|Home | About | Journals | Submit | Contact Us | Français|
Desulfurococcus kamchatkensis is an anaerobic organotrophic hyperthermophilic crenarchaeon isolated from a terrestrial hot spring. Its genome consists of a single circular chromosome of 1,365,223 bp with no extrachromosomal elements. A total of 1,474 protein-encoding genes were annotated, among which 205 are exclusive for D. kamchatkensis. The search for a replication origin site revealed a single region coinciding with a global extreme of the nucleotide composition disparity curve and containing a set of crenarchaeon-type origin recognition boxes. Unlike in most archaea, two genes encoding homologs of the eukaryotic initiator proteins Orc1 and Cdc6 are located distantly from this site. A number of mobile elements are present in the genome, including seven transposons representing IS607 and IS200/IS605 families and multiple copies of miniature inverted repeat transposable elements. Two large clusters of regularly interspaced repeats are present; none of the spacer sequences matches known archaeal extrachromosomal elements, except one spacer matches the sequence of a resident gene of D. kamchatkensis. Many of the predicted metabolic enzymes are associated with the fermentation of peptides and sugars, including more than 30 peptidases with diverse specificities, a number of polysaccharide degradation enzymes, and many transporters. Consistently, the genome encodes both enzymes of the modified Embden-Meyerhof pathway of glucose oxidation and a set of enzymes needed for gluconeogenesis. The genome structure and content reflect the organism's nutritionally diverse, competitive natural environment, which is periodically invaded by viruses and other mobile elements.
Desulfurococcus kamchatkensis strain 1221nT (also known as DSMZ 18924T or VKM B-2413T) was isolated from a hot spring of Uzon Caldera (Kamchatka Peninsula, Russia). The organism is an obligately anaerobic, heterotrophic archaeon able to grow at temperatures between 65 and 87°C, with the optimum at 85°C. Its cells are nonmotile regular cocci of 0.6 to 1 μm in diameter. D. kamchatkensis utilizes a wide range of proteinaceous substrates, including α-keratin, albumin, and gelatin, as carbon and energy sources. It is also able to ferment some monosugars (hexoses), disaccharides, and oligosaccharides. Elemental sulfur is not essential but stimulates growth. Products of glucose fermentation in the absence of sulfur include CO2, H2, acetate, and a minor amount of propionate. In the presence of elemental sulfur, H2S is formed instead of H2. A detailed phenotypic description of D. kamchatkensis was presented previously (21).
D. kamchatkensis is a member of the archaeal phylum Crenarchaeota. To date, more than 700 bacterial and 50 archaeal genomes have been sequenced completely, and 15 of them represent thermophilic crenarchaea (Fig. (Fig.1).1). In addition, the complete genome sequences of two mesophilic crenarchaea, Nitrosopumilus maritimus and Cenarchaeum symbiosum, have been determined, although the phylogenetic position of this group is under discussion.
Hyperthermophilic archaea of the genus Desulfurococcus are anaerobic organotrophs inhabiting terrestrial hot springs of Iceland, Kamchatka, Kurils, and Yellowstone National Park (56). Members of the genus Desulfurococcus are able to hydrolyze diverse substrates, including proteins and polysaccharides, obtaining energy for growth from fermentation of the monomers formed. Their genomes bear genes encoding diverse thermostable hydrolases that could find application in different fields of industry (9). The ability of these organisms to form molecular hydrogen as the end product of biopolymer degradation could be used in the process of biofuel production. The D. kamchatkensis genome presented here is the first to be reported for the genus Desulfurococcus and constitutes an important step in extending our knowledge of crenarchaeal diversity.
D. kamchatkensis 1221nT was obtained from the culture collection of the Laboratory of Hyperthermophilic Microbial Communities, Winogradsky Institute of Microbiology, Russian Academy of Sciences. Cells were grown in anaerobically prepared medium consisting of mineral background to which peptone (2 g liter−1), yeast extract (0.2 g liter−1), and elemental sulfur (10 g liter−1) were added. The headspace of the bottle was filled with an N2-CO2 (80:20) mixture. The pH of the medium was 6.5, and the incubation temperature was 85°C. Cells were harvested in the early exponential phase by centrifugation at 8,000 rpm for 15 min.
