Environ. Sci. Technol. 2007, 41, 1877-1882
Reduction of Nickel and Uranium Toxicity and Enhanced Trichloroethylene Degradation to Burkholderia vietnamiensis PR1301 with Hydroxyapatite Amendment J O Y D . V A N N O S T R A N D , †,⊥ TATIANA J. KHIJNIAK,‡ BENJAMIN NEELY,† M. ABDUS SATTAR,# ANDREW G. SOWDER,§ GARY MILLS,§ P A U L M . B E R T S C H , §,† A N D P A M E L A J . M O R R I S * ,†,| Marine Biomedicine and Environmental Sciences Center, Department of Microbiology and Immunology, and Department of Biostatistics and Bioinformatics, Medical University of South Carolina, Charleston, South Carolina 29412, Savannah River Ecology Laboratory, University of Georgia, Aiken, South Carolina 29802, and Center for Coastal Environmental Health and Biomolecular Research and Hollings Marine Laboratory, U.S. National Oceanic and Atmospheric Administration, Charleston, South Carolina 29412
The use of hydroxyapatite (HA) to sequester metals at mixed waste sites may reduce metal toxicity and facilitate microbial degradation of cocontaminant organics. The constitutive trichloroethylene (TCE) degrader, Burkholderia vietnamiensis PR1301, grew at 34.1 and 1.7 mM Ni at pH 5 and 7, respectively, with 0.01 g mL-1 HA compared to 17 and 0.85 mM Ni without HA. PR1 grew at 4.2 mM U at pH 5 and 7 with 0.01 g mL-1 HA compared to 1.1 mM U without HA. A similar decrease in the toxicity of Ni and U in combination was observed with HA. The ability of PR1 to degrade TCE at 0.85, 1.7, and 3.4 mM Ni and at 0.42 and 1.1 mM U was examined. The presence of TCE resulted in a decreased tolerance of PR1 to Ni and U; however, HA facilitated TCE degradation in the presence of Ni and U, effectively doubling the metal concentrations at which TCE degradation proceeded. These studies suggest that metal sequestration via HA amendments may offer a feasible approach to reducing metal toxicity to microorganisms at mixed waste sites, thereby enhancing the degradation of cocontaminant organics.
Introduction Binary and ternary mixtures of contaminant classes (e.g., metals, radionuclides, chlorinated hydrocarbons) have been * Corresponding author telephone: (843) 762-8803; fax: (843) 7628737; e-mail:
[email protected]. † Marine Biomedicine and Environmental Sciences Center, Medical University of South Carolina. ⊥ Current Address: Institute for Environmental Genomics, Department of Botany and Microbiology, The University of Oklahoma. ‡ Microbiology and Immunology, Medical University of South Carolina. § Savannah River Ecology Laboratory, University of Georgia. # Biostatistics and Bioinformatics, Medical University of South Carolina. | U.S. National Oceanic and Atmospheric Administration. 10.1021/es0616581 CCC: $37.00 Published on Web 02/09/2007
2007 American Chemical Society
identified at 64% and 49% of U.S. Department of Energy (DOE) facility waste sites, respectively (1). Although microorganisms can degrade many organic contaminants, cocontaminant metals can inhibit processes required for attenuation and bioremediation; however, few studies have addressed the impacts of cocontaminants on degradation of organics (2-4). Thus, a major challenge in developing remediation strategies for mixed waste sites includes understanding how the presence of metals affect microbial processes involved in contaminant degradation or sequestration (5). For over three decades the Tims Branch watershed, located on the DOE’s Savannah River Site (Aiken, SC), received metalcontaminated wastewater from production of U-Al alloy fuel and depleted U targets (6). This waste was composed primarily of U and Ni with lesser amounts of other metals. Additionally, an ∼8 km2 subsurface plume of chlorinated organic solvents, primarily trichloroethylene (TCE) and tetrachloroethylene, originating from this processing facility (7) has begun to contaminate riparian sediments along the Tims Branch corridor (8). Previous studies have examined the chemical and biological availability of Ni and U at this site. Based on sequential chemical extractions of contaminated sediments (9), uptake by indigenous plants (10), and isolation of Ni-tolerant bacteria from within the Tims Branch corridor (11), the data suggests that Ni is more available to a number of bioreceptors than U. Considering the known toxic effect of these metals, they could potentially hinder intrinsic degradation of TCE at this site. Hydroxyapatite (HA) [Ca10(PO4)6(OH)2] is recognized for its ability to sequester Pb and other metals, including U and Ni (12-15), to reduce metal toxicity. HA sequesters cationic metals (e.g., Ni, Cd, Zn) by cation exchange with Ca, surface complexation (16, 17), or metal phosphate precipitation (12, 16). Surface complexation of U was found to be the dominant reaction followed by U-phosphate precipitation (18). HA has been shown to reduce the soluble fraction of U and Ni in contaminated sediments (14, 15). Our studies focused on Ni and U toxicity to Burkholderia vietnamiensis PR1301 (PR1), a constitutive TCE-degrader that has been used in both microcosm (19) and field studies (20). We used a microbial model system in the laboratory to examine the effect of mixed wastes (e.g., Ni, U) on PR1 with pH as an environmental variable, since we demonstrated previously that PR1 exhibits a pH-dependent Ni toxicity (21). We also examined the efficacy of a sequestration agent (HA) for reducing cocontaminant metal toxicity and the resulting enhancement of microbial degradation of TCE.
