Structural Variation in the Bacterial Community Associated with Airborne Particulate Matter in Beijing, China, during Hazy and Nonhazy Days

Abstract

The structural variation of the bacterial community associated with particulate matter (PM) was assessed in an urban area of Beijing during hazy and nonhazy days. Sampling for different PM fractions (PM_2.5 [<2.5 μm], PM₁₀ [<10 μm], and total suspended particulate) was conducted using three portable air samplers from September 2014 to February 2015. The airborne bacterial community in these samples was analyzed using the Illumina MiSeq platform with bacterium-specific primers targeting the 16S rRNA gene. A total of 1,707,072 reads belonging to 6,009 operational taxonomic units were observed. The airborne bacterial community composition was significantly affected by PM fractions (R = 0.157, P < 0.01). In addition, the relative abundances of several genera significantly differed between samples with various haze levels; for example, Methylobacillus, Tumebacillus, and Desulfurispora spp. increased in heavy-haze days. Canonical correspondence analysis and permutation tests showed that temperature, SO₂ concentration, relative humidity, PM₁₀ concentration, and CO concentration were significant factors that associated with airborne bacterial community composition. Only six genera increased across PM₁₀ samples (Dokdonella, Caenimonas, Geminicoccus, and Sphingopyxis) and PM_2.5 samples (Cellulomonas and Rhizobacter), while a large number of taxa significantly increased in total suspended particulate samples, such as Paracoccus, Kocuria, and Sphingomonas Network analysis indicated that Paracoccus, Rubellimicrobium, Kocuria, and Arthrobacter were the key genera in the airborne PM samples. Overall, the findings presented here suggest that diverse airborne bacterial communities are associated with PM and provide further understanding of bacterial community structure in the atmosphere during hazy and nonhazy days.IMPORTANCE The results presented here represent an analysis of the airborne bacterial community associated with particulate matter (PM) and advance our understanding of the structural variation of these communities. We observed a shift in bacterial community composition with PM fractions but no significant difference with haze levels. This may be because the bacterial differences are obscured by high bacterial diversity in the atmosphere. However, we also observed that a few genera (such as Methylobacillus, Tumebacillus, and Desulfurispora) increased significantly on heavy-haze days. In addition, Paracoccus, Rubellimicrobium, Kocuria, and Arthrobacter were the key genera in the airborne PM samples. Accurate and real-time techniques, such as metagenomics and metatranscriptomics, should be developed for a future survey of the relationship of airborne bacteria and haze.

Keywords: PM10; PM2.5; TSP; airborne bacterial community; haze; high-throughput sequencing.

FIG 1

Dominant bacterial groups identified at the phylum level (A) and at the genus level (B). Blue text, pink text, and black text for the sample dates refer to non-, light-, and heavy-haze sample days, respectively. D, E, and F represent PM_2.5, PM₁₀, and TSP, respectively. The category “others” includes all genera for which relative abundances are lower than 2%.

FIG 2

Venn diagrams illustrating the number of unique and shared OTUs among the three PM fractions (PM_2.5S, PM₁₀S, and TSPS) (A) and the three haze levels (non-, light, and heavy haze) (B).

FIG 3

LEfSe analysis illustrating differentially abundant bacterial genera among samples with different haze levels. LDA scores (only those genera that obtain a log LDA score of >2 are ultimately considered) can be interpreted as the degree of consistent difference in relative abundance between genera in non-, light-, and heavy-haze samples.

FIG 4

Canonical correspondence analysis showing the relationships between environmental factors and airborne bacterial community composition (A), and the relationships between environmental factors and dominant genera (B).

FIG 5

Network analysis indicating interactions among dominant bacteria.