Coat proteins of necroviruses target 14-3-3a to subvert MAPKKKα-mediated antiviral immunity in plants

Mitogen-activated protein kinase (MAPK) cascades play an important role in innate immunity against various pathogens in plants and animals. However, we know very little about the importance of MAPK cascades in plant defense against viral pathogens. Here, we used a positive-strand RNA necrovirus, beet black scorch virus (BBSV), as a model to investigate the relationship between MAPK signaling and virus infection. Our findings showed that BBSV infection activates MAPK signaling, whereas viral coat protein (CP) counteracts MAPKKKα-mediated antiviral defense. CP does not directly target MAPKKKα, instead it competitively interferes with the binding of 14-3-3a to MAPKKKα in a dose-dependent manner. This results in the instability of MAPKKKα and subversion of MAPKKKα-mediated antiviral defense. Considering the conservation of 14-3-3-binding sites in the CPs of diverse plant viruses, we provide evidence that 14-3-3-MAPKKKα defense signaling module is a target of viral effectors in the ongoing arms race of defense and viral counter-defense. MAPK cascades play an important role in innate immunity. Here, the authors show that that beet black scorch virus activates MAPK signaling in Nicotiana benthamiana and the viral coat protein counteracts MAPK-mediated defence by competitively binding to a MAPKKKα interactor.

I n nature, plants are sessile stores of nutriment, thus frequently being attacked by numerous pests and pathogens, including insects, bacteria, fungi, and oomycetes. In order to survive, plants have developed a two-layered innate immune system to defend against invading pathogens 1 . Pattern-recognition receptors (PRRs) on the plasma membrane recognize pathogen/ microbe-associated molecular patterns (PAMPs/MAMPs), triggering a series of downstream defense responses, including calcium flux, production of reactive oxygen species (ROS), activation of mitogen-activated protein kinase (MAPK) cascades and upregulation of defense genes. This perception of PAMPs/ MAMPs constitutes the first layer of plant protection against pathogen attack and is referred to as pattern-triggered immunity (PTI) 2,3 . To establish successful infections, pathogens secrete effectors into the plant cell to suppress PTI, resulting in effectortriggered susceptibility (ETS) 1 . Plants have evolved resistance (R) proteins to perceive these effectors and initiate a second layer of immunity named effector-triggered immunity (ETI), which is generally stronger and lasting longer than PTI and often accompanied by hypersensitive responses (HR), a type of cell death to confine pathogens to the infection sites 4 .
It is widely accepted that RNA silencing is the first layer of plant defense against viruses 5,6 . In addition, plants also have evolved R proteins that recognize viral proteins as effectors and activate ETI similar to that employed in response to nonviral pathogens 7 . So far, no classical PAMPs have been found in plant-virus interactions. However, over the past decade, a growing body of evidence suggests that plants also deploy PTI to limit virus infection. For example, co-receptors such as brassinosteroid insensitive 1-associated kinase 1 (BAK1) and BAK1-like kinase 1 involved in PTI in Arabidopsis thaliana against nonviral pathogens are also shown to contribute to plant defense against several RNA viruses, including turnip crinkle virus (TCV) 8 , oilseed rape mosaic virus (ORMV), tobacco mosaic virus (TMV) 9 , and plum pox virus (PPV) 10 . MAPK4 homologs in soybean (Glycine max) are reported to negatively regulate SA accumulation and defense against soybean mosaic virus (SMV) 11 . Distinct from BAK1mediated PTI, the nuclear shuttle protein-interacting kinase 1 (NIK1)-mediated antiviral signaling is specific to plant DNA viruses belonging to the genus Begomovirus of Geminiviridae family, and causes translocation of the ribosomal protein L10 (RPL10) to the nucleus for translational suppression of plant defense 12,13 . Geminivirus nuclear shuttle protein (NSP) overcomes this defense by interfering with NIK1-mediated nuclear relocalization of RPL10A 12 . Some RNA viruses have also been reported to counter PTI responses. For example, coat protein (CP) of PPV and movement protein (MP) of cucumber mosaic virus (CMV) were reported to suppress flg22-induced ROS production and PTI marker genes in Arabidopsis and Nicotiana benthamiana 10,14 . Cauliflower mosaic virus (CaMV) silencing suppressor P6 was shown to suppress ROS burst and SAdependent autophagy 15 . Although these studies show some evidence suggesting an association of PTI in antiviral immunity, the functional role of PTI against plant viruses, especially positivestrand RNA viruses, and the molecular mechanisms underlying viral suppression of PTI remain to be fully characterized.
Although PTI and ETI are activated mainly through distinct immune receptors, they share some features of the signaling machinery and initiate many common responses 16,17 . For example, MAPK cascades serve as a convergent point downstream of PTI and ETI 18,19 . MAPK cascades are composed of three tiers of kinases, MAPKs-MAPKKs-MAPKKKs, and activated through phosphorylation by their upstream kinases in turn, which play a central role in stress signal transduction from upstream receptors to downstream targets 20 . Two main MAPK cascades are reported to be activated in Arabidopsis during PTI 21 , one is consisted of MAPKKK3/MAPKKK5-MKK4/MKK5-MPK3/MPK6 22,23 and the other is consisted of MEKK1-MKK1/ MKK2-MPK4 24,25 . In addition, several components of MAPK cascades are also reported to play important roles in R proteinmediated ETI and often associated with HR cell death. For example, in Arabidopsis, MPK3 and MPK6 are activated by RPS2mediated immunity 26 . In tobacco, two MAPKs, salicylic acidinduced protein kinase (SIPK) and wounding-induced protein kinase (WIPK), the orthologs of MPK6 and MPK3 in Arabidopsis 20 , are involved in N gene-mediated resistance 27,28 and cell death 29,30 . In tomato, two MAPKKs, MEK1 and MEK2, and two MAPKs, NTF6 and WIPK, contribute to Pto-mediated immunity 31 , and overexpression of a tomato MAPKKK, LeMAPKKKα, results in cell death 32 . Some pathogen effectors were reported to target MAPK cascades to inhibit defense responses such as AvrRpt2 33 , HopAI1 34 , and HopF2 35 . Recently, geminivirus tomato yellow leaf curl China virus (TYLCCNV) encoded βC1 was reported to directly target MKK2 and MPK4 to suppress plant immunity 36 . However, whether positive-strand RNA viruses target MAPK cascades remains to be determined.
In this study, we used a positive-strand RNA virus, beet black scorch virus (BBSV), as a model to explore the functional role of plant innate immune signaling in virus infection and the counterdefensive strategies employed by virus to overcome this layer of defense. We demonstrate that MAPKKKα-mediated defense against BBSV infection in N. benthamiana is correlated with the expression of a series of defense-related genes such as PR1A, ERF1B, and HIR1, whereas viral CP can subvert MAPKKKαmediated antiviral innate immune response by targeting 14-3-3a.
We also provide evidence that the CP targeting of MAPKKKα-14-3-3a defense signaling module also occurs in the infection of the other two plant RNA viruses in the family Tombusviridae.
Our study provides mechanistic insight into understanding the suppression of plant innate immunity by RNA viruses.

Results
MAPKKKα plays an antiviral role during BBSV infection. MAPKKKα has been reported to positively regulate cell-death signaling associated with plant immunity 32 . However, little is known about its role in virus infection. We used BBSV as a model to explore the role of MAPKKKα in plant-virus infection. To monitor the virus infection, we used a self-assembling split superfolder green fluorescent protein (sfGFP) system 37 . In this system, 11th β-strand of sfGFP (sfGFP11) was fused to the N-terminus of CP in the BBSV infectious clone to generate BBSV-sfGFP 11 (Fig. 1a). Co-expression of BBSV-sfGFP 11 with sfGFP1-10 βstrand (sfGFP1-10) in plant cells will result in the GFP fluorescence, which can be used to indicate the infection of BBSV. We then knocked down the expression of MAPKKKα by using intron spliced hairpin RNAi vector-mediated gene silencing 38 , and observed increased GFP fluorescence and CP accumulation in MAPKKKα-RNAi leaves compared with the empty vector (EV) control (Fig. 1b, c). Downregulation of MAPKKKα using this strategy was validated by RT-qPCR (Fig. 1d). To further confirm these results, we generated N. benthamiana MAPKKKα-knockout (KO) plants via CRISPR-Cas9 gene editing. MAPKKKα has a homolog (MAPKKKα-2) in N. benthamiana ( Supplementary  Fig. 1a) and Sanger sequencing indicated deletions at gRNA targeting site within two MAPKKKα genes in MAPKKKα-KO plants ( Fig. 1e and Supplementary Fig. 1b). MAPKKKα-KO plants accumulated increased CP upon BBSV inoculation (Fig. 1f). These results indicate that MAPKKKα plays an antiviral role during BBSV infection.
