Biochi Mic~A Et Biophysica A~Ta

BB Biochi ~mic~a et BiophysicaA~ta ELSEVIER Biochimica et Biophysica Acta 1271 (1995) 51-57

Mitochondrial bioactivation of cysteine S-conjugates and 4-thiaalkanoates: Implications for mitochondrial dysfunction and mitochondrial diseases

M.W. Anders * Department of Pharmacology, Unit,ersity of Rochester, 601 Elmwood Avenue, Boa- 711, Rochester. New York 14642 USA

Abstract

The toxicity of most drugs and chemicals is associated with their enzymatic conversion to toxic metabolites. Bioactivation reactions occur in a range of organs and organelles, including mitochondria. The toxicity of haloalkene-derived cysteine S-conjugates and related 4-thiaalkanoates is associated with their mitochondrial bioactivation. Toxic cysteine S-conjugates are formed by the glutathione S-transferase-catalyzed addition of glutathione to haloalkenes to give glutathione S-conjugates, which are hydrolyzed by y-glutamyltransferase and dipeptidases. Mitochondrial cysteine conjugate /3-1yase-catalyzed bioactivation of cysteine S-conjugates affords unstable a-halothiolates. Haloalkene-derived 4-thiaalkanoates, which are analogs of cysteine S-conjugates that lack an a-amino group, undergo bioactivation by the enzymes of fatty acid /3-oxidation to give 3-hydroxy-4-thiaalkanoates that eliminate a-halothiolates, a-Halothiolates yield alkylating and acylating agents that interact with cellular macromolecules and thereby cause cell damage. Mitochondrial dysfunction is the hallmark of cysteine S-conjugate-induced cytotoxicity: decreased respiration, decreased ATP and total adenine nucleotide concentrations, depletion of the mitochondrial glutathione content, perturbations in cellular Ca 2+ homeostasis, and damage to the mitochondrial genome are seen with cysteine S-conjugates. Similar changes are observed with cytotoxic 4-thiaalkanoates, but inhibition of the medium-chain acyl-CoA dehydrogenase and hypoglycemia are also observed.

Keywords: Mitochondrion; Bioactivation; Cysteine S-conjugate; 4-Thiaalkanoate; Toxicity

1. Introduction transferases [3,4], and some carcinogens are bioactivated by cytosolic sulfotransferases [5,6]. The mitochondrion is Many synthetic and naturally occurring chemicals in- also a site of bioactivation reactions: astrocytic monoamine duce mitochondrial dysfunction, and some of these chemi- oxidases, which are located in the mitochondrial outer cals have become established as tools in exploring mito- membrane, catalyze the oxidation of 1-methyl-4-phenyl- chondrial biochemistry. The concept that environmental 1,2,3,6-tetrahydropyridine (MPTP) to 1-methyl-4-phenyl- contaminants may also induce mitochondrial dysfunction 2,3-dihydropyridinium (MDTP) [7], which undergoes fur- has received little attention. Although the bioactivation of ther oxidation to 1-methyl-4-phenylpyridinium (MPP+); xenobiotics is most often associated with the cytochromes MPP + is taken up by the catecholamine transporter of P-450 located in the hepatic endoplasmic reticulum [1], dopaminergic neurons [8] and inhibits Complex I of the bioactivation reactions occur in several organs and or- mitochondrial electron transport chain [9]. ganelles. For example, the mutagenicity of vicinal di- This review will focus on the mitochondrial bioactiva- haloalkanes is associated with hepatic cytosolic glutathione tion and toxicity of cysteine S-conjugates and related S-transferase-catalyzed bioactivation [2], the formation of 4-thiaalkanoates. Haloalkene-derived cysteine S-conjugates reactive acyl glucuronides from nonsteroidal antiinflamma- are potent and selective nephrotoxins in vivo and are tory agents is catalyzed by the microsomal glucuronyl cytotoxic in isolated rat renal proximal tubular cells and in cultured cell lines of renal origin (LLC-PK1). 4-Thiaal- kanoates are analogs of cysteine S-conjugates that lack the a-amino group and are nephrotoxic and hepatotoxic in * Corresponding author. Fax: + l 716 2449283. vivo and cytotoxic in isolated rat hepatocytes.

