Genome-wide identification, characterization and expression profiling of purple acid phosphatases under variable phosphorus regimes in lentil (Lens culinaris Medik.) Authors: Ankita Chakraborty, Arpita Das, Joydeep Banerjee Published: 2025 ABSTRACT BackgroundPurple acid phosphatases, an important sub-class of metallo-phosphoesterase, play a key role in the uptake and homeostasis of phosphorus in plants, especially under phosphorus limited conditions. Phosphorus deficiency in legumes leads to an impaired growth, poor nodulation and a significant decline in yield potential.ResultsIn this study, 22 putative purple acid phosphatase genes in Lens culinaris (LcPAPs) were identified, distributed across seven chromosomes with chromosome 4 harbouring the highest number. In-silico analysis revealed their gene structures, physico-chemical properties, structures of the PAP proteins, and predicted sub-cellular localization. Most of the LcPAPs were found to possess a signal peptide suggesting their involvement in the secretory pathway, whereas LcPAP3 lacking any signal peptide was uniquely localized in the nucleus. Considering molecular weights, the LcPAPs clustered into three groups and were classified in five out of six clades in a phylogenetic tree constructed with PAPs from soyabean, rice and Arabidopsis. Synteny analysis of LcPAPs revealed the highest conservation with soyabean followed by Arabidopsis and least with rice, consistent with the lineage divergence of dicots and monocots that occurred million years ago. Promoter analysis of the LcPAPs revealed the presence of multiple cis elements associated with stress responsive transcription factors, including PHR1, a key regulator of phosphorus starvation response. Docking studies of the protein with ligands like para-nitrophenyl phosphate (pNPP) and phytic acid revealed the involvement of the conserved motifs and different covalent bonds in protein–ligand interaction. Gene expression analysis under differential phosphorus regimes revealed root specific upregulation of LcPAP20, while constitutive upregulation in both root and shoot tissues of LcPAP4 annotated with oxidoreductase and phosphatase enzyme activity under low as well as high phosphorus conditions as compared to control. In contrast, LcPAP9 predicted to possess dual phosphatase-phytase activity was highly upregulated in the shoots under phosphorus deficient condition as compared to phosphorus sufficiency.ConclusionsThese findings provide foundational insights into the functional diversification of LcPAPs on an intricate molecular level, and might pave the way of identifying key regulatory genes of lentil involved in phosphorus homeostasis. Further in-depth bioinformatics study on LcPAPs with their two potential interactors has unravelled their potential role and binding efficacies.Supplementary InformationThe online version contains supplementary material available at 10.1186/s12870-025-07696-z. Background Phosphorus (P) is an indispensable macro-element for plant growth and development, being integral to energy transduction, membrane biogenesis, nucleic acid synthesis, signalling pathway regulation, and the coordination of carbon and nitrogen metabolism [1]. Despite its importance, P availability is limited, as it exists in both inorganic (Pi) and organic forms in the soil. The readily available inorganic forms are the orthophosphate ions (HPO42− and H2PO4−) which are strongly influenced by soil pH. However, much of the inorganic P become immobilized by forming insoluble complexes with calcium, iron and aluminium, resulting in the formation of compounds such as apatite and oxides [2]. Organic P, on the other hand, exists mainly as esters such as phytates, where P is covalently bonded to organic carbon, further restricting bioavailability [3]. This limited P availability restricts uptake, disrupts internal distribution, and negatively impacts plant growth and yield [4, 5]. To combat this, plants have evolved adaptive mechanisms, notably, the exudation of organic acids and purple acid phosphatases (PAPs) enzyme from roots, that mobilizes the free Pi from their bound organic forms [6, 7]. The mobilized P is then absorbed and translocated through phosphate transporters [8]. PAPs belong to the metallo-phosphoesterase family and function by hydrolyzing orthophosphate monoesters to liberate Pi particularly in acidic soils [9]. Their name derived from their characteristic pink or blue colour (λ = 550–560 nm) detected through spectrophotometer, caused by a charge transfer within a bimetallic centre (Fe3⁺–Mn2⁺/Zn2⁺/Fe2⁺) from a conserved tyrosine residue [10]. PAPs are ubiquitous in occurrence available from the prokaryotic bacteria to the higher eukaryotic organisms including plants [11]. PAPs are characterized by the presence of seven conserved metal-ligating amino acid residues consisting of two aspartic acid (D), one tyrosine (Y), one asparagine (N) and three histidine (H). These conserved amino acid residues are distributed across five amino acid motifs- DXG/GDXXY/GNH(D/E)/VXXH/GHXH; where ‘X’ denotes any amino acid [12]. Plant PAPs are divided into three groups with two subgroups in each of them. Group I (subgroup a, b) and II (subgroup a, b) are comprised of PAPs having high molecular weight (~ 55 kDa or more) and are mostly available as homodimers or oligomers; while; Group III (subgroup a, b) PAPs are having low molecular weight of upto ~ 35 kDa which are monomeric in nature and lack the -NH2 domain analogous to mammalian PAPs [13, 14]. Beyond P mobilization, PAPs perform diverse physiological functions across plant species. In Arabidopsis thaliana, AtPAP17 (formerly AtACP5) is associated with ROS (reactive oxygen species) metabolism and biotic defence responses [15], while AtPAP23 is highly expressed in floral tissues suggesting its role in flower development [16]. Glycine max PAP genes such as GmPAP3 are expressed in root tissues under salt stress, independently of Pi availability [17, 18]. Overexpression of AtPAP26 in A. thaliana enhances carbon and P metabolism leading to improved growth and yield of plants, whereas the knock-out mutants exhibited stunted growth and development [19]. Additionally, CaPAP7 identified from C. arietinum contributes to phytate metabolism in seeds [20] and AtPAP7 is involved in cadmium detoxification, indicating a broader role in alleviating heavy metal stress [21]. Several PAPs have been reported across dicot and monocot plant families, highlighting their key roles in P acquisition, internal P cycling, and stress responses. In A. thaliana, 29 PAP genes have been identified of which 12 were upregulated under low P conditions. Notably, AtPAP1 and AtPAP12 were root specific suggesting their plausible role in external phosphate acquisition, while, AtPAP23 was shoot specific, likely contributing in internal P cycling [13]. In G. max, 35 PAP genes have been reported with 10 showing strong root-specific induction under P deficiency, particularly GmPAP4 and GmPAP12. GmPAP4 was found to be associated with root nodules, indicating its role in symbiotic nitrogen fixation [22]. In Jatropha curcas, 25 PAP genes were identified; 6 were upregulated in roots under P starvation and 4 were associated with heavy metal stress tolerance [23]. Similarly in Solanum lycopersicum, 25 PAP genes were reported, showing root specific upregulation under P deficiency, facilitating extracellular P acquisition [24]. In Camellia sinensis 19 PAP genes were reported; CsPAP1, CsPAP7 and CsPAP12 were strongly induced in roots under P stress, while CsPAP15 and CsPAP17 were upregulated in leaves under high iron stress, suggesting roles in ROS detoxification and metal stress mitigation [25]. Among the monocots, 26 PAP genes have been identified in Oryza sativa and 33 in Zea mays. OsPAP3 and OsPAP10 exhibited root specific expression indicating their efficient P scavenging mechanisms under low P conditions [26], while 12 ZmPAPs were reportedly up regulated under low P stress [27]. Interestingly, Triticum aestivum harbours 105 PAP genes, the highest number of PAP genes reported among plants, elucidating their extensive involvement in P metabolism and stress-responsiveness in wheat [28]. Lentil (Lens culinaris Medikus), a cool-season legume, is often termed as the “poor man’s meat” due to its affordability and high protein content, making it a vital dietary component across socio-economic groups [29]. In India and other South Asian countries, lentil is often grown as a rice-fallow crop utilizing the residual soil moistures post rice harvest [30]. However, rice fallow niches are often characterised by poor biophysical properties with reduced soil aggregation and high bulk density, limiting proper root penetration of the rice-fallow crops [31]. In addition, rice cultivation depletes nitrogen, critical for symbiotic nitrogen fixation and reduces phosphorus availability through immobilization in acidic, waterlogged conditions [32]. These constraints severely impact lentil productivity, particularly among resource-poor farmers of South Asia [33]. Therefore, identifying genes involved in P uptake and utilization is crucial, especially for improving productivity in resource-poor farming systems of rice fallow ecology. The present study aimed to: (i) perform genome-wide identification of PAP genes in lentil; (ii) characterize their physico-chemical and structural properties through in-silico analyses; (iii) assess their phylogenetic relationships and synteny with PAPs from related species to unveil the diversity and evolutionary significance of LcPAPs; and (iv) evaluate transcript-level expression of selected LcPAP genes under differential P regimes to elucidate their functional relevance in P uptake and utilization. Materials and methods Identification and bioinformatics analysis of PAP genes from lentil The Phytozome v13 database (http://www.phytozome.net, released January 2020) was explored to identify putative PAP genes in lentil. An initial keyword search using the term “purple acid phosphatase” restricted to L. culinaris, retrieved 12 PAP amino acid sequences. To expand the search, BLAST analysis was performed using previously characterized G. max PAP gene sequences [22] which yielded an additional 10 sequences, partial sequences were also considered. The finally identified 22 lentil PAPs were aligned using Clustal Omega (https://www.ebi.ac.uk/jdispatcher/msa/clustalo, accessed on 10 September, 2024) to confirm the presence of the five conserved amino acid motifs [34]. Further validation of the sequences were carried out against Pfam 36.0 database (http://pfam.xfam.org/, released March 2024) [35] and NCBI Conserved Domain Database v3.20 (https://www.ncbi.nlm.nih.gov/cdd/, released May 2024) with default parameter settings, confirming the presence of metallophosphatase domains [36]. The genomic sequences of the identified lentil PAPs including their corresponding chromosomal location IDs, start and end site of the genes, the total number of exons and introns were retrieved from the phytozome website. The prefix “LcPAP” for L. culinaris PAP was used to name the 22 lentil PAPs based on their subsequent chromosomal locations, following the nomenclature format as reported in earlier literature [13]. The ExPASy ProtParam tool (www.expasy.org, accessed October 2024) was implemented to determine the size and molecular weight of the proteins, their subsequent isoelectric point (pI), instability index, aliphatic index and grand average of hydropathy index (GRAVY) [37]. Online programmes namely, WoLF PSORT (http://www.genscript.com/wolf-psort.html, updated 2017) [38] and CELLO v2.5 (http://cello.life.nctu.edu.tw/) [39] were employed to predict the potential subcellular localizations of the identified lentil PAP proteins. For WoLF PSORT, the localization predictions were considered if raw score was >= 5 and the top-ranked localization per protein was selected. For CELLO v2.5, the predictions with reliability score >= 2.5 were assessed and the one with highest reliability score was selected. The occurrence of signal peptide was predicted by the softwares Signal P 6.0 [40] via the DTU server, the fast model was employed for large-scale screening, and borderline sequences were re-run under the slow/full model to refine cleavage site predictions. Only sequences with posterior probability ≥ 0.5 at the predicted cleavage site were retained as positive signal peptides and potential N-glycosylation sites were predicted by NetNGlyc 1.0 (https://services.healthtech.dtu.dk/services/NetNGlyc-1.0/), potential score >= 0.5 was used to define positive sites, also, the sites with ‘jury agreement’ >= 5/9 were considered high confidence glycosylation prediction sites; while the EC numbers were obtained through analysis of the standard files with TBtools v.2.0 software (https://github.com/CJ-Chen/TBtools/). Chromosomal localization and gene structure analysis MG2C v2.1 software was deployed to generate the chromosomal map illustrating the relative positions and structures of the LcPAP genes following earlier literature [41]. GSDS 2.0 software (http://gsds.cbi.pku.edu.cn/, accessed on 15 September 2024) was used to analyze the gene structures. Pfam v36.0 and MEME suite v5.5.7 (http://meme.nbcr.net/meme/, accessed on 20 September 2024) were implemented to identify and assess the domains as well as conserved motifs available in each of the LcPAP proteins, parameters were set to search for upto 10 motifs, each motif containing 6 to 50 amino acids using the zero or one occurrence per sequence model, E-value was set to < = 1e-5 and visualized them using TBtools v.2.0. (https://github.com/CJ-Chen/TBtools/, accessed on 20 September 2024) [42]. Biological sequences software illustrator v1.0 (http://ibs.biocuckoo.org/) was employed to create a schematic representation of the conserved domains of each protein [43]. Synteny analysis TBtools v.2.0. software was used to identify the orthologs of LcPAP genes using the one step MCScanX module with default parameters. Collinear areas of PAPs among L. culinaris, G. max, O. sativa and A. thaliana were visualized through Dual Synteny Plotter in TBtools v.2.0. using their corresponding General Feature Format (GFF) files [44]. Phylogenetic analysis The PAP protein sequences from L. culinaris (LcPAP), G. max (GmPAP), O. sativa (OsPAP) and A. thaliana (AtPAP) were extracted from the Phytozome v13, NCBI (https://www.ncbi.nlm.nih.gov/, accessed on 20 September, 2024) and TAIR10 (https://www.arabidopsis.org/, released December 2010) databases, respectively. The proteins were aligned using ClustalW programme [45] and a phylogenetic tree was constructed using the MEGA v12.0.7 (accessed on November 2024) programme following the neighbour joining method pertaining to computational efficiency and reliable performance for tree reconstruction for large family proteins having moderate divergence in their sequences. Poisson correction model was used as amino acid substitution model and pairwise deletion was opted for handling missing data/gaps, the bootstrap replicates were limited to 1000 [46], nodes with bootstrap values >= 50% were considered reliable. In a previous study conducted on G. max belonging to Leguminosae family, PAP proteins were also aligned using Clustal X and phylogenetic tree construction was done using neighbour- joining method [22]. To address gaps or missing information in sequences, complete deletion was employed as a data handling option of the software. Following Poisson correction, the distance between the sequences was estimated. The results were refined furthermore using iTOL v6.8 software (https://itol.embl.de/itol.cgi) following previous literature [47]. Analysis of promoter binding sites A 2000 bp upstream region from the translation start site of each predicted PAP gene was extracted from the Phytozome database. The 2000 bp upstream region from the coding sequence was obtained for all the gene sequences using Phytozome and analyzed for the presence of cis-acting regulatory elements using the PlantCARE database (https://bioinformatics.psb.ugent.be/webtools/plantcare/html/, updated 2002) following previous literature [48]. Modelling and analysis of protein structures LcPAP protein sequences were submitted to the SOPMA server (https://npsa.lyon.inserm.fr/cgi-bin/npsa_automat.pl?page=/NPSA/npsa_sopma.html, last updated 2022, accessed on 15 December 2024) for the protein secondary structure prediction using default parameters [49]. The three-dimensional structure of the identified lentil purple acid phosphatase proteins was predicted using the latest version of the AlphaFold framework, AlphaFold3 accessed via the Google DeepMind Alphafold server (https://alphafoldserver.com/, first accessed on March 2025) [50] for incorporating multiple specification formats like