To further determine the role of MAPKKKα in BBSV infection, we overexpressed MAPKKKα and a kinase-inactive mutant, MAPKKKα K236M (K236 is a conserved ATP binding site essential for the kinase activity of MAPKKKα 32 ) followed by BBSV inoculation. At 3-day post infiltration (dpi), the accumulation of BBSV CP was decreased in leaves overexpressing MAPKKKα compared with EV control. However, overexpression of MAPKKKα K236M had no significant effect on BBSV CP accumulation (Fig. 1g, top panel). Moreover, the phosphorylation levels of two MAPKs, NbSIPK, and NbWIPK, which were reported to act downstream of MAPKKKα 32 were higher in the leaves overexpressing MAPKKKα compared with EV and MAPKKKα K236M (Fig. 1g, middle panel). Furthermore, the expression of MAPKKKα induced cell death but not by  To analyze whether MAPK cascades were activated during BBSV infection, we infiltrated N. benthamiana with Agrobacterium containing BBSV infectious cDNA and monitored the phosphorylation status of NbSIPK and NbWIPK at different time points. We observed a significant increase in phosphorylation levels of NbSIPK and NbWIPK at 0.5-, 1-, and 2-day post BBSV inoculation ( Fig. 1h, i). These results indicate that the MAPK cascade is activated at the early stage of BBSV infection.
To further test whether the activated MAPK signaling is involved in defense against BBSV infection, we performed loss-offunction assays of NbMEK2, NbSIPK, and NbWIPK by using Tobacco rattle virus (TRV)-based gene silencing (VIGS). CP accumulation in NbMEK2, NbSIPK, or NbWIPK silenced plants was higher than in the control TRV-GFP inoculated plants (Fig. 1j). The silencing efficiency of NbMEK2, NbSIPK, or NbWIPK was confirmed by RT-qPCR (Fig. 1k). Together, these results indicate that BBSV infection activates the MAPKKKα-MEK2-SIPK/WIPK cascade, which in turn induces plant defense responses against BBSV infection.
MAPKKKα-mediated antiviral response associates with the expression of defense-related genes. To further investigate the mechanism underlying MAPKKKα-mediated antiviral response, we performed RNA-seq analysis in wild-type (WT) and the MAPKKKα-KO plants during BBSV infection. We collected leaves from mock-and BBSV-infected plants at 0.5, 1, and 2 dpi and performed RNA-Seq. We observed upregulation of 911, 3662, and 2570 genes in WT plants at 0.5, 1, and 2 dpi, respectively (Fig. 2a, b and Supplementary Data 1). The Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis indicated that these upregulated genes were highly enriched in plant-pathogen interaction process, MAPK signaling pathway, and plant hormone signal transduction (Fig. 2c). Further analysis showed that 370 genes, which were upregulated in WT plants after BBSV infection, were downregulated in the comparisons KO-BBSV vs WT-BBSV ( Fig. 2d, e and Supplementary data 2). KEGG analysis revealed that these downregulated genes were also enriched in plant-pathogen interaction process and MAPK signaling pathway (Fig. 2f). We selected three defense-related genes, including PATHOGENESIS-RELATED PROTEIN 1A (PR1A), ETHYLENE-RESPONSIVE TRANSCRIPTION FACTOR 1B (ERF1B) 39 , and HYPERSENSITIVE-INDUCED RESPONSE PROTEIN 1 (HIR1) 40 for further RT-qPCR analysis, and the results were consistent with the RNA-seq data (Fig. 2g). Together, these results suggest that MAPKKKα-mediated defense against BBSV infection is correlated with the expression of defense-related genes.
BBSV CP suppresses MAPKKKα-mediated antiviral immunity through its interaction with 14-3-3a. We noted that the phosphorylation levels of NbSIPK and NbWIPK declined two days after BBSV infection (Fig. 1h, i), suggesting that BBSV may have evolved with strategies to suppress MAPK activity. To test if BBSV-encoded proteins could suppress MAPKKKα-mediated signal transduction, we co-expressed different BBSV proteins and GFP control with HA-MAPKKKα, which was controlled by an estradiol-inducible system. Trypan blue staining of leaf tissues and the electrolyte-leakage assay showed that only CP of BBSV suppressed MAPKKKα-induced cell death while other viral proteins showed a similar cell-death phenotype to that of the GFP control ( Fig. 3a, b). Western blot analysis with anti-GFP antibody confirmed the expression of these proteins ( Supplementary  Fig. 3a). Together, these results demonstrate that BBSV CP plays a role in suppressing MAPKKKα-mediated defense.
We reasoned that CP may suppress MAPKKKα-mediated defense by directly interacting with MAPKKKα and inhibiting its activity. To test this, we performed bimolecular fluorescence complementation (BiFC) assay. As a positive control, CP-YFPn and CP-YFPc resulted in strong reconstituted YFP signals, yet no fluorescence signal was observed in the leaves co-expressing either CP-YFPc and MAPKKKα K236M -YFPn or CP-YFPn and MAPKKKα K236M -YFPc ( Supplementary Fig. 4a). The expression of proteins was confirmed by western blot analysis (Supplementary Fig. 3b, c). Yeast two-hybrid (Y2H) assay showed that yeast expressing AD-MAPKKKα and BD-CP or AD-CP and BD-MAPKKKα failed to grow on the drop-out media compared with  11 . 11th β-strand of super-folder GFP (sfGFP 11 ) is fused to the N-terminus of CP in BBSV infectious clone. b Knockdown of MAPKKKα promotes BBSV infection. Agrobacterium containing hairpin MAPKKKα or the control empty vector (EV) was infiltrated into half of N. benthamiana leaves, respectively. Twenty-four hours later, a mixture of Agrobacterium harboring BBSV-sfGFP 11 and sfGFP1-10 constructs was infiltrated into the pre-infiltrated leaves. GFP fluorescence was visualized by confocal microscopy at 5 dpi. Scale bars = 100 µm. c The expression level of BBSV CP in the infiltrated leaves was examined by western blot analysis using an anti-CP antibody. Actin served as the loading control. For panels b and c, the experiments were repeated three times with similar results. d RT-qPCR to confirm the downregulation of MAPKKKα. Values represent ± SD of the mean from three biological replicates. An asterisk indicates the significant difference (*P = 0.0104) based on one-sided unpaired Student's t test. e Locus of the single guide RNA target in the genomic fragment of NbMAPKKKα. Exons are represented by gray boxes, UTRs by white boxes, and introns by black lines. In T2 transformants, deletions were detected in two NbMAPKKKα-KO lines. f Western blot analysis of BBSV CP levels in the infiltrated leaves of wild-type (WT) or NbMAPKKKα-KO plants using an anti-CP antibody. RbcL served as the loading control. The experiment was repeated three times with similar results. g Overexpression of MAPKKKα inhibits BBSV infection. Agrobacterium containing MAPKKKα or MAPKKKα K236M or the control empty vector (EV) was infiltrated into different N. benthamiana leaves. Agrobacterium harboring BBSV infectious clone was infiltrated 24 h later. The accumulation levels of BBSV or phosphorylated NbSIPK and NbWIPK were detected by western blot analysis with an anti-CP or anti-phospho-p44/42 MAPKs (anti-pTEpY) antibody. Actin protein served as the loading control. Each lane represents a sample from individual experiments. h. MAPK activation in BBSV-infected N. benthamiana. Eight-leaf-stage N. benthamiana plants were infiltrated with Agrobacterium carrying empty vector (EV) or BBSV infectious clone. Infiltrated leaves were collected at indicated time points and subjected to western blot analysis with an anti-pTEpY or anti-BBSV CP antibody. Actin protein served as the loading control. i Quantification of bands from blot in panel h, with pSIPK and pWIPK abundance normalized to the intensity of EV control at 0 h. Values represent ± SD of the mean from three biological replicates. An asterisk indicates the significant difference (*P = 0.0498, 0.0284, 0.0134, respectively) based on one-sided paired Student's t test. j Western blot analysis of BBSV CP accumulation in the TRV-GFP control, TRV-NbMEK2, TRV-NbSIPK, and TRV-NbWIPK inoculated N. benthamiana plants. Actin protein served as the loading control. k RT-qPCR to confirm the downregulation of NbMEK2, NbSIPK, and NbWIPK. Values represent ± SD of the mean from three biological replicates. An asterisk indicates the significant difference (**P = 0.0043, 0.0068, 0.0085, respectively) based on one-sided unpaired Student's t test. For panels h and j, the experiments were repeated three times with similar results.
the positive control ( Supplementary Fig. 4b). These results indicate that CP does not directly interact with MAPKKKα.
