Synthesis and Structure Identification of Thiol Conjugates of

and Chung S. Yang*,‡. Department of Chemical Biology, Ernest Mario School of Pharmacy, Rutgers University,. 164 Frelinghuysen Road, Piscataway, New ...
0 downloads 0 Views 172KB Size
Download PDF
1762

Chem. Res. Toxicol. 2005, 18, 1762-1769

Synthesis and Structure Identification of Thiol Conjugates of (-)-Epigallocatechin Gallate and Their Urinary Levels in Mice† Shengmin Sang,‡ Joshua D. Lambert,‡,# Jungil Hong,‡,# Shiying Tian,§,# Mao-Jung Lee,‡ Ruth E. Stark,§ Chi-Tang Ho,⊥ and Chung S. Yang*,‡ Department of Chemical Biology, Ernest Mario School of Pharmacy, Rutgers University, 164 Frelinghuysen Road, Piscataway, New Jersey 08854-8020, Department of Chemistry, College of Staten Island and Institute for Macromolecular Assemblies, City University of New York, 2800 Victory Boulevard, Staten Island, New York 10314, and Department of Food Science, Rutgers University, 65 Dudley Road, New Brunswick, New Jersey 08901-8520 Received June 8, 2005

(-)-Epigallocatechin gallate (EGCG), the most abundant and most biologically active compound in tea, has been proposed to have many beneficial health effects. The metabolic fate of EGCG, however, is not well understood. In the present study, we found that EGCG can be oxidized by peroxidase and hydrogen peroxide and then reacted with cysteine or glutathione to form conjugates. The structures of the cysteine and glutathione conjugates of EGCG were identified using 2D NMR and MS. Two thiol conjugates of EGCG (2′-cysteinyl EGCG and 2′′cysteinyl EGCG) were identified by ESI-LC-MS/MS analysis from the urine samples of mice administered 200 or 400 mg/kg EGCG, i.p. These conjugates were not found in urine samples of mice after receiving EGCG at 50 mg/kg i.p., or 2000 mg/kg i.g., or in human urine following consumption of 3 g of decaffeinated green tea solids (containing 333 mg EGCG). At high doses, EGCG is believed to be oxidized to form EGCG quinone, which can react with glutathione to form the thiol conjugates. These results suggest that detectable amounts of thiol conjugates of EGCG are formed only after rather high doses of EGCG are given to the mice.

Introduction Tea (Camellia sinensis, Theaceae) has been proposed to have many beneficial health effects, including prevention of cancer and heart disease, based on many studies in animal models and cell lines (1-3). Many of the beneficial effects of tea have been attributed to the strong antioxidative activity of the tea polyphenols: (-)-epigallocatechin-3-gallate (EGCG) (compound 1, Figure 1), (-)-epigallocatechin (EGC), (-)-epicatechin-3-gallate (ECG), and (-)-epicatechin (EC). Nevertheless, experimental evidence for this mechanism of disease prevention is insufficient. The antioxidative activity and the metabolic fate of tea catechins in animals or humans are not fully understood (1-3). Previous studies indicate that flavonoids with an o-dihydroxy or trihydroxy B ring are the most effective antioxidants (4-7). Moreover, the antioxidative activity is further increased by the presence of the trihydroxyl structure on the D ring (gallate) in EGCG and ECG (4-7). A large number of studies have shown that tea catechins effectively suppress lipid peroxidation in animal tissues and subcellular fractions such as microsomes and low-density lipoproteins (LDL) (8-11). The scaveng† Abbreviations: EGCG, (-)-epigallocatechin gallate; i.p., intraperitoneal; i.g., intragastric; GSH, glutathione; HMQC, heteronuclear multiple quantum correlation; HMBC, heteronuclear multiple band correlation; ROS, reactive oxygen species. * Corresponding author. Phone: 732-445-3400 ext. 244. Fax: 732445-0687. E-mail: [email protected]. ‡ Department of Chemical Biology, Rutgers University. § City University of New York. ⊥ Department of Food Science, Rutgers University. # These authors contributed equally to this work.

Figure 1. Structures of EGCG (1), 2′-cysteinyl EGCG (2), 2′′cysteinyl EGCG (3), 2′-glutathionyl EGCG (4), and 2′′-glutathionyl EGCG (5).

ing effects of tea catechins on free radicals, such as 2,2diphenyl-1-picrylhydrazyl (DPPH) (11), peroxyl radicals (12), superoxide anions (13), reactive nitrogen species (14), and hydroxyl radicals (15), have also been reported and considered as the major mechanism for the antioxidant activities. Among tea catechins, EGCG is the most abundant and most effective quencher of most reactive oxygen species (ROS). During the reactions of EGCG with free radicals, several oxidation products are formed. The B ring and gallate ring appear to be the principal sites of antioxidant reactions (7, 16-18). Reactions of EGCG with peroxyl radicals lead to the formation of sevenmembered B ring anhydride dimers and ring-fission compounds (7, 16). Theasinensin A, the B ring dimer of EGCG, is the major product in the reaction of EGCG with

10.1021/tx050151l CCC: $30.25 © 2005 American Chemical Society Published on Web 11/05/2005

Thiol Conjugates of (-)-Epigallocatechin Gallate

Chem. Res. Toxicol., Vol. 18, No. 11, 2005 1763

Table 1. δH (600 MHz) NMR Spectra Data of EGCG (1), 2′-Cystenyl EGCG (2), 2′′-Cystenyl EGCG (3), 2′-Glutathionyl EGCG (4), and 2′′-Glutathionyl EGCG (5) (CD3OD) (δ in ppm, J in Hz) 2 3 4 6 8 2′ 6′ 2′′ 6′′ Cys-HR Cys-Hβ

1

2

3

4

5

4.96 s 5.52 m 2.97 dd 4.8, 16.8 2.83 dd 2.4, 16.8 5.96 s 5.96 s 6.50 s 6.50 s 6.95 s 6.95 s

