ABSTRACT
We cloned and sequenced a DNA fragment from the thermophilic archaeal strain
Sulfolobus shibatae
B12 that includes the gene
topR
encoding the reverse gyrase. The RNA of the reverse gyrase gene was
characterized indicating that the
topR
gene is fully functional
in vivo
. We showed by primer extension analysis that transcription of
topR
initiates 28 bp downstream from a consensus A-box promoter. In order to understand how this particular type I DNA
topoisomerase introduces positive superturns into the DNA, we compared the
amino acid sequence of reverse gyrase from
S.shibatae
with the two other known reverse gyrases. This comparison indicates a common
organization of these proteins: the carboxy-terminal domain is related to the type I-5
'
topoisomerase family while the amino-terminal domain possesses some motifs of proteins described as RNA or DNA
helicases. By using local alignments, we showed that (i) reverse gyrases
constitute a new and rather homogenous group within the type I-5
'
DNA topoisomerase family; (ii) a careful sequence analysis of the amino-terminal domain allows us to relate the presence of some motifs with an
ATP binding and hydrolysis reaction coupled to a DNA binding and unwinding
activity.
As noted by Watson and Crick a long time ago, the helical structure of DNA
requires enzymes with the ability to eliminate stress in this molecule (
1
). These enzymes are DNA topoisomerases, ubiquitous enzymes that are required for all DNA metabolism processes such as:
replication, transcription, recombination or chromatin assembly (for a review,
see
2
). DNA topoisomerases act by introducing transient single or double strand
breaks in DNA for type I and type II topoisomerases respectively. Whereas the
type II DNA topoisomerases appear evolutionarily related (
3
) in the three domains of living cells (Eucarya, Bacteria and Archaea) (
4
,
5
), type I DNA topoisomerases are more puzzling. From a mechanistic point of view, and irrespective of their evolutionary
history, we can define two families of type I topoisomerases as proposed by
Roca (
2
). The type I-3' family groups topoisomerases that are linked to the DNA by a 3'phosphotyrosyl link (essentially eukaryotic topoisomerase I)
while the type I-5' family is constituted by the topoisomerases that are transiently
linked to the DNA by a 5'phosphotyrosyl link. Two type I-5' topoisomerases are present in
Escherichia coli
: protein [omega], coded by the gene
topA
, relaxes the DNA in the cell (
6
) while topoisomerase III, coded by the gene
topB
, has an unknown function (
7
). This classification of topoisomerases type I-3' or I-5' on a biochemical basis is clearly consistent with
sequence data (
3
).
In order to have a better understanding of the different roles of topoisomerases
in the cell, we have been studying, for several years, topoisomerases issued
from the third domain, the Archaea (previously Archaebacteria) (
5
). In thermophilic Archaea, a new DNA topoisomerase named reverse gyrase was
first described in two different strains of
Sulfolobus
(
8
,
9
). This enzyme catalyzes the formation of positively supercoiled DNA in the
presence of ATP and magnesium. Surprisingly, although it catalyzes a gyration
reaction in an ATP-dependent manner, reverse gyrase is a type I DNA topoisomerase (
9
-
11
). Later on, reverse gyrase activity was detected in all thermophilic Archaea
tested (
12
,
13
) but also in thermophilic Bacteria (
14
,
15
) suggesting that this enzyme is a characteristic of life at high temperature
rather than an archaeal feature. These results raise questions about the
relationships between reverse gyrase and the other topoisomerases and on the
phylogenetic origin of the gene encoding it. Since Jaxel
et al.
(
16
) have shown that reverse gyrase is linked through the 5' end of the DNA, reverse gyrase is a type I-5' topoisomerase. The deduced protein sequence of the gene
coding for
Sulfolobus acidocaldarius
reverse gyrase supports this view (
17
). Moreover, reverse gyrase appears to constitute a chimeric protein with two
domains: the carboxy-terminal part is clearly related to the type I-5' DNA topoisomerase family whereas the amino-terminal part exhibits putative helicase motifs.
