Long-term fungicide exposure promotes bacterial adaptation and increases virulence of foodborne pathogen Bacillus cereus in Caenorhabditis elegans

Abstract

The widespread contamination of agricultural soils with pesticides may facilitate the development of antimicrobial resistance and phenotypic adaptations in human pathogens. However, whether one of the most common and clinically significant foodborne pathogens, Bacillus cereus, exhibits such alterations upon long-term exposure to fungicides remains unclear. This study investigated the long-term effects of eight fungicides–chlorothalonil, propineb, tebuconazole, azoxystrobin, propiconazole, mancozeb, carbendazim, and triadimefon–on B. cereus, focusing on changes in biofilm formation, antibiotic susceptibility, and pathogenicity. Results indicated that chlorothalonil, propineb, and tebuconazole completely inhibited bacterial growth at high concentrations, while 30-day exposure to chlorothalonil 8 μM, tebuconazole 500 μM, and propineb 175 μM promoted bacterial adaptation without causing mutation. Adapted strains exhibited enhanced biofilm formation, reduced susceptibility to gentamicin and tetracycline, and greater lethality toward Caenorhabditis elegans. Further qRT-PCR analyses revealed that tebuconazole exposure upregulated genes associated with biofilm development (purC, purL, and calY), multidrug efflux pumps (smr) associated with antibiotic susceptibility, and virulence factors (nheC, hblA, hblC, and hblD). These molecular changes might be associated with the observed increases in biofilm formation, reduced antibiotic susceptibility, and heightened pathogenicity in B. cereus. Overall, these findings demonstrate that prolonged exposure to fungicides can promote adaptations in foodborne pathogens. This work underscores mechanistic insights into fungicide-pathogen interactions relevant to bacterial stress adaptations.

Graphical abstract

Highlights

• •Fungicide-adapted B. cereus shows increased biofilm, tolerance, and virulence.
• •qRT-PCR analysis shows upregulation of biofilm, efflux pump, and virulence genes.
• •B. cereus treated with tebuconazole (500 μM) showed the highest virulence.
• •Fungicide exposure is linked to antimicrobial resistance and pathogen virulence risks.

1. Introduction

More than 60% of global agricultural land is at risk of pesticide contamination (Tang et al., 2021). Pesticides are indispensable for ensuring food supply and security, yet they exert broad-scale detrimental effects on non-target organisms, posing widespread ecotoxicological risks across trophic levels (Pimentão et al., 2024; Wan et al., 2025). Due to their functional effects, pesticides may facilitate the development of antibiotic resistance in microbial communities by inducing mutations, eliciting adaptive stress responses, or promoting natural transformation of resistance elements (Zhao et al., 2026). In addition to acquired resistance elements, the intrinsic resistome of some pathogens may be enhanced by sublethal concentrations of antimicrobials, resulting in phenotypic resistance (Olivares et al., 2013). Accumulating evidence suggests links between fungicide contamination and antibiotic resistance. For example, azoxystrobin and carbendazim enrich the bacterial resistome in the nematode guts (Song et al., 2022; Zhang et al., 2022). In addition, tebuconazole promotes the spread of multidrug-resistant plasmids (Wang et al., 2024), whereas chlorothalonil facilitates metabolic adaptations in the soil microbiome (Liang et al., 2011), both of which enhance bacterial resilience to bactericidal agents. While non-antibiotic agents may contribute to the selection of antibiotic-resistant bacteria (Murray et al., 2024), it remains largely unclear whether the wide array of environmental fungicides induces a similar effect. Moreover, whether phenotypic resistance arises during fungicide exposure remains to be elucidated.

Abundant in agricultural soils, antibiotic-resistant bacteria may contaminate the human food chain, posing food safety concerns worldwide (Farrukh et al., 2025; Zheng et al., 2022). The prevalence of antibiotic-resistant pathogens in food samples may range from 77% to 97% across continents (Tao et al., 2022). Among foodborne pathogens, the Gram-positive, spore-forming Bacillus cereus is prevalent at 36-45% in dairy products, vegetables, beans, and cereals (Rahnama et al., 2023). B. cereus is a spore- and toxin-forming human pathogen that causes a wide array of devastating and potentially lethal clinical conditions, including food poisoning, eye infection, anthrax-like progressive pneumonia, fulminant sepsis, and central nervous system infections (Bottone, 2010). Able to adapt to various environmental stresses and produce various biofilms via differing architectures and mechanisms (Majed et al., 2016), B. cereus represents one of the most versatile pathogens, which has been used for bioremediation of pesticide-contaminated environments (Elshikh et al., 2022; Rehman et al., 2024). However, the development of antimicrobial resistance in this species is concerning, as evidenced by the emergence of the multidrug-resistant (MDR) B. cereus in hospital wastewater (Sayem et al., 2025). It also remains unclear whether adaptation to pesticides leads to heightened virulence in B. cereus.

The nematode Caenorhabditis elegans, a widely used and genetically tractable invertebrate infection model whose intestinal epithelium provides a relevant readout of pathogen colonization and killing, was suitable for assessing the in vivo virulence of adapted strains (Irazoqui et al., 2010; Pukkila-Worley and Ausubel, 2012; McGhee, 2013). Many human pathogens, including B. cereus (Rae et al., 2010), Pseudomonas aeruginosa (Tan et al., 1999), and Salmonella enterica subspecies enterica serovar Typhimurium (S. Typhimurium) (Aballay et al., 2000), cause intestinal infection in C. elegans, underscoring its utility in elucidating bacterial virulence and host immune responses (Balla and Troemel, 2013). By examining changes in host survival and intestinal colonization, we will gain insights into how prolonged exposure to fungicides will potentially alter the virulence of B. cereus.

Herein, this study aims to investigate the effects of fungicides on bacterial growth, to evaluate whether long-term exposure induces changes in adaptability, antibiotic resistance, and pathogenicity, and to elucidate the underlying molecular mechanisms. Furthermore, pathogenicity was assessed in vivo using C. elegans as an infection host. This study provides novel insights into the possible adaptation of B. cereus and the potential emergence of antimicrobial-resistant, highly virulent phenotypes.

