Nucleotide binding sites and chemical modification of the chromaffin granule proton ATPase.

The purified proton ATPase of chromaffin granules contains five different polypeptides denoted as subunits I to V in the order of decreasing molecular weights of 115,000, 72,000, 57,000, 39,000, and 17,000, respectively. The purified enzyme was reconstituted as a highly active proton pump, and the binding of N-ethylmaleimide and nucleotides to individual subunits was studied. N-Ethylmaleimide binds to subunits I, II, and IV, but inhibition of both ATPase and proton pumping activity correlated with binding to subunit II. In the presence of ADP, the saturation curve of ATP changed from hyperbolic to a sigmoid shape, suggesting that the proton ATPase is an allosteric enzyme. Upon illumination of the purified enzyme in the presence of micromolar concentrations of 8-azido-ATP, alpha-[35S]ATP, or alpha-[32P]ATP subunits I, II, and IV were labeled. However, at concentrations of alpha-[32P]ATP below 0.1 microM, subunit II was exclusively labeled in both the purified and reconstituted enzyme. This labeling was absolutely dependent on the presence of divalent cations, like Mg2+ and Mn2+, while Ca2+, Co2+, and Zn2+ had little or no effect. About 0.2 mM Mg2+ was required to saturate the reaction even in the presence of 50 nM alpha-[32P]ATP, suggesting a specific and separate Mg2+ binding site on the enzyme. Nitrate, sulfate, and thiocyanate at 100 mM or N-ethylmaleimide and 7-chloro-4-nitrobenzo-2-oxa-1,3-diazole at 100 microM prevented the binding of the nucleotide to subunit II. The labeling of this subunit was effectively prevented by micromolar concentrations of three phosphonucleotides including those that cannot serve as substrate for the enzyme. It is concluded that a tightly bound ADP on subunit II is necessary for the activity of the enzyme.

The purified proton ATPase of chromaffin granules contains five different polypeptides denoted as subunits I to V in the order of decreasing molecular weights of 115,000, 72,000, 57,000, 39,000, and 17,000, respectively. The purified enzyme was reconstituted as a highly active proton pump, and the binding of N-ethylmaleimide and nucleotides to individual subunits was studied. N-Ethylmaleimide binds to subunits I, 11, and IV, but inhibition of both ATPase and proton pumping activity correlated with binding to subunit 11. In the presence of ADP, the saturation curve of ATP changed from hyperbolic to a sigmoid shape, suggesting that the proton ATPase is an allosteric enzyme. Upon illumination of the purified enzyme in the presence of micromolar concentrations of 8-azido-ATP, a-[36S]ATP, or CX-[~'P]ATP subunits I, 11, and IV were labeled. However, at concentrations of a-["P]ATP below 0.1 PM, subunit I1 was exclusively labeled in both the purified and reconstituted enzyme. This labeling was absolutely dependent on the presence of divalent cations, like Mg2+ and Mn2+, while Ca2+, Co2+, and Zn2+ had little or no effect. About 0.2 mM Mg2+ was required to saturate the reaction even in the presence of 50 nM CP[~~P]ATP, suggesting a specific and separate Mg' + binding site on the enzyme. Nitrate, sulfate, and thiocyanate at 100 mM or N-ethylmaleimide and 7-chloro-4-nitrobenzo-2-oxa-1,3-diazole at 100 PM prevented the binding of the nucleotide to subunit 11. The labeling of this subunit was effectively prevented by micromolar concentrations of three phosphonucleotides including those that cannot serve as substrate for the enzyme. It is concluded that a tightly bound ADP on subunit I1 is necessary for the activity of the enzyme.
