Vol. 171, No. 12JOURNAL OF BACTERIOLOGY, Dec. 1989, p. 6573-65790021-9193/89/126573-07$02.00/0Copyright © 1989, American Society for Microbiology

Gyrase Inhibitors Can Increase gyrA Expression andDNA Supercoiling

ROBERT J. FRANCO't AND KARL DRLICAl.2*

Public Health Research Institute, 455 First Avenue,' and Department of Microbiology, New York UniversitySchool of Medicine,2 New York, New York 10016

Received 25 July 1989/Accepted 19 September 1989

Treatment of bacterial cells with inhibitors of gyrase at high concentration leads to relaxation of DNAsupercoils, presumably through interference with the supercoiling activity of gyrase. Under certain conditions,however, the inhibitors can also increase supercoiling. In the case of coumermycin A1, this increase occurs atlow drug concentrations. Oxolinic acid increases supercoiling in a partially resistant mutant. We found thatincreases in chromosomal DNA supercoiling, which were blocked by treatment with chloramphenicol, wereaccompanied by an increased expression rate of gyrA. This result is consistent with gyrase being responsible forthe increase in supercoiling. In wild-type cells, increases in gyrA expression were transient, suggesting thatwhen supercoiling reaches sufficiently high levels, gyrase expression declines. Oxolinic acid studies carried outwith a AtopA strain showed that drug treatment also increased plasmid supercoiling. The levels of supercoiingand topoisomer heterogeneity were much higher when the plasmid contained one of several promoters fused togalK. Since oxolinic acid causes an increase in gyrA expression, it appears that gyrase levels may be importantin transcription-mediated changes in supercoiling even when topoisomerase I is absent.

It is becoming increasingly evident that certain environ-mental conditions can influence DNA supercoiling (1, 5, 16).Since supercoiling is an important factor in the interaction ofmany proteins with DNA, it may be that the changes insupercoiling are important for the adaptation of bacteria tochanges in growth environment. In Escherichia coli, DNAsupercoiling is controlled largely by two enzymes, DNAgyrase (13) and DNA topoisomerase I (44). Gyrase intro-duces negative supercoils into DNA in vitro, and inhibitionof the enzyme in vivo blocks supercoiling of bacteriophagelambda DNA during superinfection of a lysogen (11, 12) andleads to relaxation of supercoils in the bacterial chromosome(7, 23, 37). Topoisomerase I appears to modulate the effectof gyrase by providing a relaxing activity; mutations in topA,the gene encoding topoisomerase I (38, 41), can lead tohigher than normal levels of supercoiling (4, 34, 35). Super-coiling itself can influence expression of the genes encodinggyrase and topoisomerase I in a way thought to result in thehomeostatic regulation of supercoiling (25, 42, 43). Even thesmall changes in supercoiling expected to arise from changesin temperature or intercalating dyes appear to be correctedby corresponding changes in linking difference (8, 15). Thus,supercoiling levels are controlled; how they are altered byenvironmental changes is not known.The physiological effects of altering DNA supercoiling can

be studied by perturbing the balance of gyrase and topoiso-merase I. Studies of this type have usually involved loweringnegative supercoiling; little is known about effects associ-ated with increasing supercoiling. The ability to increasesupercoiling may be quite important in the regulation of geneexpression, since anaerobiosis and high osmolarity, the twobest-studied environmental factors that alter supercoiling,both raise supercoiling.

Levels of negative supercoiling can be raised above nor-mal in several ways. Mutation of topA is one (4, 34). Under

* Corresponding author.t Present address: Bausch and Lomb Optics Center, Rochester,

NY 14692.

special circumstances, supercoiling can also be increased byinhibitors of gyrase. Oxolinic acid, an inhibitor of the Asubunit of gyrase (11, 40), increases supercoiling in a mutantstrain partially resistant to the drug (23, 33). In wild-typecells, supercoiling can be raised by treatment with lowconcentrations of coumermycin Al (23), an inhibitor of the Bsubunit of gyrase (12). Transcription can also stimulateincreases in negative supercoiling, a phenomenon that todate has been observed only in plasmids (6, 8a, 32, 45).

Inhibitors that shift supercoiling to elevated levels providea way to study the behavior of topoisomerase genes and thecontrol of supercoiling. For example, inhibitors have beenused to show that expression of topA increases when super-coiling becomes high (43). This response by topA wouldcounter perturbations that tend to raise supercoiling. In thisstudy, we report that the rate of expression of gyrA alsoincreases when inhibitors raise supercoiling, consistent withgyrase being responsible for the increase in supercoiling. Butthe increase in gyrA expression is transient in wild-typecells, which suggests that when supercoiling becomes suffi-ciently high, gyrase expression decreases.

