A point mutation in the ATP synthase of Rhodobacter capsulatus results in differential contributions of DpH and Du in driving the ATP synthesis reaction Paola Turina and B. Andrea Melandri Department of Biology, Laboratory of Biochemistry and Biophysics, University of Bologna, Italy The interface between the c-subunit oligomer and the a subunit i n the F 0 sector of the ATP synthase is believed to form the core o f the rotating motor powered by the protonic flow. Besides the essential cAsp61 and aArg210 residues (Escherichia coli numbering), a few other residues at this interface, although nonessential, show a h igh degree of conservation, among these aGlu219. The homologous resi- due aGlu210 in the ATP s ynthase o f t he photosynthetic bacterium Rhodobacter capsulatus has been substituted by a lysine. Inner membranes prepared from the mutant strain showed approximately half of the ATP synthesis activity when driven both by light and by a cid-base transitions. As estimated with the ACMA assay, proton pumping rates in the i nner membranes were also reduced to a similar extent in the mutant. The most striking impairment of ATP synthesis in the mutant, a decrease as low as 12 times as compared to the wild-type, w as observed in the absence of a transmem- brane e lectrical m embrane potential (Du)atlowtrans- membrane pH difference ( DpH). Therefore, the mutation seems t o affect both the mechanism responsible for coupling F 1 with proton translocation by F 0 , and the mechanism determining the relative contribution of DpH and Du in driving ATP synthesis. Keywords: ATP synthase; mutagenesis; Rhodobacter cap- sulatus; DpH; Du. Membrane-bound F 0 F 1 -ATPases (ATP synthases) catalyze ATP synthesis in bacteria, chloroplasts and mitochondria at the expenses of an electrochemical potential gradient of protons (or Na + ions in so me species). The membrane- embedded h ydrophobic F 0 sector is involved in proton translocation across the membrane, and the hydrophilic F 1 sector contains the catalytic sites (reviewed in [1–3]). A wealth of high resolution structural information f or the soluble part has appeared since the first crystal structure of the mitochondrial F 1 was reported in 1994 [4], paralleled by an increasing amount of experimental evidence supporting a rotational mechanism of catalysis (reviewed in [5]). In the most investigated Escherichia c oli enzyme, F 1 consists of five types of subunits in stoichiometry a 3 b 3 cde and F 0 consists of three types of subunits in stoichiometry ab 2 c 9)12 .Thec subunit monomers span the membrane as a hairpin of two a helices [6] and are arranged in a oligomer in the form of a ring (see, for example, the crystallographic evidence in [7]). Subunit a most likely consists of five transmembrane helices [8–10], the fourth of which has been shown by extensive cross-linking analysis to pack against the second transmembrane segment of subunit c [11]. The fourth and fifth transmembrane helices, residues 206–271, house the most conserved regions of the subunit. In view of the ATP-driven rotation of the c-and e-subunit shaft within the a 3 b 3 subunit barrel in F 1 ,itis proposed that the c subunit ring in F 0 , which is connected to the ce shaft [12–14], r otates against t he a subunit, which is connected to the a 3 b 3 barrel through the b and d subunits [15,16]. Experimental evidence consistent with this idea has been presented [17–19]. A few mechanistic models for torque generation in F 0 have been proposed, which emphasize the role of electro- static interactions [2 0–22] or the r ole of conf ormational changes within the c subunit [23]. All models include a central role for the essential carboxyl group of the c subunit and for the essential Arg residue in the a subunit (cAsp61 and aArg210, respectively, in E. coli). Besides the cAsp61/aArg210 couple in the middle of t he membra ne, the remai ning a/c interface regions are believed to form the a ccess pathways for protons. P robably lining the acidic access pathway is residue aGlu219, based on cross- linking data [11]. Several lines of evidence support a close spatial and functional interaction between aG lu219 and aHis245, including the f act that in the ATP synthases of mitochondria and of photosynthetic bacteria the position of these two amino acids in the primary sequence are inverted [24], the fact that the E. coli double mutant aGlu219 fi His/aHis245 fi Glu has an ATP synthase activity signifi- cantly higher than that of either of the single mutation strains [25], and their close position in the proposed topological models [8–10]. Although t hese residues were shown to be nonessential by extensive mutagenic analysis Correspondence to B. A. Melandri, Laboratory of Biochemistry and Biophysics, Department of Biology, University of Bologna, Via Irnerio, 42, I-40126 Bologna, Italy. Fax: + 39 051 242576, Tel.: + 39 051 2091293, E-mail: melandri@alma.unibo.it Abbreviations: GTA, gene transfer agent; Bchl, bacteriochlorophyll; ACMA, 9-amino-6-chloro-2-methoxyacridine; RC, photosynthetic reaction center; Á ~ ll H þ , transmembrane difference of electrochemical potential of protons; Du, bulk-to-bulk transmembrane electrical potential difference; Dw, surface electrical potential difference. Enzyme: ATP synthase (EC 3.6.3.14). Note: a website is available at http://www.biologia.unibo.it/ (Received 12 November 2001, revised 21 February 2002, accepted 21 February 2002) Eur. J. Biochem. 269, 1984–1992 (2002) Ó FEBS 2002 doi:10.1046/j.1432-1033.2002.02843.x [25], their impo rtant functional role is indicated both by their high degree o f conservation a nd by the d eleterious effects of their mutations on the E. coli ATP syn thase. In this work, the photosynthetic bacterium Rhodobacter capsulatus has been used as a convenient system for genetic manipulation and for investigating catalytic prop- erties of the ATP synthase, as a variety of functional studies can be carried out in its well-coupled inner membrane preparations (chromatophores). The subunit composition of this ATP synthase is very similar to the E. coli enzyme [26,27], except that the additional subunit b¢, homologous to b, is found in F 0 ,asitistypicalfor photosynthetic organisms. As in other photosynthetic bacteria, in Rb. capsulatus an inverted mitochondrial-like arrangement for the residue pair aGlu219/aHis245 is found. The corresponding residues in the Rb. capsulatus are aHis173 and aGlu210 [27]. We have introduced the single point mutation aGlu210 fi Lys and have investigated detailed functional aspects of the ATP synthase in native membranes. The