Mutation of Alanine 24 to Serine in Subunit c of the Escherichia coli FIFo-ATP Synthase Reduces Reactivity of Aspartyl 61 with Dicyclohexylcarbodiimide *

Dicyclohexylcarbodiimide (DCCD) inhibits the activity of the FIFo-H+ ATP synthase of Eschiericia coli by reacting with aspartyl 61 in subunit c of the Fo sector to form a stable N-acylurea. The segment of chromosomal DNA which codes the subunits of the Fo was cloned from four independently isolated DCCDresistant mutants, and the sequence of the subunit c gene (uncE) was determined. An Alaz4 to serine (A24S) substitution was found in the subunit c gene of each mutant. The A24S uncE gene was cloned into the BamHI site of a mutant derivative of plasmid pBR322. The A24S subunit c conferred DCCD resistance to a variety of recipient E. coli strains when it was overexpressed from this plasmid. A 7-base pair deletion beginning at position 132 of the plasmid vector was responsible for the observed overexpression. Hoppe et al. (Hoppe, J., Schairer, H. U., and Sebald, W. (1980) Eur. J. Biochem. 112, 17-24) had previously shown that mutation of subunit c Ilez8 to threonine or valine resulted in DCCD resistance. The DCCD sensitivities of the membrane ATPase of these mutants and the A24S mutant were compared. DCCD sensitivity decreased in the order: wild-type >> I27V > I28T = A24S. The venturicidin sensitivities of wild-type and mutant membranes were also examined. The membrane ATPase of the I28T and I28V mutants was venturicidin resistant whereas the A24S substitution resulted in a hypersensitivity to inhibition by venturicidin. These results support a model in which subunit c folds in the membrane like a hairpin, where the region of residues 24-28 in transmembrane helix-1 is close to that of aspartyl 61 in transmembrane helix-2.

location through the membrane. When the two sectors are coupled, the enzyme functions as a reversible, H+-transporting ATPase or ATP synthase (Senior, 1988;Fillingame, 1990). The subunit composition of F1 and Fo in E. coli appears to be the simplest found in nature (Senior, 1988) wherein F1 is composed of five subunits in an a&ylG1el stoichiometry and Fo of three subunit types in a albpclo stoichiometry (Foster and Fillingame, 1982).
The membrane sector of mitochondria was originally named "Fo" (where the letter 0 subscript indicates oligomycin) to designate a detergent-solubilized preparation that restored oligomycin-sensitive ATPase activity when reconstituted with purified F1 (Racker, 1976). Dicyclohexylcarbodiimide (DCCD)' also inhibits the ATPase activity of FIFo by covalently binding to a site in FO (Beechey et al., 1967). DCCD binds covalently as the N-acylurea adduct of aspartyl 61 in a small, chloroform-methanol soluble protein of Fo termed proteolipid or subunit c (Cattell et al., 1971;Fillingame, 1975;Sebald and Hoppe, 1981). The isolation and analysis of DCCD-resistant mutants of E. coli helped to establish that ATPase inhibition was caused by the specific reaction of DCCD with subunit c (Fillingame, 1975(Fillingame, , 1980Friedl et al., 1977). Hoppe et al. (1980) sequenced the subunit c of six of the mutants isolated by Friedl et al. (1977) and found an Ilez8 to valine or threonine substitution in each mutant. To rationalize the effects of these substitutions on DCCD reactivity, subunit c was proposed to fold like a hairpin such that IleZs neighbors the DCCD-reactive Asp" in the membrane (Hoppe et al., 1980;Sebald and Hoppe, 1981;Fillingame, 1990). The amino acid substitutions in the four independently isolated DCCD-resistant mutants described by Fillingame (1975Fillingame ( , 1979 had not been examined. We report here that all four mutants contain Alaz4 +. Ser substitutions in subunit c. These substitutions support a hairpin model for subunit c in which both Ilez8 and Alaz4 must lie close to AspG1 in the membrane.
