Synthesis and Structure Elucidation of Estrogen Quinones Conjugated

Chem. Res. Toxicol. 1998, 11, 909-916

909

Synthesis and Structure Elucidation of Estrogen Quinones Conjugated with Cysteine, N-Acetylcysteine, and Glutathione Kai Cao,† Douglas E. Stack,† Ragulan Ramanathan,‡ Michael L. Gross,‡ Eleanor G. Rogan,† and Ercole L. Cavalieri*,† Eppley Institute for Research in Cancer, University of Nebraska Medical Center, Omaha, Nebraska 68198-6805, and Department of Chemistry, Washington University, One Brookings Drive, St. Louis, Missouri 63130-4899 Received December 22, 1997

Catechol estrogen quinones (CE-Q) have been implicated as ultimate carcinogenic metabolites in estrogen-induced carcinogenesis. CE-Q may covalently bind to DNA to initiate cancer. These quinones can also be conjugated with glutathione, a reaction that prevents damage to DNA by CE-Q. The glutathione conjugates are then catabolized through mercapturic acid biosynthesis to cysteine and N-acetylcysteine conjugates. This may be the most important detoxification pathway of CE-Q. The chemical synthesis and characterization of these conjugates are the first essential steps to better understand their function in biological systems. Eighteen conjugates were synthesized by reaction of estrone-3,4-quinone (E1-3,4-Q), estradiol-3,4-quinone (E2-3,4-Q), estrone-2,3-quinone (E1-2,3-Q), or estradiol-2,3-quinone (E2-2,3-Q) with various sulfur nucleophiles, RSH, in which R is the cysteine, N-acetylcysteine, or glutathione moiety. Reactions of E1-3,4-Q and E2-3,4-Q produce regiospecifically 4-OHE1-2-SR and 4-OHE2-2-SR, respectively, in almost quantitative yield. E1-2,3-Q and E2-2,3-Q react regioselectively and quantitatively to form 2-OHE1(E2)-1-SR and 2-OHE1(E 2)-4-SR, in which the 1-isomers are always the major products. The ratio between 1 and 4 isomers is 3.5 for cysteine, 2.7 for N-acetylcysteine, and 2.5 for glutathione. The synthesized conjugates will be used as standards in the identification of these compounds formed in biological systems.

Introduction The estrogens 17β-estradiol (E2)1 and estrone (E1) are metabolized by two major pathways: 2- and 4-hydroxylation to form catechol estrogens (CE) and 16R-hydroxylation (1, 2). The two CE are mainly inactivated by O-methylation catalyzed by catechol-O-methyltransferases (1). Other conjugations of CE occur via glucuronidation and sulfation. If conversion to these conjugates is incomplete, however, CE can be oxidized to semiquinones and quinones (CE-Q). Inactivation at the quinone level takes place via conjugation with glutathione (GSH) catalyzed by S-transferases or reduction to CE by quinone reductases. The CE-3,4-Q have been hypothesized to be the electrophilic species that react with DNA to generate critical mutations leading to tumor initiation (3, 4). Conjugation with GSH can serve as a protective mechanism against CE-Q-initiated carcinogenesis. * To whom correspondence should be addressed. † University of Nebraska Medical Center. ‡ Washington University. 1 Abbreviations: CAD, collisionally activated decomposition; CE, catechol estrogen(s); CE-Q, catechol estrogen quinone(s); COSY, homonuclear two-dimensional chemical shift correlation spectroscopy; Cys, cysteine; E1, estrone; E2, 17β-estradiol; E1-3,4-Q, estrone-3,4-quinone; FAB, fast atom bombardment; Gly/TFA, glycerol (1% trifluoroacetic acid); GSH, glutathione reduced form; MS, mass spectrometry; (N-Ac)Cys, N-acetylcysteine; NOE, nuclear Overhauser effect; NOESY, nuclear Overhauser effect spectroscopy; 2-OHE1, 2-hydroxyestrone; 2-OHE2 , 2-hydroxyestradiol; 4-OHE1, 4-hydroxyestrone; 4-OHE2, 4-hydroxyestradiol; TFA, trifluoroacetic acid.

GSH (γ-glutamyl-L-cysteinylglycine) is found in almost all living cells of bacteria, plants, and animal tissues (5). It is relatively abundant in liver cells, and the amounts present in various tissues depend on the growth, nutritional state, and hormonal balance of an organism. A large number of electrophilic compounds conjugate with GSH nonenzymatically or, more effectively, via S-transferase-catalyzed reactions. Once GSH conjugates are formed, catabolism occurs via mercapturic acid biosynthesis. First, the glutamyl moiety of the GSH conjugate is removed by transpeptidation catalyzed by γ-glutamyl transpeptidase. Then, the cysteinylglycine is hydrolyzed to yield the cysteine conjugate. The final step consists of acetylation to the N-acetylcysteine conjugate and excretion in the urine. The levels of these various CE-Q-GSH conjugates in tissues provide relevant information on the extent of oxidation of CE to their quinones. GSH conjugates are excreted in urine mostly as N-acetylcysteine conjugates but also as cysteine conjugates. Identification and quantitation of CE-Q conjugates in urine can provide insight into the level of CE-Q formed. Therefore, GSH conjugates can serve as biomarkers in various tissues, whereas N-acetylcysteine and cysteine conjugates can be used as urinary biomarkers for CE-Q formation. To identify these conjugates in biological systems, it is necessary to have reference compounds. They are synthesized by reaction of GSH, N-acetylcysteine, or cysteine with CE-Q. A number of these adducts have already been synthesized (6-11), but some were not

S0893-228x(97)00229-4 CCC: $15.00 © 1998 American Chemical Society Published on Web 07/21/1998

910 Chem. Res. Toxicol., Vol. 11, No. 8, 1998

completely characterized. In the present study, 18 conjugates were synthesized by reaction of the CE-Q of E1 and E2 with cysteine, N-acetylcysteine, or GSH, and their structures were unequivocally determined by NMR and tandem mass spectrometry (MS/MS).

