Insertion Scanning Mutagenesis of Subunit a of the F1F0 ATP Synthase near His245and Implications on Gating of the Proton Channel*

Subunit a of the E. coliF1F0 ATP synthase was probed by insertion scanning mutagenesis in a region between residues Glu219and His245. A series of single amino acid insertions, of both alanine and aspartic acid, were constructed after the following residues: 225, 229, 233, 238, 243, and 245. The mutants were tested for growth yield, binding of F1 to membranes, dicyclohexylcarbodiimide sensitivity of ATPase activity, ATP-driven proton translocation, and passive proton permeability of membranes stripped of F1. Significant loss of function was seen only with insertions after positions 238 and 243. In contrast, both insertions after residue 225 and the alanine insertion after residue 245 were nearly identical in function to the wild type. The other insertions showed an intermediate loss of function. Missense mutations of His245 to serine and cysteine were nonfunctional, while the W241C mutant showed nearly normal ATPase function. Replacement of Leu162 by histidine failed to suppress the 245 mutants, but chemical rescue of H245S was partially successful using acetate. An interaction between Trp241 and His245 may be involved in gating a “half-channel” from the periplasmic surface of F0 to Asp61 of subunit a.

Subunit a of the E. coli F 1 F 0 ATP synthase was probed by insertion scanning mutagenesis in a region between residues Glu 219 and His 245 . A series of single amino acid insertions, of both alanine and aspartic acid, were constructed after the following residues: 225, 229, 233, 238, 243, and 245. The mutants were tested for growth yield, binding of F 1 to membranes, dicyclohexylcarbodiimide sensitivity of ATPase activity, ATP-driven proton translocation, and passive proton permeability of membranes stripped of F 1 . Significant loss of function was seen only with insertions after positions 238 and 243. In contrast, both insertions after residue 225 and the alanine insertion after residue 245 were nearly identical in function to the wild type. The other insertions showed an intermediate loss of function. Missense mutations of His 245 to serine and cysteine were nonfunctional, while the W241C mutant showed nearly normal ATPase function. Replacement of Leu 162 by histidine failed to suppress the 245 mutants, but chemical rescue of H245S was partially successful using acetate. An interaction between Trp 241 and His 245 may be involved in gating a "halfchannel" from the periplasmic surface of F 0 to Asp 61 of subunit a.
The F 1 F 0 ATP synthase from Escherichia coli is typical of the ATP synthases found in mitochondria, chloroplasts, and many other bacteria (for recent reviews, see Refs. [1][2][3][4]. It comprises an F 1 complex, which contains the nucleotide-binding subunits involved in catalysis, and an F 0 complex, which conducts protons across the membrane. The enzyme from E. coli appears to be a minimal form of the ATP synthase, with eight essential subunits. Five different subunits are found in F 1 : ␣, ␤, ␥, ␦, and ⑀, in a stoichiometry of 3:3:1:1:1. Three different subunits, a, b, and c, form F 0 with a likely stoichiometry of 1:2:9 -12 (5). The movement of protons through F 0 , from the periplasm to the cytoplasm, drives the net synthesis of ATP by F 1 .
The pathway of protons through F 0 is thought to involve subunits a and c. Subunit b is essential for assembly of a functional F 0 (6, 7) but is not likely to participate directly in proton translocation. The subunits b are embedded in the membrane through a span of about 30 hydrophobic amino acids at the NH 2 terminus. NMR studies of c (8,9) have confirmed the ␣-helical hairpin structure of the two predicted transmembrane spans and have provided details about the environment of the essential residue Asp 61 (10,11). Questions remain about the oligomeric structure of c subunits and how they interact with subunits a and b. Atomic force microscopy (12,13) and electron spectroscopic imaging (14) have provided some evidence for a ring of 9 -12 c subunits adjacent to the subunits a and b.
Subunit a is the largest of the F 0 subunits (271 residues). Mutagenesis has revealed that in addition to Asp 61 of subunit c, three residues in subunit a seem to be important in proton translocation: Arg 210 , Glu 219 , and His 245 (15)(16)(17)(18)(19). Of these three, only Arg 210 is strictly conserved among all known sequences, which suggests a unique role for this residue. No substitutions at this position permit ATP synthesis, but an alanine substitution allows limited passive proton permeability (20). The positions of the Glu and His residues are found to be reversed in some organisms, such as human and bovine mitochondria. Other sequences lack an ionizable residue at one of these positions, such as Bacillus subtilis and spinach chloroplast. Correspondingly, some amino acid substitutions at positions 219 and 245 in E. coli have been found to be partially functional (16,21). Therefore, these two residues can be considered to be of secondary importance.
