Effects of trans-Resveratrol on Copper-Dependent Hydroxyl-Radical Formation and DNA Damage: Evidence for Hydroxyl-Radical Scavenging and a Novel, Glutathione-Sparing Mechanism of Action

Archives of Biochemistry and Biophysics Vol. 381, No. 2, September 15, pp. 253–263, 2000 doi:10.1006/abbi.2000.1973, available online at http://www.idealibrary.com on

Effects of trans-Resveratrol on Copper-Dependent Hydroxyl-Radical Formation and DNA Damage: Evidence for Hydroxyl-Radical Scavenging and a Novel, Glutathione-Sparing Mechanism of Action Mark J. Burkitt 1 and James Duncan Gray Laboratory Cancer Research Trust, P.O. Box 100, Mount Vernon Hospital, Northwood, Middlesex HA6 2JR, United Kingdom

Received March 22, 2000, and in revised form June 23, 2000

Resveratrol (3,5,4ⴕ-trihydroxy-trans-stilbene) is a natural product occurring in grapes and various other plants with medicinal properties. The phenolic antioxidant has been identified as a potential cancer chemopreventative agent and its presence in red wine has been suggested to be linked to the low incidence of heart disease in some regions of France. Recently, however, resveratrol was reported to promote DNA fragmentation in the presence of copper ions (K. Fukuhara and N. Miyata, 1998, Bioorg. Med. Chem. Lett. 8, 3187–3192), prompting us to investigate this phenomenon in mechanistic detail. By acting as a reducing agent, resveratrol was found to promote hydroxyl-radical ( •OH) formation by DNA-bound Cu(II) ions. However, in the presence of either ascorbic acid or glutathione (i.e., under more physiological conditions), the phenolic lost this property and behaved as an antioxidant. In the ascorbate system, resveratrol had no effect on the rate of •OH formation, but protected DNA from damage by acting as a radicalscavenging antioxidant. In contrast, in the glutathione system, resveratrol inhibited •OH formation via a novel mechanism involving the inhibition of glutathione disulfide formation. We have concluded, therefore, that the DNA-damaging properties of resveratrol, identified recently by Fukuhara and Miyata, will be of no significance under physiological conditions. To the contrary, we have demonstrated that the phenolic behaves as a powerful antioxidant, both via classical, hydroxyl-radical scavenging and via a novel, glutathione-sparing mechanism. © 2000 Academic Press

1

To whom correspondence should be addressed at Gray Laboratory Cancer Research Trust, P.O. Box 100, Mount Vernon Hospital, Northwood, Middlesex HA6 2JR, United Kingdom. Fax: ⫹44 (0)1923-835210. E-mail: [email protected]. 0003-9861/00 $35.00 Copyright © 2000 by Academic Press All rights of reproduction in any form reserved.

Key Words: resveratrol; DNA damage; oxygen radicals; copper; spin trapping; electron paramagnetic (spin) resonance spectroscopy.

There now exists a large body of evidence suggesting that a high dietary intake of antioxidants can confer protection against certain human diseases, including cardiovascular disease, cancer, and degenerative neurological disorders (1– 4). This has provided a considerable impetus to the identification and characterization of novel, plant-derived antioxidants present in the diet and to the investigation of their effects on health. Particular interest has been focused on polyphenolic compounds, such as the flavonoids, catechins, gallic acid derivatives, and trihydroxystilbenes (5–10). In a manner exemplified by ␣-tocopherol (vitamin E), phenolic antioxidants can quench highly reactive free radicals via their (overall) donation of a hydrogen atom (7, 10). The consequence of this “chemical repair” process is the generation of a phenoxyl radical, as shown in Reactions [1] and [2] following the repair of a peroxyl radical by the polyphenol Ph(OH) n OH. Ph共OH兲 n OH º Ph共OH兲 n O ⫺ ⫹ H ⫹

[1]

Ph共OH兲 n O ⫺ ⫹ ROO • ⫹ H ⫹ 3 Ph共OH兲 n O • ⫹ ROOH [2] Due to resonance stabilization, phenoxyl radicals are less reactive than many other radicals, and therefore the interception of free radicals by phenolic compounds generally results in the suppression of potentially 253

254

BURKITT AND DUNCAN

harmful, self-propagating chain reactions, such as lipid peroxidation (7). In a biological membrane, peroxyl radical-scavenging by polyphenolic compounds would spare ␣-tocopherol; and indeed regeneration of ␣-tocopherol through chemical repair of the ␣-tocopheroxyl radical may also enhance protection (11). It should be kept in mind that polyphenolic compounds may also suppress biological free-radical reactions via their chelation of catalytic metal ions, notably iron and copper (12, 13), and can influence biological systems through mechanisms not involving free radicals (14, 15). Although the antioxidant properties of polyphenolic compounds have been reported widely, many studies have also revealed potentially harmful, prooxidant properties (16 –22). Quercetin, for example, is a mutagen (23– 25). The prooxidant properties of polyphenolics probably result from their ease of oxidation, particularly in the presence of metal ions, resulting in the reduction of oxygen to the highly reactive hydroxyl radical, •OH (see Reactions [3] to [6] for catalysis by Cu). Ph共OH兲 n O ⫺ ⫹ Cu 2⫹ 3 Ph共OH兲 n O • ⫹ Cu ⫹ Cu ⫹ ⫹ O2 3 Cu 2⫹ ⫹ O 2•⫺

[3] [4]

2O 2•⫺ ⫹ 2H ⫹ 3 H2O2 ⫹ O2

[5]

Cu ⫹ ⫹ H2O2 3 Cu 2⫹ ⫹ • OH ⫹ OH ⫺

[6]

The catalysis of •OH generation has been suggested to be responsible for the ability of kaempferol to promote DNA fragmentation and lipid peroxidation in isolated rat-liver nuclei (20). Similar reactions are believed to account for the ability of quercetin to induce the fragmentation of isolated DNA and proteins (19). Clearly, the relative importance of the prooxidant and antioxidant properties of polyphenolic compounds, which are widely available as food supplements, must be addressed under biologically relevant conditions before any conclusions can be drawn regarding their possible impact on human health. In the present investigation, we have examined the effects of the polyphenolic compound resveratrol (3,5,4⬘-trihydroxy-trans-stilbene) on copper-dependent DNA damage. Resveratrol is of particular interest because, of hundreds of plant extracts examined, it has been identified as a potential cancer chemopreventive agent: as well as being an antimutagen, it also inhibits events associated with tumor promotion and progression (6). Furthermore, resveratrol has been suggested to be linked to the apparent inverse relationship between red wine consumption and the incidence of heart disease, the so-called French Paradox (26). However, a recent study by Fukuhara and Miyata demonstrated that resveratrol enhances the ability of copper ions to

