Salicylic acid-induced transcriptional reprogramming by the HAC–NPR1–TGA histone acetyltransferase complex in Arabidopsis

Abstract Plant immunity depends on massive expression of pathogenesis-related genes (PRs) whose transcription is de-repressed by pathogen-induced signals. Salicylic acid (SA) acts as a major signaling molecule in plant immunity and systemic acquired resistance triggered by bacterial or viral pathogens. SA signal results in the activation of the master immune regulator, Nonexpressor of pathogenesis-related genes 1 (NPR1), which is thought to be recruited by transcription factors such as TGAs to numerous downstream PRs. Despite its key role in SA-triggered immunity, the biochemical nature of the transcriptional coactivator function of NPR1 and the massive transcriptional reprogramming induced by it remain obscure. Here we demonstrate that the CBP/p300-family histone acetyltransferases, HACs and NPR1 are both essential to develop SA-triggered immunity and PR induction. Indeed HACs and NPR1 form a coactivator complex and are recruited to PR chromatin through TGAs upon SA signal, and finally the HAC−NPR1−TGA complex activates PR transcription by histone acetylation-mediated epigenetic reprogramming. Thus, our study reveals a molecular mechanism of NPR1-mediated transcriptional reprogramming and a key epigenetic aspect of the central immune system in plants.

CBP/p300-family histone acetyltransferases (HATs) are well-known transcriptional coactivators that control a variety of differentiation and developmental processes. They facilitate transcription by diverse functions: relaxing chromatin structure through histone acetylation (21)(22)(23), modulating the activity of transcriptional regulators through acetylation (24)(25)(26), acting as adaptor for numerous transcription factors (27,28) and bridging transcription factors to transcription machineries (29,30). Arabidopsis has five CBP/p300-family genes: HAC1, HAC2, HAC4, HAC5 and HAC12 (31,32). Although multiple morphological and developmental defects were observed in Arabidopsis mutants lacking multiple HACs, so far only a few studies have reported in-depth physiological analyses on these mutants and revealed the molecular functions of HACs in flowering and ethylene signaling (31,33).
Despite the essential role of NPR1 in SA-triggered transcription of pathogenesis-related genes (PRs) during plant defense, the molecular mechanism of its transcriptional coactivator role remains elusive. In this study, we show that the CBP/p300-family HATs, HAC1 and HAC5 (HAC1/5), are essential to develop SA-triggered immunity and PR induction. HAC1/5 and NPR1 form a coactivator complex and are recruited to PR chromatin through NPR1-TGA interaction, finally relaxing repressive local chromatin and facilitating transcription. In sum, our study demonstrates a mechanism of NPR1-mediated transcriptional activation and proposes epigenetic reprogramming as central part of plant immune system.

Pathogen infection
Pathogen inoculation was performed as described (24). Three days (d) after inoculation, three inoculated leaf discs each from different plants were combined and homogenized in sterile H 2 O, with at least three times of replication. Leaf extracts were plated on King's B medium and incubated at 28 • C for 2 d, and then bacterial growth was determined by counting the colony-forming units.

