The conserved Trp114 residue of thioredoxin reductase 1 has a redox sensor-like function triggering oligomerization and crosslinking upon oxidative stress related to cell death.

The selenoprotein thioredoxin reductase 1 (TrxR1) has several key roles in cellular redox systems and reductive pathways. Here we discovered that an evolutionarily conserved and surface-exposed tryptophan residue of the enzyme (Trp114) is excessively reactive to oxidation and exerts regulatory functions. The results indicate that it serves as an electron relay communicating with the FAD moiety of the enzyme, and, when oxidized, it facilitates oligomerization of TrxR1 into tetramers and higher multimers of dimers. A covalent link can also be formed between two oxidized Trp114 residues of two subunits from two separate TrxR1 dimers, as found both in cell extracts and in a crystal structure of tetrameric TrxR1. Formation of covalently linked TrxR1 subunits became exaggerated in cells on treatment with the pro-oxidant p53-reactivating anticancer compound RITA, in direct correlation with triggering of a cell death that could be prevented by antioxidant treatment. These results collectively suggest that Trp114 of TrxR1 serves a function reminiscent of an irreversible sensor for excessive oxidation, thereby presenting a previously unrecognized level of regulation of TrxR1 function in relation to cellular redox state and cell death induction.

Inhibition of malignant cell growth by 311, a novel iron chelator of the pyridoxal isonicotinoyl hydrazone class: effect on the R2 subunit of ribonucleotide reductase.

The key roles of iron and iron proteins in cell proliferation make them potential targets for cancer therapy. However, clinical trials directed toward perturbation of tumor iron homeostasis by iron chelation have been limited to the use of deferoxamine (DFO). There is thus a need to develop agents with greater efficacy. In the present study, we investigated the mechanism of cytotoxicity of 311 (2-hydroxy-1-naphthylaldehyde benzoyl hydrazone), a novel iron chelator of the pyridoxal isonicotinoyl class. We found that 311 inhibited the growth of CCRF-CEM cells in a time- and concentration-dependent fashion with an IC(50) that was approximately 20-fold lower than that of DFO. 311 also inhibited the growth of breast, bladder, and head and neck cancer cell lines. Using electron spin resonance (ESR) spectroscopy analysis, we found that a 12-h exposure of CCRF-CEM cells to 311 inhibited the tyrosyl radical ESR signal of the R2 subunit of ribonucleotide reductase. However, overproduction of the R2 subunit in hydroxyurea-resistant CCRF-CEM cells was associated with a decrease in sensitivity of cells to 311 but not to DFO. Our studies show that 311 is a more potent cytotoxic agent than DFO, with activity against both hematopoietic and nonhematopoietic cell lines regardless of their p53 status. Furthermore, the ESR studies suggest that inhibition of the R2 subunit of ribonucleotide reductase is at least one mechanism by which 311 blocks cell proliferation.

Model reactions of Cr (VI) with DNA mediated by thiol species.

Model reactions were devised to investigate the capacity of physiologically interesting thiol compounds to mediate reactions between CrO4(2-) (Cr (VI)) and DNA. The sulfhydryl containing reagents included cysteine, glutathione, apo-metallothionein (apoMT). Zinc finger 3 of transcription factor IIIA (Zn-F3) of Xenopus laevis was investigated as a potential redox active site of reaction of Cr (VI) and thiol compounds. The DNA samples were calf thymus DNA and two oligomers, one of them specific for binding Zn-F3. Results showed that in the presence of Cr (VI) apoMT readily participated in damaging DNA in a reaction that appeared to be hydroxyl radical dependent. It also became cross-linked to oligomer and native DNA samples. In comparison, the other two thiol donors were largely inactive in these assays even though they, like apoMT, were able to reduce Cr (VI) to Cr (III) under the conditions of the experiments. Direct attempts to cross link thiols with DNA in the presence of Cr3+ were unsuccessful at pH 7.4. Together, the results indicate that apoMT can effectively collaborate with Cr (VI) in reactions that are deleterious to DNA.

