Bioaerosol biomonitoring: Sampling optimization for molecular microbial ecology

Abstract

Bioaerosols (or biogenic aerosols) have largely been overlooked by molecular ecologists. However, this is rapidly changing as bioaerosols play key roles in public health, environmental chemistry and the dispersal ecology of microbes. Due to the low environmental concentrations of bioaerosols, collecting sufficient biomass for molecular methods is challenging. Currently, no standardized methods for bioaerosol collection for molecular ecology research exist. Each study requires a process of optimization, which greatly slows the advance of bioaerosol science. Here, we evaluated air filtration and liquid impingement for bioaerosol sampling across a range of environmental conditions. We also investigated the effect of sampling matrices, sample concentration strategies and sampling duration on DNA yield. Air filtration using polycarbonate filters gave the highest recovery, but due to the faster sampling rates possible with impingement, we recommend this method for fine -scale temporal/spatial ecological studies. To prevent bias for the recovery of Gram-positive bacteria, we found that the matrix for impingement should be phosphate-buffered saline. The optimal method for bioaerosol concentration from the liquid matrix was centrifugation. However, we also present a method using syringe filters for rapid in-field recovery of bioaerosols from impingement samples, without compromising microbial diversity for high -throughput sequencing approaches. Finally, we provide a resource that enables molecular ecologists to select the most appropriate sampling strategy for their specific research question.

Keywords: airborne microorganisms; bioaerosol sampling; biomonitoring; filters; impingement; next - sequencing.

Figure 1

Graphical summary of the research questions addressed for aerosol biomonitoring. DI: deionized water; GF: glass fibre filter; Gel: gelatine filter; PBS: phosphate ‐buffered saline; PC: polycarbonate filter; Tris: Tris (HCl) buffer [Colour figure can be viewed at http://wileyonlinelibrary.com]

Figure 2

Overview of the five experiments carried out in this study. Gel filters were excluded from Experiment 1 due to DNA contamination (See Results section)

Figure 3

Recovery of Gram‐negative bacteria (a) or Gram ‐positive bacteria (b) from different filter matrices. GF: glass fibre filters; PC: polycarbonate filters, and Culture = direct addition of bacterial culture to a lysis tube for comparison. Each experiment was repeated, with the repeats shown on separate facets. Pairwise comparisons of least square means were perfumed using Tukey's HSD test, and the horizontal lines show significance levels between groups (NS = not significant, *p < 0.05, **p < 0.01, ***p < 0.001). The median is marked by the line that divides the boxes, the top and bottom of the box are the 75th and 25th percentiles respectively, and the whiskers shows the minimum and maximum values (n = 3). Gel filters showed visible bands at the expected amplicon size (230 bp) on 1% (w/v) agarose gel after PCR and were excluded

Figure 4

Recovery of Gram‐negative bacteria (a) or Gram‐positive bacteria (b) spiked into different liquid matrices. DI: deionized water; PBS: phosphate‐buffered saline; Tris: Tris (HCl) buffer, and Culture: direct addition of bacterial culture to a lysis tube for comparison. Each experiment was repeated, with the repeats shown on separate facets. Pairwise comparisons of least square means were perfumed using Tukey's HSD test, and the horizontal lines show significance levels between groups (NS = not significant, *p < 0.05, **p < 0.01, ***p < 0.001). The median is marked by the line that divides the boxes, the top and bottom of the box are the 75th and 25th percentiles respectively, and the whiskers show the minimum and maximum values (n = 3)

Figure 5

Recovery of bacteria from liquid impingement samples using different methods. The lab (a) and field (b) experiment. The facets on (b) separate the two sampling locations (University of Essex campus, and an arable farm). Pairwise comparisons of least square means were perfumed using Tukey's HSD test, and the horizontal lines show significance levels between groups (NS = not significant, *p < 0.05; **p < 0.01; ***p < 0.001). The median is marked by the line that divides the boxes, the top and bottom of the box are the 75th and 25th percentiles respectively, and the whiskers show the minimum and maximum values (n = 3)

Figure 6

Effect of recovery method from impingement samples on bacterial alpha diversity measures (box plots, a–d) and community composition (NMDS, e) from air samples collected by liquid impingement. In (a) to (d) (alpha diversity) the sampling sites are separated by facets. Pairwise comparisons of least square means were perfumed using Tukey's HSD test, and the horizontal lines show significance levels between groups (NS: not significant, *p < 0.05; **p < 0.01; ***p < 0.001). The median is marked by the line that divides the boxes, the top and bottom of the box are the 75th and 25th percentiles respectively, and the whiskers show the minimum and maximum value (n = 3). For (e) (NMDS) grey circles indicate recovery by centrifugation and black squares recovery by syringe filters [Colour figure can be viewed at http://wileyonlinelibrary.com]

Figure 7

Comparison of DNA yields for varying volumes of air collected with air filtration and liquid impingement. (a, b) Environment with high biomass; (c, d) an environment with low biomass. Pairwise comparisons of least square means were perfumed using Tukey's HSD test, and the horizontal lines show significance levels between groups (NS = not significant, *p < 0.05; **p < 0.01; ***p < 0.001). The median is marked by the line that divides the boxes, the top and bottom of the box are the 75th and 25th percentiles respectively, and the whiskers show the minimum and maximum value (n = 3)

Figure 8

Comparison of 16S rRNA gene copies recovered per m air with filters (black) and liquid impingement (grey) from a range of sites with varying environmental conditions (see Supporting information Table S1). The mean value is marked by the point and the whiskers are the minimum and maximum range for each site (n = 9) [Colour figure can be viewed at http://wileyonlinelibrary.com]

Figure 9

Decision flow chart for selecting the optimal air sampling method based on the results in this study [Colour figure can be viewed at http://wileyonlinelibrary.com]