Genome-wide identification and alkaline stress response analysis of the class III peroxidase (PRX) gene family in Castanea mollissima

Abstract

Background

Class III peroxidases (PRXs) are plant-specific proteins crucial for growth, development, and stress responses. Castanea mollissima is an important woody crop; however, soil alkalization, a common environmental stressor, limits the sustainable development of its agricultural ecology. The evolution of the PRX gene family and its regulatory mechanisms under alkaline stress in this species remain unclear.

Results

In the Castanea mollissima genome, 98 CmPRX genes were identified, distributed unevenly across 12 chromosomes and classified into eight subfamilies. Tandem duplication was the primary driver of the expansion of this gene family. Under alkaline stress, key physiological indicators (SOD and CAT activities, MDA content) in leaves changed significantly, reflecting the plant’s physiological response and adaptive potential under such conditions. Integration of transcriptomic and physiological data revealed highly differentiated expression patterns among CmPRX members. The reliability of the RNA-seq data was validated by RT-qPCR analysis. Weighted Gene Co-expression Network Analysis (WGCNA) identified specific modules significantly correlated with antioxidant enzyme activities, suggesting a pivotal role for this gene family in regulating oxidative balance.

Conclusions

This study systematically clarifies the evolutionary characteristics of the CmPRX gene family and its expression regulatory network under alkaline stress. The findings provide a theoretical basis and candidate gene resources for improving alkaline tolerance in C. mollissima through genetic breeding, and offer important insights for selecting stress-tolerant varieties in regions affected by soil alkalization.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12870-026-08851-w.

Introduction

Reactive oxygen species (ROS) are inevitable by-products of normal plant metabolism; however, their levels escalate significantly under adverse environmental conditions. While ROS function as signaling molecules that regulate growth and stress responses, their excessive accumulation inflicts oxidative damage on cellular components [1–5]. Peroxidases (EC 1.11.1.X), present ubiquitously across biological organisms, catalyze the oxidation of substrates utilizing H₂O₂ as an electron acceptor. These enzymes play a crucial role in managing cellular H₂O₂ levels and mediating a variety of redox-dependent processes [6, 7]. Based on structural features and phylogenetic relationships, peroxidases are categorized into heme and non-heme types, with heme peroxidases further divided into three evolutionary classes (I-III). Class III peroxidases (PRXs) (EC 1.11.1.7, PRXs), which constitute a large, plant-specific multigene family, are predominantly found in higher plants and are integral to numerous physiological functions such as cell wall dynamics, developmental regulation, and adaptation to biotic and abiotic stresses [8, 9].

PRXs are part of a superfamily of glycoproteins, primarily localized to the cell wall and vacuole [7]. Their structure typically consists of a single polypeptide chain and a non-covalently bound heme prosthetic group. The catalytic activity of PRXs depends on conserved histidine residues, while structural stability is supported by disulfide bonds between conserved cysteine residues [8, 10, 11]. The PRX family demonstrates functional diversity, playing significant roles in cell wall lignification, ROS homeostasis, growth regulation, and pathogen defense [9, 12–14]. Under abiotic stresses such as salinity-alkalinity and drought, PRXs markedly enhance stress resistance by scavenging intracellular H₂O₂ and mediating the polymerization of phenolic compounds [10, 14–18]. Functional studies have shown that specific PRXs, such as AtPRX3 or AtPRX64, enhance salt and aluminum tolerance in Arabidopsis thaliana, respectively [15, 16]. Heterologous expression of rice OsPRX38 in A. thaliana reduces arsenic toxicity, associated with increased antioxidant enzyme activities [19]. Additionally, GsPRX40 in Glycine max, TaPRX-2 A in Triticum aestivum, and CsPRX17 in Cucumis sativus have all been demonstrated to confer enhanced resilience to drought, salt, or heavy metal stress [17, 18, 20]. Although the PRX gene family has been systematically identified in numerous species, including A. thaliana (73) [21], Oryza sativa (138) [22], Zea mays (119) [23], G. max (124) [20], C. sativus (60) [18], Litchi chinensis (77) [24], Nicotiana tabacum (210) [25], and Manihot esculenta, (91) [26], the evolutionary features and functional roles of this family in Castanea mollissima, particularly under alkaline stress, remain largely unexplored.

C. mollissima, belonging to the Fagaceae family, holds significant ecological and economic value. Its nuts are prized for their nutritional benefits and associated health advantages [27–29]. However, adapted to slightly acidic environments, C. mollissima exhibits particular vulnerability to alkaline stress. This stress manifests through combined ion toxicity, elevated pH injuries, and oxidative bursts, which collectively compromise vital physiological processes. Notably, these stressors disrupt photosynthesis by damaging the integrity of chloroplasts, thus curtailing both crop yield and the species’ geographical distribution [30–32]. Consequently, a genome-wide identification of the CmPRX gene family, coupled with analyses of its expression patterns under alkaline stress, becomes imperative. This study aims to furnish crucial genetic data and insights into the mechanisms underlying alkaline stress tolerance in C. mollissima via molecular breeding approaches.

