Arabidopsis (Arabidopsis thaliana) plants were grown in a greenhouse or culture room at 22°C under the short-day condition (8/16-h light/dark cycle). The transgenic plants were screened on a half Murashige and Skoog medium containing 2% sucrose with antibiotics (20 mg/l hygromycin or 30 mg/l kanamycin) for 3 weeks after germination and then transferred to soil for further experiments and harvesting seeds. To prepare protoplasts, Arabidopsis ecotype Columbia (Col-0) plants were cultured on B5 medium containing 2% Sucrose at 22°C under the long-day condition (16/8-h light/dark cycle) for 3 weeks. wrkyy55-1 (SAIL_861_G12) and wkry55-2 (SALK_021677) were obtained from ABRC (Arabidopsis Biological Resource Center) and genotyped as mentioned in the result.

Coding regions of WKRY55, WRKY55-N, WRKY55-C, and ORA59 were amplified by PCR using gene-specific primers with attB1 and attB2 recombination sequences as listed in Supplementary Table 1. The amplified DNAs were cloned into an entry vector pDONR201 or pDONR221 by BP Clonase Enzyme Mix (Invitrogen, Carlsbad, CA, USA). The expression clones for WRKY55, WRKY55-N, WRKY55-C, and ORA59 were obtained by LR reaction with corresponding destination vectors for their experimental purposes. The same gateway cloning technique was employed to clone promoter regions of LOX2, WRKY33, and ORA59, respectively. The one kilobase fragment of promotor for each gene was amplified with gene-specific primers designed with attB1 and attB2 recombination sequences as listed in Supplementary Table 1 and introduced into an entry vector pDONR201 (Invitrogen). In the case of the promoter of PDF1.2, the fragment was amplified with promoter-specific primers of PDF1.2 and cloned into pENTR/d-TOPO (Invitrogen).

To generate WRKY55 overexpression (OE) transgenic Arabidopsis plants (WRKY55-OE), WRKY55 entry clones were recombined with binary vectors pB2GW7 (Invitrogen). pB2GW7-WRKY55 was introduced into Arabidopsis using the Agrobacterium-mediated floral dipping method (Clough and Bent, 1998). Expression of the transgenes was confirmed by RT-PCR using primers WRKY55-RT-F and -R, as shown in Supplementary Table 1.

Three different pathogens, including Pcc, Xanthomonas citri subsp. citri (Xcc), and Pseudomonas syringae pv. tomato strain DC3000 (Pst DC3000) were inoculated to Arabidopsis wild-type (Col-0), WRKY55-OE transgenic lines, or wrky55 mutant plants as described in Park et al. (2012). For the bacterial growth assay of each pathogen (Pcc, Xcc, and Pst DC3000), bacterial stock cells were cultured in Luria-Berani (LB; Pcc and Xcc) or King’s B (Pst DC3000) broth media overnight in a shaking incubator at 28°C. The overnight cultures were transferred into fresh LB broth media for 3 h and then resuspended in 10 mM MgCl₂ solution to adjust the bacterial concentration of 10⁵ cfu/ml for Pcc or of 10⁶ cfu/ml for Xcc and Pst DC3000. Bacterial suspensions were infiltrated into the leaves of 6-week-old plants grown under short-day conditions. The bacterial colonies from three replicate samples with four leaf discs of an independent plant (n = 12) were counted and analyzed statistically. The visible disease symptoms triggered by Pcc bacterial pathogen were observed on the leaves of 6-week-old plants inoculated by dropping bacterial suspension cells (10⁷ cfu/ml) on the wound site of leaves using the toothpick inoculation method. The water-soaked lesions on the leaves were scored by the disease severity as index 1 (under 0.1 cm diameter), 2 (0.1-0.5 cm diameter), and 3 (0.5 cm diameter or greater) 2 days post-inoculation based on the previous study with slight modification.

