On the Role of Arg-210 and Glu-219 of Subunit a in Proton Translocation by the Escherichia coliF0F1-ATP Synthase*

A strain of Escherichia coli was constructed which had a complete deletion of the chromosomaluncB gene encoding subunit a of the F0F1-ATP synthase. Gene replacement was facilitated by a selection protocol that utilized the sacBgene of Bacillus subtilis cloned in a kanamycin resistance cartridge (Ried, J. L., and Collmer, A. (1987)Gene (Amst.) 57, 239–246). F0subunits b and c inserted normally into the membrane in the ΔuncB strain. This observation confirms a previous report (Hermolin, J., and Fillingame, R. H. (1995)J. Biol. Chem. 270, 2815–2817) that subunita is not required for the insertion of subunitsb and c. The ΔuncB strain has been used to characterize mutations in Arg-210 and Glu-219 of subunita, residues previously postulated to be essential in proton translocation. The aE219G and aE219K mutants grew on a succinate carbon source via oxidative phosphorylation and membranes from these mutants exhibited ATPase-coupled proton translocation (i.e. ATP driven 9-amino-6-chloromethoxyacridine quenching responses that were 60–80% of wild type membranes). We conclude that the aGlu-219 residue cannot play a critical role in proton translocation. TheaR210A mutant did not grow on succinate and membranes exhibited no ATPase-coupled proton translocation. However, on removal of F1 from membrane, the aR210A mutant F0 was active in passive proton translocation,i.e. in dissipating the ΔpH normally established by NADH oxidation with these membrane vesicles. aR210A membranes with F1 bound were also proton permeable. Arg-210 of subunit a may play a critical role in active H+transport that is coupled to ATP synthesis or hydrolysis, but is not essential for the translocation of protons across the membranes.

F 0 F 1 -ATP synthases catalyze the synthesis of ATP during oxidative phosphorylation utilizing the energy of an electrochemical proton gradient generated by H ϩ pumping electron transport complexes (1). Structurally similar F 0 F 1 -ATP synthases are also present in mitochondria, chloroplasts, and most eubacteria. The enzymes are composed of two structurally and functionally distinct sectors termed F 1 and F 0 . The F 1 sector is bound at the membrane surface and the F 0 sector spans the membrane. In Escherichia coli, F 1 consists of five subunits in a stoichiometric ratio of ␣ 3 ␤ 3 ␥ 1 ␦ 1 ⑀ 1 and F 0 consists of three different types of subunits present in a stoichiometric ratio of a 1 b 2 c 10ϩ1 (2). The c subunit spans the membrane as a hairpin of two ␣-helices and contains an essential aspartate (Asp-61) in the second transmembrane helix. Asp-61 is believed to be the key H ϩ -binding site in the proton conduction pathway through F 0 . The polar loop connecting the transmembrane helices in subunit c is thought to function in coupling of proton translocation to structural changes in F 1 leading to ATP synthesis. The bulk of the b subunit extends from the membrane surface and is responsible for binding of F 1 to F 0 . The role of subunit a in F 0 function has not been clearly defined. It is not thought to play a major role in F 1 binding but has been postulated to play a key role in proton translocation (2)(3)(4)(5).
All genetic studies on the function of subunit a have been carried out by complementation with uncB (subunit a) genes expressed from either plasmids, FЈ episomes or -transducing phage in a background strain carrying a chromosomal uncB truncation mutation, e.g. strains like CP242 (W231stop) (6) or RH305 (V239A, P240W, and W241stop) (7). The presence of the truncated subunit a in the complementation systems used might introduce complications into studies of function or assembly. For example, Hermolin and Fillingame (8) studied the insertion of subunits b and c in a strain carrying a W231stop truncation of subunit a and concluded that these subunits inserted into the membrane independently of subunit a. However, they were unable to rule out a transient role for the truncated subunit a in insertion of subunits b and c. To avoid complications that can arise due to the presence of a truncated subunit, we have constructed a strain that has a complete deletion of the chromosomal uncB gene encoding subunit a. We have re-examined the dependence of subunit insertion in the ⌬uncB background and conclude that subunits b and c do insert independently of subunit a.
