Two acid mine drainage (AMD) sites in the Appalachian Bituminous Coal Basin were selected to enrich for Fe(II)-oxidizing microbes and measure rates of low-pH Fe(II) oxidation in chemostatic bioreactors. Microbial communities were enriched for 74 to 128 d in fed-batch mode, then switched to flow-through mode (additional 52 to 138 d) to measure rates of Fe(II) oxidation as a function of pH (2.1 to 4.2) and influent Fe(II) concentration (80 to 2400 mg/L). Biofilm samples were collected throughout these operations and the microbial community structure was analyzed to evaluate impacts of geochemistry and incubation time. Alpha diversity decreased as pH decreased and as the Fe(II) concentration increased, coincident with conditions that attained the fastest rates of Fe(II) oxidation. The distribution of the seven most abundant bacterial genera could be explained by a combination of pH andFe(II) concentration. Acidithiobacillus,Ferrovum,Gallionella,Leptospirillum,Ferrimicrobium,Acidiphilium, andAcidocellawere all found to be restricted within specific bounds of pH and Fe(II) concentration. Temporal distance, defined as the cumulative number of pore volumes from the start of flow-through mode, appeared to be as important as geochemical conditions in controlling microbial community structure. Both alpha and beta diversity of microbial communities were significantly correlated to temporal distance in the flow-through experiments. Even after long-term operation under near-identical geochemical conditions, microbial communities enriched from the different sites remained distinct. While these microbial communities were enriched from sites that displayed markedly different field rates of Fe(II) oxidation, rates of Fe(II) oxidation measured in laboratory bioreactors were essentially the same. These results suggest that the performance of suspended-growth bioreactors for AMD treatment may not be strongly dependent on the inoculum used for reactor startup.
Importance This study showed that different microbial communities enriched from two sites maintained distinct microbial community traits inherited from their respective ‘seed' materials. Long-term operation (up to 128 d fed-batch enrichment followed by up to 138 d flow-through experiments) of these two systems did not lead to the same, or even more similar, microbial communities. However, these bioreactors did oxidize Fe(II) and remove Fe(T) at very similar rates. These results suggest that the performance of suspended-growth bioreactors for AMD treatment may not be strongly dependent on the inoculum used for reactor startup. This would be advantageous because system performance should be well constrained and predictable for many different sites.