DNA Damage by Dimethylformamide: Role of Hydrogen Peroxide

The enhancement of oxidative DNA damage by anti-diabetic metformin, buformin, and phenformin, via nitrogen-centered radicals. Shiho Ohnishi , Hideki ...
9 downloads 0 Views 245KB Size
Download PDF
Chem. Res. Toxicol. 2000, 13, 309-315

309

DNA Damage by Dimethylformamide: Role of Hydrogen Peroxide Generated during Degradation Kaoru Midorikawa, Mariko Murata, Shinji Oikawa, Saeko Tada-Oikawa, and Shosuke Kawanishi* Department of Hygiene, Mie University School of Medicine, Tsu, Mie 514-8507, Japan Received July 29, 1999

Dimethylformamide (DMF) has been suspected to associate with cancers in exposed workers, whereas there has been inadequate evidence for carcinogenicity in experimental animals. We demonstrated that H2O2 was generated during the degradation of DMF under aerobic conditions, and that the amount of H2O2 was enhanced by exposure to solar light or by the contamination of trace metal. Experiments using 32P-5′-end-labeled DNA fragments revealed that the degraded DMF induced DNA damage in the presence of Cu(II). However, purified DMF did not induce DNA damage even in the presence of Cu(II). Addition of purified DMF enhanced DNA damage induced by H2O2 in the presence of Cu(II). The degraded DMF caused Cu(II)-mediated DNA cleavage frequently at thymine and cytosine residues. The similar pattern of site-specific DNA damage was observed with purified DMF and H2O2. Bathocuproine and catalase inhibited the DNA damage, indicating the involvement of Cu(I) and H2O2. A typical free hydroxy radical scavenger showed no inhibitory effect on the DNA damage. Addition of purified DMF enhanced about 3-4-fold 8-oxo-7,8-dihydro-2′-deoxyguanosine formation induced by H2O2 and Cu(II). ESR spectroscopic study demonstrated that carbon-centered radicals and nitrogen-centered radicals were generated in the reaction mixture of DMF, H2O2, and Cu(II). Inhibitory effects of scavengers on radical formation and DNA damage suggest that carboncentered radicals and/or nitrogen-centered radicals may contribute to the DNA damage. These results suggest that H2O2 generation during DMF degradation is related to the possible carcinogenic activity of DMF.

Introduction Dimethylformamide (DMF)1 is widely used as a solvent. Epidemiological studies showed an excess risk for testicular germ-cell tumors among workers exposed to DMF in industries (1-4). Excess risks for cancers of the buccal cavity and pharynx were observed in workers exposed to DMF at plants (5, 6). Major et al. (7) reported significant increases in chromosome aberration and sister chromatid exchange frequencies in workers exposed to DMF. On the other hand, there has been inadequate evidence for the carcinogenicity of DMF in experimental animals (8). The discrepancy of carcinogenic potency of DMF between an experimental system and humans remains to be clarified. This discrepancy of DMF carcinogenicity may be explained by the difference between degraded DMF in the industries and purified DMF used in the laboratories. It is known that DMF begins to degrade when exposed to ultraviolet radiation or strong sunlight with the formation of dimethylamine and formaldehyde (9, 10). We found H2O2 generation during DMF degradation under several conditions. To clarify the role of H2O2 in * To whom request for reprints should be addressed. E-mail: [email protected]. 1 Abbreviations: DMF, N,N-dimethylformamide; HPLC, highperformance liquid chromatograph; HPLC-ECD, HPLC equipped with an electrochemical detector; SOD, superoxide dismutase; 8-oxodG, 8-oxo-7,8-dihydro-2′-deoxyguanosine; ESR, electron spin resonance; DTPA, diethylenetriamine-N,N,N′,N′′,N′′-pentaacetic acid; POBN, R-(4pyridyl 1-oxide)-N-tert-butylnitrone.

DNA damage by DMF, we investigated metal-mediated DNA damage by purified DMF with H2O2 using 32P-5′end-labeled DNA fragments obtained from the human c-Ha-ras-1 protooncogene and the p53 tumor suppressor gene. These two genes are known to be the targets for chemical carcinogens (11). We also measured the content of 8-oxo-7,8-dihydro-2′-deoxyguanosine (8-oxodG) in calf thymus DNA using a high-performance liquid chromatograph equipped with an electrochemical detector (HPLCECD). Formation of 8-oxodG is a relevant indicator of oxidative base damage and causes DNA misreplication that might lead to mutation or cancer (12). Furthermore, to clarify the mechanism of the DNA damage, we examined the generation of radicals from DMF by electron spin resonance (ESR) spectroscopy.