The ⌬uncB strain constructed here also provides an improved recipient background for study of plasmid-borne uncB mutations and was used here to study substitutions in Arg-210 and Glu-219, residues previously concluded to play key roles in function. Glu-219 can be mutated to Asp with full retention of function while most other substitutions tested have resulted in enzymes with very little function (9, 10). Vik and Antonio (11) suggested an interaction between Glu-219 and Gln-252 based upon the function of the three mutants with double substitutions in subunit a, i.e. E219G/Q252E, E219D/Q252E, and E219K/Q252E. However, they did not test the phenotype of the E219G and the E219K mutations by themselves. We show here that mutants with the E219G and E219K single substitutions are by themselves functional. These results severely limit the possible function of Glu-219. Arg-210 is thought to be essential as all substitutions for this residue result in loss of function. The R210A mutation, previously studied by Hatch et al. (12), is re-examined here and found to be unusual in that ATP hydrolysis-coupled proton translocation is completely abolished, whereas passive proton translocation through F 0 is unaffected. Arg-210 may be an important component for coupled proton translocation but is not an essential component of the passive proton conductance pathway through F 0 .

EXPERIMENTAL PROCEDURES
Construction of the uncB Deletion-A PCR 1 strategy as outlined in Fig. 1 was used for the construction of the uncB deletion on a plasmid. Primer 1, GTAGTGTTGGTAAATTACCCTTTGTTGTT, was designed to delete nucleotides 1015 to 1839 2 of unc DNA. DNA synthesis initiated with this primer should result in deletion of all sequences between the termination codon of uncI and the termination codon of uncB as shown in Fig. 1. Such a deletion was generated by PCR and the fragment cloned between the HpaI(61) and AvaI(1976) sites in plasmid pVF116 to replace the normal uncB gene resulting in plasmid pVF208 (Fig. 2). The presence of the uncB deletion in the plasmid pVF208 was confirmed by DNA sequencing.
Introduction of the sacRB-nptI Cartridge into the Chromosomal uncB Gene-A 3.8-kilobase sacRB-nptI cartridge (14) was used to disrupt the chromosomal uncB gene. The sacRB-nptI cartridge contains the sacB gene along with its regulatory elements, sacR and the nptI gene. The sacB gene encodes for the enzyme levansucrase which kills cells in the presence of sucrose. The nptI gene makes the cells resistant to kanamycin and can be used to score for the presence of the cartridge. The cartridge was initially cloned between the BamHI sites in the plasmid pAP55 (15), which carries a chloramphenicol resistance gene (Fig. 2). The resulting plasmid was digested with SphI and the linear fragment carrying the cartridge was used to transform a recD strain (17), VF145 (pyrE41, entA403, argH1, rpsL109, supE44, recD::Tn10), selecting for kan R transformants and screening for chromosomal recombinants based upon chloramphenicol sensitivity. The desired recombination event took place at frequency of roughly 10%. Strain VF145, which was used in the step above, was constructed by P1vir transduction of recD::Tn10 from strain RZ7365 (Dr. P. J. Kiley, University of Wisconsin-Madison) into strain MM180 (18). The presence of the cartridge in the chromosomal uncB gene was confirmed by PCR amplification of chromosomal DNA using primers GCTGATCGTTACGTGGG and CTC- 1 The abbreviations used are: PCR, polymerase chain reaction; DCCD, dicyclohexylcarbodiimide; ACMA, 9-amino-6-chloro-2-methoxyacridine; LDAO, lauryldimethylamine oxide; Tricine, N-[2-hydroxy-1,1bis(hydroxymethyl)ethyl]glycine. 2 The unc DNA numbering system corresponds to that used by Walker et al. (13).   (13). The PCR product obtained in the first PCR reaction was used along with plasmid pDF163 (23) as a template in the second PCR reaction. Plasmid pDF163 carries an unc DNA fragment from bases 870 -3216.
CAGTTTGTTTCAGT, initiating DNA synthesis at unc nucleotides 943 and 1879, respectively. The product strain, Sac-14, was concluded to have the sacRB-nptI cartridge inserted between the BamHI sites in the chromosomal uncB gene.
