Báo cáo khoa học: Interaction of selenium compounds with zinc finger proteins involved in DNA repair

doi:10.1111/j.1432-1033.2004.04251.x

Eur. J. Biochem. 271, 3190–3199 (2004) (cid:1) FEBS 2004

Interaction of selenium compounds with zinc finger proteins involved in DNA repair

Holger Blessing1, Silke Kraus1, Philipp Heindl1, Wojciech Bal2 and Andrea Hartwig1,3 1Institute of Food Chemistry and Toxicology, University of Karlsruhe, Germany; 2Institute of Biochemistry and Biophysics, Polish Academy of Sciences, Warsaw, Poland; 3Institute of Food Technology and Food Chemistry, Technical University Berlin, Germany

selenocystine, ebselen, and 2-nitrophenylselenocyanate caused a concentration-dependent decrease of Fpg activity, while no inhibition was detected with fully reduced selenomethionine, methylselenocysteine or some sulfur- containing analogs. Furthermore, reducible selenium com- pounds interfered with XPA–DNA binding and released zinc from the zinc finger motif, XPAzf. Zinc release was even evident at high glutathione/oxidised glutathine ratios pre- vailing under cellular conditions. Finally, comparative studies with metallothionein and XPAzf revealed similar or even accelerated zinc release from XPAzf. Altogether, the results indicate that zinc finger motifs are highly reactive towards oxidizing selenium compounds. This could affect gene expression, DNA repair and, thus, genomic stability.

Keywords: DNA repair; glutathione; metallothionein; selenium; zinc finger proteins.

As an essential element, selenium is present in enzymes from several families, including glutathione peroxidases, and is thought to exert anticarcinogenic properties. A remarkable feature of selenium consists of its ability to oxidize thiols under reducing conditions. Thus, one mode of action recently suggested is the oxidation of thiol groups of metal- lothionein, thereby providing zinc for essential reactions. However, tetrahedral zinc ion complexation to four thio- lates, similar to that found in metallothionein, is present in one of the major classes of transcription factors and other so-called zinc finger proteins. Within this study we investi- gated the effect of selenium compounds on the activity of the formamidopyrimidine-DNA glycosylase (Fpg), a zinc finger protein involved in base excision repair, and on the DNA- binding capacity and integrity of xeroderma pigmentosum group A protein (XPA), a zinc finger protein essential for nucleotide excision repair. The reducible selenium compounds phenylseleninic acid, phenylselenyl chloride,

As an essential element, selenium is present in enzymes from several families, including glutathione peroxidases and thioredoxin reductases [1]. Epidemiological evidence, as well as animal studies, point towards an inverse relationship between selenium intake and certain types of cancer [2,3], even though there are still some inconsistencies [4] and the levels required are a matter of debate [5]. Moreover, a multicenter, double-blind, randomized, placebo-controlled cancer prevention trial, originally started up to investigate whether nutritional supplementation with selenium can decrease the risk of skin cancer, revealed significant reductions in lung, colorectal and prostate cancers as secondary end-points [6]. Nevertheless, a follow-up of this

study demonstrated an increased incidence of squamous cell carcinoma and of total nonmelanoma skin cancer [7]. As selenocysteine is an essential constituent of glutathione peroxidases [8], selenium has been proposed to be an antioxidant, but careful consideration of its manifold chemical properties and biological activities reveals far more complex roles with respect to cancer, artherosclerosis and neurodegenerative diseases [9]. From a biochemical point of view, selenium substitutes for sulfur in defined cysteines in selenoproteins. It differs from sulfur by redox potentials and stabilities of oxidation states, leading to multiple catalytic potentials. A remarkable feature of selenium consists of its ability to oxidize thiols under reducing conditions that are present in the cytosol [2,10,11]. One mode of action recently suggested, which may be related to the protective properties of selenium, is its involvement in cellular zinc homeostasis. This assumption is based on the capacity of certain selenium compounds to catalyse thiol/disulfide interchange reactions, which mobil- ize redox-inert zinc from its binding sites, and to reduce protein disulfides, thereby generating potential binding sites for zinc in proteins. In this context, reducible selenium compounds have been shown to release zinc from metal- lothionein (MT) in isolated systems, which may thus be available for essential reactions [12–16]. However, tetrahe- dral zinc ion complexation to four thiolates, similarly to that found in MT, is present in one of the major classes of transcription factors and other so-called zinc finger proteins

Correspondence to A. Hartwig, Technical University Berlin, Institute of Food Technology and Food Chemistry, Sekr. TIB 4/3-1, Gustav-Meyer-Allee 25, D-13355 Berlin, Germany. Fax: +49 30 314 72823, Tel.: +49 30 314 72789, E-mail: Andrea.Hartwig@TU-Berlin.de Abbreviations: Fpg, formamidopyrimidine-DNA glycosylase; GSH, reduced glutathione; GSSG, oxidized glutathione; MT, metallo- thionein; NER, nucleotide excision repair; PAR, 4-(2-pyridylazo)- resorcinol; XPA, xeroderma pigmentosum group A protein; XPAzf, synthetic polypeptide corresponding to the XPA zinc finger domain. (Received 5 February 2004, revised 6 June 2004, accepted 9 June 2004)

Selenium and zinc finger DNA repair proteins (Eur. J. Biochem. 271) 3191

(cid:1) FEBS 2004

(Fig. 1), raising the question of whether reducible selenium compounds are able to release zinc also from this group of proteins. They represent a family of proteins where zinc is complexed through four invariant cysteine and/or histidine residues, forming a zinc finger domain that is mostly involved in DNA binding but also in protein–protein interactions [17]. Originally identified as being present in different transcription factors, it is now known that zinc finger structures are among the most abundant protein motifs in the eukaryotic genome and have diverse functions in many cellular processes [18]: from human genome sequencing it is estimated that (cid:1) 3% of the identified genes encode proteins with zinc finger domains [19]. It is assumed that many of these proteins are regulated by the oxidation of zinc-binding cysteine residues, which leads to a loss of DNA binding, zinc release and/or formation of disulfide bridges [20]. With respect to selenium compounds, it has recently been shown that selenite impairs the DNA-binding activity of TFIIIH, a transcription factor with three zinc finger motifs of the Cys2His2 type [21]. Furthermore, nanomolar concentrations of glutathione peroxidase mimics were demonstrated to facilitate the H2O2-induced oxidation of a Sp1 transcription factor fragment [22]. While interactions of selenium compounds with redox-regulated transcription factors have been repeatedly discussed to be involved in selenium-based chemopreventive effects [23,24], it has to be taken into account that, besides transcription factors, zinc finger structures are also present in other proteins that have essential functions in maintaining genomic stability. These include factors involved in DNA damage signaling and repair, such as poly(ADP-ribose) polymerase-1, formami- dopyrimidine-DNA glycosylase (Fpg), which is involved in the repair of certain types of oxidative DNA base damage [25], and xeroderma pigmentosum group A protein (XPA), which is essential for the assembly of the DNA damage recognition/incision complex during nucleotide excision repair (NER) in mammalian cells [26].