Materials and Methods Preparation of Solutions, Media, and Bacterial Inoculum. Glassware was washed in 25% HCl and rinsed with distilled water. Metal solutions were prepared using Ni(NO3)2‚6H2O (J T Baker; Phillipsburg, NJ) and UO2(NO3)2‚6H2O (Alfa Aesar; Ward Hill, MA) in deionized water and filter sterilized (0.2 µm nylon, Fisher Scientific, Suwanee, GA). The growth medium (4M) was designed to reduce phosphate complexation and precipitation of metals (21) by using 100 mM 2-(Nmorpholino)ethanesulfonic acid (MES) (J.T. Baker) and 3 mM β-glycerophosphate (Sigma, St. Louis, Mo) to reduce inorganic phosphate concentration. Burkholderia vietnamiensis PR1301, a mutant of B. vietnamiensis G4 that constitutively degrades TCE (19), was obtained from Dr. Malcolm Shields (Idaho State University). PR1 culture stocks were stored, and inoculum was prepared as described previously (21). The VOL. 41, NO. 6, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
1877
purity of all samples was evaluated at the beginning and end of each experiment. Toxicity of Ni and U to B. vietnamiensis PR1301 in the Presence and Absence of Hydroxyapatite. Erlenmeyer flasks (125 mL) with Teflon-lined screw caps containing 4M (25 mL; pH 5 or 7) amended with Ni and/or U were inoculated with 250 µL of culture. For experiments with HA, aliquots of HA (0.01 g mL-1) (Aldrich, Milwaukee, WI) were weighed into the flasks and autoclaved. This concentration of HA is analogous to a 1% soil application [based on a 1:1 (mL:g) solution to soil ratio], which would be practical for field applications. Media (4M; 24.75 mL; pH 5 or 7) and metals (Ni and/or U) were added to the flasks and incubated with the HA for 24 h at room temperature with shaking (200 rpm) and then inoculated. Cultures were incubated with shaking (200 rpm) at room temperature (average 24 °C) for 36-72 h. Growth was monitored by total cellular protein (22). Inhibition of growth was used as an indicator of toxicity. At time zero and at 12 h, 1 mL was removed from cultures grown at 3.4 mM Ni and/or 1.05 mM U with and without HA and the corresponding abiotic controls. Aqueous metal (Ca, Co, Cu, Fe, K, Mg, Mn, Mo, Ni, U, and Zn) concentrations were analyzed to determine the influence of HA. Samples were filtered (0.2 µm, 25 mm nylon; National Scientific Company, Duluth, GA), acidified with 1% Optima nitric acid (1:10 dilution) (Fisher), and measured by inductively coupled plasma-mass spectrometry (ELAN 6100 DRC, Perkin-Elmer Corp., Norwalk, CT). Correlation analysis was performed with SPSS 12.0 (SPSS Inc., Chicago, IL) to determine relationships among protein and nutrient metal concentrations. For nutrient metals, concentrations at 12 h were subtracted from time zero results to create univariate outcome variables (i.e., HA, PR1, pH, Ni, and U) that were used for regression analyses. Statistical p-values were not adjusted for multiple testing. Calculated Speciation of U and Ni in 4M with Hydroxyapatite. Speciation of Ni and U in 4M with HA was calculated for the pH range 5-7 using the MINTEQA2, v.3.11 thermodynamic model (EPA, Center for Exposure Assessment Modeling, Athens, GA). A revised thermodynamic database for U species (Dr. David R. Turner, Center for Nuclear Waste Regulatory Analyses, San Antonio, TX 78228) was used. Medium components