3 5 (cid:42) -Triphosphate Inhibits Activity of the Vacuolar (H (cid:49) )-ATPase from Bovine Brain Clathrin-coated Vesicles by Modification of a Rapidly Exchangeable, Noncatalytic Nucleotide Binding Site on the B Subunit*

It was previously observed that the B subunit of the tonoplast V-ATPase is modified by the photoactivated nucleotide analog 3 (cid:42) - O -(4-benzoyl)benzoyladenosine 5 (cid:42) triphosphate (BzATP) (Manolson, M. F., Rea, P. A., and Poole, R. J. (1985) J. Biol. Chem . 260, 12273–12279). We have further characterized the nucleotide binding sites on the V-ATPase and the interaction between BzATP and the B subunit. We observe that the V-ATPase iso- lated from bovine clathrin-coated vesicles possesses approximately 1 mol of endogenous, tightly bound ATP/ mol of V-ATPase complex. BzATP is not a substrate for the V-ATPase, but does act as a noncovalent inhibitor in the absence of irradiation, changing the kinetic charac-teristics of ATP hydrolysis. Irradiation of the V-ATPase in the presence of [ 3 H]BzATP results primarily in modification of the 58-kDa B subunit, with complete inhibi- tion of V-ATPase activity occurring upon modification of one B subunit per V-ATPase complex. Inhibition oc- curs as the result of modification of a rapidly ( t 1 ⁄ 2 < 2 min) exchangeable site, and yet this site does not corre- spond to a catalytic site, as indicated by the effects of cysteine-modifying reagents


mol of endogenous, tightly bound ATP/ mol of V-ATPase complex. BzATP is not a substrate for the V-ATPase, but does act as a noncovalent inhibitor in the absence of irradiation, changing the kinetic characteristics of ATP hydrolysis. Irradiation of the V-ATPase in the presence of [ 3 H]BzATP results primarily in modification of the 58-kDa B subunit, with complete inhibition of V-ATPase activity occurring upon modification of one B subunit per V-ATPase complex. Inhibition oc-
curs as the result of modification of a rapidly (t1 ⁄2 < 2 min) exchangeable site, and yet this site does not correspond to a catalytic site, as indicated by the effects of cysteine-modifying reagents which react with Cys 254 located at the catalytic sites on the A subunit. Thus, the noncatalytic nucleotide binding site modified by BzATP appears to be rapidly exchangeable. The site of [ 3 H]BzATP modification of the B subunit was localized to the region Ile 164 to Gln 171 , which from the x-ray crystal structure of the homologous F-ATPase ␣ subunit, is within 10 Å of the ribose ring of ATP bound to the noncatalytic nucleotide binding site. Thus, despite the absence of a glycine-rich loop region in the B subunit, these data are consistent with a similar overall folding pattern for the V-ATPase B subunit and the F-ATPase ␣ subunit.
Although still somewhat limited, information has begun to emerge concerning the nucleotide binding sites on the V-ATPases. Sequence homology (15)(16)(17)(18)(19)(20)(21)(22) indicates that the A and B subunits of the V-ATPases are related to the ␤ and ␣ subunits of the F-ATPases, which function in ATP synthesis in mitochondria, chloroplast, and bacteria (23)(24)(25)(26)(27). These proteins are thus all derived from a common, ancestral nucleotide-binding protein. Covalent modification by ATP-protectable inhibitors suggests that the A subunit possesses the catalytic nucleotide binding sites (for references, see Ref. 1), although the B subunit has also been shown to participate in nucleotide binding (28,29).
To discuss the nucleotide binding sites of the V-ATPase in more detail, it is first necessary to review what is known concerning these sites on the F-ATPases. The F-ATPases possess six nucleotide binding sites on the F 1 domain, three on the ␤ subunits, and three on the ␣ subunits (23)(24)(25)(26)(27). Chemical modification and mutagenesis studies indicate that the catalytic sites are located on the ␤ subunits, while the noncatalytic sites are located on the ␣ subunits. The catalytic sites show rapid nucleotide exchange while the noncatalytic sites exchange nucleotides only very slowly. The recent x-ray crystal structure of the bovine heart mitochondrial F 1 shows the nucleotide binding sites at the interface of the ␣ and ␤ subunits, with the three catalytic sites located primarily on the ␤ subunits and the three noncatalytic nucleotide binding sites located primarily on the ␣ subunits (30). The function of the noncatalytic sites remains somewhat uncertain, although results from several laboratories suggest that they may play a role in regulation (31,32) or assembly (33,34) of the F-ATPase complex. Other studies, however, suggest that the noncatalytic nucleotide binding sites do not need to be occupied in order for catalysis to occur (35).
