Genome wide identification, structural characterization, expression analysis, and regulatory network inferring of BAG gene family in barley

Abstract

The BAG (Bcl-2-associated athanogene) gene family plays a crucial role in plant stress responses by regulating programmed cell death, protein homeostasis and molecular chaperone interactions. While they have been extensively studied in model plants such as Arabidopsis and rice, the functional role and evolutionary dynamics of BAG genes in barley (Hordeum vulgare L.) are still largely unexplored. In this study, seven HvBAG genes were identified and characterised and their structural diversity, phylogenetic classification, chromosomal distribution and expression patterns were analysed. Comparative analyses with rice, maize, wheat and Arabidopsis revealed pedigree-specific expansions and conserved motifs underlining the functional importance of BAG proteins in stress adaptation. Our study revealed variations in protein stability, hydrophobicity and subcellular localisation, with HvBAG3 predominantly localised in the nucleus and HvBAG5 in the mitochondria, suggesting specialised cellular functions. Synteny and duplication analyses showed that segmental duplications and purifying selection have contributed to the expansion and evolutionary conservation of HvBAG genes. Expression profiling under heat stress revealed differential regulation in different tissues, with HvBAG3 significantly upregulated in roots and shoots, suggesting its role in heat stress resistance. Furthermore, miRNA-mediated regulation of HvBAG genes provides an additional level of post-transcriptional control that refines the mechanisms of stress response. This study provides a framework for understanding the functional diversity of BAG genes in barley and offers insights into their evolutionary history and regulatory mechanisms. These results have practical implications for the breeding of stress-resistant barley varieties and point to BAG genes as potential targets for genetic engineering strategies.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-026-57105-0.

Introduction

Barley (Hordeum vulgare L.) is one of the earliest domesticated cereal crops and continues to play a vital role in global agriculture, particularly in temperate regions. It is a primary source of food, animal feed, and raw materials for the malting and brewing industries, with global production surpassing 150 million metric tons annually¹. Despite its ability to adapt to various climatic conditions, barley productivity is increasingly threatened by abiotic stresses such as drought, salinity, and extreme temperatures^2,3. These environmental challenges significantly impair photosynthetic efficiency, disrupt nutrient uptake, and alter metabolic pathways, ultimately resulting in reduced yields⁴. Heat stress, especially during the anthesis and grain-filling stages, accelerates chlorophyll degradation, hinders carbon assimilation, and compromises kernel development, occasionally causing yield losses exceeding 30%⁵. Therefore, enhancing stress tolerance through molecular and genetic approaches is essential for ensuring sustainable barley production in a changing climate.

One promising strategy for improving stress resilience is the functional characterization of stress-associated gene families, such as the Bcl-2-associated athanogene (BAG) family. BAG proteins act as molecular co-chaperones, interacting with heat shock proteins (HSP70/HSC70) to help maintain cellular homeostasis under stress conditions⁶. In plants, BAG genes are involved in various physiological processes, including programmed cell death (PCD), stress signal transduction, and proteostasis^7,8. Based on the conserved domain architecture, plant BAG proteins are typically categorized into three structural types: (1) proteins that contain only the conserved BAG domain; (2) proteins that have an N-terminal ubiquitin-like (UBL) domain in addition to the BAG domain; and (3) proteins that have a calmodulin-binding/IQ motif in combination with the BAG domain^9,10. This structural diversification is functionally significant as the UBL domain is generally associated with proteasome-mediated degradation, whereas the IQ motif mediates calcium/calmodulin-dependent signaling. Functional studies in Arabidopsis thaliana have shown that AtBAG-4 enhances heat tolerance through regulatory interactions with stress-responsive transcription factors, while AtBAG-7 modulates unfolded protein response (UPR) pathways to alleviate endoplasmic reticulum stress^11,12. Moreover, AtBAG-6 promotes fungal resistance by facilitating autophagy-mediated defense mechanisms¹³. Similar functions have been observed in other crop species, such as rice (Oryza sativa)^14,15, maize (Zea mays)¹⁶, and tomato (Solanum lycopersicum)¹⁷, where BAG genes exhibit differential expression patterns in response to drought, salinity, and oxidative stress.

Despite increasing evidence supporting the role of BAG proteins in stress adaptation, their functional significance in barley remains largely unexplored. Given their involvement in stress response pathways, a comprehensive investigation of the BAG gene family in barley is essential. This study aims to characterize the evolutionary history, expression dynamics, and functional interactions of barley BAG proteins, providing new insights into their potential roles in abiotic stress adaptation. These findings will inform targeted breeding and genetic engineering strategies aimed at developing stress-resilient barley cultivars.

Methods and material

Identification and characterization of the BAG gene family in barley

To identify members of the BAG gene family in Hordeum vulgare, we conducted a genome-wide search using BAG protein sequences from Arabidopsis thaliana as query sequences against the Hordeum vulgare reference genome (IBSC_v3) employing the Hidden Markov Model (HMM) approach. Specifically, the BAG domain HMM profile (Pfam ID: PF02179) was retrieved from the Pfam database and used to search the barley proteome with HMMER v3.3.2 (http://hmmer.org/), applying an E-value cut-off of 1e − 5. Candidate sequences were retained only if they contained the complete BAG domain, as confirmed by SMART v9.0 (http://smart.embl.de/) and NCBI Conserved Domain Database (CDD) v3.20 (https://www.ncbi.nlm.nih.gov/Structure/cdd/cdd.shtml). We confirmed the presence of conserved BAG domains by referencing the BAG protein family domain (Pfam: PF02179) in the Pfam database v35.0 (https://pfam.xfam.org/). For the identification of additional domain classes, we also screened for the ubiquitin-like (UBL) domain (Pfam: PF00240) and the calmodulin-binding/IQ motif (Pfam: PF00612) using the Conserved Domain Database (CDD) and Pfam. This ensured accurate classification of HvBAG proteins into the three recognized structural types—BAG-only, UBL + BAG, and IQ/calmodulin-binding + BAG. For comparative analysis, we retrieved BAG protein sequences from various plant species, including Oryza sativa, Zea mays, Arabidopsis thaliana, and Sorghum bicolor using the Ensembl Plants BioMart tool (https://plants.ensembl.org/biomart/martview/). The coding sequences (CDS), protein sequences, and genomic data for H. vulgare were sourced from the Ensembl Plants database release 59 (https://plants.ensembl.org/). To further characterize the identified BAG proteins, we analyzed physicochemical properties, such as molecular weight (kDa), isoelectric point (pI), and protein length (in amino acids), using the ExPASy ProtParam tool (https://web.expasy.org/protparam/).

Subcellular localization prediction of HvBAG proteins

The subcellular localization of HvBAG proteins was predicted in silico using three complementary bioinformatic tools: Plant-mPLoc v2.0 (http://www.csbio.sjtu.edu.cn/bioinf/plant-multi/), TargetP v2.0 (http://www.cbs.dtu.dk/services/TargetP/), and WoLF PSORT v0.2 (https://wolfpsort.hgc.jp/). Each protein sequence was analyzed independently in all three tools, and a consensus prediction was recorded when at least two tools predicted the same primary localization. This majority-rule approach increases prediction robustness while accounting for algorithm-specific variability. The combined predictions provided insights into potential compartment-specific functions of HvBAG proteins and facilitated comparison with experimentally validated BAG localizations in other plant species.

