Genome-wide identification and functional characterisation of the EDS1 gene family reveals evolutionary conservation and stress-responsive regulatory roles in barley

Abstract

Enhanced Disease Susceptibility 1 (EDS1) genes are central regulators of plant innate immunity and have emerging roles in biotic stress adaptation. Although extensively characterized in dicots, a genome-wide investigation of the EDS1 gene family in barley (Hordeum vulgare L.) remains unavailable. Thirteen HvEDS1 genes were systematically identified by integrated HMM and BLASTP approaches. These genes encode structurally distinct proteins that differ in length, charge, stability and subcellular localization, with several of them reported to occur in the nucleus, chloroplasts, mitochondria and cytoskeleton. Phylogenetic analyses subdivided the HvEDS1 proteins into three major clades, revealing both a diversification of monocots and conserved ancestors. Analyses of gene structure and conserved motifs revealed clade-specific exon–intron architectures and domain configurations, suggesting functional specialization. Analysis of the genome distribution showed that the spread of HvEDS1 is primarily due to scattered duplications, with purifying selection acting on the duplicated pairs. Analysis of the promoters identified numerous cis-regulatory elements that respond to hormones (ABA, JA), light (G-box) and abiotic influences (MBS), suggesting multi-layered transcriptional regulation. The predicted miRNA–mRNA interactions revealed that hvu-miR6192, a stress-responsive microRNA, targets receptor-like kinases involved in abiotic stress signaling, suggesting a post-transcriptional regulatory role in HvEDS1-mediated stress responses. KEGG enrichment linked HvEDS1 genes to lipid signaling pathways, including glycerophospholipid and α-linolenic acid metabolism, as well as ubiquitin-mediated proteolysis. Structural modelling indicated conservation of key EDS1 domains, but showed protein-specific variations in loop flexibility and secondary structure content, suggesting different functional dynamics. The overall transition from a specialized, low-redundancy network under normal conditions to a highly coordinated and overlapping network under fungal stress condition reflects the versatility of the HvEDS1 family in rewiring of signaling cascades stress responsive pathways during different physiological conditions. This first genome-wide characterization of the HvEDS1 gene family reveals its structural diversity, evolutionary development and potential role in immunity and abiotic stress signaling. The identified candidates provide valuable targets for functional validation and molecular breeding of stress-resistant barley cultivars.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-026-41481-8.

Introduction

Barley (Hordeum vulgare L.) is one of the earliest domesticated cereal plants and currently ranks fourth in global cereal production, making a significant contribution to the food, feed and malting industries¹. Due to its diploid genome, its extensive genetic resources and its readily detectable phenotypes, barley serves as an important model for genetic and functional genomics research in temperate cereals^2,3. However, its productivity is increasingly threatened by a wide range of environmental stresses, especially the increasing pressure of global climate change⁴. Both biotic and abiotic stress factors have a negative impact on barley development and yield, often through disruption of cellular homeostasis, redox balance and hormonal regulation. While drought and salt stress affect physiological processes such as photosynthesis, osmotic adjustment and ion balance^5,6, biotic threats such as powdery mildew (Blumeria graminis f. sp. hordei) and late blight (Fusarium spp.) lead to significant yield losses and grain spoilage⁷.

To improve the resilience of barley, a detailed understanding of the molecular components that control stress perception and signaling transduction is essential⁸. Among the central components of plant innate immunity, Enhanced Disease Susceptibility 1 (EDS1) functions as a critical signal transduction site that integrates both biotic and abiotic stress responses⁹. EDS1, together with its well-characterized interaction partners Phytoalexin Deficient 4 (PAD4) and Senescence-Associated Gene 101 (SAG101), forms a highly conserved signaling module structurally characterized by an N-terminal α/β-hydrolase-like domain and a C-terminal EP domain (EDS1–PAD4)¹⁰. These domains enable the formation of functional EDS1–PAD4 and EDS1–SAG101 heterodimers, which combine pathogen perception by intracellular immune receptors with downstream transcriptional reprogramming and activation of systemic defense¹¹.

In classical effector-triggered immunity (ETI), plant resistance proteins (R-proteins) with nucleotide binding sites and leucine-rich repeat domains (NBS-LRR) recognize pathogen-derived effectors and activate robust immune responses, often resulting in localized programmed cell death¹². Based on their N-terminal domains, R proteins are categorised as coiled-coil (CC)-NBS-LRRs and Toll/interleukin-1 receptor (TIR)-NBS-LRRs¹³. EDS1 is essential for downstream signaling of TIR-type R proteins in dicotyledonous plants, where it mediates resistance to a variety of pathogens by activating transcriptional immune reprogramming¹⁴. The EDS1–PAD4 complex promotes basal and systemic acquired resistance by interacting with helper NLRs of the ADR1 family (Activated Disease Resistance 1), while the EDS1–SAG101 complex interacts with NRG1 (N Requirement Gene 1) to mediate localized hypersensitive responses¹⁵.

In addition to classical immune functions, there is increasing evidence that EDS1 complexes play a broader role in adaptation to abiotic stress and in developmental processes^10,16. In A. thaliana, EDS1 and PAD4 improve drought tolerance by regulating stomatal conductance and the expression of stress-responsive genes¹⁶. In Vitis vinifera, EDS1 transcripts are rapidly upregulated upon infection with Erysiphe necator, resulting in resistance to powdery mildew¹⁷. In monocotyledonous plants such as wheat and rice, orthologous EDS1 genes are involved in both defense responses and oxidative signaling. For example, overexpression of TaEDS1 in Triticum aestivum restricts Blumeria graminis infection and modulates hydrogen peroxide signaling, suggesting a crucial role in redox homeostasis¹⁸. Furthermore, under hypoxic conditions, EDS1, PAD4 and LSD1 regulate the formation of lysogenic aerenchyma in the hypocotyls of Arabidopsis via H₂O₂- and ethylene-dependent pathways, indicating a broader role in developmental plasticity¹⁹. In Oryza sativa, OsEDS1 contributes to heat stress tolerance by upregulating catalase activity and promoting detoxification of reactive oxygen species²⁰. Despite these advances, the functional diversification, evolutionary conservation and stress-specific regulatory functions of EDS1 homologues in monocotyledonous species, particularly barley, remain poorly understood. Given its central role in the coordination of stress responses and immune signaling, a comprehensive characterization of the EDS1 gene family in barley is crucial to unlock its potential in molecular breeding programmes to improve plant resistance. In this study, the members of the HvEDS1 family will be identified and classified, genome-wide, their gene structures, conserved domains and chromosomal positions will be investigated and phylogenetic and collinearity analyses will be performed to describe their evolutionary history. In addition, the expression dynamics of HvEDS1 genes under different abiotic and biotic stress conditions will be explored using RNA-seq datasets, providing insights into their potential regulatory divergence and stress-specific roles. Overall, this research will provide a basis for the functional validation of HvEDS1 genes and inform genetic engineering strategies for the development of stress-resistant barley cultivars.

Materials and methods

Identification and characterization of the EDS1 gene family in barley

The genome assembly (Hordeum vulgare cv. Morex v3) and its corresponding gene annotation (IBSC_v3) were retrieved from the Ensembl Plants database (https://plants.ensembl.org). To identify bona fide EDS1 proteins, Hidden Markov Model (HMM) profiles corresponding to the two conserved EDS1 domains were obtained from the Pfam database, namely the Lipase_3 domain (Pfam: PF01764) and the EDS1–PAD4 EP domain (Pfam: PF18117). Both domains were used because canonical EDS1 proteins universally harbour these two conserved regions, distinguishing them from general lipase-like proteins. The barley proteome was screened using HMMER version 3.0 with an E-value threshold of 1 × 10⁻⁵ to ensure specificity^14,21,22. In parallel, BLASTP searches were performed against the barley proteome using Arabidopsis thaliana EDS1 protein sequences obtained from TAIR (https://www.arabidopsis.org). BLASTP parameters included an E-value cut-off of 1 × 10⁻⁵, a best-hit overhang of 0.25, and a maximum of five hits per query to minimise false positives. Redundant sequences and partial or incomplete sequences lacking either the Lipase_3 or EP domain were excluded.

All non-redundant candidate sequences were further validated using multiple conserved-domain annotation platforms, including InterProScan (https://www.ebi.ac.uk/interpro/), Pfam (https://pfam.xfam.org/) via both the online interface and local PfamScan utility, SMART (http://smart.embl-heidelberg.de/), and the NCBI Conserved Domain Database using Batch CD-Search (https://www.ncbi.nlm.nih.gov/Structure/bwrpsb/bwrpsb.cgi). Proteins were retained only if they contained both PF01764 and PF18117, consistent with established criteria defining EDS1 family members in higher plants.