In order to have a better knowledge of both the phylogeny and the mechanism of
reverse gyrase, we decided to clone and sequence the reverse gyrase gene from
Sulfolobus shibatae
B12. Indeed, thermophilic Archaea constitute good models for the understanding
of fundamental biological mechanisms. For instance, the transcription machinery of these organisms may be related to the eukaryotic
one (
18
). Within the Archaea domain, the strain
S.shibatae
is particularly interesting since homologs of the eukaryotic transcription
factors, TFIIB and TBP, were described in this strain (
19
,
20
). In addition,
S.shibatae
possesses an inducible virus named SSV1 (
21
) which is able to infect
Sulfolobus
cells (
22
). Finally, SSV1 encodes a site specific recombination system (
23
) and some years ago, we showed that the DNA of SSV1 is positively supercoiled,
suggesting a major role of reverse gyrase
in vivo
(
24
). More recently, we have purified and characterized the reverse gyrase of
S.shibatae
. In particular, we showed that the partial proteolysis of the reverse gyrase
gives rise to a topoisomerase, that has only a ATP-independent relaxation activity similar to that of protein [omega] (
25
). This fact points out a possible new regulatory level in the cell.
In this paper, we describe the cloning and sequencing of a large DNA fragment of
S.shibatae
encompassing the gene encoding reverse gyrase. A messenger RNA from the reverse
gyrase gene is, for the first time, characterized. In particular, from
S.shibatae
cells, we determined the size of this RNA and we localized the transcription
initiation site 28 bp downstream from a consensus promoter sequence. Comparison
of the
S.shibatae
reverse gyrase amino acid sequence with those of reverse gyrases from
S.acidocaldarius
(
17
) and
M.kandleri
(
26
) allows us to define conserved and potentially functional regions of the
protein. All the reverse gyrases possess amino acid sequence motifs
characteristic of type I-5' topoisomerase family but exhibit additional common features. We
therefore propose that reverse gyrases constitute a new type I-5' topoisomerases group named topR beside the previously described
topA and topB groups. Finally, a more detailed comparative analysis on the
amino-terminal part of reverse gyrases suggests that this part possesses a DNA
unwinding activity.
SDS (ultragrade) was from Serva, RnaseA, Sarkosyl and DEPC from Sigma, [[alpha]-
35
S]dATP (1000 Ci/mmol) from ICN, [[alpha]-
32
P]dCTP (3000 Ci/mmol) and [[gamma]-
32
P]ATP (5000 Ci/mmol), HybondN and Hyperfilm were from Amersham. Acrylamide and
bis-acrylamide were purchased from Biorad and phenol was from Appligene.
Agarose (indubiose A 37 NA) was from Industrie Biologique Française. Other chemicals were from Carlo Erba.
Escherichia coli
alkaline phosphatase and T7 sequencing kit were from Pharmacia Biotech. Random priming kit, Dnase I-RNase free and Rnase
inhibitor were from Boehringer (Mannheim, Germany) and proteinase K was from Merck. M-MuLV reverse transcriptase, DNA ligase, restriction enzymes and polynucleotide kinase were purchased from Biolabs.
Sulfolobus shibatae
B12 frozen cells were thawed and resuspended in 50 mM Tris-HCl pH 7.9, 100 mM EDTA, 100 mM NaCl, and Sarkosyl was dissolved by
gentle mixing at room temperature (final concentration 2.5% w/v). After low
speed centrifugation, the supernatant was incubated 3 h at 50oC with proteinase K (0.25 mg/ml). The nucleic acid was extracted with phenol (twice) and chloroform/isoamyl
alcohol (24/1) followed by an ethanol precipitation. The pellet was dissolved with TE (10 mM Tris-HCl pH 8, 1 mM EDTA) and RNAs were removed by RNaseA treatment (0.1 mg/ml, 4
h at 37oC). The RNaseA was then extracted with chloroform/isoamyl alcohol and the
DNA was ethanol-precipitated in the presence of ammonium acetate.
The two degenerated oligonucleotides P1 and P2 used as primers were defined as
in Bouthier de la Tour
et al.
(
27
). Amplification was done with 400 ng of genomic DNA and 500 pmoles of each
primer in a Biomed thermocycler using
Taq
polymerase (Bioprobe). Thirty cycles of amplification were carried out at 72oC.
