Active site mutations in DNA topoisomerase I distinguish the cytotoxic activities of camptothecin and the indolocarbazole, rebeccamycin.

DNA topoisomerase I (Top1p) catalyzes topological changes in DNA and is the cellular target of the antitumor agent camptothecin (CPT). Non-CPT drugs that target Top1p, such as indolocarbazoles, are under clinical development. However, whether the cytotoxicity of indolocarbazoles derives from Top1p poisoning remains unclear. To further investigate indolocarbazole mechanism, rebeccamycin R-3 activity was examined in vitro and in yeast. Using a series of Top1p mutants, where substitution of residues around the active site tyrosine has well-defined effects on enzyme catalysis, we show that catalytically active, CPT-resistant enzymes remain sensitive to R-3. This indolocarbazole did not inhibit yeast Top1p activity, yet was effective in stabilizing Top1p-DNA complexes. Similar results were obtained with human Top1p, when Ser or His were substituted for Asn-722. The mutations altered enzyme function and sensitivity to CPT, yet R-3 poisoning of Top1p was unaffected. Moreover, top1delta, rad52delta yeast cells expressing human Top1p, but not catalytically inactive Top1Y723Fp, were sensitive to R-3. These data support hTop1p as the cellular target of R-3 and indicate that distinct drug-enzyme interactions at the active site are required for efficient poisoning by R-3 or CPT. Furthermore, resistance to one poison may potentiate cell sensitivity to structurally distinct compounds that also target Top1p.

DNA topoisomerase I (Top1p) 1 constitutes the cellular target of the antineoplastic agent camptothecin (CPT) (reviewed in Refs. 1,2). Top1p catalyzes the relaxation of supercoiled DNA by transiently cleaving a single strand of duplex DNA. This is accomplished by the nucleophilic attack of the enzyme's active site tyrosine on a phosphodiester bond in the DNA backbone. Recent structural studies suggest that the covalent linkage of enzyme to the 3Ј-phosphoryl end of the nicked DNA allows for the rotation of the noncovalently held DNA end to rewind/ unwind the DNA (3)(4)(5). A second transesterification resolves the phosphotyrosyl linkage via religation of the nicked DNA.
The reversible stabilization of the covalent enzyme⅐DNA intermediate by CPT produces irreversible DNA lesions during S-phase, as a result of collisions between advancing replication forks and drug⅐Top1p⅐DNA complexes. In yeast, deletion of the gene encoding DNA topoisomerase I (TOP1) abolishes the cytotoxic action of the drug (reviewed in Refs. 2,6). Although CPT sensitivity is restored by the expression of plasmid encoded yeast or human TOP1, cells expressing CPT-resistant mutant enzymes are viable in the presence of the drug. These studies firmly establish DNA topoisomerase I as the cellular target of CPT. Additional factors have been shown to modulate yeast cell sensitivity to CPT, such as double-stranded DNA break repair (7,8), DNA damage cell cycle checkpoints (2), sister chromatin condensation (9), ATP binding cassette transporters (10), nuclear import of Top1p (11), and DNA replication (12). Nevertheless, DNA topoisomerase I poisoning is essential for the cytotoxic action of the drug.
In recent years, a number of structurally diverse compounds have been shown to interfere with the catalytic cycle of DNA topoisomerase I (reviewed in Refs. 13,14). Among these, indolocarbazoles exhibit antitumor activity, and several analogs, such as NB-506 (15) and BMS-250749 (16), are undergoing clinical development. Rebeccamycin derivatives, such as the fluoroindolocarbazole BMS-250749, NB-506, and R-3 (see Fig.  1), contain a glucose moiety appended to an indolocarbazole chromophore, which plays a critical role in DNA binding and Top1p poisoning (17,18). Modifications of the indolocarbazole moiety also affect drug binding to DNA. For example, R-3 and NB-506 exhibit an intercalative mode of DNA binding, enhanced by the sugar group. However, studies with a NB-506 derivative indicate that DNA intercalation is not required for effective DNA topoisomerase I poisoning in vitro (19).
Indolocarbazoles have been shown to act as Top1p poisons in vitro, yet the cytotoxic mechanism of these agents in vivo remains unclear (20,21). The collateral resistance of CPT-resistant cells to CPT and indolocarbazoles (R-3 and BMS-250749) provides evidence for DNA topoisomerase I as the cellular target (16,22). In these studies, the drug resistance of the P388CPT45 cells is attributed to lower levels of DNA topoisomerase I. Furthermore, based on the cross-resistance of human DNA topoisomerase I mutants to CPT and R-3 in vitro, it has been argued that these topoisomerase I poisons share a common pharmacophore (23). In contrast, the partial resistance of CPT-resistant cell lines with known mutations in TOP1 to NB-506 and a nonintercalating derivative argue for cellular targets in addition to Top1p (21). In these studies, the levels of CPT and indolocarbazole-stabilized Top1p⅐DNA complexes were reduced in crude nuclear extracts; however, this did not correspond to a proportional decrease in the cytotoxic activity of the drugs. Although the CPT-resistant cell lines were compared with a parental cell line; the acquisition of additional genetic alterations during selection of the CPT-resistant derivative remains a possibility. A direct demonstration that DNA topoisomerase I constitutes the cellular target of rebeccamycin indolocarbazoles is further complicated by the lack of TOP1 null cell lines. In contrast to the viability of top1⌬ yeast, top1Ϫ/Ϫ mice exhibit early embryonic lethality (24).
