Conversion of phenyl methyl ethers by Desulfitobacterium spp. and screening for the genes involved

Abstract

Microbial growth coupled to O-demethylation of phenyl methyl ethers, which are lignin decomposition products, was described for acetogenic bacteria and recently also for two species belonging to the nonacetogenic genus Desulfitobacterium. To elucidate the potential role of desulfitobacteria in the O-demethylation of phenyl methyl ethers in the environment, we cultivated Desulfitobacterium chlororespirans, D. dehalogenans, D. metallireducens, and different strains of D. hafniense with phenyl methyl ethers as sole electron donors. With the exception of D. metallireducens, all species and strains tested were able to demethylate at least three of the four phenyl methyl ethers applied with fumarate, nitrate, or thiosulfate as electron acceptor. Furthermore, a high number of operons putatively encoding demethylase systems were identified in the genomes of Desulfitobacterium spp., although discrimination between O-, S-, N- and, Cl-demethylases was not possible. These findings provide evidence for the importance of the methylotrophic metabolism for desulfitobacteria and point to their involvement in the O-demethylation of phenyl methyl ethers in the environment.

O-demethylation coupled to fumarate, nitrate, thiosulfate or Fe(III) reduction was found to be a common metabolic feature of the genus Desulfitobacterium.

Open in new tabDownload slide

Introduction

Desulfitobacteria, which are most commonly described as reductively dehalogenating bacteria, are able to utilize phenyl methyl ethers as growth substrates when an appropriate electron acceptor (fumarate or a chlorinated phenol) is provided (Neumann et al., 2004). Prior to this finding, utilization of these substrates had only been described for members of the acetogenic bacteria (Bache & Pfennig, 1981; Daniel et al., 1991; Traunecker et al., 1991; Stupperich & Konle, 1993; Liesack et al., 1994). Phenyl methyl ethers are the main decomposition products of lignin and therefore widespread in the environment, especially in soil. Abundant representatives of this class include, for example, syringate, vanillate, and isovanillate (Chen et al., 1982, 1983; Kögel, 1986). Phenyl methyl ethers typically result from the conversion of lignin by ligninolytic enzymes that are produced by white rot fungi under aerobic conditions (Higuchi, 1990). Once they are released from the lignin macrostructure, they become available to the soil microbial community. In acetogens and desulfitobacteria, the methyl groups of phenyl methyl ethers are cleaved off by O-demethylase enzyme systems (Kaufmann et al., 1997; Naidu & Ragsdale, 2001; Studenik et al., 2012) and the further oxidation of the methyl moiety to CO₂ plays a key role for energy conservation (for a review see Drake et al., 2006). O-demethylases, which are induced by their respective substrates (Engelmann et al., 2001; Peng et al., 2012), consist of two methyltransferases (MTs) (MT I and MT II), a corrinoid protein (CP), and an activating enzyme (AE) (Kaufmann et al., 1997; Schilhabel et al., 2009; Studenik et al., 2012). The reaction mechanism is depicted in Fig. 1. In the first step, MT I cleaves the ether bond of the methoxylated substrate and transfers the methyl group to the super-reduced corrinoid cofactor ([Co^I]) of CP. Subsequently, MT II transfers the methyl group from the methylated corrinoid to tetrahydrofolate (FH₄) yielding methyltetrahydrofolate (CH₃-FH₄). The last protein component of the system, AE, has a repair function after inadvertent oxidation of the super-reduced corrinoid cofactor ([Co^I]) to inactive [Co^II]-CP. It catalyzes the ATP-dependent electron transfer from an unknown electron donor to [Co^II]-CP yielding physiologically active [Co^I]-CP, allowing CP to participate in further O-demethylation reactions (Siebert et al., 2005; Sperfeld et al., 2014). For acetogenic bacteria, it has been described that the phenyl methyl ether-derived methyl group in CH₃-FH₄ undergoes oxidation to CO₂. The reducing equivalents generated by methyl group oxidation are transferred to CO₂ that is reduced to enzyme-bound carbon monoxide. Acetyl-CoA is formed in the carbon monoxide dehydrogenase/acetyl-CoA synthase reaction from the enzyme-bound CO and methyl groups derived from phenyl methyl ether demethylation. Acetyl-CoA can then either be converted to acetate to conserve energy or it can be incorporated into cell carbon (Ragsdale & Pierce, 2008). In contrast to acetogens, desulfitobacteria appear to be unable to use CO₂ as a terminal electron acceptor for O-demethylation (Neumann et al., 2004; Kreher et al., 2008), which renders their reductive acetyl-CoA pathway nonfunctional for energy conservation. Instead, they rely on alternate electron acceptors that might be provided by the microbial community or by the environment. It was previously shown that Desulfitobacterium hafniense DCB-2 and D. hafniense PCE-S use fumarate as terminal electron acceptor and that D. hafniense DCB-2 was also able to couple O-demethylation of vanillate to the reductive dechlorination of 3-chloro-4-hydroxyphenylacetic acid to 4-hydroxyphenylacetic acid (Neumann et al., 2004). However, nothing is known about possible physiological electron acceptors that might be present in natural environments such as forest soil, where phenyl methyl ethers are available as substrates for the soil microbial community.

The objective of this study was to shed light on the general importance of the methylotrophic metabolism for Desulfitobacterium spp. and to elucidate their potential role in the O-demethylation of phenyl methyl ethers in the environment. For this purpose, phenyl methyl ether consumption with different electron acceptors, namely fumarate, nitrate, and thiosulfate, was followed in growth experiments of D. chlororespirans, D. dehalogenans, D. metallireducens, and different strains of D. hafniense. Additionally, bioinformatic analyses of available genome sequences of various Desulfitobacterium species were carried out to uncover the O-demethylation potential of these organisms.

Materials and methods

Bacterial strains

Desulfitobacterium hafniense strains DCB-2 (DSM-10664), DP7 (DSM-13498), G2 (DSM-16228), PCE-S (DSM-14645), PCP-1 (DSM-12420), and TCP-A (DSM-13557) as well as D. chlororespirans (DSM-11544), D. dehalogenans (DSM-9161), and D. metallireducens (DSM-15288) were obtained from the Deutsche Sammlung von Mikroorganismen und Zellkulturen (DSMZ, Braunschweig, Germany). Desulfitobacterium hafniense strain Y51 was taken from the strain collection of our laboratory and was kindly provided by Taiki Futagami (Kagoshima University, Kagoshima, Japan).

Growth of bacteria

Desulfitobacterium spp. was grown anoxically in rubber-stoppered serum bottles under 100% N₂ atmosphere as described by Neumann et al. (2004). The electron donors (syringate, vanillate, isovanillate, or 4-hydroxyanisole) and acceptors (fumarate, nitrate, Fe(III), or thiosulfate) were supplied from separately autoclaved anoxic stock solutions. The concentrations applied were as follows: phenyl methyl ethers, 2 mM; fumarate, 20 mM; nitrate and thiosulfate, 5 mM; Fe(III) citrate, 12 mM. Media were inoculated with precultures grown with pyruvate and fumarate (each 40 mM) as electron donor and acceptor, respectively. For precultures of D. metallireducens, Fe(III) (c. 20 mM) was used as electron acceptor instead of fumarate. The culture volume for inoculation was 10% (v/v) of the final volume of the main culture. Cultures were incubated in a water bath shaker at 28 °C and 150 r.p.m. Samples were taken and analyzed for growth and metabolites.

Analytical methods

The protein concentration was used as an indicator of bacterial growth and was determined according to the method of Bradford (1976) after cell lysis by alkaline treatment.

The concentrations of phenyl methyl ethers and fumarate were determined by HPLC using a LiChrospher 100 RP-8 125 × 4 mm column (Merck KGaA, Darmstadt, Germany) and 25% methanol (v/v) plus 0.3% H₃PO₄ (v/v) in water as eluent. A flow rate of 0.4 mL min⁻¹ was applied. Signals were detected at 210 nm. Under these conditions, the retention times were as follows: fumarate, 6.3 min; vanillate, 8.3 min; 4-hydroxyanisole, 8.6 min; isovanillate, 8.8 min; syringate, 9.2 min. For separation of fumarate, vanillate, and its demethylated end product 3,4-dihydroxybenzoate, a gradient with a flow rate of 0.3 mL min⁻¹ was applied: 0–11 min, 0.3% H₃PO₄; 11–21 min, 0–25% methanol plus 0.3% H₃PO₄; 21–30 min, 25% methanol plus 0.3% H₃PO₄. Under these conditions, the retention times were as follows: fumarate, 7.1 min; 3,4-dihydroxybenzoate, 14.9 min; vanillate, 24.0 min.

Nitrite was measured by a diazotization reaction described by Lunge (1904). The colorimetric reaction was started by the addition of 50 μL of 1% (w/v) sulphanilic acid and 50 μL of 0.3% (w/v) 1-naphthylamine to 50 μL of the diluted sample. After 5 min incubation at room temperature, absorption was measured at a wavelength of 525 nm on a VERSAmax tunable microplate reader (Molecular Devices, Biberach an der Riss, Germany).

For nitrate determination, the method of Bosch Serrat (1998) was followed. Diluted samples were mixed in a 1 : 1 ratio with 200 μL of a chloride solution consisting of 0.28 M NaCl dissolved in 9.35% (v/v) H₃PO₄. Subsequently, 1 mL of 0.55% (w/v) resorcinol dissolved in 62.5% (v/v) H₂SO₄ was added to each sample. The mixtures were incubated at 95 °C for 8 min. They were allowed to cool down for 20 min and absorbance was measured at 505 nm. As in this assay both nitrite and nitrate were detected, nitrate concentration was calculated by subtracting the nitrite concentration from the value obtained.

Thiosulfate was measured according to Quentin & Prachmayr (1964). Diluted samples were mixed in a 3 : 1 ratio with 0.006% (w/v) methylene blue in 6 M HCl. After 3 h of incubation at room temperature, the absorbance was measured at 660 nm.

In silico mapping of putative O-demethylase operons

Mapping of putative O-demethylase operons of Desulfitobacterium spp. was carried out with the NCBI BLAST server (Altschul et al., 1990) and the Integrated Microbial Genomes (IMG) system (Markowitz et al., 2014). As template for blastp searches, the amino acid sequences of Dhaf_4610, Dhaf_4611, Dhaf_4612, and Dhaf_2573, which represent the protein components of an O-demethylase previously identified in D. hafniense DCB-2 (Studenik et al., 2012), were used. The putative O-demethylase operons identified in the genomes of D. hafniense DCB-2 (Kim et al., 2012), DP7, PCE-S, PCP-1, TCE-1, TCP-A, and Y51 (Nonaka et al., 2006) as well as D. dehalogenans, D. dichloroeliminans, and D. metallireducens were compared in terms of organization and similarity of the encoded proteins.

Results

Growth of Desulfitobacterium spp. with phenyl methyl ethers as electron donors

Desulfitobacterium hafniense (strains DCB-2, DP7, G2, PCE-S, PCP-1 TCP-A, Y51), D. chlororespirans, D. dehalogenans, and D. metallireducens were cultivated with 4-hydroxyanisole, syringate, vanillate, or isovanillate in the presence of fumarate, nitrate, or thiosulfate as electron acceptors (Table 1). The results were obtained from at least duplicates of the experiments. All species with the exception of D. metallireducens were able to grow on at least three of the supplied phenyl methyl ethers by cleaving the ether bond. Vanillate and isovanillate were O-demethylated to 3,4-dihydroxybenzoate, 4-hydroxyanisole to hydroquinone, and syringate via 3,4-dihydroxy-5-methoxybenzoate to 3,4,5-trihydroxybenzoate. The stoichiometry of demethylated end product formed per substrate consumed was about 1 : 1. Fumarate as well as nitrate or thiosulfate served as terminal electron acceptors for methyl group oxidation following O-demethylation. The substrate spectra and the demethylation rates of the Desulfitobacterium strains used in this study are summarized in Table 1. O-demethylation was independent of the type of electron acceptor used; however, the rates of phenyl methyl ether consumption were different. For most phenyl methyl ethers, the rates of demethylation and cell growth were highest in the presence of fumarate followed by nitrate and thiosulfate. As reported before for D. hafniense DCB-2 and PCE-S (Neumann et al., 2004; Kreher et al., 2008), no demethylation of the substrates occurred when CO₂ was applied as sole terminal electron acceptor (data not shown). Figure 2 shows the consumption of vanillate by D. hafniense Y51 in the presence of fumarate, nitrate, or thiosulfate, respectively. The stoichiometries of fumarate/nitrate/thiosulfate reduced per mol methyl group consumed were usually higher than the values expected from the substrate conversion according to the following equations:

The deviant stoichiometry with fumarate can be explained by an additional fermentation/disproportionation of this substrate, which was observed in cultures with fumarate as sole energy source (data not shown). For nitrate and thiosulfate, a higher stoichiometry is feasible by incomplete reduction of these electron acceptors. All species and strains tested in this study, with the exception of D. metallireducens, were able to demethylate vanillate in the presence of Fe(III) in the media (Table S2).

Genetic background of O-demethylation in desulfitobacteria

Bacterial O-demethylases consist of four different proteins: two methyltransferases (MT I and MT II), a corrinoid protein (CP), and an activating enzyme (AE). The genes encoding CP, MT I, and MT II are usually organized in an operon. The gene encoding AE is located elsewhere (Schilhabel et al., 2009). In most cases, several O-demethylase operons are present in genomes of O-demethylating bacteria, but only the gene product of one AE gene seems to be responsible for the reduction of the different O-demethylase corrinoid proteins (Schilhabel et al., 2009; Studenik et al., 2012; Nguyen et al., 2013). For D. hafniense DCB-2, up to 17 putative demethylase operons have earlier been identified; however, it is not possible to discriminate between O-, S-, N-, and Cl-demethylases (Studenik et al., 2012). The D. hafniense strains DP7, PCE-S, PCP-1, TCE1, TCP-A, and Y51 harbor more than 10 putative demethylase operons, whereas the genomes of D. dichloroeliminans and D. dehalogenans contain 4 and 7, respectively (Table 2). In D. metallireducens, neither a demethylase operon nor a gene encoding for a corrinoid protein could be identified. The operon encoding for the previously characterized O-demethylase of D. hafniense DCB-2 (locus tags Dhaf_4610 [MT I], Dhaf_4611 [CP], and Dhaf_4612 [MT II]; Studenik et al., 2012) was present in each D. hafniense genome screened. A closely related O-demethylase operon with a sequence identity on protein level of 88–97% was also found in D. dehalogenans and D. dichloroeliminans. COG3894 protein-encoding genes, which are usually not part of the O-demethylase operons but often located in the vicinity, were also identified in the genomes. The COG3894 protein family comprises the AE proteins as well as other RACE proteins (RACE = reductive activator of corrinoid-dependent enzymes) mediating the reduction of protein-bound corrinoids involved in anaerobic methyltransferase reactions, for example, in the Wood-Ljungdahl pathway (Schilhabel et al., 2009; Hennig et al., 2012). With the exception of D. dichloroeliminans and D. metallireducens, 5–6 genes putatively encoding COG3894 proteins were detected in the genomes (Table 2). One of these genes was always found as part of the carbon monoxide dehydrogenase/acetyl-CoA synthase operon (Table S1).

