Anaerobic bacterial metabolism of hydrocarbons


*Corresponding author. Tel.: +49 (761) 203-2774; Fax: +49 (761) 203-2626; E-mail


The capacity of some bacteria to metabolize hydrocarbons in the absence of molecular oxygen was first recognized only about ten years ago. Since then, the number of hydrocarbon compounds shown to be catabolized anaerobically by pure bacterial cultures has been steadily increasing. This review summarises the current knowledge of the bacterial isolates capable of anaerobic mineralization of hydrocarbons, and of the biochemistry and molecular biology of enzymes involved in the catabolic pathways of some of these substrates. Several alkylbenzenes, alkanes or alkenes are anaerobically utilized as substrates by several species of denitrifying, ferric iron-reducing and sulfate-reducing bacteria. Another group of anaerobic hydrocarbon degrading bacteria are ‘proton reducers’ that depend on syntrophic associations with methanogens. For two alkylbenzenes, toluene and ethylbenzene, details of the biochemical pathways involved in anaerobic mineralization are known. These hydrocarbons are initially attacked by novel, formerly unknown reactions and oxidized further to benzoyl-CoA, a common intermediate in anaerobic catabolism of many aromatic compounds. Toluene degradation is initiated by an unusual addition reaction of the toluene methyl group to the double bond of fumarate to form benzylsuccinate. The enzyme catalyzing this first step has been characterized at both the biochemical and molecular level. It is a unique type of glycyl-radical enzyme, an enzyme family previously represented only by pyruvate-formate lyases and anaerobic ribonucleotide reductases. Based on the nature of benzylsuccinate synthase as a radical enzyme, a hypothetical reaction mechanism for the addition of toluene to fumarate is proposed. The further catabolism of benzylsuccinate to benzoyl-CoA and succinyl-CoA appears to occur via reactions of a modified β-oxidation pathway. Ethylbenzene is first oxidized at the methylene carbon to 1-phenylethanol and subsequently to acetophenone, which is then carboxylated to 3-oxophenylpropionate and converted to benzoyl-CoA and acetyl-CoA. Anaerobic mineralization of alkanes involves an oxygen-independent oxidation to fatty acids, followed by β-oxidation. In one strain of an alkane-mineralizing sulfate-reducing bacterium, the activation appears to proceed via a chain-elongation, possibly by addition of a C1-group at the terminal methyl group of the alkane. Finally, aspects concerned with the regulation and ecological significance of anaerobic hydrocarbon catabolic pathways are discussed.


Hydrocarbons can be classified into saturated compounds (aliphatic and alicyclic alkanes), compounds containing C–C double-bonds (alkenes), compounds with C–C triple-bonds (alkynes), and the mono- and polycyclic aromatic hydrocarbons. Details regarding the chemistry of these compounds can be found in most organic chemistry textbooks (e.g. [1–4]). Most hydrocarbon compounds exhibit high homolytic and heterolytic dissociation energies of their C–H and C–C bonds and weak chemical reactivity. Therefore, hydrocarbons do not participate in acid-base reactions in aqueous systems, as indicated by their extremely high theoretical pKa values. Protonation of alkanes (yielding carbenium ions and molecular hydrogen) requires the use of super acids. Addition reactions to unsaturated C–C bonds occur at the double- and triple-bonds of alkenes and alkynes, but not at aromatic rings.

Certain redox reactions of alkenes occur under relatively mild conditions; for example, their reduction to alkanes with hydrogen by catalytic hydrogenation. Similarly, methyl or methylene groups directly attached to aromatic rings can be catalytically oxidized to the corresponding carboxyl- or carbonyl-groups, respectively. The most common reaction of hydrocarbons is their combustion as fuel with oxygen to CO2 and water. These reactions, which proceed via radical intermediates, are initiated by oxygenation reactions, involving molecular oxygen as a direct reactant. Enzyme-catalyzed oxygenations, which also occur via radical intermediates, have until recently been the only known initial reactions for degradation of alkanes and aromatic hydrocarbons in biological systems (see below).

Radical mechanisms are also involved in some known oxygen-independent chemical reactions of hydrocarbons. Alkanes and alkyl side chains of aromatic hydrocarbons can be chemically ‘cracked’ to smaller alkanes and alkenes by pyrolysis, or halogenated with elemental halogens in the presence of light. Both reactions involve free radical intermediates. These reactions are technically performed under conditions which are not compatible with biological systems. Yet, it was recently elucidated that some bacteria actually employ oxygen-independent radical reactions to make hydrocarbons available as substrates (see below).

Aromatic hydrocarbons can be chemically derived through electrophilic substitution. Well known examples of substitution reactions at the aromatic ring are halogenation, nitration, sulfonation, diazonium coupling and reactions with carbon electrophiles, e.g. carbocations generated in Friedel–Crafts reactions. These reactions normally require conditions which probably can not exist within living organisms.

