Complete genome sequence of and proposal of Thermofilum uzonense sp. nov. a novel hyperthermophilic crenarchaeon and emended description of the genus Thermofilum

A strain of a hyperthermophilic filamentous archaeon was isolated from a sample of Kamchatka hot spring sediment. Isolate 1807-2 grew optimally at 85 °C, pH 6.0-6.5, the parameters being close to those at the sampling site. 16S rRNA gene sequence analysis placed the novel isolate in the crenarchaeal genus Thermofilum; Thermofilum pendens was its closest valid relative (95.7 % of sequence identity). Strain 1807-2 grew organothrophically using polysaccharides (starch and glucomannan), yeast extract or peptone as substrates. The addition of other crenarchaea culture broth filtrates was obligatory required for growth and could not be replaced by the addition of these organisms’ cell wall fractions, as it was described for T. pendens. The genome of strain 1807-2 was sequenced using Illumina and PGM technologies. The average nucleotide identities between genome of strain 1807-2 and T. pendens strain HRK 5T and “T. adornatus” strain 1910b were 85 and 82 %, respectively. On the basis of 16S rRNA gene sequence phylogeny, ANI calculations and phenotypic differences we propose a novel species Thermofilum uzonense with the type strain 1807-2T (= DSM 28062T = JCM 19810T). Project information and genome sequence was deposited in Genbank under IDs PRJNA262459 and CP009961, respectively. Electronic supplementary material The online version of this article (doi:10.1186/s40793-015-0105-y) contains supplementary material, which is available to authorized users.


Introduction
Currently, the archaeal phylum Crenarchaeota comprises five orders: Thermoproteales, Sulfolobales, Desulfurococcales, Acidilobales and Fervidicoccales [1]. Thermoproteales representatives are characterized as rod-shaped extreme thermophiles and hyperthermophiles, growing either auto-or heterotrophically, using different redox pairs to gain energy as well as performing fermentation of organic substrates [2]. The order consists of two families: Thermoproteaceae [3], and Thermofilaceae [4]; the second one is a deeply branching lineage, consisting so far only one validly published genus and species, Thermofilum pendens isolated from a solfataric hot spring in Iceland [5] and characterized as an anaerobic hyperthermophilic, moderately acidophilic chemoorganotrophic archaeon utilizing peptides as the energy source and sulfur as the electron acceptor. Its growth is obligatory dependent on Thermoproteus tenax polar lipid fraction. Other T. pendens strains were isolated from solfataras of Yellowstone National Park (USA) and Vulcano Island (Italy) [4]. Notably, one of them described non-validly as a separate species "Thermofilum librum" has 100 % 16S rRNA gene sequence identity with T. pendens, but does not require the addition of other organisms' cell components [6]. Recently, another species "Thermofilum adornatus" strain 1910b was isolated from a black mud pit (Kamchatka, Russia) and its genome was sequenced [7].
Here we describe another Thermofilum strain 1807-2, report its genomic sequence, what allow us to propose a novel species Thermofilum uzonense strain 1807-2 T . This new data expands the knowledge on physiology and diversity of this deep lineage in archaeal domain.

