Unique Ecophysiology among U(VI)-Reducing Bacteria as Revealed by Evaluation of Oxygen Metabolism in Anaeromyxobacter dehalogenans Strain 2CP-C

ABSTRACT Anaeromyxobacter spp. respire soluble hexavalent uranium, U(VI), leading to the formation of insoluble U(IV), and are present at the uranium-contaminated Oak Ridge Integrated Field Research Challenge (IFC) site. Pilot-scale in situ bioreduction of U(VI) has been accomplished in area 3 of the Oak Ridge IFC site following biostimulation, but the susceptibility of the reduced material to oxidants (i.e., oxygen) compromises long-term U immobilization. Following oxygen intrusion, attached Anaeromyxobacter dehalogenans cells increased approximately 5-fold from 2.2 × 107 ± 8.6 × 106 to 1.0 × 108 ± 2.2 × 107 cells per g of sediment collected from well FW101-2. In the same samples, the numbers of cells of Geobacter lovleyi, a population native to area 3 and also capable of U(VI) reduction, decreased or did not change. A. dehalogenans cells captured via groundwater sampling (i.e., not attached to sediment) were present in much lower numbers (<1.3 × 104 ± 1.1 × 104 cells per liter) than sediment-associated cells, suggesting that A. dehalogenans cells occur predominantly in association with soil particles. Laboratory studies confirmed aerobic growth of A. dehalogenans strain 2CP-C at initial oxygen partial pressures (pO2) at and below 0.18 atm. A negative linear correlation [μ = (−0.09 × pO2) + 0.051; R2 = 0.923] was observed between the instantaneous specific growth rate μ and pO2, indicating that this organism should be classified as a microaerophile. Quantification of cells during aerobic growth revealed that the fraction of electrons released in electron donor oxidation and used for biomass production (fs) decreased from 0.52 at a pO2 of 0.02 atm to 0.19 at a pO2 of 0.18 atm. Hence, the apparent fraction of electrons utilized for energy generation (i.e., oxygen reduction) (fe) increased from 0.48 to 0.81 with increasing pO2, suggesting that oxygen is consumed in a nonrespiratory process at a high pO2. The ability to tolerate high oxygen concentrations, perform microaerophilic oxygen respiration, and preferentially associate with soil particles represents an ecophysiology that distinguishes A. dehalogenans from other known U(VI)-reducing bacteria in area 3, and these features may play roles for stabilizing immobilized radionuclides in situ.

The U.S. Department of Energy (DOE) has initiated efforts to remediate 120 uranium-contaminated locations in 36 states and territories that have been impacted by former nuclear weapon production sites (25). Several dissimilatory iron-reducing bacteria (DIRB) are capable of reducing soluble U(VI) to sparingly soluble U(IV), which often precipitates as uraninite (UO 2 ) (13,14,35,51), and the feasibility of microbial U(VI) reduction and immobilization as a containment strategy has been demonstrated at field sites (57,58). A remaining challenge is controlling oxic/anoxic interface processes to ensure the long-term stability of the precipitated material because U(IV) is susceptible to reoxidation by oxidants (e.g., oxygen) (22,52,56). DIRB grow by coupling the oxidation of organic compounds or H 2 to the reduction of ferric iron and often reduce other oxidized metals such as manganese and U as well (2,17,29,30,42,44,51). Metals reduced by DIRB undergo rapid cycling between reduced and oxidized states at oxic/ anoxic transition zones due to fluctuating oxygen gradients (4). Organisms capable of respiring oxygen, as well as oxidized metals, are thus expected to be good competitors in fluctuating redox environments, taking advantage of oxygen and oxidized metals as electron acceptors (11). Several DIRB have been shown to consume oxygen, ranging from facultative aerobes (e.g., Shewanella spp.) to those that reduce oxygen only when present at very low partial pressures (e.g., some Desulfovibrio spp.) (20,23).
