NF-YA transcription factors suppress jasmonic acid-mediated antiviral defense and facilitate viral infection in rice

NF-Y transcription factors are known to play many diverse roles in the development and physiological responses of plants but little is known about their role in plant defense. Here, we demonstrate the negative roles of rice NF-YA family genes in antiviral defense against two different plant viruses, Rice stripe virus (RSV, Tenuivirus) and Southern rice black-streaked dwarf virus (SRBSDV, Fijivirus). RSV and SRBSDV both induced the expression of OsNF-YA family genes. Overexpression of OsNF-YAs enhanced rice susceptibility to virus infection, while OsNF-YAs RNAi mutants were more resistant. Transcriptome sequencing showed that the expression of jasmonic acid (JA)-related genes was significantly decreased in plants overexpressing OsNF-YA when they were infected by viruses. qRT-PCR and JA sensitivity assays confirmed that OsNF-YAs play negative roles in regulating the JA pathway. Further experiments showed that OsNF-YAs physically interact with JA signaling transcription factors OsMYC2/3 and interfere with JA signaling by dissociating the OsMYC2/3-OsMED25 complex, which inhibits the transcriptional activation activity of OsMYC2/3. Together, our results reveal that OsNF-YAs broadly inhibit plant antiviral defense by repressing JA signaling pathways, and provide new insight into how OsNF-YAs are directly associated with the JA pathway.


Author summary
In the arms race between host and viruses, plants have sophisticated defense strategies to cope with pathogenic invasion. Our previous research has shown that plant hormones, especially JA signaling, play essential roles in plant antiviral defense. Therefore, identification of the regulatory components and network of JA signaling is an important contribution to understanding plant antiviral defense. Here, we found that NF-YAs widely repress plant resistance against different types of viruses. We show that NF-YAs inhibit JA-mediated antiviral defense by dissociating the OsMYC2/3-OsMED25 complex, and inhibiting

