS-T buffer), probed with rabbit anti-MP antibody (Alpha Diagnostic #TMVMP11-A, 1:10,000 dilution), followed by the horseradish peroxidase-conjugated goat anti-rabbit antibody (Abcam #ab2057181, 20,000 dilution). The probed blots were analyzed using an Opti-4CN Substrate Kit (Bio-Rad #1708235). The western blot band intensity was quantified by densitometry and MP accumulation was interpreted as pixel value using ImageJ software (https://imagej.nih.gov/ij/) (Schneider et al., 2012).

Microsomal extracts and viral RNA immunoprecipitation and quantification
For microsomal extracts, the inoculated or uninoculated, systemic leaves from N. benthamiana plants with viral symptoms were fixed in 1% formaldehyde (plus 0.01% TrixonX-100) for 10 min in a vacuum. The fixation was stopped by replacing the formaldehyde solution with 125 mM glycine supplemented with 0.01% Trix-onX-100 and vacuum treatment for 5 min. The leaves were then washed 4 times with distilled water and dried with absorbent papers. The microsome fraction from the leaves was extracted as described (Abas and Luschnig, 2010) and resuspended by sonication in the microsome protein solubilization buffer (100mM Tris-HCl pH 7.3, 150 mM NaCl, 1mM EDTA, 10% glycerol, 20 mM NaF, 1% Triton X-100, 1mM PMSF, complete protease inhibitor cocktail 1X). Finally, the microsome extracts were diluted 10 times in the dilution buffer (16.7 mM Tris/HCl pH 8.0, 167 mM NaCl, 1.1% w/v Triton X-100, 1.2 mM EDTA pH ll OPEN ACCESS iScience Article