1. ¹Department of Microbiology, Miami University, Oxford, OH, USA
2. ²Department of Civil and Environmental Engineering, Stanford University, Stanford, CA, USA
3. ³Department of Crop and Soil Sciences, Texas A & M University, College Station, TX, USA
4. ⁴Environmental Sciences Division, Oak Ridge National Laboratory, Oak Ridge, TN, USA
5. ⁵Institute for Environmental Genomics, University of Oklahoma, Norman, OK, USA
6. ⁶Department of Microbiology, Montana State University, Bozeman, MT, USA
7. ⁷Center for Biofilm Engineering, Montana State University, Bozeman, MT, USA
8. ⁸Virtual Institute of Microbial Stress and Survival (http://vimss.lbl.gov/)

Correspondence: MW Fields, Department of Microbiology, Center for Biofilm Engineering, Montana State University, 366 EPS Building, Bozeman, MT 59717, USA. E-mail: matthew.fields@erc.montana.edu

Received 25 April 2008; Revised 3 July 2008; Accepted 9 July 2008; Published online 4 September 2008.

Abstract

Bacterial community succession was investigated in a field-scale subsurface reactor formed by a series of wells that received weekly ethanol additions to re-circulating groundwater. Ethanol additions stimulated denitrification, metal reduction, sulfate reduction and U(VI) reduction to sparingly soluble U(IV). Clone libraries of SSU rRNA gene sequences from groundwater samples enabled tracking of spatial and temporal changes over a 1.5-year period. Analyses showed that the communities changed in a manner consistent with geochemical variations that occurred along temporal and spatial scales. Canonical correspondence analysis revealed that the levels of nitrate, uranium, sulfide, sulfate and ethanol were strongly correlated with particular bacterial populations. As sulfate and U(VI) levels declined, sequences representative of sulfate reducers and metal reducers were detected at high levels. Ultimately, sequences associated with sulfate-reducing populations predominated, and sulfate levels declined as U(VI) remained at low levels. When engineering controls were compared with the population variation through canonical ordination, changes could be related to dissolved oxygen control and ethanol addition. The data also indicated that the indigenous populations responded differently to stimulation for bioreduction; however, the two biostimulated communities became more similar after different transitions in an idiosyncratic manner. The strong associations between particular environmental variables and certain populations provide insight into the establishment of practical and successful remediation strategies in radionuclide-contaminated environments with respect to engineering controls and microbial ecology.