Abstract
Recognition of microbe-associated molecular patterns (MAMPs) derived from invading pathogens by plant pattern recognition receptors (PRRs) initiates a subset of defense responses known as pattern-triggered immunity (PTI). Transcription factors (TFs) orchestrate the onset of PTI through complex signaling networks. Here, we characterized the function of ERF19, a member of the Arabidopsis thaliana ethylene response factor (ERF) family. ERF19 was found to act as a negative regulator of PTI against Botrytis cinerea and Pseudomonas syringae. Notably, overexpression of ERF19 increased plant susceptibility to these pathogens and repressed MAMP-induced PTI outputs. In contrast, expression of the chimeric dominant repressor ERF19–SRDX boosted PTI activation, conferred increased resistance to the fungus B. cinerea, and enhanced elf18-triggered immunity against bacteria. Consistent with a negative role for ERF19 in PTI, MAMP-mediated growth inhibition was weakened or augmented in lines overexpressing ERF19 or expressing ERF19–SRDX, respectively. Using biochemical and genetic approaches, we show that the transcriptional co-repressor Novel INteractor of JAZ (NINJA) associates with and represses the function of ERF19. Our work reveals ERF19 as a novel player in the mitigation of PTI, and highlights a potential role for NINJA in fine-tuning ERF19-mediated regulation of Arabidopsis innate immunity.
Introduction
Plants have adopted sophisticated defense mechanisms to fight off invading pathogens. Initiation of plant defense responses relies on the recognition of non-self organisms. Plants utilize pattern recognition receptors (PRRs) as the first line of surveillance to detect incoming threats posed by pathogens. Plant PRRs perceive microbe-associated molecular patterns (MAMPs), which are molecular structures conserved among microbes and crucial for the survival of microbes (Macho and Zipfel, 2014; Zipfel, 2014). For example, flg22, the active epitope of the bacterial MAMP flagellin, is recognized by the PRR FLAGELLIN SENSING2 (FLS2) (Felix et al., 1999; Gómez-Gómez and Boller, 2000), and the EF-Tu RECEPTOR (EFR) recognizes the conserved peptide elf18 derived from bacterial EF-Tu, which is one of the most abundant proteins in bacteria (Kunze et al., 2004; Zipfel et al., 2006). The fungal MAMP chitin, an important constituent of fungal cell walls (Silipo et al., 2010), is perceived by CHITIN ELICITOR RECEPTOR KINASE1 (CERK1) and LYSM-CONTAINING RECEPTOR-LIKE KINASE 5 (LYK5) (Miya et al., 2007; Wan et al., 2008; Cao et al., 2014). MAMP recognition induces pattern-triggered immunity (PTI), restricting the incursion and proliferation of potential pathogens (Boller and Felix, 2009; Schwessinger and Ronald, 2012; Newman et al., 2013).
Activation of PTI involves massive transcriptional reprogramming to mount defense responses against invading pathogens (Bigeard et al., 2015; Tsuda and Somssich, 2015; Garner et al., 2016; Birkenbihl et al., 2017a). General PTI responses include reinforcement of the cell wall through deposition of callose and production of defense-related proteins (Boller and Felix, 2009). Pathogenesis-related (PR) proteins and plant defensins (PDFs) represent two major classes of defense-related proteins with diverse antimicrobial activities (Thomma et al., 2002; van Loon et al., 2006). In Arabidopsis, PR1 and PR2 are induced after inoculation with the hemi-biotrophic bacterium Pseudomonas syringae pv. tomato (Pst) DC3000 and are marker genes for flg22 and elf18 treatments (Lu et al., 2009; Choi et al., 2012; Nomura et al., 2012), whereas PDF1.2 and PDF1.3, which are induced by the necrotrophic fungus Botrytis cinerea, serve as potential markers for chitin elicitation (Pieterse et al., 2009, 2012; Meng et al., 2013).
Activation of plant immunity requires a high expense of energy, and excessive immune responses reduce plant fitness, hampering plant growth and survival (Bolton, 2009; Katagiri and Tsuda, 2010; Kim et al., 2014). Transcription factors (TFs) lie at the heart of transcriptional reprogramming, and the ethylene response factor (ERF) TF family plays a key role in orchestrating the balance of defense outputs (Nakano et al., 2006; Huang et al., 2016; Jin et al., 2017). Perturbation of key immune regulators may tip the balance and lead to growth retardation. For example, direct activation of ERF6 enhances Arabidopsis resistance to B. cinerea and induces constitutive activation of defense genes (Meng et al., 2013). However, these plants exhibit a severe dwarf phenotype, which might be the result of strong defense activation (Meng et al., 2013).
In order to maintain appropriate levels of defense activation, TFs that negatively regulate immunity need to work in concert with defense-activating TFs. For example, the pathogen-induced ERF4 (ERF078) and ERF9 (ERF080) negatively regulate Arabidopsis resistance against fungal pathogens and activation of PDF1.2 (McGrath et al., 2005; Maruyama et al., 2013). In addition, transcriptional activities of TFs are modulated in a post-translational manner to ensure timely activation or repression of immune signaling cascades (Licausi et al., 2013). Typically, ETHYLENE INSENSITIVE 3 (EIN3) transactivates ERF1 (ERF092), but the transactivation function of EIN3 is repressed in the presence of JASMONATE ZIM-DOMAIN 1 (JAZ1) (Zhu et al., 2011). Notably, JAZ1 interacts with EIN3 and recruits the transcriptional co-repressor Novel Interactor of JAZ (NINJA) with TOPLESS (TPL) or TPL-related proteins (TPRs) (Pauwels et al., 2010; Zhu et al., 2011). EIN3-mediated activation of ERF1 is de-repressed when JAZ1 is degraded upon accumulation of jasmonic acid (JA) that occurs after pathogen attack (De Vos et al., 2005; Chini et al., 2007; Zhu et al., 2011). JAZ1-imposed repression on EIN3 ensures that ERF1 and ERF1-targeted defense genes such as PDF1.2 are not induced in the absence of pathogen invasion (Pieterse et al., 2012).
While there are increasing reports showing that ERFs are involved in plant defense, studies centered on ERFs regulating PTI remain sparse (Bethke et al., 2009; Meng et al., 2013; Xu et al., 2017). Here we report that the pathogen- and MAMP-induced ERF19 plays a negative role in Arabidopsis immunity against both fungal and bacterial pathogens. Notably, overexpression of ERF19 or repression of ERF19 function through expression of the chimeric dominant repressor ERF19–SRDX leads to decreased and increased PTI responses, respectively. Our data further suggest that ERF19 functions as a modulator in MAMP-mediated growth inhibition and may serve as a buffering mechanism to prevent detrimental effects of excessive PTI. Moreover, our biochemical and genetic approaches showed that NINJA associates with and represses the function of ERF19, suggesting another layer of control over PTI activation. Collectively, our functional studies on ERF19 provide novel evidence about an ERF involved in the regulation of PTI and new insights into the dynamic regulation of plant immunity.
