(C) PLOS One [1]. This unaltered content originally appeared in journals.plosone.org.
Licensed under Creative Commons Attribution (CC BY) license.
url:https://journals.plos.org/plosone/s/licenses-and-copyright

------------



Comparative genomics reveals electron transfer and syntrophic mechanisms differentiating methanotrophic and methanogenic archaea

['Grayson L. Chadwick', 'Division Of Geological', 'Planetary Sciences', 'California Institute Of Technology', 'Pasadena', 'California', 'United States Of America', 'Connor T. Skennerton', 'Rafael Laso-Pérez', 'Max-Planck Institute For Marine Microbiology']

Date: 2022-01 

Phylogenetic trees constructed from ANME genomes, sequences of closely related archaea, and a selection of sequences derived from clone libraries demonstrate the relationship between ANME and methanogens of the Halobacterota. ( A ) Phylogenetic tree built with 16S rRNA gene sequences, root leads to sequence from Sulfolobus solfataricus p2. ( B ) Phylogenomic tree built with concatenated marker set 4 (see S1 Table for list), root also to S. solfataricus p2. ( C ) Phylogenetic tree built with protein sequence of RpoB, root leads to sequences from “Ca. Methanomethyliales” and “Ca. Bathyarchaeota.” ( D ) Phylogenetic tree of McrA protein sequences. Note the divergence of proposed alkane oxidizing McrA genes in “Ca. Syntrophoarchaeum,” “Ca. Bathyarchaeota,” and “Ca. Argoarchaeum.” Branch support values of 100% are labeled with closed circles, >50% with open circles. Tree scales represent substitutions per site. Tree construction parameters are found in the Materials and methods section. For tree files and alignments, see S1 Data . Detailed tree figures are presented in S2 – S5 Figs . ANME, anaerobic methanotrophic; McrA, methyl-coenzyme M reductase subunit A; RpoB, RNA polymerase subunit beta; 16S rRNA, small subunit ribosomal RNA.

Phylogenetic reconstructions of 16S rRNA gene sequences, concatenated marker genes, DNA-directed RNA polymerase subunit beta (RpoB), and the methane-activating enzyme methyl-coenzyme M reductase subunit A (McrA) were performed to demonstrate the evolutionary relationship between ANME and other related archaea ( Fig 1 ). All marker gene sets show remarkably similar phylogenies, with ANME-1 forming the deepest branching ANME clade, while ANME-2 and ANME-3 grouped within the order Methanosarcinales. Importantly, in agreement with previous studies, ANME-3 reproducibly branches well within the family Methanosarcinaceae, in agreement with the AAI analysis described above. The McrA phylogeny deviates somewhat from this pattern, with ANME-1 falling further outside of the traditional methanogens, but grouping with some McrA from the H 2 -dependent methylotrophic methanogens as has been previously reported [ 12 , 32 , 33 ]. The McrA in all ANME clades are more similar to methanogens than the recently described divergent McrA homologs in various uncultured archaea that are thought to utilize longer chain alkane substrates [ 15 , 18 , 21 ]. With this general phylogenetic framework, we set out to characterize the major conserved features of the ANME energy metabolism.

The ANME-3 clade is the ANME group most recently diverged from known methanogens. They are closely related by their 16S rRNA genes (94% to 96% similarity) to Methanococcoides, Methanosalsum, and Methanolobus, with 65% average AAI, indicating that these organisms represent a novel genus within the family Methanosarcinaceae, for which we propose the name “Candidatus Methanovorans” ( Table 1 , S1 Fig ). The 2 genomes were both recovered from the Haakon Mosby Mud Volcano where this clade was originally described on the basis of 16S rRNA gene sequences [ 30 ]. ANME-3 forms consortia with bacteria from the Desulfobulbus group [ 31 ].

“Ca. Methanoperedens” sp. BLZ1 and nitroreducens are representatives of the family Methanoperedenaceae, formerly known as ANME-2d, within the order Methanosarcinales. These are the only known ANME that do not couple AOM in syntrophy with partner SRB, instead coupling the oxidation of methane with the reduction of nitrate, iron oxides, or manganese oxides in freshwater environments [ 20 , 26 , 27 ]. The marine sister group of “Ca. Methanoperedens”, GoM Arc I (also known as AAA), was recently described as an anaerobic ethane degrader and contains 2 genera “Candidatus Argoarchaeum” [ 15 , 21 ] and the thermophilic “Candidatus Ethanoperedens” [ 28 ]. These clades are not specifically considered in the present work as they do not appear to be marine methanotrophs and have been extensively discussed in a recent study [ 29 ].

Marine members of the ANME-2 are within the order Methanosarcinales and are comprised of subclades a, b, and c, first designated by 16S rRNA gene sequences [ 5 ]. The ANME-2a and ANME-2b together form a family-level division with 2 genus-level clades corresponding to ANME-2a and ANME-2b, recovered from 6 different locations including 2 methane cold seep sites, 2 submarine mud volcanoes, a hydrocarbon seep and shallow coastal sandy sediments ( Table 1 ). We propose the name “Candidatus Methanocomedens” for the dominant genera corresponding to ANME-2a with name propagating to the family level as Methanocomedenaceae. Recent coupled fluorescence and electron microscopy analyses have revealed distinct ultrastructural features of the ANME-2b [ 25 ], and in conjunction with our phylogenomic information, we propose the genus name “Candidatus Methanomarinus” for ANME-2b. The 6 “Ca. Methanocomedens” genomes have average amino acid identity (AAI) values, which range between 70% and 97% indicating that there are multiple distinct species, whereas the 3 “Ca. Methanomarinus” genomes have >98% AAI indicating that they represent strains of the same species ( S1 Fig ). The ANME-2c form a separate family (84% AAI among the 8 genomes analyzed in this study) representing 2 genus-level clades. For the dominant ANME-2c, we propose the genus name “Candidatus Methanogaster” with name propagating to the family level as Methanogasteraceae.

Members of the ANME-1 were originally described in 16S rRNA gene surveys of methane seep sediments [ 1 ], and representatives of the ANME-1 were among the first to be genomically characterized [ 6 ]. ANME-1 16S rRNA genes have since been identified in marine cold seep environments, diffusive margin sulfate–methane transition zones, deep sea hydrothermal vents, and select anoxic terrestrial ecosystems [ 23 ]. With our updated genomic dataset, there are now 19 representative genomes from ANME-1, recovered from 8 locations, including deep-sea methane cold seeps, hydrocarbon-impacted hydrothermal vents and cold seeps, a mud volcano and a hot, deep gold mine aquifer. Collectively, these genomes are highly diverse at the sequence level with the majority being at most 60% similar to each other, based on pairwise sequence similarity and nonhomologous gene content between genomes ( S1 Fig ). These genomes represent 6 genera as determined by analysis of their RED values. Recently proposed nomenclature based on a single ANME-1 genome from fosmid sequences placed ANME-1 within their own order Methanophagales [ 24 ], and this is consistent with GTDB release 89. The 19 ANME-1 genomes analyzed here represent a single family-level division within that order. We propose to conserve the family and genus-level designation implied by “Methanophagales” with the ANME-1 genome from early fosmids [ 7 ] belonging to the genus “Candidatus Methanophaga” within the family Methanophagaceae ( Table 1 ).

The draft ANME genomes sequenced and assembled here were between 46% and 98% complete (mean: 78%, median: 80%) as determined by the presence of a set of marker proteins common to Halobacterota. Many of the less complete genomes originated from sequencing of individual flow sorted aggregates ( Table 1 ). Most of the genomes did not contain duplicated marker genes, with the exception of the 3 genomes derived from sequenced fosmids, which encoded many duplicates ( Table 1 ). The fosmids representing these genomes have consistent nucleotide signatures and, in many cases, contain overlapping regions that suggests that they are derived from multiple strains rather than from a single clonal population. While genome incompleteness can impact the accurate reconstruction of ANME metabolism, we found reproducible trends in gene presence, absence, and synteny across the different ANME lineages, with multiple genomes for each ANME clade, often originating from different studies and habitats. Taxonomic assignments in Table 1 were made consistent with GTDB release 89 using analysis of relative evolutionary divergence (RED) [ 22 ].

We reconstructed 28 new ANME genomes from metagenomic datasets and fluorescence-activated cell sorting (FACS) of single aggregates [ 14 ], as well as from reanalyzed publicly available metagenomic data from the sequence read archive (SRA) and the MG-RAST analysis server (see Materials and methods ). These genomes were combined with 11 previously published marine ANME genomes recovered from diverse environments [ 7 , 9 , 11 , 15 – 17 ] to generate a set of 39 ANME genomes for comparative genomics, representing all of the currently described clades (ANME-1, ANME-2, and ANME-3) and subclades (e.g., ANME-2a, ANME-b, ANME-c, and multiple clades in ANME-1) frequently detected in marine sediments ( Table 1 ) [ 4 ]. In most analyses, we also include the recently described alkane-oxidizing “Candidatus Syntrophoarchaeum” [ 18 ], as well as the nitrate-reducing freshwater ANME relatives known as ANME-2d or “Candidatus Methanoperedens” [ 19 , 20 ] and their marine relatives known as “Candidatus Argoarchaeum” [ 15 , 21 ]. At least one genus from each ANME subclade was assigned a name and formal etymology can be found in Materials and methods .