Experimental Section Caution: CE-Q are toxic and were handled according to NIH guidelines (12). Chemicals. 4-Hydroxyestrone (4-OHE1), 4-hydroxyestradiol (4-OHE2), 2-OHE1, 2-OHE2, and the corresponding quinones were prepared as previously described (13-15). GSH, N-acetylL-cysteine, and L-cysteine were purchased from Aldrich Chemical Co. (Milwaukee, WI). HPLC. HPLC was performed on a Waters (Medford, MA) 600E system equipped with a Waters 990 photodiode array detector interfaced with an APC-IV Powermate computer. Analytical separations were conducted by using a YMC (Morris Plains, NJ) ODS-AQ 5-µm, 120-Å column (6.0 × 250 mm) at a flow rate of 1 mL/min. Cysteine conjugates were eluted with 30% methanol in water for 5 min and then with a linear gradient to 100% methanol in 55 min. GSH conjugates were eluted with 20% acetonitrile [0.4% trifluoroacetic acid (TFA)] in water (0.4% TFA) for 5 min, followed by a linear gradient to 100% acetonitrile (0.4% TFA) in 95 min. N-Acetylcysteine conjugates were eluted with 25% acetonitrile (0.4% TFA) in water (0.4% TFA) for 5 min and then a linear gradient to 100% acetonitrile (0.4% TFA) in 95 min. Preparative HPLC was performed by using a YMC ODS-AQ 5-µm, 120-Å column (20 × 250 mm). Cysteine conjugates were eluted with 40% methanol in water for 5 min, followed by a 55min linear gradient to 100% methanol at a flow rate of 6 mL/ min. For GSH and N-acetylcysteine conjugates, two gradients were used: the first separation started with 30% methanol, 5% ethanol, and 65% water at a flow rate of 6 mL/min for 30 min, followed by a 15-min linear gradient to 30% methanol, 30% ethanol, and 40% water at a flow rate of 9 mL/min and then a 10-min linear gradient to 50% methanol, 50% ethanol at a flow rate of 9 mL/min. The second gradient started with 30% methanol (0.4% acetic acid) in water (0.4% acetic acid) for 5 min, followed by a 55-min linear gradient to 100% methanol (0.4% acetic acid) at a flow rate of 6 mL/min. NMR. Proton and homonuclear two-dimensional chemical shift correlation spectroscopy (COSY) NMR spectra were recorded on either a Varian Unity 500 at 499.835 MHz or a Varian XL-300 at 299.938 MHz in Me2SO-d6 (with a trace of TFA-d) at 25 °C. Chemical shifts are reported relative to Me2SO-d6 (2.49 ppm), and the coupling constants (J) are given in hertz. Nuclear Overhauser effect (NOE) spectra and nuclear Overhauser effect spectroscopy (NOESY) spectra were recorded in Me2SO-d6 at 25 °C. Mass Spectrometry. All FAB mass spectra were collected by using a prototype VG ZAB-T four-sector mass spectrometer of BEBE configuration (16). Product ions were detected by using the point detector because the sample quantity was adequate. Full-scan spectra were obtained by scanning MS1 at a resolving power of 1000 over the mass range of 100-1000 at a rate of 10 s/decade (a 1-s delay between scans). Samples were dissolved in methanol (20 µL), and a 1-µL aliquot was added on the probe tip to a matrix of glycerol (1% TFA) (Gly/TFA). A primary beam of 22 keV cesium ions was used to desorb the analyte ions. The accelerating voltage of the mass spectrometer was 8 kV. Tandem mass spectra of the FABproduced precursor ions were obtained by activating them in a floated (4 kV) collision cell with helium at sufficient pressure to suppress the beam by 50%. Ten to fifteen 15-s scans were signal averaged to improve the signal-to-noise ratio. Data were acquired and processed with a DEC Alpha 3000 workstation equipped with OPUS V 3.2X software and interfaced with the mass spectrometer via a VG SIOS I unit.