Prediction of physicochemical properties of Barely HvEDS1 proteins

The physicochemical characteristics of the confirmed HvEDS1 proteins including amino acid sequence length, molecular weight (MW), theoretical isoelectric point (pI), instability index, aliphatic index, and the grand average of hydropathicity (GRAVY) score were computed using the ExPASy ProtParam tool (https://web.expasy.org/protparam/). These parameters provide preliminary insights into the stability and solubility profiles of the identified proteins. To gain further insight into their potential biological roles, the subcellular localization of each HvEDS1 protein was predicted using WoLF PSORT (https://www.genscript.com/wolf-psort.html). Predicted localizations were used to infer putative functional roles in organelle-specific signaling and stress responses.

Phylogenetic analysis of EDS1 proteins

To elucidate the evolutionary relationships among EDS1 proteins in Hordeum vulgare and two representative plant species A. thaliana and O. sativa, a multiple sequence alignment of full-length EDS1 protein sequences was conducted using ClustalW as implemented in BioEdit version 7.2. Phylogenetic trees were constructed using the Neighbor-Joining (NJ) method with 1,000 bootstrap replicates, with all other parameters set to their default values, using the MEGA X version 11 software. Two separate phylogenetic trees were generated: one comprising only H. vulgare EDS1 proteins to assess intra-species divergence, and another including all three species to investigate inter-species evolutionary relationships. The resulting trees were visualized using the Interactive Tree of Life (iTOL) online tool (https://itol.embl.de/)²³. To further evaluate domain-level conservation within the H. vulgare EDS1 protein family, sequence logo analysis was performed to identify conserved amino acid motifs. The aligned HvEDS1 sequences were analyzed using the WebLogo tool (https://weblogo.berkeley.edu), which generated graphical representations of residue conservation, thereby revealing domain-specific sequence features.

Genomic distribution, collinearity, and selection pressure analysis of EDS1 genes in barley

The chromosomal positions of the EDS1 genes were identified using annotation files from the barley reference genome and their physical distribution across chromosomes was visualized using TBtools v1.108²⁴. To investigate gene duplications and syntenic relationships, a collinearity analysis was performed using the MCScanX toolkit²⁵ integrated in TBtools. A circular map illustrating these relationships was generated using the Basic CIRCOS function. To assess the evolutionary dynamics of the duplicated gene pairs, the coding sequences (CDS) of the EDS1 genes were retrieved and analyzed using the “Simple Ka/Ks Calculator” module in TBtools v1.108²⁴. The resulting non-synonymous (Ka) and synonymous (Ks) substitution rates were used to calculate the Ka/Ks ratio, which served as an indicator of selection pressure: Ka/Ks < 1 indicated purifying selection, Ka/Ks > 1 indicated positive selection and Ka/Ks = 1 indicated neutral evolution.

Motif, domain, and gene structure analysis

Conserved motifs within the HvEDS1 protein sequences were identified using MEME Suite v5.4.1 (http://meme-suite.org/)²⁶ with a significance threshold of p < 1 × 10⁻⁵ and default motif settings. Annotation of functional domains was performed using the SMART database (http://smart.embl-heidelberg.de/)²⁷ and the Pfam database (https://pfam.xfam.org/) to confirm the presence of domains. Gene structure, including exon-intron organization, was analyzed and visualized using TBtools²⁴, which also integrated the results of the motif and domain analyzes together with the phylogenetic tree.

Cis-acting element analysis in EDS1 gene promoters

Promoter sequences (2000 bp upstream of the start codon) of HvEDS1genes were extracted and entered into the PlantCARE database (http://bioinformatics.psb.ugent.be/webtools/plantcare/html/)²⁸ to identify cis-acting regulatory elements. The identified elements were categorized and visualized using the heatmap and phylogeny visualization tools in TBtools²⁴.

Prediction of HvEDS1 Protein-Protein Interactions

The protein–protein interaction (PPI) network of HvEDS1 proteins was predicted using the STRING v12.0 database (https://string-db.org/)²⁹ based on functional orthologs in Arabidopsis thaliana. The parameters used were: complete STRING network, minimum interaction value of 0.4 (medium confidence) and a maximum display limit of 10 interactors.

Transcription factor regulatory network analysis of HvEDS1 genes

Prediction of transcription factor (TF) binding sites upstream of HvEDS1 genes was performed according to the method described by Jin et al. (2014) with minor modifications³⁰. Promoter regions extending more than 2000 bp upstream of each HvEDS1 gene were extracted and analyzed using the Plant Transcriptional Regulatory Map (PlantRegMap) (http://plantregmap.gao-lab.org/regulation_prediction.php), applying a p-value threshold of ≤ 1e^{− 631}. The resulting TF gene interaction network was then generated and visualized using Cytoscape v3.9.1³².

Prediction of putative miRNAs targeting HvEDS1 genes and functional annotation

To identify candidate microRNAs (miRNAs) targeting H. vulgare EDS1 (HvEDS1) genes, mature barley miRNA sequences were retrieved from the miRBase database (https://www.mirbase.org/). The coding sequences (CDS) of the identified HvEDS1 genes were submitted to the psRNATarget online tool (https://www.zhaolab.org/psRNATarget/) for target prediction, using the 2017 schema version with default parameters, including a maximum expectation score of 3.0. The predicted miRNA–target interactions were subsequently utilised to construct a gene–miRNA regulatory network, which was visualized using Cytoscape software (version 3.9; https://cytoscape.org/download.html).

Functional annotation and enrichment analysis

For functional annotation, the protein sequences of all HvEDS1 genes were submitted to eggNOG-mapper (http://eggnog-mapper.embl.de/; version 2.1.9) to retrieve Gene Ontology (GO) terms and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway annotations. GO and KEGG enrichment analyses were subsequently performed in R (version 4.3.1) using the ‘clusterProfiler’ package (version 4.8.1) using default parameters. Functional categories with an adjusted p-value (False Discovery Rate, FDR) < 0.05 were considered statistically significant. The enrichment results were visualized using the ‘ggplot2’ (version 3.4.4) and ‘GOplot’ (version 1.0.2) packages to facilitate graphical interpretation.

EDS1 gene regulatory network inference using the machine learning approach

To construct the gene regulatory network, a random forest regression algorithm was employed to model the expression of H. vulgare EDS1 genes under both normal and fungal stress conditions. Raw RNA-seq datasets (SRA Project ID: PRJNA315041) were retrieved from the NCBI Sequence Read Archive. Sequence quality was first assessed using FastQC to ensure that no low-quality reads or technical artifacts were present. Adapter trimming and quality filtering were then performed using Trimmomatic, which removed any contaminating adapter sequences and low-quality bases. The cleaned reads were subsequently aligned to the H. vulgare cv. Morex reference genome (IBSC_v3) using HISAT2, a fast and sensitive aligner. Transcript abundance was quantified as transcripts per million (TPM) using StringTie, which allows for accurate estimation of transcript levels while correcting for gene length and sequencing depth.

Normalisation and variance-stabilising transformation were carried out using DESeq2, ensuring that expression values were comparable across samples. Genes with low expression (TPM < 1 in > 80% of samples) or low variance were excluded from further analysis to minimise background noise and improve model robustness. This filtering step helped focus the analysis on genes with consistent and reliable expression across the dataset.

The resulting expression matrix, containing the final filtered and normalised data, was then used as input for regulatory inference. For each target HvEDS1 gene, the random forest modelled its expression as a function of all other expressed genes in the dataset. At each tree node, a random subset of predictors was selected, with the number of predictors sampled at each split defined as mtry = √p, where p is the total number of input genes. The model was trained using 1,000 trees (ntrees = 1000) to ensure convergence and stability, with implementation in R (version 4.3.1) using the randomForest package. The inferred gene regulatory network was then analysed to identify regulatory shifts associated with HvEDS1 genes under control and Cochliobolus sativus-infected conditions. Differential network analysis, comparing gene connectivity, edge weight distributions, and regulatory importance scores, enabled the detection of stress-responsive rewiring events within the HvEDS1 regulatory landscape.