Cartridge Eviction-The presence of the sacRB-nptI cartridge in the strain Sac-14 makes the cells sensitive to sucrose (14). Loss of the cartridge by a double recombination event with unc sequences present on a plasmid should make the cells resistant to sucrose and sensitive to kanamycin. Strain Sac-14 was transformed with the plasmid pVF208 (described above) and chloramphenicol-resistant transformants were isolated. A sample of a pVF208/Sac-14 transformant grown overnight in LB medium (19) was plated on LB medium agar containing 5% sucrose. Sucrose-resistant colonies were screened for sensitivity to kanamycin to identify possible candidates arising from a double recombination event. The desired recombination event took place at a frequency of roughly 60%. The cells were then cured of the plasmid by growth overnight in LB medium in the absence of chloramphenicol. This was easily accomplished due to plasmid instability in the recD background (20). Chloramphenicol-sensitive colonies were screened by replica plating and found at a frequency of roughly 10%. The presence of the uncB deletion in the chromosome was confirmed by PCR amplification of chromosomal DNA using primers GCTGATCGTTACGTGGG and CTCCAGTT-TGTTTCAGT, initiating DNA synthesis at unc nucleotides 810 and 2319, respectively, and DNA sequencing of the PCR-amplified product. The deletion was then transduced into strain MJM63 (pyrE41, entA403, argH1, rpsL109, supE44, ⌬uncE 334, ilv::Tn10) (21) using P1vir by cotransduction with Ilv ϩ . The resulting strain was made recA by cotransduction of recA56 with srl::Tn10 from strain CP242 (F(ϩ), asnA31, asnB32, thi-1, uncB108, recA56, srl:: Tn10) (6). Transduction of recA was confirmed by UV sensitivity (6).
Construction of Glu-219 Mutants-Plasmid pSBV16 derivatives containing the Glu-219 mutations along with the Q252E mutation were provided by Dr. S. B. Vik (Southern Methodist University, Dallas, TX) (11). The Glu-219 mutations were separated from the Gln-252 mutation by subcloning the BamHI fragment from the corresponding plasmid pSBV16 derivative into the BamHI sites in plasmid pCP1200 (6). Plasmid pCP1200 carries the wild type uncB gene on a HindIII(870) to AvaI(1976) fragment of unc DNA cloned between these sites in plasmid pBR322. The presence of the Glu-219 mutations in the resulting plasmids was confirmed by DNA sequencing. Plasmid pSBV16 (11) contains an uncB gene engineered to contain a number of restriction sites. Plasmid pVF247 was constructed as a wild type control for the Glu-219 mutant plasmids by cloning the BamHI fragment from plasmid pSBV16 into plasmid pCP1200.
Construction of R210A Mutant-The oligonucleotide CACTCGGTT-TGGCTCTGTTCGGTAAC corresponds to nucleotides 1640 -1665 of the wild type uncB gene except that the Arg-210 codon is substituted by GCT (underlined) which encodes Ala. The R210A mutation on the oligonucleotide was incorporated into a PCR fragment by the megaprimer method (22). The megaprimer method involves the use of a mutant primer and two wild type primers. Primers TGCTGCCGTA-CATTGCT and CACAGCACAATGCCTCT, which initiate DNA synthesis at nucleotides 1397 and 2189, respectively, were used as the wild type primers. The resulting PCR fragment was cloned between the PstI(1565) and AvaI(1976) sites in plasmid pDF163 (23). Plasmid pDF163 contains the HindIII(870) to SphI(3216) fragment of unc DNA cloned between these sites in plasmid pBR322. The presence of the R210A mutation was confirmed by DNA sequencing.