MATERIALS AND METHODSBacterial strains, plasmids, and culture conditions. The

bacterial strains used (Table 1) are derivatives of E. coliK-12. Strain GP200 is a Nalr transductant of DM800, aAtopA mutant containing a compensating gyrB mutation.Strain GP200 has the unusual properties of being onlypartially resistant to nalidixic and oxolinic acids and ofincreasing its chromosomal DNA supercoiling when ex-posed to oxolinic acid (23, 33). These properties appear to bedue to the topoisomerase mutations present, since they canbe transferred to other strains by transduction (33). We haveassigned the Nalr allele to gyrA on the basis of cotransduc-tion of this allele and temperature resistance into thegyrA43(Ts) strain KD103 (a spontaneous glpT, phosphomy-cin-resistant [Phosphor] derivative of strain KNK453 [19];phosphomycin resistance is 60% cotransducible with gyrA).Two-thirds of the Nalr temperature-resistant transductants

6573

on Decem

ber 26, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

6574 FRANCO AND DRLICA

TABLE 1. Bacterial strains

Strain Relevant genotype reference

DM4100 cysB242(Am) 38DM800 A(topA-cysB)204 acrA gyrB225 38GP200 DM800 gyrA307 (Nal') 33GP200-1 GP200 pKO100 (bla galK without 24, 27; this

promoter) workGP200-2 GP200 pGAK1 (bla galK with gyrA 27; this work

promoter)GP200-3 GP200 pGB3 (bla galK with gyrB 27; this work

promoter)GP200-4 GP200 pKG1800 (bla galK with 24, 27; this

galK promoter) workKD103 gyrA43(Ts) glpT (Phosphor) This work

from GP200 into KD103 were also Phosphos, the resultexpected if the Nalr allele maps in gyrA (9, 18). Transductionof the Nalr allele from GP200 into a wild-type strain resultsin full resistance to nalidixic acid, indicating that this gyrAallele behaves abnormally when topA is deleted (33). Sincetransduction of topA+ into DM800 raises levels of gyraseexpression (10), it is likely that GP200 has a low level ofgyrase; this may account for lowered resistance to nalidixicacid.

Plasmids that have galactokinase (galK) expression underthe control of either the gyrA (pGAK1), gyrB (pGBK3), orgalK (pKG1800) promoter or no promoter (pKO100) weregifts from Rolf Menzel (see reference 27 for constructions).These plasmids are derivatives of pBR322 in which tet hasbeen replaced by a multicloning site and galK (24) such thattranscription of galK occurs in the same orientation as tet,which in pBR322 is opposite to that of bla.

Liquid cultures were grown at 37°C in glucose M9 minimalsalts medium (28). For measurement of GyrA and GyrBsynthesis rates, the medium was supplemented as describedbelow. For DNA supercoiling assays, cells were grown inglucose M9 salts minimal medium supplemented withCasamino Acids (4.6 mg/ml; Difco Laboratories, Detroit,Mich.), cysteine (56 gig/ml), and thiamine (0.4 ,ug/ml).

Chemicals and reagents. Oxolinic acid was a gift fromWamer-Lambert Research Institute and was stored at aconcentration of 2 mg/ml in 0.1 N NaOH at 4°C. Coumer-mycin A1 was a gift from Bristol Laboratories, Syracuse,N.Y. A fresh stock solution of coumermycin (5 mg/ml indimethyl sulfoxide) was made weekly and stored at 4°C.Chloramphenicol, pyruvate kinase, catalase, and chloro-quine (diphosphate salt) were products of Sigma ChemicalCo., St. Louis, Mo. Chloramphenicol was freshly preparedfor each experiment at a concentration of 100 mg/ml in 95%ethanol. Ampholytes (Bio-Lyte 5-7 and 3-10) and agarose(ultrapure DNA grade) were purchased from Bio-Rad Lab-oratories, Richmond, Calif. [methyl-3H]thymidine (40 to 70Ci/mmol) and D-[1-14C]galactose (60 mCi/mmol) were pur-chased from Dupont, NEN Research Products, Boston,Mass. L-[4,5-3H]leucine (124 Ci/mmol) and L-[35S]methi-onine (1,210 Ci/mmol) were purchased from AmershamCorp., Arlington Heights, Ill. Purified subunits A and B ofDNA gyrase were obtained from M. Gellert, National Insti-tutes of Health, Bethesda, Md.Measurement of synthesis rates of gyrase. Synthesis rates

of the A and B subunits of gyrase were determined bymeasuring the incorporation of radioactive methionine intocellular proteins. Cells were first grown in M9 mediumlacking methionine and leucine but containing the other 18

amino acids, each at a concentration of 40 pRg/ml, and biotinat 1 ,ug/ml. They were then labeled by addition of [3H]leucine(50 giCi/ml) and grown for approximately two generations.Excess unlabeled leucine (300 ,Ig/ml) was added for 5 minbefore the addition of either oxolinic acid or coumermycin.After addition of drug, cells were pulse-labeled by additionof [35S]methionine (20 ,iCi/ml) for 1 min, followed by addi-tion of unlabeled methionine (300 ,ug/ml) for 2 min to allowcompletion of synthesis of labeled proteins. Cells were thenchilled on ice.

Radioactively labeled cells were collected by centrifuga-tion and processed as described by O'Farrell (29), with thefollowing modifications. Cells were suspended in 100 ,ul oflysis buffer (9 M urea, 0.2% Nonidet P-40, 1.6% ampholytepH 5 to 7, 0.4% ampholyte pH 3 to 10, 5% 2-mercaptoeth-anol) and lysed by three freeze-thaw cycles. The first-dimension isoelectric focusing gels were electrophoresed at400 V for 6,000 to 10,000 V-h. The gels were shaken in 5 mlof 10% glycerol-2.3% sodium dodecyl sulfate-5% 2-mercap-toethanol-62.8 mM Tris hydrochloride (pH 6.8) for 1 h andloaded onto sodium dodecyl sulfate-containing polyacryl-amide gels. After electrophoresis, gels were fixed in 50%methanol-12% acetic acid for at least 1 h, and they were thensoaked for several hours in 10% ethanol-5% acetic acid.Gels were then either silver stained (14) or prepared forautoradiography by drying on 3MM paper (Whatman, Inc.,Clifton, N.J.) under vacuum and heat.

