Pangenome-wide insights into eukaryotic translation initiation factors (eIF) genes uncover targets for sterility mosaic disease resistance engineering in pigeonpea

Abstract

Sterility mosaic disease (SMD) poses a significant threat to pigeonpea production. Although resistant cultivars are widely used, the rapid evolution of viral pathogens often renders host resistance ineffective. Targeting host susceptibility factors—particularly eukaryotic translation initiation factors (eIFs) that many plant viruses exploit for replication—offers a promising alternative. Although eIF proteins are well characterized in several model crops, their roles in pigeonpea remain largely unexplored. In this study, we conducted a comprehensive analysis of the eIF gene family in pigeonpea, supported by a pangenome-wide survey to capture cultivar-specific diversity relevant to SMD response. A total of 52 eIF genes were identified in the reference genome, while 47 homologs were consistently present across the pangenome, representing core gene members. The pangenome analysis also revealed limited sequence-level variation, highlighting strong evolutionary conservation of eIF genes in pigeonpea. Gene copy number, chromosomal distribution, and promoter cis-regulatory elements were characterized through detailed in silico analyses. Expression profiling under PPSMV infection revealed contrasting transcriptional responses: several eIF genes were upregulated in resistant plants, whereas eIF3K, eIF4G, and eIF5A showed higher expression in susceptible genotypes. Variant analysis of the susceptible genotype ICP8863 uncovered 94 polymorphisms across the 52 eIF genes, with seven predicted to have high or moderate functional impact. A stop-gained mutation in eIF3E (Cc_02836) and a frameshift indel in eIF4B (Cc_22348) are predicted to result in truncated, nonfunctional proteins. Additionally, missense mutations in eIF3E (Cc_02836), eIF5B (Cc_04638), eIF2Bδ (Cc_06116), eIF2Bα (Cc_18788), and eIF4B (Cc_22348) may influence protein structure or interaction capacity, potentially compromising antiviral defense. Overall, this study identifies key eIF family members potentially contributing to SMD susceptibility and provides candidate genes for experimental validation to engineer PPSMV resistance in pigeonpea varieties.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s12870-026-09450-5.

Introduction

Pigeonpea (Cajanus cajan) is an agriculturally important legume cultivated extensively across the semi-arid tropics and subtropical regions of Asia and Africa. However, its productivity is severely constrained by several biotic stresses, including fusarium wilt (FW), sterility mosaic disease (SMD), and phytophthora blight (PB) [1]. Among these, SMD—primarily caused by the pigeonpea sterility mosaic virus (PPSMV)—is one of the most destructive foliar diseases of pigeonpea. Yield losses caused by SMD can reach up to 95%, depending on the genotype and the stage of infection [2]. Despite decades of breeding efforts, the development of durable resistance has remained challenging. This is due to environmental variability and the limited availability of highly resistant donors within the cultivated gene pool [3]. Consequently, a deeper understanding of the molecular basis of SMD resistance is essential for accelerating resistance breeding.

Host-plant resistance is widely regarded as the most practical and economically viable strategy to combat plant viral diseases. Such resistance can arise through two major mechanisms: dominant resistance (R) genes or recessive alleles of host susceptibility genes required for successful viral infection [4]. Most dominant R genes encode nucleotide-binding leucine-rich repeat (NLR) proteins, similar to those known to confer resistance against bacterial and fungal pathogens [5–7]. In contrast, recessive resistance—commonly deployed in various crops—results from loss-of-function mutations in host factors hijacked by viruses [8, 9]. Such mutations can be introduced through crossing, natural variation, or targeted mutagenesis, offering an advantage by avoiding transgene integration and circumventing limitations in natural resistance sources [10].

Because plant viruses rely heavily on host translation machinery for viral protein synthesis, translation initiation is a crucial regulatory step in host–virus interactions [11]. Eukaryotic translation initiation factors (eIFs) therefore play central roles in viral replication and infection [12]. Many well-characterized recessive resistance genes belong to the eIF4E and eIF4G families and their isoforms [4], which have emerged as key determinants of host susceptibility [13]. In addition, several eIF family members are implicated in modulating plant defense responses [14]. Loss-of-function mutations in eIF4E/eIF(iso)4E have conferred virus resistance in numerous plant species. These include Arabidopsis [15, 16], barley [17], tomato [18], pepper [19], lettuce [20], and melon [21]. Despite the central role of translational regulators in plant–virus interactions, only a limited subset of host targets manipulated by viruses has been identified. Recent evidence shows that eIF6 expression is upregulated in Nicotiana benthamiana during turnip mosaic virus (TuMV) infection, thereby modulating host translation dynamics [22]. These findings underscore the importance of systematic investigation of eIF family members across plant species under viral infection. However, comprehensive genome-wide and functional analyses of eIF genes during PPSMV infection in pigeonpea are still lacking.

