Triclocarban (TCC, 98% purity, CAS: 101-20-2) and dimethyl sulfoxide (DMSO, 99% purity, CAS: 67-68-5) were purchased from Aladdin (Shanghai, China).

Sexual maturity samples of R. taihangensis were collected from the Taihangshan Mountain in Huixian City, Henan province, China, in April 2023, and reared in the laboratory conditions until spawning. Then, the freshly laid eggs were transferred to aquatic containers (45 cm × 33 cm × 13 cm) filled with water to a depth of 4 cm and maintained at a water temperature of 21 ± 1 °C. A photoperiod of 12 h light to 12 h dark was maintained throughout the whole experiment. The aerated water used in the experiment had a pH of 8.14 ± 0.02, conductivity of 300.8 ± 3.31 μS/cm, and dissolved oxygen levels of 8.18 ± 0.10 mg/L (mean ± SD). Tadpoles were fed daily with spirulina dry powder, and water was renewed every other day. This laboratory rearing protocol from the embryonic stage ensured that all experimental tadpoles were naïve to TCC and other environmental pollutants prior to the commencement of the exposure experiments. The developmental stages of tadpoles were determined according to Gosner (1960). Once the tadpoles reached Gosner stage (Gs) 26–28, they were exposed to TCC. A stock solution of 0.5 g/L were prepared using DMSO, stored at 4 °C, and diluted as needed for exposure experiments. The final DMSO concentration in all test solutions, including the control, did not exceed 0.05 ‰ (v/v). Previous studies have demonstrated that 0.5‰ DMSO, with the concentration used in this experiment being much lower than 0.5‰, exerted no adverse effects on the survival, metamorphosis, body size parameters (e.g., body mass, snout-vent length) and hepatic physiological functions (e.g., hepatic somatic index, antioxidant enzyme activity) of Hoplobatrachus rugulosus tadpoles (Chen J. Y. et al., 2022). All animal sampling and usage protocols were implemented in compliance with all ethical guidelines and legal regulations in China and approved by the Institutional Animal Care and Ethics Committee of Henan Normal University (protocol code IACEC20230027 and March 24, 2023).

A total of 12 tadpoles were sampled from each concentration group for intestinal microbiota analysis. From each of the three replicate containers per group, four tadpoles were collected, and every two individuals were pooled to form one sample, resulting in six replicate samples per concentration group. Total genomic DNA was extracted using the E.Z.N.A.^® stool DNA Kit (Omega Bio-Tek, Norcross, GA, United States) according to the manufacturer’s instructions. DNA integrity and purity were evaluated via electrophoresis on a 1% agarose gel and quantified using a Qubit@3.0 fluorometer (Thermo Scientific), respectively. The V3–V4 hypervariable region of the bacterial 16S rRNA gene was amplified using the forward primer 338F (5′-ACTCCTACGGGAGGCAGCA-3′) and reverse primer 806R (5′-GGACTACHVGGGTWTCTAAT-3′). The PCR reaction mixture consisted of 30 μL, including 20 ng of DNA templates, 1 μL of each primer, and 15 μL of 2 × Taq Master Mix (TaKaRa, China). The PCR cycling protocol was as follows: initial denaturation at 98 °C for 5 min, followed by 25 cycles of 98 °C for 30 s, 52 °C for 30 s, and 72 °C for 45 s, with a final extension at 72 °C for 5 min. Pair-end 2 × 250 bp sequencing was performed on the Illumina NovaSeq platform at Shanghai Personal Biotechnology Co., Ltd. (Shanghai, China). Bioinformatics analysis was conducted using QIIME2 v2019.4 (Bolyen et al., 2019) and R packages (v3.2.0). Raw sequences were processed using the DADA2 v1.20 plugin (Callahan et al., 2016) to perform quality filtering, denoising, merging, chimera removal, resulting in amplicon sequence variants (ASVs). Taxonomic annotation was carried out against the SILVA 132 database (Quast et al., 2013). Alpha-diversity indices (Chao1 richness index, Simpson diversity index and Pielou’s evenness index) were calculated from the ASV table in QIIME2, and group differences were assessed using the Kruskal–Wallis test. Beta diversity was evaluated by principal coordinate analysis (PCoA) based on both weighted UniFrac and unweighted UniFrac distances using the Vegan package in R, and statistical significance among groups was tested using PERMANOVA (Adonis test with 999 permutations). Differential microbial taxa between the solvent control and TCC-treated groups were identified at the phylum and genus levels using the Kruskal–Wallis test followed by Dunn’s post hoc test. Additionally, biomarkers were identified using linear discriminant analysis effect size (LEfSe), with a linear discriminant analysis (LDA) threshold score of 2.0.