Interactions of iron bleomycin, phosphate or cyanide, and DNA: sequence-dependent conformations and reactions.

The hypothesis was investigated that axial ligands bound to Fe(III)-bleomycin [Fe(III)Blm] are destabilized at specific 5'-guanine-pyrimidine-3' binding sites but are stable at nonselective dinucleotides. DNA oligomers and calf-thymus DNA were used in reactions with L-Fe(III)Blm, where phosphate and cyanide served as examples of large and small ligands (L). Both ligands underwent dissociation when L-Fe(III)Blm was bound to d(GGAAGCTTCC)2 (I) but not d(GGAAATTTCCC)2 (II) and at large ratios of calf-thymus DNA to drug. Fe(III)Blm is high spin in 20 mM phosphate buffer, signifying the presence of a phosphate adduct. In the titration of HPO4-Fe(III)Blm with calf-thymus DNA, a large excess of DNA was needed to reach the low-spin state, consistent with an equilibrium competition between phosphate and DNA for Fe(III)Blm. Equilibrium constants for binding Fe(III)Blm and CN-Fe(III)Blm to calf-thymus DNA (6.8x10(5) M(-1) and 5.9x10(4) M(-1), respectively, in HEPES buffer at 25 degrees C and pH 7.4) showed that the CN- ligand also reduced the affinity of DNA for the drug. The kinetics of dissociation of CN- from CN-Fe(III)Blm-DNA were slow and first order in bound drug. The reversible nature of these dissociation reactions was shown using 1H NMR spectroscopy of Fe(III)Blm-I in the absence and presence of large excesses of CN- or phosphate. The results are discussed in terms of a two-state hypothesis for the binding of L-Fe(III)Blm to specific and nonspecific dinucleotides. It is proposed that steric restrictions at specific sites inhibit binding of these ligands.

Cytochrome b(5) plays a key role in human microsomal chromium(VI) reduction.

The reduction of chromium(VI) to Cr(III) results in the formation of reactive intermediates that contribute to the cytotoxicity, genotoxicity, and carcinogenicity of Cr(VI)-containing compounds. Previous studies suggest that human microsomal Cr(VI) reduction likely proceeds through cytochrome b(5). In order to better understand Cr(VI) toxicity in humans, the role of cytochrome b(5) in combination with P450 reductase was examined in the reductive transformation of Cr(VI). Proteoliposomes containing human recombinant cytochrome b(5) and P450 reductase were constructed. The ability of P450 reductase to mediate efficient electron transfer from NADPH to cytochrome b(5) was confirmed by spectral analysis. The NADPH-dependent Cr(VI) reduction rate mediated by proteoliposomes was then compared to that of human microsomes. When these rates were normalized to equivalent cytochrome b(5) concentrations, the NADPH-dependent Cr(VI) reduction rates mediated by human microsomes were essentially identical to those for proteoliposomes containing cytochrome b(5) plus P450 reductase. Proteoliposomes containing only P450 reductase or cytochrome b(5) exhibited poor Cr(VI) reducing capabilities. Since it had been previously shown that trace amounts of iron (Fe) could dramatically stimulate microsomal Cr(VI) reduction, the ability of Fe to stimulate Cr(VI) reduction by proteoliposomes was examined. Both ferric chloride (FeCl(3)) and ferric adenosine-5'-diphosphate (FeADP) were shown to stimulate Cr(VI) reduction; this stimulation could be abolished by the addition of deferoxamine, a specific Fe(III) chelator. The NADPH-dependent reduction rates of various ferric complexes by proteoliposomes were sufficient to account for the increased Cr(VI) reduction rates seen with the addition of FeCl(3) or FeADP. Cr(V) was detected by electron paramagnetic resonance (EPR) spectroscopy as a transient intermediate formed during NADPH-dependent Cr(VI) reduction mediated by proteoliposomes containing cytochrome b(5) and P450 reductase. Overall, cytochrome b(5) in combination with P450 reductase can account for the majority of the NADPH-dependent Cr(VI) reduction seen with human microsomes.

