Stable Isotope Fractionation Caused by Glycyl Radical Enzymes during Bacterial Degradation of Aromatic Compounds

Stable isotope fractionation was studied during the degradationof m-xylene, o-xylene, m-cresol, and p-cresol with two purecultures of sulfate-reducing bacteria. Degradation of all fourcompounds is initiated by a fumarate addition reaction by aglycyl radical enzyme, analogous to the well-studied benzylsuccinatesynthase reaction in toluene degradation. The extent of stablecarbon isotope fractionation caused by these radical-type reactionswas between enrichment factors ( ${varepsilon}$ ) of –1.5 and –3.9

,which is in the same order of magnitude as data provided beforefor anaerobic toluene degradation. Based on our results, ananalysis of isotope fractionation should be applicable for theevaluation of in situ bioremediation of all contaminants degradedby glycyl radical enzyme mechanisms that are smaller than 14carbon atoms. In order to compare carbon isotope fractionationsupon the degradation of various substrates whose numbers ofcarbon atoms differ, intrinsic ${varepsilon}$ ( ${varepsilon}$ _intrinsic) were calculated.A comparison of ${varepsilon}$ _intrinsic at the single carbon atoms of themolecule where the benzylsuccinate synthase reaction took placewith compound-specific ${varepsilon}$ elucidated that both varied on averageto the same extent. Despite variations during the degradationof different substrates, the range of ${varepsilon}$ found for glycyl radicalreactions was reasonably narrow to propose that rough estimatesof biodegradation in situ might be given by using an average ${varepsilon}$ if no fractionation factor is available for single compounds.

INTRODUCTION

Top

Introduction

Many biochemical reactions are known to cause the fractionationof stable isotopes. Molecules consisting of lighter isotopesare utilized preferentially, and consequently heavier moleculesare enriched in the residual substrate pool. Well-known examplesare autotrophic CO₂ fixation by plants (22) and bacterial methanogenesis(14). Within the last few years, several studies investigatedstable isotope fractionation during bacterial degradation ofcontaminants such as toluene (17), benzene (15), and chlorinatedhydrocarbons (12) that were associated with a significant enrichmentof ¹³C in the residual substrate fraction. These findings openedthe door to assessing contaminant degradation at polluted sitesqualitatively, and under certain conditions even quantitatively,and of outlining the future development of the contaminationand its potential impact on the environment and the drinkingwater supply. This concept has been applied successfully inseveral field experiments so far (12, 16, 24, 25).

Basic features of microbial stable isotope fractionation havebeen examined for anaerobic and aerobic bacterial cultures withtoluene as a model substrate (18). It was demonstrated thatthe first enzyme reaction of toluene degradation is the rate-limitingstep and that this reaction is also the key process leadingto isotope fractionation. Isotope effects due to the transportof toluene to and into the cells were negligible. Furthermore,it was shown that various aerobic and anaerobic degradationreactions led to characteristic degrees of fractionation (19).However, it is not possible to deduce the type of the underlyingreaction mechanism from the extent of fractionation, becausefractionation factors ( ${alpha}$ ) can vary significantly between identicalreactions. Studies dealing with the stable isotope fractionationof particular reaction mechanisms have provided informationabout the span of isotope fractionation effects caused by thesemechanisms. Based on theoretical considerations and calculations,it was proposed that the rate-limiting step in the benzylsuccinatesynthase reaction is the addition of fumarate to the benzylradical (10).