2. Mitochondrial bioactivation of cysteine S-conjugates L-cysteinylglycine S-conjugates, which are hydrolyzed to the corresponding cysteine S-conjugates by aminopepti- The key role of mitochondria in xenobiotic bioactiva- dase M or cysteinylglycine dipeptidase [15]. tion was elucidated during studies on the selective nephro- The bioactivation of haloalkene-derived cysteine S-con- toxicity of haloalkenes. A multistep, multiorgan bioactiva- jugates is catalyzed by cysteine conjugate /3-1yase, which tion mechanism that explains the selective nephrotoxicity is identical with glutamine transaminase K [16]. Cysteine and, perhaps, nephrocarcinogenicity of a range of conjugate /3-1yase is a pyridoxal phosphate-dependent en- haloalkenes was subsequently elucidated. The bioactiva- zyme, which is present in both the cytosol and mito- tion of nephrotoxic haloalkenes involves hepatic glu- chondria of renal proximal tubular cells, that catalyzes tathione S-conjugate formation, hydrolysis of the glu- both transamination and /3-elimination reactions of cys- tathione S-conjugates to cysteine S-conjugates, uptake of teine S-conjugates. The transamination of cysteine S-con- the S-conjugates by the kidney via amino acid and organic jugates gives a-keto acids as products and converts the anion transport systems, and bioactivation by renal mito- enzyme-bound pyridoxal phosphate to pyridoxamine phos- chondrial cysteine conjugate/3-1yase. Several reviews about phate, which is not competent to catalyze /3-elimination the glutathione- and /3-1yase-dependent bioactivation of reactions, a-Keto acids stimulate the toxicity of cysteine haloalkenes have appeared [ 10-15]. S-conjugates, presumably by maintaining the enzyme in the pyridoxal phosphate form [ 17,18]. 2.1. Bioactivation mechanism The toxicity of cysteine S-conjugates is associated with a /3-1yase-catalyzed t-elimination reaction. The critical The microsomal glutathione S-transferases are selective role of the fl-lyase in the bioactivation of cysteine S-con- catalysts for the biotransformation of haloalkenes to glu- jugates has been demonstrated by use of selective in- tathione S-conjugates (for a review, see [15]). 1,1-Dichlo- hibitors and by nonmetabolizable analogs of cysteine S- roalkenes (Fig. 1, la) undergo addition-elimination reac- conjugates. Aminooxyacetic acid inhibits a range of pyri- tions to give S-(1-chloroalkenyl)glutathione conjugates, doxal phosphate-dependent enzymes, including cysteine which are hydrolyzed to the corresponding S-(l-chloroal- conjugate fl-lyase [19], and blocks the nephrotoxicity and kenyl)-L-cysteine conjugates (Fig. 1, 2a), and 1,1-dif- cytotoxicity of S-(l,2-dichlorovinyl)-L-cysteine (DCVC) luoroalkenes (Fig. 1, lb,c) undergo addition reactions to [19,20]. Cysteine conjugates blocked with an a-methyl give S-(1,1-difluoroalkyl)glutathione conjugates, which are group cannot undergo fl-lyase-catalyzed /3-elimination re- hydrolyzed to S-(1,1-difluoroalkyl)-L-cysteine conjugates actions, and the a-methyl analogs of DCVC, S-(2-chloro- (Fig. 1, 2b,e). The haloalkene-derived glutathione S-con- 1,1,2-trifluoroethyl)-L-cysteine and S-(2-bromo-2-chloro- jugates are excreted in the bile and appear in the portal 1,1-difluoroethyl)-L-cysteine are not toxic. circulation as the intact glutathione S-conjugates or as the fl-Lyase-catalyzed /3-elimination reactions of cysteine corresponding cysteine S-conjugates. The y-glutamyltrans- S-conjugates afford unstable thiolates as initial products. ferase-catalyzed hydrolysis of glutathione conjugates yields With 1,1-dichloroalkene-derived cysteine S-conjugates

Precursor Reactive Terminal haloalkene Cvsteine S-coniuaate Lvase oroduct intermediate product

X X x- 4oo . S X X (~H3 X X la 2,, 3e 4e 5a

X X X NH3 X X I b 2b 3b 4b 5b

Br NH3 X O I c 2c 3c 4c 5c Fig. 1. Nephrotoxic haloalkenes (la-e), corresponding cysteine S-conjugates (2a-c), products formed by cysteine conjugate /3-1yase (lyase products, 3a-c), reactive intermediates (4a-e)formed from lyase products, and terminal products (5a-c) formed from reactive intermediates. M. W. Anders / Biochimica et Biophysica Acta 1271 (1995) 51-57 53