Growth of Cells-Minimal medium 63, LB medium, and 2YT medium are described in Miller (1972). M63 minimal medium agar plates were supplemented with 2 pg/ml thiamine hydrocholoride, 0.2 mM uracil, 0.2 mM L-arginine, and 40 p M 2,3 dihydrobenzoic acid as required, and 0.2% D-glucose was used as a carbon source. A combination of 0.6% disodium succinate hexahydrate, 0.2% potassium acetate, and 0.2% potassium L-malate was used as an alternative carbon source to glucose in plates used for complementation analysis or in the analysis of DCCD resistance. DCCD was added to the succinate-acetate-malate minimal agar medium at 55 "C in 0.01 volume of ethanol at a final concentration of 5 mM (Fillingame, 1975). Cells for a biochemical characterization were grown on liquid M63 minimal medium supplemented with succinate-acetate-malate as carbon sources and any of the required supplements as indicated above. Plasmid transformants were selected, and the plasmids were maintained by the addition of ampicillin to media at 100 pg/ml. Molecular Genetic Methods-Plasmid pBR322 and phages M113mp18 and M13mp19 are described in Maniatis et al. (1982).
Chromosomal DNA was isolated by the method of Silhavy et d . (1984), and the 4.3-kilobase HindIII fragment of unc-derived DNA was cloned into plasmid pBR322. The BamHI to HpaI fragment of uncE DNA was cloned between the BamHI and HincII sites of phages M13mp19 and M13mp19, and the DNA sequence of the uncE gene was determined as described previously by Mosher et al. (1985).
Biochemical Methods-Membranes were prepared with a French press in TMDG buffer (50 mM Tris-HC1, pH 7.5, 5 mM MgC12, 1 mM dithiothreitol, 10% (w/v) glycerol) supplemented with 6 mM p-aminobenzamidine HC1 (Hermolin et al., 1983). The methods of Hermolin and  were followed for purification of subunit c in chloroform-methanol and for anion exchange HPLC analysis of charge variants of subunit c.
Inhibitor Treatment and ATPase Assays-The method used for DCCD treatment is that described by Fillingame (1979). Membranes (5-12 pg of protein) were incubated in 0.9 ml of T r i s -M F assay buffer (55.5 mM Tris H2S0,, pH 7.8, 0.222 mM MgSOJ and treated with 0.01 ml of DCCD in ethanol for 20 min at 30 "C. The ATPase assay was then initiated by the addition of 0.1 ml of 4 mM [3ZP]ATP. After a 6-min incubation the reaction was terminated by the addition of trichloroacetic acid. The molybydate complex of [y-32P]phosphate was extracted into isobutyl alcohol-benzene (Fillingame, 1975). Under these assay conditions, 20-30% of the membrane APTase proves to be DCCD insensitive because of dissociation of Fl from Fo (Hermolin and Fraga and Fillingame, 1989).
To test venturicidin inhibition, membranes (125 pg of protein in 25 pl of TMDG buffer) were incubated in 0.95 ml of 10 mM Hepes-KOH, pH 7.5, 5 mM MgCl2, 300 mM KC1 and treated with 10 pl of venturicidin in dimethyl sulfoxide, or dimethyl sulfoxide only, for 10 min at 25 "C. The ATPase assay was then initiated by addition of 25 p1 of 40 mM [y-3zP]ATP. After 5 min, the assay was terminated by the addition of trichloroacetic acid. [32P]Phosphate was extracted as described above. The Hepes-MF-KCI assay buffer used here is that described by Paule and Fillingame (1989). Less than 10% of membrane-bound Fl dissociates from Fo under these incubation conditions.

Localization of Mutations
Conferring DCCD Resistance-Chromosomal DNA was isolated from four independently selected DCCD-resistant mutants and the DNA digested with HindIII. The HindIII fragment of chromosomal DNA containing the uncBEFHA genes (bases 870-5308)2 was cloned into the HindIII site of plasmid pBR322. Recombinant plasmid transformants containing the wild-type uncB gene were selected based upon complementation of the uncB402 chromosomal recipient strain "349.
To determine whether a mutation in the uncE gene was responsible for DCCD resistance, the BamHI(1727) to The nucleotide numbering used for the unc operon is that given by Walker et al. (1984). HpaI(2162) DNA fragment containing uncE was cloned from the plasmids described above into the BarnHI and HincII sites of phage M13mp18 and the uncE DNA sequenced. Each of the four cloned uncE genes showed a CGT(A1a) to TCT(Ser) change in the codon for residue 24 of subunit c and no other changes elsewhere in the coding sequence of the protein.