To address this gap, we conducted a genome-wide identification of eIF genes in pigeonpea and extended our investigation across the pigeonpea pangenome to capture cultivar-specific diversity. In total, we identified 52 eIF genes in the reference genome and 47 homologous sequences across the pangenome. Through detailed in silico analyses, we characterized their chromosomal distribution, promoter cis-acting elements, phylogenetic relationships, genomic collinearity, and signatures of selection. Additionally, we examined the expression patterns of eIF genes in resistant and susceptible pigeonpea genotypes under PPSMV infection and mock conditions. Collectively, these findings provide fundamental insights into the strong evolutionary conservation and functional dynamics of eIF genes in pigeonpea, offering valuable targets for developing PPSMV-resistant cultivars.

Materials and methods

Identification of eIF genes and sequence retrieval

Genomic data and eIF protein sequences from pigeonpea were identified by comparative analysis with the reference model plant Arabidopsis thaliana. The eIF sequences for A. thaliana were retrieved from the TAIR database (https://www.arabidopsis.org/index.jsp) and used as query sequences. Corresponding CceIF genes in pigeonpea were then identified using local BLAST searches against the protein sequences of 59 known AteIF genes available in CajanusMine, a component of the Legume Information System (LIS) (https://mines.legumeinfo.org/cajanusmine/begin.do). Candidate sequences were filtered based on sequence similarity using an E-value threshold of ≤ 1e^⁻5, with a minimum sequence identity of ≥ 40% and query coverage of ≥ 50%. Redundant hits were removed, and the resulting non-redundant sequences were further validated for the presence of conserved eIF domains using domain analysis tools. To support evolutionary analyses, eIF protein sequences from another legume species, Medicago truncatula, were also collected. These sequences were obtained from the LIS and Phytozome databases (https://phytozome-next.jgi.doe.gov/) through local BLAST searches using A. thaliana eIF protein sequences as queries. The same strategy was applied to identify homologous eIF genes across the pigeonpea pangenome.

Phylogenetic and synteny analysis of identified eIF genes

To examine the evolutionary relationships among the identified eIF genes, a phylogenetic analysis was conducted. Multiple sequence alignment (MSA) was performed using ClustalW in MEGA v11, incorporating protein sequences from the C. cajan reference genome (v2), its pangenome homologs, and corresponding eIF sequences from A. thaliana. A phylogenetic tree was then generated using the Neighbor-Joining method with 1,000 bootstrap replicates, implemented in MEGA v11, due to its computational efficiency and its suitability for preliminary classification of gene family members [23]. The resulting tree was visualized using the iTOL v6 online platform (https://itol.embl.de/) [24]. C. cajan eIF genes were assigned to clades based on their sequence similarity to Arabidopsis eIF family members, with each clade represented by a distinct color.

Collinearity analysis was performed between C. cajan and M. truncatula, with M. truncatula used as the model legume. This analysis was conducted using the Multiple Collinearity Scan (MCScanX) toolkit integrated within TBtools, employing the following parameters: BLASTp CPU set to 2, a maximum of 5 BLAST hits, and an E-value cutoff of 1e^{− 10}, using genomic data obtained from the LIS database [25]. Additionally, synteny analysis with the model plant A. thaliana was conducted to trace the evolutionary trajectories of orthologous genes in the context of increasing genomic complexity. Furthermore, non-synonymous (Ka) and synonymous (Ks) substitution rates for identified orthologous gene pairs were calculated using the Simple Ka/Ks Calculator implemented in TBtools. Prior to this, protein sequences were aligned using MUSCLE, and CDS were aligned in a codon-aware manner using PAL2NAL. The Ka/Ks ratios were estimated based on the Nei–Gojobori method, and gene pairs with unreliable estimates (Ks = 0) were excluded to ensure robustness.

Chromosomal location and intron-exon structure analysis of eIF genes

The chromosomal locations and intron–exon structures of C. cajan eIF genes were obtained from annotation files available on CajanusMine (https://mines.legumeinfo.org/cajanusmine/begin.do). Using these chromosomal coordinates, the identified genes were mapped onto their respective chromosomes using the MapGene2Chromosome v2.1 online tool (http://mg2c.iask.in/mg2c_v2.1/) [26]. Gene structure visualization was performed with the Gene Structure Display Server 2.0 (http://gsds.gao-lab.org/index.php) [27], which was used to illustrate intron–exon organization for each eIF gene.

Cis-regulatory element and conserved domain analysis of the identified eIF genes

To identify cis-regulatory elements within the promoter regions of C. cajan eIF genes, a 2000 bp upstream sequence from the transcription start site of each gene was extracted from CajanusMine (https://mines.legumeinfo.org/cajanusmine/begin.do). These promoter sequences were analyzed using the PlantCARE database (http://bioinformatics.psb.ugent.be/webtools/plantcare/html/) [28] to annotate cis-acting elements, which were subsequently visualized as a heatmap using TBtools.

Additionally, conserved domains within eIF proteins from Arabidopsis, the C. cajan reference genome (v2), and the pigeonpea pangenome were identified using the HMMER (https://www.ebi.ac.uk/Tools/hmmer) [29]. Protein sequences were queried against the Pfam database (Pfam release 35.0) using the hmmscan module. Domain hits were filtered using a significance threshold of E-value ≤ 1e^⁻5, and only domains with reliable alignment scores were retained for further analysis. Default parameters were used for other settings. Domain architectures were visualized according to eIF family groups to assess conserved and variable features across the pangenome relative to the reference genome.

