Integrated transcriptomic and metabolomic analysis revealed the response of resistant variety Zhonghua 6 to the infection of Ralstonia solanacearum in peanut

Abstract

Peanut (Arachis hypogaea L.) is an important oil and economic crop, and its production has long been severely threatened by soil-borne bacterial wilt (BW) disease. However, the molecular mechanism of host resistance to it has not yet been systematically elucidated. In this study, the highly resistant peanut variety Zhonghua 6 was used as the research object. Through transcriptomic analysis, a total of 1,122 differentially expressed genes (DEGs) were identified between carefully designed treatment and control groups. WGCNA analysis led to the discovery of 14 hub genes, including two cytochrome P450 genes and a UGDH gene. Through metabolomic analysis, 1,614 differentially accumulated metabolites (DAMs) were identified, and 6-methylcoumarin, erucamide, and piceatannol were confirmed to inhibit the growth of R. solanacearum. Integrative transcriptomic and metabolomic analyses uncovered a comprehensive immune regulatory network consisted of genes involved in key pathways associated with R. solanacearum infection such as MAPK signaling, plant hormone signal transduction, phenylpropanoid biosynthesis, flavonoid biosynthesis, and ABC transporter. Overall, these results provide new insights into the molecular mechanisms governing peanut resistance to R. solanacearum, which might assist in the mining of resistance-related genes, developing of new disease control measures as well as breeding of novel disease-resistant cultivars in peanut.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12870-026-08883-2.

Introduction

The disease caused by Ralstonia solanacearum infection in plants is called bacterial wilt (BW), which is one of the most serious bacterial diseases in plant production. R. solanacearum is a Gram-negative bacterium and a type of soil-borne phytopathogenic bacterium [1] which can infect more than 250 plant species [2], including peanut (Arachis hypogaea L.), an allotetraploid oil and economic crop worldwide [3]. BW causes severe losses to peanut production, resulting in 10%−30% yield reduction and up to 50% in severe outbreaks [4]. Utilization of host resistance has been proven to be the most effective control method of BW in peanut production.

During the long-term evolutionary process of plants, a variety of defense mechanisms have been developed to resist the invasion of pathogenic bacteria. Physical defense serves as the first barrier against pathogenic bacterial invasion; when pathogenic bacteria invade, plants thicken or strengthen their cell walls by depositing lignin, callose, or hydroxyproline-rich glycoproteins (HRGPs) [5, 6], thus inhibiting the spread of pathogens. Secondly, the plant immune system is the core for plants to resist the invasion of pathogenic bacteria, which mainly includes two levels: PAMP (pathogen-associated molecular patterns)—triggered immunity (PTI) and effector-triggered immunity (ETI) [7]. The primary effector proteins of R. solanacearum inhibit PTI; however, some effector proteins (such as PopP2 and RipE1) are recognized by plant R proteins (e.g., RRS1-R) which activating ETI [8]. At the downstream of PTI or ETI activation, various plant hormones play central roles in triggering the plant immune signaling network [9–12]. For instance, salicylic acid (SA) biosynthesis is induced during both PTI and ETI following the recognition of PAMPs or effectors [13, 14]. However, these studies were mostly conducted in model plant, and the resistance mechanism in crops such as peanut has been largely unknown.

With the rapid development of high-throughput sequencing, transcriptomics has been widely applied to study plant-pathogen interactions [15, 16]. The release of the tetraploid reference genome of peanut [17–20] has provided a solid foundation for peanut transcriptomic analysis. In recent years, transcriptomic studies related to R. solanacearum have been conducted on many crops, including peanut [21], tobacco [22], tomato [23], and potato [24]. In addition to differentially expressed genes (DEGs), weighted gene co-expression network analysis (WGCNA) were used to analyze the expression patterns of all genes. It clusters highly co-expressed genes into functional modules, screens out key gene modules closely associated with specific traits, and further helps researchers to identify potential biomarkers. Currently, WGCNA has been widely used to identify centrally connected hub genes in humans, animals, and plants [25–27]. Recently, metabolomic profiling has been applied to reveal the metabolites involved in the resistance response of crops such as tobacco and tomato to R. solanacearum [28–30]. For example, naringenin, a flavonoid metabolite, can disrupt the structure of R. solanacearum in tobacco (KCB-1), inhibit its growth and reproduction, and exert a controlling effect on BW [31]. However, metabolomic profiling has not been conducted to reveal the metabolites involved in the resistance response of peanut to R. solanacearum.

In recent years, studies on peanut resistance to R. solanacearum have mostly focused on quantitative trait locus (QTL) mapping [32–34], while integrated omics analysis remains relatively scarce. The improved peanut variety Zhonghua 6 has been highly and stably resistant to BW disease since released by the Oil Crops Research Institute of the Chinese Academy of Agricultural Sciences in 2000. In this study, using this variety as materials, integrated transcriptomic and metabolomic analyses was conducted for the first time to investigate the resistance mechanism of peanut to R. solanacearum. Key metabolic pathways related to R. solanacearum infection were identified through the integrated analysis of transcriptomics and metabolomics, hub genes were mined via WGCNA, and a potential defense network underlying peanut responses to R. solanacearum infection were elucidated, thereby providing a theoretical basis for breeding disease-resistant peanut varieties.

Materials and methods

Experimental materials and inoculation

Zhonghua 6, a variety with high resistance to the bacterial wilt disease released by the Oil Crop Research Institute of the Chinese Academy of Agricultural Sciences [32], was used as the experimental material. Its seeds were germinated in nutrient soil. After the hypocotyls elongated and leaves emerged, the seedlings were then transferred to a hydroponic tray for cultivation at 26 °C, 60% relative air humidity, and a 16h/8h light/dark photoperiod (as shown in Figure S1A). When the seedlings grew to the three-leaves stage(as shown in Figure S1B), their root tips were removed and then inoculated with the R. solanacearum strain HA8-53, followed the previous reported protocol [35].

Sample collection

The taproots of peanut were collected at 24, 48, 72, and 96 h post inoculation (hpi). The treatment groups, which were infected with R. solanacearum, were denoted as T24ZH6, T48ZH6, T72ZH6, and T96ZH6, respectively. The control groups, which were treated with water without R. solanacearum, were marked as C24ZH6, C48ZH6, C72ZH6, and C96ZH6, respectively. For the transcriptome analysis, 3 biological replicates were collected for each group, while for the metabolome analysis, 6 biological replicates were collected for each group. Following rapid freezing in liquid nitrogen, the collected samples were stored in an ultra-low-temperature freezer at −80 °C.

