(Received for publication, October 31, 1996, and in revised form, January 26, 1997)
From the Molecular Biology Program, Memorial Sloan-Kettering Cancer Center, New York, New York 10021
Quinolones are potent broad spectrum antibacterial drugs that target the bacterial type II DNA topoisomerases. Their cytotoxicity derives from their ability to shift the cleavage-religation equilibrium required for topoisomerase action toward cleavage, thereby effectively trapping the enzyme on the DNA. It has been proposed that these drugs act by binding to the enzyme-DNA complex. Using catalytically inactive and quinolone-resistant mutant topoisomerase IV proteins, nitrocellulose filter DNA binding assays, and KMnO4 probing of drug-DNA and drug-DNA-enzyme complexes, we show: (i) that norfloxacin binding to DNA induces a structural alteration, which probably corresponds to an unwinding of the helix, that is exacerbated by binding of the topoisomerase and by binding of the drug to the enzyme and (ii) that formation of this structural perturbation in the DNA precedes DNA cleavage by the topoisomerase in the ternary complex. We conclude that cleavage of the DNA and the resultant opening of the DNA gate during topoisomerization requires the induction of strain in the DNA that is bound to the enzyme. We suggest that quinolones may act to accelerate the rate of DNA cleavage by stimulating acquisition of this structural perturbation in the ternary complex.
DNA topoisomerases are the targets of potent anti-tumor drugs and broad-spectrum antibacterial drugs (1-3). Treatment with these drugs leads to the generation of double-strand DNA breaks and cell death (2, 3). Both bacterial type II topoisomerases, DNA gyrase and topoisomerase IV (Topo IV),1 are sensitive to the quinolone family of antibacterial drugs (4-7). Spontaneous resistance of both Gram-positive and Gram-negative bacteria to the quinolones (1, 8-10) has been shown to arise in gyrA and gyrB, encoding the subunits of gyrase (11, 12), and in parC, encoding one of the subunits of Topo IV (6, 7). It has also recently been shown that a mutant parC allele (parCS80L) will confer quinolone resistance to Escherichia coli (13).
Kreuzer and Cozzarelli (11) proposed that the quinolones acted as poisons, trapping the topoisomerase on the DNA, and that this frozen complex led to the generation of a cytotoxic event. It is now clear that a number of structurally dissimilar drugs, such as the quinolones, camptothecin, and the epipodophyllotoxins, act on eukaryotic and prokaryotic topoisomerases in the same manner (2, 3, 8). These drugs affect the same point in the catalytic cycle of the enzymes, increasing the equilibrium constant for the cleavage-religation step necessary for opening and closing the DNA gate. These drugs differ in that some poisons accomplish this shift in equilibrium by inhibiting the rate of religation, whereas others appear to increase the rate of cleavage (2, 3). The molecular details of how these drugs act, however, are unknown.
The isolation of point mutations conferring quinolone resistance in the genes encoding the gyrase subunits strongly suggested that the drugs bound the enzyme (1, 8). However, Shen and Pernet (14) demonstrated that quinolones bound neither GyrA nor GyrB, but did bind double-stranded DNA. Subsequent studies showed that gyrase stimulated drug binding to DNA (15). This led to a model in which it was suggested that the quinolones bound to a gyrase-DNA complex after strand cleavage, using the exposed single strands in the staggered 4-base pair cut as a target and preventing religation (16). This model has been challenged with the demonstration that the quinolone CP-115,953 increases the rate of DNA cleavage without affecting religation (17) and the finding by Critchlow and Maxwell (18) that DNA cleavage is not required for the binding of [3H]ciprofloxacin to a complex of gyrase bound to DNA.
We have used mutant Topo IV proteins to show that DNA strand cleavage was required for replication fork arrest by a frozen quinolone-DNA-Topo IV complex (19). This collision converts the normally reversible ternary complex to a nonreversible form but does not induce a double-strand break. This presumably occurs subsequently as a result of a aborted repair attempt.
