The 20 C-terminal Amino Acid Residues of the Chloroplast ATP Synthase γ Subunit Are Not Essential for Activity*

It has been suggested that the last seven to nine amino acid residues at the C terminus of the γ subunit of the ATP synthase act as a spindle for rotation of the γ subunit with respect to the αβ subunits during catalysis (Abrahams, J. P., Leslie, A. G. W., Lutter, R., and Walker, J. E. (1994)Nature 370, 621–628). To test this hypothesis we selectively deleted C-terminal residues from the chloroplast γ subunit, two at a time starting at the sixth residue from the end and finishing at the 20th residue from the end. The mutant γ genes were overexpressed in Escherichia coli and assembled with a native α3β3 complex. All the mutant forms of γ assembled as effectively as the wild-type γ. Deletion of the terminal 6 residues of γ resulted in a significant increase (>50%) in the Ca-dependent ATPase activity when compared with the wild-type assembly. The increased activity persisted even after deletion of the C-terminal 14 residues, well beyond the seven residues proposed to form the spindle. Further deletions resulted in a decreased activity to ∼19% of that of the wild-type enzyme after deleting all 20 C-terminal residues. The results indicate that the tip of the γC terminus is not essential for catalysis and raise questions about the role of the C terminus as a spindle for rotation.

The ATP synthase enzymes of the inner membranes of mitochondria, chloroplasts and of the bacterial cytoplasmic membrane, couple the energy of a transmembrane electrochemical proton gradient to the synthesis of ATP from ADP and inorganic phosphate. The general structural features of the enzyme are highly conserved from one organism to another. It is comprised of an integral membrane-spanning H ϩ -translocating segment (F 0 or factor O) and a peripheral membrane segment (F 1 or factor 1) which contains the catalytic sites for ATP synthesis and hydrolysis. The F 1 segment is comprised of five different polypeptide subunits designated ␣ to ⑀ in order of decreasing molecular weight. The subunit stoichiometry is ␣ 3 ␤ 3 ␥ 1 ␦ 1 and ⑀ 1 . Nucleotide binding is associated with the ␣ and ␤ subunits, whereas the ␥ and ⑀ subunits play regulatory and/or structural roles. The ␦ subunit is likely to be involved in binding the F 1 segment to the F 0 segment (reviewed in Ref. 1).
A high resolution crystal structure of the core catalytic por-tion of the mitochondrial F 1 enzyme was reported recently (2). The ␣ and ␤ subunits alternate with each other to form a hexameric ring with one nucleotide binding site located at each of the six ␣␤ subunit interfaces. Part of the structure of the ␥ subunit was also solved, including well conserved regions of the N and C termini. The C terminus, from residues 209 to 272 forms a single ␣ helix that stretches from below the base to the top of the ␣␤ hexamer (see Fig. 5). The last nine residues of this remarkably long helix are predominantly hydrophobic in nature and pass through a greasy sleeve formed by a ring of hydrophobic residues provided by interacting N-terminal ␤ barrel domains of all six of the ␣ and ␤ subunits. On the basis of this unusual asymmetric structure, it was suggested that the C-terminal helix of the ␥ subunit forms a spindle around which the ␣␤ hexamer rotates, rotation being facilitated by the hydrophobic (greasy) nature of the amino acids involved. That is, the ␣␤ subunits provide a bearing through which the tip of the ␥ subunit passes and within which the ␥ subunit rotates. Although the amino acid sequences of ␥ subunits from different organisms show little overall homology, segments near the N and C termini are quite well conserved suggesting that they may be involved in forming important contacts with other F 1 subunits (3,4). The crystal structure of F 1 indicated that the three ␣␤ pairs of the ␣␤ hexamer also make direct contact with other regions of the ␥ subunit to induce different conformational states of the nucleotide binding sites at the ␣/␤ subunit interfaces. During rotation, each nucleotide binding site would sequentially alternate between three different conformational states, each state dictated by a different type of interaction with the ␥ subunit. Such rotation has been predicted from kinetic studies (5, 6), has been supported by several recent experiments (7)(8)(9), and is now widely considered to be a general mechanistic feature of all of the F 1 enzymes (reviewed in Ref. 10).
