Single-Cell Genome and Group-Specific dsrAB Sequencing Implicate Marine Members of the Class Dehalococcoidia (Phylum Chloroflexi) in Sulfur Cycling

Phylogeny and distribution of single-cell DEH-C11 in Aarhus Bay sediments.Single cells from marine sediments (10 cm below sea floor [cmbsf]) of Aarhus Bay, Denmark, were isolated by flow cytometry. After multiple-displacement amplification (MDA) and 16S rRNA gene sequencing, DNA from a single bacterial cell (designated “DEH-C11”) that was divergent from cultured members of the DEH, single amplified genomes (SAGs), and metagenomic bins was selected for deep sequencing and analyses of gene content. The cell affiliated with the previously defined “DSC-GIF3-B” subgroup of the DEH (Fig. 1) (16), which also falls within the “Dehalococcoides sister clades” (17). Sequences belonging to these Dehalococcoides sister clades are among the most prevalent DEH sequences detected in marine sediments (16). The 16S rRNA sequence of DEH-C11 is 88% similar to that of Dehalogenimonas lykanthroporepellens strain BL-DC-9, the most closely related cultivated strain, which approximately equates to family-level divergence (21). It is also highly divergent from previously described single-cell genomes (18, 19) or metagenome-derived genomes (20), which affiliate with completely different clusters of the DEH, based on 16S rRNA phylogeny (Fig. 1). Although the metagenomic bin of RBG-1351 does not contain a 16S rRNA gene (20), our comparisons of bidirectional BLASTP hits of all predicted proteins revealed that only 49% of predicted proteins from DEH-C11 had homologs in the genome of RBG-1351 and that, of these, an average protein sequence identity of only 55% was determined. This demonstrates that the degree to which the DEH-C11 genome is distinct from those of other DEH is considerable, and this is reflected in its metabolic potential compared to that of other DEH (described below). DEH subgroup DSC-GIF3 accounted for 4% of the 16S rRNA genes recovered by PCR amplification with DEH-targeted primers from all depths (10 to 300 cmbsf) of the Aarhus Bay site from which the cell had been isolated (16).

• Open in new tab
• Download powerpoint

FIG 1 

Phylogenetic analysis of the 16S rRNA gene from DEH-C11 in comparison to those of other members of the class DEH. The tree is based on the maximum likelihood algorithm. Branches in orange highlight the DSC-GIF3-B subgroup, to which the 16S rRNA gene of DEH-C11 (leaf label highlighted in red) is affiliated. Branches in green highlight the clade containing all known organohalide-respiring phylotypes, and leaf labels of sequences from cultivated bacteria or sequences implicated in organohalide respiration are highlighted in green. Branches in blue highlight the GIF9-A clade. Leaf labels in blue highlight sequences for which genomic information is available (18–20). The scale bar represents 5% sequence divergence. PCB, polychlorinated biphenyl; PCE, tetrachloroethene; SIP, stable isotope probing.

Genomic data obtained from single-cell DEH-C11.All reads from sequencing of MDA-derived DNA were assembled into 233 contigs of greater than 1,000 bp. After removal of contigs that did not contain genes with highest similarity to genes of Chloroflexi, a “high-confidence” assembly of 0.94 Mbp within 155 contigs was obtained with an average length of 6.1 kbp (Table 1). Pentanucleotide analysis of the nucleotide coding signatures of contigs showed that all but three contigs from the high-confidence DEH-C11 data set clustered together (see Fig. S1 in the supplemental material), further corroborating that the data set originated from a single genome. The high-confidence data set represented approximately 42% to 46.8% of the estimated full genome of 2.01 to 2.24 Mbp (Table 1). This is considerably larger than the streamlined genomes of the metabolically specialized, cultivated genera Dehalococcoides and Dehalogenimonas, which range from 1.34 to 1.69 Mbp (22, 23). This gave a first hint that the DEH-C11 organism may encode expanded metabolic capabilities compared to its closest cultivated relatives.

