INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Sulfate-reducing bacteria (SRB) are important in the mobility of sulfur in the environment (32, 36). Both higher plants and animals depend upon microbially produced reduced sulfur for acquisition in their own metabolism. SRB use sulfate as an oxidizing agent for the dissimilation of an organic substance (herein, lactate). Several species of SRB have been described for their ability to reduce inorganic aqueous ions in solution. SRB have been shown to metabolize iron [Fe(III)], chromium [Cr(VI)], uranium [U(VI)], manganese [Mn(IV)], and technetium [Tc(VII)], among others (20, 21, 22, 23, 26, 39). Models for the SRB reduction of these metals and radionuclides can be used to develop and design treatment systems employing SRB for bioremediation. For example, bioreduction of soluble hexavalent uranium [U(VI)] to insoluble tetravalent uranium [U(IV)] can be utilized to remove soluble uranium from groundwaters, mine waters, or secondary waste streams.

In addition to SRB, a number of bacterial species have been described as capable of reducing U(VI) to U(IV), including Geobacter metallireducens (25) (previously reported as GS-15 [11]), Shewanella putrefaciens (25), and Shewanella alga strain BrY (previously reported as BrY, or Shewanella halotolerans strain BrY [3]). A kinetic study for the reduction of U(VI) has been described for the iron-reducing S. alga strain BrY (40) but not for SRB. Before a biotreatment scheme can be designed and implemented, studies must go beyond the identification of novel processes, to quantifying the kinetics of such a process. Because of the metabolic diversity of the SRB, it is important to identify the species involved in the processes studied. Since sulfate reduction rates depend upon the species used and experimental conditions, it is important to characterize sulfate reduction for the mixed culture used in previous uranium reduction experiments (38) that were performed under repeatable, uniform conditions.

The goals of this study were to characterize the mixed SRB culture previously isolated (38) and to develop models to describe sulfate reduction alone and sulfate reduction with concurrent uranium reduction. Utilizing a mixed cell culture was viewed as an advantage in that in an operational biotreatment scheme, purity of culture in treating large volumes of water will be difficult to maintain. Understanding U(VI) reduction by SRB in a consortium is more likely to be relevant for an operational condition. The nature of the mixed SRB culture was examined via 16S ribosomal DNA (rDNA) analysis. Models for separate sulfate reduction and concurrent sulfate and uranium reduction were fit to data generated by both a mixed and a pure culture of SRB. Sulfate and cell concentrations were varied during these experiments, which were conducted at room temperature, 21 ± 3°C. This is at the upper end of temperatures expected in natural waters that could be treated. Sulfate concentration is expected to be a variable in the treatment of uranium-contaminated waters, and cell concentration is an operational variable for treatment design. These kinetic determinations were made under conditions of insignificant growth (in the absence of nutrients other than a lactate carbon source; the sulfate concentrations present could facilitate growth, but experimental time courses were too short), and under anaerobic batch conditions, to isolate the enzymatic reductive processes of both electron acceptors from those of cellular growth processes. Quantification with modeling of the sulfate reduction rate alone and the uranium reduction rate in the presence of sulfate is important for the design of a treatment system employing SRB to remove metals and radionuclides, likely to be operational under dynamic, nonstatic conditions.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial cultures. The mixed culture of SRB was obtained as previously described (38). All chemicals utilized for these studies were reagent grade or better and were used without further purification. Continuous cultivation of SRB cells in a chemostat was carried out on an insulated magnetic stirrer in an anaerobic chamber (Bactron II, Sheldon Manufacturing, Cornelius, Oreg.) fed with anaerobic mixed gas (AMG) containing 90% N₂, 5% H₂, and 5% CO₂. The growth medium used was a modified Postgate C medium (36) containing potassium phosphate (mono) at 0.5 g/liter, ammonium chloride at 1.0 g/liter, sodium sulfate at 2.0 g/liter, calcium chloride at 0.06 g/liter, magnesium chloride at 0.06 g/liter, iron sulfate at 0.005 g/liter, sodium citrate at 0.3 g/liter, yeast extract at 0.1 g/liter, and sodium lactate (60% syrup) at 15 ml/liter. Cell concentrations ranged from 0.1 to 0.15 mg (dry weight)/ml of growth medium from the 250- or 500-ml square polycarbonate chemostats with a hydraulic residence time of 8 to 12 h. The average temperature during growth in the chemostat on the chamber stage was 21 ± 3°C. The Desulfovibrio desulfuricans (ATCC 7757) culture was grown in continuous fashion with an 8-h hydraulic residence time as described for the mixed culture above. Refrigerated stocks of both cultures were transferred to freshly prepared media approximately every 2 weeks.