Within the present study we investigated the effects of selenium compounds in different oxidation states, as listed in Table 1, on the activities of the DNA repair proteins Fpg and XPA, as well as on zinc release from a synthetic polypeptide, corresponding to the XPA zinc finger domain (XPAzf). Furthermore, comparative studies with MT and XPAzf were conducted to elucidate the efficiency of zinc release, and, finally, different concentrations of reduced and oxidized glutathione (GSH and GSSG, respectively) were used to investigate the effects under reducing vs. oxidizing condi- tions. We used methylselenocysteine and selenomethionine, as well as selenocystine, as constituents of selenium-enriched yeast, broccoli and onion [27,28]. Our study also included ebselen, phenylseleninic acid, phenylselenyl chloride and 2-nitrophenylselenocyanate – synthetic organic compounds with functional selenium groups exerting cancer-preventive activities and/or which have been shown previously to release zinc from MT [13–16]. We demonstrate that the reducible selenium compounds included in this study (a) inhibit Fpg activity, (b) modulate XPA–DNA interactions, and (c) release zinc from XPAzf with equal or stronger efficiency than MT. The release of zinc also occurred at high ratios of GSH/GSSG, indicating that this reaction mechanism may also be relevant for cellular conditions.

Fig. 1. Structures of metallothionein (MT) and XPAzf (a synthetic polypeptide corresponding to the xeroderma pigmentosum group A protein zinc finger domain). (A) Crystal structure of Cd5Zn2-MT-2 from rat liver, showing the N-terminal b-domain (upper part) and the C-terminal a-domain (lower part) [56]. (B) Solution structure of XPAzf, based on the NMR structure of XPA-MBD [31,58]. Side- chains of the zinc coordinating cysteines are colored in black. Both MT and XPAzf exert tetrahedral metal ion coordination [31,57]. Coordi- nating cysteines are numbered 105, 108, 126 and 129 in XPA and 5, 7, 13, 15, 19, 21, 24, 26, 29, 33, 34, 36, 37, 41, 44, 48, 50, 57, 59 and 60 in MT. These structures are adapted from entries made to the Brook- haven protein databank (accession code 4mt2 for MT and 1xpa for XPA) and modified with CHIME (version 2.6 SP6).

3192 H. Blessing et al. (Eur. J. Biochem. 271)

(cid:1) FEBS 2004

Table 1. Structures and oxidation states of selenium compounds used in the present study.

Materials and methods

Materials

leading to saturation in DNA damage recognition, as determined for each batch of the enzyme. No nonspecific DNA cleavage was observed. Fpg activity was quantified by incision of oxidatively damaged PM2 DNA, essentially as described previously [29] with some modifications. Briefly, PM2 bacteriophage was amplified in Alteromonas espejiana and its circular supercoiled DNA of 10 kb was purified, yielding (cid:1) 90% supercoiled molecules. PM2 DNA was dissolved in enzyme buffer (40 mM sodium phosphate, 100 mM NaCl, pH 7.4) and oxidatively damaged by addition of the photoreactive thiazin dye methylene blue (final concentration 10 lgÆmL)1) and irradiation with visible light (216 JÆm)2). After precipitation (at 4 (cid:2)C for 30 min) with ethanol containing 125 mM sodium acetate, the DNA was centrifuged for 5 min at 7000 g. The supernatant was discarded and the DNA pellet resuspended in enzyme buffer. Damage induction and all subsequent steps were carried out in the dark to prevent additional DNA damage. Oxidatively damaged PM2 DNA (300 ng per sample) and Fpg (1 lgÆmL)1; 30 lL per sample) were incubated for 30 min at 37 (cid:2)C; the reaction was terminated by adding 7 lL of stop solution (0.25% bromophenol blue, 0.25% xylene cyanol, 15% Ficoll 400, w/w/w). When investigating the effects of selenium compounds, Fpg was preincubated with the respective compounds for 30 min at 37 (cid:2)C in enzyme buffer. DNA strand breaks or nicks generated by Fpg convert the supercoiled PM2 molecule into the open circular form; both forms were separated by electrophoresis in a 1% agarose gel in buffer (89 mM Trizma-Base, 89 mM boric acid, 1 mM EDTA, pH 8.2) for 2.5 h at 90 V and stained with ethidium bromide. The bands were quantified by applying a HEROLAB gel detection system (E.A.S.Y. WIN 32). For calculation of break frequencies, a Poisson distribution was assumed and a correction factor of 1.4 was applied to compensate for the relatively lower fluorescence of the supercoiled form [30]:

N ¼ (cid:2)ln½ð1:4 (cid:3) IÞ=ð1:4 (cid:3) I þ IIÞ(cid:4)

where N ¼ the number of strand breaks per molecule of PM2, I ¼ the percentage of supercoiled PM2 DNA, and II ¼ the percentage of open circular PM2 DNA. The overall number of strand breaks per 10 000 bp represents the sum of single-strand breaks and incisions generated by the repair enzyme.

Agarose type II, dimethylsulfoxide, phenylseleninic acid, methylselenocysteine, diamide, Ficoll 400, zinc metallo- thionein II and BSA were from Sigma-Aldrich (Deisenho- fen, Germany). L-cystine, phenylsulfinic acid, L-methionine, methylene blue, xylene cyanol FF and 2-2¢-dithiodipyridine were obtained from Fluka Chemie (Buchs, Switzerland). Bromophenol blue sodium salt, phenylselenyl chloride, acrylamide/bisacrylamide solution (37.5 : 1, w/w; 40%, w/v), ammonium peroxodisulfate, as well as zinc(II) chlor- ide, were products of Merck (Darmstadt, Germany). Anti- digoxigenin Fab fragments and blocking reagent were obtained from Boehringer (Mannheim, Germany), and dithiothreitol, maleic acid, TEMED and Tween-20 were from Serva (Heidelberg, Germany). Enhanced chemilumi- nescence (ECL)TM detection reagents were provided by Amersham (Bucks., UK). 4-(2-Pyridylazo)-resorcinol monosodium salt (PAR) was from Riedel-de Haen (Seelze, Germany). Ebselen, 2-nitrophenylselenocyanate, L-seleno- methionine and L-selenocystine were obtained from Acros (Geel, Belgium).

