Inhibition of ATP Hydrolysis by Thermoalkaliphilic F₁F_o-ATP Synthase Is Controlled by the C Terminus of the ε Subunit

Escherichia coli DH10B (11) was used for all cloning experiments, and E. coli mutant DK8 (28), lacking the ATP synthase genes of the unc operon, was used to overproduce the TA2F₁ complexes from the ATP synthase of Bacillus sp. strain TA2.A1.

In a previous study, the genes encoding the F₁ subunits were cloned into the expression vector pTrc99A (Amersham Biosciences) under the control of the strong tryptophan promoter to create plasmid pTrcF₁ (50). Plasmid pTrcF₁ε^ΔC, lacking the coding sequence for the C-terminal domain (residues 90 to 135) of ε, was constructed by digesting pTA2F₁ (50) with AccI and ClaI and cloning this 4.7-kb fragment into pTrc99A digested with AccI. To construct the ε mutant plasmids, pTrcε^4A and pTrcε^6A, two 0.23-kb ClaI-HindIII fragments containing codons for four and six alanine substitutions in the C-terminal domain of ε, respectively, were created by the overlap extension PCR method (16). These fragments were then cloned simultaneously with a 4.7-kb EcoRI-ClaI fragment, containing the remainder of the F₁ genes, into pTrc99A digested with EcoRI and HindIII. To create the first ClaI-HindIII fragment, two pairs of primers (viz., atpCfwd2 with atpCmutantP1 and atpCmutantP2 with pUC8rev) (Table 1) and plasmid pA3 (27) as the template were used to generate mutated DNA fragments. These overlapping fragments were then used for overlap extension PCR with the external primers atpCfwd2 and pUC8rev, and the product obtained was digested with ClaI and HindIII. To create the second ClaI-HindIII fragment containing codons for a total of six alanine substitutions, the same procedure as above was used except that plasmid pTrcε^4A served as a template and the primer pairs atpCfwd2 with atpCmutantP3 and atpCmutantP4 with pTrc99Arev (Table 1) were utilized. The PCRs were carried out with the Phusion high-fidelity PCR kit (Finnzymes) according to the manufacturer's instructions. The region amplified by PCR in the ε mutant plasmids was confirmed by DNA sequencing.

The plasmids were expressed in E. coli DK8 by growth in 2× YT medium (45) containing 2 g/liter glucose and 100 μg/ml ampicillin at 37°C with shaking to an optical density at 600 nm of 0.35 to 0.4. An inducer, isopropyl β-d-thiogalactopyranoside (IPTG) was added (1 mM final concentration), and incubation continued for 5 h. Cells were harvested, washed with precooled 100 mM MOPS (morpholinepropanesulfonic acid) buffer (pH 8.0), and resuspended in MOPS-Mg buffer (100 mM MOPS, pH 8.0, 2 mM MgCl₂, 10% glycerol). Phenylmethylsulfonyl fluoride (PMSF) was added to 0.1 mM, and the cells were disrupted by two passages through a French pressure cell at 20,000 lb/in². Pancreatic DNase was added to 0.1 mg/ml, and the mixture was kept on ice for 1 h or until viscosity decreased. Lysate was cleared of debris by centrifugation at 8,000 × g for 10 min, and the membranes were pelleted from the supernatant at 180,000 × g for 1 h at 4°C. The supernatant was retained and incubated at 60°C for 30 min to denature E. coli proteins, and these were removed by centrifugation at 180,000 × g for 20 min at 4°C. The heat-treated supernatant was then brought to 5 mM MgCl₂ and 0.5 M NaCl before application to a 50-ml column of chelating Sepharose Fast Flow (Amersham Biosciences). Before applying the TA2F₁ samples, the chelating Sepharose column was loaded with 80 ml of 0.3 M CuSO₄, washed with five column volumes of distilled water, and equilibrated with five column volumes of 20 mM Tris-HCl (pH 8.0) containing 0.5 M NaCl, 5 mM MgCl₂, and 10% glycerol. After the TA2F₁ samples (120 ml) were loaded, the column was washed with three column volumes of the same buffer. The bound TA2F₁ complexes were eluted with one column volume of the same buffer containing 100 mM imidazole. The eluted TA2F₁ complexes were dialyzed (molecular weight cutoff [MWCO], 6,000) against 20 mM Tris-HCl (pH 8.0) containing 1 mM MgCl₂ and 10% glycerol to remove excess salt. The samples (approximately 200 mg protein) were applied to a 60-ml POROS HQ 50 column (Applied Biosystems), which was equilibrated with the dialysis buffer. Bound proteins were eluted by 12 column volumes of a 0 to 600 mM linear gradient of NaCl in the same buffer. The ATPase-containing fractions were pooled and concentrated to 30 mg/ml using Amicon Ultra centrifugal filter devices (MWCO, 100,000). The ATPase samples (1 ml) were then applied in aliquots of 0.2 ml to a GPC Superose 6 column (Amersham Biosciences) and fractionated with 20 mM Tris-HCl (pH 8.0), 1 mM EDTA, and 100 mM NaCl buffer. The peak between 370 to 380 kDa was pooled and immediately used for ATPase assays and reconstitution studies.

