Subunit Interaction in the Mitochondrial H+-translocating ATPase THE ROLE OF OLIGOMYCIN SENSITIVITY CONFERRAL PROTEIN AND COUPLING FACTOR 6 IN ATPase BINDING AND Pi-ATP EXCHANGE IN MITOCHONDRIAL MEMBRANES*

oligomycin

' The abbreviations used are: FI, coupling factor 1; FB, coupling factor B; F6 or FC,, coupling factor 6; OSCP, oligomycin sensitivity conferral protein; TUA-STA particles, submitochondrial particles treated with trypsin (T) and urea (U) to remove FI, sonicated at alkaline pH (A) to remove OSCP, and treated with silicotungstic acid (STA) to remove F,; Temed, N,N,N',N'-tetramethylethylenediamine; SDS, sodium dodecyl sulfate; Tes, N-tris(hydroxymethy1)methyl-2aminoethanesulfonic acid; Mops, 4-morpholinepropanesulfonic acid. liver mitochondria appeared to have the reported properties of both OSCP and F6. Type I ATPase bound to TUA-STA particles and its ATP hydrolytic activity was sensitive to oligomycin (1). In this communication, we will demonstrate functional similarities between OSCP and the 26,500-dalton subunit and in a companion paper we will characterize further the physical similarities of the two proteins.
OSCP and F,j have been reported to stimulate the 32P,-ATP exchange (7, 8). The ability of the 26,500-dalton subunit to enhance the 32Pi-ATP exchange is compared with the stimulatory effects of OSCP.
In the present communication, we wiU demonstrate the following: (a) OSCP, in the absence of Fs, is able to bind ATPase to the TUA-STA particles with an apparent KA of lo6 M-'; ( b ) F6-dependent ATPase binding to TUA-STA particles has an apparent association constant 1-2 orders of magnitude lower than that obtained with OSCP and the bound ATPase activity is rutamycin-insensitive; (c) F1-ATPase and Type 11 ATPase require F6 in addition to OSCP and FB to reconstitute 32Pi-ATP exchange activity, the F6 requirement for 32Pi-ATP exchange being unrelated to the effect of F6 in STA particles on the binding of the ATPase; ( d ) Type I ATPase and therefore the 26,500-dalton subunit associated with it requires F6 and FB to reconstitute 32Pi-ATP exchange activity in STA particles; and (e) OSCP can be interchanged with the 26,500-dalton subunit in the stimulation of the ATPase binding and the 32Pi-ATP exchange reaction.

Materials
The following chemicals and enzymes were obtained from the indicated sources: Mops, Tris, Tes, EDTA, pyruvate kinase, lactate dehydrogenase, bovine serum albumin, and Coomassie blue R-250 from Sigma; NADH, ATP, and phosphoenolpyruvate from Boehringer Mannheim, SDS from BDH Chemical Ltd., Poole, England; ultrapure sucrose and (NH4)2S04 from Schwarz/Mann; 2 N phenol reagent from Fisher; 'acrylamide, N,N'-methylenebisacrylamide, and Temed from Eastman; ammonium persulfate and AG I-X2 resin from Bio-Rad; and digitonin from Pfaltz and Bauer. All these reagents were of reagent grade purity.
Adult male or retired breeder male albino rats (Sprague-Dawley Crl:CD(SD)BR) were obtained from the Charles River Breeding Laboratories, Wilmington, MA and were fed ad libitum Purina Rat Chow.

Methods
Preparation of Rat Liver Type I and Type II ATPase-Preparation of rat liver mitochondria, submitochondrial particles, the extraction of Type I and Type I1 ATPase from submitochondrial particles with chloroform, and their purification by a zone sedimentation was reported previously (1).
Preparation of Beef Heart OSCP, F6, Fl-ATPase, and Factor B-Beef heart OSCP was prepared according to Senior (7) with some slight modifications. The (NH4)OH extract is treated with 238 mg of (NH4)PS04 at 0 "C for 16 h. The (NH&S04 precipitate was collected and dissolved in a minimal volume of buffer containing 15 mM Tris acetate (pH 8.0), 2 mM EDTA, and 2 mM dithiothreitol (buffer 1). This material was absorbed to a column of CM-Sephadex ((3-25-120 mesh) and washed with buffer 1 containing 160 mM KC1 until no more protein eluted. At this point, the KC1 was increased to 320 mM which eluted the OSCP. The fractions containing OSCP were pooled and the OSCP precipitated with (NH4)2S04 (238 mg/ml).
