Overexpression of the Receptor‐Like Kinase BIR1 Causes SOBIR1‐ and EDS1‐Dependent Cell Death Phenotypes in Arabidopsis

ABSTRACT

The receptor‐like kinase BAK1‐INTERACTING RECEPTOR‐LIKE KINASE 1 (BIR1) negatively regulates multiple resistance signalling pathways in Arabidopsis thaliana. Previous studies showed that loss of BIR1 function causes extensive cell death and constitutive activation of immune responses. Using a dexamethasone (DEX)‐inducible system, we investigated the effects of BIR1 overexpression on plant development and immunity. Overexpression of BIR1, in the absence of microbes or elicitors, led to cell death phenotypes that resembled the effects of BIR1 depletion in knockout plants. We also observed transcriptional outputs that greatly overlap with canonical pathogen‐triggered immunity and effector‐triggered immunity (ETI), suggesting that BIR1 modulates immune responses by influencing both pathways. To investigate the genetic basis of BIR1 phenotypes, we conditionally expressed BIR1 in various Arabidopsis immune mutants including sobir1, bak1, eds1, sid2 and eds5. We found that ENHANCED DISEASE SUSCEPTIBILITY 1 (EDS1) and SUPPRESSOR OF BIR1‐1 (SOBIR1) are necessary for ETI‐type cell death seen with BIR1 overexpression. These results support the hypothesis that an excess of BIR1 may be detected by guarding NLR proteins, triggering a cell death response in which SOBIR1 and EDS1 cooperate to transduce signals downstream of R proteins.

Summary Statement

Regulation of the receptor‐like kinase BIR1 has a strong impact on plant growth and development and immune homeostasis in Arabidopsis. BIR1 overexpression causes cell death‐ and senescence‐like phenotypes that require EDS1 and SOBIR1 signalling pathways, and that resemble those observed by BIR1 depletion.

1. Introduction

Plants defend against pathogens like bacteria, fungi, oomycetes and viruses through immunity (Kørner et al. 2013; Nishad et al. 2020). Pattern‐triggered immunity (PTI) involves cell‐surface pattern recognition receptors (PRRs) that detect pathogen/microbe/damage‐associated molecular patterns (PAMP/MAMP/DAMP). Effector‐triggered immunity (ETI) relies on intracellular receptors that sense pathogen effectors or host protein manipulations (Zhou and Zhang 2020; Yu et al. 2021). PRRs are single‐transmembrane receptor‐like kinases (RLKs) or receptor‐like proteins (RLPs). RLP‐type PRRs lack kinase activity and associate with the adaptor RLK SUPPRESSOR OF BIR1‐1 (SOBIR1) in a ligand‐independent manner to transduce signals (Gust et al. 2017). PRRs form complexes with somatic embryogenesis co‐receptor kinases (SERKs), including BRASSINOSTEROID INSENSITIVE1‐ASSOCIATED RECEPTOR KINASE 1 (BAK1/SERK3) (Boutrot and Zipfel 2017; Hohmann et al. 2017; Smakowska‐Luzan et al. 2018). BAK1 positively regulates PTI (Wu et al. 2020) and plays a dual role in regulating cell death during ETI (Belkhadir et al. 2012; Domínguez‐Ferreras et al. 2015). Upon pathogen detection, PRR complexes activate receptor‐like cytoplasmic kinases (RLCKs) to trigger immune signalling cascades (Liang and Zhou 2018; Rao et al. 2018). Among them, BOTRYTIS INDUCED KINASE1 (BIK1) acts as a positive regulator with RLKs and a negative regulator with RLPs (Bi et al. 2018; Wan et al. 2019).

ETI relies on nucleotide‐binding leucine‐rich domain LRR (NLR) receptors encoded by polymorphic disease resistance genes. NLRs include TOLL/INTERLEUKIN RECEPTOR 1 (TIR1) (TIR‐NB‐LRR, TNL), proteins with a coiled‐coil (CC) domain at the N terminus (CC‐NB‐LRR, CNL) or proteins with a RESISTANCE TO POWDERY MILDEW 8‐LIKE DOMAIN (RPW8)‐type domain (RNLs) (van Wersch et al. 2020; Wang and Chai 2020). NLRs act as ‘sensors’ that detect effectors or as ‘helpers’ mediating downstream signalling (Jubic et al. 2019). Recent studies show that PTI and ETI work together to enhance immune responses (Yuan, Ngou, et al. 2021; Yuan et al. 2023). NLR signalling depends on PTI components like BAK1, SOBIR1 and BIK1, independently of PTI, while TNL genes upregulated during PTI amplify immunity (Ngou et al. 2021; Tian et al. 2021; Yuan, Jiang, et al. 2021). ENHANCED DISEASE SUSCEPTIBILITY (EDS1) and PHYTOALEXIN DEFICIENT 4 (PAD4), essential for TNL‐mediated immunity, also play key roles in PTI signalling (Pruitt et al. 2021; Tian et al. 2021). PTI and ETI outputs typically converge, though with different intensities and durations (Yuan, Ngou, et al. 2021). Early immune responses include reactive oxygen species (ROS) bursts, calcium influx, Ca⁺²‐dependent protein kinases (CPK) activation and mitogen‐activated protein kinases signalling, followed by transcriptional activation of defence processes, including phytohormone accumulation and cell wall alterations (Li et al. 2020).

Plants use various mechanisms to prevent the harmful effects of hyperactivated defences (Couto and Zipfel 2016; Withers and Dong 2017; Mithoe and Menke 2018). BAK1‐INTERACTING RECEPTOR‐LIKE KINASE (BIR) proteins are crucial for defence attenuation. The current model suggests that in the absence of microbes, BIR proteins bind constitutively to BAK1, keeping PRRs in a resting state. Upon ligand binding, BAK1 dissociates from BIR to interact with PRRs (Gao et al. 2009; Halter et al. 2014; Imkampe et al. 2017). Arabidopsis has four BIR homologues, of which BIR1 negatively regulates cell death pathways involving BAK1, SOBIR1, EDS1 and PAD4, likely linked to TNL‐mediated ETI (Gao et al. 2009). BIR2 and BIR3 suppress PRR‐mediated PTI responses (Halter et al. 2014; Imkampe et al. 2017), while the TNL protein CONSTITUTIVE SHADE AVOIDANCE 1 (CSA1) interacts with BIR3 to sense disturbances in the BIR3/BAK1 complex, triggering cell death (Schulze et al. 2022).

