A Specific Adaptation in the a Subunit of Thermoalkaliphilic F1FO-ATP Synthase Enables ATP Synthesis at High pH but Not at Neutral pH Values*

Analysis of the atp operon from the thermoalkaliphilic Bacillus sp. TA2.A1 and comparison with other atp operons from alkaliphilic bacteria reveals the presence of a conserved lysine residue at position 180 (Bacillus sp. TA2.A1 numbering) within the a subunit of these F1Fo-ATP synthases. We hypothesize that the basic nature of this residue is ideally suited to capture protons from the bulk phase at high pH. To test this hypothesis, a heterologous expression system for the ATP synthase from Bacillus sp. TA2.A1 (TA2F1Fo) was developed in Escherichia coli DK8 (Δatp). Amino acid substitutions were made in the a subunit of TA2F1Fo at position 180. Lysine (aK180) was substituted for the basic residues histidine (aK180H) or arginine (aK180R), and the uncharged residue glycine (aK180G). ATP synthesis experiments were performed in ADP plus Pi-loaded right-side-out membrane vesicles energized by ascorbate-phenazine methosulfate. When these enzyme complexes were examined for their ability to perform ATP synthesis over the pH range from 7.0 to 10.0, TA2F1Fo and aK180R showed a similar pH profile having optimum ATP synthesis rates at pH 9.0–9.5 with no measurable ATP synthesis at pH 7.5. Conversely, aK180H and aK180G showed maximal ATP synthesis at pH values 8.0 and 7.5, respectively. ATP synthesis under these conditions for all enzyme forms was sensitive to DCCD. These data strongly imply that amino acid residue Lys180 is a specific adaptation within the a subunit of TA2F1Fo to facilitate proton capture at high pH. At pH values near the pKa of Lys180, the trapped protons readily dissociate to reach the subunit c binding sites, but this dissociation is impeded at neutral pH values causing either a blocking of the proposed H+ channel and/or mechanism of proton translocation, and hence ATP synthesis is inhibited.

For nearly all aerobic life on earth, the F 1 F o -ATP synthases are the major enzymes responsible for providing ATP to drive endergonic reactions of the cell. These two domain membranebound enzymes are found in mitochondria, chloroplasts, and bacteria, coupling protons or Na ϩ ions to the synthesis of ATP (1,2). The intracellular water-soluble F 1 domain contains the catalytic sites of the enzyme, while the bulk of the hydrophobic F o domain is embedded in the cytoplasmic membrane and contains the functional center for the capture and translocation of protons. The bacterial F 1 domain has a stoichiometry of ␣ 3 ␤ 3 ␥␦⑀ in which the catalytic nucleotide-binding sites are formed by each of the ␤ subunits and the non-catalytic nucleotide-binding sites are located in the ␣ subunits (3)(4)(5). The F o domain consists of one a subunit, two b subunits making the stator, and a ring of 10 -15 c subunits depending on the species (6 -12). The coupling of ion translocation to the rotational mechanism of the F o c-ring forces open, loose, or tight conformational changes within the ␤ subunits of the F 1 domain, driving the synthesis of ATP from ADP and inorganic phosphate (13). Conversely, the F 1 domain can hydrolyze ATP causing the reverse rotation of the c-ring and pumping protons into the bulk phase, a mechanism that is absent in aerobic alkaliphilic organisms (14 -17).
With F 1 F o -ATP synthases that operate in a neutral to acidic pH environment, proton coupling is an energetically favorable process as the proton concentration outside the cell exceeds that in the cytoplasm. For example, at pH 7.1 Escherichia coli generates an electrochemical gradient of protons (⌬H ϩ ) 4 that is in the range Ϫ200 to Ϫ210 mV, with most of the ⌬H ϩ being in the form of an electrical potential (⌬ Ͼ Ϫ150 mV) but with a positive contribution of the ⌬pH (18). In bacteria that inhabit alkaline pH environments, the magnitude of the ⌬H ϩ is markedly lower (in the range Ϫ50 to Ϫ100 mV) because of the inverted pH gradient (acidic in , alkaline out ) (19 -21). Despite this apparent thermodynamic problem, the aerobic alkaliphilic bacteria studied to date employ a proton-coupled ATP synthase (14 -16). The apparent thermodynamic problem is based on a mechanism with a delocalized ⌬H ϩ in the bulk phase but could be overcome by a more localized proton pathway between the respiratory chain and the ATP synthase. Support for such a model has been provided by Guffanti et al. (22,23) who demonstrated in Bacillus firmus RAB that a respirationderived ⌬H ϩ could drive ATP synthesis at pH 9.0, but a valinomycin-mediated potassium-diffusion potential could not.
