Elucidating the metabolic landscape of Bacillus velezensis WY-65 in basic and cocktail nutrient media: a biocontrol approach against peanut web blotch

Abstract

Background

In this study, the cocktail medium significantly enhanced antifungal metabolite production, thereby improving the suppression of peanut web blotch caused by Peyronellaea arachidicola.

Results

A novel antagonistic bacterial strain was isolated from cultivated peanut soil and identified as Bacillus velezensis WY-65. Additionally, the metabolomic profile of strain WY-65 was investigated under two fermentation conditions: basic fermentation medium (BFM) and cocktail fermentation medium (CFM). In vitro assays revealed that the cell-free supernatant (CFS) from WY-65 cultured in CFM inhibited P. arachidicola by 87.68%, whereas the CFS from BFM inhibited P. arachidicola by 61.42%. Greenhouse trials targeting peanut web blotch demonstrated that CFM achieved a disease control efficiency of 92%, whereas BFM exhibited a control efficiency of 86%. Treatment with CFM reduced the disease incidence to 8.71%, and BFM treatment resulted in a disease incidence of 15%. In contrast, the control group presented a disease incidence of 95%. CFS significantly impacted the efflux pump and caused ROS in P. arachidicola. Using untargeted LC‒MS/UHPLC-based metabolomics, we identified distinct metabolic signatures influenced by medium composition, revealing an enrichment of bioactive secondary metabolites, including 3-indoleacrylic acid and L-tryptophan, in CFM. KEGG differential abundance (DA) analysis revealed that strain WY-65 cultured in CFM underwent broad metabolic reprogramming, with significant enrichment in amino acid biosynthesis and the phenylalanine, tyrosine, and tryptophan synthesis pathways. Pathway topology analysis further revealed that glycine, serine, and threonine metabolism and phenylalanine, tyrosine, and tryptophan synthesis pathways had the greatest impact and significance, highlighting their central roles in stress adaptation and antimicrobial precursor biosynthesis. In vitro confirmatory tests with synthetic chemicals confirmed the bioactivity of these compounds. At 200 μL, both 3-indoleacrylic acid and L-tryptophan exhibited peak antifungal activity (91% and 89%, respectively), whereas their activities at 150 μL (87% vs. 75%) and 300 μL (89% vs. 80%) were comparatively lower.

Conclusion

Our findings provide insight into medium-induced metabolomic modulation in strain WY-65 and highlight its potential as a robust biocontrol agent for sustainable peanut web blot disease management.

Clinical trial

Not applicable.

Graphical Abstract

Supplementary Information

The online version contains supplementary material available at 10.1186/s13036-026-00640-1.

Introduction

The global population is projected to double by 2050. Addressing the pressing need to meet the demand for agricultural production to sustainably feed the growing population is crucial [1, 2]. Food security is expected to continue to be a significant global concern in the near future [3]. Fungal diseases can have a devastating impact on crop production, resulting in a significant reduction in crop yield and deterioration of food quality [4, 5]). Oilseed crops such as peanuts are widely cultivated for their edible oil and to provide food for people worldwide [6]. The peanut crop is affected primarily by foliar diseases, such as peanut web blotch [7, 8] and peanut leaf spot [9]. Peyronellaea arachidicola is a causal agent of web blotch disease and significantly impacts peanut crops [10, 11]. Peanut web blotch poses a rising threat to peanut farming globally and is considered a significant fungal disease in China [8, 12]. Chemical fungicides are frequently employed to manage plant fungal diseases. However, chemical fungicides have a negative impact on ecosystems and biodiversity and can also pose serious risks to human health [13, 14]. Using microbes for biocontrol is a more effective approach to managing diseases [15–17]. In China, various biocontrol products are available. However, there are no biocontrol products available for peanut web blotch (http://www.chinapesticide.org.cn/ (accessed on 1 January 2021)).

Bacillus species, particularly B. velezensis, have been extensively studied for their ability to manage crop diseases [18, 19]. This genus can produce broad-spectrum bioactive secondary metabolites, particularly nonribosomal enzymatic complexes responsible for the biosynthesis of nonribosomal peptides (NRPs) and polyketides (PKs) [20, 21]). A recent study conducted by Ye et al. [22] reported that strain ZN-S10 produced 7-O-succinyl prolactin A, telocinobufagin, and surfactin secondary metabolites, which could be antifungal agents. Similarly, Armenov et al. [23] reported that antifungal secondary metabolites were found in strain R22 via UHPLC-Q-TOF LC‒MS/MS analysis of the cell-free supernatants of strain R22. B. velezensis LHSB1 mediates the biocontrol of peanut stem rot caused by Sclerotium rolfsii [24]. Ding et al. [10] identified gene clusters of B. velezensis VR-34 against P. arachidicola to control peanut web blotch. The synthesis of Bacillus-based biopesticides for sustainable crop production is not easy, as the process faces high costs and low productivity challenges [25–27]). A new nutrient medium formulation for submerged fermentation is urgently needed to save cost and increase yield while maintaining quality [28, 29].

In addition to physical parameters (pH, temperature, time, volume of flask, size of inoculum, and speed of shaker), nutrient media significantly impact the microbial fermentation process [30]. Specifically, the metabolic profile and bioactivity are directly impacted by slight variations in the nutrient medium used during fermentation [31]. According to a recent study by Nabila et al. [32], the metabolite profile of B. velezensis is significantly influenced by different culture media. Therefore, considering that fermentation is a crucial step, carbon and nitrogen sources are needed to establish a balanced formulation for high-quality products. In this context, the introduction of a new nutrient medium with improved nutritional value could harbor the chemical environment of cells in submerged fermentation to produce high-quality microbial products [33, 34]. To better understand these metabolic shifts, metabolomics enables the comprehensive profiling and identification of metabolites within a system at a specific point in time [35]. Metabolomic applications provide frameworks for systematic screening, comparative analysis, and validation of metabolites derived from bacterial fermentation [36–38].

In contrast to the widely used basic fermentation media (yeast extract, tryptone, NaCl) [10] for Bacillus cultivation, this study introduces a nutritionally optimized cocktail fermentation medium by supplementing diverse carbon and nitrogen sources to increase antifungal compound production. While B. velezensis is well recognized for its ability to promote plant growth and biocontrol, its metabolic reprogramming under different fermentation regimes remains poorly understood. Here, we applied LC–MS-based untargeted metabolomics to reveal metabolic shifts in B. velezensis WY-65 grown in basic versus cocktail media and demonstrated the superior antifungal efficacy of medium-optimized fermentation against P. arachidicola through both in vitro and in planta assays.

Materials and methods

Microorganisms and growth conditions

The pathogenic fungal strain P. arachidicola was obtained from the Institute of Plant Protection, Liaoning Academy of Agricultural Sciences, Shenyang. The antagonistic bacterial strain was isolated from cultivated peanut soil. A total of 10 soil samples were collected in sterile polythene bags, each containing 5 g of soil. A serial dilution was made to obtain a pure culture of the strain. One gram of soil was added to 9 mL of distilled water in a test tube, and a series of dilutions was prepared as follows: 10⁻², 10⁻³, and 10⁻⁴.

