Derepression of ethylene-stabilized transcription factors (EIN3/EIL1) mediates jasmonate and ethylene signaling synergy in Arabidopsis

Abstract

Jasmonate (JA) and ethylene (ET) are two major plant hormones that synergistically regulate plant development and tolerance to necrotrophic fungi. Both JA and ET induce the expression of several pathogenesis-related genes, while blocking either signaling pathway abolishes the induction of these genes by JA and ET alone or in combination. However, the molecular basis of JA/ET coaction and signaling interdependency is largely unknown. Here, we report that two Arabidopsis ET-stabilized transcription factors (EIN3 and EIL1) integrate ET and JA signaling in the regulation of gene expression, root development, and necrotrophic pathogen defense. Further studies reveal that JA enhances the transcriptional activity of EIN3/EIL1 by removal of JA-Zim domain (JAZ) proteins, which physically interact with and repress EIN3/EIL1. In addition, we find that JAZ proteins recruit an RPD3-type histone deacetylase (HDA6) as a corepressor that modulates histone acetylation, represses EIN3/EIL1-dependent transcription, and inhibits JA signaling. Our studies identify EIN3/EIL1 as a key integration node whose activation requires both JA and ET signaling, and illustrate transcriptional derepression as a common mechanism to integrate diverse signaling pathways in the regulation of plant development and defense.

Plants are sessile organisms and face different environmental changes during their lifespan. To survive various abiotic and biotic stresses, plants synthesize a number of small molecules functioning as phytohormones to elaborately regulate their growth, development, and defense. Two types of phytohormones—ethylene (ET) and jasmonate (JA)—are crucial for plant development and defense against necrotrophic fungi infections (1–3). Complicated modes of interaction between ET and JA have been documented in different processes. For example, ET strongly suppresses JA-induced wounding-responsive gene expression, but JA suppresses ET-induced apical hook formation (4, 5), indicative of their antagonisms. Upon necrotrophic fungi infections, plants can quickly produce ET and JA and induce the expression of downstream defense genes (like ERF1, ORA59, and PDF1.2) that help plants tolerate or fight against the fungal pathogens (1). Plants treated with exogenous JA or ET express high levels of defense genes (6, 7), and simultaneous treatment with JA and ET results in the highest expression (8). Nevertheless, in the ET or JA insensitive mutant (ein2 or coi1, respectively), JA and ET alone or in combination fail to induce the expression of those defense genes (8, 9), indicating that the two hormone-signaling pathways are required concomitantly for the activation of plant-defense response. These results suggest that JA and ET act synergistically and mutually dependently in regulating necrotrophic pathogen responses. However, the molecular details underlying such hormone synergy and signaling interdependency are currently unknown.

ET is a gaseous hormone, which is perceived by its receptors and represses a Raf-like kinase CONSTITUTIVE ETHYLENE RESPONSE 1 (CTR1) (10, 11). Downstream of CTR1 is ETHYLENE INSENSITIVE 2 (EIN2), which is an essential positive regulator in ET signaling (12). ETHYLENE INSENSITIVE 3 (EIN3) and its closest homolog EIN3-LIKE 1 (EIL1) are two primary transcription factors downstream of EIN2 (6, 7, 13), which are necessary and sufficient for the induction of the vast majority of ethylene response genes (13–15). In the absence of ET, EIN3/EIL1 are quickly degraded through the action of two EIN3-binding F-box proteins 1 and 2 (EBF1 and EBF2), and ET enhances EIN3/EIL1 accumulation, probably through the repression of EBF1/2 (14, 16–18).

