Introduction

Members of the genus Thermus are Gram-negative, rod-shaped, non-sporulating, thermophilic aerobic bacteria. They have been discovered from various environments with elevated temperatures, including hot springs, deep-sea hot vents, volcanic eruptions and solfatara fields [1–4]. Thermus aquaticus was first isolated in 1969 in hot springs in Yellowstone National Park, USA [5]. Thermus species and their produced enzymes, so called extremozymes, have attracted the attention of scientists from academia and industry due to their unique properties and metabolic pathways. Robust biocatalysts are attractive to various applications that often prevail in industrial processes [6–8]. The most prominent example of an industrial-relevant extremozyme is the DNA polymerase from T. aquaticus that is applied in polymerase chain reaction. Further industrial applications using enzymes from Thermus species include laundry detergents, DNA clean up prior to PCR or C-terminal sequencing [9–12]. Recently two glycoside hydrolases from T. antranikianii and T. brockianus were described and extended this group of industrial-relevant enzymes [13].

T. brockianus strain GE-1 was chosen for whole genome sequencing due to its ability to use xylan as sole carbon source and degrade xylan-rich substrates (Blank and Antranikian, unpublished results) [14]. To our knowledge the hydrolysis of xylan has not been described for any other T. brockianus strain so far, including type strain YS038T [15]. With the description of the corresponding thermostable xylanase, Xyn10, we already identified and characterized one of the key enzymes in a putative xylan degradation pathway of T. brockianus GE-1 [14]. The identification and characterization of other polymer degrading enzymes from Thermus species is of great interest since there are only few reports regarding this aspect [16, 17]. Especially in the view of finding new solutions for global challenges like degradation of xenobiotic compounds or providing novel renewable energy sources, the xylanolytic behavior of T. brockianus GE-1 justifies further examination. These findings will also contribute to the development of biotechnological processes based on lignocellulose as carbon source (biorefinery). In this paper we present the first whole genome sequence of a T. brockianus strain with finished grade status, showing a phred quality value of QV50.

Organism information

Classification and features

T. brockianus type strain YS038T has been described as a member of the family Thermaceae within the phylum Deinococcus-Thermus . The isolate GE-1 could be clearly assigned to the species T. brockianus based on sequence information and 16S rRNA pairwise alignment, indicating 100% identity in 1,476 nucleotides overlap. The phylogenetic classification of T. brockianus GE-1 within the Deinococcus-Thermus group is displayed in the phylogenetic tree shown in Fig. 1. Further alignments with closely related species of the genus Thermus were performed using the webserver LALIGN [18] and revealed the following results: T. igniterrae strain RF-4 T (96.8% identity in 1,477 nts overlap), T. aquaticus strain YT-1 (96.1% identity in 1,474 nts overlap), T. composti strain K-39 (96.1% identity in 1,445 nts overlap), T. islandicus strain PRI-3838 (95.9% identity in 1,445 nts overlap), T. arciformis strain TH92 (95.7% identity in 1,484 nts overlap), T. caliditerrae strain YIM 77925 (95.0% in 1,514 nts overlap), T. amyloliquefaciens strain YIM 77409 (94.9% identity in 1,513 nts overlap), T. scodoductus (94.9% identity in 1,476 nts overlap), T. thermophilus strain HB8 (94.9% identity in 1,518 nts overlap), T. antranikianii strain HN3-7 T (94.7% identity in 1,482 nts overlap), T. filiformis (94.0% in 1,475 nts overlap) and T. oshimai SPS-17 T (91.4% identity in 1,477 nts overlap).

Fig. 1
figure 1

Unrooted phylogenetic tree based on 16S rRNA encoding sequences from 14 species of the genus Thermus. The phylogenetic tree was generated using the program package PHYLIP (version 3.695) [58] and TreeView X [59], based on a multiple sequence alignment (1,345 nts) that was generated with clustalX [60]. The number of nucleotide replacements at each position in the sequence was estimated with the DNADIST program and trees were constructed using NEIGHBOR. Bootstrap analysis was done using 1,000 iterations. CONSENSE was used to produce a majority rule consensus tree. The position of the isolate Thermus brockianus strain GE-1 is indicated in red. The 16S rRNA encoding sequence from Marinithermus hydrothermalis was used as outgroup. Accession numbers of all sequences are indicated in the figure. For the following species sequenced genomes are available at NCBI (number of available genome sequences are given in square brackets): T. caliditerrae [1], T. amyloliquefaciens [1], T. antranikianii [1], T. scotoductus [4], T. igniterrae [1], T. brockianus [1, this study], T. aquaticus [3], T. islandicus [1], T. thermophilus [5], T. filiformis [1], T. oshimai [2] and M. hydrothermalis [1]