considered in the modeling and the modified thermodynamic database for Ni was as described previously (21). HA was entered as an infinite solid in the MINTEQA2 input file. TCE Degradation by B. vietnamiensis PR1301 in the Presence of Ni and U with and without Hydroxyapatite. A TCE stock solution was prepared in N,N-dimethylformamide and filtered (0.2 µm, nylon) (all from Fisher) prior to initiation of experiments. Serum bottles (70 mL) containing 4M (13.86 mL; pH 5 or 7; Ni or U) were inoculated with 140 µL culture, and then aliquots of TCE were added using an autoclaved 10 µL airtight syringe (Hamilton Company, Reno, NV) so that the final aqueous phase TCE concentration was 50 µM [as calculated by Yeager, et al. (23)]. Abiotic controls (pH 5 and 7, no metal amendment) were set up in the same manner (without inoculation). The bottles were immediately crimpsealed with Teflon-lined septa and incubated with shaking (200 rpm) at 24 °C. Six sets of bottles (each set included triplicates of all conditions examined and abiotic controls) were prepared. One set was sacrificed at each time point. For TCE analysis, a 4 mL aliquot was placed into a 10 mL headspace vial containing 1% sodium azide (w/v) (24). These samples were then spiked with 5 µL of 25.24 mM dibromochloromethane (DBCM, Ultra Scientific, North Kingstown, RI), an internal standard with a similar Koc used to normalize for TCE biomass partitioning, and immediately crimp-sealed with Teflon-lined septa. Headspace vials were stored in the dark at 4 °C until analyzed, but no longer than 2 weeks. 1878
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 6, 2007
FIGURE 1. Growth of Burkholderia vietamiensis PR1301 in 4M amended with Ni and/or U. Ni concentrations tested: 0 (O), 0.85 (9), 1.7 (2), 3.4 (0), 17 ([), 34.1 (×) mM Ni (A, B); U concentrations tested: 0 (O), 0.42 (2), 1.1 (0), 2.1 ([), or 4.2 (×) mM U (C, D); Ni and U concentrations tested: 0, 0 (O), 0.85, 1.1 (2), 0.85, 2.1 (0), 1.7, 0.42 ([), 1.7, 1.1 (×),1.7, 2.1 (4),3.4, 0.42 (9), 3.4, 1.1 (b), 3.4, 2.1 (]), 17, 2.1 (/) Ni and U mM, respectively (E, F) at pH 5 (A, C, E) and 7 (B, D, F) without HA. Due to increased toxicity of Ni at pH 7, 0.85 mM Ni was added to pH 7, and 17 and 34.1 mM Ni were omitted. Error bars represent standard deviation of triplicate flasks. Symbols may be larger than error bars. Note the different x-axis. Samples (200 µL) were also removed to monitor growth by protein analysis (22). Time zero samples were taken 5 min after inoculation to allow for equilibration of TCE and medium (25). The concentration of vapor-phase TCE was analyzed using an Agilent 5890 Series II gas chromatograph (Palo Alto, CA) equipped with an automated headspace sampler (Model 7694) and electron capture detector. Separation was achieved using a 30 m DB-VRX (0.25 mm i.d., 1.4 µm film) column (Agilent Technologies; Santa Clara, CA). The initial column temperature (55 °C) was held for 10 min and then ramped at 12 °C min-1 to the final temperature (190 °C). The carrier gas was He at a flow rate of 1.2 mL min-1. Data was analyzed using GC ChemStation Rev A.09.01 (Agilent Technologies). TCE concentrations were calculated using the internal standard DBCM and TCE/DCBM response factor previously determined using the same matrix and analysis conditions.
