Molybdenum Availability Is Key to Nitrate Removal in Contaminated Groundwater Environments

Abstract

The concentrations of molybdenum (Mo) and 25 other metals were measured in groundwater samples from 80 wells on the Oak Ridge Reservation (ORR) (Oak Ridge, TN), many of which are contaminated with nitrate, as well as uranium and various other metals. The concentrations of nitrate and uranium were in the ranges of 0.1 μM to 230 mM and <0.2 nM to 580 μM, respectively. Almost all metals examined had significantly greater median concentrations in a subset of wells that were highly contaminated with uranium (≥126 nM). They included cadmium, manganese, and cobalt, which were 1,300- to 2,700-fold higher. A notable exception, however, was Mo, which had a lower median concentration in the uranium-contaminated wells. This is significant, because Mo is essential in the dissimilatory nitrate reduction branch of the global nitrogen cycle. It is required at the catalytic site of nitrate reductase, the enzyme that reduces nitrate to nitrite. Moreover, more than 85% of the groundwater samples contained less than 10 nM Mo, whereas concentrations of 10 to 100 nM Mo were required for efficient growth by nitrate reduction for two Pseudomonas strains isolated from ORR wells and by a model denitrifier, Pseudomonas stutzeri RCH2. Higher concentrations of Mo tended to inhibit the growth of these strains due to the accumulation of toxic concentrations of nitrite, and this effect was exacerbated at high nitrate concentrations. The relevance of these results to a Mo-based nitrate removal strategy and the potential community-driving role that Mo plays in contaminated environments are discussed.

INTRODUCTION

Environmental contamination from industrial processes, including the release of heavy metals, organic solvents, and nitrogen-containing compounds, can have profound negative impacts on wildlife and human health (1, 2). The Department of Energy oversees the monitoring and restoration of over 90 waste sites at 18 facilities (2). They include facilities at Hanford, WA, and Oak Ridge, TN, where nuclear production and processing have resulted in the release into the environment of complex mixtures of hazardous waste. For example, between 1951 and 1983, four unlined enclosures known as the S-3 ponds were used for waste disposal at the Y-12 nuclear-processing plant on the Oak Ridge Reservation (ORR). The waste material was acidic (pH < 2.0), with nitrate; metals, such as uranium (U), aluminum, manganese (Mn), and chromium; and organic solvents (including acetone, dichloromethane, toluene, and tetrachloroethene) as major contaminants. In 1983, the waste in the ponds was treated by neutralization and biodenitrification, and the ponds were eventually capped and covered by a parking lot (3, 4). Several contaminant plumes have entered the groundwater surrounding the Y-12 S-3 ponds, and many wells have been drilled in the contaminated areas, as well as in a pristine location 6 km distant from the S-3 ponds (5). The contaminated and pristine wells have been used to conduct geochemical, bioremediation, and ecological studies (4, 6–8). Recently a comprehensive survey of the groundwater was conducted that included 16S rRNA gene sequencing. The results showed that bacterial communities could be linked to geochemical parameters, showing that bacteria could be used as biosensors to detect contamination (9).

In addition to their presence as contaminants in nitrate-rich environments, some metals play important roles in biological removal of nitrate by the process of denitrification, a major branch of the global nitrogen cycle, in which nitrate is converted to nitrogen gas. The first of four steps in denitrification, the reduction of nitrate to nitrite, is catalyzed by nitrate reductase, which contains molybdenum (Mo) at its active site (Fig. 1) (10). The second enzyme is nitrite reductase, which catalyzes the reduction of nitrite to nitric oxide. These are either iron (Fe) (NirS)- or copper (Cu) (NirK)-containing enzymes. Nitric oxide is reduced to nitrous oxide by Fe-containing nitric oxide reductase, while the terminal step, reduction of nitrous oxide to nitrogen gas, is catalyzed by Cu-containing nitrous oxide reductase (10). While denitrifiers are ubiquitous in nature, many microorganisms do not contain the complete pathway. Some nitrate-reducing organisms carry out only the first step, the dissimilatory reduction of nitrate to nitrite, while others reduce nitrate to ammonium by a process called nitrate ammonification (11). On the other hand, some microorganisms cannot reduce nitrate but contain later steps of the denitrification pathway (12).