Full size image

T. brockianus strain GE-1 is a Gram-negative, rod-shaped, non-pathogenic and non-sporulating bacterium (Fig. 2). This strain is aerobic, yellow pigmented and non-motile (Table 1). Due to their thermophilic characters, Thermus species are capable of thriving at elevated temperatures in a range between 45 °C and 83 °C and most species show an optimal growth at 80 °C [19]. While most of the species within the Thermus genus were reported to degrade a diverse set of sugars, including D-glucose, D-fructose, D-galactose, D-ribose, lactose and sucrose, only a few strains such as T. brockianus YS038T or T. thermophilus HB8 were described to utilize D-xylose [15]. However growth on xylan was only reported for the strain T. brockianus GE-1 and a corresponding xylanase-encoding gene has been identified within its genome (Blank and Antranikian, unpublished results) [14]. T. brockianus strain GE-1 was isolated from the Geysir geothermal area, Iceland. It optimally grows at a temperature of 70 °C and 160 rpm. The minimum information about the genome sequence (MIGS) [20] of that strain is shown in Table 1.

Fig. 2
figure 2

Photomicrograph of T. brockianus GE-1

Full size image
Table 1 Classification and general features of T. brockianus GE-1 according to MIGS [20]
Full size table

Genome sequencing information

Genome project history

We conceived the whole de novo genome sequencing of T. brockianus GE-1 because of its ability to degrade xylan-rich biomass that has not been described for any other T. brockianus species so far and thus it has great potential for application in future biorefineries. Comparison of its genome sequence to that of other sequenced Thermus species will also help to understand general molecular features of xylan degradation in thermophiles. Sequencing was done at GATC Biotech AG (Konstanz, Germany). The de novo assembly, annotation and finishing of the whole genome of T. brockianus GE-1 was performed at the Institute of Technical Microbiology at Hamburg University of Technology (TUHH). The finished genome sequence, including three circular replicons, has been submitted to National Center of Biotechnology Information (NCBI) in June 2016. A summary of the project information is shown in Table 2.

Table 2 Project information
Full size table

Growth conditions and genomic DNA preparation

T. brockianus strain GE-1 was obtained from the strain culture collection of the Institute of Technical Microbiology at Hamburg University of Technology (TUHH). Deposition of the strain in the German National Culture Collection (DSMZ) is in progress. The strain was grown aerobically in DSMZ medium 878 ( Thermus 162 medium) at 70 °C for at least 48 h and agitation speed of 160 rpm [19]. The genomic DNA of T. brockianus GE-1 was isolated using the PowerSoil DNA Isolation Kit (Mobio, USA). All steps were performed according to the manufacture’s instructions. Quality control of the isolated DNA was checked at GATC Biotech AG (Konstanz, Germany) prior to sequencing. A DNA concentration of 83.1 ng/μl and a 260/280 ratio of 1.87 were determined.

Genome sequencing and assembly

Third generation sequencing technology from Pacific Biosciences was chosen for whole de novo genome sequencing of T. brockianus strain GE-1 because its continuous long reads of up to 10 kb covering the longest known bacterial and archaeal repetitive regions and thus facilitate the generation of complete bacterial genome assemblies [21, 22]. Library construction, quantification and sequencing were performed at GATC Biotech AG (Konstanz, Germany). A Pacbio RS library (8–12 kb) was constructed and one SMRT cell was used for sequencing. 86,479 subreads were obtained after filtering and a total of 447.6 Mb with a N50 contig length of 2,058,948 bp were used for assembly. Pacific Biosciences sequencing data were assembled using an implemented version of PacBio SMRT Analysis, version 2.3.0 and the HGAP2 protocol (Pacific Biosciences, USA) [22]. Minimum seed read length was automatically determined by the protocol with a length cut-off of 10,819 bp. The sum of contig lengths was 2,431,825 bp. The final de novo assembly obtained three circular contigs, providing the complete genome sequence of T. brockianus strain GE-1 and genome coverage of 156.56. Each contig represented one replicon, including the chromosome (2,035,182 bp), the megaplasmid pTB1 (342,792 bp) and plasmid pTB2 (10,299 bp). Circularization of each replicon was checked and performed by using circlator [23]. Quality value of > QV50 (1 error probability in 100,000 base calls) for each replicon was reached after several polishing steps using the quiver algorithm, included in PacBio SMRT Analysis, version 2.3.0 [22].