Several studies have focused on the nucleotide binding sites of the V-ATPases. Covalent modification and peptide sequenc-ing have identified Cys 254 , located in the glycine-rich loop region of the A subunit, as the residue responsible for sensitivity of the V-ATPases to sulfhydryl reagents (36). The F-ATPase crystal structure indicates that the glycine-rich loop is in close proximity to the triphosphates of ATP (30). Moreover, Cys 254 was shown by disulfide bond formation to be within 5-6 Å of Cys 532 located in the C-terminal domain of the same A subunit (37). Because the corresponding positions in the F-ATPase ␤ subunit are also separated by 5-6 Å (30), these results suggest a similar overall folding pattern for the V-ATPase A and F-ATPase ␤ subunits. Labeling studies employing 2-azido-ATP support this suggestion (29).
Less is currently known concerning the nature and function of the nucleotide binding sites located on the V-ATPase B subunits. The B subunit does not possess a region homologous to the glycine-rich loop (16,18,19,22), suggesting that the nucleotide binding site located on the B subunit is noncatalytic. Site-directed mutagenesis of the yeast V-ATPase B subunit (38) indicated that partial (20 -50%) losses in activity resulted from modification of residues in either the region corresponding to the glycine-rich loop or the adenine binding pocket. By contrast, nearly complete loss of activity was observed upon mutation of residues postulated to be contributed by the B subunit to the catalytic nucleotide binding sites on the A subunit.
Previous studies have demonstrated that the tonoplast B subunit is covalently modified by 3-O-(4-benzoyl)benzoyladenosine 5Ј-triphosphate (BzATP) (28). In the current study we have further characterized the nucleotide binding sites on the V-ATPase and the nature and function of the nucleotide binding site on the B subunit that interacts with BzATP.
Assays-The protein concentration was measured following precipitation with trichloroacetic acid as described (39). ATPase activity was measured by a continuous spectrophotometric assay (40), and ATPdependent proton transport was measured using the fluorescent dye acridine orange (13).
Purification and Reconstitution of the Clathrin-coated Vesicle V-ATPase-Clathrin-coated vesicles were prepared from calf brain, stripped of their clathrin coat, and the V-ATPase solubilized with C 12 E 9 and isolated by glycerol density gradient sedimentation as described previously (13). The specific activity of the purified V-ATPase was typically 7-8 mol of ATP/min/mg of protein at 37°C. The purified V-ATPase was in some cases reconstituted into phospholipid vesicles containing phosphatidylcholine, phosphatidylserine, and cholesterol by dialysis as described previously (13).
Determination of Endogenously Bound Nucleotides Associated with the V-ATPase Using the Luciferin-Luciferase System-To determine endogenous nucleotides bound to the purified V-ATPase, a neutralized perchloric acid extract of the protein was prepared as described (41) with some modifications. To extract nucleotides, perchloric acid was added to a final concentration of 8% to a 1-ml solution of the purified V-ATPase containing 0.1-0.2 mg of protein. The solution was incubated on ice for 10 min followed by sedimentation for 10 min at 6,000 ϫ g. Perchloric acid was removed by four extractions with 3-4 volumes of water-saturated diethyl ether. Following removal of trace ether with a stream of nitrogen gas, the pH was adjusted to 7.0, and ATP present was assayed using the luciferin-luciferase system. Aliquots of the extracted nucleotides (2-10 l) were added to a mixture of luciferinluciferase obtained from Sigma containing luciferase, luciferin, human serum albumin, magnesium sulfate, EDTA, and glycine buffer. The emitted light was measured using a Monolight 2010 bioluminometer. Samples containing known amounts of ATP were used to construct a standard curve, and ATP conversion to ADP was checked using an ATP regeneration system containing pyruvate kinase, phosphoenolpyruvate, and MgCl 2 .
Synthesis  (43). The specific activity of [␥-32 P]ATP was 1,000 cpm/nmol and [␥-32 P]BzATP was 860 cpm/nmol. The hydrolysis incubation mixture contained 0.1 mg/ml purified V-ATPase, 1 mM [␥-32 P]ATP or [␥-32 P]BzATP, 2 mM MgSO 4 , 0.05% C 12 E 9 , 20 g of phosphatidylcholine/ ml, and 20 g of phosphatidylserine/ml. As a control, samples without protein were used. After 10-min incubation at 37°C, HClO 4 (1.5 M HClO 4 in 10 mM NaH 2 PO 4 ) was added to a final concentration of 0.5 M, followed by 10-min incubation on ice. Nucleotides were removed by treatment with activated charcoal suspended in 10 mM NaH 2 PO 4 followed by removal of the activated charcoal by sedimentation for 2 min at 10,000 ϫ g. An aliquot of supernatant was subjected to liquid scintillation counting with correction for 32 P i present in controls lacking the V-ATPase.
