Microhabitats Shape Bacterial Community Composition, Ecosystem Function, and Genome Traits

Abstract

This dissertation helps to integrate bacteria into the broader field of ecology by investigating bacterial community composition and diversity as it relates to ecosystem function in microhabitats within freshwater systems of the Great Lakes Region. Here, I combine field- and laboratory-based measurements of observational data collected from three major types of lake ecosystems: inland lakes, a freshwater estuary (Muskegon Lake), and a Great Lake (Lake Michigan). First, to determine the primary controls on lake bacterial community composition, I assessed the influence of lake layer (i.e. stratification), lake productivity, and particle-association on the bacterial community across 11 inland lakes with varying productivity in Southwestern Michigan. I found that particle-association very strongly structures freshwater bacterial community composition. Second, I studied a freshwater estuarine lake, Muskegon Lake, which has a large spatio-temporal variation in bacterial heterotrophic productivity, to test whether there was an association between heterotrophic production and bacterial biodiversity (defined as the number of taxa and taxon abundance). I specifically focused on two co-occurring freshwater habitats that my first chapter showed to be populated by very distinct communities: particle-associated and free-living. Positive biodiversity-heterotrophic productivity relationships were found only in particles. Third, I performed a genome-based analysis of free-living specialists, particle-associated bacterial specialists, and generalists to characterize the genomic architecture and genetic traits that are associated with adaptations to these specific habitats. The genomes of particle-associated specialist bacteria were about twice the size of the genomes of free-living specialists and generalists, which had streamlined genomes. Fourth, to identify the bacterial taxa driving heterotrophic productivity across the large set of lake samples, I found that high nucleic acid (i.e., HNA) functional groups identified by flow cytometry can serve as a proxy for freshwater bacterial heterotrophic productivity, whereas low nucleic acid (i.e., LNA) functional groups cannot. Then, I used a machine learning approach to identify bacterial taxa associated with HNA and LNA. This allowed me to identify the bacterial taxa, which were often members of the Phylum Bacteroidetes, that are associated heterotrophic productivity. Finally, I investigated patterns of lake specificity and phylogenetic conservation of taxonomic groups. Throughout my dissertation, I found that there was very deep (Class to Phylum-level) phylogenetic conservation of which bacteria lived in which habitats, but not of what bacterial taxa contributed to HNA and LNA functional groups, and thus heterotrophic productivity. Positive biodiversity-heterotrophic productivity relationships only existed in particle-associated, and not free-living communities, and communities composed of more phylogenetically related organisms were more productive per-capita. These differences in biodiversity-ecosystem function relationships may in part be explained by particle-associated bacteria having larger genomes, higher nitrogen content, and more unique genes that provide the potential for niche complementarity. The taxa that drove HNA and LNA cell numbers, and by proxy heterotrophic productivity, were lake and time-specific and indicated that taxa could switch between the two functional groups. Overall, my dissertation elucidates the ecological and evolutionary effects of microhabitat structure on bacterial communities and genomes in natural systems.