Jointly considering multi-medium and full-cycle to better reveal distribution and removal of antibiotic resistance genes in long-term constructed wetland.

Constructed wetlands (CWs) have been proven to effectively remove antibiotic resistance genes (ARGs) at different experimental scales; however, there is still a lack of researches on the removal and monitoring of ARGs during the actual operation of full-scale CWs. To fill this gap, this study selected the Annan constructed wetland in Beijing as a case study and utilized quantitative sequencing, metagenomic analysis, and other technical methods to determine characteristics of ARGs in CWs during different operating periods. Furthermore, we analysed the overall removal characteristics of ARGs in the CW during different operating periods and differences of ARG distribution in three media. The dominant ARGs in the CW were quinolone, β-lactam and tetracycline, with subtypes of tufA and fusA. ARG distributions are significantly influenced by anthropic activities and seasonal changes. Three periods of the CW had good removal effects on special ARGs, but there were differences in the removal characteristics of different types and subtypes of ARGs. The CW had removal effects on four types of ARGs (such as multidrugs), 16 types of fusidic acid, and nine types of ARGs (such as bleomycin) during the dormancy, start-up, and operation periods, respectively. Among ARG subtypes, the CW had removal effects on 37, 53, and 51 subtypes during the dormancy, start-up, and operation periods, respectively. The subtypes that were removed mainly included those containing tetracycline, efflux pump, and β-lactam, mcr-1, and mcr-5 (colistin ARGs). For individual parts of CWs, the removal effects on the total abundance of ARGs were as follows: forebay > surface flow wetland > subsurface flow wetland. These findings provide insights for optimizing the purification efficiency of CWs for ARGs.