Cao et al. The elemental composition of each adduct was confirmed by exact mass measurements of FAB-produced [M + H]+ ions conducted at a resolving power of 10 000 with a Kratos MS-50 three-sector mass spectrometer equipped with an argon FAB gun (17). The error of the exact mass measurement of each adduct was less than 2 ppm, as determined by peak matching using two reference ions from a glycerol/CsI mixture as mass standards. Synthesis of CE Conjugates. 4-OH-E1-2-cysteine. The syntheses of CE conjugates are summarized in Scheme 1. To a stirred solution of cysteine (32 mg, 0.26 mmol) in 3 mL of acetic acid/water (1:1 v/v) was added estrone-3,4-quinone (E1-3,4-Q) (50 mg, 0.17 mmol) in 5 mL of acetonitrile. After 30 min, the reaction mixture was filtered. The product was then analyzed on a reversed-phase column and separated on a preparative reversed-phase column to give 4-OHE1-2-cysteine (59 mg, 85% yield). UV: λmax ) 292 nm. NMR (500 MHz): 6.90 (s, 1 H, 1-H), 3.85 (dd, J ) 4.5, 7.5 Hz, 1 H, R-H-cys), 3.28 (dd, J ) 4.4, 14.3 Hz, 1 H, β-H-cys), 3.12 (dd, J ) 8.5, 14.5 Hz, 1 H, β-H-cys), 2.80 (dd, J ) 5.5, 17.5 Hz, 1 H, 6-H), 2.60-1.20 (14 H, remaining protons), 0.81 (s, 3 H, 13-CH3). Mass: [M + H]+, C21H28NO5S, calcd 406.1688, obsd 406.1684. 4-OHE2-2-cysteine. In a similar manner, 4-OHE2-2-cysteine was obtained in 83% yield. UV: λmax ) 292 nm. NMR (300 MHz): 6.89 (s, 1 H, 1-H), 3.85 (dd, J ) 4.5, 8.0 Hz, 1 H, R-Hcys), 3.49 (t, J ) 8.3 Hz, 1 H, 17R-H), 3.28 (dd, J ) 4.5, 14.3 Hz, 1 H, β-H-cys), 3.11 (dd, J ) 8.0, 14.0 Hz, 1 H, β-H-cys), 2.75 (dd, J ) 4.9, 17.8 Hz, 1 H, 6-H), 2.55-0.95 (14 H, remaining protons), 0.64 (s, 3 H, 13-CH3). Mass: [M + H]+, C21H30NO5S, calcd 408.1845, obsd 408.1836. 2-OHE1-1- and -4-cysteine. E1-2,3-Q was reacted with cysteine following the procedures used for E1-3,4-Q to afford a mixture of two isomers in 86% yield. Neither analytical nor preparative HPLC was able to separate these two isomers. The ratio between the 1- and 4-isomers was determined to be 3.5:1 from the NMR spectrum by comparing the integration of the aromatic, R-H-cys, and 13-CH3 signals. UV: λmax ) 303 nm. NMR (500 MHz): 2-OHE1-1-cysteine, 6.50 (s, 1 H, 4-H), 3.84 (dd, J ) 6.0, 6.0 Hz, 1 H, R-H-cys), 3.17 (m, 2 H, β-H-cys), 3.101.00 (15 H, remaining protons) 0.85 (s, 3 H, 13-CH3); 2-OHE14-cysteine, 6.80 (s, 1 H, 1-H), 3.97 (dd, J ) 5.5, 5.5 Hz, 1 H, R-H-cys), 3.17 (m, 2 H, β-H-cys), 3.10-1.00 (15 H, remaining protons), 0.80 (s, 3 H, 13-CH3). Mass: [M + H]+, C21H28NO5S, calcd 406.1688, obsd 406.1686. 2-OHE2-1- and -4-cysteine. These adducts were synthesized as those of the E1 adducts above and were obtained in 85% yield as a mixture of two isomers, 2-OHE2-1-cysteine and 2-OHE2-4-cysteine. The ratio between the 1- and 4-isomers was determined by NMR to be 3.5:1. UV: λmax ) 303 nm. NMR (500 MHz): 2-OHE2-1-cysteine, 6.48 (s, 1 H, 4-H), 3.86 (dd, J ) 5.0, 6.5 Hz, 1 H, R-H-cys), 3.55 (t, J ) 8.5 Hz, 1 H, 17R-H), 3.16 (m, 2 H, β-H-cys), 2.70-0.95 (15 H, remaining protons), 0.69 (s, 3 H, 13-CH3); 2-OHE2-4-cysteine, 6.79 (s, 1 H, 1-H), 3.97 (dd, J ) 4.0, 6.5 Hz, 1 H, R-H-cys), 3.50 (t, J ) 8.5 Hz, 1 H, 17R-H), 3.16 (m, 2 H, β-H-cys), 2.97 (dd, J ) 4.0, 6.5 Hz, 1 H, 6-H), 2.700.95 (14 H, remaining protons), 0.64 (s, 3 H, 13-CH3). Mass: [M + H]+, C21H30NO5S, calcd 408.1845, obsd 408.1844. 4-OHE1-2-N-acetylcysteine: To a stirred solution of Nacetylcysteine (58 mg, 0.35 mmol) in 3 mL of acetic acid/water (1:1 v/v) was added E1-2,3-Q (50 mg, 0.17 mmol) in 5 mL of acetonitrile. After 30 min, the reaction mixture was filtered, and the product was separated on preparative HPLC to give 62.3 mg of 4-OHE1-2-N-acetylcysteine in 80% yield. UV: λmax ) 292 nm. NMR (300 MHz): 6.77 (s, 1 H, 1-H), 4.27 (dd, J ) 4.4, 8.8 Hz, 1 H, R-H-cys), 3.14 (dd, J ) 4.4, 13.4 Hz, 1 H, β-Hcys), 2.93 (dd, J ) 8.9, 13.6 Hz, 1 H, β-H-cys), 2.78 (dd, J ) 5.3, 17.4 Hz, 1 H, 6-H), 2.60-0.90 (14 H, remaining protons), 1.82 (s, 3 H, COCH3), 0.75 (s, 3 H, 13-CH3). Mass: [M + H]+, C23H30NO6S, calcd 448.1794, obsd 448.1799. 4-OHE2-2-N-acetylcysteine. 4-OHE2-2-N-acetylcysteine was prepared in 79% yield by reacting E2-3,4-Q and N-acetylcysteine,

Synthesis of Estrogen Quinone Conjugates

Chem. Res. Toxicol., Vol. 11, No. 8, 1998 911

Scheme 1. Reaction of CE-Q with Cysteine, N-Acetylcysteine, and GSH

as described above. UV: λmax ) 292 nm. NMR (500 MHz): 6.74 (s, 1 H, 1-H), 4.24 (dd, J ) 4.8, 9.3 Hz, 1 H, R-H-cys), 3.50 (t, J ) 8.5 Hz, 1 H, 17R-H), 3.13 (dd, J ) 5.2, 14.3 Hz, 1 H, β-H-cys), 2.91 (dd, J ) 9.0, 14.0 Hz, 1 H, β-H-cys), 2.73 (dd, J ) 5.8, 17.8 Hz, 1 H, 6-H), 2.60-1.00 (14 H, remaining protons), 1.83 (s, 3 H, COCH3), 0.64 (s, 3 H, 13-CH3). Mass: [M + H]+, C23H32NO6S, calcd 450.1950, obsd 450.1949. 2-OHE1-1- and -4-N-acetylcysteine. E1-2,3-Q was reacted with N-acetylcysteine, affording a mixture of 2-OHE1-1- and -4N-acetylcysteine in 74% yield. The two isomers were separated by HPLC. 2-OHE1-1-N-acetylcysteine. UV: λmax ) 304 nm. NMR (500 MHz): 6.48 (s, 1 H, 4-H), 3.69 (dd, J ) 3.5, 10.5 Hz, 1 H, R-H-cys), 3.23 (dd, J ) 4.0, 13.5 Hz, 1 H, β-H-cys), 2.65 (m, 2 H, β-H-cys and 6-H), 2.60-1.00 (14 H, remaining protons), 1.77 (s, 3 H, COCH3), 0.84 (s, 3 H, 13-CH3). Mass: [M + H]+, C23H30NO6S, calcd 448.1794, obsd 448.1791. 2-OHE1-4-N-acetylcysteine. UV: λmax ) 302 nm. NMR (500 MHz): 6.75 (s, 1 H, 1-H), 4.15 (dd, J ) 5.0, 8.5 Hz, 1 H, R-H-cys), 3.15 (dd, J ) 5.0, 13.5 Hz, 1 H, β-H-cys), 2.95 (m, 2 H, β-H-cys and 6-H), 2.65-1.20 (14 H, remaining protons), 1.81 (s, 3 H, COCH3), 0.79 (s, 3 H, 13-CH3). Mass: [M + H]+, C23H30NO6S, calcd 448.1794, obsd 448.1798. 2-OHE2-1- and -4-N-acetylcysteine. E2-2,3-Q reacted with N-acetylcysteine as above, affording a mixture of the two separable isomers in 76% yield. 2-OHE2-1-N-acetylcysteine. UV: λmax ) 304 nm. NMR (300 MHz): 6.45 (s, 1 H, 4-H), 3.63 (dd, J ) 3.2, 10.9 Hz, 1 H, R-H-cys), 3.49 (t, J ) 8.0 Hz, 1 H, 17R-H), 3.20 (dd, J ) 3.2, 13.7 Hz, 1 H, β-H-cys), 2.60 (m, 2 H, β-H-cys and 6-H), 2.600.90 (14 H, remaining protons), 1.82 (s, 3 H, COCH3), 0.67 (s, 3 H, 13-CH3). Mass: [M + H]+, C23H32NO6S, calcd 450.1950, obsd 450.1951. 2-OHE2-4-N-acetylcysteine. UV: λmax ) 302 nm. NMR (300 MHz): 6.75 (s, 1 H, 1-H), 4.12 (m, 1 H, R-H-cys), 3.50 (t, J