Fpg activity

DNA-binding activity of XPA Purified recombinant mouse XPA protein was kindly provided by A. Eker (University of Rotterdam, Rotterdam, the Netherlands). The DNA-binding activity was deter- mined by gel mobility shift experiments using a digoxygenin end-labeled synthetic double-strand oligonucleotide (70 bp; MWG Biotech, Ebersberg, Germany) of the following 5¢ fi 3¢ sequence: 5¢-ATATGTGCACATGGCGCACGT ATGTATCTATAGTCTGCCATCACGCCAGTCAAT CGCTGTGGTATATGCA-3¢. XPA (500 ng) was pre- treated with selenium compounds for 30 min at 37 (cid:2)C in the gel shift buffer (final concentration 25 mM Hepes- KOH, 10% glycerol, 30 mM KCl, 4 mM MgCl2, 1 mM EDTA, 45 lgÆmL)1 BSA, 15 lM dithiothreitol, pH 8.3) previously purged with argon. Afterwards, 240 fmol of the irradiated (18 kJÆm)2 UVC; 254 nm germicidal lamp,

Fpg was a kind gift of S. Boiteux (Commissariat a l’Energie Atomique, Fontenay aux Roses, France). The concentration applied in the activity assay (1 lgÆmL)1) was selected based on the results of dose–response experiments

Selenium and zinc finger DNA repair proteins (Eur. J. Biochem. 271) 3193

(cid:1) FEBS 2004

VL-6C; Bioblock Scientific, Illkirch, France) digoxigenin- labeled oligonucleotide were added and incubated for 30 min at room temperature in the dark. The binding mixture was loaded onto a 5% polyacrylamide gel [acrylamide/bisacrylamide (37.5 : 1.0, w/w), 50 mM Tris/ HCl, 50 mM boric acid, 1 mM EDTA, pH 8.0] and electrophoresis was conducted for 75 min. Southern blotting was carried out in a semidry electroblotting apparatus using a positively charged nylon membrane (Hybond-N+ Amersham, Braunschweig, Germany), fol- lowed by fixation at 90 (cid:2)C for 1.5 h. Digoxigenin-labelled oligonucleotide was detected by chemiluminescence with an anti-digoxigenin immunoglobulin conjugated to horseradish peroxidase, using the ECLTM detection system (Amersham, UK).

Zinc release from XPAzf and MT

of chloride, phenylselenyl phenylseleninic

The XPAzf peptide with the sequence: Ac-DYVICEE CGKEFMDSYLMNHFDLPTCDNCRDADDKHK-am (purity > 95%) was custom synthesized by Schafer-N (Copenhagen, Denmark). The identity and purity of the peptide was confirmed by HPLC and ESI-MS, as described previously [31]. Reconstitution of the zinc finger structure was performed by the addition of equimolar amounts of zinc, which resulted in correct folding, as demonstrated by CD spectra and fluorescence spectroscopy [31].

Unbound zinc and salts were removed from MT by ultrafiltration and MT was further characterized by quan- tification of the thiol groups with 2-2¢-dithiodipyridine [32,33], yielding 19.2 thiol groups per molecule. Protein concentrations were determined spectrophotometrically by using the extinction coefficients published previously [34]. Determination of the metal content of MT by complete zinc release after oxidation with 10 mM H2O2 and ICP-MS revealed 6.4 zinc atoms per molecule of MT and negligible amounts of cadmium and copper.

nic/apyrimidinic sites, and converts them into DNA single- strand breaks by its associated DNA endonuclease activity [37, 38]. We applied supercoiled isolated PM2 DNA, which had been oxidatively damaged with methylene blue and visible light, a treatment shown to generate predominantly 8-oxoguanine and small amounts of Fapy-Gua [38, 39]; if Fpg is active, then oxidatively damaged supercoiled PM2 molecules are converted into the open circular form. None of the selenium compounds induced DNA strand breaks by themselves (data not shown); however, Fpg activity was inhibited to differing extents, depending on the selenium compound (Fig. 2). While fully reduced selenomethionine and methylselenocysteine did not affect enzyme activity at concentrations up to 1 mM (data not shown), phenylselenyl chloride, selenocystine, phenylseleninic acid, 2-nitrophenyl- selenocyanate and ebselen reduced the Fpg activity in a dose-dependent manner, yielding complete inhibition. The strongest inhibition was observed when using ebselen; the enzyme was almost completely inactivated at 100 nM selenium compound. To determine whether these inhibi- tions were mediated by selenium, comparative studies were conducted with some sulfur analogues. As shown in Fig. 3, neither cystine nor phenylsulfinic acid affected Fpg activity at strongly inhibitory concentrations of the respective selenium compounds. Furthermore, no inhibition was observed with up to 1 mM methionine.

Zinc release from XPAzf and MT was measured spectro- photometrically by quantifying the formation of complexes between two molecules of the chelating dye PAR and zinc(II) [35,36]. XPAzf (20 lM) or MT (5 lM) were incubated with the respective selenium compounds for 30 min at 37 (cid:2)C in 20 mM Hepes-NaOH buffer (pH 7.4), previously purged with argon. Following the addition of 100 lM PAR, the absorption was measured immediately at 492 nm. As positive controls, XPAzf or MT were exposed to 10 mM H2O2 for 30 min at 37 (cid:2)C to fully oxidize the thiol groups in the respective molecule. Zinc release occurring under these conditions was maximal and considered to be 100%.

Effect of selenium compounds on XPA-DNA binding

acid, Fig. 2. Effects 2-nitrophenylselenocyanate, selenocystine and ebselen on the activity of formamidopyrimidine-DNA glycosylase (Fpg) on methylene blue- damaged PM2 DNA. The protein was incubated with the respective selenium compound for 30 min at 37 (cid:2)C, at the concentrations indi- cated. The mean values of at least four determinations + SD are shown.

Results

Effect of selenium and corresponding sulfur compounds on Fpg activity

Next, the effects of different selenium compounds on the activity of the XPA protein were investigated. XPA binds specifically to damaged DNA, including lesions induced by UVC and benzo[a]pyrene [40–42]; in the present study we analyzed its ability to bind to a UVC-damaged oligonucle- otide by gel mobility shift assay. One representative outcome of the experiments is shown in Fig. 4, derived after a 30 min preincubation of XPA with selenocystine. In the absence of XPA, selenocystine did not affect the migration of free oligonucleotide (data not shown). In the absence of selenocystine, a shift in UVC-irradiated free

First, the effect of selenium compounds on the activity of the isolated Fpg was investigated. Fpg recognizes and removes 7,8-dihydro-8-oxoguanine (8-oxoguanine), the imidazol ring opened purines 2,6-diamino-4-hydroxy-5- formamidopyrimidine (Fapy-Gua), 4,6-diamino-5-form- amidopyrimidine (Fapy-Ade) and, to a smaller extent, 7,8-dihydro-8-oxoadenine (8-oxoadenine), as well as apuri-

3194 H. Blessing et al. (Eur. J. Biochem. 271)

(cid:1) FEBS 2004

metrically by formation of the zinc–PAR complex, as described in the Materials and methods. As a positive control, the Zn(II)-complexed peptide (20 lM) was treated with 10 mM hydrogen peroxide, leading to a saturation of zinc release caused by fully oxidized thiolates. In negative controls (bidistilled water in the absence of selenium) no considerable zinc release was observed (data not shown). Concerning the selenium compounds, (cid:1) 50% zinc release (equivalent to 10 lM Zn2+) was observed with 3 lM 2-nitrophenylselenocyanate, 6 lM selenocystine, 12 lM phe- nylseleninic acid, 16 lM ebselen or 21 lM phenylselenyl chloride. In contrast, methylselenocysteine and seleno- methionine, when used at concentrations up to 1000 lM, were unable to release zinc (Fig. 5).

