Bacteria were cultivated in a medium containing (gL^–1): (NH₄)₂SO₄ – 0.5, NaNO₃ – 0.3, CaCl₂ – 0.03, KH₂PO₄ – 0.01, K₂HPO₄ – 0.022, Na₂HPO₄.7H₂O – 0.035, MgSO₄.7H₂O – 0.05. Prior to inoculation, the following components were added to one liter of the medium as sterile solutions: trace elements and vitamins (Pfennig and Lippert, 1966; Kuznetsov and Dubinina, 1989), 1 ml; NaHCO₃, 0.5 g; Na₂S₂O₃.5H₂O, 1.0 g; sodium lactate, 0.25 g and sodium acetate, 0.25 g. The pH of the medium was adjusted to 7.5. Cultures were incubated at 27°C.

Physiological tests were carried out using a medium of the above composition using alternative carbon sources. Three passages of culture were carried out with each of the substrates to test bacterial growth. To study the ability to fix molecular nitrogen, bacteria were cultivated in a medium of the above-described composition, excluding nitrogen sources. The ability of the strains to fix nitrogen was confirmed by the acetylene reduction assay (Stewart et al., 1968).

Isolation of pure cultures of representatives of the genus Thiothrix was performed from the selected bacterial fouling, where bundles of filamentous sulfur bacteria were collected, washed with a large amount of sterile water, and transferred to a Potter manual homogenizer to prepare suspensions of colorless sulfur bacteria threads, in order to obtain Thiothrix gonidia. A number of tenfold dilutions were prepared by the method of limiting dilution from the starting material. To obtain individual colonies, serial dilutions were plated on a solid nutrient medium of the composition described above, which additionally contained (gL^–1): agar Difco – 7.0; Na₂S.9H₂O – 0.3.

Fatty acid analysis was performed as described elsewhere (Slobodkina et al., 2020). Briefly, about 10 mg of freeze-dried biomass collected at the late exponential growth phase were treated with anhydrous HCl/MeOH, extracted with hexane, and analyzed with a Trace GC Ultra DSQ II GC-MS system (Thermo Fisher Scientific, HP-5MS column, EI70 eV).

Genomic DNA was isolated from Thiothrix sp. AS and T. unzii A1^T (strain was obtained from ATCC collection ATCC 49747^T) using a DNeasy PowerSoil DNA isolation kit (Mo Bio Laboratories, Carlsbad, CA, United States). Genomic DNA was sequenced using Illumina and Oxford Nanopore platforms. For Illumina sequencing, the shotgun genome libraries were prepared using the NEBNext Ultra II DNA library prep kit (New England Biolabs, United States). The libraries were sequenced on an Illumina MiSeq (Illumina, San Diego, CA, United States) in a paired reads mode (2 × 300 nt). Low quality sequences were trimmed using Sickle v.1.33 (q = 30)¹. In addition, genomic DNA was sequenced on a MinION device (Oxford Nanopore, United Kingdom) using the ligation sequencing kit 1D and FLO-MIN110 cells. Sequencing statistics are shown in Supplementary Table 1. For each genome, nanopore reads were assembled into a single circular contig using Flye v. 2.8.2 (Kolmogorov et al., 2019). The consensus sequence of the assembled contig was corrected with two iterations of Pilon v.1.22 (Walker et al., 2014) using Illumina MiSeq reads.

The total DNA was extracted from 200 μg of a microbial fouling sample using a DNeasy PowerSoil DNA isolation kit. Metagenomic DNA was sequenced using Illumina and Oxford Nanopore platforms as described above. Nanopore reads were de novo assembled using Flye v.2.7 (Kolmogorov et al., 2019). The obtained contigs were binned into metagenome-assembled genomes (MAGs) using MetaBAT v. 2.12.1 (Kang et al., 2015). The taxonomic assignment of the obtained MAGs was performed using the Genome Taxonomy Database Toolkit (GTDB-Tk) v.1.1.1 (Chaumeil et al., 2020) and Genome Taxonomy Database (GTDB; Parks et al., 2018). Completeness of the MAGs and their redundancy (contamination) were estimated using the CheckM v.1.05 tool (Parks et al., 2015).