To directly address the mechanism of indolocarbazole action, the effects of DNA topoisomerase I active site mutations on the activity of the rebeccamycin R-3 derivative were examined in vitro and in isogenic yeast strains. We report here that yeast and human DNA topoisomerase I mutants bearing substituents of the Asn residue immediately N-terminal to the active site tyrosine (25,26) remain sensitive to R-3, despite dramatic alterations in sensitivity to CPT and catalytic activity. Results obtained with top1⌬ yeast cells expressing wild-type human Top1 and the inactive hTop1Y723F protein demonstrate that transesterification catalytic activity is required for R-3 cytotoxicity and further support human DNA topoisomerase I as the cellular target of this indolocarbazole. These studies also indicate distinct drug⅐enzyme interactions at the active site are required for efficient poisoning by the R-3 indolocarbazole or CPT and suggest that resistance to one class of DNA topoisomerase I poisons may enhance cell sensitivity to structurally distinct compounds that also target Top1p.

EXPERIMENTAL PROCEDURES
Chemicals, Yeast Strains, and Plasmids-The chemical synthesis of the rebeccamycin derivative R-3 has previously been reported (27). CPT (Sigma Chemical Co., St. Louis, MO) and R-3 were dissolved in Me 2 SO and stored at Ϫ20°C. [␣-32 P]Cordycepin 5Ј-triphosphate was purchased from PerkinElmer Life Sciences (Boston, MA).
Yeast transformations were performed by treatment with lithium acetate as described (33). The GAL1 promoted expression vectors carry the URA3 selectable marker and are maintained by selection in synthetic complete (SC)-uracil medium.
DNA Topoisomerase I Purification and Activity-DNA topoisomerase I was purified from galactose-induced cultures of EKY3 cells transformed with various TOP1 alleles under the GAL1 promoter as described (25,34). Briefly, proteins precipitated with 75% ammonium sulfate were applied to a phosphocellulose column equilibrated with TEEG buffer (50 mM Tris-HCl (pH 7.4), 1 mM EDTA, 1 mM EGTA, and 10% glycerol) containing 0.2 M KCl. The proteins were eluted with successive washes of TEEG buffer containing 0.4, 0.6, 0.8, and 1 M KCl. Top1 proteins from peak fractions (0.6 M KCl for the yeast enzymes, 0.8 M KCl for human) were diluted with 50% glycerol for storage at Ϫ20°C.
Protein integrity and relative concentrations were assayed by immunostaining with yeast or human Top1p-specific antibodies.
DNA topoisomerase I activity was assayed in plasmid DNA relaxation reactions as described (25). Serial 10-fold dilutions of Top1 proteins, corrected for concentration, were incubated with 0.3 g of negatively supercoiled plasmid pHC624 DNA in 20 mM Tris-HCl (pH 7.5), 10 mM MgCl 2 , 0.1 mM EDTA, 150 mM KCl, and 50 g/ml gelatin for 30 min at 30°C (yeast enzymes) or 37°C (human enzymes). The extent of plasmid DNA relaxation was assessed by agarose gel electrophoresis followed by staining with ethidium bromide and photography.
DNA Cleavage Reactions-Wild-type and mutant enzyme sensitivity to CPT and R-3 was determined in DNA cleavage assays (25,35). A 480-bp DNA substrate, 32 P-labeled at a single 3Ј-end, was prepared as described. Equal concentrations of the purified yeast or human Top1 proteins were incubated with ϳ10 ng (10,000 cpm) of DNA in 20 mM Tris-HCl (pH 7.5), 10 mM MgCl 2 , 0.1 mM EDTA, 50 mM KCl, 50 g/ml gelatin, and the indicated concentrations of CPT or R-3 in a final 4% Me 2 SO. Following incubation for 30 min at 30°C (for yeast Top1p) or 37°C (for hTop1p) for 30 min, the reactions were terminated with 1% SDS and heating to 75°C. The samples were treated with proteinase K and ethanol-precipitated prior to electrophoresis in an 8% polyacrylamide/7 M urea gel. The cleavage products were visualized with a PhosphorImager (Molecular Dynamics, Sunnyvale, CA).
Suicide Cleavage Assays-Relative rates of Top1p-catalyzed DNA cleavage were assessed using a suicide DNA substrate that contains a high affinity cleavage site (25,36). The scissile strand 5Ј-GATCTAAAA-GAC-TT2GGAA-3Ј (where 2 denotes the site of Top1p cleavage) was 3Ј-end-labeled with [ 32 P]cordycepin using terminal deoxynucleotidetransferase. After removal of unincorporated nucleotides, the labeled stand was annealed to the complementary unlabeled strand 5Ј-GATCTTTTTTAAAAATTTTTCCAAGTCTTTTAGATC-3Ј in annealing buffer (10 mM Tris (pH 7.8), 100 mM NaCl, 1 mM EDTA) by heating to 90°C followed by cooling to room temperature over 3 h. Reactions were performed at 25°C by incubating equal concentrations of protein with 20 fmol of DNA substrate in 20 mM Tris-HCl (pH 7.5), 10 mM MgCl 2 , 0.1 mM EDTA, 50 mM KCl, 50 g/ml gelatin, 4% Me 2 SO and, when indicated, 50 M CPT or 5 M R-3. At various times, 10-l aliquots were removed and quenched with 0.5% SDS at 75°C, followed by the addition of loading buffer (95% formamide, 1 mg/ml bromphenol blue, 1 mg/ml xylene cyanol) and electrophoresis in 20% polyacrylamide/7 M urea gels. Cleavage products were visualized using a PhosphorImager.