We have continued our analysis of the mechanism of quinolone action by studying the interaction of norfloxacin with DNA and with topoisomerase-DNA complexes. These complexes were formed using the wild-type Topo IV, a quinolone-resistant mutant, a catalytically inactive protein, or an inactive and drug-resistant version of Topo IV. Norfloxacin stabilized the binding of the wild-type and all mutant forms of Topo IV to a DNA binding site of defined sequence. In addition, we found that the drug alone induced a structural alteration in the DNA binding site that was detectable by oxidation of thymine residues with KMnO4. The extent of this structural perturbation was increased dramatically by the binding of either wild-type or inactive Topo IV and was correspondingly reduced if the enzyme was also quinolone-resistant.
Enzymes, DNAs, and Reagents
ParC S80L, ParC Y120F, ParC Y120F,S80L, and wild-type ParC and
ParE were purified as described previously (7, 19). Wild-type and
mutant Topo IV proteins were reconstituted by mixing the appropriate subunits together in their storage buffers to a final concentration of
9 µM heterotetramer and incubating on ice for 30 min.
pBS ± DNA was from Stratagene. Oligonucleotides were synthesized
on an Applied Biosystems DNA synthesizer. Oligonucleotides used for Topo IV binding assays were gel-purified through denaturing
polyacrylamide gels containing 45% urea and 35% formamide before
5-end-labeling using [
-32P]ATP (Amersham Corp.) and
T4 polynucleotide kinase (PL Biochemicals). Labeled oligonucleotides
were then gel-purified through native polyacrylamide gels. Norfloxacin
(Sigma) was dissolved and diluted in 10 mM NaOH. Stock
concentrations were adjusted such that final binding reaction mixtures
contained 1 mM NaOH.
Superhelical DNA Relaxation Assay
Reaction mixtures (20 µl) containing 40 mM HEPES-KOH (pH 8.0), 10 mM MgOAc, 10 mM dithiothreitol, 100 µg/ml BSA, 2 mM ATP, 4 µg/ml tRNA, 10 nM form I pTH201 DNA (20), and the indicated amounts of Topo IV proteins and norfloxacin were incubated at 30 °C for 30 min. EDTA was then added to 50 mM, and the incubation was continued at 37 °C for 1 min. SDS and proteinase K were then added to 1% and 100 µg/ml, respectively, and the incubation was continued at 37 °C for 15 min. A gel loading dye was then added, and the samples were analyzed by electrophoresis at 2 V/cm for 16 h through vertical 1% agarose gels using 50 mM Tris-HCl (pH 7.9 at 25 °C), 40 mM NaOAc, and 1 mM EDTA as the electrophoresis buffer. The gels were then stained with ethidium bromide and photographed.
DNA Cleavage Reactions
Cleavage Site MappingReaction mixtures (20 µl)
containing 40 mM Tris-HCl (pH 7.5 at 37 °C), 10 mM MgCl2, 10 mM dithiothreitol, 1 mM ATP, 100 µg/ml BSA, 1.5 nM 5-end-labeled
DNA fragment, and 45 nM Topo IV were incubated at 37 °C
for 10 min. SDS was then added to 1%, and the incubation was continued
for 2 min. EDTA, proteinase K, and tRNA were then added to 25 mM, 100 µg/ml, and 40 µg/ml, respectively, and the
incubation was continued for an additional 15 min. DNA products were
purified by extraction of the reaction mixture with phenol-CHCl3 (1:1, v/v) and recovered by ethanol
precipitation. The DNA was then dissolved in sequencing gel loading dye
(Amersham) and the products analyzed by electrophoresis through a 6%
sequencing gel. Sequence ladders were generated using the
oligonucleotides listed in Table I by dideoxy sequencing as described
by the manufacturer (Amersham). Gels were dried and then
autoradiographed.
|
Reaction mixtures (10 µl) were as for cleavage site mapping above except that ATP was omitted, tRNA was included at 5 µg/ml, the 24-mer oligonucleotide was at 5 nM, Topo IV was at 600 nM, and norfloxacin was as indicated. Incubation for cleavage in 1% SDS was for 10 min at 37 °C. After ethanol precipitation, the DNA was resuspended in formamide, bromphenol blue and xylene cyanol were added to 0.1%, and the cleavage products analyzed by electrophoresis through a 13.5% polyacrylamide gel (19:1 acrylamide to bisacrylamide) containing 45% urea and 35% formamide using 100 mM Tris borate (pH 8.3) and 2 mM EDTA as the electrophoresis buffer. Gels were fixed by soaking in 10% methanol, 7% HOAc, and 5% glycerol before drying onto DE-81 paper.