In this study we have tested the "bearing" hypothesis specifically suggested by the crystal structure, by selectively deleting amino acid residues from the extreme C-terminal end of the ␥ subunit, which, in the mitochondrial enzyme, extends through the greasy sleeve of the ␣␤ hexamer. To do this we utilized an efficient reconstitution system reported earlier (11) in which the native ␥ subunit isolated from CF 1 1 was reconstituted with an isolated ␣␤ subunit complex. The cloned ␥ subunit could effectively replace the native ␥ subunit in reconstitution of the core enzyme complex. 2 Eight genetically engineered ␥ subunits, lacking between 6 and 20 of the C-terminal amino acids, were tested for assembly with the ␣␤ subunits, and the catalytic activities of the assembled complexes were examined. The re-sults demonstrate that the tip of the C terminus of the ␥ subunit, from residues 304 to 323 (chloroplast numbering), is not essential for rapid turnover by CF 1 .

EXPERIMENTAL PROCEDURES
Materials-CF 1 and CF 1 lacking the ␦ and ⑀ subunits, CF 1 (Ϫ␦⑀), were prepared from fresh market spinach as described previously (12) and stored as ammonium sulfate precipitates. Prior to use the proteins were desalted on Sephadex G-50 centrifuge columns (13). The isolated ⑀ subunit (14) was stored in the isolation buffer at 4°C. An ␣␤ complex and the ␥ subunit were isolated from CF 1 (Ϫ␦⑀) as described previously (11). The ␣␤ subunit complex was recycled through the isolation procedure to ensure that trace amounts of contaminating ␥ subunit were removed.
ATP (grade I and II) and antibiotics (ampicillin, chloramphenicol, and tentoxin) were purchased from Sigma. Stock solutions of tentoxin were prepared by dissolving the inhibitor in ethanol to a final concentration of 5 mM and stored at Ϫ70°C. Pfu DNA polymerase and its reaction buffer were purchased from Strategene. T4 DNA ligase and its reaction buffer were obtained from Promega. DNase I was from Roche Molecular Biochemicals. Tryptone and yeast extract were obtained from Difco. Urea (ultra pure) was purchased from Fluka and hydroxylapatite from Bio-Rad. All other chemicals were of the highest quality reagent grade available.
Plasmid Construction-Most of the recombinant DNA methods used in this study have been described elsewhere (15,16). Escherichia coli transformation protocols were as described by Hanahan (17). Plasmid pSG101 (4), generously supplied by Dr. M. Futai, contains the fulllength cDNA for the spinach (Spinacia oleracea) chloroplast atpC gene encoding the ATP synthase ␥ subunit. A 1.1-kilobase pair BsaI-BamHI fragment of pSG101 was subcloned into the NcoI and BamHI cleaved expression vector pET8c (18) via an NcoI-BsaI adaptor. 2 The resulting plasmid pET8cgam bb1 was transformed into the expression host E. coli BL21(DE3)/pLysS (19). Plasmid DNA for sequencing was prepared by alkaline-SDS lysis and polyethylene glycol precipitation (20).