Materials and methods

Identification and phylogenetic analysis of CmPRX genes

Genomic data for C. mollissima were retrieved from the Castanea Genome Database (http://castaneadb.net/#/). The peroxidase domain’s Hidden Markov Model (HMM) profile (PF00141) was acquired from the Pfam database (http://pfam-legacy.xfam.org/). Candidate PRX proteins were initially identified using a dual strategy: (i) conducting a BLASTP (v2.2.30) search against the C. mollissima proteome, using 73 A. thaliana PRX protein sequences as queries [33], with an E-value cutoff of 1e-5; and (ii) performing an independent search with the HMMER (v3.0) tool using the PF00141 HMM profile with an E-value threshold below 1 × 10⁻³. Results from both searches were integrated to compile a preliminary list of candidates. These candidates were then verified to confirm the intact presence of the peroxidase domain through analyses using the NCBI-CDD (https://www.ncbi.nlm.nih.gov/cdd/) and SMART (http://smart.embl-heidelberg.de/) databases. The physicochemical properties of the identified proteins were predicted using the ProtParam tool on the ExPASy (https://www.expasy.org/) server. Predictions of subcellular localization were made using WoLF PSORT (https://wolfpsort.hgc.jp/), and conserved protein motifs were identified with the MEME suite (https://meme-suite.org/meme/). The gene structures were visualized using TBtools (v2.363), based on the GFF3 annotation data [34]. A phylogenetic tree, including both C. mollissima and A. thaliana PRX proteins, was constructed employing the Maximum Likelihood method in MEGA software, version 11.0 [35], with the LG + G4 amino acid substitution model, which was determined as the best-fit model by the built-in Model Selection tool. Branch reliability was assessed using 5000 bootstrap replicates.

Chromosomal localization and collinear analysis

The chromosomal positions of the CmPRX genes were determined from genome annotations and displayed using TBtools [34]. Genomic data for comparative species, including A. thaliana, Quercus robur, Solanum lycopersicum, Vitis vinifera, Z. mays, O. sativa, and Pyrus bretschneideri, were sourced from the Phytozome database (https://phytozome-next.jgi.doe.g.ov/) for interspecies collinear analysis [36, 37]. The types of gene duplications within the CmPRX family were classified using the “duplicate_gene_classifier” module in MCScanX [38]. Employing approaches established in prior studies, events of whole-genome duplication (WGD) and segmental duplication were meticulously differentiated by integrating synonymous substitution rates (Ks) of collinear blocks, homologous collinearity dot plots, and complementarity analysis of these blocks [37, 39]. Specifically, the Ka (non-synonymous substitution rate) and Ks values of homologous gene pairs on collinear blocks were calculated using KaKs_Calculator. Subsequently, a custom Python script was employed to compute the median Ks value across all homologous gene pairs within each collinear block.

Cis-acting elements, transcription factor regulation, and functional enrichment analysis

The 2000 bp sequences upstream of the start codons of CmPRX genes were analyzed to identify cis-acting elements using the PlantCARE database (https://bioinformatics.psb.ugent.be/webtools/plantcare/html/) [40]. Potential transcription factor (TF) regulatory relationships targeting CmPRX genes were predicted using the Plant Transcription Factor Database (PlantTFDB v5.0; https://planttfdb.gao-lab.org/) [41, 42]. PlantTFDB is a well-established and widely used plant TF database that integrates comprehensive, experimentally validated TF binding motif data from multiple sources, including the CIS-BP database and high-throughput DAP-seq datasets [43, 44]. The prediction is based on scanning the promoter sequences against this curated collection of TF binding motifs using a position weight matrix (PWM) approach, which is a standard and widely accepted methodology for computational identification of potential TF binding sites in plant genomics studies [45]. Specifically, the 2000 bp upstream promoter sequences of each CmPRX gene were submitted to the PlantTFDB binding site prediction tool, and a significance threshold of P-value below 1e-4 was applied to filter high-confidence TF–target interactions. The resulting regulatory network was visualized using Cytoscape version 3.9.1 [46]. Additionally, a protein-protein interaction (PPI) network for the CmPRX proteins was established using the STRING (https://cn.string-db.org/) database [47]. GO and KEGG enrichment analyses and visualizations were conducted using TBtools [34].