Hydrocarbons are widespread in the environment. Their major industrial source is petroleum and its associated natural gases, formed geochemically from biomass under conditions of high pressure and temperature. However, significant amounts of hydrocarbons are also formed by biological processes. For example, methane is produced as a metabolic end product by methanogenic bacteria. Biosynthesis of high-molecular alkanes by decarbonylation of the corresponding (n+1) aldehydes has been reported for a marine alga [5]. The simplest alkene compound, ethylene, is synthesized and released by higher plants as a ripening hormone; in addition, some ethylene-producing bacteria and fungi are known [6]. High-molecular mass alkenes are found as constituents of the cuticulas of insects and higher plants and serve as protection against loss of water as well as sex pheromones [7]. They are formed either by decarboxylation of unsaturated fatty acids [8]or by monoxygenase-catalyzed conversion of unsaturated aldehyde precursors to alkenes and CO2[9]. Another well known class of natural hydrocarbons, which often contain double bonds, are the isoprenoids, e.g. carotenoids and terpenes of many plants, insects and microorganisms [10]. Anaerobic catabolism of isoprenoids is reviewed by Hylemon and Harder in this issue. Certain aromatic hydrocarbons are also formed biologically. Low concentrations of toluene have been detected in pristine environments, such as the anaerobic hypolimnia of lakes [11]; it originates from phenylalanine degradation by several species of anaerobic bacteria. These bacteria first oxidize phenylalanine to phenylacetate, which is then decarboxylated [12, 13]. Even biological formation of naphthalene has recently been reported in some plant and animal species [14].

Catabolism of hydrocarbons has long been considered as a strictly oxygen-dependent process. Common aerobic hydrocarbon-utilizing organisms are found among fungi and bacteria. These microorganisms are capable of metabolizing virtually all naturally formed and a wide range of industrially produced hydrocarbons. In aerobic organisms, the initial attack of hydrocarbons always requires molecular oxygen as a co-substrate. The first enzymes in the metabolic pathways of alkanes are monooxygenases, while aromatic hydrocarbons are attacked by either monooxygenases or dioxygenases. These enzymes incorporate hydroxyl groups, derived from molecular oxygen, into the aliphatic chain or the aromatic ring. The alcohols formed from aliphatic hydrocarbons are then oxidized to the corresponding acids; the phenolic compounds generated by ring hydroxylation of aromatic hydrocarbons are direct precursors for oxidative ring cleavage (last reviewed in [15]).

As demonstrated throughout the last decade of microbiological research, particular microorganisms are also able to catabolize hydrocarbon compounds under anaerobic conditions. Hydrocarbons that can be degraded anaerobically include aliphatic alkenes and alkanes with chain lengths of 6-20 carbon atoms, monocyclic alkylbenzenes, such as toluene, ethylbenzene, propylbenzene, p-cymene, xylene- and ethyltoluene-isomers, as well as benzene and naphthalene. Some hydrocarbons that are degraded anaerobically by pure bacterial cultures are shown in Fig. 1. Obviously, bacteria capable of this metabolic capacity must have developed alternative, oxygen-independent reactions for the initial attack of their hydrocarbon substrates. No organism which mineralises hydrocarbons containing less than six C-atoms anaerobically has yet been discovered.

Figure 1.

Structures of some hydrocarbons which are metabolized anaerobically by pure bacterial cultures.

2Bacteria capable of metabolizing hydrocarbons under anoxic conditions

Hydrocarbons are highly reduced organic molecules. In chemotrophic organisms, the reducing equivalents generated during transformation of hydrocarbons to metabolic intermediates need to be transferred to an electron acceptor with a more positive redox potential to allow energy conservation for growth. Based on our present biochemical knowledge, energy conservation from hydrocarbon metabolism by a chemotrophic organism in pure culture is not conceivable in the absence of an external electron acceptor. In the absence of oxygen as terminal electron acceptor, energy conservation may be accomplished by anaerobic respiration with nitrate, ferric iron or sulfate (Table 1). Accordingly, all anaerobic hydrocarbon degrading strains, which are available as pure cultures, are either denitrifying, ferric iron-reducing or sulfate-reducing bacteria (Table 2). In addition, some bacteria may dispose of the reducing equivalents recovered from hydrocarbon oxidation by reducing protons to hydrogen, but this is thermodynamically feasible only in syntrophic association with hydrogen-consuming microorganisms, such as methanogens (see below). None of the ‘proton-reducing’ hydrocarbon-degrading bacteria is available in a defined coculture or in pure culture. A last group of anaerobic bacteria, which may principally use hydrocarbons as carbon- and electron sources, are the anoxygenic photosynthetic bacteria. However, although these bacteria are long known to metabolize polar aromatic compounds, no hydrocarbon-metabolizing phototrophic bacteria are yet reported.

Table 1.  Stoichiometric equations of anaerobic bacterial toluene oxidation coupled to the reduction of different electron acceptors
(1) Denitrifying bacteria:
C7H8+7.2 NO3+0.2 H+7 HCO3+3.6 N2+0.6 H2O
  ΔG°′=−3554 kJ (mol toluene)−1
(2) Iron(III) reducing bacteria:
C7H8+94 Fe(OH)37 FeCO3+29 Fe3O4+145 H2O
  ΔG°′=−3398 kJ (mol toluene)−1
(3) Sulfate reducing bacteria:
C7H8+4.5 SO2−4+3 H2O7 HCO3+2.5 H++4.5 HS
  ΔG°′=−205 kJ (mol toluene)−1
(4) Methanogenic consortia: reactions catalyzed by ‘proton-reducing’ bacteria (a) and methanogens (b, c).
(a) C7H8+9 H2OHCO3+3 H3C-COO+4 H++6 H2
  ΔG°′=+166 kJ (mol toluene)−1
(b) 6 H2+1.5 HCO3+1.5 H+1.5 CH4+4.5 H2O
  ΔG°′=−203 kJ (6 mol H2)−1
(c) 3 H3C-COO+3 H2O3 CH4+3 HCO3
  ΔG°′=−93 kJ (3 mol acetate)−1
Sum: C7H8+7.5 H2O4.5 CH4+2.5 HCO3+2.5 H+
  ΔG°′=−131 kJ (mol toluene)−1
Table 2.  Overview of bacterial strains available in pure culture, which are capable of anaerobic hydrocarbon degradation
Species/strain [reference]Hydrocarbons metabolizedOther key substrates (other than benzoate)
  1. aC.J. Krieger, M. Reinhard and A.M. Spormann, unpublished results.