(Fig. 1, 2a), l-chloroalkenylthiolates (Fig. 1, 3a) are pro- give covalently bound adducts [25-27]. 19F NMR studies duced. 1-Chloroalkenylthiolates may tautornerize to give showed that inorganic fluoride and chlorofluoroacetic acid thioacyl chlorides or may eliminate chloride to give are terminal products of the /3-1yase-catalyzed biotrans- thioketenes (Fig. 1, 4a). To explore these possibilities, formation of S-(2-chloro-l,l,2-trifluoroethyl)-L-cysteine 1,2-dichloro-3,3,3-trifluoro-1-propenyl 2-nitrophenyi disul- [24]. fide was prepared as a precursor of 1,2-dichioro-3,3,3-tri- Bromine-containing, l,l-difluoro-2,2-dihaloethene-de- fluoro-l-propenethiolate [21]. Reaction of the disulfide rived cysteine S-conjugates (Fig. 1, 2c) give 2-bromo-2- with DABCO and cyclopentadiene in tetrahydrofuran gave halo-l,l-difluoroethanethiolates (Fig. 1, 3e) as initial prod- the thianorbornene 3-(2,2,2-trifluoro-l-chloroethylidene)- ucts. The fate of the bromine-containing 1,1-dif- 2-thiabicyclo[2.2.1]hept-5-ene as a product, indicating that luoroethanethiolates differs from that of the bromine-lack- a thioketene had formed and undergone a Diels-Alder ing analogs: glyoxylate (Fig. 1, 4e), rather than dihaloac- reaction with cyclopentadiene. In the absence of cyclopen- etates, is formed as a terminal product [28]. Preliminary tadiene, syn- and anti-2,4-bis(2,2,2-trifluoro-1-chloroethyl- reaction mechanism studies indicate that o~-thiolactones idene)-l,3-dithietane were formed, indicating dimerization may be formed as intermediates during the biotransforma- of the thioketene. Several 1-chloroalkenyl 2-nitrophenyl tion of bromine-containing cysteine S-conjugates (M.B. disulfides have been prepared, and all are mutagenic in in Finkelstein, W. Dekant, A.W. Kende, and M.W. Anders, vitro test systems [22]. Thioketenes may undergo hydroly- unpublished data). sis to give haloacetic acids (Fig. 1, 5a) as terminal products [23]. 2.2. Mitochondrial toxiciO, The /3-1yase-dependent biotransformation of bromine- lacking cysteine S-conjugates (Fig. 1, 2b) gives 1,1-dif- Early studies showed that DCVC inhibits respiration in luoroalkanethiolates (Fig. 1, 3b) as initial products. Reac- rat and guinea pig kidney slices [29]. DCVC was degraded tion of S-(2-chloro-l,l,2-trifluoroethyl)-L-cysteine with to pyruvate and ammonia by rat liver mitochondria, and N-dodecylpyridoxal in cetyltrimethylammonium chloride semicarbazide blocked both the biotransformation and in- micelles in the presence of benzyl bromide showed the hibition of respiration produced by DCVC [30]. It was also formation of benzyl 2-chloro-l,l,2-trifluoroethyl sulfide demonstrated that the DCVC-induced inhibition of respira- [24], indicating that 2-chloro-l,l,2-trifluoroethanethiolate tion was associated with inhibition of mitochondrial 2- was formed and trapped. When diethyl amine was used oxoacid dehydrogenases and lipoyl dehydrogenase [31-33]. instead of benzyl bromide, N,N-diethyl chlorofluoroth- These studies were conducted with rat liver mitochondria, ioacetamide was identified as a product, indicating that the but DCVC is not hepatotoxic [34]. (In the liver, kynureni- intermediate chlorofluorothioacetyl fluoride (Fig. 1, 4b, nase catalyzes /3-elimination reactions of cysteine S-con- X = CI,F) was formed and trapped. Chlorofluorothioacetyl jugates [35].) fluoride and difluorothioacetyl fluoride (Fig. 1, 4b, X = F), Respiration is, however, also decreased in kidney mito- formed from S-(l,l,2,2-tetrafluoroethyl)-L-cysteine, also chondria isolated from rats given DCVC [29,31]. Studies react with protein-bound N'-amino groups of lysine to with isolated rat renal proximal tubular cells showed that the cytotoxicity of DCVC was associated with the inhibition of cellular oxygen consumption, perturbations in cellular adenine nucleotide concentrations, and disturbances in ~ATP Ca 2+ compartmentation [20]. Incubation of rat kidney cells E~ ADP 4- ~ AMP with DCVC resulted in a large decrease in ATP and total ~_ ~ mm ATP + ADP + AMP adenine nucleotide concentrations and a small increase in ADP and AMP concentrations (Fig. 2); the changes in adenine nucleotide concentrations were blocked by the

"2 E_ 2 /3-1yase inhibitor aminooxyacetic acid. The energy charge ((ATP + I/2ADP)/(ATP + ADP + AMP)) was 0.80 in control cells, 0.59 in cells incubated with 1 mM DCVC, and 0.71 in cells incubated with 1 mM DCVC and 0.1 mM o m i aminooxyacetic acid. Cellular oxygen consumption was 0.0 1.0 1.0 + AOAA inhibited in cells incubated with S-(1,2- DCVC (raM) dichlorovinyl)glutathione or DCVC and was blocked by Fig. 2. Effect of S-(l,2-dichlorovinyl)-L-cysteine (DCVC) on cellular acivicin, an inhibitor of y-glutamyltransferase, or adenine nucleotide concentrations. Isolated kidney cells were incubated aminooxyacetic acid, respectively. Also, DCVC reduced with no additions, with DCVC (1 raM) and, or with DCVC (1 mM) and mitochondrial Ca 2+ sequestration by 62%, but did not alter aminooxyacetic acid (AOAA, 0.1 mM) for 30 rain at 37 ° C, and ATP, ADP, and AMP concentrations were determined by pyridine nucleotide- microsomal Ca 2÷ sequestration or plasma membrane Ca 2+ linked enzyme assays. Data are shown as mean _+ S.E., n = 3. Data are transport. The effect of DCVC on cellular oxygen con- from Lash and Anders [20]. sumption with different respiratory substrates was studied 54 M. W. Anders / Biochimica et Biophysica Acta 1271 (1995) 51-57