Biochemical Assays-Cells were grown in M63 minimal medium containing 0.6% glucose supplemented with 2 mg/liter thiamine, 0.2 mM uracil, 0.2 mM L-arginine, 0.02 mM dihydrobenzoic acid supplemented with 10% LB medium (19). Ampicillin was added to a final concentration of 0.1 mg/ml when needed. Membranes were prepared in TMDG buffer (50 mM Tris-HCl, pH 7.5, 5 mM MgCl 2 , 1 mM dithiothreitol, and 10% glycerol) by rupture of cells using a French Press (18). ATPase activity was assayed in ATPase assay buffer (50 mM Tris-H 2 SO 4 , pH 7.8, 0.2 mM MgSO 4 ) using 0.4 mM [␥-32 P]ATP (24). The values reported are averages of duplicate assays. The extent of DCCD inhibition was determined after incubating membranes with 30 M DCCD for 10 min at 30°C in ATPase assay buffer. Protein concentration was determined using a modified Lowry protocol (24). ATP-driven quenching experiments were performed in HMK buffer (10 mM HEPES-KOH, pH 7.5, 5 mM MgCl 2 , 0.3 M KCl) using ACMA at a final concentration of 0.3 g/ml (25). NADH-driven quenching experiments were performed in HMK buffer using quinacrine at a final concentration of 0.375 g/ml (25). The inhibitory effects of venturicidin or DCCD on NADH-driven quenching were tested after incubation of membranes with inhibitor for 15 min at room temperature prior to addition of NADH. Rates of NADH oxidation were determined using 50 g/ml whole membranes and 50 M NADH in HMK buffer (25). All the values reported are from a single membrane preparation but are representative of several independent preparations.
Immunoblotting Analysis-Membrane vesicles were incubated in SDS sample buffer (62.5 mM Tris-HCl, pH 6.8, 2% SDS, 10% glycerol, 2.5% ␤-mercaptoethanol, and 0.01% bromphenol blue) at 100°C for 3 min. In the case of immunoblots for subunit c, the membrane sample was incubated at room temperature for 1 h since heat treatment promoted oligomerization. The solubilized membrane proteins were electrophoresed on a 12% polyacrylamide gel using the Tris-Tricine buffer described in Schagger and von-Jagow (26). After electrophoresis, proteins in the gel were electrophoretically transferred onto nitrocellulose paper (27). Rabbit antisera specific to subunit a (8), subunit b (28), and subunit c (29) were pretreated as described (8) and diluted 1:1000 prior to use. Immunoblots were developed using the "ECL System" (Amersham Corp.). F 0 , purified as described in Ref. 30, was used as a standard in the immunoblots.

Membrane Insertion of F 0 Subunits in the uncB Deletion
Strain-To reinvestigate the role of subunit a in the assembly of the F 0 sector, a strain was constructed that had a complete deletion of the chromosomal uncB gene. This was achieved in two steps. In the first step, the deletion was generated on a plasmid using the PCR protocol described in Fig. 1. The dele-  tion was then transferred from the plasmid into the chromosome using a positive selection for eviction of the sacRB-nptI cartridge inserted into the chromosomal uncB gene. The presence of the chromosomal uncB deletion in the resulting strain VF245 was confirmed by DNA sequencing. As shown in Fig. 3, immunoblots of VF245 membranes confirm the absence of subunit a (panel A, lane 3) and the presence of subunit b (panel B,  lanes 5 and 6) and subunit c (panel C, lanes 5 and 6). 3 The amounts of subunits b and c present may be slightly reduced relative to wild type.
Growth Characteristics of the uncB Deletion Strain-The ⌬uncB strain did not grow on succinate minimal medium, where a functional F 1 F 0 -ATP synthase is required. UncB ϩ transformants of strain VF245, which carry the wild type subunit a gene on a plasmid, were able to grow on succinate. The colony sizes after growth on succinate and growth yields in limiting glucose medium were studied to determine the relative function of F 1 F 0 -ATP synthase in vivo. The ⌬uncB strain showed low growth yields in glucose medium when compared with an isogenic wild type strain while the complemented strain gave growth yields similar to the wild type strain (Table I). The complemented strain also produced colonies of sizes similar to the wild type strain on succinate. We conclude that the uncB deletion within the unc operon (unc I ϩ ⌬BC ϩ F ϩ H ϩ A ϩ G ϩ D ϩ C ϩ ) does not have a major polar effect on expression of the downstream genes.