Gene expression analysis of eIF genes in pigeonpea

Plant material and source

Seeds of the pigeonpea genotypes used in this study were obtained from the pigeonpea breeding team, International Crops Research Institute for the Semi-Arid Tropics (ICRISAT), Hyderabad, India, with appropriate authorization for research use. The medium-duration variety TDRG 59 (Telangana Kandi 3) was originally developed from a cross between ICP 7118_7_5_B_B and a Banda Palera collection. In previous experiments, TDRG 59 has been characterized as moderately resistant to sterility mosaic disease (SMD). For comparison, the genotype ICP 8863, selected from the landrace ICP-7626 at ICRISAT, India, was included as the susceptible control. ICP 8863 is widely used as a reference susceptible line due to its consistently high incidence of SMD (https://oar.icrisat.org/567/1/PMD_44.pdf). All genotypes used in the study were formally identified by scientists within the ICRISAT pigeonpea Breeding Program, and authenticated seed stocks are conserved within the institutional Genebank. No wild plant materials were collected, and therefore, no field collection permits were required.

Experimental site, compliance, and sample collection

All experimental work involving pathogen screening was conducted at ICRISAT’s historical SMD sick plot, maintained under institutional, national, and international guidelines for plant experimentation. The study complied with all relevant biosafety, phytopathology, and field research regulations.

Plants were inoculated following standard PPSMV transmission procedures used in the SMD sick plot. Following inoculation, susceptible plants developed characteristic symptoms of SMD, including bushy growth and mosaic or chlorotic mottling of the leaves. In contrast, resistant plants displayed no visible symptoms. Viral load analyses confirmed the presence of PPSMV only in inoculated susceptible plants, with no detectable virus in the inoculated resistant genotype (Figure S1). Leaf tissue was collected in biological triplicates from both PPSMV-infected and mock-treated plants for each genotype, corresponding to susceptible (SP, SM) and resistant (RP, RM) treatments. Samples were harvested at the same developmental stage and immediately processed for gene expression analysis.

Expression analysis using quantitative reverse-transcription polymerase chain reaction

Total RNA was extracted using the MN kit according to the manufacturer’s instructions. RNA concentration was quantified using the Qubit 3.0 Fluorometer (Invitrogen, Thermo Fisher Scientific, USA) with the Qubit RNA Broad Range (BR) assay kit, following the manufacturer’s instructions. For RT-qPCR analysis, 2.0 µg of purified RNA was reverse-transcribed into cDNA using the ThermoScript RT-PCR system (Invitrogen, Carlsbad, CA, USA) following the recommended protocol. A total of 20 primers were designed to assess the expression of eIF genes. This included one primer set each to amplify selected members of the eIF2Bβ (Cc_11975 and 11977), eIF3B (Cc_07050), along with six sequences from other eIF3 subgroup genes, eIF4E (Cc_19714), eIFiso4E (Cc_02582), eIF4G (Cc_01505 and Cc_16013), eIFiso4G (Cc_20884 and Cc_27470), in addition to five sequences from other eIF4A subgroup genes, and eIF5A (Cc_06008) (Table S1). These sequences were chosen based on prior reports that identified these gene subgroups as involved in various virus infections across different plant species (Table S2). Primer specificity was validated using cDNA at annealing temperatures of 58 °C and 60 °C to confirm cDNA integrity and primer efficiency. Primer efficiencies were evaluated using standard curves and were within the acceptable range (90–110%). Amplification specificity was verified by melt-curve analysis, which confirmed a single peak for each primer pair (Figure S2). Quantitative RT-PCR was conducted using gene-specific primers on a CFX96 Real-Time PCR System (Bio-Rad, India) using three biological replicates. RT-qPCR reaction conditions were set in two technical triplicates as 2 min at 95 °C, 40 cycles of 10s at 95 °C, and 20s at 60 °C/ 58 °C (depending on the primer annealing temperature) followed by fluorescence measurement at each 0.5 °C variation from 60 °C to 95 °C in 20 min was included to obtain the melting curve. The GAPDH gene (XM_020358085) served as an internal control for normalizing Ct values [30]. Relative gene expression levels were calculated using the 2^^−ΔΔCt method [31]. Additionally, viral load in both mock- and PPSMV-inoculated samples of resistant and susceptible genotypes was quantified using PPSMV CP-specific primers (Table S1).

Whole genome resequencing analysis of pigeonpea susceptible genotype ICP 8863

Genomic sequencing data for the pigeonpea genotype ICP8863 were downloaded using fastq-dump v2.5.7 from the SRA toolkit (accession SRR5484735, project PRJNA383013), which is a part of the pigeonpea pangenome study [32]. Raw reads were subjected to quality assessment using FastQC v0.11.8 (http://www.bioinformatics.babraham.ac.uk/projects/fastqc/), and adapter trimming and quality filtering were performed with fastp v0.23.4 [33]. Cleaned reads were aligned to the PPSMV moderately resistant C. cajan genotype ICPL87119 (Asha) [34] reference genome assembly (v2.0) using Bowtie2 v2.2.5. Genome indices were generated from the reference genome using default settings. Read mapping was carried out using the parameters: --sensitive --end-to-end -I 0 -X 1000.