cleave DNA (22). In view of the suggestion by Jang and colleagues that resveratrol merits further investigation for use as a cancer chemopreventative agent (6), and because the compound is now available as a health food supplement, we considered it timely to examine in greater depth the reported DNA-damaging properties of resveratrol in the presence of copper ions. MATERIALS AND METHODS Reagents. Resveratrol was kindly provided by Pharmascience Inc (Montreal, Quebec). All other reagents, which were of analytical grade, were purchased from Sigma Chemical Co. (Poole, Dorset, UK). A concentrated stock solution containing 0.1 M Mops 2 buffer (pH 7.0) was prepared using Millipore-filtered deionized water and treated with chelating resin (Chelex 100) using the batch method (27). Other stock solutions were also prepared using Millipore-filtered deionized water treated with chelating resin. A concentrated stock solution of DNA (from salmon testes, sodium salt) was prepared using Millipore-filtered deionized water and treated overnight with chelating resin using the batch method. Stock solutions of ascorbic acid and glutathione were prepared on the day of use using the 0.1 M Mops buffer, which was diluted to 50 mM during adjustment to pH 7.0. The spin trap 5,5-dimethyl-1-pyrroline N-oxide (DMPO) was purified by vacuum distillation (Kugelrohr) and stored under nitrogen at ⫺80° prior to use. Determination of oxidative damage to DNA. DNA damage was determined using the ethidium-binding assay, in which DNA damage is indicated by the failure of the nucleic acid to enhance the fluorescence of ethidium bromide. Reactions were performed at room temperature in disposable, open test tubes (2 ml final vol), containing the reagents indicated in the appropriate figure and table legends. The assay was performed as described previously (28, 29) using a Perkin-Elmer LS 50 B fluorescence spectrometer, with the minor modification of terminating reactions using 10 ␮l 0.1 M EDTA plus 5 ␮l 80,000 units ml ⫺1 catalase. EPR spin-trapping studies. Spin-trapping studies were performed using the spin-traps DMPO and N-tert-butyl-␣-phenyl nitrone (PBN). The composition of each reaction mixture is given in the appropriate figure legend. All reactions (0.5 ml final vol) were prepared in microcentrifuge tubes and, unless indicated otherwise, were transferred immediately to a quartz flat cell for analysis. Electron paramagnetic resonance (EPR) spectra were recorded using a Bruker EMX spectrometer (X band), equipped with a 4103 TM cavity (Bruker Spectrospin Ltd, Coventry, UK) and using the following instrument settings: modulation frequency, 100 kHz; sweep width, 60 Gauss; microwave frequency, 9.795 GHz; microwave power, 20 mW; modulation amplitude, 0.5 G; time constant, 10.24 ms; sweep time 42 s; number of scans accumulated, 4; and an appropriate receiver gain. Hyperfine coupling constants were determined from spectral simulations performed using programs that are available through the Internet (http://epr.niehs.nih.gov/) and have been described elsewhere (30). Colorimetric determination of Cu(I). Cu(II) reduction by glutathione and resveratrol was determined colorimetrically using the Cu(I)stabilizing reagent bathocuproine disulfonic acid. DNA (100 ␮g ml ⫺1), 2 mM glutathione, and 50 ␮M CuCl 2 were mixed in the presence of 0.5 mM bathocuproine disulfonic acid and the absorption

2 Abbreviations used: Mops, 4-morpholinepropanesulfonic acid; DMPO, 5,5-dimethyl-1-pyrroline-N-oxide; PBN, N-tert-butyl-␣-phenyl nitrone; DMSO, dimethyl sulfoxide; GSH, glutathione; BC, bathocuproine disulfonic acid; GR, glutathione reductase; GSSH, glutathione disulfide; HRP, horseradish peroxidase.

RESVERATROL AND Cu-DEPENDENT •OH FORMATION

255

were performed involving Cu(II) reduction by glutathione in the presence of glutathione disulfide (see text).

RESULTS

FIG. 1. Effects of resveratrol and hydrogen peroxide on DNA damage induction by Cu(II) ions. DNA (100 ␮g ml ⫺1) was incubated at room temperature with 50 ␮M CuCl 2, in both the presence (open circles) and the absence (solid circles) of 0.1 mM resveratrol, in 25 mM Mops buffer, pH 7.0. Reactions were also performed with the inclusion of 2 mM H 2O 2, in both the presence (open triangles) and the absence (solid triangles) of 0.1 mM resveratrol. All reactions contained 1% (v/v) DMSO and were of final volume 2 ml. Following incubation for the times indicated, reactions were terminated by the addition of 10 ␮l 0.1 M EDTA plus 5 ␮l 80,000 units ml ⫺1 catalase. The integrity of the DNA was then determined by its ability to enhance the fluorescence of ethidium bromide (compared with intact DNA), as described under Materials and Methods. Values are means from three experiments. Error bars are ⫾1 SD. When not shown, error bars are smaller than the symbols.

at 480 nm was read after incubation for 5 min at room temperature. Tubes were read against appropriate blank incubations from which CuCl 2 had been omitted. When indicated, 0.1 mM resveratrol was also present or used in place of glutathione. All incubations contained 1% DMSO (v/v). The concentration of Cu(I) was calculated using the extinction coefficient of its complex with bathocuproine disulfonic acid (⑀ 480 nm ⫽ 13500 M ⫺1 cm ⫺1). Monitoring glutathione disulfide formation. This was achieved enzymatically by coupling glutathione disulfide formation to the oxidation of NADPH using glutathione reductase. NADPH oxidation was monitored against time by recording the decrease in absorption at 340 nm using a Diode Array spectrophotometer (Hewlett Packard, Model 8452 A). Reaction conditions are given in the legend to Fig. 7. EPR spectroscopy of Cu(II). Reaction mixtures (0.5 ml final vol), containing the reagents detailed in the legends to Fig. 6 and 8, were prepared in microcentrifuge tubes and then transferred to 2.5-mminternal-diameter quartz EPR tubes (Wilmad Glass Co., Inc, NJ). Samples were frozen, either immediately (Fig. 6) or after incubation at room temperature (Fig. 8), by placement of the tubes initially for 10 s in a dry-ice/ethanol mixture, followed by placement in liquid nitrogen. EPR spectra were recorded at 77 K using a finger dewar (Wilmad Glass Co., Inc). The instrument settings were modulation frequency, 100 kHz; sweep width, 1200 G; microwave frequency, 9.282 GHz; microwave power, 10 mW; modulation amplitude, 7.81 G; time constant, 20.48 ms; sweep time 84 s; number of scans accumulated, 4; and an appropriate receiver gain. To obtain g values, a small sealed capillary tube containing a dilute solution of diphenylpicrylhydrazyl ( g ⫽ 2.0036) was inserted into the sample solution before freezing. Cu(II) concentrations were estimated by double integration and comparison with a Cu(II)-EDTA standard containing 6 mM ethylenediaminetetraacetic acid (disodium salt), 0.3 mM CuCl 2, and 0.6 mg ml ⫺1 DNA in 25 mM Mops buffer, pH 7.0. Spectra were corrected for cavity contaminant signals by subtraction of the spectrum of a background solution containing 6% (v/v) DMSO and 0.6 mg ml ⫺1 DNA in 25 mM Mops buffer, pH 7.0. Additional experiments