Schematic representation of the 3 phases of ANME energy metabolism in our current model. In Phase 1, methane is oxidized to CO 2 through the reversal of the canonical seven step methanogenesis pathway. Energy is invested in this phase in the form of sodium ion translocation from the outer face of the cytoplasmic membrane to the inner face (yellow arrow). As C1 moieties are sequentially oxidized, 8 electrons are transferred to soluble electron carriers such as F 420 H 2 , NADPH, Fd 2− , and CoM-SH/CoB-SH. In Phase 2, 8 electrons on these primary electron carriers are transferred to secondary electron carriers in a process that conserves energy needed for cell growth in the form of sodium and proton motive forces (yellow arrows). These secondary electron carriers may be quinols (QH 2 ) methanophenazine (MpH 2 ) or possibly soluble electron carriers such as formate (HCOO−) or an unknown electron shuttle (XH 2 ). In Phase 3, the secondary electron carriers are relieved of their electrons in various ways depending on the environmentally available electron acceptors, which can include SRB in the case of marine ANME-SRB consortia, iron, manganese, or oxidized nitrogen species in the case of “Ca. Methanoperedens”. Humic substances and artificial electron acceptors (AQDS) have also served as electron acceptors in laboratory experiments for a variety of different ANME from fresh and marine environments. ANME, anaerobic methanotrophic; AOM, anaerobic oxidation of methane; SRB, sulfate-reducing bacteria.

It has been pointed out in recent reviews that a simple reversal of the methanogenesis pathway does not represent a viable basis for an energy metabolism in the ANME archaea, as the exact reversal of a process that results in net ATP generation must result in ATP loss [ 13 , 42 ]. Bioenergetic novelty beyond a wholesale reversal of methanogenesis therefore must exist. In the following discussion, we break down the ANME energy metabolism into 3 phases and discuss the conserved features of these phases across our collection of ANME genomes ( Fig 2 ). In the first phase, methane is oxidized to CO 2 and all 8 electrons are deposited on cytoplasmic electron carriers in an endergonic process that requires an investment of energy. In the second phase, cytoplasmic electron carriers are reoxidized in an exergonic process that reduces a set of intermediate electron carriers, recovering the energy invested in the first phase and conserving additional energy as an ion motive force. In the final phase, electrons must be discarded, likely not involving energy gain or loss. We use this division of energy metabolism as an organizing framework throughout this work.

Biochemically, it is assumed that the ANME archaea oxidize methane to CO 2 and pass electrons through an unconfirmed mechanism to their SRB partners to reduce sulfate. Metabolic reconstruction of a limited number of ANME draft genomes from environmental samples and enrichment cultures using multiple ‘omics approaches have shown that they contain and express the same genes used for methanogenesis [ 6 , 7 , 9 , 10 , 20 , 34 ]. Detailed enzymatic studies of the methanogenic pathway have shown that all 7 steps are reversible [ 35 – 37 ], including the methane producing step catalyzed by Mcr [ 38 ],which had previously been predicted to be irreversible. These findings reinforce the “reverse methanogenesis” hypothesis [ 6 , 39 – 41 ], in which ANME use the same methanogenic enzymes to oxidize, rather than produce, methane. This model offers the most likely pathway for carbon oxidation in ANME; however, the mechanisms by which these archaea conserve energy from this process and how methane-derived electrons are transferred to their syntrophic sulfate-reducing bacterial partners remain open questions.

The phylogeny of MetF proteins in ANME is best explained by 2 separate acquisitions of these genes. MetF in ANME-2 and ANME-3 were likely acquired in the common ancestor of these organisms and the other Methanosarcinaceae along with other enzymes to utilize H 4 F-bound C1 moieties for the purpose of biosynthesis. Based on its distribution in uncultured archaea of different phyla, as well as the paucity of other H 4 F-interacting proteins in ANME-1, there is good support for a different, possibly catabolic, function of MetF in ANME-1. Due to the structural similarity between H 4 F and H 4 MPT [ 53 ], it is possible that the ANME-1–type MetF has evolved to react with C1 moieties bound to H 4 MPT instead of H 4 F. A switch between H 4 F and H 4 MPT as a carbon carrier is not unprecedented, as serine hydroxymethyl transferase (GlyA) has different versions specific to either H 4 F [ 51 ] or H 4 MPT [ 54 , 55 ]. Additionally, MtdA in methylotrophic bacteria has been shown to react with either H 4 F or H 4 MPT [ 56 ].

Two phylogenetic trees were constructed with sequences of each of these MetF groups along with the most closely related homologs in the NCBI NR database ( Fig 5B and 5C ). The ANME-1 MetF homologs branched together with a diverse group of proteins found exclusively in uncultivated archaeal genomes, most of them identified as “Ca. Bathyarchaeota,” “Ca. Verstraetearchaeota,” and various members of the “Ca. Asgardarchaeota.” The MetF from the other ANME were all found within a monophyletic group containing other closely related Methanosarcinaceae. MetF and MetV in the Methanosarcinaneaeare found in gene clusters with other H 4 F-interacting enzymes that carry out important C1 reactions in biosynthesis, which led to the conclusion that H 4 F is used for biosynthesis in the Methanosarcinaceae [ 51 ]. This clustering of anabolic C1 genes is preserved in many of the ANME-2 and ANME-3, and we infer a similar function. For a detailed discussion of C1 anabolic metabolism in ANME, see below.

( A ) Amino acid sequence identity of MetF homologs found in ANME, “Ca. Argoarchaeum” and “Ca. Syntrophoarchaeum.” ANME-1 and “Ca. Syntrophoarchaeum” form one cluster based on sequence similarity, while ANME-2a, ANME-2b, ANME-2d, ANME-3, and “Ca. Argoarchaeum” form a second. Grayscale values represent percent identity. Sequences similar to the ANME-2/3 or ANME-1 clusters were retrieved via BLAST search of the NCBI nr database and used to construct phylogenetic trees of these 2 groups. ( B ) ANME-2, ANME-3, and “Ca. Argoarchaeum” cluster together with closely related members of the Methanosarcinaceae. ( C ) ANME-1 and “Ca. Syntrophoarchaeum” form a polyphyletic group within a diverse group of sequences derived from MAGs of uncultured archaea. Notably, the ANME-1 sp. SA is significantly different than the rest of the ANME-1. Roots for both trees lead to closely related MetF sequences from bacteria. Branch support values of 100% are labeled with closed circles, >50% with open circles. Tree scales represent substitutions per site. Tree construction parameters are found in the Materials and methods section. Alignments and tree files can be found in S1 Data . ANME, anaerobic methanotrophic; MAG, metagenome-assembled genome; MetF, methylenetetrahydrofolate reductase.

MetF is not only found in ANME-1, but also in other ANME groups and methanogenic members of the Methanosarcinaceae, where it is expected to function as a methylene-H 4 F-interacting enzyme involved in anabolic processes [ 51 ]. Since the potential Mer/MetF switch in ANME-1 appears to be the only significant modification to the central carbon oxidation pathway in ANME, we investigated the distribution, phylogenetic placement, and genomic context of MetF in ANME to try and better understand the evolutionary history of these proteins. MetF homologs found in ANME-1 were clearly very different at the sequence level from those found in other ANME genomes ( Fig 5A ), and, interestingly, the ANME-2c universally lack MetF of either type. All ANME MetFare found next to a MetV gene, which is a common feature of MetF in acetogenic bacteria, and evidence suggests a complex forms between the proteins encoded by these 2 genes [ 52 ].

Methylenetetrahydrofolate reductase (MetF) has been proposed to act in the third step as a replacement for Mer in ANME-1 [ 10 , 42 ] and is transcribed at similar levels as other genes in the reverse methanogenesis pathway [ 9 ] ( S3 Data ). MetF is structurally similar to Mer and completes the same step of the Wood–Ljungdahl pathway in bacteria but uses NADPH as the electron donor rather than F 420 H 2 and interacts with C1-bound tetrahydrofolate (H 4 F) instead of tetrahydromethanopterin (H 4 MPT).

Based on the above observations, we conclude that the transition from methanogenesis to methanotrophy required relatively little biochemical novelty within the central C1-carrying pathway of ANME energy metabolism. The loss of Mer in ANME-1 ( Fig 3 ) remains the single major variation to the central C1-carrying pathway of ANME energy metabolism. Some paralogs of other steps in the pathway exist, but except for Mch2 in some ANME-2, these are less well conserved and less well transcribed than those previously found in methanogenic archaea. While McrA in ANME-1 is slightly incongruous with its genome phylogeny and was found to bind a modified F 430 cofactor [ 50 ], we see little evidence for significant changes in ANME-2 or ANME-3, suggesting that these modifications are not necessary for using Mcr to activate methane. These results are broadly consistent with biochemical studies that have demonstrated the reversibility of the enzymes in this pathway [ 36 – 38 ], suggesting that there has likely been little specialization in terms of their directionality during the evolution of the ANME archaea.

The seventh and final step of the methanogenesis pathway is carried out by formyl-methanofuran dehydrogenase ( Fig 3 ). Two major variants of formylmethanofuran dehydrogenase are present in methanogens that contain either tungsten or molybdenum metal centers in their active sites (Fwd and Fmd, respectively), and multiple paralogs of both can be found in methanogens such as Methanosarcina acetivorans [ 48 ]. Based on homology to versions of these enzymes in M. acetivorans, it appears that ANME largely contain the Fwd version, which have been detected in proteomic analyses of methane seeps [ 49 ] ( Fig 4F ). Some members of the ANME-2a, ANME-2b, and ANME-2c have the Fmd version as well, and in the ANME-2c, the genes encoding Fmd and Fwd occur together in a single large gene cluster. As was observed in previous ANME-1 genomes, the FwdFG subunits are not present in any ANME-1 [ 6 , 7 , 12 ].

Divergent paralogs of formylmethanofuran-tetrahydromethanopterin N-formyltransferase (Ftr) were found in ANME-1, ANME-2a, and ANME-2b. These Ftr2 clustered together in the phylogenetic tree with Ftr genes from Archaeoglobales and deeper branching hydrogenotrophic methanogens such as Methanopyrus kandleri and Methanothermobacter marburgensis ( Fig 4E ). These archaea all contain both Ftr1 and Ftr2, and only the Ftr1 versions have been biochemically characterized to our knowledge [ 47 ]. In the cases where transcriptomic information is available, Ftr1 is more highly transcribed in ANME and is therefore expected to be the dominant version utilized in ANME energy metabolism under the AOM conditions tested ( S3 Data ).