metal ions, nucleic acids and ligands with high level of accuracy. The amino acid sequences of the proteins were uploaded to the server and metal co-factors (Fe3+ and Mn2+) were selected as input files to be incorporated in the final modelled protein tertiary structure keeping other parameters as default. The confidence of the predicted models was analysed using multiple quality metrics like pLDDT scores, Predicted Aligned Error (PAE) plots, ipTM and pTM scores [50, 51]. The modelled proteins were visualized with PyMOL v3.0.5 (https://pymol.org/2/). Docking study Two substrates namely, pNPP and phytic acid were used as ligands for docking study. pNPP is a standard artificial substrate widely reported to study the PAP activity [52] whereas; phytic acid (inositol hexakisphosphate, IP6) is a naturally occurring physiological substrate of PAPs mainly found in the form of storage compounds in plants [53]. The three-dimensional structure of ligands was downloaded from PubChem database (updated 2025) in.sdf formats. Molegro Virtual Docker v7.0 software (MVD) (released on May 5, 2019) was used for molecular docking studies of different LcPAPs with the respective ligands [54]. MolDock Scoring (GRID) was employed to evaluate protein–ligand interactions. For the MolDock scoring protocol, internal ES, internal H-bond, and Sp2–Sp2 torsions were included in the ligand evaluation. The MolDock SE search algorithm was applied with 100 runs, a maximum of 1500 iterations, and a maximum of 300 steps. The docking cavities were user-defined, the x, y, z coordinates of the grid centre was fixed at a position in between the Fe3+ and Mn2+ ion coordinates with the search-space radius set at 8 to 12 Angstrom following previous literature [55]. Post- docking refinements included energy minimization and optimization of hydrogen bonds. The software was set to generate and rank the top 5 binding poses for each ligand. The energy of interaction between the protein and ligand were assessed by MolDock score, Rerank score and HBond score. The MolDock score represents the initial interaction energy between the protein and ligand in addition to the ligand’s internal energy; the Rerank score offers a more detailed information on the functionality of the interaction, elucidating the role of electrostatic forces, steric interactions, geometrical attributes of hydrogen bonds and effects of desolvation, providing a more refined outlook of the interaction vigour between the protein and ligand; the HBond score explains the sole contribution of the hydrogen bonds in stabilizing the interaction between the ligand and protein. The higher negative values of these scoring functions assure improved protein–ligand binding [54]. The best poses exhibiting the highest MolDock, Rerank and HBond scores were selected for 3D visualization in MVD software and the possible ligand–protein interactions were visualized by PoseEdit function of the ProteinsPlus web server (https://proteins.plus/, accessed on 22 February 2025) [56]. Plant growth condition Two lentil genotypes were chosen for the experiment namely BCL1001, an interspecific advanced breeding line derived from cross between ILL 8006 (L. culinaris ssp. culinaris) × ILWL62 (L. culinaris ssp. orientalis), and Moitree (WBL 77), a popular released variety of Eastern part of India, were collected from Bidhan Chandra Krishi Viswavidyalaya (BCKV, India) centre of All India Coordinated Research Project (AICRP) for Rabi (winter) pulse. Lentil (Lens culinaris) seeds were surface sterilized, soaked for 4–6 h, and allowed to germinate on moist sterile filter paper for 3–4 days. Uniform seedlings were then moved to hydroponic tubs filled with full-strength Hoagland nutrient solution, with four seedlings placed at each site. The seedlings were cultivated in a controlled-environment chamber maintained at 25 °C ± 2 °C, with a 16-h light and 8-h dark photoperiod, approximately 300 µmol·m⁻2·s⁻1 of photosynthetically active radiation (PAR), and 60% ± 5% relative humidity. The nutrient solution was continuously aerated and replaced every 3–4 days to keep pH and nutrient levels stable. The experiments consisted of three biological replicates, with tissue samples collected at specified developmental stages and at the same time each day to reduce diurnal variation. After germination, the plants were grown for 21 days in full strength Hoagland media [57] containing varying levels of P concentrations viz 5 µM, 1 mM and 2 mM, where 1 mM was considered as control condition, 5 µM as low P condition and 2 mM as high P condition, other nutrients included 5 mM KNO3, 4 mM Ca(NO3)2, 2 mM MgSO4, 1 mM NH4NO3; 0.1 mM each of FeSO4, EDTA-Na and H3BO3; 0.13 mM MnSO4, 0.03 mM ZnSO4, 1 µM Na2MoO4 and 0.1 µM CuSO4. After 21 days, the plants were analyzed for their root morphology. RNA isolation and differential gene expression analysis After 21 days, total RNA was extracted from the shoot and root of lentil plants growing under varying P levels, using the RNeasy Plant Mini Kit (QIAGEN, CA, USA) following manufacturer’s protocol. RNA samples with A260/280 value equal to or above 1.8 were selected for further processing. The extracted RNA was put through DNase I treatment (Thermo Scientific ™) followed by cDNA synthesis with Verso cDNA synthesis kit by Thermo Scientific ™. The synthesized cDNAs were subjected to quantitative real-time PCR (qRT- PCR) analysis using PowerUp SYBR Green Master Mix by Thermo Scientific ™ to determine the relative expression of LcPAPs in root and shoot tissues under varying P regimes. All reactions were executed with QuantStudio™ 3 Real-Time PCR System. Ubiquitin, a housekeeping gene, was considered as endogenous control. The list of specific primers is provided in the Additional file 1: Supplementary Table 1. Protein–protein interaction study The protein–protein interaction analysis was performed using STRING v12 software (released April 2024). The proteome of Lens culinaris was retrieved from the phytozome database and subsequently uploaded to the STRING software with a proteome ID-STRG0A97XIQ (https://version-12-0.string-db.org/organism/STRG0A97XIQ) for functional annotation, the interactions were filtered with a confidence cut-off score 0.5. The protein–protein interaction network for all the 22 LcPAP proteins were analyzed, the obtained sequences of the interacting proteins from L. culinaris were investigated through BLAST search in the NCBI database for their identification. Satistical analysis OriginPro 2023, OriginLab Corporation, Northampton, MA, USA was used for all statistical analysis. For gene-expression study, two-tailed paired t-test was done for respective PAP genes using three biological samples and the experiment was performed in triplicates. Results Genome-wide identification of PAPs A total of 22 putative LcPAPs genes were identified through a comprehensive search in the phytozome database. These genes were unevenly distributed across all the chromosomes, with chromosome 4 harbouring the maximum number of nine PAPs, followed by chromosome 3 with four and chromosome 2 having three LcPAPs. Chromosome 1 and 7 each contained two LcPAP genes, while chromosome 5 and 6 harboured one LcPAP gene in each. Gene nomenclature was assigned based on the chromosomal location of the identified LcPAP gene sequences (Fig. 1). Gene structure analysis revealed that all the LcPAP genes possessed both 5ꞌ and 3ꞌ untranslated region (UTR), except LcPAP14 and LcPAP20, which lacked both UTRs, while LcPAP8 had only 5ꞌ UTR (Fig. 2).Fig. 1Chromosomal arrangement of the 22 LcPAP genes distributed across 7 chromosomes of lentil. Physical distance (Mb) of the chromosomes is represented by the scale ruler on the left sideFig. 