Genome annotation

Preliminary genome annotation was performed using the Prokka annotation pipeline v1.12 [24], followed by manual curation. Genes were identified by both Prodigal v2.6.3 [25] and Glimmer v3.0.2 [26]. Predicted coding sequences were translated and used to search the NCBI non-redundant database [27], UniProt [28] and Pfam [29] databases. The cmmscan and cmmsearch tools of the Hmmer3 package were used for protein similarity searches against HMM databases [30]. For COG classification RPS-BLAST was used to search against the COG database [31]. For rRNA detection we used RNAmmer v1.2 [32], while the tRNA prediction was performed by tRNAscan-SE v1.3.1 [33]. Non-coding RNAs and regulatory RNA features were identified by searching the genome for corresponding Ram profiles using INFERNAL v1.1.1 [34]. Signal peptides were identified by Signalp v4.1 [35] and clustered regularly interspaced short palindromic repeats (CRISPR) were detected by using MinCED v0.2.0 included in the Prokka annotation pipeline [36]. Analyses to identify genes that were assigned to transmembrane domains were performed by using tmhmm [37]. Circular maps were created using CGView [38].

Genome properties

The complete genome of T. brockianus GE-1 is composed of a single circular chromosome that consists of 2,035,182 bp and two circular plasmids, pTB1 and pTB2 (Fig. 3). The megaplasmid pTB1 has got 342,792 bp and the size of pTB2 is 10,299 bp (Table 3). The GC content of each replicon is 67.07 for the chromosome, 65.82 for megaplasmid pTB1 and 70.09 for plasmid pTB2. The T. brockianus GE-1 genome comprises 2,511 genes, including 2,458 protein-encoding genes (97.9%), 53 RNAs (2.1%) and 66 pseudo genes (2.6%) (Table 4). In more detail there are 2,053 protein-encoding sequences for the chromosome, 299 for megaplasmid pTB1 and 13 for plasmid pTB2. Classification of genes into the COG functional categories is shown in Table 5.

Fig. 3
figure 3

Graphical circular maps of T. brockianus GE-1 replicons. The complete genome of T. brockianus GE-1 is composed of a single circular chromosome that consists of 2,035,182 bp (a) and two circular plasmids, pTB1 (b) and pTB2 (c). The size of megaplasmid pTB1 is 342,792 bp and 10,299 bp for pTB2. These maps were generated by using CGView [38]. Data shown on those maps will be explained from the inside to the outside: Second circle represents the GC skew of both strands (green for plus strand, purple for minus strand) and the fourth circle shows the GC content. The sixth and seventh circle exhibits the protein-encoding genes for the plus and minus strand as well as RNA features. All tRNAs are highlighted in orange, rRNAs are shown in light purple and other RNAs are represented by a grey color

Full size image
Table 3 Summary of the genome of Thermus brockianus GE-1: 1 chromosome and 2 plasmids
Full size table
Table 4 Genome statistics
Full size table
Table 5 Number of genes associated with general COG functional categories
Full size table

Insights from the genome sequence

Comparison of genomes from T. brockianus GE-1 and other Thermus species

Based on the availability of their finished genomes within the NCBI genome database we compare the T. brockianus GE-1 genome with six other Thermus species and seven genomes, including T. thermophilus HB8, T. thermophilus HB27 [39], T. scotoductus SA-01 [40], Thermus sp. CCB_US3_UF1 [41], T. oshimai JL-2 [42], T. aquaticus Y51MC23 [43] and T. parvatiensis [44]. With 2.38 Mb the genome of T. brockianus GE-1 is the largest one of these finished genomes, close to the genomes of T. oshimai JL-2 (2.33 Mb), T. aquaticus Y51MC23 (2.34 Mb) and T. scotoductus SA-01 (2.36 Mb) and much bigger than Thermus sp. CCB_US3_UF1 (2.26 Mb), T. thermophilus HB8 (2.12 Mb), T. thermophilus HB27 (2.13 Mb) and T. parvatiensis (2.03 Mb). All of those finished genomes include a chromosome and at least one plasmid. The genome of T. brockianus GE-1 consists of one chromosome (2.04 Mb) and two plasmids, including megaplasmid pTB1 (0.34 Mb) and plasmid pTB2 (10 kb). In number and size of those replicons the genome of T. brockianus GE-1 is similar to T. thermophilus HB8 having a chromosome (1.85 Mb), the megaplasmid pTT27 (0.26 Mb) and the plasmid pTT8 (9.3 kb) as well as T. oshimai JL-2 with one chromosome (2.07 Mb), one megaplasmid pTHEOS01 (0.27 Mb) and one plasmid pTHEOS02 (6 kb). Megaplasmids are a common feature of Thermus spp., since they were also identified in T. thermophilus HB27 (pTT27; 0.23 Mb) and T. parvatiensis (pTP143; 0.14 Mb). Due to their thermophilic lifestyle, all finished Thermus genomes exhibit a high GC content varying between 64.9% for T. scotoductus SA-01 and 69.5% for T. thermophilus HB8 with an average value of 67.0% GC content for the genome of T. brockianus GE-1.

