Atp-bound topoisomerase ii as a target for antitumor drugs.

Topoisomerase II (TOP2) poisons interfere with the breakage/reunion reaction of TOP2 resulting in DNA cleavage. In the current studies, we show that two different classes (ATP-sensitive and -insensitive) of TOP2 poisons can be identified based on their differential sensitivity to the ATP-bound conformation of TOP2. First, in the presence of 1 mm ATP or the nonhydrolyzable analog adenosine 5'-(beta,gamma-imino)triphosphate, TOP2-mediated DNA cleavage induced by ATP-sensitive TOP2 poisons (e.g. doxorubicin, etoposide, mitoxantrone, and 4'-(9-acridinylamino)methanesulfon-m-anisidide) was 30-100-fold stimulated, whereas DNA cleavage induced by ATP-insensitive TOP2 poisons (e.g. amonafide, batracylin, and menadione) was only slightly (less than 3-fold) affected. In addition, ADP was shown to strongly antagonize TOP2-mediated DNA cleavage induced by ATP-sensitive but not ATP-insensitive TOP2 poisons. Second, C427A mutant human TOP2alpha, which exhibits reduced ATPase activity, was shown to exhibit cross-resistance to all ATP-sensitive but not ATP-insensitive TOP2 poisons. Third, using ciprofloxacin competition assay, TOP2-mediated DNA cleavage induced by ATP-sensitive but not ATP-insensitive poisons was shown to be antagonized by ciprofloxacin. These results suggest that ATP-bound TOP2 may be the specific target of ATP-sensitive TOP2 poisons. Using Lac repressor-operator complexes as roadblocks, we show that ATP-bound TOP2 acts as a circular clamp capable of entering DNA ends and sliding on unobstructed duplex DNA.

Topoisomerase II (TOP2) 1 catalyzes DNA topological reactions via a DNA breakage/reunion mechanism. The DNA topological reactions allow the enzyme to segregate interlocked chromosomal DNA at mitosis (1)(2)(3) and to remove excess DNA supercoils generated during processes such as DNA replication, RNA transcription, and chromosome condensation (4 -7). The breakage/reunion reaction of TOP2, which is ATP-dependent, can be interrupted by many antitumor drugs (TOP2 poisons) resulting in accumulation of a TOP2-DNA covalent intermediate, the cleavable complex (8). Accumulation of TOP2 cleavable complexes causes tumor cell death (8).
The molecular mechanism(s) by which TOP2 poisons interfere with the breakage/reunion reaction of TOP2 is largely unknown. Several studies have implicated a possible role of ATP in modulating the pharmacological action of TOP2 poisons; uncouplers of oxidative phosphorylation (e.g. DNP and 2-deoxyglucose) have been shown to enhance survival of adriamycin (doxorubicin)-treated Chinese hamster cells (9). Similarly, DNP, 2-deoxyglucose, and sodium cyanide, all of which affect ATP metabolism, effectively protect L1210 cells from the cytotoxic action of VM-26 and m-AMSA (10). Simultaneous cotreatment with DNP or novobiocin has been shown to abrogate m-AMSA cytotoxicity in Chinese hamster cells (11). VP-16 (etoposide)-induced chromosome-type aberrations (mainly breaks and exchanges) in cultured Chinese hamster lung fibroblasts are reduced also by cotreatment with DNP (12). Whether these effects are due to a direct effect of ATP on TOP2-mediated DNA cleavage induced by TOP2 poisons is not known.
Studies in bacteria have demonstrated also that the ATP/ ADP ratio plays a critical role in modulating the supercoiling state of chromosomal DNA as well as cytotoxicity of quinolone antibiotics (13)(14)(15). In this case, the role of ATP/ADP has been shown to directly affect TOP2-mediated DNA cleavage in the presence of quinolones (13). Studies of a drug-resistant TOP2 from mammalian cells have demonstrated also that ATP plays a direct role in modulating TOP2-mediated DNA cleavage in vitro (16,17).
