Transcriptomic and metabolomic analyses of systemic resistance against tomato bacterial wilt induced by lipopeptide from Bacillus pumilus LP-823

Abstract

Bacterial wilt, caused by Ralstonia solanacearum, poses a significant threat to global tomato production, underscoring the need for sustainable management strategies. In this study, we characterized a novel lipopeptide from Bacillus pumilus strain LP-823 and evaluated its efficacy against tomato bacterial wilt. Structural analysis using LC-QTOF MS/MS identified the lipopeptides as surfactin variants with fatty acid chains ranging from C₁₄ to C₁₉, including C₁₇–C₁₉ homologues not previously reported in this species. In pot experiments, treatment with LP-823 lipopeptides significantly suppressed disease development, delaying symptom onset by four days and reducing final disease incidence and disease index by 29.63% and 36.19%, respectively, at 20 days post-inoculation. The lipopeptides exhibited no direct antibacterial activity in vitro. Instead, they triggered plant-mediated defense responses, as evidenced by marked increases in phenylalanine ammonia-lyase (PAL) and polyphenol oxidase (PPO) activities, which peaked at 36 h post-treatment with 5.7-fold and 6.8-fold increases, respectively. Integrated transcriptomic and metabolomic analyses revealed that lipopeptide treatment orchestrates a multifaceted defense network, including modulation of ROS homeostasis (downregulation of AOX1a and CAT2), reprogramming of ABA signaling (upregulation of ABI2, downregulation of ZEP), suppression of aquaporin (PIP2-1) expression, and redirection of phenylpropanoid metabolism. These coordinated responses collectively explain the observed disease suppression and enhanced defense enzyme activities. Overall, our results establish the LP-823 lipopeptide as a potent elicitor of induced systemic resistance with considerable potential for the sustainable management of bacterial wilt in tomato.

Highlights

A novel lipopeptide from Bacillus pumilus LP-823 was identified as a surfactin with C₁₄–C₁₉ fatty acid chains, including C₁₇–C₁₉ homologues not previously reported in this species.

LP-823 lipopeptides significantly suppressed tomato bacterial wilt by delaying disease onset and reducing disease incidence and severity, without direct antibacterial activity.

Multi-omics analysis revealed that the lipopeptide activates a coordinated defense network involving ROS modulation, ABA signaling reprogramming, and redirection of phenylpropanoid and alkaloid metabolism.

This study establishes the LP-823 lipopeptide as a potent elicitor of induced systemic resistance and a promising candidate for sustainable biocontrol of bacterial wilt.

Introduction

Ralstonia solanacearum is a devastating soil-borne pathogen with a broad host range, capable of infecting over 450 plant species across 54 families, including economically critical crops such as tomato, tobacco, and potato [1]. The vascular disease it causes, bacterial wilt [2], ranks among the most widespread and economically damaging soil-borne diseases globally [3]. In tomato production, field incidence in severely affected areas can exceed 80%, leading to total yield loss and posing a major constraint to cultivation [4]. Disease management is exceptionally challenging, as conventional agricultural practices and chemical controls are often ineffective [5, 6]. These challenges underscore the imperative to develop sustainable and eco-friendly alternatives, fueling significant interest in biological control strategies that harness natural antagonists and their bioactive metabolites.

Among biocontrol agents, Bacillus species are widely employed, largely due to their ability to secrete a diverse array of lipopeptides. These metabolites exert a dual mode of action: directly antagonizing phytopathogens or inducing systemic resistance (ISR) in host plants [7–9]. Key lipopeptide families, including surfactin, fengycin, and iturin, are recognized as crucial for disease suppression. Notably, the capacity of Bacillus strains to induce plant defenses is strongly correlated with surfactin production [10]. For example, surfactin from B. subtilis enhances resistance in rice against Botrytis cinerea [11], induces ISR in tobacco against tobacco mosaic virus [12, 13], and triggers defensive responses in tomato, soybean, and rice [14–16]. In addition to their role as ISR elicitors, lipopeptides are also valued for their broad-spectrum antimicrobial activity, structural stability, and low environmental toxicity—properties inherent to their unique molecular architecture [9]. While the functions of lipopeptides from model species such as B. subtilis and B. amyloliquefaciens have been well characterized, research on those produced by B. pumilus remains scarce. Specifically, the role of B. pumilus lipopeptides as ISR elicitors and their underlying mechanisms are largely unexplored, representing a critical knowledge gap that warrants further investigation. The lipopeptide surfactin is a well-documented elicitor of induced systemic resistance (ISR). Its capacity to stimulate plant defense is well-established, with activity strongly correlating with its application content [10, 14]. For instance, surfactin from B. subtilis has been shown to enhance resistance in rice against Botrytis cinerea, concurrently modulating the activities of key defense enzymes like POD, PPO, and SOD [17]. Furthermore, it effectively induces systemic resistance in tobacco against tobacco mosaic virus [12, 13] and primes defensive responses in other crops including tomato, soybean, and rice [14–16, 18]. ISR is a physiological state of enhanced defensive capacity activated in plants upon perception of beneficial microorganisms or specific microbial compounds, such as lipopeptides, prior to pathogen attack [19]. This primed state enables a more rapid and robust response to subsequent infection. ISR establishment primarily involves the jasmonic acid (JA) and ethylene (ET) signaling pathways [20], which regulate the expression of defense-related genes. Complex crosstalk can occur with the salicylic acid (SA) pathway, typically associated with systemic acquired resistance (SAR) [21]. Lipopeptides like surfactin, with their unique properties, represent promising tools for sustainable protection strategies that harness this innate immunity.