The reaction catalyzed by Mch is a curious step of the methanogenesis pathway to have a strongly supported, ANME-specific clade of enzymes. The cyclohydrolase reaction is thought to occur essentially at equilibrium [ 37 ], so it is unclear what evolutionary pressure would result in such a stark difference between Mch2 enzymes in some ANME-2 and their closely related methanogenic relatives. The pterin moiety of the H 4 MPT analog could vary between ANME and the methanogens in the Methanosarcinaceae. At least 5 pterins are known to be found in H 4 MPT analogs in methanogenic archaea: methanopterin, sarcinopterin, tatiopterin-I, tatiopterin-O, and thermopterin [ 46 ]. However, this level of sequence variation is not observed in other enzymatic steps of the pathway, which might be expected if the divergent Mch2 was the result of a significantly different form of C1-carrier in ANME.

The fourth step of the pathway is catalyzed by the F 420 -dependent N 5 ,N 10 -methylene-H 4 MPT dehydrogenase (Mtd) enzyme ( Fig 3 ). Only a single copy of the gene encoding Mtd was found in any genome, and the phylogeny of the predicted protein sequence is largely congruent with the genome phylogenies ( Fig 4C ). The fifth step in the pathway is catalyzed by N 5 ,N 10 -methenyl-H 4 MPT cyclohydrolase (Mch). The analysis of early fosmid libraries revealed a gene on an ANME-2c–assigned fosmid encoding Mch [ 8 ], which was highly divergent from an Mch located on an ANME-2–assigned fosmid from a previous study [ 6 ]. This led the authors to question whether the Mch had diverged rapidly between ANME-2c and different ANME-2 groups, or whether multiple Mch copies were present as paralogs within ANME-2c genomes. Interestingly, ANME-2a, ANME-2b, ANME-2c, and ANME-2d all share a well-supported monophyletic group of Mch genes (Mch2) that are very different from those of closely related methanogens (Mch1) ( Fig 4D ). The Mch2 cluster corresponds to the gene identified as being closely related to Mch found in Archaeoglobus [ 8 ]. The ANME-2c genomes analyzed here also contain a copy of Mch1 that is closely related to those found in the methanogenic Methanosarcinaceae, which corresponds to the gene found in the first set of ANME-2–assigned fosmids [ 6 ]. Apparently, ANME-2c contain both copies of this gene, while the ANME-2a, ANME-2b, and ANME-2d only contain the divergent Mch2. ANME-3 contain a copy very similar to those of their close methanogenic relatives.

Phylogenetic trees constructed from protein sequences of enzymes involved in the methanogenesis pathway in ANME and related archaea. Mcr phylogeny is presented in Fig 1 . Numbers next to clades indicate whether the cluster is closely related to those found in Methanosarcinaceae [ 1 ] or are distantly related homologs [ 2 ], matching labels in Fig 3 . ( A ) MtrE, N 5 -methyl-H 4 MPT:coenzyme M methyltransferase subunit E; ( B ) Mer, methylene-H 4 MPT reductase; ( C ) Mtd, F 420 -dependent methylene-H 4 MPT dehydrogenase; ( D ) Mch, N 5 ,N 10 -methenyl-H 4 MPT cyclohydrolase; ( E ) Ftr, formylmethanofuran-H 4 MPT formyltransferase; ( F ) FmdB/FwdB, formyl-methanofuran dehydrogenase subunit B, molybdenum/tungsten variety, respectively. Branch support values of 100% are labeled with closed circles, >50% with open circles. Tree scales represent substitutions per site. Tree construction parameters are found in the Materials and methods section. Alignments and tree files can be found in S1 Data . ANME, anaerobic methanotrophic; Mcr, methyl-coenzyme M reductase.

To understand whether the transition to a methanotrophic lifestyle is accompanied by a significant diversification of the enzymes catalyzing the reactions of the reverse methanogenesis pathway, we built phylogenetic trees of the enzymes involved in each step. With regard to the first step, i.e., the presumed involvement of the McrA in methane activation, the McrA phylogeny largely tracks with phylogenetic marker trees, except for ANME-1 ( Fig 1 ). The second step of the pathway is carried out by the N 5 -methyl-H 4 MPT:coenzyme M methyltransferase (Mtr) complex ( Fig 3 ). All ANME clades contain Mtr homologs that are phylogenetically consistent with their genome phylogeny ( Fig 4A ). However, a highly divergent second copy of the entire Mtr complex exists that was first observed in the genome of an ANME-2a from an enrichment culture [ 11 ]. We find this second “Mtr2” to be sporadically distributed through the ANME-2. Absent from the other ANME-2a genomes, Mtr2 is present in ANME-2c, one ANME-2d, and both “Candidatus Argoarchaeum.” These Mtr2 complexes form a monophyletic group that is phylogenetically distinct from all previously described methanogens yet still contains the same gene synteny in the 8-subunit cluster. In addition, all the Mtr2 gene clusters are upstream of a highly divergent Mer2, which catalyzes the third step in the methane oxidation pathway. All ANME-2 and ANME-3 clades contain a less divergent copy of Mer that tracks with their genome phylogeny (Mer1) ( Fig 4B ).

The proteins responsible for the 7 steps of methanogenesis from CO 2 . Colored boxes represent presence of homologs of these proteins in ANME genomes. Missing genes are represented by gray boxes with diagonal line fill. Numbers in the second column represent estimated genome completeness. When genes are together in a gene cluster, their boxes are displayed fused together. If a gene cluster appears truncated by the end of a contig, it is depicted by a serrated edge on the box representing the last gene on the contig. Numbers following protein names indicate whether the enzyme is closely related to those found in Methanosarcinaceae [ 1 ] or are distantly related homologs [ 2 ]. Question mark represents hypothetical protein of unknown function found clustered with Mer2. Tree orienting genome order is the same as found Fig 1B . For details on paralog phylogenetic relations, see Fig 4 . Gene accession numbers can be found in S2 Data . ANME, anaerobic methanotrophic; Fmd/Fwd, formyl-methanofuran dehydrogenase; Ftr, formylmethanofuran-H 4 MPT formyltransferase; Mch, N 5 ,N 10 -methenyl-H 4 MPT cyclohydrolase; Mcr, methyl-coenzyme M reductase; Mer, methylene-H 4 MPT reductase; Mtd, F 420 -dependent methylene-H 4 MPT dehydrogenase; Mtr, N 5 -methyl-H 4 MPT:coenzyme M methyltransferase.

Our analysis of this expanded set of ANME genomes is consistent with early genomic work that identified reverse methanogenesis as the most likely pathway of carbon oxidation in the ANME archaea [ 6 , 7 , 11 ]. Genes for all 7 steps of the methanogenesis pathway were found in all ANME clades ( Fig 3 ). The only consistent exception is F 420 -dependent methylenetetrahydromethanopterin reductase (Mer), which is absent from all 19 ANME-1 genomes as well as the “Ca. Syntrophoarchaeum,” as observed previously [ 6 , 7 , 10 , 18 ]. This modification of the canonical methanogenesis pathway is a common feature of the entire class Syntrophoarchaeia. Some ANME contain paralogous copies of enzymes carrying out certain steps of the pathway, and these are indicated in Fig 3 . Notably, none of the ANME genomes contain the specific methyltransferases for methanol [ 43 ], methylamines [ 44 ], or methylated sulfur compounds [ 45 ] used for methylotrophic methanogenesis in the Methanosarcinaceae. This strongly argues against ANME archaea using methylated compounds as intermediates or end products of methane oxidation.

We can only speculate on the functions of all these modified HdrA paralogs, but it is clear that ANME have an exceptionally diverse potential of bifurcation/confurcation complexes at their disposal and that the flow of electrons through these complexes will necessarily be different than in the traditional HdrABC-MvhADG complex due to domain gain and loss within the HdrA homologs, as well as the replacement of hydrogenase subunits with various other input modules. While any electron confurcation schemes through HdrABC is speculative at this point, they seem to be viable option for Fd 2− and CoM-SH/CoB-SH oxidation in ANME due to their widespread conservation across all ANME clades, particularly in ANME-1 that lack HdrDE and most known Fd 2− oxidation systems. Additionally, an examination of previously reported transcriptomic information indicates that these complexes are often transcribed at levels on par with other components of the reverse methanogenesis pathway ( S3 Data ). A key question that remains is what substrate the FdhA homologs in ANME act upon.

Another possibility is confurcation of Fd 2− and CoM-SH/CoB-SH electrons to F 420 H 2 . FdhB and homologous proteins carry out oxidoreductase reactions with F 420 [ 81 ] in F 420 -dependent hydrogenases (FrhB), formate dehydrogenases (FdhB), F 420 -dependent sulfite reductase (Fsr), and the Fpo complex (FpoF). The observation of FdhB homologs in gene clusters with HdrA in ANME-2d lead to the suggestion that these complexes could confurcate Fd 2− and CoM-SH/CoB-SH to reduce 2 molecules of F 420 , although this model proposes no role for the FdhA homologs present in the gene clusters [ 19 ]. Some HdrA gene clusters only contain FdhB and not an FdhA homolog; however, these are found only in ANME-1 genomes ( Fig 8A , S10 Fig ). Additionally, there is evidence that certain HdrABC complexes can interact with F 420 without any additional subunits [ 62 , 74 ]. If such a reaction were to produce F 420 H 2 in ANME, its reoxidation through the Fpo complex could be coupled to energy conservation as has been previously proposed in M. acetivorans [ 82 ].