In the current studies, we show that the ATP-bound conformation is the target of a class of ATP-sensitive TOP2 poisons that includes many clinically useful TOP2-directed antitumor drugs such as doxorubicin, mitoxantrone, VP-16 (etoposide), and m-AMSA. The ATP-bound form of TOP2 has been shown to be a circular protein clamp based on biochemical and x-ray crystallographic studies (18 -20). Using Lac repressor-operator complexes as roadblocks, we have demonstrated further that the circular TOP2 protein clamp is capable of sliding on unobstructed duplex DNA.

EXPERIMENTAL PROCEDURES
Chemicals and Drugs-VM-26 was a gift from Bristol-Myers Squibb Co. m-AMSA, mitoxantrone, and amonafide were obtained from the Drug Synthesis and Chemistry Branch, Division of Cancer Treatment, National Cancer Institute. Batracylin was a gift from Dr. C. C. Cheng (University of Kansas). 5-Hydroxy-1,4-naphthoquinone was obtained from Aldrich. All drugs were dissolved in Me 2 SO (10 mM) and kept frozen in aliquots at Ϫ20°C. Ciprofloxacin was obtained from Bayer. ATP, AMPPNP, and ADP were purchased from Sigma. Except for fetal bovine serum, which was obtained from Gemini Biotech, media and other agents for tissue culture were purchased from Life Technologies, Inc. [␣-32 P]dATP (3000 Ci/mmol) was obtained from DuPont.
Construction of Mutant Topoisomerase II Overexpression Plasmid-The mutation C427A was generated by PCR-based site-directed mutagenesis. Three rounds of PCR were carried out to introduce the C427A into human TOP2␣ cDNA in YEpWob6. The resulting plasmid is named Yhtop2␣C427A. The following four primers were used (the mutations are marked in bold, and the restriction sites are underlined): primer A, 5Ј-ACGCGTCGACGAATTCGACAGGTTATC-3Ј SalI; primer B, 5Ј-ACAAGAAGGCCTCAGCTGTA-3Ј; primer C, 5Ј-TACAGCTGAGGC-CTTCTTGT-3Ј; and primer D, 5Ј-GCCTGGTACCAAACTGAC-3Ј KpnI.
Primers B and C contain the alanine codon instead of the wild-type cysteine codon. Primers A and D contain the recognition sites for SalI and KpnI, respectively. Two fragments, AC and BD, which have the alanine codon, were generated during the first round PCR in the presence of YEpWob6 DNA. After denaturation and renaturation of fragments AC and BD, two cycles of a second round of PCR without any primer were carried out to generate a small amount of the fragment containing C427A as the template for the third round of PCR. The third round of PCR amplified the fragment containing C427A using primers A and D. SalI and KpnI were used to digest both the fragment-containing C427A and YEpWob6 (partially digested with SalI), and ligation was carried out at 14°C overnight. The mutated site was confirmed by sequence analysis.
Enzymes and Nucleic Acids-TOP2 was purified to homogeneity from calf thymus glands according to the published procedure (21). Full-length human TOP2␤ cDNA (hTOP2␤ cDNA) was isolated by reverse transcription-PCR using mRNA isolated from human U937 cells and primers with sequences according to the published sequence of HeLa TOP2␤ (22). For overexpression, hTOP2␤ cDNA was used to replace the human TOP2␣ cDNA (hTOP2␣ cDNA) in YEpWob6 (23). The resulting plasmid, YEphTOP2␤, then was used to transform protease-deficient yeast BCY123 (23). Purification of both TOP2 isozymes and mutant enzyme were performed following the published procedure (23). Lac repressor was a kind gift from Dr. Kathleen S. Matthews (Rice University). YEpG (24) is a derivative of YEP24. pY␤YOm was constructed by inserting a 42-bp DNA oligomer containing a 21-bp essential Lac repressor binding site (25) into the NotI site in pBR␤Y (26). pY␤YOd, which contains two Lac repressor binding sites, was constructed by inserting the same 42-bp oligomer into the XbaI site in pY␤YOm. All plasmids were purified using the Qiagen purification kit.
