Chromatographic Purification of the Chloroplast ATP Synthase (CFo-CF1) and the Role of CFo Subunit IV in Proton Conduction*

Chromatographic procedures were developed to pu- rify chloroplast ATP synthase

cation across the thylakoid membranes and, in conjunction with CFI, couples proton flow to ATP synthesis and hydrolysis. CFo contains four subunits, designated I, II, III, and IV (l-3). We have purified four-subunit CFO and demonstrated that the four subunits were sufficient to form an active CFo which was capable of conducting protons and coupling proton flow to ATP synthesis and hydrolysis (4). Neverthrless, the roles of the individual subunits are poorly understood. The counterpart of CF,, in E. coli (F,) is so far the best studied complex among the F,-F1 type ATP synthases. It consists of three subunits: a, b, and c. Both biochemical and genetic results suggested that all three subunits of F. are required for proton translocation (5)(6)(7)(8). The c subunit, and probably the a subunit as well, directly participate in proton translocation (9)(10)(11)(12)(13). The b subunit has an extensive hydrophilic domain which has been suggested to be exposed to cytoplasm and to associate with F1 (14, 15). Subunit III of CFo, which is similar to subunit c, is a DCCD-binding protein and is directly involved in proton translocation. Subunit III itself is capable of forming active proton channels (16,17). Based on DNA sequences and the derived secondary polypeptide structures, subunit I is proposed to be analogous to subunit b (18,19), and subunit IV is similar to subunit a (2,20). Subunit II of CFo has no analog in E. coli Fo.
Direct information concerning the role of subunit IV of CF, was obtained in this study by selective removal and reconstitution of subunit IV. Subunit IV was shown to be required for DCCD-sensitive proton translocation by CFo and for stimulation of ATP-Pi exchange and DCCD-sensitive Mg2'-ATPase activities in asolectin liposomes reconstituted with CFo-CF1 deficient in subunit IV. To our knowledge, this is the first successful reconstitution of energy coupling by a CFo polypeptide. MATERIALS AND METHODS solution was applied to a DEAE-Trisacryl column (2.5 X 9 cm) equilibrated with buffer B containing the ingredients of buffer A with the addition of 0.5% Triton X-100 and 0.01% (w/v) asolectin. The column was washed with 3 column volumes of buffer B, followed by 3 column volumes of buffer C containing 10% glycerol, 1 mM MgClz, 1 mM DTT. 0.5% sodium cholate. 0.01% asolectin, and 20 mM sodium phosphate buffer (pH 7.0). Then the column was washed with a 20-190 mM sodium phosphate buffer gradient (2 X 2.5 column volumes) (pH 7.0) in buffer C. Finally, CFo-CF, was eluted by 2 column volumes of buffer D containing 10% glycerol, 0.5% sodium cholate, 1 mM MgCb, 1 mM DTT, 0.01% asolectin, 0.5 M ammonium sulfate, and 100 mM sodium phosphate buffer (pH 7.0).
If high purity was desired, the peak fractions containing greater than 0.5 mg of protein/ml were pooled and dialyzed against 2 liters of buffer A for 8 h. Triton X-100 was added to the dialyzed sample from a 10% solution to give a concentration of 0.5%. The chromatographic procedure described above was repeated.

Prevaration of Subunit IV-deficient
CFn-CF,-The crude CF,-CF, " _ " _ solution was applied to the DEAE-Trisacryl column as described in the purification of CF&!F, (method b). After being washed with 3 column volumes of buffer B, the column was washed with 3 column volumes of buffer E containing 10% glycerol, 1 mM MgCl*, 1 mM DTT, 0.08% (w/v) Zwittergent 3-12, 0.01% asolectin, and 50 mM sodium phosphate buffer (pH 7.0) to remove subunit IV. Then the column was washed with buffer C and the phosphate buffer gradient as described in the purification of CFo-CF,. Finally, subunit IVdeficient CFo-CF, was eluted from the column by buffer D.
For storage of the purified protein, asolectin and ATP were added from 4% asolectin and 0.1 M ATP (pH 7.0) solutions to the ATP synthase peak fractions obtained by the DEAE-Trisacryl chromatography to give final concentrations of 0.1% asolectin and 0.1 mM ATP. The purified CFo-CFi and subunit IV-deficient CF&!F1 were kept under liquid nitrogen.

