DNA stability at temperatures typical for ... - BioMedSearch

Download PDF
1 downloads 0 Views 1MB Size Report
Comment
Jan 4, 1994 - that a single DNA nick can stop the progression of RNA polymerase in vitro (23). Lindalh recently suggested that there is no need for an ...
.=) 1994 Oxford University Press

Nucleic Acids Research, 1994, Vol. 22, No. 9 1681-1686

DNA stability at temperatures typical for hyperthermophiles Evelyne Marguet and Patrick Forterre* Institut de Genetique et Microbiologie, CNRS URA 1354, Universite Paris-Sud, 91 405, Orsay Cedex, France Received January 4, 1994; Revised and Accepted March 30, 1994

ABSTRACT We have studied the fate of covalently-closed circular DNA in the temperature range from 95 to 1070C. Supercoiled plasmid was not denatured up to the highest temperature tested. However, it was progressively transformed into open DNA by cleavage and then denatured. Thermodegradation was not dependent on the DNA supercoiling density. In particular, DNA made positively supercoiled by an archaeal reverse gyrase was not more resistant to depurination and thermodegradation than negatively supercoiled DNA. Thermodegradation was similar in aerobic or anaerobic conditions but strongly reduced in the presence of physiological concentrations of K+ or Mg2+. These results indicate that the major problem faced by covalently closed DNA in hyperthermophilic conditions is not thermodenaturation, but thermodegradation, and that intracellular salt concentration is important for stability of DNA primary structure. Our data suggest that reverse gyrase is not directly required to protect DNA against thermodegradation or thermodenaturation.

demonstrated. In particular, the greater thermostability of positively supercoiled DNA versus negatively supercoiled DNA has never been experimentally verified. Furthermore, the effect of various histone-like proteins on linear DNA in vitro might not be significant, since linear DNA lacks the topological constraints between the two DNA strands which are typical of DNA in vivo (7). Finally, it is well known that DNA experiences several degradative processes at high temperatures, such as hydrolysis of the phosphodiester bonds, depurination and cytosine deamination (for a recent review see ref. 8). These pathways of thermodegradation should have an important effect on topologically closed DNA stability. These considerations led us to initiate experiments on the effect of very high temperatures on circular DNA duplexes with different levels of superhelicity. Our results indicate that DNA molecules are resistant to denaturation, under physiological conditions, at temperatures of up to 107°C and that positively supercoiled DNA is not significantly more resistant to thermodegradation than negatively supercoiled DNA.

MATERIALS AND METHODS INTRODUCTION Over the last ten years, many procaryotes belonging to the archaeal domain and a few bacteria have been shown to thrive at temperatures above 85°C (1). Some of these hyperthermophiles, living under pressure in a marine environment, can even sustain growth at temperatures of up to 1 100C. How do these unique microorganisms manage to maintain the DNA doublehelical structure at such temperatures? DNA in hyperthermophiles is not stabilized by a high G+C content, since there is no correlation between the G+C content of the archaeal genome and their optimal growth temperature (1). It has been suggested that the DNA in hyperthermophiles is protected against denaturation by specific mechanisms, such as positive supercoiling by reverse gyrase (for reviews see ref. 2, 3) or stabilization by histone-like proteins (4, 5). Indeed, reverse gyrase has been detected in all hyperthermophilic archaea and bacteria tested up to now (6), and the archaeal histone-like proteins HTa from Therrnoplasma acidophilum or HMf from Methanothermus fervidus increase the Tm of linear DNA molecules in vitro (4, 5). However, the role of reverse gyrase or histone-like proteins in DNA stabilization at very high temperature remains to be *To whom correspondence should be addressed

Chemicals and enzymes Ethidium bromide and chloroquine were purchased from Sigma Chemical Co., EcoRI, DNase I, nuclease SI, netropsin and bovine serum albumin from Boehringer (Mannheim, Germany), ethylene glycol and dithiothreitol from Prolabo. All other chemicals were from Merck (Darmstadt, Germany). Calf thymus DNA topoisomerase I and reverse gyrase from Pyroccocus abyssi were gifts from Dr C.Jaxel and Dr F.Charbonnier, respectively. Preparation of DNA substrates The form I (negatively supercoiled) of pTZ18 (pTZ-) was converted to the open circular form II by limited pancreatic DNase I digestion in the presence of ethidium bromide (9) and to the linear form IH by EcoRI digestion. The relaxed form of pTZ18 (pTZ°) was prepared using calf thymus DNA topoisomerase I. Positively supercoiled DNA (pTZ+) was prepared by incubating pTZ18 for 15 min at 94°C with partially purified reverse gyrase fractions from Pyrococcus albyssi (10). Highly negatively supercoiled pTZ18 (pTZ- -) was prepared by relaxing pTZ18 with calf thymus DNA topoisomerase I in the presence of 10 yg/ml ethidium bromide (11). DNA samples

