Reductive dechlorination of 1,2-dichloroethane in the presence of chloroethenes and 1,2-dichloropropane as co-contaminants

1,2-Dichloroethane (1,2-DCA) is one of the most abundant manmade chlorinated organic contaminants in the world. Reductive dechlorination of 1,2-DCA by organohalide-respiring bacteria (OHRB) can be impacted by other chlorinated contaminants such as chloroethenes and chloropropanes that can co-exist with 1,2-DCA at contaminated sites. The aim of this study was to evaluate the effect of chloroethenes and 1,2-dichloropropane (1,2-DCP) on 1,2-DCA dechlorination using sediment cultures enriched with 1,2-DCA as the sole chlorinated compound (EA culture) or with 1,2-DCA and tetrachloroethene (PCE) (EB culture), and to model dechlorination kinetics. Both cultures contained Dehalococcoides as most predominated OHRB, and Dehalogenimonas and Geobacter as other known OHRB. In sediment-free enrichments obtained from the EA and EB cultures, dechlorination of 1,2-DCA was inhibited in the presence of the same concentrations of either PCE, vinyl chloride (VC), or 1,2-DCP; however, concurrent dechlorination of dual chlorinated compounds was achieved. In contrast, 1,2-DCA dechlorination completely ceased in the presence of cis-dichloroethene (cDCE) and only occurred after cDCE was fully dechlorinated. In turn, 1,2-DCA did not affect dechlorination of PCE, cDCE, VC, and 1,2-DCP. In sediment-free enrichments obtained from the EA culture, Dehalogenimonas 16S rRNA gene copy numbers decreased 1–3 orders of magnitude likely due to an inhibitory effect of chloroethenes. Dechlorination with and without competitive inhibition fit Michaelis-Menten kinetics and confirmed the inhibitory effect of chloroethenes and 1,2-DCP on 1,2-DCA dechlorination. This study reinforces that the type of chlorinated substrate drives the selection of specific OHRB, and indicates that removal of chloroethenes and in particular cDCE might be necessary before effective removal of 1,2-DCA at sites contaminated with mixed chlorinated solvents. Electronic supplementary material The online version of this article (10.1007/s00253-019-09985-8) contains supplementary material, which is available to authorized users.


Introduction
Understanding biodegradation bottlenecks has been a major objective in efforts to harness the metabolic potential of microorganisms for bioremediation of sites contaminated with organic pollutants Meckenstock et al. 2015;Vandermaesen et al. 2016). An important class of such contaminants comprises chlorinated solvents such as chloroethenes, chloroethanes, and chloropropanes that have adverse effects on human and environmental health (EPA 2018;Weatherill et al. 2018). Organohalide respiration (OHR) is an example of a microbial metabolism that has been successfully harnessed for engineered remediation of sites contaminated with chlorinated solvents (Atashgahi et al. 2017;Edwards 2014;Ellis et al. 2000). This process is mediated by organohalide-respiring bacteria (OHRB) belonging to distinct genera within the phyla Chloroflexi (e.g., Peng Peng and Uwe Schneidewind contributed equally to this work. Electronic supplementary material The online version of this article (https://doi.org/10.1007/s00253-019-09985-8) contains supplementary material, which is available to authorized users.
Different modeling approaches of varying complexity have been developed to understand the reductive dechlorination of chloroethenes with and without competitive inhibition (Chambon et al. 2013). The description of the reaction kinetics varies from first-order (Corapcioglu et al. 2004;Da Silva and Alvarez 2008) to the more elaborate Michaelis-Menten equations (Garant and Lynd 1998;Haston and McCarty 1999) or Monod kinetics if the responsible OHRB can be sufficiently quantified (Yu and Semprini 2004). The latter two kinetic modeling approaches have been applied at lab and field scales to study competitive inhibition (Christ and Abriola 2007;Lai and Becker 2013;Yu et al. 2005), self-inhibition (Haest et al. 2010a), electron donor limitation (Cupples et al. 2004), dechlorination in the presence of multiple bacterial species (Brovelli et al. 2012;Duhamel and Edwards 2006;Haest et al. 2010b), and dechlorination in conjunction with fermentation, sulfate reduction, or methanogenesis (Kouznetsova et al. 2010;Malaguerra et al. 2011). The Michaelis-Menten and Monod kinetic modeling approaches can also be used to study the dechlorination of chloroethanes or chloropropanes; however, examples in literature are rare. Notable exceptions are Mayer-Blackwell et al. (2016) who studied concurrent dechlorination of 1,2-DCA and cDCE, and Colombani et al. (2014) who studied 1,2-DCA degradation under the influence of salt water intrusion.
