Topoisomerase II-mediated DNA Cleavage and Religation in the Absence of Base Pairing

The interaction of topoisomerase II with its DNA cleavage site is critical to the physiological functions of the enzyme. Despite this importance, the specific enzyme-DNA interactions that drive topoisomerase II-mediated DNA cleavage and religation are poorly understood. Therefore, to dissect interactions between the enzyme and its cleavage site, abasic DNA lesions were incorporated into a bilaterally symmetrical and identical cleavage site. Results indicate that topoisomerase II has unique interactions with each position of the 4-base overhang generated by enzyme-mediated DNA cleavage. Lesions located 2 bases 3′ to the point of scission stimulated cleavage the most, whereas those 3 bases from the point of scission stimulated cleavage the least. Moreover, an additive and in some cases synergistic cleavage enhancement was observed in oligonucleotides that contained multiple DNA lesions, with levels reaching >60-fold higher than the wild-type substrate. Finally, topoisomerase II efficiently cleaved and religated a DNA substrate in which apyrimidinic sites were simultaneously incorporated at every position on one strand of the 4-base overhang. Therefore, unlike classical DNA ligases in which base pairing is the driving force behind closure of the DNA break, it appears that for topoisomerase II, the enzyme is responsible for the spatial orientation of the DNA termini for ligation.

Eukaryotic topoisomerase II is an enzyme that is required for a number of indispensable nuclear processes, including DNA replication, recombination, and chromosome segregation (1)(2)(3)(4). Moreover, it is the primary target for some of the most effective and commonly employed anticancer chemotherapy regimens used for the treatment of human malignancies (5)(6)(7)(8)(9). However, in order for topoisomerase II to fulfill its pivotal role in any of these processes, it must create double-stranded breaks in the genetic material.
Topoisomerase II generates these breaks by the concerted actions of two active site tyrosyl residues, each of which induces a transient, single-stranded nick in opposite strands of the DNA helix, such that a 4-base 5Ј-overhang is generated (10,11). The integrity of the genetic material is preserved during the cleavage reaction by the formation of covalent phosphotyrosyl bonds between these residues and the newly generated 5Ј termini (12). Following ATP binding, a second helix is translocated through the topoisomerase II-associated double-stranded DNA break, and the cleaved DNA is religated (13)(14)(15). Despite the importance of the DNA cleavage/religation reaction to the functions of topoisomerase II, relatively little is known regarding the specific enzyme-DNA interactions that drive this reaction. Although consensus DNA cleavage sequences have been reported for the eukaryotic enzyme, they are relatively weak and show little similarity to one another (8,16). Thus, the molecular interactions that underlie the site specificity of topoisomerase II remain an enigma.
Previous attempts to dissect interactions between the enzyme and its cleavage site characterized the effects of DNA lesions (apurinic sites, apyrimidinic sites, or base mismatches) incorporated at specific positions in a topoisomerase II cleavage site (17)(18)(19)(20). All of these lesions displayed a strict positional specificity. When they were located within the 4-base overhang generated by cleavage, levels of DNA scission increased ϳ3-18-fold. Conversely, lesions located immediately outside the 4-base overhang inhibited DNA cleavage. Although positionspecific variations in stimulatory DNA lesions were observed, conclusions regarding nucleotide positions within the 4-base overhang were limited because the relative effects of position (e.g. removal of a base at the point of scission versus one base away) as opposed to sequence (e.g. removal of a guanine versus an adenine) could not be distinguished.
To bypass these limitations and more fully define interactions between the enzyme and specific positions within its DNA cleavage site, abasic sites were incorporated into a topoisomerase II cleavage site in which the central bases were bilaterally symmetrical and identical (see Fig. 1). Because abasic lesions contained within a given strand of this cleavage site differed only by location rather than the type of base removed, this site allowed a systematic analysis of the 4-base overhang in topoisomerase II-mediated DNA cleavage/religation. Results indicate that topoisomerase II displays a positional preference, such that the efficacy of DNA lesions is dependent on their location within the cleavage site. In addition, the enzyme efficiently cleaves and religates a DNA substrate in which apyrimidinic sites are simultaneously incorporated at every position on one strand of the 4-base overhang, demonstrating that base pairing within the cleavage site is not required either for formation or closure of the DNA break.