Apart from the thermodynamic problem of synthesizing ATP at high environmental pH, the alkaliphilic ATP synthase must also solve the problem of proton capture from an alkaline environment and subsequent translocation to the binding sites on the c-ring. Sequence analysis reveals that the atp operon gene arrangement and deduced primary structure of the ATP synthase from alkaliphilic bacteria and other eubacteria are similar (24 -26). As reported by Ivey and Krulwich (24,25), greatest variation from amino acid consensus is observed in the a and c subunits that are crucial for proton translocation. The E. coli a subunit is a hydrophobic protein, spanning the membrane five to six times (27,28). It is proposed to contain aqueous channels for protons to gain access to the binding sites in subunit c (28 -30). In the a subunits of alkaliphilic ATP synthases ( Fig. 1), an invariant lysine residue at position 180 (Lys 180 ) (Bacillus pseudofirmus OF4 numbering) is located in transmembrane helix 4 (aTMH4) (24,25). Moreover, a glycine is also found in alkaliphiles at position 212 (B. pseudofirmus OF4 numbering) in aTMH5 opposite Lys 180 . Wang et al. (31) have proposed that these residues form a gated channel residing within the proton uptake pathway of the a subunit through which protons pass onto the neighboring c subunit. In studying the role of these residues in growth of the facultative alkaliphile B. pseudofirmus OF4 on fermentable and non-fermentable carbon sources at pH 7.5 and pH 10.5 (31), the authors demonstrate that when Lys 180 is substituted for glycine, B. pseudofirmus OF4 is no longer able to grow at high pH (i.e. 10.5) on non-fermentable malate. However, the cells still grow on a fermentable carbon source at both pH 7.5 and 10.5 (31). This also correlated with low rates of ATP synthesis performed by rightside-out (RSO) membrane vesicles of the aK180G mutant at pH 10.5 when energized by ascorbate-phenazine methosulfate. Interestingly, when ATP synthesis was energized via a valinomycin-mediated potassium diffusion potential, the rates of ATP synthesis were ϳ2-fold higher in the aK180G mutant at pH 8.3, compared with the wild type, suggesting that a lysine was less effective at lower pH values (31).
We have recently initiated studies on the F 1 F o -ATP synthase from the thermoalkaliphile Bacillus sp. TA2.A1 (14,19,26,32,33). Like other alkaliphilic ATP synthases, the a subunit of strain TA2.A1 also harbors a conserved lysine residue at position 180 (Bacillus sp. TA2.A1 numbering) ( Fig. 1) and we hypothesize that its side chain amino group acts as a base to capture protons at high environmental pH. In this communication, we report on the development of a heterologous expression system to overproduce and purify recombinant F 1 F o -ATP synthase from Bacillus sp. TA2.A1 (TA2F 1 F o ) to address this hypothesis. Various amino acid substitutions in the a subunit were introduced to replace the lysine at position 180 (viz. glycine, histidine, and arginine) and the effect of external pH on ATP synthesis was studied in ADP plus P i -loaded right-sideout membrane vesicles of E. coli DK8 expressing TA2F 1 F o .  (39) supplemented with 10 g/ml thiamine, 50 g/ml each of alanine, valine, asparagine, and isoleucine, 100 g/ml ampicillin, and 35 mM succinate as a non-fermentable carbon and energy source over a pH range of 7-9.5. Control plates contained the same amount of glucose in lieu of succinate. Plates were incubated for 2-5 days at 37°C before scoring for growth.

EXPERIMENTAL PROCEDURES
Construction of an Expression Plasmid for TA2F 1 F o -In a previous study (26), the Bacillus sp. TA2.A1 genes encoding for the subunits of the F 1 F o -ATP synthase were cloned in pUC8 as  (64), and E. coli (65). Residue numbering for Bacillus sp. TA2.A1 and E. coli are indicated above and below the alignment, respectively. three overlapping fragments in plasmids pF10 (containing atpIBEFH), pB9 (containing atpHAGDЈ), and pA3 (containing atpGDC). To assemble the atp genes atpIBEFHAG and part of atpD, a 2.3-kb SalI-BamHI fragment from pF10 and a 3.4-kb BamHI-SacI fragment from pB9 were simultaneously cloned into pUC19 (36) digested with SalI and SacI to create plasmid pContigW1. The remainder of the atp genes (viz. atpD and atpC) were inserted as a 1.5-kb SacI-EcoRI fragment from pA3 into pContigW1 digested with the same restriction enzymes, to create plasmid pTA2ATP. Subsequently, the entire atp operon was cloned as a 7.2-kb SalI fragment from pTA2ATP into the expression vector pTrc99A (Amersham Biosciences) under the control of the strong tryptophan promoter to create plasmid pTrcATP9. To facilitate purification of TA2F 1 F o , a 960-bp SspI-SacI fragment, coding for a hexa-histidine tag at the N terminus of the ␤ subunit, was cloned simultaneously with a 1.7-kb BglII-SspI fragment from pB9, and a 1.3-kb SacI-ClaI fragment from pA3, into pTrcATP9 digested with BglII and ClaI. This expression plasmid was designated pATPHis5. The fragment containing the hexa-histidine tag was created by the overlap extension PCR method (40) as follows. Two pairs of primers, A3fwd2 (5Ј-CTACATGGTGATCACCTCAG-3Ј) with atpDHisrev (5Ј-GTGGTGGTGGTGGTGGTGCACGC-GTTTTCCCTCCTCGCTTTA-3Ј) and atpDHisfwd (5Ј-GTG-CACCACCACCACCACCACAATAAAGGACGTATCATCC-AAGTT-3Ј) with A3rev3 (5Ј-GGAGCCAAAAGGTCAATCAC-3Ј) were used to generate DNA fragments with six histidine codons (correspond to underlined nucleotides). These overlapping fragments were then used for overlap extension PCR with the external primers A3fwd2 and A3rev3, and the product obtained was digested with SspI and SacI.