Classification and features
In 2008, a gray mud sample was collected from the hot spring (T 83°C, pH 6.2) located in Orange Thermal Field, Uzon Caldera, Kamchatka, Russia (54.30 N 160.00 E). An enrichment culture was obtained using strictly anaerobic modified freshwater Pfennig medium with cellobiose and yeast extract (2 g l −1 and 1 g l −1 , respectively) as substrates [8]. After 4 days of incubation at 84°C at pH 5.8 two different types of cellsextremely thin rods and regular small cocciwere detected in the enrichment culture.
Strain 1807-2 T was purified by serial dilution technique on the same medium in the presence of 1/100 (v/v) Fervidicoccus fontis strain 1910a culture broth as it was performed for "Thermofilum adornatus" strain 1910b [7].
Strain 1807-2 T was a hyperthermophile and obligate anaerobe. It grew optimally at 85°C and pH 6.0-6.5 without sodium chloride in the medium. In the presence of 10 g l −1 NaCl growth of the strain ceased completely. Addition of at least 25 mg l −1 of yeast extract was mandatory. Peptone, yeast extract, starch and glucomannan were used as substrates for growth while amorphous cellulose [9] and filter paper, mannan, amorphous chitin [10], chitosan, glycerol and carbon monoxide did not support it ( Table 1). Addition of 1/100 (v/v) of culture broth filtrates of other Crenarchaeota (Fervidicoccus fontis, Desulfurococcus kamchatkensis or Pyrobaculum sp.) which served, most probably, as a source of growth factors, was obligatory required for growth of strain 1807-2 T . Filtrates could not be replaced by cell wall-fractions of these organisms, or bacterial (Caldicellulosiruptor kronotskyensis) culture broth filtrates.
Strain 1807-2 T was deposited in DSMZ (German collection of microorganisms and cell cultures) under accession number DSM 28062, and in JCM (Japan Collection of Microorganisms) under accession number JCM 19810.
16S rRNA gene-based phylogenetic analysis of strain 1807-2 T placed it into Thermofilaceae family being most closely related to Thermofilum pendens Hrk5 T showing 95.7 % sequence identity according to calculations, considered used substitution model (Fig. 2). For the analysis a complete 16S rRNA gene of strain 1807-2 T and almost complete (>1300 nt) sequences of its 50 best BLAST hits from GenBank nr/nt database, filtered through 100 % identity filter were used (final dataset included 38 sequences). 13 representative of Thermoproteaceae were used as an outgroup for the analysis (Additional file 5).
On 1st of May 2015 both RDP (Release 11, Update 3) and Silva (SSU r122) databases contain altogether 400 unique 16S rRNA genes of Thermofilaceae family clones and isolates. Most of them were partial, hence were not involved in the mentioned-above phylogenetic analysis. The analysis of distribution of their isolation sources was performed using our homemade software GetIsola-tionSources [11]. It searches for the GenBank accession numbers in SILVA, RDP and other databases, containing it, and extracts the information, residing under modifiers "description", "accession", "isolation source", "country" and "references" from the respective GenBank records. Analysis revealed that the majority of Thermofilaceae family clones and pure cultures were obtained from hot springs, both terrestrial and marine. A few others were collected from various non-thermal environments. All GenBank IDs of these 16S rRNA genes are listed in Additional file 1.

Genome project history
The sequencing project started in May 2013 and finished in August 2014. Due to the fact that both short and long insert libraries were used for sequencing, the obtained circular genomic contig can be considered as a finalized genome sequence. The complete genome sequence of T. uzonense strain 1807-2 T has been deposited in DDBJ/ EMBL/GenBank under the accession number CP009961. Related project information and sample details have been deposited in NCBI database under accession numbers PRJNA262459 and SAMN03083278, respectively ( Table 2).

Growth conditions and genomic DNA preparation
T. uzonense strain 1807-2 T was grown in medium mentioned above with glucose and yeast extract (1 g l −1 and Fig. 1 Electron micrograph of Thermofilum uzonense strain 1807-2 T . Bar, 0.5 μm. Electron micrograph of negatively stained cell of Thermofilum uzonense strain 1807-2 T . Cells were grown on glucose containing medium at 85°C and incubated for two days (late exponential phase of growth). Bar, 0.5 μm 0.5 g l −1 , respectively) as substrates and in the presence of 1/100 (v/v) Desulfurococcus kamchatkensis strain 1221n culture broth as a source of growth factors at 85°C during 4 days. Cell-free culture broth of D. kamchatkensis was obtained by filtration of grown cells through 0.22 μm filter. For that, cells of strain 1221n were grown on the same medium containing glucose and yeast extract as growth substrates at optimal conditions [12]. After cultivation, cells of strain 1807-2 T were disrupted with glass beads using Minilys homogenizer (Bertin technologies, France) and DNA was extracted using QIAamp DNA mini kit (Qiagen, Netherlands) according to the manufacturers' instructions.

