Unveiling tomato defense mechanisms against Bradysia impatiens (Diptera: Sciaridae) feeding: insights from transcriptomic and metabolomic analyses

Abstract

Background

Tomato (Solanum lycopersicum) is an important agricultural crop frequently threatened by pests such as Bradysia impatiens (B. impatiens), a globally distributed sciarid fly whose larvae inflict severe damage on roots, stems, and leaves, leading to substantial yield losses. Despite its economic importance, the molecular and metabolic mechanisms underlying tomato defense against B. impatiens remain poorly understood.

Results

In this study, integrated transcriptomic and metabolomic approaches were employed to investigate the responses of tomato leaves to B. impatiens larval infestation at one week post-infestation. In our experimental system, larvae were introduced onto the soil surface and sequentially fed on stem bases (causing seedling lodging), leaves of prostrate plants, and roots. The leaf responses therefore reflect a combination of direct larval feeding on leaf tissues and systemic effects resulting from stem and root damage. Herbivory induced extensive transcriptional reprogramming, with 2,973 differentially expressed genes (DEGs) significantly enriched in pathways related to MAPK signaling, plant hormone transduction, plant-pathogen interactions, and phenylpropanoid biosynthesis. Metabolomic analysis identified 1,462 differentially accumulated metabolites, indicating significant shifts in energy metabolism, antioxidant defense systems, and the biosynthesis of defense-related compounds such as terpenoids and phenylpropanoids. Combined analyses revealed synchronized induction of α-linolenic acid metabolism and jasmonate signaling, accompanied by increased accumulation of reactive oxygen species (ROS) and upregulation of ethylene-responsive factors (ERF), bHLH, and NAC transcription factors.

Conclusion

These findings reveal a tiered defensive regulatory network in tomato that integrates signal perception, hormonal regulation, metabolic reconfiguration, and ROS-mediated responses to counteract B. impatiens infestation. This entire defensive cascade embodies herbivory-triggered induced resistance, accompanied by sustained adaptive metabolic remodeling to withstand continuous larval feeding pressure. This study provides a comprehensive perspective on tomato plant-insect interactions and identifies potential targets for enhancing tomato resistance through molecular breeding or ecological management strategies. Notably, the hub transcription factor gene MYC, JA rate-limiting biosynthetic gene 12-oxophytodienoate reductase 3 (OPR3), α-linolenic acid-derived oxylipins and phenylpropanoid metabolites are highlighted as promising molecular biomarkers and core targets for future tomato anti-fungus gnat resistance engineering.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12870-026-09645-w.

Background

The tomato (Solanum lycopersicum) is one of the most economically important crops worldwide and a well-established model organism for plant research. As an extensively cultivated vegetable in facility-based agriculture, it is notable for its broad cultivation range, high yield, and rich nutritional profile. Tomato holds a pivotal position in both agricultural production and scientific research [1, 2]. In addition to its economic significance, the tomato contains a diverse array of metabolites—such as soluble solids, flavonoids, vitamins, and minerals—that contribute essential nutrients for human health [3, 4]. Tomato plants host more than 200 species of diverse pests and pathogens, posing substantial threats to plant health and productivity [5, 6]. Annually, various biotic stressors—including weeds, viruses, insects, bacteria, and fungi—inflict damage on tomato fruits, leading to significant yield losses [7, 8]. Commercial tomato cultivation has increasingly shifted from open-field systems to greenhouse environments. These controlled environments, while beneficial for crop growth, also promote the proliferation of certain pests, particularly Bradysia species. These insects belong to the family Sciaridae and are recognized as major pests in greenhouse production systems [9, 10]. Members of the Sciaridae family are especially prevalent in horticultural nurseries producing cucumbers, tomatoes, ornamental plants, and edible mushrooms, where they cause substantial damage and consequent economic losses [11]. Sciarid larvae exhibit polyphagous feeding behavior, typically consuming algae, fungi, and decomposed plant matter. However, they have been documented to damage healthy plants, including young carnations, cucumbers, and several crop species such as soybean, wheat, corn, and peas, by feeding on the roots and stems of seedlings, and occasionally on leaf tissues and cuttings [10, 12, 13].

Numerous Bradysia species are recognized agricultural pests, with larvae that feed on mycelia, roots, and decaying organic matter [14]. Among them, B. impatiens, commonly known as the fungus gnat, poses significant threats to the cultivation of cucumber, tomato, ornamental plants, and edible mushrooms, as its larvae are capable of destroying entire seedlings [15–17]. Adult females exhibit high fecundity, laying 50–300 eggs in moist organic substrates. The species exhibits a relatively short life cycle, lasting approximately 20–25 days, depending on environmental conditions, and encompasses the egg, four larval instars, pupa, and adult (imago) stages [10]. Consequently, their populations can increase rapidly, resulting in substantial economic losses in edible mushroom and other agricultural crop cultivation. Larval feeding causes physical damage to stem and root tissues, disrupting water and nutrient uptake and leading to symptoms such as wilting, stunted growth, chlorosis, premature leaf drop, and even plant death. In specific greenhouse environments, B. impatiens infestations have been associated with plant mortality rates of 20–40% [18, 19]. Additionally, larvae can indirectly affect plants by facilitating the entry of soil-borne pathogens. Notably, fungus gnats are well-documented vectors of fungal pathogens. Studies have demonstrated that Bradysia species are involved in the dissemination of oomycetes and fungi, such as Pythium, Fusarium, and Verticillium [10, 20].

Conventional management of sciarid flies relies primarily on chemical insecticides, an approach increasingly undermined by environmental toxicity and the emergence of resistance [13, 21]. In response to these limitations, physical control measures, such as yellow sticky traps and surface barriers, have been adopted as supplementary tools for monitoring and capturing adult flies [11, 22]. However, beyond these anthropogenic interventions, plants have developed an intricate suite of morphological and biochemical defenses, both constitutive and inducible, that are highly effective and adaptable in deterring pests and pathogens [23–25]. Constitutive defenses include physical barriers like trichomes and thickened cuticles, which can physically impede insect oviposition and feeding [26]. In contrast, plants can activate inducible defenses upon pest attack, leading to the synthesis and emission of complex blends of volatile organic compounds (VOCs) and toxic secondary metabolites such as terpenoids, alkaloids, and phenolics. These compounds function either directly to repel herbivores or indirectly by attracting their natural enemies [27–30]. This inducible defense system is regulated by an elaborate signaling network: The recognition of herbivore-associated molecular patterns (HAMPs) or damage-associated molecular patterns (DAMPs) initiates intracellular signaling cascades primarily mediated by phytohormones, including jasmonic acid (JA), salicylic acid (SA), and ethylene (ET) [31–34]. Among these, the JA pathway plays a central role in defense against chewing insects like sciarid flies, coordinating transcriptional changes that drive the biosynthesis of defensive metabolites and VOCs [35–36]. The activation of such inducible defenses represents a highly sophisticated and energy-efficient strategy, enabling plants to judiciously balance resource allocation between growth and resistance.