Preliminary evidence indicates that lipopeptides from B. pumilus strain LP-823 possess promising biocontrol activity against tomato bacterial wilt. However, their precise mechanism of action, particularly their capacity to elicit ISR, is unknown. To address this, the present study first molecularly identified strain LP-823 and characterized its lipopeptide profile via LC-QTOF MS/MS. We then employed an integrated transcriptomic and metabolomic approach, using the tomato-Ralstonia solanacearum pathosystem, to comprehensively profile the molecular and biochemical changes in tomato plants following LP-823 lipopeptide application. This multi-omics strategy aimed to delineate the key signaling and metabolic pathways involved, thereby elucidating the mechanism by which B. pumilus LP-823 lipopeptides elicit systemic resistance against bacterial wilt.

Materials and methods

Strains and culture media

Strain LP-823 was isolated from the rhizosphere soil of rapeseed in Gansu Province, China, and is preserved at the China Center for Type Culture Collection (CCTCC No. M2025925). Although initially classified as Bacillus safensis, subsequent whole-genome sequence analysis confirmed its identity as Bacillus pumilus. The pathogenic strain Ralstonia solanacearum FJAT-91, used in this study, was isolated from tomato plants exhibiting bacterial wilt symptoms and is deposited at the China General Microbiological Culture Collection Center (CGMCC No. 10692) [22]. Strain LP-823 was cultured in LB liquid medium at 30 °C with shaking at 170 rpm for 48 h. Strain FJAT-91 was routinely cultured on triphenyl tetrazolium chloride (TTC) medium for 48 h at 30 °C; a single colony was then suspended in NB medium. All culture media were purchased from AoBoXing Biological Technology Co., Ltd. (Beijing, China).

Identification of strain LP-823

Strain LP-823 genomic DNA was extracted for whole-genome sequencing on an Illumina platform. The genome sequence was subsequently submitted to the EzBioCloud database (https://www.ezbiocloud.net/) for homology comparison and average nucleotide identity (ANI) analysis to determine the precise taxonomic status.

For phylogenetic analysis, the 16S rRNA gene sequence of strain LP-823 was extracted from the genome. A phylogenetic tree based on 16S rRNA gene sequences was constructed using the neighbor-joining (NJ) method with MEGA 6.0 software. The analysis involved the sequence of strain LP-823 and the sequences of the 20 most closely related type strains obtained from the database.

To further clarify the phylogenetic position, a genome-based phylogenetic tree was also constructed using the neighbor-joining method in MEGA 6.0, based on the concatenated sequences of 31 housekeeping genes (dnaG, frr, infC, nusA, pgk, pyrG, rplA, rplB, rplC, rplD, rplE, rplF, rplK, rplL, rplM, rplN, rplP, rplS, rplT, rpmA, rpoB, rpsB, rpsC, rpsE, rpsI, rpsJ, rpsK, rpsM, rpsS, smpB, and tsf). The analysis included strain LP-823 and the 19 most closely related type strains based on genome-wide comparisons.

Isolation and compositional identification of LP-823 lipopeptides by LC-QTOF/MS

Isolation and compositional identification of LP-823 lipopeptides were performed with slight modifications based on previous studies [23]. Strain LP-823 was cultured in 500 mL of liquid medium at 30 °C with shaking at 180 rpm for 48 h. The culture was centrifuged (10,000 rpm, 10 min), and the supernatant was acidified to pH ≤ 2.0 to obtain the crude lipopeptides. Chromatographic separation was performed on an Agilent Zorbax Extend-C₁₈ column (150 × 2.1 mm, 3.5 µm) at a flow rate of 0.3 mL/min. Mobile phase A consisted of 0.1% formic acid in water, and mobile phase B was methanol. The gradient elution program was as follows: 0–60 min, linear gradient from 60 to 100% B; hold at 100% B for 5 min. ESI–MS/MS analysis was conducted in both positive and negative ion modes. The drying gas temperature was set at 300 °C with a flow rate of 5 L/min, and the nebulizer pressure was 30 psig. The fragmentor voltage was 200 V, collision energy was 100 V, and skimmer voltage was 65 V. The scan range was m/z 100–3000. Purified surfactin (Sigma-Aldrich) served as the standard.

Antibacterial activity assay

The antimicrobial activity of LP-823 lipopeptides against R. solanacearum FJAT-91 was evaluated using the filter paper disc agar diffusion method [23]. The pathogen was cultured in NB medium at 30 °C for 48 h, then suspended in semi-solid NA at a concentration of 10⁷ CFU/mL and poured into NA-coated plates. After solidification, 7-mm diameter wells were punched and filled with 100 µL of LP-823 lipopeptides solutions at concentrations of 5, 15, and 20 mg/mL. Plates were incubated at 30 °C for 48 h, and inhibition zones were measured. Experiments were performed in triplicate.

Biocontrol efficacy of Bacillus pumilus LP-823 lipopeptides against tomato bacterial wilt

The tomato cultivar used in this study was ‘Shouhe Fenguan’, which was obtained under Crop Seed Production and Operation License No. D (Lu-Wei-Shou) 0014 (2017). Plants at the five- to six-leaf stage were selected for uniformity and randomly divided into two groups: a control group and a lipopeptide-treated group. Each plant was grown in a commercial soilless potting substrate (organic matter > 60%, EC 0.8–1.2, porosity > 60%, pH 5.5–6.5; Shengsheng Agriculture Co., Ltd., Guangzhou, China). All experiments were conducted in a greenhouse under controlled conditions: 30 ± 2 °C, 80% relative humidity, and a 12 h light/12 h dark photoperiod. Plants in the treatment group received a root drench of 25 mL of B. pumilus LP-823 lipopeptide solution at a concentration of 0.2 mg/mL. Forty-eight hours post-treatment, all plants were challenge-inoculated with R. solanacearum FJAT-91 by root drenching, applying 25 mL of a bacterial suspension at a concentration of 1.0 × 10⁷ CFU/mL per plant. The inoculum was prepared by first culturing the pathogen on TTC-containing nutrient agar plates for 48 h, followed by growth in nutrient broth at 30 °C with shaking at 180 rpm for 24 h, and finally adjusting to the desired concentration. The typical symptoms of bacterial wilt were evaluated according to a disease index (di) on a scale of 0 to 4. The disease incidence rate, disease severity index and biocontrol efficacy were calculated using the following formulas [24]:

Induction of defense-related enzymes in tomato plants by the Bacillus pumilus LP-823 lipopeptide

Tomato plants were sampled at 0, 12, 24, 36, and 48 h after treatment with the B. pumilus LP-823 lipopeptide. Sampled tissues were immediately frozen in liquid nitrogen and stored at −80 °C until analysis.

