Hydrogen Thresholds as Indicators of Dehalorespiration in

Environ. Sci. Technol. 2004, 38, 1024-1030

Hydrogen Thresholds as Indicators of Dehalorespiration in Constructed Treatment Wetlands GABRIEL KASSENGA, JOHN H. PARDUE,* WILLIAM M. MOE, AND KIMBERLY S. BOWMAN Department of Civil & Environmental Engineering, Louisiana State University, Baton Rouge, Louisiana 70803

Anaerobic degradation of cis-1,2-dichloroethene (cis-1,2DCE) and 1,2-dichloroethane (1,2-DCA) was studied in microcosms derived from a laboratory-scale upflow treatment wetland system used to biodegrade chlorinated compounds present in groundwater from a Superfund site. Dechlorination kinetics of cis-1,2-DCE (0.94-1.57 d-1) and 1,2DCA (0.15-0.71 d-1) were rapid, and degradation proceeded to completion with ethene or ethane as terminal dechlorination products. Hydrogen concentrations, measured simultaneously during dechlorination, were significantly different for the two compounds, approximately 2.5 nM for cis-1,2-DCE and 38 nM for 1,2-DCA. Methanogenesis proceeded during the degradation of 1,2-DCA when H2 concentrations were high but not during the dechlorination of cis-1,2-DCE when H2 concentrations were below published thresholds for methanogenesis. A 16S rRNA genebased approach indicates that microorganisms closely related to Dehalococcoides ethenogenes were present and that they were distributed throughout the bottom, middle, and top of the upflow treatment wetland system. These results coupled with consideration of hydrogen thresholds, degradation kinetics, daughter products, and measurements of methanogenesis strongly suggest that halorespirers were responsible for dechlorination of cis-1,2-DCE and that 1,2-DCA dechlorination was co-metabolic, likely mediated by acetogens or methanogens. Rapid dechlorination potential was distributed throughout the wetland bed, both within and below the rhizosphere, indicating that reductive dechlorination pathways can be active in anaerobic environments located in close spatial proximity to aerobic environments and plants in treatment wetland systems.

Introduction Constructed treatment wetlands have the potential to passively treat groundwater contaminated with chlorinated solvents, either at the point of contact with surface water bodies or as the treatment component of a pump-and-treat system (1). A treatment wetland concept is based on the documented ability of natural wetlands to attenuate chlorinated solvents in surface water or groundwater (2-4). Microbially mediated reductive dehalogenation has been identified as a key mechanism for chlorinated solvent attenuation in wetlands (2-9) as identified in other environmental settings (10, 11). The high organic carbon content * Corresponding author phone: (225)578-8661; fax: (225)578-5043; e-mail: [email protected]. 1024

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 4, 2004

of wetland soils are conducive to production of electron donors such as hydrogen and volatile fatty acids (12-14) necessary for driving microbial reductive dechlorination reactions. Microbial reductive dechlorination can be co-metabolic [via sulfate-reducing (15), methanogenic (16) or acetogenic microorganisms (17)] or respiratory [via halorespirers such as Dehalococcoides ethenogenes (18)]. Dehalorespiration is often a favored process because of the higher rates of dechlorination possible when the process is linked with microbial growth. Of the dehalorespiring organisms isolated to date (18-22), all have been observed to use a very limited set of electron donors, primarily molecular hydrogen. Although contributions of anaerobic co-metabolic processes may be negligible in many environments (11), the large methanogenic and acetogenic populations in wetland peats may be an exception. Molecular hydrogen (H2) is produced through the action of fermentative organisms in the natural environment, and it may be consumed by methanogens, nitrate-reducing bacteria, sulfate-reducing bacteria, and ferric iron-reducing bacteria (13) in addition to halorespirers. Members of these bacterial groups compete for hydrogen, and as such, competition can potentially limit reductive dechlorination due to the limited supply of H2 in anaerobic systems (10, 19, 23-26). Experimental and theoretical studies suggest that each terminal electron-accepting process has a unique “threshold” H2 concentration (24). The magnitude of the threshold depends on the physiological characteristics of the H2-consuming microorganisms and the energy yield from hydrogen oxidation coupled to the reduction of different electron acceptors. In general, the greater the energy yield, the lower the observed H2 threshold. Measurement of H2 thresholds can be used to assess the dominant terminal electron-accepting process in sediment (24) and groundwater (23). Halorespirers have been reported to exhibit higher affinities and, therefore, lower thresholds for H2 in comparison to other H2-utilizing organisms such as methanogens and sulfate-reducing bacteria. It has been suggested that this property gives halorespirers a competitive advantage when H2 concentrations are low (19, 20). Consequently, H2 threshold measurements may be particularly useful in identifying the relative contributions of halorespirers relative to populations such as methanogens that cometabolize chlorinated solvents because a large difference in threshold should be apparent (11). The present study was conducted to evaluate hydrogen dynamics and thresholds during dechlorination of cis-1,2dichloroethene (cis-1,2-DCE) and 1,2-dichloroethane (1,2DCA) in microcosms derived from laboratory-scale constructed wetland treatment systems. The objectives of the study were 2-fold: first, to determine the relative role of halorespirers in the treatment wetland and, second, to investigate the spatial distribution of reductive dechlorination potential as a function of depth in the treatment wetland bed. Both of the test compounds, cis-1,2-DCE and 1,2-DCA, are subject to dehalorespiration through organisms such as Dehalococcoides sp. (22) and co-metabolic reductive dechlorination via methanogens (27, 28).