Labeling of the V-ATPase by [ 3 H]BzATP-Labeling of the V-ATPase in the continuous presence of [ 3 H]BzATP was carried out as follows. Purified V-ATPase (10 -100 g of protein/100 l) in solubilization buffer (50 mM NaCl, 30 mM KCl, 20 mM HEPES (pH 7.0), 0.2 mM EGTA, 10% glycerol) containing 0.02% C 12 E 9 and 20 g of phosphatidylcholine/ml was incubated with 10 -400 M [ 3 H]BzATP and 2 mM MgCl 2 for 10 min at room temperature in the dark followed by irradiation of the sample with a UV lamp (UVGL-25 Mineralight 254/366) at long wavelength at a distance of 1 cm at 4°C for 20 min with stirring. A 2-10 l aliquot was assayed for ATPase activity as described above and to a 40 -80-l aliquot was added 3 ϫ concentrated Laemmli sample buffer followed by SDS-PAGE on a 12% acrylamide gel, drying of the gel, and autoradiography. The wet gel was incubated for 1 h in 30% methanol, 7.5% acetic acid and then for 30 min in ENHANCE (Rapid Autoradiography Enchancer Enlightning™, DuPont). The gel was then dried and autoradiography performed. For some experiments an IS-1000 Digital Imaging System (Alpha Innotech Corp.) was used to quantitate label incorporation.
Modification of the V-ATPase A Subunit Cysteine 254 by Cystine-Cystine-modified V-ATPase was prepared as described (37), with some modifications. Stripped vesicles were suspended in 1 mM EDTA, 10% glycerol, and 20 mM HEPES (pH 7.0), which had been degassed to remove oxygen at a protein concentration of 4 mg/ml (total volume, 1.25 ml). The stripped vesicles were then dialyzed against two changes of 500 ml of the same buffer for 30 h at 4°C to remove the trace of 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 30 h at 4°C. The cystinetreated vesicles were then assayed for the ATP-dependent proton transport. As described previously, this procedure led to complete loss of proton transport activity, which was reversed upon incubation of the vesicles for 1 h at room temperature in the presence of 50 mM dithiothreitol. The cystine-inactivated V-ATPase was then solubilized and isolated as described above, except that reducing agents were eliminated from the solubilization buffer used throughout.
Isolation and Sequencing of [ 3 H]BzATP-labeled Peptide-Isolation and sequencing of labeled peptide fragments was carried out as described previously (36) with the following modifications. The V-ATPase (1 mg of protein) was labeled in the presence of 250 M [ 3 H]BzATP as described above. For this experiment [ 3 H]BzATP was synthesized from [3-3 H]4-benzoylbenzoic acid. The sample was lyophilized to reduce the volume to approximately 500 l and separated on a 12.5% acrylamide gel according to Fling and Gregerson (44). The labeled 58-kDa B subunit was identified by the following procedure. After the electrophoresis the wet gel was cut with a gel slicer into 1.5-mm-wide sections. The radioactivity in each slice was determined after incubation of 1 ⁄10 part of each section with 0.5 ml of 0.1 M NaOH for 6 h at 37°C. 50 l of 1 M HCl was added to neutralize the NaOH, and radioactivity of the samples was measured using a Beckman 1801 Liquid Scintillation System with liquid scintillation mixture (Ready Flow III, Beckman). The remaining part of each section was wrapped and kept at 4°C. The protein was electroeluted from the slices of the gel with the highest radioactivity as described previously (36). The eluted protein was dialyzed using 25-kDa 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 and salt. The samples were lyophilized to decrease the volume to 2 ml, and KCl was added to a final concentration 100 mM to precipitate the remaining SDS. To half of the sample (1 ml), EDTA and KH 2 PO 4 (pH 7.8) were added to final concentrations of 2 and 50 mM, respectively. V8 protease was then added to give a protease/protein ratio of 1:40, estimating the protein concentration from the absorbance at 210 nm relative to bovine serum albumin standards. To the remaining 1 ml was added CaCl 2 and NH 4 HCO 3 (pH 7.3) to final concentrations of 10 and 100 mM, respectively. Trypsin was then added at a 1:40 ratio to protein.
The samples were incubated at room temperature with rotation for 2-3 days, with a second, equal aliquot of protease added on day 2. The samples were then dialyzed against water for 2 days using 1-kDa cut-off dialysis tubing to remove salt. The sample was then concentrated to less than 100 l by lyophilization and separated by SDS-PAGE using a 20% acrylamide gel modified as follows for high salt samples. The separating gel contained 16 ml of 30% acrylamide, 0.8% bisacrylamide, 6 ml of 3 M Tris-HCl (pH 8.8), 240 l of 10% SDS, 240 l of 0.2 M EDTA, 12 l of TEMED, 1.4 ml of H 2 O, and 120 l of 20% ammonium persulfate. The stacking gel contained 1.5 ml of 30% acrylamide, 0.8% bisacrylamide, 1.12 ml of 0.5 M Tris (pH 6.8), 90 l of 10% SDS, 4.2 l of TEMED, 6.15 ml of H 2 O, 45 l of 20% ammonium persulfate. The running buffer contained 6 g of Tris, 14.4 g of glycine, and 1 g of SDS in water to 1 liter. After electrophoresis, a portion of the gel was sliced, and the position of the [ 3 H]-labeled peptide was identified using the same procedure as described above. The remaining, uncut portion of the gel was transferred to Immobilon (Millipore) at 4°C by electrophoresis at 70 V for 1 h. The portion of the Immobilon corresponding to the peak of radioactivity was cut out and subjected to amino acid sequence analysis using an Applied Biosystems gas-phase microsequencer.

Determination of Endogenous Nucleotides Bound to the V-ATPase-
To determine the level of endogenous nucleotides bound to the V-ATPase, the purified enzyme was extracted with HClO 4 and the extracted nucleotides quantitated using the luciferin-luciferase system as described under "Experimental Procedures." We observed 0.72 Ϯ 0.11 mol of ATP/mol of V-ATPase complex. This value assumes that the V-ATPase is 100% pure as isolated and may therefore be closer to a stoichiometry of 1:1 given the presence of low levels of contaminating proteins in the preparation. To determine whether any significant amount of endogenous ADP is also present, the samples were preincubated with an ATP-regenerating system and the quantification repeated. There was no significant difference in the amount of ATP detected, indicating that virtually all of the endogenous bound nucleotide was ATP.
BzATP Is Not a Substrate for the V-ATPase-To determine whether the V-ATPase could hydrolyze BzATP, the purified V-ATPase was incubated with 1 mM [␥-32 P]BzATP in the presence of 1 mM MgCl 2 and C 12 E 9 (see "Experimental Procedures"). No detectable 32 P i release was observed (data not shown). The specific activity measured under these conditions using [␥-32 P]ATP (3-4 mol of ATP/min/mg of protein) was approximately 2-fold less than that measured using the coupled spectrophotometric assay (see "Experimental Procedures"), due probably to competitive inhibition by ADP accumulated during ATP hydrolysis in the absence of an ATPregenerating system.
BzATP as an Inhibitor of the V-ATPase-To test whether BzATP could act as a reversible inhibitor of V-ATPase activity in the absence of irradiation, ATPase activity was measured in the presence of 0.025-1 mM ATP in the presence or absence of 0.1 mM BzATP. Fig. 1 shows that in the absence of BzATP, ATPase activity was characterized by two K m values of 125 and 500 M with corresponding V max values of 3.8 and 8.1 mol/ min/mg of protein, respectively. These values are similar to those previously reported for the reconstituted V-ATPase from coated vesicles (45). In the presence of 0.1 mM BzATP, ATPase activity was characterized by a single K m of 310 M and V max of 4.5 mol/min/mg. It is clear from this experiment that even without irradiation BzATP acted as an inhibitor of V-ATPase activity. The fact that BzATP also changed the kinetic characteristics of ATP hydrolysis makes evaluation of the type of inhibition observed somewhat complicated. Fig. 2 shows the result of UV irradiation of the purified coated vesicle V-ATPase in the presence of [ 3 H]BzATP. [ 3 H]BzATP labels mainly the 58-kDa B subunit, with some label associated with the 73-kDa A subunit, the 100-kDa subunit, the 33-kDa subunit, and a 50-kDa polypeptide, which corresponds to the 50-kDa subunit of the AP-2 adaptor complex (46,47). Quantitation of incorporation using an Alpha Innotech Digital Imaging System indicated that labeling of the B subunit was at least seven to eight times greater than that of any other polypeptide in the preparation. [ 3 H]BzATP labeling of the B subunit was dependent upon Mg 2ϩ ( Fig. 2A) and was partially prevented by ATP (Fig. 2B) Fig. 2C shows the effect of irradiation of the V-ATPase in the presence of BzATP on ATPase activity. As can be seen, UV irradiation alone caused only slight inhibition of activity, and this was prevented in the presence of ATP. UV irradiation in the presence of BzATP, on the other hand, caused significantly greater inhibition. As with labeling by [ 3 H]BzATP, this inhibition was concentration-dependent, dependent upon Mg 2ϩ , and was partially protected by ATP. The concentration dependence of labeling and inhibition are more thoroughly documented in inhibition observed at 20 min in the presence of 100 M of [ 3 H]BzATP was 65% relative to the activity of the nonirradiated sample. As can be seen from Fig. 4A, inhibition of activity does not correlate with incorporation of [ 3 H]BzATP into the 33-kDa subunit. Although incorporation into both the A and B subunits correlates with inhibition, the much lower incorporation observed with the A subunit makes incorporation into the B subunit the most likely cause of inhibition.

Photolabeling of the V-ATPase by [ 3 H]BzATP-
To further evaluate this relationship, the data were replotted as shown in Fig. 4B. As can be seen, complete inhibition is predicted to occur at approximately 1.3 mol of [ 3 H]BzATP incorporated into the B subunit/mol of V-ATPase complex. Because incorporation into the A subunit is 7-8-fold lower, there is insufficient A subunit labeling to account for the observed inhibition of V-ATPase activity. Moreover, the results indicate that modification of only one of the three copies of the B subunit per V-ATPase complex is sufficient to completely inhibit activity.