  2. bDoes not grow with benzoate.

I. Denitrifying bacteria (β-subclass of proteobacteria)
 Thauera aromatica K172 [42]ToluenePhenol, p-cresol, anthranilate, phenylalanine
 Thauera aromatica T1 [73]Toluenep-Cresol, 3-methylbenzoate
 Azoarcus sp. strain T [16]and unpublishedaToluene, m-xylenep-Cresol, cyclohexanecarboxylate
 Azoarcus tolulyticus Tol4 (and other strains) [74]TolueneNot reported
 Azoarcus tolulyticus Td15 [74]Toluene, m-xyleneNot reported
 Strain ToN1 [21]ToluenePhenol, p-cresol, phenylacetate
 Strain EbN1 [21]Toluene, ethylbenzeneAcetophenone, phenylalanine
 Strain PbN1 [21]Ethylbenzene, propylbenzeneAcetophenone, propiophenone, phenylacetate, phenol
 Strain EB1 [22]EthylbenzeneAcetophenone, phenylacetate, phenol
 Strain mXyN1 [21]Toluene, m-xylene3-Methylbenzoate, p-cresol
 Strain T3 [75]TolueneNot reported
 Strain M3 [75]Toluene, m-xyleneNot reported
 Strain mCyN1 [23]Toluene, p-ethyltoluene, p-cymenep-Cresol, phenylalanine, p-ethylbenzoate, p-isopropylbenzoate
 Strain mCyN2b[23]p-Cymenep-Ethylbenzoate, p-isopropylbenzoate
II. Ferric iron reducing bacterium (δ-subclass of proteobacteria)
 Geobacter metallireducens GS15 [25]ToluenePhenol, p-cresol, phenylacetate
III. Sulfate reducing bacteria (δ-subclass of proteobacteria)
 Desulfobacula toluolica Tol2 [32]ToluenePhenylacetate
 Strain PRTOL1 [33]Toluenep-Cresol, phenylacetate
 Desulfobacterium cetonicum[34]TolueneNot reported
 Strain oXyS1 [34]Toluene, o-xylene, o-ethyltolueneo-Methylbenzoate, benzylsuccinate
 Strain mXyS1 [34]Toluene, m-xylene, m-ethyltoluene, m-cymenem-Methylbenzoate
 Strain Hxd3 [29]Alkanes (C12–C20), 1-hexadecene1-Hexadecanol, 2-hexadecanol, fatty acids (C4–C18)
 Strain Pnd3 [30]Alkanes (C14–C17), 1-hexadecene1-Hexadecanol, fatty acids (C3–C18)
 Strain TD3 [31]Alkanes (C6–C16)Fatty acids (C4–C18)

The denitrifying species described mineralize a variety of alkylbenzenes, including toluene, m-xylene, ethylbenzene, propylbenzene, p-ethyltoluene and p-cymene (for structures see Fig. 1) [16–23]. Some of these strains were formerly classified as Pseudomonas sp.; based on 16S rDNA sequence comparison, they are now affiliated with the genera Thauera and Azoarcus within the β-subclass of the Proteobacteria (Table 2). The known strains exhibit a wide substrate spectrum for polar aromatic compounds, but are restricted to few aromatic hydrocarbons. Only a few of the isolated strains use more than one aromatic hydrocarbon compound as substrate (Table 2). Denitrifying bacteria grow relatively fast on alkylbenzenes: the maximum growth rates on toluene reach 0.12 h−1 (doubling time 6 h), which represents 70% of the growth rate obtained when benzoate serves as substrate [24]. The reaction balance for toluene oxidation under denitrifying conditions is given in Table 1. Hydrocarbon oxidation at the expense of nitrate yields a high amount of free energy (Table 1).

Only one species of a ferric iron-reducing bacterium which degrades an aromatic hydrocarbon, Geobacter metallireducens (Table 2), has been reported. Toluene is the only metabolisable hydrocarbon for this bacterium, which is also capable of anaerobic degradation of several other polar aromatic substrates [25]. The stoichiometry of toluene oxidation by G. metallireducens is shown in Table 1. The theoretical free energy yield of hydrocarbon utilization by ferric iron-reducing bacteria is relatively high (Table 1); growth of these bacteria is probably limited by the availability of the insoluble Fe(OH)3. Reports on degradation of benzene under ferric iron reducing conditions indicate that the metabolic potential of ferric iron reducing bacteria is probably wider than demonstrated so far [26, 27]. Other oxidized metals, e.g. Mn(IV), may also serve as electron acceptors for toluene-metabolizing bacteria, as suggested by experiments in sediments [28].

