INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

What are the molecular mechanisms involved in the adaptation of life to elevated temperatures? In terms of DNA dynamics, part of the answer was provided by the discovery in thermophilic organisms of a particular topoisomerase, the reverse gyrase, that modifies the topological state of DNA by introducing positive supercoils in an ATP-dependent process (14). It was suggested that overlinking could compensate for the effect of temperature on DNA structure (16). The enzyme is widely distributed in thermophilic archaea (6, 8). The first reverse gyrase characterized was isolated from the hyperthermophilic archaeum Sulfolobus acidocaldarius (23, 33). Mechanistic studies showed that it is transiently linked to the DNA by a 5' phosphotyrosyl bond (22, 24), classifying it in the type I 5' topoisomerase family as proposed by Roca (38). Sequence analysis further showed that it is a single polypeptide containing putative helicase and topoisomerase domains located in the amino- and carboxy-terminal, respectively, parts of the protein (9). The helicase domain exhibits motifs found in DNA and RNA helicases, and the topoisomerase domain exhibits a significant similarity with the 5' topoisomerase I (protein ) from Escherichia coli. From all of these data, a mechanism of reverse gyration involving the concerted action of two such domains was proposed (9, 14).

To date, four other archaeal reverse gyrase genes have been sequenced: from Sulfolobus shibatae (21), Methanopyrus kandleri (26), Pyrococcus furiosus (3), and Methanococcus jannaschii (7). A comparative analysis of reverse gyrases from two members of the order Sulfolobales (S. acidocaldarius and S. shibatae) and M. kandleri with the other type I topoisomerases of the 5' family (21) showed that the reverse gyrases constitute a new group within this family distinct from the previously described TopA and TopB groups, representing the equivalents of the E. coli topoisomerase I and topoisomerase III, respectively. This group was named TopR. Recent sequence information about P. furiosus and M. jannaschii reverse gyrases supported this classification. Nevertheless, although the reverse gyrases are very similar in sequence, they appear to differ in genetic organization. Whereas the two Sulfolobales and P. furiosus enzymes are single polypeptides of about 130 kDa, the enzyme from M. kandleri is a heterodimer of 138 kDa (RgyB) and 42 kDa (RgyA), with the topoisomerase domain shared between the two subunits (26). Recently, in the course of the systematic sequencing of M. jannaschii genome, the reverse gyrase gene was identified and found to have a deduced sequence of 1,613 amino acids (aa) (7). While it codes for a unique polypeptide, the authors noted the presence of an intein (494 aa) just ahead of the putative active site tyrosine of the topoisomerase domain.

Little information is available about the bacterial reverse gyrase. We have previously discovered the existence of a reverse gyrase in an order of extremely thermophilic bacteria, Thermotogales (5). Since then, another reverse gyrase, isolated from the thermophilic bacterium Calderobacterium hydrogenophilum, has been purified and biochemically characterized as a monomer of about 120 kDa (2). However, no bacterial reverse gyrase gene sequence has so far been obtained. Thus, we decided to identify the reverse gyrase gene from Thermotoga maritima, which is one of the most thermophilic bacteria, with an optimum growth temperature of 80°C. We have already isolated from this species and cloned a topoisomerase I gene whose product is clearly related to the TopA group (4).

With the identification of the T. maritima reverse gyrase gene presented in this report, we have the first bacterial reverse gyrase DNA sequence. Based both on the biochemical characterization of the purified enzyme and DNA sequence analysis, we show here that the bacterial reverse gyrase is very similar to its archaeal counterparts despite the evolutionary distance between the two domains.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Genomic DNA. T. maritima MSB8 (strain DSM 3109) cells were suspended in 100 mM Tris-HCl (pH 7.9)-50 mM EDTA-100 mM NaCl and lysed at room temperature by the addition of 1% Sarkosyl and 1% sodium dodecyl sulfate (SDS). The suspension was then incubated with proteinase K (0.5 mg/ml) for 4 h at 50°C and centrifuged for 10 min at low speed (6,000 × g). The supernatant containing DNA was precipitated with 95% ethanol, and the pellet was dissolved in TE-RNase A buffer (10 mM Tris-HCl [pH 8], 1 mM EDTA [TE], 0.5 mg of RNase A per ml). After incubation for 2 h at 37°C, the DNA was extracted three- fold with phenol and chloroform-isoamyl alcohol (24/1), ethanol precipitated, and dissolved in TE. Purification was achieved by using a cesium chloride gradient.

