The question how benzene is activated in the absence of oxygen is still not convincingly answered, although several attempts have been made since the end of the 1980s to elucidate the activation mechanism. In principle, the benzene molecule is thermodynamically very stable due to the symmetric π-electron system of the aromatic ring and the lack of potentially destabilizing or reactive substituents.
Many alkylated aromatic hydrocarbons, e.g. toluene and xylenes, are activated under anoxic conditions by a reaction sequence in which fumarate is added to the alkyl side-chain of the aromatic ring (for a review see Heider, 2007). In the first step of this reaction, a relatively stable benzyl radical is thought to be formed. One might suggest that benzene is activated by a similar mechanism; however, the formation of a phenyl radical as a reactive intermediate for a subsequent methylation or other reaction is rather unlikely for energetic reasons. The abstraction of a hydrogen atom from benzene would need an activation energy of more than 460 kJ mol−1 which is roughly 100 kJ mol−1 more compared with the formation of a benzyl radical from alkylated benzene derivatives (Widdel and Rabus, 2001; Musat and Widdel, 2008). Thus, three other activation mechanisms have been intensively discussed: (i) an anaerobic hydroxylation of benzene yielding phenol, (ii) a Friedel-Crafts-type methylation of benzene yielding toluene and (iii) a carboxylation of benzene yielding benzoate (Fig. 2). We will discuss the pros and cons for all three possible activation reactions. Generally, isotope-based methods have been used to elucidate the reaction mechanism, either by CSIA, or by detection of 13C- or 14C-labelled metabolites which are formed during transformation of 13C- or 14C-labelled benzene. Due to the lack of pure cultures, most studies aiming to elucidate the reaction mechanism were performed with enrichment cultures.
Hydroxylation of benzene had been already suggested as an initial reaction mechanism for benzene activation in one of the first reports regarding anaerobic benzene degradation (Vogel and Grbic-Galic, 1986). Addition of 18O-labelled water resulted in the formation of 18O-labelled phenol, indicating that the introduced hydroxyl group originated from water. In subsequent reports, phenol was often detected in enrichments as metabolite of benzene degradation under different electron acceptor conditions: in iron-reducing cultures (Caldwell and Suflita, 2000; Botton and Parsons, 2007; Kunapuli et al., 2008), sulfate-reducing cultures (Caldwell and Suflita, 2000; Laban et al., 2009) and methanogenic cultures (Weiner and Lovley, 1998b; Caldwell and Suflita, 2000; Ulrich et al., 2005). In some studies, benzoate was concomitantly detected with phenol (Caldwell and Suflita, 2000; Ulrich et al., 2005; Kunapuli et al., 2008).
It is noteworthy that it has recently been observed that phenol can be abiotically formed from benzene in culture media from iron and sulfate reducers by contact with air after sampling (Kunapuli et al., 2008). Hydroxyl radicals were likely generated by oxidation of iron in the sample during work up, which then reacted rapidly with benzene producing small amounts of phenol before sample analysis. The formation of 2-hydroxybenzoate and 4-hydroxybenzoate from benzoate was also explained by these mechanisms (Laban et al., 2009). These results indicate that it is generally problematic to distinguish between biotic and abiotic phenol formation in strongly reduced culture samples, probably preventing any clear evidence for benzene hydroxylation under strictly anoxic conditions by means of metabolite analysis. Hence, other methodological approaches are needed in addition. Substrate consumption tests with two highly enriched sulfate-reducing benzene-degrading cultures revealed that phenol is either not consumed (Laban et al., 2009) or only consumed after a certain lag-phase (Musat and Widdel, 2008), strongly indicating that phenol is unlikely to be an intermediate during benzene degradation in these cultures.