Generation of atpC Gene Mutants-Eight deletion mutants of ␥ were generated by "inverse" PCR with a forward primer that was complementary to the termination codon of the atpC gene and the downstream sequence of the pET8cgam bb1 plasmid. The reverse primer was complementary to the required C-terminal amino acid and its adjacent upstream sequence. PCR primers were 24 -31 base pairs long and were 5Ј-phosphorylated. Oligonucleotides were synthesized by Macromolecular Resources, Colorado State University. Plasmid DNA for PCR was prepared by ethanol precipitation after phenol:chloroform extraction (17). PCR was carried out in 50 l of cloned Pfu DNA polymerase reaction buffer, which also contained 60 ng of the pET8cgambb1 plasmid, 4 mM total MgSO 4 , 22 pmol of each primer, 0.4 mM dNTPs, and 2.5 units of cloned Pfu DNA polymerase. The components were mixed on ice and placed in a GenAmp PCR System 2400 (Perkin-Elmer) prewarmed to 94°C. Cycling parameters were: 94°C for 1 min, 56°C for 1 min, 72°C for 12 min, for 20 cycles. The PCR product was purified by agarose gel electrophoresis followed by electroelution into an ISCO micro-trap. The eluted DNA was precipitated with ethanol and circularized (14). For this 100 -200 ng of the DNA was incubated with 3 units of T4 DNA ligase in the T4 DNA ligase buffer overnight at room temperature (ϳ22°C). The resulting plasmid was transformed into E. coli XL1-Blue cells for amplification. The amplified plasmid was isolated using boiling lysis followed by isopropanol and ethanol precipitation and transformed into the expression host E. coli BL21(DE3)/pLysS (19).
Each mutant gene was isolated from the expression clone by alkaline-SDS lysis followed by ethanol precipitation after phenol:chloroform extraction (20) and sequenced by an automated fluorescence dideoxy technique (21).
Solubilization and Folding of Overexpressed ␥ Mutants-E. coli cells containing the atpC gene were grown at 37°C in LB medium containing L-ampicillin (100 mg/ml) and chloramphenicol (34 mg/ml). In mid-exponential phase growth, cells were induced with 0.1 mM isopropyl-␤-Dthiogalactopyranoside for up to 5 h. Cells were harvested by centrifugation at 4000 ϫ g for 10 min, washed once with TE50/2 buffer (50 mM Tris-HCl, 2 mM EDTA, pH 8.0), and resuspended in a small volume (10 -15 ml) of TE50/2. Cells were lysed by one to three cycles of freezing (at Ϫ70°C or in a dry ice/ethanol bath) and thawing (15). 10 mM MgCl 2 and 10 mg of DNase I were added to the lysed cells, which were incubated on ice for 20 min. DNA was then sheared by sonication with a Branson 250 sonifier for 2 ϫ 15 s at an output of 4 and a duty cycle of 10. After the sonication cells were kept on ice for additional 20 min. Inclusion bodies, together with some cell debris, were sedimented at 4000 ϫ g for 10 min. The pellet, containing mostly insoluble ␥ polypeptide, was washed three times with 25 ml of TE50/2 before solubilization.
Assembly of ␥ Mutants-The purified ␣␤ mixture was diluted to about 100 g/ml with a solution containing 20% glycerol, 50 mM Hepes-NaOH, pH 7.0, 2 mM MgCl 2 , 2 mM ATP, and 2 mM dithiothreitol and kept on ice. The ␥ subunit preparation was added dropwise to the ␣␤ mixture to give a final molar ratio of 3␥:1␣␤. The mixture was gently mixed and left to sit at room temperature (ϳ22°C) for 2 h. Unreconstituted subunits were separated from the reconstituted ␣␤␥ by anion exchange chromatography as described previously (11).
Other Procedures-ATPase activities were determined by measuring phosphate release (22) for 5 min at 37°C. The assay was carried out in 0.5-ml volumes of assay mixture containing 50 mM Tricine-NaOH, pH 8.0, and 5 mM ATP. The calcium-dependent ATPase activity was assayed in the presence of 5 mM CaCl 2 . Magnesium-dependent ATPase was carried out in the presence of 2.5 mM MgCl 2 and 25 mM Na 2 SO 3 , and manganese-dependent ATPase activity was carried out in the presence of 2.5 mM MnCl 2 and 100 mM Na 2 SO 3 . The reaction was started by addition of 1-6 g of enzyme into the assay mixture and terminated by addition of 0.5 ml of cold trichloroacetic acid. Protein concentrations were determined by the Bradford method (23). Absorbance measurements were obtained using a Beckman DU-70 spectrophotometer. Gel electrophoresis was performed on NOVEX Pre-Cast 10 -20% gradient gels.