Plant materials and phenotype determination

In October 2025, 150-day-old potted seedlings of the C. mollissima cultivar ‘Yanshanzaofeng’, which exhibited uniform growth, were acquired from the experimental base at Hebei Normal University of Science and Technology in Qinhuangdao, Hebei Province, China. These seedlings were maintained in an artificial climate chamber. Alkaline stress treatments were administered through irrigation with 150 mL of Na₂CO₃ solution at concentrations of 0.02 g/L (T0.02) and 0.5 g/L (T0.05), while seedlings irrigated with pure water served as controls (CK). Leaf samples were collected seven days post-treatment, immediately flash-frozen in liquid nitrogen, and stored at -80 °C. Physiological responses were assessed by measuring the malondialdehyde (MDA) content and the activities of catalase (CAT) and superoxide dismutase (SOD) using commercial assay kits from Suzhou Keming Biotechnology Co., Ltd., Suzhou, China, according to the manufacturer’s instructions. The proline (PRO) content was determined utilizing the sulfosalicylic acid method [48]. Statistical significance of differences among treatment groups was assessed using one-way analysis of variance (ANOVA) followed by least significant difference (LSD) test and Duncan’s multiple range test at P < 0.05, performed using SPSS software. Each physiological measurement was performed with three biological replicates.

RNA sequencing, DEG analysis, and RT-qPCR validation

Total RNA was isolated from leaf samples using the CTAB-PBIOZOL method, which was enhanced by ethanol precipitation [49]. The quality and concentration of the RNA were assessed with an Agilent 2100 Bioanalyzer (Agilent Technologies, CA, USA) and a NanoDrop spectrophotometer (Thermo Fisher Scientific, MA, USA), respectively. Sequencing libraries were prepared according to the standard protocol provided by the NEBNext^® Ultra™ RNA Library Prep Kit for Illumina (NEB, USA). All raw sequencing reads generated in this study were deposited in the NCBI BioProject repository (Accession No. PRJNA1384844). After data filtration, high-quality clean reads were aligned to the C. mollissima reference genome using HISAT2 software (v2.0.5) [37, 50–52]. Transcript abundance was quantified using FeatureCounts, with normalization performed using the Fragments Per Kilobase of transcript per Million mapped reads (FPKM) metric [53]. The DESeq2 R package (v1.4.5) was employed to identify differentially expressed genes (DEGs) across treatment groups [54]. Genes were considered significantly differentially expressed when meeting the following criteria: |log₂Fold Change| ≥ 1 and FDR < 0.05. The FDR was calculated by applying the Benjamini-Hochberg method for multiple testing correction to the raw P-values. For validation, first-strand cDNA was synthesized from total RNA using the Evo M-MLV RT Mix Kit with gDNA Clean for qPCR Ver.2. Quantitative real-time PCR (RT-qPCR) assays were conducted on an ABI 7500 system (Applied Biosystems Inc., Foster City, CA, USA) with TB Green Premix Ex Taq (Takara). The thermal cycling program included a pre-denaturation step at 95 °C for 3 min, followed by 40 cycles of 95 °C for 10 s and 60 °C for 20 s. Specific primer sequences are detailed in Table S1. The C. mollissima Actin gene served as the endogenous control, and relative expression levels were calculated using the 2^(−ΔΔCT) method [55, 56]. Additionally, Weighted Gene Co-expression Network Analysis (WGCNA) was carried out on the MetWare Cloud platform (https://cloud.metware.cn/#/tools/detail?id=247), using the WGCNA R package. A soft-thresholding power of 18 was selected based on the scale-free topology criterion. Modules were identified with a minimum module size of 50 genes. Highly similar modules were merged using a module merging threshold of 0.25. Module–trait relationships were assessed using Pearson correlation coefficients, and a correlation coefficient cutoff of 0.6 was applied for heatmap visualization.

Results

Identification, physicochemical properties, and phylogenetic analysis of CmPRX genes.

A total of 98 CmPRX genes were identified in the C. mollissima genome and were named based on their evolutionary relationship with homologs from A. thaliana (Table S2). Physicochemical analysis indicated that the lengths of CmPRX proteins ranged from 262 aa (CmPRX39b) to 413 aa (CmPRX48a), with molecular weights varying between 28.84 and 48.09 kDa. The isoelectric points (pI) of these proteins ranged from 4.33 to 9.84, with 59.18% of the proteins exhibiting a pI greater than 7, suggesting their adaptation to alkaline physiological environments. The majority (86.69%) of these proteins were hydrophilic, as indicated by GRAVY scores below zero. Predictions of subcellular localization suggested that 96 of these proteins are localized to the cell wall, plastid, or vacuole. Notably, the aliphatic index for all CmPRX proteins exceeded 70, with more than three-quarters surpassing 80, indicative of high thermostability (Table S2). Phylogenetic analysis grouped the C. mollissima and A. thaliana PRX proteins into eight subfamilies, labeled Group 1 through Group 7 (Fig. 1A). The distribution of members across these subfamilies was uneven, with Group 4 containing the most members (43 CmPRXs), while Group 6 had only two members (Figs. 1B, C).