Experimental Details Microcosm Preparation. Anaerobic microcosms were constructed from soils from vegetated laboratory-scale wetlands (1). The wetland soil columns were packed with a medium comprised of Bion Soil, a product derived from agricultural waste (Dream Maker Dairy, Cowlesville, NY), Latimer peat 10.1021/es0348391 CCC: $27.50

 2004 American Chemical Society Published on Web 01/06/2004

(Latimer’s Peat Moss Farm, West Liberty, OH), and sand mixed at a ratio of 1.3:1.1:1 (Bion Soil:peat:sand) by weight. This mixture was identified as the most promising of 10 materials evaluated for construction of a treatment wetland bed for chlorinated aliphatic compounds at a Superfund site (1). Column experiments were performed in 60 cm long, 15 cm diameter glass columns in a temperature-controlled greenhouse at 26 ( 3 °C. On a continuous basis, the laboratory-scale wetland treatment systems were supplied in an upflow mode with groundwater from a Superfund site containing chlorinated ethenes and ethanes. The laboratoryscale wetland treatment systems (50 cm in height, vegetated with Scirpus sp., and open to the atmosphere at the top) were operated for 18 months prior to the start of the present study. Soils were removed from the glass wetland treatment columns taking care to leave the material intact and minimize disturbance. The intact soil core, 50 cm in height, was cut into three sections designated as the bottom (0-10 cm), middle (20-30 cm), and top (40-50 cm). Visual inspection of the soil cores revealed that plant roots penetrated the entire depth of the column; however, root densities in the upper portion of the column were clearly higher than in the lower portion. Sectioning of the soil columns was conducted in the ambient atmosphere, and subsequently, the outer portions of the soil slices used for construction of the anaerobic microcosms were removed while under a nitrogen atmosphere in a glovebag (I2R, Cheltenham, PA) to minimize exposure to oxygen. The interiors of the soil column sections were used for construction of anaerobic microcosms. Different implements were used to collect and process samples from different spatial locations in the soil column to avoid cross-contamination. After homogenization, approximately 110 g of wet soil (51 g dry weight) from each section was placed in 160-mL serum bottles along with groundwater in a 1.5:1 volumetric ratio (wet soil:groundwater) in a manner similar to that previously reported by Lorah et al. (3). The volume of slurry in each bottle was approximately 140 mL, and the volume of headspace was approximately 20 mL. Resazurin (0.0002%) was added as a redox indicator. All reaction mixtures were then sealed with Teflon-lined rubber septa and aluminum crimp seals and incubated in an inverted position under static conditions in the dark at 25 °C. No additional nutrients or amendments were provided to the systems. Microcosms were spiked by adding either cis-1,2-DCE and 1,2-DCA stock solutions to a final aqueous-phase concentration of approximately 60 µM after first withdrawing an equivalent volume of water. Temporal monitoring of concentrations of cis-1,2-DCE, 1,2-DCA, and degradation daughter products were performed until the concentrations of the chlorinated organics dropped below the detection limit of the analytical methods (0.4-0.8 µM/L). In subsequent microcosm experiments, constructed identically to those described above, H2 and methane were monitored simultaneously with cis-1,2-DCE, 1,2-DCA, and degradation daughter products. In addition, a microcosm experiment was performed by simultaneous addition of cis-1,2-DCE and 1,2DCA, followed by monitoring of parent and daughter degradation products, H2, and methane. For all microcosm experiments, at least three replicates were used for each treatment. For aquatic sediments, Lovley and Goodwin (24) report that less than 20 min is required for dissolved H2 to come into equilibrium with gaseous H2 under static conditions. Thus, it was assumed that H2 measured in the gas headspace was at equilibrium with the liquid phase. The average H2 concentrations observed during the dechlorination of the test chemicals were calculated. Additionally, average H2 concentrations were calculated during periods when methane