To identify the targets of BBSV CP that could interfere with the MAPKKKα activity, we performed immunoprecipitation (IP) of CP-3×FLAG and separated immunoprecipitates by SDS-PAGE followed by silver staining. A distinct band was visible in the IP products from CP-3×FLAG but the corresponding band was absent in the lane of GFP-3×FLAG control (Fig. 3c). The band was excised from the gel and analyzed by liquid chromatography coupled to tandem mass spectrometry (LC-MS/MS). Several    41,42 , we hypothesized that 14-3-3 proteins might take part in the plant response to BBSV infection. Among identified 14-3-3 proteins, N. tomentosiformis 14-3-3-like protein A (gi|697111701) showed the highest score, with 100% amino acid identity with N. benthamiana 14-3-3a ( Supplementary  Fig. 5). We then cloned 14-3-3a from N. benthamiana and used it as a surrogate for further analysis.
To investigate whether CP interacts with 14-3-3a in planta, we cloned full-length 14-3-3a from N. benthamiana and performed co-immunoprecipitation (Co-IP) assay. We found that CP coimmunoprecipitates with 14-3-3a-3×FLAG, but not with GFP-3×FLAG control (Fig. 3d). BiFC experiments confirmed the interaction in planta (Fig. 3g, top panel). To see if CP interacts directly with 14-3-3a, we performed a GST pull-down assay. A GST tag was fused to the N-terminus of 14-3-3a and a 6xHis tag was fused to CP. GST and GFP-His served as negative controls. The result showed that GST-14-3-3a specifically binds to the His-CP but not to the GFP-His control (Fig. 3e). Based on these results, we conclude that CP interacts with 14-3-3a in vivo and in vitro.
There are several conserved amino acids within animal 14-3-3s protein that determine its binding to the substrate 43 and dimerization 44 . Amino acid sequence alignment indicates that the amino acids required for these functions are also conserved in plant 14-3-3s (Fig. 3f). To examine whether these amino acids are responsible for 14-3-3 dimerization and its binding to the CP, two mutants, 14-3-3a QQR , in which the amino acids at 12-LAE-14 was mutated to QQR, and 14-3-3a 3E , in which the amino acids of K52, R59, and R63 were all mutated to E, were constructed. Only the mutation in LAE eliminated the dimerization of 14-3-3a ( Supplementary Fig. 6), and both of these two mutants fail to interact with CP as shown in the BiFC and Co-IP assays (Fig. 3g, h). The expression of proteins in BiFC assay was confirmed by western blot analysis ( Supplementary Fig. 3d). These results indicate that both the dimerization and substrate binding activity of 14-3-3a is indispensable for its binding to CP.
To test whether 14-3-3a plays a functional role in the activation of cell death through MAPKKKα in N. benthamiana, we performed BiFC and Co-IP assays. The results showed that the wild-type 14-3-3a, but not the 14-3-3a 3E and 14-3-3a QQR mutants, interacts with MAPKKKα ( Fig. 4a, b). The expression of proteins in BiFC assay was confirmed by western blot analysis ( Supplementary  Fig. 3b, c). We then examined the effect of 14-3-3a on MAPKKKα-induced cell death. Wild-type or mutant 14-3-3a was co-expressed with estradiol-induced HA-MAPKKKα protein, and cell death was evaluated at 3 dpi after 12 h estradiol treatment.
Trypan blue staining showed that overexpression of wild-type 14-3-3a significantly enhanced MAPKKKα-induced cell death compared with the GFP control (Fig. 4c). In contrast, coexpression of either 14-3-3a 3E or 14-3-3a QQR mutant with HA-MAPKKKα showed an alleviated cell-death phenotype (Fig. 4c). An electrolyte-leakage assay substantiated the results shown in Fig. 4c by showing that electrolyte leakage was increased by about 25% with the expression of wild-type 14-3-3a and decreased by about 10% with the expression of 14-3-3a mutants compared with the GFP control ( Fig. 4d), indicating a possible dominantnegative effect of these two mutants on MAPKKKα-induced cell death. The expression of wild-type 14-3-3a, mutant, and GFP proteins was confirmed by western blot analysis (Fig. 4g). We next tested the MAPKKKα-induced cell death in the Nb14-3-3asilenced leaves. Trypan blue staining and the electrolyte-leakage assay showed that MAPKKKα-induced cell death was dramatically decreased when 14-3-3a was silenced (Fig. 4e, f). Downregulation of 14-3-3a by RNAi was confirmed by western blot analysis with anti-14-3-3a antibody (Fig. 4h). These results indicate that 14-3-3a functions as a positive regulator of MAPKKKα-induced cell death.
We also found that the downregulation of 14-3-3a affects the accumulation of MAPKKKα protein, which can be partially rescued by the addition of the proteasome inhibitor MG132 (Fig. 4i). SIPK and WIPK phosphorylation levels were in line with MAPKKKα protein levels ( Fig. 4i). RT-qPCR analysis showed that silencing of Nb14-3-3a had no obvious effects on the transcription of NbMAPKKKα ( Supplementary Fig. 7), suggesting that Nb14-3-3a mainly regulated the abundance of NbMAPKKKα at the protein level. Together, these results demonstrate that 14-3-3a enhances MAPKKKα-mediated signaling by increasing the stability of MAPKKKα.
