Genome-wide identification, characterization and expression analysis of the subtilisin-like protease gene family in Quercus ilex in response to Phytophthora cinnamomi Authors: Daniela M Hernández, María-Dolores Rey, Mónica Labella-Ortega, María-Ángeles Castillejo, Jesús V Jorrín-Novo, Ana Mª Maldonado-Alconada Published: 2026 ABSTRACT AbstractSubtilisin-like proteases (SBTs) are a large family of serine peptidases involved in diverse biological processses, including plant stress responses, but remain poorly characterized in forest trees. In this study, we report the genome-wide identification and characterization of the SBT gene family in Quercus ilex (QiSBT), with a focus on their transcriptional organization and response to Phytophthora cinnamomi infection. A total of 87 QiSBT genes were identified, structurally annotated, and phylogenetically classified into seven subfamilies based on comparative analyses with Arabidopsis thaliana, Populus trichocarpa, and three other Quercus species (Q. lobata, Q. robur, and Q. variabilis). Analyses of conserved domains and gene structures revealed a high degree evolutionary conservation, with one subfamily enriched in intronless or intron-poor genes. Promoter analysis identified numerous stress- and hormone-responsive cis-elements, suggesting potential transcriptional regulation under environmental stress conditions. Expression profiling under P. cinnamomi infection revealed differential transcriptional responses within the QiSBT gene family. In particular, QiSBT24, QiSBT63, and QiSBT69 showed significant induction in inoculated seedlings, with approximate 3-, 8-, and 30-fold, respectively. These results provide new insights into the structure, evolution, and potential defense-related roles of SBTs in Q. ilex, offering a foundation for future functional aimed at understanding their contribution to resistance mechanisms in Mediterranean oak species.Supplementary InformationThe online version contains supplementary material available at 10.1038/s41598-026-56449-x.Keywords: Holm oak, Quercus spp., Subtilisin-like proteins, Phytophthora cinnamomi, Gene expressionSubject terms: Computational biology and bioinformatics, Evolution, Genetics, Molecular biology, Plant sciences Proteases play a pivotal role in regulating growth, development, defense mechanisms, and stress responses, enabling organisms to adapt to changing environmental conditions1–3. Among them, serine proteases, defined by the presence of a serine residue in their active site, are involved in a wide array of biological processes, including protein processing and degradation, signal transduction, and immune responses4–6. Serine proteases are broadly classified into three main groups: subtilisin-like proteases (subtilases), Clp proteases (caseinolytic proteases), and kexin-type proteases7–9. The subtilisin-like serine proteases (SBTs) form one of the most extensively studied protease families in plants due to their functional versatility across diverse cellular processes10,11. SBTs are defined by a conserved catalytic triad, comprising serine (Ser), histidine (His) and aspartate (Asp), and a characteristic catalytic domain belonging to the peptidases_S8_S53 superfamily. This domain can exist in several variants, including Peptidases_S8_Tripeptidyl_Aminopeptidase_II, Peptidases_S8_SKI-1_like, or Peptidases_S8_312,13. In addition, SBTs contain other conserved features such as the inhibitor_I9 domain and the protease-associated (PA)_subtilisin_like domain2,5,14. The inhibitor_I9 domain, also known as the prodomain, has an auto-inhibitory function, keeping the enzyme in an inactive zymogen state and preventing premature substrate access to the active site5. The PA_subtilisin_like domain acts as a protein-protein interaction module and contributes to substrate specificity2,14. Most plant SBTs also contain a signal peptide that targets them to the secretory pathway, although a few have been predicted to localize to mitochondria or chloroplast15,16. In recent years, an increasing number of studies have demonstrated that SBTs play a crucial role in the immune response during plant–pathogen interactions2,17,18. Ramírez et al.19 identified an extracellular SBT in Arabidopsis thaliana that enhanced innate immunity, while Figueiredo et al.18 found that several SBTs were associated with plant-pathogen resistance. In Solanum lycopersicum, a subgroup of SBTs, known as P69 (P69A-F), initially annotated as a single pathogenesis-related protein (PR7)12, was documented for its role in plant defense. Among these, P69B and P69C were transcriptionally induced upon Phytophthora infestans infection20. P69C triggered a signaling loop that alleviated epigenetic silencing of defense-related genes, effectively priming the plant for a more robust response to future pathogen attacks12. Meanwhile, P69B was identified as a target for two Kazal-Like protease inhibitors from P. infestans21,22. Additionally, Paulus et al.23 reported that P69B and other SBTs activated the extracellular immune protease Rcr3 in S. lycopersicum, initiating a proteolytic cascade essential for the immune response. In Phlomis purpurea, SBT gene expression was also upregulated in response to Phytophthora cinnamomi Rands.24. Likewise, Backer et al.25 identified numerous apoplastic proteases, including two SBTs (Peame105C009g008400 and Peame105C08g006250), that were more abundant in Persea americana following P. cinnamomi inoculation. The holm oak (Quercus ilex L. subsp. ballota (Desf.) Samp.) is the most emblematic species of the Mediterranean Basin and the agrosilvopastoral ecosystem dehesa26, with substantial ecological and economic importance. Over the past decade, there has been a concerning increase in mortality among Q. ilex populations, largely attributed to infection by the soil oomycete P. cinnamomi27. This alarming trend highlights the urgent need for sustainable strategies to safeguard the species and the ecosystems it supports. Among the most promising biotechnological approaches is the exploitation of natural biodiversity to identify and select elite genotypes with enhanced resistance traits, assisted by molecular markers for subsequent clonal propagation28. Characterization of existing biodiversity through the Central Dogma of Molecular Biology framework and the omics cascade can provide key insights into phenotypic diversity