For the PAL assay, approximately 1.0 g of plant tissue was homogenized in 1 mL of ice-cold extraction buffer. The homogenate was centrifuged at 12,000 × g for 10 min at 4 °C, and the resulting supernatant was collected and kept on ice for the assay. PAL activity was determined using the PAL Assay Kit (Jiangsu ADS Biotechnology Co., Ltd., Cat. No. ADS-W-KN005-48, microplate method) by measuring the absorbance at 290 nm using a microplate reader.

For the PPO assay, approximately 0.5 g of tissue was homogenized in 1 mL of ice-cold specific assay buffer. The homogenate was centrifuged at 12,000 × g for 15 min at 4 °C, and the supernatant was collected for analysis. PPO activity was quantified using the PPO Assay Kit (Jiangsu ADS Biotechnology Co., Ltd., Cat. No. ADS-F-YH005-48, spectrophotometric method) by monitoring the increase in absorbance at 420 nm.

For the SOD assay, approximately 0.1 g of plant tissue was homogenized in 1 mL of ice-cold extraction buffer. The homogenate was centrifuged at 12,000 × g for 10 min at 4 °C, and the resulting supernatant was collected and kept on ice for the assay. SOD WST-8 Assay Kit (Jiangsu ADS Biotechnology Co., Ltd., Cat. No. ADS-W-KY011, microplate method) by measuring the absorbance at 450 nm with a microplate reader.

For the POD assay, approximately 0.1 g of plant tissue was homogenized in 1 mL of ice-cold extraction buffer. The homogenate was centrifuged at 12,000 × g for 10 min at 4 °C, and the supernatant was collected. POD Assay Kit (Jiangsu ADS Biotechnology Co., Ltd., Cat. No. ADS-W-KY003-48, microplate method) by measuring the increase in absorbance at 470 nm due to the oxidation of guaiacol in the presence of hydrogen peroxide.

Transcriptome sequencing and analysis

Total RNA was extracted and purified using VAHTS mRNA Capture Beads (Vazyme, China) according to the manufacturer’s protocol. Poly(A)-tailed mRNA was enriched through two rounds of oligo (dT) beads, eluted, and fragmented at 94 °C to generate fragments of 100–300 bp. First-strand cDNA was synthesized by reverse transcription, followed by second-strand synthesis, end-repair, and A-tailing in a single reaction with bead cleanup. Illumina adapters were ligated, and double size selection (200–500 bp) was performed. The library was PCR-amplified (12–15 cycles) and purified. Concentration and fragment size were assessed using a Qubit 3.0 fluorometer (Thermo Fisher Scientific, USA) and an Agilent 2100 Bioanalyzer (Agilent Technologies, USA). Qualified libraries (concentration > 5 ng/μL, fragment size 300–400 bp) were sequenced on the NovaSeq platform with 2 × 150 bp paired-end reads.

Raw reads were assessed using FastQC. Adapters, low-quality bases (Q < 15), and reads shorter than 40 bp were removed using fastp to obtain clean reads. Clean reads were aligned to the reference genome using STAR. Read distribution across coding, UTR, intronic, and intergenic regions was summarized using Picard. Library saturation was evaluated by RSeQC via random subsampling; transcripts were stratified by RPKM, and the proportion of transcripts with RPKM values within 15% error of the final estimate was calculated. Gene expression levels were quantified, and differential expression analysis between the LP-823 lipopeptide-treated group and the control (CK) was conducted using DESeq2. Genes with a P-value < 0.05 and |log₂(fold change)|> 1 were considered differentially expressed genes (DEGs). Among these, genes with log₂(fold change) > 1 were classified as upregulated, and those with log₂(fold change) < −1 as downregulated. Genes that did not meet these criteria were defined as non-DEGs. Based on the DEGs identified from the differential expression analysis and designated as up-regulated (Up) or down-regulated (Down), the genes were divided into three categories: All, Up, and Down (All = Up + Down). GO and KEGG enrichment analyses were performed separately for each of these three gene categories using clusterProfiler.

Quantitative real-time PCR analysis

First-strand cDNA was synthesized from 1 μg of total RNA using YALEPIC® All-in-RT RedMix (Yali Bio, China) in a 20 μL reaction under the following conditions: 25 °C for 10 min, 55 °C for 15 min, 85 °C for 5 min, and hold at 4 °C. The cDNA was diluted 5-fold and used as a template for qPCR. Gene-specific primers for the tomato actin gene (ACT51, internal control) and 11 target genes (AOX1a, CAT2, ABI2, ZEP, GAD2, WLIM1, PIP2-1, WRKY75, Ferredoxin, THI012, and THI032) were employed. Each 20 μL qPCR reaction contained 2 μL of diluted cDNA, 0.2 μM of each primer, 10 μL of PerfectStart® Green qPCR SuperMix (TransGen Biotech, China), and 7.2 μL of ddH₂O. Thermal cycling was performed on an ABI QuantStudio 3 system (Thermo Fisher, USA) as follows: 94 °C for 30 s; 40 cycles of 94 °C for 5 s and 60 °C for 30 s; followed by a dissociation stage. Relative expression levels were calculated using the comparative CT method. Relative expression levels of target genes were calculated using the 2^−ΔΔCt method [25].