Expression and Purification of TA2F 1 F o -DK8 harboring plasmid pATPHis5 was grown to an A 600 of 0.35 to 0.4. The culture was induced with 1 mM isopropyl ␤-D-thiogalactopyranoside (IPTG) and incubation continued for 5 h. Cells were harvested, washed with pre-cooled resuspension buffer A (20 mM sodium phosphate, pH 8.0, 2 mM MgCl 2 ), and resuspended in the same buffer. Phenylmethylsulfonyl flouride was added to 0.1 mM, and the cells were disrupted by two passages through a French pressure cell at 20,000 psi. Pancreatic DNaseI 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 ϫ g for 10 min and the inverted membrane vesicles were pelleted from the supernatant at 180,000 ϫ g for 1 h at 4°C. Inverted membrane vesicles were washed in resuspension buffer A, followed by another wash in resuspension buffer A containing 1% (w/v) sodium cholate. To extract TA2F 1 F o from the cytoplasmic membrane, membrane vesicles were diluted to 5 mg of protein/ml in solubilization buffer (20 mM sodium phosphate, pH 8.0, 2 mM MgCl 2 , 10% glycerol, 2% dodecyl maltoside (DDM), 500 mM NaCl, 10 mM imidazole) and incubated at room temperature with gentle stirring for 1 h. The non-solubilized material was removed by ultracentrifugation (180,000 ϫ g, 1 h, 4°C), and the supernatant was incubated with gentle agitation for 45 min at room temperature with a slurry of Nickel-Sepharose High Performance (Amersham Biosciences), that was previously washed with water and equilibrated with IMAC buffer (20 mM sodium phosphate, pH 8.0, 2 mM MgCl 2 , 10% glycerol, 0.05% DDM, 500 mM NaCl) containing 10 mM imidazole. To remove contaminating proteins, the resin was washed with IMAC buffer containing 40 mM imidazole, and TA2F 1 F o was eluted with IMAC buffer supplemented with 150 mM imidazole. To remove excess salt, the eluted TA2F 1 F o was precipitated with 15% (w/v) polyethylene glycol (PEG) 6000 overnight on ice. The precipitate was pelleted by centrifugation at 54,000 ϫ g for 20 min at 4°C and resuspended in resuspension buffer B (20 mM Tris-HCl, pH 8.0, 2 mM MgCl 2 , 10% glycerol, 0.05% DDM). The PEG-precipitated enzyme was applied to a POROS 50 HQ (Applied Biosystems) anion-exchange column, which was equilibrated with resuspension buffer B. Bound proteins were eluted with twelve column volumes of a linear gradient of 0 -600 mM NaCl in the same buffer containing 0.05% DDM. The ATPase-containing fractions were pooled and concentrated to 30 mg/ml using Amicon Ultra centrifugal filter devices (molecular weight cutoff ((MWCO), 100,000). The concentrated sample was then applied to a Superose 6 (Amersham Biosciences) gel filtration column and fractionated with 20 mM Tris-HCl, pH 8.0, 2 mM MgCl 2 , and 100 mM NaCl buffer containing 0.05% DDM. The eluted TA2F 1 F o -containing fractions were pooled and concentrated to 10 mg/ml using Amicon Ultra centrifugal filter devices (MWCO, 100,000). Native TA2F 1 F o was purified from Bacillus sp. TA2.A1 as described previously (14).
Construction and Purification of aK180 Mutants-To substitute the Lys 180 residue (Bacillus sp. TA2.A1 numbering) in the a subunit of TA2F 1 F o , oligonucleotide-directed mutagenesis was carried out using the reverse primer 5Ј-CCTGTGCTGTT-GCTTTATTG-3Ј, located upstream from a PstI site, and Phusion DNA polymerase (Finnzymes) that produces blunt-ended amplification products. Mutagenic forward primers were as follows: 5Ј-ATTTTTGCCGGAGAAACATT-3Ј for aK180G; 5Ј-ATTTTTGCCCACGAAACATT-3Ј for aK180H; and 5Ј-ATT-TTTGCCCGCGAAACATT-3Ј for aK180R, where the first three nucleotides in each primer comprise the 3Ј-half of the blunt-ended SspI recognition site and the underlined nucleotides introduce the mutation. Each of the amplified products was digested with PstI, and the resulting 370-bp product was simultaneously cloned with a 1.1-kb XbaI-SspI fragment and a 1.5-kb PstI-BglII fragment from pTrcATP9, into pATPHis5 digested with XbaI and BglII. The new plasmids were designated pK180G, pK180H, and pK180R. The regions amplified by PCR were confirmed by DNA sequencing, and the recombinant plasmids were transformed into E. coli DK8. The aK180 mutants were expressed and extracted from DK8 inverted membrane vesicles, followed by purification and PEG precipitation as outlined above for TA2F 1 F o .