Recent studies demonstrated that Anaeromyxobacter dehalogenans strain 2CP-C and Geobacter lovleyi grow with U(VI) as an electron acceptor (44,54), and Anaeromyxobacter and Geobacter lovleyi 16S rRNA gene sequences were detected at U-contaminated sites, including the Oak Ridge Integrated Field Research Challenge (IFC) site (5,36,39,48). In the current study, we enumerated Anaeromyxobacter and Geobacter lovleyi 16S rRNA gene sequences in area 3 at the Oak Ridge IFC site before (i.e., during ethanol biostimulation) and after oxygen intrusion to monitor their responses to changing redox conditions.
A few reports have mentioned growth of Anaeromyxobacter spp. under aerobic conditions (7,43,50), and the analysis of the A. dehalogenans strain 2CP-C genome supports observations that this organism is capable of oxygen utilization (49); however, we found that aerobic growth was unreliable (e.g., did not occur when the culture flasks were rigorously shaken). To better understand the oxygen metabolism of A. dehalogenans, provide growth yield and growth rate data, and speculate about the potential role of oxygen-consuming DIRB for U bioremediation, detailed laboratory experiments were performed at various oxygen partial pressures (pO 2 ). The results reveal unique physiological adaptations of A. dehalogenans to oxygen, which distinguish members of this bacterial group from other DIRB, and suggest that the ecophysiology of A. dehalogenans has implications for in situ U immobilization.

MATERIALS AND METHODS
Oak Ridge IFC site samples. Groundwater samples were collected by Oak Ridge National Laboratory (ORNL) personnel from a pilot-scale U bioreduction pilot plot in area 3 (55)(56)(57) and processed by Joy Van Nostrand at the University of Oklahoma as described previously (60). Multiport sampling wells FW101 and FW102 are located in the inner circulation loop of the bioreduction test plot (31). FW101 and FW102 wells were sampled before and after oxygen intrusion into the pilot plot area at screened intervals 2 (13.7 m below ground surface [bgs]) and 3 (12.2 m bgs) (see Fig. S1 in the supplemental material) (56). Cessation of electron donor (ethanol) and reductant (sulfite) additions initiated the intrusion of aerobic groundwater, which occurred between days 806 and 884, and samples were taken during test plot operation on days 746, 754, 810, 887, and 901 (see Fig. S1 in the supplemental material) (56). Area 3 sediment samples were collected by ORNL personnel using a well surging technique as described previously (5). Sediment samples from wells FW101 and FW102 were obtained on day 774 during the phase of active biostimulation with ethanol (days 755 to 805), which preceded oxygen intrusion (days 806 to 884) ( Fig. S1) (54,55,57). Sediment samples were collected again on day 935 during weekly ethanol stimulation, about 7 weeks after termination of the 11-week introduction of aerobic groundwater into the pilot plot (days 806 to 884) (Fig. S1). Sediments from slurries were collected by centrifugation, transferred aseptically to 50-ml Falcon plastic tubes, frozen at Ϫ80°C, shipped on dry ice, and stored at Ϫ80°C until DNA extraction (5).
Oxygen growth experiments. For cultivation with oxygen as an electron acceptor, 50 mM HEPES (4-[2-hydroxyethyl]-1-piperazine-ethanesulfonic acid [pH 7.2]; Sigma-Aldrich) replaced the bicarbonate buffer system, the acetate concentration was increased to 10 mM, and fumarate was omitted. Medium (25 ml) was dispensed into 160-ml glass serum bottles, resulting in a N 2 /CO 2 headspace (80:20, vol/vol) of 135 ml. Filter-sterilized (0.2-m membrane syringe filter; Pall Life Sciences, Ann Arbor, MI) air (16,30, and 60 ml [i.e., 3.4, 6.3, and 13 ml oxygen, respectively]) was added using a 60-ml syringe to achieve initial pO 2 values of 0.02, 0.03, and 0.07 atm, respectively. Nitrogen gas was added to balance pressure differences across treatments. To achieve higher initial pO 2 values, serum bottles containing sterilized, anaerobic medium were opened, loosely covered with sterile aluminum foil, shaken at 220 rpm for 3 h exposed to air, and resealed with sterile rubber stoppers. Sterile air (60 ml) was added via syringe to achieve equal overpressure in all treatments. The initial pO 2 in these culture vessels was 0.178 Ϯ 0.008 atm (referred to as the 0.18-atm O 2 treatment). Triplicate uninoculated vessels corresponding to each pO 2 treatment served as controls to account for oxygen loss and abiotic oxygen reduction. Triplicate, oxygen-free control cultures received 60 ml of sterile nitrogen gas instead of oxygen. Growth at atmospheric pO 2 was analyzed in foam-stoppered 250-ml Erlenmeyer flasks containing 25 ml medium. For the quantitative growth experiments, triplicate cultures for each pO 2 treatment and oxygen-free controls received 4% (vol/vol) inocula from stationary-phase, fumarate-grown cultures. Oxygen-grown cultures were repeatedly transferred to aerobic mineral salts medium amended with acetate as an electron donor to verify that the ability to respire oxygen was maintained. All culture vessels were shaken at 220 rpm and incubated at 30°C in the dark. Samples were analyzed periodically for organic acids (including acetate), oxygen, and cell numbers via quantitative real-time PCR (qPCR; see below).
