A Purified Cysteine Conjugate @-Lyase from Rat Kidney Cytosol REQUIREMENT FOR AN a-KETO ACID OR AN AMINO ACID OXIDASE FOR ACTIVITY AND IDENTITY WITH SOLUBLE GLUTAMINE TRANSAMINASE

Cysteine conjugate @-lyase has been purified from rat kidney cytosol. The enzyme is a 100,000-dalton dimer of two 55,000-dalton subunits and has an ab- sorption maximum at 432 nm. The enzyme has phenylalanine a-keto-y-methiolbutyrate transaminase ac- tivity and appears to be identical to rat kidney cytosolic glutamine transaminase K. from a-keto acids and their separation using the HPLC system described here. nondenaturing gel electrophoresis done by (22). Enzyme activity gel by values for activity were compared to the was determined an Coomassie Blue

Cysteine conjugate @-lyase has been purified from rat kidney cytosol. The enzyme is a 100,000-dalton dimer of two 55,000-dalton subunits and has an absorption maximum at 432 nm. The enzyme has phenylalanine a-keto-y-methiolbutyrate transaminase activity and appears to be identical to rat kidney cytosolic glutamine transaminase K.
The data also show that rat kidney cytosolic glutamine transaminase K catalyzes both a &elimination and a transamination reaction with DCVC when aketo-y-methiolbutyrate is present and that amino acid oxidase and a-keto-y-methiolbutyrate stimulate the enzyme activity by providing amino acceptors. When incubations were done with DCVC as substrate in the presence of excess a-keto-7-methiolbutyrate, the 8lyase catalyzed @-elimination and transamination in a ratio of 1:1.3, respectively. Under conditions where most of the a-keto-y-methiolbutyrate was consumed, the &elimination predominated indicating that the S-1,2dichlorovinyl -3mercapto -2oxopropionic acid pool was consumed by transamination after the a-ketoy-methiolbutyrate had been depleted. The data are discussed with regard to the importance of these pathways as regulators or participants in the toxicity of S-cysteine conjugates.
Glutathione S-transferases catalyze the formation of a pantheon of xenobiotic S-glutathione conjugates (l), the first step in the detoxication scheme known as mercapturic acid biosynthesis (2). The detoxication function of mercapturic acid biosynthesis is completed by the formation of an S-cysteine conjugate from an S-glutathione conjugate and the excretion of the S-cysteine conjugate after N-acetylation, i.e. as the mercapturate (3,4). Though mercapturic acid biosynthesis functions largely as a detoxication system, the formation of * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
$ To whom correspondence should be addressed W. Alton Jones Cell Science Center, Old Barn Road, Lake Placid, NY 12946.
glutathione conjugates from some xenobiotics can also initiate toxic processes rather than detoxication (5)(6)(7)(8)(9)(10)(11). Cysteine conjugate @-lyase is an enzyme which cleaves some S-cysteine conjugates by p-elimination and has been implicated in the renal toxicity of some S-glutathione conjugates (5)(6)(7)(8)(9). We have reported that the majority of the rat liver cysteine conjugate @-lyase activity is contributed by an enzyme which is in fact a kynureninase (12). In addition, cysteine conjugate @-lyase activity is present in both the cytosolic and mitochondrial fractions of rat kidney cortex (13)(14)(15), but little is known about the nature of the kidney enzymes. The proteins responsible for rat kidney cytosolic cysteine conjugate P-lyase activity do not cross-react with antibody raised against the rat liver enzyme (13), suggesting that the cysteine conjugate @-lyase enzymes present in rat kidney cytosol are different from the hepatic enzyme(s).
In a model kidney cell line, LLC-PK1 cells, the toxicity of S-1,2-dichlorovinyl-~-glutathione is dependent on both metabolism to its S-cysteine conjugate, S-1,2-dichlorovinyl-~cysteine (DCVC)', and the metabolism of DCVC, probably by a cysteine conjugate @-lyase (6). Furthermore, the cysteine conjugate @-lyase activity in LLC-PK1 cells is stimulated by the addition of pyruvate to cell homogenates, suggesting that a transamination reaction regulates the @-elimination activity in this model system. Stimulation of @-elimination activity by pyruvate resulted in an increase in the covalent binding of a sulfur-containing fragment from DCVC to cellular macromolecules. Inhibition of @-lyase activity by aminooxyacetic acid, a general inhibitor of pyridoxal phosphate-dependent enzymes (16), blocked both the covalent binding and toxicity of DCVC in the LLC-PK1 model. Aminooxyacetic acid also blocks the toxicity of DCVC in the rat (5).
