Abstract
Gyrase is the only type II topoisomerase that introduces negative supercoils into DNA. Supercoiling is catalyzed via a strand-passage mechanism, in which the gate DNA (gDNA) is transiently cleaved, and a second DNA segment, the transfer DNA (tDNA), is passed through the gap before the gDNA is religated. Strand passage requires an opening of the so-called DNA-gate by ≈2 nm. A single-molecule FRET study reported equal populations of open and closed DNA-gate in topoisomerase II. We present here single-molecule FRET experiments that monitor the conformation of DNA bound to the DNA-gate of Bacillus subtilis gyrase and the conformation of the DNA-gate itself. DNA bound to gyrase adopts two different conformations, one slightly, one severely distorted. DNA distortion requires cleavage, but neither ATP nor the presence of a tDNA. At the same time, the DNA-gate of gyrase is predominantly in the closed conformation. In agreement with the single molecule data and with the danger of dsDNA breaks for genome integrity, <5% of cleavage complexes are detected in equilibrium. Quinolone inhibitors favor DNA cleavage by B. subtilis gyrase, but disfavor DNA distortion, and the DNA-gate remains in the closed conformation. Our results demonstrate that DNA binding, distortion and cleavage, and gate-opening are mechanistically distinct events. During the relaxation and supercoiling reactions, gyrase with an open DNA-gate is not significantly populated, consistent with gate-opening as a very rare event that only occurs briefly to allow for strand passage.
Keywords: ATPase, negative supercoiling, topoisomerase, single molecule FRET
DNA topoisomerases modulate DNA structure by interconverting different DNA topoisomers (1). Gyrases are bacterial type II topoisomerases that use the chemical energy of ATP hydrolysis to introduce negative supercoils into DNA (2). The active form of gyrase is a heterotetramer formed by two GyrA and two GyrB subunits. Supercoiling is catalyzed via a strand-passage mechanism and involves the coordinated opening and closing of 3 different protein interfaces or gates, termed N-gate, DNA-gate, and C-gate (3–5) (Fig. 1A). As a first step, one segment of double-stranded DNA, the gate DNA (gDNA), binds to the DNA-gate formed by the GyrA subunits (6) that harbor the catalytic tyrosines (7). The transesterification reactions lead to the formation of a covalent protein-DNA intermediate (8). The N-gate is formed by the GyrB subunits, which dimerize in the presence of ATP and trap the transfer DNA (tDNA) segment (9–12). The tDNA can then pass through the gap created in the gDNA. After religation of the gDNA, the tDNA leaves the enzyme through the C-gate formed by the GyrA subunits (13, 14). A concerted opening and closing of these gates is a prerequisite for tight coupling of DNA cleavage and strand passage during the supercoiling reaction. The so-called double lock rule has been proposed as a simple principle to ensure unidirectional strand passage in type II topoisomerases (15, 16); each gate can only adopt an open conformation if the two others are closed.
Fig. 1.
A model gDNA substrate to study the gDNA conformation. (A) Schematic picture of gyrase and the N-, DNA- and C-gates. The active enzyme consists of two GyrA (dark gray) and two GyrB subunits (light gray). Oval, ATP binding sites in GyrB; Y, active site tyrosines in GyrA. (B) Central region of the 60-bp gDNA substrate with a preferred B. subtilis gyrase cleavage site (arrows), flanked by donor (A488, D) and acceptor (A546 or A555, A) fluorophores. The box indicates the position of the phosphorothiolate (pt) modification. (C) Anisotropy titration of A488/A546-labeled DNA with GyrA in the presence of 8 μM GyrB. The Kd value is 51 ± 5 nM. (D) Cleavage of the gDNA by gyrase, and effects of ATP, ADPNP, OXO, and CFX on cleavage. Noncleaved 60mer (upper band) and the 24mer released upon cleavage (lower band) are indicated. X equals O in nonmodified DNA, and S in pt-DNA. No cleavage products are detected with nonmodified gDNA and gyrase, gyrase/ATP, or gyrase/ADPNP. Only in the presence of the OXO or CFX the gDNA is cleaved to a significant extent. With DNA carrying a pt modification at the cleavage site the fraction of cleaved DNA is increased compared with nonmodified DNA. Controls for substrate and cleavage products for pt-DNA and nonmodified DNA are shown.
Recently reported single-molecule FRET (smFRET) experiments addressing the DNA-gate conformation of eukaryotic topoisomerase II (17, 18) demonstrated a transition of labeled gDNA between high FRET and low FRET states that were assigned to open and closed states of the DNA-gate. Similar rate constants for the transitions between these two states predicted an open DNA-gate of topoisomerase II during 50% of the time. This result contradicts previous biochemical experiments that showed little DNA cleavage complexes in equilibrium (19) and that have provided strong evidence that a tDNA is required for opening of the DNA-gate (20–22). Here, we present smFRET experiments that resolve this issue by directly monitoring the conformational state of the DNA-gate in Bacillus subtilis gyrase bound to dsDNA and during the relaxation and supercoiling reactions. These FRET experiments demonstrate that the DNA-gate is closed in the presence of linear dsDNA and during DNA relaxation or ATP-dependent DNA supercoiling.Our results are in agreement with a severe distortion of gDNA coupled to DNA cleavage. DNA binding, distortion, and cleavage are mechanistically distinct steps that precede opening of the DNA-gate. Gate opening is a rare event during relaxation and supercoiling that only occurs briefly to allow for strand passage.
Results
A Small dsDNA Substrate Functions as the gDNA for B. subtilis Gyrase.
