I n Vivo Detection and Characterization of Protein Adducts Resulting

Environmental Health Sciences Center and Department of Pharmacology, University of. Rochester School of Medicine and Dentistry, 601 Elmwood Avenue,...
0 downloads 0 Views 1MB Size
34

Chem. Res. Toxicol. 1992, 5, 34-41

I n Vivo Detection and Characterization of Protein Adducts Resulting from Bioactivation of Haloethene Cysteine S-Conjugates by ''F NMR: Chlorotrifluoroethene and Tetrafluoroethenet James W. Harris,* Wolfgang Dekant,s and M. W. Andem*?* Environmental Health Sciences Center and Department of Pharmacology, University of Rochester School of Medicine and Dentistry, 601 Elmwood Avenue, Rochester, New York 14642,and Institut fur Toxikologie, Universitat Wurzburg, Versbacherstrasse 9,0-8700 Wurzburg, Germany Received August 14, 1991

Several haloalkenes are selective nephrotoxins. The bioactivation of nephrotoxic haloalkenes involves hepatic glutathione S-conjugate formation, peptidase-catalyzed metabolism of the glutathione S-conjugates to the corresponding cysteine S-conjugates, uptake of cysteine Sconjugates by the kidneys, and renal cysteine conjugate @-lyase-catalyzed@-eliminationof a thiol. The haloalkyl and haloalkenyl thiols thus released are unstable and yield reactive intermediates whose interactions with cellular constituents are thought to contribute to the observed toxicity of S-conjugates. Tetrafluoroethene and chlorotrifluoroethene are metabolized to the cysteine (TFEC) and S-(2-chloro-1,1,2-trifluoroS-conjugates S-(1,1,2,2-tetrafluoroethyl)-~-cysteine ethy1)-L-cysteine(CTFC), respectively. Administration of TFEC (1.0 mmol/kg) or CTFC (1.0 mmol/kg) to rats resulted in acylation of renal proteins, as demonstrated with 19Fnuclear magnetic resonance spectroscopy. Single, broad resonances near 41 or 56 ppm were found in spectra of renal proteins from TFEC- or CTFC-treated rats, respectively, and these resonances were not lost on dialysis. Renal protein incubated with 2-chloro-l,l,2-trifluoroethyl2-nitrophenyl disulfide, a proreactive intermediate that yields 2-chloro-l,l,2-trifluoroethanethiol, showed the same 19FNMR spectrum as was found with CTFC-treated rats. In vitro incubation of various N*-blocked amino acids with this proreactive intermediate indicated that only lysine is stably adducted, whereas histidine is transiently acylated. In each case, proteolysis of modified protein converted a single broad NMR resonance to a doublet with little change in chemical shift and with clearly resolved, characteristic H-F couplings. The single, stable amino acid adduct formed with renal proteins of rats given CTFC or TFEC was Ne-(chlorofluorothioacety1)lysineand Nf-(difluorothioacetyl)lysine,respectively.

Introduction Glutathione and cysteine S-conjugate formation are usually associated with the detoxication of xenobiotica (2). Cysteine S-conjugates are formed by enzymatic conjugation of xenobiotics with the tripeptide glutathione and subsequent peptidase cleavage. For example, chlorotrifluoroethene is a substrate for microsomal glutathione S-transferase (3), which catalyzes the regiospecific and stereoselective addition of glutathione to the haloalkene (4).Furthermore, the cysteine S-conjugateTFECl (Figure 1; lb) was isolated from bile of rata exposed to tetrafluoroethene by inhalation (5). Enzymatic N-acetylation of cysteine S-conjugates yields the corresponding mercapturic acids, which are excreted in the urine. Cysteine S-conjugates of many industrially and environmentally important halogenated alkenes are nephrotoxic (6). Cysteine conjugate @-lyase(EC 2.6.1.64), which is identical with renal cytosolic glutamine transaminase K (7), catalyzes @-eliminationreactions of many cysteine S-conjugates,yielding a thiolate, pyruvate, and ammonia (8). Several studies have shown that the thiolates released by the action of @-lyaseon cysteine S-conjugatesof several haloalkenes afford reactive, electrophilic species (Figure + Apreliminary account of this work was presented at the Fourth International Symposium on Biological Reactive Intermediates, January 1990 (I). * Author to whom correspondence should be addressed. f University of Rochester School of Medicine and Dentistry. 8 Universitiit Wiirzburg.

1): when the metabolism of CTFC (la; 9 ) , S-(l,Zdichloroviny1)-L-cysteine,S-(1,2,2-trichlorovinyl)-~-cysteine (lo),and S-(pentachlorobutadieny1)-L-cysteine(11)was studied in enzyme model systems with diethylamine as a model nucleophile, thioamides were detected as products. In the absence of trapping agents, halocarboxylic acids were detected as terminal products. Taken together, these observations indicate the intermediacy of electrophilic thioacyl halides in S-conjugate metabolism. Recent studies show that halothioketenes may be formed from chloroalkene-derived S-conjugates (12). Covalent binding of reactive metabolites of cysteine S-conjugates to cellular proteins has been demonstrated. Darnerud et al. (13) showed that 14C from [ ~ i n y l - ' ~ C ] S-(l,%-dichlorovinyl)-~-cysteine becomes covalently bound to proteins in renal proximal tubule cells of rata. Vamvakas et al. (14)demonstrated that [14C]-S-(pentachlorobutadieny1)-L-cysteine is converted by rat liver and kidney homogenates to a metabolite that becomes covalently bound to homogenate proteins. Finally, metaboliites of 35S-labeledCTFC and TFEC bind to macromolecules of isolated rat kidney mitochondria (15)and proximal tubular cells (16). Although these studies show that covalent binding to proteins results from cysteine S-conjugate metabolism both Abbreviations: CTFC, S-(2-chloro-l,l,2-trifluoroethyl)-~-cysteine; TFEC, S-(1,1,2,2-tetrafluoroethyl)-~-cysteine; 8-lyase,cysteine conjugate @-lyase;NMR, nuclear magnetic resonance spectroscopy; FAB-MS, fast atom bombardment mass spectrometry;GC-MS, gas chromatographymass spectrometry; SDS, sodium dodecyl sulfate.