5.58 s 5.87 m 2.95 dd 4.2, 16.8 2.77 brd 17.4 5.86 d 2.4 5.87 d 2.4

4.98 s 5.61 m 3.01 dd 4.8, 17.4 2.93 brd 17.4 5.95 d 1.8 5.99 d 1.8 6.50 s 6.50 s

5.66 s 5.66 s 3.03 dd 4.8, 17.4 2.87 brd 16.8 5.98 s 5.98 s

4.99 s 5.63 m 3.00 m 2.94 m 5.94 s 5.98 s 6.52 s 6.52 s

6.70 s 6.79 s 6.79 s 3.51 m 3.25 m 2.83 m

Gly-HR Glu-HR Glu-Hβ Glu-Hγ

6.67 s 3.44 m 3.41 m 2.82 m

6.74 s 6.92 s 6.92 s 4.59 m 3.15 m 3.82 m 3.66 m 2.10 m 2.47 m

stable free radical DPPH (17). Recently, we found that in the presence of peroxidase, EGCG could react with hydrogen peroxide and then react with (-)-epicatechin to produce theaflavin-3-gallate (18). It has been demonstrated that tea and tea polyphenols can inhibit oxidative stress, such as DNA oxidative damage, in animals after receiving carcinogens or other types of oxidative stress, but such data on humans are limited (19, 20). In humans, the increase of plasma total antioxidant activity after tea ingestion has been observed in some, but not other, experiments (21). It is unclear how important these effects are in humans or animals in the absence of strong oxidative stress. A recent study reported that (+)-catechin can be metabolized by tyrosinase to form a cytotoxic o-quinone, which reacts with glutathione to form mono-, bi-, and triglutathione conjugates of (+)-catechin and mono- and biglutathione conjugates of a (+)-catechin dimer (22). When peroxidase and hydrogen peroxide were used, only monoglutathione conjugates of (+)-catechin were formed. In the presence of NADPH, rat liver microsomes also catalyzed oxidation of (+)-catechin leading to glutathione conjugate formation (22). However, detailed spectral analysis was not conducted to identify the chemical structures of these conjugates, and the occurrence of this type of conjugate in vivo is unknown. In the present study, we synthesized the glutathione and cysteine conjugates of EGCG (Figure 1) using the horseradish peroxidase/hydrogen peroxide system and determined the existence of these compounds in the urine samples of mice and human after they received different doses of EGCG or tea polyphenols. The structures of these conjugates have been identified using 2D NMR and MS.

Materials and Methods Materials. Decaffeinated green tea solids (dehydrated water extract of green tea, 1 g of powder derived from 6 g of dry leaves) were obtained from Unilever Best Foods Inc. (Englewood Cliffs, NJ). EGCG (100% pure) was provided by Mitsui Norin Co. Ltd. (Shizuoka, Japan). RP C-18 silica gel, Sephadex LH-20 gel, TLC plates (250 µm thickness, 2-25 µm particle size), horseradish peroxidase, H2O2, CD3OD, glutathione, cysteine, β-D-glucuronidase (EC 3.2.1.31, G-7896, from Escherichia coli with 9000 units/mg solid), and sulfatase (EC 3.1.6.1, S-9754, from Abalone entrails with 23 units/mg solid) were purchased from Sigma (St. Louis, MO). HPLC-grade solvents and other reagents were obtained from VWR Scientific (South Plainfield, NJ). HPLC-

6.63 s 4.31 m 2.84 m 3.86 m 3.68 m 2.10 m 2.47 m

grade water was prepared using a Millipore Milli-Q purification system (Bedford, MA). NMR. 1H (600 MHz), 13C (150 MHz), and all 2D NMR spectra were acquired on a Varian UnityINOVA 600 NMR spectrometer (Palo Alto, CA) equipped with a z-gradient inverse-detection triple resonance probe. 1H-13C HMQC (heteronuclear multiple quantum correlation) and HMBC (heteronuclear multiple band correlation) experiments were performed as described previously (23). Synthesis of EGCG-Thiol Conjugates (Compounds 2-5, Figure 1). 1. 2′-Cysteinyl EGCG (2) and 2′′-Cysteinyl EGCG (3). EGCG (1) (1.0 g) and cysteine (2.0 g) were dissolved in a mixture of acetone-pH 5 buffer (1:10, v/v, 50 mL), which contained 5 mg of horseradish peroxidase. While being stirred, 2 mL of 3.13% H2O2 was added 4 times during 45 min. The reaction mixture was then subjected to Sephadex LH-20 column chromatography eluted with 40% acetone/water to obtain the mixture of two major reaction products. This mixture was then subjected to RP C-18 column chromatography using 20% MeOH/ water to obtain 2 (54 mg) and 3 (180 mg). 1H and 13C NMR data of 2 and 3 are listed in Tables 1 and 2, and negative ESI-MS of these two compounds are m/z 576 [M - H]-. 2. 2′-Glutathionyl EGCG (4) and 2′′-Glutathionyl EGCG (5). Similar to the procedure for the synthesis of 2 and 3, 1.0 g of EGCG (1) and 2.0 g of glutathione were used to synthesize 60 mg of 4 and 270 mg of 5. 1H and 13C NMR data of 4 and 5 are listed in Tables 1 and 2, and negative ESI-MS of these two compounds are m/z 762-. All the purified compounds (2-5) were kept at -80 °C for further use. Treatment of Mice and Urine Collection. Experiments with mice were carried out according to a protocol approved by the Institutional Review Board for the Animal Care and Facilities Committee (IRB-ACFC nos. 91-024) at Rutgers University. Male CF-1 mice (30-35 g) were purchased from Charles River Laboratories and allowed to acclimate for at least 1 week prior to the start of the experiment. The mice were housed 10 per cage and maintained in air-conditioned quarters with a room temperature of 20 ( 2 °C, relative humidity of 50 ( 10%, and an alternating 12 h light/dark cycle. Mice were fed Purina Rodent Chow no. 5001 (Research Diets) and water and were allowed to eat and drink ad libitum. EGCG was administered to mice, i.p. (50, 200, or 400 mg/kg), and urine samples were collected in metabolism cages for 24 h after administration of vehicle or EGCG. In another experiment, mice were administered EGCG at 2000 mg/kg, i.g., and the mice (6 per group) were sacrificed at 90 and 180 min. Urine samples were collected and pooled within treatment groups. These samples were stabilized with 0.1 vol of an ascorbate/EDTA solution (1.14 mol/L ascorbic acid, 1.3 mmol/L EDTA) and stored at -80 °C before analysis. Human Urine Samples. The Institutional Review Board approved the protocol for human experiment for the Protec-

1764

Chem. Res. Toxicol., Vol. 18, No. 11, 2005

Sang et al.