Preparation of End-labeled DNA Fragments-3Ј end-labeling of plasmid DNA was performed as described previously (21). Briefly, 10 g of DNA was digested with a proper restriction enzyme followed by labeling at its 3Ј ends with the large fragment of Escherichia coli DNA polymerase I and [␣-32 P]dATP. Unincorporated triphosphates were removed by two cycles of ethanol precipitation in the presence of 2.5 M ammonium acetate.
TOP2 Cleavage Assay-The TOP2 cleavage assay was performed as described previously (27). The reaction mixtures (20 l each) containing 40 mM Tris-HCl, pH 7.5, 100 mM KCl, 10 mM MgCl 2 , 1.0 mM ATP, 0.5 mM dithiothreitol, 0.5 mM EDTA, 30 g/ml bovine serum albumin, 20 ng of 3Ј end-labeled DNA, 10 ng of TOP2, and various drugs were incubated at 37°C for 30 min. The reactions were terminated by the addition of 5 l of 5% SDS and 150 g/ml proteinase K and incubated for an additional 60 min at 37°C. After the addition of sucrose (5% final concentration) and bromphenol blue (0.05 mg/ml final concentration), DNA samples were loaded onto a 1% agarose gel in TPE (90 mM Tris phosphate, 2 mM EDTA, pH 8.0) buffer. Gels then were dried onto Whatman No. 3MM chromatographic paper and autoradiographed at Ϫ80°C using Kodak XAR-5 films.

RESULTS
Differential ATP Stimulation of TOP2-mediated DNA Cleavage Induced by TOP2 Poisons-The possible effect of ATP on TOP2-mediated DNA cleavage was studied using purified TOP2 and various TOP2 poisons. As shown in Fig. 1A, 1 mM ATP stimulated calf thymus TOP2-mediated DNA cleavage induced by VM-26 by about 60-fold ( Fig. 1A) (ATP stimulation is estimated as the -fold of increased drug concentration in the absence of ATP for achieving the same extent of cleavage in the presence of ATP). By contrast, TOP2-mediated DNA cleavage induced by amonafide was affected only slightly (less than 3-fold) by 1 mM ATP (Fig. 1B). The nonhydrolyzable ATP analog, AMPPNP, gave similar results as ATP (data not shown). Similar results were obtained with recombinant human TOP2␣ and TOP2␤ (data not shown). TOP2 poisons that are highly (about 30 -100-fold) stimulated by ATP in the TOP2-mediated DNA cleavage assay include VM-26, VP-16, m-AMSA, doxorubicin, and mitoxantrone (referred to as ATP-sensitive TOP2 poisons). TOP2 poisons that are affected only slightly by ATP (less than 3-fold) include amonafide, batracylin, and menadione (referred to as ATP-insensitive TOP2 poisons).
ADP Antagonizes ATP-stimulated DNA Cleavage Activity and ATP-dependent Strand-passing Activity of TOP2-Although ATP strongly stimulated TOP2-mediated DNA cleavage induced by VM-26, ADP alone had no such effect (Fig. 2). However, in the presence of ATP, ADP effectively antagonized the ATP stimulatory effect on TOP2-mediated DNA cleavage induced by VM-26 (Fig. 2). The antagonistic effect of ADP on VM-26-induced DNA cleavage was observed also with other ATP-sensitive TOP2 poisons (data not shown). By contrast, ATP had a minimum effect on TOP2-mediated DNA cleavage in the presence of amonafide, an ATP-insensitive TOP2 poison. In this case, ADP had a minimal effect on TOP2-mediated DNA cleavage induced by amonafide in the presence of ATP (Fig. 2). These results suggest that ADP specifically antagonizes the ATP-stimulatory effect on TOP2-mediated DNA cleavage induced by ATP-sensitive TOP2 poisons. We also had tested the effect of ADP on the catalytic activity of TOP2 using a P4 unknotting assay. As shown in Fig. 3, ADP also effectively antagonized the P4 unknotting activity of TOP2.