Based on the extensive conservation of FdhAB subunits in Hdr gene clusters in nearly all ANME groups, it is tempting to speculate that these enzymes may have an important role in CoM-SH/CoB-SH oxidation through electron confurcation. A reversal of electron bifurcation from formate would be a reasonable reaction to expect in ANME metabolism, as formate is a common syntrophic intermediate. Based on the ratio of formate to bicarbonate in the cell, one could envision the midpoint potential to lie appropriately between that of Fd 2− and CoM-SH/CoB-SH to receive electrons from both. FdhA belongs to the molybdopterin oxidoreductase family, members of which act on many different substrates including formate, nitrate, and DMSO among others in assimilatory and dissimilatory processes [ 79 ]. The FdhA homologs from ANME genomes are often annotated as formate dehydrogenases and retain a conserved cysteine ligand for binding the molybdenum atom (sometimes a selenocysteine in other organisms). However, they are very distantly related to any biochemically characterized enzymes, making it difficult to assign their substrate with any level of certainty. A very closely related gene cluster with HdrA and FdhAB can be found in the genomes of the methylotrophic methanogen genus Methanolobus, potentially providing a good opportunity to study the substrate specificity of this specific group of molybdopterin oxidoreductases in a pure culture organism. These are the only methanogens that encode this type of Hdr gene cluster, and it is worth noting in this context that the Methanolobus are incapable of using formate as a methanogenic substrate [ 80 ].

Another remarkable HdrA modification is a 500 amino acid insertion in the ferredoxin domain of HdrA 9, 10, and 11, present only in some ANME-1 genomes ( Fig 8 ). This insertion shares 45% amino acid sequence identity to the NfnB subunit of the bifurcating NADH-dependent reduced ferredoxin:NADP oxidoreductase crystalized from Thermotoga maritima [ 78 ]. NfnB binds the b-FAD cofactor thought to be responsible for bifurcation in the NfnAB complex, potentially giving these HdrA-NfnB genes 2 electron bifurcation sites, and suggests NADPH as a possible electron carrier in these compounds.

Many HdrA homologs in ANME were lacking various domains present in the crystalized complex from M. wolfeii ( Fig 8B ). An interesting pattern emerges in gene clusters with tandem HdrA genes (i.e., HdrA 5 and 6, or 12 and 13), in which one copy has the N-terminal domain but lacks the Cys197 ligand for the N-terminal FeS cluster in its thioredoxin reductase domain (Hdr5 and 12), while the second copy lacks the N-terminal domain but has Cys197 (Hdr6 and 13). It seems likely that these tandem HdrA genes form heterodimers, breaking the rotational 2-fold symmetry found in the crystal structure from M. wolfeii ( Fig 7B ). This pattern of Cys197/N-terminal domain complementarity can also be seen in largest HdrA homologs 3 and 4, which are fusions of 2 HdrA genes. In both cases, the N-terminal HdrA contains the N-terminal FeS cluster-binding domain but lacks Cys197, while the C-terminal HdrA contains Cys197 and no N-terminal domain, suggesting these HdrA fusions within a single continuous peptide may act as their own heterodimeric partners. Evidence for symmetry breaking in HdrABC complexes has recently been demonstrated in Methanococcus maripaludis, where a hydrogenase and a formate dehydrogenase can be simultaneously bound to dimerized HdrABC [ 77 ]. This appears to be the result of HdrABC’s ability to bind either hydrogenase or formate dehydrogenase in these methanogens, resulting in a mixture of subunits, whereas in the ANME case, this asymmetric form may be imposed by a heterodimer of 2 HdrA paralogs.

( A ) Examples of gene clusters containing HdrA genes from select ANME genomes. HdrA paralogs present in ANME have extensive modification to the domain structure as compared to the HdrA crystalized from M. wolfeii (see Fig 7 for details of HdrA structure). These domains and associated protein subunits are illustrated with the gene context and orientation. ( B ) Illustration of conserved domains and cofactor binding residues in the 13 HdrA clusters defined here. All HdrA appeared to retain residues responsible for interaction with FAD; however, the presence of FeS-binding cysteine residues and entire domains as defined on the M. wolfeii structure are variably retained. Importantly, tandem or fused HdrA appear to have complementary presence/absence of C-terminal ferredoxin domains and Cys197, suggesting the formation of a heterodimeric complex. ANME, anaerobic methanotrophic; FAD, flavin adenine dinucleotide.

By aligning the ANME HdrA paralogs and comparing the presence of domains, conserved residues, sequence similarity, and their genomic context, we clustered the dominant HdrA homologs into 14 groups ( Fig 8A , S8 Fig ). Most methanogen genomes contain 1 to 2 copies of hdrABC gene clusters, but ANME, and in particular ANME-1, have a greater abundance of HdrA homologs that exceed the number of hdrBC homologs ( S9 Fig ). Some of these HdrA homologs exceed 1,000 amino acids in length. In comparison, HdrA from methanogens is usually 650 amino acids in length, with some slightly larger homologs occurring due to the fusion of hdrA and mvhD [ 76 ]. The gene clusters containing HdrA homologs often contained HdrBC and MvhD as expected, but more unexpected was the co-occurrence of multiple copies of HdrA and the occasional presence of HdrD-like proteins ( Fig 8A , S10 Fig ). Although HdrD is a fusion of HdrB and HdrC, it is not common to find the fused form in gene clusters with HdrA in methanogens. Distant homologs of the F 420 -dependent formate dehydrogenase FdhAB were also found in HdrA gene clusters in all ANME groups. The gene cluster containing HdrA2 and 3 along with FdhAB-like genes in ANME-2a was discussed in detail in one of the earliest studies of ANME fosmid libraries [ 8 ], and the significance of this cluster has now been substantiated by its broad conservation as well as reasonably high transcription levels ( S3 Data ).

Depiction of the primary structure of HdrA and the quaternary structure of the HdrABC-MvhADG complex based on the structure from M. wolfeii. ( A ) HdrA can be broken down into 4 domains, the positions of these domains and key iron–sulfur cluster binding cysteines are illustrated, scale denotes amino acid position in the M. wolfeii sequence. ( B ) Quaternary structure of the entire HdrABC-MvhABG complex illustrating the dimeric structure. Metal cofactors involved in the oxidation/reduction of substrates or electron transport through the complex are highlighted. ( C ) Detail of HdrA domain structure highlighting cofactor position and proposed electron flow from MvhD in through the C-terminal ferredoxin, bifurcation through the FAD cofactor, with 2 electrons flowing out through HdrBC via the thioredoxin reductase domain’s FeS cluster, while 2 other electrons flow out through the inserted ferredoxin domain, presumably to free ferredoxin (Fd 2− ). Importantly, for the proposed heterodimeric HdrA discussed here, this latter electron flow passes through the FeS cluster bound through a combination of Cys residues in the N-terminal domain, combined with a single Cys from the other HdrA subunit (Cys197 highlighted in red). FAD, flavin adenine dinucleotide.

A comparison of ANME HdrA homologs to crystal structures of the entire HdrABC-MvhADG complex purified from Methanothermococcus wolfeii [ 75 ] reveals some stark differences in how these complexes may facilitate electron bifurcation. HdrA in methanogens normally consists of 4 domains, an N-terminal domain with an iron sulfur cluster, a thioredoxin reductase domain, which binds the bifurcating flavin adenine dinucleotide (FAD), and 2 ferredoxin domains ( Fig 7 ). The MvhAG hydrogenase feeds electrons through MvhD and the C-terminal ferredoxin domain to the FAD where they are bifurcated, one pair passing up through HdrBC onto CoM-S-S-CoB, and the other through the inserted ferredoxin domain to a free soluble ferredoxin. Interestingly, the heterohexameric HdrABC-MvhADG forms dimers, and one of the cysteine ligands for the N-terminal iron sulfur cluster comes from the thioredoxin-reductase domain of the other copy of HdrA (Cys197), indicating an obligate dimeric nature of the complex in M. wolfeii ( Fig 7C ).

While a strict reversal of hydrogen-dependent electron bifurcation seems unlikely in ANME, the potential involvement of alternative bifurcation/confurcation reactions in ANME metabolism is supported by the broad distribution and conservation of unusual HdrABC homologs, even in those ANME genomes containing HdrDE. Alternative bifurcation reactions have been proposed in methanogens, where formate or F 420 H 2 can serve as the electron donor in place of hydrogen. In the case of formate, MvhA and G are replaced by the F 420 H 2 -dependent formate dehydrogenase genes FdhAB [ 72 , 73 ]. In the case of F 420 H 2 , certain HdrABC complexes might be able to interact with F 420 H 2 without any additional protein subunits [ 62 , 74 ].

In most cases, a strict reversal of H 2 electron bifurcation is not possible in ANME due to a lack of MvhAG genes that encode the NiFe hydrogenase large and small subunits. The exceptions are 2 small subclades within ANME-2c and ANME-1 that contain NiFe hydrogenases similar to Mvh. These 2 subclades contain genomes recovered from different environments; a South African gold mine (SA) and the Gulf of Mexico (GoMg4) for ANME-1 versus a hydrocarbon seep off of Santa Barbara, CA (COP2) and a mud volcano from the Mediterranean (AMVER4-31) for ANME-2c. The majority of ANME-2c and ANME-1, however, lack these genes, and they are completely absent in ANME-2a, ANME-2b, and ANME-3, suggesting that confurcation of electrons from Fd 2− and CoM-SH/CoB-SH to H 2 is not a dominant process in most ANME lineages.

HdrDE gene clusters are absent in all ANME-1 genomes. This, combined with the lack of Rnf and Ech, suggests that ANME-1 in particular could rely on HdrABC complexes for both CoM-SH/CoB-SH and ferredoxin oxidation. Since its discovery just over 10 years ago [ 70 ], flavin-based electron bifurcation and confurcation have been shown to be a key energy conversion point in many anaerobic metabolisms, with a great diversity of different electron donors and acceptors [ 71 ]. In the complete HdrABC-MvhADG complex from methanogens, the endergonic reduction of Fd (E 0 ’ = −500 mV) with H 2 (E 0 ’ = −420 mV) is driven by the exergonic reduction of heterodisulfide (E 0 ’ = −145 mV) with H 2 .

