DOM composition alters ecosystem function during microbial processing of isolated sources

Dynamics of dissolved organic matter (DOM) in ecosystems are controlled by a suite of interacting physical, chemical, and biological factors. Growing recognition of the associations between microbial communities and metabolism and intrinsic DOM characteristics, highlight the potential importance of microbe-DOM relationships to modulate the role and fate of DOM, yet these relationships are difficult to isolate because they often operate across confounding environmental gradients. In a controlled laboratory incubation (44days), we integrated DOM bulk and molecular characterization, bacterial abundances, microbial assemblage composition, nutrient concentrations, and cellular respiration to discern the structural dynamics of biological processing among DOM sources from different allochthonous litters (grass, deciduous leaves, and evergreen needles). We identified two periods, consistent among DOM sources, where processing dynamics differed. Further, bulk fluorescent analyses showed shifts from low to high excitation and emission wavelengths, indicating the biological production of more complex/degraded materials over time. Molecular level analyses revealed similar temporal patterns among DOM sources in the production and consumption of individual chemical components varying in reactivity and heteroatomic content. Despite these similarities, total carbon (C)removed and carbon dioxide (CO2) accumulation differed by20% and 25% among DOM sources. This range in C processing was apparently tied to key chemical properties of the DOM (e.g., initial DOM composition, N content, and labile nature) as well as differential reorganization of the microbial populations that decomposed the DOM. We conclude that the production, transformation, and consumption of C in aquatic ecosystems is strongly dependent on the source and character of DOM as well as the structure of the microbial communities present, both of which change as DOM is processed over time. It is crucial that stream C processing models represent this complexity accurately.