JA is synthesized and conjugated with isoleucine to form the active hormone JA-Ile (19, 20), which is perceived by its receptor CORONATINE INSENSITIVE 1 (COI1), an F-box protein (21–25). JASMONATE ZIM-DOMAIN (JAZ) proteins are transcriptional repressors, which interact with MYC2 transcription factor and repress its function (23, 26, 27). Recently, one possible mechanism underlying JAZ repression on MYC2 has been demonstrated in that JAZ proteins associate with NOVEL INTERACTOR of JAZ (NINJA) to recruit TOPLESS as a corepressor to fulfill this function (28). The perception of JA-Ile induces the interaction between COI1 and JAZs, which brings JAZs for degradation and relieves their repression on MYC2 (23, 26). Two MYC2-like bHLH proteins (MYC3 and MYC4) were also able to interact with JAZs and act addictively with MYC2 in mediating a subset of JA-regulated responses, including inhibition of root elongation, wound response, and metabolism (29–31). Besides MYC2/MYC3/MYC4, two MYB transcription factors (MYB21 and MYB24) were recently identified as JAZ targets in regulating plant stamen development (32).

Histone deacetylation, which is catalyzed by histone deacetylases (HDAs), is a type of chromatin modification commonly associated with transcriptional repression (33). It has been reported that histone deacetylation affects many growth and developmental events in plants, such as flowering time, cold tolerance, embryogenesis, root hair development, abscisic acid, and salt responses (34–38). A molecular link between this epigenetic control and JA signaling has been implied, as an RPD3-type histone deacetylase 6 (HDA6) was reported to associate with COI1 in a coimmunoprecipitation (co-IP) assay (39). It was also found that the transcript levels of some JA-responsive genes are altered in HDA6 or HDA19 loss-of-function mutants and transgenic overexpression plants (40, 41). However, the action of these HDAs in JA signaling is not clear.

Here, we report that EIN3 and EIL1 integrate ET and JA signaling in plant development and defense against necrotrophic pathogens. JAZ proteins directly interact with EIN3/EIL1 and recruit HDA6 as a corepressor to repress the transcriptional activity of EIN3/EIL1. Our work defines EIN3/EIL1 as a previously unexplored class of JAZ-associated transcription factors that are subject to JAZ repression via the action of histone deacetylation, and provides unique insights into how plant hormones coordinately regulate plant development and defense response.

Results

EIN3/EIL1 Are Positive Regulators Mediating a Subset of JA Responses.

Previous studies have revealed that JA alone or in combination with ET fails to induce pathogenesis-related genes (such as ERF1, ORA59, PDF1.2) in ein2 background (8, 9). Because ein3 eil1 is phenotypically similar to ein2 and one of ET/JA-regulated genes (ERF1) is a direct target of EIN3 (7, 13), we hypothesized that EIN3 and EIL1 may function as key transcriptional regulators of JA/ET-induced gene expression. We first examined the induction of JA-responsive genes in ein3 eil1 upon JA treatment. The JA induction of pathogenesis-related genes (ERF1, ORA59, and PDF1.2) was abolished in ein3 eil1, but the expression of wound-response genes (e.g., VSP2) that are regulated by MYC2 was still induced (Fig. 1A). ein3 eil1 was also insensitive to the combination of JA and ET treatments in terms of ERF1 induction (Fig. S1A). Because the expression of pathogenesis-related genes is correlated with plant tolerance to necrotrophic fungi (42), we tested plant response to Botrytis cinerea infection. Our result showed that plants lacking EIN3, EIN3/EIL1, or EIN2 were more susceptible to B. cinerea, whereas myc2 was more tolerant to the fungus (Fig. 1B and Fig. S1 B and C) (13).

Fig. 1.

EIN3 and EIL1 are positive regulators of JA signaling. (A) Analysis of JA-responsive gene expression by quantitative RT-PCR (qRT-PCR) in 7-d-old light-grown seedlings with or without 100 μM JA treatment for 4 h. Mean ± SEM, n = 3. (B) B. cinerea infection analysis. The percentages of different infection ratios in each genetic background were scored. (C and D) Root-hair phenotypes. (C) Representative picture of roots from seedlings grown on MS medium for 4 d and then transferred onto MS (Mock) or 25 μM JA medium for additional 2 d. Arrows indicate root hairs. (Scale bars, 50 μm.) (D) Root hair densities of 7-d-old seedlings grown on MS medium supplemented with indicated concentrations of JA were scored by counting the numbers of entire root hairs. Mean ± SEM, n = 10. (E and F) Relative root growth of 7-d-old seedlings grown on indicated concentrations of JA medium. Mean ± SEM, n = 10. **P ≤ 0.01, significant differences between Col-0 and mutants.