2. Materials and methods

2.1. Chemicals

Eight fungicides, including chlorothalonil (CHT, >98%, Chemscene CS-0034415), propineb (PRO, >98%, MedChemExpress HY-119630), tebuconazole (TEB, 99.63%, MedChemExpress HY-B0852), azoxystrobin (AZX, ≥98%, ChemScene CS-0012863), propiconazole (PCZ, >95%, Cayman Chemical, 18853), mancozeb (MCZ, 85.5%, MedChemExpress HY-B0854), carbendazim (CBZ, 97%, Sigma-Aldrich 378674), and triadimefon (TDM, >98%, Sigma-Aldrich 45693) were investigated in this study. The antibiotic tetracycline (>98%, Sigma-Aldrich 87128) was purchased from Sigma-Aldrich Chemicals Co. (St. Louis, MO, USA). Other chemicals, including dimethyl sulfoxide (DMSO, ≥99.9%), crystal violet (>96%, BioShop CRY422), and gentamicin (potency >590 μg/mg, BioShop GTA202), were obtained from BioShop Canada Inc. (Burlington, Canada). Reagents related to qRT-PCR were purchased from Thermo Fisher Scientific (Waltham, MA, USA). All fungicide master stocks were dissolved in DMSO and stored at −20°C in the dark. For bacterial inhibition assays, the concentrations of the fungicides were selected based on their solubility, the need to maintain the final DMSO concentration at a level that did not affect bacterial growth, and the recommended field final spray concentrations reported by the Agricultural Chemicals Research Institute (ACRI), Ministry of Agriculture, Taiwan, in the “Plant Protection Manual - Fruit Trees, 2017 Edition (https://www.acri.gov.tw/).”

2.2. Bacterial strain and culture

Bacterial strain B. cereus was purchased from the Bioresource Collection and Research Center (BCRC), Food Industry Research and Development Institute (FIRDI), Taiwan (BCRC15850, ATCC33019), and stored at −80°C. The pathogen was cultured in nutrient broth for 20 h at 30°C before subsequent experiments.

2.3. C. elegans strain and maintenance conditions

C. elegans N2 strain (wild type) was obtained from the Caenorhabditis Genetics Center (CGC, University of Minnesota, Minneapolis, MN, USA). C. elegans were cultured on nematode growth medium (NGM) plates seeded with Escherichia coli OP50 and maintained at 20°C. After C. elegans reached adulthood, synchronization was performed using the bleaching method (1.7% NaClO in 1N NaOH) to dissolve adults and isolate eggs (Stiernagle, 2006). The harvested eggs were resuspended in M9 buffer and incubated at 20°C until hatching. The synchronized L1 larvae obtained were then used for subsequent experiments.

2.4. Growth inhibition assay

To evaluate the effects of fungicides on the growth of the foodborne pathogen B. cereus, overnight cultures were grown in nutrient broth at 30°C for 18–20 h, then diluted into Mueller-Hinton broth to an initial OD₆₀₀ of 0.01. The bacterial suspension was then exposed to eight fungicides (CHT, PRO, TEB, AZX, PCZ, MCZ, CBZ, and TDM) in the dark at 37°C for 24 h. After incubation, the OD₆₀₀ values of each group were measured, and the inhibition rate was calculated according to the following formula (Akinduti et al., 2019):

$$\text{Growth}\text{inhibition}(\%)=1‐\frac{{\text{OD}}_{600}\left(24\mathrm{h}\right)‐{\text{OD}}_{600}\left(0\mathrm{h}\right)}{{\text{OD}}_{600}\left(24\mathrm{h},\text{control}\right)‐{\text{OD}}_{600}\left(0\mathrm{h},\text{control}\right)}\times 100\%$$ (1)

The assay was conducted with at least 3 independent biological replicates and 3 technical replicates for each group.

2.5. Long-term adaptation experiment

To investigate the potential effects of long-term fungicide exposure, B. cereus was subjected to a 30-day adaptation experiment, with modifications based on (Santi et al., 2021). B. cereus was cultured in nutrient broth at 30°C for 18–20 h, diluted with Mueller-Hinton broth to an initial OD₆₀₀ = 0.1, and then exposed to fungicide concentrations previously determined to fully inhibit bacterial growth (based on the 24 h inhibition assay) for 3 h at 37°C in the dark. After exposure, cells were harvested by centrifugation, washed twice with Mueller-Hinton broth to remove residual fungicide, diluted to OD₆₀₀ = 0.01, and cultured in Mueller-Hinton broth at 37°C for 18–20 h. The following day, the bacterial suspension was again adjusted to an OD₆₀₀ of 0.1 and re-exposed to the same fungicide concentration. Residual cultures were preserved at −80°C for subsequent experiments. This procedure was repeated daily for 30 days. During the experiment, OD₆₀₀ values and colony counts were monitored before and after exposure, and survival rates were calculated as follows (Kocabay and Çetinkaya, 2020):

$$\text{Survival}(\%)=\frac{\log \text{CFU}\left(3\mathrm{h}\right)}{\log \text{CFU}\left(0\mathrm{h}\right)}\times 100\%$$ (2)

The experiment was conducted once with two technical replicates per group.

2.6. Biofilm formation assay

To examine whether long-term fungicide exposure affects the biofilm-forming ability of B. cereus, assays were performed as previously demonstrated (Zhao et al., 2021). Adapted B. cereus was cultured in nutrient broth at 30°C for 18–20 h, harvested by centrifugation, and diluted with tryptic soy broth to an initial OD₆₀₀ = 0.1. Then, 150 μL of the suspension was added to 96-well plates and incubated at 30°C for 24 h. After incubation, the wells were gently washed with phosphate-buffered saline (PBS) to remove planktonic cells without disrupting the biofilm. Wells were then stained with 175 μL of 0.1% crystal violet for 10 min, followed by PBS washes to remove excess dye, air-drying for 10 min, and destaining with 200 μL of 30% acetic acid. Absorbance was measured at 590 nm using a microplate reader (BioTek, Synergy H1, Winooski, VT, USA). The assay was conducted with at least 6 independent biological replicates and 6 technical replicates for each group.

2.7. Antibiotic challenge assay

To determine whether long-term exposure to fungicides affects antibiotic susceptibility, assays were conducted over 7 days, with modifications from a previous study (Zhang, 2023). Adapted B. cereus was cultured in nutrient broth at 30°C for 18–20 h, harvested, and diluted in Mueller-Hinton broth to an OD₆₀₀ of 0.04. Gentamicin and tetracycline were selected as test antibiotics, with concentration ranges of 32, 16, 8, 6, 4, 2, 1, 0.5, and 0.25 μg/mL for gentamicin and 32, 16, 8, 4, 2, 1, 0.5, and 0.25 μg/mL for tetracycline. In a 96-well plate, 150 μL of each antibiotic dilution was mixed with 50 μL of bacterial suspension (initial OD₆₀₀ = 0.01) and incubated at 30°C in the dark for 20 h. OD₆₀₀ values were measured at the start and end of exposure using a microplate reader (BioTek, Synergy H1, Winooski, VT, USA) to assess bacterial growth. The assay was conducted with at least 6 independent biological replicates and 2 technical replicates per group.