In a previous study (22), we probed subunit a by a series of single amino acid insertions at seven distinct locations between residues 187 and 222. Alanine insertions after residues 212, 217, and 222 were highly disruptive of function, consistent with the importance of this region in proton translocation. Studies of other membrane proteins by single-amino acid insertions have been described, including bacteriorhodopsin (23) and lac permease (24). In the present study, we have extended the analysis of subunit a from residue 229 to 245. The results indicate a single region of disruption, between residues 238 and 245, but insertions after 245 have little effect.

EXPERIMENTAL PROCEDURES
Materials-Restriction endonucleases and T4 DNA ligase were obtained from New England Biolabs, Inc. Materials for silver sequencing and plasmid minipreps were obtained from Promega Corp. Synthetic oligonucleotides were obtained from Operon Technologies, Inc. or National Biosciences. Urea was from International Biotechnologies, Inc. ACMA 1 was obtained from Molecular Probes. Dicyclohexylcarbodiimide (DCCD) and octyl glucoside were obtained from Sigma. Immunoblotting reagents and detergent-compatible protein assay materials were obtained from Bio-Rad. Nickel-nitrilotriacetic acid-agarose was obtained from Qiagen. ATP and anti-HA antibody were obtained from Boehringer Mannheim.
Bacterial Strains-Strain XL1-Blue (recA1, endA1, gyrA96, thi, hsdR17(r k Ϫ ,mk ϩ ), supE44, relA1, Ϫ ,(lac), F', proAB, lacI q Z⌬M15, Tn10(Tet R )) was obtained from Stratagene and was used for subclonings and mutagenesis. Strain RH305 (uncB205, recA56, srl::Tn10, bglR, * This work was supported by National Institutes of Health Grant GM40508 and by the Welch Foundation. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
thi-1, rel-1, Hfr PO1) (25) was used to characterize mutations in uncB. It produces an a subunit that is truncated near Pro 240 (26) and is complemented by plasmids containing a wild type uncB gene. Cultures were grown at 37°C, and cell density was monitored by optical density at 600 nm using a Milton Roy 1001 spectrophotometer. Rich medium was LB supplemented with 0.2% glucose, and minimal medium was A salts supplemented with succinate (0.2%) or with glucose, as indicated (27). Media were supplemented with chloramphenicol (34 mg/liter) or tetracycline (12.5 mg/liter) as appropriate. Growth yields were determined in minimal A supplemented with 7 mM glucose.
Construction of Plasmids and Insertions-Plasmids used in this study are shown in Fig. 1. The uncB mutations analyzed in this study were constructed using the cassette mutagenesis technique, as described previously (28). To construct pSBV18, plasmid pSBV16 (21) was digested with BsaHI and BamHI to excise a 686-bp fragment. It was replaced in a two-step procedure. First, a synthetic 69-bp fragment ( Fig.  2A) was ligated to the remaining large fragment of pSBV16. This construct retained the BsaHI site, but the BamHI site was lost. However, a BstYI site was created at the same position, which can generate compatible cohesive ends. Second, the new plasmid was digested with BstYI, producing a single fragment, and was ligated to the 617-bp BamHI fragment from pSBV11 (29). This produced pSBV18, shown in Plasmid pBJA46 was constructed from pBJA1018 by digesting with DraIII and PstI, which excised a 125-bp fragment, followed by ligation to a synthetic fragment of 127 bp (Fig. 2B), as described previously (28). This introduced several unique restriction sites via silent mutagenesis, such as AvrII, and an EcoRV site that was created by introducing two extra bases, G(AT)ATC, in the region coding for Ile 166 . Therefore, pBJA46 does not code for a full-length subunit a; but since EcoRV generates blunt ends by cleaving between the ATs, it can be used in cassette mutagenesis. Using EcoRV and AvrII, it was possible to construct the L162H mutant and to regenerate a gene coding for the wild type subunit a (Fig. 3B).
Construction of pTW1HisHA is described in an accompanying paper (30). Its features include a unique BamHI site (as in pSBV18) and a hemagglutinin epitope (HA) near the COOH terminus followed by a hexahistidine tag. The HA consists of the sequence YPYDVPDYA and can be recognized by a monoclonal antibody raised against that peptide (Boehringer Mannheim). Insertion mutations (Fig. 3A) at positions 238 and 243 and missense mutation (Fig. 3B) at 241 were constructed in pTW1HisHA.