Sulfite reduction as a potential mode of energy conservation in Chloroflexi. (i) Dissimilatory sulfite reductase in the genome of DEH-C11.A cluster of genes encoding a dissimilatory sulfite reductase (Dsr) complex and accessory subunits was identified in an assembled part of the DEH-C11 genome (Fig. 2 and 3). Two lines of evidence indicate that this gene cluster genuinely originated from the DEH-C11 genome. First, genes most closely related to Dehalogenimonas lykanthroporepellens BL-DC-9 genes flank the dsr cluster (Fig. 3). Second, pentanucleotide signature analysis shows that the dsr-harboring contig clusters closely with all other DEH-C11 contigs (see Fig. S1 in the supplemental material). The cluster contains the genes dsrABCMK, which encode a full complement of the Dsr subunits required for the reduction of sulfite (24) (Fig. 2 and 3; see also Table S1 in the supplemental material). Two additional genes, dsrD and dsrN, commonly present in dsr operons and encoding proteins not directly involved in electron shuttling during sulfite reduction, were also present in the cluster. The genes dsrD and dsrN encode a putative transcriptional regulator and an enzyme involved in the amidation of the sirohemes of DsrAB, respectively. No genes encoding DsrJOP proteins, which are typically found in Gram-negative but not in Gram-positive sulfate reducers, were present. Instead, Gram-positive bacteria are thought to use DsrMK only to transfer electrons from the menaquinone pool to DsrC (24). The presence of conserved cysteine residues required for catalytic sites in the encoded DsrAB and DsrC subunits suggests that the encoded enzymes were functional.

• Open in new tab
• Download powerpoint

FIG 2 

A simplified schematic of putative biochemical properties predicted from genomic content of DEH-C11. The redox cycling schematic for sulfite reduction is based on model proposed by Venceslau et al. (24). Aromatic substrates depicted are (left to right) phenylalanine, phenol, and pyrogallol. The heterocyclic compound depicted is creatinine. Nuo, NADH:ubiquinone oxidoreductase complex; TRAP, tripartite ATP-independent periplasmic transporters; DMSO, dimethyl sulfoxide; DMS, dimethyl sulfide; Dsr, dissimilatory sulfite reductase; S-layer, surface-layer protein coat.

• Open in new tab
• Download powerpoint

FIG 3 

Representation of the gene order present on the contig (IDBA scaffold 11) containing the dissimilatory sulfite reductase operon and associated genes from DEH-C11 (top) and D. lykanthroporepellens BL-DC-9 (bottom). Gene names in blue denote that the best BLASTP hits were to D. lykanthroporepellens BL-DC-9 or other Chloroflexi strains. Blue shaded lines between the two gene order representations show regions with high sequence similarity and synteny determined by tBLASTx as implemented using EasyFig (93). Degrees of tBLASTx sequence identity are depicted in the colored legend. The red blocks at the end of the DEH-C11 contig representation denote the ends of the contig sequence. The red and blue lines within the genes of the DEH-C11 represent the regions of the genome amplified by long-range PCR.

A dissimilatory direction was predicted by the presence of a gene for DsrD, which appears to be specific for sulfite-reducing organisms, since it is absent in all known sulfur-oxidizing organisms that use Dsr in the reverse direction (25). A reductive direction is also supported by the phylogenetic placement of the DsrAB in the reductive branch of the DsrAB phylogenetic tree (described further below). The presence of a dsr cluster with all genes required for a functional complex, as well as conserved amino acid residues in the key enzymes indicative of active enzymes, provided evidence that DEH-C11 could have used sulfite for anaerobic respiration. Sulfite could be utilized directly, since it is a common “intermediate” of the sulfur cycle (26), or could be enzymatically derived from sulfate, thiosulfate, or sulfonated organics. Genes for enzymatic transformations from sulfate (i.e., activation via adenosine 5′-phosphosulfate) were not identified in the assembly or in unassembled reads, and no genes indicating that sulfite could be derived from thiosulfate or sulfonated organics were identified in the assembled genome. Additional sequencing of related genomes will be required in order to resolve the issue regarding the source of sulfite for these organisms. Although dsrAB genes have been identified in syntrophic Desulfotomaculum strain MGP (27) and thiosulfate-disproportionating Desulfocapsa sulfexigens strain SB164P1 (Desulfobulbaceae) (28), which appear to have recently lost the ability to respire sulfate, these strains are rare among their relatives. That is, most known members of the genus Desulfotomaculum and the family Desulfobulbaceae have the ability to reduce sulfate. It is therefore plausible that even if DEH-C11 were not capable of sulfite reduction due to a recent change in lifestyle, other relatives of the organism might still be capable of sulfite reduction.

