Quinone Generated by Tyrosinase-Catalyzed Oxidation of the Skin

Jul 17, 2009 - mechanism of covalent binding could take place in the case of monohydric ..... min some residual 1 and the formation of a single main p...
0 downloads 0 Views 658KB Size
Download PDF
1398

Chem. Res. Toxicol. 2009, 22, 1398–1405

A Reactive ortho-Quinone Generated by Tyrosinase-Catalyzed Oxidation of the Skin Depigmenting Agent Monobenzone: Self-Coupling and Thiol-Conjugation Reactions and Possible Implications for Melanocyte Toxicity Paola Manini,† Alessandra Napolitano,† Wiete Westerhof,‡ Patrick A. Riley,§ and Marco d’Ischia*,† Department of Organic Chemistry and Biochemistry, UniVersity of Naples Federico II, Via Cinthia 4, I-80126 Naples, Italy, Color Foundation, Kanaalweg 23A, 1121DP Landsmeer, The Netherlands, and Totteridge Institute for AdVanced Studies, The Grange, Grange AVenue, London N20 8AB, U.K. ReceiVed January 16, 2009

Monobenzone (hydroquinone monobenzylether, 1) is a potent skin depigmenting agent that causes irreversible loss of epidermal melanocytes by way of a tyrosinase-dependent mechanism so far little understood. Herein, we show that 1 can be oxidized by mushroom tyrosinase to an unstable o-quinone (1-quinone) that has been characterized by comparison of its properties with those of a synthetic sample obtained by o-iodoxybenzoic acid-mediated oxidation of 1. Preparative scale oxidation of 1 with tyrosinase and catalytic L-DOPA, followed by reductive workup and acetylation, led to the isolation of two main products that were identified as the acetylated catechol derivative 4 and an unusual biphenyl-type dimer of 4, acetylated 5, arising evidently by coupling of 4 with 1-quinone. In the presence of L-cysteine or N-acetyl-L-cysteine, formation of 4 and 5 was inhibited, and the reaction led instead to monoadducts (6 or 9) and diadducts (7 and 8). A similar behavior was observed when the tyrosinase-promoted oxidation of 1 was carried out in the presence of sulfhydryl-containing peptides, such as reduced glutathione, or proteins, such as bovine serum albumin (BSA), as inferred by detection of adduct 9 by high pressure liquid chromatography-electrochemical detection (HPLC-ED) after acid hydrolysis. The generation and reaction chemistry of 1-quinone described in this article may bear relevance to the etiopathogenetic mechanisms of monobenzone-induced leukoderma as well as to the recently proposed haptenation hypothesis of vitiligo, a disabling pigmentary disorder characterized by irreversible melanocyte loss. Introduction Monobenzone (hydroquinone monobenzylether, 1) is a potent skin depigmenting agent which can induce a permanent loss of melanocytes by simple contact in susceptible individuals (1-4). The overall effect is the onset of a condition termed occupational vitiligo, which is clinically and histologically indistinguishable from vitiligo vulgaris (idiopathic vitiligo) (5-7). Vitiligo is a multifactorial disease that affects approximately 1% of the population worldwide (8, 9). It leads to the formation of large patches of amelanotic skin of variable size generally on hands, wrist, lips and eyelids. Although three main hypotheses have been put forward regarding the etiopathology of the idiopathic disease (10, 11), including a genetic component responsible for melanocyte fragility and susceptibility to apoptosis, the event that triggers depigmentation is still unknown. In this respect, contact or occupational vitiligo may be regarded as a unique form of vitiligo for which the etiological factor is known and is related to exposure to 1 or other phenolic compounds, such as 4-tert-butylphenol (2) and 4-hydroxyanisole (3), that are generally employed in the polymer industries and during the production of paints and lacquers, cosmetics, and pharmaceuticals (12-18). * Corresponding author. E-mail: [email protected]. † University of Naples Federico II. ‡ Color Foundation. § Totteridge Institute for Advanced Studies.

