Cytosine Arabinoside Lesions Are Position-specific Topoisomerase II Poisons and Stimulate DNA Cleavage Mediated by the Human Type II Enzymes*

Cytosine arabinoside (araC) is an important drug used for the treatment of human leukemias. In order to exert its cytotoxic effects, araC must be incorporated into chromosomal DNA. Although specific DNA lesions that involve base loss or modification stimulate nucleic acid cleavage mediated by type II topoisomerases, the effects of deoxyribose sugar ring modification on enzyme activity have not been examined. Therefore, the effects of incorporated araC residues on the DNA cleavage/religation equilibrium of human topoisomerase IIα and β were characterized. AraC lesions were position-specific topoisomerase II poisons and stimulated DNA scission mediated by both human type II enzymes. However, the positional specificity of araC residues differed from that previously reported for other cleavage-enhancing DNA lesions. Finally, additive or synergistic increases in DNA cleavage were observed in the presence of araC lesions and etoposide. These findings broaden the range of DNA lesions known to alter topoisomerase II function and raise the possibility that this enzyme may mediate some of the cellular effects of araC.

Cytosine arabinoside (araC) is an important drug used for the treatment of human leukemias. In order to exert its cytotoxic effects, araC must be incorporated into chromosomal DNA. Although specific DNA lesions that involve base loss or modification stimulate nucleic acid cleavage mediated by type II topoisomerases, the effects of deoxyribose sugar ring modification on enzyme activity have not been examined. Therefore, the effects of incorporated araC residues on the DNA cleavage/religation equilibrium of human topoisomerase II␣ and ␤ were characterized. AraC lesions were positionspecific topoisomerase II poisons and stimulated DNA scission mediated by both human type II enzymes. However, the positional specificity of araC residues differed from that previously reported for other cleavage-enhancing DNA lesions. Finally, additive or synergistic increases in DNA cleavage were observed in the presence of araC lesions and etoposide. These findings broaden the range of DNA lesions known to alter topoisomerase II function and raise the possibility that this enzyme may mediate some of the cellular effects of araC.
Beyond their established mutagenic or cytotoxic effects, physiologically relevant DNA lesions such as apurinic/apyrimidinic sites and deaminated cytosine residues profoundly influence the activity of eukaryotic type II topoisomerases (24). These lesions act as position-specific topoisomerase II "poisons" and stimulate enzyme-mediated DNA scission when they are located within the 4-base stagger that separates the two phosphodiester bonds cleaved by the enzyme (24 -28). Some nucleoside alterations, such as abasic sites, enhance doublestranded DNA scission with a potency that is 1,000-fold greater than that of etoposide (24 -26, 28).
Although DNA lesions that result from base loss or modification have been shown to alter the function of topoisomerase II (24 -28), the effects of sugar ring modification on enzyme activity have not been established. Therefore, the effects of incorporated araC residues on the DNA cleavage activity of human topoisomerase II␣ and -␤ were characterized. Substitution of a ␤-hydroxyl for a hydrogen at the 2Ј-position of deoxycytosine (converting the deoxyribose ring to an arabinose) stimulated DNA scission mediated by the ␣ and ␤ isoforms of human topoisomerase II but did so with a positional specificity that differed from those of other DNA lesions. Moreover, in contrast to previous findings with abasic sites (26), additive or synergistic increases in DNA cleavage were observed in the presence of araC lesions and etoposide. These findings suggest sugar ring modifications can alter topoisomerase II function and raise the possibility that the enzyme may mediate some of the physiological effects of araC.