The DNA fragment corresponding to the PCR product was purified and labelled with [[alpha]-
32
P]dCTP by the random priming method. This product was used in hybridization experiments. Restriction analysis of
S.shibatae
genomic DNA was performed by using the Southern method. We purified DNA of ~5-8 kb obtained after hydrolysis of genomic DNA by
Eco
RI enzyme. These fragments were ligated with pGEM3Zf(+) DNA previously
hydrolyzed by
Eco
RI and dephosphorylated. Transformation of TG1 bacteria cells was achieved by
electroporation and plated on LB, 1.5% agar, 100 [mu]g/ml ampicillin, 0.1 mM IPTG, 2
0
/
0000
X-Gal. The plates were incubated overnight at 37oC. Bacteria were replicated on Hybond N filters and grown for 4 h.
Cells were lysed with NaOH as described by the manufacturer. After
prehybridization (6 h), filters were hybridized with
32
P labelled P 0.85 DNA fragment at 42oC for 16 h in 6* SSC, 5* Denhardts and 50% formamide. Filters were washed at 42oC, three times with 2* SSC and five times with 0.2* SSC, 0.1% SDS. The filters were submitted
to an autoradiography by using Hyperfilm (Amersham).
The positive clones were isolated and a second round of selection was done.
The double strand DNA fragment was sequenced by using the dideoxy method.
Usually, a chase was performed with 200 [mu]M of each dNTP before stopping the reaction (3.5 min at room temperature for
the elongation step, 5 min for the termination reaction at 37oC and 2 min at 37oC for the chase). The sequence was performed either with the
restriction fragments subcloned in pGEM3Zf(+) or directly on the 6046 bp DNA
fragment cloned in pGEM3Zf(+). We used universal or synthetic oligonucleotides as primers. Sequencing reactions were analyzed by polyacrylamide gels (6% acrylamide; 0.2% bis-acrylamide; 7 M urea) in TBE buffer.
Sulfolobus shibatae
B12 cells were grown essentially as described by Zillig
et al.
(
28
). A mid-log culture (50 ml, OD
600nm
= 0.68) was quickly cooled by frozen fresh medium at pH 5.5 and centrifuged at
10 000
g
for 10 min. The cells were resuspended in 50 mM of EDTA and disrupted by
addition of SDS (0.25%). Na-acetate (pH 5.2) was then added to a final concentration of 50 mM; RNAs were extracted twice with 1 vol of phenol saturated with water, followed by an extraction with chloroform/isoamyl alcohol.
The supernatant was neutralized by 60 mM Tris-HCl pH 8, 150 mM NaCl and then ethanol-precipitated. The pellet was dissolved with DEPC-treated water. The purity and concentration of the RNA
preparation were checked spectroscopically.
Primer extension was essentially realized as described by Kingston (
29
). Briefly, 3 and 15 [mu]g of
S.shibatae
RNA (eventually pretreated by RNaseA or RNase-free DNase I) was denatured. A 38mer oligonucleotide described in Figure
1
A (position 874-911) was
32
P-labelled at the 5' terminus. The RNA was then hybridized with it during 20 h at 30oC in 80% formamide, 2.6* SSC. After ethanol precipitation, the dried pellets
were resuspended in reverse transcriptase buffer. The extension reactions were
performed at 37oC for 90 min. The RNA was hydrolyzed by RnaseA treatment and the DNA was
ethanol-precipitated and dried. Sequence controls were realized with the same
oligonucleotide by using the 6046 bp DNA fragment cloned in pGEM3Zf(+) as a
matrix and by the dideoxy method. A part of the reaction products (1/6) were
analyzed by sequencing gel electrophoresis.
Sequence assembly, G+C content, dinucleotide frequency, restriction analyses,
ORF translation, codon usage, amino acid composition and deduced molecular mass
were performed by using DNA Strider (
30
) or LGBC (
31
) softwares.
FASTA (
32
), BLAST (
33
), BLITZ (
34
) and PATTERN (
35
) softwares were used for retrieving sequences in banks. For this, we used
computer facilities of CITI2 (
36
), NCBI or EBI.
