The assumption that antibiotic resistance spreads primarily through clinical medicine or direct antibiotic overuse is increasingly difficult to defend. Agricultural fields — where herbicides, fungicides, and organic amendments interact with living microbial ecosystems — may be quietly generating and disseminating resistance genes that eventually reach human pathogens. This is no longer a theoretical concern.
A controlled field study using 64 experimental plots of processing tomatoes systematically tested how dairy manure, glyphosate, copper, streptomycin, and the triazole fungicide propiconazole each shaped the soil and leaf microbiome, the abundance of culturable antibiotic-resistant bacteria (CARB), and the prevalence of 87 distinct antimicrobial resistance genes (ARGs). The results are counterintuitive in places: glyphosate-treated plots showed the lowest CARB counts yet harbored the highest leaf-level concentrations of specific resistance genes — notably tetA, tetB, OXA-50, and OXA-58. Dairy manure drove the strongest overall resistance burden across both soil and plant tissue, enriching genes including ACT-1, LAT, MIR, aadA1, and aphA6. Intriguingly, streptomycin and propiconazole treatments were associated with lower prevalence of multiple ARGs compared to untreated controls, a finding that complicates simple narratives about antibiotic use in agriculture.
This research sits at an important junction in AMR science. Herbicides like glyphosate have previously been shown in laboratory settings to induce stress responses that upregulate resistance pathways — this field-scale evidence adds meaningful ecological weight to that mechanism. The dissociation between culturable resistant bacteria and ARG prevalence under glyphosate is particularly significant: it suggests that resistance genes can propagate through horizontal gene transfer even when the resistant bacterial population appears suppressed. Limitations include the single-season, single-crop design, which restricts generalizability across climates and cropping systems. Still, the finding that a common fungicide can influence resistance gene ecology at a level comparable to antibacterial agents is potentially paradigm-shifting for how regulators approach pesticide risk assessment.
