This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution license (http://creativecommons.org/licenses/by/3.0/).

Chlorinated ethenes are prevalent groundwater contaminants. To better constrain (bio)chemical reaction mechanisms of reductive dechlorination, the position-specificity of reductive trichloroethene (TCE) dehalogenation was investigated. Selective biotransformation reactions (i) of tetrachloroethene (PCE) to TCE in cultures of

_{12}

Chlorinated organic compounds have natural and anthropogenic sources and are represented in nearly every organic chemical class [

Considering the importance of biotic reactions involved in the degradation of chlorinated hydrocarbons, surprisingly little is known about the underlying reaction mechanisms. Even the detailed reductive dehalogenation mechanisms of tetrachloroethene (PCE) and trichoroethene (TCE)—the two most abundant dry cleaning and degreasing agents and notorious groundwater pollutants—remain imperfectly understood. Some bacteria produce

Typical stepwise reaction sequence in microbial dechlorination of PCE leading to less chlorinated ethenes and to non-toxic ethene.

On the most fundamental level, this product formation depends on the initial reaction step of cob(I)alamin (coenzyme B_{12}), the transition-metal cofactor present in reductive dehalogenases [

Specifically, it remains unclear whether the transformation of TCE to

Proposed degradation pathways of TCE catalyzed by cobalamin.

A potential solution are measurements of kinetic isotope effects (KIEs), either on reacting bonds (primary isotope effects) or on adjacent bonds (secondary isotope effects). If only one C-Cl bond in TCE is reactive, a primary isotope effect would occur specifically in this position leading to a pronounced difference to the other positions where only secondary isotope effects would be expected. In contrast, if two positions are reactive, they would take turns in the reaction so that both would reflect a combination of primary and secondary effects. If isotope effects in both positions are compared, a smaller difference would therefore be expected. On the most fundamental level, however, general knowledge about typical primary and secondary isotope effects of chlorine would first be warranted, since so far only very limited knowledge exists on isotope effects in reductive dechlorination of chlorinated ethenes.

To close this research gap, KIEs on specific positions in a target molecule must be determined. Typical techniques are isotope labelling, or determination of position-specific KIEs by NMR-measurements [^{35}Cl and ^{37}Cl) are NMR active, they show broad chemical shifts and poor precision in signal integrations so that position-specific ^{37}Cl/^{35}Cl isotope analysis with NMR is challenging. [

A potential solution to the problem is offered by recent instrumental developments in gas chromatography-isotope ratio mass spectrometry (GC-IRMS). Compounds are separated by gas chromatography and isotopologue ion multiplets of individual chlorinated ethene molecules are recorded simultaneously in dedicated Faraday cups of isotope ratio mass spectrometers [

Even though the recent introduction of chlorine isotope GC-IRMS—combined with routine carbon isotope GC-IRMS—has brought about first dual (carbon and chlorine) isotope effect investigations, these studies have, therefore, only targeted

In this study, we take advantage of the fact that Cl^{−} is released during reductive dechlorination. Since the chlorine substituents of the parent chlorinated ethene end up in different products (^{−} and the less chlorinated ethene), they are subject to different isotope effects (primary effects in the cleaved bond, secondary effects in non-reacting positions). Information on the magnitude of either ^{−} and the less chlorinated ethene). This approach was discussed in previous work [

In this study, a more rigorous evaluation was enabled by (i) investigations of one-step dehalogenation reactions only and (ii) application of an appropriate mathematical framework. Cultures of

Selective reductive dechlorination of PCE to TCE was accomplished in anaerobic biodegradation of PCE with the Firmicute

Mass spectrometry can measure the proportion of different stable isotopes of element E in a given molecule. When looking at a given molecular position, this ratio of heavy isotopes ^{h}E to light isotopes ^{l}E, denoted with R_{0} for the starting material, typically changes during the progress of a reaction to a different ratio R_{t} at time t:

Owing to the kinetic isotope effect (KIE), slower reacting isotopes become enriched during the reaction compared to the original starting material. This KIE_{E} is given by the ratio of rate constants of the light isotope ^{l}k and heavy isotope ^{h}k:

For the sake of simplicity, we stick to this this definition even though, strictly speaking, bacterial transformation gives isotope effects on (V/K)—V: maximum enzyme velocity, K: Michaelis-Menten constant—rather than on elementary rate constants k [

The proportion of _{t}/R_{0}

In the case of compound-specific isotope analysis, _{Compound-Average}_{Compound-Average}

The exponent of Equation (3) can alternatively be called enrichment factor ε and 1/KIE be named fractionation factor α with the relationship [_{Compound-Average}

A _{Compound-Average}

The isotope ratios of the heavy (^{h}E) and light (^{l}E) element in a compound are typically stated relative to reference materials, (Vienna Pee Dee Belemnite (VPDB) for carbon, Standard Mean Ocean Chloride (SMOC) for chlorine) and the isotope ratio of a substance is typically expressed in the delta notation:

Equation (6) can be rearranged to:
_{Compound-Average}^{h}E_{Standard}

This can be introduced into Equation (5) according to:

Equation (8) is typically used in its logarithmic form as the common Rayleigh equation:
^{h}E^{h}E_{0} + 1) ≈ δ^{h}E^{h}E_{0} = ε ln ^{h}E^{h}E_{0} + ε ln

The obtained epsilon (ε) represents the isotope fractionation as a compound average and therefore expresses the average positions

For the products formed during the reaction, an isotopic mass balance must be fulfilled in a closed system:

Here, the reactant contains _{S}^{h}E_{0}^{h}E_{i}^{h}E_{P,i}

In the conversion from PCE to TCE or from TCE to

An equation that allows fitting product isotope trends is obtained by combination of Equation (12) with Equation (9) to yield

The parameter ε_{carbon}

When chlorine isotope values are measured, reductive dechlorination of chlorinated ethenes releases chloride and simultaneously forms the less chlorinated ethene so that two chlorine-containing products are formed at the same time (

In the case of PCE, the molecular positions of all atoms are chemically equivalent so that the same chlorine atoms may potentially end up in TCE or as Cl^{−}. Accordingly, isotopes partition according to the kinetic isotope effects associated with the formation of either product _{i}_{i} = 1/KIE_{i}

With Equation (7), the fractionation factors α_{i}_{Diff}^{−}) and the average secondary isotope effects (in the three molecular positions that become TCE). This equation can be rearranged and simplified according to:

Accordingly, the difference between primary and secondary isotope effects ε_{Diff}^{−} and TCE are always separated by ε_{Diff}

For the reaction of PCE with four chlorine atoms to TCE with three chlorine atoms and Cl^{−} the enrichment trends of these products must follow Equation (17), as well as an isotopic mass balance. According to the derivation in the supporting information, the following equations apply:

When _{PCE}

For interpretation of the experimental PCE degradation data, the product isotope trends of chloride and TCE were fitted in Sigma Plot according to Equations (18) and (19), respectively, with ε_{PCE}_{Diff}

In the case of PCE, ε_{Diff}_{PCE}, in contrast, relates to changes in the four chemically equivalent positions of PCE (average of primary and secondary isotope effects) and is the same in fits of δ^{37}_{PCE}^{37}_{Cl−}^{37}_{TCE}

In contrast to PCE, molecular positions in TCE are not chemically equivalent. In addition, formation of _{β} because it does not experience C-Cl bond cleavage.

Although the two chlorine atoms in ^{37}Cl/^{35}Cl ratios change due to a secondary isotope effect—and one from the geminal α-positions of TCE, in which ^{37}Cl/^{35}Cl ratios may change due to a combination of primary and secondary effects.

One step reductive dechlorination of TCE with possible locations of primary isotope effects (dotted line) and the location of a secondary isotope effect (Cl_{β}).