Purification
of Subunit IV-deficient CFO-The crude CFo-CF, was applied to the DEAE-Trisacryl column as described in the purification of CFo-CF, (method b). After being washed with buffer B and buffer E to remove subunit IV as described in the purification of subunit IV-deficient CFo-CFi, the column was washed with 3 column volumes of 10% glycerol, 1 mM EDTA, 0.1 mM DTT, 0.01% asolectin, 0.5% Triton X-iO0, and 30 mM sodium phosphate buffer (pH 7.0). Subunit IV-deficient CF,, was eluted from the column by 10% glycerol, 1 mM EDTA, 0.1 mM DTT, 0.01% asolectin, 20 mM Zwittergent 3-12, and 30 mM sodium phosphate buffer (pH 7.0) and concentrated on a small DEAE-Trisacryl column (1 X 4 cm) as previously described in the purification of CFo (4).
Purification of Subunit IV-The crude CFo-CF1 was applied to the DEAE-Trisacryl column as described in the purification of CFo-CF, (method b). After being washed with 3 column volumes of buffer B, the column was washed with 3 column volumes of 10% glycerol, 1 mM EDTA. 0.1 mM DTT. 0.01% asolectin. 0.5% Triton X-100. and 5 rnM sodium phosphate buffer (pH 7.0). Then subunit IV was'eluted by 3 column volumes of 10% glycerol, 1 mM EDTA, 0.1 mM DTT, 0.01% asolectin, 20 mM Zwittergent 3-12, and 5 mM sodium phosphate buffer (pH 7.0). Fractions containing subunit IV were pooled and concentrated on a small DEAE-Trisacryl column as previously described in the purification of CF,, (4) with the following modifications. The small DEAE-Trisacryl column (1 X 4 cm) was equilibrated with a buffer containing 10% glycerol, 1 mM EDTA, 0.1 mM DTT, 0.01% asolectin, 0.5% cholate, and 5 mM sodium phosphate buffer (pH 8.0) instead of 20 mM phosphate. Sodium cholate (20%. uH 8.0) was added to the pooled s&unitIV-containing solution to give a final concentration of 0.5%. After the mixture was applied to the column, it was washed with the same buffer used for equilibrating the column. Then subunit IV was eluted from the column with a buffer containing 10% glycerol, 1 mM EDTA, 0.1 mM DTT, 0.01% asolectin, 0.5% cholate, 0.5 M ammonium sulfate, and 100 mM phosphate (pH 8.0) as previously described in the purification of CFO (4). The concentrated subunit IV was dialyzed against 1 liter of 10% glycerol, 1 mM EDTA, 50 mM sodium phosphate (pH 7.0) for 2 days. The dialysis buffer was changed after 24 h. The dialyzed subunit IV was concentrated by ultrafiltration through Amicon YM 5 if necessary and kept under liquid nitrogen.
All operations were performed at 4 "C. The pH of the solutions was adjusted at room temperature. Reconstitution of Subunit IV-deficient CFO-CF, and Measurement of ATP-P, Exchange and Mg2'-ATPase Activities To reconstitute subunit IV with subunit IV-deficient CFo-CFi, less than 100 ul of the dialvzed subunit IV (0.2 ma/ml) was added to 250 ~1 of subunit IV-deficient CFo-CF, containing?.5 mg of protein, The final volume was adjusted to 360 ~1 by adding a buffer containing 10% glycerol, 3 mM MgCl,, and 50 mM sodium phosphate (pH 7.0). Sonicated asolectin was added to the mixture from a 4% stock solution to give a final concentration of 0.1% in 360 ~1. The protein/lipid mixture was then dialyzed against 1 liter of 10% glycerol, 3 mM Mg&, 1 mM DTT, and 50 mM sodium phosphate buffer (pH 7.0) for lo-16 h.
