Evaluation of bacterial sampling methods for use with the bacterial tag-encoded flexible (FLX) amplicon pyrosequencing (bTEFAP) technique

Abstract

Background: The need to enumerate airborne microorganisms during infectious disease outbreaks, indoor air quality evaluations, and agricultural health studies has identified limitations in culture-based or viable sampling and characterization of bioaerosols. Pyrosequencing promises to be a novel, molecular-based technology that is exceptionally sensitive, low-cost, and provides a reasonable turnaround in the identification, distribution and concentration of aerosolized microorganisms. However, bioaerosol sampling methods for use with pyrosequencing have not been thoroughly evaluated. The intent of this project was to investigate a standardized sampling protocol for use with bTEFAP that would ultimately provide occupational scientists a novel and effective tool in the quest to characterize bioaerosol exposure and its subsequent relationship to worker health. Methods: Four filter types (Millipore Durapore® Membrane Filter, SKC water-soluble gelatin filter, SKC PTFE, SKC PVC) were prescreened for low-background DNA content using Pyrosequencing. Studies comparing the performance of the SKC Polyvinylchloride (PVC) and SKC gelatin filters in IOM samplers to an impinger - the SKC biosampler - were conducted in a previously characterized bioaerosol chamber using a Collision nebulizer. The challenge organism was a spore former, Bacillus atrophaeus. Tag-encoded FLX amplicon pyrosequencing analyses utilized Roche 454 FLX instrument with DNA extraction, massively parallel bTEFAP and bacterial identification data analysis was performed at the Research and Testing Laboratory (Lubbock, TX). Results: From an initial filter analysis, both the SKC PVC and SKC gelatin filters were selected for use in this project based on low-background DNA content, ease of use and cost. The two filter types and the SKC biosampler were challenged against B. atrophaeus for 30 minute sampling times in a series of six trials. Post pyrosequencing of detectable samples, it was demonstrated that the biosampler performed less effectively when compared to the PVC (p=0.0002) and gelatin filter (p=0.0006) based on an alpha value of 0.05. No significant difference was demonstrated between the two filter types (p=0.8). Of the original n=66 samples analyzed through pyrosequencing, only n=15 were reported to have counts for the challenge organism. In comparison to the pyrosequencing data, the cultured count demonstrated a significant difference when compared to the filters and biosampler media (p=0.003) in countable spores. Conclusions: The results indicate that with the model used in this study, the biosampler performed significantly different when compared to two filter types, the SKC PVC and the SKC gelatin, when challenged with B. atrophaeus. In addition, the microbial results suggest that there is possible significant contamination in the pyrosequencing methods used and or in the handling methods prior to analysis. Method analysis needs to be completed before further studies are completed.