Co-IP assay
Co-immunoprecipitation (co-IP) assay was performed as previously described (12) with minor modifications. Briefly, total proteins were extracted from 4-w-old plants by grinding in liquid N 2 and homogenizing in extraction buffer (50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 10 mM MgCl 2 , 5 mM ethylenediaminetetraacetic acid (EDTA), 10% glycerol, 60 M MG132, 100 mM ␤-glycerophosphate, 20 mM sodium fluoride, protease inhibitors and phenylmethylsulfonyl fluoride (PMSF)). After preclearing with protein-A agarose beads, proteins were incubated with ␣-HA agarose beads (Sigma-Aldrich A2095) or ␣-GFP (Roche 11814460001) coupled to protein-A agarose (Santa Cruz sc-2001) at 4 • C for 3 h. For protein elution, the beads were boiled in 2× SDS sample buffer, and the supernatant obtained after centrifugation was saved and used for protein detection.  Table S3). NPR1 and HAC1 cDNA fragments were cloned into pGADT7 (Clontech 630442) and pGBKT7 (Clontech 630443) vectors, respectively, and introduced into yeast strain AH109 by lithium acetate method as described in the Clontech yeast protocol handbook. Interactions were assessed by yeast growth on synthetic drop-out medium lacking leucine, tryptophan, adenine and histidine in the presence of 1 or 3 mM 3-AT. Protein extraction from yeast was carried out as described previously (43). Briefly, cells were suspended in lysis buffer (50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 0.01% NP-40, 1 mM EDTA, 100 mM PMSF, 1 mM benzamidine, 1 g/ml leupeptin and 1 g/ml pepstatin) followed by bead-beating. Cell extracts were centrifuged at 1600 g for 10 min at 4 • C and the supernatant was subjected to SDS-PAGE. BDfusion proteins were detected by using anti-Myc antibody (Merck 05-724) at 1:1500 dilution, whereas AD-fusion proteins were detected by using anti-HA antibody (Roche 11867423001) at 1:1500 dilution.

Gel filtration assay
Proteins were prepared by homogenizing 4-w-old plant tissues in extraction buffer (20 mM Tris-HCl, pH 7.5, 200 mM NaCl, 10% glycerol, 60 M MG132, 100 mM ␤glycerophosphate, 20 mM sodium fluoride, protease inhibitors and PMSF) followed by 20 min of incubation at 4 • C. After centrifugation at 13 000 rpm for 5 min, the supernatant was saved and filtered through a 0.45 m filter (Millipore SLHP033RS). About 1.5 mg of total proteins were injected on the Superdex 200 10/300GL column (GE Healthcare Life Sciences GE17-5175-01) and fractionated by the AKTA fast protein liquid chromatography system (Amersham Biosciences). Proteins in each fraction were concentrated using acetone, separated by SDS-PAGE, and transferred onto nitrocellulose membranes (Millipore HATF00010) for immunoblot analysis.

RNA extraction and RT-qPCR analysis
Total RNA extraction and reverse transcription were performed as described previously (35). The sequences of primers used for reverse transcription followed by quantitative real-time PCR (RT-qPCR) are provided in Supplementary Table S4.

RNA sequencing analysis
Total RNA was isolated from leaves of 4-w-old short daygrown plants treated with DW or INA for 12 h using Tri Reagent (MRC TR118) and further purified with RNeasy MiniKit (QIAGEN 74106) to obtain OD 260/280 ratio of 1.8 to 2.2. RNAs obtained from three biologically independent experiments were combined and used for RNA-seq preparation. RNA-seq library was constructed and sequenced on the Illumina HiSeq ™ 2000 at Beijing Genomics Institute (Hong Kong). Reads were aligned to the Arabidopsis reference genome using SOAPaligner/soap2 allowing mismatches of no more than 2 bases. Gene-expression level was calculated by using RPKM (reads per kb per million reads) method. Differentially expressed genes (DEGs) were selected with False discovery rate (FDR) ≤ 0.01 and |log 2 Ratio| ≥ 1 as thresholds.

ChIP sequencing analysis
ChIP was performed as previously described (31,35) with minor modifications. Protein-DNA immune-complex was precipitated using agarose A beads (Santa Cruz sc-2001) instead of salmon sperm DNA/Protein A agarose beads to avoid the contamination of ChIPed DNA with salmon sperm DNA. About 12-20 ng of DNA pooled from six independent ChIPs was used for library construction after quality check with 2100 Bioanalyzer (Agilent). Library construction and sequencing on Illumina HiSeq ™ 2000 were performed at Beijing Genomics Institute (Hong Kong). Reads were aligned to the TAIR10 Arabidopsis genome by using SOAP2 aligner and BWA, and uniquely mapped reads were used for further analysis. Using MACS2 version 2.1.0, normalized signals respective to Col input were obtained, and H3Ac-enriched peaks were identified (P < 1.00e-02). The wiggle files obtained from peak scanning were visualized and analyzed by using Integrative Genomics Viewer (IGV). Differential peaks between genotypes and/or treatments were identified by using MACS2 bdgdiff (45) (log 10 likelihood ratio = 1) and annotated by PAVIS (46) (https: //manticore.niehs.nih.gov/pavis2/). H3Ac-distribution analysis was performed by using computeMatrix and plotprofile installed in the public server at the Galaxy (https:// usegalaxy.org/) (47).