In this study, in cultures amended with one or two chlorinated compounds (1,2-DCA with either tetrachloroethene (PCE), cDCE, VC, or 1,2-DCP), the impact of cocontaminants on OHRB was investigated by quantifying 16S ribosomal RNA (rRNA) genes of known OHRB. The dechlorination reactions were approximated with Michaelis-Menten kinetics taking into account competitive inhibition. Parameter estimation was performed using AMALGAM (Vrugt and Robinson 2007), a multi-objective, multi-method (ensemble) evolutionary optimization procedure to account for the high correlation among the parameters describing dechlorination kinetics and the existence of multiple solutions. Results showed that all applied chlorinated compounds inhibited 1,2-DCA dechlorination, whereas 1,2-DCA had no pronounced inhibitory effect on the dechlorination of other chlorinated compounds.

Materials and methods
Chemicals 1,2-DCA, chloroethenes, 1,2-DCP, ethene, and propene were purchased from Sigma-Aldrich, and were used directly in the following experiments. Lactate stock solution (1 M) was prepared from 60% sodium DL-lactate solution (Sigma-Aldrich). Other organic and inorganic chemicals were of analytical grade and were used without further purification.

Sediment collection and enrichment set-up
Surface sediment samples (down to 15 cm below surface) were collected from a wetland in Estarreja, Portugal. This site has a long history of contamination with agrochemical and fine chemistry effluents (Carvalho et al. 2005). Sediment enrichment cultures were set up in 120-mL serum bottles using 10 g of wet sediment and 50 mL of an anoxic medium as described previously (Stams et al. 1993). Resazurin (0.005 g/L) and Na 2 S·9H 2 O (0.48 g/L) were added as redox indicator and reducing reagent, respectively. The headspace of the bottles was exchanged with N 2 and CO 2 (80:20%, 140 kPa). Lactate (3 mM) was used as the carbon source and electron donor. The electron acceptors for the sediment cultures were PCE (10 μmol/bottle, designated EA culture) and PCE plus 1,2-DCA (10 μmol/bottle each, designated EB culture). The bottles were sealed with Teflon lined butyl rubber stoppers and aluminum crimp caps (GRACE, MD, USA) and incubated statically in the dark at 20°C. The EA and EB sediment cultures were spiked 21 times with their respective chlorinated electron acceptors during enrichment. After dechlorination of each spike of the chlorinated substrates to ethene, the headspace of the cultures was flushed with N 2 and CO 2 (80:20%) for three times (3 min for each run and 10 min rest in between each flushing cycle) before re-amendment of the respective chlorinated substrate(s). In the last three spikes, the concentration of each chlorinated substrate(s) was increased stepwise from 10 μmol/bottle to 25 and 40 μmol/bottle. To avoid toxicity of the chlorinated compounds in the following experiments, each single chlorinated substrate was added at 25 μmol/bottle unless otherwise stated.
Sediment-free cultures were obtained by transferring the EA and EB sediment cultures as the following: a 5% slurry from the EA and EB sediment cultures was transferred into bottles containing fresh anoxic medium with lactate (5 mM) and single or double chlorinated substrates, i.e., 1,2-DCA either alone or with PCE, cDCE, or VC ( Fig. 1(a)). Each transfer culture was amended with three spikes of the respective chlorinated substrate(s), and then diluted to duplicate cultures (50% inoculum, Fig. 1(b)) and amended with another three spikes of the respective chlorinated substrate(s). In parallel, a 5% slurry from the EA and EB sediment cultures was also transferred into bottles containing 1,2-DCP (10 μmol/bottle). Only the EA transfer culture showed 1,2-DCP dechlorination. Therefore, only this culture was subsequently transferred (5% inoculum) to fresh anoxic medium with either 1,2-DCP, 1,2-DCA, or a mixture of 1,2-DCP and 1,2-DCA ( Fig. 1(c)). After depletion of three spikes of the respective chlorinated substrate, these cultures were diluted to duplicate cultures (50% inoculum, Fig. 1(d)) and amended with another three spikes of the respective chlorinated substrate(s).