A combined strategy including Sanger sequencing of a clonal shotgun library and 454 pyrosequencing was used. For Sanger sequencing, the shotgun library was constructed in the plasmid vector pUC19 (average insert fragment size, 2 kb), followed by sequencing of 5,000 clones from both ends. The resulting sequences were assembled into about 200 large contigs with the Consed software package (14). The 454 pyrosequencing part of the project was performed on a GS FLX sequencer and resulted in the generation of about 50 Mb of sequences, with an average read length of 220 bp. The GS FLX reads were assembled into 13 large contigs (>2,000 bp) by use of a GS de novo assembler. The 454 contigs were oriented into scaffolds, and the complete genome sequence was generated upon adding the set of Sanger contigs to the assembly. Several sequence ambiguities were resolved by generating and sequencing appropriate PCR fragments. The assembly of the genome at sites with insertion sequence (IS) elements was verified by PCR amplification and sequencing of these regions.
The rRNA genes were identified by a search against the Rfam database (15). tRNA genes were located with tRNAscan-SE (26). Protein coding genes were identified with the GLIMMER gene finder (11). Whole-genome annotation and analysis were performed with the AutoFACT annotation tool (20). Frameshifts or premature stop codons within coding sequences were identified by comparison to other species and confirmed to be “authentic” by either their high-quality sequencing reads or resequencing. Clusters of regularly interspaced repeats (CRISPR) were identified using CRISPR Finder (16); putative transposon-related proteins were found by a search against the IS database (http://www-is.biotoul.fr/is.html). Signal peptides were predicted with SignalP v. 3.0 (http://www.cbs.dtu.dk/services/SignalP/), using the HMM algorithm. Transmembrane protein topology was predicted with TMHMM Server v. 2.0 (http://www.cbs.dtu.dk/services/TMHMM/).
The annotated genome sequence has been deposited in GenBank under accession no. CP001140.
D. kamchatkensis has a single circular chromosome of 1,365,223 bp with no extrachromosomal elements (Fig. (Fig.2).2). There is a single copy of the 16S-23S rRNA operon and a single distantly located 5S rRNA gene. A total of 47 tRNA genes are scattered over the genome. By a combination of coding potential prediction and similarity searches, 1,474 potential protein-encoding genes were identified, with an average length of 818 bp, covering 88.4% of the genome. These values are in good accordance with the general correlation between the microbial genome size and the predicted gene numbers.
Through similarity and domain searches for the predicted protein products, functions of 75% (1,111 genes) of them may be predicted with different degrees of confidence and generalization. The functions of the remaining 363 genes (25%) cannot be predicted from the deduced amino acid sequences; among them, 205 genes are unique to D. kamchatkensis, with no significant similarity to any known sequences.
The numbers of predicted protein genes that are homologous with genes from the other sequenced representatives of the family Desulfurococcaceae, namely, Staphylothermus marinus (NC_009033) and Ignicoccus hospitalis (NC_009776), are illustrated in Fig. Fig.33 (left panel). The data show that D. kamchatkensis and S. marinus share about 55 to 58% of the total gene pool, but only about half of these common genes are also present in the I. hospitalis genome. Consistent with this observation, fewer than 5% of D. kamchatkensis genes have homologs in I. hospitalis but not in the S. marinus genome; the same is true for S. marinus. In this trio, each genome carries a large number of genes exclusive to it, about 40% for D. kamchatkensis and S. marinus and 63% for the genome of I. hospitalis. These results serve to underline the considerable diversity even among members of one family of hyperthermophilic crenarchaea.
The second plot (Fig. (Fig.3,3, right panel) demonstrates the numbers of homologs shared between D. kamchatkensis, S. marinus, and a representative of the evolutionary distant euryarchaeal order Thermococcales, Thermococcus kodakaraensis. Interestingly, T. kodakaraensis carries even more homologs of D. kamchatkensis and S. marinus genes than does I. hospitalis, a member of the same crenarchaeal family, Desulfurococcaceae. This may be attributed partly to the larger number of genes in the T. kodakaraensis genome, but the primary reason is probably the similarity of the main metabolic pathways of D. kamchatkensis and T. kodakaraensis: both organisms are fermenters growing on proteinaceous substrates, while I. hospitalis grows chemolithoautotrophically (37). These results show that genome diversity does not directly correlate with evolutionary distances between organisms, as measured by 16S rRNA sequence divergence, but may be influenced strongly by environmental adaptation requirements.