The function of vacuolar ATPases is to pump protons into organelles of the vacuolar system of eukaryotic cells. This activity is vital for the function of several organelles, such as lysosomes, endosomes, and neurosecretory granules (1). The enzyme has been purified from clathrin-coated vesicles (2, 3), plant vacuoles (4-8), fungal vacuoles (8,9), and chromaffin granules (10,11). However, the quality of the various preparations and the small amounts of enzymes that were obtained limited the thoroughness of the studies conducted with them. Therefore, even the basic information on the subunit structure and function of these enzymes is not readily available. Recently, we reported on a purification of a proton ATPase enzyme from chromaffin granules that can be reconstituted * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
3 To whom correspondence and reprint requests should be addressed.
into vesicles with a high ATP-dependent proton uptake activity (12). The preparation contains five major polypeptides with apparent molecular weights of 115,000, 72,000, 57,000, 39,000, and 17,000. These polypeptides were denoted as subunits I to V in the order of decreasing molecular weights. Subunits I and V were shown to bind DCCD,' and subunits I, 11, and IV bind NEM with various affinities. It is the aim of this study to reveal more information on the various binding sites on the enzyme and to study the nucleotide binding sites as a first step for understanding the mechanisms of action of vacuolar ATPases.
Analytical Methods-Published procedures were used for determination of protein concentrations (13,14), purification of -y-[32P]ATP and assay of ATPase activity (X), SDS-gel electrophoresis, and fluorography in the presence of Amplify (16). ATP-dependent proton uptake was assayed by following the absorbance changes of acridine orange at 492-540 nm by Aminco DW-2a spectrophotometer as previously described (17,18). Binding of nucleotides to individual subunits was assayed by UV illumination of unmodified or azido-modified nucleotides (19,20). The reaction was performed in a 50-pl reaction mixture containing 10 mM MOPS/Tris (pH 7.0), 50 mM KCl, 2 mM MgCl', about 2 pg of the reconstituted enzyme, and the labeled nucleotide at the specified concentration and specific activity. The reaction mixture was placed in flat bottom plastic test tubes, on ice, and at a 5-cm distance from a UV lamp (UVP Inc. 0.90A) without filter. After a 10-min illumination, the reaction was terminated by the addition of 10 p1 of concentrated dissociation buffer to give a final concentration of 2% SDS and 8-mercaptoethanol. Samples of 30 pl were electrophoresed in the presence of SDS, and, following fluorography for 35S or without fluorography for 32P, were exposed to x-ray film.
Preparations-Chromaffin granule membranes were prepared from bovine adrenal glands as previously described, with the protease inhibitors pepstatin A at 2 pg/ml and leupeptin at 5 pg/ml present during all the purification steps (12,18,21). The membranes were frozen in liquid nitrogen and kept at -85°C. The proton ATPase was purified as previously described, and usually the enzyme was reconstituted by dilution immediately following the purification or after thawing the purified enzyme frozen at liquid nitrogen. Typically The control 100% activities were 4 pmol of Pi released/min/mg of protein for the ATPase activity and initial rate of acridine orange absorbance changes of 0.015 A at 492-540 nm/min for proton uptake activity. The relative labeling of the various subunits was obtained from x-ray film depicted in the right part of this figure by the method of Suissa (35). The intensity of 150 p~ NEM for each subunit was taken as 100%.
suspension was centrifuged at 200,000 X g for 60 min. The pellet was homogenized in 0.5 ml of the same buffer and stored at 0°C for up to 3 days with little loss in proton pumping activity. Occasionally, the reconstituted preparation was frozen in liquid nitrogen and stored at -85°C. This preparation could be stored indefinitely, but, after thawing, it retained less than 50% of its original activity. When NEM is to be used, DTT was omitted during the reconstitution of the enzyme.

RESULTS
One of the main characteristics of vacuolar ATPases is inhibition of their activity by NEM (1). In isolated chromaffin granules, the proton pumping activity was found to be much more sensitive to the alkylating agent than the ATPase activity (11,22,23). A high affinity binding site for NEM was recently identified on the M , = 72,000 subunit of the enzyme while sites with lower affinity were detected on the M , = 115,000 and 39,000 subunits (12). Fig. 1 shows that the labeling of the M , = 72,000 polypeptide correlates well with inhibition of both the proton pumping and ATPase activities of reconstituted vesicles. At saturation, about 2 eq of NEM were bound to the M , = 72,000 polypeptide per one enzyme molecule. Fig. 2 shows that prolonged incubation of the reconstituted enzyme with oxidized glutathione caused a complete inhibition of the proton uptake activity of the enzyme. This effect could be reversed by the addition of DTT. Moreover, the binding of glutathione protected against inhibition by NEM (not shown). This experiment indicates that a reduced "SH outside the vesicles is required for the proton pumping activity of the enzyme.