Beef heart Fs was prepared as described by Kanner et al. (8). Beef heart F1-ATPase was prepared as described by Horstman and Racker Preparation ofSubmitochondria1 Particles Depleted in ATPase-Rat liver submitochondrial particles were treated sequentially with trypsin and urea and then passed through a French pressure cell at pH 10 to remove OSCP. These submitochondrial particles were then treated with 1.5% silicotungstic acid to remove FCZ (3). Beef heart TUA-STA particles and STA particles were prepared exactly as described (3,5). The ATPase-depleted submitochondrial particles were stored under liquid Nz until used.
Measurement of the Binding of ATPase to the TUA-STA Particles-Binding of Type I ATPase was measured by incubating the TUA-STA particles (1 mg/ml) with various amounts of Type I ATPase in a medium containing 20% ethylene glycol, 10 mM Tris/ Tes, pH 7.5, and 2 mM EDTA total volume was 50 or 100 pl. The incubations were carried out at room temperature for 15 min. At the end of the incubation, samples were centrifuged at 220,000 X g for 30 min in a 42.2 Ti rotor at 4 "C. Aliquots of the supernatant fluid were withdrawn and ATPase activity was measured. The pellet residues were rinsed 3 times with ATPase-assay medium, then suspended in 40 pl of the same medium for ATPase activity and rutamycin sensitivity measurement.
For measuring the effect of OSCP and Fs on the binding of ATPase, rat liver Type I1 ATPase was incubated with OSCP (pretreated with 50 mM dithiothreitol at room temperature for 30 min) or Fti in a medium containing 20% ethylene glycol, 10 mM Tris/Tes, pH 7.5, and 2 mM EDTA at room temperature for 15 min. TUA-STA particles were added to the mixture at a final concentration of 1 m g / d (50 pl total volume) and the incubation was carried out for an additional 15 rnin. At the end of the incubation, the samples were centrifuged and ATPase activity in the supernatant pellet was assayed as described above.
Assay of ATPase Activity-ATPase was assayed with a coupled enzymatic reaction as described previously (1). Rutamycin sensitivity was measured by adding 10 pg of rutamycin in the 3-ml assay mixture; the subsequent ATPase activity was determined.
Determination of Protein-The submitochondrial particle protein was determined in the presence of 0.5% SDS (12) and the soluble protein was determined according to Lowry et al. (13).

RESULTS
The results shown in Fig. 1 confirm the data in the literature (2,3) that F6 (FC2) causes binding of F1-ATPase to TUA-STA particles. However, in contrast to Ref. 2, OSCP alone was sufficient to bind F,-ATPase. The bound ATPase activity was oligomycin-sensitive with OSCP-induced binding and oligomycin-insensitive with Fs-induced binding. Moreover, we found a rather large difference in the association constants for binding F,-ATPase to TUA-STA particles. The KA for Fsinduced binding of F,-ATPase was at least 1-2 orders of magnitude less than the KA found for OSCP-induced binding of FI-ATPase. When Fs and OSCP were used in combination, the stoichiometry and KA of ATPase binding were exactly the same as 0scp alone. We must conclude from these data that Fs does not participate in the binding of F1 to TUA-STA particles. In order to show that this conclusion is valid for the rat liver ATPase preparation, we made hybrid ATPase corn-extent of ATPase binding varied from 0.3-0.8 m o l of ATPase bound per mg of TUA-STA particle. The variation in stoichiometry was not due to the TUA-STA particles but undefined differences in the ATPase preparations, As shown in Table I, F6-induced binding was always rutamycin-resistant and low affinity binding, while OSCP-induced binding was always high affinity binding and the bound ATPase activity was rutamycin-sensitive. Again we must conclude that Fs is not involved in the specific binding of the ATPase to TUA-STA particles regardless of source of either the ATPase or particles. The Binding of beef heart F1-ATPase to TUA-STA submitochondrial particles derived from beef heart mitochondria. Beef heart FI (8.1 pg) and beef heart TUA-STA particles (49.3 pg) were incubated at room temperature with various amounts of Fs or OSCP. The total volume of the reaction mixtures was 50 pl and it contained, in addition to the protein components, 10 mM Tris/Tes, pH 7.5,2 mM EDTA, 20% ethylene glycol, and 13.5 mM dithiothreitol when OSCP was present. The F6 and FI and OSCP and FI were incubated for 15 min and then, after addition of the TUA-STA particles, were incubated 15 min longer. The incubation mixtures were then centrifuged at 220,000 X g for 20 min at 23 "C. The supernatant fluid was assayed for ATPase activity and the pellet was assayed for rutamycin-sensitive ATPase activity. The units remaining in the supernatant fluid were designated as free ATPase. The bound ATPase was derived by subtracting the free ATPase from a control to which no particles had been added. Using a one-binding site model, the data were directly fitted using a nonlinear regression, inerative technique. The experimental points (0) corresponded well with the calculated points (0). The calculated binding parameters are shown.  (8). The data in Table I1 c o n f i i this observation.