Detection of dsr genes in DEH bacteria indicates a potential for sulfite reduction in members of the phylum Chloroflexi and suggests a previously unexpected role for these bacteria in marine sulfur cycling. It may also provide hints with regard to the widespread nature of DEH within marine environments, since they are rich in sulfur compounds. A recent study reported the presence of genes for DsrAB in a metagenome-derived genome of an Anaerolineae (Chloroflexi) organism and suggested these organisms might therefore be capable of sulfate reduction (29). Our examination of the encoded proteins for comparative purposes identified these as homologs of small iron-sulfur-containing ferredoxin-like proteins, most similar to proteins from methanogens, while no genes for bona fide DsrAB and DsrC (key catalytic enzymes required for sulfite reduction) were found. We therefore conclude that those Anaerolineae-affiliated organisms may not have the genetic potential for dissimilatory sulfite or sulfate reduction.

Because the reduction of oxidized sulfur compounds drives the oxidation of vast amounts of organic matter in marine sediments on a global scale (30, 31), a contribution to this process by widespread DEH could be biogeochemically significant. The abundance of DEH-related dsrAB relative to total dsrAB in marine sediments could not be estimated reliably, since previously used dsrAB primers contained up to 6 mismatches to the dsrAB of DEH-C11 (32, 33). This would therefore lead to severe underestimation of the relative abundances of DEH-related dsrAB genes. Future studies using updated primer sets targeting dsrAB or metagenomic data are required to gain reliable estimates of the relative abundances of DEH-related dsrAB genes and of their potential contribution to sulfur cycling in marine sediments.

Phylogenetic analysis of the DsrAB sequence of DEH-C11 showed that it is a member of a sister clade to the recently denominated “uncultured family-level lineage 3” (Fig. 4) (34). The DEH-C11-related clade, along with uncultured family-level lineages 2, 3 and 4, is composed almost entirely of DsrAB sequences retrieved by PCR-based approaches from unidentified microorganisms inhabiting marine sediments (32, 33, 35, 36). The aforementioned clades are also part of the broadly defined “Firmicutes-related” sensu lato group (34).

• Open in new tab
• Download powerpoint

FIG 4 

Phylogenetic analysis of DsrAB from DEH-C11 (highlighted in red) and cloned DsrAB sequences as determined by the evolutionary placement algorithm (88), which was used to place sequences onto a previously constructed DsrAB consensus tree (34). Orange leaf labels indicate sequences retrieved by long-range PCR in this study. L.A.-Firmicutes, DsrAB proteins that were laterally acquired by certain Firmicutes species of the Deltaproteobacteria (34). The red dots on leaf labels indicate that the DsrAB sequences are derived from amplified sequences that were sequenced by primer walking and with pentanucleotide signatures similar to that of the DEH-C11 genome (see Fig. S1 in the supplemental material). The grey scale bar represents 5% sequence divergence.

The dsr operon in DEH-C11 is flanked by genes encoding proteins involved in the synthesis of a siroheme cofactor required by DSR for catalytic activity, i.e., genes stretching from the uroporphyrinogen-III methyltransferase/synthase gene to the siroheme synthase gene (Fig. 3) (37). Intriguingly, predicted protein sequences for synthesis of the siroheme cofactor are most similar to homologs from Dehalogenimonas lykanthroporepellens BL-DC-9, a cultivated member of DEH (Fig. 3). This finding therefore opens an evolutionary question as to whether Dehalogenimonas and other related DEH previously harbored the genetic capacity for sulfite reduction. The lack of other genes required for sulfite reduction in Dehalogenimonas lykanthroporepellens BL-DC-9 might be the result of genomic streamlining due to a more contemporary specialization for organohalide respiration.