Several nucleotides can protect against inhibition by NEM. Table I depicts the effects of added nucleotides during the incubation of the reconstituted enzyme with NEM. Inclusion of AMP or Mg2+ had no effect on the inhibition by NEM.
However, high Mg2+ + AMP concentrations slightly prevent.ed the inhibition. Mg2+ + ADP was the most effective in protecting against NEM inhibition.
The observation that protection by ATP was much more pronounced in the pres-  Table  I. Both ADP and ATP prevented only about 50% of the labeling of this subunit, but fully protected against inactivation by NEM. This observation suggests that modification of only one of the two sulfhydryl groups on the M , = 72,000 polypeptide can cause a complete inhibition of the ATPdependent proton uptake activity of the enzyme. However, the possibility that the residual label is due to a population of enzyme that was not properly reconstituted cannot be excluded. The effects of nucleotides on the labeling by ["CINEM indicated the possibility that more than one nucleotide binding site is present on the enzyme. Fig. 4 shows that the presence of ADP modifies the dependence of ATPase activity on ATP concentrations. The data strongly suggest that the enzyme is an allosteric enzyme and more than one nucleotide binding site may be operating. A possible involvement of a regulatory site was also suggested by the effect of UTP and CTP on the proton pumping activity of the enzyme. While these nucleotides could not catalyze proton uptake by themselves, they stimulated proton pumping catalyzed by low concentrations of ATP up to 3-fold (data not shown). Further support for the existence of more than one nucleotide binding site was obtained by labeling of the enzyme by 8-azido-ATP and CY-[~~S]ATP. As depicted in Control, -, 100 p~ ADP were present during the assay, W. The ATPase activity was assayed by measuring the release of Pi from y-[3ZP]ATP. Less than 10% of the total ATP was hydrolyzed in each reaction tube. the labeling by the nucleotide, but DCCD had no effect.

Similar results were obtained when (~-[~'P]8-azido-ATP was used instead of CY-[~~S]ATP.
When (Y-[~'P]ATP a t concentrations below 0.1 p~ was used for the labeling of the purified or reconstituted enzyme, mostly subunit I1 was labeled (Fig. 6). At higher concentrations, subunits I and IV were also labeled (not shown). This phenomenon enabled a thorough study of a single nucleotide binding site on the enzyme. The presence of a divalent cation was necessary for the labeling of subunit 11. Mg2+ and Mn'+ were the best, and Ca'+, Co'+, and Zn'+ had much smaller effects. Even though the ATP concentration was only 50 nM, the Mg2f saturation appears to be about 0.2 mM. This observation may suggest the presence of a cation binding site on the enzyme and that Mg2+ was not required merely for the formation of Mg2+ -ATP as a substrate. Fig. 7 shows the effect of anions on the labeling subunit I1 by (Y-[~'P]ATP. While organic anions like acetate had no effect, anions that are required for the proton uptake activity of the enzyme, such as C1-and Br-, inhibited the labeling. Much stronger inhibition was obtained by anions which were shown to inhibit the proton uptake activity of the enzyme. The nitrate, nitrite, thiocyanate, and sulfate nearly prevented the labeling of subunit 11. This may suggest the presence of the anion binding site on subunit 11, but long distance effects of conformational changes cannot by excluded. Fig. 8 shows that pretreatment with NEM or NBD-Cl competed with the binding of (Y-[~'P] ATP on subunit 11, and NEM at 50 p~ and NBD-C1 a t 10 p~ decreased the labeling by about 90%. This experiment is consistent with the protection against inhibition by NEM in the presence of nucleotides.
[l4C]NBD-C1 labeled all of the ATPase subunits, and it may act as a sulfhydryl reagent with this enzyme rather than a tyrosine reagent in the eubacterial type proton ATPases.