Results

Identification and characterization of the EDS1 gene family in barley

A total of thirteen HvEDS1 genes were identified in the barley genome using BLASTp and HMMER3 searches with known EDS1 protein sequences, using Arabidopsis thaliana and Oryza sativa as queries. (Table S1). These genes are distributed across multiple chromosomes and are transcribed from either the sense or antisense strand, with no evidence of strand-specific clustering, highlighting a dispersed genomic arrangement. The HvEDS1 gene models exhibited substantial variation in their genomic and transcript structures. Transcript lengths, including untranslated regions (UTRs) and coding sequences (CDS), ranged from 1,398 to 3,350 bp, while CDS lengths spanned from 1,035 to 3,048 bp. Notably, all HvEDS1 transcripts were associated with a single isoform, except for HvEDS1-8, which exhibited two transcript variants, indicating potential alternative splicing events. Protein lengths ranged from 344 to 1,015 amino acids, with corresponding molecular weights between 38.66 and 112.85 kDa. The broad range of predicted isoelectric points (pI), from 5.51 to 9.25, suggests variability in charge properties and solubility profiles under physiological pH conditions. The grand average of hydropathy (GRAVY) values ranged from − 0.491 to 0.124, indicating that most HvEDS1 proteins are predominantly hydrophilic, with HvEDS1-6 being the most hydrophilic member. The aliphatic index values, indicative of protein thermostability, ranged from 76.70 to 100.02, while the instability index spanned from 25.00 (stable) to 53.94 (moderately unstable), suggesting diverse structural stabilities across the gene family. Collectively, these physicochemical properties reflect the evolutionary diversification and potential functional specialization of HvEDS1 proteins.

Subcellular localization predictions indicated a broad distribution of HvEDS1 proteins across various compartments, including the nucleus, cytosol, mitochondria, chloroplasts, endoplasmic reticulum (ER), peroxisomes, vacuole, Golgi apparatus, cytoskeleton, and extracellular space (Fig. 1). Several members most notably HvEDS1-3, HvEDS1-5, HvEDS1-8, and HvEDS1-11 displayed high confidence scores for nuclear localization (≥ 7.0), implying potential roles in transcriptional regulation and nuclear signal transduction pathways. Chloroplast-specific localization was most prominent in HvEDS1-4 and HvEDS1-7, suggesting involvement in plastid-associated stress responses or redox signaling mechanisms. In addition, HvEDS1-5 exhibited strong association with the cytoskeleton, possibly implicating it in structural integrity or protein trafficking. HvEDS1-6 presented a notably diverse localization profile, with predicted presence in the nucleus, mitochondria, and chloroplasts, suggesting a multifaceted regulatory role at the intersection of organellar and nuclear signaling networks. Similarly, HvEDS1-9 and HvEDS1-13 demonstrated multi-compartmental localization patterns, reflecting functional versatility and potential cross-talk between cellular compartments.

Fig. 1.

Predicted subcellular localization of HvEDS1 proteins in barley (Hordeum vulgare). The heatmap illustrates the frequency scores of predicted HvEDS1 protein localization across different cellular compartments based on WoLF PSORT analysis. Abbreviations: Cyto: cytoplasm; Cyto–nucl: cytoplasm and nucleus; Mito: mitochondria; E.R.: endoplasmic reticulum; Pero: peroxisome; Chlo: chloroplast; Extr: extracellular space; Vacu: vacuole; Plas: plasma membrane; Nucl: nucleus; Golg: Golgi apparatus; Cysk–nucl: cytoskeleton and nucleus; Cysk: cytoskeleton. The colour gradient from yellow to red represents increasing prediction confidence or occurrence frequency, with yellow indicating low frequency and red indicating high frequency. Numerical values within each cell denote the number of predictive hits supporting localization in the respective compartment. Hierarchical clustering along the left dendrogram groups HvEDS1 proteins with similar subcellular distribution profiles.

Table 1.

Basic information of EDS1 gene family members in barley.

Gene name	Gene ID	Protein length (aa)	Isoelectric point (pI)	Molecular weight (kDa)	GRAVY	Aliphatic index	Instability index (II)
HvEDS1-1	HORVU.MOREX.r3.1HG0023010	519	9.12	56.61	0.124	100.02	25.00
HvEDS1-2	HORVU.MOREX.r3.1HG0059850	344	5.82	38.66	– 0.122	87.53	43.34
HvEDS1-3	HORVU.MOREX.r3.2HG0114390	351	5.83	39.33	– 0.030	92.22	31.22
HvEDS1-4	HORVU.MOREX.r3.2HG0118060	360	6.91	39.67	0.092	90.22	36.65
HvEDS1-5	HORVU.MOREX.r3.3HG0233200	524	6.53	57.21	0.076	94.96	26.69
HvEDS1-6	HORVU.MOREX.r3.3HG0276120	394	5.51	44.22	– 0.491	76.70	45.11
HvEDS1-7	HORVU.MOREX.r3.3HG0314780	350	6.04	39.17	0.006	92.20	39.46
HvEDS1-8	HORVU.MOREX.r3.4HG0347000	613	8.62	66.72	– 0.089	82.40	41.85
HvEDS1-9	HORVU.MOREX.r3.4HG0364120	361	5.85	40.15	0.011	92.11	36.56
HvEDS1-10	HORVU.MOREX.r3.5HG0427650	458	6.19	49.21	– 0.070	89.74	39.11
HvEDS1-11	HORVU.MOREX.r3.5HG0478590	619	5.92	68.70	– 0.169	83.89	53.94
HvEDS1-12	HORVU.MOREX.r3.6HG0584930	1,015	8.12	112.85	– 0.280	87.82	50.63
HvEDS1-13	HORVU.MOREX.r3.7HG0640520	465	9.25	51.12	0.115	95.20	28.45

GRAVY: Grand average of hydropathicity.

Phylogenetic analysis of EDS1 domain-containing proteins in barely

A maximum-likelihood phylogenetic tree was constructed using full-length EDS1 protein sequences from Hordeum vulgare (Hv), Oryza sativa (Os), and Arabidopsis thaliana (At) to examine evolutionary relationships within the EDS1 gene family (Fig. 2). The resulting topology resolved three well-supported lineages, designated Clade A (orange), Clade B (pink), and Clade C (blue), each representing a distinct branch derived from an ancestral diversification event that predated the monocot–dicot split. Clade A comprised predominantly monocot sequences from barley and rice, with several HvEDS1 paralogues forming closely related subgroups, indicating limited species-specific expansion within the Poaceae. Clade B likewise contained only Hv and Os members, forming a compact grouping that suggests an independent monocot lineage distinct from Clade A. Clade C included representatives from all three species, encompassing the canonical Arabidopsis homologues (At3G48080 and At3G48090). Within this clade, Oryza and Arabidopsis proteins were interspersed with Hordeum members: HvEDS1-8 and HvEDS1-10 formed a well-supported cluster near the clade core, whereas HvEDS1-12 occupied a longer branch, indicating greater sequence divergence while remaining within the conserved Clade C lineage.

Fig. 2.

Phylogenetic analysis of the EDS1 gene family in Hordeum vulgare and related species. A maximum-likelihood phylogenetic tree was constructed using full-length EDS1 protein sequences from Hordeum vulgare (Hv), Oryza sativa (Os), and Arabidopsis thaliana (At) to infer evolutionary relationships among EDS1 family members. The tree is divided into three major clades: Clade (A) (orange), Clade (B) (pink), and Clade (C) (blue) each representing distinct evolutionary lineages. Bootstrap support values are indicated by triangular node symbols of varying sizes, where larger triangles denote higher confidence (bootstrap > 90%). Barley proteins are highlighted in bold to facilitate interspecies.

A barley-specific phylogenetic tree constructed from the thirteen HvEDS1 proteins (Fig. 3) resolved four clusters, designated Hv-Clades A–D (Fig. 3A). Hv-Clade A comprised only HvEDS1-12, which consistently appeared as the most divergent member of the family, forming an isolated branch in both the barley-specific and interspecific analyses (Fig. 2). Hv-Clade B grouped HvEDS1-8 and HvEDS1-11, while Hv-Clade C contained HvEDS1-10 and HvEDS1-6. The remaining eight HvEDS1 genes formed Hv-Clade D, representing a more conserved assemblage. This gene distribution pattern mirrored the overall relationships observed in the interspecific tree (Fig. 2): HvEDS1-12 remained distinct within the broader Clade C lineage, whereas HvEDS1-8, HvEDS1-10, and HvEDS1-11 clustered more closely with the core EDS1 members of rice and Arabidopsis. The consistent placement of HvEDS1-12 as a long-branch singleton in both analyses suggests early functional divergence within the barley EDS1 family, potentially reflecting subfunctionalisation or neofunctionalisation following species-specific expansion. Overall, the phylogenetic distribution of HvEDS1 proteins reveals clear clustering patterns that reflect both conserved and species-restricted evolutionary relationships among monocots and dicots, consistent with previously reported diversification trends in the EDS1 family^33,34.

Fig. 3.