Molecular characterization. For molecular characterization of the mixed cell culture used in this and a previous study (38), the methods of Hugenholtz et al. were followed (12). Bead beating for cellular disruption and DNA extraction was followed by PCR with the primers 8F universal forward and 1492R universal reverse (18). Taq polymerase-amplified PCR products were directly inserted into a PCR4-TOPO vector, followed by reaction with One Shot Competent Cells in a TOPO TA Cloning Kit (Invitrogen, Carlsbad, Calif.). Restriction fragment length polymorphism (RFLP) analysis was carried out on a mini-prepped, T3-T7 primer-amplified PCR (Invitrogen) clone colony DNA product, using the restriction enzymes HinPI and MspI (New England Biolabs, Beverly, Mass.). After bead beating, all steps were performed using a 96-well format.

Kinetics studies. As described elsewhere, a method was developed to examine the enzymatic reduction kinetic of U(VI) to U(IV) by SRB using the radionuclide ²³³U as a tracer (38). For sulfate reduction experiments, this method, in combination with the methods of Ingvorsen and colleagues (13, 14) was applied by substituting Na₂³⁵SO₄ as the radionuclide tracer. Briefly, a selected amount of the SRB cell mass (0.2 to 1.3 mg [dry weight]/ml per experiment) was obtained from a growth chemostat, washed in a sulfate-free sodium bicarbonate (2.5 g/liter) buffer, and then suspended anaerobically in a sterile medium (10 mM lactate-20 mM sodium bicarbonate) with a small magnetic stir bar, in sterile 30-ml polycarbonate septum flasks sealed with Teflon-lined butyl rubber stoppers (38). The headspace in these sealed reaction vessels contained ~10 ml of the anaerobic chamber's AMG. The turbid culture was in contact with the buffered carbon source/electron donor for 15 min prior to the addition of sulfate. Sulfate was added from a 10 mM stock solution spiked with Na₂³⁵SO₄ as a tracer, typically 25 µl of a 45,000-dpm/ml stock Na₂³⁵SO₄ solution (purchased from Isotope Products Laboratory, Burbank, Calif.). The Na₂SO₄ electron acceptor was mixed with the Na₂³⁵SO₄ spike in a syringe and fed to the cells by injection through the septum. For experiments with both electron acceptors, U(VI) was added as uranyl acetate, UO₂(CH₃COO)₂ · 2H₂O, from a 10 mM stock solution spiked with ²³³U(VI), typically 200 µl of a 23,670-dpm/ml stock ²³³U(VI) solution (purchased from Isotope Products Laboratory). The U(VI) electron acceptor was mixed with the ²³³U(VI) spike in a syringe and fed to the cells by the same method as the sulfate. The injection of electron acceptors set the time at t = 0 and marked the start of the kinetic experiment. The initial pH of these kinetics experiments was 7.2 ± 0.2.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

16S rRNA sequence. Cells from the isolated mixed culture initially appeared to be all of vibrio shape, all stained gram negative, and exhibited an active production of sulfide. With time (months), small gram-positive rods were observed and spores were periodically present. At any one time the gram-positive species represented 0 to 10% of the total cells present. For these reasons molecular characterization was needed to define the culture. The decision to sequence 500 bp of the 16S rRNA gene fragments from clones of this mixed culture was made because of differences in banding patterns on RFLP gels. Two hundred twenty clones were subjected to RFLP analysis, and there appeared to be two distinct banding patterns, with one pattern far more prevalent than the other, by a ratio of 10:1. The most prevalent pattern was found to be 99% identical to that of a Desulfovibrio vulgaris strain (PT-2; accession number M98496 as described by Kane et al. [16]). The other pattern was found to be similar to that of a unique, low-G+C, gram-positive, anaerobic genus (94% identical to Clostridium butyricum over 500 bp considered; accession number M59085 as described by C. R. Woese, D. Yang, and L. Mandelco [unpublished data]). Sequences were manually aligned with similar sequences using the BLAST server. RFLP analysis performed on the culture 1 year earlier yielded the same patterns, indicating no change in culture constituents. The SRB culture was initially isolated by picking one colony from an agar deep tube method (36) at a high dilution; with time, the culture came to be contaminated by the spore-forming Clostridium sp. This was viewed as an advantage because such a contaminant would likely come to be present in a bioremediation treatment scheme. This mixed cell culture was stable over a period of 5 years and showed little variability.