FIG 1.

Model of the effects of NO₃⁻, Mo, and W on the growth of Pseudomonas isolates under denitrifying conditions. The first enzyme in the denitrification pathway, nitrate reductase, requires a molybdopterin guanine dinucleotide cofactor for activity. The activity of nitrate reductase can be increased during growth of a Pseudomonas isolate by supplementing the growth medium with Mo or NO₃⁻. This effect is beneficial to growth up to a point; however, oversupplementation with Mo and NO₃⁻ can result in an inhibitory buildup of nitrite, decreasing growth. Supplementation of the growth medium with W results in antagonistic interactions with Mo that decrease nitrate reductase activity.

Since Mo is a key element in the biological conversion of nitrate to nitrite, the primary goals of this project were to measure the concentrations of Mo in the ORR-contaminated groundwater and to determine if Mo could influence the role that biological denitrification plays in removing nitrate from such environments. There was also an interest in the concentration of tungsten (W) in the groundwater. Tungstate and molybdate have very similar physical properties, resulting in an inability of many biological systems to fully discriminate between the two (13), and tungstate has been shown to inhibit dissimilatory nitrate reduction by several different microorganisms (14–18), as well as by mixed cultures obtained from the rumen (19). Tungsten could therefore impede environmental nitrate utilization if present in significant concentrations relative to Mo. Here, we show that the contaminated ORR groundwater contains extremely high concentrations of a range of metals. Unexpectedly, however, both Mo and W are present at extremely low concentrations. Moreover, laboratory studies of the Mo requirements of nitrate-reducing microorganisms, including two isolated from the ORR groundwater, demonstrated that Mo availability is likely a major determinant in the efficient removal of environmental nitrate. Consequently, Mo supplementation could stimulate biological nitrate reduction in groundwater. We also show that addition of Mo to groundwater could lead to a toxic buildup of nitrite, so Mo supplementation experiments require careful optimization. This project represents the first step in understanding the potentially community-shifting effects that Mo might have on the groundwater community at ORR and will inform bioreactor experiments that may lead to an improved method of nitrate bioremediation involving Mo supplementation.

MATERIALS AND METHODS

Strain isolation and sequencing.

Strain N2A2 was isolated from groundwater collected from well FW-300 on basal medium (using 10 mM sodium acetate as the sole carbon source and electron donor and 10 mM sodium nitrate as the electron acceptor). Strain N2E2 was isolated from groundwater collected from the same well using the same basal medium but 10 mM sodium propionate as the sole carbon source and electron donor and 10 mM sodium nitrate as the electron acceptor.

DNA was extracted from the isolates using the MoBio UltraClean Microbial DNA isolation kit (MoBio Inc., Carlsbad, CA). We PCR amplified the 16S rRNA gene using primers 27F and 1492R and performed Sanger sequencing directly on the PCR products. We prepared Illumina sequencing libraries for N2A2 and N2E2 using two methods: the Nextera XT kit, following the manufacturer's (Illumina) instructions, and 800-bp inserts labeled with Illumina TruSeq adapters. For the Nextera XT libraries, we performed paired-end sequencing (2 by 250 bp) on the Illumina MiSeq. For the 800-bp insert libraries, we performed paired-end sequencing (2 by 150 bp) on the Illumina MiSeq. We merged the Nextera and 800-bp-insert paired-end reads using PEAR (https://pear.php.net/index.php) and assembled the genomes with the SPAdes genome assembler (http://bioinf.spbau.ru/spades).

Growth conditions.

