The transcriptional regulator BZR1 mediates trade-off between plant innate immunity and growth

1The trade-off between plant growth and immunity needs to be finely regulated to ensure proper allocation of resources in an efficient and timely manner upon effective integration of environmental cues. A key aspect of plant immunity is the perception of pathogen-associated molecular patterns (PAMPs) by surface-localized pattern-recognition receptors (PRRs), leading to PAMP-triggered immunity (PTI). PRRs of the leucine-rich repeat receptor kinases (LRR-RKs) class rely on the regulatory LRR-RK BAK1 (BRASSINOSTEROID INSENSITIVE 1-ASSOCIATED KINASE 1) for signaling; that is the case of FLS2 (FLAGELLIN SENSITIVE 2) and EFR (EF-TU RECEPTOR), which perceive bacterial flagellin (or the active peptide flg22) and EF-Tu (or the active peptide elf18) respectively. BAK1 also interacts with the LRR-RK BRI1 (BRASSINOSTEROID INSENSITIVE 1), the main receptor for the growth-promoting steroid hormones brassinosteroids (BR), and is a positive regulator of BR-mediated growth. Hence, a crosstalk between the BR- and PAMP-triggered signaling pathways resulting from competition for BAK1 was hypothesized. While a unidirectional antagonism between BR and PTI signaling has been recently described in Arabidopsis, the exact underlying mechanisms are still controversial. Activation of the BR signaling pathway via either transgenic overexpression of BRI1 or the BR biosynthetic gene DWF4 or expression of the activated BRI1 allele BRI1sud suppresses PTI outputs. One such output, the PAMP-triggered callose deposition, could be restored by over-expression of BAK1-HA, suggesting that BAK1 is a limiting factor. However, exogenous BR treatment of wild-type plants does not affect the FLS2-BAK1 complex formation upon FLS2 activation, while it results in decreased PTI responses.

2In order to clarify the role of BAK1 in the BR-PTI crosstalk, we investigated FLS2-BAK1 complex formation in the transgenic Arabidopsis lines overexpressing BRI1 or DWF4 or expressing BRI1sud. Upon treatment with flg22, FLS2 associated normally with BAK1 in these transgenic plants, and neither FLS2 nor BAK1 accumulation was altered (Figure 1—figure supplement 1A). Moreover, these plants displayed a weaker reactive oxygen species (ROS) burst in response to chitin (Figure 1—figure supplement 1B), whose signaling pathway is BAK1-independent. This result is consistent with the previous finding that exogenous BR treatment can also inhibit the chitin-induced ROS burst. BAK1-HA is not fully functional in BR signaling and exerts a dominant-negative effect on the endogenous BAK1 (Figure 1—figure supplement 1C), which may explain that introduction of the BAK1-HA transgene can override the suppression of immunity triggered by overexpression of BRI1; BAK1-HA does not exert such a dominant negative effect, however, on PTI signaling (Figure 1—figure supplement 1D). Taken together, these results indicate that the BR-mediated suppression of PTI is triggered independently of a competition between BRI1 and PRRs for BAK1.

3We sought to determine at which level of the BR signaling pathway the antagonism initiates. After BR perception by BRI1 and activation of the BRI1-BAK1 complex, the BR signal transduction cascade includes inactivation of BIN2 (BR INSENSITIVE 2) and BIN2-like kinases, a family of GSK3-like kinases acting as negative regulators of the pathway. This leads to dephosphorylation of BZR1 (BRASSINAZOLE RESISTANT 1) and BES1/BZR2 (BRI1-EMS-SUPPRESSOR 1/BRASSINAZOLE RESISTANT 2), two bHLH transcription factors acting as major regulators of BR-induced transcriptional changes, which then become active. Treatment with the chemicals LiCl and bikinin, which inhibit GSK3-like kinases, resulted in impaired flg22-triggered ROS burst (Figure 1A,B), as observed upon genetic or ligand-induced activation of the BR pathway. Furthermore, a triple mutant in BIN2 and the two closest related GSK3-like kinases, BIL1 (BIN2-LIKE 1) and BIL2 (triple GSK3 mutant; ), shows a similar impairment in response to either flg22 or chitin (Figure 1C). Interestingly, in spite of regulating MAPKs involved in stomata development, neither BR treatment nor loss of function of BIN2 affect flg22-triggered MAPK activation (Figure 1—figure supplement 2), contrary to what has been recently suggested. These results indicate that the BR-PTI crosstalk occurs downstream of BIN2.