Drug Sensitivity Assays-To determine cellular sensitivity to CPT or R-3, isogenic top1⌬ yeast strains that were RAD52 (EKY3) or rad52⌬ (MBY3) were transformed with the indicated YCpGAL1-TOP1 or YCp-GAL1-hTOP1 constructs and selected on SC-uracil, dextrose medium. Exponentially growing cultures of individual transformants were adjusted to A 595 ϭ 0.3, serially 10-fold diluted and spotted onto SC-uracil plates containing 25 mM HEPES (pH 7.2), 2% dextrose or galactose, and the indicated concentrations of CPT or R-3. All plates contained a final 0.25% Me 2 SO concentration. Cell viability was determined following incubation at 30°C for 3 days.
DNA Unwinding Assays-R-3 inhibition of Top1p catalyzed relaxation of plasmid DNA was examined under two experimental conditions. First, 0.3 g of negatively supercoiled plasmid DNA and increasing concentrations of R-3 were combined in reaction buffer (20 mM Tris-HCl (pH 7.5), 10 mM MgCl 2 , 0.1 mM EDTA, 150 mM KCl, and 50 g/ml gelatin) prior to enzyme addition. The reactions were then incubated at 30°C for 30 min and terminated with 1% SDS. In a second set of reactions, supercoiled plasmid DNA was treated with equivalent units of Top1p or Top1N726Dp in the absence of drug to allow for complete relaxation of the plasmid DNA. Increasing concentrations of R-3 were then added for an additional 30 min at 30°C. The reactions were terminated with SDS, and the drug was extracted with phenol. DNA topoisomer products were resolved by agarose gel electrophoresis, stained with ethidium bromide, and photographed.

RESULTS
Yeast DNA Topoisomerase I Is Sensitive to R-3-Substitutions of residues N-terminal to the active site tyrosine (Tyr-727) of yeast DNA topoisomerase I have profound effects on the stability of the Top1p⅐DNA covalent complex and enzyme sensitivity to CPT (see Refs. 25,26,37 and Table I). For instance, mutation of Asn-726 to Ser (in Top1N726Sp) renders the enzyme resistant to CPT, with no detectable effect on enzyme catalysis. Other substitutions, Thr-722 to Ala (T722A), Asn-726 to Asp (N726D), and Asn-726 to His (N726H), adversely affect the catalytic cycle of Top1p and induce cell lethality when overexpressed in the absence of drug (25). This results from increased stabilization of covalent mutant Top1p⅐DNA complexes: Top1T722Ap mimics CPT in diminishing the rate of DNA religation, whereas Top1N726Hp exhibits enhanced rates of DNA scission (25). 2 The DNA binding defect of Top1N726Dp also appears to generate irreversible DNA lesions, although the mechanism by which this is accomplished is unclear. The distinct alterations in enzyme catalysis and CPT sensitivity exhibited by this family of Top1p mutants present a unique opportunity to investigate the cytotoxic mechanism of structurally diverse Top1p poisons.
To examine the effect of active site mutations on Top1p sensitivity to the rebeccamycin R-3 derivative (Fig. 1), the mutant enzymes were purified, corrected for concentration, and assayed for CPT and R-3-induced DNA cleavage. As reported (25), the specific activities of Top1N726Sp and Top1N726Hp were equivalent to wild-type Top1p ( Fig. 2A). Top1T722Ap activity was diminished about 2-fold, whereas the relaxation activity of Top1N726Dp was reduced by ϳ50-fold. Top1Y727Fp, in which the active site tyrosine is mutated to Phe, was inactive.
To assess the effects of single amino acid substituents on enzyme sensitivity to Top1p poisons, the proteins were incubated with a 3Ј-end-labeled DNA fragment in the presence or absence of R-3 or CPT. Following incubation at 30°C for 30 min, the covalent enzyme⅐DNA intermediates were trapped by SDS, treated with proteinase K, and resolved in sequencing gels. As shown in Fig. 2B and Table I, R-3 proved a potent enhancer of covalent complex formation with each of the catalytically active yeast enzymes tested. The extent of DNA cleavage obtained with wild-type Top1p in the presence of 50 M CPT was equivalent to that induced by 5 M R-3 (compare Top1 lanes C and R, Fig. 2B). Although CPT and R-3 stabilized cleavable complexes at unique sites within the DNA substrate, the majority of strong cleavage sites were common to both. This held true for each enzyme examined. In contrast, although the extent of CPT-induced DNA cleavage by the mutant enzymes varied considerably, R-3 stabilized complexes were largely unaffected by these substitutions.