One MAG, designated A52, and assigned to Thiothrix, was manually assembled into a single circular contig using the Flye assembly graph visualized in Bandage v. 0.8.1 tool (Wick et al., 2015). The assembly sequence was polished using Illumina reads with two iterations of Pilon v. 1.22 (Walker et al., 2014).

Genomic DNAs from Thiothrix sp. Ku-5 and T. fructosivorans Q^T (strain was obtained from ATCC collection ATCC 49748^T) were purified using a Monarch Genome Purification kit (T3010S, NEB, MA, United States) and sequenced using the Pacific Biosciences (PacBio) RSII sequencing platform. SMRTbell libraries were prepared using a modified PacBio protocol adapted for NEB reagents. Genomic DNA samples were sheared to an average size of ∼ 10–20 kb using the G-tube protocol (Covaris; Woburn, MA, United States), treated with FFPE, end repaired, and ligated with hairpin adapters. Incompletely formed SMRTbell templates were removed by digestion with a combination of exonuclease III and exonuclease VII (New England Biolabs, Ipswich, MA, United States). The qualification and quantification of the SMRTbell libraries were made on a Qubit fluorimeter (Invitrogen, Eugene, OR, United States) and a 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA, United States). Single molecule real time sequencing (SMRT) sequencing was performed using a PacBio RSII (Pacific Biosciences; Menlo Park, CA, United States) based on standard protocols for 10--20 kb SMRTbell library inserts. Sequencing reads were collected and processed using the SMRT Analysis pipeline from Pacific Biosciences² (Chin et al., 2013).

In addition, genomic DNA of T. fructosivorans Q^T was sequenced on a GridION device (Oxford Nanopore, United Kingdom) using the ligation sequencing kit 1D and R9.4 flow cell for 12 h run. Sequencing statistics are shown in Supplementary Table 1. Nanopore reads allowed us to overcome two large, approximately 50 kb repeats and the chromosome was assembled into a single circular contig using Basecaller 4.2.3 and MinKNOW 20.10.6 software.

Gene search and annotation were performed using the National Center for Biotechnology Information (NCBI) Prokaryotic Genome Annotation Pipeline (Tatusova et al., 2016; Haft et al., 2018). ANI was calculated using an online resource³ based on the OrthoANIu algorithm, using USEARCH (Lee et al., 2016). AAI between the genomes was determined using the aai.rb script from the enveomics collection (Rodriguez-R and Konstantinidis, 2016). dDDH calculation was performed using the GGDC⁴ online platform.

The core genome of analyzed strains was determined by clustering all genes selected by the BLASTclust v.2.2.26 program from the BLAST package (Korf and Gish, 2000). Clustering was carried out by the single link method. Genes whose predicted protein products had more than 70% amino acid sequence identity across more than 80% of the length were combined into one cluster. Clusters containing at least one protein from each genome formed the core genome.

The Genome Taxonomy Database toolkit (GTDB-Tk; Chaumeil et al., 2020), release 04-RS89 was used to identify the 120 single-copy phylogenetically informative conservative marker genes used in the GTDB classification system (Parks et al., 2018) in the genomes of T. lacustris BL^T, Thiothrix sp. AS, T. fructosivorans Q^T, Thiothrix sp. Ku-5, T. caldifontis G1^T, Thiothrix sp. A52, T. unzii A1^T, “Ca. Thiothrix moscowensis RT,” “Ca. Thiothrix singaporensis SSD2,” Thiothrix singaporensis SSD2,” Thiothrix nivea JP2^T, Thiolinea disciformis B3-1^T, Thiothrix eikelboomii AP3^T, Thiofilum flexile EJ2M-B^T, and Leucothrix mucor DSM 2157^T. These genes were used to construct a multiple alignment of concatenated amino acid sequences in GTDB-Tk. The multiple alignment was used to construct a maximum-likelihood phylogenetic tree in PhyML v. 3.3 (Guindon et al., 2010) with default parameters (LG amino-acid substitution model was used, equilibrium amino-acid frequencies are defined by the substitution model, gamma distribution with estimated shape parameter was used to model 4 substitution rate categories, no invariant sites). The levels of support for internal branches were assessed using the Bayesian test in PhyML.

