Comparative Study on the Bioactivation Mechanisms and Cytotoxicity

have synthesized Te-phenyl-L-tellurocysteine and evaluated its bioactivation and ... reflect the bond strength of the corresponding C-S, C-Se, and C-T...
0 downloads 0 Views 195KB Size
Download PDF
1610

Chem. Res. Toxicol. 2002, 15, 1610-1618

Comparative Study on the Bioactivation Mechanisms and Cytotoxicity of Te-Phenyl-L-tellurocysteine, Se-Phenyl-L-selenocysteine, and S-Phenyl-L-cysteine Martijn Rooseboom, Nico P. E. Vermeulen,* Fatma Durgut, and Jan N. M. Commandeur Leiden/Amsterdam Center for Drug Research (LACDR), Division of Molecular Toxicology, Department of Pharmacochemistry, Vrije Universiteit Amsterdam, De Boelelaan 1083, 1081 HV Amsterdam, The Netherlands Received May 6, 2002

Tellurium compounds are effective antioxidants and chemoprotectors, even more active than their selenium and sulfur analogues. In addition to these properties, some selenium compounds, such as selenocysteine Se-conjugates, possess significant chemopreventive and antitumor activities, and selenol metabolites are considered as active species. In the present study, we have synthesized Te-phenyl-L-tellurocysteine and evaluated its bioactivation and cytotoxicity. The activities of this compound were compared with those of the corresponding selenium and sulfur analogues. Te-Phenyl-L-tellurocysteine is bioactivated into its corresponding tellurol, as detected by GC-MS, by cysteine conjugate β-lyase and amino acid oxidase, analogously to what has been shown previously for Se-phenyl-L-selenocysteine. The rate of β-elimination may reflect the bond strength of the corresponding C-S, C-Se, and C-Te bond. Bioactivation of Te-phenyl-L-tellurocysteine and its selenium and sulfur analogues by oxidative enzymes was evaluated by measuring NADPH-dependent activation of hepatic mGST and inhibition of EROD. Te-Phenyl-L-tellurocysteine and Se-phenyl-L-selenocysteine displayed strong and timedependent mGST activation, while S-phenyl-L-cysteine resulted in no significant activation. Te-Phenyl-L-tellurocysteine was also a strong inhibitor of EROD activity. In addition to EROD inhibition, Te-phenyl-L-tellurocysteine was the strongest inhibitor of several human cytochrome P450 isoenzymes followed by Se-phenyl-L-selenocysteine, while S-phenyl-L-cysteine was the weakest inhibitor. Interestingly, Te-phenyl-L-tellurocysteine selectively inhibited cytochrome P450 1A1 directly, which is, for example, responsible for the activation of several procarcinogens. Preliminary cytotoxicity studies with Te-phenyl-L-tellurocysteine in freshly isolated rat hepatocytes showed a time-dependent depletion of GSH and LDH leakage comparable with the relatively nontoxic drug paracetamol, while the selenium and sulfur analogues were nontoxic toward rat hepatocytes. In conclusion, because the chemopreventive and antitumor activities of selenium compounds are thought to be mediated via their selenol metabolites and tellurium compounds might be even more active than selenium compounds, tellurocysteine Te-conjugates might be an interesting novel class of prodrugs for the formation of biologically active tellurols.

Introduction Previously, it has been observed that several tellurium (Te) compounds are potent antioxidants by scavenging hydrogen peroxide, hydroxyl radicals, and peroxynitrite (1, 2). Furthermore, micromolar levels of tellurides were shown to inhibit both thioredoxin reductase and the growth of human cancer cells (3). Tellurium compounds also were shown to be much more effective than their corresponding selenium and sulfur analogues (two other elements in Group VI of the periodic table). Thus, the glutathione peroxidase mimetic activity of bis(4-aminophenyl)telluride was 20-fold higher than that of the corresponding selenide and three times more effective than ebselen, a well-known glutathione peroxidase mimetic agent (4). Bis(4-aminophenyl)telluride was also * To whom correspondence should be addressed. Telephone: +31 20 4447590. Fax: +31 20 4447610. E-mail: [email protected].

much more effective in the reduction of butylated hydroxytoluene-induced cytotoxicity in human lung fibroblastic cells (IC50 < 2 µM) than its selenium and sulfur analogue (IC50 . 50 µM) (5). Diphenyltelluride was also more effective in this system than the corresponding selenium and sulfur compounds (5). Furthermore, bis(4aminophenyl)telluride reacted much faster with the stable radical 2,2-diphenyl-1-picrylhydrazyl (DPPH), a frequently used compound in antioxidant studies, than the corresponding selenide and sulfide (6). While the importance of sulfur and selenium in biology is clear, the biological role of tellurium is unclear and much research in this area remains to be performed (7). Selenium compounds and to a lesser extent sulfur compounds have been shown to be interesting compounds in chemopreventive and antitumor therapies. During recent years, it has been shown that several selenium compounds can function as antioxidants, chemoprotec-

10.1021/tx020034f CCC: $22.00 © 2002 American Chemical Society Published on Web 12/16/2002