Incubation of DNA with Cu(II) ions and resveratrol resulted in the induction of very modest levels of oxidative damage, as indicated by a decrease in the ability of the nucleic acid to enhance the fluorescence of ethidium bromide. This was not observed when resveratrol was omitted from incubations (Fig. 1). This finding is consistent with the report of Fukuhara and Miyata, who found that resveratrol enhances the ability of Cu(II) ions to cleave plasmid DNA (22). These authors reported a 90% inhibition of DNA fragmentation by catalase, indicating a requirement for hydrogen peroxide. Under these conditions, it is believed that the H 2O 2 necessary for cleavage is generated via the reduction of oxygen by Cu(I) (Reactions [4] and [5]). When H 2O 2 was added to the incubations containing Cu(II) and resveratrol, DNA oxidation was enhanced markedly, but the omission of resveratrol had little effect on the level of damage induction (Fig. 1). The addition of Cu(II) to resveratrol in the presence of the spin-trap DMPO resulted in the detection of an EPR signal from the DMPO hydroxyl radical-adduct, DMPO/ •OH (a N ⫽ 14.96 G, a ␤H ⫽ 14.65 G) (Fig. 2A). The six weaker lines in the spectrum are from a

FIG. 2. EPR spectra of DMPO radical-adducts generated in the presence of Cu(II) ions and resveratrol. (A) The reaction mixture consisted of 50 ␮M CuCl 2, 2 mM resveratrol, 5% (v/v) ethanol, and 100 mM DMPO in 50 mM Mops buffer, pH 7.0. (B) As A, plus 1600 units ml ⫺1 catalase. (C) As A, but the ethanol concentration was increased to 40% (v/v). (D) Computer simulation of the spectrum in C, consisting of the following species: DMPO/ •OH (a N ⫽ 14.71 G, a ␤H ⫽ 13.66 G), DMPO/ •CH(OH)CH 3 (a N ⫽ 15.56 G, a ␤H ⫽ 22.32 G), and DMPO/ •OCH 2CH 3 (a N ⫽ 14.30 G, a ␤H ⫽ 9.32 G, a ␥H ⫽ 1.46 G), with relative concentrations of 0.482, 0.081, and 0.437, respectively.

256

BURKITT AND DUNCAN

trapped carbon-centered radical, probably the ␣-hydroxyethyl radical-adduct of DMPO, DMPO/•CH(OH)CH3, derived from the small amount of ethanol included in the incubation. On the basis of their finding that the DMPO/ •OH signal was insensitive to the presence of either 0.2 M DMSO or 0.5 M mannitol, both of which scavenge hydroxyl radicals, Fukuhara and Miyata suggested that the DMPO/•OH adduct did not result from the reaction of free •OH radicals with DMPO (22). The inclusion of catalase in the incubation system had no effect on the EPR spectrum obtained (Fig. 2B), indicating that H 2O 2 is not required for generation of DMPO/•OH. It is now established that Cu(II) can catalyze the nucleophilic addition of water to DMPO, resulting in formation of the DMPO/ •OH adduct. The observation that catalase had no effect on the intensity of the DMPO/•OH adduct generated in the presence of Cu(II) and resveratrol suggests that it was generated by this route. In order to validate this hypothesis, the reaction was performed in the presence of 40% ethanol. As shown in Fig. 2C, this resulted in the detection of an additional EPR signal, attributed to the DMPO/ •OCH 2CH 3 adduct on the basis of its hyperfine coupling constants (see Fig. 2) (31, 32). Unlike DMPO/ •CH(OH)CH 3, which is generated following the oxidation of ethanol by •OH, the DMPO/ •OCH 2CH 3 adduct is generated by the Cu(II)catalyzed nucleophilic addition of the alcohol (in competition with H 2O) to the spin trap via its lone pair electrons on the oxygen atom (31, 32). Only a very weak contribution from a six-line signal, which may be from DMPO/ •CH(OH)CH 3, was present, remaining unchanged upon increasing the ethanol concentration from 5 to 40%. It is obvious that the radical adducts observed using DMPO as a spin trap in this system are spin-trapping artifacts and therefore provide no insight into the identity of the oxidant(s) responsible for DNA damage. In order to overcome this problem, further experiments were performed using a secondary spin-trapping system, consisting of DMSO and the spin trap PBN. The • OH radical can react directly with PBN, but the adduct formed is short lived. However, in the presence of excess DMSO, •OH radicals are converted into methyl radicals ( •CH 3), which form a relatively stable adduct with PBN, thereby providing a convenient method for the detection of •OH (29, 33). When DMSO and PBN were included in incubations containing DNA, Cu(II), and resveratrol, only very weak background signals, together with a degradation product of the spin trap (di-tert-butyl nitroxide) were observed (data not shown), reflecting the very low levels of DNA damage observed under such conditions (Fig. 1). However, when H 2O 2 was included, a spectrum consisting of signals from two species was observed: a weak signal from the PBN methyl radical-adduct (PBN/ •CH 3) and a stronger signal from the PBN methoxyl radical-adduct

FIG. 3. Effect of resveratrol on hydroxyl-radical formation by DNA-bound Cu(II) ions in the presence of H 2O 2. (A) The reaction mixture, consisting of 2 M DMSO, 100 mM PBN, 100 ␮g ml ⫺1 DNA, 2 mM H 2O 2, and 50 ␮M CuCl 2 in 25 mM Mops buffer, pH 7.0, was incubated at room temperature for 30 min before recording the EPR spectrum. (B) As A, but with the addition of 0.1 mM resveratrol. Hydroxyl-radical formation is indicated by the detection of the sixline signals from the PBN/ •CH 3 (a N ⫽ 16.35 G, a ␤H ⫽ 3.57 G) and PBN/ •OCH 3 (a N ⫽ 15.02 G, a ␤H ⫽ 3.29 G) radical adducts, the latter being the stronger signal in these spectra.

(PBN/ •OCH 3) (Fig. 3). The latter species is generated via the reaction of oxygen with methyl radicals before their reaction with the spin trap (33). This finding is good evidence for the generation •OH. When resveratrol was included, the intensity of both signals increased almost threefold, demonstrating that the phenolic can enhance •OH generation by Cu(II) ions in the presence of DNA (Fig. 3B). We postulated that because the above prooxidant activity of resveratrol results from its ability to reduce Cu(II) ions (see Reactions [3] to [6]), the effects of the compound on •OH formation and DNA damage should also be examined in the presence of typical cellular reducing agents. Indeed, it is already well established that ascorbic acid, glutathione (GSH) and NAD(P)H can promote DNA oxidative damage in the presence of copper ions (28, 34, 35). When DNA was incubated in the presence of Cu(II), H 2O 2, and ascorbic acid, extensive DNA damage occurred (Fig. 4). The inclusion of 0.1 mM resveratrol resulted in a marked decrease in the rate of DNA oxidation: in the absence of resveratrol the DNA totally lost its ability to enhance the fluorescence of ethidium bromide after approximately 30 min, whereas this was delayed to 60 min in the presence of the phenolic (Fig. 4). In parallel spin-trapping experiments, incubation of DNA with Cu(II), hydrogen peroxide and ascorbic acid resulted in the detection of a strong signal from the PBN/ •CH 3 radical adduct (Fig. 5A). However, the EPR spectrum was unaffected by the presence of 0.1 mM resveratrol (Fig. 5A⬘), despite the

257

RESVERATROL AND Cu-DEPENDENT •OH FORMATION TABLE I

Effect of Resveratrol on DNA Damage Induced by Cu(II) Ions in the Presence of Hydrogen Peroxide and Glutathione Fluorescence (%)