Cellular adaptation to down-regulated iron transport into lymphoid leukaemic cells: effects on the expression of the gene for ribonucleotide reductase.

Ribonucleotide reductase is an iron-containing enzyme that is essential for DNA synthesis. Whereas previous studies have used various iron chelators to examine the relationship between cellular iron metabolism and ribonucleotide reductase activity in cells, they have not elucidated the relationship between iron transport into cells and the expression of the gene for ribonucleotide reductase. To investigate this, we examined ribonucleotide reductase mRNA, protein and enzyme activity in a novel line of CCRF-CEM cells (DFe-T cells) that display an approx. 60% decrease in their uptake of iron compared with the parental wild-type cell line. We found that DFe-T cells displayed an approx. 40% decrease in ribonucleotide reductase specific enzyme activity relative to wild-type cells without a change in their proliferation. Kinetic analysis of CDP reductase activity revealed an approx. 60% decrease in V(max) in DFe-T cells without a change in K(m). Despite the decrease in enzyme activity, the mRNA and protein for the R1 and R2 subunits of ribonucleotide reductase in DFe-T cells were similar to those of wild-type cells. ESR spectroscopy studies revealed that DFe-T cells had a 22% decrease in the tyrosyl free radical of the R2 subunit, suggesting that a larger amount of R2 protein was present as functionally inactive apo-R2 in these cells. Our studies indicate that ribonucleotide reductase activity in CCRF-CEM cells can be down-regulated by more than 50% in response to down-regulated iron transport without an adverse effect on cell proliferation. Furthermore, our studies suggest a regulatory link between ribonucleotide reductase activity and iron transport into these cells.

Orientation of iron bleomycin and porphyrin complexes on DNA fibers.

Bleomycin (Blm) is an antitumor agent that requires iron and oxygen for strand cleavage of DNA. In this study, ferric bleomycin, Fe(III)Blm, or the nitric oxide adduct of ferrous bleomycin, ON-Fe(II)Blm, were bound to one-dimensionally oriented DNA fibers. Reductive nitrosylation of Fe(III) complexes took place in situ on B-form DNA fibers. Electron paramagnetic resonance (EPR) spectra were obtained as a function of the angle phi between the magnetic field B and the fiber axis Zf. For comparison, EPR spectra were acquired for ON-Fe(II)TMpyP and ON-Fe(II)TMpyP-Im on oriented DNA fibers, where TMpyP is 5,10,15,20-tetrakis(1-methyl-4-pyridino)porphyrin and Im is imidazole. EPR spectra showed both low-spin Fe(III)Blm and ON-Fe(II)Blm bound to B-form DNA in two slightly different binding orientations in the ratio of 1:0.2. With A-form DNA, a fraction of bound Fe(III)Blm was high spin. Specifically, the angle beta between the fiber axis Zf and the g axis, gz, perpendicular to or nearly perpendicular to the equatorial plane of the iron complex was estimated as 20 degrees and 25 degrees for ON-Fe(II)Blm and 30 degrees and 25 degrees for Fe(III)Blm, respectively. The angle gamma that determines the orientation of gx and gy axes was estimated as 90 degrees for the two ON-Fe(II)Blm species and 10 degrees for the two Fe(III)Blm species, respectively. The NO was held rigidly in place as the temperature increased from 123 K to room temperature for ON-Fe(II)Blm but not for ON-Fe(II)TMpyP or ON-Fe(II)TMpyP-Im. It is hypothesized that the NO is structurally oriented by hydrogen bonding like the peroxide is held in HO2(-)-Co(III)Blm (Wu et al. J. Am. Chem. Soc. 1996, 118, 1281-1294). The EPR parameters are consistent with a six-coordinate complex for ON-Fe(II)Blm, although the superhyperfine structure from the trans nitrogen was not detected. The increase in g value anisotropy upon binding ON-Fe(II)Blm to DNA fiber may be caused by an increase in the overlap of d pi and 2p pi* orbitals induced by an interaction of NO with DNA and/or by a perturbation of d orbitals due to the pyrimidine-guanine interaction. It is concluded that the EPR parameters of ON-Fe(II)Blm and Fe(III)Blm bound to oriented DNA support the hypothesis that FeBlm species bind to DNA with adduct structures similar to those formed by related CoBlm species and DNA.