Two enzyme systems from methanogens could potentially produce heterodisulfide by running in reverse: soluble HdrABC complexes using a confurcation mechanism as mentioned above, or HdrDE, a membrane-bound system, which would deposit electrons on methanophenazine ( Fig 6C ). ANME-2a, ANME-2b, ANME-2c, ANME-2d, and ANME-3 genomes contained HdrDE genes similar to ones from closely related methanogens. The reaction carried out by this complex in methanogens is expected to be readily reversible, with electrons from CoM-SH/CoB-SH oxidation being deposited on methanophenazine. This electron transfer reaction is slightly endergonic at standard state due to the redox potential of methanophenazine (E 0 ’ = −165 mV) being lower than that of heterodisulfide by 20 mV. This reaction may be exergonic under physiological concentrations of oxidized and reduced species. Alternatively, the slightly endergonic nature of the electron transfer could be overcome by a quinol loop-like mechanism, where 2 protons are released in the cytoplasm from CoM-SH/CoB-SH oxidation and 2 protons are consumed on the outer face of the cytoplasmic membrane to form reduced MpH 2 , thus dissipating proton motive force [ 69 ]. Electron transfer from the cytoplasm to the outer face of the membrane occurs through the b-type cytochromes in the HdrE subunit. The HdrDE complexes are well transcribed in the ANME-2a, ANME-2c, and ANME-2d ( S3 Data ).

The last oxidation that needs to occur during this part of ANME energy metabolism is CoM-SH/CoB-SH oxidation to the heterodisulfide (CoM-S-S-CoB). This oxidation is the most energetically challenging because the relatively high midpoint potential of the heterodisulfide (E 0 ’ = −145 mV) rules out most methanogenic electron carriers as acceptors without the input of energy to force electrons “uphill” to a lower redox potential. It also represents the second net reaction in the ANME energy metabolism that does not occur during canonical methanogenesis (the first step being methane activation through Mcr). In methylotrophic methanogens, heterodisulfide is the terminal electron acceptor for the 6 electrons that have passed through F 420 H 2 and Fd 2− discussed in the previous sections.

A final possible mechanism of Fd 2− oxidation is through soluble Hdr-mediated flavin-based electron confurcation. This process has some biochemical support for operating in this direction [ 62 ] but would constitute a reversal of the well-accepted electron bifurcation mechanism used by many hydrogenotrophic methanogens. In the best-characterized examples of this process, an enzyme complex of soluble Hdr and a hydrogenase (hdrABC-mvhADG) reduce ferredoxin and heterodisulfide with 4 electrons sourced from 2 hydrogen molecules [ 68 ]. If reversed, this reaction could potentially oxidize Fd 2− . As this mechanism would also involve CoM-SH/CoB-SH oxidation, it is discussed in detail in the following section.

Another possibility would be for soluble FqoF to act as a Fd 2− :F 420 oxidoreductase and, then, the subsequent oxidation of F 420 H 2 by a normal Fqo complex ( Fig 6B ). This pathway has been proposed in Methanosarcina mazei based on the Fd 2− :F 420 oxidoreductase activity of soluble FpoF [ 67 ]. In either of these Fpo/Fqo-dependent Fd 2− oxidation pathways, the oxidation of Fd 2− would result in the reduction of a membrane-bound electron carrier coupled to energy conservation in the form of proton translocation by the Fpo/Fqo complexes.

Because the majority of ANME-1 do not contain an Rnf complex, an alternative mechanism is needed to explain how ferredoxin is recycled. One option is the Fpo-dependent mechanism proposed in Methanosaeta thermophila where ferredoxin is oxidized using an Fpo complex without the FpoF subunit [ 61 ]. If this pathway operates in ANME-1, the reduced ferredoxin would donate electrons directly to the iron sulfur clusters found in the FqoBCDI subunits. Since FqoF homologs are encoded in the ANME-1 genomes and are expected to be used in complete Fqo complexes to oxidize F 420 H 2 as described above, this Fd 2− oxidation strategy would require some Fqo complexes to have FqoF bound, while others do not.

An additional ANME-specific modification to the rnf gene cluster is the inclusion of a membrane-bound b-type cytochrome in ANME-2a, ANME-2b, and ANME-3. This type of cytochrome is generally involved in electron transfer between membrane-bound and soluble electron carriers and has no closely related homologs in methanogens. If the protein encoded by this gene is incorporated into the Rnf complex, it could be of great importance to the flow of electrons in these groups. This observation is particularly striking in ANME-3 since their close methanogenic relatives in the Methanosarcinaceae lack this b-type cytochrome. This means that ANME-3 has acquired this subunit by horizontal gene transfer from an ANME-2a or ANME-2b or that all of the related methanogens have lost it. Whichever evolutionary scenario is correct, this represents an important ANME-specific modification to a key bioenergetic complex.

Rnf was not found in the ANME-1 genomes analyzed here, except for 2 genomes from a genus-level subclade recovered from a South African gold mine aquifer (SA) and from a hydrocarbon seep from the Gulf of Mexico (GoMg4) ( Table 1 , S6 Fig ). Rnf gene clusters in ANME-2a, ANME-2b, and ANME-3 contain homologs of the cytochrome c subunit and MA0665 normally found in methanogens, but surprisingly, both were missing in the Rnf gene clusters from all ANME-2c and the 2 ANME-1. Homologs for these subunits were not identified elsewhere in any of these genomes. Based on the current information, it is unclear what reaction the ANME-2c and ANME-1 Rnf are performing since it has been demonstrated that this cytochrome c is involved in the electron transfer to methanophenazine [ 66 ]. It is possible that Rnf retains the ability to transfer electrons from ferredoxin to a membrane-bound electron carrier in these lineages or, alternatively, they could function as ferredoxin:NAD + oxidoreductase, as found in bacteria. It is unclear what role the latter function could play in our current model of ANME metabolism.

Genomes from many members of the ANME-2a, ANME-2b, ANME-2c, and ANME-3 contained homologs of the Rnf complex ( S6 Fig ). Rnf was first characterized in bacteria as an enzyme that performed the endergonic transfer of electrons from NADH to ferredoxin by dissipating sodium motive force https://paperpile.com/c/aoT9UA/PFhT [ 63 ]. In methanogens, however, Rnf is thought to couple the exergonic electron transfer from ferredoxin to methanophenazine with the endergonic translocation of sodium ions to the positive side of the cytoplasmic membrane. This activity in methanogens involves a novel multiheme cytochrome c subunit encoded in their Rnf gene clusters [ 60 ], along with a small conserved integral membrane protein (MA0665 in M. acetivorans). This enzyme complex operates in methylotrophic members of the Methanosarcinaceae, which do not conduct hydrogen cycling, and in some cases is required for energy conservation during acetoclastic methanogenesis [ 64 , 65 ].

Similar to F 420 H 2 oxidation, several mechanisms are known for coupling ferredoxin oxidation to energy conservation in methanogens. Four major pathways have been proposed: (1) a hydrogen cycling mechanism using energy-conserving hydrogenase (Ech) [ 57 ]; (2) oxidation with a modified version of the Rhodobacter nitrogen fixation (Rnf) complex [ 60 ]; (3) an Fpo-dependent pathway [ 61 ]; and (4) a heterodisulfide reductase (Hdr)-mediated electron confurcation [ 62 ] ( Fig 6B ). The Ech model is easily ruled out by the lack of Ech homologs in all marine ANME genomes, although these are present in the freshwater ANME-2d genomes [ 19 , 20 ].

This leaves Fpo as the much more likely candidate for F 420 H 2 oxidation in ANME. Fpo is a homolog of respiratory complex I, and in the Methanosarcinaceae, it couples the transfer of electrons from F 420 H 2 to the membrane soluble electron carrier methanophenazine with the translocation of protons across the cell membrane. In sulfate-reducing Archaeoglobales, the homologous F 420 H 2 :quinone oxidoreductase complex (Fqo) utilizes a membrane-soluble quinone acceptor instead of methanophenazine [ 58 , 59 ]. These complexes are conserved across all ANME clades ( S6 Fig ), and phylogenetic analysis of Fpo shows that homologs from ANME-2a, ANME-2b, ANME-2c, and ANME-3 are most similar to the Methanosarcinaceae, suggesting that they also may utilize methanophenazine. The homologs from ANME-1 were distinct from those of the other ANME clades and were most similar to the Fqo described from Archaeoglobales as previously reported [ 6 ] ( S7 Fig ). Consistent with this possibility, ANME-1 genomes also contain homologs of the futalosine pathway used for menaquinone biosynthesis that are absent in the other marine ANME clades. This suggests that ANME-1 use a quinone as their membrane-soluble electron carrier as was previously suggested [ 42 ]. In either case, Fpo and Fqo are expected to be important points of F 420 H 2 oxidation and membrane energization in the ANME archaea.

Some energy conservation systems discovered in methanogenic archaea are conserved in ANME archaea (colored fill), while others appear absent (transparent gray with diagonal line fill). ( A ) F 420 H 2 oxidation is coupled to proton translocation in methylotrophic methanogens via the Fpo/Fqo complex or by the production of H 2 by Frh and subsequent oxidation by Vht. In either case, electrons are ultimately deposited on MpH 2 . Neither Frh or Vht complexes have been observed in any ANME genomes analyzed here. ( B ) Fd 2− oxidation can be coupled either to sodium motive force or proton motive force in methylotrophic methanogens. The Rnf complex catalyzes Fd 2− :Mp oxidoreductase reaction coupled to sodium translocation and is found in a number of methanogens and ANME. The ANME-2c contain most of the complex but lack the cytochrome c, cytochrome b, and MA0665 subunits, so their activity is difficult to predict. Ech and Vht can combine to produce net proton translocation via H 2 diffusion in methylotrophic methanogens, but neither complex is found in ANME. FpoF can catalyze a Fd 2− :F 420 oxidoreductase reaction, and F 420 H 2 could then pass through the Fpo/Fqo complex. Various HdrABC complexes are present in all ANME genomes and could in principle oxidize Fd 2− and CoM-SH/CoB-SH through a reversal of electron bifurcation reaction. The electron acceptor in this process is likely to not be H 2 in most ANME groups due to the absence of MvhG and MvhA. ( C ) Besides, the HdrABC complexes mentioned above and second possible CoM-SH/CoB-SH oxidation strategy would be a reversal of the HdrDE reaction found in methylotrophic methanogens. In ANME, the reaction would have to proceed in the direction illustrated and therefore would dissipate proton motive force by consuming a proton on the positive side of the membrane. For presence/absence of these systems in ANME genomes analyzed here, see S6 Fig . ANME, anaerobic methanotrophic; Ech, energy-conserving hydrogenase; Fpo, F 420 H 2 :methanophenazine oxidoreductase; Fqo, F 420 H 2 :quinone oxidoreductase complex; Frh, F 420 -reducing hydrogenase.

