Cysteine 254 of the 73-kDa A Subunit Is Responsible for Inhibition of the Coated Vesicle (H+)-ATPase Upon Modification by Sulfhydryl Reagents*

The vacuolar class of (H+)-ATPases are highly sen- sitive to sulfhydryl reagents, such as N-ethylmaleim-ide. The cysteine residue which is responsible for in- hibition of the coated vesicle (H+)-ATPase upon modification by N-ethylmalemide is located in subunit A and is able to form a disulfide bond with the cysteine moiety of cystine through an exchange reaction. This unique property distinguishes this cysteine residue from the remaining cysteine residues of the (H+)-ATP-ase. Using this reaction, we selectively labeled the cystine-reactive cysteine residue of subunit A with fluorescein-maleimide. After complete digestion of the labeled subunit A by V8 protease, a single labeled fragment of moleculer mass 3.9 kDa was isolated and the amino-terminal sequence was determined. This fragment contains 2 cysteine residues, Cys240 and Cysza4. Since Cyszar is conserved among all vacuolar (H+)-ATPases whereas Cys240 is not, it is likely that C Y S ~ " ~ is the residue which is responsible for the sen- sitivity of the vacuolar (H+)-ATPase to sulfhydryl reagents. are inhibited by higher concentrations of NEM (0.1-1 mM), and the F-type ATPases, which are virtually resistant to inhibition by NEM. The cysteine residue(s) which are responsible

are inhibited by higher concentrations of NEM (0.1-1 mM), and the F-type ATPases, which are virtually resistant to inhibition by NEM. The cysteine residue(s) which are responsible for NEM sensitivity are present in subunit A (2,lZ-14). Based on sequence homology with the a and / 3 subunits of F1, subunit A contains a putative nucleotide binding site (15)(16)(17)(18). Binding of ATP to subunit A prevents reaction of the cysteine residue(s) of subunit A with NEM or 7-chloro-4-nitrobenz-2oxa-l,3-diazole and thus protects the enzyme from inhibition (2, 12-14). On the other hand, modification of subunit A by NEM also prevents binding of nucleotides to this subunit (12). Identification of the cysteine residues which are responsible for NEM inhibition will provide information concerning the role of particular cysteine residues in catalysis.
A cDNA encoding subunit A of the bovine vacuolar ATPase has been cloned and sequenced recently in this laboratory (18). According to the deduced amino acid sequence, this subunit contains 8 cysteine residues. In the present work, we have identified CysZ5* as the cysteine residue responsible for NEM sensitivity of the coated vesicle (H')-ATPase.
All vacuolar ATPases share the property of being inhibited by low concentrations of N-ethylmaleimide (NEM) (1-2 p~) (1). This distinguishes them from the P-type ATPases, which 9-Amino-6-chloro-2-methylacridine and fluorescein maleimide were purchased from Molecular Probes, Inc. [3H]NEM was obtained from Du Pont-New England Nuclear. Most of the other chemicals were purchased from Sigma.
Preparation of Clathrin-coated Vesicle ATPase Modified by Cystine-Clathrin-coated vesicles were prepared from calf brain as previously described (2). Vesicles were stripped of their clathrin coat according to Adachi et at. (19). Stripped vesicles were suspended in 1 mM EDTA, 10% glycerol, and 20 mM HEPES (pH 7.0) at a protein concentration of 2 mg/ml. The stripped vesicles were then dialyzed against two changes of the above suspension buffer (300 volumes) for 30 h a t 4 "C to remove trace dithiothreitol which is present in the coated vesicle preparation. The vesicles were then dialyzed against 1 mM cystine, 1 mM EDTA, 10% glycerol, and 20 mM HEPES (pH 7.0) for 24-30 h at 4 "C. The cystine-treated vesicles were then placed on ice, and an aliquot was assayed for the ATP-dependent proton transport described below. The vesicles were kept on ice until proton transport was completely inhibited (usually less than 24 h). The (H')-ATPase was then solubilized from the cystine-treated vesicles and purified by glycerol gradient centrifugation as previously described (2) with the exception that mercaptoethanol was omitted from solutions for solubilization and purification.