C427A Mutant TOP2␣ Is Cross-resistant to ATP-sensitive but Not ATP-insensitive TOP2 Poisons-To characterize the role of ATP in the action of TOP2 poisons further, we generated C427A mutant TOP2␣. The cysteine 427 is located in the ATPase domain of TOP2␣ (31). In the absence of ATP, C427A mutant TOP2␣ exhibited almost identical sensitivity to both VM-26 and amonafide compared with wild-type TOP2␣ in a standard DNA cleavage assay (Fig. 4). Strikingly, in the presence of 1 mM ATP, C427A mutant TOP2␣ was at least 10-fold more resistant to VM-26 as compared with the wild-type enzyme. The resistance of C427A mutant TOP2␣ to VM-26 is apparently caused by a reduced ATP-stimulatory effect on DNA cleavage as compared with the wild-type enzyme (Fig. 4). By contrast, both mutant and wild-type TOP2␣ were equally sensitive to amonafide (Fig. 4). We also tested other ATPsensitive drugs such as doxorubicin, m-AMSA, mitoxantrone, and CP115,953. Similar results to VM-26 were observed (data not shown). These results suggest that C427A mutant TOP2␣ can distinguish between ATP-sensitive and -insensitive TOP2 poisons, possibly because of its altered interaction with ATP.

C427A Mutant Enzyme Exhibits Reduced ATPase Activity and Increased ATP Requirement for Its Catalytic Activity-
Previous studies on another multidrug-resistant mutant TOP2␣ mutant have shown that the mutant enzyme exhibits an increased ATP requirement for catalysis (16). To test whether C427A behaves similarly, the catalytic activity of the C427A mutant TOP2␣ was measured by a P4 unknotting assay in the presence of two different concentrations of ATP. As shown in Fig. 5, C427A mutant TOP2␣ was about 5-fold less active than the wild-type enzyme in the presence of 1 mM ATP. However, in the presence of 50 M ATP, C427A was at least 25-fold less active than the wild-type enzyme. This result is similar to the result from the experiment performed on another mutant TOP2␣ enzyme, R450Q TOP2␣, which is cross-resistant to ATP-sensitive TOP2 poisons and exhibits an increased ATP requirement for enzyme catalysis (17).
The DNA-stimulated ATPase activity of C427A mutant TOP2␣ also was measured and shown to be much reduced relative to the wild-type enzyme. The V max was reduced from 60 to 11 mM min Ϫ1 , and K m was increased from 0.78 to 3.2 mM.
Ciprofloxacin Antagonizes TOP2-mediated DNA Cleavage Induced by ATP-sensitive but Not ATP-insensitive TOP2 Poisons-Ciprofloxacin is known to interact with TOP2 but does not induce significant TOP2-mediated DNA cleavage (32). Consequently, ciprofloxacin has been used to compete with other TOP2 poisons in a standard DNA cleavage assay to assess possible overlap of interaction domains on TOP2 (32). Based on this kind of ciprofloxacin competition assay, it has been suggested that various TOP2 poisons including etoposide, m-AMSA, genistein, and the antineoplastic quinolone, CP-115,953, share a common interaction domain with ciprofloxacin on TOP2 (32). To test whether ATP-sensitive and -insensitive TOP2 poisons may interact with different domains on TOP2, we performed the ciprofloxacin competition assay (32). As shown in Fig. 6, ciprofloxacin reduced TOP2-mediated DNA cleavage induced by VM-26 as evidenced by the gradual increase in band intensity of the uncleaved DNA bands (see the ATP-bound TOP2 Is a Sliding Protein Clamp-Previous studies have suggested that yeast TOP2 when bound to AMP-PNP can undergo a conformational change into a circular protein clamp (19,20), which is consistent with results from x-ray crystallographic studies (18). To test whether calf thymus TOP2 also can form a circular protein clamp in its ATP-bound form, we have designed a more stringent assay requiring the TOP2 protein clamp to slide on unobstructed DNA under physiological conditions. As shown in Fig. 7, a linear DNA (8310-bp, 32 P end-labeled) with two internally bound Lac repressor molecules at their respective Lac operator sites was used to demonstrate entry and sliding of AMPPNP-bound TOP2. Calf thymus TOP2 was reacted first with AMPPNP to form a circular protein clamp and then incubated with the linear DNA bound by Lac repressors. VM-26 or amonafide was used subsequently to locate the TOP2 sliding clamps on DNA by inducing TOP2mediated DNA cleavage. As shown in Fig. 7 3 and 5, respectively). The lower part of the gel was overexposed to reveal cleavage sites from the other end (Fig. 7). The results from this experiment support the previous claim that AMPPNP-bound TOP2 is in the form of a circular protein clamp. In addition, this experiment has demonstrated that AMPPNP-bound TOP2 is a protein clamp capable of entering DNA ends and sliding on unobstructed duplex DNA while retaining sensitivity to VM-26.