in traits such as growth, productivity, and stress tolerance, establishing a robust foundation for the selection of elite or plus genotypes for use in reforestation, conservation, and breeding programs. Recent integrative omics studies, including transcriptomics, proteomics and metabolomics, have identified candidate molecular markers linked to stress responses, with proteases emerging as consistently relevant elements in response to pathogen infection29–32. Notably, Escandón et al.33 highlighted the involvement of proteases and protease inhibitors in Q. ilex seeds during germination, although their specific role in the defense response remains unknown. To date, a systematic study of the SBT gene family in Q. ilex has not been conducted. Given the well-established roles of SBTs in plant defense, including their involvement in immune signaling and proteolytic cascades in response to P. cinnamomi, a comprehensive analysis is essential to uncover their functional significance and explore their potential for enhancing resilience in Q. ilex. The recent release of the Q. ilex draft genome provides a valuable resource for exploring and characterizing this gene family34. Here, we present the first genome-wide identification, structural characterization and expression analysis of the SBT gene family in Q. ilex (QiSBT). This study includes a comparative phylogenetic analysis involving Q. ilex, Arabidopsis thaliana (L.) Heynh., Populus trichocarpa Torr. & A.Gray ex. Hook., and three other Quercus species (Q. lobata Née, Q. robur L. and Q. variabilis Blume.). We further examine chromosomal distribution, gene duplication events, prediction of cis-acting regulatory elements, and miRNA target sites. Finally, we present expression profiling of QiSBT genes in response to P. cinnamomi, offering valuable insights into their potential roles in stress adaptation and defense. Materials and methods Identification of SBT gene family members and sequence analysis The protein sequences of the subtilisin-like protease gene family in Q. ilex (hereafter QiSBT) were retrieved from the National Center for Biotechnology Information database (NCBI, https://www.ncbi.nlm.nih.gov/datasets/genome/GCA_032727855.1/, accessed on 28 November 2023)34. Initially, a BlastP search was performed against the Q. ilex genome using 55 A. thaliana subtilisin-like proteases (AtSBT) protein sequences (excluding the pseudogene AtSBT4.10), obtained from the TAIR database (https://www.arabidopsis.org/) as queries. Hits were filtered using an E-value threshold of ≤ 1e− 10. In parallel, a Hidden Markov Models (HMM)-based search was also conducted using the HMMER package (http://hmmer.org/). Two HMM profiles were used: (i) the Subtilase family domain profile (PF00082) from the Pfam database35, and (ii) a custom profile built from the alignment of the 55 curated AtSBT protein sequences. HMMER hits with an E-value ≤ 1e − 3 were retained. Candidate sequences were manually validated for the presence of the conserved S8 peptidase domain using InterProScan36(PF00082; IPR000209) and NCBI Conserved Domain Database (CDD) CD-search37,38 including Peptidases_S8_3 (cd04852), Peptidases_S8_SKI-1_like (cd07479), and Peptidases_S8_Tripeptidyl_Aminopeptidase_II (cd04857), all belonging to the Peptidase S8_S53 superfamily (cl10459). Proteins lacking SBT-related domains were excluded. Physicochemical properties of QiSBT proteins were estimated using ProtParam (ExPASy Server)39(accessed on 02 October 2024). Subcellular localization and signal peptide prediction were performed using SignalP and DeepLoc tools40,41 (accessed on 24 April 2023). Exon-intron structures were obtained from the Q. ilex genome annotation34. Conserved motifs were identified MEME42 with the following parameters: classic mode, zero or one occurrence per sequence, 20 motifs and motif widths of 6–50 amino acids. Gene structure, motif, and domain data (filtered with a E-value ≤ 1e− 5) visualized using TBtools v2.043. Phylogenetic analysis To investigate the phylogenetic relationships of SBT proteins, amino acid sequences from Q. ilex and A. thaliana were aligned using MUSCLE v544. Additionally, a multispecies alignment was performed including SBT sequences from Q. robur (PM1N assembly, https://www.oakgenome.fr), Q. lobata (ValleyOak3.0 assembly, https://valleyoak.ucla.edu), Q. variabilis v2.0 (Haplotype A assembly)45, and P. trichocarpa v4.1 Phylogenetic trees were constructed using the maximum-likelihood (ML) method with the best-fit model and 1,000 bootstrap replicates in IQ-TREE46. Subfamilies were classified according to Rautengarten et al.47, with modifications based on Taylor and Qiu48. Chromosome distribution and gene duplication analysis QiSBT genomic coordinates, including start position, chromosomal localization, and gene length, were retrieved from the Q. ilex GFF3 annotation file34. Genes were mapped using TBtools v2.043 and named according to chromosomal order. For synteny analysis, genome assemblies and annotations of A. thaliana (TAIR10) and Q. lobata (ValleyOak3.0) were downloaded from TAIR database and EnsemblPlants (https://plants.ensembl.org/), respectively. Inter-species syntenic relationships were analyzed using MCScanX (One-Step module) with default parameters and visualized using TBtools Dual Synteny Plot43. Selective pressure was assessed using Ka/Ks ratios calculated with the Simple Ka/Ks Calculator (NG) module in TBtools43. A Ka/Ks ratio > 1 indicates positive selection, < 1 purifying selection, and = 1 neutral evolution. Cis-acting regulatory element analysis of QiSBT genes For each QiSBT gene, 1.5 kb upstream sequences from the start codon (ATG) were extracted from the Q. ilex genome34. Cis-acting regulatory elements were identified using PlantCARE49(accessed on 05 May 2024). miRNA prediction in the QiSBT gene family miRNA target prediction for the 87 QiSBT genes was performed using the Plant Small RNA Target Analysis Server v2 (psRNATarget) (http://plantgrn.noble.org/psRNATarget/), with default parameters and A. thaliana miRNA sequences50(accessed on 04 October 2024). Due to the absence of a Q. ilex protein database in the psRNATarget server, A. thaliana was used as a reference for prediction. Interactive protein analysis of QiSBT genes The protein interactive network analysis for QiSBTs were predicted querying the STRING database version 12 (https://string-db.org/) against the translated genome of Q. ilex, previously indexed to the served with the “Add any organism to STRING” (accessed on 02 February 2025). Protein interaction networks with a confidence score ≥ 0.4 were considered. The predicted networks integrate multiple lines of evidence, including phylogenetic co-occurrence, gene co-expression, curated database annotations, and/or experimentally validated interactions. Inoculation of Q. ilex seedlings with P. cinnamomi The inoculation procedure followed the protocol previously described in Hernández-Lao et al.32. Briefly, roots of six-month-old Q. ilex seedlings were immersed in carrot–agar liquid inoculum containing P. cinnamomi (P90) at a concentration of 43 chlamydospores µL− 1. Non-inoculated control seedlings were treated with pathogen-free medium. Sampling was performed on day 12, when inoculated seedlings showed a 35% reduction in leaf fluorescence measured using a FluorPen FP100 portable leaf fluorimeter (Photon Systems Instruments, Drasiv, Czech Republic). Infection was confirmed on day 16 by pathogen re-isolation and microscopic identification. RNA Extraction, RNA-Seq, and RT-qPCR Q. ilex individuals located at 37° 56’ 46” N, 6° 11’ 33” W (Santa Olalla del Cala, Huelva, Spain) were used for expression analysis. Individuals were identified by the Andalusian Network for Damage Monitoring in Forest Ecosystems (RED SEDA; Junta de Andalucía, 2016). Collection compiled with all relevant institutional, national, and international guidelines and legislation. Voucher specimen was deposited at the Global Biodiversity Information Facility of University of Cordoba, Spain (deposition number. 36840-1)51. Total RNA was extracted from 50 mg of fresh Q. ilex seedling leaves (6-month-old seedlings grown under pot conditions), previously macerated in liquid nitrogen, using the InviTrap® Spin Plant RNA Mini kit (Stratec Molecular GmbH, Berlin, Germany). The extracted RNA was treated with DNAase I (ThermoFisher Scientific, Waltham, USA) to eliminate potential genomic DNA contamination. RNA quality was assessed on 1% agarose gel electrophoresis, while both RNA quantity and integrity were evaluated by electrophoresis using the Agilent 2100 Bioanalyser kit (ThermoFisher Scientific, Waltham, USA). All samples showed RIN (RNA Integrity Number) values ≥ 7. RNA sequencing (RNA-seq) was performed to compare gene expression between P. cinnamomi-inoculated seedlings and non-inoculated controls, using the NovaSeq PE150 platform (Illumina, Inc., San Diego, CA, USA). Subsequently, quality assessment of the raw sequencing data was performed using FastQC v.0.11.952. Trim-Galore version 1 was used to remove residual adapters, low quality portions and short reads53. Filtered reads were mapped to the Q. ilex reference genome34 using HiSAT2 (Hierarchical indexing for Spliced Alignment of Transcripts 2) version 2.2.1.54. The StringTie transcript assembler55 was then used to assemble RNA-seq alignments into potential transcripts and gene expression levels were estimated with RSEM (Expectation–Maximization method56. Differential gene expression analysis between inoculated and non-inoculated seedlings was performed in R using the edgeR package. Raw count data were normalized using the Trimmed Mean of M-values (TMM)57 method. Genes with a p-value < 0.05 were considered differentially expressed. A heatmap of differentially expressed genes was generated in R using pheatmap package version 4.2.0. Complementary DNA (cDNA) from 1 µg RNA was synthesized by a Reliance Select cDNA Synthesis Kit (Bio-Rad, California, United States) according to the manufacturer’s instructions and used in the RT-qPCR reaction in an iCycler system using SYBR-Green (Bio-Rad, California, USA). Pairs of primers were designed using the Primer3-Plus web-based program (https://primer3plus.com). QiSBT24, QiSBT63, and QiSBT69 were amplified using primers specifically designed for this study: Forward: 5’ ATGTCACCAGTGCCACAAAA 3’ and Reverse: 5’ TACCTGCAGCTTGGAGTCCT 3’; Forward: 5’ CAACATTCTCAAGCCCGATT 3’ and Reverse: 5’ GAACCATGAGGGCAAGACAT 3’; Forward: 5’ AAGAAAGAGATTGCCGCCGA 3’ and Reverse: 5’ TCCCGTCCGATGTAACTTGC 3’, respectively. Each 10 µL reaction mixture contained 2× SsoAdvanced Universal SYBR® Green Supermix (Bio-Rad, CA, USA), 0.2 µM of each forward and reverse primer, 50 ng of cDNA template, and nuclease-free water. A two-step amplification protocol was employed, beginning with an initial denaturation step at 95 °C for 30 s, followed by 45 cycles of denaturation at 95 °C for 15 s and annealing/polymerization at 60 °C for 60 s. Subsequently, a melting analysis ranging from 60 °C to 97 °C at a rate of 0.1 °C/s was conducted to validate the specificity of the target amplicon. Gene expression levels were determined based on the threshold cycle (Ct value) and relative differential expression were determined using the 2-ΔΔCt method58. Elongation factor 1-alpha (EF-1α) and Histone 3 (H3), which showed stable expression under the experimental conditions tested, were used as internal reference genes30. Each reaction was performed in duplicate. Comparison of co-expression and protein–protein interaction networks Gene co-expression analysis was performed using normalized logCPM expression values derived from RNA-seq data. Pearson correlation coefficients among QiSBT family members were calculated in R using the Hmisc package. Correlation matrices and associated p-values were generated in R using the rcorr function, and significant associations were defined using thresholds of |r| ≥ 0.7 and p-value ≤ 0.05. Adjacency matrices were constructed from significant correlations and used to generate undirected co-expression networks with the igraph package in R. Network modules were identified using the Louvain community detection algorithm59. Network topology parameters, including degree, betweenness, and weighted strength centrality, were calculated to identify highly connected hub genes within the co-expression network. Heatmaps of Pearson correlation coefficients were generated in R using the pheatmap package. Significant correlations (p < 0.05) were visualized, and expression values were standardized using row-wise Z-score transformation prior to clustering. A target-based co-expression analysis was performed using QiSBT69 as a query gene. Pearson correlation coefficients between