Since no _{α,E} should be the only reactive position, in analogy to the considerations for Cl_{β} above. Previous studies, however, considered that selective formation of _{TCE}_{−> chloride} = (_{α,E}_{α,Z}_{α,E}, ε_{α,Z}, and ε_{β} are the position-specific isotope effects in the different bonds according to _{PCE}_{TCE->chloride}_{TCE->cis-DCE}

The introduced constants

The parameters ε_{TCE->chloride}_{TCE->cis-DCE}^{37}Cl curves of chloride and _{α,E}, ε_{α,Z} and ε_{β} the position-specific isotope effects, which can tell whether only one, or both positions may react. Specifically, ε_{α,E} is a weighted average of ε_{α,E primary} and ε_{α,E secondary}. Here, ε_{α,E primary} is defined as the primary isotope effect when position E reacts to release chloride (percentage _{α,E secondary} is the secondary isotope effect in E when position Z reacts to release chloride (percentage (1 − _{α,E}_{α,E,prim}_{α,E,sec}_{α,E,sec}_{Diff,α,E}

In the same way ε_{α}_{,Z} is expressed as
_{α,Z}_{α,Z,prim}_{α,Z,sec}_{α,Z,sec}_{Diff,α,Z}

Substitution into Equation (26) gives:
_{TCE}_{−> chloride} = _{α,E}_{α,Z}_{α,E,prim}_{α,E,sec}_{α,Z,prim}_{α,Z,sec}

In the same way, the contribution of _{α}_{,Z} + (1 − _{α}_{,E} in Equation (26)—the one which describes the chlorine atoms of _{α,Z}_{α,E}_{α,Z,prim}_{α,Z,sec}_{α,E,prim}_{α,E,sec}

_{α,E} reacting to Cl) assuming exemplary numeric values for chlorine isotope effects (ε_{primary}_{Cl} = 1.008) and ε_{secondary}_{Cl} = 1.001)), which are further assumed to be identical in both positions (

Representation of parameters ε_{α,E}, ε_{α,Z}, ε_{TCE->chloride}_{TCE->cis-DCE}^{−}_{α,E} reacting to Cl^{−}. Assumed values for primary and secondary isotope effects were ε_{primary}_{secondary}

_{α,E} and ε_{α,Z} vary linearly with _{TCE->chloride}_{TCE‑>cis-DCE}_{α} is stronger in the position from which more chloride is formed with a primary isotope effect; (ii) in addition, more atoms in these positions are released as chloride so that the product curve of chloride more strongly reflects this higher enrichment. The opposite trend can be observed in the product curve of _{TCE‑>chloride}_{TCE->cis-DCE}

_{TCE->chloride}_{α,Z} + (1 − _{α,E} to ε_{TCE‑>cis‑DCE}_{TCE->chloride}_{TCE->cis-DCE}_{TCE->chloride}_{primary}_{TCE‑>cis‑DCE}_{secondary}

As illustrated in the supporting information, the difference K in chlorine isotope signatures of the initially formed chloride and ^{37}Cl_{0,i} of the position(s), from which the respective product is formed (in contrast to PCE, the positions in TCE are not chemically equivalent and may show relevant variations of ^{37}Cl/^{35}Cl between each other); (ii) the kinetic isotope effect from the reaction (primary for Cl^{−}, secondary for

Even though this kinetic isotope effect information is desirable, a direct interpretation like in the case of PCE is not possible because TCE internal isotope distributions cannot experimentally be determined. In contrast to interpretations of the PCE data, insight into position-specific chlorine isotope effects of TCE is therefore _{TCE->chloride}_{TCE‑>cis‑DCE}_{PCE}_{TCE->chloride}_{TCE‑>cis-DCE}

Selective reductive dechlorination of PCE to TCE and chloride was performed with the microorganism _{PCE}_{TCE}^{13}C than the substrate from which it was formed, ^{13}C/^{12}C increased in the remaining reactant pool. Evaluation of reactant isotope data according to the Rayleigh equation [Equation (9)] was therefore an alternative way of determining ε (from reactant data: ε_{PCE}_{TCE}^{13}C curves, ε could alternatively be determined from product isotope data according to Equation (13): ε_{PCE}_{TCE}

Carbon isotope data from degradation of PCE to TCE by

After complete conversion at

For the selective reductive dechlorination of PCE to TCE and chloride by

Enrichment factors ε_{PCE} were obtained from PCE reactant data (−5.0‰ ± 0.1‰) according to Equation (9), from chloride product data (−4.1‰ ± 3.7‰) according to equation (18), and from TCE product data (−5.3‰ ± 0.3‰) according to Equation (19) (uncertainties are 95% confidence intervals). This confirms that the products TCE and chloride reflect the chlorine isotope enrichment trend of the four chemically equivalent positions of PCE from which they are formed. After full conversion at _{Diff}_{Diff}