Genome sequencing and assembly
For sequencing of T. uzonense genome both fragment and mate-paired libraries were used. Fragment library for Miseq paired-end sequencing was prepared from 50 ng genomic DNA with NexteraTM fragment library kit (Illumina, San Diego, CA, USA) according to manufacturer instructions. Analysis of library size distribution showed the mean insert size of 175 bp. Sequencing of this library using MiSeq instrument resulted in 932,095 pairs of 250 bp reads which were subjected to stringent quality and adapter trimming with the corresponding tool of CLCBio Genomics Workbench 7.5 (Qiagen, Netherlands). After trimming and filtering procedures 911,060 read pairs were used for assembly.
The mate-paired library with target insert size of 2400 bp for Ion Torrent sequencing was prepared from 3 μg of DNA with SOLiD 5500 mate-paired library kit (Life Technologies, Carlsbad, CA, USA) using Ion Torrent PGM adapters for final amplification. Library was sequenced on a 314 chip resulting in 300,306 single sequences. The internal adapter used for library construction was trimmed during import procedure to CLC Genomics Workbench. After quality and length filtering totally 197,320 read pairs were used for de novo assembly guidance and scaffolding (single reads were excluded from further analysis).
The initial assembly was made with CLCBio de novo assembler using word size of 64 and bubble size of 500 bp [13]. Since overall quality of MiSeq reads is generally much higher than PGM reads quality, the Illumina  [39]. The tree with the highest log likelihood (−5739.3765) is shown. The percentage of trees in which the associated taxa clustered together (bootstrap test of 1000 replications) is shown next to the nodes. Initial tree(s) for the heuristic search were obtained by applying the Neighbor-Joining method to a matrix of pairwise distances estimated using the Maximum Composite Likelihood (MCL) approach. A discrete Gamma distribution was used to model evolutionary rate differences among sites (4 categories (+G, parameter = 0.2411)). The tree is drawn to scale, with branch lengths measured in the number of substitutions per site. The analysis involved 38 nucleotide sequences; all were longer than 1300 nucleotides. All positions containing gaps and missing data were eliminated. There was a total of 1.227 positions in the final dataset. Evolutionary analyses were conducted in MEGA6 [40]. Thermoproteaceae branch includes 13 sequences (Additional file 5). Cultivated strains are underlined. Thermofilum uzonense strain 1807-2 T is in bold. Complete 16S rRNA gene of Desulfurococcus kamchatkensis (NC_011766.1, Desulfurococcales order) was chosen as an out-group. Bar, 2 substitutions per 100 nucleotides reads were used for assembly of contigs, while PGM reads used only in "guidance only" mode for de Bruijn graph ambiguities resolution. Twenty two scaffolds with L50 of 456,782 bp were obtained. After in silico closure of gaps with GapFiller v1.11 [14] contigs were subjected to the second round of scaffolding and gapfilling with SSPACE v.2.0 [15] and GapFiller. Finally one gapless contig of 1,610,790 bp was obtained. Assembly validation was accomplished by mapping of all high-quality MiSeq reads to the final chromosome. Analysis of the mapping revealed 3 problematic regions characterized by the enrichment of mapping conflicts and unaligned read ends. These regions were amplified by PCR and sequenced by Sanger method with ABI 3730XL DNA analyzer (Life Technologies, Carlsbad, CA, USA). Chromosome circularity was confirmed by PCR with outward-oriented primers and subsequent Sanger sequencing. After all correction procedures the length of Thermofilum uzonense strain 1807-2 T circular chromosome was 1,611,988 bp.

Genome annotation
Identification of genes and primary annotation was performed with PGAAP pipeline [16]. As a part of the pipeline, protein-coding genes were detected with GeneMarkS+ [17], tRNA genes were identified with , not directly observed for the living, isolated sample, but based on a generally accepted property for the species, or anecdotal evidence). These evidence codes are from the Gene Ontology project [38] tRNA-scan-SE [18] and rRNA genes were identified with blastn search against a curated NCBI rRNA reference set. Assignment of predicted CDS to the clusters of orthologous groups (COGs) was made using BLAST [19] against the latest version of COG database [20] with maximal e-value of 10 −5 . Identification of conserved domains families was performed using HMMER hmmscan service [21]. Proteins harboring signal peptides and transmembrane helices were identified with Phobius server [22] Twin-arginine signal peptides were predicted using TatP server [23], while non-classical signal peptides were predicted using SecretomeP server [24]. Other predictions were done according to the genome annotation protocol [25].