Comparison between Te, Se, and S

tors, apoptosis inducers, and effective chemopreventors in a variety of organ systems, including brain, mammary gland, liver, skin, colon, lung, and prostate (8-10). In comparison to sulfur compounds the corresponding selenium analogues were 500-750-fold more effective in cancer prevention (11, 12). On the basis of the observation that several tellurium compounds are more effective antioxidants and chemoprotectors than their corresponding selenium analogues (1-6), tellurium compounds might be an interesting novel class of potential chemopreventive and antitumor agents. A particularly interesting group of selenium compounds exerting the above-mentioned pharmacological effects are the selenocysteine Se-conjugates (SeCysconjugates)1 (13). Some of these compounds are important constituents of selenium-enriched garlic and are prodrugs that can be activated by β-lyase enzymes, resulting in the formation of pharmacologically active selenols, ammonia and pyruvate (13-15). It has been reported that the activation rates of these compounds, catalyzed by rat renal cysteine conjugate β-lyase/glutamine transaminase K (β-lyase/GTK), were much higher for SeCys-conjugates than for their sulfur analogues, i.e between 30- and 200fold (16). Another enzyme that was shown to activate SeCys-conjugates was amino acid oxidase (AAO). This enzyme catalyzed β-elimination of SeCys-conjugates, while the corresponding cysteine S-conjugates were not β-eliminated (17). Both SeCys-conjugates and cysteine S-conjugates were oxidatively deaminated by AAO, albeit, the selenium analogues at a much higher rate than the corresponding sulfur analogues (17). Triggered by the observed differences between bioactivation SeCys-conjugates and the corresponding cysteine S-conjugates, we were interested in the bioactivation of the corresponding tellurium compounds. Therefore, in the present study, we synthesized and evaluated the bioactivation of Te-phenyl-L-tellurocysteine by cysteine conjugate β-lyase and amino acid oxidase and compared it with the bioactivation of the corresponding sulfur and selenium analogues. Previously, we showed that SeCysconjugates can be activated to thiol reactive selenoxides, a reaction mainly catalyzed by flavin-containing monooxygenases (FMOs) (18). To investigate the protein alkylation of reactive tellurium intermediates, the activation of microsomal glutathione S-transferase (mGST), the inhibition of cytochrome P450 1A1 and the consumption of thiocholine were determined. These effects were previously used as model systems to study the protein alkylation of metabolites of thiourea-containing compounds (19). The ability of the compounds to inhibit human cytochrome P450s was also studied. Finally the cytotoxicity, lipid peroxidation (LPO), and GSH depletion of these agents was investigated in freshly isolated rat hepatocytes.

Experimental Procedures Materials. Diphenyl ditelluride was obtained from Aldrich (Beerse, Belgium). L-Amino acid oxidase (L-AAO) type I from Crotalus adamanteus (0.46 units/mg), p-hydroxyphenylacetic 1 Abbreviations: β-lyase/GTK, cysteine conjugate β-lyase/glutamine transaminase K; EROD, 7-ethoxyresorufine O-dealkylation; FMO, flavin-containing monooxygenases; AAO, amino acid oxidase; LDH, lactate dehydrogenase; LPO, lipid peroxidation; mGST, microsomal glutathione S-transferase; PTU, N-phenylthiourea; SeCys-conjugate, selenocysteine Se-conjugate.

Chem. Res. Toxicol., Vol. 15, No. 12, 2002 1611 acid (HPA), β-chloro-L-alanine, R-keto-γ-methiolbutyrate (ΚΜΒ), horseradish peroxidase (HRP) type I from horseradish (116 units/mg), 7-ethoxyresorufin, 7-methoxy-4-(trifluoromethyl)coumarin (MTFC), GSSG reductase from bakers yeast (62 500 units/mL), and 5,5′-dithiobis(2-nitrobenzoic acid) (DTNB) were purchased from Sigma Chemical Co. (St Louis. MO). o-Phenylenediamine was obtained from Janssen Chimica (Geel, Belgium). Catalase from beef liver (25 000 units/mg) and glucose-6-phosphate dehydrogenase from yeast (grade II) were obtained from Boehringer (Mannheim, Germany). 3-Cyano-7ethoxycoumarin (CEC) was from Molecular Probes (Eugene, OR). N-Phenylthiourea (PTU) was purchased from Merck (Munich, Germany). Hydrogen peroxide was from J. T. Baker (Deventer, The Netherlands). 5,5′-Dithiobis(2-nitrobezoic acid) was obtained from Fluka. 7-Benzyloxy-4-(trifluoromethyl)coumarin (BTFC) and human P450s (Supersomes) were obtained from Gentest (Woburn, MA). Rat renal cytosolic β-lyase/GTK was purified 1036-fold from rat kidney by Yamauchi et al. (20). Thiocholine was prepared as described by Guo and Ziegler (21). S-Phenyl-L-cysteine and Se-phenyl-L-selenocysteine were prepared as described previously (13). N-Methyl-7-methoxy-4(aminomethyl)coumarin (MMAMC) was synthesized by Venhorst et al. (22). All other chemicals were of the highest grade commercially available. Synthesis of Te-Phenyl-L-tellurocysteine. Te-Phenyl-Ltellurocysteine was synthesized by a procedure comparable to that described previously by Esaki and Soda (23). Diphenyl ditelluride (1.5 mmol, 614 mg) was dissolved in 20 mL of ethanol and 8 mL of 0.5 N NaOH. At 0 °C, 15 mmol (570 mg) of NaBH4 was added. The mixture was stirred for 30 min at 50 °C during which the color changed from orange to colorless. The reaction mixture was subsequently cooled to 0 °C and β-chloroalanine (4 mmol, 512 mg) dissolved in 4 mL of 0.5 N NaOH was added. After stirring the reaction mixture overnight, the mixture was acidified until pH 2 using concentrated HCl and extracted with 20 mL of diethyl ether. The aqeous layer was brought to pH 6-7 with NaOH after which the product precipitated. The product was recrystallized from hot water giving light-yellow crystals. The yield was 28%, and the purity, determined by 1H NMR, was >98%. The 1H NMR was similar to that obtained for Se-phenyl-L-selenocysteine (13). 1H NMR (D2O, Na2CO3): δ (ppm) 3.2 (2H, dd, CH2-CH), 3.5 (1H, t, CH2-CH), 7.4 (3H, m, m,p-Ar-H), 7.8 (2H, d, o-Ar-H). Elemental analysis was consistent with that expected for Te-phenyl-L-tellurocysteine and demonstrated that the synthesized compound was obtained in pure form. Elemental analysis, calcd (C9H11NO2Te): C, 37.07; H, 3.77; N, 4.69. Found: C, 37.12; H, 3.76; N, 4.71. HPLC analysis revealed a single peak belonging to Te-phenyl-Ltellurocysteine also analogous to previously obtained results with SeCys-conjugates (14). Isolation of Rat Liver Microsomes and Rat Hepatocytes. Male Wistar rats (200-220 g) were obtained from Harlan (Zeist, The Netherlands) and were acclimatized for 1 week after delivery to the laboratory. The animals were fed a standard laboratory diet from Hope Farms (Woerden, The Netherlands) and had access to food and water ad libitum. Rats were pretreated intraperitoneally with β-naphthoflavone (60 mg/kg in arachidis oil) for two consecutive days. Rats were sacrificed by decapitation and rat liver microsomes (washed twice) were isolated as described previously (18). For the isolation of rat hepatocytes, rats pretreated with β-naphthoflavone as described above were fasted overnight. The liver parenchymal cells were isolated by a two-step collagenase perfusion method essentially according to the procedure described by Seglen (24) as modified by Nagelkerke et al. (25). The viability of the hepatocytes was above 95% as determined by trypan blue exclusion. Bioactivation of Te-Phenyl-L-tellurocysteine by Purified Rat Renal β-Lyase/GTK and L-AAO. (1) β-Lyase/GTK. The time course and kinetic parameters (kcat and Km) of β-elimination of Te-phenyl-L-tellurocysteine by β-lyase/GTK were determined by measuring pyruvate formation as described for S-phenyl-L-cysteine and Se-phenyl-L-selenocysteine by Com-