To prove that the Ala'" + Ser substitution was the mutation causing DCCD resistance, the BamHI(1727) to HpaI(2162) fragment of uncEIOI(A24S) DNA was subcloned from the larger plasmid described above into the BmHI and PuuII sites of plasmid pBR322. This plasmid (pMO101) only weakly complemented uncE recipient strains and did not confer DCCD-resistant growth. In contrast, plasmid pCP35 (Fillingame et al., 1984), which was thought to be equivalent to plasmid pMOlOl but carries the wild-type uncE gene, gave robust complementation with the same uncE recipient strains. A control plasmid was constructed in which the wild-type BamHI to HpaI DNA fragment was cloned into the BamHI and PuuII sites of plasmid pBR322, and, in contrast to plasmid pCP35, this plasmid (pM0100) gave weak complementation of uncE recipient strains. To determine whether the vector DNA was responsible for the differences in complementation, segments of the pCP35 and pMOlOl plasmids were exchanged. The plasmids were cut with BamHI and at the NdeI site (position 2296) in the pBR322 vector DNA (Fig.  1). The 655-base pair BamHI/NdeI fragment containing the uncE gene was separated from the 2449-base pair BarnHIl NdeI vector fragment by agarose gel electrophoresis, and the purified fragments were mixed and ligated. The product plasmid pM0351 contains the uncElO1 (A24S) gene in the pCP35 BamHIINdeI vector DNA. Plasmid pM0351 gave robust complementation when transformed into several uncE mutants (Table I) and conferred DCCD-resistant growth to both mutant and wild-type transformants. On the other hand, the plasmid containing the wild-type uncE gene in the pMO101derived vector DNA complemented poorly and confirmed that the differences in complementation efficacy were caused by the vector DNA.
The EcoRI(4361) to BamHI(375) segment of pCP35 vector DNA was sequenced. When compared with pBR322 DNA, the pCP35-derived DNA was found to have a 7-base pair deletion of nucleotides (T)132GGATGC(T)139. It is unclear from the DNA sequence whether T132 or T139 was deleted. The was constructed by ligating the 435-base pair BamHI/HpaI fragment of wild-type uncE DNA between the BamHI and PuuII sites of plasmid pBR322. The black area shows the uncE structural gene, and the hatched area is other unc DNA. The isolated pCP35 vector DNA was found to contain a 7-base pair deletion between bases 132 and 139 clockwise of the EcoRI site (A). The initiation site and direction of transcription from the tetracycline resistance promoter are shown. Plasmids pMOlOO and pMOlOl contain the BamHI/HpaI fragment of uncE' or uncEIOf DNA in a normal pBR322 vector. Plasmid pM0351 was constructed by ligating the BamHI + NdeI fragment of uncEIOl DNA from plasmid pMOlOl into the B Q~H I and NdeI sites of vector DNA from plasmid pCP35. bp, base pairs.  (2162)  ' unc DNA cloned between RamHI(375) and PuuII(2065) sites of plasmid pBR332. Scoring (+++, ++, +) indicates the extent of growth on succinate as carbon source; (?) indicates very weak to negligible growth; (-) indicates no growth. 0.20 0.6 "Cells were grown on LB-ampicillin medium as described under "Results." The culture of pM0101/MM994 'Size of colonies produced on 2YT-ampicillin plates after growth of liquid culture for subunit c purification.
which formed >95% small colonies was grown directly from small single colonies. deletion lies in the tetracycline resistance gene and conceivably could affect expression of the cloned uncE gene from the tetracycline resistance promoter (Stuber and Bujard, 1981).

Efficiency of Complementation Related to Expression of
Plasmid-coded Subunit c-It seemed possible that the difference in the complementation efficacy of the plasmids shown in Table I  which carries the uncElI4 (Q42E) mutation, was transformed with plasmid pMOlOl (uncEIOI, A24S) or plasmid pM0351 (uncEIOl, A24S) so that the amounts of chromosomal versus plasmid-borne products could be compared.
The Q42E mutation results in a charge difference that permits separation of the two proteins by anion exchange HPLC (Hermolin and . When pMOlOl and pM0351 transformant colonies of strain "994 [uncEIIC (Q42E)l were compared on 2YT-ampicillin-rich medium, the pMOlOlO transformants were distinctly smaller.3 In the process of growing cells to do this experiment, we noticed that cells from cultures of strain pM0101/MM994 eventually formed a mixture of large and small colonies on LB-ampicillin or 2YT-ampicillin plates. When a small single colony from a n LB-ampicillin plate was grown in 5 ml of LB-ampicillin medium overnight and 2.4 ml of that culture used to inoculate 1,200 ml of LB-ampicillin medium, the product cells formed small and large colonies in approximately a 50/50 ratio. A more homogeneous population of small colony-forming cells was obtained by scraping two small colonies from an agar plate and directly inoculating 200 ml of LB-ampicillin medium. The yield of subunit c from these pM0101/MM994 cultures is compared with that from a pM0351/MM994 culture in Table 11, and the proportions of plasmid-derived A24S protein to chromosomal-derived Q42E protein are indicated " The pMOlOl and pM0351 plasmid transformants of wild-type strain "833 (uric+) did not differ in size.
