Human Topoisomerase IIα Rapidly Relaxes Positively Supercoiled DNA

Movement of the DNA replication machinery through the double helix induces acute positive supercoiling ahead of the fork and precatenanes behind it. Because topoisomerase I and II create transient single- and double-stranded DNA breaks, respectively, it has been assumed that type I enzymes relax the positive supercoils that precede the replication fork. Conversely, type II enzymes primarily resolve the precatenanes and untangle catenated daughter chromosomes. However, studies on yeast and bacteria suggest that type II topoisomerases may also function ahead of the replication machinery. If this is the case, then positive DNA supercoils should be the preferred relaxation substrate for topoisomerase IIα, the enzyme isoform involved in replicative processes in humans. Results indicate that human topoisomerase IIα relaxes positively supercoiled plasmids >10-fold faster than negatively supercoiled molecules. In contrast, topoisomerase IIβ, which is not required for DNA replication, displays no such preference. In addition to its high rates of relaxation, topoisomerase IIα maintains lower levels of DNA cleavage complexes with positively supercoiled molecules. These properties suggest that human topoisomerase IIα has the potential to alleviate torsional stress ahead of replication forks in an efficient and safe manner.

One of the most striking features of DNA is the intertwining of the two complementary strands of the double helix (1). Discovery of this characteristic led to the immediate recognition that biological processes such as replication would be severely affected by the topological state of the genetic material (2). DNA in all species ranging from bacteria to humans is globally underwound (i.e. negatively supercoiled) (3)(4)(5)(6). This underwinding makes it easier to separate complementary DNA strands from one another and greatly facilitates initiation of replication and the assembly of replication forks. Once the fork begins to travel along the DNA template, however, the deleterious effects of topology manifest themselves (Fig. 1). Because helicases separate, but do not unwind the two strands of the double helix, fork movement results in acute overwinding (i.e. positive supercoiling) of the DNA ahead of the replication machinery ( Fig. 1B) (3,(5)(6)(7). This overwinding has two major consequences. First, it increases the difficulty of separating duplex DNA into individual strands. There-fore, accumulation of positive supercoils presents a formidable block to fork movement (5,(7)(8)(9)(10)(11). Second, DNA overwinding ahead of the fork leads to a compensatory underwinding behind the replication machinery. If the replisome rotates around the helical axis of the DNA, this underwinding allows some of the torsional stress in the prereplicated DNA to be translated to the newly replicated daughter molecules in the form of precatenanes ( Fig. 1C) (6,7,11). If these precatenanes are not resolved, they ultimately lead to the formation of catenated duplex daughter chromosomes.
The topological state of DNA in the cell is modulated by enzymes known as topoisomerases (5,7,(12)(13)(14)(15)(16). There are two classes of topoisomerases that are distinguished by their catalytic mechanisms. Type I enzymes act by generating a transient single-stranded break in the double helix, followed by either a single-stranded DNA passage event or controlled rotation about the break (14,17,18). As a result, these enzymes are able to alleviate torsional stress (i.e. remove superhelical twists) in duplex DNA.
Type II topoisomerases act by generating a transient double-stranded DNA break, followed by a double-stranded DNA passage event (14 -16, 19). Consequently, enzymes in this latter class are able to remove superhelical twists from DNA and resolve knotted or tangled duplex molecules. Whereas lower eukaryotes such as yeast and Drosophila encode only a single type II topoisomerase (20,21), bacteria and vertebrate species express two discrete forms of the enzyme (14,(22)(23)(24). Bacteria express DNA gyrase and topoisomerase IV. Although these two enzymes are closely related, they play distinct physiological roles. Gyrase is the only known topoisomerase that can actively underwind DNA (14,23,24). It functions to maintain the superhelical density of the bacterial chromosome and removes positive supercoils that accumulate ahead of DNA tracking systems. In contrast, topoisomerase IV is a potent decatenase that resolves knots in the genome and untangles newly replicated daughter chromosomes (14,25). As discussed below, this type II enzyme also may play a role ahead of replication forks.