Plants are frequently exposed to pathogenic attacks, and one of their earliest defense responses involves the production of reactive oxygen species (ROS), including superoxide anions (O₂•⁻) and their derivative, hydrogen peroxide (H₂O₂) [37]. This oxidative burst can induce localized cell death by damaging DNA, proteins, and cell membranes near the infestation site, thereby limiting nutrient availability for the proliferation of pathogens [38]. The regulation of ROS metabolism during pathogen invasion involves numerous antioxidant enzymes, including ascorbate peroxidase, peroxidase, superoxide dismutase (SOD), catalase, polyphenol oxidase, and glutathione reductase [39]. During the initial stress response, ethylene-responsive factors (ERFs) enhance ROS synthesis by regulating genes of the respiratory burst oxidase homolog (RBOH) family, thus facilitating rapid signal transduction. In Arabidopsis thaliana, AtERF enhances H₂O₂ production by modulating RBOH family genes, thereby accelerating stress signal transmission and completing early stress responses [40, 41]. The bHLH transcription factor AtbHLH112 upregulates AtPOD and AtSOD genes while downregulating P5CDH and ProDH genes, contributing to enhanced salt stress tolerance in Arabidopsis thaliana [42]. Members of the NAC transcription factor family also play crucial roles in disease resistance by directly regulating defense-related genes, modulating plant hormone signaling pathways, managing ROS homeostasis, triggering the hypersensitive response (HR), and interacting with pathogen effectors [43]. Despite extensive research on tomato defense mechanisms against various chewing pests, the specific molecular and metabolic responses to B. impatiens larval attack—particularly within an integrated transcriptomic and metabolomic framework—remain largely unexplored. In our experimental system, larvae were introduced onto the soil surface and sequentially fed on stem bases (causing seedling lodging), leaves of prostrate plants, and roots. Consequently, the leaf responses analyzed in this study reflect a combination of direct larval feeding on leaf tissues and systemic effects arising from stem and root damage. Plant defense responses to herbivory unfold as a temporal cascade, progressing from early signal perception (minutes to hours) to sustained metabolic reprogramming (days to weeks). However, most multi-omics studies, including the present work, have relied on single time-point designs. This approach provides a robust snapshot of the transcriptomic and metabolomic landscape at a specific stage but cannot capture the full temporal dynamics of the response. Nonetheless, such a snapshot is essential for identifying key pathways and candidate genes, thereby establishing a foundation for subsequent time-resolved investigations.

In this study, we employed combined transcriptomic and metabolomic approaches to investigate changes in gene expression and metabolite profiles in tomato leaves one week after infestation by B. impatiens larvae under the experimental conditions described above. Our objectives were to identify key defense-related genes and metabolites, delineate the signaling pathways and metabolic remodeling activated by herbivory, and establish a multi-omics foundation for understanding tomato–sciarid fly interactions. These findings may inform future efforts to enhance tomato resistance through molecular breeding or ecological management strategies.

Materials and methods

Growth of tomato plants and infestation with Bradysia impatiens

Tomato plants (Solanum lycopersicum cv. Micro-Tom) were used in this study. Seeds were first soaked in deionized water for 30 min, followed by surface sterilization with a 10% (v/v) hypochlorous acid solution for 10 min. After thorough rinsing with deionized water, the seeds were placed on sterile, moistened filter paper and incubated in darkness to promote germination. After 4–5 days, uniformly germinated seedlings were transferred to a potting mixture composed of Pindstrup Substrate (Pindstrup Mosebrug A/S, Denmark) and cultivated in a growth chamber under controlled conditions: 25 ± 2℃ with a 16-hour light/8-hour dark photoperiod for 2–3 weeks.

The pest larvae used in this experiment were obtained from the laboratory greenhouse. Species identification was confirmed by DNA extraction and sequencing, which revealed the pest to be B. impatiens.

Tomato seedlings exhibiting uniform growth were randomly assigned to either an infested group or an untreated control group. For the infested group, approximately 20 third-instar B. impatiens larvae were introduced onto the soil surface around the base of each tomato seedling in each pot (each pot containing five seedlings). The larvae were allowed to freely disperse and feed. Under these conditions, the larvae first fed on the stem bases, causing the seedlings to lodge (fall over) and become prostrate on the soil surface. Subsequently, the larvae continued feeding on the leaf tissues of the lodged seedlings, as well as on the root systems after penetrating the soil. Both groups were maintained under identical growth conditions for one week. Subsequently, plant materials were harvested for morphological observations and further analyses. After one week of B. impatiens larval infestation, comprehensive phenotypic quantification was conducted with no less than 15 biological replicates for each treatment (control and B. impatiens larvae groups). For detached single-leaf damaged area quantification, uniform functional leaves from each seedling were photographed under consistent light conditions. ImageJ software was utilized to calculate the total leaf area and lesion area, and the damaged area ratio (%) was calculated as (total lesion area / total leaf area) × 100. All tomato seedlings were uprooted gently to separate aboveground tissues (stems + all leaves), and fresh weight was recorded immediately using an electronic analytical balance. Seedling survival rate was calculated as (number of surviving seedlings / total initial seedlings per group) × 100.

At harvest, the presence of larvae and feeding damage on stems, leaves, and roots was visually confirmed. For the infested group, the five seedlings in each pot were combined into one biological sample, and a total of four biological samples were collected. Similarly, four biological replicates were collected from the control plants. All harvested samples were immediately snap-frozen in liquid nitrogen and stored at -80 °C until use in transcriptomic, metabolomic, and RNA extraction analyses.

Transcriptome sequencing and data analysis

Total RNA was extracted from four biological replicates of both treated and untreated tomato seedlings using the TRIzol Reagent Kit (Invitrogen, USA). The integrity and purity of RNA samples were assessed through agarose gel electrophoresis, a NanoPhotometer spectrophotometer (Implen, USA), and an Agilent 2100 Bioanalyzer (Agilent Technologies, USA). Only samples that fulfilled the quality requirements were selected for sequencing on the Illumina PE150 platform (Illumina, San Diego, USA) in collaboration with Novomagic (Beijing, China).

Raw sequencing reads in FASTQ format were processed using the fastp software to generate clean reads. This step involved the removal of reads containing adapters, poly-N stretches, or low-quality bases. The Q20, Q30, and GC content of the resulting clean reads were calculated, and all downstream analyses were based on this high-quality dataset. The files for the reference genome and gene model annotation were sourced from NCBI (https://www.ncbi.nlm.nih.gov/datasets/genome/GCF_036512215.1/, accessed on 8 August 2025). An index for the reference genome was built using HISAT2 (v2.2.1), and paired-end clean reads were aligned to the genome using the same tool.

Analysis of differentially expressed genes (DEGs)

DEG analysis was performed using the DESeq2 R package (v1.42.0), which employs a negative binomial distribution model for statistical evaluation. P-values were adjusted using the Benjamini and Hochberg method to control the false discovery rate (FDR). Genes were considered significantly differentially expressed if they met the following criteria: adjusted P-value (padj) ≤ 0.05 and |log₂(fold change)| ≥ 1. This analysis was performed for datasets containing biological replicates. In cases where biological replicates were unavailable, DEG analysis was carried out using the edgeR R package (v4.0.16). Prior to analysis, read counts from each sequencing library were normalized by applying a scaling factor to account for differences in sequencing depth across samples, ensuring accurate comparison of expression levels. Following normalization, DEGs were identified using the same statistical criteria as with DESeq2, with adjusted P-values determining significance.