PCR. PCR was carried out in a total volume of 100 µl which contained 0.5 µg of genomic DNA, all four deoxynucleoside triphosphates (each at 0.2 mM), 10 µl of 10× PCR buffer (Eurogentec), 2.5 U of Taq polymerase (Eurogentec Goldstar), and 100 pmol of each oligonucleotide primer. The forward primer P1 sequence was 5'CGCGGATCCMGNATHGARGAYMGNTGGAT3' (Y = C + T; N = A + C + G + T; M = A + C; H = A + T + C; R = A + G), and the reverse primer P2 sequence was 5'CGGGGTACCTCNGTNCKRTGRTANGTDAT3' (K = G + T; D = G + A + T). The BamHI and KpnI sites were introduced to allow subcloning of the PCR fragment. PCR conditions were 30 cycles of 15 s at 94°C, 30 s at 50°C, and 2 min at 72°C. The 389-bp product was isolated from an agarose gel with Qiaex silica gel particles (Qiagen) according to the manufacturer's protocol. It was then cloned into a pGEM plasmid which had been previously digested by BamHI and KpnI.

Construction and screening of size-selected plasmid libraries. The cloning procedures, including the preparation of DNA libraries, identification and isolation of clones, and Southern blot analysis, were performed by using standard procedures (39). The cloning strategy is summarized in Fig. 3A. The cloned PCR fragment was labeled with [-³²P]dCTP by the random priming method. It was used as a radioactive probe against various endonuclease digestions of T. maritima genomic DNA. A SacI/KpnI fragment of about 2.2 kbp was found to hybridize with the probe. This genomic region was subsequently cloned into plasmid pBluescript SK (+), previously digested by SacI/KpnI, and transformed into E. coli DH5. From about 3,000 colonies screened with the radioactive probe, 20 positive clones were found. Analysis of their restriction maps showed that they were identical. The sequence of the SacI/KpnI 2.2-kbp insert was then determined. As the 5' part of the gene was missing, a SacI/PstI restriction fragment located on the 5' end of the 2.2-kbp fragment was used to probe the restriction pattern of the T. maritima DNA. A ClaI/PstI genomic fragment of about 3.5 kbp was found to hybridize with the SacI/PstI radioactive probe. This region was then cloned into the ClaI/PstI sites of pBluescript SK (+). The size-selected library obtained was screened, and the positive recombinant clones were isolated (6 of 500 colonies screened). After having checked that the DNA inserts presented the same restriction map, we sequenced one of them.

DNA sequencing. The inserts of recombinant plasmids were sequenced by the dideoxy-chain termination method with a T7 sequencing kit (Pharmacia) and universal or synthetic oligonucleotides as primers.

Computer analysis. Protein pairwise sequence similarities were calculated with the Bestfit program with a gap weight parameter of 3.0 and a length weight parameter of 1.0. This program makes an optimal alignment of the best segment of similarity between two sequences by using the algorithm of Smith et al. (41). Multiple alignments of amino acid sequences were obtained by using the Pileup program of the Genetics Computer Group package, version 7.2 (11). The parameters were the following: gap creation penalty, 1; gap extension penalty, 0.1. The phylogenetic tree was constructed by using the DARWIN (data analysis and retrieval with indexed nucleotide) program based on PAM (point accepted mutations) distances (18). For all computer analyses, the intein of the M. jannaschii reverse gyrase was omitted, and the two sequences RgyA and RgyB from M. kandleri were combined.