corresponding mutation in the E. coli enzyme, aGlu219 fi Lys, was examined i n two previous studies [28,29], where it was shown to cause reduced cell growth and reduced proton pumping by an ATP hydrolytic activity similar to wild-type. The r esults of the present work are consistent with the data obtained in E. coli and reveal novel functional properties of the mutated enzyme. EXPERIMENTAL PROCEDURES Bacterial strains, growth conditions, membrane preparations Rb. c apsulatus strain B100 is wild-type strain B10 cured of phages. A s pontaneous rifampicin-resistant o f B100, J1, was used in the GTA 1 procedures. Rb. capsulatus strains were grown photoeterotrophically on a synthetic medium containing malate as a carbon source [30]; kanamycin and tetracycline were added at 25 and 2 lgÆmL )1 , respectively. Cultures were illuminated by two opposite panels e ach carrying nine 100-W i ncandescent light bulbs; excessive warming was prevented by water cooling. The mutant and wild-type strain we re grown in p arallel and cells were harvested at D 600 ¼ 1.2–1.4. Intra-cytoplasmic mem- branes (chromatophores) were prepared according to the method described previously [30], r esuspended in 50 m M glycyl-glycine/NaOH, 5 m M MgCl 2 , pH 7.5, rapidly frozen as small d roplets in liquid n itrogen, and stored a t )80 °C. Subunit a mutagenesis The wild-type copy of subunit a was c arried by the pFo16 plasmid which contained the whole atp2 operon of Rb. c apsulatus cloned into the pTZ19U plasmid [27]. The aE210 fi K mutation was introduced into this plasmid by using the QuickChange Site-Directed Mutagenesis K it (Stratagene), based on linear PCR, using the following mutagenic oligonucleotides: 5¢-CGCGATGTATGCGC TC AAGATCCTCGTGGCC-3¢ and 5¢-GGCCACGAGG AT CTTGAGCGCATACATCGCG-3¢. The mutation introduced an additional restriction site for XhoII, therefore its p resence was confirmed by restriction site analysis. The mutated atp2 operon was then cloned into the broad-host- range plasmid pRK415 [31] carrying the tetracycline resistance, as described previously [27]. The new plasmid, pKFo102, was subsequently introduced into Rb. capsulatus B100 strain by triparental conjugation [32]. The wild-type chromosomal copy of the atp2 operon was deleted by taking advantage of the so-called GTA, bacteriophage-like parti- cles produced by Rb. capsulatus [33] as described previously [26]. GTA particles are produced by Rb. c apsulatus cells which pack, randomly, pieces of DNA about 4.6-kb long, either from the chromosome o r from resident p lasmids of donor cells, and transfer them to acceptor cells, where they are integrated into the chromosome by homologous recombination. This results in a n exchange between the incoming DNA and the corresponding chromosomal DNA which is lost in the process. In this work, the GTA exchange donor was the J1 strain carrying the p Fo39 plasmid. This plasmid was a gift from R. Borghese in our department and had been constructed by inserting the kanamycin resistence cassette of Tn903 in place of the atpBEXF g enes (subunits a, c, b¢, b) leaving only a C-terminal truncated atpI gene (subunit i). Af ter GTA transfer, kanamycin-resistant colo- nies appeared that could contain the kanamycin resistance cassette (i.e. the F 0 deletion) either on the chromosomes or on the pKFo102 plasmid. Restriction analysis of the plasmid isolated from such colonies allowed the selection of those carrying the F 0 deletion on the chromosomes. Two mutated s trains were selected, w hich carried mutated plasmids originated from two different PCR runs. A pseudo-wild-type strain was constructed in parallel, which carried the wild-type F 0 operon on the plasmid and the deletion of the chromosomal F 0 operon. The cells used for chromatophores preparations were routinely checked for the presence of the mutation by XhoII restriction analysis of the resident plasmid. Western blot The amount of ATP synthase in the membrane was evaluated by quantitative Western blot on SD S/PAGE isolated chromatophores protein, using a yeast anti-(b subunit) antiserum (kindly provided by J. Velours, Bor- deaux), the luminol assay for detection, and a purified ATP synthase (isolated from Rb. capsulatus as described previ- ously [34]) as a standard. The amounts of chromatophores and standard protein in the different lanes of a single gel were kept in the linear range of the luminol assay response. Light-induced ATP synthesis Light-driven ATP synthesis was carried out at 30 °Cin the following buffer: 100 m M glycyl-glycine/NaOH, pH 7.7 5, 50 m M KCl, 10 m M Mg-acetate, 0 .1% bovine serum albumin, 5 m M P i ,0.2m M succinic acid, 0.8 m M AMP, 10 l M Bchl. The chromatophores suspension was illuminated from two opposite sides by two 100 W incandescent bulbs. The reaction was started by addition of 100 l M ADP. After s topping the r eaction with 7 % perchloric acid, t he ATP c oncentration i n each s ample was measured in a luminometer (LKB 1 250) with the ATP-Monitoring Kit ( Bioorbit). The small amount of ATP synthesized in the dark (due to the adenylate kinase reaction) was subtracted. For the experiments reported in Ó FEBS 2002 The aE210K mutation in Rb. capsulatus ATP synthase (Eur. J. Biochem. 269) 1985 Fig. 1, the cuvette contained 2 0 l M ADP, 6 0 l M lucif- erine and 2–10 lgÆmL )1 purified luciferase (32–160 · 10 3 light units ÆmL )1 ) from Sigma (L9009) in the reaction mixture described above, and the luminescence was detected in real-time a t room temperature essentially as described previously [35]. The assay mixture was illumi- nated by a halogen lamp (160 WÆm )2 light intensity, filtered through 1 cm water and two layers of 8 8 A Wratten fi lters) and different illumination times were determined by an electronic shutter controlled by a Uniblitz T132 Driver . The photomultiplier was shielded against actinic light by a copper sulfate solution. The amount of sy nthetized ATP w as evaluated by a dding 10–25 n M standard ATP. ATP synthesis induced by acid-base transitions Acid-base driven ATP synthesis was carried out similarly as described in [ 36] by rapidly injecting an acidified chromatophores suspension into a l uminometer cuvette containing a basic solution and monitoring the ATP concentration c ontinuously with luciferine/luciferase i n the luminometer. The chromatophores were first resuspended in 5 m M P i /NaOH, pH 7.3, 2 m M MgCl 2 ,1m M AMP, 5 l M valinomycin, 10% sucr ose, and either 1 m M KCl (+Du)or100m M KCl (–Du) and left incubating in this medium  1 h at room temperature. This time was enough to equilibrate the K + concentration across the