0a93-22ax/92/2~05-003~~03.00/00 1992 American Chemical Society

Chem. Res. Toxicol., Vol. 5, No. 1, 1992 35

Bioactiuation of Haloethene Cysteine S-Conjugates F

:

.F

F

.F

NO,

H * + O X

F

NH,

+

r

(8)

1

x = CI

X = F X = CI; R E H X = C1; R = ( ~ ) x = F ; R = H (f) X = F ; R = (b) (c) (d)

1

4' 4-

Jp+H20

J $O .HX

F

/

-H2S

-F-

\

+Amine

HX & NF H U R

Figure 1. Proposed pathway of CTFC and TFEC bioactivation. la,CTFC; lb,TFEC; 2,2-chloro-1,1,2-trifluoroethyl2-nitrophenyl disulfide; 3a, 2-chloro-1,1,2-trifluoroethanethiolate;3b, 1,1,2,2,-tetrafluoroethanethiolate;4a, chlorofluorothioacetyl fluoride;4a,ditluorothioacetylfluoride;5a,chlorofluoroacetic acid; 5b,difluoroacetic acid; 6c, (ch1orofluorothioacetamido)ethane; 6d, Nf-(chlorofluorothioacety1)lysine; 6e, (difluorothioacetamido)ethane; 6f, Ne-(difluorothioacety1)lysine.Cleavage of 2chloro-l,l,2-trifluoroethyl2-nitrophenyl disulfide (2)generates the same thiolate (3a)that is released by 0-lyase-catalyzed metabolism of CTFC (la).

in vivo and in vitro, the adducts formed with proteins have not been characterized, and the question of selectivity in amino acid adduct formation has not been addressed. The objective of these experiments was to investigate the fate of reactive intermediates resulting from S(haloalky1)cysteine conjugate metabolism in vivo by characterizing the adducts formed with renal proteins.

Experimental Procedures Analyses. Proton and fluorine NMR spectra were obtained with a Bruker WP-27OSYspectrometer equipped with a dedicated 5-mm 19F probe operating at 254.18 MHz for fluorine. For 19F NMR, the pulse width was 3 ps, the interpulse time was 0.7 s, and the spectral width was 50 kHz (16Kdata points). Exponential multiplication of the free-induction decay was not employed. Spectra were acquired at room temperature with sample spinning. Depending on signal strength and desired signal-to-noise ratio, 2000 or more transients were acquired during assays for covalent binding of fluorinated metabolites to proteins. Fluorine NMR chemical shifts are reported as ppm upfield from trifluoroacetamide in D20 (6 = 0 ppm) contained in a sealed coaxial tube; proton NMR chemical shifta are reported as ppm downfield from intemal tetramethylsilane (6 = 0 ppm). Fast atom bombardment m a spectra were acquired from a glycerol matrix with a TS-250 instrument (VG Inc., Manchester, U.K.). Electron impact mass spectra were recorded with a Hewlett-Packard 5880A gas chromatograph equipped with an HP-1dimethylpolysiloxane gum capillary column (25m X 0.2 mm X 0.5 pm film thickness) and coupled to an HP-5970 mass-selective detector (70eV). Chemicals. CTFC and TFEC were prepared by stirring an alkaline, ethanolic solution (4:lethanol/water; 500 mL; pH 10) of L-cysteine (0.40mol) contained in a vessel attached to a balloon

filled with chlorotrifluoroetheneor tetrafluoroethene,respectively. The haloethene atmosphere was maintained until unreacted cysteine was no longer detected by thin-layer chromatography (silica gel; 41:l butanol/acetic acid/water; ninhydrin detection) and normally required 24-30 h. Reaction mixtures were contaminated only by cystine (typically 5%). The conjugates were purified by repeated recrystallization from ethanol/water (5:2, pH 5) and resulted in yields of 20-30%. Spectral data for the products ('H NMR, '9 NMR,and FAB-MS) were consistent with values previously reported (3, 17). The mercapturic acids of CTFC and TFEC were similarly prepared except that N-acetyl-L-cysteine (10 mmol) replaced cysteine. The products, recovered as oils, were purified by flash chromatography on silica gel eluted with 9:l ethanol/acetic acid (yield = 80%). 'H and 19FNMR values for mercapturic acids and GC-MS values for mercapturic acid methyl esters (prepared with diazomethane)were consistent with those previously reported (17, 18). 2-Chloro-1,1,2-trifluoroethyl2-nitrophenyl disulfide was prepared as described previously (12). (Chlorofluoroacetamido)ethane and (difluoroacetamido)ethane were prepared by acylation of ethylamine (Aldrich, Milwaukee, WI) with chlorofluoroacetyl chloride or difluoroacetyl chloride, respectively. Ethyl chlorofluoroacetate, used as the starting material for the preparation of chlorofluoroacetylchloride (19), was prepared by the method of Englund (20, 21). Ethyl difluoroacetate was purchased from PCR, Inc. (Gainesville, FL). Caution(!): The preparation and distillation of the acetyl chlorides (19) should be carried out in an efficient hood, and the corrosive, volatile byproducts of the reaction should be vented continuously during distillation to avoid overpressurization of the closed system. The freshly distilled dihaloacetyl chloride (150 mmol) was added to a closed, stirred solution of ethylamine (100 "01) and triethylamine (100"01) dissolved in chloroform (75 0 OC) over a 15min period. After stirring the reaction mixture overnight at room temperature, the chloroform and haloacetic acids were removed in vacuo. Ethyl ether (100mL) was added to the residue, the solution was filtered to remove triethylamine hydrochloride, and the ether was removed in vacuo. Products were chromatographed on silica gel columns eluted with chloroform to yield colorless oils. (Chlorofluoroacetamido)ethane[MS, m/z (re1intensity) 141 (41,139(12),126 (31,124(lo), 104 (6),72 (100),69(6),67(191,44 (83);'H NMR (CDC13) 6 1.21 (t, 3 H, J = 7.3 Hz), 3.38 (d oft, 2 HI J = 6.8 and 7.3 Hz), 6.37 (d, 1 H, J = 50.6 Hz);'?F NMR (4:10.4M phosphate buffer, pH 7.4/ethanol) 6 69.9 (d, J = 50.6 Hz)] was recovered in a yield of 76%. (Difluoroacetamido)ethane [MS, m/z (re1 intensity) 123 (31),108 (261,72 (681,51 (39),44 (100);'H NMR (CDCls) 6 1.20 (t, 3 HI J = 7.3 Hz), 3.37 (d of t, 2 H, J = 6.8 and 7.3 Hz),5.94 (t, 1 HI J = 54.2 Hz); 19F NMR (4:l 0.4 M phosphate buffer, pH 7.4/ ethanol) 6 51.9 (d, J = 54.2Hz)] was recovered in a yield of 50%. (Ch1orofluorothioacetamido)ethaneand (difluorothioacetamido)ethane were prepared by reacting the corresponding amides (20"01) with 2,4bis~-methoxyphenyl)-l,3-dithiadiphospheta11e 2,4disulfide (Lawesson's reagent; Aldrich; 12 "01) in chloroform (75mL). Reactions were conducted in sealed, stirred flasks at 50 "C for 12 h. The progress of the reaction was monitored by GC-MS, which indicated complete conversion from amide to thioamide in all cases. The solvent was removed in vacuo, and the residue was chromatographed on silica gel columns eluted with chloroform to yield red oils. The silica gel column packing was used once and then discarded in these experiments. (Chlorofluorothioacetamido)ethane [MS, m/z (re1 intensity 157 (29),155 (77),140 (l),120 (loo), 88 (17),76 (56),60 (80), 44 (39); 'H NMR (CDC13) 6 1.33 (t, 3 H, J = 7.3 Hz), 3.73 (d o f t , 2 HI J = 6.4 and 7.3 Hz),6.64 (d, 1 H, J = 51.8 Hz)] was recovered in a yield of 68%. (Difluorothioacetamido)ethane[MS, m / z (re1 intensity) 139 (sa),124 (2),95 (24),88 (39),60 (loo),51 (29),44 (38);'H NMR (CDCI,) 6 1.30 (t, 3 H, J = 7.3 Hz), 3.70 (d o f t , 2 H, J = 6.4 and 7.3Hz), 6.16 (t, 1H, J = 56.2 Hz)] was recovered in a yield of 85%. 19F NMR data for these compounds are presented in Table I. In Vivo Experiments. Adult male Fischer 344 rats (Charles River Laboratories,Wilmington, MA; 200-300g) were anesthetized with a mixture of 90 mg/kg ketamine and 8 mg/kg xylazine and were given 0.1,0.5, or 1.0 mmol/kg body weight of either CTFC