DISCUSSION
Our results have demonstrated that different TOP2-mediated DNA cleavage induced by various TOP2 poisons exhibits a different degree of ATP dependence. The differences in ATP dependence among various TOP2 poisons may reflect differences in their interaction with TOP2 and/or TOP2-DNA complexes. Based on our current results, there seems to be two distinct classes of TOP2 poisons, ATP-sensitive and ATP-insensitive.
These two classes of TOP2 poisons show quite different responses to ATP stimulation in the standard DNA cleavage assay. The specific antagonistic effect of ADP against ATPsensitive but not -insensitive TOP2 poisons has demonstrated further the differences between these two classes of TOP2 poisons. The fact that ADP also strongly inhibits ATP-dependent catalytic activity of TOP2 suggests that ADP may compete with ATP both in enzyme catalysis and cleavable complex formation by the same mechanism. Studies in bacteria have established that the ATP/ADP ratio is a critical determinant for the supercoiling state in cells probably because of the sensitivity of DNA gyrase to the ATP/ADP ratio (14,15). More recent studies have demonstrated also that quinolone-induced DNA cleavage depends strongly on the ATP/ADP ratio, both in cells and using purified gyrase (13). These results suggest that the ATP/ADP ratio may be a common determinant for sensitivity/resistance to both antibiotics and antitumor drugs directed against TOP2. Although ATP-sensitive TOP2 poisons used in this work have very disparate structures, their interaction domains with TOP2 have been suggested to overlap (32).
Previous studies have demonstrated that a multidrug-resistant mutant TOP2␣ is cross-resistant to all ATP-sensitive TOP2 poisons (17). This multidrug-resistant mutant TOP2␣ was shown to exhibit an increased requirement of ATP for catalysis and cleavage. It has been suggested that the R450Q mutation on this mutant TOP2␣ is responsible for altered ATP utilization and cross-resistance to ATP-sensitive TOP2 poisons (17). This mutation is located in a Walker consensus motif (17). In the current study, we have created another mutation C427A on TOP2␣. Like the R450Q mutant TOP2␣, C427A mutant TOP2␣ also exhibits multidrug resistance to all ATP-sensitive poisons. Interestingly, C427A mutant TOP2␣ retains the same sensitivity to ATP-insensitive TOP2 poisons as compared with the wild-type enzyme. C427A mutant TOP2␣ exhibits reduced ATPase activity and an increased requirement of ATP for catalysis. Taken together, these results suggest that ATP-sensitive and -insensitive TOP2 poisons interfere with the breakage/ reunion reaction of TOP2 by distinct mechanisms and that ATP-sensitive TOP2 poisons may interfere specifically with a step in TOP2 catalysis requiring ATP utilization.
Results from the ciprofloxacin competition experiment have suggested that the ATP-insensitive TOP2 poisons do not share the same interaction domain on TOP2 with ATP-sensitive TOP2 poisons. This result suggests that ATP-sensitive TOP2 poisons may target TOP2 with a distinct conformation compared with ATP-insensitive poisons.