GO and KEGG enrichment analyses

Gene Ontology (GO) enrichment analysis of the DEGs identified in this study was performed using the clusterProfiler R package (v4.8.1), with adjustments incorporated to account for potential biases associated with gene length. GO terms with a corrected P-value < 0.05 were considered significantly enriched among the DEGs. The Kyoto Encyclopedia of Genes and Genomes (KEGG) database, a comprehensive resource for understanding the high-level functions and interactions of biological systems, provides information spanning from the cellular level to entire ecosystems. It compiles molecular-level information, largely derived from genome sequencing and various high-throughput experimental approaches (http://www.genome.jp/kegg/). The clusterProfiler R package was also used to evaluate the statistical significance of DEGs in the context of KEGG pathways.

qRT-PCR analysis

Eight genes were randomly selected from the four pathways enriched in response to B. impatiens larval infestation, and their expression levels were validated using quantitative reverse transcription PCR (qRT-PCR), with the tomato actin gene serving as the internal reference. The primer sequences used are listed in Table S4. Total RNA extraction and qRT-PCR procedures were performed according to the methods outlined by Luo et al. (2024) [44].

Metabolites extraction and UHPLC-MS/MS analysis

Four leaf tissues (100 mg each) were collected individually from control plants and from tomato plants infested by B. impatiens larvae. Samples were homogenized in liquid nitrogen and resuspended in prechilled 80% methanol with vigorous vortexing to ensure thorough mixing. The samples were then incubated on ice for 5 min and centrifuged at 15,000 g for 20 min at 4 °C. The supernatant was collected, diluted with LC-MS grade water to achieve a final methanol concentration of 53%, transferred to new Eppendorf tubes, and centrifuged again under the same conditions. The final supernatant was used for UHPLC-MS/MS analysis.

UHPLC-MS/MS was performed using a Vanquish UHPLC system (ThermoFisher, Germany) coupled to either an Orbitrap Q Exactive™ HF or HF-X mass spectrometer (Thermo Fisher, Germany) at Novogene Co., Ltd. (Beijing, China). The samples were loaded onto a Hypersil Gold column (100 × 2.1 mm, 1.9 μm) and separated using a 12-minute linear gradient at a flow rate of 0.2 mL/min. Mobile phases consisted of eluent A (0.1% formic acid in water) and eluent B (methanol), applied in both positive and negative polarity modes, with the following gradient program: 2% B for 1.5 min, 85% B over 3 min, 100% B over 10 min, decrease to 2% B over 10.1 min, and hold at 2% B for 12 min. The Q Exactive™ HF mass spectrometer was operated in both positive and negative polarity modes, with the following settings: spray voltage of 3.5 kV, capillary temperature of 320℃, sheath and auxiliary gas flow rates of 35 psi and 10 L/min, respectively, S-lens RF level of 60, and auxiliary gas heater temperature of 350℃.

Measurement of antioxidant enzyme activities and MDA content

To quantitatively evaluate the oxidative stress response, the activities of superoxide dismutase (SOD), peroxidase (POD), and catalase (CAT), as well as the content of malondialdehyde (MDA), were measured in tomato leaves from both control and B. impatiens-infested groups. Leaf samples (approximately 0.1 g) were homogenized in 1 mL of pre-chilled extraction buffer and centrifuged at 12,000 × g for 10 min at 4 °C. The supernatant was collected for all subsequent assays. All measurements were performed using commercial kits (Aidisheng Biological Technology Co., Ltd.) following the manufacturer’s protocols, with three biological replicates per treatment. SOD activity was determined using the WST-8 method, which detects the inhibition of superoxide anion-induced formazan dye formation at 450 nm. One unit of SOD was defined as the amount of enzyme required to achieve 50% inhibition of the reaction, and activity was expressed as U per gram of fresh weight (U/g FW). POD activity was assayed by monitoring the oxidation of guaiacol at 470 nm. One unit of POD was defined as a 0.01 increase in absorbance per minute per gram of fresh weight (∆OD₄₇₀/min/g FW). CAT activity was measured using a chromogenic probe that reacts with residual H₂O₂ after catalase-mediated decomposition, with absorbance recorded at 510 nm. The decrease in H₂O₂ was calculated using a standard curve. One unit of CAT was defined as the amount of enzyme catalyzing the decomposition of 1 µmol of H₂O₂ per minute per gram of fresh weight (µmol/min/g FW). MDA content was determined using the thiobarbituric acid (TBA) reaction method. The absorbance was measured at 532 nm and 600 nm, and the MDA concentration was calculated using an extinction coefficient of 155 mM⁻¹·cm⁻¹, with results expressed as nmol per gram of fresh weight (nmol/g FW). All data are presented as mean ± standard deviation (SD). Statistical significance was determined by one-way analysis of variance (ANOVA) followed by Tukey’s honestly significant difference (HSD) post-hoc test, with p < 0.05 considered statistically significant.

Metabolomics data analysis

The raw data generated by UHPLC-MS/MS were analyzed using Compound Discoverer 3.3 (CD3.3, ThermoFisher) for peak alignment, peak detection, and metabolite quantification. The key parameters were configured as follows: the peak area was adjusted relative to the first QC sample, with a mass tolerance of 5 ppm, a signal intensity tolerance of 30%, and a defined minimum intensity threshold applied, among other settings. Subsequently, peak intensities were normalized to the total spectral intensity. This normalized data was then used to predict molecular formulas by taking into account additive ions, molecular ion peaks, and fragment ions. Peak alignment and annotation were performed using the mzCloud (https://www.mzcloud.org/), mzVault, and MassList databases to obtain accurate qualitative and relative quantitative information. Statistical analyses were conducted using R (v3.4.3), Python (v2.7.6), and CentOS (v6.6). For data not conforming to a normal distribution, standardization was applied according to the formula: sample raw quantification value / (sum of sample metabolite quantification values / sum of QC1 sample metabolite quantification values) to obtain relative peak areas. Metabolites exhibiting coefficients of variation (CVs) of relative peak areas greater than 30% in QC samples were excluded from further analyses.

The identified metabolites were annotated using the KEGG (https://www.genome.jp/kegg/pathway.html), HMDB (https://hmdb.ca/metabolites), and LIPIDMaps (http://www.lipidmaps.org/) databases. For multivariate analysis, metabolomic data were transformed via the metaX software, followed by principal component analysis (PCA) and partial least squares discriminant analysis (PLS-DA) to extract variable importance in projection (VIP) values for each metabolite. Following this, univariate analyses were performed using t-tests and fold change (FC) values. Metabolites were defined as differentially expressed if they satisfied the criteria VIP > 1 and P < 0.05, and metabolic pathways were considered significantly enriched when P < 0.05.

Results

Morphology of tomato plants after feeding by Bradysia impatiens larvae

To evaluate the damage caused by B. impatiens larvae (BILA) on tomato plants, we examined the macroscopic symptoms following larval infestation. Larvae introduced onto the soil surface initially fed on the stem bases, causing the seedlings to lodge (fall over) and become prostrate on the soil surface. Subsequently, larvae continued feeding on the leaf tissues of the lodged plants and also attacked the root systems. Compared with the control group (CK), infested leaves developed irregularly shaped necrotic patches and chlorosis that expanded over time. Most of the affected leaves eventually wilted and senesced. In some plants, severe stem and root damage led to overall plant decline and, in extreme cases, plant death (Fig. 1A). These observations indicate that B. impatiens larval feeding on multiple tissues (stems, leaves, and roots) can severely threaten the overall health and survival of the plant. To quantitatively characterize the phenotypic differences between CK and BILA plants under herbivory stress, we first assessed leaf damage at the single-leaf level. BILA exhibited a significantly higher proportion of damaged leaf area than CK (Fig. 1B). In line with this pronounced leaf injury, BILA plants showed markedly suppressed vegetative growth. Both the shoot fresh weight and the seedling survival rate of BILA were significantly reduced compared with those of CK (Fig. 1C, D). Collectively, these quantitative phenotypic results provide robust biological evidence for the herbivory-induced transcriptomic and metabolic remodeling observed in BILA plants, and validate the biological relevance of the omics changes reported in this study.