Vertebrates express two closely related isoforms of topoisomerase II, ␣ and ␤ (22,26). These enzymes display a high degree of amino acid sequence identity and similar enzymological characteristics. However, they differ in their protomer molecular masses (170 versus 180 kDa, respectively) and are encoded by separate genes (14,24,27). Either topoisomerase II␣ or ␤ can complement the loss of topoisomerase II in yeast (28 -30), but the two enzymes have distinct patterns of expression and physiological functions in vertebrate cells. The ␣ isoform is up-regulated dramatically during cell proliferation and is tightly associated with mitotic chromosomes (13,(31)(32)(33)(34). In contrast, expression of the ␤ isoform is independent of proliferative status, and the enzyme dissociates from chromosomes during mitosis (13,27,31,35). Thus, topoisomerase II␣ is believed to be the isoform that functions in growth-dependent processes, such as DNA replication and chromosome segregation (7,13). Although cells can survive the absence of topoisomerase II␤, this isoform cannot compensate for the loss of topoisomerase II␣ (27,36,37).
Beyond their critical physiological functions, prokaryotic and eukaryotic topoisomerases are the targets for some of the most successful antibacterial (23, 38 -41) and anticancer (15)(16)(17)(42)(43)(44) drugs currently in clinical use. The common feature of all of these drugs is that they act by increasing levels of covalent topoisomerase-cleaved DNA complexes (i.e. cleavage complexes) that are requisite catalytic intermediates (15-17, 23, 40, 45). When DNA tracking systems, such as replication forks, collide with these complexes, transient enzyme-associated DNA breaks are converted to permanent breaks in the genome (15-17, 23, 40, 45). As a result, antibacterial and anticancer agents convert topoisomerases from essential enzymes to potent cellular toxins.
Topoisomerases play critical roles during DNA replication in eukaryotic cells (5,7,(12)(13)(14)(15)(16). Because type I topoisomerases can relax positive or negative superhelical twists (14,17,18,46), it has been assumed that these enzymes act ahead of replication forks to alleviate torsional stress. Conversely, because type II topoisomerases can untangle duplex DNA molecules (12, 14 -16, 19), these enzymes are believed to act exclusively behind forks to resolve precatenanes or later in replication to unlink catenated daughter chromosomes. In marked contrast to this proposed segregation of function, models for anticancer drug action place topoisomerase II ahead of approaching replication forks (15,16,45,(47)(48)(49). This discrepancy raises the question of whether eukaryotic type II topoisomerases have normal physiological functions ahead of the DNA replication machinery.
Two lines of evidence suggest that this may be the case. First, yeast topoisomerase II can compensate for the loss of topoisomerase I in Saccharomyces cerevisiae, but loss of both enzymes abruptly halts DNA synthesis (8,9). This finding indicates that the type II enzyme can assume the role of topoisomerase I ahead of the replication machinery. Second, Escherichia coli topoisomerase IV (whose functions in bacteria appear to parallel those of topoisomerase II in eukaryotes) can partially compensate for the loss of DNA gyrase during replication elongation (50,51). Moreover, topoisomerase IV relaxes positive DNA supercoils ϳ20-fold faster than it does negative supercoils (52,53).
If type II topoisomerases play a role ahead of replication forks in vertebrate species, then positively supercoiled DNA should be the preferred relaxation substrate for topoisomerase II␣, the isoform that is involved in replicative processes. Results indicate that human topoisomerase II␣ relaxes positively supercoiled plasmids Ͼ10-fold faster than negatively supercoiled molecules. In contrast, human topoisomerase II␤, which is not required for DNA replication, shows no such preferential relaxation. Finally, topoisomerase II␣ maintains lower levels of DNA cleavage complexes with positively supercoiled molecules. These enzymological properties are consistent with those of a topoisomerase that functions ahead of the replication machinery.

EXPERIMENTAL PROCEDURES
Enzymes and Plasmid DNA-Human topoisomerase II␣ and topoisomerase II␤ were expressed in S. cerevisiae (54) and purified as described previously (55). Based on Coomassie-or silver-stained protein gels, both enzymes were at least 95% homogeneous. Archaeoglobus fulgidus reverse gyrase was expressed in E. coli C41(DE3) and purified according to Rodriguez (56). Negatively supercoiled pBR322 DNA was prepared from E. coli using a Plasmid Mega Kit (Qiagen) as described by the manufacturer.
