Permanent draft genome of Thiobacillus thioparus DSM 505T, an obligately chemolithoautotrophic member of the Betaproteobacteria

Thiobacillus thioparus DSM 505T is one of first two isolated strains of inorganic sulfur-oxidising Bacteria. The original strain of T. thioparus was lost almost 100 years ago and the working type strain is Culture CT (=DSM 505T = ATCC 8158T) isolated by Starkey in 1934 from agricultural soil at Rutgers University, New Jersey, USA. It is an obligate chemolithoautotroph that conserves energy from the oxidation of reduced inorganic sulfur compounds using the Kelly-Trudinger pathway and uses it to fix carbon dioxide It is not capable of heterotrophic or mixotrophic growth. The strain has a genome size of 3,201,518 bp. Here we report the genome sequence, annotation and characteristics. The genome contains 3,135 protein coding and 62 RNA coding genes. Genes encoding the transaldolase variant of the Calvin-Benson-Bassham cycle were also identified and an operon encoding carboxysomes, along with Smith’s biosynthetic horseshoe in lieu of Krebs’ cycle sensu stricto. Terminal oxidases were identified, viz. cytochrome c oxidase (cbb3, EC 1.9.3.1) and ubiquinol oxidase (bd, EC 1.10.3.10). There is a partial sox operon of the Kelly-Friedrich pathway of inorganic sulfur-oxidation that contains soxXYZAB genes but lacking soxCDEF, there is also a lack of the DUF302 gene previously noted in the sox operon of other members of the ‘Proteobacteria’ that can use trithionate as an energy source. In spite of apparently not growing anaerobically with denitrification, the nar, nir, nor and nos operons encoding enzymes of denitrification are found in the T. thioparus genome, in the same arrangements as in the true denitrifier T. denitrificans.


Introduction
In 1902, Thiobacillus thioparus was one of the first two obligately chemolithoautrophic sulfur-oxidizing Bacteria to be isolated, (along with what is now Halothiobacillus neapolitanus), and was named in 1904 [1,2]. The original isolates were lost, but Thiobacillus thioparus is now the type species of the genus (a member of the Betaproteobacteria) and the type strain is an isolate from Starkey (1934) (= Culture C T = DSM 505 T = ATCC 8158 T = CIP 104484 T = JCM 3859 T = NBRC 103402 T ) [3,4]. Originally, the characteristic of utilizing inorganic sulfur compounds as an energy source was thought to be a taxonomic trait unique to Thiobacillus and at its height the genus contained in at least 32 different 'species' [4]. With phylogenetic methods however, many of these strains have since been reassigned to different, often new genera with T. thioparus DSM 505 T being one of four species with validly published names left in the genus. Compared to other inorganic sulfur-oxidisers, surprisingly little research has been conducted on T.
thioparus DSM 505 T in terms of physiology and biochemistry or genetics, possibly due to the low growth yields of this species when compared to T. denitrificans and T. aquaesulis [5,6] making it more challenging to study. It may, however, give extended and contrasted insights into autotrophic sulfur-oxidation in the Bacteria. It was selected for genome sequencing as part of the Department of Energy DOE-CSP 2012 initiativeas type species of a genus.