Unlike the binding of ATPase to the membrane, the 32Pi-ATP exchange required F6 in the presence of the ATPase binding protein (Type I ATPase) or OSCP. The only exception was with crude Type I1 ATPase where F6 was not required for the exchange activity. The Fg dependence was regained after purification of the crude Type I1 ATPase by zone sedimentation (data not shown). A requirement for F6 in the 32Pi exchange reaction catalyzed by beef heart F1-ATPase in STA particles has been reported previously (8). These data demonstrate that the F 6 prepared in our laboratory by published procedures is active in the 32Pi-ATP exchange reaction.
The ability of the ATPase binding protein to enhance the 32Pi-ATP exchange reaction was evident by the ability of the Type I ATPase to catalyze the exchange reaction. The Type I1 ATPase was unable to catalyze 32Pi-ATP exchange unless supplemented with OSCP. Beef heart OSCP is therefore functionally interchangeable with the rat liver ATPase binding protein in the 32Pi-ATP exchange reaction. The hybrid ATPase complex formed from rat liver Type I1 ATPase and beef heart OSCP behaved similarly to the homologus enzyme (Type I ATPase) in both the ATPase binding to TUA-STA particles and the 32Pi-ATP exchange reaction in STA particles. We used beef heart particles prepared with 1% silicotungstic acid to measure the 3ZPi-ATP exchange reaction (5) because, as others have also found (3), the TUA-STA particles are unable to carry out 32Pi-ATP exchange even when supplemented with coupling factors, despite their ability to reconstitute oligomycin-sensitive ATPase activity.
Since Fg has enhancing effects on both the 32Pi-ATP ex-  The incubation was continued at 37 "C for 15 min and then the reaction was terminated by the addition of 0.5 ml of 10% trichloroacetic acid. After removal of denatured protein by low speed centrifugation, the inorganic phosphate was extracted as the phosphomolybdate complex and the organic phosphate was determined as described (11 33 0.66 0.99 1.32 1.65 nmoles Fs / mg STA   FIG. 2. Lack of correlation between Fa-dependent PI-ATP exchange and ATPase binding in STA particles. The Pi-ATP exchange reaction was measured as described in Table I1 and the ATPase binding to STA particles was measured as described in Fig.   1.
change reaction and on low affinity binding of the ATPase to submitochondrial particles, it was important to ascertain if this effect of F6 on the 32Pi-ATP exchange was due to promotion of binding of the ATPase to the membrane by F6. We undertook the experiment shown in Fig. 2, where part of the sample was used to measure 32Pi-ATP exchange and the other part was used to measure the amount of ATPase bound to the membrane. It was evident that there is no correlation between the 32Pi-ATP exchange reaction and binding of F1-ATPase to STA particles. There was a %fold increase in the 32Pi-ATP exchange and essentially no change in the binding of F1-ATPase to the STA particles. Moreover, as was shown in Fig. 1, in order to see any effect with F6 on ATPase binding, one needs 10-20 nmol of Fc per mg of TUA-STA particles while in the 32Pi-ATP exchange, 1 nmol of F6 per mg of STA particles gives maximum effect.
Similar results were obtained with rat liver Type I ATPase (data not shown).