2.8. C. elegans infection assay

To assess whether long-term fungicide exposure altered bacterial virulence, C. elegans was used as a model host, and assays were adapted from a previous study (Rae et al., 2010). Adapted B. cereus was cultured in nutrient broth at 30°C for 18–20 h, harvested, and adjusted to OD₆₀₀ = 3 with M9 buffer. Then, 40 μL of the suspension was spread evenly onto 3.5 cm NGM plates (Stiernagle, 2006) and incubated at 37°C for 72 h to induce toxin production. Adult C. elegans were washed three times with M9 buffer, and 20 worms were placed on each plate, incubated at 25°C, and monitored for survival. Survival rates were recorded at 24, 27, 30, and 33 h, and then daily until day 7. Worms were considered dead if they failed to respond to gentle prodding with a worm pick, and no pharyngeal pumping was observed. The assay was conducted with at least 3 independent biological replicates and 2 technical replicates per group, with a minimum of 120 worms per group.

2.9. Bacterial colonization of C. elegans

To evaluate whether long-term exposure to fungicides alters the intestinal colonization capacity of B. cereus, an intestinal colonization assay was conducted using C. elegans as an infection model (Ayala et al., 2017). Adapted B. cereus was cultured in nutrient broth at 30°C for 18–20 h, harvested, and adjusted to OD₆₀₀ = 3 with M9 buffer. Subsequently, 40 μL of the bacterial suspension was evenly spread onto 3.5 cm NGM plates and incubated at 37°C for 72 h to induce toxin production. After incubation, adult worms were washed three times with M9 buffer, approximately 20 worms were transferred to each plate, and the plates were incubated at 25°C for 29 h. The worms were then collected and transferred into M9 buffer containing 0.25 mM tetramisole and 100 μg/mL gentamicin and gently shaken for 15 min to remove surface-adhered bacteria. The worms were pelleted by centrifugation and washed three additional times with M9 buffer. Cleaned worms were homogenized using a grinder; the resulting bacterial suspension was then serially diluted and plated on LB agar supplemented with 50 μg/mL ampicillin. After overnight incubation, CFU counts were determined, and bacterial colonization levels were quantified as log₁₀ CFU per worm. The assays were conducted with 3 biological replicate and 2 technical replicates.

2.10. Quantitative real-time PCR (qRT-PCR) analysis

Adapted and non-adapted B. cereus were cultured in nutrient broth at 30°C for 18–20 h, harvested, and diluted in Mueller-Hinton broth to OD₆₀₀ = 0.01, followed by incubation at 30°C for 20 h. Cells were collected by centrifugation, washed three times with sterile water, and resuspended in TRIzol™ Reagent (Invitrogen, Carlsbad, CA, USA). Total RNA was extracted using the Welprep cell/tissue RNA isolation kit (Welgene, Taipei, Taiwan). cDNA was synthesized from RNA using the SuperScript III First-strand synthesis SuperMix (Invitrogen, Carlsbad, CA, USA). qRT-PCR was performed with Power SYBR™ Green PCR Master Mix (Applied Biosystems, Waltham, MA, USA) on a Roche LightCycler 480 (Basel, Switzerland). Primers for the targeted genes, including nheA, nheB, nheC, hblA, hblC, hblD, purC, purL, calY, codY, smr, mate, and mfs were listed in Table S1. Expression levels were normalized to 16S rRNA (Yuan et al., 2013). Data were analyzed using the comparative cycle threshold (Ct) method 2^−ΔΔCt (Schmittgen and Livak, 2008). The assay was conducted with three biological replicates per group and at least two technical replicates.

2.11. Bacterial whole-genome sequencing and bioinformatics processing

The non-treated B. cereus BCRC15850 was used as the control to identify potential genetic mutations. B. cereus strains exposed to 0.1% DMSO (solvent control), CHT (8 μM), TEB (500 μM), and PRO (175 μM) for 30 days were collected for total DNA extraction and whole-genome sequencing. Whole-genome sequencing libraries were prepared using the NEBNext® Ultra™ II DNA Library Prep Kit. After library preparation, the DNA concentrations were determined by Qubit, and the fragment lengths were determined by Fragment Analyzer. Sequencing was conducted on an Illumina NovaSeq X Plus by the Roy J. Carver Biotechnology Center. After sequencing, raw paired-end reads were processed by removing adapter sequences using Trimmomatic v0.40 (Bolger et al., 2014) and subsequently subjected to quality filtering with Fastp v1.0.1 (Chen et al., 2018). High-quality reads were de novo assembled using Unicycler v0.5.1 (Wick et al., 2017). The assembled contigs were analyzed using PhaME v1.0.2 (Shakya et al., 2020) for phylogenetic and molecular evolutionary analyses. The genome assembly and annotations of the ATCC33019 reference strain (GenBank: GCA_031316765.1) were used as the reference dataset. Genome assemblies from randomly selected B. cereus strains, including ATCC10876 (CM000715.1), NC7401 (AP007209.1), ROCK1-15 (CM000729.1), BDRD-ST26 (CM000724.1), 95/8201 (CM000727.1), AH1273 (CM000741.1), and BCRC15850, were included for comparison. Single-nucleotide polymorphisms (SNPs) were identified in both coding and non-coding regions. Phylogenetic trees were constructed using the maximum-likelihood method implemented in RAxML within the PhaME pipeline (Stamatakis, 2014). In addition, variant analyses were performed by mapping cleaned reads from the treatment groups to the reference genome using the Burrows–Wheeler Aligner, followed by variant calling with FreeBayes (Garrison and Marth, 2012), as detailed in the Supporting Information.