For metabolomic analysis, each tadpole was treated as an independent sample and homogenized in 200 μL of ice-cold water using a high-throughput tissue grinder. A total of 24 samples were processed, comprising six biological replicates per treatment group across the four experimental groups. Metabolites were extracted by adding 800 μL of methanol-acetonitrile mixed solution (1:1, v/v), followed by ultrasonication for 30 min, incubation at −20 °C for 30 min, and centrifugation at 12,000 rpm for 10 min. The resulting supernatant was collected, concentrated under vacuum, and reconstituted in 150 μL of 50% methanol containing 5 ppm 2-chlorophenylalanine. After vortexing for 30 s and recentrifugation, 10–20 μL of supernatant from each sample was pooled into a quality control (QC) sample to assess instrumental stability and data reliability. Metabolite profiling was performed on an ACQUITY UPLC HSS T3 column (100 Å, 1.8 μm, 2.1 mm × 100 mm) and analyzed using a Thermo Orbital Trap Exploris 120 mass spectrometer controlled by Compound Discoverer^™ v3.3.2.31 (Thermo, Waltham, United States). Orthogonal partial least squares discriminant analysis (OPLS-DA) was conducted using the Ropls package in R to reduce data dimensionality and visualize group separation. Differential expressed metabolites were identified based on a combination of univariate and multivariate criteria: p-values <0.05 (Student’s t-test), variable importance in projection (VIP) >1 from the OPLS-DA model, and fold change (FC). Model quality was assessed using R²Y (goodness of fit) and Q² (predictive ability), with values closer to 1 indicating stronger model performance. Venn diagrams generated with the VennDiagram v1.7.3 package were used to illustrate common and unique metabolites across comparison groups. Shared differential metabolites identified across all TCC-treated groups were subjected to KEGG pathway enrichment analysis using the clusterProfiler package (v4.6.0). Metabolic pathways with p-value <0.05 and containing at least five differentially abundant metabolites were considered significantly enriched and further analyzed.

LEfSe was employed to identify potential microbial and metabolite markers, with a LDA score threshold set at 2.0. Spearman correlation analysis was then performed to evaluate the associations between the identified microbial and metabolite biomarkers. Only marker pairs showing strong correlations (|rho| >0.7) were retained for subsequent analysis. Additionally, bar plots illustrating differentially expressed metabolites between the solvent control group and the TCC-treated groups were generated using Origin software.

In the acute toxicity assays, probit analysis (Finney, 1978) was employed to estimate the LC₅₀ (with a 95% confidence interval) using mortality data recorded at 24-h intervals. The Kolmogorov–Smirnov and Shapiro–Wilk tests were used to evaluate whether the data significantly deviated from a normal distribution, while Levene’s test was conducted to assess homoscedasticity across groups. As the data violated parametric assumptions, differences in morphological traits among groups were analyzed using the Kruskal–Wallis test, followed by Dunn’s post hoc test for multiple comparisons. All data are presented as mean ± standard error (SE), and statistical significance was set at p < 0.05.

Probit regression analysis determined the median LC₅₀ of TCC for R. taihangensis tadpoles to be 343.708, 228.342, 197.477, and 169.863 μg/L at 24, 48, 72, and 96 h, respectively (see Supplementary Table S1, Supplementary Figure S1, and Table 1 for detailed data). Morphological traits, including body mass, total length, SVL and body condition score, were compared in the solvent control and TCC-treated groups (5, 15, 45 μg/L), with results presented in Figure 1. While no statistically significant differences in body mass (H = 3.006, df = 3, p = 0.391) or total length (H = 5.874, df = 3, p = 0.118) were observed between TCC-exposed and control tadpoles, a non-significant increasing trend in total length was detected with TCC exposure (Figures 1a,b). In contrast, the body condition score was reduced in all three TCC treatment groups compared to the control (H = 11.093, df = 3, p < 0.011), reaching statistical significance at 5 μg/L (p = 0.009) and 45 μg/L (p = 0.002) (Figure 1c). In addition, SVL was significantly increased in all TCC-treated groups relative to the control (H = 18.354, df = 3, p < 0.001) (Figure 1d). These morphological changes may further affect the normal physiological processes in tadpoles.

