Cysteine Conjugate S-Oxidase CHARACTERIZATION OF A NOVEL ENZYMATIC ACTIVITY IN RAT HEPATIC AND RENAL

Cysteine conjugate S-oxidase activity, with S-benzyl-L-cysteine as substrate, was found mostly in the microsomal fractions of rat liver and kidney. In the presence of oxygen and NADPH, S-benzyl-L-cysteine is converted to S-benzyl-L-cysteine sulfoxide; no S-benzyl-L-cysteine sulfone was detected. The Vmax for S-benzyl-L-cysteine sulfoxide formation by kidney microsomes was nearly 3-fold greater than the rate measured with liver microsomes. Inclusion of catalase, superoxide dismutase, glutathione, butylated hydroxyanisole, the peroxidase inhibitor, potassium cyanide, the cytochrome P-450 inhibitors, 1-benzylimidazole and metyrapone, or a monoclonal antibody to cytochrome P-450 reductase did not inhibit the metabolic reaction. Flavin-containing monooxygenase alternate substrates, N,N-dimethylaniline, n-octylamine, and methimazole inhibited the S-oxidase activities. Analogues of S-benzyl-L-cysteine, S-methyl-L-cysteine, and S-(1,2-dichlorovinyl)-L-cysteine inhibited the S-benzyl-L-cysteine S-oxidase activities, whereas S-carboxymethyl-L-cysteine and S-benzyl-L-cysteine methyl ester had no effect. These results provide clear evidence against the involvement of reactive oxygen intermediates or cytochrome P-450 in the sulfoxidation of S-benzyl-L-cysteine and indicate that the S-oxidase activities may be associated with flavin-containing monooxygenases which exhibit selectivity in the interaction with cysteine S-conjugates.

Cysteine conjugate S-oxidase activity, with S-benzyl-L-cysteine as substrate, was found mostly in the microsomal fractions of rat liver and kidney.
In the presence of oxygen and NADPH, S-benzyl-L-cysteine is converted to S-benzyl-L-cysteine sulfoxide; no Sbenzyl-L-cysteine sulfone was detected.
The V,,,,x for S-benzyl-L-cysteine sulfoxide formation by kidney microsomes was nearly 3-fold greater than the rate measured with liver microsomes.
Inclusion of catalase, superoxide dismutase, glutathione, butylated hydroxyanisole, the peroxidase inhibitor, potassium cyanide, the cytochrome P-450 inhibitors, 1-benzylimidazole and metyrapone, or a monoclonal antibody to cytochrome P-450 reductase did not inhibit the metabolic reaction.
Analogues of S-benzyl-L-cysteine, S-methyl-L-cysteine, and S-(1,2-dichlorovinyl)-~cysteine inhibited the Sbenzyl-L-cysteine S-oxidase activities, whereas S-carboxymethyl-L-cysteine and S-benzyl-L-cysteine methyl ester had no effect. These results provide clear evidence against the involvement of reactive oxygen intermediates or cytochrome P-450 in the sulfoxidation of S-benzyl-L-cysteine and indicate that the S-oxidase activities may be associated with flavin-containing monooxygenases which exhibit selectivity in the interaction with cysteine S-conjugates.
Glutathione S-conjugate formation is an important mechanism for detoxication of many chemicals (1,2). This metabolic reaction is catalyzed by glutathione S-transferases which are present in several tissues, with activity in the liver being much higher than in the kidney. Glutathione S-conjugates are eliminated in the bile or, after further metabolism to mercapturic acid, in the urine. The activities of the enzymes that catalyze mercapturic acid formation, namely, y-glutamyl transpeptidase, cysteinylglycine dipeptidase, and cysteine conjugate N-acetyltransferase, are much higher in the kidney than in liver (l-5).