Mechanism of superoxide dismutase/H(2)O(2)-mediated nitric oxide release from S-nitrosoglutathione--role of glutamate.

S-Nitrosoglutathione (GSNO), a physiologically relevant nitric oxide ((*)NO) donor, exhibits antioxidant, anti-ischemic, and antiplatelet properties. The exact mechanism of (*)NO release from GSNO in biological systems has not been determined. Both copper ions and copper-containing enzymes have been shown to catalyze (*)NO release from GSNO. In this study we observed that copper-zinc superoxide dismutase (Cu,ZnSOD) in the presence of H(2)O(2) caused a rapid decomposition of GSNO, forming oxidized glutathione (GSSG) and (*)NO. The cupric ions (Cu(2+)) released from Cu,ZnSOD were bound to the glutamate moiety of GSNO, yielding a 2:1 (GSNO)(2)Cu(2+) complex. Strong chelators of cupric ions, such as histidine and diethylenetriaminepentaacetic acid, inhibited the formation of (GSNO)(2)Cu(2+) complex, GSSG, and (*)NO. GSSG alone inhibited Cu(2+)-induced decomposition of GSNO. This effect is attributed to complexation of copper by GSSG. We conclude that binding of copper to GSNO is obligatory for (*)NO release from GSNO; however, the rate of this reaction was considerably slowed due to binding of Cu(2+) by GSSG. The glutamate moiety in GSNO and GSSG controls copper-catalyzed (*)NO release from GSNO. Cu,ZnSOD and H(2)O(2) enhanced peroxidation of unsaturated lipid that was inhibited by GSNO. The antioxidant function of GSNO is related to the sequestering of copper by GSNO and its ability to slowly release (*)NO. Implications of these findings are discussed in relation to GSNO-induced cardioprotection and to neuropathological processes.

Complexation of type 2 copper by cytochrome c oxidase: probing of metal-specific binding sites by electron paramagnetic resonance.

Cytochrome c oxidase, CcO, contains at least four, probably five type 2 copper binding sites per monomer in addition to the mixed valence [CuA(1.5+)CuA(1.5+)], S = 1/2 center and the EPR-silent CuB. Electron paramagnetic resonance (EPR) parameters for these site are g parallel = 2.22 and A parallel = 195 G. Nitrogen superhyperfine structure is observed in the g perpendicular region, with A perpendicular N of around 15 G. The EPR parameters for Cu(2+) bound to a synthetic peptide, AHGSVVKSEDYALPS, are similar to the parameters for the type 2 sites in CcO. The lines in the EPR spectrum of the type 2 site in the synthetic peptide are better resolved at low microwave frequency (3.4 GHz). Resolved lines in the expansion of the MI = -1/2 line in the g parallel region of the low frequency spectrum are attributed to superhyperfine structure from three almost equivalent nitrogen donor atoms bound to Cu(2+) in a square planar configuration. The MI = -1/2 line in the g parallel region for excess Cu(2+) bound to CcO is not as well resolved as for the synthetic peptide, presumably because the four or five binding sites per monomer are similar, but not exactly equivalent. These binding sites are proposed to be at the N-terminus of subunits of CcO, for example, at subunit IV where the sequence is AHGS-. Nitrogen donor atoms from the alpha-amino group of the amino terminal residue, the imidazole group of histidine, and a peptide nitrogen are predicted to comprise the binding site.

Fe- and Co-bleomycins bound to site specific and nonspecific DNA decamers: comparative binding and reactivity of their metal centers.