Genes required for protein export systems are present in D. kamchatkensis, and the SignalP algorithm predicts a total of 97 proteins carrying N-terminal signal sequences. Most of them have been annotated as transporters, exported binding proteins, proteinases, or hypothetical proteins.
Analysis of the repeated sequences and a search against the IS database revealed seven putative transposons belonging to two IS families. We found six almost identical complete copies of an ISC1913-like transposon of the IS607 family, which was named ISDka1. This 1,992-bp transposon contains two genes, orfA and orfB, encoding homologs of resolvase and transposase of ISC1913. In addition, we found eight almost identical copies of 267-bp miniature inverted repeat transposable elements (MITEs) derived from ISDka1. Such MITEs often occur in crenarchaeal genomes where IS elements are present (3). The almost identical sequences of different copies of ISDka1 and MITE-1, derived from it, suggest that this IS element invaded the D. kamchatkensis genome recently on the evolutionary time scale and probably remains functional. Another IS family, IS200/IS605, is represented by a single copy of an ISSis2-like transposon. This IS element, named ISDka2, is 1,890 bp long and also carry two genes encoding homologs of transposase and resolvase. Unlike most IS elements, the transposons present in D. kamchatkensis have no inverted repeats in their terminal regions, as has been reported for the IS200/IS605 family (30).
The third set of IS elements comprise eight copies of another MITE. These repeats (MITE-2) are more divergent in both sequence and length, and some of them contain internal deletions; the lengths of individual repeats are between 225 and 260 bp, and the identities of sequences are in the range of 74 to 98%. The parental IS element cannot be identified in the D. kamchatkensis genome. Probably, these MITEs are remnants of an ancient invasion of an IS sequence whose full copy has already been eliminated, while its MITE derivatives are in different stages of decay.
Searches against the Sulfolobus database (5) yielded no clear evidence for the presence of integrated archaeal viruses and plasmids in the genome. We found, however, two phage-related integrase genes (Dkam_1130 and Dkam_1004) that could represent ancient integration events.
The D. kamchatkensis genome contains 47 tRNA genes, and 16 of them contain introns. No introns are present in the rRNA genes. Until now, only one intron has been found to occur within an archaeal protein-encoding gene (52), but several others have been predicted (4). This intron is located in homologs of the eukaryotic cbf5 gene, encoding RNA pseudouridine synthase. Its excision from mRNAs in Aeropyrum pernix, Sulfolobus solfataricus, and Sulfolobus tokodaii restores the integrity of the coding region of the cbf5 gene (52). The D. kamchatkensis cbf5 gene (Dkam_1147) also contains an internal stop codon, suggesting the presence of a 22-bp intron whose excision would restore the full coding potential of the cbf5 gene (see Fig. S1 in the supplemental material). The search for other “split” genes whose coding potential could be restored by excision of hypothetical introns revealed only either potential points of programmed frameshifts, known to occur in some archaea (8), or genes broken by insertions of IS elements.
Archaeal genomes often contain inteins, in-frame sequences excised by a self-catalytic mechanism after translation. Searches against the intein database (38) gave only one significant hit, a 201-amino-acid intein in a predicted anaerobic ribonucleoside triphosphate reductase (Dkam_1118). This intein contains a self-splicing domain but not a homing endonuclease domain and, consequently, is not able to spread further. Its closest homolog is an intein located within the gene encoding the large subunit of DNA polymerase II of the euryarchaeon Pyrococcus abyssi.
The search for potential restriction enzymes identified an unusual hybrid endonuclease-methyltransferase fusion protein (Dkam_1203) similar to enzymes found in T. kodakaraensis (TK1158 and TK1460). Both adenine- and cytosine-specific DNA methylases are present (encoded by genes Dkam_0485 and Dkam_1265, respectively). Another gene, Dkam_0306, is similar to the adenine-specific DNA methylase gene of the Sulfolobus virus STSV1 and may have appeared in the genome as a result of a horizontal gene transfer event.