Stable isotope fractionation during degradation reactions initiatedby glycyl radical enzymes is of particular interest becausethis reaction mechanism seems to be predominant in the anaerobicdegradation of contaminants such as aromatic hydrocarbons, alkylphenols, and alkanes. Regardless of the electron acceptor employed,anaerobic bacterial toluene degradation proceeds via benzylsuccinateformation in all cases investigated so far (8, 26). The enzymemechanism was first described for anaerobic toluene degradationby denitrifying bacteria, where the enzyme benzylsuccinate synthasecatalyzes the addition of a fumarate molecule to the methylgroup to form benzylsuccinate (3). Fumarate addition and subsequentactivation and ß oxidation convert the former methylgroup to a carbonylic function that acts as an entry port forsingle electrons in the subsequent ring reduction (4). The glycylradical formed in benzylsuccinate synthase and related enzymesderives from 5'-deoxyadenosyl that is a cleavage product ofS-adenosylmethionine (7). The enzyme reaction withdraws onehydrogen atom from a highly conserved cysteine residue of theenzyme. This thiyl radical then takes one hydrogen atom fromthe methyl group of the aromatic substrate, which later is transferredfrom the cysteine residue to the C_ß of the succinylside chain of the produced benzylsuccinate (7, 13) (Fig. 1).

View larger version (11K):
[in this window]
[in a new window]
FIG. 1. Reaction mechanism proposed for benzylsuccinate synthase modified according to the work of Frey (7) for m-xylene, o-xylene, m-cresol, and p-cresol, with R₁, R₂, and R₃ equal to H, CH₃, and OH, respectively. E, polypeptide chain of the enzyme.

We investigated stable isotope fractionation in degradationpathways initiated by a glycyl radical mechanism with two bacterialstrains. Desulfobacterium cetonicum, a toluene-degrading sulfatereducer, was shown to degrade m-cresol and p-cresol via fumarateaddition (20, 21). Enzyme activities in cell extracts showedthat the cresols are probably not converted by the toluene-degradingbenzylsuccinate synthase, but whether the two different cresolisomers are attacked by the same enzyme remains undetermined.

Degradation of m-xylene and o-xylene by the sulfate-reducingstrain OX39 also proceeds via fumarate addition, which was confirmedby the identification of methylbenzylsuccinate derivatives inculture supernatants (B. Morasch, unpublished data). Inductionexperiments with strain OX39 and with m-xylene, o-xylene, andtoluene showed that every substrate needed a specific enzymefor degradation (B. Morasch, unpublished data).

The objective of the present study was to systematically investigatestable carbon isotope fractionation during radical enzyme reactionsof aromatic compounds which are prominent groundwater contaminants.The study should elucidate whether this type of reaction producesconsistent isotope fractionation during degradation. For bettercomparison of isotope fractionations of the various substratemolecules, data for the intrinsic isotope fractionation at themolecular site of the reactions are provided.

MATERIALS AND METHODS

Top

Materials and Methods

Cultivation of bacteria.
The sulfate-reducing strain OX39 was isolated from soil froma site contaminated with BTEX (benzene, toluene, ethylbenzene,and xylene) and polycyclic aromatic hydrocarbons near Stuttgart,Germany (B. Morasch, unpublished data). D. cetonicum (DSM 7267)was taken from the lab culture collection.

Bacteria were cultivated at 30°C in bicarbonate-bufferedfreshwater (strain OX39) or brackish (D. cetonicum) mineralmedium, pH 7.4, with sulfate (10 mM) as the electron acceptor.The medium was prepared under an atmosphere of N₂-CO₂ (80:20)and reduced with Na₂S (1 mM) (27). A sterile, anoxic FeCl₂ solutionwas added to the medium of strain OX39 to a final concentrationof 3 mM.

Strains were grown in 120-ml serum bottles half filled withmineral medium and tightly sealed with Viton rubber stoppers(Maag Technic, Dübendorf, Switzerland). Cultures for isotopefractionation experiments were inoculated with 6 ml of precultures.m-Xylene and o-xylene were injected directly into the culturebottles through the rubber stoppers with microsyringes. m-Cresoland p-cresol were added from aqueous stock solutions (100 mM).Metabolic activity was monitored by observing sulfide production(6).