Table 2. δC (150 MHz) NMR Spectra Data of EGCG (1), 2′-Cystenyl EGCG (2), 2′′-Cystenyl EGCG (3), 2′-Glutathionyl EGCG (4), and 2′′-Glutathionyl EGCG (5) (CD3OD) (δ in ppm) 2 3 4 5 6 7 8 9 10 1′ 2′ 3′ 4′ 5′ 6′ 1′′ 2′′ 3′′ 4′′ 5′′ 6′′ 7′′ Cys-CR Cys-Cβ Cys-CON Gly-CR Gly-COO Glu-CR Glu-Cβ Glu-Cγ Glu-COO Glu-CON

1

2

3

4

5

78.5 d 69.9 d 26.8 t 157.7 s 95.8 d 157.8 s 96.5 d 157.1 s 99.4 s 130.8 s 106.8 d 146.6 s 133.7 s 146.6 s 106.8 d 121.4 s 110.2 d 146.2 s 139.7 s 146.2 s 110.2 d 167.6 s

76.6 d 69.1 d 27.1 t 157.8 s 95.9 d 157.9 s 96.6 d 157.4 s 99.4 s 133.7 s 106.7 s 149.1 s 135.0 s 148.3 s 109.2 d 121.4 s 110.2 d 146.2 s 139.7 s 146.2 s 110.2 d 167.5 s 55.5 d 39.0 t 172.6 s

78.3 d 70.8 d 26.7 t 157.7 s 96.0 d 157.8 s 96.7 d 157.1 s 99.4 s 130.7 s 107.0 d 146.7 s 133.8 s 146.7 s 107.0 d 128.0 s 108.9 s 149.9 s 138.2 s 147.4 s 111.4 d 168.9 s 55.3 d 37.7 t 173.1 s

76.7 d 69.1 d 27.0 t 157.7 s 96.7 d 157.8 s 96.7 d 157.3 s 99.5 s 133.7 s 108.4 s 148.6 s 134.4 s 148.0 s 108.6 d 121.4 s 110.3 d 146.2 s 139.7 s 146.2 s 110.3 d 167.5 s 55.5 d 38.1 t 173.0 s 42.3 t 174.1 s 55.4 d 27.5 t 32.8 t 173.4 s 175.2 s

78.3 d 70.6 d 26.9 t 157.6 s 96.0 d 157.6 s 96.7 d 157.1 s 99.5 s 130.8 s 106.9 d 146.7 s 133.7 s 146.7 s 106.9 d 128.6 s 110.1 s 149.0 s 137.4 s 147.2 s 111.3 d 168.4 s 54.7 d 38.7 t 172.4 s 42.4 t 173.8 s 54.2 d 27.5 t 32.9 t 172.6 s 175.0s

tionof Human Subjects in Research (IRB nos. 92-034R) at Rutgers University (Piscataway, NJ). Two healthy male volunteers (30-32 years old, weighing 60-80 kg, nonsmokers) participated in the study. The subjects did not consume tea, tea-related products, or other herbal products for at least 2 weeks before the experiments. Urine samples were collected just before ingestion of a cup of tea (3 g of decaffeinated green tea solids containing 11.10% EGCG, 6.32% EGC, 3.80% ECG, and 3.44% EC in 150 mL of water) and at 0-3, 3-6, 6-9, 9-12, 12-18, and 18-24 h thereafter. Aliquots of 20 mL of each urine sample were transferred into plastic bottles containing 20 mg of ascorbic acid (as a preservative) and stored at -80 °C before analysis. Urine Sample Preparation. Urine sample (200 µL) was hydrolyzed with β-glucuronidase/sulfatase as described previously (24). The sample was then added to 1 mL of acetone to precipitate protein. After centrifugation, the supernatant was dried under vacuum, and the solid was resuspended in 50% aqueous acetonitrile. The recovery of compounds 2 and 3 was more than 90%. These samples were analyzed by LC-MS and HPLC-ECD. LC-ESI-MS Analysis of EGCG Cysteine Conjugates in Urine. LC-MS analysis was carried out with a Finnigan Spectra System which consisted of a Finnigan model P4000 pump, a model AS3000 refrigerated autosampler, and a Finnigan LCQ Deca mass detector (ThermoFinigan, San Jose, CA) incorporated with electrospray ionization (ESI) interface. A Supelco Discovery HS C18 column (75 mm × 2.1 mm i.d.; particle size, 3 µm) was used for separation. The column elution, at a flow rate of 0.2 mL/min, started with 6 min of isocratic phase of 100% solvent A (5% aqueous methanol with 0.2% acetic acid), followed by progressive, linear increases of solvent B (95% aqueous methanol with 0.2% acetic acid) to 20% at 15 min, 50% at 25 min, 55% at 30 min, and 100% at 36 min. The mobile phase was maintained at 100% B for 8 min and then was reequilibrated to 100% A at 45 min. The LC elute was introduced into the ESI interface. The negative ion polarity mode was set for the ESI ion source with the voltage on the ESI interface