Extensive studies on 2 showed that melanocytes are more sensitive than keratinocytes to the cytotoxic effect of this phenol, which can cause apoptosis, DNA fragmentation, and membrane blebbing (19). Despite the reported dermatological use of 1 to treat remaining disfiguring pigmentation in individuals affected by idiopathic vitiligo (20, 21), the precise mechanism of toxicity of 1 and related phenolic compounds has remained so far little understood. The structural similarity with tyrosine, the key melanin precursor in melanocytes, has led to the suggestion that these compounds can be oxidized by tyrosinase with a mechanism akin to the first step of melanin biosynthesis (22-25), leading to quinone compounds (26, 27). Quinones behave as Michael acceptors, e.g., with thiol compounds, and can cause cellular damage through adduct formation with crucial proteins and/or DNA (28). An attractive consequence in this regard, which has recently been addressed in detail (9), is that the covalent binding of o-quinones to proteins, including tyrosinase itself, enables the generation of a catecholic substrate, which

10.1021/tx900018q CCC: $40.75  2009 American Chemical Society Published on Web 07/17/2009

Monobenzone-Quinone and Thiol Adducts

Figure 1. Suggested mechanisms of melanocyte toxicity of phenolic skin depigmenting agents.

acts as an electron donor for the bicupric active site of Mettyrosinase, and may lead to the formation of an antigen. This mechanism of covalent binding could take place in the case of monohydric phenol substrates of tyrosinase which are unable to undergo an intramolecular cyclization reaction. A related mechanism proposed for the irreversible inactivation of tyrosinase has been ascribed to the cresolase activity of the enzyme capable of effecting the ortho-hydroxylation also of catecholic substrates, during which process reduction of Cu(II) to Cu(0) at the active site occurs with consequent release from the active site (reductive elimination) (29, 30). Alternatively, quinones can take part to redox cycle with their semiquinone radicals, leading to the formation of reactive oxygen species (ROS1), including superoxide, hydrogen peroxide, and ultimately hydroxyl radical, that can induce an inflammatory response in the skin. Moreover, within functionally active melanocytes, quinone generation would interfere with melanin biosynthesis by engaging tyrosinase in competing processes with possible onset of toxicity (Figure 1). Evidence of the formation of o-quinones by the action of tyrosinase on skin depigmenting phenols has been reported in the case of 2, 3 and 4-tert-butylcatechol (31-33), and the formation of alkylthio-adducts (e.g., cysteinyl-adducts) has been demonstrated in all cases, suggesting that these quinones have the chemical potential to bind covalently to protein targets. This reactivity was proposed as early as 1970 by Riley who reported on the selective incorporation of radiolabeled 3 into melanocytes, particularly into melanosomes where tyrosinase is compartimentalized (34). This result has been corroborated by further experiments carried out by the same group using 14C-labeled 3, demonstrating the covalent binding of the corresponding o-quinone to tyrosinase (35). Although consistent evidence points to tyrosinase-mediated transformation of phenolic substrates as the main factor 1 Abbreviations: ROS, reactive oxygen species; L-DOPA, L-3,4-dihydroxyphenylalanine; HPLC-ED, high pressure liquid chromatographyelectrochemical detection; BSA, bovine serum albumin; PLC, preparative thin layer chromatography; IBX, o-iodoxybenzoic acid; ESI, electrospray ionization; HSQC, heteronuclear single quantum coherence; HMBC, heteronuclear multiple bond correlation; COSY, correlation spectroscopy; TTA, tyramide-based tyrosinase assay.

Chem. Res. Toxicol., Vol. 22, No. 8, 2009 1399

responsible for contact/occupational vitiligo, the actual role of this enzyme has still to be assessed. The situation is complicated by the fact that some phenolic agents exhibit effects that are independent of tyrosinase-catalyzed oxidation. For example, 3 has been shown to interfere with mitochondrial oxidation and microsomal enzyme systems, to inhibit ribonucleotide reductase, to affect cell surface properties, to interfere with the cell cycle, and to initiate S-phase cell death in the absence of tyrosinase activity (36). In this regard, Boissy and co-workers (1) concluded that tyrosinase has no role in mediating the cytotoxic effects of 2, suggesting an apoptotic process as the cause of cell death. In addition, Picardo et al. found that tyrosinase does not alter the sensitivity of cells to catecholic compounds (37). It should be noted, however, that the above results were obtained in in Vitro systems lacking cell-mediated cytotoxicity (e.g., granulocytes or lymphocytes). Whether 1 is also acted upon by tyrosinase to give a reactive quinone has remained so far an open issue. To fill this void, we have undertaken an investigation into the oxidation of 1 by tyrosinase, with a view to assessing on a chemical basis the viability of a tyrosinase-mediated mechanism of toxicity to melanocytes. Specific aims of the study were to chemically characterize the primary quinone and its decay products under biomimetic conditions and to explore its reactions with sulfhydryl-containing amino acids, peptides, and proteins.