The dechlorination pattern of the last (third) spike of chlorinated substrates(s) ( Fig. 1(b, d)) was used for kinetic modeling. To study the relief of inhibition by different chlorinated compounds on 1,2-DCA dechlorination, the cultures amended with double chlorinated substrates were subsequently spiked only with 1,2-DCA.

DNA extraction and quantitative PCR
After dechlorination of each spike of the chlorinated compounds during the last three spikes (amended respectively with 10, 25, and 40 μmol/bottle each compound) in the EA and EB sediment cultures ( Fig. 1), 2 mL slurry samples were taken for DNA extraction. After dechlorination of the third spike of chlorinated compounds in the sedimentfree transfer cultures that were used for kinetic modeling, 2 mL samples were also taken for DNA extraction. DNA was extracted using a DNeasy PowerSoil Kit (QIAGEN, Hilden, Germany) following the manufacturer's instructions. The copy number of 16S rRNA genes was determined by real-time quantitative PCR (qPCR) with primers targeting total bacteria (Muyzer et al. 1993) and OHRB including Dehalococcoides  (Table S1). Assays were performed in triplicates using a CFX384 Real-Time system in a C1000 Thermal Cycler (Bio-Rad Laboratories, USA) with iQ TM SYBR Green Supermix (Bio-Rad Laboratories, USA).

Analytical methods
Chloroethenes, 1,2-DCA, 1,2-DCP, ethene, and propene were analyzed using a gas chromatograph-mass spectrometer (GC-MS) composed of a Trace GC Ultra (Thermo Fisher Scientific, Waltham, MA, USA) equipped with an Rt®-Q-BOND column (Retek, PA, USA) and a DSQ MS (Thermo Fisher Scientific). Helium was used as a carrier gas with a flow rate of 2 ml min −1 . The inlet temperature was 100°C. The split ratio was 30. The temperature program of the column was 40°C hold for 1 min, followed by an increase at 40°C min −1 to 260°C and hold for 1.5 min.

Modeling and parameter estimation
Reductive dechlorination of the chloroethenes in cultures shown in Fig. 1(b) and (d) was modeled using a Michaelis-Menten model with competitive inhibition following Eq. (1) where r n (μM d −1 ) is the dechlorination rate that depends on the respective chlorinated compound, k max,n (μM d −1 ) is the compound-specific maximum utilization rate or degradation constant, K s,n (μM) is the compound-specific half-velocity constant, I n (μM) is the compound-specific competitive inhibition rate, and C n (μM) is the aqueous concentration of the chlorinated compounds. The index i represents the number of parent compounds considered in the dechlorination during competitive inhibition that varied with enrichment set-up. Dihaloelimination of 1,2-DCA to ethene and of 1,2-DCP to propene were also modeled using Eq. (1). In the enrichment cultures containing both chloroethenes/1,2-DCP and 1,2-DCA, competitive inhibition between 1,2-DCA and the other compounds was also included in the models.
As the Michaelis-Menten parameters describing dechlorination are highly correlated, model calibration was performed was not able to dechlorinate 1,2-DCP and hence not used for the kinetic study performed using cultures in box d using AMALGAM (Vrugt and Robinson 2007), a multi-objective, multi-method (ensemble) evolutionary optimization procedure executable in MATLAB. AMALGAM attempts to find for each culture a set of optimal solutions, i.e., an ensemble of optimized kinetic parameters. These solutions have to adhere to the Pareto-principle, which means that all objectives (here, concentrations of the individual dechlorination products) must be met with equal efficiency. The number of Pareto-efficient solutions depends on model complexity and the size of the parameter space as well as the number of model runs and varies for each culture. As such we chose to provide in the results section ranges based on the 50 best parameter combinations of each culture. These were ranked according to their Euclidean distance to the zero-objective point of our n-dimensional space, where n is the number of culture-specific objectives. Using a compromise solution (Werisch et al. 2014;Wöhling et al. 2008), the parameter set representing the solution with the smallest Euclidean distance on the Pareto front to the zeroobjective point was then utilized to create graphical representations as previously outlined (Schneidewind et al. 2014).