Biochemical Characterization of the uncB Deletion Strain-The level of membrane bound ATPase activity present in the ⌬uncB strain was about 50% of that observed in the isogenic wild type strain (Table II). The ATPase activity was activated to a similar extent by LDAO, as compared with the wild type, which indicates that the reduction in ATPase activity was not due to an abnormal inhibition of the membrane-bound form. Rather the reduction is probably due to a slight polar effect of the uncB deletion on downstream gene expression. The presence of membrane-bound ATPase in the uncB deletion strain indicates proper insertion of the b and c subunits into the membrane and assembly into a complex that is competent in binding F 1 . The UncB ϩ plasmid complemented strain showed ATPase activity similar to that in the wild type membranes. The ATPase activity of the ⌬uncB strain was DCCD insensitive, whereas the complemented strain had a DCCD sensitivity similar to wild type. The complemented strain showed normal ATP-driven proton translocation activity as indicated by quenching of ACMA fluorescence (Fig. 4).
Characterization of the aGlu-219 Mutants-The ⌬uncB strain was transformed with plasmids containing single Glu-219 mutations. Growth of the transformants was tested on succinate minimal medium. The E219D and the E219K mutants showed growth similar to wild type and the E219G mu-tant showed reduced growth relative to wild type (Table III). The E219A mutant was unable to grow on succinate after 3 days but pinpoint colonies were observed after incubation for 6 days. The growth yields in glucose liquid medium were also determined and correspond roughly with the colony size on succinate, i.e. wild type Ն E219D Ͼ E219K Ͼ E219G Ͼ E219A (Table III). Each of the Glu-219 mutants showed membranebound ATPase activity with a DCCD sensitivity similar to the wild type, indicating normal assembly of the F 0 sector in all the mutants (Table III). Immunoblots indicated the presence of roughly similar amounts of subunit a in the membranes of the VF245(⌬uncB) strains carrying the various Glu-219 mutant plasmids (Fig. 5). Each of the Glu-219 mutants tested also showed ATP-driven proton translocation as indicated by quenching of ACMA fluorescence (Fig. 6). E219D mutant membranes gave a quenching response similar to wild type membranes as has been previously demonstrated (9). The E219G mutant membranes consistently showed greater quenching than the E219K mutant membranes. E219A mutant membranes also showed significant ATP-driven quenching despite being derived from cells that could not grow on succinate. The cell culture used in preparation of these membranes was checked for succinate positive revertants and none were found. The NADH-driven quinacrine quenching response of each of the mutant membrane preparations was equivalent to that of the wild type preparation. The reduced ATP-driven quenching response therefore cannot be attributed to increased proton leakiness of the membrane.
Characterization of the aR210A Mutant-The ⌬uncB strain was transformed with plasmids carrying the R210A mutation or an UncB ϩ control plasmid. The R210A mutant was unable to grow on succinate and gave low growth yields on glucose (Table  IV). The R210A mutant showed very low levels of membrane bound ATPase (Table IV). The ATPase could be activated to wild type levels by LDAO treatment which indicated that F 1 was inhibited when bound to the R210A mutant F 0 . The R210A 3 The subunit a antiserum used here also detects a protein in the crude membrane fraction with a mobility greater than subunit a (panel A, lane 3). We have observed this immunoartifact in several other strains where the uncB gene has been disrupted and have concluded that it is not subunit a. The antiserum needs to be used with caution in studies of E. coli membranes.  mutant did not show any ATP-driven proton translocation (data not shown). The relative proton permeability of stripped membranes was determined by NADH-driven quenching of quinacrine fluorescence. Stripped membranes from the ⌬uncB strain showed a maximal NADH-driven quenching response which indicated that they were impermeable to protons (Fig. 7). Stripped membranes from the ⌬uncB strain transformed with the UncB ϩ plasmid pDF163 showed reduced NADH-driven quenching, which indicated that the membranes were proton permeable and contained an F 0 that was functional in proton translocation. Stripped membranes from the ⌬uncB strain transformed with plasmid pVF270, carrying the R210A mutation, showed the smallest NADH driven quenching response which suggested that the R210A mutant F 0 was fully functional in passive proton translocation. The proton permeability of the R210A stripped membranes could be inhibited by venturicidin or DCCD treatment and this confirmed that the proton leak took place through the mutant F 0 . The R210A F 0 is less sensitive to DCCD treatment than wild type F 0 . 4 The proton permeability of whole (unstripped) membranes was also tested. Whole membranes from the UncB ϩ complemented strain showed maximal NADH-driven quenching, indicating that the membranes were impermeable to protons whereas membranes from the R210A mutant showed very little NADHdriven quenching indicating that they were proton permeable (Fig. 8). Under these conditions, the rates of NADH oxidation were 0.58, 0.86, and 0.85 mol min Ϫ1 mg Ϫ1 for VF245(⌬uncB), VF245/pDF63(UncB ϩ ), and VF245(⌬uncB)/pVF270(aR210A) whole membranes, respectively. The reduced NADH-driven quenching response seen in the aR210A mutant membranes therefore cannot be attributed to diminished rates of NADH oxidation. We conclude that the passive proton pathway through the R210A mutant F 0 is fully functional and that this pathway is not blocked by the binding of F 1 .