60 from mitochondria, and the rapid formation of membrane =o~ 5o ==ocvc blebs. In isolated rat kidney mitochondria, PCBC induced a rapid loss of mitochondrial Ca 2÷ and the collapse of the 0"~ 30 mitochondrial membrane potential; these effects were blocked by the /3-1yase inhibitor aminooxyacetic acid. 0 ~ 10 PCBC also inhibited state 3 respiration with succinate as the substrate, indicating that succinate dehydrogenase was o Glu + Mal Succ Ase + TMPD inhibited. Respiratory Substrate PCBC also induces mitochondrial dysfunction in rabbit Fig. 3. Effect of S-(l,2-dichlorovinyl)-L-cysteine (DCVC) on cellular renal proximal tubules [39]. PCBC (0.2 mM) increased oxygen consumption with different respiratory substrates. Isolated kidney basal and ouabain-insensitive respiration in rabbit renal cells were incubated in the presence and absence of DCVC (1 mM) with glutamate (Glu, 4 mM)+ malate (Mal, 2 mM), with succinate (Succ, 3.3 proximal tubules after 15 min incubation; this was fol- raM), or with ascorbate (Asc, 1 mM)+ N,N,N',N'-tetramethylene-p- lowed by a decrease in basal, nystatin-stimulated, and phenylenediamine (TMPD, 0.2 mM) for 2 h at 37 ° C; cellular oxygen ouabain-insensitive respiration and a decrease in glu- consumption was measured with a Clark-type electrode. Data are shown tathione concentrations after 60 min of incubation. PCBC as mean + S.E., n = 3. Data are from Lash and Anders [20]. is cytotoxic in rabbit renal proximal tubules, and additional studies showed that the initial increase in respiration resulted from an uncoupling of oxidative phosphorylation, (Fig. 3). DCVC had no effect on cellular oxygen consump- whereas the later changes were associated with gross tion with glutamate and malate or with ascorbate and mitochondrial damage, including inhibition of state 3 res- N,N, N',N'-tetramethyl-p-phenylenediamine as substrates, piration, cytochrome c-cytochrome oxidase, and electron but inhibited oxygen consumption with succinate as the transport. Further studies on the uncoupling of mitochon- substrate. drial oxidative phosphorylation showed that PCBC in- Further studies showed that DCVC produced a concen- creased state 4 respiration with pyruvate + malate or with tration-dependent reduction in mitochondrial state 3 oxy- succinate as the substrates and in the presence of the gen consumption in rat kidney mitochondria in the pres- FoFI-ATPase inhibitor oligomycin [40]. PCBC induces ence of succinate and ADP, but had no effect on state 4 swelling in rabbit renal cortical mitochondria in media respiration [36]. In addition, DCVC caused a time-depen- containing ammonium chloride or sodium chloride, and the dent release of Ca 2÷ and impaired the ability of protonophore carbonyl cyanide p-trifluoromethoxyphenyl succinate-energized mitochondria to generate a membrane hydrazone produces similar effects, indicating increased potential. In addition, incubation of rat kidney mito- mitochondrial proton permeability. These data show that chondria with 1 mM DCVC resulted in the leakage of PCBC initially uncouples oxidative phosphorylation by malic dehydrogenase into the medium. In addition to the dissipating the proton gradient. effects of DCVC on mitochondrial function and integrity, Recent studies with fluorescence digital imaging mi- mitochondrial metabolism was also perturbed. Mitochon- croscopy showed that DCVC produced perturbations in drial isocitrate, succinate, and aspartate concentrations were Ca 2÷ homeostasis in LLC-PK1 cells [41]. LLC-PK1 cells increased and the concentrations of glutamate and a-keto- incubated with DCVC showed lowered Ca 2÷ concentra- glutarate were decreased after incubation with 1 mM tions in the mitochondrial region, and the mitochondria DCVC. The changes in citric acid cycle metabolites were could not be stained with rhodamine-123, indicating severe accompanied by decreases in the activities of both isoci- mitochondrial damage and the loss of the mitochondrial trate dehydrogenase and succinate:cytochrome c oxido- membrane potential. The disturbances in Ca 2+ homeosta- reductase. Mitochondrial glutathione concentrations de- sis and mitochondrial function preceded the loss of cell creased rapidly and markedly (70%) in mitochondria incu- viability, as indicated by propidium iodide staining. Addi- bated with 1 mM DCVC, and the decrease in glutathione tional studies showed that DCVC-induced Ca 2+ efflux concentrations was paralleled by an increase in glutathione from mitochondria is preceded by oxidation of mitochon- disulfide concentrations. Finally, the decrease in mitochondrial pyridine nucleotides [42]. The DCVC-induced Ca 2+ drial glutathione concentrations was followed by an in- efflux is blocked by ATP, which inhibits the hydrolysis of crease in lipid peroxidation, as measured by the formation pyridine nucleotides, and by meta-iodobenzylguanidine, an of thiobarbituric acid-reactive substances. ADP-ribose acceptor. The elevations in cytosolic Ca 2+ The nephrotoxic conjugate S-(l,2,3,4,4-pentachloro- concentrations are followed by an increase in DNA dou- butadienyl)-L-cysteine (PCBC; Fig. 1, 2a, X = C1, CC1 = ble-strand breaks, which is associated with the activation CC12), which is formed by the glutathione-dependent of Ca 2+- and Mg2+-dependent endonucleases. The in- bioactivation of hexachlorobutadiene [37], is cytotoxic in crease in DCVC-induced DNA double-strand breaks is rat renal epithelial cells [38]. Exposure of rat kidney cells followed by increased poly(ADP-ribosylation) of nuclear to PCBC (0.1 mM) resulted in inhibition of respiration, proteins. Inhibitor studies show that DCVC-induced cyto- depletion of cellular ATP concentrations, loss of Ca 2+ toxicity in LLC-PKI cells is blocked by Quin-2, which M. W. Anders / Biochimica et Biophysica Acta 1271 (1995) 51-=57 55 chelates cytosolic Ca ~+, by aurintricarboxylic acid, which (~1 ~H H .0, inhibits DNA fragmentation, and by 3-aminobenzamide, which inhibits poly(ADP-ribosyl)transferase. CI b H H~--B:

Mitochondrial DNA is also as target for haloalkene-de- Pa .~ a rived cysteine S-conjugates. Metabolites of [1,2,3,4- MC ]4C]hexachlorobutadiene, which is metabolized to O C[ CI 0

[l,2,3,4-butadienyl-~4C]PCBC, become covalently bound ~'~SCoA + CI (2795 _ 410 pmol ~4C/mg DNA) to kidney mitochondrial 6 CI 4 2

DNA (mtDNA), whereas little binding to renal nuclear H20 ] Crotonase~FH20 DNA or to hepatic nuclear DNA or mtDNA was detected ~ [43]. PCBC also inhibits chloramphenicol-sensitive mito- 0 '~cP cL o~ H o chondrial protein, mtDNA, and mtRNA synthesis, and this HOJ~,~SCoA 8 effect is inhibited by aminooxyacetic acid [44]. In rat renal CI)x • S CI 3 mitochondria incubated with PCBC, much of the high- 7 molecular-weight mtDNA was degraded, and the super- CI 0 0 coiled form was converted to the relaxed circular form and o s~ + to shorter linear fragments. CI~.~OH CI 6 In summary, a range of haloalkene-derived, nephrotoxic $ 4 cysteine S-conjugates undergo mitochondrial bioactivation Fig. 4. Bioactivation of 5,6-dichloro-4-thia-5-hexenoyl-CoA (1). MCAD, and induce mitochondrial dysfunction. Indeed, mitochon- medium-chain acyl-CoA dehydrogenase; 2, 5,6-dichloro-4-thia-2,5- drial dysfunction is the hallmark of cysteine S-conjugate- hexadienoyl-CoA; 3, 5,6-dichloro-3-hydroxy-4-thia-5-hexenoyl-CoA; 4, induced cytotoxicity. The steps leading from mitochondrial 1,2-dichloroethenethiolate; 5, malonylsemialdehyde-CoA; 6, acryoyl- CoA; 7, chlorothioketene; 8, 3-hydroxypropionyl-CoA; 9, chloroacetic dysfunction to cell death have not been elucidated, but acid. recent studies indicate that mitochondrial HSP60 and a HSP70-1ike protein (mortalin) are targets for reactive intermediates formed from cysteine S-conjugates [45]; hence, that the 4-thiaalkanoate DCTH may also undergo bioacti- perturbations in vital housekeeping functions may be an vation by the pathway. important step in conjugate-induced cell death. 3.1. Bioactivation mechanism