1682 Nucleic Acids Research, 1994, Vol. 22, No. 9 were deproteinized by adjusting to 1 % (w/v) sodium dodecyl sulfate and 1 mM NaCl and extracted with one volume chloroform-isoamyl alcohol (24/1, v/v). DNA was precipitated and dialysed against 25 mM sodium phosphate buffer, pH 7.5. Positively supercoiled DNA was dialysed for two days with several changes of buffer to remove all contaminating ATP which could interfere with the measurement of optical density.

Determination of DNA superhelical densities DNA superhelical densities were estimated by the band counting method in 0.7% two-dimensional gel electrophoreses, with control DNAs of known superhelical density values (10). Hypernegatively supercoiled DNA (pTZ- -) was relaxed in the first dimension by different chloroquine concentrations (15-100 gg/ml) and compared with either pTZ- or with DNA relaxed in the presence of 6 ,ug/ml ethidium bromide. In each cases, twice this chloroquine concentration was used during the run in the second dimension. To determine the superhelical density of positively supercoiled DNA, chloroquine was replaced by netropsin (5 and 10 ,g/ml in the first and the second dimension, respectively) and relaxed pTZ18 was used as reference. The superhelical density of this relaxed DNA was determined by twodimensional agarose gel electrophoresis with no drug in the first dimension and 5 /tg/ml netropsin in the second dimension. DNA temperature treatment DNA thermodegradation was analysed by incubating 1.25 ,ug of DNA in 25 IAI of incubation mixture covered with H20-saturated parafim oil to prevent evaporation. The pH and the temperature of the incubation mixtures were controlled with temperature and pH probes. For experiments under anaerobic conditions, the solutions used to prepare the incubation mixture were reduced by flushing with a mixture of N2 and CO2 and the reactions were carried out inside an anaerobic chamber (La Calhene, Velizy, Villacoublay, France).

Agarose gel analyses For neutral gel electrophoresis, 20 11 DNA samples were run at room temperature in 0.7% agarose gel in TBE buffer for 16 h at 2 V cm-'. For alkaline gel electrophoresis, the gels were prepared in 50 mM NaCl, 4 mM EDTA and soaked for at least 1 h in running buffer containing 30 mM NaOH and 2 mM EDTA. The DNA samples were incubated for 2 h in 0.5 M NaOH (12) and run for 16 h at 1 V cm-' in a 0.7% agarose gel (with recirculation of the buffer). The gels were washed with water and stained with ethidium bromide. Polaroid photographs were taken under UV transillumination at 254 nm.

samples were run in a 0.7% agarose gel with 10 ng/ml ethidium bromide to partly relax supercoiled molecules (FI) [Fig. IA, five left lanes (-)]. The plasmid preparation (time 0) contained a small amount of denatured random coiled molecules (form *) which were not relaxed by ethidium bromide. The data indicates that the negatively supercoiled topoisomers were progressively destroyed upon incubation at high temperature up to 45 min, with the simultaneous appearence of two new bands migrating below the covalently closed form I. The amount of form II slightly increased, whereas the random coiled DNA rapidly disappeared. The two bands appearing at high temperature were shown to correspond to single-stranded plasmid molecules, since they disappeared after nuclease SI treatment, they produced form II upon renaturation and their migration was similar in gels with and without ethidium bromide (data not shown). Denaturation of form II produced two similar bands migrating at the same position, whereas denatured linearized pTZ18 migrated at the position of the lower band, indicating that the upper and lower bands corresponded to circular and linear single-stranded plasmid molecules, respectively. In this experiment with pTZ-, we did not detect the appearence of denatured covalently closed DNA (random coiled) at high temperature. However, residual form I DNA observed

Ai

I

A.

g'.a,

11

v

4

Spectrophotometry Between 20 and 30 ytg/ml DNA were incubated in a final volume of 2 ml in closed and thermostated cuvettes (optical path 1 cm) in a Perkin Elmer X15 spectrophotometer run by an Epson-PCe computer. The temperature was determined by a Digital controller using the program Dos temp.