Introduction
Among the various viruses that cause severe economic losses in rice production around the world, Southern rice black-streaked dwarf virus (SRBSDV) and Rice stripe virus (RSV) are currently widespread in east Asia and relatively well-studied. SRBSDV belongs to the genus Fijivirus in the family Reoviridae and is mainly transmitted in a persistent propagative manner by Sogatella furcifera (the white-backed planthopper, WBPH) [1,2]. The SRBSDV genome has 10 segments (S1-S10) of linear double-stranded RNA (dsRNA). Plants infected by SRBSDV are severely stunted and have excessive tillering. RSV is classified in the genus Tenuivirus (family Phenuiviridae) and is transmitted in a persistent propagative manner by Laodelphax striatellus (the small brown planthopper, SBPH) [3,4]. RSV has a single-stranded RNA (ssRNA) genome of four segments. RSV infection usually causes chlorotic or necrotic stripes on newly developed leaves, which is followed by wilting and stunting. Plant hormone signaling pathways play important roles in response to infection of rice by viruses. For example, rice STV11 encodes a sulfotransferase, which can catalyze the conversion of salicylic acid (SA) into the more effective sulfonated SA (SSA) and enhance the resistance of rice to RSV [5]. Rice dwarf virus (RDV)-encoded Pns11 protein promotes ethylene production which weakens the plant resistance to viral infection [6]. The multiple ARF transcription factors in rice have diverse regulatory functions on rice antiviral defense [7]. Our previous studies showed that different phytohormones play diverse roles in antiviral defense response. For example, the jasmonic acid (JA) pathway positively regulated rice defense, and suppressed brassinosteroids (BR)-mediated susceptibility to viral infection [8]. Abscisic acid (ABA) increased the susceptibility of rice to virus by inhibiting JA pathways [9]. BR mediates the susceptibility to Rice black-streaked dwarf virus (RBSDV) infection by inhibiting the plant defense response mediated by JA and SA and reducing the accumulation of reactive oxygen species (ROS) mediated by peroxidase [10]. Activation of the JA pathway was part of the auxin signalling-mediated antiviral defense [11]. It has recently been shown that JA signaling activates RNA silencing, and synergistically promotes rice antiviral defense [12]. The accumulation of miR319 by Rice ragged stunt virus (RRSV) in rice suppresses JA-mediated defense to promote viral infection and symptom development [13]. These studies all illustrate that JA signaling plays essential roles in rice antiviral defense. Therefore, identification of the regulatory components and network of JA signaling is an important contribution to understanding plant antiviral defense.
Nuclear factor Y (NF-Y) transcription factors act as essential regulators of growth developmental process and stress responses. Because they recognize the CCAAT-box, NF-Ys are also called CCAAT-binding factors (CBFs) or Heme Activator Proteins (HAPs) [14][15][16]. Classically, NF-Y is a heterotrimeric transcription factor complex consisting of three subunits: NF-YA (HAP2/CBF-B), NF-YB (HAP3/CBF-A) and NF-YC (HAP5/CBF-C) subunits, which are highly conserved in most species [17,18]. Usually, NF-YB and NF-YC dimerize in the cytoplasm and move to interact with NF-YA in the nucleus [19]. This hetero-trimeric complex then binds to the CCAAT-box element to activate or repress gene expression. NF-Y transcription factors are encoded by a single-gene in animals and yeast. However, in plants, each NF-Y subunit type is encoded by 10 or more genes due to gene duplication events [20]. For example, the Arabidopsis genome encodes 10 NF-YA, 13 NF-YB and 13 NF-YC genes [21], while rice has 11 NF-YA, 11 NF-YB and 12 NF-YC genes [22]. This multiplicity suggests a substantial potential regulatory capacity.
NF-Y family members are characterized by the presence of highly conserved domains. NF-YA proteins have two conserved domains (subunit interactions and DNA binding domains). The subunit interactions domain is crucial for the interaction with NF-YB and NF-YC, and the DNA binding domain is required for sequence-specific binding to CCAAT-boxes. Numerous studies have identified the NF-Y genes that play vital roles in plant developmental and physiological responses. The first plant NF-Y regulating development was identified and reported in Arabidopsis [23] and was found to be involved in embryogenesis [24]. In Arabidopsis, NF-YC3, NF-YC4 and NF-YC9 can interact with CONSTANS (CO), a key gene regulating flowering, to promote the photoperiod-dependent flowering process [25]. The NF-Y complex promotes flavonoid biosynthesis by binding CCAAT elements and recruiting histone methyltransferase or demethylase to regulate H3K27me3 levels in tomato [26]. In rice, OsNF-YB1 forms a heterotrimeric complex with OsNF-YC12 and bHLH144 to directly regulate the transcription of Wx and regulate the synthesis of starch [27]. OsNF-YA7 was induced by drought and improved rice tolerance to drought stress [28]. In addition, some NF-YA family members were shown to modulate plant defense against pathogens. For example, overexpression of OsHAP2E (OsNF-YA2) conferred resistance in rice to the fungal pathogen Magnaporthe oryzae and the leaf blight bacterium Xanthomonas oryzae pv. Oryzae [29]. Down-regulation of the OsNF-YA family members by overexpressing miR169a makes rice hypersusceptible to M. oryzae [30]. These studies showed that OsNF-YAs positively regulate plant defense against fungal and bacterial pathogens but the relationship between NF-YAs and viral pathogens has never been investigated, and it is unclear whether and how NF-YAs affect JA-mediated antiviral defense.
In this study, we explored the roles of OsNF-YAs in defense against viruses. In contrast to their positive roles against fungal and bacterial pathogens, OsNF-YAs negatively regulated rice antiviral defense against two different RNA viruses, RSV and SRBSDV. We found that OsN-F-YAs directly interacted with OsMYC2/3, which was shown to be the key transcription factor in the JA pathway. We provide evidence that OsNF-YAs destroy the interaction between OsMYCs-OsMED25 complexes, and repress the transcriptional activation activity of OsMYC2/3, finally impairing JA-mediated antiviral defense.

Infection by RSV or SRBSDV induces the expression of OsNF-YAs
To investigate the roles of OsNF-YA in defense against viruses, the expression levels of OsNF-YA family genes were analyzed following infection by viruses. Oryza. sativa L. japonica Nipponbare (NIP) was used for viral inoculation experiments because it is susceptible to both RSV and SRBSDV. Results of qRT-PCR showed that a number of OsNF-YA family genes, including OsN-F-YA1, OsNF-YA2 and OsNF-YA10, were significantly induced 30 day after inoculation with RSV, a single-stranded RNA virus (Fig 1A). Similarly, the expression level of most OsNF-YAs was obviously increased 30 dpi with SRBSDV, a double-stranded RNA virus (Fig 1B). The finding that two different viruses both induced the expression of OsNF-YA family genes, suggested that OsNF-YAs might be generally involved in the interaction between rice and viruses.