Preparation of RSO ADP plus P i -loaded Membrane Vesicles-Plasmid transformants of E. coli DK8 were grown to an A 600 of 0.25, before addition of 1 mM IPTG to induce expression. RSO membrane vesicles were prepared from mid-log phase cultures (A 600 of 1.1) by previously described methods (41,42), with the exception that cellular debris and unbroken spheroplasts were removed by low speed centrifugation at 480 ϫ g, 15 min, instead of 730 ϫ g, 30 min. Spheroplasts were prepared in a hypertonic buffer (33 mM Tris-HCl, pH 8.0, 22% sucrose, 20 mM EDTA) containing 1.0 mg/ml lysozyme and gently stirred at room tem-perature for 40 min. Spheroplasts were pelleted by centrifugation at 11,600 ϫ g, resuspended with an 18-gauge needle in 100 mM potassium phosphate buffer (pH 7.8) containing 20 mM MgSO 4 , 20% sucrose (pH 7.8), and 5 mg/ml DNaseI and lysed by osmotic shock, by diluting 100-fold into 50 mM potassium phosphate buffer (pH 7.8) containing 5 mM ADP. The suspension was stirred for 15 min at room temperature before the addition of 10 mM EDTA and 15 mM MgSO 4 , stirring for 15 min at room temperature between steps. The suspension was centrifuged at 22,300 ϫ g, and the pellet was resuspended in precooled 100 mM potassium phosphate buffer (pH 7.8) containing 10 mM EDTA and 5 mM ADP. Cellular debris and unbroken spheroplasts were removed by low speed centrifugation (480 ϫ g, 15 min). RSO membrane vesicles were harvested by centrifuging the supernatant at 180,000 ϫ g for 45 min and resuspending in 100 mM potassium phosphate buffer (pH 7.8) containing 2 mM MgCl 2 and 10% glycerol. Estimates of the fraction of vesicles in the opposite orientation using these methods vary, but they generally range from 10 to 20% (43,44). To confirm vesicle orientation, NADH oxidation activity was measured in RSO membrane vesicles and compared with the same activity in inverted membrane vesicles. NADH (final concentration 0.25 mM) oxidation activity was monitored continuously at 340 nm.
Functional Assays-ATP hydrolysis activity was measured using a spectrophotometric ATP-regenerating assay at 45°C. The assay mixture contained 100 mM Bis-Tris propane (various pH values), 2 mM MgCl 2 , 3 mM phosphoenolpyruvate, 0.25 mM NADH, 0.57 units/ml pyruvate kinase, 3.2 units/ml lactate dehydrogenase, and 2.5 mM ATP. The reaction was initiated by the addition of enzyme into 1 ml of assay mixture, and the rate of NADH oxidation was monitored continuously at 340 nm using a Cary 50 (Varian) spectrophotometer. Approximately 21 g of recombinant TA2F 1 F o was used for measurements. The activity that hydrolyzed 1 mol of ATP per min is defined as 1 unit.
ATP synthesis in inverted membrane vesicles was carried out by the method of Tomashek et al. (45) at 37°C in 1 ml of 10 mM MOPS/Tris-HCl (pH 8.0) buffer containing 2 mM MgCl 2 with stirring. Approximately 0.5 mg of membrane vesicles were incubated with stirring at 37°C for 2 min, followed by incubation in the presence of 2.5 mM NADH for 2 min. When performing inhibition experiments, 100 M DCCD was added 1 min prior to the addition of NADH. ATP synthesis was initiated with the concurrent addition of 0.75 mM ADP and 2.5 mM potassium phosphate (pH 8.0). At various time intervals, 100-l aliquots were removed and transferred to 400 l of stop solution (1% trichloroacetic acid, 2 mM EDTA). Each sample was diluted 500-fold in water prior to the measurement of ATP (see below).
ATP synthesis in ADP (5 mM) plus P i (50 mM)-loaded RSO membrane vesicles was based on the protocol of Tsuchiya (46). Membrane vesicles were diluted 20-fold to a concentration of 0.5 mg/ml into prewarmed (35°C) 25 mM Bis-Tris propane buffer (various pH values) containing 0.25 M sucrose, 5 mM MgCl 2 , and 200 mM K 2 SO 4 . When performing inhibition experiments, membranes were incubated with 100 M DCCD for 2 min prior to dilution. The reaction was initiated within 10 s of dilution by energizing the system with 20 mM potassium ascorbate and 0.1 mM phenazine methosulfate (ascorbate-PMS) with aerated vesicles. Aliquots were removed at various time points, transferred to precooled stop solution (30% perchloric acid, 9 mM Na 2 EDTA), and incubated on ice for 20 min. Denatured protein was removed by centrifugation (13,000 ϫ g for 5 min), and the supernatant neutralized by the addition of 1 M NaOH plus 0.5 M NaCO 3 , followed by snap freezing. Salt was removed from the frozen supernatant by rapid thawing and centrifugation for 5 min at 13,000 ϫ g. To ensure residual perchloric acid did not inhibit subsequent ATP measurements, each sample was diluted 50-fold in water.
Using the diluted samples from inverted and RSO membrane vesicle preparations, the amount of ATP was determined by the luciferin-luciferase assay as described previously (14). To measure ATP, each sample was diluted into 400 l of Tris acetate buffer (50 mM Tris acetate, pH 7.8, 2 mM EDTA, 50 mM MgCl 2 ) in a luminometer tube. 50 l of luciferin-luciferase reagent (Sigma) was added to the tube, and the fluorescence monitored with a chemiluminometer (FB 12 luminometer; Berthold). The amount of ATP synthesized was calculated from a standard curve performed on the day of each set of ATP measurements. For each individual experimental set, the presence of background ATP was measured using non-energized vesicles and subtracted from total ATP measured.
