Inhibition of Geobacter Dechlorinators at Elevated Trichloroethene

Jan 2, 2013 - ... in predicting the reactivation of the dechlorination reaction in treatments in which the inhibiting TCE concentration was lowered af...
0 downloads 0 Views 1MB Size
Download PDF
Article pubs.acs.org/est

Inhibition of Geobacter Dechlorinators at Elevated Trichloroethene Concentrations Is Explained by a Reduced Activity Rather than by an Enhanced Cell Decay Jo Philips,†,§ Pieter Jan Haest,‡ Dirk Springael,† and Erik Smolders*,† †

Division of Soil and Water Management, Department of Earth and Environmental Sciences, KU Leuven (University of Leuven), Kasteelpark Arenberg 20, B-3001 Heverlee Belgium ‡ Flemish Institute for Technological Research (VITO), Boerentang 200, B-2400 Mol Belgium S Supporting Information *

ABSTRACT: Microbial dechlorination of trichloroethene (TCE) is inhibited at elevated TCE concentrations. A batch experiment and modeling analysis were performed to examine whether this self-inhibition is related to an enhanced cell decay or a reduced dechlorination activity at increasing TCE concentrations. The batch experiment combined four different initial TCE concentrations (1.4−3.0 mM) and three different inoculation densities (4.0 × 105 to 4.0 × 107 Geobacter cells·mL−1). Chlorinated ethene concentrations and Geobacter 16S rRNA gene copy numbers were measured. The time required for complete conversion of TCE to cis-DCE increased with increasing initial TCE concentration and decreasing inoculation density. Both an enhanced decay and a reduced activity model fitted the experimental results well, although the reduced activity model better described the lag phase and microbial decay in some treatments. In addition, the reduced activity model succeeded in predicting the reactivation of the dechlorination reaction in treatments in which the inhibiting TCE concentration was lowered after 80 days. In contrast, the enhanced decay model predicted a Geobacter cell density that was too low to allow recovery for these treatments. Conclusively, our results suggest that TCE self-inhibition is related to a reduced dechlorination activity rather than to an enhanced cell decay at elevated TCE concentrations.

1. INTRODUCTION Groundwater contaminations of trichloroethene (TCE) pose serious threats to human health since this chemical is a suspected mutagen and carcinogen. Remediation of such contaminations is complicated, because TCE forms dense nonaqueous phase liquids (DNAPL) in the subsoil. One of the proposed strategies to remediate DNAPL source zones is bioremediation through microbial reductive dechlorination. Several anaerobic bacteria are known to gain energy from the reduction of TCE to cis-dichloroethene (cis-DCE).1 In the vicinity of a DNAPL, this dechlorination reaction can enhance the DNAPL dissolution rate and, as such, reduce the remediation time.2 However, the saturated TCE concentrations in the vicinity of a TCE DNAPL are toxic for dechlorinating bacteria.2,3 This TCE self-inhibition, i.e. inhibition of the dechlorination reaction at high TCE concentrations, likely limits the biological enhancement of DNAPL dissolution.4 As such, an assessment of the potential of bioenhanced DNAPL dissolution requires that the dechlorination kinetics at elevated TCE concentrations is well understood. Several models have been developed to describe dechlorination at high chlorinated ethene concentrations. Most of these models assume that elevated chlorinated ethene concentrations reduce the dechlorination activity. Amos et al.,5 for instance, © 2013 American Chemical Society

used a general threshold concentration to model perchloroethene (PCE) self-inhibition, while Yu and Semprini6 incorporated Haldane inhibition to describe dechlorination at high PCE and TCE concentrations. Haest et al.7 used a log− logistic expression to describe TCE self-inhibition, as Haldane inhibition failed to describe their experimental results. Alternatively, self-inhibition has been modeled by assuming toxic effects on the dechlorinating biomass. Sabalowsky and Semprini,8 for instance, used an enhanced cell decay to model the toxicity of elevated cis-DCE concentrations. In addition, their model well described the experimental results of Yu and Semprini,6 which were previously modeled with Haldane inhibition.8 Other studies rather suggested a combination of both modeling approaches. Huang and Becker,9 for instance, used Haldane inhibition in combination with a biomass inactivation term to model PCE self-inhibition. None of these studies, however, could validate the used model assumptions, as they did not quantify the cell densities of the dechlorinating bacteria. Received: Revised: Accepted: Published: 1510

August 21, 2012 December 18, 2012 January 2, 2013 January 2, 2013 dx.doi.org/10.1021/es303390q | Environ. Sci. Technol. 2013, 47, 1510−1517