To detect H₂O₂, Pcc-treated or mock-treated leaves of six-week-old plants were prepared and subject to DAB (3,3-diaminobenzidine; Sigma, St. Louis, MO, USA) staining as described previously (Torres et al., 2002). Briefly, Pcc- or mock-treated plant leaves were collected 24 h post-treatment, followed by vacuum-infiltration with 1 mg/ml DAB solution containing 0.05% Tween-20. The samples were incubated under high humid condition approximately 6 h until brown color developed and fixed with a solution of 3:1:1 ethanol/lactic acid/glycerol incubated in a water bath at 95°C for 15 min. Samples were bleached until chlorophyll was completely depleted followed by visualization under white light.

To test the localization of WRKY55, WRKY55 entry was recombined with pEarleygate104 (for 35s- YFP::WRKY55) (Earley et al., 2006). For bimolecular fluorescent complementation (BiFC), the entry clone containing WRKY55 and ORA59 were recombined with pE-SPYCE-GW and pE-SPYNE-GW kindly provided by W. Dröge-Laser, University of Göttingen (Walter et al., 2004) for tagging at the C-terminal or N-terminal half of YFP, respectively. For transient expression analysis, plasmid DNAs were purified using NucleoBond Xtra Midi EF Kit as manufacturer’s protocol (Macherey Nagel, Duren, Germany) and transfected into protoplasts derived from 3-week-old Arabidopsis leaf tissues using polyethylene glycol-mediated transformation as previously described (Bae et al., 2006). The protoplasts were transfected with plasmid DNAs to observe the fluorescent signal at 24 h after transfection. Images were captured using an Olympus FV300 laser scanning confocal microscope, followed by rendering images with green for YFP and red for RFP, respectively (Olympus, Tokyo, Japan).

Leaves of 6-week-old plants were sprayed with 50 μM methyl jasmonate (Sigma). At 6,12-, and 24-h post-treatment, total RNA from harvested tissues samples was isolated according to the manufacturer’s instruction (PureLink kit, Invitrogen). Two micrograms of total RNAs were reversely transcribed using M-MLV reverse transcriptase (Promega, Madison, WI, USA) and subjected to quantitative RT-PCR (qRT-PCR) and semi-quantitative RT-PCR with the gene-specific primers shown in Supplementary Table 1. qRT-PCR was carried out using SYBR green PCR master mix (Bio-Rad, Hercules, CA, USA) with the MyiQ RT-PCR system (Bio-Rad). Threshold cycles (Ct) values from each gene were subtracted with Ct values of Actin for normalization by 2^−ΔΔCt method. The experiment was performed three times. Semi-quantitative RT-PCR was performed as same as qRT-PCR except using ExTaq-polymerase (Takara, Seoul, Korea) and visualized by ethidium bromide staining.

Each construct of pPDF1.2-LUC, pLOX2-LUC, pWKRY33-LUC, or pORA59-LUC was transfected with 35-RLUC (Renilla luciferase) (Bhaumik and Gambhir, 2002) to protoplasts from Col-0 and WRKY55-OE transgenic plants. In addition, pAtPDF1.2-LUC was co-transfected into protoplasts from Col-0 and WKRY55-OE transgenic plants. Luciferase activity was measured by using the dual luciferase assay system according to the manufacturer’s instructions (Promega). In brief, the protoplasts harvested 24 h after transfection were analyzed to quantify the relative LUC activity. R-LUC was co-transfected to normalize the efficiency, and the normalized value (F-LUC/R-LUC) of Col-0 sample was set to 1. Firefly luciferase activity was normalized to Renilla luciferase activity.