Classification and features
Thiobacillus thioparus DSM 505 T was isolated from sandy loam soil from the New Jersey Agricultural Experimental Station planted with unspecified vegetable crops using 20mM thiosulfate as sole energy source in basal medium at pH 8.5 by Starkey (1934) [3] and is currently one of four species with validly published names within the genus. It forms small white colonies of 1-3 mm diameter after 2-3 days that turn pink or brown with age and which become coated with yellow-ish elementary sulfur. When grown in liquid media, finely divided (white) elementary sulfur is formed during early stages of growth, particularly when thiosulfate is used as the energy source. This disappears when growth approaches stationary phase. Thiosulfate is oxidized stoichiometrically to tetrathionate after 24 h accompanied by a rise in pH, characteristic of the Kelly-Trudinger pathway. Tetrathionate is subsequently oxidized to sulfate with culture pH falling to pH 4.8 by stationary phase. During continuous culture, no intermediates are detected in the medium once steady-state has been established. If the dilution rate of a thiosulfate limited chemostat is increased, a large production of elementary sulfur is observed, which disappears as a new steadystate is established at the faster dilution rate. General features of T. thioparus DSM 505 T are summarized in Table 1. A phylogenetic tree based on the 16S rRNA gene sequence showing this organisms position within the Betaproteobacteria and rooted with Thermithiobacillus tepidarius is given in Fig. 1.
Cells are 1.5-2.0 by 0.6 to 0.8 μm and stain Gram negative. They are motile by means of a single polar flagellum between 6 and 10 μm in length, as shown in Fig. 2. The dominant respiratory quinone is ubiquione-8 [7][8][9][10] and fix carbon dioxide using the Calvin-Benson-Bassham cycle at the expense of inorganic sulfur oxidation. Cells accumulate polyphosphate ('volutin') granules in batch culture and to a lesser extent in an energylimited chemostat [8]. Carboxysomes are present in cells regardless of growth conditions employed (Fig. 1), but are apparently not found in T. denitrificans cells, at least not under growth conditions employed in previous studies [9]. T. thioparus can only use oxygen as a terminal electron acceptornitrate, nitrite, thiosulfate, sulfate, elementary sulfur and ferric iron do not support growth as terminal electron acceptors. The genomic DNA G + C content has been estimated using the thermal denaturation [11] at 61-62 mol% [7,10]. T. thioparus DSM 505 T does not grow on any organic carbon compound tested, including sugars (glucose, ribose, fructose, sucrose), intermediates of Krebs' cycle (citrate, succinate, fumarate, malate, oxaloacetate), carboxylates (glycolate, formate, acetate, propionate, pyruvate), one-carbon compounds (monomethylamine, dimethylamine, trimethylamine, methanol, methane), structural amino acids (all 20), substituted thiophenes (thiophene-2-carboxylate, thiophene-3-carboxylate) or complex media (yeast extract, nutrient broth, brain-heart infusion, Columbia sheep or horse blood agar, chocolate agar). Energy sources that support autotrophic growth of DSM 505 T include thiosulfate, trithionate, tetrathionate, pentathionate, hexathionate, thiocyanate and dithionate. Some bone fide strains of T. thioparus (Tk-m and E6) grow autotrophically on carbon disulfide, dimethylsulfide, dimethyldisulfide and Admidate (O,O-dimethylphosphoramidothioate) [4], but it is not known if the type strain DSM 505 T is capable of this. Autotrophic growth is not supported by Fe(II), Mn(II), Cu(I), U(IV), sulfite, dimethylsulfoxide, dimethylsulfone, pyrite or formate. During batch growth on thiosulfate the intermediate production of tetrathionate is observed during early stages of growth, indicative of the Kelly-Trudinger pathway [12]. Compared to the other two members of the genus [5,6], T. thioparus DSM 505 T has relatively low growth yields on thiosulfate (Hutt and Boden, manuscript in preparation) which may give insight into the physiological variances of Kelly-Trudinger pathway organisms even within one genus.

Genome project history
This organism was selected for sequencing on the basis of its role in sulfur cycling, physiological, biochemical, evolutionary and biogeochemical importance, and is part of the GEBA-KMG project at the U.S. Department of Energy JGI. The genome project is deposited in the Genomes OnLine Database [13] and a high-quality permanent draft genome sequence in IMG (the annotated genome is publically available in IMG under Genome ID 2515154076) [14]. Sequencing, finishing and annotation were performed by the JGI using state of the art sequencing technology [15]. A summary of the project information is given in Table 2.
Growth conditions and genomic DNA preparation T. thioparus DSM 505 T DNA was obtained from Dr Hans-Peter Klenk at the DSMZ, having been grown on basal salts medium pH 6.6, supplemented with 40 mM thiosulfate as the sole energy source (DSM Medium 36), under air at 26°C for 72 h. DNA was extracted using the JETFLEX Genomic DNA Purification Kit from Genomed (Löhne, Germany) into TE Buffer. Quality was checked by agarose gel electrophoresis.

