Antibiotic resistance has an obvious suspect: antibiotics. The more we use them across human and animal medicine, the more we select for the bacteria that survive them (1). But that pressure doesn’t only come from the drugs we designed to kill bacteria. A growing line of research points to the fields, soils and waterways where our food is grown, and to chemicals that were never meant to be antibiotics at all (2,3). It is a reminder of what researchers mean by One Health: human, animal and environmental health are one connected system, not three separate problems, and resistance travels the connections between them (1,3).
Two fields, one experiment
A study in Frontiers in Microbiology makes this case. Clinical and environmental resistance research usually run on separate tracks, with different standard methods, so this team did something rarer: they put both under identical glyphosate exposure (4). Alongside multidrug-resistant isolates from hospital infections, they tested environmental bacteria from a protected wetland not directly exposed to herbicides (4).
The clinical strains told the clearest story. Every multidrug-resistant isolate tolerated glyphosate at high concentrations, while the environmental bacteria spread across a gradient from sensitive to highly resistant, with the strongest effects coming from a formulated herbicide rather than the pure compound (4). The revealing part: the environmental strains at the resistant end were the ones most closely related to the clinical pathogens (4). Enterobacter, a genus that already worries hospital infection-control teams, appeared repeatedly among the most glyphosate-resistant (4).
The shared machinery
You would expect glyphosate resistance to come from changes to its target, an enzyme called EPSPS (4). Instead, the genomes pointed to something more general: efflux pumps and degradative genes, which mattered more to resistance than the target enzyme did (4). Efflux pumps in particular, the machinery bacteria use to push toxic molecules out, also expel antibiotics, which the authors flag as the likely route of cross-selection (4). A bacterium that survives glyphosate by pumping it out may already carry the equipment to survive antibiotics the same way.
This isn’t a one-study curiosity. In soil experiments, three common herbicides, glyphosate, glufosinate and dicamba, increased the abundance of antibiotic resistance genes and mobile genetic elements, with glyphosate strongest at roughly a ninefold rise over 60 days (2). The likely mechanisms were selection for tolerant genotypes and increased movement of resistance plasmids between bacteria (2). The pattern held in the field, too: across 11 Chinese provinces, glyphosate use and residue levels tracked with higher resistance-gene abundance versus herbicide-free soils (2).
The route runs through water
The geography sharpens the concern. The Argentine researchers mapped where soybean cultivation and glyphosate use overlap with dense population and with carbapenem-resistant Klebsiella pneumoniae in bloodstream infections (4). That overlap doesn’t prove cause and effect, but it sketches out a plausible route, and the route runs through water: clinical bacteria enter the environment through wastewater, and irrigation, horticulture and livestock runoff all circulate resistant strains and genes (4). This is why the environment now sits at the center of how scientists think about resistance, which moves between humans, animals and the environment (3). Antimicrobial use in agriculture is itself a recognized driver, and the same manure and wastewater streams that carry resistance out of farms and herds carry it back toward people (1).
What the study shows
It’s worth being precise to the research article. Across the full set of strains, the researchers found no blanket correlation between glyphosate resistance and antibiotic resistance (4). The signal was narrower: clinical strains were uniformly glyphosate-resistant, the most resistant environmental strains were genetically close to them and shared efflux machinery is the likely link (4). The lab concentrations were also high, above typical field rates, though residues in that range have turned up in some real soils (4). The authors call this association plus a credible mechanism, not proof, and name glyphosate a potential driver (4).
What comes next
What comes next, the authors argue, is genomic surveillance that tracks bacteria and mobile genetic elements as they move through irrigation water, horticulture and livestock wastewater (4). That rests on a practical foundation: extracting and identifying bacterial DNA reliably and at scale, including from the water and wastewater where these genes travel. The researchers identified their environmental strains using Promega’s GoTaq® DNA Polymerase (4), and the broader toolkit for this work could run from automated extraction built for environmental samples, like the Maxwell® RSC platform, through the sequencing and quantitation that resistance surveillance depends on.
Conclusion
The shift these findings argue for is simple. When we assess a chemical for environmental risk, we ask what it does to its target: the weed, the pest, the plant. This work adds a second question: what is it doing to everything else living in that soil and water? Resistance is usually framed as a problem of medicine and stewardship. Evidence like this suggests it belongs to agriculture, animal health and water quality too.
References
- McEwen SA, Collignon PJ (2018) Antimicrobial resistance: a One Health perspective. Microbiol. Spectr.6:ARBA-0009-2017. doi: 10.1128/microbiolspec.ARBA-0009-2017
- Liao H et al. (2021) Herbicide selection promotes antibiotic resistance in soil microbiomes. Mol. Biol. Evol.38:2337–2350. doi: 10.1093/molbev/msab029
- Larsson DGJ, Flach CF (2022) Antibiotic resistance in the environment. Nat. Rev. Microbiol. 20:257–269. doi: 10.1038/s41579-021-00649-x
- Knecht CA et al. (2026) Glyphosate resistance as a potential driver for the dissemination of multidrug-resistant clinical strains. Front. Microbiol. 17:1740431. doi: 10.3389/fmicb.2026.1740431
