Nuclear Overexpression of NAD(P)H:Quinone Oxidoreductase 1 in Chinese Hamster Ovary Cells Increases the Cytotoxicity of Mitomycin C under Aerobic and Hypoxic Conditions*

The effects of the subcellular localization of overexpressed bioreductive enzyme NAD(P)H:quinone oxidoreductase 1 (NQO1) on the activity of the antineoplastic agent mitomycin C (MC) under aerobic and hypoxic conditions were examined. Chinese hamster ovary (CHO-K1/ dhfr (cid:1) ) cells were transfected with NQO1 cDNA to produce cells that overexpressed NQO1 activity in the nucleus (148-fold) or the cytosol (163-fold) over the constitutive level of the enzyme in parental cells. Subcellular localization of the enzyme was confirmed using antibody-assisted immunofluorescence. Nuclear localization of transfected NQO1 activity increased the cytotoxicity of MC over that produced by overexpression in the cytosol under both aerobic and hypoxic conditions, with greater cytotoxicity being produced under hypoxia. The greater cytotoxicity of nuclear localized NQO1 was not attributable to greater metabolic activation of MC but instead was the result of activation of the drug in close proximity to its target, nuclear DNA. A for

The effects of the subcellular localization of overexpressed bioreductive enzyme NAD(P)H:quinone oxidoreductase 1 (NQO1) on the activity of the antineoplastic agent mitomycin C (MC) under aerobic and hypoxic conditions were examined. Chinese hamster ovary (CHO-K1/ dhfr ؊ ) cells were transfected with NQO1 cDNA to produce cells that overexpressed NQO1 activity in the nucleus (148-fold) or the cytosol (163-fold) over the constitutive level of the enzyme in parental cells. Subcellular localization of the enzyme was confirmed using antibody-assisted immunofluorescence. Nuclear localization of transfected NQO1 activity increased the cytotoxicity of MC over that produced by overexpression in the cytosol under both aerobic and hypoxic conditions, with greater cytotoxicity being produced under hypoxia. The greater cytotoxicity of nuclear localized NQO1 was not attributable to greater metabolic activation of MC but instead was the result of activation of the drug in close proximity to its target, nuclear DNA. A positive relationship existed between the degree of MC-induced cytotoxicity and the number of MC-DNA adducts produced. The findings indicate that activation of MC proximal to nuclear DNA by the nuclear localization of transfected NQO1 increases the cytotoxic effects of MC regardless of the degree of oxygenation and support the concept that the mechanism of action of MC involves alkylation of DNA.
Mitomycin C (MC) 1 is a naturally occurring antibiotic that was isolated originally from the microorganism Streptomyces caspitosus (1). MC exhibits a broad spectrum of antitumor activity and is an important component in the combination chemotherapy of malignancies such as early stage head and neck cancer, early stage cervical cancer, and intravesicle ther-apy of superficial bladder cancer. Specific MC-DNA lesions associated with the action of MC consist of both monofunctional and bifunctional alkylations (2)(3)(4)(5)(6)(7)(8)(9). Monoalkylations initially occur through the linkage of the C-1 position of MC to the amino function in the 2-position of guanine bases in DNA and may proceed to a DNA cross-link through the C-10 position of MC to an amino entity in the 2-position of an adjacent DNA guanine (6). Although monoalkylations are potentially cytotoxic, compelling evidence in both bacterial and mammalian systems implicates MC-induced cross-links as the primary event responsible for cell death (10 -12).
The difference in the production of these two reactive species, MCH 2 and MC semiquinone anion radical, gives rise to the difference in the survival curves that are observed for cells treated with MC under aerobic and hypoxic conditions. This separation, known as the "aerobic/hypoxic differential," is reflected in the cytotoxicity profiles for Chinese hamster ovary (CHO) cell transfectants overexpressing NPR (27) and NBR (28,29) but not for those overexpressing NQO1 (30). Thus, although NPR and NBR are not important enzymes in the activation of MC under aerobic conditions, they contribute to the preferential activation of MC in hypoxia. Bioactivation of MC by NQO1, a two-electron donating enzyme, directly generates MCH 2 ; therefore, NQO1 contributes to the cytotoxicity of MC under both aerobic and hypoxic conditions.