To compare zinc release by selenium compounds with zinc release by the cellular oxidant GSSG, and to investigate the potential physiological relevance of zinc release by selenium compounds, experiments were per- formed by applying 3 mM GSSG, 1 mM GSH or 10 lM phenylseleninic acid, as well as different GSH/GSSG ratios. As expected, only marginal zinc release was observed in the presence of 1 mM GSH. In contrast, 3 mM GSSG induced (cid:1) 22% zinc release, while 10 lM phenylseleninic acid (in the absence of GSH and GSSG) resulted in (cid:1) 40% zinc release (data not shown), indica- ting that reducible selenium is a stronger oxidant than GSSG. At GSH/GSSG concentration ratios of 1 : 1 to 3 : 1, which is observed in the endoplasmic reticulum [43], zinc release mediated by phenylseleninic acid was accel- erated, from 40% in the absence of GSH/GSSG to 50% or 70%, respectively. At high GSH/GSSG ratios between 40 and 100 prevailing in the overall cell, zinc release was still evident, although to a smaller extent (Fig. 6).

Comparative studies on zinc release from MT and XPAzf

oligonucleotide (lane 1, band 1) mobility was observed, thus demonstrating specific binding of XPA (lane 2, band 2). With increasing concentrations of selenocystine (lanes 3–10), the intensity of band 2 decreased. Nevertheless, instead of free oligonucleotide, which would be expected in the event of diminished XPA–DNA binding, a new band of even slower migration appeared (band 3), indicative of high molecular mass DNA–protein complexes. Similarly, com- plete inhibition of oligonucleotide migration was seen with 25 lM phenylseleninic acid or 75 lM phenylselenyl chloride, as well as with the thiol-oxidizing compound, diamide (data not shown), indicating that thiol oxidation, and probably disulfide formation, may account for this effect. In contrast, oligonucleotide–XPA binding was not affected at up to 1 mM concentrations of fully reduced methylselenocysteine and selenomethionine (data not shown).

Zinc release from XPAzf by selenium compounds

As stated in the Introduction, previous studies by Maret and co-workers have demonstrated zinc release from MT, induced by reducible selenium compounds, as a potential indication for the involvement of selenium in zinc homeo- stasis [12–16]. Thus, in the present study, comparative experiments were conducted to elucidate whether, in the presence of selenium compounds, zinc is released more, equal to or less efficiently from XPAzf than MT. As shown in Fig. 7, all reducible selenium compounds mediated zinc

To investigate the interactions with the zinc finger structure more directly, a peptide consisting of 37 amino acids (XPAzf) was applied as a model, representing the amino acid sequence and structure of the zinc finger domain of human XPA [31]. Zinc release was followed spectrophoto-

Fig. 3. Effects of selenomethionine, selenocystine and phenylseleninic acid and the sulfur analogous methionine, cystine and phenylsulfinic acid on formamidopyrimidine-DNA glycosylase (Fpg) activity. Fpg was incubated with the respective compound for 30 min at 37 (cid:2)C, at the concentrations indicated. The mean values of at least four determi- nations + SD are shown.

Fig. 4. Effect of selenocystine on the DNA- binding activity of xeroderma pigmentosum group A protein (XPA) to a UVC-irradiated oligonucleotide. XPA was incubated with selenocystine for 30 min at 37 (cid:2)C and its binding activity was analysed by gel mobility shift assay, as described in the Materials and methods. One representative experiment is shown.

Selenium and zinc finger DNA repair proteins (Eur. J. Biochem. 271) 3195

(cid:1) FEBS 2004

Fig. 7. Release of zinc from metallothionein (MT) by phenylselenyl chloride, phenylseleninic acid, 2-nitrophenylselenocyanate, selenocystine, ebselen, selenomethionine and methylselenocysteine. MT was incubated with the respective selenium compound for 30 min at 37 (cid:2)C. Zinc release was measured spectrophotometrically by the addition of 100 lM 4-(2-pyridylazo)-resorcinol (PAR). The mean values of at least six determinations + SD are shown. Fig. 5. Release of zinc from XPAzf by phenylselenyl chloride, phenyl- seleninic acid, 2-nitrophenylselenocyanate, selenocystine, ebselen, sele- nomethionine and methylselenocysteine. The XPAzf peptide was incubated with the respective selenium compound for 30 min at 37 (cid:2)C. Zinc release was measured spectrophotometrically after the addition of 100 lM 4-(2-pyridylazo)-resorcinol (PAR). The mean values of at least six determinations + SD are shown.

chloride, zinc release occurred at similar concentrations from both molecules (Fig. 9).

Fig. 8. Release of zinc from metallothionein (MT) or XPAzf by 2-nitrophenylselenocyanate and selenocystine. The data are derived from the experiments shown in Figs 5 and 7, but plotted against the number of cysteines present in XPAzf and MT. The mean values of at least six determinations + SD are shown.

Discussion

release also from MT, while fully reduced selenomethionine and methylselenocysteine were inactive, as seen previously with XPAzf. As MT contains 20 zinc-binding cysteines as compared to four cysteines in XPAzf, for better compar- ison, zinc release was plotted against the ratio of selenium compound/cysteine. In the case of selenocystine and 2-nitrophenylselenocyanate, zinc was released with an even higher efficiency from XPAzf than from MT (Fig. 8); with respect to ebselen, phenylseleninic acid and phenylselenyl

The cancer-preventive activities of selenium compounds have long been discussed, such as the effects of different dietary levels as a result of geographical variation, the potential benefits of dietary supplements, and the clinical applications as chemopreventive agents [2–4,6,9]. Further- more, selenium deficiency has been linked to increased viral pathogenicity in humans (Keshan disease) and in animal experiments [44]. One obvious function of selenium relates to its role as an antioxidant as a constituent part of glutathione peroxidases in the detoxification of peroxides, which leads to a reduction in the level of reactive oxygen species in cells and tissues. In addition, in recent years,

Fig. 6. Release of zinc from XPAzf by various ratios of reduced gluta- thione (GSH)/oxidized glutathione (GSSG), in the presence (s) or absence (h) of 10 lM phenylseleninic acid. GSSG was applied from 10 lM to 3 mM, while GSH was maintained at 1 mM. The peptide was incubated with the respective compounds for 30 min at 37 (cid:2)C. Zinc release was measured spectrophotometrically by the addition of 100 lM 4-(2-pyridylazo)-resorcinol (PAR). About 40% zinc release was observed with 10 lM phenylseleninic acid in the absence of GSH and GSSG (data not shown). The mean values of at least nine deter- minations + SD are shown.

