N,N’-Dicyclohexylcarbodiimide-binding Proteolipid of the Vacuolar H+-ATPase from Oat Roots*

The inhibitor N,N’-dicyclohexylcarbodiimide (DCCD) was used to probe the structure and function of the vacuolar H+-translocating ATPase from oat roots (Avena sativa var. Lang). The second-order rate constant for DCCD inhibition was inversely related to the concentration of membrane, indicating that DCCD reached the inhibitory site by concentrating in the hydrophobic environment. [ 14C]DCCD preferentially labeled a 16-kDa polypeptide of tonoplast vesicles, and the amount of [“CIDCCD bound to the 16-kDa peptide was directly proportional to inhibition of ATPase activity. A 16-kDa polypeptide had previously been shown to be part of the purified tonoplast ATPase. As predicted from the observed noncooperative inhibition, binding studies showed that 1 mol of DCCD was bound per mol of ATPase when the enzyme was completely inactivated. The DCCD-binding 16-kDa polypeptide was purified 12-fold by chloroform/methanol extraction. This protein was thus classified as a proteolipid, and its identity as part of the ATPase was confirmed by positive reaction with the antibody to the purified ATPase on immunoblots. From the purification studies, we estimated that the 16-kDa subunit was present in multiple (4-8) copies/holoenzyme. The purification of the proteolipid is a first step towards testing its pro- posed role in H+ translocation.

dependent proton pumping are inhibited by N-ethylmaleimide, Nbd-Cl,l DIDS, and DCCD (9,10). These properties are not only shared by the vacuolar ATPase of many other higher plants and of fungi (3), but also by H'-ATPases associated with endosomes (7), such as clathrin-coated vesicles (11) of animal cells. In spite of this, it is still unclear how similar the plant and animal enzymes are at the molecular level.
[14C]Nbd-C1 specifically labeled the 72-kDa subunit of the oat tonoplast ATPase (18); this labeling could be protected by the substrate ATP or a potent competitive inhibitor, TNP-ATP, indicating that the 72-kDa subunit contains a substratebinding site. The antibody to the 72-kDa polypeptide specifically inhibited the tonoplast ATPase and H' pumping activities, consistent with the idea that the 72-kDa polypeptide contains a catalytic domain (18). Similar results had been reported for the vacuolar ATPase from corn coleoptiles (19) and Neurospora (17). In red beets, Bz-ATP bound to the 57-kDa polypeptide of the tonoplast ATPase, suggesting that this subunit also had a nucleotide-binding site (12). Since Bz-ATP was not a simple competitive inhibitor of the ATPase, Manolson et al. (12) concluded that this binding site might be a regulatory site.

EXPERIMENTAL PROCEDURES
Plant Material-Oat seeds (Avena sativa L. var. Lang) were germinated in the dark over an aerated solution of 0.5 mM CaS04. After 3 or 4 days of growth, the roots were harvested.
Preparation of Tonoplast Vesicles, Tonoplast ATPase, and Mitochondria-Tonoplast vesicles were prepared from oat roots by the procedure of Churchill and Sze (8) with minor modifications (10). The microsomal pellet was further separated by centrifugation on a 6% dextran cushion. Low-density vesicles collected from the turbid 0/6% (w/w) dextran interface were referred to as tonoplast vesicles (2,9). The vesicles could be stored at -70 "C for several weeks without loss of ATPase activity. The ATPase was purified by gel filtration (Sepharose CL-GB) after solubilization of tonoplast membrane proteins with 5% Triton X-100 (10).
Mitochondria obtained in the 8,000 X g pellet were further purified on sucrose step gradients consisting of 0.6, 1.2, 1.5, and 1.8 M sucrose (21). More than 90% of the ATPase activity of the 1.2/1.5 M interface was sensitive to 0.2 mM azide. Less than 1% of the activity originated from the tonoplast ATPase as judged by azide-resistant and nitratesensitive activity.