Genome annotation
Genes were identified using Prodigal [19], followed by a round of manual curation using GenePRIMP [20] for finished genomes and Draft genomes in fewer than 10 scaffolds. The predicted CDSs were translated and used to search the NCBI nonredundant database, UniProt, TIGR-Fam, Pfam, KEGG, COG, and InterPro databases. The tRNAScanSE tool [21] was used to find tRNA genes, whereas ribosomal RNA genes were found by searches against models of the ribosomal RNA genes built from SILVA [22]. Other non-coding RNAs such as the RNA components of the protein secretion complex and the RNase P were identified by searching the genome for the corresponding Rfam profiles using INFERNAL [23]. Additional gene prediction analysis and manual functional annotation was performed within the IMG platform [24] developed by the JGI, Walnut Creek, CA, USA [25,26].

Genome properties
The genome of T. thioparus DSM 505 T is 3,201,518 bplong with a 62.3 mol% G + C content (Table 3). Of the 3,197 predicted genes, 3,135 were protein-coding genes a Evidence codes -IDA Inferred from Direct Assay, TAS Traceable Author Statement (i.e., a direct report exists in the literature), NAS Non-traceable Author Statement (i.e., 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 [51,52] and 62 were RNA genes. A total of 2,597 genes (81.23%) have predicted function. A total of 538 (16.83%) were identified as pseudogenesthe remainder annotated as hypothetical proteins. The properties and the statistics of the genome are given in Table 3, the distribution of genes into COG functional categories is given in Table 4. The genome is the second largest genome of obligate chemolithoautotrophs sequenced to date and is 89% (bp/bp) or 90% (protein coding genes/protein coding genes) of the size of that of T. denitrificans DSM 12475 T [12].  [29], which varies between Classes of the 'Proteobacteria' , for example, genes for fumarase, succinate dehydrogenase and the E1 subunit of  Sequences pertaining to organisms for which a publically available genome sequence exists are underlined. Accession numbers for the GenBank database are in parentheses. Alignment and tree were constructed in MEGA 6 [53]. Tree was drawn using the Tamura-Nei model for maximum-likelyhood trees [54].

Insights from the genome sequence
Values at nodes are based on 5,000 bootstrap replicates, with values <70% omitted. Scale-bar indicates 2 substitutions per 100. Thermithiobacillus tepidarius DSM 3134 T from the Acidithiobacillia is used as the outgroup α-ketoglutarate dehydrogenase were recently found to be missing from Thermithiobacillus tepidarius DSM 3134 T [12,30,31] of the Acidithiobacillia. Activities of Krebs' cycle enzymes assayed by Smith et al. [29] in cellfree extracts of T. thioparus found activities of all enzymes except for α-ketoglutarate dehydrogenase, indicating the E1 and E2 subunits alone were not sufficient for activity. Interestingly, Smith detected fumarase activity; however, this strain of T. thioparus was isolated by the authors themselves and was not DSM 505 T , so there may be further variation of Smith's horseshoe at strain level. With many species having been erroneously classified as strains of T. thioparus in the past that have subsequently been proven to belong to other genera [4] it is also plausible that Smith's strain was from another specie, genus or even Class. The E1, E2 and E3 subunit genes for α-ketoglutarate dehydrogenase were identified in the genome of T. denitrificans ATCC 25259 T [32], though enzyme activity was also absent in this strain [29,33]. It is important to stress that the full suite of enzymes of Krebs' cycle or Smith's horseshoe have never been assayed in T. thioparus DSM 505 T -clearly this is needed in order to identify if genomic data reflect true activities in vivo [27].