RNA extraction and transcriptome data collection

Total RNA were extracted from 24 samples using the MJZol total RNA extraction kit (Majorbio, China). The Illumina® Stranded mRNA Prep, Ligation kit was used to construct RNA-seq libraries, which were then sequenced on the DNBSEQ-T7 platform. The raw sequencing data were first filtered using the fastp software (https://github.com/OpenGene/fastp) to obtain high-quality sequencing data (clean data), thereby providing a basis for downstream analyses. The transcriptome data have been uploaded to the National Genomics Data Center (NGDC) under project number PRJCA061968.

Identification of DEGs and functional enrichment

The clean data of each sample was aligned to the reference genome arahy.Tifrunner.gnm2.J5K5 (https://data.legumeinfo.org/Arachis/hypogaea/genomes) using the HiSat2 software (http://ccb.jhu.edu/software/hisat2/index.shtml) to obtain mapped reads, which were used in subsequent transcript assembly, expression level calculation, and other analyses. DESeq2 (http://bioconductor.org/packages/stats/bioc/DESeq2) was then used to identify genes with differential expression between treatment and control group pairs, i.e. T24ZH6 vs C24ZH6, T48ZH6 vs C48ZH6, T72ZH6 vs C72ZH6, and T96ZH6 vs C96ZH6. The criteria for DEGs were: FDR < 0.05 and |log2FC|≥ 1. The Python scipy package [36] was used to perform KEGG PATHWAY enrichment analysis, and Fisher's exact test was adopted for relevant calculations. To effectively control the false positive rate during the calculation process, the Benjamini-Hochberg (BH) method was employed for multiple test correction to adjust the FDR. KEGG pathways with corrected p-values after adjustment less than 0.05 were defined to be significantly enriched.

Weighted gene co-expression network analysis

In the construction of the co-expression network, to reduce background noise, the gene expression matrix was first subjected to standardized preprocessing: low-expression and low-variation genes with an average expression level (TPM) < 1 and a coefficient of variation < 0.1 were excluded. To make the network conform to the scale-free characteristic, the minimum value of soft threshold (soft power β) that satisfying R² > 0.8 or reaching a plateau was selected. The dynamic tree cutting algorithm was used to divide co-expression modules, with the parameter settings as follows: minModuleSize = 30, minKMEtoStay = 0.3, and mergeCutHeight = 0.25. Finally, co-expression modules were obtained and marked with different colors, among which the gray module contained unclassified genes. Correlation analysis was performed between the obtained modules and the treatment (ZH6Tr) or control (ZH6Ck) groups. Genes within the most associated module with treatment that met the screening conditions (|MM|> 0.8 and |GS|> 0.5) were identified as hub genes [37].

Metabolites extraction and non-targeted metabolomics analysis

For each sample, 50 mg root was placed into a 2 mL centrifuge tube, and one grinding bead (6 mm in diameter) was added. Then, 400 μL of extraction solution [methanol:water = 4:1 (v:v)] containing an internal standard (L-2-chlorophenylalanine) at a concentration of 0.02 mg/mL was added into the tube. The sample solution was ground in a frozen tissue grinder for 6 min (−10 °C, 50 Hz), followed by low-temperature ultrasonic extraction for 30 min (5 °C, 40 kHz). The tube was then incubated at −20 °C for 30 min and centrifuged at 13,000 × g for 15 min at 4 °C. The supernatant (200 μL) was collected and subjected to ultra-high performance liquid chromatography-tandem mass spectrometry (UHPLC-MS/MS) analysis using the Thermo Scientific UHPLC-Q Exactive HF-X system (ultra-high performance liquid chromatography coupled with quadrupole-Orbitrap high-resolution mass spectrometry). Quality control (QC) samples were prepared by mixing equal volumes of metabolite extracts from all samples. During the instrumental analysis, one QC sample was inserted every 5–15 samples to evaluate the reproducibility of the entire analytical process.

Identification and analysis of differentially accumulated metabolites (DAMs)

After the LC–MS analysis was completed, the raw LC–MS data were imported into the metabolomics processing software Progenesis QI (Waters Corporation, Milford, USA) for baseline filtering, peak detection, integration, retention time correction, and peak alignment. A data matrix containing retention time, mass-to-charge ratio (m/z), and peak intensity was finally generated. Meanwhile, the MS and MS/MS spectral information were matched against public metabolomic databases [HMDB (http://www.hmdb.ca), Metlin (https://metlin.scripps.edu)] and the in-house database of the Majorbio company to obtain metabolite information.

Firstly, the data matrix was preprocessed: variables with a relative standard deviation (RSD) > 30% in quality control (QC) samples were removed, and the remaining data were subjected to log10 transformation to generate the final data matrix for subsequent analyses. Secondly, the ropls package (Version 1.6.2) in R language was used to perform principal component analysis (PCA) and orthogonal partial least squares discriminant analysis (OPLS-DA) on the preprocessed data matrix, and a 7-fold cross-validation was applied to evaluate the stability of the models. A metabolite was defined as a significantly DAMs when VIP > 1.0 and p < 0.05 [38]. The DAMs were annotated to metabolic pathways using the KEGG database (https://www.kegg.jp/kegg/pathway.html) to determine the pathways involved. Pathway enrichment analysis was conducted using the scipy.stats package as well.

Analysis of cis-acting elements

Cis-acting elements in the 2000 bp promoter region upstream of the start codon of candidate genes were predicted using Plant CARE database (https://bioinformatics.psb.ugent.be/webtools/plantcare/html). Subsequently, a figure representing the combination of the predicted elements was generated using the TBtools software [39].

qRT-PCR analysis

Reverse transcription and quantitative real-time PCR (qRT-PCR) experiments were conducted using the HiScript III 1 st Strand cDNA Synthesis Kit PASS (+ gDNA wiper, Cat: R312-02, Vazyme company) and the ChamQ SYBR Color qPCR Master Mix (Company name:Vazyme, cat. Q411), respectively. In the qRT-PCR experiment, actin was used as the reference gene (internal control gene) [40], and 3 biological replicates and 3 technical replicates were performed for each sample. The relative expression levels of genes were calculated using the 2 ^−ΔΔCt method [41]. Experimental data were graphed using GraphPad Prism 8. The primers and gene IDs were provided in Table S1.