The resulting BAM files were sorted and indexed by genomic coordinates using SAMtools v1.8 (http://samtools.sourceforge.net/). Variant calling was performed with bcftools v1.17, using the mpileup and call functions. Identified variants were annotated with SnpEff v5.1, and only those with a quality score ≥ 20 and classified as high or moderate impact were retained for downstream analysis.

Results

Identification and sequence analysis of the pigeonpea eIF family genes

A total of 52 C. cajan eIF (CceIF) genes were identified in the reference genome as orthologs to 59 A. thaliana eIF (AteIF) genes retrieved from the TAIR database, following the removal of redundant sequences. Within the pigeonpea pangenome, 47 homologous eIF genes were detected, all of which were consistently present across all 89 accessions analyzed. These genes were further validated through InterProScan analysis and manual curation to eliminate redundant entries and false positives. The identified genes were classified into six major groups corresponding to eIF1 through eIF6, each comprising the respective AteIF genes, their orthologous CceIF counterparts, and homologous sequences from the pangenome. Each group was further divided into subgroups based on sequence variation and the presence of additional domains. Detailed sequence information for all identified genes is provided in Table 1.

Table 1.

Identification of eIF family genes in Arabidopsis, pigeonpea reference genome and pigeonpea pangenome

eIF groups	eIF subgroups	Numbering	AteIF	CceIF ref genome	Pangenome seq
eIF1 group	eIF1	eIF1-1	At4g27130	Cc_09668, Cc_09899, Cc_09898	g50603,
		eIF1-2	At5g54760		g50598
		eIF1-3	At1g54290
		eIF1-4	At5g54940	Cc_12555	g36543
	eIF1A	eIF1A-1	At5g35680	Cc_08506
		eIF1A-2	At2g04520
eIF2 group	eIF2alpha	eIF2-A1	At2g40290	Cc_00264	g55748
		eIF2-A2	At5g05470
	eIF2beta	eIF2-B1	At5g20920	Cc_11975, Cc_11977	g46510,
		eIF2-B2	At3g07920		g46536
		eIF2-B3	At5g01940
	eIF2gamma	eIF2-G1	At1g04170	Cc_24468	g32891
		eIF2-G2	At4g18330
		eIF2-G3	At2g18720
	eIF2Balpha	eIF2B-A1	At1g53880	Cc_08622	g55763
		eIF2B-A2	At1g72340
		eIF2B-A3	At2g05830	Cc_18788	g26136
	eIF2Bbeta	eIF2B-B1	At3g07300	Cc_17642, Cc_24897	g42832,
					g31660
	eIF2Bdelta	eIF2B-D1	At5g38640	Cc_20699, Cc_17210, Cc_27260, Cc_06116	g23968,
		eIF2B-D2	At1g48970		g50866,      g53692,
		eIF2B-D3	At2g44070		g52662
	eIF2Bepsilon-2Bgamma	eIF2B-E1	At2g34970	Cc_08386	g55164
		eIF2B-E2	At3g02270
		eIF2B-E3	At4g18300
eIF3 group	eIF3A	eIF3A-1	At4g11420	Cc_07682	g51489
	eIF3B	eIF3B-1	At5g27640	Cc_07050	g08658
		eIF3B-2	At5g25780
	eIF3C	eIF3C-1	At3g56150	Cc_23082, Cc_08019	g35895,
		eIF3C-2	At3g22860		g54882
	eIF3D	eIF3D-1	At4g20980	Cc_19884	g51256
		eIF3D-2	At5g44320
	eIF3E	eIF3E-1	At3g57290	Cc_02836	g51473
	eIF3F	eIF3F-1	At2g39990	Cc_00317, Cc_12196	g28180,
					g39851
	eIF3G	eIF3G-1	At3g11400	Cc_02371	g51227
		eIF3G-2	At5g06000
	eIF3H	eIF3H-1	At1g10840	Cc_17238	g51090
	eIF3I	eIF3I-1	At2g46280	Cc_13610	g46851
		eIF3I-2	At2g46290
	eIF3K	eIF3K-1	At4g33250	Cc_27747	g57667
eIF4 group	eIF4A	eIF4A-1	At3g13920	Cc_07109, Cc_23422, Cc_18851, Cc_25520, Cc_27658, Cc_19025	g48223,
		eIF4A-2	At1g54270		g40683,       g36856
	eIF4B	eIF4B-1	At3g26400	Cc_22348	g20439
		eIF4B-2	At1g13020
	eIF4E	eIF4E-1	At4g18040	Cc_19714	g51170
		eIF4E-2	At1g29590
		eIF4E-3	At1g29550
	eIFiso4E	eIFiso4E-1	At5g35620	Cc_02582	g55997
	eIF4G	eIF4G-1	At3g60240	Cc_01505, Cc_16013	g15410,
					g46123
	eIFiso4G	eIFiso4G-1	At5g57870	Cc_27470, Cc_20884	g36562,
		eIFiso4G-2	At2g24050		g11769
eIF5 group	eIF5	eIF5-1	At1g77840	Cc_20274, Cc_17330	g27831,
		eIF5-2	At1g36730		g51866
	eIF5A	eIF5A-1	At1g26630	Cc_13380, Cc_27150, Cc_06008, Cc_27456	g53428,
					g43832,
		eIF5A-2	At1g69410		g51683,
		eIF5A-3	At1g13950		g37706
	eIF5B	eIF5B-1	At1g76810	Cc_04638	g22070
		eIF5B-4	At1g21160
eIF6 group	eIF6	eIF6-1	At3g55620	Cc_04146, Cc_17894	g19347,
					g16078