ATPase Assay-The release of ADP was monitored by measuring NADH oxidation spectrophotometrically at 340 nm in a coupled lactate dehydrogenase-pyruvate kinase reaction and ATP-regenerating system (10) at 25 "C in a 0.5 ml of reaction mixture. All assays contained 0.1 mM ammonium molybdate to inhibit any contaminating acid phosphatase activity. The mitochondrial ATPase was assayed in the presence of 0.05% Triton X-100 and determined as azide-sensitive activity (f0.2 mM sodium azide) (9). The membrane-bound tonoplast ATPase was assayed in the presence of sodium vanadate (200 p~) and sodium azide (200 p~) to inhibit activity from the plasma membrane and mitochondrial ATPases, respectively. Gramicidin (2.5 pg/ml) was added to prevent any inhibitory effect of the proton electrochemical gradient (2). The assay was initiated by addition of 10-15 pg of vesicle protein. After 4 min, 40 mM BTP-NO, (pH 7) was added. The tonoplast ATPase was routinely expressed as NO;-sensitive activity. Generally, more than 75% of the activity was NO;sensitive.
To measure inhibition of the tonoplast ATPase by DCCD, tonoplast vesicles were diluted with resuspension buffer (2.5 mM Hepes-BTP (pH 7.4), 250 mM sorbitol, 0.5 mM DTT) to 200 or 400 pg protein/ml. DCCD (0-80 p~) was added and the tubes incubated at 20 "C for various periods. Aliquots (10-15 pg of protein) were diluted 20-fold into the assay mixture. Further inactivation by DCCD was quenched by addition of 0.05% bovine serum albumin or 0.1 mg/ml soybean phospholipids. These additions did not reduce ATPase activity.
Protein concentrations were estimated by the method of Lowry et al. (22) as modified by Bensadoun and Weinstein (23).
To determine ["CIDCCD binding to the 16-kDa polypeptide, tonoplast vesicles (150 pg protein/ml) were incubated with [14C]DCCD (5-500 nmol of DCCD/mg of protein) at 0 or 20 "C. After incubation, 150 p1 (22.5 pg of protein) were washed with acetone and trichloroacetic acid as described above. The proteins were separated by SDS-PAGE, and bound ['4C]DCCD was visualized by fluorography. The molecular masses of the DCCD-binding proteins were estimated by use of radiolabeled standards. Gel regions corresponding to the 16-kDa peptide (determined by alignment of the exposed fluorogram) were excised, solubilized in tissue solubilizer, and the amount of radioactivity was determined by liquid scintillation counting (efficiency = 88%). To analyze the relationship between binding and inactivation, aliquots (50 pl = 7.5 pg of protein) of the same incubation mixtures were also analyzed for ATPase activity as described above.
Partial Purification of the DCCD-binding Proteolipid-Tonoplast vesicles (3-10 mg protein/ml) were injected into 20-25 volumes of chloroform/methanol (volume ratio 2/1 or 6/11, containing 5 mM DTT or butylated hydroxy toluene. The solution was vigorously stirred for 30 min at room temperature, transferred to Corex tubes, and centrifuged at 20,000 X g for 10 min. The chloroform/methanol insoluble material was called PI. The chloroform/methanol phase (SI) was mixed with 4 volumes of diethyl ether (precooled to -20 "C), incubated at -20 "C for at least 1 h, and centrifuged at 20,000 X g for 15 min at -20 'C. The pellet (containing the proteolipid fraction) is referred to as P,, the supernatant as S,. All fractions were washed with acetone and trichloroacetic acid as described above, dried under vacuum, solubilized in electrophoresis buffer, and analyzed by SDS-PAGE. The protein concentrations of the fractions were determined using the Amido Black method (26).
To detect [14C]DCCD-bound polypeptides, tonoplast membranes (200 pg/ml) were incubated with 300 nmol of [14C]DCCD/mg protein for 30 min at 20 'C. The vesicles were then diluted at least 20-fold with resuspension buffer and spun at 130,000 X g (SW 28, rms.) for 90 min at 4 "C. The tonoplast vesicle pellet was taken up in resuspension buffer to a final concentration of 2 mg/ml and extracted with organic solvents as described above. The protein composition and relative DCCD labeling in the fractions were analyzed by SDS-PAGE and fluorography.
The proteolipid fraction from mitochondria was extracted using the same procedure (chloroform/methanol2:1).
Gel Electrophoresis and Fluorography-Samples were boiled 2-4 min in electrophoresis buffer (25) containing 8 M urea and 100 mM DTT. The high concentration of urea was necessary to reduce aggregation of the DCCD-binding proteins (see Fig. 5). Proteins were separated by SDS-PAGE (80 X 100 X 0.5 mm) according to Laemmli (25) as described previously (10) and silver-stained (27) or prepared for fluorography. Fixed gels were washed twice (10 min) in deionized water, equilibrated in Autofluor for 40 min, and dried. Gels were exposed to preflashed Kodak XAR-5 film for 1-3 weeks at -70 "C.