was accumulating in the headspace, after the complete dechlorination of the test chemical. In all studies, triplicate microcosms were utilized, and H2 concentrations were averaged for all replicate serum bottles. To minimize the chance of errors associated with using the first-order model, variations in initial concentrations of the test chemicals were minimized during the kinetic studies as suggested by previous researchers (29). Kinetic rate constants for transformation of the test chemicals were calculated by fitting a pseudo-first-order model to the time course data for removal of the parent compounds. Equations were fit using nonlinear regression performed using Sigma Plot 6.0 (SPSS Inc., Chicago IL), which employs a MarquardtLevenberg algorithm to calculate the kinetic parameters of interest. Analytical Procedures. Chlorinated solvents were analyzed using EPA Method 8260B, and methane was measured using GC/FID as previously described in Kassenga et al. (1). Hydrogen was analyzed using a reduction gas chromatograph (Trace Analytical, Menlo Park, CA). Headspace samples were injected into a 1-mL gas sampling loop and were separated with a molecular sieve analytical column (Trace Analytical, Menlo Park, CA) at an oven temperature of 40 °C; ultrahighpurity nitrogen (BOC Gases, Baton Rouge, LA) was used as a carrier gas after it was passed through a catalytical combustion converter (Trace Analytical, Menlo Park, CA) to remove traces of H2. The detection limit for H2 under these conditions was 1 ppb. Aqueous H2 concentrations were then calculated as described by Lo¨ffler et al. (11). Statistical Analysis. Statistical analysis included ANOVA (hydrogen data), linear regression (calibration curves), and nonlinear regression using a first-order decay model (cis1,2-DCE and 1,2-DCA monitoring data). Results were considered statistically significant if P e 0.05. Extraction of Community DNA. A total of 1.5 mL of TE buffer (10 mM Tris, 1 mM EDTA, pH 8) containing 20 mg of polyvinyl polypyrrolidone (PVPP) (Agros Organics, Geel, Belgium) was added to 0.25-0.4 g of wet soil slurry samples collected from the microcosm bottles. After vortexing for 5 s, tubes were centrifuged at 2000g for 10 min, and then the supernatant was decanted and discarded. DNA was extracted from resulting soil pellets using a MoBio Ultraclean Soil DNA Isolation Kit (Salona Beach, CA) with a modified version of the manufacturer’s protocol. The first modification was use of a Biospec Mini-Beadbeater 3110BX (Biospec Products Inc., Bartlesville, OK), operated for 3 min at 4800 rpm, in place of the MoBio Vortex Adapter. The second modification involved rinsing the spin filter with MoBio S4 solution three times, as opposed to a single wash in the manufacturer’s protocol. PCR Amplification of Extracted DNA. PCR was performed using an Eppendorf Thermocycler (Eppendorf GmbH, Hamburg, Germany) under conditions previously reported by Hendrickson et al. (30): initial denaturation at 95 °C for 2 min, followed by 30 cycles consisting of denaturation at 94 °C for 1 min, annealing at 55 °C for 1 min, and extension at 72 °C for 1 min. Each reaction (100 µL) contained 1X Taq Buffer with Mg2+, 2.5 U Taq DNA Polymerase, 0.75X TaqMaster PCR Enhancer (Brinkmann Instruments, Inc., Westbury, NY), 200 µM of each dNTP (10 mM total) (Applied Biosystems, Forster City, CA), 50 pmol of each primer, and 1 µL of purified community DNA. Table 1 lists the oligonucleotide PCR primers used in this study. Two sets of primers, the first comprised of DHC1f and DHC1377r and the second comprised of DHC946f and DHC1212r, specific to variable regions of the 16S rRNA gene of Dehalococcoides group bacteria (30) were used to amplify DNA in separate samples. These primer sets produce PCR products of 1377 and 266 bp, respectively, for Dehalococcoides sp. In cases where no PCR products were detected using VOL. 38, NO. 4, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

