Purification and Characterization of Lysosomal H+-ATPase AN ANION-SENSITIVE V-TYPE H+-ATPase FROM RAT LIVER LYSOSOMES*

Lysosomal H+-ATPase was purified to homogeneity from rat liver lysosomes. It is a bafilomycin Al-sensi-tive Mg2+-ATPase, which reacts with antibodies against the 16- and 70-kDa subunits of vacuolar H+-ATPase (Nezu, J., Motojima, K., Tamura, H., and Ohk-uma, S. (1992) J. Biochern. (Tokyo) 112, 212-219), and has been separated from both the N-ethylmaleim-ide (NEM)-sensitive/bafilomycin AI-insensitive Mg2+-ATPase (ATPase I) and the NEM-insensitive M$+/ Ca2+-ATPase (ATPase 11) (Hayashi, H., Arai, K., Sato, O., Shimaya, A., Sai, Y., and Ohkuma, S. (1992) Chem. Pharm. Bull. 40, 2783-2786). The purified enzyme had the subunit structure of vacuolar H+-ATPase, consisting of 110-, 70-, 56-, 42-, 39-, 34-, (32-,) and 16- kDa proteins. It had optimal activity at a pH of 7.0-8.0, with an apparently single K , value for ATP of 95 HM. It hydrolyzed ATP 2 dATP >> GTP, ITP >> UTP, but not CTP,

We and others have also detected Me-ATPase activity on lysosomal membranes (26-28). However, purification and structural clarification of the lysosomal H+-ATPase have been hampered (29,30). Only recent studies using "cold inactivation" (31) and immunoprecipitation (32) have demonstrated the presence on lysosomal membranes of putative subunit proteins for the H+-ATPase with sizes similar to those of vacuolar H'-ATPases. The reasons for the difficulties in the isolation of the lysosomal H+-ATPase are that it is highly unstable, probably due to the presence of various proteases within the lysosome, and that there is a high background of ATPases on rat liver lysosome that obscure the H+-ATPase, which constitutes only one-tenth of the total lysosomal M$+-ATPase activity (33) unless assayed under undisrupted conditions (28). The background ATPases are the NEM-sensitive but bafilomycin AI-insensitive Me-ATPase (ATPase I) and the NEM-insensitive Mg2+/Ca2+-ATPase (ATPase 11) (27,30,33). Furthermore, it is difficult to separate the H+-ATPase from ATPases I and 11.

5650
Purification of Lysosomal H+-A TPme the 16-and 70-kDa subunits of vacuolar H+-ATPases (36) for as a bafilomycin AI-sensitive Me-ATPase and suggested the biochemical and immunological identification of lysoso-that it is an anion-sensitive v-type H+-ATPase. mal H+-ATPase. In this study, we have purified and characterized, for the first time, lysosomal H+-ATPase from rat liver from Sigma. Affinity-purified anti-16 kDa antibody was obtained from antiserum raised against the N-terminal 16-residue of the poly-50

A B
peptide sequence deduced from a cDNA for the 16-kDa subunit of rat vacuolar H+-ATPase, and affinity-purified anti-70-kDa antibody was obtained from antiserum raised against a 16-residue consensus amino acid sequence of 70-kDa subunit of vacuolar H+-ATPases (36). (1 mg/ml, ~500-600 before excision, which resulted in a tritosome preparation with a milliunits of total M?-ATPase/mg of protein (~5 5 milliunits of relative specific activity (relative to homogenate) for 8-N-acetyl-Dbafilomycin AI-sensitive MC-ATPase/mg of protein)) were solubi-glucosaminidase close to 80. Protease inhibitors (pepstatin A, chylized as described under "Experimental Procedures," and the ATPase mostatin, leupeptin, and antipain, all at 5 pg/ml) were included in activity and protein content of each of the supernatants (Extract) every solution used during the course of the preparation of the and the pellets (ppt) were determined. A , MF-ATPase activity; tritosomes and membrane ghosts. For the preparation of membrane shaded bars indicate bafilomycin Al-sensitive MP-ATPase. B, pro-ghosts, tritosomes were added by pipette to 10  EDTA, 1 mM DTT, and the protease inhibitors. Solubilization of TMG-TMG were solubilized with octylthioglucoside (38). The TMG suspension which had been adjusted to 1 mg protein/ml in the solubilization buffer (20 mM Mops-TMAH (pH 7.0), 10 mM DTT, 20% (v/v) glycerol, 1 mM EDTA, and 5 gg/ml protease inhibitors) was mixed with 0.2 mg/ml asolectin and then with 1.0 or 1.5% (w/v) n-octyl-P-D-thioglucoside. Immediately thereafter, the mixture was vortexed for 15 s, incubated on ice for 10 min, and then sedimented at 106,000 X g for 1 h. The resulting supernatant was used as the solubilized fraction.
