The Genome Characteristics and Predicted Function of Methyl-Group Oxidation Pathway in the Obligate Aceticlastic Methanogens, Methanosaeta spp

In this work, we report the complete genome sequence of an obligate aceticlastic methanogen, Methanosaeta harundinacea 6Ac. Genome comparison indicated that the three cultured Methanosaeta spp., M. thermophila, M. concilii and M. harundinacea 6Ac, each carry an entire suite of genes encoding the proteins involved in the methyl-group oxidation pathway, a pathway whose function is not well documented in the obligately aceticlastic methanogens. Phylogenetic analysis showed that the methyl-group oxidation-involving proteins, Fwd, Mtd, Mch, and Mer from Methanosaeta strains cluster with the methylotrophic methanogens, and were not closely related to those from the hydrogenotrophic methanogens. Quantitative PCR detected the expression of all genes for this pathway, albeit ten times lower than the genes for aceticlastic methanogenesis in strain 6Ac. Western blots also revealed the expression of fwd and mch, genes involved in methyl-group oxidation. Moreover, 13C-labeling experiments suggested that the Methanosaeta strains might use the pathway as a methyl oxidation shunt during the aceticlastic metabolism. Because the mch mutants of Methanosarcina barkeri or M. acetivorans failed to grow on acetate, we suggest that Methanosaeta may use methyl-group oxidation pathway to generate reducing equivalents, possibly for biomass synthesis. An fpo operon, which encodes an electron transport complex for the reduction of CoM-CoB heterodisulfide, was found in the three genomes of the Methanosaeta strains. However, an incomplete protein complex lacking the FpoF subunit was predicted, as the gene for this protein was absent. Thus, F420H2 was predicted not to serve as the electron donor. In addition, two gene clusters encoding the two types of heterodisulfide reductase (Hdr), hdrABC, and hdrED, respectively, were found in the three Methanosaeta genomes. Quantitative PCR determined that the expression of hdrED was about ten times higher than hdrABC, suggesting that hdrED plays a major role in aceticlastic methanogenesis.


Introduction
Methanogenic Archaea are the only organisms known to produce abundant CH 4 for energy metabolism. Therefore, they exert a significant ecological impact on global carbon cycling. Cultured methanogens are categorized into four metabolic types based on methanogenic precursors, including hydrogenotrophic, methylotrophic, aceticlastic and methanol plus H 2 methanogenesis [1]. Since an estimated two-thirds of the methane in nature is from acetate [2], aceticlastic methanogenesis makes a major contribution to global methane production. So far, the methanogens utilizing acetate for methanogenesis are confined to the order Methanosarcinales. The genus Methanosarcina (Msr.) consists of the most metabolically diverse methanogenic species, most of them conduct three types of methanogenic metabolism [3]. In general, Methanosarcina strains have large genomes, e.g. 5.8 Mb for Msr. acetivorans, 4.8 Mb for Msr. barkeri and 4.1 Mb for Msr. mazei [4,5,6], and are about two times larger than other sequenced methanogen genomes.
In contrast, species of the other genus of aceticlastic methanogens Methanosaeta are obligately aceticlastic. As specialists, Methanosaeta strains possess a higher affinity for acetate than Methanosarcina stains and are favored in environments with low concentrations of acetate [7]. Hence, they are deemed the principal players in aceticlastic methanogenesis in nature.
Thus far, only three Methanosaeta species have been cultured. M. harundinacea is a mesophilic species isolated from an upflow anaerobic sludge blanket reactor treating beer-manufacturing wastewater in Beijing [8]. To gain insight into the genetic background of the aceticlastic methanogens, the complete genome of M. harundinacea 6Ac was sequenced and compared to that of the two other species. In addition to the other common genome characteristics among the three species, unexpectedly the methyl-group oxidation pathway was present in all three genomes. In methylotrophic methanogens, this pathway provides reducing equivalents for methanogenesis [9]. It also plays a role in the anaerobic oxidation of methane [10,11]. To understand the possible function of the pathway, in this study, the expression of the genes in the pathway was tested in M. harundinacea 6Ac. The gene expression and the preliminary physiological study suggest that the methyl-group oxidation pathway can be used to generate the extra reducing equivalents in the obligate aceticlastic methanogens.