In the gut microbiota of R. taihangensis tadpoles, the dominant phyla across all groups (i.e., the three TCC treatments and the solvent control) were Pseudomonadota (Proteobacteria), Fusobacteriota, Bacillota (Firmicutes) and Bacteroidota (Figure 2a). At the genus level, the most abundant taxa were Xanthobacter, Cetobacterium, Ancylobacter, [Anaerorhabdus]_furcosa_group, Proteocatella and Aeromonas (Figure 2b). Compared to the solvent control, Kruskal–Wallis tests showed that low and medium TCC concentrations did not induce significant changes in the phylum-level abundance. However, the high-concentration TCC group exhibited a significant increase in Pseudomonadota and a significant decrease in Fusobacteriota (Supplementary Figure S2a), suggesting that Pseudomonadota and Fusobacteriota were more sensitive to TCC, whereas Bacillota and Bacteroidota appeared relatively insensitive. At the genus level, Cetobacterium and Citrobacter showed notable sensitivity to TCC. Cetobacterium abundance decreased significantly only under high TCC exposure, whereas Citrobacter was significantly reduced in all TCC-treated groups (Supplementary Figure S2b).

Alpha-diversity of the gut microbiota was evaluated using the Chao1, Simpson and Pielou’s evenness indices (Figure 2c). The Chao1 index, reflecting ASV richness, increased significantly (p < 0.01; Figure 2c) only in the low-concentration TCC group (TL, 5 μg/L). Overall, species richness was lowest in the solvent control group (CS) and highest in the TL group, declining gradually as TCC concentrations increased from TL to medium (TM, 15 μg/L) and high (TH, 45 μg/L) levels. Nevertheless, median richness in all three TCC-exposed groups remained above that of the CS group (Figure 2c). Pielou’s evenness index decreased progressively from CS to TL, TM, and TH (p = 0.0049), indicating a gradual reduction in community evenness relative to the control (CS) (Figure 2c). Similarly, the Simpson index, which reflects species dominance and distribution uniformity, also decreased continuously across the same gradient (p = 0.0029), with a statistically significant reduction observed only in the TH group (Figure 2c). Together, these indices revealed a shift in the gut microbiota structure: species richness initially increased then decreased, while dominance increased and evenness declined, reflecting a transition from a “basically balanced” state to “rich but imbalanced,” and finally to a “depleted and imbalanced” microbial community. Beta-diversity, evaluated by PCoA based on both weighted UniFrac and unweighted UniFrac distances, indicated that the high-concentration TCC group (TH) diverged significantly from the control in both metrics (p < 0.05, Figure 2d). Unweighted UniFrac analysis revealed significant structural shifts even in the TL and TM groups compared to the CS group (p < 0.05, Figure 2d).

LEfSe analysis was performed across multiple taxonomic levels (from phylum to genus) to identify specific bacterial taxa associated with TCC exposure. At the phylum level, Verrucomicrobiota and Pseudomonadota were microbial markers in the 5 μg/L and 45 μg/L TCC groups, respectively (Figure 3). At the genus level, Citrobacter was identified as a marker genus in the control group (Figure 3). In TCC-exposed tadpoles, the 5 μg/L group was characterized by enrichment of Exiguobacterium, Anaerofustis and Paracoccus; the marker microbial genera in 15 μg/L group were Tyzzerella, Pseudomonas and Leptolyngbya_RV74; while in the 45 μg/L group, the marker microbial genera were Citrobacter, Bilophila and Thiobacillus (Figure 3). Notably, the 5 μg/L of TCC exposure yielded a more distinct set of bacterial markers compared to the other exposure concentrations (Figure 3). In total, 19 microbial markers were identified at the genus level.

Non-targeted metabolomic profiling detected a total of 28,202 and 18,488 features in positive and negative ionization modes, respectively. OPLS-DA revealed a clear and distinct separation among all groups in both positive and negative modes (Figures 4a,b). Model quality had high robustness, as indicated by R²Y and Q² values close to 1 in positive (R²Y = 0.996, Q² = 0.705) and negative (R²Y = 0.995, Q² = 0.666) ion modes. Permutation tests further confirmed model validity and absence of overfitting (Figures 4c,d).