Chromatography-Solubilized lysosomal ATPases were separated by two successive chromatographic steps of fast protein liquid chromatography (FPLC). In the first step, a Mono Q anion-exchange column (Pharmacia Biotech, Tokyo, Japan) was used, followed by a TSK-gel GX4000SWx~ gel filtration column (Tosoh, Tokyo, Japan). The octylthioglucoside-solubilized membranes (0.65 mg/ml, 2 ml) were applied first onto a HR 5/5 Mono Q column equilibrated with running buffer (50 mM Tris-C1 (pH 8.0), 0.04% (w/v) polyoxyethylene 9-laurylether (CIZES), 0.04% (w/v) octylthioglucoside, 2 mM DTT, 10% (v/v) glycerol, 0.5 mM EGTA and 8 pg/ml asolectin). The sample was eluted at a flow rate of 1.0 ml/min with a three-phase salt gradient of 0-1 M NaCl in a running buffer. One-ml fractions were collected, and aliquots of these were assayed for total and bafilomycin AI-sensitive M%+-ATPase and for protein. Aliquots were also analyzed by Mg+-ATPase activity staining after electrophoresis on an native polyacrylamide (4-20%) gel and by immunoblotting after SDSpolyacrylamide gel electrophoresis (PAGE) (see below).
Immunoblotting-Proteins that had been subjected to electrophoresis on gels (usually, 15% (w/v) or 12% (w/v) polyacrylamide) were transferred to nitrocellulose filters (0.1 pm, Schleicher & Schull) or polyvinylidene difluoride membranes (0.45 gm, Millipore) at 150 mA for 1 h at room temperature in a Trans-Blot transfer cell (Nihon Eido) in transfer buffer (192 mM glycine, 25 mM Tris, and 20% (v/v) methanol) (39). The filters were washed three times with phosphatebuffered saline containing 0.05% (v/v) Tween 20, blocked overnight with 5% (w/v) nonfat dry milk in phosphate-buffered saline, and then incubated for 2 h with antibody diluted in Tris-buffered saline containing 0.1% (w/v) BSA. The filters were then washed with three changes of Tris-buffered saline containing 0.05% (w/v) Tween 20 and then incubated for 1 h with alkaline phosphatase-conjugated antirabbit IgG goat antibody in Tris-buffered saline. The filters were washed exhaustively with Tris-buffered saline. Alkaline phosphatase activity was detected using 5-bromo-4-chloro-3-indolyl phosphate and nitro blue tetrazolium.
ATPase Activity Staining-Native PAGE gels were stained for the detection of M%+-ATPase by incubation at 37 "C overnight in a substrate solution containing 1 mM ATP, 20 mM Tris-maleate (pH 8.5), 5 mM MgC12, and 1 mM Pb(N03)2. To visualize ATPase, the white Pb3(P04)2 precipitates formed were converted to dark brown PbS by incubation in 0.5% Na2S after rinsing with H20 according to the established procedure (42).
Labeling of Purified Lysosomal H+-ATPase with f4C]DCCD-The bafilomycin AI-sensitive ATPase fraction of the Mono Q column eluate (2.5 pg) was incubated with [14C]N,N"dicyclohexylcarbodiimide (DCCD) (0.6 pCi, 11.4 nmol) at 30 "C for 23 min, washed with buffer, and then subjected to the 15% SDS-PAGE by the method of Laemmli, and autoradiographed.