Fig. 1.

Phylogenetic analysis of the PRX genes of C. mollissima and A. thaliana. (A) Maximum Likelihood phylogenetic tree of PRX members from C. mollissima (98) and A. thaliana (73). A complete version with bootstrap support values displayed at all nodes is provided in Supplementary Figure S1. Different background colors represent different subfamilies. (B) Population distribution of 171 PRX genes from C. mollissima and A. thaliana, in eight subfamilies. (C) The proportion of the PRX genes in C. mollissima across eight subfamilies

Conserved domains, motifs, and gene structure analysis

Domain analysis confirmed that all CmPRX proteins possess the typical Secretory_peroxidase domain (Figs. 2A, B). Motif analysis revealed that all members contained at least seven motifs, with motif composition within each subfamily showing high consistency, reflecting strong evolutionary conservation (Fig. 2C). Gene structure analysis demonstrated a relatively stable exon-intron pattern within the family, predominantly characterized by genes containing three to four exons (78.57%), and very few genes having fewer than two or more than five exons (Fig. 2D; Table S2).

Fig. 2.

The conserved domains, conserved motifs distribution, and gene structure of CmPRX genes. (A) Phylogenetic analysis of CmPRX genes. (B) The conserved domains of CmPRX proteins. (C) The conserved motifs of CmPRX proteins. (D) Gene structure of CmPRX genes

Chromosomal distribution and duplication analysis

The CmPRX genes were unevenly distributed across the 12 chromosomes of C. mollissima. Chromosomes 8, 6, and 1 had the highest numbers of genes (containing 17, 16, and 15 genes, respectively), whereas Chromosomes 3, 5, 7, 10, 11, and 12 each contained only four genes (Fig. 3A). Intraspecific collinear analysis identified nine collinear gene pairs (Fig. 3B). Using MCScanX in combination with analysis of collinear blocks and synonymous substitution sites (Ks), we differentiated between WGD and segmental duplication, consistent with previous studies [39, 57] (Fig. 3C). Tandem duplication was identified as the primary mechanism driving the expansion of the CmPRX gene family, involving 41 genes (41.84%). The proportions of other duplication types were as follows: Dispersed (27, 27.55%), Proximal (12, 12.24%), WGD (12, 12.24%), and Segmental (6, 6.12%) (Fig. 3D). These results suggest that tandem duplication has played a critical role in the rapid evolution and environmental adaptation of the CmPRX gene family.

Fig. 3.

Chromosome distribution and duplication type analysis of CmPRX genes. (A) Chromosome distribution of CmPRX genes. The color of segments in the chromosomes shows the gene density of the corresponding region. (B) Collinear relationship within C. mollissima genome. (C) Homologous collinear dot-plot within C. mollissima genome. The collinear blocks from WGD containing CmPRX genes are marked in the pale yellow boxes of the figure, and the blocks length and median Ks of the collinear blocks are marked. The genes highlighted in green are identified as from WGD events. (D) The pie chart displaying the proportion of CmPRX quantity for different duplication types

Evolutionary analysis of CmPRX in several plant species

To further explore the evolution of the CmPRX genes, we conducted a collinear analysis between C. mollissima and seven representative species (O. sativa, Z. mays, S. lycopersicum, V. vinifera, A. thaliana, P. bretschneideri, Q. robur) (Fig. 4). We identified 26, 8, 49, 55, 28, 70, and 49 homologous gene pairs between C. mollissima and these species, respectively. These homologous blocks contained 15, 8, 36, 43, 21, 41, and 43 CmPRX genes, respectively (Figs. 4A, B; Tables S3-S9). Notably, CmPRX51b and CmPRX52h were found to have homologs in all seven species. According to the gene dosage hypothesis, genes that are highly conserved across multiple species typically participate in critical biological processes and are sensitive to dosage changes, which favors their retention during evolution [58, 59]. Furthermore, CmPRX52c has five and four homologous copies in O. sativa and P. bretschneideri, respectively. CmPRX52a and CmPRX67a each have four copies in O. sativa and P. bretschneideri, respectively. In contrast, 33 CmPRX genes lacked homologs in any of the seven species, suggesting that these genes may have originated from lineage-specific events after speciation or underwent rapid sequence divergence potentially linked to adaptive traits specific to C. mollissima [60] (Figs. 4B, C; Tables S3-S9). Statistical analysis of collinear blocks revealed that the number of blocks between C. mollissima and the other species was 21, 7, 46, 53, 28, 63, and 15, respectively (Fig. 4D; Tables S10-S16).