Antibody Production and Zmmunoblotting-The purified ATPase was first treated with Amberlite XAD-2 to remove Triton X-100 and then dialyzed and lyophilized. Polyclonal antibodies to the Sepharosepurified tonoplast ATPase (250 pg of protein in Freund's complete adjuvant) were raised in New Zealand White rabbits as described (18). Booster shots were given at 3-week intervals. After the titer built up, blood was collected, and the serum was stored at -70 "C.
To detect immunoreactive polypeptides, unstained SDS-polyacrylamide gels were blotted onto diazophenylthioether paper as described (28,29). The paper was incubated with anti-ATPase (dilution 1:400) for 3 h at 22 "C, and the blots were thoroughly washed and probed with goat anti-rabbit IgG conjugated to alkaline phosphatase. The latter activity was detected by the colored product (indigo) after hydrolysis of 5-bromo-4-chloro-3-indoyl phosphate.
Reagents-[I4C]DCCD was obtained from Research Products International. "Autofluor" and tissue solubilizer "TS 1" were purchased from National Diagnostics. Other reagents were from Sigma or were analytical grade.

Kinetics of DCCD Inhibition of the Tonoplast H+-ATPme-
Incubation of the membrane-bound ATPase with DCCD resulted in rapid inactivation of the enzyme activity (Fig. U).
The pseudo-first-order rate constants for DCCD inhibition determined from the semilog plot of Fig. 1A were proportional to the DCCD concentration (Fig. lB), indicating noncooperative inhibition by 1 mol of DCCD/mol of ATPase (30). The second-order rate constants for DCCD inhibition (K") determined from Fig. 1B were inversely proportional to the concentrations of membranes, being 1225 M-' min" for 200 pg protein/ml; and 638 M" min" for 400 pg protein/ml. Identical second-order rate constants were obtained for both protein concentrations when the concentration of DCCD was expressed as nmols/milligrams of protein. Thus, the rate of enzyme inactivation was not dependent on the DCCD concentration in molar units, but rather in units of nanomoles of DCCD/milligrams of membrane protein, which are used in subsequent experiments. The dependence of the second-order rate constant on the concentration of vesicles suggests that DCCD reached the inhibitory site by partitioning into the DCCD Binding to the 16-kDa Subunit of the ATPase Is Directly Proportional to Inhibition of ATPase Activity-Although ["CC]DCCD labeled the 16-kDa polypeptide of the purified tonoplast ATPase (lo), the relationship between binding and inactivation has not been studied previously. T o determine whether the inactivation of the ATPase was caused by covalent DCCD binding, the effect of incubation time on DCCD binding to membranes was compared to the inhibition of ATPase activity (Fig. 2). Both binding and activity assays were conducted using the same tonoplast vesicles and under the same experimental conditions. After 30 min incubation with low concentrations of DCCD (29-63 nmol/mg), there was a fairly good correlation between DCCD binding and ATPase inhibition. After incubation for 30 min with 486 nmol of DCCD/mg of protein, the ATPase activity was completely inhibited (Fig. 2B); however, the binding of DCCD to membranes continued to increase (Fig. 2 A ) . Since DCCD forms covalent bonds with hydrophobically located carboxyl groups (32), the results suggest that DCCD labeled numerous tonoplast proteins unselectively at high concentrations.