membrane, as judged by the d isappearance of the K + /valinomycin diffusion potential (monitored by the carotenoid shift signal) i nduced by the initial K + gradient [36]. The chromatophores were then mixed with the acidic solution [30 m M succinic acid/NaOH, pH 4.6–6.5, 2 m M MgCl 2 , 5m M P i ,either1m M (+Du) or 100 m M KCl (–Du), 1 m M AMP, 5 l M valinomycin] and incubated at room tempera- ture at variable t imes between 2 and 30 min, depending on the pH of t he suspending buffer, prior t o injection of 100 lL i nto the basic solution. This latter contained 850 lL o f basic solution so that the final concentrations after chromatophores addition would be 200 m M Tricine, 2m M MgCl 2 ,5m M P i ,either150m M KOH + 30 m M NaOH (+Du)or100m M KOH + 80 m M NaOH (–Du), pH 8.65, 1 m M AMP, 100 l M ADP, 50 lLoftheATP- Monitoring Kit and 5 lgÆmL )1 of purified luciferase (80 · 10 3 light units ÆmL )1 ). The final Bchl c oncentration varied between 1 and 8 l M . The ATP concentration was evaluated by adding 100–200 n M standard ATP in each cuvette. The basic solution was thermostated so that the ATP synthesis reaction took place a t 13 °C. The p H measured after mixing the chroma tophores with the acidic solution was taken as the internal pH, the pH measured after mixing the acidified chromatophores with the basic solution (8.55 ± 0.05) was taken as the external pH. Their difference is the indicated DpH. F or the lowest DpH differences the Ôacidic Õ solution contained 20 m M Tricine instead of succinic acid. Assuming complete equilibration of the K + during the 1 h preincubation (see above), the value of t he K + /valinomycin diffusion potential deter- mined by the K + transmembrane concentration difference during the acid-base transition can be approximated, ( on the basis of the Nernst equation or the Goldman equation for monovalent ions). For T ¼ 12 °C, Du ¼ 124 mV or Du ¼ 0 mV are obtained b ased on the Nernst equation, Du ¼ 75 mV or Du ¼ 0.4 mV results by applying the Goldman equation as described previously [36], for [K + ] in /[K + ] out ¼ 1/150 m M and [ K + ] in /[K + ] out ¼ 100/100 m M , respectively. ACMA assay ACMA fluorescence quenching assay was carried out in a Jasco FP 500 spectrofluorimeter (wavelen gth 412 and 482 nm for excitation and emission, respectively) at 1 5 °C in the following mixture: 20 m M Tricine/KOH, pH 8.0, 50 m M KCl, 0.5 m M MgCl 2 ,0.2m M succinic acid , 5 l M Antimycin, 0.2 l M myxothiazol, 2 l M valinomycin, 1.5 l M ACMA, 20 l M Bchl, 400 l M ATP. The response of ACMA to DpH was empirically calibrated using artificially induced protonic g radients, established by HCl and NaOH a ddi- tions, under similar temperature and buffer c onditions, except for the presence of 20 m M succinic acid, as described previously [37]. Fig. 1. ATP s ynthesis in c hromatophores i nduced by light. Light- induced ATP synthesis in the absence (A) or in the presence of 2 l M valimomycin (B). The increase in luminescence associated with ATP production follow ing a short period of illumination has been mon- itored. The reactio n assays, describe d in the Experime ntal proce dures section, contained luciferase ( 32–160 · 10 3 light unitsÆmL )1 )and luciferine (60 l M ) as reported previously [35]. ADP and Bchl concen- trations were 20 l M and 10 l M , respectively. The temperature was 28 °C. (d,m) wild-type and (s,n) mutant chromatophores. (B) The data points h ave been fitted with a n arbitrary function. The d ashed lines extrapolate the linear part of the curves. (C) The first derivatives of the fi tting functions were calculated b etween 0 a nd 1000 ms (extrapolating the fittin g function when necessary). The ratios of the wild-type derivatives over the mutant, indicating the ratios of the ATP synthesis rates, are plotted for the exp eriments in the absence (dashed line) or in the presence (continu ous line) of valinomycin. As n o ATP synthesis could be detected with added valinomycin for short illumin- ation times, the rat io function was truncated for times £ 150 ms. 1986 P. Turina and B. A. Melandri (Eur. J. Biochem. 269) Ó FEBS 2002 ATP hydrolysis assays ATP hydrolysis was measured routinely at 30 °Cinthe following buffer: 20 m M Tricine/KOH, pH 8.0, 50 m M KCl, 2 m M MgCl 2 ,0.2m M succinic acid, 20 l M Bchl. Th e reaction was started by adding 1 m M ATP. After stopping the reaction at different time s with 5% trichloroacetic acid, the P i concentration was measured by molybdate colori- metric assay as described previously [38]. For more sensitive measurements, the re leased P i was measu red w ith t he EnzCheck Phosphate Assay Kit (Molecular Probes) or with the malachite green assay [39], both methods giving similar results. Other methods Bchl concentration was measured in acetone/methanol extract [40]. The protein concentration of c hromatophores was determined using the BCA assay (Pierce). An aliquot of the c hromatophores was e xtracted with acetone/meth- anol (7 : 2 , v/v). After centrifugation, the p rotein pellet was dissolved in 0.1 M NaOH/1% SDS for determination of protein concentration. The light-induced transmem- brane electric potential difference was evaluated following the electrochromic signal of endogenous carotenoid [41]. The c oncentration of photo-oxidizable reaction centre (RC) and of total photo-oxidizable cytochrome (c 1 + c 2 ) were measured as described previously [42], following trains of closely spaced flashes. The amount of phosp- holipid was determined by t he method described previ- ously [43]. RESULTS The single point mutation aE210K was introduced into the atp2 operon o f Rb. capsulatus, c ontaining the F 0 genes, cloned in an E. coli strain. The mutated operon was then transferred into a broad-host-range vector and the resulting plasmid introduced by conjugation into Rb. capsulatus wild-type cells. Finally, a GTA transfer was allowe d to take place , wh ich generated the deletion o f the chromoso- mal atp2 operon by substitution with a kanamycin resistance cassette. Therefore, the resulting strain carried the deletion of the chromosomal atp2 operon and several copies of a plasmid carrying the mutated atp2 operon. As a control, a parallel at p2-deleted strain w as created, in which the r esident p lasmid carried the w ild-type operon. This pseudo-wild-type strain is referred to as wild-type i n the following procedures. Characterization of mutant chromatophores Phototrophic g rowth of the aE210K mutant cells was slower than the wild-type cells. Accordingly, the light- induced ATP synthesis r ate catalyzed by the mutant chromatophores was about 