Overexpression of OsNF-YAs increased susceptibility to infection by RSV or SRBSDV
We next constructed transgenic rice lines overexpressing OsNF-YAs and inoculated them with viruses. Stable transgenic lines expressing three representative OsNF-YA genes (OsNF-YA1, OsNF-YA2, OsNF-YA10) were created in the NIP background and the expression levels of these genes were verified by qRT-PCR (S1A Fig). The homozygous T2 generation transgenic rice lines OE-OsNF-YA1#3 and #5, OE-OsNF-YA2#4 and #6, OE-OsNF-YA10#1 and #2 were then first inoculated with RSV. The severity of RSV symptoms that developed on each plant was assessed using the four grades: healthy (N), mild mosaic (I), severe mosaic (II) or wilting (III) (S2 Fig). Symptoms were more severe in the overexpression lines than in the NIP control (Figs 1C and S3A). The transgenic lines had severe dwarfing, and many plants were wilting (grade III) or had severe mosaic (grade II) (Figs 1D and S3B). Similarly, the accumulation of RSV RNAs (CP and P2) and CP protein in the overexpressed transgenic lines were significantly higher than those of NIP (Figs 1E and 1F, and S3C and S3D).
The OE-OsNF-YAs transgenic lines were then inoculated with the double-stranded RNA virus SRBSDV. More severe stunting symptoms were observed in SRBSDV-infected transgenic

PLOS PATHOGENS
NF-YAs repress JA-mediated antiviral immunity in rice lines than in control plants ( Fig 1G). Viral incidence was significantly higher in both OE-OsN-F-YA1 and OE-OsNF-YA2 transgenic lines than in NIP plants ( Fig 1H). qRT-PCR and western blot experiments showed that viral RNA transcripts (S2, S4 and S6) and caspid protein P10 were all obviously more abundant in the transgenic lines than in the controls (Fig 1I and 1J). Thus, overexpression of OsNF-YAs made rice more susceptible to both the ssRNA virus RSV and the dsRNA virus SRBSDV.
To investigate any effect of the OsNF-YAs transgenic plants on the insect vectors, we tested the survival of SPBH and WBPH on the different plants over a 5-day period but found no significant differences in survival between the transgenic and control plants (S1C and S1D Fig).

Suppressing the expression of OsNF-YAs increased resistance to viruses
To further study the relationship between OsNF-YAs and viral infection, we next constructed stable transgenic rice lines with decreased expression of OsNF-YAs by RNA interference. Because of the functional redundancy among OsNF-YAs, we constructed these lines using a conserved region of the OsNF-YA family genes and obtained lines named RNAi-OsNF-YA#7 and RNAi-OsN-F-YA#8 that had efficiently suppressed expression of OsNF-YAs (S1B Fig). When these lines were challenged with RSV, symptoms were less severe than in the NIP controls, with less stunting, and a larger percentage of plants that were healthy (grade N) or which had only slight mosaic symptoms (grade I) (Fig 2A and 2B). Consistently, RSV RNAs (CP and P2) and CP protein were significantly lower in RNAi-OsNF-YA lines than in wild type NIP (Fig 2C and 2D). The RNAi-OsNF-YA lines were also more resistant to SRBSDV, with less dwarfism (Fig 2E) and significantly lower viral incidence ( Fig 2F) than in the control NIP ( Fig 2F). qRT-PCR and western blot experiments also confirmed that the accumulation of SRBSDV RNA transcripts (S2, S4 and S6) and CP protein were significantly reduced in the transgenic lines (Fig 2G and 2H). Thus repression of the expression of OsNF-YAs increased rice resistance to both RSV and SRBSDV.

Transcriptomic analysis suggests that OsNF-YAs modulate the JA pathway
To understand how OsNF-YAs negatively regulate the antiviral immune response, we performed a genome-wide transcriptomic analysis on NIP and OE-OsNF-YA2#4 transgenic plants infected with RSV at 20 dpi. In order to reduce the likelihood of false positives, we used a pvalue less than 0.05 as the cutoff criterion. We identified 4620, 971 and 7013 differentially upregulated genes in the respective subsets OE-OsNF-YA2-CK versus NIP-CK, OE-OsNF-YA2-RSV versus NIP-RSV, and NIP-RSV versus NIP-CK ( Fig 3A). Venn diagram analysis showed that about half of the genes (429/971) were specifically up-regulated in OsNF-YA2 transgenic plants after RSV infection ( Fig 3A). Hierarchical clustering analysis showed these 429 genes were indeed specially activated in the OsNF-YA2 plants compared to the control NIP plants ( Fig  3B). Gene ontology (GO) analysis clearly revealed that these genes are highly enriched in the categories of plasma membrane, chloroplast, DNA-binding transcript factor and protein phosphorylation ( Fig 3C). Meanwhile, we identified 1483, 754 and 2753 differentially down-regulated genes in the respective subsets OE-OsNF-YA2-CK versus NIP-CK, OE-OsNF-YA2-RSV versus NIP-RSV, and NIP-RSV versus NIP-CK. Interestingly, almost all the down-regulated genes (634/754) were manipulated by the OsNF-YA2 in the context of RSV infection ( Fig 3D). Hierarchical clustering revealed that these genes were obviously suppressed in the OE-OsN-F-YA2 plants compared to the control NIP plants under RSV infection ( Fig 3E). Further gene ontology (GO) analysis identified that a large number of defense-related genes, especially in the JA pathway, were specifically down-regulated in OE-OsNF-YA2 plants infected by RSV ( Fig  3F). These data imply that OsNF-YA2 may repress the JA-related pathway in response to RSV infection, powerfully hinting at a role for OsNF-YA2 in the JA-mediated immune response.