FIG. 4. Effect of resveratrol on DNA damage induction by the Cu(II)-H 2O 2-ascorbate system. DNA (25 ␮g ml ⫺1) was incubated at room temperature with 2 mM H 2O 2, 2 mM ascorbic acid, and 10 ␮M CuCl 2, in both the presence (closed circles) and the absence (open circles) of 0.1 mM resveratrol, in 25 mM Mops buffer, pH 7.0. All reactions contained 1% (v/v) DMSO and were of final volume 2 ml. Following incubation for the times indicated, reactions were terminated by the addition of 10 ␮l 0.1 M EDTA plus 5 ␮l 80,000 units ml ⫺1 catalase. The integrity of the DNA was then determined by its ability to enhance the fluorescence of ethidium bromide (compared with intact DNA), as described under Materials and Methods. Values are means from three experiments. Error bars are ⫾1 SD. When not shown, error bars are smaller than the symbol.

protection afforded the nucleic acid (Fig. 4). Only at very high (millimolar) concentrations was resveratrol found to suppress formation of the PBN/ •CH 3 adduct (not shown). When DNA was incubated with Cu(II) and GSH, in either the absence or the presence of resveratrol, no

FIG. 5. Effects of resveratrol on hydroxyl-radical formation by DNA-Cu(II)-H 2O 2 in the presence of either ascorbic acid or glutathione. Reaction mixtures, consisting of 2 M DMSO, 100 mM PBN, 25 ␮g ml ⫺1 DNA, 2 mM H 2O 2, 10 ␮M CuCl 2, and either 2 mM ascorbic acid (A) or 2 mM glutathione (B) in 25 mM Mops buffer, pH 7.0, were incubated at room temperature for 30 min before analysis by EPR spectroscopy. Where indicated, 0.1 mM resveratrol was also present. Hydroxyl-radical formation is indicated by the detection of the sixline EPR signals from the PBN/ •CH 3 (a N ⫽ 16.35 G, a ␤H ⫽ 3.57 G) and PBN/ •OCH 3 (a N ⫽ 15.02 G, a ␤H ⫽ 3.29 G) radical adducts. Spectra B and B⬘ are expanded 16-fold in the y axis relative to A and A⬘.

Incubation time (min)

No resveratrol

Plus resveratrol

30 60

86.68 ⫾ 2.71 79.90 ⫾ 1.88

86.24 ⫾ 2.23 79.37 ⫾ 2.73

Note. DNA (25 ␮g ml ⫺1) was incubated at room temperature with 2 mM glutathione, 2 mM H 2O 2, and 10 ␮M CuCl 2, in both the presence and the absence of 0.1 mM resveratrol, in 25 mM Mops buffer, pH 7.0, for the times indicated. All reactions contained 1% (v/v) DMSO and were of final volume 2 ml. Reactions were terminated by the addition of 10 ␮l 0.1 M EDTA plus 5 ␮l 80,000 units ml ⫺1 catalase. The integrity of the DNA was then determined by its ability to enhance the fluorescence of ethidium bromide (compared with intact DNA), as described under Materials and Methods. Values are means from three experiments ⫾ 1 SD.

DNA damage occurred, as judged by ethidium bromide fluorescence (data not shown). When H 2O 2 was included, very modest levels of DNA damage occurred (Table I). In contrast to the experiments employing ascorbic acid as the reductant, 0.1 mM resveratrol had no effect on the level of DNA damage induction (Table I). However, and contrasting further with the experiments using ascorbic acid, 0.1 mM resveratrol was found to suppress •OH formation (by about 50%) in the corresponding spin-trapping experiment (Figs. 5B and B⬘). Resveratrol had no effect on the intensity of the PBN/ •CH 3 signal when added at the end of the incubation, thereby excluding the possibility that this effect is due its reduction of the radical adduct to a hydroxylamine (not shown). The finding that resveratrol inhibited Cu-dependent DNA damage in the ascorbate system, but had no effect on the level of the PBN/ •CH 3 adduct detected by EPR spectroscopy (unless present at far greater concentration), suggests the operation of competitive radical scavenging. In the incubations assessed for DNA damage, resveratrol is suggested to offer the DNA partial protection by scavenging •OH. In contrast, in the spintrapping incubations, resveratrol must compete with DMSO (and PBN) for reaction with •OH to cause any suppression of the EPR signals. Since DMSO, present at a concentration of 2 M, reacts with •OH with a rate constant of 7 ⫻ 10 9 M ⫺1 s ⫺1 (36), it is certainly not surprising that the inclusion of 0.1 mM resveratrol had no effect on the intensity of the PBN/ •CH 3 signal. Using the same argument, it is reasonable to conclude that the approximate 50% inhibition of radical-adduct formation by 0.1 mM resveratrol in the experiments employing GSH cannot be due to •OH scavenging. Therefore, at this stage, our working hypothesis was that resveratrol must suppress radical-adduct forma-

258

BURKITT AND DUNCAN

FIG. 6. Low-temperature EPR spectra showing the reduction and stabilization of DNA-bound Cu(II) ions by glutathione. (A) Spectrum obtained from a mixture containing 0.6 mg ml ⫺1 DNA, 6% (v/v) DMSO, and 0.3 mM CuCl 2 in 25 mM Mops buffer, pH 7.0. (B) As A, but with the inclusion of 12 mM GSH. Spectra were recorded at 77 K. The vertical bar marks the center of the DPPH spectrum ( g ⫽ 2.0036).

tion in the presence of GSH by directly inhibiting the generation of •OH. This could involve the inhibition of either: (i) the reduction of Cu(II) to Cu(I); or (ii) the reduction of H 2O 2 to the •OH radical by Cu(I). Cu(II) ions added to GSH in the presence of DNA and the Cu(I)-stabilizing agent bathocuproine disulfonic acid (BC) underwent reduction to Cu(I), as indicated by the formation of the Cu(I) complex of BC. The yield of Cu(I) ions was 100% (data not shown). Only about 10% of the Cu(II) underwent reduction in the absence of GSH, but the replacement of GSH with resveratrol also resulted in a 100% yield of Cu(I). The addition of resveratrol to the GSH system had no effect on Cu(II) reduction (not shown). Time-course experiments indicated that Cu(II) reduction was complete within the first minute of incubation, irrespective of the presence of resveratrol (not shown). Since BC stabilizes Cu(I), it can be argued that its presence favors Cu(II) reduction when used to assay Cu(I) (37). Therefore, Cu(II) reduction was also examined directly using EPR spectroscopy. Unlike Cu(I), Cu(II) is paramagnetic and therefore detectable by EPR. When Cu(II) was added to DNA, an EPR signal from Cu(II) was detected at 77 K [ g 储 ⫽ 2.29, a(Cu) 储 ⫽ 157 G], which is assigned to the DNA-bound metal ion (Fig. 6A). When GSH was present, no EPR signal was seen, reflecting the reduction of Cu(II) to Cu(I) and its stabilization as Cu(I) by excess thiol (Fig. 6B). In contrast, when resveratrol was present instead of GSH, the Cu(II) signal remained (data not shown). Rather than indicating that resveratrol cannot reduce Cu(II) (and contradicting the finding from the BC assay), this is believed to reflect the absence of a chelating agent (e.g., GSH or BC) that can stabilize the reduced metal ion: Cu(I) is unstable in