D. kamchatkensis contains two CRISPR located close to each other and containing 8 and 85 repeat spacer units. The spacer regions are supposed to be derived from mobile elements, such as viruses (34), and their spacer transcripts may inactivate mobile element propagation by a mechanism somewhat similar to eukaryotic RNA interference (2, 24). We found no matches between the spacer sequences and any known archaeal extrachromosomal genetic elements, but this may merely reflect the unavailability in public databases of sequences of viruses and plasmids that can propagate in D. kamchatkensis. There was also no CRISPR spacer match in the sequences of the IS elements present in the D. kamchatkensis genome.
The only match of spacer sequences of the CRISPR locus is in another region within the D. kamchatkensis genome itself. One 41-bp spacer sequence matches the 3′ region of the Dkam_1260 gene, so the predicted mRNA of the spacer produced by transcription of the CRISPR locus is complementary to the mRNA of Dkam_1260. The function of Dkam_1260 is unknown, although homologous genes are present in the S. marinus (Smar_0909) and Hyperthermus butylicus (Hbut_0467) genomes. Matching of a spacer sequence of the CRISPR locus and a sequence of an endogenous gene opens the question of whether CRISPR activity may be a mechanism for controlling expression of the organism's resident genes, similar to RNA interference in eukaryotes.
A single superoperon containing a set of cas and csa genes adjoins a larger repeat cluster. These proteins are supposed to be involved in the development and activity of the CRISPR locus (19, 24). The gene order is crenarchaeon specific (24), except for the absence of apparent homologs of the cas4 and cas6 genes.
The archaeal chromosomal replication initiation site (oriC) was first identified in P. abyssi within the noncoding region located upstream of a gene encoding a homolog of eukaryotic Orc1/Cdc6 cell division control proteins (36). Some archaea carry multiple cdc6 genes, and for Sulfolobus and Pyrococcus species, multiple chromosomal replication origins were found to be located close to them (27, 33, 40). Z-curve analysis (55) of the D. kamchatkensis genome showed one major peak, at around 1,150 kb, where a nucleotide compositional deviation change occurred (Fig. (Fig.4).4). This peak was found for the Y component of the Z curve (MK disparity), as observed for Sulfolobus acidocaldarius (6) and Pyrobaculum aerophilum (55) chromosomal ori sites. The search for crenarchaeal origin recognition box (ORB) sequences, the binding sites for Orc1/Cdc6 proteins (40, 41), revealed that the noncoding region located between genes Dkam_1234 (putative DNA binding protein Alba2) and Dkam_1235 (unknown function) contains four copies of the ORB motif (Fig. (Fig.4).4). The position of this region coincides with the location of the MK disparity peak (Fig. (Fig.4),4), further supporting the hypothesis that the ori site is located at this point. This site also contains two copies of another ori-specific motif, UCM (41). Another known ori-binding protein is the archaeal WhiP initiator protein, a homolog of the eukaryotic protein Cdt1 (41). The whiP genes are present in genomes of Sulfolobus species, A. pernix, and H. butylicus (41), but we found no apparent homolog of this protein in D. kamchatkensis. Two orc1/cdc6 homologs, encoding replication initiation proteins, are present in the D. kamchatkensis genome (Dkam_1377 and Dkam_1427), but both are located distantly from the potential oriC site, and no ORB-like sequences were found around these genes.
The normal set of genes associated with the crenarchaeal replication apparatus is present in the D. kamchatkensis genome (32), as well as two homologs (Dkam_1233 and Dkam_0826) of reverse gyrase, the hyperthermophile-specific enzyme responsible for introduction of positive supercoils in the DNA molecule (13).
The full set of crenarchaeal RNA polymerase subunits (53) is present, as are archaeal transcription factors, including a single TATA-box-binding protein (Dkam_0275) and two transcription initiation factors IIB (Dkam_0056 and Dkam_0922). A single copy of the 16S/23S rRNA gene operon and a single distantly located 5S rRNA gene are present, as well as the usual set of crenarchaeal ribosomal protein genes (25).