As shown in Table 1, several pure cultures of sulfate-reducing bacteria were isolated which were capable of utilizing alkanes and alkenes (Table 2; [29–31]). In addition, four strains of alkylbenzene-metabolizing sulfate-reducing bacteria have been described (Table 2). Two of those were restricted to toluene as the only hydrocarbon substrate [32, 33]. Two new strains were recently isolated on o-xylene and m-xylene, respectively; the o-xylene-degrading strain also metabolises toluene and o-ethyltoluene, the m-xylene-degrading strain toluene, m-ethyltoluene and m-cymene (for structures see Fig. 1) [34]. The reaction equation for toluene oxidation by sulfate-reducing bacteria is given in Table 1. This process yields only small amounts of free energy (Table 1) and relatively low growth rates of the bacteria are observed [30, 32, 33]. The metabolic potential of sulfate-reducing bacteria for degradation of alkylbenzenes is probably much wider than presently known; even benzene and naphthalene appear to be degraded in anaerobic sediments under sulfate-reducing conditions [35, 36].

Finally, a fourth group of anaerobic bacteria utilizing various alkylbenzenes have been found in syntrophic association with methanogenic archaea [37, 38]. Hydrocarbons are converted to CO2, H2, and acetate. This disproportionation is thermodynamically feasible only if the steady state concentration of hydrogen (and possibly also of acetate) is kept at a low level. This is achieved by hydrogen and acetate consumption by methanogens. The low free energy available from conversion of alkylbenzenes to CO2 and methane (Table 1) must sustain all organisms involved in the syntrophic association. The hydrogen partial pressure, which just allows growth of both organisms of the consortium, is in the range of 1 Pa (10−5 atm). As a result, the syntrophic consortia disproportionate aromatic hydrocarbons to methane and bicarbonate (see Table 1).

3Anaerobic catabolism of toluene

Interest in anaerobic toluene mineralization resulted initially from the observation that toluene was readily degraded in sewage sludge and in anaerobic hydrocarbon contaminated sediments [37, 39, 40]. In an anaerobic enrichment culture with crude oil under conditions of sulfate reduction, toluene was the most rapidly consumed of the utilisable alkylbenzenes [31, 41]. Since 1990, pure cultures of bacteria mineralizing toluene anaerobically have been isolated. These include denitrifying, sulfate-reducing, as well as iron-reducing bacteria that belong to the β- and δ-subclasses of the Proteobacteria (e.g. [18–20, 25, 32, 33, 42]).

Benzoate (or its CoA-thioester) has been recognized as a central intermediate in anaerobic mineralization of numerous aromatic compounds. In many anaerobic toluene-mineralizing cultures, benzoate has been detected as a transiently excreted product [17, 43–46]. Thus, the initial series of reactions in anaerobic toluene degradation apparently involves the conversion of toluene to benzoate (or benzoyl-CoA). The reactions leading to de-aromatization of benzoyl-CoA and further degradation of the alicyclic intermediates are discussed by Harwood et al. in this volume.

3.1Pathway of anaerobic toluene conversion to benzoate

Based on in vitro studies with two denitrifying bacteria, Thauera aromatica and Azoarcus sp. strain T [24, 47], a pathway of toluene oxidation to benzoyl-CoA was proposed (Fig. 2).

Figure 2.

Proposed pathway of anaerobic oxidation of toluene to benzoyl-CoA. The initial reaction is the addition of fumarate to the methyl group of toluene, catalyzed by benzylsuccinate synthase (enzyme 1). The hypothetical further steps are analogous to β-oxidation of α-methyl-branched fatty acids, with a CoA-transferase (enzyme 2) initiating the pathway. The proposed enzymes catalyzing these further reactions are given below (3–6).

The first reaction is the addition of toluene to fumarate to form benzylsuccinate (Fig. 2). Free benzylsuccinate was identified as a transient intermediate of anaerobic toluene oxidation by both in vitro experiments and in vivo isotope trapping experiments in T. aromatica[24]. Kinetic studies of benzylsuccinate formation from toluene and fumarate in Azoarcus sp. strain T demonstrated that the in vitro rate of benzylsuccinate formation was about 30% of the in vivo rate of toluene consumption [47], suggesting that this reaction actually represents the first step in anaerobic toluene mineralization. Accumulation of benzylsuccinate in culture media was observed in earlier studies of anaerobic toluene mineralization, but was interpreted as the result of dead-end metabolism [33, 43–46].