CPT and R-3 induced comparable levels of DNA cleavage by the lethal mutant proteins, Top1T722Ap and Top1N726Hp.
The residues immediately N-terminal to the active site tyrosine (Tyr-727 in yeast Top1p and Tyr-723 in human Top1p) are shown. Single amino acid substitutions encoded by the top1 or htop1 mutant alleles in the YCpGAL1 vectors are indicated in boldface. b Mutant enzyme sensitivity to CPT or R-3, relative to the respective wild-type enzyme in DNA cleavage assays, are shown. Average band intensity for each wild-type enzyme in the presence of CPT (50 M) or R-3 (5 M) was arbitrarily set at ϩϩ. Each ϩ indicates an average ϳ5to 10-fold increment in drug-stabilized complexes at multiple sites.
These mutants exhibited enhanced covalent complexes in the absence of drug, yet, the patterns of DNA cleavage reverted to those of wild-type Top1p in the presence of CPT or R-3 (Fig.  2B). As reported (25), Top1N726Dp and Top1N726Sp were CPT-resistant, with decreased levels of DNA cleavage in the presence of the drug. Remarkably, however, these enzymes remained sensitive to R-3. In the case of the Asn-726 to Asp substitution (N726D), there was no significant increment in cleavable complexes upon addition of CPT (Fig. 2B, compare N726D, lanes Ϫ and C). In contrast, R-3 did enhance covalent complex formation (see lane R), albeit at lower levels than that observed with equal concentrations of wild-type Top1p. This differential in N726D versus Top1p⅐DNA complexes is consistent with the diminished catalytic activity of Top1N726Dp apparent in Fig. 2A. Quantitation of band intensities in drugtreated Top1p and Top1N726Dp samples indicated that R-3-

FIG. 2. Mutation of conserved residues in yeast Top1p have distinct effects on enzyme sensitivity to camptothecin and rebeccamycin.
A, equal concentrations of yeast DNA topoisomerase I (Top1p) and mutant Top1N726D, Top1N726H, Top1N726S, Top1T722A, and Top1Y727F proteins were serially 10-fold diluted and incubated in a plasmid DNA relaxation assay (see "Experimental Procedures"). Top1Y727Fp was only assayed at higher concentrations. After 30 min at 30°C, the reactions were terminated with 1% SDS, and the extent of plasmid DNA relaxation assessed by agarose gel electrophoresis, followed by ethidium bromide staining. The relative positions of relaxed (R) and supercoiled (Sc) DNA topoisomers are indicated. B, equal concentrations of the same proteins were incubated with a 3Ј-32 P-end-labeled DNA substrate in a DNA cleavage reaction (see "Experimental Procedures"). Stabilization of covalent protein⅐DNA complexes was assayed in the absence of drug (Ϫ), and in the presence of 50 M CPT (C) or 5 M R-3 (R). Following incubation at 30°C for 20 min, covalent DNA⅐protein complexes were trapped with SDS and digested with proteinase K. The reaction products were denatured, resolved in a 7 M urea/8% polyacrylamide gel and visualized using a PhosphorImager. The leftmost lane (D) is substrate DNA without enzyme. The position of the high affinity cleavage site is indicated with an asterisk.
induced DNA cleavage was proportional to enzyme specific activity under the same conditions (data not shown). This is in contrast with the greater reduction in CPT-induced complexes, reflecting the intrinsic CPT resistance conferred on DNA topoisomerase I by the N726D substitution.
When Ser was substituted for Asn-726 (N726S), there were no detectable alterations in enzyme activity, other than a reduction in CPT sensitivity (25) (in Fig. 2B, compare Top1 and N726S, lane C). This is in contrast with the efficient stabilization of covalent Top1N726Sp⅐DNA complexes by R-3 (lane R). Thus, mutations of Asn-726 that render Top1p resistant to CPT did not compromise enzyme sensitivity to R-3.
Rebeccamycin Does Not Inhibit Yeast Top1p Catalytic Activity-As rebeccamycin R-3 intercalates into DNA, alterations in DNA structure induced by drug binding might affect enzyme activity and contribute to covalent complex formation. Other drugs exhibiting an intercalative mode of DNA binding, such as saintopin, have been shown to inhibit DNA topoisomerase I catalytic activity (34), although such effects are observed at concentrations greater than those needed to stabilize the covalent complex. High concentrations of CPT also inhibit human DNA topoisomerase I activity, yet the relaxation of plasmid DNA by yeast Top1p is unaffected by the drug (34). Because R-3 has been reported to inhibit human Top1p-catalyzed relaxation of plasmid DNA (23), the effects of increasing concentrations of R-3 on yeast Top1p and Top1N726Dp activity were assessed.
For a direct comparison of drug action on wild-type and mutant enzyme activity, Top1p and Top1N726Dp were corrected for specific activity and two experimental strategies were followed. First, negatively supercoiled plasmid DNA was mixed with increasing R-3 concentrations prior to the addition of enzyme. As shown in Fig. 3A, this produced similar dose-dependent alterations in plasmid linking number. The shift in mobility from relaxed topoisomers at 0 -5 M R-3 to the supercoiled molecules observed at higher R-3 concentrations could reflect an inhibition of enzyme activity. Alternatively, unwinding of closed circular DNA upon intercalation of R-3 would induce positive supercoils. DNA topoisomerase I catalyzed relaxation of these supercoils, followed by drug removal prior to electrophoresis, would result in the accumulation of negatively supercoiled DNA. In either case, the net effect would be the same, a dose-dependent shift in plasmid mobility from that of relaxed topoisomers to that of negatively supercoiled DNA.