The D. kamchatkensis genome contains 47 tRNA genes carrying 43 different anticodons coding for 19 amino acids. We were not able to identify a gene for tRNATrp(CCT), although the corresponding tryptophanyl-tRNA synthetase gene is present (Dkam_1457). As in most other archaeal genomes, three tRNAMet(CAT) genes are present, and one of them contains an intron. A selenocysteine incorporation system is absent, as in other crenarchaeal genomes. Searches for other untranslated RNAs revealed a gene encoding the RNA component of RNase P.
Aminoacyl tRNA synthetases were identified for all amino acids except glutamine. Two genes encoding subunits of glutamyl-tRNAGln amidotransferase were found; thus, as in many other archaea and gram-positive bacteria (10), glutamine-tRNAs are probably aminoacylated with glutamic acid, followed by transamidation to yield the glutaminyl-tRNA. Two genes were found for methionyl-tRNA synthetase: the first one encodes the full-size enzyme of 499 amino acids (Dkam_0547), while the second gene (Dkam_0847) encodes a shorter, 109-amino-acid protein corresponding to the C-terminal part of the long methionyl-tRNA synthetase gene. Close homologs of the first enzyme are present in the genomes of S. marinus (Smar_1511) and many thermophilic bacteria (Aquifex, Thermus, Petrotoga, Thermoanaerobacter, etc.) but not in other archaea. This gene was probably horizontally transferred from a thermophilic bacterium to a common progenitor of D. kamchatkensis and S. marinus. The short variant of methionyl-tRNA synthetase is of archaeal origin.
The particular property of D. kamchatkensis that prompted us to sequence its genome is the ability of this archaeon to grow on a variety of proteinaceous substrates, including peptone, tryptone, beef extract, albumin, gelatin, and such resistant protein as α-keratin. This suggests that D. kamchatkensis should produce heat-stable extracellular peptidases. A number of peptidases have been isolated from crenarchaeons of the genera Desulfurococcus, Sulfolobus, Staphylothermus, Pyrobaculum, and Aeropyrum (reviewed in reference 51).
The genome of D. kamchatkensis contains more than 30 different endo- and exopeptidases that belong to different families, and genes for peptide transport into the cell. At least five proteases carry putative N-terminal signal peptides, suggesting that they may be active extracellularly: 70-kDa serine protease-like protein Dkam_0359, 49-kDa putative membrane-bound serine protease Dkam_0447, 17-kDa signal peptidase Dkam_1437, and two subtilisin-like serine proteases, Dkam_1274 (141 kDa) and Dkam_1144 (44 kDa). Intracellular proteolysis consists in joint action of several intracellular proteases, including aminopeptidases of different families, Dkam_0225, Dkam_0465, Dkam_1239, and others; metallopeptidases, Dkam_0358, Dkam_1341, and Dkam_1138; trypsin-like serine protease, Dkam_0433; and intracellular proteasome subunits, Dkam_1288 and Dkam_1278.
The comparison of amino acid sequences of putative extracellular proteases from D. kamchatkensis with those available in GenBank showed that their closest homologs are peptidases inferred from genome sequences of Crenarchaeota species. Dkam_0359 belongs to S16-type family of serine proteases ubiquitously present in cren- and euryarchaeal genomes (COG1750). Protease Dkam_0447 belongs to NfeD-like membrane-bound serine proteases; its amino acid sequence was predicted to contain six transmembrane domains. Proteases Dkam_1274 and Dkam_1144 belong to subtilisin-like family of serine proteases (COG1404). The closest homologs of Dkam_1274 are protease Tpen_1714 from T. pendens and a surface layer-associated stable protease from P. aerophilum (30 to 38% identity). Protease Dkam_1144 has more than 50% identity with a previously characterized subtilisin-like serine protease from P. aerophilum, known as aerolysin (50).