Sulfate reduction kinetics. The reduction of sulfate concentrations ranging from 0.1 to 100 mM was examined for the mixed cell culture. Cell concentrations were similar for all sulfate concentrations except 0.1 mM, where the cell concentration was approximately 1/10 of that in the other sulfate reduction experiments (Table 1). For an initial sulfate concentration (S₀) of 100 mM, there was no measurable reduction over 3 h. For an S₀ of 10 mM, 45% of the sulfate was reduced in 3 h by a similar cell concentration. There was 100% removal of sulfate when S₀ was 1.0, 0.25, or 0.1 mM, as shown in Fig. 1.

View larger version (21K): [in this window] [in a new window]   FIG. 1.   Time course of sulfate reduction by the mixed cell culture fit with a zero-order model. Model lines are based on the coefficients of Table 1. Each set of data points represents an average of at least two experiments with the same mixed cell culture. , S0 = 1.0 mM and X0 = 0.53 mg (dry weight) of cells/ml; , S0 = 0.25 mM and X0 = 0.57 mg (dry weight) of cells/ml; , S0 = 0.1 mM and X0 = 0.1 mg (dry weight) of cells/ml. Error bars, standard errors.

Modeling of sulfate reduction. A zero-order model, with respect to sulfate concentration, best fit the experimental data where

(1)

and S is the model predicted millimolar concentration of sulfate, S₀ is the initial millimolar concentration of sulfate, k₀ is the maximum specific reaction rate coefficient expressed as the millimolar concentration of SO₄² per milligram (dry weight) of cells per milliliter per minute, X is the bacterial cell concentration in milligrams (dry weight) per milliliter, t is time in minutes, and t_L is the lag time until the onset of sulfate reduction. A 30-min lag time was observed only with a low bacterial cell concentration (0.07 mg [dry weight] of cells/ml or lower) and an S₀ of 0.1 mM sulfate, as shown in Fig. 2. For an S₀ of 0.1 mM and an X₀ of 0.1 mg (dry weight) of cells/ml, there was no lag time. Reduction of sulfate alone proceeded without a lag for initial sulfate concentrations of 1 to 10 mM. The zero-order model is a simplification of Michaelis-Menten and Monod type kinetics at high substrate concentrations. Parameter equivalence between models can be given by

(2)

where k₀ is defined, µ_m is the Monod maximum specific growth rate constant in units of 1 h¹, Y represents cell yield, expressed as mass of cells in milligrams per milligram of substrate used, and V_max is the Michaelis-Menten maximum substrate utilization rate constant, expressed as the millimolar concentration of SO₄² per milligram (dry weight) of cells per milliliter per minute. Zero-order model fits to sulfate reduction data are shown in Fig. 1, 2, and 3; correlation coefficients, r², were 0.92.

View larger version (23K): [in this window] [in a new window]   FIG. 2.   Reduction of 0.1 mM sulfate by low mixed-cell concentrations. , 0.1 mg (dry weight) of cells/ml (no lag time); , 0.07 mg (dry weight) of cells/ml (30-min lag time); , 0.06 mg (dry weight) of cells/ml (no sulfate reduction). Zero-order model lines are fit through the 0.1- and 0.07-mg (dry weight)/ml data points, representing model values given in Table 1. Each data set represents averages from two experiments. Error bars, standard errors.

View larger version (19K): [in this window] [in a new window]   FIG. 3.   Model simulations of variable sulfate concentration with approximately equal cell concentrations. , S0 = 10 mM and X0 = 0.49 mg (dry weight) of cells/ml; , S0 = 1.0 mM and X0 = 0.53 mg (dry weight) of cells/ml; , S0 = 0.25 mM and X0 = 0.57 mg (dry weight) of cells/ml. Error bars, standard errors.

2

1

2

1

2

1

View larger version (23K): [in this window] [in a new window]   FIG. 4.   Reduction of 1 mM sulfate by the mixed cell culture and by the pure culture of D. desulfuricans. , 323.0 mg (wet weight) of cell mass, which is equivalent to 0.53 mg (dry weight)/ml of medium used in the batch experiment, for the mixed cell culture; , 298.0 mg (wet weight) of cell mass, which is equivalent to 0.30 mg (dry weight)/ml of medium used in the batch experiment for the pure culture of D. desulfuricans (ATCC 7757). Zero-order model lines are fit to the plotted data. Error bars, standard errors.