Preliminary results from the investigation of DCVC metabolism in rat kidney cytosol indicated that a-keto acids stimulated @-lyase activity 2-3-fold. In addition, a protein factor which separated from the @-lyase during purification, and appeared to be an amino acid oxidase, was able to stimulate DCVC metabolism. After the amino acid oxidase was separated from the @-lyase, the @-lyase activity became almost totally dependent on the presence of an a-keto acid. If the cytosolic cysteine conjugate @-lyase were a pyridoxal phosphate-dependent transaminase, the stimulation of activity by a-keto acids and an amino acid oxidase could be explained by the hypothesis represented schematically in Fig. 1. According to the hypothesis, a single enzyme might catalyze both @elimination and transamination reactions, in accord with the proposal for the enzyme in LLC-PK1 cells (6). In the absence of added a-keto acid or an amino acid oxidase, which directly oxidizes amino acid substrates to a-keto acids, the @-lyase would accumulate in the inactive pyridoxamine form. The data presented in this report support this hypothesis and show that under certain conditions, transamination of DCVC is in fact the predominant reaction catalyzed by the purified enzyme. The implications of these observations are discussed with regard to the role of this enzyme in nephrotoxicity.

EXPERIMENTAL PROCEDURES
Materials-Radiolabeled and unlabeled DCVC were prepared as previously described (6). ["CC]DCVC was synthesized from [U-"C] cysteine HCl, and 13%]DCVC was synthesized from [35S]cystine. All other chemicals were of at least reagent grade and were used without further purification.
Enzyme Purification-Rat kidney cytosolic cysteine conjugate /3lyase was purified by a modification of the procedure used for the hepatic enzyme (12), a procedure which is similar to the method of Cooper and Meister (17,18) for cytosolic glutamine transaminase K. Kidneys were obtained from male Sprague-Dawley rats (300-400 g), and cortex cytosol was prepared by differential centrifugation in 0.25 M sucrose containing 10 mM Tris-HC1, pH 7.5, 0.1 mM phenylmethylsulfonyl fluoride, 1 mM EDTA, and 1 mM dithiothreitol (Buffer A). The cytosol was adjusted to pH 5.0 with 1 M acetic acid and stirred at 4 "C for 15 min. After centrifugation for 20 min at 10,000 X g, the pellet was discarded and the supernatant adjusted to pH 7.0 with 1 M Tris (pH 9.0). The supernatant was further fractionated with solid ammonium sulfate, and the protein which precipitated between 35 and 50% saturation was collected by centrifugation and dissolved in Buffer A containing no sucrose (Buffer B). After dialysis against two changes of Buffer B, the protein was charged onto a 40 X 2.5-cm column of DEAE-Sephacel (Pharmacia Fine Chemicals, Uppsala, Sweden) equilibrated with Buffer B, and the enzyme eluted with a 1.5-liter gradient of 0-0.35 M KC1 in Buffer B. Active fractions, which eluted at 7.8 mmho, were pooled and concentrated by precipitation with ammonium sulfate at 60% saturation. The precipitated protein was dissolved in 5 ml of Buffer B and applied to a 2.5 X 90-cm column of Sephacryl S-300 superfine grade (Pharmacia Fine Chemicals, Uppsala, Sweden) equilibrated with Buffer B. Enzyme was eluted with Buffer B, and the active fractions were concentrated by precipitation with ammonium sulfate at 60% saturation. The precipitated protein was dialyzed against three changes of 5 mM potassium phosphate (pH 6.5) containing 1 mM dithiothreitol, 1 mM EDTA, and 0.1 mM phenylmethylsulfonyl fluoride (Buffer C), and charged onto a 1.5 X 15-cm column of hydroxylapatite (Bio-Rad) equilibrated with Buffer C. The enzyme was eluted with a 300-ml linear gradient of 5-250 mM potassium phosphate (pH 6.5) in Buffer C. The enzyme eluted at 3.4 mmho and was concentrated by vacuum dialysis against Buffer C adjusted to pH 7.5 (Buffer D). The enzyme was charged onto a 0.4 X 25-cm Synchrompak Q300 anion exchange column (Synchrompak, Linden, IN) equilibrated with Buffer D and operated at a flow rate of 1 ml/min with a Gilson programmable high pressure liquid chromatograph (Gilson Medical Electronic, Middleton, WI). The enzyme was eluted with a 30-ml gradient of 5-200 mM potassium phosphate (pH 7.5) in Buffer D. The purified enzyme was stored at a concentration of 1 mg/ml at 4 "C for at least 12 weeks with no loss of activity. Protein was determined according to the method of Bradford (19) using the Bio-Rad kit and bovine IgG as standard.