) 8.3 Hz, 1 H, 17R-H), 3.14 (dd, J ) 5.0, 13.2 Hz, 1 H, β-H-cys), 2.92 (m, 2 H, β-H-cys and 6-H), 2.80-0.90 (14 H, remaining protons), 1.81 (s, 3 H, COCH3), 0.65 (s, 3 H, 13-CH3). Mass: [M + H]+, C23H32NO6S, calcd 450.1950, obsd 450.1953. 4-OHE1-2-SG. To a stirred solution of GSH (108 mg, 0.352 mmol) in 3 mL of acetic acid/water (1:1 v/v) was added E1-3,4-Q (50 mg, 0.175 mmol) in 5 mL of acetonitrile dropwise at room temperature. After 30 min, the reaction mixture was filtered, and the product was analyzed by HPLC and separated on preparative HPLC to give 4-OHE1-2-SG (85.2 mg, 83% yield). UV: λmax ) 292 nm. NMR (500 MHz): 6.76 (s, 1 H, 1-H), 4.36 (dd, J ) 5.0, 8.0 Hz, 1 H, R-H-cys), 3.99 (m, 1 H, R-H-glu), 3.70 (s, 2 H, CH2-gly), 3.15 (m, 1 H, β-H-cys), 2.86 (m, 1 H, β-H-cys), 2.83 (m, 1 H, 6-H), 2.60-1.20 (18 H, remaining protons), 0.81 (s, 3 H, 13-CH3). Mass: [M + H]+, C28H38N3O9S, calcd 592.2329, obsd 592.2328. 4-OHE2-2-SG. 4-OHE2-2-SG was prepared by reacting E23,4-Q and GSH as described above and was obtained in 84% yield. UV: λmax ) 292 nm. NMR (300 MHz): 6.77 (s, 1 H, 1-H), 4.37 (dd, J ) 4.5, 9.3 Hz, 1 H, R-H-cys), 3.99 (m, 1 H, R-H-glu), 3.70 (s, 2 H, CH2-gly), 3.49 (t, J ) 8.5 Hz, 1 H, 17R-H), 3.13 (m, 1 H, β-H-cys), 2.88 (m, 1 H, β-H-cys), 2.75 (m, 1 H, 6-H), 2.601.00 (18 H, remaining protons), 0.62 (s, 3 H, 13-CH3). Mass: [M + H]+, C28H40N3O9S, calcd 594.2485, obsd 594.2486. 2-OHE1-1- and -4-SG. In a manner similar to the above two compounds, the mixture of 2-OHE1-1- and -4-SG was obtained in 84% yield. The two isomers were subsequently separated by HPLC using an acidic gradient. 2-OHE1-1-SG. UV: λmax ) 304 nm. NMR (500 MHz): 6.46 (s, 1 H, 4-H), 4.02 (dd, J ) 4.0, 10.5 Hz, 1 H, R-H-cys), 3.89 (t, J ) 8.5 Hz, 1 H, R-H-glu), 3.66 (s, 2 H, CH2-gly), 3.08 (dd, J ) 4.0, 13.0 Hz, 1 H, β-H-cys), 2.77 (dd, J ) 10.5, 13.0 Hz, 1 H, β-H-cys), 2.68 (m, 1 H, 6-H), 2.60-1.00 (18 H, remaining protons), 0.84 (s, 3 H, 13-CH3). Mass: [M + H]+, C28H38N3O9S, calcd 592.2329, obsd 592.2324.

912 Chem. Res. Toxicol., Vol. 11, No. 8, 1998 2-OHE1-4-SG. UV: λmax ) 302 nm. NMR (500 MHz): 6.74 (s, 1 H, 1-H), 4.31 (dd, J ) 4.7, 9.7 Hz, 1 H, R-H-cys), 3.95 (m, 1 H, R-H-glu), 3.68 (s, 2 H, CH2-gly), 3.10 (dd, J ) 4.5, 13.0 Hz, 1 H, β-H-cys), 3.00 (dd, J ) 6.0, 16.5 Hz, 1 H, 6-H), 2.85 (dd, J ) 9.7, 13.2 Hz, 1 H, β-H-cys), 2.60-1.00 (18 H, remaining protons), 0.80 (s, 3 H, 13-CH3). Mass: [M + H]+, C28H38N3O9S, calcd 592.2329, obsd 592.2329. 2-OHE2-1- and -4-SG. E2-2,3-Q reacted with GSH in a manner similar to E1-2,3-Q, affording a mixture of the two isomers in 83% yield. They were subsequently separated by HPLC. 2-OHE2-1-SG. UV: λmax ) 304 nm. NMR (500 MHz): 6.44 (s, 1 H, 4-H), 4.01 (dd, J ) 3.8, 11.0 Hz, 1 H, R-H-cys), 3.92 (m, 1 H, R-H-glu), 3.67 (m, 2 H, CH2-gly), 3.53 (t, J ) 8.0 Hz, 1 H, 17R-H), 3.06 (dd, J ) 3.8, 13.2 Hz, 1 H, β-H-cys), 2.76 (dd, J ) 11.0, 13.0 Hz, 1 H, β-H-cys), 2.60 (m, 1 H, 6-H), 2.60-0.90 (18 H, remaining protons), 0.68 (s, 3 H, 13-CH3). Mass: [M + H]+, C28H40N3O9S, calcd 594.2485, obsd 594.2483. 2-OHE2-4-SG. UV: λmax ) 302 nm. NMR (500 MHz): 6.75 (s, 1 H, 1-H), 4.32 (dd, J ) 5.0, 9.5 Hz, 1 H, R-H-cys), 3.97 (m, 1 H, R-H-glu), 3.69 (s, 2 H, CH2-gly), 3.50 (t, J ) 8.5 Hz, 1 H, 17R-H), 3.06 (dd, J ) 5.0, 13.0 Hz, 1 H, β-H-cys), 2.93 (dd, J ) 4.7, 17.3 Hz, 1 H, 6-H), 2.86 (dd, J ) 9.5, 13.0 Hz, 1 H, β-Hcys), 2.60-1.00 (18 H, remaining protons), 0.65 (s, 3 H, 13-CH3). Mass: [M + H]+, C28H40N3O9S, calcd 594.2485, obsd 594.2483.