QiSBT69 and all other QiSBT genes were computed across samples. Genes showing strong correlation (|r| ≥ 0.99, p < 0.01) were defined as the core co-expression module associated with QiSBT69. Functional annotation of the QiSBT69 module was performed based on gene-to-Gene Ontology (GO) and gene-to- Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway mappings using the available Q. ilex genome annotation34. Protein–protein interaction data were retrieved from the STRING database and filtered to retain interactions among QiSBT genes. Co-expression networks and STRING-derived interaction networks were compared by identifying shared interaction pairs. Statistical significance of overlap was assessed using Fisher’s exact test. Statistical analysis Statistical analysis of gene expression data was performed using the STATISTIX v.10.0 software (Analytical Software, FL, USA). Gene expression data were subjected to Student’s t-test with p < 0.05. Data not fulfilling the assumptions of a parametric test were subjected to the Kruskal-Wallis´s test (p < 0.05 and means separated by the Dunn test at p < 0.05). Results Identification and characterization of QiSBT genes A total of 87 QiSBT genes containing the conserved S8 peptidase domain were identified in the Q. ilex genome and named QiSBT-1 to QiSBT-87 according to their chromosomal localization (Supplementary Table 1). Detailed information for each QiSBT gene, including gene name, gene ID, subfamily, coding DNA sequence (CDS), amino acids (AA) length, protein molecular weight (MW) and isoelectric point (IP), instability index (II), aliphatic index (AI), grand average of hydropathy (GRAVY), signal peptide (SP), and subcellular localization (SL), is provided in Supplementary Table 1. The QiSBT CDS lengths ranged from 381 to 4188 bp, corresponding to proteins of 126 and 1395 amino acids. The predicted MW of the proteins ranged from 13,385.48 to 154,165.98 Da, with pI values between 4.97 and 9.87. The II varied from 20.32 to 43.85, the AI ranged from 75.08 to 103.05 and GRAVY values varied from − 0.332 to 0.228. Out of 87 identified QiSBT proteins, 47 (54%) were predicted to contain signal peptides. SL predictions indicated that most QiSBT proteins were localized in the extracellular or apoplastic space, lysosomes, and vacuoles (Supplementary Table 1). However, a few proteins, such as QiSBT28, QiSBT37, QiSBT38, QiSBT50, QiSBT69, and QiSBT80, were predicted to localize in the cytoplasm, plastids, endoplasmic reticulum, or nucleus. Phylogenetic analysis of SBT gene family A phylogenetic tree was constructed using the 87 QiSBT protein sequences to investigate their evolutionary relationships. Arabidopsis thaliana sequences (AtSBT) were used as a reference for QiSTB classification. Based on sequence homology, QiSBTs were clustered into seven groups (SBT1-SBT7), consistent with the classification of the 55 AtSBT proteins (Fig. 1a). Among these, the SBT4 subfamily was the largest, comprising 42 QiSBT and 14 AtSBTs, followed by SBT1 (19 QiSBTs and 9 AtSBTs) and SBT3 (8 QiSBTs and 18 AtSBTs). Two QiSBT proteins, QiSBT2 and QiSBT46, formed a poorly supported clade that could not be confidently assigned to any established group and were provisionally categorized as a separate group (SBT0). This assignment should be interpreted with caution, as the phylogenetic diversity analyzed is restricted to eudicot species, which may not fully capture the broader evolutionary diversity of the SBT family, including potential lineage-specific gene losses in A. thaliana. Phylogenetic relationships of SBT proteins from Q. ilex (QiSBT) and A. thaliana (AtSBT) (a), and from Q. ilex (QiSBT), A. thaliana (AtSBT), P. trichocarpa (PtSBT), Q. robur (QrSBT), Q. lobata (QlSBT), and Q. variabilis (QvSBT) (b). The phylogenetic tree was constructed using the maximum-likelihood method. SBT proteins are grouped into seven clades, each represented by a different color. To further explore the evolutionary relationships of the SBT family, an expanded phylogenetic analysis was conducted including 357 additional SBT protein sequences from Q. lobata (109 sequences, QlSBT), Q. variabilis (93, QvSBT), Q. robur (95, QrSBT), and P. trichocarpa (60, PtSBT), together with QiSBT and AtSBT sequences (Fig. 1b). The resulting phylogenetic tree was divided into seven subfamilies (SBT1-SBT7), with SBT4 again emerging as the largest group, encompassing 200 members, including 14 AtSBTs, 7 PtSBTs, 44 QrSBTs, 47 QlSBTs, and 46 QvSBTs. Interestingly, four QiSBTs were found to be unique to Q. ilex, while 30 were exclusive to the Quercus genus. Despite their initial ambiguous placement, QiSBT2 and QiSBT46 were found to have orthologues in both Quercus species and P. trichocarpa. Chromosomal localization, collinearity analysis and selective pressure analysis of QiSBT Based on the genomic localization data from the Q. ilex genome, the 87 QiSBT genes were unevenly distributed across the 12 chromosomes of Q. ilex (Fig. 2). Chromosomes 1 and 8 harbored the highest number of QiSBT genes. Notably, many of these genes were organized in clusters rather than being randomly or evenly distributed. Chromosomal distribution of QiSBT genes. Labels correspond to the 87 identified QiSBT genes. (a) Genes are color-coded based on their subfamily classification (SBT1–SBT7), with ungrouped genes labeled as SBT0. Tandem duplicate genes are indicated by red lines. To explore evolutionary relationships and conservation, a comparative syntenic analysis was performed between A. thaliana, Q. ilex and Q. lobata (Fig. 3). The analysis revealed 13 homologous gene pairs between Q. ilex and A. thaliana, and 42 between Q. ilex and Q. lobata. Selective pressure on orthologous gene pairs was assessed using Ka, Ks and Ka/Ks ratios (Supplementary Table 2). The Ka/Ks ratios for most QiSBT gene pairs were < 1, indicating purifying selection. However, two gene pairs, QlSBT30/QiSBT31 and Ql12p016678/QiSBT80, showed Ka/Ks > 1, suggesting positive selections. Interestingly, Ql12p016678 is a previously predicted SBT lacking the conserved Peptidase S8/S53 domain PF00082 (IPR000209) and the corresponding superfamily annotation (cl10459). Synteny and interspecific collinearity analysis of the SBT gene family in A. thaliana, Q. ilex and Q. lobata. Gray lines in the background indicate collinear blocks between Q. ilex and other plant genomes, while