Next, we sought to determine whether loss of EIN3/EIL1 affects other JA-regulated responses. JA has been reported to induce root hair development (43). However, such induction in ein3 eil1 was largely abolished (Fig. 1 C and D). We also found that ein3 eil1 was partially insensitive in JA-induced inhibition of root elongation (Fig. 1E). Conversely, transgenic plants overexpressing EIN3 or EIL1 (EIN3ox or EIL1ox) (6) were hypersensitive to JA in root-growth inhibition (Fig. 1F). On the other hand, ein3 eil1 was indistinguishable from WT in terms of fertility and JA-induced anthocyanin biosynthesis (Fig. S2). Collectively, we conclude that EIN3 and EIL1 are positive regulators of a subset of JA responses, including pathogenesis-related gene expression, plant resistance to necrotrophic fungi, and root development, but not fertility and pigment metabolism.

JAZ Proteins Interact with EIN3/EIL1.

To understand how EIN3/EIL1 mediate JA signaling, we speculated that EIN3 and EIL1 might be targeted by JAZ proteins, and thus tested the interaction between JAZ proteins and EIN3/EIL1. We found that JAZ1 interacted with EIN3 in yeast cells, and an EIN3 fragment containing amino acid residues 200 to 500 was sufficient for JAZ1 binding (Fig. 2A). The interaction between JAZ1 and an EIL1 fragment (amino acids 201–501) was also observed (Fig. S3A). To test if EIN3/EIL1 interacts with other JAZ proteins, we chose JAZ3 and JAZ9 and found that both JAZ3 and JAZ9 were able to interact with EIN3 (amino acids 200–500) and EIL1 (amino acids 201–501) in yeast cells (Fig. S3B). A pull-down assay also verified JAZ1-EIN3 interaction in which Escherichia coli-expressed GST-JAZ1 successfully pulled down EIN3 from plant total protein extracts (Fig. 2B). To confirm JAZ1-EIN3 interaction in vivo, we performed a co-IP assay in the 35S:JAZ1-GUS transgenic plants (23) and found that the anti-GUS antibody was able to precipitate EIN3 along with JAZ1-GUS, indicating their association in planta (Fig. 2C). A bimolecular fluorescence complementation (BiFC) assay in onion epidermis cells also confirmed JAZ1-EIN3 interaction in vivo (Fig. 2D). To further characterize the interaction domain in JAZ1, we used an E. coli-expressed GST-fused N terminus of JAZ1 (containing the ZIM domain) or a GST-fused C terminus of JAZ1 (containing the Jas domain) to pull down EIN3 protein from extracted plant total proteins. Our result revealed that the C terminus of JAZ1 was mainly responsible for EIN3 binding (Fig. S3C). It has been demonstrated that the C terminus of JAZs (Jas domain) is the interaction domain for JAZs-MYC2 binding (26), therefore JAZ proteins seem to use the same Jas domains to interact with various transcription factors. Together, these results indicate EIN3 and EIL1 as a previously unexplored class of JAZs-associated transcription factors.

Fig. 2.

JAZ proteins physically interact with EIN3. (A) Yeast two-hybrid assays showing the interaction between EIN3 and JAZ1. White colony indicates protein interaction in yeast cells grown on selective medium. (B) Pull-down assay. E. coli-expressed GST or GST-JAZ1 proteins were incubated with protein extracts from Col-0 and further immobilized by glutathione Sepharose 4B. Half of the pull-down products were probed with anti-EIN3 antibody to test interaction (Upper) and the other half of pull-down products were subjected to Coomassie blue staining as loading controls (Lower). (C) EIN3 was coimmunoprecipitated with JAZ1. The 35S:JAZ1-GUS seedlings were treated with 100 μM MG132 for 4 h to stabilize JAZ1-GUS and EIN3 proteins. Protein extracts from JAZ1-GUS were immunoprecipitated with anti-GUS antibody (+) or no antibody (−). Precipitated products were probed with anti-GUS or anti-EIN3 antibody to detect the interaction. (D) BiFC assay in onion epidermis cells. YFP fluorescence indicates protein interaction. (Upper) Two arbitrary independent fields showing JAZ1-EIN3 interaction. (Lower) Negative controls. (Scale bars, 50 μm.)