Nb14-3-3a-mediated antiviral defense depends on the MAPKKKα. To determine the functional role of 14-3-3a in viral infection, we transiently overexpressed wild-type and mutant 14-3-3a in N. benthamiana leaves and then mechanically inoculated BBSV virions at 24 h post infiltration (hpi). The accumulation of BBSV CP was greatly reduced in the leaves overexpressing wildtype 14-3-3a compared with the control leaves expressing GFP (Fig. 5a), indicating that overexpression of 14-3-3a inhibits BBSV infection. In contrast, we found increased CP accumulation in leaves overexpressing 14-3-3a mutants that fail to interact with CP ( Fig. 5a), suggesting a dominant-negative effect of these mutants on BBSV infection. We also transiently silenced 14-3-3a followed by infiltration with Agrobacterium harboring the BBSV infectious clone at 24 hpi. Western blotting with anti-CP or anti-14-3-3a antibodies showed that the accumulation of viral CPs was increased along with the downregulation of 14-3-3a expression (Fig. 5b). To further confirm the role of 14-3-3a in BBSV infection, we generated 14-3-3a overexpression (14-3-3a-OE) and knockdown (14-3-3a-KD) transgenic N. benthamiana plants. We then compared the symptoms of WT and transgenic N. benthamiana plants upon BBSV infection. The symptoms of BBSVinfected plants were arbitrarily divided into four grades based on the severity of symptoms on systemic leaves ( Supplementary  Fig. 8). When these transgenic plants were inoculated with BBSV virions, we found that a higher percentage of 14-3-3a-OE plants showed attenuated symptoms (grade I) and a lower percentage of plants showed severe disease symptoms (grade III) in the systemic leaves. In contrast, BBSV induced more severe symptoms in 14-3-3a-KD plants compared with that of the control WT plants Fig. 3 CP suppresses MAPKKKα-induced cell death by interacting with 14-3-3a. a BBSV CP suppresses the cell death induced by MAPKKKα. Five different BBSV-encoded proteins or GFP control were co-expressed with HA-MAPKKKα controlled by an estradiol-inducible system in indicated leaf regions. β-estradiol was infiltrated into the leaves at 48 h after agroinfiltration. Leaves were stained by trypan blue, and representative photographs were taken at three days after estradiol treatment. b Quantification of cell death by measuring electrolyte leakage of leaves shown in (a). Error bars indicate ± SD of the mean (n = 5 biologically independent plants). Asterisks indicated the significant difference based on one-way ANOVA analysis with Dunnett's multiple comparison test (**P = 0.0013). c Silver-stained SDS-PAGE gel image of CP-3×Flag and GFP-3×Flag protein immunoprecipitated using anti-Flag beads from BBSV-infected N. benthamiana leaf tissues. The band indicated by the rectangle was analyzed by LC-MS/MS. d Co-IP analysis of the interaction between 14-3-3a and CP. N. benthamiana leaves transiently expressing combinations of different proteins indicated above the panel were harvested at 3 dpi. Total proteins were immunoprecipitated with anti-Flag beads and detected by western blot with an anti-CP or anti-Flag antibody. e GST pull-down assay to detect the interaction between CP and 14-3-3a. Purified His-CP or His-GFP was incubated with GST-14-3-3a or GST protein control. After incubation with glutathione-Sepharose beads, the pull-down products were analyzed by western blot with an anti-His or anti-GST antibody. For panels c-e, the experiments were repeated three times with similar results. f Depiction of 14-3-3 N-terminal sequences alignment generated by UniPort (https://www.uniprot.org/ align/). 14-3-3 isoforms from different species were indicated on the left side. The underlined amino acids are required for dimerization of 14-3-3 proteins 44 . The amino acids responsible for 14-3-3-ligand-binding activity are highlighted with asterisks 43 . g BiFC analysis of the interaction between wildtype or mutant 14-3-3a and CP. CP and wild-type or mutant 14-3-3a fused to N or C-terminus of YFP were transiently co-expressed in N. benthamiana leaves. Confocal analysis was performed at 3 dpi. Representative results of at least three independent experiments are shown. Scale bars = 50 µm. h Co-IP analysis of the interaction between wild-type or mutant 14-3-3a and CP. N. benthamiana leaves transiently expressing combinations of different proteins were harvested at 3 dpi. Total proteins were immunoprecipitated with anti-Flag beads and detected by western blot with an anti-GFP or anti-Flag antibody. The experiment was repeated three times with similar results. Error bars indicate ± SD of the mean (n = 7 biologically independent plants). Different letters in the chart indicate statistically significant differences among different groups according to the one-way ANOVA analysis with Tukey's multiple comparison test (P < 0.05). e Knockdown of 14-3-3a eliminates MAPKKKα-induced cell death. Agrobacterium containing hairpin 14-3-3a or the control empty vector was co-infiltrated with HA-MAPKKKα controlled by an estradiol-inducible system into the opposite halves of N. benthamiana leaves. β-estradiol was applied to the leaves 48 h after agroinfiltration. Leaves were stained by trypan blue and representative photographs were taken 3 days after estradiol treatment. f Quantification of cell death by measuring electrolyte leakage of leaves treated as described in (e). Error bars indicate ± SD of the mean (n = 6 biologically independent plants). Asterisks indicate the significant difference (**P = 0.0024) based on one-sided paired Student's t test. g Western blot analysis to confirm the expression of the GFP and 14-3-3a proteins shown in panel c with an anti-Flag antibody. h Western blot analysis to confirm the downregulation of 14-3-3a at protein level shown in panel e with an anti-14-3-3a antibody. i 14-3-3a is important for maintaining the stability of MAPKKKα. Agrobacterium containing hairpin 14-3-3a or empty vector were co-infiltrated with HA-MAPKKKα controlled by an estradiol-inducible system into the leaves. In total, 30 μM βestradiol with 50 μM MG132 or DMSO (control) was infiltrated into the leaves 48 h after agroinfiltration. Samples were collected at 0, 12, 24 h after treatment with a chemical reagent, and subjected to western blot with an anti-HA or anti-pTEpY antibody. For panels g-i, the experiments were repeated three times with similar results.
Based on the data presented above, we reasoned that 14-3-3amediated defense against BBSV might depend on MAPKKKα. To test this hypothesis, we overexpressed 14-3-3a in WT and MAPKKKα-KO plants, and performed BBSV infection assay. Western blot analysis showed that overexpression of 14-3-3a in MAPKKKα-KO plants failed to impair the BBSV infection based on the CP accumulation, in contrast to that of the WT plants (Fig. 5f). These results suggest that 14-3-3a positively regulates antiviral defense in a MAPKKKα-dependent manner.
CP suppresses the MAPKKKα-mediated cell-death signaling by targeting Nb14-3-3a. BBSV CP is divided into two regions: a N-terminal region responsible for RNA binding and nuclear localization 45,46 , and a shell domain 47 . To determine the region of CP protein that is important for its binding to 14-3-3a, we constructed N-terminal or C-terminal truncated CP mutants respectively (Fig. 6a), and tested their interactions with 14-3-3a by BiFC assay. The result showed that CP 46-232 retains the ability to interact with 14-3-3a, whereas CP 1-187 was unable to interact with 14-3-3a (Fig. 6b). The expression of proteins was confirmed by western blot analysis (Supplementary Fig. 3e). These results indicate that the C-terminal region between amino acids 188 and 232 is required for CP interaction with 14-3-3a. By aligning CP sequences of BBSV and TNV-D H , we found a motif in this region, which is similar to a conventional 14-3-3-binding motif (Rxx[pS/ pT]xP) 48 . Intriguingly, this motif in BBSV CP contains tyrosine (Y) instead of S/T and TNV-D H contains phenylalanine (F) (Fig. 6c). The Y and F amino acids are structurally similar, implying that the phosphorylation of the 14-3-3-binding motif in BBSV CP may not be required for its binding to 14-3-3a and the maintenance of the structure may be critical for its binding. To test this hypothesis, we mutated Y194 in BBSV CP into F and alanine (A) and tested their binding abilities to 14-3-3a. BiFC, Co-IP, and GST pull-down assays showed that CP Y194F mutant retains the ability to interact with 14-3-3a, whereas CP Y194A mutant did not (Fig. 6d-f). The expression of proteins in BiFC assay was confirmed by western blot analysis ( Supplementary  Fig. 3e). Together, these data indicate that the conformation but not the phosphorylation state of Y194 of CP is important for its interaction with 14-3-3a.
To investigate whether CP suppressed MAPKKKα-mediated cell death by binding to 14-3-3a, wild-type or mutant CPs were co-expressed with estradiol-induced HA-MAPKKKα. Trypan blue staining and electrolyte-leakage assays showed that leaf tissues expressing CP and CP Y194F , but not CP Y194A , exhibited reduced cell death induced by MAPKKKα compared with the EV control (Fig. 7a, b). The expression of CP proteins was confirmed by western blot analysis (Fig. 7c). Since CP Y194A barely interact with 14-3-3a ( Fig. 6d-f), we conclude that inhibition of MAPKKKα-induced cell death by viral CP is dependent on its interaction with 14-3-3a.
To investigate the molecular basis underpinning the CP inhibition of MAPKKKα-induced cell death, we detected the levels of MAPKKKα proteins and activation of MAPKs (phosphorylation of SIPK and WIPK) at different time points after induction of MAPKKKα expression in the presence of CP or CP Y194A . Western blot analysis showed that, compared with EV control, the abundance of HA-MAPKKKα was decreased in the leaves expressing CP, but showed no obvious changes in the leaves expressing CP Y194A (Fig. 7d). Consistently, the levels of SIPK and WIPK phosphorylation were also decreased in leaves expressing CP (Fig. 7d). To determine whether CP was able to impair the enhancement of 14-3-3a on MAPKKKα-mediated signaling, we conducted a similar assay in 14-3-3a-OE transgenic N. benthamiana plants. Levels of both MAPKKKα proteins and MAPKs phosphorylation were decreased in the leaves expressing CP (Fig. 7e). We also generated stable N. benthamiana transgenic lines overexpressing CP. Compared with the WT plants, two independent N. benthamiana plants stably overexpressing CP showed a decreased abundance of MAPKKKα and reduced phosphorylation of MAPKs after induction of MAPKKKα expression (Fig. 7f). Together, these results indicate that CP suppresses the activation of MAPKs by reducing the abundance of MAPKKKα.