2Motif, domain and gene structure analysis of the 22 LcPAP genes Chromosomal arrangement of the 22 LcPAP genes distributed across 7 chromosomes of lentil. Physical distance (Mb) of the chromosomes is represented by the scale ruler on the left side Motif, domain and gene structure analysis of the 22 LcPAP genes The amino acid sequences of LcPAP1, LcPAP4, LcPAP10, LcPAP11, LcPAP14, LcPAP15, LcPAP16, LcPAP17 and LcPAP18 retrieved from phytozome database, contained all the seven conserved metal ligating residues, characterizing PAP metallo-esterase (DXG/GDXXY/GNH(D/E)/VXXH/GHXH), distributed across five amino acid blocks. In contrast, LcPAP2 was devoid of the motif 2, while motif 3 was absent in LcPAP20. Additionally, LcPAP2, 3, 5, 6, 7, 8, 9, 12, 13, 19, 20, 21 and 22 were found to be lacking the conserved amino acid residues of motif 4. Interestingly, LcPAP3 and LcPAP8a showed a complete absence of motif 4, while, LcPAP20 lacked both motif 4 and motif 5 (Fig. 3).Fig. 3Diagrammatic representation of the conserved amino acid motifs containing the metal ligating amino acids of the LcPAP proteins. Metal ligating amino acids are represented as bold among the amino acid motifs Diagrammatic representation of the conserved amino acid motifs containing the metal ligating amino acids of the LcPAP proteins. Metal ligating amino acids are represented as bold among the amino acid motifs Although, many LcPAPs lacked the complete set of consensus motifs, many exhibited conserved sequences retaining most of the metal ion binding residues at exact positions. For instance, LcPAP2 and LcPAP3 displayed a modified version of motif 1 preserving the consensus pattern with meagre number of amino acid substitutions. The metal ligating tyrosine (Y) residue of motif 2 was missing in LcPAP2 forming a consensus sequence of GDXXT. Similarly, LcPAP20 exhibited an alteration in motif 3, represented by consensus sequence ‘GNQE’ instead of the conserved GNH(D/E). Furthermore, several LcPAPs namely LcPAP2, LcPAP5, LcPAP6, LcPAP7, LcPAP9, LcPAP12, LcPAP13, LcPAP19, LcPAP21 and LcPAP22 displayed an alteration in the fourth amino acid block forming a consensus sequence ‘XXXH’ (Additional file 1: Supplementary Table 2). The physico-chemical properties of the 22 identified LcPAPs were analysed extensively. It was detected that 15 of these proteins had a molecular weight above ~ 45 kDa, classifying them as high molecular PAPs. Their molecular masses ranged from 46.84 kDa (LcPAP3) containing 415 amino acids to 73.57 kDa (LcPAP15) with 654 amino acids. The remaining 7 were categorized as low molecular weight PAPs, ranging from 33.27 kDa (LcPAP14) having 287 amino acids to 43.10 kDa (LcPAP2) with 383 amino acids. The lowest and the highest molecular weight was recorded for LcPAP14 (33.27 kDa) and LcPAP15 (73.57 kDa), respectively. Among the identified LcPAPs, pI ranged from 5.14 (LcPAP9) to 8.1 (LcPAP2). The instability index spanned from 28.84 (LcPAP4) to 46.63 (LcPAP5), while, the range of aliphatic index which reflects thermostability was from 64.88 (LcPAP14) to 90.66 (LcPAP17). The GRAVY, indicative of overall hydrophobicity or hydrophilicity, varied from −0.22 (LcPAP2) to −0.601 (LcPAP8), suggesting that all the identified LcPAP protein were hydrophilic. Signal peptide analysis revealed peptide length ranged from 16 (LcPAP16) to 34 (LcPAP17) amino acids. Notably, LcPAP3 was the only member lacking any signal peptide and was predicted to be located in the nucleus through bioinformatics analyses. In contrast, the remaining LcPAPs were predicted to be localized predominantly in the extracellular region or organelle like lysosome, chloroplast and vacuole along with their presence in the cytoplasm. LcPAP8 and LcPAP17 were exclusively localized in the cytoplasm whereas LcPAP5 was targeted to the plasma membrane (Additional file 1: Supplementary Table 3). Prediction of potential N-linked glycosylation sites revealed considerable variability among the LcPAP proteins. The number of glycosylation sites ranged from one (LcPAP4, LcPAP16, LcPAP17, LcPAP18, LcPAP20) to eight (LcPAP5). Specially, LcPAP2, LcPAP3, LcPAP8 harboured three sites; LcPAP6, 12 and 14 possessed four. Five glycosylation sites were present in LcPAP1, 7, 10, 11, 13, 19 and 21; six were observed in LcPAP15 and 22, while, seven in LcPAP9 (Additional file 1: Supplementary Table 3). Synteny and phylogeny analysis Synteny analysis depicted that twelve out of the twenty-two identified LcPAP genes, located on chromosomes 2, 3, 4, 5 and 7 of L. culinaris shared collinear relationship with twelve chromosomes of G. max (Fig. 4, Additional file 1: Supplementary Table 4), while, eight of the LcPAP genes from chromosomes 2, 3, 4, 5 and 7 were having syntenic association with five chromosomes of A. thaliana. Notably, only LcPAP16 of chromosome 4 displayed synteny with one chromosome of O. sativa (Fig. 4). Among the LcPAP family members, LcPAP16 was emerged as the most versatile, possessing the highest number of orthologous pairs, two with G. max, two with A. thaliana and one with O. sativa. LcPAP22 also exhibited a broad range of evolutionary relationship with two orthologous pairs each in G. max and A. thaliana. Similarly, LcPAP3, LcPAP9, LcPAP16 and LcPAP18 showed 2 orthologous in G. max, and one in A. thaliana, whereas, LcPAP15 displayed the reverse pattern. LcPAP19 was found to be linked with a single gene each from G. max and A. thaliana. The remaining LcPAPs viz. LcPAP4, LcPAP5, LcPAP6, LcPAP10, LcPAP17 and LcPAP21 revealed orthologous pairs with either G. max or A. thaliana but not both. Overall, the highest covariance existed between G. max genome, followed by A. thaliana, and then O. sativa, indicating a higher degree of evolutionary conservation of the LcPAP gene family with other leguminous species (Additional file 1: Supplementary Table 4).Fig. 4Synteny analysis between LcPAPs obtained from Lens culinaris and PAPs from Glycine max, Oryza sativa and Arabidopsis thaliana. The syntenic blocks are represented by grey lines while red lines represented duplicated PAP gene pairs among species Synteny analysis between LcPAPs obtained from Lens culinaris and PAPs from Glycine max, Oryza sativa and Arabidopsis thaliana. The syntenic blocks are represented by grey lines while red lines represented duplicated PAP gene pairs among species The amino acid sequences of LcPAPs were aligned with the representative amino acid sequences of PAPs from G. max (GmPAP), A. thaliana (AtPAP) and O. sativa (OsPAP) to construct a phylogenetic tree. The resulting dendrogram delineated six major clades, out of which five clades (Clade- I to IV and VI) were comprised of 22 LcPAP proteins, whereas, clade V was devoid of any lentil PAP proteins. Clade III exhibited the highest representation, comprising 10 LcPAP members, suggesting a possible expansion or conservation of this group in lentil. Clade I consisted of LcPAP2 and 3, both sharing similar molecular weights. Clade II comprised of LcPAP4, 16, 17 and 18, with molecular weights predominantly ranged in between 35- 40 kDa. On contrary, the high molecular weight LcPAPs were clustered mostly in clade III, indicative of their close evolutionary proximity. Clade IV comprised of LcPAP5 along with 9, and clade VI comprised of LcPAP1, 10, 11 and 14 (Fig. 5A). Complementary to the phylogenetic classification, LcPAPs were further divided into three groups and six subgroups according to their molecular weight. Group I encompassed predominantly high molecular weight PAPs, with subgroup Ib having members with molecular weights more than 70 kDa. Group II was represented by PAPs with intermediate molecular weights ranging mainly from 50 to 65 kDa. Group III consisted of all the low molecular weight PAPs, weights ranging from 35 to 47 kDa (Fig. 5B).Fig. 5A Phylogenetic relations of purple acid phosphatase genes from Lens culinaris, Glycine max, Oryza sativa and Arabidopsis thaliana, number of bootstraps were limited to 1000. Red circles represent the lentil PAPs occupying Clades I, II, III, IV and VI. B A phylogenetic tree representing the position of the LcPAPs into groups and subsequent sub-groups according to their molecular weights A Phylogenetic relations of purple acid phosphatase genes from Lens culinaris, Glycine max, Oryza sativa and Arabidopsis thaliana, number of bootstraps were limited to 1000. Red circles represent the lentil PAPs occupying Clades I, II, III, IV and VI. B A phylogenetic tree representing the position of the LcPAPs into groups and subsequent sub-groups according to their molecular weights Study of transcription factor (TF) binding sites A comprehensive analysis of the 2000 base-pair upstream promoter region from TSS of the 22 LcPAPs detected 368 TF binding sites, indicating a complex and diverse regulatory landscape. Binding site of MYC was the most prevalent, followed by MYB binding site. A total of 97 MYC, 47 MYB binding sites were detected along with 9 MYB recognition sites and 15 MBS. Additionally, 27 ERE (ethylene-responsive elements), 48 ABRE (abscisic acid-responsive elements), 5 ABRE3a, and 10 ABRE4 variants were