PLOS PATHOGENS
NF-YAs repress JA-mediated antiviral immunity in rice

OsNF-YAs negatively regulate the JA pathway
To confirm the association of OsNF-YA with the JA pathway, we used qRT-PCR to determine the expression levels of the JA-related genes (OsJAZ8, OsJAZ9, OsJAZ11, OsJAZ12, OsLOX2 and JiOsPR10) in OsNF-YA2 transgenic lines challenged by RSV. In control NIP plants, the These results indicate that OsNF-YA2 negatively regulates the expression of JA-related genes in response to RSV or SRBSDV infection.
We then examined whether OsNF-YA2 affects JA sensitivity by exogenous application of methyl jasmonate (MeJA) to rice in hydroponic culture. The root lengths of wild-type and OsNF-YA2 transgenic plants were measured 5 days after treatment with 2 or 5 μM MeJA. In the presence of 2 μM MeJA, the root lengths of OE-OsNF-YA2 lines were less reduced than those of the wild-type plants (40% vs 55%). However, MeJA-treated RNAi-OsNF-YA lines had a greater reduction in root length than NIP plants (65% vs 55%) (Fig 4B and 4C). There was a similar effect from 5 μM MeJA treatment. We also performed JA sensitivity assays using stable transgenic rice lines OE-OsNF-YA1 (#3 and #5) and OE-OsNF-YA10 (#1 and #2) and found that the root lengths of these OE-OsNF-YAs lines were less sensitive to MeJA treatment than those of the wild-type plants (S5 Fig). These results indicate that overexpression of OsNF-YAs makes rice plants less sensitive to JA treatment whereas repression of OsNF-YAs increases sensitivity and therefore show that OsNF-YAs have negative roles in regulating the JA pathway.

OsNF-YAs interact with JA key transcription factor OsMYC2/3
Given the negative roles of OsNF-YAs in the JA pathway, we wondered whether OsNF-YAs directly modulate the JA pathway by interacting with the essential components of JA signaling. We have recently shown that JA signaling key transcription factors OsMYC2/3 play vital roles in antiviral immunity against both RSV and SRBSDV [32]. Thus, we hypothesized that OsN-F-YAs might physically associate with OsMYC2/3. To test this hypothesis, we first used a yeast two-hybrid (Y2H) assay which showed that OsMYC2 and OsMYC3, but not OsMYC4, interacted with several OsNF-YA family genes, including OsNF-YA1 and OsNF-YA2 (Fig 5A). Coimmunoprecipitation (Co-IP) assays further confirmed that OsMYC2/3 but not the negative