Oxidation of benzylsuccinate to benzoyl-CoA was proposed to proceed via a modified β-oxidation pathway (Fig. 2). The existence of this pathway is supported by several experimental findings: (1) CoA-dependent conversion of benzylsuccinate to benzoyl-CoA did not require ATP [24]; (2) E-phenylitaconate (or the CoA thioester) was identified as an oxidation product of benzylsuccinate [47]; (3) benzylsuccinate oxidation to E-phenylitaconate and benzoyl-CoA was significantly increased when succinyl-CoA served as the source of CoA [48]; and (4) benzylsuccinate:succinyl-CoA CoA-transferase activity was detected in toluene-grown cells of T. aromatica (C. Leutwein and J. Heider, unpublished). Therefore, activation of benzylsuccinate to the CoA-thioester is apparently catalyzed by a CoA-transferase rather than an ATP-dependent CoA ligase. Benzylsuccinyl-CoA is thought to be subsequently oxidized to E-phenylitaconyl-CoA by a benzylsuccinyl-CoA dehydrogenase. The next three postulated enzymatic reactions are hydration to 2-carboxymethyl-3-hydroxy-phenylpropionyl-CoA, oxidation to benzoylsuccinyl-CoA, and thiolytic cleavage to benzoyl-CoA and succinyl-CoA (Fig. 2). None of these postulated intermediates or enzyme activities have yet been detected in cell-free extracts. Considering the expected redox potentials in analogy to β-oxidation, the reducing equivalents generated from oxidation of benzylsuccinyl-CoA to E-phenylitaconyl-CoA (Fig. 2, enzyme 3) are probably transferred to the quinone pool, whereas the 3-hydroxyacyl-CoA dehydrogenase (Fig. 2, enzyme 5) probably reduces NAD+ or NADP+. The third two-electron oxidation step required for toluene oxidation to benzoyl-CoA is accomplished by utilizing the oxidized co-substrate fumarate in the first step, which is released in the reduced form as succinate (Fig. 2). Regeneration of fumarate from succinate by succinate dehydrogenase involves another transfer of reducing equivalents to the quinone pool.

Elucidation of the unusual pathway of anaerobic toluene degradation described above in T. aromatica and Azoarcus strain T raises the question of whether this pathway is unique to these species or a general mode for anaerobic toluene mineralization. Benzylsuccinate formation from toluene and fumarate was also found in two toluene-mineralizing sulfate-reducing strains [49, 50]. The sulfate-reducing strains represent a phylogenetically distant group of bacteria (δ-subclass of proteobacteria), suggesting that anaerobic toluene metabolism generally proceeds via benzylsuccinate.

3.2Benzylsuccinate synthase, the enzyme catalyzing the initial reaction in anaerobic toluene mineralization

Benzylsuccinate formation from toluene and fumarate appears to be the initial reaction in anaerobic toluene oxidation [24, 47]. This enzymatic reaction, catalyzed by benzylsuccinate synthase, has several novel features. Firstly, enzymatic toluene addition to fumarate does not involve a net redox reaction. This is in contrast to all known toluene-transforming oxygenases, which catalyze an oxidation of the hydrocarbon substrate [51, 52]. Secondly, benzylsuccinate formation represents a unique biochemical reaction of forming a carbon–carbon bond, which differs from carboxylations, aldolase-type and oxo-acid lyase-type reactions [53].

3.2.1Properties of benzylsuccinate synthase

The novel enzyme benzylsuccinate synthase catalyzes the addition of toluene to fumarate (Fig. 2). Benzylsuccinate synthase activity was extremely sensitive to inactivation by air, exhibiting a half life time of only 20–30 s. The enzyme was reversibly inhibited by the substrate analogs benzyl alcohol, benzaldehyde or phenylhydrazine [24, 54]. Benzylsuccinate synthase was purified under anoxic conditions from toluene-grown cells of T. aromatica. The pure enzyme was extremely oxygen-sensitive and did not require further co-substrates. The enzyme contained a redox-active flavin cofactor, but no iron–sulfur clusters. It had a native mass of 220 kDa and consisted of three subunits of apparent masses of 98 kDa (α), 8.5 kDa (β) and 6.4 kDa (γ). Based on the native mass, an α2β2γ2 composition of benzylsuccinate synthase is assumed. Half of the α-subunits showed a C-terminal truncation of 4 kDa, producing an α′-fragment of 94 kDa with identical N-terminal amino acid sequences.

The amino acid sequences of the subunits of benzylsuccinate synthase were derived from the corresponding genes (see below). The α-subunit showed strong similarity to enzymes containing glycyl radicals, which have so far been represented only by pyruvate-formate lyase and anaerobic ribonucleotide reductase (Sawers, this issue). Based on the catalyzed reactions and the recorded similarity scores, benzylsuccinate synthase represents a new subclass of the glycyl-radical containing enzymes. The presence of a glycyl radical in benzylsuccinate synthase was inferred from characterization of the observed truncation product (α′) of the large subunit, which is thought to be generated by oxygenolytic cleavage of the polypeptide backbone at the site of the glycyl radical [54]. The observed pattern of nearly equal amounts of intact and truncated α-subunits suggests that only one of the two large subunits of the benzylsuccinate synthase holoenzyme carries a glycyl-radical. The same type of oxygenolytic cleavage is known to exist for the other glycyl-radical enzymes ([55, 56]; Sawers, this issue).

3.2.2Proposed reaction mechanism of benzylsuccinate synthase

A reaction mechanism of benzylsuccinate synthase involving radical intermediates is suggested by a number of experimental observations: firstly, benzylsuccinate synthase was identified as a possible new glycyl-radical enzyme by sequence similarity [54, 57]. Secondly, the predicted glycyl radical site was identical with the determined site of oxygenolysis [54]. Thirdly, the proposed radical-carrying glycine and a conserved cysteine of benzylsuccinate synthase, which are also part of the active center of pyruvate formate-lyase [58, 59], were essential for growth on toluene, as shown by mutagenesis and genetic complementation studies [57](see below). Finally, GC/MS analysis of benzylsuccinate formed from [methyl-D3]toluene showed that the deuterium atom abstracted from the methyl group of toluene is retained in the succinyl moiety of benzylsuccinate [47]. Considering these observations, a radical reaction mechanism is proposed as shown (Fig. 3).

