Chem. Res. Toxicol. 1992,5, 231-237 duced by copper (11) ion and hydrogen peroxide. J. Biol. Chem. 264, 15435-15440. (25) Arai, H., Kogure, K., Sogioka, K., and Nakano, M. (1987)Importance of two iron-reducing systems in lipid peroxidation of rat brain: Implications for oxygen toxicity in the central nervous system. Biochem. Znt. 14,741-749. (26) Jamieson, D. (1989)Oxygen toxicity and reactive oxygen metabolites in mammals. Free Radical Biol. Med. 7, 87-108. (27) Bondy, S.C., McKee, M., and LeBel, C. P. (1990)Changes in synaptosomal pH and rates of oxygen radical formation induced by chlordecone. Mol. Chem. Neuropathol. 13,95-106. (28) LeBel, C. P., and Bondy, S. C. (1991)Oxygen radicals: Com-
231
mon mediators of neurotoxicity. Neurotoxicol. Teratol. 13, 341-346. (29)Bondy, S.C., and LeBel, C. P. (1991)Oxygen radical generation as an index of neurotoxic damage. Biomed. Environ. Sci. 4, 217-223. (30) LeBel, C. P.,Ali, S. F., and Bondy, S. C. (1991)Deferoxamine inhibits methylmercury-induced increases in reactive oxygen species formation in rat brain. Toxicol. Appl. Pharmacol. (in press). (31) Mattia, C. J., LeBel, C. P., and Bondy, S. C. (1991)Effecta of toluene and its metabolites on cerebral reactive oxygen species generation. Biochem. Pharmacol. 42,879-882.
Formation of Mitochondrial Phospholipid Adducts by Nephrotoxic Cysteine Conjugate Metabolites Patrick J. Hayden,tJ Clement J. Welsh,+#§ Yun Yang,l William H. Schaefer,l Anthony J. I. Ward,'[and James L. Stevens*-? W.Alton Jones Cell Science Center, Lake Placid, New York 12946,Department of Chemistry, Clarkson University, Potsdam, New York 13699-5548,and SmithKline Beecham Pharmaceuticals, King of Prussia, Pennsylvania 19406 Received September 6,1991
Nephrotoxic cysteine conjugates derived from a variety of halogenated alkenes are enzymatically activated via the 0-lyase pathway to yield reactive sulfur-containing metabolites which bind covalently to cellular macromolecules. Mitochondria contain @-lyaseenzymes and are primary targets for binding and toxicity. Previously, mitochondrial protein and/or DNA have been considered as molecular targets for cysteine conjugate metabolite binding. We now report that metabolites of nephrotoxic cysteine conjugates form covalent adducts with rat kidney mitochondrial phospholipids. Rat kidney mitochondria were incubated with the Tdabeled conjugates S-(1,1,2,2-tetrafluoroethyl)cysteine (TFEC),S- (2-chloro-1,1,2-trifluoroethyl) cysteine (CTFC), S-(1,2-dichlorovinyl)-~-cysteine, and S-(1,2,3,4,4-pentachlorobutadienyl)-~-cysteine. Quantitation of metabolite binding to whole mitochondria and to mitochondrial protein and lipid fractions revealed that as much as 42% of the 35S-labelassociated with the mitochondria was found in the lipid fraction. Total lipids were also extracted from %-treated mitochondria and separated by thin-layer chromatography. 35S-Containingmetabolites were found in the lipid fractions from mitochondria treated with each of the conjugates. Lipids from both [35S]CTFC- and [35S]TFEC-treated mitochondria contained major %Llabeled lipid adducta which had similar mobility by thin-layer chromatography. Fatty acid analysis, 19Fand 31PNMR spectroscopy, and mass spectrometric analyses confirmed that the major TFEC and CTFC adducts are thioamides of phosphatidylethanolamine.