DNA extraction. DNA from laboratory cultures was extracted using InstaGene matrix (Bio-Rad Laboratories, Hercules, CA). Aliquots (0.5 ml) from laboratory cultures were transferred to 1.5-ml plastic microcentrifuge tubes (Westbury, NY), and cells were collected by centrifugation at 16,000 ϫ g for 3 min at room temperature. The supernatant was decanted, and the pellets were frozen at Ϫ20°C for at least 24 h. InstaGene matrix solution (0.2 ml; Bio-Rad Laboratories, Hercules, CA) was added to each frozen pellet, and DNA was prepared according to the manufacturer's specifications. The DNA was utilized immediately or stored at Ϫ20°C until qPCR analysis.
DNA extraction from area 3 sediment samples used 500-mg aliquots of sediment with the FastDNA SPIN kit for soil (Bio 101, Vista, CA), and dry DNA aliquots (100 ng each) were shipped to the Georgia Institute of Technology. The DNA was suspended in 100 l of sterile, nuclease-free, deionized water to yield a homogeneous solution with a DNA concentration of 1 ng l Ϫ1 . qPCR analysis. The SYBR green approach was used to quantify organisms closely related to Geobacter lovleyi strain SZ as described previously (1). TaqMan reactions targeting the A. dehalogenans strain 2CP-C 16S rRNA gene were performed as described previously (48). To quantify all Anaeromyxobacter 16S rRNA gene sequences in field samples, genus level-targeted TaqMan-based qPCR tools were utilized (48). For all TaqMan reactions, qPCR 1-fold-concentrated master mix (Applied Biosystems [ABI], Foster City, CA), 125 nM probe, and 100 nM (each) primer were combined in sterile, nuclease-free water. Aliquots (18 l) of the reaction mix were dispensed into an ABI MicroAmp Fast Optical 96-well reaction plate held on ice. Template DNA (2 l) was added to each tube, and the tubes were sealed with an ABI optical adhesive cover. The quantification limit was considered to be the lowest standard in a linear standard curve that produced measurable fluorescence. The quantification limit for Anaeromyxobacter 16S rRNA gene-targeted TaqMan qPCR was two to three gene copies per reaction, which equates to about 2 ϫ 10 2 16S rRNA gene copies/liter of groundwater. Detection of G. lovleyi 16S rRNA genes used SYBR green chemistry and required 1 ϫ 10 3 16S rRNA gene copies per reaction, or about 1 ϫ 10 5 G. lovleyi 16S rRNA gene copies/liter of groundwater to yield positive signals.