DCCD-binding Proteolipid
T o determine which polypeptides were labeled by ["C] DCCD, the tonoplast membrane proteins were separated by gel electrophoresis and analyzed by fluorography. Most of the radioactivity was associated with a 16-kDa polypeptide after labeling 30 min with low DCCD concentrations (80 nmol/mg protein, Fig. 3). After 30-min incubation a t 150 or 500 nmol of DCCD/mg of protein, many other polypeptides were labeled. We have found that labeling was more specific for the 16-kDa polypeptide a t 0 "C than at 20 "C when tonoplast vesicles were incubated with [14C]DCCD to give the same degree of inhibition of ATPase activity (data not shown). Since previous studies showed that a 16-kDa polypeptide is associated with the purified tonoplast ATPase (lo), the region of the gel corresponding to the 16-kDa was excised, solubilized, and the radioactivity was directly measured. Fig. 4A shows that with increasing DCCD concentrations, [14C]DCCD bound to the 16-kDa polypeptide increased as ATPase activity decreased. A strong correlation (0.97) between inactivation of ATPase and DCCD binding to the 16-kDa polypeptide was observed (Fig. 4B). This correlation was independent of the DCCD concentration and the length or the temperature of incubation. These results strongly support the notion that inhibition of the tonoplast H+-ATPase results from the co- We estimated the number of DCCD bound per ATPase complex required to confer maximum inhibition. Fig. 4B shows that about 0.08 nmol of DCCD was bound to the 16-kDa subunit per mg of tonoplast membrane protein when ATPase activity was completely inhibited. According to the 17-to 25-fold purification (lo), we estimated that the tonoplast ATPase made up about 4-6% of the total membrane protein. Assuming a molecular mass of about 500 kDa for the tonoplast ATPase (10,13), there would be about 0.1 nmol of ATPase/mg of membrane protein. These estimates indicate that only 1 mol of DCCD bound per mol of ATPase is sufficient for complete inactivation.
Extraction of the DCCD-binding Subunit with Chloroform/ Methanol-As a first step towards understanding the properties of the DCCD-binding subunit and its proposed function in H' translocation (lo), we have started to purify this polypeptide. Since the DCCD-binding 16-kDa polypeptide might be analogous in structure and function to the 8-kDa subunit of the mitochondrial FIFO-ATPase, organic solvent extraction procedures were tried. The butan-1-01 extraction method, successfully applied for the purification of the DCCD-binding proteolipid of rat liver mitochondria and the proteolipid from lettuce chloroplast ATPase (33,34), did not extract any proteolipid from tonoplast vesicles as judged by silver-stained polyacrylamide gels (results not shown). The chloroform/ methanol procedure described by Sebald et al. (35) for the purification of the mitochondrial proteolipid from yeast and Neurospora crassa was effective in extracting a proteolipid fraction from tonoplast vesicles. A simplified version of this method (as described under "Experimental Procedures") proved to be more satisfactory. The chloroform/methanol ratio (v/v) was varied from 2:l to 6:l. Although all the treatments were effective, it appeared that higher yields with the 2:l extraction procedure were accompanied by less purity. About 3-4 polypeptides were detected in the chloroform/ methanol soluble fraction precipitated by ether (lanes 3 and 4 in Fig. 5A). Chloroform/methanol a t a volume ratio of 6:l extracted 2 major polypeptides, with molecular masses of 26 and 16 kDa.
To test whether the 16-kDa proteolipid bound DCCD, tonoplast vesicles were labeled with [14C]DCCD and then extracted with organic solvents. Although numerous polypeptides in the membrane fraction formed covalent bonds with DCCD, most of the radioactivity extracted by chloroform/ methanol (61) and subsequently precipitated by ether was associated with a polypeptide of 16-18 kDa (lane 9 in Fig.  5B). These results demonstrate that the 16-kDa proteolipid is a DCCD-binding protein.
Based on the total radioactivity recovered from each fraction, the DCCD-binding proteolipid after ether precipitation was only purified 4.5-fold relative to the tonoplast vesicles (Table I). But careful examination of the fluorogram (Fig. 5B,  lane 6) clearly demonstrates that this would be an underestimate of the purification, since much of the radioactivity in the tonoplast vesicles was associated with other polypeptides. By determining the relative specific activity of [14C]DCCD associated with the 16-kDa polypeptide alone in all fractions, the purification of the DCCD-binding proteolipid in the ether pellet would be about 12-fold. Since only about 36% of the radioactivity and 11.5% of the protein in the ether pellet was attributable to the 16-kDa band, the actual purification of the 16-kDa polypeptide in the gel may be as high as 105-fold (see legend to Table I). A similar maximal purification -fold was obtained when the 16-kDa proteolipid was extracted with chloroform/methanol a t 2:l (v/v) (three experiments, results not shown).
