Desulfitobacterium dichloroeliminans Strain DCA1 during

in situ bioaugmentation test at an industrial site (Belgium). .... groundwater velocity between wells LV22 and LV51 amounted ..... Internet at http://...
1 downloads 0 Views 730KB Size
Download PDF
Environ. Sci. Technol. 2006, 40, 5544-5552

Transport and Activity of Desulfitobacterium dichloroeliminans Strain DCA1 during Bioaugmentation of 1,2-DCA-Contaminated Groundwater ANN MAES,† HILDE VAN RAEMDONCK,† KATHERINE SMITH,‡ WENDY OSSIEUR,§ L U C L E B B E , ‡ A N D W I L L Y V E R S T R A E T E * ,† Laboratory of Microbial Ecology and Technology, Faculty of Bioscience Engineering, Ghent University, Coupure Links 653, B-9000 Ghent, Belgium, Cell of Groundwater Modelling, Faculty of Science, Ghent University, Krijgslaan 281 S8, B-9000 Ghent, Belgium, and Avecom n.v., Bloemendalestraat 138, B-8730 Beernem, Belgium

The transport and activity of Desulfitobacterium dichloroeliminans strain DCA1 in 1,2-dichloroethane (1,2-DCA)contaminated groundwater have been evaluated through an in situ bioaugmentation test at an industrial site (Belgium). The migration of strain DCA1 was monitored from an injection well toward a monitoring well, and the effect of the imposed groundwater flow on its distribution was assessed by means of transport model MOCDENS3D. The results of the real-time PCR (16S rRNA gene) quantification downstream from the injection point were used to evaluate the bacterial distribution pattern simulated by MOCDENS3D. In the injection well, the 1,2-DCA concentration in the groundwater decreased from 939.8 to 0.9 µM in a 35 day time interval and in the presence of a sodium lactate solution. Moreover, analyses from the monitoring well showed that the cells were still active after transport through the aquifer, although biodegradation occurred to a lesser extent. This study showed that strain DCA1 can be successfully applied for the removal of 1,2-DCA under field conditions and that its limited retardation offers perspectives for largescale cleanup processes of industrial sites.

1. Introduction 1,2-Dichloroethane (1,2-DCA) is used as an intermediate for industrial polyvinyl chloride production, and the combined production in the United States, Western Europe, and Japan is estimated to be 17.5 × 106 tons/year (1). This compound ranks among the most abundant chlorinated C2 groundwater pollutants (Tri-explorer, EPA) (2). Because of the high water solubility (8 g/L), the recalcitrance in reductive environments, the environmental half-life of 50 years in anoxic aquifers, and the putative carcinogenic effects, this pollutant poses a long-lasting threat to humans and wildlife (3). Hence, natural attenuation of this chloroalkane, which has already been studied in the past (4-7), seems not to be a good option for * Corresponding author phone: +32 (0)9 264 5976; fax: +32 (0)9 264 6248; e-mail: [email protected]. † Faculty of Bioscience Engineering, Ghent University. ‡ Faculty of Science, Ghent University. § Avecom n.v. 5544

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 17, 2006

the remediation of contaminated sites (3). In most cases, pump-and-treat technologies are too costly and timeintensive to remediate expanded contamination plumes, and with respect to the prevailing reducing conditions in groundwater, in situ reductive technologies are more appealing to engineer (3). Chemical catalysts have been developed for the conversion of some chloroalkanes in reductive conditions (8, 9), although their use in groundwater is too costly and the complex water and subsoil composition often disturbs smooth catalysis (10). While several chloroalkanes, perchloroethene (PCE), and trichloroethene (TCE) are rather easily dechlorinated by Fe0 (11, 12) or bimetallic particles (13, 14), 1,2-DCA is resistant to reductive dechlorination by metals (3, 11). In view of the successes that have been reported for the stimulation of autochthonous PCE and TCE-metabolizing anaerobic bacteria (15, 16), the in situ biostimulation of 1,2DCA-degrading bacteria would offer an attractive alternative. However, until now, only one report described a biostimulation field trial, in which the decrease in the aqueous 1,2DCA concentration correlated with the production of methane (17). Given the full characterization of some 1,2-DCA dehalorespiring isolates, their injection into the groundwater as highly efficient detoxification catalysts should offer a cleanup strategy. Tests with Dehalococcoides ethenogenes strain 195 revealed that 1,2-DCA can be converted into ethene (18, 19) at rates of 7.5 µmol min-1 (mg of TCE reductive dehalogenase)-1 (20), indicating a rapid dehalorespiratory process [>100 nmol of Cl- released min-1 (mg of protein)-1] (3, 21). However, this dichloroelimination reaction concomitantly produces up to 1% of the carcinogenic vinyl chloride (VC) (20), which can be removed by strain 195 only in a cometabolic process and in the presence of higher chlorinated ethenes (22). Given the fact that this organism has unravelled complex nutrient requirements (19, 23) and that it shows extreme oxygen sensitivity, its application possibilities as a pure culture for in situ 1,2-DCA detoxification seem to be restricted (3). De Wildeman et al. (10) described the isolation of Desulfitobacterium dichloroeliminans strain DCA1, a nutritionally defined dehalorespirator that converts 1,2-DCA and vicinal dichloropropane and -butane isomers into completely dechlorinated end products. Unlike those of other dehalorespirators, its degradation pathways are exclusively based on the presence of a stereoselective dehalogenase that catalyzes an energy-conserving anti dichloroelimination process. The dechlorination rate of strain DCA1 exceeded 350 nmol of Cl- min-1 (mg of total bacterial protein)-1 (10), and lab-scale experiments with 1,2-DCA-containing groundwater revealed a complete removal (24). Since strain DCA1 survives aerobic conditions for at least 24 h and as VC is not formed during 1,2-DCA dechlorination (10), the dehalorespiratory biochemistry of this strain might be a promising tool for in situ bioremediation applications. In 2003, the isolation of Dehalococcoides sp. strain BAV1 was reported, a strictly anaerobic bacterium that degrades dichloroethenes, VC, and 1,2-DCA as part of its energy metabolisms and generates stoichiometric amounts of ethene as the end product (23), yet its 1,2-DCA dechlorination rate has not yet been described (23, 25). Despite many (successful) studies on the biodegradation of PCE and TCE by means of in situ bioaugmentation (15, 26, 27), no reports are available on the anaerobic degradation of 1,2-DCA using this cleanup strategy. Therefore, an experiment was set up to find out whether D. dichloroeliminans strain DCA1 could be actively transported in a contaminated aquifer. This study describes a pilot-scale test 10.1021/es060953i CCC: $33.50