It has been previously shown that the CP of another RNA virus, potato virus X (PVX) accumulate to a higher level in NbMAPKKKα-silenced N. benthamiana plants 49 . Therefore, we tested whether BBSV CP mediated reduction of MAPKKKα could lead to enhanced PVX infection. For this, we engineered BBSV CP or CP Y194A into PVX-based vector and agroinfiltrated into N. benthamiana. The symptoms caused by PVX-BBSV-CP were more severe than that of either PVX-BBSV-CP Y194A or PVX-GFP (Fig. 7g). Consistently, the accumulations of both PVX and BBSV CP in the systemic leaves of PVX-BBSV-CP-infected N. benthamiana were higher than that in PVX-BBSV-CP Y194A or PVX-GFP-infected plants (Fig. 7h). These results further revealed that BBSV CP has the ability to inhibit MAPKKKα-mediated antiviral immunity.
To further explore the biological significance of the CP-14-3-3a interaction on MAPKKKα-mediated defense against BBSV infection, we generated BBSV mutant by replacing CP with CP Y194A (BBSV Y194A ) and inoculated them onto WT or MAPKKKα-KO plants. We observed cell-death phenotype on the leaves of WT plants inoculated with BBSV Y194A at 2 dpi but not for the leaves inoculated with WT-BBSV ( Supplementary  Fig. 9a). In contrast, neither BBSV nor BBSV Y194A induced celldeath phenotype on the inoculated leaves of MAPKKKα-KO plants ( Supplementary Fig. 9a), indicating that the BBSV Y194Ainduced cell death depends on the MAPKKKα. Consistent with this, leaves inoculated with BBSV Y194A showed a lower level of CP accumulation compared with BBSV, and knocking out of the NbMAPKKKα gene partially rescues the accumulation level of BBSV Y194A (Supplementary Fig. 9b). We did not observe the systemic infection of BBSV Y194A in either WT or MAPKKKα-KO plants ( Supplementary Fig. 9b, c), suggesting the Y194 in CP is critical for BBSV systemic infection. Since the Y194A mutation eliminates the interaction of CP with 14-3-3a, these results indicate that the interaction of CP with 14-3-3a is essential for suppressing MAPKKKα-mediated cell death during BBSV infection.
CP competitively interferes with the MAPKKKα binding to 14-3-3a. Given that both CP and MAPKKKα interact with 14-3-3a but CP does not interact with MAPKKKα (Figs. 3 and 4a, b and Supplementary Fig. 4), we wonder whether CP interferes with the interaction of MAPKKKα with 14-3-3a. To test this hypothesis, we performed a competitive Co-IP assay. Agrobacterium containing 14-3-3a-3×Flag and HA-MAPKKKα K236M expression cassette were co-infiltrated into the N. benthamiana leaves. Leaf homogenates were prepared and mixed with E. coli-expressed His-tagged CP proteins. After incubation for 40 min, the homogenates were precipitated with anti-Flag beads, and the eluates were then examined by western blot. The results showed that the association between 14-3-3a-3×Flag and HA-MAPKKKα K236M was reduced with the addition of increasing amounts of E. coliexpressed His-tagged CP (His-CP). However, CP Y194A mutant protein, which barely interacts with 14-3-3a, had little effect on the interaction of 14-3-3a-3×Flag with HA-MAPKKKα K236M (Fig. 8a). Furthermore, we performed GFP competitive pull-down assay. Because E. coli-expressed MAPKKKα was insoluble, GFPfused MAPKKKα K236M was expressed in N. benthamiana followed by purification with GFP-Trap beads. An excessive amount of E. coli-expressed His-14-3-3a was used to saturate the beads, and the supernatant was removed. With the addition of an increasing amount of E. coli-expressed His-CP to the above  Fig. 6 Determination of key amino acids that are responsible for CP-14-3-3a interaction. a Schematic representation of CP truncated mutants used for BiFC assays. b BiFC analysis of regions within CP that interacted with 14-3-3a. Wild-type or truncated CP mutants fused to N-terminus of YFP and 14-3-3a fused to C-terminus were transiently co-expressed in N. benthamiana leaves. Confocal analysis was performed at 3 dpi. Representative results of at least three independent experiments are shown. Scale bars = 50 µm. c Multiple sequence alignment of the C-terminus of CPs from Betanecroviruses. The 14-3-3-binding-like motif is shaded in gray, and the putative 14-3-3-binding site is highlighted in red. d BiFC analysis of the interaction between wild-type or mutant CP and 14-3-3a. Wild-type or mutant CP and 14-3-3a fused to N or C-terminus of YFP were transiently co-expressed in N. benthamiana leaves. Confocal analysis was performed at 3 dpi. Representative results of at least three independent experiments are shown. Scale bars = 50 µm. e Co-IP analysis of the interaction between wild-type or mutant CP and 14-3-3a. N. benthamiana leaves transiently expressing combinations of different proteins were harvested at 3 dpi. Total proteins were immunoprecipitated with anti-Flag beads and detected by western blot with an anti-CP or anti-Flag antibody. f GST pull-down assay to detect the interaction between wild-type or mutant CP and 14-3-3a. Purified His-tagged CP or its derivatives was incubated with GST-14-3-3a or GST protein. His-GFP protein served as the negative control. After incubation with glutathione-Sepharose beads, the pull-down products were analyzed by western blot with an anti-His or anti-GST antibody. For panels e and f, the experiments were repeated three times with similar results.
GFP-Trap beads, more and more His-14-3-3a was pulled off from the beads concomitant with increased levels of His-14-3-3a in the supernatant (Fig. 8b). These data indicate that CP competitively interferes the interaction between MAPKKKα and 14-3-3a, leading to a decrease in the abundance of MAPKKKα (Fig. 8c).
Perturbation of 14-3-3a-MAPKKKα functional module by the CP may be a common strategy employed by necroviruses. Sequence alignments showed that the 14-3-3a binding-like motif is conserved in the CPs of alphanecrovirus and betanecrovirus ( Supplementary Fig. 10), suggesting that the perturbation of 14-3-3a-MAPKKKα functional module by CP may also be employed by other necroviruses. To test this hypothesis, the conserved 14-3-3a binding site F 236 in the CP of Tobacco necrosis virus-A (TNV-A C , alphanecrovirus) and F 228 in the CP of TNV-D H (betanecrovirus) were mutated to alanine. BiFC assay showed that coexpression of 14-3-3a with wild-type TNV-A C and TNV-D H CPs, but not mutant CPs, resulted in the reconstitution of fluorescence signals (Fig. 9a, b). Expression of the test proteins was confirmed by western blot analysis (Supplementary Fig. 3f). These results indicate that both TNV-A C and TNV-D H CPs interact with 14-3-3a, and the amino acids F 236 or F 228 within the CPs are crucial for their interaction with 14-3-3a.
To examine the effect of TNV-A C and TNV-D H CPs on MAPKKKα-induced cell death, wild-type or mutant CPs was coexpressed with HA-MAPKKKα controlled by an estradiol- inducible system. Trypan blue staining and the electrolyte-leakage assays showed that compared with the EV control, MAPKKKαinduced cell death was significantly decreased in leaf tissues expressing wild-type CPs, but not mutant CPs (Fig. 9c-f). Consistently, the levels of MAPKKKα proteins and downstream MAPK phosphorylation level were decreased in the leaves expressing wild-type CPs, but not mutant CPs, compared with the EV control (Fig. 9g, h). We also investigated whether 14-3-3a and MAPKKKα play antiviral roles in TNV-A C and TNV-D H infection. 14-3-3a-OE and MAPKKKα-KO plants were inoculated with TNV-A C or TNV-D H , respectively. TNV-A C and TNV-D H caused more severe symptoms in MAPKKKα-KO plants than WT plants (Fig. 9i, j). Western blot and RT-qPCR analysis showed that both TNV-A C and TNV-D H accumulation was increased in MAPKKKα-KO plants (Fig. 9k, l) and significantly decreased in 14-3-3a-OE plants ( Supplementary Fig. 11). These findings indicate that other necroviruses may utilize the same strategy as that of BBSV to counteract MAPKKKα-mediated antiviral immunity.

Discussion
MAPK cascades play a central role in defense signaling against pathogens 20 . In animals, both DNA and RNA viruses are known to induce the MAPK cascades 50,51 . They utilize and fine-tune MAPK cascades to promote viral replication 52,53 . However, the relationship between MAPK cascades and plant viruses remains understudied. Here, we show that MAPKKKα plays an antiviral role. Our findings showed that the MAPK cascade is activated during the early stage of BBSV infection as a defense mechanism. Consistent with this, downregulation of MAPKKKα, as well as the downstream MAPKs, enhanced the accumulation of BBSV. Likewise, MAPK cascades are also reported to be activated in tomato yellow leaf curl China virus (TYLCCNV)-infected N. benthamiana 36 Supplementary Fig. 12a), consistent with the RNAseq data (Supplementary data 4). Although phosphorylation levels of SIPK and WIPK were significantly increased in response to BBSV infection (Fig. 1h, i), the protein levels of SIPK and WIPK were also not altered during BBSV infection (Supplementary Fig. 12c). These results suggest that BBSV-induced activation of MAPK signaling pathway is through modulation of the phosphorylation status of MAPKKKα-MEK2-SIPK/WIPK components but not through the increased protein expression.
