The 16S rRNA gene sequences of the candidate phylum OP1 were originally revealed in the sediments of the Obsidian Pool hot spring (75‒95°C), Yellowstone National Park, United States (Hugenholtz et al., 1998). This is one of the first 12 candidate phyla described as a results of molecular analysis of microbial diversity. The 16S rRNA gene sequences assigned to this candidate phylum were subsequently detected in various terrestrial and marine ecosystems, including deep-sea hydrotherms (Teske et al., 2002; Stott et al., 2008; Kato et al., 2009), hot springs (Costa et al., 2009; Tobler and Benning, 2011), anaerobic bioreactors (Nobu et al., 2015), and groundwater (Hirayama et al., 2005). In spite of its wide occurrence in nature, the share of OP1 in most habitats did not exceed 5%. The SILVA database (Quast et al., 2013) contains presently over 1000 16S rRNA gene sequences assigned to the candidate phylum ОР1.

Although no OP1 member has been isolated in pure culture, some information on the properties of bacteria of this candidate phylum was obtained by sequencing of metagenomes and single-cell genomes. The Genome Taxonomy Database (GTDB, Parks et al., 2018) contains presently 24 ОР1 genomes, although only one of them is complete. The first incomplete genome (MAG, metagenome-assembled genome) of an organism of this group was obtained from the metagenome of a hot spring microbial mat (Takami et al., 2012). Its analysis revealed that this bacterium is acetogenic and possess the Wood–Ljungdahl folate-dependent pathway of CO2 fixation. The name “Candidatus Acetothermum autotrophicum” was therefore proposed for this organism (Takami et al., 2012), and the name Acetothermia was proposed for the candidate phylum OP1 (Rinke et al., 2013). Another MAG (Acetothermia bacterium 64_32) was obtained from the metagenome of the sediments of a sea shelf oil deposit (Hu et al., 2016). These genome lacked the key genes encoding autotrophic CO2 fixation pathways, which indicated heterotrophic metabolism. The only complete genome (one circular chromosome) of an OP1 bacterium was obtained from the metagenome of an anaerobic bioreactor for organic waste treatment (Hao et al., 2018). Genome annotation and reconstruction of the metabolic pathways indicated anaerobic chemoheterotrophic metabolism, with carbon and energy obtained via fermentation of peptides, amino acids, and simple sugars to acetate, formate, and hydrogen. The name “Candidatus Bipolaricaulis anaerobius” was proposed for this bacterium due to its unusual cell morphology, and renaming of the OP1 phylum to “Candidatus Bipolaricaulota” was proposed due to nomenclature problems with the previously used name (Hao et al., 2018).

Deep subsurface ecosystems are extreme habitats with high temperature and pressure. Microbial communities of these ecosystems are usually characterized by the presence of various uncultured bacterial and archaeal groups. Partial or complete genomes have been determined for many of them using metagenomics or single-cell genome sequencing (Anantharaman et al., 2016; Magnabosco et al., 2016; Hernsdorf et al., 2017; Probst et al., 2017). We have previously investigated microbial communities of groundwater in Western Siberia, Russia. These subsurface thermal waters located in Cretaceous sedimentary rocks at the depth of 1 to 3 km are available due to numerous oil exploration wells, through which groundwater flows to the surface under natural pressure. We analyzed the composition of microbial communities (Kadnikov et al., 2017a, 2017b) and sequenced the metagenome of groundwater flowing out through the 5P borehole in Tomsk region, Russia (Kadnikov et al., 2017c). Approximately 50% of this community consisted of methanogenic archaea, while its other half was represented by bacteria of various uncultured lineages (Kadnikov et al., 2017a). Analysis of the community composition based on the 16S rRNA gene sequences revealed bacteria of the phylum Bipolaricaulota, and their share in the community was about 1.5%.

The goal of the present work was to use the results of metagenomic analysis in order to determine the complete genome (at the single chromosome level) of a new member of Bipolaricaulota and to characterize its metabolic pathways and global distribution.