Fig. 4.

Collinear analyses between the CmPRX genes and genes in seven representative plant species (A. thaliana, O. sativa, Z. mays, Q. robur, V. vinifera, S. lycopersicum, and P. bretschneideri). (A) The genetic relationship between eight plant species. (B) The dual collinear plot between C. mollissima and seven representative plant species. Gray lines in the background indicate collinear blocks within C. mollissima and other plant genomes, while red lines highlight collinear PRX gene pairs. Species names correspond to chromosome colors. (C) Heatmap of orthologous gene pairs between C. mollissima and seven other plant genomes. (D) The block number, median block length, average block length and orthologous gene pair number between C. mollissima and other seven plant genomes

Cis-acting elements, TF regulatory networks, and functional enrichment analysis

In the 2,000 bp promoter regions of the 98 CmPRX genes, we identified 2,286 cis-acting elements belonging to 59 categories, which were classified into four functional groups (Fig. 5; Table S17). Multiple elements related to stress response, light signaling, and hormone regulation were distributed across almost all members, including the stress-responsive elements ARE (anaerobic induction) and MBS (MYB binding site involved in drought inducibility); light-responsive elements Box 4, G-box, GT1-motif, and TCT-motif; and hormone-responsive elements ABRE (abscisic acid) and CGTCA-motif (MeJA) (Figs. 5A, B). The number of cis-elements varied significantly among the genes. For instance, CmPRX47p, CmPRX47c, and CmPRX20 contained the most elements (60, 40, and 40, respectively), while CmPRX59b and CmPRX72a contained the fewest (12 and 10, respectively) (Fig. 5C). Further analysis predicted that 412 transcription factors (TFs) from 43 families could bind to CmPRX promoters (Fig. 6A; Table S18). Families such as ERF, MYB, and NAC were the most abundant, suggesting that CmPRX genes are regulated by multiple stress signaling pathways, consistent with their core functions in stress response [61–63]. Additionally, CmPRX48c, CmPRX47h, and CmPRX47p exhibited high TF binding density, indicating their potential key roles in the regulatory network (Figs. 6A, B). GO and KEGG enrichment analyses confirmed that family members are primarily involved in phenylpropanoid biosynthesis, cell wall organization, and oxidative detoxification (Figs. 6C, D).

Fig. 5.

Prediction of cis-acting elements in the promoters of CmPRX genes (A) Cis-acting acting elements in the promoters of CmPRX genes. Various color symbols present different elements, and their position in the figure indicates their relative position on the promoter. (B) The relative proportions of different cis regulatory elements in the promoters of CmPRX genes are indicated in the chart. The same color represents cis-acting elements sharing identical or similar functions. (C) The number of various cis-acting elements in the promoters of each CmPRX genes

Fig. 6.

GO/KEGG enrichment and TFs regulatory network analysis of CmPRXs. (A) The putative TFs regulatory network analysis of CmPRXs. (B) Wordcloud of predicted TFs interacting with CmPRX genes. The font size is positively correlated with the number of corresponding TFs. (C) GO function enrichment analysis of CmPRXs. (D) KEGG function enrichment analysis of CmPRXs

RNA sequencing and DEGs analysis

To investigate the transcriptional regulatory mechanisms of C. mollissima under alkaline stress, plants were exposed to solutions of Na₂CO₃ at concentrations of 0.02 g/L (T0.02) and 0.5 g/L (T0.5) for seven days, with plants treated with pure water serving as the CK. Physiological assays revealed that alkaline stress significantly reduced the activity of SOD. In contrast, CAT activity and MDA content increased notably at the lower concentration of 0.02 g/L, indicating oxidative damage and the activation of the antioxidant defense system. The content of PRO exhibited a decreasing trend (Fig. 7A). Nine cDNA libraries were constructed, comprising three biological replicates per treatment. Transcriptome sequencing yielded 73.59 Gb of high-quality data. The percentages of Q20 and Q30 quality scores were above 98.88% and 96.85%, respectively, across all samples. The GC content remained stable, ranging from 43.34% to 43.55%, and the mapping rates to the reference genome varied from 90.72% to 91.23% (Tables S19-20). Principal Component Analysis (PCA) showed a distinct separation between the treatment groups (Fig. 7B), confirming the quality and reliability of the data. In total, 2,126 DEGs were identified. In the comparison between T0.02 and CK, 360 genes were up-regulated, and 331 were down-regulated. In the comparison between T0.5 and CK, the number of up-regulated genes increased to 552, while down-regulated genes reached 796 (Figs. 7C, D, E, G). The number of DEGs rose significantly with increasing stress concentration, with up-regulation consistently surpassing down-regulation. Functional enrichment analysis revealed that the DEGs were significantly enriched in pathways related to plant-pathogen interaction, phenylpropanoid biosynthesis, and MAPK signaling. Under high-concentration stress, pathways related to hormone signal transduction and hydrogen peroxide metabolism were also significantly activated, suggesting a complex regulatory network in response to alkaline stress (Figs. 7F, H).