RESULTS
Overexpression of the Spinach atpC Gene in E. coli-The atpC gene encoding the full-length ␥ subunit of the spinach chloroplast ATP synthase was inserted into the pET8c expression vector and overexpressed at high levels (Ͼ100 mg/liter of cells at the end of log-phase growth). The overexpressed protein was solubilized from insoluble inclusion bodies into 4 M urea and recovered by slow dialysis. The cloned protein was identical to the native protein (11) in its ability to reconstitute with native ␣␤ subunits to form a fully active core enzyme complex. 2 Eight deletion mutants of the atpC gene were prepared and transformed into the overexpression host and the deletions verified by sequencing each entire gene. The truncated polypeptides were designated ␥ D6 to ␥ D20 , according to the number of amino acid residues missing from the C terminus. Amino acid sequences of the C-terminal fragment of the fulllength ␥ subunit and the eight deletion mutants are shown in Fig. 1. Also shown in Fig. 1 is the corresponding sequence at the C terminus of the bovine mitochondrial F 1 ␥ subunit. The sequence underlined corresponds to that part of the ␥ subunit that is in the immediate vicinity of the hydrophobic sleeve, the last seven residues (267-273) actually passing through the sleeve (2). Deletion of all ten C-terminal residues would arguably be sufficient to test the bearing hypothesis. The C-terminal segment of ␥ shown in Fig. 1 is one of the most highly conserved regions among ␥ subunits from different species. This is evident from the more than 50% direct sequence identity between the bovine mitochondrial and chloroplast subunits (Fig. 1).
Assembly of the ␥ Mutants-Each of the ␥ constructs was tested for its ability to organize the ␣␤ subunits into a stable ␣␤␥ core enzyme complex. For this, folded ␥ polypeptide was incubated with the isolated ␣␤ complex, and the resulting ␣␤␥ assembly was purified by DEAE-cellulose column chromatography as described earlier for purifying ␣␤␥ assembled using the native F 1 subunits (11). Incubation of each of the ␥ constructs with the ␣␤ subunits resulted in formation of an ␣␤␥ complex, which is eluted from DEAE-cellulose at the same salt concentration as the native complex and which is significantly higher than that required to elute unassembled subunits (11).
The polypeptide profiles of all of the assemblies were very similar to each other as indicated for the ␣␤␥ D12 and ␣␤␥ D20 assemblies, which are compared with the ␣␤␥ WT assembly in Fig. 2. This suggests that all of the ␥ mutants were capable of assembling with the ␣␤ subunits.
The results shown in Table I compare the ATPase activities of protein assemblies reconstituted with the first two mutants, ␥ D6 and ␥ D8 with the wild-type ␥. Remarkably, both mutant assemblies were significantly more active than the wild-type assembly in calcium-dependent ATP hydrolysis. The magnesium-dependent activities of the two mutants, however, were significantly reduced. The apparent K m and K cat for Ca-ATP hydrolysis of the ␥ D6 mutant were measured and compared with the wild-type assembly (Table I). Only the K cat exhibited a measurable change in the mutant. Fig. 3 compares the relative rates of ATP hydrolysis of the remaining mutants, ␥ D10 through ␥ D20 , in the presence of either Ca 2ϩ , Mg 2ϩ , or Mn 2ϩ as the divalent cation substrate. The ␥ D10 , ␥ D12 , and ␥ D14 mutant assemblies all showed similar responses to those of the ␥ D6 and ␥ D8 mutant assemblies in that their Ca-ATPase activities were significantly higher than that of the wild-type enzyme. The maximum activity was obtained with the ␥ D14 mutant, which had a specific activity of 55 mol⅐min Ϫ1 ⅐mg Ϫ1 , which is the highest rate of Ca-ATP hydrolysis that we have ever observed with the chloroplast enzyme. However, deletion of 16 residues from the ␥C terminus resulted in a sharp decrease in Ca-ATPase activity, which continued upon deletion of additional residues ending with an activity that was ϳ19% of the wild-type control at ␥ D20 . In contrast to the Ca-ATPase activity, the Mg-ATPase and Mn-ATPase activities declined continuously with each additional pair of residues deleted. Nevertheless, even after deleting 20 residues from the C terminus, the enzyme exhibited significant rates of catalysis: 17% of the wild-type Mg-ATPase activity and 20% of the wild-type Mn-ATPase activity.
