Despite its emerging significant public health concern, the presence of antibiotic resistance genes (ARGs) in urban air has not received significant attention. Here, we profiled relative abundances (as a fraction, normalized by 16S rRNA gene) of 30 ARG subtypes resistant to seven common classes of antibiotics, which are quinolones, β-lactams, macrolides, tetracyclines, sulfonamides, aminoglycosides, and vancomycins, in ambient total particulate matter (PM) using a novel protocol across 19 world cities. In addition, their longitudinal changes in PM samples in Xi'an, China as an example were also studied. Geographically, the ARGs were detected to vary by nearly 100-fold in their abundances, for example, from 0.07 (Bandung, Indonesia) to 5.6 (San Francisco, USA). The β-lactam resistance gene blaTEM was found to be most abundant, seconded by quinolone resistance gene qepA; and their corresponding relative abundances have increased by 178% and 26%, respectively, from 2004 to 2014 in Xi'an. Independent of cities, gene network analysis indicates that airborne ARGs were differentially contributed by bacterial taxa. Results here reveal that urban air is being polluted by ARGs, and different cities are challenged with varying health risks associated with airborne ARG exposure. This work highlights the threat of urban airborne transmission of ARGs and the need of redefining our current air quality standards in terms with public health.

Air quality is often assessed using particulate matter (PM) mass concentration without considering its toxicity, thus possibly leading to improper control policies or inadequate health protection. Here, we studied differences in oxidative potentials (OPs) of PM samples collected using automobile air conditioning (AC) filters from 19 global cities, including influences from microbial contents. Dithiothreitol (DTT) assay showed remarkable differences in the PM OPs among cities (p-values <= 0.001, Kruskal-Wallis test). For example, the normalized index of oxidant generation (NIOG) of PM samples in San Francisco (2.20×10-2 , annual average PM 10 =16 µg/m 3) was found to be twice that in Beijing (1.14×10-2 , annual average PM 10 =135 µg/m 3). Limulus amebocyte lysate (LAL) assay found that PM-borne endotoxin ranged from 12.16 EU/mg (Florianopolis, Brazil) to 2518.23 EU/mg (Chennai, India) among cities. Besides, culturing method and real-time qPCR revealed significant differences up to ~100-fold in both bacterial and fungal levels among 19 cities. Spearman's correlation analysis implied that PM-borne microbes such as bacteria and fungi as well as metals could strongly influence the PM OP. As an example, our results in Xi'an, China further suggest that the PM 2.5 OP evolves for a particular city over the time, which is attributable to both the urbanization and air pollution control measures. This work highlights the importance in optimizing the current air quality control measures by considering the toxicity factor and its microbial constituents.

No abstract