I ntroductlon Halogenated alkenes induce nephrotoxicity after enzymatic conjugation with GSHl in the liver. The GSH conjugates are metabolized to the corresponding cysteine conjugates and mercapturates during a complex pathway of interorgan disposition ( I , 2;Figure 1). 0-Elimination of a toxic sulfur-containing metabolite from the cysteine conjugate occurs in the kidney via the action of cysteine conjugate 0-lyase (3; EC 4.4.1.13). Covalent binding of the reactive sulfur-containing metabolite to cellular macromolecules is presumed to initiate a cascade of events which *Towhom correspondence should be addressed at the W. Alton Jones Cell Science Center, 10 Old Barn Rd.,Lake Placid, NY 12946. W. Alton Jones Cell Science Center. Present address: Laboratory of Molecular Biophysics, National Institute of Environmental Health Sciences, P.O. Box 12233,Research Triangle Park, NC 27709. 8 Present address: Laboratory of Nutritional and Molecular Regulation, National Cancer Institute-Frederick Cancer Research Facility, Frederick, MD 21702. 11 Clarkson University. SmithKline Beecham Pharmaceuticals.
*
eventually leads to cell death (4-7)and, in some cases, mutagenesis (8,9).However, the identity of the critical targets for binding and the mechanisms which couple binding to cell death remain unclear. For recent reviews, see refs 9-12. A considerable amount of evidence implicates the mitochondrion as a primary target for cysteine conjugate toxicity. Mitochondria contain &lyase enzymes (13-16), and recently, binding of TFEC metabolites to kidney protein was shown to be localized to specific proteins of the mitochondrial fraction in vivo (17). Additionally, metabolism of cysteine conjugates has been shown to result inhibition of in inhibition of respiration (6,7,13,18-21), 2-oxoacid dehydrogenases (19),isocitrate dehydrogenase, and succinate dehydrogenase (21),loss of lipoyl del Abbreviations: S-(1,2-dichlorovinyl)-~-cysteine, DCVC; S42-chloro1,1,2-~rifluoroethyl)-~-cysteine, CTFC; S-(1,1,2,2-tetrafluoroethyl)-~cysteine, TFEC; S-(1,2,3,4,4-pentachlorobutadienyl)-~-cysteine, PCBC;
thin-layer chromatography, TLC; phosphatidylethanolamine, PE; (chlo-
rofluorothioacetamido)phosphatidylethanolamine,CFTA-PE (difluorothioacetamido)phosphatidylethanolamine,DFTA-PE; glutathione,GSH;
trichloroacetic acid, TCA; fast atom bombardment, FAB; (ethylenedinitri1o)tetraacetic acid, EDTA.
0S93-22Sx/92/2~05-0231~Q3.00/0 0 1992 American Chemical Society
232 Chem. Res. Toxicol., Vol. 5, No. 2, 1992
Hayden et al.
Structures of cysteine +conjugates ( II 1
TFEC and CTFC2 form stable thioamide adducts with protein lysyl residues and may form unstable adducts with sulfhydryl nucleophiles such as cysteinyl residues or GSH. However, adducts other than thioamides or dithioesters may form as well, since the nature of the reactive intermediate produced may be quite different depending on the halogen substitution. The chlorinated olefin conjugates may form both thioketenes (III)and thionoacyl halides (IV; 33) while thionoacyl halides are favored for the fluorinated aliphatic conjugates (30, 33). Thioketenes form cycloaddition products with cyclopentadiene (VI; 33),suggesting that this pathway could yield novel adducts, other than thioamides, with biological molecules. In addition, adducts other than thioamides could also form from conjugates which are metabolized to reactive thiirane intermediates (34);however, these adducts have not been characterized. Given the several types of reactive intermediates which may be formed from cysteine conjugates, structural analysis of the adducts formed in biological systems would shed light on both the chemistry and the toxicity of the halogenated ethylenes. Although proteins are important molecular targets for cysteine conjugate metabolite binding, lipids also contain nucleophilic functional groups which could react with the thionoacylating species. Indeed, Derr and Schultze (35) showed that while %-metabolites from [Y3]DCVC bound predominately to protein, a significant portion (up to 16%) of the metabolites was bound to the lipid fraction of rat tissues after treatment in vivo. In the present study we investigated the interaction of cysteine conjugate metabolites with rat kidney mitochondrial lipids. Four conjugates, PCBC, DCVC, TFEC, and CTFC, produced apparent covalent adducts with mitochondrial lipids. TFEC and CTFC produced significant amounts of a thioamide derivative of phosphatidylethanolamine.