40% lower on a Bchl basis than the rate c atalyzed by wild-type c hromatophores (Table 1). T he same reduction res ulted also when the ADP concentration was varied between 20 and 500 l M , indicating that the mutation does not affect the apparent K m for ADP. In contrast, no significant difference could be observed in t he ATP hydrolysis r ate. The concentration of ATP synthase was estimated on a Bchl basis by quantitative Western blot analysis and was found to be the same within experimental error for both mutant and wild-type chro- matophores. The specific activity of ATP s ynth esis was 8±2and13±3ATP/(F 1 F 0 Æs) for mutant and wild-type, respectively, whereas the specific ATP hydrolysis activities were 2.0 ± 0.7 and 2.3 ± 1.0 ATP/(F 1 F 0 Æs). These values are from a single preparation of chromatophores but are representative of several preparations, obtained from strains carrying mutated plasmids originating from two different PCR runs. The lower ATP synthesis rate of the mutant was not due to a lower efficiency of the electron transport chain or higher permeability of the membrane as the ATP synthesis rate induced by acid-base transitions was similarly reduced (see below and Fig. 3A). The possibility of a higher percentage of open, and therefore uncoupled, membrane fragments o r of r ight-side-out vesicles in the mutant was also r uled out by measuring t he extent of flash-induced carotenoid shift and the extent of cyto- chrome c accessible photo-o xidizable RC, which were comparable to those of w ild-type chromatophores (not shown). The mutant and wild-type preparations were further characterized as to their RC, phospholipid and protein content ( Table 1 ). The ATP synthase/Bchl ratio, the size of the a ntenna (Bchl/RC) and the phospholipid content were very similar in both strains. The most striking difference was found in the protein content, which w as approximately 1.6-fold lower in the mutant chromatophores on a B chl basis. It is possible that this difference affects the adsorption of ACMA to the m embrane and therefore t he probe response to DpH (see below). Table 1. Catalytic activities and composition of chromatophores from wild-type and mutant cells. All values reported are from a s ingle chromatophores preparation but are representative of several different preparations. Wild-type Mutant aE210 fi K Photophosphorylation rate a 74 ± 4 44 ± 3 (mM ATP/M Bchl/s) ATP Hydrolysis Rate b 13 ± 3 11 ± 2 (mM P i /M Bchl/s) Bchl/ATP Synthase Ratio c 178 ± 35 180 ± 31 (moles/mole) Protein/Bchl 48.4 29.7 (mg/lmole) Bchl/RC 71 89 (moles/mole) phospholipids/Bchl 13 ± 1.8 12 ± 2.0 (moles/mole) a The ATP synthesized at each illumination time (n ¼ 5) was measured after denaturation in a luminometer with the ATP Monitoring Kit as described in the Experimental procedures. b The P i released at each time point (n ¼ 5) was analyzed after dena- turation by the colorimetric assay described previously [38]. c After SDS/PAGE and transfer onto nitrocellulose paper of chromato- phores and known amounts of isolated Rb. capsulatus ATP synthase, the unknown amount of ATP synthase was evaluated by detecting the b subunit with an anti-(b subunit) Ig in a linear range of response. Ó FEBS 2002 The aE210K mutation in Rb. capsulatus ATP synthase (Eur. J. Biochem. 269) 1987 Light-induced ATP synthesis at low D ~ ll H þ The ATP synthesis rates reported in Table 1 were measured under light close to saturation at times ranging from 2 to 10 s. Under these conditions, a high steady-state D ~ ll H þ was obtained, which resulted in a linear increase of the ATP yields with illumination time. In order to investigate the activity of the mutated enzyme at lower D ~ ll H þ values, shorter illumination times were chosen (from 100 ms to 2 s ). The assays were also supplemented with the ionophore valino- mycin, which largely prevents the onset of the electrical component of D ~ ll H þ , thus further reducing its total extent during short illumination times. Due to the low ATP yields expected in these experiments, the luciferine/luciferase ATP detection system was added into t he assay cuvette at high concentration (32–160 · 10 3 light units ÆmL )1 of luciferase), so that the ATP-induced luminescence was directly detected. Figure 1A shows the ATP yields obtained from m utant and wild-type chromatophores as a function of illumination time in the absence of valinomycin. A linear relationship was observed under these conditions, and the ATP yield by the mutant chromatophores was  50% of that by the wild- type at every time point. W hen the assay c ontained valinomycin, the data reported in Fig. 1 B were obtained. In this case, the ATP yield of both wild-type and mutant as a function of illumination time presented a lag phase before becoming linear. A similar lag phase had a lready been observed in wild-type c hromatophores [35] and c an be interpreted as being due to a lag phase in the onset of D ~ ll H þ when its e lectrical c omponent is being dissipated b y valinomycin. Strikingly, this lag phase was much more pronounced in thecaseofthemutant.Thelargedifferencerelativetothe wild-type can be best appreciated by taking the derivative of the fi tting f unctions of the mutant and wild-type data a nd plotting their ratio, as in Fig. 1C. This derivative represents the r ate o f ATP synthesis at each time point. The twofold difference between the mutant and wild-type observed in the presence of a Du is approached o nly for illumination times longer than 1 s, whereas between 200 and 500 ms the ATP synthesis rate catalyzed by the mutant is much decreased, up to 12-fold lower with respect to the wild-type. Clearly, the effect of the mutation is much larger in the absence of a significant Du and at low D ~ ll H þ values. ATP synthesis induced by acid-base transitions The functioning of the mutated enzyme was also studied by using the technique of acid-base transitions. This technique allows one to control the extent of DpH across vesicle Fig. 2. ATP synthesis driven by acid-base transitions. Chromatophores were preincubated in resuspending and acidic media as described in the Experimental procedures, and injected into the basic medium as indicated by the arrow. The ATP synthesis following chromatophores injection was mo nitored continuously with t he luciferine/luciferase ATP Monitoring Kit in a luminometer. The high signal-to-noise ratio was obtained due to ad ded purifie d luciferase (80 · 10 3 light unitsÆmL )1 ). The internal and external pH’s were 4.96 and 8.54, respectively, and the [K + ] out and [K + ] in were 150 m M and 1 m M , respectively, thus inducing a K + /valinomycin diffus ion potential (+Du). Bchl concentration in t he luminometer cuvette was 1.1 l M . The reaction temperature was 1 3 °C. Oligomycin was added during the preincubation time at a concentratio n of 20 lgÆmL )1 and was present at the same concentration in the luminometer cuvette. Fig. 3. Rates of ATP synthesis induced by acid-base transitions in the presence and in t he absence of a diffusion potential. (A) Data are obtained from measurements similar to those reported in Fig. 2. The compositio n of the preincubation and reaction mixtures are detailed in the Experimental procedures. Bchl concentration varied between 1 and 8 l M . The reaction temperature was 13 °C. The ade nylate kinase rate has been subtracted . Data are averages of 2–3 measurements. (d,m) wild-type and ( s,n) mutant chromatophores. Rates w ere measured both in the p re sence (d,s) and in the ab sence (m,n)ofadiffusion potential (Du), i.e. with [K + ] out and [K + ] in equal to 1 and 150 m M , respectively, or 100 and 100 m M , respectively. The data points have been fitted with arbitrary functions. (B) An enlarged view of (A), showing only the data obtained in the absence of Du.(C)Theratioof the best-fitting functions and of the data points of wild-type over mutant are plotted for data in the presence (d)andabsence(m)ofDu. 1988 P. Turina and B. A. Melandri (Eur. J. Biochem. 269) Ó FEBS 2002 membranes and to independently superimpose an electrical diffusion potential Du, by varying the K + gradient across the vesicle membrane in the presence of valinomycin. In the present work, the [K + ] in and [K + ] out for + Du was 1 and 150 m M , respectively, and the [K + ] in and [ K + ] out for –Du was 100 and 100 m M , respectively. The DpH values were varied up to 3.8 units by decreasing the internal pH at constant external pH. Rb. c apsulatus chromatophores had been shown to catalyze, by this technique, rates of ATP synthesis comparable to those obtained by illumination [36]. At [K + ] in ¼ [K + ] out ¼ 100 m M (–Du condition), the absence of any significant diffusion potential due to other ionic species, e.g. monoionic succinate [44] was e xcluded observing the electrochromic spectral shift of carotenoids. We therefore conclude that in chromatophores of Rb. c ap- sulatus, under these conditions, DpH represents the only driving force for ATP synthesis. When carried out at room temperature, the linear phase of the D ~ ll H þ -induced ATP synthesis decays within a few hundred milliseconds [36] , due to the r elatively h igh permeability o f t he chromatophores membrane and to the high H + flow through the ATP synthase, thus requiring for its measurement a quench-flow apparatus. In the present work, the du ration of this linear phase was increased to a few seconds by decreasing the reaction temperature to 13 °C, and the transition was carried out manually by rapidly injecting the acidic chromatophores suspension into the luminometer cuvette containing the b asic solution and the luciferine/luciferase detection kit (plus additional 80 · 10 3 light units ÆmL )1 of luciferase). This method has already been applied at room temperature for measuring the ATP synthesis activity of ATP synthases incorporated into liposomes [45,46]. Figure 2 shows some of the traces obtained by this method. The initial high rates of ATP synthesis decay within a few seconds mainly due to the decay of the D ~ ll H þ -induced by the acid-base transitio n. The slow increase of ATP concentration seen thereafter was uncoupler- (not shown) and oligomycin-insensitive, and can be attributed to the adenylate kinase a ctivity present in chromatophores pre- parations [36]. Such a rate has been routinely subtracted from the initial ATP synthesis rate, which is uncoupler- and oligomycin-sensitive and can be attributed to the ATP synthase. This initial rate has been measured as a function of DpH in the presence and absence of Du for both mutant and wild-type chromatophores preparations. The resulting data are shown in Fig. 3A. In the presence of a Du,theATP synthesis rates of the m utants wer e about twofold lower with respect to the wild-type at every DpH tested. Figure 3B is an enlarged view of Fig. 3A showing only the r ates obtained in the absence of a Du. In this case, the impairment of the ATP synthesis rates in th e mutant was higher than twofold, especially in the low DpH range. Again, this latter comparison can be best appreciated by plotting the ratio of the fi tting f unctions of wild-type over mutant, either in the presence or in the absence of a Du,as shown in Fig. 3 C. In the presence of a Du, the difference between the two strains was twofold over the entire DpH range tested. In the absence of a Du, the ATP synthesis rate of the mutant was up to eightfold lower with respect to the wild-type for DpH values r anging between 2 .4 and 3.3, whereas a twofold factor was approached at the h ighest DpH tested. The trace –Du in Fig. 3C can b e d irectly compared to the +valinomycin trace in Fig. 1C, because, in the absence of a Du,theDpH is low in the short illumination times (200–500 ms), and it increases at longer times. Efficiency of proton pumping as estimated with the ACMA assay The proton-pumping activities of the mutant and wild-type chromatophores were compared first by measuring the ATP-induced fluorescence quenching of ACMA. For better comparison of the in itial rates of quenching, t he ATP hydrolysis rates were slowed down by decreasing the reaction temperature to 15 °C. The ACMA fluorescence quenching as a function of time after addition of ATP is shown i n Fig. 4A f or chromatophores of both strains. N o significant difference between the two chromatophores preparations could be detected. However, when the ACMA quenching was calibrated as a function of known DpH’s generated during acid-base transitions, as described previ- ously [37], the r esponse turned out to be sign ifican tly different for the mutant and wild-type c hromatophores (Fig. 4 B). This different respon se was systematically found in different experimental sessions and i n different chro- matophores preparations. Therefore, after converting the fluorescence quenching values into DpH values according to this calibration procedure, the initial rate of DpH formation resulted to be higher in the wild-type w ith respect to the mutant chromatophores by a factor of about 1.8, as shown in Fig. 4C. The rates of ATP hydrolysis measured under the same conditions of the ACMA quenching did not differ significantly (see Fig. 4D). Under t he assumption that the