Assessment of Norfloxacin-induced DNA Fragment CleavageReaction mixtures were as for assessment of norfloxacin-induced oligonucleotide cleavage above except that the DNA fragment was at 3 nM and Topo IV was at 60 nM. Analysis was as for cleavage site mapping above.
Nitrocellulose Filter DNA Binding Assay
DNA-binding reaction mixtures (10 µl) containing 50 mM Tris-HCl (pH 7.5 at 30 °C), 20 mM KCl, 10 mM MgCl2, 5 mM dithiothreitol, 50 µg/ml BSA, 20 µg/ml tRNA, 6% gycerol (from the Topo IV fractions), 2 nM 32P-labeled duplex oligonucleotide, and
the indicated amounts of either wild-type or mutant Topo IV were
incubated at 30 °C for 15 min. The reactions were then diluted to
200 µl with washing buffer (25 mM Tris-HCl (pH 7.5 at
30 °C), 20 mM KCl, 10 mM MgCl2, and 10 mM -mercaptoethanol) and filtered through
nitrocellulose filters (Millipore HAWP) at 1 ml/min. The filters were
washed three times with washing buffer (1 ml) and then dried under a heat lamp, and the radioactivity retained was determined by liquid scintillation spectrometry. In experiments utilizing norfloxacin, both
the reaction mixture and the washing buffer contained the indicated
concentration of drug. Filters were also soaked in binding buffer
containing the appropriate concentration of drug prior to use.
KMnO4 Sensitivity
Reaction mixtures (25 µl) containing 50 mM
Tris-HCl (pH 7.5 at 30 °C), 20 mM KCl, 10 mM
MgCl2, 50 µg/ml BSA, 5 µg/ml tRNA, 2 nM
5-32P-labeled duplex oligonucleotide, 0.9 µM
Topo IV protein, and 0.5 mM norfloxacin (when indicated)
were incubated at 30 °C for 10 min. KMnO4 was then added
to 1 mM, and the incubation was continued for 3 min at room
temperature.
-Mercaptoethanol was then added to 600 mM,
and the incubation was continued for 2 min at room temperature. EDTA
was then added to 50 mM, and the incubation was continued
for 3 min at 30 °C. The volume of the reaction mixture was then
adjusted to 50 µl with 50 mM EDTA. tRNA (1 µg in 1 µl) was added, and the DNA products were purified by extraction with a phenol-CHCl3 mixture (1:1, vol:vol) and then recovered by
ethanol precipitation. The DNA was resuspended in 50 µl of 1 M piperidine, covered with mineral oil, and incubated at
90 °C for 40 min. The samples were then lyophilized, resuspended in
100 µl of H2O, and lyophilized again. Electrophoretic
analysis was as described above in the assessment of
norfloxacin-induced oligonucleotide cleavage for the DNA cleavage
reactions. Maxam-Gilbert cleavages of the oligonucleotides were
prepared as described elsewhere (21).
To understand the effects of quinolones on the type II bacterial enzymes, we exploited a series of mutant topoisomerases and a defined DNA binding site. A characterization of the mutant topoisomerases is described in this section, and the derivation of a DNA binding sequence for Topo IV is described in the next section.
Our experiments required Topo IV proteins that were drug-resistant, catalytically inactive, or both. We have already described the construction of such mutant ParC proteins (19). These are: (i) ParC S80L that, when combined with ParE, forms a Topo IV that is resistant to quinolones (13, 19); (ii) ParC Y120F that, when combined with ParE, forms an inactive Topo IV (19); and (iii) the double mutant ParC Y120F,S80L. Our initial report (19) using these proteins described their activity only during oriC DNA replication, we describe here a direct characterization of their catalytic activities.