Results and Discussion Toxicity of Ni and/or U. The toxicity to PR1 of Ni, U, and a combination of the two was examined with and without HA. In the absence of HA, PR1 grew at higher concentrations of Ni at pH 5 (17 mM) than pH 7 (0.85 mM) (Figure 1A,B). This pH-dependent Ni toxicity has been observed previously with PR1 (21) and with Gram-positive isolates from Steed Pond (11) but was not observed with U in this study. PR1 was able to grow at 1.1 mM U but not at 2.1 mM U at pH 5 and 7 (Figure 1C,D). An increase in toxicity when PR1 was exposed to both Ni and U was evident from the longer lag phase compared to
FIGURE 2. Growth of Burkholderia vietnamiensis PR1301 in 4M amended with Ni and/or U with 0.01 g mL-1 HA. Ni concentrations tested: 0 (O), 0.85 (9), 1.7 (2), 3.4 (0), 17 ([), 34.1 (×) mM Ni (A, B); U concentrations tested: 0 (O), 0.42 (2), 1.1 (0), 2.1 ([), or 4.2 (×) mM U (C, D); Ni and U concentrations tested: 0, 0 (O), 0.85, 1.1 (2), 0.85, 2.1 (0), 1.7, 0.42 ([), 1.7, 1.1 (×),1.7, 2.1 (4),3.4, 0.42 (9), 3.4, 1.1 (b), 3.4, 2.1 (]), 17, 2.1 (/) Ni and U mM, respectively (E, F) at pH 5 (A, C, E) and 7 (B, D, F) with 0.01 g mL-1 HA. Error bars represent standard deviation of triplicate flasks. Symbols may be larger than error bars. Ni and U individually (Figure 1). At pH 5, the lag phase at 1.7 mM Ni was 6 h (Figure 1A) and at 1.1 mM U (Figure 1C) was 30 h. When exposed to both Ni and U, the lag phase was over 50 h for the same concentrations (Figure 1E). This trend was also observed at pH 7. This increased lag phase suggests that more time was required for the cells to adjust to the increased stress of both metals. Additionally, at pH 5, PR1 grew at 17 mM Ni and at 0.42 mM U; however, when Ni and U were added together, total growth inhibition occurred at these concentrations even after 72 h (Figure 1), suggesting a synergistic toxicity. Synergistic toxicity with Ni has been observed with Cu (26, 27), Cd, and Cr (27). Fulladoas et al. (28) speculated that a synergistic toxicity could be the result of different metals acting on the same “energy supplying pathway”, with one metal acting early in the pathway and the other acting later. Toxicity of Ni and U with Hydroxyapatite Amendment. Addition of 0.01 g mL-1 HA decreased Ni and U toxicity to PR1. With HA, PR1 grew at 34.1 or 1.7 mM Ni at pH 5 and 7 and at 4.2 mM U at pH 5 and 7, respectively. The increase in protein concentration at these concentrations in the presence of HA was significantly higher than in the absence of HA (T-test, p < 0.01). Similar decreases in toxicity were observed when PR1 was exposed to both Ni and U (Figure 2). With respect to toxicity in other systems, amendment of soils with HA reduced the concentration of extractable metals (e.g., Cu, Zn, Cd) and decreased metal uptake by maize plants (29). HA also reduced the soluble and the toxicity characteristic leaching procedure (TCLP) extractable fractions of U and Ni in sediments (14, 15). A decrease in bioavailable metals is presumed to have resulted in the decreased Ni and U
toxicity observed in the current study (vide infra). At pH 5 in the absence of HA, the Ni concentration at which 50% growth inhibition (IC50) occurred after 24 h for PR1 was 17.3 mM ((1.5). A significant (T-test, p < 0.01) increase in IC50 occurred with HA, to >34.1 mM Ni, a decrease in toxicity of >50% (>17 mM Ni). At pH 7, the IC50 without HA is 1.2 mM Ni ((0.03), while with HA the IC50 is 2.8 mM Ni ((1.12), a decrease in toxicity of 1.6 mM Ni or approximately 50%. The differences in Ni toxicity reduction by HA at pH 5 and 7 could be explained by the dissolution of HA, which occurs more readily at lower pH values (30, 12). Dissolution of HA results in an increased concentration of Ca and phosphate (Figure S1C,D), allowing for more metal phosphate precipitation, thereby decreasing toxicity. Aqueous phase speciation generated by MINTEQ indicated similar speciation patterns at pH 5 and pH 7 at 3.4 mM Ni (21), and the presence of HA did not alter predicted Ni speciation (data not shown). Additionally, the aqueous Ni concentration decreased by a similar concentration with HA at pH 5 and 7 (vide infra). However, reduction of U toxicity was not similarly affected by pH, suggesting that differential dissolution of HA at pH 5 and 7 is not affecting toxicity. Therefore, differential dissolution of HA does not explain the differences in the level of Ni toxicity in the presence of HA at pH 5 