Materials and methods
Biological materials and growth conditions
Growth conditions of Arabidopsis thaliana (L. Heyhn.) and Nicotiana benthamiana were described previously (Yeh et al., 2016). Arabidopsis ecotype Col-0 was used as the wild-type (WT) for the experiments unless stated otherwise. We obtained mutants npr1-1 from X. Dong (Duke University, Durham, NC, USA), ein2-1 from the Arabidopsis Biological Resource Center (https://abrc.osu.edu/), coi1-16 (Col-6 background) from J.G. Turner (University of East Anglia, Norwich, UK), and ninja-1 from E.E. Farmer (University of Lausanne, Switzerland). The Arabidopsis transgenic line 35S:GFP was obtained from K. Wu (National Taiwan University, Taipei, Taiwan). The fungus B. cinerea was obtained from C.-Y. Chen (National Taiwan University, Taipei, Taiwan) and was grown on potato dextrose broth (PDB)–agar plates in the growth chamber where Arabidopsis plants were grown (Zimmerli et al., 2001). The bacterium Pst DC3000 was provided by B.N. Kunkel (Washington University, St. Louis, MO, USA) and was grown at 28 °C, 200 rpm in King’s B medium with 50 mg l–1 rifampicin.
Preparation of chemicals
Chitin (#C9752, Sigma), and flg22 and elf18 peptides (Biomatik) were suspended in deionized water. β-Estradiol (β-Est, #E2758, Sigma) was prepared in DMSO.
Pathogen infection assays
Droplet inoculation with B. cinerea and assessment of disease symptoms were performed as previously described (Catinot et al., 2015), except that 8 µl of B. cinerea inoculum per leaf were used in this study. For spray inoculation with B. cinerea, the spore suspension (105 spores ml–1 in 1/4 PDB) was evenly sprayed on the leaves of 4-week-old plants until run-off occurred. The infected plants were kept at 100% relative humidity, and disease development was scored at 5 days post-inoculation (dpi). Dip inoculation with Pst and assessment of bacterial populations were performed as described (Yeh et al., 2016). To assess PTI-mediated resistance to Pst, assays were performed as previously described with slight modifications (Liu et al., 2015). Briefly, five leaves per plant were syringe infiltrated with deionized water or 10 nM elf18 prior to syringe infiltration of 106 cfu ml–1Pst solution. The inoculated plants were kept at 100% relative humidity overnight. Bacterial titers were determined at 2 dpi as described (Zimmerli et al., 2000).
Generation of transgenic plants
The coding sequence (CDS) of ERF19 without a stop codon was amplified from Col-0 cDNA with ERF19-F1 and ERF19-R1 primers and cloned into pCR8-TOPO vector (Invitrogen) to create pCR8-ERF19. The ERF19 CDS was subcloned into pMDC83 (Curtis and Grossniklaus, 2003) and pEarleyGate103 (Earley et al., 2006) vectors via LR reaction (Thermo Fisher Scientific) to create pMDC83-ERF19 and pEarleyGate103-ERF19 constructs, respectively. To create the inducible construct, the ERF19–green fluorescent protein (GFP) CDS was partially digested from pEarleyGate103-ERF19 with XhoI and PacI. The ERF19–GFP fragment was ligated with pMDC7 vector (Zuo et al., 2000; Curtis and Grossniklaus, 2003) digested with the same enzymes to create pMDC7-ERF19. To construct chimeric ERF19–SRDX, the genomic fragment of ERF19 including its promoter region (base pairs –1 to –1535) was amplified by PCR using ERF19-F2 and ERF19-R2 primers. The product was digested with AscI and SmaI, and then introduced into the same enzyme-treated VB0227 vector. The complete ProERF19:ERF19-SRDX:HSP part was transferred into the pBCKH(VB0047) (Mitsuda et al., 2011) binary vector by LR reaction to create pBCKH-ERF19-SRDX. Agrobacterium tumefaciens GV3101 was used to deliver the constructs into plants (Martinez-Trujillo et al., 2004). Constructs pMDC83-ERF19, pEarleyGate103-ERF19, pMDC7-ERF19, and pBCKH-ERF19-SRDX were used to generate transgenic ERF19-OE, ERF19-OE/ninja-1, ERF19-iOE, and ERF19–SRDX lines, respectively. Independent homozygous T3 lines with a single T-DNA insertion were used for the experiments. All primers used in this study are summarized in Supplementary Table S1 at JXB online.
Treatment with β-Est
Twelve-day-old seedlings and 5-week-old plants were treated with 20 µM β-Est by submergence in liquid half-strength Murashige and Skoog (1/2 MS) and syringe infiltration, respectively, 24 h before downstream experiments.
Subcellular localization
Β-Est-treated, 12-day-old ERF19-iOE1 and 35S:GFP seedlings were vacuum infiltrated with DAPI solution (5 µg ml–1) for 2 min and washed three times with distilled water. The GFP and DAPI signals in the roots were imaged with a Zeiss LSM 780 confocal microscope.
RT–PCR
To monitor MAMP- or pathogen-induced ERF19, 12-day-old seedlings were incubated in liquid 1/2 MS for one night before treatments with 200 µg ml–1 chitin, 100 nM flg22, 100 nM elf18, 5 × 105B. cinerea spores ml–1, or 107 cfu ml–1Pst. To prepare the microbial inoculants, B. cinerea spores and Pst were pelleted by centrifugation at 3000 g for 5 min and resuspended in 1/2 MS. Total RNA isolation, reverse transcription, and real-time PCR (RT–PCR) analyses were performed as described (Catinot et al., 2015). The gene UBIQUITIN 10 (UBQ10) was used for normalization. For RT–PCR, 2 µl of cDNA were used as template, and standard PCR conditions were applied as described (Huang et al., 2014). UBQ10 was used as a loading control.
Callose deposition assays
Fourteen-day-old seedlings were incubated in liquid 1/2 MS for one night before treatments with 200 µg ml–1 chitin, 100 nM flg22, 100 nM elf18, or deionized water. Twenty-four hours later, callose deposits were stained and quantified as previously described (Kohari et al., 2016; Yeh et al., 2016).
Protoplast preparation and transfection
Arabidopsis protoplasts were prepared from the leaves of 5-week-old plants as previously described (Wu et al., 2009). Polyethylene glycol-mediated protoplast transfection was performed as described (Yoo et al., 2007).
Protoplast transactivation (PTA) assays
PTA assays were performed as previously described (Hsieh et al., 2013). The reporter plasmid consists of the gene encoding firefly luciferase (fLUC) under the control of upstream activation sequence (UAS) targeted by the yeast GAL4 TF. The reference plasmid carries the gene encoding Renilla luciferase (rLUC) under the control of the 35S promoter. Effector plasmid harboring the DNA-binding domain of GAL4 expressed from the 35S promoter was used as the empty vector control (GAL4DB). The fragment of ERF19 CDS amplified by PCR using ERF19-F3 and ERF19-R3 primers was digested with XmaI and SalI, and then introduced into GAL4DB to create GAL4DB-ERF19 effector plasmid. To construct GAL4DB-ERF19-SRDX effector plasmid, the fragment ERF19–SRDX, amplified from pBCKH-ERF19-SRDX with primers ERF19-F and SRDX-R, was ligated with the vector backbone, amplified from GAL4DB with primers pGAL4-F and pGAL4-R, by blunt-end cloning. To create the NINJA, HDA6, and HDA19 effector plasmids, the CDS of NINJA, HDA6, and HDA19 were amplified from Col-0 cDNA, introduced into pCR8-TOPO vector, and subcloned into pGWHA, a plasmid modified from p2FGW7 (Karimi et al., 2002), by substituting the GFP tag with a single HA tag, by LR reaction. The effector plasmids, reporter plasmids, and reference plasmids were transfected to Arabidopsis protoplasts at the ratio of 5:4:1. After 24 h, the luciferase activities were analyzed using the Dual-Luciferase Reporter Assay System (Promega). Data are presented as the normalized fLUC activities relative to the no effector control (set as 1).
