Heating-season dynamics of the airborne microbiome, resistome and mobilome in Belgrade, Serbia

Abstract

Antimicrobial resistance (AMR) and air pollution are critical global health challenges, but their interplay remains poorly understood, particularly in Europe. Serbia, characterized by extensive antibiotic use, high prevalence of multidrug-resistant isolates and severe air pollution, provides a relevant model to study airborne AMR dissemination. During the heating season, air samples were collected at eight locations in Belgrade, representing industrial, traffic loaded and background environments. Shotgun metagenomics, co-occurrence networks and NMDS ordinations were applied to investigate the relationships between atmospheric pollutants, antibiotic resistance genes (ARGs), biocide resistance genes (BRGs), metal resistance genes (MRGs) and mobile genetic elements (MGEs). Autumn microbiomes were dominated by Lactococcus spp., whereas winter lacked such dominance. ARGs associated with antibiotic inactivation accounted for > 50% in autumn and > 75% in winter, with β-lactam resistance (blaTEM) predominating in both seasons. Winter resistomes also showed more consistent patterns of BRGs and MRGs, with multibiocide/acid and multimetal resistance prevailing. Integron analysis revealed predominance of class 1 integrons (intI1) commonly associated with Escherichia coli. Plasmid-related contigs were most similar to sequences reported in Acinetobacter baumannii and E. coli, while plasmid signatures related to Lactococcus lactis were also detected in autumn. Crucially, the network analysis revealed a seasonal restructuring of the airborne resistome. Autumn networks displayed fragmented structure, showing antagonism between Lactococcus and Escherichia, whereas winter networks coalesced into a densely interconnected superhub that could facilitate horizontal gene transfer and co-selection of resistance determinants. These findings suggest that prolonged air pollution and seasonality jointly shape airborne resistomes, reinforcing the need for integrated environmental and AMR surveillance in highly polluted urban areas.