Protein concentrations were determined by using the bicinchoninic acid protein assay kit from Sigma, with bovine serum albumin as the standard. Purified TA2F₁ complexes were analyzed by 13% polyacrylamide gel electrophoresis (PAGE) in the presence of 0.1% sodium dodecyl sulfate (SDS) (30), and protein bands were visualized by staining with Coomassie brilliant blue.

All protease treatments of TA2F₁ complexes and TA2F₁F_o were performed in an assay mixture containing 50 mM Tricine [N-tris(hydroxymethyl)methylglycine]-NaOH, pH 8.0, 10% sucrose, 10% glycerol, 2 mM EDTA, and 2.5 mM ATP. TA2F₁ and TA2F₁F_o preparations (1 mg/ml) were incubated at 30°C with either trypsin (DPCC [diphenyl carbamyl chloride] treated) or α-chymotrypsin (TLCK [Nα-p-tosyl-l-lysine chloromethyl ketone] treated) at a ratio of 1:250 to 1:1,000 (wt/wt). At various time intervals 5-μl aliquots were withdrawn and immediately assayed for ATPase activity in the presence and absence of LDAO. At intervals of 60 to 120 min, proteolysis was stopped by a 15-fold dilution of the sample into buffer (50 mM MOPS [pH 7.5] and 2 mM MgCl₂) containing PMSF (100-fold excess by weight over proteases) and stored at −70°C. The subunit composition was analyzed by SDS-PAGE.

Inverted membrane vesicles of strain TA2.A1 were prepared as previously described (7). To strip F₁ from wild-type membranes, approximately 15 mg of washed membranes (500 μl) was resuspended in 100 ml of 5 mM Tris-sulfate buffer (pH 8.0) containing 0.5 mM EDTA and the mixture was incubated for 30 min at room temperature with gentle agitation. The homogenate was subsequently centrifuged at 180,000 × g for 45 min. The pellet was resuspended in 100 ml of 0.5 mM EDTA (pH 8.0) with stirring for 1 h at room temperature, followed by high-speed centrifugation (180,000 × g for 45 min) to harvest the membranes. The pellet was resuspended in 0.5 ml of 50 mM potassium phosphate buffer (pH 7.8) containing 2 mM MgCl₂ and 0.1 mM PMSF. The ATPase activity (LDAO stimulated) of the F₁-stripped membranes indicated that the protocol was effective at removing >95% of the F₁ moiety.

Purified TA2F₁ complexes (1.3 mg) were incubated with freshly prepared F₁-stripped membranes (1.5 mg) in a total volume of 300 μl, and the volume was increased to 5 ml with 50 mM MOPS (pH 7.5) containing 100 mM KCl and 2 mM MgCl₂ (buffer A). The mixture was incubated on ice for 60 min and then centrifuged at 180,000 × g for 45 min. The membranes were washed with buffer A (10 ml) to remove unbound TA2F₁ complexes. This step was repeated three or four times or until the supernatant contained no ATPase activity.