MAMP-induced growth inhibition
Growth inhibition experiments were performed as described (Ranf et al., 2011). Briefly, ten 5-day-old seedlings of the same genotype were transferred into 6-well plates supplemented with liquid 1/2 MS (0.5 g l–1 MES, 0.25% sucrose, pH 5.7). The seedlings were treated with water or MAMPs at the indicated concentration. The treated seedlings were further grown for another 10 d under normal growth conditions. Ten seedlings in a single well were blotted dry on tissue paper and weighed as a whole.
Co-immunoprecipitation (Co-IP) assay in Arabidopsis protoplasts
Full-length CDS of NINJA, HDA6, and HDA19 were cloned into pCR8-TOPO entry vector, and subcloned into pEarleyGate103 (Earley et al., 2006) by LR reaction. The GFP empty vector control was created by digesting pEarleyGate103 with XhoI to remove the Gateway cassette. The ERF19 CDS from pCR8-ERF19 was first introduced into pGWB14 vector with a C-terminal triple HA fusion (Nakagawa et al., 2007) via LR reaction, and the fragment 35S:ERF19-HA3:NOS was amplified with 35S-F and NOS-R primers by PCR. This fragment was cloned into pCR8-TOPO to create a plant expression plasmid with high copy number. Protoplast transfection and Co-IP were performed as previously described (Yeh et al., 2015).
Bimolecular fluorescence complementation (BiFC) in N. benthamiana
Using LR reaction, the CDS of ERF19, NINJA, HDA6, and HDA19 were introduced into BiFC vectors carrying split yellow fluorescent protein (YFP) fragments (Waadt et al., 2008). To create the construct for the nuclear marker, the nuclear localization signal (NLS) was fused to the N-terminus of mCherry by PCR with primers NLS-mCherry-F and mCherry-R. This fragment was cloned into pENTR/D-TOPO vector and digested with SmaI to create a blunt-end vector. The vector was then ligated with the PCR fragment mCherry–NLS, amplified with primers mCherry-F and mCherry-NLS-R, to create the complete NLS–mCherry–mCherry–NLS sequence. This sequence was introduced into pEarleyGate100 (Earley et al., 2006) by LR reaction to create the construct for the nuclear marker. The constructs were transformed into A. tumefaciens GV3101 by electroporation. Transient expression in Nicotiana benthamaian was performed as described (Roux et al., 2011), except that the Agrobacterium strains carrying the BiFC constructs were mixed 1:1 to a final OD600 of 0.4 for each strain, and the nuclear marker strain was added to a final OD600 of 0.1. Two days later, the transiently expressing leaves were imaged with a Zeiss LSM 780 confocal microscope.
Protein extraction in Arabidopsis seedlings
Extraction of total proteins from Arabidopsis seedlings was performed as previously described (Tsugama et al., 2011).
Immunoblotting
Immunoblotting was performed as previously described (Yeh et al., 2016). The primary antibodies used in this study were anti-GFP (#sc-9996, Santa Cruz Biotechnology) and anti-HA (#sc-7392, Santa Cruz Biotechnology).
Yeast two-hybrid (Y2H) assays
Using LR reaction, the full-length CDS of ERF19 was subcloned into the pGADT7 vector, and NINJA, HDA6, and HDA19 CDS were introduced into the pGBKT7 vector. The constructs were transformed into yeast strain AH109 based on the LiAc-mediated transformation protocol following the manufacturer’s instructions (Clontech). At least 10 co-transformed yeast colonies were plated on Synthetic Drop-Out (SD) medium supplemented with X-α-Gal (Clontech) but without leucine, tryptophan, and histidine (-L-W-H). The plates were incubated at 30 °C for 3 d to test the nutritional marker gene expression and galactosidase activity of the MEL1 reporter protein.
Accession numbers
Sequence data from this article can be found in the Arabidopsis Genome Initiative under the accession numbers: ERF19 (AT1G22810), NINJA (AT4G28910), HDA6 (AT5G63110), HDA19 (AT4G38130), UBQ10 (AT4G05320), PDF1.2 (AT5G44420), PDF1.3 (AT2G26010), PR1 (AT2G14610), and PR2 (AT3G57260).
Results
Overexpression of ERF19 enhances Arabidopsis susceptibility to pathogens
To identify TFs involved in the regulation of Arabidopsis defenses against the fungal pathogen B. cinerea, we designed a screen to evaluate the resistance of Arabidopsis from the AtTORF-Ex collection (Weiste et al., 2007; Wehner et al., 2011) to this pathogen. Notably, we found a transgenic line overexpressing ERF19/ERF019 (At1g22810, HA-ERF19) that developed increased disease lesions after drop inoculation with B. cinerea spores (Supplementary Fig. S1A–C). To confirm that the increased susceptibility phenotype of the HA-ERF19 line to B. cinerea was not due to multiple transformation events (Weiste et al., 2007), we generated additional Arabidopsis lines expressing the CDS of ERF19 fused with GFP under the control of the Cauliflower mosaic virus 35S (CaMV 35S) promoter in the Col-0 background. Two independent lines (ERF19-OE1 and -OE2), expressing high levels of ERF19 mRNA and ERF19–GFP proteins (Supplementary Fig. S2A, B), were selected for further analyses. Confirming the increased susceptibility to B. cinerea observed in HA-ERF19 (Supplementary Fig. S1B, C), ERF19-OE1 and -OE2 developed larger disease lesions than Col-0 after B. cinerea drop inoculation (Fig. 1A). In addition to ERF19-OEs, we generated transgenic lines expressing the CDS of the ERF19-GFP fusion under the control of the β-Est-inducible XVE system (ERF19-iOEs). Overexpression of ERF19 and ERF19–GFP was β-Est dependent (Supplementary Fig. S2C, D). Confirming data observed in lines constitutively overexpressing ERF19 (Fig. 1A), increased susceptibility to B. cinerea was observed in ERF19-iOEs treated with β-Est, but not in mock controls treated with DMSO (Supplementary Fig. S3). Importantly, β-Est treatment did not alter Col-0 resistance to B. cinerea as compared with the DMSO-treated control (Supplementary Fig. S3), indicating that the increased susceptibility to B. cinerea in ERF19-iOEs is specifically linked to overexpression of ERF19 rather than to the β-Est treatment. In summary, our phenotypic analyses on HA-ERF19, ERF19-OEs, and ERF19-iOEs show that overexpression of ERF19 enhances Arabidopsis susceptibility to B. cinerea. Confirming earlier work (Scarpeci et al., 2017), the rosette leaves of 5-week-old ERF19-OEs exhibited different degrees of inward curling, and the rosette biomass of ERF19-OEs was smaller than that of the WT Col-0 (Supplementary Fig. S4A, B). However, unlike ERF19-OEs, the rosettes of ERF19-iOE and Col-0 plants were indistinguishable when grown in laboratory conditions (Supplementary Fig. S4C). Since ERF19-OE and β-Est-treated ERF19-iOE lines showed similar enhanced susceptibility to B. cinerea, the observed enhanced susceptibility phenotype to B. cinerea in ERF19-OEs (Fig. 1A) is probably not linked to the altered growth phenotype of these OE lines.