RESULTS Thermodegradation and denaturation of native pTZ18 We investigated the effect of high temperature upon a native negatively supercoiled DNA, the bacterial plasmid pTZ18 (pTZ-), following incubation for various time periods at 100°C in 25 mM sodium phosphate buffer, pH 7.8. Heated DNA

Figure 1. Kinetics of thermodegradation of DNA with different degrees of superhelicity analysed on neutral and aLkaline agarose gels. Negatively supercoiled pTZ (-), hypernegatively supercoiled pTZ (- -) and positively supercoiled pTZ (+) were incubated for various times at 100°C in 25 mM potassium phosphate buffer, pH 7.5. (A) Neutral agarose gel containing 10 ng/ml ethidium bromide. The band indicated by * corresponds to random coiled DNA. (B) Alkaline agarose gel. RC, random coiled DNA. SSc and SSl, circular and linear singlestranded DNA, respectively.

Nucleic Acids Research, 1994, Vol. 22, No. 9 1683 after incubation for up to 45 min at 100°C [Fig. IA (-)] could have been previously denatured in the test tube at high temperature and subsequently renatured during the electrophoresis at room temperature. To test this hypothesis, we followed the absorbance of pTZ18 forms I and II by spectrophotometry at temperatures increasing from 30 up to 107°C. The DNA was incubated in 25 mM potassium phosphate buffer, pH 7.8, and the temperature was increased at a rate of 2°C per minute. Figure 2 shows that approximately 90% of the DNA in our preparation of negatively supercoiled pTZ18 remained double-stranded up to 107°C, whereas open circular pTZ18 was completely melted with a Tm of 81 + 0.5°C. The slight increase in optical density observed around 80°C and above 105°C with the supercoiled DNA preparation corresponded to the denaturation of contaminating form II and to DNA cleaved at temperatures above 105°C, respectively (see corresponding gels in Fig. 2). This result indicated that covalently closed circular DNA was not denatured at temperatures of up to 107°C, even when negatively supercoiled. However, when incubation was prolonged at this temperature, DNA progressively melted, roughly following first order kinetics (Fig. 2, lower panel) and resulting in the

Time (min) 0 10 20 30451 0 10 20 3045 1 0 10 20 3045

A FlI

Fl

Time 0

I

I

I

10

(min) 0 10 20 30 45 10 10 20 30 45

20 30 45

B

SSc ssI

RC

I

I

I

Figure 2. Variation of optical densities at 258 nm of nicked DNA (upper panel) and negatively supercoiled DNA (lower panel) upon incubation at increasing temperatures from 30 to 1070C, and then at various times at 107°C (see Materials and Methods). The inserts correspond to DNA samples taken before the treatment (left) or after reaching 107°C (right) and then run on agarose gels. DNA recovered at 107°C has been slowly cooled to room temperature to allow renaturation of form

II.

appearence of single-stranded plasmid molecules, as shown previously in Fig. 1. DNA cleavage was probably induced by depurination, since the DNA phosphodiester bond is weakened near apurinic sites (13). To confirm this, we have studied the effect of pH upon thermodegradation. As expected for a process triggered by depurination (14), the rate of thermodegradation increased with decreasing pH (from 8.5 to 5.5) in all buffers tested (sodium phosphate, citrate, borate and Hepes) (data not shown). To identify apurinic sites in plasmids incubated at high temperature, heated DNA samples were treated with alkali, in order to cleave phosphodiester bonds near putative apurinic sites, and analysed on alkaline agarose gels (12). Three bands were observed in non-heated control DNA sample [Fig. iB, five left lanes (-)]. The two upper bands (SSc and SSl) corresponded to circular and linear single-stranded DNA, respectively, and the lower band to random coiled forms (RC). Upon incubation at high temperature, SSc and SSl forms include both molecules previously cleaved and denatured at high temperature and exsupercoiled molecules containing apurinic sites cleaved and denatured in alkaline conditions. In contrast, the residual random coiled forms correspond to supercoiled molecules lacking apurinic sites. Comparison of DNA degradation in neutral and alkaline gels (Fig. IA,B, five left lanes) shows that the extent of heatinduced DNA cleavage was slightly higher in alkaline gels, suggesting that a few uncleaved apurinic sites are present in residual form I incubated at high temperature.