36 Chem. Res. Toxicol., Vol. 5, No. 1, 1992 Table I. Summary of NMR Data" experiment in vivo protein in vitro protein proteolysis of protein

TFEC -41 ppm, broad not done 41.5 ppm, doublet, J H F = 56 HZ not done

CTFC -56 ppm, broad -56 ppm, broad 56.6 ppm, doublet, JHF = 52 HZ in vitro lysine 56.6 ppm, doublet, J H F = 52 HZ in vitro, absence of not done no resonances nucleophiles near 56 ppm (chlorofluorothioacetnac 56.6 ppm, doublet, JHF = 51.8 HZ amido)ethaneb(612) (difluorothioacetamido)- 41.5 ppm, doublet, nac J H F = 56.2 HZ ethaneb (6e)

" Upfield of trifluoroacetamide with D20as solvent; in vitro experimenta use the proreactive intermediate 2-chloro-1,1,2-trifluoroethyl 2-nitrophenyl disulfide (2). bSolvent was 0.4 M phosphate buffer/ethanol (41,pH 7.4). cNot applicable. or TFEC by intraperitoneal (ip) injection of a 110 mM aqueous solution. Solutions of S-conjugates were prepared just before injection by dissolving the S-conjugate in a minimum volume of 0.5 M HC1, adjusting the pH to 7.0 with 1.0 M NaOH, and diluting with 0.9 M NaCl to the final volume. The S-conjugate injection solutions were sterilized by passage through 0.2-pm membrane filters. One hour after giving the Sconjugates, the anesthetizedanimals were killed by puncturing the diaphragm, and the kidneys were excised. Urine was collected by syringe from the bladders of killed animals and was stored frozen until analyzed.

Preparation of Protein Fractions for NMR Analysis. Kidneys were homogenized with a Potter-Elvehjem homogenizer, and mitochondrial fractions were prepared according to the method of Johnson and Lardy (22) in 0.1 M phosphate buffer, pH 7.0 (buffer A). The supernatant from the mitochondrial fraction was centrifuged for 60 min at 100000g, yielding the cytosolic fraction as the Supernatant and the microsomal fraction as the pellet. Subcellular fractions were purified from constituents of higher or lower sedimentation coefficienta by resuspension in fresh buffer A and centrifugation. The fractions were suspended in water, placed in dialysis tubing (Spectrapor 3, Spectrum Medical Industries, Loa Angeles, CA; 3500 MW cutoff), and dialyzed against 10 mM phosphate buffer, pH 7.0, containing 0.1% (w/v) SDS to remove weakly bound metabolites (23).The dialysis buffer was changed several times during at least 48 h of dialysis at 4 OC. The dialyzed protein was lyophilized and then dissolved in D20for '9F NMR analysis. A minimum of 40 mg of protein/mL of D20was used. Protein concentrationswere measured (24) with bovine serum albumin as the standard. Protein or pOly(DL-lySine) synthetically acylated with 2chloro-l,l,2-trifluoroethyl2-nitrophenyl disulfide was prepared as follows: renal homogenates from untreated rata were fractionated as described above; no resonances were found in the 19F NMR spectra of these fractions (Figure 2A). Neat 2-chloro1,1,2-trifluoroethyl2-nitrophenyl disulfide (12.5 pmo1/100 mg of polypeptide) and 2% Triton X-100 (v/v; to aid in the dissolution of the hydrophobic disulfide) were added to an aqueous suspension of cytosolic or mitochondrial protein or of pOly(DL-lysine) (Sigma, St. Louis, MO; average molecular weight = 11.3 kDa). The mixture (pH 7.4) was stirred overnight at 4 "C and dialyzed as described above. In some experiments, protein fractions that had been treated with 2-chloro-1,1,2-trifluoroethyl2-nitrophenyl disulfide were incubated with dithiothreitol after initial dialysis and NMR analysis. Treated protein was dissolved in 10 mM phosphate buffer, pH 7.0, containing 0.1% (w/v) SDS, and 10 equiv (based dion the amount of 2-chloro-1,1,2-trifluoroethyl2-nitrophenyl sulfide originally used) of dithiothreitol was added; the solution was stirred overnight at 4 OC. The protein was again dialyzed, and 19F NMR spectra were acquired. Subcellular protein fractions (either from CTFC- or TFECtreated rata or protein treated with 2-chloro-l,l,2-trifluoroethyl 2-nitrophenyl disulfide) or pOly(DL-lySine) (treated with 2chloro-l,l,2-trifluoroethyl2-nitrophenyl disulfide) was proteolyzed by incubating the treated and dialyzed polypeptides with pro-

Harris et al. teinase K (5 mg,Sigma, from Tritirachium album) at 37 "C. The course of the incubation was followed by placing the mixture (pH 7.0, 0.60 mL) in an NMR tube and acquiring 19Fspectra periodically. Subcellular fractions that contained lipid (e.g., mitochondria or microsomes) did not undergo complete proteolysis, as determined by the line width of the adduct NMR signal, unless the lipid was extracted with chloroform and an additional portion of proteinase K added. Analysis of Urinary Metabolites. Urine samples from treated animals were analyzed by '?F NMR and by GC-MS. For NMR analysis,urine samples were fdtered, diluted with D20,and analyzed directly. Haloacetic acid analysis by GC-MS was conducted according to the phenyldiazomethane derivatization procedure reported by Karashima et al. (2.5). Urine from treated or control animals was extracted four times with 3 volumes of ethyl ether, phenyldiazomethane was added to the pooled ether fractions, the ether was evaporated, and the residue was dissolved in methanol for GC analysis (temperature programming: 50-200 OC at 10 "C/min). GC retention times of benzyl esters of chlorofluororacetic acid and difluoroacetic acid were 10.76 0.01 and 8.19 0.01 min, respectively. Urinary mercapturic acid excretion was confirmed by GC-Ms: urine was treated in a manner identical to that just described, except that samples were derivatized with diazomethane. Electron impact mass spectra of mercapturic acid methyl esters contain several characteristic fragment ions; the fragment ion a t m/z 88 (+H2N=CHC02CH3;26) was used for single ion monitoring of urine extracts. GC retention times of methyl esters of N-acetyl-CTFC and N-acetyl-TFEC were 14.58 A 0.02 and 12.26 f 0.02 min, respectively. GC retention times for derivatized haloacetic acids and mercapturic acids were established by dissolving authentic samples in urine followed by ether extraction and derivatization.