 2006 American Chemical Society Published on Web 08/08/2006

in which the distribution pattern of strain DCA1 was simulated by means of code MOCDENS3D and in which the in situ activity of this dehalorespirator was assessed for the first time.

2. Experimental Procedures 2.1. Site Description. For more than 30 years, the top layers of the phreatic aquifer at the industrial site (LVM, Tessenderlo, Belgium) contain a 1,2-DCA plume as the result of former product operations and old leakages from storage tanks and pipelines. This aquifer is composed of relatively homogeneous medium-grained sands rich in glauconite, and the water table is situated approximately 7 m below the subsurface (mbs). The naturally occurring horizontal groundwater velocities are limited to ∼1.7 m/year due to the position of the test site on the groundwater divide in the phreatic aquifer. Through the absence of indications of a decrease in the contaminant load and an increase in degradation products since 1992-1998, evidence of naturally occurring reductive dechlorination activity was thus far not detected. 2.2. Inoculum Description and Preparation. A pure culture of D. dichloroeliminans strain DCA1 was used as the inoculum. This anaerobic 1,2-DCA-dechlorinating strain was obtained from microcosms from the LVM site (temporary water-saturated layer at 1-2 mbs) and was isolated by serial transfers and dilution to extinction in a medium containing lactate and pyruvate as the available electron donors and 1,2-DCA as the sole electron acceptor (10). The production of strain DCA1 on a 10 L scale was performed as described previously (24), using the electron donor sodium lactate (40 mM) and a regular redosing of 1,2-DCA (400 µM) after depletion. To meet the demand of strain DCA1 for menaquinone as an essential cofactor for growth, vitamin K2 (1 µM) was supplemented. The anaerobic Schott 10 L bottles (GL45, Germany), in which the cultures exhibited high dechlorination activity and yielded a cell density of ∼5 × 107 bacteria/ mL, were used as inoculation material for the bioaugmentation test and transported to the site on day 1. 2.3. Plot Construction and Operation. (i) Installation. The configuration of the three collinear observation wells on the site is schematized in Figure 1. Their screens lie between 13 and 20 mbs and are situated within the sandy deposits of the Diest Formation (Upper Miocene). At a depth of 1317 mbs, the horizontal hydraulic permeability (Kh) was equal to 13.5 m/day, and between 17 and 20 mbs, it reached 7.5 m/day. LV22 and LV51 were used as monitoring wells, while LV22 also served as the injection well for the inoculum. Through the extraction of groundwater from LV52 with a pump (Grundfos, Belgium), an artificial groundwater flow was imposed on the system which resulted in the distribution of the inoculum in the direction of LV52. Every contact between oxygen and the extracted groundwater was avoided to ensure an anaerobic re-injection of the groundwater into LV22 and LV51. (ii) Inoculum Injection and the Transport Test. The bioaugmentation plot was inoculated according to Table 1. To support the inoculum exposed to oxygen stress during the injection, a yeast extract solution was added during part a. In part c, the supply of the lactate solution occurred at a rate of 1.6 L/h during six cycles per day, and one cycle comprised a time interval of no injection for 3 h and injection for 1 h. During this part, the average purely advective groundwater velocity between wells LV22 and LV51 amounted to 3.21 m/day. The transport test was terminated by shutting down all the pumps on day 29. (iii) Activity of the Inoculum. The inoculum activity was monitored during four parts (Table 1). During the injection of the appropriate electron donor and buffer solution at the beginning of every part, an artificial groundwater flow was imposed once more on the system by extracting groundwater

FIGURE 1. Collinear observation wells. Permeable screens (filters) are represented by the black parts. The water table is situated approximately 7 m below the subsurface. LV52 was used as the extraction well; LV22 and LV51 were used as monitoring wells, and LV22 also served as the injection well for the inoculum. from LV52. After these supplies, the pumps were shut down and the groundwater became almost stagnant as a result of the limited naturally occurring horizontal groundwater velocity. (iv) Geochemical Monitoring. More information regarding the groundwater sampling strategy is found in the Supporting Information. (v) GC Analysis and TOC Content. The procedures for the determination of VOCl’s, ethane, ethene, methane, fatty acids, and the TOC content are described in the Supporting Information. (vi) Microbial Monitoring. Groundwater samples (2 × 1.5 L) were collected in airtight plastic sterile recipients according to the geochemical sampling procedure. To collect microbial biomass, filtration of the samples was performed using a 500 mL polysulfon filter holder (Nalgene, Rochester, NY) with a 0.2 µm pore filtration membrane (type GS, Millipore, Brussels, Belgium). The filtration paper was stored in a 14 mL polypropylene round-bottomed tube (Greiner bio-one, Kremsmuenster, Austria), and the DNA from the cell pellet was extracted following the protocol described by Boon et al. (28). The real-time amplification of the 16S rRNA gene by means of species-specific primers and the quantitative evaluation of D. dichloroeliminans strain DCA1 performed by means of an ABI PRISM SDS 7000 Sequence Detection System (PE Applied Biosystems, Nieuwerkerk a/d IJssel, The Netherlands) have been described by Van Raemdonck et al. (29). A serial dilution of known amounts of strain DCA1 was spiked in 1.5 L of groundwater (Sint Niklaas, Belgium) to determine the lower limit of quantification (LLQ) and the lower limit of detection (LLD) according to the method described by Seurinck et al. (30). The presence of strain DCA1 VOL. 40, NO. 17, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