JAZ Proteins Repress the Function of EIN3/EIL1.

We next determined whether JAZ proteins repress the function of EIN3/EIL1. First, we carried out a transient transcriptional activity assay in Arabidopsis protoplasts using ERF1:LUC (firefly luciferase-coding gene driven by the ERF1 promoter) as a reporter. Overexpression of EIN3 (35S:EIN3) induced ERF1:LUC expression, but overexpression of JAZ1 (35S:JAZ1) suppressed EIN3-induced expression of ERF1:LUC (Fig. 3A), indicating that JAZ1 somehow represses EIN3 function. Second, we introduced 35S:JAZ1D3A (a stable form of JAZ1) (23) into EIN3ox plants and found that overaccumulation of JAZ1 dramatically reduced the sensitivity of EIN3ox to JA (Fig. 3 B and C). Third, we found that the coi1 mutation also reduced the sensitivity of EIN3ox to JA in terms of ERF1 gene expression and root hair development (Fig. 3D). Finally, we introduced 5×EBS::GUS (the glucuronidase gene is driven by a promoter containing five tandem repeats of the EIN3-binding sites) (44) into Col-0, ein3-1, or ein3 eil1 backgrounds. Upon ET or JA treatment, the transcriptional activity of EIN3 indicated by the GUS level was augmented in Col-0 but diminished or abolished in ein3-1 or ein3 eil1, respectively (Fig. 3E). Together, these results demonstrate that JAZs repress the function of EIN3/EIL1 and suggest that JA might relieve such repression by removing JAZ proteins.

Fig. 3.

JAZ proteins repress the transcriptional activity of EIN3. (A) Transient assay of ERF1:LUC expression. ERF1:LUC-35S:REN reporter constructs were transiently expressed in Arabidopsis protoplasts together with control vector, 35S:EIN3, or 35S:EIN3 and 35S:JAZ1 effectors, respectively. The expression of REN was used as an internal control. LUC/REN ratio represents the relative activity of the ERF1 promoter. Results are repeated three times with the same pattern. Mean ± SEM, n = 3. (t test: ** P ≤ 0.01). (B and C) Root elongation phenotypes of 7-d-old seedlings grown on 20 μM JA medium. (Scale bar in B, 1 mm.) Mean ± SEM, n = 10. (D) Root hair densities of 7-d-old seedlings grown on MS (Mock) or 10 μM JA medium (Upper, mean ± SEM, n ≥ 5), and qRT-PCR results showing relative expression of ERF1 from 7-d-old seedlings treated without (Mock) or with 100 μM JA for 4 h (Lower, mean ± SEM, n = 3). (E) GUS staining. 5×EBS::GUS expression in 7-d-old seedlings treated with Mock solution, 100 μM ACC (ET) or 100 μM JA (JA) for 4 h. ACC (1-aminocyclopropane-1-carboxylic acid) is the ethylene biosynthetic precursor. (Scale bar, 1 mm.) (F) ChIP-PCR result showing that JA enhances the DNA binding ability of EIN3 in vivo. Chromatin was extracted from 12-d-old seedlings after either 100 μM JA or 100 μM ACC treatment for 3.5 h and was precipitated using anti-EIN3 antibody. Precipitated DNA was amplified by primers corresponding to sequences neighboring the EIN3 binding sites in the promoter of ERF1. The ein3 eil1 samples and no-antibody panels were negative controls showing the specificity of precipitation. Input indicates samples before immunoprecipitation.

Next, we investigated how JAZ proteins repress EIN3/EIL1. Because JAZ1 interacts with the EIN3 (amino acids 200–500) and EIL1 (amino acids 201–501) fragments that overlap with their DNA binding domains (amino acids 59–359 in EIN3) (7), we determined whether JAZ1 interferes with EIN3 binding to DNA. ChIP-PCR assay showed that the in vivo association of EIN3 to the ERF1 promoter was increased by JA or ET (Fig. 3F). However, an in vitro binding experiment using electrophoretic mobility shift assay failed to detect the interference of JAZ1 on EIN3 binding to the ERF1 promoter element. One explanation for this discrepancy is that JAZ repressors possibly need additional components (corepressors) to suppress the DNA binding of EIN3 in vivo.

