Coal-bed methane (CBM), a form of natural gas distributed in coal seams or adjacent sandstones, is a relatively untapped energy source with a large potential: The global reserves in 2014 were estimated at 50 trillion m
3, equivalent to 11% of conventional natural gas resources (
1). Large-scale CBM production has been implemented in the United States, Canada, Australia, and other countries worldwide. The contribution of biogenic methane to CBM is quite large (
2,
3); geochemical studies have estimated that 40% of CBM produced in the United States is of microbial origin (
4). Live microbial communities are present in coal seams and are associated with methanogenesis from coal in subsurface environments (
5–
10). Geomicrobiological studies have shown that enhanced CBM production in coal seams might be achieved by the stimulation of methanogenic activity (
11). Although extensive efforts have been made to develop this technology, very little is known about what components of coal can be used for methanogenesis and which microorganisms possess the metabolic capabilities to do so.
Coal is an extremely complex and heterogeneous material whose structure consists of single and condensed aromatic rings (
12,
13). Aromatic compounds in coal are derived from lignin monolignols and are often substituted with hydroxyl, methoxy, and carboxyl groups (
14,
15). Methoxy groups are especially abundant and common in immature coal (
14,
16). Because methanogenesis from coal tends to occur in immature coal rather than in mature coal (
4,
17), coal-bed microorganisms may produce methane from methoxy groups. Methanogenic microorganisms in CBM fields are commonly dominated by methylotrophic methanogens belonging to the archaeal order
Methanosarcinales (
5,
18,
19). The methylotrophic methanogens are capable of using methyl compounds such as methanol, methylamines, and/or dimethylsulfide (
20), but it is unclear whether they can directly use methoxylated aromatic compounds (MACs) as substrates.
To investigate the possibility of MACs as substrates for methylotrophic methanogens, we tested the methane production ability of one archaeal isolate (
Methermicoccus shengliensis strain AmaM) obtained from a high-temperature deep subsurface oil reservoir in this study (fig. S1), and of 10 type strains belonging to the genera
Methanosarcina,
Methanolobus,
Methanohalophilus,
Methanosaeta,
Methanomicrococcus,
Methanococcoides,
Methanohalobium,
Methanosalsum,
Methanomethylovorans, and
Methermicoccus (i.e.,
M. shengliensis strain ZC-1) (
21) in the order Methanosarcinales using seven types of MACs (2-methoxy-benzoate, 3-methoxy-benzoate, 4-methoxy-benzoate, 3,4,5-trimethoxy-benzoate, 3,4,5-trimethoxy-cinnamate, 1,2,3-trimethoxy-benzene, and 3,4,5-trimethoxy-benzylalcohol) as substrates. We only observed substantial methane production in the incubation of
Methermicoccus shengliensis strains AmaM and ZC-1 with seven and six types of MACs, respectively (
Fig. 1A). To investigate their substrate ranges, we incubated the strains AmaM and ZC-1 with 40 commercially available MACs. The results showed that the strains AmaM and ZC-1 used 35 and 34 types of MACs, respectively (
Fig. 1B and table S1).
To confirm whether these “methoxydotrophic methanogens” could produce methane from coal, we incubated
M. shengliensis AmaM with coals of different maturity levels: lignite (most immature) and subbituminous and bituminous (most mature) coals (fig. S2). A small but substantial amount of methane (7.5 to 10.8 μmol/g-coal) was produced in all of the three lignites (L-A, L-B, and L-C) and even in the subbituminous coal S-A and the bituminous coal B-A (
Fig. 2A).
M. shengliensis AmaM used MACs, methanol, and methylamines as substrates for methanogenesis (
Fig. 1 and fig. S1), which suggests that coal may have provided some of these substrates. We analyzed growth media with coal before incubation for the 26 types of MACs usable for
M. shengliensis AmaM (marked with asterisks in
Fig. 1B) by gas chromatography–mass spectrometry (GC-MS). We detected either two or three types of methoxylated benzoates in each medium from which methane production was observed (
Fig. 2B). The total concentrations of methanol and methylamines as well as of MACs detected in the media were too low to account for the concentrations of methane produced in the coal cultures alone (
Fig. 2); for example, in bituminous coal culture B-A, 0.09 μmol/g-coal of MACs was detected but 9.39 μmol/g-coal of methane was produced. This result suggests that
M. shengliensis AmaM produced methane from undetectable MACs dissolved in the media as well as those chemically or physically bound to the coal surface. This was supported by the ability of
M. shengliensis AmaM to use a wide variety of MACs (>30 types of MACs) for methanogenesis (
Fig. 1B).