aqueous solution and, therefore, in the absence of a suitable chelating agent, will undergo rapid oxidation or disproportionation (38). When both GSH and resveratrol were present, the Cu(II) signal was not seen, reflecting, again, the stabilization of Cu(I) by the thiol (not shown). These observations suggest that Cu(II) undergoes facile reduction in the presence GSH and resveratrol. When Cu(II) was added to GSH in the presence of DNA, glutathione reductase (GR) and NADPH, the absorption at 340 nm decreased, indicating NADPH oxidation (Fig. 7A). This NADPH oxidation reflects the reduction of glutathione disulfide (GSSG) to GSH, catalyzed by GR. The GSSG is formed via the reduction of Cu(II) to Cu(I) by GSH. Although Cu(II) can oxidize NADPH directly, it has been shown previously that this reaction is of negligible importance in the presence of GSH (39). Thus, GR serves to couple GSH oxidation to the oxidation of NADPH. When resveratrol was included, almost no NADPH oxidation was observed, suggesting its prevention of GSH oxidation by Cu(II) (Fig. 7A). A much higher rate of NADPH oxidation occurred when H 2O 2 was present, but this was inhibited by some 50% in the presence of resveratrol (Fig. 7B). Since the EPR measurements confirmed that

FIG. 7. Effects of resveratrol on glutathione oxidation to its disulfide (GSSG) by Cu(II) ions in the presence and absence of hydrogen peroxide. (A) The reaction system contained 100 ␮g ml ⫺1 DNA, 2 mM GSH, 0.5 mM NADPH, 3.5 units ml ⫺1 glutathione reductase, and 50 ␮M CuCl 2 in 25 mM Mops buffer, pH 7.0. (B) as A, but with the inclusion of 2 mM H 2O 2. Where indicated (upper traces), 0.1 mM resveratrol was present. All incubations contained 1% (v/v) DMSO and were performed at room temperature. GSSG formation is indicated by the decrease in absorption at 340 nm.

RESVERATROL AND Cu-DEPENDENT •OH FORMATION

FIG. 8. Low-temperature EPR spectra showing the effect of resveratrol on formation of the Cu(II)-GSSG complex during the oxidation of glutathione by Cu(II) ions in the presence of hydrogen peroxide and DNA. (A) Incubations contained 0.6 mg ml ⫺1 DNA, 6% (v/v) DMSO, 12 mM H 2O 2, 12 mM GSH, and 0.3 mM CuCl 2 in 25 mM Mops buffer, pH 7.0. (B) As A, but including 0.6 mM resveratrol. Incubations were performed at room temperature. At the times indicated, solutions were frozen and spectra were recorded at 77 K. The Cu(II) concentrations given in parentheses below each spectrum were determined by comparison of the doubly integrated spectra with that of a Cu(II)-EDTA standard. The vertical bar marks the center of the DPPH spectrum ( g ⫽ 2.0036) relative to the spectra in column A.

Cu(II) undergoes reduction in the presence of GSH and resveratrol, the finding that resveratrol prevents GSH oxidation by Cu(II) must mean that the phenolic is acting as the preferred reductant for Cu(II). When Cu(II) was added to DNA in the presence of GSH and H 2O 2, a highly distinctive EPR signal from Cu(II) was seen after freezing to 77 K [ g 储 ⫽ 2.25, a(Cu) 储 ⫽ 175 G]. The intensity of this signal increased upon incubation at room temperature prior to freezing, reaching a maximum after about 2.5 min (Fig. 8A). This signal is characteristic of the spectrum of Cu(II) complexed by GSSG (40, 41). Indeed, an identical spectrum was observed following the addition of Cu(II) to GSSG (data not shown). This Cu(II) signal was also observed when resveratrol was present (Fig. 8B), but its appearance was delayed compared with the system without resveratrol: whereas essentially all of the Cu(II) added appeared as the GSSG complex after 2.5 min in the absence of resveratrol, it took approximately 15 min for all of the added Cu(II) to form this complex in the presence of resveratrol. The finding that resveratrol delayed the appearance of the Cu(II)-GSSG signal further supports the conclusion that the phenolic protects GSH from oxidation. Additional experiments were performed to investigate the ability of GSH to reduce Cu(II) in the presence

259

of increasing proportions of GSSG. As described earlier (Fig. 6), no Cu(II) EPR signal was detected following the addition of 12 mM GSH to 0.3 mM Cu(II) in the presence of DNA (but not H 2O 2), indicating Cu(II) reduction and its stabilization as Cu(I) by GSH. The replacement of GSH with GSH:GSSG mixtures at ratios of 1:1, 1:3, and 1:9 (total concentration, 12 mM) was without effect (data not shown). An extremely weak signal from GSSG-Cu(II) was seen following the addition of 0.6 mM GSH to 0.3 mM Cu(II) in the presence of DNA (not shown), suggesting that a minimum of 2 equiv of thiol are required to generate and stabilize Cu(I). However, a GSSG-Cu(II) signal, corresponding to approximately 10% of the total Cu present, was detected following the addition of 0.6 mM GSH to 0.3 mM Cu(II) in the presence of 11.4 mM GSSG (not shown). These findings indicate that GSH, providing its concentration is more than twice that of Cu(II), can generate and stabilize Cu(I) in the presence of excess GSSG. Glutathione disulfide may be formed from GSH via the combination of two glutathionyl radicals (GS •), generated following the one-electron oxidation of the thiol. Alternatively, in the presence of excess GSH, GSSG can be formed via the addition of a glutathionyl radical to the thiolate anion of glutathione (GS ⫺), forming the highly reducing glutathione disulfide radical anion (GSSG •⫺), which would undergo rapid oxidation to GSSG (42). When Cu(II) was added to DNA, GSH and H 2O 2 in the presence of DMPO, a signal from the DMPO glutathionyl radical adduct (DMPO/ •SG) was detected (a N ⫽ 15.16 G, a ␤H ⫽ 15.74 G), indicating GSH oxidation to GS • (Fig. 9A) (the weaker, six-line signal in this spectrum is from a trapped carbon-centered radical, probably the methyl radical, formed upon DMSO oxidation by •OH). The DMPO/ •SG adduct was not detected when H 2O 2 was omitted, being replaced by a very weak signal from DMPO/ •OH (Fig. 9B), which has been shown to be formed as an artifact (see above and Fig. 2). The DMPO/ •SG adduct was also observed when resveratrol was included in the H 2O 2-containing system, but the signal intensity was reduced by approximately 30% (Fig. 9C); again, generation of DMPO/ • SG was H 2O 2 dependent (Fig. 9D). The finding that DMPO/ •SG was generated only when H 2O 2 was present suggests that GSH is oxidised to GS • by •OH. In addition to its reaction with GSH, • OH could also react with resveratrol, forming the resveratrol phenoxyl radical, Res(OH) 2O •. Some phenoxyl radicals have been shown to oxidize GSH to GS •, which may provide an additional route to GS • (43– 45). In order to explore this possibility, the resveratrol phenoxyl radical was generated directly in high yield using horseradish peroxidase (HRP) and H 2O 2. When HRP and H 2O 2 were added to resveratrol in the presence of GSH and the spin-trap DMPO, a prominent signal