Then, 1 mL of the suspension was transferred into the next 3 test tubes, and at the final transfer, 1 mL from the last dilution of the soil sample was transferred by pipette and poured onto LB agar media (10 g of tryptone, 5 g of yeast extract, 5 g of NaCl, and 15 g of agar/L), and on nutrient agar media (10 g of peptide, 10 g of beef extract, 8 g of NaCl, and 15 g of agar/L) separately. The mixture was subsequently incubated for 7 days at 28 °C.

Finally, a colony was isolated and purified. The 6 purified strains were screened against the pathogen P. arachidicola. On the basis of its strong antifungal activity, the most potent antagonistic strain was selected. The pure antagonistic strain WY-65 was cultivated for 3 days at 28 °C in the dark on LB agar medium. Pathogenic P. arachidicola fungi were cultured for 14 days at 28 °C in the dark on potato dextrose agar (PDA). Additionally, an inoculum suspension of P. arachidicola was prepared in potato dextrose broth (PDB), which consisted of 10 g potato extract and 10 g dextrose per 1000 mL of distilled water. For suspension preparation, 40 mL of PDB was dispensed into each 200 mL flask, and a 5 mm diameter agar plug containing actively growing P. arachidicola mycelium was inoculated into each flask. The cultures were incubated at 28 °C in the dark for 14 days [10].

Molecular and morphological identification of strains

After DNA extraction with a commercial kit from Sangon Biotech, the 16S rRNA, rpoB, and gyrA genes were amplified via primers, and the polymerase chain reaction (PCR) conditions are detailed in Supplementary Table S1. The annealing temperature conditions for 16S rRNA were as follows: initial temperature 95 °C for 5 min; 30 cycles at 94 °C for 30 s, 57 °C for 30 s, and 72 °C for 90 s; and final extension at 72 °C for 10 min. Similarly, for the gyrA gene, the initial temperature conditions started at 94 °C for 5 min; 35 cycles at 94 °C for 30 s, 55 °C for 30 s, and 72 °C for 45 s; and a final extension at 72 °C for 10 min. For the rpoB gene, the initial temperature conditions were as follows: 94 °C for 3 min; 30 cycles at 94 °C for 1 min, 56 °C for 1 min, and 72 °C for 2 min; and a final extension at 72 °C for 10 min. The resulting sequences were compared to GenBank reference sequences via the National Center for Biotechnology Information (NCBI) Basic Local Alignment Search Tool (BLAST). The sequencing data were then used to construct a phylogenetic tree via the maximum likelihood (ML) method in MEGA v.11 software. The nucleotide sequence was deposited in GenBank under the accession number PV770029. For morphological observation, strain WY-65 was grown under defined conditions. The samples were fixed with 2.5% glutaraldehyde. A dehydration procedure was performed using a sequence of ethanol solutions (30%, 50%, 70%, 90%, and 100%) to eliminate water and avoid structural collapse. Dehydrated samples were coated with gold. To observe the morphology of strain WY-65, a scanning electron microscope (SEM, Hitachi, Japan) was used [39]. Strain WY-65 was streaked on LB media and incubated at 27 °C for 24 hours. Colony morphology was observed, and physiological and biochemical identification methods were performed according to standard protocols, as previously described [40].

Fermentation process

Basic fermentation media

The biphasic submerged fermentation procedure was executed following the method described by Wang et al. [9]. First, to synthesize the spore suspension, a 250 mL Erlenmeyer flask containing 40 mL of Luria–Bertani (LB) medium was inoculated with two 5-mm spore cakes of the WY-65 strain. Consequently, the basic fermentation medium was prepared with exact quantities of 7 g/L yeast extract, 10 g/L tryptone, and 2.5 g/L NaCl. The pH was carefully adjusted from 6.6 to 6.8 by adding 0.5 M HCl and 1.0 M NaOH solution. Afterward, the fermentation medium was subjected to sterilization via an autoclave. To initiate the submerged fermentation process, a 5 mL aliquot of spore suspension, maintaining a concentration of 1 × 10⁹ CFU/mL, was poured into each flask containing previously sterilized fermentation nutrient medium. After inoculation, the fermentation medium was incubated in a rotary shaker operating at 30 °C with shaking at 165 rpm for 96 hours. The fermentation broth of strain WY-65 obtained from both BFM and CFM cultures was first centrifuged at 10,000×g rpm for 15 min at 4 °C to remove cells and debris.

Cocktail fermentation medium

The same biphasic submerged fermentation procedure described above was executed. A novel cocktail fermentation medium (CFM) was formulated in this study. The composition was as follows: 4 g/L semolina flour, 4 g/L soybean meal, 4 g/L corn syrup, 4 g/L peanut root powder, 12.5 g/L beef extract, and 0.5 g/L MgSO₄. The initial pH of the medium was adjusted to 6.6–6.8 prior to sterilization. The peanut root extract was prepared by washing fresh peanut roots thoroughly with distilled water, drying them in an oven at 50 °C, and grinding them into a fine powder. To initiate the submerged fermentation process, a 5 mL aliquot of spore suspension of strain WY65, maintaining a concentration of 1 × 10⁹ CFU/mL, was poured into each flask containing sterilized CFM. After inoculation, the fermentation medium was incubated in a rotary shaker operating at 30 °C with shaking at 165 rpm for 96 hours. These nutrient media were selected on the basis of antifungal activity, as described in a previous report [41]. The fermentation broth of strain WY-65 obtained from both BFM and CFM cultures was first centrifuged at 10,000×g rpm for 15 min at 4 °C to remove cells and debris.

In vitro bioassays of strain WY-65 against P. arachidicola

The antifungal activities of CFM and BFM were assessed via the poisoned food method described by Grover and Moore (1962), with modifications. The CFSs derived from CFM and BFM were incorporated into molten potato dextrose agar (PDA) cooled to 40 °C. The medium was mixed thoroughly, dispensed into 90 mm sterile Petri plates at 20 mL per plate, and then allowed to solidify. The control plates contained PDA without the test agent. A 5 mm mycelial disc from the actively growing margin of a 7-day-old pathogen culture on PDA was placed at the center of each plate. The plates were incubated at 28 °C for 14 days or until the control plates exhibited nearly complete fungal growth. The radial growth of fungal colonies was measured along two perpendicular axes, and the mean colony diameter was recorded. The percentage inhibition of mycelial growth was calculated as follows:

The percentage inhibition of mycelial growth was calculated via the following formula:

% Inhibition = DC−DT/DC × 100, where DC is the mean colony diameter in control plates and DT is the mean colony diameter in treated plates. Each treatment was performed in triplicate.