To distinguish between inhibitory effects on catalytic activity and drug intercalation, negatively supercoiled DNA was first incubated with Top1p or Top1N726Dp to ensure complete relaxation of the plasmid DNA. Increasing concentrations of R-3 were then added, and the reactions were incubated for an additional 30 min. Under these conditions, intercalation of R-3 into the relaxed topoisomers would still induce positive supercoiling. However, if enzyme activity were inhibited, there would be little or no alteration in plasmid linking number; the plasmids would still migrate as relaxed topoisomers. In contrast, if enzyme activity were unaffected, the positive supercoils would be relaxed and the distribution of topoisomers in the gel would resemble that obtained in panel A. As shown in Fig. 3B, the latter was observed. Thus, as with CPT, R-3 did not inhibit wild-type yeast Top1p-or Top1N726Dp-catalyzed relaxation of supercoiled DNA.
The CPT-resistant Top1N726S Mutant Enzyme Exhibits Enhanced Sensitivity to R-3-To further investigate the sensitivity of CPT-resistant Top1p mutants to R-3, DNA cleavage was examined over a range of drug concentrations. To avoid alterations in catalytic activity, Top1p and Top1N726Sp were used. In Fig. 4, wild-type Top1p displayed a dose-dependent increase in cleavage complex formation in the presence of R-3 and CPT. Top1N726Sp was highly resistant to CPT: The level of cleavage stimulated by 50 M was comparable to that of Top1p at 1 M. Surprisingly, substituting Ser for Asn-726 not only failed to confer cross-resistance to R-3, but quantitation of cleavage products revealed a stimulation of Top1N726Sp⅐DNA cleavage by R-3. With the exception of the high affinity site, the magnitude of this increase varied between cleavage sites, ranging from 2-to 6-fold over that obtained with Top1p (data not shown).
CPT stabilizes the enzyme⅐DNA intermediate by reversibly inhibiting the DNA religation step of the catalytic cycle. However, enhanced DNA cleavage would also increase covalent complex steady-state levels. To determine if the N726S mutation altered the mechanism of R-3 action, the effects of CPT and R-3 on the rate of covalent complex formation were evaluated using a suicide DNA substrate (Fig. 5A). Cleavage of the scissile strand at the high affinity cleavage site (indicated by the arrow) liberates a 5-mer, labeled at the 3Ј-end. Because this short segment of DNA is not covalently attached to the enzyme, it dissociates from the covalent complex. Without a 5Ј-OH nucleophile to resolve the covalent intermediate, the cleavage and religation reactions are uncoupled and the relative rates of DNA scission may be approximated by the accumulation of the labeled 5-mer (arrowhead in panels B and C).
The addition of CPT or R-3 had no detectable effect on the rate of DNA cleavage by wild-type Top1 or Top1N726S enzymes (Fig. 5B). For comparison, assays were also performed with Top1N726Hp, where substitution of His for Asn-726 enhances the rate of Top1p-catalyzed cleavage (Fig. 5C). Neither agent enhanced the rate of formation of enzyme⅐DNA complexes, consistent with an inhibition of DNA religation. Indeed, in experiments with longer scissile strands, where a significant level of religation could be detected, CPT and R-3 inhibited wild-type Top1p-catalyzed resolution of the covalent com- plexes. In contrast, R-3 inhibited Top1N726Sp religation, whereas CPT had no effect (data not shown).

R-3 Poisons DNA Topoisomerase I in Yeast-
To determine if the R-3 hypersensitivity of the CPT-resistant Top1N726S enzyme translates into enhanced cytotoxicity, yeast strains deleted for TOP1 (top1⌬) were transformed with GAL1-promoted TOP1, top1N726S, and top1Y727F vectors and assayed for CPT or R-3 sensitivity. In all cases, cell viability was unaffected by CPT or R-3 on dextrose-containing media, where the GAL1 promoter is repressed ( Fig. 6 and see Fig. 8 below, data not shown). Galactose-induced expression of wild-type yeast TOP1 was sufficient to restore cell sensitivity to CPT, whereas cells expressing the CPT-resistant mutant enzyme, Top1N726Sp, or the catalytically inactive Top1Y727Fp, were viable in the presence of the drug (Fig. 6, upper panel). However, cells expressing TOP1 or top1N726S were insensitive to R-3.
The permeability of yeast cells to a wide range of drugs is a significant issue in the study of drug action (6). The pleiotropic drug resistance network of ATP binding cassette transporters has been shown to modulate cell sensitivity to CPT through the action of the Pdr1 transcriptional activator and the Snq2 transporter (10). Other permeability mutants, such as ise1 and ISE2, also exhibit greater sensitivity to drugs that target DNA topoisomerase I or II (7,38). However, none of these factors affected cell sensitivity to R-3 (data not shown), suggesting distinct mechanisms of drug uptake/efflux from CPT.