Recently, the nephrotoxicity, mutagenicity, and carcinogenicity of halogenated hydrocarbons such as ethylene dichloride, ethylene dibromide, trichloroethylene, chlorotrifluoroethylene, and hexachlorobutadiene have been attributed to the formation of glutathione and cysteine S-conjugates (6)(7)(8)(9)(10)(11)(12)(13)(14). Evidence for the existence of two classes of nephrotoxic cysteine conjugates, i.e. direct acting through the formation of reactive episulfonium ions, and indirect acting which require activation by cysteine conjugate /3-lyase to generate reactive intermediates, was obtained (12)(13)(14). Metabolism of cysteine S-conjugates by cysteine conjugate P-lyase may also result in the formation of stable thiols (1,15). This metabolic reaction has been exploited recently for site selective delivery of B-mercaptopurine, an anti-tumor drug, to the kidneys (16). Evidence for the presence of another metabolic pathway which involves the conversion of cysteine S-conjugates to sulfoxides has been reported. For example, in viuo treatments of rats and mice with ethylene dibromide or ethylene dimethanesulfonate resulted in the excretion of S-(2-hydroxyethyl)-N-acetyl-L-cysteine sulfoxide in urine (17); ethyl mercapturic acid sulfoxide was identified as a urinary metabolite of bromoethane in rats (18). Furthermore, a number of cysteine S-conjugates such as S-methyl-L-cysteine, S-carboxymethyl-L-cysteine, S-ethyl-L-cysteine, S-(2-hydroxyethyl)-Lcysteine, S-propyl-L-cysteine, S-butyl-L-cysteine, and S-pentyl-L-cysteine were metabolized in oiuo to sulfoxides (18-22). When rats were given S-ethyl-L-cysteine, approximately 18 and 35% of the dose was excreted as ethylmercapturic acid and the corresponding sulfoxide, respectively (22). A sulfenic acid species which could arise by the cysteine conjugate /?lyase-dependent metabolism of S-(pentachlorobutadienyl)-Lcysteine sulfoxide had also been identified in the urine of rats given hexachlorobutadiene (8). Although sulfoxidation appears to be a major metabolic pathway for cysteine S-conjugates in viuo, the nature of the enzymes involved and their role in the metabolism and toxicity of cysteine S-conjugates were not previously investigated. Therefore, the present study was performed to characterize the hepatic and renal S-oxidation of cysteine Sconjugates, using S-benzyl-L-cysteine (SBC)' as substrate, in terms of the metabolites formed and the mechanisms involved. The results of these studies provide clear evidence against the involvement of reactive oxygen intermediates or cytochrome P-450 in this metabolic reaction. Furthermore, the results indicate the involvement of flavin-containing monooxygenases. The rate of SBC sulfoxide formation by kidney microsomes was nearly s-fold greater than the rate obtained with liver microsomes. Materials-NADPH,  catalase,  superoxide  dismutase,  horse heart  cytochrome  c, S-methyl-L-cysteine,  S-carboxymethyl-L-cysteine,  Sbenzyl-L-cysteine  methyl  ester,  glutathione,  N,N-dimethylaniline,  butylated  hydroxyanisole,  and (2 mM) and incubating at 31 "C for 20 min. Protein concentrations were determined by the method of Lowry et al. (30) using bovine serum albumin as standard. Antibody protein concentrations were determined by using E:,$'"' = 1.35 cm-' (31). RESULTS A highly sensitive HPLC assay, which involves the formation, separation, and detection of nanomolar amounts of N-2,4-dinitrophenyl derivatives of SBC and its potential metabolites, SBC sulfoxide and SBC sulfone, was developed to characterize hepatic and renal cysteine conjugate S-oxidases. HPLC analyses of SBC incubation mixtures with rat hepatic and renal microsomes in the presence of NADPH and oxygen resulted in the appearance of a new component on the HPLC chromatogram (Fig. lB), which had an electronic absorption spectrum (Fig. 1C) and retention time similar to that obtained with a reference sample of SBC sulfoxide (data not shown). These results demonstrate the presence of cysteine conjugate S-oxidases in rat liver and kidney. Whereas no enzymatic formation of SBC sulfone was detected, SBC sulfoxide formation was dependent on the presence of NADPH, and oxygen ( Fig. 1). Sulfoxidation of SBC by hepatic and renal microsomes was also dependent upon incubation time and protein concentrations (Fig. 2). Kidney microsomal SBC sulfoxidation exhibited a temperature optimum of 45 "C and a pH optimum of 7.2; liver SBC S-oxidase exhibited a similar trend (Fig. 2). Subcellular localizations of the NADPH-dependent SBC Soxidase activities in both rat liver and kidney showed that all of the S-oxidase activity was present in the particulate fractions (Table I). Because the renal S-oxidase activity was found in both the microsomal and mitochondrial fractions, the possibility of contamination of the kidney mitochondrial fraction with microsomes was investigated by studying the distribution of glucose-6-phosphatase, a microsomal marker enzyme, in kidney subcellular fractions. The results show that the distribution of glucose-6-phosphatase activity paralleled that of the S-oxidase activity (data not shown). These results indicate that the renal mitochondrial fraction was contaminated significantly with microsomes. Attempts to carry out the Soxidase assay using subcellular fractions which were prepared by a method reported to yield a pure kidney mitochondrial fraction (32) were not successful. This fractionation method reqiired the use of sucrose, and triethanolamine, which was found to inhibit the hepatic and renal S-oxidase activities (data not shown). The addition of only sucrose to the fractionation buffer, described under "Experimental Procedures," did not improve the purity of the kidney mitochondrial fraction (data not shown). The total recovery of SBC S-oxidase in liver or kidney subcellular fractions was nearly 100% with the nuclear and cell debris fractions containing approximately 30% of the total activity (data not shown).

EXPERIMENTAL PROCEDURES
The kinetics of SBC sulfoxidation in hepatic and renal microsomes were studied using double-reciprocal plots. Renal microsomal SBC S-oxidase exhibited a nearly 3-fold higher V,,, than hepatic microsomal S-oxidase (Table II).
To determine the type of enzyme catalyzing SBC oxidation, the effects of selected cofactors or inhibitors on SBC Soxidase activities were determined in liver and kidney microsomes (Table III). The absence of NADPH or removal of oxygen, by a 5-or lo-min nitrogen purge, significantly inhibited SBC S-oxidation in both liver and kidney microsomes. Inclusion of 1-benzylimidazole, a selective cytochrome P-450 inhibitor which does not affect flavin-containing monooxy- C, electronic absorption spectrum of peak II.
Because SBC is readily oxidized by hydrogen peroxide,  , had no effect on hepatic and renal SBC S-oxidase activities. Furthermore, when microsomes were incubated with NADPH for 20 min and ethanol and SBC were added simultaneously, no SBC sulfoxide was detected. This indicates that NADPH does not simply initiate lipid peroxidation or generation of reactive oxygen species (35,46) which lead to SBC sulfoxidation, but rather is required as a cofactor. Chemical inhibition data indicated that cytochrome P-450 might not be involved in SBC sulfoxidation.
A monoclonal antibody to cytochrome P-450 reductase has been shown to be effective in determining the relative contribution of cytochrome P-450 and flavin-containing monooxygenases in Sand N-oxidations of various chemicals in microsomes (37,38,41). Therefore, to provide further evidence against cytochrome P-450 involvement in SBC S-oxidase activity, a monoclonal antibody to rat cytochrome P-450 reductase was incubated with hepatic or renal microsomes and its effect on SBC sulfoxidation was compared with its effect on microso-ma1 cytochrome c reduction. A concentration of cytochrome P-450 reductase antibody which inhibited nearly 70-80% of hepatic or renal microsome-catalyzed cytochrome c reduction had no effect on SBC sulfoxidation (Fig. 3). The addition of an equal amount of bovine serum albumin to incubations without antibody also had no effect on SBC S-oxidase activities, thereby ruling out the possibility that the increase in protein concentration due to the presence of the cytochrome P-450 reductase antibody nonspecifically aided in SBC sulfoxidation.