Controlled environment study on plant disease control

In a controlled greenhouse environment, two peanut seeds were sown in each of four 4-inch plastic pots containing nutrient-rich, sterilized soil. There were three pots for each treatment. The plants were grown under a temperature regime of 30 °C during the day and 24 °C at night, with adequate sunlight and humidity to support optimal growth. After 30 days of growth, when the plants reached the five-leaf stage (comprising approximately 20 leaflets), inoculation treatments were applied by soaking the leaves for 5 minutes. For the positive control, a solution was prepared by dissolving 100 mL of commercial tebuconazole fungicide in 20 mL of water. The following treatments were administered:

• A)Pa: Inoculated with 10 mL of P. arachidicola spore suspension
• B)Pa+CFM: Inoculated with 10 ml of P. arachidicola and soaked in 10 ml of CMF broth
• C)Pa+BFM: Inoculated with 10 ml of P. arachidicola and soaked with 10 ml of BFM broth
• D)Pa+CF: Inoculated with 10 ml of P. arachidicola and soaked with a commercial fungicide (tebuconazole)

The following formula calculates disease incidence and disease control efficiency [9, 39].

Action of the WY-65 strain against P. arachidicola

SEM analysis of mycelia and spores

The effects of CFS derived from CFM and BFM on the morphology of P. arachidicola hyphae were assessed by collecting mycelial samples from peripheral regions. Both treated and untreated control mycelia were excised and fixed in 2% glutaraldehyde at 4 °C for 4 hours. After fixation, the samples were rinsed with phosphate-buffered saline (PBS, 0.5 M, pH 7.0), subjected to secondary fixation with 1% aqueous osmium tetroxide, and dehydrated through a graded ethanol series. The samples were then immersed in isoamyl acetate overnight, subjected to critical point drying with carbon dioxide, and coated with a thin layer of gold via a sputter coater. Hyphal structural alterations were examined by scanning electron microscopy (SEM) [42].

Efflux pump analysis

To assess the effects of CFM and BFM on the efflux pump activity of P. arachidicola, a Rh6G efflux assay was employed. Briefly, pathogen cells were incubated in potato dextrose broth (PDB) overnight. Then, in the later stages, the cells were harvested. The samples were subsequently washed with glucose-free PBS and adjusted to 10⁷ CFU/mL. Later, a solution of Rh6G was added to the cell suspension at a final concentration of 10 μM, and the suspension was incubated at 35 °C for 50 minutes. At the same time, an ice water bath was used for 10 minutes to expose the sample. After harvesting, the cells were washed with glucose-free PBS. Glucose/PBS (5%) was added to resuspend the cells, and CFM and BFM were added to different samples. The fluorescence intensity was measured every 60 min for 4 h every 60 min with excitation at 488 nm and emission at 530 nm by flow cytometry [43].

ROS analysis

To determine the effects of CFM and CBM on the level of ROS in the pathogen. After being treated with both CFSs, the cells of the pathogen were processed with the probe 20,70-dichlorodihydrofluorescein diacetate (H2DCF-DA) at a final concentration of 3 µM. The probe 20,70-dichlorodihydrofluorescein (H2DCF) is produced when the esterase in the cell cleaves the acetate groups of H2DCF-DA. Later, the oxidation of H2DCF was caused by intracellular ROS. Later, it yielded the fluorescent product 20,70-dichlorofluorescein (DCF). The measurement of fluorescence was performed via a fluorescence microplate reader at excitation and emission wavelengths of 485 and 535 nm, respectively. The method was followed from [44] with modification.

Extraction of metabolites

The cell-free supernatant of strain WY-65 was used for metabolite analysis. For each condition, a 100 μL aliquot of the clarified broth was transferred into a 2 mL Eppendorf tube and subjected to metabolite extraction. Later, an extraction solvent with a volume of 400 μL of methanol:acetonitrile (1:1, v/v) was added, which was labeled an isotopic internal standard mixture. Then, homogenization of the samples was carried out at 35 Hz for 4 minutes, and the samples were placed in a water bath for 5 minutes for the sonication process. For an efficient extraction process, the above mentioned whole cyclic process was repeated three times. The samples were incubated for 1 h at 40 °C, followed by centrifugation at 12,000 rpm (RCF = 13800 ×g; rotor radius = 8.6 cm) for 15 min at 4 °C. The obtained supernatant was placed with certain care into clean glass vials for further analysis. To prepare a quality control (QC) sample, equal volumes of supernatants from each sample were mixed. In this investigation, four quality control (QC) samples and six experimental samples were examined.

LC‒MS/MS analysis

A Vanquish ultrahigh-performance liquid chromatography (UHPLC) system (Thermo Fisher Scientific) connected to an Orbitrap Exploris 120 mass spectrometer (Thermo Fisher Scientific) was used to conduct liquid chromatography‒tandem mass spectrometry (LC‒MS/MS) investigations. For the chromatographic separation process, a Waters BEH amide column (2.1 × 50 mm, 1.7 μm particle size) was used. The mobile phase was set up with two solutions: (A) 25 mmol L⁻¹ ammonium hydroxide and 25 mmol L⁻¹ ammonium acetate in water at pH 9.75 and (B) acetonitrile. The injection volume was set at 2 μL, and the autosampler was maintained at 4 °C. In “Xcalibur software (Thermo Fisher Scientific),”, which continually monitors the full-scan mass spectrum to initiate MS/MS scans, data acquisition was carried out in “information-dependent acquisition (IDA) mode.” The gas flow rate was set up in the following mode: the auxiliary gas flow rate was 15 Arb, whereas the sheath gas flow rate was 50 arbitrary units (Arb). The temperature of the capillary was set at 350 °C. These were the electrospray ionization (ESI) source characteristics. The full-scan MS resolution was 60,000, and the MS/MS resolution was 30,000. In stepped mode, normalized collision energies (NCE) of 20, 30, and 40 were applied. In positive ion mode, the spray voltage was set to + 3 kV, whereas in negative ion mode, it was set to −3 kV.

Statistical analysis of differentially abundant metabolites and pathway annotation

A total of 24,329 peaks remained after denoising on the basis of the relative standard deviation from the initial 37,491 peaks that were found. Half of the lowest detected value was used to impute missing data. All the data were normalized according to the internal standard.

The processed dataset, including peak number, sample identity, and normalized peak areas, was imported into SIMCA 16.0.2 software (Sartorius Stedim Data Analytics AB, Umeå, Sweden) for multivariate statistical analysis. To reduce noise and account for variable variance, the data were scaled and log-transformed prior to analysis. Principal component analysis (PCA), an unsupervised dimensionality reduction technique, was employed to visualize the sample distribution and clustering. A 95% confidence interval in the PCA score plot was used to identify potential outliers.

To further distinguish metabolic differences between groups, supervised orthogonal projections to latent structures-discriminant analysis (OPLS-DA) was performed. Model robustness and predictive power were assessed through 7-fold cross-validation, yielding R² (explained variance) and Q² (predictive accuracy) values. To validate the OPLS-DA model, a permutation test was conducted with 200 iterations. Intercept values for R² and Q² were obtained to evaluate the risk of overfitting, with a lower Q² intercept indicating greater model reliability.