4Transgenic expression of two different constitutively active versions of BZR1, BZR1Δ and BZR1S173A, results in impaired flg22- or chitin-triggered ROS burst (Figure 1D, Figure 1—figure supplement 3A). Consistent with previous results (; Figure 1—figure supplements 1A and 2), plants expressing BZR1Δ or BZR1S173A display normal FLS2-BAK1 complex formation and MAPK activation upon flg22 treatment (Figure 2A,B, Figure 2—figure supplement 1A,B), but are impaired in PAMP-triggered marker gene expression, seedling growth inhibition (SGI) (Figure 2C–E) and induced resistance to P. syringae pv. tomato (Pto) DC3000 (Figure 2F, Figure 2—figure supplement 1C), and are more susceptible to the non-host strain Pseudomonas syringae pv. cilantro (Pci) 0788-9 (Figure 2G). Notably, transgenic expression of a constitutively active form of BES1, BES1S171A, does not impact the flg22-triggered ROS burst (Figure 1—figure supplement 3B). We then tested if activation of BZR1 is sufficient to inhibit PTI signaling. Induction of BR signaling by bikinin treatment still represses elf18-induced ROS burst in the BRI1 mutant bri1-5 (we used elf18 in this experiment because bri1-5 is in the Ws-2 background, which is a natural fls2 mutant) (Figure 1—figure supplement 3C). bri1-5/BZR1Δ plants still exhibited reduced PAMP-triggered ROS burst (Figure 1E), and treatment with the BR biosynthetic inhibitor brassinazole (BRZ) did not affect the BZR1Δ effect (Figure 1—figure supplement 3D). Interestingly, BRZ treatment of wild-type Col-0 plants results in increased ROS production (Figure 1—figure supplement 3D), which is consistent with the fact that BR inhibits PTI responses and suggests that endogenous concentrations of the hormone exert this effect. These results demonstrate that activation of BZR1 affects PTI signaling independently of BR perception or synthesis.

5To understand how BZR1 mediates the BR-PTI crosstalk, we performed meta-analysis of microarray and ChIP-chip data containing BR-regulated and BZR1 or BES1 target genes. Functional enrichment of BR-regulated genes unveiled a statistically significant over-representation of defense-related GO terms of the Biological Process ontology (Table 1), indicating that BR signaling regulates the expression of defense-related genes. Independent analysis of BR-regulated BZR1 or BES1 targets confirmed BZR1 as the main transcription factor involved in the regulation of defense gene expression (Table 1). Two out of four over-represented GO terms of the Molecular Function ontology among the BR-regulated BZR1 targets are transcription factor and transcription repressor activity (Table 2). Interestingly, several defense-related GO terms are also over-represented in the subset of BR-regulated BZR1-targeted transcription factors (Table 3), pointing at a BZR1-mediated secondary transcriptional wave of defense-related genes.

6To identify BZR1-regulated transcription factors with a prominent role in defense, we performed promoter enrichment analysis on the subset of defense-related BR-regulated genes, and found the W-box motif as the only significantly over-represented motif (Table 4). The W-box motif is the binding site for the WRKY family of transcription factors, and several members of this family are BR-regulated BZR1-targets (Table 5). We hypothesized that WRKYs that are BR-induced and BZR1 targets may be involved in PTI signaling. Notably, wrky11, wrky15, wrky18 and wrky70 mutants displayed enhanced PAMP-triggered ROS (Figure 3A), suggesting that these transcription factors act as negative regulators of early PTI signaling. This is in accordance with their role as negative regulators of immunity (Figure 3—figure supplement 1A; ). Therefore, the BZR1-mediated inhibition of PTI might be partially explained by the up-regulation of genes encoding WRKY transcription factors that negatively control the expression of genes involved in early PTI signaling.