All cultures were modeled individually, and models were calibrated on the data obtained from the third spike, during which stable dechlorination patterns were noted. Initially, the cultures with the simplest dechlorination sequences (i.e., VC to ethene or 1,2-DCA to ethene) were modeled, and subsequently model complexity was gradually increased by including additional dechlorination reactions (i.e., cDCE to VC to ethene, etc.). Following this procedure, additional prior information could be used for more complex models to decrease the parameter space (defined by user-set upper and lower boundaries) from which AMALGAM retrieves optimal solutions. Initial boundary values were based on literature (Garant and Lynd 1998;Haest et al. 2010a;Haston and McCarty 1999;Mayer-Blackwell et al. 2016;Schaefer et al. 2009;Schneidewind et al. 2014;Yu and Semprini 2004) for chloroethenes and 1,2-DCA. For 1,2-DCP, boundary conditions were derived from simple Lineweaver-Burk plots (Lineweaver and Burk 1934). Dimensionless, species-dependent Henry coefficients at 20°C were used to account for volatilization of the chlorinated com-
The 16S rRNA gene copy number of Dehalococcoides in EA transfer cultures (~10 8 copies/mL, Fig. 3h-j) was about one order of magnitude higher than in the original EA sediment culture (Fig. 2b), whereas the 16S rRNA gene copy numbers of Dehalococcoides in the EB culture (10 8 -10 9 copies/mL, Fig. 2d) and its transfer cultures (Fig. 4h-j) were similar. In EA transfer cultures, Dehalogenimonas 16S rRNA gene copy numbers were 1-3 orders of magnitude higher in the cultures fed 1,2-DCA, VC, or 1,2-DCA plus chloroethenes (PCE, cDCE, VC) than in the cultures fed only PCE or cDCE ( Fig. 3h-j). The 16S rRNA gene copy numbers of Dehalogenimonas were below 10 3 copies/mL in the EB transfer cultures (Fig. 4h-j), in line with the pattern in the EB sediment culture (Fig. 2d). The 16S rRNA gene copy numbers of Geobacter in the EA and EB transfer cultures were 10 7 -10 8 copies/mL (Figs. 3h-j and 4h-j).

Dechlorination and co-contaminant effect of 1,2-DCA and 1,2-DCP
The original EA and EB sediment cultures had not been amended with 1,2-DCP. To study the co-contaminant effect between 1,2-DCA and 1,2-DCP, EA and EB cultures were first transferred (5% inoculum) to fresh media containing only 1,2-DCP (10 μmol/bottle). During 70 days of incubation, more than 90% of the 1,2-DCP in the EA transfer culture was dechlorinated to propene (Fig. S1), whereas no 1,2-DCP dechlorination was observed in the EB transfer culture (data not shown). Therefore, the EA transfer culture fed 1,2-DCP was used to study the co-contaminant effect (Fig. 1d). In the subsequent transfer cultures, 1,2-DCP dechlorination was stably maintained (cultures EA-T10, Fig. 5c), while 1,2-DCA was also dechlorinated (cultures EA-T8, Fig. 5a). Similar to the cocontaminant effect between 1,2-DCA and chloroethenes, 1,2-DCA dechlorination was inhibited in the transfer cultures concurrently amended with 1,2-DCP, whereas no obvious inhibitory effect of 1,2-DCA on 1,2-DCP dechlorination was observed (culture EA-T9, Fig. 5b). Specifically, the time to complete dechlorination of 1,2-DCA was increased from around 3 days (Fig. S17-S18) to 9 days in the presence of 1,2-DCP (Fig. S19-S20), while 1,2-DCP dechlorination was not inhibited (Fig. S19-S22). When these cultures were amended only with 1,2-DCA, its dechlorination was completed in 2 days (Fig.  S2G), which was strongly enhanced compared with its dechlorination in the presence of 1,2-DCP (same amount of 1,2-DCA was dechlorinated in 9 days) (Fig. 5b). Dehalococcoides (~10 8 16S rRNA gene copies/mL), Geobacter (~10 7 16S rRNA gene copies/mL), and Dehalogenimonas (10 6 -10 7 16S rRNA gene copies/mL) were the predominant known OHRB in these transfer cultures, similar to EA transfer cultures amended with 1,2-DCA and chloroethenes (Figs. 3h-g and 5d). Table 1 provides a summary of parameter ranges for all compounds as compared with the values from previous studies. Those parameter ranges were obtained by modeling dechlorination after the third spike only. Maximum and minimum values of the 50 best parameter combinations for individual cultures are shown in Table S2. K s and I values of the chloroethenes are in the same range as in the previous studies listed in Table 1. A comparison of values of k max is less straight forward as many studies provide k max in units related to the bacterial cell or protein Fig. 2 Reductive dechlorination of 1,2-DCA and PCE in the sediment cultures EA (a) and EB (c) during the last three spikes, and 16S rRNA gene copy numbers of total bacteria, Dehalococcoides, Geobacter, Dehalogenimonas, Dehalobacter, Desulfitobacterium, and Sulfurospirillum at the end of each spike (b, d). The arrows in panels a and c indicate re-spike of the chlorinated substrates. Error bars of the qPCR values indicate standard deviations of triplicate qPCRs performed on one sample of each culture mass. k max estimates for PCE and TCE obtained here are similar to those in Haston and McCarty (1999) and Schneidewind et al. (2014), while k max estimates for cDCE and VC in this study were about one order of magnitude higher.