5545

5546 9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 17, 2006

TABLE 1. Inoculation and Supplementation Scheme during the Transport and Activity Test transport part

LV22

part b: day 1, 1-8 h

part c: day 1 (8 h) to day 29

activity part part I: day 35 to day 98

LV22 groundwater injection (day 35) (2 h) at 1 m3/h, together with 7 L of sodium lactate (50 vol %) (Fluka) and 3.6 L of NaHCO3 buffer (1 M) (VWR)

part II: day 99 to day 105

part III: day 106 to day 168

part IV: day 169 to day 212

LV52

groundwater injection (1 h) at 2 m3/h, together with a 50 L inoculum (5 × 107 DCA1 bacteria/mL), a 5 L sodium lactate solution (40 mM) (Fluka), 2 L of yeast extract (50 g/L) (Oxoid), and 4.8 L of NaHCO3 buffer (1 M) (VWR) groundwater injection (7 h) at 2 m3/h, together with a 28 L sodium lactate solution (40 mM) (Fluka), 2 L of yeast extract (50 g/L) (Oxoid), and 4.8 L of NaHCO3 buffer (1 M) (VWR) groundwater injection (continuously) at 0.2 m3/h together with a 280 L sodium lactate solution (10 mM) (Fluka)

part a: day 1, 0-1 h

groundwater injection (day 106) (2 h) at 1 m3/h, together with 4 L of lactate (80 vol %) (Galactic), 3 L of NaOH (50 vol %), and 7 L of NaHCO3 buffer (1 M) (VWR)

LV51 groundwater injection (day 35) (1.5 h) at 1 m3/h, together with 5 L of sodium lactate (50 vol %) (Fluka) and 2.8 L of NaHCO3 buffer (1 M) (VWR) groundwater injection (day 99) (2 h) at 1 m3/h, together with 3.75 L of lactate (80 vol %) (Galactic) and 7 L of NaHCO3 buffer (1 M) (VWR) groundwater injection (day 106) (2 h) at 1 m3/h, together with 10 L of NaHCO3 buffer (1 M) (VWR) groundwater injection (day 169) (1.5 h) at 1 m3/h, together with 6 L of sodium lactate (60 vol %) (Galactic) and 4 L of NaHCO3 buffer (1 M) (VWR)

groundwater extraction (1 h) at 2.6 m3/h

groundwater extraction (7 h) at 2.6 m3/h

groundwater extraction (continuously) at 5 m3/h

LV52 groundwater extraction (day 35) (2 + 1.5 h) at 1 m3/h groundwater extraction (day 99) (2 + 2 h) at 2 m3/h groundwater extraction (day 106) (2 + 2 h) at 1 m3/h

groundwater extraction (day 169) (1.5 + 1.5 h) at 1 m3/h

TABLE 2. QPCR Detection of Strain DCA1 during the Transport Test and the Activity Testa LV22 day

a

0 3 5 7 9 13 17 21 30 34 43 55 71 85 98 101 106 113 120 129 134 141 225

NDb

2.28 × 108 7.63 × 107 7.75 × 107 1.40 × 108 1.83 × 107 1.48 × 107 3.88 × 106 3.35 × 106 1.16 × 107 2.51 × 107 DNQc 9.90 × 105 DNQc 1.95 × 106 3.46 × 106 DNQc DNQc DNQc DNQc

b 6/6 6/6 3/6 6/6 6/6 6/6 6/6 6/6 6/6 6/6 6/6 6/6 6/6 6/6 3/3 4/6 6/6 5/6 1/6 3/6

LV51 c 1.92 × 107 6.45 × 106 8.89 × 106 1.98 × 107 8.55 × 105 4.00 × 106 1.98 × 105 1.23 × 105 1.59 × 106 9.42 × 105 4.81 × 104 1.56 × 105 5.54 × 105

a NDb DNQc 2.65 × 106 DNQc DNQc DNQc DNQc DNQc DNQc DNQc DNQc DNQc DNQc DNQc DNQc DNQc DNQc DNQc NDb NDb NDb

b 6/6 3/6 6/6 2/6 3/6 2/6 5/6 2/6 6/6 6/6 6/6 3/6 3/3 3/3 3/3 1/6 3/6 1/6 6/6 6/6 6/6

LV52 c

a NDb

8.71 × 105

b

NDb NDb DNQc DNQc DNQc DNQc -

6/6 6/6 6/6 4/6 3/6 5/6 5/6 -

-

-

NDb

6/6

a

Column a shows the average amount of strain DCA1 (number of cells per liter of groundwater), column b the number of positive samples out of the total number of samples, and column c the standard error (number of cells per liter of groundwater). b No detection. c Detection but no quantification.

was always checked by the profile of the melting curve obtained during the quantitative real-time PCR (qPCR) analysis. A LLQ of 5 × 105 bacterial cells/L of groundwater and a LLD of 5 × 10 bacterial cells/L of groundwater were obtained. All measurements were taken in triplicate. 2.4. Groundwater Modeling. To gain an insight into the distribution of strain DCA1, a computer simulation was performed using numerical density-dependent three-dimensional solute transport model MOCDENS3D. Transport equations for advection, dispersion (both mechanical dispersions and diffusion), external sources (mixing or dilution from fluid sources), and linear sorption are incorporated in the groundwater flow model based on the three-dimensional method-of-characteristics groundwater flow and transport model MOC3D (31-33). The evolution of the solute concentration (i.e., strain DCA1) is represented by eq 1

Rf ×

∂C0 (C′ - C0)Rw ) -∇(C0V h ) + ∇[D h ∇(C0)] + (1) ∂t ne

where C0 is the concentration of the solutes in the aqueous phase [ML-3], V h is the average interstitial velocity vector [LT-1] (V h )q j /ne (2), where q j is the average Darcian flow velocity [LT-1] and ne the effective porosity [L3L-3]], D h is the secondorder hydrodynamic dispersion tensor [L2T-1], C′ is the concentration of the injected water [ML-3], Rw is an external source (recharge and pumping) [LT-1], and Rf is the retardation factor [Rf ) 1 + FbKd/ne (3), where Fb is the bulk density [ML-3] and Kd the distribution coefficient of the solute between the solid and the aqueous phase [L3M-1]]. More methodological details regarding MOCDENS3D are provided in the Supporting Information. In the numerical model, the longitudinal dispersivity was set at 0.08 m, the transverse dispersivity at 0.008 m, the effective porosity at 0.37, and the bulk density at 1.67 kg/L. The initial concentration of bacteria in the aquifer was set at 0%. A 100% concentration injection of bacteria (2.5 × 1012 bacteria/16 m3 of groundwater) was assumed.