Hydrocarbon analysis.
Xylene concentrations were determined by high-performance liquidchromatography (Bischoff Chromatography, Leonberg, Germany)with a C₁₈ reversed-phase column (Prontosil, 200 by 3 mm, 3-µmfilm thickness; Bischoff) at 30°C and by UV detection at210 nm with a mix of acetonitrile (Chromasol V super gradientgrade; Fluka, Buchs, Switzerland) and demineralized water (70:30[vol/vol]) as eluents. Cresol concentrations were determinedby using a mix of acetonitrile and ammonium phosphate buffer(100 mM, pH 3.5) (50:50 [vol/vol]) as the eluent. Culture sampleswere diluted 1:5 with ethanol (99.9% gradient grade) and centrifuged(20,000 x g, 5 min) to remove precipitates before analysis.

Isotope analysis.
¹³C/¹²C stable carbon isotope ratios were determined by isotoperatio-monitoring gas chromatography-mass spectrometry. The systemconsisted of a gas chromatograph (HP-5890; Hewlett-Packard Co.,San Diego, Calif.) connected via a combustion unit (gas chromatograph-combustioninterface; Finnigan, Bremen, Germany) with an isotope mass spectrometer(MAT 252; Finnigan). Samples were measured as described previously(18).

Calculations.
The ¹³C/¹²C isotope ratios of the substrate were calculatedas relative ${delta}$ ¹³C values (per mille) according to equation 1 below,where R_sample is the ¹³C/¹²C isotope ratio of the sample andR_std is the isotope ratio of the international Pee Dee Belemnitestandard (11). Kinetic isotope ${alpha}$ ( ${alpha}$ C) were calculated by usingequation 2, which is derived from the Rayleigh equation forclosed systems (11, 23). C_t/C₀ is the fraction of the substrateremaining in the sample at time t. If ln(R_t/R₀) is plotted overln(C_t/C₀), for the time intervals (t), the slope of the linearregression curve gives the ${alpha}$ C as ( ${alpha}$ C – 1). Enrichment factors( ${varepsilon}$ ) are a convenient expression of stable isotope fractionationthat can be retrieved directly from ${alpha}$ C using equation 3 (5).

\mathrm|<||<|\delta|>|^|<|13|>||>|C\mathrm|<| []|<|=|>||>|\left(\frac|<|R\mathrm|<|_|<|sample|>||>||>||<|R\mathrm|<|_|<|std|>||>||>|\mathrm|<||<|\mbox|<|---|>||>|1|>|\right)\mathrm|<||<|\times|>|1,000|>| (1)

(2)

(3)

Derivation of ${alpha}$ _intrinsic.
The general definition of the isotope fractionation factor ${alpha}$ is

(4)
where k_H and k_L are the reactionrate constants of the heavier and lighter isotopomeres, respectively.The intrinsic isotope fractionation factor ${alpha}$ ( ${alpha}$ _intrinsic) is definedas

(5)
with k_H,intrinsic being the reactionrate constant at the position of the molecule where the chemicalreaction takes place. For our model compound toluene, the heavieratom may be located at the methyl group (C-7) or at any positionof the aromatic ring. Therefore, the overall reaction rate constant(k_H) is the sum of the individual reaction rate constants (k_H,i)for the location of the heavy atom at any position (i) of thelabeled compound divided by the number of atoms (n) of the molecule,as shown in general equation 6.

(6)
Inthe specific case of the carbon stable isotope fractionationof toluene [k_H = (k_H,1 + k_H,2 +... + k_H,7)/7], the chemicalreaction (fumarate addition) takes place at the methyl groupof toluene (the C-7 atom). Therefore, only the reaction ratek_H,7 (when the ¹³C atom is located at the methyl group) is subjectto a kinetic isotope effect (k_H,intrinsic). For all other reactionrate constants, k_H,i, the ¹³C atom is located elsewhere in themolecule. With the assumption that secondary isotope effectsare negligible compared to primary isotope effects, the k_H,irate constants for ¹²C atoms at the methyl group and ¹³C atomsat nonreactive positions of the molecule are assumed to be identicalto the reaction rate constant k_L for nonlabeled molecules. Inaddition, we assume that the probability of there being two¹³C carbon atoms per toluene molecule is statistically unlikelydue to the low natural abundance of ¹³C. With the exceptionof k_H,intrinsic, the k_H,i reaction rate constants are thereforeequal to the reaction rate constant k_L for the ¹²C isotopomeres.Equation 6 can therefore be written as