Gene structure, conserved motifs, and domain organisation of the HvEDS1 gene family in barley (H. vulgare). (A) Phylogenetic relationships among HvEDS1 genes, with branch colours indicating four distinct clades: Hv-Clade-A (pink), Hv-Clade-B (dark blue), Hv-Clade-C (light green), and Hv-Clade-D (orange). (B) Gene structure of HvEDS1 genes showing untranslated regions (UTRs; green boxes), coding sequences (CDS; orange boxes), and introns (black lines). (C) Conserved domain architecture of HvEDS1 proteins, with functional domains represented by distinct colours. (D) Distribution of conserved motifs among HvEDS1 proteins identified using MEME. Each motif is indicated by a unique colour and numbered 1–10, as shown in the legend. Motifs are displayed according to their relative positions along each protein sequence (from N- to C-terminus), and motif numbering reflects MEME identifiers rather than positional order. (E) Sequence logos of the ten conserved motifs generated by MEME, illustrating amino acid conservation patterns across HvEDS1 proteins.

Gene structure, conserved motifs, and domain architecture of eds1 gene family members in barley

Exon–intron structures, conserved motifs, and domain features of the HvEDS1 family were examined to assess structural divergence among paralogues (Fig. 3). In accordance with the phylogenetic relationships (Fig. 3A), the thirteen HvEDS1 genes were resolved into four clades (Hv-Clades A–D), and members within each clade generally exhibited similar structural features. Gene-structure analysis (Fig. 3B) revealed marked variability in exon–intron organisation, with exon numbers ranging from two (HvEDS1-2) to seven (HvEDS1-10). UTR regions were detected in several genes but were unevenly distributed across the clades. Such variation likely reflects intron gain or loss events that have contributed to the structural diversification of HvEDS1 members. Clade A (HvEDS1-1, -3, -4, -7, and − 9) generally contained three to five exons and lacked extended UTRs, whereas Clade D (HvEDS1-10) displayed a more complex exon–intron structure with extended intronic regions, indicating potential divergence in transcriptional regulation. Structural conservation within each clade may therefore correspond to shared evolutionary ancestry and similar regulatory features.

Domain-composition analysis (Fig. 3C) confirmed the ubiquitous presence of conserved EDS1-associated domains in all family members. Each HvEDS1 protein possessed the characteristic α/β-hydrolase fold domain (IPR029058), lipase 3 domain (IPR033973), and EP domain (IPR005747), consistent with canonical EDS1 proteins involved in plant immune signalling. Subtle differences in domain architecture, particularly in domain length and spatial arrangement, were detected among paralogues. For instance, HvEDS1-6 and HvEDS1-13 exhibited truncated C-terminal EP domains, suggesting potential differences in subcellular localisation or interaction dynamics compared with other family members.

Motif analysis using the MEME suite identified ten distinct motifs within the HvEDS1 protein sequences (Fig. 3D). Motifs 1, 2, and 3 were universally distributed across all clades and corresponded to the conserved functional core of EDS1 proteins. Motif 7 occurred exclusively within Clade C (HvEDS1-8, -10, -11), whereas motif 9 was restricted to Clade A (HvEDS1-3, -4). These clade-specific motifs may underpin functional specialisation related to stress responses or regulatory interactions. Sequence logos (Fig. 3E) showed strong conservation of polar and charged residues essential for protein stability and partner recognition. Motifs MEME-1 and MEME-3 contained residues forming the catalytic core of the lipase 3 domain, while MEME-6 and MEME-9 exhibited higher variability, reflecting evolutionary adaptation to divergent regulatory contexts. Integration of the phylogenetic, structural, and motif evidence suggests that the HvEDS1 gene family in barley has retained the canonical EDS1 framework while undergoing clade-specific diversification. The resulting structural and sequence variability may contribute to functional specialisation and stress-responsive adaptability within barley’s immune and signalling networks.

Chromosomal distribution, collinearity, and synteny analysis of the HvEDS1 gene family in barley

Thirteen HvEDS1 genes were mapped across all seven chromosomes of the barley (H. vulgare) genome, revealing an uneven but widespread chromosomal distribution (Fig. 4A). Chromosomes 1 H, 2 H, 4 H, and 5 H each harbour two HvEDS1 genes, while chromosome 3 H contains three. The presence of multiple HvEDS1 genes on chromosomes 1 H, 2 H, 3 H, 4 H, and 5 H suggests these genes may have arisen from duplication events, although the precise nature of these events (whether local or segmental) is not fully resolved by chromosomal distribution alone. Notably, chromosome 2 H contains two closely positioned HvEDS1 genes (HvEDS1-3 and HvEDS1-4), which may indicate a local duplication event in this region. Chromosomes 6 H and 7 H each contain a single HvEDS1 gene (HvEDS1-11 on 6 H and HvEDS1-13 on 7 H), suggesting these loci may have been retained as unique copies. The distribution of HvEDS1 genes suggests a complex expansion pattern, likely involving both segmental and local duplications.

Fig. 4.

Chromosomal distribution, gene duplication, and syntenic relationships of the HvEDS1 gene family in barley (H. vulgare). (A) Chromosomal mapping of HvEDS1 genes showing their physical positions across the seven barley chromosomes. (B) Intraspecies collinearity analysis identifying two segmentally duplicated HvEDS1 gene pairs, indicating potential gene expansion within the barley genome. (C) Comparative synteny between H. vulgare, Oryza sativa, and Arabidopsis thaliana, highlighting conserved orthologous gene pairs. Grey lines represent genome-wide syntenic relationships, while red lines denote syntenic connections specific to HvEDS1 genes.

To further characterise the duplication mechanisms underlying the expansion of the HvEDS1 gene family, intragenomic collinearity and duplication type analyses were conducted (Fig. 4B; Table S2). A single segmental duplication event, classified by MCScanX as “WGD or segmental,” was identified between HvEDS1-2 and HvEDS1-7. Here, “WGD” stands for “whole genome duplication,” a phenomenon often linked to ancient polyploidy events, which lead to large-scale duplications in the genome followed by subsequent chromosomal rearrangements. In this context, we refer to this duplicated pair as “segmentally duplicated” to distinguish it from other types of duplication. All other HvEDS1 genes were classified as products of dispersed duplication (Table S2), suggesting that dispersed duplication mechanisms predominantly contributed to the HvEDS1 family expansion. The absence of tandem or proximal duplications further supports the conclusion that both dispersed and segmental mechanisms have shaped the genomic distribution of HvEDS1 genes. To evaluate selective pressure, Ka and Ks values were calculated for the identified paralogous HvEDS1 pairs. As only one segmental paralogue pair (HvEDS1-2/HvEDS1-7) was detected in the barley genome, Ka/Ks analysis was only applicable to this pair. All other HvEDS1 loci occurred as dispersed singletons, precluding pairwise evolutionary rate estimation. The HvEDS1-2/HvEDS1-7 pair exhibited a Ka/Ks ratio of 0.36 (Table S3), which is consistent with purifying selection and suggests functional conservation and evolutionary constraint at both the protein and regulatory levels.

To further investigate the evolutionary conservation of HvEDS1 loci across lineages, a comparative synteny analysis was performed with Arabidopsis thaliana and Oryza sativa (Fig. 4C). Several HvEDS1 genes were found in conserved syntenic blocks with O. sativa, particularly on chromosomes 1, 3, 5, 7, and 9, suggesting that orthologous relationships are retained within monocots. In contrast, only limited syntenic correspondence was observed between barley and A. thaliana, consistent with the significant divergence between monocot and dicot lineages. These results highlight that the HvEDS1 gene family underwent monocot retention, followed by functional diversification within the Poaceae.

Cis-acting regulatory element analysis of HvEDS1 genes

To elucidate the transcriptional regulatory landscape of the HvEDS1 genes in Hordeum vulgare, the 2 kb upstream promoter regions were analysed using the PlantCARE database (Fig. 5A, B). The detected cis-elements were grouped into four functional categories: cell cycle and development-related, hormone-responsive, light-responsive, and stress-responsive motifs. Developmental and cell cycle–related elements were comparatively scarce. CAT-box elements were detected in several HvEDS1 promoters, including HvEDS1-1, HvEDS1-4, HvEDS1-5, and HvEDS1-9 to HvEDS1-12, but were absent from others (Fig. 5A). The GCN4-motif was detected only in HvEDS1-5. GAP-box elements were detected in HvEDS1-6, HvEDS1-7, HvEDS1-9 to HvEDS1-11, and HvEDS1-13, whereas F-box–related elements were observed in all HvEDS1 promoters except HvEDS1-3, HvEDS1-8, HvEDS1-10, and HvEDS1-11 indicating restricted developmental regulation.