Variable importance in projection (VIP) scores from the first principal component of the OPLS-DA model were calculated. Metabolites with VIP > 1.0 and p values < 0.05 (Student’s t test) were considered significantly altered. These differentially abundant metabolites were subjected to metabolic pathway enrichment analysis via the Kyoto Encyclopedia of Genes and Genomes (KEGG; http://www.genome.jp/kegg) and MetaboAnalyst (http://www.metaboanalyst.ca). Univariate receiver operating characteristic (ROC) curve analysis was also performed, and the area under the curve (AUC) was calculated to evaluate the diagnostic potential of key metabolites.

Confirmation of the compounds

To test the antagonism of the main compounds synthesized by the WY-65 strain, we selected only the commercially available compounds from the VIP compounds from the metabolomic analysis, i.e., 3-indoleacrylic acid (CAS no. 1204–06-4) and L-tryptophan (CAS no. 73–22-3), purchased from Sigma Aldrich. A food poisoning method was adopted to determine the bioactivity of these compounds. The stock solution was prepared at a concentration of 10 mg/mL by dissolving 1 g of each compound in 100 mL of sterile distilled water. For bioactivity in 20 mL of molten PDA medium, upon cooling, 150 μL, 200 μL, or 300 μL of each compound was added, and for both compounds, the same method was used. This mixture was subsequently inoculated with 5 mL of a spore suspension (1 × 10^×7 spores/mL) of P. arachidichola. The plates were incubated at 28 °C for 14 days or until the control plates exhibited nearly complete fungal growth. The radial growth of fungal colonies was measured along two perpendicular axes, and the mean colony diameter was recorded. The percentage inhibition of mycelial growth was calculated as follows:

The percentage inhibition of mycelial growth was calculated via the following formula:

% Inhibition = DC−DT/DC × 100

where DC is the mean colony diameter in the control plates and DT is the mean colony diameter in the treated plates. Each treatment was performed in triplicate.

Statistical analysis

All the statistical analyses were performed via GraphPad Prism v10.5. In vitro data are presented as the means ± standard deviations (SDs). Comparisons between BFM and CFM treatments were conducted via an unpaired two-tailed Student’s t test. A p value < 0.05 was considered statistically significant. For disease incidence and control, repeated-measures ANOVA with the Greenhouse–Geisser correction and one-way ANOVA with the Brown–Forsythe test were applied. All the data represent the means ± SEs of three replicates. The ROS data were analyzed via one-way ANOVA, followed by Tukey’s multiple comparison test. Differences were considered statistically significant at p < 0.05. The efflux pump data were processed by nonlinear regression with a biphasic decay model. The goodness of fit was determined by the residual error and R2. For the compound confirmation test, two-way ANOVA was performed, and significance was defined as p < 0.05. The data are the means of three replicates. All the experiments were conducted in triplicate.

Results

Molecular and morphological identification of the WY-65 strain

The colony morphology results revealed a whitish appearance of the WY-65 strain on LB agar media (Fig. 1A). Similarly, SEM revealed that strain WY-65 exhibited a typical rod shape, resembling those of other strains of B. velezensis (Fig. 1B). The phylogenetic tree based on the 16S rRNA (Fig. 1C), rpoB, and gyrA genes (Figure S1a, b) indicated that strain WY-65 belongs to B. velezensis. The 16S rRNA, rpoB, and gyrA genes of strain WY-65 were amplified and sequenced, resulting in amplicons of 1492 bp, 750 bp, and 1100 bp, respectively. The biomarkers of these genes are displayed in Figure S2.

Fig. 1.

Taxonomic identification of strain WY-65. (a) Morphological characteristics of colonies of strain WY-65 on LB medium. (b) SEM micrograph of strain WY-65. (c) Maximum likelihood phylogenetic tree of strain WY-65 on the basis of 16S rRNA sequences

The WY-65 isolate grew in media containing 2 to 10% sodium chloride at temperatures ranging from 5 to 45 °C, demonstrating substantial environmental adaptability. The isolate exhibited positive results for starch hydrolysis, gelatin liquefaction, and esculin hydrolysis, indicating the secretion of extracellular enzymes involved in nutrient cycling. The strain also exhibited oxidase and catalase activity, fermented glucose under aerobic conditions, and reduced nitrate to nitrite but did not generate hydrogen sulfide (Table S2).

In vitro bioassays of strain WY-65 against P. arachidicola

In vitro experiments revealed that strain WY-65 exhibited potent antifungal activity against P. arachidicola. An illustration of the negative control with other treatments is shown in Fig. 2A. In the negative control, P. arachidicola grew strongly, while it was strongly inhibited in the treatment groups. The results revealed that the CFM fermentation broth from strain WY-65 had an 87.68 mm inhibition zone against P. arachidicola, whereas that of the BFM fermentation broth was 61.42 mm (Fig. 2B).

Fig. 2.

In vitro bioassay of cell-free supernatant (CFS) from strain WY-65 in basic (BFM) and cocktail (CFM) fermentation media against P. arachidicola (PA). (A) An illustration of the negative control (CK: PA only) shown in comparison with other CFS treatments (BFM, CFM). (B) Comparison of the inhibition zones (%) between the BFM and CFM treatments. The bars represent the means ± SDs, with individual data points shown (n = 3 per group). Statistical analysis was performed via an unpaired two-tailed Student’s t test

Controlled environment study on plant disease control

Greenhouse experiments demonstrated that strain WY-65 was highly effective in managing peanut web blotch disease. Figure S1 visually compares the negative control group with the treated groups. The results indicate a high presence of web blotch where the peanut plants were only inoculated with PA and no treatment was given, whereas fewer disease symptoms were observed in the peanut plants inoculated with PA and then treated with CFS (CFM, BFM). In terms of disease incidence, the negative control group (CK) presented the highest incidence of 95%. However, treatment with CFM significantly reduced the incidence to 8.71%. Similarly, BFM decreased the percentage to 15%, as presented in Fig. 3A. Treatment had a highly significant effect on disease control (F(3,8) = 1553, p < 0.0001, R² = 0.9983). The untreated control (CK) group presented no disease control (0%), whereas all the CFS (CFM+PA, BFM+PA) and CF+PA treatment groups presented high levels of disease control, ranging from approximately 86% to 92%. Specifically, CFM achieved a disease control efficiency of 92%, whereas BFM displayed a slightly lower control efficiency of 86%, as shown in Fig. 3B. The positive control formulation CF also strongly suppressed disease, with a control efficiency of 90%.

Fig. 3.