260

BURKITT AND DUNCAN

FIG. 9. Effects of resveratrol on the DMPO radical-adducts detected during the oxidation of glutathione by DNA-bound Cu(II) ions in the presence and absence of hydrogen peroxide. (A) The reaction mixture contained 100 mM DMPO, 25 ␮g ml ⫺1 DNA, 1% (v/v) DMSO, 2 mM H 2O 2, and 10 ␮M CuCl 2 in 25 mM Mops buffer, pH 7.0. (B) As A, but without H 2O 2. (C) As A, but including 0.1 mM resveratrol. (D) As C, but without H 2O 2. The predominant four-line EPR signal (a N ⫽ 15.16 G, a ␤H ⫽ 15.74 G) is assigned to the glutathionyl radical-adduct of DMPO.

from the DMPO/ •SG adduct was observed (a N ⫽ 15.13 G, a ␤H ⫽ 16.18 G) (Fig. 10A). The intensity of this signal was approximately sevenfold less when resveratrol was omitted from the reaction (Fig. 10B). These observations indicate that GSH can be oxidized to GS • by the resveratrol phenoxyl radical (the weaker signal observed in the absence of resveratrol reflects the direct oxidation of the thiol by HRP-H 2O 2).

copper ions to DNA is site specific, it is reasonable to expect DNA damage also to be site specific, which has been demonstrated experimentally (47). When Aruoma and colleagues analyzed the damaged-base products formed in DNA by Cu(II), ascorbic acid, and H 2O 2, they found the pattern of products to be similar to that induced by •OH radicals formed via the radiolysis of water, commenting that no other reactive oxygen species or metal ion-oxygen complex so far studied can generate this range of products from the DNA bases (48). This would not be the case had DNA damage been induced by a less reactive, more selective oxidant. Although we have shown that the DMPO/ •OH adduct is not a reliable indicator of •OH formation from Cu(II)/resveratrol mixtures, our secondary spin-trapping experiments do confirm that resveratrol promotes Cu-dependent •OH formation. The finding that resveratrol increased radical-adduct formation approximately threefold (Fig. 3), but did not cause an increase in DNA damage (Fig. 1), is suggested to reflect radical scavenging by the phenolic. When resveratrol was included in a biomimetic DNA-damaging system consisting of Cu(II), H 2O 2, and ascorbic acid, protection of the nucleic acid was seen (Fig. 4), which is believed to be due to •OH scavenging by resveratrol. This proposal is consistent with the finding that resveratrol did not have any effect on the level of radical-adduct (primarily PBN/ •CH 3) generated when included in the parallel spin-trapping experiment: the •OH radical reacts with DMSO with a rate constant of 7 ⫻ 10 9 M ⫺1 s ⫺1 (36), forming •CH 3, which is detected by spin trapping with PBN; resveratrol, at the concentration used here (0.1 mM), simply cannot compete with 2 M DMSO for reaction with •OH.

DISCUSSION

Consistent with the findings of Fukuhara and Miyata, resveratrol was found to enhance DNA damage when added to a simple solution of the nucleic acid and Cu(II) ions. The identity of the oxidant(s) responsible for the promotion of DNA damage in Cu(I)/H 2O 2generating systems is a matter of some contention in the literature (32). Some authors suggest that hydroxyl radicals ( •OH) are involved (46 –50), whereas others favor the formation of copper-peroxide complexes (22, 35, 51–53) or Cu(III) species (54 –56). The proponents of a major role for oxidants other than •OH often refer to the site specificity of Cu-dependent DNA oxidation. However, because the binding of

FIG. 10. Detection of the glutathionyl radical-adduct of DMPO (a N ⫽ 15.13 G, a ␤H ⫽ 16.18 G) during the oxidation of resveratrol by the horseradish peroxidase-H2O 2 system. (A) The reaction mixture contained 100 mM DMPO, 5 mM glutathione, 1 mM resveratrol, 2% (v/v) DMSO, 1.1 unit ml⫺1 horseradish peroxidase (HRP), and 50 ␮M H 2O 2 in 50 mM Mops buffer, pH 7.0. (B) As A, but without resveratrol.

RESVERATROL AND Cu-DEPENDENT •OH FORMATION

In the DNA oxidation experiment, resveratrol must compete with the nucleic acid for reaction with •OH if it is to provide protection. From a purely kinetic point of view, it is perhaps not unexpected that resveratrol, at a concentration of 0.1 mM, was found to offer the DNA (25 ␮g/ml, corresponding to ⬃60 ␮M nucleotide) some protection from •OH (Fig. 4). In contrast to the Cu(II)-H 2O 2-ascorbate system, • OH generation by the Cu(II)-H 2O 2-GSH system was suppressed markedly by resveratrol. For the kinetic reasons given above, the lower signal intensities of the PBN/ •CH 3 and PBN/ •OCH 3 adducts seen in the presence of resveratrol in the GSH system cannot be due to • OH scavenging. Instead, it has to be concluded that the phenolic suppresses the rate of •OH formation. In principle, the rate of •OH formation could be suppressed by the inhibition of either (i) the initial reduction of Cu(II) to Cu(I) or (ii) the subsequent reaction of Cu(I) with H 2O 2. It is clear from the low-temperature EPR studies (Fig. 6) and the Cu(I) measurements made using BC that Cu(II) undergoes facile reduction in the presence of both GSH and resveratrol, thereby ruling out (i). Indeed, the GSSG measurements revealed that the phenolic is the preferred reductant for Cu(II). The reduction of Cu(II) by GSH is believed to proceed without the formation of thiol radicals (57, 58). Gilbert and colleagues have proposed that one equivalent of the thiol is used to stabilize the Cu(I) formed, even when the metal is bound to DNA (Reaction [7]), and suggested that GSSG, a product of the reaction, will remove Cu(II) from DNA (Reaction [8]) (59, 60). Cu 2⫹ -DNA ⫹ 2 GSH 3 Cu ⫹ -共DNA兲共SG兲 ⫹ 12 GSSG

[7]

Cu 2⫹ -DNA ⫹ GSSG 3 Cu 2⫹ -GSSG ⫹ DNA

[8]

Although reaction of the Cu(I) complex with H 2O 2 might be expected to generate •OH (Reaction [9]), Gilbert et al. have suggested that this oxidation occurs primarily without the generation of •OH (Reaction [10]) (60). Cu ⫹ -共DNA兲共SG兲 ⫹ H2O2 3 Cu 2⫹ -DNA ⫹ • OH ⫹ OH ⫺ ⫹ GSH [9] Cu ⫹ -共DNA兲共SG兲 ⫹ H2O2 3 Cu 2⫹ -DNA ⫹ 12 GSSG ⫹ 2 OH ⫺

[10]

Reaction [10] is stoichiometrically equivalent to Reaction [9] being followed (or accompanied) by reduction of •OH by the bound thiol equivalent, forming OH ⫺ and GS • (21 GSSG), the latter of which we have detected as

261

SCHEME I. Summary of reactions proposed to lead to •OH generation in the DNA-Cu(II)-H 2O 2-glutathione system.