Yeast strains defective in homologous recombination, due to deletion of RAD52 or other members of the RAD52 epistasis group, exhibit enhanced sensitivity to CPT-induced Top1p⅐DNA lesions (6). Thus, rad52⌬ cells, expressing yeast or human TOP1, die at lower concentrations of CPT than isogenic, repair-proficient RAD52 strains. In Fig. 6, rad52⌬ cells expressing wild-type TOP1 were more sensitive to CPT than the isogenic RAD52 cells. Even the ϳ50-fold lower CPT sensitivity of Top1N726Sp (Fig. 2B) was able to elicit a 10-to 20-fold drop in rad52⌬ cell viability in the presence of the drug. With R-3, expression of wild-type Top1p had no effect. However, the hypersensitive Top1N726S enzyme produced an R-3-sensitive phenotype, with an ϳ2-log drop in cell viability. The increment in R-3-mediated DNA cleavage varied considerably from site to site in the DNA substrate used in vitro (see Figs. 2B and 4), complicating the extrapolation from in vitro assays to cell kill. Nevertheless, these data indicate that expression of an R-3sensitive DNA topoisomerase I is sufficient for the cytotoxic activity of R-3 in a rad52⌬ yeast strain.
Expression of Human DNA Topoisomerase I Sensitizes top1⌬ Yeast Cells to R-3-Eukaryotic DNA topoisomerase I is highly conserved in terms of reaction mechanism, structure, and drug FIG. 4. The camptothecin-resistant yeast Top1N726S protein exhibits enhanced sensitivity to R-3. As described in Fig. 1, equal concentrations of yeast Top1 and Top1N726S proteins were incubated in a DNA cleavage assay with the indicated concentrations of CPT or R-3, in a final 4% Me 2 SO concentration. Lane D contains substrate DNA alone. The reaction products were resolved in a 7 M urea/8% polyacrylamide gel and visualized by phosphorimaging. sensitivity (3,6,39). Expression of human TOP1 (hTOP1) is sufficient to restore top1⌬ yeast cell sensitivity to CPT. Furthermore, substitution of conserved residues have similar effects on enzyme activity and drug action. As with yeast Top1T722Ap, mutation of the corresponding residue in human Top1p (Thr-718 to Ala) produces similar effects on enzyme catalysis (37). Substitutions of other conserved residues immediately N-terminal to the active site tyrosine also produce similar effects on enzyme sensitivity to CPT (34,40). However, despite the overriding similarity in enzyme structure and function, differences do exist. For example, yeast cells expressing the human enzyme are more sensitive to CPT than cells expressing comparable levels of the yeast enzyme (30). Furthermore, the human Top1N722S enzyme in extracts of CPT-resistant CEM/C2 cells was reported to be resistant to another indolocarbazole, NB-506 (21), whereas mutation of residue 361 in human Top1p conferred resistance to both R-3 and CPT (23).
To determine whether the effects of active site mutations in Top1p on R-3 sensitivity were specific to the yeast enzyme or reflected structural features particular to R-3 interactions within the active site of Top1p, the same substitutions of Asn-722 were made in GAL1-promoted hTOP1 constructs (see Table  I). The mutant proteins were purified from top1⌬ yeast cells, corrected for protein concentration, and assayed in DNA relaxation and cleavage assays. As with yeast top1N726H, replacing Asn-722 with His in hTop1N722Hp enhanced the stability of the covalent complex in the presence and absence of CPT (Fig.  7, hN722H, and Table I). The htop1N722H mutant also induced yeast cell lethality in the absence of CPT (data not shown). However, in contrast to the yeast mutant, the His substituent in human hTop1N722Hp enhanced enzyme sensitivity to R-3-induced DNA cleavage (compare hTop1 and hN722H, lane R, in Fig. 7). As reported (41), replacing Asn with Ser (in hTop1N722Sp) reduced enzyme sensitivity to CPT, without affecting enzyme activity (Fig. 7, data not shown). However, the CPT resistance of the yeast Top1-Ser mutant was greater than that observed for the human Ser mutant enzyme. Yet, despite differences in wild-type and mutant enzyme sensitivity to CPT, both exhibited comparable levels of R-3-stabilized covalent complexes (compare hTop1 and hN722S, lane R). As with the yeast Top1-Ser mutant, resistance to CPT did not correspond to cross-resistance to R-3.
These effects on human Top1p function were also manifest in isogenic RAD52 and rad52⌬ yeast strains. Expression of hTOP1 restored yeast RAD52, top1⌬ cell sensitivity to CPT (see Fig. 8). Cells expressing htop1N722S remained sensitive to high concentrations of CPT; however, a resistant phenotype was observed at lower drug concentrations. As with yeast Top1p, the repair-proficient RAD52 cells were resistant to R-3 irrespective of human DNA topoisomerase I. In the rad52⌬ strain, expression of the human Ser mutant restored CPT sensitivity at low drug doses. However, even in the absence of drug, galactose-induced expression of wild-type hTOP1 compromised rad52⌬ cell viability as evidenced by a 10-fold drop in cell number relative to cells expressing the Ser mutant (htop1N722S) or the catalytically inactive htop1Y723F mutant. In this genetic background, expression of catalytically active human Top1p or the Ser mutant enzyme rendered the cells sensitive to high concentrations of R-3. The 10-fold decrease in viable cells expressing hTOP1 versus htop1N722S across the entire range of R-3 concentrations tested corresponded to the 10-fold decrease in cell viability attendant with hTOP1 expression in the absence of drug.