With its total gene number of 2,511 and 2,458 protein-encoding genes, the genome of T. brockianus GE-1 showed high-ranking numbers in comparison to the available genomes, comparable to T. oshimai JL-2 (2,580 in total and 2,436 protein-encoding genes), T. scotoductus SA-1 (2,511 and 2,458), T. aquaticus Y51MC23 (2,484 and 2,325) and higher than Thermus sp. CCB_US3_UF1 (2,333 and 2,279), T. thermophilus HB8 (2,226 and 2,173), T. thermophilus HB27 (2,263 and 2,210) and T. parvatiensis (1,573 and 2,190). The genome of T. brockianus GE-1 encodes 47 tRNA genes and 6 rRNA genes, similar to most of the other finished genomes. Additionally, eight clustered regularly interspaced short palindromic repeats (CRISPRs) were identified in the genome of T. brockianus GE-1, suggesting the presence of a defense mechanism against phage DNA invasion, equivalent to other finished Thermus genomes with reported CRISPR sequences, including Thermus sp. CCB_US3_UF1 (8), T. thermophilus HB8 (11) and HB27 (10) and T. scotoductus SA-01 (3) [41].

By whole genome comparison analyses we identified the highest number of protein orthologs in the genomes of T oshimai (85.86%) and T. aquaticus (85.34%). These two genomes shared 2,156 and 2,143 of the 2,511 total proteins with T. brockianus GE-1. Further comparisons revealed 83.07% protein orthologs in comparison to Thermus sp. CCB_US3_UF1 (2,086 of 2,511), 82.44% to T. scotoductus (2,070 of 2,511), 81.96% to T. thermophilus HB8 (2,058 of 2,511) and 81.76% to T. thermophilus HB27 (2,053 of 2,511). With 1,661 and 66.15% we identified the lowest numbers of protein orthologs between T. parvatiensis and T. brockianus GE-1.

The number of total and protein-encoding genes on megaplasmid pTB1 were 314 and 299 as well as 13 for both in case of pTB2. Especially, the number of genes on the megaplasmid pTB1 is much higher compared to other available megaplasmids, since their total gene numbers vary between 150 (T. parvatiensis) and 251 ( T. thermophilus HB8). These differences are explicable by the smaller size of both megaplasmids (0.14 Mb for pTP143 and 0.26 MB for pTT27) in comparison to pTB1 (0.34 Mb). In contrast, the size of pTB2 (10 kb) is smaller than most other additional plasmids, which were reported to be 6–60 kb.

General metabolic pathways were investigated by KEGG analysis and revealed complete sets of genes for glycolysis, gluconeogenesis, citrate cycle, pentose phosphate pathway as well as genes involved in the lipid-, nucleotide-, amino acid-, cofactor- and vitamin-metabolism. For nutrient uptake, we identified 14 ABC transporters. All of these genes were localized on the chromosome of T. brockianus GE-1.

Xylan degradation pathway

Thermophilic bacteria like T. brockianus are of great industrial relevance, because they produce heat-stable and heat-active enzymes, so called thermozymes that perfectly match harsh process conditions. With regard to biocatalysts with a great potential for biotechnological processes, e.g. biorefinery, we identified sequences encoding putative lipases, subtilisin-like proteases, glucosidases and galactosidases in the genome of T. brockianus GE-1. The observed ability of this strain to degrade xylan-rich substrates and the identification of the responsible xylanase Xyn10 in our previous study directed our interest towards the investigation of the xylan degradation pathway by performing an analysis of the whole genome sequence of T. brockianus GE-1 [14]. Interestingly, there is no homologue xylanase sequence detectable in any other Thermus genome that is currently available. Thus, the amino acid sequence of the xylanase Xyn10 from T. brockianus strain GE-1 displays the highest identity (57%) to a 1,4-β xylanase from Streptomyces sp. NRRL WC-3723. These two organisms are not closely related but it can be hypothesized that due to similar environmental conditions an ancestor of T. brockianus obtained the coding sequence of xyn10 by horizontal gene transfer. Another interesting fact to consider is that the localization of the corresponding gene xyn10 was not detected on the bacterial chromosome like other hydrolases, including lipases, peptidases and ATPases.