Chlorine isotope data of PCE degradation to TCE and chloride by ^{37}Cl_{Cl}^{−} were calculated by error propagation including uncertainties in δ^{37}Cl_{PCE}, δ^{37}Cl_{TCE} and

In contrast to the isotope pattern of carbon, these intercepts between PCE as starting material and the instantaneously formed products reveal additional information. In the case of a one-step scenario, the secondary and primary isotope effects are accessible from the intercepts according to _{sec}

Interpretation of chlorine isotope data in a one-step scenario with respect to intercepts between PCE, TCE and chloride of the applied mathematical fits.

In an alternative approach the primary isotope effect may be extracted in higher precision when considering that the enrichment factor of PCE is a weighted average of primary and secondary effects:

The average secondary isotope effect is further given by

In contrast, if a two-step scenario prevails, no

Interpretation of chlorine isotope data in a two-step scenario with respect to intercepts between PCE, TCE and chloride of the applied mathematical fits.

In this case, the observed isotope effect of PCE with −5.0‰ ± 0.1‰ reflects the first step to the intermediate (I). A second rate-limiting step forms TCE and Cl^{−}. The difference between primary and secondary chlorine isotope effects for the second step is obtained from the difference of the intercepts (ε_{Diff}

Independent of the prevailing scenario, the difference between primary and secondary isotope effects in reductive dehalogenation of PCE by

In the case of TCE, the intercepts do not provide useful information about primary and secondary isotope effects because the three positions are distinguishable and TCE may have an unequal isotope distribution, which cannot be directly measured. For instance, if the reactive position in TCE contains more ^{37}Cl/^{35}Cl than the average molecule, a respective “lighter” signature would be found in the formed ^{37}Cl/^{35}Cl. Such a scenario would bias the interpretation of intercepts. The following discussion is therefore only based on the fitted parameters of ε, which are independent of the intercepts in the isotope patterns. Since these parameters reflect enrichment trends in molecular positions of the reactant (even if extracted from product data), they reflect precisely those reaction steps that lead up to and include the first irreversible step. Consequently, they incorporate the initial reaction steps and their interpretation does not require consideration of the one

Since no 1,1-DCE is formed in _{β}_{β}. In contrast, as discussed above, we considered the possibility that both α-chlorines may undergo C-Cl cleavage.

The resultant product isotope enrichment for the formed _{TCE->cis-DCE}

Likewise, data on chloride were fitted according to Equation (24) to yield
_{TCE->Chloride}

Chlorine isotope data of TCE degradation to

As derived above, our considerations show that these parameters reflect enrichment trends in the molecular positions of TCE according to
_{TCE}_{−> chloride} = (_{α,E}_{α,Z}

Assuming that only Cl_{α,E}_{Diff}_{max}/K_{m}) rather than to elementary rate constants [

A more consistent picture arises when the possibility of two reactive positions is considered, as encountered for any 1 >

Our results therefore indicate that the two chlorine substituents in the α-positions are accessible for reductive dechlorination of TCE. In this context, the selective formation of

In previous mechanistic studies, the nucleophilic substitution of a chlorine substituent by the cobalt centre has been discussed as a potential pathway, leading to the formation of a dichlorovinyl-anion in the subsequent step of reduction [

Recent studies pointed out that radical intermediates are involved in reductive dechlorination with cobalamin as model system [

Mechanistic proposal for dechlorination of TCE via a single electron transfer with selective formation of

Based on theoretical considerations, such a conformational change could also occur in a pathway of nucleophilic addition, where the addition of the cobalt centre of the corrinoid would lead to a change of hybridisation from sp^{2} to sp^{3} and create a freely rotating bond (

Mechanistic proposal for dechlorination of TCE via nucleophilic addition followed by

This intermediate was computed by Pratt and van der Donk [

With these considerations, the scenario of nucleophilic addition would combine several unusual attributes, including a strict conformational selectivity at an sp^{3} hybridized carbon centre and an unexpectedly large magnitude of secondary chlorine isotope effects. Nonetheless, this pathway cannot be strictly ruled out as a possible transformation pathway.