Not in COGs
The total is based on the total number of protein coding genes in the genome variant Calvin-Benson-Bassham cycle [34]. Adjacent to these genes are cbbO and cbbQ, consistent with form IAq RuBisCO [35], which is canonically cytoplasmic, rather than carboxysomal. A cluster of 11 genes that encode carboxysome shell proteins and a carboxysome carbonic anhydrase were also identified. This cluster is located on the forward strand while on the reverse strand the RuBisCO cluster is located c.9 kb upstream, between which is a divergently transcribed transcriptional regulator (cbbR) [32]. Further evidence of carboxysome expression can be seen in transmission electron micrographs (Fig. 1). The carboxysome gene cluster (experimental data would be required to demonstrate if it is an operon or not) in T. thioparus DSM 505 T does not start with cbbLS, (as would be found in Halothiobacillus, Acidithiobacillus and Thermithiobacillus spp. [36]) -the lack of these genes has also been observed in T. denitrificans ATCC 25259 T [32] and may be a diagnostic property of this genus. Whilst T. thioparus DSM 505 T cannot grow anaerobically with denitrification but T. denitrificans does, putative operons encoding the nitrate reductase (nar), nitrite reductase (nir), nitric oxide reductase (nor) and nitrous oxide reductase (nos) proteins of canonical denitrification are found in both T. denitrificans and T. thioparusthis may indicate a potential for the latter to grow with denitrification albeit not under conditions previously employed for example, it may occur only under micro-oxic conditions rather than fully anoxic conditions, alternatively, some unknown factor may prevent detection of nitrate and/or the expression of these genes. No evidence for any alternative denitrification pathways [37] was found With regard to respiration, five cytochromes c 553 , one cytochrome c 556 and three cytochrome b were present. The two high-affinity terminal oxidases were both foundwith multiple gene copies of cytochrome c oxidase (cbb 3 , EC 1.9.3.1) and a single copy of cytochrome bd-type quinol oxidase (EC 1.10.3.10). It is worth noting that T. denitrificans also possess the aa 3 variant of cytochrome c oxidase (EC 1.9.3.1), which may permit it greater metabolic diversity, for example, growth under more variable oxygen partial pressures -this could potentially explain why T. thioparus cannot denitrify under anoxic conditions, even though it possesses the operons encoding enzymes of denitrification. It is known that microaerophilic organisms usually employ cbb 3 or bd-type high-affinity terminal oxidases [37] and this may further evidence that T. thioparus and T. denitrificans may grow under such conditions, potentially with denitrification, though we cannot find any studies that demonstrate this in vivo thus far.