The lack of complete inhibition of microsomal cytochrome c reduction may have been due to the inability of the antibody to reach the enzyme at some of its sites within the microsomes or its inability to completely block the active site of cytochrome P-450 reductase.
In an effort to characterize the SBC-binding site on SBC S-oxidase, structural analogues of SBC were evaluated for their abilities to inhibit the hepatic and renal S-oxidase activities. As shown in Table IV, S-methyl-L-cysteine and S-(1,2-dichlorovinyl)-L-cysteine inhibited the hepatic and renal S-oxidase activities, whereas S-carboxymethyl-L-cysteine or S-benzyl-L-cysteine methyl ester had no effect. Although the mechanism of inhibition of the S-oxidase activities by Smethyl-L-cysteine and S-(1,2-dichlorovinyl)-L-cysteine was not investigated, the results suggest selectivity in the interaction of the hepatic and renal S-oxidases with cysteine Sconjugates.

DISCUSSION
In the present study, the presence of cysteine conjugate soxidases which catalyzed the conversion of SBC to SBC sulfoxide was demonstrated in rat liver and kidney homoge-   nates.
The cysteine conjugate S-oxidases were characterized by studying the distribution of the activities in rat hepatic and renal subcellular fractions, and by studying the biochemical mechanism and other selected properties of the reaction. As shown in Tables I-III and Figs. 1 and 2, the S-oxidase activity which was present mostly in the microsomal fractions of both liver and kidney, was dependent on the presence of oxygen, NADPH, incubation temperature, time, pH, and protein concentration.
S-Methyl-L-cysteine and S- (1,2-dichlo-rovinyl)-L-cysteine inhibited the hepatic and renal SBC Soxidase activities, whereas S-carboxymethyl-L-cysteine and S-benzyl-L-cysteine methyl ester had no effect (Table IV). These results suggest that cysteine S-conjugates which have a non-ionizable alkyl group on the sulfur atom, or which do not have an esterified carboxyl group on the cysteine moiety, are likely to act as substrates for the S-oxidases. Renal microsomes catalyzed SBC S-oxidation at a rate that was nearly 3fold higher than the rate obtained with liver microsomes (Tables I and II, Fig. 2). Although the reason for this rate difference is not clear, inclusion of the cytochrome P-450 reductase antibody did not alter the preferential sulfoxidation of SBC in kidney microsomes compared with liver microsomes (Fig. 3). This indicates that reduction of SBC sulfoxide by cytochrome P-450 reductase did not contribute to the observed rate difference. SBC sulfoxidation is, to our knowledge, the first example of a sulfoxidation reaction that is preferentially catalyzed by kidney microsomes compared with liver microsomes. It should be noted, however, that metabolism of cysteine S-conjugates to mercapturic acids is also preferentially catalyzed by kidney microsomes compared with liver microsomes (5,6).
Microsomal enzymes such as the cytochrome P-450 family of monooxygenases, the flavin-containing monooxygenases, and prostaglandin synthetase are known to catalyze the conversion of many different sulfides to sulfoxides (35,(46)(47)(48)(49)(50)(51)(52). The findings that the S-oxidase activities were dependent on NADPH, and that inclusion of catalase or KCN in the incubation mixture did not affect the SBC S-oxidase activities, provide evidence against the participation of peroxidases such as prostaglandin synthetase in the NADPH-dependent Soxidase activities. Similarly, involvement of reactive oxygen intermediates, such as hydrogen peroxide and superoxide anion, which can be generated from the cytochrome P-450 containing monooxygenases (46) or lipid peroxides, generated from peroxidative processes of microsomes (35,(43)(44)(45)(46) was ruled out by demonstrating that SBC S-oxidase activities were not inhibited by catalase, superoxide dismutase, butylated hydroxyanisole, or glutathione (Table III). In addition, the findings that the S-oxidase activities were not inhibited by the cytochrome P-450 inhibitors, metyrapone and l-benzylimidazole (Table III), or by a monoclonal antibody against cytochrome P-450 reductase (Fig. 3) provide clear evidence against the involvement of cytochrome P-450 in the metabolic reaction.