Phylogenetic analysis of eIF family gene sequences

To elucidate the evolutionary relationships and functional classification of eIF genes in pigeonpea, a neighbor-joining phylogenetic tree was constructed using protein sequences from A. thaliana, C. cajan genome v2, and pangenome assemblies (Fig. 1). The analysis revealed that the identified eIF genes clustered into 29 distinct subgroups, each forming separate clades. These subgroups were further grouped into six major categories based on sequence similarity and domain composition: eIF1, eIF2, eIF3, eIF4, eIF5, and eIF6 and previous classification.

Fig. 1.

Phylogenetic relationships of eIF proteins from Cajanus cajan (genome and pangenome) with Arabidopsis thaliana homologs. Distinct eIF subfamilies form separate clades, indicated by different colors. Red dots represent bootstrap support, with larger sizes indicating higher values (maximum = 100)

The eIF1 group included eIF1 and eIF1A clades. The eIF2 group was represented by seven clades: eIF2α, eIF2β, eIF2γ, eIF2Bα, eIF2Bβ, eIF2Bδ, and eIF2Bε–2Bγ. The eIF3 group comprised ten clades: eIF3A, eIF3B, eIF3C, eIF3D, eIF3E, eIF3F, eIF3G, eIF3H, eIF3I, and eIF3K. The eIF4 group included eIF4A, eIF4B, eIF4E, eIFiso4E, eIF4G, and eIFiso4G clades. The eIF5 group consisted of eIF5, eIF5A, and eIF5B clades, while the eIF6 group was represented by a single eIF6 clade.

Subgroups with high sequence similarity were visually distinguished using unique colors and labelled accordingly. Notably, the eIF2Bα subgroup exhibited divergence, with its members distributed across multiple clades. This separation was attributed to domain-level differences within the subgroup, indicating potential functional diversification.

Collinearity and Ka/Ks analysis of identified eIF genes

To further elucidate the syntenic relationships between pigeonpea CceIF genes and those from other Fabaceae members, a collinearity analysis was performed using M. truncatula (model legume) and A. thaliana (reference plant). Out of the 52 CceIF genes identified, only 30 exhibited collinearity with their Arabidopsis orthologs, highlighting evolutionary divergence and potential speciation events within the gene family (Fig. 2A). In contrast, analysis with M. truncatula revealed collinearity for 44 CceIF genes, suggesting closer evolutionary conservation within the legume lineage (Fig. 2B).

Fig. 2.

Synteny analysis of eIF genes between Cajanus cajan and Arabidopsis thaliana (A), and Medicago truncatula (B). Green bars represent chromosomes of C. cajan, while orange bars represent chromosomes of the respective comparison species. Red curves indicate collinear eIF gene pairs between genomes

To assess the evolutionary pressures acting on these gene pairs, we calculated the ratio of non-synonymous to synonymous substitutions (Ka/Ks) between orthologous pairs in Arabidopsis and pigeonpea (Table S3). All Ka/Ks values were below 1, indicating that the eIF gene family has undergone strong purifying selection during its evolution in pigeonpea.

Chromosomal distribution of eIF family genes in pigeonpea

The 52 identified CceIF genes were mapped across all eleven chromosomes of pigeonpea (Table S4). Chromosome 3 (Chr03) contained the highest number, with nine eIF genes. Both Chr08 and Chr11 followed, each harboring seven genes. Chr09 contained five genes, while Chr01, Chr02, Chr05, and Chr07 each had four eIF genes. Three genes were found on both Chr04 and Chr10, and Chr06 had the fewest, with only two genes (Fig. 3). This distribution reflects the broad chromosomal representation and suggests functional diversification of the eIF gene family across the pigeonpea genome.

Fig. 3.

Chromosomal distribution of 52 eIF genes in Cajanus cajan. Chromosomes are shown as yellow bars, with eIF genes marked in red. The scale indicates chromosome length (Mb)

Analysis of intron–exon structure and conserved domains analysis

Structural analysis of the CceIF gene family revealed considerable variation in exon-intron organization, with exon numbers ranging from 2 to 26 across different genes, highlighting substantial structural diversity (Fig. 4; Table S4). Notably, 13 genes (Cc_01505, Cc_02836, Cc_07050, Cc_09668, Cc_11977, Cc_12196, Cc_13610, Cc_17330, Cc_18788, Cc_24468, Cc_24897, Cc_27456, and Cc_27658) were identified as lacking one or both untranslated regions (UTRs), likely due to incomplete gene models or annotation errors. Despite this, these genes retained conserved functional domains and showed sequence homology with their corresponding reference eIF family members and were thus used for further analysis.