Environmental Science & Technology

Article

2.4. Analytical Methods. Concentrations of TCE and cisDCE were analyzed using GC-FID.7 Stock solutions of TCE and cis-DCE in methanol were used to calibrate for the chlorinated ethene concentrations in the aqueous phase. The distribution between the aqueous phase and the gas phase was quantified by adding precise amounts of TCE and cis-DCE to serum bottles with the same volume of medium as in the batch experiment. These batch bottles were stored under the same conditions as the experimental bottles (20 °C). Sampling of the aqueous phase allowed calculation of dimensionless Henry coefficients for TCE and cis-DCE of 0.50 ± 0.03 and 0.074 ± 0.015, respectively. In addition, GC-FID was used to measure the concentrations of VC, ethene, and methane, but these concentrations were always below the detection limit. Formate concentrations were measured using anion exclusion chromatography as described by Philips et al.12 DNA was extracted using the DNeasy Blood and Tissue kit (Qiagen)11 and Geobacter and bacterial 16S rRNA gene copy numbers were quantified with real-time qPCR.10 Dehalococcoides 16S rRNA gene copy numbers were determined according to Dijk et al.,13 but were always at the detection limit. A control sample was included in each combined DNA extraction and qPCR run as described by Philips et al.11 to reduce the variability in the 16S rRNA gene copy number quantification. The 16S rRNA gene copy numbers correspond to cell numbers as was previously described.11 The detection limit for the qPCR analysis was 1 × 104 16S rRNA gene copy numbers per mL. 2.5. Modeling. Dechlorination of TCE and growth by Geobacter were modeled using the system of ordinary differential equations:

The objective of this study was to examine whether selfinhibition is related to a reduced dechlorination activity or to an enhanced cell decay at elevated TCE concentrations. A batch experiment was performed combining four different initial TCE concentrations and three different inoculation densities. Chlorinated ethene concentrations and dechlorinator cell densities were monitored. In addition, the survival of the dechlorinators after exposure to toxic TCE concentrations was tested. The experimental results were modeled with either an enhanced cell decay or a reduced activity at elevated TCE concentrations.

2. MATERIALS AND METHODS 2.1. Medium and Culture. Anaerobic defined medium was prepared as described by Haest et al.,10 except that the concentration of yeast extract was reduced to 10 mg·L−1. A subculture of the KB-1 culture was used to inoculate the experiment. This KB-1 subculture was grown with formate as electron donor as previously described.11 Briefly, this culture was stimulated to dechlorinate TCE only to cis-DCE by weekly flushing and fresh TCE addition. Philips et al.11 demonstrated that this KB-1 subculture consisted of dechlorinating bacteria highly similar to Geobacter lovleyi SZ and fermenting bacteria related to Clostridium, while Dehalococcoides species were outcompeted from this culture. 2.2. Batch Experiment. Twenty-four serum bottles of 158 mL each were filled with 80 mL of inoculated anaerobic medium. Volumes of the KB-1 subculture corresponding to 0.6, 5.9, and 59.6% were mixed with medium, resulting in initial cell densities of, respectively, 4.0 × 105, 4.0 × 106, and 4.0 × 107 Geobacter 16S rRNA gene copies per mL and, respectively, 1.0 × 106, 1.0 × 107, and 1.0 × 108 bacterial 16S rRNA gene copies per mL, as determined with quantitative PCR (qPCR). We will refer to these different inoculation densities as LD (low inoculation density), ND (normal inoculation density, i.e. the inoculation density normally applied in our experiments), and HD (high inoculation density). Mixing of the medium and the inoculum and filling of the bottles was performed in a glovebox with N2/H2 (95/5) atmosphere. Afterward, the headspace of the batch bottles was flushed with nitrogen gas to remove H2. Neat TCE volumes were injected that corresponded to total amounts of TCE in the bottle of, respectively, 0.17, 0.24, 0.30, and 0.36 mmol and TCE concentrations in the aqueous phase of, respectively, 1.4, 2.0, 2.5, and 3.0 mM. All bottles were amended with 0.32 mmol formate, i.e. an amount almost equimolar to the highest initial TCE concentration. All treatments were performed in duplicate (A and B). Bottles were incubated on a horizontal shaker (100 rpm) at 20 °C. Samples (0.5 mL) of the aqueous phase were taken over time to monitor TCE, cis-DCE, and formate concentrations and microbial growth. 2.3. Reactivation Experiment. Some treatments of the batch experiment did not show dechlorination by day 80 after the start of the experiment. At day 82, the TCE concentration was lowered in the replicates B of these treatments, in order to test if this could reactivate the dechlorination. The TCE was removed by flushing the aqueous phase with nitrogen gas. Afterward, fresh TCE was added corresponding to a total amount of 0.17 mmol and an aqueous concentration of 1.4 mM. In addition, 0.32 mmol formate was added, as well a volume of fresh medium to compensate the loss of liquid volume due to previous samplings. The replicates A were not flushed and were further monitored for comparison.