Metabolomic analysis

An appropriate amount of sample was weighed into a 2 mL Eppendorf tube, to which 600 µL of a 2-chloro-L-phenylalanine solution (4 ppm, prepared in methanol) was added. The mixture was vortexed for 30 s, then a steel bead was introduced, and the tube was placed in a tissue grinder and ground at 55 Hz for 60 s. The sample was subsequently sonicated at room temperature for 15 min and centrifuged at 12,000 rpm at 4 °C for 10 min. The supernatant was collected, filtered through a 0.22 µm membrane, and the filtrate was transferred to a vial for analysis.

Metabolomics analysis was performed using an ultra-performance liquid chromatography-quadrupole Orbitrap mass spectrometer (Thermo Fisher Scientific, USA) equipped with an ACQUITY UPLC® HSS T3 column (2.1 × 100 mm, 1.8 µm) (Waters, Milford, MA, USA). The column was maintained at 40 °C, and the flow rate was set at 0.30 mL/min. For positive ion mode, the mobile phase consisted of 0.1% formic acid in acetonitrile (B2) and 0.1% formic acid in water (A2). For negative ion mode, the mobile phase consisted of acetonitrile (B3) and 5 mM ammonium formate in water (A3). The injection volume was 2 µL.

The gradient elution program was as follows: Positive ion mode: 0–1 min, 10% B2; 1–5 min, linear gradient from 10 to 98% B2; 5–6.5 min, 98% B2; 6.5–6.6 min, linear gradient from 98 to 10% B2; 6.6–8 min, 10% B2; Negative ion mode: 0–1 min, 10% B3; 1–5 min, linear gradient from 10 to 98% B3; 5–6.5 min, 98% B3; 6.5–6.6 min, linear gradient from 98 to 10% B3; 6.6–8 min, 10% B3.

Mass spectrometry was conducted using a Thermo Orbitrap Exploris 120 mass spectrometer (Thermo Fisher Scientific, USA). The instrument was equipped with an electrospray ionization (ESI) source and operated in both positive and negative ion modes. The positive ion spray voltage was set at 3.50 kV, while the negative ion spray voltage was set at −2.50 kV. The sheath gas flow rate was set at 40 arb, and the auxiliary gas flow rate was 10 arb. The capillary temperature was maintained at 325 °C. The mass spectrometer was configured for full scan resolution at 60,000 with a scan range of m/z 100–1000. Higher-energy collisional dissociation was employed for fragmentation with a collision energy of 30%, and the MS/MS resolution was set at 15,000. The top four ions were selected for fragmentation, and dynamic exclusion was applied to remove unnecessary MS/MS information. The criteria for screening differential metabolites were p < 0.05 and OPLS-DA VIP score > 1. Five replicates were performed for each condition.

Statistical analysis

Results were presented as the mean ± standard deviation (SD) of replicates. Statistical analyses were performed using a t-test and one-way ANOVA to evaluate differences between treatments. Data analysis and graphing were conducted using GraphPad Prism 8. All experiments were conducted from May 2024 to October 2025.

Results

Identification of strain LP-823

The genome of strain LP-823 comprises 3,631,276 bp with a chromosomal DNA G + C content of 41.88%. Strain identification was conducted using a polyphasic approach. First, 16S rRNA gene sequence analysis showed that LP-823 exhibited 100% identity with B. pumilus (Fig. 1A). Furthermore, housekeeping gene analysis supported this assignment, revealing 99.4% sequence identity with the same reference strain (Fig. 1B). The average nucleotide identity (ANI) values between strain LP-823 and B. licheniformis ATCC 14580^T, B. siamensis KCTC 13613^T, B. amyloliquefaciens DSM7^T, B. altitudinis DSM 21631^T, B. xiamenensis HYC-10^T, B. zhangzhouensis DW5-4^T, B. safensis FUA2117^T and B. pumilus DSM27^T were calculated as 70.62, 70.46, 70.98, 89.30, 89.68, 91.48, 92.40, and 95.79%, respectively. The last one displayed ANI value of > 95%. These results collectively confirm that strain LP-823 represents a member of the species B. pumilus.

Fig. 1.

Phylogenetic tree of strain LP-823 and related strains. A phylogenetic tree of 16S rRNA genes; B phylogenetic tree of house-keeping genes

Identification of Bacillus pumilus LP-823 lipopeptides

The lipopeptide composition of B. pumilus LP-823 was analyzed by LC-Q/TOF–MS/MS (Fig. 2, Table 1). The components of each type lipopeptide were putatively annotated by comparison of their accurate mass and MS/MS fragmentation patterns with those in the relevant literature [22, 26].

Fig. 2.

The total ion flow spectra of Bacillus pumilus LP-823 lipopeptides

Table 1.

Identification of Bacillus pumilus LP-823 lipopeptides by LC-QTOF-MS/MS

fract-ion	Retention time (min)	MS m/z [M + Na]+	MS2 m/z [M + Na]+	Identified compounds
1	52.23	1044	931.6, 818.5, 728.5, 657.5, 594.3, 481.3, 391.3	C14 surfactin
2	53.46	1058	945.6, 832.5, 742.5, 671.5, 594.4, 481.3, 391.3	C15 surfactin
3	54.50	1072	959.6, 846.6, 756.5, 685.5, 594.5, 481.3, 391.3	C16 surfactin
4	56.62	1086	973.6, 860.5, 770.5, 699.5, 594.5, 481.3, 391.3	C17 surfactin
5	57.45	1100	987.6, 874.5, 784.5, 713.5, 594.5, 481.3, 391.3	C18 surfactin
6	58.80	1114	1001.6, 888.5, 798.5, 594.5, 481.3, 391.3	C19 surfactin

Only one class of cyclic lipopeptides was detected, namely surfactin (retention time, 52.23–58.80 min), MS and MS² spectral data and identification results of B. pumilus LP-823 lipopeptides are summarized in Table 1. The results show that the lipopeptides consisted of C₁₄-C₁₉ surfactin. To our knowledge, this series of compounds with the same C₁₇–C₁₉ surfactin structure has not previously been reported in B. pumilus.