ATPase Assay-ATPase activity was determined by measuring the release of inorganic phosphate from ATP at 30 "C, at which temperature the ATPase activity was stable and linearly expressed for up to 40 min. The reaction medium contained 0.5 mM ATP, 40 mM Hepes-TMAH (pH 7.5), 0.5 mM MgC12 and 0.1 M KCl. The reaction was initiated by the addition of the enzyme and stopped by the addition of one volume of 0.162% (w/v) malachite green, 5.72% (w/v) ammonium molybdate in 6 N HCl, and 2.32% (w/v) polyvinyl alcohol, according to the method of Chan et al. (43). One unit was defined as 1 pmol of Pi released/min. For ATPase assay, 1-10 milliunits/ml enzyme was typically used.
Proton Pump Activity-Lysosomal proton pump activity was measured using lysosomes loaded with fluorescein-dextran obtained in the lysosome-rich fraction of a rat starved overnight after an injection of fluorescein-dextran as reported previously (7,44). The conditions were exactly the same as in the assay of ATPase activity except that 0.2 M sucrose was added to stabilize the lysosomes. Proton pump activity was indicated by the slope of fluorescence quenching after the addition of ATP.
Incorporation of Isolated H+-ATPase into Liposomes and Reconstitution of Proton Pump Actiuity-The bafilomycin A,-sensitive ATPase fraction of the Mono Q column eluate was used as the source of enzyme for reconstitution of the proton pump activity. Reconstitution procedures were as follows. The bafilomycin A,-sensitive ATPase fraction of the Mono Q column eluate (20 pg/ml) was mixed with a 20-fold excess (by weight) of asolectin liposomes and incubated on ice for 30 min. The mixture (100 pl) was added to 2 ml of assay buffer containing 20 mM Bicine-TMAH (pH 8.5), 0.15 M KCl, 2.5 mM MgC12, and stirred for 5 min at 37 "C. 1 p~ acridine orange and 2.5 mM ATP-Na2 were added to the assay buffer containing the reconstituted liposomes. Proton pump activity was determined by measuring the fluorescence quenching of acridine orange at 30 "C.
Protein-For most assays, the Amido Black/solid-phase method was performed according to Schaffner and Weissmann (45) using BSA as a standard.

Solubilization and Purification of Lysosomal H+-ATPase-
After extensive testing of detergents, we chose octylthioglucoside for the solubilization of lysosomal H+-ATPase. AS shown in Fig. lA, this detergent selectively and almost completely (>go%) solubilized the bafilomycin Al-sensitive M P -ATPase from lysosomal membranes. This ATPase constitutes only 10-15% of the total M$+-ATPase activity on tritosomal membranes. On the other hand, most of the other ATPases (ATPase I and 11) were inactivated by this extraction procedure. This resulted in an extract in which the bafilomycin AIsensitive Mg2"ATPase (specific activity of 0.108 unit/mg protein from 50-60 milliunits/mg protein on TMG) constituted about 55-65% (4-fold enrichment) of the total activity. This extract showed 70-80% sensitivity to NEM but no sensitivity to oligomycin and efrapeptin. On activity staining at pH 8.5, however, the extract showed very little bafilomycin

Purification of Lysosomal H+-ATPase
Kinetics of lysosomal H+-ATPase. A , ATPase activity. ATPase activities were measured as described under "Experimental Procedures," except that MgCl, and ATP-Na2 were added at 5 mM and at the designated concentrations, respectively. The inset shows a Lineweaver-Burk plot. B, proton pump activity. ATP-dependent acidification of lysosomes was monitored by FD fluorescence as described under "EXperimental Procedures," except that MgCl, was added at 5 mM and the concentration of ATP was varied as indicated. The initial velocity of acidification was defined by the ATP-dependent quenching of FD fluorescence (AF in 30 s after the addition of ATP) and plotted as AF(%)/30 s. A Lineweaver-Burk plot is shown as an inset. AI-sensitive M$+-ATPase (due to inhibition by Pb2+, see below), but showed two bafilomycin .Al-resistant bands of ATPase I (500-650 kDa) and ATPase I1 (360 kDa). However, a band of molecular mass 700-800 kDa was recognized on immunoblotting with anti-16 kDa (see Fig. 4 and Ref. 30) and anti-70 kDa antibodies (data not shown).