xi x2 x3
Materials and Methods
x1-
x3
n
x2
(I)
+GSH
x4
CTFE TFE
TCE HCBD R-S
R-S x1
x3
UX3
n
1
P-lyase
DCVC(Ua)
-
CI CI
Ci CI
CTFC (Ub)
F
F
F
TFEC (Ub)
F
F
F
PCBC (Ua)
x4 W C'
Synthesis of Cysteine Conjugates. Cysteine conjugates and cinci 35S-labeledcysteine conjugates were synthesized as previously H reported (16). The radiochemical purity was >98% by HPLC CI F
Figure 1. Scheme of halogenated alkene metabolism. The scheme depicts the metabolism of halogenated alkenes (I) after conjugation with glutathione (GSH). The GSH conjugates are degraded to the cysteine conjugates (11),which may be vinylic (IIa) or aliphatic (IIb) depending on the haloalkene, with loss of glutamate (glu) and glycine (gly). The cysteine conjugates are acted on by cysteine conjugate @-lyase(@-lyase)to give either a thioketene (111)or a thionoacyl halide (IV) reactive intermediate, depending on the structure of the conjugate. Either of these reactive species could give rise to an adduct upon reaction with a nucleophile (V). A nucleophilic amine would give a thioamide adduct while a thiol would yield a dithioester adduct. In addition, the thioketenes have been shown to give rise to cycloaddition products with cyclopentadiene (VI; 33). For the cysteine conjugates (11), R = +NH3CH(CH2)COO-.TCE = trichloroethylene, HCBD = hexachlorobutadiene, CTFE = chorotrifluoroethylene, TFE = tetrafluoroethylene. hydrogenase (22), collapse of membrane potential and ability of cells or mitochondria to sequester calcium (6, 23-25), and depletion of ATP and mitochondrial GSH (5, 6 , 2 1 ) . Although numerous mechanisms have been proposed to account for these effects, at present their cause remains uncertain. Recently, several groups have used nucleophilic amines (diethylamine, aniline, or benzylamine) to trap reactive metabolites of several cysteine conjugates as thioamides (V;Figure 1; 26-31). Subsequently, Hayden et al. (32) showed that the thionoacetylating metabolites (IV) of
for all the conjugates. There was no significant loss of purity (1-270) over a 1-year period although the 36S-specificactivities did decrease. The specific activities reported were determined at the time the experiment was performed. All other chemicals were obtained from commercial sources. Isolation of Rat Kidney Mitochondria. Mitochondria were isolated from the kidney cortex of male Sprague-Dawley rats as previously reported (16) according to the procedure of Schnaitman and Greenawalt (36). This procedure typically produces mitochondria with respiratory control ratios of 5-7 using site I substrates. For large-scale isolation of lipid adducts, 20 kidneys were used, yielding approximately 500 mg of mitochondrial protein. Isolation of Total Mitochondrial Lipids. Mitochondrial lipids were extracted after resuspending mitochondrial pellets in water by addition of 1 volume of 2% acetic acid in CH30H and 1volume of CHC1, according to the method of Bligh and Dyer (37). After separation of the liquid solvent phases by centrifugation, the lower phase containing the total lipids was collected and concentrated to dryness under a stream of nitrogen gas. The dried lipid residue was dissolved in CHC13/CH30H (2/1) for further analysis. Thin-Layer Chromatography and Autoradiography. TLC was performed on 250-wm silica gel G layers (Analtech, Newark, DE) using (A) CHC13/CH30H/NH40H(65/35/3), (B) CHC13/ CH30H/ethyl acetate/40% aqueous CH3NH2(20/20/20/10), or (C) hexane/ether/acetic acid (60/40/1) as developing solvent systems. Lipids were visualized by spraying with a 10% CuS04/10% reagent and heating at 180 OC. For preparative TLC, continuous sample bands were applied at the origin Unpublished results for CTFC.
