Sources of bacteria in outdoor air across cities in the midwestern United States

Abstract

Bacteria are abundant in the atmosphere, where they often represent a major portion of the organic aerosols. Potential pathogens of plants and livestock are commonly dispersed through the atmosphere, and airborne bacteria can have important effects on human health as pathogens or triggers of allergic asthma and seasonal allergies. Despite their importance, the diversity and biogeography of airborne microorganisms remain poorly understood. We used high-throughput pyrosequencing to analyze bacterial communities present in the aerosol fraction containing fine particulate matter of ≤2.5 μm from 96 near-surface atmospheric samples collected from cities throughout the midwestern United States and found that the communities are surprisingly diverse and strongly affected by the season. We also directly compared the airborne communities to those found in hundreds of samples representing potential source environments. We show that, in addition to the more predictable sources (soils and leaf surfaces), fecal material, most likely dog feces, often represents an unexpected source of bacteria in the atmosphere at more urbanized locations during the winter. Airborne bacteria are clearly an important, but understudied, component of air quality that needs to be better integrated into efforts to measure and model pollutants in the atmosphere.

Fig. 1.

Most abundant bacterial groups identified using barcoded pyrosequencing at the phylum level (A) and at the order level (B). Proteobacterial groups are designated by the letters α, β, γ, and Δ for the Alphaproteobacteria, Betaproteobacteria, Gammaproteobacteria, and Deltaproteobacteria, respectively.

Fig. 2.

Relationships between bacterial communities in samples from the four locations (A) and across the summer and winter seasons (B). Bacterial communities were clustered using principal-coordinate analysis of the unweighted UniFrac distance matrix, with symbols that are closer together indicating air samples with more phylogenetically similar bacterial communities.

Fig. 3.

Total abundances of bacteria present in air samples collected from one small town (Mayville, WI) and three metropolitan cities (Chicago, IL; Detroit, MI; and Cleveland, OH) in the midwestern United States during the summer and winter. An asterisk denotes a statistically significant difference (P < 0.05) between summer and winter samples. Error bars indicate ±1 standard deviation.

Fig. 4.

Relationships between summer and winter air samples (maroon and blue, respectively) and samples collected from a variety of possible source environments, including soils, leaf surfaces and various animal fecal samples. Communities were clustered using principal-coordinate analysis of the unweighted UniFrac distance matrix from a combined sequence data set of approximately 630,000 sequences. (A) Midwestern air samples alongside all three potential source environments (soil, leaf surfaces, and feces). (Inset) The same principal-coordinate analysis plot rotated along the first and third principal-coordinate axes. (B) Phylogenetic similarity between midwestern airborne bacterial communities and the communities of various fecal samples, including dog feces.

Fig. 5.

Relative contributions of bacteria from likely source environments, i.e., soil, leaf surfaces, and animal feces. (A) Ternary plot representing the relative abundances of the indicator taxa in midwestern air samples color coded by season (maroon, summer; blue, winter), where the shading designates a given source environment (red, soil; green, leaf surface; brown, potential fecal origin). (B) Relative abundances of taxa indicative of the three source environments. (Inset) Relative abundances of these same indicator taxa in their native habitats, i.e., soil, leaf surfaces, and animal feces.