BIR1 expression is activated by microbial and viral pathogens in a salicylic acid (SA)‐ and jasmonic acid (JA)‐dependent manner (Guzmán‐Benito et al. 2019; Robinson et al. 2025). Under normal conditions, BIR1 transcription is negatively regulated by RNA‐directed DNA methylation, with post‐transcriptional silencing reinforcing this epigenetic regulation when BIR1 is activated (Guzmán‐Benito et al. 2019). Overexpression of BIR1 results in senescence and cell death phenotypes in Arabidopsis (Guzmán‐Benito et al. 2019). While the autoimmune response in bir1‐1 knockout mutants is well‐characterized (Gao et al. 2009; Liu et al. 2016), the molecular basis of BIR1 overexpression phenotypes remains unclear. Using a dexamethasone (DEX)‐inducible system, we investigated the immune pathways involved in BIR1 overexpression in the absence of microbes or elicitors. We found that senescence and cell death phenotypes in BIR1 overexpressing lines coincide with improper expression of genes encoding immune receptors and co‐receptors involved in the perception and downstream signalling of ligands and effectors. Genetic depletion of key immune regulators revealed that TNL‐associated EDS1 and/or the RLP‐co‐receptor SOBIR1 are required for these cell death phenotypes.

2. Methods

2.1. Plant Material and Growth Conditions

Arabidopsis thaliana plants were grown in controlled chambers under long‐day conditions (16 h day/8 h night) at 22°C. The bak1‐5, sobir1‐12, eds1‐2, eds5‐3 and sid2‐2 mutants and the BIR1‐overexpression lines L6 and L9 under a DEX‐inducible system were as described previously (Jirage et al. 1999; Nawrath and Metraux 1999; Bartsch et al. 2006; Gao et al. 2009; Schwessinger et al. 2011; Guzmán‐Benito et al. 2019). The same system expressing BIR1‐mCherry fusion was used to transform mutant backgrounds via floral dipping (Clough and Bent 1998). Independent homozygous lines were selected in T3 for further experiments. Morphological phenotypes after DEX treatments were examined in at least two independent transgenic lines per genotype.

2.2. RNA and Protein Analyses

Total RNA was extracted using TRIzol reagent (Invitrogen) and treated with DNase I (Invitrogen) to remove genomic DNA. One‐step quantitative RT‐PCR (RT‐qPCR) was performed using Brilliant III Ultra‐Fast SYBR Green QRT‐PCR Master Mix (Agilent Technologies) in a Rotor‐Gene 6000/Rotor‐Gene Q real‐time PCR machine (Corbett/Qiagen). Relative gene expression was determined using the Delta‐delta cycle threshold method and Rotor‐Gene 6000 Series Software (Corbett). Transcript levels of the target genes were normalized to the transcript levels of the housekeeping gene CBP20 (At5g44200) as a reference gene. Gene‐specific primers are given in Supporting Information S2: Table S1.

For protein analysis, leaf tissue was ground in liquid nitrogen and homogenized in extraction buffer (65 mM Tris‐HCl, pH 8; 3% SDS; 1% ß‐mercaptoethanol; 10% glycerol). Samples were diluted in Laemmli buffer, heated at 95°C for 5 min, and loaded onto 10% SDS‐PAGE gels. After electrophoresis, proteins were transferred to P0.45 Polyvinylidine Fluoride (PVDF) blotting membranes (Amersham) and detected with HRP‐conjugated secondary antibodies and chemiluminescent substrate (LiteAblot Plus).

2.3. Library Preparation for Transcriptome Sequencing and Bioinformatic Analyses

Total RNA was extracted using the RNeasy Plant Mini Kit (Qiagen) and treated with DNase according to the manufacturer's instructions. RNA from six individual plants was pooled per sample. RNA purity and concentration were assessed using a Nanodrop spectrophotometer (ThermoFisher Scientific), and integrity was measured with the Agilent Bioanalyzer 2100 (Agilent Technologies). Sequencing libraries were prepared at NOVOGENE (HK) using the NEBNext Ultra RNA Library Prep Kit for Illumina (New England Biolabs), following the manufacturer's guidelines, and index codes were added to attribute sequences to each sample. First‐strand cDNA was synthesized from mRNA using random hexamer primers and M‐MuLV Reverse Transcriptase (RNase H‐). Second‐strand cDNA synthesis was performed with DNA Polymerase I and RNase H. NEBNext adaptors with a hairpin loop structure were ligated, and library fragments of 150–200 bp were selected using the AMPure XP system (Beckman Coulter). PCR amplification was done using Phusion High‐Fidelity DNA polymerase, Universal PCR primers and Index (X) Primer. The PCR products were purified, and library quality was assessed on the Agilent Bioanalyzer 2100. Clustering was performed on a cBot Cluster Generation System using the TruSeq PE Cluster Kit v3‐cBot‐HS (Illumina). Paired‐end reads were generated via the Illumina PE150 Hiseq platform at NOVOGENE (HK). Clean reads were obtained by removing adaptor tags, poly‐N and low‐quality reads. Paired‐end clean reads were mapped to the Arabidopsis TAIR v10 reference genome using HISAT2. Gene expression levels were quantified using HTSeq v0.6.1, and the abundance of transcripts from different samples was normalized using rlog and FPKM (Fragments Per Kilobase of transcript per Million mapped reads). Differential expression analysis was performed with DESeq. 2 (1.18.0), and p‐values were adjusted using the Benjamini‐Hochberg method for False Discovery Rate (FDR). Genes with an adjusted p‐value ≤ 0.05 were considered differentially expressed genes (DEGs). Functional classification of DEGs, including Gene Ontology (GO) and Kyoto Encyclopaedia of Genes and Genomes (KEGG) pathway analysis, was performed. GO enrichment was analyzed using the tGOseq R package, with gene length bias corrected. KOBAS software tested the statistical enrichment of DEGs in KEGG pathways and GO terms, with corrected p‐value ≤ 0.05 considered significant.