*

*

Reaction of a-Amino-BlockedAmino Acids and Imidazole with 2-Chloro-1,1,2-trifluoroethyl2-Nitrophenyl Disulfide. Solutions (50 mM) of Na-acetyl amino acids (L-cysteine,L-lysine, L-histidine, L-methionine, L-tyrosine, L-arginine, L-serine, Lglutamine, and L-asparagine; Sigma) and of imidazole (Aldrich) were prepared in 0.40 M phosphate buffer, pH 7.4, containing 2% Triton X-100 (w/v) and 20% D20 (v/v). Neat 2-chloro1,1,2-trifluoroothyl 2-nitrophenyl disulfide (10 mM final concentration) was added to these solutions, and the mixtures were transferred to sealed NMR tubes. The reaction mixtures were incubated at 40 "C in a shaking water bath, and the progress of the reaction was monitored periodically by 19FNMR. Additions of candidate nucleophiles had no obvious effect on the rate of 2-chloro-l,l,2-trifluoroethyl 2-nitrophenyl disulfide hydrolysis under these conditions, and in all cases reactions were complete within 3 h.

Results Identification of Cysteine S-Conjugate Derived Protein Adducts in Vivo. Single, broad 19FNMR resonances centered near 56 or 41 p p m were observed in kidney subcellular fractions of rats given 1.0 mmol/kg CTFC or TFEC, respectively. These NMR resonances were observed in mitochondrial, microsomal, and cytosolic subcellular fractions of kidneys from treated rats. No resonances were observed in the 19FNMR spectra of renal fractions from control rats (Figure 2A). Spectra were acquired over a chemical shift range (350 ppm) encompassing organofluorine compounds (27),thus ensuring that all resonances present were detected. Figure 2B is a representative 19F NMR spectrum of dialyzed kidney mitochondria from a rat given CTFC. The resonance shown in Figure 2B,which was n o t lost on dialysis, is centered near 56 p p m and is several hundred hertz in width. This resonance did not deviate from a central position near 56 p p m in a n y of the subcellular fractions studied. Dialyzed kidney cytosolic, mitochondrial, and microsomal protein from rats given TFEC showed a single, broad resonance centered near 41 ppm that was not lost on dialysis (Figure 3A). The chemical shift values for each of these resonances were highly reproducible and were separated by

Bioactivation of Haloethene Cysteine S-Conjugates

Chem. Res. Toxicol., Vol. 5, No. 1, 1992 37

1

A

B

C

D

1

'

"

'

1

" ' 40

'

1

'

' ~ 45

'

1

~

'

'

'

1

"

" 55

50

1

'

'

"

1 60

~

'

'

' 65

PPM

Figure 2. 19FNMR spectra of dialyzed kidney mitochondrial fractions from untreated rats (A) or from rats given 1.0 mmol/kg CTFC (B). Renal cytosolic protein from untreated rats was incubated with 2-chlor~l,l,2-trifluoroethyl2-nitrophenyl disulfide and dialyzed; '9 NhfR spectra were recorded before (C) and after (D) subsequent dithiothreitol treatment of the modified protein.

C

54.0

56.0

PPM

Figure 4. I9F NMR spectra of kidney mitochondrial (A) or cytosolic (C)protein from untreated rats after incubation with 2-chloro-l,l,2-trifluoroethyl2-nitrophenyl disulfide and dialysis. (B) Spectrum A after proteinase K treatment. (D) Spectrum C after proteinase K treatment. Small amounta of chlorofluoroacetic acid (6 = 61.3 ppm) appeared in some incubation mixtures. PPM

Figure 3. 19FNMR spectra of dialyzed TFEC-modified renal proteins. (A) Cytosolic proteins from a rat given 1.0 mmol/kg TFEC. Protein shown in panel A incubated in the presence (B) or absence (C) of proteinase K. Small amounta of difluoroacetic acid (6 = 49.1 ppm) appeared in some incubation mixtures.

approximately 15 ppm. Tissue-associated I9FNMR resonances due to CTFC and TFEC metabolism were either significantly reduced or undetectable when renal subcellular fractions from animals given either 0.5 or 0.1 mmol/kg, respectively, were analyzed (data not shown). 2-Chloro-1,1,2-trifluoroethyl2-nitrophenyl disulfide the same produces 2-chloro-1,1,2-trifluoroethanethiolate, thiolate that is released by the 8-lyase-catalyzed P-elimination reaction of CTFC (Figure 1;12). Figure 2C shows the spectrum of a synthetically prepared protein adduct obtained by reacting 2-chloro-l,l,2-trifluoroethyl2-nitrophenyl disulfide with renal proteins from untreated rats. A single, broad resonance centered near 56 ppm remained after dialysis, and this resonance was unchanged after treatment of the acylated protein with dithiothreitol (Figure 2D). Proteolysis of mitochondrial and cytosolic fractions acylated with 2-chloro-1,1.2-trifluoroethyl2-nitrophenyl disulfide converted the broad NMR resonance (Figure 4,

panels A and C, respectively) to a sharp doublet (Figure 4; panel B, mitochondrial protein; panel D, cytaolic protein). Similarly, when renal cytosolic proteins from rats given TFEC (Figure 3A) were proteolyzed, a significant reduction in line width was observed (Figure 3B). Control experiments were conducted to examine the role of proteinase K in the observed decreases in NMR line width. First, when renal cytosol from TFEC-treated rats was incubated in the absence of proteinase K, the change in the adduct NMR spectrum was far less pronounced (Figure 3C) than in the presence of proteinase K (Figure 3B). Second, acylation of poly(DL-lysine) with 2-chloro1,1,2-trifluoroethyl2-nitrophenyldisulfide gave a broadened adduct resonance (data not shown), whose chemical shift was identical with that seen with renal protein from CTFC-treated rats (Figure 2B). The line width of the adduct resonance was not altered by incubation with proteinase K (data not shown), presumably because the recognition sequences required for proteinase K catalyzed peptide hydrolysis (28) are not present in poly(DL-lysine). Because proteolysis converts the broadened adduct resonances to single signals with clearly resolved, characteristic geminal H-F coupling, these data indicate that (1) the broadened NMR resonances found in spectra of renal