Experimental Procedures Oxidation of 1 with Tyrosinase: General Procedure. A solution of 1 (4 mg, 20 µmol) in methanol (1 mL) was added to 20 mL of 100 mM sodium phosphate buffer (pH 7.4) and then treated with tyrosinase (270 U/mL) at 37 °C under a flux of oxygen. Aliquots of the reaction mixture were periodically withdrawn and subjected to chromatographic or spectroscopic analysis. Similar experiments were carried out in the presence of L-3,4-dihydroxyphenylalanine (L-DOPA) (5% mol). Oxidation of 1 with Tyrosinase in the Presence of Sulfhydryl Compounds: General Procedure. A solution of 1 (4 mg, 20 µmol) in methanol (1 mL) was added to 20 mL of 100 mM sodium phosphate buffer (pH 7.4) and then treated with the appropriate sulphydryl compound (2 mol equiv) and tyrosinase (270 U/mL) at 37 °C under a flux of oxygen. Aliquots of the reaction mixture were periodically withdrawn and subjected to HPLC-UV and LC-MS analysis. With reduced glutathione, the mixture was subjected to acid hydrolysis and alumina purification (see below) prior to high pressure liquid chromatography-electrochemical detection (HPLC-ED) analysis. When bovine serum albumin (BSA) was used, the reaction was carried out as follows: a solution of 1 (1 mg, 5 µmol) in methanol (100 µL) was added to 890 µL of 100 mM sodium phosphate buffer (pH 7.4) and treated at 37 °C and under a flux of oxygen with BSA (66 mg, 1 µmol) and 10 µL of a solution of tyrosinase (2 mg in 200 µL) in 100 mM sodium phosphate buffer (pH 7.4) (270 U/mL in the resulting mixture). After 5 min, the reaction was stopped by adding 1 mL of 10% trichloroacetic acid, and the precipitate was collected after 1 h at 4 °C by centrifugation. The precipitate was washed twice with 5% trichloroacetic acid and subjected to acid hydrolysis, purification on alumina, and HPLC-ED analysis (see below). Acid Hydrolysis and Adduct Analysis: General Procedure. The acid hydrolysis and product analysis of the reaction mixtures were carried out as reported (38). The appropriate solution or precipitate was treated with 1 mL of 6 M HCl containing 5% thioglycolic acid at 110 °C under an argon atmosphere. After 16 h, the reaction mixture was brought to room temperature and purified on alumina. In a conical plastic tube were placed 50 mg of alumina, 0.1 mL of 2% Na2S2O5, and 0.2 mL of the hydrolyzate. Catechols were adsorbed onto alumina at pH 8.6 by adding 1 mL of 2.7 M Tris containing 2% Na2EDTA and immediately shaking for 5 min.

1400

Chem. Res. Toxicol., Vol. 22, No. 8, 2009

Manini et al.

Table 1. 1H and 13C NMR Data of Compounds 6, 7, 8 (CD3OD), and 9 (D2O/DCl) 6 13

carbon 1 2 3 4 5 6 OCH2 C-1′ C-2′, C-3′, C-4′, C-5′, C-6′ CH2S CH COOH CONH CH3 a

C NMR

147.8 142.3 123.3 111.3 154.5 106.1 72.7 138.8 126.6, 127.6, 128.4 28.9 57.7 177.1 173.8 22.3

1

H NMR (mult.)

6.58 (d, 2.7 Hz) 6.49 (d, 2.7 Hz) 4.95 (s) 7.2-7.4 (m) 2.90 (m) 4.46 (t, 5.1 Hz) 1.97 (s)

7 13

C NMR

147.8 142.1 124.3 112.1 153.8 105.1 72.7 137.7 126.6, 127.6, 128.4 37.3a, 37.8a 53.0a, 53.1a 177.2 173.7 22.3

8

1

H NMR (mult.)