Gene structure, domain architecture and promoter analysis in Hordeum vulgare

The genomic and coding sequences of the identified HvBAG genes were analyzed using the Gene Structure Display Server (GSDS) v2.0 (http://gsds.gao-lab.org/) to determine exon–intron organization. Conserved protein domains were identified using the NCBI Conserved Domain Database (CDD) v3.20 (https://www.ncbi.nlm.nih.gov/Structure/cdd/cdd.shtml) and aligned with Pfam (PF02179) for confirmation. For promoter analysis, 2 kb upstream sequences (relative to the annotated transcription start sites) were retrieved from Ensembl Plants (IBSC_v3). Promoter sequences were scanned for transcription factor binding sites (TFBS) and cis-regulatory motifs using a multi-database approach to ensure contemporary coverage and cross-validation. Specifically, (i) JASPAR 2024 CORE Plant database (https://jaspar.elixir.no/) profiles were scanned using FIMO from MEME Suite v5.5.5 (https://meme-suite.org/meme/tools/fimo) with a significance threshold of p < 1 × 10⁻⁴ and Benjamini–Hochberg correction; (ii) PlantPAN v4.0 (http://plantpan.itps.ncku.edu.tw/) was interrogated using its web interface; and (iii) PlantRegMap (http://plantregmap.gao-lab.org/) predictions were used to complement the PWM-based calls. The old PlantCARE database¹⁸(http://bioinformatics.psb.ugent.be/webtools/plantcare/html/) was also consulted to ensure comparability with previous literature. To assess concordance between resources, we assigned motif calls to promoter coordinates and considered a motif to be concordant if the same TF family or PWM was detected in at least two databases within a ± 25 bp window. Redundant motif calls were summarized by TF family (e.g. MYB, bZIP, NAC) in order to subsequently categorize them. Motifs were grouped by functional categories (hormone-, stress-, light- and development-related) according to the PlantPAN/PlantCARE ontologies and frequencies were log normalized. Graphical representations of gene structure, domain composition and promoter element distribution were generated using TBtools v1.108 (https://github.com/CJ-Chen/TBtools/releases).

Phylogenetic analysis and sequence logos in barley

A phylogenetic tree was constructed using Molecular Evolutionary Genetics Analysis X (MEGA-X) v10.2.6 (https://www.megasoftware.net/) to compare members of the BAG gene family across five plant species H. vulgare, O. sativa, Z. mays, A. thaliana, and S. bicolor. BAG protein sequences for each species were identified via BLASTp searches against UniProtKB (https://www.uniprot.org/) or Ensembl Plants, followed by confirmation of the BAG domain (Pfam: PF02179) using the Conserved Domain Database (CDD) and HMMER. Multiple sequence alignments were performed using the MUSCLE algorithm implemented in MEGA-X, and the phylogenetic tree was generated using the Maximum Likelihood method with the JTT substitution model. Bootstrap support values were calculated from 1,000 replicates to assess branch reliability. Conserved amino acid motifs were visualized with the WebLogo v3.7.12 online tool (https://weblogo.berkeley.edu/), highlighting key functional regions of BAG proteins in A. thaliana and H. vulgare.

Chromosomal localization, Ka/Ks ratio, and collinearity analysis

To analyze the evolutionary dynamics of the BAG gene family members in barley, non-synonymous (Ka) and synonymous (Ks) substitution rates were calculated for duplicated gene pairs using TBtools v1.108. The divergence time (in millions of years ago) was estimated using the formula T = Ks/(2λ), where λ = (2 × 6.5 × 10⁻⁹) × 10⁶. Collinearity relationships between H. vulgare and other species (O. sativa and A. thaliana) were analyzed and visualized using TBtools to investigate conserved syntenic regions and potential evolutionary events that have shaped the BAG gene family in barley.

Expression pattern of HvBAGs under high temperature stress

For the expression analysis of HvBAGs genes, RNA-Seq data related to heat stress in barley were retrieved from the NCBI Sequence Read Archive (SRA) database (https://www.ncbi.nlm.nih.gov/sra). Specifically, the dataset PRJNA324116 was chosen to examine the gene expression profiles of HvBAGs under heat stress conditions in shoot and root tissues. The study of shoot and root tissues enhances our understanding of tissue-specific regulatory mechanisms governing HvBAGs and stress adaptation. For data analysis, raw RNA-Seq datasets were retrieved and the initial quality assessment of the data was performed using FastQC v0.11.5 (https://www.bioinformatics.babraham.ac.uk/projects/fastqc/) to evaluate sequence quality, GC content, and potential adapter contamination. To ensure the integrity of the data, adapter sequences and reads with a Phred quality score below 30 were removed using Trimmomatic v0.32 (http://www.usadellab.org/cms/?page=trimmomatic). The cleaned reads were subsequently mapped to the barley reference genome using Hierarchical Indexing for Spliced Alignment of Transcripts 2 (HISAT2) v2.2.1 (https://daehwankimlab.github.io/hisat2/). Following this, transcript abundance was quantified using HTSeq-count v0.13.5 (https://htseq.readthedocs.io/en/master/count.html), enabling downstream expression analysis. Expression levels were quantified as fragments per kilobase of exon per million mapped reads (FPKM) and subsequently visualized using TBtools v1.108 through heat map representations to facilitate comparison of expression profiles under heat stress conditions.

Validation of expression pattern by Leave-One-Out Cross-Validation (LOOCV) approach

To evaluate the expression patterns and infer the potential functional relevance of the predicted genes, an in-silico validation was conducted utilizing the Leave-One-Out Cross-Validation (LOOCV) framework. This approach was selected for its suitability in small-sample contexts and its ability to rigorously assess the predictive contribution of features at the single-sample level. For each tissue type (shoot and root), expression matrices were organized with samples as observations and genes as predictor variables. The classification task was defined as a binary comparison between the Control and Treated groups. Prior to model fitting, all features were standardized using the StandardScaler function implemented in the scikit-learn machine learning library v1.3.0 (https://scikit-learn.org/) to eliminate scale-related bias and ensure equal weighting across variables. Throughout the LOOCV iterations, predicted class probabilities were aggregated to generate Receiver Operating Characteristic (ROC) curves, from which Area Under the Curve (AUC) values were computed as a quantitative measure of model discrimination. Overall classification accuracy was assessed by comparing predicted versus actual sample classes across all iterations.

Prediction of putative miRNAs targeting HvBAG genes and functional annotation

To identify potential microRNAs (miRNAs) that target H. vulgare BAG (HvBAG) genes, mature barley miRNA sequences were retrieved from publicly available databases. The coding sequences of the identified HvBAG genes were analyzed using the psRNATarget server (2017 release; https://www.zhaolab.org/psRNATarget/) with default parameters. The interaction network between the predicted miRNAs and HvBAG genes were constructed and visualized using Cytoscape v3.9.1 (https://cytoscape.org/). For functional annotation, the protein sequences of all HvBAG genes were submitted to the eggNOG-mapper v2 database (http://eggnog-mapper.embl.de/) to retrieve Gene Ontology (GO) terms and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway annotations. KEGG pathway mapping and annotation were interpreted according to the KEGG resource developed by Kanehisa Laboratories^19,20. Further enrichment analysis was carried out, and the results were graphically represented using TBtools.

BAG gene regulatory network inference using the machine learning approach

To construct the gene regulatory network, the Random Forest regression algorithm was employed on gene expression matrix from genes with a p-value > 0.05 in both shoot and root tissues. This method facilitated the identification of potential regulatory relationships under both normal and heat stress conditions. At each decision tree node, a random subset of candidate predictors was selected (K = “sqrt”), corresponding to the square root of the total number of genes. This approach optimized the balance between computational efficiency and predictive accuracy. The model was trained with an ensemble of 1000 decision trees (nTrees = 1000), ensuring robust and stable predictions of regulatory interactions. Further, the inferred network was analyzed to explore the rewiring of regulatory networks associated with BAG genes across different conditions and tissues. By comparing interactions under normal and heat stress, we identified key regulatory shifts that provide insights into the dynamic modulation of BAG gene functions. Additionally, functional annotation and enrichment analyses were conducted to elucidate the biological significance of the identified regulatory modules, highlighting their potential roles in stress adaptation and response mechanisms.