2.12. Data analysis

Overall, the data analyses were conducted primarily in R, with the data checked for normality, stratified to investigate interactions, reviewed for dose dependence, and controlled for false positives using p-value adjustments. In this study, data are displayed as violin plots (first quartile, median, and third quartile), with individual observations or technical replicates treated as independent data points during visualization and statistical analyses. Distributional normality was examined using the Shapiro–Wilk test. Because the assumption of normality was not met, differences among treatment groups were analyzed using nonparametric Kruskal–Wallis rank-sum tests, followed by Dunn's post hoc comparisons with Holm-adjusted p-values, as implemented in the rstatix package in R. Dose–response relationships for bacterial growth (OD₆₀₀) to increasing concentrations of CHT, PRO, TEB, and other compounds, as well as OD₆₀₀ responses to graded concentrations of gentamicin or tetracycline, were evaluated using the Jonckheere–Terpstra (JT) test (coin package). Analyses were conducted with a rank transformation of the response variable and an ordered-scores transformation of the dose factor. For antibiotic susceptibility datasets, JT tests were stratified by chemical treatment condition (Control, CHT 8 μM, TEB 500 μM, and PRO 175 μM), thereby assessing monotonic dose dependence within each treatment group. The main effects of chemical treatment and time (Days 5, 10, 20, and 30) on biofilm formation, as well as their interaction (Chem × Day), were examined using the Scheirer–Ray–Hare test implemented in the rstatix package. An analogous analytical framework was applied to the gentamicin and tetracycline susceptibility datasets to evaluate the main effects of chemical treatment, antibiotic concentration, and their interaction (Chem × Antibiotic) on bacterial growth (OD₆₀₀). Differences in expression of virulence- and toxin-associated genes (nheA, nheB, nheC, hblA, hblC, hblD, purC, purL, calY, codY, smr, mate, and mfs) between TEB-treated and control groups were assessed on a gene-by-gene basis using the Kruskal–Wallis rank-sum test and cross-validated using both log₂-scale and linearized data (2^−ΔΔCt) to confirm that all the statistics remain the same. To account for multiple testing across genes and control the false discovery rate, p-values were adjusted using Holm's procedure. Differences in gut colonization (CFU) between the control and TEB-treated groups were evaluated using the Wilcoxon rank-sum test. For all analyses, a two-sided significance level of α = 0.05 was adopted. Test results are reported based on Holm-adjusted p-values, as summarized in Appendix 1. Statistically significant differences relative to the control group are denoted as * p < 0.05, **p < 0.01, and ***p < 0.001, or using different capital letters. Survival curves from the C. elegans infection assay were generated and analyzed using GraphPad Prism 8.0.2 (GraphPad Software, CA, USA), and differences in survival were evaluated using the Mantel–Cox log-rank test, with significance reported using the same notation as asterisks.

3. Results and discussion

3.1. Effects of fungicide exposure on the growth of B. cereus

As shown in Fig. 1A, exposure to CHT for 24 h caused dose-dependent growth inhibition (p < 0.001, JT tests), culminating in complete suppression at 8 μM (p < 0.001). PRO similarly exhibited concentration-dependent inhibition (p < 0.001, JT tests) with complete suppression at 175 μM (p < 0.001) (Fig. 1B). In contrast, TEB (50–250 μM) did not inhibit growth (p = 0.134-1.000), whereas higher concentration (500 μM) caused potent inhibition (p = 0.013), while the overall effects were dose-dependent (p < 0.001, JT tests) (Fig. 1C). Across Fig. 1D–H, AZX exhibited minimal effects on growth, with only slight inhibition at higher concentrations (1000-1200 μM) (p = 0.001 and 0.032) (Fig. 1D). PCZ showed no effect at 100-200 μM (p = 0.062-1.000), whereas 250-290 μM significantly promoted growth (p < 0.001) (Fig. 1E). MCZ induced measurable inhibition only at 7.5 μM (p = 0.040) (Fig. 1F). CBZ showed no effect at 50-150 μM (p = 0.451-1.000) and caused only minor inhibition at 200 and 250 μM (p = 0.002 and p < 0.001) (Fig. 1G). TDM (50-500 μM) did not affect bacterial growth (p = 0.846-1.000) (Fig. 1H).

Fig. 1.

Effect of fungicide exposure on the growth of foodborne pathogen B. cereus 15850

B. cereus was co-cultured with different concentrations of (A) CHT, (B) PRO, (C) TEB, (D) AZX, (E) PCZ, (F) MCZ, (G) CBZ, and (H) TDM. After 24 h of incubation at 37°C in the dark, OD₆₀₀ was measured to assess bacterial growth, and the inhibition rate was calculated accordingly. The quartiles of the data were indicated as horizontal lines within the violin plot. Statistical significance was determined using Kruskal-Wallis tests, followed by Dunn's post hoc tests with Holm's p-value adjustments, with different letters indicating statistically significant differences (p < 0.05). The concentration dependence of the effects was analyzed using Jonckheere–Terpstra tests. Different capital letters mean statistically significant differences at p < 0.05. Three independent biological replicates were performed, with at least three technical replicates per treatment.

Previous studies have found that common pesticides, including copper hydroxide, acetamiprid, cypermethrin, and permethrin, significantly inhibited the growth of E. coli O157:H7 and Salmonella spp. (Dobhal et al., 2014), and chlorothalonil (CHT) has been shown to suppress B. cereus growth (Shi et al., 2013), consistent with our findings. We further investigated whether prolonged exposure to fungicides induces phenotypic changes in B. cereus. Because AZX, PCZ, MCZ, CBZ, and TDM did not exhibit consistent concentration-dependent inhibition within the tested range, and PCZ promoted growth at 250-290 μM, a comparable inhibitory concentration for the acclimation protocol could not be defined. Their potential sub-lethal effects under alternative exposure designs warrant future study.

Unlike the other seven fungicides tested, PCZ did not inhibit B. cereus growth across the concentration range examined; instead, growth was significantly stimulated at 250-290 μM compared to the untreated control (p < 0.001, Fig. 1E). This suggests a biphasic, hormetic dose-response, in which low or sub-inhibitory doses stimulate growth, whereas higher doses become inhibitory (Calabrese and Baldwin, 2003). Similar hormetic growth stimulation has been observed in several fungal plant pathogens exposed to azoles and other fungicides, where low doses increased mycelial growth, sporulation, and virulence traits before inhibition at higher concentrations (Garzon and Flores, 2013). A related phenomenon was reported for soil bacteria exposed to triazole fungicides: TEB within good agricultural practice doses (0.01-1.0 mg kg⁻¹) stimulated bacterial proliferation (up to 1.6-fold increase) and increased soil enzyme activity, rather than inhibiting growth (Baćmaga et al., 2022). Generally, fungicides affecting ergosterol biosynthesis, membrane integrity, or respiration can have variable effects on non-target bacteria, which depend on concentration, exposure time, and bacterial physiological state, ranging from inhibitory to stimulatory (Yang et al., 2011; Roman et al., 2021). Since PCZ, like TEB and PRO, is a triazole that inhibits fungal lanosterol 14α-demethylase, its failure to kill or inhibit B. cereus, along with the observed growth stimulation, may indicate a weak or nonexistent inhibitory effect on bacterial targets at the tested concentrations, possibly coupled with low-level metabolic stimulation, similar to hormetic responses. The physiological reasons for this stimulatory effect in B. cereus were not directly studied here but suggest future research directions, such as transcriptomic or metabolomic profiling across the full stimulatory-to-inhibitory range. The differing responses across eight fungicides point to a context-dependent relationship with non-target bacteria. For example, Dobhal et al. (2014) reported similar findings with eight pesticides against E. coli O157:H7 and Salmonella, with AZX stimulating E. coli growth, while others had inhibitory or neutral effects. CHT applied at various soil doses has been shown to stimulate bacterial and actinobacterial growth but inhibit fungi (Baćmaga et al., 2018). Some B. cereus strains can degrade CHT enzymatically, indicating detoxification rather than susceptibility (Zhang et al., 2007). These results highlight that fungicide effects are strain-specific and should not be generalized without testing in specific environments.