Amino acid oxidase was purified from the S-300 column eluate by pooling the fractions which had amino acid oxidase activity and concentrating them by precipitation with ammonium sulfate at 60% saturation. The protein was dialyzed against two changes of Buffer D and applied to a 1 X 25-cm Synchrompak Q300 column equilibrated with buffer D and operated at 3 ml/min. Enzyme was eluted with a 75-m15-200 mM gradient of potassium phosphate in Buffer D.
Enzyme Assays-Cysteine conjugate &lyase activity was assayed by the extraction method previously reported (6) using ["C]DCVC as the substrate. Since [U-"Clcystine was used for the preparation of ["CC]DCVC, the assay detects those extractable products containing the cysteine carbon skeleton. As will be shown later, metabolism of DCVC can yield two a-keto acid products, pyruvate and the a-keto acid of DCVC, S-(1,2-dichlorovinyl)-3-mercapto-2-oxopropionic acid. The extraction assay does not differentiate between the two products. Separation of the two products is achieved with the HPLC assay described below. Enzyme was incubated in a volume of 250 pl for 10 min at 37 "C with 1 mM ["C]DCVC and 100 mM Tris.HC1 (pH 8.0) with 0.1 mM a-keto-y-methiolbutyrate, unless stated otherwise. The reaction was stopped by the addition of 125 pl of 2 N HCl and extracted with 2 ml of ethyl acetate. Radioactivity in 1 ml of the ethyl acetate was determined by liquid scintillation counting. Glutamine transaminase K was assayed by the method of Cooper and Meister (18) using phenylalanine and a-keto-y-methiolbutyrate as the substrates. Amino acid oxidase activity was assayed with leucine (20 mM) using the dinitrophenylhydrazine method of Greenberg et al. (20) to determine a-keto acid product with a-ketoisocaproic acid as the standard. Incubations with amino acid oxidase also included 50 mM sodium pyrophosphate (pH 9.0) and 40 pg/ml catalase.
HPLC Analysis of Incubation Mixtures-Products from reaction mixtures were separated by HPLC and quantitated by both radiochemical detection and absorbance at 210 nm. Incubations were in a final volume of 150 pl and contained 1 mM DCVC; other additions are as described in the figure legends. Following incubation, 10 p1 of phosphoric acid was added to stop the reaction, and 50 p1 of the incubation mixture were injected onto a 0.4 X 30-cm C18 reversed phase column (Waters Associates, Milford, MA) operating at a flow rate of 1 ml/min. Products were eluted with one of the following solvent programs where solvent A is 0.1 M potassium phosphate buffer (pH 3.0) and solvent B is methanol: System A, 0-5 min 100% A, 5-15 min ramp to 40% B, plateau at 40% B through 30 min; or System B, 0-10 min 100% A, 10-20 min ramp to 40% B, plateau at 40% B through 30 min.

Identification of S-1,2-Dichlorovinyl-3-mercapto-2-oxopropionic
Acid-In order to gain structural information about the products from the amino acid oxidase and cysteine conjugate B-lyase-dependent metabolism of DCVC, the product from the amino acid oxidase reaction was isolated as a 2,4-dinitrophenylhydrazone derivative. Amino acid oxidase (0.36 mg), isolated as described above, was incubated with DCVC (9 pmol), catalase (24 pg), and 25 pmol of sodium pyrophosphate (pH 9.0) in a final volume of 0.5 ml. After 1 h, 200 pl of a 1% solution of 2,4-dinitrophenylhydrazine in 2 N HCl was added and allowed to react for 15 min at room temperature. The unreacted dinitrophenylhydrazine and any hydrazone derivatives were extracted into 2 ml of ethyl acetate and the solvent removed in uacllo. The yellow residue was dissolved in 500 pl of methanol, and 500 pl of 0.1 M potassium phosphate was added. The sample was applied to a 0.4 X 30-cm Water pBondapak C18 column, and the column was eluted with 60% methanol in 0.1 M potassium phosphate (pH 3.0) while monitoring both 280-and 360-nm wavelengths with a Gilson model 116 detector. Unreacted dinitrophenylhydrazine eluted just after the front, and the derivative of S-1,2-dichlorovinyl-3-mercapto-2-oxopropionic acid was recovered as two peaks which eluted at 19 and 24 min. The two peaks are believed to correspond to cis and trans (E and Z) isomers of the hydrazone derivative. The larger peak from the HPLC analysis of the product at 19 min was collected and the solvent removed in vacuo with the aid of additional methanol. The residue was extracted into 2 ml of ethyl acetate and exchanged two times with a mixture of 1 ml of acetone and 0.5 ml of 'H20. The sample was dried overnight under high vacuum, dissolved in 0.5 ml of 'Hs-acetone, and the 400-MHz 'H-NMR spectrum was collected with a JEOL GX-400 spectrometer.