Materials and Methods Materials. Restriction enzymes (BamHI, AvaI, XbaI, and PstI) and T4 polynucleotide kinase were purchased from New England Biolabs (Beverly, MA). [γ-32P]ATP (222 TBq/mmol) was from New England Nuclear (Boston, MA). Calf intestine phosphatase was from Boehringer Mannheim GmbH (Mannheim, Germany). Diethylenetriamine-N,N,N′,N′′,N′′-pentaacetic acid (DTPA) and bathocuproinedisulfonic acid were from Dojin Chemicals Co. (Kumamoto, Japan). Cu(I) chloride, Cu(II) chloride dihydrate, ferrous ammonium sulfate hexahydrate, Fe(III) chloride, ethanol, and scopoletin were from Nacalai Tesque, Inc. (Kyoto, Japan). Calf thymus DNA, SOD (3000 units/mg from bovine erythrocytes), and catalase (45 000 units/mg from bovine liver) were from Sigma Chemical Co. (St. Louis, MO). Nuclease P1 was from Yamasa Shoyu Co. (Chiba, Japan). Horseradish

10.1021/tx990139r CCC: $19.00 © 2000 American Chemical Society Published on Web 03/28/2000

310

Chem. Res. Toxicol., Vol. 13, No. 4, 2000

peroxidase was from Toyobo Co. (Osaka, Japan). R-(4-Pyridyl 1-oxide)-N-tert-butylnitrone (POBN) and methional were from Tokyo Kasei Co. (Tokyo, Japan). DMF, acrylamide, bisacrylamide, piperidine, and H2O2 were from Wako Chemicals Co. (Osaka, Japan). DMF of the highest purity available packaged under nitrogen was used as purified DMF. Degraded DMF was prepared from purified DMF by being exposed to solar light for 2 weeks under air, or storage under air in the dark for more than 2 years. Detection of H2O2 Generation during the Degradation of DMF. The amounts of H2O2 in DMF were determined by assessing the extinction of scopoletin fluorescence during its oxidation by horseradish peroxidase (13-15). Reactions were performed in cuvettes containing 10 µM scopoletin and 5% (v/ v) DMF in 10 mM phosphate buffer (pH 7.4) at 37 °C. The reaction was initiated by the addition of 1 µM horseradish peroxidase. Fluorescence was measured with a spectrofluorometer (RF-5300PC spectrofluorophotometer, Shimadzu) with sample excitation at 365 nm and emission at 450 nm. Preparation of 32P-5′-End-Labeled DNA Fragments. DNA fragments were obtained from the human p53 tumor suppressor gene (16) and the c-Ha-ras-1 protooncogene (17). The DNA fragment of the p53 tumor suppressor gene was prepared from the pUC18 plasmid. A singly 32P-5′-end-labeled 443 bp fragment (ApaI 14179-EcoRI* 14621) and a 211 bp fragment (HindIII* 13972-ApaI 14182) were obtained according to the method described previously (18). DNA fragments were also prepared from plasmid pbcNI, which carries a 6.6 kb BamHI chromosomal DNA segment containing the human c-Ha-ras-1 protooncogene (19-21). A singly labeled 337 bp fragment (PstI 2345-AvaI* 2681) and a 261 bp fragment (AvaI* 1645-XbaI 1905) were obtained according to the method described previously (19-21). The asterisk indicates 32P-labeling. Detection of Damage to 32P-5′-End-Labeled DNA. The reaction mixture in a 1.5 mL microtube contained the indicated concentrations of DMF, H2O2, CuCl2, 32P-5′-end-labeled DNA fragment, and calf thymus DNA in 200 µL of 10 mM sodium phosphate buffer (pH 7.8) containing 5 µM DTPA. After incubation at 37 °C for the indicated period of time, the DNA fragments were heated at 90 °C in 1 M piperidine for 20 min where indicated and treated as described previously (19). The preferred cleavage sites were identified by direct comparison of the positions of the oligonucleotides with those produced by the chemical reactions of the Maxam-Gilbert procedure (22) using a DNA sequencing system (LKB 2010 Macrophor). A laser densitometer (LKB 2222 UltroScan XL) was used for the measurement of the relative amounts of oligonucleotides from treated DNA fragments. Analysis of 8-OxodG Formation in Calf Thymus DNA. Calf thymus DNA (50 µM/base) was incubated with indicated concentrations of H2O2 and DMF in the presence of CuCl2 in 4 mM sodium phosphate buffer (pH 7.8) at 37 °C for the indicated period of time. After ethanol precipitation, the DNA was digested to the nucleosides by incubation with nuclease P1 and alkaline phosphatase and analyzed with an HPLC-ECD as described previously (23). ESR Spectra Measurements. ESR spectra were measured at room temperature (25 °C) by using a JES-TE-100 (JEOL, Tokyo, Japan) spectrometer with 100 kHz field modulation according to the method described previously (24). Spectra were recorded with a microwave power of 4 mW, a modulation amplitude of 0.050 mT, a receiver gain of 500, a time constant of 1 s, and a sweep time of 15 min. The magnetic fields were calculated by the splitting of Mn(II) in MgO (∆H3-4 ) 8.69 mT). POBN was used as the radical trapping reagent. The ESR software [Isotropic EPR Simulation, version 2.2 A (Labotec Co., Ltd.)] was used for analysis and simulation of ESR data.