The dialyzed CFo-CF, preparations were incubated with 50 mM DTT on ice for 30 min and then incorporated into asolectin vesicles according to Krupinski and Hammes (22). ATP-Pi exchange was measured as described by Pick (23). For the Ma+-ATPase assay, the reconstituted proteoiiposomes containing IO-20 pg of protein were added to 1 ml of 5 mM MeCl,. 5 mM ATP. and 50 mM Na-Tricine (pH 8.0). DTT (10 mM) was added to the mixture prior to assaying. The reaction mixture was incubated at 37 "C for 30 min. When treated with DCCD, the mixture was incubated with 50 gM DCCD on ice for 30 min before DTT was added. The reaction was stopped by adding 0.1 ml of 40% (w/v) trichloroacetic acid. The reaction mixture was centrifuged for 2 min in a clinical centrifuge to remove the precipitate. P, in the supernatant was determined spectrophotometrically by the method of Taussky and Shorr (24). To solubilize residual asolectin that precipitated in the acid molybdate solution, 0.33% SDS was included in the Pi assay. SDS did not affect the color yield.

Measurement of ATP-induced Proton Uptake
The purified ATP synthase was precipitated at 45% saturation of ammonium sulfate by adding solid salt and centrifuging the mixture at 12,000 X g for 10 min. The pellets, containing 2 mg of protein, were solubilized in 1 ml of 1.4% sodium cholate (pH 8.0) and 4.0% asolectin and incubated with 50 mM DTT on ice for 30 min. Incorporation of the ATP synthase into liposomes and measurement of ATP-induced proton uptake were carried out according to Admon et al. (25), with the exception that the ApH-sensitive fluorescence probe 9-aminoacridine was replaced by quinacrine. Quinacrine fluorescence was monitored with a Farrand model 801 spectrofluorometer. Excitation was 426 nm, and emission was 505 nm with lo-nm slits for both. The quinacrine fluorescence was passed through an Oriel LP .47 filter, which transmits wavelengths >458 nm. Concentrated subunit IV (20 pg in 200 ~1) was added to the concentrated subunit IV-deficient CFo (100 pg in 200 ~1). The mixture was dialyzed against 1 liter of 10% glycerol, 1 mM MgCl*, 0.1 mM DTT, and 50 mM sodium phosphate (pH 8.0) in 6000-8000 molecular weight cut-off dialysis tubing for 3 days. The dialysis buffer was changed every 24 h. Duplicate samples containing either subunit IVdeficient CFO (100 rg) or subunit IV (20 pg) were prepared. The final volume was adjusted to 400 ,ul by adding the dialysis buffer. Reconstitution of CFo or subunit IV into asolectin liposomes was carried out as described by Okamoto et al. (26) with the following modifications. The weight ratio of CFO or subunit IV-deficient CFO to asolectin was about 1:130 and subunit IV to asolectin was about 1:650. The final concentration of asolectin was 20 m&ml in the protein/lipid/ detergent mixture. The mixtures were dialyzed against 1 liter-of 2.5 mM M&l,, 0.1 mM DTT, and 10 mM Tricine-NaOH (PH 8.0) for 18 h at 4 'C. The reconstituted proteoliposomes were loaded with KCI, and the proton translocation was measured as previously described (4).

Other Analytic Methods
Protein was determined according to Bensadoun and Weinstein (27); 0.3% SDS was included in the protein assay to solubilize lipids. SDS-polyacrylamide gel electrophoresis was carried out as described by Fling and Gregerson (28) Other biochemicals were reagent grade. Cholic acid was purified by recrystallization (30).

RESULTS
Purification of CFO-CF,-Although the sucrose gradient centrifugation procedure is a well developed method for purifying CFo-CFI (1, 21), it is not suitable for large scale preparation and subsequent isolation of CFO subunits. The chromatographic procedure described in this paper overcomes these limitations. In our procedure, the crude chloroplast ATP synthase (48% ammonium sulfate precipitate) was bound to a DEAE-Trisacryl anion exchange column in the presence of 0.5% Triton X-100. Decreasing the concentration of Triton X-100 to less than 0.5% significantly reduced the amount of the ATP synthase that bound to the column.* The majority of the contaminating proteins and colored materials in the crude preparation was eluted from the column by 50 mM sodium phosphate buffer in the presence of Triton X-100. Subsequent washing of the column with a 20-190 mM sodium phosphate buffer gradient in the presence of 0.5% sodium cholate efficiently removed ribulose bisphosphate carboxylase/oxygenase and free CF, without loss of the ATP synthase.