Co- and Fe-bleomycins (Blms) have been reacted with DNAa, d(GGAAGCTTCC)2, containing a specific site for cleavage, and DNAb, d(GGAAATTTCC)2, a closely related nonspecific 10-mer, to survey whether features of structure and reactivity of these adducts vary systematically as a function of the base sequence of the DNA oligomer. The ESR spectrum of NO-Fe(II)BlmDNAa is rhombically perturbed in comparison with that of NO-Fe(II)BlmDNAb, which is nearly identical to the spectrum of NO-Fe(II)Blm. The ESR spectrum of Fe(III)BlmDNAa in phosphate buffer is low-spin; that of Fe(III)BlmDNAb is high-spin as seen with Fe(III)Blm alone. According to absorbance spectroscopy, O2-Fe(II)BlmDNAa is stabilized in comparison with the DNAb adduct. Similar stabilization of O2-Co(II)Blm bound to DNAa but not to DNAb was also observed by ESR spectroscopy. HO2(-)-Co(III)Blm A2 binds in slow exchange on the NMR time scale to DNAa at its 5'-G-pyrimidine-3' site of cleavage. In contrast, fluorescence and NMR spectroscopy demonstrate that most of HO2(-)-Co(III)Blm A2 binds stoichiometrically in fast exchange to DNAb. The reactions of Fe(III)BlmDNAa and Fe(III)BlmDNAb with ascorbate and O2 reveal that the latter becomes activated and cleaves its 10-mer, producing base propenals, at a faster initial rate. Thus, in two series of metallobleomycins, (A) NO-Fe(II)Blm, O2-Fe(II)Blm, Fe(III)Blm in phosphate buffer, and HO2(-)-Fe(III)Blm and (B) O2-Co(II)Blm and HO2(-)-Co(III)Blm, the metal domain of each species interacts differently with DNA depending upon its base sequence.

Metal coordination environment and dynamics in 113cadmium bleomycin: relationship to zinc bleomycin.

The 13C chemical shifts of Cd- and ZnBlm A2 are almost identical throughout the entire molecule, suggesting that these structures adopt similar conformations. Nuclear magnetic resonance experiments with 113Cd-bleomycin have defined part of the metal-ligand environment of the molecule. Nitrogen atoms from the primary amine, pyrimidine, and imidazole are bound to 113Cd according to 13C spectra showing 113Cd-13C spin-spin couplings. Bound and free forms of the secondary amine nitrogen may be in equilibrium, as suggested by temperature-dependent 13C studies with Cd-bleomycin. In addition, a number of other carbon resonances are in chemical exchange over the temperature range 5-54 degrees C. The temperature dependence of the line widths of carbon atoms of Zn-bleomycin strongly resembles that of Cd-bleomycin. Examination of the 113Cd resonance as a function of temperature also supports the presence of at least two differently coordinated forms of cadmium in the molecule. According to the position of the 113Cd chemical shift, at most four nitrogen atoms are bound to Cd at low temperature. Titrations of 113Cd-bleomycin with chloride or acetate demonstrate that these anions can bind to major and minor forms of the structure and that a minor species exists which does not associate with chloride.

Iron requirement for cellular DNA damage and growth inhibition by hydrogen peroxide and bleomycin.

Studies with Euglena gracilis and HL-60 cells have assessed the need for intracellular iron in the mechanisms of inhibition of cell growth and DNA damage by H2O2 and bleomycin. Cell culture media were directly depleted of iron in order to deprive cells of nutrient iron. Major pools of cellular iron were reduced in both cell types. Nevertheless, iron bound in e.s.r.-observable haem protein and ribonucleotide diphosphate reductase in HL-60 cells was not decreased. In both control cell populations, there was a concentration-dependent reduction in proliferation and cell survival caused by H2O2. In comparison, the proliferation rates of both iron-deficient cell types were significantly less sensitive to H2O2. H2O2 caused concentration-dependent single-strand breakage in DNA in control HL-60 and Euglena gracilis cells. Iron deficiency reduced the amount of strand breaks in HL-60 cells at each concentration of H2O2 used. Single-strand breakage caused by H2O2 in Euglena gracilis was a direct function of the concentration of iron in which the cells had been grown. Growth inhibition and both single- and double-strand DNA damage caused by bleomycin were substantially reduced or eliminated in iron-deficient cells. Copper bleomycin behaved like metal-free bleomycin when assayed for the capacity to cause DNA damage in iron-normal and iron-deficient HL-60 cells. In contrast, iron bleomycin was equally active under the two conditions in these cells.