Results H2O2 Formation during the Degradation of DMF. H2O2 was formed during the degradation of DMF (Figure

Midorikawa et al.

Figure 1. Amount of H2O2 formed from DMF under several conditions. The amounts of H2O2 generated from DMF were determined by assessing the extinction of scopoletin fluorescence during oxidation of the fluorophor by horseradish peroxidase. The reactions were performed in cuvettes containing 10 µM scopoletin and 5% (v/v) DMF in 10 mM phosphate buffer (pH 7.4) at 37 °C, and were initiated by the addition of 1 µM horseradish peroxidase. Fluorescence was measured with an RF5300PC fluorescence spectrophotometer with sample excitation at 365 nm and emission at 450 nm: control, aerobic conditions in the dark; Fe(III), 10 mM FeCl3 added under aerobic conditions in the dark; Cu(II), 10 mM CuCl2 added under aerobic conditions in the dark; Light, under aerobic conditions and exposed to solar light during the daytime; and Light (anaerobic), under anaerobic conditions and exposed solar light during the daytime.

1). The extent of H2O2 formation increased time-dependently under aerobic conditions for 4 weeks. A large amount of H2O2 was detected in DMF exposed to solar light. In contrast, under anaerobic conditions, a very small amount of H2O2 was detected in DMF even after it had been exposed to solar light. Among metal ions, addition of Cu(II) enhanced H2O2 generation. When DMF stored in the dark under aerobic conditions for more than 2 years, H2O2 was detected at a concentration of ∼75 µM (data not shown). Damage to 32P-Labeled DNA Fragments Induced by Degraded DMF in the Presence of Metal Ions. The extent of damage to isolated DNA induced by degraded DMF or purified DMF in the presence of Cu(II) was estimated by gel electrophoretic analysis (Figure 2). Oligonucleotides were detected on the autoradiogram as a result of DNA damage. Degraded DMF caused DNA damage in the presence of Cu(II). The intensity of DNA damage increased depending on concentrations of degraded DMF. Purified DMF alone did not caused DNA damage in the presence of Cu(II). Effect of Purified DMF on DNA Damage Induced by H2O2 and Cu(II). H2O2 caused DNA damage in the presence of Cu(II). The level of DNA damage increased with increasing concentrations of H2O2 (Figure 3A). Addition of purified DMF markedly enhanced the DNA damage induced by H2O2 and Cu(II) (Figure 3B). The extent of DNA damage also increased with increasing concentrations of purified DMF (Figure 3C). When Cu(I) was used instead of Cu(II), similar results were obtained. However, no DNA damage was observed when Fe(III) was added at the same concentration (data not shown). This suggests that Cu(II) reacts more efficiently with H2O2 than Fe(III), and therefore, copper ions medi-

DNA Damage by Dimethylformamide

Figure 2. Autoradiogram of the 32P-labeled DNA fragment incubated with degraded DMF or purified DMF in the presence of Cu(II). The reaction mixture contained the 32P-5′-end-labeled 337 bp DNA fragment, sonicated calf thymus DNA (10 µM/base), the indicated concentrations of DMF, and 20 µM CuCl2 in 200 µL of 10 mM sodium phosphate buffer (pH 7.8) containing 5 µM DTPA and was incubated at 37 °C for 1 h. After piperidine treatment, the DNA fragments were electrophoresed, and the autoradiogram was obtained. Degraded DMF contained about 500 µM H2O2.