The basal growth medium had the following composition: 20 mM sodium fumarate, 4.7 mM NH₄Cl, 1.3 mM KCl, 2 mM MgSO₄, 0.2 mM NaCl, 1.2 mM NaHCO₃, and 5 mM NaH₂PO₄, with sterile vitamins and trace elements prepared as described by Widdel and Bak but lacking Mo and W (20). The NaHCO₃ (>99.7% pure) used was from Sigma-Aldrich, St. Louis, MO (other sources contained elevated contaminating levels of Mo that interfered with the growth experiments). Either NaNO₃ or NaNO₂ was used as an electron acceptor at the indicated concentrations. Initial cultures were grown aerobically in a medium containing fumarate (20 mM) and nitrate (20 mM) but lacking added Mo and W. They were then diluted 20-fold in the experimental growth medium. For each experiment, the indicated amounts of nitrate, Mo, and/or W were added to the basal growth medium. Cultures were grown in a 100-well Bioscreen plate, with each well containing 400 μl of diluted preculture. The Bioscreen plate was incubated anaerobically at 30°C with continuous shaking in a Bioscreen C (Thermo Labsystems, Milford, MA) within an anaerobic chamber (Plas Laboratories, Lansing, MI) in an atmospheric composition of 95% Ar and 5% H₂. Growth was monitored at an optical density (OD) of 600 nm. All experiments were performed in biological duplicate or triplicate. The error for the growth experiments was calculated as the standard deviation.

To condense large amounts of growth data, the maximum growth OD or percent maximum growth OD is reported. The maximum growth OD was defined as the OD at the time point of a curve that corresponds to the same time point at which maximum growth was first obtained by the fastest-growing culture. The percent maximum growth OD is the maximum growth OD that was obtained normalized to the fastest-growing growth condition (taken as 100%).

Nitrite detection and nitrite reductase assay.

Nitrite was measured in the supernatant after cultures had been centrifuged aerobically for 20 min at 5,000 rpm in an Allegra 25R Centrifuge (Beckman Coulter, Fullerton, CA). Samples were mixed with an equal volume of Griess reagent (Sigma-Aldrich, St. Louis, MO), and after incubating aerobically for 15 min at 22°C, the A₅₄₀s of the samples were measured in a Spectra Max 190 (MTX Lab Systems, Inc., Vienna, VA). For whole-cell nitrite reductase assays, the strains were grown anaerobically for 16 h in 50 ml of basal medium containing fumarate (20 mM), nitrate (20 mM), and 10 nM Mo. Cells were harvested aerobically and washed with 10 ml of wash buffer (50 mM KPO₄ buffer, pH 7.0) and were then resuspended to a final volume of 1 ml with wash buffer. An aliquot of cell suspension that contained 1 to 10 mg protein, as determined by the Bradford method (21), was added to assay buffer (50 mM potassium phosphate containing 40 mM fumarate and 1 mM nitrite, pH 7.0) to give a final volume of 5 ml. The mixture was incubated at 30°C, and samples were removed every 5 min for 30 min, diluted 1:10 with double-distilled water (ddH₂O), and mixed 1:1 with Griess reagent (Sigma-Aldrich, St. Louis, MO). After incubating for 15 min at 22°C, the A₅₄₀s of the samples were measured in a Spectra Max 190 (MTX Lab Systems, Inc., Vienna, VA). One unit of nitrite reductase activity is defined as the amount catalyzing the reduction of 1 nmol of nitrite/min.

ICP-MS analysis.

Groundwater samples were stored in sealed serum bottles at 4°C prior to analysis. A Corning MP-3A distillation apparatus was used to produce pure glass double-distilled water (gddH₂O), which was used in all dilution and washing steps. All containers used for inductively coupled plasma mass spectrometry (ICP-MS) analysis were acid washed by submersion in a 2% (vol/vol) solution of concentrated nitric acid in gddH₂O for 24 h and rinsed twice by submersion in gddH₂O for 24 h. Trace metal grade concentrated (70%) nitric acid (Fisher; A509-212) was used to acidify samples. To measure the soluble elements present, groundwater samples (6 ml) in acid-washed 17- by 20-mm Sarstedt polypropylene screw-cap conical tubes (62.554.002-PP) were centrifuged at 7,000 × g for 15 min at 4°C in a Beckman-Coulter Allegra 25R centrifuge. The supernatant was decanted into acid-washed polypropylene tubes and acidified with 120 μl of concentrated nitric acid. All samples were briefly vortexed (30 s) and incubated with shaking at 37°C for 1 h in a New Brunswick Scientific G24 Environmental Incubator Shaker. All samples were centrifuged at 2,000 × g for 10 min in a Beckman Allegra 6R centrifuge at 25°C. Metal analyses were performed in triplicate using an Agilent 7500ce octopole ICP-MS in FullQuant mode using an internal standard with in-line addition and a multielement external standard curve as previously described (22). Samples were loaded via a Cetac ASX-520 autosampler. Control of sample introduction, data acquisition, and processing was performed using Agilent MassHunter version B.01.01.