Preparation of Cell Fractions and Assays-Fractionation of cells and isolation of membranes and stripped membranes were carried out as described previously (15). Protein concentrations were determined by a detergent-compatible protein assay (Bio-Rad) using bovine serum albumin as standard. ATP hydrolysis assays and fluorescence quenching assays were performed essentially as described previously (28). ATP hydrolysis was measured in 50 mM Tris-HCl (pH 9.1), 1 mM MgCl 2 , 3 mM ATP at 37°C, and fluorescence assays were measured using 400 mg/liter of membrane protein in a solution of 50 mM MOPS (pH 7.3), 10 mM MgCl 2, 1 M ACMA, and either 0.5 mM NADH or 0.1 mM ATP, as appropriate. The excitation wavelength was 410 nm, and the emission wavelength was 490 nm. Inhibition of ATP hydrolysis by DCCD was measured as described previously (28).
Immunoblotting-Membranes were prepared as above from 250-ml cultures. The membrane pellets were suspended in 1.5 ml of a detergent solution containing 0.1% deoxycholate, 0.5% cholate, 1.5% octyl glucoside, and 50 mM Tris-HCl, pH 7.5, and incubated for 1 h at 25°C. The solubilized membranes were centrifuged for 10 min at 16,000 ϫ g in a microcentrifuge, and the supernatant fraction was mixed with 0.4 ml of nickel-nitrilotriacetic acid-agarose and incubated overnight at 4°C. The agarose was washed according to the manufacturer's (Qiagen) batch method, using the detergent solution above, and eluted by suspending the agarose in 80 l of detergent solution plus 0.5 M imidazole. After a 10-min incubation at 25°C and 10 min of microcentrifugation at 16,000 ϫ g, the supernatant fraction was collected as partially purified subunit a. The samples were electrophoresed and transferred to nitrocellulose as described previously (21). The a subunits were detected by probing first with mouse anti-HA antibody (5 mg/liter) followed by goat anti-mouse IgG-alkaline phosphatase conjugate, according to the manufacturer's instructions.

RESULTS
A series of single amino acid insertions, both alanine and aspartic acid, were constructed after residues 225, 229, 233,  238, 243, and 245 in the a subunit of the F 1 F 0 ATP synthase, as described under "Experimental Procedures." These sites span the two residues of secondary importance in proton translocation, Glu 219 and His 245 . Growth yields of E. coli strains carrying these mutations were determined and are shown in Tables I and II. The four mutants with the lowest growth yields, those at positions 238 and 243, were all unable to grow on succinate minimal medium, indicating a deficiency in oxidative phosphorylation.
Membrane vesicles were prepared from cells carrying each of the 12 insertion mutations. The membrane and supernatant fractions were tested for ATPase activity to determine the relative binding of the F 1 sector to the membranes. The results are reported in Tables I and II. Alanine insertions (Table I) had little effect on F 1 binding after residues 225, 243, and 245, but after residue 238 significant loss of F 1 binding occurred. The ATPase activities of the membrane fractions were also tested for sensitivity to DCCD, a reagent that reacts specifically with the subunit c of the F 0 complex (31). These results are also presented in Tables I and II. Alanine insertions (Table I) after residues 225 and 245 had no detectable effect, while those after residues 238 and 243 caused a loss of sensitivity. The membrane vesicles from all 12 insertion mutants were also tested for ATP-dependent proton translocation activity, using the fluorescent dye ACMA, and the results are shown in Fig. 4. Nearly normal ATP-driven proton translocation was seen with the mutants 225ٌA, 225ٌD, and 245ٌA. Little or no proton translocation was seen with mutants 238ٌA, 238ٌD, 243ٌA, and 243ٌD. The other insertion mutants showed intermediate levels of proton translocation that were at least 50% of the wild type level. The membranes were also stripped of F 1 and assayed for passive proton permeability, using NADH to generate a proton gradient. These results are presented in Fig.  5. In general, the same trends were observed, but the aspartate insertions had a noticeably more significant effect on passive proton permeability than did the alanine insertions. Membranes from the four mutants unable to grow on succinate minimal medium (238ٌA,D, and 243ٌA,D) were tested for the level of subunit a by immunoblotting. The results, presented in Fig. 6, indicate that subunit a is present at somewhat reduced levels in the 238ٌA,D mutants and at even lower levels in the 243ٌA,D mutants.