Since many attempts to purify lysosomal H+-ATPase by ultracentrifugation resulted in poorly reproducible results: we adopted a chromatographic procedure for the separation of H+-ATPase from rat liver lysosomes. Fig. 2 shows the elution profile of the octylthioglucoside extract on Mono Q FPLC. Most of the proteins were eluted in flow-through and 0.15-0.17 M NaCl fractions, where most of the bafilomycin AI-resistant ATPases were also eluted ( Fig. 2 A ) . ATPase activity staining clearly showed the presence of ATPase I and ATPase I1 in the 0.15-0.17 M NaCl fractions (Fig. 2B)  have caused some inactivation, see below and "Discussion") resulted in a fraction of bafilomycin AI-sensitive ATPase of a nominal specific activity of 0.55 unit/mg protein, which corresponded to a 10-fold purification from TMG, with a recovery of 13.4%. Immunoblotting with anti-16-kDa and anti-70-kDa antibodies (Fig. 2C) showed that the reactivity against both antibodies was localized only to the bafilomycin Al-sensitive ATPase fractions. This confirmed the presence of v-type H+-ATPase in these fractions. However, the reactivity to anti-16-kDa but not to anti-70-kDa antibodies was also observed in fractions 8-11 eluted at 0.15-0.17 M NaC1. These fractions were devoid of any bafilomycin Al-sensitive ATPase activity. This immunoreactivity most probably represents the 16-kDa subunit of the V, domain of lysosomal H+-ATPase that is devoid of the catalytic VI domain, because these fractions were also reactive with anti-39-kDa antibody (kindly supplied by Y. Moriyama and N. Nelson; data not shown) (for more details, see "Discussion").

Purification of
Lysosomal H+-ATPase Fig. 3 shows the FPLC elution profile on TSK-gel GX4000SWxL size exclusion gel chromatography of the peak bafilomycin AI-sensitive ATPase fractions of the Mono Q column eluate. The bafilomycin AI-sensitive ATPase was eluted in fractions that were well separated from the major proteins ( Fig. 3) as well as from ATPase I1 (as detected by ATPase activity staining, data not shown), and showed 95-97% sensitivity to bafilomycin AI. The bafilomycin Al-sensitive fraction also reacted with both the anti-16-kDa and the anti-70-kDa antibodies (Fig. 3B). This fraction, however, still contained AMPase activity, although the AMPase activity was insensitive to bafilomycin A, and was probably not associated with the H+-ATPase. The specific activity (and therefore the degree of purification) of the bafilomycin A1sensitive ATPase fraction was hard to estimate because of its limited availability. The overall recovery of bafilomycin A1sensitive ATPase was 8.9%.
Behavior of Lysosomal H+-ATPase on Native PAGE- Fig.  4 shows the protein migration patterns on native PAGE of the fractions obtained during the course of purification of the H+-ATPase, compared with the immunoblotting pattern. The relative amount of the 700-800-kDa protein band increased during the course of the purification of H+-ATPase. This band corresponded to the immunoblot band (anti-16-kDa antibody, Fig. 4B) of the same molecular mass, suggesting that the molecular mass of the intact H+-ATPase is 700-800 kDa. The 1,200-1,400-kDa band detected by immunoblotting may represent a dimer of the H+-ATPase protein. Size-exclusion HPLC also resulted in an apparent molecular mass of 710 kDa for the bafilomycin Al-sensitive ATPase (data not shown). However, ATPase activity staining barely detected a band corresponding to the H+-ATPase (Fig. 4A), probably because it is sensitive to Pb2+ (see below), which is required for activity staining.