maintained at approximately -4 kV. Nitrogen gas was used as the sheath gas at a flow rate of 80 arb and the auxiliary gas at 2 arb, respectively. The heated capillary temperature and voltage were maintained at 260 °C and -24 V, respectively. The tube lens offset voltage was -55 V. The structural information of the standard cysteine conjugates of EGCG was obtained by tandem mass spectrometry (MS/MS) through collision-induced dissociation (CID) with a relative collision energy setting of 25%. The limit of detection is ∼100 pg/mL for both 2′-cysteinyl EGCG (2) and 2′′-cysteinyl EGCG (3). HPLC-ECD Analysis. The level of EGCG cysteine conjugates was analyzed using an HPLC system, consisting of a Waters 717 refrigerated autosampler, a HITACHI L-6200A intelligent pump, and an ESA 5600 coulochem electrode array system (CEAS). The potentials of the CEAS were set at 0, 200, 400, and 600 mV. Separation was achieved using previously described methods with a slight modification (24). In brief, a Supelcosil C18 reversed-phase column (150 mm × 4.6 mm inner diameter; Supelco Co., Bellefonte, PA) was used. The column and coulochem electrode array system detector were housed in a temperature-regulated compartment maintained at 35 °C. The autosampler was maintained at 6 °C. For binary gradient elution, mobile phase A (1.75% acetonitrile and 0.12% tetrahydrofuran in 30 mM NaH2PO4, pH 3.35) and B (58.5% acetonitrile and 12.5% tetrahydrofuran in 15 mM NaH2PO4, pH 3.45) were used. The flow rate was maintained at 1 mL/min, and the mobile phase began with a 5 min isocratic phase of 96% A and 4% B. It was followed by progressive, linear increases in B to 15% at 15 min, 22% at 25 min, 80% at 32 min, and 100% at 33 min. The mobile phase was maintained at 100% B for 6 min and then was reequilibrated to 4% at 40 min for another run. The limit of detection is ∼5 ng/mL, and the limit of quantification is 50 ng/mL, for both 2′-cysteinyl EGCG (2) and 2′′-cysteinyl EGCG (3).

Results Synthesis of Thiol Conjugates of EGCG (Compounds 2-5). EGCG (1) was oxidized by peroxidase/ hydrogen peroxide to form o-quinone, which reacted with glutathione or cysteine to form thiol conjugates, compounds 2-5. These compounds were purified through column chromatography, and their structures were identified by NMR (1H, 13C, HMQC, HMBC) and MS. Compound 2 was assigned the molecular formula C25H23O13NS based on negative ion ESI-MS ([M - H]at m/z 576) and 13C NMR data. Its 1H and 13C NMR spectra showed very similar patterns as those of EGCG (Tables 1 and 2). The 1H NMR spectrum of 2 exhibits signals for the A ring (H-6, 8 at δH 5.86 d 2.4 and 5.87 d 2.4), C ring (H-2 at δH 5.58 s, H-3 at δH 5.87 m, and H-4 at δH 2.95 dd 4.2, 16.8 and 2.77 brd 17.4), and D ring (H-2′′, 6′′ at δH 6.79 s), which were similar to those of EGCG, indicating that the A, C, and D rings of 2 do not undergo any change during oxidation. In comparison with the 1H NMR spectrum of EGCG, compound 2 only showed one singlet signal for one proton of the B ring (δH 6.70 s 1H), instead of the one singlet signal for two protons of EGCG (δH 6.50 s 2H), and three additional proton signals for cysteine group (δH 3.51 m, 1H, Cys-HR; 3.25 m, 2.83 m, 2H, Cys-Hβ). The major differences in the 13C spectra of compound 2 and EGCG were (a) the presence of three additional carbons at δC 172.6 s, 55.5 d, and 39.0 t for cysteine group in 2 and (b) the quaternary carbon observed at δC 106.7 in lieu of an unsubstituted aromatic carbon from the B ring of EGCG (Tables 1 and 2). In addition, the molecular weight of 2 was 119 (M-2 of cysteine (MW 121)) mass units higher than that of EGCG. All of these spectral features sup-

Thiol Conjugates of (-)-Epigallocatechin Gallate

Chem. Res. Toxicol., Vol. 18, No. 11, 2005 1765

Figure 2. Significant HMBC (H f C) correlations of 2′-cysteinyl EGCG (2), 2′′-cysteinyl EGCG (3), 2′-glutathionyl EGCG (4), and 2′′-glutathionyl EGCG (5).

ported the presence of a cysteine group in compound 2, which was located at C-2′ of the B ring. This was confirmed by the cross-peaks in the HMBC spectrum (Figure 2) between δC 106.7 and H-2 (δH 5.58 m), H-6′ (δH 6.70 s) and cys-HR (δH 3.51 m). Thus, the cysteine group was located at the position C-2′ of the B ring. Compound 2 was identified as 2′-cysteinyl EGCG (Figure 1). The negative ion ESI-MS of 3 displayed a molecular ion peak at m/z [M - H]- 576, supporting a molecular formula of C25H23O13NS, as noted above for 2. The NMR spectra of 3 displayed signal patterns similar to those of 2 (Tables 1 and 2). The 1H NMR spectrum of 3 showed signals for the A ring (H-6, 8 at δH 5.95 d 1.8 and 5.99 d 1.8), B ring (H-2′, 6′ at δH 6.50 s), C ring (H-2 at δH 4.98 s, H-3 at δH 5.61 m, and H-4 at 3.01 dd 4.8, 17.4 and δH 2.93 brd 17.4), only one proton (singlet signal) for the D ring (δH 6.67 s 1H), and three additional proton signals for the cysteine group (δH 3.44 m, 1H, cys-HR; 3.41 m, 2.82 m, 2H, cys-Hβ). The 13C NMR spectrum of 3 displayed 25 carbon signals, 15 of which were assigned to the A, B, and C rings of flavan-3-ols, 7 of which were the signals for the gallate group, and three were for the cysteine group (δC 173.1 s, 55.3 d, and 37.7 t) (Table 2). Similarly, one quaternary carbon and one unsubstituted aromatic carbon were observed at δC 108.9 and 111.4 ppm instead of two unsubstituted aromatic carbons from the D ring. These spectra indicated the presence of a cysteine group located at the D ring in compound 3. The major differences between 2 and 3 were the location of the cysteine group. In compound 3, the cysteine group conjugated at the 2′′ position of the D ring, rather than the 2′ position of the B ring in 2. Compound 3’s HMBC spectrum showed correlations between δC 108.9 and H-6′′ (δH 6.67 s) and cys-HR (δH 3.44 m) and δC 168.9 (C-7′′ CO) and H-6′′ (Figure 2). This confirmed that the cysteine group was conjugated at 2′′ position of the D ring in 3. Therefore, compound 3 was identified as 2′′-cysteinyl EGCG (Figure 1).