identified. Elements involved in wound and pathogen response like 8 WRE3, 13 W box, 11 Wun motif were also present. Photo-regulatory and stress-related elements like 53 G-box along with rarer motifs such as 1 E2Fb, 2 Sp1, 1 P-box and 2 MRE binding sites, further enriched the cis-regulatory landscape. Importantly, 19 PHR1 Binding Site (P1BS), a key promoter binding element for genes involved in Pi starvation response was also reported. LcPAP2 and LcPAP12 exhibited the highest number of P1BS, indicating their likely pivotal roles in Pi-responsive regulatory networks (Fig. 6 and Additional file 1: Supplementary Table 5).Fig. 6Transcription factor binding sites present in the upstream 2000 bp promoter region of LcPAP genes Transcription factor binding sites present in the upstream 2000 bp promoter region of LcPAP genes Protein structure prediction and protein–ligand interaction analysis Based on the phylogenetic analysis, two selected PAPs from each of the clades were further analysed in detail- LcPAP2 and 3 from clade I, LcPAP4 and 17 from clade II, LcPAP15 and 20 from clade III, LcPAP5 and 9 from clade IV and LcPAP1 and 11 from clade VI. Except for LcPAP1, 2, 3, 4 and 5 in which beta-turns were absent, the secondary structures of most of the LcPAP proteins were composed of alpha helix, beta- sheets, beta-turns and random coils (Fig. 7) (Table 1). In all the three-dimensional structures generated for the analysed LcPAPs except LcPAP20, four metal-ion binding domains were identified in the central core region of the protein along with two metal ions (Fig. 8). In case of LcPAP20, one metal ion binding domain was found to be located away from the central core region (Fig. 8). Furthermore, in case of LcPAP15, along-with the existing four domains, another new metal-ion binding domain was identified which was located away from the central core region. (Fig. 8). Model confidence was evaluated using the predicted local distance difference test (pLDDT) results. High to very high confidence regions were defined as having pLDDT > 70, whilst moderate confidence regions were defined as having pLDDT ≤ 70. Regions with pLDDT < 50 were classified as low confidence/disordered and were excluded from further analysis. The model with the highest average pLDDT and the most consistent overall fold (as determined by secondary structure integrity and metal-binding site geometry) was chosen when several model predictions were produced for each protein. Conserved active-site residues were compared with homologous empirically determined PAP structures from the Protein Data Bank (PDB) in order to reach a consensus in cases where there were slight differences between models. The predicted TM-score (pTM) and interface predicted TM-score (ipTM) > 0.7 were considered for the protein models suggesting reliable global folding, consistent with the topologies of the known purple acid phosphatases.Fig. 7Secondary structure of the LcPAP proteins predicted using SOPMA software depicting the presence of alpha helix, beta sheet, turns and coilsTable 1Proportion of alpha helix, beta sheets, turns and coils present in the secondary structure of LcPAP proteins obtained using SOPMA softwareProtein templateAlpha helixBeta sheetTurnCoilLcPAP197 (20.59%)101 (21.44%)0273 (57.96%)LcPAP290 (23.50%)71 (18.54%)0222 (57.96%)LcPAP393 (22.41%)67 (16.14%)0255 (61.45%)LcPAP471 (20.82%)75 (21.99%)0195 (57.18%)LcPAP5111 (19.34%)100 (17.42%)0363 (63.24%)LcPAP6101 (16.19%)135 (21.63%)23 (3.69%)365 (58.49%)LcPAP793 (15.10%)134 (21.75%)23 (3.73%)366 (59.42%)LcPAP8114 (23.75%)88 (18.33%)32 (6.67%)246 (51.25%)LcPAP987 (15.96%)111 (20.37%)21 (3.85%)326 (59.82%)LcPAP1071 (15.24%)110 (23.61%)19 (4.08%)266 (57.08%)LcPAP1176 (16.34%)107 (23.01%)21 (4.52%)261 (56.13%)LcPAP1290 (14.33%)142 (22.61%)28 (4.46%)368 (58.60%)LcPAP1399 (15.79%)145 (23.13%)23 (3.67%)360 (57.42%)LcPAP1464 (22.30%)60 (20.91%)12 (4.18%)151 (52.61%)LcPAP1594 (14.37%)135 (20.64%)25 (3.82%)400 (61.16%)LcPAP1670 (21.60%)73 (22.53%)16 (4.94%)165 (50.93%)LcPAP1782 (23.70%)74 (21.39%)16 (4.62%)174 (50.29%)LcPAP1877 (22.85%)78 (23.15%)17 (5.04%)165 (48.96%)LcPAP19103 (16.83%)139 (22.71%)25 (4.08%)345 (56.37%)LcPAP2034 (10.79%)69 (21.90%)12 (3.81%)200 (63.49%)LcPAP21105 (17.07%)146 (23.74%)26 (4.23%)338 (54.96%)LcPAP22112 (18.09%)138 (22.29%)26 (4.20%)343 (55.41%)Fig. 8Tertiary structure of the LcPAP proteins modelled using Alphafold3 software depicting the presence of alpha helix, beta sheet, turns and coils. The orange sphere represents the Fe3⁺ ion and the purple sphere represents the Mn2⁺ ion forming the bi-metallic nuclear centre. The red, green, purple, cyan and yellow colours represent the motifs 1, 2, 3, 4 and 5, respectively Secondary structure of the LcPAP proteins predicted using SOPMA software depicting the presence of alpha helix, beta sheet, turns and coils Table 1Proportion of alpha helix, beta sheets, turns and coils present in the secondary structure of LcPAP proteins obtained using SOPMA softwareProtein templateAlpha helixBeta sheetTurnCoilLcPAP197 (20.59%)101 (21.44%)0273 (57.96%)LcPAP290 (23.50%)71 (18.54%)0222 (57.96%)LcPAP393 (22.41%)67 (16.14%)0255 (61.45%)LcPAP471 (20.82%)75 (21.99%)0195 (57.18%)LcPAP5111 (19.34%)100 (17.42%)0363 (63.24%)LcPAP6101 (16.19%)135 (21.63%)23 (3.69%)365 (58.49%)LcPAP793 (15.10%)134 (21.75%)23 (3.73%)366 (59.42%)LcPAP8114 (23.75%)88 (18.33%)32 (6.67%)246 (51.25%)LcPAP987 (15.96%)111 (20.37%)21 (3.85%)326 (59.82%)LcPAP1071 (15.24%)110 (23.61%)19 (4.08%)266 (57.08%)LcPAP1176 (16.34%)107 (23.01%)21 (4.52%)261 (56.13%)LcPAP1290 (14.33%)142 (22.61%)28 (4.46%)368 (58.60%)LcPAP1399 (15.79%)145 (23.13%)23 (3.67%)360 (57.42%)LcPAP1464 (22.30%)60 (20.91%)12 (4.18%)151 (52.61%)LcPAP1594 (14.37%)135 (20.64%)25 (3.82%)400 (61.16%)LcPAP1670 (21.60%)73 (22.53%)16 (4.94%)165 (50.93%)LcPAP1782 (23.70%)74 (21.39%)16 (4.62%)174 (50.29%)LcPAP1877 (22.85%)78 (23.15%)17 (5.04%)165 (48.96%)LcPAP19103 (16.83%)139 (22.71%)25 (4.08%)345 (56.37%)LcPAP2034 (10.79%)69 (21.90%)12 (3.81%)200 (63.49%)LcPAP21105 (17.07%)146 (23.74%)26 (4.23%)338 (54.96%)LcPAP22112 (18.09%)138 (22.29%)26 (4.20%)343 (55.41%) Proportion of alpha helix, beta sheets, turns and coils present in the secondary structure of LcPAP proteins obtained using SOPMA software Tertiary structure of the LcPAP proteins modelled using Alphafold3 software depicting the presence of alpha helix, beta sheet, turns and coils. The orange sphere represents the Fe3⁺ ion and the purple sphere represents the Mn2⁺ ion forming the bi-metallic nuclear centre. The red, green, purple, cyan and yellow colours represent the motifs 1, 2, 3, 4 and 5, respectively The interactions involved in pNPP- LcPAP binding were mainly non-covalent interactions like hydrogen bonds, metal interactions, pi-pi interactions, ionic interactions and hydrophobic contacts (Additional file 1: Supplementary Table 6). Exceptionally, apart from the other types of interactions, LcPAP4 reported a cation-pi interaction with pNPP (carbon atom of pNPP like C2, 5, 7, 9, 11, 13 with the NH2 atom of Arg 277) (Additional file 1: Supplementary Table 6). The docking study revealed that pNPP bound directly in the bimetallic active site pocket of the proteins forming metal interactions between the ligand (Figs. 9, 10 and 11). Both metal ions of the LcPAPs were actively involved in interaction with the ligand pNPP except for LcPAP2 and LcPAP3 in which at least one of the metal ions was involved in interaction with pNPP. Except for LcPAP20, all other pNPP- LcPAP interactions were stabilized by hydrophobic contacts. Interestingly, LcPAP1, 9, 11 and 20 did not depict any ionic interactions with pNPP (Fig. 10). Among the LcPAPs, the best poses of LcPAP- pNPP interaction according to MolDock score as well as Rerank score were reported by LcPAP15 (Table 2). On the other hand, hydrogen bond, ionic or electrostatic bonds and metal bonds were involved in stabilizing the phytic acid-LcPAP interactions. LcPAP1 and LcPAP3 were devoid of any metal interactions during their binding to phytic acid ligand while; LcPAP2, 9 and 20 did not exhibit ionic interactions for the same (Fig. 11). According to MolDock score, the best interaction status was reported by LcPAP4 with phytic acid ligand while; LcPAP17 reported the most stable interaction with phytic acid ligand according to the Rerank score (Table 2). Expression profiles of the selected genotypes were evaluated in both shoot and root tissues, normalized against their control condition (1 mM P). Morphological variations in root structure were evident when grown under varying levels of phosphorus availability (Fig. 12).Fig. 9Three-dimensional docking poses of the ligands (pNPP and phytic acid) with LcPAPsFig. 