ATP hydrolysis activity was measured at 45°C by using an ATP-regenerating assay. The assay mixture contained 100 mM MOPS (pH 7.5), 3 mM phosphoenolpyruvate, 2 mM MgCl₂, 0.25 mM NADH, 0.57 U/ml pyruvate kinase, 3.2 U/ml lactate dehydrogenase, and 50 μM or 2 mM ATP, as specified in the figure legends. The reaction was initiated by the addition of the sample into 1 ml of assay mixture, and the rate of NADH oxidation was monitored continuously at 340 nm. Nine micrograms of TA2F₁ subunit complexes was used for measurements. The activity that hydrolyzed 1 μmol of ATP per min is defined as 1 unit.

ATP-dependent proton translocation was determined at 45°C by the quenching of 9-amino-6-chloro-2-methoxyacridine (ACMA). The 1.5-ml reaction mixture contained 5 mM potassium phosphate buffer (pH 7.5), 5 mM MgCl₂, 1 μM ACMA, and 150 to 200 μg of inverted membranes or reconstituted TA2F_oF₁ complexes. After the fluorescence signal had stabilized (100%), the reaction was initiated by the addition of neutralized ATP (1.7 mM). Fluorescence was measured with an excitation wavelength of 410 nm and an emission wavelength of 480 nm (slit width, 10 nm). At the times indicated in Fig. 7, 1.5 μM of carbonyl cyanide m-chlorophenylhydrazone (CCCP) was added.

To study the role of the ε subunit in the ATP synthase of Bacillus sp. strain TA2.A1, we developed a heterologous expression system in E. coli to produce recombinant TA2F₁ complexes with various mutations in atpC. The genes encoding TA2F₁ were overexpressed in DK8 (Δatp) and purified using a three-step procedure. The purity of the recombinant TA2F₁ complexes was assessed by SDS-PAGE and compared to the subunit composition of purified TA2F₁F_o from native membranes (Fig. 2). The five protein bands from the recombinant TA2F₁(ε^WT) complex, corresponding to the α, β, γ, δ, and ε subunits, coincided with those from the native TA2F₁F_o-ATP synthase (Fig. 2, compare lanes 1 and 4). The TA2F₁(ε^ΔC) complex, containing the ε subunit from which the C-terminal domain had been removed (Fig. 2, lane 3), and the TA2F₁(ε^6A) complex, where the basic residues in ε (i.e., Arg¹¹⁶, His¹¹⁷, Lys¹¹⁸, Arg¹¹⁹, Arg¹²³, and Arg¹²⁷) were substituted with nonpolar alanines (Fig. 2, lane 2), exhibited the expected changes at a protein level. The ε subunit of TA2F₁(ε^6A) migrated at a slightly lower molecular weight than TA2F₁(ε^WT).

To determine whether the recombinant TA2F₁ complexes were functional, ATP hydrolysis assays were carried out. Like the native F₁F_o-ATP synthase, the TA2F₁(ε^WT) was blocked in ATP hydrolysis activity with a specific activity of 0.9 units mg protein⁻¹. ATPase activity was activated 30-fold by the addition of LDAO (Fig. 3A), and maximal ATP hydrolysis activity was observed at 0.04% LDAO with a specific activity of 28.5 units mg protein⁻¹. Purified TA2F₁F_o-ATP synthase was also stimulated by LDAO (20-fold, 19.4 units mg protein⁻¹), but 10-fold more LDAO was required to reach near-saturation ATPase activity (Fig. 3B).