by the HPLC profiles of Fig. 2. In the culture producing uniformly small colonies, plasmid pMOlOl generated less subunit c than that produced by the chromosomal gene coding (chromosomal Q42E) were separated on an AX300 anion exchange HPLC column by elution with an ammonium acetate gradient in chloroformmethanol-water (4:4:1) solvent. The conditions are given in Hermolin and . Trace 1, plasmid pM0351-transformed strain "994; trace 2, plasmid pM0101-transformed strain "994 from culture producing 50% large and 50% small colonies (see "Experimental Procedures"); truce 3, plasmid pM0101-transformed strain "994 from culture producing >95% small colonies (see Table 11).
the Q42E protein. Considerably more subunit c was produced by the culture of large and small colony-producing cells, and this suggests that mutation of the plasmid led to greater A24S subunit c production and more vigorous growth because of biochemical complementation of the nonfunctional Q42E subunit c protein.
In these experiments, plasmid pM0351 produced four times as much subunit c as that produced by the chromosomal gene.4 Mixing of A24S and Q42E Subunits in FO Restores DCCD Sensitivity to ATPase-The uncElOl (A24S) mutant shows DCCD-resistant ATPase activity because of reduced reactivity of DCCD with AspG1 (Fillingame, 1975). The uncEll4 (Q42E) mutant exhibits DCCD-resistant ATPase activity because of uncoupling of Ft and Fo (Mosher et al., 1985).
Membranes from uncElOl plasmid transformants of uncE114 cells showed an intermediate sensitivity to DCCD, i.e. greater inhibition was observed than with A24S mutant membranes but less inhibition than with wild-type membranes (Fig. 3). This intermediate sensitivity occurs with membranes containing one Q42E subunit/8-9 A24S subunits.* Comparison of DCCD Resistance Conferred by the A24S, I28T, and I28V Mutations- Hoppe et al. (1980) had previously reported DCCD-resistant mutants with substitutions at Hez8 of subunit c. The relative DCCD sensitivity of the I28T, I28V, and A24S mutants was compared. The ATPase of I28T and A24S membranes was equally resistant to inhibition by DCCD (Fig. 4). The I28V membrane ATPase showed an intermediate sensitivity between that of the aforementioned mutants and that of wild type (Fig. 4). It is of interest that all three DCCD-resistant mutants had membrane ATPase activities that were elevated relative to wild type. An elevated membrane ATPase activity was also reported in previous studies of the A24S mutant (Fillingame, 1975;Fillingame and Wopat, 1978).
Venturicidin Resistance Is Conferred by the I28T and I28V Mutations but Not by the A24S Mutation-The venturicidin sensitivity of the ATPase of mutant and wild-type membrane When strain pM0351/MM994 was grown on a glucose minimal medium containing 5% LB, the ratio of plasmid to chromosomal product increased to 8-9, and the total yield of subunit c increased to approximately 5 pg/mg cellular protein. Despite the large overproduction of subunit c, only tiny amounts were found in the cytoplasmic fraction of transformant cells (<3% of that found in the membrane), i.e. the amount found in the cytoplasm did not exceed that of normal monoploid cells. Membranes were treated with venturicidin in Hepes-Mg2+-KCI assay buffer and assayed for ATPase activity as described under "Experimental Procedures." 0, wild-type membrane (strain "180);
is compared in Fig. 5. As originally demonstrated by Perlin et al. (1985), venturicidin is an effective inhibitor of the wildtype membrane ATPase. The I28T and I28V mutations resulted in venturicidin resistance, with the I28T membrane ATPase being most resistant. On the other hand, the A24S mutation led to a membrane ATPase that was hypersensitive to inhibition by venturicidin (Fig. 5 ) .