MATERIALS AND METHODS

Well characterization, sampling, and metagenomic DNA isolation. Water was collected from the oil-exploration well 5P, drilled in the 1950s to the depth of 2.8 km in the vicinity of the Chazhemto settlement, Tomsk oblast, Russia. The water sample (20 L) was collected in April 2016 (Kadnikov et al., 2017с). The water temperature was ~20°C, it had near-neutral pH (7.43‒7.6), and a negative redox potential (–304 to ‒338 mV). Sulfate concentration was low (90.4 mg/L; Kadnikov et al., 2017b).

To collect microbial biomass, the sample was filtered through 0.22-μm cellulose nitrate membranes. The filters were homogenized by grinding with liquid nitrogen, and the total community DNA was isolated using the Power Soil DNA Isolation Kit (MO BIO Laboratories, Carlsbad, United States). A total of ~1 µg DNA was obtained.

Sequencing of metagenomic DNA usingIllumina platform and assembly of microbial genomes. Sequencing of metagenomic DNA using Illumina HiSeq2500 (Illumina, United States) was described previously (Kadnikov et al., 2017с). Sequencing (in the paired reads format, 2 × 250 nt) with subsequent filtration (Q > 33) resulted in obtaining DNA sequences with total length of ~16.9 Gbp. The sequences were assembled into contigs using metaSPAdes v. 3.7.1. The contigs were then binned into clusters corresponding to individual microbial genomes (MAG) using CON-COСT (Alneberg et al., 2014). Completeness of assembled MAGs and the degree of their contamination (presence of the contigs belonging to other microorganisms) were determined using CheckM (Parks et al., 2015).

Taxonomic position of assembled genomes was determined according to the GTDB database using GTDB-Tk v. 0.1.3 (Parks et al., 2018). One of the genomes, designated Ch78, was identified as belonging to a Bipolaricaulota member.

Sequencing of metagenomic DNA using theMin-ION system and assembly of a complete genome of a member ofBipolaricaulota. Metagenomic DNA was additionally sequenced on MinION (Oxford Nanopore, United Kingdom) using Ligation Sequencing kit 1D protocol according to the manufacturer’s recommendations. Sequencing resulted in 1418419 reads with the total length of ~1.54 Gbp. These long reads were used to join the contigs belonging to the Ch78 genome into longer sequences. For this purpose, the MinION reads exhibiting high homology to the sequences of the Ch78 contigs were selected using BWA v. 0.7.15 (Li and Durbin, 2009). The contigs were combined using npScarf (Cao, 2017). The possible errors in the consensus sequence were corrected with Pilon 1.22 (Walker et al., 2014) basing on mapping of Illumina reads performed using Bowtie 2 (Langmead and Salzberg, 2012). The genome sequence of the Ch78 bacterium was deposited to NCBI GenBank under accession no. CP034928.

Similarity of genomic sequences. The average amino acid identity (AAI) between the genomes was determined using the aai.rb script from the Enveomics Collection (Rodriguez-R and Konstantinidis, 2016). The values of DNA‒DNA hybridization in silico were calculated using the GGDC2 tool (Meier-Kolthoff et al., 2013) available at http://ggdc.dsmz.de/, using the recommended formula 2.

Genome annotation and analysis. Search for the genes and their annotation were carried out using the RAST server; the annotation was then checked and manually corrected by comparing the predicted protein sequences with the NCBI databases. The N-terminal signal peptides were predicted using Signal P v. 4.1 (http://www.cbs.dtu.dk/services/SignalP/) and PRED-TAT (http://www.compgen.org/tools/PRED-TAT/); transmembrane domains were identified using TMHMM Server v. 2.0 (http://www.cbs.dtu.dk/services/TMHMM/). The index of genomic DNA replication was calculated using the iRep software (Brown et al., 2016). Hydrogenases were classified using the HydDB online service (https://services.birc. au.dk/hyddb/) (Søndergaard et al., 2016).