3. Results 3.1. Goal. During the transport test, the distribution of strain DCA1 through the aquifer was realized by pushing the cells from LV22 in the direction of LV52. As the naturally occurring horizontal groundwater velocity is limited to 1.7 m/year, groundwater had to be extracted from LV52 at a flow of 5 m3/h for 30 days to impose a pulling flow to the inoculum. Sodium lactate was continuously added to maintain reductive redox conditions in the aquifer and to support the injected inoculum. At the end of the transport test, all pumps were shut down and the distributed cells of strain DCA1 could start to degrade 1,2-DCA in the zone wherein the spreading was realized. During the activity test, a re-injection of extracted groundwater was performed four times (on days 35, 99, 106, and 169) during a few hours to effect the supplementation of additional amounts of electron donor to strain DCA1. Through MOCDENS3D, the determination of the required supplementation flow was performed to provide enough amounts of the electron donor to the cells of strain DCA1 distributed around the monitoring wells. Hence, since the groundwater was almost stagnant during the majort part of the activity test, a closed-loop recirculation system was not present during the activity test. 3.2. Results from the Transport Test. To predict the moment of detection of strain DCA1 in LV51, the transport test was simulated by means of MOCDENS3D. Initially, an Rf value of 14 was used in eq 1, since in this way the geometric average for the bacterial distribution coefficient (Kd) of 2.79 L/kg was incorporated. On the basis of a modeling study using the flow rates mentioned in Table 1, detectable amounts of strain DCA1 should arrive in LV51 after ∼7-8 days. By applying a safety factor of 4, the transport test covered 1 month to ensure the detection of strain DCA1 in LV51. The concentrations of strain DCA1 determined by means of qPCR are presented in Table 2. The first detection of the 16S rRNA gene of strain DCA1 in LV51 and LV52 took place on days 3 and 7, respectively, whereas the quantification of strain DCA1 in LV51 was only feasible on day 5. Variations of the retardation parameter in MOCDENS3D revealed that a VOL. 40, NO. 17, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

5547

FIGURE 2. Simulation of the evolution of the spreading of (A) a conservative tracer (Rf ) 1) and (B) D. dichloroeliminans strain DCA1 (Rf ) 9) 17-18 m below the subsurface. View after 5 days. Concentration contour lines (in percent) are depicted by bold lines and numerically indicated in the rectangular labels. The hydraulic head contour lines are represented every 0.25 m and represented with circular labels. The arrows in the picture represent the horizontal groundwater velocity vectors. maximal fit between the molecular analyses and the simulated distribution patterns was obtained for an Rf value of ∼9. This value was consistent with the maximal concentration of strain DCA1 in LV51 5 days after the start of the test and its appearance time in LV52. The simulation of the distribution pattern on day 5 at a depth of 17-18 mbs is depicted in Figure 2. To evaluate the impact of the Rf value, the comparison was made with the distribution of a conservative tracer (such as Cl-), to which an Rf value of 1 can be attributed. On the basis of the simulation mentioned above and eq 2, a distribution coefficient (Kd) of ∼1.45 L/kg can be assigned to strain DCA1. At the end of the transport test, the distribution was most extensive around well LV22, with a 5548

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 17, 2006

radius of ∼0.9 m. This distribution is directly related to the stagnation point at a distance of 0.9 m of LV22 as a consequence of the imposed groundwater flow through the pulsing injection flow in LV22 and the pulling extraction flow out of LV52 (Figure 2). No significant biodegradation of 1,2-DCA was observed during this test. The contaminant concentration fluctuated strongly in every well, mainly between 400 and 1000 µM. Almost invariable ethene measurements with an average concentration of 21.0 ( 6.2 µM confirmed that no enhanced degradation of 1,2-DCA occurred. The average acetic acid concentration amounted to 4.1 ( 2.9 mg/L, and most of the time, no propionic acid could be detected. Furthermore, the

FIGURE 3. 1,2-DCA (b), ethene (O), and VC (1) concentration (micromolar) in the groundwater of (A) LV22 and (B) LV51 during the activity test. The electron donor and buffer supplementations are indicated with vertical arrows. TOC content (milligram of C per liter) fluctuated between 3 and 29 mg of C/L, confirming the low availability of substrate for the cells. A stable pH profile was observed in every well with an average pH value of ∼5.7 ( 0.4. The redox potential (Eh) and the dissolved oxygen concentration (DO) measurements indicated that micro-aerophilic to moderate reducing conditions were present (in the range of 30 to -100 mV). 3.3. Results from the Activity Test. The detection of the 16S rRNA gene of strain DCA1 during the activity test is presented in Table 2. At the start of the activity test, a strong increase in the level of 1,2-DCA was observed in LV22 and LV51 (Figure 3). After this increased profile was sustained for ∼1 week, a sharp drop occurred. Fifty days later, the 1,2DCA concentration amounted to 15.2 and 57.6 µM for LV22 and LV51, respectively. Quickly after the second electron donor supplementation in LV51 and LV22, the increased profile was observed again, although the augmentation for LV22 was not as pronounced as that with the one after the first electron donor supply. Soon, the contaminant decrease was once more conspicuous with the detection of 26.3 µM 1,2-DCA in LV51 on day 129 and 0.9 µM 1,2-DCA in LV22 on day 155. In contradistinction with the preceding electron donor supplementations, a moderate contaminant increase was monitored in LV51 after the final electron donor addition. As indicated in Figure 3, the 1,2-DCA concentration in LV22 remained quite low during the rest of the activity test (11.7 ( 4.3 µM), whereas no considerable 1,2-DCA decrease was observed in LV51. The increased ethene production at the start and after the second electron donor supply unambiguously confirmed the stimulated activity of strain DCA1 in