Accession numbers.

The draft genome sequences for strains N2A2 and N2E2 are available in the Integrated Microbial Genomes system (accession numbers 2600255294 and 2600255295, respectively).

RESULTS

Concentrations of metals and nitrate in ORR groundwater.

In a previous study, 93 groundwater samples taken from 80 wells at various distances from the S-3 ponds on the ORR were analyzed for a number of geochemical parameters, including pH, nitrate, and dissolved O₂ and several metals (9). Here, the same groundwater samples were further processed by centrifugation, and the soluble fractions were then analyzed by ICP-MS to determine the concentrations of 26 metals, including Mo, W, Fe, and Cu, using a previously described method (23). The metal concentrations are reported in Table S1 in the supplemental material, with standard deviations reported in Table S2 and detection limits in Table S3 in the supplemental material.

Our analysis showed high concentrations of a range of metals in the ORR groundwater, in addition to the expected U. This might have been anticipated, due to the complexity and diversity of the waste deposited in the S-3 ponds. For example, Fig. 2A shows the concentrations of metals in a groundwater sample from one of the most highly contaminated wells (FW-126), together with the data from a pristine well (FW-300) located 6 km from the contaminated S-3 ponds. Both groundwater samples were acquired on the same date. As might be expected, the concentrations of most metals (21 out of 26) were higher in the FW-126 well, with Mn, U, and nickel (Ni) above 0.1 mM. Interestingly, the concentration of Mo was extremely low, below 10 nM, in both wells.

FIG 2.

Metals in pristine and contaminated wells at the ORR. (A) Concentrations of metals and nitrate (inset) measured in the soluble fractions of groundwater samples from pristine well FW-300 (black) and contaminated well FW-126 (gray). Measurements at or below detection limits are indicated with asterisks. The error bars indicate standard deviations. (B) The groundwater samples were divided into two subsets based on the soluble U concentration: those below (black) and those above (gray) the maximum amount (126 nM) allowed in drinking water by the EPA. For each subset of groundwater samples, the low, median, and high concentrations are represented. Metals whose median concentrations are higher in U-contaminated than in noncontaminated wells are represented by circles, and those whose median concentrations are lower are represented by triangles.

In order to look at the effects of contamination on the concentrations of the 26 metals across all 93 groundwater samples, the samples were divided into two categories: highly contaminated, which had levels of U above the U.S. Environmental Protection Agency (EPA) maximum contaminant level (MCL) for drinking water (126 nM; 27 samples), and samples with smaller amounts of U (66 samples). The minimum, maximum, and median concentrations of each metal in the two categories of groundwater samples were calculated and are reported in Fig. 2B. The median concentration of U in the highly contaminated samples was 1,600-fold higher than in the other groundwater samples. For all other metals, there is significant overlap in their ranges of concentrations between the two categories. However, the median concentrations of many of the metals vary greatly between the two subsets of groundwater samples. For example, Mn was present at the highest concentration (maximum, 3.2 mM) and the median concentration of Mn was 1,700-fold higher in the highly contaminated wells. Other metals that had significantly greater median concentrations in the highly contaminated groundwater samples were cadmium (Cd) (2,700-fold), cobalt (Co) (1,300-fold), Ni (230-fold), thallium (Tl) (60-fold), rhenium (Re) (38-fold), Cu (25-fold), and Fe (20-fold).

Although most of the metals were present in higher median concentrations in samples from wells highly contaminated with U (≥126 nM), there were three metals that had lower median concentrations, including Mo, which was 7-fold lower. The other metals were vanadium (V) (2-fold lower) and gallium (Ga) (1.4-fold lower). Indeed, the concentrations of Mo were very low in all the wells. They ranged from <10 pM to 330 nM, with 85% of the groundwater samples containing Mo at concentrations below 10 nM. The concentrations of W, an antagonist of Mo utilization in biological systems, were even lower in the groundwater samples, ranging from 10 pM to 20 nM, with 98% of the samples below 10 nM.