6.47 (s) 4.93 (s) 7.2-7.4 (m) 3.1-3.5 (m) 4.6-4.7 (m) 1.90 (s)

13

C NMR

147.7 142.5 125.5 104.9 155.1 116.3 72.7 137.7 126.6, 127.6, 128.4 36.9a, 38.1a 53.4a, 53.7a 177.2 173.7 22.3

9

1

H NMR (mult.)

6.65 (s) 4.93 (s) 7.2-7.4 (m) 3.1-3.5 (m) 4.6-4.7 (m) 1.90 (s)

13

C NMR

146.5 141.1 119.5 112.8 153.3 106.7 71.7 137.7 128.9, 129.1, 129.7 25.2 53.4 177.8 -

1

H NMR (mult.)

6.49 (d, 2.7 Hz) 6.37 (d, 2.7 Hz) 5.28 (s) 6.8-7.1 (m) 3.21 (m) 4.45 (t, 4.8 Hz)

Interchangeable.

The alumina was washed three times with water, and then catechols were eluted with 0.3 mL of 0.4 M HClO4. The fraction eluted was subjected to HPLC-ED analysis using an octadecylsilane-coated column (Sphereclone C18-4.6 × 250 mm, 5 µm) and the following elution conditions: 0.2% formic acid/methanol 55:45 (v/v) and 1 mL/min flow rate. The detector was set at +450 mV versus an Ag/AgCl reference electrode. Isolation of 4-Benzyloxy-1,2-diacetoxybenzene (Acetylated 4) and 2,2′-Dibenzyloxy-4,4′,5,5′-tetraacetoxybiphenyl (Acetylated 5). A solution of 1 (100 mg, 0.5 mmol) in methanol (1 mL) was added to 100 mM sodium phosphate buffer (pH 7.4) (55 mL) at 37 °C and was treated with L-DOPA (5 mg, 0.025 mmol). Tyrosinase (270 U/mL) in water (1 mL) was then added, and the reaction mixture was fluxed vigorously with oxygen. After 1 h, the mixture was reduced with sodium dithionite and extracted twice with ethyl acetate. The organic layers were collected, dried over anhydrous sodium sulfate, and evaporated under reduced pressure. The residue was treated with acetic anhydride (1 mL) and pyridine (50 µL) overnight. Preparative thin layer chromatography (PLC) (eluant a: cyclohexane/ethyl acetate 4:6 (v/v)) of the resulting mixture afforded, besides some acetylated 1, pure acetylated 4 (20 mg, yields 13%, Rf ) 0.89 eluant a) and acetylated 5 (40 mg, yields 26%, Rf ) 0.77 eluant a). Acetylated 4. 1H NMR (400 MHz, CDCl3) δ (ppm): 2.27 (3H, s, CH3), 2.28 (3H, s, CH3), 5.03 (2H, s, CH2O), 6.82 (1H, d J ) 2.8 Hz, H-3), 6.85 (1H, dd J ) 8.8, 2.8 Hz, H-5), 7.08 (1H, d J ) 8.8 Hz, H-6), 7.32-7.39 (5H, m, H-2′, H-3′, H-4′, H-5′, H-6′). 13C NMR (75 MHz, CDCl3) δ (ppm): 20.6 (2 × CH3), 70.6 (OCH2), 110.0 (C-3), 112.7 (C-5), 123.6 (C-6), 127.5 (2 × CH), 128.1 (CH), 128.6 (2 × CH), 135.8 (C-1), 136.4 (C-1′), 142.5 (C-2), 156.9 (C-4), 168.2 (OCOCH3), 168.7 (OCOCH3). ESI(+)/MS: m/z 301 ([M + H]+). Acetylated 5. 1H NMR (400 MHz, CDCl3) δ (ppm): 2.22 (2 × 3H, s, 2 × CH3), 2.24 (2 × 3H, s, 2 × CH3), 4.91 (2 × 2H, s, 2 × OCH2), 6.78 (2 × 1H, s, 2 × H-3), 7.12 (2 × 1H, s, 2 × H-6), 7.15-7.25 (2 × 5H, m, 2 × H-2′′, 2 × H-3′′, 2 × H-4′′, 2 × H-5′′, 2 × H-6′′). 13C NMR (50 MHz, CDCl3) δ (ppm): 20.5 (2 × CH3), 20.7 (2 × CH3), 70.9 (2 × OCH2), 108.3 (2 × C-3), 124.9 (2 × C-1), 125.8 (2 × C-6), 126.6 (4 × CH), 127.6 (2 × CH), 128.4 (4 × CH), 135.3 (2 × C-5), 136.5 (2 × C-1′′), 141.9 (2 × C-4), 154.2 (2 × C-2), 168.1 (2 × OCOCH3), 168.4 (2 × OCOCH3). ESI(+)/ MS: m/z 599 ([M + H]+). Isolation of 3-S-[(N-Acetyl)cysteinyl]-5-benzyloxy-1,2-dihydroxybenzene (6), 3,4-Di-S-[(N-acetyl)cysteinyl]-5-benzyloxy-1,2-dihydroxybenzene (7), and 3,6-Di-S-[(N-acetyl)cysteinyl]-5-benzyloxy1,2-dihydroxybenzene (8). A solution of 1 (100 mg, 0.5 mmol) in methanol (1 mL) was added to 100 mM sodium phosphate buffer (pH 7.4) (500 mL) at 37 °C and treated with N-acetyl-L-cysteine (163 mg, 1.0 mmol). Tyrosinase (270 U/mL) in water (1 mL) was then added, and the reaction mixture was fluxed vigorously with oxygen. After 30 min, the reaction mixture was acidified to pH 3 with 2 M HCl and lyophilized. The residue taken up in H2O was fractionated by preparative HPLC using 1% acetic acid/acetonitrile