Sequential ChIP assay
Sequential ChIP was performed as previously described (48) with minor modifications. Chromatin was isolated from cross-linked samples by using 450 ml of nuclei lysis buffer (50 mM Tris-HCl, pH 8.0, 10 mM EDTA, 1% SDS, 0.1 mM PMSF and protease inhibitors), fragmented by sonication and subjected to immunoprecipitation with anti-HA antibody (Abcam ab9110). Immune complexes were eluted by gentle agitation in 100 l of elution buffer (16.7 mM Tris-HCl, pH 8.0, 1.2 mM EDTA, 20 mM DTT and 1% SDS) at 37 • C for 30 min. Eluted chromatin was diluted with 20-fold of ChIP dilution buffer (16.7 mM Tris-HCl, pH 8.0, 1.2 mM EDTA, 167 mM NaCl and 1.1% Triton X-100), subjected to the second immunoprecipitation with anti-GFP (Roche 11814460001) or control anti-FLAG (Sigma-Aldrich A8592) antibody and then eluted with elution buffer (1% SDS and 0.1 M NaHCO 3 ). DNA was isolated by reverse-crosslinking and proteinase K treatment and purified using QIAquick PCR Purification Kit (QIA-GEN 28106). Quantification of immunoprecipitated DNA Nucleic Acids Research, 2018, Vol. 46, No. 22 11715 and the evaluation of the relative amount of amplified products in samples were performed as described in the ChIP assay section.