Protein spots were excised from dried gels and incubatedin 500 gil of hydrogen peroxide at 70°C for 6 h. Catalase (400,ul of a 250-gIg/ml solution) was added, and incubation wascontinued for 15 min. Radioactivity was then measured afterthe samples were shaken in 19 ml of Liquiscint scintillationfluid for 24 h at room temperature. The rate of synthesis of agiven protein was taken as the 35S/3H ratio in the excisedspot (3H label corrects for differences in protein recoveryfrom gels). Quench curves were determined for each exper-iment to correct for differences in counting efficiency amongsamples. Specific activities from different regions of singlespots were identical, reducing the possibility that neighbor-ing spots influenced the determinations.Measurement of DNA supercoiling. The effects of oxolinic

acid and coumermycin on average chromosomal DNA su-percoiing were assessed by determining the concentrationof ethidium bromide necessary to titrate the negative super-coils present in nucleoids extracted from drug-treated cells.The cell lysis procedure is a modification of that of Stoning-ton and Pettijohn (39); titration of supercoils was measuredby sedimentation of 3H-labeled nucleoids into sucrose den-sity gradients. Both procedures have been described previ-ously (7) but were modified in the following ways: solution Ccontained 0.2% Sarkosyl (CIBA-GEIGY Corp., Summit,N.J.) in addition to the other detergents, and cell lysateswere loaded directly onto the sucrose gradients.

Plasmid supercoiling was compared by gel electrophoresisin agarose containing chloroquine at concentrations indi-cated in the figure legends. Plasmid DNA was isolated by themethod of Holmes and Quigley (17), followed by centrifuga-tion in CsCl.

RESULTS

Correlation of increased chromosomal supercoiling withincreased gyrA expression. In principle, elevated levels ofsupercoiling could arise from an increase in gyrase, from adecrease in topoisomerase I, or both. Increased supercoilinggenerated by oxolinic acid occurs only in strains carrying a

J. BACTERIOL.

on Decem

ber 26, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

INHIBITORS OF GYRASE 6575

IL

S

-5

0-

0

0

Al

FIG. 1. Two-dimensional gel electrophoresis of GyrA and GyrB.Purified GyrA and GyrB were mixed with small amounts of[35S]methionine-labeled E. coli cell lysates and electrophoresed ontwo-dimensional O'Farrell-type gels as described in Materials andMethods. The first dimension was electrophoresed for 6,600 V-h.The gels were then equilibrated for 1 h and loaded onto 7.5% sodiumdodecyl sulfate-polyacrylamide gels, which were electrophoresed at30 mA for 6 h. The gels were fixed and silver stained to locate theGyrA (A) and GyrB (B) spots. Autoradiographs were prepared toidentify the radioactive spots corresponding to GyrA or to GyrB.Electrophoresis was from left to right (acidic to basic) in the firstdimension and from top to bottom (high to low molecular weight) inthe second dimension.

topA mutation (33); therefore, in this case changes in topo-isomerase I levels cannot be involved. To examine thepossibility that gyrase increases supercoiling, we measuredthe rate of synthesis of gyrA after oxolinic acid treatment ofstrain GP200, a AtopA mutant partially resistant to oxolinicacid (see Materials and Methods for strain description). Cellswere pulse-labeled with [35S]methionine, and cell extractswere subjected to two-dimensional gel electrophoresis (29)(Fig. 1). The GyrA protein spot, identified as described in thelegend to Fig. 1, was excised from gels, and the amount of3p incorporated into the protein was determined. Sincegyrase is a stable protein (25), determining the amount of 3tSincorporated into the GyrA protein during a short timeshould reliably estimate the rate of GyrA synthesis. Oxolinicacid increased the rate of synthesis of GyrA by twofold inGP200 (Fig. 2A) under conditions previously shown toincrease average levels of chromosomal DNA supercoilingby about 50% (23, 33). The gyrB gene is mutant in strainGP200, and GyrB does not migrate to the same position asdoes the wild-type protein during electrophoresis; therefore,we could not unambiguously measure the effect of oxolinicacid on GyrB synthesis by using this system.

Since supercoiling can also be increased by treatment ofwild-type cells with low concentrations of coumermycin(23), rates of GyrA synthesis were measured under theseinhibitor conditions. Treatment of strain DM4100 with 1 tLg

time (min)FIG. 2. Effect of inhibitors of gyrase on GyrA and GyrB synthe-

sis. (A) An exponentially growing culture of strain GP200 wastreated with 0 (0) or 5 (0) jig of oxolinic acid per ml for theindicated times and then pulse-labeled with [35S]methionine. Celllysates were electrophoresed in two dimensions, and the amount ofradioactivity in the GyrA protein was determined. Rates of GyrAsynthesis are expressed as a percentage of the control value at timezero. Radioactive labeling and electrophoresis conditions are de-scribed in Materials and Methods. (B) Exponentially growing cul-tures of DM4100 (wild type) were prepared as described for panel A.Rates of protein synthesis for GyrA (0, *) and GyrB (0) areexpressed as a percentage of the untreated control value at timezero. Cultures were treated with coumermycin (1 ,ug/ml; *) oroxolinic acid (2 jig/ml; 0, 0) at time zero.