The separation of two isomers of dinitrophenylhydrazones by thin layer chromatography is well known (21). To confirm this behavior in the HPLC system described here, the dinitrophenylhydrazone of pyruvate was prepared by adding 2.0 g of pyruvate to 50 ml of 85% phosphoric acid and recrystallizing the product from ethanol/water. The recrystallized dinitrophenylhydrazone of pyruvate gave a single peak by HPLC analysis. Upon heating an aqueous solution of the hydrazone, two peaks were observed, both of which were found to be the dinitrophenylhydrazone of pyruvate by NMR analysis (data not shown). When the dinitrophenylhydrazone of pyruvate was prepared from incubation mixtures spiked with authentic pyruvate, using the method described above, two peaks were seen by HPLC as well. Furthermore, when dinitrophenylhydrazone derivatives were prepared from incubations containing [35S]DCVC and amino acid oxidase, two peaks, both of which contained 33S label, were found. These data confirm the formation of cis and trans isomers of dinitrophenylhydrazones from a-keto acids and their separation using the HPLC system described here. Electrophoresis-SDS-PAGE and nondenaturing gel electrophoresis were done by the method of Maize1 (22). Enzyme activity was determined after nondenaturing gel electrophoresis by incubating gel slices in 200 pl of 10 mM potassium phosphate buffer at 4 "C (pH 7.5) and assaying aliquots of the supernatant for enzyme activity. The RF values for enzyme activity were compared to the RF for protein which was determined by staining an identical gel with Coomassie Brilliant Blue R-250 (22).

RESULTS
Purification and Characterization of Rat Kidney Cytosolic Cysteine Conjugate 8-Lyase-Cysteine conjugate p-lyase activity was purified 890-fold from rat kidney cytosol (Table I).
Phenylalanine-a-keto-y-methiolbutyrate transamination activity copurified with the p-lyase activity, and both activities comigrated during ion exchange chromatography by HPLC ( Fig. 2 4 ) and nondenaturing gel electrophoresis (Fig. 2B). When phenylalanine was included in incubation mixtures containing the purified enzyme, a-keto-y-methiolbutyrate and ["CC]DCVC, the formation of product from DCVC was inhibited with kinetic behavior characteristic of competitive inhibition (Fig. 3). The K,,, and V, , , for DCVC metabolism in the absence of inhibitors were 0.26 mM and 22.0 pmol/mg. 10 min, respectively. The dashed line in Fig. 2 illustrates that substrate inhibition was observed with high DCVC concentrations. The data show that both the phenylalanine-a-ketoy-methiolbutyrate transaminase and cysteine conjugate plyase activities are the property of the same protein and suggest that both reactions are catalyzed at the same active site on the enzyme.
The protein has an M, of about 100,000, as determined by gel filtration chromatography (data not shown), and two subunits of M , = 55,000 determined by SDS-PAGE (Fig. W). A minor band which was not seen by nondenaturing gel electrophoresis (Fig. 2B) was apparent by SDS-PAGE.
Although pyridoxal phosphate is not required for stability during the purification, the purified enzyme has absorption maxima at 432 and 282 nm (Fig. 4), a spectrum characteristic of some pyridoxal phosphate-containing proteins (23). Furthermore, agents that are known to inhibit pyridoxal phosphate-dependent enzymes inhibited the purified enzyme (Table 11).
During purification (Table I), the ability of the enzyme to metabolize DCVC became increasingly dependent on the presence of a-keto-y-methiolbutyrate. Fig. 5 shows the dependence of the purified enzyme on the a-keto-y-methiolbutyrate concentration. Activity reaches a plateau at about 0.5 mM aketo-y-methiolbutyrate and then decreases; increasing the concentrations (1-5 mM) of a-keto-y-methiolbutyrate actually resulted in a further decrease in enzyme activity. In initial experiments with enzyme from Step 5, maximal stimulation was seen with 0.1 r n~ a-keto-y-methiolbutyrate, and this concentration was chosen for the standard assay. Taken together, the substrate inhibition kinetics (Fig. 3) and the inhibition of the enzyme at high a-keto-y-methiolbutyrate concentrations suggest that substrate inhibition occurs.
The structure-activity relationship for the stimulation of DCVC metabolism by a-keto acids was also investigated (Table 111). Those a-keto acids with more hydrophobic substituents at the &carbon were most effective in maintaining @-elimination activity. The ability to stimulate enzyme activity increased at higher concentrations (5 mM) for those aketo acids which, at a concentration of 0.1 mM, were not as effective as a-keto-y-methiolbutyrate (data not shown).