Chakraborty and Coates (2005) suggested that Dechloromonas strain RCB hydroxylated benzene to phenol, which was subsequently transformed to benzoate; the reaction was dependent on the presence of nitrate as electron acceptor. Here, the authors did not report phenol formation or benzene degradation in anoxic, nitrate-free control cultures amended with benzene. The origin of the introduced hydroxyl group could not be identified. When cells degraded benzene in H218O-enriched mineral salt medium, the formed phenol was only slightly enriched with 18O, suggesting that the hydroxyl group did not originate from water. On the other hand, hydroxyl free radical scavengers strongly inhibited benzene degradation and phenol formation, indicating that hydroxyl radicals were the source of the hydroxyl group in phenol. In an additional study, Chakraborty and colleagues (2005) showed that strain RCB could degrade benzene and several other aromatic hydrocarbons with nitrate, chlorate or oxygen as electron acceptor. Surprisingly, none of the known genes for anaerobic degradation of aromatic compounds could be found in the genome of strain RCB, which was recently sequenced (Salinero et al., 2009). Due to the lack of these genes, anaerobic benzene degradation in strain RCB ‘remains enigmatic’ (Salinero et al., 2009). On the other hand, strain RCB encodes several aerobic pathways for aromatics degradation. In the presence of chlorate, this organism releases oxygen during chlorate respiration allowing aromatics degradation by means of oxygenases even in the initial absence of oxygen in the culture medium, as shown also for the chlorate-reducing benzene degrader Alicycliphilus denitrificans (Weelink et al., 2008). Interestingly, it has been recently demonstrated that oxygen can be released during reduction of nitric oxide (NO) (Ettwig et al., 2010), an intermediate of the classical nitrate reduction pathway to dinitrogen. As also suggested by Weelink and colleagues (2010), strain RCB may contain this enzyme, allowing the use of oxygen for the initial attack of benzene and other aromatics even under nitrate-reducing conditions, explaining the apparently inconsistent physiological and genetic data. Unfortunately, the oxygen-releasing enzyme has not been characterized and can therefore not yet be identified in the genome of strain RCB. The genome of strain RCB contains the genes for the classical nitrate reduction to dinitrogen pathway, including those for nitric oxide reductase which catalyses nitric oxide reduction to nitrous oxide (N2O) (Salinero et al., 2009).
Benzene methylation via Friedel-Crafts-type reaction is exergonic using the unique biological methyl donors S-adenosyl-methionine or methyl-tetrahydrofolate (Coates et al., 2002), which open the doors for another hypothesis for anaerobic benzene activation. Actually, S-adenosyl-methione-dependent alkylation of benzene (and substituted aromatics) has been observed in bone marrow (Flesher and Myers, 1991). Methylation has been also proposed for the anaerobic activation of the non-substituted aromatic hydrocarbon naphthalene (Safinowski and Meckenstock, 2006). If benzene is methylated by anaerobes, the reaction product toluene could be further activated by addition of fumarate to the methyl group of toluene catalysed by the enzyme benzylsuccinate synthase (BSS), leading to the characteristic compound benzylsuccinate as intermediate. BSS has been detected in several anaerobic toluene-degrading pure and mixed cultures, and fumarate addition seems to be a unique activation mechanism for anaerobic toluene degradation (for an overview see Heider, 2007). PCR primers for the gene encoding the protein subunit which contains the reactive centre, bssA, have been also developed (Winderl et al., 2007). Thus, reasonable strategies for verifying benzene activation by methylation are detecting the intermediates toluene and benzylsuccinate or detecting the presence or expression of bssA-like genes or the induction or activity of BSS.
Ulrich and colleagues (2005) detected [ring-13C]-labelled toluene and [ring-13C]-labelled benzoate as intermediates in [13C6]-benzene-spiked nitrate-reducing and methanogenic enrichment cultures. The formation of [13C6]-phenol was observed only in the methanogenic culture. The nitrate-reducing culture degraded toluene rapidly and at higher rates than benzene, supporting the hypothesis that toluene might be an intermediate during benzene degradation in this culture. In contrast, toluene was only slightly degraded by the methanogenic culture. The authors concluded that two degradation pathways exist: (i) a methylation pathway leading to toluene with subsequent transformation to benzoate operating in the nitrate-reducing and methanogenic culture, and (ii) a hydroxylation pathway leading to phenol with subsequent formation of benzoate operating only in the methanogenic culture. This hypothesis was supported by studies in which CSIA was used for characterizing the initial step of benzene activation in different cultures (Mancini et al., 2003; 2008; Fischer et al., 2008). Mancini and colleagues (2008) showed that the ratio of hydrogen isotope fractionation (Δδ2H) versus carbon isotope fractionation (Δδ13C) – a value defined as lambda: Λ = Δδ2H/Δδ13C (Fischer et al., 2008) – for anaerobic benzene degradation was significantly higher for the methanogenic culture (Λ = 39 ± 5) compared with the nitrate-reducing enrichment culture (Λ = 16 ± 2) investigated by Ulrich and colleagues (2005). Simplified, the lambda value can be seen as a biochemical fingerprint of a given biochemical reaction. Concordantly, other nitrate-reducing cultures showed lambda values in the range between 8 and 19, whereas for other methanogenic or sulfate-reducing cultures lambda values between 28 and 31 were determined (Mancini et al., 2003; 2008; Fischer et al., 2008). Thus, the CSIA data indicate that benzene activation under nitrate-reducing conditions is different from benzene activation under sulfate-reducing and methanogenic conditions. Nevertheless, ‘different reaction mechanism’ means that either the reactions are truly different (different products are formed), or the reactions are similar on paper (same products are formed) but proceed via different reaction mechanisms catalysed by different enzymes or cofactors leading to different fractionation patterns; the latter has been recently shown for toluene activation by benzylsuccinate synthase (Vogt et al., 2008; Herrmann et al., 2009). Hence, further research is needed to conclusively demonstrate that benzene can be methylated under nitrate-reducing conditions. Some highly enriched strictly anaerobic benzene-degrading cultures cannot degrade toluene (Kunapuli et al., 2008; Musat and Widdel, 2008; Laban et al., 2009), probably excluding biomethylation of benzene to form toluene as activation mechanism.