Results

Identification and characterization of the BAG gene family in barley

To systematically identify members of the BAG gene family in the barley genome, a comprehensive analysis was performed using BLASTp searches in combination with HMM-based approaches, using reference BAG protein sequences from the model plants rice and Arabidopsis. As a result, seven BAG genes encoding proteins with different sequence lengths ranging from 194 to 1317 amino acids were identified (Table 1). These genes were designated as HvBAG1–HvBAG7 according to their physical positions on the barley reference genome (Morex v3): numbering followed ascending chromosome number (2 H to 7 H), and for multiple genes on the same chromosome, ascending genomic coordinate from the short (p) to the long (q) arm. This chromosomal-order convention is widely adopted in plant gene-family studies (e.g. CER genes in cotton²¹ and ensures transparency, reproducibility, and ease of comparative mapping. Among these, HvBAG5 was the shortest, while HvBAG3 was the longest protein. The molecular weights ranged from 20.79 to 144.08 kDa, while the predicted isoelectric points ranged from 5.01 to 10.97. Based on conserved domain architecture, the identified HvBAG proteins were classified into three recognized structural types: (1) BAG-only proteins (HvBAG1, HvBAG4, HvBAG5), (2) UBL + BAG proteins (HvBAG2, HvBAG3), and (3) calmodulin-binding/IQ + BAG proteins (HvBAG6, HvBAG7) (Table S1, Supplementary Fig. 1). The UBL domain is typically associated with proteasome-mediated protein degradation and quality control, whereas the IQ motif participates in calcium/calmodulin-mediated signaling pathways, often linked to abiotic stress responses. BAG-only proteins act primarily as co-chaperones that directly regulate HSP70/HSC70 activity without additional domain-mediated signaling. This classification reflects the evolutionary and functional diversification of the BAG gene family in plants.

Table 1.

Basic information of BAG gene family members in barley.

Gene Name	Gene ID	GC Content (%)	Protein Length (aa)	Molecular Weight (KDa)	Isoelectric Point	GRAVY	Instability Index	Aliphatic Index	Transcript Length (bp)
HvBAG1	HORVU.MOREX.r3.2HG0195140	54.15	268	29.15	9.4	− 0.448	47.01	81.53	1191
HvBAG2	HORVU.MOREX.r3.3HG0303840	42.88	259	28.93	5.68	− 0.562	35.73	77.88	1016
HvBAG3	HORVU.MOREX.r3.4HG0418350	48.09	1317	144.08	5.01	− 0.956	55.86	55.19	4228
HvBAG4	HORVU.MOREX.r3.5HG0497770	61.26	349	37.67	9.71	− 0.543	40.67	72.55	1779
HvBAG5	HORVU.MOREX.r3.6HG0609940	68.93	194	20.79	10.97	− 0.28	63.46	77.63	782
HvBAG6	HORVU.MOREX.r3.7HG0642360	63.02	357	38.25	9.79	− 0.525	57.57	76.39	1829
HvBAG7	HORVU.MOREX.r3.7HG0686900	59.19	323	35.37	9.63	− 0.554	52.75	73.16	1647

In addition, the instability index values varied between 35.73 and 63.46, indicating differences in protein stability under physiological conditions. The GRAVY (Grand Average of Hydropathy) values ranged from − 0.956 to -0.28, indicating different levels of hydrophobicity among the HvBAG proteins. In addition, the aliphatic index ranged from 55.19 to 81.53, reflecting variations in thermostability.

The transcriptional properties of the HvBAG genes were also analyzed, including transcript start and end positions, transcription start site (TSS) and transcript length (including untranslated regions (UTRs) and coding sequences (CDS)) (Table 1 & Supplementary Table S1). Furthermore, in silico predictions of the subcellular localization of HvBAG proteins generated using three complementary bioinformatics tools (Plant-mPLoc, TargetP and WoLF PSORT) indicate their distribution in different cellular compartments, including cytoplasm, nucleus, mitochondria, chloroplasts, peroxisomes and plasma membrane (Fig. 1; Supplementary Table S2). Predictions were consistent across all tools, with HvBAG3 most frequently localized to the nucleus, a compartment typically associated with transcriptional regulation and integration of stress signals, while HvBAG5 was most likely localized to the chloroplasts, suggesting a possible role in stress responses associated with photosynthesis. These results are consistent with experimentally validated localizations by BAG orthologues in A. thaliana^9,10 and Solanum lycopersicum¹⁷, which show similar compartmentalization patterns, supporting the plausibility of our predictions. However, the computational predictions fail to capture potential stress- or tissue-specific re-localization events. Although direct experimental validation in barley (e.g. transient GFP fusion expression) is beyond the scope of the present study, we have highlighted this as a limitation and indicated that such experiments are planned in future work to confirm the predicted localizations and refine functional hypotheses.

Fig. 1.

Heat map depicting the predicted subcellular localization of all HvBAG genes within various cellular compartments, including the nucleus, cytoplasm, chloroplast, Golgi apparatus, mitochondria, plasmodesmata, and peroxisomes. The colour scheme represents the relative functional presence of each gene in the respective organelle: grey indicates the absence of the gene in a given compartment, blue signifies minimal functional presence, while pink denotes maximal functional significance, highlighting the predominant localisation and potential activity of the HvBAG genes within the plant cell.

Gene structure, conserved motifs, domains, and chromosomal localization of the BAG gene family in barley

To investigate the evolutionary relationships between the barley BAG (HvBAG) gene family, a phylogenetic tree was constructed showing the division of the HvBAG genes into three different clades (Fig. 2A). This clustering pattern suggests an evolutionary diversification of HvBAG genes, possibly leading to functional divergence. Genes within the same clade may have conserved functions, while those located in separate groups may have undergone structural and functional differentiation to adapt to specific cellular processes in barley.

Fig. 2.

Phylogenetic relationships, gene structure, Domain analysis, conserved motifs, and motif sequences of the BAG gene family in barley. (A) Phylogenetic analysis of HvBAG genes. (B) Gene structure analysis. (C) Domain analysis, (D) Conserved motif distribution. (E) Corresponding motif sequences of HvBAG genes in barley.

Structural analysis of the HvBAG genes revealed considerable variation in exon-intron organization, illustrating the evolutionary complexity of this gene family (Fig. 2B). The number of exons varied among the HvBAG genes, with some members exhibiting a compact gene structure while others contained extensive intronic regions. This structural diversity suggests that exon-intron shuffling, gain or loss contributed to the functional specialization of HvBAG genes. Interestingly, genes within the same clade tend to have more conserved exon-intron patterns, emphasizing their evolutionary relatedness and potential functional similarity.

Analysis of the conserved domains within the HvBAG proteins (Fig. 2C) revealed the presence of important functional domains associated with molecular chaperone activity. Along with the canonical BAG domain, additional structural features, including ubiquitin-like motifs, were identified in five HvBAG genes (clades I and III), suggesting a potential expansion of their functional roles beyond traditional BAG-mediated protein homeostasis. Conversely, clade II genes exhibited additional coiled-coil regions that may enhance protein-protein interactions, further supporting the hypothesis of functional diversification within the family. This variation in domain architecture suggests that HvBAG proteins may have specialized roles under different stress conditions and developmental processes in barley.

Motif analysis identified several conserved motifs in the HvBAG proteins, highlighting their evolutionary conservation and functional significance (Fig. 2D). Motifs 1, 3 and 7 were present in almost all HvBAG proteins, suggesting their fundamental role in BAG protein function, while certain motifs were only present in specific clades, indicating possible functional specialization within subgroups. The conservation of these motifs emphasizes their possible involvement in cellular stress regulation and protein quality control mechanisms. Further motif sequence analysis (Fig. 2E) provided insights into the amino acid composition of the conserved domains and revealed key residues that are probably crucial for the function of the BAG protein. The sequence logos show a high degree of conservation between essential amino acids within these motifs, suggesting their importance in maintaining protein stability and mediating protein-protein interactions. The observed sequence divergence between different HvBAG proteins may indicate adaptive modifications possibly related to specific stress responses or regulatory pathways in barley. Overall, the integrated phylogenetic, structural and motif analyses provide compelling evidence for the evolutionary diversification of the HvBAG gene family leading to both conserved and specialized functional roles.