In the presence of sodium cholate, the ATP synthase bound to DEAE-Trisacryl very tightly and was only eluted from the column by a sudden increase in the ionic strength of the elution buffer. The fractions containing the ATP synthase were slightly yellowish. The presence of phosphate in the buffer system is crucial for obtaining highly active ATP synthase. The purification procedure described under "Materials and Methods" is for medium scale preparation. This purification can readily be scaled up to a final yield of 100 mg of ATP synthase by starting with 400-500 mg of the 48% saturation of ammonium sulfate precipitate on a 5 X lo-cm DEAE-Trisacryl column.
The SDS-polyacrylamide gel electrophoresis pattern of the purified ATP synthase is shown in Fig. 1. Like the ATP synthase purified by the sucrose gradient centrifugation procedure, the chromatographically prepared protein complex contains five CF1 polypeptides and four CFo subunits, I, II, III, and IV, with apparent molecular masses of 18, 16, 8, and 20 kDa, respectively. Subunit III did not stain well with Coomassie Blue, and its presence was confirmed by ['"Cl DCCD labeling (data not shown). The purity of the chromatographic preparation was 95% as judged by densitometric scanning of Coomassie Blue-stained gels. The purity of our preparation was very slightly lower than that of the sucrose gradient centrifugation preparation (97% for the ATP synthase peak fraction). This preparation is suitable for most experiments that require high activity. If higher purity is desired, repeating the anion exchange chromatographic procedure can increase the purity above 98%, but the specific ATP-Pi exchange activity will be reduced by about 50%. Table I shows a comparison of the purification of CFO-CFi by sucrose gradient centrifugation and anion exchange chromatography. Recovery of the ATP-Pi exchange activity by our chromatographic purification procedure is usually more than 50%, which is more than four times that obtained by the sucrose gradient centrifugation procedure. The typical specific ATP-Pi exchange activity of the ATP synthase purified by our chromatographic method was in the range of loo-250 ' Y. Feng  nmol/mg ATP synthaselmin, which was equivalent to or often higher than that purified by the centrifugation method. The protein yield of the chromatographic preparation was typically 2-3 times that of the sucrose gradient preparation. The chromatographically purified preparation of chloroplast ATP synthase, when reconstituted into liposomes, catalyzes ATP hydrolysis. The DCCD-sensitive ATPase activity was about 200 nmol of ATP hydrolyzed/mg of ATP synthase/ min. ATP-induced ApH formation in reconstituted CFo-CF1containing liposomes can be measured by determining the quenching of quinacrine fluorescence (Fig. 2). Typically, for such reconstituted proteoliposomes, 40-50% quenching of the quinacrine fluorescence can be obtained when ATP is added. The ATP-induced quenching was abolished by the ionophore gramicidin and completely inhibited by preincubation of the proteoliposomes with DCCD (Fig. 2).

Preparation
of Subunit IV-deficient CFO-CF,, Subunit ZVdeficient CF,, and Subunit IV-By modification of the chromatographic method for purifying CFo-CF1, subunit IV-deficient CF&Fi can be obtained. We found that subunit IV dissociates from CFo-CF1 at a very low concentration (0.08%) of Zwittergent 3-12, which has a critical micellar concentration of 0.12%. The dissociated subunit IV can be eluted from the DEAE-Trisacryl column by 50 mM phosphate buffer in the presence of Zwittergent 3-12 while subunit IV-deficient CFo-CFi remains on the column and can be collected separately. All remaining subunits in subunit IV-deficient CFO-CF1 appeared to be associated together as judged by sucrose density gradient centrifugation (31). The SDS-polyacrylamide gel electrophoresis pattern of subunit IV-deficient CFo-CF1 is shown in Fig. 3. Subunit IV stained much better with silver " The detergent extract of chloroplast thylakoid membranes was fractionated by ammonium sulfate. Precipitate between 37.5 and 45% saturation of ammonium sulfate was centrifuged on a sucrose gradient in the presence of 0.2% Triton X-100 according to Nelson (21).