Figure 3.

Proposed reaction mechanism of benzylsuccinate synthase. The radical-containing enzyme produces a benzyl radical from toluene, which adds to fumarate to form a benzylsuccinyl radical. The enzyme converts the product radical to benzylsuccinate, and the radical form of benzylsuccinate synthase is regenerated. The H-atoms originating from the methyl group of toluene are highlighted to indicate their retention in the benzylsuccinate formed.

The proposed reaction mechanism of benzylsuccinate synthase begins with the radical-containing, activated form of the enzyme (Fig. 3). It appears plausible that an enzyme-based radical of active benzylsuccinate synthase first abstracts a hydrogen atom from toluene to yield a benzyl radical at the active site. This radical would then add to the double bond of fumarate, forming a benzylsuccinyl radical. This type of radical addition to a C–C double bond is well known in organic chemistry and is employed in free-radical polymerization reactions [4]. Finally, the enzyme would donate the abstracted hydrogen atom back to the benzylsuccinyl radical, thus forming benzylsuccinate and at the same time regenerating the enzyme radical (Fig. 3). The strong homology to pyruvate formate-lyase suggests that the glycyl radical may react with the conserved cysteine residue to form an intermediate thiyl radical, which actually abstracts the hydrogen atom.

Formation of the radical form of pyruvate formate-lyase in Escherchia coli is catalyzed by pyruvate formate-lyase activase with S-adenosyl-methionine and reduced flavodoxin as co-substrates [56]. The discovery that the benzylsuccinate synthase operon contains an essential open reading frame which exhibits significant similarity to pyruvate formate-lyase activase, suggests that the protein encoded by this open reading frame may well generate the radical (active) form of benzylsuccinate synthase by a similar mechanism [54, 57].

3.3Genetics of anaerobic toluene mineralization

The genes coding for the subunits of benzylsuccinate synthase (bssA, B and C) were cloned and sequenced from T. aromatica. A single gene codes for both forms of the large subunit (α and α′), whereas the small β and γ subunits are encoded by separate genes [54]. Together with another gene, bssD (Fig. 4), these genes form a toluene-inducible operon (Fig. 4). The predicted bssD gene product shows strong similarity with activating enzymes required for radical generation in other glycyl radical enzymes. The bssD gene is cotranscribed with the structural genes of benzylsuccinate synthase and is probably regulated at the translational level [54].

Figure 4.

Organization of the bss operon, coding for benzylsuccinate synthase. The gene products encoded by the different genes are indicated below the genes.

An independent genetic analysis of anaerobic toluene metabolism was conducted in Thauera sp. strain T1. Screening of 10 000 chemically induced mutants for loss of the ability to grow anaerobically with toluene resulted in the identification of four mutants (named tut for toluene utilization), which were unable to grow with toluene anaerobically, but still metabolized benzoate. Three of these mutants have so far been characterized. They are affected in three different genes, which are closely clustered on one cosmid clone [57, 60]. One of these genes, tutB, appears to be involved in regulation of gene expression [60]. The other two genes are virtually identical to the genes coding for the large subunit of benzylsuccinate synthase (tutD and bssA; 80% identity) and the activating enzyme (tutE and bssD; 62% identity) from T. aromatica. In addition, equally close homologs of the genes for the two small subunits of benzylsuccinate synthase (bssB and bssC) are linked to tutD and tutE in Thauera sp. strain T1, and the operon organization is identical in these two strains. Substitutions of the predicted active site residues glycine 828 and cysteine 492 by alanine in TutD/BssA of Thauera sp. strain T1 failed to complement the mutant phenotype [57]. Thus, a biochemical and a genetic approach have identified independently the same genes for benzylsuccinate synthase and its putative activase.

The gene products of the bssABC genes of T. aromatica have been shown to be the subunits of benzylsuccinate synthase; the bssD gene probably codes for the activating enzyme required for converting benzylsuccinate synthase into the active (radical-containing) state. Although it has not yet been demonstrated, it appears likely that Thauera sp. strain T1 utilises a pathway of anaerobic toluene catabolism identical to the benzylsuccinate pathway shown in Fig. 2. This, along with the above mentioned identities of the tut and bss genes, leads us to suggest the adoption of the bss nomenclature for the structural genes involved in toluene conversion to benzylsuccinate also for strain T1.

4Anaerobic catabolism of ethylbenzene

Anaerobic mineralization of ethylbenzene has been reported in three denitrifying bacteria, strains EbN1, PbN1 and EB1 [21, 22]. The first two strains were isolated from freshwater mud, while the latter strain originated from an oil refinery treatment pond. All three isolates are closely related to each other, and are affiliated with the genus Azoarcus in the β-subclass of Proteobacteria, as indicated by 16S rDNA sequence analysis.

In ethylbenzene-mineralizing cultures, benzoate was detected as a transient intermediate [22]. This suggested that benzoate or benzoyl-CoA was an intermediate in the catabolism of ethylbenzene, as found in the degradation of many other aromatic compounds. The pathways employed for anaerobic oxidation of the two alkylbenzenes toluene and ethylbenzene to the oxidation level of benzoate, however, appear to be quite different.