Figure 2. Reaction of 1 with tyrosinase as monitored at 420 nm in the presence (9) or in the absence ([) of added L-DOPA.

Figure 3. HPLC analysis of the reaction mixture of 1 with tyrosinase after 5 min, with UV detection set at 280 or 420 nm.

60:40 (v/v) as eluant. The fractions eluted at 15 and 7 min were collected, evaporated under reduced pressure, and found to consist of pure 6 (57 mg, yields 30%) and a mixture of 7 and 8 (40 mg, 15% yields), respectively. Compound 6. 1H NMR (300 MHz, CD3OD) δ (ppm): see Table 1. 13C NMR (50 MHz, CDCl3) δ (ppm): see Table 1. ESI(+)/MS: m/z 378 ([M + H]+). Compounds 7 and 8 (Mixture). 1H NMR (400 MHz, CD3OD) δ (ppm): see Table 1. 13C NMR (50 MHz, CDCl3) δ (ppm): see Table 1. ESI(+)/MS: m/z 539 ([M + H]+). Isolation of 3-S-Cysteinyl-5-benzyloxy-1,2-dihydroxybenzene (9). A solution of 1 (100 mg, 0.5 mmol) in methanol (1 mL) was added to 100 mM sodium phosphate buffer (pH 7.4) (100 mL) at 37 °C and treated with L-cysteine (121 mg, 1.0 mmol). Tyrosinase (270 U/mL) in water (1 mL) was then added, and the reaction mixture was fluxed vigorously with oxygen. After 5 min, the mixture was acidified to pH 3 with 2 M HCl and lyophilized. The residue taken up in H2O was fractionated by preparative HPLC using 1% acetic acid/acetonitrile 55:45 (v/v) as eluant. The fraction eluted at 10

Monobenzone-Quinone and Thiol Adducts

Chem. Res. Toxicol., Vol. 22, No. 8, 2009 1401

Figure 4. LC-MS analysis of the reaction mixture of 1 and N-acetyl-L-cysteine (a) or L-cysteine (b) with tyrosinase after 30 min and ESI(+) mass spectra of peaks I (c), II (d), and III (e).