DNA cleavage is independent of DNA superhelicity To test the contribution of DNA superhelicity to thermodegradation, we prepared positively (pTZ+) and hypernegatively supercoiled pTZ18 (pTZ- -), as well as relaxed pTZ18 (pTZ°). pTZ+ was prepared by incubating pTZ18 with reverse gyrase from the hyperthermophilic archaeon Pyrococcus abysii, formerly GE5 (10, 15) whereas pTZ - and pTZ° were prepared by relaxation with calf thymus DNA topoisomerase I with and without ethidium bromide, respectively (see Materials and Methods). The superhelical densities of these DNAs were determined by two-dimensional electrophoresis in agarose gels with either chloroquine or netropsine. Chloroquine relaxes negatively supercoiled DNA by intercalating between the base pairs, whereas netropsin relaxes positively supercoiled DNA by binding into the helix minor groove (16). Table I shows that the superhelical density values ranged from -0.062 to +0.054, when estimated at 107°C, corresponding to a linking number difference of 31 between hypernegatively and positively supercoiled DNA. The extrapolations at 107°C were performed using the equation of Depew and Wang (17), which has been recently validated up to 90°C (10, 18). The absence of detectable denaturation of pTZ18 up to 107°C suggests that this equation might be valid at even higher temperatures. We first compared the rates of DNA thermodegradation of the four DNA samples (pTZ- -, pTZ-, pTZ°, pTZ+) by spectrophotometry. We measured the initial rate of denaturation observed after reaching 107°C (see Fig. 2, lower panel). These values were normalized by the amount of double-stranded DNA present at the beginning of incubation at 107°C, the latter quantity being deduced from the total increase in optical density observed after complete denaturation. Table I shows that all DNA samples were degraded at about the same rate (between 3.3 and 3.6% DNA denatured per min, following cleavage) in spite of their very different superhelical densities. We performed the same

1684 Nucleic Acids Research, 1994, Vol. 22, No. 9 experiment at low salt (0.5 mM potassium phosphate, pH 7.8) to determine if negatively supercoiled DNA was more prone to degradation under conditions destabilizing the double helix. In this case, cleavage and denaturation started around 90°C instead of 1050C, but the initial cleavage rates remained similar for pTZ+ and pTZ - (8.5 and 9.4% DNA denatured per min, respectively). We also compared the thermodegradation rates at 100°C of pTZ- -, pTZ- and pTZ+ by analysing heated DNA samples on agarose gels (see Fig. 1, lanes underlined by -, - -, +). The heated plasmids were run half-sample on neutral gel (panel A) and the other half on aLkaline gel after alkali treatment (panel B). In contrast to pTZ-, pTZ- - was not relaxed by 10 ng/ml ethidium bromide and appeared as a single band migrating at the position of single-stranded circular DNA (Fig. IA, five central lanes). Positively supercoiled DNA was compacted by ethidium bromide and migrated as a single band at the position of singlestranded linear DNA (five right lanes). In alkaline gels, all covalently closed forms (pTZ-, pTZ- - and pTZ+) give the same band of random coiled DNA (RC) after denaturation (panel B). By visual inspection, the three DNAs were degraded at roughly the same rate, either in neutral or in alkaline gels. This was confirmed by densitometric analysis (not shown). The data from neutral gels was in agreement with the results obtained by spectroscopic analysis, i.e. DNA cleavage is independent of superhelicity. Furthermore, the data from the alkaline gel indicated that DNAs with various levels of superhelicity do not exhibit different levels of depurination. Protection of DNA against thermodegradation by high salt concentrations Since positively supercoiled DNA is no more resistant than negatively supercoiled DNA to thermal degradation, we looked for other conditions which could stabilize DNA at high temperature. We observed that resistance to thermodegradation increased at high concentrations of either monovalent or divalent salts. This was in agreement with previous reports that depurination is reduced at high ionic strength (13). Interestingly, meaningful effects were observed at physiological salt concentrations. Figure 3A shows that optimal protection of DNA incubated for 30 min at 950C was obtained at 100 mM KCI (see disappearence of the band SSc, corresponding to circular singlestranded DNA). In the case of MgCl2, we observed optimal protection between 5 and 10 mM (Fig. 3, lower panel). Incubation at high temperature with MgCl2 concentrations above 15 mM induced DNA aggregation, suggesting that high MgCl2