3. Mitochondriai bioactivation of 4-thiaalkanoates DCTH-induced time- and concentration-dependent cytotoxicity in isolated rat hepatocytes [48]. The hypothesis Structure-toxicity studies with DCVC showed that the that DCTH is bioactivated by the enzymes of fatty acid analog lacking the carboxy group, i.e., S-(1,2-dichloro- fl-oxidation was studied with chain-length analogs of vinyl)-2-mercaptoethylamine, was not toxic, whereas the DCTH. The 5-thia and 7-thia analogs of DCTH (Fig. 5, analog lacking the amino group, i.e., 5,6-dichloro-4-thia- DCTHep and DCTN, respectively), elongated by addition 5-hexenoic acid (Fig. 4, 1, DCTH; also termed S-(l,2-di- of one or three methylene units, were not cytotoxic at chlorovinyl)-3-mercaptopropionic acid) was more toxic concentrations 20-fold higher than DCTH, whereas the than DCVC and also induced mitochondrial dysfunction 6-thia analog of DCTH (Fig. 5, DCTO), elongated by by inhibiting 2-oxoacid dehydrogenases [29,31,32]. addition of two methylene units was as cytotoxic as DCTH Studies were undertaken to test the hypothesis that DCTH is bioactivated by the enzymes of fatty acid /3- 100 oxidation: the initial step in the bioactivation of DCTH may be CoA thioester formation to give DCTH-CoA (Fig. 4, 1). DCTH-CoA may be a substrate for the medium-chain acyl-CoA dehydrogenase and give 5,6-dichloro-4-thia- 40- 2,5-hexadienoyl-CoA (Fig. 4, 2) as a product. Enoyl-CoA 20 2 may be a substrate for enoyl-CoA hydratase (crotonase) and yield the thiohemiacetal 5,6-dichloro-4-thia-3-hy- 0 Con DCTHep DCTN PCTOD droxy-5-hexenoyl-CoA (Fig. 4, 3). Thiohemiacetal 3 may DCTH DCTO CTFTH eliminate 1,2-dichloroethenethiolate (Fig. 4, 4) to afford Fig. 5. Cytotoxicity of DCTH (50 /xM), 6,7-dichloro-5-thia-6-heptenoic malonylsemialdehyde-CoA (Fig. 4, 6). Ethenethiolate 4 is acid (DCTHep, l raM), 7,8-dichloro-6-thia-7-octenoic acid (DCTO, 50 the same product that is presumably formed by the /3- /xM), 8,9-dichloro-7-thia-8-nonenoic acid (DCTN, 1 mM), 6-chloro- lyase-catalyzed /3-elimination reaction of DCVC (Fig. 1, 5,5,6-trifluoro-4-thiahexanoic acid (CTFTH, 100/xM), and 5,6,7,8,8-pentachloro-4-thia-5,7-octadienoic acid (PCTOD, 50 /xM) in isolated rat 3a, X = H, CI). 4-Thiaoctanoic acid is a substrate for the hepatocytes; Con = control. Data are shown as mean_+ S.D., n = 3. Data medium-chain acyl-CoA dehydrogenase [46,47], indicating are from Fitzsimmons and Anders [48]. 56 M. W. Anders / Biochimica et Biophysica Acta 1271 (1995)51-57

(Fig. 5). These data indicate that the 6-thia analog DCTO medium-chain acyl-CoA dehydrogenase activity is reduced undergoes one cycle of/3-oxidation to give DCTH, which in rat liver mitochondria incubated with DCTH. Also, a is cytotoxic. In contrast, the 5- and 7-thia analogs DCTHep prominent feature of DCTH-induced toxicity is hypo- and DCTN cannot form thiohemiacetals that eliminate glycemia; this may be the result of impaired gluconeo- 1,2-dichloroethenethiolate and are, therefore, not cytotoxic. genesis, which is a consequence of decreased fatty acid Moreover, the 4-thia analogs of the cysteine S-conjugates /3-oxidation. Methylenecyclopropylacetic acid, a metabo- S-(2-chloro-l,l,2-trifluoroethyl)-L-cysteine (Fig. 5, lite of hypoglycin A, inactivates the short- and medium- CTFTH) and PCBC (Fig. 5, PCTOD) are cytotoxic in rat chain acyl-CoA dehydrogenases and causes hypoglycemia hepatocytes, indicating that /3-oxidation results in the elim- [50]. Finally, organic aciduria was observed in DCTH- ination of a toxic metabolite. The cytotoxicity of DCTH treated rats. Increased urinary lactic acid concentrations was blocked by sodium benzoate, which depletes CoA may be the result of inhibition of the pyruvate dehydro- concentrations, and octanoic acid, which is an alternative genase complex [51 ]. substrate for the medium-chain acyl-CoA dehydrogenase; these observations support a role for /3-oxidation in the toxicity of DCTH. 4. Biomedical consequences of mitochondriai bioactiva- The bioactivation mechanism for DCTH outlined above tion