With chlorine isotope effects from GC-IRMS measurements, our study brings forward a new perspective for investigating initial mechanisms of reductive chlorinated ethene dehalogenation. Mathematical equations accurately describe reactant and product isotope data, allow extracting position-specific chlorine isotope effect information and may serve as a benchmark for similar evaluations in the future. In reductive biotransformation of PCE, this evaluation allowed us to constrain the difference between primary and average secondary chlorine isotope effects to an unexpectedly large value of −16.3‰ ± 1.4‰ (standard error). This novel insight on primary and secondary chlorine isotope effects in chlorinated ethenes falls outside the range of typical chlorine isotope effects [

Our evaluation further allowed us to test whether one or two C-Cl bonds in TCE were reactive. In the first case (only one C-Cl position reacts), isotope values of chloride would be expected to exclusively reflect the primary isotope effect of the C-Cl bond cleavage, whereas isotope values of

This insight, in turn, significantly constrained the mechanistic scenarios for the initial step in TCE reductive dehalogenation. Direct nucleophilic substitution via dichlorovinyl-anion intermediates could be ruled out for reduction of TCE by

Finally, this insight can support current interpretations of dual element (C and Cl) isotope fractionation during TCE and PCE dehalogenation.

Dual element (carbon and chlorine) isotope plots for degradation of PCE toTCE by

The value of ∆δ^{13}C_{TCE}/∆δ^{37}Cl_{TCE} = 3.4 ± 0.2 (95% confidence interval) during dehalogenation of TCE by _{12}. The present study complements this finding with underlying mechanistic insight. Taken together, our results suggest that this insight may also be of relevance for reductive dehalogenation by vitamin B_{12}. In addition, our calculations can explain the different numerical values in the TCE degradation experiment when ∆δ^{13}C/∆δ^{37}Cl of TCE is compared to ∆δC^{13}C/∆δ^{37}Cl of _{TCE,carbon}_{TCE,chlorine}_{TCE,carbon}_{TCE‑>cis‑DCE}_{chlorine}. The observation that ε_{TCE->cis‑DCE,chlorine}_{TCE,chlorine}

In the case of PCE, the dual element isotope slope of ∆δ^{13}C_{PCE}^{37}Cl_{PCE}^{13}C_{TCE}^{37}Cl_{TCE}_{PCE}_{TCE,chlorine}_{chlorine}_{carbon}_{TCE,carbon}_{TCE,chlorine}^{13}C_{TCE}^{37}Cl_{TCE}^{13}C_{PCE}/ ∆δ^{37}Cl_{PCE}_{chlorine}_{carbon}^{13}C_{PCE}^{37}Cl_{PCE}

Supporting Information can be accessed at:

A1 MATERIALS and METHODS: Experimental protocols for biodegradation of PCE with ^{37}cl_{cl}− and propagated errors; A6 equations to calcuate δ^{37}cl_{cl}− and propagated errors; A7 all fitting procedures and regression reports.

We thank Kris McNeill from the ETH Zürich and Stefan Haderlein from University of Tübingen for helpful discussions and critical comments. We further thank Wolfgang zu Castell-Rüdenhausen and Michael Hagen from the Institute of Bioinformatics, Helmholtz Zentrum Munich, for helpful initial discussions on alternative approaches to mathematical fitting. This work was supported by the German Research Foundation (DFG), EL 266/3-1, as well as by the Initiative and Networking Fund of the Helmholtz Association. A.B. was supported by a fellowship of the Minerva Foundation, Max-Planck-Gesellschaft.

SC, AB and ME designed the study, SC performed the research and analyzed the data. OSS contributed isotope standards, and FEL provided materials. ME and SC derived the mathematical framework for data interpretation and wrote the manuscript. All authors contributed to manuscript preparation and approved the final version.

The authors declare no conflict of interest.

_{12}-catalyzed reductive dehalogenation of perchloroethylene and trichloroethylene

^{35}Cl NMR Spectroscopy and Exact Spectral Line-Shape Simulations

^{13}C/

^{12}C,

^{37}Cl/

^{35}Cl)