concentrations promoted hybridization of multiple single-stranded DNA molecules. The effect of NaCl and MgSO4 were similar to those of KCl and MgCl2, respectively. We also observed salt protection when the DNA was incubated at high temperature and low pH (Tris-HCl, pH 5.5). The experiments illustrated in Fig. 3 were performed in 25 mM Hepes buffers, pH 7.5, because magnesium produced DNA smears in phosphate buffers. To monitor the effect of salts on depurination, DNA samples were incubated for 30 min at 95°C in the presence of different KCI concentrations followed by treatement with alkali and analysis on alkaline gel (Fig. 3B). Compared to the alkaline gel of Fig. iB, we observed an additional band migrating slightly above the random coiled form. This band probably corresponded to residual forms I, as suggested by the migration in alkaline gel of pTZ18 pre-treated or untreated with alkali (Fig. 3B, lanes b and a, respectively). Addition of KCI had a dramatic effect on the degree of depurination of DNA molecules exposed to heat: such that very little DNA remained visible in lanes corresponding to samples incubated at high temperature without KCI. In contrast, the DNA persisted in the presence of KCI. The degree of protection was already significant at 50-100 mM and increased up to 500 mM. This indicates that protection of DNA by KCI against thermodegradation is primarily due to the inhibition of depurination. We were unable to repeat the same experiment with MgC12 since the DNA precipitated upon alkaline gel electrophoresis in the presence of this salt.

Effect of anaerobiosis on the thermodegradation of DNA It has been previously claimed that the rate of DNA hydrolysis is extremely low in an N2 atmosphere (cited in ref. 19). Since most hyperthermophiles are anaerobes (1), we tested whether the thermodegradation of covalently closed circular DNA was reduced in anaerobiosis. We determined the thermodegradation rate of pTZ- in the temperature range of 95-105°C in an anaerobic chamber with a N2 atmosphere, using reduced solutions. We obtained the same rates of thermodegradation and salt effects as those previously obtained under aerobic conditions (data not shown). It is likely that the low solubility of oxygen in an incubation mixture at very high temperature creates anaerobic-like conditions, even in an oxygen atmosphere.

DISCUSSION We have shown that covalently closed circular DNA is not denatured at temperatures typical for hyperthermophiles, as long as the sugar-phosphate backbone remains intact. This result

Table I. Relative rates of thermodegradation at 107°C of DNAs with different superhelicities Plasmids

Superhelical density

Superhelical density

at 250C

at 107oCa

:

pTZ-pTZpTZ°

pTZ+

0.002

-0.088 -0.052 +0.003 +0.025

A 0.002

-0.062 -0.025 +0.031 +0.054

Initial rates of cleavageb at 107°C 0.6

experiments

3.6 3.7 3.3 3.6

3 2 2 2

No. of

aThe superhelical density was extrapolated at 107°C using the equation of Depew and Wang (17). However, this equation has been validated only up to 95°C (10) and its use does not take into account the effect that the change in superhelicity should induce on the twist. bThe values are the average of two or three experiments (No. of experiments), they are expressed in percentage of total DNA denatured per minute.

e:~ ~s

Nucleic Acids Research, 1994, Vol. 22, No. 9 1685

confirms and extends an earlier report by Vinograd et al. (20). These authors observed that 20% of polyoma virus form I DNA was denatured at 104°C in SSC (0. 15 M NaCl, 0.015 M sodium citrate, pH 7.4) and they extrapolated a Tm of 107°C from experiments performed at 73°C in 7.2 M NaCl04. In our experiments, we could not detect melting at temperatures of up to 107°C in 25 mM phosphate buffer, pH 7.5. The extraordinary resistance of topologically closed DNA to denaturation is due to the very high configurational entropy of the random coiled DNA which is produced by melting form I DNA (20). Almagor

KCI (mM)

A)

o 25 50 100250500 C

F II

sscF F

I

B)

:FF 1

0 25 50 100 250 500

SSS

F'_

j

FI

ab

MgCI2 (mM)

