Distributed Metagenomic Sampling of River Corridors Identifies Shared Microbial Functional Traits Underpinning Transferable Hydrobiogeochemical Processes

Abstract:

To accurately incorporate the membership and trait distribution patterns of microbes into reaction to watershed scale models, an understanding of microbial roles in hydrobiogeochemical processes is required. We used metagenomic sequencing obtained as part of the Worldwide Hydrogeochemistry Observation Network for Dynamic Rivers systems (WHONDRS) to identify metabolic potential harbored in rivers. This included pore- and surface-water samples from four rivers, located in Germany (Erpe River) and in the United States: Washington (Columbia River), Tennessee (East Fork Poplar Creek), and Georgia (Altamaha River). Student participants in the 2020 PNNL Multiscale Microbial Dynamics Modeling Summer School assembled, binned, and analyzed 204 Gbp of sequencing data using the KBase platform. This resulted in a bacterial and archaeal genome database of 169 medium and high-quality riverine metagenome-assembled genomes (MAGs). Microbial communities from surface and porewater within a river were more similar than sample types across rivers. Only 7 MAGs (>99.5% identity) were detected across rivers, however at broader taxonomic levels microbial membership was more conserved. MAGs of the genus Fonsibacter dominated in sediment and planktonic samples from all but one river, demonstrating the prevalence of aerobic heterotrophy in this ecosystem. The capacity for degrading polyphenolics and other litter derived carbohydrates was encoded in all samples, consistent with the ubiquity of these compounds in paired metabolite data. Our genomic analyses revealed this carbon was likely oxidized by microbial aerobic respiration and denitrification, using processes conserved across rivers. We identified broad taxonomic and metabolic patterns shared across rivers, despite finer-scale variances in hydrogeochemistry. Our findings demonstrate how community-engaged scientific efforts can be harnessed; these results will form the basis of the first Genome Resolved Open Watershed (GROW) database. Ultimately, data generated here will provide a framework enabling distributed modeling experiments across watersheds to reveal transferable principles.