A number of factors are involved in modulating the therapeutic efficacy of MC. These include the levels of the individual reductase enzymes and the cofactors, NADH and NADPH; the extent of formation of the exceedingly cytotoxic DNA cross-link; and the damaging oxygen radicals, superoxide, hydrogen peroxide, and/or hydroxyl radical that are formed by redox cycling reactions. Reductive activation of the MC present in the cytosol in air to form the highly reactive electrophile MCH 2 minimizes the formation of the exceedingly lethal DNA cross-links, since MCH 2 reacts with a variety of cellular nucleophiles, including water, during its diffusion into the nucleus and reaction with DNA. Thus, activation of MC in the nucleus close to the DNA target should result in a greater number of DNA cross-links and increased cell kill. Such a prediction was realized in studies with CHO cells transfected with a cDNA encoding rat NBR, in that nuclear localization of the enzyme resulted in greater cell kill and increased numbers of MC-DNA adducts over those occurring with overexpressed NBR localized predominantly in its normal mitochondrial and endoplasmic recticulum locations (29). Since NQO1, being a two-electron reducing system that directly generates MCH 2 (4,7,10,15,22,(31)(32)(33)(34)(35), is considered to be a more important bioactivator of MC than the one-electron activating system, NBR, we have measured the cytotoxicity of MC and the number of MC-DNA adducts formed from this agent in CHO cells overexpressing NQO1 activity in the cytosolic and nuclear compartments under aerobic and hypoxic conditions.

EXPERIMENTAL PROCEDURES
Plasmid Construction-The cDNA mammalian expression constructs of nuclear localized NQO1 (NLS-NQO1) and cytoplasmic localized NQO1 (CYTO-NQO1) were prepared by PCR using a plasmid encoding the cDNA for the rat NQO1 gene (30) as the template. To create the NLS-NQO1, the NQO1 gene was modified by PCR using the following upstream oligonucleotide primer: 5Ј-CGC GGA GCT AGC CCG GTG AGA AGA GCC CTG ATT GTA-3Ј. This oligonucleotide featured the sequences for a unique NheI site (shown above in italics) upstream of the coding sequences for rat NQO1 truncated to remove the start codon. The NheI site was used to fuse NQO1 in frame behind the nuclear localization signal of the SV40 large T antigen (36 -38) in the pRC-CMV plasmid (Invitrogen).
The 3Ј-end of the gene was modified to include the sequences for the 15 amino acids of the muscle actin epitope (highlighted below by the double underline) (39), which is recognizable by the HUC 1-1 monoclonal antibody (ICN, Costa Mesa, CA), followed by a unique XbaI site (represented in italics) in succession with the complementary sequences for NQO1 (shown by the single underline) using the following downstream oligonucleotide primer: 5Ј-CGC GGA TCC TCT AGA CTA AAT ACT TGG CCC TTC ATC ATA TTC TTG TTT GGA TAT CCA CAT CCC TCT AGC TTT GAT CTG GTT ATC GGC-3Ј. The resulting PCR product was cloned in frame behind the sequence for the nuclear localization signal of the SV40 large T antigen residing in the mammalian expression vector, pRC-CMV (Invitrogen), created for these studies.
The CYTO-NQO1 plasmid was prepared to include a HindIII site (represented below in italics) and a Kozak sequence (highlighted by the double underline) upstream of the NQO1 coding sequence (highlighted by a single underline) using the following upstream oligonucleotide primer: CYTO-NQO1 forward, 5Ј-CGC GGA AAG CTT ACC GCC ATG GCG GTG AGA AGA GCC CTG ATT GTA TTG 3Ј. The reverse primer for the muscle actin epitope was identical to the one employed for the creation of the NLS-NQO1. This gene product was unidirectionally subcloned into the HindIII and XbaI polylinker sites of the mammalian expression vector, pRC-CMV (Invitrogen).
Both of these plasmids, which insert stably into the genome of transfected cell lines, contain the appropriate sequences for polyadenylation and selection by G418 (neomycin resistance). In the case of NLS-NQO1, the plasmid also contains the sequence for the nuclear localization signal from the SV40 large T antigen. Expression of both NLS-NQO1 and CYTO-NQO1 fusion constructs were driven by the cytomegalovirus promoter to produce fusion proteins.