To determine the effect of ε on ATP hydrolysis activity, TA2F₁(ε^WT) was compared to TA2F₁(ε^ΔC). At low concentrations of ATP (50 μM), TA2F₁(ε^WT) exhibited very low levels of ATP hydrolysis with a specific activity of 0.12 units mg protein⁻¹ (Fig. 4A). In contrast, TA2F₁(ε^ΔC) exhibited levels of ATPase activity that were approximately sevenfold higher (0.85 units mg protein⁻¹) (Fig. 4A). At higher concentrations of ATP (2 mM), TA2F₁(ε^WT) had a rate of ATP hydrolysis activity of 1.2 units mg protein⁻¹ and the ATP hydrolysis activity of TA2F₁(ε^ΔC) was 4.3 units mg protein⁻¹ (Fig. 4B). To identify the residues in the C-terminal domain of the ε subunit important in this process, we changed the cluster of basic residues at positions Arg¹¹⁶, His¹¹⁷, Lys¹¹⁸, and Arg¹¹⁹, as well as Arg¹²³ and Arg¹²⁷, to alanine residues (Fig. 1). In a mutant in which four of the basic residues (viz., Arg¹¹⁶, His¹¹⁷, Lys¹¹⁸, and Arg¹¹⁹) were changed to alanines [i.e., TA2F₁(ε^4A)], ATP hydrolysis activity was still blocked at 50 μM ATP, but at 2 mM the ATPase activity was similar to TA2F₁(ε^ΔC) (Fig. 4C and D). The TA2F₁(ε^6A) mutant in which all six basic residues were changed to alanines exhibited levels of ATP hydrolysis activity that were equal to that of TA2F1(ε^ΔC) at both low and high concentrations of ATP (Fig. 4C and D). The kinetics of ATP hydrolysis for TA2F₁(ε^WT), TA2F₁(ε^ΔC), and TA2F₁(ε^6A) were determined (Fig. 4E). With increasing concentrations of ATP, TA2F₁(ε^WT) hydrolyzed ATP at a rate that was almost proportional to the external ATP concentration, and the kinetic analyses using the Lineweaver-Burk equation indicated that the apparent K_m and V _max were 0.52 mM and 1.27 units mg protein⁻¹, respectively. In contrast, TA2F₁(ε^ΔC) and TA2F₁(ε^6A) showed high rates of ATP hydrolysis at low ATP concentrations (Fig. 4E). For TA2F₁(ε^ΔC), the apparent K_m and V _max were 0.20 mM and 3.03 units mg protein⁻¹, respectively. Similar values were obtained for TA2F₁(ε^6A) (i.e., the K_m and V _max were 0.17 mM and 3.11 units mg protein⁻¹, respectively). Based on the observation that TA2F₁(ε^6A) behaved like TA2F₁(ε^ΔC), we decided to use TA2F₁(ε^6A) for further characterization studies. When the rates of ATP hydrolysis for TA2F₁(ε^WT) and TA2F₁(ε^6A) were measured at increasing concentrations of ATP in the presence of 0.05% LDAO, typical saturation kinetics were observed (Fig. 4F). The V _max of TA2F₁(ε^6A) was significantly higher (twofold) than that of TA2F₁(ε^WT) (Fig. 4F), but the affinity constants were in similar ranges (i.e., between 0.2 to 0.3 mM ATP).

The above results demonstrate that the C-terminal arm of the ε subunit inhibits ATP hydrolysis activity of TA2F₁, and the most likely mechanism for this inhibition is a strong electrostatic interaction between the cluster of basic residues in the second α-helix of ε and another subunit in the F₁ complex. The other subunit in this interaction remains to be identified, but, on the basis of previous studies, the DELSEED region of the β subunit seems a likely possibility (12). Studies with the chloroplast ATP synthase have shown that the C-terminal domain of the ε subunit is a regulator of ATP hydrolysis activity both in CF₁ and CF₁F_o, and most of this inhibitory activity can be attributed to the last 45 C-terminal amino acids (38, 39). While the ε subunit is a regulator of ATP hydrolysis activity, ATP synthesis experiments revealed that the C-terminal domain of the ε subunit is not required for photophosphorylation (39).