(ii) Amplification of putative DEH-related dsrAB from marine sediments.To investigate if dsr genes can be found in other DEH-related bacteria in marine sediments and to explore further the diversity of dsr genes in DEH, we designed degenerate primer sets to directly amplify DEH-related dsr loci from DNA extracted from different sediment sites. The aim was to amplify “long” (up to 3.9 kbp) DNA sequences containing novel DEH-C11-related dsrAB sequences and flanking DEH-related “marker” genes in order to support the phylogenetic affiliation of the amplified dsrAB sequences. As flanking marker genes, a gene for a siroheme cofactor synthesis enzyme and a gene for a LuxR-related “transcriptional regulator,” both with highest sequence similarity to genes in Dehalogenimonas spp. (Fig. 3), were targeted. With primers targeting luxR and dsrB, a 2.5-kbp fragment (Fig. 3) was successfully amplified from sediments of Aarhus Bay. The retrieved clone sequences were closely related to dsrAB of DEH-C11, showing nucleotide sequence identities of 97% to 100% (Fig. 4; see also Table S2 in the supplemental material). Although no products were obtained in attempts to amplify fragments with genes for siroheme cofactor synthesis enzymes and dsrAB, two alternative degenerate primer pairs targeting dsrA and dsrC that amplify dsrABDNC (Fig. 3) proved to successfully amplify ~3.9-kbp fragments and enabled recovery of the more divergent DEH-C11-related dsrAB sequences (Fig. 4). Sixteen end-sequenced clones obtained from different sediment depths of Aarhus Bay at the location where DEH-C11 was isolated, as well as from the Baffin Bay, Greenland and from tidal flat sediments of the Wadden Sea, Germany (Fig. 4), featured dsrA nucleotide sequences that were unique; i.e., they were not found in dsrA of DEH-C11. Only two sediment samples that were tested failed to provide amplicons, namely, sediments from the Peru margin and from the Black Sea. Phylogenetic placement of DsrA or DsrB sequences onto a previously constructed consensus reference tree (34) by the evolutionary placement algorithm (EPA), as well as by de novo constructed trees based on partial DsrA sequences using both neighbor-joining and maximum likelihood algorithms (results not shown), revealed that eight sequences formed a monophyletic clade with the DsrAB of DEH-C11 (Fig. 4). Based on the distinct monophyletic clustering of the eight clone sequences with DsrA of DEH-C11 compared to other stable family-level lineages as shown by using all three approaches and on the fact that the best BLASTX hits of corresponding dsrN sequences were to genes encoding the DsrN of DEH-C11 (see Table S1), we infer that these clones are likely derived from DEH organisms. Six clones from the monophyletic DEH-C11-related clade were also further sequenced by primer walking to obtain sufficient sequence information for nucleotide signature analysis. These clones showed pentanucleotide signatures very similar to those of DEH-C11 (see Fig. S1), suggesting DEH-related bacteria as their sources.

Some of the obtained dsrAB sequences of the monophyletic DEH-C11-related clade shared as little as 76% nucleotide sequence identity with the dsrAB of DEH-C11, while minimum intragroup sequence identities of 73% for nucleotide sequences and 86% for amino acid sequences were determined (see Table S1 in the supplemental material). The minimum intragroup sequence identity determined here is well within the range of intrafamily amino acid sequence identities of known families of sulfate-reducing microorganisms, which range from 64% to 89% (34, 38). On the basis of previously constructed linear regression plots of corresponding pairs of 16S rRNA and dsrAB genes derived from many sequenced isolates (34), we deduced that the organisms harboring the cloned DEH-related dsrAB genes with 76% nucleotide sequence identity to DEH-C11 may harbor 16S rRNA genes with less than 94.5% identity (see Fig. S2). This is below a recently defined genus-level taxonomic threshold (21). It could therefore be inferred that the organisms harboring these genomic fragments belong to different genera, and we propose that a family-level clade within the DEH exists that harbors genes for Dsr. No evidence for DEH-related dsr could be identified in the metagenomes searched. Ongoing research aims to uncover the diversity of dsr in DEH and to determine whether other genes can be identified that may indicate the source of intracellular sulfite.