of cysteine S-conjugates and 4-thiaalkanoates has recently been validated (Fitzsimmons, M.E., Thorpe, C. and Anders, M.W., unpublished data). Incubation of Mitochondrial bioactivation accounts for the observed DCTH-CoA with purified medium-chain acyl-CoA dehy- cytotoxicity of a range of cysteine S-conjugates in isolated drogenase in the presence of ferricenium hexafluorophos- kidney cells or in kidney-derived cell lines (i.e., LLC-PK1). phate as an electron acceptor lead to the formation of The organ-selective toxicity of haloalkene-derived cysteine 5,6-dichloro-4-thia-2,5-hexadienoyl-CoA (Fig. 4, path a, S-conjugates is associated with the ability of the kidney to 2). Enoyl-CoA hydratase-catalyzed hydration of enoyl-CoA transport and concentrate amino acids and amino acid 2, which would be expected to give an intermediate thio- analogs [52] and with the presence of /3-1yase in the hemiacetal (Fig. 4, 3), lead to the formation of mal- kidney [ 16]. Moreover, several haloalkenes are nephrotoxic onylsemialdehyde-CoA (Fig. 4, 5) and chloroacetic acid [53,54], and the glutathione-dependent bioactivation of (Fig. 4, 9); chloroacetic acid was presumably formed by these compounds explains their selective nephrotoxicity. In hydrolysis of the 1,2-dichloroethenethiolate-derived addition, the /3-1yase-dependent bioactivation of cysteine chlorothioketene (Fig. 4, 7 ~ 9). When DCTH-CoA was S-conjugates has been implicated in the nephrocarcino- incubated with the acyl-CoA dehydrogenase in the absence genicity of chloroalkenes [55]. of an electron acceptor, 3-hydroxypropionyl-CoA (Fig. 4, The target-organ selective toxicity of cysteine S-con- 8) was identified as a product. In the absence of an jugates amounts to a delivery system for toxic metabolites, electron acceptor (Fig. 4, path b), DCTH-CoA apparently and, in principle, the same system could be exploited to eliminates 1,2-dichloroethenethiolate 4 to give acryoyl- deliver drugs to the kidney. This possibility has been CoA (Fig. 4, 6), which is hydrated to give 3-hydroxypro- successfully tested: S-(6-purinyl)-L-cysteine undergoes /3- pionyl-CoA. Also, in the absence of an electron acceptor, lyase-dependent metabolism and produces elevated con- DCTH-CoA rapidly and irreversibly inactivates the centrations of 6-mercaptopurine in the kidney [56]. medium-chain acyl-CoA dehydrogenase. The bioactivation of 4-thiaalkanoates has potential for the organ- and organelle-selective delivery of drugs. The 3.2. Mitochondrial toxicity mitochondrial location of the enzymes of fatty acid /3- oxidation indicates that 4-thiaalkanoate-based drugs or The mitochondrial toxicity of 4-thiaalkanoates has not antioxidants could be exploited to release selectively drugs been explored in detail, but earlier studies indicate that or antioxidants in mitochondria. Also, if the kidney selec- mitochondria are a site of action [29,31,32]. Also, DCTH tivity of DCTH can be extrapolated to other 4- is activated by the enzymes of fatty acid /3-oxidation, thiaalkanoate-based drugs, it may be possible to deliver which are located in mitochondria. In studies in isolated rat drugs to renal mitochondria. These possibilities merit ex- hepatocytes, DCTH decreased cellular ATP concentra- amination. tions, increased ADP and AMP concentrations, decreased the energy charge, and decreased cellular glutathione concentrations, but had no effect on glutathione disulfide concentrations [48]. These perturbations in cellular adenine Acknowledgements nucleotide homeostasis are also consistent with a role for mitochondrial as a critical cellular target. Research in the author's laboratory was supported by In vivo studies also point to mitochondria as a target for National Institute of Environmental Health Sciences grant DCTH [49]. Fatty acid /3-oxidation is reduced in both liver ES03127 and by NATO grant 901032. Sandra E. Morgan and kidney mitochondria of DCTH treated rats, and assisted in the preparation of the manuscript. M. W. Anders / Biochimica et Biophysica Acta 1271 (1995) 51-57 57

