Antibiotic resistance poses a significant threat to human health, and wastewater treatment plants (WWTPs) are important reservoirs of antibiotic resistance genes (ARGs). Here, we analyze the antibiotic resistomes of 226 activated sludge samples from 142 WWTPs across six continents, using a consistent pipeline for sample collection, DNA sequencing and analysis. We find that ARGs are diverse and similarly abundant, with a core set of 20 ARGs present in all WWTPs. ARG composition differs across continents and is distinct from that of the human gut and the oceans. ARG composition strongly correlates with bacterial taxonomic composition, with Chloroflexi, Acidobacteria and Deltaproteobacteria being the major carriers. ARG abundance positively correlates with the presence of mobile genetic elements, and 57% of the 1112 recovered high-quality genomes possess putatively mobile ARGs. Resistome variations appear to be driven by a complex combination of stochastic processes and deterministic abiotic factors.

In this study, 295 antibiotic resistance genes (ARGs) and mobile genetic elements (MGEs) from the influent, activated sludge (AS), and membrane bioreactor (MBR) permeate were primarily examined in the wastewater treatment plant (WWTP) biweekly over 13 months. The absolute concentrations of ARGs and MGEs respectively ranged from 1.27 × 1010 to 1.94 × 1011 and 8.00 × 109 to 1.24 × 1011 copies/L in the influent, of which were reduced by 2 to 3 orders of magnitude in the permeate. No significant seasonal variation of ARGs and MGEs was found in the WWTP, except that the absolute abundance of ARGs and MGEs in the AS was peaked during spring. The antibiotics affected neither ARGs nor MGEs significantly, suggesting their concentrations may be not high enough to pose a selective pressure. In contrast, the bacterial community had direct effect on the MGEs variation, meanwhile the MGEs influenced the ARG abundance directly. Class 1 integron-integrase gene (intI1), clinical intI1, and Tn21 associated more frequently with ARGs in the AS over long-term, suggesting the potential of them involved in horizontal gene transfer. Both intI1 and clinical intI1 had significantly positive associations with the overall abundance of ARGs, as well as significantly negative relationships with the overall removal rates of ARGs in the MBR. However, the abundances between intI1 and clinical intI1 were significantly different. Meanwhile, clinical intI1 remained rather consistent proportion with the ARG abundance in the AS and permeate, was stronger correlated with human pathogens, and was associated with greater number of ARGs over time. Moreover, clinical intI1 was significantly associated with the removal efficiency of ARGs from all classes. Taken together, clinical intI1 can be adopted as an indicator for the abundance and removal efficiency of ARGs in the WWTP.