⎧ dc ⎛ ⎞ Vw ⎪ TCE , w = −k ·c ·⎜ ⎟ cell GEO ⎜ ⎪ dt Vw + HTCE·Vg ⎟⎠ ⎝ ⎨ ⎪ dc ⎪ GEO = YGEO·kcell·cGEO − kd·cGEO ⎩ dt

(1)

where cTCE,w is the aqueous phase TCE concentration (mmol·L−1), cGEO is the Geobacter cell density (cells·L−1), t is the time (day), and YGEO is the growth yield of Geobacter (cells·mmol−1). The factor Vw/(Vw + HTCE·Vg) accounts for the distribution of TCE between the aqueous and the gas phase, whereas TCE is only degraded in the aqueous phase. In this factor is HTCE the dimensionless Henry coefficient for TCE, and Vw and Vg are the volumes of the aqueous and gas phase (L), respectively. TCE self-inhibition was incorporated by assuming that either the microbial decay coefficient kd (day−1) or the cell specific dechlorination rate kcell (mmol·cell−1·day−1) was affected by the TCE concentration. A first model relates the toxicity of TCE to an enhanced cell decay (kd) at increasing TCE concentrations. Sabalowsky and Semprini8 previously used the following relationship proposed by Rittmann and Saez:14 ⎛ cTCE , w ⎞ kd = kd ,min·⎜1 + ⎟ Kt ⎠ ⎝

(2)

where kd,min is the minimum or uninhibited microbial decay coefficient (day−1) and Kt is an inhibition coefficient (mmol·L−1). This model further assumed that high TCE concentrations did not affect the dechlorination activity and 1511

dx.doi.org/10.1021/es303390q | Environ. Sci. Technol. 2013, 47, 1510−1517

Environmental Science & Technology

Article

Figure 1. Experimental and modeled TCE and cis-DCE concentrations and experimental formate concentrations in total mmol in the bottle for the batch experiment combining four different initial TCE concentrations (concentrations in the aqueous phase: 1.4, 2.0, 2.5, and 3.0 mM) and three different inoculation densities (LD: 4.0 × 105, ND: 4.0 × 106, and HD: 4.0 × 107 Geobacter 16S rRNA gene copies per mL). Data for both replicates are plotted till day 30. The reduced activity model (eqs 1 and 4) was used to model the TCE and cis-DCE concentrations.

concentrations. Two parameters of the reduced activity model (YGEO and kd) were directly calculated from the experimental results, while the three remaining parameters (kcell,max, EC50, and bi) were optimized. Optimization was performed using Optipa16 which is a tool built in Matlab to optimize models based on ordinary different equations. A sensitivity analysis of both models was performed as described in the SI.

Monod kinetics was used to describe the cell specific dechlorination rate:

kcell =

kcell ,max ·cTCE , w Ks , TCE + cTCE , w

(3)

where kcell,max is the maximum TCE dechlorination rate (mmol·cell−1·day−1) and Ks,TCE is the half saturation concentration (mmol·L−1). In addition, this model set the Geobacter growth yield (YGEO) to zero once the Geobacter cell density decreased below one Geobacter cell per bottle. We will further refer to this model as the enhanced decay model. A second model relates TCE self-inhibition to a reduced dechlorination activity (kcell) at elevated TCE concentrations. Haest et al.7 previously modeled TCE self-inhibition by extending Monod kinetics with the log−logistic relationship of Doelman and Haanstra:15 kcell = (Ks , TCE

kcell ,max ·cTCE , w ⎛ + cTCE , w) ·⎜1 + ⎝

cTCE , w

bi ⎞

( ) ⎟⎠ EC50

3. RESULTS 3.1. Batch Experiment. The time for complete dechlorination of TCE to cis-DCE slightly differed between the replicates of some treatments, but, overall a good reproducibility was found (Figure 1). The results demonstrated that the time required for complete conversion of TCE increased with an increase of the initial TCE concentration and a decrease of the inoculation density. No dechlorination was found by day 30 in three treatments that combined a high initial TCE concentration and a low inoculation density (LD + 2.5 and 3.0 mM TCE, ND + 3.0 mM TCE) (Figure 1). Sampling of these treatments was continued, but even by day 80 no dechlorination was observed (Figure 3). Dechlorination did proceed in the LD bottles with low initial TCE concentrations (1.4 and 2.0 mM TCE), but only after a lag phase (Figure 1). A shorter lag phase and a faster degradation was found in the corresponding ND treatments (1.4 and 2.0 mM TCE). Dechlorination also started in the ND treatment with 2.5 mM TCE, but only after a lag phase of 20 days. In the HD bottles dechlorination was found at all initial TCE concentrations (1.4−3.0 mM TCE). Dechlorination was fast in these treatments and an almost linear TCE concentration change was shown. In all treatments formate was consumed (Figure 1). This consumption was related to the dechlorination of TCE, as the residual formate concentration was lower if more TCE was converted. However, in the treatments showing a lag phase, the