Enzyme purification. All purification steps were carried out at about 6°C. Frozen cells of T. maritima (15 g) were suspended in 100 ml of buffer A (50 mM Na₂HPO₄ · KH₂PO₄ [pH 7.4], 1 mM dithiothreitol, 1 mM EDTA, 0.4 mM phenylmethylsulfonyl fluoride) containing 1.2 M NH₄Cl, 1 mM EGTA, 1 mM sodium bisulfite, 1 µg of leupeptin per ml, and 1 µg of pepstatin per ml. They were crushed by passage through a French pressure cell (680 bars), and the resulting solution was centrifuged at 17,400 × g for 30 min. Polymyxin P was added to the supernatant (106 ml, 2,027 mg of protein) to a final concentration of 0.36%. After being mixing for 30 min, the solution was centrifuged at 90,000 × g for 1 h. Subsequently, ammonium sulfate (65% saturation) was added to the supernatant and centrifuged at 24,000 × g for 30 min. The pellet was dissolved in 40 ml of buffer A containing 10% ethylene glycol (buffer B) and 0.1 M NaCl and dialyzed overnight against this buffer. The dialysate (43 ml, 522 mg of protein) was clarified by centrifugation at 24,000 × g for 20 min and applied to a phosphocellulose P11 (Whatman) column (18.5 by 3.2 cm) equilibrated with this same buffer. After being washed with the equilibration buffer, the bound proteins were eluted with 1 M NaCl in buffer B. The active fractions (134 ml) were pooled and diluted with buffer B to give a solution whose conductimetry was equal to that of buffer C (buffer B containing 0.5 M NaCl). This solution (192 ml, 66 mg of protein) was loaded onto a single-stranded DNA agarose column (14.5 by 0.9 cm; Bethesda Research Laboratories) equilibrated with buffer C. The bulk of ATP-dependent DNA topoisomerase activity was recovered in the nonretained pool. It was adjusted to final concentrations of 0.8 M ammonium sulfate and 1.0 M NaCl in buffer B (200 ml, 48 mg of protein) and applied to a phenyl-Sepharose column (14 by 1.1 cm; Pharmacia) equilibrated with this buffer. The column was successively washed with 62 ml of equilibration buffer and 200 ml of 0.25 M NaCl in buffer B (buffer D). Then the bound proteins were eluted twice with 30 ml of an ethylene glycol gradient, 10 to 65% in buffer D. The most active fractions eluted around 45% ethylene glycol. They were combined, adjusted to 0.2 M NaCl by dilution with buffer A (3.7 ml, 2.7 mg of protein), and applied to a heparin-Sepharose column (5 by 0.5 cm; IBF) equilibrated with buffer B containing 0.2 M NaCl. The column was washed with this buffer and eluted twice with 3.5 ml of a 0.2 to 1.2 M NaCl linear gradient. Active fractions were pooled, concentrated, and equilibrated with buffer A containing 0.6 M NaCl (buffer E) in an Amicon concentrator (Centricon 30). Finally, the resulting solution (210 µl, 135 µg of protein) was loaded on a 5 to 20% sucrose gradient (11 ml) in buffer E and centrifuged at 175,000 × g for 41 h. Fractions of 600 µl were collected and stored at 20°C until required.

Reverse gyrase activity assay. The reverse gyrase assays were performed as previously described (33). In the course of purification, the ATP-dependent DNA relaxation test was used. A standard reaction mixture (10 µl) contained 50 mM Tris-HCl (pH 8.0), 0.5 mM dithiothreitol, 0.5 mM EDTA, 10 mM MgCl₂, 120 mM NaCl, 30 µg of bovine serum albumin per ml, and 0.25 µg of negatively supercoiled DNA (pTZ18; Pharmacia). Each assay was performed in either the presence or the absence of 1 mM ATP to evaluate the contaminating ATP-independent topoisomerase activity present in the fractions to be tested. After addition of 1 µl of the fraction, the mixture was incubated for 30 min at 75°C (in some cases, 80 to 90°C). Paraffin oil (4 µl) was layered onto the reaction mixture to avoid evaporation. The reaction was stopped by rapid cooling on ice followed by addition of 0.5% SDS, 10 mM EDTA, 5% glycerol, and 0.04% bromophenol blue. The reaction products were analyzed by 1.2% agarose gel electrophoresis for 3 h at 3.6 V/cm in TEP buffer (36 mM Tris, 30 mM NaH₂PO₄, 1 mM EDTA [pH 7.8]). Subsequently, the gel was stained with ethidium bromide (1 µg/ml) and photographed under UV illumination.