buffering capacity and inner volume of chromatophores were the same for both preparations, these data can be taken to indicate that the mutated enzyme was less efficient than the wild-type in ATP-driven proto n tr anslocation ( ATP slipping). The reason for this d ifferent response of ACMA to calibration is presently unclear, but it could be related to the only gross structural difference found between wild-type and mutant chromatophores, i.e. the different protein/Bchl ratio (see Table 1), which might affect the adsorption equilibria between the free and membrane-bound probe [37]. DISCUSSION The main finding of this work is t hat a single point mutation can a ffect the ATP synthesis rate o f an ATP synthase differently according to whether a Du is present or not. To our knowledge, no previous reports of a similar effect exist in the literature on ATP synthases . Two different lines of evidence reported in this work support this conclusion. In the light-driven ATP synthesis measurements (Fig. 1), the Du was c ollapsed by the ionophore valinomycin and a significantly higher impair- ment in the catalysis by the mutated enzyme (up to 12-fold) was seen during the first 1000 ms of illumination, i.e. when a DpH was building up but had not yet reached its maximum, stationary value. In the ATP synthesis induced by acid-base transitions ( Fig. 3 ), a Du in the presence o f the K + - transporter valinomycin was either imposed by a [K + ] gradient across the membrane, or prevented by imposing equally high K + concentrations in both intra- and extra- vesicular compartments. In the latter case, the A TP Ó FEBS 2002 The aE210K mutation in Rb. capsulatus ATP synthase (Eur. J. Biochem. 269) 1989 synthesis r ate was m uch lower (up to eightfold) in the mutant with respect to the wild-type in the DpH range between 2.4 and 3.3. The fact that this higher impairment is seen only within a limited pH in range strongly suggests that a step in the ATP synthase functioning, which is rate-limiting at th e low H + in concentrations, gradually ceases to b e r ate-limiting f or increasing H + in concentrations. It is this rate-limiting step, apparently, which is strongly affected in the mutant. What could be rate-limiting at t he lowest H + in concentrations is most likely the rate of H + binding to some of the H + binding sites in the H + translocation pathway, ultimately cAsp61. Models of the F 0 rotation driven by the protonmotive force [20,21] suggests a periplasmic (acidic) and a cytoplasmic (basic) half-channel, placed between subunit a and the c-subunit oligomer, connecting t he bulk aqueous compartments to the cAsp61 in the middle of the membrane. I n these models, an e lectrostatic con straint implies that c arboxyl groups on the r ing are always protonated ( electroneutral) when facing the lipid phase, while they can be unprotonated and charged when facing the a subunit. The t wo half-channels are spatially offset, so that this electrostatic constraint imposes a directionality to the c-oligomer rotational fluctuations. A ccording to the model p roposed by Oster and co workers [ 21], the rate constant k in for proton hopping into the channel from the periplasmic s ide d epends upon [H + ] in and upon the potential drop Dw due to surface charges, i.e. k in a10 –pHin exp(Dw/RT). As a Glu fi Lys substitution is expected to change the e lectrostatic profile o f t he nearby region, a possible explanation for the data presented is that residue a210 (a219 in E. co li) belongs to the periplasmic half- channel and that one of i ts roles is to create a favorable electrostatic profile f or incoming protons. This role is drastically altered when a lysine side-chain substitutes the carboxylate. Within this framework, the results obtained in the presence of a Du indicate that different steps of the reaction cycle b ecome rate-limiting under these conditions. According to the cited model [21], the main effects of a Du in speeding up the enzymatic rate are an increased rate o f H + releaseatthecytoplasmicsideandadecreasedrateof H + release at the periplasmic side. Particularly this last effect can presumably compensate for the low bulk H + in concentration b y effectively i ncreasing [H + ] w ithin the periplasmic half-channel. Under these conditions, i.e. in the presence of a high bulk-to-bulk Du, the overall reaction rate would be relatively less sensitive to the extent of the surface p otential drop Dw at the entrance of this half- channel. Fig. 4. Proton pumping and ATP hydrolysis. The composition o f the reaction mixtures are detailed in the Experimental procedures. ATP and Bchl concentrations were 400 and 20 l M , respectively. The t emperature was 15 °C. Photosynthetic electron transfer was inhibited by the addition of myxothiazol and antimycin to avoid any proton pumping due to the measuring light. The onset of a membrane potential was prevented b y the addition of 2 l M valinomycin. (d) wild-t ype and (m) mutant chromatophores. (A) ACMA fluorescence quenching as a function of time aft er addition of ATP. The original recorder trac es have been digitalized and data po ints have been fitted with arbitrary exponential functions. (B) Calibration of the ACMA fluorescence response to transmembrane pH differe nces. Th e ACMA fl uorescenc e que nching h as be en measure d after imposing various DpH jumps. The reaction m ixture was identical to that in Fig. 4A, except that 20 m M succina te was added to i ncr ease the buffering power at acidic pH and no ATP was added. The curves are the best-fit of data to the function described previously [37]: DpH ¼ (A Æ Q%) / (B ) Q%) Æ exp[Q%/(B ) Q%) + C Æ Q%]. The values obtained for the empirical fitting parameters were A ¼ 17.8, B ¼ 118 .8, C ¼ )0.046 and A ¼ 10.9, B ¼ 13.6, C ¼ )0.086 for wild-type and mutant, respectively. (C) The fluorescence quenching values of (A) have been converted to DpH values according to the best-fit functions of (B). (D) P i concentration as a function of time measured after addition of ATP to wild-type or mutant chromatophores, under experimental conditions identical to those of the fluorescence measurements of (A). After trichloroacetic acid denaturation, th e released P i was analyzed with the EnzCh eck Phosphate Assay. The rates of AT P hydrolysis resulted 4.7 m M ATP/(M BchlÆs) for both wild-type and the E210K mutant. 1990 P. Turina and B. A. Melandri (Eur. J. Biochem. 269) Ó FEBS 2002 A s econd major effect of the aE210K mutation is the generally lower ATP synthesis f ound regardless of D ~ ll H þ extent and composition, which might also explain t he lower growth rate, as found here and in E. coli strains carrying the homologous aE219K mut ation [28,29]. This general impairment of ATP synthesis amounted here ap proxi- mately to a factor of 2. This is seen for light-driven ATP synthesis, both for long and short illumination times (Table 1 and Fig. 1A, respectively), a s well as f or ATP synthesis driven b y acid-base transitions, both in the presence of a Du andinitsabsenceathighH + in concentrations (Fig. 3C). Several hypotheses can be put forward t o e xplain this impairment. One possibility is that the mutation perturbs the electrostatic profile sensed b y t he unproto nated Asp61 during i ts r adial translocation, thereby i ncreasing t he probability of protons leaking through F 0 without perform- ing any work [21]. This perturbation would also reduce the H + pumping efficie ncy, in line with t he conclusions suggested by our (Fig. 4) and other [28,29] data. Evidence for ATP slipping caused by the aE210K mutation, if confirmed, w ould b e a n i ndependent element for assessing theroleofGlu210intheprotonpathwaywithinF 0 . Alternatively, it can be t hat the mutation has an even longer-range effect, affecting t he conformation at the interface between the c-ring and the ce shaft. The conformational modification of t his interface could also lead to a similar phenotype of decreased coupling efficiency both in ATP synthesis and in ATP hydrolysis direction. As a direct connection between subunit a and F 1 has not yet been seen, s uch l ong-distance e ffect would p resumably o ccur through a change of the c-ring conformation. This hypoth- esis cannot be excluded especially in view of the results recently obtained by Cain and coworkers [47], which showed that structural changes resulting in the E. coli aR2 10 fi I, aA217 fi R, a nd a G218 fi D mutants are propagated to the e subunit, affecting its trypsin digestion and cross-linking pattern. In addition, a change in the e conformation was previously seen, a gain by trypsin digestion and cross-linking, following dicyclohexylcarbodiimide modification of cAsp61 [48], and a movement in the cytoplasmic loop of the isolated c subunit, which is in contact with the e subunit in the whole enzyme [12,14], had been detected by NMR spectroscopy on protonation of cAsp 61 [23]. Finally, it should be considered that the ATP synthase of Rb. c apsulatus had been shown t o undergo significant D ~ ll H þ -induced activation phenomena [49], whose possible interplay w ith the catalytic activity is not yet c lear. Very similar activation phenomena have recently been seen in the E. coli enzyme [50]. S trikingly, in both cases, the Du component was driving this activation significantly more than the DpH, suggesting t he presence o f an allosteric Du sensor in the transmembrane portion of F 0 . It cannot be ruled out that the aGlu210 is part of this activation switch, whose perturbation could lead both to a decrease in the functional efficiency of the enzyme population and to a different sensitivity towards DpH and Du, respectively, as observed in the present work. ACKNOWLEDGEMENTS This work has been supported by t he grant PRIN/01, Processi Ossidoriduttivi e Trasduzione di Energia in Membrane Procariotiche ed Eucariotiche, from the Italian M inistery for E ducation of University and Research (MIUR). We are grateful to G. Venturoli for many discussions and to F. Federici and D. Giovannini for excellent help in the experiments. REFERENCES 1. Weber, J. & Senior, A .E. (1997) Catalytic mechanism of F 1 - ATPase. Biochim. Biophys. Acta 1319, 19–58. 2. Boyer, P.D. (1998) Catalytic site forms and controls in ATP synthase catalysis. Biochim. Biophys. Acta 1365, 3–9. 3. Fillingame, R.H., Jiang, W. & Dmitr iev, O.Y. (2000) Coupling H + transport to rotary catalysis in F-type ATP synthases: struc- ture and organization of the transmembrane rotary motor. J. Exp. Biol. 203, 9–17. 4. Abrahams, J.P., Leslie, A.G., Lutter, R. & Walker, J.E. (1994) Structure at 2.8 A ˚ resolution of F 1 -ATPasefrombovineheart mitochondria. Nature 370, 621–628. 5. Yoshida, M., Muneyuki, E. & Hisabori, T. (2001) ATP synthase – a marvellous rotary engine of the cell. Nat. Rev. Mol. Cell. Biol. 2, 669–677. 6. Girvin, M.E., Rastogi, V.K., Abildgaard, F., Markley, J.L. & Fillingame, R.H. (1998) Solution structure of the transmembrane H + -transporting subunit c of the F 1 F 0 ATP synthase. Biochem- istry 37, 8817–8824. 7. Stock, D., Leslie, A.G. & Walker, J.E. ( 1999) Molecular archi- tecture of the rotary motor in ATP synthase. Science 286, 1700– 1705. 8. Valiyaveetil, F.I. & Fillingame, R.H. (1998) Transmembrane topography of subunit a in th e Escherichia c oli F 1 F 0 ATP synthase. J. Biol. Chem. 273, 16241–16247. 9. Long, J.C., Wang, S. & Vik, S.B. (1998) Membrane topology of subu nit a of the F 1 F 0 ATP synthase as d etermined b y labeling of uniq ue cysteine re sidues. J. Biol. Che m. 273, 16235– 16240. 10. Wada, T., Long, J.C., Zhang, D. & Vik, S.B. (1999) A novel labeling approach supports the five-transmembrane model of subunit a of the Escherichia coli ATP synthase. J. Biol. Chem. 274, 17353–17357. 11. Jiang, W. & Fillingame, R.H. (1998) I nteracting helical faces of subunits a and c in the F 1 F 0 ATP syn thase of Escherichia coli defined by disulfide cross-linking. Proc. Natl Acad. Sci. USA 95, 6607–6612. 12. Zhang, Y. & Fillingame, R.H. (1995) Changing the ion binding specificity of the Escherichia coli H + -transporting AT P s ynthase by directed mutagenesis of subunit c. J. Biol. Chem. 27 0, 24609– 24614. 13. Watts, S.D., Tang, C. & Capaldi, R.A. (1996) The stalk region of the Escherichia coli ATP s ynthase. Tyrosine 205 of the gamma subunit is in the interface between the F 1 and F 0 parts and can interact with bot h the epsilon and c oligome r. J. Biol. Chem. 271, 28341–28347. 14. Hermolin, J., Dmitriev, O.Y., Zhang, Y. & Fillingame, R.H. (1999) Defining the domain of binding of F 1 subunit ep silon with the polar loop of F 0 subunit c in the Escherichia coli ATP synthase. J. Biol. Chem. 274, 17011–17016. 15. Dunn, S.D., McLachlin, D.T. & Revington, M. (2000) The second stalk of Escherichia coli ATP synthase. Biochim. Biophys. Acta 1458, 356–363. 16. Capaldi, R.A., Schulenberg, B., Murray, J. & Aggeler, R. (2000) Cross-linking and electron microscopy studies of the structure and functioning of the Escherichia coli ATP synthase. J. Exp. Biol. 203, 29–33. 17. Sambongi, Y., Iko, Y., Tanabe, M., Omote, H., Iwamoto-Kihara, A., Ueda, I., Y anagida, T., Wada, Y. & Futai, M. (1999) Mechanical rotation of the c subunit oligomer in ATP s ynthase (F 0 F 1 ): direct observation. Science 286, 1722–1724. Ó FEBS 2002 The aE210K mutation in Rb. capsulatus ATP synthase (Eur. J. Biochem. 