Effect of treatments on disease incidence and control. (a) Disease incidence of peanut web blotch under CFS (BFM, CFM) and commercial fungicide (CF); CK inoculated only with PA. Repeated-measures ANOVA with greenhouse–geisser correction revealed a significant treatment effect (F(1.18, 2.35) = 623.5, p = 0.0006). The bars represent the means ± standard errors (SEs) of three replicates. Individual differences were not significant (F(2, 6) = 0.119, p = 0.8899). (b) Disease control (%) of peanut web blotch by CFS (BFM, CFM) and commercial fungicide (CF); CK inoculated only with PA. One-way ANOVA confirmed a significant treatment effect (F(3, 8) = 1553, p < 0.0001, R² = 0.9983), with equal variances supported by the brown–forsythe test (p = 0.5323). The bars represent the means ± standard errors (SEs) of three replicates. All the experiments were performed in triplicate

Action of the WY-65 strain against P. arachidicola

SEM analysis of mycelia and spores

SEM analysis revealed morphological alterations in the ultrastructures of P. arachidicola. The hyphae appeared ruptured and fragmented with a roughened surface. Furthermore, spore production was significantly reduced due to the action of WY-65. CFM and WY-65. BFM (Fig. 2C). Conversely, in the control group (CK), the mycelia appeared smooth and regular in shape, and abundant spores were produced (Fig. 4A).

Fig. 4.

Action mechanism of CFM and BFM. (a) SEM images of PAs treated with CFS (BFM, CFM) and untreated CK. (b) Effects of CFM and BFM on the efflux pump of P. arachidicola. (c) Effects of CFM and BFM on the levels of ROS in P. arachidicola. The data are expressed as the means ± SEs (n = 5). The asterisks indicate significant differences among the treatments (one-way ANOVA, p < 0.05)

Efflux analysis

The results revealed that the fluorescence intensity decreased actively with time in CK, CFM, and BFM-treated fungal cells. This revealed active efflux of the fluorescent probe. However, compared with untreated CK cells, CFM- and BFM-treated cells consistently presented greater intracellular fluorescence at all time points. The greater residual fluorescence in CFM-treated cells revealed a decrease in efflux activity or increased intracellular retention of the dye, indicating a potential inhibitory effect of CFM on fungal efflux mechanisms (Fig. 4B).

ROS analysis

The results revealed that, in treated pathogenic cells, DCF fluorescence was significantly reduced. The results of the present study revealed that CFM induced ROS in pathogen cells, indicating a strong influence on the metabolites produced. Similarly, BFM-treated fungi presented less significant ROS generation in response to the pathogen. However, CK, without any treatment, presented a lower accumulation of ROS (Fig. 4C).

Metabolomic analysis

Multivariate analysis of metabolites

The score plot obtained from orthogonal projections to latent structures (OPLS) analysis effectively separates the profiles of WY-65. CFM and WY-65. BFM (Fig. 5A). In this figure, t½1 P represents the predicted principal component score of the first principal component, whereas t½1° corresponds to the orthogonal principal component. The purple squares depict WY-65. For the CFM fermentation broth group, yellow dots represent WY-65. BFM group, demonstrating a clear distinction between them. elling’s T² test indicates that all samples in both groups are within a 95% confidence interval. The OPLS-DA model was validated through permutation tests (n = 200) to determine the R² and Q² values of the random models, confirming the model’s reliability.

Fig. 5.

(A) OPLS score plot model of CFM vs BFM (R2X = 0.833). (B) Permutation test results of the OPLS-DA model of CFM vs BFM

The permutation test results for the OPLS-DA model are illustrated in Fig. 5B. As shown in the figure, the R²Y value for the original model is 1, which is very close to 1, indicating that the model accurately represents the actual sample data. The Q² value is 0.998, which is also nearly equal to 1, suggesting that the model would likely maintain similar distribution patterns if new samples were introduced. Additionally, the permuted R² values on the left side of the intercept are consistently lower than the original R² value on the right side, and the same trend is observed for all the Q² values in the permuted data. These findings demonstrate that the original model is robust and not overfit.

Univariate analysis of differentially abundant metabolites

The identified metabolites were categorized and quantified according to their chemical class and attribution data. In HMDB, the secondary categorization of these diverse metabolites is represented by a three-tiered proportional chart. Different color blocks in the donut plots denote distinct categories, while the proportions illustrate the extent to which each classification contributes to the overall set of identified metabolites. The analysis revealed that organic acids and their derivatives had the highest abundance, accounting for 29.55% of the total metabolites (Fig. 6A). A comprehensive comparison of WY-65. CFM and WY-65. BFM revealed 475 (red highlighted Table S3) differentially expressed metabolites (DEMs) in positive ion mode and 80 DEMs in negative ion mode, including MS1 data (blue highlighted Table S3).

Fig. 6.

(A) Donut plot classification chart for all metabolites between the CFM and BFM groups. (B) Volcano plot visually displaying the DEMs between the CFM and BFM groups. *p < 0.05, **p < 0.01, and ***p < 0.001. p values were calculated via two-sample t tests

Metabolomic profiling via OPLS-DA and variable importance in projection (VIP) scoring revealed a distinct upregulation of several low-molecular-weight compounds in B. velezensis WY-65 under cocktail fermentation media. Notably, 3-indoleacrylic acid and L-tryptophan were among the top discriminating metabolites, each with known or putative antimicrobial properties. The elevated levels of pantothenic acid and L-tryptophan suggest a shift toward enhanced precursor biosynthesis, potentially priming the bacterium for secondary metabolite production even in the absence of classical NRPS/PKS-derived lipopeptides in the detected range.

In contrast, several metabolites were markedly downregulated, indicating strategic metabolic reallocation. Sugars such as 1-kestose and maltotetraose were reduced, suggesting diminished carbon flux toward stored carbohydrates and a shift away from extracellular polysaccharide accumulation. Antioxidant and defense-related compounds, including 1,2,3-trihydroxybenzene, benzene-1,2,4-triol, and capsoside A, also presented decreased abundance, implying a decrease in oxidative stress responses under stabilized fermentation conditions. Furthermore, the suppression of arginyl-valine, N-ethylacetamide, and pyruvic acid reflects a potential decline in amino acid catabolism and central carbon metabolism, which is consistent with a rerouting of metabolic resources toward biosynthetic activities. Reduced levels of specialized metabolites such as lansiumamide A and dide-O-methylsimmonides may also indicate decreased mimicry of plant-derived chemical defenses or the uptake of exogenous compounds. Collectively, these changes highlight coordinated metabolic reprogramming in B. velezensis WY-65, which favors secondary metabolite biosynthesis while suppressing energetically costly or nonessential pathways (Fig. 6B). A higher VIP score for metabolites indicates greater potential for discrimination. The volcano diagram encompasses all the metabolites measured in this study, with each point representing a peak. The criteria for inclusion in the diagram are p < 0.05 and a VIP > 1. The abscissa denotes the comparison between the CFM and BFM controls. The log_2-fold change in the expression of the substance is represented by the abscissa, whereas the -log₁₀p value from Student’s t test is represented by the ordinate. The metabolites that have undergone significant upregulation are visually represented in red, indicating their heightened expression levels. The figure displays metabolites that have undergone significant downregulation in blue, whereas metabolites that have not exhibited significant differences are depicted in gray (Fig. 6B).