Lowry and Frasch (33) have highlighted that other subunits of the soluble E. coli F₁-ATP synthase (ECF₁) play a role in regulating ATP hydrolysis activity, namely, the salt bridges between the residues βD372 and γK9. We noted some variation in the apparent K_m values for ATP of the TA2F₁(ε^WT) and TA2F₁(ε^6A) complexes, suggesting a potential role for ε in modulating the K_m of the β subunits for ATP. ε has been proposed to act as a sensor of intracellular ATP in vivo for some bacterial species (26). This could be mediated through an effect whereby ε converts the enzyme from a form with a high affinity for binding ATP to one where the affinity is greatly reduced. Under physiological conditions within the cell, ATP hydrolysis could simply be impaired by the lowered affinity of the enzyme for ATP. ATP hydrolysis would therefore be observed only if the intracellular ATP concentration increased to very high levels. In strain TA2.A1, the intracellular ATP content of exponential-phase cells is in the range 0.5 to 3.5 mM and this drops to below 0.1 mM in stationary-phase cells (41). A role for ε in modulating the affinity of the β subunits for Mg-nucleotide complexes has been reported for ECF₁ (56). In ECF₁, ε alters the binding affinity of high-affinity catalytic nucleotide binding site 1 for MgADP and MgATP. The affinity for MgADP and MgATP is 10-fold higher in an ε-replete ECF₁ compared to a ε-depleted form of the enzyme (56). The authors suggest that a potential explanation for this difference could be the reduced off rate for the Mg-nucleotide complex in the presence of ε and hypothesize that the ε subunit decelerates a conformational change step that is necessary to release ligands from the high-affinity binding site.

The above results indicate that alanine substitutions in the C-terminal arm of the ε subunit prevent inhibition of ATPase activity in TA2F₁. To determine if these alanine substitutions induced a change in the conformation of the C-terminal arm, we studied the effect of protease treatment on TA2F₁, TA2F₁(ε^6A), and TA2F₁F_o (Fig. 5). When TA2F₁F_o was treated either with chymotrypsin or trypsin, low ATP hydrolysis activities were observed (i.e., 0.32 and 0.25 units mg protein⁻¹, respectively), indicating that the enzyme was still blocked in ATP hydrolysis activity. SDS-PAGE revealed some proteolytic digestion products in both the chymotrypsin- and trypsin-treated TA2F₁F_o enzyme when compared to the untreated control; however, the ε subunit was unaffected (Fig. 5A). Extensive proteolytic digestion was noted in TA2F₁(ε^WT), but no proteolysis of the ε subunit was discernible (Fig. 5B, lanes 2 and 3), and the ATPase activity of the protease-digested TA2F₁(ε^WT) remained low (i.e., 0.58 to 0.9 units mg protein⁻¹). In contrast, the ε subunit in TA2F₁(ε^6A) was completely degraded by trypsin (Fig. 5C, lane 3) but not by chymotrypsin (Fig. 5C, lane 2). In neither case did protease treatment further activate ATP hydrolysis in TA2F₁(ε^6A) (2.8 to 2.92 units mg protein⁻¹) compared with the untreated control (3.8 units mg protein⁻¹). These data imply that the conformation of the ε subunit in both TA2F₁ and TA2F₁F_o is protected by proteolytic enzymes compared to TA2F₁(ε^6A), where ε is degraded. Since no further activation of ATP hydrolysis in TA2F₁(ε^6A) was observed, despite extensive proteolytic digestion, the latent ATPase activity of the enzyme seems to be controlled by the ε subunit, at least in terms of protein-protein interactions. The conformation of ε in strain TA2F₁ would appear to be different to that in ECF₁ since, under similar experimental conditions, trypsin was shown to activate ATPase activity by the removal of ε from the ECF₁ complex (32, 49). High-resolution structural data have indicated that the ε subunit can exist in two conformations with different arrangements of the two α-helices. In the “up-state form” the two α-helices extend into the catalytic head of the F₁. In the “down-state” or retracted form the helices lie in a plane parallel to the oligomeric c ring (10, 43, 52). A third conformation, where the α-helices are fully extended, has been suggested by cross-linking studies (51). Biochemical data have accumulated that indicate that the ε subunit adopts the extended form in the absence of added nucleotide or in the presence of ADP, in which ATPase activity is inhibited, and that ATP counteracts ADP by favoring the retracted form through the binding of ATP to the catalytic β subunit, which is a noninhibitory form (19, 51). The blockage in ATP hydrolysis activity of the TA2F₁F_o-ATP synthase, even under conditions of increasing ATP, suggests that under physiological conditions the enzyme never works in the ATPase direction, and thus it is tempting to speculate that the ε subunit α-helices are fully extended, resulting in ATP hydrolysis inhibition and protection from trypsin.