Antibacterial activity of DAMs against R. solanacearum

The antibacterial activity of DAMs against R. solanacearum strain HA5-83 was performed as the previous study [42] with minor modification. R. solanacearum strain HA8-53 was retrieved from a −80°C freezer. A 30 μL of the bacterial suspension was inoculated into 3 mL of CPG medium and incubated with shaking at 28 °C and 200 rpm until the optical density (OD) value reached approximately 0.3 for bacterial activation. Prepare the metabolite as a stock solution at 100 mg/mL for later use. Take 5 μL of the metabolite stock solution and corresponding solvent were separately added to 4.95 mL of CPG medium. Additionally, 50 μL of the activated bacterial suspension was inoculated into each treatment group, and each treatment was performed in triplicate. The working concentration of metabolite is 100 mg/L. After 24 h of incubation, measure the OD₆₀₀ value and calculate the inhibition rate based on the OD₆₀₀ value. The inhibition rate was calculated using the formula:

The antimicrobial metabolites used for validation were high-purity reagents: 6-methylcoumarin (Aladdin, purity: 99%, CAS: 92–48-8), erucamide (J&A, purity: 90%, CAS: 112–84-5), and piceatannol (Aladdin, purity 98%, CAS: 10,083–24-6).

Results

Transcriptome profiling of peanut root under R. solanacearum infection

To reveal the molecular basis of peanut resistance to R. solanacearum, a total of 24 samples were collected from root at 24, 48, 72, and 96 hpi with R. solanacearum as well as sterile water control. Transcriptome libraries were constructed for each sample and sequenced to generate a total of 154.63 Gb of data (Table S2). The number of raw reads for each library ranged from 39.87 million to 47.10 million. After quality control, the clean reads for each library ranged from 39.64 million to 46.83 million. The average values of Q20, Q30 and GC contents were 98.7%, 95.7% and 44.63%, respectively. The Clean Reads of each sample were aligned to the designated reference genome, and the alignment rates ranged from 97.81% to 98.39% (Table S2). These results indicate that the data qualities are reliable and meet the requirements for subsequent analysis.

The results of the expression levels of 83,107 genes were obtained (Table S3). Among these genes, a total of 23, 407, 590, 153 DEGs (FDR < 0.05 and |log2FC|≥ 1) were identified between treatment and control groups at the 24, 48, 72 and 96 hpi (Figure S2) using DESeq2. At the 24 hpi stage, only 15 genes were up-regulated and 8 genes were down-regulated (Fig. 1A, Table S4), which was the fewest DEGs among the four stages. At the 48 hpi stage, 170 genes were up-regulated and 237 were down-regulated (Fig. 1A, Table S5). The 72 hpi stage had the most DEGs, with 413 up-regulated genes and 177 down-regulated genes (Fig. 1A, Table S6). At the 96 hpi stage, 134 genes up-regulated and 19 genes were down-regulated (Fig. 1A, Table S7). Venn analysis revealed that these DEGs related to 1,122 genes in total and only 46 of the 1,122 genes showed differential expression at two or more stages (Fig. 1B). For instance, the gene Ah0g001800 (ATP synthase family) exhibited differential expression from 24 to 72 hpi. Additionally, the genes Ah13g129000 (Cytochrome P450) and Ah19g347600 (Xylanase inhibitor) exhibited differential expression at 48, 72 and 96 hpi. In addition, there were eight condidate R genes in the DEGs dataset. The gene Ah02g036200 was only upregulated at 48 hpi, while the other seven genes were upregulated at 48 and 72 hpi (Fig. 1C). Especially, Ah16g532000, enocoding homolog of the RPW8 gene, showed the most significant upregulation, with a 337-fold increase in expression in T72ZH6 compared with C72ZH6 at 72 hpi (Fig. 1C, Table S6).

Fig. 1.

Analysis of DEGs under R. solanacearum infection. A The number of up-regulated and down-regulated DEGs in different comparison groups. B Venn analysis of DEGs between different comparison groups. C Heatmap of R genes expression levels. D KEGG enrichment analysis of DEGs in different comparison groups

KEGG enrichment analysis of the identified DEGs

To investigate the involvement of DEGs in biological pathways under R. solanacearum infection, KEGG enrichment analysis were conducted for DEGs at each stage. There was no pathway enriched for the 23 DEGs identified at 24 hpi (Table S8). The 407 DEGs identified at 48 hpi were mainly enriched in phenylpropanoid biosynthesis (p = 6.23 × 10⁻¹¹), MAPK signaling pathway—plant (p = 1.52 × 10⁻⁵), and plant hormone signal transduction (p = 4.54 × 10⁻³) (Fig. 1D, Table S9). In terms of 72 hpi, the 590 DEGs were mainly enriched in flavonoid biosynthesis (p = 4.20 × 10⁻³⁴), and tropane, piperidine and pyridine alkaloid biosynthesis (p = 2.43 × 10⁻³⁷) (Fig. 1D, Table S10). At 96 hpi, the 153 DEGs were mainly enriched in flavonoid biosynthesis (p = 3.78 × 10⁻⁹) and isoflavonoid biosynthesis (p = 3.16 × 10⁻¹²) (Fig. 1D, Table S11). Among them, flavonoid biosynthesis and isoflavonoid biosynthesis were enriched at all three stages (48, 72, and 96 hpi), indicating that they are continuously involved in the peanut resistance response to R. solanacearum invasion.

WGCNA and identification of hub genes

After filtering the low-expression and low-variations genes of the 89,519 genes identified through transcriptomics, 25,316 genes were obtained and used in WGCNA. Based on the results of scale-free topology fitting index and average connectivity analysis, β = 4 was selected to construct the co-expression network (Figure S3). A total of 29 co-expression modules were finally obtained using the dynamic tree cutting method, and modules with similar expression patterns were merged (Figure S4). Finally, a total of 29 modules were obtained, with the number of genes in each module ranging from 36 to 12,702 (Fig. 2A). Module-phenotype correlation analysis showed that the orange module exhibited the most significant positive correlation with the R. solanacearum treatment group (Tr) (R = 0.77, p = 1.0 × 10^–5) (Fig. 2A). KEGG enrichment analysis was performed on the 70 genes in orange module and found that these genes were mainly enriched in flavonoid biosynthesis (Fig. 2B).

Fig. 2.