The annotated genome sequences of Thiothrix sp. AS, T. unzii A1^T, and Thiothrix sp. A52 were submitted to the NCBI GenBank database and are accessible via BioProject PRJNA719636. The annotated genome sequences of Thiothrix sp. Ku-5 and T. fructosivorans Q^T were submitted to the NCBI GenBank database and are accessible via the BioProject PRJNA631701.

The samples for metagenomic analysis, as well as for the isolation of new Thiothrix isolates were taken from microbial fouling in the place of the outflow of sulfidic water from a drained well from the closed and flooded coal mine “Severnaya” in Kemerovo region of Russia (55°19.59′N, 84°4.59′E), a sulfidic spring in the Volgograd region of Russia (49°8.921′N, 44°6.236′E) and on the shore of the White Sea, Russia at the border of the freshwater stream flowing into the sea (66°33.17′N, 33°5.58′E) (Figures 1A,C,D). In the latter case, bacterial fouling was taken at low tide in the channel of a fresh stream; hydrogen sulfide came from the bottom sediments.

The studied biotopes were characterized by a low redox potential (from −112 to −120 mV), the oxygen concentration varied from 2 to 5 mg L^–1, and the concentration of hydrogen sulfide was from 1 to 6 mg L^–1. The water temperatures at all sampling sites was between 9 and 16°C (Table 1).

The Thiothrix sp. Ku-5 was isolated from bacterial fouling in the drainage water flowing out of the coal mine in Kemerovo, Russia. The genome sequence of the Thiothrix sp. Ku-5 assembled into four closed-circular contigs; one circular 4,036,668 bp long chromosome (CP053482) and three circular plasmids, – pTSUKu5_27_unt8 (CP053484), pTSUKu5_12_unt9 (CP053483), and pTSUKu5_2_unt34 (CP053485) with lengths of 26,592 bp, 12,118 bp, and 2,180 bp, respectively. Annotation of the genome of Thiothrix sp. Ku-5 revealed four copies of the 16S rRNA gene, 46 tRNA genes, and 3,965 potential protein-coding genes.

The Thiothrix sp. AS was isolated from bacterial fouling in the shore zone of the White Sea. The genome of the Thiothrix sp. AS was assembled into a single closed circular contig (CP072801) with a total length of 4,177,302 bp. Genome annotation revealed three copies of the 16S rRNA gene, 44 tRNA genes, and 4,045 potential protein-coding genes.

The genome of the T. unzii A1^T was assembled into eight contigs; one circular 3,648,856 bp long chromosome (CP072793.1) and seven circular plasmids: pTunz1 (CP072796.1, 34,476 bp), pTunz2 (CP072795.1, 22,458 bp), pTunz3 (CP072792.1, 21,012 bp), pTunz4 (CP072799.1, 18,228 bp), pTunz5 (CP072797.1, 14,837 bp), pTunz6 (CP072798.1, 14,317 bp), and pTunz7 (CP072794.1, 14,102 bp). The genome annotation revealed two copies of the 16S rRNA gene, 44 tRNA genes, and 3,729 potential protein-coding genes.