SMARTIES softwares package (unpublished, Atelier de BioInformatique, Paris) was used on Macintosh computer for retrieving sequences in
banks by FASTA, for primary multialignments and for consensus determination. Primary multialignments were also realized by
using the VIZZ program (
37
).
The data bank used is essentially Swissprot but also GenBank, EMBL library and
non-redundant database of the NCBI. The matrices used are essentially PAM (50,
250 or 500) and BLOSUM62.
The sequence of the 6046 bp DNA fragment of
S.shibatae
(Ssh) was deposited with EMBL library with annotations. The accession number is
X98420.
The local primary alignments realized by using computer programs were modified
manually.
The Swissprot accession number of the proteins used are: for reverse gyrases:
S.acidocaldarius
(Sac) topR: Q08582,
M.kandleri
(Mka) topRa: U41058*,
M.kandleri
topRb: U41059*; for topAs:
E.coli
(Eco): P06612,
H.influenzae
: (Hin): P43012,
Bacillus subtilis
(Bsu): P39814,
Synecoccochus
sp (Ssp): P34185,
T.maritima
: (Tma): P46799; for topBs:
E.coli
: P14294,
H.influenzae
: P43704,
Saccharomyces cerevisiae
(Sce) topIII: P13099, human (Hum) topIII: U43431*. We also used the poorly
characterized cellular type I-5' topoisomerases from
B.anthracis
: P40114,
M.genitalium
: P47368,
B.firmus
(partial): P34184* and Trae from RP4 plasmid: 437697** only for strictly
conserved regions.
The unwinding proteins used are: human eIF4A: P04765,
S.cerevisiae
DBP1: P24784 and DBP2: P24783,
Vaccinia virus
(Vacc) NPH-II: M35027**, PriA from
E.coli
: P17888 and
H.influenzae
: P44647,
E.coli
LHR: P30015, RecG from
E.coli
: P24230,
S.cerevisiae
SGS1: P35187, BLM from human: U39817*, RecQ from
E.coli
: P15043 and human: P46063.
*indicated a EMBL library and ** a GenBank accession number instead of a
Swissprot accession number.
Comparison of
S.acidocaldarius
reverse gyrase amino acid sequence with bacterial topoisomerase I sequence
allowed us to define conserved amino acids (
17
). On this basis, Bouthier de la Tour
et al.
(
27
) designed a couple of degenerated oligonucleotides (P1, motif 5, and P2, motif
10) (Figs
1
A and
4
A) used for PCR amplification. By using
S.shibatae
DNA, a DNA fragment with a size of ~0.85 kb was amplified (Fig.
1
A, P 0.85). The amino acid sequence deduced from the nucleotide sequence of this
fragment exhibited large homology with
S.acidocaldarius
reverse gyrase. We therefore used this DNA fragment as a probe in order to
screen a sub-bank of
Eco
RI
S.shibatae
genomic fragments, ranging in size from 5 to 8 kb. We cloned a DNA fragment of
6046 bp (for details, see Materials and Methods). By using the P 0.85 cloned
fragment as a probe, restriction analyses show that the cloned DNA fragment and
the genomic DNA have the same restriction maps (Fig.
1
A). The sequencing was performed by using, either the 6046 bp DNA fragment or
Bgl
II restriction fragments cloned in pGEM 3Z(+) vector as matrices. Sequence
analysis of the 6046 bp DNA fragment shows that it contains five ORFs (Fig.
1
B). The ORF3 located on the upper strand corresponds to reverse gyrase as seen
by homology with
S.acidocaldarius
(see below). ORFs 1, 2, 4 and 5 are located on the bottom strand and the
deduced lengths in amino acids are respectively: >102, 97, 89, >453. ORFs 1 and
5 are partial and the ORF 4 overlaps that of reverse gyrase. For these ORFs, we
did not find sequences with any significant similarity in data banks by using
BLAST or FASTA softwares.