Discussion

In this study, it was shown that O-demethylation is a common metabolic feature of the genus Desulfitobacterium. Among the strains tested, only D. metallireducens failed to utilize phenyl methyl ethers as electron donors even when an appropriate electron acceptor was supplied. As reported earlier for D. hafniense DCB-2 and PCE-S (Neumann et al., 2004; Kreher et al., 2008), all strains tested were not able to use CO₂ as electron acceptor. This finding distinguishes the O-demethylation pathway of desulfitobacteria from that of acetogenic bacteria. Fumarate and chlorinated compounds, which were reported to serve as electron acceptors for O-demethylation in desulfitobacteria (Neumann et al., 2004), can be substituted by nitrate, thiosulfate, or Fe(III). In most growth experiments, no clear preference was observed for one of the electron acceptors applied. As in natural habitats of Desulfitobacterium spp. such as forest soils (Lanthier et al., 2001), nitrate and Fe(III) are expected to be the dominant compounds among the electron acceptors tested; O-demethylation by desulfitobacteria in soil is most likely coupled to nitrate or Fe(III) reduction.

Of the four substrates tested, syringate, vanillate, and isovanillate were preferred by most desulfitobacteria and O-demethylated to their corresponding phenolic derivatives 3,4,5-trihydroxybenzoate and 3,4-dihydroxybenzoate, respectively. The ability of the species and strains tested to degrade these compounds might have resulted from an adaptation to the environment, as syringate, vanillate, and isovanillate are among the most abundant products of lignin degradation (Kögel, 1986).

Several putative O-demethylase operons were identified in the genomes of Desulfitobacterium spp. The highest numbers were found in the D. hafniense strains analyzed. As upon lignin degradation different phenyl methyl ethers are released into the soil matrix (Chen et al., 1982, 1983), the presence of multiple putative O-demethylases may be favorable for the conversion of different substrates as it was shown for A. dehalogenans (Engelmann et al., 2001; Schilhabel et al., 2009). It is possible that the high number of putative O-demethylase operons in D. hafniense may be derived from gene duplication that resulted in an adaptation to the natural habitat, for example, forest soils (Lanthier et al., 2001), where lignin degradation products are abundant carbon and energy sources (Whitehead, 1964).

From studies with A. dehalogenans and M. barkeri, it was assumed that only one AE protein is responsible for the reduction of inadvertently oxidized cofactors of different demethylase corrinoid proteins (Ferguson et al., 2009; Schilhabel et al., 2009; Nguyen et al., 2013). In desulfitobacteria, usually 5–6 copies of AE-encoding genes were identified. As shown before for D. hafniense DCB-2 (Studenik et al., 2012), it is expected that in desulfitobacteria also, just one gene product is responsible for the O-demethylase corrinoid reduction. Further gene products may be involved in the reductive activation of other corrinoid-containing proteins like the corrinoid-iron/sulfur protein of the reductive acetyl-CoA pathway (Hennig et al., 2012).

Desulfitobacteria were first isolated from soils contaminated with halogenated organic compounds (for a review see Villemur et al., 2006). The capability of these microorganisms to cleave off methyl groups from aromatic compounds was discovered as a side reaction during cultivation on methoxylated chloroaromatics (Dennie et al., 1998; Milliken et al., 2004a, b). Later, it was demonstrated that two strains of D. hafniense can couple the O-demethylation of phenyl methyl ethers to energy conservation and growth (Neumann et al., 2004). In this study, O-demethylation was found to be a common metabolic feature of the genus Desulfitobacterium, which was coupled to nitrate, thiosulfate, or Fe(III) reduction. Hence, besides acetogens, also desulfitobacteria might contribute to the degradation of phenyl methyl ethers in soil, where nitrate, thiosulfate, and/or Fe(III) are present. This assumption is supported by the enrichment of desulfitobacteria from forest topsoils using O-demethylation as the growth-selective process (unpublished data).

Acknowledgements

This work was supported by the International Leibniz Research School (ILRS) for Microbial and Biomolecular Interactions (Jena, Germany). We are grateful to Hauke Smidt, Thomas Kruse (Wageningen University, the Netherlands), and Tobias Goris (Friedrich Schiller University Jena, Germany) for providing the genome sequence of Desulfitobacterium hafniense strain PCE-S (Contig Accession Number LK996017-LK996040, European Nucleotide Archive EMBL-EBI). All authors declared that they have no conflict of interests.

References

Author notes