LV22 and LV51 (Figure 3). However, the final supplementation had no significant impact on ethene production in LV51. During the activity test, the 1,2-DCA concentration in LV52 fluctuated around the initially measured concentration of the transport test, except after the first electron donor supply when an increased profile was also observed (Figure 1 of the Supporting Information). The absence of significant activity of strain DCA1 in this well was confirmed by the ethene profile (Figure 1 of the Supporting Information). During the activity test, an almost invariable methane and ethane profile was monitored. The average methane concentrations amounted to 21.8 ( 11.5, 14.9 ( 8.1, and 30.2 ( 20.0 µM and the average ethane concentrations to 0.28 ( 0.15, 0.18 ( 0.08, and 0.14 ( 0.04 µM in LV22, LV51, and LV52, respectively. An upward or downward trend was missing in the VC concentrations for the greatest part of the activity test, although an increase was measured after every electron donor supplementation (Figure 3, and Figure 1 of the Supporting Information). The average concentration measured 30.0 ( 15.9 and 20.4 ( 13.0 µM in LV22 and LV51, respectively. On day 36, the TOC content in LV22 and LV51 amounted to 914 and 710 mg of C/L, respectively. The initial injected amount of sodium lactate (Table 2) represented ∼74 and ∼91% of the TOC, respectively. The fatty acid concentration (FAC ) acetic acid concentration + propionic acid concentration) amounted to 0.1 and 0.7% of the TOC, respectively. Very soon, a gradual decrease in the TOC content and an increase in FA production took place. At the end of the first part, ∼741 and ∼694 mg of C/L was consumed and the FAC totally represented the TOC content. After the second electron donor supplementation, a similar increase of ∼600 mg of C/L was seen in LV22 and LV51 [763 mg of C/L (85% FAC) and 661 mg of C/L (77% FAC), respectively]; however, a faster TOC depletion took place in LV51 as compared with that in LV22. Indeed, a gradual decrease was monitored in LV22 until the end of the activity test [day 168, 354 mg of C/L (90% FAC); day 212, 216 mg of C/L], whereas the TOC content in LV51 already decreased to 54 mg of C/L (100% FAC) after 1 week. An extra electron donor supplementation was essential, since the 1,2-DCA concentration in LV51 gradually increased from day 141 on. This resulted in the detection of 476 mg of C/L on day 176. The TOC content of 44 mg of C/L on day 197 showed that a fast depletion took place again. Throughout the activity test, the redox potential evolved from moderate to stronger anaerobic conditions and a gradual acidification of the groundwater occurred. At the end of the activity test, the average pH amounted to 4.8 ( 0.4, 3.9 ( 0.5, and 4.6 ( 0.5 in LV22, LV51, and LV52, respectively. The time course of the pH and the redox potential of the groundwater are described in detail in the Supporting Information.

4. Discussion The results from the transport test combined with modeling experiments in MOCDENS3D afforded a deeper understanding of the distribution of strain DCA1 in the aquifer (Figure 2). The 16S rRNA gene of strain DCA1 was already detected in LV51 and LV52 after 1 week, and MOCDENS3D simulations revealed a relatively low Rf value of ∼9. Rudimentary site information (ne ) 0.37; Fb ) 1.67 kg/L) applied in eq 2 resulted in a distribution coefficient (Kd) of ∼1.45 L/kg for strain DCA1 in the particular aquifer. However, the Kd value is mainly determined by microbial properties such as the bacterial size and shape and the cell wall structure. The Gram-positive DCA1 cells have been described as curved rods 0.5-0.7 µm in diameter and 2-5 µm in length (10). Lindqvist and Bengtsson proposed that this distribution coefficient can vary between 0.2 and 39 L/kg (34). A more accurate determination of the Kd parameter can be realized VOL. 40, NO. 17, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