Activation of the latent Mg-ATPase and Mn-ATPase activities of CF 1 normally requires, in addition to removing the inhibitory ⑀ subunit (14), the presence of oxyanions such as ethanol, carbonate, or sulfite, which overcome a strong inhibition caused by free metal ions binding to and stabilizing bound ADP at the catalytic site(s) (24). The degree of stimulation by oxyanions usually varies between 10-and 100-fold depending on the divalent cation and the oxyanion concentrations. The Ca-ATPase activity, however, is already high once ⑀ is removed and is slightly inhibited by oxyanions (25). So the magnesiumand manganese-dependent ATPase activities listed in Table I and shown in Fig. 3 were measured in the presence of high concentrations (25 mM) of sulfite ions. It was of interest to examine the affects of the ␥ deletions on the Mg-ATPase activities in the absence of the stimulatory oxyanions. The results of this study are shown in Table II. The Mg-ATPase activity in the absence of sulfite was, like the Ca-ATPase activity, stimulated by deletion of residues from the ␥C terminus and was highest in the ␥ D14 mutant. The activity of this mutant was almost 4-fold that of the wild-type enzyme, and in parallel to the Ca-ATPase activity, it decreased markedly upon deletion of 16 or more residues. The ␥ D20 mutant still retained a readily measurable activity, which was ϳ45% of that of the wild-type enzyme (Table II).
Sensitivity of the Mutant Assemblies to Inhibitors-The responses of the different assemblies to the inhibitory ⑀ subunit and to the fungal inhibitor tentoxin were examined, in part to evaluate the effect of the deletions on the ability of the two inhibitors to block activity and in part to verify that the observed activities are representative of the normal activity of CF 1 , which responds to these inhibitors with absolute specificity. The inhibitory responses of the Ca-ATPase activities of the different constructs to a fixed concentration (10-fold molar excess) of added ⑀ subunit are summarized in Table III. All of the enzyme assemblies, including the enzyme assembled with the ␥ D20 mutant, were strongly inhibited by ⑀, although there was a significant variation (between 64 and 83%) in the extent of inhibition observed, and all were less inhibited than the wildtype assembly (91%). The ␥ D14 mutant, which exhibited the highest activity, was the least inhibited in the presence of a 10-fold molar excess of ⑀. However, in the presence of a 30-fold molar excess of the ⑀ subunit, the ␥ D14 mutant was inhibited by the same extent as the wild-type enzyme (results not shown), indicating that the deletion had reduced the apparent affinity of the enzyme for ⑀ but not the maximal extent of inhibition.
The results of titrating the ␣␤␥ assemblies with tentoxin are shown in Fig. 4. All of the assemblies, with the exception of the ␥ D20 mutant, were sensitive to inhibition by tentoxin. There were, however, significant differences among the mutant en- , wild-type spinach chloroplast (CF 1 -␥ WT ), and C terminus deletion mutants of spinach chloroplast ␥ subunit beginning with deletion of 6 residues from the C terminus (␥ D6 ) continuing with successive deletion of 2 residues up to 20 residues (␥ D20 ).

FIG. 2. Gel electrophoresis profile of the reconstituted, purified ␣␤␥ assemblies and isolated CF 1 (؊␦⑀). Wild-type and mutant
␥ subunits were reconstituted with the native ␣␤ subunits and the protein assemblies purified by ion exchange chromatography as described previously (9). Electrophoresis was performed on a 10 -20% Tris-glycine gel, and proteins were stained with Coomassie Brilliant Blue R. Each lane contained 4 g of protein.
A, isolated CF 1 lacking the two small subunits, ␦ and ⑀ (CF 1 (Ϫ␦⑀)); WT, D12 and D20, the ␣␤␥ assemblies containing the wild-type ␥, ␥ D12 , and ␥ D20 mutants, respectively. zymes in the concentrations of tentoxin required to reach maximum inhibition. The most obvious differences were with the longer deletions. For example, a greater than 20-fold higher concentration was required for 90% inhibition of the activities of the ␥ D16 and ␥ D18 mutants than that required to inhibit the ␥ WT to the same extent. DISCUSSION A cross-sectional view through the structure of the beef heart mitochondrial F 1 is shown in Fig. 5. The tip of the C terminus of the ␥ subunit, more specifically the last 7-10 residues, is surrounded by a sleeve of residues formed by part of the tightly packed ␤ barrel domains of the six ␣ and ␤ subunits. The sleeve residues, located in the region marked A on the ␤ subunit in Fig. 5, have an overall hydrophobic character as do the nearby residues on the ␥ subunit. A hydrophobic contact between ␥ and the surrounding sleeve could allow the ␥ subunit to act as a spindle around which the ␣␤ hexamer could rotate with minimum frictional resistance (2). The base of the C-terminal helix  a Complexes were reconstituted and purified by anion exchange chromatography according to Ref. 11. The values listed are averages of three independent measurements. b Mg-ATPase activity was determined in an assay mixture containing 50 mM Tricine-NaOH, pH 8.0, 5 mM ATP, 2.5 mM MgCl 2 , 5 g of enzyme, and sodium sulfite as indicated.
FIG. 3. Relative ATPase activity of the reconstituted, purified ␣␤␥ constructs. All of the assays were carried out as described under "Experimental Procedures." The columns represent the relative ATPase activities of the different ␣␤␥ assemblies in the presence of calcium chloride: white, magnesium chloride; gray, manganese chloride; black, D10 to D20 are designations for ␣␤␥ D10 to ␣␤␥ D20 assemblies. Activities of the ␣␤␥ WT (100% controls) were: Ca-ATPase, 35.5 Ϯ 1.1; Mg-ATPase, 47.5 Ϯ 1.5; and Mn-ATPase, 67.9 Ϯ 5.4 mol⅐min Ϫ1 ⅐mg Ϫ1 . Values shown are the averages and S.D. for five separate determinations. a Assemblies were incubated with a 10-fold molar excess of the ⑀ subunit or an identical volume of the buffer (control) used to isolate the ⑀ subunit (7). Samples were incubated for 8 min at room temperature and 2 min at 37°C immediately prior to assay. Other conditions are the same as described in Table I. FIG. 4. The effect of tentoxin on the Ca-ATPase activities of the reconstituted, purified ␣␤␥ assemblies. Ca-ATPase activity was measured as described under "Experimental Procedures." Each assembly was incubated with tentoxin to give the indicated tentoxin/enzyme ratios in the ATPase assay medium for 10 min at room temperature (ϳ22°C) then 2 min at 37°C. The reactions were started by addition of 5 mM CaCl 2 . The ATPase activities in the absence of tentoxin were: ␣␤␥ WT , 16.9 mol⅐min Ϫ1 ⅐mg Ϫ1 (f); ␣␤␥ D10 , 35.1 mol⅐min Ϫ1 ⅐mg Ϫ1 (q); ␣␤␥ D12 , 25.9 mol⅐min Ϫ1 ⅐mg Ϫ1 (Ⅺ); ␣␤␥ D14 , 33.9 mol⅐min Ϫ1 ⅐mg Ϫ1 (OE); ␣␤␥ D16 , 24.9 mol⅐min Ϫ1 ⅐mg Ϫ1 (Ⅺ); ␣␤␥ D18 , 15.0 mol⅐min Ϫ1 ⅐mg Ϫ1 (E); ␣␤␥ D20 , 3.1 mol⅐min Ϫ1 ⅐mg Ϫ1 (‚). of ␥ is offset from the central axis of the hexamer by ϳ7 Å, so that, provided it remained rigid, it would sequentially and reversibly come into contact with regions of the ␣ and ␤ subunits during rotation to create the required asymmetry among the nucleotide binding sites.
The remarkably high amino acid sequence conservation among the ␣ and ␤ subunits of F 1 enzymes from different species, together with the fact that the structures of the ␣ and ␤ subunits of a thermophilic bacterium can be essentially superimposed upon those of the mitochondrial F 1 subunits (26), are cogent reasons for assuming that all of the F 1 enzymes have a very similar overall structure and utilize the same basic mechanism for ATP synthesis. There is, however, some evidence suggesting that there may be some minor structural differences among the F 1 enzymes. For example, site mapping studies of the chloroplast F 1 using fluorescence resonance energy transfer (27) as well as chemical cross-linking experiments (28) have indicated that cysteine 322, which is the second last amino acid residue at the C terminus of the CF 1 ␥ subunit, is located near the base of the ␣␤ hexamer, more than 60 Å away from its position in the mitochondrial F 1 . The reason for this difference is not understood at this time but is particularly intriguing given the significant amino acid sequence homology which is apparent in the C-terminal domains of ␥ subunits from different organisms (Ref. 4; also see Fig. 1). Moreover, a different location for the ␥C terminus implies that the idea that the C-terminal helix of ␥ acts as a spindle for rotation is probably not correct, at least not as originally envisioned based on the mitochondrial F 1 structure (2).
We have selectively deleted part of ␥C terminus reasoning that if this region of ␥ was indeed acting as the tip of a spindle for rotation, or if it was in any way critical for catalysis by CF 1 , the deletion should result in a complete loss of catalytic activity. However, the enzyme containing mutant ␥ subunits missing up to 20 amino acids from the C terminus was still capable of significant catalytic activity, which, with the exception of the D20 mutant, was sensitive to specific allosteric inhibitors of CF 1 ; a strong indication that the mutant enzymes followed the usual cooperative catalytic pathway and that their activities were not artifactual. It is noteworthy that a similar result was obtained for E. coli F 1 (29). In that case, membranes containing the mutant enzyme retained a limited catalytic activity (ϳ10% of both ATP hydrolysis and synthesis) following deletion of 10 residues from the C terminus. Deletion more than 10 residues from the C terminus resulted in a complete loss of activity in that case, although the enzyme still apparently correctly assembled on the membrane. The greater sensitivity of the E. coli enzyme to deletion of the C terminus may reflect a different structural requirement for catalysis by the F 0 -F 1 complex than for the isolated F 1 . The activity of the E. coli F 1 mutants following isolation from the membrane was not investigated in that study.
The initial activation of enzyme turnover upon deletion of up to 14 residues from the C terminus of ␥ occurred for both the calcium-and the magnesium-dependent ATPase activities. One likely explanation for this effect is that the deletions resulted in a partial loosening of the structure of the enzyme to a point where it weakened binding of the cation-ADP reaction product at the catalytic site(s) in the interfacial region between ␣ and ␤ subunits. Since the off-rate of the cation-ADP limits the overall reaction rate, the end result is to increase the K cat of the enzyme. This would also explain why the sulfite-stimulated activity is inhibited rather than stimulated by the deletions. Assuming that the presence of high concentrations of sulfite maximally reduce the off-rate of cation-ADP so that the on-rate of the cation-ATP now becomes rate-limiting, a further reduction in nucleotide affinity at the catalytic site caused by the ␥ deletions would result in a reduction in the on-rate for the cation-ATP substrate and thus a reduction in the K cat . This could also explain the reduced apparent affinities for the ⑀ subunit and for tentoxin, which resulted from the C-terminal deletions. Both inhibitors are known to block cooperative release of bound nucleotides, probably by stabilizing a rigid inhibited conformation of CF 1 (1,12). Thus a loosening of the F 1 structure might favor the activated conformation over the inhibited conformation. The structural change resulting in altered inhibitor affinity does not necessarily have to be large, since small perturbations of ␥ structure, such as reduction of the ␥ disulfide bond or single-site cleavage of ␥ by trypsin, are known to markedly decrease the affinity of CF 1 for the ⑀ subunit (30).