Fig. 7.

Effect of alkaline stress treatments on the physiological index of C. mollissima and transcriptome analysis results. (A) The antioxidant enzyme activity (SOD, CAT), MDA content, and proline content in C. mollissima leaves under alkaline stress treatment. Error bars represent the standard error (SE) of the mean from three biological replicates. Statistical significance was analyzed by one-way ANOVA followed by LSD and Duncan’s multiple range tests (*p < 0.05, **p < 0.01; ns, not significant). (B) PCA analysis of transcriptome data. (C) DEGs distribution in three comparisons at different treatment groups. (D) DEGs distribution in four comparisons at different treatment groups. (E) DEGs of T0.02 vs. CK. (F) GO and KEGG enrichment analysis of DEGs of T0.02 vs. CK. (G) DEGs of T0.5 vs. CK. (H) GO and KEGG enrichment analysis of DEGs of T0.5 vs. CK

WGCNA analysis

Using WGCNA in conjunction with transcriptomic and phenotypic data, we identified critical gene co-expression modules and hub genes associated with the response to alkaline stress (Figs. 8A, B) [64]. Among the 14 modules identified, the “blue” and “green” modules demonstrated significant correlations with various physiological indices. The blue module exhibited a positive correlation with SOD (r = 0.81, p = 8.1 × 10⁻³) and PRO (r = 0.88, p = 1.7 × 10⁻³), and a negative correlation with MDA (r = -0.75, p = 0.02) and CAT (r = -0.51, p = 0.11). Conversely, the green module was negatively correlated with SOD (r = -0.91, p = 6.6 × 10⁻⁴) and PRO (r = -0.74, p = 0.023), and positively correlated with MDA (r = 0.92, p = 4.4 × 10⁻⁴) and CAT (r = 0.75, p = 0.02). These findings suggest that the blue and green modules might be involved in distinct regulatory mechanisms in response to alkaline stress. Functional enrichment analysis revealed that genes within the blue module were primarily associated with chloroplast functions, photosynthesis, and the regulation of gene expression. In contrast, the green module was significantly enriched for protein degradation and endomembrane transport pathways. Notably, 19 out of 98 CmPRX genes were located within these two pivotal modules, underscoring their essential roles in the adaptation of C. mollissima to alkaline stress (Figs. 8C-F).

Fig. 8.

WGCNA and functional enrichment analysis of the blue and green modules. (A) The clustering dendrogram of genes identifying the WGCNA modules. (B) The correlation of the identified modules with the phenotypic traits at different developmental stages. (C) GO enrichment analysis of the blue module. (D) KEGG enrichment analysis of the blue module. (E) GO enrichment analysis of the green module. (F) KEGG enrichment analysis of the green module

CmPRX gene expression, physiological correlation, and RT-qPCR validation

The CmPRX genes showed significant differential expression. Genes such as CmPRX52l, CmPRX53c, and CmPRX52k were significantly up-regulated as stress intensified. For instance, CmPRX53c’s expression (FPKM) approximately doubled from 12.99 to 24.02, while CmPRX52k’s expression tripled from 2.82 to 8.06. In contrast, several genes such as CmPRX12, CmPRX47n, CmPRX67a, and CmPRX51b were down-regulated, with CmPRX47n displaying the most significant decrease from 66.43 to 11.37 (Fig. 9A). Correlation analysis indicated potential associations between gene expression and phenotypic traits. For example, CmPRX52g expression was significantly positively correlated with SOD activity, whereas CmPRX25 and CmPRX64c showed negative correlations. Moreover, the expression levels of CmPRX26, CmPRX11, CmPRX21, CmPRX58, CmPRX47h, CmPRX52j, and CmPRX47k were positively correlated with PRO content, while CmPRX20 was negatively correlated. CmPRX72c exhibited a negative correlation with MDA content. Remarkably, CmPRX31 and CmPRX10a showed extremely significant positive correlations with MDA and CAT content, respectively (Fig. 9B). To validate the reliability of the RNA-seq data, four representative CmPRX genes were selected for RT-qPCR validation based on their significant and representative changes in FPKM values observed under alkaline stress treatments in the RNA-seq data, representing distinct expression response patterns (Fig. 9C). Notably, significant upregulated observed in CmPRX52k and CmPRX53c was highly consistent with the transcriptomic profiles.

Fig. 9.