Gyrases exhibit limited sequence specificity for DNA cleavage. To facilitate studies on the DNA gate conformation by FRET, we designed a 60-bp substrate for gyrase that contains a preferred cleavage site in the center (23), flanked by a donor (A488) and acceptor fluorophore (A546 or A555) (Fig. 1B). Cleavage at the preferred site would produce two fragments with 24/28 and 36/32 bases. When the anisotropy of A546 was used as a probe for DNA binding to gyrase, a Kd of the gyrase/DNA complex of 51 ± 5 nM was obtained (Fig. 1C), confirming high affinity. Next, we tested whether the substrate is cleaved by gyrase and if cleavage occurs at the predicted site (Fig. 1D). Upon incubation with gyrase, no cleavage of the 60-bp DNA substrate was detected, indicating that the cleavage complex is not significantly populated. Quinolones inhibit DNA relaxation, supercoiling, catenation, and decatenation activities of gyrase (24, 25) by favoring the cleavage reaction (19). In the presence of oxolinic acid (OXO) or ciprofloxacin (CFX), the 60-bp DNA substrate was cleaved, indicating that the lack of cleavage products is due to efficient religation. The main product comprises 24 bases, confirming that cleavage occurs preferentially at the sequence identified by Bashkirov and Zvingila (23). Altogether, our results validate the 60-bp dsDNA as an adequate gDNA for B. subtilis gyrase.
dsDNA Bound to the Gyrase DNA-Gate Adopts Two Different Conformations.
smFRET experiments with the donor/acceptor-labeled dsDNA yielded histograms with mean FRET efficiencies of 0.53 ± 0.01 (A488/A546), and 0.42 ± 0.01 (A488/A555) (Fig. 2A and Fig. S1). These FRET efficiencies correspond to donor/acceptor distances of 6.4 or 7.0 (± 0.03) nm (see SI Text and Table S1), in excellent agreement with the expected distance for the dye positions assuming B-DNA geometry (6.7 nm). When the GyrA and GyrB subunits were added to the DNA, a significant broadening of the peak was observed in the FRET histograms for both dsDNA constructs (Fig. 2B and Fig. S1). This broadening strictly depends on the formation of gyrase heterotetramers (Fig. S1) and is independent of the order in which the subunits are added. The histogram is well-described by a sum of two FRET distributions, centered at EFRET = 0.14 ± 0.01 and EFRET = 0.30 ± 0.01 (7.3 and 8.7 nm, respectively). Eighty-five percent of the DNA molecules (85 ± 6%) are in the high FRET conformation, the remaining 15 ± 3% adopt the low FRET conformation. The appearance of a second population raises the question about the identity of the two observed species: The two FRET distributions could either reflect free and bound DNA, or bound DNA in two different conformations. However, free DNA would be expected at EFRET of 0.42 ± 0.01 (not at 0.53 ± 0.01 as in histograms of free DNA because of different correction parameters, see SI Text, Fig. S2, and Table S1). Similar smFRET histograms are obtained when the protein concentrations are doubled, also arguing against two FRET peaks because of free and bound DNA. Hence, the two FRET species reflect different conformations of gDNA in complex with gyrase.
Fig. 2.
gDNA bound to gyrase adopts two different conformations. smFRET histograms of A488/A546-labeled gDNA. (A) DNA only. (B) +gyrase. (C) +gyrase (GyrA_Y123F). (D) +gyrase, ATP. (E) + gyrase, ADPNP. (F) + gyrase (GyrB_E44Q), ATP. Black line in B, D, and E, fit with two Gaussian distributions; light and dark gray lines, individual contributions. Numbers refer to the fractions of each species. DNA bound to gyrase adopts two different conformations, with FRET efficiencies of 0.3 and 0.14, respectively. ATP (filled arrow) does not significantly affect the fraction of the low FRET species, whereas ADPNP (open arrow) disfavors its population. In the presence of ADPNP, with the cleavage-deficient Y123F mutant and with the ATPase-deficient E44Q mutant in the presence of ATP, DNA only adopts the high FRET DNA conformation.
The difference in donor/acceptor distances of 1.4 nm between the two conformations is smaller than the diameter of dsDNA (≈2 nm), rendering it unlikely that the high and low FRET species reflect DNA bound to gyrase with a closed or open DNA-gate, respectively. The low FRET population was not observed with the cleavage-deficient GyrA_Y123F mutant (Fig. 2C, Fig. S3, and Fig. S4). Thus, the appearance of the additional DNA conformation is linked to DNA cleavage. However, with ≈15%, the population of this species is higher than the fraction of DNA that is cleaved under similar conditions (Fig. 1D), and it thus does not reflect the open state of the DNA-gate (see next section). Our smFRET data instead suggest that DNA bound to gyrase in the absence of nucleotides can adopt two different conformations.
We next tested the influence of nucleotides on the distribution of the two populations. Nucleotide binding does not significantly affect the DNA affinity of gyrase [Kd = 25 ± 3 nM (ADPNP), and 75 ± 12 nM (ATP)] (Fig. S5). In the presence of ATP (Fig. 2D and Fig. S3), the fraction of the low FRET population was not significantly affected. The nonhydrolyzable ATP analog ADPNP has been established as a suitable ATP analog for gyrase (3, 9, 11, 26). Addition of ADPNP reduces the low FRET population (Fig. 2E and Fig. S3). The same behavior was observed with ATP and gyrase containing the hydrolysis-deficient GyrB_E44Q mutant (27) (Fig. 2F, Fig. S3, and Fig. S6). Thus, the population of the low FRET conformation is disfavored in the ATP state.
In summary, our data show that the conformations of free DNA and DNA bound to gyrase are different and that DNA bound to cleavage-competent gyrase exists in two distinct conformations. Bound DNA is distorted, with increased inter-dye distances of 7.3 nm (high FRET species) and 8.7 nm (low FRET species) compared with B-DNA (6.7 nm).
Populating Cleavage Complexes by Gyrase Inhibitors and Phosphorothiolate-Modified DNA Does Not Open the DNA-Gate.