The original polished assembly of T. fructosivorans Q^T generated three linear contigs (1,938,437; 1,467,464 and 312,998 bp) (JAFMPM000000000.1) for the main chromosome with 51.37% GC and four closed circular plasmids pTfr153 (CP072750, 153,578 bp), pTfr21 (CP072751, 21,013 bp), pTfr12 (CP072749, 12,657 bp), pTf2 (CP072752, 2,341 bp). Therefore, T. fructosivorans Q^T was additionally sequenced on a GridION device (Oxford Nanopore, United Kingdom). Nanopore reads allowed us to overcome two large, approximately 50 kb repeats and the chromosome was finally assembled into a single circular contig (CP072748, 4,137,062 bp). Annotation of the genome of T. fructosivorans Q^T revealed two copies of the 16S rRNA gene, 44 tRNA genes, and 4,071 potential protein-coding genes.

Metagenome-assembled genome of Thiothrix sp. A52 was assembled from the environmental DNA isolated from bacterial fouling from a sulfidic spring in the Volgograd region. The metagenome sequencing resulted in 11 MAGs with completeness higher than 70%; three of these MAGs were assigned to the genus Thiothrix. One MAG was sequenced at 201-fold coverage and represented about half of the entire metagenome. Using long reads obtained by nanopore sequencing, the complete closed circular sequence of the chromosome with a length of 3,546,315 bp (CP072800) was obtained. The genome annotation revealed two copies of the 16S rRNA gene, 30 tRNA genes, and 3,613 potential protein-coding genes.

The main characteristics of the obtained genomes are shown in Table 2.

The sequence identities of the 16S rRNA genes among the members of the genus Thiothrix are in the range from 93.3 to 98.9% (Figure 2). For example, the 16S rRNA genes of Thiothrix sp. AS and T. lacustris BL^T are completely identical (Figure 2). Strain AS previously was used as an example of the ability of representatives of the genus Thiothrix to grow anaerobically on nitrates (Trubitsyn et al., 2014). However, the ANI and dDDH values clearly show that Thiothrix sp. AS and T. lacustris BL^T are different species. According to ANI and dDDH data Thiothrix sp. AS is sufficiently distant from other Thiothrix species (<92.3% and <56.2%, respectively) to be classified as a new species (Figure 2 and Supplementary Figure 1). The ANI values between MAG Thiothrix sp. A52, Thiothrix sp. Ku-5, and other Thiothrix genomes are less than 90.1%, which is below the species boundary cutoff of 95% (Konstantinidis and Tiedje, 2005; Figure 2). These data are also consistent with the results of dDDH (<46.9%) (Figure 2).

The phylogenetic position of the new Thiothrix bacteria was also analyzed by building a phylogenetic tree based on concatenated sequences of 120 conservative marker genes. All new genomes appeared to form a distinct lineage within the genus Thiothrix, along with previously described species (Figure 3).

Overall, genome-based phylogenetic reconstructions and pairwise ANI and dDDH values clearly confirmed that Thiothrix sp. AS, Thiothrix sp. Ku-5, and MAG of Thiothrix sp. A52 can be classified as new species within the genus Thiothrix, for which we propose names Thiothrix litoralis sp. nov. AS^T, Thiothrix subterranea sp. nov. Ku-5^T, and “Candidatus Thiothrix anitrata” sp. nov. A52.

Thiothrix sp. Ku-5 and Thiothrix sp. AS are Gram negative rods forming immobile filaments (trichomes) surrounded by a mucous sheath. They form mobile gonidia, which give white colonies, which are rosettes with a diameter of 1–3 mm (Figure 1B).

The bacteria can grow in a temperature range of 4–28°C with an optimum growth at 20–22°C. pH range 6.8–8.0, with an optimal at pH 7.4–7.5. The growth range with NaCl for Thiothrix sp. Ku-5 is up to 1%, however, Thiothrix sp. AS is able to survive up to 3% NaCl in the medium. Both are capable of lithotrophic growth in the presence of reduced sulfur compounds (H₂S, Na₂S₂O₃). When growing with thiosulfate or hydrogen sulfide, elemental sulfur accumulates in the cells (Figure 1B). They are also capable of lithoautotrophic, lithoheterotrophic, and organoheterotrophic growth. Comparative characteristics of the phenotypic and cultural properties of Thiothrix sp. Ku-5 and Thiothrix sp. AS with other phylogenetic closely related species of the genus Thiothrix are presented in Table 3. It should be noted that the phenotypic and cultural properties of all pure cultures of representatives of the genus Thiothrix are largely consistent. The composition of major fatty acids determined in this work is consistent with the data obtained earlier for other species (Chernousova et al., 2009; Supplementary Table 2 and Supplementary Figure 2).