Cell Culture-The cell line used in this study is the dihydrofolatedeficient variant of the CHO-K1 cell line called CHO-K1/dhfr Ϫ (40) and was obtained from the American Type Culture Collection (Manassas, VA). These cells were maintained in Iscove's modified Dulbecco's medium (Invitrogen) supplemented with 10% fetal bovine serum (Sigma), 2 mM glutamine, 0.1 mM hypoxanthine, 0.01 mM thymidine, penicillin (100 units/ml), and streptomycin (100 g/ml). Transfected cell lines were maintained in identical medium in the presence of 1 mg/ml G418 (Geneticin; Invitrogen) to provide a positive selection pressure for the expression vector. Cells were grown and treated as monolayers in a variety of cell culture vessels in a humidified atmosphere of 95% air and 5% CO 2 at 37°C.
Transfections-CHO-K1/dhfr Ϫ cells were transfected with either NLS-NQO1 or CYTO-NQO1 constructs using calcium phosphate methodology essentially as described by the manufacturer (Invitrogen). Cells containing NLS-NQO1 or CYTO-NQO1 were selected using G418, and individual positive populations were expanded and screened for NQO1 activity. Cell populations with elevated enzyme activity were subjected to single cell sorting by flow cytometry and expanded. Expanded clones were rescreened and selected for high levels of NQO1 activity and basal levels of NPR and NBR activities. Relatively matched clones with respect to NQO1 activity were selected for study.
Enzyme Activity Assays-Monolayers of exponentially growing cells incubated under aerobic (95% air and 5% CO 2 ) conditions were harvested in phosphate-buffered saline and lysed by sonication, and NQO1 activity was measured essentially as described by Ernster (41). NQO1 activity inhibitable by dicumarol (Sigma) was quantified by measuring the reduction of dichlorophenolindophenol (Sigma) at 600 nm with a Beckman model 25 UV-visible spectrophotometer (Beckman Instruments Inc., Fullerton, CA); activities were calculated using an extinction coefficient of 21 mM Ϫ1 cm Ϫ1 at 30°C. NPR activity was quantified by the reduction of ferricytochrome c measured at 550 nm at 30°C; activities were calculated using an extinction coefficient of 21 mM Ϫ1 cm Ϫ1 at 30°C (42). NBR activity was quantified as NADH:ferricyanide reductase measured at 420 nm at 30°C; activities were calculated using an extinction coefficient of 1.02 mM Ϫ1 cm Ϫ1 at 30°C (32) using a final concentration of 0.34 mM NADH. Protein concentrations were determined by bicinchoninic acid assay (Pierce) (43).
Indirect Immunofluorescence-Cells were seeded at 1000 cells/well on a poly-D-lysine-treated 8-chambered glass slide (BD Falcon, Boston, MA). Twenty-four h later, the cells were fixed in 20% formaldehyde for 15 min and were permeabilized with ice-cold acetone for 5 min. Samples were incubated with primary antibody, anti-muscle actin (HUC 1-1) monoclonal antibody (ICN, Costa Mesa, CA) to the muscle actin epitope at a 1:120 dilution, washed, and then incubated with goat anti-mouse IgG-fluorescein isothiocyanate-conjugated antibody at a 1:128 dilution (Sigma). Each incubation was performed at 4°C for 18 h. The samples were then treated with 50% glycerol in phosphate-buffered saline, pH 9.0, and examined using a Nikon Optiphot microscope equipped with a Nikon Episcopic-Fluorescence attachment EF-D and a Nikon UFX-IIA MicroFlex Camera.
Aerobic and Hypoxic Cell Survival Experiments-Cells were seeded in glass milk dilution bottles at 2.5 ϫ 10 5 cells/bottle and grown for 3 days in a humidified atmosphere of 95% air and 5% CO 2 . Hypoxia was established by gassing the cultures with a humidified mixture of 95% N 2 plus 5% CO 2 (containing Ͻ10 ppm O 2 ) (AirGas, Cheshire, CT) at 37°C for 2 h through a rubber septum fitted with 13-gauge (inflow) and 18-gauge (outflow) needles. MC at 2.5, 5.0, 10, 12.5, and 15 M was then introduced into the cultures using a Hamilton syringe without compromising the hypoxic environment, and cultures were incubated for 1 h. Cells under aerobic conditions were treated identically but gassed with a humidified atmosphere of 95% air plus 5% CO 2 . MC-treated cells were then washed, harvested by trypsinization, and assayed for survival using a clonogenic assay (27). Macroscopic colonies consisting of more than 40 cells were counted, and the plating efficiencies, defined as the number of macroscopic colonies counted divided by the number of cells plated, was determined. The surviving fractions were calculated by normalizing the plating efficiencies of the drug-treated groups to those of the vehicle-treated control groups.