Although it has been reported that MAPK cascades play an important role in plant defense against a variety of pathogens 18 , the mechanism underlying MAPK-mediated defense against plant-virus infection remains largely unclear. Here, we performed RNA-seq analysis to reveal that the expression of a large number of genes were transcriptionally regulated in response to BBSV infection, and some of these genes are specifically regulated through MAPKKKα (Fig. 2). For example, PR1A (a salicylic acid (SA)-responsive gene), ERF1B (a transcription factor that is involved in ethylene signaling), and HIR1 (specifically associated with the HR cell death) are all reported to be regulated during plant defense against virus 40,54,55 . Moreover, the SA pathway, ethylene pathway, and HR cell death are reported to cross-talk with each other and also with MAPK signaling pathway 26,[56][57][58] . Hence, our results demonstrated that MAPKKKα plays an antiviral role by activating MAPK signaling and inducing downstream defense-related gene expression, which provides a global insight into the mechanisms underlying the MAPKKKα-mediated antiviral immunity. Nevertheless, more details on how these genes perceive the MAPK signaling and induce antiviral response needs further investigation. Besides, how viruses activate MAPK cascades is a remaining gap to be addressed in the future. It is likely that intracellular or extracellular receptors recognize some conserved structures such as the virions, viral ribonucleoprotein complexes, or viral-derived dsRNAs to initiate innate immune signaling 59 .
Although the activation of MAPK signaling has some effect on BBSV infection, it is not sufficient to block virus infection suggesting that BBSV may have evolved a strategy to suppress MAPK signaling. Our data showed that BBSV CP is responsible for counteracting MAPKKKα-mediated antiviral defense. Interestingly, unlike Pseudomonas bacterial effectors HopF2, HopAI1, and AvrRpt2, or the βC1 encoded by a satellite associated with a Fig. 7 CP suppresses the MAPKKKα-mediated immune signaling by targeting Nb14-3-3a. a Interaction with 14-3-3a is required for CP to inhibit MAPKKKα-induced cell death. Agrobacterium carrying wild-type CP or its variants was co-infiltrated with Agrobacterium containing HA-MAPKKKα controlled by an estradiol-inducible system into different leaf regions. β-estradiol was used to induce protein expression 48 h after agroinfiltration. Leaves were stained by trypan blue and representative photographs were taken three days after estradiol treatment. b Quantification of cell death by measuring electrolyte leakage of leaf regions shown in (a). Error bars indicate ± SD of the mean (n = 6 biologically independent plants). Different letters in the chart denote statistically significant differences among different groups according to the one-way ANOVA analysis with Tukey's multiple comparison test (P < 0.05). c Western blot analysis to confirm the expression of the CP or its variants with an anti-CP antibody. d CP suppresses MAPKKKα protein accumulation and MAPK activation in WT N. benthamiana. Agrobacterium carrying CP, CP Y194A , or empty vector (EV) control was co-infiltrated with Agrobacterium containing HA-MAPKKKα controlled by an estradiol-inducible system into leaves. 30 μM β-estradiol was infiltrated into the leaves 48 h after agroinfiltration. Samples were collected at 0, 12, 24 h after β-estradiol treatment and subjected to western blot analysis with different antibodies indicated on the right. Actin served as the loading control. e CP suppresses MAPKKKα protein accumulation and MAPK activation in 14-3-3a-OE transgenic N. benthamiana. Agrobacterium carrying CP, CP Y194A or empty vector was co-infiltrated with Agrobacterium containing HA-MAPKKKα controlled by an estradiol-inducible system into leaves. In total, 30 μM β-estradiol was applied to the leaves 48 h after agroinfiltration. Samples were collected at 0, 12, 24 h after β-estradiol treatment and subjected to western blot analysis with different antibodies indicated on the right. CBB-stained RbcL served as the loading control. f CP-transgenic N. benthamiana exhibited reduced accumulation of HA-MAPKKKα protein and MAPK phosphorylation compared with WT plants.
Agrobacterium carrying HA-MAPKKKα controlled by an estradiol-inducible system was infiltrated into leaves. In all, 30 μM β-estradiol was applied to the leaves 48 h after agroinfiltration. Samples were collected at 0, 12, 24 h after β-estradiol treatment and subjected to western blot with different antibodies indicated on the right. Actin served as the loading control. For panels c-f, the experiments were repeated three times with similar results. g Symptom observation of engineered PVX-inoculated N. benthamiana plants. Agrobacterium carrying engineered PVX infection clones were infiltrated into two leaves of each plant. Representative photographs were taken at 6 dpi. h Western blot analysis of viral CP in the systemic leaves shown in panel g by using anti-PVX-CP or anti-BBSV CP antibody. Actin served as the loading control. The experiment was repeated three times with similar results.
DNA virus that directly interacts with MAPKs and inactivates them to compromise defense responses [33][34][35][36] , BBSV CP does not interact with MAPKKKα. Instead, it interacts with 14-3-3a to modulate MAPKKKα-mediated antiviral immunity. The 14-3-3 proteins are small acidic proteins (∼30 kDa) that are highly conserved in eukaryotes 60,61 14-3-3 proteins play important roles in signal transduction pathways to regulate diverse biological processes, such as cell division, primary metabolism, biotic and abiotic stress responses by interacting with various proteins in both plants and animals [62][63][64] . Accumulating evidence indicates that 14-3-3 proteins are involved in plant immunity 41,42,65 . 14-3-3 proteins interact with various signaling components of the plant immune system and modulate their activities, such as NADPH oxidase NtrbohD in N. tabacum 66 , plasma membrane H + -ATPase in maize and Arabidopsis 67 , R protein RPW8.2 in Arabidopsis 68 , and MAPKKKα in tomato 69 . Although these studies unveil the important roles of 14-3-3 proteins in plant defense to nonviral pathogens, their roles in plant virus infection remain unexplored. Our findings described here, for the first time, revealed that 14-3-3a protein plays an antiviral role. Although we cannot rule out the possibility that other 14-3-3 isoforms may also participate in the plant immune response, our results indicate that 14-3-3a is at least one of the key regulators.
In animals, 14-3-3 interacts with and activates two MAPKKKs, Raf-1 and MEKK3 70,71 . In plants, ZmMPK6 is reported to interact with a maize 14-3-3 protein GF14-6 72 . SIMAPKKKα is reported to interact with a tomato 14-3-3 protein TFT7 69 . These results prompted us to test whether NbMAPKKKα could interact with 14-3-3a in N. benthamiana. Our results confirmed their interaction and revealed that 14-3-3a is required for NbMAPKKKα-mediated cell death in N. benthamiana, which is in accordance with the positive regulation of SIMAPKKKα by TFT7 in tomato 69 . Our results also showed that 14-3-3a could protect the MAPKKKα from ubiquitination-mediated degradation (Fig. 4i). Genetic analysis using MAPKKKα-KO plants indicated that 14-3-3a participate in MAPKKKα-mediated defense against BBSV infection. As a counter-defensive strategy, BBSV uses its CP to suppress MAPKKKα-mediated defense by directly interacting with 14-3-3a that reduces the MAPKKKα accumulation. Previous studies also showed that downregulation of MAPKKKα expression increases the accumulation of PVX in N. benthamiana 49 . Interestingly, expression of BBSV CP in the PVX infectious clone resulted in enhanced accumulation of PVX and symptoms, further supporting the BBSV CP suppression of MAPKKKα-mediated plant defense. We also performed a timecourse analysis of the Nb14-3-3a expression during BBSV infection and found no significant changes in the Nb14-3-3a expression levels ( Supplementary Fig. 12b, c), consistent with the RNAseq data (Supplementary Data 5). Considering that the Nb14-3-3a-mediated antiviral defense depends on the MAPKKKα (Fig. 5) and the relatively high abundance of the endogenous plant 14-3-3 proteins ( Supplementary Fig. 12c), we speculate that the endogenous level of 14-3-3 might be sufficient to activate MAPK signaling.  Fig. 8 CP interferes with the MAPKKKα interaction with 14-3-3a in a dose-dependent manner. a Competitive Co-IP assay. His-CP and His-CP Y194A were purified from E. coli. 14-3-3a-3×Flag and HA-MAPKKKα K236M were co-expressed in N. benthamiana leaves and immunoprecipitated with anti-Flag beads. Anti-Flag beads with 14-3-3a-3×Flag and HA-MAPKKKα K236M proteins was incubated with increasing amounts of His-CP protein (0, 20, and 40 μg) or His-CP Y194A (40 μg). Input and IP proteins were analyzed by western blot with different antibodies indicated on the right. Asterisk indicates nonspecific bands. The experiment was repeated three times with similar results. b Competitive pull-down assay. MAPKKKα K236M -GFP was expressed in N. benthamiana leaves and immunoprecipitated with GFP-Trap agarose. After removing the supernatant, GFP-Trap agarose was incubated with E. coli-expressed His-14-3-3a followed by removal of the supernatant. GFP-Trap agarose was incubated with a serial dilution of His-CP or His-CP Y194A (1, 1/2, 1/4). The resultant agarose was analyzed by western blot with different antibodies indicated on the right. The experiment was repeated three times with similar results. c A proposed model for CP-Nb14-3-3a-MAPKKKα functional module in plant-virus interaction. Virus infection leads to the activation of the MAPK cascade-mediated antiviral defense. In turn, viral CP disturbs the interaction between 14-3-3a and MAPKKKα, resulting in the degradation of MAPKKKα and the attenuation of antiviral response.