Labeling of the Coated Vesicle (H+)-ATPase with Fluorescein Maleimide-Stripped vesicles were incubated with cystine until no proton transport could be detected as described above. NEM (5 mM) was then added to the cystine-treated stripped vesicles, and the reaction was allowed to proceed for 1 h at 23 "C. Dithiothreitol (50 mM) was then added to the vesicles to neutralize unreacted NEM and to cleave disulfide bonds formed between cysteine residues of the protein and the cysteine moiety of cystine. The mixture was kept at 23 "C for 30 min before being dialyzed against 1 mM EDTA, 10% glycerol, and 20 mM HEPES (pH 7.0) to remove dithiothreitol. The dialysis was allowed to proceed for 2 days with four changes of dialysis buffer. Fluorescein maleimide (1 mM) was then added to the dialyzed vesicles, and the reaction was allowed to proceed at 23 "C for 1 h. Dithiothreitol (20 mM) was then added the vesicles to neutralize excess fluorescein maleimide and the mixture was incubated at 23 "C for 30 min. The fluorescein maleimide-labeled stripped vesicles were then centrifuged for 1 h at 100,000 X g in a Beckman SW-50.1 rotor. The pellet was suspended in 50 mM NaCI, 30 mM KC1,0.2 mM EGTA, 10% glycerol, and 20 mM HEPES (pH 7.0). The fluorescein maleimide-labeled (H+)-ATPase was solubilized from the stripped vesicles and purified by glycerol gradient centrifugation as previously described (2).
Separation and Sequencing of the Fluorescein Maleimide-labeted Fragment-SDS was added to the fluorescein maleimide-labeled (H+)-ATPase prepared as described above to a final concentration of 0.2%. The protein sample was transferred to a 25,000 molecular weight cut-off dialysis tubing and dialyzed against 150 mM NaCl, 0.02% SDS, and 50 mM Tris-HC1 (pH 8.0) for 2 days and then against 0.02% SDS for 1 day at 23 "C. The (H+)-ATPase sample containing 1 mg of protein was then lyophilized to reduce the volume to 500 pl and run on a 10% acrylamide gel according to Fling and Gregerson (20). The M, 73,000 band (subunit A) labeled by fluorescein maleimide was cut out when illuminated by long UV light. Subunit A was then electroeluted from the gel as previously described (21). The eluted subunit A was dialyzed in 25,000 molecular weight cut-off dialysis tubing against 150 mM NaCl and 0.02% SDS for 2 days and then against 0.02% SDS for 1 day at 23 "C to remove excess SDS. The protein concentration of the subunit A solution was estimated from the absorbance at 210 nm with bovine serum albumin as a standard. 50 mM sodium phosphate (pH 7.8) and 2 mM EDTA were added to the dialyzed subunit A followed by V8 protease. The mixture was incubated at 37 "C for 2-3 days until subunit A was completely digested. V8 protease was added to the solution by three additions over the course of incubation so that the final ratio of proteinm8 protease is 201. The V8 fragments were dialyzed in a 1,000 molecular weight cut-off tubing against 0.02% SDS at 23 "C for 3 days to reduce the concentration of phosphate to about 1 mM. The dialyzed sample was then concentrated to 100 pl and run on a 20% acrylamide gel according to Fling and Gregerson (20). The fluorescent band of 3.9 kDa was cut out and washed in 10% methanol and 10 mM CAPS (pH 11) for 30 min. The 3.9-kDa peptide was then electrotransferred to Immobilon for 30 min as previous described (10). The fluorescent band was then cut out from the Immobilon and sequenced using an Applied Biosystems 477A protein Sequencer.