Quantitative correlation of metabolites

To investigate the metabolic interplay and coregulation among the intracellular metabolites of WY-65 cultured in BFM and optimized CFM media, a comprehensive correlation analysis was performed. The resulting p value matrix revealed several statistically significant pairwise associations among the 100 detected metabolites. A substantial number of metabolite pairs demonstrated highly significant correlations (p < 0.05), indicating coordinated metabolic responses under varying nutritional conditions. Notably, metabolites such as maltol, 3’-sialyllactose, dihydroxyacetone (dimer), and arginylvaline strongly covaried with multiple other metabolites, suggesting their central roles in media-specific metabolic reprogramming. These compounds may serve as key biomarkers reflecting nutrient-dependent shifts in primary and secondary metabolism. The presence of distinct clusters of significant correlations further implies functional associations within specific metabolic modules, such as amino acid biosynthesis, sugar metabolism, and energy-related intermediates. These patterns highlight the adaptive ability of an organism to tailor its metabolic network in response to optimized nutrient formulations. Overall, the correlation matrix offers valuable insights into the dynamic metabolic architecture of B. velezensis WY-65, revealing the influence of medium composition on metabolic coordination and pathway interactions. These findings support the potential for media optimization strategies to selectively modulate desired metabolic outputs in biotechnological applications (Figure S4 and Table S4).

KEGG pathway annotation analysis

Upon annotation of 788 small molecules (SMs) to KEGG pathways via the KEGG database (https://www.kegg.jp/kegg/kegg1.htm), 788 SMs were annotated to KEGG pathways for a total of 250 times. The annotation times varied between one and twelve times per metabolite. After eliminating duplicate KEGG pathways, a total of 88 KEGG pathways remained (Supplementary Table S5).

KEGG pathway classification and differential abundance score analysis

Using Fisher’s exact test and the Kyoto Encyclopedia of Genes and Genomes (KEGG) database, we performed pathway analysis to determine which metabolic pathways were affected, as shown in Fig. 7A. The KEGG pathway differential abundance score revealed substantial differences in metabolite profiles between CFM and BFM. The “metabolic pathways” category was the most abundant pathway, reflecting broad upregulation of core cellular metabolism. In particular, pathways such as “Biosynthesis of amino acids”, “Aminoacyl-tRNA biosynthesis”, “D-Amino acid metabolism”, and “ABC transporters” were significantly enriched, indicating increased amino acid synthesis, protein biosynthesis, and nutrient transport under optimized conditions. According to the results, the phenylalanine, tyrosine and tryptophan synthesis pathways were significantly enriched, indicating the synthesis of 3-indoleacrylic acid and L-tryptophan.

Fig. 7.

(A) KEGG pathway enrichment analysis revealed substantial differences in metabolite profiles between the CFM and BFM conditions. (B) Pathway topology analysis of significantly altered metabolites between the CFM and BFM conditions. Bubble size indicates pathway impact; color intensity corresponds to statistical significance (–log₁₀P)

Pathway topology analysis

Pathway topology analysis of the significantly altered metabolites revealed several high-impact biochemical routes modulated in B. velezensis WY-65 under CFM fermentation media. As shown in the bubble plot (Fig. 7B), the glycine, serine, and threonine metabolism pathways presented the most significant alterations, with the highest -log_₁₀ (p value) and maximal pathway impact scores. This suggests substantial reprogramming of amino acid biosynthesis, potentially linked to stress adaptation and precursor provisioning for antimicrobial metabolite synthesis.

Similarly, the phenylalanine, tyrosine and tryptophan synthesis pathways also impact the synthesis of 3-indoleacrylic acid and L-tryptophan. Other notably impacted pathways include the following: beta-alanine metabolism, which is strongly perturbed, likely reflecting alterations in pantothenic acid and coenzyme A biosynthesis, which are essential for fatty acid and polyketide production; alanine, aspartate, and glutamate metabolism, which is indicative of nitrogen flux adjustments and possibly involved in the biosynthesis of various precursors for secondary metabolism; and nicotinate and nicotinamide metabolism, which reflects NAD⁺/NADH cofactor adjustments under fermentation-driven redox pressure. Riboflavin metabolism is possibly involved in modulating the cellular energy status and oxidative stress resistance, although with a lower overall impact. The bubble size correlates with the pathway impact (from 0 to 1), whereas the color gradient (pale to red) indicates statistical significance (–log_₁₀p). The dominant clustering of bubbles toward the top-right quadrant demonstrated that both statistical significance and pathway centrality were highest in amino acid-associated networks.

Metabolic pathway and network analysis

Metabolite-centric KEGG network analysis revealed extensive remodeling of metabolic pathways in B. velezensis WY-65 grown in CFM compared with those in BFM. The main compounds 3-indoleacrylic acid (C12283) and L-tryptophan (C00078) were upregulated through the phenylalanine, tyrosine and tryptophan synthesis pathways. An L-phenylalanine carboxylyase (oxidative-deam) reaction was involved. Chorismate is first converted to anthranilate by anthranilate synthase (EC4.1.3.27), followed by phosphoribosylation to N-(5’-phosphoribosyl)-anthranilate (EC 2.4.2.18), isomerization (EC 5.3.1.24), and cyclization to indole-3-glycerol phosphate (EC 4.1.1.48). This intermediate is cleaved by the tryptophan synthase α-subunit (EC 4.1.2.8) to indole, which is finally condensed with serine by the β-subunit (EC 4.2.1.20) to yield L-tryptophan. In addition to serving as a proteinogenic amino acid, L-tryptophan also contributes to indole alkaloid biosynthesis, where its transamination to indole-3-pyruvate and subsequent decarboxylation produce 3-indoleacrylic acid (C12283), linking primary metabolism to indole-derived secondary metabolites.

Similarly, other compounds, such as pyruvate (C00022), glutamine (C00212), and acetyl-CoA (C00130), emerged as central hubs, reflecting upregulated flux through glycolysis, the TCA cycle, and nitrogen assimilation routes. Enzymes such as aldehyde dehydrogenase (EC 1.2.1.8) and uridine phosphorylase (EC 2.4.2.1) further increase amino acid and nucleotide turnover. The enrichment of pathways related to pyrimidine metabolism (bsu00240) and glycine-serine-threonine metabolism (bsu00260) support the hypothesis that optimized media facilitates a metabolically versatile and biosynthetically active phenotype (Figure S5).

ROC analysis of metabolites

The ROC curve analysis of the upregulated metabolites revealed a suite of compounds with perfect discriminatory power (AUC = 1.0; CI = 1–1), including 1,5-anhydrosorbitol, 3-indoleacrylic acid, D-pantothenic acid, L-tryptophan, and glycyrrhizin. These results highlight the consistent and significant upregulation of these metabolites in strain WY-65 cultured under optimized cocktail conditions (Table S6). The elevation of indole derivatives, amino acids, and cofactor-related metabolites suggests enhanced biosynthetic capacity, stress adaptation, and possibly increased secondary metabolite production.