Genome properties
The genome of T. uzonense strain 1807-2 T consists of one circular chromosome of a total length of 1,611,988 base pairs (47.9 % G+C content). A total of 1697 genes were predicted, 1455 of which are protein-coding genes (CDS), 50 are RNA genes and 192 are pseudogenes. The genome has one ribosomal operon consisting of single copies of 16S and 23S ribosomal RNA genes, while 5S gene is located 300 kb apart from it. A search for IS elements using IS Finder database [26] showed that genome contains two IS200/IS605 family IS elements, one of which is inactive due to disrupted transposase gene. The chromosome also contains two CRISPR stretches of total length about 13 kb [27].
Approximately 70 % of CDS (1019 of 1455) were associated with Clusters of Orthologous Groups (COGs). Since the majority of individual domains of multidomain proteins were assigned to different COGs the number of COGs is higher than the number of proteins with COGs: 1389 ( Table 3). The distribution of genes into COGs functional categories is presented in Table 4. Graphical map of T. uzonense strain 1807-2 T circular chromosome is presented on Fig. 3.

Insights from the genome sequence
Average nucleotide identity was calculated by ANI calculator [28] with default parameters. ANIs between the genomes of strain 1807-2 T and T. pendens strain HRK 5 T and "T. adornatus" strain 1910b were 85 and 82 %, respectively which is below the species border, proposed to be 95 % [29]. These results, together with the data, obtained during 16S rRNA gene-based phylogenetic analysis and phenotypic differences, support the proposal of the novel species. Genome analysis of T. pendens revealed a massive loss of biosynthetic pathways, supporting the observation that T. pendens grows only in nutrient-rich environments [30]. To check if this is a common property of Thermofilaceae we analyzed the presence of crenarchaeal biosynthetic COGs [30] in in silico translated proteomes of all members of this family with currently available genome sequences (assembly accession numbers GCA_000015225.1, GCA_000446015.1, GCA_00 0813245.1, GCA_000993805.1). Analysis revealed that all Thermofilum representatives have only 20-30 of 125 biosynthetic COGs present in almost all Bacteria and Archaea (Additional file 2). Interestingly, the number of biosynthetic protein domains which have been lost during evolution negatively correlates with genome size of Thermofilaceae family members (Additional file 3). Comparative analysis of COG functional groups distribution didn't reveal any significant deviations showing general metabolic resemblance of Thermofilaceae (Additional file 4). The analysis of T. pendens genome [30] revealed the possibility of growing on various polysaccharides including starch and cellulose, however none of them were tested so far. T. uzonense is able to grow on starch and glucomannan, and the genes encoding corresponding glycosidases were found in its genome. In total, 17 genes are predicted to encode various glycosidases. Two of them are predicted to be extracellular: MA03_06285 (GH57) and MA03_03765 (GH3). The first one is putative alphaamylase, most probably responsible for extracellular starch hydrolysis, while the activity of the second one is not particularly obvious, but most likely, is a beta-glucosidase. Intracellular utilization of starch hydrolysis products could occur by means of four GH57 family glycosidases: MA03_00470, MA03_00770, MA03_05655 and one GH13 maltogenic amylase MA03_06280. Final intracellular hydrolysis of maltose could be catalyzed by alphaglucosidases MA03_03630 (GH4) and MA03_04210 (GH122). Glycosidases MA03_06285 and MA03_06280 are localized in a gene cluster together with transporters MA03_06265 (MalG) and MA03_06270 (MalF) and maltooligosaccharide-binding proteins MA03_06275 and MA03_06290 indicating their synergetic action. Enzymes presumably involved in glucomannan hydrolysis are MA03_02580 (DUF377 domain-containing putative glycosidase), MA03_03185 (GH38), MA03_04200 (GH1), MA03_02300 (GH113). All predictions of their localization (SignalP, TatP, SecP) did not reveal any signatures of extracellular proteins, indicating rather unusual motifs, recognized by signal peptidases, or another modifications in their secretion system than the intracellular state of all these enzymes. Interestingly, putative glycosidase MA03_02580 is located in the cluster of ABC and MFS families of putative sugar transporters and solute-binding proteins: MA03_2555-2575 and MA03_2585 indicating their possible co-regulation.
Two genes encoding proteins MA03_04265 and MA03_06125 are homologous to NAD-dependant oxidoreductases, some of which could be involved in degradation of glycosidic bonds in polysaccharides (GH109); however their function in T. uzonense is unclear.
Apart from glycoside hydrolases two genes encoding carbohydrate esterases (CEs) were also found: MA 03_03235 (CE14) family and MA03_04280 (CE9). The total is based on the total number of protein coding genes in the genome Characterized enzymes of these families are predicted to be involved in metabolism of N-acetyl-glucosamines and relative compounds. The genome contains more than 40 genes encoding peptidases of different families. Ten of them are predicted to be extracellular and four possibly are responsible for the hydrolysis of peptides during growth on peptone: subtilase (S08A family) MA03_03015, archaeal serine endopeptidase MA03_01720, thermopsin (A5 family) MA03_03850 and metalloendopeptidase (M48B family) MA03_08235.