The specific activity of the ParC S80L Topo IV during the relaxation of
superhelical DNA was identical to that of the wild-type enzyme (Fig.
1A). However, this enzyme was only inhibited
in this assay by about 50% at 150 µM norfloxacin (Fig.
1B), a concentration nearly 20-fold greater that the
apparent Ki for the wild-type enzyme (13). Neither
the ParC Y120F nor the ParC Y120F, S80L Topo IV protein could relax
superhelical DNA even when present at concentrations 100-fold greater
than that required for complete relaxation by the wild-type enzyme
(Fig. 2A).2
Because this inactivity could be the result of the failure to execute
one of a number of different steps in the topoisomerization cycle, we
investigated whether these mutant proteins could cleave DNA.
Disruption of the cleavage-religation equilibrium by the addition of a
protein denaturant results in the covalent attachment of the cleavage
subunit to the DNA via a 5-phosphotyrosine linkage (22,
23). If the DNA is 32P-labeled, this results in transfer of
the label to GyrA, in the case of gyrase (24), or ParC, in the case of
Topo IV (6, 7). To assess this, gyrase, ParC, and ParC Y120F and
wild-type Topo IV were bound to uniformly labeled pBS ± DNA, the
DNA was then treated with DNase I, and the protein was denatured by the addition of SDS. The products were analyzed by SDS-polyacrylamide gel
electrophoresis (Fig. 2B). Bands migrating at molecular
weights of 105 and 83 kDa were generated when gyrase and wild-type Topo IV, respectively, were present in the reaction mixtures. This corresponds to the molecular weight of GyrA (25) and ParC (7), respectively. Under the same reaction conditions, no labeled
protein-DNA complexes could be detected when either ParC alone or the
ParC Y120F Topo IV was present in the reaction mixtures, indicating that these proteins do not possess a functioning DNA cleavage and
religation activity.
To be able
to appreciate the DNA sequence characteristics preferred by Topo IV
when it binds, we generated a series of Topo IV-catalyzed DNA cleavages
on pBS ± DNA and compared the sequences of the cleavage sites.
The DNA was first digested independently with five of the seven
restriction endonucleases, which all cleave the DNA only once, shown in
Fig. 3A. (EcoRI and
BsaI were not used for the initial linearization.) The
linear DNAs were dephosphorylated and then 5-end-labeled using
[
-32P]ATP and polynucleotide kinase. Each DNA was then
digested again as described in the legend to Fig. 3A to
generate two fragments (right (downstream of the original cleavage
site) and left (upstream of the original cleavage site)) that were each
labeled uniquely on one 5
-end. These fragments were recovered after
gel electrophoresis and used as substrates for norfloxacin-induced,
Topo IV-catalyzed DNA cleavage. The cleavage products were then
separated by electrophoresis through a sequencing gel (see Fig.
3B for an example). The sequences of the cleavage sites were
determined by comparison to dideoxy sequence ladders generated from
primers (listed in Table I) that had the same 5
-ends as
the DNA substrates.
Derivation of a defined DNA binding site for
Topo IV. A, map of pBS ± DNA showing the restriction
enzymes used in the analysis. DNAs linearized by digestion with
NdeI, BamHI, AflIII, ScaI,
and AatII were labeled at their 5-ends and then redigested with EcoRI, NdeI and AlwNI,
EcoRI and BsaI, NdeI and
AlwNI, and EcoRI and ScaI,
respectively, to generate the DNA fragments used for mapping Topo IV
cleavage sites. B, sample gels of cleavage site mapping.
Left, a segment of the the region upstream of the BamHI cleavage site corresponding to nucleotides 804
(top arrow) to 865
(bottom arrow).
Right, a segment of the region downstream of the
BamHI cleavage site corresponding to nucleotides 1016 (bottom arrow) to 1102 (top arrow). C,
top, a compilation of all 61 cleavage sites used arranged by
frequency of base occurrence centered about the cleavage site
(arrow). Middle, a compilation of the ten
strongest cleavage sites, determined as described under "Results."
