Microbial diversity and groundwater chemistry in a pristine aquifer

The Mahomet aquifer, a pristine water resource spanning much of central Illinois, is home to a diverse ecosystem of bacteria and archaea. We examined this microbial community at 18 wells across the aquifer by filtering cells suspended in groundwater and trapping sediment-bound populations on sterile samplers incubated in the wells themselves. After extracting DNA directly from these samples, we characterized both the suspended and attached communities with terminal restriction fragment length polymorphism (T-RFLP) and large-scale sequencing of the 16S rRNA gene. Using multivariate statistics, we quantified how the distribution of bacterial populations corresponds to the geochemical zonation of groundwater. Groundwater in this area of the Mahomet aquifer is divided into areas of high and low sulfate, and the concentration of dissolved methane in these areas is inversely proportional to that of sulfate. Although groundwater in every well we sampled was considerably richer in ferrous iron than sulfide, we found that the presence of iron-reducing bacteria does not exclude sulfate reducers as previously assumed. Instead, where the concentration of sulfate is high, sulfate reducers comprise a proportion of the bacterial community (18%) nearly equivalent to that of iron reducers (23%). Iron reducers only dominate the bacterial community in wells with sulfate concentrations of 0.2 mM or less, where clones classified as Geobacter, Geothrix, or Desulfuromonas represent more than half of all sequences analyzed. We observed, in contrast, no statistically significant relationship between the structure of the bacterial community and the concentration of either ferrous iron or dissolved sulfide. This result calls into question the use of iron and sulfide as indicators of the nature of subsurface bacterial activity. Sulfate reducers and iron reducers do not appear to be segregated into discrete zones in the aquifer, as commonly believed to occur as a result of competitive exclusion. Instead, we found the two groups co-existing in the subsurface in a relationship that we argue is not only competitive, but mutualistic. Through further use of sediment traps to evaluate the attached microbial community, we found that overall structure of bacterial communities in the Mahomet is resistant to a shift in the availability of sulfate. After one year of incubation within a well where the concentration of sulfate was either high (1.5 mM) or low (0.04 mM), sediment traps were switched between the two wells for an additional year of incubation. Despite the more than 40-fold change in the concentration of sulfate, these switched traps remained more similar to the community from the well in which they were initially incubated. While the overall community composition did not change significantly, certain bacterial groups associated with sulfate reduction (Desulfobacter and Desulfobulbus) were found to increase or decrease along with the concentration of sulfate. These results show that while the abundance of many populations is at least partly controlled by the evolutionary history of that particular community, certain functional groups of critical biogeochemical importance are sensitive to local changes. While measures of overall community similarity are useful, the specific abundance of these taxa must be accounted for