Rapidly and Slowly Exchangeable Nucleotide Binding
Sites-It was shown recently that the V-ATPase possesses both rapidly (t1 ⁄2 Ͻ 2 min) and slowly (t1 ⁄2 Ͼ 2 min) exchangeable nucleotide binding sites (29). To determine which type of nucleotide binding site was responsible for [ 3 H]BzATP inhibition, the following experiments were carried out. The purified V-ATPase was first incubated with [ 3 H]BzATP in the presence of Mg 2ϩ for 1 h followed by removal of unbound nucleotide using a Sephadex spin column. The sample was then incubated with ATP and Mg 2ϩ for 2 min to remove label from rapidly exchangeable sites, unbound nucleotides were again removed by gel filtration, and the V-ATPase was irradiated followed by assay of ATPase activity or SDS-PAGE and autoradiography. The enzyme should thus be labeled by [ 3 H]BzATP only at slowly exchangeable sites under these conditions. To label the V-ATPase at rapidly exchangeable sites, the same experiment is carried out, except that ATP is employed during the first incubation and [ 3 H]BzATP is used in the second incubation. As can be seen from Fig. 5, significant incorporation of [ 3 H]BzATP was observed only under conditions designed to label rapidly exchangeable sites (lane 2). Moreover, the inhibition observed following modification of rapidly exchangeable sites was 45% as compared with only 8% upon modification of slowly exchangeable sites. Thus, BzATP appears to inhibit the V-ATPase principally through modification of a rapidly exchangeable site.

Role of the Catalytic and Noncatalytic Sites in [ 3 H]BzATP
Binding to the V-ATPase-Previous studies (38, 29) have suggested that, by analogy with the ␣ subunit of the F-ATPase, the  3 and 4), ATP was 1 mM, and Mg ϩ2 was 2 mM). The samples were irradiated for 20 min using a UV Mineralight 254/ 366 lamp at long wavelengths. 80 l of labeled V-ATPase from each incubation mixture was separated by SDS-PAGE on a 12.5% acrylamide gel followed by autoradiography (A, B), and 5 l was assayed for ATPase activity (C). The data shown are a combination of results from several experiments where the activity of each sample is expressed relative to the activity of a nonirradiated sample (100%), which had a specific activity 6.8 -7.2 mol/min/mg of protein. B subunit of the V-ATPase forms the major portion of the noncatalytic nucleotide binding sites, but contributes several residues to the catalytic sites that are located principally on the A subunit. In addition, we have demonstrated that Cys 254 of the A subunit is located at the catalytic nucleotide binding site and that modification of this residue by NEM blocks both nucleotide binding and activity, whereas modification by thiodisulfide exchange with cystine or by oxidation blocks activity without interfering with nucleotide binding (36,37,48). We thus used selective modification of Cys 254 with NEM or cystine to determine whether inhibition and labeling by [ 3 H]BzATP resulted from modification of catalytic or noncatalytic nucleotide binding sites on the V-ATPase. As shown in Fig. 6A (36). Interestingly, reduction of the oxidized enzyme (in which a disulfide bond is formed between Cys 254 and Cys 532 of the A subunit (37)) results in a significant increase in labeling of the B subunit (Fig. 6B). This same increase in labeling is observed upon reduction of cystine inactivated enzyme (Fig. 6C). These results suggest that the noncatalytic site on the B subunit is sensitive to changes that occur in the oxidation state of cysteine residues located at the catalytic site on the A subunit. It is also interesting to note in Fig. 6A that NEM modification actually increases [ 3 H]BzATP labeling of the A subunit (the band immediately above the 58-kDa B subunit in A). The reason for this is uncertain, but it is possible that NEM modification of the catalytic site on the A subunit alters residues contributed by the A subunit to the noncatalytic sites on B in such a way as to make them more reactive toward [ 3 H]BzATP.