Module identification and WGCNA. A Correlation between modules and specific phenotypes. The left column of numbers indicates the gene count in each module, while the right columns show the correlation coefficients (and significance p-values in parentheses) between modules and phenotypes. B KEGG enrichment analysis of genes in the ME Orange module. C Scatter plot of Module Membership (MM) vs Gene Significance (GS) for the orange module. Each dot represents a gene, with yellow dots indicating hub genes

A MM-GS scatter plot was generated for the 70 genes in the orange module, and 14 hub genes meeting the screening criteria were identified (Fig. 2C), all of which are DEGs. Detailed functional annotations of the 14 hub genes are provided in Table S12. To elucidate the potential regulatory mechanisms of these hub genes, a cis-acting element analysis was performed in the promoter (2,000 bp upstream) regions of these gene sequences (Fig. 3A). A total of 126 elements, representing 18 types, were identified, among which phytohormone-related elements were the most abundant with 67 in total. This suggests that phytohormones may play a crucial role in the regulatory network of these genes. As shown in the heatmap of expression levels (Fig. 3B), these hub genes exhibited upregulated expression at 48 hpi and 96 hpi comparing to control. For example, the expression level of Ah03g095600 increased by 5.2-fold at 48 hpi(Table S5) and 6.9-fold at 96 hpi(Table S7). After performing genetic interaction network analysis on the 70 genes within the orange module, three core regulatory nodes were identified, and all of them were hub genes. These include two P450 genes, Ah03g095600 (connectivity = 8.6) and Ah13g129000 (connectivity = 7.41), as well as the UDP-glucose dehydrogenase (UGDH) gene Ah11g339000 (connectivity = 8.1) (Fig. 3C).

Fig. 3.

Hub genes analysis. A Cis-acting elements of hub gene promoters. B Expression heatmap of hub genes. C Protein–protein interaction (PPI) network of the orange module genes

Identification of root DAMs under R. solanacearum infection

To investigate the metabolome changes after R. solanacearum infection, a total of 48 samples were collected from root at 24, 48, 72, and 96 hpi with R. solanacearum as well as sterile water control. Non-targeted metabolomics analysis was performed for each sample. A total of 3,718 annotated metabolites were identified in this study, and the expression levels of each metabolite in each sample are shown in Table S13. Partial least squares discriminant analysis (PLS-DA) showed that significant biochemical changes occurred between the treatment and the control groups at the four stages (Figure S5), indicating that variable importance in projection (VIP) analysis can be used to screen for DAMs [43].

A total of 1,614 DAMs (VIP > 1.0 and p < 0.05) were identified in the metabolome comparisons. The fewest DAMs (74 in total) were detected in the T24ZH6 vs C24ZH6 comparison, including 61 up-regulated DAMs and 13 down-regulated DAMs. In the T48ZH6 vs C48ZH6 comparison, 750 DAMs were identified, with 605 up-regulated DAMs and 145 down-regulated DAMs. For the T72ZH6 vs C72ZH6 comparison, 406 DAMs were found, consisting of 117 up-regulated DAMs and 289 down-regulated DAMs. The largest number of DAMs (769 in total) was observed in the T96ZH6 vs C96ZH6 comparison, including 422 up-regulated DAMs and 347 down-regulated DAMs (Fig. 4A, Table S14-S17).

Fig. 4.

Analysis of DAMs Under R. solanacearum Infection. A Number of up-regulated and down-regulated DAMs in different comparison groups. B Categories of DAMs compounds. C Venn analysis of DAMs in different comparison groups. D The expression abundance of DAMs. E KEGG enrichment analysis of DAMs at different comparison groups

Among all of the 1,614 DAMs, the main categories were carboxylic acids and derivatives, fatty acyls, and organooxygen compounds. Examples of these metabolites include L-phenylalanine, Ketoace, Jasmonic acid (JA), Arachidic acid, Rosavin, and Trans-p-coumaric acid (Fig. 4B). Moreover, the DAMs were compared across the four stages, revealing that eight DAMs were commonly identified in all of the four stages (Fig. 4C). Interestingly, four of these common DAMs were consistently up-regulated, whereas the remaining four were consistently down-regulated. For instance, after R. solanacearum infection, Linatine (a type of carboxylic acid and derivative) showed increased expression, whereas 2-Phenylethylamine (a type of benzenoid) showed decreased expression (Fig. 4D). Additionally, glyceollin II and glyceollin III, which belong to isoflavonoids, were upregulated in T96ZH6 compared with C96ZH6 at 96 hpi (Fig. 4D, Table S17). These compounds function as phytoalexins in soybean and contribute to defense against pathogen invasion [44]. Studies have shown that peanut stilbenoids are the key phytoalexins responsible for maintaining its disease resistance [45]. These stilbenoids are abundantly synthesized when peanuts are infected by pathogenic microorganisms [46].In this study, piceatannol, a DAMs of the stilbenoid class, was identified and also upregulated at 96 hpi (Fig. 4D, Table S17). Notably, among the DAMs identified in this study, 6-methylcoumarin was upregulated at 48 hpi (Fig. 4D, Table S15). Previous studies have reported that 6-methylcoumarin exhibits inhibitory activity against R. solanacearum growth [42]. Furthermore, studies have shown that erucamide can inhibit the growth activity of R. pseudosolanacearum in mulberry seedlings and delay the occurrence of bacterial wilt [47]. This metabolite was upregulated in T72ZH6 compared with C72ZH6 (Fig. 4D, Table S16) [47]. Conversely, N-acetylputrescine whose increase in content accelerated the growth of R. solanacearum in tomato xylem [48], was also identified as a DAM that was downregulated at 48 and 96 hpi (Fig. 4D, Table S15 and Table S17).

KEGG analysis of DAMs at different stages

To investigate the involvement of DAMs in metabolic pathways, KEGG analysis was performed on DAMs at different stage (Fig. 4E, Table S18-S21). At 24 hpi stage, DAMs were mainly enriched in pathways including glycerophospholipid metabolism (p = 6.71 × 10^–4) and autophagy–other pathways (p = 9.56 × 10^–3). At 48 hpi and 72 hpi stages, DAMs were primarily enriched in pathways including ABC transporters (p = 1.91 × 10^–3) and tryptophan metabolism (p = 7.76 × 10^–8). At 96 hpi stage, DAMs were mainly enriched in metabolic pathways such as nucleotide metabolism (p = 1.89 × 10^–5) and biosynthesis of cofactors (p = 6.27 × 10^–3). Additionally, it was found that pathways like ABC transporters and plant hormone signal transduction were enriched from 48 to 96 hpi stages.