of coumermycin per ml produced a transient, threefoldstimulation of GyrA production, with maximal stimulationoccurring 20 min after addition of the drug (Fig. 2B). Aftertreatment with coumermycin and separation of proteins byelectrophoresis, the protein spot corresponding to GyrBdisappeared. A more intense protein spot having the samemolecular weight appeared which behaved as though it weremore acidic than GyrB. Although we cannot say unequivo-cally that this new spot is GyrB, it is reasonable to assumethat GyrB production is stimulated by coumermycin, since ithas been shown by other assays that synthesis of bothsubunits of gyrase is stimulated to similar degrees by higherconcentrations of this drug (25).Although high concentrations of the inhibitors also stimu-

late gyrase expression, supercoiling does not increase. Thiswas first seen with coumermycin (25) under conditions thatrelax chromosomal DNA (7). An intermediate effect couldbe seen after oxolinic acid treatment of wild-type cells. Atoxolinic acid concentrations that maximally inhibit DNAsynthesis, little change in supercoiling occurs (23, 36; datanot shown), and the stimulation of gyrA and gyrB expressionis transient (Fig. 2B). Therefore, inhibitors of gyrase in-crease rates of gyrase expression, but the degree and direc-tion of supercoiling change probably depend on whether thedrug concentration is high enough to block the supercoilingactivity of gyrase and on the response of topoisomerase I, ifpresent, to the perturbation.

Blockage by chloramphenicol of increases in supercoilinginduced by gyrase inhibitors. If increased synthesis of gyrase

VOL. 171, 1989

0:

6 01104a

044

A.. I

on Decem

ber 26, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

6576 FRANCO AND DRLICA

x

04-

0

co0-

E._

EEC

c0-0

ac

E.)-a

0

rE

-0--ac

[EBrl (jig/ml)

FIG. 3. Effect of oxolinic acid and chloramphenicol on chromo-somal DNA supercoiling in GP200. Exponentially growing culturesof GP200 labeled with [3H]thymidine (10 ,Ci/ml for one generation)were treated with chloramphenicol (200 ,ug/ml), oxolinic acid (5,ug/ml), or both. Cells were lysed, and nucleoid sedimentation rateswere determined at various concentrations of ethidium bromide. Foreach gradient, the sedimentation rate of the nucleoids was deter-mined relative to that of a '4C-labeled bacteriophage T4B marker(1025 S [3]), using the means of the radioactivity peaks. (A)Untreated (0) and chloramphenicol treated (0); (B) treated withoxolinic acid (0) and chloramphenicol plus oxolinic acid (0). Thedashed line marks the sedimentation minimum of nucleoids fromuntreated wild-type cells.

is responsible for the oxolinic acid-induced increase insupercoiling, then the increase in supercoiling should beblocked by the inhibition of protein synthesis. Chloramphen-icol, at a concentration sufficient to inhibit protein synthesisby more than 95%, blocked the ability of oxolinic acid toincrease supercoiling in strain GP200 (Fig. 3). In thesemeasurements, nucleoids were sedimented into sucrose den-sity gradients containing various concentrations of ethidiumbromide. Under the conditions used, the ethidium bromideconcentration at the sedimentation minimum is proportionalto the superhelix density (2). Since chloramphenicol doesnot affect the ability of oxolinic acid to rapidly inhibit DNAsynthesis in GP200 (data not shown), it is unlikely thatchloramphenicol blocks the increase in supercoiling by pre-venting cellular uptake of oxolinic acid. Chloramphenicolalso blocked the ability of low concentrations of coumermy-cin (1 or 3 ,ug/ml) to increase supercoiling (data not shown).

Time (min)FIG. 4. Kinetics of increase in supercoiling induced by inhibitors

of gyrase. (A) 3H-labeled cultures of strain GP200 were treated withoxolinic acid (5 ,ug/ml) for the indicated times and then rapidlychilled. Cells were lysed, and sedimentation rates of the nucleoidswere determined at various concentrations of ethidium bromide asdescribed in the legend to Fig. 3. The relative superhelical density ofthe DNA, expressed as the ethidium concentration at the sedimen-tation minimum, was determined from plots similar to those shownin Fig. 3. (B) Cultures of strain DM4100 were treated with coumer-mycin (1 pug/ml) for the indicated times, and the relative superhelicaldensity of the nucleoids was determined as described for panel A.

Increase in gyrase expression without detectable DNA re-laxation. Expression of the gyrase genes in vitro is higherwhen DNA is relaxed than when it is supercoiled (25).Therefore, it is possible that the inhibitor treatments de-scribed above cause a decrease in supercoiling that thenstimulates gyrase expression and subsequently increasessupercoiling. We were unable to detect transient relaxationat the whole-chromosome level when strain GP200 wastreated with oxolinic acid (Fig. 4A) or when strain DM4100(wild type) was treated with low concentrations of coumer-mycin (Fig. 4B). In these experiments, drug incubationswere stopped by rapidly chilling the cultures. Since thisprocedure blocks gyrase activity (37), it is unlikely that thetime (10 min) needed to harvest and lyse the cells contrib-uted significantly to incubation times. Rapid (completedwithin a few minutes), slight (less than 10%), or localrelaxation could have escaped detection by the assay used.