Dechlorination kinetics
A comparison of parameter estimates for 1,2-DCP dechlorination was not possible due to the lack of kinetic models in literature. Parameter estimates of 1,2-DCA were comparable with those found by Mayer-Blackwell et al. (2016). However, in our study, they span a rather wide range, defined by cultures EA-T9, where concurrent dechlorination of 1,2-DCA and 1,2-DCP occurred (see Fig. S19 and S20 and Table S2), as further discussed below.
In general, modeled and observed results showed a good fit for cultures amended with a single chlorinated compound (e.g., EA-T1, EA-T7, Figs. S3-S4, S15-S16) indicated by low root-mean-square errors (RMSE, data not shown). Model fits decreased (higher RMSE) for the cultures with more complex reaction networks (e.g., EB-T2, Fig. S25-S26). Duplicate batches showed comparable parameter ranges, and parameters differed by less than one order of magnitude. A notable exception is experiment EA-T9 (Fig. S19, S20) where one of the replicate cultures (culture A) showed relatively narrow parameter ranges for 1,2-DCA, whereas parameter ranges for culture B for 1,2-DCA varied by several orders of magnitude.

Discussion
The present study revealed inhibition of 1,2-DCA dechlorination in the presence of chloroethenes and 1,2-DCP using Sulfurospirillum at the end of the third spike in these cultures (h, i, j). Each concentration value represents the average measured from duplicate cultures. The arrows in panels a-g indicate re-spike of the chlorinated substrates. Error bars were not included in panels a-g for clarity. Error bars of the qPCR values indicate standard deviations of triplicate qPCRs performed on one sample of each of the duplicate cultures (n = 2 × 3) organohalide-respiring microbial consortia obtained from a wetland contaminated with agrochemical products. Among the tested chlorinated substrates, cDCE showed the strongest inhibitory effect on 1,2-DCA dechlorination. Dechlorination of 1,2-DCA started only when cDCE was completely depleted (culture EA-T3, Fig. 3c and culture EB-T3, Fig. 4c). This is consistent with previous findings showing that cDCE strongly inhibited 1,2-DCA dechlorination using a continuous enrichment culture containing Dehalococcoides (Mayer-Blackwell et al. 2016). We also noted inhibition of 1,2-DCA dechlorination in the presence of PCE, VC, and 1,2-DCP, further supported by the decreased k max,DCA and increased K s,DCA values compared with the cultures amended with 1,2-DCA only. Notably, the inhibitory pattern of PCE, VC, and 1,2-DCP on 1,2-DCA dechlorination was different from that of cDCE. In cultures containing 1,2-DCA with either PCE, VC, or 1,2-DCP as cocontaminants, the cultures concurrently dechlorinated both amended chlorinated compounds. However, delayed dechlorination of 1,2-DCA due to a decreased dechlorination rate was observed during concurrent dechlorination of 1,2-DCA and PCE (culture EA-T2, Fig. 3b and culture EB-T2, Fig. 4b), probably because of the cDCE production from PCE.
