Characterization of the ATP-dependent proton pump of clathrin-coated vesicles.

The ATP-dependent proton pump which was previously identified in clathrin-coated vesicles isolated from calf brain (Forgac, M., Cantley, L., Wiedenmann, B., Altstiel, L., and Branton, D. (1983) Proc. Natl. Acad. Sci. U.S.A. 80, 1300-1303) is further characterized. 7-Chloro-4-nitrobenz-2-oxa-1,3-diazole (NBD-Cl) was identified as a potent inhibitor of both ATP-dependent proton uptake and Mg2+-ATPase activity of coated vesicles. Thus, incubation with 10 microM NBD-Cl for 10 min at 23 degrees caused the loss of 80% of the Mg2+-ATPase activity and 95% of the proton pumping activity. The observed protection from NBD-Cl inhibition by ATP suggests that NBD-Cl may react at the catalytic site, and reversal of NBD-Cl inhibition by 2-mercaptoethanol is consistent with reaction at either a tyrosine or cysteine residue. In addition, no stable phosphorylated intermediate was observed during turnover of the coated vesicle proton pump and neither Na+ nor K+ was countertransported by the pump during ATP-dependent proton uptake.

diazole (NBD-Cl) was identified as a potent inhibitor of both ATP-dependent proton uptake and Mg2+-ATPase activity of coated vesicles. Thus, incubation with 10 pM NBD-Cl for 10 min at 23 O C caused the loss of 80% of the Mga+-ATPase activity and 95% of the proton pumping activity. The observed protection from NBD-Cl inhibition by ATP suggests that NBD-Cl may react at the catalytic site, and reversal of NBD-Cl inhibition by 2-mercaptoethanol is consistent with reaction at either a tyrosine or cysteine residue. In addition, no stable phosphorylated intermediate was observed during turnover of the coated vesicle proton pump and neither Na+ nor K+ was countertransported by the pump during ATP-dependent proton uptake.
Various lines of evidence have suggested that exposure to low pH is the signal which activates ligand-receptor dissociation following receptor-mediated endocytosis (1-6), and a number of studies have indicated that this acidification occurs in a prelysosomal compartment (6)(7)(8)(9). We recently reported that clathrin-coated vesicles contain an ATP-dependent proton pump capable of acidifying the vesicle interior (10) and suggested that this pump is responsible for the acidification event required for ligand-receptor dissociation during receptor recycling back to the cell surface. A similar result, again in brain-coated vesicles, has also been reported by Stone et al. (11). In addition, evidence has been obtained for the existence of an ATP-dependent proton pump with essentially identical properties in endosomes derived from a number of cell types (12,13). In the present communication, we provide a more detailed characterization of the ATP-dependent proton pumping activity of clathrin-coated vesicles.