3.2. Effects of long-term fungicide exposure on B. cereus

The long-term exposure design consisted of repeated 3-h daily challenge-and-recovery cycles, adapted from experimental acclimation approaches (Santi et al., 2021). A 24-h exposure was not feasible because 8 μM CHT, 500 μM TEB, and 175 μM PRO strongly inhibited or completely suppressed growth in the screening assay, preventing their propagation over 30 days. AZX, PCZ, MCZ, CBZ, and TDM did not yield a consistent inhibitory concentration within the tested range; PCZ promoted growth at 250-290 μM. Because we hypothesized that prolonged sub-inhibitory fungicide levels may promote bacterial adaptations, we excluded those that showed no significant effects. Therefore, based on the results from Section 3.1, CHT 8 μM, TEB 500 μM, and PRO 175 μM, the concentrations that completely inhibited growth within 24 h were selected for the 30-day repeated exposure assay. According to Fig. 2A, OD₆₀₀ measurements indicated that control groups grew stably throughout the experiment, whereas B. cereus growth was consistently suppressed during 3 h of daily exposure to CHT, TEB, or PRO. Colony counts further revealed differential responses to adaptation. After exposure to 8 μM CHT, survival decreased to ∼70% on day 1 but rapidly recovered, stabilizing at 85–100% from day 2 onward. In the case of TEB 500 μM, survival decreased to ∼50% on day 1, fluctuated slightly during first 7 days, but recovered by day 8 and subsequently remained stable at 80–100%. Under PRO 175 μM exposure, survival rates declined markedly to 5–60% during the first 11 days, with the lowest level (<5%) observed on day 6; however, survival increased thereafter and stabilized at 70–100% from day 12 onward. (Fig. 2B). We further analyzed the genomes of B. cereus before and after exposure to 0.1% DMSO (control), CHT 8 μM, TEB 500 μM, and PRO 175 μM, in relation to the original strains BCRC15850 (sequenced in this study) and ATCC33019 (sequences obtained from GenBank) (Fig. 2C). The results indicated that fungicide treatments over our current experimental duration of 30 generations did not induce meaningful genetic mutations in B. cereus, as evidenced by highly similar sequences between the original and treated strains (Fig. 2C). The results were also supported by the variant-calling procedures, which failed to detect significant mutations in the B. cereus genomes (Table S2).

Fig. 2.

Effect of long-term fungicide exposure on the growth of foodborne pathogen B. cereus 15850

B. cereus 15850 (initial OD₆₀₀ = 0.1) was co-cultured with three fungicides—CHT 8 μM, TEB 500 μM, and PRO 175 μM—under dark conditions at 37°C for 3 h. After exposure, the bacterial cells were collected, resuspended in fresh medium to an OD₆₀₀ of 0.01, and cultured for 18–20 h without fungicides. The following day, cells were re-exposed to the same fungicide concentrations. This repeated-exposure experiment was conducted continuously for 30 days. During each daily cycle, bacterial growth after 3 h of exposure was evaluated by measuring OD₆₀₀ (A), and the survival rate was determined by counting CFU before and after fungicide exposure (B). One independent biological replicate was performed, with two technical replicates per treatment. (C) Phylogenetic tree based on SNPs among B. cereus strains before and after exposure to 0.1% DMSO (control), CHT 8 μM, TEB 500 μM, and PRO 175 μM.

Previous studies have shown that long-term exposure of Listeria monocytogenes isolates from food, humans, and the environment to quaternary ammonium compound (QAC) disinfectants, such as benzalkonium chloride (BAC), increased tolerance, with MIC values increasing 6- to 7-fold. This tolerance remained stable for at least 7 days (approximately 50 generations) even after BAC removal (Aase et al., 2000). Similarly, long-term exposure of eight clinical isolates of uropathogenic E. coli (UPEC) to BAC, silver nitrate, and triclosan enhanced adaptation, with four strains showing increased tolerance to BAC and all eight exhibiting higher tolerance to silver nitrate and triclosan, reflected in a twofold increase in MIC values (Henly et al., 2019). The findings indicated that prolonged exposure to fungicides (CHT, PRO, and TEB) might induce phenotypic changes without significant evidence of genetic mutations. The lack of detectable mutations in fungicide-adapted B. cereus, despite stable phenotypic changes, is consistent with non-mutational adaptation in bacteria (Windels et al., 2020). Similar rapid tolerance increases without resistance mutations have been observed in E. coli and other pathogens during antibiotic cycles, often through extended lag phases or the formation of persister cells (Balaban et al., 2004). Similar findings have been demonstrated using various bacterial species, including L. monocytogenes and uropathogenic E. coli, treated with antibacterials (Aase et al., 2000; Henly et al., 2019). Since tolerance and persistence are reversible and non-heritable, they typically do not alter the genome, consistent with our sequencing results in different adapted lineages. Our findings suggest this concept extends to fungicide stress, resulting in increased biofilm and reduced antibiotic susceptibility. However, this acclimation procedure was conducted for a single independent lineage per fungicide, with two technical replicates at each time point. Accordingly, Fig. 2 is presented as a descriptive record of the acclimation trajectory, and the generalization of these adaptation trajectories requires confirmation using independent acclimation lineages.