Preparation of Positively Supercoiled Plasmid DNA-Positively supercoiled plasmid DNA was prepared by treating negatively supercoiled molecules with reverse gyrase (56). Reaction mixtures contained 35 nM negatively supercoiled pBR322 DNA and 420 nM reverse gyrase in a total of 500 l of 50 mM Tris-HCl (pH 8.0), 10 mM NaCl, 10 mM MgCl 2 , and 1 mM ATP. Reactions were incubated at 95°C for 5 min, halted by the addition of 13 l of 375 mM EDTA, and cooled on ice. Proteinase K was added (10 l of 4 mg/ml), and the reactions were incubated at 45°C for 30 min to digest the enzyme. Samples were extracted with phenol: chloroform:isoamyl alcohol (25:24:1), and positively supercoiled DNA was precipitated with cold ethanol. Plasmids were resuspended in 100 l of 5 mM Tris-HCl (pH 7.4) containing 500 M EDTA. To make certain that differences between negatively and positively supercoiled substrates were not influenced by the temperature or other conditions used in the preparation protocol, negatively supercoiled plasmids were treated in a parallel fashion, except that reverse gyrase was omitted from reaction mixtures.
The average number of superhelical twists present in DNA substrates and the resulting values were determined by electrophoretic band counting relative to fully relaxed molecules. For negatively supercoiled substrates, time courses for the relaxation of pBR322 by topoisomerase I were resolved by electrophoresis in 1% agarose gels in TBE (100 mM Tris borate (pH 8.3), 2 mM EDTA) containing 1-2 g/ml chloroquine (Sigma) as running buffer. Optimal resolution was observed in 1 g/ml chloroquine. For positively supercoiled substrates, time courses for the generation of positive superhelical twists by reverse gyrase were resolved by electrophoresis as above in TBE containing 5-15 g/ml netropsin B (Roche Applied Science). Optimal resolution was observed in 7.5 g/ml netropsin B. Positively supercoiled bands also were FIGURE 1. Model for topoisomerase function and topological stress associated with DNA replication. This figure was adapted from Wang (7). The replication machinery is represented by a rod moving through the double helix. DNA ends are anchored to hypothetical immobile structures existing in the nucleus. A, upon initiation of DNA replication, the two strands of duplex DNA are separated, and the replication fork is formed. B, movement of the replication machinery through the immobilized DNA template strands induces acute overwinding (i.e. positive supercoiling) ahead of the fork. C, if the replisome rotates around the helical axis of the DNA, compensatory underwinding (i.e. negative supercoiling) behind the replication machinery allows some of the torsional stress in the prereplicated DNA to be translated to the newly replicated daughter molecules in the form of precatenanes. If these precatenanes are not resolved, they ultimately lead to the formation of catenated duplex daughter chromosomes. Topoisomerase I is proposed to work ahead of the replication fork to remove positive DNA supercoils, whereas topoisomerase II is proposed to work primarily behind the fork to remove precatenanes. In contrast to replication models, drug models place topoisomerase II ahead of DNA tracking systems.
counted in time courses of relaxation of positively supercoiled plasmid by topoisomerase I. Calculated values were consistent with those obtained from the reverse gyrase time courses.
DNA Relaxation-DNA relaxation assays were based on the procedure of Fortune and Osheroff (57). Unless stated otherwise, reaction mixtures contained 1 nM human topoisomerase II␣ or topoisomerase II␤, 5 nM negatively or positively supercoiled pBR322 DNA, and 1 mM ATP in a total of 20 l of relaxation buffer (10 mM Tris-HCl (pH 7.9), 0.1 mM EDTA, 175 mM KCl, 5 mM MgCl 2 , and 2.5% glycerol). Samples were incubated at 37°C, and DNA relaxation was stopped by the addition of 3 l of 0.77% SDS, 77 mM EDTA (pH 8.0). Samples were mixed with 2 l of agarose gel loading buffer (60% sucrose in 10 mM Tris-HCl (pH 7.9)), heated at 45°C for 5 min, and subjected to electrophoresis in a 1% agarose gel in 100 mM Tris borate (pH 8.3), 2 mM EDTA.
Alternatively, samples were analyzed by two-dimensional gel electrophoresis. The first dimension was run for 2 h as described in the preceding paragraph. The gel was soaked for 2 h with gentle shaking in 200 ml of TBE containing 4.5 g/ml chloroquine and run in the orthogonal dimension (90°clockwise) for 2 h in fresh TBE containing 4.5 g/ml chloroquine. All of the gels were stained with 1 g/ml ethidium bromide, and DNA bands were visualized with ultraviolet light and quantified using an Alpha Innotech digital imaging system.