Isolation and Sequencing of [ 3 H]BzATP-labeled Peptides-
To determine the site of [ 3 H]BzATP labeling on the B subunit, the labeled polypeptides were first separated by SDS-PAGE followed by electroelution. The electroeluted B subunit was then subjected to proteolysis and the peptides isolated by a second round of SDS-PAGE on a 20% acrylamide gel. A portion of the gel was sliced and used for quantitation of radioactivity, while the remainder was transferred to Immobilon. Tryptic cleavage of the labeled B subunit gave a single labeled peptide of approximately 18 -20 kDa, beginning at the residue Ile 164 . V 8 protease cleavage of the labeled B subunit gave a peptide of approximately 5 kDa starting at Asp 126 and ending at either Glu 167 , Glu 168 , or Gln 171 . The overlap of these two peptides localizes the site of [ 3 H]BzATP labeling to the region between residues 164 and 171 of the B subunit, which has the sequence IYPE(EMIQ). The peptide sequenced in each case was the major peak of radioactivity on the gel, possessing at least 10-fold more radioactivity than any other labeled band. The 0.02% C 12 E 9 , and 20 g of phosphatidylcholine/ml. The sample was irradiated, and at the indicated times, 100-l aliquots were applied to the second Sephadex column. 2-l aliquots of the eluted sample were used to measure ATPase activity, as described under "Experimental Procedures." To 50-l aliquots was added 3 ϫ Laemmli sample buffer, and samples were separated by SDS-PAGE on a 12.5% acrylamide gel followed by autoradiography. The autoradiogram was analyzed using an Alpha Innotech Digital Imaging System. A, the incorporation of [ 3 H]BzATP into the A (q), B (f), and 33-kDa subunits (ࡗ) was quantitated at each time and expressed relative to the incorporation observed at 20 min (defined as 100%). Also shown is the percent inhibition observed at each time point (E) expressed relative to the inhibition observed at 20 min (defined as 100%) in order to facilitate comparison of the curves. B, the incorporation of [ 3 H]BzATP into the B subunit expressed as mol/mol of V-ATPase complex is plotted against V-ATPase activity.

FIG. 5. Labeling of the V-ATPase by [ 3 H]BzATP at rapidly and slowly exchangeable sites.
To label the V-ATPase at slowly exchangeable sites, 0.2 mg of purified V-ATPase was first incubated with 100 M [ 3 H]BzATP and 1 mM MgSO 4 in a final volume 200 l for 1 h at 4°C. Unbound [ 3 H]ATP was removed using a Sephadex G-25 centrifuge column as described in the legend to Fig. 4. The sample was then incubated with 0.5 mM ATP and 1 mM MgSO 4 for 2 min. Unbound nucleotides were removed using a second centrifuge column under the same conditions, and the V-ATPase was irradiated followed by assay of ATPase activity (2-l aliquot) or SDS-PAGE on a 12.5% acrylamide gel and autoradiography (80-l aliquot) (lane 1). To label the V-ATPase at rapidly exchangeable sites, 0.2 mg of purified V-ATPase was incubated with 0.5 mM ATP and 1 mM MgSO 4 for 1 h, followed by removal of unbound nucleotide and incubation with 100 M [ 3 H]BzATP and 1 mM MgSO 4 for 2 min. Following a second spin column to remove unbound nucleotide, the V-ATPase was photolyzed and analyzed as described above (lane 2). ATPase activity after labeling of slowly exchangeable sites was 78%, while activity after labeling of rapidly exchangeable sites was 41% relative to control samples incubated in the absence of nucleotides and without irradiation, but which had been passed through the same gel filtration columns used to remove unbound nucleotides. fraction of radioactivity recovered in each peptide was 20 -28% of the total radioactivity incorporated into the B subunit after electroelution. The recovery of radioactivity in the B subunit during electroelution was approximately 33% due to loss of both protein and label. Thus, the yield of labeled peptide to the original labeled B subunit was approximately 7-9%. DISCUSSION Among the unresolved questions for the V-ATPases is the existence and number of endogenous, tightly bound nucleotides. The F-ATPases have been shown to possess such endogenous tightly bound nucleotides, which vary in number depending upon the conditions employed to isolate the enzyme (49,24,27). These tightly bound nucleotides remain associated with the F 1 through various procedures, including gel filtration, ammonium sulfate precipitation, and treatment with activated charcoal, but could be extracted by treatment with HClO 4 (41,50). Release of tightly bound nucleotides is also enhanced by treatment of the enzyme with low pH, glycerol, high ionic strength in the presence of EDTA, and cold treatment (50 -54). Under certain conditions, mitochondrial F 1 -ATPase can be isolated containing 3 mol of tightly bound nucleotides, 2 mol bound at nonexchangeable sites and 1 mol bound at an exchangeable site (55).
We observe that the V-ATPase isolated from clathrin-coated vesicles contains approximately 1 mol of endogenous tightly bound nucleotide/mol of V-ATPase complex. Because the nucleotide is present as ATP rather than ADP, and because the vesicles are initially isolated in the presence of MgCl 2 , it is most likely that the endogenously bound nucleotide is present at a noncatalytic site, since otherwise ATP hydrolysis might have been expected. While the level of endogenously bound nucleotides for the V-ATPase is lower than that observed for the F-ATPases, it is possible that one or more endogenously bound nucleotides were lost during isolation of the V-ATPase. This is particularly possible given the fact that the enzyme is isolated in the presence of glycerol, which is known to promote dissociation of tightly bound nucleotides from F 1 . Nevertheless, the V-ATPase containing a single mole of endogenously bound nucleotide is active, and there is no indication that the activity of the V-ATPase is reduced on exposure to glycerol. On the contrary, glycerol is observed to considerably enhance the level of ATPase activity observed following detergent solubilization of the V-ATPase, presumably through stabilization of the solubilized enzyme.