1612

Chem. Res. Toxicol., Vol. 15, No. 12, 2002

mandeur et al. (16). Briefly, Te-phenyl-L-tellurocysteine was incubated with β-lyase/GTK (1 µg/mL) and R-keto-γ-methiolbutyric acid (KMB) (0.2 mM) in 50 mM Tris-HCl buffer (pH 8.6) at 37 °C. The amount of pyruvate was detected by HPLC equipped with a fluorescence detector (excitation wavelength of 336 nm and emission wavelength of 420 nm) after derivatization with o-phenylenediamine (OPD) (16). To obtain enzyme kinetic parameters a 20 min incubation time was used and experiments were performed in triplicate. Nonenzymatic degradation was not observed. (2) L-AAO. The time course and kinetic parameters (kcat and Km) of oxidative deamination and β-elimination of Te-phenylL-tellurocysteine by L-AAO were determined by measuring hydrogen peroxide and pyruvate formation, respectively, as described for S-phenyl-L-cysteine and Se-phenyl-L-selenocysteine by Rooseboom et al. (17). Briefly, Te-phenyl-L-tellurocysteine was incubated with L-ΑΑΟ from Crotalus adamanteus (5 µg/ mL), p-hydroxyphenylacetic acid (HPA) (0.94 mM), and horseradish peroxidase (HRP) (28.5 µg/mL) in 100 mM potassium phosphate buffer (pH 7.2) at 37 °C. The amount of pyruvate was detected by HPLC equipped with a fluorescence detector detector (excitation wavelength of 336 nm and emission wavelength of 420 nm) after derivatization with o-phenylenediamine (OPD) (16). Hydrogen peroxide was detected by measuring fluorescence of 2,2′-dihydroxybiphenyl-5,5′-diacetate detector (excitation wavelength of 323 nm and emission wavelength of 400 nm) (17). To obtain enzyme kinetic parameters a 20 min incubation time was used. Nonenzymatic degradation was not observed. Identification of Phenyltellurol. Incubations of Te-phenylL-tellurocysteine with either β-lyase/GTK, L-AAO, or in the absence of enzyme were performed as described above on a larger scale (1 mL) and for 4 h. The reaction was stopped by the addition of 0.5 mL of 2 N HCl and extracted with ethyl acetate. The organic layer was subsequently treated with Etheral diazomethane and analyzed by GC-MS. GC-MS analyses were performed on a HP 5890 gas chromatograph equipped with a 25 m BPX5 column (0.25 mm i.d., 0.25 µm film thickness, SGE, Amstelveen, The Netherlands) coupled to a Hewlett-Packard MSD 5970 mass spectrometer (ion source, electron impact ionization, electron energy of 70 eV). Temperatures of the injection port and transfer line were 270 °C. The column temperature was programmed from 40 (5 min) to 270 °C (20°C/min) and kept at 270 °C for 10 min. Full scanning analyses were performed in the range of m/z 30-650. Using these GC-MS conditions, the following retention time and mass spectrum was recorded for methylphenyltelluride: Retention time, 13.7 min; mass spectrum m/z (relative intensity, tellurium isotope, assignment) 222 (100, 130Te, M•+), 207 (62, 130Te, M•+- CH ), 140 (38), 77 (97, C H +), 51 (56). 3 6 5 Methylphenyltelluride was prepared as a reference compound by reduction of diphenyl ditelluride with NaBH4 and subsequent methylation by Etheral diazomethane as has been described previously for diphenyl diselenide (17). Activation of mGST. Conjugates (100 µM) were incubated in the presence of 0.5 mM NADPH, 1 mM EDTA, 5 mM MgCl2, and 1 mg/mL rat liver microsomes in 50 mM potassium phosphate buffer (pH 7.4) at 37 °C, as described by Onderwater et al. (19). At specific time points, a 100 µL aliquot was taken and added to a cuvette containing 25 µL of 40 mM 1-chloro-2,4dinitrobenzene (CDNB) in ethanol, 100 µL of 50 mM GSH and 775 µL of 125 mM potassium phosphate buffer (pH 6.5) and the rate of the increase in absorbance at 340 nm (GSH-CDNB conjugate) was determined as described previously by Habig and Jacoby (26). Experiments were also carried out in the absence of NADPH or in the absence of test compounds and all experiments were performed in triplicate. Inhibition of 7-Ethoxyresorufin O-Dealkylation (EROD). Compounds (100 µM) were incubated in the presence of 0.5 mM NADPH, 1 mM EDTA, 5 mM MgCl2, and 1 mg/mL rat liver microsomes in 50 mM potassium phosphate buffer (pH 7.4) at 37 °C, as described by Onderwater et al. (19). At specific time