Cysteine Conjugate-Phospholipid Adducts of 250-pm plates. Representative areas at either side of the plate were exposed to iodine vapor in order to visualize the separated lipids and to guide in selection of appropriate areas for recovery of modified lipids. Lipids were recovered after scraping the silica gel from the plate by eluting with CHC13/CH30H (2/1) and concentrating the solvent by evaporation under nitrogen. In some experiments, dried extracts of 35S-labeledlipids from mitochondria treated with 35S-conjugates were dissolved in CHCl3/CHBOH(2/1), and -200000 cpm/lane were applied to the TLC plates. Plates were developed, dried, and exposed to X-ray film (Du Pont, Wilmington, DE) a t -70 "C using an enhancer screen (Fisher Biotech, Pittsburgh, PA) for 4 days. In each experiment, lipid standards were run as intemal controls and were visualized by charring as described above. Treatment of Isolated Mitochondria with Cysteine Conjugates. Binding of [35S]metabolites to whole mitochondria, protein, or lipids was determined in isolated mitochondria (2.0 mg) which were incubated with 50 nmol of [35S]PCBC (4.2 pCi/pmol), [%]DCVC (1.9 pCi/pmol), [%]TFEC (3.1 pCi/pmol), or [35S]CTFC (2.4 pCi/pmol) in a total volume of 500 pL of isolation buffer for 1h at 37 "C. After incubation, whole mitochondria were collected by centrifugation at l6000g for 5 min and the mitochondrial pellets were resuspended in isolation buffer and recentrifuged. The pellets were then dissolved in 500 pL of 0.1 N KOH, and total binding was quantitated by scintillation counting. Binding of [%]metabolites to protein was determined after adding 2.0 mL of 10% TCA directly to incubation mixtures. Protein was allowed to precipitate for 1h at 4 "C, and precipitated protein was collected onto Whatman GF/C glass fiber filters by vacuum filtration. After washing successively with 2 volumes (5 mL) of 5.0% TCA, 95% ethanol, and CHC13, the filters were immersed in scintillation cocktail and the 35S-labelretained on the filter was quantitated by scintillation counting. Control samples either were kept on ice for an equivalent amount of time or were precipitated immediately after adding radiolabel to exclude nonenzymatic formation of adducts during the procedure. Binding of [35S]metabolites to mitochondrial lipids was determined after collecting mitochondria by centrifugation at 16OOOg for 5 min. The mitochondrial pellet was resuspended in 500 pL of water, and lipids were extracted by addition of 500 pL of 2.0% acetic acid in CH30H and 500 pL of CHC13 The solvent phases were separated by centrifugation a t l6000g for 5 min, and the lower phase containing the lipids was collected. The aqueous (top) phase was extracted again by addition of 500 pL of CHC13. The lower phases from both extractions were combined, and 36S-label associated with the lipid fraction was quantitated by scintillation counting. Control incubations were carried out as described above to exclude nonenzymatic formation of adducts. When TLC and autoradiography experiments were performed, mitochondria were diluted to a concentration of -40 mg of protein/mL in isolation buffer. Aliquots from stock solutions of 35S-labeled conjugates, [35S]DCVC (16 mM; 119 pCi/pmol), [?3]TFEC (34 mM; 63 pCi/pmol), or [35S]CTFC (11 mM; 26 pCi/pmol), were added to 250 pL of mitochondria to give a final concentration of 1 mM. Due to its poor solubility, the final concentration of [%]PCBC (stock solution 2.3 mM, 105 pCi/pmol) was limited to 200 pM. All the samples were incubated for 1 h at 37 "C. When large-scale isolations of lipid adducts derived from CTFC and TFEC were performed, mitochondria (20 mg/mL) were incubated for 1h a t 37 "C in 20 mL of isolation buffer containing 5.0 mM conjugate. Lipids were isolated and separated by TLC as described above. Analytical Methodologies and Instrumentation. Gas chromatography was performed on a Varian 3200 instrument equipped with a 6-ft X l/&in. stainless steel column packed with 5% DEGS-PS (100/120 mesh Supelcoport; Supelco, Bellefonte, PA), with flame ionization detection. The carrier gas was helium. The temperature program was as follows: 100 "C for 4 min; ramp at 5 "C/min to 188 "C; and maintain at 188 "C to the end of the run. Injector and detector temperature were 250 "C. Fatty acid methyl ester derivatives of phospholipids were obtained by dissolving approximately 50 pg of phospholipid in benzene and adding 3 pL of METH-PREP I1 [0.2 N methanolic [m-(trifluoromethy1)phenylltrimethylammonium hydroxide] (AlltechApplied Science, Deerfield, IL). After incubation for 15 min at
Chem. Res. Toxicol., Vol. 5, No. 2, 1992 233 Table I. Association of Cysteine Conjugate Metabolites with Protein and Lipid Fractions of Isolated Rat Kidney Mitochondriaa nmol bound conjugate total protein lipid PCBC - AOA 33.2 f 7.8 19.5 f 1.3 6.9 f 0.4 PCBC + AOA 7.7 f 1.3 5.2 f 0 2.9 f 0.7 DCVC - AOA 17.6 f 0.2 12.0 f 0.6 5.3 f 0.4 DCVC + AOA 1.4 f 0.1 1.0 f 0.3 0.6 f 0.8 CTFC - AOA 18.2 f 1.4 8.9 f 0.7 7.0 f 0.1 CTFC + AOA 2.8 f 0.8 1.3 f 0 1.0 f 0.3 TFEC - AOA 17.9 f 0.7 10.0 f 1.9 7.2 f 0.6 TFEC + AOA 1.4 f 0.6 1.0 f 0.5 1.0 f 0.3 2.0 mg of mitochondria was incubated with 50 nmol of the indicated [%3]cysteine conjugate for 1 h at 37 "C with or without 200 pM AOA. Binding of [35S]metabolitesto whole mitochondria, protein, or lipid fractions was then determined as described in Materials and Methods. Results are the mean f the range of 2 experiments (n = 2) and are presented as nmol of 35S-labelfound in whole mitochondria (total) or in the protein or lipid fractions. room temperature, 3 pL of the sample was injected onto the GC column. 19Fand 31PNMR spectra were recorded with an IBM NR/250 FTNMR spectrometer operating at 235.33 and 101.26 MHz, respectively. Samples were dissolved in CHCl3/CH3OH (2/1) for 19Fexperiments, and in CHC13/CH30H/0.2 M EDTA, pH 6.0 (10/4/1), as described by Meneses and Glonek (38) for 31Pexperiments. Fluorine chemical shifts were referenced to trifluoroacetic acid. Phosphorous chemical shifts were obtained by first assigning purified egg phosphatidylethanolamine (Avanti Polar Lipids, Alabaster, AL) a chemical shift of 0 ppm and obtaining the chemical shift of phosphoric acid (-1.108 ppm) relative to PE. Chemical shifts were referenced to phosphoric acid (-1.108 ppm) in subsequent experiments. Mass spectra were acquired on a VG 7070E-HF instrument operated with an accelerating potential of 6 kV using FAB ionization. Fast xenon atoms were generated using a saddle-field fast atom gun operated at 8 kV and 1mA. Samples were ionized from a matrix of triethanolamine. Mass spectra were obtained in continuous mode with the magnet scanning at a rate of 20 s/dec from m/z 1300 to 100. Isotopic distributions were determined from summed continuous mass spectra obtained using linear voltage scans over a narrow mass range (m/z