Figure 3. Autoradiogram of 32P-labeled DNA fragments incubated with H2O2 and Cu(II) in the presence and absence of purified DMF. The reaction mixture contained the 32P-5′-endlabeled 337 bp DNA fragment (A and B) and the 443 bp DNA fragment (C), sonicated calf thymus DNA (10 µM/base), 20 µM CuCl2, and the indicated concentrations of H2O2 (A), the indicated concentrations of H2O2 with 10% (v/v) purified DMF (B), or 50 µM H2O2 with the indicated concentrations of purified DMF (C) in 200 µL of 10 mM sodium phosphate buffer (pH 7.8) containing 5 µM DTPA. The mixture was incubated for 1 h at 37 °C, and treated via the method described in the legend of Figure 2.

ate more extensive DNA damage under the same conditions (25). Effects of Scavengers and a Metal Chelator on the DNA Damage Induced by Degraded DMF in the Presence of Cu(II). Catalase and bathocuproine inhib-

Chem. Res. Toxicol., Vol. 13, No. 4, 2000 311

Figure 4. Effects of scavengers and a metal chelator on Cu(II)-dependent DNA damage induced by degraded DMF or purified DMF and H2O2. The reaction mixture contained the 32P-5′-end-labeled 211 bp fragment, sonicated calf thymus DNA (10 µM/base), 20 µM CuCl2, and 5% (v/v) degraded DMF or purified DMF with 25 µM H2O2 in 200 µL of 10 mM sodium phosphate buffer (pH 7.8) containing 5 µM DTPA. The mixture was incubated for 1 h at 37 °C, and treated via the method described in the legend of Figure 2. The concentrations of scavengers and metal chelator were as follows: 30 units of catalase, 30 units of SOD, 0.1 M methional, 5% (v/v) ethanol, and 50 µM bathocuproine. Degraded DMF contained about 500 µM H2O2.

ited DNA damage, suggesting the involvement of H2O2 and Cu(I). Methional which scavenges not only the free •OH but also crypto-OH radicals (26, 27) also inhibited DNA damage. SOD and typical free •OH scavenger (ethanol) did not inhibit DNA damage (Figure 4A). Similar results were obtained with purified DMF and H2O2 in the presence of Cu(II) (Figure 4B). The same amount of H2O2 in degraded DMF was added to purified DMF. Comparison of the Site Specificity of Cu(II)Dependent DNA Damage by Degraded DMF with That by Purified DMF and H2O2. The DNA cleavage sites were identified by direct comparison of the positions of the oligonucleotides with these produced by the chemical reactions of the Maxam-Gilbert procedure (22). The 32P-5′-end-labeled DNA fragment treated with H O , 2 2 degraded DMF, or purified DMF and H2O2, in the presence of Cu(II), followed by piperidine treatment, was electrophoresed, and an autoradiogram was obtained (Figure 5). Purified DMF enhanced DNA damage induced by H2O2 and Cu(II). The DNA cleavage pattern of purified DMF was similar to that of degraded DMF. For the measurement of the relative intensity of DNA cleavage, the autoradiograms were scanned with a laser densitometer (Figures 6 and 7). The degraded DMF (Figure 6A) induced DNA cleavage frequently at thymine and cytosine residues in the presence of Cu(II). A similar DNA cleavage pattern was observed when purified DMF and H2O2 were used (Figure 6B). H2O2 and Cu(II) weakly induced DNA cleavage at cytosine and guanine residues

312

Chem. Res. Toxicol., Vol. 13, No. 4, 2000

Midorikawa et al.

Figure 6. Site specificity of Cu(II)-dependent DNA cleavage induced by degraded DMF or purified DMF and H2O2. The reaction mixture containing the 32P-5′-end-labeled 337 bp DNA fragment, calf thymus DNA (10 µM/base), 20 µM CuCl2, and 20% (v/v) degraded DMF (A) or 10% (v/v) purified DMF and 50 µM H2O2 (B) in 10 mM phosphate buffer (pH 7.8) containing 5 µM DTPA was incubated for 1 h at 37 °C. After piperidine treatment, DNA cleavage sites were identified as described in Materials and Methods. The horizontal axis shows the nucleotide number of the c-Ha-ras-1 protooncogene (17). Degraded DMF contained about 75 µM H2O2.

Figure 5. Site specificity of Cu(II)-dependent DNA cleavage induced by degraded DMF or purified DMF and H2O2. The reaction mixture containing the 32P-5′-end-labeled 337 bp DNA fragment, calf thymus DNA (10 µM/base), 20 µM CuCl2, and 20% (v/v) degraded DMF or 10% (v/v) purified DMF with 50 µM H2O2 or 50 µM H2O2, in 10 mM phosphate buffer (pH 7.8) containing 5 µM DTPA was incubated for 1 h at 37 °C, and treated via the method described in the legend of Figure 2. Lanes G + A and T + C represent the patterns obtained for the same fragment after cleavage with piperidine treatment by the chemical methods of Maxam and Gilbert (22). Degraded DMF contained about 75 µM H2O2.