Phylogenetic analysis. The dataset used for genome-based phylogenetic analysis included the Ch78 genome, genomes of 21 other members of Bipolaricaulota (Table 1), genomes of the members of Synergistetes, Deinococcus-Thermus, and Thermotogae, members of the candidate phyla Aerophobetes and Calescamantes, and the genome of Aquifex aeolicus. In each of these genomes, 43 conservative single-copy marker genes were identified using CheckM, and multiple alignment of concatenated sequences of these marker genes was carried out. This multiple alignment was used to construct the phylogenetic tree by the Maximum Likelihood method using PhyML v. 3.3 (Guindon et al., 2010) with the default parameters.

Table 1.   Main characteristics of the Bipolaricaulota genomes
Full size table

RESULTS AND DISCUSSION

Genome assembly of a member of the candidate phylum Bipolaricaulota. Sequencing of the metagenome of the microbial community from well 5P groundwater using the Illumina technology resulted in generation of ~16.9 Gbp assembled into contigs (Kadnikov et al., 2017c). The contigs were clustered using CONCOCT. One of the MAGs obtained, designated Ch78, was represented by 6 contigs with a total length of 1.7 Mbp, with an average 55-fold sequence coverage. The relative abundance of this genotype in the community, defined as a fraction of this MAG in the metagenome, was about 0.6%. Analysis of the taxonomic position of this genome using Genome Taxonomy database supported its affiliation with the candidate phylum Bipolaricaulota (ОР1).

Complete genome of the Ch78 bacterium was assembled using 1.4 × 106 long reads (~1.5 Gbp in total) obtained by monomolecular nanopore sequencing. As a result, six contigs were assembled into a single circular 1701655 nt-long sequence. According to CheckM assessment, this genome was 100% complete in the absence of possible contamination. This is the second known complete genome of a member of Bipolaricaulota.

Analysis of the Ch78 genome revealed the 5S‒23S rRNA gene operon, a separately located 16S rRNA gene, and 47 transfer RNA genes. Based on genome annotation, 1668 potential protein-encoding genes were identified, of which the functions of 808 could be predicted by comparison with NCBI databases. The Ch78 genome contained a single CRISPR (Clustered Regularly Interspaced Short Palindromic Repeats) locus, containing 50 repeat‒spacer units, and a set of genes encoding the type 1-C CRISPR system. Genome size and gene number of Ch78 were close to those of other Bipolaricaulota (Table 1).

The rate of DNA replication in situ (in groundwater) for Ch78 was evaluated using the iRep software (Brown et al., 2016). The iRep replication index of 1.15 was calculated for the Ch78 genome, indicating slow growth of these bacteria (~15% of cells were actively replicating at the time of sampling), but confirming them to be a metabolically active part of the microbial community.

Phylogenetic position of Ch78. Search for phylogenetically related microorganisms based on genome similarity revealed “Candidatus Bipolaricaulis anaerobius” Ran1 (Hao et al., 2018) to be the closest relative of Ch78 with an average AAI of 76.08%. The AAI values between Ch78 and other Bipolaricaulota genomes did not exceeded 60.1%. According to the AAI thresholds proposed by Konstantinidis et al. (2017) for determination of the phylogenetic position of uncultured microorganisms, Ch78 and “Ca. Bipolaricaulis anaerobius” belong to different species of the same genus. The estimated value of their DNA‒DNA hybridization in silico (18.4‒23.0%) also supports assigning these microorganisms to different species. Identity of the 16S rRNA gene sequences of Ch78 and “Ca. Bipolaricaulis anaerobius” Ran1 was 97%.

A phylogenetic tree based on concatenated sequences of 43 conservative marker genes from Ch78 and other available Bipolaricaulota genomes was constructed in order to elucidate the phylogeny of this phylum. The results obtained (Fig. 1) confirm that “Ca. Bipolaricaulis anaerobius” is the closest relative of Ch78. The previously characterized “Ca. Acetothermum autotrophicum” falls into a cluster together with “Ca. Fraserbacteria bacterium” RBG_16_55_9 and forms a separate remote branch within the candidate phylum Bipolaricaulota.