Low dietary iron prevents free radical formation and heart pathology of copper-deficient rats fed fructose.

Two studies were conducted to determine whether hepatic iron overload in rats fed fructose plays a role in the exacerbation of the signs associated with copper deficiency. When fed the adequate iron diet (50 micrograms Fe/g), copper-deficient rats fed either fructose or starch exhibited hepatic iron overload of similar magnitude. However, only livers of copper-deficient rats fed fructose exhibited the presence of high peaks associated with an iron compound detected by electron spin resonance. In addition, only copper-deficient rats fed fructose developed anemia, pancreatic atrophy, and heart hypertrophy with histopathologic changes, and they died prematurely of heart-related abnormalities. Lowering dietary iron from 50 micrograms/g to 30 micrograms/g was not sufficient to protect the animals against the pathologic consequences of copper deficiency. In contrast, the consumption of a fructose diet inadequate in both copper (0.6 micrograms/g) and iron (17 micrograms/g) resulted in the reduction of hepatic iron, which in turn caused the amelioration of the deficiency, compared with rats fed the adequate iron (50 micrograms/g) diet. None of these rats developed pancreatic atrophy, none exhibited myocardial lesions, and none died of the deficiency. Electron spin resonance spectra of their livers did not show the presence of free radicals. The data suggest that hepatic iron overload plays a role in the exacerbation of copper deficiency only when fructose diets are consumed.

Effect of gallium on the tyrosyl radical of the iron-dependent M2 subunit of ribonucleotide reductase.

Gallium, a pharmacologically important metal which resembles iron, was shown in previous studies to inhibit ribonucleotide reductase. To better understand its mechanism of action, we have examined the interaction of gallium with the iron-dependent M2 subunit of ribonucleotide reductase. In its active form, M2 contains an iron center and a tyrosyl free radical which is detectable by ESR spectroscopy. In the present study, cytoplasmic extracts prepared from murine leukemic L1210 cells after an 18-hr incubation with 960 microM gallium nitrate displayed a > 60% inhibition in their M2 tyrosyl radical ESR signal. However, this signal was restored within 15 min to levels greater than that of controls by the addition of increasing concentrations of ferrous ammonium sulfate. Gallium citrate added directly to cytoplasmic extracts from control cells also decreased the tyrosyl radical signal, an effect which could be reversed by iron. Immunoblot analysis revealed that incubation with gallium did not diminish the amount of M2 protein in cells, thus indicating that the decrease in the tyrosyl radical signal was not due to a decrease in cellular M2 content. In immunoprecipitation studies of 59Fe-labeled M2, gallium displaced 55-60% of the 59Fe incorporated into M2. Our studies suggest that gallium displaces iron from the M2 subunit of ribonucleotide reductase, resulting in a loss of the tyrosyl radical and an accumulation of inactive M2 within the cell.

Oxidation-reduction reactions in Ehrlich cells treated with copper-neocuproine.