DISCUSSION
The uncE genes of four independently isolated, DCCDresistant mutants were sequenced in this study, and each of the four genes coded A24S substitutions in subunit c. In contrast, Hoppe et al. (1980) had sequenced the subunit c protein from six DCCD-resistant mutants, and all six proteins contained either I28V or I28T substitutions. The I28T and A24S mutation confer equivalent DCCD resistance as judged by growth of cells on succinate-acetate-malate in the presence of DCCD and by the resistance of the membrane ATPase. It is unclear why the two collections of mutants differ with respect to the mutated residue. The mutants isolated by Fillingame (1975Fillingame ( , 1979 were selected on a combined succinate-acetate-malate carbon source on plates containing 5 mM DCCD. The cells were mutagenized with N-methyl-N"nitronitrosoguanidine. The mutants isolated by Fried1 et al. of F1Fo-ATP Synthase (1977) were selected after N-methyl-N'-nitronitroguanidine mutagenesis on plates containing 24 mM DCCD and a succinate carbon source. The codon changes predicted for the I28V and I28T mutations are ATC to GTC and ACC, respectively. We have confirmed these substitutions by DNA sequence analysis of chromosomal DNA amplified by the polymerase chain reaction. These mutants were therefore generated by AT -GC transition mutations. Each of the A24S substitutions was caused by a GCT to TCT codon change, i.e. by GC + TA transversion mutations. It is of interest that the major and direct mode of MNNG mutagenesis is via GC + AT transitions (Glass, 1982) and that none of the mutants contain this type of substitition. GC + AT transitions would lead to silent changes in the IleZ8 codon but would cause threonine or valine substitutions in the Ala24 codon. One might predict that threonine or valine substitutions at position 24 will either have a deleterious effect on ATP synthase function or not confer DCCD resistance.
The finding that mutations in both Alaz4 and Ile28 affect reactivity of DCCD with AspG1 most simply suggests a binding site for DCCD between residues 24 and 28 in the first transmembrane helix of subunit c. The binding site must obviously be close enough to AspG1 in transmembrane helix-2 for DCCD to react with the aspartyl P-carboxyl group. The evidence that subunit c does fold like a hairpin with two transmembrane helices has been reviewed elsewhere (Miller $iet al., 1990;Fillingame, 1990). The region surrounding residues 24-28 is highly conserved in all subunit c (see Fig. 6), and we have suggested that this region may interact with the region around Asp61 during the proton transfer cycle (Miller et al., 1990;Fillingame, 1990). The Gly23 and GlyZ7 residues of transmembrane helix-1 are absolutely conserved, and it is of interest that these glycines immediately precede the two positions giving rise to DCCD resistance. In addition, glycine is usually Gly Leu Thr Gly Ala Gly Ile Gly Ile Gly A. nidulans mito. Gly Leu Gly Gly Ala Gly lle Gly Thr Gly Bovine mito.
Gly Val Ala Gly Ser Gly Ala Gly lle Gly Consensus Ala/Gly X X Gly X Gly X Gly X Gly FIG. 6. Sequence comparisons for the conserved glycinerich segment of transmembrane helix-1 of subunit c from different species. The amino acid numbering system is for the E. coli protein. Residues that are identical to those in E. coli are printed in boldface. The sequences are from the following sources. E. coli, Streptococcus faecalis, Mastigocladus laminosus, Acido caldarius, thermophilic bacterium PS-3, R. rubrum, spinach chloroplast, Saccharomyces cerevisiae mitochondria, Neurospora crassa mitochondria, Aspergillus nidulans mitochondria, and bovine mitochondria: Hoppe and Sebald, 1984;Synechococus 6301: Cozens and Walker, 1987; tobacco mitochondria and Zea mays mitochondria: Bland et al., 1986.