Integrated KEGG enrichment analysis of transcriptomic and metabolomic after R. solanacearum infection

KEGG enrichment analysis was also performed on all identified DEGs and DAMs, which were co-enriched in seven pathways, including ABC transporters, phenylpropanoid biosynthesis, plant hormone signal transduction, tropane, piperidine and pyridine alkaloid biosynthesis, biosynthesis of various plant secondary metabolites, ascorbate and aldarate metabolism, and linoleic acid metabolism (Fig. 5, Table S22-S23), indicating that R. solanacearum infection reshapes the gene expression network and metabolite profile of Zhonghua 6.

Fig. 5.

Integrated analysis of transcriptome and metabolome. KEGG pathways commonly enriched by all DEGs and DAMs

To further investigate the regulatory network of peanut in response to R. solanacearum, a correlation analysis was performed between the top 100 DEGs and DAMs from each of the four comparison groups, using Pearson's algorithm. Where |R|> 0.8 and p < 0.05, a significant correlation was indicated between the DEGs and the DAMs. The strongest positive correlation (R = 0.999, p = 8.77 × 10⁻⁷) was identified between the upregulated gene Ah12g456200 (encoding a pathogenicity island component) at 24 hpi and the metabolite 5-(1-Adamantyl)−4H-1,2,4-Triazole-3-Thiol which differentially accumulated across all four time points (Figure S6, Table S24). In contrast, the strongest negative correlation (R = −0.996, p = 1.41 × 10^–5) was identified between the function unknown gene Ah05g092200 upregulated at 48 hpi and the metabolite D-Phe-L-Pro-L-Arg-(Chloromethyl) Ketone (Figure S7, Table S25). The upregulated gene Ah06g462900, which encodes ESL1 (ERD six-like 1), showed a significant positive correlation with dattelic acid (R = 0.878, p = 0.021) enriched in the phenylpropanoid biosynthesis pathway at 24 hpi (Figure S6, Table S24). The downregulated genes Ah03g298700 at 48 hpi, which encode Jasmonate ZIM-domain (JAZ) proteins, showed a significantly negative correlation with SA (R = −0.877, p = 0.021) enriched in the plant hormone signal transduction. The upregulated gene Ah06g022200 at 48 hpi encoding the MYC2 (MYC2 transcription factor) domain exhibited a significantly positive correlation with L-Threonine (R = 0.827, p = 0.041) and 2'-Deoxyuridine (R = 0.855, p = 0.029) enriched in the ABC transporter pathway (Figure S7, Table S25). At 72 hpi stage, the upregulated gene Ah06g035100 encoding barwin (PR2-like) showed a significantly positive correlation with erucamide (R = 0.855, p = 0.068) (Figure S8, Table S26). At 96 hpi stage, the upregulated gene Ah20g075800 encoding Pathogenesis-Related Protein 1 (PR1) exhibited a significantly negative correlation with the metabolite L-phenylalanine (R = −0.874, p = 0.022, which was co-enriched in both the ABC transporter and phenylpropanoid biosynthesis pathways) (Figure S9, Table S27).

Analysis of ABC transporters pathway

In the KEGG enrichment analysis of the DAMs, the ABC transporter pathway was significantly enriched at 48–96 hpi stages and exhibited a co-enrichment phenomenon in both transcriptomics and metabolomics analyses. This pathway is associated with 22 DEGs. These DEGs include 7 genes encoding ABCC2 (ATP-binding cassette subfamily C member 2), 6 genes encoding PDR (Pleiotropic Drug Resistance protein), 3 genes encoding ABCB1 (ATP-binding cassette subfamily B member 1), and others. For instance, the expression level of the gene Ah07g229100 encoding PDR in the T96ZH6 group was increased by 40 folds to that of the C96ZH6 group, the gene Ah06g162100 encoding ABCB1 in the T96ZH6 group was increased by 2 folds to that of the C96ZH6 group (Fig. 6A, Table S7). At the metabolomic level, 16 DAMs were identified in this pathway. The nucleoside class is the most abundant, accounting for 7 metabolites (e.g. deoxyadenosine, deoxyinosine, deoxycytidine), followed by amino acids and derivatives with 6 metabolites (e.g. L-aspartic acid, L-lysine, L-histidine) (Fig. 6B).

Fig. 6.

DEGs and DAMs analysis of ABC transporters. A Heatmap of DEGs expression levels in the ABC transporters pathway. B Pathway map (map02010) of the ABC transporters, where DEGs are represented in pink and DAMs in blue. The heatmap of expression abundances of DAMs in the ABC transporters pathway is included in the pathway map

Analysis of plant hormone signal transduction pathway

Plant hormone signal transduction pathway exhibited a co-enrichment phenomenon in both transcriptomics and metabolomics analyses, In addition, in the cis-acting element analysis of hub genes identified by WGCNA, phytohormone-related elements were the most abundant. KEGG enrichment analysis of DAMs that the plant hormone signal transduction pathway was significantly enriched from 48 to 96 hpi. This pathway involves 34 DEGs, including 6 genes encoding JAZ (Jasmonate ZIM-domain protein), 5 genes encoding MYC2, 5 genes encoding GH3 (Gretchen Hagen 3 protein), 3 genes encoding PR1 (Pathogenesis-related protein 1), and others. JAZ proteins exert transcriptional repressive functions in the JA signaling pathway [49]. Five genes encoding JAZ proteins (Ah15g050900, Ah14g003800, Ah01g284400, Ah05g369000, and Ah03g298700) were upregulated at 72 hpi (Fig. 7A). Compared with the control group, the expression level of the gene MYC2 (Ah03g271100) in the T72ZH6 group was increased by 3.7 folds to that of the C72ZH6 group (Fig. 7A, Table S6); the expression level of the gene PR1 (Ah10g048100) in the T96ZH6 group was increased by 19 folds to that of the C96ZH6 group (Fig. 7A, Table S7). At the metabolomic level, 6 DAMs were identified in this pathway, including auxin (IAA), SA, brassinolide (BR), JA, and others. Among them, SA was differentially downregulated at 24 hpi and differentially upregulated at both 48 hpi and 96 hpi, whereas JA was upregulated at 24–48 hpi but differentially downregulated at 72–96 hpi (Fig. 7B).

Fig. 7.