Calculation of growth rate, growth yield, and f e and f s values. Growth rate values were determined using nonlinear regression analysis by plotting the number of 16S rRNA gene copies per ml of culture over time. Instantaneous specific growth rates () were determined using the exponential growth equation y ϭ y 0 ϫ e t where y is the number of 16S rRNA gene copies per ml at time t and y 0 is the number of 16S rRNA gene copies per ml at t ϭ 0. Triplicate cultures were used to determine growth rates at pO 2 values of 0.02 atm, 0.03 atm, 0.07 atm, and 0.18 atm. Growth yields were determined on the basis of initial and final numbers of 16S rRNA gene copies per ml of culture. 16S rRNA gene copies were used as an approximation of cell numbers. The four sequenced Anaeromyxobacter genomes (GenBank accession numbers CP001131, CP000769, CP000251, and ABKC00000000) harbor duplicate rRNA operons (49), suggesting that members of the Anaeromyxobacter group contain two 16S rRNA gene copies per cell. The apparent fraction of electrons utilized for energy generation (i.e., oxygen reduction) (f e ) and the fraction of electrons released in electron donor oxidation and used for biomass production (f s ) were calculated from growth yields and the electron balance of cell biosynthesis associated with oxygen reduction by the methods of McCarty (33) and Criddle et al. (9).
Analytical methods and calculations. Organic acids were quantified using a Waters model 1525 high-performance liquid chromatography system equipped with an Aminex HPX-87H ion exclusion column and connected to a Waters model 2487 dual-wavelength absorbance detector (15). Oxygen was quantified using an Agilent Technologies model 6890N gas chromatograph equipped with a thermal conductivity detector as described previously (37). Oxygen quantification used a 5-point calibration curve established by injecting an ambient air sample and samples from 160-ml bottles containing 25 ml of reduced medium (initial N 2 /CO 2 headspace [80:20, vol/vol] of 135 ml) amended and equilibrated with 0, 16, 30, and 60 ml of air. Standards and samples were withdrawn using a 1,000-l glass syringe without a valve, and 100-l samples equilibrated to atmospheric pressure were injected. Equivalent amounts of nitrogen were injected VOL. 76, 2010 OXYGEN METABOLISM IN ANAEROMYXOBACTER 177 into the bottles at each sampling event to maintain constant overpressure. The detector response (i.e., peak area) was correlated to the calculated pO 2 in the headspaces of the 160-ml bottles based on the calibration curves. Peak areas from all samples and standards were corrected to account for the analytical method baseline response, which was determined by injecting oxygen-free nitrogen gas. pO 2 is reported in atmosphere (atm) and in total micromoles per bottle. The total mass of oxygen per bottle (n, in micromoles) was calculated from pO 2 (atm) according to the ideal gas law [n ϭ (pO 2 ϫ V ϫ 10 6 )/(R ϫ T), where V equals the culture headspace volume (0.135 liter), R is the ideal gas constant (0.082057 liter atm K Ϫ1 mol Ϫ1 ), T is the incubation temperature in Kelvin (i.e., 303 K), and the multiplier 10 6 converts moles to micromoles]. For example, a pO 2 of 0.02 atm equates to 110 mol of oxygen per bottle. Due to oxygen's very low Henry's law constant (K H ) of 1.26 ϫ 10 Ϫ3 M/atm, dissolved oxygen did not significantly contribute to the total amount of oxygen present in each bottle (e.g., less than 1% of the total amount of oxygen would be dissolved in the aqueous phase of the experimental system at a pO 2 of 0.02 atm). Statistical analysis. Nonlinear and linear regression analyses, including significance tests for the regressions, were performed using GraphPad Prism version 5.00 (GraphPad Software, San Diego, CA). For modeling exponential growth of strain 2CP-C, the instantaneous specific growth rate was constrained to less than 1, and default settings were used for all other analyses.

Response of two DIRB populations to oxygen intrusion at the Oak Ridge IFC site.