Results and Discussion Synthesis of CE-Q-Thiol Conjugates. Because CE-Q are unstable, coupling of these compounds with GSH or N-acetylcysteine was obtained by trapping the freshly formed quinones in situ with the various thiols (6, 9, 10). Abul-Hajj and Cisek (8) oxidized 4-OHE1 with activated manganese dioxide in chloroform and obtained a high yield of the GSH conjugate after addition of GSH in 50% aqueous acetic acid. Following Abul-Hajj’s method of oxidation (18), Dwivedy et al. (13) synthesized in high yield and characterized all of the quinones of the CE of E1 and E2. Oxidation of CE to CE-Q is virtually quantitative in several solvents, which include chloroform, acetone, Me2SO, dimethylformamide, and acetonitrile. Oxidation of CE was carried out at -30 °C with activated manganese dioxide in acetonitrile. After filtration, the product CE-Q was added dropwise to a stirred solution of the thiol in 50% aqueous acetic acid, pH 1.9. Decoloration of the quinone occurred in about 30 s. The reaction was continued for 30 min at room temperature. Under these homogeneous reaction conditions, formation of thiol conjugates was almost quantitative. The conjugates were analyzed and purified by HPLC. Cysteine conjugates were analyzed by HPLC with a methanol/water gradient. 4-OHE1(E2)-2-cysteine was the only product obtained from reaction of E1(E2)-3,4-Q with cysteine. E1(E2)-2,3-Q afforded, instead, a mixture of 2-OHE1(E2)-1- and -4-cysteine, which could not be separated by HPLC, even under conditions in which the mobile phase or the pH of the mobile phase was changed. Quantitative determination of the two isomers was achieved by NMR, based on integration of three pairs of signals: aromatic protons, R-H-cysteine protons, and 13methyl protons. HPLC analysis of the GSH and N-acetylcysteine conjugates with a methanol/water or an acetonitrile/water gradient resulted in broad peaks and poor separation. After addition of 0.4% acetic acid or TFA to the mobile phase, sharp peaks were obtained, and the 1- and 4-isomers were separated.

Cao et al.

E1(E2)-3,4-Q produced regiospecifically 4-OHE1(E2)-2SR, whereas E1(E2)-2,3-Q afforded 2-OHE1(E2)-1-SR and 2-OHE1(E2)-4-SR, in which the 1-isomers were the major products. The ratio between the 1- and 4-isomers was 3.5 for cysteine, 2.7 for N-acetylcysteine, and 2.5 for GSH. To examine the effect of pH on the conjugation of GSH with the four CE-Q, reactions were conducted at pH 3.5, 5, and 7.4 by using phosphate buffers and at pH 6.7 (unbuffered water). The yields of GSH conjugates were nearly the same at all pH, with the exception of a slight decrease at higher pH, owing to the presumed decomposition of CE-Q. The ratio between the 1- and 4-isomers from reaction of GSH with E1(E2)-2,3-Q remained the same over the range of pH, indicating that pH does not affect the relative formation of the regioisomers. The synthesis of these compounds sheds light on the mechanism of their formation, as well as providing standards for analysis of biologically formed conjugates. Structural Analysis of the Conjugates. NMR Spectroscopy. Determination of the structures of the CE-thiol conjugates presents some difficulties, and the assignments reported in the literature have not always been correct. Abul-Hajj and Cisek (7, 8) proved that E13,4-Q reacts with GSH regiospecifically at C-2 by using C-1-tritiated E1-3,4-Q and demonstrating that the conjugate does not lose tritium. More recently, Iverson et al. reported the addition of GSH at C-2 of E2-3,4-Q and addition of GSH at C-1 and C-4 of E2-2,3-Q after horseradish peroxidase-catalyzed oxidation of the corresponding CE (11). Butterworth et al. (9) mistakenly assigned the conjugate of GSH with E2-3,4-Q as being substituted at C-1 instead of C-2. Furthermore, the two regioisomers at C-1 and C-4 obtained from reaction of E2-2,3-Q with GSH had assignments reversed from those obtained by Iverson et al. (11) and Abul-Hajj et al. (7). In typical NMR spectra of the thiol conjugates (Figure 1), the most characteristic chemical shifts for structure assignment are the ones corresponding to the aromatic protons (H-1: 6.73-6.90 ppm, H-4: 6.44-6.50 ppm). The 13-methyl signals (0.64-0.69 ppm for estradiol conjugates, 0.75-0.84 ppm for estrone conjugates) are indicative of the purity of the samples. All chemical shifts and coupling constants are summarized in Table 1. The homonuclear COSY technique was used to assign the proton signals. The most characteristic coupling is that between R-H-cys and β-H-cys. All three protons generally display characteristic doublet-doublet signals. For the synthesized CE-thiol conjugates, integration showed only one aromatic proton (H-1 or H-2 for 4-hydroxy estrogens, H-1 or H-4 for 2-hydroxy estrogens), indicating a Michael addition of the thiol to the aromatic ring. Yet, the regiochemistry of the reaction cannot be completely determined by one-dimensional NMR and COSY spectra. The best way to assign unequivocally the structures of CE-Q conjugated with various biological nucleophiles derives from studies of the nuclear Overhauser effect of various neighboring protons. This has allowed us to determine the structures of the two rotational isomers at C-1 of 4-OHE1 (E2) bound to the N-7 of guanine (3). The same technique has been used by Tabakovic et al. to establish the structure of 4-methylimidazole bound to C-1 of 2-OHE1 (19). The protons involved in NOE experiments are the pairs H-1, H-11 and H-4, H-6. As shown in Figure 2, irradiation of the H-1 resonance enhances the H-11 resonance and vice versa. Spatial closeness between H-4 and H-6