Other methods. Protein concentration was determined by the method of Shaffner and Weissmann (40), with bovine serum albumin as a standard. Electrophoresis of proteins was performed under denaturing conditions on an SDS-8% polyacrylamide gel as described by Laemmli (27). Protein bands were visualized by silver staining. Molecular mass markers were myosin (205 kDa), -galactosidase (116 kDa), phosphorylase b (97.4 kDa), bovine serum albumin (66 kDa), ovalbumin (45 kDa), and carbonic anhydrase (29 kDa).

Nucleotide sequence accession number. The sequence of the 5,516-bp DNA fragment of T. maritima has been submitted to GenBank under accession no. AF013268.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

T. maritima reverse gyrase properties. We were first interested in comparing the enzymatic properties of the T. maritima reverse gyrase with those of the other reverse gyrases. To date, one bacterial (2) and four archaeal (3, 25, 32, 33) reverse gyrases have been purified to homogeneity, essentially by using the protocol described for the S. acidocaldarius enzyme (33). Thus, we used this purification scheme for isolating the reverse gyrase from T. maritima cells. This scheme included successive phosphocellulose, phenyl-Sepharose, and heparin-Sepharose chromatographies and a sucrose gradient ultracentrifugation (see Materials and Methods). We also introduced an additional single-stranded DNA agarose chromatography after the phosphocellulose column to remove the important ATP-independent topoisomerase activity present in the extracts (data not shown and reference 5). At 0.5 M NaCl, the ATP-independent topoisomerase activity remained bound to the resin, while the unretained fraction contained the reverse gyrase activity (not shown).

View larger version (79K): [in this window] [in a new window]   FIG. 1.   Analysis of protein fractions following sucrose density gradient fractionation. The fractions were collected from the bottom of the gradient. (A) Activities of the fractions (4 to 18) visualized by one-dimensional agarose gel electrophoresis of DNA. Incubations were performed at 75°C with 1 µl of each fraction, in the presence of 1 mM ATP (see Materials and Methods). T, pTZ18 DNA incubated alone; FIs, negatively supercoiled DNA; FII, nicked circular DNA. (B) SDS-polyacrylamide gel electrophoresis of fractions 7 to 13. Lane M, molecular mass markers; lane P, fraction layered on the sucrose gradient corresponding to the heparin pool. Loadings represent 1 to 2 µg of proteins.

View larger version (58K): [in this window] [in a new window]   FIG. 2.   Enzymatic properties of the T. maritima reverse gyrase. Fraction 10 of the sucrose gradient was used for the assays and the activities were visualized by two-dimensional agarose gel electrophoresis (see Materials and Methods). (A) Requirement for ATP. pTZ18 DNA was incubated without (control DNA) or with enzyme, with 0, 0.1, 0.5, 1, 5, 10, 100, and 1,000 µM ATP. (B) Temperature optimum. The enzyme was incubated with the DNA in a standard mixture at 80, 85, and 90°C. (C) Thermostability. The enzyme was preincubated at 85°C in a standard mixture without DNA for different times (0, 2, 3, and 4 h); then the reaction was initiated by addition of the DNA, and incubation was carried out for 30 min at 85°C. In each panel, the left branch of the arc corresponds to negatively supercoiled topoisomers and the right branch corresponds to positively supercoiled topoisomers.