For yeast trans-activation assay, entry clones containing WRKY55, WRKY55-N, and WRKY55-C were transferred to a Gal4 DNA-binding domain (BD)-based bait vector (pGBKT7; Clontech, L. Deslandes). Yeast transformation was performed in Saccharomyces cerevisiae strain AH109 using the LiAc method based on the manufacturer’s protocol (Matchmaker gold yeast two-hybrid system, Clontech, L. Deslandes). The yeast two-hybrid analysis was done using a Gal4 DNA-BD-based bait vector (pGBKT7-GW; Clontech, L. Deslandes) for WKRY55 and a Gal4 activation domain-encoding prey vector (pGADT7-GW; Clontech, L. Deslandes) for AtORA59, respectively. By employing the LiAc method, pGBKT7-WRKY55 and pGADT7-ORA59 were transformed into AH109 yeast cell line. The transformed yeast cells were grown overnight in yeast peptonedextrose liquid medium to an optical density at 600 nm of 0.1 and diluted in a 10⁴ dilution. Ten microliters of diluted transformants were spotted onto synthetic dextrose (SD) medium -Leu -Trp to select co-transformed cells and SD medium -His -Leu -Trp. Plates were incubated 4 days at 28°C. Autoactivation of some hybrid proteins was suppressed by supplementing the interaction medium with 3 mM 3-amino-1,2,4-triazole.

GraphPad Prism ver. 7 (GraphPad Software, San Diego, CA, USA) was used for statistical analyses. Statistical significance was analyzed by one-way ANOVA with the least significant difference or student’s t-test for multiple or two samples comparison, respectively.

Sequence data used in this manuscript can be found in the Arabidopsis Genome Initiative databases under the following accession numbers: WRKY55 (At2g40740), WRKY33 (At2g38470), ORA59 (At1g06160), PDF1.2 (At5g44420), LOX2 (At3g45140).

To test the role of WRKY55 in immune response against Pcc, the expression of WRKY55 transcripts under Pcc infection was monitored. We observed that the transcripts level of WRKY55 was induced upon Pcc-infection at 24, and 48 h after Pcc infection (Fig. 1A), while no increased transcripts of WRKY55 were observed upon the infection of Xcc (Supplementary Fig. 1A), suggesting that WRKY55 may be induced in response to Pcc and function during Pcc-Arabidopsis interaction.

We then generated Arabidopsis transgenic plants overexpressing WRKY55 (WRKY55-OE) under constitute 35S promoter and confirmed similar expression of WRKY55 in three independent stable transgenic lines (Supplementary Fig. 1B). Two independent WRKY55-OE plants were tested upon Pcc-infection to compare the expression level with induced WRKY55 transcripts in wild-type Col-0 plants, confirming that both plants exhibited two or three-fold higher expression than Col-0 plants (Supplementary Fig. 1C). With expression-confirmed WRKY55-OE lines, bacterial growth was monitored at day 1 and 2 post-Pcc-infection. All three WRKY55-OE plants reduced Pcc growth, proposing that WRKY55 may function as a positive regulator during Pcc-Arabidopsis interaction (Fig. 1B). To further investigate the role of WRKY55, we obtained two independent T-DNA insertion mutant lines (wrky55-1 and wrky55-2) (Supplementary Fig. 1D) and verified null mutation of WRKY55 in both knock-out lines by genotyping and qRT-PCR (Supplementary Fig. 1E and F). Both wrky55-1 and wrky55-2 mutant displayed enhanced disease susceptibility with more bacterial growth at day 2 upon Pcc-infection, consistent with the enhanced resistance phenotype observed in WRKY55-OE plants (Fig. 1C). Reactive oxygen species (ROS) burst is one of the hallmarks during pathogen infection (Torres et al., 2002). Therefore, in situ ROS production in the Pcc-infected leaves was visualized by DAB staining 24 h post-Pcc-inoculation in wild-type Col-0, WRKY55-OE, and wrky55 mutant plants. As shown in Fig. 1D, hyper-ROS production with more brown color development in WRKY55-OE was observed compared to Col-0, and no ROS accumulation was detected in wrky55 mutants. These results strongly suggested that WRKY55 can modulate disease resistance positively against Pcc-infection. To address the progress of Pcc-triggering soft rot symptoms dependent on WRKY55 expression, we performed Pcc-inoculation into leaves in same plant genotypes (wild-type Col-0, WRKY55-OE, and wrky55-1 mutants) followed by comparing the disease severity based on soft rot symptom with three categories (Level 0, 1, and 2) measuring the diameter of the macerated region at one day after inoculation. Three independent WRKY55-OE transgenic plants displayed reduced soft rot symptoms, while increased disease phenotype in wrky55-1 was observed (Fig. 1E). Lastly, other hemibiotrophic bacterial pathogens, Xcc and Pseudomonase syringae pv. tomato (Pst), were inoculated to wild-type Col-0, WRKY55-OE, and wrky55 mutant plants, and bacterial growth was monitored (Supplementary Fig. 2). Importantly, we confirmed that the expression level of WRKY55 did not change the disease resistance phenotype in response to Xcc (Supplementary Fig. 2A) and Pst (Supplementary Fig. 2B). Together, our data demonstrate that the expression of WRKY55 can be Pcc-specific and may positively regulate the plant immune responses against Pcc-infection in Arabidopsis.