The residue His 245 was replaced by Gly, Ser, and Cys to see if small or polar residues at that position would retain any function. None of the three mutants was able to grow on succinate minimal medium after 2 days of growth at 37°C. Some growth was seen when the H245S strain was supplemented with 10 -100 mM sodium acetate. This had no effect on the inability of the uncB strain RH305 to grow on succinate minimal medium. Attempts to demonstrate enhanced ATP-dependent proton translocation, by preparing membranes or performing assays in the presence of 10 -100 mM acetate, all failed to FIG. 3. Oligonucleotides used to construct mutations. A, insertion mutations. The site of the insertion follows the residue number shown at the left. The restriction sites used in the cassette mutagenesis are also indicated. The nucleotides inserted are GMC or GMT (M represents A plus C), resulting in an equal mixture of alanine and aspartic acid. B, missense mutations. The residue number of the missense mutation is indicated at the left, and the restriction sites used are also shown. At position 245, nucleotides KST (K represents G plus T; S represents G plus C) code for glycine, alanine, cysteine, and serine. At position 241, nucleotides TGT code for cysteine. At position 162, nucleotides CWC (W represents A plus T) code for leucine (wild type) and histidine.

FIG. 4. ATP-driven proton translocation assays of alanine insertions (A) and aspartate insertions (B).
Proton gradient formation is indicated by quenching of the fluorescence of ACMA, as described under "Experimental Procedures." show significant differences. In contrast, membranes prepared from W241C showed normal rates of proton translocation, although passive proton permeability was somewhat reduced from the wild type level (results not shown).
In a previous study (21), residue Leu 162 of subunit a in E. coli was identified as corresponding to a histidine that is conserved among all known Bacillus species. Three double mutants were constructed to see if relocating His 245 to position 162 would retain function: L162H/H245G, L162H/H245S, and L162H/ H245C. It was determined that none could grow on succinate minimal medium. DISCUSSION All subunits of the E. coli F 1 F 0 ATP synthase are essential for ATP synthesis. Subunit a has been implicated in proton translocation through F 0 , but its precise role remains largely unknown. Single amino acid insertion-scanning mutagenesis has been applied to a region of subunit a between the important residues Glu 219 and His 245 . The goal was to detect amino acid residues in subunit a that are important in proton translocation and regions of subunit a that are in close contact with other subunits. This study is a continuation of a previous study (22) in which insertions of alanine and aspartate were made after residues 187,193,198,202,212,217, and 222 of subunit a. In that study, it was found that insertions of both alanine and aspartate after residues 212, 217, and 222 reduce or abolish ATP-driven proton translocation. In particular, the 222 insertions were the most severe and showed disruption of F 1 binding to membranes. Interestingly, this region of subunit a had also been found to be the location of several second-site suppressors of mutations in subunit c (32). Most of the rest of the insertions had little or no effect on function; therefore, no other residues were implicated in function.
In this study, a similar pattern of results was obtained. Insertions of alanine or aspartate after residues 238 or 243 disrupted oxidative phosphorylation such that the strains could not grow on succinate minimal medium. Assays of the membrane fractions indicated that F 1 binding was decreased significantly in the case of 238ٌA, while some ATP-driven proton translocation remained. In contrast, membranes from the 243ٌA mutant had normal F 1 binding but no ATP-driven proton translocation. In both cases, the DCCD sensitivity of ATPase activity was very low. These results indicate two different types of disruption by the insertions of alanine at 238 and at 243.
The phenotype of 238ٌA is consistent with a disruption of a-b or a-c interactions. In a previous study, Kumamoto and Simoni (33) discovered that Pro 240 was the site of two suppressor mutations of subunit b mutation G9D. This subunit b mutation had been isolated and characterized by two groups (34,35) and was further analyzed by immunoprecipitation (36). Its properties include inability to grow on succinate minimal medium, reduced binding of F 1 to membranes, and greatly diminished ATP-driven proton translocation and sensitivity of ATPase activity to DCCD. Two virtually identical partial suppressors were identified as P240A and P240L in subunit a (33). The suppressors permitted growth on succinate minimal medium and significant ATP-driven proton translocation but had little effect on F 1 binding or DCCD sensitivity. Detergent solubilization and immunoprecipitation studies of the original mutant, G9D, indicated very low levels of subunit a in F 1 F 0 complexes, suggesting that the mutation affects a-b interactions. In the presence of the suppressor mutation, a normal complex of subunits a, b, and c was detected (36). The results presented here, including loss of F 1 binding and DCCD sensitivity, are also consistent with an important interaction between subunit a in the region of Pro 240 and either the membrane-spanning NH 2 terminus of subunit b or subunit c.