Fig. 1.

Phenotypic characterization of tomato plants under B. impatiens larval infestation. A Representative qualitative symptoms of infested tomato plants, showing leaf necrosis, chlorosis, and seedling lodging caused by larval feeding on stems, leaves and roots. B Damaged area ratio of detached single leaves. C Aboveground fresh weight per plant. D Seedling survival rate. Data are mean ± SD (n ≥ 15). Statistical significance was determined by one-way analysis of variance (ANOVA). ***P < 0.001. CK, control, BILA, B. impatiens infestation

Transcriptome sequencing analysis and identification of DEGs

The clean read data were generated through a filtering process that involved assessing sequencing error rates and evaluating GC content distribution. The Q20% scores (representing a sequencing error rate below 0.01) ranged from 99.34% to 99.43%, while the Q30% scores (corresponding to a sequencing error rate below 0.001) ranged from 97.07% to 97.53% (Table S1). These high-quality reads were subsequently used for downstream transcriptome analysis. Over 97% of the sequencing reads successfully aligned to the reference genome (Table S2), demonstrating the high reliability of the sequencing data.

To evaluate the reproducibility of biological replicates and global transcriptional divergence between groups, we performed principal component analysis (PCA) and Pearson correlation analysis based on normalized gene expression data. The three-dimensional PCA plot revealed clear separation between the CK and BILA groups, whereas biological replicates within each group clustered closely (Fig. 2A). Consistently, the Pearson correlation coefficients of gene expression within each group were all greater than 0.88 (Fig. 2B), confirming strong intra-group reproducibility and substantial transcriptional differences between the two experimental groups.

Fig. 2.

Transcriptional profiling and global gene expression variation between CK and BILA groups. A Three-dimensional principal component analysis (PCA) plot of normalized gene expression data from tomato leaves under control (CK) and (B) impatiens larval feeding (BILA) conditions. Each point represents a biological replicate, with clear separation observed between the two groups and close clustering within each group. B Pearson correlation heatmap showing the correlation coefficients of gene expression among all samples. Coefficients within each group were greater than 0.88, indicating high reproducibility among biological replicates and substantial transcriptional divergence between the CK and BILA groups

Utilizing these sequencing data, RNA-Seq analysis was performed to identify the DEGs between tomato leaves subjected to B. impatiens larval feeding and control conditions. The Venn diagram illustrating the DEGs reveals a core set of 18,373 co-expressed genes, along with 679 uniquely expressed genes in CK and 716 uniquely expressed genes in the BILA group (Fig. 3A). A total of 2,973 genes were found to be significantly differentially expressed between the CK and BILA groups. Of these, 1,988 were upregulated (P < 0.05), and 985 were downregulated (P < 0.05) in response to larval feeding (Fig. 3B). The substantial number of DEGs indicates a pronounced reprogramming of the tomato transcriptome following B. impatiens larval infestation (Fig. 3C), reflecting extensive transcriptional changes in the host plant in response to herbivory.

Fig. 3.

Statistical analysis of differentially expressed genes (DEGs) between plants fed by B. impatiens larvae and the control group. A Venn diagram showing the number of DEGs. B Volcano plot depicting up- and downregulated genes in CK versus B. impatiens larvae-infested plants. C Cluster heat map of DEGs

GO and KEGG analyses

GO enrichment analysis was conducted to elucidate the biological roles of DEGs, which were categorized into biological process, cellular component, and molecular function. Among these, the molecular function category contained the highest number of genes (Fig. S1). Overall, a greater number of DEGs were upregulated compared to those that were downregulated (Fig. S2). In response to B. impatiens larval attack, several functional categories—including “response to stress,” “transcription regulator activity,” “DNA-binding transcription factor activity,” “molecular function regulator,” and “enzyme regulator activity”—showed significantly elevated expression levels (Fig. 4A and S2).

Fig. 4.

GO and KEGG analyses of DEGs in leaves under B. impatiens larval attack and CK. A GO enrichment analysis of DEGs between (B) impatiens-infested and CK leaves. B KEGG pathway enrichment analysis of DEGs between B. impatiens-infested and CK leaves

KEGG pathway analysis was performed to explore the influence of B. impatiens larval attack on metabolic pathways associated with DEGs. The results revealed significant enrichment in multiple pathways, notably “phenylpropanoid biosynthesis,” “plant-pathogen interaction”, “plant hormone signal transduction”, and “MAPK signaling pathway-plant” (Fig. 4B and S3).

Expression analysis and qRT-PCR validation of representative genes involved in “MAPK signaling pathway-plant”, “plant hormone signal transduction”, “plant-pathogen interaction”, and “phenylpropanoid biosynthesis” under B. impatiens larval attack

To further elucidate the transcriptional dynamics of key defense-related pathways to B. impatiens larval attack, we examined the expression patterns of representative DEGs involved in the four significantly enriched KEGG pathways (Fig. 5A-D). In the “MAPK signaling pathway-plant” a marked alteration was observed in the expression of 16 genes (log₂FC > 2), all of which were upregulated (Fig. 5A). In the “plant hormone signal transduction pathway” 26 genes showed significant transcriptional alterations, including 20 genes that were upregulated and 6 that were downregulated following larval attack (Fig. 5B). In the“plant-pathogen interaction pathway” 27 defense-related genes were identified, the expression levels of 26 genes, including PR1-1a1 and WRKY31, were upregulated, whereas only LOC138342247 was significantly downregulated (Fig. 5C). In the “phenylpropanoid biosynthesis pathway” 35 genes exhibited significantly upregulated expression, while only LOC101247048 and LOC101266953 exhibited downregulated expression (Fig. 5D). These results suggest that upon B. impatiens infestation, tomato plants strengthen their defense capacity by activating key genes in the MAPK cascade, hormone signaling, pathogen interaction, and phenylpropanoid biosynthetic pathways, with the majority of representative genes showing significant upregulation.

Fig. 5.

Clustering of differentially expressed genes (DEGs) associated with four significantly enriched KEGG pathways. A MAPK signaling pathway-plant. B Plant hormone signal transduction pathway. C Plant-pathogen interaction pathway. D Phenylpropanoid biosynthesis pathway. E qRT-PCR validation of eight randomly selected DEGs in plants after attack by B. impatiens larvae

To validate the RNA-sequencing results, qRT-PCR analyses were performed for two representative genes from each of the four pathways. The qRT-PCR results were consistent with the RNA-seq expression profiles, confirming the reliability of the sequencing data (Fig. 5E). Collectively, these findings indicate that B. impatiens larval attack triggers orchestrated induction of the MAPK signaling, hormone signal transduction, pathogen defense, and phenylpropanoid metabolism pathways, potentially contributing to enhanced resistance in the host plant.