CBP/p300-family histone acetyltransferases (HACs) activate SA-dependent plant immunity by promoting PR transcription through histone acetylation
We and others have found that histone H3 acetylation (H3Ac) at the Arabidopsis PR1 locus is increased by pathogen attack or SA treatment, and this increase is tightly associated with PR1 transcription (35,(49)(50)(51). Interestingly, the H3Ac increase at PR1 is undermined by the loss of either NPR1 or the three related Class II TGAs (TGA2, TGA5 and TGA6; TGA2/5/6) (50) ( Figure 1A). These inspired us to identify HATs responsible for the SA-induced H3Ac (Supplementary Figure S1A). As H3Ac acts as an active epigenetic mark, first we searched for Arabidopsis HAT mutants with impaired PR1 and PR2 induction upon 2,6dichloroisonicotinic acid (INA; synthetic SA analog) treatment. The mutants lacking HAG1 (hag1-6) (52) Figure S1B). Further, the INAinduced H3Ac increase at PR1 was barely detectable in hac1/5 ( Figure 1C) as in npr1 and tga2/5/6, whereas in hag1-6 it was comparable to wild-type (WT) in the promoter regions but reduced in the gene body ( Supplementary Figure S1C), suggesting that HACs are likely to be the responsible HATs.
Consistent with the above results, upon infection of Pseudomonas syringae pv. tomato DC3000 (Pst DC3000), the PR1/2 induction and H3Ac increase were all severely impaired in hac1/5 and to lesser extents in hac1 and hac1/12 ( Figure 1D and E; Supplementary Figure S1D and E). Basal resistance to Pst DC3000 was also substantially decreased by the hac1/5 mutations ( Figure 1F). Moreover, a HAC1:HA fusion protein was targeted to the PR1/2 promoters in both INA-and pathogen-dependent manners ( Figure  S2D), consistent with its reported dual roles as repressor and activator depending on SA signal (7,10,53,54). Interestingly, INA treatment further increased the TGA2:FLAG targeting to these regions and induced a significant targeting to the P3 region of the PR1 promoter also, resulting in the targeting pattern of TGA2 similar to those of NPR1 and HAC1. This result suggests that INA might induce changes in the biochemical property of TGAs that affect the binding affinity of TGAs to the PR1/2 promoters or to the antibody used. Sequential ChIP assays using HAC1:HA-and NPR1:GFP-containing transgenic plants showed the presence of PR1 promoter-bound NPR1:GFP within the HAC1:HA immunoprecipitate ( Figure 2D), indicating the co-localization of HAC1 and NPR1 on PR1 upon INA treatment. These findings, together with the well-known NPR1-TGA interaction and the lack of INAinduced H3Ac increase in npr1 and tga2/5/6 mutants (Figure 1A), led us to hypothesize that HACs, NPR1 and TGAs might form a complex on PR promoters and modulate transcription through chromatin modification. In support of this view, hac1/5, npr1 and npr1 hac1/5 mutants (Supplementary Figure S6) showed comparable INA-induced PR1 transcript levels and susceptibilities to Pst DC3000 ( Figure  2E and F).
To study whether HAC1, NPR1 and TGAs interact with each other, and, if they do, how the SA signal affects their interactions, we examined the subcellular localization of each protein and their interactions before and after INA treatment using stable transgenic Arabidopsis plants. HAC1 and TGA2/5 were always localized within nucleus, whereas the abundance and localization of NPR1 were affected by INA (Supplementary Figure S7) as previously reported (12,13,16). HAC1 and TGA2/5 were detected in the NPR1:GFP immunoprecipitate, and reciprocally NPR1 and TGA2/5 were also detected in the HAC1:HA immunoprecipitate ( Figure 3A and B; Supplementary Figure S8), revealing the existence of a complex containing HAC1, NPR1 and TGA2/5. TGA2/5, but not HAC1 enrichment within NPR1:GFP immunoprecipitate, was increased by INA ( Figure 3A and Supplementary Figure S8A), suggesting that HAC1 might be limiting in complex formation. In contrast, both NPR1 and TGA2/5 enrichments within the HAC1:HA immunoprecipitate were increased by INA ( Figure 3B and Supplementary Figure S8B), and similar increases were also observed after pathogen attack (Supplementary Figure S9), implying the possibility of one HAC1 molecule engaging multiple NPR1 and TGA2/5 molecules as SA-bound nuclear NPR1 level increases in response to SA or pathogen signal. This model might be a reminiscent of the interaction of p300 and MEF2 on DNA in which the highly conserved TAZ domain of p300 binds to three MEF2:DNA complexes (55). We could observe interactions between the two TAZ domains of HAC1 and the C-terminal region of NPR1 in yeast ( Supplementary Figure S10), suggesting that, similar to p300, the HAC1 TAZ domains might be important for the assembly of the HAC-NPR1-TGA complex.
ChIP assays were then used to study the binding hierarchy of HAC, NPR1 and TGA to PR1 chromatin. INAinduced targeting of NPR1 and HAC1 to PR1 chromatin was completely abolished in tga2/5/6 triple mutants (Figure 4A and B), and notably, INA-induced HAC1 targeting to PR1 was undetectable in npr1 mutants ( Figure 4C). These results point to that HAC1 and NPR1 are recruited to PR1 chromatin via the interaction between NPR1 and the DNA-binding protein TGA as expected from the co-IP results ( Figure 3A-F). Strikingly, in contrast to the HAC1/5-independent NPR1-TGA2/5 interaction ( Figure  3E), NPR1 but not TGA2 targeting was reduced largely in PR1 and to a lesser extent in PR2 chromatin in the absence of HAC1/5 ( Figure 4D and Supplementary Figure S11A-C). Therefore, although HAC1/5 may not be required for the interaction between NPR1 and free TGAs, they are likely required for efficient NPR1 binding to TGAs in the chromatin context. One possibility is that HAC recruited via NPR1 to PR chromatin might modify local chromatin landscape by acetylating histones, which in turn might allow more stable association of the HAC-NPR1-TGA complex with chromatin. Alternatively, HAC might act as an adaptor forming multivalent interactions with transcription factors and thus stabilizing NPR1 association with PR chromatin. It is also possible that SA-binding to NPR1 might induce a conformational change to the HAC-NPR1 complex or to the ternary HAC-NPR1-TGA complex rendering more efficient interaction with DNA-bound TGAs or PR promoters. In sum, one role of HAC might be to facilitate or/and stabilize the establishment of the functional HAC-NPR1-TGA complex on PR chromatin.