TPMP ϩ Accumulation in RSO Membrane Vesicles-The membrane potential (⌬) generated by ascorbate-PMS energization of RSO membrane vesicles over the pH range 7-10 was determined by measuring the accumulation of [ 3 H]methyltriphenylphosphonium iodide (TPMP ϩ ) (30 -60 Ci/mmol) using filtration assays (0.45-M cellulose-acetate filters) (47). A value of 2.2 l of intravesicular volume per mg of protein was used (48). RSO membrane vesicles (0.6 -0.9 mg/ml) were energized with ascorbate-PMS under identical conditions used for measuring ATP synthesis in RSO membrane vesicles (see above) and the accumulation of [ 3 H]TPMP ϩ (1 M final concentration) measured after 1 min incubation at 35°C. The ⌬ was calculated from the Nernst equation (⌬ ϭ 61 ϫ log [TPMP] in / [TPMP] out ). Nonspecific TPMP ϩ binding was estimated from RSO membrane vesicles that had been treated with either CCCP or a combination of valinomycin and nigericin (10 M each) for 10 min. The results reported are the mean values of three biological replicates and the standard error of the mean associated with these measurements is shown.
SDS-PAGE and Immunoblotting-TA2F 1 F o preparations were routinely analyzed on 16% SDS-polyacrylamide gels in the presence of 0.1% SDS using the buffer system of Laemmli (49). Polypeptide bands were visualized using either Simply Blue Safe Stain (Invitrogen) or Coomassie Brilliant Blue. During immunoblotting, RSO membrane vesicles were subjected to 14% SDS-polyacrylamide gel electrophoresis (SDS-PAGE) followed by electroblotting onto a polyvinylidene difluoride membrane ensuring efficient transfer by including 0.02% SDS in the running buffer. Detection was achieved using a penta-His antibody conjugate (Qiagen) directed against the hexa-histidine tag of the ␤ subunit of the recombinant TA2F 1 F o . The antibodyspecific bands were visualized using the SuperSignal West Pico chemiluminescence system.
Protein Assay-Protein concentrations were determined using a bicinchoninic acid (BCA) protein assay kit (Sigma) with bovine serum albumin as the standard.

RESULTS
Expression of TA2F 1 F o in E. coli-The atp operon, coding for TA2F 1 F o with a hexa-histidine tag at the N terminus of the ␤ subunit, was cloned into the expression plasmid pTrc99A, and expressed in the unc deletion mutant E. coli DK8. TA2F 1 F o was extracted from E. coli membranes and purified via a three step purification procedure. SDS-PAGE analysis of the purified recombinant enzyme identified seven subunits (viz. ␣, ␤, ␥, ␦, b, ⑀, and c) that corresponded to those of the native F 1 F o purified from Bacillus sp. TA2.A1 ( Fig. 2A, lanes 1 and 2). The a subunit was difficult to visualize in both the native and recombinant TA2F 1 F o -ATP synthase using the staining methods employed here. However, chloroform-methanol extraction as previously described (14) demonstrated that the a subunit was present in both our preparations (data not shown). Moreover, ATP synthesis was inhibited by DCCD confirming that the ATP synthase was indeed coupled (see Fig. 3C). A unique feature of the TA2F 1 F o enzyme purified from Bacillus sp. TA2.A1 is its specific blockage in ATP hydrolysis activity, and this activity can be stimulated Ͼ15-fold with 0.4% LDAO (14). Like the native TA2F 1 F o , the purified recombinant enzyme was blocked in ATP hydrolysis activity, with a specific activity of 0.7 units/mg protein (Fig. 2B). This ATPase activity could be stimulated 25-fold with 0.4% LDAO (Fig. 2B), corresponding to a specific activity of 17.6 units/mg protein, similar to that observed for the native TA2F 1 F o (14). The pH profile of the recombinant TA2F 1 F o enzyme was comparable to that of the native TA2F 1 F o enzyme showing a maximum rate of ATP hydrolysis activity at pH 7.5 (Fig. 2C). To determine whether the TA2F 1 F o enzyme could complement E. coli DK8 ⌬atp, the strain harboring the entire atp operon of Bacillus sp. TA2.A1 (i.e. DK8/pATPHis5) was plated onto M13 minimal medium containing succinate as the sole carbon and energy source over the pH range 7-9.5. As a positive control, E. coli DK8 was transformed with plasmid pBWU13 (37) encoding for ECF 1 F o . The positive control grew on either glucose or succinate over the pH range 7-8.5. E. coli DK8 harboring TA2F 1 F o grew on glucose but no growth was observed on succinate over the pH range tested. Hence, the atp operon encoding for TA2F 1 F o was unable to produce an ATP synthase in the host cells with proper function to support growth of E. coli DK8 on non-fermentable carbon sources over the pH range 7-9.5.