Fig. 5.

Cis-regulatory element analysis of HvEDS1 promoters. (A) Heatmap showing the frequency of cis-acting regulatory elements identified within the 2 kb upstream promoter regions of HvEDS1 genes. Numbers inside boxes represent the count of each motif, and cell shading intensity reflects increasing motif abundance (yellow–orange scale). (B) Genomic distribution of identified cis-regulatory elements across HvEDS1 promoter regions. Elements are grouped into four functional classes: cell cycle and development-related, hormone-responsive, light-responsive, and stress-responsive.

Hormone-responsive elements were the second most abundant and widespread. Abscisic acid-responsive ABRE motifs were present in almost all promoters except HvEDS1-3. Jasmonic acid-responsive TGACG- and CGTCA-motifs were frequently detected, except in HvEDS1-5, with stronger representation in HvEDS1-4 compared with HvEDS1-7 and HvEDS1-1 (Fig. 5A). Auxin-responsive TGA elements were detected in several HvEDS1 promoters, including HvEDS1-1, HvEDS1-2, HvEDS1-4, HvEDS1-6, and HvEDS1-12, whereas AuxRR-core motifs were not identified. Gibberellin-responsive elements, including the P-box and GARE motif, were detected in multiple HvEDS1 promoters: the P-box was found in HvEDS1-13, HvEDS1-6, HvEDS1-7, and HvEDS1-4, while the GARE motif was present in HvEDS1-2, HvEDS1-5, and HvEDS1-12. Salicylic acid-responsive TCA elements and ethylene-responsive EREs were present at low frequencies.

Light-responsive motifs were the third most prevalent category, with Box 4, G-box, GT1 motif, and TCT motif widely distributed. Six HvEDS1 promoters contained GT1 motifs, including HvEDS1-3, HvEDS1-4, HvEDS1-6, HvEDS1-8, HvEDS1-10, and HvEDS1-13. Among all HvEDS1 genes, HvEDS1-7 and HvEDS1-11 contained fewer light-responsive motifs compared to other family members, whereas HvEDS1-8 and HvEDS1-13 contained the highest numbers of light-responsive motifs.

Stress-responsive elements were the most highly represented across HvEDS1 promoters, including MYB recognition sites, MYB and MYB-like sequences, MYC motifs, ARE (anaerobic induction), as-1, W-box and WRE3 (WRKY-related elements), STRE (stress response elements), LTR (low-temperature response), JERE (jasmonate and elicitor-responsive element), TC-rich repeats, and CAG motifs (Fig. 5A, B). STREs were particularly abundant in HvEDS1-2 to HvEDS1-8, HvEDS1-10 and HvEDS1-11, and HvEDS1-13. MYB- and MYC-related motifs were detected at variable frequencies across multiple promoters. LTR elements were detected in five HvEDS1 genes, with higher counts observed in HvEDS1-13. TC-rich repeats were detected at low frequencies in HvEDS1-1, HvEDS1-6 to HvEDS1-8, and HvEDS1-13. The CAG-motif was found only in HvEDS1-7, whereas, JERE was the least frequent stress-responsive elements, detected only in HvEDS1-11.

Predicted tertiary structures and secondary structure variation among HvEDS1 proteins

To explore the structural diversity of the HvEDS1 gene family, tertiary structures of all 13 HvEDS1 proteins were modelled using the SWISS-MODEL homology-modelling platform (Fig. 6). The predicted models displayed Global Model Quality Estimation (GMQE) values ranging from 0.60 (HvEDS1-12 protein) to 0.91 (HvEDS1-10 protein), reflecting moderate to high confidence levels. SWISS-MODEL typically generates core domain structures and truncates unresolved N- and C-terminal regions when no suitable homologous templates are available; all 13 models were inspected for such truncation. Accordingly, several models particularly HvEDS1-2, HvEDS1-6, HvEDS1-8, and HvEDS1-12 proteins represent partial structures limited to conserved domains rather than full-length proteins. The absence of terminal regions therefore reflects template unavailability rather than intrinsic disorder.

Fig. 6.

Predicted three-dimensional structures of HvEDS1 proteins and their phylogenetic grouping. Homology-based structural models of all HvEDS1 proteins were generated using SWISS-MODEL. The structures are displayed using the default SWISS-MODEL colour scheme, which reflects the linear progression of the amino acid sequence along each model. Proteins are grouped according to their corresponding phylogenetic clades (A–D), with clade colours matching those used in Fig. 3A. GMQE (Global Model Quality Estimation) scores for each HvEDS1 model are indicated and were used to assess model quality.

All HvEDS1 proteins exhibited a conserved globular architecture typical of EDS1 family members, featuring a central α/β-hydrolase core linked to an EP domain. Peripheral regions varied in loop length, orientation, and flexibility, suggesting potential divergence in regulatory or interaction interfaces. HvEDS1-12 protein displayed the lowest GMQE score, corresponding to incomplete template coverage rather than structural instability. By contrast, HvEDS1-10 protein exhibited the most compact and well-defined fold within its resolved domain region; its N- and C-termini were, however, truncated by the modelling algorithm, indicating that the observed compactness applies to the conserved core rather than an intact full-length fold.

To assess the relationship between structural and evolutionary divergence, models were grouped according to the phylogenetic clades identified in Fig. 2. The predicted structures segregated consistently into Hv-Clade A, Hv-Clade B, Hv-Clade C, and Hv-Clade D, as defined by the phylogenetic analysis. Hv-Clade A, represented by a single member (HvEDS1-12 protein), exhibited extended loop regions and an enlarged surface area relative to other HvEDS1 proteins, likely reflecting incomplete template coverage rather than intrinsic structural expansion. Hv-Clade B proteins formed compact and well-defined α/β-folds with relatively limited loop extensions, consistent with their closer evolutionary relatedness. Hv-Clade C proteins displayed mixed structural architectures, frequently accompanied by partially unresolved N- or C-terminal extensions. Hv-Clade D, comprising the largest number of HvEDS1 members, showed substantial structural heterogeneity, characterised by variation in loop length, secondary-structure composition, and the extent of terminal truncation. This diversity likely reflects functional diversification following gene duplication events, as well as differences in template availability during modelling. Nevertheless, all Hv-Clade D proteins retained the conserved α/β-hydrolase core and EP domain, indicating preservation of the canonical EDS1 structural framework despite clade-specific divergence.

Secondary-structure analysis (Table 2) revealed considerable variation in the proportions of α-helices, β-strands, β-turns, and random coils across HvEDS1 proteins. Random coils were the most abundant structural element (ranging from 34.89% in the HvEDS1-11 protein to 51.74% in the HvEDS1-2 protein), consistent with a generally flexible conformation. α-Helices ranged from 31.69% (HvEDS1-2 protein) to 54.44% (HvEDS1-11 protein), while extended β-strands varied between 6.85% and 16.90%, being most prominent in HvEDS1-3 and HvEDS1-9 proteins. β-turns accounted for 0.65–3.05%, with HvEDS1-6 protein showing the highest proportion, indicative of local folding flexibility.

Table 2.

Predicted secondary-structure composition of HvEDS1 proteins.

Protein	α-helix (%)	Extended β-strand (%)	β-turn (%)	Random coil (%)	Structural template
HvEDS1-1	43.93	11.75	2.70	41.62	K7VXF0.1.A, triacylglycerol lipase-like 1
HvEDS1-2	31.69	14.83	1.74	51.74	Q6I5Q3.1.A, cDNA clone J013159D11
HvEDS1-3	34.19	16.52	1.71	47.58	A0A2T7F017.1.A, fungal lipase-type protein
HvEDS1-4	35.00	13.89	1.67	49.44	W5C4H6.1.A, fungal lipase-type protein
HvEDS1-5	41.98	8.02	1.91	48.09	A0A453E205.1.A, fungal lipase-type protein
HvEDS1-6	36.55	13.20	3.05	47.21	7 × 0c.1.A, phospholipase A1-IIγ
HvEDS1-7	36.29	14.57	1.43	47.71	Q6I5Q3.1.A, cDNA clone J013159D11
HvEDS1-8	53.02	6.85	0.65	39.48	8zf0.1.C, lipase-like PAD4
HvEDS1-9	37.40	16.90	1.94	43.77	K7TUH0.1.A, lipase
HvEDS1-10	33.19	13.54	2.18	51.09	8zf0.1.B, EDS1
HvEDS1-11	54.44	7.92	2.72	34.89	F2D2D4.1.A, predicted protein
HvEDS1-12	40.89	9.46	1.77	48.88	F2DI63.1.A, predicted protein
HvEDS1-13	45.59	7.96	1.72	44.73	A0A446RJL0.1.A, putative lipase