Compounds 4 and 5 were the oxidation products of EGCG in the presence of glutathione. Compounds 4 and 5 had the same molecular formula, C32H33O17N3S, based on the analysis of their negative ion ESI-MS ([M - H]at m/z 762) as well as their 13C NMR data. The 1H NMR spectrum of 4 exhibited signals for the A ring (H-6, 8 at δH 5.98 s, 2H), only one proton (singlet signal) for the B ring (δH 6.67 s 1H), C ring (H-2 at δH 5.66 s, H-3 at δH 5.66 s, and H-4 at δH 3.03 dd 4.8, 17.4 and 2.87 brd 16.8), D ring (H-2′′, 6′′ at δH 6.92 s 2H), and the glutathione group (δH 4.59 m, 1H, Cys-HR; 3.15 m, 2H, Cys-Hβ; 3.82 m, 1H, Gly-HR; 3.66 m, 2H, Glu-HR; 2.10 m, 2H, Glu-Hβ; 2.47 m, 1H, Glu-Hγ), which showed a similar pattern as those of 2. These observations indicated that the glutathione group conjugates at position C-2′ of the B ring. This was further confirmed by the analysis of its 13C and APCI-MS spectra. The major difference in the 13C spectra of 4 and 2 was the presence of signals for the glutathione group in 4 rather than for the cysteine group in 2 (Table 2). In addition, the molecular weight of 4 was 305 (M-2 of glutathione (MW 307)) mass units higher than that of EGCG. All of these spectral features supported the presence of a glutathione group in compound 4, which was located at C-2′ of the B ring. The crosspeaks in the HMBC spectrum of 4 (Figure 2) also confirmed this by the signals between δC 108.4 and H-2 (δH 5.66 s), H-6′ (δH 6.74 s), and cys-HR (δH 4.59 m). Therefore, the glutathione group was located at the position C-2′ of the B ring. Compound 4 was identified as 2′-glutathionyl EGCG (Figure 1). The 1H and 13C NMR spectra of 5 also showed the presence of the glutathione group. The major differences between 5 and 4 were the location of glutathione group, which was similar to the differences between 3 and 2. In compound 5, the glutathione group conjugated at the 2′′ position of the D ring, rather than the 2′ position of the B ring in 4. The HMBC spectrum of compound 5 showed correlations between δC 110.1 and H-6′′ (δH 6.63 s) and cys-HR (δH 4.31 m) and δC 168.4 (C-7′′ CO) and H-6′′ (Figure 2). This confirmed

1766

Chem. Res. Toxicol., Vol. 18, No. 11, 2005

Sang et al.

Figure 3. HPLC-ECD spectra of urine samples of mice after receiving EGCG, i.p., at 400, 200, or 50 mg/kg, or vehicle (control).

that the glutathione group was conjugated at the 2′′ position of the D ring in 5. Therefore, compound 5 was identified as 2′′-glutathionyl EGCG (Figure 1). Full assignments of the 1H and 13C NMR signals of 2-5 were made by analyzing the signals in the HMBC and HMQC spectra (Tables 1 and 2). Analysis of EGCG Cysteine Conjugates in Urine Samples. Mouse urine samples were collected for 24 h after i.p. doses of EGCG (400, 200, and 50 mg/kg) or vehicle and analyzed after hydrolysis with β-D-glucuronidase and sulfatase. In the HPLC-ECD spectra, two peaks with the same retention times as those of 2 and 3, respectively (Figure 3), were observed in the samples from mice treated with EGCG at 400 and 200 mg/kg, but not in the 50 mg/kg group or the control group. Corresponding to the ESI negative molecular ion 576 [M - 1](the molecular ion of EGCG cysteine conjugate is 577), two peaks showed up in the LC spectrum of mouse urine samples from the 400 mg/kg and 200 mg/kg groups. These two peaks could not be detected in the urine samples from the control group or the 50 mg/kg group. To identify these peaks, these samples were analyzed together with our two standards by LC-MS/MS under the same conditions. In comparison to the EGCG cysteine conjugates (2 and 3), the two urinary metabolite peaks showed, respectively, almost the same retention time as well as the same molecular masses and fragment ion mass spectra (Figures 4 and 5). The urinary levels of 2′-cysteinyl EGCG and 2′′-cysteinyl EGCG in the 400 mg/kg group were 37.0 and 1.6 µg/mL, respectively, and in the 200 mg/kg group were 6.3 and 0.36 µg/mL, respectively. These two compounds could not be detected in mouse urine after the mice received i.g. a high dose of EGCG (2000 mg/kg). The conjugates were also not detected in human urine samples after individuals drank a solution containing 3 g of decaffeinated green tea solids (equal to 4 to 5 cups of tea).