The interaction of 2,9-dimethyl-1,10-phenanthroline (neocuproine or NC) and its copper complex with Ehrlich ascites tumor cells was studied. NC is frequently used as a negative control in studies of in vitro DNA degradation by copper phenanthroline and has also found use as a potential inhibitor of damage from oxidative stress in biological systems. NC inhibited Ehrlich cell growth in monolayer culture over 48 h treatment by 50% at 0.05 nmol/10(5) cells. Addition of 5- to 100-fold ratios of CuCl2 to NC (at 0.035 nmol NC/10(5) cells) produced progressively more growth inhibition. Addition of 1:0.5 ratios of NC to CuCl2 over the range of NC concentrations 0.08-0.2 nmol/10(5) cells/mL resulted in DNA single-strand breakage during 1-h treatments as measured by DNA alkaline elution. Concomitant addition of catalase or dimethyl sulfoxide (DMSO) inhibited DNA strand scission, while superoxide dismutase enhanced breakage. Catalase and DMSO also inhibited induction of membrane permeability by the copper complex of NC. These cellular effects apparently result from the intracellular generation of hydroxyl radical from H2O2. NC facilitated the uptake of copper into cells, though it was initially bound as a copper-histidine-like complex. The internalized copper was reduced to Cu(I), bound mostly as (NC)2Cu(I). To explain the (NC)2Cu-dependent generation of hydroxyl radical, it is hypothesized that glutathione successfully competes for Cu(I), converting it to a redox-active form that can catalyze the reduction of molecular oxygen to .OH. Model studies support this view. Radical scavengers did not reverse growth inhibition produced by NC or NC + CuCl2.

Properties of redox-inactivated bleomycins. In vitro DNA damage and inhibition of Ehrlich cell proliferation.

Blenoxane, bleomycin A2, bleomycin B2, and demethyl bleomycin A2 and products of their reactions with Fe2+ and oxygen were used to explore the relationship between their capacity to carry out in vitro DNA strand scission and their growth inhibitory activity against Ehrlich cells. Reaction of Fe2+, bleomycin and O2 in the absence of DNA decreased the subsequent effectiveness of various bleomycin congeners to degrade DNA in the presence of Fe2+ and oxygen. In comparison with controls, this loss of strand scission activity was not paralleled by equivalent decreases in growth inhibition. Demethyl bleomycin A2 retained full biological activity relative to bleomycin A2, despite being only 30% as effective as bleomycin A2 in its ability to cleave DNA in vitro. Prior reaction of bleomycins with Fe2+ did not alter their capacity to reduce oxygen or affect their ability to generate the activated intermediate which, for native bleomycin structures, is competent to cleave DNA in vitro.

Interactions of 1,10-phenanthroline and its copper complex with Ehrlich cells.

Mechanistic details of the interaction of 1,10-phenanthroline and its copper complex with Ehrlich ascites tumor cells were examined, using inhibition of cell proliferation, DNA breakage, and increased membrane permeability as indices of cellular damage. The metal chelating agent, 1,10-phenanthroline (OP), the 1:0.5 complex of 1,10-phenanthroline and CuCl2 [(OP)2Cu], and CuCl2 inhibited growth of Ehrlich ascites tumor cell monolayers during 48-h treatments by 50% at about 3.5, 2, and 70 nmol/10(5) cells/mL, respectively. (OP)2Cu at 10 nmol/10(5) cells also enhanced uptake of trypan blue dye during 6 h of treatment, while dye uptake in OP- and CuCl2-treated cells remained similar to controls. DNA breakage, measured by DNA alkaline elution, was produced during 1-h treatments with (OP)2Cu at drug/cell ratios similar to those producing growth inhibition. Copper uptake was similar for both (OP)2Cu and CuCl2. Electron spin resonance (ESR) spectroscopy suggested that cellular ligands bind copper added as (OP)2Cu or CuCl2 and then undergo time-dependent reductions of Cu(II) to Cu(I) for both forms. Inhibition of (OP)2Cu-induced single-strand scission and trypan blue uptake by scavengers of activated oxygen is consistent with participation of superoxide and H2O2 in both processes. In contrast, superoxide dismutase (SOD) did not reduce the magnitude of the fraction of cellular DNA appearing in lysis fractions prior to alkaline elution of (OP)2Cu-treated cells. Dimethyl sulfoxide (DMSO) inhibited uptake of trypan blue dye but did not inhibit DNA strand scission produced by (OP)2Cu. Thus, multiple mechanisms for generation of oxidative damage occur in (OP)2Cu-treated cells. Growth inhibition produced by OP or (OP)2Cu, as well as the low levels of strand scission produced by OP, was not reversed by scavengers.