Increase in plasmid supercoiling by oxolinic acid treatment.To determine whether the oxolinic acid-induced increase insupercoiling is a general property of bacterial replicons instrain GP200, we also examined supercoiling of plasmids. Tominimize the topological effects that might be generated bytranscription from divergent promoters (20, 32), plasmidpKO100 was examined. In this derivative of pBR322, apromoterless galK gene has replaced tet. Plasmid DNA waselectrophoresed in the presence of chloroquine under con-ditions in which the more negatively supercoiled topoiso-mers migrate more rapidly. Oxolinic acid caused plasmid

J. BACTERIOL.

on Decem

ber 26, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

INHIBITORS OF GYRASE 6577

1 2 3 4 5 6 7 8

FIG. 5. Increased plasmid supercoiling induced by oxolinic acid.Plasmid DNA was isolated from cultures of GP200 transformed withpKO100 (lanes 1 and 2), pGAK1 (lanes 3 and 4), pGBK3 (lanes 5 and6), and pKG1800 (lanes 7 and 8) (31), and electrophoresis wascarried out in 1% agarose at 4°C with continuous recirculation ofbuffer containing 10 ,ug of chloroquine per ml (31). The direction ofmigration is from top to bottom; at this chloroquine concentration,more negatively supercoiled plasmids migrate more rapidly. Plas-mids were isolated from cultures that were treated with oxolinic acid(5 ,ug/ml) for 30 min (lanes 1, 3, 5, and 7) or from untreated cultures(lanes 2, 4, 6, and 8).

DNA supercoiling to increase (Fig. 5, lanes 1 and 2), a resultconsistent with studies on chromosomal DNA (23).

In previous work, it had been discovered that transcrip-tion of the tet gene of pBR322 causes a large increase innegative supercoiling in the AtopA strain DM800 (32). Liuand Wang (20) subsequently proposed that tracking of RNApolymerase along DNA could establish two topologicaldomains. Negative supercoils would be generated behind thetranscription complex, and positive supercoils would begenerated in front of it. The negative and positive supercoilswould be relaxed by topoisomerase I and gyrase, respec-tively. As observed, loss of topoisomerase I allows theaccumulation of negative supercoils (32), and inhibition ofgyrase allows the accumulation of positive supercoils (21,45). An increase in gyrase expression also correlated with atranscription-dependent increase in negative supercoiling:oxolinic acid caused a much greater increase in plasmidsupercoiling in transformants of strain GP200 if the galpromoter had been fused to galK (Fig. 5; compare plasmidspKO100 [lane 1] and pKG1800 [land 7]). In the absence ofthe drug, the two plasmids exhibited little difference (Fig. 5,lanes 2 and 8).The topoisomers of pKG1800 exhibited extensive linking-

number heterogeneity after oxolinic acid treatment of strainGP200-4 (Fig. 6A), as had been seen previously with tran-scription-stimulated increases in supercoiling of pBR322 instrain DM800 (32). In this type of experiment, electrophore-sis was first carried out at a moderately high concentration ofchloroquine to separate the plasmid topoisomers into aseries of bands. The less negatively supercoiled speciesbecame positively supercoiled at this chloroquine concen-tration; the more negatively supercoiled species remainednegatively supercoiled. A second dimension of electropho-resis was then carried out at a higher chloroquine concen-tration to separate species that had remained negativelysupercoiled (left arc) in the first dimension from those thatwere positively supercoiled (right arc). With pKO100 fromstrain GP200-1, extensive heterogeneity was seen only whengels were overloaded (Fig. 6B).

FIG. 6. Two-dimensional gel electrophoresis of plasmids iso-lated from GP200 treated with oxolinic acid. Plasmid DNA wasisolated from cultures of GP200 transformed with pKG1800 (A),pKO100 (B), pGBK3 (C), or pGAK1 (D) and subjected to electro-phoresis in two dimensions as described by Pruss (31). The directionof migration is from top to bottom in the first dimension and from leftto right in the second dimension. The concentrations of chloroquinewere 120 ,ug/ml in the first dimension and 600 ,ug/ml in the seconddimension. The bright smear that appears at the top left of eachtopoisomer distribution is open circular DNA. In each panel, the leftlane is plasmid DNA isolated from untreated control cultures andthe right lane is plasmid DNA isolated from cultures that weretreated with oxolinic acid (5 ,ug/ml) for 30 min.

We also examined plasmid supercoiling in derivatives ofpKO100 in which the gyrA and gyrB promoters were fused tothe coding region of galK. Flanking DNA extended from-194 to +264 for gyrA and from -110 to +69 for gyrB (+1is the start of transcription) (27). Oxolinic acid increased thenegative supercoiling and the topoisomer heterogeneity ofthese plasmids in a manner similar to that observed when thegalK promoter was present (Fig. 5 and 6). Therefore, thegyrase promoters were sufficiently active for transcription-dependent increases in supercoiling to occur in strainsGP200-2 and GP200-3. Using the method of McKenney et al.(24), we then measured galactokinase production to assessthe effect of oxolinic acid and the very large increase innegative supercoiling on the activity of the gyrase promot-ers. The level of galactokinase detected in cells containingplasmids with the gyrA, gyrB, or galK promoter fused togalK (strain GP200-2, GP200-3, or GP200-4, respectively)was about 27 times that of the control (strain GP200-1containing pKO100), and the addition of oxolinic acid re-duced it by only 10 to 25% for each promoter (data notshown). Therefore, the promoters were relatively insensitiveto the very high levels of negative supercoiling in theseplasmids.