5549

through the determination of cell surface characteristics such as hydrophobicity [e.g., by the BATH test (35)]. The fast movement of the bacterial front through the aquifer also results from the small fraction of large pores contributing to the porosity of the homogeneous medium-grained sands and the high permeability of the soil matrix (Kh ) 7.5-13.5 m/day). The input of processes such as advective transport, dispersion, mixing from fluid sources, and linear sorption throughout MOCDENS3D enables us to gain a more precise idea of the bacterial distribution, since the vast majority of a microbial tracer is trapped by filtration or adsorption or is dispersed in the subsurface between the source and the vector (36). Nevertheless, some improvements to the program such as the input of microbial processes (growth, decay, chemotaxis, etc.) are required to obtain a detailed representation of the time-dependent distribution of strain DCA1. Approximately 3.1% of the initial injected amount of strain DCA1 should be detected in LV51 on day 5 (Figure 2). In practice, however, only 1.7% was monitored and many more DCA1 cells were counted in LV22 in comparison with the value predicted by MOCDENS3D (Table 2). These low numbers are acceptable, since the loss of bacteria due to interaction with grain surfaces in sandy aquifer sediments can be at least one log unit per 10 in. (25.4 cm) of travel (37). However, the difference between the modeled and measured value is considerable and might result from the different depth position of the screens of LV52 and LV51 and, to a lesser extent, from the loss of microbial mass during the sampling strategy. The large population density measured around LV22 (Table 2), presumably resulting from the inoculum accumulation in the screen, can be explained by microbial attachment processes and a hindered mass transfer from the screen of LV22 to the surrounding groundwater and requires an individual modeling (38). In conclusion, it is clear that MOCDENS3D adds significant value to the results obtained in the pilot test, since it resulted in a better understanding of the influence of the imposed flow path on the distribution of strain DCA1 through the aquifer. This is valuable since the success of bioaugmentaion depends as much on the ability of the introduced microorganism to reach the contaminants as on its in situ survival. The limited retardation of strain DCA1 is significantly important since it allows us to create a biozone or biocurtain in a reasonable time interval. The fluctuating 1,2-DCA concentration during the transport test showed that no homogeneous contamination was present in the water surrounding the monitoring wells. The absence of any degradation activity of strain DCA1 was probably related to the insufficient anaerobic conditions and to the lack of sufficient available electron donor amounts, which is reflected in the low TOC content and the absence of degradation products of sodium lactate. The microaerophilic conditions can be explained by the inflow of groundwater with a slightly increased Eh during the sodium lactate injection, although every contact between oxygen and the extracted groundwater had been avoided. It has been demonstrated that in laboratory experiments anoxic conditions (Eh < -180 mV) are essential for the dechlorination activity of strain DCA1 (10). The decrease in the 1,2-DCA contamination load during the activity test was very distinct and incontestable the result of the metabolic degradation activity of strain DCA1 in LV22 (99.9%) and LV51 (98.3%), confirmed by the nearly equimolar production of ethene. Moreover, the almost invariable methane measurements point to the absence of any methanogenic activity. Additionally, the VC measurements corroborated the absence of stimulated 1,2-DCA-degrading microorganisms which coproduce vinyl chloride. Finally, prior to this in situ pilot test, groundwater samples from the contamination plume of the test site were incubated in the 5550

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 17, 2006

lab at 28 °C with 1 mM 1,2-DCA and 15 mM sodium lactate. No 1,2-DCA degradation and ethene production could be measured, which indicates that the indigenous microbial populations failed to degrade 1,2-DCA, despite the electron donor supplementation. The increase in the contaminant load after every electron donor supplementation resulted partially from the use of the extraction water from LV52. However, since the 1,2-DCA concentration in LV52 amounted to 858.9, 535.6, and 374.9 µM on days 36, 101, and 113, respectively, an additional explanation is needed. In all probability, the electron donor supplementation gave rise to an enhanced mass transfer of 1,2-DCA from the nonaqueous phase to the aqueous phase. Similar results were noted in a field study in which sodium lactate injection resulted in an increase of a factor of 23 in TCE concentrations due to a decreased interfacial tension with nonaqueous TCE (39). After the last electron donor supplementation, no 1,2DCA decrease was observed in LV51, although a sufficient electron donor amount was provided according to the TOC analyses. Additionally, strain DCA1 could not be detected anymore in LV51 by means of qPCR from day 134 on (Table 2). These observations imply that the stimulation of the remaining populations for the 1,2-DCA degradation could not be established. Different hypotheses can account for the lack of biodegradation activity during the last part. First, it is clear that the insufficient amount of strain DCA1 cells can be considered as the principal cause, although these cells were not outcompeted by methanogenic populations or sulfate-reducing bacteria. Indeed, almost invariable methane concentrations were measured, and the sulfate concentrations at the beginning and end of the pilot test were equal to 103 and 95 mg/L, respectively. The measurements of other VOCl’s (TCE, cis-DCE, trans-DCE, VC, and 1,1-DCA) exclude the possibility that strain DCA1 was overruled by other stimulated dechlorinating populations. For these reasons, it is more likely to assume that the repeated injection pulses of the electron donor supplementations caused a radial displacement of strain DCA1 in the direction of the other wells, which resulted in the continual increasing dilution of the cells present in LV51. In this way, insufficient strain DCA1 cells were present to derive energy from the degradation of 1,2-DCA for the creation of an ecological niche in the microbial community. A second reason for the lack of biodegradation capacity in LV51 during the last part of the activity test can be the accumulation of the degradation product HCl. In a previous study, strain DCA1 was still active at pH 6.0, in addition to its optimal pH range of 7.2-7.8 (24). The results from this test showed the robustness of strain DCA1, since its degradation activity was sustained at very low pH values of ∼5.4. However, the gradual acidification in LV51 to pH 3.9 must have contributed to the stagnant degradation activity. It is important to stress that the experimental setup of this study is different from most other bioaugmentation experiments on sites contaminated with chlorinated ethenes (26, 40). Since no continuous inoculum recirculation has been carried out, it was demonstrated that 1,2-DCA degradation in LV51 was not the result of transport of decontaminated water from the bioreactive zone created around LV22. In this study, only groundwater was analyzed since 1,2-DCA is characterized by a considerable solubility in water (8 g/L) as compared with other chlorinated solvents. However, further research in which the full degradation capacity of strain DCA1 is evaluated by the addition of soil samples to the analytical procedures and in which other 1,2-DCAdegrading microorganisms [such as Dehalococcoides sp. strain BAV1 (23)] are included to provide clues about their environmental performance in the presence of other (dechlorinating) populations must be performed.

This research has resulted for the first time in a detailed characterization of the transport and activity of strain DCA1 during bioaugmentation of 1,2-DCA-contaminated groundwater, pointing to the conclusion that this strain offers promising perspectives for larger-scale cleanup processes of anaerobic aquifers due to its limited retardation and its in situ ability to successfully degrade 1,2-DCA.

Acknowledgments We acknowledge the cooperation with LVM (Tessenderlo, Belgium), Avecom NV (Beernem, Belgium), and ESA BVBA (Hasselt, Belgium). The research was funded by a Ph.D. grant of the Institute for the Promotion of Innovation through Science and Technology in Flanders (IWT-Vlaanderen) and funded by LVM. Special thanks to Nico Boon, Tom Van de Wiele, Wim De Windt, Birgit Mertens and Matthias D’hooghe for a critical review of the manuscript.