(7)
Equation7 inserted into equation 4 gives

(8)
Equation5 inserted into equation 8 results in

(9)
Theexponent of the Rayleigh equation 2 can therefore be writtenas

(10)
or, according to equation 3,as

(11)

RESULTS AND DISCUSSION

Top

Results and Discussion

Stable ¹³C/¹²C isotope fractionation was determined during thegrowth of D. cetonicum with m-cresol and p-cresol that had initial ${delta}$ ¹³C signatures of –29.32

±0.36

and –27.50

± 0.36

, respectively. Anaerobic m-cresoldegradation caused a strong increase in ${delta}$ ¹³C in the residualsubstrate fraction, which is shown in a representative experimentwhere degradation of 68% of the initial 1,030 µM led toan isotope shift from –29.04

±0.47

to –24.53

± 0.41

(Fig. 2A). The average isotope ${varepsilon}$ of –3.9

± 0.5

for anaerobic m-cresol degradation was determined from fourreplicates (Fig. 3). Anaerobic degradation of p-cresol produceda shift in ${delta}$ ¹³C from –27.11

±0.17

to –23.44

± 0.25

during the degradation of89% of the initial 1,052 µM (data not shown). The smallerincrease in ${delta}$ ¹³C per amount of substrate degraded resulted inan average ${varepsilon}$ of –1.6

± 0.1

calculated from three replicates (Fig. 3). For comparison, theisotope fractionation produced during toluene degradation byD. cetonicum resulted in an ${varepsilon}$ of –2.2

± 0.4

(18).

View larger version (20K):
[in this window]
[in a new window]
FIG. 2. ¹³C enrichment in the residual substrate fraction during the anaerobic degradation of m-cresol ( ${blacktriangleup}$ ) by D. cetonicum (A) and o-xylene (•) by the sulfate-reducing strain OX39 (B). Changes in sulfide concentration ( ${Delta}$ and ${circ}$ ) and ${delta}$ ¹³C ( ${blacklozenge}$ ) were monitored over time. The carbon isotope composition is presented as an average of six individual measurements, with error bars indicating the standard deviations. The diagram shows the results of one representative experiment out of four and of three replicates for m-cresol and o-xylene. d, days.

View larger version (15K):
[in this window]
[in a new window]
FIG. 3. Stable carbon isotope fractionation during the degradation of m-xylene (—) and o-xylene (---) by strain OX39 and of m-cresol (·-·-) and p-cresol (···) by D. cetonicum. Regression lines of the ¹³C/¹²C isotope data over the respective concentrations are plotted according to equation 2, with r² values of 0.968, 0.893, 0.942, and 0.946 for m-xylene, o-xylene, m-cresol, and p-cresol, respectively.