DNA Binding-The ability of human topoisomerase II␣ to bind negatively and positively supercoiled DNA was assessed using a competitive nitrocellulose filter binding assay. Binding mixtures contained 400 nM topoisomerase II␣, 5 nM linear pBR322 DNA that was cleaved with Hind III and terminally labeled with [ 32 P]phosphate and 0 -20 nM negatively or positively supercoiled DNA in a total of 20 l of binding buffer (10 mM Tris-HCl (pH 7.9), 100 or 175 mM KCl, 0.1 mM EDTA, and 2.5% glycerol). Samples were incubated at 37°C for 6 min. Under the conditions of the assay, a DNA binding equilibrium was established in less than 1 min. Nitrocellulose membranes (0.45 m HA, Millipore) were prepared by incubation in binding buffer for 10 min. Samples were applied to the membranes and filtered in vacuo. Membranes were washed three times with 1 ml of binding buffer, dried, and submerged in 8 ml of scintillation fluid (Econo-Safe Research Products International). Radioactivity remaining on the membranes was quantified using a Beckman LS 5000 TD scintillation counter. The percentage of linear DNA bound to topoisomerase II␣ was determined based on the ratio of radioactivity on the membranes versus that of the input DNA.
It should be noted that DNA binding experiments were performed in the absence of ATP and Mg 2ϩ . This was done to prevent the formation of concatenated DNA multimers, which are too large to pass through the filter, or the generation of covalent enzyme-DNA cleavage complexes during the course of the assay.
ATP Hydrolysis-ATPase assays were performed as described by Osheroff et al. (58). Reaction mixtures contained 1 nM human topoisomerase II␣, 55 nM negatively or positively supercoiled pBR322 DNA, and 1 mM [␥-32 P]ATP in a total of 40 l of relaxation buffer. Mixtures were incubated at 37°C, and 2-l samples were removed at time intervals up to 12 min and spotted on polyethyleneimine-impregnated thin layer cellulose chromatography plates (EMD Chemicals). Plates were developed by chromatography in freshly made 400 mM NH 4 HCO 3 and analyzed using a Bio-Rad molecular imager FX. ATP hydrolysis was monitored by the release of free phosphate.
Plasmid DNA Cleavage-DNA cleavage reactions were carried out using the procedure of Fortune and Osheroff (57). Reaction mixtures contained 0 -800 nM topoisomerase II␣, 10 nM negatively or positively supercoiled pBR322 DNA, and 5 mM MgCl 2 or CaCl 2 in a total of 20 l of cleavage buffer (10 mM Tris-HCl (pH 7.9), 100 mM KCl, 0.1 mM EDTA, and 2.5% glycerol). Mixtures were incubated at 37°C for 6 min, and enzyme-DNA cleavage complexes were trapped by the addition of 2 l of 5% SDS followed by 1 l of 375 mM EDTA (pH 8.0). Proteinase K (2 l of 0.8 mg/ml) was added, and the samples were incubated at 45°C for 30 min to digest the type II enzyme. Samples were mixed with 2 l of agarose gel loading buffer, heated at 45°C for 5 min, and subjected to electrophoresis in 1% agarose gels in 40 mM Tris acetate (pH 8.3), 2 mM EDTA containing 0.5 g/ml ethidium bromide. DNA bands were visualized and quantified as described above, and cleavage was monitored by the conversion of supercoiled plasmid DNA to linear molecules.
DNA cleavage reactions were performed in the absence of ATP so that the topological state of the DNA did not change during the course of the reaction. It should be noted that the nucleotide cofactor does not influence the mechanism of topoisomerase II-mediated DNA scission (15).