ATP Synthesis in E. coli Inverted Membrane Vesicles-Inverted membrane vesicles were prepared from DK8, DK8-expressing ECF 1 F o , and DK8-expressing TA2F 1 F o (Fig. 3). ATP synthesis was energized by the addition of NADH 2 min prior to initiating ATP synthesis with ADP and P i . Inverted membrane vesicles of DK8 containing ECF 1 F o synthesized ATP at a rate of 403 Ϯ 17 nmol of ATP/min/mg protein and the majority of this activity was inhibited by the addition of DCCD (Fig. 3A). The rate of ATP synthesis observed here agrees well with previous reports, which showed the synthesis rate of ECF 1 F o in inverted membrane vesicles to be 520 nmol ATP/ min/mg protein at pH 7.5 (45). Inverted membrane vesicles of DK8 synthesized ATP at a rate of 174 Ϯ 22 nmol of ATP/ min/mg protein that was not inhibited by DCCD, suggesting that this ATP synthesis was via some endogenous route as these vesicles do not contain ECF 1 F o (Fig. 3B). It is noteworthy that this level of ATP synthesis is similar to the rate of ATP synthesis in ECF 1 F o in the presence of DCCD (Fig. 3A). Inverted membrane vesicles of DK8 containing TA2F 1 F o showed an ATP synthesis rate of 486 Ϯ 19 nmol of ATP/min/mg protein and DCCD inhibited this rate to similar levels seen for ECF 1 F o (Fig.  3C). These data indicated that the TA2F 1 F o enzyme was functional in membrane vesicles of E. coli DK8 and therefore we proceeded to make various mutations in the a subunit of this enzyme for functional studies.
Expression of aK180 Mutants in E. coli DK8-To characterize the role of the conserved lysine residue at position 180 (Bacillus sp. TA2.A1 numbering) in the a subunit of the TA2F 1 F o enzyme (identified in Fig. 1), three amino acid substitutions were made in pATPHis5 (viz. aK180H, aK180G, aK180R). Mutant constructs were expressed in E. coli DK8, and aK180H, aK180G, and aK180R were solubilized and purified to determine their biochemical properties. To examine if the aK180 mutant complexes assembled properly, SDS-PAGE analysis was used to compare these to TA2F 1 F o (Fig. 4). SDS-PAGE analysis revealed that all mutant aK180 complexes were assembled in the E. coli membranes and functional assays showed that the ATPase activity (LDAO-stimulated) of these mutant forms was comparable to the recombinant TA2F 1 F o complex (data not shown). The apparent K m for ATP and Mg 2ϩ were also in the range of the values reported for the native TA2F 1 F o enzyme (data not shown). The pH profile (pH 6.5-9.0) of each aK180 mutant was determined and each mutant had a pH optimum of ϳ7.5 (data not shown), indicating that the substitution of the lysine residue for histidine, glycine, and arginine did not change the pH profile of the mutant enzyme in the ATP hydrolysis direction. None of the aK180 mutants were able to complement E. coli DK8 for growth on minimal media containing succinate as the sole carbon source and energy source.
ATP Synthesis of aK180 Mutants in E. coli DK8 Inverted Membrane Vesicles-Inverted membrane vesicles of DK8 containing aK180H, aK180G, and aK180R were prepared with an internal pH of 8.0. Time course ATP synthesis assays energized by NADH were conducted independently for each set of vesicles with or without the addition of the specific ATPase inhibitor DCCD at pH 8.0 (Fig. 5). aK180H and aK180G showed ATP synthesis rates of 973 Ϯ 11 and 735 Ϯ 8 nmol of ATP/ min/mg protein, respectively (Fig. 5, A and B). These rates were 1.5-to 2-fold higher than for the TA2F 1 F o enzyme (see Fig. 3C). With the aK180R enzyme, a rate of ATP synthesis similar to TA2F 1 F o was observed (i.e. 386 Ϯ 11 nmol of ATP/min/mg protein) (Fig. 5C). The addition of DCCD caused a significant inhibition of ATP synthesis in all aK180 mutants (Fig. 5, A-C). These experiments were performed at an external pH of 8.0 that mimics the intracellular environment of the catalytic F 1 portion of the enzyme using inverted membrane vesicles (i.e. F 1 on the outside). In order to assess the effect of external pH on ATP synthesis, we prepared ADP plus P i -loaded RSO membrane vesicles where the F 1 portion was on the inside of the vesicles (maintained at pH 7.5-8.0) and the F o (i.e. a subunit) could be exposed to different external pH values (e.g. pH 7.0 -10.0). E. coli has been shown to grow at an external pH of 8.7 (50,51), and experiments have been performed previously with RSO membrane vesicles at external pH values of 2.5-8.5 (46,52).
ATP Synthesis of TA2F 1 F o and aK180 Mutants in ADP plus P i -loaded RSO Membrane Vesicles-RSO membrane vesicles were prepared using a standard E. coli procedure (41,42) from each strain with an internal pH of 7.5-8.0 containing 5 mM ADP and 50 mM potassium phosphate. While the protocol is designed to generate RSO membrane vesicles, it has been esti-  mated that 10 -20% of these vesicles may in fact be orientated opposite to the direction intended (43,44). Importantly, in terms of this study, the fraction orientated opposite to the intended direction should be constant from E. coli mutant to mutant. A comparison of NADH oxidation activity at 340 nm between the RSO membrane vesicle and inverted membrane vesicle preparations revealed that while we could detect significant NADH oxidation (880 nmol NADH/min/mg protein) by the inverted membrane vesicles, there was negligible activity by the RSO membrane vesicles. On the basis of this data, the vesicles are in the correct orientation (i.e. RSO).
To assess proper subunit assembly within RSO membrane vesicles and TA2F 1 F o expression, an SDS-PAGE of RSO membrane vesicles revealed that all forms of the TA2F 1 F o were assembled and an immunoblot of the purified proteins was performed which targeted the hexa-histidine tag on the N terminus of the ␤ subunit. Expression was equal across all strains (Fig. 6).