Barley BAG protein’s phylogenetic classification

Using the protein sequences of Hordeum vulgare (Hv), Oryza sativa (Os), Arabidopsis thaliana (At), Zea mays (Zm), and Sorghum bicolor (Sb), we constructed a phylogenetic tree that classified the BAG gene family into four distinct clades (Fig. 3). For Arabidopsis and rice, eight BAG proteins were identified through BLASTp searches against UniProt, followed by confirmation of the BAG domain (Pfam: PF02179) using the Conserved Domain Database (CDD) and HMMER. This result is consistent with the recent genome-wide analysis by ¹⁷, replacing earlier reports of six BAG members in each species. For barley, seven BAG proteins were identified in the present study using the same BLASTp–CDD–HMMER pipeline. For maize (Zea mays), 21 BAG protein sequences were retrieved from the Ensemble Plants database and validated for the BAG domain using CDD and HMMER. This comparatively high number, which exceeds previous monocot BAG family reports, may reflect lineage-specific expansion, alternative isoforms, or differences in genome annotation. For sorghum (Sorghum bicolor), eight BAG proteins were identified using the same approach. Applying an identical identification procedure to all species ensured methodological consistency and accurate cross-species comparison. The complete list of BAG protein sequences for all species used in the phylogenetic analysis is provided in Supplementary file S1 for full transparency.

Fig. 3.

Phylogenetic analysis of the BAG gene family in Hordeum vulgare (Hv), Oryza sativa (Os), Arabidopsis thaliana (At), Zea mays (Zm), and Sorghum bicolor (Sb). The tree was constructed using the maximum likelihood method with 1,000 bootstrap replicates. Branch values represent bootstrap support percentages, while branch lengths indicate the number of amino acid substitutions per site. Coloured arcs represent the four phylogenetic clades: Clade I (magenta), Clade II (dark blue), Clade III (orange), and Clade IV (green). The BAG domain (Pfam: PF02179) was confirmed for all sequences using the Conserved Domain Database (CDD) and HMMER.

The phylogeny illustrates the evolutionary relationships between these BAG genes, highlighting both conserved and divergent patterns between monocotyledons and dicotyledons. Clade I (magenta) contains BAG genes from both dicots and monocots, including AtBAG-4, which clusters separately from monocot members, suggesting functional divergence. Clade II (dark blue) separates monocot and dicot genes, with HvBAG and OsBAG genes clustering closely, implying conserved roles distinct from AtBAG homologues. Clade III (orange) is largely monocot-specific but includes certain Arabidopsis BAG members, suggesting ancient conservation of specific functions. Clade IV (green) is almost exclusively monocot, containing only one Arabidopsis representative, and likely represents a monocot-specific lineage that has undergone lineage-specific specialization, possibly in stress tolerance and protein homeostasis.

Chromosomal distribution, collinearity, and Synteny analysis of the HvBAG gene family in barley

The chromosomal distribution of seven HvBAG genes in Hordeum vulgare was identified and showed scattered localization across five chromosomes (Fig. 4A). Specifically, the HvBAG genes were assigned to chromosomes 2, 3, 4, 5, 6 with one gene each and chromosome 7, with chromosome 7 harboring two genes (HvBAG-6 and HvBAG-7). This revised number of seven genes corrects the earlier underestimation. This scattered arrangement suggests that the expansion of HvBAG genes in barley was primarily caused by segmental duplications rather than tandem duplications, as no physically clustered gene pairs were identified.

Fig. 4.

Chromosomal distribution, collinearity, and synteny analysis of HvBAG genes in barley. (A) Chromosomal localization of HvBAG genes in barley. The chromosomes are represented as vertical bars, with gene positions marked in red. (B) Collinearity analysis of HvBAG genes in barley. The circular diagram illustrates collinear gene pairs connected by red lines, indicating potential evolutionary relationships. (C) Synteny analysis of HvBAG genes among Arabidopsis thaliana, Hordeum vulgare, and Oryza sativa. Syntenic gene pairs are highlighted with red connecting lines.

To investigate the evolutionary relationships within the HvBAG gene family, a collinearity analysis was performed to detect duplicated gene pairs within the barley genome (Fig. 4B). This analysis identified two pairs of collinear HvBAG genes, suggesting that their expansion occurred mainly through segmental duplications. The collinear pairs detected include HvBAG-3 ↔ HvBAG-5 and HvBAG-4 ↔ HvBAG-7, with an additional scattered duplication event between HvBAG-1 and HvBAG-6. These gene duplication patterns are clearly depicted in Fig. 4B, emphasizing the role of segmental duplications in the evolution of the HvBAG family. The identification of these collinear pairs suggests that whole genome duplication (WGD) or large-scale chromosomal rearrangements have contributed to the evolution of the HvBAG family in barley. In particular, the absence of tandem duplications supports the hypothesis that segmental duplications and not local gene duplications have played a predominant role in the expansion of the HvBAG gene family.

Further insight into the selection pressure acting on these genes was gained by analyzing the Ka/Ks ratio, which revealed that all identified gene pairs have a Ka/Ks ratio of less than 1, indicating that they are predominantly subject to purifying selection (Table 2). The HvBAG-3/HvBAG-5 pair had a Ka/Ks ratio of 0.86, indicating a relatively high rate of non-synonymous substitutions that may be indicative of functional divergence. Estimated duplication events range between 8.5 and 11.28 million years ago (mya), with segmental duplications being older (~ 11 mya) than dispersed duplications (~ 8.5–9 mya). This evolutionary timeframe emphasizes the long-term conservation of HvBAG genes and their functional importance for barley adaptation. The presence of scattered duplications suggests that certain HvBAG genes may have been duplicated by transposons or genome rearrangements, contributing to the structural and functional diversity of this gene family.

Table 2.

Nonsynonymous and synonymous (Ka/Ks) substitution rate.

Gene I	Gene II	Ka	Ks	Ka/Ks	Type of Duplication	Time (mya)
HvBAG4	HvBAG7	0.03	0.09	0.33	WGD or Segmental	11.28
HvBAG3	HvBAG5	1.34	1.56	0.86	WGD or Segmental	11.14
HvBAG6	HvBAG1	0.03	0.16	0.19	Dispersed	8.5
HvBAG1	HvBAG5	0.05	0.18	0.28	Dispersed	9.16

To investigate the evolutionary relationships between the BAG genes of barley, Arabidopsis and rice, a synteny analysis was performed (Fig. 4C). The results showed a stronger syntenic relationship between barley and rice compared to Arabidopsis, consistent with their closer phylogenetic relationship as monocotyledons. A total of three orthologous BAG gene pairs were identified between barley and rice, while only two orthologous pair was found between barley and Arabidopsis. This can be clearly seen in Fig. 4C, where red lines indicate the conserved BAG gene synteny between the species. The extensive syntenic relationship between barley and rice suggests that the BAG genes in these species have retained homologous genomic regions despite evolutionary divergence, supporting the hypothesis that monocotyledons have a more recent common ancestor compared to dicotyledons. The relatively weaker syntenic relationship between barley and Arabidopsis may suggest that the BAG genes in dicotyledons have undergone independent evolutionary changes leading to functional diversification. Furthermore, the presence of multiple collinear BAG gene pairs between barley and rice suggests functional conservation and evolutionary stability within monocotyledons. The retention of these conserved genomic regions suggests a possible role for BAG genes in fundamental cellular processes, particularly stress response and protein homeostasis.

Examination of the Cis-acting elements in Barley’ BAG genes

The study of cis-acting regulatory elements within gene promoter regions is fundamental to understanding transcriptional regulation and the functional role of gene families (Hamid et al., 2024). To obtain a robust and contemporary map of putative cis-regulatory motifs, we scanned the 2 kb upstream promoter regions of HvBAG genes using a multi-database approach: JASPAR 2024 CORE (plant) via FIMO (p < 1 × 10⁻⁴, Benjamini–Hochberg correction)²², PlantPAN 4.0 [PWM-based and ChIP-based predictions]²³, and PlantRegMap [functional TF target predictions and motif conservation]²⁴, while retaining PlantCARE¹⁸ for historical comparability. The multi-database strategy confirmed the majority of motifs originally detected by PlantCARE and yielded additional TFBSs absent from the older repository. Quantitatively, 85% of high-confidence motifs were detected in at least two independent databases, while 2% were found in only one platform (Supplementary Table S3). The overlaps between the databases are illustrated in the newly added Venn diagram (Fig. 5A). The identified motifs were categorized into four primary functional groups: hormone-responsive, growth- and development-responsive, stress-responsive, and light-responsive elements (Fig. 5B; Supplementary Table S3).