To unambiguously assign the two conformational states observed in the FRET histograms of gyrase-bound DNA, we attempted to shift the balance from the closed DNA-gate to the open state by using the quinolone inhibitors OXO or CFX. These compounds populate cleavage-complexes and may stabilize gyrase with an open DNA-gate. OXO was used to populate cleavage complexes for the identification of the preferred cleavage site of B. subtilis gyrase (23), and we therefore tested the inhibitory effect of OXO and its derivative CFX on DNA supercoiling activity (Fig. S7). Both compounds inhibited B. subtilis gyrase, with CFX being the more potent inhibitor. Both inhibitors favor the population of cleavage complexes (Fig. 1D). In smFRET experiments of donor/acceptor-labeled DNA and gyrase in the presence of OXO or CFX (Fig. S7), FRET histograms revealed a predominant DNA species with EFRET = 0.38 ± 0.02 (donor/acceptor distance of 6.9 nm), similar to the high FRET species observed before. Opening of the DNA-gate would inevitably lead to the appearance of a low FRET species. The absence of a low FRET species thus provides direct experimental evidence that the DNA-gate must be in the closed state in the presence of quinolones, as suggested from studies with gyrase with a cross-linked DNA-gate (4). Quinolones only weakly interact with DNA (28, 29). When donor/acceptor-labeled DNA was mixed with inhibitors in the absence of gyrase, the FRET efficiency was not affected (Fig. S7), confirming that CFX had no effect on DNA conformation. Interestingly, the donor/acceptor distance in the gyrase-bound DNA substrate in the presence of CFX (6.9 ± 0.1 nm) is identical to the donor/acceptor distance of DNA bound to the cleavage-deficient mutant GyrA_Y123F (6.8 ± 0.1 nm) (Fig. 2C and Fig. S4), suggesting similar conformations in both cases. The data also demonstrate that quinolones do not act via stabilization of an open DNA-gate.
3′-bridging phosphorothiolates populate cleavage complexes of DNA topoisomerase II by disfavoring religation (30). We therefore used the same DNA substrate as before, but with a 3′-bridging phosphorothiolate at the cleavage site of one strand to promote DNA cleavage by gyrase. In the presence of CFX, phosphorothiolate-modified DNA (pt-DNA) is cleaved at the preferred site for gyrase (Fig. 1D). In contrast to nonmodified dsDNA, cleavage complexes are detected in the absence of inhibitors (Fig. 1D). smFRET histograms of the pt-DNA labeled with donor and acceptor (A488/A546) show a peak at EFRET = 0.50 ± 0.01 (6.5 ± 0.03 nm) (Fig. 3A), similar to the nonmodified DNA. As for nonmodified DNA, in the presence of both gyrase subunits, a significant broadening of the peak was observed independent of the order in which the subunits are added. The FRET histograms can be described by two overlapping peaks at EFRET = 0.30 ± 0.05 (7.3 ± 0.3 nm) and 0.14 ± 0.01 (8.7 ± 0.14 nm) (Fig. 3B). Importantly, the fraction of the species with EFRET = 0.14 is increased ≈3-fold (to 40% ± 10%) compared with the experiments with nonmodified DNA (Figs. 2B and 3B). Thus, an increased fraction of cleaved DNA is paralleled by an increased population of the low FRET DNA conformation. With the cleavage-deficient mutant GyrA_Y123F (Fig. S8), only the high FRET species is observed, confirming that the conformational change to the low FRET conformation is linked to strand cleavage. In the presence of CFX, the DNA bound to the cleavage-deficient mutant remains in the high FRET state. As with nonmodified DNA, addition of ATP does not change the distribution of the two populations (33 ± 3% low FRET). ADPNP addition reduces the amplitude of the low EFRET species (6 ± 5%) (Fig. 3 C and D), and experiments with the ATPase-deficient GyrB_E44Q yield similar histograms to wild-type gyrase with ADPNP (Fig. 3D and Fig. S6). Thus, the population of the DNA bound to gyrase with EFRET = 0.14 requires cleavage and increases with the fraction of cleavage complexes. The question remains whether this population reflects the open conformation of the gyrase DNA-gate.
Fig. 3.
A phosphorothiolate (pt)-modification at the cleavage site favors the low FRET conformation. smFRET histograms of A488/A546-labeled gDNA with a pt modification. (A) pt-DNA only. (B) +gyrase. (C) +gyrase, ATP. (D) + gyrase, ADPNP. The low FRET DNA conformation is more prominent with the pt-DNA than with nonmodified DNA. Again, ATP does not affect the distribution, whereas ADPNP disfavors the low FRET conformation. See legend of Fig. 2 for details of the depiction.
The DNA-Gate of Gyrase Is Predominantly Closed with Linear dsDNA During Relaxation and Supercoiling of Plasmid DNA.