These metabolites serve as robust biomarkers of media-induced metabolic activation and reflect the profound reprogramming of B. velezensis metabolism under nutrient-enriched conditions. To identify key metabolic markers distinguishing strain WY-65 grown in basic media from those grown in optimized cocktail media, ROC curve analysis was performed on the significantly downregulated metabolites. Remarkably, several compounds—including 1-kestose, arginyl-valine, capsoside A, and pyruvic acid—exhibited perfect classification performance, with AUC values of 1.0 and 95% confidence intervals of 1–1. This finding indicates a highly consistent and statistically robust downregulation pattern across all biological replicates (Table S6).

Confirmation of the compounds

Confirmatory tests demonstrated that both 3-indoleacrylic acid and L-tryptophan exhibited strong antifungal activity against the pathogen. At a dose of 200 μL, 3-indoleacrylic acid resulted in a significant inhibition zone with 91% activity. It also exhibited substantial antifungal activity at 100 μL and 300 μL doses, with 87% and 89% inhibition, respectively. Similarly, L-tryptophan showed high antifungal activity at a 200 μL dose, with 89% inhibition, whereas the 100 μL and 300 μL doses resulted in 75% and 80% inhibition, respectively (Fig. 8a). Figure 8b provides a visual depiction of this antifungal activity. These results confirm that 3-indoleacrylic acid and L-tryptophan are the active compounds responsible for the observed bioactivity.

Fig. 8.

(A) Bioactive confirmatory test of VIP compounds against P. arachidicola by IA (3-indoleacrylic acid) and L-tr (L-tryptophan). All the values are the means of three replicates. Mean values for each row level under the IA and L-tr treatments. Two-way ANOVA detected significant main effects of treatment and concentration factors and a significant interaction between them (p < 0.0001 for all). Points represent individual measurements; bars indicate the means ± SE. (B) Antifungal results of the bioactive confirmatory test of VIP compounds against P. arachidicola by (A) IA and L-tr. In Ck, there was no treatment involving only P. arachidicola

Discussion

The morphological and molecular results indicated that strain WY-65 belongs to the B. Velezensis family. The physiological and biochemical characteristics of strain WY-65 closely align with those reported in a previous study [40]. According to the in vitro results, the WY-65 strain showed potent antagonistic activity against P. arachidichola. A greenhouse study revealed that strain WY-65 has the ability to control peanut web blotch disease. As previously reported, B. velezensis is a potential biocontrol agent against crop fungal pathogens [45, 46]. Compared with BFM, CFM positively impacted strain WY-65 and increased its bioactivity. The bioactivity of microbial antagonists, such as B. velezensis, is directly impacted by nutrient composition during fermentation [47, 48]. Similarly, nutrient media also impact metabolic pathways [49].

SEM revealed that strain WY-65 altered the morphology of the pathogen. The hyphae and spore shapes appeared to be deformed. As reported by Ahsan et al. [50], microbial antagonists can alter and deform the morphology of fungal pathogens. The efflux pump results revealed a decrease in fluorescence intensity due to the effects of CFM, and BFM gradually decreased over time. These findings suggest that both CFS treatments attenuate both phases, supporting their role in modulating efflux pumps. A similar mechanism was reported previously by Lu et al. [43] in Candida albicans. The results of the ROS assay revealed that CFM may cause a ROS burst in pathogenic cells. These results support the antifungal mechanism by which CFM inhibits pathogens. CFM is assumed to cause a xenobiotic effect, which leads to the death of pathogenic fungal cells. As previously reported by Blenkey et al. [51], fungicidal compounds cause excessive ROS in fungal cells, which leads to the death of these cells.

According to the overall profile of the metabolites, most secondary metabolites produced by strain WY-65 are organic acids. The enrichment of 3-indoleacrylic acid and L-tryptophan, which are molecules with antimicrobial or signaling functions, indicates that the bacterium may adopt a competitive, dominant and substrate-interacting biochemical profile. Recently, Wang et al. [52] reported that 3-indoleacrylic acid is an antifungal compound. Similarly, L-tryptophan is an antifungal low-molecular-weight precursor compound that can act as an antifungal agent [53].

The increased levels of pantothenic acid and L-tryptophan indicate a significant shift in the biosynthetic process, enhancing the production of key precursors for downstream metabolites. This intriguing development occurs even in the notable absence of lipopeptides typically associated with NRPS/PKS systems, highlighting a fascinating divergence in metabolic pathways. This is probably indicative of either alternative biosynthetic pathways or posttranslational control over canonical secondary metabolism under these conditions [54, 55]. Moreover, the significant downregulation of metabolites (1-kestose, maltotetraose, pyruvic acid, etc.) suggests a metabolic energy-saving strategy. A decrease in oligosaccharides and central metabolic intermediates may lead to carbon flux to other priorities, such as specialized metabolite synthesis, rather than storage and general energy metabolism [56]. The reduction in the levels of antioxidant and defense-related molecules, such as 1,2,3-trihydroxybenzene (Pyrogallol) [57] and benzene-1,2,4-triol [58], could indicate reduced oxidative stress or conscious regulation of nonessential defense pathways. Moreover, reduced levels of alkaloids and amides, such as lansiumamide A and dide-O-methylsimmonidin, might indicate reduced uptake or synthesis of plant-like chemical defenses, which could reduce redundancy in metabolism [59].

Correlation analysis of the intracellular metabolites in B. velezensis WY-65 in basic media with those in optimized cocktail medium revealed prominent metabolic network remodeling in response to nutrient medium composition. The observation of a number of statistically significant correlations (p < 0.05) also implies that concentration levels are not regulated individually but rather as coordinated modules, reflecting generalized changes in biosynthetic and degradation activity [60]. KEGG enrichment analysis revealed that the cocktail fermentation medium plays a major role in regulating microbial metabolism, mainly through increasing the expression of genes involved in amino acid biosynthesis and related pathways. The most enriched term, “Biosynthesis of amino acids,” emphasized the requirement for anabolic functions to provide conducive conditions to increase protein synthesis and biomass production. Specifically, the phenylalanine, tyrosine and tryptophan biosynthesis pathways are directly linked for the synthesis of 3-indoleacrylic acid and L-tryptophan.

This finding is in agreement with earlier reports of increased amino acid flux in high-emt-value microbial systems [61]. The enrichment of the “Aminoacyl-tRNA biosynthesis” pathway indicates increased translation ability, which may be required to meet the increased demand of protein synthesis during fermentation [62]. Accurate charging of tRNAs by amino acids is required for fidelity and speed of translation, leading to increased productivity of microorganisms.

ABC transporters were also enriched, suggesting increased nutrient acquisition and metabolite efflux. These transporters are essential for cellular homeostasis, particularly under stress and/or high productivity. The upregulation of ABC transporter genes has also been documented in other studies aimed at improving the microbial production of secondary metabolites [63]. Moreover, the enrichment of pathways such as “D-Amino acid metabolism” and “2-Oxocarboxylic acid metabolism” mirrors more global changes in nitrogen and carbon flow. Presumably, these prophylactic changes allow optimization of primary metabolism for energy generation in a scenario where rapid growth or increased secondary metabolite biosynthesis is needed. Overall, metabolic reprogramming under optimized conditions reflects a change to more efficient biosynthetic and nutrient transport processes, which support enhanced fermentation performance. These are useful targets for future metabolic engineering for further increasing yield and robustness of the process [61].