269) 1991 18. Pa ¨ nke, O., Gumbiowski, K., Junge, W. & Engelbrecht, S. (2000) F-ATPase: spec ific observation of the rotating c subunit oligomer of EF 0 EF 1 . FEBS Lett. 472, 34–38. 19. Tsunoda, S.P., Aggeler, R., Yoshida, M. & Capaldi, R.A. (2001) Rotation of the c subunit oligomer in fully functional F 1 F 0 ATP synthase. Proc. Natl Acad. Sci. USA 98, 898–902. 20. Junge, W., Lill, H. & Engelbrecht, S. (1997) A TP s ynthase: an electrochemical transducer with rotatory mechanics. Trends Bio- chem. Sci. 22, 420–423. 21. Elston, T., Wang, H. & Oster, G. (1998) Energy transduction in ATP synthase. Nature 391, 510–514. 22. Vik, S.B., Long, J.C., Wada, T. & Zhang, D. (2000) A model for the structure of subunit a of the Escherichia coli ATP synthase and its role in proton translocation. Biochim. Biophys. Acta 1458,457– 466. 23. Rastogi, V.K. & Girvin, M.E. (1999) Structural changes linked to proton translocation by subunit c of the ATP synthase. Nature 402, 263–268. 24. Hartzog, P.E. & Cain, B.D. (1994) Second-site suppressor muta- tions at glycine 218 and histidine 245 in the alpha subunit of F 1 F 0 ATP synthase in Escherichia coli. J. Biol. Chem. 269, 32313–32317. 25. Cain, B.D. & Simoni, R.D. (1988) Interaction between Glu-219 and His-245 within the a subunit of F 1 F 0 -ATPase in Escherichia coli. J. Biol. Chem. 263, 6606–6612. 26. Borghese, R., Crimi, M., Fava, L. & Melandri, B.A. (1998) The ATP s ynthase atpHAGDC (F 1 ) operon from Rhodobacter cap- sulatus. J. Bacteriol. 180, 416–421. 27. Borghese, R., Turina, P., Lambertini, L. & Melandri, B.A. (1998) The atpIBEXF operon coding for the F 0 sector of the ATP syn- thase from the purple nonsulfur photosynthetic bacterium Rho- dobacter capsulatus. Arch. Microbiol. 170 , 385–388. 28. Valiyaveetil, F.I. & Fillingame, R.H. (1997) On the role of Arg-210 and Glu-219 of subunit a in proton translocation by the Escher- ichia coli F 0 F 1 -ATP synthase. J. Biol. Chem. 272, 32635–32641. 29. Hatch, L.P., Cox, G.B. & Howitt, S.M. (1998) Glutamate residues at positions 219 and 252 in the a-subun it of the Escherichia coli ATP synthase are not functionally equivalent. Biochim. Biophys. Acta 1363, 217–223. 30. Baccarini-Melandri, A. & Melandri, B.A. (1971) Partial resolution of phosphorylating system of Rps. Capsulata. Methods Enzymol. 23, 556–561. 31. Keen, N.T., Tamaki, S., Kobayashi, D. & Trollinger, D. (1988) Improved broad-host-range plasmids for DNA cloning in gram- negative bacteria. Gene 70, 191–197. 32. Sharak Genthner, B.R. & Wall, J.D. (1985) Ammonium uptake in Rhodopseudomonas capsulata. Arch. Microbiol. 141, 219–224. 33. Yen, H.C. & Marrs, B. (1976) Map of genes for carotenoid and bacteriochlorop hyll biosynthesis in Rhodopseudomonas capsulata. J. Bacteriol. 126, 619–629. 34. Gabellini, N., Gao, Z., Eckerskorn, C., Lottspeich, F. & Oesterhelt, D. (1988) Purification of the H + -ATPase from Rho- dobacter capsulatus, identification of the F 1 F 0 components and reconstitution of the active enzyme. Biochim. Biophys. Acta 934, 227–234. 35. Melandri, B.A., De Santis, A., Venturoli, G. & Baccarini- Melandri, A. (1978) The rates of onset of photophosphorylation and of the protonic electrochemical potential difference in chro- matophores. FEBS Lett. 95, 130–134. 36. Turina, P., Melandri, B.A. & Gra ¨ ber, P. (1991) ATP synthesis in chromatophores driven by artificially induced ion gradients. Eur. J. Biochem. 196, 225–229. 37. Casadio, R. (1991) Measurements of transmembrane pH differ- ences of low extents in bacterial chromatophores Eur. Biophys. J. 19, 189–201. 38. Taussky, M. & Schorr, J. (1953) A m icrocolorimetric method for the determination of inorganic phosphate. J. Biol. Chem. 202,675– 686. 39. Lanzetta, P.A., Alvarez, L.J., Reinach, P.S. & Candia, O.A. (1979) An improved assay for nanom ole amounts of inorganic phos- phate. Anal. Biochem. 100, 95–97. 40. Clayton, R.K. (1963) Toward the isolation of a photoche mical reaction center in Rhodopseudomonas sphaeroides. Bioc him. Bio- phys. Acta 75, 312–323. 41. Jackson, B.J. & Nicholls, D.G. (1986) Methods for the determ ina- tion of membrane p otential in bioenergetic s ystems. Methods Enzymol. 127, 557–577. 42. Barz, W.P., Francia, F., Venturoli, G., Melandri, B.A., Vermeglio, A. & Oesterhelt, D. (1995) Role of PufX protein in photosynthetic growth of Rhodobacter sphaeroides. 1. PufX is required for efficient light-driven electron transfer and p hotophosphorylation u nder anaerobic conditions. Biochemistry 34, 15235–15247. 43. Petitou, M., Tuy, F. & Rosenfeld, C. (1978) A simplified proce- dure for organic phosphorus dete rmination fro m phospho lipids. Anal. Biochem. 91, 350–353. 44. Kaim, G. & Dimroth, P. (1999) ATP synthesis by F-type ATP synthase is obligatorily dependent on the transmembrane voltage. EMBO J. 18, 4118–4127. 45. Slooten, L. & V andenbranden, S. (1989) ATP-synthesis by proteoliposomes inc orporat ing Rhodospirillum rubrum F 0 F 1 as measured with firefly luc iferase: dependence on delta psi and delta pH. Biochim. Biophys. Acta 976, 150–160. 46. Fischer, S ., Etzold, C., T u rina, P., Deckers-Hebestreit, G., Altendorf, K. & Gra ¨ ber, P. (1994) ATP synthesis catalyzed by the ATP synthase of Escherichia coli reconstituted into liposomes. Eur. J. Biochem. 225, 167–172. 47. Gardner, J.L. & Cain, B.D. (1999) Amino acid substitutions in the a subunit affect the epsilon subunit of F 1 F 0 ATP synthase from Escherichia coli. Arch. Biochem. Biophys. 361, 302–308. 48. Mendel-Hartvig, J. & Capaldi, R.A. (1991) Nucleotide-dependent and dicyclohexylcarbodiimide-sensitive conformational changes in the epsilon subunit of Escherichia coli ATP synthase. Biochemistry 30, 10987–10991. 49. Turina, P., Rumberg, B., Melandri, B.A. & Gra ¨ ber, P. (1992) Activation of the H + -ATP syn thase in the ph otosynthetic bacter- ium Rhodobacter capsulatus. J. Biol. Chem. 267, 11057–11063. 50. Fischer, S., Gra ¨ ber, P. & Turina, P. (2000) The activity of the ATP synthase from Escherichia coli is regulated by the transmembrane proton motive force. J. Biol. Chem. 275, 30157–30162. 1992 P. Turina and B. A. Melandri (Eur. J. Biochem. 269) Ó FEBS 2002 . A point mutation in the ATP synthase of Rhodobacter capsulatus results in differential contributions of DpH and Du in driving the ATP synthesis reaction Paola. solution. The amount of sy nthetized ATP w as evaluated by a dding 10–25 n M standard ATP. ATP synthesis induced by acid-base transitions Acid-base driven ATP synthesis