The observed inhibitory effect of VC on 1,2-DCA dechlorination is in contrast to what was reported by Mayer-Blackwell et al. (2016) using a continuous enrichment culture containing Dehalococcoides where VC had a negligible inhibitory effect on 1,2-DCA dechlorination. Interestingly, Mayer-Blackwell et al. (2016) found that long-term exposure of the continuous culture to 1,2-DCA shifted the Dehalococcoides population, leading to , cDCE (EB-T6, f), and VC (EB-T7, g) in the sediment-free enrichment cultures obtained from EB sediment culture, and 16S rRNA gene copy numbers of total bacteria, Dehalococcoides, Geobacter, Dehalogenimonas, Dehalobacter, Desulfitobacterium, and Sulfurospirillum at the end of the third spike in these cultures (h, i, j). Each concentration value represents the average measured from duplicate cultures. The arrows in panels a-g indicate respike of the chlorinated substrates. Error bars were not included in panels a-g for clarity. Error bars of the qPCR values indicate standard deviations of triplicate qPCRs performed on one sample of each of the duplicate cultures (n = 2 × 3) diminished VC dechlorinating ability. In contrast, we found no inhibitory effect of 1,2-DCA on the dechlorination of chloroethenes and 1,2-DCP in batch cultures.
Based on the data presented here, it is likely that the original sediments in EA and EB cultures contained at least two different Dehalococcoides populations. The 1,2-DCP dechlorinating population was maintained during enrichment in the presence of 1,2-DCA (EA culture) but likely lost during incubation in the presence of 1,2-DCA plus PCE (EB culture). This suggests that 1,2-DCA but not PCE was also a growth substrate for the 1,2-DCP dechlorinating Dehalococcoides population, whereas PCE was likely a substrate for the other Dehalococcoides population. The EA transfer cultures were also able to dechlorinate PCE (Fig.  3b, e). This indicates that feeding 1,2-DCA alone in the EA sediment culture also maintained the Dehalococcoides population capable of PCE dechlorination, and therefore, 1,2-DCA was also a growth substrate for the PCE-dechlorinating Dehalococcoides population. Feeding 1,2-DCA plus PCE likely promoted selective growth of the 1,2-DCA/PCE-dechlorinating Dehalococcoides over the 1,2-DCA/1,2-DCP dechlorinating Dehalococcoides. This is also likely the reason for the loss of Dehalogenimonas in the EB culture, as Dehalogenimonas is known to dechlorinate 1,2-DCA and 1,2-DCP via dihaloelimination but does not dechlorinate PCE (Bowman et al. 2013;Moe et al. 2009). Accordingly, Dehalogenimonas was only maintained in the EA transfer cultures amended with substrates known to support Dehalogenimonas growth (e.g., 1,2-DCA or 1,2-DCA plus VC or 1,2-DCA plus 1,2-DCP) (Maness et al. 2012;Martín-González et al. 2015;Moe et al. 2016;Yang et al. 2017) (Fig. 3h-j). Notably, feeding solely PCE or cDCE that are not known growth substrates for Dehalogenimonas decreased Dehalogenimonas 16S rRNA gene copy numbers by 1-3 orders of magnitude (Fig. 3h, i), whereas feeding solely VC or 1,2-DCP that are known substrates to support Dehalogenimonas growth (Bowman et al. 2013;Moe et al. 2009;Yang et al. 2017) did not strongly reduce its 16S rRNA gene copy numbers (Figs. 3j and 4d). Geobacter, which is known to dechlorinate PCE/ TCE (Sung et al. 2006), and 1,2-DCA (Duhamel and Edwards 2007), was stably maintained in both EA (Fig. 3h-j) and EB  sediment-free transfer cultures. Our results indicate that the type of chlorinated substrate drives the selection of OHRB. Likewise, a recent study showed that a Dehalococcoides population shift was driven by different chlorinated electron acceptors in enrichment cultures containing Dehalococcoides and at a contaminated site (Pérez-de-Mora et al. 2018).
Michaelis-Menten kinetics was successfully used to model dechlorination after the third spike when dechlorination followed a concave pattern. Parameter estimates obtained here compare well with those obtained in previous studies (see Table 1). Different kinetic models could be more suited to model dechlorination after the first and the second spikes. Especially dechlorination after the first spike might be modeled more successfully using Monod kinetics (see also Schneidewind et al., 2014) to better account for an apparent lag phase before the onset of dechlorination. However, Monod  -T10, c) in sediment-free cultures derived from the EA transfer culture amended with 1,2-DCP, and 16S rRNA gene copy numbers of total bacteria, Dehalococcoides, Geobacter, Dehalogenimonas, Dehalobacter, Desulfitobacterium, and Sulfurospirillum at the end of the third spike in these cultures (d). Each concentration value represents the average measured from duplicate cultures. The arrows in panels a-c indicate re-spike of the chlorinated substrates. Error bars were not included in panels a-c for clarity. Error bars of the qPCR values indicate standard deviations of triplicate qPCRs performed on one sample of each of the duplicate cultures (n = 2 × 3) modeling was not used in this study due to limited information on microbial interactions (growth and decay patterns).