EXPERIMENTAL PROCEDURES
Preparation of Oligonucleotides-A 40-base single-stranded oligonucleotide that corresponds to residues 1072-1111 of the MLL oncogene (21,22) and its complementary oligonucleotide were prepared by Cruachem Inc. The sequences of the top and bottom oligonucleotides were 5Ј-GCCTGGGTGACAAAGC2AAAACACTGTCTCCAAAAAAAATT-3Ј and 5Ј-AATTTTTTTTGGAGACAGTG2TTTTGCTTTGTCACCCAGG-C-3Ј, respectively. The points of topoisomerase II-mediated DNA cleavage are denoted by the arrows. Single-stranded oligonucleotides containing a tetrahydrofuran abasic site analogue were prepared in a similar manner utilizing a tetrahydrofuran phosphoramidite. Single-stranded oligonucleotides were labeled on their 5Ј termini with [ 32 P]phosphate, purified by polyacrylamide gel electrophoresis, and annealed as described previously (18).
Topoisomerase II-mediated DNA Cleavage-Topoisomerase II-mediated DNA cleavage reactions were carried out by a protocol similar to that of Kingma et al. (19). Reactions contained 100 nM oligonucleotide in 19 l of cleavage buffer (10 mM Tris-HCl, pH 7.9, 0.1 mM EDTA, 100 mM KCl, and 2.5% glycerol) that contained 5 mM MgCl 2 and were initiated by the addition of 1 l of human topoisomerase II␣ (final concentration, 150 nM). (Human topoisomerase II␣ was purified from Saccharomyces cerevisiae as described previously (19).) Reactions were incubated for 10 min at 37°C and stopped with 2 l of 10% SDS followed by 1.5 l of 250 mM EDTA. When appropriate, cleavage reactions were reversed by the addition of 1.5 l of 250 mM EDTA for 5 min at 37°C prior to the detergent. Cleavage products were digested with proteinase K, precipitated twice with ethanol, and resolved by electrophoresis in denaturing 7 M urea, 14% polyacrylamide gels as described previously (18). Reaction products were visualized and quantified using a PhosphorImager system. In all cases (other than those that incorporated an abasic site on both strands), cleavage was monitored on the complementary wildtype strand. Levels of DNA cleavage were calculated either relative to that obtained with the wild-type substrate or from the percentage of total substrate that was cleaved.
Topoisomerase II-mediated DNA Religation-DNA religation assays were carried out by a modification of the procedure of Osheroff and Zechiedrich (23). Cleavage/religation equilibria were established as described above in cleavage buffer that contained 5 mM CaCl 2 . Kinetically competent topoisomerase II-DNA cleavage complexes were trapped by the addition of EDTA (final concentration, 6 mM). NaCl was added (final concentration, 500 mM) to prevent recleavage. Religation was initiated by the addition of MgCl 2 (final concentration, 0.1 mM) and terminated by the addition of 2 l of 10% SDS at various times up to 60 s. Samples were prepared and analyzed as described above. The apparent first order rate of DNA religation was determined by quantifying the loss of the cleavage product.

RESULTS
To dissect the recognition of the DNA cleavage site by topoisomerase II and the role of base pairing in the cleavage/religation reaction of the enzyme, abasic lesions were incorporated into a 40-base pair oligonucleotide that corresponds to residues 1072-1111 of the MLL gene and includes a topoisomerase II cleavage site at nucleotide position 1087 (21,22). This bilaterally symmetrical site was utilized because it contains identical bases along each strand of the 4-base overhang (i.e. all adenine residues on the top strand and thymine residues on the bottom) ( Fig. 1). The positions of DNA lesions are designated relative to the point of cleavage with the points of scission on both the top and bottom strands located 5Ј to the ϩ1 base. Cleavage of the top and bottom strands of this 5Ј-labeled oligonucleotide by human topoisomerase II␣ results in the formation of radioactive 16-and 20-base products, respectively, such that a 4-base 5Ј-cleavage overhang is generated.
Topoisomerase II Displays a Positional Preference for DNA Lesions in the 4-base Overhang-The sensitivity of human topoisomerase II␣ toward alterations at specific positions in the 4-base cleavage overhang was characterized by incorporating abasic DNA lesions at each position within the top and bottom strands (Fig. 1). Consistent with our previous work (17,19), each individual lesion stimulated enzyme-mediated DNA cleavage, and greater levels of cleavage were observed with apurinic sites (top strand) than with apyrimidinic sites (bottom strand). However, levels of cleavage stimulation for either type of lesion were not uniform across the cleavage site. Consistently, lesions at the ϩ2 position stimulated DNA scission the most, ϳ9 -11-fold above the wild-type substrate. In contrast to the strong preference for the ϩ2 position, only a weak preference (ϳ2-3-fold above wild-type) for the ϩ3 position was detected on both strands. The preference for lesions placed at the ϩ1 versus ϩ4 positions was irregular. When comparing these two positions, a slight preference for the ϩ1 position was observed on the top strand, whereas on the bottom strand the ϩ4 position was preferred. As discussed below, these findings suggest a predisposition of the enzyme for the "left side" of the cleavage site.