HACs are essential components of the SA-induced NPR1and TGA-containing high molecular-weight complex
To gain further insight into the HAC-NPR1-TGA complex in vivo, we performed gel-filtration chromatography assays. Without INA treatment, HAC1:FLAG, NPR1:GFP and TGA2/5 were predominantly identified in fractions with molecular weights greater than their respective predicted monomeric sizes ( Figure 5A-D and Supplementary Figure S12), suggesting their presence within complexes in vivo (12,13,54). Noticeably, INA treatments broadened and shifted the elution profiles of HAC1:FLAG toward both larger and smaller mass ranges ( Figure 5A). NPR1:GFP expressed by the native promoter of NPR1 was greatly increased in abundance by INA treatment in all fractions where NPR1:GFP was detected ( Figure 5B). When NPR1:GFP was overexpressed, its elution profile was clearly shifted toward larger mass ranges by INA treatment, although its abundance was not greatly increased (Supplementary Figure S12A). INA treatment also substantially affected the elution profile of TGA2/5 to form another peak at much higher mass range (∼fraction #19 in Figure  5C  For deeper understanding of the role of HACs in the assembly of the HAC-NPR1-TGA complex, we then compared the elution profiles of NPR1:GFP and TGA2/5 in WT versus hac1/5 mutants ( Figure 5B and C). Without INA, the elution profiles of NPR1:GFP and TGA2/5 were similar between WT and hac1/5. However, after INA treatment, the abundance of NPR1:GFP in higher molecularweight fractions was evidently decreased by hac1/5 mutations ( Figure 5B and Supplementary Figure S12A). Furthermore, TGA2/5 abundance in fractions >669 KD was also drastically reduced or eliminated in hac1/5 mutants and instead TGA2/5 were detected in smaller-weight fractions ( Figure 5C and Supplementary Figure S12B). Thus, HACs are essential components of the INA-induced high molecular-weight complex containing NPR1 and TGAs. Similarly, after INA treatment, TGA2/5 were not detected in the >669 KD fractions in npr1 mutants ( Figure 5D), consistent with the co-IP results showing NPR1-dependent HAC1-TGA2/5 interaction ( Figure 3F).

Several hundred genes are commonly regulated by NPR1 and HAC1/5
To assess the role of the collaboration between NPR1 and HACs in SA-induced transcriptional reprogramming at genome-wide level, we performed RNA-seq analyses of transcriptomes of WT, npr1 and hac1/5 either treated with INA or not ( Figure 6A). About 71% and 18% of the genes significantly upregulated by INA in WT were not upregulated in npr1 and hac1/5, respectively. Among the NPR1dependent genes (2742), 21% (584) also showed HAC1/5dependency (Group 1; Supplementary Datasets S1 and S2), whereas the remaining 79% (2158) did not (Group 2; Supplementary Datasets S1 and S2). The RNA-seq results were confirmed by RT-qPCR analyses of dozens of randomly selected genes from each group (Supplementary Figures S13 and S14). Thus, a small but considerable fraction (15%) of the INA-induced transcriptome in WT is dependent on both NPR1 and HAC1/5 (Group 1), whereas a larger fraction (56%) requires only NPR1 (Group 2).