The effect of various nucleotides on the labeling of subunit I1 by (Y-[~'P]ATP is depicted in Fig. 9. Phosphate and AMP at a concentration of 50 p~ had little effect. Pyrophosphate at the same concentration inhibited about 80% of the labeling. ATP, GTP, and AMP-PNP at 1 p~ inhibited up to 90% of the labeling, and similar effects were obtained by 10 p~ ADP and CTP. The very high affinity for nucleotide binding and the binding of CTP which cannot serve as a substrate for the enzyme, may suggest that the nucleotide binding site revealed on subunit I1 by UV illumination of (Y-[~'P]ATP is a regulatory site rather than catalytic site of the enzyme.

DISCUSSION
Several lines of evidence indicated that proton pumping into chromaffin granules is not tightly coupled to the ATPase activity of the enzyme catalyzing these reactions. It was observed that much higher concentrations of NEM or nitrate are required for inhibiting the ATPase activity of chromaffin granules than those required to inhibit proton uptake (11,12,22,23). The presence of chloride markedly enhanced the proton uptake activity, while marginally affecting the ATPase activity of the same preparation (12,24). Sulfate concentrations, which strongly inhibit proton uptake, do not affect the ATPase activity. The ATPase activity of the purified enzyme was inhibited by NEM via two distinct affinities with Ki values in the micro-and millimolar ranges (11). The loose coupling between ATPase and proton pumping may serve as a regulatory mechanism for the limited acidification required in the various organelles. The data presented in Fig. 1 show that, in reconstituted vesicles, the proton uptake and the  Table I). If this is the case, a complete elimination of NEM binding by M$+ + ADP would be expected. However, the experiment depicted in Fig. 3 did not substantiate this suggestion. The most probable explanation for this phenomenon may be the presence of two sulfhydryl groups on the M , = 72,000 polypeptides having the same high affinity to NEM, but only one of them is responsible for the inhibition of the enzyme.
The saturation curve for substrate may give some indication on the catalytic mechanism of the enzyme. Thus, the first indication that eucaryote type H+-ATPases have more than one nucleotide binding site was obtained by observing the changes of the ATP saturation curve with the chloroplast proton ATPase (25). As shown in Fig. 4, for the ATPase activity of the reconstituted enzyme from chromaffin granules, ADP changed the saturation curve of ATP from a hyperbolic to a sigmoid shape. The apparent reaction order changed from 1.0 in the absence of ADP to 1.6 in the presence of ADP, for both the ATPase and the initial rates of proton uptake activities of the enzyme (not shown). This suggests that more than a single nucleotide binding site is operating in the proton ATPase from chromaffin granules. This sugges- polypeptide. This effect together with the strong competition of ADP and the protection against inactivation by NEM suggests that ADP tightly bound to the M , = 72,000 subunit is required for the catalytic activity of the enzyme. This notion is supported by the observation that preincubation with ADP markedly enhanced the proton uptake activity of the reconstituted enzyme (not shown).
It is quite difficult to ascribe a precise function for the nucleotide binding site on subunit 11. The extremely high affinity to the nucleotides tested, revealed by the competition for labeling by a-["PIATP ( Fig. 9), may suggest a regulatory or mechanistic function for this binding site. Similar tight binding of slow turnover ADP was observed in the chloroplast ATPase (26,27). The regulatory function of this binding site is supported by the strong competition of nucleotides that cannot serve as substrates for the enzyme (see Fig. 9). Those nucleotides also enhanced the proton uptake activity of the enzyme in the presence of low ATP concentrations. Therefore, it is suggested that the nucleotide binding site on the M , = 72,000 polypeptide should be occupied in order for the enzyme t,o be catalytically active. This effect may mimic the single site catalysis and the cooperativity that was recently discovered in the mitochondrial proton ATPase; however, several other explanations may apply for the same set of observations described in this work (28-34).
The location of the active site and the mechanism of proton uptake by the vacuolar ATPases is yet to be discovered. Several questions should be solved before we are able to get a more definite idea on the function of each individual subunit in this class of ATPases.