Polypeptide Composition of Lysosomal H+-ATPase- Fig. 5A shows the SDS-PAGE analysis of the bafilomycin AI-sensitive ATPase fractions. The polypeptide migration patterns were similar to those of the vacuolar H+-ATPases, as suggested previously (31, 32). Specifically, the Mono Q column fraction showed a t least eight polypeptide bands of 110, 70,56, 42, 39, 34, 32, and 16 kDa. However, the most purified TSK gel fraction showed seven clear polypeptide bands (110,70,56,42,39,34,and 16 kDa) and was devoid of the 32-kDa band (Fig. 5A, lane 3 ) . The 110-kDa band was dim, probably because it is glycosylated like similar bands (110-115-kDa protein) in H+-ATPases from other sources (46) (31) was hard to detect on this autoradiogram. Enzymatic Properties of Lysosomal H+-ATPase-Because of the limited availability of the highly purified TSK gel fraction, the Mono Q fraction containing the bafilomycin AIsensitive ATPase was used for the enzymatic characterization of lysosomal H+-ATPase. The enzymatic properties of the isolated lysosomal H+-ATPase were then compared with those of the proton pump activity on intact lysosomes as measured by the fluorescence quenching of FD fluorescence under exactly the same conditions as that in the assay of ATPase activity (7,44). Since proton pump activity is not always easy to express quantitatively, it is shown as recorder traces. Fig. 6 shows the pH dependence of the ATPase activity. Optimal activity was observed over a broad pH range of pH 7.0-8.0 with little activity below pH 5.5. All subsequent ATPase and proton pump assays were performed at pH 7.5.
The substrate (ATP) concentration dependence values of the ATPase and the proton pump activities are shown in Fig.  7. Both ATPase and proton pump activities showed Michaelis-Menten type saturation kinetics showing in each case an apparently single K,, value for ATP, of 95 and 74 FM, respectively. These values are very close. Unless otherwise described, the concentration of ATP was fixed a t 0.5 mM in the following experiments.

Substrate specificity of partially purified lysosomal H'-ATPase
The bafilomycin Al-sensitive ATPase fractions of the Mono Q column eluate were incubated a t 30 "C for 40 min in assay buffers containing 40 mM Hepes-TMAH (pH 7.5), 0.1 M KCl, 0.5 mM MgC12, and one of the listed substrates (0.5 mM), or a combination, and the liberated inorganic phosphate levels were measured as described under "Experimental Procedures."

Substrate sensitive ATPase
Bafilomycin AI-Substrate In the presence of 0.5 mM ATP, ATPase activity was maximally expressed at 0.2-0.5 mM M$+ (optimal ATP/M$+ ratio 2/1), while inhibited at higher concentrations of M e . Unless otherwise described, the concentration of M$+ was fixed at 0.5 mM in the following experiments. The preliminary study on the phospholipid requirement (tested at 5 pg/ml) of the bafilomycin Al-sensitive M$+-ATPase suggested that phospholipids tested so far (including asolectin) were all effective in supporting ATPase activity, except phosphatidic acid and phosphatidylinositol, which hardly supported the ATPase activity (data not shown). Unless otherwise described, asolectin (200 pg/ml) was included in the ATPase assay mixture in the following experiments.
The nucleotide substrate specificity of the ATPase is shown in Table I and that of the proton pump is shown in Fig. 9.
The hydrolytic activity of the enzyme on different substrates decreased in the following order: ATP = 2-deoxy ATP >> AMP-PCP, and AMP-PNP were not hydrolyzed, while ADP (ICso 100 p~ at 0.5 mM M e -A T P ) and AMP-PNP were inhibitory. Other substrates including pyrophosphate and 0glycerophosphate were not hydrolyzed, either. These findings are in agreement with the properties of the proton pump activity on intact lysosomes.
The divalent cation dependence values are shown in Table  I1 and Fig. 10 for the ATPase and proton pump, respectively. M e and Mn2+ promoted the highest ATPase activity, followed by Fez+ and Co2+. Ca2+ supported ATPase activity only slightly, which, however, seemed to exceed the very low level of proton pump activity promoted by Ca2+ as detected on intact lysosomes (Fig. 10). In contrast, NiZ+ did not support ATPase activity at all, but supported proton pump activity to some extent. This suggests that the binding site(s) for inhibitory Ni2+ is located on a domain of the enzyme that is normally not exposed to the cytosol. Neither ATPase nor proton pump activity was detected in a divalent cation-free medium (EDTA medium). Zn2+, Cu2+, Pd2+, Cdz+, H e , and Ni2+ all inhibited both M$+-ATPase and Me-ATP-dependent proton pump activities, probably directly by blocking essential sulfhydryls or, in some case(s) (e.g. Ni2+), indirectly by competing with M e for ATP. Ca2+ inhibited 25% of the Me-ATPase activity and similarly inhibited the activity on intact lysosomes (28), which was reflected by the slightly inhibitory effect of Caz+ on M$+-ATP-dependent proton pump activity (Fig. 10). Table I11 shows the monovalent cation dependence of the ATPase. Little difference in effect on ATPase activity was detected among a variety of monovalent cations tested so far, including choline. This was also the case for the ATP-dependent proton pump activity (data not shown; see also Ref.