Rooseboom et al. points, a 50 µL aliquot was added to 450 µL of (100 mM potassium phosphate buffer (pH 7.8). Subsequently, 15 µL of this mixture was added to a cuvette containing 33 µL of 0.25 mM 7-ethoxyresorufin in ethanol, 33 µL of 5 mM NADPH and 919 µL of 100 mM potassium phosphate buffer (pH 7.8) and the rate of the increase in fluorescence at an excitation wavelength of 530 nm and an emmision wavelength of 586 nm (resorufin) was determined as described previously by Onderwater et al. (19). Experiments were also carried out in the absence of NADPH or in the absence of test compounds and all experiments were performed in triplicate. Oxidation of Thiocholine. The extent of oxidation of thiocholine was determined according to a method of Guo and Ziegler (21). Briefly, 1 mg/mL rat liver microsomes was preincubated for 5 min at 37 °C in 1 mM EDTA, 5 mM glucose-6phosphate, 0.5 mM NADPH, glucose-6-phosphate dehydrogenase (1 units/mL), catalase (2 units/mL), 500 µM thiocholine, and 100 mM potassium phosphate buffer (pH 7.5). The test compounds (final concentration 100 µM) were added to the incubation and aliquots were taken at fixed time points. The amount of remaining thiocholine was determined spectrophotometrically with 5,5′-dithiobis(2-nitrobenzoic acid) (DTNB) at 405 nm. All experiments were carried out in triplicate and in the absence or presence of test compounds. Inhibition of Human Cytochrome P450 Isoenzymes. All experiments were performed in Costar 3595 96-well plates using a 100 mM potassium phosphate buffer (pH 7.4) and a total volume of 200 µL. The final concentrations of the different fluorescent probes, serving as substrates for the P450s studied, were approximately equal to their Km value. All analyses were caried out on a Victor 1420 multilabel counter (Wallac Oy., Turku, Finland) equipped with a plate heater set at 37 °C. In the case of P450 1A1, 0.5 pmol of P450 1A1 and 2 µM 3-cyano-7-ethoxycoumarin (CEC) in the presence or absence of test compound. The reaction was subsequently started by the addition of a preincubated NADPH-regenerating system, resulting in final concentrations of 1.3 mM NADPH, 3.3 mM glucose6-phosphate, and 0.4 units/mL glucose-6-phosphate dehydrogenase. The increase in fluorescence was followed in time for 20 min with intervals of 5 min. The excitation wavelength was 405 nm, and the emission wavelength was 460 nm. The same procedure was used for P450 1A2; however, a final concentration of 5 µM 3-cyano-7-ethoxycoumarin (CEC) and 1 pmol/well enzyme was used. 3-Cyano-7-ethoxycoumarin (CEC) was also used to detect P450 2C19 activity, but with a final concentration of 25 µM. The incubation and analyses were performed as described above for P450 1A2, except for the incubation time (45 min) and the amount of enzyme per well (2 pmol). In the case of P450, 2D6 N-methyl-7-methoxy-4-(aminomethyl)coumarin (MMAMC) was used as the fluorescent probe at a final concentration of 5 µM as described by Venhorst et al. (22). The experiments were performed as described for P450 2C19, using 2 pmol of P450 2D6/well. To abolish interference from background fluorescence, a final concentration of 0.1 mM NADPH and 0.3 mM glucose-6-phosphate was used. In the case of P450 2C9, an end-point measurement was used due to the low fluorescent yield of the product formed at physiological pH. Incubations were performed as described above in the presence of 1 pmol/well P450 2C9 and 75 µM 7-methoxy-4(trifluoromethyl)coumarin (MTFC). After 45 min the reaction was stopped by the addition of 10 µL 5 M trichloroacetic acid. Subsequently, 100 µL of 1.5 M Tris-HCl (pH 9.0) was added, and the fluorescence was determined (excitation wavelength of 405 nm and emission wavelength of 535 nm). To measure P450 2E1 inhibition, the same method was used, although in this case, 100 µM MTFC and 4 pmol/well P450 2E1 was used. The same method carried out for P450 2C9 was used to measure P450 3A4 activity; however, in this case, a final concentration of 50 µM 7-benzyloxy-4-(trifluoromethyl)coumarin (BTFC), 2 pmol/well, and an incubation time of 30 min were used.