under the condition that was used (Figure 7A). Addition of purified DMF induced extensive DNA cleavage and especially enhanced cleavage at thymine and cytosine residues (Figure 7B). Formation of 8-OxodG by Purified DMF in the Presence of H2O2 and Cu(II). Using an HPLC-ECD, we measured the 8-oxodG content in calf thymus DNA treated with various concentrations of H2O2 and Cu(II) in the presence and absence of purified DMF (Figure 8). The amount of Cu(II)-dependent 8-oxodG formation increased with increasing concentrations of H2O2. The addition of purified DMF enhanced 3-4-fold 8-oxodG formation induced by H2O2 and Cu(II). Formation of the Radicals from Purified DMF and H2O2 in the Presence of Cu(II). An ESR spectrum of a spin adduct was observed when POBN was added to a mixture solution of purified DMF, H2O2, and Cu(II) (Figure 9A-a). It is estimated by reference to the reported constants (28) that this complex spectrum may consist of two components. The first component is the carboncentered adduct of POBN with the following splitting constants: aN ) 1.49 mT and aH ) 0.23 mT; the second overlapping component is the nitrogen-centered adduct of POBN with the following splitting constants: aN ) 1.49 mT and aH(β) ) aN(β) ) 0.21 mT. When DMF was omitted, the signals were not observed. Addition of catalase, methional, or bathocuproine decreased the yield of the radical adducts of POBN (Figure 9A-b, -d, and -f). SOD and ethanol did not inhibit them (Figure 9A-c and -e). It was verified by computer simulation that the signals

DNA Damage by Dimethylformamide

Chem. Res. Toxicol., Vol. 13, No. 4, 2000 313

Figure 7. Site specificity of Cu(II)-dependent DNA cleavage by H2O2 or purified DMF and H2O2. The reaction mixture containing the 32P-5′-end-labeled 211 bp DNA fragment, calf thymus DNA (10 µM/base), 20 µM CuCl2, and 50 µM H2O2 (A) or 10% (v/v) purified DMF and 50 µM H2O2 (B) in 10 mM phosphate buffer (pH 7.8) containing 5 µM DTPA was incubated for 1 h at 37 °C. After piperidine treatment, DNA cleavage sites were identified as described in Materials and Methods. The horizontal axis shows the nucleotide number of the human p53 tumor suppressor gene (16). Figure 9. ESR spectra of radicals derived from purified DMF in the presence of H2O2 and Cu(II) and effects of scavengers and a metal chelator on the radicals. (A) The reaction mixture contained 400 mM POBN, 20% (v/v) purified DMF, 1 mM H2O2, and 20 µM CuCl2 in 100 µL of 10 mM Tris-HCl buffer (pH 7.8) containing 5 µM DTPA, and the spectra were recorded immediately after mixing had occurred by ESR spectroscopy: (Aa) no scavenger, (A-b) 30 units of catalase, (A-c) 30 units of SOD, (A-d) 0.1 M methional, (A-e) 5% (v/v) ethanol, and (A-f) 500 µM bathocuproine. (B) Computer simulation of the experimental spectra (B-a): 67% of a spectrum with hyperfine splitting carbon-centered radicals and 33% nitrogen-centered radicals. Computer simulation with the following coupling constants: (Bb) aN ) 1.49 mT and aH ) 0.23 mT; (B-c) aN ) 1.49 mT and aH(β) ) aN(β) ) 0.21 mT. (B-a) The mixture of spectra B-b and B-c with a 2:1 proportion, using Isotropic EPR Simulation, version 2.2 A (Labotec Co., Ltd.). Figure 8. Cu(II)-dependent formation of 8-oxodG by H2O2 or purified DMF and H2O2. The reaction mixture contained calf thymus DNA (50 µM/base), 20 µM CuCl2, and the indicated concentrations of H2O2 (0) or H2O2 and 10% (v/v) purified DMF (O) in 400 µL of 4 mM phosphate buffer (pH 7.8) containing 5 µM DTPA. After incubation for 1 h at 37 °C, the DNA fragment was enzymatically digested into nucleosides, and 8-oxodG formation was assessed with an HPLC-ECD as described in Materials and Methods.

degrade when it is exposed to ultraviolet radiation or strong solar light with the formation of dimethylamine, formaldehyde, and the related free radicals (9, 10). On the basis of these findings and the previous reports (9, 10), we propose possible mechanisms of the formation of H2O2 during degradation of DMF.

HCON(CH3)2 f •CON(CH3)2 + •H

were assigned as a mixture of carbon-centered radicals (Figure 9B-b) and nitrogen-centered radicals (Figure 9Bc) with the proportion being 2:1 (Figure 9B-a). The computer simulation pattern provided a satisfactory fit of the signals obtained in this study.