To assign unambiguously whether the observed low FRET population corresponds to DNA bound to gyrase with an open DNA-gate, we directly monitored the conformation of the DNA-gate in gyrase. To this end, we constructed mutants of GyrA with cysteines for fluorescent labeling in the vicinity of the DNA-gate (GyrA_H107C, GyrA_S117C, GyrA_T140C, GyrA_D145C, and GyrA_E211C) (Fig. 4A). FRET efficiencies for donor/acceptor-labeled GyrA were 0.52 ± 0.02 (H107C), 0.94 ± 0.01 (S117C), 0.95 ± 0.01 (T140C), 0.63 ± 0.01 (D145C), and 0 ± 0.01 (E211C) (Tables S2 and S3). The corresponding distances agree reasonably well with the expected distance from the crystal structure of the homologous E. coli GyrA (PDB-ID 1ab4; see SI Text and Table S3), confirming that the conformation of the DNA-gate in B. subtilis GyrA resembles the closed conformation captured in the E. coli GyrA structure. To compare identical conditions between experiments with donor/acceptor-labeled gDNA and unlabeled gyrase, smFRET experiments with (unlabeled) gDNA and labeled gyrase were performed. Consistent results were obtained for all mutants. Data is shown GyrA_T140C only because the expected distance changes for the open and closed conformations of the DNA-gate would be largest, as judged from a comparison of distances between corresponding residues in the structure of E. coli GyrA (PDB-ID 1ab4, closed conformation) and a Saccharomyces cerevisiae topoisomerase II fragment (1bjt, partially open conformation). For this construct, an increase in the interdye distance of ≈2 nm would be expected, leading to an EFRET for the open conformation of ≈0.56 that would be clearly distinguishable from the closed conformation (EFRET = 0.95). The relaxation and supercoiling activities of gyrase containing labeled GyrA_T140C were unchanged to wild-type gyrase (Fig. S9). Upon addition of GyrB, the FRET histograms remained unchanged, with a single peak at EFRET = 0.95 (Fig. 4B), indicating that the formation of the gyrase heterotetramer does not induce any detectable conformational changes in GyrA around the DNA-gate. In the presence of dsDNA, the FRET efficiency also remained at 0.95 (Fig. 4C), and no additional species appeared when ATP or ADPNP were added (Fig. 4 D and E). Hence, these experiments do not show any indication for a gyrase conformation with an open DNA-gate. Strikingly, the FRET efficiency also remained at 0.95 when the pt-DNA was added to gyrase containing GyrA_T140C (Fig. 4F). The pt-DNA elicits a higher fraction of cleavage and a second, low FRET conformation when the DNA conformation is probed. Thus, we can conclude that mechanistically different events are observed with labeled gDNA and with labeled gyrase.
Fig. 4.
The DNA-gate of gyrase is predominantly closed when gDNA is bound. (A) Positions used for fluorophore attachment. Side-chains are highlighted by space-filling in the structure of E. coli GyrA (PDB-ID 1ab4). The catalytic tyrosines are depicted as stick model in black. (B–F) smFRET histograms of A488/TMR-labeled gyrase (GyrA_T140C). (B) gyrase. (C) gyrase + dsDNA. (D) gyrase + dsDNA, ATP. (E) gyrase + dsDNA, ADPNP. (F) gyrase + pt-DNA. The FRET efficiency of 0.95 is in agreement with a closed DNA-gate in all cases.
Using linear dsDNA as a model substrate for gyrase imposes limitations on the interpretation of the data. Most importantly, a 60-bp DNA as a gDNA cannot at the same time serve as a tDNA because of its small size. To elucidate the role of tDNA for the conformational cycle of the DNA-gate, negatively supercoiled pUC18 plasmid was used as a substrate for gyrase labeled at the DNA gate, and the DNA-gate conformation was monitored (Fig. 5A). The data unambiguously showed no significant fraction of a second species with increased donor/acceptor distance, consistent with a negligible population of gyrase with an open DNA-gate during DNA relaxation.
Fig. 5.
The DNA-gate is predominantly closed during relaxation and supercoiling. smFRET histograms of A488/TMR-labeled gyrase (GyrA_T140C) during DNA relaxation and supercoiling. (A) gyrase + pUC18−sc, relaxation reaction. (B) gyrase + pUC18rel. (C) gyrase + pUC18rel, ATP, supercoiling reaction. No significant population of gyrase with an open DNA-gate is detected. (D) Supercoiling of pUC18rel under smFRET conditions: 1, pUC18−sc; 2, pUC18rel, beginning of supercoiling reaction; 3, pUC18rel, 20′ after addition of ATP to start supercoiling; 4, pUC18rel, 20′ after addition of ADPNP; 5, independent repetition of 3; 6, pUCrel, after incubation with 50 pM unlabeled gyrase (GyrA_T140C, corresponding to the unlabeled fraction in smFRET experiments), and ATP. Relaxed plasmid is supercoiled by labeled gyrase during the smFRET experiment. The direct appearance of maximally supercoiled plasmid indicates a high processivity of B. subtilis gyrase. (E) Model for conformational changes in the bound DNA and gyrase during the catalytic cycle. The tDNA is omitted for clarity. Gyrase with an open DNA-gate is a transiently populated short-lived intermediate.
Finally, we monitored the DNA-gate conformation during DNA supercoiling using relaxed pUC18 plasmid as a substrate (Fig. 5 B and C). Gyrase activity under smFRET conditions was confirmed by the appearance of supercoiled DNA after 20 min (Fig. 5D). This activity can be attributed to the donor/acceptor-labeled gyrase, because neither dye affects the activity in control experiments (Fig. S9). Addition of relaxed plasmid and GyrB in the absence and presence of ATP does not affect the FRET histograms of donor/acceptor-labeled GyrA_T140C significantly, indicating that the DNA-gate remains in the closed conformation.
Discussion
DNA bound to the DNA-gate of gyrase can adopt two different conformations. The appearance of the distorted DNA conformation strictly requires DNA cleavage, but not ATP binding or the presence of a tDNA. Its population increases when DNA cleavage is favored, but exceeds the fraction of DNA cleaved, as observed with pt-DNA. Thus, DNA cleavage is a required for distortion, but does not appear to be a quantitative measure for DNA distortion. For topoisomerase II, two similar states of bound DNA have been observed. They were assigned to open and closed conformations of the DNA-gate based on the requirement for DNA cleavage and for ATP hydrolysis and on the extrapolated change in distances consistent with the diameter of a dsDNA (17). The distances involved were indirectly estimated by combining steady state FRET efficiencies and equilibrium populations as derived from kinetic traces. Here, we directly quantified the interdye distances in the two DNA conformers bound to the DNA-gate of gyrase and determined a distance change of 1.4 nm. A 1.4-nm gap in the gDNA would not be sufficient for tDNA passage. When labels are placed on gyrase to directly monitor the DNA-gate conformation, no corresponding conformational change of the protein is detected, excluding that the low FRET species reflects gyrase with an open DNA-gate.
