Anaerobic Benzene Oxidation via Phenol in Geobacter metallireducens

The bacterial strains and plasmids used are listed in Table S1 in the supplemental material. Geobacter metallireducens (ATCC 53774 and DSM 7210) (46) was obtained from our laboratory culture collection and was routinely cultured under strict anaerobic conditions with 50 mM Fe(III) citrate as the electron acceptor, as previously described (47). Unless otherwise noted, the electron donor (concentration) for growing cultures was benzene (0.25 mM), phenol (0.5 mM), benzoate (1 mM), toluene (0.5 mM), or acetate (10 mM). The exception to this was that in the studies in which phenol concentrations during growth on benzene were monitored, the initial concentration of benzene was 0.1 mM.

For control studies, G. sulfurreducens cultures that were grown with acetate (10 mM) as the electron donor and Fe(III) citrate (50 mM) as the electron acceptor served as the inoculum, and a 10% inoculum was transferred into medium with benzene (100 μM) and Fe(III) citrate (50 mM). The acetate remaining in the inoculum (ca. 4 mM) served as an additional electron donor.

Studies of the metabolism of ¹⁴C- or ¹⁸O-labeled compounds were performed with cell suspensions to increase cell density and provide more rapid metabolic flux. Cells grown in acetate-Fe(III) citrate medium were concentrated under anaerobic conditions via centrifugation (4,400 × g for 10 min at 15°C). The cell pellets were washed with Fe(III) citrate medium devoid of an electron donor and then resuspended to provide a 30-fold concentration of cells. Aliquots (3 ml) of the cell suspension were incubated in 15-ml anaerobic pressure tubes. For studies on the origin of oxygen in phenol, H₂¹⁸O (95 atom% ¹⁸O; Sigma-Aldrich, St. Louis, MO) was mixed with water to provide 9.5% H₂¹⁸O. Additions for ¹⁴C-labeled studies were 2.59 × 10⁵ Bq [U-¹⁴C]benzene (39 μM, 2.78 × 10⁹ Bq mmol⁻¹; Moravek Biochemicals, Brea, CA), 2.22 × 10⁵ Bq [U-¹⁴C]benzoate (40 μM, 2.22 × 10⁹ Bq mmol⁻¹; Moravek Biochemicals, Brea, CA), or 3.7 × 10⁵ Bq [U-¹⁴C]phenol (44 μM, 2.96 × 10⁹ Bq mmol⁻¹; American Radiolabeled Chemicals Inc. [ARC], St. Louis, MO).

Cells were grown in 1-liter bottles and harvested during mid-exponential phase by centrifugation. The cell pellet was immediately frozen in liquid nitrogen and stored at −80°C.

RNA was isolated from triplicate cultures grown on each electron donor with a modification of a previously described method (17). Briefly, cell pellets were resuspended in HG extraction buffer (18) preheated to 65°C. The suspension was incubated for 10 min at 65°C to lyse the cells. Nucleic acids were isolated with a phenol-chloroform extraction followed by ethanol precipitation. The pellet was washed twice with 70% ethanol, dried, and resuspended in sterile diethylpyrocarbonate-treated water. RNA was then purified with an RNA cleanup kit (Qiagen, Valencia, CA) and treated with DNA-free DNase (Ambion, Woodward, TX). The RNA samples were tested for genomic DNA contamination by PCR amplification of the 16S rRNA gene. cDNA was generated with a TransPlex whole-transcriptome amplification kit (Sigma).

The sequences of all primers used for quantitative reverse transcription-PCR are listed in Table S2 in the supplemental material. Each reaction consisted of forward and reverse primers at a final concentration of 200 nM, 5 ng of cDNA, and 12.5 μl of Power SYBR green PCR master mix (Applied Biosystems, Foster City, CA). Primer pairs with amplicon sizes of 100 to 200 bp were designed for the following: bamY (Gmet 2143), ppsA (Gmet 2100), and bssA (Gmet 1539). Expression of these genes was normalized to the expression of proC, a gene shown to be constitutively expressed in Geobacter species (48). Relative levels of expression of the studied genes were calculated with the 2^−ΔΔCT threshold cycle (C_T) method (49).