Pathway topology analysis revealed the results from pathway impact analysis through a bubble plot. Glysine, serine, and threonine metabolism and phenylalanine, tyrosine and tryptophan biosynthesis were the pathways with the greatest influence, with the highest impact scores and statistical significance. These amino acids are involved in carbon metabolism, are a major source of methyl groups, and are precursors of various secondary metabolites. The activation of this pathway might be interpreted as increased metabolic flux into an essential biosynthetic route that is essential for both growth and bioactivity [64].

Other affected and notable pathways included beta-alanine metabolism; alanine, aspartate and glutamate metabolism; and nicotinate and nicotinamide metabolism. This can be correlated with adaptations in redox balance, amino acid availability, and co-factor biosynthesis needed for optimal secondary metabolite production. Moreover, the downregulation of riboflavin metabolism may suggest that gene resource reallocation to other biosynthetic routes may be a priority following the fermentation medium applied. A similar pattern of pathway impact under culture conditions was described [65] in a study of the metabolic profile of Lactobacillus casei Zhang.

Comparative KEGG-based metabolic network analysis of B. velezensis WY-65 in BFM and CFM revealed significant differences in central metabolic pathways, providing evidence of metabolic flexibility to adapt to resource availability. The elevated levels of L-tryptophan (C00078) and 3-indoleacrylic acid (C12283) in strain WY-65 in cocktail fermentation medium (CFM) indicate increased flux through the shikimate and tryptophan biosynthetic pathways. Recently, Liang et al. [66] reported that these two compounds, L-tryptophan (C00078) and 3-indoleacrylic acid, are produced via the phenylalanine, tyrosine and tryptophan biosynthesis pathways. Pyruvate, acetyl-CoA glutamate, glutamine and aspartate emerged as potential key metabolites that act as hubs in global networks, indicating significant shifts in carbon, nitrogen and amino acid metabolism. The increased metabolic flux through glycolysis and the TCA cycle was further reflected by the accumulation of pyruvate (C00022) and its generation of acetyl-CoA (C00130) in the optimized cocktail medium. This finding suggests increased energy production and biosynthetic activity, a metabolic signature of bacterial growth in nutrient-rich or nutrient-engineered environments [67]. In addition, the position of glutamate (C00213) and glutamine (C00212), which are central to nitrogen assimilation and transamination reactions, implies nitrogen-dependent metabolic reorganization in response to optimized media [68]. The participation of uridine phosphorylase (EC 2.4.2.1) and thymidine phosphorylase (EC 2.4.2.2), together with enriched pyrimidine and purine metabolism pathways (bsu00240, bsu00250), is indicative of enhanced nucleotide turnover and DNA precursor production. These metabolic features are generally accompanied by a high cell proliferation rate and a high level of secondary metabolite production in bacterial species [59].

The cocktail medium also increased the flux in the amino acid biosynthesis pathways, especially in glycine, serine and threonine metabolism (bsu00260). These modifications are also known to be involved in increased protein biosynthesis and bioactive compound production, particularly in B. velezensis, a producer of antimicrobial peptides and lipopeptides [69]. The reaction-centered approach demonstrated by this analysis focuses on the reactions R00355 (pyruvate → acetyl-CoA) and R00510 (glutamine ↔ glutamate) and acknowledges that metabolic profiles are strictly controlled by enzymatic activities and the availability of substrates [70]. Interestingly, these changes might also be programmed for balanced primary and bioactive secondary metabolism under cocktail media, which is frequently observed in bioengineering studies toward biocontrol and biosynthesis.

A higher AUC, closer to 1, indicates a more favorable diagnostic effect. The accuracy of a given AUC value decreases when it falls within the range of 0.5–0.7. Conversely, when the AUC value falls within the range of 0.7–0.9, it exhibits a level of accuracy that can be considered certain. Furthermore, an AUC value exceeding 0.9 indicates a higher degree of accuracy. When the area under the curve (AUC) is equal to 0.5, the diagnostic modality is completely ineffective and has no diagnostic utility [71].

Conclusion

Our findings revealed that the cocktail medium significantly enhanced the production of key antifungal compounds and led to improved disease suppression, thus providing a rational basis for fermentation optimization in microbial biocontrol formulations. This work offers new insights into the metabolic adaptability of B. velezensis WY-65 and underscores the potential of medium engineering as a strategy to improve the efficacy of microbial biopesticides in sustainable agriculture [72–74].

Electronic supplementary material

Below is the link to the electronic supplementary material.

Abbreviation

ANOVA

Analysis of variance

AUC

Area under the curve

CV

Coefficient of variation

MS

Mass Spectrometry

MVA

Multivariate Analysis

NEG

Negative Ion Mode

OPLS-DA

Orthogonal projections to latent structures Discriminant analysis

POS

Positive Ion Mode

PCA

Principal component analysis

OE

Orbitrap Exploris

QC

Quality Control

ROC

Receiver operating characteristic curve

RSD

Relative standard deviation

SVR

Support Vector Regression

TIC

Total Ion Current

UHPLC

Ultra-High-Performance Liquid Chromatography

UVA

Univariate Analysis

VIP

Variable Importance in the Projection

Author contributions

L.C. and T.A. designed the project, L.C. supervised the experiments, T.A. performed all the experiments, and T.A. X.P. M.I. Z.C. and H.Q. analyzed the data. T.A. and Z.C. wrote the initial manuscript. T.A. L.C. and Z.C. edited the manuscript. T.A. and M.I. prepared the final draft. H.Q. visualized and approved the study. All the authors checked and verified the final manuscript.

Funding

National Key Research and Development Program (2024YFD1400802); Basic research business expenses of Liaoning Academy of Agricultural Sciences (2025XTCX0404005); Integration and Demonstration of Supporting Technologies for New Peanut Varieties (2023JH1/10200002); National Modern Agricultural Industry Technology System Peanut Industry Liaoning Innovation Team Construction (2025); Liaoning Province Germplasm Innovation Grain Storage and Technology Special Plan (2023020524-JH1/102-05); Plant Protection Research Institute Science and Technology Innovation Fund Basic Research Special Project (2025ZBJC002); Plant Protection Research Institute Science and Technology Innovation Fund Key Technology R&D Special Project (2025ZBGJ002); President’s Foundation Project of the Liaoning Academy of Agricultural Sciences (2020QN2407).

Data availability

The main data presented in this work have been added to the supplementary material, and other minor data are available upon reasonable request. The 16S rRNA sequence data were deposited in GenBank under the accession number PV770029. The sequences of the gyrA and rpoB genes were deposited in GenBank under Submission # 3,014,669 and Submission # 3,014,679, respectively. The accession will be released in the later stage of the article.

Declarations

Ethics approval and consent to participate

All methods complied with relevant institutional, national, and international guidelines and legislation. The peanut seeds were provided by the Institute of Plant Protection, Liaoning Academy of Agricultural Sciences, Shenyang, China.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.