Other respiratory modes inferred from genome annotations. (i) Molybdenum-containing oxidoreductases.Four genes encoding catalytic alpha subunits of complex iron-sulfur molybdoenzyme (CISM) oxidoreductases (39) that may function as terminal reductases were present in the genome of DEH-C11. Phylogenetic analysis showed that two CISM alpha subunits from DEH-C11 affiliated with known dimethyl sulfoxide (DMSO)-reducing enzymes and branched closest to homologs from another DEH single-cell genome, that of DEH-J10 (see Fig. S3 in the supplemental material) (19). DMSO and trimethylamine N-oxide (TMAO) were previously discussed as candidate substrates for these predicted enzymes (19), although experimental validation is needed. One of the predicted CISM alpha subunits (DEH-C11.384) had no close phylogenetic affiliation, while another (DEH-C11.399) was affiliated with the broad arsenate/polysulfide/thiosulfate oxidoreductase branch of the CISM phylogeny. A putative role as an arsenate reductase alpha subunit was inferred by various lines of evidence, including the presence of a twin-arginine transport (TAT) peptide “leader” signal sequence predicted to guide the subunit to or across the cytoplasmic membrane, which thereby distinguishes it from the cytoplasmic arsenate reductases typically used for resistance to arsenic toxicity (40, 41). Additionally, colocalization of genes encoding an arsenical resistance operon repressor (ArsR family), as well as those encoding an adenosylmethionine-dependent methyltransferase and an arsenite efflux transporter, which may act to methylate and extrude the reduced and toxic As(III) from the cell, strongly suggests that this enzyme could reduce arsenate (42) (Fig. 2). Further genes encoding homologs of formate dehydrogenase-like (FDH) beta and gamma subunits were present in adjacent locations and may function to transfer electrons and embed the complex into the membrane, respectively, if analogous to other known three-subunit CISM complexes (43) (Fig. 2). Although arsenic resistance genes were previously identified in other DEH genome data, i.e., the RBG-2 and RBG-1351 metagenome-derived genomes (20), this is to our knowledge the first report of a potential respiratory arsenate reductase in DEH and also in the phylum Chloroflexi.

Although very little is known about the existence and activity of microorganisms involved in arsenic cycling in marine sediments, various arsenic species are naturally prevalent in marine waters and sediments, even in marine environments that are free from direct anthropogenic impact (44). Recent metagenomic analyses have identified genes encoding enzymes predicted to catalyze transformations in the arsenic cycle, including arsenate reduction, in various marine sediments (45, 46). Our results therefore indicate that so-far-overlooked arsenic cycling microorganisms occur in marine sediments and suggest that microorganisms capable of performing such transformations fill a functional niche in marine sediments.

(ii) Reductive dehalogenase homologous genes.A single gene encoding a reductive dehalogenase subunit A homolog (RdhA) was identified in the DEH-C11 genomic content. The predicted enzyme contains conserved cysteine residues required for two Fe-S cluster binding motifs typical of known RdhA enzymes, suggesting that the encoded enzyme was functional. No TAT peptide leader signal sequence required for protein translocation across the cytoplasmic membrane was detected. Furthermore, no genes were identified for predicted membrane-anchoring RdhB subunits, which are often encoded by sequences adjacent to rdhA. Other rdhA genes that do not encode TAT sequences and without adjacently encoded RdhB have been previously found, e.g., in D. lykanthroporepellens BL-DC-9 (23) and in the aerotolerant nonrespiratory reductive dehalogenase of Nitratireductor pacificus pht-3B (NprdhA) (47).