DISCUSSION

Inhibitors of gyrase generally relax bacterial DNA (7, 23);however, in special circumstances they can also increase

VOL. 171, 1989

on Decem

ber 26, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

6578 FRANCO AND DRLICA

supercoiling in chromosomes (23, 33; Fig. 3 and 4) and inplasmids (Fig. 5 and 6). Increased gyrA expression (Fig. 2)accompanies inhibitor-related increases in supercoiling, sug-gesting that increased gyrase expression accounts for theincrease in supercoiling. As expected, chloramphenicol, aninhibitor of protein synthesis, blocks the increase in super-coiling caused by the two inhibitors of gyrase (Fig. 3 anddata not shown).How gyrA expression is increased by inhibitors is not well

understood. In vitro transcription from gyrA and gyrB ismuch higher from relaxed templates (25), with the gyrA andgyrB promoters being the supercoiling-sensitive regions ofthe genes (26). However, both oxolinic acid and coumermy-cin can also increase gyrA expression under conditions inwhich supercoiling increases (Fig. 2 and 4). One testable ideais that the SOS response is somehow involved, since bothdrugs induce that response. It could also be argued that atransient relaxation occurred that was too rapid or too small(less than 10%) for us to detect. That view would beconsistent with the observations obtained with extensiveDNA relaxation (25-27). Also consistent is the possibilitythat a local relaxation occurs around the gyrA gene, one thatwould have escaped detection by our whole-chromosomemeasurements. Such a phenomenon, if peculiar to the chro-mosome, might help explain why oxolinic acid increasesgyrA expression from the chromosome in GP200 (Fig. 2A)but decreases expression slightly when the promoter is fusedto galK in plasmid pGAK1 (data not shown). Indeed, thegyrase genes might be good regions of the chromosome toexplore for local changes in supercoiling.The transient nature of the increase in gyrase expression

(Fig. 2B) fits with the hypothesis that supercoiling is homeo-statically regulated. According to that model (25), decreasedsupercoiling (7) causes expression of gyrase to increase (25)and expression of topoisomerase I to decrease (42). Thiswould tend to bring supercoiling back to normal levels,which would in turn reduce gyrase expression and increasetopoisomerase I expression. Raising supercoiling above nor-mal levels should have the opposite effect. Topoisomerase Ipromoters, cloned onto a plasmid, do increase their activityin strain GP200 after oxolinic acid treatment (43), and inwild-type cells the rate of gyrA expression declines after theinitial increase generated by coumermycin (Fig. 2B). Atransient increase in expression is also seen after oxolinicacid treatment of wild-type cells (Fig. 2B), but so far we havenot observed changes in average chromosomal supercoilingunder these conditions. Transitory or local changes in super-coiling could have escaped detection. When both gyrasegenes are mutant and topA is deleted (strain GP200), anaspect of control appears to be lost: during oxolinic acidtreatment of strain GP200, the rate of gyrA expression failedto drop once chromosomal supercoiling became elevated(Fig. 2A).Plasmid supercoiling increases much more dramatically

after oxolinic acid treatment of a partially resistant mutant(GP200) if the galK gene in the plasmid is fused to apromoter (Fig. 5 and 6). Promoter-dependent increases innegative supercoiling had been reported previously for thetet gene in pBR322 in strain DM800 (32), from which GP200was derived. In the case of DM800, rifampin treatmenteliminates the high level of supercoiling (6), and it is likelythat with both DM800 and GP200 the increase in supercoilingis stimulated by transcription. Our experiments with strainGP200 extend the previous studies in two ways. First,transcription-stimulated increases in supercoiling in topAmutants can occur when galK is substituted for tet (during

the course of this work, a similar observation was made byWu et al. [45]; for discussion of differences between tet andgal, see reference 22). This finding supports the idea thattranscription-mediated changes in supercoiling are commonto many genes. Second, the absence of topoisomerase I isnot a sufficient condition for transcription to increase nega-tive supercoiling: strain GP200 contains a deletion of topA,but plasmids do not exhibit the high degree of supercoilingwithout oxolinic acid treatment (Fig. 5 and 6). In strainDM800, oxolinic acid treatment was not required to observethe very high level of supercoiling (31, 32). We speculate thatin strain GP200 the gyrA307 allele lowers gyrase activity anddiminishes the effect of transcription on supercoiling inplasmids (Fig. 5). Oxolinic acid stimulates production ofgyrA and presumably gyrase (Fig. 2); then very high levels ofsupercoiling with considerable topoisomer heterogeneity areobserved in appropriate plasmids (Fig. 5 and 6). Thus, itappears that the imbalance between gyrase and topoisomer-ase I must exceed a threshold value before the high levels ofsupercoiling are observed. Since transcription itself seems tobe able to create topoisomerase imbalances in wild-typebacteria (8a) and in yeast cells (30), it is likely that transcrip-tion-mediated alteration of DNA supercoiling is a generalphenomenon. Factors important in transcription-mediatedsupercoiling, other than topoisomerase imbalance, includethe orientation of transcription units (20, 45), the level oftranscription (8a), and anchorage of transcription units (22).We now need to determine whether transcription is animportant factor in determining levels of chromosomal DNAsupercoiling.