Extended insights
Enzymes of the Kelly-Trudinger pathway remain poorly understood and many of the genes are yet to be identified [12]. The oxidation of thiosulfate to tetrathionate takes place via a cytochrome c-linked thiosulfate dehydrogenase (EC 1.8.2.2), one gene (tsdA) for which was identified in Allochromatium vinosum, a member of the Gammaproteobacteria [38]. However, tsdA is not present in T. thioparus DSM 505 T supporting the hypothesis that more than one thiosulfate dehydrogenase is present in the 'Proteobacteria' or could vary at Class level. Numerous Kelly-Friedrich pathway genes (soxXYZAB) were present but the remaining of the conserved soxTRS-VW-XYZABCDEFGH genes being absent [39][40][41]. In Paracoccus spp. (Alphaproteobacteria) soxYZ encodes a protein complex that binds thiosulfate via a cysteine residue in the initial stage of the Kelly-Friedrich pathway while soxXA encode cytochromes c 551 and c 552.5 which capture two electrons from thiosulfate oxidation. Finally soxB encodes a hydrolase that removes the terminal sulfone group as sulfate. Missing the SoxCD or sulfur dehydrogenase protein from the multi-enzyme system would leave a sulfur atom attached to the SoxYZ residue, preventing the action of SoxB to liberate this residue to act further with another thiosulfate molecule. An unidentified protein may participate in the release of these sulfur atoms and potentially may explain the deposits of sulfur seen during initial stages of growth. This does not explain the production of tetrathionate which still would require the activity of thiosulfate dehydrogenase. If soxXYZAB genes are being expressed then they may be required as a functional part of the Kelly-Trudinger pathway. The action of Sox proteins (if any) in T. thioparus DSM 505 T in conjunction and potentially collaboration with additional Kelly-Trudinger pathway proteins would undoubtedly be essential in resolving its chemolithoautotrophic metabolism, which remains poorly understood. An unidentified gene encoding a putative DUF302-family protein is present between soxA and soxB genes of Thermithiobacillus tepidarius DSM 3134 T , Acidithiobacillus thiooxidans ATCC 19377 T and Acidithiobacillus caldus ATCC 51756 T , and Thiohalorhabdus denitrificans DSM 15699 T , the function of which may be important in the Kelly-Trudinger pathway [12]; however, DUF302 is not present on the sox operon of T. thioparus DSM 505 T , although six unidentified genes annotated as DUF302 family proteins are present elsewhere in the genome. The soxEF genes encode and a flavocytochrome c sulfide dehydrogenase (EC 1.8.2.3) and are both found in the T. denitrificans genome, separate from the main sox gene cluster, but they are not found in T. thioparus, whereas dissimilatory sulfite reductase (dsr) genes are found in both genomes, as are adenylyl sulfate reductase (aprAB, EC 1.8.99.2) genes, the presence of which has also been confirmed in T. aquaesulis [42].
It was previously noted that 5.9% of the genome (178 genes) of Thermithiobacillus tepidarius DSM 3134 T were potential horizontally transferred genes from Thiobacillus thioparus, Thiobacillus denitrificans and Sulfuricella denitrificans of the Betaproteobacteria [12]; 96 of these genes came from the two Thiobacillus spp. However, very little gene transfer has taken place from members of Acidithiobacillia to T. thioparus DSM 505 T with only 6 genes from Ttb. tepidarius DSM 3134 T , 4 from Acidithiobacillus spp. This is perhaps not surprising due to the thermophilic and acidophilic nature of these three Acidithiobacillia compared to the mesophilic requirements of T. thioparus and the unlikelihood of these species co-inhabiting the same environments. A far larger portion (143 genes; 4.82%) of genes were attributed to transfer from members of Gammaproteobacteria, many of which grow at more neutral pH and mesophilic temperatures. There was no distinct pattern of any particular metabolism pathways or resistances etc. being encoded by these potentially transferred genes.

Conclusions
The genome of Thiobacillus thioparus DSM 505 T gives insights into many aspects of its physiology, biochemistry and evolution. This organism uses the transaldolase variant of the Calvin-Benson-Bassham cycle and produces carboxysomes for carbon dioxide fixation, evident from both the genome and transmission electron microscopy. The expression of both carboxysomes and Ru-BisCO may be regulated by the same divergently transcribed transcriptional regulator. Smith's biosynthetic horseshoe is present in lieu of Krebs' cycle sensu stricto, but this is unusually large as only 2 genes are missing, though T. denitrificans [32] is only missing 1this may in part explain the heterotrophic growth of T. aquaesulis since just one additional gene would convert Smith's horseshoe into a functional version of Krebs' cycle [6]. Many inorganic sulfur-oxidation genes of the sox cluster were found but soxC, soxD, soxE and soxF are absent. The tsdA gene for a thiosulfate dehydrogenase identified in Allochromatium vinosum is absent and confirmation of the presence of a different thiosulfate dehydrogenase enzyme and gene will require further study. The genome sequence will enable evolutionary studies into the nature of Thiobacillus and chemolithoautotrophs in general, in particular reigniting the obligate versus facultative and the autotrophic versus mixotrophic debates that have been largely absent from the literature in recent years, but genome sequences becoming available will now answer many questions proposed over 10 years ago [27,29]. KMG: 1,000 microbial genomes project; NBRC: NITE biological resource center; NCBI: National center for biotechnology information; RuBisCO: Ribulose-1,5-bisphosphate carboxylase/oxygenase