OsNF-YAs inhibit the transcriptional activation of OsMYC2/3
We next tested which domain of OsMYC2/3 was responsible for their association with OsN-F-YAs. The transcription factors MYC2/3 usually have a putative transcriptional activation domain (TAD) and a basic-helix-loop-helix domain (bHLH) [33,34]. Based on the conserved sequence characteristics, we constructed several OsMYC3 mutants and found TAD domain was required for the association with OsNF-YA1 and OsNF-YA2 (Fig 6A). Given that the TAD domain is responsible for OsMYC2/3 transcriptional activation, we tested whether OsN-F-YAs affect the transcriptional activation activity of OsMYC2/3 using a dual-LUC reporter system in N. benthamiana leaves. The control BD-Ven (BD, which specifically binds the 5×GAL4 cis element, fused to the fluorescent protein Venus) as the control, BD-OsMYC2, BD-OsMYC3, OsNF-YA1 and OsNF-YA2 act as the effectors (Fig 6B). Consistent with our previous reports [32], BD-OsMYC2 and BD-OsMYC3 have obvious transcriptional activation activity compared with the control BD. However, the transcriptional activation activity of BD-OsMYC2/3 was significantly impaired in the presence of OsNF-YA1 or OsNF-YA2 ( Fig  6C-6F). Together, these results suggest that OsNF-YAs inhibit the transcriptional activation of OsMYC2/3.
In the JA signaling pathway, MED25 (Mediator complex subunit 25) is a subunit of the mediating complex, and constitutes the transcription-activating complex with MYC2/3 by binding to the TAD domain [35,36]. Y2H assays showed the TAD motif of OsMYC3 is required for interaction with both OsMED25 and OsNF-YA1/2 (Figs 6A and S6). We then hypothesized that OsNF-YAs might compete with OsMED25 to bind OsMYC2/3. We first designed protein competition Co-IP assays in N. benthamiana. Immunoblot analysis indicated that OsMED25 can bind to OsMYC2/3 when co-infiltrated with the negative control HA-GFP, but the association between OsMED25 and OsMYC2/3 was significantly reduced in the presence of OsNF-YAs (Fig 6G-6J). A BiFC experiment was also used as a protein competition assay in N. benthamiana leaves. The fluorescence in the combinations of OsMED25-nYFP and cYFP-OsMYC2/3 was observed in presence of the control Gus-myc, but was obviously weakened in the presence of OsNF-YAs (S7A and S7B Fig). In addition, we then conducted a split luciferase complementation imaging assay which showed that the co-expression of either nLUC-OsMYC2 or nLUC-OsMYC3 with cLUC-OsMED25 resulted in effective luciferase activity in N. benthamiana leaves when co-infiltrated with the negative control HA-GFP, but that this activity was reduced in the presence of OsNF-YAs (S7C and S7D Fig). Results of qRT-PCR assays showed that there were no significant differences in transcript expression of MYC2/3 and MED25 in these BiFC and LCI assays (S8 Fig). Taken together, these results show that OsNF-YAs repress JA signaling by directly disturbing the OsMYC2/3-OsMED25 complex and impeding the transcriptional activation activity of OsMYC2/3.

PLOS PATHOGENS
NF-YAs repress JA-mediated antiviral immunity in rice

PLOS PATHOGENS
NF-YAs repress JA-mediated antiviral immunity in rice

Discussion
The NF-Y transcription factors complex is composed of three subunits, NF-YA, -YB and -YC, and NF-YA plays diverse roles in a variety of developmental processes including plant growth, development and stress responses [37]. For instance, lines overexpressing AtNF-YA5 were more resistant to drought, while nf-ya5 lines were hypersensitive to drought in comparison with wild-type plants [38]. In rice, overexpression of OsNF-YA7, but not OsNF-YA4, mediates drought tolerance in an ABA-independent manner, suggesting an important role for OsN-F-YA7 in the water acquisition signaling pathway [28]. In Arabidopsis, lines overexpressing AtNF-YA3, AtNF-YA7 and AtNF-YA10 were more sensitive to ABA [39]. In wheat (Triticum aestivum), overexpression of TaNF-YA10 increased sensitivity to salt stress [40]. While numerous studies have reported that OsNF-YAs participate in plant growth and developmental processes, their role(s) in plant resistance to pathogens have rarely been reported. Recently, overexpression of miR169a and suppression of its target OsNF-YAs was shown to enhance rice susceptibility to M. oryzae by reducing hydrogen peroxide accumulation, indicating that OsN-F-YAs positively modulate rice defense against fungal pathogens [30]. However, it is unclear whether OsNF-YAs have similar roles in the interactions between rice and viruses. Here, we showed that the expression of most OsNF-YAs was induced when rice was infected by either of two distinct viruses, RSV and SRBSDV (Fig 1A and 1B). Overexpression of three representative OsNF-YAs increased the susceptibility of rice to both viruses while OsNF-YAs RNAi mutant plants were more resistant than the controls (Figs 1 and 2 and S3). These results suggest that, in contrast to their positive roles in plant defense against fungal pathogens, OsN-F-YAs negatively regulate rice antiviral immunity.
Recent evidence has shown that NF-Y family members associate with plant hormone pathways to regulate plant growth development and stress responses. AtNF-YCs negatively regulate BR biosynthesis genes by stabilizing BIN2 to inhibit light-induced hypocotyl elongation [41]. Simultaneously, AtNF-YCs mediate seed germination by interacting with the ABA and GA pathway [42]. AtNF-YCs interact with the master regulators of ABA signaling ABF3 and ABF4 to regulate the expression of the SOC1 gene and accelerate flowering under drought stress in Arabidopsis thaliana [43]. AtNF-YA2 and AtNF-YA10 participate in the regulation of Arabidopsis leaf growth and development by directly binding to cis-CCAAT in the YUC2 promoter to inhibit auxin biosynthesis [44]. However, little has been reported about the direct interplay between NF-Y family proteins and the JA pathway. Here, transcriptome analysis and qRT-PCR showed that, in response to infection by rice viruses, the JA pathway was down-regulated when OsNF-YA2 was overexpressed but up-regulated in OsNF-YA RNAi mutants (Figs 3D-3F, 4 and S4B). Overexpression of OsNF-YAs made plants less sensitive to JA treatment (Figs 4A and S5). JA treatment significantly enhanced the resistance of wild-type rice plants to RSV infection compared to OE-OsNF-YA2 transgenic plants (43% VS 12%) (S9A Fig). qRT-PCR and Western blotting experiments confirmed that JA treatment significantly reduced the accumulation of RSV CP level in wild-type rice plants (51%), while the extent of JA treatment was much alleviated in OE-OsNF-YA2 transgenic plants (28% or 25%) (S9B and S9C Fig). These results indicated that OsNF-YAs make rice insensitive to JA treatment and inhibit JA-mediated antiviral defense. OsNF-YAs were shown to physically interact with OsMYC2/3 (Fig 5), directly disturbing the OsMYC2/3-OsMED25 complex by competing for the TAD domain (Figs S6 and 6G-6J). Together, our results showed that OsNF-YAs participate in and repress JA signaling by directly interacting with and impeding the transcriptional activity of OsMYC2/3.
Phytohormones play critical roles in various aspects of plant growth regulation and defense response. In rice, auxin and the JA pathway participate positively in defense against viruses,