Isolation of the genomic DNA fragment of T. maritima containing the reverse gyrase gene. To identify the reverse gyrase gene (topR) from T. maritima, we used the PCR-based approach. Two degenerate oligonucleotide primers (P1 and P2) were defined from highly conserved regions of archaeal reverse gyrase sequences (21). P1 corresponds to the amino acid sequence RIEDRWI located within motif 4 described in reference 21, and P2 corresponds to the amino acid sequence ITYHRTD (motif 7b in reference 21). A fragment of 389 bp was amplified from the genomic DNA of T. maritima, consistent with the approximate distance existing between the two corresponding motifs in S. acidocaldarius and S. shibatae topR. The deduced amino acid sequence of the fragment exhibited a large degree of similarity to Sulfolobales reverse gyrases. Therefore, we used this PCR-cloned fragment as a probe to identify the gene encoding T. maritima reverse gyrase. The complete coding sequence and its flanking regions were isolated in two steps (Fig. 3A) (for details, see Materials and Methods). By using the PCR probe, a 2.2-kbp fragment containing the 3' region of the gene was isolated from the size-selected SacI/KpnI library. The 5' region of the gene was identified by screening a size-selected ClaI/PstI library of 3.5 kbp with the 265-bp SacI/PstI restriction fragment located at the 5' end of the 2.2-kbp fragment. In all, a 5,516-bp genomic region was sequenced. A perfect correlation was found between the restriction map of genomic DNA and that of cloned DNA fragments (Fig. 3B).

View larger version (18K): [in this window] [in a new window]   FIG. 3.   Cloning and sequencing of the T. maritima DNA fragment containing the reverse gyrase gene. (A) Strategy of cloning. The 389-bp fragment is the PCR fragment obtained by using primers P1 and P2. The 2.2-kbp fragment was obtained by screening a size-selected SacI/KpnI library with the radioactive 389-bp fragment. The 3.5-kbp fragment was obtained by screening a size-selected ClaI/PstI library with the SacI/PstI restriction product of the 2.2-kbp fragment. Abbreviations for restriction enzymes: S, SacI; Ps, PstI; Pv, PvuI; K, KpnI; C, ClaI; EV, EcoRV. (B) Organization of the ORFs contained in the sequenced fragment. The dashed lines correspond to the unsequenced regions of ORF1 and ORF4. The hatched box represents the coding region of the reverse gyrase gene (T. maritima topR [Tma topR]). <a> represents the sequence covering the 3' end of ORF2 and 5' end of T. maritima topR; <b> represents the sequence covering the 3' end of T. maritima topR and the 5' end of ORF4. (C) Sequences of T. maritima topR overlapping regions, <a> and <b>. The initiation codons for T. maritima topR and ORF4 are underlined. The Shine-Dalgarno (ribosome binding) sequences (RBS) are overlined. The 10 and 35 promoter sequences for ORF4 are represented by asterisks.

Identification of ORFs. Sequence analysis of the 5,516-bp DNA fragment revealed the presence of four open reading frames (ORFs) on the upper strand (Fig. 3B). The amino acid sequences for these ORFs were compared to entries in the protein data banks by using the BLAST software. The 3' end of ORF1 codes for a 57-aa partial sequence too short to be significantly analyzed. ORF2 encodes a 404-aa protein which exhibits high similarities to archaeal adenosylhomocysteine hydrolases: 54% identity with the enzyme from Sulfolobus solfataricus (36) and 57% with that from M. jannaschii (7).

Comparison of the amino acid sequence of T. maritima reverse gyrase with its archaeal homologs. The amino acid sequence of the T. maritima reverse gyrase is shown in Fig. 4. It is the first known sequence of a bacterial reverse gyrase, and we were interested in comparing it to the five corresponding archaeal sequences. The first feature is that the enzyme from T. maritima consists of a single polypeptide as is the case for the archaeal equivalents (except M. kandleri). In contrast to M. jannaschii, the sequence does not contain an intein. Its size (1,104 aa) is comparable to those of reverse gyrases of S. acidocaldarius (1,247 aa), S. shibatae (1,166 aa), and P. furiosus (1,214 aa). This size is about the same for the M. jannaschii enzyme (1,119 aa) when the intein sequence is removed.