7One of the WRKY genes targeted by BZR1 is WRKY40. Interestingly, all described targets of WRKY40 are also targets of BZR1 (Table 6). The over-representation of the W-box motif among BZR1 targets (Table 7) suggests that BZR1 may interact with WRKY transcription factors (such as WRKY40) to cooperatively regulate transcription. WRKY40 has been described as a negative regulator of defense against biotrophic pathogens and insects. In agreement with this, we found that a null wrky40 mutant is more resistant to Pto DC3000 (Figure 3—figure supplement 1B). Strikingly, wrky40 plants are partially impaired in the BR-mediated suppression of PAMP-triggered ROS (Figure 3B), suggesting that WRKY40 may act coordinately with BZR1 to suppress immunity. Indeed, we found that BZR1 associates with WRKY40, but not WRKY6, in co-immunoprecipitation experiments when transiently co-expressed in Nicotiana benthamiana leaves (Figure 3C) or Arabidopsis protoplasts (Figure 3D). Collectively, these results indicate that BZR1 and WRKY40 form a protein complex that may participate in the transcriptional inhibition of PTI signaling.

8BZR1, together with DELLAs and PIF4, is part of a core transcription module that integrates hormonal (gibberellin [GA] and BR) and environmental (light) signals. In the dark, BZR1 is activated by endogenous BR and GA to promote growth, partially through the synergistic interaction with PIF4. Because etiolation requires rapid growth, we hypothesized that plants may prioritize growth over immunity in dark conditions. In fact, we found that PAMP-triggered SGI was partially impaired in dark-grown seedlings (Figure 4A–D). This impairment was abolished in the BR-insensitive mutants bri1-301 and bin2-1 (Figure 4A, Figure 4—figure supplement 2A), indicating that BR signaling is responsible for the dark-induced suppression of this PTI response. Activation of BZR1 in the BZR1Δ line mimicked the dark-induced suppression of SGI in both light and dark (Figure 4B). However, activation of BES1 in the BES1S171A line did not impact SGI (Figure 4—figure supplement 3A). Consistent with the previous results, exogenous BR treatment suppressed SGI in both light and dark (Figure 4C, Figure 4—figure supplement 2B,C). While treatment with GA alone did not have a dramatic effect on SGI, co-treatment with BL and GA resulted in an enhancement of the BR-mediated suppression of SGI (Figure 4C, Figure 4—figure supplement 2B), indicating an additive effect of these two hormones when applied together. Moreover, treatment with the GA synthesis inhibitors paclobutrazol (PAC) or uniconazole (Uni) abolished the effect of BL on SGI (Figure 4—figure supplement 1A,B and Figure 4—figure supplement 2B,C), and this effect was reduced in the GA biosynthetic mutant ga1-3 (Figure 4—figure supplement 1C). Taken together, these results demonstrate that BR suppress at least one PTI output, SGI, in the dark in a GA-dependent manner, most likely through activation of BZR1. Notably, although the wrky40 mutant undergoes etiolation normally (Figure 4—figure supplement 2D), it shows a diminished suppression of SGI in the dark (Figure 4D, Figure 4—figure supplement 2D), supporting the idea that WRKY40 is required for the BZR1-mediated inhibition of PTI.

