Longitudinal survey of microbiome associated with particulate matter in a megacity

Abstract

Background: While the physical and chemical properties of airborne particulate matter (PM) have been extensively studied, their associated microbiome remains largely unexplored. Here, we performed a longitudinal metagenomic survey of 106 samples of airborne PM_2.5 and PM₁₀ in Beijing over a period of 6 months in 2012 and 2013, including those from several historically severe smog events.

Results: We observed that the microbiome composition and functional potential were conserved between PM_2.5 and PM₁₀, although considerable temporal variations existed. Among the airborne microorganisms, Propionibacterium acnes, Escherichia coli, Acinetobacter lwoffii, Lactobacillus amylovorus, and Lactobacillus reuteri dominated, along with several viral species. We further identified an extensive repertoire of genes involved in antibiotic resistance and detoxification, including transporters, transpeptidases, and thioredoxins. Sample stratification based on Air Quality Index (AQI) demonstrated that many microbial species, including those associated with human, dog, and mouse feces, exhibit AQI-dependent incidence dynamics. The phylogenetic and functional diversity of air microbiome is comparable to those of soil and water environments, as its composition likely derives from a wide variety of sources.

Conclusions: Airborne particulate matter accommodates rich and dynamic microbial communities, including a range of microbial elements that are associated with potential health consequences.

Keywords: Air pollution; Archaea; Bacteria; Eukaryotes; Microbiome; Particulate matter (PM); Viruses.

Fig. 1

Taxonomic and functional characteristics of air microbiota. a Temporal distribution of daily PM concentration variations during the sampling period. b Relative abundance of different domains in air microbiome. c Taxonomic Shannon index of the PM_2.5 (red) and PM₁₀ (blue) samples. d Gene number of the PM_2.5 (red) and PM₁₀ (blue) samples. e Temporal distribution of relative abundance from the top 10 most abundant phyla across the sampling period of the PM_2.5 (left) and PM₁₀ (right) samples. Asterisks denote Wilcoxon signed-rank test results; P values were adjusted using Benjamini and Hochberg false discovery rate (FDR) (**adjusted P < 0.01)

Fig. 2

Characteristics of drug resistance and detoxification genes in PM samples. a Box plot showing the numbers of antibiotic resistance gene types in PM_2.5 (red) and PM₁₀ (blue) samples. b Box plot showing the RPKM values of total antibiotic resistance genes in PM_2.5 (red) and PM₁₀ (blue) samples. c, d Box plots showing the top 10 most abundant antibiotic resistance targets (c) and types (d) across PM_2.5 (red) and PM₁₀ (blue) samples. Labels 1–10 represent TEM beta-lactamase, major facilitator superfamily (MFS) antibiotic efflux pump, resistance-nodulation-cell division (RND) antibiotic efflux pump, Erm 23S ribosomal RNA methyltransferase, tetracycline-resistant ribosomal protection protein, lincosamide nucleotidyltransferase (LNU), sulfonamide resistant sul, ABC-F ATP-binding cassette ribosomal protection protein, chloramphenicol acetyltransferase (CAT), and ANT (6), respectively. e Bar plot showing the numbers of detoxification genes in PM_2.5 (red) and PM₁₀ (blue) samples. f Box plot showing the relative abundance of detoxification genes across PM_2.5 (red) and PM₁₀ (blue) samples. g, h Box plot showing the number of antibiotic resistance gene types (g) and RPKM values of the total antibiotic resistance gene types (h) across different environments. i, j Box plot showing the number of detoxification gene types (i) and RPKM values of the total detoxification gene types (j) across different environments. Asterisks denote Wilcoxon signed-rank test results; P values were adjusted using Benjamini and Hochberg false discovery rate (FDR) (*adjusted P < 0.05; **adjusted P < 0.01; ***adjusted P < 0.001)

Fig. 3

Comparative analysis for five different classes of PM_2.5 and PM₁₀ samples. a Stratification of classes I–V for PM_2.5 and PM₁₀ samples. b, c PCoA analysis based on the Bray-Curtis distance metric of species abundance in PM_2.5 (b) and PM₁₀ (c) samples. d–g Taxonomic species number (d, e) and taxonomic Shannon index (f, g) for PM_2.5 (d, f = red) and PM₁₀ (e, g = blue) samples, respectively. h–k Gene number (h, i) and gene Shannon index (j, k) for PM_2.5 (red) and PM₁₀ (blue) samples, respectively. Asterisks denote Wilcoxon rank-sum test results (**P < 0.05; ***P < 0.01). l Pairwise Spearman’s correlation matrix of the portion of airborne microorganisms associated with different environmental sources correlating with PM concentrations (*adjusted P < 0.05; **adjusted P < 0.01)