Antibacterial activity of LP-823 lipopeptide against Ralstonia solanacearum

The antimicrobial activity of LP-823 lipopeptides against R. solanacearum FJAT-91 was evaluated, and the lipopeptides exhibited no direct antibacterial activity against the pathogen, as no inhibition zones were observed at concentrations of 5, 15, and 20 mg/mL (Fig. 3).

Fig. 3.

Antibacterium activity of LP-823 lipopeptide against Ralstonia solanacearum. A 5 mg/mL LP-823 lipopeptide; B 15 mg/mL LP-823 lipopeptide; C 20 mg/mL LP-823 lipopeptide; D CK (sterile water)

Biocontrol efficacy of LP-823 lipopeptide against tomato bacterial wilt

The biocontrol efficacy of B. pumilus LP-823 lipopeptide against tomato bacterial wilt was evaluated in greenhouse pot experiments (Fig. 4A and B). Wilt symptoms appeared in control plants at 8 days post-inoculation but were delayed until 11 days in LP-823-treated plants, and by 20 days post-inoculation, the lipopeptide-treated plants had a final disease incidence of 52.78% and a disease index of 46.53%, corresponding to reductions of 29.63% and 36.19%, respectively, relative to the control (Fig. 4C and D).

Fig. 4.

Biocontrol efficacy of Bacillus pumilus LP-823 lipopeptides against tomato bacterial wilt in pot experiments A LP-823 lipopeptide-treated; Bcontrol plants; C Disease incidence rate (DIR, %); D Disease severity index (DSI, %)

Induction of defense-related enzymes in tomato plants by Bacillus pumilus LP-823 lipopeptides

To elucidate the mechanism underlying the biocontrol activity of B. pumilus LP-823 lipopeptides against tomato bacterial wilt, we first confirmed that the lipopeptides exhibited no direct antibacterial activity against the pathogen (Sect. “Antibacterial activity of LP-823 lipopeptide against Ralstonia solanacearum”). Therefore, we focused on their potential role in activating host resistance in tomato. Tomato plants were sampled at 0, 12, 24, 36, and 48 h after treatment with the LP-823 lipopeptide. The treatment significantly increased the activity of four defense-related enzymes: phenylalanine ammonia lyase (PAL), polyphenol oxidase (PPO), superoxide dismutase (SOD), and peroxidase (POD). However, the magnitude and temporal patterns of these increases differed markedly between enzymes.

PAL activity in the LP-823-treated plants was significantly higher than in the control group at 24 and 36 h post-treatment (p < 0.0001). The activity peaked at 36 h with a value of 953.43 U/g FW, which was 5.7-fold higher than that of the control (Fig. 5 A). Similarly, PPO activity in the LP-823-treated plants was significantly higher than in the control group at 36 and 48 h post-treatment (p < 0.0001). The highest PPO activity, observed at 36 h, reached 367.62 U/g FW, representing a 6.8-fold increase over the control level (Fig. 5B).

Fig. 5.

Induction of defense enzyme activities in tomato plants by LP-823 lipopeptides. A phenylalanine ammonia-lyase (PAL) activity; B polyphenol oxidase (PPO) activity; C superoxide dismutase (SOD) activity; D peroxidase (POD) activity. Note: Significance levels are indicated as follows, **** (p < 0.0001), *** (p < 0.001), ** (p < 0.01), ns, not significant, (p ≥ 0.05)

In striking contrast, SOD and POD activities peaked at 12 h after lipopeptide treatment (increased by 3.9% and 14.8%, respectively), then gradually declined(Fig. 5C and D). Both enzymes reached their lowest levels at 36 h (SOD: 45.48 U/g FW, 0.70-fold of 0 h; POD: 78.44 U/g FW, 0.53-fold of 0 h). Notably, despite this decrease, the activities at 48 h partially recovered to 0.88-fold (SOD) and 0.72-fold (POD) of the control, indicating that the overall fold changes throughout the 48 h period remained relatively limited.

Overall, unlike the robust, multi-fold activation of PAL and PPO, the fluctuations of SOD and POD activities remained within a narrow range (all fold changes < 1.5-fold) throughout the 48 h period.

Transcriptome analysis of the tomato stress response to Bacillus pumilus LP-823 lipopeptides

The sequencing data obtained in this study exhibited high quality, with over 96.6% of bases achieving a Q30 score, confirming the reliability of the data for subsequent analyses. Transcriptomic profiling revealed that tomato plants treated with LP-823 lipopeptides exhibited 6,219 DEGs compared to the control group, comprising 2,396 upregulated and 3,360 downregulated genes (Fig. 6A). The plot clearly indicates that downregulated genes were more prevalent than upregulated ones.

Fig. 6.