its adduct to DMPO (Fig. 9A). Therefore, it is apparent that any •OH that is generated will be scavenged by the bound thiol. These reactions are summarized by the reaction cycle shown in the top half of Scheme I, in which Cu(I) is oxidized without the generation of •OH [see also Gilbert et al. (59)]. With each turnover of this reaction cycle, GSSG is generated. As the disulfide accumulates, increasing amounts of Cu(II) will be removed from the DNA (Scheme I), as indicated by the gradual appearance of the GSSG-Cu(II) EPR signal (Fig. 8A). The observation that GSH oxidation continued to occur long after all of the Cu present appeared as Cu(II)-GSSG (compare Figs. 7B and 8A) suggests that Cu remains redox active when complexed to the disulfide. This conclusion is also supported by the EPR studies showing that Cu(I) is formed when GSH is added to Cu(II) in the presence of excess GSSG (not shown). The stabilization of Cu(I) by GSH is believed to be responsible for the very low rate at which Cu(I) complexed to the thiol is oxidized by H 2O 2. Gilbert et al. have suggested that a large, negative entropy of activation underlies the slow reaction rate (59). We believe that the eventual chelation of Cu(II) by GSSG is an important factor in •OH generation. Although GSH was found to be capable of forming and stabilizing Cu(I) in the presence of excess GSSG, the Cu(I) was not stabilized completely when only 2 equiv of GSH (relative to Cu) were added in the presence of excess disulfide. Moreover, the detection of the GSSG-Cu(II) EPR signal (Fig. 8A) during the course of GSH oxidation (Fig. 7B) indicates that the metal is not stabilized in the cuprous form when H 2O 2 is also present. Therefore, we propose that the ability of GSH to suppress Cu(I) oxidation is progressively compromised by the GSSG accumulating in the reaction system, allowing •OH generation to occur in reactions predominated by the

262

BURKITT AND DUNCAN

disulfide-bound metal ion (Scheme I, lower cycle). The inhibition of •OH formation by resveratrol may now be explained in terms of its protection of GSH from oxidation, thereby delaying the transfer of Cu from DNA to GSSG. It is also important to consider the involvement of the glutathionyl radical (GS •), which can react with the glutathione anion (GS ⫺) to form the highly reducing glutathione disulfide radical anion, GSSG •⫺, which would be expected to undergo immediate reaction with either oxygen or Cu(II), forming superoxide (42, 61) or Cu(I), respectively. Glutathionyl radicals were indeed detected in the Cu(II)-H 2O 2-GSH system, as their adduct to the spin-trap DMPO, DMPO/ •SG. The fact that GS • was not detected in the absence of H 2O 2 is consistent with the widely held view that GS • radicals are not generated during the reaction of Cu(II) with GSH. Instead, this finding indicates that GS • is generated via the reaction of •OH with GSH (Reaction [11]), the lower yield of DMPO/ •SG in the presence of resveratrol reflecting the effect of the phenolic on •OH formation. GSH ⫹ • OH 3 H2O ⫹ GS •

[11]

We have also shown that the resveratrol phenoxyl radical, generated using the HRP-H 2O 2 system, can oxidize GSH to GS •. Therefore, in addition to the direct oxidation of GSH to GS • by •OH (Reaction [11]), the oxidation of GSH by the phenoxyl radical following the scavenging of •OH by resveratrol must be considered. This latter route to GS • formation could, of course, contribute to the DMPO/ •SG adduct formed under the conditions shown in Fig. 9C. However, the fact that this signal is weaker than that seen in the absence of resveratrol (Fig. 9A) suggests that this indirect route to GS • is not as efficient as the direct oxidation of GSH by • OH. The resveratrol phenoxyl radical is also expected to be generated following the initial reduction of Cu(II) by the phenolic (see Reaction [3], above). However, the failure to detect GS • when resveratrol was added to Cu(II)-DNA in the presence of GSH (but in the absence of H 2O 2) indicates that this is not an important route to GS • formation, probably because the phenoxyl radical is rapidly oxidized by Cu(II), as reported for other phenolics (62, 63). Thus, in sparing GSH, the oxidation of resveratrol by Cu(II) and •OH radicals still results in a net decrease in the rate of GSSG formation. Important conclusions arise from this study: (i) in addition to confirming that resveratrol can promote Cu-dependent oxidative damage to DNA under simple, nonphysiological conditions (i.e., in the absence of biological reducing agents), we have provided EPR spintrapping evidence for •OH generation and demonstrated that the spin-trapping system used by Fukuhara and Miyata to eliminate •OH as the main species

responsible for damage is flawed; (ii) crucially, we have demonstrated that resveratrol loses its ability to promote Cu-dependent •OH formation and DNA damage, switching to antioxidant behavior, in the presence of the reducing agents ascorbic acid and GSH; and (iii) as well as displaying classical, radical-scavenging activity in the ascorbate system, resveratrol can suppresses • OH formation via a novel, GSH-sparing mechanism. Finally, although GSSG appears to play an important role in •OH formation from Cu/GSH/H 2O 2 mixtures, it should be noted that the disulfide has been shown to protect DNA from Cu-dependent damage (even in CuH 2O 2-ascorbate systems) via its removal of the metal ion from the nucleic acid (64). ACKNOWLEDGMENTS The authors thank Pharmascience Inc (Montreal, Quebec) for funding this research and a summer student position for J. Duncan through the Resverin Grant Program for Research on trans-Resveratrol. Additional support was provided by the Cancer Research Campaign.

REFERENCES 1. 2. 3. 4. 5. 6.

7.

8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19.

Gey, K. F. (1993) Brit. Med. Bull. 49, 679 – 699. Halliwell, B. (1996) Free Radical Res. 25, 57–74. Diplock, A. T. (1997) Free Radical Res. 27, 511–532. Floyd, R. A. (1999) Proc. Soc. Exp. Biol. Med. 222, 236 –254. Salah, N., Miller, N. J., Paganga, G., Tijburg, L., Bolwell, G. P., and Rice-Evans, C. (1995) Arch. Biochem. Biophys. 322, 339 –346. Jang, M., Cai, L., Udeani, G. O., Slowing, K. V., Thomas, C. F., Beecher, C. W. W., Fong, H. H. S., Farnsworth, N. R., Kinghorn, A. D., Mehta, R. G., Moon, R. C., and Pezzuto, J. M. (1997) Science 275, 218 –220. Mitchell, J. H., Gardner, P. T., McPhail, D. B., Morrice, P. C., Collins, A. R., and Duthie, G. G. (1998) Arch. Biochem. Biophys. 360, 142–148. Paganga, G., Miller, N., and Rice-Evans, C. A. (1999) Free Radical Res. 30, 153–162. Rehman, A., Bourne, L. C., Haliwell, B., and Rice-Evans, C. A. (1999) Biochem. Biophys. Res. Commun. 262, 828 – 831. Bors, W., and Michel, C. (1999) Free Radical Biol. Med. 27, 1413–1426. Jia, J. S., Zhou, B., Yang, L., Wu, L. M., and Liu, Z. L. (1998) J. Chem. Soc., Perkin Trans. 2, 911–915. Iwahashi, H., Ishii, T., Sugata, R., and Kido, R. (1990) Arch. Biochem. Biophys. 276, 242–247. Brown, J. E., Khodr, H., Hider, R. C., and Rice-Evans, C. A. (1998) Biochem. J. 330, 1173–1178. Miksicek, R. J. (1993) Mol. Pharmacol. 44, 37– 43. Ghazali, R. A., and Waring, R. H. (1999) Life Sci. 65, 1625–1632. Rahman, A., Shahabuddin, Hadi, S. M., Parish, J. H., and Ainley, K. (1989) Carcinogenesis 10, 1833–1839. Laughton, M. J., Halliwell, B., Evans, P. J., and Hoult, J. R. (1989) Biochem. Pharmacol. 38, 2859 –2865. Aruoma, O. I., Murcia, A., Butler, J., and Halliwell, B. (1993) J. Agr. Food Chem. 41, 1880 –1885. Ahmed, M. S., Ainley, K., Parish, J. H., and Hadi, S. M. (1994) Carcinogenesis 15, 1627–1630.