ERF19-OEs are hypersusceptible to B. cinerea and Pst DC3000. (A) B. cinerea-mediated lesions. Leaves of 5-week-old ERF19-OEs were droplet inoculated with 8 µl of B. cinerea spore suspension (105 spores ml–1 in 1/4 PDB). Disease symptoms were photographed and lesion perimeters were measured at 3 days post-inoculation (dpi). Data represent the average ±SE of at least 72 lesion perimeters pooled from three independent experiments each with at least six plants per line. Asterisks indicate a significant difference from Col-0 based on a t-test (***P<0.001). (B) Pst growth and symptoms. Five-week-old plants were dip inoculated with 106 cfu ml–1Pst, and symptoms were photographed at 3 dpi. Bacterial populations in the leaves were evaluated at 2 dpi. Values represent the average ±SE from three independent experiments pooled, each with five plants per line (n=15). Asterisks indicate a significant difference from Col-0 based on a t-test (**P<0.01; ***P<0.001).
To dissect the role of ERF19 in Arabidopsis resistance to microbial pathogens further, ERF19-OEs and Col-0 plants were dip inoculated with Pst DC3000, and disease symptoms were evaluated 3 d later. ERF19-OEs developed increased disease symptoms as indicated by widespread chloroses on the leaves of ERF19-OEs (Fig. 1B). Consistently, bacterial growth assays revealed that ERF19-OEs harbored at least 10 times more bacteria than Col-0 plants (Fig. 1B), indicating that ERF19-OEs were hypersusceptible to Pst bacteria. Collectively, these data suggest that overexpression of ERF19 in Arabidopsis induces hypersusceptibility to both fungal and bacterial pathogens.
ERF19 is transiently induced by MAMPs
To evaluate further the role of ERF19 in Arabidopsis immunity, we first monitored the expression of ERF19 in Col-0 seedlings after inoculation with B. cinerea spores or treatment with the fungal MAMP chitin over a 24 h period. ERF19 transcripts were up-regulated by B. cinerea spores or chitin within half an hour, and steadily declined at later time points (Fig. 2A). These results are consistent with previous reports showing that ERF19 is rapidly induced by chitin and chitin derivatives (Ramonell et al., 2005; Libault et al., 2007; Fakih et al., 2016). Signaling pathways of phytohormones such as salicylic acid (SA), JA, and ethylene (ET) are important for transcriptional regulation of immune regulators (Pieterse et al., 2009, 2012). To dissect the regulation of chitin-induced ERF19, we examined the expression of ERF19 after chitin treatment in npr1-1, coi1-16, and ein2-1 mutants, which are defective in SA, JA, and ET signaling pathways, respectively (Guzmán and Ecker, 1990; Cao et al., 1994; Ellis and Turner, 2002). Chitin-induced ERF19 transcripts in ein2-1, npr1-1, and coi1-16 were similar to their respective WT within 1 h post-treatment (Fig. 2B, C). These data indicate that rapid induction of ERF19 by chitin is unaffected when SA, JA, or ET signaling is impaired.
Expression analyses of ERF19. (A) Time course expression of ERF19 after inoculation with B. cinerea or treatment with chitin. Twelve-day-old seedlings were inoculated with a suspension of 5 × 105B. cinerea spores ml–1 or treated with 200 µg ml–1 chitin. Samples were collected at the indicated time points, and ERF19 expression was determined by qRT-PCR. After normalization with UBQ10, ERF19 expression levels were compared with time 0 (defined value of 1). Data represent the mean ±SD of three replicates (n=3). Asterisks denote values significantly different from time 0 based on a t-test (*P<0.05). (B) Chitin-induced ERF19 in ein2-1, npr1-1, and WT Col-0. Twelve-day-old seedlings were treated with 200 µg ml–1 chitin for 30 min and 60 min. Samples were collected at the indicated time points, and ERF19 expression was analyzed as in (A). Data represent the mean ±SD of four replicates (n=4). No significant differences in ERF19 expression were found between Col-0 and the mutants at different time points (t-test; P>0.05). (C) Chitin-induced ERF19 expression in coi1-16 mutant and its WT Col-6 was evaluated as in (B). Data represent the mean ±SD of four replicates (n=4). No significant differences of ERF19 expression were found between Col-6 and the coi1-16 mutant at different time points (t-test; P>0.05). (D) Time course expression of ERF19 after inoculation with Pst or after treatment with flg22 or elf18. Twelve-day-old seedlings were inoculated with 107 cfu ml–1Pst, or treated with 100 nM flg22 or 100 nM elf18, and samples were collected at the indicated time points. Analysis of ERF19 expression was performed and presented as in (A). Asterisks denote values significantly different from the respective time 0 based on a t-test (*P<0.05).
Since overexpression of ERF19 induced hypersusceptibility to Pst bacteria, we also monitored the expression of ERF19 in Col-0 seedlings after inoculation with Pst, or after treatment with the bacterial MAMPs flg22 or elf18. Similarly to B. cinerea spores or chitin, inoculation with Pst or treatments with flg22 or elf18 transiently up-regulated ERF19 for 1 h, but ERF19 transcripts declined steadily afterwards (Fig. 2D). To ensure that ERF19 expression levels observed in Fig. 2A, D are not a consequence of the experimental conditions, we also performed a time course study of ERF19 expression after mock (water or 1/2 MS) treatment. No up-regulation of ERF19 was observed in the mock controls (Supplementary Fig. S5). Together these data show that ERF19 is transiently up-regulated upon activation of Arabidopsis immunity.
PTI responses are down-regulated in ERF19 overexpression lines
Plants utilize PTI as a defense mechanism to ward off diverse pathogens (Boller and Felix, 2009; Huang and Zimmerli, 2014), and perturbation of PTI compromises plant defense against both fungal and bacterial pathogens (Tsuda et al., 2009; Kim et al., 2014). Since ERF19-OEs showed an increased susceptibility to both B. cinerea and Pst DC3000 and since ERF19 was up-regulated by fungal and bacterial MAMPs (Fig. 2A, B), we evaluated whether ERF19 is involved in PTI. Towards this goal, we first measured callose deposition, a PTI output activated by fungal and bacterial MAMPs (Millet et al., 2010; Shinya et al., 2014), in ERF19-OEs and Col-0. While the water-treated callose deposits were similar between ERF19-OEs and Col-0, callose deposition induced by chitin, flg22, or elf18 was significantly impaired in ERF19-OEs (Fig. 3A). Next, the expression of PTI maker genes was monitored in ERF19-OEs and Col-0 after MAMP treatments. Transcripts of chitin-induced PDF1.2 and PDF1.3, as well as flg22- or elf18-induced PR1 and PR2 were lower in ERF19-OEs than in Col-0 (Fig. 3B–D; Supplementary Fig. S6A–C), indicating a defective up-regulation of these PTI marker genes when ERF19 is overexpressed. Lastly, we tested the plant sensitivities toward flg22- and elf18-mediated growth arrest, a well-documented feature of PTI (Gómez-Gómez and Boller, 2000; Zipfel et al., 2006; Ranf et al., 2011). While treatment with flg22 or elf18 profoundly inhibited the growth of Col-0 seedlings, the MAMP-mediated growth inhibition effect was significantly lower in ERF19-OEs (Fig. 3E, F). Interestingly, smaller sizes of adult ERF19-OEs were not observed at an early developmental stage (compare Fig. 3E, F and Supplementary Fig. S4A, B). Taken together, these results show that overexpression of ERF19 alters the activation of common PTI responses and MAMP-mediated growth inhibition.