An Amino Acid Oxidase Is Required for the Expression of Rat Kidney Cytosolic Cysteine Conjugate B-Lyase Activity in Vitro-Addition of a-keto-y-methiolbutyrate produced only

Purification of rat kidney cysteine conjugate &lyase
Glutamine transaminase K (transamination) was assayed with phenylalanine and a-keto-y-methiolbutyrate as substrates, and cysteine conjugate @-lyase activity was assayed with DCVC (@-lyase) as substrate in the presence of 0.1 mM a-keto-y-methiolbutyrate as described under "Experimental Procedures." The ratio of activity with DCVC as substrate in the presence and absence of a-keto-y-methiolbutyrate (+mtb/-mtb) and the ratio of the phenylalanine transamination activity (TRA) to the activity with DCVC (CBL) activity are shown (TRA/CBL), A unit of enzyme is 1 pmol of product produced in 10 min at 37 "C, and specific activity is expressed as microunits/ mg of protein.  FRACTION  FIG. 2. Co-migration of phenylalanine-a-keto-y-methiolbutyrate transaminase and cysteine conjugate &lyase activities by HPLC anion exchange chromatography. A shows the co-chromatography of phenylalanine-a-keto-y-methiolbutyrate transaminase activity (0---0 ) and cysteine conjugate @-lyase (M) activity on a Synchrompak Q300 anion exchange column and SDS-PAGE analysis of the purified protein; the major band has an M , of 55,000. Protein (-) is expressed as absorbance at 280 nm relative to the absorbance of the major peak, and nmol refers to the amount of product produced from a 10-min incubation. B shows the comigration of transaminase and cysteine conjugate 8-lyase activity with protein (13 pg) upon nondenaturing gel electrophoresis of the purified protein. An identical gel was stained to show protein. a 2.5-fold stimulation of @-lyase activity in rat kidney cortex cytosol. Dialysis of kidney cortex cytosol or purification through Step 4 (Table I) did not result in an increase in the ratio of independent to a-keto-y-methiolbutyrate-dependent cysteine conjugate @-lyase activity with DCVC as the substrate, suggesting that the presence of a-keto acids in the cytosol was not responsible for the modest dependence on aketo-y-methiolbutyrate. Since gel filtration on Sephacryl s-300 (Step 5, Table I) resulted in an increase in the ratio of aketo-y-methiolbutyrate stimulated to unstimulated activity from 2-to 10-fold, we investigated the possibility that a factor which stimulates P-lyase activity was separated by gel filtration from the a-keto-y-methiolbutyrate-stimulated P-lyase activity. Fig. 6 shows that a small peak (Peak 1) of apparent @-lyase activity, which was insensitive to a-keto-y-methiolbutyrate, eluted prior to the main peak of a-keto-y-methiolbutyrate-stimulated activity and phenylalanine-a-keto-y- methiolbutyrate transaminase activity. When aliquots from Peak 1 were added back to the fractions which contained the a-keto-y-methiolbutyrate-dependent @-lyase, activity with DCVC was restored in the absence of a-keto-y-methiolbutyrate (data not shown).
The protein in Peak 1 was purified further by ion exchange HPLC. The purified protein oxidized leucine to an a-keto acid (specific activity, 30 nmol/min. mg) and had a spectrum which was characteristic of a flavin-containing protein ( Fig.  7 (24)), suggesting that the factor from Peak 1 is a flavin TABLE I1 Inhibition of cysteine conjugate 0-lyase acitvity Inhibition studies were done using enzyme with a specific activity of 0.8 pmol/mg.min. The enzyme was incubated in the standard assay mixture with the inhibitor for 15 min prior to the addition of 1 mM DCVC and 0.1 mM a-keto-r-methiolbutvrate.

Structure-activity relationship for the stimulation of enzyme activity
Metabolism was assayed with DCVC as described under "Experimental Procedures." All a-keto acids were added at 0.1 mM, and the stimulation is expressed as a percentage of the stimulation seen with a-keto-y-methiolbutyrate. containing amino acid oxidase (25). SDS-PAGE (Fig. 7) revealed a major diffuse band of about 38,000 daltons suggesting a high molecular weight protein with multiple subunits, consistent with data reported for renal amino acid oxidase (26). The ability of the amino acid oxidase to stimulate enzyme activity was compared to a-keto-y-methiolbutyrate stimulation (Fig. 8). Stimulation by high concentrations of amino acid oxidase was almost as effective as a-keto-y-methiolbutyrate (0.1 mM; single point in Fig. 8). The stoichiometry, on a molar basis, of amino acid oxidase to @-lyase which produced maximal stimulation was 36:1, but ratios as low as 3:l stimulated activity a t a level which was 25% of the level achieved Amino acid oxidase activity was measured with L-leucine as substrate; specific activity, 30 nmol/min/mg. with 0.1 mM a-keto-y-methiolbutyrate. The calculations are based on an M , of 314,000 (25) for amino acid oxidase and 100,000 for the cysteine conjugate @-lyase.