DEGs and DAMs analysis of plant hormone signal transduction. A Heatmap of DEGs expression levels in the plant hormone signal transduction pathway. B Pathway map (map04075) of the plant hormone signal transduction, where DEGs are represented in pink and DAMs in blue. The heatmap of expression abundances of DAMs in the plant hormone signal transduction pathway is included in the pathway map

Analysis of phenylpropanoid biosynthesis and flavonoid biosynthesis pathways

Phenylpropanoid biosynthesis was co-enriched in both transcriptomics and metabolomics analyses, and its downstream flavonoid biosynthesis pathway was significantly enriched in the transcriptomics at the 48–96 hpi stages. A total of 81 genes are enriched in phenylpropanoid biosynthesis and flavonoid biosynthesis. These DEGs included 2 genes encoding CYP73A (E1.14.1491, cinnamic acid 4-hydroxylase),3 genes encoding E2.1.1.104, 4 genes encoding HCT (E2.3.1.133, shikimate O-hydroxycinnamoyltransferase), 2 genes encoding 4CL (E6.21.12, 4-coumarate-CoA ligase), 23 genes encoding E1.11.1.7 (peroxidase), 4 genes encoding CHR (E23.1.170, Chalcone Reductase) and 34 genes encoding CHS (E2.3.1.74, Chalcone Synthase). For example, compared with the C96ZH6 control group, the expression levels of the genes encoding E2.1.1.104 (Ah15g473100 and Ah05g320700) in the T96ZH6 treatment group increased by 10 fold and 7 fold, respectively (Fig. 8A, Table S7). the expression level of the gene Ah02g388600 (4CL) in the T48ZH6 group was increased by 2 folds to that of the C48ZH6 group (Fig. 8B, Table S5); and the expression level of the gene CHR (Ah17g008300) in the T96ZH6 group was increased by 10 folds to that of the C96ZH6 group (Fig. 8C, Table S7). At the metabolome level, 9 DAMs are enriched in phenylpropanoid biosynthesis, including phenylalanine, ferulic acid, P-coumaroylquinic acid, sinapic acid, and others. Among them, ferulic acid and sinapic acid were upregulated at 24, 48 and 96 hpi (Fig. 8D). In addition, butin, a DAM of flavonoid biosynthetic pathway which can protect cells from hydrogen peroxide (H₂O₂)-induced apoptosis [50], were upregulated at 48 and 72 hpi during the response of Zhanghua 6 to R. solanacearum (Fig. 8D).

Fig. 8.

Analysis of phenylpropanoid biosynthesis and flavonoid biosynthesis. A DEGs shared by phenylpropanoid biosynthesis (map00940) and flavonoid biosynthesis (map00941). B DEGs exclusively involved in phenylpropanoid biosynthesis. C DEGs exclusively involved in flavonoid biosynthesis. The color of circles indicated the relative expession of genes in the form of heatmap. D The orange pathway represents the common pathway of phenylpropanoid biosynthesis and flavonoid biosynthesis, and the orange squares indicate the DEGs, and the orange dots represent the DAM. The blue pathway represents phenylpropanoid biosynthesis, where the blue squares denote the DEGs, and the blue dots represent the DAMs. The green pathway represents flavonoid biosynthesis, with the green squares indicating the DEGs, and the green dots represent the DAM. The color of squares indicated the relative contents of metabolites in the form of heatmap

qRT-PCR validation of DEGs and resistance verification of DAMs

To validate the accuracy of the transcriptome data, we randomly selected 8 DEGs for qRT-PCR validation (Fig. 9A). Among these genes, two were hub genes (Ah03g095600 and Ah11g339000); two were from plant hormone signal transduction (Ah02g044400 and Ah15g050900);two were from phenylpropanoid biosynthesis (Ah02g388600 and Ah05g320700); two from flavonoid biosynthesis (Ah15g473100 and Ah17g008300). The expression trends of these 8 DEGs were highly consistent with the expression patterns of the transcriptome data.

Fig. 9.

qRT-PCR validation and resistance verification of DAMs. A Comparison of expression levels of genes in RNA-seq (line chart) and qRT-PCR (bar chart) analysis. Relative expression levels (mean ± standard error) of genes before and after R. solanacearum infection. B Growth status of R. solanacearum HA8-53 in CPG medium supplemented with 100 mg/L,6-methoxycoumarin (solvent: DMSO), erucamide (solvent: chloroform), piceatannol (solvent: DMSO), or the corresponding solvents alone, with OD₆₀₀ measured after 24 h incubation. C Inhibition rates calculated based on the OD₆₀₀. The inhibition rates of 6-methoxycoumarin, erucamid, and piceatannol were 68.12%, 29.59%, and 36.49%, respectively. Note: For the preparation of erucamide stock solution, ultrasonic-assisted dissolution at 80 Hz should be employed to achieve adequate solubilization; given that erucamide has a relatively high freezing point (22–24 ℃), the temperature of the CPG medium must be maintained above its freezing point prior to the addition of the stock solution

In addition, in vitro inhibitory effects verifications were conducted for four identified DAMs, including 6-methylcoumarin (annotated based on in silico MS/MS spectral library), erucamide (annotated with authentic standard), piceatannol (annotated with authentic standard). As shown in Fig. 9B and C, at a concentration of 100 mg/L, the in vitro inhibitory effects of 6-methoxycoumarin, erucamide, and piceatannol against R. solanacearum strain HA8-53 at 24 h were 68.12%, 29.59%, and 36.49%, respectively. The results confirmed that 6-methoxycoumarin, erucamide, and piceatannol could inhibit the proliferation of R. solanacearum strain HA8-53 isolated from peanut.

Discussion

Transcriptomic studies on peanut resistance to BW

Research on the resistance mechanism of bacterial wilt is of great significance, however, there are still obvious shortcomings in the relevant studies to date. Transcriptome technology has been applied to dissect the progression of disease-resistant responses in diverse crop species. A commonly adopted approach is to use the 0-h time point as the control, which compared samples from different post-inoculation stages with pre-inoculation samples. For instance, in tobacco, transcriptomic datasets generated from susceptible cultivar C048 and resistant cultivar C244 at 3, 9, 24, 48, and 72 hpi were compared with those at 0 hpi. A total of 14,858 DEGs (7266 upregulated and 7592 downregulated) were identified in C244, while 13,751 DEGs (6899 upregulated and 6852 downregulated) were detected in C048 [51]. However, considering that genes in plants undergo alterations during normal growth, this study established more stringent controls. Peanut samples inoculated with R. solanacearum at 24, 48, 72, and 96 h were directly compared with the corresponding stage samples treated with water in the same period. Based on these precise comparisons, only 1,122 DEGs were identified across the four distinct post-inoculation stages, which reduced DEGs caused by natural growth of the plants as well as treatments. Among them, eight candidate R genes were activated by the R. solanacearum infection (Fig. 1C).