Results and Discussion
General genome features of Methanosaeta species In this study, the complete genome of Methanosaeta harundinacea 6Ac, which contains a single circular chromosome and a circular plasmid designated as pH 6AC, was obtained. The chromosome is 2,559,043 bp in length with an average G+C content of 60.6% ( Figure 1). It contains one rRNA operon (5S, 16S, and 23S), a distinct 7S rRNA gene, and 39 tRNA genes (Table 1)  (3) genes included in the genome of M. harundinacea but not in the genomes of M. concilii and M. thermophila; (4) genes involved in methane metabolism; (5) genes found in all the three Methanosaeta genomes; (6) core genes found in the methanogen genomes sequenced; (7) RNA genes, including tRNA (red) and rRNAs (yellow); (8) genes on the plus and minus strand, respectively. All the genes were colored by functional categories according to COG classification. doi:10.1371/journal.pone.0036756.g001 observed in terms of the genome sizes, resulting in the difference in gene content (Table 1), with M. thermophila PT having the smallest genome, and M. concilii having more genes in category of DNA replication, recombination and repair, which is contributed by the multiple copies of the genes in this category. The coding regions in the three genomes of the Methanosaeta strains (.80%) account for the higher percentage of total genes compared to those of the Methanosarcina strains (70-75%).
Comparison of the gene function categories among the three species (Table 2) showed that while the two mesophilic strains contain similar gene numbers in each functional category, the thermophilic one has remarkably reduced gene numbers in cell motility and secretion, defense systems, and post-translational modification (PTM). This is probably consistent with its hot niche, where the biodiversity is relatively lower.
A suite of genes for the methyl-group oxidation pathway present in the Methanosaeta genomes It has been long determined that Methanosaeta strains are obligately aceticlastic methanogens, acetate being the exclusive substrate for CH 4 formation and energy biosynthesis. However, an entire suite of genes for methyl-group oxidation pathway was found in the three genomes. Two types of formylmethanofuran dehydrogenase, tungsten formylmethanofuran dehydrogenase (Fwd) and the molybdenum isoenzyme (Fmd), were found. As shown in Figure 2, M. thermophila possesses only the tungsten formylmethanofuran dehydrogenase genes, implying that the tungsten formylmethanofuran dehydrogenase is essential for these methanogens. This assumption could be supported by the observation of Hochheimer et al. [12] that the tungsten enzymes are constitutively expressed, while the molybdenum ones show a molybdate-induced expression in two thermophilic methanogens Methanothermobacter wolfeii and Methanothermobacter thermoautotrophicus.
To gain an insight into the gene organization for methyl-group oxidation pathways in the Methanosaeta genome, we found that genes for the formylmethanofuran dehydrogenase complex were arranged differently in the three strains. While the three operons for the Fwd complex in M. harundinacea 6Ac are arrayed in a similar pattern to those of Methanosarcina spp., only two fmd/fwd operons were present in M. concilii and M. thermophila, with additional fwdE genes scattered on the chromosomes ( Figure 2).
Based on the protein phylogenetic analysis, we found that genes encoding Fwd, Mtd, Mch, and Mer, from the three Methanoseata strains were closely related with those from the methylotrophic methanogens, which branched off from those of the hydrogenotrophic methanogens ( Figure 3A). However, other proteins that functioned in different methanogenic pathways did not form similar clades ( Figure 3B). The different phylogenetic clustering patterns for the proteins can be attributed to their varied roles, such as Fwd, Mtd, Mch, and Mer in the methylotrophic methanogens mainly act in methyl oxidation, while in the hydrogenotrophic methanogens they are involved in CO 2 reduction to methane formation.
Genes encoding the proteins involved in methyl-group oxidation pathway expressed but at relatively low levels To examine whether the genes for methyl-group oxidation pathway were expressed during growth, transcription of the genes in M. harundinacea 6Ac was detected by means of quantitative PCR. The assay showed that the transcript abundances were at an average level of about 14% equivalence to that of the N 5 -methyl-H 4 SPT: CoM methyltransferase (mtr) genes (Table 3), but were 1.58 times above the average levels of the total genes based on microarray data (unpublished data). Furthermore, Western blot also determined the expression of the C subunit of Fwd and Mch in the acetate-growing culture of M. harundinacea 6Ac ( Figure 4). Therefore, an active methyl-group oxidation pathway was present in strain 6Ac even during aceticlastic growth.