Quality control of metabolomic data

Metabolomic profiling of tomato plants following infestation by B. impatiens larvae was conducted using UHPLC-MS/MS coupled with a high-resolution mass spectrometer to investigate metabolic alterations induced by larval attack. An unsupervised PCA was performed separately for the datasets acquired in positive and negative ionization modes. The resulting PCA score plots, encompassing 11 samples (3 quality control (QC) samples and 8 biological replicates), illustrated clear distinctions between the infested and control groups. In the negative ion mode, the first two principal components accounted for 59.45% of the total variance (PC1: 46.50%, PC2: 12.95%). A pronounced separation between the infested and control groups was observed primarily along PC1, with the infested group clustering in the positive region and the control group in the negative region (Fig. 6A). Similarly, in the positive ion mode, the PCA model revealed a clear segregation along PC1, which explained 40.28% of the total variance (53.72% cumulative for PC1 and PC2) (Fig. 6B). These contrasting PCA patterns suggest that distinct groups of metabolites, exhibiting ionization-dependent responses, were significantly modulated following B. impatiens larval infestation. This finding reflects widespread rewiring of metabolic homeostasis in tomato plants triggered by larval attack. Furthermore, the tight clustering of the QC samples near the origin in both ionization modes confirmed high analytical reproducibility, thus validating the reliability of the observed metabolic differences.

Fig. 6.

Identification of metabolites accumulated in plants under B. impatiens larval attack. A PCA score plot of metabolite profiles in negative ion mode for CK, B. impatiens-infested (BILA), and quality control (QC) samples. B PCA score plot of metabolite profiles in positive ion mode for CK, B. impatiens-infested (BILA), and QC samples. C Classification of metabolites based on the HMDB database. D Classification of metabolites based on the LIPIDMaps database

A total of 3854 metabolites were identified in tomato leaves infested by B. impatiens larvae, with 2084 detected in positive ion mode and 1711 in negative ion mode (Fig. S4). The identified metabolites were annotated using the HMDB and LipidMaps databases. Within the HMDB database, 908 metabolites were classified into 16 categories, with “Lipids and lipid-like molecules” representing the largest group (458 metabolites), while “Hydrocarbon derivatives” comprised only a single metabolite (Fig. 6C). The LipidMaps database annotated 482 metabolites into seven main classes. The “Fatty acyls” class was the most diverse, encompassing 11 subclasses and 221 metabolites. The next most abundant classes, “Polyketides” and “Sterol Lipids”, contained 92 and 74 metabolites, respectively, each further divided into five subclasses. In contrast, the “Glycerolipids” class contained only six metabolites within subclasses, while the “Sphingolipids” class was represented solely by the “Sphingoid bases” subclass, comprising eight metabolites (Fig. 6D). Collectively, these results demonstrate that infestation by B. impatiens larvae triggers broad-scale remodeling of leaf metabolic profiles in tomato leaves, with particularly pronounced changes in lipid metabolism. These findings provide valuable metabolomic insights into the molecular mechanisms underlying plant-insect interactions.

Differentially expressed metabolites (DEMs) and functional enrichment

A total of 1,462 metabolites were differentially expressed between the two experimental groups. Of these, 733 metabolites exhibited significant alterations in the negative ion mode, including 237 upregulated and 496 downregulated (Fig. 7A). In the positive ion mode, 729 metabolites showed significant alterations, with 311 upregulated and 418 downregulated (Fig. 7B).

Fig. 7.

Statistical analysis of differentially expressed metabolites (DEMs) in tomato leaves under control conditions (CK) and B. impatiens larval attack, along with functional enrichment and correlation analyses. Comparison of upregulated and downregulated metabolites under CK and B. impatiens larvae-infested plants. A Metabolites in negative ion mode and B Metabolites in positive ion mode. Functional enrichment and correlation analyses of DEMs under CK and B. impatiens larval attack. C Metabolites in negative ion mode. D Metabolites in positive ion mode. In panels C and D, the top 20 KEGG enrichment pathways with the smallest P-values are displayed by default. Dots located farther to the right indicate a higher enrichment of differential metabolites in the corresponding pathway. The color and size of the dots correspond to the P-value and number of differential metabolites, respectively. Darker, larger dots indicate smaller P-values and greater numbers of metabolites

KEGG pathway enrichment analysis of the DEMs revealed the top 20 enriched pathways (Table S3), indicating profound reshaping of leaf metabolic networks following herbivory by B. impatiens larvae. In the negative ion mode, significantly enriched pathways were primarily associated with energy metabolism (including carbon metabolism, the TCA cycle, and the pentose phosphate pathway), antioxidant responses (e.g., ascorbate and aldarate metabolism, and cysteine and methionine metabolism), and the biosynthesis of defense-related compounds (such as terpenoid backbone biosynthesis, diterpenoid biosynthesis, and phenylalanine metabolism) (Fig. 7C). In the positive ion mode, enrichment was observed in pathways related to nitrogen-containing compound metabolism (e.g., tryptophan metabolism and indole alkaloid biosynthesis), lipid signaling (specifically α-linolenic acid metabolism), cofactor metabolism (notably vitamin B₆ metabolism), and specialized metabolite biosynthesis (including stilbenoid, diarylheptanoid, and gingerol biosynthesis) (Fig. 7D). Collectively, these results indicate that herbivory by B. impatiens larvae induces a synergistic defensive cascade in tomato plants, characterized by energy reallocation, activation of antioxidant systems, and enhanced production of diverse secondary metabolites.

Correlation and enrichment analyses between DEGs and DEMs

Metabolomic and transcriptomic analyses were conducted to compare samples from the two experimental groups, providing a comprehensive understanding of the physiological responses of tomato plants under B. impatiens larval attack. The results revealed that, in the negative ion mode, 47 metabolic pathways were significantly enriched. In contrast, in the positive ion mode, 18 metabolic pathways were significantly enriched. Specifically, in the negative ion mode, three metabolites and 74 genes were significantly enriched in the “Plant hormone signal transduction pathway”, one metabolite and 57 genes in the “Phenylpropanoid biosynthesis pathway”, and one metabolite and 19 genes in the “Phenylalanine metabolism pathway” (Fig. 8A). In the positive ion mode, one metabolite and 11 genes were enriched in the “α-Linolenic acid metabolism pathway”, one metabolite and 10 genes in the “Galactose metabolism pathway”, and one metabolite and 15 genes in the “Phosphatidylinositol signaling system” (Fig. 8B). Moreover, pathways associated with “Cysteine and methionine metabolism”, “Diterpenoid biosynthesis”, “Stilbenoid, diarylheptanoid, and gingerol biosynthesis”, and “Arginine and proline metabolism” were also significantly enriched in plants infected by B. impatiens larvae.

Fig. 8.

Association analysis of KEGG pathways with differentially expressed genes (DEGs) and differentially expressed metabolites (DEMs) in tomato leaves following attack by B. impatiens larvae. A Negative ion mode. B Positive ion mode

Biosynthesis of plant hormone signal transduction and α-linolenic acid metabolism

Analysis of the DEGs and DEMs following B. impatiens larval infestation revealed significant enrichment in the plant hormone signal transduction and α-linolenic acid metabolism pathways. Therefore, we further explored the expression patterns and network interactions of the DEGs and DEMs associated with these two pathways. Among the 74 DEGs associated with plant hormone signal transduction, 44 showed upregulated expression, whereas 30 were downregulated after B. impatiens infestation (Fig. 9A). Additionally, all 11 DEGs linked to α-linolenic acid metabolism were upregulated (Fig. 9B).

Fig. 9.

Analysis of differentially expressed genes (DEGs) and differentially expressed metabolites (DEMs) in tomato leaves following B. impatiens larval infestation. A DEGs involved in plant hormone signal transduction. B DEGs associated with α-linolenic acid metabolism. C DEMs related to plant hormone signal transduction and α-linolenic acid metabolism

Regarding the metabolic changes, in the plant hormone signal transduction pathway, the levels of gibberellin A1 and gibberellin A4 were significantly downregulated, while the bioactive defense hormone (-)-Jasmonoyl-L-isoleucine (JA-Ile) exhibited a marked upregulation (Fig. 9C). This inverse correlation between gibberellins and JA-Ile reflects the classical growth-defense trade-off in tomato plants, wherein the activation of jasmonate-mediated defense responses suppresses gibberellin biosynthesis to reallocate resources from vegetative growth to herbivore resistance.

In the α-linolenic acid metabolism pathway, the precursor α-linolenic acid was downregulated, likely due to its rapid consumption for downstream defense signaling. Concurrently, the downstream products 9,10-EOT, 9(S)-HpOTrE, and Heptadecatrienal were all upregulated (Fig. 9C), indicating that the lipoxygenase (LOX) cascade and oxidative burst pathways were vigorously activated upon insect infestation. Collectively, these results demonstrate that B. impatiens feeding triggers a systemic reprogramming of phytohormone and fatty acid metabolism, shifting the plant from a growth-oriented state to a defensive posture.

Accumulation of ROS in B. impatiens larvae-infested plants

Upon detection of insect-derived elicitors, plants activate defense responses, including calcium ion (Ca²⁺) influx, plasma membrane depolarization, activation of mitogen-activated protein kinases (MAPKs), stimulation of NADPH oxidase, ROS generation, and nitric oxide (NO) synthesis [45]. To assess ROS accumulation in tomato leaves following B. impatiens larval infestation, we performed NBT and DAB staining to detect superoxide (O₂•⁻) and hydrogen peroxide (H₂O₂) levels. The results revealed significant accumulation of O₂•⁻ and H₂O₂ in response to larval feeding (Fig. 10A). To quantify this oxidative response, we measured the activities of three key antioxidant enzymes, SOD, POD, and CAT, as well as the content of MDA, a reliable marker of membrane lipid peroxidation. Larval feeding significantly elevated SOD, POD, and CAT activities by 362.0%, 171.7%, and 39.3%, respectively, relative to the control group. Concurrently, MDA content increased by 72.5% in infested plants, confirming extensive membrane lipid peroxidation (Fig. 10B). Collectively, these quantitative data corroborate the histochemical staining results and demonstrate that B. impatiens herbivory induces a strong oxidative burst, accompanied by the activation of enzymatic antioxidant systems to counteract oxidative damage.

Fig. 10.

Histochemical staining and quantitative detection of reactive oxygen species (ROS) and oxidative stress-related indicators in CK and BILA tomato leaves after B. impatiens larval infestation. A Bright-field images, DAB staining for hydrogen peroxide (H₂O₂), and NBT staining for superoxide anion radical (O₂•⁻) in detached functional leaves of CK and infested BILA seedlings. Dark brown precipitates in DAB staining represent accumulated H₂O₂; blue formazan precipitates in NBT staining indicate O₂⁻ accumulation. Scale bar = 1 cm. B Quantitative analysis of antioxidant enzyme activities and MDA content. From left to right: Catalase (CAT) activity, superoxide dismutase (SOD) activity, peroxidase (POD) activity, and malondialdehyde (MDA) content. Data are shown as mean ± SD, n = 3 biological replicates. Asterisks denote significant differences relative to CK (**P < 0.01, ***P < 0.001, statistical significance was determined by one-way analysis of ANOVA). CK, untreated control group. BILA, B. impatiens larval infestation group

Our results indicate significant alterations in the expression of ERF, bHLH, and NAC transcription factor family genes within the enriched plant hormone signal transduction and α-linolenic acid metabolism pathways (Fig. 9A, B). Specifically, among ERF family genes, GH3.2 and EREB were upregulated, whereas GH3-18 was downregulated. In the bHLH family, LOC138347874, LOC101250310, MYC2, and LOC101250954 were upregulated, while LOC101251725 and LOC101252930 were downregulated. All NAC family genes analyzed, including LOC101260391, LOC101253664, LOC100316879, LOC101251665, and OPR3 were upregulated (Fig. 9A, B). Collectively, these findings indicate that B. impatiens larval infestation in tomato leaves induces significant ROS accumulation, potentially mediated by the coordinated transcriptional regulation of ERF, bHLH, and NAC family genes.

Discussion

Tomato plants employ complex defense mechanisms to resist herbivore attacks, combining both physical barriers and inducible biochemical responses. In this study, a combined transcriptomic and metabolomic approach was used to comprehensively investigate the molecular and metabolic remodeling in tomato leaves following infestation by B. impatiens larvae. The results revealed that herbivory activates a hierarchically organized defensive cascade, mediated by key signaling pathways and resulting in the accumulation of defensive metabolites and ROS (Fig. 11). In this study, B. impatiens larvae exhibited a sequential feeding behavior, attacking stem bases (causing lodging), followed by leaves of prostrate plants, and finally roots after penetrating the soil. The observed transcriptional and metabolic changes in tomato leaves therefore represent an integrated response to multiple modes of attack: direct leaf tissue damage from larval feeding, systemic signals arising from stem wounding and lodging-induced stress, and root-derived systemic signals triggered by below-ground herbivory. This multifaceted pattern of larval feeding is consistent with the polyphagous and opportunistic behavior of sciarid larvae, which exploit accessible plant tissues depending on environmental conditions and larval mobility.

Fig. 11.

A schematic diagram illustrating the tomato plant’s response to larval infestation by Bradysia impatiens

Transcriptomic analysis revealed significant alterations in gene expression after B. impatiens attack, identifying 2,973 DEGs. Enrichment analyses highlighted the involvement of critical pathways, including the “MAPK signaling pathway-plant”, “plant hormone signal transduction”, “plant-pathogen interaction”, and “phenylpropanoid biosynthesis”. The upregulation of genes in the MAPK cascade reflects a conserved early stress response, facilitating the activation of downstream transcription factors and the induction of defense-related genes [8, 46]. Additionally, the enrichment of plant hormone signaling pathways, particularly those linked to JA and ET, underscores their pivotal role in coordinating defenses against herbivorous insects [47, 48]. Furthermore, activation of the “plant-pathogen interaction” pathway suggests a broad-spectrum defense response, potentially priming the plant for secondary pathogen infestation commonly vectored by sciarid flies [20].