In the case of the ORF encoding reverse gyrase, we found two putative ATG
initiation codons separated by three amino acids. Both are included in a
putative ribosome binding site (
38
) as described in Figure
2
. By using the first ATG codon, the length of this ORF is 3498 bp corresponding
to a protein of 1166 amino acids with a molecular mass of 132 kDa. This
predicted molecular mass is slightly higher than that of the purified protein (
25
), as described for other thermophilic proteins. The amino acid composition is
in good agreement with the biochemical data (
25
) and in particular, the cysteine content is very low, as for the other
Sulfolobales proteins. Upstream from the first initiation codon, we found a
putative A-box (Fig.
2
) that is a characteristic of the archaeal promoter (
39
). All these data suggest that this ORF corresponds to the functional gene
encoding reverse gyrase.
In order to map the transcribed DNA region, we prepared total RNA from mid-log culture of
S.shibatae
cells. By using the DNA fragment 3BI as a probe (Fig.
1
A), Northern blot analysis reveals a messenger with a size of ~3700 nt (not shown). This result indicates that this described reverse
gyrase gene is functional
in vivo
. In order to map the transcription initiation site of the reverse gyrase gene,
a 38 base oligonucleotide was used in a primer extension assay with
S.shibatae
RNA. This yielded a 75 nucleotide run-off product (Fig.
2
) mapping to the A (position 798) that is at 28 bp downstream from the 3' terminus of the putative A-box element (Fig.
2
). Finally, the
S.shibatae
reverse gyrase promoter agrees well with the archaeal promoter consensus,
emphasing the importance of a TATA-like A-box in transcriptional initiation as proposed by Reiter
et al.
(
39
). We assume that the first putative initiation codon, located 4 bp downstream
from the 5' end of the messenger, is used in the cell. Indeed, the promoter
sequence, the transcription start, the position of the ATG initiation codon and
the putative ribosome binding site, exhibit the same features as those
described for the transcripts of SSV1, especially for T3 (
40
). As transcription termination sites in
S.shibatae
genes generally map to thymine-rich regions (
41
), we looked for this kind of feature in our sequence. We found that downstream
from the stop codon, the dinucleotide TT frequency is increased and four putative termination sites are found. They are localized between the positions 4310 and 4531, in
agreement with the size of the messenger.
In order to have a better understanding of the reverse gyrase unusual
topoisomerase activity, we compared (Fig.
3
) the amino acid sequence obtained in
S.shibatae
with the sequences previously determined in
S.acidocaldarius
and in
M.kandleri
. The two
Sulfolobale
reverse gyrases are constituted by one polypeptide whereas the
M.kandleri
counterpart is composed of two subunits. Nevertheless, except for an additional
domain in the
M.kandleri
enzyme (
26
), the three reverse gyrases are highly homologous, with short insertions or
deletions. They are composed of a carboxy-terminal domain (amino acids 582-1167 in
S.shibatae
reverse gyrase) related to the type I-5' topoisomerases family (motifs 1-10) and an amino-terminal part (amino acids 1-581) exhibiting some motifs (motifs I, Ia, II,
III, V and VI) of some DNA or RNA helicases (Fig.
4
).
In this manuscript, we report the sequence of a 6046 bp DNA fragment from the
archaeal strain
S.shibatae
B12 containing the gene encoding reverse gyrase.
Analysis of this DNA fragment shows that it exhibits 33.25% of G+C. This very
low G+C content appears as a characteristic of the
Sulfolobus
genus. Interestingly, we found that, in the five coding DNA sequences, the
dinucleotide AA frequency is very high (13.3-16.38%) compared to the TT frequency (4.44-8.92%). The reverse gyrase coding region presents the lowest AA frequency
(13.3%). We searched AA frequency bias for four other genomic coding sequences
reported in
S.shibatae
. We found the same high frequency for AA dinucleotide (11-16.11%) compared with TT (5.88-6.88%). To date, the meaning of this unusual AA content is
unknown. Nevertheless, this may be a new criterion for searching putative
coding regions in thermophilic organisms.
In order to define the genomic organization, we mapped the transcription
initiation site and upstream, we found the consensus sequence of the promoter A-box of
Sulfolobus
. The two putative initiation codons of reverse gyrase are ATG instead of GTG
for the
S.acidocaldarius
counterpart (
17
). The genomic organization is not conserved between
Sulfolobus acidocaldarius
and
shibatae,
since the flanking sequences have no similarity. In
M.kandleri
, the organization is also completely different since reverse gyrase is encoded
by two separate genes (
26
).