FIl SS _

F

0 1 2 5 10 25 c Figure 3. Effect of KCl and MgCl2 on thermodegradation of supercoiled DNA. pTZ18 plasmid was incubated for 30 min at 95'C in 25 mM Hepes buffer, pH 7.5 in the presence of various KCl (panels A and B) or MgCl2 concentrations (lower panel). After incubation, the DNA samples were run in either neutral (panel A and lower panel) or alkaline (panel B) agarose gels. Lanes a, b and c are nonheated DNA controls. The control a has not been alkali-treated before loading. In this condition, the superhelical DNA is not denatured in the alkaline gel. After alkali treatment, most of the supercoiled DNA has been transformed into the random coiled form (RC), but residual form I can still be detected. SSc and SSl correspond to circular and linear single-stranded pTZ18, respectively.

and Cole reported a Tm of 105°C by scanning calorimetry for chromatin DNA in intact nuclei (21). This putative Tm might be due to the denaturation of the proteins which maintain the linear DNA in eucaryotic nuclei in a topologically closed structure. The complete resistance of covalently closed DNA to denaturation up to at least 107°C indicates that DNA thermodenaturation per se should not be a problem for hyperthermophiles. This may explain why there is no correlation between the G+C content of DNA and the growth temperature of hyperthermophiles, whereas stable RNAs, which are not topologically closed molecules, are more GC-rich in hyperthermophiles than in mesophiles (reviewed in ref. 22). This also indicates that histone-like proteins or reverse gyrase are probably not required to prevent DNA denaturation in hyperthermophiles. The weak point of topologically closed DNA at very high temperatures is the heat-induced hydrolysis of phosphodiester bonds. Potentially, DNA thermodegradation could pose serious problems in hyperthermophiles. For example, it has been shown that a single DNA nick can stop the progression of RNA polymerase in vitro (23). Lindalh recently suggested that there is no need for an exaggerated DNA repair process in hyperthermophiles, since the efficient repair system for correcting chemically-induced depurination in E.coli would be just sufficient to withstand the roughtly 3,000-fold increase in depurination expected when temperature is shifted from 37 to 100°C (8). Nevertheless, a specific problem might occur if depurination is rapidly followed by DNA hydrolysis and denaturation at 100°C, since this would prevent any repair system from operating efficiently. However, we have shown that thermodegradation can be significantly prevented in vitro by high salt concentrations. This effect is especially striking since it becomes meaningful precisely at physiological salt concentrations (50-100 mM KCI, 5-10 mM MgCl2). In particular, the very high intracellular potassium concentrations (from 500 mM to 1 M) observed in some extreme and hyperthermophilic methanogens and Thermococcales (1) may afford even better DNA protection at very high temperatures. We have shown that KCI protects the DNA against thermodegradation by inhibiting depurination. This is probably also the case for MgCl2, since Lindalh and Andersson reported that depurination was reduced by 30% in a buffer with 1 mM MgCl2, but that this salt increases the rate of chain breakage at apurinic sites (13). In our temperature conditions, DNA cleavage occurred very rapidly after depurination, which became the rate limiting step of the process. Lindalh and Andersson suggested that that the inhibition of depurination by magnesium was due to the stabilization of the DNA secondary structure (13). However, we do not detect destabilization of supercoiled DNA at high temperature, in 25 mM phosphate buffer, even in the absence of magnesium (see Fig. 2). The inhibition of DNA cleavage by K+ and Mg2+ is therefore more probably due to a direct effect on depurination. Finally, we have shown that thermodegradation of covalently closed circular DNA is not dependent upon its superhelicity. This was a priori surprising, since negatively supercoiled DNA contains transiently melted regions (24) which could have been more prone to depurination and cleavage than duplex DNA (14, 25). The difference in susceptibility of single-stranded versus double-stranded DNA is probably not sufficient (about four times) and the time during which some regions are transiently melted too short to introduce a difference in sensitivity in our

1686 Nucleic Acids Research, 1994, Vol. 22, No. 9 experiments. Our result contradicts the role usually postulated for reverse gyrase. This prompted us to look for another function for this enzyme. Since reverse gyrase is only detected in hyperthermophilic organisms, we still suspect a role related to temperature. We have recently observed a correlation between the superhelicity of reporter plasmids from Archaea and the optimal growth temperature of the host strain, i.e. the higher the temperature, the lower the negative superhelicity (26). We are currently working on a model in which the increase in the linking number value mediated by reverse gyrase is required to maintain the DNA helical pitch in hyperthermophiles at the same length as in mesophiles.