MC Co., Tokyo, Japan) for 2 h. Genomic DNA was isolated from 1 ϫ 10 7 cells using the PURGENE DNA purification system (Gentra Systems, Minneapolis, MN) as described by the manufacturer. Briefly, cells were lysed and treated with 100 g/ml proteinase K overnight followed by 20 g/ml RNase A for 2 h at 37°C. Isolated DNA was washed two times with 70% ethanol to remove noncovalently bound [ 3 H]MC, and the DNA was resuspended in 10 mM Tris-HCl, 1 mM EDTA (pH 7.0). An aliquot was used to quantify the number of [ 3 H]MC-DNA adducts using a Beckman scintillation spectrometer, and the DNA concentration was determined spectrophotometrically at A 260 nm. Radioactivity in the sample was normalized to the total DNA concentration.

Isolation of Stable CHO Cell Line Transfectants
Overexpressing NQO1-To examine the effects of the overexpression of NQO1 activity in different subcellular compartments on the sensitivity of aerobic and hypoxic cells to MC, clonal populations of CHO cells were isolated and assayed for NQO1 activity. Clones that exhibited high levels of NQO1 activity relative to parental CHO cells, with no significant differences in the levels of expression of both NBR and NPR were chosen for these studies; the enzymatic activity profiles for NLS-NQO1, CYTO-NQO1, and parental CHO cells are summarized in Table I. CYTO-NQO1 and NLS-NQO1 CHO-K1/dhfr Ϫ cell transfectants that expressed significantly more NQO1 activity (163-fold and 148-fold, respectively) than parental CHO cells were isolated. This level of overexpression in transfected CHO-K1/dhfr Ϫ cells is roughly comparable with previously reported increases in NQO1 activity of 126-and 133-fold over parental CHO cells (30) and is markedly higher than the increase in expression obtained for cells transfected with the one-electron bioreductive enzyme systems NPR (44,45) and NBR (28,29). No significant changes occurred in the NBR and NPR activities of the CYTO-NQO1 and NLS-NQO1 CHO-K1/dhfr Ϫ transfectants and parental cells. Xanthine:oxygen oxidoreductase and xanthine:NAD ϩ oxidoreductase activities were not measured, since CHO-K1/dhfr Ϫ cells do not show detectable levels of these enzymes (27). Additionally, measurements of cellular growth indicated that CHO-K1/dhfr Ϫ transfectants and parental cells had comparable growth rates (data not shown).
Subcellular Distribution of Overexpressed NQO1 Activity in CHO-K1/dhfr Ϫ Parental, CYTO-NQO1, and NLS-NQO1 Cells-To ensure that the nuclear localization signal included in the fusion construct of NLS-NQO1 directed the overexpressed enzyme to the nucleus, NQO1 was visually assessed by indirect immunofluorescence techniques in NLS-NQO1, CYTO-NQO1, and parental CHO-K1/dhfr Ϫ cells (Fig. 1). Since both the CYTO-NQO1 and NLS-NQO1 fusion proteins are epitopetagged, the use of antibody-directed immunofluorescence resulted in visualization of only the fusion protein and not the endogenous NQO1 protein. Staining with the HUC1-1 monoclonal antibody for the epitope tag revealed intense nuclear staining in the overexpressing NLS-NQO1 cells, indicating that nuclear localization of NQO1 was achieved (Fig. 1A). In contrast, CYTO-NQO1 cells demonstrated intense staining, which was predominantly cytoplasmic, with minimal nuclear staining (Fig. 1B). As expected, parental cells exhibited minimal fluorescence staining, since the HUC1-1 antibody specifically recognized the fusion protein expressed by the transfectants (Fig. 1C).