7).
The anion dependence of the ATPase is shown in Fig. 11. Some anions stimulated ATPase activity, whereas others The anion effects on the ATPase activity roughly correlated with their effects on the proton pump activity on intact lysosomes (Fig. 12). However, discrepancies between ATPase and proton pump activities are obvious for F-, NO;, and I-; ATPase-stimulating Fsupported proton pump activity only moderately, while NO; and I-effectively supported proton pump activity but inhibited isolated ATPase activities. Tentative explanations for these phenomena are as follows. Fmay have some uncoupling effect, and NO; and Imay inhibit ATPase activity only from within lysosomes. For more detailed explanations of these discrepancies, see "Discussion." Table IV summarizes the effects of inhibitors on the ATPase activity. The sensitivity profile of the ATPase to the inhibitors was quite similar to that of the proton pump on intact lysosomes (Fig. 13). Sensitivities of both ATPase and proton pump to bafilomycin A1 are shown as dose-response curves in Fig. 14A, showing ICso values of about 0.5 nM for both activities. Fig. 14B shows the dose-response curves to NEM of both the ATPase and proton pump activities. These are also parallel, with IC5o values of about 100 wM. 7-Chloro-4-nitrobenzo-2-oxa-l,3-diazole (NBD-Cl), quercetin, 4,4'-diisothiocyanatostilbene-2,Z'-disulfonic acid (DIDS), and DCCD also inhibited both the ATPase and the proton pump activities. The effect of DCCD on the proton pump activity was not as pronounced as the effects on tritosomes (8)  activity was slightly inhibited by sodium azide but was not significantly affected by ouabain, vanadate, and efrapeptin.
Reconstitution of Proton Pump Activity from Isolated H+-ATPase-The bafilomycin AI-sensitive ATPase fraction of the Mono Q column eluate was successfully reconstituted into artificial liposomes, showing Mg+-ATP-and K'-valinomycin-dependent, bafilomycin AI-and protonophore-sensitive proton pump activity as detected by fluorescence quenching of acridine orange (Fig. 15). This suggests that the lysosomal H+-ATPase were isolated, at least in part, in a reconstitutively active state.  13. Effect of various inhibitors on lysosomal proton pump. Procedures were as described under "Experimental Procedures." Extralysosomal FD fluorescence quenched by the addition of the indicated (*) inhibitors has been taken into account. Inhibitors: bafilomycin A,, 100 nM; DCCD, 25 p~; NEM, 1 mM; NBD-C1,25 p~; DIDS, 10 pM; quercetin, 100 pM; ouabain, 100 pM; vanadate, 100 pM; efrapeptin, 5 pg/ml; NaN3, 10 mM. Treatment of lysosomes with DCCD was started 10 min before the proton pump assay by preincubation at 30 "C in 0.25 M sucrose, 1 mM EDTA, and 0.1% ethanol.

DISCUSSION
Previously, we and others have shown that neutral-alkaline M$+-ATPase activity is associated with membrane ghosts derived from rat liver lysosomes (tritosomes) (26-28). Its characteristics are similar to those of the proton pump. However, there are some differences between the proton pump and the ATPase activities in sensitivity to chemicals, unless assayed under undisrupted conditions (28). Specifically, in contrast to proton pump activity, ATPase activity was only marginally inhibited by SH-reagents (for example, NEM (maximum 20% inhibition at 1 mM with IC6o = 0.1 mM) and p-chloromercuribenzoic acid (no inhibition)) and by bafilomycin AI (maximum 10% inhibition with IC,, + 1 nM).