The DEH-C11 RdhA sequence has low sequence similarity to known RdhA enzymes used for respiration by cultivated DEH organisms. Instead, it is most similar to a homolog derived from Desulfobacula toluolica Tol2, an aromatic-compound-degrading sulfate reducer isolated from marine sediments, which to our knowledge has not been tested for growth via organohalide respiration (48). The RdhA sequence in strain Tol2 also does not contain a TAT leader peptide and has no adjacently encoded RdhB. Phylogenetic analysis shows the DEH-C11 RdhA forms a clade with (i) the RdhA of D. toluolica Tol2, (ii) clade “Smkt-3” RdhA from metagenomic contigs of unknown phylogenetic origin derived from marine sediments of Shimokita Peninsula, Japan (49); (iii) RdhA sequences from Chloroflexi-binned metagenomic scaffolds from a terrestrial aquifer sediment-derived metagenome (20); (iv) a single RdhA from a contig (that we deduced is derived from a DEH organism) from a metagenome of deep subsurface sediments of the Canterbury Basin, New Zealand (50); and (v) RdhA sequences from estuary sediment-derived metagenomic contigs (51), which we identified as belonging to DEH or Chloroflexi in this study (Fig. 5). The clade of mostly DEH-derived RdhA is distinct from previously described clades of RdhA derived from DEH such as the “conserved sytenic” RdhA of Dehalococcoides spp., as well as the diverse “Dehalococcoidales” RdhA clade (52) (Fig. 5). In 8 of the 12 rdhA-containing metagenomic contigs analyzed, we observed the gene for tRNA-His-GTG (and often also the gene for tRNA-Arg-TCG) adjacent to rdhA and yet observed few other syntenic genes, hinting at only a small degree of conservation among the genomes (see Fig. S4 in the supplemental material). Only two RdhA sequences from the estuary sediment-derived metagenomic contigs (51), which were derived from nearly identical contigs, affiliated with the Dehalococcoidales RdhA clade (Fig. 5). The DEH-C11-related RdhA clade is a sister clade to RdhA derived from various aerobic marine bacteria and includes strains such as Comamonas sp. strain 7D-2 and N. pacificus pht-3B, which have been described to contain nonrespiratory reductive dehalogenases (47, 53). Additionally, a MarR-type transcriptional regulator was encoded by a gene directly adjacent to rdhA in DEH-C11, and the gene was most similar in sequence to MarR regulator genes from cultivated DEH that are typically associated with regulation of rdhA expression (22).

• Open in new tab
• Download powerpoint

FIG 5 

Phylogenetic analysis of RdhA derived from DEH-C11 (highlighted in red), DEH-derived RdhA retrieved from metagenomes, and RdhA from pure reference strains. The tree is based on the maximum-likelihood algorithm. Branches of metagenome-derived RdhA sequences determined to originate from DEH- or Chloroflexi-related contigs are highlighted in orange (50, 51, 74). Branches corresponding to the “conserved syntenic” Dehalococcoides species RdhA are highlighted in dark blue, while the Dehalococcoidales RdhA clade are highlighted in light blue. Nearly full-length RdhA sequences from the Smkt-3 clade (two of six available were nearly full length) from subsurface sediments of Shimokita peninsular were included (49). Branches of RdhA derived from aerobic bacteria are highlighted in purple. The grey scale bar represents 20% sequence divergence. WOR, White Oak River.

Together, the results indicate that the RdhA in DEH-C11 was located in the cytoplasm and was therefore not likely involved in a membrane-bound respiratory chain. Possible alternative functions include the reoxidation of respiratory cofactors for a “facilitated fermentation,” detoxification of substrates, or the removal of halogens from organics to enable further catabolism. Dehalogenation of substrates to enable further catabolism was recently described for the aerobic strains Comamonas sp. strain 7D-2 and N. pacificus pht-3B, which dehalogenate aromatic compounds prior to further catabolic processing (47, 53). The phylogenetic relatedness of the RdhA of DEH-C11 to the RdhA of Comamonas sp. strain 7D-2 and N. pacificus pht-3B could suggest similar functions for these enzymes in anaerobes, although such functions will need to be experimentally verified. The apparent high frequency of rdhA (49) and of haloacid dehalogenase genes (50) in the genetic content of marine subsurface microbes suggests that there is an evolutionary advantage to possessing the capacity to dehalogenate organics in marine subsurface environments.