Bottom, sequence of the 50-mer. The arrow pointing
down indicates the cleavage site on the top strand, whereas the
arrow pointing up indicates the cleavage site on the bottom
strand. The black dots under the sequence indicate the limits of the 24-mer. The
nucleotides in bold face indicate the position of the
AccI site and the underlined nucleotides indicate
the positions of the HpaII sites.
The sequences of 61 distinct cleavage sites (Table II)
were determined and are shown in Fig. 3C as a function of
nucleotide frequency aligned about the Topo IV cleavage site. Summaries
of all the sites and of the 10 strongest cleavage sites (as determined by comparing the relative intensities of the cleaved DNA fragment to
the uncleaved DNA at the top of the gel) are shown. Inspection of these
data revealed a preference for cleavage after a purine with the 3-base
being either an A or a T. C was the least preferred 3
-base. There
seemed to be little else in the way of distinguishing sequence features
downstream of the cleavage site. On the other hand, preferences could
be detected for a G-rich track just upstream of the cleavage site as
well as an A-rich track starting about 10-12 bases upstream of the
cleavage site.
|
We have shown that Topo IV protects about 34 bp from DNase I digestion when it is bound to the DNA (26). Modeling of a duplex DNA across the putative DNA gate binding site in the crystal structure of the large fragment of yeast topoisomerase II suggests that about 18-20 bp of DNA could be contained within the van der Waals radius of the protein (27). We therefore designed two oligonucleotides for the study of Topo IV binding to DNA, a 50-mer, and a 24-mer. The oligonucleotides were designed as perfect palindromes with the cleavage sites offset 2 bp from the center. In addition, an AccI site was placed at the center of the oligonucleotide and a HpaII site 5 bp offset from the center. This allowed us to easily distinguish between the hairpin form (digested by HpaII but not by AccI) and the duplex form (digested by both enzymes).
The binding of the wild-type and various mutant Topo IV proteins to the
duplex forms of these two oligonucleotides was tested using a
nitrocellulose filter DNA-binding assay (Fig. 4). All four Topo IV proteins bound the 50-mer with similar
KD values (between 0.4 and 0.9 µM),
with the versions carrying the ParC Y120F mutation having a roughly
2-fold greater affinity than those that did not carry this mutation
(Fig. 4A). This difference was much more dramatic when the
binding of the enzymes to the 24-mer was compared (Fig. 4B).
In this case, at 600 nM enzyme, the ParC Y120F and ParC
Y120F,S80L Topo IV proteins bound the DNA with affinities 9-18-fold
greater than both the wild-type and the ParC S80L Topo IV. However, all
proteins displayed an overall decrease in binding compared with that
observed when the 50-mer was used as a substrate. None of these
proteins showed any affinity for the hairpin form of either
oligonucleotide (data not shown).
The stabilization of DNA binding by the presence of the ParC Y120F
mutation seemed counterintuitive, because the enzymes with a wild-type
active site can form covalent complexes with the DNA (cleavage of the
24-mer is shown in Fig. 6). However, it could be explained if the
kinetic pathway of DNA binding is different for the enzymes carrying
the active site mutation. It is possible that with the active enzymes,
a branch point exists after religation that is dissociation from the
DNA. This branch point may not exist for the active site mutant
proteins because the conformational change required to trigger it
(actuated by DNA cleavage) is never executed.
DNA Strand Cleavage Is Not Required for Norfloxacin Stimulation of the Binding of Topo IV to DNA
The effect of norfloxacin on Topo
IV binding to the 24-mer was examined. Norfloxacin stimulated the
binding of both the wild-type enzyme and the ParC Y120F Topo IV (Figs.
5, A and B). Stimulation of the
wild-type enzyme was 6-fold greater than that observed with the ParC
Y120F enzyme and had a [norfloxacin]1/2 of 104 µM. However, the extent of stimulation should not be
taken as anything other than a relative effect, because the binding of
the wild-type enzyme to the 24-mer is so poor in the absence of drug
(Fig. 4B). In addition, some portion (as determined by the
equilibrium constant) of the extent of norfloxacin stimulation of the
binding of the wild-type enzyme results from covalent complex
formation. Binding of the ParC Y120F Topo IV to the DNA was maximally
stimulated about 8-fold by norfloxacin, with a
[norfloxacin]1/2 of 23.6 µM. This
stimulation was reduced by more than half when the ParC S80L quinolone
resistance-conferring mutation was also present (Fig. 5B).