An interpretation of competitive inhibition is not straightforward from the obtained inhibition constants I (Table S2). For example, cDCE seems to have a pronounced effect on 1,2-DCA dechlorination only at starting concentrations of cDCE (~I cDCE ). As soon as cDCE concentrations drop by a factor ≥ 10, inhibition of 1,2-DCA dechlorination by cDCE becomes much less important. The calibrated inhibition constants suggest that VC is a stronger inhibitor on 1,2-DCA dechlorination than cDCE in all cultures except EB-T2 (I VC is substantially below I cDCE ). However, I VC is largest in cultures EA-T4 and EB-T4 where no higher chlorinated compounds were present. A higher competitive inhibition constant indicates a smaller inhibitive effect. The parameter for VC probably lumps inhibition effects from higher chlorinated parent products when present, resulting in a higher simulated inhibition effect of VC in mixed compound tests.
In addition, interpreting parameter estimates (especially inhibition constants) obtained from cultures with multiple chlorinated compounds proved challenging, as strong crosscorrelation exists among the parameters of the Michaelis-Menten kinetics. For example, we observed comparable K s and k max values for chloroethenes and 1,2-DCP in cultures with and without 1,2-DCA. In contrast, k max,DCA values decreased and K s,DCA values increased in multi-compound cultures compared with the cultures amended with 1,2-DCA only. This hints towards an effect of the chloroethenes and 1,2-DCP on 1,2-DCA dechlorination. However, the best parameter combination did not put this effect in the inhibition constants, but rather in the dechlorination constants of 1,2-DCA itself. Another example is culture EA-T9: a small K s of 1,2-DCA is counteracted by a small inhibition constant of 1,2-DCP in the duplicate cultures. In other words, the obtained parameters suggest that dechlorination of 1,2-DCA could occur at a maximal rate at low substrate concentration, but would then be more inhibited by 1,2-DCP; or in contrast, the maximal dechlorination rate of 1,2-DCA is not attained under the current experimental conditions but its dechlorination would be less inhibited by 1,2-DCP. However, both parameter combinations would yield a proper  Table 3 of the reference f μmol/d and g cells g μmol/d and mg protein h,i,j For 1,2-DCA in batches without 1,2-DCP: 157.9 ≤ k max,DCA ≤ 1428.9, 2.9 ≤ K s,DCA ≤ 1032.5, 2.5 ≤ I DCA ≤ 250.0 fit of the experimental observations, proving the nonuniqueness of the solution (Beven 2001). The use of results from less complex culture set-ups in the refinement of the parameter space, from which AMALGAM choses viable solutions for more complex set-ups (e.g., EA-T1, EA-T4 and EA-T6 for EA-T2), allowed us to reduce the uncertainty on the parameter estimates. However, the problem of non-uniqueness still remains, e.g., the existence of multiple parameter combinations producing an equally good fit. Uncertainty on the parameter estimates arises due to incomplete or insufficient information on the dechlorination processes in the individual cultures (e.g., information on microbial interactions or on necessary micro-nutrients).
In conclusion, the inhibitory effect of chloroethenes and 1,2-DCP on 1,2-DCA dechlorination identified in this study has important implications for understanding the persistence of 1,2-DCA at many contaminated sites. For effective bioremediation of such contaminated sites, it will be necessary to first remove potential inhibitors such as cDCE as well as its parent compounds PCE and TCE, which can delay 1,2-DCA dechlorination and even cause loss of 1,2-DCA/1,2-DCP dechlorinating Dehalococcoides and Dehalogenimonas populations. Further studies are needed to better understand the inhibitory mechanisms. Possible experimental approaches include identification of the genes and enzymes involved in 1,2-DCA dichlorination, study the transcriptional regulation of these genes, and assess potential competitive inhibition of the enzymes. Theoretical modeling experiments that more rigorously look into the effect of the size of the parameter space or the use of different estimation algorithms could further improve our understanding of parameter/model uncertainty.