Comparison between Te, Se, and S

Chem. Res. Toxicol., Vol. 15, No. 12, 2002 1613

Figure 1. Chemical structures of Te-phenyl-L-tellurocysteine, Se-phenyl-L-selenocysteine and S-phenyl-L-cysteine. All experiments were performed in the presence and absence of the test compounds and in triplicate. Incubations with Rat Hepatocytes. Freshly isolated rat hepatocytes (106 cells/mL), obtained as described above, were preincubated in Hanks-HEPES buffer (pH 7.6) containing 1.5% BSA in a total volume of 6 mL. Cells were preincubated in a rotary shaker (140 rpm) at 37 °C and under a 95% oxygen and 5% carbon dioxide atmosphere for 1 h. Subsequently, the test compounds (dissolved in incubation buffer) or incubation buffer was added. In this study, paracetamol was used as a positive control. At specific time points, aliquots were taken: 0.5 mL for measuring lipid peroxidation (LPO) and 0.5 mL for GSH and LDH (lactate dehydrogenase) determinations. Samples for GSH and LDH determinations were centrifuged for 2 min at 150 rpm. The supernatant was used to measure LDH activity, and the pellet was used to determine GSH levels. All experiments were performed in quadruplicate. (1) LPO. LPO was determined by assessment of thiobarbituric acid-reactive substances as described by Haenen and Bast (27). To 0.4 mL of incubation sample was added 0.7 mL of 2-thiobarbituric acid/trichloroacetic acid solution, consisting of 15% (w/v) trichloroacetic acid and 0.375% (w/v) 2-thiobarbituric acid in 0.25 N HCl. Samples were heated for 15 min at 95 °C and centrifugated for 15 min at 4000 rpm. Subsequently, the absorbance of the supernatant was recorded at 535 and 590 nm. The amount of LPO is expressed as A535nm - A590nm. (2) LDH. LDH activity was determined by measuring the conversion of pyruvate and NADH into lactate and NAD+, respectively, as described by Molde´us et al. (28). A 20 µL sample was added to a cuvette containing 1 mL of 50 mM triethanolamine, 10 mM MgCl2, 5 mM EDTA, 0.2 mM NADH, and 1 mM pyruvate at pH 7.4, and the decrease in absorbance at 340 nm (NADH) was determined for 30 s. The obtained values are expressed as percentages of samples treated with Triton X-100, indicating 100% LDH leakage. (3) GSH. Cellular levels of GSH were determined by the method of Baker et al. based on the reduction of 5,5′-dithiobis(2-nitrobenzoic acid) (DTNB) to 5-thio-2-nitrobenzoate (TNB) by GSH. GSSG reductase is added to increase the sensitivity of this method. To the cell pellets of the samples was added 0.5 mL of 3% HClO4, and samples were subsequently centrifuged for 15 min at 4000 rpm. A total of 50 µL supernatant was added to 100 µL of reaction buffer, consisting of 0.5 mM NADPH, 4.5 mM 5,5′-dithiobis(2-nitrobenzoic acid) (DTNB), and 12.5 units/ mL GSSG reductase in 100 mM potassium phosphate buffer (pH 7.4). The increase in absorbance at 405 nm was measured in time. Linearity of the assay was confirmed using a standard curve. Statistical Analysis. Statistical evaluation of the results was performed using the Student’s t-test. Differences were considered significant if P was less that 0.05 (*).

Results and Discussion Bioactivation of Te-Phenyl-L-tellurocysteine by Rat Renal β-Lyase/GTK and L-AAO. Incubation of Tephenyl-L-tellurocysteine with β-lyase/GTK and L-AAO followed by derivatization of the product with Etheral diazomethane resulted in the formation of an ion at m/z 222 (130Te) with a characteristic tellurium isotope pattern as determined by GC-MS (Figure 2). The mass spectrum and retention time (13.7 min) of synthetic standard methylphenyltelluride was identical to that obtained after incubation of Te-phenyl-L-tellurocysteine with β-lyase/

Figure 2. Typical mass spectrum of incubation of Te-phenylL-tellurocysteine with β-lyase/GTK or L-AAO. Samples were acidified, extracted and derivatized with Etheral diazomethane as described in Experimental Procedures. Mass spectrum is identical with methylphenyltelluride, synthesized as a reference. Fragment patterns are described in the Experimental Procedures.