ATP synthesis in RSO membrane vesicles was energized using a number of standard methods (e.g. valinomycin-induced K ϩ diffusion potential, NADH oxidation, or ascorbate-PMS). Ascorbate-PMS was the most effective energization source for these experiments. This was validated by measuring the accumulation of [ 3 H]triphenylmethylphosphonium iodide in RSO membrane vesicles of ECF 1 F o and TA2F 1 F o under identical conditions used to measure ATP synthesis over the pH range 7-10. The membrane potential over this pH range, ranged from Ϫ66 to Ϫ75 mV (interior negative) for RSO membrane vesicles of ECF 1 F o and at pH 10 the ⌬ was Ϫ73 Ϯ 4 mV. RSO membrane vesicles of TA2F 1 F o generated a ⌬ of Ϫ69 to Ϫ83 mV over the pH range tested and at pH 10 the ⌬ was Ϫ81 Ϯ 8. These data demonstrate that the external pH, during the course of the incubation, did not have a significant effect on the ability of RSO membrane vesicles to generate a significant ⌬ that could be used to drive ATP synthesis. Furthermore, the ⌬ values reported here are in good agreement with previous reports using RSO membrane vesicles of E. coli energized via the oxidation of ascorbate-PMS (i.e. Ϫ75 mV over the pH range 7-8.5) (47,53). The ⌬ generated under these conditions has been shown to be sufficient to drive energetic processes (e.g. lactose, proline, and serine transport) (54).
ATP synthesis experiments were initially performed with DK8 membrane vesicles, and DK8 membrane vesicles harboring ECF 1 F o and TA2F 1 F o at an external pH of 8.0 using ascorbate-PMS energization. These experiments showed ECF 1 F o FIGURE 5. ATP synthesis of aK180 mutants in E. coli inverted membrane vesicles. A-C, time course ATP synthesis assays at pH 8.0 and 37°C with 0.5 mg of inverted membrane vesicles using the ATP synthesis inverted membrane vesicle assay described under "Experimental Procedures." Membranes were preincubated for 2 min with 2.5 mM NADH with stirring before the reaction was initiated using 0.75 mM ADP and 2.5 mM P i . Open circles, with no DCCD; closed circles, a 1-min preincubation with 100 M DCCD. A, aK180H; B, aK180G; C, aK180R. Each point is the result of three biological replicates, and the standard error associated with this measurement is shown. performed ATP synthesis at a rate of 1.66 Ϯ 0.12 nmol of ATP/ min/mg protein with most of this activity being inhibited by DCCD (Fig. 7A). RSO membrane vesicles of DK8 vesicles showed low rates of ATP synthesis both in the absence and presence of DCCD (0.25 Ϯ 0.11 and 0.12 Ϯ 0.01 nmol of ATP/ min/mg protein, respectively) (Fig. 7B). The rate of ATP synthesis for DK8 membrane vesicles containing TA2F 1 F o was 1.14 Ϯ 0.08 nmol of ATP/min/mg protein, and this rate was almost completely abolished by DCCD (Fig. 7C).
Having confirmed that the experimental system was reproducible and in agreement for E. coli ATP synthesis rates published (46), time course ATP synthesis assays were conducted independently over a pH range from 7.0 to 10.0 for each set of RSO membrane vesicles (Fig. 8). ECF 1 F o showed high ATP synthesis activity at pH 7.0 and 7.5 (6.23 Ϯ 0.12 and 5.52 Ϯ 0.14 nmol ATP/min/mg protein, respectively) (Fig. 8A). As the external pH was increased, the rate of ATP synthesis decreased dramatically and no ATP synthesis was measurable at pH Ͼ 8.5 (Fig. 8A). RSO membrane vesicles of DK8 had no detectable levels of ATP synthesis over the entire pH range studied (Fig.  8B). The ATP synthesis pH profile of TA2F 1 F o was opposite to ECF 1 F o with a pH optimum between 9.0 -9.5 and no significant rates of ATP synthesis observed below pH 8.0 (Fig. 8C). All of the ATP synthesis under these conditions was sensitive to DCCD (Fig. 8C). aK180R had a pH profile for ATP synthesis (all DCCD sensitive) that was comparable to that of TA2F 1 F o (Fig.  8F). When the K180 in the a subunit was substituted for either histidine or glycine, the pH profile of the TA2F 1 F o enzyme was shifted toward the neutral pH range with ATP synthesis rates peaking at pH 8.0 and pH 7.5 for aK180H and aK180G, respectively (Fig. 8, D and E).
Based on the observation that the aK180H and aK180G mutants exhibited a similar pH profile to ECF 1 F o for ATP synthesis, we tested the ability of these constructs to complement growth of strain DK8 ⌬atp over the pH range 7-9.5 on M13 minimal medium containing succinate as the sole carbon and energy source. Under no conditions was complementation (i.e. growth) observed on succinate-containing medium (data not shown).