MATERIALS AND METHODS
Clathrin-coated vesicles were prepared from calf brain by the procedure of Wiedenmann and Mimms (14). As with the procedure previously employed (lo), electron microscopy indicated that >95% of the vesicles were coated (with virtually all of the clathrin-coated * This work was supported by National Institutes of Health Grant GM 26199. 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.
$ Leukemia Society Fellow. To whom correspondence should be addressed at, Department of Physiology, Tufts University School of Medicine, 136 Harrison Ave., Boston, MA 02111. 5 American Heart Association Established Investigator.
Proton transport using the ["C]methylamine trapping procedure and ATPase activity were measured as previously described (14). All experiments were carried out in 60 mM KC1, 10 mM NaCI, 10 mM HEPES (pH 7.5), 0.10 mM EGTA at 23 "C unless otherwise indicated. Proton transport activity is expressed as the difference between the ["C]methylamine trapped after 2 min in the presence of 1.0 mM ATP (tripotassium salt), 2.0 mM MgSO, and that trapped in the presence of 3.0 mM KCl, 2.0 mM MgSO,. Typically, addition of ATP resulted in the generation of a 5-7-fold increase in the amount of ["C] methylamine trapped. M$+-ATPase activity is defined as the fraction of the total ATPase activity that is resistant to 100 p~ strophanthidin (typically 7040%). The specific activity of the M$+-ATPase ranged from 0.015 to 0.025 pmol of ATP/min/mg of protein at 23 "C. Relative activities represent the activity observed at a given inhibitor concentration relative to a control which received an equal concentration of solvent (1.0% ethanol unless otherwise indicated). 22Na+ and @Rb+ trapping by coated vesicles were measured using a 10-ml Sephadex G-50 column as previously described (15). Protein concentrations were determined by the method of Lowry et al. (16) in the presence of 1.0% sodium dodecyl sulfate.
Phosphorylation by [-y-32P]ATP was carried out as follows. Coated vesicles were stripped of their clathrin coats by incubation in 5.0 mM Tris (pH 8.5), 150 mM sucrose, 0.10 mM EGTA at 23 "C for 1 h (17), followed by centrifugation of the stripped vesicles for 1 h at 100,OOO X g . This procedure resulted in loss of approximately 40% of the Me-ATPase activity, 60% of the proton transport activity, and 80% of the protein and was employed to reduce the amount of labeling due to an endogenous protein kinase activity associated with the coated vesicles (18). Essentially identical results were obtained using intact coated vesicles. Stripped vesicles were resuspended to a protein concentration of approximately 2.0 mg of protein/ml in 70 mM KCl, 10 mM HEPES (pH 7.5), 0.10 mM EGTA. Alternatively, samples were prepared which contained an equivalent amount (in ATPase activity) of purified canine kidney (Na+,K+)-ATPaseZ in 70 mM NaC1, 10 mM HEPES (pH 7.5), 0.10 mM EGTA (KC1 was used in place of NaCl with the coated vesicles to prevent phosphorylation of any endogenous (Na+,K+)-ATPase pre~ent).~ To 50-pl samples were added 2.0 mM MgSO, and 100 p~ cold ATP (prelabeling with cold ATP was employed to reduce labeling at kinase phosphorylation sites).
Following a 30-s incubation, 100 PM [ Y -~~P~A T P (5 Cilmmol) was added, and the samples were incubated for 15 s followed by either quenching with 0.90 ml of ice-cold 5% trichloroacetic acid or a 15-9 chase with 1.0 rnM cold ATP and then quenching with 5% trichloroacetic acid. Samples were spun for 2 min at 10,000 X g in a Beckman microfuge, the supernatants were discarded, and the surface of the (Na+,K+)-ATPase was purified from canine kidney by the procedure of Jorgensen (19) and had a specific activity of approximately 10 pmol of ATPlminjmg of protein at 23 "C. K+ is unlikely to activate dephosphorylation of the coated vesicle proton pump (as it does with the (Na+,K+)-ATPase) since K+ is not required for ATPase activity (10) and @Rb+ (a K+ analog) is Dot transported during ATP-dependent proton uptake by the coated vesicles (see below). pellets were washed with 0.15 M potassium phosphate (pH 2). Samples were dissolved in 50 pl of 5% 2-mercaptoethanol, 0.25 M sucrose, 0.10 M potassium phosphate (pH 4), 70 mM hexadecylpyridinium chloride, and 1% pyronin Y. Acid gel electrophoresis was carried out as described by Amory et al. (20). Samples were applied to an 8% acrylamide, 0.2% N,N"methylenebisacrylamide running gel with a 4% acrylamide, 0.9% N,N"methylenebisacrylamide stacking gel. Electrophoresis was performed in 75 mM glycine, 0.125% hexadecylpyridinium chloride, (adjusted to pH 2.9 with phosphoric acid) for 4.5 h at 30 mA. Following electrophoresis, the gel was incubated for 10 rnin in 10% isopropyl alcohol, 10% acetic acid, dried under vacuum, and exposed to Kodak XAR-5 film for 15 h at -70 "C using an intensifier screen.
Phosphorylation experiments with inorganic phosphate were also carried out using stripped vesicles (2.0 mg of protein/ml), except that the reaction was done in the presence of 200 p M 32Pi (5 Ci/mmol, neutralized with imidazole), 5 mM MgC12, 20 mM imidazole HCl (pH 7.5), 0.8 mM EGTA for 1 min at 23 'C. In order to control for phosphorylation of the (Na+,K+)-ATPase present in the coated vesicle preparation, the reaction was also carried out in the presence or absence of 200 p~ strophanthidin (which increases phosphorylation ofthe (Na+,K+)-ATPase by 32Pi (21)) and in the presence or absence of 60 mM NaCl (which reduces phosphorylation of the (Na+,K+)-ATPase by 32Pi (22)). Quenching and washing with trichloroacetic acid and acid gel electrophoresis were carried out as described above.