F 420 H 2 oxidation for the purpose of energy conservation in methylotrophic methanogens can occur via the F 420 H 2 :methanophenazine oxidoreductase complex (Fpo) or the F 420 -reducing hydrogenase (Frh) ( Fig 6A ). F 420 H 2 oxidation by the cytoplasmic Frh produces H 2 , which diffuses out of the cell and is subsequently oxidized on the positive side of the membrane by a membrane-bound hydrogenase (Vht) [ 57 ]. Our comparative analysis of ANME genomes does not support this mechanism of electron flow given the lack of both the NiFe hydrogenase subunit of Frh and Vht-like hydrogenases.

This is very similar to the situation in the energy metabolism of methylotrophic methanogens in the Methanosarcinaceae, which are close relatives of ANME-2 and ANME-3. In these methanogens, methyl groups are transferred from substrates such as methanol or methylamines onto CoM via substrate-specific methyltransferases. The methyl group is subsequently oxidized to CO 2 via a reversal of the first 6 steps of the methanogenesis pathway, consuming sodium motive force at Mtr and producing 2 F 420 H 2 and 1 Fd 2− that must be reoxidized coupled to energy conservation. In most cases, these electrons are transferred onto membrane-bound methanophenazine as an intermediate electron carrier in processes that lead to ion motive force generation. This ion motive force is then used to produce ATP via ATP synthase. The only oxidation reaction in ANME that does not normally occur in any methanogen is the production of the heterodisulfide (CoM-S-S-CoB) from the free sulfides CoM-SH and CoB-SH. Many of the strategies of energy conservation coupled to cytoplasmic electron carrier oxidation that have been characterized in the methylotrophic methanogens appear to be conserved in the ANME archaea.

After methane is oxidized to CO 2 in the pathway described above, 4 electrons will be found on 2 molecules of F 420 H 2 , 2 electrons on reduced ferredoxin (Fd 2− ), and 2 electrons on the reduced forms of coenzyme M and coenzyme B (CoM-SH, CoB-SH). If the proposed Mer/MetF switch in ANME-1 is correct, then it is possible that MetF may have produced a reduced NADPH instead of one of the F 420 H 2 ( Fig 2 ). Importantly, these C1 oxidation reactions do not produce energy for the cell and will in fact consume sodium motive force at the step catalyzed by Mtr. This investment of energy helps drive the reaction in the oxidative direction, and the effect of energy investment can be seen in the decreased redox potential of the 8 methane-derived electrons: The CH 4 /CO 2 redox couple has a standard state midpoint potential of −240 mV, and the average midpoint potential of the electrons once transferred to their cytoplasmic electron carriers is approximately −340 mV (2 on Fd 2− (−500 mV), 4 on F 420 H 2 (−360 mV), and 2 on CoM-SH/CoB-SH (−145 mV)) [ 37 ]. The next phase of ANME energy metabolism, the reoxidation of these reduced cytoplasmic electron carriers, must conserve sufficient energy to overcome the loss at Mtr and support growth.

Energy metabolism phase 3: Genomic evidence for mechanisms of syntrophic electron transfer

The most enigmatic phase of marine ANME metabolism is the interspecies electron transfer to the sulfate-reducing partner, a process that appears to necessitate the formation of conspicuous multicellular aggregates of the 2 organisms [4,83]. The cytoplasmic electron carrier oxidation described in the previous section will result in 8 electrons on a combination of membrane-bound MpH 2 or QH 2 , and, possibly, some soluble electron carrier formed through an electron confurcation reaction oxidizing Fd2− and CoM-SH/CoB-SH. Based on energetic considerations and precedent from other known syntrophies, the hypothesized mechanisms for the AOM syntrophy have included the diffusive exchange of small molecules such as hydrogen, formate, acetate, methanol, methylamine [84], methyl-sulfides [85], zero-valent sulfur [86], and direct electron transfer using multiheme cytochrome c proteins [7,87,88]. We assessed the genomic potential for each of these syntrophic electron transfer strategies across our expanded sampling of ANME clades.

Hydrogen transfer. Hydrogen transfer is one of the most common forms of syntrophic electron transfer [89]. A classic mode of syntrophic growth involves hydrogenotrophic methanogens consuming H 2 produced by the fermentative metabolism of a syntrophic partner. A direct reversal of the methanogenic side of this syntrophy is not possible as the majority of ANME genomes lack any identifiable hydrogenases with the notable exception of MvhA homologs in a small set of ANME-1 and ANME-2c genomes as mentioned in the preceding section (S10 Fig). The first report of an ANME-1 genome from fosmid libraries contained a gene that appeared to encode an FeFe hydrogenase [7]; however, homologs of this gene were not found in any of the other ANME-1 genomes analyzed here. This lack of hydrogenases is consistent with the lack of Ech in the genomes of their syntrophic sulfate-reducing bacterial partners from cold seeps [90] and previous experimental results showing that the addition of excess hydrogen does not inhibit AOM [91,92]. Hydrogen has been shown to stimulate sulfate reduction in AOM sediments, suggesting that at least some portion of the sulfate-reducing community can utilize this electron donor [84,91,93]. In the case of the syntrophic thermophilic ANME-1-“Candidatus Desulfofervidus auxilii” consortium, hydrogen amendment suppresses growth of ANME-1, because the “Ca. D. auxilii” can grow alone on hydrogen and stops investing in electron transferring structures [94]. Together, genomic and physiological data suggest that the vast majority of ANME are incapable of producing hydrogen as an electron shuttle.

Formate transfer. Formate is another possible small molecule intermediate commonly involved in syntrophic electron transfer [95] and was predicted to be a possible intermediate in the ANME-SRB syntrophy based on early modeling studies [96]. As described above, FdhAB-like genes in association with Hdr-containing gene clusters are broadly distributed in ANME and could potentially be used to produce formate from Fd2− and CoM-SH/CoB-SH and are relatively highly transcribed (S3 Data), but their substrate is currently unknown. In addition to these complexes, 5 of the ANME-1 genomes contained a membrane-bound formate dehydrogenase with signal sequences predicted to mediate their secretion to the positive side of the cytoplasmic membrane. One of the ANME-1 genomes containing this formate dehydrogenase is based on early fosmid libraries that were analyzed extensively, and it was suggested to play a role in electron transfer, depicted as either being free floating in the pseudoperiplasm or anchored to the membrane via a c-type cytochrome [7]. Interestingly, the gene cluster encoding these formate dehydrogenases in 4 other ANME-1 genomes also contain DmsC-like genes that traditionally serve as membrane anchor subunits for members of the molybdopterin oxidoreductase family. The presence of these genes in the other ANME-1 genomes makes it likely that this is how the complex is anchored and interacts with the ANME-1 membrane-bound quinol, since this is a fairly common pattern for members of this family of oxidoreductases [97]. Interestingly, these formate dehydrogenase genes are distributed very sporadically through the ANME-1 genomes instead of being confined to a specific subclade (S10 Fig). This genomic evidence from different subclades provides a more plausible role for formate than hydrogen in ANME metabolism. However, like hydrogen, formate addition was not shown to inhibit AOM [84,91,92], which would be expected if it was the major electron donor to the sulfate-reducing partner in those sediments. The addition of formate to AOM sediments and enrichment cultures of ANME-SRB consortia has also had mixed results in terms of stimulating sulfate reduction [84,91,93]. Additionally, gene-encoding FdhC, a member of the formate–nitrate transporter family thought to be responsible for formate transport in many formate-utilizing methanogens [98,99], was absent from all ANME genomes.

Other soluble electron carriers. Acetate has been proposed as a syntrophic intermediate in sulfate-coupled AOM [7,100], and one could easily envision an energy conservation strategy in ANME archaea whereby methane and CO 2 are combined in a reversal of acetoclastic methanogenesis and energy is conserved via substrate-level phosphorylation through Pta and AckA as in acetogenic bacteria. Acetate, thus produced, could be used by the SRB partner for sulfate reduction. Two problems exist with this model that make it unlikely in any of the ANME groups represented here. First, Pta and AckA were absent from all ANME genomes. Second, this model requires that the 2 electrons transferred from methane to heterodisulfide (E 0 ’ = -−145 mV) would be used for the reduction of CO 2 , which requires ferredoxin (E 0 ’ = −500 mV). This would mean that the single ATP produced by substrate level phosphorylation would have to overcome the sodium motive energy lost in the Mtr step, force electrons energetically uphill by approximately 350 mV from the free disulfides onto ferredoxin, and have energy left over for growth. Experimental addition of acetate to AOM enrichments has not been found to decouple the partners and inhibit methane oxidation, as would be expected for a soluble extracellular intermediate [84,91,92]. Methylsulfides were suggested to be a potential intermediate in AOM [85], but as noted above, the methyltransferase genes for methylsulfides, methanol, and methylamines were not recovered in any ANME genomes, and experiments have not found methylsulfide-stimulated sulfate reduction in ANME-SRB enrichments [91]. Zero-valent sulfur has also been suggested as an intermediate in marine AOM using an ANME-2 enrichment [86]. However, in vitro studies with other ANME consortia did not show a similar response [88,93], and recent investigations into sulfur utilizing genes in ANME found no evidence for dissimilatory sulfate reduction [17]. If this process occurs, it is in a restricted group of ANME through a novel biochemical mechanism. Electron transfer is also possible via soluble organic shuttles. Shewanella oneidensis MR-1 was determined to utilize small molecule shuttles derived from menaquinone [101] or flavins [102] that help electrons pass through the external environment to their ultimate electron acceptor. The mechanism of producing these compounds is very poorly understood. Only recently, a major menaquinone-like shuttle was identified and its biosynthesis elucidated [103]. Such a shuttle-based electron transfer strategy could be potentially used for accepting electrons from the ANME Hdr complexes, but predicting the occurrence of these shuttling mechanisms from genomic information is not possible with our current understanding of these processes.