Inhibition of iron uptake in HL60 cells by 2-formylpyridine monothiosemicarbazonato Cu(II).

The electron paramagnetic resonance (EPR) signal of the tyrosyl radical attributed to ribonucleoside diphosphate reductase decreases after treatment of promyelocytic leukemic HL60 cells with 2-formylpyridine thiosemicarbazonato copper(II) (CuL). According to EPR studies, CuL forms adducts with both histidine and cysteine-like Lewis bases associated with isolated membranes from HL60 cells. After the addition of CuL, the EPR signal for the cysteine-like adduct decreases substantially over a 15-min period. The reduction of signal is consistent with oxidation of thiols as shown by an analysis of sulfhydryl content. It is hypothesized that receptor-mediated transferrin internalization is inhibited by oxidation of critical thiols. Since the uptake of 59Fe-transferrin is greatly inhibited by the treatment of HL60 cells with CuL, the reduced uptake of iron by cells, in the presence of CuL, may lead to decreased iron availability for the activity of the M2 subunit of ribonucleotide reductase and a subsequent decrease in the tyrosyl radical signal of the enzyme. Moreover, the intact subunit M2 is no longer detected by EPR, even after the addition of excess iron.

Spin-trapping studies of the oxidation-reduction reactions of iron bleomycin in the presence of thiols and buffer.

The reaction of ferrous bleomycin with dioxygen is reexamined to clarify whether radical species derived from molecular oxygen are generated. Detection of low levels of spin-trapped oxyradicals confirm the production of OH during this reaction when bleomycin is present in excess, but not when iron and drug concentrations are equal. In phosphate buffer, hydroxyl radicals continue to be spin trapped for at least 15 min after Fe(II)bleomycin has been oxidized to Fe(III)bleomycin. In HEPES buffer, detection of a HEPES radical in the absence of spin trap over the same period independently supports the conclusion that reactive radicals are present after the initial oxidation of Fe(II)bleomycin is complete. When glutathione is included in the aerobic reaction mixture, thiyl radical species are spin trapped. The reaction of Fe(III)bleomycin with cysteine produces thiyl radical without spin-trapped hydroxyl radical.

Development of drug resistance to gallium nitrate through modulation of cellular iron uptake.

We have shown that transferrin-gallium (Tf-Ga) blocks DNA synthesis through inhibition of cellular iron incorporation and a diminution in the activity of the iron-dependent M2 subunit of ribonucleotide reductase. To examine the mechanisms of drug resistance to gallium, we developed a subline of HL60 cells (R cells) which is 29-fold more resistant to growth inhibition by gallium nitrate than the parent line (S cells). R cells displayed a 2.5-fold increase in transferrin (Tf) receptor expression, without a change in receptor affinity for Tf. The uptake and release of 67Ga were similar for both S and R cells. The uptake of 59Fe-Tf by S cells was inhibited by gallium nitrate over 24-48 h of incubation. In contrast, 59Fe-Tf uptake by R cells, although initially inhibited by gallium nitrate at 24 h, was no longer inhibited at 48 h of incubation. 59FeCl3 uptake by R cells was significantly greater than that of S cells, regardless of the time in culture. Despite the increase in 59Fe uptake by R cells, the ferritin content of these cells was lower than that of S cells. The ribonucleotide reductase electron spin resonance signal of R cells was comparable to that of S cells. R cells were not cross-resistant to Adriamycin, vincristine, cis-platinum or hydroxyurea. Resistance to gallium nitrate in this subline of HL60 cells results primarily from the ability of cells to overcome the gallium-induced block in iron incorporation. In addition, intracellular iron in R cells appears to traffic preferentially to a non-ferritin compartment.