the red lines highlight the syntenic SBT genes pairs. Gene structure, conserved motifs, and structural domains of QiSBTs The gene structure, conserved motifs, and structural domains of QiSBTs are summarized in Fig. 4. The number of exons ranged from 1 to 20. QiSBT69 (SBT6 subfamily) showed the most complex structure with 20 exons, whereas 16 genes contained one single exon. The SBT1 subfamily displayed the simplest gene architecture, with 15 out of 19 genes being intronless. All QiSBT proteins contained the conserved Peptidases_S8_S53 superfamily domain (cl10459). Additionally, the PA_subtilisin_like domain (cd02120) was found in 77 QiSBTs, the fn3_6 superfamily domain (cl05663) in 77, and the Inhibitor_I9 superfamily domain (cl05470) in 73 proteins. A smaller number (6 proteins) contained the AprE superfamily domain, while the TPPII superfamily domain (cl13955) was uniquely found in QiSBT69. Conserved motif analysis identified 20 distinct motifs across QiSBT sequences. Motifs, 1, 2, 5, 6, 7, 8, 9, 10, and 15 corresponded to Peptidases_S8_S53 superfamily domain (cl10459), while motifs 14 and 20 were associated with the PA_subtilisin_like domain. Motifs 12, 17 and 18 matched the fn3_6 domain, and motif 11 corresponded to the Inhibitor_I9 domain. The remaining motifs did not match known functional domains. Conserved motifs, structural domains, and gene structure of QiSBTs. Conserved motif composition is shown, which different motifs displayed in colored boxes (a). Conserved domain composition is shown, with different colors representing different domains (b). Exon–intron structure is shown, where yellow boxes represent coding sequences (CDS), blue boxes represent untranslated regions (UTRs), and black lines represent introns (c). Cis-acting element analysis of QiSBT gene promoters Cis-acting regulatory elements identified in promoters were classified into five categories: promoter-related (6361), development-related (246), environmental stress-responsive (1139), hormone-responsive (750), and light-responsive elements (750) (Fig. 5a). Within the environmental-stress responsive category (Fig. 5b), defense- and stress-related elements were the most abundant (732), followed by anaerobic-responsive (122), wound- and pathogen-responsive (103), and fungal-responsive elements (66). Among the hormone-responsive elements, abscisic acid-responsive elements were the most abundant (223), followed methyl jasmonate (96) and salicylic acid-responsive elements (87) (Fig. 5c). A comprehensive count matrix detailing the distribution of each cis-element subcategory across individual QiSBT genes is provided in Supplementary Table 3. Cis-acting regulatory elements in the promoter regions of QiSBT genes. Classification and proportion of cis-acting elements (a). Cis-elements related to environmental stress responses (b) and hormone-responsiveness (c). miRNA target prediction in the QiSBT gene family The miRNA target prediction identified 10 QiSBT genes as potential targets of miRNAs (expectation ≤ 3.0) (Supplementary Table 4). QiSBT14 was the most frequently targeted gene, interacting with seven miRNAs. QiSBT41 was targeted by two miRNAs, while the remaining genes each showed a single predicted interaction. The most frequent miRNA was ath-miR166, primarily targeting QiSBT14. Interaction network analysis of QiSBT genes The interaction network analysis revealed potential functional associations between QiSBTs proteins and several Q. ilex proteins (Fig. 6). Predicted interactors included a subtilisin-like protease, SBT1.8 (lacking the Peptidase S8/S53 domain), as well as proteins involved in signaling and regulation, such as SCARECROW, SCARECROW-like protein 18, MADS-box transcription factor 17, pectinesterase inhibitor 9, RPM1-interacting protein 4, and nucleotide exchange factors 1 and 2. Predicted protein–protein interaction network of QiSBTs. Color of nodes represents the QiSBT subfamily. Color lines between nodes refer to the evidence of interaction; data from curated databases in cyan, experimentally determined interactions in magenta, phylogenetic co-occurrence in dark blue, co-expression in black, and protein homology in light purple. Expression patterns of QiSBT genes in response to P. cinnamomi RNA-seq analysis revealed the expression dynamics of QiSBTs genes in response to P. cinnamomi infection. Among the 87 QiSBT genes, 45 were undetectable, 15 showed very low expression (TMM < 1) and 27 genes were considered actively expressed under the experimental conditions (Fig. 7; Supplementary Table 5). Most expressed QiSBT genes showed low or constitutive expression in both inoculated and non-inoculated seedlings. However, several genes exhibited clear differential expression patterns. QiSBT22, QiSBT38, and QiSBT69 showed relatively high expression under both conditions, with QiSBT69 significantly induced upon inoculation. In contrast, QiSBT24 and QiSBT63 were weakly expressed in non-inoculated seedlings but were significantly upregulated after infection. Heatmap showing the expression profiles of actively expressed QiSBT genes in leaves of non-inoculated and P. cinnamomi-inoculated Q. ilex seedlings based on RNA-seq data (p-value < 0.05). Color intensity represents normalized transcript abundance (TMM values), with darker blue indicating higher expression levels. RNA-seq validation confirmed the transcriptomic results for three selected genes (QiSBT24, QiSBT63, and QiSBT69), showing approximately 3-, 8-, and 30-fold induction, respectively, in inoculated seedlings (Fig. 8). RT-qPCR validation of upregulated QiSBT genes in leaves of 6-month-old Q. ilex seedlings in response to P. cinnamomi. Bars represent mean expression values, and error bars indicate standard error. Different letters denote statistically significant differences between treatments (p < 0.05). QiSBT24 (a), QiSBT63 (b), and QiSBT69 (c). Co-expression network analysis of QiSBT genes with protein–protein interaction evidence A heatmap was constructed to visualize significant co-expression relationships among QiSBT genes (p < 0.05; Supplementary Fig. 1), revealing distinct expression patterns within the gene family. Several well-defined clusters of positively correlated genes (red) were observed, together with a smaller number of negatively correlated relationships (blue), reflecting structured co-expression organization among QiSBT