RESULTS AND DISCUSSION
To assist in the purification and characterization of any ion transport protein, it is clearly of value to have an inhibitor of ion transport. NBD-Cl has been previously observed to inhibit the H+-ATPases of mitochondria (23), chloroplasts (24), and bacteria (25), as well as the plasma membrane (Na+,K+)-ATPase (26). As shown in Fig. 1, NBD-Cl is also a potent inhibitor of proton transport and Mg2f-ATPase activity in clathrin-coated vesicles. Thus, reaction with 10 PM NBD-Cl for 10 min at 23 "C resulted in the loss of 95% of the proton transport activity and 80% of the Mg2f-ATPase activity. Even at 50 ~L M NBD-C1, a fraction of the M$+-ATPase activity (ie. 15%) remained uninhibited, suggesting that this 15% of the

TABLE I Effect of 2-mercaptoethunol on NBD-C1 inhibition of Mg'+-ATPase
actiuity of clathrin-coated vesicles Coated vesicles (2.0 mg of protein/ml) were incubated in 60 mM KCl, 10 mM NaCl, 10 mM HEPES (pH 7.5), 0.10 mM EGTA at 23 "C for 10 min in the presence of the ligands listed under Initial incubation conditions. Samples were then either assayed for Me-ATPase activity or were treated with 2% 2-mercaptoethanol for 30 min at 23 "C and then assayed for activity. Me-ATPase activity was assayed as described under "Materials and Methods," and each value shown represents the average of two measurements relative to the activity measured under control conditions (the numbers in parentheses are the average deviations from the mean).

Initial incubation Secondary incubation
Relative M e - total activity is not associated with proton transport in coated vesicles. The half-time of inhibition of both proton transport and M$'-ATPase activity at 10 p~ NBD-C1 was approximately 2 min (Fig. 2), as compared with half-times of 2 and 11 min at 100 p~ NBD-C1 for the (Na+,K+)-ATPase (26) and the mitochondrial ATPase (23), respectively. As can also be seen from Fig. 2, proton transport and M$+-ATPase activity are protected from NBD-C1 inhibition by the presence of 2.5 mM ATP, suggesting that NBD-Cl may be inactivating the enzyme by reaction at the catalytic site. The effect of reducing agents (ie. 2-mercaptoethanol) on NBD-Cl inhibition of the coated vesicle Me-ATPase was also studied. As can be seen from Table I, if vesicles which had been reacted with NBD-Cl were then treated with 2% 2mercaptoethanol for 30 min at 23 "C, a complete reversal of the inhibition of Mg2"ATPase activity due to NBD-C1 was observed. The reversibility of NBD-Cl inhibition by reducing agents suggests that NBD-Cl is causing inhibition of the Mg2"ATPase by reaction at a tyrosine phenolic group or a cysteine sulfhydryl group rather than a lysine amino group (26). It is interesting in light of this result to note that the sulfhydryl reagent N-ethylmaleimide has also been observed to inhibit the coated vesicle proton pump (11): Vesicles which were treated only with 2-mercaptoethanol ( Table I) showed some stimulation of M$+-ATPase activity relative to control vesicles, possibly due to a partial uncoupling effect of the 2mercaptoethanol or to a reversal of inhibition of Mg2"AT-Pase activity due to partial oxidation of essential sulfhydryl groups during purification of the coated vesicles. If 1 mM 2mercaptoethanol was included prior to reaction with NBD-C1, a partial (but not complete) protection of M$+-ATPase activity was observed ( Table I). The ability of NBD-Cl to cause partial inhibition of activity even in the presence of excess reducing agent is similar to its effect on the ATPdependent proton pump in Golgi-derived vesicles (27).
Experiments were carried out to determine whether a phosphorylated intermediate occurred during ATP hydrolysis by the clathrin-coated vesicle proton pumping ATPase. Vesicles from which the clathrin coat had been stripped (see "Materials and Methods") were prelabeled with cold ATP and then given a 15-s pulse of [ T -~~P I A T P followed by quenching or a 15-s chase with excess cold ATP and then quenching. Prelabeling with cold ATP was carried out to reduce the amount of labeling at kinase phosphorylation sites due to the endogenous protein kinase activity associated with coated vesicles (18). Samples were then analyzed using an acid gel electrophoresis system under conditions in which phosphoaspartate bonds are stable (20). As can be seen from Fig. 3, although the 96,000-dalton phosphorylated intermediate of the ((Na+,K+)-ATPase could be readily observed, no difference before and after the cold ATP chase could be seen with the coated vesicle proton pump, despite the presence of equal levels of ATPase activity in the two cases. A phosphorylated intermediate in the 60,000-dalton region could have gone undetected due to the heavy kinase labeling which occurred in this region of the gel. However, it should be noted that all cation pumps previously characterized which form a phosphorylated intermediate have catalytic subunits with molecular masses of approximately 95,000 daltons or higher (28)(29)(30)(31)(32). No rapidly chaseable label was detected in this higher molecular weight range. Alternatively, the phosphorylated intermediate of the coated vesicle proton pump may be appreciably less stable than that of the (Na+,K+)-ATPase. However, the technique described has been shown to be capable of detecting the phosphorylated intermediates of not only the (Na+,K+)-ATPase (33, 34) but also the Ca2'-ATPase (34), the gastric (H+,K+)-ATPase (30), and the plasma membrane H+-ATPases from Neurospora (31) and yeast (32).
We have also attempted to detect a phosphorylated intermediate of the coated vesicle proton pump using 32Pi in the presence of M P . This phosphorylation reaction has been observed for both the (Na+,K')-ATPase (22) and the Ca2+-ATPase (35). In the case of the (Na',K+)-ATPase, phosphorylation by 32Pi is increased by cardiac glycosides (21) and is inhibited by the presence of Na+ (which shifts the enzyme to the El conformation (22)). As can be seen in Fig. 4 However, proton pumping in coated vesicles has been shown not to be due to mitochondrial contamination by the resistance of proton transport and ATPase activity to oligomycin and aurovertin (10).
In was of interest to determine whether the coated vesicle proton pump carried out unidirectional transport of protons or whether countertransport of another cation occurred. Since Na+ and K+ are the most likely cations to be countertransported in uiuo, we have measured 22Na+ and =Rb+ (a K' analog) movement during ATP-dependent proton uptake by coated vesicles ( Table 11). The transport studies were carried out in the presence of strophanthidin and vanadate to prevent movement of 22Na+ and =Rb+ due to the (Na',K+)-ATPase activity present. Addition of ATP caused a small decrease in the amount of =Rb+ trapped, but this was abolished by the addition of the proton ionophore carbonyl cyanide ptrifluoromethoxyphenylhydrazone, suggesting that =Rb' was