3196 H. Blessing et al. (Eur. J. Biochem. 271)

(cid:1) FEBS 2004

complexing thiol groups in the enzyme and a simultaneous reduction of the selenium compound. This interpretation is in agreement with the mutational analysis of Fpg, which revealed that substitution of any of the cysteines in the zinc finger destroys DNA-binding capacity and enzyme func- tion, while substitution of the other two cysteines outside the zinc finger has little effect [25]. The participation of selenium is further demonstrated by the lack of inhibition by the sulfur-containing analogues cystine and phenylsulfinic acid. Similarly, selenocystamine reacted with the zinc/sulfur clusters of MT at much lower concentrations than cysta- mine [12,47]. Although sulfur shares similar chemical properties with selenium [48], one difference is the redox chemistry, which allows selenium to act at much lower physiological concentrations than sulfur.

the complexes,

greater emphasis has been given to specific cellular reac- tions, based on selenium catalysis. This concerns reversible cysteine/disulfide transformations in redox-regulated pro- teins, such as transcription factors [24], and MT, as a zinc storage protein [11,14]. Nevertheless, there is increasing evidence that zinc finger structures are among the most common protein motifs, present not just in transcription factors, but in basically all families of proteins involved in maintaining genomic stability, including DNA repair pro- teins and cell cycle control proteins [18]. This prompted us to investigate the effect of selenium compounds in different oxidation states on the integrity and function of two zinc finger proteins (Fpg and XPA) with tetrahedral zinc ion complexation to four cysteines involved in DNA repair, as compared to zinc release from MT. Our experiments demonstrate, for the first time, that reducible selenium compounds (i.e. with an oxidation state of –I or higher) including selenocystine, phenylselenyl chloride, ebselen, 2-nitrophenylselenocyanate and phenylseleninic acid, are able to react with the thiolates of these zinc finger DNA-repair proteins, resulting in enzyme inactivation and release of zinc.

The effects of different selenium compounds on the mammalian XPA protein were investigated in the second model. In vivo, XPA is absolutely required for incision during NER by co-ordinating the binding of ERCC1– XPF and presumably activating XPG [49]. In vitro, it binds specifically to bulky DNA lesions, such as (6–4)- photoproducts induced by UVC [40]. Even though its zinc finger motif is not directly involved in DNA binding, it is required for the correct folding of the minimal DNA- binding domain [50]; substitution of any of its zinc- complexing cysteines (Cys105, Cys108, Cys126 and Cys129) leads to diminished DNA binding and a severe reduction in NER activity [26]. In the present study, all reducible selenium compounds investigated in this set of experiments (selenocystine, phenylselenyl chloride, phenyl- seleninic acid) led to profound changes in the gel mobility shift experiments performed to investigate XPA–DNA interactions. If DNA–protein interactions would have been disturbed, one would expect the appearance of free oligonucleotide with increasing concentrations of redu- cible selenium compounds. Surprisingly, this was not the case; however, a third band appeared with almost no migration in the gel, indicative of high molecular mass DNA–protein complexes. As selenium compounds exert higher reactivity towards free cysteines as compared to zinc-bound thiolates [16], intermolecular disulfide bridges may be formed between two or more XPA molecules which still retain their DNA-binding activity. It cannot be excluded that selenium compounds also react with the zinc finger thiol groups, but, owing to the completely retarded migration of this cannot be discriminated by gel mobility shift analysis.

To elucidate interactions with the zinc finger domain of XPA more directly, the effects of selenium compounds on a synthetic 37 amino acid peptide, representing the zinc finger domain of the human XPA protein (XPAzf), were inves- tigated by determining zinc release. Similarly to the results obtained with Fpg, all reducible selenium compounds caused zinc release in a concentration-dependent manner, starting at the low micromolar range, while fully reduced selenomethionine and methylselenocysteine were inactive. For comparison, a 10 mM H2O2 solution was required to mediate complete zinc release.

The experiments demonstrate that

The bacterial Fpg was used as the first model. Fpg is a glycosylase that initiates base excision repair in Escheri- chia coli, which recognizes and removes several oxidative DNA base modifications. Fpg has the highest affinity for 8-oxoguanine, which, owing to its mutagenic potential, is believed to be the biologically most relevant substrate [37,38]. DNA binding of Fpg is mediated by a single zinc finger domain in the C-terminal region of the enzyme, where zinc is complexed tetrahedrally by four cysteines (Cys244, Cys247, Cys264 and Cys267) [45,46]. Within the present study we demonstrated that all reducible selenium com- pounds inhibited Fpg activity completely, albeit at different concentrations. As fully reduced methylselenocysteine and selenomethionine were not inhibitory, the observed enzyme inactivation is probably caused by the oxidation of zinc-

in principle any reducible selenium compound should be able to release zinc from zinc finger structures. Nevertheless, there were clear differences with respect to the reactivities of different

Fig. 9. Release of zinc from metallothionein (MT) or XPAzf (a synthetic polypeptide corresponding to the xeroderma pigmentosum group A protein zinc finger domain) by various concentrations of phenylselenyl chloride, phenylseleninic acid and ebselen. The data are derived from the experiments shown in Figs 5 and 7, but plotted against the number of cysteines present in XPAzf and MT. The mean values of at least six determinations + SD are shown.

Selenium and zinc finger DNA repair proteins (Eur. J. Biochem. 271) 3197

(cid:1) FEBS 2004

consensus recognition sequence when cells were treated with either 1,4-phenylenebis(methylene)selenocyanate (p-XSC) or selenite, even though the authors did not investigate whether the zinc finger was affected directly [23].

selenium compounds. In the experiments with Fpg, inhi- bitory concentrations were not related to the respective oxidation state. Thus, for selenium compounds of the oxidation state 0, the strongest inhibition was observed with ebselen, where full inhibition was evident at 100 nM, while 2-nitrophenylselenocyanate and phenylselenyl chloride exer- ted similar inhibition at 2.5 and 10 lM, respectively. Full inhibition was obtained with 100 lM selenocystine (oxida- tion state –I) or 100 lM phenylseleninic acid (oxidation state +II). In experiments with XPAzf, the strongest effect was observed with 2-nitrophenylselenocyanate, while compar- able reactivities were found for the other selenium com- pounds in different oxidation states. As the results were somehow more uniform in the isolated zinc finger peptide, XPAzf, one reason for the differences observed with Fpg may be the higher order protein structures and differences in accessibility in the intact protein, one major determinant for the sensitivity of zinc finger thiolates towards oxidation [20]. Nevertheless, zinc release is not merely a stoichiometric process, but also exerts a catalytic component owing to the oxidation of selenolates, for example by trace amounts of oxygen. This catalytic component has been shown to occur in 2-nitrophenylselenocyanate with respect to zinc release from MT [14] and may differ depending on the actual selenium compound.