Fig. 5.

Cis-regulatory element analysis in HvBAG promoters. (A) Venn diagram illustrating the overlap of predicted cis-elements identified by PlantCARE, JASPAR 2024 CORE, and PlantPAN 4.0 (motif-based and TFBS-based searches), with overlapping areas indicating shared motifs and non-overlapping areas indicating database-specific predictions. (B) Distribution of elements within the 2000 bp upstream regions of HvBAG genes, colour-coded as stress-responsive (green), light-responsive (yellow), hormone-responsive (pink), and cell cycle & development-related (blue). (C) Functional categorisation matrix showing the number of elements in each promoter.

Among the hormone-responsive elements, abscisic acid-responsive elements (ABRE) were the most abundant, suggesting involvement in ABA-mediated stress responses (Fig. 5C). In addition, gibberellin-responsive elements, including the GARE motif, P-box, and TATC-box, were detected, indicating a possible role in growth regulation. The presence of ethylene-responsive elements (ERE) and auxin-responsive elements (AuxRR core) suggests that HvBAG genes may be involved in multiple hormones signaling pathways. The frequent occurrence of CGTCA and TGACG motifs associated with methyl jasmonate (MeJA) signaling was confirmed by all four databases, while PlantPAN 4.0 and JASPAR 2024 additionally identified NAC-, bZIP-, and ARF-family TFBSs, which are absent in PlantCARE, strengthening the link between HvBAG regulation and hormone-mediated stress adaptation.

Stress-responsive elements were prevalent in HvBAG promoters, highlighting their potential role in adaptation to environmental stress. Prominent motifs included anaerobic response elements (ARE), MYB-binding sites (MBS) for drought response, stress response elements (STRE), low-temperature response elements (LTR), and TC-rich repeats associated with defense and stress responses (Fig. 5C). MYB and MYC binding sites known to regulate responses to abiotic and biotic stresses were highly abundant, suggesting possible regulatory interactions with MYB/MYC transcription factors. The presence of W-box elements, which serve as binding sites for WRKY transcription factors, suggests that HvBAG genes may play a role in pathogen defense. Other stress-related motifs such as WRE3 and CCGTCC boxes were also identified, supporting the hypothesis that HvBAG genes contribute to stress tolerance. Although no classical heat shock elements (HSE consensus nGAAn) were detected below the stringent threshold, JASPAR 2024 and PlantRegMap predicted potential HSF-related motifs at lower information content scores, and MYC elements known to interact with heat stress-responsive factors were frequent, suggesting indirect regulation in heat stress. The JASPAR dataset also identified potential binding motifs for DREB/CBF proteins and members of the ERF/AP2 family, important regulators of abiotic stress responses, providing additional evidence for HvBAG involvement in adaptation to heat, cold, and oxidative stress.

Among the growth- and development-related elements, multiple copies of the CAT box (associated with meristem expression), the RY element (involved in seed-type regulation), and the O₂ site (associated with the regulation of zein metabolism) were identified, suggesting a possible role in embryogenesis and seed development (Fig. 5C). In addition, the discovery of circadian regulatory elements suggests that some HvBAG genes may be subject to circadian control, which may be important for temporal gene expression in response to environmental stimuli. Light-dependent elements were also widely distributed across the promoters of HvBAG genes, suggesting a possible role in light-dependent gene regulation. The most frequently identified motifs included Box 4, ACE, and G-box, all of which are involved in light-dependent transcriptional activation (Fig. 5C). Additional light-dependent TFBSs predicted by PlantPAN 4.0 and JASPAR 2024 included HY5- and GBF-binding sites, suggesting that these promoters are involved in photomorphogenesis and light-driven stress signaling. In summary, the combination of PlantCARE with recently curated TFBS resources (PlantPAN 4.0, JASPAR 2024, and PlantRegMap) not only confirmed the identification of core motifs, but also revealed a broader repertoire of cis-elements, providing a more comprehensive and timely understanding of HvBAG transcriptional regulation. The integration of several independent databases, together with concordance statistics and complete motif lists (see Table S3), increases the reliability of the promoter analysis and brings the study in line with current bioinformatics standards in plant regulatory genomics.

3D structural modelling of HvBAG proteins

To gain insight into the structural diversity and functional significance of the HvBAG gene family, we performed tertiary protein structure prediction using the Phyre2 online server. The predicted models were generated using homology modeling, with GMQE values ranging from 0.67 to 0.75, indicating high structural reliability (Fig. 6). The tertiary structures showed a conserved alpha-helical BAG domain across all HvBAG members, accompanied by variable beta-sheet and loop regions that could contribute to functional diversification.

Fig. 6.

Predicted Three-Dimensional Structures of HvBAG Proteins in Barley. Structural models of the seven HvBAG proteins (HvBAG-1 to HvBAG-7) were generated using homology-based modelling. The ribbon representations illustrate the conserved BAG domain, depicted in a spectrum colour gradient from the N-terminal (blue) to the C-terminal (red). The Global Model Quality Estimation (GMQE) values for each model are indicated in parentheses, reflecting the reliability of the predicted structures.

We also investigated the secondary structural composition of HvBAG proteins and identified four major structural elements: Alpha helices, extended strands, beta-turns and random coils (Table 3). The proportion of these elements varied considerably between HvBAG members, suggesting possible structural and functional divergence. Alpha helices were the predominant structural feature and ranged from 34.38% in HvBAG4 to 56.7% in HvBAG5, emphasizing their essential role in protein-protein interactions. In contrast, elongated strands were relatively rare, with HvBAG3 having the lowest proportion (1.52%) and HvBAG-1 the highest (12.69%). Beta-turns were rare, accounting for 0.68% to 4.48% of the total structure, while random coils were the second most abundant element, ranging from 35.82% to 58.47%.

Table 3.

Secondary structure composition of HvBAG proteins.

Gene Name	Alpha Helix (%)	Extended Strand (%)	Beta Turn (%)	Random Coil (%)	Template with Highest Similarity
HvBAG1	47.01	12.69	4.48	35.82	4hwi.1.B (BAG family molecular chaperone regulator 1)
HvBAG2	43.24	8.88	3.86	44.02	4hwi.1.B (BAG family molecular chaperone regulator 1)
HvBAG3	39.33	1.52	0.68	58.47	4hwd.1.A (BAG family molecular chaperone regulator 2)
HvBAG4	34.38	8.02	2.87	54.73	4hwi.1.B (BAG family molecular chaperone regulator 1)
HvBAG5	56.70	0.00	1.03	42.27	4hwh.4.A (BAG family molecular chaperone regulator 4)
HvBAG6	36.13	8.12	2.52	53.22	4hwi.1.B (BAG family molecular chaperone regulator 1)
HvBAG7	34.67	9.29	3.72	52.32	4hwi.1.B (BAG family molecular chaperone regulator 1)

Among the individual proteins, HvBAG3 had the highest content of random coils (58.47%), suggesting a highly flexible structure that could facilitate dynamic conformational changes. In contrast, HvBAG5 had the highest alpha-helix content (56.7%) and had no elongated strands, suggesting a more rigid and stable conformation. The predominance of random coils in most HvBAG members suggests that these proteins retain structural flexibility, a property that is essential for their interactions with molecular chaperones and stress-related proteins.

The modelling results also showed that HvBAG proteins have structural similarities within closely related subgroups, indicating a conserved evolutionary origin. The BAG domain was predominantly helical, consistent with its known role in binding heat shock proteins (HSPs) and modulating stress response pathways. Of note, most HvBAG proteins were modelled with the template 4hwi.1.B (molecular chaperone regulator 1 of the BAG family), while HvBAG3 matched 4hwd.1.A (molecular chaperone regulator 2 of the BAG family), suggesting a possible functional divergence within the family (Fig. 6). These results emphasize the structural adaptability of HvBAG proteins, which likely contributes to their functional role in stress adaptation, protein quality control and cell survival mechanisms. The observed variations in secondary structure composition and tertiary organization highlight the complexity of the evolution of the HvBAG gene family, which may have contributed to their molecular functions in barley stress tolerance.