members. The QiSBT co-expression network consisted of 35 genes (nodes) connected by 81 significant interactions (|r| ≥ 0.7, p ≤ 0.05), forming a moderately to highly connected network. Louvain community detection identified 12 co-expression modules of variable size, indicating a modular organization of the network (Supplementary Fig. 2). Network topology analysis revealed a subset of highly connected hub genes based on degree and weighted strength centrality, including QiSBT21 and QiSBT39, which exhibited high centrality within the network. QiSBT20 emerged as a shared hub supported by both RNA-seq-derived co-expression and STRING protein–protein interaction data, suggesting a central role within the QiSBT stress-responsive network (Supplementary Fig. 3). No overlapping edges were detected between the co-expression and STRING-derived protein–protein interaction networks, and Fisher’s exact test did not indicate significant enrichment between both datasets. QiSBT69 core co-expression module and functional annotation The QiSBT69 core co-expression module comprised a set of genes showing very strong positive transcriptional correlations with the target gene (|r| ≥ 0.99, p < 0.01; Supplementary Table 6). Genes within the module included proteins annotated with functions related to signaling processes, transcriptional regulation, protein modification, and intracellular metabolism. Several genes including receptor-like kinases, transcription factors, and components associated with ubiquitin- and SUMO-mediated protein regulation were present. In addition, genes associated with chloroplast and mitochondrial functions, RNA processing, and proteolytic activity were identified among the co-expressed set. GO annotation of the QiSBT69 module revealed a structured distribution of functional categories (Supplementary Fig. 4a). The most represented GO terms corresponded to cellular localization categories, including integral component of membrane, plasma membrane, nucleus, and general membrane-associated components. Functional binding-related categories were also prominent, including protein binding, ATP binding, and metal ion binding, together with nucleic acid binding. Transcription-related GO terms were also represented, including DNA-binding transcription factor activity and regulation of transcription, DNA-templated. At the biological process level, lignin biosynthetic process and transmembrane transport were detected. At KEGG level, the most represented pathway was phenylpropanoid biosynthesis (Supplementary Fig. 4b). Discussion Due to the increasing mortality of Q. ilex individuals, P. cinnamomi has emerged as one of the most critical threats to the survival of this key Mediterranean forest species27,31,60,61. As a result, the identification of resilient individuals, through molecular markers, is required for breeding and conservation programs. SBTs are known to play diverse roles in plant defense, particularly against Phytophthora species21,22, including P. cinnamomi24,25. Therefore, a comprehensive analysis of the SBT gene family may facilitate the discovery of candidate molecular markers involved in pathogen resistance. SBTs are considerably more prevalent in plants than in other organisms12, with expanded gene families reported across multiple species, including A. thaliana (56), Oryza sativa (63), Solanum tuberosum (74), Vitis vinifera (82), Glycine max (118), and Triticum aestivum (255)15,16,47,61–64. Despite this expansion, SBTs remain poorly characterized in long-lived forest species, where they may play key roles in stress adaptation and developmental plasticity. The identification of 87 SBT genes provides new insight into the complexity of this protease family in Q. ilex. The structural and physicochemical properties of QiSBT proteins, ranging in size from 126 to 1395 amino acids, with broad isoelectric points and generally stable hydrophilic profiles, are consistent with functional diversity observed in other plant SBT families and support their potential involvement in multiple cellular processes64,65. More than half of the QiSBT proteins are predicted to contain signal peptides, and most are targeted to the extracellular space or endomembrane-associated compartments. This is particularly relevant given that apoplastic proteases play central roles in plant defense signaling and host–pathogen interactions. In this regard, the predicted extracellular localization of a large fraction of QiSBTs suggests that this gene family may contribute to defense responses in Q. ilex, particularly against soil-borne pathogens such as P. cinnamomi. These findings support the hypothesis that SBT expansion in Q. ilex is not merely a genomic feature but may reflect functional adaptation linked to environmental stress resilience, warranting further functional characterization of specific members involved in biotic stress responses66,67. In addition to QiSBT genes, we identified 60 SBT genes in P. trichocarpa, 109 in Q. lobata, 93 in Q. variabilis, and 95 in Q. robur. Notably, four SBT genes were unique to Q. ilex, and 30 were exclusive to the Quercus genus. These species- and genus-specific SBTs may serve as valuable molecular markers for the identification of natural Quercus hybrids, a common occurrence for which currently few reliable diagnostic tools currently exist. The SBT gene family in all analyzed species was classified into seven subgroups (SBT1-SBT7), consistent with previous reports62,63. The pattern of gene copy expansion observed within SBTs may reflect the evolutionary significance of this family in plants, particularly in processes such as immunity48, programmed cell death68 and responses to abiotic and biotic stress62,63. Comparison of SBT subfamily distribution revealed a conserved pattern in Q. ilex and A. thaliana, where SBT6 and SBT7 were each represented by a single gene (QiSBT69 and QiSBT21 in Q. ilex, and AtTPP2 and AtS1P, in A. thaliana). This distribution is consistent with observations in other species such as O. sativa, S. tuberosum, and V. vinifera15,16,61. However, this pattern was not conserved in P. trichocarpa and other Quercus species. Specifically, the SBT6 type ortholog (TTP2) was absent in Q. lobata but duplicated in P. trichocarpa, whereas the SBT7 type S1P ortholog was missing in Q. robur. Although SBT6 TTP2 was initially considered non-essential for