Fig. 1.
figure 1

Position of “Candidatus Bipolaricaulis sibiricus” Ch78 on the phylogenetic tree constructed using concatenated sequences of 43 conservative single-copy marker genes.

Full size image

Analysis of metabolic pathways of Ch78. Analysis of the Ch78 genome revealed the genes of АВС type transporters responsible for uptake of sugars, amino acids, and peptides into the cell. Search for glycosyl hydrolases containing the N-terminal secretion signal peptides revealed only two α-amylases, probably involved in extracellular hydrolysis of starch and its derivatives. Ch78 also possessed ABC type transporters for maltose/maltodextrin import and intracellular enzymes for utilization of these carbohydrates (maltodextrin glucosidase, α-amylase, amylomaltase, and α-glucosidase). No genes of other hydrolytic enzymes involved in polysaccharide hydrolysis were detected. Imported sugars may be metabolized via the Embden−Meyerhof pathway (Fig. 2). Genome analysis revealed the genes of all enzymes of glycolysis, gluconeogenesis, and the nonoxidative branch of the pentose phosphate pathway. Glycogen may be used by Ch78 as a storage polysaccharide; the pathways for its synthesis and degradation are encoded in its genome. Pyruvate produced by glycolysis may be converted to acetyl-CoA by pyruvate-ferredoxin oxidoreductase. Oxidation of acetyl-CoA with formation of acetate and generation of ATP may be carried out by ADP-producing acetyl-CoA synthetase. The tricarboxylic acids cycle is not closed (succinate dehydrogenase is absent) and is probably used only for biosynthetic purposes.

Fig. 2.
figure 2

The main metabolic pathways of “Candidatus Bipolaricaulis sibiricus” Ch78. Enzyme designations: Atr, aminotransferases; POR, pyruvate-ferredoxin oxidoreductase; OOR, ferredoxin-dependent oxidoreductases of 2-keto acids; ACD, acetyl-CoA synthetase (ADP-forming); Hdr/Mvh, complex of heterodisulfide reductase and group 3c hydrogenase; Mbh, group 4d hydrogenase; Hox, group 3d hydrogenase; PPase, pyrophosphatase. Other designations: EMP, Embden−Meyerhof pathway; PEP, phosphoenolpyruvate; Fd(ox), oxidized ferredoxin; Fd(red), reduced ferredoxin; CoA, coenzyme A; PP, pyrophosphate; Р, phosphate.

Full size image

Amino acids and peptides may also be used as carbon source; their import may be carried out by the ABC type transporters specific for amino acids and oligo- or dipeptides. The genes encoding potentially secreted proteolytic enzymes were not found. Amino acids (imported or formed by intracellular peptide hydrolysis) may be deaminated and converted to 2-keto acids (pyruvate, 2-oxoglutarate, etc.). The 2-keto acids may be oxidized to acyl-CoA derivatives by ferredoxin-dependent oxidoreductases with various specificities; the latter may be then cleaved with production of ATP and relative acids (Fig. 2).

Since the genome of “Ca. Acetothermum autotrophicum”, the first characterized OP1 member, encoded the complete Wood–Ljungdahl pathway of СО2 fixation, the possibility of its autotrophic metabolism was proposed (Takami et al., 2012). The key enzymes of this pathway, including the CO dehydrogenase and acetyl-CoA synthetase complex, were not found in the genome of Ch78. Genome analysis of other ОР1 members revealed the Wood–Ljungdahl pathway in the closest relative of “Ca. Acetothermum autotrophicum,” ”Ca. Fraserbacteria” bacterium RBG_16_55_9, but not in other OP1 members. The metabolic capabilities of Ch78 are probably limited to fermentation of organic compounds, while the known pathways of aerobic or anaerobic respiration are absent.