Previous studies indicated that bacterial pathogen effectors target 14-3-3 proteins. For example, the XopQ effector from Xanthomonas oryzae pv. oryzae interacts with several rice 14-3-3 proteins to regulate immunity 73 . The effectors XopQ, XopE1, XopE2, and XopO from Xanthomonas euvesicatoria interact with multiple tomato 14-3-3 proteins to suppress ETI 74,75 . The effector XopN from Xanthomonas campestris pv. vesicatoria interacts with a tomato atypical receptor-like kinase 1 (TARK1) and TFT1 to suppress PTI 76,77 . Here, for the first time, we demonstrate that 14-3-3a is targeted by a viral CP effector. Based on our findings, we propose a model illustrating the MAPKKKαmediated defense against viruses and how viral CP counteracts ARTICLE the MAPKKKα-mediated immunity (Fig. 8c). Upon perception of viral infection, MAPKKKα is activated by some unknown components and 14-3-3a binds to and enhances the stability of MAPKKKα, leading to the activation of MAPK signaling and concomitant induction of plant defense-related genes, which may contribute to immunity against virus infection. To counteract the MAPKKKα-mediated antiviral immunity, CP is translated from viral RNA (vRNA) and increased amount of CP accumulates during the viral infection, it interacts with 14-3-3a and disrupts the 14-3-3a-MAPKKKα module. The released MAPKKKα is degraded by the 26S proteasome, leading to the impairment of MAPKKKα-mediated antiviral resistance (Fig. 8c). Our findings provide novel insights towards a better understanding of defense and counter-defense during plant-virus interplay. Our results also show that CPs of other necroviruses interfere with 14-3-3a-MAPKKKα immune signaling module as a counter-defensive strategy to promote viral infection. Moreover, we also noted that CPs of numerous viruses in the family Tombusviridae contain the 14-3-3-binding-like motif 78 , implying a general mechanism employed by the viruses in the Tombusviridae to subvert plant immunity.
In summary, our findings described here deepen our understanding of the MAPKKKα-mediated defense against viruses and reveal a novel mechanism for the viral CP in counteracting MAPKKKα-mediated antiviral immunity.

Methods
Plant materials and growth conditions. Nicotiana benthamiana plants were grown in a growth chamber at 23-24°C with a 15-h-light/9-h-dark cycle.
Generation of transgenic N. benthamiana plants. 14-3-3a-OE and 14-3-3a-KD transgenic N. benthamiana plants were generated by Agrobacterium-mediated transformation with binary vectors 14-3-3a-3×Flag and 14-3-3a-RNAi, respectively. MAPKKKα-KO transgenic N. benthamiana plants were generated by Agrobacterium-mediated transformation with binary vector BGK01-MAPKKKα. Single guide RNA targeting open reading frame of MAPKKKα was designed using CRISPR-P 2.0 design tool (http://crispr.hzau.edu.cn/CRISPR2/). Plasmids were transformed into A. tumefaciens strain EHA105. Leaf disc transformation was performed to generate the transgenic N. benthamiana plants 79 . Western blots with anti-14-3-3a antibodies were performed to screen the positive 14-3-3a-OE and 14-3-3a-KD transgenic plants. To screen the positive NbMAPKKKα knockout plants, DNA fragment corresponding to the gRNA-targeted DNA sequence regions was amplified from genomic DNA of transgenic plants using corresponding primers listed in Supplementary Data 6, and DNA sequencing was carried out to examine the insertion or deletion at the target site.
Construction of plasmids. BBSV derivatives used in this study were generated from pCB301-BBSV 80 . The sequence of sfGFP11 and linker was amplified using synthetic oligos and introduced into N-terminal of CP in BBSV infectious clone using reverse PCR to generate BBSV-sfGFP 11 . Single amino acid mutation within the BBSV CP was constructed by using the QuickChange Site-Directed Mutagenesis Kit (Agilent Technologies, Santa Clara, CA, USA) according to the manufacturer's instructions.
For transient expression in plants, plasmid pMDC32-CP Y194A and pMDC32-CP Y194F were constructed by amplifying the CP fragments from infectious clones of mutant BBSV containing Y194A or Y194F respectively and recombined into Kpn I and Spe I-digested pMDC32 plasmid 81 using the Seamless Assembly Cloning Kit (Clone Smarter). Nb14-3-3a was amplified from the cDNA of N. benthamiana and cloned into pMD19-T vector (Takara). The obtained plasmid was used for sitedirected mutagenesis by using a QuickChange Site-Directed Mutagenesis Kit. The DNA fragment encoding wild-type or mutant Nb14-3-3a was amplified and cloned into pMDC32-3×Flag using the Seamless Assembly Cloning Kit (Clone Smarter) to generate 14-3-3a-3×Flag, 14-3-3a 3E -3×Flag, and 14-3-3a QQR -3×Flag respectively. The construction of the plasmid pMDC32-3×Flag was described previously 82 . For the construction of HA-MAPKKKα K236M , NbMAPKKKα was amplified from the cDNA of N. benthamiana and cloned into the pGD-3×HA vector, an updated version of pGD vector 83 that was engineered with an N-terminal triple HA epitope. The resultant plasmid was used for site-directed mutagenesis using a QuickChange Site-Directed Mutagenesis Kit. To generate XVE::HA-MAPKKKα, MAPKKKα with a N-terminal triple HA epitope was amplified and recombined into pDONR207 vector (Invitrogen) by BP recombination followed by LR recombination into destination vector SPDK2130 under the estradiol-inducible system 84 . The other plasmids including pMDC32-CP, CP-3×Flag, and GFP-3×Flag were described previously 80 .
For TRV-mediated gene silencing, the cDNA fragments corresponding to the putative NbMEK2, NbSIPK and NbWIPK were amplified using cDNA of N. benthamiana and cloned into pYL156 85 .
To generate 14-3-3a-RNAi and MAPKKKα-RNAi vector, 300 nucleotides from 14-3-3a cDNA and 1-322 nucleotides from MAPKKKα cDNA were amplified and recombined into pDONR207 vector (Invitrogen) by BP recombination respectively. The interference fragments were then introduced into the destination vector pHELLSGATE 8 87 via LR reaction to generate RNAi expression vector.
For prokaryotic expression experiments, Nb14-3-3a gene was amplified and cloned into EcoR I-digested pGEX-KG plasmid to generate GST-14-3-3a. His-CP, His-CP Y194A , and His-CP Y194F were constructed by amplifying the CP fragments from the infectious clones of wild-type or mutant BBSV and ligated into pET30a (+) at the BamHI and SalI restriction sites. The construction of GFP-His was described previously 80 .
For PVX infection assay, BBSV CP and CP Y194A were cloned into the pND108 vector 88 at the XhoI and ApaI sites to yield PVX-CP and PVX-CP Y194A .