WRKY55 contains a group III WRKY domain, a 60 amino acid region with the conserved WRKYGQK sequence with Cys₂HisCys zinc-finger-like motif (Eulgem et al., 2000). The WRKY domain of WRKY55 is located close to the C-terminal region (Fig. 2A). The C-terminal part of WRKY proteins is more responsible for DNA binding than N-terminal region (Maeo et al., 2001; Yamasaki et al., 2005). We first tested the role of WRKY55 as a transcription factor based on the structural prediction and observed that GAL4-BD-fused full-length WRKY55 (BD::WRKY55) had a transcription activation property (Fig. 2B). To confirm the potential site of WRKY55 as a transcriptional activator, we generated the N- or C-terminal form of WRKY55 (BD::WRKY55-N or BD::WRKY55-C) (Fig. 3A, middle and bottom) with full-length WRKY55. The BD::WRKY55-C displayed a similar activation as full-length BD::WRKY55 (Fig. 2B), indicating that the C-terminal of WRKY55 is required for the transcriptional activation of WRKY55. The nuclear localization of WRKY55 was confirmed in Arabidopsis protoplasts by expressing YFP-fused WRKY55 protein with co-expression of NLS::RFP (Fig. 2C). Together, we conclude that WRKY55 can play a role in transcriptional activation in the nucleus.

Pcc is a major causal agent of the soft-rot disease, which secrets the massive amount of pectolytic enzymes resulting in tissue maceration. The damage caused by Pcc-infection turns on the JA/ET-dependent defense signaling in plants (Brader et al., 2001; Davidsson et al., 2013). JA-mediated defense genes are activated by necrotrophic pathogens and herbivorous insects (Egusa et al., 2009; Howe and Jander, 2008). To address the positive role of WRKY55 during the resistance to Pcc-infection in Arabidopsis plants, we analyzed the expression of marker genes related to JA-mediated defenses. PDF1.2, a well-known marker gene specifically activated in the JA pathway (Po-Wen et al., 2013; Zarei et al., 2011), was highly upregulated in two independent WRKY55-OE plants 1-day post-Pcc-infection (Fig. 3A, left). The transcriptional activation of PDF1.2 was further confirmed with the native promoter of the 1 kb upstream part of PDF1.2 coding sequence fused with luciferase (pPDF1.2-LUC). Increased PDF1.2 promoter activity was observed in WRKY55-OE plants compared to Col-0, consistent with upregulation of PDF1.2 gene expression (Fig. 3A, right). To further address whether WRKY55 plays a role in the expression of JA-mediated marker genes, another JA-responsive gene, LIPOXYGENASE 2 (LOX2) expression in WRKY55-OE plants as tested with PDF1.2. LOX2 was highly upregulated, and promoter activity of pLOX2-LUC was increased in WRKY55-OE plants (Fig. 3B), suggesting the increase of WRKY55 expression upon Pcc-infection can affect JA-dependent hormonal signaling pathway. The GCC-box motifs in PDF1.2 promoter are responsible for the recognition by ORA59, an AP2/ERF-type transcription factor, and the regulation of SA/JA crosstalk (Van der Does et al., 2013; Zarei et al., 2011), leading us to infer that overexpressed WRKY55 can induce PDF1.2 transcripts possibly via GCG box in PDF1.2 promoter. Thus, we tested the expression level of additional necrotrophic pathogen-responsive transcription factors with GCG box in the promoter, such as WRKY33 and ORA59 in WRKY55-OE plants. The expression of all tested genes was increased in WRKY55-OE plants (Fig. 3C). The promoter activity of WRKY33 was increased. In contrast, the activity of ORA59 was increased slightly in the protoplasts from WRKY55-OE transgenic plants (Supplementary Fig. 3A and B). Together, our data suggest that WRKY55 can play a role in the transcriptional activation of key JA- and necrotrophic pathogen-responsive genes during Arabidopsis-Pcc interaction possibly via WRKY55-DNA and/or WRKY55-ORA59 interaction.