EXPERIMENTAL PROCEDURES
Preparation of Oligonucleotides-A 42-base single-stranded oligonucleotide that corresponds to residues 1050 -1091 of the MLL oncogene (29) and its complementary oligonucleotide were synthesized. The sequences of the top and bottom oligonucleotides were 5Ј-ATGATTGTACCACTGCAG2TCCAGCCTGGGTGACAAAGCAAAA-3Ј and 5Ј-TTTTGCTTTGTCACCCAGGC2TGGACTGCAGTGGTACAAT-CAT-3Ј, respectively. Points of topoisomerase II-mediated DNA cleavage are denoted by arrows (28,29). Modified top and bottom oligonucleotides containing a cytosine arabinoside residue (Glen Research, Sterling, VA) were synthesized using the corresponding phosphoramidite. For DNA cleavage assays, single-stranded oligonucleotides were 32 P-labeled on their 5Ј termini with T4 polynucleotide kinase, purified by electrophoresis in a 7 M urea, 14% polyacrylamide gel, visualized by UV shadowing, excised, and eluted using the Qiagen gel extraction protocol. Complementary oligonucleotides were annealed by incubating equimolar amounts of each at 70°C for 10 min and cooling to 25°C (28). Etoposide was obtained from Sigma and stored as a 20 mM stock in Me 2 SO at room temperature.
Topoisomerase II-mediated DNA Cleavage-Human topoisomerase II␣ and ␤ were purified from Saccharomyces cerevisiae as described (28). DNA cleavage reactions contained 100 nM oligonucleotide in 19 l of cleavage buffer (10 mM Hepes-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 (150 nM final concentration). Reactions were incubated at 37°C for 10 min and stopped by the addition of 2 l of 10% SDS and 1.5 l of 250 mM EDTA. When appropriate, cleavage reactions were reversed by the addition of NaCl (500 mM final concentration) at 37°C for 5 min prior to the addition of detergent and EDTA. Reactions containing etoposide contained 100 M drug and 1% Me 2 SO (final concentration). DNA cleavage products were digested with proteinase K, precipitated with ethanol, and resolved by * This work was supported by Grant GM53960 from the National Institutes of Health. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
§ Trainee under National Institutes of Health Grant 5 T32 GM08320. ʈ To whom correspondence should be addressed: Dept. of Biochemistry, 654 Medical Research Bldg. I, Vanderbilt University School of Medicine, Nashville, TN 37232-0146. Tel.: 615-322-4338; Fax: 615-343-electrophoresis in denaturing 7 M urea, 14% polyacrylamide gels. Reaction products were visualized and quantified using a Molecular Dynamics PhosphorImager system. DNA cleavage was monitored on the complementary wild-type strand. Levels of DNA scission were calculated relative to that obtained with the wild-type substrate.
Topoisomerase II-mediated DNA Religation-DNA religation assays used a procedure modified from Osheroff and Zechiedrich (30). Cleavage/religation equilibria were established as described above in cleavage buffer containing 5 mM CaCl 2 . Kinetically competent topoisomerase II-DNA cleavage complexes were trapped by the addition of EDTA (6 mM final concentration). Sodium chloride (500 mM final concentration) was added to prevent recleavage, and religation was initiated by the addition of MgCl 2 (0.15 mM final concentration). Reactions were terminated at times up to 80 s by the addition of 1% SDS, and samples were analyzed as described above. The apparent first order rate of DNA religation was determined by quantifying the loss of cleaved DNA.

RESULTS AND DISCUSSION
AraC is a widely prescribed anticancer drug that has long been used for the treatment of human leukemias (1,2). In order to induce cytotoxicity, this agent must be converted to its activated triphosphate form and incorporated into chromosomal DNA (1)(2)(3)(4)(5)(6)(7)(8). Pretreatment of human cells with araC increases levels of DNA breaks induced by topoisomerase II-targeted anticancer drugs, suggesting possible physiological interactions between topoisomerase II and incorporated araC residues (31). Since a variety of DNA lesions have been shown to stimulate the DNA cleavage activity of the type II enzyme, the effects of incorporated araC residues on the DNA cleavage/ religation equilibrium of human topoisomerase II␣ and ␤ were characterized (24,32). AraC lesions, which contain an arabinose sugar in place of deoxyribose, represent the first sugar ring modification examined for activity against type II topoisomerases.
AraC Lesions Stimulate DNA Cleavage Mediated by Human Topoisomerase II␣ and ␤-AraC residues were incorporated into a 42-base pair oligonucleotide that contained a centrally located cleavage site for human type II topoisomerases (Fig. 1). The sequence was derived from a region of the MLL oncogene proximal to a leukemic breakpoint at chromosomal band 11q23 (29). Individual substrates contained single araC lesions incorporated at the indicated positions, and the points of DNA scission on the top and bottom strands are indicated by arrows (28,29).