Discussion The horseradish peroxidase/hydrogen peroxide system can oxidize a large variety of phenolic compounds. When

EGCG was incubated with horseradish peroxidase/ hydrogen peroxide, EGCG dimers were formed (25), whereas theaflavin-3-gallate was the major product when EC was also added to the incubation (26). In the present study, we found that in the presence of glutathione or cysteine, the glutathione or cysteine conjugates of EGCG were formed. Monoglutathione or cysteine conjugates, but not biglutathione or cysteine conjugates, were formed as the major products, possibly because EGCG is more easily oxidized than the monoconjugates. In addition, modification of the reaction conditions, such as increasing the amount of glutathione or cysteine and extending the reaction time, may increase the formation of bithiol conjugates. Further studies are needed to clarify this issue. N-acetyl cysteine could also react with EGCG, but less efficiently than cysteine and glutathione. Only a small amount of N-acetyl cysteine conjugates of EGCG were detected by LC-MS after reaction for 1 h. On the basis of the structures of these conjugates, a mechanism of peroxidase-catalyzed oxidation of EGCG with hydrogen peroxide is proposed (Figure 6). EGCG is oxidized to form semiquinone radical and then o-quinone at the B ring or gallate ring. These o-quinones would react with the sulfydryl group of glutathione or cysteine to form the respective adducts. The gallate ring is the more favorable site for the formation of thiol conjugates, and more 2′′-thiol conjugates (compounds 3 and 5) were formed than 2′-thiol conjugates (compounds 2 and 4). To determine whether the EGCG-thiol conjugates are formed in vivo, we analyzed for compounds 2 to 5 in urine samples collected under different experimental conditions. 2′-Cysteinyl and 2′′-cysteinyl EGCG were detected in the urine samples of mice that were treated i.p. with EGCG at doses of 200 and 400 mg/kg. Those compounds were not detected in urine samples collected from humans after consumption of an equivalent of 4 or 5 cups of green tea. They were also not detected in urine samples from mice after receiving a very high oral dose of EGCG at 2000 mg/kg or from mice treated with EGCG at a dose of 50 mg/kg, i.p. It is interesting to note that the major conjugate observed was 2′-cysteinyl EGCG. This is dif-

Thiol Conjugates of (-)-Epigallocatechin Gallate

Chem. Res. Toxicol., Vol. 18, No. 11, 2005 1767

Figure 4. LC chromatograms of standard 2′-cysteinyl EGCG (2), 2′′-cysteinyl EGCG (3) (bottom), and urine samples (top) after receiving 400 mg/kg EGCG, i.p.

Figure 5. Negative ion ESI-MS/MS spectra of standard 2′-cysteinyl EGCG (2), 2′′-cysteinyl EGCG (3) (right), and mice urine samples (left) after receiving 400 mg/kg EGCG, i.p.

ferent from the peroxidase/hydrogen peroxide system which produced mainly 2′′-cysteinyl EGCG. We found that the administration of high doses of EGCG (greater than 200 mg/kg) by intraperitoneal injection resulted in severe toxicity and rapid lethality. All the mice (5 mice/ group) died within 24 h at a dose of 400 mg/kg EGCG, i.p., and 4 out of 5 mice died within 2-3 days after

receiving 200 mg/kg EGCG, i.p. Although there are no previous toxicological studies in vivo, Schmidt et al. have reported that treatment of freshly isolated rat hepatocytes with green tea extract and EGCG resulted in dosedependent cytotoxicity (27). The concentrations at which cell death occurred were quite high (greater than 500 µg/mL) but may be achievable in the perihepatic space

1768

Chem. Res. Toxicol., Vol. 18, No. 11, 2005

Sang et al.

Figure 6. Proposed mechanism for the formation of 2′-cysteinyl EGCG (2), 2′′-cysteinyl EGCG (3), 2′-glutathionyl EGCG (4), and 2′′-glutathionyl EGCG (5) in the horseradish peroxidase/H2O2 system. EGCG was oxidized by horseradish peroxidase/H2O2 to form semiquinone radicals, which disproportionately form the o-quinone at the B ring and gallate ring that react with GSH or cysteine to form related thiol conjugates following a classical mechanism of conjugation involving nucleophilic addition of GSH or cysteine to the quinone through the sulfydryl group.

following intraperitoneal injection of EGCG. Further studies in vivo are needed to clearly define the toxicological parameters of EGCG and the other green tea polyphenols. Such studies are particularly necessary as green tea polyphenols move from the dietary to the pharmacological setting as disease preventive agents. The formation of detectable EGCG thiol conjugates appears to result from administration of toxic doses of EGCG. Under normal physiological conditions, EGCG is metabolized through methylation, glucuronidation, and sulfation. At toxic doses of EGCG, these pathways may be saturated, and the excessive amount of EGCG is oxidized to form EGCG quinone, which can react with glutathione to form glutathione conjugates. In the mercapteuric acid pathway, the glutamyl moiety of the glutathione conjugate is removed by γ-glutamyl transpeptidase, and the cysteinylglycine derivative is then hydrolyzed to yield the cysteine conjugate, which is excreted in urine (28). It is not clear how EGCG is oxidized in this proposed pathway. EGCG, with eight phenolic groups, appears to be too hydrophilic to be a good substrate for cytochrome P450 enzymes. It is possible that high concentrations of EGCG may be oxidized nonspecifically by oxygenases/peroxidases. The resulting quinone then undergoes redox cycling to produce oxidative stress beyond the capacity of antioxidative enzymes (superoxide dismutase, glutathione peroxidase, and catalase). The glutathione conjugate formation may represent a mechanism to detoxify EGCG quinone. Such phenomena have been observed in previous studies with other flavonoids, dopamine, and catechol estrogens, which undergo peroxidase- or tyrosinase-catalyzed oxidation chemically, in primary cultured astrocytes and in the liver (29-34). The thiol conjugates of catechol estrogens were detected in