Similar to hydroxylation or methylation of benzene, benzene carboxylation is slightly exergonic or close to the thermodynamic equilibrium depending on the carboxyl donor and thus feasible even in sulfate-reducing or methanogenic cultures (Musat and Widdel, 2008). Additionally, for some non-substituted aromatic compounds, e.g. naphthalene (Zhang and Young, 1997; Musat et al., 2009; DiDonato et al., 2010) or phenanthrene (Zhang and Young, 1997; Davidova et al., 2007), a carboxylation reaction was suggested for ring activation, indicating that carboxylation might be an important activation principle for the degradation of non-substituted aromatic compounds in nature. Indeed, benzoate has been detected as intermediate of anaerobic benzene degradation in sulfate-reducing (Caldwell and Suflita, 2000; Phelps et al., 2001; Laban et al., 2009), iron-reducing (Caldwell and Suflita, 2000; Kunapuli et al., 2008), nitrate-reducing (Ulrich et al., 2005) or methanogenic (Caldwell and Suflita, 2000; Ulrich et al., 2005) enrichment cultures, concomitantly with the intermediates phenol (Caldwell and Suflita, 2000; Ulrich et al., 2005; Kunapuli et al., 2008; Laban et al., 2009) or toluene (Ulrich et al., 2005). Phelps and colleagues (2001) found deuterated benzoate (D5) as sole intermediate of deuterated benzene (D6) degradation in their highly enriched marine sulfate-reducing culture. Interestingly, incubation in the presence of a 13C-labelled bicarbonate buffer system did not lead to 13C incorporation in the carboxyl group of the benzoate intermediate, indicating that the introduced carboxyl group did not originate from carbon dioxide. This was in accordance with results presented by Caldwell and Suflita (2000) for their benzene-degrading sulfate-reducing freshwater culture. Here, fully [13C7]-labelled benzoate was formed when the culture was spiked with fully [13C6]-labelled benzene, showing that the carboxyl group of benzoate was stemming from transformation products of [13C6]-benzene itself, but not from the non-labelled bicarbonate buffer system. In contrast, Kunapuli and colleagues (2008) found both [13C6]-benzoate and [13C7]-benzoate in their iron-reducing enrichment culture during incubation with fully labelled [13C6]-benzene. Furthermore, the authors detected 13C-carboxy group-labelled benzoate if cells were incubated in medium prepared with non-labelled benzene and 13C-labelled bicarbonate buffer, favouring the hypothesis that the bicarbonate buffer was the carboxyl group donor for benzoate formation.
However, it is generally difficult to interpret all these observations since benzoyl-CoA, the activated form of benzoate, is a common intermediate within the anaerobic degradation pathways of several aromatic compounds including toluene and phenol (for an overview see Carmona et al., 2009; see also Fig. 2); in addition, benzoate has been reported as an excreted intermediate during phenol degradation under methanogenic conditions (Knoll and Winter, 1987; Kobayashi et al., 1989; Bechard et al., 1990; Karlsson et al., 2000). Hence, the metabolite benzoate might be formed directly from benzene by a carboxylation step, but could be formed as well in later steps during anaerobic benzene degradation pathways starting, e.g. with a methylation or hydroxylation step.
Lately, Laban and colleagues (2010) investigated their highly enriched iron-reducing benzene-degrading culture using an approach combining metagenomics and metaproteomics. Subcultures were grown with benzene, phenol or benzoate as sole substrates, and peptide sequences were subsequently identified based on the metagenome which had been sequenced before. Proteins similar to the phenylphosphate carboxylase subunits PpcA and PpcD of Azoarcus sp. strain EbN1 and to the benzoate-CoA ligase of Geobacter metallireducens were specifically expressed during anaerobic benzene degradation. Based on these results, the authors suggested that benzene is directly carboxylated by a putative anaerobic benzene carboxylase. The formed benzoate might be further activated by a benzoate-CoA ligase to benzoyl-CoA. However, an enzyme activity test for the putative anaerobic benzene carboxylase, ultimately proving this hypothesis, could not be established yet.