To more directly assess the requirement of human Top1p transesterification catalytic activity for the cytotoxic activity of R-3, exponentially growing cultures of rad52, top1⌬ cells expressing wild-type human Top1p or the catalytically inactive Top1Y723F protein were treated with R-3 or no drug. As shown in Fig. 9, the viability of cells expressing htop1Y723F was unaffected by R-3. In contrast, drug treatment induced a significant drop in the number of viable cells expressing wild-type hTOP1, with a greater than 10-fold decrease apparent by 24 h. Consistent with the plating experiments shown in Fig. 8, these data indicate that human TOP1 may also provide a cellular target for R-3 in yeast and that the cleavage/religation activity of the enzyme is essential for the cytotoxic action of R-3 and CPT. Because the active site tyrosine is dispensable for the purported kinase activity of human Top1p (22), the cytotoxic activity of R-3 appears unrelated to any inhibitory effects on kinase activity. These results further establish that, although substitutions near the active site tyrosine abolish enzyme sensitivity to CPT, these mutations do not confer cross-resistance to the rebeccamycin derivative, R-3. DISCUSSION DNA topoisomerase I has emerged as an important cellular target in the development of novel chemotherapeutics, including CPT analogs (topotecan and CPT-11) and rebeccamycin derivatives (BMS-250749, NB-506, J-107088, and R-3) (1,6,13,14,16,42). Top1p catalyzes changes in DNA topology via the formation of a phosphotyrosyl linkage between the active site tyrosine and the 3Ј-end of the nicked single strand of DNA (39,43). This transient, covalent intermediate is an obligate step in the unwinding or rewinding of duplex DNA. However, it also poses a threat to cell survival when, for example, advancing replication forks collide with CPT-stabilized Top1p⅐DNA complexes. This produces potentially lethal DNA lesions that induce cell cycle arrest and death. Studies in mammalian and yeast cells have firmly established that the cytotoxic activity of camptothecins derive from their ability to interfere with the cycle of DNA breakage and rejoining catalyzed by DNA topoisomerase I (1,2,6,39). Although rebeccamycin derivatives act as Top1p poisons to stabilize covalent enzyme⅐DNA complexes in vitro, the antiproliferative mechanism of these compounds is unclear. Mammalian cell lines deficient in DNA topoisomerase I or expressing CPT-resistant enzymes argue for Top1p as the cellular target of BMS-250749 and R-3 (16,22), whereas related studies with NB-506 and a nonintercalating derivative suggest cellular targets in addition to Top1p (21). These contrary findings may reflect the functional consequences of different indolocarbazole substituents (as shown for NB-506 and R-3 in Fig. 1) and/or result from other genetic alterations acquired during the selection of CPT-resistant cell lines. In either case, the lack of isogenic top1Ϫ/Ϫ and TOP1ϩ/ϩ mammalian cells complicates a direct determination of the antiproliferative effects of Top1p poisoning by these agents.
Using a yeast strain deleted for TOP1 (top1⌬) and defective for recombinational repair of double-strand DNA breaks (due to deletion of RAD52), we were able to show that expression of a R-3-sensitive yeast or human DNA topoisomerase I was sufficient to confer R-3 cytotoxicity. As with CPT (2, 6), the R-3 sensitivity of these cells was dependent on the expression of a catalytically active enzyme and correlated with the relative levels of drug-stabilized complexes observed in vitro. For example, with yeast DNA topoisomerase I, only cells expressing the CPT-resistant Top1N726S enzyme were sensitive to R-3, consistent with the enhanced levels of R-3-induced DNA cleavage obtained in vitro, relative to wild-type yeast Top1p. Furthermore, because R-3 did not inhibit the catalytic activity of the yeast enzyme, and TOP1 is nonessential in yeast, the cytotoxic activities of CPT and R-3 cannot be attributed to differences in DNA binding.
Cells expressing wild-type hTop1p or the CPT-resistant human Top1N722S enzyme exhibited a similar R-3 dose-dependent decrease in cell viability, whereas cells expressing comparable levels of the catalytically inactive hTop1Y723F protein were resistant to the drug (Figs. 8 and 9). Although, the indolocarbazole derivatives, R-3 and NB-506, have been reported to inhibit the kinase activity of human Top1p (22,44), the active site tyrosine, Tyr-723, is dispensable for this activity (22).
Thus, as with camptothecins, these results demonstrate that the transesterification reactions required for DNA cleavage and religation by hTop1p are also required for the antiproliferative activity of R-3.