According to polyphase analysis, the new isolates of Thiothrix sp. Ku-5 and Thiothrix sp. AS were assigned as new species of the genus Thiothrix, T. subterranea sp. nov. Ku-5^T, and T. litoralis sp. nov. AS^T, respectively.

In this work, representatives of the genus Thiothrix were studied using a pangenomic approach. We analyzed all publicly available genomes of taxonomically described species of the genus Thiothrix. In addition, the newly obtained genome sequences of T. subterranea Ku-5^T, T. litoralis AS^T, “Ca. Thiothrix anitrata” A52, T. fructosivorans Q^T and T. unzii A1^T were included in the analysis (Table 2). All genomes were of high quality except for MAG of “Ca. Thiothrix singaporensis” SSD2 which contained multiple frameshifts and therefore was excluded from the analysis.

The pangenome analysis of nine species indicated that the genus Thiothrix comprises 11,549 gene clusters. The core genome of the genus Thiothrix contains 1,355 genes and the auxiliary genome comprises 3,894 gene clusters present in 2–8 genomes. The number of unique genes varies in different species from 397 to 1,360 (Figure 4). These genes were predicted to encode mostly hypothetical proteins with unknown functions, transporters, transcriptional regulators, methyltransferases, transposases, and the toxin-antitoxin systems.

The genus Thiothrix is characterized by the presence of a number of genes for dissimilatory sulfur metabolism. Thiothrix species are classical lithotrophs, capable of using reduced sulfur compounds as an energy source. The hydrogen sulfide oxidation systems, SQR (sqrA and sqrF) and FCSD (fccAB), are present in the core genome and were found in all species. The branched SOX system, represented by the soxAXYZB genes and involved in the oxidation of thiosulfate to sulfur and sulfate, was also found in all members of the genus. The genes of the reverse Dsr (rDsr) dsrABEFHÑEMKLJONR complex involved in the oxidation of sulfur to sulfite were also present in all species, as well as genes for direct (soeABC) and indirect (aprAB, sat) pathways of oxidation of sulfite to sulfate. Thus, the presence of genes of several systems of dissimilatory sulfur metabolism in all known species of the genus Thiothrix indicates their exceptional role in the life of this group of sulfur bacteria.

The core genome contains the hyaA and hyaB genes encoding the large and small catalytic subunits of membrane-bound [NiFe] -hydrogenase of type I, and the hyaC gene, which encodes the cytochrome b subunit, which attaches hydrogenase to the cytoplasmic membrane. In addition, the core genome contains hypABCDEF genes, which are involved in the maturation of [NiFe] – hydrogenases.

The core genome includes all genes for the glycolysis, the Krebs cycle and the glyoxylate cycle. All members of the genus Thiothrix lack the classical NAD-dependent malate dehydrogenase (mdh), but instead contain genes encoding FAD-dependent malate: quinone oxidoreductase (mqo). The glk gene for glucokinase was found in all members of the genus, however, it was not included in the core genome due to the lower sequence similarity, which suggests horizontal transfer or an increase rate of evolution of this gene.

All Thiothrix species have a complete set of genes for the Calvin-Benson-Bassham cycle of carbon fixation. Genes for the type IAq RuBisCO were in the core genome, while type IAc was found in all genomes with the exception of T. lacustris BL^T and T. fructosivorans Q^T. All the species except “Ca. Thiothrix anitrata” A52 also have type II RuBisCO (Figure 5).