ACKNOWLEDGMENTS

We thank Martin Gellert for providing the A and B subunits ofgyrase, Rolf Menzel for providing plasmids, and James Wang forcommunicating results prior to publication. We also thank thefollowing for critical comments on the manuscript: Richard Burger,Marila Gennaro, John Kornblum, Rolf Menzel, Ellen Murphy, andJames Wang.

This work was supported by Public Health Service grants fromthe National Institutes of Health, grant NP565 from the AmericanCancer Society, and grant PMB 8718115 from the National ScienceFoundation.

LITERATURE CITED1. Balke, V., and J. Grafla. 1987. Changes in the linking number of

supercoiled DNA accompany growth transitions in Escherichiacoli. J. Bacteriol. 169:4499-4506.

2. Bauer, W., and J. Vinograd. 1968. The interaction of closedcircular DNA with intercalative dyes. J. Mol. Biol. 33:141-176.

3. Cummings, D. 1964. Sedimentation and biological properties ofT-phages of Escherichia coli. Virology 23:408-418.

4. DiNardo, S., K. A. Voelkel, R. Sternglanz, A. E. Reynolds, andA. Wright. 1982. Escherichia coli DNA topoisomerase I mu-tants have compensatory mutations in DNA gyrase genes. Cell31:43-51.

5. Dorman, C., G. Barr, N. Bhriain, and C. Higgins. 1988. DNAsupercoiling and the anaerobic growth phase regulation of tonBgene expression. J. Bacteriol. 170:2816-2826.

6. Drlica, K., R. Franco, and T. Steck. 1988. Rifampicin and rpoBmutations can alter DNA supercoiling in Escherichia coli. J.Bacteriol. 170:4983-4985.

7. Drlica, K., and M. Snyder. 1978. Superhelical Escherichia coliDNA: relaxation by coumermycin. J. Mol. Biol. 120:145-154.

8. Esposito, F., and R. Sinden. 1987. Supercoiling in prokaryoticand eukaryotic DNA: changes in responses to topologicalperturbation of plasmids in E. coli and SV40 in vitro, in nuclei,and in CV-1 cells. Nucleic Acids Res. 15:5105-5123.

8a.Figueroa, N., and L. Bossi. 1988. Transcription induces gyrationofthe DNA template in Escherichia coli. Proc. Natl. Acad. Sci.

J. BACTERIOL.

on Decem

ber 26, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

INHIBITORS OF GYRASE 6579

USA 85:9416-9420.9. Fuchs, J., and H. Karistrom. 1976. Mapping of nrdA and nrdB in

Escherichia coli K-12. J. Bacteriol. 128:810-814.10. Geliert, M., R. Menzel, K. Mizuuchi, M. H. O'Dea, and D.

Friedman. 1983. Regulation of DNA supercoiling in E. coli.Cold Spring Harbor Symp. Quant. Biol. 47:763-767.

11. Gellert, M., K. Mizuuchi, M. H. O'Dea, T. Itoh, and J. Tomi-zawa. 1977. Nalidixic acid resistance: a second genetic charac-ter involved in DNA gyrase activity. Proc. Natl. Acad. Sci.USA 74:4772-4776.

12. Gellert, M., M. H. O'Dea, T. Itoh, and J. Tomizawa. 1976.Novobiocin and coumermycin inhibit DNA supercoiling cata-lyzed by DNA gyrase. Proc. Natl. Acad. Sci. USA 73:4474-4478.

13. Gellert, M., M. H. O'Dea, K. Mizuuchi, and H. Nash. 1976.DNA gyrase: an enzyme that introduces superhelical turns intoDNA. Proc. Natl. Acad. Sci. USA 73:3872-3876.

14. Goldman, D., S. Sedman, M. Ebert, and H. Lyon. 1980. Ultra-sensitive stain for protein in polyacrylamide gels shows spinalregional variation in cerebral fluid. Science 211:1437-1438.

15. Goldstein, E., and K. Drlica. 1984. Control of DNA supercoil-ing: temperature shifts change DNA linking numbers. Proc.Natl. Acad. Sci. USA 81:4046-4050.

16. Higgins, C. F., C. J. Dorman, D. A. Stirling, L. Waddell, I. R.Booth, G. May, and E. Bremer. 1988. A physiological role forDNA supercoiling in the osmotic regulation of gene expressionin S. typhimurium and E. coli. Cell 52:569-584.

17. Holmes, D., and M. Quigley. 1981. A rapid boiling method forthe preparation of bacterial plasmids. Anal. Biochem. 114:193-197.

18. Kistler, W., and E. Lin. 1971. Anaerobic L-a-glycerophosphatedehydrogenase of Escherichia coli: its genetic locus and itsphysiological role. J. Bacteriol. 108:1224-1234.

19. Kreuzer, K. N., and N. R. Cozzarelli. 1979. Escherichia colimutants thermosensitive for deoxyribonucleic acid gyrase sub-unit A: effects on deoxyribonucleic acid replication, transcrip-tion, and bacteriophage growth. J. Bacteriol. 140:424-435.

20. Liu, L., and J. Wang. 1987. Supercoiling of the DNA templateduring transcription. Proc. Natl. Acad. Sci. USA 84:7024-7027.

21. Lockshon, D., and D. R. Morris. 1983. Positively supercoiledplasmid DNA is produced by treatment of Escherichia coli withDNA gyrase inhibitors. Nucleic Acids Res. 11:2999-3016.