3.3. Impact of long-term fungicide exposure on biofilm formation by B. cereus

As shown in Fig. 3, on day 5 of the acclimation experiment, the groups treated with CHT (8 μM), TEB (500 μM), and PRO (175 μM) exhibited significantly lower biofilm production than the control (p < 0.001), indicating that all three fungicides suppressed biofilm formation during the early phase of long-term exposure. However, different trends emerged as acclimation time increased. In the CHT-treated group, no significant differences were found in biofilm formation after day 10, day 20, or day 30 (p = 0.053-0.178). In the TEB-treated group, significant increases were observed on days 10, 20, and 30 (p < 0.001), with an upward trend throughout the acclimation period. In the PRO-treated group, biofilm formation increased significantly by day 10 (p < 0.001). However, biofilm formation was markedly lower than in the controls on day 20 (p < 0.001), and there was no significant difference on day 30 (p = 0.061). SRH analysis indicated significant effects of the tested fungicides (p < 0.001), days of fungicide pretreatment (p < 0.001), and their interaction (p < 0.001) on biofilm formation. Taken together, these results indicate that long-term fungicide exposure significantly enhances the biofilm-forming capacity of B. cereus.

Fig. 3.

Effect of long-term fungicide exposure on biofilm formation of foodborne pathogen B. cereus 15850

B. cereus 15850 adapted to CHT 8 μM, TEB 500 μM, and PRO 175 μM for 5, 10, 20, and 30 days were cultured under dark conditions at 30°C for approximately 24 h prior to biofilm staining. Biofilm formation was quantified by measuring the absorbance at 590 nm (OD₅₉₀) using a microplate reader. The quartiles of the data were indicated as horizontal lines within the violin plot. Statistical significance was determined using Kruskal-Wallis tests, followed by Dunn's post hoc tests with Holm's p-value adjustments. The significance levels relative to the control were as follows: p < 0.05 (*), p < 0.001 (***), and not significant (n.s.). Scheirer-Ray-Hare tests were conducted to determine the contribution of each factor on the changes in bacterial growth. At least six independent biological replicates were performed, with at least six technical replicates per treatment.

Biofilm formation can be influenced by unfavorable environmental conditions, and several studies have demonstrated that exposure to stress may promote biofilm development. For instance, clinical isolates of UPEC showed significantly increased biofilm formation after long-term exposure to triclosan, while partial increases were also observed following prolonged exposure to BAC or silver nitrate (Henly et al., 2019). Similarly, B. subtilis exhibited elevated biofilm production in response to sublethal doses of chlorine dioxide (Shemesh et al., 2010). On the other hand, biofilm formation depends on a balance between matrix production and dispersal in response to environmental and nutritional cues (McDougald et al., 2012; Guilhen et al., 2017). These changes might be triggered by nutrient limitation, oxygen changes, or population density shifts, and involve signaling systems that activate enzymes to degrade the matrix and release cells. The daily exposure, recovery, and regrowth likely caused the population to cycle through stress, nutrient depletion, and regrowth phases, leading to temporary biomass drops and subsequent recovery (on day 30). Alternatively, these fluctuations might reflect selection within a diverse population, favoring different subpopulations at various stages. Since gene expression and signaling were not measured at intermediate points, this remains a hypothesis to be tested with time-resolved transcriptomics. Nonetheless, these findings suggest that both short- and long-term environmental stress can increase biofilm formation in bacteria.

3.4. Impact of long-term fungicide exposure on antibiotic susceptibility in B. cereus

B. cereus is generally susceptible to antibiotics, including tetracycline, gentamicin, erythromycin, chloramphenicol, ciprofloxacin, and imipenem (Hwang et al., 2022). Tetracycline is a broad-spectrum antibiotic that remains widely used in human and veterinary medicine due to its low cost, minimal side effects, and wide availability (Amangelsin et al., 2023). Gentamicin is widely applied in the treatment and prevention of various bacterial infections, with relatively low rates of resistance emergence (Chen et al., 2014). Gentamicin and tetracycline were selected as non-β-lactam agents because B. cereus commonly produces β-lactamase and is commonly less susceptible to β-lactams, while susceptibility to gentamicin and tetracycline is more informative but can vary among isolates (Fiedler et al., 2019; McDowell et al., 2023). Based on these considerations, tetracycline and gentamicin were selected as representative antibiotics for testing in this study.

As shown in Fig. 4A and B cereus subjected to long-term exposure to CHT 8 μM exhibited significantly higher growth compared with the control when exposed to gentamicin at 4 (p = 0.001), 6 (p = 0.012), and 8 μg/mL (p < 0.001). B. cereus subjected to long-term exposure to TEB 500 μM displayed significantly elevated growth at 6 (p = 0.048) and 8 μg/mL (p < 0.001) compared to the control. B. cereus subjected to long-term exposure to PRO 175 μM showed significantly higher growth at 8 μg/mL (p < 0.001). For tetracycline exposure, only the CHT (8 μM) (p = 0.028) and TEB (500 μM)-treated groups (p = 0.009) exhibited slightly improved growth compared with the controls at 8 μg/mL tetracycline (Fig. 4B). SRH analyses suggested significant effects of fungicide treatments on survival under gentamicin and tetracycline challenges (p = 0.002 and 0.005).

Fig. 4.

Effect of long-term fungicide exposure on antibiotic resistance of foodborne pathogen B. cereus 15850

B. cereus 15850 adapted for 30 days to CHT 8 μM, TEB 500 μM, and PRO 175 μM were cultured with different concentrations of (A) gentamicin and (B) tetracycline under dark conditions at 30°C for 20 h. Bacterial growth was evaluated by measuring the optical density at 600 nm (OD₆₀₀). The quartiles of the data were indicated as horizontal lines within the violin plot. Statistical significance was determined using Kruskal-Wallis tests, followed by Dunn's post hoc tests with Holm's p-value adjustments. The significance levels relative to the control were as follows: p < 0.05 (*), p < 0.001 (***), and not significant (n.s.). Scheirer-Ray-Hare tests were conducted to determine the contribution of each factor on the changes in bacterial growth. The concentration dependence of the effects was also analyzed using Jonckheere–Terpstra tests. At least six independent biological replicates were performed, with at least two technical replicates per treatment.