Methyl-group oxidation pathway could act as a methyl oxidation shunt of acetate in Methanosaeta strains
To reveal the possible function of the methyl-group oxidation pathway in the obligate aceticlastic methanogens, 13 CH 4 or NaH 13 CO 3 was added to cultures of strain 6Ac to determine if trace methane oxidation or production occurred. However, neither the 13 CO 2 in the 13 CH 4 supplemented culture nor the 13 CH 4 in the NaH 13 CO 3 supplemented culture was detected after 60 days of  incubation. These tests indicated that the methyl-group oxidation pathway in M. hurandiacea 6Ac was not used for trace methane oxidation or production. Cultures were then supplemented with 1-13 C-labeled acetate or 2-13 C-labeled acetate. Upon acetate depletion by incubation at 37uC for 60 days, as expected in the [1-13 C] acetate-culture only labeled 13 CO 2 but not 13 CH 4 was found. Surprisingly, in the [2-13 C] acetate-culture, 13 CH 4 and 13 CO 2 were both detected. The generated 13 CO 2 accounted for about 1% of the methyl carbon of acetate (Table 4). This suggested that 1% methyl carbon of acetate was oxidized to CO 2 through methyl-group oxidation pathway in the obligate aceticlastic Methanosaeta strains. This oxidative carbon flux shunt, like in methylotrophic methanogenesis, was probably used by Methanosaeta strains to generate essential reducing equivalents such as reduced ferredoxin or reduced F 420 for other biological processes, e.g. cell biomass synthesis. This hypothesis is supported by the fact that Msr. barkeri and Msr. acetivorans fail to grow on acetate upon inactivation of mch, which is required for methyl group oxidation [13].

Characteristics of the genes involved in aceticlastic methanogenesis pathways in Methanosaeta strains
Acetate in undissociated form (pKa = 4.75, 25uC) is believed to diffuse freely across the cytoplasmic membrane of bacteria [14]. However, under neutral pH, in which Methanosaeta strains grow optimally, acetate is present in a dissociated state, thus, a transport protein is necessary. A putative acetate transporter, the Ady2 gene (Mhar_0433), was present in the three Methanosaeta genomes and has been identified as an active acetate transporter in Saccharomyces cerevisiae based on microarray analyses of an ady2D strain [15].
It is believed that the first reaction, i.e. acetate activation to acetyl-CoA, is a rate-limiting reaction for aceticlastic methanogenesis [16,17]. Acetate activation is accomplished through different reactions in the two types of aceticlastic methanogens, i.e. Methanosaeta strains use AMP-forming acetyl-CoA synthetase (ACS), while Methanosarcina spp. employ the combined actions of acetate kinase (AK) and phosphotransacetylase (PTA). Surprisingly, three to five acs gene homologs were identified in the genomes of Methanosaeta spp., and most of them were organized in tandem ( Figure 5). Quantitative PCR assay showed that the three acs genes in M. harundinacea 6Ac were expressed differentially. The abundance of acs2 (Mhar_0751) and acs3 (Mhar_0752) transcripts were about nine fold higher than those of acs1 (Mhar_0749) ( Table 3). This suggests that acs2 and acs3 might be the key proteins for aceticlastic methanogenesis, while acs1 may be more highly expressed under growth conditions not examined here. As biochemically characterized, following the acetate activation step, Methanosarcina and Methanosaeta employ the same enzymes in the remaining reactions in aceticlastic methanogens. The corresponding genes encoding the proteins involved in those reactions are all present in the three Methanosaeta genomes; these include the genes for the carbon monoxide dehydrogenase/acetyl-CoA decarbonylase complex (CODH/ACDS), methyltetrahydrosarcinapterin: CoM methyltransferase (Mtr), methyl-CoM methylreductase (Mcr), and heterodisulfide reductase (Hdr).

Differential expression of two operons encoding two types of heterodisulfide reductase
As in the Methanosarcinales genomes, two gene classes of coenzyme B-coenzyme M heterodisulfide reductase, hdrABC and hdrED, are present in the genomes of Methanosaeta strains. Quantitative PCR assay showed that the transcript abundances of hdrED are about ten times higher than those of hdrABC in strain 6AC (Table 3), indicating that hdrED is the primary Hdr in methanogenesis in M. harundinacea 6AC. According to that HdrA1B1C1 is used specifically in the methylotrophic methanogenesis pathway in Methanosarcina acetivorans [18], we hypothesize that hdrED and hdrABC obtain reducing equivalents from CODH and the methyl-group oxidation in Methanosaeta strains, respectively; and those obtain from methylgroup oxidation by hdrABC can be used to compensate the biosynthesis-consumed reducing equivalents.

An fpo operon for a truncate electron transport complex possibly functions in the reduction of CoM-CoB heterodisulfide
In Methanosarcina species, a heterodisulfide reductase is involved in releasing coenzyme M and coenzyme B from the heterodisulfide using reduced ferredoxin as electron donor as well as energy conservation in acetate metabolism [19]. To implement this biochemical reaction, Msr. mazei and Msr. barkeri employ the Ech complex and F 420 non-reducing hydrogenase, while Msr. acetivorans uses the Rnf-like complex [20]. However, neither the genes for Ech nor the Rnf-like complex were found in the three Methanosaeta genomes; instead, a gene cluster comprised of 11 genes (Mhar_1410-Mhar_1420) for a F 420 H 2 dehydrogenase (Fpo) complex was present in the three genomes.
The F 420 H 2 dehydrogenase complex (Fpo) functions specifically in methylotrophic methanogenesis in Methanosarcina spp. and the obligate methylotrophic methanogens, like Methanohalophilus mahii [21]. However, no gene for the FpoF subunit was found in the three genomes. This cytoplasmic protein accepts reducing equivalents from F 420 H 2 . Welte and Deppenmeier (2011) determined that M. thermophila uses reduced ferredoxin to reduce coenzyme M-coenzyme B heterodisulfide and proposed that M. thermophila uses the Fd: heterodisulfide oxidoreductase encoded by an Fpo operon as the energy-conserving system [22]. The activity of Fd: heterodisulfide oxidoreductase was also detected in the membrane fraction of an Msr. mazei Dech mutant [23], implying that other protein complexes, like the Fpo complex, may complete the activity in place of Ech.

DNA extraction
Methanosaeta harundinacea 6Ac T ( = JCM 13211, CGMCC 1.5026, and DSM 17206) was isolated from a UASB reactor in our lab, the strain information and growth conditions were described previously [7]. Cells were harvested by centrifugation at 15,0006g for 15 min at 4uC from 1000 ml culture. The cell pellets were frozen at 280uC. Then the frozen cell pellets were placed into a sterile, pre-cooled mortar and put into liquid N 2 . After the liquid N 2 had evaporated, the cells were ground to a powder with a rod. Upon being ground 5 times, the genomic DNA was extracted using the TIANamp Bacteria DNA Kit (TIANGEN Biotech, Beijing, China). The genomic DNA was quantified on 0.8% agarose gel stained with ethidium bromide and spectrophotometrically assessed.

Genome sequencing, assembly, and gap closure
The genome sequence of Methanosaeta harundinacea 6Ac T was determined using the Roche GS 454 system [24]. 220,740 reads containing up to 52,294,588 bases (averaged read length as 236 bp), were obtained resulting in a 26-fold coverage of the genome. Assembly was performed using the GS de novo Assembler software (http://www.454.com/) producing 62 contigs ranging from 500 bp to 160,686 bp (the N50 contig size is 92,019 bp). The relationship of the contigs was determined by multiplex PCR [25]. Gaps were then closed by sequencing the PCR products using ABI 3730xl capillary sequencers. Phred, Phrap, and Consed software packages (http://www.phrap.org/phredphrapconsed.html) were used for the final assembly and editing. Low quality regions of the genome were resequenced.

Genome analysis and annotation
Putative CDSs were identified by GeneMark [26] and Glimmer [27]. Peptides shorter than 30 aa were eliminated. Sequences from the intergenic regions were compared to GenBank's nonredundant (nr) protein database [28] to detect pseudogenes and to identify genes missed by the Glimmer or GeneMark prediction. Insert sequences were first detected using the IS Finder database (http://www-is.biotoul.fr/is.html) with default parameters selected manually. Transfer RNA genes were predicted by tRNAScan-SE [29], while ribosomal DNAs (rDNAs) and other RNA genes were identified by comparing the genome sequence to the rRNA database [30] using the Infernal program [31]. Functional annotation of CDSs was performed through searching against the nr protein database using BLASTP [32]. The protein set was also searched against the COG (http://www.ncbi.nlm.nih.gov/  COG/) [33] and KEGG (Kyoto encyclopedia of genes and genomes; http://www.genome.jp/kegg/) [34] databases for further function assignment. The criteria used to assign function to a CDS were (1) a minimum cutoff of 40% identity and 60% coverage of the protein length and (2) at least two best hits among the COG, KEGG, or nr protein database. A search for gene families in the genome was performed by BLASTCLUST. Subcellular localization of the proteins was predicted by the PSORTb program (v2.0.1) [35]. The TatP 1.0 server (v2.0) [36] and TATFIND 1.2 program [37] were used to detect the potential substrates of the Tat secretion system.

Data availability
The sequence and annotation of the M. harundinacea chromosome and the plasmid were submitted to the GenBank database under accession numbers CP003117 and CP003118, respectively.