A. thaliana under normal growth conditions69, later studies revealed that loss-of-function mutants in both SBT6 TTP2 and SBT7 S1P exhibited an increased root gravitropic index compared to wild-type plants, linked to their role in maturation and activation of a small secretory peptide GOLVEN 1 (GLV1), which regulates auxin transport6,70. Additionally, SBT7 S1P is implicated in the unfolded protein response and stress signaling pathways12. SBT4 emerged as the most expanded subfamily within the Quercus genus, in contrast to P. trichocarpa and A. thaliana, where SBT1 and SBT3 were the most abundant subfamilies, respectively. Furthermore, clustering of AtSBT3 genes without identifiable orthologs in Quercus species or P. trichocarpa suggests that this subfamily may have emerged after the divergence of A. thaliana from Q. ilex. Similarly, the reduced number of SBT4 genes in A. thaliana and P. trichocarpa compared to Quercus species may reflect a lineage-specific expansion of Quercus, potentially associated with functional diversification in response to environmental pressures or developmental needs. The Ka/Ks ratios calculated for 55 gene pairs among QiSBTs, AtSBTs, and QlSBTs were all < 1, indicating strong purifying selection and functional constraint throughout evolution. This is consistent with findings in other plant species such as cotton65, supporting the idea that the SBT gene family has been evolutionarily conserved due to essential biological roles. Notably, two gene pairs between Q. ilex and Q. lobata (QiSBT31-Ql04p053765 and QiSBT80-Ql12p016678 pairs) showed Ka/Ks > 1, suggesting potential recent adaptive evolution and species-specific functional divergence. Regarding gene structure, members of the SBT1-SBT7subfamilies generally exhibited conserved exon–intron organization, although exceptions were observed in subfamilies with reduced intron content. The SBT family is among the few plant genes known to contain intronless or intron-poor (≤ 3 introns) genes12,71. In Q. ilex, all 19 QiSBT genes classified withing the SBT1 subfamily were either intronless or intron-poor. This structural feature may have functional implications, as intronless genes are often associated with rapid transcriptional responses and roles in stress adaptation, epigenetic regulation, and development. Moreover, SBT1 subfamily members have been involved in biotic interaction in species such as A. thaliana, S. lycopersicum, and S. tuberosum12,20,61, highlighting their relevance in defense responses. Cis-acting regulatory elements play a crucial role in gene expression regulation, particularly under biotic stress conditions72,73. To investigate regulatory potential, 1.5 kb upstream promoter regions of QiSBT genes were analyzed. Identified cis-elements were grouped into environmental stress-responsive, hormone-responsive, development-related and light-responsive categories, highlighting the involvement of QiSBTs in multiple signaling pathways, growth, development, and defense responses12,47,65. A large proportion of cis-elements were associated with environmental stress responses, suggesting that QiSBT genes may contribute to defense against wounding, pathogens and fungal elicitors. Many of these elements were also linked to key phytohormones, particularly abscisic acid, methyl jasmonate acid and salicylic acid, which are known to regulate plant immunity and stress adaptation, including response to P. cinnamomi74,75. In P. americana, salicylic acid signaling peaks early after infection, followed by jasmonic acid activation, triggering defense-related gene expression74. In silico miRNA analysis predicted seven miRNAs targeting ten QiSBT genes, indicating potential post-transcriptional regulation76. QiSBT14 was targeted by seven members of the miR166 family, a conserved miRNA group involved in development and stress responses across multiple species77–80. In P. dulcis, miR166 expression negatively correlates with stress-responsive genes, suggesting regulatory roles in abiotic and biotic stress adaptation79. Protein interaction network analysis revealed that QiSBT23, QiSBT24, and QiSBT26 co-occur with gene-21,037 encoding RPM1-interacting protein 4, a negative regulator of PAMP-triggered immunity81. QiSBT69 was predicted to interact with TTM1 and TTM2, enzymes involved in developmental regulation and immunity, where TTM2 acts as a negative regulator of oomycete resistance82. QiSBT21 was predicted to interact with PMEI9, previously identified as a positive regulator of fungal resistance in cotton83. These interactions suggest that QiSBTs may participate in both developmental regulation and immune signaling pathways. To date, and to the best of our knowledge, transcriptional responses of Q. ilex to P. cinnamomi infection remain largely unexplored. This study provides the first RNA-seq dataset generated from Q. ilex seedlings inoculated with P. cinnamomi, offering novel insights into the involvement of QiSBTs in defense-related processes. Among the 87 QiSBT genes identified, 45 were not transcriptionally detected under the experimental conditions, whereas several others exhibited relatively low expression levels. Nevertheless, a subset of QiSBT genes showed clear pathogen-responsive induction. In particular, QiSBT24, QiSBT63, and QiSBT69 were significantly upregulated following P. cinnamomi infection, suggesting their potential involvement in defense mechanisms. Beyond individual gene responses, co-expression analysis revealed a structured transcriptional organization within the QiSBT gene family84. At the network level, the QiSBT co-expression network displayed relatively high connectivity, comprising 35 genes connected by 81 significant interactions. This topology indicates that QiSBT genes do not function as isolated transcriptional units but instead form an interconnected regulatory system. Consistently, Louvain community detection identified 12 modules of variable size, supporting a modular architecture in which subsets of genes may be co-regulated and potentially associated with distinct functional roles. Integration of co-expression and predicted protein–protein interaction data highlighted QiSBT20 as a central hub within the QiSBT interaction network. This gene exhibited high connectivity in the RNA-seq-derived co-expression network and was also supported by interactions reported in the STRING