•

H + O2 f O2•- + H+

(1-2a)

2H+ + 2O2•- f H2O2 + O2

(1-3)

•

Discussion This study revealed that H2O2 was formed in DMF under aerobic conditions and that the amount of H2O2 formation increased depending on time. Exposure to solar light or contamination of trace metal enhanced H2O2 formation in DMF. It is known that DMF begins to

(1-1a)

CON(CH3)2 f CO + •N(CH3)2

(2)

•

N(CH3)2 + HCON(CH3)2 f HN(CH3)2 + •CON(CH3)2 (3)

The formation of carbon-centered radicals and O2•- is efficiently enhanced in DMF by exposure to UV light

314

Chem. Res. Toxicol., Vol. 13, No. 4, 2000

Midorikawa et al.

(reactions 1-1a and 1-2a). Then, H2O2 is formed by dismutation of O2•- (reaction 1-3). The carbon-centered radicals dissociate into CO and •N(CH3)2 (reaction 2). The nitrogen-centered radicals react with DMF itself to form dimethylamine [HN(CH3)2] and the carbon-centered radicals again (reaction 3). Similarly, a possible mechanism for enhancement of carbon-centered radical and O2•formation by Cu(II) can be proposed as follows.

HCON(CH3)2 + Cu(II) f •

CON(CH3)2 + Cu(I) + H+ (1-1b)

Cu(I) + O2 f C(II) + O2•-

(1-2b)

Using ESR spin trapping methods, we detected carboncentered radicals and nitrogen-centered radicals from DMF induced by H2O2 and Cu(II). It was previously suggested that the radical •CON(CH3)2 was generated from DMF by the addition of H2SO4, H2O2, and FeSO4 (29). The mechanism for the formation of these radicals can be envisioned as follows. Since Cu(I)OOH formed from H2O2 and Cu(II) has the potential of •H abstraction like •OH or UV irradiation, •CON(CH3)2 can be formed from DMF (reaction 1-1a), and the radical can form •N(CH3)2 (reaction 2). Alternatively, there was a report that • OH reacted with DMF to form HCON(C4 H2)CH3 (30). However, it is generally accepted that the abstraction of a hydrogen from the CH3 group is more difficult than that from the CHO group (29). Therefore, it is reasonably considered that •CON(CH3)2 is more likely to be a candidate of the formed carbon-centered radicals than HCON(C4 H2)CH3. This study demonstrated that degraded DMF caused Cu(II)-mediated DNA damage, but purified DMF did not. Purified DMF markedly enhanced DNA damage by H2O2/ Cu(II). The amount of 8-oxodG formation also increased with the addition of purified DMF to the reaction system of H2O2 and Cu(II). Cu(II)-dependent DNA damage by degraded DMF was inhibited by catalase and bathocuproine. Similar scavenger effects were observed with purified DMF and H2O2/Cu(II). These results have suggested that Cu(I)OOH obtained with H2O2 and Cu(II) participates in the DNA damage. Typical free •OH scavenger showed no inhibitory effect on the DNA damage. Usually, there is no site specificity in free •OHmediated DNA damage (24, 31). DMF induced the sitespecific DNA damage mainly at thymine and cytosine residues in the presence of H2O2 and Cu(II), suggesting that free •OH might not play an important role. Carboncentered radicals and nitrogen-centered radicals detected by ESR in the reaction mixture of DMF, H2O2, and Cu(II) were suppressed by catalase and bathocuproine. The same results were obtained for the inhibitory effects of scavengers on DNA damage. Therefore, we speculate that in addition to a complex of Cu(I) with H2O2, these radicals may participate in the DNA damage. Relevantly, it was reported that carbon-centered radicals were responsible for the site-specific damage at cytosine residues (32). However, carbon-centered radicals seem not to be the main species to react with DNA, because N-methylformamide (a metabolite of DMF) did not enhance DNA damage by H2O2 and Cu(II) despite generating carboncentered radicals (data not shown). We previously reported that nitrogen-centered radicals caused guaninespecific modification (33). On the basis of the result of