@-
Lyase-It seemed possible that in the crude extracts from rat kidney cortex, an amino acid oxidase provides the necessary a-keto acid required to maintain @-lyase activity by oxidizing substrate to an a-keto acid. Walsh et al. (27) have shown that with @-chloro-D-alanine as substrate, D-amino acid oxidase from pig kidney can catalyze a @-elimination reaction producing ammonia, pyruvate, and chloride as the products or oxidize @-chloro-D-alanine to chloropyruvate and ammonia with the concomitant reduction of molecular oxygen. The ratio of pyruvate to chloropyruvate produced is dependent on the availability of oxygen as an electron acceptor, and a mixture of pyruvate and chloropyruvate was produced at 20% oxygen. Our data suggested that pyruvate was not particularly effective in stimulating the @-lyase activity indicating that the stimulation might result from the oxidation of DCVC by rat kidney amino acid oxidase to the corresponding a-keto acid, S-(1,2-dichlorovinyl) (DMOP) which then serves as a cosubstrate for transamination (Fig. 1). We investigated this possibility by HPLC analysis of the metabolites produced from incubations with DCVC and the purified rat kidney cytosolic amino acid oxidase. In addition, we characterized the products from the metabolism of DCVC by rat kidney cytosolic cysteine conjugate @-lyase. Fig. 9A shows the HPLC chromatogram from a reaction mixture which contained the purified rat kidney amino acid oxidase and DCVC. A new peak, labeled DMOP, with a retention time of 19 min, is seen when the enzyme is incubated with DCVC for 30 min at 37 "C; pyruvate has a retention time of 3.5 min on this column. When identical incubations were done using either 35Sor I4C-labeled DCVC and radioactivity in the peaks was quantitated, both labels were associated with the DMOP peak. Fig. 10 shows that the product in the DMOP peak is produced in a time-dependent manner concomitant with the loss of DCVC. The combined recovery of radioactivity in the DCVC and DMOP peaks was 97, 95, 96, and 97% at 10,20,30, and 60 min, respectively, suggesting that DMOP was the only major product and that the product had lost none of the radioactivity present in the parent DCVC. Table   IV summarizes results from incubations with 35Sand 14Clabeled DCVC and shows that similar results are obtained with either 35Sor 14C-labeled DCVC indicating that all of the label from DCVC is in the product. Furthermore, when the 35S-labeled product was derivatized with 2,4-dinitrophenylhydrazine, the radioactivity shifted to two peaks which eluted at a much higher methanol concentration (65%) and had absorbance at 380 nm, suggesting that the product had formed a 2,4-dinitrophenylhydrazone derivative of an a-keto acid (data not shown). The 400-MHz 'H-NMR spectrum of the dinitrophenylhydrazone isolated from incubations ( Fig. 11) was consistent with the structural assignment shown. The intensity of the vinyl proton signal increased when the acquisition time was increased from 4 to 30 s. This is consistent with the behavior of a proton which resides on a carbon with chlorine substituents and no nearby protons. The long relaxation time of the chlorovinyl proton accounts for the 1:0.8

TABLE IV Metabolism of S-(l,2-dichlorouiny~-~-cysteine by amino acid oxidase
Radioactivity which was lost from either ["C]DCVC or ["S]DCVC as well as radioactivity appearing in DMOP was measured by reversed phase HPLC analysis as described under "Experimental Procedures." Amino acid oxidase (0.36 mg) was incubated for 30 min at 37 "C in a final volume of 150 p1 with 1 mM DCVC, 50 mM sodium pyrophosphate buffer (pH 9.0), and 40 pg/ml catalase. Reactions were stopped by the addition of 10 p1 of 5 M phosphoric acid, and 50 p1 were injected onto the column. n = 2. 1 " " l " " l " " l " " I " " I " "~" " l " " O.a integration for the pheny1:vinyl and methylene:vinyl protons, respectively. Therefore, the product in the DMOP peak was identified as S-1,2-dichlorovinyl-3-mercapto-2-oxopropionic acid. The data suggest that S-1,2-dichlorovinyl-3-mercapto-2-oxopropionic acid is the a-keto acid which is responsible for the stimulation of @-lyase activity by amino acid oxidase. Fig. 9B shows a chromatogram for the reaction products from an incubation containing 1 mM DCVC, 0.5 mM a-ketoy-methiolbutyrate, and purified cysteine conjugate @-lyase. The absorbance in the DCVC and the a-keto-y-methiolbutyrate peaks decreases with incubation, while the peaks labeled DMOP, pyruvate, and methionine increase. Table V summarizes the results from incubations of cysteine conjugate @-lyase with 35Sand 14C-labeled DCVC and shows that the DMOP peak contains both the cysteine carbons and the sulfur from DCVC, as seen in the amino acid oxidase experiment. As expected, I4C label but no 35S label is found in the pyruvate peak. The data show that DMOP and pyruvate are produced in a 1.3:l ratio, respectively, in the presence of 0.5 mM aketo-y-methiolbutyrate. When 0.1 mM a-keto-y-methiolbutyrate is used (Table V), almost all the a-keto-y-methiolbutyrate is consumed, and the ratio decreases to 1:4. Therefore, it appears that with DCVC as substrate, the rat renal cytosolic cysteine conjugate @-lyase catalyzes both @-elimination and transamination reactions.