TABLE I1
=Na+ and =Rb+ movement during ATP-dependent H+ transport by clathrin-coated vesicles Coated vesicles (2.5 mg of protein/ml) were equilibrated for 2 h at 23 'C in the standard buffer containing either ,'Na+ (16 pCi/ml), "Rb' (12 pCi/ml), or ["Clmethylamine (12 pCi/ml) and the additions listed under Conditions. After equilibration, vesicles were incubated for 5 min at 23 "C with either 1.0 mM ATP (tripotassium salt), 2.0 mM MgSO,, or 3.0 mM KCl, 2.0 mM MgS04, and trapping was measured as described under "Materials and Methods". The results are expressed as the average value of the ratio (+ATP/-ATP) with the numbers in parentheses the average deviation from the mean. FCCP, carbonyl cyanide p-trifluoromethoxyphenylhydrazone.

Isotope
Conditions Gradient (+ATP/-ATP) simply moving in response to a membrane potential (inside positive) generated during proton uptake (10). Surprisingly, ATP caused a small increase in the amount of '*Na+ trapped.
The observation that this increase could be eliminated by amiloride (an inhibitor of Na+/H+ antiport (39)) suggests that this small rise is due to Na' influx coupled to the ATPgenerated proton gradient via a Na+/H+ antiporter. From the permeability of the vesicles to "Na+ and =Rb+ (the half-time FIG. 5. Equilibration of coated vesicles with "Na* and 88Rb*. Coated vesicles (2.0 mg of protein/ml) were incubated at 23 "C in the standard buffer containing *,Na+ ( 0 30 pCi/ml) or "Rb+ ( 0 ; 50 pCi/ml). At the indicated times, 50-pl aliquots were assayed for trapped isotope as described under "Materials and Methods". The final level of trapping reached (approximately 0.04%/mg of protein/ ml) was the same for both ,'Na+ and =Rb+ and was very close to the equilibrated level of trapping of the more permeable species ["C] methylamine (0.036%/mg of protein/ml (10)). of equilibration at 23 "C being 17 and 19 min for ,'Na+ and 06Rb+, respectively; Fig. 5) and the rate of ATP hydrolysis, one can calculate that the transport of one ,*Na+ or =Rb+ per ATP hydrolyzed should have resulted in the generation of a substantial (i.e. greater than 2.5-fold) gradient of the transported The absence of such ion fluxes suggests that proton transport in coated vesicles is not accompanied by the countertransport of another cation.
In summary, we have identified NBD-C1 as a potent and ATP-protectable inhibitor of the coated vesicle proton pump. This pump does not appear to form a stable phosphorylated intermediate during turnover and does not appear to catalyze countertransport of another cation during proton uptake. Purification and reconstitution of this enzyme should provide additional information concerning the structure of this important proton transport system.
' The magnitude of the gradient depends in part on the proportion of coated vesicles containing a proton pump. If the rate of turnover of the proton pump is comparable to that of other transport ATPases (i.e. the (Na+,K+)-ATPase), the specific activity of the enzyme gives 1-2 copies/vesicle. Assuming a random distribution, this implies that 14-37% of the vesicles would contain no cation pump, which would result in apparent cation gradients of 2.7-and 7.1-fold, respectively.