10Two-dimensional representation of the docking interactions of LcPAPs with the ligand pNPPFig. 11Two-dimensional representation of the docking interactions of LcPAPs with the ligand phytic acidTable 2Details of the MolDock score, Rerank score and Hbond score obtained from different LcPAP protein–ligand interactionsProtein nameLigand nameMolDock scoreRerank scoreHbond scoreLcPAP1pNPP−105.48−70.9−7.22LcPAP1Phytic acid−203.12−110.93−14.08LcPAP2pNPP−52.15−46.44−2.74LcPAP2Phytic acid−203.64−88.06−8LcPAP3pNPP−130.35−79.25−9.34LcPAP3Phytic acid−261.59−147.32−11.67LcPAP4pNPP−128.5−49.91−7.48LcPAP4Phytic acid−338.95−181.98−13.24LcPAP5pNPP−104.57−74.34−4.85LcPAP5Phytic acid−213.97−148.79−14.64LcPAP9pNPP−104.12−50.56−5.22LcPAP9Phytic acid−233.53−68.41−12.25LcPAP11pNPP−104.35−76.99−2.45LcPAP11Phytic acid−171.98−57.63−3.15LcPAP15pNPP−132.48−95.56−4.06LcPAP15Phytic acid−267.22−20.59−3.36LcPAP17pNPP−98.28−68.17−7.79LcPAP17Phytic acid−336.93−195−22.03LcPAP20pNPP−83.52−62.36−4.04LcPAP20Phytic acid−214.21−122.77−8.66Fig. 12Morphological root variations in genotypes BCL 1001 (A, B, C) and Moitree (D, E, F) when grown under varying conditions of phosphorus availability. A, D Represents the genotypes grown under P-deficient conditions (5 µM P); B, E represents the genotypes grown under control conditions (1 mM P); while, C, F represents the genotypes grown under P-sufficient conditions (2 mM P) Three-dimensional docking poses of the ligands (pNPP and phytic acid) with LcPAPs Two-dimensional representation of the docking interactions of LcPAPs with the ligand pNPP Two-dimensional representation of the docking interactions of LcPAPs with the ligand phytic acid Table 2Details of the MolDock score, Rerank score and Hbond score obtained from different LcPAP protein–ligand interactionsProtein nameLigand nameMolDock scoreRerank scoreHbond scoreLcPAP1pNPP−105.48−70.9−7.22LcPAP1Phytic acid−203.12−110.93−14.08LcPAP2pNPP−52.15−46.44−2.74LcPAP2Phytic acid−203.64−88.06−8LcPAP3pNPP−130.35−79.25−9.34LcPAP3Phytic acid−261.59−147.32−11.67LcPAP4pNPP−128.5−49.91−7.48LcPAP4Phytic acid−338.95−181.98−13.24LcPAP5pNPP−104.57−74.34−4.85LcPAP5Phytic acid−213.97−148.79−14.64LcPAP9pNPP−104.12−50.56−5.22LcPAP9Phytic acid−233.53−68.41−12.25LcPAP11pNPP−104.35−76.99−2.45LcPAP11Phytic acid−171.98−57.63−3.15LcPAP15pNPP−132.48−95.56−4.06LcPAP15Phytic acid−267.22−20.59−3.36LcPAP17pNPP−98.28−68.17−7.79LcPAP17Phytic acid−336.93−195−22.03LcPAP20pNPP−83.52−62.36−4.04LcPAP20Phytic acid−214.21−122.77−8.66 Details of the MolDock score, Rerank score and Hbond score obtained from different LcPAP protein–ligand interactions Morphological root variations in genotypes BCL 1001 (A, B, C) and Moitree (D, E, F) when grown under varying conditions of phosphorus availability. A, D Represents the genotypes grown under P-deficient conditions (5 µM P); B, E represents the genotypes grown under control conditions (1 mM P); while, C, F represents the genotypes grown under P-sufficient conditions (2 mM P) The expression patterns of LcPAP5 and LcPAP20 were consistently down-regulated in shoots under both P sufficient and deficit condition. The expression of LcPAP15 exhibited a notable up regulation under low P conditions, while down regulation under high P conditions in shoot tissues. The expression of LcPAP2, 3, 9, 11 and 17 were upregulated in shoot tissue and subsequently downregulated in root tissues across both P regimes. In root tissues, LcPAP4 and 20 were upregulated under both low and high P conditions, but interestingly, LcPAP1, 5 and 17 were upregulated under high P while down regulated under low P conditions almost similar in root tissues (Fig. 13). Additionally, in shoot tissues, the expression of LcPAP2 was significantly higher under high P condition, also, the expression of LcPAP17 was significantly higher under both P regimes. While in root tissues, the expression of LcPAP2 was significantly lower under low P conditions, but the expression of LcPAP9 was significantly downregulated across both P regimes (Fig. 13).Fig. 13Relative gene expression of LcPAP genes in shoot tissues (A) and root tissues (B) grown under P-deficient (5 µM P) and P-sufficient (2 mM P) conditions. Data in the figure are mean of three replicates ± S.E. * denotes p-value < 0.05, ** denotes p-value < 0.01 and *** denotes p-value < 0.001 Relative gene expression of LcPAP genes in shoot tissues (A) and root tissues (B) grown under P-deficient (5 µM P) and P-sufficient (2 mM P) conditions. Data in the figure are mean of three replicates ± S.E. * denotes p-value < 0.05, ** denotes p-value < 0.01 and *** denotes p-value < 0.001 Predicted interaction of LcPAPs with other proteins Protein–protein interaction study revealed that LcPAP2, 4, 5, 9, 15, 16, 17, 18 and 21 displayed acceptable confidence scores of interactions (> 0.5) with other target proteins and those were studied in detail (Fig. 14). LcPAP4, 16, 17 and 18 showed interactions with a bi-functional riboflavin kinase/flavin mononucleotide (FMN) phosphatase, riboflavin synthase, flavin adenine dinucleotide (FAD) synthetase, FMN hydrolase and nudix hydrolase (Fig. 14). LcPAP5, 9, 15 were found to show interaction with callose-synthase protein and DL-glycerol-3-phosphatase with a good confidence score (> 0.5). LcPAP4 displayed an interaction with another PAP protein LcPAP16, with a good confidence score of 0.86, while LcPAP17 and LcPAP18 was also observed to be interacting with LcPAP4 with a confidence score of 0.86 (Additional file 1: Supplementary Table 7).Fig. 14Protein–protein interaction study revealing the interactions of LcPAP2, 4, 5, 9, 15, 16, 17, 18 and 21 with the respective proteins of lentils, obtained through STRING software. The red enclosure represents the common interacting proteins among LcPAP4, 16, 17 and 18; while the blue enclosure represents the common interacting proteins among LcPAP5, 9 and 15. The identified LcPAPs interacting with each other are enclosed in black Protein–protein interaction study revealing the interactions of LcPAP2, 4, 5, 9, 15, 16, 17, 18 and 21 with the respective proteins of lentils, obtained through STRING software. The red enclosure represents the common interacting proteins among LcPAP4, 16, 17 and 18; while the blue enclosure represents the common interacting proteins among LcPAP5, 9 and 15. The identified LcPAPs interacting with each other are enclosed in black Discussion Although, the PAP genes have previously been reported in several plant species like O. sativa, G. max, A. thaliana, J. curcas, C. arietinum etc. [21], the present study marks the first comprehensive identification of 22 PAP genes in L. culinaris. Gene duplication events are pivotal drivers of genomic evolution and expansion of gene families [58]. The predominant localization of PAP genes in chromosomes 3 and 4 of L. culinaris might prove to be supportive evidence of the tandem and segmental gene duplication, likely associated with various structural rearrangements of the chromosomes [59, 60] (Additional file 1: Supplementary Table 8). Consistent with earlier studies in A. thaliana and G. max [13, 26] regarding the categorisation of PAPs considering their molecular weights, LcPAPs were also cladded into three main groups and six subsequent sub-groups (Ia & b, IIa & b and IIIa & b). The molecular weights of LcPAP ranging from 33 to 74 kDa and pI ranging from 5.14 to 8.1, aligning with PAPs identified in tomato [24]. The identified LcPAPs proteins were distributed in five out of six clades in the phylogenetic tree constructed using PAP sequences from G. max, A. thaliana and O. sativa, indicating a conserved yet diversified evolutionary relationship among these orthologs across monocot and dicot lineages. Clade III consisted of several PAPs possessing mutated motif 4 (VXXH to XXXH) as displayed by the presence of LcPAP6, 7, 12, 13, 19, 21, 22, or complete absence of motif 4 as observed in LcPAP8 and 20. Notably, LcPAP15 of clade III carried an additional motif 3 (GNHD/E) with highest molecular weight (73.57 kDa), formed a subgroup with other large PAPs such as GmPAP35, AtPAP2, AtPAP9, and OsPAP9a reported in earlier study [22]. Clade IV members, LcPAP5 