Effects of MC under Aerobic and Hypoxic Conditions on the Survival of Stable CHO Cell Transfectants
Overexpressing NQO1 in the Cytosol or the Nucleus-The effects of overexpression of NQO1 activity, either in the cytosol ( Fig. 2A) or nucleus (Fig. 2B), on the cytotoxicity of MC under aerobic and hypoxic conditions were measured by comparing the survival curves for the CHO transfectants with that of parental cells. In the parental cell line, greater levels of cell kill were obtained in hypoxia than under aerobic conditions, a finding consistent with previous reports (27,29,30). Comparative survival curves for CHO-K1/dhfr Ϫ parental and CYTO-NQO1 transfectants ( Fig. 2A) and CYTO-NQO1 and NLS-NQO1 transfectants (Fig. 2B) exposed to MC under aerobic and hypoxic conditions are shown. Overexpression of cytosolic localizing NQO1 activity in CHO cells resulted in a significant increase in the cytotoxicity of MC relative to parental cells at concentrations of MC of 2.5-10 M under both aerobic and hypoxic conditions. Furthermore, a highly significant increase in MC-induced toxicity was observed in cells that overexpressed NQO1 activity in the nucleus over those overexpressing NQO1 activity in the cytosol under both aerobic and hypoxic conditions at a concentration of 10 M MC (Fig. 2B). Exposure of CHO-K1/dhfr Ϫ parental, CYTO-NQO1, and NLS-NQO1 cells to higher concentrations of MC in the range of 10 -15 M clearly demonstrated that overexpression of NQO1 in either the cytosol or nucleus increased cell kill under both aerobic (Fig. 3A) and hypoxic (Fig. 3B) conditions, with nuclear localization of the enzyme activity decreasing cell survival more than cytosolic localization, although cytosolic overexpression of NQO1 activity was increased by 163-fold, whereas nuclear overexpression was increased somewhat less by 148-fold.
Metabolic Activation of MC-Since metabolic activation of MC is required for the cytotoxic effects of this agent, differences in activation between CHO-K1/dhfr Ϫ parental, CYTO-NQO1, and NLS-NQO1 cells may be responsible for the observed differences in the cytotoxicity produced by MC. To measure the bioreduction of MC in intact cells, the comparative ability of these cell lines to metabolically activate the drug was assessed by measuring the disappearance of MC from the cultures. Fig.  4, A and B, depicts the time course for the activation of MC under aerobic and hypoxic conditions, respectively, for CHO-K1/dhfr Ϫ parental, CYTO-NQO1, and NLS-NQO1 cells treated with 12.5 M MC. The aerobic studies (Fig. 4A) were performed using 1.0 ϫ 10 8 cells/ml, whereas the hypoxic studies (Fig. 4B) were performed using 5.0 ϫ 10 6 cells/ml. Thus, although not evident from the data presented in Fig. 4, A and B, the rate of disappearance of MC under hypoxia was 20 times greater than in air, requiring the use of 20 times fewer cells under hypoxia to generate the results shown. Nonetheless, it is clear that  Table II. Overexpression of NQO1 activity in the cytosol produced 37 and 13% increases in the number of MC-DNA adducts formed in air and hypoxia, respectively, compared with parental cells. In cells that overexpressed NQO1 activity in the nucleus, 40 and 34% increases in MC-DNA adducts were observed under aerobic and hypoxic conditions, respectively, compared with cells that overexpressed the enzyme in the cytosol. All of the cell lines (i.e. parental CHO-K1/dhfr Ϫ cells and the CYTO-NQO1 and NLS-NQO1 transfectants) exhibited increases of ϳ50% in the production of MC-DNA adducts when exposed to MC under conditions of hypoxia, reflecting preferential alkylation of DNA by [ 3 H]MC under hypoxic conditions (46,47). DISCUSSION The effects of overexpressed NQO1 activity in different subcellular locations on the sensitivity of CHO-K1/dhfr Ϫ cells to the antineoplastic agent MC was ascertained to determine the contribution of this activating enzyme to the cytotoxic action of this agent. Overexpression of NQO1 activity by 163-fold in the cytosol caused significant increases in sensitivity of these cells to the cytotoxic effects of MC under both aerobic and hypoxic conditions (Fig. 2). These results are consistent with the finding that the intracellular concentration of NQO1 activity is important to the cytotoxic action of MC, independent of the degree of oxygenation, providing further evidence for a role for NQO1 in the bioreductive activation of MC. In earlier studies from our laboratory, in which dicumarol was used to protect NQO1-rich cells from the cytotoxic actions of MC, a role for this enzyme was assigned in activating MC under aerobic but not hypoxic conditions (15,24). However, it is now widely recognized that dicumarol is not a specific inhibitor of NQO1 in that it can inhibit other enzymes (48). Additionally, dicumarol has been shown to induce the expression of xanthine:NAD ϩ oxidoreductase and xanthine:oxygen oxidoreductase (49), although we do not believe that the induction of these enzymes occurs in CHO cells. Therefore, it is likely that the increased sensitivity to MC observed in cells treated with dicumarol under hypoxic conditions reflects contributions to the activation of MC by bioreductive enzymes other than NQO1.