View larger version (53K): [in this window] [in a new window]   FIG. 4.   Amino acid sequence of the T. maritima reverse gyrase deduced from the T. maritima topR gene. The vertical dash delimits the helicase and topoisomerase domains. The motifs are numbered according to Jaxel et al. (21). The boxed amino acids correspond to the amino acids found in the reverse gyrase consensus sequence (at least four identical amino acids out of six), obtained from a multiple alignment comprising the six available reverse gyrase sequences from S. acidocaldarius, S. shibatae, P. furiosus, M. jannaschii, M. kandleri, and T. maritima. Among these conserved amino acids, we have shaded those found in DNA or RNA helicases (for the helicase domain) and those found in the type I 5' topoisomerases of the TopA family (for the topoisomerase domain).

View larger version (20K): [in this window] [in a new window]   FIG. 5.   Unrooted phylogenetic tree derived from sequence comparison of type I 5' topoisomerases. The analysis includes the topoisomerase domains of reverse gyrases (TopR with the residues analyzed as indicated), eight TopA sequences, six TopB sequences, the M. jannaschii topoisomerase I (Mja Top1), the Saccharomyces cerevisiae Top3 sequence (Sce Top3), and the human Top3 sequence (Hsa Top3). Abbreviations for the TopR family: Pfu, P. furiosus; Sac, S. acidocaldarius; Tma, T. maritima; Ssh, S. shibatae; Mja, M. jannaschii. Abbreviations for the TopA family: the Tma, T. maritima; Fis, Fervidobacterium islandicum; Bsu, Bacillus subtilis; Scc, Synechococcus sp.; Mtu, Mycobacterium tuberculosis; Eco, E. coli; Hin, Haemophilus influenzae; Mge, Mycoplasma genitalium. Abbreviations for the TopB family: Ban, Bacillus anthracis; Spy, Streptococcus pyrogenes; TraE, RP4 plasmid-encoded topoisomerase; TrsI, Staphylococcus aureus pSK41 plasmid-encoded topoisomerase.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

The knowledge of reverse gyrases isolated from hyperthermophilic archaea is increasing. However, before this work, no primary structure of a reverse gyrase of bacterial origin was known. In this report, we describe the properties and gene structure of the reverse gyrase from the hyperthermophilic bacterium T. maritima.

The gene begins with the unusual initiation codon TTG as in the case of the T. maritima L10 ribosomal protein gene (28) or trp (G-D) genes (42). No promoter consensus characteristic of Thermotoga genes was found upstream of the reverse gyrase coding sequence, probably because the promoter sequence is divergent from the consensus (fewer than four conserved nucleotides). Such a situation was also observed for other Thermotoga genes. Another alternative would be that the reverse gyrase gene is part of a transcriptional unit with the overlapping ORF2. However, this seems unlikely since there is apparently no functional connection between the two gene products.

Analysis of the sequence of the T. maritima reverse gyrase shows that it is similar to the archaeal sequences: it can be classified both in the superfamily of helicases within its N-terminal half and in the type I 5' topoisomerase family within its C-terminal half. Similar to the enzymes from Sulfolobus and Pyrococcus sp., it is made of a single polypeptide and thus appears more canonical than the enzyme from M. kandleri, which differs in its polypeptide organization. Moreover, determination of the sequence of the bacterial reverse gyrase strengthens the previously defined consensus motifs (21), since the sequence exhibits all of the helicase and topoisomerase blocks already identified. For instance, within the helicase-like domain, all reverse gyrases share the sequences AP(T/P)GXGK(T/S) (equivalent of the helicase motif I [AX₄GKT]), DDVD(A/T) (equivalent of the DEAD helicase motif II), SAT (motif III), XRGXDXP (helicase motif V [ARGXD]), and TYXQ(A/G)SGRXSR (helicase motif VI [QXXGRXGR]). In DNA and RNA helicases, all of these motifs are assumed to form the active site: motifs I and II have been shown to be involved in the binding and hydrolysis of ATP, motif III seems responsible for the unwinding activity, and motif VI may interact with DNA or RNA in relation to ATP hydrolysis (34, 35). Other motifs exclusively found in the helicase domain of reverse gyrases are the N-terminal putative zinc finger and several other amino acid blocks distributed along the sequence.