How do these data relate to the inhibition by norfloxacin of the superhelical DNA relaxation activity of Topo IV, which occurs at a Ki of roughly 2 µM (13)? At first glance, because the concentrations of drug required for the effects with the 24-mer DNA substrate were considerably greater than that required for inhibition of activity, it would seem that there was little connection. However, we noted that the affinity of wild-type Topo IV for superhelical plasmid DNA (26) was nearly 50-fold greater than that measured here for the 24-mer, similar to the difference between the Ki and the [norfloxacin]1/2 for stimulation of binding to the oligonucleotide. We therefore considered that the results with the 24-mer did, in fact, reflect accurately the formation of a ternary complex between enzyme, drug, and DNA that leads to inhibition of activity and that the requirement for higher concentrations of drug was a result of a decreased stability of the complex.
The [norfloxacin]1/2 required for stimulation of DNA cleavage by Topo IV should be a direct measure of the affinity of the drug for the ternary complex. Thus, if our explanation was correct, we should observe that the [norfloxacin]1/2 for cleavage of the 24-mer was similar to that for stimulation of binding of the enzyme, whereas the [norfloxacin]1/2 for cleavage of a large DNA fragment was similar to the Ki. This proved to be the case.
Norfloxacin-induced, Topo IV-catalyzed cleavage of the 24- mer was
measured by gel electrophoresis (Fig. 6A).
Three distinct Topo IV-catalyzed cleavage sites could be detected.
These were the predicted site, between the G and the T at positions 10 and 11 (Fig. 3C), respectively, as well as at each
nucleotide flanking the predicted site (see Fig. 8 for the sequence
alignment).3 A similar pattern of cleavage
was found on the 50-mer (data not shown). The reason for this, at the
moment, is unclear. It could imply either that there exists a mixture
of Topo IV molecules bound to the 24-mer that are out of register with
one another by one nucleotide or that there is some play in the
position of the scissile bond as a result of fluidity of the bound DNA
across the active site of the enzyme. We cannot distinguish presently between these possibilities.
The [norfloxacin]1/2 for cleavage of the 24-mer was 130 µM (Fig. 6B), nearly identical to the [norfloxacin]1/2 for stimulation of DNA binding (Fig. 5A).
To test the second half of our prediction, we measured
norfloxacin-induced, Topo IV-catalyzed cleavage of the
BamHI-left DNA fragment used for determination of the DNA
sequence of the oligonucleotide substrates (Fig. 7). The
[norfloxacin]1/2 for the appearance of the major cleavage
product closest to the 5-end of the fragment (the arrow in Fig.
7A) was 5.3 µM (Fig. 7B), very
similar to the Ki. Therefore, our findings on the
formation of a ternary complex between norfloxacin, the 24- mer, and
Topo IV, are an accurate reflection of the molecular events that
underlie inhibition of the topoisomerization activity of the
enzyme.
Thus, based on the data presented in Figs. 5, 6, 7, we conclude that formation of the quinolone-Topo IV-DNA ternary complex does not require a functioning DNA strand cleavage and religation activity. Neither does it require the presence of ATP (Table III). Moreover, this suggests that, at least in part, quinolones act to affect a step in the catalytic topoisomerization cycle that precedes strand cleavage. The most likely steps that could be affected are those that involve the interaction of the enzyme with the DNA. We therefore investigated whether the ternary complex displayed an altered interaction with DNA compared with in the absence of drug.
|
Several chemical methods
of probing the state of the DNA in the ternary complex were
investigated. The most revealing proved to be KMnO4
oxidation of thymine residues followed by cleavage of the DNA with
piperidine (28). In these experiments (Figs. 8 and
9), Topo IV binding to the DNA was allowed to come to
equilibrium in the presence or absence of 1 mM norfloxacin.