The flavin-containing monooxygenase alternate substrates, N,N-dimethylaniline, n-octylamine, and methimazole, significantly inhibited SBC S-oxidase activities in both hepatic and renal microsomes (Table III). These results indicate that SBC S-oxidase activities may be associated with flavin-containing monooxygenases.
Methimazole saturates the flavin-containing monooxygenases at concentrations less than 200 pM, and at concentrations less than 2 mM, methimazole is S-oxygenated at measurable rates only by the flavin-containing monooxygenases (35). Thus, the finding that methimazole is an effective competitive inhibitor of the hepatic and renal Soxidase activities with K, values of 60 and 40 PM, respectively, provides further evidence for the involvement of flavin-containing monooxygenases in SBC S-oxidation. The results presented in this report indicate that cysteine S-conjugates may act as substrates for flavin-dependent monooxygenase enzymes. It should be noted, however, that a conclusive statement regarding this apparent new type of substrate of flavin-containing monooxygenases or the exact identity of the SBC S-oxidase cannot be made before purifying this enzymatic activity. Studies with purified enzymes may provide explanations for the differences (outlined below) that exist between the properties of the S-oxidases reported here and those reported for known flavin-containing monooxygenase-dependent reactions. First, the relative tissue distribution of SBC S-oxidase (Table I) is inconsistent with the reported immunochemical and biochemical studies on the relative distribution of the N,N-dimethylaniline oxidase, a flavin-containing monooxygenase, in rat tissues (35,41,53). Second, the pH optimum for the hepatic and renal SBC S-oxidase was slightly lower than those reported for the flavin-containing monooxygenase-dependent S-oxidation of thiobenzamide and methimazole (35,38,54). These latter substrates, however, differ from SBC in containing no ionizable group which may have significant effects on both the binding of the substrate to the enzyme and on the rate of the enzymatic reaction. Third, the observed temperature optimum for SBC S-oxidase in liver and kidney microsomes (Fig. 2) was higher than those reported for flavin-containing monooxygenases (35,38,40). Interestingly, in regard to the apparent thermal stability and the effects of n-octylamine, the liver and kidney SBC Soxidases appear to be more similar to the lung flavin-containing monooxygenases rather than the liver and kidney isoenzymes (37)(38)(39)55).
The role of S-oxidases in the disposition and toxicity of cysteine S-conjugates is not clear. The products of cysteine conjugate S-oxidase, i.e. the sulfoxides, would be expected to be more polar than the sulfides, and hence S-oxidase may affect the distribution and excretion of cysteine S-conjugates. Oxidation of the sulfur atom of S-(2-haloethyl)-L-cysteine, a potential metabolite of 1,2-dihaloethane, is expected to prevent the non-enzymatic rearrangement of the molecule to form a reactive episulfonium ion. Thus, S-oxidation of S-(2haloethyl)-L-cysteine may represent a detoxication reaction. Sulfoxidation may also affect the activities of enzymes, e.g. cysteine conjugate N-acetyltransferase and cysteine conjugate P-lyase, which are involved in the detoxication and bioactivation of cysteine S-conjugates (5)(6)(7)(8)(9)(10)(11)(12)(13)(14). Thus, it is also possible that sulfoxides of cysteine S-conjugates may be more toxic than the sulfide forms.
In conclusion, the results presented in this report describe the development of a highly sensitive HPLC assay which was used to characterize cysteine conjugate S-oxidase activities in rat liver and kidney microsomes. The usefulness of cofactor dependence, selective inhibitors, and monoclonal antibodies in studying the mechanism of SBC S-oxidase was demonstrated. Future studies, which examine the metabolism and toxicity of nephrotoxic and mutagenic cysteine S-conjugates and their corresponding sulfoxides, should elucidate the role of this enzyme in the bioactivation of cysteine S-conjugates. Enzyme purification studies and immunochemical comparison with known forms of flavin-containing monooxygenases should also reveal the characteristics of the hepatic and renal microsomal cysteine conjugate S-oxidases.