Stable carbon isotope fractionation was determined during thedegradation of m-xylene and o-xylene by strain OX39 in threeindependent growth experiments per substrate. The initial ${delta}$ ¹³Ccarbon signatures of the substrates were –27.66± 0.44 for m-xylene and –28.19± 0.06 for o-xylene. During the degradationof 82% of the initial m-xylene (203 µM), the ¹³C isotopebecame enriched to a level of –24.81± 0.16 in one of three growth experiments(data not shown). Three replicate m-xylene degradation experimentsyielded an average ${varepsilon}$ of –1.8 ±0.2 (Fig. 3). Degradation of an initialo-xylene concentration of 257 µM by strain OX39 resultedin a ${delta}$ ¹³C shift from –28.28 ±0.13 to –24.72± 0.19 when 85% of the substratewas used (Fig. 2B). The average ${varepsilon}$ of three replicates was –1.5± 0.1 (Fig. 3) and thus similar tothe factor found for the degradation of the m-xylene isomer.In previous degradation experiments using a sediment columnfilled with contaminated aquifer material under sulfate-reducingconditions, carbon isotope fractionation by an ${varepsilon}$ of –1.1for o-xylene, which was lower than the ${varepsilon}$ reported here, was determined(24). Furthermore, carbon isotope fractionation resulting inan ${varepsilon}$ of –0.9 ± 0.1was found for the degradation of 2-methylnaphthalene by a sulfate-reducingenrichment culture (9) which also initiates degradation viafumarate addition (2).

By far the most information on isotope fractionation by glycylradical reactions is available concerning toluene degradationunder various redox conditions. In a comparative study, carbonisotope ${alpha}$ determined for bacterial pure cultures using Fe(III),NO₃^–, or SO₄^2– as electron acceptors were foundto be between an ${varepsilon}$ of –1.8 and an ${varepsilon}$ of –1.7 (17). These results were inthe same order as those we obtained for p-cresol, m-xylene,and o-xylene. Another study reported isotope fractionation upontoluene degradation yielding ${varepsilon}$ of –0.8and –0.5 by sulfate-reducing and methanogenicenrichment cultures, respectively, which were significantlylower than those measured in our experiments (1).

To summarize the data available on isotope fractionation duringdegradation reactions employing glycyl radical mechanisms, thelowest fractionation ( ${varepsilon}$ = –0.5) wasfound for toluene degradation under methanogenic conditions,and the strongest fractionation ( ${varepsilon}$ = –3.9)was found for m-cresol degradation by D. cetonicum. Most ${varepsilon}$ obtainedwere about –1.8. It should be emphasizedthat isotope fractionation in the radical reactions of the benzylsuccinatesynthase-type enzymes involves just one methyl carbon atom atthe reactive site. However, determination of ${varepsilon}$ by compound-specificisotope analysis gives overall ${delta}$ ¹³C values of the entire molecule.This discrepancy can be overcome by calculating the fractionationfactor site-specific isotopes ${alpha}$ _intrinsic, which refers to theatom of the target molecule where the enzymatic reaction takesplace (equation 9).

The use of this equation allows the normalization of isotopefractionation of aromatic compounds with different numbers ofcarbon atoms in order to study specific reactions in detail.However, the comparison of intrinsic enrichment factors ( ${varepsilon}$ _intrinsic)upon reactions of the benzylsuccinate synthase type in Table1 shows the same extent of variation of the average of the intrinsicenrichment factors ( ${varepsilon}$ _intrinsic = –12.3 ± 5.8)as the average of the overall enrichment factors ( ${varepsilon}$ = –1.67± 0.86). The standard deviationsare 51 and 47%, respectively (Table 1). Obviously, the specificproperties of every single enzyme catalyzing the same type ofreaction but taking other substrates influence the degree ofisotope fractionation. Thus, the use of the ${varepsilon}$ _intrinsic does notlead to additional information.

View this table:
[in this window]
[in a new window]
TABLE 1. Comparison of compound-specific carbon isotope enrichment factor ${varepsilon}$ to the ${varepsilon}$ _intrinsic calculated for the carbon atoms at the molecular site of the reaction of the respective substrate^a