Conclusions
Analysis of complete genome sequence of strain 1807-2 T indicates that it can be proposed as a novel species Thermofilum uzonense strain 1807-2 T . The majority of T. uzonense CDSs have homologs in the genomes of other Thermofilum members, and their distribution among COGs functional categories shows high level of metabolic homogeneity. As it was found for other Thermofilum representatives, massive reduction in number of proteins, contributing to biosynthetic pathways [30], has occurred in T. uzonense genome, hence, like other members of the family, its lifestyle could be characterized as opportunistic heterotroph, growing in nutrient-rich environments.
Apart from peptidic substrates, T. uzonense is able to grow on α-linked (starch) and β-linked (glucomannan) polysaccharides. While this capability was speculated in T. pendens genome paper [30], until now, it was never proven experimentally. Comparative analysis of all available Thermofilum genomes revealed two genes coding alpha-glucosidase of GH122 family and putative carbohydrate esterase of CE4 family, were present exclusively in genome of T. uzonense. Another two, coding extracellular Fig. 3 Circular map of Thermofilum uzonense strain1807-2 T generated with CIRCOS [41]. From outside to inside: positive strand CDSs colored by COG FCs (Clusters of Orthologous Genes Functional Categories), negative strand CDSs colored by COG FCs, RNA-genes and CRISPRs (tRNApurple; rRNAblue; riboswitchesgreen, CRISPR repeats -yellow), G+C-content, GC-skew. On the right -COG functional categories color codes putative amylase of GH57 family and putative endomannanase of GH113 family were found only in T. uzonense and "T. adornatus" genomes. These results suggest that the ability to hydrolyze various polysaccharides could be characteristic for Thermofilum representatives with some substrate-specific peculiarities of each individual strain. This allows one to conclude that Thermofilum members should not be regarded solely as consumers of simple organic substrates but also as primary destructors of the complex organic matter which occurs in their environments.

Taxonomic and nomenclatural proposals
Description of Thermofilum uzonense sp. nov.
Cells are non-motile thin straight or curved filaments, 0.15-0.3 μm in width and 2-100 μm in length. Strict anaerobe. Hyperthermophile growing optimally at 85°C and pH 6.0-6.5 in freshwater medium. Utilizes starch, glucomannan, peptone and yeast extract as the substrates. Amorphous cellulose, filter paper, mannan, amorphous chitin, glycerol and carbon monoxide do not support growth. Yeast extract and culture broth filtrate of other Crenarchaeota are required for growth. The type strain is 1807-2 T (= DSM 28062 T = JCM 19810 T ), was isolated from a mud sample of Uzon Caldera, Kamchatka (Russia). Genome size of the type strain is 1.6 Mb. The G+C content of DNA is 47.9 mol %. The genome sequence of the strain has been deposited in DDBJ/EMBL/GenBank under the accession number CP009961.

Emended description of the genus Thermofilum
The description is based on that provided by Zillig and colleagues [5], with the following amendments. The genus contains slightly or moderately acidophilic species utilizing peptides and polysaccharides and requiring polar component lipid from Thermoproteus tenax or culture broth of other Crenarchaeota for growth. Type species: Thermofilum pendens.