HACs also regulate SA biosynthesis or accumulation-related genes in an NPR1-independent manner
Despite relatively small portion of the Group 1 genes, the susceptibilities of hac1/5 and npr1 against Pst DC3000 were comparable ( Figure 2F). This suggests that the genes coregulated by HAC1/5 and NPR1 might be crucial in plant immunity. Alternatively, HACs might affect plant immunity in an NPR1-independent as well as NPR1-dependent manners. Lately, it was reported that the ethylene-signaling pathway is hyper-activated in hac1/5 (33). As ethylene (Et) and jasmonic acid (JA) antagonistically crosstalk with SA in general (56)(57)(58)(59), the activation of Et/JA-signaling could suppress the SA-dependent defense pathway (60). Thus, first we compared the expressions of ERF1 and ERF2 (56,61), genes encoding ethylene-response factors and CHIB and VSP2 (62), the JA-and Et-signaling pathway markers, respectively, in WT, npr1 and hac1/5 (Supplementary Figure S15A). Although, the expression levels of ERF1 and ERF2 in hac1/5 were higher than in WT and npr1 after and before Pst DC3000 infection, respectively, pathogen-induced expressions of CHIB and VSP2 were significantly reduced in hac1/5 but not in npr1, indicating that the susceptibility of hac1/5 was not likely caused by the activated Et/JA pathways.
We then examined the effect of hac1/5 mutations on the pathogen-induced expression of several SA biosynthesis or accumulation-related genes, namely ICS1 (63), EDS5 (64), PAD4 (65) and GDG1 (66,67) as the induction of ICS1 and EDS5 by pathogen infection was known to be NPR1-independent (68)(69)(70). Induction of ICS1, EDS5 and PAD4 by Pst DC3000 were significantly reduced by hac1/5, but not by npr1 mutations (Supplementary Figure S15B). Further, pathogen-induced targeting of HAC1:HA but not NPR1:GFP was observed in the examined regions of ICS1 and EDS5 promoters (Supplementary Figure S16). These results are consistent with the previously reported NPR1independent pathogen-induced expression of these genes and indicate that HACs also promote SA-dependent immunity by NPR1-independently regulating SA biosynthesis or accumulation-related genes, explaining part of the severe pathogen-susceptible phenotype of hac1/5.

HACs are epigenetic partners of NPR1 and the HAC-NPR1-TGA complex constitutes part of the genome-wide SA-induced transcriptional reprogramming system
Next, by ChIP seq we studied how H3Ac levels are affected by INA, npr1 and hac1/5 mutations at the Group 1-and Group 2-gene loci ( Figure 6B). Reproducibility of the ChIPseq data was confirmed by ChIP-qPCR analyses of 11 selected loci (Supplementary Figure S17). At the Group 1gene loci, H3Ac levels at the downstream of the transcription start sites were substantially increased by INA treatment in WT. However, this INA-induced increase was not observed both in npr1 and hac1/5 mutants. The Group 2gene loci showed ∼1.5-fold higher basal H3Ac levels and clear but less substantial INA-induced H3Ac increase compared to the Group 1-gene loci in WT. In hac1/5 mutants, the INA-induced H3Ac increase was still obvious at the At 56% of the further selected loci, H3Ac levels were substantially increased by INA, and 79% of these loci showed compromised H3Ac increases in both npr1 and hac1/5 mutants. Considering diverse mechanisms other than histone acetylation for the transcriptional coactivator role of CBP/p300 HATs as mentioned in the Introduction, we believe it is plausible that at least part of the Group 1 genes are expected to show no good correlations between H3Ac and RNA expression, even though they are the direct targets of HACs and NPR1. Taken all together, our ChIP-seq and RNA-seq analyses indicate that the HAC-NPR1-TGA complex constitutes part of the genome-wide transcriptional activator system responsible for the SA-induced transcriptional reprogramming.

