AUTHOR=Byappanahalli Muruleedhara N. , Nevers Meredith B. , Shively Dawn , Nakatsu Cindy H. , Kinzelman Julie L. , Phanikumar Mantha S. TITLE=Influence of Filter Pore Size on Composition and Relative Abundance of Bacterial Communities and Select Host-Specific MST Markers in Coastal Waters of Southern Lake Michigan JOURNAL=Frontiers in Microbiology VOLUME=12 YEAR=2021 URL=https://www.frontiersin.org/journals/microbiology/articles/10.3389/fmicb.2021.665664 DOI=10.3389/fmicb.2021.665664 ISSN=1664-302X ABSTRACT=

Water clarity is often the primary guiding factor in determining whether a prefiltration step is needed to increase volumes processed for a range of microbial endpoints. In this study, we evaluate the effect of filter pore size on the bacterial communities detected by 16S rRNA gene sequencing and incidence of two host-specific microbial source tracking (MST) markers in a range of coastal waters from southern Lake Michigan, using two independent data sets collected in 2015 (bacterial communities) and 2016–2017 (MST markers). Water samples were collected from river, shoreline, and offshore areas. For bacterial communities, each sample was filtered through a 5.0-μm filter, followed by filtration through a 0.22-μm filter, resulting in 70 and 143 filter pairs for bacterial communities and MST markers, respectively. Following DNA extraction, the bacterial communities were compared using 16S rRNA gene amplicons of the V3–V4 region sequenced on a MiSeq Illumina platform. Presence of human (Bacteroides HF183) and gull (Gull2, Catellicoccus marimammalium) host-specific MST markers were detected by qPCR. Actinobacteriota, Bacteroidota, and Proteobacteria, collectively represented 96.9% and 93.9% of the relative proportion of all phyla in the 0.22- and 5.0-μm pore size filters, respectively. There were more families detected in the 5.0-μm pore size filter (368) than the 0.22-μm (228). There were significant differences in the number of taxa between the two filter sizes at all levels of taxonomic classification according to linear discriminant analysis (LDA) effect size (LEfSe) with as many as 986 taxa from both filter sizes at LDA effect sizes greater than 2.0. Overall, the Gull2 marker was found in higher abundance on the 5.0-μm filter than 0.22 μm with the reverse pattern for the HF183 marker. This discrepancy could lead to problems with identifying microbial sources of contamination. Collectively, these results highlight the importance of analyzing pre- and final filters for a wide range of microbial endpoints, including host-specific MST markers routinely used in water quality monitoring programs. Analysis of both filters may increase costs but provides more complete genomic data via increased sample volume for characterizing microbial communities in coastal waters.