and 9 showed a different mutation in motif 4 (VXXH to ATXH). LcPAP4 clustered with AtPAP17, a protein which was reported to exhibit peroxidase activity [61]. The Enzyme Commission (EC) numbers predicted for LcPAP4 were EC 1.11.1.7 and EC 3.1.3.2. EC:1.11.1.7 encodes a class of oxidoreductases, specifically peroxidases involved in hydrogen peroxide metabolism, whereas, EC 3.1.3.2 denotes acid phosphatases, a subclass of hydrolases that catalyzes the hydrolysis of phosphate monoesters under acidic conditions, reflecting the dual functional potential of LcPAP4 (Additional file 1: Supplementary Table 9 and Supplementary Table 10). Moreover, the predicted sub-cellular localization of LcPAP4 was found to be either in extracellular matrix or cytoplasmic, in alignment with AtPAP17 localization [61]. Similarly, LcPAP9 formed a group with GmPAP29 and AtPAP15, both known for high phytase activity [22]. The predicted EC number (EC:3.1.3.26) of LcPAP9 might support the hypothesis that LcPAP9 could possess a dual phosphatase-phytase activity. Glycosylation is a vital post-translational modification that affects folding of proteins, its localization, trafficking, action and stability [9]. All the LcPAPs had at least one N-glycosylation sites indicating their targeted localization through Golgi apparatus as reported previously for GmPAPs in G. max [22]. The synteny relation among lentil, soyabean, Arabidopsis and rice described the evolutionary divergence among the dicot and monocot families depicting notable genomic variations in structures and composition over evolutionary time [57]. Promoter analysis of LcPAP genes revealed the presence of multiple cis-regulatory elements linked to Pi starvation and stress responses. Key TFs involved in Pi starvation responses include basic-leucine zipper (bZIP), basic helix-loop-helix (bHLH), myeloblastosis (MYB), WRKY, zinc-finger of A. thaliana 6 (ZAT6) and phosphate starvation response 1 (PHR1), as reported in A. thaliana, Nicotiana tabacum and O. sativa [62–66]. In A. thaliana a MYB TF, PHR1 regulates the expression of AtPT1 and AtACP5 (renamed AtPAP17), and its mutation reduces their expression, indicating its pivotal role in Pi signalling [61]. Similarly, bZIP homolog phi-2 gene in N. tabacum is induced under Pi starvation and modulates ABA signalling [62], while OsPTF1, a bHLH TF in rice, contributes to Pi stress adaptation [64]. WRKY75 and ZAT6, both zinc finger TF reported in A. thaliana, are implicated in Pi acquisition, homeostasis, root development, and ROS-mediated signalling [65, 66]. In the present study, multiple cis-regulatory motifs associated with stress-responsive TFs were identified in the promotor regions of LcPAPs. The G-box (CACGTG), a well-known cis element recognized by the bZIP family TF, plays a crucial role in mitigating various environmental stresses like, response to light and deficit of nutrients [67]. Notably, this motif also represents a variant of the E-box (CANNTG) typically targeted by the bHLH TFs, which are key regulators of plant growth and development and stress tolerance [68]. Moreover, W box (TTGACC/TTGACT), along with its variants WRE3 and WUN-motif, observed in the present study, are known binding sites for WRKY TFs. WRKY75, in particular, has been implicated in Pi starvation response and ROS signalling in A. thaliana [65], while WRE3 and WUN-are associated with several biotic stress responses [69]. P-box and MYB-recognition element (MRE), detected in the promoter region are the binding sites of MYB TF. PHR1, a key regulator of Pi starvation response and modulator of hormonal pathways in Arabidopsis [70], also plays a key role in response to abiotic stress in O. sativa [71]. Sp1, a zinc-finger binding site (C2H2), was identified in two LcPAPs promoters in the present study. The zinc finger TFs are integral to the floral development process, photomorphogenesis and pathogen responsiveness [72]. Notably, the P1BS motif with a consensus sequence of AAATATTAG was identified in the promoter regions of 13 LcPAPs genes (LcPAP1, 2, 3, 4, 5, 8, 11, 12, 13, 15, 17, 20 and 22), suggesting their regulation with P uptake and its remobilization consistent with the role of AtPAP17 [62]. Furthermore, the ERE and ABRE motifs (ABRE, ABRE3a, ABRE4), which mediate signalling of ethylene and abscisic acid in A. thaliana and S. lycopersicum, were also observed, indicating hormonal cross-talk during stress responses [69, 73]. In corroboration to the present findings associated to the interaction of P-containing ligands with LcPAPs, previous reports clearly states that the divalent metal site plays a very crucial role in anchoring the substrates containing phosphorus to the protein. The metal bond, which is a coordinate bond between the metal ions and phosphate-oxygen of the substrate like pNPP, formed in a non-bridging manner, to stabilize the ligand or substrate binding to the active site of proteins for further catalytic activities by the enzyme [74, 75]. In the present study either Asparagine or Histidine residue, or both of the residues belonging to motif 3 (GNH(D/E) or GNQE) were involved in interaction with pNPP in the case of LcPAP1, 2, 3, 4, 5, 9, 11, 15, 17 and 20; while one of the Histidine residues of motif 5 (GHXH) of LcPAP4 and LcPAP9 was found to be binding to pNPP. Furthermore, either Asparagine or Histidine residues of motif 3 (GNH(D/E) or GNQE) without or with the second Histidine residue of motif 5 (GHXH) belonging to LcPAP1, 2, 3, 4, 5, 9, 11, 15, 17 and 20 were also found to be involved in the interaction with the ligand phytic acid. Moreover, exclusively in case of LcPAP3 Tyrosine residue of motif 2 (GDXXY) was also found to be involved in binding with phytic acid. Previous study reported the binding of various pNPP analogues like phosphate, sulfate, tungstate and molybdate to the catalytic active centre of the red-kidney bean purple acid phosphatase via covalent bonds like electrostatic interactions and hydrogen bonds with the conserved amino acids like histidine, tyrosine and aspartic acid [76]. Faba-Rodriguez et al. 2022 [77] reported the first crystallized structure of myo-inositol hexakisphosphate (IP6) or phytate binding to the active site of purple acid phytases (PAPhy) in cereals. According to that report, the phosphate group of the phytate molecule formed coordinate bonds directly with the bi-metallic nuclear centre. Furthermore, the covalent bonds like hydrogen bonds, electrostatic bonds and hydrophobic interactions formed with the conserved amino acid residues played a crucial role in stabilizing the binding of phytate to PAPhy enzyme [77]. Similar instances have been observed in our study, as depicted by the binding of phytic acid to the LcPAPs through metal coordinate bonds (LcPAP2, 4, 5, 9, 11, 15, 17 and 20) and hydrogen bonding was the common covalent bond describing its pivotal role in the protein–ligand interaction stabilization. The present study also reports the binding of pNPP as well as phytic acid to the conserved amino acids like asparagine (Asn) apart from the other interacting amino acids (Figs. 10 and 11, Additional file 1: Supplementary Table 6). Though, Sträter et al. 1995 [78] stated the involvement of the metal coordinating Asn residue available within the conserved motif 3 of red kidney bean purple acid phosphatase in binding with sulfate, till now to the best of our knowledge, no identified PAPs were found to be interacted with the pNPP or phytic acid via the Asn residue of motif 3. The root structure architecture (RSA) of the genotypes BCL1001 and Moitree (WBL 77) exhibited a striking morphological difference when grown under different P regimes. BCL1001, an interspecific derivate of cultivated L. culinaris and wild species L. orientalis, was found to possess a characteristic feature of exuberant lateral roots under variable P regimes, as compared to Moitree with limited lateral root branching. Earlier literature highlighted the unique root traits of L. orientalis with better RSA and total root surface area [79], supporting the observed phenotype of BCL1001 and its suitability for functional characterization of LcPAP genes (Fig. 12). Tissue specific transcript expression profiling revealed that, LcPAP4, potentially encoding both phosphatase and oxidoreductase, was upregulated in shoot as well as root tissue under both P sufficient and deficient conditions, suggesting a constitutive