Using the criteria of VIP >1 and p < 0.05, 145 differentially expressed metabolites were identified in the low-concentration TCC group (37 down-regulated, 108 up-regulated), 124 in the medium-concentration group (24 down-regulated, 100 up-regulated), and 143 in the high-concentration group (18 down-regulated, 125 up-regulated) (Figures 5a–c). A Venn diagram showed that 77 metabolites were common to all three comparison groups (Figure 5d and Supplementary Table S2). These differential metabolites were primarily classified at the HMDB superclass level as lipids and lipid-like molecules, organic acids and derivatives, benzenoids and organoheterocyclic compounds (Supplementary Table S2). KEGG enrichment analysis of the 77 shared metabolites identified 21 significantly altered metabolic pathways. The top 20 pathways, ranked by significance, were presented in a bubble chart (Figure 6 and Supplementary Table S3). Among these, five pathways (Arachidonic acid metabolism, tyrosine metabolism, ABC transporters, purine metabolism, and biosynthesis of amino acids) were selected based on a p-value <0.05 and the inclusion of at least five differential metabolites. A total of 23 metabolites enriched in these pathways were sorted out (Supplementary Table S4).

An integrated multi-omics approach was performed to investigate the relationship between gut microbiota and metabolic alterations following TCC exposure. Across the three comparison groups, 210 differential metabolites were identified (Figure 5d). LEfSe analysis further detected 86 significant metabolic markers across all experimental groups (Supplementary Table S5). Spearman correlation analysis revealed 15 strongly correlated microbe-metabolite pairs (|rho| >0.7), comprising 12 positive and three negative correlations (Supplementary Table S6). Differences in the metabolic markers across groups are presented in Supplementary Figure S3.

Although chronic TCC exposure did not significantly affect body mass or total length of R. taihangensis tadpoles, it induced two conspicuous phenotypic changes: an increase in SVL and a reduction in body condition score. The significant increase in SVL, without a corresponding change in total length, suggested that TCC may disrupt developmental allometry. Such a premature shift in body proportions could desynchronize morphological development from critical environmental cues (e.g., food availability), potentially resulting in asynchronous metamorphosis—a known factor that reduces post-metamorphic survival (Székely et al., 2020). This risk is particularly critical for R. taihangensis, a species whose mountain-stream habitat already imposes strict constraints on metamorphic timing due to factors such as short growing seasons.

The observed reduction in body condition score, calculated as 100 × (body mass/body length³), indicated depleted somatic energy reserves (Ruthsatz et al., 2020). This energy deficit undermines the tadpoles’ capacity to cope with environmental stresses, including food deprivation, pathogen infection, and environmental pollution (Ruthsatz et al., 2020). For R. taihangensis, which inhabits food-limited mountain streams, this TCC-induced energy deficit could lead to heightened population vulnerability through reduced overwintering survival or metamorphic success. Furthermore, the non-significant trend of increased total length without a corresponding increase in body mass suggested that TCC might disrupt energy allocation, favoring linear growth over mass accumulation. This trade-off could compromise long-term fitness, given that body mass is a key determinant of post-metamorphic reproductive success in amphibians (Tejedo, 1992). Collectively, these morphological changes demonstrated that TCC disrupts the normal developmental trajectory of R. taihangensis tadpoles, with potential implications for individual survival and population persistence.

The intestinal microbiota plays a critical role in host nutrient absorption, digestion and immune function (Dawood and Koshio, 2016; Gao et al., 2020). Numerous studies have shown that environmental pollutants can alter gut microbiota diversity and composition (Lv et al., 2023; Zhang Q. et al., 2023; Zhu et al., 2023). In this study, Pseudomonadota, Fusobacteriota, Bacillota and Bacteroidota collectively constituted approximately 97% of the intestinal bacterial communities in R. taihangensis tadpoles. We found that TCC exposure significantly reduced gut microbial diversity and disrupted intestinal homeostasis, which aligns with previously reported effects in R. chensinensis (Yang et al., 2020b) and mice (Song et al., 2022; Zhang et al., 2021).

Non-targeted metabolomics revealed that TCC exposure disrupts the metabolome of R. taihangensis tadpoles, with 145, 124, and 143 differential metabolites identified in the low-, medium-, and high-concentration groups, respectively. Among these, 77 metabolites were shared across all TCC groups, primarily classified as lipids/lipid-like molecules, organic acids, and benzenoids. KEGG enrichment analysis of these shared metabolites identified five key perturbed pathways: arachidonic acid metabolism, tyrosine metabolism, ABC transporters, purine metabolism, and biosynthesis of amino acids. These pathways are critical for immune function, neuroendocrine signaling, energy homeostasis, and nutrient utilization—providing molecular insights into TCC’s sublethal toxicity.

To validate the link between gut microbiota and metabolic perturbations, we performed Spearman correlation analysis on LEfSe-identified microbial and metabolic markers (LDA score >2.0). This analysis revealed 15 strongly correlated microbe-metabolite pairs (|rho| >0.7), confirming that TCC-induced gut dysbiosis contributes to metabolic dysfunction.