Synthesis of Estrogen Quinone Conjugates

Chem. Res. Toxicol., Vol. 11, No. 8, 1998 913

Figure 1. Typical NMR spectra of CE-thiol conjugates: (A) 4-OHE1-2-SG and (B) 2-OHE1-1- and -4-cysteine. Spectra were recorded in Me2SO-d6 with 0.4% TFA-d.

allows us to perform similiar NOE experiments. For the CE-thiol conjugates, a positive NOE between the aromatic proton and H-11 will establish the aromatic proton as H-1, thereby indicating that the thiol addition occurs at C-2 for E-3,4-Q or at C-4 for E-2,3-Q. Similarly, an NOE effect between the aromatic proton H-4 and the two H-6 indicates thiol addition at C-1 for E-2,3-Q (Figure 2). Before the structures of the thiol conjugates were elucidated, NOE experiments were carried out with CE. The NMR spectrum of 4-OHE1(E2) in Me2SO-d6 showed a singlet at 6.55 ppm (H-1 and H-2 merge as one peak, with integration corresponding to two protons). Irradiation at the resonance of 6.55 ppm (H-1) resulted in a positive NOE of the signal at 2.3 ppm (H-11). 2-OHE1(E2) in Me2SO-d6 showed two singlet signals at 6.4 and 6.7

ppm. Irradiation at the resonance of 6.7 ppm caused a significant enhancement of signal at 2.3 ppm (H-11), and irradiation at the resonance of 6.4 ppm showed an NOE at 2.68 ppm (H-6). Thus, the signal at 6.7 ppm was assigned to H-1 and the signal at 6.4 ppm to H-4. In summary, NOE experiments allow us to assign the aromatic protons of CE and indicate at the same time the positions of thiol addition to the CE-Q in the conjugates. To elucidate, for example, the regiochemistry of the GSH conjugates, NOE and NOESY experiments were carried out. Irradiation of 4-OHE1-2-SG at 6.76 ppm significantly enhanced the signal at 2.2 ppm (H-11), enabling the assignment of H-1 to the chemical shift at 6.76 ppm. The NOESY spectrum of 4-OHE2-2-SG also showed an enhancement of H-11 (2.2 ppm) after irradia-

914 Chem. Res. Toxicol., Vol. 11, No. 8, 1998

Cao et al.

Table 1. 300 and 500 MHz NMR Dataa of the Thiol Conjugates in Me2SO-d6 (Trace TFA-d) H-1 4-OHE1-2-Cys

6.90 (s)

4-OHE2-2-Cys

6.89 (s)

2-OHE1-1-Cys 2-OHE1-4-Cys

6.50 (s) 6.80 (s)

2-OHE2-1-Cys

6.48 (s)

2-OHE2-4-Cys

6.79 (s)

4-OHE1-2-(N-Ac)Cys

6.77 (s)

4-OHE2-2-(N-Ac)Cys

6.74 (s)

2-OHE1-1-(N-Ac)Cys 2-OHE1-4-(N-Ac)Cys

6.48 (s) 6.75 (s)

2-OHE2-1-(N-Ac)Cys

6.45 (s)

2-OHE2-4-(N-Ac)Cys

6.75 (s)

4-OHE1-2-SG

6.76 (s)

4-OHE2-2-SG

6.77 (s)

2-OHE1-1-SG 2-OHE1-4-SG

6.46 (s) 6.74 (s)

2-OHE2-1-SG 2-OHE2-4-SG a

H-4

6.44 (s) 6.75 (s)

R-H-cys 3.85 (dd) J ) 4.5, 7.5 3.85 (dd) J ) 4.5, 8.0 3.84 (dd) J ) 6.0, 6.0 3.97 (dd) J ) 5.5, 5.5 3.86 (dd) J ) 5.0, 6.5 3.97 (dd) J ) 4.0, 6.5 4.27 (dd) J ) 4.4, 8.8 4.24 (dd) J ) 4.8, 9.3 3.69 (dd) J ) 3.5, 10.5 4.15 (dd) J ) 5.0, 8.5 3.63 (dd) J ) 3.2, 10.9 4.12 (m) 4.36 (dd) J ) 5.0, 8.0 4.37 (dd) J ) 4.5, 9.3 4.02 (dd) J ) 4.0, 10.5 4.31 (dd) J ) 4.7, 9.7 4.01 (dd) J ) 3.8, 11.0 4.32 (dd) J ) 5.0, 9.5

17-R-H

3.49 (t) J ) 8.3

3.55 (t) J ) 8.5 3.50 (t) J ) 8.5 3.50 (t) J ) 8.5

3.49 (t) J ) 8.0 3.50 (t) J ) 8.3 3.49 (t) J ) 8.5

3.53 (t) J ) 8.0 3.50 (t) J ) 8.5

β-H-cys

β-H-cys

H-6

13-CH3

3.28 (dd) J ) 4.4, 14.3 3.28 (dd) J ) 4.5, 14.3 3.17 (m)

3.12 (dd) J ) 8.5, 14.5 3.11 (dd) J ) 8.0, 14.0 3.17 (m)

2.80 (dd) J ) 5.5, 17.5 2.75 (dd) J ) 4.9, 17.8 2.70 (m)

0.81 (s)

3.17 (m)

3.17 (m)

2.94 (m)

0.80 (s)

3.16 (m)

3.16 (m)

0.69 (s)

3.16 (m)

3.16 (m)

3.14 (dd) J ) 4.4, 13.4 3.13 (dd) J ) 5.2, 14.3 3.23 (dd) J ) 4.0, 13.5 3.15 (dd) J ) 5.0, 13.5 3.20 (dd) J ) 3.2, 13.7 3.14 (dd) J ) 5.0, 13.2 3.15 (m)

2.93 (dd) J ) 8.9, 13.6 2.91 (dd) J ) 9.0, 14.0 2.65 (m)

2.65 (dd) J ) 5.0, 14.0 2.97 (dd) J ) 4.0, 16.5 2.78 (dd) J ) 5.3, 17.4 2.73 (dd) J ) 5.8, 17.8 2.65 (m)