38 Chem. Res. Toxicol., Vol. 5, No. 1, 1992

tissue from rata given CTFC and TFEC represent fluorinated metabolites covalently attached to protein and that (2) a single amino acid is stably acylated as a result of CTFC and TFEC metabolism in vivo. The doublets revealed by proteolysis of CTFC- or TFEC-modifiedproteins were identical in chemical shift and coupling constant with synthetic (chlorofluorothioacetamido)ethane or (difluorothioacetamido)ethane, respectively (Figure 1, Table I). Thus the single, stable amino acid adduct formed with renal proteins of rats given CTFC or TFEC was Ne(chlorofluorothioacety1)lysine(6d) and N-(difluorothioacety1)lysine (60, respectively (Figure 1). As a result of the processing or storage of protein fractions, a resonance was sometimes observed at approximately 69 ppm (FIgure 2C,D) or near 52 ppm (Figure 3C) in the NMR spectra of CTFC- or TFEC-modified protein fractions, respectively. After incubation of protein fractions with proteinase K, the chemical shifts of these new resonances were identical with those of (chlorofluoroacetamidolethane and (difluoroacetamid0)ethane (see Experimental Procedures). Because thioamides can be converted to amides through a thioamide S-oxide intermediate by treatment with hydrogen peroxide (29),the thioamide protein adducts produced by CTFC and TFEC metabolism (this work) may also be converted to amides by hydrogen peroxide. This was studied by incubating synthetic (ch1orofluorothioacetamido)ethane with 1% hydrogen peroxide (pH 7.4; 37 "C) and following the course of the reaction. The conversion from thioamide to amide was efficient under these conditions (data not shown). In Vitro Acylation of Amino Acids. To test the in vivo observation that the E-amino group of lysine was the only target for stable acylation by metabolites of CTFC and TFEC, in vitro incubations of 2-chloro-1,1,2-trifluoroethyl2-nitrophenyl disulfide in the presence or absence of Na-acetyl amino acids were conducted. The spectrum in Figure 5A shows the products resulting from the reaction of 2-chloro-l,l,2-trifluoroethyl2-nitrophenyl disulfide with 5 molar equiv of N"-acetyl-L-lysine in aqueous buffer. In addition to the formation of inorganic fluoride and chlorofluoroacetic acid, a new doublet of J H F = 52 Hz at 56.6 ppm was formed (n = 6). The 19FNMR chemical shift and coupling constant of this resonance were identical with that of (ch1orofluorothioacetamido)ethane (Table I). Furthermore, the chemical shift and coupling constant of the small doublet at 69.9 ppm were identical with those of (ch1orofluoroacetamido)ethane(see Experimental Procedures). In the absence of added nucleophiles, hydrolysis of 2chloro-1,1,2-trifluoroethyl2-nitrophenyl disulfide gave inorganic fluoride and chlorofluoroacetic acid as stable producta (Figure 5B). Additionally, Figure 5B shows the spectral pattern of 2-chloro-1,1,2-trifluoroethyl2-nitrophenyl disulfide before hydrolysis (multiplets near 12 and 72 ppm). Only inorganic fluoride and chlorofluoroacetic acid were observed as products when 2-chloro-1,1,2-trifluoroethyl 2-nitrophenyl disulfide was incubated with 5 molar equiv of "-blocked L-tyrosine, L-glutamine, L-asparagine, Lserine, L-arginine, L-methionine, or L-cysteine (n1 3) (data not shown). Attempts to detect transient products of acylation by recording NMR spectra both during and after the completion of 2-chloro-1,1,2-trifluoroethyl2-nitrophenyl disulfide hydrolysis were unsuccessful in the presence of all amino acids named above. Incubation of 2-chloro-1,1,2-trifluoroethy12-nitrophenyl disulfide with 5 molar equiv of Ne-acetyl-L-histidine or imidazole ( R 2 3) produced new, transient products.

Harris et al.

C

D

10

20

30

40

60

SO

70

80

90

P PU

Figure 5. 19FNMR spectra of reaction mixture containing 2chloro-l,l,2-trifluoroethyl2-nitrophenyl disulfide and various

amino acid-based nucleophiles. (A) Incubation in the presence (B)Incubation in the absence of added nucleophiles. Identical results were obtained in the presence of N"-blocked L-tyrosine, L-glutamine, L-asparagine, L-serine, Larginine, L-methionine, and L-cysteine (data not shown). The multiplets near 12 and 72 ppm are assigned to 2-chloro-1,1,2trifluoroethyl2-nitrophenyldisulfide (2). (C)Incubation in the presence of N"-acetyl-Lhistidine. The resonance centered at 98.6 ppm is enlarged in the figure inset. (D)Sample in panel C after incubation at 40 "C overnight. Inorganic fluoride (6 = 44.4 ppm) and chlorofluoroacetic acid (5a; b = 61.3 ppm) were found in all experiments (see Figure 1). of Ne-acetyl-L-lysine.

Figure 5C shows a new product resonance centered at 98.6 ppm, which was lost by incubating the mixture overnight at 40 "C (Figure 5D).The spectral pattern observed, which consisted of two closely spaced doublets, each with Jm= 71 Hz (inset, Figure 5C), is consistent with single, equal chlorofluorothioacylation of each of two nonequivalent nitrogens of the imidazole ring of histidine. Consistent with this hypothesis is the finding that incubation of 2chloro-l,l,2-trifluoroethyl2-nitrophenyl disulfide with imidazole yielded only a single doublet, J H F = 71 Hz, at 98.6 ppm (data not shown). This is the expected result because the imidazole nitrogens are equivalent. However, mass spectral evidence for an N-(chlorofluorothioacety1)imidazole adduct could not be obtained. Furthermore, addition of primary amines [i.e., N*-acetyl-L-lysine or poly(DL-lysine)] to reactions containing the unknown, histidine-associated resonance at 98.6 ppm did not result in detectable fluorinated products other than inorganic fluoride and chlorofluoroacetic acid. Urinary Metabolites of CTFC and TFEC. Mercapturic acids, dihaloacetic acids, and inorganic fluoride were identified in the urine of rats given CTFC and TFEC. The presence of chlorofluoroacetic acid as a metabolite of CTFC and of difluoroacetic acid as a metabolite of TFEC was confirmed by NMR coresonance with authentic samples and by GC-MS analysis (data not shown). Urinary fluoride excretion in treated, but not control, animals was detected by NMR (data not shown). Mercapturic acid excretion was indicated by the NMR spectra of urine from

Bioactivation of Haloethene Cysteine S-Conjugates

rats treated with CTFC and TFEC; however, the '?E' NMR spectra of these cysteine S-conjugates and their corresponding mercapturic acids did not differ substantially (data not shown). In order to confirm urinary mercapturic acid excretion, urine samples were extracted, methylated with diazomethane, and analyzed by GC-MS. Mass spectra with characteristic fragment ions and retention times identical with those of authentic N-acetyl-S-(2chloro-1,1,2-trifluoroethyl)-~-cysteine and N-acetyl-S(1,1,2,2-tetrafluoroethyl)-~-cysteine were obtained, thus confirming the formation and urinary excretion of the mercapturic acids of CTFC and TFEC.