A comparison of the survival curves for NLS-NQO1 and CYTO-NQO1 cells revealed that nuclear overexpression of NQO1 activity in CHO-K1/dhfr Ϫ cells results in greater cytotoxicity from exposure to MC. This finding is in agreement with that reported for cytosolic (28) and nuclear expression of NBR (29) in CHO-K1/dhfr Ϫ cells. Since one of the features of NQO1 that separates it from NPR and NBR is that NQO1 activates MC by a two-electron reduction process to directly generate the oxygen-insensitive MCH 2 intermediate, we anticipated that overexpression of NQO1 activity would increase the sensitivity of cells to the cytotoxic actions of MC under both aerobic and hypoxic conditions. Thus, although an increase in cell kill occurred in both CYTO-NQO1 and NLS-NQO1 cells over that obtained in CHO-K1 dhfr Ϫ parental cells, when cells were treated with 12.5 M MC under aerobic conditions an ϳ35-fold increase in cytotoxicity was observed in cells overexpressing NQO1 activity in the nucleus compared with those overexpressing the enzyme in the cytosol. Similarly, in cells exposed to the same concentration of this agent under hypoxic conditions, an approximately 35-fold increase in cell kill was obtained in cells that overexpressed nuclear localized NQO1 activity compared with cytosolic localized NQO1. Interestingly, significant increases were observed only at concentrations of 10 M MC or higher. Since NQO1 activity was overexpressed at levels in the CYTO-NQO1 and NLS-NQO1 transfectants that were greater than 100-fold more than that of parental CHO cells, it is conceivable that MC may be essentially completely depleted by activation by CHO-K1/dhfr Ϫ parental cells when used at relatively low concentrations, accounting for the fact that differences in cytotoxicity were not observed at concentrations below 10 M in both aerobic and hypoxic conditions. In an analogous fashion when NBR was overexpressed in the nucleus by 3-fold, greater differences between aerobic (5-fold) and hypoxic (10-fold) toxicities were observed than seen with mitochondrial/endoplasmic recticulum-localized NBR, overexpressed by 5-fold. The fact that aerobic cell toxicity is lower than hypoxic toxicity is a reflection of the oxygen-sensitive one-electron reductive activation of MC by NBR.
NQO1 activates MC to directly generate the oxygen-insensitive MCH 2 -reactive intermediate. Therefore, we anticipated the increase in the cytotoxic action of MC that occurred under both aerobic and hypoxic conditions by the overexpression of NQO1 activity. A 10-fold increase in cytotoxicity was observed in cells that overexpressed NQO1 activity in the nucleus compared with cells that overexpressed the enzyme in the cytosol, when exposed to 10 M MC under aerobic conditions. Similarly, at the same concentration of this agent under hypoxic conditions, a 7-fold increase in cytotoxicity was observed in cells that overexpressed nuclear localized NQO1 activity compared with . Cells were grown as described under "Experimental Procedures," stained with the HUC1-1 monoclonal antibody specific for the rat smooth muscle actin epitope, and photographed using a fluorescence microscope. cells with cytosolic localized NQO1 activity. Additionally, at higher concentrations of MC, comparison of the curves suggests a trend wherein nuclear overexpression of NQO1 activity increased cytotoxicity to MC to a greater extent than cytosolic overexpression of the enzyme under both aerobic (Fig. 3A) and hypoxic (Fig. 3B) conditions. Collectively, the finding that sensitivity occurs regardless of the degree of oxygenation is consistent with the concept that the MCH 2 mediates the cytotoxic effects of the drug.