Focusing on the coding sequences of the three known reverse gyrases, we observed
that apart from the highly conserved motifs, the proteins are different with
some insertions or deletions. In particular, the two related
Sulfolobus
genes exhibit slight differences. We conclude that the highly conserved motifs
reflect the selective pressure and consequently correspond to the regions
involved in the enzymatic activity of the protein: positive supercoiling of the
DNA in an ATP-dependent process. Finally, the sequence comparison allows us to define
three groups of type I-5' topoisomerases. The first is represented by topA-related enzymes, the second by topB-related proteins and the third by reverse gyrases. In
order to clarify the nomenclature, we propose to name the gene coding for
reverse gyrase,
topR
. In the case of
M.kandleri
,
topRb
is the gene coding for the amino-terminal domain and
topRa
the gene coding for the carboxy-terminal domain [topRa is the protein containing the tyrosine of the
active site as proposed by Krah
et al.
(
26
) and by analogy with the gyrase gene].
The carboxy-terminal part of reverse gyrase corresponds approximatively (see Figs
3
and
4
A) to the truncated 67 kDa of
E.coli
topA used by Lima
et al.
(
50
) for the three dimensional structure determination. We have localized the
strictly conserved amino acids on this three dimensional structure (not shown).
We observed that these amino acids are clustered to a pocket around the
tyrosine involved in the transesterification catalytic reaction (
51
). In terms of enzymatic activity, it is possible to correlate some particular
enzymatic reactions with particular regions of the primary structure of the
proteins. Indeed, since the 67 kDa of topA is only able to cleave single strand
DNA (
52
), we hypothesize that the 67 kDa part of topA is responsible for the
transesterification reaction. The removed carboxy-terminus is presumably responsible for double strand DNA binding (through
zinc fingers anchorage) as suggested by Tse-Dinh (
53
). In the case of topB, the single strand RNA or DNA binding reaction is
performed by the basic amino acids in the carboxy-terminal part (
54
). Recently, Zhang
et al.
(
55
) have shown that the carboxy-terminal domain of TopA and TopB is responsible for the substrate binding
whereas the rest of the protein is responsible for the transesterification
reaction. In the case of reverse gyrase, we showed previously that partial
proteolytic products possess an ATP-independent activity which can only relax negatively supercoiled DNA like
topA does (
25
). We assume that these proteolytic products possess all the motifs
characteristic of the type I-5' topoisomerases (motifs 1-10 in Fig.
3
) and may bind to DNA by the putative zinc finger Zn2.
Comparison of the amino-terminal conserved regions of the three known reverse gyrases with some
DNA or RNA helicases shows that some helicase motifs are present (Fig.
4
B). Motif I and motif III are strictly conserved. Moreover, the sequences of
motifs II and VI seem characteristic of reverse gyrases with DDxD consensus for
motif II and QxxGRxSR for motif VI. Nevertheless, the enzymatic activity of
this putative helicase domain is speculative. Indeed, what is a helicase? The
first definition is a DNA-dependent ATPase that unwinds DNA and moves along it (
56
). From a mechanistic point of view, Lohman (
57
) proposed a `rolling' mechanism for dimeric DNA helicases like Rep (which
possesses a 3'-5' polarity and a low processivity). The enzymatic cycle can
be divided in two parts: the first is the translocation of the unbound monomer
to a double strand DNA, the second step is the unwinding of this double strand
region coupled to ATP hydrolysis. This ATP hydrolysis permits the recycling of
the enzyme. This model may be also proposed for eIF4a (
58
), NPH-II (
59
), PriA (
60
), RecQ (
61
) and RecG (
62
) since their biochemical properties are similar to those of Rep. On the other
hand, most advanced biochemical studies on these proteins succeeded in relating
the different sequence motifs with a precise activity. Thus, the motifs Ia, Ib
and II of eIF4a are responsible for ATP binding and hydrolysis, and the motifs
III and VI connect the ATP hydrolysis with double strand nucleic acid binding
and unwinding (
49
). Since the conserved motifs extend to ~300-400 amino acids in length, it is possible that the additional
translocation activity of helicase is carried out by another part of these
proteins or requires oligomerization. In particular, the helicase activity of
eIF4A is increased by addition of the eIF4F protein (
48
). Moreover, by using vaccinia virus RNA helicase NPH-II, Gross and Shuman (
59
) recently showed that the motif VI is required only for ATP hydrolysis and RNA
unwinding. Consequently, we think that the motifs described as helicase motifs
in fact define an ATP-dependent unwinding activity rather than a helicase activity
per se
. We therefore propose the amino-terminal domain of reverse gyrase as a DNA unwinding domain.