Other Methods-ATPase activity was measured by a continuous spectrophotometric assay as previously described (22). The proton transport activity was measured by 9-amino-6-chloro-2-methyoxyacridine fluorescence quenching using a Perkin-Elmer LS-5 spectrofluorometer as previously described (2) with minor modifications. Briefly, less than 200 pl of the ATPase samples were diluted with 50 mM NaCI, 30 mM KC1, and 20 mM HEPES (pH 7.0) to a final volume of 500 pl. 9-Amino-6-chloro-2-methyoxyacridine (1 p M ) and valinomycin (1 p~) were then added to the solution. The ATP hydrolysis was started by addition of 20 pl of 10 mM ATP and 20 mM MgSOI. When fluorescence quenching reached the maximum, 1 pM of carbonyl cyanide rn-chlorophenylhydrazone was added to the reaction mixture. The proton transport activity was determined from the carbonyl cyanide rn-chlorophenylhydrazone-sensitive fluorescence quenching. Reconstitution of the purified (H+)-ATPase was carried out as previously described (2). Protein concentration was determined by the method of Lowry et al. (23) or, for samples containing CI& by the method of Schaffner and Weissman (24). SDS-polyacrylamide gel electrophoresis was carried out according to Laemmli (25) or Fling and Gregerson (20). Following electrophoresis, gels were fixed in 30% methanol and 7.5% acetic acid for 2 h before silver staining as described by Oakley et al. (26). For radiolabeled samples, the fixed gel was soaked in En3Hance (Du Pont-New England Nuclear), dried under vacuum and exposed to Kodak XAR-5 film using an intensifier screen.

RESULTS
The clathrin-coated vesicle (H')-ATPase is highly sensitive to sulfhydryl reagents. When treated with NEM, the ATPase activity of this enzyme is irreversibly abolished. This enzyme is also sensitive to cystine. Before treatment with cystine, coated vesicles stripped of clathrin were dialyzed over 30 h to remove sulfhydryl-reducing reagents present during isolation of the coated vesicles. The dialyzed vesicles were incubated with 1 mM cystine. This resulted in complete inhibition of the proton transport (Table I). However, unlike NEM, the cystine inhibition can be reversed by dithiothreitol. Moreover, when the ATPase is inactivated by cystine, it becomes resistant to NEM. The proton transport of the cystine-treated enzyme can be restored by dithiothreitol after NEM treatment (Table I). This result indicates that the cysteine residue of the coated vesicle ATPase which is responsible for NEM inhibition is able to form a disulfide bond with the cysteine moiety of cystine through an exchange reaction. Formation of the disulfide bond thus prevents this cysteine residue from reacting with NEM. Stripped vesicles which were not treated with cystine also showed dithiothreitol-enhanced proton transport activity after incubation in a solution depleted of sulfhydryl reagents over 30 h. Presumably the coated vesicle ATPase becomes less active due to formation of internal disulfide bonds between the cysteine residues of the ATPase under moderate oxidation and can be fully reactivated by reducing these disulfide bonds. However, these disulfide bonds did not protect the ATPase from NEM inhibition. After NEM treatment, no proton transport could be restored by subsequent dithiothreitol reduction. Therefore, these cysteine residues can be distinguished from that which forms a disulfide bond with the cysteine moiety of cystine.
Formation of the disulfide bond between the cysteine residue of subunit A and the cysteine moiety of cystine is a slow process. Typically, 24-30 h are required to reach complete inhibition of the proton-pumping activity (Fig. 1). ATP (2.5 mM) protects the cysteine residue of the ATPase from reaction with cystine as it protects the enzyme from NEM inhibition ( Fig. 1). In addition to ATP, increasing the ionic strength of the medium also protects the ATPase from cystine inhibition but is less effective than ATP. This suggests that the conformation of the ATPase is effected by ionic strength and that either as a result of some conformational change or through direct stearic hindrance by ATP, certain cysteine residues are hindered from reacting with sulfhydryl reagents.