As stated in the Introduction, one mechanism suggested to contribute to the cancer-preventive properties of selenium compounds is the involvement in cellular zinc homeostasis by mediating zinc release from MT, thus providing it for essential reactions. However, significant protection would require preferential reaction with MT, compared with zinc finger proteins, to minimize toxic reactions. Our experiments confirm zinc release by reducible selenium compounds described previously by Maret and co-workers [12–15], but revealed additionally that the zinc finger domain of XPA is at least as susceptible towards thiol oxidation by reducible selenium compounds as MT. During incubation with selenocystine and 2-nitrophenylselenocyanate, zinc was released at even lower concentrations from XPAzf compared with MT. The reason for this is unclear at the moment; potential contributing factors are better assessibility of the peptide vs. MT combined with more pronounced catalytic reaction components, as discussed above. Whether or not zinc finger DNA repair proteins are affected under cellular conditions has yet to be elucidated. In the cytosol of cells, MT is present in excess over DNA repair proteins; however, owing to the coupling of zinc-binding structures (either in MT or in zinc finger proteins) by selenium compounds with the cellular redox system GSH/GSSG, reducible selenium may be regenerated within catalytic cycles and thus is not necessarily inactivated by an excess of MT.

Taken together, the results presented in this study demonstrate that low concentrations of selenium com- pounds in reducible oxidation states may inactivate DNA repair processes by the oxidation of zinc finger structures in DNA repair proteins. As these observations are derived from subcellular systems, experiments on interactions with DNA repair systems in intact cells are urgently needed. The hypothesis on DNA repair inhibition seems to contradict observations published recently, where selenomethionine even stimulated DNA repair after UV irradiation of human fibroblasts [53]. Nevertheless, selenomethionine is fully reduced (oxidation state –II) and was found not to interfere with zinc finger proteins in the present study. In principle, our results are in agreement with the growing experimental evidence that the role of selenium compounds in biological systems is not merely antioxidative (by detoxification of reactive oxygen species as part of glutathione peroxidases), but rather complex, owing to multiple interactions with cellular thiol groups. Importantly, it was shown that not only free SH groups are targets of reducible selenium compounds, but also zinc-complexed thiol groups, which are commonly thought to be more resistant towards redox reactions [54, 55]. While interactions with isolated DNA repair proteins are demonstrated for the first time, recent studies by other groups also showed inhibition of transcrip- tion factor–DNA binding by reducible selenium com- pounds [21, 23]. In fact, these interactions with either transcription factors or MT have repeatedly been discussed to contribute to the protective properties of selenium compounds with respect to tumor prevention or tumor therapy. However, as shown in the present study, these reactions may also have detrimental consequences: because

In cells there is an excess of free SH groups, predomin- antly provided by GSH, raising the question of whether under conditions of high concentrations of GSH and/or GSSG, the cysteines in zinc finger structures remain as targets for oxidation by selenium compounds. As shown in the present study, zinc release is even accelerated in the presence of GSSG and at GSH/GSSG ratios between 0.3 and 3, and is only partly prevented even at a 100-fold excess of GSH over both phenylseleninic acid and GSSG. This indicates that reducible selenium compounds are able to attack zinc-sulfur bonds also under cellular conditions, where GSH/GSSG ratios between 1 and 3 for the endoplasmatic reticulum and between 40 and 100 for the overall cell have been reported [43]. Similarly, Maret and co- workers observed zinc release from zinc-sulfur clusters in MT, catalyzed by reducible selenium compounds, at a high excess of GSH over MT [13]. One possible explanation relates to the formation of mixed selenodisulfides between selenium compounds and GSH, which are still sufficiently reactive towards zinc finger thiols. Furthermore, in the presence of GSSG, redox cycling could take place by oxidation of reduced selenium species, which would explain the enhanced zinc release in the presence of GSSG observed in our experiments. In this context, reducible selenium coumpounds have been shown to efficiently couple the GSH/GSSG redox pair with the MT/thionein system [15] and a similar mechanism would be plausible in the case of zinc finger proteins. Under cellular conditions, redox changes of selenium are part of the essential reactions, for example within the catalytic cycle of glutathione peroxidases [51,52]. In an isolated test system, recent investigations demonstrated that glutathione peroxidase mimics in the presence of hydrogen peroxide are able to oxidize the zinc finger peptide fragment of transcription factor, SP1 [22]. With respect to oxidized selenium species, Youn et al. reported a concentration-dependent decreased binding of zinc finger transcription factors Sp1 and Sp3 to their

3198 H. Blessing et al. (Eur. J. Biochem. 271)

(cid:1) FEBS 2004

11. Jacob, C., Giles, G.I., Giles, N.M. & Sies, H. (2003) Sulfur and selenium: the role of oxidation state in protein structure and function. Angew Chem. Int. Ed. Engl. 42, 4742–4758.

12. Jacob, C., Maret, W. & Vallee, B.L. (1998) Control of zinc transfer between thionein, metallothionein, and zinc proteins. Proc. Natl Acad. Sci. USA 95, 3489–3494.

13. Jacob, C., Maret, W. & Vallee, B.L. (1999) Selenium redox bio- chemistry of zinc-sulfur coordination sites in proteins and enzymes. Proc. Natl Acad. Sci. USA 96, 1910–1914.

14. Chen, Y. & Maret, W. (2001) Catalytic oxidation of zinc/sulfur coordination sites in proteins by selenium compounds. Antioxid. Redox Signal. 3, 651–656.

zinc finger proteins are involved in basically all cellular reactions required to maintain genomic stability, their inactivation may lead to increased genetic instability. Thus, functioning DNA repair processes are urgently required to protect the genome not only from DNA damage induced by environmental agents such as UV radiation, food mutagens and polycyclic aromatic hydrocarbons, but also from endogenous DNA damage generated continuously, for example by reactive oxygen species arising in the course of oxygen consumption. As redox reactions are important for the regulation of zinc finger proteins and thus the cellular pathways that are dependent on these proteins, an imbal- ance in selenium compounds as powerful mediators of cellular redox reactions, provoked by either selenium deficiency or oversupply, may considerably decrease genomic stability.