(4)

where kcell,max is the uninhibited or maximum cell specific dechlorination rate (mmol·cell−1·day−1), EC50 is the TCE concentration at which the cell specific dechlorination rate is half of the uninhibited rate (mmol·L−1) and bi is a parameter proportional to the slope of the cell specific dechlorination rate at the EC50 concentration (-). A graphical interpretation of eq 4 is provided in the Supporting Information (SI). This model further assumed a constant microbial decay coefficient. We will refer to this model as the reduced activity model. The system of ordinary different equations (eq 1) was solved using the ode15s solver in Matlab (MathWorks). The parameter value for Ks,TCE was taken from Haest et al.7 The remaining parameters of the enhanced decay model (kcell,max, YGEO, kd, and Kt) were fitted to the experimental TCE 1512

dx.doi.org/10.1021/es303390q | Environ. Sci. Technol. 2013, 47, 1510−1517

Environmental Science & Technology

Article

Figure 2. Experimental and modeled Geobacter 16S rRNA gene copies per mL and experimental bacterial 16S rRNA copies per mL for the batch experiment combining four different initial TCE concentrations (concentrations in the aqueous phase: 1.4, 2.0, 2.5, and 3.0 mM) and three different inoculation densities (LD: 4.0 × 105, ND: 4.0 × 106, and HD: 4.0 × 107 Geobacter 16S rRNA gene copies per mL). Data for both replicates are plotted till day 30. The reduced activity model (eqs 1 and 4) was used to model the Geobacter cell densities.

formate concentration declined before TCE dechlorination started. Moreover, also in the treatments lacking TCE dechlorination a decrease of the formate concentration was observed. The formate converted in the latter treatments potentially accumulated as hydrogen. However, this could not be verified, as hydrogen concentrations were not measured. No growth of Geobacter was found in the treatments lacking dechlorination (Figure 2). In contrast, 16S rRNA gene copy numbers of Geobacter increased with 1 and 2 orders of magnitude in the dechlorinating ND and LD treatments, respectively. No significant increase of the Geobacter 16S rRNA gene copy numbers was observed in the HD batches. The stable Geobacter copy numbers in the latter treatments were likely due to the high initial cell density and could explain the linear TCE concentration change observed for these treatments. In most treatments bacterial 16S rRNA gene copy numbers increased along with the increase of the Geobacter 16S rRNA gene copy numbers (Figure 2). However, in the treatments showing a lag phase the bacterial 16S rRNA gene copy numbers increased before the Geobacter 16S rRNA gene copy numbers started to rise. Moreover, bacterial densities increased more than 1 order of magnitude in the treatments lacking dechlorination and growth of Geobacter. The increase in the bacterial copy numbers in combination with the consumption of formate in the treatments lacking dechlorination suggest that the fermenting species present in the used KB-1 subculture are less vulnerable to elevated TCE concentrations than the Geobacter dechlorinators. 3.2. Reactivation Experiment. As explained above, no dechlorination was observed by day 80 in three treatments (LD + 2.5 and 3.0 mM TCE and ND + 3.0 mM TCE) (Figure 3). By this time, the 16S rRNA gene copy numbers for Geobacter in these treatments had clearly declined (Figure 4). Sampling of the replicates A of these treatments was continued till day 110. Dechlorination eventually started in the replicate A of the

treatment LD + 2.5 mM TCE around day 100 (Figure 3) and was associated with a 2 orders of magnitude increase of the Geobacter cell density (Figure 4). In contrast, no dechlorination was observed in the two other treatments by day 110. In the replicates B of the treatments lacking dechlorination, the aqueous phase TCE concentration was lowered to 1.4 mM at day 82. This lowering of the TCE concentration triggered the dechlorination (Figure 3) and the growth of Geobacter (Figure 4) in all three treatments after an initial lag phase. These results suggest that the Geobacter cells were able to survive the elevated TCE concentrations and that lowering of the TCE concentration reversed the inactivation of the dechlorination reaction. 3.3. Modeling. Optimization was performed by fitting the experimental TCE concentrations. The data points of the treatment HD + 3.0 mM TCE were excluded from the optimization, since none of the models could fit the linear TCE concentration decrease in this treatment. In addition, the data points obtained from the replicates B after lowering of the TCE concentration were omitted from the fit. All parameters of the enhanced decay model were optimized, except for the parameter Ks,TCE (Table S1). Using an experimentally determined growth yield, this model could not describe the experimental results and, therefore, the parameter YGEO was included in the optimization. This model fitted the experimental TCE concentrations well (Figures S2 and S6A) (correlation coefficient (R2) of 0.87) and adequately described the growth of Geobacter in the treatments with dechlorination (Figure S3). However, this model failed to fit the onset of dechlorination after 100 days in the replicate A of the treatment LD + 2.5 mM TCE (Figure S4) and described a much sharper decrease of the Geobacter cell density than experimentally observed for the treatments LD and ND + 3.0 mM (Figure S5). In addition, the model largely underestimated the time required for the complete conversion of TCE to cis-DCE in the treatment HD + 3.0 mM (Figure S2). The optimized parameter values were further used to model the effect of the lowering of 1513