Based on our results, it seems plausible that ATP-sensitive TOP2 poisons may specifically target an ATP-bound conformation of TOP2. Our current studies have suggested that AMP-PNP-bound TOP2 is capable of entering duplex DNA only from its ends, consistent with the closed circular clamp conformation of ATP-bound TOP2 proposed previously on the basis of studies of yeast TOP2 (19). Our results also indicate that upon entry, AMPPNP-bound TOP2 can slide on unobstructed DNA under physiological conditions. Previous studies on Drosophila TOP2 and yeast TOP2 have implicated linear diffusion in high salt conditions that presumably weaken protein-DNA interactions to allow mobility of the protein circular clamp (19,33). Our results, however, show that AMPPNP-bound mammalian TOP2 is able to linearly diffuse under physiological conditions. The ability of TOP2 to slide on DNA under physiological conditions may imply a role of limited linear diffusion, dictated by ATP binding and hydrolysis, in its strand-passing reaction. Our limited understanding of the role of ATP in TOP2 catalysis precludes us from any meaningful speculation on the mechanistic and/or functional implications of the sliding action of ATP-bound TOP2. The resistance of TOP2 poisons has been studied in cells under stress conditions (e.g. hypoxia and glucose deprivation) that are associated often with solid tumors (9 -12, 34 -38). Reduced TOP2␣ levels in stressed cells have been found and suggested to be responsible in part for the FIG. 7. AMPPNP-bound TOP2 is a sliding protein clamp. pY␤YOd DNA, containing two Lac repressor binding sites, was digested with HindIII and 3Ј end-labeled with [␣-32 P]dATP by the Klenow polymerase. The labeled DNA then was digested with ScaI, resulting in two one-end-labeled DNA of the sizes of 8310 and 546 bp, respectively. The two Lac repressor binding sites are located at 808 -829 and 4527-4548 bp from the labeled end of the 8310-bp fragment. Demonstration of entry and sliding of AMPPNP-bound TOP2 was performed in three sequential steps: (a) calf thymus TOP2 (30 ng each) was incubated with 2 mM AMPPNP or ATP at 37°C for 10 min; (b) 3 ϫ 10 Ϫ13 M labeled DNA was incubated with 2 ϫ 10 Ϫ11 M Lac repressor protein in 40 mM Tris-HCl, pH 7.5, 100 mM KCl, 10 mM MgCl 2 , 0.5 mM dithiothreitol, 0.5 mM EDTA, and 30 g/ml bovine serum albumin at room temperature for 7 min; (c) AMPPNP-bound TOP2 and repressor-bound DNA were mixed together followed by incubation at 37°C for 7 min in the presence of 0.5 M VM-26. After treatment with SDS (final concentration 1%) and proteinase K (final concentration 200 g/ml) for 1 h at 37°C, the samples were subjected to electrophoresis in 1% agarose gel. Lane 1, DNA alone; lane 2, contained Lac repressor and TOP2; lane 3, contained Lac repressor, TOP2, and VM-26. Lanes 4 and 5 were identical to lanes 2 and 3, respectively, except that AMPPNP rather than ATP was present in each reaction. Lane 6, DNA size markers. To the right of the panel, the 8310-bp DNA, which is 3Ј end-labeled with 32 P (see the asterisk), is aligned schematically with the gel to indicate the cleavage sites relative to the Lac repressor binding sites on DNA. In the presence of AMPPNP, TOP2 is shown as a protein doughnut that can enter linear DNA only through DNA ends. Drug-induced (also background) cleavage sites mark the accessible regions of DNA to the TOP2 sliding clamp. observed resistance (38 -40). Our results raise the possibility that alteration in ATP/ADP ratios, which is known to occur in hypoxic and nutrient-depleted cells (41), may contribute to the overall resistance mechanisms through its modulation on TOP2-mediated DNA cleavage. Thus, it seems plausible that ATP-insensitive TOP2 poisons may be useful particularly for treating hypoxic tumors that have compromised ATP/ADP ratios.