ACKNOWLEDGEMENTS We thank Christine Jaxel and Franck Charbonnier for the gifts of calf thymus DNA topoisomerase I and Pyrococcus abyssi reverse gyrase, respectively, Agnes Bergerat and Yvan Zivanovic for critical discussions and advice, Patrick Hughes, Michel Duguet and Mark Blight for reading and correction of the manuscript, and Eric Adam who initiated some of the experiments. We are indebted to D.Prieur (CNRS, Roscoff) and W.Guschelbauer (CEA, Saclay) for the usage of the anaerobic chamber and the Perkin-Elmer spectrophotometer, respectively. This work was supported by a grant from the Association de la Recherche sur le Cancer (ARC) and an EC grant from the Generic Project 'Biotechnology of extremophiles', contract BIO-CT93-02734.

REFERENCES 1. Stetter,K.O., Fiala,G., Huber,G., Huber,R. and Segerer,A. (1990) FEMS Microbiol. Rev. 75, 117-124. 2. Kikuchi,A., (1990) In Cozzarelli,N.R. and Wang,J.C. (eds), DNA Topology and its Biological Effects, Cold Spring Harbor Laboratory Press, NY, pp

285-298. 3. Forterre,P. and Elie,C. (1993) In Kates,M., KushnerD. and Matheson,A. (eds), The Biochemistry of Archaea (Archaebacteria)-New Comprehensive Biochemistry, Vol. 26, pp 325-361. 4. Stein,D.B. and Searcy,D.G., (1978) Science 202, 219-221. 5. Sandman,K., Krzycki,J.A., Dobrinski,B., Lurz,R. and Reeve,J.N. (1990) Proc. Natl. Acad. Sci. USA 87, 5788-5791. 6. Bouthier de la Tour,C., Portemer,C., Huber,R., Forterre,P. and Duguet,M.

(1991) J. Bacteriol. 173, 3921-3923.

7. Sinden,R.R. and Petitjohn,D.E., (1981) Proc. Natl. Acad. Sci. USA 78, 224-228. 8. Lindahl,T. (1993) Nature 362, 709-715. 9. Barzilai,R.J. (1973) J. Mol. Biol. 74: 739-742. 10. Charbonnier,F., Erauso,G., Barbeyron,T., Prieur,D. and Forterre,P. (1992) J. Bacteriol. 174, 6103-6108. 11. Keller,W. (1975) Proc. Natl. Acad. Sci. USA 72, 4876-4880. 12. Rosa,S., Fortini,P., Karran,P., Bignami,M. and Dogliotti,E. (1991) Nucl. Acids Res. 19, 5569-5574. 13. Lindahl,T. and Andersson,A. (1972) Biochemistry 11, 3618-3623. 14. Lindahl, T. and Nyberg, B. (1972) Biochemistry, 11, 3610-3618. 15. Erauso,G., Reysenbach,A.-L., Godfroy,A., Meunier,J.-R., Crump,B., Partensky,F., Baross,J.A., Marteinsson,V., Barbier,G., Pace,N.R. and Prieur,D. (1993) Arch. Microbiol. 160, 338-3497. 16. Snounou,G. and Malcom,A.D.B. (1983) J. Mol. Biol. 167, 211-216. 17. Depew,R.E. and Wang,J.C. (1975) Proc. Natl. Acad. Sci. USA 72, 4275-4279. 18. Duguet,M. (1993) Nucl. Acids Res. 21, 463-468. 19. Freifelder,D. and Dewitt,R. (1977) Gene 1, 385-387. 20. Vinograd,J., Lebowitz,J. and Watson,R. (1968) J. Mol. Biol. 33,173 - 197. 21. Almagor,M. and Cole,R.D. (1987) J. Biol. Chem. 262, 15071-15075. 22. Dalgaard,J.Z. and Garrett,R.A. (1993) In Kates,M., KushnerD. and Matheson,A. (eds), The Biochemistry of Archaea (Archaebacteria)-New Comprehensive Biochemistry, Vol. 26, pp 535-573.

23. Zhou,W. and Doetsch,P.W. (1993) Proc. Natl. Acad. Sci. USA 90, 6601 -6605. 24. Wang,J.C. (1974) J. Mol. Biol. 87, 797-816. 25. Eigner,J., Boedtker,H. and Michaels,G. (1961) Biochim. Biophys. Acta 51, 165-168. 26. Charbonnier,F. and Forterre,P. (1994) J. Bacteriol. 176, 1251-1259.