In this study, to further identify key genes associated with peanut BW resistance, we performed WGCNA based on transcriptome sequencing data, identifying a total of 14 hub genes, including 3 genes with high connectivity. Two of the three genes, namely Ah03g095600 and Ah13g129000, were annotated as members of the CYP81E subfamily based on KO annotation, homologous sequence alignment and phylogenetic analysis (Figure S9). The protein sequences encoded by Ah03g095600 and Ah13g129000 showed 42.51 to 45.93% similarities to CYP81E genes in other spcieces, including the CYP81E1 whose function has been experimentally validated in Licorice (Glycyrrhiza echinata L.) (Figure S10) [52]. As a cytochrome P450 subfamily, CYP81E plays a key hydroxylation role in the biosynthesis of the glyceollin, an isoflavonoid phytoalexin [44]. It catalyzes the conversion of the precursor daidzein into 2’-hydroxydaidzein, which provides the essential substrate for the subsequent reduction reaction mediated by isoflavone reductase, thus acting as a important step that initiates the glyceollin biosynthetic pathway [44, 53, 54]. Notably, Ah13g129000 was consistently upregulated at all three infection stages (48, 72, and 96 hpi), suggesting that it may continuously contribute to the biosynthesis of the phytoalexin glyceollin during pathogen infection.

Two UGDH genes, Ah11g339000 and Ah01g430600, were identified among the 14 hub genes, with Ah11g339000 exhibiting high connectivity (connectivity = 8.1). UGDH is involved in synthesizing hemicellulose and pectin precursors found in newly formed plant cell walls [55, 56]. It serves as a key rate-limiting enzyme that catalyzes the conversion of UDP-glucose to UDP-GlcUA (UDP-glucuronic acid) [57]. UDP-GlcUA plays an essential role in plants, as approximately half of the monosaccharides constituting matrix polysaccharides—which are critical for maintaining cell wall integrity—are derived from UDP-GlcUA metabolites [58]. Therefore, these two UGDH genes may involve in the resistance response against R. solanacearum infection by reinforcing the cell wall.

Furthermore, genes in the phenylpropane and flavonoid biosynthesis pathways were also activated. CYP73A hydroxylates cinnamic acid to form p-coumaric acid, and 4CL subsequently catalyzes p-coumaric acid to produce p-coumaroyl-CoA, which provides substrates for the flavonoid biosynthesis pathway. For example, in this study, the biosynthesis of the stilbene compound piceatannol (KEGG: map00945) and the flavonoid compound butin (KEGG: map00941) were involved. In addition, PR-1 and its upstream regulatory TGA transcription factors were significantly upregulated, which are core components of the SA-mediated systemic acquired resistance (SAR) pathway [59, 60]. Moreover, several ABC transporter genes (ABCG2, ABCB1, and PDR) also showed significantly upregulated expression. These genes may be induced by SA [61] and are responsible for delivering defensive substances such as antimicrobial metabolites and PR proteins. In summary, these upregulated genes may form a multi-level immune response through synergistic effects, including signal transduction activation and antimicrobial substance synthesis, which collectively mediate the resistance of Zhonghua 6 to R. solanacearum.

Metabolomic studies on peanut resistance to BW

To date, metabolomic research on resistance to BW has not been reported in peanut. Plants infected by pathogens produce a large number of secondary metabolites, among which some may be resistant metabolites that can damage the structure of pathogens and inhibit their growth and reproduction [62]. This study identified 1,614 differential metabolites through metabolomic analysis, 6-methylcoumarin, erucamide and piceatannol might be involved in the inhibition of R. solanacearum.

Coumarins are a class of secondary metabolites synthesized through the phenylpropane metabolic pathway, with a benzopyran-2-one ring as their core structure. Studies have shown that infection of plants by various pathogens can induce the accumulation of coumarins [63]. 6-methoxycoumarin causes abnormal cell elongation and interferes with normal cell division. Pot and field experiments demonstrated that the application of 6-methoxycoumarin effectively suppresses the development of BW in tobacco and significantly reduces the population of R. solanacearum in the stems [42]. The 6-methylcoumarin achieved an in vitro inhibitory effects of 76.79% against R. solanacearum at a concentration of 100 mg/L [42]. In this study, the in vitro inhibitory effects of 6-methylcoumarin against R. solanacearum strain HA8-53 isolated from peanut was validated to be 68.12% at a concentration of 100 mg/L.

In studies on mulberry seedlings, erucamide demonstrated inhibitory effects on R. pseudosolanacearum growth activity by modulating cell morphology and extracellular polysaccharide levels. At a concentration of 1500 mg/L, it effectively delayed the onset of bacterial wilt in mulberry seedlings [47]. The inhibitory effect of erucamide on activity R. solanacearum may be achieved by disrupting its T3SS, i.e. key virulence apparatus responsible for injecting effector proteins into host cells and is vital for the pathogenicity of many Gram-negative bacteria, including R. solanacearum [64]. Studies have shown that erucamide is a broad-spectrum anti-virulence and phytoalexin that targets the HrcC protein, a key component of the bacterial T3SS. This targeting disrupts the membrane localization of the pathogen and the assembly of the injectisome, preventing the secretion of effector proteins and thereby inhibiting pathogenicity [65]. In this study, the in vitro inhibitory effects of erucamide against R. solanacearum strain HA8-53 isolated from peanut was validated to be 29.59% at a concentration of 100 mg/L.

At present, extensive studies have been conducted on the role of stilbenoids in combating bacteria and fungi [46, 66, 67]. Among stilbenoids, resveratrol (3,4’,5-trihydroxystilbene) has been most extensively studied; it not only has numerous research applications in the field of antibacterial activity but also involves fields such as food science and medicine [68–70]. In peanut, resveratrol has been found to inhibit the growth of Aspergillus flavus and reduce the production of aflatoxins [71, 72]. In this study, piceatannol (3,3',4',5-tetrahydroxystilbene), a stilbenoid upregulated at 96 hpi, was identified. It is a hydroxylated analog of resveratrol, which can be converted into piceatannol by cytochrome P450 enzymes [73]. Additionally, studies have shown that piceatannol is superior to resveratrol in terms of antioxidant activity [74]. Some studies have determined that the minimum inhibitory concentration (MIC) of piceatannol against Staphylococcus aureus is 64–128 mg/L [75]. However, it remains unclear whether piceatannol can inhibit the growth of R. solanacearum. In this study, the in vitro inhibitory effects of piceatannol against R. solanacearum strain HA8-53 isolated from peanut was validated to be 36.49% at a concentration of 100 mg/L. The experimental results illustrated that piceatannol may contribute to peanut defense against R. solanacearum infection, and enriched current understanding of stilbene compounds against R. solanacearum.