urine and kidney samples of hamsters (35, 36), but no thiol conjugates of flavonoids were reported in vivo previously. The formations of thiol conjugates were dependent on the redox potential and the structure of the flavonoids (29, 30). Flavonoids with low antioxidant ability, such as flavone apigenin and flavanone naringenin, could cause extensive GSH oxidation to form GSSG instead of the formation of thiol conjugates (29). Flavonoids containing a catechol-type substituent in their B ring would be oxidized by peroxidase to form related thiol conjugates through the quinone/quinone methide pathway (30). Quercetin was the only flavonoid reported so far to form thiol conjugates at positions C6 and C8 of the A ring due to the involvement of quinone methide-type intermediates (29, 37). The absence of the C3-OH group (luteolin), the C2dC3 double bond (taxifolin), or the C5-OH group (3,3′,4′-trihydroxylflavone) would prevent the tautomeric shift of the o-quinone to the corresponding quinone methide and form the thiol conjugates in the B ring (29). Thus, theoretically EGCG would form the thiol conjugates in the B ring or gallate ring, and this was confirmed by the present results. Further studies are needed to assess the toxicological aspect of the EGCG thiol conjugate formation. Taken together, our findings indicate that, in the peroxidase/hydrogen peroxide system, EGCG can be oxidized to form the thiol conjugates in the presence of glutathione and cysteine. However, the formation of detectable levels of these conjugates occurs in vivo only at toxic doses of EGCG. To our knowledge, this is the first report on the formation and structure elucidation of thiol conjugates of EGCG and the occurrence of these reactions in vivo. Due to the poor oral bioavailability, the safety of orally administered EGCG may not be a concern.

Thiol Conjugates of (-)-Epigallocatechin Gallate

However, if EGCG is administered by i.p. or i.v. injection, the dose of EGCG should be carefully considered to avoid adverse effects.

Acknowledgment. This study was supported by NIH Grant PO1 CA88961. Funds to purchase the 600 MHz NMR spectrometer at CUNY College of Staten Island were provided by the New York State Dormitory Authority and the New York State Higher Education Applied Technology Program. The LC-MS analysis was conducted in the analytical center (directed by Dr. Brian Buckley) at the Environmental Health Sciences Center supported by NIEHS Center Grant ES-05022.

References (1) Yang, C. S., Maliakal, P., and Meng, X. (2002) Inhibition of carcinogenesis by tea. Annu. Rev. Pharmacol. Toxicol. 42, 25-54. (2) Hou, Z., Lambert, J. D., Chin, K. V., and Yang, C. S. (2004) Effects of tea polyphenols on signal transduction pathways related to cancer chemoprevention. Mutat. Res. 555, 3-19. (3) Lambert, J. D., Hong, J., Yang, G. Y., Liao, J., and Yang, C. S. (2005) Inhibition of carcinogenesis by polyphenols: evidence from laboratory investigations. Am. J. Clin. Nutr. 81, 284S-291S. (4) Jovanovic, S. V., Steenken, S., Tosic, M., Marjanovic, B., and Simic, M. G. (1994) Flavonoids as antioxidants. J. Am. Chem. Soc. 116, 4846-4851. (5) Jovanovic, S. V., Steenken, S., Hara, Y., and Simic, M. G. (1996) Reduction potential of flavonoid and model phenoxyl radicals. Which ring in flavonoids is responsible for antioxidant activity? J. Chem. Soc., Perkin Trans. 2, 2497-2504. (6) Rice-Evans, C. A., Miller, N. J., and Paganga, G. (1996) Structureantioxidant activity relationships of flavonoids and phenolic acids. Free Radical Biol. Med. 20, 933-956. (7) Valcic, S., Burr, J. A., Timmermann, B. N., and Liebler, D. C. (2000) Antioxidant chemistry of green tea catechins. New oxidation products of (-)-epigallocatechin gallate and (-)-epigallocatechin from their reactions with peroxyl radicals. Chem. Res. Toxicol. 13, 801-810. (8) Terao, J., Piskula, M., and Yao, Q. (1994) Protective effect of epicatechin, epicatechin gallate, and quercetin on lipid peroxidation in phospholipid bilayers. Arch. Biochem. Biophys. 308, 278284. (9) Yang, T. T., and Koo, M. W. (2000) Inhibitory effect of Chinese green tea on endothelial cell-induced LDL oxidation. Atherosclerosis 148, 67-73. (10) Salah, N., Miller, N. J., Paganga, G., Tijburg, L., Bolwell, G. P., and Rice-Evans, C. (1995) Polyphenolic flavanols as scavengers of aqueous phase radicals and as chain-breaking antioxidants. Arch. Biochem. Biophys. 322, 339-346. (11) Nanjo, F., Goto, K., Seto, R., Suzuki, M., Sakai, M., and Hara, Y. (1996) Scavenging effects of tea catechins and their derivatives on 1,1-diphenyl-2-picrylhydrazyl radical. Free Radical Biol. Med. 21, 895-902. (12) Kondo, K., Kurihara, M., Miyata, N., Suzuki, T., and Toyoda, M. (1999) Scavenging mechanisms of (-)-epigallocatechin gallate and (-)-epicatechin gallate on peroxyl radicals and formation of superoxide during the inhibitory action. Free Radical Biol. Med. 27, 855-863. (13) Guo, Q., Zhao, B., Shen, S., Hou, J., Hu, J., and Xin, W. (1999) ESR study on the structure-antioxidant activity relationship of tea catechins and their epimers. Biochim. Biophys. Acta 1427, 13-23. (14) Panzella, L., Manini, P., Napolitano, A., and d’Ischia, M. (2005) The acid-promoted reaction of the green tea polyphenol epigallocatechin gallate with nitrite ions. Chem. Res. Toxicol. 18, 722729. (15) Hanasaki, Y., Ogawa, S., and Fukui, S. (1994) The correlation between active oxygens scavenging and antioxidative effects of flavonoids. Free Radical Biol. Med. 16, 845-850. (16) Valcic, S., Muders, A., Jacobsen, N. E., Liebler, D. C., and Timmermann, B. N. (1999) Antioxidant chemistry of green tea catechins. Identification of products of the reaction of (-)-epigallocatechin gallate with peroxyl radicals. Chem. Res. Toxicol. 12, 382-386. (17) Zhu, N., Wang, M. F., Wei, G. J., Lin, J. K., Yang, C. S., and Ho, C. T. (2001) Identification of reaction products of (-)-epigallocatechin, (-)-epigallocatechin gallate and pyrogallol with 2,2-diphenyl-1-picrylhydrazyl radical. Food Chem. 73, 345-349.