RNA isolation and qRT-PCR
M. harundinacea 6Ac was grown in conditions described previously [8]. The cells were harvested for RNA isolation during the logarithmic growth phase at an OD 600 nm of 0.35-0.45. Total RNA was extracted using TRIzolH Reagent (Invitrogen) according to the manufacturer's instructions, modifications of the method included grinding in liquid nitrogen before TRIzol reagent was added. RNA was further purified by following the RNA cleanup protocol of the RNeasy Mini Kit (Qiagen). Contaminating DNA was digested twice with 0.1-0.2 U?ml 21 DNase I (Promega), according to the manufacturer's instructions, including 0.4-0.8 U?ml 21 RNasin (Promega) in the reaction. RNA was purified for a second time using the RNeasy Mini Kit prior to cDNA synthesis. 210-350 ng RNA and 500 ng random hexamers (Promega) were incubated for 10 min at 70uC, and subsequently cooled on ice. Synthesis of cDNA was performed in 1 mM of each dNTP (Promega), 16 U?ml 21 M-MLV reverse transcriptase (RNase H Minus, Point Mutation, Promega), 1.6 U?ml 21 RNasin Inhibitor (Promega) and one fold concentrated first strand buffer (Promega) for 10 min at room temperature, followed by 50 min at 45uC. Real-time PCR oligonucleotide primers (Table S1) were designed using the software Beacon Designer 5.0 (Premier Biosoft, Palo Alto, USA) to obtain maximal amplification efficiency and sensitivity. The specificity of primers was verified by evaluating the qPCR melting curve and PCR products by DNA gel electrophoresis. For quantification of gene expression, a DNA fragment including the target gene was amplified by PCR and quantified using a UV800 spectrophotometer (Beckman). These fragments were then serially diluted 10-fold (10 23 -10 29 ) to be used for the standard curve, which was performed in triplicate.
Transcript quantification was performed using an ABI 7000 SDSH (Applied Biosystems TM ) with SYBR Green supermix (TAKARA, Dalian, China) using the following thermocycling program: 40 cycles of 95uC for 10 s and 60uC for 30 s. The copy number in each cDNA sample was calculated according to the calibration curve generated by the PCR products including the target gene.
Western blot for the proteins participating in methylgroup oxidation pathway 59-AATACATATGGTCAC-TAAATCA-GACACG-39 and 59-GAATTCATGTGTCACCGCCTG-39, respectively. The resulting construct was checked by sequencing (Biosune, Beijing, China). Then protein production was induced and purified as described previously [38]. Polyclonal antibodies against FwdC, McrD, and Mch were prepared at the Laboratory Animal Center, Institute of Genetics and Developmental Biology, Chinese Academy of Sciences, by injecting rats with the three purified recombinant proteins.
M. harundinacea 6Ac cells at different growth phases were harvested by centrifugation at 12,0006g and washed three times with PBS buffer. Cell suspensions were lysed by sonication and centrifuged at 21,0006g to obtain the cell fractions. Supernatant fractions were used as all protein of cell extract. The fractions were subjected to SDS-PAGE on a 12% gel and then transferred to nitrocellulose membranes (Amersham, Little Chalfont, UK). Western blot and immunodetection were performed as described previously [39]. The non-specific binding of antibodies was blocked by incubation with Tris-buffered saline and Tween-20 (TBST) with 5% skim milk. Membranes were then probed with a 1,000-fold dilution of the following polyclonal antibodies: rat anti-FwdC, rat anti-Mch, and rat anti-McrD. Bound antibodies were visualized using goat anti-rat IgG (Cwbio, China) conjugated to horseradish peroxidase (HRP), followed by enhanced chemiluminescence (Amersham) according to manufacturer's instructions Incorporation of 13 C-labeled acetate, methane, and bicarbonate M. harundinacea was cultured in basal medium omitting NaHCO 3 containing 50 mM acetate by following the procedure previously described [8]. The 13 C labeled acetate or 13 C labeled bicarbonate was added to the culture at a final concentration of 5% (w/w) of non-labeled substrate. Cells were cultured in 100-ml serum vials containing 50 ml of medium (pH 7.0 at 25uC) under an atmosphere of N 2 and incubated at 37uC. Methane and acetate were determined by gas chromatograph GC-14B (Shimadzu) and  2 was determined by gas chromatograph GC-14C (Shimadzu) [8,40]. The stable isotope composition was determined by Trace GC/IsoLink/Delta V Advantage GC/IRMS (Thermo Fisher Scientific, Bremen, Germany) [41,42]. Table S1 The primers used in this study.

Supporting Information
(DOC) Table S2 The accession numbers of the proteins included in Figure 3.