JAZ Proteins Recruit HDA6 to Associate with EIN3/EIL1.

We then tested the possibility whereby corepressors are required for JAZ repression of EIN3 function. It was reported that HDA6 is coimmunoprecipitated with COI1 and its expression is up-regulated by both JA and ET (39, 41). In addition, histone deacetylation is generally associated with transcriptional repression (33). These features make HDA6 a good candidate to be a corepressor of JAZ proteins. We thus tested the physical interactions among JAZs, HDA6, and EIN3/EIL1. We found that HDA6 interacted with EIN3 (amino acids 200–500) and EIL1 (amino acids 201–501) in yeast cells (Fig. 4A). We also carried out a co-IP assay using an inducible EIN3-3×FLAG in ein3 eil1 ebf1 ebf2 background, in which the induced FLAG-tagged EIN3 becomes more stabilized because of the lack of EBF1 and EBF2 and thus easier to detect (14). Our results showed that the anti-FLAG antibody efficiently coprecipitated HDA6 together with FLAG-tagged EIN3, suggesting their association in vivo (Fig. 4B).

Fig. 4.

JAZ proteins recruit HDA6 to associate with EIN3. (A) Yeast two-hybrid assays indicating that HDA6 interacts with EIN3 and EIL1. (B) HDA6 was coimmunoprecipitated with EIN3-3×FLAG. EIN3-3×FLAG is preinduced in pER8-EIN3-3×FLAG/ein3 eil1 ebf1 ebf2 transgenic plants by 10 μM β-estrogen for 4 h and then immunoprecipitated with anti-FLAG antibody (+Ab) or no antibody (−Ab), followed by immunoblots with anti-FLAG or anti-HDA6 antibody. (C) Yeast two-hybrid assays indicating that HDA6 interacts with JAZ1. (D and E) Pull-down assays. His-JAZ1, His-APRT or His-APT3 proteins (D) and GST-JAZ1ΔC or GST-JA1ΔN proteins (E) were incubated with Col-0 protein extracts and further immobilized by Ni-NTA Agarose (D) or glutathione Sepharose 4B (E). Half of the pull-down products were probed with anti-HDA6 antibody to test interaction (D and E) and the other half of pull-down products were probed with anti-His (D) or anti-GST (E) as loading controls. His-APRT and His-APT3 are proteins as negative controls. (F and G) JA treatment reduces EIN3-HDA6 association. (F) pER8-EIN3-3×FLAG/ein3 eil1 ebf1 ebf2 seedlings were initially treated with 10 μM β-estrogen for 4 h and then treated with 100 μM JA (+JA) or mock (−JA) for additional 1 h. Extracted protein was immunoprecipitated with anti-FLAG antibody (+Ab) or no antibody (−Ab), followed by immunoblots with anti-FLAG or anti-HDA6 antibody. Ten percent of protein extracts were used as input controls. Protein extracts from Col-0 and axe1-4 were used as controls showing the specificity of anti-HDA6 antibody. (G) The relative intensities of HDA6 and EIN3 were calculated with Image J software (National Institutes of Health) from input or precipitated products (+Ab) to show the association of HDA6 and EIN3. Results from JA-treated samples (+JA) were normalized to samples without JA treatment (−JA). Similar results were obtained from two independent experiments and error bar represents SEM. (t test: *P = 0.026.)