GTK and L-AAO. Other tellurium-containing metabolites were not observed. Methylphenyltelluride was not observed in incubations performed in the absence of β-lyase/ GTK or L-AAO. Analogy between the metabolism of tellurium and selenium has been observed previously. Exposure of workers to tellurium dust, TeO2, and gaseous H2Te resulted in detection of the unpleasant garlic-like smell resulting from (CH3)2Te in breath, which is analogous with metabolites observed after selenium exposure (29). In addition to the formation of phenyltellurol, upon incubation of Te-phenyl-L-tellurocysteine with β-lyase/ GTK and L-AAO, the formation of pyruvate was also observed (Table 1 and Figure 3), as was previously reported for the selenium analogue (16, 17). Enzyme kinetic studies showed that Te-phenyl-L-tellurocysteine has a lower Km for β-lyase/GTK than Se-phenyl-L-selenocysteine, with values of 0.48 and 0.92 mM, respectively (Table 1). The kcat values were comparable, resulting in a 1.7-fold higher kcat/Km value for Te-phenyl-L-tellurocysteine (4552 min-1 mM-1) than for Se-phenyl-L-selenocysteine (2740 min-1 mM-1) in case of β-lyase/GTK (Table 1). As shown previously, determination enzyme kinetics could not be carried out accurately for S-phenyl-Lcysteine due to a too low turnover. The higher rate of β-elimination of Te-phenyl-L-tellurocysteine compared to its selenium analogue might be explained by the weaker bond strength of the C-Te bond compared to the C-Se bond.2 Similarly, the higher β-elimination rate of Sephenyl-L-selenocysteine compared to its sulfur analogue was previously explained by differences in bond strength between C-Se (234 kJ mol-1) and C-S (272 kJ mol-1) (16, 30). As shown previously, L-AAO catalyzes both the β-elimination and oxidative deamination of Se-phenyl-Lselenocysteine (17). The fact that the β-elimination pathway (measured as pyruvate formation) was not affected by catalase excluded the possibility that this pathway results from selenoxidation by the hydrogen peroxide formation of the oxidative transaminase path2 Although the bond energy of the C-Te bond has not been analyzed, the bond is weaker than the C-Se bond (234 kJ mol-1) due to the larger atomic radius of Te (137 pm) than that of Se (117 pm) and because of the decreasing bond strength in Group VI of the periodic table from C-O (358 kJ mol-1) > C-S (272 kJ mol-1) > C-Se (234 kJ mol-1).

1614

Chem. Res. Toxicol., Vol. 15, No. 12, 2002

Rooseboom et al.

Table 1. Enzyme Kinetic Parameters for S-Phenyl-L-cysteine, Se-Phenyl-L-selenocysteine, and Te-Phenyl-L-tellurocysteinea specific activity (nmol min-1 mg-1) S-phenyl-L-cysteine Se-phenyl-L-selenocysteine Te-phenyl-L-tellurocysteine

53 ( 30 9130 ( 1300 11103 ( 1932

Km (mM)

β-lyase/GTK ND 0.92 ( 0.20 0.48 ( 0.12

kcat (min-1)

kcat/Km (min-1 mM-1)

ND 2430 ( 360 2041 ( 129

ND 2740 ( 490 4252 ( 269

L-AAO

S-phenyl-L-cysteine oxidative deamination β-elimination Se-phenyl-L-selenocysteine oxidative deamination β-elimination Te-phenyl-L-tellurocysteine oxidative deamination β-elimination

63.5 ( 3 ND

0.20 ( 0.04 ND

6 ( 0.5 ND

30 ( 7 ND

200 ( 6 466 ( 14

0.04 ( 0.01 0.11 ( 0.02

13 ( 0.3 57 ( 4.1

307 ( 43 532 ( 84

6.8 ( 1 477 ( 24

ND 1.51 ( 0.19

ND 240 ( 20

ND 159 ( 7

a Experiments were performed as described in Experimental Procedures. Results are presented as mean value ( SD for three preparations. Specific activity determined at a substrate concentration of 0.5 mM. Data for S-phenyl-L-cysteine and Se-phenyl-Lselenocysteine were obtained from ref 16 in case of β-lyase/GTK and (17) in case of L-AAO. ND, not detectable due to a too low turnover of the enzyme.

Figure 3. Hypothetical scheme for bioactivation of Te-phenylL-tellurocysteine. Reaction A, hypothetical oxidation of Tephenyl-L-tellurocysteine into its corresponding telluroxide by NADPH-dependent oxidative enzymes analogous to that observed for SeCys-conjugates (18). Reaction B, bioactivation into phenyltellurol by β-lyase/GTK and L-AAO. Reaction C, Oxidative deamination into the corresponding R-keto-acid by L-AAO, which has been shown to be a minor pathway (Table 1).

way. These two pathways are therefore believed to occur via a common enzyme-substrate intermediate, analogous to the biotransformation of β-chloroalanine (31). It was shown that at low oxygen levels β-chloroalanine was exclusively β-eliminated by L-AAO resulting in the formation of pyruvate, ammonia, and chloride, while at high oxygen levels only oxidative deamination, leading to chloropyruvate and ammonia, was observed (31). At intermediate oxygen levels, mixtures of β-elimination and oxidative deamination products from β-chloroalanine were observed. The sum of β-elimination and oxidative deamination products was independent of the oxygen level indicating a common intermediate (31). The results of the present study show that Te-phenyl-L-tellurocysteine is also β-eliminated by L-AAO, with a Km value of 1.5 mM and a kcat value of 240 min-1 (Table 1 and Figure 3). In addition to β-elimination, L-AAO also catalyzes the oxidative deamination of various amino acids. Analogous to this observation, S-phenyl-L-cysteine and Se-phenylL-selenocysteine were both shown to be metabolized by