' The detergent extract was precipitated at 48% saturation of ammonium sulfate. The precipitate was fractionated on a DEAE-Trisacryl column as described under "Materials and Methods." The purified ATP synthase was reconstituted with asolectin by gel filtration. Proteoliposomes (1 ml) containing 60 pg of CFO-CFi were diluted with 30 mM KCl, 3 mM MgCl?, 1 mM KH,PO,, 20 mM Na-Tricine (pH 8.0), and 10 nM valinomycin to a final volume of 2 ml. The final concentration of quinacrine was 1 PM. When added, ATP was 0.16 mM and gramicidin (gram) was 0.2 FM. A, ATP-dependent quenching of quinacrine fluorescence in ATP synthase-containing liposomes. B, as in A except that ATP synthase-containing liposomes were incubated with 50 pM DCCD for 30 min on ice before assay. than with Coomassie Blue. To quantitate the residual subunit IV in the subunit IV-deficient CFo-CF, preparations, a gel of CFo-CF, and subunit IV-deficient CFo-CF1 was lightly stained with silver and scanned with a densitometer. Quantitative measurements of four preparations indicated that 80-90% of the subunit IV was removed in the subunit IV-deficient preparations. The protein yield of subunit IV-deficient CFo-CF1 was quite similar to that of CFo-CF1 purified by the chromatographic method. The subunit IV thus obtained, however, was very dilute, and a separate procedure was used to isolate it. For efficiently eluting subunit IV, the CF,-CF, was loaded onto a DEAE-Trisacryl column which was then washed to remove the majority of impurities, followed by elution with 20 mM Zwittergent 3-12 at a low concentration (5 mM) of phosphate. The elution profile is shown in Fig. 4 Electrophoresis was carried out on a 15% polyacrylamide gel. Samples in lanes 1 and 2, containing 60 pg of protein, were run on the gel stained by Coomassie Blue, and samples in lanes 3 and 4 containing 30 rg of protein were run on the gel stained by silver nitrate. Lanes 1 and 3, CF&Fi purified by the chromatographic method; lanes 2 and 4, subunit IV-deficient CFo-CR.
came off the column as a sharp peak followed by subunit I and a trace of subunit III. The early fractions containing subunit IV were combined to minimize the contamination by subunits I and III. The subunit IV preparation was treated with ['4C]DCCD, and no contaminating subunit III was detected (data not shown). By this method, about 240 pg of subunit IV was obtained from 70 mg of crude CFo-CF1 precipitated at 48% saturation of ammonium sulfate. Assuming that subunit IV is approximately 3% of the total protein in CFo-CF,, the overall yield of subunit IV is about 40%. Subunit IVdeficient CFo was obtained by first removing subunit IV from the CF&F, followed by dissociation of subunit IV-deficient CF, from CF, by Zwittergent 3-12 as described in more detail in a previous paper (4). Some subunit III may be lost during the preparation of subunit IV-deficient CFo. The SDS-polyacrylamide gel electrophoresis patterns of CFo-CF1, subunit IV, and subunit IV-deficient CF, are shown in Fig. 5. Reconstitution of ATP-Pi Exchange and Mg2'-ATPase Actiuities-Both ATP-Pi exchange and Mg"-ATPase activities were inhibited when subunit IV was removed from CFo-CF, ( Table II). The results from several independent measurements showed that while IO-20% of subunit IV was present in the subunit IV-deficient preparation, the residual ATP-Pi exchange was 15-30%, and the residual DCCD-sensitive M$'-ATPase was 45-60% of the activities in the corresponding CFo-CF, preparation. The residual ATP-Pi exchange activity in the subunit IV-deficient CF,-CF, preparation is quite constant (about 30 nmol/mg protein/min) and, therefore, to a great extent the percentage of the residual activity in the subunit IV-deficient CFo-CFI preparation depends on the activity of the corresponding CFo-CF1 preparations, which varied from 100 to 250 nmol/mg protein/min. Strong inhibition of ATP-P, exchange which is consistent with the deficiency of subunit IV is always observed only when the activity of the corresponding CFo-CF1 preparation is over 150 nmol/ mg protein/min. Although the DCCD-sensitive Mg2'-ATPase is also impaired upon removal of subunit IV, the percentage of residual DCCD-sensitive Mg2'-ATPase activity is always higher than that of the ATP-P, exchange. Removal of 80-90% of subunit IV caused only a 40-55% inhibition of Mg"-ATPase.
Reconstitution of subunit IV with subunit IV-deficient CFo-CF1 by a dialysis procedure partially restored the ATP-P, exchange activity (Fig. 6). However, when the amount of subunit IV was increased above a certain level during reconstitution, the ATP-Pi exchange was inhibited. The reason for inhibition of exchange at higher subunit IV levels is unclear but may be the result of traces of detergent in the preparation. Residual Zwittergent appears not only to induce proton leakiness of the membrane but also to prevent correct reconstitution of subunit IV with subunit IV-deficient CFo-CF1. The direct addition of subunit IV to subunit IV-deficient CFo-CF, only caused further inhibition of the residual activity in the subunit IV-deficient CFo-CF, preparation (data not shown). It is necessary to reconstitute subunit IV with subunit IVdeficient CFo-CF1 by the dialysis procedure to reduce the Zwittergent concentration. Like the ATP-P, exchange reaction, the Mg2'-ATPase activity was also restored by dialysis of subunit IV-deficient CFo-CF1 with subunit IV (Fig. 7). In contrast to ATP-P, exchange, Mg2'-ATPase was not inhibited at high concentrations of subunit IV. This is understandable because Mg2+-ATPase activity is much less sensitive to the proton leak caused by the residual detergent than ATP-P, exchange. The restoration of ATP-Pi exchange and Mg2'-ATPase activities to subunit IV-deficient CFo-CF, by the addition of subunit IV provides strong evidence that the impairment of the activities in the subunit IV-deficient CFo-CF1 preparation is the direct consequence of the loss of subunit IV.
Reconstitution of Proton Translocating Activity-We have shown that purified CFo is capable of translocating protons when incorporated into asolectin liposomes (4). Proton translocation was measured after loading CFO-containing liposomes with potassium and induction of a potassium diffusion potential by valinomycin. The proton uptake due to a negative inside membrane potential was monitored by ACMA, a fluorescent pH indicator. When the same experiments were performed with the proteoliposomes containing either subunit IV-deficient CFo or subunit IV, there was almost no DCCDsensitive proton translocation (Fig. 8, B and C). Reconstituting subunit IV and subunit IV-deficient CFo restored proton conducting activity, which was inhibited by DCCD (Fig. 8A). The slow rate of proton transfer by the reconstituted CFo is probably partially due to loss of some subunit III during purification.' DISCUSSION The chromatographic procedure described in this paper is a convenient way to purify CFo-CF, in large quantity with high purity. By modifications of this procedure, we obtained subunit IV-deficient CFo-CF,, CFo (4), subunit IV-deficient CFo, and subunit IV. The CFo-CF1 purified by this procedure has high activity in comparison with that purified by the of subunit IV from CFo-CF1 Freshly prepared CFo-CFI and subunit IV-deficient CFo-CF1 were incorporated into asolectin liposomes by gel filtration. Aliquots of proteoliposomes were assayed for the activities as described in Table I and under "Materials and Methods." The residual subunit IV in the subunit IV-deficient CFa-CFi preparation was 12%. For determination of DCCD-sensitive Mg2'-ATPase activity, the mean activity of duplicate samples treated with 50 FM DCCD was subtracted from the total activity of DCCD-untreated samples. Subunit IV (0, 1, 5,10, and 20 pg) was reconstituted with 0.5 mg of subunit IV-deficient CFO-CFi by dialysis as described under "Materials and Methods." Then the reconstituted CFO-CF, was incorporated into asolectin liposomes by the gel filtration procedure. Aliquots of CFO-CF,-containing liposomes were assayed for ATP-Pi exchange activity as described in Table I. sucrose gradient centrifugation procedure. The results presented here show that removal of subunit IV from the CFo-CF, impaired both ATP-Pi exchange and Mg2'-ATPase activities.
The impaired activities were partially restored by adding back purified subunit IV to subunit IV-deficient CFo-CFr. Therefore, there is little doubt that the impairment in the activities is directly reIated to the loss of subunit IV from CFo-CFr. The measurements of proton conductivity of subunit IV-deficient CF, and the reconstituted CFo indicate that CF, loses its DCCD-sensitive proton conductivity when subunit IV is depleted. These results suggest that loss of ATP-Pi exchange and Mg*'-ATPase activities is caused by impairment of proton conductivity in CFo upon removal of subunit IV. After dissociation of CF1, CFo in thylakoid membranes gradually converted into an inactive form and lost its proton conductivity (32)(33)(34). The mechanism underlying this conversion is not clear. Nelson  Reconstitution of subunit IV with subunit IV-deficient CFo-CFI and subsequent reconstitution of CFo-CFi with asolectin liposomes were carried out as described in Fig. 6. Aliquots of CFo-CFi-containing liposomes were assayed for Mg2f-ATPase activity as described in Table II. Duplicate samples were treated with 50 pM DCCD prior to the assay. 0, subunit IV-deficient CFo-CF1+ subunit IV; H, subunit IV-deficient CFo-CFl+ subunit IV + DCCD. The DCCD-insensitive Mg2'-ATPase activity was subtracted from the total activity in DCCD-untreated samples; the DCCD-sensitive Me-ATPase activity is shown in the inset.
subunits of CFO becomes very weak. Thus subunit IV more easily dissociates from the other CF, subunits than in CF1associated CFO. CFc may undergo further conformational change when subunit IV is dissociated. A consequence of dissociation of subunit IV is loss of proton conductivity by CF, as demonstrated in our experiments. Subunit IV of CF, has similar secondary and tertiary structural features to the a subunit of F. in E. coli (19). The role of subunit a in proton translocation is still under investigation in regard to whether subunit a directly participates in proton translocation or only stabilizes the conformation of the proton channel built up by oligomers of subunit c (9,35). Recently, Cain and Simoni (10-12) reported that several missense mutations in the uric B (subunit a) gene resulted in the substitution of certain polar amino acids with nonpolar amino acids in the carboxyl-terminal region of subunit a. These substitutions impaired proton conductivity but did not alter F1 binding. Based on these studies, it was proposed that these polar amino acid residues might directly participate in proton transport (10-13). However, this may not be an unequivocal explanation. A single amino acid change in one subunit of the complex may lead to major alterations in biogenesis, subunit assembly, conformation, and activity of the complex. Maximum quenching of the fluorescence was obtained by adding 2 ~1 of 2 mM carbonyl cyanide m-chlorophenylhydrazone, a protonophore. When treated with DCCD, CFo liposomes were incubated with 50 pM DCCD for 1 h on ice. A, reconstituted CFo from subunit IV-deficient CFo and subunit IV; B, subunit IV-deficient CFo; C, subunit IV.
In our experiments, subunit IV was selectively removed by a mild treatment to minimize disruption of CF, structure. Our results showed that ATP-Pi exchange activity was inhibited more by subunit IV depletion than was Mg2+-ATPase activity. For ATP-Pi exchange, CFI must first hydrolyze ATP, pumping protons into lipid vesicles, and then use the generated proton gradient to synthesize ATP. Thus to synthesize one molecule of ATP by the exchange reaction, protons must move across membranes via CFo twice (36). For DCCDsensitive ATP hydrolysis, in contrast, protons only need to cross CFO once. Thus ATP-Pi exchange activity should be more sensitive to any impairment in CFO than Mg2'-ATPase activity.
The partial inhibition of DCCD-sensitive Mg2+-ATPase by removal of subunit IV is, however, perplexing. CF1-associated CFo could be partially active in the absence of subunit IV. If this is true, an essential role of subunit IV in proton translocation would be unlikely. Instead, subunit IV could be required for organizing CFo structure. It is also possible that, even though the Mg2'-ATPase is inhibited by DCCD, ATP hydrolysis is not tightly coupled to proton transport. The /3 subunit of Rhodospirillum rubrum F1 may be selectively removed from chromatophore membranes and high rates of ATPase activity reconstituted by the addition of either E. coli (37) or spinach chloroplast (38) fi subunit. Although reconstitution with R. rubrum fi subunit restores both ATPase activity and ATP-driven proton translocation, little proton translocation is linked to ATP hydrolysis by the hybrid enzymes. The hydrolysis of Ca'+-ATP by R. rubrum FO-FI also does not drive proton transport.
Similar results were reported for E. coli ATP synthase (10-12). While proton translocation was impaired upon substitution of certain amino acids in subunit a of Fo, the DCCD-sensitive ATPase activity of the membrane-bound Fo-F1 was much less affected.