Previously (21), using a sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis zymogram method, we identified a ~120-kDa thermostable endopeptidase in the cell envelope fraction of D. kamchatkensis grown on α-keratin. The endopeptidase was active at 85°C and pH 6.6 to 9.0. The only extracellular endopeptidase of this size could be subtilisin-like serine protease Dkam_1274, taking into account that subtilisin-like peptidases are usually synthesized as longer propeptides subsequently self-cleaved to produce shorter mature active protein (46). However, attribution of this endopeptidase to gene Dkam_1274 will require further confirmation since proteins from hyperthermophiles may migrate slowly in SDS polyacrylamide gel electrophoresis due to incomplete unfolding and failure to bind the full amount of SDS.
D. kamchatkensis contains a variety of poly- and oligosaccharide-degrading enzymes, as expected from its growth substrate spectrum. The set of enzymes that act upon starch and related polymers include four different putative α-amylases (Dkam_0349, Dkam_0350, Dkam_0351, Dkam_0353), two pullulanases (Dkam_0273 and Dkam_0979), amylopullulanase (Dkam_0638), arabinogalactan endo-1,4-β-galactosidase (Dkam_0406), and α-glucosidase (Dkam_0582). Four of these proteins (Dkam_0273, Dkam_0979, Dkam_0638 and Dkam_0406) contain putative signal sequences at the N-terminal end, suggesting that they may be secreted and cleave polysaccharides outside the cell.
This agrees with the ability of D. kamchatkensis to grow on the linear α-(1,4)-linked d-glucose polymer dextrin and the α-(1,4), α-(1,6)-linked d-glucose polymer dextran (21). Presumably, the resulting lower molecular weight dextrins can be imported into the cells by the function of the ABC-type maltose and maltodextrin transport system. The imported oligosaccharides are further degraded to monosaccharides (d-glucose) by intracellular sugar-degrading enzymes, including the ones mentioned above. Microbial α-glucosidases are conventionally classified into two groups on the basis of their substrate specificities. Group I enzymes hydrolyze aryl glucosides more rapidly than maltooligosaccharides. Group II enzymes hydrolyze maltooligosaccharides, unlike the group I enzymes (54). Maltose cannot be utilized by D. kamchatkensis (21); consequently, Dkam_0582 most likely represents group I of microbial α-glucosidases.
D. kamchatkensis possesses proteins relevant to cellulose degradation, such as putative endoglucanase (Dkam_0589) and β-glucosidase (Dkam_0253). None of these enzymes is predicted to contain a N-terminal signal sequence, and thus they are not expected to act outside the cell. This agrees with the observation that D. kamchatkensis is unable to grow on microcrystalline cellulose, carboxymethyl cellulose or cellobiose (21). Probably, the above enzymes are involved in intracellular processing of polysaccharides. No genes for xylan or chitin degradation were identified, consistent with the inability of D. kamchatkensis to grow on these substrates (21).
The genome of D. kamchatkensis contains genes encoding enzymes of the modified Embden-Meyerhof pathway of glucose oxidation, which is characteristic of Archaea (45). Not surprisingly, genes found in D. kamchatkensis correspond to the modification variant shown enzymatically for Desulfurococcus amylolyticus (see Table S1 in the supplemental material) (45). Comparisons of the deduced amino acid sequences of the D. kamchatkensis genes with those from A. pernix and P. aerophilum showed a significant degree of correlation to enzymes, many of which were shown to be responsible for particular enzymatic reactions (39).
The oxidized product of glucose fermentation by D. kamchatkensis is acetate (21), which is formed from pyruvate by means of pyruvate:ferredoxin oxidoreductase (Dkam_0581 to Dkam_578 [γ, δ, α, and β]) and acetyl-coenzyme A (acetyl-CoA) synthetases (Dkam_0846 [α], Dkam_0725 [β], and Dkam_0930-Dkam_0929 [α and β]).
In the course of heterotrophic growth on proteinaceous substrates, the organism needs to generate sugars; consistently, the D. kamchatkensis genome encodes a set of enzymes that are needed to reverse the direction of modified Embden-Meyerhof pathway for gluconeogenesis (see Table S1 in the supplemental material).