Similar to the observations in the presence of excess dsDNA, a single gyrase conformer with a closed DNA-gate was detected during DNA relaxation and during ATP-dependent supercoiling. Our results are in agreement with several lines of evidence that the DNA-gate of gyrase (and of topoisomerase II) should be predominantly in the closed state during the catalysis of DNA supercoiling. Trapping experiments have consistently revealed a very low fraction of DNA-cleavage complexes (19). In addition, efforts to thread a linear DNA through topoisomerase II with a bound gDNA when the N-gate was closed by binding of ADPNP were not successful, consistent with a closed DNA-gate. Only in the presence of inhibitors such as quinolones, cleavage-complexes were highly populated (31, 32). Originally, it had been inferred that these inhibitors favor the open state of the DNA-gate. As we show here, quinolones such as OXO and CFX in fact do not favor opening of the DNA-gate, but prevent large distortions of gyrase-bound DNA that accompany cleavage. These results are in agreement with biochemical data that demonstrated quinolone binding to gyrase/DNA complexes in the absence of cleavage (33) and to gyrase with the DNA-gate cross-linked in the closed conformation (4).
The large conformational change during opening of the gyrase DNA-gate would not have gone unnoticed in FRET histograms. The absence of a protein conformational change in parallel with the conformational change of the DNA demonstrates that the two different DNA conformations do not represent open and closed DNA-gates. Instead, our data show that DNA bound to gyrase is strongly distorted. This DNA distortion may reflect an “unlocking” event preceding gate opening that has been suggested to involve DNA bending, twisting, or stretching (15, 16, 21, 34). The requirement of rearrangements in gyrase/DNA complexes has been inferred from structures of type II topoisomerases in cleavage-incompetent conformations (7, 35–37), where the distance of the two catalytic tyrosines largely exceeds the distance of the scissile bonds in the DNA. Consequently, either the bound DNA or the protein has to undergo a conformational change for DNA cleavage and gate-opening to occur. In the recently determined structure of DNA topoisomerase II in complex with DNA nicked at the cleavage sites (38), the DNA is indeed sharply bent, and the bases around the cleavage site adopt A-DNA conformation. Such a bending has been suggested as a prerequisite for cleavage (15, 38). In-line with structural evidence for conformational adjustments of the DNA, biochemical data show that the cleavage specificity of topoisomerase II is determined in a separate step after DNA binding (39). Possibly, the binding energy from cognate interactions is used for a sequence-specific conformational change toward a cleavage-competent state (39). Kinetic analyses of gDNA dissociation from topoisomerase II (40) also provided evidence for two noncovalent topoisomerase II/DNA complexes, a cleavage-competent state and a cleavage-incompetent state. The interpretation of the two DNA conformations we observed with gyrase as DNA bound to a cleavage-incompetent (high FRET) and a cleavage-competent (low FRET) state would rationalize the parallel increase of cleavage and the distorted DNA conformer. It would also be consistent with the higher DNA affinity of the cleavage-deficient GyrA_Y123F mutant compared with wild-type gyrase (Fig. S5) that has also been reported for the corresponding topoisomerase II mutant (41). The mutants do not convert binding energy into a distortion of the DNA. Altogether, our findings support the notion that DNA-gate opening is a two-step process with a chemical step (cleavage), associated with a severe distortion of the DNA (unlocking) (15, 16, 21, 34), and a subsequent mechanical step (gate-opening). DNA distortion, cleavage, and DNA-gate opening are thus mechanistically distinct events. Only cleavage and unlocking are observed with linear dsDNA, in agreement with the tDNA being required for gate-opening.
Interestingly, the distortion of the DNA bound to gyrase is already observed in the absence of nucleotides. ATP binding does not change the distribution between the two conformers. For topoisomerase II, different DNA-complexes were interpreted as complexes with open and closed N-gate conformation (40), but the lacking response to ATP contradicts a similar interpretation for gyrase. Similarly, it contradicts the assignment to open and closed forms of the DNA-gate (17) because the coupling of strand passage to hydrolysis would predict a response of this equilibrium to ATP. In contrast to gyrase, the two DNA conformations bound to topoisomerase II were only observed in the presence of ATP/Mg2+ (17, 18), and the fraction of the distorted DNA species is much higher for topoisomerase II (50%) than for gyrase (15%). Although it is not clear whether the two FRET species observed with topoisomerase II reflect similar DNA conformations, a different behavior of these two enzymes may represent a genuine feature that distinguishes ATP-driven relaxation by topoisomerase II from ATP-dependent supercoiling by gyrase.
Topoisomerase II and gyrase also appear to respond differently to ADPNP. For topoisomerase II, a reduced FRET efficiency of the gDNA was observed in the presence of ADPNP, which was interpreted as either a mixture of open and closed forms or as a single conformation with a “partially open” gate (17). With gyrase, we can unambiguously distinguish and quantify the two DNA conformations and find that ADPNP disfavors the low FRET conformation of gDNA. ADPNP binding to the GyrB subunits leads to a closure of the N-gate. ADPNP binding counteracts the DNA distortion associated with DNA cleavage, at least if the DNA-gate remains closed in the absence of a tDNA. Most likely, this effect is a direct consequence of the closure of the N-gate. An effect of ADPNP binding on the conformation of DNA bound to gyrase is in agreement with previous evidence from probing experiments (3, 42, 43). Although the 60-bp gDNA substrate is too short for wrapping to occur, the distortion of bound gDNA may be linked to DNA-wrapping by gyrase during the supercoiling reaction. The observed loss of DNA wrapping by ADPNP binding to gyrase (44) may thus be a consequence of ADPNP disfavoring distortion of gDNA. The DNA-gate remains predominantly closed when ADPNP is present and even during the relaxation or supercoiling reactions. Consequently, N-gate closure that captures a tDNA is not sufficient for opening of the DNA-gate.