KMnO4 was added to the reaction mixtures for a brief period
and then inactivated by the addition of mercaptan. Any cleavable
complexes still present were allowed to reversibly dissociate by a
subsequent incubation in the presence of EDTA before purification of
the DNA products by extraction with phenol-CHCl3. This
accounts for the absence of Topo IV-catalyzed cleavage products in the
lanes containing the active enzymes.
In the case of both the active and inactive enzymes, formation of a
ternary complex resulted in an increase in the reactivity of the T
residues within the 4-bp stagger between the cleavage sites,
10GTATA
C15. In the
case of the active enzyme, this reactivity was diminished when the
quinolone resistance-conferring ParC S80L mutation was present. The
extent of this reduction in reactivity was consistent with the
magnitude of norfloxacin-induced stimulation of the binding of the
wild-type Topo IV to the 24-mer (Fig. 5A). Similarly, the ParC S80L mutation decreased the reactivity of the thymines by about
50% in the ParC Y120F background, also consistent with its effect on
the extent of stimulation of binding of the inactive Topo IV to the
24-mer by norfloxacin (Fig. 5B). Reactivity could also be
detected, using different exposures of the autoradiogram, at the T
residues at positions 4 and 21-24. Moreover, we noted, although this
was variable from experiment to experiment, that norfloxacin alone
caused slight, but significant, reactivity at the same residues
(compare the dil lanes in Fig. 8). Thus, it seemed as if the
binding of norfloxacin to the DNA caused a structural distortion that
resulted in the increased reactivity of some T residues and that this
was magnified considerably in the presence of Topo IV. We investigated
whether either the effect of norfloxacin on the DNA alone might be more
apparent or the pattern of reactivity in the presence of
topoisomerase might be different with the longer oligonucleotide.
When the 50-mer was used as a substrate (Fig. 9), the effect of norfloxacin alone on the DNA could be observed clearly as an increase in reactivity of the T residues at positions 8, 11, and 35-39 (compare lanes 3 and 4). The reactivity of the T residues in the 4-bp stagger (positions 24 and 26) between the Topo IV-catalyzed cleavage sites was much less obvious. As with the 24 mer, the presence of the ParC Y120F Topo IV significantly increased the reactivity of the same T residues. In addition, the reactivity of the T residues in the 4-bp stagger could now be observed. In this case, the presence of the ParC S80L mutation appeared to reduce the reactivity of the T residues to that observed in the presence of the drug alone.
It was interesting that with the 50-mer, unlike with the 24- mer, the majority of the reactivity was observed distal from the scissile bonds. The reason for this is not immediately obvious; however, it could reflect the local stability of the base pairs in these short DNA fragments, with the more breathable (A + T)-rich clusters in the 50-mer, which are not present in the 24-mer, serving as focuses for release of any strain in the DNA.
DNA topoisomerases have proved to have extraordinary clinical relevance. Both the most potent broad-spectrum antibacterial drugs and some of the most promising anti-tumor chemotheraputic agents target these enzymes in prokaryotes and eukaryotes, respectively. Intensive research has produced a large body of data that has shed considerable light on the mechanism of action of these drugs. These cellular poisons act by affecting the cleavage-religation equilibrium required for strand passage during topoisomerization (1-3). The net effect is that, in the presence of drug, the topoisomerase spends considerably more time covalently bound to the DNA. This significantly increases the likelihood of the generation of a double-stranded DNA break, the presumed cytotoxic event.
The mechanism by which these double strand breaks arise is still uncertain, although we have recently demonstrated that a ternary complex of norfloxacin, DNA, and Topo IV will arrest progression of a replication fork (19). This collision converts the normally reversible cleavable complex to a nonreversible form. DNA breakage then must occur by a second pathway that might, for example, involve an aborted repair event. In the current report, we have extended our investigation on the mechanism of quinolone action.