Further investigations of the genomic region revealed an unique set of genes related to cellulose degradation and xylose metabolism adjacent to xyn10, including a putative endoglucanase as well as membrane ABC sugar transporter encoded by xylH, xylF encoding the xylose binding protein, the xylose repressor encoding gene xylR, the xylose isomerase encoding gene xylA and the gene xylB coding for a xylulokinase (Fig. 4). Especially the latter ones are of great importance, since XylA catalyzes the first step of the xylose metabolism by isomerization of xylose to xylulose, while XylB is responsible for the second step, phosphorylating xylulose to xylulose-5-phosphate [45, 46]. These xylose metabolism-related genes are conserved in other Thermus spp. genomes and a similar set has been described for plasmid pVV8 in the genome of T. thermophilus HB8 [46, 47]. Genome comparison studies of pVV8 and pTB1 revealed 75% identity of the genomic region encoding the xylose metabolism-related genes. However, the xylanase, the endoglucanase and the ABC transporter system-associated genes are described for the first time in a Thermus species genome. These genes showed highest similarities (43% to 57%) to distantly related genera including Streptomyces and Alicyclobacillus . The absence of a β-xylosidase in the genome of T. brockianus GE-1 is explicable by the β-xylosidase side activity of xylanase Xyn10, thus breaking down xylan directly to D-xylose [14]. Altogether the localization of this whole set of genes on the megaplasmid pTB1 in combination with the non-essential character of the xylan degradation pathway for the lifestyle of T. brockianus GE-1 indicates that T. brockianus GE-1 has gained the xylanolytic ability as a beneficial advantage, probably via horizontal gene transfer in an ancestor of T. brockianus . In accordance to the bioconversion of lignocellulosic biomass, another interesting point to consider is the identification of a putative endoglucanase. By linking genes encoding endoglucanase and xylanase on its megaplasmid pTB1, these enzymes might act in concert in a synergistically fashion. Especially, in regard to a recent report of an evolved strain of T. thermophilus co-utilizing xylose and glucose [16], T. brockianus GE-1 and its unique genomic linkage of sequences encoding key enzymes for xylan and cellulose degradation as well as for xylose metabolism seems to be of great interest for biotechnical applications and thus will be examined in prospective studies.

Fig. 4
figure 4

Genomic organization of genes encoding proteins for xylan and cellulose degradation as well as xylose metabolism located on the megaplasmid pTB1 of T. brockianus GE-1. Sizes, localization and orientation of the genes on megaplasmid pTB1 section are displayed proportionally. All genes highlighted with a star are not detectable in any other Thermus spp. genome except T. brockianus GE-1. Genes marked with a diamond are conserved in Thermus spp. ABC transporter system associated genes include sugar ABC transporter substrate-binding protein and two sugar ABC transporter permeases

Full size image

Conclusions

Thermus spp. and their extremozymes are of great interest for a wide set of industrial applications. Here we present the first whole genome sequence of T. brockianus GE-1, providing further insights into the biotechnological potential of the genus Thermus spp. in general and T. brockianus GE-1 specifically. The genome of T. brockianus GE-1 consists of a chromosome and two plasmids, including the megaplasmid pTB1. Sequences coding for essential metabolism pathways like glycolysis, gluconeogenesis, pentose phosphate pathway or citrate cycle were assigned to the bacterial chromosome just as well as sequences encoding industrial relevant enzymes, including galactosidases, glucosidases, lipases and subtilisin-like proteases. These novel extremozymes will be targets of prospective characterization studies to prove their industrial relevance. However, localization of gene xyn10 coding for a previously described xylanase from T. brockianus GE-1 was not detected on the chromosome but on the megaplasmid pTB1 adjacent to sequences encoding key enzymes for cellulose degradation and xylose metabolism. Thus, in accordance to a reported β-xylosidase side activity of xylanase Xyn10 the complete breakdown of xylan to D-xylose is genetically linked to the xylose metabolism in the genome of T. brockianus GE-1. These findings are consistent with the described xylanolytic activity of T. brockianus GE-1. The described combination of the identification of novel sequences encoding putative biocatalysts on the one hand and the description of a xylanolytic degradation pathway on the other hand emphasize the importance of Thermus spp. as promising sources of extremozymes with potential industrial value.