Whole-genome microarray hybridizations were carried out by Roche NimbleGen, Inc. (Madison, WI). Triplicate biological and technical replicates were conducted for all microarray analyses. Cy3-labeled cDNA was hybridized to oligonucleotide microarrays based on the G. metallireducens genome and resident plasmid sequences (GenBank accession numbers NC_007515 and NC_007517). The microarray results were analyzed with Array 4 Star software (DNASTAR, Madison, WI). A gene was considered differentially expressed only if the P value determined by Student's t test analysis was less than or equal to 0.01.

Sequences for all primers used for construction of the mutants are listed in Table S2 in the supplemental material. Genomic DNA was extracted with an Epicentre MasterPure DNA purification kit (Epicentre Biotechnologies, Madison, WI). Plasmids were extracted with a QIAprep Spin miniprep kit (Qiagen). PCR amplification was done with Taq polymerase (Qiagen). DNA gel purification was done with a QIAquick gel extraction kit (Qiagen). Mutant alleles were constructed by replacing the coding sequences with a spectinomycin resistance cassette as described previously (45). Briefly, upstream and downstream flanking regions of the genes to be deleted (ca. 500 bp) were amplified with G. metallireducens genomic DNA as the template. PCR products were mixed, digested with AvrII (NEB, Beverly, MA), and ligated with the T4 DNA ligase (NEB). The resulting construct (ca. 1kb) was cloned into pCR2.1-TOPO with a TOPO TA cloning kit (Invitrogen, Carlsbad, CA). The spectinomycin resistance cassette with AvrII sites at both ends was amplified with pRG5 (50) as the template. The spectinomycin resistance cassette was AvrII digested and ligated into the AvrII site located between the flanking regions of the genes to be deleted.

The sequences of the mutant alleles were verified by Sanger sequencing. Plasmids bearing the mutant alleles were linearized by restriction enzyme digestion. The linearized plasmids were concentrated by ethanol precipitation. The linearized plasmids were electroporated into G. metallireducens as described earlier (45). The genotypes of the mutant strains were confirmed by PCR using the genomic DNA of the mutant strains as the template.

For complementation studies, ppsA and ppcB with their respective native ribosome binding sites (RBSs) were cloned under the control of a constitutive lac promoter into pCM66 (51).

Fe(III) reduction was monitored by measuring the accumulation of Fe(II) over time with ferrozine, as previously described (52).

The benzene and toluene concentrations in the headspace were quantified with a gas chromatograph equipped with a flame ionization detector. The samples were run through a Supelco VOCOL fused silica capillary column (60 m by 0.25 mm by 1.5 μm) held at 50°C for 0.5 min, followed by an increase to 200°C at 10°C/min, as previously described (16). The concentrations of benzene in the aqueous phase were calculated with Henry's law using the constant at 25°C of 0.25 for benzene and 0.27 for toluene (53).

The acetate, benzoate, and phenol concentrations in the medium were measured with high-performance liquid chromatography as previously described (16). The [¹⁴C]carbon dioxide in the headspace was quantified with a gas proportional counter, as previously described (54, 55).

Samples (5 ml) of culture for analysis of transient phenol were collected over time from benzene-degrading cultures and immediately extracted with hexane (2 ml) by vigorous shaking for 2 min. Hexane extracts from unlabeled water and ¹⁸O experiments were immediately analyzed with a gas chromatograph-mass spectrometer (Shimadzu GCMS-QP2010Ultra). The injector port temperature was set at 250°C, and 1-μl sample extracts were injected through a SHRX1-5MS column (30 m by 0.25 mm by 0.25 μm) with the splitless mode. The column temperature was initially held at 60°C for 1 min, followed by an increase to 200°C at 4°C/min. Compounds were identified by comparison with chemical standards, and mass spectral data were characterized by comparison with data in the National Institute of Standards and Technology (NIST) library.

Microarray data have been deposited with the NCBI GEO database under accession number GSE33794.