Fig. 4.

Gene structure of eIF genes in Cajanus cajan. Exons are shown as red boxes, introns as black lines, and UTRs as blue boxes. Absence of UTRs in some genes likely reflects annotation limitations rather than true absence

Conserved domain analysis revealed that all AteIF, CceIF, and pangenome eIF sequences within each subgroup exhibited unique domain signatures, distinguishing them from other subgroups and enabling clear classification into distinct subfamilies based on domain architecture (Fig. 5). Specifically, the eIF1 group featured SUI1_MOF2 and eIF1_SUI1-like domains, while the eIF2 group predominantly contained IF-2 domains. The eIF3 group included not only canonical eIF-3 domains but also PCI, Smc, PINT, RRM, BRCT_TopBP1_rpt2, MPN, and SAC3_GANP domains across various subgroups. Within the eIF4 group, the eIF4A subgroup displayed DEAD-box helicase domains, whereas the eIF4G subgroup exhibited the presence of MIF4G and MA3 domains. The eIF5 subgroup displayed P-loop_NTPase, W2, and eIF5-specific domains, whereas the eIF6 group was characterized by the IF6 domain.

Fig. 5.

Domain analysis of eIF genes in pigeonpea. Different domains present in six groups of eIF family genes were represented by distinct colored blocks

Cis-regulatory element analysis of the identified eIF genes

To explore the regulatory potential of pigeonpea eIF genes, cis-regulatory elements (CREs) were identified by analyzing 2000 bp upstream sequences from the transcription start sites (Fig. 6). A total of 96 distinct CREs were detected, many of which were associated with responses to environmental stimuli, including light, biotic, and abiotic stresses (e.g., low temperature, drought, anaerobic, and anoxic conditions), as well as binding motifs for transcription factors such as MYB and AT-rich DNA-binding proteins. Additionally, elements related to hormonal regulation—responsive to gibberellin, auxin, methyl jasmonic acid, abscisic acid, and salicylic acid—were also identified, along with CREs linked to growth processes such as circadian rhythms, endosperm expression, and palisade mesophyll cell differentiation. The frequency and distribution of these elements varied across genes, with core promoter motifs (notably TATA-box and CAAT-box) accounting for over half of the elements identified. Notably, stress- and hormone-responsive elements were the most prevalent, suggesting a regulatory role of eIF genes in stress adaptation and hormonal signaling pathways in pigeonpea. Elements associated with salicylic acid responsiveness (TCA-element), jasmonic acid responsiveness (CGTCA-motif and TGACG-motif), and abscisic acid responsiveness (ABRE) showed higher frequency in selected gene clusters. Additionally, defense-related elements such as W-box and stress-responsive motifs (STRE, MYB, and MYC binding sites) were prominently observed.

Fig. 6.

The distribution of regulatory cis-elements in the pigeonpea eIF promoter region

Differential expression of eIF genes in pigeonpea

To elucidate the functional roles of eIF genes in response to PPSMV infection, expression profiling was performed using leaf tissues from both pathogen- and mock-inoculated resistant (RP and RM) and susceptible (SP and SM) pigeonpea genotypes (Fig. 7). While eIF genes were expressed in both genotypes, their expression patterns diverged markedly under pathogen stress. Notably, expression levels were comparable between mock-inoculated samples, but pathogen-inoculated samples exhibited significant genotype-specific differential expression.

Fig. 7.

Relative expression of eIF genes in contrasting pigeonpea genotypes following PPSMV infection, normalized to GAPDH. RP: resistant + PPSMV; RM: resistant + mock; SP: susceptible + PPSMV; SM: susceptible + mock. Error bars indicate mean ± SE (n = 3). Asterisks denote significant differences (* p < 0.05, ** p < 0.01; one-way ANOVA with Tukey’s test)

A focused analysis was conducted on specific eIF subgroups—eIF2 group, eIF3 group, eIF4A, eIF4E, eIFiso4E, eIF4G, eIFiso4G, and eIF5A—previously linked to viral defense in other crops (Table S2). Several genes were significantly upregulated in the pathogen-infected resistant genotype, including Cc_11977 (eIF2β; 14.52 FC), Cc_07682 (eIF3A; 4.24 FC), Cc_07050 (eIF3B; 10.81 FC), Cc_25520 (eIF4A; 32.68 FC) Cc_27658 (eIF4A; 17.50 FC), Cc_18851 (eIF4A; 9.05 FC) Cc_19714 (eIF4E; 9.00 FC), Cc_02582 (eIF(iso)4E; 3.45FC), Cc_16013 (eIF4G; 7.22 FC), and Cc_20884 (eIF(iso)4G; 15.61 FC), among others.

Conversely, genes such as Cc_01505 (eIF4G; 23.52 FC), Cc_27747 (eIF3K; 2.63 FC), and Cc_06008 (eIF5A; 2.66 FC) were upregulated in susceptible genotypes upon pathogen exposure. These results suggest differential activation of eIF gene subgroups may possibly link to resistance and susceptibility mechanisms during PPSMV infection in pigeonpea.