PS-3
Gly Ala Leu Gly Ala Gly lle Gly Asn Gly found at positions 25 and 29. The high frequency of conserved glycine suggests that a compact structure may be necessary. The region is also very hydrophobic. The introduction of a serine or threonine residue could obviously disrupt DCCD binding by decreasing the hydrophobicity of the binding pocket. The I28V mutation is more difficult to rationalize but may again indicate the necessity for a precisely packed, compact structure in this region. Alaz4 and Ile2' are clearly important for DCCD binding in E. coli subunit c but are unlikely to be the only determinants of binding specificity. Several species of subunit c have serine at the equivalent of position 24, and others have valine or threonine at the equivalent of position 28 (Fig. 6). These changes by themselves do not lead to DCCD resistance, e.g. both bovine mitochondria with the equivalent of Ser24 and Rhodospirillum rubrum with the equivalents of Ser24 and Val2' are DCCD sensitive (Beechey et al., 1967;Bengis-Garber and Gromet-Elhanan, 1979). In addition, the binding site must have components that discriminate between structurally different carbodiimides. This was indicated by the studies of Abrams and Baron (1970) in which a specificity for nonpolar carbodiimides was first shown. However, the inhibitory potence of various carbodiimides does not appear to be a simple function of their hydrophobicity. The relative inhibitory potence of apolar carbodiimides does vary widely (Pate1 and Kaback, 1976;Beechey and Knight, 1978;Hoppe and Sebald, 1984) and seems to differ somewhat depending upon the species of FIFo examined. These observations are most easily explained by a very specific or rigidly shaped binding site within subunit c.
Interpretation of the carbodiimide inhibition experiments is complicated by the complexity of the reaction. DCCD must first react with the carboxylate form of Asp6' to yield on 0acylisourea and the adduct then rearrange to form the stable N-acylurea product (Williams and Ibrahim, 1981): Reaction 1 might only occur after DCCD was bound at its binding site. Alternatively, reaction 1 might occur between an Asp61 carboxylate group that was exposed to the lipid bilayer and DCCD that was simply dissolved in that lipid bilayer. The carbodiimide binding site might then function in promoting the acyl migration occurring in reaction 2. Binding might simply prevent the competing hydrolysis reaction of the 0-acylisourea to dicyclohexylurea that would regenerate a free carboxyl group. Perlin et al. (1985) have shown that venturicidin inhibits the activity of E. coli FIFo-ATPase and that its binding reduces the reaction of DCCD with subunit c. Similar results are seen with both venturicidin and oligomycin in mitochondria (Enns and Criddle, 1977;Kiehl and Hatefi, 1980), and the binding sites for all three compounds are thought to overlap. Oligomycin is a very weak inhibitor of the E. coli FIFo-ATPase (Perlin et al., 1985). A number of substitutions in the mitochondrial subunit c result in oligomycin and venturicidin resistance in yeast and Neurospora (Sebald and Hoppe, 1981;Galanis et al., 1989), and the sites leading to resistance overlap. Interestingly, venturicidin resistance in yeast results from mutations in residues corresponding to position 25, 27, and 29 of the E. coli protein (Galanis et al., 1989). The yeast venturicidin resistance site on transmembrane helix-1 thus surrounds the I28T venturicidin resistance site found in the E. coli protein. A second venturicidin resistance site in yeast lies in the equivalent of residues 58 and 59 of E. coli subunit c. The binding site thus seems to span or lie between the two membrane-traversing helices. It will be of interest to see if mutation of residues causing venturicidin resistance in yeast will lead to a venturicidin-resistance for the E. coli F,Fo-ATPase.
Our unpublished NMR studies on isolated subunit c in chloroform-methanol solvent indicate an extended a-helix from Asp7 through L~u~' .~ The structure of the purified subunit in chloroform-methanol solution must resemble that found in the membrane since the I28T mutation reduces the reactivity of AspG1 with DCCD even when the protein is purified. The conserved GlyZ3 and GlyZ7 residues would lie on one face of this a-helical structure and Gl~(Ala)'~ and Gly(A1a)" on the opposite face of the same helix. The glycine residues are separated by bulky, hydrophobic residues and would form a screwlike groove around the helix of the protein.
The groove could accommodate the packing of bulky side chains from another a-helix. During the course of proton translocation and the consequent conformational changes leading to ATP synthesis (Fillingame, 1990), one could easily envision a rotation of helices relative to each other. For example, in one state the ValG0 and Met57 side chains of one subunit c could lie in the groove formed by GlyZ3 and GlyZ7 in the same, or a second, subunit c and in the alternate state rotate up the helix to the part of the groove formed by Alaz5 and GlyZ9.
It is of interest that the substitutions at positions 24 and 28 which cause DCCD resistance also activate the ATPase activity of membrane-bound F1 (Fig. 4). The small conformational changes that result from the DCCD resistance mutations in this region of subunit c must ultimately then promote conformational changes at the catalytic site of F,.
Mutations in AspG1 on the opposite helix also alter the properties at the catalytic site (Fillingame et al., 1984). These changes in conformation may obviously be related to those required for the coupling of proton translocation to ATP synthesis.