The Mo and nitrate concentrations within each groundwater sample are compared in Fig. 3. As reported previously (9), the concentrations of nitrate in the wells varied over an extremely large range, from 100 nM to 230 mM. Of note is a cluster of groundwater samples in the lower-right quadrant of Fig. 3 that have extremely low concentrations of Mo (<1 nM) and very high concentrations of nitrate (>10 mM). The extent to which subnanomolar levels of Mo might limit nitrate removal in the contaminated groundwater is discussed below, along with the effect nitrate has on Mo solubility. The uranium concentrations are also depicted in Fig. 3 and show that, in general, the wells containing U above the drinking water standard (126 nM) also contained high nitrate levels (>10 mM), although not all wells contaminated with nitrate were contaminated with U.

FIG 3.

Nitrate and Mo concentrations in groundwater wells at the ORR. Each point represents a groundwater sample taken from a well at the ORR. A total of 93 groundwater samples were taken from 80 wells. Duplicate samples from the same well were taken either on different dates or at different well depths. The concentrations of nitrate and Mo are plotted on logarithmic scales. Wells containing U above (blue) or below (green) a concentration of 126 nM are indicated. The red dashed line crosses the Mo concentration axis at 10 nM, and the purple dashed line crosses at the limit of detection for Mo. The nitrate data were collected as part of a previous study (9).

Isolation of bacteria.

The finding of low to subnanomolar concentrations of Mo in the vast majority of the ORR groundwater samples begged the question of whether such concentrations would limit microbial denitrification. To address this issue, two different nitrate-reducing microbial strains were isolated from groundwater samples taken from a pristine well (FW-300) at the ORR. The well contained 60 μM nitrate and 9 nM Mo (and 1 nM W). The strains were designated N2A2 and N2E2, and by 16S rRNA gene analysis, their closest relatives were Pseudomonas migulae strain T19 (EU111700.2; 99.72% similarity) and Pseudomonas corrugata strain P94 (EF153018.1; 99.13% similarity), respectively. Also included in the following analysis of Mo requirements was Pseudomonas stutzeri strain RCH2, a model denitrifying microorganism that was isolated from chromate-contaminated groundwater at the nuclear site at Hanford, WA (24). The denitrification pathway in P. stutzeri has been well studied (10). All three strains exhibit nitrate-dependent growth using fumarate as the carbon and electron source (see Fig. S1 in the supplemental material).

Effect of Mo and W concentrations on nitrate-dependent growth.

The three Pseudomonas strains were grown on fumarate in the presence of nitrate (60 mM) in a medium containing added Mo (100 pM to 10 μM). In addition, W was also added (0 to 1 μM W). The results are shown in Fig. S2 in the supplemental material. In the absence of added Mo and W, the medium contained 3 nM Mo and 2 nM W as contaminants from other medium components. For each strain, maximal growth required at least 10 nM Mo (in the absence of added W). This is much higher than the concentrations of Mo measured in most of the groundwater samples (Fig. 3). When 10 nM W was added to the growth medium, an amount greater than that seen in 98% of the groundwater samples, growth of all strains was largely unaffected compared to when W was not added. Concentrations of W of at least 100 nM were required to inhibit growth, but the inhibition was relieved by adding comparable concentrations of Mo (see Fig. S2 in the supplemental material). Hence, for efficient nitrate reduction, the three strains require much higher concentrations of Mo than are found in the vast majority of the ORR wells, but nitrate removal would not be affected by W at the concentrations found in those wells.

Effect of high concentrations of Mo on denitrification.