The expression profiles of CmPRX genes, their correlation with phenotypic indicators and physiological index of C. mollissima. (A) The expression profiles of CmPRXs at different alkaline stress treatments. (B) The correlation between phenotypic indicators and CmPRXs expression levels at different alkaline stress treatments. ***, ** and * denote genes and physiological indicators exhibiting significant differences (p < 0.001, p < 0.01, p < 0.05, respectively). Pearson correlation coefficients were used. Benjamini-Hochberg FDR correction was applied for multiple testing adjustment. (C) RT-qPCR analysis of CmPRXs in C. mollissima at different alkaline stress treatments. Error bars represent the standard error (SE) of the mean from three biological replicates, each containing three technical replicates. The lowercase letters indicate the significance of differences between FPKM of transcriptome data from different treatment groups, while uppercase letters indicate the significance of differences between RT-qPCR results from different treatment groups (one-way ANOVA followed by Duncan’s multiple range test, P < 0.05)

Discussion

PRXs are enzymes exclusive to plants, playing crucial roles in the regulation of plant growth and development, response to abiotic stresses, and maintenance of ROS metabolic balance [6, 7, 9, 12–14, 65, 66]. Although researchers have identified and functionally characterized the PRX gene family in various plant species [15–20, 22, 33], its specific composition, evolutionary traits, and biological functions in the C. mollissima have yet to be fully explored. To bridge this knowledge gap, the present study undertook the first genome-wide systematic analysis of the C. mollissima PRX gene family, aiming to elucidate its potential roles in the response to alkaline stress and to provide a theoretical foundation for subsequent functional research.

In this study, 98 CmPRX genes were identified within the C. mollissima genome. Phylogenetic analysis segregated these genes into eight subfamilies (Fig. 1). The encoded proteins demonstrate substantial variation in amino acid length and molecular weight; nevertheless, the majority are characterized as alkaline, hydrophilic, and thermostable, localizing predominantly to the cell wall, plastids, or vacuoles. These attributes align with those observed in the PRX superfamily across various species [7, 18, 67, 68]. The elevated aliphatic index of these proteins suggests a robust structural backbone, likely supported by conserved disulfide bonds and calcium-binding sites. These structural features are essential for maintaining enzymatic activity under variable environmental conditions [69]. Moreover, the alkaline peroxidases display a heightened affinity for the negatively charged matrix of the cell wall, positioning them as principal contributors to lignification and cell wall reinforcement under stress conditions [70–72]. Their extracellular localization strategically situates CmPRXs at the forefront of defense mechanisms. This positions them to directly regulate apoplastic ROS homeostasis and thus mitigate oxidative damage induced by alkaline stress [13]. The evolutionary adaptation of these alkaline and thermostable CmPRX proteins to the apoplastic environment enables them to function effectively as ROS scavengers or cell wall modifiers in response to alkaline stress. The structural consistency among them is further affirmed by the universal presence of the Secretory_peroxidase domain and the highly conserved motifs identified in our analysis (Fig. 2). Additionally, nearly 80% of CmPRX genes are composed of only 3–4 introns, suggesting a compact intron-exon structure that provides insights into the transcriptional dynamics of the CmPRX family. According to the “genome economy” hypothesis, genes that require rapid induction in response to environmental stimuli typically evolve with fewer introns and shorter lengths [73]. Alkaline stress rapidly triggers a substantial and potentially harmful ROS burst necessitating swift molecular defenses [74]. Accordingly, the streamlined genomic structure of CmPRX genes likely supports increased transcriptional elongation rates and more efficient splicing. Consequently, this facilitates the rapid accumulation of PRX proteins and activation of antioxidant systems upon stress detection, thereby minimizing oxidative damage to cellular components.

Interspecific collinear analysis has revealed complex evolutionary trajectories for the CmPRX gene family within angiosperms. Consistent with expectations, C. mollissima exhibited significantly greater collinearity with dicots than with monocots, reflecting major divergence events in early angiosperm evolution [75]. However, the number and length of collinear blocks did not exhibit a strict correlation with phylogenetic distance. This phenomenon, observed in other gene families of C. mollissima, is likely attributable to lineage-specific rates of genome fractionation and chromosomal rearrangement [37, 51, 57, 76]. For instance, the genome of V. vinifera retains a highly conserved ancestral karyotype structure and thus often exhibits stronger collinear signals compared to other related species [75]. Despite variations in macro- collinear, gene retention patterns at the micro-level revealed critical functional constraints. Notably, CmPRX51b and CmPRX52h retained homologs across all seven tested species (Figs. 4C, D; Tables S10-S16). This deep conservation across the monocot-dicot divide suggests that these core members likely perform indispensable housekeeping functions and are thus preserved from loss through strong purifying selection [77]. Conversely, C. mollissima-specific non-collinear genes likely originated from rapid adaptive evolution following speciation to occupy specific ecological niches.