Transcriptomic profiling and validation of LP-823 lipopeptide-induced responses. A Volcano plot of DEGs. Red: upregulated (log₂FC > 1, p-value < 0.05, 2,396 genes); green: downregulated (log₂FC < −1, p-value < 0.05, 3,360 genes); blue: not significant; B GO enrichment analysis of DEGs. Longer bars indicate smaller P-values, representing greater statistical significance; C KEGG enrichment analysis of DEGs. Point color represents p-value (redder = smaller p, higher significance); point size represents gene count (larger = more genes); D Heatmap of top 50 DEGs. Columns: samples; rows: genes. color: red = high expression, green = low expression; E qRT-PCR validation of selected DEGs. Relative expression levels of eleven genes (THI012, WLIM1, GAD2, ABI2, AOX1a, WRKY75, ZEP, THI032, PIP2-1, CAT2, and Ferredoxin) in CK vs. LP-823 (normalized to ACT51; mean ± SD, n = 5). Significance levels are indicated as follows, **** (p < 0.0001), *** (p < 0.001), ** (p < 0.01), ns, not significant, (p ≥ 0.05). The qRT-PCR results were consistent with the transcriptome sequencing data, further supporting the reliability of the RNA-seq analysis; (F) Correlation analysis between RT-qPCR and RNA-seq data. The regression curve (Y = 0.9869X + 0.07564) demonstrates excellent consistency between the two methods, with an R² of 0.9329

Differential expression analysis between the LP-823 lipopeptides treatment group and the control (CK) was conducted using DESeq2, with genes satisfying |log₂FC|> 1 defined as DEGs. Among these, 2,396 genes were identified as upregulated (log₂FC > 1) and 3,360 as downregulated (log₂FC < −1).

To functionally characterize these DEGs in the context of tomato response to LP-823 lipopeptide stress, Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses were performed. As shown in Fig. 6B, DEGs under LP-823 lipopeptides treatment were widely distributed across GO categories. Within biological processes, enriched terms included “response to hormone,” “response to endogenous stimulus,” “response to abiotic stimulus,” and “hormone-mediated signaling pathway.” Cellular component terms were mainly associated with chloroplast structures, such as “thylakoid,” “photosynthetic membrane,” “chloroplast thylakoid,” and “plastid thylakoid.” Molecular function categories showed enrichment for “tetrapyrrole binding,” “heme binding,” “oxidoreductase activity,” “monooxygenase activity,” and “iron ion binding.”

KEGG pathway analysis indicated significant enrichment across 135 metabolic and signaling pathways. Key pathways included “plant hormone signal transduction,” “phenylpropanoid biosynthesis,” “plant—pathogen interaction,” “biosynthesis of amino acids,” “flavonoid biosynthesis,” “glyoxylate and dicarboxylate metabolism,” “photosynthesis-antenna proteins,” and “circadian rhythm—plant” (Fig. 6C).

Among the top 50 most significantly altered DEGs, 15 were upregulated and 35 were downregulated. The upregulated DEGs included LOC101256771, LOC101263256, THI012 (LOC101055512), WLIM1 (LOC101257855), LOC101264428, LOC101259710, LOC101247272, PTPKIS1, LOC101265590, DDTFR10A, ABI2, LOC101258673, LOC101249886, LOC101267825, and LOC101259262. In contrast, the downregulated DEGs included LOC101246031, LOC101266223, LOC101265598, LOC101246032, LOC101256576, PIP2-1, PERC2, CAT2, ZEP, ABCB21, GAD2 and AOX1a among others (Fig. 6D).

To validate the transcriptomic findings, 11 DEGs were selected for qPCR analysis (Fig. 6E), including AOX1a, CAT2, ABI2, ZEP, GAD2, WLIM1, PIP2-1, WRKY75, Ferredoxin, THI012, and THI032. The qRT-PCR results were consistent with the transcriptome sequencing data, further supporting the reliability of the RNA-seq analysis.

Correlation analysis between RT-qPCR and RNA-seq data revealed a strong positive correlation, with an R² of 0.9329, a slope of 0.9869, and an intercept of 0.07564, indicating excellent consistency between the two methods (Fig. 6F).

Metabolomics profiling of tomato response to Bacillus pumilus LP-823 lipopeptides

Metabolomic profiling of tomato plants was conducted using UPLC-Q-Orbitrap MS, leading to the identification of 3,650 metabolites. Major compound classes included prenol lipids (238 metabolites), organooxygen compounds (185 metabolites), fatty acyls (174 metabolites), carboxylic acids and derivatives (172 metabolites), benzene and substituted derivatives (144 metabolites), and steroids and steroid derivatives (99 metabolites) (Fig. 7A).

Fig. 7.

Metabolic profiling of tomato plants in response to LP-823 lipopeptide. A The number of distinct types of identified metabolites; B Bubble plot of metabolic pathway enrichment. Bubble size = number of differential metabolites; bubble color = p-value (redder = smaller p, more significant; C Boxplot of differential metabolites; D Heatmap of the top 50 differential metabolites. Redder color indicates higher expression, and bluer color indicates lower expression; columns represent samples, rows represent metabolites

Comparative analysis revealed significant metabolic differences between the control and lipopeptide-treated plants. Principal component analysis (PCA) demonstrated clear separation between the two groups, indicating distinct metabolic profiles. Further untargeted metabolomic comparison between surfactin-induced resistant plants and untreated controls identified 1,173 differential metabolites, among which 175 were upregulated and 998 were downregulated. Key upregulated metabolites included 4α-hydroxymethyl-ergosta-7,24 (241)-dien-3β-ol, norlaudanosoline, 2,6-dihydroxybenzoic acid (2,6-DHBA), L-tyrosine, p-hydroxyphenylethanolamine, FAMP, 4-hydroxytryptamine, L-tryptophan, 2-oxosuberate, and desethyl atrazine (Fig. 7B and C). Conversely, notably downregulated compounds comprised (R)-naproxen, isoquercitrin, 5-hydroxypentanoic acid, 2-methylnaphthalene, bovinocidin, and 4-hydroxychlorothalonil.

To elucidate the biological relevance of these metabolic changes, pathway enrichment analysis was performed. The most significantly altered pathways included “biosynthesis of amino acids,” “2-oxocarboxylic acid metabolism,” “biosynthesis of various other secondary metabolites,” “alanine, aspartate and glutamate metabolism,” and “pyruvate metabolism” (Fig. 7D), suggesting coordinated reprogramming of primary and secondary metabolism in response to lipopeptide-induced resistance.