We demonstrated that the overexpression and mutant plants of WRKY55 contribute to the expression of PDF1.2 gene during Pcc-infection (Fig. 3). We then hypothesized that WRKY55 might form a complex with transcription factors binding to cis-element in JA- and necrotrophic pathogen-responsive genes. We assessed whether WRKY55 could interact with ORA59, a master regulator of PDF1.2, by employing yeast two-hybrid protein-protein interaction assay with WRKY55-C, which did not show autoactivation, and ORA59 fused with a BD and activation domain (AD), respectively. As shown in Fig. 4A, the yeast expressing BD::WRKY55-N and AD::ORA59 grew in selection media (SD-LWH), representing that WRKY55 and ORA59 interact directly in vitro. To monitor the interaction of full-length WRKY55 with ORA59, we performed BiFC in Arabidopsis protoplasts. Arabidopsis protoplast cells co-expressing the C-terminal half of YFP with WRKY55 (WRKY55-cY) and ORA59 fused to the N-terminal half of YFP (ORA59-nY) exhibited strong BiFC fluorescence in the nucleus and cytoplasm (Fig. 4B, left). In comparison, cells co-expressing empty vector of either cYFP or nYFP with ORA59-nYFP or WRKY55-cYFP (Fig. 4B, middle and right) did not complement. To further investigate the function of this interaction in PDF1.2 gene expression, we monitored the transcript level of PDF1.2 with HA:ORA59 expressed in Arabidopsis protoplasts from wild-type or WKRY55-OE plants (Fig. 4C). The expression of PDF1.2 was significantly higher with co-expression of HA:ORA59 in the protoplasts from WRKY55-OE plants compared to Col-0 protoplasts. These results strongly support the hypothesis that WRKY55 may play a role in enhancing PDF1.2 expression by interaction with ORA59 to increase disease resistance to Pcc.

The F-box protein CORONATINE INSENSITIVE1 (COI1) is a crucial receptor of JA-mediated signaling in Arabidopsis (Van der Does et al., 2013). Our data suggest that WRKY55 can play a vital role in the expression of JA- and necrotrophic pathogen-responsive genes (Fig. 3). Thus, we reckoned that WRKY55 could involve in the JA-mediated defense signaling pathway via COI1. To confirm the epistasis between WRKY55 and COI1, we generated WRKY55-OE plants in coi1-1 mutant background (WRKY55-OE coi1-1). We first tested root growth inhibition in response to JA. Upon JA-treatment on Col-0, coi1-1, and WRKY55-OE coi1-1 seedlings, WRKY55-OE coi1-1 demonstrated sensitivity to JA as observed in Col-0, suggesting that overexpression of WRKY55 in coi1-1 can bypass JA-insensitive phenotype of coi1-1 (Fig. 5A). PDF1.2 expression was also increased in WRKY55 coi1-1 plants (Fig. 5B). The suppression of bacterial growth upon Pcc-infection in the identical genotypes tested for JA-response and PDF1.2 expression (Col-0, coi1-1, and WRKY55-OE coi1-1) exhibited enhanced disease resistance in WRKY55-OE coi1-1 compared to coi1-1 mutant plants (Fig. 5C). We further confirmed that the expression of WRKY55 was suppressed in another coi1 mutant allele (coi1-16) post-Pcc-infection (Fig. 5D). Our results propose that WRKY55 can work downstream of COI1 via JA-mediated signaling cascade upon Pcc-infection.