Site-specific DNA Cleavage-DNA sites cleaved by human topoisomerase II␣ in negatively and positively supercoiled DNA were mapped using a modification of the procedure of O'Reilly and Kreuzer (59,60). DNA cleavage mixtures contained 2.2 M human topoisomerase II␣, 10 nM negatively or positively supercoiled pBR322 DNA, and 1 mM MgCl 2 or CaCl 2 in a total of 160 l of cleavage buffer. Samples were incubated at 37°C for 6 min, and enzyme-DNA cleavage complexes were trapped by the addition of 16 l of 1% SDS followed by 8 l of 375 mM EDTA (pH 8.0). Proteinase K (16 l of 0.8 mg/ml) was added, and the mixtures were incubated at 45°C for 30 min to digest the type II enzyme. DNA products were purified by passage through Qiaquick spin columns (Qiagen) as described by the manufacturer. DNA cleavage products were linearized by treatment with HindIII. Terminal 5Ј-phosphates were removed by treatment with calf intestinal alkaline phosphatase and replaced with [ 32 P]phosphate using T4 polynucleotide kinase and [␥-32 P]ATP. Samples were treated with EcoRI, and the singly endlabeled DNA products were purified by passage through a CHROMA SPINϩTE-100 column (Clontech). Loading buffer (40% formamide, 8.4 mM EDTA, 0.02% bromphenol blue, and 0.02% xylene cyanole FF) was added, and the samples were subjected to electrophoresis in 6% sequencing gels. Gels were fixed in 10% methanol, 10% acetic acid and

Generation of Positively Supercoiled DNA Substrates-Steady-state
interactions between eukaryotic topoisomerases and positively supercoiled DNA molecules have been examined previously (46,58,(61)(62)(63)(64). However, in most cases, positive superhelical twists were introduced in the DNA by the addition of intercalating agents such as ethidium bromide or chloroquine (58,(61)(62)(63). Because these chemicals locally underwind the double helix, they generate a compensatory global overwinding elsewhere in covalently closed circular DNA substrates. Although previous studies have provided useful information, they suffer from a common deficiency; it is not always possible to dissociate the effects of DNA topology on topoisomerase action from those of the intercalating agent. Thus, these studies have not been able to assess interactions between eukaryotic topoisomerases and positively supercoiled DNA in a quantitative manner.
To overcome this critical deficiency, we prepared positively supercoiled molecules by incubating pBR322 plasmid DNA with A. fulgidus reverse gyrase. This type I topoisomerase is the only enzyme unique to hyperthermophiles and the only known topoisomerase that can actively introduce positive superhelical twists into DNA (56,(65)(66)(67)(68).
A time course for the conversion of negatively to positively supercoiled pBR322 DNA is shown in Fig. 2. The initial plasmid substrate contained ϳ15-17 negative superhelical twists/molecule ( Ϸ Ϫ0.035 to Ϫ0.039) as determined by electrophoretic band counting in agarose gels containing chloroquine. This superhelical density is typical of plasmids isolated from E. coli. Reverse gyrase rapidly relaxed pBR322, and positive superhelical twists began to appear within the first minute of the reaction. As determined by electrophoretic band counting in gels containing netropsin B, positively supercoiled plasmids generated following a 5-min incubation with reverse gyrase contained ϳ15-17 positive superhelical twists/molecule ( Ϸ ϩ0.035 to ϩ0.039). The handedness of positively supercoiled DNA was confirmed by twodimensional gel electrophoresis (Fig. 3, left panel). Thus, the substrates employed for the experiments in this study contained equivalent numbers of superhelical twists but were of opposite handedness.
When resolved by one-dimensional gel electrophoresis, relaxation mixtures were run in the absence of intercalating dyes and subsequently stained with ethidium bromide. Despite the opposite handedness of the DNA substrates employed, similar resolution of negatively and positively supercoiled plasmids was observed. However, it is notable that positively supercoiled substrates bind less ethidium bromide than negatively supercoiled plasmids. Therefore, to ensure that equal amounts of initial substrates were used in all of the experiments, the DNA concentration was independently assessed by spectrophotometric analysis and  Human Topoisomerase II␣ Preferentially Relaxes Positively Supercoiled DNA-If type II topoisomerases play a role ahead of replication forks in vertebrates, then positively supercoiled DNA should be the preferred relaxation substrate for topoisomerase II␣, the isoform that is involved in replicative processes. Therefore, to characterize interactions between human topoisomerase II␣ and DNA substrates found ahead of replication forks, we assessed the ability of the enzyme to relax positively supercoiled molecules (Figs. 3 and 4).
Compared with relaxation of negatively supercoiled substrates, the enzyme removed positive superhelical twists at a much higher rate. In addition, topoisomerase II␣ relaxed positive supercoils in a more distributive fashion. Because of this difference, relaxation rates were quantified by one-dimensional gel analysis using two complementary methods: gain of fully relaxed product or loss of fully supercoiled substrate (Fig. 4). As determined by the former method, the rate of relaxation of positively supercoiled DNA by topoisomerase II␣ is Ͼ30-fold higher than the rate of relaxation of negatively supercoiled molecules. As determined by the latter method, it is nearly 10-fold higher.