Consistent with the finding of Critchlow and Maxwell (18), who demonstrated that quinolone could bind to a complex of DNA and an active site mutant DNA gyrase, we have shown that quinolone can likewise stimulate the binding of both wild-type and inactive Topo IV to DNA. In addition, the extent of quinolone-stimulated DNA binding could be attenuated in the ParC Y120F background if the enzyme carried a quinolone resistance-determining mutation. This therefore establishes a connection between the ability of quinolone to stimulate topoisomerase binding to DNA in the absence of a strand cleavage and religation activity and the genetic determinant of quinolone resistance. This suggests that under normal circumstances, i.e. with the wild-type enzyme, much of what quinolones do transpires before actual DNA cleavage occurs. Thus, models that propose an interaction of quinolone with the single-stranded DNA in the covalent complex of topoisomerase and cleaved DNA, thereby preventing religation, are unlikely to be true.
The studies of Robinson et al. (17) demonstrated that, at least for the Drosophila type II topoisomerase, a quinolone, CP-115,953, affected the cleavage-religation equilibrium by increasing the rate of cleavage. The rate of religation remained constant. This is in contrast to the action of etoposide, which decreased the rate of religation. Similar studies have not been performed with the bacterial type II topoisomerases because of the lack of defined substrates. However, we have shown that Topo IV interacts with DNA as if it were the prototypical eukaryotic enzyme rather than a DNA gyrase (26). Quinolones could therefore act in a similar manner with the bacterial enzymes to enhance the rate of DNA strand cleavage.
In this light, it is interesting to consider the molecular events suggested by our data. Using KMnO4 oxidation of thymine residues, we have detected a distortion in the DNA bound to the DNA gate. This structural perturbation requires the presence of quinolone and was observed with both wild-type and catalytically inactive Topo IV. Thus, it precedes actual strand cleavage. The nature of the structural perturbation is unclear. Increased reactivity to KMnO4 oxidation of thymine residues in double-stranded DNA is generally interpreted as resulting from unpairing of the base pair (28). Thus, it is possible that the DNA in the quinolone-Topo IV-DNA complex is unwound or kinked.
Generation of this structural distortion in the DNA in the ternary complex appears to have at least two components. The drug alone seems capable of distorting the DNA somewhat. However, the extent of distortion is much greater in the ternary complex and use of a quinolone-resistant enzyme obviates the increased distortion observed in the presence of the topoisomerase. Thus, it is likely that in the ternary complex, drug binds both the DNA and the enzyme. This is consistent with amino acid homologies between the site of quinolone resistance-conferring mutations in the bacterial type II enzymes, GyrA S83 and ParC S80, and S741 of the yeast type II enzyme that locates to a helix-turn-helix structure close to the active site Tyr.
It seems likely, therefore, that binding of DNA to the DNA gate requires unwinding of the substrate. This distortion may result from an induced fit of the DNA across the binding site or could result from a conformational change in the enzyme that is triggered by DNA binding. Because the structural distortion was not observed in the absence of norfloxacin under conditions where the DNA was bound to the enzyme, we currently favor the latter possibility. This step could be necessary for the enzyme to proceed to the far more dramatic conformational changes required for actual strand cleavage and strand passage. This requires that a duplex DNA, with a diameter of about 20 Å, transit between the ends of a 4-bp staggered cut in another section of the duplex. Even disregarding the steric obstacle of the overlapping single strands at the cut, this would require separation of the scissile bonds by at least 7 or 8 Å. The actual separation is likely to be much greater. This can only be accomplished by significant distortion of the axis of the duplex DNA that is bound to the DNA gate binding site. This is supported by the structure of a fragment of the yeast type II enzyme reported by Berger et al. (27) that presumably represents the form of the enzyme in which the DNA gate is open. In this structure the active site tyrosines are separated by over 30 Å.
It is interesting to speculate, then, how quinolones could act to shift the cleavage-religation equilibrium. If, during the wild-type catalytic cycle, strand cleavage is preceded by a conformational change in the enzyme that distorts the DNA bound to the DNA gate, quinolones may act to accelerate strand cleavage by mimicking this structural distortion when they are bound to DNA such that the activation energy required in the drug-topoisomerase-DNA ternary complex to assume this state decreases compared with that which obtains in the absence of the drug. Thus, an enhancement in the rate of topoisomerase-catalyzed strand cleavage would be expected in the presence of drug.
We thank Stewart Shuman for his critical reading of the manuscript and David Valentine for the artwork.