PLOS PATHOGENS
NF-YAs repress JA-mediated antiviral immunity in rice the BR pathway negatively regulates resistance to RBSDV by interfering with the JA pathway and ABA inhibits JA-mediated antiviral defense [8][9][10][11]32,45,46]. Thus, the various synergistic or antagonistic interactions between phytohormones play important roles in the outcome of antiviral responses. To successfully infect plants, rice viruses encode pathogenicity factors that target the essential components of antiviral phytohormones to achieve infection. For example, we recently showed that different proteins encoded by viruses interact with auxin response factor OsARF17 and weaken the auxin-mediated antiviral defense response [46]. These different viral proteins can also hijack and repress the key components of the JA pathway (OsJAZs and OsMYC2/3), benefiting virus pathogenicity and vector transmission [32]. Recently, RSV coat protein (CP) was shown to act as an elicitor to trigger JA accumulation, and JA signaling synergized with RNA silencing to promote antiviral defense. These studies suggest that the JA pathway plays essential roles in the battlefield between rice and viruses. Our work here has identified some new regulators of the JA pathway, OsNF-YAs, which directly associate with the TAD domain of OsMYC2/3 disturb the OsMYC2/3-OsMED25 complex and repress their transcriptional transactivation.
Whereas OsNF-YAs have been shown to positively regulate plant defense against fungal and bacterial pathogens, our results show that they negatively impact defense against viruses by interfering with the JA-mediated antiviral defense. We propose a model in which the enhanced expression of OsNF-YAs following virus infection leads OsNF-YAs to interact with OsMYC2/3 transcription factors. This impedes their transcriptional activation activity by interfering with the transcriptional activation complex OsMYC2/3-OsMED25, finally resulting in repression of JA-mediated antiviral immunity (Fig 7).

Plant materials and growth conditions
All rice plants were grown in the greenhouse at 28-30˚C under 10 h light/14 h dark. To generate transgenic plants over-expressing OsNF-YAs, the full length of these genes was cloned and the expression vector was constructed by homologous arm recombination. The recombinant

PLOS PATHOGENS
NF-YAs repress JA-mediated antiviral immunity in rice plasmid was electroporated into Agrobacterium tumefaciens strain GV3101 and transformed into O. sativa L. japonica Nipponbare (NIP) background. RNAi-OsNF-YA lines with silenced OsNF-YAs expression were also constructed in the NIP background. Nicotiana benthamiana plants were grown in a growth chamber at 25˚C and with 16 h light/8 h darkness. The primers used are listed in S1 Table.