The cystine modified coated vesicle ATPase can be solubilized from the native membranes and reconstituted into artificial membranes as with the normal vacuolar ATPase except that sulfhydryl reagents must be avoided. The reconstituted ATPase-containing liposomes did not show any proton transport nor ATPase activity (Table 11). However, both activities can be restored by dithiothreitol treatment. Like the cystinetreated vacuolar ATPase on the native membrane, the enzyme purified from the cystine-treated vesicles is also resistant to NEM. About 70% of the activity can be recovered by dithiothreitol following NEM treatment. Although reaction of NEM with cysteine residues other than the cystine-reactive one causes partial inhibition of the enzyme (about 30% in our experiments, Table 11), modification of the cystine-reactive cysteine residue results in complete inhibition of the ATPase regardless of whether other cysteine residues are modified by sulfhydryl reagents. Thus only the cystine-reactive cysteine residue is responsible for NEM sensitivity.  (1 mg protein/ml) were dialyzed against 10% glycerol, 1 mM EDTA, and 20 mM HEPES (pH 7.0) for 30 h. The dialyzed stripped vesicles were then divided into two parts. One part continued to be dialyzed against the same buffer. The other part was dialyzed against 1 mM cystine, 10% glycerol, 1 mM EDTA, and 20 mM HEPES (pH 7.0). The dialysis proceeded for 24 h. Aliquots were then assayed for ATP-dependent proton transport as described under "Experimental Procedures." Where indicated, the vesicles (1 mg of protein/ml) were incubated with 5 mM NEM or 50 mM dithiothreitol for 1 h at 23 "C in the order indicated.  To determine which cysteine residues in the vacuolar (H+)-ATPase form disulfide bonds with the cysteine moiety of cystine, the cystine-reactive residues were labeled by [3H] NEM. The cystine-treated vacuolar ATPase was purified from coated vesicles and treated with NEM t o block cysteine residues in the reduced form. The cysteine moiety which forms disulfide bonds with the cysteine residues of the vacuolar ATPase was then removed by dithiothreitol. The reduced cysteine residues were subsequently labeled by [3H]NEM. The autoradiograph in Fig. 2 shows that [3H]NEM predominantly reacts with cysteine residues in the A subunit and reacts slightly with the B subunit as well. Nevertheless, only the cysteine residues in subunit A are protected from [3H]NEM labeling by forming disulfide bonds with the cysteine moiety of cystine (Fig. 2). To evaluate the total NEM-accessible cysteine residues of the ATPase, the cystine-modified ATPase was first treated with dithiothreitol to remove the cysteine moiety and then labeled with t3H]NEM after removal of  I1 ATP-dependent proton transport and ATPase activity of reconstituted uesicles containing the cystine-modified coated vesicle (H+)-ATPase Stripped vesicles were incubated with cystine until proton transport was completely inhibited as described in Table I. The cystinemodified coated vesicle (H+)-ATPase was then solubilized, purified, and reconstituted into liposomes. Proton transport and ATPase activity were assayed as described under "Experimental Procedures." Where indicated, the reconstituted vesicles were incubated with 1 mM NEM or 50 mM dithiothreitol in the order shown. The specific ATPase activity of the reconstituted vesicles which had been incubated with dithiothreitol was 9.8 pmol of ATP/min (mg of protein)" at 37 "C. dithiothreitol. It is shown that the radioactivity incorporated into the cystine-reactive cysteine residues is much less than the total NEM-accessible cysteine residues in subunit A (Fig.   2). Therefore, it appears that the cystine-reactive cysteine residues are only a small fraction of the NEM-reactive cysteine residues in subunit A.