The phenotype of 243ٌA was quite different, in that F 1 remained bound to the membranes, but ATP-driven proton translocation and sensitivity to DCCD were completely missing. The immunoblot using anti-HA showed at least a low level of subunit a, which presumably is necessary for normal F 1 FIG. 5. Passive proton permeability assays of stripped membranes prepared from alanine insertions (A) and aspartate insertions (B). Proton permeability through F 0 stripped of F 1 is indicated by diminished ability to quench fluorescence of ACMA in the presence of NADH, which rapidly builds a proton gradient. Preparation of stripped membranes and assay conditions are described under "Experimental Procedures." binding. Therefore, the loss of function is not consistent with a large scale disruption of subunit interactions but rather suggests a more specific effect. Further insight into the 243ٌA,D mutants may be gained by considering the insertions after residue 245 and the conservation of residues 241-245. Although His 245 seems to be a functionally important residue from the results of mutagenesis (15,19), the insertion of alanine following it has very little effect on function. One interpretation is that His 245 does not interact with other residues toward the carboxyl terminus of subunit a but that the insertion at 243 disrupts an important interaction between the side chain of His 245 and a nearby residue toward the amino terminus. The best candidate would be Trp 241 , for the following two reasons. First, a tryptophan is found at this position in all subunit a sequences that have a histidine at position 245 (Fig.  7). Second, a recent study (37) of model peptides has revealed a pH-dependent stabilization of an ␣-helical conformation mediated by Trp n and His n ϩ 4 . This stabilization was not seen for peptides with Trp n /His n ϩ 3 or for His n /Trp n ϩ 4 . An analysis of proteins of known three-dimensional structure has revealed a significant number of instances of interactions in ␣-helices between Trp n and His n ϩ 4 . The side chains were found to be oriented for an interaction between the protonated imidazolium of His and the planar indole ring of tryptophan. Such an interaction at the periplasmic surface of F 0 could function as a proton gate, regulated by the pK a of His 245 . In this way, at sufficiently low pH, the imidazole of His 245 would be protonated, allowing an interaction with Trp 241 , and thereby stabilizing the ␣-helix. This interaction might, through conformational or pK a shifts, permit access of protons to Asp 61 of subunit c. In such a scheme, neither His 245 nor any other residues of subunit a, would be on the pathway of protons from the periplasm to Asp 61 of subunit c. Rather, the pathway would be a pH-gated, water-filled "half-channel." It is interesting to note that another second-site suppressor of the subunit b G9D mutation was isolated in subunit c, A62S, near Asp 61 (38).
Analysis of the other insertions from residue 225 to 233 revealed, at most, intermediate effects on function. Both insertions after residue 225 were nearly identical to the wild type, suggesting that this position is not near a catalytic residue or an important subunit interaction. Insertions after residues 229 and 233 showed some loss of function, in particular the 233ٌD mutant. These results seem to reflect the proximity of these positions to the 238 region and may be affecting subunit interactions or, indirectly, the conformation of His 245 .
The replacement of His 245 with the small side chain amino acids glycine, serine, and cysteine led to total loss of function, as indicated by an inability to grow on succinate minimal medium. This suggested that water or fortuitous, neighboring side chains could not substitute for His 245 . Chemical rescue has been achieved for certain Asp mutants of bacteriorhodopsin using acetate and azide (39) and with carbonic anhydrase His mutants using imidazole (40), but our success was limited. Only when acetate, but not imidazole, was added to the growth medium of H245S was growth detectable on succinate minimal medium. These results are consistent with His 245 being involved with gating rather than being on the proton pathway, because it is unlikely that an imidazole-Trp 241 interaction would induce a conformational change, but the smaller acetate might facilitate proton entry to a suboptimal channel.
The recent results of labeling studies (30) essentially eliminated the possibility that histidine at position Leu 162 could substitute for the mutated His 245 , since the findings reported here were that no suppression occurred. In light of the findings here and of the membrane topology indicated by labeling studies, a revised model for a-c interactions and proton translocation is presented in Fig. 8. In this model, the half-channel from the periplasm to Asp 61 of subunit c is gated by both His 245 and Glu 219 . Alanine insertions at residues 238, 241, and 222 are more disruptive than those at 245 and 217, indicating important conformations on the periplasmic side of these residues. Arg 210 resides at the other end of this half-channel, interacting with Asp 61 of a subunit c. The half-channel from the second Asp 61 to the cytoplasmic surface is drawn accessible to solvent, and this would keep the aspartate unprotonated. Glu 196 , a conserved residue, is located near the cytoplasmic surface of this half-channel as a facilitator of proton transport. Mutagenesis of this residue (28) showed a charge-and polarity-dependent effect on rates of proton translocation. High resolution structural information about F 0 subunits will be necessary for further insight into the mechanism of proton translocation.
Acknowledgments-We thank Lata Narawane and Takaaki Wada for construction of some of the plasmids and for assistance in immunoblots.