Based on the findings, it is verified that the compounds 6-methylcoumarin, erucamide and piceatannol possess the potential to inhibit R. solanacearum in peanuts. Although the precise in vivo concentrations of these metabolites in peanut tissues remain to be further quantified, the significant inhibitory effects observed in vitro at least indicate their intrinsic antibacterial activity, which supports their potential roles in plant defense. This discovery provides an important theoretical basis for the development of bio-based antibacterial agents, and holds particular application potential for the prevention and control of soil-borne BW.

Integrated transcriptomic and metabolomic analyses reveal plant pathways conferring resistance to BW

Previous studies have reported that the temporal relationship between transcripts and metabolites is not always synchronous [76]. In this study, we also found asynchrony between the identified DEGs and DAMs. For example, at 48 hpi, the number of upregulated DEGs (170) was lower than that of downregulated DEGs (237), but the number of upregulated DAMs (605) in the same period was higher than that of significantly downregulated DAMs (145); a reverse trend was also observed at 72 hpi. This phenomenon is speculate to arise from the complex regulatory mechanisms, such as negative correlations between certain genes and metabolites, or the inherent time-lag effect from gene expression to metabolite synthesis [77].

Most pathways associated with BW resistance have been identified solely through transcriptomic analyses. For instance, Chen et al. [78] reported that the resistance pathways in tomatoes against BW include plant-pathogen interaction, plant hormone signal transduction, and MAPK signaling pathways. Regarding tobacco resistance-related pathways, consistent identification of the phenylpropanoid biosynthesis pathway has been achieved across different studies using diverse materials and methodologies [79–81]. However, inconsistent pathways have also been documented, such as glutathione metabolism [79] and hormone signal transduction pathways [81]. Recently, integrated transcriptomic and metabolomic analysis has been conduct in eggplant to reveal its resistance to BW, and the plant hormone signal transduction pathway was identified as a key mediator [82]. However, integrated transcriptomic and metabolomic analysis has not been conducted in other crops including peanut.

In this study, the integrated transcriptomics and metabolomics analyses were used for the first time to investigate the resistance mechanism of a highly resistant germplasm Zhonghua 6 to R. solanacearum infection. A hypothesis could be proposed according to the integrated transcriptomics and metabolomics analyses (Fig. 10). After plants of Zhonghua 6 were incolulated with R. solanacearum, The MAPK signaling and plant hormone signal transduction pathways were activated. Phytohormones, especially SA and JA, play an important role in responding to pathogen infection [14, 83, 84]. The antagonistic interaction between SA and JA pathways is the important mechanism for plants to regulate different defense responses and optimize resource allocation [14, 85]. During the infection of Zhonghua 6 peanut by R. solanacearum, JA content was upregulated in the early stage of infection (24–48 hpi), but downregulated in the middle and late stages (72–96 hpi); however, SA content was downregulated at 24 hpi, and then upregulated at 48 hpi and 96 hpi, reflecting the antagonistic effect between SA and JA pathways. Previous studies have shown that activation of the SA signaling pathway can inhibit the expression of JA [86]. SA plays a central role in regulating the plant immune system [87, 88]. Its activation at the middle and late stages may be the key phytohormone underlying the resistance of Zhonghua 6. Its increases upregulated the expression of TGA transcription factors which directly bind to the PR-1 gene promoter to drive its expression (KEGG: map04075). Upon these signal transductions, the expression of numerous genes (such as CHS, CHR, 4CL and CYP81E in the phenylpropanoid biosynthesis, flavonoid biosynthesis and isoflavonoid biosynthesis pathways) were upregulated, driving the accumulation of antimicrobial metabolites (e.g. erucamide, 6-methylcoumarin, piceatannol, glyceollin, and others) and inhibiting the biosynthesis of metabolites that were required for bacterial nutrition (e.g. N-acetylputrescine). Previous research also found that SA regulated the phenylpropanoid biosynthesis pathway in Vigna mungo response to Fusarium oxysporum infection [89]. ABC transporters (such as the ABCG2, ABCB1, and highly expressed PDR) might function as a logistical transport system and may be responsible for translocating these PR proteins and antimicrobial metabolites to their sites of action [90], thereby forming a systemic defense barrier to inhibit the spread of R. solanacearum (Fig. 10).

Fig. 10.

The regulatory network of Zhonghua 6 in response to R. solanacearum infection

Conclusions

Based on transcriptomic analysis, a total of 1,122 DEGs were identified in this study, and 14 hub genes were screened out. These genes are speculated to be closely associated with peanut bacterial wilt resistance, providing a crucial molecular basis for subsequent molecular breeding of bacterial wilt-resistant peanut varieties and in-depth research. Meanwhile, in the metabolomic analysis, 1,614 DAMs were identified, among which 3 metabolites, namely 6-methylcoumarin, erucamide, and piceatannol, were confirmed to inhibit the growth of Ralstonia solanacearum. The integrated transcriptomics and metabolomics analyses revealed a potential immune regulatory network which provides an important research basis for systematically clarifying the resistance mechanism of peanuts against Ralstonia solanacearum infection. Further studies should be conducted to reveal specific functions of the identified key DEGs and DAMs.

Authors’ contributions

PY contributed to the experimental design and data analysis as the major contributor of this study, and drafted the original manuscript. HYL provided numerous valuable suggestions and performed the major revisions on the original draft. QY played a key role in the linguistic polishing of the manuscript. All authors read and approved the final version of the manuscript.

Funding

This work was supported by the National Natural Science Foundation of China (Grant No. 32171997, 32161143006 and 32341039), the Natural Science Foundation of Hubei Province (2024AFA098), the Earmarked Fund for China Agriculture Research System (CARS-13), the Nanfan special project of CAAS (Grant No. YBXM2552), the Central Public-interest Scientific Institution Basal Research Fund (Grant No. Y2025YC112), and the Agricultural Science and Technology Innovation Program of the Chinese Academy of Agricultural Sciences (Grant No. CAAS-ASTIP-2021-OCRI).

Data availability

All data generated or analysed during this study are included in this published article and its supplementary materials. Transcriptomic and metabolomic analyses were conducted using the Majorbio Cloud Platform (www.majorbio.com) which was previously reported [91].

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.