In the crystal structures of human DNA topoisomerase I and DNA (3)(4)(5), the two "Lip" domains form a salt bridge to close the protein clamp around duplex DNA on the opposite face of the DNA from the active site tyrosine. Mutation of residue 361 or adjacent residues in the upper Lip domains of hTop1p was previously shown to confer resistance to both R-3 and CPT (23). These data suggest that CPT and rebeccamycin share a common pharmacophore. However, this is in contrast to our findings that mutations around the active site tyrosine, which abrogate yeast or human Top1p sensitivity to CPT, did not confer cross-resistance to R-3. Indeed, in yeast Top1N726Sp, where substitution of Ser for Asn-726 produces about a 50-fold decrease in Top1p sensitivity to CPT, there was a 2-to 6-fold enhancement of R-3-induced DNA cleavage. The same substitution of the corresponding residue in the human enzyme, Top1N722Sp, also failed to confer cross-resistance to CPT and R-3, yet replacing this residue with Ser or His produced similar effects on yeast and human enzyme activity (Ref. 25, Table I, Figs. 2, 5, 7, and data not shown). This corresponded to similar alterations in the cytotoxic activity of CPT and R-3 in yeast cells expressing yeast top1N726S or human top1N722S (Figs. 6 and 8). Taken together, these data indicate a conservation in the active site architecture of the yeast and human enzymes that distinguishes enzyme⅐DNA interactions with CPT from R-3. Although CPT and rebeccamycin may share common structural elements that govern interactions with enzyme⅐DNA complexes around the Lip domains, distinct structural elements of the enzyme appear to dictate drug interactions within the active site.
Several lines of evidence also suggest that specific substituents in the indolocarbazole moiety of rebeccamycin can affect Top1p poisoning. For example, P388/CPT45 cells are crossresistant to CPT and the rebeccamycin derivatives R-3 and BMS-250749 (16,22), yet they exhibit limited cross-resistance to NB-506 (21). Indeed, the lack of detectable NB-506-induced DNA cleavage in crude extracts of several CPT-resistant cell lines was at odds with the limited resistance of these cell lines to NB-506, suggesting additional cellular targets of this rebeccamycin derivative (21). However, one of the cell lines used in these studies expresses the human Top1N722S mutant enzyme, which is cross-resistant to NB-506 (21), yet, in our hands, remains sensitive to R-3.
Because NB-506-induced DNA cleavage was assessed in crude nuclear extracts (21), the possibility remains that additional genetic alterations acquired during selection for CPT resistance could also contribute to differences in enzyme activity and/or cell sensitivity to the rebeccamycin analogs used in the studies. By way of example, in yeast, deletion of RAD52 FIG. 8. Yeast cells expressing human top1N722S are resistant to camptothecin, yet remain sensitive to R-3. Isogenic RAD52 (EKY3) or rad52⌬ (MBY3) yeast strains were transformed with ARS/CEN vectors expressing the indicated human TOP1 cDNA (hTOP1, htop1N722S, or htop1Y723F) from the GAL1 promoter. Individual transformants were grown in selective media, serially 10-fold diluted, and spotted onto SC-uracil plates supplemented with 25 mM HEPES (pH 7.2), 2% galactose, or dextrose and either Me 2 SO alone, CPT (0.1 or 5 M), or R-3 (1, 5, or 25 M). All plates contained 0.25% Me 2 SO.

FIG. 9. Yeast cell sensitivity to R-3 is dependent on the expression of catalytically active human DNA topoisomerase I.
Exponentially growing cultures of top1⌬, rad52⌬ (MBY3) yeast cells, galactose-induced to express plasmid borne GAL1-hTOP1 (hTOP1) or GAL1-htop1Y723F (hY723F) constructs, were treated with 1% Me 2 SO or 50 M R-3 at t ϭ 0. At the times indicated, aliquots were serially diluted and plated onto selective media containing dextrose. The number of cells forming colonies was assessed at 30°C and plotted relative to the number obtained at t ϭ 0. The data were obtained from four experiments.
increased cell sensitivity to CPT and R-3 (Ref. 6,and Figs. 6 and 8). In contrast, disruption of the pleiotropic drug resistance pathway by deleting PDR1 or deletion of the ABC transporter gene, SNQ2, both enhanced yeast cell sensitivity to CPT (6,10) yet had no detectable effect on R-3 cytotoxicity (data not shown). Along these lines, it would be of interest to determine whether the related human BCRP transporter, which induces cellular resistance to topotecan (45), is also able to transport specific rebeccamycin derivatives.
However, we favor an alternative explanation, where the discrepancy in hTop1N722S enzyme sensitivity to NB-506 and R-3 is a consequence of indolocarbazole modifications (diagrammed in Fig. 1) that alter the interaction of these rebeccamycin analogs with Top1p. For example, R-3 lacks the phenolic OH groups present at positions 1 and 11 on the indolocarbazole chromophore of NB-506, which play a major role in DNA binding, Top1p poisoning, and the cytotoxic activity of this agent (46,47). Whether amino acid substitutions in Top1p affect enzyme sensitivity to Top1p poisons by altering drug binding or indirectly through perturbations in protein structure will be resolved as the crystal structures of Top1p⅐DNA⅐drug complexes become available. Nevertheless, the data suggest that different substituents within the indolocarbazole moiety of rebeccamycin can affect the interaction of these drugs with active site residues and alter the ability of these agents to stabilize the covalent Top1p⅐DNA complex. This interpretation is further consistent with our findings that support Top1p as the cellular target of R-3 and raises the possibility that resistance to one class of DNA topoisomerase I poisons could potentiate the cytotoxic activity of structurally distinct agents that also target Top1p.