Variation in secondary-structure composition among HvEDS1 proteins reflects both evolutionary divergence and differences in template coverage across the predicted structural models; all secondary-structure contents were computed consistently from homology-based predictions, and comparisons are therefore qualitative and model-derived rather than experimental. Nevertheless, consistent patterns, including the dominance of the conserved α/β-hydrolase core and the inverse relationship between helical and coil content, support the biological relevance of the observed variation. The diversity of structural templates used for modelling further underscores the evolutionary plasticity of this family. Several HvEDS1 proteins (HvEDS1-3, HvEDS1-4, HvEDS1-5) aligned with fungal lipase-like domains, whereas HvEDS1-8 and HvEDS1-10 proteins corresponded to canonical plant immune regulators PAD4 (8zf0.1.C) and EDS1 (8zf0.1.B), respectively. This combination of plant- and fungal-like templates highlights the structural versatility of HvEDS1 proteins while preserving the conserved lipase-3–EP domain framework typical of the EDS1 family.

miRNA-mediated regulation of the HvEDS1 gene family in barley

To investigate the post-transcriptional regulation of the HvEDS1 gene family in barley, an in-silico prediction of miRNA–mRNA interactions was performed. This analysis revealed 20 interaction pairs involving 15 unique barley miRNAs targeting specific HvEDS1 transcripts (Table S4), suggesting a complex and gene-specific regulatory network. Predictions were filtered based on expectation values ranging from 3.5 to 5.0, which represent dimensionless scores derived from sequence complementarity penalties in the psRNATarget al.gorithm. These values reflect the strength of miRNA–mRNA pairing, where lower scores indicate stronger complementarity and higher values correspond to weaker interactions. These thresholds are consistent with established criteria for biologically plausible miRNA–target interactions in plants³⁵. Several miRNAs exhibited multi-targeting potential, notably hvu-miR6192, which was predicted to regulate HvEDS1-7 and HvEDS1-12, and hvu-miR6205, which targets three members (HvEDS1-3, HvEDS1-4, and HvEDS1-12). Similarly, hvu-miR6184 and hvu-miR6214 each targeted two transcripts. This multi-targeting behaviour suggests potential cross-regulatory functions within the HvEDS1 network, enabling coordinated fine-tuning of gene expression among functionally redundant or paralogous family members.

The predicted interactions were categorised according to their mode of inhibition, with the vast majority (90.5%) involving transcript cleavage, an expected outcome for miRNA–mRNA interactions in plants, while only two (9.5%) were associated with translational repression, namely HvEDS1-1, HvEDS1-5, and HvEDS1-8. This cleavage dominance is consistent with the AGO1-mediated slicing activity reported in barley and other cereals³⁶. Although unpaired energy (UPE) values were uniformly reported as − 1, likely a placeholder default, the observed complementarity patterns strongly support functional miRNA–mRNA interactions. For example, hvu-miR6214 targeted HvEDS1-10 at two discrete loci (positions 63–82 and 979–998), both displaying high complementarity and canonical pairing, indicative of efficient cleavage.

Multiplicity values showed that hvu-miR6214 and hvu-miR5053 each targeted two HvEDS1 transcripts, suggesting coordinated regulation of evolutionarily conserved paralogues. The distribution of target sites spanned the entire coding sequence, from near the 5′-end to the distal 3′-region, suggesting spatial diversity in miRNA regulation that may influence mRNA decay rates or ribosome accessibility. Several miRNAs, including hvu-miR6192, hvu-miR6205, and hvu-miR6184, have been previously associated with abiotic or biotic stress responses and were found to target HvEDS1 transcripts annotated with defence-related and signalling functions. These findings support the hypothesis that miRNA-mediated post-transcriptional silencing contributes to stress-dependent modulation of HvEDS1 expression.

Enrichment analyses of HvEDS1 genes using GO and KEGG pathway analysis

To investigate the functional roles of the HvEDS1 gene family in Hordeum vulgare, Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses were conducted, targeting biological processes (BP), molecular functions (MF), and cellular components (CC) associated with HvEDS1 genes (Fig. 7A, B). GO analysis revealed a broad spectrum of significantly enriched terms (FDR < 0.05). For clarity, only the most significantly enriched GO categories are described below, whereas the complete set of enriched terms is presented in Fig. 7A. Within the biological process (BP) category, HvEDS1 genes were predominantly associated with signal transduction (GO:0007165), defence response (GO:0006952), and phosphorylation (GO:0016310). Notably, the protein kinase C-activating G protein-coupled receptor signalling pathway (GO:0007205; p = 8.23 × 10⁻²⁷) represented one of the most significantly enriched BP terms, indicating a strong association with signalling pathways involved in stress adaptation. In addition, enrichment of the response to ozone term (GO:0010193) suggests a potential role for HvEDS1 genes in oxidative stress responses.

Fig. 7.

Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses of HvEDS1 genes. (A) Significantly enriched GO terms categorised into biological process (BP), cellular component (CC), and molecular function (MF). Bubble size represents the number of HvEDS1 genes associated with each term, and colour intensity indicates statistical significance (− log₁₀ FDR). (B) Top significantly enriched KEGG pathways associated with HvEDS1 genes. Bubble size reflects gene count per pathway, and colour scale denotes false discovery rate (FDR).

In the cellular component (CC) category, significant enrichment was observed for structures involved in lipid metabolism and signalling, including lipid droplets (GO:0005811), the endoplasmic reticulum (GO:0005783), and membrane-associated components (GO:0016020). In the molecular function (MF) category, HvEDS1 genes were significantly enriched for enzymatic activities related to lipid metabolism, such as O-acyltransferase activity (GO:0008374) and lipase activity (GO:0016298), supporting roles in lipid-mediated signalling and metabolic regulation.

KEGG pathway analysis corroborated the GO enrichment results and identified significantly enriched pathways (FDR < 0.01), with glycerophospholipid metabolism (hvu00564; FDR = 1.0 × 10⁻⁹), glycerolipid metabolism (hvu00561), and α-linolenic acid metabolism (hvu00592) representing the most highly enriched pathways (Fig. 7B). These pathways are central to membrane biogenesis and lipid-derived signalling, including jasmonate biosynthesis, further supporting the involvement of HvEDS1 genes in lipid metabolism and stress responses. In addition to pathway-level trends, gene-specific functional associations were observed. HvEDS1-3, HvEDS1-5, and HvEDS1-10 were preferentially associated with lipid metabolic pathways, whereas HvEDS1-7 and HvEDS1-12 showed stronger associations with signalling and defence-related processes. Together, these results indicate functional diversification within the HvEDS1 gene family, with individual members contributing distinctly to lipid metabolism, signal transduction, and stress regulation in barley.

Regulatory network rewiring of EDS1 gene family during fungal stress condition

Although EDS1 genes have traditionally been associated with plant immune responses and stress signalling, recent findings indicate that HvEDS1 members are also active during non-stressed growth, playing crucial roles in basal physiological regulation, cellular homeostasis, and developmental coordination. Under normal conditions, HvEDS1 paralogues exhibit diverse yet partially redundant functions. HvEDS1-2, HvEDS1-7, and HvEDS1-12 act as key regulatory hubs, targeting a wide range of genes essential for fundamental cellular processes (Fig. 8). Only HvEDS1 genes exhibiting statistically significant co-expression or regulatory interactions (|r| > 0.7, p < 0.05) were visualised, whereas members without detectable connections were excluded to enhance clarity and interpretability. In contrast, other HvEDS1 members such as HvEDS1-3, HvEDS1-5, HvEDS1-8, and HvEDS1-9 demonstrated more specific regulatory patterns, likely fine-tuning physiological pathways related to membrane transport, signalling, and metabolism.

Fig. 8.

Regulatory network of HvEDS1 genes under non-stress conditions. The network illustrates predicted regulatory interactions between HvEDS1 genes (yellow nodes) and their downstream targets in Hordeum vulgare. Each node represents a gene, and edges denote significant co-expression or regulatory associations (|r| > 0.7, p < 0.05). Only HvEDS1 genes with detectable interactions are visualised to enhance clarity. The network structure highlights modular organisation, where individual HvEDS1 paralogues regulate distinct subsets of genes involved in chloroplast development, cell cycle regulation, metabolism, and transport, suggesting a role in maintaining physiological homeostasis under normal growth conditions.