ERF19 is involved in PTI. (A) MAMP-induced callose deposition in ERF19-OEs. Fourteen-day-old seedlings were treated with deionized water (mock control), 200 µg ml–1 chitin, 100 nM flg22, or 100 nM elf18, and samples were collected 24 h later for aniline blue staining. Data represent the average numbers of callose deposits per square millimeter ±SE pooled from four independent experiments each with at least six biological repeats (n >24). Asterisks denote values significantly different from the respective Col-0 controls based on a t-test. (*P<0.01). (B–D) Activation of PTI marker genes in ERF19-OEs. Chitin-induced PDF1.2 (B), flg22-induced PR1 (C), and elf18-induced PR1 (D) in ERF19-OEs were determined by qRT-PCR. Twelve-day-old seedlings were treated with 200 µg ml–1 chitin, 1 µM flg22, or 1 µM elf18. Samples were collected at the indicated time points. After normalization with UBQ10, expression levels of PTI marker genes were compared with Col-0 at time 0 (defined value of 1). Data represent the mean ±SD of three replicates (n=3). Asterisks denote values significantly different from the respective Col-0 controls based on a t-test (*P<0.05). (E) MAMP-mediated growth inhibition in ERF19-OEs. Five-day-old seedlings were grown in liquid 1/2 MS supplemented with 1 µM flg22 or 100 nM elf18. Seedlings were weighed 10 d after treatment. Data represent the average fresh weight of 10 seedlings ±SE from three independent experiments (n=3). Asterisks indicate a significant difference from the respective Col-0 controls based on a t-test (*P<0.05). (F) Representative seedlings treated as in (E).
Expression of the dominant-negative ERF19–SRDX transgene enhances Arabidopsis PTI responses
To determine further the biological role of ERF19, we aimed to investigate the dominant-negative actions of ERF19 in planta, a commonly used strategy for studying TF functions (Mitsuda and Ohme-Takagi, 2009). We first examined the transcriptional activity of ERF19 by using PTA assays based on the GAL4/UAS and dual-luciferase reporter system. In Arabidopsis protoplasts, expression of ERF19 fused to the GAL4DB showed higher luciferase activity than expression of GAL4DB alone (Fig. 4A, B), suggesting that ERF19 acts as a transcription activator. Importantly, PTA assays revealed that the fusion of a plant-specific EAR-motif repression domain (SRDX) (Hiratsu et al., 2003; Mitsuda et al., 2011) to ERF19 successfully converted the activator feature of ERF19 into a repressor (Fig. 4A, B), indicating that the chimeric repressor ERF19–SRDX is appropriate for studying the dominant-negative actions of ERF19.
Expression of the dominant repressor ERF19–SRDX enhances PTI. (A) Schematic diagrams of reporter, effector, and reference plasmids used in the PTA assay. (B) PTA assay. Relative luciferase activities were evaluated in Arabidopsis protoplasts co-transfected with the reporter plasmid (UAS:fLUC), the effector plasmids (35S:GAL4DB, 35S:GAL4DB-ERF19, or 35S:GAL4DB-ERF19-SRDX), and a calibrator plasmid encoding rLUC. Protoplasts transfected without the effector plasmids were used as a control (no effector). All the values were normalized to the rLUC activity and were relative to the values of the no effector control. Values are means ±SE of four independent experiments (n=4). Different letters denote significant differences between groups based on a one-way ANOVA (P<0.01). (C) MAMP-induced callose deposition in ERF19–SRDXs. Fourteen-day-old seedlings were treated with deionized water (mock control), 200 µg ml–1 chitin, 100 nM flg22, or 100 nM elf18, and samples were collected 24 h later for aniline blue staining. Data represent the average numbers of callose deposits per square millimeter ±SE pooled from three independent experiments each with at least six biological repeats (n >24). Asterisks denote values significantly different from the respective Col-0 controls based on a t-test (*P<0.01). (D–F) Activation of PTI marker genes in ERF19–SRDXs. Chitin-induced PDF1.2 (D), flg22-induced PR1 (E), and elf18-induced PR1 (F) in ERF19–SRDXs were determined by qRT-PCR. Twelve-day-old seedlings were treated with 200 µg ml–1 chitin, 1 µM flg22, or 1 µM elf18. Samples were collected at the indicated time points, and UBQ10 was used for normalization. Relative gene expression levels were compared with Col-0 at time 0 (defined value of 1). Data represent the mean ±SD of three replicates (n=3). Asterisks denote values significantly different from the respective Col-0 controls based on a t-test (*P<0.05). (G) MAMP-mediated growth inhibition in ERF19–SRDXs. Five-day-old seedlings were grown in liquid 1/2 MS supplemented with 100 nM flg22 or 25 nM elf18. Seedlings were weighed 10 d after treatment. Data represent the average fresh weight (FW) of 10 seedlings ±SE from three independent experiments (n=3). Asterisks indicate a significant difference from the respective Col-0 controls based on a t-test (*P<0.05). (H) B. cinerea-mediated lesions in ERF19–SRDXs. Leaves of 5-week-old plants were droplet inoculated with 8 µl of B. cinerea spores (105 spores ml–1 in 1/4 PDB). Lesion perimeters were measured at 3 dpi. Data represent the average ±SE of 138 lesion perimeters (n=138) pooled from four independent experiments each with at least six plants per line. Asterisks indicate a significant difference from Col-0 based on a t-test (***P<0.001). (I) Pst DC3000 growth in ERF19–SRDXs. Five-week-old plants were syringe infiltrated with H2O or 10 nM elf18 6 h before syringe infiltration with 106 cfu ml–1Pst. Bacterial populations in the leaves were evaluated at 2 dpi. Values represent the average ±SE from three independent experiments each with three plants per line pooled (n=9). Different letters denote significant differences between groups based on a two-way ANOVA (P<0.01).
To assess further the biological function of ERF19–SRDX, the ERF19 genomic sequence, consisting of the intergenic promoter region (base pairs –1 to –1535), the 5'-untranslated region, and the CDS of ERF19 fused to the SRDX CDS, was expressed in Col-0 to generate ERF19–SRDX lines. The use of a native promoter of ERF19 better reflects the biological function of ERF19–SRDX than a constitutive promoter (Mitsuda and Ohme-Takagi, 2009). Two independent lines of ERF19–SRDX, of which the transgene ERF19–SRDX was chitin responsive, were selected for further analyses (Supplementary Fig. S7A). Unlike ERF19-OEs, the rosettes of ERF19–SRDXs were indistinguishable from those of the Col-0 WT (Supplementary Fig. S7B). To confirm the role of ERF19 in PTI and pathogen resistance, we first analyzed MAMP responses of ERF19–SRDX lines. Remarkably, MAMP-induced callose deposits were higher in ERF19–SRDXs than in Col-0 (Fig. 4C). Similarly, chitin-induced PDF1.2 and PDF1.3, flg22-induced PR1, and elf18-induced PR1 and PR2, were higher in ERF19–SRDXs than in Col-0 plants (Fig. 4D–F; Supplementary Fig. S7C, E). Surprisingly, despite enhanced expression of flg22-induced PR1, ERF19–SRDXs showed WT expression levels of flg22-induced PR2 (Supplementary Fig. S7D). Confirming the augmented PTI responses, MAMP-induced growth arrest was much more severe in ERF19–SRDXs (Fig. 4G). Together, these results suggest that transgenic expression of ERF19–SRDX enhances Arabidopsis PTI responses and MAMP-induced inhibition of growth.