Public databases were used to compare transcriptomic data from BIR1 overexpression with other PTI‐ and ETI‐related responses, including infection with Pseudomonas syringae pv. tomato (Pst) DC3000 empty vector (EV) and Pst DC3000 AvrRps4 (GSE225073) (Jacob et al. 2023), flg22 treatment (GSE246173) (Mooney et al. 2024), elf18 treatment (GSE40354) (Tintor et al. 2013), DEX‐induced expression of MILDEW RESISTANCE LOCUS A coiled‐coil (MLAcc) (GSE80422) (Jacob et al. 2018) and CONSTITUTIVE EXPRESSER OF PR GENES 5 (CPR5) (GSE72742) (Gu et al. 2016). Data from DEX‐treated BIR1 L9 plants after 4 days served as a negative control. The 1000 DEGs between DeL9 and DeWT conditions with the lowest adjusted p‐values were used to construct a matrix of log₂ fold change values across six data sets. Gene expression heatmaps were generated using the ComplexHeatmap package (v2.14.0) in R (v4.2.0) with the Euclidean distance method (Gu 2022). GO enrichment analysis was performed with the clusterProfiler package (v4.6.2), using org.AT.tair.db (v3.16.0) for annotation and all genes in the database as the background (Wu et al. 2021). Multiple testing correction was applied using the Benjamini‐Hochberg method (adjusted p‐value < 0.05). Redundancy was reduced with the Wang method (threshold 0.7). The raw data were deposited in the Gene Expression Omnibus under accession number GSE234036.

2.4. DEX Treatments and Sampling Procedure

Arabidopsis plants, grown in soil for 3–4 weeks, were sprayed with 30 µM DEX (Sigma) every 24 h for optimal transgene induction as described (Guzmán‐Benito et al. 2019). Samples were collected at the onset of symptoms, typically around 12 days after DEX treatment. However, the timing may vary between genotypes and treatments, meaning plants of different genotypes may be harvested at different developmental stages. Therefore, quantitative comparisons of gene expression across genotypes should be made with caution.

3. Results

3.1. BIR1 Overexpression Causes Transcriptome Reprogramming in Arabidopsis

We conducted a transcriptomic analysis of Arabidopsis BIR1 L9 plants, which express a DEX‐inducible BIR1‐mCherry transgene, to explore the biological processes and pathways associated with BIR1 overexpression under pathogen‐free conditions. After DEX treatment, BIR1 L9 plants exhibited growth defects, including stunting, leaf deformation, premature senescence and cell death starting around Day 12 (Figure 1a) (Guzmán‐Benito et al. 2019). At this time, mCherry‐tagged BIR1 protein levels were markedly higher than after 4 days of DEX treatment, indicating gradual protein accumulation (Figure 1b). In the WT genotype (without the BIR1 cassette), DEX treatments caused a slight growth delay when applied to young plants (Boyes stage 1.5) but had no effects in later developmental stages (Boyes stage 1.7) (Figure 1a and Supporting Information S1: Figure S1a,b). Cell death or senescence phenotypes were not observed in DEX‐treated WT (Figure 1a and Supporting Information S1: Figure S1a). Consistent with previous results, phenotypes in the BIR1‐mCherry inducible system were similar to those observed when a BIR1‐HA construct was overexpressed using a tobacco rattle virus (TRV)‐based expression system (Guzmán‐Benito et al. 2019), confirming that cell death was due to BIR1 overexpression, and not to DEX effects or protein tagging (Figure 1a).

Figure 1.

Transcriptional reprogramming associated with BIR1 overexpression in Arabidopsis. (a) Morphology of dexamethasone (DEX)‐inducible BIR1‐mCherry transgenic Arabidopsis (L9) after 10–12 days of 30 µM DEX treatment. Spraying began at Boyes' growth stage 1.7. Phenotypes in DEX‐treated wild‐type (WT) plants and plants expressing an HA‐tagged version of BIR1 via TRV infection (TRV‐BIR1) are shown. (b) Western blot showing relative accumulation of BIR1‐mCherry protein in BIR1 L9 plants after 4 and 12 days of DEX treatment. Ponceau S was used as a protein loading control. (c) Relative expression of BIR1 transcripts based on normalized r log counts (left) and fragments per kilobase of transcript sequence per million base‐pairs sequenced (FPKM) (right) determined by RNA‐seq in mock‐ and DEX‐treated WT and BIR1 L9 plants. (d) Number of up‐ or downregulated transcripts with |log₂ (Fold‐Change)| ≥ 1 and adjusted p‐value ≤ 0.01 in each pairwise comparison. (e) Principal component analysis (PCA) of RNA‐Seq data from differentially expressed genes (DEGs) under various conditions: Mock‐treated WT (Mock‐WT), Mock‐treated BIR1 L9 (Mock‐L9), DEX‐treated WT (DEX‐WT) and DEX‐treated BIR1 L9 (DEX‐L9). Samples were collected after 4 and 12 days of DEX treatment. Biological replicates are shown in the score plot, with variance explained by each component (%) indicated.

For RNA‐seq, 3‐week‐old plants were treated with DEX for 12 days, and control plants were mock‐treated with H₂O. Three biological replicates were used for each condition: DEX‐treated WT (DeWT_), mock‐treated WT (MoWT_), DEX‐treated BIR1 L9 (DeL9_) and mock‐treated BIR1 L9 (MoL9_) (Supporting Information S3: Table S2). To avoid transcriptome damage, samples were collected before cell death symptoms appeared. RNA‐seq data revealed high expression levels of the BIR1 transgene in DEX‐treated plants after 12 days (Figure 1c, Supporting Information S3: Table S2). BIR1 overexpression induced significant transcriptional reprogramming in Arabidopsis (Figure 1d), with DEGs summarized in Supporting Information S4: Table S3. DEGs with various statistical thresholds are shown in Supporting Information S1: Figure S2a,b and Supporting Information S5: Table S4. Principal component analysis (PCA) revealed that DEX‐treated BIR1 L9 samples (DeL9_12) were distinct from other conditions, with the BIR1 expression level explaining 28.37% of the variance (Figure 1e). After 12 days of treatments, WT (mock‐ and DEX‐treated) and mock‐treated BIR1 L9 samples clustered in the proximity, indicating similar gene expression profiles (Figure 1e). PC2 highlighted a shift in gene expression due to plant age or developmental stage (16.83%). The impact of BIR1 induction on gene expression was minimal after 4 days of DEX treatment, as samples from both genotypes were close to each other in the PCA plot (Figure 1e).