In this study, we showed that purified lysosomal H+-ATPase (bafilomycin Al-sensitive Mg2"ATPase) is, like the lysosomal proton pump, sensitive to SH-reagents (NEM, p-chloromercuribenzoic acid (data not shown), and NBD-Cl), stilbene derivatives (DIDS), and quercetin. Its activity is expressed at a very low level with Ca2+ as the divalent cation cofactor, and it had no activity in a divalent cation-free medium ( Table 11). Although Ca2+ supported a low level of ATPase activity, its effect on proton pump activity was negligible (Fig. 10). This minor disparity between the ATPase and proton pump activities is probably due to the uncoupling effect of Caz+, as is the case for the F,Fl-ATPase (49). The extremely low ATPase activity observed in the divalent cation-free medium may represent CaZ+-ATPase activity promoted by residual Ca2+ in the medium, since it was completely abolished by the addition of EDTA. The other properties of the bafilomycin Al-sensitive ATPase also agreed well with those of the ATP-driven proton pump activity (Table V), including the K,,, value for ATP (95 PM for ATPase compared with 74 W M for proton pump), substrate specificity (driven by the nucleotide in the order of ATP > GTP >> CTP, UTP, and inhibited by ADP), insensitivity to monovalent cations, and sensitivity to anions (activated by chloride, as is the ATPase on intact lysosomes; Ref. 28), suggesting that lysosomal acidification is maintained by the bafilomycin AIsensitive ATPase separated and described in this paper.
Sensitivity to anions requires further comments, because some H+-ATPases are activated by some anions (e.g. chloride). For instance, the H+-ATPases of chromaffin granules (50), plant tonoplasts (vacuoles) (51,52), and Golgi apparatus (531, but not those of kidney microsomes (18, 19) and coated vesicles (54), have sensitivities like those of the lysosomal ATPase. There were also some discrepancies between the effects of anions on the ATPase and on the proton pump. In addition, the effects of some anions were monophasic and of others, biphasic.
First, chloride ions usually provide the necessary conductance to support continuous acidification by dissipating the membrane potential established by the electrogenic proton pump (H+-ATPase) and the co-existence of v-type H+-ATPase and chloride channels (e.g. Ref. 55) is a common feature of endomembrane compartments, including lysosomes (30). However, purified lysosomal H+-ATPase itself was activated by chloride (Fig. 11), showing that it required chloride for maximum expression of its ATP-hydrolyzing activity. This suggests the presence on lysosomal H+-ATPase of one or more binding sites for activating anions (chloride, bromide, and fluoride). This characteristic is shared by chromaffin granule and plant tonoplast H+-ATPases (52, 56).
Second, there were some discrepancies between the effects of anions on the ATPase and proton pump activities. Sensitivity to inhibition by anions (nitrate, iodide, sulfate, and sulfite) is not unique to lysosomal H'-ATPase but rather a general characteristic of vacuolar H+ -ATPases (3, 4, 13, 14). However, proton pump activity was only moderately or negligibly affected by nitrate and iodide (Fig. 10). This is probably because nitrate affects H+-ATPases from within lysosomes (56, 57), as its inhibitory effects on the lysosomal proton pump (both acidification and depolarization) emerge only after prolonged incubation with lysosomes (data not shown). However, its binding sites may be present also on the outside of lysosomes, since "cold inactivation" by nitrate is observed (31,53,58,59). Similarly, iodide may interact with the normally membrane-embedded region of the bafilomycin AIsensitive ATPase and thereby affect its activity. On the other hand, fluoride activated the bafilomycin Al-sensitive ATPase, while it supported the ATP-dependent acidification of lysosomes only marginally. This is not explained by the permeability properties of fluoride, because it did not support the ATP-dependent depolarization (A# increase) of lysosomes, either (data not shown). Therefore, the effects of fluoride might be related to uncoupling (60). Furthermore, F-may activate ATPase activity more efficiently from within lysosomes, because its activation effect on the ATPase of intact lysosomes is not strong (28). Third, the effects of some anions (SO$and SO:-) are biphasic, activating activity at low concentrations (e.g. 10 mM) and inhibiting activity at higher concentrations (e.g. 100 mM). Similar biphasic effects have been reported for sulfite on bovine kidney plasmalemma1 H'-ATPase (19). The biphasic effects suggest that these anions have two binding sites, one for activation and the other for inhibition. More precise studies, especially using reconstitution systems (56, 57), will clarify the mechanisms of the regulation of lysosomal H'-ATPase by anions.