dx.doi.org/10.1021/es303390q | Environ. Sci. Technol. 2013, 47, 1510−1517

Environmental Science & Technology

Article

Figure 3. Experimental and modeled TCE and cis-DCE concentrations and experimental formate concentrations in total mmol in the bottle for the treatments that did not show dechlorination by day 80. The monitoring of replicates A was continued after day 82, whereas the TCE concentration was lowered to 1.4 mM at day 82 in the replicates B to test reactivation. Lowering of the TCE concentration triggered the dechlorination in all these treatments. The reduced activity model (eqs 1 and 4) was used to model the TCE and cis-DCE concentrations.

Figure 4. Experimental and modeled Geobacter 16S rRNA gene copies per mL and experimental total bacterial 16S rRNA gene copies per mL for the treatments that did not show dechlorination by day 80. The monitoring of replicates A was continued after day 82, whereas the TCE concentration was lowered to 1.4 mM at day 82 in the replicates B to test reactivation. Lowering of the TCE concentration triggered the dechlorination in all these treatments. The reduced activity model (eqs 1and 4) was used to model the Geobacter cell densities.

on the increase of the Geobacter copy numbers in the LD and ND treatments with initial TCE concentrations between 1.4 and 2.5 mM (Figures 2 and 4). No correlation between the growth yield and the initial TCE concentration was found and, therefore, an average value for the growth yield was calculated. The decrease of the Geobacter cell densities in the treatments lacking dechlorination during the first 80 days (Figure 4) was used to calculate a microbial decay coefficient for Geobacter. Three other parameters of the reduced activity model (kcell,max, EC50, and bi) were optimized (Table S1). The resulting model described the TCE dechlorination and the growth of Geobacter in the batch experiment very well (Figures 1, 2, and S6B) (R2 of 0.92). This model succeeded in fitting the lack of dechlorination in the treatments LD and ND + 3.0 mM TCE, as well as the start of the dechlorination after 100 days in

the TCE concentration in the treatments initially lacking dechlorination. The predicted Geobacter cell density at day 82 for the different treatments was used as initial condition to solve the system of differential equations (eq 1) over a second time interval starting at day 82. The enhanced cell decay model predicted reactivation for the replicate A of treatment LD + 2.5 mM (Figure S4), but failed to predict reactivation for the treatments LD and ND + 3.0 mM (Figure S4), because the modeled Geobacter cell density by day 82 in these treatments was too low to allow recovery (Figure S5). A second model assuming a reduced dechlorination activity at elevated TCE concentrations was subsequently tested. Two parameters of this model (YGEO and kd) were directly calculated from the experimental Geobacter 16S rRNA gene copy numbers (Table S1). Growth yields for Geobacter were calculated based 1514

dx.doi.org/10.1021/es303390q | Environ. Sci. Technol. 2013, 47, 1510−1517

Environmental Science & Technology

Article

(Figures 3 and 4). The enhanced decay model, however, only predicted reactivation for the treatment LD + 2.5 mM, as the modeled cell density by day 80 for the other treatments was too low to sustain recovery (Figure S4). Consequently, our experimental and modeling analysis shows that TCE selfinhibition at elevated TCE concentrations is explained by a reduced dechlorination activity rather than by an enhanced cell decay of the Geobacter dechlorinators. These results are in agreement with the successful use of the reduced activity model to describe dechlorination in diffusion-cells,18 flow-through columns,19 and 2D sand boxes.20 The toxicity of solvents has often been ascribed to their disturbing effects on the cytoplasmic membrane.21 Our study shows that elevated TCE concentrations (up to 3.0 mM) led to a reversible inactivation of the Geobacter dechlorinators, which could potentially result from a nonlethal membrane disturbance. Singh et al.22 demonstrated a rapid loss of viability of Pseudomonas putida F1 cells at saturated TCE concentrations. Therefore, it seems plausible that higher TCE concentrations than applied in our study are lethal to the Geobacter dechlorinators and may increase their cell decay. Future research, therefore, should discriminate among dead, inactive, and active cells at a full range of concentrations to clarify the biochemical mechanism of TCE toxicity. Furthermore, it should be noted that several bacterial adaptation mechanisms to high solvent concentrations have been demonstrated.23−26 Adaptation of dechlorinating bacteria to chlorinated ethenes has not yet been observed, but this could largely affect the dechlorination kinetics at elevated chlorinated ethene concentrations and, therefore, warrants future investigation. 4.2. High Geobacter Cell Densities Overcome TCE SelfInhibition. The batch experiment demonstrated dechlorination for the treatment HD + 3.0 mM, whereas at the same initial TCE concentration dechlorination was completely inhibited at lower inoculation densities (LD and ND + 3.0 mM) (Figure 1). Similar findings were obtained for PCE by Huang and Becker,9 whereas also Haest et al.7 argued, based on modeling, that self-inhibition depends on the number of dechlorinators. This dependency on the dechlorinator cell density can be explained from the structure of both the enhanced decay and the reduced activity model used in this study. A first situation arises at low initial TCE concentrations or initial net growth, i.e. if:

the treatment LD + 2.5 mM TCE (Figure 3). In addition, the model was able to predict the complete conversion of TCE to cis-DCE in 25 days in the treatment HD + 3.0 mM TCE, although the predicted decline of the TCE concentration clearly differed in shape from the experimentally recorded decline (Figure 1). The parameter values optimized for the reduced activity model were also used to describe the effect of the lowering of the TCE concentrations in the treatments initially lacking dechlorination. This model succeeded in predicting the reactivation of the dechlorination reaction for all three treatments (Figures 3 and 4). Results of the sensitivity analysis are described in the SI.

4. DISCUSSION 4.1. Survival of Geobacter at Elevated TCE Concentrations Is Described by a Reduced Activity Rather than by an Enhanced Decay. In this study, two models, assuming either an enhanced cell decay or a reduced dechlorination activity at increasing TCE concentrations, were tested to model the dechlorination and growth of Geobacter. Both models fitted the experimental TCE concentrations rather well (Figure 1, Figure S2). In addition, the combination of both models was tried, but this extended model did not fit the experimental results any better than the individual models (results not shown). This is in contrast to the findings of Huang and Becker,9 which showed that the incorporation of both a biomass and an activity reduction at elevated PCE concentrations was required to model PCE self-inhibition. Alternatively to the two model concepts tested in this study, TCE self-inhibition could be due to a reduction of the growth yield parameter Ym at increasing TCE concentrations.7 Isken et al.,17 for instance, demonstrated decreasing Pseudomonas putida growth yields at increasing toluene concentrations. This concept, however, is insufficient to explain TCE self-inhibition in our experiment, as different dechlorination rates were observed for the HD batches with different initial TCE concentrations, while the Geobacter cell densities in these batches were high and relatively constant. The enhanced cell decay model well described the dechlorination and growth of Geobacter in most treatments, but for the treatments lacking dechlorination this model predicted a much sharper decline of the Geobacter cell density than experimentally observed (Figures S2−S5). This difference may be due to the quantification of 16S rRNA gene copy numbers, which may include gene copies originating from dead cells. The reduced activity model, however, better described the decrease of the Geobacter cell densities in the treatments lacking dechlorination (Figure 4). In addition, the latter model succeeded in fitting the 100 days lag phase for the treatment LD + 2.5 mM TCE (Figure 3), in contrast to the enhanced decay model (Figure S4). The most convincing argument for the reduced activity model, however, is the fact that this model was able to predict the reactivation of the dechlorination reaction after exposure to inhibiting TCE concentrations (Figures 3 and 4). Huang and Becker9 also lowered the PCE concentration in their batches lacking dechlorination, but they did not observe reactivation, maybe because they monitored the chlorinated ethene concentrations only for an additional 24 h. Our findings, however, demonstrate that the Geobacter dechlorinators were able to survive the inhibiting TCE concentrations and that the dechlorinating biomass at day 80 was sufficiently high to allow recovery. This recovery was predicted for all three treatments by the reduced activity model

Ym·kcell > kd

at

t=0

(5)

In this case, decay can never catch up with growth, because the decline of the TCE concentration together with the increase of the dechlorinator cell density decreases kd or increases kcell in the enhanced decay or the reduced activity model, respectively. As such, in this case, the complete dechlorination of TCE to cis-DCE is predicted for any initial Geobacter cell density. The dechlorination rate, however, depends on the initial cell density and in the case of low Geobacter cell densities dechlorination is predicted to proceed only after a long lag phase. Using the model parameters optimized for the reduced activity model (Table S1), relation 5 becomes true at TCE concentrations below 2.6 mM. The experimental results agree with this finding, since all treatments amended with TCE concentrations below this threshold showed dechlorination (Figure 1) after a lag phase of maximal 100 days for the treatment LD + 2.0 mM TCE (Figure 3). 1515

dx.doi.org/10.1021/es303390q | Environ. Sci. Technol. 2013, 47, 1510−1517

Environmental Science & Technology

Article

information is available free of charge via the Internet at http:// pubs.acs.org/

A second situation is that of high initial TCE concentrations or initial net decay, i.e. if: Ym·kcell < kd

at

t=0

■

(6)