Uranium and sulfate reduction. Time courses for the reduction of soluble U(VI) to insoluble U(IV) and for the reduction of soluble sulfate to sulfide for a typical experiment are shown in Fig. 5. Rate constants were determined only for the removal of the U(VI) and sulfate. Uranium reduction was more rapid in the presence of sulfate than in its absence for both the mixed and pure cell cultures. Figure 6 shows data for the reduction of U(VI) alone (38) and for uranium reduction with sulfate [at electron equivalent amounts of U(VI) and sulfate] for the mixed cell culture. Because the reduction of sulfate to sulfide is an 8-electron transfer and the reduction of U(VI) to U(IV) is a 2-electron transfer, the sulfate concentration used was equivalent to 25% of the uranium concentration. A 90-minute lag time for the reduction of U(VI) only was decreased to 5 ± 5 min for U(VI) in the presence of sulfate for an initial U(VI) concentration (U₀) of 1 mM, an S₀ of 0.25 mM, and an X₀ of 0.5 mg (dry weight) of cells/ml.

View larger version (28K): [in this window] [in a new window]   FIG. 5.   Time course of 1 mM uranium and sulfate reduction by the mixed cell culture dominated by a Desulfovibrio sp. Insoluble U(IV) (as uraninite [38]) was collected in pellet fractions of sample aliquots, and insoluble 35S was collected as insoluble sulfides after reacting with zinc acetate in sample aliquots. Solid circles at time zero and 180 min represent total activity of uranium, 233U [U(VI) plus U(IV)], for mass balance accountability; starbursts at the same time points show total activity for 35S (35SO42 plus un-ionized sulfides [35S2]) for mass balance. Data are for one experiment with 0.51 mg (dry weight) of cells/ml.

View larger version (25K): [in this window] [in a new window]   FIG. 6.   Time course of uranium and sulfate reduction by the Desulfovibrio sp.-dominated mixed culture, with a U0 of 1.0 mM and an S0 of 0.25 mM. , U(VI) reduction in the absence of sulfate, with a U0 of 1.0 mM and an X0 of 0.46 mg (dry weight) of cells/ml (38); , U(VI) reduction (U0 = 1.0 mM) in the presence of 0.25 mM sulfate; , X0 = 0.48 mg (dry weight) of cells/ml. Each data set is an average from three experiments with the mixed cell culture. Error bars, standard errors.

Modeling of uranium and sulfate reduction. Under all experimental conditions tested, sulfate reduction was best fit by a zero-order model (equation 1), and uranium reduction was best fit by a first-order model. The first-order model fit to the U(VI) reduction data was

(3)

where U is the model predicted millimolar concentration of uranium, U₀ is the initial millimolar concentration of uranium, k₁ is the first-order rate constant, expressed as milligrams (dry weight) of cells per milliliter per minute, and X and t are as defined above. A linearized form of equation 3 was fit to the data to determine the rate constant k₁,

(4)

with units and terms as defined above. The model fits using average coefficients, and the data are shown in Fig. 7A for the mixed cell culture and in Fig. 7B for D. desulfuricans (ATCC 7757). In all cases the models fit the data with coefficients of determination, r², of 0.96 or higher. Figure 8 shows the respective model fits for uranium reduction in the presence of a sulfate concentration that might be found in freshwater. Figure 9 shows the uranium reduction kinetics in the presence of a higher, 10 mM sulfate concentration relevant for high-sulfate natural waters containing uranium.

View larger version (29K): [in this window] [in a new window]   FIG. 7.   (A) Time course of concurrent uranium and sulfate reduction for a U0 of 1.0 mM and an S0 of 1.0 mM by the mixed cell culture. , sulfate concentration; , U(VI) concentration. Lines represent zero-order and first-order models fit to sulfate and uranium reduction data, respectively. Data points are average values from two separate experiments. Average X0 = 0.51 mg (dry weight) of cells/ml. , reduction of 1 mM U(VI) alone by 0.50 mg (dry weight) of cells/ml by the same mixed cell culture (38). Error bars, standard errors. (B) Time course of concurrent uranium and sulfate reduction for a U0 of 1.0 mM and an S0 of 1.0 mM by the pure culture D. desulfuricans (ATCC 7757). Symbols are as described for panel A. Data points are average values from two separate experiments. Average X0 = 0.29 mg (dry weight) of cells/ml. , reduction of 1 mM U(VI) alone by 0.32 mg (dry weight) of cells/ml (38). A model for this reduction is reported elsewhere (38). Error bars, standard errors.