Discussion The bioactivation mechanism previously elucidated for CTFC (9) demonstrated the protential for acylation of cellular nucleophiles (Figure 1). This potential was confirmed in the in vivo experiments described herein: the 19FNMR spectra of renal subcellular fractions from rats given CTFC (Figure 1; la) or TFEC (lb) contained a broadened signal of low itensity in addition to the sharp resonances due to free metabolites present in the tissue. Because reduced molecular tumbling (30) or changes in chemical environment, or both, lead to broadened signals in NMR spectra, the observation of broad resonances in protein fractions from S-conjugate-treated rats suggested that these resonances were due to CTFC and TFEC metabolites that became covalently bound to proteins. Dialysis of macromoleculeswith a SDS-containing buffer removes weakly bound xenobiotic metabolites, but covalently bound metabolites are not removed (23). The effectiveness of dialysis in the presence of SDS in removing fluorine-containing CTFC and TFEC metabolites from protein fractions of treated animals was monitored by comparing '9F NMR spectra both before and after dialysis. A single resonance remained in the NMR spectra of dialyzed renal proteins isolated from rats given CTFC (la) or TFEC (lb). These data indicate that the broad resonances observed in the spectra of CTFC- or TFEC-modified proteins are due to covalent adduct formation. A similar approach has been used to demonstrate tissue acylation by a metabolite of halothane and of the chlorofluorocarbon substitute 2,2-dichloro-l,l,l-trifluoroethane (31). The NMR resonances of CTFC- and TFEC-modified protein fractions (Figures 2 and 3, respectively) are significantly broadened and thus may contain resonances resulting from adduct formation with one or more amino acid side chains in protein. To identify the amino acids that are targets for acylation, proteolysis of dialyzed protein fractions from rats treated with CTFC (la) and TFEC (lb) was used to decrease the molecular size of the adducted proteins and thus decrease the observed line width in the NMR spectra. In proteolyzed samples, the broadened 19F NMR resonances associated with TFEC- and CTFC-modified proteins were converted to single, wellresolved doublets (Figures 3 and 4,respectively). Control experiments confirmed the role of proteolysis in the reduction of observed NMR line width: in the absence of added proteinase, the degree of spectral change was reduced (Figure 3, compare panels B and C), whereas the spectrum of CTFC-modified poly(m-lysine) was not changed by proteinase K treatment (data not shown). Upfield movements of the adduct chemical shift were observed during proteolysis of CTFC- and TFEC-modified proteins. For example, the midpeak chemical shift of renal proteins from TFEC-treated rats was 40.9 ppm before proteolysis (Figure 3A) and 41.5 ppm after incubation with

Chem. Res. Toxicol., Vol. 5, No. 1, 1992 39

proteinase K (Figure 3B). This phenomenon is apparently another NMR spectral indicator of increased molecular motion, along with the observed decrease in line width. Similar upfield movements in chemical shifts are seen in the NMR spectra of mixtures of phosphoglycerate kinase containing 5-fluorotryptophan residues and free 5-fluorotryptophan (32) and in the NMR spectra of intercalating agents in the presence or absence of nucleic acids (33). The well-resolved doublets that result from proteolysis of CTFC- and TFEC-modified protein fractions were identical in chemical shift and coupling constant to those for synthetic (ch1orofluorothioacetamido)ethane (Figure 1,6c) and (difluorothioacetamido)ethane(6e), respectively (Table I). These derivatives of ethylamine represent chemical-shift standards for Nf-(chlorofluorothioacety1)L-lysine (6d) and Ne-(difluorothioacety1)-L-lysine(6f). Derivatives 6c and 6e model the chemical shifts of lysine adducts 6d and 6f, respectively, because the side chain of lysine is an alkylamine and because the fluorine atoms monitored in these studies are separated by five bonds from the atoms at which they differ structurally. Reaction of 2-chloro-1,1,2-trifluoroethyl2-nitrophenyl disulfide (2) with N"-acetyl amino acids showed that stable products were formed only with lysine (Figure 5A); the chemical shift and coupling constant of this product were identical with those of (chlorofluorothioacetamido)ethane (6c; see Table I). This observation, combined with the in vivo data described above, indicates that primary amines, particularly the e-amino group of protein-incorporated lysine, are the sole target of stable protein acylation due to CFTC and TFEC bioactivation; these data do not exclude adducts with N-terminal amino groups of proteins, although such adducts may be comparatively few in number. The spectroscopic evidence described herein indicates that the adducts formed in vivo are Nf-(chlorofluorothioacety1)lysine (6d) and Ne-(difluorothioacety1)lysine(60, respectively. The identification of N'-(difluorothioacety1)lysine as a modified amino acid agrees with the recent report of Hayden et al. (161, who found the same adduct when TFEC was incubated with Nu-acetyl-L-lysine and a pyridoxal 5'-phosphate and copper(I1) catalysis system. Additionally, incubation of [%]TFEC with bovine serum albumin and /3-lyase or with isolated rat kidney proximal tubules gave similar results (16). Aa indicated above, Ne-(dihalothioacety1)lysineswere the only stable protein adducts formed in CTFC- or TFECtreated rats. Similarly, when the S-conjugate proreactive intermediate 2-chloro-1,1,2-trifluoroethyl 2-nitrophenyl disulfide (2) was incubated with a range of amino acids, the analogous adduct was formed. A transient product was, however, formed when 2-chloro-1,1,2-trifluoroethyl 2-nitrophenyl disulfide was incubated with histidine and imidazole. The I9FNMR resonances of these products are consistent with thioacylation of imidazole nitrogens. Addition of target amines [N"-acetyl-L-lysine or PO~Y(DLlysine)] to reaction mixtures that contained the putative imidazole adduct did not result in thioacylation of the added nucleophiles. Under the conditions of the present experiments, no evidence for nucleophilic catalysis, as reported by Hayden et al. (16), was found; it should be noted that the experimental conditions employed in the present studies and by Hayden et al. (16) were different. Urinary metabolites of CTFC and TFEC found in these studies were inorganic fluoride, the mercapturic acid of each S-conjugate, and chlorofluoroacetic acid or difluoroacetic acid. The metabolites reported here for TFEC are identical with those found by Commandeur et al. (17)in the urine of rats given the mercapturic acid of TFEC.