Several target genes within this network are known to be involved in vital biological processes. For example, TIP2-1, a tonoplast aquaporin, likely contributes to maintaining vacuolar water homeostasis; AHK4, a cytokinin receptor, is essential for hormone perception and cell division; STN7, a kinase involved in photosynthetic regulation, supports light adaptation and energy distribution; SERK2 and CRK23, receptor-like kinases, are implicated in developmental signalling; PYRD-2, associated with vitamin B6 biosynthesis, contributes to cellular metabolism; and ZIFL2, a transporter, may help maintain ion balance. The presence of shared target genes among multiple HvEDS1 regulators suggests a degree of transcriptional buffering or cooperative regulation. For instance, HORVU.MOREX.r3.7HG0711310 and HORVU.MOREX.r3.3HG0280530 are regulated by three or more HvEDS1 genes, underscoring their importance in maintaining cellular stability (Fig. 8). This interconnectivity likely ensures robust expression control and protection against fluctuations in developmental or metabolic signals.

Under fungal stress conditions, the HvEDS1 regulatory network functions as a coordinated system for transcriptional defence and adaptation. The HvEDS1 family is a well-established component of plant immune and stress-response pathways, particularly in salicylic acid signalling, systemic acquired resistance (SAR), and abiotic stress mitigation. In this context, multiple HvEDS1 genes especially HvEDS1-2, -3, -4, -8, -9, and − 11 regulate distinct yet overlapping sets of target genes, suggesting functional specialisation (Fig. 9). As in Fig. 8, only HvEDS1 members with statistically significant stress-responsive regulatory associations were visualised, while non-responsive genes were omitted to highlight biologically relevant interactions. Each member fine-tunes responses to specific stress types such as drought, oxidative stress, heat, or pathogen attack while also providing redundant control to enhance network robustness.

Fig. 9.

Regulatory network of HvEDS1 genes under fungal stress conditions. This network depicts predicted interactions between HvEDS1 genes (yellow nodes) and their stress-responsive target genes. Each node represents a gene, and edges indicate significant co-expression or regulatory associations inferred from expression data. Only HvEDS1 members showing significant stress-induced interactions are displayed. The network demonstrates increased connectivity and redundancy compared to the non-stressed state, with key stress-responsive genes such as SAG12, MPK9, AZF2, VDE1, and SWEET17 regulated by multiple HvEDS1 paralogues. This coordinated architecture reflects a robust transcriptional response, highlighting the adaptive role of HvEDS1 genes in defence and resilience during fungal challenge.

Many identified target genes are directly involved in stress perception, signalling, and response, including MPK9 (mitogen-activated protein kinase), AHL10 (AT-hook transcription factor), AZF2 (zinc-finger transcription factor), and NAC103, indicating the activation of signalling cascades and stress-responsive transcriptional networks. The senescence and defence markers SAG12 and SAG20, regulated by HvEDS1-2 and HvEDS1-9, highlight the network’s role in controlled cell death and nutrient remobilisation during stress. Genes associated with chloroplast and energy regulation, such as GLK1, VDE1, and CHS, suggest protective roles in photosynthesis, photoprotection, and secondary metabolite production. Additionally, transporters and detoxification-related genes, including SWEET17, AVT6E, BGLU44, and ABCA7, contribute to restoring cellular homeostasis, detoxifying reactive compounds, and maintaining osmotic balance. Genes involved in DNA repair (RECG) and protein folding (SBH1) further emphasise the importance of safeguarding genome integrity and protein quality under adverse conditions.

Notably, several genes, including HORVU.MOREX.r3.7HG0721540, HORVU.MOREX.r3.4HG0341060, and HORVU.MOREX.r3.2HG0205780, were co-regulated by multiple HvEDS1 members, suggesting their potential role as fungal stress-response hubs (Fig. 9). This redundant regulation likely evolved to ensure network stability and responsiveness, even when one or more regulators are functionally compromised. During fungal stress, HvEDS1 genes themselves are often upregulated, activating downstream gene cascades. Consequently, this regulatory network represents a fungal stress-induced expression module, wherein HvEDS1 paralogues act as upstream sensors and master regulators linking external stress signals to internal transcriptional reprogramming. The HvEDS1-centred regulatory network under fungal stress exemplifies a strategically organised transcriptional framework that governs essential processes such as signal transduction, cellular detoxification, senescence, chloroplast protection, and metabolic adaptation. Its architecture combining specialisation with redundancy ensures flexibility, robustness, and precise control of stress-responsive transcriptional regulation, ultimately enhancing the plant’s resilience and survival under hostile environmental conditions.

Discussion

The Enhanced Disease Susceptibility 1 (EDS1) gene family is central to TIR-NLR-mediated immune signaling and functions as an important integrator of basal and effector-triggered immunity (ETI) in angiosperms^10,37. Although it is well characterized in dicotyledonous species such as A. thaliana, its organization, evolutionary diversification and functional modulation in monocotyledonous plants remain largely unexplored^38,39. In this study, we report the first comprehensive genome-wide identification and characterization of EDS1 family members in barley (H. vulgare), finding 13 HvEDS1 homologues with conserved domain architectures. Remarkably, all HvEDS1 proteins possess the canonical lipase 3 and EP domains (EDS1–PAD4), structural features that facilitate EDS1 complex formation and downstream immune signaling^40,41. Despite the sequence divergence between the paralogues, the conservation of domains indicates strong purifying selection and functional constraint over evolutionary timescales. In contrast to dicotyledonous plants, where the EDS1–PAD4–SAG101 signaling module is well established, there is no identifiable SAG101 orthologue in barley, a feature consistent with other monocotyledonous plants such as rice and maize^10,21. This absence suggests that monocotyledons may have evolved alternative EDS1–PAD4 signaling interfaces or compensatory molecular mechanisms, although experimental validation is required to confirm this hypothesis.

Gene duplication analysis shows that the expansion of the HvEDS1 gene family has occurred primarily through scattered duplication, with only a single segmental duplication event (HvEDS1-2/7) detected. This finding is consistent with the calculated Ka/Ks ratios (< 1), indicating purifying selection that maintains the integrity of essential functional regions. The wide chromosomal dispersion of HvEDS1 genes combined with the absence of tandem and proximal duplications supports the role of dispersed duplication in shaping the architecture of the gene family^42,43. This distribution pattern could have arisen through transposon-mediated events or chromosomal rearrangements⁴⁴. Dispersed duplication is a well-documented driving force for the evolution of stress-responsive gene families^33,45. It contributes to both paralogue maintenance and functional innovation while preserving key defense mechanisms^46–48. Although no tandem duplications were identified, the spatial proximity of certain HvEDS1 genes, particularly those on chromosomes 2 H and 5 H, suggests possible localized coregulatory evolution. Such genomic clustering has been associated with shared cis-regulatory elements or higher order chromatin organization in other R gene clusters^49,50. The retention of lipase 3 and EP domain architecture in the HvEDS1 paralogues, combined with the substantial differences in cis-acting regulatory elements and stress-responsive expression profiles, argues for a model of under-functionalization. This partitioning likely allows HvEDS1 genes to respond to different environmental signals or developmental cues, reflecting the transcriptional diversification trends observed in the NLR gene families of cereal plants^51,52.

Cis-element analysis of HvEDS1 promoter regions revealed a diverse repertoire of hormone- and stress-responsive motifs, including ABRE, MBS, ARE and G-box elements, suggesting potential transcriptional regulation by ABA-, ROS- and JA-associated signaling pathways⁵³. This cis-regulatory complexity reflects the potential of HvEDS1 genes to be involved in multiple stress response networks. The promoters of HvEDS1-10 and HvEDS1-11 were enriched in ABRE and MBS motifs, suggesting possible involvement in ABA-mediated and drought-responsive regulatory modules through AREB/ABF-type transcription factors⁵⁴. Similarly, HvEDS1-4 and HvEDS1-6 contained G-box motifs typically recognized by bHLH transcription factors such as MYC2, which link jasmonate signaling to immune- and light-dependent pathways⁵⁵. The enrichment of GT1 and Box 4 elements in several HvEDS1 promoters also suggests that they respond to light and photoperiodic cues⁵⁶. This regulatory feature is consistent with findings from Arabidopsis, where EDS1–PAD4 expression is modulated by photoperiods and circadian rhythm^57,58. The possible convergence of hormone and light signaling pathways through common transcriptional hubs such as HY5 may represent a conserved but understudied mechanism by which barley coordinates immune readiness with environmental cycles^59,60.