Small quantities of phenol but not benzoate were detected in cultures of G. metallireducens growing on benzene (Fig. 1), suggesting that G. metallireducens metabolized benzene via a phenol intermediate. It was previously suggested that phenol may be abiotically produced from benzene in the presence of iron (30), but this was not observed in subsequent studies with F. placidus (17). In order to further evaluate this possibility, an inoculum of G. sulfurreducens, a Geobacter species which cannot metabolize aromatic compounds (56), was grown in acetate-Fe(III) citrate medium, and a 10% inoculum of this culture was added to Fe(III) citrate medium that contained 100 μM benzene. The Fe(II) from the inoculum, as well as the Fe(II) produced from the residual acetate in the inoculum, ensured the presence of substantial (>7 mM) Fe(II), yet no phenol was detected over 27 days (Fig. 1). These results further demonstrate that phenol is not abiotically produced from benzene reacting with iron under the culture conditions employed.

In order to investigate the source of oxygen for phenol production, cell suspensions were provided benzene and Fe(III) citrate in medium in which the water included 9.5% H₂¹⁸O (Fig. 2). The m/z 96 peak, representing ¹⁸O-labeled phenol, had a relative intensity that was 5% of the m/z 94 peak, representing unlabeled phenol (Fig. 2A). There was no m/z 96 peak when G. metallireducens was incubated in benzene medium without H₂¹⁸O (Fig. 2B). The production of ¹⁸O-labeled phenol suggested that the hydroxyl group introduced into benzene during benzene metabolism was derived from water. This result is consistent with the fact that G. metallireducens does not possess monooxygenase genes, which would be required for activation of benzene with molecular oxygen. These results, coupled with the fact that strict anaerobic culturing techniques and the presence of abundant Fe(II) in the cultures ensured that molecular oxygen was not available, suggested that the formation of phenol in benzene-degrading cultures was likely to be an anaerobic enzymatic reaction.

In order to gain additional insight into the pathway for benzene metabolism, the transcript abundance for genes coding for subunits of the enzymes catalyzing the first step of benzoate, phenol, or toluene metabolism was quantified with reverse transcription-quantitative PCR during growth with one of these aromatic compounds. Results were compared with expression during growth with acetate. As might be expected, each gene was the most highly expressed when its respective substrate was the electron donor that was added to support growth (Fig. 3). For example, bamY (Gmet 2143) encodes the benzoate coenzyme A (benzoate-CoA) ligase, an enzyme catalyzing the conversion of benzoate to benzoyl-CoA (57–59), and the transcript abundance of bamY was the highest during growth on benzoate (Fig. 3A). The transcript abundance for ppsA (Gmet 2100), which encodes the alpha subunit of phenylphosphate synthase, responsible for the conversion of phenol to phenylphosphate (60, 61), was the highest during growth on phenol (Fig. 3B). The transcript abundance of bssA (Gmet 1539), which encodes the alpha subunit of benzylsuccinate synthase, responsible for the activation of toluene with fumarate (60, 62), was the highest during growth on toluene (Fig. 3C).

As has previously been observed (63, 64), the genes for enzymes responsible for initiating the metabolism of benzoate, phenol, or toluene also had some increased expression during growth on the alternative aromatic compounds. This result demonstrates that the regulation of expression of aromatic degradation genes is not absolute but that relative gene expression levels might be diagnostic of the metabolic pathway employed.

During growth on benzene, the greatest increase in transcript abundance was for ppsA (Fig. 3D). There was a slight increase in transcript abundance for bamY, but this was much less than that in cells grown on any other aromatic compound. Transcript levels for bssA were comparable to those in acetate-grown cells. These results suggest that phenol, but not benzoate or toluene, is an intermediate in benzene degradation. The transcript abundance of ppsA was lower in the benzene-grown cells than in the phenol-grown cells, which may be linked to the higher phenol concentrations available to phenol-grown cells, as well as the greater phenol flux in phenol-grown cells, which grow much faster than benzene-grown cells.