The electron transport chain (ETC) was represented in all species by complex I NuoABCDEFGHIJKLMN, complex II SdhABCD, complex III (CytB, Cyc1, Ryp1), and cytochrome c oxidase of complex IV (CcoNOPGQS). The presence of all ETC genes in all genomes additionally confirms the respiratory type of metabolism in the genus Thiothrix (Chernousova et al., 2009).

All members of the genus Thiothrix except for “Ca. Thiothrix anitrata” A52, are capable of anaerobic respiration with nitrogen compounds. However, the genes for dissimilatory nitrogen metabolism were not found in the core genome, and their set differs markedly among Thiothrix species (Figure 5). The nirS nitrite reductase gene is presented in T. litoralis AS^T, T. unzii A1^T, T. fructosivorans Q^T, T. caldifontis G1^T, but was probably lost in T. lacustris BL^T, T. nivea JP2^T, “Ca. Thiothrix moscowensis” RT, T. subterranea Ku-5^T, and “Ca. Thiothrix anitrata” A52. Assimilatory ammonium-forming nitrite reductase nirBD is absent only in “Ca. Thiothrix anitrata” A52. The cnorBC nitric oxide (NO) reductase genes were found in T. litoralis AS^T, T. unzii A1^T, T. fructosivorans Q^T, T. caldifontis G1^T, and “Ca. Thiothrix moscowensis” RT. All members of the genus Thiothrix lack the nosZ gene encoding the dissimilatory N₂O reductase. In the cases of nirS, nirBD and cnorBC, the corresponding genes found in different Thiothrix species were highly homologous; therefore, their absence in individual species probably indicates a loss in course of evolution rather than an independent acquisition as a result of horizontal gene transfer.

The distribution pattern is more complex for dissimilatory nitrate reductase genes. In the Thiothrix genomes two types of genes for NarGHI reductase were found, the identity of their deduced amino acid sequences is less than 60%. Genomes of T. lacustris BL^T, T. litoralis AS^T, T. fructosivorans Q^T, T. subterranea Ku-5^T, and T. caldifontis G1^T contain only one type of narG, the genome of “Ca. Thiothrix moscowensis” RT harbored only another type, and the genome of T. unzii A1 ^T contains both variants of the narG gene. The narG genes are missing in the genomes of “Ca. Thiothrix anitrata” A52 and T. nivea JP2^T; in the later species the periplasmic nitrate reductase napAB is present instead of the narGHI (Figure 5). Such distribution pattern indicates the events of the acquisition of additional copies of the nar genes as a result of horizontal transfer with the subsequent loss of one of the two copies, as well as the replacement of nar operon with Nap type of nitrate reductase.

The nifASUBNXX2YB2ENQVWMHDKZTO gene cluster for molecular nitrogen fixation is present in all species of the genus, with the exception of T. fructosivorans Q^T, “Ca. Thiothrix anitrata” A52, and T. lacustris BL^T. The amination process for all members of the Thiothrix genus is represented by glutamine synthetase (glnB), glutamate synthase (gltBD), and aspartate aminotransferase (aspB).

The genes pstDCABS and phoURB, involved in inorganic phosphorus transport into the cell, were found in the core genome. The core genome of the genus Thiothrix contains gene for polyphosphate kinases PPK1 that can catalyzes the synthesis of the poly-P chain. Also The core genome contains gene for poly-AMP phosphotransferase (PAP), which belongs to class II of polyphosphate kinases PPK2 and that can catalyze the synthesis of the ADF (Kimura and Kamatani, 2021). Most likely, in Thiothrix, PAP is responsible for ATP regeneration in the cell (Kameda et al., 2001). In addition, the epp gene for exopolyphosphatase, which enables the cleavage of polyphosphates, was found in all genomes. The presence of polyphosphate kinase PPK1 and exopolyphosphatase suggests that Thiothrix species could accumulate and degrade polyphosphates which are probably used for temporal ATP storage.