A pathway for the anaerobic oxidation of ethylbenzene to benzoyl-CoA has been proposed, based on growth experiments [21]and cell suspension studies ([22]; K. Zengler, C. Champion, R. Rabus and F. Widdel, unpublished data). Strains EbN1 and EB1 were able to grow with 1-phenylethanol and acetophenone as sole carbon and electron sources under denitrifying conditions [21, 22]. Formation of both compounds from ethylbenzene was demonstrated in cell suspensions and cell extracts ([22, 50]; H.A. Johnson and A.M. Spormann, unpublished data). The proposed initial reaction for this pathway is the oxidation of ethylbenzene to 1-phenylethanol. The oxygen atom of the hydroxyl group of 1-phenylethanol is derived from water, as shown by labeling studies with 18O-water and GC/MS analysis, confirming that the reaction definitely occurred under strict exclusion of molecular oxygen [22]. 1-Phenylethanol is further oxidized to acetophenone. As inferred from the calculated redox potentials of analogous compounds (e.g. isopropanol/propane, E°′=−28 mV; acetone/isopropanol, E°′=−323 mV), the redox equivalents generated from oxidation of ethylbenzene to 1-phenylethanol (enzyme 1, Fig. 5) are probably transferred to the quinone pool; those generated from oxidation of 1-phenylethanol to acetophenone may have a sufficiently negative potential to be transferred to NAD+ or NADP+.

Figure 5.

Proposed pathway of anaerobic ethylbenzene degradation to the level of benzoyl-CoA. The enzymes needed for catalyzing the reactions of the proposed pathway are given (1–5). Note that input of energy is required for carboxylation of acetophenone and for CoA-thioester formation of benzoylacetate. The co-substrate requirements of these reactions are not known.

Only indirect evidence is available for the further reactions involved in acetophenone conversion to benzoyl-CoA. It is proposed that acetophenone is carboxylated to benzoylacetate (3-oxophenylpropionate) in a reaction analogous to reactions found in aerobic and anaerobic degradation of aliphatic ketones [61, 62]. This is supported by several experimental findings. Firstly, growth of both organisms on ethylbenzene or acetophenone occurred only in the presence of CO2 ([22]; K. Zengler, C. Champion, R. Rabus and F. Widdel, unpublished data). Furthermore, an unknown compound, which exhibited a similar UV-visible spectrum to acetophenone and may be the proposed carboxylation product benzoylacetate, was found to be transiently formed during acetophenone metabolism [22]. Benzoylacetate is proposed to be activated to the CoA thioester and to be cleaved thiolytically to acetyl-CoA and benzoyl-CoA (Fig. 5).

Strain EbN1 is able to metabolize two different hydrocarbon substrates, toluene and ethylbenzene [21]. Cells grown on either of the two substrates exhibited the enzyme activities necessary for metabolism of the respective growth substrate, but did not metabolize the other hydrocarbon compound. This is consistent with the use of completely different initial metabolic routes for ethylbenzene and toluene ([21, 22, 50]; H.A. Johnson and A.M. Spormann, unpublished data; K. Zengler, C. Champion, R. Rabus and F. Widdel, unpublished data).

5Anaerobic metabolism of alkanes

The first pure culture of an alkane-degrading bacterium was reported in 1991. This bacterium, sulfate-reducing strain Hxd3, grows on hexadecane and other long chain alkanes under strictly anaerobic conditions. The degradation balance showed that hexadecane was completely oxidized to CO2 at the expense of sulfate [29]. Since then, several further sulfate-reducing alkane-degrading isolates have been obtained which grow on alkanes with chains of six and more carbon atoms [30, 31]. The known alkane-degrading sulfate-reducing bacteria are nutritionally and phylogenetically unrelated to Desulfovibrio species [30]. The biochemical basis of hydrocarbon metabolism in these organisms is still poorly understood. Experiments with cell suspensions have indicated that anaerobic alkane degradation probably does not occur via dehydrogenation to a 1-alkene and hydration to an alcohol [30], a hypothetical mechanism presented in former literature. Most interestingly, one of the isolated alkane-degrading sulfate reducers, strain Hxd3 (Table 2), was found to produce membrane lipids containing odd numbers of C-atoms from alkanes with an even number of C-atoms (and vice versa). In contrast, control cells grown on 1-alkenes or fatty acids did not show this shift in the carbon chain length of cellular fatty acids. This indicates that alkanes and alkenes have different metabolic routes in this strain, and that addition or removal of a metabolite containing an odd number of C-atoms is involved in anaerobic alkane catabolism. The most plausible explanation for this would involve an initial carboxylation or carbonylation of the activated alkane to produce the Cn+1-fatty acid or aldehyde [30]; this would formally correspond to a reversal of the assumed biosynthetic reactions involved in alkane biosynthesis [5]. However, the described alteration of the chain length of fatty acids upon growth on alkanes has not been observed in any of the other known strains capable of growing anaerobically on alkanes (see Table 2). This suggests that there are different mechanisms for initiating anaerobic alkane metabolism, and that the mechanism employed depends on the bacterial strain [30].