Corresponding Author

In this case, it depends on the initial Geobacter cell density whether growth can or can not catch up with decay, as the decrease of kd or the increase of kcell with decreasing TCE concentrations indirectly depends on the microbial density. At low Geobacter cell densities, the TCE concentrations slowly decline and the decrease of kd or the increase of kcell is insufficient to obtain net growth, whereas at higher Geobacter cell densities the dechlorination rate is faster and kd decreases or kcell increases to the point where growth exceeds decay. Using the model parameters for the reduced activity model (Table S1), dechlorination or complete inhibition, depending on the initial Geobacter cell density, is predicted for initial TCE concentrations above 2.6 mM. This is in agreement with the experimental results, since dechlorination of 3.0 mM TCE only started in the treatment with the highest initial Geobacter cell density (Figure 1). The above interpretation is valid for batch systems, in which kd or kcell is indirectly related to the Geobacter cell density, since the TCE concentrations decline concurrently with the increase of the Geobacter cell densities in these systems. However, in the presence of a TCE DNAPL or in flow-through systems with a constant influent TCE concentration, the relation between kd or kcell and the dechlorinator cell density is different. In these systems, TCE is continuously replenished and, as such, growth of the dechlorinators does not correspond with a similar decrease of the TCE concentration as in batch. Therefore, additional interpretations are required to understand the effect of the dechlorinator cell density on self-inhibition in systems other than batch. 4.3. Clostridium are Less Vulnerable to Elevated TCE Concentrations than Geobacter. In the KB-1 subculture, formate is likely converted by Clostridium species.11 The dehydrogenation of formate to hydrogen, however, does not yield energy and it was suggested that the Clostridium in the used KB-1 subculture grow on the fermentation of yeast extract and cystein, which are two medium constituents.11 As such, the conversion of formate is rather a cometabolic process, which probably depends on the metabolic activity of the Clostridium. In the batch experiment of the current study, a decrease of the formate concentration and an increase of the total bacterial cell densities were observed in the treatments lacking dechlorination (LD and ND + 3.0 mM TCE) (Figures 1 and 2). These results suggest that the Clostridium were able to grow and convert formate at TCE concentrations inhibiting the Geobacter dechlorinators. Bowman et al.27 similarly reported that tolerance to elevated PCE concentrations is widespread among the genus Clostridium, although the mechanisms for this solvent tolerance are still unknown. These findings demonstrate that the lack of dechlorination in some treatments with an elevated initial TCE concentration was not due to an inhibited electron donor conversion.

■

AUTHOR INFORMATION

*E-mail: [email protected]; tel: +3216321609; fax: +3216321997. Present Address §

Department of Microbiology, University of Massachusetts, 639 North Pleasant Street, Amherst MA 01003-9298, USA. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS We thank Fanny Hamels for her assistance in performing the experimental work. This research was funded as a Ph.D. Fellowship of the Research Foundation − Flanders (FWO).

■

REFERENCES

(1) Bradley, P. M. History and ecology of chloroethene biodegradation: A review. Bioremediat. J. 2003, 7 (2), 81−109. (2) Yang, Y. R.; McCarty, P. L. Biologically enhanced dissolution of tetrachloroethene DNAPL. Environ. Sci. Technol. 2000, 34 (14), 2979− 2984. (3) Duhamel, M.; Wehr, S. D.; Yu, L.; Rizvi, H.; Seepersad, D.; Dworatzek, S.; Cox, E. E.; Edwards, E. A. Comparison of anaerobic dechlorinating enrichment cultures maintained on tetrachloroethene, trichloroethene, cis-dichloroethene and vinyl chloride. Water Res. 2002, 36 (17), 4193−4202. (4) Chu, M.; Kitanidis, P. K.; McCarty, P. L. Possible factors controlling the effectiveness of bioenhanced dissolution of nonaqueous phase tetrachloroethene. Adv. Water Res. 2004, 27 (6), 601− 615. (5) Amos, B. K.; Christ, J. A.; Abriola, L. M.; Pennell, K. D.; Loffler, F. E. Experimental evaluation and mathematical modeling of microbially enhanced tetrachloroethene (PCE) dissolution. Environ. Sci. Technol. 2007, 41 (3), 963−970. (6) Yu, S.; Semprini, L. Kinetics and modeling of reductive dechlorination at high PCE and TCE concentrations. Biotechnol. Bioeng. 2004, 88 (4), 451−464. (7) Haest, P. J.; Springael, D.; Smolders, E. Dechlorination kinetics of TCE at toxic TCE concentrations: Assessment of different models. Water Res. 2010, 44 (1), 331−339. (8) Sabalowsky, A. R.; Semprini, L. Trichloroethene and cis-1,2dichloroethene concentration-dependent toxicity model simulates anaerobic dechlorination at high concentrations: I. Batch-fed reactors. Biotechnol. Bioeng. 2010, 107 (3), 529−539. (9) Huang, D.; Becker, J. G. Dehalorespiration model that incorporates the self-inhibition and biomass inactivation effects of high tetrachloroethene concentrations. Environ. Sci. Technol. 2011, 45 (3), 1093−1099. (10) Haest, P. J.; Philips, J.; Springael, D.; Smolders, E. The reactive transport of trichloroethene is influenced by residence time and microbial numbers. J. Contam. Hydrol. 2011, 119 (1−4), 89−98. (11) Philips, J.; Hamels, F.; Smolders, E.; Springael, D. Distribution of a dechlorinating community in relation to the distance from a trichloroethene dense non-aqueous phase liquid in a model aquifer. FEMS Microbiol. Ecol. 2012, 81, 636−647. (12) Philips, J.; Springael, D.; Smolders, E. A three-layer diffusion-cell to examine bio-enhanced dissolution of chloroethene dense nonaqueous phase liquid. Chemosphere 2011, 83 (7), 991−996. (13) Dijk, J. A.; Breugelmans, P.; Philips, J.; Haest, P. J.; Smolders, E.; Springael, D. Catalyzed reporter deposition-fluorescent in situ hybridization (CARD-FISH) detection of Dehalococcoides. J. Microbiol. Methods 2008, 73 (2), 142−147.