and 7. This pH-dependent Ni toxicity was also observed in the absence of HA, suggesting an underlying biological mechanism (21). The differences in the influence of pH on reduction of Ni and U toxicity by HA may be due to the mechanism by which HA sequesters each metal. Cationic metals (Cd, Zn, Ni) are sequestered by substitution for structural Ca, surface complexation (16, 17, 31), or precipitation as metal phosphates (12, 16). Surface complexation of U is the dominant reaction followed by U-phosphate precipitation (18). Alternatively, since HA is better able to reduce the soluble fraction of U compared to Ni (15), higher concentrations of U might demonstrate differences in the toxicity reduction of U at pH 5 and 7 similar to that observed by Ni. Aqueous Ni and U Concentrations. To evaluate the ability of HA to sequester Ni and U, the aqueous concentrations (defined as the fraction of metal able to pass through a 0.2 µm filter) of each was examined at time zero and after 12 h of growth. Without HA, the average aqueous Ni concentrations in the 3.4 mM Ni biotic samples at time zero, when the number of bacteria was low, were 2.6 ((0.1) and 2.6 ((0.2) mM Ni at pH 5 and 7, respectively (Figure S1A). Aqueous Ni concentrations in abiotic samples were similar, and Ni concentrations in both biotic and abiotic samples remained relatively constant over 12 h, suggesting that PR1 does not accumulate significant amounts of Ni, as was observed previously (21). Aqueous U concentrations decreased over time. At time zero in the absence of HA, U concentrations for the 1.1 mM biotic samples averaged 0.69 ((0.01) and 0.02 ((0.001) mM U at pH 5 and 7, respectively (Figure S1B). Following 12 h of growth with PR1, U concentrations decreased to 0.08 ((0.01) mM U at pH 5, and below the detection limit at pH 7, a similar decrease did not occur in abiotic controls. This decrease in aqueous U could be the result of sorption by PR1. Several Gram-negative bacteria (e.g., Escherichia coli, Pseudomonas sp., Serratia marcescens) accumulate large concentrations of U (70-300 mg g-1 dry weight at pH 5.8 and 125 mM U) (32). PR1 may be accumulating U or causing U-precipitation thus removing it from the aqueous phase. The calculated speciation of 1.1 mM U in 4M predicts that 53% of the U is precipitated with 98% of the U is predicted to have precipitated with no U present as the free ion. The concentration of measured aqueous phase U at 1.1 mM in abiotic samples [0.24 ((0.01) and 0.08 ((0.01) mM U at pH 5 and VOL. 41, NO. 6, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
1879
TABLE 1. TCE Degradation at 36 h by PR1 in 4M at pH 5 and 7a pH 5
TCE lossb (growthc) TCE loss (growth)
0 mM
3.4 mM Ni
NDd (100)
99 (100)
ND (100)
ND (112)
17 mM Ni
34.1 mM Ni
0 mM HA 23 (18)
0.42 mM U
2.1 mM U
4 (0)
7 (11)
2 (0)
0.01 mM HA 29 (20) 15 (13)
ND (78)
ND (93)
pH 7
TCE loss (growth) TCE loss (growth) a
0 mM
0.85 mM Ni
ND (100)
78 (63)
0 mM HA 8 (0)
55 (56)
9 (0)
ND (92)
0.01 mM HA 34 (61)
ND (88)
ND (82)
ND (100) b
1.7 mM Ni
0.42 mM U
2.1 mM U
c
All results expressed in percentage. Percent of TCE degraded as compared the concentration at time zero. Percent of growth as compared to growth of the 0 mM metal control. d ND, Below detection limit.
7, respectively] is less than expected at pH 5 but more than expected at pH 7 based on the model predictions (0.50 and 0.01 mM U), suggesting that the system may have not reached equilibrium. The observed lag phase of PR1 at 1.1 mM U and the inability to grow at 2.1 mM U at both pH 5 and 7 (Figure 1C,D) in spite of the predicted absence of free U (Table S1), suggests that more free U is available than predicted or that other U species (e.g., U-NTA, UO2-lactate) may be toxic. Bioavailability of U to a green alga decreased when the concentration of EDTA, citrate, or phosphate was increased, suggesting that these U species were not bioavailable (33). An initial increase in U was noted after addition of PR1 at pH 5 (Figure S1). If PR1 were producing metal binding substances, then the U may have been solubilized by these compounds. The effect of such compounds would not be accounted for by the thermodynamic speciation modeling. Pseudomonas fluorescens was able to solubilize U from U ores; release was attributed to production of the siderophore pyoverdine (34). The CAS-shuttle assay (35) was used to evaluate PR1 for production of siderophores in 4M (Van Nostrand, unpublished data). PR1 produced