In addition to transcriptional regulation, our findings suggest a potential layer of post-transcriptional control of HvEDS1 genes mediated by 15 stress-associated barley miRNAs. These predicted interactions, identified through in silico miRNA-mRNA pairing, propose that HvEDS1 expression may be fine-tuned under abiotic and biotic stresses via miRNA-guided transcript cleavage or translational inhibition. Several miRNAs including hvu-miR6205 and hvu-miR6214, both previously implicated in salt stress responses⁶¹ were predicted to target multiple HvEDS1 transcripts. Among these, HvEDS1-10 emerged as a candidate for multi-level regulation, exhibiting predicted miRNA-binding sites at multiple positions and concurrent promoter enrichment of ABA-responsive motifs. This suggests a potential dual-layered regulatory mechanism, balancing activation through ABA-responsive transcription factors and repression through miRNA-mediated silencing. Such regulatory buffering may prevent excessive or mistimed immune activation, aligning with broader models in plant stress biology where miRNAs modulate defense gene expression to maintain homeostasis while enabling rapid responsiveness^62,63. In monocots such as barley, where the canonical EDS1-PAD4-SAG101 module is partially or entirely absent, such post-transcriptional regulation may serve as an adaptive mechanism to modulate EDS1 activity with enhanced specificity⁶⁴. This highlights a potentially critical role for miRNA-guided control in shaping the plasticity and robustness of the barley immune system under fluctuating environmental conditions.

Functional predictions based on GO and KEGG analyses are limited by FDR thresholds but offer valuable hypotheses. Molecular functional terms such as lipase and phospholipase activity support the idea that HvEDS1 proteins may be involved in lipid-mediated signalling. This is consistent with studies in Arabidopsis linking EDS1-PAD4 complexes to lipidome remodeling and defense metabolite synthesis^21,65,66. Even more compelling evidence comes from the enrichment of KEGG pathways (FDR < 0.01), revealing a role in glycerophospholipid metabolism, α-linolenic acid metabolism and ubiquitin-mediated proteolysis key processes in membrane signalling, JA biosynthesis and proteostasis during stress. These results suggest that HvEDS1 proteins are involved in broader cellular stress responses, potentially bridging immune, hormone and redox signalling networks through membrane-anchored signal transduction and turnover of regulatory proteins^11,67.

The HvEDS1 gene family in barley plays a significant role in regulating gene expression in response to both normal growth conditions and fungal stress, with a clear shift in its regulatory architecture under different physiological contexts. Under normal growth conditions, HvEDS1 genes specifically HvEDS1-2, HvEDS1-3, HvEDS1-8, and HvEDS1-9 modulate a diverse set of target genes involved in crucial physiological processes. These include chloroplast development (GLK1), cell cycle regulation (CYCU4-2), metabolic pathways (HMGS, CHS, ETFA), and intracellular transport (SWEET17, AVT6E, ABCA7). The regulatory network under normal conditions is modular, where each HvEDS1 member typically regulates a distinct subset of genes, allowing precise control over metabolic and developmental pathways.

Under fungal stress conditions, however, the HvEDS1 genes exhibit a markedly different pattern of connectivity. The network becomes more highly integrated, with target genes frequently shared among multiple HvEDS1 paralogues. For example, stress-associated genes such as SAG12, MPK9, AZF2, SWEET17, RECG, and VDE1 are co-regulated by two or more HvEDS1 members. This shift towards a redundant regulatory configuration likely reflects an inherent mechanism of resilience, ensuring that essential defence responses remain active even if part of the regulatory system is compromised. These genes are linked to key stress-related functions, including signalling, senescence, detoxification, and the protection of photosynthetic machinery^11,41.

Notably, several target genes appear in both the normal and fungal stress-responsive networks. For instance, SWEET17, CHS, and VDE1 act as shared targets under both conditions, although their regulatory interactions and expression dynamics may vary according to environmental cues. Their dual presence indicates that certain genes function as central hubs, facilitating roles in both metabolism and adaptive stress responses. The continued regulation of these hub genes by multiple HvEDS1 paralogues reinforces the concept that this gene family provides flexible transcriptional control, enabling the plant to modulate its gene-expression landscape as required. The transition from a specialised, low-redundancy network under normal conditions to a highly coordinated and overlapping configuration under stress underscores the versatility of the HvEDS1 family. These genes sustain growth and metabolic homeostasis in favourable environments but rapidly shift to prioritise defence during stress exposure. Such regulatory flexibility appears to result from both evolutionary divergence and functional adaptation within the HvEDS1 family¹⁰. The HvEDS1 transcriptional network thus operates as a dynamic system that reconfigures its regulatory logic based on the plant’s physiological state. Under normal conditions, HvEDS1 genes promote growth and metabolic balance through precise, non-redundant interactions; under stress, they activate a broader and more interconnected array of defence-related genes, ensuring robust transcriptional responsiveness. This dual functionality highlights the central role of HvEDS1 paralogues in balancing growth and stress adaptation and emphasises their potential value in breeding programmes aimed at enhancing stress resilience in barley.

Conclusion

This study provides the first comprehensive genome-wide characterization of the Enhanced Disease Susceptibility 1 (EDS1) gene family in barley, elucidating its structural conservation, evolutionary trajectory, and regulatory complexity. We identified 13 HvEDS1 homologs, each harbouring the conserved lipase 3 and EP domains, indicative of strong functional constraint and purifying selection over evolutionary time. Gene duplication analysis revealed that the expansion of the HvEDS1 gene family was primarily driven by dispersed duplication events, with only a single segmental duplication (HvEDS1-2/HvEDS1-7) detected. The wide chromosomal dispersion of these genes and the absence of tandem duplications further support the role of dispersed duplication in shaping the genomic architecture of HvEDS1.

The Ka/Ks analysis of HvEDS1-2/HvEDS1-7 showed a ratio of 0.36, consistent with purifying selection, which suggests functional conservation and evolutionary constraint at both the protein and regulatory levels. Comparative synteny analysis with Arabidopsis thaliana and Oryza sativa revealed conserved syntenic relationships in monocots, while divergence between monocot and dicot lineages was observed, further highlighting the evolutionary history of the HvEDS1 family. Promoter analysis uncovered a broad spectrum of hormone- and stress-responsive cis-regulatory elements, including ABRE, MBS, ARE, and G-box motifs, pointing to the involvement of HvEDS1 genes in diverse signalling networks, such as ABA-, JA-, ROS-, and light-mediated pathways. Notably, HvEDS1-10 and HvEDS1-11 were enriched for ABA- and drought-related motifs, suggesting a role in stress adaptation. Furthermore, the predicted targeting of HvEDS1 genes by stress-responsive miRNAs such as hvu-miR6205 and hvu-miR6214 indicates a potential layer of post-transcriptional control, enabling a finely tuned balance between activation and repression under stress conditions. In addition to transcriptional regulation, our functional annotations suggest that HvEDS1 proteins are involved in lipid metabolism, proteostasis, and stress-related signalling. KEGG pathways, such as glycerophospholipid metabolism, α-linolenic acid metabolism, and ubiquitin-mediated proteolysis, support a role in integrating immune and hormone signalling via membrane-associated processes. Network analysis under normal and fungal stress conditions revealed a shift in HvEDS1 regulatory architecture, with increased redundancy in stress conditions to ensure robust defense responses. Some genes (e.g. SWEET17, CHS, VDE1) acted as central hubs, suggesting their involvement in both growth and stress response pathways.

These findings present the HvEDS1 gene family in barley as a flexible, condition-dependent regulatory module. Under normal conditions, it supports growth and metabolic homeostasis through specific gene interactions, while during stress, it adopts a more integrated network to coordinate defense responses. The discovery of stress-responsive regulatory elements and miRNA targets offers exciting possibilities for improving disease resistance and stress tolerance in barley and related cereals. The versatility of HvEDS1 genes positions them as strong candidates for genetic engineering and targeted breeding strategies. However, our conclusions are based largely on computational predictions, and experimental validation such as gene knockout, overexpression studies, and stress-induced expression profiling is essential to confirm the physiological roles of individual HvEDS1 genes. Future research should focus on experimental assays to validate the regulatory relationships and explore how HvEDS1 interacts with other components of the plant immune system. Integrating transcriptomics, proteomics, and genome-editing technologies will provide deeper insights into HvEDS1’s role in barley immunity and adaptation, laying the groundwork for the development of stress-resilient barley cultivars.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

**BP** and **RH** Conceptualization, Data curation, Formal analysis, Writing original draft. **ZGH** Investigation, Validation. **FJ** review & editing. The author(s) read and approved the final manuscript.

Funding

This study was funded by Agricultural Biotechnology Research Institute of Iran (ABRII).

Data availability

All relevant data are provided within the manuscript and the supplementary information files.

Declarations

Competing interests

The authors declare no competing interests.

Ethics approval and consent to participate

Data will be available upon reasonable request.