oxidative deamination resulting in the corresponding R-keto acids with a 10-fold higher kcat/Km value for Sephenyl-L-selenocysteine (17). The present study indicates that the specific activity for oxidative deamination of Tephenyl-L-tellurocysteine by L-AAO was much lower (6.8 nmol min-1 mg-1) than that obtained for S-phenyl-Lcysteine and Se-phenyl-L-selenocysteine (Table 1). Due to a slow turnover, the enzyme kinetic parameters (kcat and Km) for the oxidative deamination pathway could not be determined for Te-phenyl-L-tellurocysteine. S-PhenylL-cysteine is not β-eliminated at all by L-AAO, possibly due to the relatively high bond strength of the C-S bond (272 kJ mol-1) (30). Interestingly, Te-phenyl-L-tellurocysteine is almost exclusively β-eliminated by L-AAO (Table 1 and Figure 3). Se-Phenyl-L-selenocysteine is both β-eliminated and oxidatively deaminated at a similar rate by L-AAO. Interestingly, L-AAO can therefore apparently function as a complete oxidative deaminating enzyme or a complete β-eliminating enzyme dependent on the substrate used. Microsomal Bioactivation of Conjugates to Protein and Thiol-Reactive Metabolites. Previously, we have shown that FMO-dependent oxidation of SeCysconjugates results in the formation of thiol-reactive selenoxides (18). To investigate protein alkylation by reactive intermediates formed from the conjugates (i.e., electrophilic or oxidative), the activation of microsomal glutathione S-transferase (mGST), the inhibition of P450 1A1 and the consumption of thiocholine were used as model systems. These systems have previously been used to study the thiol and protein reactivity of reactive intermediates from thiourea-containing compounds, generated by FMO (19). To investigate the alkylating potential of reactive intermediates of Te-phenyl-L-tellurocysteine and its sulfur and selenium analogues, the activation of mGST and the inhibition of EROD activity were studied in the present study. mGST is a membranebound protein consisting of three individually catalytic active subunits (32). Alkylation of the sulfhydryl group of Cys-49 of these subunits is known to result in an increased catalytic activity (33). As shown in Figure 4A, 100 µM Te-phenyl-L-tellurocysteine resulted in a timedependent activation of mGST, indicative for alkylation of the cysteine residue of mGST. The extent of activation reached a maximal value of 1.5-fold after 10 min and then

Comparison between Te, Se, and S

Figure 4. Activation of mGST, inhibition of EROD activity and oxidation of thiocholine by Te-phenyl-L-tellurocysteine (Te), Sephenyl-L-selenocysteine (Se), and S-phenyl-L-cysteine (S). Substrates (0.1 mM) were incubated as described in Experimental Procedures. (A) mGST activity; (B)EROD activity, and (C) thiocholine levels. PTU (0.1 mM) was used as a positive control (19). Results are presented as mean value (SD for three preparations. Asterisks indicate statistically significant difference (P < 0.05) from control incubations, as determined by Student’s t-test.

Chem. Res. Toxicol., Vol. 15, No. 12, 2002 1615

decreased. The activation of mGST was dependent on the presence of NADPH as was shown for PTU, which was used in this study as a positive control (19). S-PhenylL-cysteine did not activate mGST. Se-Phenyl-L-selenocysteine also showed a strong activation of mGST which was not significantly different from that of Te-phenyl-Ltellurocysteine (Figure 4A). Since NADPH was essential for the activation of mGST, the tested compounds are not capable of alkylating mGST directly but require bioactivation to alkylating metabolites. In addition to the formation of sulfoxides from cysteine S-conjugates, selenoxides were previously shown to be generated form SeCys-conjugates catalyzed by FMOs (18). These selenoxides were reactive with thiols, such as GSH and thiocholine. Therefore, activation of Te-phenyl-L-tellurocysteine to the corresponding telluroxides by oxidative enzymes is likely responsible for the mGST activation; however, this remains to be established (Figure 3). Inhibition of EROD activity was used previously as a second tool to investigate the protein alkylating potential of reactive intermediates of the tested compounds (19). In this assay, cysteine residues of P4501A1 and P4501A2 are alkylated resulting in a decrease in enzyme activity (34). The currently tested compounds Te-phenyl-L-tellurocysteine, Se-phenyl-L-selenocysteine, and S-phenylL-cysteine all inhibited EROD activity, although with a varying potency (Figure 4B). The inhibition of EROD activity was time-dependent and most profound for Tephenyl-L-tellurocysteine, while the inhibition levels by Sephenyl-L-selenocysteine and S-phenyl-L-cysteine were comparable. As was demonstrated for mGST activation, EROD inhibition was also dependent on the presence of NADPH. This NADPH-dependent protein alkylation was shown previously for the positive control PTU (19). In addition to protein alkylation, the reactivity of oxidative products of Te-phenyl-L-tellurocysteine, Sephenyl-L-selenocysteine and S-phenyl-L-cysteine with thiols was investigated by measuring time-dependent thiocholine oxidation (Figure 4C). While 100 µM Sphenyl-L-cysteine and Se-phenyl-L-selenocysteine did not result in a significant decrease in thiocholine levels, Tephenyl-L-tellurocysteine showed a time-dependent oxidation of thiocholine. The extent of thiocholine oxidation by reactive intermediates of Te-phenyl-L-tellurocysteine was comparable to that of PTU, which was used as a positive control (19). It has been reported previously that tellurite (TeO32-) and diaryl tellurides in the presence of oxidative enzymes can rapidly oxidize thiol compounds such as GSH (35, 36). Moreover, following oxidation, diaryl tellurides react more rapidly with GSH than their corresponding selenium analogues in biological and chemical systems, as was also observed in the present study (35). Previously, selenides were shown to be activated by FMO enzymes into their corresponding selenoxides, which were highly reactive with various thiols (18, 40). Analogous to this, telluroxides may very well be responsible for the depletion of thiocholine observed in this study (Figure 3). In conclusion, the microsomal activation studies of the three conjugates seem to follow the trend Te g Se g S. The tellurocysteine Te-conjugate showed the highest effect in all three tests, whereas the cyseine S-conjugate consistently showed the lowest effects. The selenocysteine Se-conjugate had similar effect as the tellurocysteine Teconjugate in the mGST-activation assay, probably because the activation by these compounds was already

1616

Chem. Res. Toxicol., Vol. 15, No. 12, 2002

Rooseboom et al.