GO analysis revealed DEGs with kinase activity enriched only in BIR1 overexpressing plants compared to mock‐treated BIR1 L9 or DEX‐treated WT plants (Supporting Information S1: Figures S1c,d and S2c) (Supporting Information S6: Table S5). The KEGG pathway analysis highlighted the enrichment of DEGs related to plant hormone signal transduction only in DEX‐treated BIR1 L9 plants, but not in mock‐treated BIR1 L9 plants or DEX‐treated WT plants (Supporting Information S1: Figures S1d and S3, Supporting Information S7: Table S6).

3.2. BIR1 Overexpression Triggers the Expression of Immunity‐Related Genes

RNA‐Seq results were validated in mock‐ and DEX‐treated BIR1 L9 plants after 12 days using RT‐qPCR. Expression of DEGs encoding BIR1, PR1, RESPIRATORY BURST OXIDASE HOMOLOG PROTEIN D (RBOHD), FRK1 and CYTOCHROME P450 MONOOXYGENASE 79B2 (CYP79B2) was upregulated in BIR1‐overexpressing plants compared to mock‐treated controls, confirming the RNA‐Seq profiles and reproducibility of these expression patterns (Figure 2a). Consistent with transcript levels, Western blot analysis showed enhanced accumulation of SOBIR1, PR1 and EDS1 proteins in DEX‐treated BIR1 L9 plants compared to controls (Supporting Information S1: Figure S4).

Figure 2.

Differential expression of immune‐related genes in BIR1 overexpressing plants. (a) RT‐qPCR analysis of selected BIR1‐responsive differentially expressed genes (DEGs) based on fold‐change values (log₂ DeL9_12/DeWT_12 and log₂ DeL9_12/MoL9_12). Values were normalized to the CBP20 internal control and related to untreated (DEX‐) plants, set to 1. Data are mean ± SD, analyzed by Mann–Whitney test; *p < 0.05. (b–f) Cluster Heatmap representation of genes overrepresented in the pool of differentially expressed genes (DEGs) encoding cysteine‐rich and L‐type lectin‐domain containing receptor‐like kinases (CRK, LECRK) (b), receptor‐like proteins (RLP) (c), disease resistance proteins (d), PBS1‐LIKE kinases (PBL) (e) and Ca⁺²‐dependent protein kinases (CPK). Each row represents a DEG, and each column represents an experimental group. The colour scale represents z‐scores of the median normalized count ratios across conditions. Dendrograms were generated using Euclidean distance and complete linkage hierarchical clustering for both rows and columns. Gene expression levels are coded, with blue and red representing low and high‐expressed genes, respectively.

RNA‐Seq analysis revealed that overexpression of BIR1 misregulated multiple genes involved in pathogen perception and defence activation, a pattern not observed in DEX‐treated WT or mock‐treated BIR1 L9 plants (Supporting Information S5: Table S4). For instance, genes encoding the NADPH oxidases RBOHD and RBOHF involved in ROS production, the co‐receptors BAK1 and SOBIR1 or the heterotrimeric G proteins ARABIDOPSIS G PROTEIN α‐SUBUNIT 1 (GPA1), β‐SUBUNIT 1 (AGB1), γ‐SUBUNIT 2 (AGG2) and EXTRA‐LARGE G PROTEINS 1 and 2 (XLG1/2) were more abundant in BIR1‐overexpressing plants compared to controls after 12 days of DEX treatment (Supporting Information S5: Table S4). Two‐tailed Fisher's exact tests were used to identify group of genes overrepresented in the set of DEGs (Supporting Information S8: Table S7 and Supporting Information S9: Table S8) and cluster heatmaps to illustrate changes in the expression of overrepresented DEGs (both induced or repressed) between our four experimental groups. Thus, we observed a significant enrichment of LRR‐type RLKs, particularly cysteine‐rich RLKs (CRKs) and L‐type lectin‐domain CRKs (LECRKs) (Figure 2b). BIR1 overexpression significantly increased the transcription of RLP‐type PRRs (Figure 2c) and genes encoding RLCKs like BIK1 and several AvrPphB SUSCEPTIBLE1 (PBS1)‐LIKE kinases (PBL) (Figure 2e). CPKs were not collectively regulated, but many individual CPK‐coding genes, including CPK4‐6 and CPK11, were DEGs in BIR1‐overexpressing plants (Figure 2f). A significant number of DEGs (induced or repressed) were annotated as disease resistance protein, particularly TNL proteins (Figure 2d). This finding is consistent with the induction in DEX‐treated BIR1 L9 plants, but not in controls, of several genes important for PTI and TNL‐receptor function in ETI, such as those encoding EDS1, PAD4, SENESCENCE ASSOCIATED GENE 101 (SAG101) and ACTIVATED DISEASE RESISTANCE 1 (ADR1). Conversely, neither genes encoding NON‐RACE SPECIFIC DISEASE RESISTANCE 1 (NDR1) nor NDR1‐dependent CNL intracellular receptors were differentially expressed under BIR1 overexpression (Supporting Information S4: Table S3).

In conclusion, BIR1 overexpression disrupts immune homeostasis under pathogen‐free conditions, leading to the overexpression of multiple genes involved in immune activation and signalling. We propose that the cell death‐like phenotypes associated with BIR1 overexpression stem from the intrinsic deregulation of multiple defence and resistance pathways.