Chem. Res. Toxicol., Vol. 18, No. 11, 2005 1769 (18) Sang, S., Tian, S. Y., Meng, X. F., Stark, R. E., Rosen, R. T., Yang, C. S., and Ho, C. T. (2002) Theadibenzotropolone A., a new type pigment from enzymatic oxidation of (-)-epicatechin and (-)epigallocatechin gallate and characterized from black tea using LC/MS/MS. Tetrahedron Lett. 43, 7129-7133. (19) Hashimoto, R., Yaita, M., Tanaka, K., Hara, Y., and Kojo, S. (2000) Inhibition of radical reaction of apolipoprotein B-100 and R-tocopherol in human plasma by green tea catechins. J. Agric. Food. Chem. 48, 6380-6383. (20) Higdon, J. V., and Frei, B. (2003) Tea catechins and polyphenols: health effects, metabolism, and antioxidant functions. Crit. Rev. Food Sci. Nutr. 43, 89-143. (21) Hodgdon, J., Proudfoot, J. M., Croft, K. D., Puddey, I. B., and Mori, T. A. (1999) Comparison of the effects of black and green tea on in vitro lipoprotein oxidation in human serum. J. Sci. Food Agric. 79, 561-566. (22) Moridani, M. Y., Scobie, H., Salehi, P., and O’Brien, P. J. (2001) Catechin metabolism: glutathione conjugate formation catalyzed by tyrosinase, peroxidase, and cytochrome P450. Chem. Res. Toxicol. 14, 841-848. (23) Fang, X., Qiu, F., Yan, B., Wang, H., Mort, A. J., and Stark, R. E. (2001) NMR studies of molecular structure in fruit cuticle polyesters. Phytochemistry 57, 1035-1042. (24) Lee, M. J., Prabhu, S., Meng, X., Li, C., and Yang, C. S. (2000) An improved method for the determination of green and black tea polyphenols in biomatrixes by high-performance liquid chromatography with coulometric array detection. Anal. Biochem. 279, 164-169. (25) Zhu, N., Sang, S. M., Huang, T. C., Bai, N. S., Yang, C. S., and Ho, C. T. (2000) Antioxidant chemistry of green tea catechins: Oxidation products of (-)-epigallocatechin gallate and (-)-epigallocatechin with peroxidase. J. Food Lipids 7, 275-282. (26) Sang, S., Lambert, J. D., Tian, S., Hong, J., Hou, Z., Ryu, J. H., Stark, R. E., Rosen, R. T., Huang, M. T., Yang, C. S., and Ho, C. T. (2004) Enzymatic synthesis of tea theaflavin derivatives and their antiinflammatory and cytotoxic activities. Bioorg. Med. Chem. 12, 459-467. (27) Schmidt, M., Schmitz, H. J., Baumgart, A., Guedon, D., Netsch, M. I., Kreuter, M. H., Schmidlin, C. B., and Schrenk, D. (2005) Toxicity of green tea extracts and their constituents in rat hepatocytes in primary culture. Food Chem. Toxicol. 43, 307-314. (28) Knapen, M. F., Zusterzeel, P. L., Peters, W. H., and Steegers, E. A. (1999) Glutathione and glutathione-related enzymes in reproduction. A review. Eur. J. Obstet. Gynecol. Reprod. Biol. 82, 171184. (29) Awad, H. M., Boersma, M. G., Boeren, S., van Bladeren, P. J., Vervoort, J., and Rietjens, I. M. (2001) Structure-activity study on the quinone/quinone methide chemistry of flavonoids. Chem. Res. Toxicol. 14, 398-408. (30) Galati, G., Moridani, M. Y., Chan, T. S., and O’Brien, P. J. (2001) Peroxidative metabolism of apigenin and naringenin versus luteolin and quercetin: glutathione oxidation and conjugation. Free Radical Biol. Med. 30, 370-382. (31) Awad, H. M., Boersma, M. G., Boeren, S., van der Woude, H., van Zanden, J., van Bladeren, P. J., Vervoort, J., and Rietjens, I. M. (2002) Identification of o-quinone/quinone methide metabolites of quercetin in a cellular in vitro system. FEBS Lett. 520, 30-34. (32) Schipper, H. M., Kotake, Y., and Janzen, E. G. (1991) Catechol oxidation by peroxidase-positive astrocytes in primary culture: an electron spin resonance study. J. Neurosci. 11, 2170-2176. (33) Elovaara, E., Mikkola, J., Luukkanen, L., Antonio, L., FournelGigleux, S., Burchell, B., Magdalou, J., and Taskinen, J. (2004) Assessment of catechol induction and glucuronidation in rat liver microsomes. Drug Metab. Dispos. 32, 1426-1433. (34) Bolton, J. L., Yu, L., and Thatcher, G. R. (2004) Quinoids formed from estrogens and antiestrogens. Methods Enzymol. 378, 110123. (35) Devanesan, P., Todorovic, R., Zhao, J., Gross, M. L., Rogan, E. G., and Cavalieri, E. L. (2001) Catechol estrogen conjugates and DNA adducts in the kidney of male Syrian golden hamsters treated with 4-hydroxyestradiol: potential biomarkers for estrogeninitiated cancer. Carcinogenesis 22, 489-497. (36) Nakagomi, M., and Suzuki, E. (2000) Quantitation of catechol estrogens and their N-acetylcysteine conjugates in urine of rats and hamsters. Chem. Res. Toxicol. 13, 1208-1213. (37) Boersma, M. G., Vervoort, J., Szymusiak, H., Lemanska, K., Tyrakowska, B., Cenas, N., Segura-Aguilar, J., and Rietjens, I. M. (2000) Regioselectivity and reversibility of the glutathione conjugation of quercetin quinone methide. Chem. Res. Toxicol. 13, 185-191.

TX050151L