View larger version (22K): [in this window] [in a new window]   FIG. 8.   Zero- and first-order models applied to the enzymatic reduction of 0.25 mM sulfate and 1 mM uranium by the mixed cell culture, respectively. Each set of data points is an average from three experiments with the same mixed cell culture. Models are fit to plotted data with values given in Table 2. X0 = 0.48 mg (dry weight) of cells/ml. Error bars, standard errors.

View larger version (21K): [in this window] [in a new window]   FIG. 9.   Zero- and first-order models applied to the enzymatic reduction of 10 mM sulfate and 1 mM uranium by the mixed cell culture, respectively. The average X0 for the three experiments with the same mixed cell culture represented is 0.46 mg (dry weight) of cells/ml. Error bars, standard errors.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Various aspects of the SRB reduction of U(VI) to U(IV) have been studied. Lovley and Phillips (21) found that D. desulfuricans was capable of uranium reduction. They later demonstrated that cytochrome c₃ was an essential component of uranium reduction by D. desulfuricans (22). Ganesh et al. (9) considered the SRB reduction of U(VI) in organic complexes. Tebo and Obraztsova (39) identified an SRB capable of growth with U(VI) as an electron acceptor. Spear et al. (38) established the rate constants for uranium reduction by a mixed SRB culture and by D. desulfuricans. In all cases U(IV) was precipitated from solution as the mineral uraninite, UO₂. Three studies have considered the reduction of U in the presence of the native sulfate electron acceptor, and that was for a pure culture of D. desulfuricans (21, 30, 31, 40a). These studies however, have not modeled the kinetics involved across a range of solution conditions. A few reports have described the kinetics for the enzymatic reduction of sulfate alone under various conditions (8, 13, 14, 21, 28, 29, 33, 34, 37). Ingvorsen and Jørgensen (14) provided kinetics information for four SRB pure cultures at 20°C; Ingvorsen et al. (13) provided the kinetics for both batch and chemostat cultures at 30°C; and Sonne-Hansen et al. (37) provided the kinetics for sulfate reduction by two species of thermophilic SRB at 70°C.

Sulfate reduction. For Desulfovibrio vulgaris (Hildenborough), Ingvorsen and Jørgensen (14), found a V_max of 1.1 µmol of SO₄²/mg (dry weight) · h¹. For D. postgatei, Ingvorsen et al. obtained a V_max of 4.2 µmol of SO₄²/mg (dry weight) · h¹ (13). The Desulfovibrio sp. identified in the mixed cell culture for this report has a V_max range of 2 to 5 µmol of SO₄²/mg (dry weight) · h¹ (Table 1), similar to the V_max for D. postgatei and for D. vulgaris (Hildenborough) at 21 ± 3°C (13, 14). Vester and Ingvorsen report that 4.1 × 10¹⁴ mol of SO₄²/cell · day¹ could be reduced by a pure culture of Desulfobulbus propionicus by using a direct cell count method, and a value of 2.43 × 10¹³ using a T-MPN (tracer most-probable-number) method to calculate cell number (42). The range in this study was 0.73 × 10¹⁴ to 1.2 × 10¹⁴ mol of SO₄²/cell · day¹, consistent with those of Vester and Ingvorsen and others (13, 15).

Concurrent uranium and sulfate reduction. The design of a uranium removal biotreatment system employing SRB requires a knowledge of the individual and concurrent rates of U(VI) and sulfate reduction. Bioreactor systems can be designed for sequential growth and U(VI) reduction or for concurrent growth and U(VI) reduction, depending upon the system layout and the SRB employed. Rate information is needed to design a growth reactor that integrates into its design the potential for the competitive effects of concurrent U(VI) and SO₄² reduction in a combined growth and U(VI) reduction system.

Conclusion. For the mixed cell culture, a reproducible and predictable lag time until the onset of reduction for both the native electron acceptor, sulfate and U(VI) is possible. A cell concentration-based threshold until the onset of sulfate reduction can begin was reproducibly found for the mixed cell culture, resulting in a described lag time. This culture exhibited a similar lag time in reducing U(VI) alone, though at a higher cell concentration. Zero- and first-order models best fit the data for the concurrent removal of sulfate and uranium, respectively, suggesting that the rate-limiting step for each electron acceptor's reduction is not the same. These studies were performed at room temperature for both cultures, the upper-end temperature of natural waters. For a bio-based treatment system this is important, as U(VI)-containing waters are naturally cool. For the cultures tested herein, reduction of aqueous U(VI) was enhanced by the presence of aqueous sulfate. The presence of sulfate both minimizes the lag time and increases the overall rate.