The current picture of strand passage by gyrase supports an active role of the tDNA in DNA-gate opening. The cavity between a closed N- and DNA-gate is too small to accommodate a tDNA (10, 20, 22), suggesting that a captured tDNA will exert strain onto the DNA-gate once the N-gate has closed and thus potentially aid in DNA-gate opening (12, 15, 16). In contrast to gate-opening, gDNA cleavage is possible in the absence of tDNA (40). The single molecule study on topoisomerase II was carried out in the absence of a tDNA and hence in the absence of strand passage. The assignment of open and closed states suggested ATP-dependent opening and closing of the DNA-gate that does not require tDNA (17). However, these studies can only link DNA conformational changes in the DNA to cleavage, but not to gate-opening. Here, we present experiments that cover conditions in the absence (labeled DNA) and in the presence of tDNA and strand passage (labeled gyrase). Gyrase distorts bound DNA severely without a requirement for ATP and tDNA, but with a strict dependence on DNA cleavage. We therefore propose that this DNA distortion represents what has been called unlocking of the DNA-gate (15, 21, 34) (Fig. 5E). It is likely that local conformational changes of the DNA-gate accompany the distortion of the DNA, but do not lead to changes in FRET efficiencies. The large-scale movement that results in opening of the DNA-gate, however, is a mechanistically distinct, subsequent step. The DNA-gate is not opened by gDNA binding, presumably because it strictly requires a tDNA. Even during the relaxation and supercoiling reactions, gyrase with an open DNA-gate is not significantly populated, consistent with gate-opening as a very rare event that only occurs briefly to allow for strand passage. A default of the DNA-gate in the closed state has been put forward as the driving force for the expulsion of tDNA (21, 36). In-line with its potential danger for genome integrity, opening of the DNA-gate thus appears to be under a tight two-layered control that ensures strict coupling of gate-opening to strand passage: Binding and distortion of a cognate gDNA unlatches the DNA-gate, such that tDNA can transiently open the DNA-gate for strand passage.
Materials and Methods
Protein Production and Purification.
The GyrA and GyrB subunits of B. subtilis gyrase wild-type and mutants were produced and purified as described in ref. 45.
DNA Substrates and DNA Cleavage Assays.
The donor/acceptor-labeled 60-bp DNA substrate was prepared by annealing the complementary strands A (5′-GACGGCCAGT GCCAAGCTAT GCATGATCAT ACGXCGACTC TAGAGGATCC CCGGGTACCG-3′, X denotes Alexa488 (A488) attached to an amino-linker at the C6 of T), and B (5′-CGGTACCCGG GGATCCTCTA GAGTCGACGT ATGATCATGC AXAGCTTGGC ACTGGCCGTC-3′, X denotes A546 or A555 attached to an amino-linker at the C6 of T). In pt-DNA, a 3′-bridging pt-DNA was introduced in strand A at the cleavage site (italicized). All substrates contain a preferred cleavage site for B. subtilis gyrase (23) in the center (bold) (Fig. 1D). For details on DNA cleavage assays, see SI Text.
Fluorescence Measurements and Fluorescence Labeling of Gyrase.
Dissociation constants of the gyrase/DNA complex were determined in fluorescence equilibrium titrations of A488/A546-labeled dsDNA with GyrA in the presence of 8 μM GyrB at 37 °C in 50 mM Tris/HCl, pH 7.5, 100 mM KCl, 10 mM MgCl2 using the fluorescence anisotropy of A546 as a probe.
Cysteine mutants of GyrA were labeled by incubating a 3-fold molar excess of A488- and a 5-fold excess of TMR- or A546-maleimide with 70 μM GyrA in 50 mM Tris/HCl, pH 7.5, 200 mM NaCl, 10 mM MgCl2, 1 mM TCEP for 1 h at 25 °C. Free dye was removed by size-exclusion chromatography using NAP-5 columns space (GE Healthcare).
Quantum Yields, Förster Distances, and smFRET Experiments.
Donor quantum yields and Förster distances were determined as described (46 and references therein) (Tables S1 and S2). The orientation factor κ2 was set to (Table S4).
SmFRET experiments were performed at 37 °C with 50 pM donor/acceptor-labeled DNA and 2 μM GyrA and 5 μM GyrB, or with 180 pM donor/acceptor-labeled GyrA, 8 μM GyrB, and 0.5 μM DNA or 20 nM pUC18 in the presence of 2 mM ATP or ADPNP in 50 mM Tris/HCl, pH 7.5, 100 mM KCl, 10 mM MgCl2 using a home-built confocal microscope, and FRET histograms were calculated as described in ref. 46. (see SI Text).
Supplementary Material
Acknowledgments.
We thank Alex Burgin (deCODE biostructures, WA) for generously providing the phosphoamidite for synthesis of the DNA substrates containing a 3′-bridging phosphorothiolate; Andreas Schmidt, Ines Hertel, and Diana Blank for expert technical assistance; and Thomas Göttler for preliminary smFRET experiments with labeled gyrase. This work was funded by the VolkswagenStiftung and the Swiss National Science Foundation. M.H. is the recipient of a fellowship from the Roche Research Foundation.
Footnotes
The authors declare no conflict of interest.
This article is a PNAS Direct Submission.
This article contains supporting information online at www.pnas.org/cgi/content/full/0902493106/DCSupplemental.