Vertebrate cells contain two isoforms of topoisomerase II, ␣ and ␤ (32-35). Topoisomerase II␣ is dramatically up-regulated during periods of rapid cell proliferation, reflecting important and potentially specific roles in DNA replication and chromosomal segregation (32, 36 -38). In comparison, topoisomerase II␤ levels are relatively constant across both cell and growth cycles, indicating that this isoform may have a more general function in DNA and/or RNA processes (32,34,37). Because of the association of topoisomerase II␣ with replicating cells, initial studies focused on interactions between this enzyme isoform and incorporated araC residues.
As seen in Fig. 1 (closed bars), araC lesions stimulated DNA cleavage mediated by human topoisomerase II␣ in a positionspecific fashion. Consistent with previous findings for other DNA lesions, araC residues located within the 4-base cleavage overhang (i.e. the ϩ2 and ϩ3 positions on the top strand) increased levels of DNA scission (24, 26 -28, 39, 40). AraC incorporation at the ϩ2 position displayed the greatest effect on enzyme activity and increased cleavage ϳ5-fold.
AraC lesions located outside of the cleavage overhang 3Ј to the scissile bonds (i.e. the ϩ6 position on the top strand and the ϩ5 and ϩ8 positions on the bottom strand) decreased DNA scission. These findings are similar to those reported for other DNA lesions (24, 26 -28, 40). However, araC residues located outside of the cleavage overhang 5Ј to the scissile bonds (i.e. the Ϫ3 position on the top strand and the Ϫ1 position on the bottom strand) stimulated DNA scission as much as 3.5-fold.
AraC residues are the first DNA lesions reported to significantly enhance topoisomerase II-mediated scission when located outside the cleavage site of the enzyme. The basis for this novel positional specificity is not known; however, the structural distortions induced in DNA by incorporated araC residues are considerably different from those generated by other lesions, such as apurinic/apyrimidinic sites, that stimulate DNA scission mediated by topoisomerase II (41)(42)(43)(44)(45)(46)(47)(48)(49)(50)(51)(52)(53). In addition, the trans-2Ј-hydroxyl moiety of araC lesions projects into the major groove of the double helix and forms intrastrand hydrogen bonds (51, 52) that potentially could affect topoisomerase II-DNA interactions.
Three additional points should be noted. First, although incorporation of araC residues altered levels of DNA cleavage, the points of scission remained unchanged. Second, DNA cleavage induced by the presence of araC lesions was reversed by treatment with salt prior to protein denaturant (Fig. 1, open  bars). Salt reversibility is a hallmark of topoisomerase II-mediated DNA scission (32,54). Third, similar increases in DNA cleavage were observed when scission was monitored on either the wild-type oligonucleotide or the lesion-containing strand. This result implies that the observed DNA cleavage was double-stranded in nature. Taken together, these findings provide strong evidence that the observed DNA cleavage was generated by topoisomerase II␣ and was not due to a chemical degradation of lesion-containing oligonucleotides.
Since topoisomerase II␤ appears to be present throughout the cell cycle of all mammalian tissues, it also has the potential to interact with araC lesions in chromosomal DNA (32-34, 37, 55). Therefore, the effects of selected araC residues on the DNA cleavage activity of topoisomerase II␤ were characterized. As seen in Fig. 2, significant cleavage stimulation (similar to that seen with the ␣ isoform) was observed with the Ϫ3 lesion (top strand). Although cleavage enhancement also was observed with the ϩ2 lesion, it was considerably lower than that seen with topoisomerase II␣, and the effects of the ϩ3 lesion appeared to be negligible. As above, the points of DNA scission did not change in lesion-containing oligonucleotides, and cleavage was salt-reversible. Due to the greater effects of araC lesions on topoisomerase II␣ and the probable role of this enzyme in DNA replication, the ␣ isoform was used for further studies.
AraC Lesions Do Not Inhibit DNA Religation Mediated by Human Topoisomerase II␣-As defined by their ability to inhibit enzyme-mediated DNA religation, topoisomerase II poisons increase levels of enzyme-generated DNA breaks by one of two potential mechanisms (32,56). Whereas some poisons act primarily by impairing the ability of topoisomerase II to religate cleaved DNA molecules, others have little effect on this reaction and are presumed to act primarily by enhancing the formation of enzyme-DNA cleavage complexes.