If methoxydotrophic methanogenesis proceeds in analogy to methylotrophic methanogenesis, it is expected from stoichiometry that ¾ mol of methane is produced from 1 mol of the methoxy group (4Ar-OCH
3 + 2H
2O → 4Ar-OH + 3CH
4 + CO
2, where Ar denotes any aromatic group). During incubation with 2-methoxy-benzoate,
M. shengliensis AmaM produced methane and 2-hydroxy-benzoate with a decrease in 2-methoxy-benzoate (
Fig. 3). The 2-hydroxy-benzoate produced was nearly equivalent to the 2-methoxy-benzoate consumed, which suggests that
M. shengliensis AmaM produced methane via
O-demethylation of the methoxy group.
We conducted stable isotope tracer experiments to elucidate the mode of metabolism in the methoxydotrophic methanogenesis. In the incubation of
M. shengliensis AmaM with 2-[
13C]methoxy-benzoate, the
13C contents of methane increased with increasing contents of
13C at the methoxy group (
Fig. 4A), indicating that the methoxy carbon was incorporated into methane. However, the incorporation efficiency from the methoxy group to methane, estimated as the slope of the regression line between the
13C contents of methane and the methoxy group, was 63.6% (
Fig. 4A). By contrast, in the incubation with [
13C]methanol, we estimated that nearly all (96.4%) of the methane carbon came from the substrate methanol. This implies that additional carbon (other than methoxy carbon) is incorporated into methane during MACs-driven methanogenesis.
To identify this additional carbon, we evaluated whether CO
2 was incorporated into methane via CO
2 reduction. We incubated
M. shengliensis AmaM with 2-methoxy-benzoate or methanol in the medium amended with [
13C]bicarbonate. Although the
13C contents of methane increased only slightly in the presence of methanol, those of methane in the presence of 2-methoxy-benzoate increased far more substantially (
Fig. 4B), indicating that CO
2 is also incorporated into methane via CO
2 reduction in methoxydotrophic methanogenesis. The incorporation efficiency from CO
2 into methane, estimated as the slope of the regression line (
Fig. 4B), was 29.6%. Considering that the incorporation efficiency from the methoxy group to methane is 63.6% (
Fig. 4A), approximately one-third of methane carbon is derived from CO
2 and two-thirds of methane carbon from the methoxy group. In the incubation of
M. shengliensis AmaM with a variety of coals in the presence of [
13C]bicarbonate, we observed the production of highly
13C-enriched methane, which indicates that methane produced from coal was mostly derived from MACs (fig. S3).
Further tracer experiments incubating
M. shengliensis AmaM with 2-methoxy-benzoate in the medium amended with [2-
13C]acetate revealed a small but substantial incorporation of the acetate methyl carbon into methane (fig. S4). The acetate concentration in the medium did not change during growth, implying no intentional uptake of extracellular acetate through acetoclastic methanogenesis. We therefore infer that acetyl–coenzyme A (CoA) could be a catabolic intermediate in the methoxydotrophic methanogenesis.
M. shengliensis AmaM genes encode acetyl-CoA synthesis, acetyl-CoA oxidation, and CO
2-reducing methanogenesis, but its genome lacks known acetogen-associated genes for
O-demethylation of the methoxy group (e.g., Mtv system) (
22,
23) and electron transport systems (e.g., Rnf, complete Fpo complex, etc.) (
20) necessary for conventional methanogenesis (table S2). Although details of the metabolic pathway in the methoxydotrophic methanogenesis remain elusive, all the results indicate that the mode of metabolism is clearly different from the conventional methylotrophic methanogenesis.
Our finding that MACs serve as a direct substrate for methanogens may not be limited to coal-bed environments. In the deep subsurface, MACs are contained in sedimentary organic matter derived from lignin in higher plants, namely kerogen, with quantitative variation depending on the maturity (
24). In fact, alkyl-methoxy-phenols with a short C
1-C
3 chain have been detected in the pyrolysates of immature kerogen extracted from a Cretaceous (Cenomanian) black shale (
25). Kerogen is ubiquitous in sediments and accounts for most of the organic matter in subsurface environments (
26). Microorganisms from the genus
Methermicoccus and related clones have often been detected in deep subsurface environments worldwide (fig. S5). Methoxydotrophic methanogenesis may therefore play an important role in the biogeochemical carbon cycle in Earth as well as in the formation of biogenic gas, which accounts for more than 20% of natural gas resources, including CBM (
27).
Acknowledgments
We thank Japan Petroleum Exploration Co., Ltd. (JAPEX) for providing samples from an oil reservoir; F. Nozawa, K. Shuin, Y. Shinotsuka, T. Ujiie, and X. Meng for technical support; and M. Nobu for valuable comments. Supported by JSPS KAKENHI grants JP26709070, JP25289333, JP26710012, and JP26106004. M. shengliensis strain AmaM has been deposited as accession number NBRC 112467 in the Biological Resource Center, National Institute of Technology and Evaluation (NBRC). Genomic data of M. shengliensis strain AmaM are available in the Integrated Microbial Genomes system of the U.S. Department of Energy Joint Genome Institute with ID no. 2516653088 (Gold Project ID: Gp0021722).