Our results show that [ 3 H]BzATP is a photoactivated inhibitor of the V-ATPase from clathrin-coated vesicles and that this inhibition is dependent upon Mg 2ϩ and partially protected by ATP. Irradiation in the presence of [ 3 H]BzATP results predominantly in labeling of the 58-kDa B subunit, and this labeling, like inhibition, is stimulated in the presence of Mg 2ϩ and decreased in the presence of ATP.
Our results are similar to those reported for the tonoplast V-ATPase (28), although the concentration of [␣-32 P]BzATP, which gave saturating labeling of the 57-kDa subunit (10 M), was much less than for the coated vesicle V-ATPase (Ͼ200 M). Moreover, no evidence was presented in the previous study to indicate that covalent labeling of the V-ATPase by [␣-32 P]BzATP led to inhibition of activity, and no information about the structure and function of this nucleotide binding site was provided.
In our experiments BzATP did not act as substrate for the V-ATPase. Nevertheless, BzATP did act as an inhibitor of V-ATPase activity, even in the absence of irradiation. As reported previously (45), the V-ATPase from clathrin-coated vesicles displays two kinetically distinguishable sites for ATP hydrolysis, with K m values reported in the current study of 125 and 500 M ATP and corresponding V max values of 3.8 and 8.1 mol of ATP/min/mg of protein, respectively. The presence of BzATP (100 M) changed the kinetics of ATP hydrolysis such that a single K m of 310 M and V max of 4.5 mol of ATP/min/mg of protein were now observed. It is possible that BzATP inhibits either the high or low affinity ATP sites (each of which contribute approximately half of the observed maximal velocity at saturating ATP concentrations) and simply changes the affinity for ATP of the remaining site. Nevertheless, it is clear that BzATP is not simply a competitive inhibitor of the V-ATPase, since the original maximal velocity cannot be achieved in the presence of BzATP, even at saturating ATP. This behavior is somewhat different than for the tonoplast V-ATPase (28), where it was suggested that BzATP was "somewhat like" a competitive inhibitor given the fact that the same maximal velocity was approached even in the presence of the inhibitor at sufficiently high ATP concentrations.
These results are very different than those obtained for the F-ATPases. In that case, BzATP acts as a substrate and a competitive inhibitor for the mitochondrial F-ATPase, with a specific activity 10 -200-fold less than for ATP (42). In addition, BzATP binds predominantly to the ␤ subunit, which possesses the catalytic nucleotide binding sites (42, 56 -58).
A plot of the incorporation of [ 3 H]BzATP into the B subunit versus activity indicates that complete inhibition is achieved upon modification of approximately 1 mol of B subunit/mol of V-ATPase complex. The amount of the radioactivity incorpo- In A, V-ATPase was reconstituted into liposomes (as described under "Experimental Procedures") and treated as follows before labeling with [ 3 H]BzATP. The reconstituted V-ATPase (fully active and reduced) was dialyzed against thoroughly deoxygenated solubilization buffer for 4 h at 4 C to remove 2-mercaptoethanol employed during reconstitution. An aliquot (containing 6 g of protein) was labeled with [3H]BzATP as described below (lane 1), and an equal aliquot was first treated with 1 mM NEM for 30 min at room temperature prior to labeling (lane 2). In B, reconstituted V-ATPase was first oxidized by dialysis against 20 mM HEPES (pH 7.0), 0.2 mM EGTA, 10% glycerol containing normal levels of oxygen for 2 days at 4°C. An aliquot (containing 6 g of protein) was labeled with [ 3 H]BzATP (lane 1), and an equal aliquot was reduced by incubation with 20 mM dithiothreitol for 30 min at room temperature prior to labeling (lane 2). In C, the V-ATPase in stripped vesicles was inactivated with 1 mM cystine as described under "Experimental Procedures" followed by solubilization and purification of the cystine-inactivated V-ATPase. An aliquot (containing 5 g of protein) was reacted with [ 3 H]BzATP (lane 1) and an equal aliquot was reduced by treatment with 20 mM dithiothreitol as described above prior to labeling (lane 2). Relative to the fully reduced enzyme, the V-ATPase activity of the NEM-modified enzyme was 14%, that of the oxidized enzyme was 11%, and that of the cystine-modified enzyme was 8%. Dithiothreitol restored complete activity of both the oxidized and cystine-modified enzymes. For all panels, the V-ATPase was labeled with 100 M [ 3 H]BzATP for 20 min followed by SDS-PAGE and autoradiography as described under "Experimental Procedures." rated into the A subunit was 7-8-fold less than that incorporated into the B subunit, so that even though inhibition of V-ATPase activity parallels incorporation into both the A and B subunits, it is very unlikely that activity is inhibited as a result of modification of the A subunit.