Analysis of Oak Ridge IFC area 3 sediment and groundwater samples collected before and after oxygen intrusion (5, 56) suggested a change in the Anaeromyxobacter population size (Fig. 1). The number of Anaeromyxobacter 16S rRNA gene copies per g of sediment increased as much as 5-fold in response to oxygen intrusion. For example, the number of Anaeromyxobacter cells increased from 2.2 ϫ 10 7 Ϯ 8.6 ϫ 10 6 to 1.0 ϫ 10 8 Ϯ 2.2 ϫ 10 7 cells per g of sediment collected from well FW101-2 (i.e., FW101 at screened interval 2 [13.7 m bgs]). The number of Anaeromyxobacter cells in sediment samples ranged from 1 ϫ 10 7 to 1 ϫ 10 8 cells per g (Fig. 1A) and were substantially higher than the cell titers in groundwater, which ranged from 1 ϫ 10 3 to 1 ϫ 10 4 cells per liter (Fig. 1B), suggesting that the majority of Anaeromyxobacter cells were attached to the aquifer solids. The number of unattached Anaeromyxobacter cells either remained relatively constant (well FW101-2) or decreased (FW102-2 and FW102-3) after oxygen intrusion (Fig. 1B). In contrast to observations made with Anaeromyxobacter, the number of sediment-associated cells of G. lovleyi, a U(VI)-reducing deltaproteobacterium present at area 3 at the Oak Ridge IFC site (1), decreased by as much as 50% following the intrusion of oxygenated groundwater (Fig. 1C). Conversely, the percentage of unattached G. lovleyi cells increased in two of the three samples (FW101-2 and FW102-3) following oxygen intrusion (see Table 2) (Fig. 1D).
Oxygen consumption in laboratory cultures. Laboratory experiments were performed under precisely controlled conditions to study the different behaviors of Anaeromyxobacter dehalogenans and Geobacter lovleyi in more detail. Stationaryphase A. dehalogenans strain 2CP-C cultures previously grown with 10 mM fumarate consumed three consecutive oxygen feedings (each at a pO 2 of 0.02 atm) within 50 h per addition ( Fig. 2A). Oxygen reduction coincided with acetate oxidation (Fig. 2A), and neither substrate was consumed in uninoculated controls (data not shown). No measurable net increase in biomass occurred during oxygen reduction in the stationary-phase cultures as indicated by constant 16S rRNA gene copy numbers ( Fig. 2A). The ability to consume oxygen in stationaryphase cultures was not shared by G. lovleyi strain SZ (Fig. 2B). Under the conditions tested, neither oxygen reduction nor acetate consumption was observed in cultures of strain SZ, and cell numbers decreased by an order of magnitude during exposure to oxygen (pO 2 of 0.02 atm) (Fig. 2B). Previous studies have shown that A. dehalogenans and Geobacter lovleyi use acetate as an elec- tron donor, which both organisms oxidize to carbon dioxide (16,17,47). Our analyses supported full oxidation of acetate, as no organic breakdown products were detected in any of the cultures. Microaerophilic growth by A. dehalogenans strain 2CP-C. While strain 2CP-C grew readily under anaerobic conditions, growth in liquid medium at atmospheric oxygen tension (i.e., pO 2 of 0.21 atm and cultures shaken in foam-stoppered Erlenmeyer flasks) was unreliable and occurred only in about one third of the cultures after long lag phases of 3 to 5 weeks (data not shown). To quantify oxygen depletion and to correlate oxygen consumption to biomass production, experiments in closed serum bottles were performed. At an initial pO 2 of 0.02 atm, oxygen consumption began after a 45-hour lag