0.75 (s)

2.95 (m)

2.95 (m)

0.79 (s)

2.60 (m)

2.60 (m)

0.67 (s)

2.92 (m)

2.92 (m)

0.65 (s)

2.86 (m)

2.83 (m)

0.81 (s)

3.13 (m)

2.88 (m)

2.75 (m)

0.62 (s)

3.08 (dd) J ) 4.0, 13.0 3.10 (dd) J ) 4.5, 13.0 3.06 (dd) J ) 3.8, 13.2 3.06 (dd) J ) 5.0, 13.0

2.77 (dd) J ) 10.5, 13.0 2.85 (dd) J ) 9.7, 13.2 2.76 (dd) J ) 11.0, 13.0 2.86 (dd) J ) 9.5, 13.0

2.68 (m)

0.84 (s)

3.00 (dd) J ) 6.0, 16.5 2.60 (m)

0.80 (s)

2.93 (dd) J ) 4.7, 17.3

0.65 (s)

0.64 (s) 0.85 (s)

0.64 (s)

0.64 (s) 0.84 (s)

0.68 (s)

J values are reported in Hz and chemical shifts in ppm.

Mass Spectrometry. We used mass spectrometry in two ways to confirm the structures of the conjugates. Exact mass measurements at 10 000 resolving power confirmed molecular formulas, and these measurements are within 2 ppm of the theoretical values. Second, tandem four-sector MS substantiated the structural conclusions drawn from NMR. For the GSH conjugates, a systematic nomenclature that was previously used to describe product ions (20) was also applied to the estrogen quinone conjugates of cysteine and N-acetylcysteine, although not all reactions pertain to these simpler systems.

Figure 2. Possible NOEs for CE and CE-thiol conjugates.

tion of the signal at 6.77 ppm; thus, the proton was assigned as H-1. These results clearly indicate that addition of thiols to E1(E2)-3,4-Q occurs at C-2. The NOESY spectrum of 2-OHE1-4-SG also showed an enhancement of the H-11 signal (2.25 ppm) after irradiation of H-1 (6.74 ppm), indicating that the SR group is bonded to the C-4 position. For 2-OHE1-1-SG, enhancement of the H-6 signals (2.68 ppm, 2.48 ppm) was observed after irradiation of H-4 (6.46 ppm), suggesting that the thiol is substituted at C-1. In this way, the results from these NMR techniques unequivocally prove the regiochemistry of CE-thiol conjugates.

4-OHE1-2-SG, 2-OHE1-4-SG, and 2-OHE1-1-SG and Corresponding E2 Analogues. A series of ions from peptide-backbone fragmentation (called “type 1”) are produced, confirming that the Cys moiety of GSH is modified, as is seen in the product ion spectra of the subject molecules (Figure 3). Type-1 fragments that are modified by attachment of the intact steroid are at m/z 517 and 463 (modified b2 and y2, respectively, and designated as *b2 and *y2), and those without the steroid moiety are observed inter alia at m/z 130, 173, and 179 (b1, d2, and y2). Internal

fragment ions of the structure and +CHdCH-S-Rx (where Rx is the estrone moiety) are observed at m/z 360 and 343, respectively, and were previously classified as “type 2” (20). Highly abundant fragment ions of m/z 317 (“type 3”) form by cleavage of the bond between the sulfur and peptide chain with the charge abiding on the steroid. The +NH dCHCH S-Rx 2 2

Synthesis of Estrogen Quinone Conjugates

Chem. Res. Toxicol., Vol. 11, No. 8, 1998 915

Figure 3. Tandem mass spectra of FAB-produced [M + H]+ ions from (A) 4-OHE1-2-SG, (B) 2-OHE1-4-SG, and (C) 2-OHE1-1-SG, obtained by using a tandem four-sector mass spectrometer.

Figure 4. Tandem mass spectra of FAB-produced [M + H]+ ions from (A) 4-OHE2-2-N-acetylcysteine and (B) 2-OHE2-1-Nacetylcysteine, obtained by using a tandem four-sector mass spectrometer.

916 Chem. Res. Toxicol., Vol. 11, No. 8, 1998

fragment ions at m/z 299 and 284 form by subsequent losses of H2O and the HS radical. The fragment ions of m/z 315 and 297 (“type 4”) are important for isomer distinction (20). They are not detected when the GSH is at position 2 of the steroid (i.e., for 4-OHE1-2-SG), whereas the abundance is moderate when the attachment is at position 4 (i.e., for 2-OHE14-SG) and most abundant for 2-OHE1-1-SG. The mechanism involves C-S bond cleavage with oxidation of the steroid and reduction of the GSH moiety, as was discussed previously (20). The FAB-produced [M + H]+ ions of the conjugated E2 analogues give nearly identical spectra as the estrone conjugates except the masses of steroid-containing fragments are upshifted by 2 u. 4-OHE1-2-, 2-OHE1-4-, and 2-OHE1-1-N-acetylcysteine and Corresponding E2 Analogues. Collisional activation of the [M + H]+ ions of the E1 analogues gives a modified y1 (*y1), *z1 ion, and internal or “type-2” fragment ions at m/z 360 and 343, which are consistent with the assigned structures. The cleavage of the C-S bond results in both a radical ion, RxSH+ (of m/z 318), and the expected RxS+ (of m/z 317), and both subsequently lose H2O. The loss of an HS radical from the two species gives ions of m/z 285 and 284, respectively. Cleavage of the C-S bond accompanied by loss of the steroid gives protonated N-acetylcysteine (m/z 164) at low abundance and an oxidized form at m/z 162, which is the most abundant fragment. The fragment ion of m/z 130 is probably protonated N-acetyldehydroalanine. Production spectra of the [M + H]+ (Figure 4) and [M + Na]+ ions of the E2 analogues show the expected 2-u shift for fragments containing the steroid. Unlike those of GSH, N-acetylcysteine conjugates do not give significantly abundant “type-4” fragment ions (20), precluding isomer distinction by tandem MS. 4-OHE1-2-, 2-OHE1-1-, and 2-OHE1-4-cysteine and Corresponding E2 Analogues. Collisional activation of the FAB-produced [M + H]+ yields simple spectra that are consistent with general structural features. The expected RxS+ fragments of m/z 317 are dominant. Only three other product ions of m/z 389 [M + H - NH3]+, 284 [RxS - SH]+, and 155 (a steroid fragment) are significant. Tandem mass spectra of the E2 analogues are similar except the fragment ions are shifted higher by 2 u. Once again, “type-4” product ions do not form, and isomer distinction is not possible by MS.