Direct interspecies electron transfer. If soluble electron carriers were produced by ANME, they could simply diffuse through the intracellular space between the ANME to the SRB without any specialized protein systems. Specific transporters or permeases would only be needed if the compounds produced in the ANME cytoplasm are not membrane permeable. On the other hand, electrons on MpH 2 /QH 2 would require an oxidation system with no analog in methanogens. In cytochrome-containing methanogens, the eventual electron acceptor for these membrane-bound electrons is the heterodisulfide located in the cytoplasm, whereas the terminal electron acceptor of marine ANME would be their partner SRB located outside of the cell. This is a respiratory challenge that is similar to the one facing bacteria that carry out extracellular electron transfer (EET) to utilize insoluble metal oxides or electrodes as terminal electron acceptors. Instead of having a soluble terminal electron acceptor that can diffuse to the cytoplasmic membrane and be reduced by a terminal oxidase, these organisms make conduits for their electron transport chain to extend through the periplasm and outer membranes to interact directly with an electron acceptor in the extracellular space [104]. Various systems have been discovered in gram-negative bacteria that perform EET. These generally consist of a quinol:cytochrome c oxidoreductase enzyme, small soluble cytochrome c intermediates between the inner and outer membranes, and a beta barrel/decaheme cytochrome c protein complex that is embedded in the outer membrane (Fig 9A) [105]. In bacteria that form large conductive biofilms, the extracellular space is also densely packed with secreted c-type cytochromes, which can form large conductive complexes [106]. Based on the similarity of the metabolic challenge facing ANME and EET-capable bacteria, as well as the discovery of analogs of bacterial EET systems in ANME genomes [7,87,88], an attractive hypothesis is that the ANME-SRB syntrophy is based on direct electron transfer between the syntrophic partners. However, in contrast to other EET-capable bacteria that use metals as terminal electron acceptors, the highly specific interspecies interaction observed in the AOM consortia may require the partner SRB to encode cognate electron-accepting systems. PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 9. Overview of proposed EET pathways. Comparison between EET systems known from gram-negative bacteria and proposed analogous systems in ANME archaea. (A) EET systems in gram-negative bacteria involve membrane-bound quinol:cytochrome c oxidoreductases (CbcL, ImcH, CymA, NetD), small soluble cytochromes apparently involved in electron transport through the periplasmic space (PpcA, Stc, PdsA), and a beta-barrel/decaheme cytochrome c protein complex (MtrCAB) that acts as an electron conduit by which electrons can transit through the outer membrane to the extracellular space filled with additional cytochrome c such as OmcZ and filaments of OmcS. (B) Analogous protein complexes found in ANME genomes that appear optimized for the challenges associated with EET in the archaeal cell architecture. MpH 2 :cytochrome c oxidoreductases are likely present in the form of gene clusters containing VhtC cytochrome b subunits together with large 7 or 11 heme-binding MHC proteins (Mco). Other potential options for this step could include the NapH homologs sporadically distributed through ANME genomes or through the action of the unique cytochrome b gene found in ANME Rnf clusters. Electron transfer through the outer proteinaceous S-layer requires a different mechanism than the beta-barrel/decaheme cytochrome strategy evolved in the EET-capable bacteria containing an outer membrane. This step is expected to be overcome by the giant ANME-specific MHC proteins containing S-layer domains allowing them to integrate into the S-layer structure. Very close homologs of OmcZ are found in ANME (see Fig 10). For details of S-layer MHC fusions, see Fig 11. ANME, anaerobic methanotrophic; EET, extracellular electron transfer; Mco, methanophenazine-cytochrome c oxidoreductase; MHC, multiheme c-type cytochrome; Rnf, Rhodobacter nitrogen fixation. https://doi.org/10.1371/journal.pbio.3001508.g009

Potential methanophenazine:cytochrome c oxidoreductase complexes. In bacteria capable of EET, the quinol:cytochrome c oxidoreduction step is carried out by a diverse group of nonhomologous enzymes. These can be as simple as CymA in S. oneidensis MR-1, which is a 4 heme-binding cytochrome c protein with a single-transmembrane helix [107], or more complex, involving multiple subunits such as the recently described NetBCD system in Aeromonas hydrophila [108]. These complexes are relatively easy to replace, as CymA knockouts in S. oneidensis MR-1 can be rescued by suppressor mutants that turn on the expression of completely unrelated quinol oxidoreductase complexes [109]. In Geobacter sulfurreducens, at least 2 quinol:cytochrome c systems coexist, ImcH and CbcL, which appear to be tuned to the redox potential of different terminal electron acceptors [110,111]. The only common features in these systems is the presence of periplasmic cytochrome c or FeS binding proteins associated with, or fused to, membrane anchor subunits that facilitate electron transfer from the membrane bound quinol to the periplasmic acceptor (Fig 9A). Due to the nonhomologous nature of these electron transport systems in bacteria, we examined the ANME genomes for genes and gene clusters with a potential for analogous function but adapted to the specifics of archaeal cell biology (Fig 9B). ANME-2a, ANME-2b, and ANME-2c were found to encode a membrane-bound cytochrome b that is closely related to the membrane-bound subunit of the Vht hydrogenase (VhtC), which in methanogens mediates the reduction of methanophenazine with electrons from H 2 [112] (S11 Fig). In the genomes of ANME-2a and ANME-2b, this cytochrome b is followed by a multiheme c-type cytochrome (MHC) containing between 7 and 11 heme-binding motifs (CxxCH), instead of the vhtAG genes encoding the hydrogenase subunits found in Methanosarcina. In ANME-2c, a closely related homolog for this MHC protein was found elsewhere in the genome. This conspicuous gene clustering is not found in any methanogenic archaea, and the importance of this system is supported by the high transcript levels of the cytochrome b subunit reported in both ANME-2a and ANME-2c [9,11]. Phylogenetic analysis shows these VhtC homologs to be a closely related sister group to those found in Methanosarcina (S12 Fig). These gene clusters are the clearest examples of biological novelty well conserved in ANME that could explain the evolution of electron transfer capabilities, and we refer to them here as methanophenazine-cytochrome c oxidoreductase (Mco). Notably, ANME-1 and ANME-3 genomes did not contain homologs to Mco, so this does not appear to be a universal mechanism of electron transport in all ANME. ANME-1 lacked any identifiable cytochrome b, while the only ones apparent in ANME-3 were multiple copies of hdrE and the gene associated with the Rnf gene cluster mentioned above. These Rnf-associated cytochrome b in ANME-2a, ANME-2b, and ANME-3 are not found in any related methanogens. It is conceivable that these genes are involved in the oxidation of MpH 2 and in the transfer of these electrons onto the Rnf-associated cytochrome c for the purpose of EET. Two recent studies on M. acetivorans have implicated the Rnf-associated cytochrome c in electron transfer to Fe(III) or the artificial electron acceptor AQDS [113,114]. Such a process may be occurring in the ANME-SRB syntrophy via a similar mechanism and explain previous results of marine ANME utilizing AQDS, Fe(III), and Mn(IV) [115,116]. Another possible MpH 2 /QH 2 oxidizing systems could be represented by diverse homologs related to the membrane-bound periplasmic nitrate reductase subunit NapH, which are sporadically distributed across ANME genomes. NapH typically contains 4 transmembrane helices, 2 conserved cytoplasmic FeS clusters, and mediates electron flow from a menaquinol to the periplasmic nitrate reductase NapAB [117]. The NapH homologs in ANME are not found in suggestive gene clusters except in one of the ANME-2c genomes in which NapH was followed immediately by an 8 heme MHC, reminiscent of quinol:cytochrome c oxidoreductase gene clusters described in bacteria (Fig 9B). NapH homologs are found in both ANME-3, as well as 2 ANME-1, and are worthy of investigating further due to the lack of other obvious candidates for menaquinol/methanophenazine oxidation in those genomes. However, under standard laboratory AOM conditions [9], the NapH homolog found in ANME-2c E20 had very low transcript levels (S3 Data). Due to their uneven distribution in ANME genomes, these NapH homologs may have other nonessential functions. While no other gene clusters could be identified that seemed likely to mediate MpH 2 /QH 2 :cytochrome c oxidoreduction, the example of CymA in S. oneidensis MR-1 highlights how simple these systems can be while still being vitally important respiratory proteins and how easily one system can be functionally complemented by a completely unrelated one [107,109]. Numerous small multiheme cytochromes are encoded and transcribed in ANME genomes that either have single membrane anchors on their C-terminus or PGF-CTERM archaeosortase motifs that are predicted to covalently link them to membrane lipids. These small MHC proteins could potentially act in a similar fashion to CymA. However, without an associated or fused large membrane anchor that is homologous to previously characterized systems, it is difficult to implicate them directly in membrane-bound electron carrier oxidation through genomic evidence alone.