Gene duplication is a significant mechanism driving gene family expansion and functional diversification [78–82]. In this study, 41.84% of CmPRX genes originated from tandem duplications, underscoring its substantial role in the expansion of this family. Interestingly, four distinct gene clusters on Chr4, Chr6, and Chr8 were primarily formed by tandem duplication (Fig. 3). Phylogenetic analysis showed that members of the Chr4 cluster reside in the same branch as AtPrx47 (Fig. 1). Sequence similarity suggests functional similarity [83–85]. Given that AtPrx47 is associated with stem tissue support strength in Arabidopsis [21], the members of the Chr4 cluster may fulfill a similar function in C. mollissima. Likewise, a cluster of seven members on Chr6 (CmPRX53b, CmPRX53d, CmPRX59a-59e) clustered with AtPrx53 and AtPrx59. AtPrx53 is known to be primarily responsible for secondary cell wall lignification, providing mechanical strength [9], while AtPrx59 is involved in cell wall remodeling and fortification [72]. These findings further support the hypothesis that the seven members of the Chr6 cluster likely participate in cell wall lignification and mechanical strengthening in C. mollissima. Notably, as a perennial woody plant, C. mollissima has a much higher demand for lignin synthesis and cell wall strength compared to the herbaceous plant Arabidopsis. Tandem duplication has likely specifically increased the copy number of certain CmPRX members to meet the specific growth, development, and environmental adaptation requirements of C. mollissima.

Alkaline stress disrupts cellular redox homeostasis by inducing ion toxicity and elevated pH, which leads to a rapid accumulation of ROS. This accumulation results in membrane lipid peroxidation and structural damage to cells [30–32]. In our study, we observed an increase in MDA content alongside dynamic shifts in the activities of SOD and CAT, indicating that the C. mollissima seedlings experienced considerable oxidative stress and activated their enzymatic defense systems [86]. Our combined transcriptomic and physiological analyses identified distinct functional roles for specific CmPRX members (Figs. 7, 8 and 9). Notably, CmPRX52g and CmPRX10a exhibited significant positive correlations with the activities of SOD and CAT, respectively. Given that SOD catalyzes the dismutation of superoxide anions into H₂O₂, which is subsequently detoxified by CAT and PRXs, these co-expression patterns suggest a coordinated regulatory mechanism. This mechanism likely involves these CmPRXs acting in concert with the canonical antioxidant machinery to establish an efficient ROS scavenging module [87]. Furthermore, CmPRX26 and six other genes showed positive correlations with PRO content. Considering PRO’s role as a critical osmoprotectant, the synchronous upregulation of these genes likely represents a comprehensive stress response strategy that manages osmotic potential while potentially reinforcing the cellular barrier against the physiological drought induced by alkaline conditions [88]. Importantly, the correlation with MDA levels revealed divergent response patterns within the PRX family. The significant negative correlation of CmPRX72c with MDA suggests that its upregulation contributes to mitigating oxidative damage. Conversely, CmPRX31 showed a strong positive correlation with MDA. This indicates that CmPRX31 may function as a stress-inducible gene whose expression increases with stress severity, or it may be involved in lignification processes that produce ROS as a byproduct [69]. These findings demonstrate that the CmPRX family orchestrates a multifaceted response to alkaline stress through specific functional modules rather than acting as a single unit. The results from WGCNA further anchored these genes within co-expression modules tightly linked to stress traits. The enrichment of binding motifs for stress-responsive TFs, such as ERF and MYB, in the promoters of these genes suggests an upstream regulatory hierarchy that fine-tunes CmPRX expression in response to oxidative and osmotic challenges [61, 62]. Through genome-wide identification and combined physiological-transcriptomic analysis, this study systematically delineates the evolutionary patterns of the C. mollissima PRX gene family and its differentiated expression in response to alkaline stress. This discovery not only enhances our understanding of the regulatory mechanisms of stress resistance in the CmPRX family but also provides a theoretical basis for identifying key genes associated with tolerance to saline-alkaline conditions and for conducting genetic improvements in C. mollissima.

Authors’ contributions

LY and XW conceived and designed the project. XL, TZ, and GL performed the bioinformatics analysis. QL, YZ, DW, XuW, GZ and HZ performed the experiments. XL wrote the manuscript. XW and LY revised the manuscript. LY and XW supervised the manuscript. All authors read and approved the manuscript.

Funding

This research was funded by the Science Research Project of Hebei Education Department (QN2025036), and the Scientific Research Foundation of Hebei Normal University of Science and Technology (2023YB027).

Data availability

The datasets generated and/or analysed during the current study are available in the NCBI repository, https://www.ncbi.nlm.nih.gov/bioproject/PRJNA1384844.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.