Interaction between HvBAG Genes and miRNAs in Barley

The interaction between the HvBAG gene family and microRNAs (miRNAs) was systematically analysed to explore post-transcriptional regulatory mechanisms affecting gene expression. The results revealed several miRNA targets within the HvBAG gene family that are predominantly associated with cleavage-mediated inhibition, indicating their role in transcriptional degradation and turnover (Supplementary Table S4).

Among the identified interactions, hvu-miR6192 exhibited significant target specificity toward HvBAG-7, which matched the target site at positions 1337–1358. The alignment suggests a strong cleavage-mediated regulatory mechanism characterized by highly conserved miRNA-target pairing with minimal sequence mismatches. Similarly, hvu-miR6196 was predicted to be the target of HvBAG-6 with an alignment spanning positions 1526–1546, further supporting miRNA-directed cleavage as a key regulatory process. Of note, hvu-miR168-5p was predicted to interact with HvBAG-1, albeit with a slightly higher expected value (4.5), indicating a comparatively weaker binding affinity. The nature of this interaction suggests a predominant role in translational repression rather than direct transcriptional cleavage, indicating a possible functional divergence between miRNA-mediated regulatory pathways. Furthermore, hvu-miR6180 was observed to target multiple HvBAG genes, including HvBAG-6, HvBAG-4 and HvBAG-3, supporting the hypothesis of coordinated miRNA-driven regulation of BAG gene expression. In addition, hvu-miR6190 and hvu-miR6198 were identified as regulators of HvBAG-4 and HvBAG-3, respectively, with sequence complementarities indicative of cleavage-mediated gene silencing. The large number of miRNA interactions indicates a highly redundant and intricate regulatory network in which individual HvBAG genes can be modulated by several miRNAs, ensuring tight post-transcriptional control of cellular responses to environmental and developmental factors.

Gene Ontology (GO) Enrichment analysis of the HvBAG gene family in barley

To clarify the functional role of the HvBAG gene family, a Gene Ontology (GO) enrichment analysis was performed, which revealed a strong association with molecular chaperone activity, protein homeostasis, stress response and cellular regulatory processes (Fig. 7). The results show that HvBAG genes are primarily involved in the regulation of protein stability (GO:0031647) and chaperone binding (GO:0051087), highlighting their established role as co-chaperones that interact with heat shock proteins (HSP70/HSC70) to mitigate proteotoxic stress. Furthermore, the enrichment of heat shock protein binding (GO:0031072) and protein stabilization (GO:0050821) suggests a crucial function in maintaining protein homeostasis under adverse environmental conditions.

Fig. 7.

Gene Ontology (GO) enrichment analysis of the HvBAG gene family in barley. The bubble plot represents the GO enrichment analysis of the HvBAG gene family. The x-axis indicates fold enrichment, while the size of the circles corresponds to the number of genes associated with each GO term. The colour gradient represents statistical significance, expressed as -log10(FDR), with darker shading indicating a higher level of significance.

In addition to their role in protein quality control, HvBAG genes show significant involvement in the regulation of metal ion transport (GO:0010959) and the regulation of stomatal movement (GO:0010119), indicating their contribution to mechanisms of abiotic stress tolerance, including adaptation to drought and oxidative stress. The presence of adenyl nucleotide exchange factor activity (GO:0000774) and ATPase regulatory activity (GO:0060590) also suggests that HvBAG proteins modulate energy-dependent cellular responses and may coordinate stress adaptation via ATP-driven pathways. The functional enrichment also highlights nucleotide binding (GO:0000166), adenyl nucleotide binding (GO:0030554) and purine nucleotide binding (GO:0017076), consistent with the known interaction of BAG proteins with ATP-dependent chaperones. This suggests a possible role in signaling transduction and stress memory, where BAG proteins may influence cellular responses by modulating ATP-driven molecular processes. Interestingly, the GO terms positive regulation of transport (GO:0051050) and molecular function regulator activity (GO:0098772) suggest that HvBAG genes may be involved in transport or regulatory mechanisms related to intracellular stress signaling. In addition, their cytoplasmic localization (GO:0005737) is consistent with reports of BAG proteins functioning within stress granules and proteostasis networks, further supporting their role in stress adaptation and regulation of programmed cell death.

Expression pattern of BAG gene under heat stress conditions across different tissue

The gene expression analysis of the HvBAG gene family revealed distinct regulatory patterns in control versus treated shoot and root tissues. HvBAG3 exhibited minimal expression in control shoots but showed a significant increase in treated shoots and a dramatic upregulation in treated roots, indicating a strong response to treatment, particularly in root tissues. In contrast, HvBAG5 maintained consistently low expression across all groups, with only a slight increase observed in treated roots, suggesting a minor role in the treatment response. HvBAG7 displayed moderate expression levels in both control and treated shoots, along with higher levels in control roots; however, its expression decreased in treated roots, indicating downregulation under treatment conditions (Fig. 8).

Fig. 8.

Expression patterns of identified HvBAG genes under normal growth conditions and heat stress across shoot and root tissues. The heatmap illustrates the relative transcript abundance of HvBAG genes in shoots and roots under control (unstressed) and heat stress conditions. The color gradient represents the expression levels, with red indicating upregulation and blue indicating downregulation. Genes such as HvBAG3 show significant induction in both roots and shoots under heat stress, whereas HvBAG7, HvBAG4, and HvBAG6 exhibit downregulation in heat-treated roots. The observed differential expression suggests tissue-specific regulatory mechanisms and highlights the role of HvBAG genes in heat stress adaptation.

HvBAG-2 exhibited relatively stable expression across all groups, with slightly higher levels in control and treated roots, reflecting minimal sensitivity to treatment or tissue type. HvBAG-1 showed higher expression in shoot tissues, with moderate levels in both control and treated shoots, but low expression in control and treated roots, highlighting a tissue-specific regulation that is independent of treatment. HvBAG-4 had moderate expression in control shoots, lower levels in treated shoots, and higher expression in control roots compared to treated roots, suggesting downregulation in treated root tissues. Similarly, HvBAG-6 displayed moderate expression in control and treated shoots, with higher levels in control roots than in treated roots, reinforcing the trend of downregulation in treated roots (Fig. 8).

In total, HvBAG-3 was strongly upregulated in treated roots and moderately upregulated in treated shoots, indicating a significant role in the treatment response. HvBAG-5 showed a minor but consistent increase in treated roots. HvBAG-7, HvBAG-4, and HvBAG-6 were downregulated in treated roots, suggesting suppression under treatment conditions. HvBAG1 exhibited preferential expression in shoot tissues regardless of treatment, while HvBAG-2 remained stable across groups, reflecting minimal responsiveness to treatment or tissue type. These findings highlight the differential regulation of the HvBAG gene family under treatment conditions, emphasizing their potential roles in stress adaptation and tissue-specific responses.

The leave-one-out cross-validation (LOOCV) analysis of the BAG gene family expression profiles achieved a perfect accuracy score of 1.0 across multiple iterations. The mean LOOCV accuracy was also 1.0, demonstrating the model’s ability to distinguish between experimental conditions without misclassifications (Fig. 9). This consistent performance highlights the strong discriminatory capacity of BAG gene expression patterns in differentiating control and treated samples, suggesting their significant role in stress response mechanisms. However, the perfect accuracy should be interpreted cautiously, as smaller datasets may lead to overly optimistic performance estimates. These findings provide computational evidence for the biological significance of the BAG gene family in this context.

Fig. 9.

Validation of BAG gene family expression patterns using the Leave-One-Out Cross-Validation (LOOCV) approach. The LOOCV analysis yielded perfect classification performance across all folds, resulting in a mean LOOCV accuracy of 1.0 for both the root and shoot tissues.