In analogous studies where NPR is overexpressed in the nucleus by 9-fold, greater differences between aerobic (11-fold) and hypoxic (90-fold) toxicities were observed than were seen with endoplasmic recticulum-localized NPR overexpressed by 16-fold (45). Likewise, when NBR was overexpressed by 3-fold in the nucleus, the increases in sensitivity under aerobic and hypoxic conditions were 5-and 10-fold greater, respectively, than those seen for mitochondrial/endoplasmic recticulum-localized enzyme overexpressed by 5-fold. Taken together, since aerobic cytotoxicity is less than hypoxic toxicity, the findings reflect the fact that activation of MC by NBR and NPR occurs by an oxygen-sensitive mechanism that results in redox cycling of the MC semiquinone anion radical.
Cell survival assays indicated that the closer the activation of MC to the nuclear DNA target, the greater the extent of cell kill. This conclusion was solidified by the finding that activation of MC, measured by the disappearance of the prodrug, was somewhat greater when overexpression of NQO1 activity occurred in the cytosol than when it occurred in the nucleus, despite the fact that the transfectants with NQO1 in the nucleus displayed greater sensitivity to MC and more MC-DNA adducts. These findings suggest that activation of MC in the cytoplasm results in the alkylation of nucleophiles other than nuclear DNA, since the reactive MCH 2 electrophile migrates toward its nuclear DNA target.
Since it is widely accepted that nuclear DNA is the target of MC (6,10,12), the ability of this agent to produce genomic DNA alkylations in cells with nuclear and cytosolic overexpressed NQO1 activity was measured. Using [ 3 H]MC, we found that the degree of sensitivity to MC corresponded to the number of MC-DNA adducts in CHO-K1/dhfr Ϫ parental, CYTO-NQO1, and NLS-NQO1 cells. Under aerobic conditions, the number of DNA adducts generated from [ 3 H]MC results primarily from activation of the antineoplastic agent by NQO1, since the oneelectron reducing systems NPR and NBR generate the semiquinone anion radical, which in the presence of oxygen is rapidly converted back to MC and superoxide. The greater number of DNA adducts formed under hypoxic conditions reflects the contribution of NBR and NPR to that of the two-electron reducing system(s). The difference between the number of MC-DNA adducts under hypoxic and aerobic conditions for parental, CYTO-NQO1, and NLS-NQO1 cells results in relatively constant increases in cpm/g of DNA under hypoxia of 30, 26, and 32, respectively, over that in air, reflecting the constancy of the added activities of NBR and NPR under hypoxic conditions in transfectants. Furthermore, in agreement with expectations, overexpression of NQO1 activity in the cytosol increased the number of DNA alkylations over that in parental cells, and overexpression in the nucleus caused a further increase in MC-DNA adducts.
In a previous analogous study where the subcellular distribution of NBR was altered, a similar correlation between MC sensitivity and the number of MC-DNA adducts was obtained when comparisons were made between CHO-K1/dhfr Ϫ parental cells and transfectants overexpressing NBR activity in either the nucleus or the mitochondrial/endoplasmic reticulum membranes. Despite the fact that 5-fold differences in cell kill were observed between CHO-K1/dhfr Ϫ cells expressing NBR in the nucleus and in the mitochondria/endoplasmic reticulum, there was no difference in the number of MC-DNA adducts observed between the two cell types when treated with 10 M MC under aerobic conditions (29). Since NBR reduces MC by a one-electron pathway to an oxygen-sensitive species, it is probable that under aerobic conditions, the observed enhancement in cell kill produced by MC involved a different mechanism. Thus, MC sensitivity under these conditions probably was not solely a consequence of MC-DNA adduct formation, but in addition reflected increases in damage caused by oxygen radi-  cals produced as by-products of the interaction between the MC semiquinone anion radical and molecular oxygen that have been shown to have toxic effects in EMT6 cells (50,51).
Taken together, the findings show that activation of MC proximal to its nuclear DNA target through nuclear expression of NQO1 activity or NBR activity results in enhanced cell kill as a consequence of increases in genomic DNA alkylations. Thus, the subcellular localization of the bioactivating reducing enzymes, as well as the levels of these enzymes, is important in determining the cytotoxicity of MC.