Given that the concentration of Mo in the ORR groundwater potentially limits nitrate reduction, the process could potentially be stimulated by Mo supplementation. To determine the potential effects of high Mo concentrations, the three Pseudomonas strains were grown on fumarate media containing up to 2 mM Mo with 60 mM or 120 mM nitrate. Note that, as shown in Fig. 3, many wells contain at least 60 mM nitrate. The results of the growth study are summarized in Fig. 4. In order to condense the large amount of data, composite growth graphs are shown based on the percentage of the maximum growth observed. Surprisingly, the growth of the two ORR isolates was inhibited at Mo concentrations above 100 nM (strain N2A2) and 1 μM (strain N2E2), with 2 mM Mo resulting in less than 10% of maximum growth (in the presence of 120 mM nitrate). Strain RCH2 was less sensitive to inhibition by high Mo concentrations, and inhibition was observed only in the presence of the higher nitrate concentration (120 mM). While strains RCH2 and N2E2 displayed increased Mo inhibition at the higher nitrate concentration, strain N2A2 was severely inhibited by Mo at both of the nitrate concentrations that were used.

FIG 4.

Effect of high concentrations of Mo on the growth of Pseudomonas strains in the presence of high concentrations of nitrate. The three strains (RCH2, N2A2, and N2E2) were grown anaerobically on basal medium with 60 mM nitrate (black circles) or 120 mM nitrate (gray squares) at various Mo concentrations (expressed on a logarithmic scale). Growth is represented as a percentage of the maximum growth measured for the strain. The error bars indicate standard deviations.

To confirm that the inhibitory effects of Mo were related to nitrate reduction, the effect of adding W, the Mo antagonist, was investigated. When cells were grown in the presence of added W (1 μM) under conditions where Mo caused inhibition of growth (1 mM Mo, 120 mM nitrate), the inhibition was relieved (see Fig. S3 in the supplemental material). Since W is a Mo antagonist and results in inactive nitrate reductase (14), it is proposed that the inhibition by high concentrations of Mo is due to a buildup of toxic nitrite (Fig. 1). High concentrations of Mo in the growth medium result in increased amounts of active nitrate reductase, which in turn can lead to increased nitrite production. If these nitrite concentrations were beyond the capacity of the rest of the denitrification pathway to remove, then nitrite would accumulate and inhibit growth. Hence, we propose that W has a positive effect on growth by inhibiting nitrate reductase, thereby lowering the amount of nitrite produced.

To investigate the proposed model of W stimulation of growth and its relevance to environmental nitrate removal, the concentration of nitrite was measured in the medium after all three Pseudomonas strains had been grown with increasing concentrations of added nitrate. As shown in Fig. S4 in the supplemental material, all three strains accumulated nitrite, and the two ORR strains, which were more readily inhibited by high Mo concentrations, accumulated significantly higher concentrations of nitrite than did RCH2. Accordingly, as shown in Fig. S5 in the supplemental material, the two ORR strains were significantly more sensitive to inhibition by added nitrite than was strain RCH2. Whole-cell nitrite reductase assays were performed for the three strains grown anaerobically on fumarate nitrate (20 mM) medium in the presence of 10 nM Mo. Strains RCH2 and N2A2 had higher nitrite reductase activities (15.0 ± 1.0 and 29 ± 2 U/mg, respectively) than did strain N2E2, which had little to no activity (1.0 ± 0.2 U/mg). These results are consistent with genome-sequencing data showing that strains RCH2 (NC_019936.1) and N2A2 have the complete denitrification pathway while strain N2E2 lacks dissimilatory nitrite reductase (although it contains both nitric oxide reductase and nitrous oxide reductase [Fig. 1]).

The observed accumulation of nitrite by all three nitrate-reducing strains, along with the nitrite sensitivity data, therefore, supports the proposed nitrite toxicity model for W-stimulated growth at the higher Mo concentrations. Ironically, while W, in its role as a Mo antagonist, inhibits growth when the concentration of Mo is low (see Fig. S2 in the supplemental material), it has a beneficial effect on the growth of all three strains at high Mo concentrations by inhibiting nitrate reductase, thereby reducing the amount of toxic nitrite that is produced.