The amino acid metabolism in D. kamchatkensis proceeds via mechanisms well studied in P. furiosus and assumed to operate also in other thermophilic archaea. Amino acids can be deaminated in a glutamate dehydrogenase coupled manner by aminotransferases (see Table S1 in the supplemental material). The resulting 2-oxoacids are oxidized by ferredoxin-dependent oxidoreductases of the following four types, with different substrate specificities: pyruvate:ferredoxin oxidoreductase, 2-oxoglutarate:ferredoxin oxidoreductase, 2-oxoisovalerate:ferredoxin oxidoreductase, and indolepyruvate:ferredoxin oxidoreductase (see Table S1 in the supplemental material). These oxidoreductases produce reduced ferredoxin and CoA derivatives, which are converted, with concomitant ATP production, to corresponding acids by acetyl-CoA synthetases (see Table S1 in the supplemental material), whose homologs in Pyrococcus furiosus were shown to have broad substrate specificity (31).
Alternatively, 2-oxoacids derived from amino acids may be converted to corresponding aldehydes in ferredoxin-independent reactions catalyzed by the above-listed oxidoreductases, as proposed in (28). The aldehydes may be oxidized by the tungsten-containing aldehyde:ferredoxin oxidoreductase (Dkam_0593).
As shown above, the metabolism of monosaccharides and amino acids in D. kamchatkensis produces much reduced ferredoxin which should be reoxidized. The only reduced fermentation product detected during the growth of D. kamchatkensis on glucose in the absence of sulfur is molecular hydrogen (21). The only hydrogenases found in D. kamchatkensis genome are two six-subunit membrane-bound Ni,Fe-hydrogenases, presumably, energy-converting ones (see Table S1 in the supplemental material). The hydrogenase nature of their catalytic subunits (EchE) is supported by the presence of two CXXC Ni-binding motifs in each of their amino acid sequences, absent in NADH-ubiquinone oxidoreductases, rather closely related to them (49). The energy converting hydrogenases (ECHs) form a subclass within the class of Ni,Fe-hydrogenases (49). ECHs form multisubunit membrane-bound enzyme complexes able to couple proton reduction with low-potential electrons to pumping ions out of the cells or to mediate reverse transfer of electrons at the expense of transmembrane ion gradient. In D. kamchatkensis, ECHs are evidently involved in the direct reaction of hydrogen production from ferredoxin (encoded by Dkam_0514 and Dkam_0628), reduced, e.g., via the operation of glyceraldehyde-3-phosphate ferredoxin oxidoreductase and pyruvate:ferredoxin oxidoreductases during growth on glucose or ferredoxin-dependent oxidoreductases of oxoacids and aldehyde oxidoreductase during growth on peptides. Among Crenarchaeota, the presence of ECH genes has been reported for T. pendens (1). Our search showed that ECH gene clusters are also present in S. marinus but not in A. pernix, I. hospitalis, H. butylicus, C. maquilingensis, Pyrobaculum spp., M. sedula, and Sulfolobus sp. genomes.
Although the presence of ECHs provides for the possibility of oxidative phosphorylation and sufficiently good growth in the absence of electron acceptors, the growth of D. kamchatkensis is inhibited at high hydrogen concentration (21). As many other organotrophic hyperthermophilic archaea, D. kamchatkensis can use elemental sulfur as an electron acceptor, thus avoiding formation of hydrogen. The mechanisms of sulfur reduction by Crenarchaeota are poorly studied, especially at the genomic level, an exception being the autotrophic crenarchaeon Acidianus ambivalens (22). We identified no cytochromes or sets of subunits of membrane-bound sulfur reductases found in some bacteria (17) and A. ambivalens (22). Neither does D. kamchatkensis genome encode close homologs of soluble Ni,Fe-containing hydrogenases of P. furiosus, which were shown to exhibit sulfur reductase activity in vitro and termed sulfhydrogenases but were later demonstrated to be downregulated by S0 (43). The genome of D. kamchatkensis does not encode close homologs of the proposed NAD(P)H:sulfur oxidoreductase from P. furiosus (44), present in many crenarchaeal genomes, or of subunits of the Nso NAD(P)H-oxidizing polysulfide-reducing complex unique to Thermococcus litoralis and T. kodakaraensis (47).