3.3. The Transcriptional Response of BIR1 Overexpressing Plants Overlaps With PTI and TNL‐Dependent ETI

We next examined publicly available Arabidopsis microarray and RNA‐seq data sets to compare the expression profile of the 1000 most significant DEGs in BIR1‐overexpressing plants with the transcriptional changes elicited under several defence‐induced conditions. We observed a significant overlap and a strong positive correlation between the transcriptional responses to BIR1 overexpression and infection with either virulent Pst DC3000 EV or avirulent Pst DC3000 expressing AvrRps4, which co‐activates PTI and ETI responses via the TNL protein RPS4 (Figure 3a,b) (Jacob et al. 2023). In both cases, GO analysis of overlapping DEGs revealed the upregulation of genes involved in PTI and ETI‐related processes such as responses to external stimuli, systemic acquired resistance, response to ROS, senescence, cell death, immune response, response to hypoxia or defence hormone responses (Figure 3c). The BIR1‐mediated transcriptome also showed considerable overlap and strong correlation with transcriptional outputs in plants exposed to flg22 peptide derived from the flagellin protein (Mooney et al. 2024) or elf18 peptide derived from elongation factor Tu (EF‐Tu) (Tintor et al. 2013), which activate basal PRR‐mediated PTI in pathogen‐free conditions (Figure 3a,b). These genes were similarly enriched in GO terms related to the defence responses mentioned above (Figure 3c) (Supporting Information S3: Table S9). No positive correlation was found with genes deregulated in DEX‐treated BIR1 L9 transgenic plants after 4 days, which served as a negative control in our analysis (Figure 3a,b). Furthermore, while overexpression of BIR1 and CPR5, which activate ETI, programmed cell death and defence gene expression (Gu et al. 2016), shared fewer common DEGs, they exhibited the strongest positive correlation, suggesting similar transcriptomic responses (Figure 3a,b). In contrast, little resemblance was found with the transcriptome linked to DEX‐inducible overexpression of the coiled‐coil (CC) domain of MLAcc, which encodes a CNL recognizing effectors of Blumeria graminis f. sp. hordei isolate K1 (Bgh K1) (Jacob et al. 2018). In conclusion, our comparative analysis indicates that BIR1 overexpression induced transcriptional responses that closely resemble the activation of PRR‐mediated PTI and TNL‐mediated ETI pathways, each associated with the activation of an overlapping set of defence genes. Interestingly, although the correlation is high in all cases, the overlap in the PTI transcriptional responses appears to be greater in the presence of infection than in a pathogen‐free system, where microbial elicitors directly trigger the response.

Figure 3.

Transcriptome‐wide comparison of differentially expressed genes (DEGs) associated with BIR1 overexpression and transcriptional responses related to pattern‐triggered immunity (PTI) and effector‐triggered immunity (ETI). (a) Cluster Heatmap showing log₂ fold‐change (LFC) gene expression values for the 1000 DEGs in BIR1‐overexpressing plants common across all the conditions tested. BIR1 overexpression was compared with gene expression patterns triggered by infection with virulent Pseudomonas syringae pv. tomato (Pst) DC3000 and avirulent Pst DC3000 AvrRps4, treatment with PTI elicitors flg22 or elf18, and activation of the coiled‐coil nucleotide‐binding leucine‐rich domain receptor (CNL) MILDEW RESISTANCE LOCUS A coiled‐coil (MLAcc) and CONSTITUTIVE EXPRESSER OF PR GENES 5 (CPR5). (b) Gene ontology (GO) terms (Biological Process) enriched in common up‐ and downregulated (or both) DEGs across all conditions used for comparison. N.E., not enriched. (c) Pairwise global similarity of expression patterns across conditions analyzed using Spearman's correlation based on log₂ fold‐change values of all commonly DEGs.

3.4. Disruption of SOBIR1, but not BAK1, Partially Rescues BIR1 Overexpression Phenotypes

To investigate the genetic basis of phenotypes associated with BIR1 overexpression, we examined whether mutations in key PTI and ETI signalling components could alleviate cell death‐like morphologies and immune gene upregulation. The DEX‐inducible BIR1‐mCherry cassette was transformed into various mutant backgrounds, with two independent BIR1‐mCherry transgenic lines analyzed for each genotype to minimize insertion position effects. The phenotypes resulting from BIR1 overexpression were consistently reproduced across both independent transgenic lines and experiments. None of the observed morphological and growth defects in BIR1 L9 plants were present in mutant controls (without the BIR1‐mCherry construct) or transgenic lines with only mCherry after DEX treatment (Supporting Information S1: Figure S5).

SOBIR1 and BAK1 are essential for activating cell death and defence responses in the bir1‐1 mutant (Gao et al. 2009; Liu et al. 2016). To determine if they are also required for BIR1 overexpression phenotypes, we transformed the DEX‐inducible BIR1‐mCherry cassette into the sobir1‐12 and bak1‐5 mutants. DEX‐treated sobir1‐12 BIR1 and bak1‐5 BIR1 plants displayed WT‐like appearance and accumulated similar levels of BIR1‐mCherry protein as BIR1 L9 plants after 7 days, confirming comparable BIR1 expression (Figure 4a). Then, morphological phenotypes were registered after approximately 12 days of DEX treatment. At this time, BIR1 transcript levels remained elevated in both genotypes (Figure 4b and Supporting Information S1: Figure S6c). Consistent with previous data, DEX‐treated BIR1 L9 plants showed severe stunting, accelerated senescence and cell death (Figure 4c). Overexpression of BIR1 in sobir1‐12 mutants resulted in milder phenotypes, suggesting partial suppression of the morphological defects by the sobir1‐12 mutation (Figure 4c and Supporting Information S1: Figure S6a,b). In contrast to what was observed in BIR1 L9 plants, defence markers such as RBOHD, FRK1 and WRKY DNA‐BINDING PROTEIN 29 (WRKY29) showed no significant induction in DEX‐treated sobir1‐12 BIR1 plants overexpressing BIR1 (Figure 4d), suggesting that BIR1 overexpression activates immune pathways that depend on SOBIR1. In contrast, bak1‐5 BIR1 plants exhibited severe morphological defects and yellowing in BIR1 overexpressing plants (Figure 4b and Supporting Information S1: Figure S6a,b), as well as elevated RBOHD, FRK1 and WRKY gene expression (Figure 4e). These findings suggest that SOBIR1, but not BAK1, contributes to the morphological phenotypes and defence gene activation during BIR1 overexpression.