In ERF19-OEs, the enhanced susceptibility to fungal and bacterial pathogens was correlated with reduced PTI responses. We thus hypothesized that the heightened PTI activation in ERF19–SRDX plants will confer pathogen resistance. As expected, ERF19–SRDXs exhibited smaller disease lesions than Col-0 WT upon B. cinerea infection (Fig. 4H), indicating that ERF19–SRDXs were more resistant to B. cinerea than Col-0 plants. However, Col-0 and ERF19–SRDXs developed similar Pst-mediated disease symptoms (Supplementary Fig. S7F). To highlight the role of ERF19–SRDX in PTI-mediated defense against Pst DC3000, we activated Arabidopsis PTI by treatment with 10 nM of the MAMP elf18 prior to Pst DC3000 inoculation. In water-treated controls, bacterial growth was similar in Col-0 and ERF19–SRDXs (Fig. 4I), confirming our previous observation showing that ERF19–SRDXs infected with Pst DC3000 exhibit WT disease symptoms (Supplementary Fig. S7F). Strikingly, a decrease of Pst DC3000 growth by elf18 pre-treatment was significantly stronger in ERF19–SRDXs than in Col-0 plants (Fig. 4I), suggesting that elf18-induced resistance to Pst DC3000 was enhanced in ERF19–SRDXs. Together, these results show that the expression of the dominant repressor ERF19–SRDX boosts PTI responses and, consequently, can confer increased resistance to fungal and bacterial pathogens. In summary, our phenotypic analyses on ERF19-OEs and ERF19–SRDXs provide genetic evidence that ERF19 plays a negative role in the regulation of Arabidopsis PTI and defense towards pathogens.
ERF19 is a nuclear TF
To determine the subcellular localization of ERF19, we took advantage of the high expression levels of ERF19–GFP in β-Est-treated ERF19-iOE1 (Supplementary Fig. S2C). Confocal microscope images revealed that strong GFP signals co-localized with DAPI-stained nuclei in the seedling roots of β-Est-treated ERF19-iOE1 (Fig. 5), indicating that ERF19–GFP is enriched in the nucleus. In contrast, the GFP alone control roots of transgenic seedlings showed a dispersed nuclear and cytoplasmic fluorescence (Fig. 5). These data suggest a nuclear localization for ERF19.
Subcellular localization of ERF19–GFP. Pictures were taken from seedlings of 12-day-old ERF19-iOE1 treated with 20 µM β-Est for 24 h and 35S:GFP transgenic lines. DAPI staining was used to determine the position of nuclei. Strong green fluorescence (green) of ERF19–GFP was co-localized with the DAPI-stained (magenta) nuclei. The scale bar represents 5 µm.
NINJA associates with and represses ERF19
The activities of TFs can be regulated via protein–protein interactions. The identification of TF-interacting proteins is thus crucial to unravel the regulation of TF regulatory networks (Licausi et al., 2013). ERFs were reported to form complexes with co-repressors and histone deacetylases (HDAs) (Kagale and Rozwadowski, 2011). We thus tested whether ERF19 associates with the well-studied HDA6 and HDA19 (Liu et al., 2014) via BiFC assays. Since HDA6 and HDA19 are components of the NINJA co-repressor complex (Zhang et al., 2017), we also included NINJA in the assay. Reconstitution of YFP in the nucleus was observed when ERF19 fused to the N-terminus of YFP was co-expressed with NINJA, HDA6, or HDA19 fused to the C-terminus of YFP in the leaves of N. benthamiana (Fig. 6A), suggesting that ERF19 can interact with these proteins in planta. Co-expression of ERF19–nYFP and cYFP alone did not show any yellow fluorescence (Fig. 6A). Consistently, Co-IP analyses revealed that ERF19-HA3 proteins could be pulled down along with NINJA–GFP, HDA6–GFP, and HDA19–GFP proteins, but not GFP alone (Fig. 6B, C), further strengthening the idea that ERF19 associates with these proteins in planta. However, in our Y2H assays, only NINJA is capable of associating with ERF19 in vitro (Fig. 6D), suggesting that ERF19–HDA6 and ERF19–HDA19 association requires plant-specific factors. As NINJA, HDA6, and HDA19 are probably part of a co-repressor complex, we tested via PTA analysis whether NINJA, HDA6, or HDA19 alters the transcriptional activity of ERF19. Interestingly, only co-transfection of NINJA with GAL4DB-ERF19 strongly and significantly repressed ERF19-activated luciferase activity (Fig. 6E). Taken together, these data suggest that NINJA associates with ERF19 and plays a negative role in the transcriptional activity of ERF19.
NINJA associates with and represses the transcriptional activity of ERF19. (A) BiFC analysis. N. benthamiana plants were co-transformed with the indicated split YFP constructs and a nuclear marker construct carrying NLS–mCherry–mCherry–NLS. YFP fluorescence (yellow), nucleus (blue), chlorophyll autofluorescence (red), bright field, and overlay images are shown. This experiment was performed at least three times with similar results. Scale bars represent 20 µm. (B and C) Analysis of ERF19–NINJA, ERF19–HDA6, and ERF19–HDA19 association by Co-IP. Total proteins from protoplasts expressing GFP, NINJA–GFP, HDA6–GFP, or HDA19–GFP with ERF19-HA3 were immunoprecipitated (IP) with anti-GFP antibodies. Total proteins before (input) and after IP (GFP-IP) were immunoblotted with anti-GFP and anti-HA antibodies. Similar results were obtained from three independent experiments. (D) Analysis of ERF19–NINJA, ERF19–HDA6, and ERF19–HDA19 association by Y2H assays. Ten-fold serial dilutions of yeasts expressing the indicated protein fusion to the activation domain (AD) or binding domain (BD) of GAL4 were plated on control (-L-W) or selective (-L-W-H/+X-α-Gal) SD media. Growth and blue staining of the colonies on selective SD medium indicate association between the two fusion proteins. The experiment was performed three times with similar results. (E) PTA assay. Relative luciferase activities of Arabidopsis protoplasts co-transfected with the reporter plasmid (UAS:fLUC), the effector plasmids (35S:GAL4DB, 35S:GAL4DB-ERF19, or 35S:GAL4DB-ERF19 with 35S:NINJA, 35S:HDA6, or HDA19), and a normalization plasmid encoding rLUC. All the values were normalized to the rLUC activity and were relative to the values of the no effector control. Different letters denote significant differences between groups based on a one-way ANOVA (P<0.05).