Similar differences in relaxation rates were observed when reaction mixtures were resolved by two-dimensional gel electrophoresis (Fig. 3). The center panel shows the results of a 30-min relaxation assay with negatively supercoiled plasmid, and the right panel shows a 5-min assay with positively supercoiled substrate. It should be noted that as the DNA molecules approached their fully relaxed state, underwound plasmids remained slightly negatively supercoiled, and overwound plasmids remained slightly positively supercoiled. However, because the buffer conditions of the DNA relaxation assays and the gel electrophoresis are not identical, it is possible that the final distribution of reaction products differs somewhat.
The degree of DNA overwinding that occurs ahead of DNA replication forks in human cells is not known. However, single molecule experiments suggest that DNA ahead of tracking systems can reach a value as high as ϩ0.110 (69). DNA buckling (i.e. the transition to non-B form structures) was not observed in these experiments below a value of ϩ0.058 (69). Therefore, we believe that the positively supercoiled substrates used in our DNA relaxation assays ( Ϸ ϩ0.035 to ϩ0.039) should be free of unusual non-B structures. However, to make certain that this was the case, additional DNA relaxation experiments were carried out using positively supercoiled substrates with a lower superhelical density ( Ϸ ϩ0.022). Similar high rates of DNA relaxation were observed with this substrate (data not shown). Thus, preferential relaxation by human topoisomerase II␣ does not require extensive overwinding of the DNA.
The salt concentration utilized for DNA relaxation experiments, 175 mM KCl, represents the optimal condition for the relaxation of negatively supercoiled substrates. To determine whether the switch from a processive to a distributive reaction contributed to the preferential relaxation of positively supercoiled plasmids, additional assays were carried out at lower ionic strengths. The relaxation of positively supercoiled DNA switched to a fully processive reaction at 100 mM KCl (Fig.  5). Even under fully processive relaxation conditions, positively supercoiled DNA substrates were relaxed at a rate that was ϳ10-fold faster than that observed with negatively supercoiled plasmids. Therefore, the processivity of DNA relaxation by human topoisomerase II␣ does not appear to be the underlying basis for the preferential removal of positive superhelical twists.
Human Topoisomerase II␤ Does Not Preferentially Relax Positively Supercoiled DNA-Although topoisomerase II␣ and ␤ differ dramatically in their physiological regulation and functions, the two isoforms display only minimal differences in their enzymatic properties (15,16,22,27,70). Because topoisomerase II␤ is not believed to play a role in DNA replication, it was of interest to determine whether this isoform also displayed a preference for positively supercoiled molecules. As seen in Fig. 6, topoisomerase II␤ relaxed positive and negative DNA supercoils at similar rates (less than 2-fold difference). Once again, relaxation of positively supercoiled molecules appeared to be more distributive. The recognition of positively supercoiled DNA by topoisomerase II␣ represents the first major enzymological difference between the two isoforms. Furthermore, it is consistent with the proposal that topoisomerase II␣, and not topoisomerase II␤, functions in DNA replication.
Effects of the Geometry of DNA Supercoils on Substrate Binding by Human Topoisomerase II␣-Previous studies indicate that the ability of human topoisomerase II␣ to recognize DNA sequence or damage is controlled by chemical steps in the catalytic cycle of the enzyme rather than by alterations in substrate binding affinity (71)(72)(73). In contrast, the ability of eukaryotic type II topoisomerases to distinguish negatively supercoiled DNA from relaxed substrates appears to be governed primarily by binding interactions (61, 74 -76). These enzymes differentiate between supercoiled and relaxed DNA topoisomers by preferentially interacting with DNA cross-overs, which are more prevalent in supercoiled molecules (61,75). However, because both negatively and positively supercoiled molecules contain DNA nodes, albeit of opposite handedness, it is not known how topoisomerase II␣ discriminates between underwound and overwound DNA.
As a first step toward addressing this critical issue, the relative affinity of the human enzyme for negatively and positively supercoiled plasmids  NOVEMBER 25, 2005 • VOLUME 280 • NUMBER 47 was determined using a competitive DNA binding assay (Fig. 7). In this system, the ability of supercoiled plasmids to compete with radiolabeled linear pBR322 molecules for binding to topoisomerase II␣ was monitored on nitrocellulose filters. The enzyme was exposed to both DNA substrates simultaneously, and the 6-min time point used for the assay ensured that a binding equilibrium was established (data not shown). DNA binding was analyzed at the two salt conditions, 175 and 100 mM KCl, used for DNA relaxation assays.