Although the cleavage site utilized in this study permits a unique analysis of the positional preferences of the enzyme, the generality of these conclusions must be tempered by the fact that only a single site was examined. However, these preferences are supported by results obtained with at least two other DNA cleavage sites (19). 1 Even with the above caveat, it is clear that topoisomerase II does not have uniform interactions with each base of the overhang; rather the enzyme appears to have a positional preference for DNA lesions and displays specific responses to alterations at each position.
Multiple Abasic DNA Lesions Stimulate Topoisomerase IImediated DNA Cleavage-Several clinically relevant anticancer drugs exert their chemotherapeutic actions by increasing levels of covalent topoisomerase II-cleaved DNA complexes (6 -9). As a result of their actions, drugs that "poison" topoisomerase II (i.e. shift the cleavage/religation equilibrium toward the cleaved state) increase levels of protein-associated DNA breaks and are lethal to rapidly dividing cancer cells. The original models proposed for the actions of these poisons generally placed a drug molecule at the point of scission on each strand of the double helix with two molecules acting in concert to stimulate double-stranded DNA cleavage (6, 8, 16, 24 -26).
Based on the results of previous studies with DNA lesions, an alternative model known as the "positional poison model" was proposed that encompassed the actions of both anticancer drugs and DNA lesions (18). Unlike previous drug models, the positional poison model suggested that two drug molecules were not required to stimulate double-stranded DNA cleavage, because a single, strand-specific DNA lesion was capable of enhancing enzyme-mediated DNA cleavage at both points of scission. Therefore, to extend our use of DNA lesions as a probe of drug mechanism, a series of oligonucleotides containing two abasic sites was analyzed to determine whether multiple lesions might also display a positional preference.
In most cases, the stimulatory effects of multiple DNA lesions appeared to be at least additive (i.e. levels of cleavage 1 S. D. Cline and N. Osheroff, unpublished results. with two DNA lesions were as high as the theoretical sum of cleavage levels observed with the individual abasic sites) (Fig.  2, compare closed and open bars). This additive effect was observed when abasic sites were positioned adjacent to each other (17), separated by a gap, or located at both points of scission in either the ϩ1 top/ϩ1 bottom or the ϩ4 top/ϩ4 bottom positions (the postulated locations of drugs (6, 8, 16, 24 -26)). Therefore, although only one topoisomerase II poison is required to induce double-stranded DNA cleavage, two appear to produce greater levels of cleavage enhancement.
Remarkably, in some cases a synergistic effect was observed. When lesions were incorporated at the ϩ1/ϩ2 or the ϩ2/ϩ3 positions on the top strand or the ϩ2/ϩ3 or the ϩ3/ϩ4 positions on the bottom strand levels of cleavage stimulation reached Ͼ60-fold and were ϳ2-3-fold greater than the predicted additive effects of the individual DNA lesions. Because this superadditive effect was only observed with abasic lesions at these positions, it appears that synergism within this cleavage site is dependent on two primary factors. First, the abasic sites must be located adjacent to each other. Second, lesions must be positioned on the left side of the cleavage site, such that the ϩ1 or ϩ2 positions on the top strand or the ϩ3 or ϩ4 positions on the bottom strand are included.
Topoisomerase II-mediated DNA Cleavage and Religation in the Absence of Base Pairing-Previous work by Westergaard and co-workers examined the importance of base pairing in ligation reactions between a topoisomerase II-cleaved DNA complex and external, acceptor fragments of DNA (27). In this intermolecular ligation reaction, base pairing of the external fragment to the cleaved DNA was important for the positioning of the incoming termini and enhanced rates of ligation as much as 8-fold. In contrast, studies that incorporated single abasic sites or base mismatches within a topoisomerase II cleavage site indicated that intramolecular DNA religation mediated by the enzyme (i.e. resealing the original DNA break covalently linked to topoisomerase II) was not inhibited by disruption of one of the 4 base pairs within the overhang (17,18). Therefore, to further define the importance of base pairing within the 4-base overhang, enzyme-mediated DNA cleavage and intramolecular religation were examined in the presence of multiple abasic sites.
As seen in Fig. 3, topoisomerase II tolerated the removal of all four pyrimidines on the bottom strand (apyrimidinic sites were used because they are less disruptive to the structure of the double helix than apurinic sites (28 -30)). In fact, cleavage increased significantly with the removal of each successive base and reached levels Ͼ60-fold above the wild-type substrate when all 4 bases on the bottom strand were removed.