40 Chem. Res. Toxicol., Vol. 5, No. 1, 1992

The bioactivation scheme elucidated for CTFC (la) is shown in Figure 1 (9). a-Fluoroalkyl thiolates, such as 2-chloro-l,1,2-trifluoroethanethiolate(3a), eliminate an a-fluorine with ease (34). The chlorofluorothioacetyl fluoride (4a) thus formed may react with nucleophilic amines, such as diethylamine ( 9 , 3 5 ) . Chlorofluorothioacetyl fluoride may also react with water to produce the observed stable end products chlorofluoroaceticacid (5a), inorganic fluoride, and hydrogen sulfide (36). The work reported here demonstrates that TFEC (lb) bioactivation occurs in an analogous manner (Figure 1). 2-Chloro-l,l,2-trifluoroethyl2-nitrophenyl disulfide (2) is designed to release the same thiolate 3a as is released by the enzymatic cleavage of CTFC (la) by &lyase (Figure 1). Because identical covalent adducts with protein result (compare panels B and C of Figure 2) and because inorganic fluoride, chlorofluoroacetic acid (5a), and chlorofluorothioacetamidesare formed in the presence of amines (Figure 5A), reaction of 2-chloro-1,1,2-trifluoroethyl2nitrophenyl disulfide (2) provides a valid model for the metabolism of CTFC. Further, the identification of 3fluoro-3-(chlorofluoromethyl)-2-thiabicyclo[2.2.1] hept-5ene in reaction mixtures containing 2-chloro-1,1,2-trifluoroethyl2-nitrophenyldisulfide (2) and cyclopentadiene (12) demonstrates the formation of chlorofluorothioacetyl 2-nitrofluoride (4a) from 2-chloro-1,1,2-trifluoroethyl phenyl disulfide. The observation that CTFC- and TFEC-modified proteins were observed in all subcellular fractions studied (mitochondria, microsomes, and cytosol) indicates that haloethanethiolates (3a and 3b) or thioacetyl fluorides (4a and 4b) released by metabolism of CTFC and TFEC may cross membranes separating organelles or that cysteine S-conjugates are metabolized in each of the three cellular compartments studied. @-Lyaseactivity has been found only in cytosol and mitochondria of rat kidney, but is present in human kidney microsomes (37). Although covalent modification of renal mitochondrial, microsomal, and cytosolic proteins by the dihalothioacetyl fluorides produced by metabolism of TFEC and CTFC was detected in the present study, quantification of adduct formation is difficult due to the inherent insensitivity of NMR. Experiments with radiolabeled CTFC and TFEC may allow quantification of acylation in these cellular compartments and would allow the question of whether compartmental acylation coincides with compartmental P-lyase activity to be addressed. The observation of renal protein alkylation resulting from the treatment of animals with known nephrotoxins provides evidence for a contributing role of protein alkylation in S-conjugate-induced cytotoxicity. A role for protein alkylation in toxicity is supported by the requirement for metabolism of cysteine S-conjugates in order to produce toxicity: the a-methyl derivatives of several conjugates are not toxic in vitro (3) or in vivo (38). These analogues no longer posses the a proton that must be abstracted for @lyase to complete its catalytic cycle. Furthermore, administration of (amino0xy)acetic acid, which inhibits pyridoxal 5’-phosphate-dependent enzymes such as P-lyase, blocks the nephrotoxicity of haloalkyl S-conjugates in vivo (1 7,38). Alkylation of cellular proteins by xenobiotics or their metabolites is linked to the toxicities of many compounds (reviews: ref 39-41). The covalent interaction of S-conjugate-derived thiols with protein thiols to form mixed disulfides has been suggested as a mechanism for cysteine S-conjugate nephrotoxicity (42). The demonstration herein that stable adducts are formed in vivo only with lysine and the failure to detect transient

Harris et al.

intermediates formed with the thiol of cysteine in vitro argue against an important role for mixed disulfide, thioester, or dithioester formation with cellular thiols in Sconjugate-induced toxicity. The protein adducts resulting from CTFC and TFEC metabolism are similar in structure, differing by a single chlorine or fluorine (Figure 1, compare 6d and 60. The observation that the adduct resonances for each chemical differ by 15 ppm (compare Figures 2B and 3A) demonstrates the ability to discern fine structural differences by 19FNMR. This methodology may prove useful in characterizing cellular targets of reactive intermediates in vivo for several classes of compounds. Acknowledgment. This work was supported by National Institutes of Environmental Health Sciences Granta ES03127 and RR03829 (M.W.A.) and ES07026 (J.W.H.) and by Deutsche Forschungsgemeinschaft Sonderforschungsbereich 172 (W.D.). We thank S. E. Morgan for assistance in the preparation of the figures.

References (1) Harris, J. W., and Anders, M. W. (1990)The use of 19FNMR in the study of protein alkylation by fluorinated reactive intermediates. Adv. Exp. Med. Biol. 283,735-738. (2) Anders, M. W. (1988)Glutathione-dependent toxicity: Biosynthesis and bioactivation of cytotoxic S-conjugates. ISZ Atlas Sci.: Pharmacol. 2,99-104. (3) Dohn, D.R., Leininger, J. R., Lash, L. H., Quebbemann, A. J., and Anders, M. W. (1985)Neprotoxicity of S-(2-chloro-1,1,2-trifluoroethy1)glutathione and S-(2-chloro-l,l,2-trifluoroethyl)-Lcysteine, the glutathione and cysteine conjugates of chlorotrifluoroethene. J. Pharmacol. Exp. Ther. 235,851-857. (4) Harp,S.J.,Fitzsimmons, M. E., Aniya, Y., and Anders, M. W. (1991) Stereochemistry of the microsomal glutathione Stransferase catalyzed addition of glutathione to chlorotrifluoroethene. Biochemistry 30, 717-721. (5) Odum, J., and Green, T. (1984)The metabolism and nephrotoxicity of tetrafluoroethylene in the rat. Toxicol. Appl. Pharmacol. 76,306-318. (6) Dekant, W., Vamvakas, S., and Anders, M. W. (1989)Bioactivation of nephrotoxic haloalkenes by glutathione conjugation: Formation of toxic and mutagenic intermediates by cysteine conjugate @-lyase. Drug Metab. Reu. 20,43-83. (7) Stevens, J. L., Robbins, J. D., and Byrd R. A. (1986)A purified cysteine conjugate @-lyasefrom rat kidney cytosol. J. Bioi. Chem. 261. 15529-15537. (8) Cooper, A. J. L., and Anders, M. W. (1990)Glutamine transaminase K and cysteine conjugate @-lyase.Ann. N.Y. Acad. Sci. 585, 118-127. (9) Dekant, W., Lash, L. H., and Anders M. W. (1987)Bioactivation mechanism of the cytotoxic and nephrotoxic S-conjugate S42chloro-l,l,2-trifluoroethyl)-L-cysteine.Proc. Natl. Acad. Sci. U.S.A. 84,7443-7447. (IO) Dekant, W., Berthold, K., Vamvakas, S., Henschler, D., and Anders, M. W. (1988)Thioacylating intermediates aa metabolites and S-(1,2,2-trichlorovinyl)-Lof S-(1,2-dichlorovinyl)-L-cysteine cysteine formed by cysteine conjugate @-lyase.Chem. Res. Toxicol. 1, 175-178. (11) Dekant, W., Berthold, K., Vamvakas, S., and Henschler, D. (1988)Thioacylating agents as ultimate intermediates in the @lyase catalysed metabolism of S-(pentachlorobutadieny1)-L-cysteine. Chem.-Biol. Interact. 67,139-148. (12) Dekant, W., Urban, G., Gorsmann, C., and Anders, M. W. (1991)Thioketene formation from a-haloalkenyl 2-nitrophenyl disulfides: Models for biological reactive intermediates of cytotoxic S-conjugates. J. Am. Chem. SOC. 113,5120-5122. (13) Darnerud, P.O.,Brandt, I., Feil, V. J., and Bakke, J. E. (1988) S-(1,2-Di~hloro-[~~C]vinyl)-L-cysteine (DCVC) in the mouse kidney: Correlation between tissue-binding and toxicity. Toxicol. Appl. Pharmacol. 95,423-434. (14) Vamvakas, S.,Dekant, W., Berthold, K., Schmidt, S., Wild, D., and Henschler, D. (1987)Enzymatic transformation of mercap turic acids derived from halogenated alkenes to reactive and mutagenic intermediates. Biochem. Pharmacol. 36,2741-2748. (15) Hayden, P.J., and Stevens, J. L. (1990)Cysteine conjugate toxicity, metabolism, and binding to macromolecules in isolated