Meanwhile, we found that HDA6 also interacted with JAZ1 in yeast cells (Fig. 4C). To verify JAZ1-HDA6 interaction, we did a pull-down assay and found that His-JAZ1 could pull down HDA6 purified from plant extracts (Fig. 4D). To further confirm HDA6-JAZ1 association, we performed a BiFC assay and demonstrated their in vivo interaction (Fig. S4B). We also tested the interactions between HDA6 and other JAZ proteins (e.g., JAZ3 and JAZ9) and found that HDA6 interacted with both JAZ3 and JAZ9 (Fig. S4A). To define the interaction domain of JAZ1 for HDA6 binding, we used GST-JAZ1 ΔN or GST-JAZ1 ΔC to do a pull-down assay and found that GST-JAZ1 ΔC was sufficient for HDA6 binding (Fig. 4E), suggesting that JAZ1 uses distinctive interaction modules for EIN3- and HDA6-binding. Because both JAZ1 and HDA6 interact with EIN3 (amino acids 200–500), we further performed yeast two-hybrid analysis with EIN3 deletion fragements and found that EIN3 (amino acids 300–500) was sufficient for JAZ1-EIN3 interaction but not for HDA6-EIN3 interaction (Fig. S4C), suggesting that JAZ1 and HDA6 interact with different domains of EIN3. To investigate the biological consequence of JAZ interactions with EIN3 and HDA6, we used anti-FLAG antibody to coprecipitate HDA6 in transgenic plants expressing FLAG-tagged EIN3 with or without JA treatment. We found that JA treatment reduced the in vivo HDA6-EIN3 association (Fig. 4 F and G), suggesting that JAZ proteins or other components are necessary for maximal interaction between HDA6 and EIN3 in vivo.

HDA6 Is a Repressor for EIN3-Mediated Transcription and JA Signaling.

Histone deacetylases usually associate with transcriptional repression. To investigate whether HDA6 is involved in the repression of EIN3-mediated transcription and JA signaling, we studied JA responses in HDA6 loss-of-function mutant (axe1-4) and transgenic overexpressing plant (HDA6ox) (Fig. S5A). The JA-induced ERF1 gene expression, root hair development, and root growth inhibition were enhanced in axe1-4 (Fig. 5 A and B and Fig. S5B) but were attenuated in HDA6ox (Fig. 5A and Fig. S5 B and C). Meanwhile, we used a specific histone deacetylase inhibitor, trichostatin A (TSA) (37) to mimic the loss of HDA activities. TSA treatment was sufficient to induce root hair formation and ERF1 expression in Col-0, similar to the effect of JA (Fig. S6 A–C). Interestingly, TSA induced root hair formation in the coi1-2 mutant but not in the ein3 eil1 mutant (Fig. S6A). Consistent with this observation, ein3 eil1 completely suppressed the JA-hypersensitivity phenotypes of axe1-4 (Fig. 5 A and B). These results suggest that HDA6 (and probably other histone deacetylases) is a negative regulator of the JA-regulated processes (e.g., ERF1 gene expression and root hair development) acting in an EIN3/EIL1-dependent manner.

Fig. 5.

HDA6 is a repressor for EIN3-mediated transcription and JA signaling. (A) qRT-PCR results showing relative expression of ERF1 from 7-d-old seedlings treated without (Mock) or with 100 μM JA for 4 h. Mean ± SEM, n = 3. Significant differences between Col-0 and axe1-4 or HDA6ox were indicated: **P ≤ 0.01. (B) Root-hair numbers. Four-day-old seedlings grown on MS were transferred to MS or 25 μM JA medium for 2 d. Root-hair numbers were counted from the 2-mL region of root tip under an Olympus microscope. Mean ± SEM, n ≥ 5. Significant differences between Col-0 and axe1-4 are indicated: **P ≤ 0.01. (C and D) Histone acetylation analysis by ChIP-PCR. Before ChIP analysis, seedlings were treated with or without 100 μM JA for 4 h. Antibodies specifically recognizing either acetylated histone H3 (AcH3) or H4 (AcH4) were used to precipitate chromatins. Precipitated DNA was amplified by primers corresponding to sequences neighboring the EIN3 binding sites in the promoter of ERF1. qPCR (C) or PCR product separated on ethidium bromide stained gel (D) shows the histone acetylation levels of the ERF1 promoter region. Mean ± SD, n = 3 (t test: *P ≤ 0.05, **P ≤ 0.01). Actin was used as a negative control. Input indicates samples before immunoprecipitation.