database, indicating partial concordance between transcriptomic associations and predicted protein-level connectivity. Notably, QiSBT20 did not show significant differential expression between inoculated and non-inoculated seedlings, suggesting that its network centrality is not driven by transcriptional regulation under infection conditions. Instead, its stable expression together with its high topological importance points to a potential role in maintaining the structural coherence of the network. In contrast, comparison between the RNA-seq-derived co-expression network and the STRING-based protein–protein interaction network revealed no statistically significant overlap, as confirmed by Fisher’s exact test. This result suggests that transcriptional co-regulation and predicted physical interactions represent distinct and only partially convergent layers of biological organization, rather than reflecting a fully overlapping interaction framework. RT-qPCR validation further confirmed the RNA-seq results, revealing approximately 3-, 8-, and 30-fold induction for QiSBT24, QiSBT63, and QiSBT69, respectively. Among these genes, the strong induction observed for QiSBT69 supports its responsiveness under P. cinnamomi infection conditions. Functional annotation of the QiSBT69 core co-expression module revealed a broad representation of genes associated with membrane-localized components, intracellular trafficking, and nuclear- and organelle-associated functions85. In particular, GO terms related to cellular localization, including integral component of membrane, plasma membrane, nucleus, and general membrane-associated compartments, were among the most represented categories within the module. In addition, functional categories associated with molecular binding activities were highly represented, including protein binding, ATP binding, metal ion binding, nucleic acid binding, and DNA binding. These functions are commonly found in regulatory and enzymatic proteins. Consistently, transcription-related GO terms, including DNA-binding transcription factor activity and regulation of transcription, DNA-templated, were also detected, suggesting the presence of genes potentially involved in transcriptional regulation. At the biological process level, lignin biosynthetic process and transmembrane transport were among the most represented GO categories. This suggests that genes involved in cell wall-associated metabolism and transport processes are part of the co-expression module86. This pattern was further supported by KEGG pathway analysis, where phenylpropanoid biosynthesis emerged as the most represented pathway86. These results describe the functional composition of the QiSBT69 co-expression module under P. cinnamomi infection conditions. Further experimental validation will be required to clarify the specific roles of individual genes within this module, as well as to assess its potential as a molecular marker in resistance breeding programs. Conclusions In this study, we performed the first genome-wide identification and characterization of the subtilisin-like serine protease (SBT) gene family in Q. ilex, identifying 87 QiSBT genes distributed across 12 chromosomes. Phylogenetic analysis grouped these genes into seven subfamilies, consistent with other plant species such as A. thaliana, P. trichocarpa, Q. lobata, Q. robur, and Q. variabilis, but showing lineage-specific expansion patterns in Quercus. Cis-regulatory and expression analyses suggest that QiSBT genes may be involved in multiple biological processes, including responses to P. cinnamomi infection. In particular, QiSBT24, QiSBT63, and QiSBT69 showed significant transcriptional induction in inoculated seedlings, supporting their potential as infection-responsive genes of interest for further functional characterization. Network-based co-expression analysis further revealed a structured and modular organization of QiSBT transcriptional relationships. QiSBT20 exhibited high network centrality within the co-expression network, indicating a non-random organization of QiSBT transcriptional relationships under infection conditions. Co-expression analysis of QiSBT69 revealed its association with a gene module enriched in functions related to membrane components, transcriptional regulation, transport processes, and pathways linked to cell wall-associated metabolism, including phenylpropanoid biosynthesis. These results indicate that QiSBT69 is part of a transcriptional network that includes genes potentially involved in structural and metabolic responses in the plant. Overall, this study provides a comprehensive framework for understanding the evolution and potential functional diversification of QiSBT genes and offers a basis for future studies aimed at elucidating their roles in the response to P. cinnamomi. Supplementary Information Supplementary Material 1 Author contributions Daniela M. Hernández: Acquisition of data and analysis, written draft, research execution, revised the manuscript. Mónica Labella-Ortega: Research execution, revised the manuscript. María-Ángeles Castillejo: Conception, revised the manuscript. Jesús V. Jorrín-Novo: Conception, project design, revised the manuscript. María-Dolores Rey: Acquisition of data and analysis, written draft, research execution, project design, revised the manuscript. Ana Mª Maldonado-Alconada: Conception, revised the manuscript. Funding This work was supported by Grant ENCINOMICS-3 PID2022-141599OB-I00 funded by MICIU/AEI/10.13039/501100011033 and by “ERDF/EU”, and ProyExcel_00881 from the Council of Economic Transformation, Industry, Knowledge, and Universities of the Andalusian Regional Government. Data availability Transcriptome datasets have been deposited in the NCBI SRA repository under accession number PRJNA1265199. All data generated or analyzed during the present study are available within this article and its supplementary materials. Declarations Competing interests The authors declare no competing interests. Footnotes Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Supplementary Materials Supplementary Material 1 Data Availability Statement Transcriptome datasets have been deposited in the NCBI SRA repository under accession number PRJNA1265199. All data generated or analyzed during the present study are available within this article and its supplementary materials.