Despite these dramatic increases in DNA scission, Ն90% of the cleavage was reversed by the addition of EDTA, even when all the base pairing was removed from the 4-base overhang (Fig. 3). In addition, under conditions in which the enzyme was capable of carrying out DNA strand passage (i.e. in the presence of ATP), cleavage was still reversible in the absence of base pairing. Therefore, base pairing within the overhang is not required for either DNA cleavage or religation mediated by topoisomerase II.
Although not necessary for religation, the potential contributions of base pairing to the rate of this reaction are unknown. Therefore, the kinetics of religation were examined using a series of apyrimidinic substrates (Fig. 4). Removal of one (ϩ1 position) or two bases (either at the ϩ1/ϩ2 or ϩ3/ϩ4 positions) from the bottom strand did not inhibit rates of religation. In fact, rates appeared to increase slightly, which is consistent with earlier findings with single DNA lesions (17). Removal of 3 or 4 bases decreased rates of religation ϳ2-3-fold, a relatively minor effect when compared with the Ͼ60-fold stimulation of cleavage induced by four apyrimidinic sites. It is unclear whether this small decrease in the rate of religation reflects a minor contribution of base pairing to this reaction or whether it reflects a lesion-induced structural distortion of the cleavage site in addition to the potential effects of base pairing. This notwithstanding, it is clear that under the conditions used in this system, base pairing is not the driving force behind topoisomerase II-mediated DNA religation. DISCUSSION Determining the specific molecular interactions that govern the topoisomerase II-mediated DNA cleavage/religation reac- tion is vital to our understanding of this essential enzyme. In this study, abasic DNA lesions were used to reveal some of the critical features of this reaction.
Results indicate that topoisomerase II has a positional preference for DNA lesions located within the 4-base cleavage overhang. Although every DNA lesion introduced within this region stimulated DNA scission, levels of cleavage with abasic sites at the ϩ2 position were consistently the highest, and those at the ϩ3 position were the lowest. Moreover, an additive and in some cases synergistic effect was observed in oligonucleotides that contained more than one abasic site. Considering the dramatic cleavage stimulation (Ͼ60-fold above wild-type) observed with certain combinations of DNA lesions, it is tempting to speculate that the efficacy of anticancer drugs might be substantially improved if they were capable of distorting multiple base pairs within the 4-base overhang. Finally, two lines of evidence suggest that the interactions between topoisomerase II and this DNA cleavage sequence are biased toward the left side of this site. First, cleavage levels observed with single DNA lesions at the ϩ1 top or ϩ4 bottom positions on the left side of the cleavage site were ϳ2-fold higher than levels observed with lesions at the ϩ4 top or ϩ1 bottom positions, respectively, on the right side. Second, the synergistic effects of multiple abasic sites were only detected with adjacent DNA lesions that included at least one of the two positions on the left side of the cleavage site. Because the leftward bias was observed in independent experiments that examined either apurinic or apyrimidinic sites, it does not appear to be specific to the type of lesion utilized. Beyond this observation, the specific molecular interactions that underlie the leftward bias of this cleavage site are unclear. However, the homodimeric structure of the enzyme and the symmetrical nature of the central bases of this site suggest that this bias must originate from the nonsymmetrical region immediately outside the 4-base overhang.
Although removing the bases from the cleavage overhang significantly stimulates the DNA scission reaction of topoisomerase II, base pairing in this region is not required for religation of the break and in relative terms has only a minor effect on the reaction rate. This is in marked contrast to most classical DNA ligases, which are dramatically inhibited by the loss of base pairing within their substrate. How does topoi-somerase II religate a cleaved DNA substrate in the absence of base pairing? The defining property that distinguishes topoisomerase II from DNA ligases is its covalent linkage to the DNA strands containing the unpaired bases. This may explain why DNA termini covalently linked to topoisomerase II can be positioned and religated correctly by the enzyme without the assistance of base pairing. Thus, the driving force behind topoisomerase II-mediated DNA religation appears to be conformational changes within the enzyme rather than complementarity between the cohesive ends.
Finally, this work may address how some topoisomerase II poisons can stimulate DNA cleavage without inhibiting DNA religation (5,7). Although topoisomerase II preferentially recognizes and cleaves the double helix when it is distorted by a drug molecule or DNA lesion (as suggested in the positional poison model (18)), religation of the single-stranded DNA termini apparently is orchestrated primarily by the enzyme independent of the DNA structure. Thus, the same distortion of the double helix that alters the forward DNA cleavage reaction would have no effect on the reversal of this process.
In summary, the interaction of topoisomerase II with its DNA cleavage site is critical to the physiological and therapeutic functions of the enzyme. DNA lesions are rapidly becoming a powerful tool for understanding the factors that dictate these interactions.