According to the Clinical and Laboratory Standards Institute (CLSI) guidelines, B. cereus isolates with MIC values ≤ 4 μg/mL are considered susceptible, while those with MIC ≥ 16 μg/mL are classified as resistant (CLSI, 2015). In this study, no growth was observed in any group at gentamicin or tetracycline concentrations ≥16 μg/mL (Data not shown). Although these findings do not definitively classify the isolates as resistant strains, they clearly demonstrate that long-term fungicide exposure can reduce B. cereus susceptibility and enhance its tolerance to gentamicin and tetracycline. The reduced susceptibility to gentamicin and tetracycline in fungicide-adapted B. cereus occurs without detectable genomic mutation. This suggests a tolerance-type phenotype rather than resistance, as resistance involves elevated MICs and is usually genetically encoded, while tolerance allows bacteria to survive longer without increased MIC and is often physiologically mediated (Windels et al., 2020). The absence of genomic mutations, along with increased biofilm formation and decreased antibiotic susceptibility in TEB-adapted strains, supports a physiological reduction in susceptibility rather than classical resistance (Mah and O’Toole, 2001). This aligns with studies showing that environmental stressors can induce cross-tolerance via shared stress-response pathways rather than specific resistance mechanisms (Aase et al., 2000; Henly et al., 2019). We interpret this as an adaptive, tolerance-related reduction in susceptibility, though MIC testing is needed to confirm if true resistance has emerged.

3.5. Impact of long-term fungicide exposure on the pathogenicity of B. cereus

As shown in Fig. 5A and B cereus after 30-day long-term exposure to CHT 8 μM, TEB 500 μM, or PRO 175 μM exhibited significantly higher lethality in C. elegans compared with the control group (p < 0.001). In the CHT 8 μM and PRO 175 μM groups, a gradual but sustained increase in mortality was observed. In contrast, in the 500 μM TEB group, C. elegans survival declined rapidly over the first two days, and by day 3, most worms were dead. These findings demonstrate that long-term exposure to fungicides markedly enhances B. cereus pathogenicity toward its host, as evidenced by increased virulence in the C. elegans infection model. Following the survival assay observations, intestinal colonization experiments were conducted. As shown in Fig. 5B, our results demonstrated that prolonged exposure to 500 μM TEB led to significantly higher intestinal colonization levels in C. elegans than in the control group (p = 0.025), suggesting that chronic fungicide exposure promoted bacterial persistence within the host gut.

Fig. 5.

Effects of long-term fungicide exposure on the pathogenicity and intestinal colonization ability of foodborne pathogen B. cereus 15850

(A) B. cereus 15850 adapted for 30 days to CHT 8 μM, TEB 500 μM, or PRO 175 μM were co-exposed with the host C. elegans, and worm survival was monitored daily. Statistical differences between the control and exposure groups were analyzed using the log-rank (Mantel–Cox) test. The significance levels, relative to the control, were as follows: p < 0.001 (***). (B) B. cereus 15850 strains adapted to TEB 500 μM for 30 days were co-incubated with C. elegans for 29 h. Worms were collected, washed, and homogenized, then serially diluted and plated onto LB agar supplemented with ampicillin for CFU enumeration. The quartiles of the data were indicated as horizontal lines within the violin plot. Statistical significance was determined using the Kruskal-Wallis test. Asterisk (*) indicates p < 0.05. At least three independent biological replicates were performed, with two technical replicates per group.

Previous studies have shown that clinical isolates of UPEC exhibit significantly increased lethality in the Galleria mellonella infection model following long-term exposure to disinfectants such as BAC, silver nitrate, and polyhexamethylenbiguanide (PHMB) (Henly et al., 2019). The increased lethality of CHT-, TEB-, and PRO-adapted B. cereus against C. elegans reflects a broader phenomenon in which prolonged exposure of bacteria to sublethal stress enhances both survival and pathogenicity. This link is documented for foodborne pathogens subjected to food-processing or disinfectant stressors, which trigger protective responses and boost virulence traits such as adhesion, invasion, toxin production, and infectivity (Wesche et al., 2009; Begley and Hill, 2015). Such stress responses overlap with those encountered within the host gastrointestinal tract, effectively priming bacteria for infection (Horn and Bhunia, 2018). Our findings support this framework, as B. cereus lineages adapted to CHT, TEB, and PRO exhibited increased lethality in C. elegans, biofilm formation, and reduced susceptibility to biocides such as gentamicin and tetracycline, without detectable genomic mutations (Fig. 2C; Table S2). Similar phenotypic shifts have been observed in L. monocytogenes exposed to benzalkonium chloride (Jiang et al., 2022). Overall, these data imply that fungicide adaptation in B. cereus may result in a broader stress-adapted phenotype driven by non-mutational regulatory mechanisms. This supports the idea that prolonged fungicide exposure might prime pathogens for increased virulence without requiring the acquisition of new resistance genes.

The environmental relevance of the repeated-exposure concentrations requires careful interpretation. CHT, TEB, and PRO were selected because they significantly inhibited B. cereus growth during the initial screen, whereas the daily exposure duration was reduced to 3 h to retain a viable population for repeated adaptation and recovery cycles. Therefore, the long-term experiment was designed as a controlled laboratory model of recurrent, acute fungicide stress rather than a direct reconstruction of exposure in crops or food products. The 8 μM CHT concentration corresponds to approximately 2.13 mg/L and is of the same order of magnitude as CHT residues reported for some food commodities (Sun et al., 2022). However, 500 μM TEB (approximately 154 mg/L) and 175 μM PRO (approximately 51 mg/L) pose stronger experimental challenges than the residues reported in many foods after treatment. Importantly, direct comparison of aqueous molar concentrations with food residues expressed as mass concentrations is inherently approximate because the locally bioavailable concentration depends on moisture content, partitioning, adsorption onto soil or organic material, formulation, degradation, and the microenvironment inhabited by bacterial cells. Accordingly, our results demonstrate that repeated fungicide stress can generate adaptive phenotypic responses under controlled conditions, but they do not establish that identical responses occur at typical environmental residue concentrations. Future studies should combine residue measurements from agricultural and food-associated matrices with experiments using environmentally measured concentrations, realistic exposure pulses, mixtures, and multiple B. cereus isolates.

3.6. Impact of long-term fungicide exposure on the expression levels of genes in B. cereus

The mRNA levels of virulence genes were measured to assess the elevated virulence and resilience of B. cereus in the absence of genetic mutations. TEB was prioritized as the most consistent phenotype for an initial mechanistic screen, and because TEB showed the most significant enhancement of virulence against C. elegans, we investigated the TEB-adapted B. cereus for transcriptional changes.