15. Chen, Y. & Maret, W. (2001) Catalytic selenols couple the redox cycles of metallothionein and glutathione. Eur. J. Biochem. 268, 3346–3353.

16. Jacob, C., Maret, W. & Vallee, B.L. (1998) Ebselen, a selenium- containing redox drug, releases zinc from metallothionein. Bio- chem. Biophys. Res. Commun. 248, 569–573. 17. Mackay, J.P. & Crossley, M. (1998) Zinc fingers are sticking

Acknowledgements

together. Trends Biochem. Sci. 23, 1–4.

18. Laity, J.H., Lee, B.M. & Wright, P.E. (2001) Zinc finger proteins: new insights into structural and functional diversity. Curr. Opin. Struct. Biol. 11, 39–46. 19. Maret, W. (2003) Cellular zinc and redox states converge in the metallothionein/thionein pair. J. Nutr. 133, 1460S–1462S.

The Fpg protein was a kind gift of Dr Serge Boiteux, Commisariat Energie Atomique, Fontanay aux Roses, France and XPA was kindly provided by Dr Andre´ Eker, Erasmus University of Rotterdam, the Netherlands. This work was supported by grants Ha 2372/1-3 and Ha 2372/3-2 from the Deutsche Forschungsgemeinschaft, and by a grant from the Alexander von Humboldt Foundation (a maintenance grant to W.B.). 20. Wilcox, D.E., Schenk, A.D., Feldman, B.M. & Xu, Y. (2001) Oxidation of zinc-binding cysteine residues in transcription factor proteins. Antioxid. Redox Signal. 3, 549–564.

References

21. Larabee, J.L., Hocker, J.R., Hanas, R.J., Kahn, F.M. & Hanas, J.S. (2002) Inhibition of zinc finger protein–DNA interactions by sodium selenite. Biochem. Pharmacol. 64, 1757–1765.

1. Ko¨ hrle, J., Brigelius-Flohe´ , R., Bock, A., Gartner, R., Meyer, O. & Flohe´ , L. (2000) Selenium in biology: facts and medical per- spectives. Biol. Chem. 381, 849–864. 22. Giles, N.M., Gutowski, N.J., Giles, G.I. & Jacob, C. (2003) Redox catalysts as sensitisers towards oxidative stress. FEBS Lett. 535, 179–182.

2. Ganther, H.E. (1999) Selenium metabolism, selenoproteins and mechanisms of cancer prevention: complexities with thioredoxin reductase. Carcinogenesis 20, 1657–1666. 23. Youn, B.W., Fiala, E.S. & Sohn, O.S. (2001) Mechanisms of organoselenium compounds in chemoprevention: effects on tran- scription factor-DNA binding. Nutr. Cancer 40, 28–33. 3. El-Bayoumy, K. (2001) The protective role of selenium on genetic 24. Kim, Y.S. & Milner, J. (2001) Molecular targets for selenium in damage and on cancer. Mutat. Res. 475, 123–139. cancer prevention. Nutr. Cancer 40, 50–54.

25. O’Connor, T.R., Graves, R.J., de Murcia, G., Castaing, B. & Laval, J. (1993) Fpg protein of Escherichia coli is a zinc finger protein whose cysteine residues have a structural and/or functional role. J. Biol. Chem. 268, 9063–9070. 4. Vinceti, M., Rovesti, S., Bergomi, M. & Vivoli, G. (2000) The epidemiology of selenium and human cancer. Tumori 86, 105–118. 5. Thomson, C.D. (2004) Assessment of requirements for selenium and adequacy of selenium status: a review. Eur. J. Clin. Nutr. 58, 391–402.

26. Miyamoto, I., Miura, N., Niwa, H., Miyazaki, J. & Tanaka, K. (1992) Mutational analysis of the structure and function of the xeroderma pigmentosum group A complementing protein. Iden- tification of essential domains for nuclear localization and DNA excision repair. J. Biol. Chem. 267, 12182–12187.

6. Clark, L.C., Combs, G.F. Jr, Turnbull, B.W., Slate, E.H., Chal- ker, D.K., Chow, J., Davis, L.S., Glover, R.A., Graham, G.F., Gross, E.G., Krongrad, A., Lesher, J.L. Jr, Park, H.K., Sanders, B.B. Jr, Smith, C.L. & Taylor, J.R. (1996) Effects of selenium supplementation for cancer prevention in patients with carcinoma of the skin. A randomized controlled trial. Nutritional Prevention of Cancer Study Group. JAMA 276, 1957–1963.

27. Cai, X.-J., Block, E., Uden, P.C., Zhang, X., Quimby, B.D. & Sullivan, J.J. (1995) Allium chemistry: identification of seleno- amino acids in ordinary and selenium-enriched garlic, onion, and broccoli using gas chromatography with atomic emission detec- tion. J. Agric. Food Chem. 43, 1754–1757.

7. Duffield-Lillico, A.J., Slate, E.H., Reid, M.E., Turnbull, B.W., Wilkins, P.A., Combs, G.F. Jr, Park, H.K., Gross, E.G., Graham, G.F., Stratton, M.S., Marshall, J.R. & Clark, L.C. (2003) Sele- nium supplementation and secondary prevention of non- melanoma skin cancer in a randomized trial. J. Natl Cancer Inst. 95, 1477–1481. 8. Flohe´ , L., Gunzler, W.A. & Schock, H.H. (1973) Glutathione 28. Bird, S., Uden, P.C., Tyson, J.F., Block, E. & Denoyer, E. (1997) Speciation of selenoamino acids and organoselenium compounds in selenium-enriched yeast using high-performance liquid chro- matography-inductively coupled plasma mass spectrometry. J. Anal. Atomic Spectrom. 12, 785–788. peroxidase: a selenoenzyme. FEBS Lett. 32, 132–134.

29. Asmuss, M., Mullenders, L.H., Eker, A. & Hartwig, A. (2000) Differential effects of toxic metal compounds on the activities of Fpg and XPA, two zinc finger proteins involved in DNA repair. Carcinogenesis 21, 2097–2104. 9. Brigelius-Flohe´ , R., Maiorino, M., Ursini, F. & Flohe, L. (2001) Selenium: An antioxidant? In Handbook of Antioxidants, 2nd edn. (Cadenas, E. & Packer, L., eds), pp. 633–664. Marcel Dekker, Inc., New York, Basel.

30. Mu¨ ller, E., Boiteux, S., Cunningham, R.P. & Epe, B. (1990) Enzymatic recognition of DNA modifications induced by singlet oxygen and photosensitizers. Nucleic Acids Res. 18, 5969–5973. 10. Turan, B., Fliss, H. & Desilets, M. (1997) Oxidants increase the intracellular free Zn2+ concentration in rabbit ventricular myo- cytes. Am. J. Physiol. 272, 2095–2106.

Selenium and zinc finger DNA repair proteins (Eur. J. Biochem. 271) 3199

(cid:1) FEBS 2004

(II)-associated Escherichia coli formamidopyrimidine-DNA gly- cosylase: extended X-ray absorption fine structure evidence for a metal-binding domain. Biochemistry 39, 12441–12449. 31. Bal, W., Schwerdtle, T. & Hartwig, A. (2003) Mechanism of nickel assault on the zinc finger of DNA repair protein XPA. Chem. Res. Toxicol. 16, 242–248.

46. Zharkov, D.O., Shoham, G. & Grollman, A.P. (2003) Structural characterization of the Fpg family of DNA glycosylases. DNA Repair (Amst). 2, 839–862. 32. Grassetti, D.R. & Murray, J.F. Jr (1967) Determination of sulf- hydryl groups with 2,2¢- or 4,4¢-dithiodipyridine. Arch. Biochem. Biophys. 119, 41–49. 33. Vasak, M. (1991) Criteria of purity for metallothioneins. Methods Enzymol. 205, 44–47. 47. Maret, W. (1995) Metallothionein/disulfide interactions, oxidative stress, and the mobilization of cellular zinc. Neurochem. Int. 27, 111–117. 48. Barceloux, D.G. (1999) Selenium. J. Toxicol. Clin. Toxicol. 37, 145–172. 34. Schaeffer, A. (1991) Absorption, circular dichroism, and magnetic circular dichroism spectroscopy of metallothionein. Methods Enzymol. 205, 529–540.

35. Hunt, J.B., Neece, S.H., Schachman, H.K. & Ginsburg, A. (1984) Mercurial-promoted Zn2+ release from Escherichia coli aspartate transcarbamoylase. J. Biol. Chem. 259, 14793–14803.

49. Volker, M., Mone, M.J., Karmakar, P., van Hoffen, A., Schul, W., Vermeulen, W., Hoeijmakers, J.H., van Driel, R., van Zeeland, A.A. & Mullenders, L.H. (2001) Sequential assembly of the nucleotide excision repair factors in vivo. Mol. Cell. 8, 213– 224. 36. McCall, K.A. & Fierke, C.A. (2000) Colorimetric and fluorimetric assays to quantitate micromolar concentrations of transition metals. Anal. Biochem. 284, 307–315.

50. Buchko, G.W., Ni, S., Thrall, B.D. & Kennedy, M.A. (1998) Structural features of the minimal DNA binding domain (M98– F219) of human nucleotide excision repair protein XPA. Nucleic Acids Res. 26, 2779–2788. 37. Tchou, J., Kasai, H., Shibutani, S., Chung, M.H., Laval, J., Grollman, A.P. & Nishimura, S. (1991) 8-oxoguanine (8-hydro- xyguanine) DNA glycosylase and its substrate specificity. Proc. Natl Acad. Sci. USA 88, 4690–4694.

51. Wendel, A., Pilz, W., Ladenstein, R., Sawatzki, G. & Weser, U. (1975) Substrate-induced redox change of selenium in glutathione peroxidase studied by x-ray photoelectron spectroscopy. Biochim. Biophys. Acta 377, 211–215.

38. Boiteux, S., Gajewski, E., Laval, J. & Dizdaroglu, M. (1992) Substrate specificity of the Escherichia coli Fpg protein (for- mamidopyrimidine-DNA glycosylase): excision of purine lesions in DNA produced by ionizing radiation or photosensitization. Biochemistry 31, 106–110.

52. Aumann, K.D., Bedorf, N., Brigelius-Flohe´ , R., Schomburg, D. & Flohe´ , L. (1997) Glutathione peroxidase revisited – simulation of the catalytic cycle by computer-assisted molecular modelling. Biomed. Environ. Sci. 10, 136–155. 39. Floyd, R.A., West, M.S., Eneff, K.L. & Schneider, J.E. (1990) Mediation of 8-hydroxy-guanine formation in DNA by thiazin dyes plus light. Free Radic. Biol. Med. 8, 327–330.

53. Seo, Y.R., Kelley, M.R. & Smith, M.L. (2002) Selenomethionine regulation of p53 by a ref1-dependent redox mechanism. Proc. Natl Acad. Sci. USA 99, 14548–14553. 54. Bettger, W.J. (1993) Zinc and selenium, site-specific versus general antioxidation. Can. J. Physiol. Pharmacol. 71, 721–724. 40. Robins, P., Jones, C.J., Biggerstaff, M., Lindahl, T. & Wood, R.D. (1991) Complementation of DNA repair in xeroderma pigmen- tosum group A cell extracts by a protein with affinity for damaged DNA. EMBO J. 10, 3913–3921.

41. Jones, C.J. & Wood, R.D. (1993) Preferential binding of the xeroderma pigmentosum group A complementing protein to damaged DNA. Biochemistry 32, 12096–12104. 55. Knoepfel, L., Steinkuhler, C., Carri, M.T. & Rotilio, G. (1994) Role of zinc-coordination and of the glutathione redox couple in the redox susceptibility of human transcription factor Sp1. Biochem. Biophys. Res. Commun. 201, 871–877.

56. Robbins, A.H., McRee, D.E., Williamson, M., Collett, S.A., Xuong, N.H., Furey, W.F., Wang, B.C. & Stout, C.D. (1991) Refined crystal structure of Cd, Zn metallothionein at 2.0 A resolution. J. Mol. Biol. 221, 1269–1293.

42. Asahina, H., Kuraoka, I., Shirakawa, M., Morita, E.H., Miura, N., Miyamoto, I., Ohtsuka, E., Okada, Y. & Tanaka, K. (1994) The XPA protein is a zinc metalloprotein with an ability to recognize various kinds of DNA damage. Mutat. Res. 315, 229–237. 43. Hwang, C., Sinskey, A.J. & Lodish, H.F. (1992) Oxidized redox state of glutathione in the endoplasmic reticulum. Science 257, 1496–1502. 57. Romero-Isart, N. & Vasak, M. (2002) Advances in the struc- ture and chemistry of metallothioneins. J. Inorg. Biochem. 88, 388– 396.

44. Beck, M.A. (2001) Antioxidants and viral infections: host immune response and viral pathogenicity. J. Am. Coll. Nutr. 20, 384S– 388S.

45. Buchko, G.W., Hess, N.J., Bandaru, V., Wallace, S.S. & Kennedy, M.A. (2000) Spectroscopic studies of zinc (II)- and cobalt 58. Ikegami, T., Kuraoka, I., Saijo, M., Kodo, N., Kyogoku, Y., Morikawa, K., Tanaka, K. & Shirakawa, M. (1998) Solution structure of the DNA- and RPA-binding domain of the human repair factor XPA. Nat. Struct. Biol. 5, 701–706.