22. Lodge, J., T. Kazic, and D. Berg. 1989. Formation of supercoil-ing domains in plasmid pBR322. J. Bacteriol. 171:2181-2187.

23. Manes, S. H., G. J. Pruss, and K. Drlica. 1983. Inhibition ofRNA synthesis by oxolinic acid is unrelated to average DNAsupercoiling. J. Bacteriol. 155:420-423.

24. McKenney, K., H. Shimatake, D. Court, U. Schmeissner, C.Brady, and M. Rosenberg. 1981. A system to study promoterand terminator signals recognized by Escherichia coli RNApolymerase, p. 383-415. In J. C. T. Papas (ed.), Gene amplifi-cation and analysis, vol. 2. Structural analysis of nucleic acids.Elsevier Science Publishing, Inc., New York.

25. Menzel, R., and M. Gellert. 1983. Regulation of the genes for E.coli DNA gyrase: homeostatic control of DNA supercoiling.Cell 34:105-113.

26. Menzel, R., and M. Geliert. 1987. Modulation of transcription byDNA supercoiling: a deletion analysis of the Escherichia coligyrA and gyrB promoters. Proc. Natl. Acad. Sci. USA 84:4185-4189.

27. Menzel, R., and M. Geliert. 1987. Fusions of the Escherichiacoli gyrA and gyrB control regions to the galactokinase gene areinducible by coumermycin treatment. J. Bacteriol. 169:1272-1278.

28. Miller, J. 1972. Experiments in molecular genetics. Cold SpringHarbor Laboratory, Cold Spring Harbor, N.Y.

29. O'Farrell, P. 1975. High resolution two-dimensional electropho-resis of proteins. J. Biol. Chem. 250:4007-4021.

30. Osborne, B. I., and L. Guarente. 1988. Transcription by RNApolymerase II induces changes of DNA topology in yeast.Genes Dev. 2:766-772.

31. Pruss, G. J. 1985. DNA topoisomerase I mutants: Increasedheterogeneity in linking number and other replicon-dependentchanges in DNA supercoiling. J. Mol. Biol. 185:51-63.

32. Pruss, G., and K. Drlica. 1986. Topoisomerase I mutants: thegene on pBR322 that encodes resistance to tetracycline affectsplasmid DNA supercoiling. Proc. Natl. Acad. Sci. USA 83:8952-8956.

33. Pruss, G., R. Franco, S. Chevalier, S. Manes, and K. Drlica.1986. Escherichia coli topoisomerase I mutants: effects ofinhibitors of DNA gyrase. J. Bacteriol. 168:276-282.

34. Pruss, G. J., S. H. Manes, and K. Drlica. 1982. Escherichia coliDNA topoisomerase I mutants: increased supercoiling is cor-rected by mutations near gyrase genes. Cell 31:35-42.

35. Richardson, S. M. H., C. F. Higgins, and D. M. J. Lilley. 1984.The genetic control of DNA supercoiling in Salmonella typhi-murium. EMBO J. 3:1745-1752.

36. Snyder, M., and K. Drlica. 1979. DNA gyrase on the bacterialchromosome: DNA cleavage induced by oxolinic acid. J. Mol.Biol. 131:287-302.

37. Steck, T. R., G. J. Pruss, S. H. Manes, L. Burg, and K. Drlica.1984. DNA supercoiling in gyrase mutants. J. Bacteriol. 158:397-403.

38. Sternglanz, R., S. DiNardo, K. A. Voelkel, Y. Nishimura, Y.Hirota, A. K. Becherer, L. Zumstein, and J. C. Wang. 1981.Mutations in the gene coding for Escherichia coli DNA topo-isomerase I affecting transcription and transposition. Proc.Natl. Acad. Sci. USA 78:2747-2751.

39. Stonington, 0. G., and D. E. PettUohn. 1971. The folded genomeof Escherichia coli isolated in a protein-DNA-RNA complex.Proc. Natl. Acad. Sci. USA 68:6-9.

40. Sugino, A., C. Peebles, K. Kruezer, and N. Cozzarelli. 1977.Mechanism of action of nalidixic acid: purification of Esche-richia coli nalA gene product and its relationship to DNA gyraseand a novel nicking-closing enzyme. Proc. Natl. Acad. Sci.USA 74:4767-4771.

41. Trucksis, M., and R. Depew. 1981. Identification and localiza-tion of a gene that specifies production of Escherichia coli DNAtopoisomerase I. Proc. Natl. Acad. Sci. USA 78:2164-2168.

42. Tse-Dinh, Y. 1985. Regulation of the Escherichia coli DNAtopoisomerase I gene by DNA supercoiling. Nucleic Acids Res.13:4751-4763.

43. Tse-Dinh, Y., and R. Beran. 1988. Multiple promoters fortranscription of the E. coli DNA topoisomerase I gene and theirregulation by DNA supercoiling. J. Mol. 13iol. 202:735-742.

44. Wang, J. C. 1971. Interaction between DNA and an Escherichiacoli protein. J. Mol. Biol. 55:523-533.

45. Wu, H., S. Shyy, J. C. Wang, and L. F. Liu. 1988. Transcriptiongenerates positively and negatively supercoiled domains in thetemplate. Cell 53:433-440.

VOL. 171, 1989

on Decem

ber 26, 2020 by guesthttp://jb.asm

.org/D

ownloaded from