Fig. 6 demonstrates that prolonged TEB exposure significantly elevated the expression of all tested genes except for nheA (p = 0.092), nheB (p = 0.061), codY (p = 0.106), and mate (p = 0.106). Among the most significantly affected genes, purC (p = 0.042) and purL (p = 0.042) are involved in purine biosynthesis and extracellular DNA release, both of which are essential components of EPS during early biofilm formation (Vilain et al., 2009). The bifunctional protein calY (p = 0.005) contributes to matrix composition and adhesion to host tissues (Candela et al., 2019; Sonenshein, 2005). These results indicate that TEB exposure is associated with biofilm-associated gene expression. Moreover, efflux pump–related genes in B. cereus were induced by smr and mate. These findings suggest that long-term fungicide acclimation enables B. cereus to reduce antibiotic susceptibility by upregulating SMR efflux transporters. Similar regulatory patterns are observed in M. tuberculosis under antibiotic stress, in which multiple ABC, MFS, and SMR pumps are upregulated (Gupta et al., 2010). Furthermore, virulence-related genes nheC (p = 0.042), hblA (p = 0.042), hblC (p = 0.042), and hblD (p = 0.042) were markedly upregulated and encode the non-hemolytic enterotoxin (NHE) and hemolysin BL (HBL), which disrupt intestinal epithelial cells (Tuipulotu et al., 2021). Along with upregulation of purC (p = 0.042), purL (p = 0.042), and calY (p < 0.005), further suggesting enhanced potential for adhesion and colonization (Candela et al., 2019; Vilain et al., 2009). Previous studies have shown that L. monocytogenes exposed to long-term BAC treatment exhibited significantly elevated expression of key virulence genes, including prfA, plcA, hly, mpl, actA, and plcB (Jiang et al., 2022).

Fig. 6.

Effect of long-term fungicide exposure on the expression of virulence-related genes in B. cereus 15850

B. cereus 15850 adapted to TEB 500 μM for 30 days was cultured in Mueller–Hinton broth at 30°C under dark conditions for 20 h. Total RNA was then extracted and subjected to quantitative real-time PCR (qRT-PCR) analysis. The 16S rRNA gene was used as the internal reference. The quartiles of the data were indicated as horizontal lines within the violin plot. Statistical significance was determined using Kruskal-Wallis tests, followed by Dunn's post hoc tests with Holm's p-value adjustments, with statistically significant differences indicated using asterisks: p < 0.05 (*), p < 0.01 (**), p < 0.001 (***), and not significant (n.s.). Three independent biological replicates were performed, with at least two technical replicates per group.

Interestingly, we also discovered that B. cereus adapted to TEB consistently exhibited high expression levels of biofilm, virulence, and resistance genes, even after exposure to gentamicin (6 μg/mL) (Fig. S1), suggesting that the enhanced bacterial strains might complicate antibiotic treatment. Therefore, repeated fungicide challenge may select for a non-mutational stress-adaptation state. In TEB-adapted cells, smr upregulation is consistent with altered efflux-mediated susceptibility, while concurrent changes in biofilm- and toxin-associated transcripts are consistent with stress-linked co-regulation. Because qRT-PCR profiling in this study was performed only for the TEB-adapted lineage, the gene-expression-based mechanistic interpretations discussed above, including the proposed roles of purC, purL, calY, smr, nheC, hblA, hblC, and hblD in biofilm formation, antibiotic susceptibility, and virulence, are specific to TEB-adapted B. cereus and should not be extrapolated to CHT- or PRO-adapted strains without direct molecular confirmation. Whether the phenotypic changes observed in the CHT- and PRO-adapted lineages arise from similar, overlapping, or distinct regulatory pathways remains an open question, and comparative transcriptomic profiling of all three fungicide-adapted lineages is identified as a priority for future work.

This study has several limitations to consider when interpreting the results. First, all experiments used only the B. cereus reference strain BCRC15850. Responses to long-term fungicide exposure may vary between strains and should not be generalized to all foodborne B. cereus. This is important because B. cereus isolates display significant genomic diversity and different profiles of virulence genes, antibiotic susceptibility, and pathogenicity across environmental, food, and clinical sources (Ehling-Schulz et al., 2004; Okutani et al., 2025). Second, this research was a controlled lab hazard identification study. The fungicide levels and repeated 3-h exposures were chosen to create a reproducible model of chronic sub-lethal stress but do not necessarily reflect the actual concentrations, exposure patterns, or persistence that B. cereus experiences in soil, crops, food processing, or the food supply. Therefore, the findings suggest a potential biological hazard but will need further validation to inform environmental risk or regulatory assessments. Future research should include multiple genetically and ecologically diverse B. cereus strains, use real-world fungicide residue data and exposure scenarios, determine standardized MICs via broth microdilution, and examine whether the phenotypes persist after fungicide removal. Nonetheless, the findings of this study revealed an understudied scenario in which persistent stress might induce adaptations in B. cereus that could improve resilience and worsen virulence without evidence of mutations, a pattern that could be tested in other relevant species and strains across stressors.

4. Conclusion

This laboratory study demonstrates that long-term repeated exposure to the fungicides CHT, TEB, or PRO can induce persistent phenotypic alterations in B. cereus BCRC15850. These alterations include modified biofilm formation, decreased susceptibility to gentamicin and tetracycline, and heightened lethality towards the infection host C. elegans. Notably, these phenotypic changes occurred in the absence of detectable genomic mutations, suggesting that the observed phenotypes result from non-mutational stress adaptation rather than antibiotic resistance. In the lineage adapted to TEB, altered expression of genes associated with biofilm formation (purC, purL, and calY), efflux mechanisms (smr), and virulence (nheC, hblA, hblC, and hblD) corresponded with the phenotypic changes observed; however, these transcriptional associations do not establish causality and should not be extrapolated to CHT- or PRO-adapted lineages without further molecular characterization. Given that this research employed a single B. cereus strain and controlled laboratory conditions designed to produce a reproducible repeated-stress model, the findings should be regarded as hazard identification rather than evidence of environmental exposure, food safety risk, or regulatory necessity. Future investigations should include multiple B. cereus isolates, assessments of environmentally relevant fungicide residues and realistic exposure scenarios, standardized MICs, evaluation of the reversibility of the adapted phenotypes following fungicide withdrawal, and comparative molecular analyses across different fungicide treatments, prior to determining the real-world implications of these findings.

Ethics declaration

This study was conducted in accordance with the ARRIVE (Animal Research: Reporting of In Vivo Experiments) guidelines. Ethics committee approval was not required. Ethical approval was not required for this study because the use of Caenorhabditis elegans, an invertebrate model organism, is exempt from institutional animal ethics review under the applicable institutional guidelines.

Declaration of generative AI and AI-assisted technologies in the writing process

During the preparation of this work, the authors used ChatGPT-4 and Grammarly Pro to improve the quality of the English writing. After using this tool, the authors reviewed and edited the content as needed and took full responsibility for the publication.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Handling Editor: Dr. Siyun Wang

Appendix A. Supplementary data

The following are the Supplementary data to this article: