Photoinactivation of the F1-ATPase from Spinach Chloroplasts by Dequalinium Is Accompanied by Derivatization of Methionine β183*

In contrast to the F1-ATPases from bovine mitochondria and the thermophilic Bacillus PS3, which are reversibly inhibited by dequalinium in the absence of irradiation, the Mg2+-ATPase activity of heat- or dithiothreitol-activated chloroplast F1 (CF1) from spinach chloroplasts is slightly stimulated by dequalinium. Conversely, dequalinium is a partial inhibitor (maximal inhibition is 85–90%) of the Ca2+-ATPase of CF1 activated by heat, dithiothreitol, or octylglucoside. The Mg2+- and Ca2+-ATPase activities of CF1 respond differently in the presence of lauryl dimethylamine oxide (LDAO) in the assay medium. Whereas the Mg2+-ATPase activity of heat- or dithiothreitol-activated CF1 is stimulated up to 14-fold by increasing concentrations of LDAO, the Ca2+-ATPase is inhibited in a biphasic manner by increasing concentrations of LDAO. In the presence of LDAO, dequalinium does not stimulate the heat-activated Mg2+-ATPase over that promoted by LDAO alone. That dequalinium slightly stimulates Mg2+-ATPase activity although it inhibits Ca2+-ATPase activity can be reconciled by assuming that dequalinium binds to two sites in CF1, a stimulatory site that also binds LDAO and an inhibitory site. By acting as a partial inhibitor of the Mg2+-ATPase activity that it activates, the combined effect of dequalinium is modest stimulation. Irradiation of heat- or dithiothreitol-activated CF1 or the α3β3γ subcomplex of CF1 in the presence of 12 μm dequalinium led to rapid photoinactivation. ATP and ADP, separately or in combination with Mg2+, protect against photoinactivation. After photoinactivating the α3β3γ subcomplex of CF1 with [14C]dequalinium, tryptic and peptic digests of the isolated, derivatized β subunit were fractionated by high performance liquid chromatography. Sequencing of the isolated, radioactive tryptic and peptic peptides revealed that Metβ183, which is at or near the catalytic site, is derivatized in a single β subunit when CF1 is photoinactivated with [14C]dequalinium.

The ATP synthase complex of chloroplast thylakoid membranes catalyzes ATP synthesis coupled to a proton electrochemical gradient across the thylakoid membrane generated by light. The membrane-bound enzyme comprises an integral membrane protein complex, CF 0 , 1 and a peripheral membrane protein complex, CF 1 . CF 0 mediates transmembrane proton conduction, whereas CF 1 contains the catalytic sites for ATP synthesis.
The CF 1 portion of ATP synthase can be easily removed from the thylakoid membrane. Isolated CF 1 is composed of five different subunits, designated ␣, ␤, ␥, ␦, and ⑀ in order of decreasing molecular weight with a stoichiometry of ␣ 3 ␤ 3 ␥␦⑀, and has a molecular mass of 400 kDa (1). CF 1 contains six binding sites for adenine nucleotides (2), three of which participate directly in catalysis, whereas the other three, which are called noncatalytic sites, do not have a well defined functional role. Isolated CF 1 is a latent ATPase, which can be activated with dithiothreitol (DTT) (3), trypsin (4), heat (4,5), organic solvents (6), or octylglucoside (7). DTT treatment reduces the disulfide bond within the ␥ subunit (8), while heat activation relieves inhibition caused by the ⑀ subunit (9,10). Trypsin digestion partially cleaves the ␥ subunit, which significantly decreases the affinity of CF 1 for the ⑀ subunit (11,12). The effects of octylglucoside and organic solvents on CF 1 are more complicated. Octylglucoside may remove the inhibitory ⑀ subunit, decrease inhibition by free Mg 2ϩ , and also increase the affinity for substrate (7,13). One effect of organic solvent is to overcome inhibition by the ⑀ subunit. A second activating effect may be a decrease of the affinity for inhibitory MgADP (14).
The F 1 -ATPases are inhibited by a variety of amphipathic cations that include substituted phenothiazines (15,16), substituted xanthenes (17)(18)(19)(20), substituted acridines (17,19,21), and mono-and bisalkylquinaldiniums (17,19,(22)(23)(24). The binding sites for quinacrine mustard and dequalinium have been partly defined by labeling studies on TF 1 and MF 1 . Inactivation of MF 1 with quinacrine mustard is due, at least in part, to modification of one or more of the carboxylic acid side chains in the DELSEED segment of ␤ subunit (25), which, according to the x-ray structure, participates in a catch contact with Lys 87 and Lys 90 on the ␥ subunit (26). Dequalinium inhibits MF 1 noncompetitively in the absence of light (17,22). Irradiation of MF 1 in the presence of dequalinium at 350 nm rapidly inactivates the enzyme. Photoinactivation of MF 1 with [ 14 C]dequalinium results in the labeling of Phe ␣403 and Phe ␣406 in mutually exclusive reactions and a side chain within residues 440 -459 of the ␤ subunit (23). To a certain extent, dequalinium cross-links the ␣ and ␤ subunits of MF 1 , indicating that the reagent spans the distance between Phe ␣403 or Phe ␣406 and a side chain within residues 440 -459 of the ␤ subunit. Dequalinium interacts differently with TF 1 . Although it inhibits TF 1 and MF 1 noncompetitively, a long lag is introduced when TF 1 is assayed in the presence of dequalinium, a phenomenon not observed with MF 1 . Irradiation of TF 1 in the presence of dequalinium inactivates the enzyme, which is accompanied by derivatization of Phe 420 (equivalent to Phe ␤424 of MF 1 ) in the nucleotide binding domain of a single ␤ subunit (24).
Groth and Junge (27) reported that dequalinium blocks proton release from the CF 0 CF 1 -ATP synthase under conditions of proton slip, a conducting state observed when ADP and P i are not present to consume the proton electrochemical potential generated when thylakoids are illuminated. They assumed that the DELSEED segment of the ␤ subunit, which is derivatized when MF 1 is inactivated with quinacrine mustard, is part of the binding site for dequalinium on CF 1 (27). Given that MF 1 and TF 1 respond differently to dequalinium and that different residues in the two enzymes are derivatized when they are photoinactivated with [ 14 C]dequalinium, it was of interest to determine if the interaction of dequalinium with CF 1 resembles more its interaction with MF 1 or TF 1 .

EXPERIMENTAL PROCEDURES
Materials-Reagents for gel electrophoresis were purchased from Bio-Rad. Enzymes and biochemicals used in assays and dequalinium dichloride were purchased from Sigma. LDAO (30% aqueous solution) was purchased from Calbiochem. Pepsin was purchased from Sigma. Sequence grade modified trypsin was purchased from Promega. HPLC solvents were purchased from Fisher. Centricon 30 was obtained from Amicon. [ 14 C]Dequalinium with a specific radioactivity of 2.9 cpm/pmol was synthesized as described previously (23). CF 1 was prepared by modification of previously described methods (28,29). After CF 1 was purified by anion exchange on DEAE-cellulose (Whatman) and DEAE-Sephadex (Pharmacia Biotech Inc.), considerable Rubisco remained, which was removed as follows. A solution of CF 1 (150 mg) in buffer A (50 mM Tris-SO 4 , pH 8.0, 1 mM EDTA, and 1 mM ATP) containing 0.5 M (NH 4 ) 2 SO 4 was applied to a 2.5 ϫ 8 cm column of Butyl-Toyopearl-650S (Toso-Haas), which was equilibrated with the same buffer. After application, the column was washed with two volumes of the same buffer. Rubisco was eluted by washing with 100 ml of 0.2 M (NH 4 ) 2 SO 4 in buffer A. The column was then washed with 100 ml of 0.15 M (NH 4 ) 2 SO 4 in buffer A, which eluted a small amount of CF 1 deficient in the ␦ subunit, CF 1 (Ϫ␦), a trace amount of residual Rubisco, and contaminating proteins that migrate behind the ␣ subunit on SDS-PAGE. Finally, pure, intact CF 1 as assessed by SDS-PAGE was eluted with buffer A. The specific activity of CF 1 was 26 mol of ATP hydrolyzed min Ϫ1 mg Ϫ1 when assayed at 30°C in the presence of 30 mM octylglucoside (7).
When CF 1 , prepared as described above, was reapplied to the Butyl-Toyopearl-650S column in buffer A containing 0.5 M (NH 4 ) 2 SO 4 and then developed with a linear gradient of 0.5-0 M (NH 4 ) 2 SO 4 in buffer A in a total volume of 600 ml, CF 1 (Ϫ␦), as assessed by SDS-PAGE, eluted as a single peak in fractions 65-90 of the 6.5-ml fractions collected. The column was then washed with 50 ml of buffer A, which removed a single peak containing mostly ␦ subunit and trace amounts of the ␣, ␤, ␥, and ⑀ subunits as assessed by SDS-PAGE. This peak probably represents free ␦ subunit containing a trace of intact CF 1 . CF 1 (Ϫ␦) prepared in this manner was stored at 4°C after precipitation with 50% saturated (NH 4 ) 2 SO 4 . The ␣ 3 ␤ 3 ␥ subcomplex, CF 1 depleted of the ␦ and ⑀ subunits, was prepared from CF 1 (Ϫ␦) by the method described for preparing CF 1 (Ϫ⑀) (10).
Methods-Peptide separations by HPLC were conducted as described previously (23,24). Protein concentrations were determined by the method of Bradford (30) with Coomassie Blue Plus from Pierce, but using a correction factor of 1.5 based on an A 277 of 0.483 cm 2 /mg determined for CF 1 (31). Molar concentrations of CF 1 were based on a molecular mass of 400 kDa (1). Radioactivity was determined by scintillation counting in Ecoscint from National Diagnostics. CF 1 was activated with DTT (8) or heat in the absence of 10 mM DTT (32). Excess ATP and DTT were removed using centrifuge columns of Sephadex G-50 (33) equilibrated with 50 mM Tris-HCl, pH 8.0, for heat-activated CF 1 and 50 mM Tris-HCl, pH 8.0, and 1 mM DTT for DTT-activated CF 1 . The Mg 2ϩ -ATPase was determined spectrophotometrically using the ATP regeneration system described previously (34). Ca 2ϩ -ATPase was determined by either inorganic phosphate release using the method of Taussky and Shorr with 5 mM ATP and 5 mM CaCl 2 (35) or the enzyme-coupled assay method of Bruist and Hammes (36). Octylglucoside-dependent Mg 2ϩ -and Ca 2ϩ -ATPase were assayed using the conditions of Pick and Bassilian (7).
Isolation and Proteolytic Digestion of the 14 C-Labeled ␤ Subunit after Photoinactivating the ␣ 3 ␤ 3 ␥ Subcomplex of CF 1 with [ 14 C]Dequalinium-A solution of the ␣ 3 ␤ 3 ␥ subcomplex of CF 1 at about 1 mg/ml in 50 mM Tris-Cl, pH 8.0, containing 7 mM ␤-mercaptoethanol and 14 M [ 14 C]dequalinium was irradiated at 350 nm in a Rayonet photochemical reactor. Samples were withdrawn with time and assayed for residual Mg 2ϩ -ATPase activity in the presence of octylglucoside (7). When the ATPase was photoinactivated by 80%, the reaction mixture was concentrated to 0.5 ml with a Centricon 30 membrane and passed through a 5-ml centrifuge column of Sephadex G-50 equilibrated with 50 mM Tris-HCl, pH 8.0, to remove unbound [ 14 C]dequalinium. The labeled ␣ 3 ␤ 3 ␥ subcomplex was dialyzed against 2 liters of medium consisting of 50 mM sodium succinate, 1 M NaCl, 0.25 M NaNO 3 , 0.1 mM DTT, 4 mM EDTA, pH 6.1, overnight at 4°C as described by Michel et al. (37). Protein precipitated during dialysis, and was recovered by centrifugation. Analysis of the supernatant and precipitate revealed that the supernatant contained mostly ␤ subunit with small amounts of the ␣ and ␥ subunits, whereas the precipitate contained mostly ␣ and ␥ subunits and a trace of the ␤ subunit. The supernatant was applied to a hydroxyapatite column (Bio-Rad HTP, 1.5 ϫ 5 cm) equilibrated with 25 mM Tricine-NaOH, pH 8.0, 0.2 mM EDTA, 0.1 mM ATP, and 0.1 mM DTT at 4°C. The column was washed with two volumes of equilibration buffer. The ␤ subunit was eluted with the same buffer supplemented with 30 mM sodium phosphate, pH 8.0 (38). The solution containing ␤ subunit was dried under vacuum and dissolved in 6 M guanidine HCl. About 1.5 mg of pure ␤ subunit was obtained from 6 mg of ␣ 3 ␤ 3 ␥ subcomplex. Based on the amount of 14 C incorporated per mol of ␤ subunit, 27% of the isolated ␤ subunit was labeled by [ 14 C]dequalinium.
Prior to pepsin digestion, the denatured ␤ subunit in guanidine HCl was dialyzed against 2 liters of 1% (v/v) formic acid overnight at 4°C. An aqueous solution of pepsin was added to a final pepsin:protein ratio of 1:50. Digestion was performed at 30°C for 4 h with continuous stirring and was interrupted by freezing at Ϫ20°C. Prior to trypsin digestion, the denatured ␤ subunit in guanidine HCl was dialyzed against 2 liters of distilled water overnight at 4°C. The ␤ subunit precipitated during dialysis and was collected by centrifugation. The protein was suspended in 500 l of Tris-HCl, pH 8.0. Trypsin was added to a final protease:protein ratio of 1:50. Digestion was carried out at 37°C for 4 h with continuous stirring and was interrupted by freezing at Ϫ20°C.

RESULTS
The Effects of LDAO on the Mg 2ϩ -and Ca 2ϩ -ATPase Activities of CF 1 -The neutral detergent LDAO stimulates the ATPase activity of F 1 -ATPase preparations from Escherichia coli with and without the ⑀ subunit (39), and it also stimulates TF 1 -ATPase preparations with and without the ␦ and ⑀ subunits (40). Since the ATPase activity of latent CF 1 is activated by octylglucoside (7) or methanol (14), the effects of LDAO on the Mg 2ϩ -ATPase activity of CF 1 and the ␣ 3 ␤ 3 ␥ subcomplex of CF 1 were examined. Fig. 1A illustrates that LDAO stimulates the Mg 2ϩ -ATPase activity of heat-or DTT-activated CF 1 , latent CF 1 and the latent or DTT-activated ␣ 3 ␤ 3 ␥ subcomplex of CF 1 . Each was stimulated maximally at LDAO concentrations between 0.2 and 0.4%. In this range of LDAO concentration, heat-activated CF 1 , latent CF 1 , and the latent ␣ 3 ␤ 3 ␥ subcomplex were stimulated about 12-, 14-, and 6-fold, respectively. DTT-activated CF 1 and the DTT-activated ␣ 3 ␤ 3 ␥ subcomplex were stimulated 14-and 6-fold, respectively. The finding that LDAO stimulates the latent ␣ 3 ␤ 3 ␥ subcomplex only half as much observed for latent CF 1 and heat-activated CF 1 indicates that part of the simulation of intact CF 1 by LDAO is caused by displacement of the ⑀ subunit from an inhibitory position. The same argument applies to the difference in LDAO-induced stimulations observed for DTT-activated CF 1 and the DTTactivated ␣ 3 ␤ 3 ␥ subcomplex. A similar difference is observed when stimulation of E. coli F 1 by LDAO is compared with that of E. coli F 1 deleted of the ⑀ subunit (39).
In contrast to the Mg 2ϩ -ATPase activity, which is stimulated by LDAO, the Ca 2ϩ -ATPase shows a biphasic response to increasing concentrations of LDAO illustrated in Fig. 1B. The Ca 2ϩ -ATPase activity of heat-activated, DTT-activated CF 1 and the latent ␣ 3 ␤ 3 ␥ subcomplex are progressively inhibited by increasing concentrations of LDAO up to about 0.025%. Thereafter, the activity progressively activates, reaching a plateau where little further activation is observed with increasing LDAO concentration. A possible explanation for this apparent anomaly is as follows. The sharp dip in Fig. 1B occurring at about 0.025% LDAO may represent the critical micelle concentration of LDAO under the ionic conditions of the experiment. In water, the critical micelle concentration of LDAO is 0.05% (41). At LDAO concentrations over 0.025%, increasing concen-trations of micelles are present, which might stimulate ATPase activity.
Dequalinium Stimulates the Mg 2ϩ -ATPase Activity of CF 1 , Whereas It Inhibits Ca 2ϩ -ATPase Activity-When isolated from spinach chloroplasts, CF 1 is a latent ATPase that can be activated by a variety of methods. After heat activation or DTT activation, CF 1 has low Mg 2ϩ -ATPase activity of 0.54 mol of P i mg Ϫ1 min Ϫ1 or 0.59 mol of P i mg Ϫ1 min Ϫ1 , respectively, illustrated in Fig. 2A. Since dequalinium inhibits the ATPase activities of MF 1 and TF 1 (23,24), it was surprising to find that DTT activation of the ␣ 3 ␤ 3 ␥ subcomplex and DTT and heat activation of CF 1 were performed as described under "Experimental Procedures." A, stimulation of the Mg 2ϩ -ATPase. Samples of the activated forms of CF 1 and the ␣ 3 ␤ 3 ␥ subcomplex (7.5 g each) and 20-g samples of latent CF 1 were assayed with an ATPregenerating system containing 2 mM ATP, 1 mM MgCl 2 , and the concentrations of LDAO indicated. B, inhibition of the Ca 2ϩ -ATPase. Samples (15 g each) of the enzymes were assayed with 2.5 mM ATP, 2.5 mM CaCl 2 , and the concentrations of LDAO indicated using the ATP regeneration system described by Bruist and Hammes (36).
FIG. 2. The effects of dequalinium on the Mg 2؉ -ATPase and Ca 2؉ -ATPase activities. A, effects of dequalinium on Mg 2ϩ -ATPase activity. Heat and DTT activation were performed as described under "Experimental Procedures." Mg 2ϩ -ATPase was assayed as described in the legend of Fig. 1. B, inhibition of Ca 2ϩ -ATPase activity by dequalinium. Activation with octylglucoside was accomplished by incubating enzyme at 1.0 mg/ml with 30 mM octylglucoside for 10 min at 37°C in the presence of 5 mM ATP before diluting it 10-fold into 50 mM Tris-HCl, pH 8.0 (7). Heat-and DTT-activated CF 1 were also diluted to 0.1 mg/ml with 50 Tris-HCl and 50 mM Tris-HCl plus 1 mM DTT, respectively. Samples (50 l each) of the diluted stock solution were assayed for Ca 2ϩ -ATPase at 37°C in a medium containing 50 mM Tris-HCl, 5 mM ATP, and 5 mM CaCl 2 in the presence of the dequalinium concentrations indicated.
it stimulates the Mg 2ϩ -ATPase activity of heat-activated CF 1 and DTT-activated CF 1 , whereas it has no effect on the Mg 2ϩ -ATPase activity of latent CF 1 . Heat-activated CF 1 and DTTactivated CF 1 behaved differently in the presence of dequalinium. The heat-activated Mg 2ϩ -ATPase was maximally stimulated by about 50% with 10 M dequalinium. At higher concentrations, stimulation declined. In contrast, the Mg 2ϩ -ATPase activity of DTT-activated CF 1 was stimulated maximally by about 80% with 30 M dequalinium. However, in this case, stimulation did not decline with increasing dequalinium in the concentration range examined. In the presence of 0.2% LDAO, dequalinium does not stimulate the Mg 2ϩ -ATPase activity of heat-activated CF 1 over that stimulated by LDAO alone. In fact, concentrations of dequalinium greater than 15-20 M inhibit Mg 2ϩ -ATPase activity in the presence of 0.2% LDAO.
Whereas dequalinium inhibits the Mg 2ϩ -ATPase activity of the ␣ 3 ␤ 3 ␥ subcomplex of CF 1 before it is activated with DTT, the subcomplex is stimulated over 2-fold in the presence of 30 M dequalinium after it is activated with DTT. In the presence of 0.2% LDAO, dequalinium only slightly stimulated the Mg 2ϩ -ATPase activity of the DTT-activated ␣ 3 ␤ 3 ␥ subcomplex over that observed in the presence of LDAO alone. Fig. 2B shows that dequalinium inhibits the Ca 2ϩ -ATPase activity of CF 1 activated by treatment with DTT, heat, or octylglucoside. Unlike the Mg 2ϩ -ATPase activity, the Ca 2ϩ -ATPase activity is inhibited at all concentrations of dequalinium examined. A K i value of 4.0 M was determined from a Dixon plot for inhibition of heat-activated Ca 2ϩ -ATPase activity by dequalinium.
Photoinactivation of CF 1 with Dequalinium and the Effects of Ligands on Photoinactivation-When heat-activated CF 1 was irradiated at 350 nm in the presence of dequalinium, and then assayed for Mg 2ϩ -ATPase activity in the presence of octylglucoside (7), first order decay of enzyme activity was observed. Although in the presence of LDAO, low concentrations of dequalinium no longer stimulate the Mg 2ϩ -ATPase of heat-activated CF 1 , it has no effect on the rate of photoinactivation of heat-activated CF 1 by dequalinium (data not shown). Pseudofirst order rate constants, summarized in Table I, were determined for photoinactivation of heat-activated CF 1 by 12 M dequalinium in the presence or absence of ligands. Inorganic phosphate and pyrophosphate had almost no effect on the rate of photoinactivation, whereas ADP, ATP, and Mg 2ϩ provided protection against photoinactivation. The combination of ADP plus Mg 2ϩ and ATP plus Mg 2ϩ were the most effective, each decreasing the rate of inactivation nearly 10-fold. Ca 2ϩ alone or in combination with ADP or ATP also provides protection against photoinactivation, but is not as effective as Mg 2ϩ alone or in combination with ADP or ATP. Since ATP in the presence of Mg 2ϩ or Ca 2ϩ is hydrolyzed by heat-activated CF 1 , the latter combinations contained a mixture of MgADP and MgATP or CaADP and CaATP. Fig. 3 shows the pattern of radioactive peptides obtained when a sample of heat-activated CF 1 that had been photoinactivated by 80% with [ 14 C]dequalinium was submitted to SDS-PAGE. About 1.4 mol of reagent/mol of CF 1 was incorporated. The percentages of the 14 C in the labeled CF 1 applied to the gel that were recovered in the ␣, ␤, ␥, ␦, and ⑀ subunits are shown in Fig. 3. The 20% 14 C found in slices 2 and 3 apparently represents derivatized CF 1 that resisted depolymerization by SDS. A comparable band was not observed in the lane containing unmodified CF 1 submitted to SDS-PAGE on the same gel. It is clear that the ␤ subunit was predominantly labeled with [ 14 C]dequalinium. With the exception of the ␣ subunit, which was slightly labeled compared with ␤, very little radioactivity was covalently bound to the other subunits.

Photoinactivation of Heat-activated CF 1 with [ 14 C]Dequalinium Is Caused by Derivatization of Met 183 -
A tryptic digest of the isolated ␤ subunit prepared from the ␣ 3 ␤ 3 ␥ subcomplex derivatized with [ 14 C]dequalinium as described under "Experimental Procedures" was submitted to HPLC on a C 4 column equilibrated with 12 mM HCl. After the column was washed with 12 mM HCl for 10 min, it was developed with a linear gradient of 0 -15% acetonitrile in 12 mM HCl for 10 min and then with a linear gradient from 15 to 53% acetonitrile in 12 mM. A single radioactive peak eluted from the

FIG. 3. Resolution of labeled subunits by SDS-PAGE after inactivating CF 1 with [ 14 C]dequalinium.
Following heat activation of 1.0 mg of CF 1 in 0.5 ml, excess ATP was removed by passing the sample through two consecutive 5.0-ml centrifuge columns of Sephadex G-50 equilibrated with 50 mM Tris-HCl, pH 8.0. After adding ␤-mercaptoethanol and [ 14 C]dequalinium to final concentrations of 2 mM and 8 M, respectively, the protein solution was irradiated at 350 nm. When the octylglucoside-activated Mg 2ϩ -ATPase activity was photoinactivated by over 80%, the enzyme solution was passed through two consecutive centrifuge column of Sephadex G-50 to remove unbound [ 14 C]dequalinium. The resulting sample was applied to two lanes (30 g each) of a 12% SDS-polyacrylamide gel. After staining the gel with Coomassie Brilliant Blue G-250, the lanes containing photoinactivated CF 1 were sliced and corresponding slices from the two lanes were combined. The combined gel slices were digested in scintillation vials with H 2 O 2 -NH 4 OH (95:5) as described previously (23). The radioactivity in the digested samples was determined by liquid scintillation counting. column at 270 min and accounted for about 5% of the radioactivity applied. The resulting peptide was submitted to automatic Edman degradation, which revealed the sequence T-V-L-I-X-E-L-I-N-N corresponding to residues 179 -188 of the ␤ subunit. The "X" represents a cycle in which no detectable phenylthiohydantoin-amino acid was released. This is the position occupied by Met ␤183 in CF 1 .
Since the labeled tryptic peptide was obtained in only 5% yield, a peptic digest of the labeled ␤ subunit was fractionated by HPLC using the scheme illustrated in Fig. 4. Two radioactive fractions, P1-A and P1-B, were isolated, which accounted for 9.1% and 32.1%, respectively, of the total 14 C in the peptic digest submitted to the fractionation procedure. Automatic Edman degradation of the materials in P1-A and P1-B revealed the same tripeptide with the sequence: I-X-E. There are five segments in the ␤ subunit of CF 1 containing an I-X-E sequence. Only one of the tripeptide segments contains Met in the position designated X. This is Met ␤183 . The finding that the sequence of the labeled peptic peptides corresponds to the sequence of the labeled tryptic peptide strongly implicates Met ␤183 as the site of derivatization when CF 1 is photoinactivated with [ 14 C]dequalinium.
To ensure that photoinactivation of the different forms of CF 1 examined in this study proceeds with modification of Met ␤183 , 1-mg samples of latent CF 1 , heat activated CF 1 in the presence or absence of 0.1% LDAO, DTT-activated CF 1 , and the latent ␣ 3 ␤ 3 ␥ subcomplex of CF 1 were inactivated with 12 M [ 14 C]dequalinium using the spectrophotometric assay in the presence of octylglucoside to monitor inactivation. After photoinactivation, each sample was denatured and digested with pepsin. The peptic digests were submitted to HPLC under the conditions described above for the large scale preparation of peptic peptides derived from the isolated ␤ subunit derivatized with [ 14 C]dequalinium. The profiles of radioactivity in the effluents of the peptic digests were essentially identical to the profiles obtained in the large scale fractionation of peptic peptides derived from the isolated ␤ subunit derivatized with [ 14 C]dequalinium. From these analyses, we conclude that Met ␤183 is the major site of derivatization when each form of CF 1 was photoinactivated with [ 14 C]dequalinium. Given that 27% of the ␤ subunit was derivatized on 80% photoinactivation of the ␣ 3 ␤ 3 ␥ subcomplex of CF 1 by [ 14 C]dequalinium, we also conclude that modification of Met 183 in a single ␤ subunit is sufficient for complete photoinactivation. DISCUSSION It is clear from the results presented that photoinactivation of the various forms of CF 1 with dequalinium is caused by derivatization of Met ␤183 . This residue corresponds to Met ␤167 of MF 1 (42). Fig. 5 shows that Met ␤167 of MF 1 is among a cluster of apolar amino acid residues that line a cavity surrounding the side chain of Glu ␤199 , a conserved residue near the catalytic nucleotide binding site. It is possible that the positive charge on one of the quinaldinium moieties of dequalinium interacts with the side chain of the equivalent of Glu ␤199 in CF 1 . Comparison of the unliganded catalytic site, ␤ E , illustrated in Fig. 5A, with the catalytic site of MF 1 liganded with AMP-PNP clearly shows that the side chain of Met ␤167 shifts away from the entrance of this cavity when the catalytic site is liganded. In ␤ DP (data not shown), the side chain of Met ␤167 is also shifted away from the entrance to the cavity. This is consistent with the observation that ADP and ATP, especially when complexed with Mg 2ϩ , protect CF 1 against photoinactivation by dequalinium. It is interesting that photoinactivation of TF 1 with dequalinium, which has a glutamine residue in the position corresponding to Met ␤167 of MF 1 (43), leads to derivatization of Phe ␤420 , which corresponds to Phe ␤424 of MF 1 . Fig. 5 shows that Phe ␤424 is near the entrance of the hydrophobic cavity. The change in orientation of helix B with respect to helix C is extraordinarily large in going from the unliganded catalytic site shown in Fig.  5A to the liganded catalytic site shown in Fig. 5B. Derivatization of either Met ␤183 in CF 1 or Phe ␤420 in TF 1 probably blocks this conformational change.
The finding that dequalinium binds at or near a catalytic site of CF 1 is difficult to reconcile with the results of Groth and Junge (27), who reported that ADP blocks proton slip mediated by CF 0 F 1 at the electrogenic step, whereas dequalinium blocks the proton release step. To explain the differential effects of dequalinium and ADP on proton slip, Groth and Junge (27) assumed that dequalinium does not bind to a catalytic site.
The anomalous finding that, in the absence of light, dequalinium stimulates Mg 2ϩ -ATPase activity, whereas it inhibits Ca 2ϩ -ATPase activity of CF 1 activated by dithothreitol or heat, can be explained by assuming that, in addition to binding to an inhibitory site at or near a catalytic site, dequalinium binds to a stimulatory site that also binds LDAO. The finding that dequalinium stimulates the Mg 2ϩ -ATPase activity of heat-activated CF 1 at lower concentrations of LDAO, whereas it inhibits the ATPase activity at higher concentrations of LDAO, is consistent with this assumption. LDAO stimulates the ␣ 3 ␤ 3 ␥ subcomplex of TF 1 by promoting dissociation of inhibitory MgADP from a catalytic site during turnover (40). Since isolated CF 1 contains ADP tightly bound to a single catalytic site, which is converted to inhibitory MgADP in the presence of medium Mg 2ϩ (44,45), that part of activation of CF 1 by LDAO that is independent of the ⑀ subunit is likely to occur by the same mechanism. In experiments not shown, it was found that addition of low concentrations of dequalinium to heat-or DTTactivated CF 1 in the presence of 0.2% LDAO does not stimulate Mg 2ϩ -ATPase activity above the level observed in the presence of LDAO alone. This suggests that dequalinium also stimulates the Mg 2ϩ -ATPase activity of CF 1 in the same manner as LDAO. However in the presence of LDAO, higher concentrations of dequalinium still slightly inhibit the Mg 2ϩ -ATPase of heat-activated CF 1 . This also suggests that CF 1 contains a low affinity inhibitory site for dequalinium that does not bind LDAO. In contrast, the Ca 2ϩ -ATPase of CF 1 is not inhibited by CaADP bound to a single catalytic site (46). Fig. 2B shows that FIG. 4. Summary of the purification of the major radioactive peptides in a peptic digest of ␤ subunit from CF 1 inactivated with [ 14 C]dequalinium. The ␤ subunit was purified and digested as described under "Experimental Procedures." The peptic digest was first submitted to HPLC on a C 4 column that was equilibrated with 12 mM HCl and eluted with a gradient (10 -42%) of acetonitrile in 12 mM HCl in 160 min. Peak P1, which contained the majority of 14 C, eluted from the C 4 column was concentrated under vacuum and submitted to HPLC on a phenyl column, which was equilibrated with 12 mM HCl and eluted with a gradient (15-35%) of acetonitrile in 12 mM HCl in 200 min. The material in peaks P1-A and P1-B were submitted to automatic sequence analysis. The percentages in parentheses represent yields of radioactivity obtained for each column step. dequalinium is a partial inhibitor of the Ca 2ϩ -ATPase activity of CF 1 . Dequalinium is also a partial inhibitor of MF 1 and TF 1 (22,24). Maximal inhibition of each enzyme by dequalinium is not greater than 85-90%. Therefore, dequalinium may act both as an activator of the Mg 2ϩ -ATPase activity of CF 1 , by promoting dissociation of inhibitory MgADP from a catalytic site, and as a partial inhibitor of the Mg 2ϩ -ATPase activity that it activates. The combined effect of dequalinium is the relatively modest stimulation of Mg 2ϩ -ATPase activity observed, compared with that promoted by LDAO.
In contrast to its effect on Mg 2ϩ -ATPase activity, dequalinium only inhibits the Ca 2ϩ -ATPase activity of heat-, DTT-, and octylglucoside-activated CF 1 . Unlike the Mg 2ϩ -ATPase activity, the Ca 2ϩ -ATPase activities of heat-activated CF 1 , DTTactivated CF 1 , and the latent ␣ 3 ␤ 3 ␥ subcomplex are inhibited by low concentrations of LDAO. These observations support an earlier report (46) that the Ca 2ϩ -ATPase activity of CF 1 is not subject to inhibition caused by binding of CaADP in a single catalytic site. Consistent with this argument, Leckband and Hammes (47) reported that the initial rate of Ca 2ϩ -ATP hydrolysis by heat-activated CF 1 is triggered by free Ca 2ϩ . In contrast, free Mg 2ϩ inhibits the Mg 2ϩ -ATPase activity of CF 1 by interacting with ADP bound to a catalytic site to form the inhibited CF 1 ⅐ADPMg complex.
The finding that derivatization occurs at or near a catalytic site when CF 1 or TF 1 are photoinactivated with dequalinium, whereas a different site is derivatized when MF 1 is inactivated with the reagent, might be related to the insensitivity of spinach CF 1 and TF 1 and sensitivity of MF 1 to inhibition by aurovertin (48). In the case of MF 1 , photoinactivation with dequalinium is accompanied by derivatization of Phe ␣403 or Phe ␣406 , in mutually exclusive reactions, and a side chain within residues 440 -459 of the ␤ subunit. To a limited extent, ␣-␤ crosslinking accompanies photoinactivation (23). The crystal struc-ture of MF 1 liganded with 2 aurovertin molecules has been solved (49). One aurovertin molecule is bound to ␤ TP , and the other is bound to ␤ E . In ␤ TP , the pyrone ring of bound inhibitor interacts with the side chain of Tyr ␤458 and the methoxy oxygen attached to the pyrone is in van der Waals contact with a carboxylate oxygen of Glu ␣199 of ␣ TP . At an ␣ TP /␤ TP interface, Phe ␣403 is 16.4 Å from Tyr ␤458 , which is within the spanning distance of the two 4-amino-2-methylquinaldinium moieties of dequalinium. Whether or not Tyr ␤458 is derivatized when MF 1 is photoinactivated with dequalinium has not been established. However, an investigation is in progress to identify the residue derivatized within residues 440 -459 of the ␤ subunit when MF 1 is inactivated with [ 14 C]dequalinium.