Previous studies indicate that a variety of DNA lesions do not inhibit DNA religation and appear to act by this latter mechanism (24 -27, 40). Because of the altered positional specificity of araC residues compared with that of other lesions, the effects of this nucleoside analog on the DNA religation activity of human topoisomerase II␣ were characterized.
Rates of DNA religation were unaffected by incorporation of araC residues at the Ϫ3 or ϩ2 positions (Fig. 3). Therefore, it appears that araC lesions act primarily by enhancing the for-mation of topoisomerase II-DNA cleavage complexes. Furthermore, the mechanistic basis for the actions of araC on topoisomerase II␣ is independent of the location of the lesion relative to the scissile bonds.
Multiple AraC Lesions Induce Additive Levels of DNA Cleavage-Previous studies indicate that multiple abasic sites within a topoisomerase II DNA cleavage overhang induce levels of scission that are greater than those generated by the individual lesions (26,40). In some cases, the stimulatory effects were additive or even synergistic in nature.
Under clinically relevant conditions, ϳ100,000 araC residues/genome can be incorporated into human leukemic cells (1). Although it is unlikely that clusters of araC lesions will be observed routinely under these conditions, it is reasonable that two residues occasionally could be incorporated in close proximity to one another. Therefore, to determine whether multiple araC lesions also induce additive levels of DNA cleavage, the ability of human topoisomerase II␣ to cleave oligonucleotides that contained two stimulatory araC residues was characterized. Combinations that coupled the internal (relative to the DNA cleavage overhang) ϩ2 araC lesion with either the adja- cent ϩ3 lesion or the external Ϫ3 araC were examined (Fig. 4). Both combinations enhanced DNA scission in an additive fashion. As found for the individual lesions, cleavage induced by multiple araC residues was salt-reversible (not shown). Thus, levels of DNA cleavage mediated by human topoisomerase II␣ can be further stimulated by the incorporation of additional araC lesions proximal to the points of scission.
Etoposide and AraC Lesions Stimulate DNA Cleavage Mediated by Human Topoisomerase II␣ in an Additive or Synergistic Fashion-Chemotherapeutic regimens that combine araC with etoposide display synergistic effects in murine leukemia models and are used routinely for the treatment of patients with relapsed or refractory acute leukemias (1,57,58). Although abasic lesions dominate over etoposide and negate the enhancement of topoisomerase II-mediated DNA cleavage by the drug (not shown) (27), DNA distortions induced by the inclusion of an arabinose ring differ from those that accompany base loss (41)(42)(43)(44)(45)(46)(47)(48)(49)(50)(51)(52)(53). Therefore, the effects of etoposide on the ability of human topoisomerase II␣ to cleave oligonucleotides that contained incorporated araC residues were characterized.
As seen in Fig. 5, the DNA cleavage enhancement by etoposide and araC residues located within the cleavage overhang (at the ϩ2 or ϩ3 positions) was additive. Furthermore, when the lesion was located 5Ј to the scissile bond at the Ϫ3 position, synergistic DNA cleavage enhancement was observed. Levels of scission in the presence of etoposide and the Ϫ3 lesion were nearly three times higher than predicted based on simple additivity. The mechanistic basis for this unexpected finding is not known. However, it may be related to the usual ability of araC lesions to enhance topoisomerase II-mediated DNA scission when located external to the cleavage overhang.
In conclusion, araC is an important anticancer drug that must be incorporated into DNA in order to exert its cytotoxic effects. Cell lines with reduced levels of topoisomerase II␣ display resistance to araC (59), and pretreatment of cells with araC increases the concentration of DNA breaks induced by the topoisomerase II-targeted drugs, etoposide and amsacrine (31). Results of the present study indicate that incorporated araC residues are position-specific poisons of human type II topoisomerases. This work broadens the spectrum of DNA lesions known to stimulate topoisomerase II-mediated DNA cleavage and suggests potential roles for the type II enzyme in araC cytotoxicity or mutagenicity.