9Previously, a unidirectional negative crosstalk between the growth-promoting hormone BR and PTI had been described. In this work, we show that activation of one of two major BR-activated transcription factors, BZR1, is sufficient to suppress PTI, measured as PAMP-triggered ROS production, PAMP-triggered gene expression, SGI or induced resistance (Figures 1 and 2, Figure 1—figure supplement 3, Figure 2—figure supplement 1). Of note, another PTI output, MAPK activation, is not affected by activation of the BR pathway (Figure 2B, Figure 1—figure supplement 2, Figure 2—figure supplement 1B). BR treatment results in BZR1-dependent changes in the expression of defense-related genes, among which several members of the WRKY family of transcription factors can be found. Because the promoters of BR-regulated defense-related genes are enriched in the W-box motif (Table 4), BZR1-targeted WRKY transcription factors could be responsible for a secondary wave of transcription, ultimately leading to the suppression of PTI. In agreement with this idea, a subset of WRKYs induced by BR (WRKY11, WRKY15 and WRKY18) (Figure 3A) act as negative regulators of PAMP-triggered ROS, potentially by controlling the steady-state expression of genes encoding components required for this response. The over-representation of the W-box motif among the BZR1 targets (Table 7) raises the possibility that, additionally, WRKY transcription factors could also act together with BZR1 to cooperatively regulate gene expression. We found that WRKY40 associates with BZR1 directly or indirectly in planta (Figure 3C,D); in the absence of WRKY40, the BR-mediated suppression of PAMP-triggered ROS burst is partially impaired (Figure 3B). Therefore, WRKYs may play a dual role in the BZR1-mediated suppression of defense, as both co- and secondary regulators of defense gene expression. Given that the loss of BR-mediated suppression of PAMP-triggered ROS burst in the wrky40 mutant is only partial, BZR1 may interact with other members of the WRKY family, such as WRKY18 or WRKY60, to repress immunity.

10Furthermore, we recently described that the bHLH transcription factor HBI1, which is a BRZ1 target, negatively regulates PTI (Malinovsky et al., under revision). All together, these results illustrate that BZR1 controls the expression of transcription factors (e.g. WRKY11, WRKY15, WRKY18 and HBI1), which themselves might control the expression of PTI components (see model in Figure 4E) whose identities remain to be identified.

11Plants need to finely regulate allocation of resources upon integration of environmental cues, both biotic and abiotic, in order to rapidly and readily adapt to changing conditions and ensure survival in a cost-efficient manner. Dark conditions impose an energetic limitation due to lack of photo-assimilates; in this situation, the restoration of normal photosynthesis by reaching light is an essential requirement to guarantee perpetuation, and as such must be given priority. We hypothesize that when plants face conditions that require rapid growth, such as when germinating in soil or when under a canopy, limited resources are invested in this developmental process at the expense of immunity in what must be a quantitative choice. Indeed, we show that etiolated seedlings do not arrest their growth in response to PAMPs as light-grown seedlings do, as measured by total fresh weight (Figure 4A–D). In addition, BR signaling, acting cooperatively with GA signaling, is required for the dark-induced suppression of this PTI response (Figure 4C, Figure 4—figure supplement 1), and activation of BZR1 is sufficient to exert this effect regardless of light conditions (Figure 4B). Although seedlings were used in these experiments due to technical reasons, BR also regulate growth at later developmental stages, so this phenomenon may be more general. Based on these findings, we propose a model in which BZR1 regulates the expression of defense genes, assisted by WRKY40 (and potentially other WRKYs), ultimately resulting in a quantitative suppression of immunity (Figure 4E). Because the activation status of BZR1 depends on BR, GA and light signaling, BZR1 would act as a molecular integrator of these inputs to effectively regulate the trade-off between growth and immunity.

12Col-0 plants were used as control. The transgenic lines BZR1Δ, bri1-5/BZR1Δ and BES1S171A, BZR1S173A and BZR1-CFP, 35S:BRI1-cit, BRI1p:BRI1sud-cit, 35S:DWF4 and BAK1-HA are published. The mutant lines Triple GSK3 mutant, bri1-5, bri1-301, bin2-1, wrky11, wrky18, wrky53 and wrky70 wrky27, wrky40 and ga1-3 are published. wrky15 mutant was identified in the ZIGIA population; wrky48 and wrky57 are from the SALK collection.

13Arabidopsis plants and seedlings were grown as described in.