The biochemical studies on reverse gyrase are consistent with this analysis.
Indeed, Shibata
et al
. (
42
) showed that reverse gyrase possesses a DNA-dependent ATPase activity and we have previously shown that the binding of
reverse gyrase induced a DNA unwinding or a left-handed DNA wrapping in an ATP-independent manner (
16
). Finally, DNA cleavage analysis suggests that reverse gyrase does not
translocate along the DNA axis (
16
,
63
and unpublished results of C. J. and M. N.).
Both sequence comparison data and biochemical data are consistent with a very
simple hypothetical model. Starting from a relaxed circular DNA, the DNA helix
is unwound by the amino-terminal part of the reverse gyrase, defining two topological domains in the DNA molecule: the unwound domain and the rest of the
molecule. This unwinding consequently introduces a positive supercoiling in the rest of
the molecule. The topoisomerase part of reverse gyrase would relax the unwound DNA domain. After dissociation
of reverse gyrase, the result would be an increase in the linking number of the
complete DNA molecule and a production of positively supercoiled DNA. In this
model, ATP hydrolysis would occur for recycling the amino-terminal part of reverse gyrase.
Finally, reverse gyrase activity, first discovered in thermophilic Archaea, is
also present in thermophilic Bacteria, indicating that life at high temperature
requires a positive DNA supercoiling activity. It would be interesting to know
the relationships between archaeal and bacterial genes, since the genomic
organization of
M.kandleri
is very different from that of
Sulfolobus
. In addition, Gangloff
et al.
(
64
) point out the possible existence of reverse gyrase in yeast, since they
demonstrated a direct interaction between SGS1 (a protein which possesses
similarities to the reverse gyrase unwinding domain) and TOPIII (a type I-5' topoisomerase) using the double hybrid method. Genetic studies
indicated that
SGS1
and
TOPIII
are involved in the recombination pathway; it is possible that reverse gyrase
activity is specialized in the control of this cellular process in mesophilic
organisms. This idea is supported by the existence in
E.coli
of a similar pair of proteins, RecQ and topB. Finally, the same kind of
interaction may exist in human cells between the
RecQ
or
BLM
gene [involved in Bloom's syndrome (
65
), a repair and recombination desease] and the
TOPIII
gene product recently reported by Hanai
et al.
(
66
). Finally, it is interesting to remember that 10 years ago, a not well
characterized positive supercoiling activity was described in hypermutating
myeloma cell line (
67
). It is possible that positive DNA supercoiling is not limited to thermophilic
organisms but is more crucial for these organisms.
We thank Valérie Borde, Frédérique Gallison, Joël Pothier and Seamus O'Regan for critical reading
of the manuscript, Joël Pothier and Alain Viari (Atelier de BioInformatique, Paris) for the
SMARTIES package softwares and for assistance and instruction in the use of
MOLSCRIPT, René Perard for modification of the shaker for the
Sulfolobus
cultures, Christophe Cullin for his advices about bank screening and Christiane
Portemer for her technical assistance. We acknowledge Christiane Elie for
helpful discussions about ATPase sequences comparison. We thank Wolfram Zillig
and its team for the gift of
Sulfolobus
strain and for their advices about
Sulfolobus
culture. We thank Alfonso Mondragon for providing coordinates of the
E.coli
topoisomerase I crystal structure prior to their submission to the Brookhaven
Protein Data Base. This research was supported by funds from CNRS. M. N. is
supported by Université Versailles Saint-Quentin en Yvelines.
REFERENCES
Return