min was collected and evaporated under reduced pressure to give pure 9 (67 mg, yield 40%). Compound 9. 1H NMR (200 MHz, D2O/DCl) δ (ppm): see Table 1. 13C NMR (50 MHz, D2O/DCl) δ (ppm): see Table 1. ESI(+)/ MS: m/z 336 ([M + H]+). Preparation of 4-Benzyloxy-1,2-benzoquinone (1-quinone). A 4 mM solution of 1 (40 mg, 0.2 mmol) in a 9:1 (v/v) mixture of chloroform/methanol was treated with o-iodoxybenzoic acid (IBX) (112 mg, 0.4 mmol) at room temperature and under vigorous stirring. After 30 min, the reaction mixture was extracted with 100 mM sodium phosphate buffer (pH 7.4)/ chloroform, and the organic layers were collected, dried over anhydrous sodium sulfate, and carefully evaporated under reduced pressure to afford pure 1-quinone (34 mg, 80% yields). 1-Quinone. 1H NMR (200 MHz, CDCl3) δ (ppm): 5.04 (2H, s, OCH2), 5.88 (1H, d J ) 3.0 Hz, H-3), 6.42 (1H, d J ) 10.6 Hz, H-6), 6.93 (1H, dd J ) 10.6, 3.0 Hz, H-5), 7.39-7.43 (5H, m, H-2′, H-3′, H-4′, H-5′, H-6′). 13C NMR (50 MHz, CDCl3) δ (ppm): 71.9 (OCH2), 103.9 (CH), 127.9 (2 × CH), 128.8 (2 × CH), 129.0 (CH), 130.6 (CH), 133.6 (C), 139.1 (CH), 167.6 (C), 177.7 (C), 179.9 (C); LC-MS 13 min. ESI(+)/MS: m/z 237 ([M + Na]+), 253 ([M + K]+).

Results Tyrosinase-Catalyzed Oxidation of 1. The reaction of 1 with mushroom tyrosinase under biomimetic conditions was carried out in 100 mM sodium phosphate buffer (pH 7.4) at 37 °C, by incubating the phenolic compound (1 mM) with the enzyme (270 U/mL) under constant air bubbling and vigorous stirring. After the addition of the enzyme, the reaction mixture immediately turned yellow due to the development of an absorption maximum centered at 420 nm. Kinetic monitoring of the reaction course at 420 nm showed the expected lag phase which was removed in the presence of a catalytic amount (5%) of L-DOPA, a known cofactor of the enzyme (35, 39) (Figure 2). HPLC-UV analysis of the reaction course revealed after 5 min some residual 1 and the formation of a single main product (A) detectable both at 280 and 420 nm, which was unstable and completely decomposed over a period of time of 10 min (Figure 3). LC-MS (electrospray ionization, ESI(+)) analysis of the mixture indicated for product A pseudomolecular ion peaks at m/z 237 and 253 ([M + Na]+ and [M + K]+) compatible with

1402

Chem. Res. Toxicol., Vol. 22, No. 8, 2009

a quinone of 1. To secure this formulation, authentic 1-quinone was prepared chemically by oxidation of 1 with IBX, a hypervalent iodine compound typically used to prepare oquinones from the corresponding phenols (40-43).

The yellow organic extract of the reaction mixture revealed on TLC analysis the presence of a single compound (λmax ) 420 nm) featuring in the 1H NMR spectrum a distinct ABX spin system at δ 5.88 (d, J ) 3.0 Hz), 6.42 (d, J ) 10.6 Hz), and 6.93 (dd, J ) 10.6, 3.0 Hz), and the signals expected for the benzyloxyl moiety. The 13C NMR spectrum revealed the presence of 13 signals, including three deshielded quaternary carbon resonances at δ 167.6, 177.7, and 179.9, consistent with the expected o-quinone structure. When kept in CDCl3 solution, 1-quinone rapidly decomposed preventing further spectral characterization. Both HPLC-UV and LC-MS analysis then provided definitive evidence for the structural identity between product A and the synthetic 1-quinone. In another series of experiments, the oxidation reaction of 1 was carried out on a preparative scale. To this purpose, a catalytic amount of L-DOPA proved to be necessary to remove the initial lag period in the tyrosinase-catalyzed oxidation and to hasten the reaction. After 1 h, when the starting material was almost completely consumed, the reaction mixture was treated with sodium dithionite to reduce quinones, extracted with ethyl acetate, and then acetylated. TLC analysis revealed the presence of two main products, B (Rf ) 0.89) and C (Rf ) 0.77), which could be isolated by preparative TLC and subjected to spectroscopic analysis. Compound B (pseudomolecular ion peak ([M + H]+) at m/z 301) was readily identified as the acetyl derivative of 4-benzyloxybenzene-1,2-diol (acetylated 4), which provided further evidence for o-quinone formation. Quite unexpectedly, compound C was dimeric in nature (pseudomolecular ion peak ([M + H]+) at m/z 599). It showed in the 1H NMR spectrum a multiplet at δ 7.2 (5H) and two singlets at δ 7.12 and 6.78 (1H each), for aromatic protons. Close inspection of the 1H,13C heteronuclear single quantum coherence (HSQC) and 1 13 H, C heteronuclear multiple bond correlation (HMBC) spectra allowed formulation of the compound as the acetyl derivative of 2,2′-dibenzyloxy-4,4′,5,5′-tetrahydroxybiphenyl (acetylated 5), i.e., the biphenyl-type dimer of 4. The position of the linkage between the two aromatic units was deduced by the lack of coupling between the aromatic proton signals at δ 7.12 and 6.78, as confirmed by the 1H,1H COSY spectrum (see Supporting Information), that rules out all other modes of coupling.