In addition to sensitivity to anions, there are some differences between the enzymatic properties of the lysosomal H'-ATPase and of other vacuolar H+-ATPases. A single K,,, value for ATP of 95 p~ was obtained (Fig. 7), like the value for plant tonoplast H+-ATPase, but unlike that of the ATPases on coated vesicles (54) and chromaffin granules (61), which showed two K,,, values. The reason for this enzymatic difference is not clear but might be related to functional differences between vacuo-lysosomal and endo-/exocytic vesicles. In a preliminary study on the phospholipid dependence, we found that little ATPase activity was expressed with phosphatidic acid (as in kidney microsome ATPase; Ref. 18) and phosphatidyl inositol, but reasonable activity was detected with other phospholipids, including acidic cardiolipin. This unique phospholipid dependence is not inconsistent with the preference of the reconstitution of lysosomal proton pump for Escherichia coli phospholipids (rich in cardiolipin) (62). Obviously, more sophisticated investigations using reconstituted proton The structure of the H'-ATPase isolated from rat liver lysosomes resembles the vacuolar type of H+-ATPases isolated from a variety of other vacuolar organelles. It had three main subunits of molecular mass of 70, 56, and 16 kDa, with several accessory subunits including 110-, 42-, 39-, 34-, and 32(31)-kDa proteins. This is consistent with the results of immunological studies using antibodies against individual subunits of the H+-ATPase from chromaffin granules on lysosomal membrane proteins released by "cold inactivation," studies of lysosomal membranes labeled with ['4C]DCCD (31), and in immunoprecipitation studies (32). However, a putative 19-kDa subunit (31) was hard to detect in our preparations. Still to be solved is the possible involvement with lysosomal H'-ATPase of other proteins with lower molecular masses (12)(13)(14)(15) that have been suggested as components of the H+-ATPases of plant tonoplasts and kidney microsomes (18, 21).
A remarkable structural feature of our preparation of lysosomal H'-ATPase is its association with a 31 (32)-kDa subunit. This subunit protein was detected in the Mono Q column fraction, but was not observed in our final TSK gel analysis of lysosomal H'-ATPase. In the vacuolar H'-ATPase of the bovine kidney brush border, micro-heterogeneity has been observed not only in the molecular masses of the 56-and 31-kDa subunits but also in the immunoreactivity of the 31-kDa subunit in different tissues, in contrast to the reasonably uniform immunoreactivity of the 70-kDa subunit (64, 65). This suggests that the 31-kDa subunit has some role in sorting or regulating the H+-ATPase. Since the 31-kDa subunit has been detected on lysosomes biosynthetically and immunologically by the immunoblotting technique (31,32), it is plausible that the 31-kDa subunit dissociated from the proton pump during the purification procedure. Nevertheless, it is an interesting hypothesis that such behavior (loose attachment to the H+-ATPase) of the 31-kDa subunit is evidence that the subunit is concerned with sorting and regulation of the lysosomal H+-ATPase. Immunoprecipitation studies using different monoclonal antibodies (17) will help identify the true components of lysosomal H'-ATPase. In addition, quantitative reconstitution study (66, 67) will serve to clarify which components are necessary for proton pump activity.
A finding also notable from structural and regulatorypoints of view is the structures, found in Mono Q column chromatography, that are reactive to anti-16-kDa and anti-39-kDa but not to anti-70-kDa antibodies and are devoid of any bafilomycin Al-sensitive ATPase activity (Fig. 2). This structure most probably represents the V, domain of lysosomal H+-ATPase that is devoid of the catalytic V1 domain. Similar findings (excess V, domain) have recently been made on the coated vesicle membranes of bovine brain by glycerol gradient analysis of solubilized proteins (68). Detachment of the V1 domain may occur during Mono Q column chromatographic steps and may explain the low recovery of bafilomycin A1sensitive ATPase. However, we have also detected a pool of unassembled 70-kDa subunit (probably in the form of VI domain) in liver This finding is of potential interest from the view of regulation of lysosomal acidification, especially in the control of assembly of V1 and V,. Now that the properties of lysosomal H+-ATPase are clear, analyses of its assembly, the mechanism of lysosomal acidification, and the maintenance of acidic pH established by its activity can be performed. Combined studies of the characteristics of H+-ATPases in reconstituted systems and of intact lysosomes will provide data on these topics. Such studies are now in progress in our laboratory.