Insect vectors and virus inoculation assay
The rice virus inoculation methods have been detailed in our previous reports [11,32,46]. For RSV inoculation, 2-3 instar larvae of SBPH were allowed to feed on rice seedlings at the 4-leaf stage for 3 days, and then removed. The inoculated plants were then grown in the greenhouse for observation and were tested at 30 dpi by qRT-PCR and western blotting. Similar methods were used for SRBSDV inoculation. The specific primers used to detect SRBSDV/RSV are listed in S1 Table. The number of infected plants was determined following RT-PCR to calculate the viral incidence. All experiments were repeated at least three times with similar results.

Hormone treatments
Stock solutions of MeJA (Sigma) in 100% ethanol were diluted with sterile distilled water containing 0.1% Triton X-100. RSV-inoculated rice seedlings were then sprayed with 50 μM MeJA or 0.1% Triton X-100 as the mock control. Each treatment used at least 30 seedlings. Disease symptoms were observed 20 days post inoculation (dpi). Leaf samples were frozen in liquid nitrogen and stored at −80˚C for protein and RNA extraction.

Total RNA extraction and quantitative real-time PCR (qRT-PCR) analysis
Total RNA was extracted from rice leaves using the TRIzol reagent (Invitrogen, Carlsbad, CA, USA, Cat. no. 15596-026) according to the manufacturer's instructions. 1 μg of total RNA was mixed with 4 � gDNA wiper mix to eliminate genomic DNA at 42˚C for 2 minutes and then reverse transcribed to cDNA using 5 � HiScript III qRT Super Mix (Vazyme). qRT-PCR analysis was performed using the Hieff qPCR SYBR Green Master Mix (Low ROX Premixed) by the ABI7900HT Sequence Detection System (Applied Biosystems, Carlsbad, CA, USA). The reference gene OsUBQ5 (AK061988) was used to normalize the relative expression levels [47], and the method of 2 -ΔΔCt was used for analysis [48]. The experiments were repeated at least three times. Each biological sample consisted of more than 10 plants. The qRT-qPCR primers used in this study are listed in S1 Table.

Western blot analysis
Total proteins were extracted from infected rice leaves with 10% SDS lysis buffer (0.1M Tris-HCl (pH = 6.8), 10% SDS, 2% β-mercaptoethanol) and separated in 12% SDS-PAGE gels before transfer to a methanol pre-activated PVDF membrane. This was then incubated for 2h at room temperature with anti-CP or anti-P10 (provided by Prof. Jianxiang Wu) monoclonal antibodies to specifically recognize RSV and SRBSDV, respectively. After incubating with IgG-HRP antibody, the protein membranes were imaged using ECL substrate by the BIO-RAD ChemiDoc MP Imaging System.

RNA-seq analysis
The methods of RNA library construction were described in our previous reports [8,49]. Briefly, RNA sequencing used three samples each of mock and RSV-infected rice plants at 20 dpi. Total RNA was extracted using TRIzol reagent (Invitrogen, Carlsbad, CA, USA, Cat. no.

PLOS PATHOGENS
NF-YAs repress JA-mediated antiviral immunity in rice 15596-026), the RNA quantity and purity were determined by Hangzhou Lianchuan (Hangzhou, China) and the Illumina HiSeq 2000 platform was used for RNA sequencing. The rice genome (MSU Rice Genome Annotation Project database version 7.0) was used for mapping of sequencing reads by Bowtie software. Blast2go software was used for the GO functional classes and KEGG pathways. Significant differential expression of genes was assessed using the absolute value of log 2 (fold change) ratio �1 and p � 0.05 in Tukey's multiple comparison test.

Root growth inhibition assay
Seeds (� 20 seeds per line) were germinated at 37˚C and then grown in rice nutrient solution plant hydroponic boxes containing different concentrations of MeJA (2 and 5 μM) with a regime of 8 h light at 25˚C/16 h dark at 30˚C). Five days later, the root lengths were measured and recorded, and the relative root length was used to evaluate the sensitivity to JA. Each measurement was repeated three times.

Yeast Two-Hybrid Assay
Full-length coding sequences of the OsNF-YAs (OsNF-YA1, OsNF-YA2, OsNF-YA10, OsN-F-YA11) genes were amplified by PCR using the primers listed in S1 Table and inserted into the bait vector pGBKT7. The full-length coding sequences of the OsMYCs (OsMYC2/3/4) and their different truncated variants were described in our previous reports [32]. Different combinations of plasmids were co-transformed into the yeast strain AH109 according to the product specification (Takara, Japan). The transformants were cultivated on SD/-Leu/-Trp plates for about 3 days, and then positive clones were transferred to SD/-Leu/-Trp/-His/-Ade plates for interaction tests. Yeast cells were photographed after 3 days at 30˚C to record growth. All experiments were repeated three times with similar results.