1025

TABLE 1. Oligonucleotide Primers Used in This Study target regiona

primer

sequence (5′-3′)

DHC1f DHC774f DHC946f DHC1212r DHC1377r

Dehalococcoides sp. GATGAACGCTAGCGGCG GGGAGTATCGACCCTCTC AGTGAACCGAAAGGGAAA GGATTAGCTCCAGTTCACACTG GGTTGGCACATCGACTTCAA

1-17 774-791 946-963 1220-1199 1385-1366

8f 1492r

Universal Primers AGAGTTTGATCCTGGCTCAG GGTTACCTTGTTACGACTT

8-27 1510-1492

M13f M13r

M13 Primers GTAAAACGACGGCCAGT GGAAACAGCTATGACCATG

cloning vector cloning vector

a For the Dehalococcoides sp. primers, the coordinates are the D. ethenogenes strain 195 16S rRNA base position coordinates, and for the universal primers, the coordinates are the E. coli base position coordinates.

these primers, a nested approach was used in which bacterial DNA was first amplified using universal primers (see Table 1), and then 1 µL of the resulting products was amplified using the two Dehalococcoides-specific primer sets. PCR reactions using universal primers were carried out under conditions identical to those using Dehalococcoides-specific primers. All primers were synthesized, purified, desalted, and lyophilized by AlphaDNA (Montreal, PQ, Canada). Primers were reconstituted with TE buffer (10 mM Tris, 1 mM EDTA, pH 8) and then stored at -20 °C prior to use. PCR products were detected and quantified using 1-µL samples of the PCR reactions in conjunction with an Agilent 2100 Bioanalyzer and DNA 1000 and DNA 12000 LabChip kits (Agilent Technology, Willington, DE). Cloning and Sequencing of Amplified DE Sequences. To verify that PCR products produced using Dehalococcoidesspecific primers were in fact from Dehalococcoides sp., PCR products from select samples were cloned and sequenced. Specifically, the PCR products cloned were those produced from DNA extracted from microcosms derived from the bottom of the wetland treatment system. Cloning was accomplished with the TOPO TA Cloning Kit for Sequencing (with pCR4-TOPO) with One Shot TOP10 Chemically Competent E. coli (Invitrogen, Carlsbad, CA) following a modification of the manufacturer protocol. Randomly selected clone colonies were transferred to PCR tubes containing 25 µL of water, and DNA was extracted by heating to 98 °C for 10 min. Tubes were then centrifuged at 14000g for 1 min. Next, plasmid inserts were reamplified by PCR using primers M13f and M13r. The PCR temperature program for amplification of inserts included initial denaturation at 94 °C for 3 min, followed by 25 cycles consisting of 52 °C for 30 s, 72 °C for 1 min, and 94 °C for 30 s, and then final extension at 72 °C for 7 min. Each 50-µL reaction contained 1X GeneAmp Buffer with Mg2+ (Applied Biosystems, Foster City, CA), 2.5 U AmpliTaq DNA Polymerase (Applied Biosystems, Foster City, CA), 0.02 µM of each dNTP (Roche Applied Science, Indianapolis, IN), 250 ng of each primer, and 0.5 µL of clone DNA. Following PCR, the products were purified using an UltraClean PCR Clean-up Kit (MoBio, Carlsbad, CA). Sequencing reactions were performed on the purified products using an ABI Prism Big Dye Terminator Cycle Sequencing Ready Reaction Kit (Applied Biosystems, Forester City, CA) using minor modifications of the manufacturer’s protocol. For inserts corresponding to the 1377-bp fragment, two sequencing reactions were performed. The first used the DHC1f primer, and the second used the DHC774f primer described by Hendrickson et al. (30). Sequencing reactions associated with inserts corresponding to the 266-bp amplicon 1026