We had estimated before that there would be about 0.1 nmol of ATPase (50 pg)/mg of membrane protein, when the molecular mass of the holoenzyme is about 500 kDa. If a 100fold purification is required to purify the 16-kDa polypeptide, then there are about 10 pg or 0.6 nmol of the 16-kDa subunit/ mg of membrane protein. Although these values are based on several estimates, we can conclude that there are several copies (perhaps as many as four to eight) of the 16-kDa subunit/ATPase holoenzyme.
To confirm that the 16-kDa DCCD-binding protein is part of the tonoplast ATPase, the proteolipid fraction was probed  Although the functions of all these polypeptides have not been determined, it is possible that the ATPase complex may consist of all these subunits. The anti-ATPase antibody reacted with a single band of mass 16 kDa from the tonoplast proteolipid fraction (Fig. 6, lane 6), but not with the 8-kDa proteolipid from the mitochondria (Fig. 6, lane 5 ) . We can therefore eliminate the possibility that the 16-kDa proteolipid was an aggregate of the 8-kDa proteolipid from the mitochon-drial ATPase. These results support the idea that the 16-kDa DCCD-binding proteolipid is part of the tonoplast ATPase (10,12) and also suggest that the DCCD-binding proteolipid of the tonoplast ATPase is immunologically different from the mitochondrial one.

DISCUSSION
DCCD is a very useful probe for understanding the structure and function of the tonoplast H+-translocating ATPase. Here we have shown the following: (i) DCCD inhibits the tonoplast ATPase by partitioning into the hydrophobic environment of the membrane. (ii) Kinetic analyses of DCCD inhibition of the membrane-bound tonoplast ATPase suggests that 1 mol of DCCD bound per ATPase is sufficient to inactivate the enzyme (Fig. 1). This prediction was confirmed by binding experiments that showed 1 mol of DCCD was bound to the 16-kDa polypeptide when the tonoplast ATPase was completely inhibited (Fig. 4). (iii) The 16-kDa subunit of the tonoplast ATPase is a proteolipid, supporting the idea that it is an integral membrane protein.
(iv) Purification studies indicate that the 16-kDa proteolipid is present in multiple copies per holoenzyme (perhaps four to eight copies). How DCCD interacts with the 16-kDa subunit(s) to inhibit ATP hydrolysis and H+ translocation is still unclear.
The similarities between the 16-kDa proteolipid of the tonoplast ATPase and the 8-kDa proteolipid of the mitochondrial, chloroplast, and bacterial F,Fo-ATPase are striking. DCCD has been a useful probe for studying the structure and function of transport proteins (32), especially the FIFO-AT-Pase (36). Although DCCD can interact with carboxyls, sulfhydryls, and tyrosines, it has been shown mostly to form covalent bonds with hydrophobically located carboxyl groups, such as glutamate and aspartate residues (32). When DCCD reacts with a carboxyl residue, an unstable 0-acylurea adduct is first formed which can then rearrange to a stable Nacylurea. One of the results suggesting the participation of the 8-kDa proteolipid of the FIFO-ATPases in proton translocation came from bacterial mutants which had lost their ability to catalyze proton transport via the ATPase after the DCCD-sensitive carboxyl group had been eliminated by conversion to a glycine or asparagine (reviewed in Ref. 37).
Based on our preliminary findings, it is tempting to infer that the 16-kDa proteolipid of the tonoplast ATPase is analogous in structure and function to the 8-kDa proteolipid of the FIFO-ATPases and that DCCD inhibits both enzymes by a similar mechanism. However, there are indications that the two proteolipids vary. The tonoplast ATPase is less sensitive to DCCD than the mitochondrial enzyme from oat roots (9). The second-order inactivation rate constant of the tonoplast ATPase reported here is about a hundred times smaller than that of bovine heart mitochondria (30). The 16-kDa proteolipid of the tonoplast ATPase is two times larger than the mitochondrial proteolipid from oats, and they appear to be distinct immunologically (Fig. 6). Thus, determination of the complete molecular structure of the 16-kDa proteolipid is needed to establish its structural organization and its proposed role in proton translocation.
Note Added in Proof-After submission of this paper, two publications reporting similar characteristics of the DCCD-binding proteolipids of the vacuolar H'-ATPase from red beet tonoplast (38) and bovine brain clathrin-coated vesicles (39) have appeared. Another study has shown by reconstitution into proteoliposomes that the DCCD-binding polypeptide from bovine clathrin-coated vesicles has a role in H+-translocation (40).