In the topoisomerase domain, nearly all of the conserved motifs are shared by the type I 5' topoisomerases of the TopA family, with the exception of the putative zinc finger found in the N-terminal part. This motif is present in the bacterial reverse gyrase, although its function is still a matter of speculation. In the E. coli topoisomerase I, there is no zinc finger in this position, but three such motifs are clustered in the C-terminal region (44). Since the 67-kDa fragment of TopA lacking these motifs is able to cleave DNA but does not present a catalytic activity of DNA relaxation, it was suggested that the zinc fingers were important to anchor DNA for the strand passage event (29). Similarly, the zinc finger located in the N-terminal part of the topoisomerase domain of reverse gyrases might perform the same function, since there is no zinc finger in its C terminus. However, the M. jannaschii reverse gyrase does not possess a zinc finger at this position. Alternatively, the other zinc finger motif, located in the N-terminal part of the helicase domain, may play the same role.

From the sequence data, as well as from the enzymatic properties, the bacterial T. maritima reverse gyrase is not distinguishable from the archaeal enzymes. A phylogenetic tree constructed by using the topoisomerase domains suggests that reverse gyrases form a very homogeneous group (TopR) including the bacterial enzyme. This finding suggests that reverse gyrase was already present in the last common ancestor of bacteria and archaea (although lateral gene transfer cannot be excluded). One could imagine that a fusion between a helicase module and a topoisomerase gave rise to reverse gyrase before the divergence between bacteria and archaea and that this new gene is conserved in thermophiles of both branches.

The order Thermotogales is one of the deepest branches of bacteria (43), and these organisms have evolved slowly, suggesting that the ancestral reverse gyrase was relatively similar to the enzyme from T. maritima. This would explain the extent of similarity with the archaeal enzymes. Indeed, T. maritima presents a mixture of bacterial and archaeal features: rRNAs and the entire transcription-translation machineries are clearly of a bacterial type (1, 31), while several proteins from T. maritima are closer to their archaeal homologs (ORF2 of this work and reference 10). Moreover, T. maritima differs from other bacteria in its insensitivity to rifampin (an RNA polymerase inhibitor) (20) and to aminoglycosides (translation inhibitors) (30). It also differs in the use of some metabolic pathways. For instance, L-alanine production in T. maritima is similar to that of P. furiosus and Thermococcus profundus (37).

Regarding the topoisomerase content of T. maritima, three different topoisomerases from this species have now been cloned and sequenced: two type I topoisomerases, a reverse gyrase (this work), and an ATP-independent topoisomerase similar to E. coli protein  (4). Recently, a thermophilic type II topoisomerase has been isolated from T. maritima and determined to be a gyrase (19). These results raise the question of the function of these enzymes. It has been found that in mesophilic bacteria, DNA topology is tightly controlled by the antagonist action of gyrase and topoisomerase I (protein ) and results in a negative superhelical density of about 0.05 (12). As a member of the bacterial domain, T. maritima should also contain a gyrase and a topoisomerase I. This is indeed the case, and the topology of plasmid pRQ7 isolated from a Thermotoga sp. shows that the DNA is again negatively supercoiled (19). This addresses the question of the role of a reverse gyrase in thermophilic bacteria. DNA topology may be controlled by gyrase and reverse gyrase antagonist activities. However, this hypothesis is not in good agreement with the fact that topoisomerase I represents the major activity in the T. maritima crude extracts (5). A more likely possibility is that the gyrase and topoisomerase I activities regulate topology, and in this case, reverse gyrase would act on more specific processes. For instance, it has been suggested that it may be involved in genetic stability by removing illegitimate recombination intermediates, mismatched structures, hairpins, or other noncanonical DNA structures formed during replication or transcription (15).