14Flg22 and elf18 peptides were purchased from Peptron, and chitin oligosaccharide from Yaizu Suisankagaku. epiBL was purchased from Xiamen Topusing Chemical. LiCl, bikinin, brassinazole and GA were purchased from Sigma (St Louis, MO, USA). Paclobutrazol was purchased from Duchefa (Haarlem, NL). Uniconazole was purchased from Sigma.

15The measurement of ROS generation was performed as described in. Leaf discs from five-week-old Arabidopsis plants were used in each experiment, as indicated in the figure legends. Total photon counts were measured over 40 min by using a high-resolution photon counting system (HRPCS218) (Photek, St Leonards on Sea, UK) coupled to an aspherical wide lens (Sigma Imaging, Welwyn Garden City, UK).

16Protein extraction and immunoprecipitation of Arabidopsis was performed as described in. Arabidopsis mesophyll protoplasts were prepared from 4 to 5-week-old plants, transfected with the indicated constructs and incubated for 16 hr prior to BL treatment. Protein extraction of N. benthamiana was performed as described in; immunoprecipitations were performed using the μMACS GFP Isolation Kit (Miltenyi Biotec, Church Lane Bisley, UK), following the manufacturer’s instructions. In N. benthamiana, BZR1-GFP was expressed from the pUb-cYFP-Dest vector; WRKY40-HA and WRKY6-HA were expressed from the pAM-PAT vector (AY436765; GeneBank). In protoplasts, WRKY40-HA was expressed from the pGWB414 vector; the construct to express BZR1-GFP has been described elsewhere. In both cases, samples were treated with 1 μM epiBL solution for 1 hr prior to protein extraction.

17MAP kinase activation assays were performed as described in. Phospho-p44/42 MAPK (Erk1/2; Thr202/Tyr204) rabbit monoclonal antibodies (Cell Signaling Technologies, Hitchin, UK) were used according to the manufacturer’s protocol.

18RNA isolation was performed from ten-day-old seedling following the protocol described in. First-strand cDNA synthesis was performed with the SuperScript III RNA transcriptase (Invitrogen, Paisley, UK) and oligo(dT) primer, according to the manufacturer’s instructions. For qPCR reactions, the reaction mixture consisted of cDNA first-strand template, primers (5 nmol each) and SYBR Green JumpStart Taq ReadyMix (Sigma). qPCR was performed in a BioRad CFX96 real-time system. UBQ10 was used as the internal control; expression in mock-treated Col-0 seedlings was used as the calibrator, with the expression level set to one. Relative expression was determined using the comparative Ct method (2-ΔΔCt). Each data point is the mean value of three experimental replicate determinations. Primers for At2g17740 are described in; for NHL10 are described in; for LOX2 are described in; for UBQ10 (U-box) are described in.

19Seedling growth inhibition assays were performed as described in. In brief, four-day-old Arabidopsis seedlings were grown in liquid Murashige–Skoog medium containing 1% sucrose supplemented with flg22 and the appropriate chemicals. Seedlings were weighed between 6 and 10 days after treatment.

20Induced resistance assays were performed as described previously. In brief, water or a 1 μM flg22 solution were infiltrated with a needleless syringe into leaves of four-week-old Arabidopsis plants 24 hr prior to bacterial inoculation (Pto DC3000, 105 cfu/ml). Bacterial growth was determined 2 days after inoculation by plating serial dilutions of leaf extracts.

21Spray inoculation of P. syringae pv. cilantro (Pci) 0788-9 was performed as described in. In brief, bacteria were grown in an overnight culture in LB medium, cells were harvested by centrifugation, and pellets were re-suspended to OD600 = 0.02 in 10 mM MgCl2 with 0.04% Silwet L-77. Bacterial suspensions were sprayed onto leaf surfaces and plants were kept uncovered. Bacterial growth was determined 3 days after inoculation by plating serial dilutions of leaf extracts.

22Functional enrichment analyses of the Biological Process ontology were performed using VirtualPlant. Functional enrichment analysis of the Molecular Function ontology was performed using the Classification SuperViewer tool of the Bio-Array Resource for Arabidopsis Functional Genomics, BAR. Promoter analyses were performed using Athena.