Table 2. Inhibition of P450 Activities by S-Phenyl-L-cysteine, Se-Phenyl-L-selenocysteine, and Te-Phenyl-L-tellurocysteinea P450

S-phenyl-L-cysteine

Se-phenyl-L-selenocysteine

Te-phenyl-L-tellurocysteine

remaining activity in % 1A1 1A2 2C9 2C19 2D6 2E1 3A4 a

25 µM

250 µM

25 µM

250 µM

25 µM

250 µM

98.7 ( 0.1 100.0 ( 2.0 100.0 ( 6.7 99.3 ( 7.0 94.4 ( 1.3 97.4 ( 5.2 101.1 ( 3.6

75.3 ( 0.1 94.1 ( 3.7 79.7 ( 6.0 91.5 ( 0.8 85.5 ( 1.2 103.0 ( 5.4 94.3 ( 2.2

97.0 ( 1.3 101.2 ( 9.7 104.1 ( 5.5 99.4 ( 4.1 97.4 ( 5.8 97.3 ( 2.8 100.6 ( 3.9

43.9 ( 0.9 86.3 ( 3.9 64.6 ( 1.3 89.7 ( 4.9 67.6 ( 3.8 70.7 ( 0.3 70.2 ( 5.4

15.2 ( 0.5 49.6 ( 1.6 101.1 ( 8.0 46.9 ( 4.3 53.8 ( 1.3 96.5 ( 13.2 83.1 ( 1.2

1.0 ( 0.1 8.0 ( 0.1 53.7 ( 2.9 32.6 ( 3.6 10.8 ( 1.7 58.6 ( 2.0 47.1 ( 0.8

Results are presented as mean value ( SD for three individual preparations.

maximal in 5-10 min. The fact that the extent of effects in the three assays is quantitatively not exactly the same may be a reflection of the different characteristics of the target molecules (mGST, CYP1A, and thiocholine) and of the different chemical properties and quantities of the reactive species involved. Reactive species involved may be the oxides of the conjugates itself, which in case of selenium and tellurium are oxidative species and/or the reactive species resulting from oxidative β-elimination. The consequences of protein and thiol reactivity of metabolites of Te-phenyl-L-tellurocysteine remains to be established. As reviewed by Talalay et al. (41), various chemopreventive agents, that share almost no structural similarities, are able to react with sulfhydryl groups by virtue of their electrophilicity or their redox properties. It has been proposed that an interaction with sulfhydryl groups of a yet unidentified “target” protein that can subsequently react with the so-called antioxidant/electrophile response element (ARE/EpRE) is a signal for enzyme induction. By this mechanism monofunctional induction of phase II enzymes, like glutathione Stransferases, quinone reductase, epoxide hydrolase, glucuronosyltransferases, aldehyde reductase, and others, is accomplished (41). Indeed, recently we were able to demonstrate that SeCys-conjugates can induce mRNA levels of glutathione S-transferase isoenzymes (42). Analogous to this, Te-phenyl-L-tellurocysteine may be a chemopreventive agent by induction of phase II enzymes. Inhibition of Human Cytochrome P450 Isoenzymes. In addition to the NADPH-dependent inhibition of EROD activity in rat liver microsomes, the direct inhibition of several human P450s was investigated. Recently, we have shown that SeCys-conjugates caused a relative selective inhibition of P450 1A1 (Venhorst et al., submitted for publication). In the present study, P450 1A1 was strongly inhibited by Te-phenyl-L-tellurocysteine resulting in a complete inhibition at 250 µM and 85% inhibition at 25 µM (Table 2). P450 1A1 inhibition by Sephenyl-L-selenocysteine was less profound than that by Te-phenyl-L-tellurocysteine, while S-phenyl-L-cysteine only moderately inhibited P450 1A1 activity (25% inhibition at 250 µM). For P450 1A2 and P450 2D6, a similar inhibition pattern was observed; again Te-phenyl-Ltellurocysteine showed the most profound inhibition, i.e., 92 and 90% inhibition at 250 µM for P450 1A2 and P450 2D6, respectively (Table 2). Although the profile of inhibition was similar to that obtained for P450 1A1, the levels of inhibition of P450 1A2 and P450 2D6 were much lower than that obtained for P450 1A1. P450 2C9, P450 2E1, and P450 3A4 activity were moderately inhibited by each of the three compounds tested (Table 2). Only at a high concentrations (250 µM) was a significant inhibi-

tion observed. The profile of inhibition observed for P450 1A1 and P450 1A2 was also found for P450 2C9, P450 2E1, and P450 3A4. P450 2C19 was more strongly inhibited by Te-phenyl-L-tellurocysteine than P450 2C9 with inhibition percentages of 53 and 68% at substrate concentrations of 25 and 250 µM, respectively. As was observed for all other P450s tested, P450 2C19 inhibition was most profound for Te-phenyl-L-tellurocysteine (Table 2). The relatively selective inhibition of P450 1A1, previously shown for Se-phenyl-L-selenocysteine (Venhorst et al., submitted for publication), was also observed for Tephenyl-L-tellurocysteine, however, the reason for this selective inhibition remains unclear. It is of interest that P450 1A1 is responsible for the activation of procarcinogens, such as benzo[a]pyrene, and that selective inhibition of P450 1A1 is therefore an interesting tool in chemoprevention (37, 38). Moreover, based on the present results, human P4501A1 appears to be more susceptible to inhibition by Te-phenyl-L-tellurocysteine than rat P4501A1, since human P4501A1 was inhibited by 85% at 25 µM, while EROD activity in rat liver microsomes was only very weakly inhibited (