Since only NINJA strongly repressed ERF19 transactivation, we focused on studying the biological impact of NINJA on ERF19. To this end, disease resistance of the NINJA loss-of-function mutant ninja-1 overexpressing ERF19 was tested (Acosta et al., 2013). Two ERF19-OEs/ninja-1 independent lines overexpressing ERF19-GFP and with increased ERF19–GFP proteins, that demonstrated comparable expression levels to ERF19-OEs in the Col-0 background (Supplementary Fig. S8A, B), were selected for phenotypical analyses. The ninja-1 mutant appeared to have a long petiole phenotype when grown in our laboratory conditions, and ERF19-OEs/ninja-1 plants showed reduced rosette and leaf sizes (Supplementary Fig. S8C). We first evaluated the B. cinerea resistance of ERF19-OEs/ninja-1 lines by droplet inoculation of B. cinerea. The ninja-1 mutant developed slightly, but significantly smaller lesions than Col-0 WT plants (Supplementary Fig. S8D), and this may be due to the de-repression of the JA signaling pathway that contributes to B. cinerea resistance, in ninja-1 (Gasperini et al., 2015; Zhang et al., 2017). In contrast, overexpression of ERF19 in ninja-1 reversed the resistant phenotype of ninja-1 to susceptible levels comparable with ERF19-OEs (Supplementary Fig. S8D). We speculated that the droplet inoculation method did not faithfully reflect the susceptibility of ERF19-OEs/ninja-1 to B. cinerea as the disease evaluation was limited by the leaf size, with ERF19-OEs/ninja-1 being much smaller than the other lines (Supplementary Fig. S8C). The disease resistance of ERF19-OEs/ninja-1 against B. cinerea was thus assessed through spray inoculation, and progression of B. cinerea was ranked according to disease symptoms. After spray inoculation with B. cinerea spores, ERF19-OEs/ninja-1 lines developed dramatic disease symptoms. Most of the plants were indeed heavily or completely macerated at 5 dpi (Fig. 7A, B). In contrast, ERF19-OE plants exhibited only several macerated leaves, and symptoms were less severe than in ERF19-OEs/ninja-1 (Fig. 7A, B). While the majority of the Col-0 and ninja-1 plants developed symptoms with necrotic spots, they showed the least severe symptoms of the lines tested (Fig. 7A, B). These results indicate that a loss of NINJA function strongly enhanced susceptibility to B. cinerea in ERF19 overexpression lines. Since overexpression of ERF19 increased Arabidopsis sensitivity to B. cinerea, a further increase in sensitivity by a loss of NINJA function in ERF19-OEs/ninja-1 plants implies that NINJA represses the function of ERF19 in Arabidopsis immunity against B. cinerea. In summary, our data based on biochemical and genetic approaches strongly suggest that NINJA associates and negatively regulates the function of ERF19 in Arabidopsis immunity.
Hypersusceptibility to B. cinerea by ERF19 overexpression is enhanced in ninja-1. (A and B) B. cinerea resistance in transgenic lines overexpressing ERF19. Four-week-old plants were spray inoculated with a B. cinerea spore suspension (105 spores ml–1 in 1/4 PDB). Symptoms were photographed (A) and disease ranks were determined (B) at 5 dpi. Data in (B) represent 90 biological replicates (n=90) pooled from three independent experiments. The distribution of the disease rank proportions among the lines was analyzed using the χ2 test. Groups that do not share a letter are significantly different in the distribution of disease ranks (P<0.01).
Discussion
ERF19 negatively regulates PTI
ERF19 was first identified as one of the genes highly induced by chitin (Libault et al., 2007) and is used as a marker for chitin elicitation (Fakih et al., 2016). ERF19 is also involved in the regulation of plant growth, flowering time, and senescence, and positively regulates drought tolerance (Scarpeci et al., 2017). Here we report that ERF19 functions as a negative regulator of Arabidopsis immunity. The fact that ERF19 positively regulates drought tolerance and negatively regulates immunity suggests a potential role for ERF19 in modulating the crosstalk between abiotic and biotic stress signaling pathways (Atkinson and Urwin, 2012). Notably, our phenotypic studies of ERF19-OEs and ERF19–SRDXs show that ERF19 negatively regulates disease resistance against the fungus B. cinerea and Pst DC3000 bacteria. Although ERF19-OEs exhibited curly leaves and reduced rosette size, the increased disease susceptibility of ERF19-OEs is probably not linked to the altered developmental habitus of ERF19 overexpression. Indeed, we showed that ERF19-iOEs with appearance and morphology indistinguishable from those of the WT Col-0 were also hypersusceptible to B. cinerea when ERF19 overexpression was induced by β-Est. These observations suggest that an altered plant growth pattern is not the major determinant of ERF19-mediated susceptibility. In line with this argument, small size plants, as a result of overexpression of TFs, could display either increased or decreased resistance against pathogens (Chen and Chen, 2002; Xing et al., 2008; Tsutsui et al., 2009), further suggesting that plant growth habitus is not a decisive measure of plant resistance. Importantly, the altered B. cinerea and Pst DC3000 resistance in ERF19-OEs and ERF19–SRDXs was correlated with an altered activation of PTI. PTI functions through common signaling pathways to activate transcriptionally defense responses against invading pathogens (Kim et al., 2014). The necrotrophic fungus B. cinerea and the hemi-biotrophic bacterium Pst DC3000 are distinct microorganisms and therefore the observed altered resistance to different types of pathogens may be the result of perturbations of a broad spectrum immunity such as the PTI signaling network. Up-regulation of MAMP-specific marker genes was indeed repressed in ERF19-OEs and enhanced in ERF19–SRDXs, suggesting that ERF19 negatively regulates the PTI signaling network. In addition, ERF19 was induced by fungal and bacterial MAMPs, and the diverse natures of these MAMPs further imply that ERF19 is a critical, downstream regulator in a common, general PTI signaling network. Since ERF19 acted as a transcriptional activator when analyzed by PTA assays and PTI was negatively correlated with ERF19 function (ERF19-OE versus ERF19–SRDX), we propose that the repression of PTI signaling by ERF19 is likely to be mediated through the transcriptional activation of negative regulators of PTI. These negative regulators, which may consist of repressors, co-repressors, kinases, phosphatases, E3 ligases, histone modification enzymes, and miRNAs (Couto and Zipfel, 2016; Li et al., 2010; Schwessinger and Zipfel, 2008), could in turn transcriptionally, post-transcriptionally, and/or post-translationally suppress PTI signaling pathways (Fig. 8).
Proposed model for ERF19 and NINJA roles in PTI. MAMP perception initiates PTI repression signaling, triggering the induction of ERF19, in parallel with PTI activation signals. Accumulation of ERF19 may transcriptionally induce negative regulators of PTI, which are likely to be involved in the suppression of PTI signaling. PTI responses such as callose deposition, induction of PDF and PR genes, and MAMP-induced growth arrest are turned down by the ERF19-mediated pathway. The repressor NINJA provides another layer of control on PTI signaling through negative regulation of ERF19 function.
Transcriptional regulation of ERF19
Rapid and transient up-regulation of ERF19 by pathogens and MAMPs may seem paradoxical, since ERF19 plays a negative role in PTI activation. In fact, positive and negative regulators of immunity work in concert to mount appropriate levels of defense responses (Couto and Zipfel, 2016). In line with this, ERF4, ERF9, rice OsERF922, and potato StERF3 are induced by pathogens and function as negative regulators in plant immunity (McGrath et al., 2005; Liu et al., 2012; Maruyama et al., 2013; Tian et al., 2015). In addition, the L-type lectin receptor kinase-V.5 (LecRK-V.5), which is induced specifically in stomatal guard cells by Pst DC3000 and flg22, negatively regulates pathogen- and MAMP-induced stomatal closure, a common response of PTI (Melotto et al., 2006; Arnaud et al., 2012; Desclos-Theveniau et al., 2012). Furthermore, flg22-induced WRKY18 and WRKY40 act redundantly to regulate flg22-triggered genes negatively (Birkenbihl et al., 2017b). Collectively, these studies show that recognition of pathogens or MAMPs can transcriptionally induce negative regulators of immunity, which are necessary to buffer plant defense outputs.