Subsequent genome-scale analysis of gene expression with whole-genome microarrays (NCBI GEO accession number GSE33794) revealed that genes for the putative phenol pathway in G. metallireducens that are upregulated in cells grown on phenol versus their regulation in cells gown on acetate are also upregulated during growth on benzene (Fig. 4A). The one exception was the gene for PpcD, which is thought to be a subunit in the enzyme phenylphosphate carboxylase (65). However, expression of the gene for the other subunit of this enzyme was upregulated in benzene-grown cells. Genes in the pathway for toluene degradation, which were highly upregulated in toluene-grown cells, had relatively little or no increase in transcript abundance in benzene-grown cells (Fig. 4B). In a similar manner, the microarray analysis confirmed the earlier findings from quantitative PCR analysis that bamY expression levels were much lower in benzene-grown cells than in cells metabolizing benzoate (Fig. 4C). Like bamY, bct (Gmet 2054), a gene coding for a succinyl-CoA:benzoate-CoA transferase involved in a recently discovered alternative mode of benzoate activation (66), was highly upregulated in benzoate-grown cells but not in cells grown on benzene (Fig. 4D). The results are consistent with a metabolism in which benzene is anaerobically hydroxylated to phenol, followed by a two-step carboxylation process to 4-hydroxybenzoate via phenylphosphate as an intermediate. 4-Hydroxybenzoate is further reductively converted via its coenzyme A ester to benzoyl-CoA. Free benzoate itself is not an expected intermediate.

The possibility that benzene was metabolized via a phenol intermediate was further evaluated by deleting either ppsA, a gene predicted to encode a subunit for the enzyme necessary for the first step in phenol metabolism, or ppcB (65), a gene predicted to encode a subunit for the second step in phenol metabolism (Fig. 4). Cell suspensions of the wild-type strain readily utilized phenol, as expected (67), but cell suspensions of the ppsA- or ppcB-deficient strains could not (Fig. 5A), even though they metabolized benzoate as well as the wild type (Fig. 5B). Benzene oxidation was completely inhibited (Fig. 5C) in the absence of ppsA or ppcB. Complementation of either gene deletion by expression of the deleted gene on a plasmid restored the capacity for anaerobic phenol and benzene oxidation (Fig. 5). Rates of metabolism in the complemented strains were lower than those in the wild type, consistent with the general finding in previous studies with Geobacter species that the strategy for in trans expression does not yield wild-type activities when complementing deletions of genes for a wide variety of functions (45, 68–71). These results demonstrate that the phenol degradation pathway is essential for benzene degradation, further suggesting that phenol is an important intermediate in the metabolism of benzene.

The results demonstrate that phenol is a key intermediate in the anaerobic degradation of benzene by Geobacter metallireducens. The specific upregulation of genes for phenol metabolism during growth on benzene and the finding that the capacity for anaerobic benzene metabolism is lost when genes for phenol metabolism are deleted indicate that phenol is produced from benzene. The incorporation of oxygen from water in phenol demonstrates that molecular oxygen is not involved in the benzene activation. Potential mechanisms for conversion of benzene to phenol, an endergonic reaction, have previously been proposed (7, 24), but unlike the clustering of genes for the degradation of other monoaromatic compounds (60), no genomic regions specific for benzene degradation have been identified through genome annotation or gene expression patterns. Thus, further speculation on activation mechanisms are not warranted at this time.

The production of phenol as an intermediate in anaerobic benzene degradation in G. metallireducens contrasts with the production of benzoate as an intermediate in the hyperthermophile Ferroglobus placidus (17). A highly enriched benzene-oxidizing, Fe(III)-reducing enrichment culture also appears to activate benzene via carboxylation (19, 30).

It seems likely that benzene conversion to phenol will be the major route for anaerobic benzene degradation in petroleum-contaminated environments in which Geobacter species are specifically enriched (37). The ability to discern pathways for anaerobic benzene degradation through analysis of gene transcript abundance in pure cultures, as demonstrated previously in studies with F. placidus (17) and in this study with G. metallireducens, might be extended with metatranscriptomic analysis to elucidate how benzene is anaerobically degraded in other environments.