Discussion

Plants possess innate immune systems that can be activated by specific elicitors. Unlike pathogen-dependent SAR, ISR is triggered by beneficial microorganisms or abiotic stimuli, offering a sustainable strategy for disease management [27, 28]. Among microbial elicitors, the lipopeptide surfactin has been extensively documented as a potent inducer of plant defense responses, capable of triggering ISR in various plant species [10, 13, 29]. In the present study, we found that lipopeptides from B. pumilus LP-823 activate ISR in tomato against R. solanacearum. Specifically, Cawoy et al. [10] demonstrated a concentration-dependent defense induction by surfactin; our results extend this by showing that even at low concentrations, the C₁₄–C₁₉ surfactin homologues from LP-823 elicit robust systemic resistance. Consistent with previous reports in tomato [14–16, 18], rice [30], and tobacco [12, 13, 16], LP-823 lipopeptide treatment significantly suppressed disease development, delayed symptom onset by four days, and reduced final disease incidence and disease index by 29.63% and 36.19%, respectively. Moreover, the lipopeptides exhibited no direct antibacterial activity against R. solanacearum in vitro, and triggered marked increases in the defense-related enzymes PAL (5.7-fold) and PPO (6.8-fold) at 36 h post-treatment. These combined observations—absence of direct antibiosis, systemic protection, and activation of classical ISR markers—provide strong evidence that LP-823 lipopeptides act as genuine ISR elicitors rather than through direct pathogen antagonism.

Historically, the ISR-inducing function of surfactin was first characterized in commercially available biocontrol strains such as B. subtilis [31, 32]. However, accumulating evidence indicates that B. pumilus is also a significant source of structurally novel surfactin homologues. For instance, surfactin from B. pumilus W-7 induced defense responses in potato by upregulating pod, pal, and cat transcription and their corresponding enzyme activities [33]. Specific isoforms, including C₁₅ and C₁₆ surfactins, have been identified in other B. pumilus strains [34], and B. pumilus HY1 was found to produce predominantly C₁₂–C₁₆ surfactins [35]. These observations suggest that B. pumilus represents an underexplored reservoir of structurally diverse surfactins. Our lipopeptide profiling of strain LP-823 revealed a series of C₁₄–C₁₉ homologues, confirming that B. pumilus is indeed a source of diverse surfactins. Notably, the C₁₇–C₁₉ derivatives, identified by diagnostic fragments and conserved peptide sequence, have not been previously reported in B. pumilus. The presence of these long-chain homologues suggests they may possess unique physicochemical properties or modes of interaction with plant cell membranes, potentially underpinning their ability to elicit broad-spectrum and effective resistance in tomato.

To elucidate how these lipopeptides orchestrate plant defense, we employed integrated transcriptomic and metabolomic analyses. Our results revealed a multi-layered molecular network underlying lipopeptide-ISR against R. solanacearum in tomato. This network is characterized by the coordinated reprogramming of redox, hormonal, physical, and chemical defense components.

At the redox level, we observed significant downregulation of AOX1a (log₂FC = −5.06), CAT2 (log₂FC = −1.96), and Ferredoxin (log₂FC = −2.24). AOX1a and CAT2 are key regulators of mitochondrial and peroxisomal ROS homeostasis; their downregulation promotes controlled ROS bursts that serve as signaling molecules for defense priming. Ferredoxin downregulation may further modulate photosynthetic electron flow and cellular redox status. Together, these changes indicate that LP-823 lipopeptides prime the plant for a rapid and amplified oxidative response upon pathogen challenge, without causing constitutive oxidative damage.

Concerning hormonal signaling, our data showed downregulation of ZEP (log₂FC = –2.89) and upregulation of ABI2 (log₂FC = 1.72), which relieve ABA-mediated growth-defense trade-offs, enabling rapid immune activation. Additionally, upregulation of GAD2 (log₂FC = 1.37) suggests a link between glutamate metabolism and GABA-mediated responses, further expanding the hormonal and metabolic crosstalk involved in ISR establishment.

Physical barriers are critical against vascular pathogens like R. solanacearum, which invades through the xylem. Based on our transcriptomic data, we propose two complementary mechanisms: (i) WLIM1 upregulation (log₂FC = 2.53) likely drives lignin deposition to reinforce cell walls, and (ii) PIP2-1 downregulation (log₂FC = –4.33), may reduce xylem water conductance, potentially restricting pathogen spread.

Chemical defense involves ABCB21 downregulation (log₂FC = –1.31) promoting intracellular retention of antimicrobial alkaloids such as norlaudanosoline (log₂FC = 4.08). This is supported by the accumulation of phenolic precursors L-tyrosine (log₂FC = 0.57) and L-tryptophan (log₂FC = 0.92). Both phenylpropanoids and alkaloids play central roles in plant immunity. Phenylpropanoid synthesis is a key adaptation in land plants for coping with abiotic stresses and microbial pathogens [36]. This pathway supplies precursors for defense-related compounds such as salicylic acid, flavonoids, lignin, and suberin, and its core enzymes (PAL, C4H, 4CL) are present across land plant lineages and consistently upregulated during pathogen infection in both bryophytes and vascular plants [37]. Alkaloids serve as chemical defenses in angiosperms against bacterial pathogens by targeting DNA and membranes via their planar, amphiphilic structures, and potent examples like cryptolepine and tryptanthrin are promising leads for antibiotic development [38].