The only D. kamchatkensis genes that we can currently relate to sulfur reduction are Dkam_0441, encoding a homolog of the presumably soluble NADH:polysulfide oxidoreductase from Thermotoga spp. (7) (EMBL/GenBank/DDBJ accession no. AAM76054), and genes of putative ferredoxin:NAD(P) oxidoreductases, encoding homologs of soluble (Dkam_0099 and Dkam_0100) and membrane-bound (Dkam_0956-Dkam_0945) enzymes of P. furiosus (PF1327-PF1328 and PF1441-PF1453) (29, 44). The presumable membrane-bound ferredoxin:NAD(P) oxidoreductase involves subunits homologous to those of NADH:ubiquinone oxidoreductases (PF1441 to PF1447; Dkam_0956 to Dkam_0953 and Dkam_0945) and subunits homologous to those of cation/proton antiporters (PF1448-PF1453; Dkam_0952-Dkam_0947). In P. furiosus, it is strongly upregulated by S0 and is thought to be an energy-converting enzyme complex that supplies NADPH for NAD(P)H:sulfur oxidoreductase and concomitantly generates transmembrane potential (44). The genes of the D. kamchatkensis gene cluster Dkam_0956 to Dkam_0945 all have homologous counterparts in the P. furiosus gene cluster PF1441 to PF1453 and vice versa, with the only exception being Dkam_0946.
The presence of cation/proton antiporters in D. kamchatkensis and clustering of their genes with genes of ECH and membrane-bound ferredoxin:NAD(P) oxidoreductase became easily explainable after our analysis of the gene cluster encoding the ATP synthase of this organism. It turned out that, like in some other hyperthermophilic archaea, the membrane rotor subunit (subunit K) (Dkam_0993) of this V-type ATP synthase contains all Na+-ligands described by Mulkidjanian et al. (35) as necessary to predict Na+ translocation. Thus, assuming that D. kamchatkensis ECHs and membrane-bound ferredoxin:NAD(P) oxidoreductase act as proton pumps, the cation/proton antiporters are needed to transform proton gradient to sodium gradient, utilized for ATP synthesis. The proposed pathways of electron flow and ion translocation in D. kamchatkensis are shown in Fig. Fig.55.
In the genome of D. kamchatkensis, we found the gene encoding ribulose 1,5-bisphosphate carboxylase type III. RuBisCO III is a homolog of RuBisCO I and II, which are the key enzymes of the Calvin-Benson-Bassham cycle of autotrophic CO2 fixation in plants and diverse bacteria. RuBisCO III genes are known to be present in methanogens (12); they are also present in genomes of some other Euryarchaeota (sulfate-reducing Archaeoglobus fulgidus and organotrophic Thermococcales). Among the Crenarchaeota with sequenced genomes, RuBisCO III was identified in the genome of T. pendens (1). We found RuBisCO III genes in S. marinus and H. butylicus, but not in A. pernix, I. hospitalis, C. maquilingensis, Pyrobaculum spp., M. sedula, and Sulfolobus spp. Ribulose bisphosphate carboxylase activity has been found in two Pyrodictium species (18); however, the sequencing of P. abyssii genome is still in progress and no data on the genes encoding RuBisCo in this organism are available.
The role of RuBisCO in archaea is so far unclear, since phosphoribulokinase, the enzyme supplying RuBisCO with its substrate, ribulose 1,5-bisphosphate, has never been found in archaea. We also failed to find this gene in D. kamchatkensis genome. Attempts have been made to design a cycle of CO2 fixation that would involve RuBisCO but would do without phosphoribulokinase (23, 42); however, the cycle still cannot be closed with the enzymes encoded by the genomes involved. The understanding of the role of RuBisCO in archaeal metabolism is a challenging task for further genome research.
We thank Taisia Strakhova for expert technical advice.
This work was supported by the Federal Agency for Science and Innovations of Russia (contract 02.512.11.2080). N.R. was supported by a grant from the National Science Support Foundation. The work of I.V.K., A.V.L., N.A.C. and E.A.B.-O. on the analysis of metabolism of D. kamchatkensis was supported by Program “Molecular and Cellular Biology” of RAS.
Published ahead of print on 29 December 2008.
†Supplemental material for this article may be found at http://jb.asm.org/.