DISCUSSION
Cysteine conjugate @-lyase was purified from rat kidney cortex cytosol and found to have @-lyase activity with Scysteine conjugates and phenylalanine-a-keto-y-methiolbutyrate transamination activity. Cooper (28) has shown that phenylalanine-a-keto-y-methiolbutyrate transamination activity in rat kidney is a property of glutamine transaminase K. The size and subunit composition of rat kidney cysteine conjugate @-lyase activity, as well as the predilection for aketo acids with hydrophobic substituents at the @-carbon, suggests that the cysteine conjugate @-lyase reported here is identical to the soluble form of rat kidney glutamine transaminase K. Our results agree with those of Cooper and Meister (17) in that the enzyme is not stabilized by pyridoxal phosphate during the purification, unlike the hepatic form of cysteine conjugate @-lyase which has been shown to be a kynureninase and is stabilized by pyridoxal phosphate during purification (12). However, the inhibition of the enzyme by agents which inhibit pyridoxal phosphate enzymes and the UV-visible spectrum of the purified enzyme suggests that pyridoxal phosphate is a cofactor. Odum and Green (29) have reported a 24-fold purification of a soluble form of cysteine conjugate @-lyase from rat kidney. They include pyridoxal phosphate in some of the purification steps but do not report whether this stabilized the enzyme. In addition, these authors do not mention a role for either a-keto acid or an amino acid oxidase.
The results show that the enzyme is capable of catalyzing both transamination and p-elimination reactions with DCVC as substrate. In fact the ability to catalyze a p-elimination is dependent on a source of a-keto acids which act as amino group acceptors and complete the transamination reaction. According to the scheme in Fig. 1, when a source of a-keto acids is absent the enzyme would accumulate in the inactive pyridoxamine form which cannot accept another substrate. The requirement for an a-keto acid is fulfilled by the presence of an amino acid oxidase in the cytosol which complements the transamination activity by oxidizing DCVC to S-(1,2dichlorovinyl)-3-mercapto-2-oxopropionic acid which then serves as the amino acceptor to complete the transamination. A key feature of the scheme is that both reactions can occur through the formation of a common intermediate, the Schiff base between the pyridoxal phosphate and its amino acid substrate. Following Schiff base formation, pyridoxal phosphate enzymes are thought to abstract the a-proton of the substrate amino acid and stabilize the carbanion through resonance with the pyridine ring of the pyridoxal phosphate. The resulting quinoneimine intermediate (not shown in Fig.  1) is common to enzymes which catalyze @-elimination, @replacement, and transamination reactions (30).
If a good leaving group is present at the @-carbon, the quinoneimine can collapse via @-elimination to an enzyme-bound aminoacrylate, which is released from the enzyme and hydrolyzes to pyruvate nonenzymatically ( Fig. 1 (30, 31)). Alternatively,

Rat Kidaey Cortex Cys~eine C o~j~u t e @-Lyase
addition of water to the a-carbon of the quinoneimine results in the formation of an a-keto acid and an enzyme-bound pyridoxamine, completing the first half of a transamination reaction. The partitioning between @-elimination and transamination from a common quinoneimine intermediate can be involved in the mechanism-based inactivation of some pyridoxal phosphate-dependent enzymes (31). A similar scheme was proposed for the stimulation of DCVC metabolism by pyruvate in LLC-PK1 cells (6). In LLC-PK1 cells, DCVC metabolism to pyruvate was stimulated 2-3-fold by the addition of 10 mM pyruvate to incubation mixtures containing cell homogenate and DCVC. The data presented here are consistent with this model. The 2-3-fold stimulation of activity in LLC-PK1 cell homogenates is similar to the stimulation seen in rat kidney cytosol (Table I).