Metabolomic profiling corroborated the observed transcriptional changes, revealing broad-scale remodeling of core metabolic pathways. A total of 1,462 metabolites exhibited differential accumulation. KEGG enrichment analysis of these metabolites indicated significant perturbations in energy metabolism pathways (e.g., the TCA cycle), antioxidant responses (e.g., ascorbate metabolism), and the biosynthesis of specialized defensive compounds (such as terpenoids and phenylpropanoids). The pronounced enrichment of “Lipids and lipid-like molecules” aligns with the essential role of fatty acid derivatives, including those from the α-linolenic acid metabolism pathway, as precursors for JA biosynthesis and other signaling molecules [36]. Our results showing increased levels of α-linolenic acid and associated oxylipins (e.g., 9,10-EOT, 9(S)-HpOTrE) further support the activation of the JA biosynthesis pathway, a hallmark of plant defense against herbivory. In the phenylpropanoid biosynthesis pathway, the coordinated upregulation of key enzyme-encoding genes, including PAL and 4CL, suggests an enhanced metabolic flux toward lignin precursors and flavonoid compounds. Phenylpropanoid metabolism is a central secondary metabolic pathway in plants, and its products serve a dual function in herbivore defense [49, 50]. On one hand, lignin, as a major structural component of the cell wall, reinforces the physical barrier against larval penetration through lignification. On the other hand, flavonoids and phenolic acids act as chemical defense compounds that can directly inhibit insect feeding, interfere with digestive enzyme activities, or impair reproductive capacity [49]. 4CL functions as a key branch-point enzyme in the phenylpropanoid pathway, and its expression level significantly influences the biosynthetic flux toward downstream defense metabolites [51]. The pronounced upregulation of PAL and 4CL observed in this study, together with the enrichment of the entire phenylpropanoid biosynthesis pathway, suggests that these regulatory nodes play a central role in tomato’s response to B. impatiens herbivory, coordinating the reinforcement of physical barriers with the production of chemical defense compounds to enhance overall resistance.

Integrated transcriptomic and metabolomic analysis offers a more cohesive understanding of the defense responses. The co-occurrence of genes and metabolites in key pathways, such as “plant hormone signal transduction” and “α-linolenic acid metabolism,” highlights the tight coupling between transcriptional regulation and metabolic outcomes. The jasmonic acid (JA) signaling pathway occupies a central position in plant defense. In the biosynthesis of JA-Ile, OPR3 catalyzes the reduction of 12-oxo-phytodienoic acid (OPDA), representing a key rate-limiting step in JA biosynthesis [52]. The bHLH transcription factors MYC2/3/4, together with their JAZ-mediated derepression, constitute the core transcriptional switch of the JA signaling pathway [53]. In this study, the significant upregulation of OPR3 and MYC2, coupled with the marked accumulation of JA-Ile, collectively confirms the robust activation of JA-mediated defense signaling in response to B. impatiens infestation.

It is noteworthy that JA-Ile functions as a rapidly turning-over signaling molecule, whose steady-state levels are governed by the dynamic equilibrium between biosynthesis and metabolic degradation [54]. Under sustained herbivore attack, JA-Ile is rapidly synthesized and subsequently cleared through hydrolysis or conversion into downstream products — a “rapid synthesis-turnover” pattern documented in other plant-insect interaction systems [54, 55]. The concurrent upregulation of JA biosynthetic genes (LOXD, AOC, OPR3) and signaling-responsive genes (MYC2), along with the net accumulation of JA-Ile, collectively indicates an overall activation state of the JA signaling pathway. Furthermore, a positive feedback loop involving MYC2 and specific ERF members, in which MYC2 activates ERF transcription and ERFs in turn upregulate JA biosynthetic genes, may further amplify JA signaling outputs [48]. This regulatory circuitry enables plants to mount and sustain robust defense responses upon insect infestation.

This study demonstrates that tomato leaves infested with B. impatiens larvae exhibit a significant increase in ROS levels. DAB and NBT staining confirmed elevated accumulation of H₂O₂ and O₂•⁻, respectively. ROS are well-established as critical signaling molecules in plant defense, enabling rapid responses to environmental stimuli. They play essential roles in detecting both abiotic and biotic stress, integrating diverse environmental signals, and activating stress-responsive networks. Through these functions, ROS contributes to the establishment of plant defense systems and overall resilience, supporting acclimatization and adaptation [56]. Genes encoding members of the ERF, bHLH, and NAC transcription factor families were significantly upregulated, indicating the presence of a transcriptional regulatory framework governing ROS dynamics. These transcription factors are known regulators of ROS metabolism and stress response pathways [40, 42, 43]. Specifically, ERF transcription factors can modulate RBOH gene expression, which drives ROS production, while NAC and bHLH family members regulate the antioxidant system, thereby fine-tuning ROS homeostasis during defense [47]. Based on the coordinated upregulation of ERF, bHLH, and NAC transcription factor families in this study and their known functions in ROS regulation, we propose a ROS-transcription factor homeostatic model. At the early stage of B. impatiens herbivory, ERF transcription factors (such as the upregulated EREB and GH3.2 in this study) activate Rboh-mediated ROS burst, initiating defense signal transduction [40, 41]. Simultaneously, bHLH (such as MYC2) and NAC family members upregulate antioxidant enzyme genes (e.g., genes encoding superoxide dismutase, catalase, and peroxidase) to prevent excessive ROS accumulation that could cause cellular damage [42, 43]. This “production-scavenging” balance mechanism enables ROS to serve as effective defense signals to activate downstream defense responses while avoiding oxidative damage to host cells. Furthermore, ROS themselves can further activate MAPK cascades and transcription factor activity, forming a signal amplification loop that enhances the efficiency and sustainability of the overall defense response [56]. The concurrent upregulation of SOD, POD, and CAT activities, alongside the elevated MDA content (Fig. 10B), indicates that despite the activation of enzymatic antioxidant defenses, ROS production exceeded the scavenging capacity, resulting in measurable oxidative damage to membrane lipids. This is consistent with the metabolomic observation of α-linolenic acid depletion and accumulation of oxylipin derivatives (Fig. 9C), further supporting the notion that oxidative stress is a central component of tomato’s defense response to B. impatiens. Concurrent activation of phenylpropanoid biosynthesis suggests reinforcement of both physical and chemical barriers. Phenylpropanoids serve as precursors for lignin, which strengthens cell walls, and for various antimicrobial and antifeedant phenolic compounds [30]. Their accumulation likely reduces tissue palatability and digestibility for larvae while fortifying the plant against opportunistic pathogens.

Overall, this multi-omics investigation reveals that tomato defense responses to B. impatiens are not governed by a single mechanism but instead involve a coordinated interplay of multiple processes. All these responses fall into the category of induced resistance: they are absent in undamaged control plants and activated only after larval feeding is perceived via HAMPs/DAMPs. The persistent rewiring of energy, lipid and secondary metabolic pathways detected at one week post-infestation further constitutes an adaptive defense strategy, enabling tomatoes to reallocate resources from vegetative growth to anti-herbivore resistance and tolerate prolonged pest stress. These include rapid herbivore detection, activation of MAPK and hormone signaling pathways (notably JA and ET), and extensive transcriptional reprogramming mediated by ERF, bHLH, and NAC transcription factors. This is accompanied by a metabolic shift toward the production of defensive specialized metabolites, such as phenylpropanoids and terpenoids, and signaling molecules like oxylipins. ROS are strategically deployed as both signaling molecules and direct defensive agents. This integrated response is likely designed to deter herbivores, minimize tissue damage, and prepare the plant for future biotic stress. These findings provide a valuable molecular resource for developing resistance-enhancing strategies, including marker-assisted breeding and targeted elicitor applications, to sustainably manage sciarid fly populations in tomato cultivation.