Transcriptional coordination is marked by upregulation of THI012 (log₂FC = 3.56), downregulation of THI032 (log₂FC = –1.20), and downregulation of WRKY75 (log₂FC = –2.94), which together fine-tune ISR establishment and prevents hypersensitive responses. These genetic and metabolic changes converge on phenylpropanoid and isoquinoline alkaloid biosynthesis pathways. Key metabolites supporting a shift from primary to secondary metabolism include 2-oxosuberate (log₂FC = 0.97) and 2,6-dihydroxybenzoic acid (log₂FC = 1.00), the latter contributing to direct antimicrobial activity, as well as reduced AMP levels (log₂FC = –0.88). In the present study, the lipopeptides from B. pumilus LP-823 triggered plant-mediated defense responses, as evidenced by marked increases in PAL and PPO activities, which peaked at 36 h post-treatment with 5.7-fold and 6.8-fold increases, respectively. Collectively, these observed changes correlate with the delayed disease onset, reduced disease severity, and enhanced defensive enzyme activities observed in treated plants.

Beyond directly activating plant immunity, the lipopeptides produced by Bacillus species exert dual functions in plant protection. First, they exhibit direct bactericidal activity, reducing pathogen damage and inhibiting bacterial growth, thereby indirectly relieving the burden on the plant immune system. Second, as demonstrated in this study and supported by previous reports [39], surfactin stimulates induced systemic resistance (ISR) across different plant-pathogen systems, including tomato against Botrytis cinerea [10, 14]. Notably, the resistance mechanism elicited by LP-823 lipopeptides operates through a systems-level integration of controlled oxidative signaling, ABA-independent defense priming, and lifestyle-specific physical barriers. This strategy confers disease resistance without the growth penalties typically associated with constitutive defense activation, thereby distinguishing it from a purely antimicrobial mode of action and highlighting the potential of B. pumilus LP-823 lipopeptides as sustainable plant protection agents.

The role of secondary metabolites in plant defense is further underscored by recent metabolomic advances. Pathogen-infested plants produce numerous secondary metabolites, some of which function as resistance metabolites that disrupt pathogen structure and inhibit their growth and reproduction. With advances in mass spectrometry and metabolomics, untargeted metabolomics has emerged as a powerful tool for screening disease resistance-related chemicals, understanding underlying mechanisms, and developing innovative disease control approaches. Tang et al. [40] applied untargeted metabolomics to potato spindle tuber viroid (PSTVd)-infected and control plants, identifying xanthohumol and echinocystic acid as inhibitors of PSTVd titer. Similarly, Mareya et al. [41] used untargeted metabolomics on two sorghum cultivars with differential disease tolerance infected with Burkholderia andropogonis, monitoring metabolic changes and identifying salicylic acid, jasmonic acid, and zeatin as significant defense-related biomarker metabolites. Consistent with these approaches, the present study employed untargeted metabolomics on tomato root exudates induced by virulent and avirulent strains of R. solanacearum, successfully identifying potential anti-bacterial wilt metabolites and further validating the utility of this approach in unraveling plant defense mechanisms.

In summary, this study demonstrates that the novel lipopeptides produced by B. pumilus LP-823, particularly the previously unreported C₁₇–C₁₉ surfactin homologues, establish a robust and synergistic defense network in tomato against R. solanacearum. This network integrates redox regulation, hormonal reprogramming (including GABA-related signaling), physical barrier reinforcement, and activation of specific secondary metabolic pathways (phenylpropanoid and isoquinoline alkaloid), providing a multi-layered defense strategy that effectively impedes this vascular pathogen. Collectively, these findings not only deepen our understanding of B. pumilus as a source of bioactive lipopeptides but also provide a theoretical foundation for developing novel, environmentally friendly plant protection agents.

Conclusions

This study identifies a novel lipopeptide from B. pumilus LP-823 as an effective elicitor of ISR against tomato bacterial wilt. Structural analysis revealed C₁₇–C₁₉ surfactin homologues not previously reported in this species, expanding the known diversity of lipopeptide variants with potential biocontrol relevance. Multi-omics analysis further elucidated the underlying molecular network, demonstrating that LP-823 lipopeptides trigger coordinated defense responses across multiple layers: controlled oxidative signaling through downregulation of AOX1a and CAT2, hormonal reprogramming via ABI2 upregulation and ZEP downregulation; physical barrier reinforcement through WLIM1-driven lignification and PIP2-1-mediated vascular restriction, and metabolic shifts toward phenylpropanoid biosynthesis and intracellular accumulation of antimicrobial alkaloids such as norlaudanosoline. These integrated responses collectively contribute to delayed disease onset, reduced disease severity, and enhanced defense enzyme activities without relying on direct antibacterial effects. Collectively, these results indicate that LP-823 lipopeptides trigger ISR in tomato, as evidenced by systemic protection, absence of direct antimicrobial activity, and activation of defense-related responses.

Our findings establish the LP-823 lipopeptide as a promising candidate for lipopeptide-based biopesticides or plant immune primers, aligning with sustainable agriculture principles. Future research should focus on identifying the plant receptors recognizing these novel lipopeptides, validating key regulatory components such as ABCB21 and WRKY75 through genetic approaches, and assessing practical efficacy under field conditions to support the development of sustainable strategies for bacterial wilt management.

Authors’ contributions

**Zheng Chen:** Conceptualization, Writing—original draft, Data curation, Investigation, Formal analysis, Funding acquistion, Writing – review & editing, Project administration. **Junjie Feng:** Methodology, Validation. **Tao Lin:** Writing- review & editing, Formal analysis. **Dongjun Xie:** Validation, Software, Methodology. **Hui Wei:** Writing- review and editing, Supervision. **Enquan Lin:** Methodology, Visualization. **Meichun Chen:** Resource, Writing- review and editing, Supervision.

Funding

The authors gratefully acknowledge the financial support from the Fujian Provincial Special Project for Public Welfare Research Institutes (Grant No. 2026R1023002); the Fujian Provincial Natural Science Foundation (Grant No. 2023J01378) and the Program of the Fujian Academy of Agricultural Sciences (Grant No. YCZX202408).

Data availability

The sequencing data generated in this study have been deposited in the NCBI Genome repository, under the BioProject accession number PRJNA1447142.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.