6Regulation of anaerobic hydrocarbon metabolism

In all cases studied, the enzymes of anaerobic hydrocarbon metabolism are clearly substrate-induced. Catabolic enzymes for toluene are only present in cells grown on toluene, not in cells grown on other substrates, as demonstrated for T. aromatica[63], strain EbN1 and the sulfate-reducing D. toluolica[50]. Genes coding for two-component regulatory systems possibly involved in induction of gene expression by toluene have been described in Thauera strain T1 [60]and T. aromatica[64]. The exact regulatory mechanisms remain unknown.

Although ethylbenzene is chemically very similar to toluene, no cross-induction of catabolic enzymes for the two alkylbenzenes was observed in cells of strain EbN1 grown either on toluene or on ethylbenzene [50]. Details of the regulatory systems are not known.

Anaerobic alkane metabolism is induced only in cells grown on alkanes. It has been shown that synthesis of new proteins is necessary in fatty acid-grown cells in order to gain the ability to metabolize alkanes [30]. All known anaerobic alkane-degrading bacteria use a limited and clearly defined range of chain lengths; this range differs between the various strains [29–31].

7Ecological aspects of anaerobic hydrocarbon degradation

Hydrocarbon compounds as substrates for aerobic and anaerobic bacteria have probably always been available near natural petroleum deposits or petroleum formation sites, e.g. the Guaymas Basin [65]. Since one can assume continuous spreading of hydrocarbons into anaerobic environments over geological periods, the existence of bacteria capable of anaerobic hydrocarbon degradation is understandable from an ecological standpoint. In modern environments, polluted by human activity, these organisms are probably enriched, together with the long-known aerobic hydrocarbon degraders thriving in the oxic zones.

Hydrocarbon catabolism by anaerobic bacteria may cause serious problems in the petroleum industry. Secondary extraction of petroleum deposits is often performed by injection of sea water, which may introduce a bacterial inoculum, along with sulfate as an electron acceptor, into the sulfate-depleted reservoir. The extracted oil–water mixtures are suitable growth media for sulfate-reducing bacteria. Several constituents of petroleum serve directly as electron donors for sulfate reduction to sulfide [31], which causes ‘souring’ of the petroleum and gas. Furthermore, precipitation of insoluble metal sulfides may interfere with extraction of the oil and its separation from water, and free sulfide corrodes pipelines and storage tanks even in the absence of oxygen. Hydrogen sulfide escaping into the air presents a hazard for the workers in oil production plants.

Anaerobic bacterial hydrocarbon degradation may be technically exploited in the bioremediation of some hydrocarbon-polluted sites. Because the anaerobic processes are usually slower and less efficient than degradation under aerobic conditions, this application is only appropriate at sites with limited access of air, or which can not be aerated easily. Examples are contaminated groundwater aquifers, for which treatment with nitrate as anaerobic electron acceptor has been shown to favor the rate and the extent of bioremediation [66, 67].


The field of anaerobic microbial degradation of hydrocarbon compounds has developed only recently, but we are already beginning to understand some of the underlying mechanisms by which these rather inert substrates are attacked without the aid of molecular oxygen. The few anaerobic initiation reactions known are surprisingly diverse, compared to aerobic pathways, which are always initiated by an oxygenation reaction. Only in the case of hydrocarbons with similar structure and reactivity can we expect bacteria to make use of similar anaerobic degradation pathways. For example, propylbenzene catabolism is probably analogous to that of ethylbenzene [21], and m-xylene is apparently degraded by certain bacterial strains via a pathway analogous to that outlined for toluene (C.J. Krieger, M. Reinhard and A.M. Spormann, unpublished data). The o- and p-xylene isomers can also be co-metabolized with toluene and converted to dead-end products by some bacterial strains [33, 44, 47, 68]. So far, only one pure culture growing on o-xylene has been obtained [34], and no pure cultures growing on p-xylene have yet been reported. The metabolic pathways used for the mineralization of these xylene isomers remain unknown. Anaerobic degradation of some other hydrocarbons probably proceeds via unique and novel pathways. Examples include benzene and naphthalene, which have been considered to be very recalcitrant to anaerobic degradation in the laboratory in the past, but which have recently been shown to be degraded in anaerobic environments [35–37, 69]. The initial attack of these compounds under anoxic conditions is thought to be accomplished by novel mechanisms, for example by direct oxidation or carboxylation of the aromatic ring, as suggested by metabolic studies of mixed cultures [37, 70]. Finally, there have been reports for about 20 years that there is significant anaerobic oxidation of the most abundant alkane in nature, methane, in anoxic sediments and sewage sludge [71]. Some reports suggest that methanogenic archaea participate in this reaction, formally via a reversal of methanogenesis [71, 72], but there is no definitive information available on the organisms and the biochemical reactions involved in anaerobic methane oxidation in natural environments.


J.H. acknowledges Professor G. Fuchs (Mikrobiologie, Universität Freiburg) for his constant support and encouragement, as well as the financial support of the Deutsche Forschungsgemeinschaft. Research in the laboratory of A.M.S. was supported by NSF Grants MCB 9723312 and MCB 9733535, by Grant R-815738 of the Office of Research and Development, U.S. Environmental Protection Agency through the Western Region Hazardous Research Center, and by a Terman Award to A.M.S. The work of F.W. was supported by the Deutsche Forschungsgemeinschaft, the Max-Planck-Gesellschaft and the Fonds der chemischen Industrie.