Significantly higher formation yields of 5 were obtained by treating 1-quinone with equimolar amounts of the catechol 4 rather than using 1-quinone alone. Moreover, small amounts of 5 were also observed when the tyrosinase-catalyzed reaction was run in the absence of L-DOPA, although under these latter conditions substrate consumption was much slower.

Manini et al.

Reaction of 1-Quinone with Thiol Compounds. Further experiments were aimed at assessing the reaction behavior of 1-quinone toward representative sulfhydryl compounds, including amino acids such as L-cysteine and N-acetyl-L-cysteine, small peptides, such as reduced glutathione, and proteins such as bovine serum albumin (BSA). When the tyrosinase-catalyzed oxidation of 1 (1 mM) was carried out in the presence of 2 mol equiv of N-acetyl-L-cysteine or L-cysteine, LC-MS analysis of the reaction mixtures after 30 min revealed the presence of, besides residual 1, two main detectable peaks, I and II (eluant II), in the case of N-acetyl-L-cysteine (Figure 4a), and one main peak (eluant III), in the case of L-cysteine (Figure 4b). In the former reaction, peak I eluted at 9.7 min gave a pseudomolecular ion peak at m/z 378 compatible with an N-acetyl-L-cysteine-1 adduct (Figure 4c), whereas peak II eluted at 5.2 min gave a pseudomolecular ion peak at m/z 539 suggesting a diadduct (Figure 4d). With L-cysteine, the main species, eluted at 9.4 min, gave a pseudomolecular ion peak at m/z 336 compatible with an adduct to 1-quinone (Figure 4e). For product isolation, two preparative scale reactions were carried out, and the resulting mixtures were subjected to preparative HPLC. In the case of N-acetyl-L-cysteine, two

Figure 5. HPLC-ED analysis of the tyrosinase-mediated oxidation mixture of 1 and reduced glutathione (a) or BSA (b) after acid hydrolysis and purification onto alumina. For comparison, the elutograms of standard 9 (d) and of a control mixture obtained as in b but in the absence of BSA (c) are also reported. All of the experiments have been carried out at least in duplicate.

Monobenzone-Quinone and Thiol Adducts

fractions were obtained, one containing 3-S-[(N-acetyl)cysteinyl]-5-benzyloxy-1,2-dihydroxybenzene (6), corresponding to peak I, and the other consisting of a mixture of the regioisomeric diadducts 3,4-di-S-[(N-acetyl)cysteinyl]-5-benzyloxy-1,2-dihydroxybenzene (7) and 3,6-di-S-[(N-acetyl)cysteinyl]-5-benzyloxy-1,2-dihydroxybenzene (8) (peak II). The similar elutographic behavior of 7 and 8 under a variety of conditions made all attempts at separation fruitless. In the case of L-cysteine, a single main product was consistently obtained, which was identified as 3-S-cysteinyl-5-benzyloxy-1,2-dihydroxybenzene (9). Structural assignments for all compounds were based on extensive mono- and bidimensional NMR analysis (Table 1).