DISCUSSION
Although NPR1 is a well-known master regulator of the SA-dependent immunity and systemic acquired resistance, how it acts as a transcriptional coactivator for over 2000 downstream genes is not fully understood at the molecular level. Our study demonstrates a molecular mechanism for the coactivator role of NPR1 in which NPR1 acts in concert with HACs as epigenetic partners and that the HAC-NPR1-TGA complex is involved in genome-wide transcriptional reprogramming through histone acetylation-based mechanism (Figure 7). Interestingly, the SA-induced transcriptional reprogramming model we propose here is a reminiscent of the steroid hormone-induced epigenetic and transcriptional reprogramming model prevalent in animal systems. It is also possible that HAC in the ternary complex might also acetylate transcriptional regulators including NPR1/TGAs and affect their transcription activities or might act as a scaffold leading to the formation of a large transcription-activator complex required for PR expression. Further, our work indicates that epigenetic reprogramming is a central feature of the immune system in plants which, unlike animals, lack specialized immune cells.
Our finding of both HAC-dependent (Group 1) and independent (Group 2) NPR1-regulated genes suggests that NPR1 might act in different modes depending on target chromatin contexts. For example, the degree of chromatin compaction could be a factor in the HAC requirement. Our finding of higher basal H3Ac levels at the Group 2-gene loci compared to the Group 1-gene loci ( Figure 6B) supports this hypothesis. The Group 2 genes might be in open chromatin conformation with enriched basal H3Ac and poised to respond to SA-activated NPR1 without assistance from HATs. In this case, the INA-induced increased H3Ac levels observed in WT and hac1/5 might be consequences rather than causes of the increased transcriptional activities in those plants. Alternatively, HATs other than HACs might act as epigenetic partners of NPR1 for the Group 2 genes. The Group 1 and Group 2 genes do not seem to act in different biological processes as our preliminary gene ontology analysis did not reveal significant differences between them. Thus, it would be of interest in the future to understand the chromatin features of the Group1 and Group 2 genes or the dependency of the Group 2 transcription on chromatin factors other than HACs. Comprehensive evalu- Figure 7. Model for the epigenetic reprogramming of PR genes by the HAC-NPR1-TGA complex. Under normal condition (left), NPR1 (blue oval) preferentially presents within the cytoplasm as oligomers while its minor fraction is within the nucleus and interacts with TGA (pink oval) and HAC (yellow oval). Although TGA, which is not in the ternary HAC-NPR1-TGA complex, binds to PR promoters and represses PR transcription, the HAC-NPR1-TGA complex is not recruited to PR chromatin under this condition. Upon pathogen challenge and following SA surge (right), the nuclear fraction of NPR1 is increased due to enhanced stability and translocation of part of the cytoplasmic NPR1 into the nucleus. With increased concentration and SA binding, NPR1 interacts with HAC possibly in a multiple:one fashion. The SA-bound HAC-NPR1-TGA complex is now recruited to PR promoters or alternatively the SA-bound HAC-NPR1 complex is recruited by TGA on PR promoters, and the ternary complex induces transcriptional activation through histone acetylation (Ac)-dependent chromatin reprogramming. SA might induce a conformation change to the HAC-NPR1-TGA complex and activate it during this process.
ations on the genome-wide role of the HAC-NPR1-TGA complex in SA-induced transcriptional reprogramming will be possible through integrative analyses of data from RNA seq, H3Ac ChIP seq and the genome-wide association studies of HACs, NPR1 and TGAs.

DATA AVAILABILITY
The ChIP-seq and RNA-seq data have been deposited in the Gene Expression Omnibus (GEO) under the SuperSeries accession number GSE101572. All other data are available from the authors upon reasonable request.