Topoisomerase II␣ Rapidly Relaxes (؉) Supercoiled DNA
At either ionic strength, topoisomerase II␣ displayed a slightly higher affinity (ϳ2-fold) for negatively supercoiled plasmids as compared with positively supercoiled substrates (Fig. 7). This finding indicates that the rapid relaxation of positively supercoiled molecules by human topoi-somerase II␣ is not due to a higher binding affinity of the enzyme for its initial DNA substrate.
Effects of the Geometry of DNA Supercoils on ATP Hydrolysis by Human Topoisomerase II␣-DNA-bound topoisomerase II hydrolyzes more ATP than does free enzyme (58,(77)(78)(79)(80)(81). It has been proposed that the stimulation of ATP hydrolysis by DNA represents a nonproductive cycling of the enzyme in the "closed clamp" form and actually decreases the efficiency of ATP utilization (81). Therefore, if topoisomerase II␣ relaxes positively supercoiled substrates more efficiently than it does negatively supercoiled DNA, it might hydrolyze lower levels of ATP during the course of the reaction. As seen in Fig. 8, the rate of ATP hydrolysis in the presence of positively supercoiled plasmid was ϳ2-fold  lower than the rate observed in reactions that contained negatively supercoiled DNA.
Human Topoisomerase II␣ Maintains a Lower Level of DNA Cleavage Complexes with Positively Supercoiled Substrates-As a prerequisite to the strand passage event, topoisomerase II creates transient doublestranded breaks in its DNA substrate. To maintain genomic integrity during this process, the enzyme forms covalent bonds between active site tyrosyl residues and the 5Ј-terminal phosphates of the cleaved DNA (5,15,16,19,82). However, two negative outcomes are possible if a DNA tracking system, such as the replication machinery, collides with one of these cleavage complexes. First, the collision can disrupt the cleavage complex, making it impossible for the enzyme to ligate the cleaved strands (49,83). Second, the collision can arrest the progress of the replication fork, which triggers fork restart and recombination pathways (84). Both of these outcomes eventually lead to the formation of double-stranded breaks in the genome (15, 16, 47-49, 85, 86).
Because the DNA that precedes the replication fork is overwound, we characterized the ability of human topoisomerase II␣ to cleave positively supercoiled plasmid molecules (Fig. 9). Over a range of enzyme:plasmid ratios, levels of DNA scission with positively supercoiled substrates were 3-4-fold lower than observed with negatively supercoiled plasmids. Similar results were seen when the physiological divalent cation, Mg 2ϩ , was replaced with Ca 2ϩ (Fig. 9, inset). This latter divalent cation supports higher levels of DNA cleavage (60,87). Although DNA scission was reduced with overwound substrates, the site specificity of cleavage, as well as the relative site utilization, was identical with positively and negatively supercoiled DNA (Fig. 9).

DISCUSSION
The movement of the replication machinery on DNA results in the accumulation of positive superhelical twists ahead of the fork and precatenanes behind (3,(5)(6)(7)11). Although the torsional stress of DNA overwinding can be alleviated by the actions of an enzyme that generates single-stranded breaks in the double helix, the untangling of daughter chromosomes can only be accomplished by an enzyme that creates double-stranded breaks (5, 7, 12, 14 -16, 19). Thus, it has been assumed that topoisomerase I functions ahead of the replication fork, whereas topoisomerase II acts behind.
Previous in vivo studies on S. cerevisiae topoisomerase II (8,9) and E. coli topoisomerase IV (50, 51) suggest a role for type II enzymes ahead of the replication fork. This suggestion is supported by in vitro studies, which indicate that topoisomerase IV preferentially relaxes positively   NOVEMBER 25, 2005 • VOLUME 280 • NUMBER 47 supercoiled DNA (52,53). To determine whether mammalian type II topoisomerases also display catalytic properties consistent with functions ahead of a replication fork, we characterized the ability of human topoisomerase II␣ to relax positively supercoiled DNA.