The SA, JA, and ET pathways are known to play important roles in regulating the pathogen-induced TF network. For example, expression of ERF1 after Fusarium oxysporum f. sp. conglutinans inoculation depends on JA and ET signaling pathways and is independent of SA (Berrocal-Lobo and Molina, 2004). Similarly, B. cinerea-induced ERF96 requires intact JA and ET pathways (Catinot et al., 2015). In contrast, JA and ET signaling negatively regulate Pst-induced WRKY48 (Xing et al., 2008). By using appropriate mutants, we showed that rapid induction of ERF19 by chitin was unaffected when SA, JA, and ET signaling were individually impaired. It is possible that SA, JA, and ET act redundantly in the transcriptional control of chitin- (or MAMP-) induced ERF19 so that the loss of one defense pathway is compensated by other functional signaling pathways. Indeed, it has been shown that the transcriptional network of PTI signaling is highly buffered, robust, and tunable (Kim et al., 2014; Hillmer et al., 2017). The induction of ERF19 by chitin (or MAMPs) could also be regulated in addition to or independently of SA, JA, and ET.
ERF19 buffers MAMP-induced growth inhibition
Plant growth and immunity are maintained at a fine balance to ensure plant survival. In the presence of invading pathogens, positive and negative regulators of immunity together tailor this balance to ensure appropriate levels of defense outputs. Exaggerated defense responses that tip the balance towards immunity can hamper plant growth and survival. For example, constitutive activation of ERF6 or overexpression of ERF11 results in direct activation of defense genes, but these transgenic plants suffer from severe growth defects (Tsutsui et al., 2009; Meng et al., 2013). In addition, the L-type lectin receptor kinase-VI.2 (LecRK-VI.2) associates with FLS2 and functions as a positive regulator of PTI (Singh et al., 2012; Huang et al., 2014). Plants with high expression of LecRK-VI.2 show constitutive PTI responses but display a dwarf phenotype (Singh et al., 2012). Furthermore, loss of BAK1-INTERACTING RECEPTOR-LIKE KINASE 1 (BIR1), a negative regulator of plant immunity, leads to constitutive activation of defense responses and cell death, which dramatically hampers plant growth (Gao et al., 2009). These studies illustrate that genetic disruption of crucial immune regulators can deleteriously affect plant growth. Although ERF19 functions as a negative regulator of PTI, unlike the bir1 mutant (Gao et al., 2009), the ERF19–SRDX lines showed WT growth under normal conditions and did not exhibit constitutive activation of PTI responses. The dominant repressor ERF19–SRDX was regulated by the native promoter of ERF19. This basal expression of ERF19–SRDX might thus be insufficient to trigger constitutive PTI activation. In spite of normal growth, flg22- or elf18-induced growth inhibition was much more severe on ERF19–SRDX lines than on Col-0 WT, even at low concentrations of flg22 or elf18. The high sensitivity of ERF19–SRDXs to MAMP-mediated growth arrest implies that in response to MAMPs, ERF19 acts as a buffering regulator to prevent exaggerated growth arrest, which could negatively impact plant growth. In agreement with this, ERF19-OEs showed diminished growth inhibition imposed by high concentration of MAMPs. Taken together, our data suggest that ERF19 is part of a buffering mechanism to avoid exaggerated PTI activation and MAMP-mediated growth arrest to maintain a proper balance between growth and immunity upon MAMP recognition.
NINJA negatively regulates ERF19
Post-translational regulation such as protein–protein interaction is known to alter the transcriptional activities of TFs (Licausi et al., 2013). For example, EIN3 and MYC2, a crucial TF regulating JA signaling, interact and reciprocally affect each other’s functions (Song et al., 2014; Zhang et al., 2014). In addition, JAZ1 and JAZ proteins negatively regulate the functions of EIN3 and MYC TFs, respectively (Chini et al., 2007; Pauwels and Goossens, 2011; Zhu et al., 2011; Zhang et al., 2015). Such negative regulations are thought to modulate fine-tuning mechanisms to achieve rigorous transcriptional controls. NINJA was originally identified as the adaptor between JAZ proteins and the transcriptional co-repressors TPL and TPRs and was demonstrated to act as a negative regulator of JA signaling (Pauwels et al., 2010). Later studies showed that NINJA is also involved in the regulation of root growth (Acosta et al., 2013; Gasperini et al., 2015) and, together with topoisomerase II-associated protein PAT1H1, NINJA participates in the maintenance of root stem cell niche (Yu et al., 2016). In this study, we found a novel function for NINJA in the negative regulation of ERF19. The repression mechanism(s) of NINJA on ERF19 may be linked to ERF19 association with NINJA that in turn recruits other co-repressors such as TPL (Pauwels et al., 2010), and thus suppresses the transcription of the ERF19-bound loci. In addition, association with NINJA may change the conformation of ERF19 and subsequently inhibit the transcriptional function of ERF19 as observed in MYC3–JAZ9 regulation (Zhang et al., 2015). Such a conformational change may hinder the ability of ERF19 to recruit co-activators and/or to bind to DNA. Our data provide evidence that NINJA is involved in the regulation of ERF19 function and further suggest that through modulation of ERF19 at transcriptional and post-translational levels, plants can fine-tune PTI to cope with the vast variety of environmental stimuli they face.
Supplementary Data
Supplementary data are available at JXB online.
Fig. S1. Characterization of the HA-ERF19 line.
Fig. S2. Characterization of lines overexpressing ERF19.
Fig. S3. B. cinerea-mediated lesions in ERF19-iOE lines.
Fig. S4. Growth phenotypes of ERF19-OE and ERF19-iOE lines.
Fig . S5. Time course study of ERF19 expression after treatment with 200 µg ml–1 chitin, water, or 1/2 MS.
Fig. S6. Expression of PTI marker genes in ERF19-OEs.
Fig. S7. Characterization of ERF19–SRDXs.
Fig. S8. Characterization of ERF19-OEs/ninja-1.
Table S1. Primers used in this study.
Author contributions
P-YH and LZ designed the research; P-YH performed most experiments; Y-PL started the project by screening the AtTORF-Ex collection; JZ generated ERF19-iOE lines and performed Co-IP analyses; CC performed the time course study of ERF19 expression after mock treatments; BJ conducted Y2H and BiFC experiments and rough phenotyping of ERF19–SRDX and ERF19-OE/ninja-1 lines; J-HY performed flg22-induced ERF19 analysis; KC cloned ERF19-SRDX and generated ERF19–SRDX lines. P-YH and LZ analyzed the data and wrote the manuscript; and LZ supervised the project.
Acknowledgements
We thank ABRC, X. Dong, J.G. Turner, E.E. Farmer, and K. Wu for providing seeds, and C.-Y. Chen and B.N. Kunkel for B. cinerea and Pst bacteria, respectively. We also acknowledge members of the Zimmerli laboratory for critical comments. We thank the Technology Commons (TechComm), College of Life Science, National Taiwan University for providing qRT-PCR equipment and excellent technical assistance with the confocal laser scanning microscopy. This work was supported by the Ministry of Science and Technology of Taiwan, grants 106-2917-I-564-005-A1 (to P.-Y.H.), 98-2311-B-002-008-MY3, 102-2311-B-002-027, 103-2311-B-002-004, and 104-2311-B-002-003 (to L.Z.).
Comments