(iii) NADH-ubiquinone oxidoreductase.The DEH-C11 bacterium encoded an 11-subunit NADH-ubiquinone oxidoreductase (nuo), while it lacked genes for a cytoplasmic electron input “N-module," i.e., NuoEFG subunits. This is analogous to genomes of cultured DEH that encode similar 11-subunit Nuo complexes but that were hypothesized to function in combination with an archaeon-like cofactor F420 input module (FpoF) (54). The predicted complex may play a role in anaerobic respiration by transferring reducing equivalents from cytosolic oxidation reactions to an electron transport chain that is connected with proton translocation. The Nuo complex was incomplete in the genome recovery of the metagenome-derived genome of RBG-1351, which gave the first described genomic information derived from the Dehalococcoides sister clades (20). In combination with a lack of genes for terminal respiratory reductases in RBG-1351, this led the authors to infer that the organism was likely an obligate fermentative organism and that fermentative metabolism may be widespread among the DEH, since genomic content from other DEH clades (i.e., GIF-9) also suggested fermentative lifestyles (19, 20). In a further search, we identified several other full nuo gene clusters from the previously generated Rifle Aquifer metagenome (20) that were most similar in protein sequence identity to those of cultured DEH, single-cell DEH-C11, and metagenome-derived genome RBG-1351 (results not shown). This similarity was exemplified by phylogenetic analysis of corresponding NuoL subunits (the largest subunit of the Nuo complex) from the full nuo gene clusters from the metagenome and reference Chloroflexi (see Fig. S5 in the supplemental material). Together, these results provide evidence that members of the Dehalococcoides sister clades are not limited to fermentative modes of energy conservation.

Electron donors. (i) Catabolism of homocyclic aromatics.Multiple genes provided evidence that the DEH-C11 bacterium had the capacity to oxidize aromatic organic molecules with various substituents. These include three copies of genes encoding pyrogallol hydroxytransferases similar in amino acid sequence to predicted homologs from bacteria known to degrade hydroxylated aromatic molecules (see Fig. S3 in the supplemental material) (55, 56). Pyrogallol hydroxytransferases could enable the bacterium to perform transhydroxylation or hydroxylation reactions required for catabolism of aromatic molecules with metapositioned hydroxyl groups (57). Genes encoding pyrogallol hydroxytransferases were also previously identified in the metagenome-derived genomes of RBG-1351 and RBG-2 (20).

Genes required for catabolism of aromatics without metapositioned hydroxyl groups were also identified. These included genes encoding two subunits of a putative benzylsuccinate coenzyme A (CoA) transferase (DEH-C11.996 and DEH-C11.997), which have also been identified in DEH-J10 and RBG-1351. This indicates the DEH-C11 cell could activate aromatics to arylcarboxyl-CoA esters, which could be further catabolized via a central ring reduction and cleavage pathway. Downstream were two genes encoding alpha and delta subunits (DEH-C11.982 and DEH-C11.983) of a phenylphosphate carboxylase-like enzyme, which is one of the key enzymes for activation of phenol to benzoyl-CoA. Further genes identified that may be related to enzymes catalyzing steps of phenol degradation include candidate genes for a putative 4-hydroxybenzoyl-CoA reductase (A, B, and C subunits), although this enzyme is also known to function in analogous steps in degradation pathways of other substituted aromatic molecules (58). Genes encoding subunits of a phenylglyoxylate:acceptor oxidoreductase (DEH-C11.203-205) and a phenylacetyl-CoA:acceptor oxidoreductase (not present in the high-confidence data set) indicate that an anaerobic phenylacetate degradation pathway exists which could facilitate the degradation of the aromatic amino acid phenylalanine. Degradation of phenylalanine appears to be common among aromatic compound-degrading anaerobes (58), since the carbon can be funneled into the benzoyl-CoA degradation pathway used for the final breakdown of aromatic molecules.