Multiheme cytochrome c protein abundance and transcript levels. ANME genomes contained many (5 to 8 times) more genes encoding MHCs than any of their methanogenic relatives (S13 Fig). Only the “Ca. Syntrophoarchaeum” and some members of the Archaeoglobales that are known to conduct EET contain a similar number of MHC genes among the archaea. Many homologous groups of these MHCs are only present in ANME and the aforementioned archaea. All MHC proteins are predicted to reside in the extracellular space as heme attachment to the cytochrome c apoprotein occurs here [118]. Representatives of these MHCs are among the highest transcribed proteins in multiple ANME lineages. Small MHC in ANME-2c E20, ANME-1 GB37, and ANME-1 GB60 were the 14th, 10th, and 18th highest transcribed proteins, respectively, and contain 8, 4, and 4 heme-binding motifs, respectively (S3 Data). One specific group of small MHC proteins has widespread distribution in the ANME, “Ca. Syntrophoarchaeum,” and some members of the Archaeoglobales, with no relatives in methanogenic archaea. Multiple paralogs exist in most ANME groups, and these include some of the highest transcribed proteins in ANME, many exceeding all methanogenesis pathway genes except Mcr (S14 Fig, S3 Data). As with the quinol oxidation step, the intermediate small soluble cytochromes c vary greatly between different EET-capable bacteria. Any number of these ANME cytochromes could be capable of carrying electrons between the cytoplasmic membrane and outermost layer of the cells. A small 6 heme-binding cytochrome c protein OmcS was recently shown to form the conductive extracellular “nanowires” in G. sulfurreducens that are thought to imbue the biofilms with conductive properties (Fig 9) [106,119]. No close homolog of OmcS were identified in ANME genomes, although any number of these relatively small cytochromes could be carrying out a similar function in the extracellular space between ANME and SRB, consistent with ultrastructural observations made with heme-reactive staining [9,87]. An 8-heme MHC in G. sulfurreducens known as OmcZ was shown to be required for optimal anodic current in biofilms grown on electrodes and is secreted in biofilms and in culture supernatant [120]. It was recently shown that OmcZ will also polymerize, forming conductive nanowires [121]. Very closely related homologs of OmcZ were found in multiple ANME-2a, ANME-2b, and ANME-2c genomes (Fig 10). The sequence similarity shared between OmcZ and the proteins found in these ANME groups is not limited to CxxCH motifs, but also extensive N- and C-terminal regions with many completely conserved residues. This level of sequence conservation is quite remarkable for homologs found in archaea and bacteria and suggests an important conserved function of these non-heme-binding domains, as well as a relatively recent interdomain horizontal transfer. The gene encoding the OmcZ homolog in ANME-2c E20 is the 14th highest transcribed gene in cultures grown under standard laboratory AOM conditions (S3 Data). In ANME-2a and ANME-2b, the genes encoding these OmcZ homologs are found next to the enormous MHCs described below. Investigations into the properties of ANME OmcZ homologs, specifically whether they can undergo the same polymerization observed in G. sulfurreducens, will be an important future area of research. PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 10. OmcZ homologs in ANME archaea. Protein sequence alignment of OmcZ homologs from various Geobacter species and ANME genomes reported here using muscle 3.8.31 with default settings. The 8 CxxCH-binding motifs are highlighted in gray. Regions of significant sequence identity are present throughout the protein, not just associated with the CxxCH motifs, suggesting conserved function. Alignment file can be found in S1 Data. ANME, anaerobic methanotrophic. https://doi.org/10.1371/journal.pbio.3001508.g010

S-layer conduits. ANME-2a, ANME-2b, ANME-2c, ANME-2d, and ANME-3 genomes also contain exceptionally large MHCs with 20 and 80 heme-binding sites (Fig 11, S13 Fig). Similarly large cytochromes in archaea are only found in Geoglobus and Ferroglobus [87]. Both of these groups are capable of extracellular iron reduction [122,123]. The largest and best conserved of these MHCs were classified into 3 major groups (ANME MHC Type A, B, and C) based on sequence identity, conserved domains, and heme-binding motif distribution (Fig 11). The most notable feature of ANME MHC-A and MHC-B is the presence of predicted S-layer domains, which commonly make up the outer proteinaceous shell surrounding the cytoplasmic membrane in many archaea [124]. ANME MHC-B additionally contain an N-terminal domain free of heme-binding motifs and annotated as “peptidase M6-like domains.” The presence of an S-layer domain in these cytochromes suggests their use for electron transfer through this outermost layer. In M. acetivorans, the S-layer domain structure has been determined by x-ray crystallography and contains 2 subdomains connected by a flexible linker [125]. In all ANME MHC with S-layer domains, this flexible linker also contains a single heme-binding motif, which would place a heme group within the plane of the S-layer. Both ANME MHC-A and MHC-B encode for C-terminal regions predicted to be transmembrane helices that may anchor them cytoplasmic membrane. We suggest that ANME MHC-A and MHC-B are functionally analogous to the MtrCAB complexes found in the outer membranes of the EET-capable gram-negative bacteria (Fig 9A) [104]. In ANME, an alternative mechanism needs to be employed since the outermost layer is proteinaceous. PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 11. Large multiheme cytochrome c proteins in ANME archaea. Schematic of protein structure highlighting the position of heme-binding motifs and other conserved features of the large ANME-specific multiheme cytochromes. The ANME MHC were divided into 3 major groups based on sequence similarity and conserved domains structure. ANME MHC-A contains an S-layer domain and C-terminal transmembrane helix. In ANME-2a and ANME-2b, these proteins are extended by an additional transmembrane helix and more heme-binding motifs. ANME MHC-B contains an S-layer domain and C-terminal transmembrane helix as well, but an N-terminal region devoid of heme-binding domains has similarity to peptidase M6-like domains. ANME MHC-C do not contain S-layers or C-terminal transmembrane helices but instead contain a large N-terminal region with a predicted pectin lyase-type domain. Domains predicted with InterProScan and are displayed with colored boxes. Large MHC proteins from ANME-3 sp. HMMV2 and ANME-2c sp. ERB4 that do not clearly fit into these categories are also shown (Note: ANME-2c sp. ERB4 is a single-peptide split between 2 lines due to its size). ANME, anaerobic methanotrophic; MHC, multiheme c-type cytochrome. https://doi.org/10.1371/journal.pbio.3001508.g011 ANME MHC type C do not encode for an S-layer domain or C-terminal transmembrane helices but do encode for a large N-terminal “pectin lyase-like domain.” It is difficult to predict the function of this additional domain. However, it is interesting to note that the pectin lyase fold typically occurs in proteins that attach to and/or degrade carbohydrates and has been found in bacteriophage tail spike proteins for attachment to hosts [126]. It is possible then that this domain is involved in recognizing the outer cell wall of partner bacteria. Some ANME genomes contained exceptionally large MHC that did not fall clearly into these 3 categories. ANME-3 HMMV2, for example, contained a single polypeptide containing C-terminal features of ANME MHC-C and N-terminal features of ANME MHC-B (Fig 11). A fosmid assigned to ANME-2c ERB4 encoded the largest MHC in our dataset containing an S-layer domain, a pectin lyase-like domain, and 86 heme binding motifs. In actively growing ANME-2c cultures, genes encoding these large ANME MHC were transcribed at lower levels than the enzymes of the reverse methanogenesis pathway and some of the smaller MHCs leading the authors to conclude that these were of little importance to electron transfer in this culture [9]. If 2 proteins are expected to carry out the same function in a pathway, transcription levels may be useful in determining which may be the dominant one operating under certain conditions. But this is not the case for the different classes of MHCs in the model of ANME EET presented here. The small MHCs are thought to serve as intermediates between the inner membrane and S-layer, as well as assist in conveying conductivity to the extracellular matrix, while the large S-layer proteins specifically provide a conduit through the S-layer. The small MHCs may cover longer distances and would therefore be expected to be numerically dominant but still could not make a functional EET system if no specific mechanisms existed for the short transfer through the S-layer to the extracellular environment. This is exactly the scenario observed in the electrogenic model bacterium G. sulfurreducens that requires MHC conduits through the outer membrane for EET formed by MtrCAB and other analogous porin-cytochrome systems [104]. While absolutely necessary for efficient EET, these genes show lower transcript levels than those of some of the smaller soluble MHCs encoded in the genome [127]. We therefore find this previously reported transcript information to be consistent with the model of ANME-2 and ANME-3 EET presented here in which the S-layer fusions play a key role (Fig 9B). ANME-1 contained fewer MHCs and at most had 10 predicted heme-binding motifs, and these proteins did not contain any identifiable S-layer domains. From the available genomic data, it is unclear how these cytochromes fit into the cell structure of ANME-1. It was recently suggested that ANME-1 lack an S-layer due to the absence of predicted S-layer domain containing proteins [42]. However, the same analysis found no Archaeoglobus proteins with S-layer domains, yet the Archaeoglobus S-layer has been visualized and extensively characterized [128]. The S-layer domain in the large MHC from ANME-2 and ANME-3 are only recognizable because they are homologous to the S-layer protein of M. acetivorans that was recently crystalized [125]. Before this study, the S-layer domain was only identified as a domain of unknown function (DUF1608). The ANME-1 ultrastructure is usually cylindrical and highly reminiscent of Methanospirillum, which contains S-layers [129]. How the MHC in ANME-1 fit into this cell structure will require further investigation. Gram-positive bacteria are also capable of EET and in some cases have been found to use more modestly sized MHC proteins (6 to 9 heme-binding motifs) for electron transport through the outermost cell wall [130]. Such a model could very well apply to the cell wall and cytochromes encoded in the ANME-1 genomes.

[END]

[1] Url: https://journals.plos.org/plosbiology/article?id=10.1371/journal.pbio.3001508

(C) Plos One. "Accelerating the publication of peer-reviewed science."
Licensed under Creative Commons Attribution (CC BY 4.0)
URL: https://creativecommons.org/licenses/by/4.0/


via Magical.Fish Gopher News Feeds:
gopher://magical.fish/1/feeds/news/plosone/