Johnson and McCarty (23) have provided evidence that the chloroplast ATP synthase undergoes a conformational shift during its activation by the Δμ_H⁺ in light conditions, which specifically induces a change in the conformation of the C-terminal domain of the ε subunit from a protected to a solvent-exposed state. Studies on the orientation of the ε subunit within CF₁ demonstrate that it shares a common structure and positioning within the ATP synthase to that of the E. coli ε subunit (4, 22).

To determine what effect the alanine substitutions in ε had on the entire TA2F₁F_o-ATP synthase, purified TA2F₁(ε^WT) and TA2F₁(ε^6A) were reconstituted with F₁-stripped native membranes of Bacillus sp. strain TA2.A1. Native membranes exhibited approximately 4.4 units mg protein⁻¹ of ATP hydrolysis activity in the presence of 0.5% LDAO. After F₁ was stripped off the membranes, only 13% of the original LDAO-stimulated ATPase activity could be detected, indicating that the procedure removed up to 87% of the F₁ moiety (data not shown).

After reconstitution of TA2F₁(ε^WT) to F_o membranes, proteins were visualized by SDS-PAGE (Fig. 6). Even when stripped membranes were loaded in excess onto the SDS-PAGE gel, no F₁ subunits were visible (Fig. 6, lane 3). These are clearly visible in native membranes that are overloaded on the same gel (Fig. 6, lane 5). When TA2F₁(ε^WT) was reconstituted with F₁-stripped F_o membranes [i.e., TA2F_oF₁(ε^WT)], all F₁ subunits are clearly visible, indicating that reconstitution of F₁ to F_o had indeed taken place (Fig. 6, lane 2).

When proton transport was measured by the quenching of ACMA fluorescence in native membranes, only very low levels of proton pumping (fluorescence quenching) could be detected (Fig. 7A), which is consistent with the latent ATPase activity of this enzyme in strain TA2.A1. The residual ATP-dependent proton pumping in native membranes was dissipated by the addition of the protonophore CCCP (Fig. 7A). Importantly, no proton pumping could be detected in the stripped membranes (Fig. 7B). When proton pumping measurements were performed with reconstituted TA2F_oF₁(ε^WT), ACMA fluorescence was quenched to some extent by the addition of ATP, and this was dissipated by CCCP addition (Fig. 7C). Identical experiments were performed with TA2F_oF₁(ε^6A) and TA2F_oF₁(ε^ΔC) membranes (Fig. 7D and E). Under these conditions, proton pumping was faster than with TA2F_oF₁(ε^WT) and the quenching of ACMA fluorescence was again reversed by CCCP. While TA2F_oF₁(ε^ΔC) membranes showed a greater level of proton pumping when compared to TA2F_oF₁(ε^WT), the level was lower than that of TA2F_oF₁(ε^6A) despite both mutant forms having similar ATP hydrolysis activities. ATP-dependent proton pumping of TA2F_oF₁(ε^WT), TA2F_oF₁(ε^6A), and TA2F_oF₁(ε^ΔC) was sensitive to DCCD (N,N′-dicyclohexylcarbodiimide) (data not shown). ATPase activity was determined in TA2F_oF₁(ε^WT), TA2F_oF₁(ε^6A), and TA2F_oF₁(ε^ΔC), and the ATP hydrolysis rates at high ATP concentration (i.e., 2 mM) (Fig. 7F) agree with those observed in the corresponding TA2F₁ complexes (Fig. 4D).