References [27] Fisher, M.B., Hayden, P.J., Bruschi, S.A., Dulik, D.M., Yang, Y., Ward, A.J.I. and Stevens, J.L. (1993) Chem. Res. Toxicol. 6, 223-230. [1] Anders, M.W. (ed.) (1985) Bioactivation of Foreign Compounds, [28] Finkelstein, M.B,, Dekant, W. and Anders, M.W. (1994) Toxicolo- Academic Press, New York. gist 14, 110. [2] Guengerich, F.P. (1994) Adv. Pharmacol. 27, 211-236. [29] Parker, V.H. (1965) Food Cosmet. Toxicol. 3, 75-84. [3] Bock, K.W. (1994) Adv. Pharmacol. 27, 367-383. [30] Stonard, M.D. and Parker, V.H. (1971) Biochem. Pharmacol. 20, [4] Zia-Amirhosseini, P., Spahn-Langguth, H. and Benet, L.Z. (1994) 2429-2437. Adv. Pharmacol. 27, 385-397. [31] Stonard, M.D. and Parker, V.H. (t971) Biochem. Pharmacol. 20, [5] Falany, C.N. and Wilborn, T.W. (1994) Adv. Pharmacol. 27, 301- 2417-2427. 329. [32] Stonard, M.D. (1973)Biochem. Pharmacol. 22, 1329-1335. [6] Michejda, C.J. and Kroeger Koepke, M.B. (1994) Adv. Pharmacol. [33] Lock, E.A. and Schnellmann, R.G. (1990) Toxicol. Appl. Pharma- 27, 331-363. col. 104, 180-190. [7] Singer, T.P., Salach, J.I., Castagnoli, J.N. and Trevor, A. (1986) [34] Terracini, B. and Parker, V.H. (1965) Food Cosmet. Toxicol. 3, Biochem. J. 235, 785-789. 67-74. [8] Javitch, J., D'Amato, R., Nye, J, and Javitch, J. (1985) Proc. Natl. [35] Stevens, J.L. (1985) J. Biol. Chem. 260, 7945-7950. Acad. Sci. USA 82, 2173-2177. [36] Lash, L.H. and Anders, M.W. (1987) Mol. Pharmacol. 32, 549-556. [9] Ramsay, R.R., Krueger, M.J., Youngster, S.K., Glick, M.R., Casida, [37] Dekant, W., Vamvakas, S., Henschler, D. and Anders, M.W. (1988) J.E. and Singer, T.P. (1991) J. Neurochem. 56, 1184-1190. Drug Metab. Dispos. 16, 701-706. [10] Anders, M.W., Lash, L., Dekant, W., Elfarra, A.A. and Dohn, D.R. [38] Jones, T.W., Wallin, A., Thor, H., Gerdes, R.G., Ormstad, K. and (1988) CRC Crit. Rev. Toxicol. 18, 311-341. Orrenius, S. (1986) Arch. Biochem. Biophys. 251,504-513. [11] Lock, E.A. (1988)CRC Crit. Rev. Toxicol. 19, 23-42. [39] Schnellmann, R.G., Lock, E.A. and Mandel, L.J. (1987) Toxicol. [12] Koob, M. and Dekant, W. (1991) Chem.-Biol. Interact. 77, 107-136. Appl. Pharmacol. 90, 513-521. [13] Anders, M.W., Dekant, W. and Vamvakas, S. (1992) Xenobiotica [40] Schnellmann, R.G., Cross, T.J. and Lock, E.A. (1989) Toxicol. 22, 1135-1145. Appl. Pharmacol. 100, 498-505. [14] Dekant, W., Vamvakas, S. and Anders, M.W. (1992) In Tissue- [41] Vamvakas, S., Sharma, V.K., Sheu, S.-S. and Anders, M.W. (1990) Specific Toxicity: Biochemical Mechanisms (Dekant, W. and Neu- Mol. Pharmacol, 38, 455-461. mann, H.-G., eds.), pp. 163-194, Academic Press, London. [42] Vamvakas, S., Bittner, D., Dekant, W. and Anders, M.W. (1992) [15] Dekant, W., Vamvakas, S. and Anders, M.W. (1994) Adv. Pharrna- Biochem. Pharmacol. 44, 1131-1138. col. 27, 117-164. [43] Schrenk, D. and Dekant, W. (1989) Carcinogenesis 10, 1139-1141. [16] Cooper, A.J.L. and Anders, M.W. (1990) Ann. NY Acad. Sci. 585, [44] Banki, K. and Anders, M.W. (1989) Carcinogenesis 10, 767-772. 118-127. [45] Bruschi, S.A., West, K.A., Crabb, J.W., Gupta, R.S. and Stevens, [17] Stevens, J.L., Robbins, J.D. and Byrd, R.A. (1986) J. Biol. Chem. J.L. (1993)J. Biol. Chem. 268, 23157-23161. 261, 15529-15537. [46] Lau, S.-M., Brantley, R.K. and Thorpe, C. (1988) Biochemistry 27, [18] Elfarra, A.A., Lash, L.H. and Anders, M.W, (1987) Mol. Pharmacol. 5089-5095. 31, 208-212. [47] Lau, S.-M., Brantley, R.K. and Thorpe, C. (1989) Biochemistry 28, [19] Elfarra, A.A., Jakobson, I. and Anders, M.W. (1986) Biochem. 8255-8262. Pharmacol. 35, 283-288. [48] Fitzsimmons, M.E. and Anders, M.W. (1993) Chem. Res. Toxicol. [20] Lash, L.H. and Anders, M.W. (1986) J. Biol. Chem, 261, 13076- 6, 662-668. 13081. [49] Fitzsimmons, M.E., Baggs, R.B. and Anders, M.W. (1994) J. Phar- [21] Dekant, W., Urban, G., GSrsmann, C. and Anders, M.W. (1991) J. macol. Exp. Ther. 271,515-523. Am. Chem. Soc. 113, 5120-5122. [50] Wenz, A., Thorpe, C. and Ghisla, S. (1981) J. Biol. Chem. 256, [22] Miiller, D.-A., Urban, G. and Dekant, W. (1991) Chem.-Biol. Inter- 9809-9812. act. 77, 159-172. [51] Haworth, J.C., Perry, T.L., Blass, J.P., Hansen, S. and Urquhart, N. [23] Dekant, W., Berthold, K., Vamvakas, S., Henschler, D. and Anders, (1976) Pediatrics 58, 564-572. M.W. (1988) Chem. Res. Toxicol. 1, 175-178. [52] Lash, L.H. and Anders, M,W. (1989) J. Pharmacol. Exp. Ther. 248, [24] Dekant, W., Lash, L.H. and Anders, M.W. (1987) Proc. Natl. Acad. 531-537. Sci. USA 84, 7443-7447. [53] Yang, R.S.H. (1988)Rev. Environ. Contam. Toxicol. 101, 121-137. [25] Hargus, S.J. and Anders, M.W. (1991) Biochem. Pharmacol. 42, [54] Kennedy, J., G.L. (1990)Crit. Rev. Toxicol. 21, 149-170. R 17-R20. [55] Dekant, W. and Vamvakas, S. (1992) Adv. Pharmacol. 22, 297-337. [26] Harris, J.W., Dekant, W. and Anders, M.W. (1992) Chem. Res. [56] Hwang, I.Y. and Elfarra, A.A. (1991) J. Pharmacol. Exp. Ther. 258, Toxicol. 5, 34-41. 171-177.