Further, correlation analysis revealed positive linear associations between eIF gene expression and viral load. eIF3K showed a strong correlation (R² = 0.71; y = 0.863x − 1.0235), indicating that increased expression corresponds with higher viral titers. Similarly, eIF4G exhibited a comparable positive relationship (R² = 0.71), with expression increasing in parallel with viral accumulation. eIF5A also displayed a positive correlation (R² = 0.71; y = 0.6516x + 0.175), suggesting coordinated expression with viral progression (Figure S3). Collectively, these results indicate that eIF3K, eIF4G, and eIF5A are positively associated with PPSMV accumulation, supporting their potential roles as susceptibility-associated factors or infection-responsive markers. However, these relationships are correlative and require functional validation to confirm causality.

Variants in the susceptible genotype ICP8863

Uniquely mapped reads were used to identify variants in the susceptible genotype ICP8863. A total of 94 variants were detected across the 52 eIF genes when compared with the reference genome. Of these, only seven variants were classified as of HIGH or MODERATE impact (Table S5). A stop-gained mutation was identified in eIF3E (Cc_02836), suggesting a potential loss-of-function due to the introduction of a premature stop codon. Additionally, an indel in eIF4B (Cc_22348) resulted in a frameshift mutation, indicating premature truncation and likely functional disruption.

Five eIF genes carried missense variants, each involving a single nucleotide substitution that altered the encoded amino acid and potentially affected protein structure (Fig. 8). These included variants in eIF3E (Cc_02836) (Met→Leu), eIF5B (Cc_04638) (Thr→Pro), eIF2Bδ (Cc_06116) (Thr→Arg), eIF2Bα (Cc_18788) (Arg→Trp), and eIF4B (Cc_22348) (Gln→His). Together, these seven genes represent promising candidates for further functional validation to elucidate their roles in PPSMV susceptibility.

Fig. 8.

Circos plot of Cajanus cajan showing chromosomes (C1–C11) in the outer circle, 52 eIF genes in the middle layer, and five genes harboring seven high- and moderate-impact variants in ICP8863 in the inner layer

Discussion

Plant viruses rely on host cellular machinery to complete their infection cycles due to their limited coding capacity. Consequently, mutations that disrupt essential host factors often confer recessive resistance, and such genes are generally termed susceptibility genes or recessive resistance genes [35]. Among these, the eukaryotic initiation factors (eIFs) constitute the most widely recognized group, given their central role in translation, plant development, and stress responses [13, 36]. Several eIFs—including eEF1A, eEF1B, PABPs, eIF2Bβ, eIF3B, eIF4B, eIF4E, and eIF4G—have been implicated in facilitating virus replication and viral protein translation [37]. The absence or disruption of these factors often renders plants resistant to viral infection [38, 39]. Although eIF-mediated resistance has been characterized in several crop species through mutation, RNAi, and genome editing, the eIF gene family in pigeonpea has remained unexplored [12].

In this study, we characterized 52 eIF genes in pigeonpea, examined their phylogenetic relationships, investigated their structural features, and analyzed their transcriptional responses to PPSMV infection in contrasting genotypes. Remarkably, sequence surveys across 89 pigeonpea accessions revealed the consistent presence of all 47 eIF homologs, highlighting the evolutionary stability and likely indispensability of these genes in core cellular processes. Phylogenetic and collinearity analyses further reaffirmed the conserved evolutionary trajectory of the eIF family, with pigeonpea showing stronger synteny with the legume Medicago truncatula than with Arabidopsis. The uniformly low Ka/Ks ratios (< 1) for orthologous pairs underscore strong purifying selection acting on these genes, preserving their essential functions. However, this interpretation should be considered in light of potential biases, including alignment quality and model assumptions.

The eIF genes were distributed across all 11 pigeonpea chromosomes, with notable clusters on chromosomes 3, 8, and 11, suggesting potential local duplication events. Gene structure analysis revealed considerable diversity in exon–intron architecture (2–26 exons), which may promote sub-functionalization or regulatory flexibility. Conserved domain and motif analyses confirmed the presence of canonical elements such as DEAD-box and RNA helicase domains, key components implicated not only in translation but also in early defense responses [38, 40]. Furthermore, promoter analysis identified 96 cis-regulatory elements associated with various biotic and abiotic stresses. The cis-element heatmap (Fig. 6) revealed that several eIF genes harbor multiple stress- and hormone-responsive motifs, including ABRE (abscisic acid-responsive), CGTCA/TGACG motifs (jasmonic acid-responsive), and TCA-elements (salicylic acid-responsive), along with defense-associated elements such as W-box and MYB/MYC binding sites. These elements are known to play central roles in plant defense signaling pathways activated during viral infection [41–43]. The enrichment of such motifs in specific eIF genes, especially the ones upregulated in the resistant genotype, suggests that their transcriptional regulation may be modulated by host defense signaling networks. In particular, the genes that are upregulated in susceptible genotypes under PPSMV infection (e.g., eIF3K, eIF4G, and eIF5A) lacked the defense related motifs indicating their role in successful infection.

