Article pubs.acs.org/est
Vanadate and Acetate Biostimulation of Contaminated Sediments Decreases Diversity, Selects for Specific Taxa, and Decreases Aqueous V5+ Concentration Alexis P. Yelton,†,‡,○ Kenneth H. Williams,§ John Fournelle,∥ Kelly C. Wrighton,⊥ Kim M. Handley,§,⊥,#,△ and Jillian F. Banfield*,†,§,⊥ †
Department of Environmental Science, Policy, and Management, University of California, Berkeley, California 94720, United States Department of Marine Biology, Institute of Marine Sciences, Barcelona, 08003, Spain § Earth Sciences Division, Lawrence Berkeley National Laboratory, Berkeley, California 94720, United States ∥ Department of Geoscience, University of Wisconsin−Madison, Madison, Wisconsin 53706, United States ⊥ Earth and Planetary Science, University of California, Berkeley, California 94720, United States # Computation Institute, University of Chicago, Chicago, Illinois 60637, United States and Computing, Environment and Life Sciences, Argonne National Laboratory, Lemont, Illinois 60439, United States ‡
S Supporting Information *
ABSTRACT: Vanadium is a commercially important metal that is released into the environment by fossil fuel combustion and mining. Despite its prevalence as a contaminant, the potential for vanadium bioremediation has not been widely studied. Injection of acetate (as a carbon source) directly into an aquifer to biostimulate contaminated sediments in Colorado, United States, resulted in prolonged removal of aqueous vanadium for a period of at least two years. To further investigate this process, we simultaneously added acetate and vanadate (V5+) to columns that were packed with aquifer sediment and inserted into groundwater wells installed on the Colorado River floodplain. This allowed evaluation of the microbial response to amendments in columns that received an influx of natural groundwater. Our results demonstrate the removal of up to 99% of the added V5+(aq) and suggest microbial mediation. Most probable number measurements demonstrate up to a 50-fold increase in numbers of V5+-reducing cells in vanadium-amended columns compared to controls. 16S rRNA gene sequencing indicates decreased diversity and selection for specific taxa in columns that received vanadate compared to those that did not. Overall, our results demonstrate that acetate amendment can be an effective strategy for V removal, and that V bioremediation may be a viable technology.
■
V3+). Reduction to V4+ or V3+ can result in precipitation of Vbearing minerals,14,15 making V less bioavailable and reducing its toxicity. Vanadium bioreduction can occur in one of two ways. First, it can be respired by bacteria (via electron transfer). Previous studies have demonstrated microbial V5+ respiration in a few cases. More frequently, vanadium bioreduction has been reported without evidence of respiration.2,14−20 In these cases, bacteria may be reducing vanadium in order to detoxify it or inadvertently as the result of vanadium binding to reductases of other electron acceptors. A handful of studies showing V5+ bioreduction also report precipitation of an insoluble V4+ phase.14,15
INTRODUCTION Vanadium(V) is both a widespread environmental contaminant from mining and fossil fuel combustion and a commercially important metal. Due to its value to the steel industry and its status as an impurity in fossil fuels, vanadium contamination of the environment continues to grow with increasing industrialization. Substantial vanadium contamination in the U. S. has been recorded at 283 superfund sites,1 but its remediation has generally not been addressed, and little is known about the potential for its immobilization by bacteria.2 Vanadium is moderately toxic to animals3−6 and becomes toxic to animal cells at concentrations greater than 1−10 nM.7−11 Despite these findings, it is not currently regulated under the U.S. Environmental Protection Agency’s Safe Drinking Water Act. Vanadium naturally occurs in three oxidation states: V3+, V4+, and V5+.12 Under oxic conditions, V5+ is usually found in aqueous solution as vanadate (VO43−).13 In shallow aquifer sediments, where conditions fluctuate between oxic and anoxic, vanadium may be biotically reduced (either from V5+ to V4+ or © XXXX American Chemical Society
Received: February 11, 2013 Revised: May 19, 2013 Accepted: May 28, 2013
A
dx.doi.org/10.1021/es4006674 | Environ. Sci. Technol. XXXX, XXX, XXX−XXX
Environmental Science & Technology
Article
tracer concentrations minus colorimetric measurements of effluent V5+(aq) concentrations. In-Well Columns. For community analysis experiments, flow-through columns30,31 were dispatched into a monitoring well (MNA-01) in the aquifer that had not previously been acetate-amended. Sediments used for the field column study were recovered from the aquifer at a depth of ca. 4 m below ground surface using a backhoe excavator. Upon recovery, sediments were sieved (0.00003