DISCUSSION

The results presented here show that it is likely that the availability of Mo, an essential metal for nitrate reductase, limits nitrate reduction in the environment, and particularly in contaminated ecosystems. In general, the concentrations of Mo present in the ORR groundwater samples are much lower than those required for optimum nitrate-dependent growth of Pseudomonas strains in the laboratory, including two strains isolated from the ORR groundwater. Of the ORR groundwater samples investigated, 85% contained less than 10 nM Mo, and Mo was below the limit of detection (<10 pM) in many wells. Moreover, Mo was one of the few metals that had a lower median concentration in the subset of highly contaminated wells. Such wells also tended to contain higher concentrations of nitrate and to have a lower pH. The sorption of molybdate to soil in aqueous suspensions is dependent on the pH, with increased sorption in the lower pH range, which could further limit Mo solubility and availability (25). However, Mo seems to be limited in the ORR environment in general, since the median concentration of Mo measured in the sediment portion of the highly contaminated ORR groundwater samples was still extremely low (equivalent to 0.4 nM).

Both Fe and Cu are also involved in environmental nitrate removal, as, like Mo, they are essential metal cofactors of enzymes of the denitrification pathway. The median concentrations of Fe and Cu in the noncontaminated wells were 30 nM and 4 nM, respectively, and their median concentrations were increased 20- and 25-fold in the highly contaminated wells. Both Fe and Cu are used in numerous other processes in Pseudomonas strains, and it is possible that in some wells containing very low concentrations of these metals (≪10 nM), growth may be limited by the metals. Like Mo, W is also present at very low concentrations (typically ≪10 nM) in the ORR environment and likely does not impact nitrate reduction to any extent. In the laboratory, 10 nM W, which is higher than that measured in 98% of the ORR groundwater samples, did not have an inhibitory effect on nitrate-dependent growth of the Pseudomonas strains (see Fig. S2 in the supplemental material). On the other hand, it is also shown here that W can stimulate the growth of some nitrate-reducing strains under conditions of excess Mo (≥1 μM) by relieving the accumulation of toxic nitrite (see Fig. S3 in the supplemental material). It is noteworthy that the EPA MCL for nitrite in drinking water is 22 μM, compared to 161 μM for nitrate.

During microbial nitrate removal in natural environments, the product of the nitrate reductase enzyme is nitrite, which in an organism containing the complete denitrifying pathway is further reduced to nitric oxide, nitrous oxide, and finally nitrogen gas (Fig. 1). Of these metabolites, both nitrite and nitric oxide are known to be toxic to biological systems, with nitric oxide being highly toxic (10). However, here, we show that under Mo-sufficient conditions (10 nM Mo), Pseudomonas strains are inhibited by nitrite rather than by nitric oxide. In particular, strain N2E2, which is readily inhibited by increasing Mo concentrations, lacks dissimilatory nitrite reductase activity and cannot produce nitric oxide. The presence of nitrite in the growth media of the Pseudomonas strains (see Fig. S4 in the supplemental material) supports its role as the growth inhibitor. A factor that could be contributing to the accumulation of nitrite is that nitrate may be competing with nitrite for the active site of nitrite reductase. During denitrification by the bacterium Azospirillum brasilense SP7, nitrite reduction occurred only when nitrate concentrations were below 50 μM. In fact, the addition of nitrate caused nitrite to accumulate, and it was proposed that nitrate inhibited nitrite reductase (14, 26, 27). Preferentially using nitrate over nitrite may be competitively advantageous under certain circumstances, since the action of nitrate reductase conserves energy but that of nitrite reductase does not (11). Under conditions where nitrate is limiting, it may be beneficial for microorganisms to first use only nitrate as an acceptor in the environment, as the ORR isolates appear to do, whereas RCH2, which uses nitrite concurrently with nitrate, may have an advantage in contaminated environments by having the complete denitrification pathway, thereby avoiding the accumulation of nitrite.

The high nitrate and high Mo concentrations used in this study represent a very different environment than the pristine groundwater from which the ORR Pseudomonas strains were isolated and to which they were presumably adapted. Indirect control of gene expression can commonly result in fitness defects for organisms exposed to nonnative conditions, such as those used in the laboratory (28). Another difference between the experimental conditions and the ORR groundwater environment is carbon availability. The strains were grown in high concentrations (20 mM) of a readily usable carbon source (fumarate), whereas the groundwater environment is limiting in carbon (9), which in turn limits nitrate reduction. The experiments described here are probably more relevant to nitrate-contaminated environments, where carbon stimulation experiments could potentially have different results depending on whether Mo is supplemented. In terms of Mo supplementation of groundwater environments, it is noteworthy that chronic-toxicity tests conducted on multiple diverse freshwater organisms found that molybdate had a low degree of toxicity compared to other metals, with the most sensitive organism tested being the freshwater fish Pimephales promelas, with a chronic-effect level of 162 μM (29).