Rewiring of regulatory networks of BAGs in response to heat stress in shoot

The regulatory impact of BAG genes in shoot tissues under control and heat-treated conditions showed distinct differences, reflecting their adaptive roles in maintaining cellular stability and responding to stress. Under control conditions, BAG genes primarily regulate genes involved in basal cellular processes such as protein folding, metabolism, and development. For example, HvBAG-6 targets genes like CCT1-2 and HSP70-17, while HvBAG-2 regulates HPD and HSP90-1, indicating roles in protein stability and metabolic processes. HvBAG1 targets VHA-G1 and HSP70-4, highlighting its involvement in cellular homeostasis. HvBAG4 and HvBAG7 regulate genes like HSP18.1 and CPN60A1, pointing to their roles in protein folding and stress responses (Fig. 10). The regulatory network of BAG genes in shoot tissue under control conditions is characterized by a focus on maintaining cellular stability, growth, and development, with minimal stress-related targets. Under heat stress, the regulatory focus of BAG genes shifts significantly toward stress-responsive pathways. HvBAG6 targets genes like HSFB2A, APX1, and ERDJ3A, indicating roles in heat shock transcription factors, antioxidant defense, and protein folding under stress. HvBAG2 regulates UBQ10, HSP90-2, and HSP17.6 C, emphasizing its role in protein degradation and heat shock responses. HvBAG1 targets RSZ21 and His1-3, suggesting involvement in RNA processing and histone regulation under stress. HvBAG4 and HvBAG7 regulate genes like APX3 and ELIP1, highlighting their roles in oxidative stress management and photoprotection. HvBAG3 and HvBAG5, which show minimal activity under control conditions, become prominent regulators under heat stress, targeting genes such as SHM4, NCL, and HSFB2A, indicating their roles in metabolic adaptation and stress responses. Key differences between the two conditions include the upregulation of stress-responsive genes under heat stress, such as heat shock proteins (HSPs), antioxidant enzymes (e.g., APX1, APX3), and heat shock transcription factors (e.g., HSFB2A), which are minimally targeted under control conditions. Heat stress also induces the regulation of genes involved in metabolic pathways (e.g., SHM4, CPT1) and enhances protein folding (e.g., CPN60A1, ERDJ3A) and degradation (e.g., UBQ10) pathways to cope with stress-induced damage (Fig. 9). Additionally, oxidative stress management becomes a priority under heat stress, with the regulation of antioxidant genes (e.g., APX1, APX3) to mitigate oxidative damage. HvBAG3 and HvBAG5, which show limited activity under control conditions, become key regulators under heat stress, targeting genes involved in stress adaptation and metabolic processes (Fig. 10). Despite these differences, there are overlapping regulatory roles between the two conditions. Both conditions involve the regulation of chaperones and heat shock proteins (e.g., HSP70-4, HSP90-2), though this is more pronounced under heat stress. Genes like VHA-G1 and HSP70-4 are targeted under both conditions, indicating their roles in maintaining cellular stability.

Fig. 10.

Inferred regulatory networks of HvBAG genes in barley shoots under control (A) and heat-treated (B) conditions. The network diagrams depict predicted interactions between HvBAG genes (red nodes) and their associated regulatory elements, including transcription factors, molecular chaperones, and stress-responsive genes. The changes in network connectivity highlight the role of HvBAG genes in maintaining protein homeostasis and cellular resilience under heat stress.

Regulatory networks rewiring of BAGs in response to heat stress in Roots

The regulatory network of BAG genes in root tissues shows significant differences between control growth conditions and heat-treated conditions, highlighting their adaptive roles in maintaining cellular homeostasis and responding to stress. Under control conditions, BAG genes primarily regulate genes involved in essential cellular processes such as growth, development, and metabolism. In contrast, during heat stress, their regulatory focus shifts towards pathways that respond to stress, including heat shock responses, metabolic adaptation, and oxidative stress management. Under control conditions, specific BAG genes target distinct sets of genes associated with cellular maintenance. For example, HvBAG6 regulates HORVU.MOREX.r3.7HG0701900 and TFIIS, indicating its role in transcriptional regulation and RNA processing. HvBAG4 targets CTL1, NADP-ME3, and CESA6, linking it to cell wall integrity, metabolic pathways, and cellulose biosynthesis. Similarly, HvBAG1 is associated with HORVU.MOREX.r3.2HG0156450, KCS11, and XERICO, suggesting its involvement in stress responses, lipid metabolism, and cellular signaling (Fig. 11). The regulatory influence of HvBAG3 extends to CAF1-7, DEGP9, and NF-YC10, indicating its roles in chromatin remodeling, protein degradation, and transcriptional regulation. Other BAG members, including HvBAG5 and HvBAG2, regulate genes related to calcium signaling, protein ubiquitination, and metabolic processes, reinforcing their contribution to cellular stability under normal growth conditions. Under heat stress, the regulatory dynamics of BAG genes shift towards pathways crucial for stress adaptation. HvBAG4 exerts regulatory control over ASK8, SC35, and TET4, implicating its role in signal transduction, RNA splicing, and metabolic adaptation. HvBAG7 targets genes such as T6D22.20, MBF1A, and BZIP17, which are involved in transcriptional regulation, stress responses, and protein folding. HvBAG2 regulates MRO11.1, BFRUCT3, and PUP1, underscoring its role in metabolic pathways and stress adaptation, while HvBAG6 modulates NIR1 and F14P3.5, suggesting involvement in nitrogen metabolism and stress responses. The regulatory network of HvBAG3 under heat stress encompasses RPOT3, BZIP17, and UBQ10, which are associated with transcription, protein degradation, and stress signaling. Similarly, HvBAG5 influences UGE1, CWINV1, and NIT4, highlighting its role in carbohydrate metabolism and nitrogen assimilation. A comparison of BAG gene regulatory networks under normal and heat-stressed conditions reveals significant differences. Under heat stress, BAG genes notably upregulate stress-responsive genes, including BZIP17, UBQ10, and NIR1, which are only minimally targeted under normal conditions (Fig. 11). This regulatory shift induced by heat stress also enhances metabolic adaptation, as seen in the activation of genes such as BFRUCT3 and CWINV1, which are essential for carbohydrate metabolism and energy balance. Additionally, BAG genes play a role in maintaining protein homeostasis, as demonstrated by their regulation of MBF1A and UBQ10, which are involved in protein folding and degradation, respectively. Oxidative stress management is also intensified under heat stress, with genes like NIR1 and BZIP17 playing crucial roles in alleviating stress-induced cellular damage. Notably, certain BAG members, such as HvBAG3 and HvBAG5, shift from a relatively minor role under normal conditions to key regulatory functions under heat stress, targeting genes critical for stress adaptation and metabolic reprogramming. Despite these differences, some regulatory functions of BAG genes remain consistent across both conditions. Genes associated with protein folding and degradation, such as UBQ10 and DEGP9, are regulated continuously, although their importance is heightened under stress. Similarly, genes involved in metabolic pathways, including NADP-ME3 and BFRUCT3, are targeted in both scenarios, with increased regulatory dynamics during heat stress. Furthermore, cellular homeostasis is a shared regulatory focus, as genes such as GULLO2 and UBQ10 are targeted under both conditions, reflecting their vital roles in maintaining cellular stability.

Fig. 11.

Inferred regulatory networks of HvBAG genes in barley root under control (A) and heat-treated (B) conditions. The network diagrams depict predicted interactions between HvBAG genes (red nodes) and their associated regulatory elements, including transcription factors, molecular chaperones, and stress-responsive genes. The changes in network connectivity highlight the role of HvBAG genes in maintaining protein homeostasis and cellular resilience under heat stress.

Discussion

The identification and characterization of the BAG gene family in barley offer important insights into their functional roles, evolutionary dynamics, and potential applications for stress tolerance. Studies in model plants such as A. thaliana and O. sativa have emphasized the role of BAG proteins in regulating programmed cell death, maintaining protein homeostasis, and responding to stress^25–27. The discovery of seven HvBAG genes with varying physicochemical properties such as molecular weights ranging from 20.79 to 144.08 kDa and isoelectric points between 5.01 and 10.97 suggests potential functional specialization within this gene family. Importantly, the classification of HvBAG proteins into the three established domain types BAG only, UBL + BAG and calmodulin-binding/IQ + BAG provides additional evidence for functional diversification. The UBL-containing HvBAG2 and HvBAG3 are likely involved in ubiquitin–proteasome-mediated degradation processes, as reported for the UBL-type BAG proteins in rice and maize, which function in the removal of misfolded proteins under heat stress^16,28. The IQ motif-containing HvBAG6 and HvBAG7 are likely linked to the calcium/calmodulin-dependent signaling pathway, a pathway known to modulate abiotic stress responses and programmed cell death in Arabidopsis and tomato^9,17. BAG-only proteins (HvBAG1, HvBAG4, HvBAG5) may act as central co-chaperones with direct HSP70/HSC70 regulation, which is consistent with reports from wheat where BAG-only members are upregulated during early heat stress²⁹. This domain-based classification therefore not only refines the structural annotation of HvBAG proteins but also provides a predictive framework for hypothesizing their stress-related functional roles in barley. These findings are consistent with previous reports in rice, where OsBAG proteins display a wide range of biochemical properties and subcellular localizations, indicating functional diversification³⁰.