site specificity, it is considered that nitrogen-centered radicals from DMF may not participate in the DNA damage. There is a possibility that DMF affects the conformation of Cu(II)-bound DNA, resulting in enhancement of DNA damage by H2O2 and Cu(II). We examined whether this is a DMF-specific effect or not, using other organic solvents instead of DMF. Ethanol, methanol, dimethyl sulfoxide, and N-methylformamide did not enhance DNA damage by H2O2 and Cu(II) (data not shown). This result weakens the possibility of the common organic solvent effect which affects the conformation of Cu(II)-bound DNA. The enhancing effect of DMF on DNA damage appears quite specific among these organic solvents. As mentioned above, DNA damage observed in our experiments in vitro was due to reactive species generated from Cu(I) and H2O2, likely Cu(I)OOH. However, there is little evidence that carbon-centered radicals and nitrogen-centered radicals from DMF can exist as reactive species not only in vivo but also in vitro. H2O2 formation in degraded DMF and metal-mediated DNA damage are primarily interesting chemical processes that could potentially be related to the possible carcinogenic activity of degraded DMF. There is not yet clear evidence of carcinogenicity in experimental animals. On the other hand, excess risks for cancers in workers exposed to DMF have been demonstrated. This discrepancy of DMF carcinogenicity in experimental animals and humans can be explained with the following hypothesis. The formation of H2O2 from degraded DMF used in the industries might be concerned. The labor environment in the 1980s had not been sufficiently arranged in industries (1-4). The DMF to which the workers were exposed might be degraded and contain H2O2. Ito et al. (34) reported that H2O2 caused carcinomas in mice. The relatively low activity of testicular catalase (35) might predispose the testis to H2O2-mediated oxidative stress. It may be concluded that purified DMF itself is not carcinogenic, but H2O2 formation during DMF degradation plays a part in the expression of possible carcinogenicity of degraded DMF.

References (1) Ducatman, A. M., Conwill, D. E., and Crawl, J. (1986) Germ cell tumors of the testicle among aircraft repairmen. J. Urol. 136, 834-836. (2) Levin, S. M., Baker, D. B., Landrigan, P. J., Monagham, S. V., Frumin, E., Braithwaite, M., and Towne, W. (1987) Testicular cancer in leather tanners exposed to dimethylformamide. Lancet 2, 1153. (3) Calvert, G. M., Fajen, J. M., Hills, B. W., and Halperin, W. (1990) Testicular cancer, dimethylformamide, and leather tanneries. Lancet 336, 1253-1254. (4) Ducatman, A. M. (1989) Dimethylformamide, metal dyes, and testicular cancer. Lancet 1, 911. (5) Chen, J. L., Fayerweather, W. E., and Pell, S. (1988) Cancer incidence of workers exposed to dimethylformamide and/or acrylonitrile. J. Occup. Med. 30, 813-818. (6) Chen, J. L., Fayerweather, W. E., and Pell, S. (1988) Mortality study of workers exposed to dimethylformamide and/or acrylonitrile. J. Occup. Med. 30, 819-821. (7) Major, J., Huda´k, A., Kiss, G., Jakab, M. G., Szaniszlo´, J., Na´ray, M., Nagy, I., and Tompa, A. (1998) Follow-up biological and genotoxicological monitoring of acrylonitrile- and dimethylformamide-exposed viscose rayon plant workers. Environ. Mol. Mutagen. 31, 301-310. (8) IARC Working Group (1999) Dimethylformamide. In IARC Monographs on the Evaluation of Carcinogenic Risks to Humans, Vol. 71, pp 545-574, International Agency for Research on Cancer, Lyon, France.