Cipriano et al. (5) have put forward a model that proposes that the ε subunit plays an important role in keeping the ATP synthase efficient by preventing uncoupled ATP hydrolysis. This is further supported by single-molecule experiments with F₁-ATPase from the thermophilic Bacillus sp. strain PS3 which demonstrate unequivocally that the ε subunit is essential for coupling catalysis with mechanical rotation in the ATP synthesis direction (44). Dunn and coworkers showed that if the E. coli ATP synthase is blocked in rotation by the fusion of a 20-kDa protein to the N terminus of ε, ATPase activity is not eliminated (6). Moreover, the residual ATPase activity of this enzyme is doubled if the C-terminal domain is truncated. In the wild-type enzyme, truncation of the C-terminal domain resulted in an increased rate of ATP hydrolysis activity but a decreased rate of proton pumping, indicating uncoupled activity (6). Rondelez et al. (44) report that ε is dispensable for ATP hydrolysis activity using α₃β₃γ subcomplexes where the mechanochemical coupling efficiency was three ATPs hydrolyzed per turn either in the presence or absence of ε. However, these studies did not measure proton-pumping activity, and therefore it is not known whether ATP hydrolysis was coupled to proton translocation. Our data suggest that an activation of ATPase activity in TA2F₁(ε^6A) does correlate with increased proton pumping. However, in TA2F₁(ε^ΔC), we observed a 50% decrease in proton pumping, suggesting that, in this form, ATP hydrolysis was partially uncoupled from proton-pumping activity.

We have previously demonstrated that the F₁F_o-ATP synthase of the thermoalkaliphile Bacillus sp. strain TA2.A1 is blocked in ATP hydrolysis activity, and this is intrinsic to the F₁ moiety (7). The results of the present study demonstrate that the ε subunit plays an important role in blocking ATP hydrolysis activity of both TA2F₁ and TA2F₁F_o. This inhibition of ATPase activity was attributed to a number of basic residues located in the C-terminal arm of the ε subunit (viz., Arg¹¹⁶, His¹¹⁷, Lys¹¹⁸, Arg¹¹⁹, Arg¹²³, and Arg¹²⁷). As shown in Fig. 1, various members of the genus Bacillus (alkaliphilic and nonalkaliphilic) show a higher proportion of basic residues in the C-terminal arm of the ε subunit when compared to E. coli. We hypothesize that this serves an important regulatory function when cells become either limited for oxygen (decrease in Δμ_H⁺ ) or grow at high pH (low Δμ_H⁺ due to an inverted pH gradient). Notably, both conditions are prevalent in the lifestyle of alkaliphilic bacilli. Under these conditions, there is a need to overcome the thermodynamic problem of uncontrolled ATPase activity, and kinetically these bacteria have solved this problem by blocking ATP hydrolysis activity (7, 14, 15, 17). We have recently performed side-by-side experiments with inverted membrane vesicles of all Bacillus species listed in Fig. 1 and report that they all show various degrees of latent ATPase activity, but importantly all were able to synthesize ATP at high rates (B. Seale, S. Keis, and G. M. Cook, unpublished data). For alkaliphilic bacteria, the blockage in the ATP hydrolysis direction, but not in the ATP synthesis (proton import) direction, is further magnified by the requirement for intracellular pH homeostasis, both in the retention and acquisition of protons to prevent intracellular alkalinization. The role of the ATP synthase in alkali adaptation has been reported for E. coli, where the enzyme has been shown to be up-regulated at pH 8.7 as a mechanism of proton import (34). Conversely, systems involved in proton export were down-regulated in E. coli at high pH (e.g., cytochrome o and NADH dehydrogenases I and II).

The precise molecular mechanism whereby the enzyme can block rotation in one direction but rotates in the ATP synthesis direction remains unknown but clearly involves the ε subunit and other protein partners. Notwithstanding this important role, the ε subunit alone is not sufficient for ATP synthase function at high pH. Wang and coworkers have identified the presence of amino acid motifs that are unique to the c and a subunits of alkaliphilic bacilli (55). Studies by this group with the F₁F_o-ATP synthase from the facultative alkaliphile B. pseudofirmus OF4 have demonstrated that these alkaliphilic motifs are critical for growth on nonfermentable carbon sources and ATP synthesis at pH 10.5 but not pH 7.5 (55).

If the ATP synthase of strain TA2.A1 is specifically adapted to work only in the ATP synthesis direction, then one could speculate that the C-terminal arm of ε is obligatory for growth under oxidative phosphorylation conditions at high pH. To our knowledge, the role of ε on the energetics of growth of E. coli on nonfermentable carbon sources has not been assessed, and therefore the physiological role of ε in vivo remains unknown. Future research will focus on developing a genetic system to make in vivo mutations in the ε subunit of strain TA2.A1 and assess the effect on growth at high pH.