To further explore the role of histone deacetylation in EIN3-mediated transcription, we analyzed the histone acetylation levels in the promoter region of the EIN3 target gene, ERF1. Upon JA treatment, the levels of both histone H3 and H4 acetylation were enhanced, but in axe1-4, this enhancement was more pronounced, particularly for histone H4 acetylation (Fig. 5C), which agrees with JA induction of ERF1 expression in axe1-4 (Fig. 5A). Meanwhile, the levels of histone H4 acetylation in coi1-2 and ein3 eil1 were greatly reduced compared with that of Col-0 upon JA treatment (Fig. 5D), which again was in agreement with the levels of ERF1 expression and JA sensitivity in these mutants (Fig. 1A and Fig. S1A). Taking these data together, we conclude that JAZ proteins recruit HDA6 (and probably other HDAs as well) to deacetylate histones (especially H4) and obstruct the chromatin binding of EIN3/EIL1, but JA treatment removes JAZs-HDA6 corepressors away from EIN3/EIL1 by degrading JAZs.

Discussion

Emerging evidences show that the action of JA and ET is synergistic and mutually dependent in regulating tolerance to necrotrophic fungi (8, 9), although not much is known about the molecular details of this hormonal synergy and signaling interdependency. In this study, we report EIN3 and EIL1 as two JAZ-associated transcription factors integrating JA and ET signaling to regulate gene expression, pathogen defense, and root development. Under normal conditions, EIN3/EIL1 are repressed by JAZs, which recruit HDA6 as a corepressor to decrease EIN3/EIL1 transcriptional ability. ET treatment enhances EIN3/EIL1 protein stability (14, 16–18), while JA treatment leads to JAZs degradation and transcriptional derepression, providing a nice explanation for the hormonal synergy in ET/JA-regulated gene expression and developmental processes (Fig. S7). When ET signaling is blocked (e.g., in ein2), EIN3/EIL1 proteins are rapidly degraded (14, 18); therefore, the absence of EIN3/EIL1 accumulation fails to induce downstream gene expression and leads to ET/JA insensitivity. When JA signaling is blocked (e.g., in coi1), JAZs are highly accumulated (23, 26), which strongly repress EIN3/EIL1 function even in the presence of ET and JA.

The discovery of JAZ proteins has successfully linked JA receptors to downstream transcription factors. bHLH transcription factors (MYC2/MYC3/MYC4) and MYB transcription factors (MYB21/MYB24) are reported primary transcription factors associated with and repressed by JAZ protein (23, 30–32, 45). Our study has identified EIN3/EIL1 as a unique class of JAZ-targeted transcription factors. EIN3 and EIL1 physically interact with a number of JAZ proteins including JAZ1, JAZ3, and JAZ9, and JAZ proteins use the same module containing the Jas domain to interact with EIN3/EIL1 and MYC2 (26). Unlike MYC2/3/4 that substantially contribute to JA-mediated wound response and metabolism (30, 46), EIN3/EIL1 largely regulate JA-induced root hair development and resistance to necrotrophic fungi. We also examined the phehotypes of ein3 eil1 for other types of JA responses, such as fertility and anthocyanin biosynthesis, and found that EIN3/EIL1 have little role in these processes, which are more likely regulated by MYC2 or MYB transcription factors (32, 45–47). Consistently, JA was still able to robustly induce the expression of VSP2 in ein3 eil1, a MYC2-regulated gene invloved in wounding response (45). Interestingly, both EIN3/EIL1 and MYC2-mediated pathways participate in JA inhibition of root elongation, although EIN3/EIL1 seem to play a minor role in this growth response. Therefore, EIN3/EIL1 work separately and cooperatively with other JAZs-associated transcription factors to mediate a diverse range of JA responses. EIN3/EIL1 are predominant in JA regulation of fungal defense and root hair development, whereas MYC2-like transcription factors are indispensible for JA-mediated wounding response and metabolism, and MYB21/MYB24 function mainly in stamen development (32).