DNA Damage by Dimethylformamide (9) Bolton, R., Miller, J., and Parker, A. J. (1963) The reaction of p-fluoronitrobenzene in dimethylformamide with sodium azide (further observations) and with dimethylamine. Chem. Ind., 492493. (10) Pukhal’skaya, G. V., Kotov, A. G., and Pshezhetski, S. Ya. (1969) The photochemistry of free radicals: the action of light on radicals in γ-irradiated formamide and dimethylformamide. Khim. Vys. Energ. 3, 340-345. (11) Denissenko, M. F., Pao, A., Tang, M., and Pfeifer, G. P. (1996) Preferential formation of benzo[a]pyrene adducts at lung cancer mutational hotspots in p53. Science 274, 430-432. (12) Shibutani, S., Takeshita, M., and Grollman, A. P. (1991) Insertion of specific bases during DNA synthesis past the oxidationdamaged base 8-oxodG. Nature 349, 431-434. (13) Root, R. K., Metcalf, J., Oshino, N., and Chance, B. (1975) H2O2 release from human granulocytes during phagocytosis. I. Documentation, quantitation, and some regulating factors. J. Clin. Invest. 55, 945-955. (14) Starkebaum, G., and Harlan, J. M. (1986) Endothelial cell injury due to copper-catalyzed hydrogen peroxide generation from homocysteine. J. Clin. Invest. 77, 1370-1376. (15) Stamler, J. S., Osborne, J. A., Jaraki, O., Rabbani, L. E., Mullins, M., Singel, D., and Loscalzo, J. (1993) Adverse vascular effects of homocysteine are modulated by endothelium-derived relaxing factor and related oxides of nitrogen. J. Clin. Invest. 91, 308318. (16) Chumakov, P. (1990) EMBL Data Library, accession number X54156. (17) Capon, D. J., Chen, E. Y., Levinson, A. D., Seeburg, P. H., and Goeddel, D. V. (1983) Complete nucleotide sequences of the T24 human bladder carcinoma oncogene and its normal homologue. Nature 302, 33-37. (18) Yamashita, N., Murata, M., Inoue, S., Hiraku, Y., Yoshinaga, T., and Kawanishi, S. (1998) Superoxide formation and DNA damage induced by a fragrant furanone in the presence of copper(II). Mutat. Res. 397, 191-201. (19) Yamamoto, K., and Kawanishi, S. (1989) Hydroxyl free radical is not the main active species in site-specific DNA damage induced by copper(II) ion and hydrogen peroxide. J. Biol. Chem. 264, 15435-15440. (20) Kawanishi, S., and Yamamoto, K. (1991) Mechanism of sitespecific DNA damage induced by methylhydrazines in the presence of copper(II) or manganese(III). Biochemistry 30, 3069-3075. (21) Yamamoto, K., and Kawanishi, S. (1991) Site-specific DNA damage induced by hydrazine in the presence of manganese and copper ions; the role of hydroxyl radical and hydrogen atom. J. Biol. Chem. 266, 1509-1515.

Chem. Res. Toxicol., Vol. 13, No. 4, 2000 315 (22) Maxam, A. M., and Gilbert, W. (1980) Sequencing end-labeled DNA with base-specific chemical cleavages. Methods Enzymol. 65, 499-560. (23) Ito, K., Inoue, S., Yamamoto, K., and Kawanishi, S. (1993) 8-Hydroxy-deoxyguanosine formation at the 5′ site of 5′-GG-3′ sequences in double-stranded DNA by UV radiation with riboflavin. J. Biol. Chem. 268, 13221-13227. (24) Kawanishi, S., Inoue, S., and Sano, S. (1986) Mechanism of DNA cleavage induced by sodium chromate(VI) in the presence of hydrogen peroxide. J. Biol. Chem. 261, 5952-5958. (25) Ito, K., Yamamoto, K., and Kawanishi, S. (1992) Manganesemediated oxidative damage of cellular and isolated DNA by isoniazid and related hydrazines: non-fenton-type hydroxyl radical formation. Biochemistry 31, 11606-11613. (26) Pryor, W. A., and Tang, R. H. (1978) Ethylene formation from methional. Biochem. Biophys. Res. Commun. 81, 498-503. (27) Youngman, R. J., and Elstner, E. F. (1981) Oxygen species in paraquat toxicity: the crypto-OH radical. FEBS Lett. 129, 265268. (28) Buettner, G. R. (1987) Spin trapping: ESR parameters of spin adducts. Free Radical Biol. Med. 3, 259-303. (29) March, J. (1992) in Advanced organic chemistry, 4th ed., pp 720733, Wiley-Interscience, New York. (30) Misˇ´ık, V., and Riesz, P. (1996) Peroxyl radical formation in aqueous solutions of N,N-dimethylformamide, N-methylformamide, and dimethylsulfoxide by ultrasound: implications for sonosensitized cell killing. Free Radical Biol. Med. 20, 129-138. (31) Celander, D. W., and Cech, T. R. (1990) Iron(II)-ethylenediaminetetraacetic acid-catalyzed cleavage of RNA and DNA oligonucleotides. Biochemistry 29, 1355-1361. (32) Liang, G., Gannett, P., and Gold, B. (1995) The use of 2-hydroperoxytetrahydrofuran as a reagent to sequence cytosine and to probe non-Watson-Crick DNA structures. Nucleic Acids Res. 23, 713-719. (33) Yamamoto, K., and Kawanishi, S. (1991) Free radical production and site-specific DNA damage induced by hydralazine in the presence of metal ions or peroxidase/hydrogen peroxide. Biochem. Pharmacol. 41, 905-914. (34) Ito, A., Naito, M., Naito, Y., and Watanabe, H. (1982) Induction and characterization of gastro-duodenal lesions in mice given continuous oral administration of hydrogen peroxide. Gann 73, 315-322. (35) Zini, A., and Schlegel, P. N. (1996) Catalase mRNA expression in the male rat reproductive tract. J. Androl. 17, 473-480.

TX990139R