DISCUSSION
The present studies imply that Fg does not play a significant role in the binding of ATPase to the mitochondrial membrane. F,j is, however, involved in the 32Pi-ATP exchange. It is OSCP or the 26,500-dalton protein subunit which is responsible for both the ATPase binding and the conferring of the rutamycin sensitivity to the membrane-bound ATPase activity. Experimental evidence against the involvement of Fs in the binding of ATPase include the fact that the F6-dependent binding of ATPase has an apparent association constant 1-2 orders of magnitude lower than that of the OSCP-dependent or the 26,500-dalton subunit-dependent binding of ATPase. The KA of ATPase for F6 therefore is 1-2 orders of magnitude weaker than that for OSCP or the 26,500-dalton subunit. The binding of ATPase to membrane in the presence of both OSCP and F6 (at a concentration of F6 approximately 7-10-fold higher than OSCP) is similar to that seen with OSCP alone. The presence of excess F6 does not affect the binding parameters with respect to OSCP (data not shown). This observation again demonstrates that ATPase binds preferentially to OSCP sites in the membrane.
The lack of agreement on the role of Fs in the binding of the ATPase to TUA-STA particles as suggested in the literature may arise from at least two uncertainties: qualitative observations and heterogeneous F6 preparations. The F6 effect had not previously been quantiatively analyzed except by Vandineau et al. (2). Our observation that the KA for the OSCP-induced F1-ATPase binding is 1-2 orders of magnitude greater than the binding in the presence of F 6 alone may be due to some differences between our preparations and that of Vadineau et al. (2). The OSCP used in their experiments was isolated from FIX (F1.OSCP) complex and is not homogeneous. We have isolated OSCP by the method published by Senior (7) and this preparation is homogeneous in gradient polyacrylamide gel (14-23%) in the presence of SDS.
The reported observation that OSCP does not increase the F1 binding but only increases the rutamycin sensitivity (6) can be interpreted in light of our findings. In the absence of OSCP, FI binds to the unextracted F6 sites (or some other yet unidentified F1 binding site) in TUA particles (2), urea-Fo (6), or NaBr Fo (6,14,15); the bound ATPase therefore is insensitive to rutamycin. Upon addition of OSCP, F1 now binds preferentially to OSCP sites instead of F6 sites (or unidentified ATPase binding sites) in the membrane. The resultant particle ATPase activity may not be changed significantly, but the bound ATPase becomes rutamycin-sensitive. It is the presence of F6 sites or yet undefined other ATPase binding sites on the membrane that marks the later binding effect of OSCP. By using TUA-STA particles which are depleted of OSCP as well as F6, we can reduce the background binding of ATPase due to F6, and can demonstrate the binding effect due to OSCP. Since the F6-dependent binding of ATPase is much weaker than the OSCP-dependent binding, the subsequent addition of OSCP in a separate incubation would provide a high affinity binding site for the ATPase. The high affinity site competes for the F F , binding site, leading to the transfer of ATPase from F6 binding sites to the OSCP binding sites and rutamycin sensitivity is then restored to the bound ATPase.
The requirement for F6 for the 32Pi-ATP exchange in STA particles would suggest that F6 acts as an energy coupling factor. It has been suggested that F6 is part of the H'-translocating ATPase complex based on its presence in various ATPase complex preparations (6,15). The actual mechanism for the F 6 effect on the 32Pi-ATP exchange remains to be elucidated. Type I ATPase catalyzes the 32Pi-ATP exchange in STA particles. This reaction requires F6 and FB, but not OSCP. In fact, addition of OSCP causes about 24% inhibition of the 32Pj-ATP exchange. In view of the functional equivalency of the OSCP and the 26,500-dalton subunit, the inhibition could result from the competition between these two proteins for the same site in STA particles.
Rat liver TUA-STA particles appear to recognize OSCP and the 26,500-dalton protein equally well, since the binding of the Type I1 ATPase to rat liver particles in the presence of OSCP shows similar binding characteristics to that of the Type I ATPase. On the other hand, rat liver ATPase also recognizes OSCP and the 26,500-dalton protein equally well (data shown in the following paper). Further observations on the 32Pi-ATP exchange in beef heart STA particles reconstituted with rat liver Type I, rat liver Type I1 and OSCP, or beef heart F1 and OSCP indicate that as far as rat liver and beef heart mitochondria are concerned, the soluble ATPase, binding protein component, and the membrane components are completely interchangeable. Beef heart OSCP and yeast OSCP have been reported to be not interchangeable in yeast submitochondrial particles (16) with regard to the rutamycin sensitivity conferring activity, but partially interchangeable with regard to the ATPase-binding activity.