phase and 110 Ϯ 29 mol of oxygen were consumed after 82 h of incubation ( Fig. 3A; see Fig. S2 in the supplemental material). Oxygen was not consumed in uninoculated control bottles (data not shown), indicating that strain 2CP-C cells were responsible for oxygen consumption. A. dehalogenans strain 2CP-C grew during oxygen reduction as indicated by increases in the number of 16S rRNA gene copies from 2.66 ϫ 10 7 Ϯ 8.78 ϫ 10 6 to 5.18 ϫ 10 8 Ϯ 7.20 ϫ 10 7 per ml of culture suspension (Fig. 3A). In culture vessels not amended with oxygen, the number of A. dehalogenans 16S rRNA gene copies decreased from 3.16 ϫ 10 7 Ϯ 9.16 ϫ 10 6 (cells introduced with the inoculum) to 6.48 ϫ 10 6 Ϯ 3.64 ϫ 10 6 per ml of culture suspension over the 86 h of incubation period (Fig. 3B), indicating that cell growth in the oxygen-amended cultures was attributable to oxygen respiration (Fig. 3A). The mean numbers of 16S rRNA gene copies in experimental and control cultures increased from 2.7 ϫ 10 7 Ϯ 8.8 ϫ 10 6 to 6.9 ϫ 10 7 Ϯ 1.2 ϫ 10 7 and from 3.2 ϫ 10 7 Ϯ 9.2 ϫ 10 6 to 5.8 ϫ 10 7 Ϯ 2.5 ϫ 10 7 16S rRNA gene copies per ml culture, respectively, in the first 27 h with no measurable decrease in acetate or oxygen concentration during the first 45 h. The modest increases in cell numbers were likely due to residual growth and were within experimental error (i.e., standard deviations of individual data points) and, thus, within the error reported for biomass calculations. Following 45 h of incubation, oxygen and the electron donor acetate were consumed concomitantly (Fig.  3A). No acetate consumption occurred in control cultures without oxygen (Fig. 3B), and oxygen consumption ceased in bottles that had consumed all acetate (data not shown). Growth of A. dehalogenans was also observed after comparable lag times at pO 2 values of 0.03, 0.07, and 0.18 atm (see Fig. S2 in the supplemental material). When oxygen-grown cells were transferred to fresh medium, oxygen reduction and growth resumed with a lag time of Ͻ45 h, and the ability to grow with oxygen as the sole electron acceptor was maintained over repeated transfers.
Effect of pO 2 on growth. Increasing pO 2 affected growth rates of strain 2CP-C (see Table S1 and Fig. S2 in the supple-   (Fig. 4). The slope of the -pO 2 regression line in Fig. 4 was significantly not zero (P Ͻ 0.05). Doubling times of fumarate-grown and oxygen-grown (pO 2 of 0.02 atm) cultures were not statistically different, with 95% confidence intervals of 10 to 19 h and 12 to 16 h, respectively (see Table S1 in the supplemental material). In addition to growth rate, the O 2 utilization efficiency of A. dehalogenans strain 2CP-C decreased with increasing pO 2 ( Table 1). The growth yield of strain 2CP-C at an initial pO 2 of 0.02 atm was 6 Ϯ 2 g of biomass (dry weight) per mol of electrons consumed, which is comparable to the growth yield with fumarate (5.7 Ϯ 0.6 g of cells per mol of electrons consumed; Table 1). When pO 2 was increased to 0.18 atm, the growth yield decreased by 78% to 1.3 Ϯ 0.5 g per mol of electrons consumed. The measurement of biomass revealed that the fraction of electrons released in acetate oxidation and used for cell growth (f s ) decreased from 0.52 to 0.19 as the pO 2 values increased from 0.02 to 0.18 atm. Accordingly, the fraction of electrons used for oxygen reduction (f e ) increased from 0.48 to 0.81 as the pO 2 increased from 0.02 atm to 0.18 atm (Table 1).

