Data_Sheet_1_Changes in Carbon Oxidation State of Metagenomes Along Geochemical Redox Gradients.PDF

There is widespread interest in how geochemistry affects the genomic makeup of microbial communities, but the possible impacts of oxidation-reduction (redox) conditions on the chemical composition of biomacromolecules remain largely unexplored. Here we document systematic changes in the carbon oxidation state, a metric derived from the chemical formulas of biomacromolecular sequences, using published metagenomic and metatranscriptomic datasets from 18 studies representing different marine and terrestrial environments. We find that the carbon oxidation states of DNA, as well as proteins inferred from coding sequences, follow geochemical redox gradients associated with mixing and cooling of hot spring fluids in Yellowstone National Park (USA) and submarine hydrothermal fluids. Thermodynamic calculations provide independent predictions for the environmental shaping of the gene and protein composition of microbial communities in these systems. On the other hand, the carbon oxidation state of DNA is negatively correlated with oxygen concentration in marine oxygen minimum zones. In this case, a thermodynamic model is not viable, but the low carbon oxidation state of DNA near the ocean surface reflects a low GC content, which can be attributed to genome reduction in organisms adapted to low-nutrient conditions. We also present evidence for a depth-dependent increase of oxidation state at the species level, which might be associated with alteration of DNA through horizontal gene transfer and/or selective degradation of relatively reduced (AT-rich) extracellular DNA by heterotrophic bacteria. Sediments exhibit even more complex behavior, where carbon oxidation state minimizes near the sulfate-methane transition zone and rises again at depth; markedly higher oxidation states are also associated with older freshwater-dominated sediments in the Baltic Sea that are enriched in iron oxides and have low organic carbon. This geobiochemical study of carbon oxidation state reveals a new aspect of environmental information in metagenomic sequences, and provides a reference frame for future studies that may use ancient DNA sequences as a paleoredox indicator.