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 4, 2004

FIGURE 1. Dechlorination of cis-1,2-DCE and associated H2 and methane concentrations in microcosms derived from the bottom of the constructed wetland. Data points represent means of results from analysis of triplicate microcosms; bars indicate the range of results for individual bottles. used only the DHC 946f primer (see Table 1). Sequencing was performed using an ABI 377 Automated DNA Sequencer, and resulting sequences were edited using BioEdit (North Carolina State University, Raleigh, NC). For inserts corresponding to the 1377-bp amplicon, complete sequences were constructed from partial sequences generated using individual primers DHC1f and DHC774f. Sequences were then analyzed by BLAST in July 2003. The sequence of a clone originating from the microcosm derived from the bottom section of the treatment wetland has been deposited in the NCBI database (Genbank Accession No. AY351859).

Results and Discussion cis-1,2-DCE Dechlorination. Constructed wetland soil microbial populations degraded cis-1,2-DCE under anaerobic conditions without a noticeable lag period, consistent with the long exposure period and degradation observed in the treatment wetland (1) prior to the start of the microcosm studies (Figure 1). Pseudo-first-order rate constants for dechlorination of cis-1,2-DCE ranged from 0.94 to 1.57 d-1 (Table 2). This corresponds to half-lives of between 10 and 17 h. Two additional spikes of cis-1,2-DCE were sequentially added as described above, and results followed similar trends (data not shown). Dechlorination rate constants for cis-1,2-DCE decreased with distance from the bottom of the soil columns [ANOVA and pairwise comparisons (P < 0.05); Table 2]. Higher rate constants were expected in microcosms constructed from the lower portions of wetland column (those closest to the inlet) because contaminated groundwater was introduced to the laboratory-scale wetland treatment system in an upflow mode and rapid, complete dechlorination was observed near the inlet (1). Despite these spatial differences, half-lives were less than 1 d in microcosms constructed of wetland soil from all of the depths examined.

TABLE 2. Summary Results of Pseudo-First-Order Rate Constants of cis-1,2-DCE and 1,2-DCA in Constructed Wetland Soil and the Associated Quasi-Steady-State Hydrogen Concentrationsa biodegradation rate constant, k (d-1) (half-life, t1/2, d)

hydrogen concentration (nM)b

section

cis-1,2-DCE

1,2-DCA

cis-1,2-DCE dechlorination

1,2-DCA dechlorination

methanogenesis

bottom

1.70 ( 0.19 (0.41 ( 0.05) 1.41 ( 0.14 (0.49 ( 0.06) 1.15 ( 0.19 (0.60 ( 0.11)

0.77 ( 0.18 (0.90 ( 0.14) 0.19 ( 0.02 (3.61 ( 0.46) 0.17( 0.03 (4.13 ( 2.17)

2.49 ( 1.08

43.4 ( 8.8

45.7 ( 15.2

2.62 ( 0.45

37.8 ( 16.4

37.8 ( 14.3

2.39 ( 1.58

31.7 ( 12.9

46.8 ( 11.7

middle top a

Data represent means ( standard error.

b

Hydrogen concentrations were determined based on at least 16 observations,

Wetland soil in the present study was able to sustain high concentrations of H2 (between 35 and 226 nM) in the absence of the chlorinated solvents, cis-1,2-DCE, and 1,2-DCA (data not shown). Aqueous H2 concentrations decreased drastically after spiking with cis-1,2-DCE (Figure 1). In one representative set of microcosms from the bottom section, H2 concentrations decreased from approximately 45 nM to approximately 2 nM within 8 h of spiking cis-1,2-DCE (Figure 1). Thereafter, H2 concentrations stayed nearly constant at 2.71 ( 0.73 nM, during the period when dechlorination was occurring. After cis-1,2-DCE was completely transformed, H2 concentrations increased gradually to levels comparable to the initial concentrations. These observations strongly suggest that microbial reductive dechlorination of cis-1,2-DCE was driven by H2 as an electron donor. The fact that comparable, low H2 concentrations were maintained while vinyl chloride (VC) was transformed to ethene (after all cis-1,2-DCE had been removed) indicates that dechlorination of VC was probably also driven by H2. Other potential electron donors (the organic acids acetate, formate, lactate, succinate, benzoate, propionate, and butyrate) were measured using HPLC, but results were below detection (