Lactobacillus contributes to host health and microbial balance through multiple mechanisms, including immune homeostasis regulation (Rizzello et al., 2011), intestinal barrier maintenance (Zhao et al., 2021), metabolic function modulation (Teng et al., 2020) and pathogens inhibition (Alakomi et al., 2000). Our results revealed a significant negative correlation between Lactobacillus and Isovitexin (rho = −0.76, p < 0.001). This relationship is supported by recent studies demonstrating bidirectional interaction between the two: Lactobacillus can enhance the bioavailability of isovitexin by secreting β-glucosidases that hydrolyze its glycosidic form into more absorbable aglycones, while isovitexin selectively promotes the growth of Lactobacillus, thereby reinforcing its antioxidant and anti-inflammatory effects (Baky et al., 2022). In addition, Lactobacillus reuteri Zj617 has been shown to serve as a substrate source for microbial spermidine biosynthesis, counteracting the loss of spermidine-producing taxa and protecting against metabolic disorders (Ma et al., 2025). At the same time, we also found a significant positive correlation between Lactobacillus and Prostaglandin A1 (PGA1) (rho = 0.72, p < 0.001). In line with this, Zhao et al. (2022) reported that both Lactobacillus and PGA1 were positively associated with pro-inflammatory cytokines (e.g., TNF-α, IL-6, CXCL1) in brain tissues, suggesting that Lactobacillus may promote sepsis-related neuroinflammation via PGA1 up-regulation (Zhao et al., 2022). Conversely, a recent study by Shen et al. (2024) demonstrated that oral administration of Lactobacillus plantarum L168 increased PGA1 levels in a rat model of bronchopulmonary dysplasia (BPD), where both were implicated in anti-inflammatory and lung injury repair processes (Shen et al., 2024). Furthermore, Limosilactobacillus has been shown to protect intestinal epithelial integrity, alleviate inflammation (Xu et al., 2025), modulate oxidative responses (Choi et al., 2024), and regulate immune function (Abramov et al., 2022). A recent study reported that L-asparagine (L-ASNase, E.C.3.5.1.1) produced by Limosilactobacillus secaliphilus catalyzes the hydrolysis of L-asparagine to L-Aspartic acid and NH₃ (Zhang et al., 2024), consistent with our findings of strong positive correlation between Limosilactobacillus and L-aspartic acid (rho = 0.72, p < 0.001).

Integrating our findings across morphology, microbiome, and metabolome, we propose a causal chain explaining TCC’s sublethal toxicity to R. taihangensis tadpoles. (1) TCC exposure disrupts gut microbiota: TCC preferentially inhibits the growth of beneficial taxa (e.g., Cetobacterium, Citrobacter) and promotes the abundance of pro-inflammatory taxa (e.g., Pseudomonadota). This leads to reduced microbial diversity, impaired intestinal barrier integrity, and loss of key microbial functions (e.g., vitamin B₁₂ synthesis). (2) Microbial dysbiosis drives metabolic perturbations: The depletion of Cetobacterium reduces vitamin B₁₂ availability, impairing purine metabolism and leading to the accumulation of purine intermediates (e.g., hypoxanthine) and reduced ATP production. Concurrently, the decline in Citrobacter and other beneficial taxa disrupts amino acid biosynthesis and ABC transporter function, leading to perturbations in tyrosine metabolism (e.g., norepinephrine up-regulation) and arachidonic acid metabolism (e.g., pro-inflammatory metabolite up-regulation). (3) Metabolic perturbations induce phenotypic changes: Reduced ATP production leads to a decline in body condition score, as energy is diverted from somatic growth to stress responses. Disrupted tyrosine metabolism causes behavioral abnormalities (e.g., hyperactivity), increasing predation risk. Perturbed arachidonic acid metabolism impairs skin barrier function, enhancing TCC absorption and exposure to other contaminants. Additionally, increased SVL (metamorphic asynchrony) may result from energy allocation trade-offs, reducing post-metamorphic survival. This chain of events highlights the interconnectedness of biological levels: TCC’s impact on gut microbiota is a proximal driver of metabolic dysfunction, which in turn manifests as phenotypic toxicity. Importantly, this interplay is concentration-dependent: low TCC concentrations (5 μg/L) induce mild microbial shifts and metabolic compensation, while high concentrations (45 μg/L) lead to severe dysbiosis, irreversible metabolic damage, and significant phenotypic impairment.