At present we cannot determine how such S-1,2-dichlorovinyl-3-mercapto-2-oxopropionic acid might be produced in vivo from the metabolism of DCVC by glutamine transaminase K and/or amino acid oxidase, since it is both a substrate and a product. The data in Table V show that the ratio of S-1,2-dichlorovinyl-3-mercapto-2-oxopropionic acid to pyruvate is 1:4, respectively, when the incubation was done under conditions where the a-keto-y-methiolbut~ate was limiting (0.1 mM). At this concentration, at least 80% o f the a-ketoy-methiolbutyrate was metabolized, while at a concentration of 0.5 mM, only 30% of the a-keto-y-methiolbutyrate was metabolized (data not shown). Therefore, the apparent decrease in the ratio of S-1,2-dichlorovinyl-3-mercapto-2-oxopropionic acid to pyruvate in the presence of 0.1 mM a-ketoy-methiolbutyrate is probably due to the reutilization of the S-1,2-dichlorovinyl-3-mercapto-2-oxopropionic acid as a cosubstrate for transamination when ff-keto-y-methiolbutyrate becomes limiting due to metabolism. The recycling of S-1,2dichlorovinyl-3-mercapto-2-oxopropionic acid to DCVC and further metabolism via P-elimination and t r a n s~i n a t i o n would eventually result in the accumulation of pyruvate.
Therefore, the ratio of products in uiuo may also depend on the availability of a-keto acid cosubstrates. This recycling hypothesis is supported by the observation that amino acid oxidase stimulates the metabolism of DCVC in the absence of another amino acid or a-keto acid substrate.
When the metabolism of DCVC by rat kidney amino acid oxidase was measured by the formation of a dinitrophenylhydrazone derivative, a product was formed which, under basic conditions, had absorption at 440 nm (data not shown), a wavelength at which, under basic conditions, the dinitrophenylhydrazone of pyruvate and a-ketoisocaproate absorb. Therefore, since 2,4-dinitrophenylhydrazine will form a hydrazone with any a-keto acid, it may not be appropriate for the detection of an a-keto acid metabolite produced from an Scysteine conjugate if structural information is not available.
The observation that S-1,2-dichlorovinyl-3-mercapto-2-oxopropionic acid can constitute 60% of the product from the metabolism of DCVC by glutamine transaminase K and is the sole product from the amino acid oxidase-catalyzed metabolism of DCVC raises the obvious question. Do a-keto acids of S-cysteine conjugates play any role in toxicity? At this point, we have no information on the toxicity of a-keto acid products from any cysteine conjugates. However, in this regard it is interesting to note that Kaczorowski and Walsh (32) have suggested that chloropyruvate produced from @chloroalanine may be the species which inhibits active transport in Escherichia coli membrane vesicles.
Whether the oxidation of DCVC by the kidney amino acid oxidase plays a role in toxicity is not clear. However, aminooxyacetic acid inhibits DCVC toxicity in the rat (5) and in LLC-PK1 cells (6) but does not inhibit the amino acid oxidase isolated from rat kidney cytosol (data not shown). This suggests that the cysteine conjugate @-lyase arm of the metabolic scheme is necessary. It is possible, however, that inhibition of the amino acid oxidase would produce a similar blockade of toxicity if the metabolic cooperation between amino acid oxidase and cysteine conjugate @-lyase operates in the whole cell. Amino acid oxidase is also present in the mitochon~ial fraction of rat kidney (25) along with cysteine conjugate plyase activity (14,15). Mitochondria have been proposed as a site of S-cysteine conjugate toxicity; therefore, the purification of the mitochondrial @-lyase is an important issue. Moreover, it is possible that metabolic cooperation between amino acid oxidase and mitochondrial cysteine conjugate @-lyase occurs as well.
Cooper and Meister (17,18,28) have characterized a mitochondrial form of glutamine transaminase K which is very similar to the cytosolic form of glutamine transaminase K.
However, Lash et al. (15) have performed mitochondrial fractionation studies in which they find that glutamine transaminase K is in the matrix, as reported by Cooper and Meister (18), but that cysteine conjugate @-lyase activity is present only in the outer membrane fraction indicating that glutamine transaminase K in mitochondria is not a P-lyase. Our preliminary data suggest that a cysteine conjugate @-lyase which is active with DCVC and is present in the rat kidney mitochondrial matrix is also dependent on the presence of an a-keto acid and that the enzyme cross-reacts with antibody raised against the cytosolic enzyme, which we have shown is in fact glutamine transaminase K (33). Moreover, when assayed with DCVC, the mitochondrial matrix enzyme is also dependent on the presence of ~-keto-~-methiolbutyrate. Therefore, the matrix enzyme could be missed if one does not assay DCVC activity in the presence of an a-keto acid.
In conclusion, we have shown that metabolism of DCVC in rat kidney cytosol is the result of cooperation between an amino acid oxidase and glutamine transaminase K. A novel a-keto acid metabolite of DCVC has been identified. The data suggest that metabolic cooperation between an amino acid oxidase and a cysteine conjugate @-lyase activity may be important as a regulator and/or participant in the bioactivation of S-cysteine conjugates to toxic species. This hypothesis is under investigation.