Based on the observed isotope enrichment factor ${varepsilon}$ , we estimatedthe maximal molecular mass of the substrate that would stillallow us to measure isotope fractionation accurately in thefield. As shown above, the ${varepsilon}$ _intrinsic of enzyme reactions ofthe benzylsuccinate type are more or less equal. However, thelarger the molecular mass of the substrate, the less overallisotope fractionation can be measured, because the isotope effectis diluted with an increasing number of carbon atoms. Takingan average inaccuracy of isotope analysis of ±0.5at the detection limits, the absolute isotope shift needed toreliably detect isotope fractionation would be about 2if about 90% of the substrate was degraded. This value correlatesto an overall ${varepsilon}$ of –0.895. If we takethe average intrinsic ${varepsilon}$ ( ${varepsilon}$ _intrinsic) from Table 1, the presumptionswould be fulfilled for molecules such as toluene with sevencarbon atoms ( ${varepsilon}$ _overall = –1.76) orxylene with eight carbon atoms ( ${varepsilon}$ _overall = –1.54).The calculated borderline would be at molecules of 13 carbonatoms, with ${varepsilon}$ of –0.949. Beyond 14carbon atoms, the overall ${varepsilon}$ (–0.881)is probably too low to allow significant isotope shifts upondegradation to be analyzed. 2-Methylnaphthalene with 11 carbonatoms (Table 1) is close to the analytical limit, but isotopefractionation can be measured under ideal conditions. However,the variations of the intrinsic isotope effects reported aboveemphasize the insecurity of exact measurements in the field,and the limitation for field applications will probably be at11 to 12 carbon atoms.

In situ, the enrichment of heavier isotopes along a groundwaterflow can be taken as qualitative evidence for biodegradation.In addition, the extent of bacterial degradation can be quantifiedby combining the isotope signatures determined in the fieldwith the laboratory-derived ${alpha}$ and using the Rayleigh equation(equation 2) (16, 24). The ${varepsilon}$ for m-cresol, p-cresol, m-xylene,and o-xylene from this study were used together with assumedchanges in the isotope signatures between ${delta}$ ¹³C levels of –30and –20 to calculate the portion ofsubstrate remaining compared to the initial concentration (Fig.4). The percentage of biodegradation is defined according toequation 12 (24):

(12)
To give animpression of how accurate an average ${alpha}$ would be, we used an ${varepsilon}$ of –1.6 ± 0.86from Table 1 to calculate the extent of biodegradation basedon assumed isotope shifts. At small isotope shifts of 1% inthe field, the extent of biodegradation would be 45.5% ±16.15% (Fig. 4). The error becomes smaller as the extent ofbiodegradation increases; e.g., an isotope shift of 5results in an average biodegradation of 95.1% ± 6.87%(Fig. 4). Thus, it might be possible to apply an average ${varepsilon}$ tocalculate the biodegradation of methylated compounds by glycylradical enzymes in contaminated aquifers if the shifts in ${delta}$ ¹³Care sufficiently high.

View larger version (32K):
[in this window]
[in a new window]
FIG. 4. Calculation of the percentage of bacterial biodegradation based on the assumed amount of ${delta}$ ¹³C (per mille) isotope shifts. Curves with symbols are based on the isotope enrichment factors determined here for D. cetonicum degrading m-cresol ( ${blacktriangleup}$ ) and p-cresol ( ${blacktriangledown}$ ) and for strain OX39 degrading m-xylene ( ${circ}$ ) and o-xylene ( ${square}$ ). The error range of biodegradation calculated for the average glycyl radical type enrichment factor is marked in grey.

ACKNOWLEDGMENTS

We are grateful to Stefan Haderlein for continuous support andto Christian Griebler for assistance with the calculation oferror propagations.

This work was financially supported by the Deutsche Forschungsgemeinschaft(grant Schi 180/7) and by the Bundesministerium für Bildungund Forschung (grant 02WT0022).

FOOTNOTES

* Corresponding author. Mailing address: Institut für Grundwasserökologie, GSF-Gesellschaft für Gesundheit und Umwelt, Ingolstädter Landstrasse 1, D-85764 Neuherberg, Germany. Phone: 49-(0)89-31872560. Fax: 49-(0)89-3187-3361. E-mail: rainer.meckenstock{at}gsf.de.

REFERENCES

Top