Environmental contamination from synthetic chemicals like pesticides, industrial pollutants, and persistent organic compounds has increased significantly in recent decades. This rise has raised concerns about their largely overlooked impact on the human gut microbiome. Although many of these chemicals are not designed to target bacteria, they routinely reach the gastrointestinal (GI) tract through water, food, and environmental exposure, frequently at micromolar concentrations. It is essential to understand how chemical pollutants impact gut bacteria as the microbiome plays a central role in metabolism, immunity, and host physiology. While chemical safety assessments increasingly consider microbiome effects, existing data on xenobiotic-microbe interactions remain limited.
The aim of this study was to evaluate the effects of a large and chemically diverse set of pollutants on common human gut bacteria, to determine how these effects translate to microbial communities, and to identify the genetic determinants of pollutant sensitivity, and also explore whether chemical structures can predict antibacterial activity using machine learning (ML). The researchers aimed to provide foundational knowledge to evaluate environmental chemicals as possible disruptors of the gut microbiome by integrating high-throughput screening, community-level tests, chemical feature analysis, and bacterial genetics.
The researchers created a comprehensive library of 1076 chemicals relevant to the human diet and drinking water. This included 829 pesticides from 291 classes, 119 pesticide metabolites, 75 related compounds, 48 industrial chemicals like bisphenols and nitrosamines, and many mycotoxins. All compounds were screened at 20 ÎĽM, an environmentally relevant dosage, against 22 human gut bacterial strains, which represent the main phylogenetic and metabolic groups. Growth was monitored anaerobically for 24 hours, and inhibitory interactions were defined by statistically significant reductions in the area under the growth curve.
Further investigations examined pollutant absorption using LC-MS/MS to assess the impact on 20 20-species synthetic gut community and tested specific substances at many concentrations. They performed a genome-wide transposon mutant fitness screen in Parabacteroides merdae and validated conserved results in Bacteroides thetaiotaomicron to detect genetic mechanisms of xenobiotic tolerance. ML models based on chemical fingerprints and deep learning embeddings were trained to predict antibacterial activity.
The results showed widespread and earlier unrecognised antibacterial effects in environmental contaminants. Of the 1,076 chemicals tested, 168 inhibited at least one bacterial species with fungicides, industrial chemicals, and acaricides, which showed the highest rate of activity. Bacteroidale species were the most sensitive, whereas E. coli, Fusobacterium nucleatum, and Akkermansia muciniphila were more resistant. Most compounds showed narrow-spectrum effects, but 24 chemicals like closantel, the flame retardant TBBPA, chlordecone, and bisphenol AF showed broad toxicity across multiple strains. Many pollutants inhibited key beneficial microbes like Eubacterium rectale at concentrations below 2.5 ÎĽM, which indicates strong potency at real-world exposure levels. Pollutant sensitivity correlated with previously reported sensitivity to human-targeted drugs, and this suggests shared, nonspecific mechanisms of chemical interaction. In community experiments, chemicals like TBBPA and BPAF changed species composition with bioaccumulation, providing cross-protection to some species.
The ML models achieved a balanced accuracy of about 0.77 to predict toxicity in the pesticide dataset. Predictive performance was lower for mixed pesticide–drug datasets due to limited chemical overlap. Genetic screening revealed that efflux pumps, specifically RND family transporters and their regulators, are key determinants of resistance. Loss-of-function mutations in the acrR regulatory gene enhance resistance to TBBPA, closantel, and antibiotic ciprofloxacin. Additional conserved tolerance mechanisms were detected, like components related to ATP synthase and lipid A modification. Exposure to pollutants appeared to induce mutations that disrupt metabolic pathways, potentially impacting the production of beneficial metabolites, which are important for host cardiovascular and immune health.
Overall, this study demonstrated that many environmental chemicals, including pesticides, industrial pollutants, and continuous compounds, possess unanticipated antibacterial activity capable of reshaping gut microbiota composition, metabolism, and antibiotic resistance profiles. The findings identify efflux systems as a key defense mechanism that relates to pollutant exposure to cross-resistance against clinically important antibiotics. The researchers underscore the major gap in current chemical safety assessment, which often overlooks microbiome impacts, and advocate for integrating computational and experimental approaches to improve hazard evaluation and guide the development of safer chemicals. It provides a comprehensive framework for understanding how environmental pollutants influence the gut microbiome and, ultimately, human health.
Reference: Roux I, Lindell AE, GrieĂźhammer A, et al. Industrial and agricultural chemicals exhibit antimicrobial activity against human gut bacteria in vitro. Nat Microbiol. 2025. doi:10.1038/s41564-025-02182-6