Topoisomerase II␣ Rapidly Relaxes (؉) Supercoiled DNA
Topoisomerase II␣ displayed two important enzymological characteristics that would be beneficial to an enzyme that operates ahead of the replication machinery. First, the enzyme removed positive superhelical twists at a rate that was Ͼ10-fold faster than the rate for negative superhelical twists. Thus, topoisomerase II␣ displays preferential activity with the DNA substrates that accumulate ahead of the fork. Furthermore, the rapid rate of catalysis with overwound substrates makes it less likely that the replication machinery would collide with a molecule of topoisomerase II␣ acting on the prereplicated DNA. Second, the enzyme maintained lower levels of DNA cleavage complexes with positively supercoiled substrates. This decreases the probability that a collision with a replication fork would result in the formation of a topoisomerase II-associated double-stranded break in the genetic material. Taken together, these properties suggest that human topoisomerase II␣ has the potential to alleviate torsional stress ahead of replication forks in an efficient and safe manner.
In contrast to the ␣ isoform, topoisomerase II␤, which is not believed to function in DNA replication, relaxed positively and negatively supercoiled substrates at similar rates. Relatively few studies on the enzymatic properties of topoisomerase II␤ have been reported. However, in every fundamental mechanistic aspect that has been compared, topoisomerase II␣ and ␤ are virtually identical (22,27,70). Results of the present study indicate a major difference between these two isoforms that is consistent with their proposed physiological roles. It is not clear how topoisomerase II␣, but not ␤, is able to distinguish the handedness of DNA supercoils. However, based on structural studies on DNA gyrase and topoisomerase IV, coupled with nucleic acid modeling and binding experiments, it has been proposed that this ability resides in the C-terminal domain of bacterial type II enzymes (64, 88 -91).
The amino acid sequence of the C-terminal domain varies considerably from species to species. Although topoisomerase II␣ and ␤ possess a high degree of amino acid sequence identity in their catalytic cores (ϳ79% identity), the two enzymes diverge significantly in their C-terminal domains (ϳ31% identity) (92). The present observations make it tempting to speculate that the C-terminal domain of human topoisomerase II␣ plays a role in sensing the geometry of DNA substrates. Studies with deletion mutants of the human enzyme currently are underway to test this hypothesis.
Lower eukaryotic species, such as S. cerevisiae and Drosophila contain only one isoform of topoisomerase II (20,21). Single molecule experiments suggest that Drosophila topoisomerase II removes DNA supercoils of different handedness at similar rates (64). Steady-state experiments conducted as part of the present work (data not shown) support this conclusion. Drosophila topoisomerase II relaxed positively supercoiled plasmids at a rate that was marginally (Ͻ2-fold) higher than that obtained with negatively supercoiled molecules. Similar results were observed with S. cerevisiae topoisomerase II, which relaxed positively supercoiled plasmids ϳ2-3 times faster. It is not known why the ability to discern the handedness of DNA supercoils is not extended (or is only weakly extended) to eukaryotic organisms that contain only a single isoform of topoisomerase II. However, it may be related to the distinct separation of physiological functions of topoisomerase II␣ and ␤ in vertebrate species.
Human topoisomerase II␣ and E. coli topoisomerase IV (52, 53) both preferentially relax positively supercoiled DNA. Although neither enzyme displays an increased affinity for positive DNA supercoils, they exhibit different attributes with overwound molecules with regard to DNA cleavage. Whereas topoisomerase II␣ maintains lower levels of cleavage complexes with positively supercoiled substrates, topoisomerase IV displays higher amounts of DNA scission (52). Thus, there appears to be at least one fundamental difference between the modalities that these two enzymes use to distinguish the handedness of superhelical twists.
Numerous reports indicate that topoisomerase II␣ is an important physiological target for anticancer drugs (15-17, 42, 43, 45). If the enzyme spends a significant period of time ahead of replication forks, the relevant nucleic acid substrate would be positively supercoiled DNA. However, previous studies used either negatively supercoiled or linear substrates to characterize the effects of drugs on the DNA cleavage activity of topoisomerase II␣. Because the handedness of superhelical twists has a strong influence on the ability of the human enzyme to cleave DNA, it will be of interest to re-examine drug actions when positively supercoiled molecules are used as cleavage substrates.
In summary, human topoisomerase II␣ preferentially removes positive superhelical twists from DNA. This finding suggests that positively supercoiled DNA is the preferred physiological substrate for this enzyme and implies that topoisomerase II␣ plays a role in relieving torsional stress that accumulates in front of the replication machinery or other DNA tracking systems. The high rates of DNA relaxation observed with positively supercoiled substrates, coupled with the low levels of DNA cleavage, makes topoisomerase II␣ ideally suited to function ahead of replication forks.