References
- 1.Champoux JJ. DNA topoisomerases: Structure, function, and mechanism. Annu Rev Biochem. 2001;70:369–413. doi: 10.1146/annurev.biochem.70.1.369. [DOI] [PubMed] [Google Scholar]
- 2.Gellert M, Mizuuchi K, O'Dea MH, Nash HA. DNA gyrase: An enzyme that introduces superhelical turns into DNA. Proc Natl Acad Sci USA. 1976;73:3872–3876. doi: 10.1073/pnas.73.11.3872. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Williams NL, Howells AJ, Maxwell A. Locking the ATP-operated clamp of DNA gyrase: Probing the mechanism of strand passage. J Mol Biol. 2001;306:969–984. doi: 10.1006/jmbi.2001.4468. [DOI] [PubMed] [Google Scholar]
- 4.Williams NL, Maxwell A. Locking the DNA gate of DNA gyrase: Investigating the effects on DNA cleavage and ATP hydrolysis. Biochemistry. 1999;38:14157–14164. doi: 10.1021/bi991478m. [DOI] [PubMed] [Google Scholar]
- 5.Williams NL, Maxwell A. Probing the two-gate mechanism of DNA gyrase using cysteine cross-linking. Biochemistry. 1999;38:13502–13511. doi: 10.1021/bi9912488. [DOI] [PubMed] [Google Scholar]
- 6.Liu LF, Wang JC. DNA-DNA gyrase complex: The wrapping of the DNA duplex outside the enzyme. Cell. 1978;15:979–984. doi: 10.1016/0092-8674(78)90281-7. [DOI] [PubMed] [Google Scholar]
- 7.Morais Cabral JH, et al. Crystal structure of the breakage-reunion domain of DNA gyrase. Nature. 1997;388:903–906. doi: 10.1038/42294. [DOI] [PubMed] [Google Scholar]
- 8.Horowitz DS, Wang JC. Mapping the active site tyrosine of Escherichia coli DNA gyrase. J Biol Chem. 1987;262:5339–5344. [PubMed] [Google Scholar]
- 9.Brino L, et al. Dimerization of Escherichia coli DNA-gyrase B provides a structural mechanism for activating the ATPase catalytic center. J Biol Chem. 2000;275:9468–9475. doi: 10.1074/jbc.275.13.9468. [DOI] [PubMed] [Google Scholar]
- 10.Wigley DB, Davies GJ, Dodson EJ, Maxwell A, Dodson G. Crystal structure of an N-terminal fragment of the DNA gyrase B protein. Nature. 1991;351:624–629. doi: 10.1038/351624a0. [DOI] [PubMed] [Google Scholar]
- 11.Ali JA, Orphanides G, Maxwell A. Nucleotide binding to the 43-kilodalton N-terminal fragment of the DNA gyrase B protein. Biochemistry. 1995;34:9801–9808. doi: 10.1021/bi00030a018. [DOI] [PubMed] [Google Scholar]
- 12.Roca J, Wang JC. The capture of a DNA double helix by an ATP-dependent protein clamp: A key step in DNA transport by type II DNA topoisomerases. Cell. 1992;71:833–840. doi: 10.1016/0092-8674(92)90558-t. [DOI] [PubMed] [Google Scholar]
- 13.Roca J, Wang JC. DNA transport by a type II DNA topoisomerase: Evidence in favor of a two-gate mechanism. Cell. 1994;77:609–616. doi: 10.1016/0092-8674(94)90222-4. [DOI] [PubMed] [Google Scholar]
- 14.Roca J, Berger JM, Harrison SC, Wang JC. DNA transport by a type II topoisomerase: Direct evidence for a two-gate mechanism. Proc Natl Acad Sci USA. 1996;93:4057–4062. doi: 10.1073/pnas.93.9.4057. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Roca J. Topoisomerase II: A fitted mechanism for the chromatin landscape. Nucleic Acids Res. 2009;37:721–730. doi: 10.1093/nar/gkn994. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Roca J. The path of the DNA along the dimer interface of topoisomerase II. J Biol Chem. 2004;279:25783–25788. doi: 10.1074/jbc.M402555200. [DOI] [PubMed] [Google Scholar]
- 17.Smiley RD, Collins TR, Hammes GG, Hsieh TS. Single-molecule measurements of the opening and closing of the DNA gate by eukaryotic topoisomerase II. Proc Natl Acad Sci USA. 2007;104:4840–4845. doi: 10.1073/pnas.0700342104. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Collins TR, Hammes GG, Hsieh TS. Analysis of the eukaryotic topoisomerase II DNA gate: A single-molecule FRET and structural perspective. Nucleic Acids Res. 2009;37:712–720. doi: 10.1093/nar/gkn1059. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Wilstermann AM, Osheroff N. Stabilization of eukaryotic topoisomerase II-DNA cleavage complexes. Curr Top Med Chem. 2003;3:321–338. doi: 10.2174/1568026033452519. [DOI] [PubMed] [Google Scholar]
- 20.Tingey AP, Maxwell A. Probing the role of the ATP-operated clamp in the strand-passage reaction of DNA gyrase. Nucleic Acids Res. 1996;24:4868–4873. doi: 10.1093/nar/24.24.4868. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Wang JC. Moving one DNA double helix through another by a type II DNA topoisomerase: The story of a simple molecular machine. Q Rev Biophys. 1998;31:107–144. doi: 10.1017/s0033583598003424. [DOI] [PubMed] [Google Scholar]
- 22.Schoeffler AJ, Berger JM. Recent advances in understanding structure-function relationships in the type II topoisomerase mechanism. Biochem Soc Trans. 2005;33:1465–1470. doi: 10.1042/BST0331465. [DOI] [PubMed] [Google Scholar]
- 23.Bashkirov VI, Zvingila DJ. Sequence specificity of Bacillus subtilis DNA gyrase in vivo. Genetica. 1991;85:3–12. doi: 10.1007/BF00056101. [DOI] [PubMed] [Google Scholar]