DISCUSSION
The ATP synthases from bacteria, mitochondria, and chloroplasts are highly conserved enzymes that have evolved to capture and translocate protons or sodium ions from the bulk phase to drive rotation of an oligomeric c-ring coupled to the synthesis of ATP at the catalytic ␤ subunits. In bacteria, two different classes of F-type ATP synthases have been reported, those that are coupled to protons and another class that employ sodium ions (1). In those ATP synthases that use sodium ions, a specific sodium-binding motif has been identified in the c subunit (viz. Q 28 , E 61 , and S 62 ) (7,(55)(56)(57)(58). For those bacteria growing at alkaline pH, very few protons (e.g. pH 10 [H ϩ ] ϭ 1 ϫ 10 Ϫ10 M) exist in the bulk phase and therefore it would seem disadvantageous to couple membrane-bound bioenergetic processes to proton translocation. The inverted pH gradient in these bacteria leads to an overall low electrochemical potential of protons and the ATP synthase works against a large Z⌬pH (i.e. 2 pH units) to synthesize ATP. Despite these apparent ther- modynamic problems and the increased energetic cost for pH homeostasis, aerobic alkaliphilic bacteria grow optimally at high pH values on non-fermentable carbon sources and couple ATP synthesis to protons and not sodium ions.
Current studies from our laboratory on the thermoalkaliphile Bacillus sp. TA2.A1 has shown that this bacterium is unable to grow below pH 9.0 in pH-controlled batch culture on strictly non-fermentable carbon sources (e.g. succinate and malate). 5 However, strain TA2.A1 is able to grow from pH 7.5 to 10 on fermentable carbon sources (e.g. sucrose and glutamate) (19) suggesting that the oxidative phosphorylation machinery of strain TA2.A1 may be specialized to function at high pH only. These data question the nomenclature used to describe alkaliphilic bacteria as obligate or facultative, based on their pH profile for growth (59). For example, strain TA2.A1 could be termed a facultative alkaliphile (i.e. pH 7.5-10) during growth on fermentable carbon sources (19), but an obligate alkaliphile during growth on non-fermentable carbon sources (i.e. ϾpH 9.0). 5 In the current study, we hypothesized that the ATP synthase in strain TA2.A1 may be better adapted to function at high pH versus neutral pH values thus providing an explanation for the lack of growth at pH values below 9 on non-fermentable carbon sources.
To address the above hypothesis we developed a heterologous system to express the TA2F 1 F o enzyme in E. coli DK8 (⌬atp) for functional studies and used ADP plus P i -loaded RSO membrane vesicles of E. coli to measure ATP synthesis. In the absence of a genetic system for Bacillus sp. TA2.A1, the recombinant TA2F 1 F o system enabled us to make various amino acid substitutions to the invariant lysine residue at position 180 of the a subunit. Mutations were created substituting aK180 for either histidine (pK a ϭ 6.1), arginine (pK a ϭ 12.84), or glycine and the effect of external pH on ATP synthesis was studied. TA2F 1 F o synthesized ATP between external pH values of 8.5 to 9.5 that was DCCD-sensitive. A similar pH profile was observed for aK180R. When histidine was substituted for lysine 180, the pH profile of the enzyme was much broader (pH 7-9.5) and shifted into the neutral pH range (pH optimum of 8). The aK180G mutant exhibited a pH profile that was comparable to that of E. coli. On the basis of these findings, we propose that a residue with a strong base, such as lysine or arginine, is ideally suited to function at alkaline pH, but is inefficient at pH values well below its pK a in solution (e.g. 10.3 for lysine). A residue at position 180 without a base in its side chain, such as glycine, was on the other hand favorable for ATP synthesis at neutral pH but did not support ATP synthesis at high pH. Histidine with a pK a in the neutral range supported ATP synthesis at pH values from neutral to alkaline. It should be pointed out, that in general, there was a poor quantitative correlation between the pH maximum of ATP synthesis activity and the pK a of the residue at position 180. This was particularly evident for histidine where significant ATP synthesis activity was observed at pH 9.0. However, the pK a values assumed here are for the residues in solution and the ionization states under physiological conditions often differ, particularly in a membrane environment (60). For example, Cymes et al. (60) revealed that histidine could participate in proton transfer reactions over the broad pH range 6 to 9 when engineered in the ␣-helical lining of the transmembrane pore of the nicotinic acetylcholine receptor.
According to accessibility studies for hydrophilic probes, glycine 218 of E. coli, which corresponds to aK180 of TA2F 1 F o , is proposed to lie at the entrance of the periplasmic access channel for protons to gain access to the binding sites on the c-ring (29,30). A possible explanation for the observations described above is that Lys 180 with its high pK a is ideally suited for proton acquisition from the bulk phase at high pH. Furthermore, at pH values near the pK a of the side chain amino group of Lys 180 , the captured proton could easily dissociate again and continue its journey through the channel to the c subunit binding sites. At pH values well below the pK a of lysine 180, however, the trapped proton would largely be retained on its amino group and this would cause either a blocking of the proposed proton channel and/or mechanism of proton translocation, and there-  fore ATP synthesis is inhibited. Consistent with this hypothesis is the lack of ATP synthesis in TA2F 1 F o (i.e. aK180) and aK180R enzymes at pH values less than 8.0. A glycine at position 180 can obviously not contribute to proton acquisition from an alkaline environment, nor will it act as a permanent proton trap and hence, with this residue ATP synthesis is most favorable in the neutral pH range. The model presented above is corroborated by observations made with the ATP synthase of E. coli whereby an E. coli aG218K mutant was able to grow on a fermentable carbon source like glucose, but unable to grow on succinate at neutral pH values (61). The effect of higher pH on the growth of this mutant on succinate was not tested.