Supporting Information Available Description of the geochemical monitoring strategy, GC analysis, TOC content determination, methodological details regarding groundwater modeling, results of the pH and redox potential analyses during the activity test, and a figure of the groundwater analyses for 1,2-DCA, VC, and ethene in LV52 (Figure 1). This material is available free of charge via the Internet at http://pubs.acs.org.

Literature Cited (1) Field, J. A.; Sierra-Alvarez, R. Biodegradability of chlorinated solvents and related chlorinated aliphatic compounds. Rev. Environ. Sci. Biotechnol. 2004, 3, 185-254. (2) Hughes, K.; Meek, M. E.; Caldwell, I. 1,2-Dichloroethane: Evaluation of risks and health from environmental exposure in Canada. J. Environ. Sci. Health, Part C: Environ. Health Sci. 1994, 12, 293-303. (3) De Wildeman, S.; Verstraete, W. The quest for microbial reductive dechlorination of C2 to C4 chloroalkanes is warranted. Appl. Microbiol. Biotechnol. 2003, 61, 94-102. (4) Kassenga, G.; Pardue, J. H.; Moe, W. M.; Bowman, K. S. Hydrogen tresholds as indicators of dehalorespiration in constructed treatment wetlands. Environ. Sci. Technol. 2004, 38, 1024-1030. (5) Bosma, T. N. P.; van Aalst-van Leeuwen, M. A.; Gerritse, J.; van Heiningen, E.; Taat, J.; Pruijn, M. Intrinsic dechlorination of 1,2-dichloroethane at an industrial site. In Natural attenuation; Wickramanayake, G. B., Hinchee, R. E., Eds.; Battelle Press: Columbus, OH, 1998; Vol. 3, pp 7-12. (6) Nobre, R. C. M.; Nobre, M. M. M. Natural attenuation of chlorinated organics in a shallow sand aquifer. J. Hazard. Mater. 2004, 110, 129-137. (7) Lee, M. D.; Sehayek, L.; Sleep, B. E.; Vandell, T. D. Investigation and remediation of a 1,2-dichloroethane spill part II: Documentation of natural attenuation. Ground Water Monit. Rev. 1999, 19, 82-88. (8) Job, N.; Heinrichs, B.; Ferauche, F.; Noville, F.; Marien, J.; Pirard, J. P. Hydrodechlorination of 1,2-dichloroethane on Pd-Ag catalysts supported on tailored texture carbon xerogels. Catal. Today 2005, 102, 234-241. (9) Lambert, S.; Ferauche, F.; Brasseur, A.; Pirard, J. P.; Heinrichs, B. Pd-Ag/SiO2 and Pd-Cu/SiO2 cogelled xerogel catalysts for selective hydrodechlorination of 1,2-dichloroethane into ethylene. Catal. Today 2005, 100, 283-289. (10) De Wildeman, S.; Diekert, G.; Van Langenhove, H.; Verstraete, W. Stereoselective microbial dehalorespiration with vicinal dichlorinated alkanes. Appl. Environ. Microbiol. 2003, 69, 56435647. (11) Song, H.; Carraway, E. R. Reduction of chlorinated ethanes by nanosized zerovalent iron: Kinetics, pathways, and effects of reaction conditions. Environ. Sci. Technol. 2005, 39, 6237-6245. (12) Dries, J.; Bastiaens, L.; Springael, D.; Agathos, S. N.; Diels, L. Combined removal of chlorinated ethenes and heavy metals by zerovalent iron in batch and continuous flow column systems. Environ. Sci. Technol. 2005, 39, 8460-8465. (13) He, F.; Zhao, D. Y. Preparation and characterization of a new class of starch-stabilized bimetallic nanoparticles for degradation of chlorinated hydrocarbons in water. Environ. Sci. Technol. 2005, 39, 3314-3320.