The presence of several NiFe hydrogenases is an evidence of important role of hydrogen in Ch78 metabolism (Fig. 2). The first one, belonging to group 3c, is a component of a cytoplasmic complex with heterodisulfide reductase (Hdr-Mvh complex), which may carry out Н2 oxidation coupled to heterodisulfide and ferredoxin reduction due to electron bifurcation (Greening et al., 2016). The second hydrogenase (Mbh), of group 4d, is a multisubunit complex associated with the cytoplasmic membrane. Hydrogenases of this group oxidize reduced ferredoxin with production of molecular hydrogen and transport the H+ or Na+ ions through the membrane, generating the transmembrane ion gradient (Greening et al., 2016). The third enzyme (Hox) belongs to group 3d. Depending on the cell redox status, these cytoplasmic hydrogenases may either oxidize NADH and reduce protons to Н2 or to carry out the reverse reaction (Greening et al., 2016). Activities of these hydrogenases can not only drive the reoxidation of NADH and reduced ferredoxin produced during the course of fermentation, but also generate the transmembrane ion gradient, that can be used by the membrane F0F1-type ATP synthase for ATP synthesis. Apart from the group 4d NiFe hydrogenase, Na+-transporting membrane pyrophosphatase may also be involved in formation of the transmembrane ion gradient. Overall, Ch78 has limited abilities of generation of the transmembrane ion gradient because it has no NADH dehydrogenase or other ion pumps.

Interestingly, the genome of “Ca. Bipolaricaulis anaerobius” encodes a group 4d (NiFe)-hydrogenase capable of generating the transmembrane ion gradient but does not encode ATP synthase. Homologs of the Ch78 ATP synthase were found in many OP1 members, which may indicate its loss in “Ca. Bipolaricaulis anaerobius.” Comparison of the genome region flanking the ATP synthase gene cluster in Ch78 and the corresponding region in “Ca. Bipolaricaulis anaerobius” genome revealed their similarity and deletion of ATP synthase locus in “Ca. Bipolaricaulis anaerobius.” Thus “Ca. Bipolaricaulis anaerobius” probably uses the transmembrane gradient only for active transport, while ATP could bes formed only in the reactions of substrate phosphorylation.

Global distribution of Bipolaricaulota and their ecological role. Although the SILVA database (Quast et al., 2013) contains 1054 16S rRNA gene sequences representing the phylum Bipolaricaulota, their length usually does not exceed 300‒700 bp. Out of 366 near complete sequences (over 1200 bp long), most were found in deep sea and lake sediments, hot springs, oil reservoirs, and salt lake sediments. Organic-rich environment and anoxic conditions are common to all these habitats, which may favor development of fermenting organotrophs of the phylum Bipolaricaulota.

The microbial community of the subsurface aquifer in which Ch78 was found consist of methanogenic archaea (~50%) and various uncultured bacterial lineages phylogenetically distant from the known species and belonging mainly to the phyla Firmicutes, Chloroflexi, Proteobacteria, Bacteroidetes, and Ignavibacteriae (Kadnikov et al., 2017a). Members of these phyla are known to be able to hydrolyze complex polymers. Particularly, members of the phyla Ignavibacteriae and Chloroflexi (family Anaerolineaceae) found in subsurface waters in the Western Siberian region may be involved in the degradation of polysaccharides and proteins (Kadnikov et al., 2018). In such ecosystems, members of the Bipolaricaulota probably act as scavengers, consuming low-molecular organic compounds produced by hydrolytic microorganisms and producing hydrogen and acetate, the substrates for methanogenic archaea.

Description of “Candidatus Bipolaricaulis sibiricus” Ch78. Based on analysis of its complete genome, a new species, “Candidatus Bipolaricaulis sibiricus,” is proposed. Bipolaricaulis sibiricus sp. nov. (si.bi’ri.cus N.L. masc. adj. sibiricus, originating from Siberia). Uncultured. The organism was detected in a deep subsurface thermal aquifer in Western Siberia. Probably anaerobic, capable of fermenting sugars and proteinaceous substrates. The DNA G + C content is 67.23 mol %. The organism is represented by complete genome (GenBank CP034928) obtained from the metagenome of subsurface thermal water from the 5P borehole (Chazhemto, Tomsk region, Russia).