RESVERATROL AND Cu-DEPENDENT •OH FORMATION 20. Sahu, S. C., and Gray, G. C. (1994) Cancer Lett. 85, 159 –164. 21. Yamanaka, N., Oda, O., and Nagao, S. (1997) FEBS Lett. 401, 230 –234. 22. Fukuhara, K., and Miyata, N. (1998) Bioorg. Med. Chem. Lett. 8, 3187–3192. 23. Bjeldanes, L. F., and Chang, G. W. (1977) Science 197, 577–578. 24. Gaspar, J., Laires, A., Monteiro, M., Laureano, O., Ramos, E., and Rueff, J. (1993) Mutagenesis 8, 51–55. 25. Cao, G., Sofic, E., and Pryor, R. L. (1997) Free Radical Biol. Med. 22, 749 –760. 26. Kopp, P. (1998) Eur. J. Endocrinol. 138, 619 – 620. 27. Buettner, G. R. (1988) J. Biochem. Biophys. Methods 16, 27– 40. 28. Milne, L., Nicotera, P., Orrenius, S., and Burkitt, M. J. (1993) Arch. Biochem. Biophys. 304, 102–109. 29. Burkitt, M. J. (1994) Methods Enzymol. 234, 66 –79. 30. Duling, D. (1994) J. Magn. Res. B 104, 105–110. 31. Hanna, P. M., Chamulitrat, W., and Mason, R. P. (1992) Arch. Biochem. Biophys. 296, 640 – 644. 32. Burkitt, M. J., Tsang, S. Y., Tam, S. C., and Bremner, I. (1995) Arch. Biochem. Biophys. 323, 63–70. 33. Burkitt, M. J., and Mason, R. P. (1991) Proc. Natl. Acad. Sci. USA 88, 8440 – 8444. 34. Spear, N., and Aust, S. D. (1995) Arch. Biochem. Biophys. 317, 142–148. 35. Oikawa, S., and Kawanishi, S. (1996) Biochemistry 35, 4584 – 4590. 36. Veltwisch, D., Janata, A., and Asmus, K. D. (1980) J. Chem. Soc., Perkin Trans. 2, 146 –153. 37. Sayre, L. M. (1996) Science 274, 1933–1934. 38. Cotton, F. A., and Wilkinson, G. (1988) Advanced Inorganic Chemistry, 5th ed., pp. 757–758, Wiley, New York. 39. Burkitt, M. J., Bishop, H. S., Milne, L., Tsang, S. Y., Provan, G. J., Nobel, C. S. I., Orrenius, S., and Slater, A. F. G. (1998) Arch. Biochem. Biophys. 353, 73– 84. 40. Kroneck, P. (1975) J. Am. Chem. Soc. 97, 3839 –3841. 41. Singh, R. J., Hogg, N., Goss, S. P., Antholine, W. E., and Kalyanaraman, B. (1999) Arch. Biochem. Biophys. 372, 8 –15. 42. Asmus, K. D. (1990) Methods Enzymol. 186, 168 –180. 43. Sipe, H. J., Jr., Corbett, J. T., and Mason, R. P. (1997) Drug Metab. Dispos. 25, 468 – 480. 44. Sturgeon, B. E., Sipe, H. J., Jr., Barr, D. P., Corbett, J. T., Martinez, J. G., and Mason, R. P. (1998) J. Biol. Chem. 273, 30116 –30121.

263

45. Galati, G., Chan, T., Wu, B., and O’Brien, P. J. (1999) Chem. Res. Toxicol. 12, 521–525. 46. Stoewe, R., and Pru¨tz, W. A. (1987) Free Radical Biol. Med. 3, 97–105. 47. Sagripanti, J. L., and Kraemer, K. H. (1989) J. Biol. Chem. 264, 1729 –1734. 48. Aruoma, O. I., Halliwell, B., Gajewski, E., and Dizdaroglu, M. (1991) Biochem. J. 273, 601– 604. 49. Gunther, M. R., Hanna, P. M., Mason, R. P., and Cohen, M. S. (1995) Arch. Biochem. Biophys. 316, 515–522. 50. Burkitt, M. J., and Milne, L. (1996) FEBS Lett. 379, 51–54. 51. Yamamoto, K., and Kawanishi, S. (1989) J. Biol. Chem. 264, 15435–15440. 52. Yamamoto, K., Inoue, S., and Kawanishi, S. (1993) Carcinogenesis 14, 1397–1401. 53. Hiraku, Y., and Kawanishi, S. (1996) Cancer Res. 56, 1786 – 1793. 54. Johnson, G. R. A., Nazhat, N. B., and Saadalla-Nazhat, R. A. (1985) J. Chem. Soc., Chem. Commun., 407– 408. 55. Johnson, G. R. A., Nazhat, N. B., and Saasalla-Nazhat, R. A. (1988) J. Chem. Soc., Faraday Trans. 84, 501–510. 56. Masarwa, M., Cohen, H., Meyerstein, D., Hickman, D. L., Bakac, A., and Espenson, J. H. (1988) J. Am. Chem. Soc. 110, 4293– 4297. 57. Davis, F. J., Gilbert, B. C., Norman, R. O. C., and Symons, M. C. R. (1983) J. Chem. Soc., Perkin Trans. 2, 1763–1771. 58. Scrivens, G., Gilbert, B. C., and Lee, T. C. P. (1995) J. Chem. Soc., Perkin Trans. 2, 955–963. 59. Gilbert, B. C., Silvester, S., and Walton, P. H. (1999) J. Chem. Soc., Perkin Trans. 2, 1115–1121. 60. Gilbert, B. C., Silvester, S., Walton, P. H., and Whitwood, A. C. (1999) J. Chem. Soc., Perkin Trans. 2, 1891–1895. 61. Surdhar, P. S., and Armstrong, D. A. (1987) J. Phys. Chem. 91, 6532– 6537. 62. Utaka, M., and Takeda, A. (1985) J. Chem. Soc., Chem. Commun., 1824 –1826. 63. Yamashita, N., Murata, M., Inoue, S., Burkitt, M. J., Milne, L., and Kawanishi, S. (1998) Chem. Res. Toxicol. 11, 855– 862. 64. Tsang, S. Y. (1996) Ph.D. thesis, Department of Chemistry, University of Aberdeen, Scotland.