The primers used for constructing these plasmids are listed in Supplementary Data 6 and DNA sequencing was performed to confirm the correctness of all the plasmids.
Agroinfiltration and viral inoculation. All plasmids were transformed into Agrobacterium tumefaciens strains GV3101 or EHA105 by freeze-thaw transformation 89 . After being cultured in LB medium containing 25 μg/mL kanamycin and 100 μg/mL rifampicin at 28°C for 14-16 h, Agrobacterium cells were harvested and resuspended in infiltration media (10 mM MgCl 2 , 150 μM Fig. 9 Perturbation of Nb14-3-3a-MAPKKKα functional module by the CP is a common strategy employed by necroviruses. a, b BiFC analysis of the interaction between CP of TNV-A C (a) or TNV-D H (b) and 14-3-3a. The indicated combinations of different proteins were transiently co-expressed in N. benthamiana leaves. Confocal analysis was performed at 3 dpi. The experiments were repeated three times and representative results are shown. Scale bars = 50 µm. c, d CP inhibits MAPKKKα-induced cell death. Agrobacterium carrying wild-type TNV-A C (c) or TNV-D H (d) CP or their variants was coinfiltrated with Agrobacterium containing HA-MAPKKKα controlled by an estradiol-inducible system into different leaf regions. Empty vector (EV) served as the negative control. β-estradiol was applied 48 h after agroinfiltration. Leaves were stained by trypan blue and representative photographs were taken 3 days after estradiol treatment. e, f Quantification of cell death by measuring electrolyte leakage of leaf regions shown in (c) and (d). Error bars indicate ± SD of the mean (n = 6 biologically independent plants). Different letters in the chart denote statistically significant differences among different groups according to the one-way ANOVA analysis with Tukey's multiple comparison test (P < 0.05). g, h CP affects MAPKKKα protein stability and MAPK activation in non-transgenic N. benthamiana. Agrobacterium carrying CP or CP Y194A was co-infiltrated with Agrobacterium containing HA-MAPKKKα controlled by an estradiol-inducible system into the leaves. Empty vector (EV) served as the negative control. In all, 30 μM β-estradiol was infiltrated into the leaves 48 h after agroinfiltration. Samples were collected at 0, 12, 24 h after β-estradiol treatment and subjected to western blot with different antibodies indicated on the right. Actin served as the loading control. The experiments were repeated three times with similar results. i, j Phenotype observation of TNV-A C (i) or TNV-D H (j) -infected MAPKKKα-KO plants. N. benthamiana leaves were inoculated with 100 ng TNV-A C virions. Representative photographs were taken at 9 dpi. k Western blot analysis of TNV-A C CP levels in the plants shown in panel i using an anti-TNV-A C antibody. Actin served as the loading control. The experiment was repeated three times with similar results. l RT-qPCR analysis of TNV-D H genomic RNA levels in the plants shown in panel j. Values represent ± SD of the mean (n = 6 biologically independent plants). An asterisk indicates the significant difference based on one-sided Student's t test (*P = 0.0357).
For viral infection analysis, 300 ng of BBSV or TNV-A C virions mixed with FES inoculation buffer (0.1 M glycine, 0.06 M dipotassium phosphate, 1% sodium pyrophosphate decahydrate, 1% bentonite, 1% celite, pH 8.5) at a 1:1 ratio followed by mechanical inoculation onto 4-5 leaf stage N. benthamiana plants. RNA transcripts of TNV-D H were inoculated as described previously 82 . The SmaI-linearized plasmid A172S 90 (200 ng) were used as templates for in vitro transcription with T7 RNA polymerase (Promega). Freshly prepared in vitro transcripts were analyzed by 1% (wt/vol) agarose gel electrophoresis to evaluate the quality of the RNA transcripts. The RNA transcripts were then mixed with an equal volume of FES inoculation buffer at a 1:1 ratio followed by mechanical inoculation onto 4-5 leaf stage N. benthamiana plants.
RNA-Seq analysis. The leaf samples from Mock-and BBSV-inoculated plants were collected at the indicated dpi. Three leaves from different plants were collected as one biological replicate, and three biological replicates were used for each group. Total RNAs were extracted using TRIzol reagent (Invitrogen). After quantification and qualification, total RNAs were pooled for cDNA library construction. cDNA library preparation and sequencing were carried out on an Illumina NovaSeq 6000 platform with 150-bp pair-end reads in Novogene (Beijing, China). Total reads were mapped to the Nicotiana benthamiana transcriptome (http://sefapps02.qut.edu.au/downloads/Nbv0.5.genome.fa.gz) using Hisat2 (v2.0.5), and the mapped reads of each sample were assembled by StringTie (v1.3.3b) in a reference-based approach. The expression level of each gene was calculated based on expected number of Fragments Per Kilobase of transcript sequence per Million's base pairs sequenced (FPKM). Differential expression analysis of two groups was performed using the DESeq2 R package (1.20.0). The resulting P values were adjusted using Benjamini and Hochberg's approach for controlling the false discovery rate (Padj). Padj < 0.05 and |log2(foldchange)| >1 were set as the threshold for significantly differential expression.
Mass spectrometry analysis. Silver-stained protein band as indicated by the rectangle in Fig. 3c was excised from SDS-PAGE gels, the gel band was destained with Sigma Proteosilver kit (PROTSIL1, Sigma), and then reduced with 10 mM DTT, alkylated with 55 mM iodoacetamide, and digested with trypsin (pH 8.5) at 37°C for 12 h. The digested peptides were separated by nanoscale C18 reversephase liquid chromatography (Waters), and then electro-sprayed into a Q-Exactive high-resolution mass spectrometer (Thermo Fisher Scientific) at the Mass Spectrometry Facility of China Agricultural University. Protein identification was performed by searching against the protein database of either the National Center for Biotechnology Information (NCBI) or the N. benthamiana sequence v1.0.1 proteome (ftp://ftp.solgenomics.net/genomes/Nicotiana_benthamiana) using Mascot Server (version 2.5.1, Matrix Science).
BiFC assay. Agrobacterium harboring various BiFC constructs were mixed with a final OD 600 = 0.3 and co-infiltrated into 4-week-old N. benthamiana leaves. At 3 dpi, the leaf samples were observed under a Zeiss LSM880 confocal microscope. YFP signals were excited at 514 nm.
Expression and purification of recombinant proteins. Protein purification was performed as described previously 82 . Expression constructs were transformed into E. coli strains BL21 or Rosetta. Cultures were grown at 37°C until OD 600 = 0.8, and protein expression was induced by 0.2 mM isopropyl-β-D-thiogalactopyranoside (IPTG) at 18°C overnight. Cells were then collected and subjected to protein purification using Glutathione Sepharose TM 4 Fast Flow (GE Healthcare) or Ni-NTA agarose (QIAGEN) according to the manufacturer's instructions.
Trypan blue staining. Trypan blue staining was performed as previously described with minor modifications 92 . The infiltrated leaves were pretreated with ethanol for 3 min, and then combined with trypan blue staining solution prepared in 40 mL of ethanol, 10 mL of lactic acid, 10 mL of water phenol, 10 mL of glycerol, and 10 mL of sterile water with 15 mg of Trypan blue. The samples were placed in boiling water for 10 min and incubated at room temperature for 8 h. The samples were destained using chloral hydrate (2.5 g/mL sterile water) followed by photography.
Electrolyte-leakage assay. The electrolyte-leakage assay was performed as described previously with minor modifications 93 . Two leaf discs (9 mm in diameter) were collected from infiltrated areas of six to eight plants and floated in 15-mL tubes containing 10 mL of MilliQ-water for 1 h at 25°C with shaking at 165 rpm. Conductivity was measured using a DDS-12DW conductivity meter (BANTE) to obtain S1 value. The samples were placed in boiling water for 45 min, followed by shaking at 25°C for 1 h, and measured to obtain the S2 value. The ratio of S1 to S2 was calculated as electrolyte leakage.
Reporting summary. Further information on research design is available in the Nature Research Reporting Summary linked to this article.

Data availability
Data supporting the findings of this work are available within the paper and its Supplementary Information files, or from the corresponding author upon request. The raw RNA-seq data are available in the NCBI database under accession code PRJNA767174. Source data are provided with this paper.
Received: 2 March 2021; Accepted: 17 January 2022; Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/ licenses/by/4.0/.