Conclusions Eighteen CE-thiol conjugates were successfully synthesized in high yield by reaction of E1-3,4-Q, E2-3,4-Q, E1-2,3-Q, or E2-2,3-Q with cysteine, N-acetylcysteine, or GSH. Their structures were characterized by HPLC, UV, NMR, and MS. Reaction of E1(E2)-3,4-Q with the sulfur nucleophiles, RSH, yields regiospecifically 4-OHE1 (E2)2-SR, whereas E1(E2)-2,3-Q affords 2-OHE1 (E2)-1-SR, and 2-OHE1 (E2)-4-SR, in which the 1-isomer is always the major product. The ratio between 1 and 4 isomers is 3.5 for cysteine, 2.7 for N-acetylcysteine, and 2.5 for glutathione. These compounds will serve as standards in biological studies that use the analytical HPLC methods developed.

Acknowledgment. This research was supported by a U.S. PHS grant from the National Cancer Institute

Cao et al.

(P01 CA49210). Core support at the Eppley Institute was provided by NCI Laboratory Cancer Research Center Support (Core) Grant CA36727 and at Washington University by Grant No. P41RR00954. K.C. received fellowship support from the University of Nebraska Center for Environmental Toxicology.

References (1) Ball, P., and Knuppen, R. (1980) Catechol oestrogens (2- and 4-hydroxyestrogens): Chemistry, biogenesis, metabolism, occurrence and physiological significance. Acta Endocrinol. (Copenhagen) 93 (Suppl. 232), 1-127. (2) Martucci, C. P., and Fishman, J. (1993) P450 enzymes of estrogen metabolism. Pharmac. Ther. 57, 237-257. (3) Stack, D. E., Byun, J., Gross, M. L., Rogan, E. G., and Cavalieri, E. L. (1996) Molecular characteristics of catechol estrogen quinones in reactions with deoxyribonucleosides. Chem. Res. Toxicol. 9, 851-859. (4) Cavalieri, E. L., Stack, D. E., Devanesan, P. D., Todorovic, R., Dwivedy, I., Higginbotham, S., Johansson, S. L., Patil, K. D., Gross, M. L., Gooden, J. K., Ramanathan, R., Cerny, R. L., and Rogan, E. G. (1997) Molecular origin of cancer: Catechol estrogen3,4-quinones as endogenous tumor initiators. Proc. Natl. Acad. Sci. U.S.A. 94, 10937-10942. (5) Boyland, E., and Chasseaud, L. F. (1969) The role of glutathione and glutathione S-transferases in mercapturic acid biosynthesis. Adv. Enzymol. 32, 173-219. (6) Jellinck, P. H., and Elce, J. S. (1969) Synthesis of estrogen glutathione and cysteine derivatives. Steroids 13, 711-718. (7) Abul-Hajj, Y. J., and Cisek, P. L. (1986) Regioselective reaction of thiols with catechol estrogens and estrogen-O quinones. J. Steroid. Biochem. 25, 245-247. (8) Abul-Hajj, Y. J., and Cisek, P. (1988) Catechol estrogen adducts. J. Steroid Biochem. 31, 107-110. (9) Butterworth, M., Lau, S. S., and Monks, T. J. (1996) 17-β-Estradiol metabolism by hamster hepatic microsomes: Comparison of catechol estrogen O-methylation with catechol estrogen oxidation and glutathione conjugation. Chem. Res. Toxicol. 9, 793-799. (10) Suzuki, E., Iwasaki, R., Goto, J., Matsuki, Y., and Nambara, T. (1996) Synthesis of N-acetylcysteine conjugates of catechol estrogens. Steroids 61, 296-301. (11) Iverson, S. L., Shen, L., Anlar, N., and Bolton, J. L. (1996) Bioactivation of estrone and its catechol metabolites to quinoidglutathione conjugates in rat liver microsomes. Chem. Res. Toxicol. 9, 492-499. (12) NIH Guidelines for the Laboratory Use of Chemical Carcinogens (1981) NIH Publication No.81-2385, U.S. Government Printing Office, Washington D. C. (13) Dwivedy, I., Devanesan, P., Cremonesi, P., Rogan, E. and Cavalieri, E. (1992) Synthesis and characterization of estrogen 2,3and 3,4-quinones. Comparison of DNA adducts formed by the quinones versus horseradish peroxidase-activated catechol estrogens. Chem. Res. Toxicol. 5, 828-833. (14) Chen, S.-H., Luo, G.-R., Wu, X.-S., Chen, M., and Zhao, H.-M. (1986) A new synthetic route to 2- and 4-methylestradiols by nucleophilic substitution. Steroids 47, 63-66. (15) Xie, R.-G., Deng, L.-S., Gu, H.-Q., Fan, Y.-M., and Zhao, H.-M. (1981) A new route to 2-hydroxyl steroidal estrogens via acetylation and Dakin oxidation. Steroids 40, 389-392. (16) Gross, M. L. (1994) Tandem mass spectrometric strategies for determining structure of biologically interesting molecules. Acc. Chem. Res. 27, 361-369. (17) Gross, M. L., Chess, E. K., Lyon, P. A., Crow, F. W., Evans, S., and Tudge, H. (1982) Triple analyzer mass spectrometry for highresolution MS/MS studies. Int. J. Mass Spectrom. Ion Phys. 42, 243-254. (18) Abul-Hajj, Y. J. (1984) Synthesis of 3,4-estrogen-o-quinone. J. Steroid Biochem. 21, 621-622. (19) Tabakovic, K., Gleason, W. B., Ojala, W. H., and Abul-Hajj, Y. J. (1996) Oxidative transformation of 2-hydroxyestrone. Stability and reactivity of 2,3-estrone quinone and its relationship to estrogen carcinogenicity. Chem. Res. Toxicol. 9, 860-865. (20) Ramanathan, R., Cao, K., Cavalieri, E., and Gross, M. L. (1998) Mass spectrometric methods for distinguishing structural isomers of glutathione conjugates of estrone and estradiol. J. Am. Soc. Mass Spectrom. 9, 612-619.

TX9702291