As has been observed previously for the F-ATPases (23)(24)(25)(26)(27), we have recently demonstrated that the V-ATPases possess nucleotide binding sites that exchange at different rates (29). Using modificaton by 2-azido-ATP we were able to distinguish between rapidly exchangeable sites (those which exchanged nucleotide with a t1 ⁄2 Ͻ 2 min) and slowly exchangeable sites (which showed a t1 ⁄2 for exchange of Ͼ2 min). Inhibition of V-ATPase activity due to covalent modification by [ 3 H]BzATP occurred upon modification of a rapidly exchangeable site. Nevertheless, the subunit labeled under these conditions was the B subunit.
It is clear from the recent x-ray crystal structure of the bovine heart mitochondrial F 1 that although the catalytic sites are located principally on the ␤ subunits, there are some ␣ subunit residues that are contributed to these sites (30). It is therefore necessary to determine whether the observed labeling of the B subunit by [ 3 H]BzATP is occurring at catalytic or noncatalytic sites of the V-ATPase. To distinguish between these sites on the V-ATPase, we employed modification by NEM, cystine, and oxidation. We have demonstrated previously that cystine modification of Cys 254 or disulfide bond formation between Cys 254 and Cys 532 at the catalytic sites is able to block ATPase activity without inhibition of nucleotide binding at these sites (36,37). By contrast, modification of Cys 254 with NEM blocks both ATPase activity and nucleotide binding to these sites.
When the effects of these reagents on BzATP labeling were tested, it was found that NEM had no effect on BzATP modification of the B subunit, whereas oxidation or cystine modification inhibited BzATP labeling. These results clearly indicate that modification of the B subunit by BzATP is not occurring through occupancy of a catalytic site, but rather a noncatalytic site. Taken together with the data described above, these results indicate that the V-ATPases possess a rapidly exchangeable, noncatalytic nucleotide binding site on the B subunit whose modification by BzATP results in inhibition of activity. This site is therefore qualitatively different from the noncatalytic nucleotide binding sites on the F-ATPases, which exchange nucleotides only very slowly.
The decrease in labeling of the B subunit sites observed upon modification of Cys 254 of the A subunit with cystine or on oxidation also indicates that nucleotide binding to the noncatalytic sites is sensitive to changes which occur at the catalytic sites. This is important as we have suggested previously that disulfide bond formation between Cys 254 and Cys 532 at the catalytic sites serves as a mechanism of regulation of V-ATPase activity in vivo (48,37). This cross-talk between the A and B subunit nucleotide binding sites suggests the possibility that nucleotide binding or hydrolysis at the A subunit catalytic sites may also be sensitive to changes which occur at the noncatalytic sites on the B subunit. This is further supported by the observation that modification of a single noncatalytic site on the B subunit is sufficient to completely inhibit ATP hydrolysis by the V-ATPase complex. Additional evidence for a cooperative interaction between nucleotide binding sites on the V-ATPase complex has come from studies of the yeast V-ATPase, where release of tightly bound [␥-32 P]ATP was accelerated by addition of ATP (59), and by the decreased 18 O exchange observed at increasing ATP concentrations (60), although in these cases the subunit location of the sites involved was not identified.
We identified the region labeled by [ 3 H]BzATP as a IYPE(EMIQ), starting at Ile 164 and ending at Glu 167 , Glu 168 , or Gln 171 . We did not attempt to identify a single modified amino acid residue within this sequence as the benzophenone radical generated is not specific for the type of reside modified (56). Sequence alignment indicates that the corresponding region of the yeast V-ATPase B subunit (Ile 145 -Glu 148 ) is perfectly conserved (19) and that this region corresponds to the sequence ISVR beginning at Ile 140 in the bovine heart mitochondrial F-ATPase ␣ subunit (61). From the crystal structure of F 1 (30), this peptide is located at the interface between the ␣ and ␤ subunits in a random coil stretch that links a ␤ sheet stretch near the top of the molecule (that ends with residue 130) to ␤ sheet 3 (residue 163-170) that is immediately upstream of the glycine-rich loop. This places the Ile 140 -Arg 143 peptide within 10 Å of the ribose ring of the ATP bound at the noncatalytic site on ␣. Thus, labeling of this region in the B subunit is consistent with a similar overall folding pattern of the ␣ and B subunits. Interestingly, the glycine-rich loop region itself, while present in the ␣ subunit, is not conserved in the V-ATPase B subunit (16,18,19,22). The corresponding region in the bovine B subunit (SAAGLPHN beginning at Ser 174 ) is almost perfectly conserved in the yeast B subunit, but bears no similarity to the GX 4 GKT sequence of the ␣ subunit. Mutagenesis studies have indicated that the glycine-rich loop sequence of the ␣ subunit is crucial for activity of the E. coli F-ATPase (34). Mutagenesis studies of the corresponding region in the yeast V-ATPase B subunit suggest that, while bearing no sequence similarity to the ␣ subunit, changes in this region do result in marked, if less dramatic, decreases in ATPase activity and proton pumping (38). Thus, results from both the previous and current studies suggest that changes (either through mutagenesis or chemical modification) in the noncatalytic nucleotide binding sites on the B subunit result in changes in activity of the V-ATPase complex.