Figure 4.

Phenotypes caused by BIR1 overexpression are partially rescued in sobir1‐12 mutants. (a) Western blot analysis of BIR1‐mCherry proteins in dexamethasone (DEX)‐inducible BIR1‐mCherry transgenic Arabidopsis lines: wild‐type (WT) BIR1 L9, sobir1‐12 BIR1 (L6) and bak1‐5 BIR1 (L10) after 7 days of DEX treatment. Each sample is a pool of two plants, with two samples per genotype shown. Ponceau S was used as a protein loading control. A composite image of the same blot is presented. (b) RT‐qPCR analysis of BIR1 transcripts in WT BIR1 L9, sobir1‐12 BIR1 L6 and bak1‐5 BIR1 L10 plants. (c) Morphology of representative WT BIR1 L9, sobir1‐12 BIR1 L6 and bak1‐5 BIR1 L10 plants. Images from independent transgenic lines are shown in Supporting Information S1: Figure S6. (d and e) RT‐qPCR analysis of defence genes RBOHD, FRK1 and WRKY29 in sobir1‐12 BIR1 L6 (d) and bak1‐5 BIR1 L10 (e) plants. Samples in (b)–(d) were analyzed after 12 days of mock (DEX−) or 30 µM DEX treatment (DEX+). qRT‐PCR data were normalized to the CBP20 internal control and related to mock‐treated plants (set to 1). Data are mean ± SD, analyzed by Mann–Whitney test; *p < 0.05. Experiments were repeated at least twice with similar results.

3.5. The Absence of EDS1 Partially Rescues BIR1 Overexpression Phenotypes

A previous study found that stunting, cell death and defence responses in the bir1‐1 mutant are partially dependent on EDS1 (Gao et al. 2009). Since TNL‐coding transcripts were significantly deregulated in BIR1‐overexpressing plants, we speculated that EDS1‐dependent TNL‐mediated resistance responses might be altered by BIR1 overexpression. To investigate whether BIR1 overexpression phenotypes depend on EDS1, we generated DEX‐inducible BIR1‐mCherry transgenic lines in eds1‐2. After 7 days of DEX treatment, eds1‐2 BIR1 and BIR1 plants displayed a WT‐like phenotype and accumulated similar levels of BIR1‐mCherry protein (Figure 5a). After 12 days, eds1‐2 BIR1 plants accumulated high BIR1 transcript levels (Figure 5b) but, unlike BIR1 L9 plants, maintained a largely WT‐like appearance with occasional yellowing, suggesting partial rescue of the BIR1 overexpression phenotypes (Figure 5c and Supporting Information S1: Figure S6a). FRK1, RBOHD and WRKY29 transcript levels were elevated in these eds1‐2 BIR1 plants compared to non‐treated controls (Figure 5b,d), indicating that loss of EDS1 attenuated BIR1 overexpression phenotypes without significantly affecting defence expression.

Figure 5.

Phenotypes caused by BIR1 overexpression are partially rescued in eds1‐2 mutants. (a) Western blot analysis of BIR1‐mCherry proteins in dexamethasone (DEX)‐inducible BIR1‐mCherry transgenic Arabidopsis lines: wild‐type (WT) BIR1 L9 and eds1‐2 BIR1 L1b, treated with DEX for 7 days. Each sample is a pool of two plants, with three samples per genotype. Ponceau S was used as a protein loading control. A composite image of the same blot is presented. (b) RT‐qPCR analysis of BIR1 transcripts in eds1‐2 BIR1 L1b plants. (c) Morphology of representative eds1‐2 BIR1 L1b plants. Images from independent transgenic lines are shown in Supporting Information S1: Figure S6. (d) RT‐qPCR analysis of defence genes RBOHD, FRK1 and WRKY29 in eds1‐2 BIR1 L1b plants. Samples were analyzed after 12 days of mock (DEX−) or 30 µM DEX treatment (DEX+). qRT‐PCR data were normalized to the CBP20 internal control and compared to mock‐treated plants (set to 1). Data are mean ± SD, analyzed by Mann–Whitney test; *p < 0.05. Experiments were repeated at least twice with similar results.

3.6. Phenotypes of BIR1 Overexpressing Lines Are Unrelated to SA Signalling Pathways

Previous studies indicated that high SA levels contribute to bir1‐1 phenotypes (Gao et al. 2009). To investigate the role of SA in BIR1 overexpression phenotypes, we used Arabidopsis mutants deficient in SA biosynthetic enzymes SALICYLIC ACID INDUCTION DEFICIENT 2 (SID2)/ISOCHORISMATE SYNTHASE 1 (ICS1) and EDS5, to assess their effects on cell death. We generated DEX‐inducible BIR1‐mCherry transgenic plants in sid2‐2 and eds5‐3 mutants. The eds5‐3 BIR1 plants accumulated BIR1‐mCherry protein similar to BIR1 L9 controls, while sid2‐2 BIR1 plants accumulated less protein after 7 days of DEX treatment (Figure 6a). However, both sid2‐2 BIR1 and eds5‐3 BIR1 plants exhibited severe phenotypes and high BIR1 transcript levels after 12 days of DEX treatment (Figure 6b,c and Supporting Information S1: Figure S6a). This suggests that the BIR1 overexpression phenotypes are not dependent on SA levels. Interestingly, transcripts of PR1, a canonical SA‐related defence marker, were increased in sid2‐2 BIR1 and, more significantly, in eds5‐3 BIR1 plants (Figure 6d), suggesting that SA‐independent mechanisms contribute to PR1 expression in BIR1 overexpressing plants. Although PR1 upregulation coincides with BIR1 overexpression, some responsiveness to DEX was also observed in WT plants under our conditions (Figure 6d).

Figure 6.