From the perspective of translational breeding and pest management, we further pinpoint two representative hub genes and two key metabolite clusters uncovered in this study as candidate biomarkers and resistance engineering targets. The bHLH transcription factor MYC2 and JA biosynthetic rate-limiting gene OPR3 were robustly upregulated upon larval infestation, serving as central regulatory nodes linking upstream MAPK-ROS signaling and downstream secondary metabolic defense. These two genes can be utilized as DNA molecular markers to screen for high-resistance tomato germplasm, and transgenic overexpression of MYC2 or OPR3 is a feasible strategy to constitutively activate jasmonate-mediated anti-insect defense. For metabolic biomarkers, significantly accumulated α-linolenic acid oxylipins (9(S)-HpOTrE, 9,10-EOT) and phenylpropanoid phenolic compounds can be used as physiological indicators to rapidly quantify the defense activation level of tomato seedlings. Meanwhile, exogenous elicitors that trigger phenylpropanoid and oxylipin biosynthesis can be developed to achieve ecological control of B. impatiens in greenhouses.

A key limitation of this study is the use of a single time point (one week post-infestation) for both transcriptomic and metabolomic analyses. Plant defense responses to herbivory are inherently dynamic, progressing through multiple temporal phases: early perception of herbivore-associated molecular patterns (HAMPs) and rapid signaling events, such as Ca²⁺ influx, MAPK activation, and ROS burst, occurring within minutes to hours, transcriptional reprogramming and phytohormone accumulation (e.g., jasmonates) peaking within hours to days, and sustained metabolic shifts, including the accumulation of secondary metabolites and cell wall modifications, developing over days to weeks. By analyzing a single endpoint at one week, our data capture the integrated, late-stage outcome of these processes rather than the temporal sequence of regulatory events. Consequently, we cannot definitively distinguish between primary signaling responses and secondary or feedback-regulated changes. For instance, the observed upregulation of MAPK-related genes and accumulation of JA-Ile at one week post-infestation may represent sustained activation or secondary signaling waves rather than the initial inductive events. Despite this limitation, our integrated multi-omics dataset provides a valuable resource for identifying candidate genes, metabolites, and pathways involved in tomato defense against B. impatiens, and serves as a foundation for future time-course studies aimed at dissecting the regulatory hierarchy of the defense response. Furthermore, because larvae sequentially attacked stem bases, leaves (after lodging), and roots, the observed leaf responses integrate direct leaf damage with systemic signals from wounded stems and roots. Our current experimental design does not allow us to quantitatively partition the relative contributions of each tissue-specific attack to the overall leaf transcriptomic and metabolomic profiles. Future studies employing spatially separated or tissue-specific infestation (e.g., leaf-only, stem-only, or root-only) would be valuable for disentangling the distinct signaling pathways activated by each mode of attack. Another important limitation of the current multi-omics dataset is the absence of genetic functional validation for core hub genes identified in our integrated analysis. The coordinated upregulation of MYC2, OPR3, WRKY31, PR1, and key phenylpropanoid biosynthetic genes only provides correlational evidence linking these genes to tomato defense against B. impatiens, and cannot establish causal regulatory relationships. As master regulators of JA signaling, plant immunity, and physical/chemical defense metabolism, these genes require genetic confirmation of their in vivo biological functions against fungus gnat larvae. For instance, transgenic overexpression or knockout lines are required to verify whether elevated MYC2 and OPR3 can boost JA-Ile accumulation and enhance larval resistance, and silencing assays of WRKY31 and PR1 will clarify their roles in systemic defense and tolerance to secondary pathogens vectored by B. impatiens. In addition, functional characterization of pathway key enzymes such as PAL and 4CL is needed to confirm their contributions to lignification and anti-feeding phenolic metabolite production.

Conclusions

In this study, we employed an integrated transcriptomic and metabolomic approach to elucidate the complex defense mechanisms of tomato (Solanum lycopersicum) in response to infestation by B. impatiens larvae under the experimental conditions described above. Our results demonstrate that herbivory triggers extensive transcriptional reprogramming, with 2,973 differentially expressed genes significantly enriched in key pathways including MAPK signaling, plant hormone signal transduction, plant-pathogen interaction, and phenylpropanoid biosynthesis. Concurrently, metabolomic profiling revealed substantial shifts in 1,462 metabolites, highlighting enhanced energy metabolism, antioxidant defense, and the biosynthesis of defense-related compounds such as terpenoids and phenylpropanoids. Synchronized induction of α-linolenic acid metabolism and jasmonate signaling, along with the accumulation of ROS and the upregulation of ERF, bHLH, and NAC transcription factors, underscores a tiered defensive regulatory network. This defensive cascade integrates signal perception, hormonal regulation, metabolic reconfiguration, and ROS-mediated signaling to mitigate insect damage and prepare the plant for subsequent biotic challenges. Collectively, these coordinated molecular and metabolic shifts represent a combination of rapid induced resistance signals and long-term adaptive defense adjustments deployed by tomato against B. impatiens larval herbivory. Our findings provide a comprehensive multi-omics foundation for understanding tomato-insect interactions at an established stage of infestation (one week post-feeding), identifying key pathways and candidate targets for enhancing crop resistance. Future time-resolved studies building upon this foundation will be essential to dissect the temporal regulatory cascade and distinguish early signaling events from downstream adaptive responses. These insights could inform the development of sustainable management strategies, such as marker-assisted breeding or targeted elicitor applications, to reduce reliance on chemical pesticides and improve tomato resilience against sciarid fly infestations. The identified MYC2, OPR3, oxylipin and phenylpropanoid candidates lay solid molecular foundations for subsequent tomato resistance genetic modification and rapid resistance evaluation.

In future work, we will conduct systematic genetic functional validation of representative candidate genes screened in this study, including MYC2, OPR3, WRKY31, PR1, and rate-limiting enzymes in the phenylpropanoid biosynthesis pathway. We plan to generate CRISPR/Cas9 knockout and overexpression transgenic Micro-Tom tomato lines, combined with VIGS transient silencing technology. After infestation with B. impatiens larvae, we will compare phenotypic indexes including leaf damaged area ratio and seedling survival rate between transgenic materials and wild-type plants to evaluate pest resistance capacity. Biochemical indicators such as JA-Ile content, ROS accumulation, antioxidant enzyme activity and phenylpropanoid secondary metabolites will be quantified to dissect downstream metabolic reprogramming mediated by these genes. Moreover, yeast one-hybrid and dual-luciferase reporter assays will be applied to uncover transcriptional regulatory interactions between transcription factors and downstream defense genes.

Authors’ contributions

Conceptualization, Y.S. and C.L.; methodology, Y.S., Y.W., L.H., G.L., Q.L., S.M., M.F., H.C., and C.L.; software, Y.S. and C.L.; validation, Y.S.; data curation, C.L.; writing—original draft preparation, Y.S. and C.L.; writing—review and editing, Y.S., Y.W., L.H., G.L., Q.L., S.M., M.F., H.C., and C.L.; project administration, C.L.; funding acquisition, Y.S. and C.L. All authors read and approved the final manuscript.

Funding

This work was supported by grants from the National Natural Science Foundation of China (grant number 32560077, 32460457), the Yunnan Fundamental Research Projects (grant number 202501AU070172, 202401AU070002), the Special Basic Cooperative Research Innovation Programs of Qujing Science and Technology Bureau & Qujing Normal University (grant numbers KJLH2024ZD04), and the Program of Innovation Research Team from Qujing Normal University.

Data availability

Transcriptional and metabolic data were generated by Novomagic (Novogene Bioinformatics Technology Co., Ltd). The transcriptome data can be accessed from the Biological Project Database of the National Center for Biotechnology Information (NCBI), with the accession number PRJNA1400255. The original metabolomics data can be found in the supplementary data section labeled “Supplementary Data Metabolomics Raw”.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.