The chloroplast ␥ subunit has a glycine residue at position 310, 14 residues in from the C-terminal end (Fig. 1). Most secondary structural prediction algorithms predict a break in the C-terminal helix of ␥ at Gly-310. If the CF 1 ␥ were to turn back on itself at this point, the cysteine at position 322 would face toward the bottom part of CF 1 (i.e. toward the membrane in CF 1 -F 0 ) and come close to the position of this residue determined by fluorescence distance mapping (27). This would create a three-helix bundle in the central cavity of the enzyme rather than the two-helix bundle identified in the mitochondrial enzyme (2). If this is the case, the important binding interactions between ␥ and the ␣␤ subunits identified in the mitochondrial enzyme, and which are likely to be primarily responsible for creating asymmetry among the different catalytic sites, might also be preserved in CF 1 . Deleting the 14 C-terminal residues from CF 1 ␥ would remove the third helix from the central bundle, possibly decreasing the number of contacts between ␥ and the ␣␤ subunits. This could feasibly have the effect of loosening the structure, thereby weakening the nucleotide affinity. The sharp decrease in catalytic activity, which occurred upon deleting residues beyond the first 14, may have resulted from an interference with the important ␥-␣␤ interactions. For example, the arginines at positions 254 and 256 in MF 1 are surrounded by a ring of 9 charged residues located on six loop segments of the ␣ and ␤ subunits marked "B" on one of the ␤ subunits in Fig. 5. In the MF 1 structure, FIG. 5. Cross-section through part of the beef heart mitochondrial F 1 structure indicating sites of interaction between the ␣, ␤, and ␥ subunits. The ␥ subunit contacts a hydrophobic sleeve formed by the structures marked A on all six of the ␣ and ␤ subunits. A second site of contact involves a salt link between the ␥ subunit and the structure marked B on an adjacent ␤ subunit (2).
Arg-254 and Gln-255 form hydrogen bonds with adjacent residues in the ␤ subunit loop to form one of the few sites of direct contact between ␥ and the ␣ and ␤ subunits (2). Assuming that CF 1 has the same arrangement in this region of ␥, deleting residues in the near vicinity of the site of contact would be expected to significantly compromise catalytic function as was observed.
Interestingly, an earlier study of the E. coli enzyme (31) showed that mutations near the C terminus of the ␥ subunit were able to restore catalytic function to functionally impaired enzymes which contained mutations near the N terminus. The original interpretation of these results was that the two mutations are in close proximity to each other. If this is true it would place the C terminus of the E. coli ␥ in a location very close to that of the chloroplast ␥ as determined by the fluorescence mapping experiments, assuming of course that the position of the N terminus of the E. coli ␥ is similar to that of the mitochondrial enzyme.
In conclusion, the results of this study have demonstrated that the extreme C-terminal 20 residues of the ␥ subunit of CF 1 are not essential for normal cooperative catalytic turnover by the isolated enzyme. The results eliminate the possibility of a catalytic mechanism that is universal to all F 1 enzymes in which the tip of the C terminus of the ␥ subunit must act as a spindle for rotational catalysis. The lack of functional importance of this part of the ␥ subunit for rapid turnover by CF 1 is consistent with earlier work indicating that the conformation of the C-terminal portion of the ␥ subunit of the chloroplast ATP synthase may differ from that of the mitochondrial enzyme. The results further suggest that the contacts between the ␣, ␤, and ␥ subunits of the enzyme, which are essential for rotational catalysis must be provided by regions of the ␥ subunit other than the extreme C terminus.