role in P metabolism. LcPAP9, predicted to have both phosphatase and phytase activity, showed strong transcript level expression in shoot tissue under low P level, but was down-regulated in the root tissue. This pattern implies a shoot-specific role in mobilizing internal P reserves in the aerial plant parts, consistent with its ortholog AtPAP15 which were expressed specifically in the seed and pollen tissues in Arabidopsis [80]. It is to be recalled here that phylogenetic analysis performed in this study placed both LcPAP9 and AtPAP15 in clade IV, supporting this functional similarity. Notably, certain LcPAPs were upregulated even under high P conditions, implying additional roles in other biochemical pathways beyond P acquisition [81] (Fig. 13). LcPAP2, LcPAP3 and LcPAP11 were upregulated in shoot whereas downregulated in root tissues, indicating a likely involvement in intracellular phosphate mobilization rather than soil P acquisition. Conversely, LcPAP20 exhibited root-specific expression under P deficiency, suggesting a role in hydrolysing and mobilizing Pi from organic P reserves present in the soil, analogous to AtPAP10, which is known for its strong root expression under P-starvation [82]. Callose deposition is one of the markers for identifying plants growing under conditions having low phosphorus availability. Overexpression of cell-wall localized NtPAP12, isolated from N. tabacum demonstrated elevated activity of callose and cellulose synthase enzymes indicating the role of phosphatases in cell-wall biosynthesis [83]. Based on the protein–protein interaction study it could be hypothesized that LcPAP5, 9 and 15 might be associated with the process of callose deposition in lentils. Moreover, numerous probable LcPAP protein interactors were identified in the present study other than callose which were also found to be associated with phosphorus cycling as well as utilization. AtG3Pp1 and AtG3Pp2, which are functionally analogous to DL-glycerol-3-phosphatase were found to be involved in maintaining phosphorus homeostasis under conditions of low phosphorus availability [84]. Enzymes like FAD synthetase are known to play an important role in plant stress-responses, maintaining the functionality of the electron transport chain while; the enzymes like riboflavin kinase, FMN phosphatase, riboflavin synthase and FMN hydrolase involved in flavin metabolism involving ATP hydrolysis, might play a role in making Pi available to the plants for utilization [85, 86]. Furthermore, acid phosphatases have been reported to catalyse hydrolysis of FMN in spinach leaves [87]. Acid phosphatases extracted from leaves and nodules of Phaseolus vulgaris reportedly hydrolysed FMN when provided as substrate [88]. Hence it could be inferred that LcPAP4, 16, 17 and 18 might be associated with FMN hydrolysis in lentils to make Pi available to the plant upon requirement. Transient over-expression of nudix hydrolases encoding genes (NUDTs) from A. thaliana in Nicotiana benthamiana leaves revealed that nudix hydrolase accelerate the breakdown of PP-InsPx to InsP6 (Inositol-6-phosphate), the reduced concentrations of PP-InsP enabled the binding of the PHR1 TF to the P1BS binding site of a SPX promoter driving the sequential catalysis of a red-pigment betalanin [89]. It could be hypothesized that nudix hydrolases being associated with the phosphorus starvation responses might be interacting with LcPAP4, 16, 17 and 18 as revealed by the protein–protein interaction study (Fig. 14). Also, the interaction of LcPAP4, 16, 17 and 18 among themselves might indicate their collaborative involvement in the process of maintaining phosphorus homeostasis in lentils. Conclusion In conclusion, the present study offers valuable insights into the genome-wide identification and characterization of the 22 PAP genes present in lentil, advancing our understanding of their structural and functional diversity. Phylogenetic analysis grouped the identified LcPAPs into five clades reflecting evolutionary relationships and conserved motif variations among PAPs from multiple plant species. Expressional profiling under differential P regimes suggested that LcPAP1, 2, 3, 4, 5, 9, 11, 15, 17 and 20 might be involved in either P acquisition, internal translocation, and homeostasis, eventually contributing Phosphorus Use Efficiency (PUE) under nutrient limited conditions. These findings provide a valuable foundation for functional validation of the key PAP candidates through heterologous expression or by genome-editing of the targeted PAP genes and their promoters. The present study might pave the way of identifying key regulatory genes of lentil involved in P homeostasis for subsequent allele mining considering a large lentil population. This finding will aim in the development of suitable gene-based markers associated with P acquisition facilitating the development of nutrient-resilient lentil cultivars through advanced molecular breeding strategies, an essential step toward improving the productivity of this vital protein-rich crop in resource-constrained agroecosystems. Supplementary Information Additional file 1: Supplementary Table 1. List of primers used for observing the relative gene expression of the LcPAP genes. Supplementary Table 2. Amino acid motifs of 22 LcPAPs harboring the metal binding domains that are highlighted in bold. Supplementary Table 3. Physico-chemical properties of the 22 predicted LcPAPs including the transcript IDs, gene names, number of amino acids, number of exons and introns, weight of the proteins, pI, Instability Index, Aliphatic Index, Grand Average of Hydropathy (GRAVY) Index, sub-cellular localization, length and cleavage site of signal peptide and number of N-Glycosylation sites. Supplementary Table 4. Orthologous gene pairs forming syntenic blocks amongLens culinaris, Glycine max, Oryza sativa and Arabidopsis thaliana in respect to PAP genes. Supplementary Table 5. Number of the mentioned transcription factor binding sites present in the upstream 2000 bp promoter region from the translation start site of LcPAP genes. Supplementary Table 6. Details of docking interactions between LcPAPs and ligands pNPP and phytic acid. Supplementary Table 7. Details on protein-protein interaction study. Supplementary Table 8. Representation of the 22 LcPAPs transcript IDs, gene names, chromosome IDs, start and end sites of the genes and the strand on which the genes are present. Supplementary Table 9. The EC numbers of the respective LcPAP proteins obtained with TBtools v.2.0 software (https://github.com/CJ-Chen/TBtools/) through the analysis of the GFF and annotation files of lentils obtained from phytozome database. Supplementary Table 10. Details of the EC numbers according to the KEGG database. Additional file 1: Supplementary Table 1. List of primers used for observing the relative gene expression of the LcPAP genes. Supplementary Table 2. Amino acid motifs of 22 LcPAPs harboring the metal binding domains that are highlighted in bold. Supplementary Table 3. Physico-chemical properties of the 22 predicted LcPAPs including the transcript IDs, gene names, number of amino acids, number of exons and introns, weight of the proteins, pI, Instability Index, Aliphatic Index, Grand Average of Hydropathy (GRAVY) Index, sub-cellular localization, length and cleavage site of signal peptide and number of N-Glycosylation sites. Supplementary Table 4. Orthologous gene pairs forming syntenic blocks amongLens culinaris, Glycine max, Oryza sativa and Arabidopsis thaliana in respect to PAP genes. Supplementary Table 5. Number of the mentioned transcription factor binding sites present in the upstream 2000 bp promoter region from the translation start site of LcPAP genes. Supplementary Table 6. Details of docking interactions between LcPAPs and ligands pNPP and phytic acid. Supplementary Table 7. Details on protein-protein interaction study. Supplementary Table 8. Representation of the 22 LcPAPs transcript IDs, gene names, chromosome IDs, start and end sites of the genes and the strand on which the genes are present. Supplementary Table 9. The EC numbers of the respective LcPAP proteins obtained with TBtools v.2.0 software (https://github.com/CJ-Chen/TBtools/) through the analysis of the GFF and annotation files of lentils obtained from phytozome database. Supplementary Table 10. Details of the EC numbers according to the KEGG database.