(14) Nutt, M. O.; Hughes, J. B.; Wong, M. S. Designing Pd-on-Au bimetallic nanoparticle catalysts for trichloroethene hydrodechlorination. Environ. Sci. Technol. 2005, 39, 1346-1353. (15) Lendvay, J. M.; Lo¨ffler, F. E.; Dollhopf, M.; Aiello, M. R.; Daniels, G.; Fathepure, B. Z.; Gebhard, M.; Heine, R.; Helton, R.; Shi, J.; Krajmalnik-Brown, R.; Major, C. L.; Barcelona, M. J.; Petrovskis, E. A.; Hickey, R.; Tiedje, J. M.; Adriaens, P. Bioreactive barriers: Bioaugmentation and biostimulation for chlorinated solvent remediation. Environ. Sci. Technol. 2003, 37, 1422-1431. (16) Macbeth, T. W.; Cummings, D. E.; Spring, S.; Petzke, L. M.; Sorenson, K. S. Molecular characterization of a dechlorinating community resulting from in situ biostimulation in a trichloroethene-contaminated deep, fractured basalt aquifer and comparison to a derivative laboratory culture. Appl. Environ. Microbiol. 2004, 70, 7329-7341. (17) Dyer, M.; van Heiningen, E.; Gerritse, J. A field trial for in-situ bioremediation of 1,2-DCA. Eng. Geol. 2003, 70, 315-320. (18) Maymo-Gatell, X.; Anguish, T.; Zinder, S. H. Reductive dechlorination of chlorinated ethenes and 1,2-dichloroethane by “Dehalococcoides ethenogenes” strain 195. Appl. Environ. Microbiol. 1999, 65, 3108-3113. (19) Maymo-Gatell, X.; Chien, Y. T.; Gossett, J. M.; Zinder, S. H. Isolation of a bacterium that reductively dechlorinates tetrachloroethene to ethene. Science 1997, 276, 1568-1571. (20) Magnuson, J. K.; Romine, M. F.; Burris, D. R.; Kingsley, M. T. Trichloroethene reductive dehalogenase from Dehalococcoides ethenogenes: Sequence of tceA and substrate range characterization. Appl. Environ. Microbiol. 2000, 66, 5141-5147. (21) Holliger, C.; Wohlfarth, G.; Diekert, G. Reductive dechlorination in the energy metabolism of anaerobic bacteria. FEMS Microbiol. Rev. 1998, 22, 383-398. (22) Maymo-Gatell, X.; Nijenhuis, I.; Zinder, S. H. Reductive dechlorination of cis-1,2-dichloroethene and vinyl chloride by “Dehalococcoides ethenogenes”. Environ. Sci. Technol. 2001, 35, 516521. (23) He, J. Z.; Ritalahti, K. M.; Yang, K. L.; Koenigsberg, S. S.; Lo¨ffler, F. E. Detoxification of vinyl chloride to ethene coupled to growth of an anaerobic bacterium. Nature 2003, 424, 62-65. (24) De Wildeman, S.; Linthout, G.; Van Langenhove, H.; Verstraete, W. Complete lab-scale detoxification of groundwater containing 1,2-dichloroethane. Appl. Microbiol. Biotechnol. 2004, 63, 609612. (25) Krajmalnik-Brown, R.; Ho¨lscher, T.; Thomson, I. N.; Saunders, F. M.; Ritalahti, K. M.; Lo¨ffler, F. E. Genetic identification of a putative vinyl chloride reductase in Dehalococcoides sp. strain BAV1. Appl. Environ. Microbiol. 2004, 70, 6347-6351. (26) Major, D. W.; McMaster, M. L.; Cox, E. E.; Edwards, E. A.; Dworatzek, S. M.; Hendrickson, E. R.; Starr, M. G.; Payne, J. A.; Buonamici, L. W. Field demonstration of successful bioaugmentation to achieve dechlorination of tetrachloroethene to ethene. Environ. Sci. Technol. 2002, 36, 5106-5116. (27) Morrill, P. L.; Lacrampe-Couloume, G.; Slater, G. F.; Sleep, B. E.; Edwards, E. A.; McMaster, M. L.; Major, D. W.; Lollar, B. S. Quantifying chlorinated ethene degradation during reductive dechlorination at Kelly AFB using stable carbon isotopes. J. Contam. Hydrol. 2005, 76, 279-293. (28) Boon, N.; Goris, J.; De Vos, P.; Verstraete, W.; Top, E. M. Bioaugmentation of activated sludge by an indigenous 3-chloroaniline-degrading Comamonas testosteroni strain, I2gfp. Appl. Environ. Microbiol. 2000, 66, 2906-2913. (29) Van Raemdonck, H.; Maes, A.; Ossieur, W.; Verthe´, K.; Vercauteren, T.; Verstraete, W.; Boon, N. Real Time PCR quantification in groundwater of the dehalorespriring Desulfitobacterium dichloroeliminans strain DCA1. J. Microbiol. Methods 2006, in press (doi: 10.101b/j.minet.2006.04.001). (30) Seurinck, S.; Defoirdt, T.; Verstraete, W.; Siciliano, S. D. Detection and quantification of the human-specific HF183 Bacteroides 16S rRNA genetic marker with real-time PCR for assessment of human faecal pollution in freshwater. Environ. Microbiol. 2005, 7, 249-259. (31) Oude Essink, G. H. P. Salt water intrusion in a three-dimensional groundwater system in The Netherlands: A numerical study. Transp. Porous Media 2001, 43, 137-158. (32) Konikow, L. F.; Goode, D. J.; Hornberger, G. Z. A threedimensional characteristics solute-transport model (MOC3D). U.S. Geological Survey Water-Resources Investigations Report; U.S. Geological Survey: 1996; 96-4267, pp 87. (33) Lebbe, L.; Oude Essink, G. H. P. MOCDENSITY/MOCDENS3Dcode. In Seawater intrusion in coastal aquifers, concepts, methods and practices; Bear, J., Cheng, A. H. D., Sorek, S., Quazar, D., Herrera, I., Eds.; Kluwer Academic Publishers: Dordrecht, The Netherlands, 1999; pp 434-439. VOL. 40, NO. 17, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

5551

(34) Lindqvist, R.; Bengtsson, G. Dispersal Dynamics of Groundwater Bacteria. Microbial Ecol. 1991, 21, 49-72. (35) Rosenberg, M.; Gutmik, D.; Rosenberg, E. Adherence of bacteria to hydrocarbons: A simple method for measuring cell-surface hydrophobicity. FEMS Microbiol. Lett. 1980, 91, 29-33. (36) Taylor, R.; Cronin, A.; Pedley, S.; Barker, J.; Atkinson, T. The implications of groundwater velocity variations on microbial transport and wellhead protection: Review of field evidence. FEMS Microbiol. Ecol. 2004, 49, 17-26. (37) Harvey, R. W.; Harms, H. Transport of microorganisms in the terrestrial subsurface: In situ and laboratory methods. In Manual of Environmental Microbiology, 2nd ed.; Hurst, C. J., Knudsen, G. R., McInerney, M. J., Stetzenback, L. D., Crawford, R. L., Eds.; ASM Press: Washington, DC, 2001; pp 753-776.

5552

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 17, 2006

(38) MacDonald, T. R.; Kitanidis, P. K.; McCarty, P. L.; Roberts, P. V. Mass-transfer limitations for macroscale bioremediation modeling and implications on aquifer clogging. Ground Water 1999, 37, 523-531. (39) Sorenson, K. S. Enhanced bioremediation for treatment of chlorinated solvent residual source areas. ACS Symp. Ser. 2003, 837, 119-131. (40) Ellis, D. E.; Lutz, E. J.; Odom, J. M.; Buchanan, R. J.; Bartlett, C. L.; Lee, M. D.; Harkness, M. R.; Deweerd, K. A. Bioaugmentation for accelerated in situ anaerobic bioremediation. Environ. Sci. Technol. 2000, 34, 2254-2260.

Received for review April 20, 2006. Revised manuscript received June 22, 2006. Accepted June 23, 2006. ES060953I