Co-immunoprecipitation (Co-IP) assay
N. benthamiana leaves (four-weeks-old) were co-infiltrated by agro-infiltration. After 48 hours, the leaf tissue was collected, quick-frozen with liquid nitrogen and ground into powder. Total proteins were extracted using cold IP buffer (40 mM Tris-HCl (pH = 7.5), 100 mM NaCl, 0.5% Triton X-100, 1 mM EDTA, 1% glycerol with 1 mM DTT, 1 mM PMSF) for 20 min on ice and centrifuged twice at highest speed for 10 min at 4˚C. Then, 60μl was used as input control for western blot and the remaining supernatant was transferred to a new microcentrifuge tube, where it was incubated with 5μl FLAG-trap beads or GFP-trap beads (Sigma-Aldrich, USA) at 4˚C for approximately 2 h. The supernatant was then removed and the beads were washed at least three times in ice-cold IP buffer to remove the non-specifically adsorbed proteins. Subsequently, all protein samples were boiled for 10 minutes, briefly centrifuged and then separated on a 12% SDS-PAGE gel followed by western blotting analysis.

BiFC analysis
The coding sequences of OsNF-YA1/2, OsMYC2/3 were amplified and cloned into the nYFP and cYFP vectors, respectively. The primers used are listed in S1 Table. Equal amounts of the plasmids were co-transformed into Agrobacterium strain GV3101 and then a combination of Agrobacterium samples, each with OD 600 = 0.4, was co-injected into N. benthamiana. The YFP fluorescence signals were detected under a confocal laser scanning microscope (Nikon). Three biological repeats were done with similar results.

PLOS PATHOGENS
NF-YAs repress JA-mediated antiviral immunity in rice

Luciferase complementation imaging (LCI) assay
For the LUC complementation assays, the coding sequences of OsNF-YA1/2 and OsMYC2/3 were combined into pCAMBIA1300-cLUC and pCAMBIA1300-nLUC vectors, respectively. Equal amounts of the plasmids were co-transformed into Agrobacterium strain GV3101 and then a combination of Agrobacterium samples, each with OD 600 = 0.4, was co-infiltrated into different areas of the same N. benthamiana leaf. The plants were cultivated in the greenhouse for 48h, and then 0.1 mM luciferase substrate was infiltrated into the same positions for 10 minutes. Imaging was performed in a low-light CDD imaging apparatus (LUMAZONE SOPHIA2048B, USA) with pre-cooling. All the experiments were independently repeated at least 3 times and produced similar results.

Dual luciferase transient transcriptional activity assay
For dual-LUC determination, empty pCV-BD was used as a vector to link with the coding sequences of OsMYC2 and OsMYC3. OsNF-YA1-MYC and OsNF-YA2-MYC were used as effectors. The 5 � gal promoter was used to drive the firefly luciferase gene (LUC) as a reporter and the Renillia luciferase (REN) reporter gene controlled by Cauliflower mosaic virus promoter (35S) was the reference. The BD domain can bind to the 5 � gal promoter. All plasmids were transformed into Agrobacterium strain GV3101, and then a combination of Agrobacterium samples, each at the concentration of OD 600 = 1.0, was co-infiltrated into different areas of the same N. benthamiana leaf. The detection methods were described in our previous reports [32]. The relative luciferase activity was analyzed used LUC/REN ratios. All the experiments were repeated more than 3 times to calculate the Mean ± s.d. of relative ratio of LUC/REN. The primers used for these constructs are listed in S1 Table.

Statistical analysis
Differences were analyzed using one-way or two-way ANOVA with Fisher's least significant difference tests. A p-value � 0.05 was considered statistically significant. All analyses were performed using ORIGIN 9.1 software. OsNF-YA1/2. Data are means from 40 to 60 transfected cells. Error bars represent SD. Significant differences were identified using Fisher's least significant difference tests. � At the top of columns indicates significant difference at p � 0.05. C, D. LCI assays showing that OsN-F-YA1/2 interferes with the interaction between OsMED25 and OsMYC2/OsMYC3. Fusion proteins were transiently expressed in leaves of N. benthamiana and the measurements of luciferase activity due to LUC reconstitution for the different combinations. Quantification of the fluorescence intensity from leaves co-expressing cLUC-OsMED25 and OsMYC2-nLUC or OsMYC3-nLUC with HA-GFP or OsNF-YA1/2. Error bars represent SD of 3 biological replicates. Significant differences were identified using Fisher's least significant difference tests.

PLOS PATHOGENS
NF-YAs repress JA-mediated antiviral immunity in rice