The virtual absence of Mo in nitrate-contaminated groundwater effectively limits the use of nitrate as an electron acceptor for microbial action. This could hinder nitrate removal and result in the preferential survival of fermentative organisms and organisms that can grow using alternate electron acceptors. Previously, chromate (CrO₄²⁻)-contaminated groundwater was used as an inoculum for a continuous-flow reactor supplemented with fumarate (30 mM) and only a low concentration of sulfate (0.05 mM) as the electron acceptor (30). Diversity in the reactor decreased from 63 to 11 bacterial genera, and the reactor was dominated by Pelosinus species (30), which are known to ferment various compounds, including fumarate (31). They are also able to use nitrate, Fe(III), CrO₄²⁻, and UO₂²⁺ as electron acceptors (30, 31). The diversity of mechanisms that they use to dispose of reductant may be one of the reasons they are found in great abundance in many supplemented-groundwater studies, including lactate-amended enrichments, some of which were conducted with ORR groundwater (30, 32, 33). The results presented here suggest that Mo supplementation of ORR nitrate-contaminated groundwater enrichments may favor a different set of organisms that efficiently use nitrate as a terminal electron acceptor.

Nitrate contamination of groundwater is an issue in many other locations besides the ORR and is a result of geologic origin, poor waste disposal, and, most predominantly, the use of nitrate-containing fertilizers (34). The concentrations of Mo and W may be an important factor to consider when looking at other such nitrate-contaminated sites. Removal of toxic chemicals, such as nitrate, UO₂²⁺, and CrO₄²⁻, from environments by stimulation of the microbial community is a complex process that depends on the types and concentrations of electron donors and acceptors that are used. A factor not yet considered in these types of in situ studies is the pivotal role that Mo might play due to its essential role in the dissimilatory nitrate reductase. In a previous study at the ORR, the contaminated groundwater was stimulated with ethanol over a 2-year period (6). The process was dependent on a preemptive ex situ pH adjustment (from pH 3.6 to 6.0) and nitrate removal with a denitrification bioreactor (from >100 mM to 0.5 mM). This resulted in the reductive immobilization of U(VI) to below EPA drinking water standards (126 nM). Community analysis after the experiment revealed an increase in the diversity of microorganisms present, including denitrifying, iron-reducing, and sulfate-reducing species of Desulfovibrio, Geobacter, Anaeromyxobacter, Desulfosporosinus, and Acidovorax (6).

One wonders how the results of this and other in situ studies (6, 35, 36) might have been affected by the presence of sufficient concentrations of Mo. The Mo concentrations in wells in the same area of the ORR (the FW-100 area) in which the ethanol stimulation experiment was performed were in most cases much lower (pM range) than the Mo concentration required for laboratory denitrifying growth of the Pseudomonas isolates. The laboratory experiments reported here, in which the carbon source was more than sufficient (20 mM fumarate) but Mo was limiting, are closer to conditions in the ethanol-supplemented groundwater than in unstimulated groundwater (6). Clearly, a future direction is to use groundwater-fed bioreactors supplemented with carbon and various concentrations of Mo to investigate Mo-induced community changes and the effect they have on the removal of nitrate. From our results using three different Pseudomonas species, Mo in the concentration range of 10 to 100 nM is needed for efficient nitrate-dependent growth, although as we demonstrate, a further increase in the Mo concentration could have inhibitory effects due to nitrite accumulation. We predict that in such bioreactor experiments, high concentrations of Mo would result in community changes favoring nitrite-utilizing or nitrite-resistant microorganisms, like strain RCH2. The properties of these organisms, including nitrate, nitrite, and metal tolerance, may make them important players in a more comprehensive study of in situ nitrate removal.