JAZ proteins have been identified as a group of respressors in JA signaling (23, 26, 27). We have presented several lines of evidence to demonstrate that JAZ proteins act to inhibit the function of EIN3/EIL1, and JA initiates a transcriptional derepression event by degrading JAZ proteins (Fig. 3). HDA6/19 had been implicated in the regulation of JA-responsive genes (40, 41), although it was not clear how histone deacetylation modulates JA signaling. Here, we reveal a previously unexplored mode of action for JAZ repressors, in which JAZ proteins recruit HDA6 (and probably other HDAs) to switch chromatin structure into a less accessible state. Treatment of a specific HDA inhibitor, TSA, was sufficient to drastically induce ERF1 expression and root hair formation, suggesting that histone deacetylation is a major repression mechanism of these JA-regulated processes. Consistent with this notion, the acetylation levels of histones (particularly H4) in the ERF1 promoter region were significantly increased by JA treatment, and such enhancement was more pronounced in the hda6 mutant, which were in good correlation with the expression levels of ERF1. Interestingly, we noted that the H4 acetylation level of the ERF1 promoter was greatly reduced in ein3 eil1, which could hardly be explained by the action of HDAs, implying the involvement of histone acetyltransferases in EIN3/EIL1-mediated gene activation. It thus remains to be investigated whether and how histone acetylation and deacetylation coordinately modulate JA/ET-regulated processes.

It was recently reported that JAZs interact with an adaptor protein NINJA that recruits Groucho/Tup1-type corepressor TOPLESS and its related proteins to suppress MYC2-dependent transcription (28). As TOPLESS-mediated repression has been implicated to link with histone deacetylation (33), it would be interesting to ascertain whether JAZs-NINJA-TOPLESS and JAZs-HDA6 repressors work as a large repressive module in modulating JA signaling. Alternatively, JAZ proteins might adopt distinct repression mechanisms to attenuate the function of their associated transcription factors in different JA responses. In support of this view, HDAs seem not involved in repressing MYC2 as TSA treatment does not induce VSP2 expression, which is a MYC2 responsive gene.

Transcriptional derepression has also been implicated in other signaling integration. Eliminating repressors (DELLA proteins) from transcriptional regulators (PIFs) has been reported in the coordinated regulation of plant growth and development by gibberellins and light (48, 49). Our work provides further evidence to highlight transcriptional derepression as a possibly common mechanism for the integration of multiple signaling pathways in plants. Given the existence of a number of repressors in plant hormone signaling (DELLA proteins for gibberellin, AUX/IAA proteins for auxin, JAZ proteins for JA, A-type Arabidopsis-response regulators for cytokinin) (50), further research should focus on the identification of repressor-associated transcription factors that might be integration points of plant hormones and other environmental or developmental signals.

Experimental Procedures

Root Hair Density.

We counted the entire root hair numbers from the root tip to the root-hypocotyl junction under an Olympus dissecting microscope, and divided this number by root length to obtain the value of root hair density. Each root hair observation was carried out for at least three times with the same result.

Yeast Two-Hybrid Assays.

The coding sequences of COI1, JAZ1, and HDA6 were amplified from wild-type cDNA and cloned into pGBKT7 vectors, and the coding sequences of EBF2, MYC2, HDA6, EIN3, EIN3 (200–500), EIN3 (1–300), EIN3 (200–400), EIN3 (300–500), and EIL1 (201–501) were cloned into pGADT7 vectors and transformed into yeast strain AH109 following the Matchmaker user's manual protocol (Clontech). Yeast transformants were streaked onto SD (-Leu/-His/-Ade/-Trp) medium (Clontech) and grown at 30 °C for 4 d. The white colonies represented interactions. Each yeast two-hybrid assay has been independently repeated at least twice with the same results. All primers used in this study are summarized in Table S1.

Protein Expression and Purification.

The coding sequences of JAZ1, JAZ1ΔN, and JAZ1ΔC were cloned into pGEX 5×-1 vectors (GE Healthcare) for GST fusion or pET-16b vectors (Novagen) for His fusion and transformed into BL21 (DE3)-competent cells. Protein expression was induced by 0.1 mM isopropyl-beta-d-thiogalactopyranoside and proteins were purified by glutathione Sepharose 4B (GE Healthcare) or Ni-NTA Agarose (Qiagen) following the manufacturer's instructions.