BIR1 overexpression phenotypes are independent of salicylic acid (SA)‐mediated signalling pathways. (a) Western blot analysis of BIR1‐mCherry proteins in dexamethasone (DEX)‐inducible BIR1‐mCherry transgenic Arabidopsis lines: wild‐type (WT) BIR1 L9, sid2‐2 BIR1 L5 and eds5‐3 BIR1 L10, treated with DEX for 7 days. Each sample is a pool of two plants, with three samples per genotype. Ponceau S was used as a protein loading control. (b) RT‐qPCR analysis of BIR1 transcripts in sid2‐2 BIR1 L5 and eds5‐3 BIR1 L10 mutants. (c) Morphology of representative sid2‐2 BIR1 L5 and eds5‐3 BIR1 L10 plants. Images from independent transgenic lines are shown in Supporting Information S1: Figure S6. (d) RT‐qPCR analysis of PR1 transcripts in WT, eds1‐2 BIR1 L1b, pad4‐1 BIR1 L0, sobir1‐12 BIR1 L6, bak1‐5 BIR1 L10, sid2‐2 BIR1 L5 and eds5‐3 BIR1 L10 plants. Samples were analyzed after 12 days of mock (DEX−) or 30 µM DEX treatment (DEX+). qRT‐PCR data were normalized to the CBP20 internal control and compared to mock‐treated plants (set to 1). Data are mean ± SD, analyzed by Mann–Whitney test; *p < 0.05. Experiments were repeated at least twice with similar results.

3.7. Expression of Phytohormone Defence Markers in Compromised‐Defence Mutants Overexpressing BIR1

RNA‐seq data revealed significant transcriptional changes in genes involved in phytohormone metabolism and signalling in BIR1‐overexpressing plants. To assess the impact of plant hormone signalling on BIR1‐related phenotypes, we analyzed several well‐characterized phytohormone‐related markers in BIR1‐overexpressing Arabidopsis mutants. We used RT‐qPCR to measure transcript levels of genes encoding JA‐responsive PLANT DEFENSIN 1.2 (PDF1.2), ethylene (ET) and JA‐responsive PR4, and proteins involved in camalexin (CA) biosynthesis, such as CYP71B15 (PAD3) and CYP79B2. None of these genes were affected by DEX treatment in WT plants (Supporting Information S3: Table S2). PDF1.2 induction in bak1‐5 BIR1 plants suggests a role for BAK1 in regulating JA signalling during BIR1 overexpression, as reported for JA‐dependent resistance to wounding and herbivory (Figure 7a) (Yang et al. 2011; Tungadi et al. 2021). EDS1 acts as a repressor of JA/ET defence signalling (Brodersen et al. 2006), which was consistent with PDF1.2 induction in eds1‐2 BIR1 mutants (Figure 7a). The ET/JA marker PR4 was activated in all BIR1 overexpressing mutants, including the SA‐deficient sid2‐2 and eds5‐3 mutants, suggesting that the SA/ET‐JA crosstalk is unaffected by BIR1 overexpression (Figure 7b) (Stroud et al. 2022). The expression of genes encoding CA‐related PAD3 and CYP79B2 was upregulated in all defence‐compromised mutants (Figure 7c). In conclusion, our data suggest that the cell death‐like phenotypes in BIR1 overexpressing plants are not directly linked to hormone defence signalling activation or inhibition.

Figure 7.

Expression analyses of phytohormone‐related genes in BIR1 overexpressing plants. (a–c) RT‐qPCR analysis of PDF1.2 (a), PR4 (b), PAD3 and CYP79B2 (c) gene expression in sobir1‐12 BIR1 L6, bak1‐5 BIR1 L10, eds1‐2 BIR1 L1b, sid2‐2 BIR1 L5 and eds5‐3 BIR1 L10 mutant lines. Measurements were taken in mock‐treated (DEX−) versus DEX‐treated (DEX+) plants using the same set of replicates. Data were normalized to the CBP20 internal control and compared to mock‐treated plants (set to 1). Values are mean ± SD, analyzed by Mann–Whitney test; *p < 0.05. (d) Proposed model for BIR1 homeostasis: SA‐mediated BIR1 transcription is regulated by RNA‐directed DNA methylation (RdDM) and post‐transcriptional RNA silencing (PTGS) (Guzmán‐Benito et al. 2019). Constitutive BIR1 overexpression is sensed by TNLs, such as CSA1, triggering deregulation of immune‐related and defence hormone genes that overlap with typical PTI and ETI responses. TNL activation then induces ETI‐type cell death, requiring EDS1 and SOBIR1.

4. Discussion

Plants regulate the expression of various genes to execute defence responses to microbes and viruses (Nishad et al. 2020). BIR1, for example, is upregulated during bacterial, oomycete and viral infections in a SA‐ and JA‐dependent manner (Gao et al. 2009; Guzmán‐Benito et al. 2019; Robinson et al. 2025). The increased resistance of bir1‐1 mutants to the oomycete Hyaloperonospora parasitica Noco2 and the RNA virus TRV suggests a role for BIR1 in plant immunity (Gao et al. 2009; Guzmán‐Benito et al. 2019; Robinson et al. 2025). Overexpression of BIR1 in Arabidopsis, using both a viral vector and a transgenic inducible system, led to senescence and cell death‐like phenotypes, similar to those seen in BIR1 knockout plants, suggesting that improper regulation of BIR1 disrupts its function (Guzmán‐Benito et al. 2019). To investigate the genetic components responsible for these phenotypes, we used a DEX‐inducible system to express a mCherry‐tagged BIR1 transgene in Arabidopsis mutants defective in key PTI and ETI regulators. These transgenic plants accumulated comparable BIR1‐mCherry protein levels after DEX treatment, indicating consistent induction between genotypes. Nevertheless, although certain variability in transgene expression between plants was observed, we believe that their contribution to the observed phenotypic differences is minor. Plants with different expression levels exhibited similar morphologies, and those with high BIR1 levels were indistinguishable from those with even higher levels. This observation suggests that phenotype severity is not directly correlated with the BIR1 overexpression levels. Instead, we hypothesize that cell‐death phenotypes likely result from sustained, unregulated and constitutive overexpression of BIR1 over time. Once an optimal expression threshold is exceeded, autoimmune‐like responses are activated, and beyond this threshold, further increases in BIR1 levels do not produce additional phenotypic effects.