Previous research has highlighted the significance of these eIFs in conferring tolerance to biotic stress [44, 45]. Expression profiling under PPSMV infection revealed contrasting transcriptional responses between resistant and susceptible pigeonpea genotypes. Several factors, including eIF2β, eIF3B, eIF4A, eIF4E, and eIFiso4G showed increased expression in the resistant genotype. These findings are consistent with previous evidence linking eIF4E and eIFiso4G allelic variation to virus resistance in other crops [13, 17, 46]. Conversely, eIF3K, eIF4G, and eIF5A were upregulated in the susceptible genotype following infection. These factors have been previously associated with susceptibility in different plant–virus systems—for example, eIF3 in BYDV infection [47], eIF4G in Triticum mosaic virus [48], and eIF5A in SMV- and Pseudomonas-induced defense and programmed cell death [49, 50]. Targeted gene knockouts have successfully conferred virus resistance in several crop species in previous studies. Disruption of eIF4E in winter barley resulted in resistance to Bymovirus by altering the translational reading frame [17], while knockout of eIF4G in rice provided partial resistance to rice black-streaked dwarf virus [51]. Similarly, loss of eIF4E function in melon conferred resistance to Moroccan watermelon mosaic virus [21].

Our analysis of the susceptible genotype ICP8863 revealed a total of 94 sequence variants across the 52 eIF genes, of which seven were classified as having high or moderate functional impact. Of these, a stop-gained mutation in eIF3E and a frameshift indel in eIF4B are likely to result in truncated proteins, implying a loss of function. Moreover, five genes—eIF3E, eIF5B, eIF2Bδ, eIF2Bα, and eIF4B —bear missense mutations that result in amino acid substitutions potentially perturbing protein structure or interactions, which could compromise antiviral defense. These observations align with evidence from other crops, where disruption of eIF genes confers viral resistance. A natural recessive resistance in pepper involves mutations in eIF4E and its isoform (pvr2/pvr6 loci) [52]. In tomato, CRISPR-Cas9 knockout of eIF4E2 successfully restricted potyvirus infection [53], and in wheat, editing of eIF4E improved resistance to wheat yellow mosaic virus (WYMV) without yield penalty [54]. Based on our results, loss-of-function mutations in eIF genes can confer recessive resistance in some plant–virus systems; however, their effects are highly context-dependent, varying with the specific eIF member, mutation type, and virus. Accordingly, we interpret the identified variants as putative associations rather than definitive determinants of susceptibility, noting that some may be non-functional or compensatory. Functional validation is required to establish causality.

Together, our variant data and published functional studies suggest that targeted manipulation of identified pigeonpea eIF genes could be a viable strategy to confer durable resistance to PPSMV. While the present study identifies candidate eIF genes harboring predicted high-impact variants in the susceptible genotype, their direct functional roles in mediating susceptibility to PPSMV remain to be experimentally validated. Future studies could employ targeted genome editing approaches, such as CRISPR/Cas9-mediated knockout of candidate susceptibility genes (e.g., eIF3E and eIF4B) in susceptible backgrounds to assess their contribution to viral infection. Conversely, allele replacement strategies introducing variant alleles into resistant genotypes may help determine whether these mutations alter host–virus interactions.

In addition, complementary approaches such as virus-induced gene silencing (VIGS) or RNA interference (RNAi) could provide rapid functional insights into gene function during infection. Protein–protein interaction assays, including yeast two-hybrid or co-immunoprecipitation, may further elucidate interactions between eIF proteins and viral components. Together, these approaches would enable a comprehensive validation framework to establish causal relationships between eIF gene variation and PPSMV susceptibility.

Conclusions

Overall, our study provides the first comprehensive pan-genome based genomic and comparative analysis of the eIF gene family in pigeonpea. The integrative evaluation of evolutionary features, gene structure, stress-responsive regulatory elements, transcriptional dynamics, and sequence variation highlights several eIF candidates that may contribute to PPSMV resistance or susceptibility. Notably, eIF4G (Cc_01505), eIF3K (Cc_27747), and eIF5A (Cc_06008) showed elevated expression in susceptible plants following infection, highlighting their potential roles in mediating host–virus interactions. These findings offer a valuable foundation for future molecular breeding and genome editing efforts aimed at enhancing durable resistance to sterility mosaic disease in pigeonpea.

Authors’ contributions

Conceptualization: PS and KY; Methodology: PS, NS, AT, HG, NB, and KY; Software and data analysis: PS, NS, AT, and KY; Materials and resources: MS, NB, WT and PG; Writing—original draft preparation: PS, NS and AT; Writing—review and editing: MS, PG, WT, and KY; supervision and funding acquisition: KY.

Funding

This work was supported by core funding from the Indian Council of Agricultural Research (ICAR) and the Breeding for Tomorrow (B4T) initiative from CGIAR to ICRISAT, India.

Data availability

All datasets analysed in this study are publicly available. Arabidopsis thaliana eIF sequences were obtained from the TAIR database. Pigeonpea (Cajanus cajan) eIF sequences and genome annotations were retrieved from CajanusMine (LIS). eIF sequences for other legumes, including Medicago truncatula, were accessed through LIS and Phytozome. Whole-genome resequencing data for ICP 8863 were downloaded from the NCBI SRA under accession SRR5484735 (BioProject PRJNA383013).

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.