Bioactioation of Haloethene Cysteine S-Conjugates rat kidney mitochondria. Mol. Pharmacol. 37, 468-476. (16) Hayden, P. J., Yang, Y., Ward, A. J. I., Dulik, D. M., McCann, D. J., and Stevens, J. L. (1991) Formation of difluorothionoacetyl-protein adducts by S-(1,1,2,2-tetrafluoroethyl)-L-cysteine metabolites: Nucleophilic catalysis of stable lysyl adduct formation by histidine and tyrosine. Biochemistry 30, 5935-5943. (17) Commandeur, J. N. M., Brakenhoff, J. P. G., De Kanter, F. J. J., and Vermeulen, N. P. E. (1988) Nephrotoxicity of mercapturic acids of three structurally related 2,2-difluoroethylenes in the rat. Biochem. Pharmacol. 37,4495-4504. (18) Boogaard, P. J., Commandeur, J. N. M., Mulder, G. J., Vermeulen, N. P. E., and Nagelkerke, J. F. (1989) Toxicity of the cysteine S-conjugates and mercapturic acids of four structurally related difluoroethylenes in isolated proximal tubular cells from rat kidney. Biochem. Pharmacol. 38,3731-3741. (19) Middleton, W. J. (1979) One-step method for converting esters to acyl chlorides. J. Org. Chem. 44, 2291-2292. (20) Englund, B. (1963) 2-Chloro-1,1,2-trifluoroethyl ethyl ether. Organic Syntheses, Collect. Vol. 4, pp 184-186, Wiley, New York. (21) Englund, B. (1963) Ethyl chlorofluoroacetate. Organic Syntheses, Collect. Vol. 4, pp 423-424, Wiley, New York. (22) Johnson, D., and Lardy, H. (1967) Isolation of liver or kidney mitochondria. Methods Enzymol. 10,94-96. (23) Sun,J. D., and Dent, J. G. (1980) A new method for measuring covalent binding of chemicals to cellular macromolecules. Chem.-Biol. Interact. 32, 41-61. (24) Read, S. M., and Northcote, D. H. (1981) Minimization of variation in the response to different proteins of the Coomassie blue G dye-binding assay for protein. Anal. Biochem. 116,5344. (25) Karashima, D., Shigematsu, A., Furukawa, T., Nagayoshi, T., and Mataumoto, I. (1977) Esterification of trifluoroaceticacid with phenyldiazomethane for quantitative gas chromatographic analysis: Methods involving separation from biological materials. J. Chromatogr. 130, 77-86. (26) Vetter, W. (1972) Amino acids. In Biochemical Applications of Mass Spectroscopy (Waller, G. R., Ed.) pp 387-399, Wiley-Interscience, New York. (27) Ghther, H. (1987) NMR Spectroscopy, pp 343-344, Wiley, New York. (28) Ebeling, W., Hennrich, N., Klockow, M., Metz, H., Orth, H. D., and Lang, H. (1974) Proteinase K from Tritirachium album limber. Eur. J. Biochem. 47, 91-97. (29) Walter, W., and Bauer, 0. H. (1976) On the oxidation products of carbothioamides, XXXIV-A simple method for the preparation of tertiary thioamide S-oxides. Liebigs Ann. Chem., 1584-1597.

Chem. Res. Toxicol., Vol. 5, No. 1, 1992 41 (30) Gadian, D. G. (1982) Nuclear Magnetic Resonance and Its Applications to Living Systems, pp 40-41, Oxford University Press, Oxford. (31) Harris, J. W., Pohl, L. R., Martin, J. L., and Anders, M. W. (1991) Tissue acylation by the chlorofluorocarbon substitute 2,2,-dichlorel,l,l-trifluoroethane.Proc. Natl. Acad. Sci. U.S.A. 88, 1407-1410. (32) Brindle, K., Williams, S. P., and Boulton, M. (1989) l9FNMR detection of a fluorine-labelled enzyme in vivo. FEBS Lett. 255, 121-124. (33) Mirau, P. A., Shafer, R. H., James, T. L., and Bolton, P. H. (1982) Fluoroquinacrine binding to nucleic acids: Investigation of the l9F-NMR, optical, and fluorescence properties in the presence of DNA, poly(A), and tRNA. Biopolymers 21,909-921. (34) Middleton, W. J. (1966) Organic sulfwfluorine compounds. In Organic Chemistry of Bivalent Sulfur (Reid, E. E., Ed.) Vol. 6, pp 331-412, Chemical Publishing Co., New York. (35) Fokin, A. V., Skladnev, A., and Knunyants, I. (1961) Reactions of fluoro olefins: Reaction of fluoro olefins with hydrogen sflide. Dokl. Akad. Nauk SSSR 138,1132-1135. (36) Banki, K., Elfarra, A. A,, Lash, L. H., and Anders, M. W. (1986) to hyMetabolism of S-(2-chloro-1,1,2-trifluoroethyl)-L-cysteine drogen sulfide and the role of hydrogen sulfide in the S42chloro-1,1,2-trifluoroethyl)-L-cysteine-induced mitochondrial toxicity. Biochem. Biophys. Res. Commun. 138, 707-713. (37) Lash, L. H., Nelson, R. M., Van Dyke, R. A., and Anders, M. W. (1990) Purification and characterization of human kidney cytosolic cysteine conjugate &lyase activity. Drug Metab. Dispos. 18, 50-54. (38) Elfarra, A. A., Jakobson, I., and Anders, M. W. (1986) Mechanism of S-(l,2-dichlorovinyl)glutathione-inducednephrotoxicity. Biochem. Pharmacol. 35, 283-288. (39) Monks, T. J., and Lau, S. S. (1988) Reactive intermediates and their toxicological significance. Toxicology 52, 1-53. (40) Nelson, S. D., and Pearson, P. G. (1990) Covalent and noncovalent interactions in acute lethal cell injury caused by chemicals. Annu. Rev. Pharmacol. Toxicol. 30, 169-195. (41) Smith, C. V., Lauterberg, B. H., and Mitchell, J. R. (1985) Covalent binding and acute lethal injury in vivo: How has the original hypothesis survived a decade of critical examination? In Drug Metabolism and Disposition: Considerations in Clinical Pharmacology (Wilkinson, G. R., and Rawlins, D. M., Eds.) pp 161-181, MTP Press, Lancaster. (42) Green, T., and Odum, J. (1985) Stucture/activity studies of the nephrotoxic and mutagenic action of cysteine conjugates of chlore and fluoroalkenes. Chem.-Biol. Interact. 54, 15-31.