- 24.Gellert M, Mizuuchi K, O'Dea MH, Itoh T, Tomizawa JI. Nalidixic acid resistance: A second genetic character involved in DNA gyrase activity. Proc Natl Acad Sci USA. 1977;74:4772–4776. doi: 10.1073/pnas.74.11.4772. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Hallett P, Maxwell A. Novel quinolone resistance mutations of the Escherichia coli DNA gyrase A protein: Enzymatic analysis of the mutant proteins. Antimicrob Agents Chemother. 1991;35:335–340. doi: 10.1128/aac.35.2.335. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Tamura JK, Bates AD, Gellert M. Slow interaction of 5′-adenylyl-beta,gamma-imidodiphosphate with Escherichia coli DNA gyrase. Evidence for cooperativity in nucleotide binding. J Biol Chem. 1992;267:9214–9222. [PubMed] [Google Scholar]
- 27.Jackson AP, Maxwell A. Identifying the catalytic residue of the ATPase reaction of DNA gyrase. Proc Natl Acad Sci USA. 1993;90:11232–11236. doi: 10.1073/pnas.90.23.11232. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Shen LL, Baranowski J, Pernet AG. Mechanism of inhibition of DNA gyrase by quinolone antibacterials: Specificity and cooperativity of drug binding to DNA. Biochemistry. 1989;28:3879–3885. doi: 10.1021/bi00435a038. [DOI] [PubMed] [Google Scholar]
- 29.Shen LL, Pernet AG. Mechanism of inhibition of DNA gyrase by analogues of nalidixic acid: The target of the drugs is DNA. Proc Natl Acad Sci USA. 1985;82:307–311. doi: 10.1073/pnas.82.2.307. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Deweese JE, Burgin AB, Osheroff N. Using 3′-bridging phosphorothiolates to isolate the forward DNA cleavage reaction of human topoisomerase IIalpha. Biochemistry. 2008;47:4129–4140. doi: 10.1021/bi702194x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Liu LF. DNA topoisomerase poisons as antitumor drugs. Annu Rev Biochem. 1989;58:351–375. doi: 10.1146/annurev.bi.58.070189.002031. [DOI] [PubMed] [Google Scholar]
- 32.Maxwell A. DNA gyrase as a drug target. Trends Microbiol. 1997;5:102–109. doi: 10.1016/S0966-842X(96)10085-8. [DOI] [PubMed] [Google Scholar]
- 33.Critchlow SE, Maxwell A. DNA cleavage is not required for the binding of quinolone drugs to the DNA gyrase-DNA complex. Biochemistry. 1996;35:7387–7393. doi: 10.1021/bi9603175. [DOI] [PubMed] [Google Scholar]
- 34.Wang JC. Unlocking and opening a DNA gate. Proc Natl Acad Sci USA. 2007;104:4773–4774. doi: 10.1073/pnas.0701070104. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Fass D, Bogden CE, Berger JM. Quaternary changes in topoisomerase II may direct orthogonal movement of two DNA strands. Nat Struct Biol. 1999;6:322–326. doi: 10.1038/7556. [DOI] [PubMed] [Google Scholar]
- 36.Berger JM, Gamblin SJ, Harrison SC, Wang JC. Structure and mechanism of DNA topoisomerase II. Nature. 1996;379:225–232. doi: 10.1038/379225a0. [DOI] [PubMed] [Google Scholar]
- 37.Laponogov I, et al. Breakage-reunion domain of Streptococcus pneumoniae topoisomerase IV: Crystal structure of a Gram-positive quinolone target. PLoS ONE. 2007;2:e301. doi: 10.1371/journal.pone.0000301. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Dong KC, Berger JM. Structural basis for gate-DNA recognition and bending by type IIA topoisomerases. Nature. 2007;450:1201–1205. doi: 10.1038/nature06396. [DOI] [PubMed] [Google Scholar]
- 39.Mueller-Planitz F, Herschlag D. DNA topoisomerase II selects DNA cleavage sites based on reactivity rather than binding affinity. Nucleic Acids Res. 2007;35:3764–3773. doi: 10.1093/nar/gkm335. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Mueller-Planitz F, Herschlag D. Coupling between ATP binding and DNA cleavage by DNA topoisomerase II: A unifying kinetic and structural mechanism. J Biol Chem. 2008;283:17463–17476. doi: 10.1074/jbc.M710014200. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Mueller-Planitz F, Herschlag D. Interdomain communication in DNA topoisomerase II: DNA binding and enzyme activation. J Biol Chem. 2006;281:23395–23404. doi: 10.1074/jbc.M604119200. [DOI] [PubMed] [Google Scholar]
- 42.Kampranis SC, Bates AD, Maxwell A. A model for the mechanism of strand passage by DNA gyrase. Proc Natl Acad Sci USA. 1999;96:8414–8419. doi: 10.1073/pnas.96.15.8414. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.Orphanides G, Maxwell A. Evidence for a conformational change in the DNA gyrase-DNA complex from hydroxyl radical footprinting. Nucleic Acids Res. 1994;22:1567–1575. doi: 10.1093/nar/22.9.1567. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44.Heddle JG, Mitelheiser S, Maxwell A, Thomson NH. Nucleotide binding to DNA gyrase causes loss of DNA wrap. J Mol Biol. 2004;337:597–610. doi: 10.1016/j.jmb.2004.01.049. [DOI] [PubMed] [Google Scholar]
- 45.Gottler T, Klostermeier D. Dissection of the nucleotide cycle of B. subtilis DNA gyrase and its modulation by DNA. J Mol Biol. 2007;367:1392–1404. doi: 10.1016/j.jmb.2007.01.055. [DOI] [PubMed] [Google Scholar]
- 46.Theissen B, Karow AR, Kohler J, Gubaev A, Klostermeier D. Cooperative binding of ATP and RNA induces a closed conformation in a DEAD box RNA helicase. Proc Natl Acad Sci USA. 2008;105:548–553. doi: 10.1073/pnas.0705488105. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.