When the same reaction was carried out with reduced glutathione, LC-MS analysis showed complete consumption of the starting compound and the formation of two main little retained products, giving pseudomolecular ion peaks at m/z 522 and 827, corresponding to a mono and a diadduct (see Supporting Information). Formation of the adduct of reduced glutathione to 1-quinone was supported by analysis of the products obtained after acid hydrolysis of the reaction mixture. After 16 h at 110 °C in 6 M HCl, the reaction mixture was chromatographed on alumina, and the fraction eluted with 0.4 M HClO4 was subjected to HPLC-ED analysis. This revealed the presence of a main peak eluted at 9.4 min (Figure 5a) that was identified as 9 by comparison with an authentic sample prepared by reaction with L-cysteine as described above (Figure 5d). Finally, the tyrosinase-catalyzed reaction of 1 with BSA was investigated according to a literature procedure (38) using the substrate at 5 mM concentration. After 5 min, the reaction was stopped by the addition of 10% trichloroacetic acid which led to precipitation of the proteic fraction. This was collected after 1 h at 4 °C by centrifugation and was subjected to acid hydrolysis and purification on alumina, as described above. The fraction thus obtained was analyzed by HPLC-ED and was shown to contain a peak eluted at 9.4 min (Figure 5b). The identity of the compound as 9 was confirmed by coinjection with an authentic sample. No detectable 9 was formed in control experiments in which the tyrosinase-catalyzed conversion of 1 was carried out in the absence of BSA as thiol donor (Figure 5c).

Chem. Res. Toxicol., Vol. 22, No. 8, 2009 1403

Discussion The clinical picture of occupational vitiligo is one of itching, redness, and scaliness, similar to contact dermatitis, which can be observed when treating vitiligo patients with 1 to remove remaining disfiguring pigmentation (1). This picture appears to be more compatible with an immunological response rather than the effect of a simple irritant. An irritant response, such as diaper dermatitis (ammoniac dermatitis), is a dose-dependent effect and can occur in every individual, whereas contact dermatitis and 1-induced vitiligo are not dose-dependent and occur in predisposed persons, e.g., vitiligo patients are predominantly HLAA2 positive (10). It has recently been hypothesized that haptenation by tyrosinase-mediated generation of o-quinones is an important mechanism in both idiopathic and occupational vitiligo (9). According to this hypothesis, the observed skin reaction of 1 can be explained by the specific interaction with the enzyme tyrosinase, leading to the formation of a reactive o-quinone which can react with tyrosinase or other melanosomal proteins by covalent binding to the sulfhydryl group or other residues (29, 37, 44-46). Being self-proteins, melanosomal proteins do not induce immune reactions; however, following modification by quinone binding they may be recognized by the immune system as nonself and may initiate a T-cell response. Regulatory T-cells, having a role in maintaining tolerance to melanocyte self-antigens (47), normally suppress CD8+ T-cells (including autoreactive T-cells), resulting in a transient immune reaction; but in vitiligo, the T-regulatory cells are decreased in number or are immature so that the tolerance against antigens (tyrosinase/o-quinone complex) is decreased and can result in autoimmunity. This is in line with an increased expression of the activation-associated surface antigen CD25. These changes presumably reflect increased antigen-mediated activation (48). Several papers support the covalent binding to proteins of o-quinones produced by tyrosinase-catalyzed oxidation of phenols. Naish-Byfield and Riley (35), for example, incubated tyrosinase with 14C ring-labeled 3 and demonstrated the covalent binding of 3-quinone to tyrosinase, suggesting reaction with nucleophilic groups, such as thiols, or amino functions of the enzyme. Han et al. (49) reassessed cutaneous melanogenesis using a tyramide-based tyrosinase assay (TTA), a simple test for tyrosinase activity in situ, which is based on the presumed tyrosinase-catalyzed conversion of tyramide to an unstable o-quinone, which then binds to nearby molecules of the enzyme on the tissue section. Whether phenol 1 is involved in a haptenation process by the action of tyrosinase cannot be asserted on the basis of present knowledge. The results of this study provide the first demonstration on chemical grounds that 1 can indeed be oxidized by tyrosinase to an o-quinone which exhibits a sufficiently long lifetime to interact with a range of biologically relevant thiol compounds and form covalent adducts. Experiments with BSA, in particular, have shown that under appropriate conditions very little tyrosinase is sufficient to mediate adduct formation with cysteine residues of proteins. BSA is a representative thiol-containing protein commonly used for binding studies (38) which has only one free, not easily accessible cysteine residue available for chemical coupling. The demonstration that in the absence of BSA no detectable amount of adduct is formed by reaction of 1 with tyrosinase would indicate that the observed adduct does come from BSA alkylation and would rule out any competition by the enzyme under the specific conditions of the assay (