The transcriptional outputs induced by BIR1 overexpression closely resemble PTI and TNL‐mediated ETI activation, suggesting that BIR1 modulates immunity by influencing both pathways. TNL‐based immunity involves two signalling modules, EDS1/SAG101/NRG1 and EDS1/PAD4/ADR1 (Sun et al. 2021). We found that BIR1‐induced phenotypes partially depend on EDS1. We observed that BIR1 overexpression in a pad4‐1 mutant background produced cell‐death phenotypes similar to BIR1 L9 plants (Supporting Information S1: Figure S6a), suggesting that EDS1, potentially with SAG101, triggers cell death‐like phenotypes upon TNL activation in BIR1 overexpressing plants (Figure 7d). Although the mechanism of TNL activation under BIR1 overexpression remains unclear, one model suggests that loss of BIR1 activates hypothetical guarding R proteins, initiating SOBIR1‐dependent resistance pathways (Gao et al. 2009). We propose that excess BIR1 may affect protein maturation or multimeric complex formation, which unknown TNL proteins could sense to activate EDS1‐dependent mediated resistance without effectors. We also show that SOBIR1 is essential for the BIR1 overexpression phenotype, but it remains unclear whether SOBIR1 and EDS1 act independently, synergistically or in parallel pathways in transmitting the immune signals initiated by BIR1 overexpression. Interestingly, previous reports, as well as our own unpublished data, indicate that BIR1 does not interact directly with SOBIR1 or EDS1, despite both being essential for the cell death and constitutive defence responses observed in bir1‐1 mutants (Gao et al. 2009) and BIR1 overexpression lines. Overexpression of BIR1 activates RLPs, many of which are primarily expressed during leaf senescence, supporting the idea that BIR1 overexpression accelerates a senescence programme (Wang et al. 2008). Notably, BIR1‐responsive RLP23, RLP30 and RLP42 are known to interact with SOBIR1 (Wang et al. 2008; Zhang et al. 2013; Albert et al. 2019). Furthermore, the RLP23 receptor forms a supramolecular complex with EDS1/PAD4/ADR1 via SOBIR1, mediating ligand‐independent defence signalling (Pruitt et al. 2021). These observations suggest that the phenotypic rescue in the sobir1‐12 mutant may result from the disruption of SOBIR1‐dependent RLP signalling, which is constitutively activated by BIR1 overexpression. The TNL protein CSA1 detects disturbances in BIR3, BAK1 or BIR3/BAK1 complexes, triggering ETI‐type autoimmune responses (Schulze et al. 2022). CSA1 interacts with BIR1, but does not respond to BIR1 overexpression in our assay, and the csa1‐2 mutation does not revert the cell death phenotypes observed in bir1‐1 mutants (Schulze et al. 2022). Further investigation is needed to determine if CSA1 guards BIR1 integrity, possibly in cooperation with SOBIR1 or other SERKs (Figure 7d). Finally, BIR1 overexpression phenotypes were unaffected by the bak1‐5 mutation, suggesting that other SERK family members may compensate for BAK1 loss in PTI/ETI signalling.

We found that BIR1 overexpression upregulated genes encoding FRK1, WRKY29 and RBOHD, an effect reduced by mutations in SOBIR1, but independent of EDS1. Since EDS1 and SOBIR1 are crucial for both PRR‐ and TNL‐mediated signalling (Pruitt et al. 2021; Tian et al. 2021), further experiments are required to determine whether PTI gene expression changes involve PTI pathways, TNL activation or both. Our study also reveals significant changes in plant hormone gene expression due to BIR1 overexpression. Genes associated with ET, JA, CA and SA, which regulate immune responses, remained upregulated in eds1‐2 BIR1 and sobir1‐12 BIR1 plants, where phenotypes were partially rescued. This suggests that hormone regulation in BIR1 overexpressing does not require SOBIR1‐ and/or EDS1‐mediated signalling, and that the observed hormonal imbalance minimally impacts the phenotypes.

Arabidopsis G protein AGB1 (SOBIR2) is a common signalling component in PTI mediated by various RLK‐PRRs, functioning alongside PAD4 downstream of SOBIR1 to regulate cell death and pathogen resistance in bir1‐1 mutants (Liu et al. 2013; Liang et al. 2016). Similarly, Gγ‐proteins AGG1/AGG2 act as positive regulators of cell death and constitutive defence in bir1‐1 mutants (Liu et al. 2013). AGB1 and AGG2‐coding genes were upregulated in BIR1 overexpressing plants, suggesting they may regulate EDS1‐mediated cell death when BIR1 integrity is compromised. Additionally, BIR1‐responsive GPA1 and XLG, associated with Gβγ modules, are critical for growth and development (Wang et al. 2023). Thus, overactivation of these G proteins may somehow contribute to cell‐death phenotypes due to BIR1 overexpression.

In conclusion, our study explores the implications of BIR1 homeostasis in the immune response of Arabidopsis. BIR1 overexpression induces cell death phenotypes similar to those in bir1‐1 loss‐of‐function mutants and causes transcriptomic changes that overlap with canonical PTI and TNL‐dependent ETI responses. We demonstrate that both EDS1 and SOBIR1 are required for the observed cell death phenotypes in BIR1 overexpressing plants. We propose that an excess of BIR1 resulting from BIR1 misregulation may cause perturbations in BIR1 function, sensed by unknown TNLs, with SOBIR1 and EDS1, likely together with SAG1, transducing signals downstream of these TNLs (Figure 7d). Further research is needed to test these hypotheses and explore their relevance in other plant species. Identifying novel BIR1 interactors will also enhance our understanding of the mechanisms underlying BIR1‐mediated immunity regulation.

Supporting information

GB et al PCE Supplemental Figures 2.

Table S1.

Table S2.

Table S3.

Table S4.

Table S5.

Table S6.

Table S7.

Table S8.

Table S9.

GB et al PCE Supplemental data final 2.

Data Availability Statement

The data supporting this study's findings are available from the corresponding author upon reasonable request.