ASSOCIATED CONTENT

S Supporting Information *

Parameter values obtained for the enhanced decay and the reduced activity self-inhibition model; a graphical interpretation of eq 4; modeling results for the enhanced decay model; and description of a sensitivity analysis of both models. This 1516

dx.doi.org/10.1021/es303390q | Environ. Sci. Technol. 2013, 47, 1510−1517

Environmental Science & Technology

Article

(14) Ritmann, B. E.; Saez, P. B., Modeling biological processes involved in degradation of hazardous organic substrates. In Biotreatment of Industrial and Hazardous Waste; Levin, M. A., Gealt, M. A., Eds.; McGraw-Hill, 1993; pp 113−136. (15) Doelman, P.; Haanstra, L. Short-term and long-term effects of heavy-metals on phosphatase-activity in soils- An ecological doseresponse model approach. Biol. Fertil. Soils 1989, 8 (3), 235−241. (16) Hertog, M. L. A. T. M.; Verlinden, B. E.; Lammertyn, J.; Nicolai, B. M. OptiPa, an essential primer to develop models in the postharvest area. Comput. Electron. Agric. 2007, 57 (1), 99−106. (17) Isken, S.; Derks, A.; Wolffs, P. F. G.; de Bont, J. A. M. Effect of organic solvents on the yield of solvent-tolerant Pseudomonas putida S12. Appl. Environ. Microbiol. 1999, 65 (6), 2631−2635. (18) Philips, J. Bioremediation of Groundwater Contaminated with Trichloroethene: Assessment of the Biological and Chemical Factors Affecting Bio-enhanced Free Phase Dissolution. PhD thesis, KU Leuven ,2012. (19) Haest, P. J.; Springael, D.; Smolders, E. Modelling reactive CAH transport using batch experiment degradation kinetics. Water Res. 2010, 44 (9), 2981−2989. (20) Haest, P. J.; Springael, D.; Seuntjens, P.; Smolders, E. Selfinhibition can limit biologically enhanced TCE dissolution from a TCE DNAPL. Chemosphere 2012, 89, 1369−1375. (21) Duldhardt, I.; Nijenhuis, I.; Schauer, F.; Heipieper, H. J. Anaerobically grown Thauera aromatica, Desulfococcus multivorans, Geobacter sulfurreducens are more sensitive towards organic solvents than aerobic bacteria. Appl. Microbiol. Biotechnol. 2007, 77 (3), 705− 711. (22) Singh, R.; Olson, M. S. Kinetics of trichloroethylene and toluene toxicity to Pseudomonas putida F1. Environ. Toxicol. Chem. 2010, 29 (1), 56−63. (23) Isken, S.; de Bont, J. A. M. Bacteria tolerant to organic solvents. Extremophiles 1998, 2 (3), 229−238. (24) Ramos, J. L.; Duque, E.; Gallegos, M. T.; Godoy, P.; RamosGonzalez, M. I.; Rojas, A.; Teran, W.; Segura, A. Mechanisms of solvent tolerance in gram-negative bacteria. Annu. Rev. Microbiol. 2002, 56, 743−768. (25) Sardessai, Y.; Bhosle, S. Tolerance of bacteria to organic solvents. Res. Microbiol. 2002, 153 (5), 263−268. (26) Duldhardt, I.; Gaebel, J.; Chrzanowski, L.; Nijenhuis, I.; Haertig, C.; Schauer, F.; Heipieper, H. J. Adaptation of anaerobically grown Thauera aromatica, Geobacter sulfurreducens and Desulfococcus multivorans to organic solvents on the level of membrane fatty acid composition. Microb. Biotechnol. 2010, 3 (2), 201−209. (27) Bowman, K. S.; Rainey, F. A.; Moe, W. M. Production of hydrogen by Clostridium species in the presence of chlorinated solvents. FEMS Microbiol. Lett. 2009, 290 (2), 188−194.

1517

dx.doi.org/10.1021/es303390q | Environ. Sci. Technol. 2013, 47, 1510−1517