In order to identify which cysteine residues in subunit A form disulfide bonds with the cysteine moiety of cystine, the cystine-reactive cysteine residues were labeled by fluorescein maleimide, which provides a visual marker to follow the labeled cysteine residues during isolation. Stripped vesicles were first dialyzed to remove trace sulfhydryl reagents and to let putative disulfide bonds form. The cystine-reactive cysteine residues will still be in the reduced form under these conditions (Table I, Fig. 1). Subsequently, the stripped vesicles were incubated with cystine until the ATP-dependent proton translocation was completely inhibited. The stripped vesicles were then treated with NEM to block reduced cysteine residues followed by dithiothreitol treatment to cleave the disulfide bonds between the cysteine residues and the cysteine moiety of cystine. Before labeling of the newly reduced cysteine residues, dithiothreitol was completely re- moved to allow those putative disulfide bonds to form again so that those cysteine residues involved would not be labeled. The cysteine residues which are responsible for the NEM sensitivity will not be able to form disulfide bonds with other cysteine residues under these conditions (Table I). The resulting vesicles were then treated with fluorescein maleimide to label the remaining reduced cysteine residues. The labeled ATPase was then solubilized from the vesicle membranes by C12E9 and purified by glycerol gradient centrifugation. The coated vesicle ATPase thus prepared was predominantly labeled by fluorescein maleimide in subunit A (Fig. 3A). A sample treated with NEM after removal of the cysteine moiety had no available cysteine residues to react with fluorescein maleimide, and was therefore not labeled (Fig. 3A).
The fluorescein maleimide-labeled subunit A was isolated and subjected to complete digestion by V8 protease which specifically hydrolyzes peptide bonds on the carboxylic side of glutamic acid and aspartic acid in a phosphate buffer (pH 7.8) (27). The proteolytic fragments were separated by SDSpolyacrylamide gel electrophoresis. Fig. 3B shows that among the fragments separated on the SDS gel, only one peptide with an apparent molecular mass of 3.9 kDa is labeled by fluorescein maleimide. The labeled peptide was then transferred to Immobilon and sequenced. The first 11 amino acids from the amino terminus of this peptide are listed in Table  111. It matches residues 220-230 of the amino acid sequence deduced from the cDNA encoding subunit A of the bovine vacuolar ATPase (18). The molecular mass of the labeled peptide is 3.9 & 0.1 kDa (n = 2). According to the cDNA sequence, a peptide of molecular mass 3.9 kDa beginning at residue 220 would have the following sequence: KLPANHPL LTGQRVLDALFPCVQGGTTAIPGAFGC. This peptide, after labeling by fluorescein maleimide, has a molecular mass of 3.987 kDa. A peptide which continued through the next acidic residue (Aspz7') would have a molecular mass of 5.681

TABLE I11
Amino acid sequence determined from the proteolytic fragment of subunit A which was labeled by fluorescein maleimide The cystine-reactive cysteine residue of subunit A was labeled by fluorescein maleimide as described under "Experimental Procedures." The labeled subunit A was isolated and digested by V8 protease. The proteolytic fragments were then separated on a 20% acrylamide Fling and Gregerson gel (20). The fluorescein maleimide-labeled peptide was then transferred to Immobilon and sequenced. kDa, while one which continued through GluZ7' would have a mass of 6,614 kDa, both of which are much larger than the observed molecular mass. It thus appears that the labeled fragment was generated by cleavage at Glu219 and Cys2". It is likely that addition of the negatively charged fluorescein group to C Y S '~~ has generated a cleavage site for V8 protease. T o test this possibility, we modified the cysteine residue of a peptide, CVRWKQPRTYQKL, with fluorescein maleimide. This peptide contains no glutamic and aspartic acid residues. However, digestion of the modified peptide with V8 protease resulted in cleavage of the peptide at the carboxylic side of the fluorescein maleimide-modified cysteine residue. Therefore, it is evident that it is C Y S '~~ of subunit A which reacts with fluorescein maleimide and becomes hydrolyzable by V8 protease. Moreover, CysZ4O, which is one of the 2 cysteine residues in the fragment from LyszZo to CYS'~~, is unlikely to be modified by fluorescein maleimide because such modification would result in cleavage at the carboxyl side of Cys240 by V8 digestion to generate a fragment of 2.728 kDa which is much smaller than the observed molecular mass.

DISCUSSION
Vacuolar (H+)-ATPases are characterized by their sensitivity to inhibition by sulfhydryl reagents, such as NEM (1). The cysteine residue which is responsible for NEM inhibition is located on the A subunit and is able to form a disulfide bond with the cysteine moiety of cystine through an exchange reaction which results in complete inactivation of the (H+)-ATPase. After removal of the blocking cysteine group, this residue can be selectively labeled by fluorescein maleimide. Isolation and cleavage of the labeled A subunit with V8 protease generated a single-labeled fragment of molecular mass 3.9 +. 0.1 kDa which contains 2 cysteine residues, Cys240 and CYS'~~. Cys254 is likely the only cysteine residue to form a disulfide bond with the cysteine moiety of cystine.
