Hindawi Archaea Volume 2021, Article ID 8865133, 13 pages https://doi.org/10.1155/2021/8865133

Research Article Analysis of a and an Actinobacterium Dominating the Thermophilic Microbial Community of an Electromethanogenic Biocathode

Hajime Kobayashi ,1 Ryohei Toyoda,1 Hiroyuki Miyamoto,1 Yasuhito Nakasugi,1 Yuki Momoi,1 Kohei Nakamura,2 Qian Fu,3 Haruo Maeda,4 Takashi Goda,1 and Kozo Sato1,5

1Department of Systems Innovation, Graduate School of Engineering, The University of Tokyo, Tokyo 113-8656, Japan 2Faculty of Applied Biological Sciences, Gifu University, Yanagido, Gifu 501-1193, Japan 3Key Laboratory of Low-Grade Energy Utilization Technologies and Systems, Chongqing University, Ministry of Education, Chongqing 400044, China 4INPEX Corporation, 9-23-30 Kitakarasuyama, Setagaya-ku, Tokyo 157-0061, Japan 5Frontier Research Center for Energy and Resource (FRCER), Graduate School of Engineering, The University of Tokyo, Tokyo 113-8656, Japan

Correspondence should be addressed to Hajime Kobayashi; [email protected]

Received 6 July 2020; Revised 9 February 2021; Accepted 15 February 2021; Published 2 March 2021

Academic Editor: William B. Whitman

Copyright © 2021 Hajime Kobayashi et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Electromethanogenesis refers to the bioelectrochemical synthesis of methane from CO2 by biocathodes. In an electromethanogenic system using thermophilic microorganisms, metagenomic analysis along with quantitative real-time polymerase chain reaction and fluorescence in situ hybridization revealed that the biocathode microbiota was dominated by the methanogen Methanothermobacter sp. strain EMTCatA1 and the actinobacterium Coriobacteriaceae sp. strain EMTCatB1. RNA sequencing was used to compare the transcriptome profiles of each strain at the methane-producing biocathodes with those in an open circuit and with the inhibitor 2-bromoethanesulfonate (BrES). For the methanogen, genes related to hydrogenotrophic methanogenesis were highly expressed in a manner similar to those observed under H2-limited conditions. For the actinobacterium, the expression profiles of genes encoding multiheme c-type cytochromes and membrane-bound oxidoreductases suggested that the actinobacterium directly takes up electrons from the electrode. In both strains, various stress-related genes were commonly induced in the open-circuit biocathodes and biocathodes with BrES. This study provides a molecular inventory of the dominant species of an electromethanogenic biocathode with functional insights and therefore represents the first multiomics characterization of an electromethanogenic biocathode.

1. Introduction conversion (power to gas) and CO2 utilization have been proposed [2, 3]. Electromethanogenesis refers to the bioelectrochemical Hydrogenotrophic , particularly those synthesis of methane (CH4) from carbon dioxide (CO2) belonging to the family Methanobacteriaceae, appear to play at the biocathodes of bioelectrochemical systems [1]. In a primary role in electromethanogenesis and are com- such systems, catalytic microbes present on the cathode monly detected as the dominant methanogen in biocathodes surface typically utilize electrons from the electrodes and [1, 4, 5]. Recent studies of biocathodes inoculated with reduce CO2. Because these biocathodes enable highly pure and cocultures revealed electron transfer pathways efficient conversion of electrical energy into methane, from the electrodes to methanogens. For example, direct promising applications related to renewable electricity electron uptake from negatively polarized electrodes was 2 Archaea demonstrated in a methanogen of the family Methanobac- using 250 mL glass bottles. Two-chamber reactors compris- teriaceae, namely, the iron-corroding Methanobacterium- ing two identical 300 mL glass bottles separated by a pre- like archaeon strain IM1 [6]. While some methanogens, treated proton exchange membrane (12.5 cm2,Nafion 117, including Methanococcus maripaludis, lack this ability [7], DuPont Co., Wilmington, DE, USA) were also constructed. such as hydrogenases and the heterodisulfide The anodes and cathodes were composed of plain carbon reductase complex from M. maripaludis [7–9] adsorbed cloth (4 × 10 cm, TMIL Ltd., Ibaraki, Japan). Each electrode on the cathode surface have been shown to catalyze the was connected to the circuit via a titanium wire (0.5 mm in production of soluble electron mediators such as H2 and diameter, Alfa Aesar, Ward Hill, MA, USA), which was formate using electrons from the electrodes, which can in directly fastened to the end of the electrode without glue. turn be utilized by methanogens. Such mediators can also The internal resistance between the electrodes and titanium be produced by other microbes capable of direct electron wires was less than 3.0 Ω. All reactors were sealed with butyl uptake, such as the iron-corroding sulfate reducer Desulfo- rubber stoppers and aluminum seals, and their headspaces fi pila corrodens strain IS4 [10, 11]. were lled with N2/CO2 (80 : 20). The inoculated reactors Despite the basic knowledge gained from defined culture were operated at 55°C in the fed-batch mode, in which the systems, for practical applications, it is important to under- medium was exchanged with the fresh medium when current stand the mechanisms of electromethanogenesis in the production was attenuated to the background level. A multispecies microbial consortia enriched on biocathodes. magnetic stirrer was continuously used in each chamber to Characterizing the functions of these constituent species is provide sufficient mixing during the incubation. expected to lead to the identification of new catalysts, includ- ing methanogens and other species capable of electron 2.1.1. Operation of the Single-Chamber Reactor. The con- uptake, as well as microbes with auxiliary functions (e.g., oxy- struction of active biocathodes was first initiated in the gen scavengers) and detrimental species (e.g., producers of single-chamber reactor. The initial source of the microorgan- undesirable products). These microbes may represent poten- isms was the effluent of a preexisting bioelectrochemical tial targets for the functional engineering of biocathodes for reactor, which was originally inoculated with formation better performance and robustness. In two types of bio- water from a petroleum reservoir [16]. Then, 25 mL of fi cathodes, namely, a CO2- xing aerobic biocathode and a bio- inoculum and 125 mL of sterile anaerobic medium contain- cathode primarily producing acetate, metagenomic analyses ing 0.136 g/L of KH2PO4, 0.54 g/L of NH4Cl, 0.2 g/L of have revealed the compositions and metabolic capabilities MgCl2·6H2O, 0.147 g/L of CaCl2·2H2O, 2.5 g/L of NaHCO3, of the surface microbial consortia [12–15]. In addition, active and 10 mL/L of Wolfe’s Solution supplemented with fi metabolic pathways, including those involved in CO2 xation 0.8 g/L sodium acetate were added to the reactor. A constant and electron transfer, and possible interspecies interactions voltage of 0.7 V was applied using a digital power supply have been inferred via metatranscriptomic and metaproteo- (Array 3645A, Array Electronics, Nanjing, China). A fixed mic analyses [12, 15], providing crucial insights into the in external resistance of 1.0 Ω was connected to the circuit. situ functions of the various community members present The voltage across the resistance was recorded every 5 min at those biocathodes. using a multimeter (34970A, Agilent Technologies, Santa We previously reported the identification of a thermo- Clara, CA, USA). philic microbial consortium that was capable of catalyzing electromethanogenesis at 55°C with a cathode poised at 2.1.2. Operation of the Two-Chamber Reactor. After three −0.35 V versus a standard hydrogen electrode (SHE) [16]. fed-batch cycles in the single-chamber reactor, the bio- The results revealed that both methanogenesis and electron cathode was gently rinsed using sterile anaerobic medium consumption at the biocathode were dependent on the and then transferred to the cathode chamber of the two- presence of CO2 and were strongly inhibited by the methano- chamber reactor for further analyses. In this setup, each genesis inhibitor 2-bromoethanesulfonate (BrES). These anode and cathode chamber was filled with 200 mL of the findings suggested that the electrons from the cathode were anaerobic medium (with no sodium acetate). For the anode, primarily consumed for methanogenesis. Initial evaluation a new abiotic electrode was used. An Ag/AgCl reference elec- of relevant 16S rRNA clone libraries suggested that a metha- trode (1 M KCl) with a potential of +0.20 V versus an SHE at nogen related to Methanothermobacter, along with several 55°C was inserted into the cathode chamber. The biocathode, other bacterial species, was enriched on the biocathode sur- anode, and Ag/AgCl electrode were connected to a potentio- face. Therefore, in this study, we aim to characterize the stat (HSV-110, Hokuto Denko, Japan) as the working, coun- primary constituents of this consortium and intend to gain ter, and reference electrodes, respectively. The biocathode insight into their respective roles in the electromethanogen- was poised at a constant potential of −0.5 V versus an SHE. esis process. The reactor was operated in the fed-batch mode. In the study, the biocathodes were sacrificed for nucleic acid extraction 2. Materials and Methods and microscopic analyses after three fed-batch cycles in the two-chamber reactors. 2.1. Reactor Design and Operation. The specific characteris- tics and operating conditions of the reactors used in this 2.2. Analytical Measurements. The gas composition in the study were generally the same as those described in our pre- reactor headspaces was analyzed using a gas chromato- vious study [16]. Single-chamber reactors were constructed graph (GC-2014 equipped with a ShinCarbon ST column; Archaea 3

Shimadzu, Kyoto, Japan) for each experiment. The pressure 5 μL of Premix Ex Taq (TaKaRa, Kyoto, Japan), and PCR- in the reactor headspace was measured using a digital grade sterilized water. PCR amplification was performed pressure sensor (AP-C40; Keyence, Osaka, Japan). In the as follows: an initial 30 s of incubation at 95°C, 40 cycles of two-chamber reactor, cyclic voltammetry (CV) was per- denaturation for 5 s each at 95°C, and annealing/extension formed using a standard three-electrode system. A potentio- for 30 s at 60°C. The amplification efficiency of the primer- stat (HSV-110) was used in conjunction with the following probe sets was 1.80–1.86. Three separate trials were parameters: equilibrium time of 99 s, scan rate of 1 Mv/s, conducted for each sample. Standard curves for each and scanning range of −0.7 to −0.2 V versus an SHE. assay were constructed using a synthetic 928 bp DNA fragment containing the target regions of the 16S rRNA 2.3. Metagenome Analysis. Whole-genome shotgun sequenc- genes of the Methanothermobacter sp. strain EMTCatA1 ing of the biocathode-associated microbial community, (for the archaeal and Methanobacteriales assays) and assembling, and annotation of the metagenome have been Coriobacteriaceae sp. strain EMTCatB1 (for the bacterial described in previous reports [17, 18]. DNA was extracted and Coriobacteriaceae species assays). from two independent biocathodes actively producing meth- ane at −0.5 V versus an SHE using a DNeasy PowerMax Soil 2.5. Microscopic Analyses. Fluorescence in situ hybridization Kit (Qiagen, Hilden, Germany). The extracted DNA was (FISH) and scanning electron microscopy (SEM) were sequenced in an Illumina HiSeq 2000 sequencer (150 bp performed. FISH was used to identify Methanothermobac- paired-end sequencing). Adapter and quality trimming of ter-related methanogens using a probe specific to the 16S the reads was performed using Cutadapt (version 1.8.3) rRNA of the order Methanobacteriales (MB311, Table S1) [19]. The 16S rRNA gene amplicons of the biocathode- [30]. To improve the penetration of the probe into the associated communities were sequenced using the extracted methanogens having pseudomurein cell walls, the FISH DNAs as the template, as previously described [20]. procedure was modified to include an enzymatic Approximately 395 million trimmed reads, approxi- pretreatment of 4% (w/v) paraformaldehyde-fixed samples mately 60 gigabase pairs (Gbp), were used for the metage- with recombinant pseudomurein endopeptidase (rPeiW), as nomic binning. The MetaPhlAn2 tool (version 2.2.0) [21] previously described [31]. A probe specific to the 16S rRNA was used to reveal the composition of the biocathode- of the Coriobacteriaceae sp. strain EMTCatB1 (B1_648, associated consortium from the unassembled metagenomic Table S1) was designed using the ARB program [28]. The reads. For assembling, reads were first downsampled to 400 specificity and efficiency of the probe were preevaluated in megabase pairs (Mbp); as a result, sequences from relatively silico using the SILVA TestProbe (version 3.0) [32] and the minor species were reduced. Then, the reads were assembled mathFISH tool [33]. For the permeabilization of gram- using the Velvet package (version 1.2.10) [22], followed by positive cell walls of the Actinobacteria, samples were fixed gap filling using the Sealer tool (ver.2.0.2) [23] and a quality in 96% ethanol (without formaldehyde fixation) and check using the REAPR tool (version 1.0.18) [24]. The scaf- pretreated with 10 mg/mL of lysozyme (Sigma-Aldrich, St. folds were annotated using Prokka (version 1.13) [25]. For Louis, MO, USA) for 60 min, followed by digestion with the metabolic pathway analysis, the proteins encoded in the 10 U/mL of achromopeptidase (Sigma-Aldrich) for 30 min draft genomes were mapped onto the Kyoto Encyclopedia before hybridization [34]. Both probes were labeled with of Genes and Genomes pathway database using the KEGG Alexa Fluor 488 (Thermo Fisher Scientific, Waltham, MA, Mapper [26]. Mauve was used for alignment of linealized USA), and 5 μM SYTO59 (Thermo Fisher Scientific) was genomes [27]. used for counterstaining. The micrographs were analyzed using Fiji [35] to estimate the size distributions of the 2.4. Quantitative Real-Time Polymerase Chain Reaction microbial cells on the cathode surfaces. For the SEM (qPCR). The shotgun-sequenced DNA and DNA samples analysis, the electrodes were first fixed with 2.5% (w/v) extracted from other six independent biocathodes were used glutaraldehyde and 2% (w/v) paraformaldehyde in 0.1 M as templates in the qPCR analyses, which were performed in phosphate buffer solution (PBS, pH 7.4) and processed as a LightCycler 480 system (Roche Diagnostics, Mannheim, described previously [16]. Germany). Group-specific primers and probes designed by Yu et al. [28] were used for the methanogen of the order 2.6. RNA Extraction and Transcriptome Analysis. RNA sam- Methanobacteriales (i.e., Methanothermobacter sp. strain ples were extracted from six independent biocathodes that EMTCatA1), total archaea, and total bacteria (listed in produced methane actively at −0.5 V versus an SHE. In this Table S1). The primers and probes specific to the experimental setup, biocathodes were clustered into two by Coriobacteriaceae sp. strain EMTCatB1 were designed establishing three experimental conditions, namely, closed- according to the 16S rRNA gene sequence of the strain and circuit (CC), open-circuit (OC), and BrES conditions. For its three closely related sequences (FJ638596, KM819482, this purpose, two biocathodes were directly subjected to and AY753404) using the ARB program [29]. The DNA RNA extraction (CC condition), whereas two other bio- concentrations were quantified using a Qubit 3.0 fluorometer cathodes were left in the open circuit for 5 h before RNA (Thermo Fisher Scientific) in conjunction with the Qubit extraction (OC condition). For the next two biocathodes, dsDNA HS Assay Kit (Thermo Fisher Scientific). Each 10 μL BrES was anoxically injected into the cathode chambers at a reaction mixture comprised 1 μL of the template DNA, final concentration of 12 mM, and the biocathodes were 300 nM of the specific primers, 200 nM of TaqMan probe, incubated with a poised potential of −0.5 V versus an SHE 4 Archaea for 5 h before RNA extraction (BrES condition). Before being tion, DNA was isolated from two independent biocathodes aseptically crushed, the biocathodes were soaked in Life- that were operated for more than 60 days with a poised Guard Soil Preservation Solution (Qiagen) to stabilize the potential of −0.5 V versus an SHE. Subsequent methane pro- 2 RNAs. Total RNA was then extracted using an RNeasy duction rates (20.2 and 24.2 mmol CH4/day/cm ) and CV PowerSoil Total RNA Kit (Qiagen). Residual DNA was scans showed that the electromethanogenic activities of the removed by DNase treatment using a TURBO DNA-free two biocathodes were similar to each other (Figure S1A). Kit (Thermo Fisher Scientific). From the total extracted Furthermore, 16S rRNA sequencing results suggested that RNA, mRNA was enriched by removing rRNA using a the microbial compositions of the two biocathodes were Ribo-Zero Kit (Illumina, San Diego, CA, USA). The enriched also similar (Figure S1B). Thus, the DNA from the two mRNA was then amplified using a MessageAmp II-Bacteria biocathodes was combined at the library preparation step RNA Amplification Kit (Thermo Fisher Scientific) and for whole-metagenome shotgun sequencing. further converted into cDNA using a SuperScript Double- The assembly and binning of the Illumina sequencing Stranded cDNA Synthesis Kit (Thermo Fisher Scientific). reads indicated that the metagenome of the consortium was The cDNA was then used to prepare a sequencing library almost exclusively dominated by the sequences derived using a TruSeq Stranded mRNA Library Prep Kit (Illumina). from the two dominant species (Figure 1(a)). Further Metatranscriptome sequencing was performed by using an assembling of the contigs, followed by gap filling, resulted Illumina HiSeq 2500 system (Illumina), which yielded in the reconstruction of the circular draft genomes of the 100 bp pair-end reads totaling 51.7–108.3 million reads two species (Figure S2) [17, 18]. Phylogenetic analyses of (Table S2). Eurofins Scientific Co. constructed the relevant the marker genes demonstrated that one species was an libraries and performed all the sequencing reactions. archaeon that was closely related to methanogens of the genus The RNA-seq reads were quality-filtered using Trim- Methanothermobacter (thus named Methanothermobacter sp. momatic (version 0.36) [36] and aligned against the strain EMTCatA1). The genome of this species was genomes of the Methanothermobacter sp. strain EMTCatA1 identified to encode enzymes needed for hydrogenotrophic fi and Coriobacteriaceae sp. strain EMTCatB1 (AP018336 and methanogenesis as well as a CO2 xation pathway that BDLO01000001 in the DDBJ/EMBL/GenBank database) proceeds via the incomplete reductive citrate cycle. The using BWA (version 0.7.17) [37]. The number of reads genome did not encode any apparent homolog of mapped onto the respective reference genomes was counted formate transporter or cytochrome. In particular, the using SAMtools [38] and is shown in Table S3. Kaiju Methanothermobacter sp. strain EMTCatA1 is closely (version 1.7.2) [39] was used for the taxonomic assignment related to the M. thermautotrophicus strain ΔH, a of the unmapped reads. StringTie (version 1.3.4d) [40] was model organism of thermophilic methanogens, sharing used to assemble the mapped reads into transcripts and 99% and 98% sequence identity of the 16S rRNA and calculate the relative abundances of the assembled mcrA genes, respectively. Moreover, the genome of the transcripts. Statistical analyses were performed using the Methanothermobacter sp. strain EMTCatA1 is highly edgeR package (version 3.16.1) [41]. Transcripts per million similar to the genome of the M. thermautotrophicus (TPM) was used to normalize the read counts to compare strain ΔH, sharing most of its genes with the M. the expression levels across the genes for transcriptomes thermautotrophicus strain ΔH in almost identical gene obtained from the same condition (Tables S4 and S5). orders (Figure S3). Another species identified was a Furthermore, the trimmed mean of M-values (TMM) was bacterium distantly related to Actinobacteria of the used to normalize the expression levels of each gene family Coriobacteriaceae (named Coriobacteriaceae sp. between the transcriptomes from the different conditions strain EMTCatB1) (Figure S4). The draft genome of the (Tables S6 and S7). The likelihood ratio test was used to Coriobacteriaceae sp. strain EMTCatB1 was found to determine the statistical significance of differences in gene encode homologs of the enzymes required for anaerobic expression. The Benjamini–Hochberg adjustment was used respiration (nitrite reduction), along with many putative to control the false discovery rate (FDR) due to multiple proteins (e.g., 18 c-type cytochromes). These hypothesis testing. For this, genes were considered to be results were largely unexpected because 16S rRNA gene differentially expressed if the expression level changed by amplicon sequencing had suggested that although the more than two-fold (log2 fold change = 1) and the FDR Methanothermobacter-related methanogen was shown to was <0.05. Hierarchical clustering with average linkage be a primary archaeal constituent, the bacterial population was performed using the Pearson correlation dissimilarity was composed of more diverse bacteria (Figure S1B). Indeed, metric, in which the cut-off distance, or dissimilarity, was in the 16S rRNA gene amplicons, actinobacterial species 0.25. represented only a relatively minor proportion of the sequences (Figure S1B). Underestimation of actinobacterial 3. Results and Discussion species in 16S rRNA gene amplicons has previously been reported and is presumably due to the high GC content of 3.1. Metagenomic Analysis of the Microbial Consortium of the these genomes [42, 43]. For example, the GC content is 67.2% Biocathodes. The biocathodes were subjected to metage- inthecaseoftheCoriobacteriaceae sp. strain EMTCatB1. nomic analysis to characterize the microbial composition and the metabolic potential of the surface microbial consor- 3.2. qPCR Analysis of the Biocathode Microbial Consortia. To tium. To minimize any potential effects due to sample varia- examine the abundances of these two species on the Archaea 5

Thermotoga (2%) Others(1%) 6

Methanothermobacter

sp. strain EMTCatA1

Archaea (31%) 5

Bacteria 395119242 reads (ca. 60 Gbp) 16S rRNA gene copy number (log10) 4 MG C1 C2 C3 C4 C5 C6 DNA samples Coriobacteriaceae sp. strain EMTCatB1 Arc Bac (66%) MBT COR (a) (b)

Figure 1: (a) Relative abundances and inferred taxonomies of the unassembled metagenome reads from the biocathode consortium. Kingdom-, genus-, and species-level identifications representing at least 1% relative abundance are shown. (b) The 16S rRNA gene copy numbers in the shotgun-sequenced DNA (MG) and the DNA samples extracted from the biocathodes (C1-C6). Copy numbers were quantified using primer-probe sets to detect the 16S rRNA genes of the domains Archaea (Arc) and bacteria (Bac), the order Methanobacteriales (MBT), and the Coriobacteriaceae sp. strain EMTCatB1 and related Actinobacteria (COR). biocathode surfaces, 16S rRNA gene copy numbers were esti- fluorescence micrographs showed that approximately 26% mated using qPCR, along with group-specific primers of the microbes were labeled with the probe for methanogens (Figure 1(b)). In addition to the shotgun-sequenced DNA of the family Methanobacteriales, thereby targeting the strain (MG in Figure 1(b)), DNA samples were extracted from bio- EMTCatA1 (Figures 2(a)–2(c)), and 68% of the cathode- cathodes of six independent electromethanogenic reactors associated microbes were labeled with the probe targeting (named C1~C6) and analyzed (C1-C6 in Figure 1(b)). CV the Coriobacteriaceae sp. strain EMTCatB1 (Figures 2(e)– confirmed the ability of the six biocathodes to catalyze 2(g)). In particular, the cells labeled with probes targeting electrochemical reactions (Figure S5), which is consistent EMTCatA1 were relatively longer filamentous cells or rods, with previously reported results [16]. The surface microbial with a median length of 2.8 μm, compared with the unlabeled colonization was confirmed by SEM (Figure S6). In cells, which had a median length of 1.5 μm (Figure 2(d)). This 20 pg of DNA, the copy numbers of archaeal 16S rRNA finding was consistent with the previous reports of other genes ranged from 3:4±0:05 × 105 to 4:9±0:2×105. Methanothermobacter species [44, 45]. In contrast, the cells The copy numbers of the 16S rRNA genes of the order labeled with probes targeting the EMTCatB1-labeled rod Methanobacteriales (including the genus Methanothermobacter) cells having a median length of 1.2 μm were mostly shorter ranged from 1:4±0:02 × 105 to 3:1±0:7×105,which than the unlabeled cells, which had a median length of corresponded to an average of 51% of the copy numbers of 2.5 μm (Figure 2(h)). Therefore, it can be concluded that the total archaeal 16S rRNA genes. Bacterial 16S rRNA the surface communities of the biocathodes were primarily gene copy numbers ranged from 9:3±0:07 × 104 to 2:9± composed of two types of cells: long rods or filamentous cells 0:03 × 105.FortheCoriobacteriaceae-related species, gene (typically longer than 1.6 μm) of the Methanothermobacter copy numbers ranged from 5:8±0:1×104 to 1:1±0:02 × sp. strain EMTCatA1 and relatively short rods (typically μ 105, which corresponded to an average of 52% of the shorter than 1.6 m) of the Coriobacteriaceae sp. strain EMT- total bacterial 16S rRNA gene copy numbers. It should CatB1 (Figures 2(d) and 2(h)). be noted that the primers of the 16S rRNA genes likely overestimated the abundance of Methanobacteriales, 3.4. Transcriptome Analysis of the Dominant Species on the and therefore, Archaea,asthedraftgenomeofthe Biocathode Surfaces. Metatranscriptomes of the biocathodes Methanothermobacter sp. strain EMTCatA1, contained under the CC, OC, and BrES conditions were analyzed to two copies of the 16S rRNA gene. Nonetheless, absolute gain insight into the respective roles of the dominant species quantification of the 16S rRNA gene copy numbers in electromethanogenesis. As we previously observed [16], supported the dominance of the two species on the methanogenesis ceased at the biocathodes under the OC biocathode surfaces. condition, in addition to the BrES condition, in which both methanogenesis and electron consumption processes at 3.3. FISH Analysis of Microbial Cells on the Biocathode the biocathode were inhibited (Figure S7). For all the Surfaces. FISH analysis further confirmed that the two metatranscriptomes, 69%–92% of the reads were mapped species represented the major constituents of the microbial onto the genomes of the two species (Table S2), further populations on the biocathode surfaces (Figure 2). The epi- indicating that they were the main metabolically active 6 Archaea

SYTO 59 Alexa Fluor 488 (n) (n) (2409) (50) (50) 100 5

4

3 50 MB311 2 Cell length ( � m) (Methanobacteriales) 1 Relative abundance (%)

0 0 (a) (b) (c) (d)

(2395) (60) (60) 100 5

4

3 50 B1_648 2 Cell length ( � m)

( Coriobacteriaceae sp.) 1 Relative abundance (%)

0 0 (e) (f) (g) +− Alexa 488 signal (h)

Figure 2: FISH analysis of the microbial populations present on the cathode surfaces and in the supernatants of the cathode chambers. Alexa Fluor 488-labeled probes targeting the order Methanobacteriales (MB311) and Coriobacteriaceae sp. strain EMTCatB1 (B1_648) were used. SYTO59 was used for counterstaining. (a, b, e, f) Epifluorescent micrographs of two representative fields separately capturing the fluorescence of Alexa Fluor 488 (green) and SYTO59 (red) using the probes MB311 and B1_648, respectively (scale bar: 1.0 μm). (c, g) Stacked bar charts (100%) of the relative abundances of the cells with the Alexa Fluor 488 signal (green-colored stacks) and without the signal (gray-colored stacks) on the cathode surfaces. (d, h) Box and whisker plots of the cell lengths with the Alexa Fluor 488 signal (+) and without the signal (−). The number of counted/measured cells (n) is indicated at the top of the panels (c), (d), (g), and (h).

species at the biocathode. The unmapped reads were reductase I (MRI) and two F420- assigned to diverse taxa (Figure S8). No taxon appeared dependent enzymes, respectively, were found to be expressed to be commonly overrepresented among the unmapped to a greater extent than the genes for their isofunctional reads. Therefore, as this study focused on the dominant enzymes (e.g., MRII and H2-dependent enzymes) (Figure 3). species, the unmapped reads were excluded from further In closely related M. thermautotrophicus strains, the analyses. expression of MRI and enzymes involved in cofactor F420- fi The transcriptome pro les under the CC condition were dependent reactions was induced under H2-limited conditions analyzed, with a particular focus on highly transcribed genes (e.g., syntrophic cocultures) [46–54]. related to energy metabolism and electron transfer to identify Among the transcriptomes of the Methanothermobacter the candidate genes involved in electromethanogenesis. sp. strain EMTCatA1, 146 genes were found to be differen- Hierarchical clustering of the differentially expressed genes tially expressed (Table S6). Based on the hierarchical (significance criteria: FDR < 0:05, fold change > 2) among clustering of the differential expression patterns, six clusters the conditions was used to estimate the influence of the elec- of the differentially expressed genes were identified tron supply from the cathode and the methanogenic activity (Figure 4, Table S8). The largest cluster (A1-I) consisted of on the physiology of the dominant species. 49 genes that were expressed at higher levels under the CC condition than at those under the BrES or OC conditions. 3.4.1. Transcriptome Analysis of the Methanothermobacter sp. Overall, 28 genes in the A1-I cluster encoded hypothetical Strain EMTCatA. For the Methanothermobacter sp. strain proteins of unknown functions. A1-IV, the second-largest EMTCatA1, the mRNAs related to hydrogenotrophic metha- cluster, consisted of 42 genes that were expressed at higher nogenesis were among the most highly abundant under the levels under both the BrES and OC conditions than under CC condition, with 15 of 46 methanogenesis-related genes the CC condition. Eight genes in the A1-IV cluster in the top 10% of abundant transcripts (Table S4). Notably, encoded homologs of various stress-related proteins, such mcrA (tca_01121), mtd (tca_01413), and mer (tca_01698), as chaperones and proteasomes (tca_00660, tca_00698, and which encoded the homologs of a subunit of methyl- tca_00826), antioxidant enzymes, and alternative redox Archaea 7

TPM normalized read counts in CC condition 0 1000 2000 3000 600050004000 tca_01502 fwdH tca_01503 fwdF tca_01504 CO fwdG 2 tca_01505 fwdD Fwd MFR, Fdred tca_01506 fwdA Fwd Fmd tca_01507 fwdC Fdox tca_01508 fwdB tca_00881 fmdE Formyl-MFR tca_00882 fmdC Fmd tca_00883 fmdB H4MPT tca_01215 fr Ftr FtrII Ftr MFR tca_00370 ftrII tca_00740 mch Mch Formyl-H4MPT tca_01413 mtd Mtd tca_01099 hmd Mch tca_01459 hmdII Hmd H2O tca_00476 hmdIII tca_01698 mer Mer Methenyl-H4MPT tca_01113 mtrH tca_01114 mtrG F420red H 2 tca_01115 mtrF F420 tca_01116 mtrA ox Mtr tca_01117 mtrB H4 Mtd Hmd tca_01118 mtrC Frh tca_01119 mtrD Methylene-H4MPT H2 tca_01120 mtrE F420red tca_01121 mcrA Mer tca_01122 mcrG F420ox tca_01123 mcrC MRI tca_01124 mcrD Methyl-H MPT 4 tca_01125 mcrB CoM-SH tca_01085 mrtA tca_01086 mrtG Mtr MRII H4MPT tca_01087 mrtD Hdr tca_01088 mrtB Methyl-S-CoM tca_01251 frhB tca_01252 CoB-SH frhG Frh MRI MRII tca_01253 frhD Mvh CoM-CoB tca_01254 frhA H2 2[H] tca_01335 hdrA CH4 tca_01814 hdrC Hdr tca_01815 hdrB tca_01089 mvhB tca_01090 mvhA Mvh tca_01091 mvhG tca_01092 mvhD (a) (b)

Figure 3: The methanogenesis pathway and gene expression patterns of methanogenesis-related genes of the Methanothermobacter sp. strain EMTCatA1. (a) Enzymes catalyzing respective reactions in the pathway are indicated in boxes. Fwd: tungsten-containing formyl-MFR dehydrogenase; Fmd: molybdenum-containing formyl-MFR dehydrogenase; Ftr: formyl-MFR:H4MPT formyltransferase; Mch: N5N10- methenyl-H4MPT cyclohydrolase; Mtd: F420-dependent N5N10-methylene-H4MPT dehydrogenase; Hmd: H2-dependent N5N10- methylene-H4MPT dehydrogenase; Mer: F420-dependent N5N10-methylene-H4MPT reductase; Mtr: N5-methyl-H4MPT methyltransferase; MRI: methyl-CoM reductase I; MRII: methyl-CoM reductase II; Hdr: heterodisulfide reductase; Mvh: methyl viologen-reducing hydrogenase; Frh: F420-reducing hydrogenase; MFR: ; Fd: ferredoxin; H4MPT: ; CoM-SH: coenzyme M; CoB-SH: . Isofunctional enzymes expressed at higher levels are highlighted in green. (b) TPM-normalized read counts of methanogenesis-related genes in CC condition.

proteins (tca_00140, tca_00141, tca_00142, tca_00723, and Although methanogenesis ceased under the OC and BrES tca_00821) (marked by red arrowheads in Figure 4) conditions, methanogenesis-related genes did not match (Table S8). This likely reflected the process of cellular our criteria for differential expression. For other methano- energy depletion due to the lack of methanogenesis under gens, the expression of methanogenesis genes was reported fi the OC and BrES conditions. No gene encoding an to be controlled by H2 availability and was not signi - apparent homolog related to direct electron uptake was cantly affected by treatments that inhibited methanogen- identified. esis [46, 55]. This was also consistent with the vital role 8 Archaea

CDS Gene name Annotated gene product Cluster z-score tca_00350 tca_00350 hypothetical protein 1.5 tca_01648 trm1_2 tRNA (guanine(26)-N(2))-dimethyltransferase tca_01517 nasC Assimilatory nitrate reductase catalytic subunit tca_01838 tca_01838 hypothetical protein tca_01277 tca_01277 hypothetical protein tca_00612 tca_00612 hypothetical protein tca_01576 SERB Phosphoserine phosphatase tca_00346 tca_00346 hypothetical protein tca_01387 tca_01387 hypothetical protein tca_00311 tuaB Teichuronic acid biosynthesis protein TuaB tca_00663 tca_00663 hypothetical protein tca_01001 tca_01001 hypothetical protein tca_00294 tca_00294 hypothetical protein tca_00339 tca_00339 hypothetical protein tca_00466 tca_00466 DUF5591 domain-containing protein tca_00626 tca_00626 hypothetical protein tca_01012 tca_01012 hypothetical protein tca_00128 tca_00128 hypothetical protein tca_00915 tca_00915 hypothetical protein tca_01548 tca_01548 hypothetical protein tca_00069 tca_00069 hypothetical protein A tca_01360 tca_01360 hypothetical protein tca_00304 mfF Putative biosynthesis glycosyltransferase MfF tca_00403 tca_00403 hypothetical protein tca_01111 tca_01111 hypothetical protein A1-I (49) tca_00057 tca_00057 hypothetical protein tca_01816 tca_01816 hypothetical protein tca_01821 cysP Sulfate permease CysP tca_01756 tca_01756 hypothetical protein tca_00127 mobA Molybdenum cofactor guanylyltransferase tca_00873 tca_00873 hypothetical protein tca_01626 rplK 50S ribosomal protein L11 tca_01696 tca_01696 hypothetical protein tca_00779 tca_00779 hypothetical protein tca_00331 tagF_2 CDP-glycerol:poly(glycerophosphate) glycerophosphotransferase tca_00044 rpsB 30S ribosomal protein S2 tca_00335 rfX_2 Putative O-antigen transporter tca_00421 tca_00421 hypothetical protein tca_01401 tca_01401 hypothetical protein tca_00064 tca_00064 hypothetical protein tca_00188 tca_00188 hypothetical protein tca_00338 mshA_2 D-inositol 3-phosphate glycosyltransferase tca_01573 tca_01573 hypothetical protein tca_00066 moaA_1 Cyclic pyranopterin monophosphate synthase tca_00094 tca_00094 hypothetical protein tca_01541 tca_01541 hypothetical protein tca_00058 tca_00058 Undecaprenyl-phosphate mannosyltransferase tca_01375 tca_01375 Non-canonical purine NTP pyrophosphatase tca_00359 tca_00359 hypothetical protein tca_00799 tca_00799 hypothetical protein tca_01043 ygcB CRISPR-associated endonuclease/helicase Cas3 tca_01789 tca_01789 hypothetical protein tca_01040 cas2 CRISPR-associated endoribonuclease Cas2 tca_00290 tca_00290 hypothetical protein tca_01611 cmk_3 Cytidylate kinase A1-II (10) tca_01045 tca_01045 CRISPR-assoc_Cas7/Cst2/DevR tca_01616 tca_01616 hypothetical protein tca_01644 tca_01644 hypothetical protein tca_00042 tca_00042 hypothetical protein tca_01832 yjjX Non-canonical purine NTP phosphatase tca_00825 obg_1 GTPase Obg tca_01149 tca_01149 hypothetical protein tca_01635 tca_01635 hypothetical protein tca_00254 tca_00254 hypothetical protein tca_00815 tca_00815 hypothetical protein tca_00237 tca_00237 hypothetical protein tca_00322 tca_00322 Oxalate-binding protein tca_00819 tca_00819 hypothetical protein tca_01833 tca_01833 hypothetical protein tca_00814 tca_00814 hypothetical protein tca_00817 pyrI Aspartate carbamoyltransferase regulatory chain tca_01028 tca_01028 hypothetical protein tca_01834 tca_01834 hypothetical protein tca_00823 tldD Metalloprotease TldD 0 tca_01319 tca_01319 hypothetical protein tca_00816 tca_00816 hypothetical protein tca_00110 guaB_1 Inosine-5'-monophosphate dehydrogenase tca_00809 tca_00809 hypothetical protein tca_00806 tca_00806 hypothetical protein tca_00993 tca_00993 Ferredoxin-2 A1-III (40) tca_00803 tca_00803 hypothetical protein tca_00798 moaA_2 Cyclic pyranopterin monophosphate synthase tca_00800 tca_00800 hypothetical protein tca_00808 eamA putative amino-acid metabolite efux pump tca_00646 cobN_5 Aerobic cobaltochelatase subunit CobN tca_00106 nudI Nucleoside triphosphatase NudI tca_00801 tca_00801 hypothetical protein tca_00802 tca_00802 hypothetical protein tca_00255 cyoE Protoheme IX farnesyltransferase tca_00810 hisA_2 1-(5-phosphoribosyl)-5-imidazole-4-carboxamide isomerase tca_00830 tca_00830 hypothetical protein tca_00807 truA tRNA pseudouridine synthase A tca_00804 wecC UDP-N-acetyl-D-mannosamine dehydrogenase tca_00805 wbpI UDP-2,3-diacetamido-2,3-dideoxy-D-glucuronate 2-epimerase tca_01047 tca_01047 hypothetical protein tca_01837 tca_01837 hypothetical protein tca_01086 mrtG Methyl-coenzyme M reductase subunit gamma tca_00428 tca_00428 hypothetical protein tca_00832 tca_00832 hypothetical protein tca_00099 tca_00099 hypothetical protein tca_00141 fprA_1 Nitric oxide reductase tca_01772 tca_01772 hypothetical protein tca_00820 tca_00820 putative GMC-type oxidoreductase tca_01274 pyrE_2 Orotate phosphoribosyltransferase tca_00822 tca_00822 Magnesium transporter MgtE tca_00498 tca_00498 hypothetical protein tca_00140 rd Rubredoxin tca_00821 tca_00821 Ferredoxin-2 tca_00529 tca_00529 hypothetical protein tca_00698 fsH_1 ATP-dependent zinc metalloprotease FtsH tca_00723 rbr1 Rubrerythrin-1 tca_00029 tca_00029 hypothetical protein tca_00714 ndhI_2 NAD(P)H-quinone oxidoreductase subunit I, chloroplastic tca_00813 argC N-acetyl-gamma-glutamyl-phosphate reductase tca_00790 pdtaS_9 putative sensor histidine kinase pdtaS tca_01836 tca_01836 hypothetical protein tca_00708 tca_00708 Putative nickel-responsive regulator tca_00944 tca_00944 hypothetical protein tca_01540 gpmI 2,3-bisphosphoglycerate-independent phosphoglycerate mutase tca_00828 thiDN Bifunctional thiamine biosynthesis protein TiDN tca_00818 tca_00818 hypothetical protein A1-IV (42) tca_00826 hspA HSP20 family protein tca_00092 tca_00092 hypothetical protein tca_00620 tca_00620 hypothetical protein tca_00920 atpC V-type ATP synthase subunit C tca_00827 glmS_2 Glutamine--fructose-6-phosphate aminotransferase tca_00921 atpE V-type ATP synthase subunit E tca_00091 tca_00091 hypothetical protein tca_00093 tca_00093 hypothetical protein tca_01441 tca_01441 hypothetical protein tca_01415 yafJ_2 Putative glutamine amidotransferase YafJ tca_01066 tca_01066 hypothetical protein tca_00142 fnA Bacterial non- ferritin tca_01279 tca_01279 hypothetical protein tca_01143 tca_01143 hypothetical protein tca_01771 tca_01771 hypothetical protein tca_01307 tca_01307 hypothetical protein tca_00660 prcB_1 Proteasome subunit beta tca_00502 tca_00502 hypothetical protein tca_01038 tca_01038 hypothetical protein tca_00289 mepA Multidrug export protein MepA tca_00378 tca_00378 hypothetical protein A1-V (1) tca_01013 rpoB_1 DNA-directed RNA polymerase subunit beta tca_00124 lutB Lactate utilization protein B tca_01569 tca_01569 Putative thiazole biosynthetic A1-VI (4) tca_00869 tca_00869 hypothetical protein –1.5 CC BrES OC

Figure 4: Heatmap and hierarchical clustering of the differentially expressed genes of the Methanothermobacter sp. strain EMTCatA1. The scale of the heatmaps was the average of the TMM-normalized count values transformed so that the mean was 0 and the standard deviation was 1 (z-score). The assigned clusters were indicated at the rightmost column, with the number of genes contained in each cluster shown in brackets. The red arrowheads indicate the stress-related genes. Archaea 9 of methanogenesis in methanogens. An exception was methanogenesis via direct electron uptake from the cathode, mrtG (tca_01086), which encoded an MRI gamma- a phenomenon that was previously reported for the Metha- subunit homolog. Although this gene demonstrated higher nobacterium-like archaeon strain IM1 [6] and various expression under the OC condition than under the other Methanosarcina species [58]. However, this might not be conditions (in cluster A1-III in Figure 4), the associated the case for the Methanothermobacter sp. strain EMTCatA1 transcript abundances were barely detectable and, there- as the pure cultures of the M. thermautotrophicus strain fore, unlikely to have biological significance (Table S8). ΔH, a close relative of the Methanothermobacter sp. strain EMTCatA1, demonstrated no catalytic ability on a cathode 3.4.2. Transcriptome Analysis of the Coriobacteriaceae sp. poised at a potential higher than −0.6 V versus an SHE Strain EMTCatB1. For the Coriobacteriaceae sp. strain EMT- [16]. Only 14 genes of the Methanothermobacter sp. CatB1, genes encoding the putative multiheme c-type cyto- strain EMTCatA1, including two CRISPR-associated genes, chromes (1d0125, 1c0363, 1c0406, 1d0898, and 1c0642), a namely, tca_01044 and tca_01045, had no apparent homolog NiFe hydrogenase (1c0061, 1c0060, and 1c0059), and a in the M. thermautotrophicus strain ΔH (Table S11). formate dehydrogenase (1c0408 and 1c0407 with the above- Although some of these genes might confer the capability mentioned 1c0406) were found to be highly expressed under for direct electron uptake in the methanogen, we presumed the CC conditions (in the top 10% abundant transcripts: that the Methanothermobacter sp. strain EMTCatA1 was a Tables S5 and S9). Both multiheme c-type cytochromes hydrogenotrophic methanogen highly similar to the M. and membrane-bound oxidoreductases have been proposed thermautotrophicus strain ΔH and was likely unable to to constitute an extracellular electron transfer conduit in catalyze electromethanogenesis by itself. Thermincola potens, an exoelectrogenic gram-positive The role of the Coriobacteriaceae sp. strain EMTCatB1 in bacterium [56, 57]. In addition, a gene cluster encoding the electromethanogenesis remained more speculative. Based on subunits of a V-type ATPase (1d0031-37, Table S5) was its gene expression profile, it is presumable that the Coriobac- highly transcribed. teriaceae sp. strain EMTCatB1 was capable of direct electron Among the transcriptomes of the Coriobacteriaceae sp. uptake from the cathode via multiheme c-type cytochromes strain EMTCatB1, 103 differentially expressed genes were (e.g., those encoded by 1d0898 and 1c0642). The electrons identified (Figure 5, Table S7). The general cluster were then likely conducted to the relevant membrane- organization of the differentially expressed genes was bound oxidoreductases, such as hydrogenase and formate similar to that in the Methanothermobacter sp. strain dehydrogenase. At the least, part of the electrical energy from EMTCatA1 (Figure 5, Table S10). The largest cluster (B1- this electron flow was possibly utilized to create a proton II), which consisted of 43 genes, was expressed at higher motive force to drive ATP synthesis via V-type ATPase. In levels under the CC condition, and conversely, the second- addition, NADH:ubiquinone oxidoreductase (complex I) largest cluster (B1-V), which consisted of 38 genes, was might be involved in the generation of a proton motive force expressed at higher levels under the BrES and OC as two genes encoding subunits of the enzyme (nuoCD: conditions. Cluster B1-II contained the abovementioned 1c0337 and 1c0336) were highly transcribed under the CC genes encoding two multiheme c-type cytochromes condition (Table S5). Therefore, cellular energy was likely (1d0898 and 1c0642), a hydrogenase component of formate depleted under the OC condition, which was consistent dehydrogenase (1c0407), and two subunits of V-type with the induction of stress-related genes. ATPase (1d0032 and 1d0037) (blue arrowheads in Figure 5) (Table S10). These results suggest that these genes play a 3.6. A Possible Model of the Electromethanogenesis Process on role in electron consumption at the biocathode surfaces. the Biocathode Surface. Based on our results and discussions, Furthermore, similar to the A1-IV cluster of the strain it is tempting to speculate that a significant proportion of the EMTCatA1, the B1-V cluster consisted of stress-related electrons from the cathode was channeled to proton reduc- genes encoding homologs of chaperones (1d0043, 1d0806, tion in the Coriobacteriaceae sp. strain EMTCatB1, resulting and 1d0807), antioxidant enzymes (1c0451, 1c0554, and in H2 evolution. The Methanothermobacter sp. strain EMT- 1c0665), and an alternative redox protein (1c0602) (red CatA1 then consumed the released H2 for hydrogenotrophic arrowheads in Figure 5) (Table S10), suggesting that the methanogenesis. In the Coriobacteriaceae sp. strain EMT- bacterium was under stress in both the BrES and OC CatB1, higher expression of stress-related genes was observed conditions. under the BrES condition, suggesting that the addition of BrES affected the physiology of the bacterium. This observa- 3.5. Possible Metabolic Functions of the Dominant Species on tion could be interpreted that the bacterium needed metha- the Biocathode Surface. Based on our present study results nogenesis for its metabolic activity on the cathode. In other and those from our previous study [16], we investigated the words, hydrogenotrophic methanogenesis by the methano- possible roles of the dominant species in the electro- gen served to reduce the H2 partial pressure. Therefore, it methanogenesis process. We concluded that the Metha- kept its metabolism thermodynamically favorable. Thus, the nothermobacter sp. strain EMTCatA1 was responsible for two dominant species might be metabolically interdependent methanogenesis at the biocathodes via the hydrogeno- at the biocathode surface, performing obligately mutualistic trophic methanogenesis pathway, which was operated in metabolism [59] that serves to catalyze electromethanogen- a manner similar to that under H2-limited conditions. It esis. Bacteria related to the family Coriobacteriaceae have is possible that this methanogen alone catalyzed electro- been detected as the dominant species in another biocathode 10 Archaea

CDS Gene name Annotated gene product Cluster z-score 1d0012 infA Translation initiation factor IF-1 {ECO:0000255 1.5 1d0187 atpD F0F1-type ATP synthase, beta subunit B1-I (2) 1d0937 tufB Elongation factor Tu {ECO:0000255 1d0208 hisJ ABC-type amino acid transport/signal transduction systems, peri plasmic component/domain 1c0483 spoVG Uncharacterized protein, involved in the regulation of septum location 1d0684 1d0684 hypothetical protein 1c0642 1c0642 Multiheme cytochrome c domain-containing protein 1d0563 1d0563 hypothetical protein 1d0898 1d0898 Multiheme cytochrome c domain-containing protein 1d0592 wecB UDP-N-acetylglucosamine 2-epimerase 1c0449 livK ABC-type branched-chain amino acid transport systems, periplasm ic component 1c0710 1c0710 Hypothetical 66.3 kDa protein in hag2 5'region 1c0031 1c0031 hypothetical protein 1d0923 tufB Elongation factor Tu {ECO:0000255 1d0053 1d0053 hypothetical protein 1d0814 mdlB ABC-type multidrug transport system, ATPase and permease components 1d0363 holA DNA polymerase III, delta subunit 1c0460 1c0460 hypothetical protein 1d0485 acpP Acyl carrier protein {ECO:0000255 1d0925 1d0925 hypothetical protein 1d0921 rpoE DNA-directed RNA polymerase specialized sigma subunit, sigma24 homolog 1d0074 1d0074 hypothetical protein 1d0009 secY Preprotein translocase secY subunit 1c0052 1c0052 hypothetical protein B1-II (43) 1c0018 1c0018 hypothetical protein 1c0028 rfaG Glycosyltransferase 1c0487 1c0487 Hypothetical protein yabE 1c0407 hybA Fe-S-cluster-containing hydrogenase components 1 1d0032 1d0032 Archaeal/vacuolar-type H+-ATPase subunit E 1c0533 1c0533 hypothetical protein 1d0037 ntpD Archaeal/vacuolar-type H+-ATPase subunit D 1d0115 ilvB Tiamine pyrophosphate-requiring enzymes 1d0936 fusA Elongation factor G {ECO:0000255 1d0017 rpoA Transcriptase subunit alpha {ECO:0000255 1d0196 lytB N-acetylmuramoyl-L-alanine amidase cwlB precursor 1d0048 guaB IMP dehydrogenase/GMP reductase 1d0934 rpsL 30S ribosomal protein S12 {ECO:0000255 1d0831 metH Methionine synthase I, cobalamin-binding domain 1d0391 1d0391 hypothetical protein 1c0578 1c0578 Predicted permeases 1d0285 1d0285 hypothetical protein 1c0077 1c0077 hypothetical protein 1d0741 hemN Coproporphyrinogen III oxidase and related Fe-S oxidoreductases 1c0671 modF ABC-type molybdenum transport system, ATPase component/photorepair protein PhrA 1c0323 thrA Homoserine dehydrogenase 1c0006 thrC Treonine synthase 1d0802 1d0802 hypothetical protein 1d0067 1d0067 hypothetical protein 1c0393 ccmB ABC-type transport system involved in cytochrome c biogenesis, permease component 1c0259 1c0259 hypothetical protein 1d0767 ccmF Cytochrome c biogenesis factor 1c0272 1c0272 hypothetical protein B1-III (11) 1d0089 1d0089 hypothetical protein 1c0118 1c0118 hypothetical protein 1d0827 1d0827 Predicted ATPase (AAA+ superfamily) 1c0119 1c0119 hypothetical protein 1c0491 1c0491 Predicted methyltransferases 1c0015 fer Ferredoxin 1c0664 ahpC Selenocysteine-containing peroxiredoxin PrxU 1d0453 1d0453 6-pyruvoyl-tetrahydropterin synthase 1c0085 livG ABC-type branched-chain amino acid transport systems, ATPase component B1-IV (9) 1c0419 1c0419 hypothetical protein 1d0109 rfaG Glycosyltransferase 1d0163 elaC Metal-dependent hydrolases of the beta-lactamase superfamily II I 1d0311 hemE Uroporphyrinogen decarboxylase {ECO:0000255 0 1d0318 1d0318 hypothetical protein 1c0602 napF Ferredoxin 1c0646 sufC ABC-type transport system involved in Fe-S cluster assembly, ATPase component 1c0673 pgi Phosphohexose isomerase {ECO:0000255 1c0516 1c0516 Uncharacterized conserved protein 1d0029 1d0029 FOG: CBS domain 1c0554 1c0554 Peptide methionine sulfoxide reductase MsrB 1c0108 gcd1 Nucleoside-diphosphate-sugar pyrophosphorylase involved in lipopolysaccharide biosynthesis 1d0569 mviN Uncharacterized membrane protein, putative virulence factor 1d0577 1d0577 hypothetical protein 1d0633 abrB Regulators of stationary/sporulation gene expression 1d0503 xerD Site-specific recombinase XerD 1d0239 pfA 6-phosphofructokinase 1c0451 grxC Glutaredoxin and related proteins 1c0644 1c0644 hypothetical protein 1d0709 rhaT Permeases of the drug/metabolite transporter (DMT) superfamily 1d0172 hemK Protein-glutamine N-methyltransferase PrmC {ECO:0000255 1d0724 1d0724 hypothetical protein 1c0514 nifS Cysteine desulfurase IscS {ECO:0000255 B1-V (38) 1c0129 1c0129 Transcriptase subunit omega {ECO:0000255 1d0739 1d0739 Predicted membrane protein 1d0806 dnaK Heat shock protein 70 {ECO:0000255 1c0083 paaK Coenzyme F390 synthetase 1c0590 rpsF 30S ribosomal protein S6 {ECO:0000255 1c0674 1c0674 hypothetical protein 1d0807 grpE HSP-70 cofactor {ECO:0000255 1c0045 proP Permeases of the major facilitator superfamily 1d0604 thiH Tiamine biosynthesis enzyme TiH and related uncharacterized e nzymes 1d0869 malK ABC-type sugar transport systems, ATPase components 1d0793 rimI Acetyltransferases 1d0730 1d0730 hypothetical protein 1c0515 1c0515 Predicted transcriptional regulator 1d0670 1d0670 hypothetical protein 1d0043 groS Protein Cpn10 {ECO:0000255 1d0790 pflA Pyruvate-formate lyase-activating enzyme 1c0109 pgsA Phosphatidylglycerophosphate synthase 1c0665 ahpC Selenocysteine-containing peroxiredoxin PrxU 1d0211 1d0211 Ribosome-associated protein Y (PSrp-1) –1.5 CC BrES OC

Figure 5: Heatmap and hierarchical clustering of the differentially expressed genes of the Coriobacteriaceae sp. strain EMTCatB1. The scale of the heatmaps was the average of the TMM-normalized count values transformed so that the mean was 0 and the standard deviation was 1 (z-score). The assigned clusters were indicated at the rightmost column, with the number of genes contained in each cluster shown in brackets. The blue arrowheads indicate the genes potentially involved in cathodic electron consumption. The red arrowheads indicate the stress-related genes.

[60], supporting the notion that Actinobacteria play a role in be due to the direct effect; i.e., BrES was toxic to or electromethanogenesis. metabolized by the bacterium, thereby altering its gene However, the above model was highly speculative and expression pattern. Moreover, as the genome does not fi required future investigations. A considerable limitation encode conserved enzymes for CO2 xation, it is unclear was the current lack of physiological knowledge regarding how the bacterium grows on the cathode. In this regard, it the dominant species. In particular, the Coriobacteriaceae is plausible that the Coriobacteriaceae sp. strain EMTCatB1 sp. strain EMTCatB1 had no close relatives among the utilized acetate as its carbon source, which was present in cultured species (Figure S4), and its metabolic properties the medium only during the initial development of the remained mostly unknown. The response of the biocathode (in the single-chamber reactor) and then Coriobacteriaceae sp. strain EMTCatB1 to BrES could well omitted from the medium (in the two-chamber reactor). In Archaea 11 other words, the bacterium might not be able to propagate Supplementary Materials on the biocathode (in the absence of acetate) but still Supplementary figures (characterization of the metagenome- metabolically active and able to produce H2. Therefore, future studies should perform isolation of sequenced biocathodes (Figure S1), genome maps of the two the dominant species, together with biochemical analyses dominant species (Figure S2), whole genome alignment of [56, 57, 61]. Such an approach would also be useful to exam- the Methanothermobacter sp. strain EMTCatA1 and M. ther- Δ ine the possible contribution of relatively minor species, mautotrophicus strain H (Figure S3), phylogenetic tree of which were excluded from this study, to elaborate the electro- the Coriobacteriaceae sp. strain EMTCatB1 and other methanogenesis process. Further transcriptome/proteome Actinobacteria (Figure S4), cyclic voltammograms of the bio- analyses using different conditions, such as using various cathodes of the electromethanogenic reactors C1~C6 (Figure electrical potential values, are also required to understand S5), representative scanning electron micrographs of the bio- the biocathode mechanism more comprehensively. cathode surfaces of the electromethanogenic reactors C1~C6 fi (Figure S6), current generation and CH4 production pro les of the biocathodes used for the transcriptome analysis 4. Conclusions (Figure S7), and taxonomic assignments of unmapped RNA-seq read (Figure S8)) and tables (details on probes The primary constituents of a novel thermophilic consor- and primers (Table S1), RNA sequencing and mapping tium enriched on an electromethanogenic biocathode were (Tables S2 and S3), TPM- (Tables S4 and S5) and TMM- characterized in the present study. The results indicated that (Tables S6 and S7) normalized read counts and differentially the metagenome of the consortium was mainly dominated by expressed gene clusters (Tables S8 and S10) of the transcrip- the Methanothermobacter sp. strain EMTCatA1 and Corio- tomes of the dominant species, c-type cytochromes of the bacteriaceae sp. strain EMTCatB1. The dominance of the Coriobacteriaceae sp. strain EMTCatB1 (Table S9), and genes two species on the biocathodes was further confirmed using of the Methanothermobacter sp. strain EMTCatA1 not pres- qPCR and FISH, leading us to analyze the transcriptomes ent in the M. thermautotrophicus strain ΔH (Table S11)) are in the biocathodes under different conditions (CC, OC, and available online. (Supplementary Materials) BrES). Based on the expression profile of the genes involved in hydrogenotrophic methanogenesis by the methanogen References and those encoding c-type cytochromes and membrane- bound enzymes of the bacterium, these strains were sug- [1] S. Cheng, D. Xing, D. F. Call, and B. E. Logan, “Direct biolog- gested to have functions in methane production and electron ical conversion of electrical current into methane by electro- uptake, respectively, in the electromethanogenesis process. methanogenesis,” Environmental Science & Technology, This study therefore represents the first multiomics charac- vol. 43, no. 10, pp. 3953–3958, 2009. terization of an electromethanogenic biocathode, providing [2] F. Geppert, D. Liu, M. van Eerten-Jansen, E. Weidner, complementary information for previous studies on various C. Buisman, and A. Ter Heijne, “Bioelectrochemical power- bioelectrochemical systems. to-gas: state of the art and future perspectives,” Trends in Bio- technology, vol. 34, no. 11, pp. 879–894, 2016. [3] K. Sato, H. Kawaguchi, and H. Kobayashi, “Bio-electrochemi- Data Availability cal conversion of carbon dioxide to methane in geological stor- age reservoirs,” Energy Conversion and Management, vol. 66, Raw reads of the metagenome, metatranscriptomes, and pp. 343–350, 2013. 16S rRNA gene amplicons were deposited in the Sequence [4] M. Siegert, M. D. Yates, A. M. Spormann, and B. E. Logan, Read Archive of the DNA Data Bank of Japan (DDBJ) “Methanobacterium dominates biocathodic archaeal commu- under the BioProjects PRJDB8994, PRJDB8993, and nities in methanogenic microbial electrolysis cells,” ACS PRJDB8998, respectively. Sustainable Chemistry & Engineering, vol. 3, no. 7, pp. 1668– 1676, 2015. Conflicts of Interest [5] M. C. A. A. Van Eerten-Jansen, A. B. Veldhoen, C. M. Plugge, A. J. M. Stams, C. J. N. Buisman, and A. Ter Heijne, “Microbial The authors have no direct financial relationship with the community analysis of a methane-producing biocathode in a commercial identities mentioned in the paper that might lead bioelectrochemical system,” Archaea, vol. 2013, Article ID to a conflict of interest. 481784, 12 pages, 2013. [6] P. F. Beese-Vasbender, J. P. Grote, J. Garrelfs, M. Stratmann, and K. J. Mayrhofer, “Selective microbial electrosynthesis of Acknowledgments methane by a pure culture of a marine lithoautotrophic archaeon,” Bioelectrochemistry, vol. 102, pp. 50–55, 2015. The authors would like to thank Dr. Hiroshi Sagara (Institute [7] J. S. Deutzmann, M. Sahin, and A. M. Spormann, “Extracellu- of Medical Science, the University of Tokyo) for the SEM lar enzymes facilitate electron uptake in biocorrosion and analysis. The authors would like to thank Enago (http:// bioelectrosynthesis,” MBio, vol. 6, no. 2, 2015. www.enago.jp) for the English language review. This study [8] S. T. Lohner, J. S. Deutzmann, B. E. Logan, J. Leigh, and A. M. was supported by INPEX Corporation and the Japan Society Spormann, “Hydrogenase-independent uptake and metabo- for the Promotion of Science (JSPS) Grant-in-Aid for Scien- lism of electrons by the archaeon Methanococcus maripalu- tific Research (C) 17K07713 (to Hajime Kobayashi). dis,” The ISME Journal, vol. 8, no. 8, pp. 1673–1681, 2014. 12 Archaea

[9] M. Lienemann, J. S. Deutzmann, R. D. Milton, M. Sahin, and [24] M. Hunt, T. Kikuchi, M. Sanders, C. Newbold, M. Berriman, A. M. Spormann, “Mediator-free enzymatic electrosynthesis and T. D. Otto, “REAPR: a universal tool for genome assembly of formate by the Methanococcus maripaludis heterodisulfide evaluation,” Genome Biology, vol. 14, article R47, 2013. reductase supercomplex,” Bioresource Technology, vol. 254, [25] T. Seemann, “Prokka: rapid prokaryotic genome annotation,” pp. 278–283, 2018. Bioinformatics, vol. 30, no. 14, pp. 2068-2069, 2014. [10] P. F. Beese-Vasbender, S. Nayak, A. Erbe, M. Stratmann, and [26] M. Kanehisa and Y. Sato, “KEGG Mapper for inferring cellular K. J. Mayrhofer, “Electrochemical characterization of direct functions from protein sequences,” Protein Science, vol. 29, electron uptake in electrical microbially influenced corrosion no. 1, pp. 28–35, 2019. of iron by the lithoautotrophic SRB Desulfopila corrodens [27] A. C. Darling, B. Mau, F. R. Blattner, and N. T. Perna, “Mauve: ” – strain IS4, Elcetrochimica Acta, vol. 167, pp. 321 329, 2015. multiple alignment of conserved genomic sequence with rear- [11] J. S. Deutzmann and A. M. Spormann, “Enhanced microbial rangements,” Genome Research, vol. 14, no. 7, pp. 1394–1403, electrosynthesis by using defined co-cultures,” The ISME Jour- 2004. – nal, vol. 11, no. 3, pp. 704 714, 2017. [28] Y. Yu, C. Lee, J. Kim, and S. Hwang, “Group-specific [12] Z. Wang, D. H. Leary, A. P. Malanoski et al., “A previously primer and probe sets to detect methanogenic communities uncharacterized, nonphotosynthetic member of the Chroma- using quantitative real-time polymerase chain reaction,” fi tiaceae is the primary CO2- xing constituent in a self- Biotechnology and Bioengineering, vol. 89, no. 6, pp. 670– regenerating biocathode,” Applied and Environmental Micro- 679, 2005. – biology, vol. 81, no. 2, pp. 699 712, 2015. [29] W. Ludwig, O. Strunk, R. Westram et al., “ARB: a software [13] B. J. Eddie, Z. Wang, W. J. T. Hervey et al., “Metatranscrip- environment for sequence data,” Nucleic Acids Research, tomics supports the mechanism for biocathode electroautotro- vol. 32, no. 4, pp. 1363–1371, 2004. “ ” ” phy by Candidatus Tenderia electrophaga , mSystems, vol. 2, [30] G. Crocetti, M. Murto, and L. Bjornsson, “An update and opti- no. 2, 2017. misation of oligonucleotide probes targeting methanogenic [14] D. H. Leary, W. J. T. Hervey, A. P. Malanoski et al., “Metapro- Archaea for use in fluorescence in situ hybridisation (FISH),” teomic evidence of changes in protein expression following a Journal of Microbiological Methods, vol. 65, no. 1, pp. 194– change in electrode potential in a robust biocathode micro- 201, 2006. ” – biome, Proteomics, vol. 15, no. 20, pp. 3486 3496, 2015. [31] K. Nakamura, T. Terada, Y. Sekiguchi et al., “Application of [15] C. W. Marshall, D. E. Ross, K. M. Handley et al., “Metabolic pseudomurein endoisopeptidase to fluorescence in situ reconstruction and modeling microbial electrosynthesis,” Sci- hybridization of methanogens within the family Methanobac- entific Reports, vol. 7, no. 1, article 8391, 2017. teriaceae,” Applied and Environmental Microbiology, vol. 72, [16] Q. Fu, Y. Kuramochi, N. Fukushima, H. Maeda, K. Sato, and no. 11, pp. 6907–6913, 2006. H. Kobayashi, “Bioelectrochemical analyses of the develop- [32] C. Quast, E. Pruesse, P. Yilmaz et al., “The SILVA ribosomal ment of a thermophilic biocathode catalyzing electromethano- RNA gene database project: improved data processing and genesis,” Environmental Science & Technology, vol. 49, no. 2, web-based tools,” Nucleic Acids Research, vol. 41, pp. D590– pp. 1225–1232, 2015. D596, 2013. [17] H. Kobayashi, Q. Fu, H. Maeda, and K. Sato, “Draft genome [33] L. S. Yilmaz, S. Parnerkar, and D. R. Noguera, “mathFISH, a sequence of a novel Coriobacteriaceae sp. strain, EMTCatB1, web tool that uses thermodynamics-based mathematical reconstructed from the metagenome of a thermophilic electro- models for in silico evaluation of oligonucleotide probes for methanogenic biocathode,” Genome Announcements, vol. 5, fluorescence in situ hybridization,” Applied and Environmen- no. 10, 2017. tal Microbiology, vol. 77, no. 3, pp. 1118–1122, 2011. [18] H. Kobayashi, X. Sun, Q. Fu, H. Maeda, and K. Sato, [34] R. Sekar, A. Pernthaler, J. Pernthaler, F. Warnecke, T. Posch, “Draft genome sequence of Methanothermobacter sp. strain and R. Amann, “An improved protocol for quantification of EMTCatA1, reconstructed from the metagenome of a thermo- freshwater Actinobacteria by fluorescence in situ hybridiza- philic electromethanogenesis-catalyzing biocathode,” Genome tion,” Applied and Environmental Microbiology, vol. 69, Announcements, vol. 5, no. 35, 2017. no. 5, pp. 2928–2935, 2003. [19] M. Martin, “Cutadapt removes adapter sequences from high- [35] J. Schindelin, I. Arganda-Carreras, E. Frise et al., “Fiji: an open- throughput sequencing reads,” EMBnet Journal, vol. 17, source platform for biological-image analysis,” Nature no. 1, pp. 10–12, 2011. Methods, vol. 9, no. 7, pp. 676–682, 2012. [20] H. Kobayashi, A. Nagashima, M. Kouyama et al., “High-pres- [36] A. M. Bolger, M. Lohse, and B. Usadel, “Trimmomatic: a flex- sure thermophilic electromethanogenic system producing ible trimmer for Illumina sequence data,” Bioinformatics, methane at 5 MPa, 55°C,” Journal of Bioscience and Bioengi- vol. 30, no. 15, pp. 2114–2120, 2014. neering, vol. 124, no. 3, pp. 327–332, 2017. [37] H. Li and R. Durbin, “Fast and accurate short read alignment [21] D. T. Truong, E. A. Franzosa, T. L. Tickle et al., “MetaPhlAn2 with Burrows-Wheeler transform,” Bioinformatics, vol. 25, for enhanced metagenomic taxonomic profiling,” Nature no. 14, pp. 1754–1760, 2009. Methods, vol. 12, no. 10, pp. 902-903, 2015. [38] H. Li, B. Handsaker, A. Wysoker et al., “The sequence align- [22] D. R. Zerbino and E. Birney, “Velvet: algorithms for de novo ment/map format and SAMtools,” Bioinformatics, vol. 25, short read assembly using de Bruijn graphs,” Genome no. 16, pp. 2078-2079, 2009. Research, vol. 18, no. 5, pp. 821–829, 2008. [39] P. Menzel, K. Ng, and A. Krogh, “Fast and sensitive taxonomic [23] D. Paulino, R. L. Warren, B. P. Vandervalk, A. Raymond, S. D. classification for metagenomics with Kaiju,” Nature Commu- Jackman, and I. Birol, “Sealer: a scalable gap-closing applica- nications, vol. 7, no. 1, article 11257, 2016. tion for finishing draft genomes,” BMC Bioinformatics, [40] M. Pertea, G. M. Pertea, C. M. Antonescu, T. C. Chang, J. T. vol. 16, article 230, 2015. Mendell, and S. L. Salzberg, “StringTie enables improved Archaea 13

reconstruction of a transcriptome from RNA-seq reads,” growth,” The ISME Journal, vol. 6, no. 11, pp. 2045–2055, Nature Biotechnology, vol. 33, no. 3, pp. 290–295, 2015. 2012. [41] M. D. Robinson, D. J. McCarthy, and G. K. Smyth, “edgeR: a [54] Q. Xia, T. Wang, E. L. Hendrickson, T. J. Lie, M. Hackett, and Bioconductor package for differential expression analysis of J. A. Leigh, “Quantitative proteomics of nutrient limitation in digital gene expression data,” Bioinformatics, vol. 26, no. 1, the hydrogenotrophic methanogen Methanococcus maripalu- pp. 139-140, 2010. dis,” BMC Microbiology, vol. 9, no. 1, p. 149, 2009. [42] F. O. Glockner, E. Zaichikov, N. Belkova et al., “Comparative [55] X. Zhou, M. J. A. Stevens, S. Neuenschwander et al., “The tran- 16S rRNA analysis of lake bacterioplankton reveals globally scriptome response of the ruminal methanogen Methanobrevi- distributed phylogenetic clusters including an abundant group bacter ruminantium strain M1 to the inhibitor lauric acid,” of actinobacteria,” Applied and Environmental Microbiology, BMC Research Notes, vol. 11, no. 1, p. 135, 2018. – vol. 66, no. 11, pp. 5053 5065, 2000. [56] H. K. Carlson, A. T. Iavarone, A. Gorur et al., “Surface multi- [43] F. Warnecke, R. Sommaruga, R. Sekar, J. S. Hofer, and -type cytochromes from Thermincola potens and impli- “ J. Pernthaler, Abundances, identity, and growth state of acti- cations for respiratory metal reduction by gram-positive ff ” nobacteria in mountain lakes of di erent UV transparency, bacteria,” Proceedings of National Academy of Sciences of the Applied and Environmental Microbiology, vol. 71, no. 9, United States of America, vol. 109, no. 5, pp. 1702–1707, 2012. pp. 5551–5559, 2005. [57] N. L. Costa, B. Hermann, V. Fourmond et al., “How thermo- “ [44] J. G. Zeikus and R. S. Wolfe, Fine structure of Methanobacter- philic Gram-positive organisms perform extracellular electron ff ium thermoautotrophicum: e ect of growth temperature on transfer: characterization of the cell surface terminal reductase ” morphology and ultrastructure, Journal of Bacteriology, OcwA,” MBio, vol. 10, no. 4, 2019. vol. 113, no. 1, pp. 461–467, 1973. [58] M. O. Yee, O. L. O. Snoeyenbos-West, B. Thamdrup, L. D. M. [45] A. I. Majernik, M. Lundgren, P. McDermott, R. Bernander, Ottosen, and A.-E. Rotaru, “Extracellular electron uptake by and J. P. Chong, “DNA content and nucleoid distribution in ” ” two Methanosarcina species, Frontiers in Energy Research, Methanothermobacter thermautotrophicus, Journal of Bacte- vol. 7, p. 29, 2019. riology, vol. 187, no. 5, pp. 1856–1858, 2005. [59] B. E. Morris, R. Henneberger, H. Huber, and C. Moissl-Eichin- [46] S. Kato, T. Kosaka, and K. Watanabe, “Comparative ger, “Microbial syntrophy: interaction for the common good,” transcriptome analysis of responses of Methanothermobacter FEMS Microbiology Reviews, vol. 37, no. 3, pp. 384–406, 2013. thermautotrophicus to different environmental stimuli,” Envi- “ ronmental Microbiology, vol. 10, no. 4, pp. 893–905, 2008. [60] Q. Fu, S. Xiao, Z. Li et al., Hybrid solar-to-methane conver- sion system with a Faradaic efficiency of up to 96%,” Nano [47] M. Enoki, N. Shinzato, H. Sato, K. Nakamura, and – Y. Kamagata, “Comparative proteomic analysis of Metha- Energy, vol. 53, pp. 232 239, 2018. nothermobacter themautotrophicus ΔH in pure culture and in [61] S. Ardissone, H. Kobayashi, K. Kambara et al., “Role of BacA in co-culture with a butyrate-oxidizing bacterium,” PLoS One, lipopolysaccharide synthesis, peptide transport and nodula- vol. 6, no. 8, article e24309, 2011. tion by Rhizobium sp. NGR234,” Journal of Bacteriology, – [48] H. W. Luo, H. Zhang, T. Suzuki, S. Hattori, and Y. Kamagata, vol. 93, no. 9, pp. 2218 2228, 2011. “Differential expression of methanogenesis genes of Metha- nothermobacter thermoautotrophicus (formerly Methanobac- terium thermoautotrophicum) in pure culture and in co- cultures with fatty acid-oxidizing syntrophs,” Applied and Environmental Microbiology, vol. 68, no. 3, pp. 1173–1179, 2002. [49] R. M. Morgan, T. D. Pihl, J. Nolling, and J. N. Reeve, “Hydro- gen regulation of growth, growth yields, and methane gene transcription in Methanobacterium thermoautotrophicum del- taH,” Journal of Bacteriology, vol. 179, no. 3, pp. 889–898, 1997. [50] E. L. Hendrickson, A. K. Haydock, B. C. Moore, W. B. Whitman, and J. A. Leigh, “Functionally distinct genes regu- lated by hydrogen limitation and growth rate in methanogenic Archaea,” Proceedings of National Academy of Sciences of the United States of America, vol. 104, no. 21, pp. 8930–8934, 2007. [51] E. L. Hendrickson and J. A. Leigh, “Roles of coenzyme F420-

reducing hydrogenases and hydrogen- and F420-dependent methylenetetrahydromethanopterin dehydrogenases in reduc- tion of F420 and production of hydrogen during methanogen- esis,” Journal of Bacteriology, vol. 190, no. 14, pp. 4818–4821, 2008. [52] E. L. Hendrickson, Y. Liu, G. Rosas-Sandoval et al., “Global responses of Methanococcus maripaludis to specific nutrient limitations and growth rate,” Journal of Bacteriology, vol. 190, no. 6, pp. 2198–2205, 2008. [53] C. B. Walker, A. M. Redding-Johanson, E. E. Baidoo et al., “Functional responses of methanogenic archaea to syntrophic