An additional indication of the uptake and catabolism of aromatic molecules was found after examination of BLASTP results that revealed numerous gene clusters for tripartite ATP-independent periplasmic (TRAP) transporters with highest amino acid sequence identities to encoded proteins from well-characterized aromatic compound-degrading bacteria. For instance, five of eight TRAP “solute-binding” P subunits were most similar to protein sequences from the polyaromatic-degrading deltaproteobacterial strain NaphS2 (59, 60). We interpret this as an indication that some of the TRAP transporters of DEH-C11 could allow specific uptake of aromatic substrates from the environment and thereby feed these substrates into the aromatic catabolic pathways described above.

The emerging finding of various genes encoding key enzymes of aromatic-molecule-catabolizing pathways among multiple DEH genomes indicates that catabolism of aromatic molecules may be a common metabolic route for carbon and energy acquisition in various DEH organisms and therefore may be ecologically and biogeochemically important. Sources of aromatic compounds in marine sediments include aromatic amino acids, aromatic fermentation products (e.g., benzoate), or hydrocarbon seeps or spills. Additionally, evidence suggests that a large fraction of the dissolved organic matter pool in the oceans is comprised of diverse aromatic molecules (61), which may be derived from sources such as plant lignin (62, 63) or bacterium-derived metabolites (64). These compounds likely become buried in marine sediments when associated with sinking marine particles. Indeed, recent organic geochemical analyses of subsurface marine sediments have shown that relatively high proportions of lignin-derived aromatic compounds, as well as other aromatic molecules formed by diagenetic processes or of unknown sources, are present in sediments of the Namibian margin (65) or tidal flats of Helgoland, North Sea (66). Diverse aromatic substrates may therefore sustain specialized microorganisms such as the DEH in marine sediments and, importantly, could serve as promising substrates for postgenomic experimentation.

(ii) Alternative carbon and energy sources.Several genes that were unique to DEH-C11 versus cultivated DEH gave results that indicate that the capacity to cleave and catabolize different heterocyclic aromatic compounds exists in DEH-C11. These include a gene encoding a putative “creatinine amidohydrolase,” which is known to catalyze hydrolytic cleavage of amide bonds in the heterocyclic compound creatinine, as well as additional genes encoding subunits of a hydantoinase family enzyme (cyclic amidohydrolases) that were detected on contigs not included in the high-confidence data set. Heterocyclic organics that are dearomatized by such enzymes can be converted to central intermediates or directly to amino acids and can also be fermented by some organisms. Additionally, we identified seven predicted proteins with high sequence similarity and conserved domains of the amidohydrolase family that are unique to DEH-C11 versus cultured DEH. Although it is difficult to predict their precise functions, they gave indications of an expanded capacity to hydrolyze organic molecules. The potential to utilize acetone or other short-chain ketones, e.g., butanone, was indicated by genes encoding homologs of alpha and gamma subunits of acetone carboxylases. Gamma subunits of acetone carboxylases are considered to differentiate them from related hydantoinase-type enzymes (67), and phylogenetic analysis also showed that the alpha subunit from DEH-C11 affiliates with known acetone carboxylase-harboring bacteria (see Fig. S6 in the supplemental material). These genes have also been previously identified in RBG-1351 (20). Acetone can be a fermentation product and could therefore be made available by fermenting comembers of the microbial community.

Cellular functions (cell membrane).No indications related to peptidoglycan synthesis were found in the DEH-C11 genome content, which is in line with previous studies of DEH isolates and their genomes (22, 54, 68, 69). Various genes encoding enzymes required for the synthesis and modification of glycoproteins, which we hypothesize may be conjugated to surface-layer proteins, were present on a single scaffold (scaffold_7), e.g., genes encoding UDP-glucose 4-epimerase, GDP-l-fucose synthase, GDP-mannose 4,6-dehydratase, and various glycosyltransferases with sequence identities similar to those of various surface-layer-containing archaea. In many surface-layer-containing organisms, such surface-attached glycoprotein chains can have a capsulation effect and may therefore provide extra stability to the cell wall. Such information regarding cell walls could give insights into the ability of DEH to survive in deep marine sediments and into why DEH and Chloroflexi are apparently resistant to various cell disruption methods (70) and alkaline lysis procedures in single-cell genome-based studies (18).