DISCUSSION
The implication of in situ growth of Anaeromyxobacter spp. following oxygen intrusion in area 3 at the Oak Ridge IFC site motivated detailed laboratory studies of the oxygen metabolism of A. dehalogenans strain 2CP-C. Oxygen growth studies revealed specific adaptations of A. dehalogenans to environments with fluctuating redox conditions. The fraction of electrons released during electron donor oxidation and directed toward reduction of the terminal electron acceptor (f e ) indicates if the electron acceptor is used for energy generation (i.e., respiration) or if electron acceptor reduction is uncoupled from growth (e.g., to achieve detoxification or to serve as a fortuitous electron sink). Previous studies have demonstrated that an organism's metabolic efficiency, as measured by f e and f s , is constant for a given set of substrates (9,33). For example, characteristic f e values for aerobic respiration range between 0.4 and 0.88 (depending on the electron donor) (32) and are governed by the energetics of the energy-generating redox reaction (33). The growth yield (f s ) and efficiency of oxygen utilization (f e ) in A. dehalogenans strain 2CP-C at a pO 2 of 0.02 atm are comparable to those observed with fumarate or nitrate as an electron acceptor (Table 1) and are consistent with other acetate-utilizing aerobes (16,17). At a pO 2 of 0.18 atm, less biomass was produced per mole of transferred electrons, and f e and f s values were in line with those measured for soluble ferric iron respiration (Table 1), an energetically less favorable process than aerobic respiration (28). Growth experiments performed at atmospheric oxygen concentrations indicated that oxygen toxicity and detoxification are a likely explanation for the observed shift in the use of reducing power by the cells. The fraction of reducing power generated from acetate oxidation and used for oxygen reduction (i.e., f e ) increased significantly under high pO 2 conditions and reached values of Ͼ0.8.  Table S1 in the supplemental material. b Cell yield is reported per mole of electrons transferred to the electron acceptor based on oxygen reduction to water, 2-chlorophenol (2-CP) reduction to phenol, and fumarate reduction to succinate. Cell yield estimates are calculated by multiplying the number of cells per micromoles of electrons transferred to the electron acceptor times the dry weight of a single cell (3.80 Ϯ 0.27 ϫ 10 Ϫ14 g) estimated previously for fumarate-grown Anaeromyxobacter cells (44). This high f e does not indicate an increase in metabolic energy gain but reflects an apparent f e in the sense that not all the free energy release associated with the redox reaction is coupled to growth. A significant fraction of the oxygen is not respired but is consumed without energy capture, presumably to avoid oxygen toxicity. Hence, the true f e linked to metabolic energy gain from oxygen respiration is observed only with a low pO 2 when electron consumption for oxygen detoxification is minimal. Consequently, the evaluation of f e for an aerobic process should be determined at different pO 2 values to ensure that a true f e , as opposed to an apparent f e , is measured. This procedure is pertinent for microorganisms whose relationships to oxygen are unclear. Genome analysis of A. dehalogenans strain 2CP-C suggested that two oxygen-reducing pathways exist: one that shares homology with anaerobes and microaerophiles and includes a cytochrome cbb 3 oxidase and a second pathway that includes a cytochrome aa 3 oxidase characteristic of the Myxobacteria, which are primarily aerobes (49). Cytochrome oxidases can play roles in both respiratory (energy-yielding) and/or oxygen-detoxifying (non-energy-yielding) reactions (40). The Alphaproteobacteria Bradyrhizobium japonicum and Paracoccus denitrificans both possess cytochrome aa 3 oxidases as well as cytochrome cbb 3 oxidases and are thought to maintain both protein complexes for quick response to changing pO 2 (3,41). It is conceivable that one cytochrome oxidase is used for detoxification while a separate, respiratory cytochrome oxidase is involved in the organism's energy metabolism at lower pO 2 . Instantaneous oxygen reduction by anaerobically grown A. dehalogenans strain 2CP-C cells that had reached stationary phase suggests a constitutive oxygen detoxification pathway, an adaptation beneficial for anaerobic or microaerophilic organisms near fluctuating oxic/anoxic interfaces. Geobacter spp. represent another group of environmentally relevant metalreducing deltaproteobacteria that are found at the oxic/anoxic interface but whose relationship to oxygen remains unclear. G. lovleyi strain SZ failed to consume oxygen under the conditions tested, but growth under aerobic conditions has been reported for Geobacter sulfurreducens; however, growth of G. sulfurreducens with oxygen could not be maintained through repeated transfers without intermediate, anoxic cultivation with Fe(III) as an electron acceptor (24). In contrast, A. dehalogenans grows with oxygen as the sole electron acceptor over repeated transfers. On the basis of the existing genome annotation for G. lovleyi strain SZ (GenBank accession number CP001089), the electron transport chain of strain SZ contains a cytochrome bd-type oxidase, but not the aa 3 type, present in both G. sulfurreducens (24) and A. dehalogenans, or the cbb 3 type present in A. dehalogenans. Therefore, differences in oxygen metabolism between G. lovleyi, G. sulfurreducens, and A. dehalogenans would be expected.