The variability in the instability index and GRAVY values among HvBAG proteins suggests differences in stability and hydrophobicity, which may affect protein interactions and cellular localization. Similar observations have been made in wheat (Triticum aestivum), where TaBAG proteins exhibited distinct structural features that aligned with their functions in responding to abiotic stress, particularly heat and drought^29,31. Subcellular localization analysis further reinforces this functional divergence, with HvBAG3 primarily found in the nucleus and HvBAG-5 located in the mitochondria. This pattern is reminiscent of Arabidopsis BAG proteins, where AtBAG6 is localized in the nucleus and interacts with chaperone proteins to facilitate proteostasis, while AtBAG7 is associated with the endoplasmic reticulum and plays a role in the unfolded protein response¹¹.

Phylogenetic analysis divided the HvBAG genes into five clades, highlighting both evolutionary conservation and divergence³². The distribution of domain classes among clades supports the hypothesis of subfunctionalisation by gene duplication³³, with BAG-only types clustered separately from UBL- and IQ-containing types. The presence of conserved BAG domains, together with additional structural motifs such as ubiquitin-like motifs and coiled-coil regions, suggests that these genes have specialized regulatory functions. Similar findings have been reported in rice and maize (Z. mays), where BAG genes show lineage-specific expansion and domain adaptations associated with stress responses^16,28. Importantly, motif analysis indicated that specific motifs, such as Motifs 1, 3, and 7, are shared among all HvBAG proteins, underscoring their essential roles in regulating cellular stress. These results align with studies in Arabidopsis and wheat, which demonstrate that conserved BAG motifs play a critical role in chaperone activity and cellular defense mechanisms²⁹. The phylogenetic classification of barley BAG proteins, together with those from rice, maize, sorghum (Sorghum bicolor), and A. thaliana, provides deeper insights into their evolutionary history. The clear separation of several clades between monocotyledons and dicotyledons supports the hypothesis that lineage-specific diversification, likely driven by gene duplication and subsequent neofunctionalization, has contributed to the acquisition of specialized functional roles, particularly in stress adaptation and protein homeostasis^34,35.

Chromosomal distribution and synteny analysis shed light on the evolutionary history of the HvBAG family. The HvBAG genes are scattered across five chromosomes and show evidence of segmental duplications, similar to findings in rice and maize, where BAG gene expansion was largely driven by whole-genome duplication events²⁸. The presence of collinear gene pairs indicates that segmental duplications were significant in the evolution of HvBAG genes, a trend also observed in wheat and Arabidopsis, where the retention of BAG genes is linked to adaptive functions in stress resilience⁹. Additionally, the Ka/Ks ratio analysis suggests purifying selection, indicating functional conservation and reinforcing the evolutionary pressure to maintain BAG gene functions, akin to observations in rice and Arabidopsis¹¹.

The synteny analysis of barley, rice, and Arabidopsis demonstrated a stronger conservation of BAG genes between barley and rice, highlighting their common monocot lineage. Specifically, six orthologous BAG gene pairs were identified between barley and rice, while only one pair was found between barley and Arabidopsis. This finding supports previous comparative genomics studies that emphasize the closer evolutionary relationship between barley and rice, stemming from their shared ancestry within the Poaceae family^17,36.

In plants, miRNAs play a crucial role in fine-tuning responses to various abiotic stresses, including heat stress, by targeting stress-associated genes^36,37. The identification of specific miRNAs that target HvBAG genes in barley offers new insights into the post-transcriptional control of BAG-mediated stress responses. For example, hvu-miR6192 is predicted to target HvBAG-7, while hvu-miR6196 appears to regulate HvBAG-6, suggesting a potential cleavage-mediated regulatory mechanism. This type of miRNA-guided gene regulation is a well-established mechanism in plant stress adaptation, with miRNAs serving as key modulators of stress-responsive gene networks^37–39. The expression patterns of HvBAG genes under heat stress reveal a complex regulatory landscape that varies across different tissues. Notably, HvBAG-3 is significantly upregulated in both treated roots and shoots, indicating its active role in heat stress adaptation. This finding aligns with previous studies on BAG genes in other plant species, where homologs of BAG3 have been associated with stress mitigation through interactions with heat shock proteins (HSPs) and antioxidant enzymes^40,41. Conversely, the downregulation of HvBAG-7, HvBAG-4, and HvBAG-6 in heat-treated roots suggests that these genes may be suppressed under specific stress conditions, potentially to prevent excessive protein turnover or energy expenditure in stressed tissues.

The rewiring of regulatory networks involving BAG genes in response to heat stress highlights their essential role in maintaining cellular homeostasis. BAG proteins interact with molecular chaperones like HSP70 and HSP90, facilitating protein refolding and degradation processes. This interaction is particularly critical during heat stress, as the accumulation of misfolded proteins can lead to cellular damage^26,42,43. The differential expression of HvBAG genes in shoots and roots indicates tissue-specific regulatory mechanisms. Shoots, being directly exposed to heat stress, may require a more robust BAG-mediated response, while roots might focus on other protective strategies, such as osmolyte accumulation and the scavenging of reactive oxygen species (ROS) process.

Conclusion

This study was performed to identify and characterize the BAG gene family in barley, providing crucial insights into their evolutionary conservation, functional specialization, and roles in stress adaptation. By conducting comparative genomic analyses with Arabidopsis, rice, wheat, and maize, we uncovered the structural and biochemical variability of HvBAG proteins, highlighting their specialized roles in programmed cell death, protein homeostasis, and stress response mechanisms. Phylogenetic classification and chromosomal distribution analyses suggest that segmental duplications and purifying selection have significantly influenced the expansion and functional retention of BAG genes in barley, reflecting similar evolutionary patterns observed in other plant species.

A key finding of this study is the miRNA-mediated regulation of HvBAG genes, introducing an additional layer of post-transcriptional control in response to heat stress. The observed differential subcellular localization and tissue-specific expression patterns further emphasize the diverse regulatory roles of HvBAG genes in plant stress resilience. Additionally, the predicted interactions between BAG proteins and molecular chaperones (HSP70 and HSP90) indicate a conserved mechanism for maintaining protein stability and cellular homeostasis, particularly under environmental stress conditions. The findings of this study contribute to the existing knowledge of the BAG gene family in plants and offer new insights into their regulatory mechanisms and stress-associated functions. The results have important theoretical and practical implications, especially for crop improvement programs. Understanding the role of HvBAG genes in stress tolerance pathways can inform biotechnological approaches aimed at enhancing stress resilience in barley and other cereal crops. In conclusion, this study establishes a foundational framework for further research on HvBAG genes, with potential applications in genetic engineering and molecular breeding strategies to enhance barley’s stress resilience. By uncovering the evolutionary dynamics and stress-related functions of the BAG gene family, this work paves the way for future efforts to develop climate-resilient crop varieties, contributing to sustainable agriculture and global food security.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

BP and RH conceived and design the study, performed the data analysis. RH prepared the graphs and wrote the manuscript. FJ contributed in manuscript writing. 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. Moreover, we used the public data set with PRJNA324116 ID available at European nucleotide archive (ENA) database (https://www.ebi.ac.uk/ena/browser/home).

Declarations

Competing interests

The authors declare no competing interests.

Ethics approval and consent to participate

All experimental studies on plants were conducted in compliance with relevant institutional, national, and international guidelines and legislation.