DISCUSSION
The interdependence of insertion of the various F 0 subunits into the membrane was tested by Hermolin and Fillingame (8).
Subunits b and c were found to insert independently, whereas stable insertion of subunit a required the presence of both subunits b and c. These studies were carried out in a strain carrying a truncated subunit a with a W231stop mutation. The truncated subunit was not detected in immunoblots of membranes from this strain. The lack of the truncated subunit may be strain dependent as truncated subunits could be detected in the experiments of Eya et al. (31), who used the same antiserum. Recently, subunit a was shown to insert into the membrane in the absence of the other F 0 subunits where it is subject to degradation by the FtsH protease (32). In the experiments of Hermolin and Fillingame (8), it is possible that a truncated subunit a was inserted into the membrane and then degraded in the absence of subunits b and/or c. Furthermore, if the W231stop truncated subunit a was inserted transiently, then conclusions regarding the independence of insertion of subunits b and c would need to be re-evaluated. The study here of the ⌬uncB strain clearly indicates that subunit a is not required for the membrane insertion of subunits b and c or the assembly of the subunits into a F 0 like complex capable of binding F 1 .
Subunit a is required for the assembly of a functional F 0 and therefore the strain carrying the uncB deletion is unable to grow on succinate. The uncB deletion strain gives abnormally low growth yields on glucose for an unc mutant. A similar effect on glucose growth yield has been observed in an uncC mutant lacking the ⑀ subunit (33). The low growth yield in the case of the uncC mutant was thought to result from an increase in active ATPase in the cytoplasm due to the lack of the ⑀ subunit.
The ⌬uncB strain described here shows levels of cytoplasmic  ATPase similar to a wild type strain (data not shown) so excess, unregulated cytoplasmic ATPase is apparently not the explanation for the low growth yield. The effect on growth yield is clearly related to the loss of subunit a since close to wild type growth yields are observed when the uncB deletion is complemented by a wild type subunit a gene carried on a plasmid. Previous mutagenesis studies on Glu-219 of subunit a indi-cated that the residue was not tolerant to mutation. For example, the E219Q mutation led to almost complete loss of function (10), and the Glu-219 residue was concluded to be a likely participant in the proton translocation pathway. It was therefore surprising to find that the E219K mutant demonstrated such robust growth and function. The E219G mutation is also functional. A third mutant, the E219A mutant, did not form colonies on succinate during a prolonged incubation of 3 days, but did show significant ATP-driven quenching. We have no explanation for the imperfect correlation between growth on succinate and ATP-dependent quenching in these mutants. However, in summary, these results clearly indicate that Glu-219 cannot play a critical part in the proton translocation pathway. If the residue were a critical component of the proton translocation pathway, then mutations which alter or eliminate the charge on the side chain, i.e. E219K or E219G mutations, would be expected to be non-functional. Hartzog and Cain (34) have shown that the functionless H245G mutation is suppressed by substitution of Asp or Lys for Gly-218. Furthermore, the experiments of Hatch et al. (12) indicate that the functionless R210Q mutation can be suppressed by a Q252R substitution. These two sets of results suggest a close interaction between the final two transmembrane helices in subunit a, i.e. transmembrane helix-4 and helix-5 in the model of Hatch et al. (12), with Gly-218 opposite His-245 and Arg-210 opposite Gln-252. If Arg-210 and Glu-219 exist on a continuous stretch of ␣-helix, then these residues would be positioned on opposite faces of helix-4. The helices would also be oriented so that Arg-210 on helix-4 and Gln-252 on helix-5 are in structurally similar positions. This orientation would preclude any interaction between Glu-219 and Gln-252 as was suggested by Vik and Antonio (11). Residues adjacent to Gly-218 and Glu-219 have been identified by suppressor analysis as being important for the functional interaction between subunit a and subunit c (35). The substitutions at Glu-219 and His-245 which disrupt function may do so by perturbing interactions between the a and c subunits.
The aR210A mutation was initially described by Hatch et al. (12). They reported an abnormally low NADH-driven quenching response for mutant membranes and suggested that the mutant membranes were partially permeable to protons, although other possible explanations were not ruled out. The data reported by Hatch et al. (12) led us to re-investigate the properties of the aR210A substitution. Using the ⌬uncB complementation system developed here, we do observe significantly greater proton leakiness in the aR210A mutant membranes than was suggested by the work of Hatch et al. (12). Indeed, our experiments consistently indicate that aR210A mutant-stripped membranes have a greater proton permeability than wild type-stripped membranes. The proton conductance is mediated by F 0 as it is blocked by the specific inhibitors, venturicidin and DCCD. Higher concentrations of DCCD are required for inhibition of proton conductance by the aR210A membranes than with wild type membranes. This may indicate that the aR210A mutation directly effects the chemical reactivity of cAsp61, or that it alters the DCCD-binding site, although less direct effects are also possible. The proton leakiness of the aR210A mutant membranes indicates that Arg-210 is not essential for passive proton conductance through F 0 . Previously, several reports have suggested that Arg-210 may directly contribute to the proton conductance pathway, based upon the properties of the mutants studied (12,31,36,37). All other mutations in this residue (i.e. R 3 K, Q, E, V, I) were concluded to block passive proton translocation mediated by F 0 .
We have previously discussed a model for ATPase-coupled proton transport by F 1 F 0 -ATPase in which the H ϩ carrier, i.e.  Membranes were diluted to 0.25 mg/ml in HMK buffer containing 0.375 g/ml quinacrine. NADH was added to a final concentration of 50 M at the time indicated. Traces: 1, VF245(⌬uncB); 2, VF245/pDF163(UncB ϩ ); 3, VF245/pVF270(aR210A). the carboxyl side chain of Asp-61 of subunit c, undergoes a pK a change during the active transport cycle (2). The key aspect of this model, namely the change in pK a during ATP synthesis is now supported by kinetic studies (38). In the proposed model, the high pK a form of Asp-61 is proposed to facilitate the passive proton conductance of uncoupled F 0 , an idea also supported by kinetic evidence (39,40). The passive proton conductance through F 0 would normally be prevented by the binding of F 1 to F 0 . The passive proton translocation cycle is proposed to be mediated by a single subunit c, whereas the translocation of 3-4 H ϩ during coupled synthesis/hydrolysis of one ATP is proposed to be mediated by multiple subunit c. The model explicitly makes the passive H ϩ translocation cycle a side path that need not share kinetic characteristics with the H ϩ transport coupled ATP synthesis/hydrolysis cycle. In the model, we have proposed that the role of the aArg-210 may be to transiently lower the pK a of cAsp-61 during coupled H ϩ translocation. The lowering of pK a would be coupled to conformational changes traversing the entire F 1 F 0 complex which would ultimately lead to product release from the catalytic sites. The properties of the aR210A mutant can be rationalized by the model if, in the absence of the Arg side chain, subunit c is locked in a high pK a conformation where the Asp-61 carboxyl group has access to conducting channels on both sides of the membrane. As a result, the mutant cannot carry out coupled H ϩ translocation but would be unaffected in passive H ϩ transport. In this high pK a conformation, there are structural changes in the polar loop of subunit c which result in an uncoupling of F 1 from F 0 , i.e. generate a proton leaky membrane due to improper binding of F 1 to F 0 . The uncoupled phenotype observed here is also seen in mutations in subunit c mapping to the polar loop (18,41). The three key properties of the aR210A mutant, i.e. complete lack of ATPase-coupled H ϩ transport, normal passive F 0 -mediated H ϩ transport, and an uncoupled proton leak through F 1 F 0 , are consistent with this view.