(0.003%) to exclude low coverage OTUs (Figure S3) and those with a high proportion of Ns (Figure S4). The genes were aligned to each other, using the SSU-align software.37 The alignment was automatically masked with the SSU-mask program. OTUs were clustered at a 97% nucleotide identity cutoff, using USEARCH.38 A phylogenetic tree was constructed with the aligned sequences via the FastTree maximum likelihood method with options −gtr −nt and 1000 iterations of the FastTree bootstrap.39,40 Rarefaction curves were generated using the relative abundance data from the normalized priors of each EMIRGE operational taxonomic unit (OTU; Figure S5). These abundances were divided by a minimal unit of abundance, 0.003%, in order to normalize for detection limit. Rarefaction, rank abundance, and all statistical tests were carried out with the R statistical software,41 using the vegan and picante packages.42,43
■
RESULTS V Removal and V5+ Concentration. ICP-MS results from the multiyear in situ experiment indicate that two successive acetate additions led to the removal of almost all vanadium from groundwater for a period of 23 months or longer after an initial small influx of vanadium into groundwater (Figure 1). Ongoing monitoring continues to demonstrate prolonged removal of V from groundwater in the absence of additional acetate loadings (data not shown). For the in situ sediment columns amended with both vanadate and acetate (V1, V2), short-term amendment resulted in a decrease in V5+ in the column groundwater (Figure 2). After one day of injection, the groundwater turned bluethe color diagnostic of aqueous V4+14,44and the color persisted for the duration of the experiment (Figure S6). After breakthrough (first detection of bromide in the column), the vanadium influent was on average ∼6.0 mM in V1 (±2.4 mM) and 3.4 mM in V2 (±1.5 mM). Thus, these samples allow us to C
dx.doi.org/10.1021/es4006674 | Environ. Sci. Technol. XXXX, XXX, XXX−XXX
Environmental Science & Technology
Article
The results indicate that the column receiving the most vanadium influent (V1) contained 2.4 × 105 V-reducing cells/g sediment, whereas the column with less vanadium influent (V2) contained 1.1 × 106 cells/g sediment. No statistical difference (p < 0.05, Cochran’s t test35) was noted in cell numbers between V1 and V2 treatments. In comparison the acetate only and control columns contained only 2.3 × 104 (A1), 9.3 × 104 (A2), and 2.8 × 104 V-reducing cells/g (C1). These MPN counts indicate that vanadium-reducing populations were enriched in V-stimulated columns, with a statistically significant increase (p < 0.05) between the V2 column and a subset of columns not receiving vanadium (A1 and C1). The increase in V reducers in V-stimulated columns ranged from approximately 2.6-fold to 48-fold, as compared to acetate-stimulated and unstimulated columns. Community Composition and Diversity. To examine the response of the microbial community to vanadium treatment, 16S rRNA genes were reconstructed from DNA extracted from the background sediment and sediment columns. Normalized rarefaction curves constructed from these genes begin to level off, particularly those from the Vamended communities, suggesting a high (though not complete) level of sampling of the community diversity (Figure S5). 16S rRNA rank abundance curves show high levels of diversity in all samples (Figure 3). We report richness values ranging from 1340 to 1462 OTUs in vanadate-amended columns (V1, V2) to 4136 in background sediments (B). This difference in richness is largely due to the contribution of rare members in the background community. Fewer rare taxa and a higher level of dominance were detected in the acetate samples
Figure 2. Aqueous V(V) concentrations in flow-through column effluent. Injection began on 7/23/2010. MNA-01 is groundwater from the surrounding well. A1 and A2 are acetate addition columns. V1 and V2 are vanadium and acetate addition columns. C1 and C2 are no flow controls in the well.
explore the effects of different vanadium (and acetate) amendment levels on bacterial community structure. Although inflow rates differed between the vanadateamended columns due to different degrees of sediment accumulation in the effluent column tubing, the rates of V5+ removal after signal stabilization were similar (Table 1). A total Table 1. Removal Rates of V5+ and Acetatea sample
average V5+ removal rate (μM/h/g)
peak V5+ removal rate (μM/h/g)
average acetate removal rate (μM/h/g)
peak acetate removal rate (μM/h/g)
V1 V2 A1 A2
17.8 20.7 NA NA
22.5 23.7 NA NA
16.1 16.5 12.7 7.7
39.3 33.7 17.7 47.6
a
Rates are in mM/h/g of sediment and were generated on the basis of the average vanadium concentration removal from 8/5/10 to 8/9/10 multiplied by the column flow rate and divided by the mass of sediment per column.
of 17.8 and 20.7 μM V5+/h/g of sediment were removed on average from V1 and V2 (Table 1), which accounted for up to 99% of the V5+ in the column influent. Vanadium removal was concurrent with acetate oxidation as measured via ion chromatography (data not shown), with acetate oxidation rates of 16.1 and 16.5 μM/h/g. The maximum amount of V5+ removal was 7.6 mM (V1) and 6.6 mM (V2). In addition to the decrease in effluent V5+ concentration, vanadium and acetate treatment sediments were enriched in solid/sorbed vanadium as determined by acid extraction and ICP-AES (Table S1). Scanning electron microscopy imaging indicates that the vanadium can be found in all samples tested (V1, A1, MNA-01) in the form of thin (