Other members of the Deltaproteobacteria, including sulfatereducing Desulfovibrio spp., flourish at the oxic/anoxic interface in the presence of sulfide/sulfate and are capable of oxygen detoxification, although they were classically considered strict anaerobes (21). Aerobic growth of Desulfovibrio spp. has not been demonstrated, but cultures of Desulfovibrio desulfuricans ATCC 27774 consumed oxygen, and the expression of catalase and superoxide dismutase increased when grown with nitrate under a high pO 2 (up to 0.18 atm) (26). Several transcriptomic and proteomic studies with Desulfovibrio spp. have been performed to identify the metabolic changes in response to oxygen (12,38,45,59), including a study indicating that exposure to air (pO 2 of 0.21 atm) caused downregulation of central metabolic pathways and that the organisms dedicated less reducing power for biomass production (34). Growth measurements of A. dehalogenans strain 2CP-C indicated that the respiratory efficiency decreased at higher pO 2 , resulting in less biomass synthesis ( Table 1). Downregulation of central metabolic pathways in response to high pO 2 , as seen in D. vulgaris, could explain the decreasing biomass synthesis observed in A. dehalogenans strain 2CP-C cultures.
A major physiological difference between A. dehalogenans and other metal-reducing Deltaproteobacteria is its demonstrated use of type IV pilus-based surface motility (49). Consistent with the classification of Anaeromyxobacter as a myxobacterium with surface motility, the quantitative analysis of area 3 sediment and groundwater samples demonstrated that Anaeromyxobacter cells occur predominantly associated with the soil surfaces. In area 3, Anaeromyxobacter cells were 1,000to 10,000-fold more prevalent on solids than Geobacter lovleyi cells ( Fig. 1 and Table 2). In contrast, metal-reducing Geobacter spp., which use flagellar motility (6,47), were more evenly distributed between solids and the aqueous phase (Table 2), similar to what was observed in another U-contaminated aquifer (Rifle site, CO) (18). The ratio of attached to unattached G. lovleyi cells decreased in two out of three samples following oxygen intrusion (Table 2). A possible explanation is detachment of G. lovelyi cells from the solids in response to oxygen intrusion due to toxicity or a shift of lifestyle from a solid-associated lifestyle to an unattached lifestyle. Thus, the apparent increase in the number of G. lovleyi cells in ground- water may reflect a redistribution of cells from the attached state to the unattached state rather than aerobic growth of unattached cells. The uneven distribution of DIRB between solids and groundwater emphasizes the need for investigation of how organism-specific characteristics and environmental factors influence attached versus unattached growth. Obviously, the analysis of Anaeromyxobacter population dynamics based on groundwater sampling data alone will provide an incomplete picture of the true abundance of members of this group. Bioreduction and immobilization of U are a promising treatment strategy, but reoxidation of immobilized U(IV) affects stability and mobility of the mineralized material and complicates long-term site management. In anaerobic subsurface environments, different DIRB coexist (5,10,39,46,53), and area 3 at the Oak Ridge IFC site harbors at least Geobacter spp., Anaeromyxobacter spp., and Desulfovibrio spp. Ecological exclusion theory predicts that organisms occupying the same ecological niche do not cooccur in the same environment (19). Therefore, physiological features must distinguish the DIRB present in area 3, and a relevant distinguishing factor is their response to oxygen. A. dehalogenans responds quickly to oxygen intrusion via a constitutive detoxification system, grows with oxygen under microaerophilic conditions, predominantly associates with solids, and uses oxidized metals and radionuclides, including U(VI), as respiratory electron acceptors. These combined characteristics are unique and not shared with other known bacterial populations contributing to U(VI) reduction in area 3 at the Oak Ridge IFC site. Continued monitoring of U concentrations in the area 3 bioreduction plot suggested that U initially oxidized by oxygen was quickly immobilized (Wei-min Wu, personal communication), possibly related to DIRB activity. Future efforts should therefore explore not only the abilities of organisms to reduce U(VI) under anaerobic conditions but also focus on the ecophysiologies of the organisms, as those features may affect the long-term stability of reduced radionuclides.