Comparison of the available A subunit sequences from bovine (18), Neurospora (16), carrot (15), and yeast (17), indicates that CyP4 is conserved in all four species while Cys240 is cysteine in bovine and yeast but serine in Neurospora and carrot. Because all of the vacuolar (H+)-ATPases identified have been shown to be sensitive to micromolar concentrations of NEM (l), the reactive cysteine residue would be expected to be common to all of them. In fact, only 3 cysteine residues are conserved in all of the V-ATPase A subunit sequences obtained thus far: CyP4, CYS'~~, and C Y S~~' (using the bovine A subunit residue numbers). Our data indicate that it is Cys254 which is responsible for the NEM sensitivity of the vacuolar (H+)-ATPases.
Further support for this comes from sequence comparisons with the archaebacterial and F-type ATPases. The A subunit of the V-ATPases has been shown to be approximately 50% identical to the A subunit of the archaebacterial ATPases   (15-18, 28, 29) and approximately 25% identical to the p subunit of F, (15-18, 30). The conservation is even higher in regions which are believed, from chemical labeling and mutagenesis studies of the / 3 subunit of F1, to be involved in nucleotide binding or hydrolysis (for review, see Ref. 31). Consistent with the insensitivity of both the archaebacterial and F-type ATPases to NEM, the residue corresponding to Cys254 is serine in the archaebacterial ATPases (28, 29) and valine in the F-type ATPases (30).
Residues 250-258 of the bovine A subunit correspond to the consensus sequence GXXGXGKTV, termed the Walker "A" sequence (32). This region forms a glycine-rich loop which is perfectly conserved in all of the V, F, and archaebacterial ATPases and has been shown from x-ray crystallographic studies to form part of the nucleotide binding site of adenylate kinase (33). Chemical labeling and mutational analysis of the p subunit of Fl have strongly implicated this region in nucleotide binding or hydrolysis (34-37). In addition, a 50-amino acid peptide containing this sequence has been shown to bind nucleotides (38). Thus, modification of CYS'~~, which is located in the middle of this consensus sequence, would be expected to disrupt binding of ATP. Similarly, ATP might be expected to block reaction of Cys254 with sulfhydryl reagents. These predictions are both born out for the V-ATPases since NEM labeling of the A subunit is ATP-protectable (2, 12-14) and inhibition of ATP binding to the A subunit occurs on reaction with NEM (12).
The coated vesicle (H+)-ATPase contains nine subunits. Six of these subunits (100, 73, 58, 40, 38, and 33 kDa) have cysteine residues which are able to react with NEM (Fig. 2)  or 7-chloro-4-nitrobenz-2-oxa-1,3-diazole (2). Cystine reversibly reacts with a single cysteine residue (CYS'~~) in subunit A (73 kDa). This property makes cystine a useful sulfhydryl reagent to study the various cysteine residues of the vacuolar ATPases. C y P 4 of subunit A is responsible for NEM sensitivity. As long as C Y S '~~ is modified by sulfhydryl reagents, the ATPase is completely inhibited, in spite of whether other cysteine residues are modified. On the other hand, NEM modification of cysteine residues of the coated vesicle ATPase other than Cys254 results in partial loss of activity. In addition, the coated vesicle ATPase in the native membrane has at least 2 cysteine residues which are able to form a disulfide bond under moderate oxidation. Reducing this disulfide bond enhances the proton pumping activity up to 50% (Table I). However, oxidative formation of disulfide bonds between these cysteine residues does not protect the vacuolar ATPase from NEM inhibition. Therefore, these 2 cysteine residues can be distinguished from C y P 4 which renders NEM resistance to the ATPase once protected by formation of a disulfide bond.