The Role of the βDELSEED-loop of ATP Synthase*

ATP synthase uses a unique rotational mechanism to convert chemical energy into mechanical energy and back into chemical energy. The helix-turn-helix motif, termed “DELSEED-loop,” in the C-terminal domain of the β subunit was suggested to be involved in coupling between catalysis and rotation. Here, the role of the DELSEED-loop was investigated by functional analysis of mutants of Bacillus PS3 ATP synthase that had 3–7 amino acids within the loop deleted. All mutants were able to catalyze ATP hydrolysis, some at rates several times higher than the wild-type enzyme. In most cases ATP hydrolysis in membrane vesicles generated a transmembrane proton gradient, indicating that hydrolysis occurred via the normal rotational mechanism. Except for two mutants that showed low activity and low abundance in the membrane preparations, the deletion mutants were able to catalyze ATP synthesis. In general, the mutants seemed less well coupled than the wild-type enzyme, to a varying degree. Arrhenius analysis demonstrated that in the mutants fewer bonds had to be rearranged during the rate-limiting catalytic step; the extent of this effect was dependent on the size of the deletion. The results support the idea of a significant involvement of the DELSEED-loop in mechanochemical coupling in ATP synthase. In addition, for two deletion mutants it was possible to prepare an α3β3γ subcomplex and measure nucleotide binding to the catalytic sites. Interestingly, both mutants showed a severely reduced affinity for MgATP at the high affinity site.

F 1 F 0 -ATP synthase catalyzes the final step of oxidative phosphorylation and photophosphorylation, the synthesis of ATP from ADP and inorganic phosphate. F 1 F 0 -ATP synthase consists of the membrane-embedded F 0 subcomplex, with, in most bacteria, a subunit composition of ab 2 c 10 , and the peripheral F 1 subcomplex, with a subunit composition of ␣ 3 ␤ 3 ␥␦⑀. The energy necessary for ATP synthesis is derived from an electrochemical transmembrane proton (or, in some organisms, a sodium ion) gradient. Proton flow down the gradient through F 0 is coupled to ATP synthesis on F 1 by a unique rotary mechanism. The protons flow through (half) channels at the interface of the a and c subunits, which drives rotation of the ring of c subunits. The c 10 ring, together with F 1 subunits ␥ and ⑀, forms the rotor. Rotation of ␥ leads to conformational changes in the catalytic nucleotide binding sites on the ␤ subunits, where ADP and P i are bound. The conformational changes result in the formation and release of ATP. Thus, ATP synthase converts electrochemical energy, the proton gradient, into mechanical energy in the form of subunit rotation and back into chemical energy as ATP. In bacteria, under certain physiological conditions, the process runs in reverse. ATP is hydrolyzed to generate a transmembrane proton gradient, which the bacterium requires for such functions as nutrient import and locomotion (for reviews, see Refs. 1-6). F 1 (or F 1 -ATPase) has three catalytic nucleotide binding sites located on the ␤ subunits at the interface to the adjacent ␣ subunit. The catalytic sites have pronounced differences in their nucleotide binding affinity. During rotational catalysis, the sites switch their affinities in a synchronized manner; the position of ␥ determines which catalytic site is the high affinity site (K d1 in the nanomolar range), which site is the medium affinity site (K d2 Ϸ 1 M), and which site is the low affinity site (K d3 Ϸ 30 -100 M; see Refs. 7 and 8). In the original crystal structure of bovine mitochondrial F 1 (9), one of the three catalytic sites, was filled with the ATP analog AMP-PNP, 2 a second was filled with ADP (plus azide) (see Ref. 10), and the third site was empty. Hence, the ␤ subunits are referred to as ␤ TP , ␤ DP , and ␤ E . The occupied ␤ subunits, ␤ TP and ␤ DP , were in a closed conformation, and the empty ␤ E subunit was in an open conformation. The main difference between these two conformations is found in the C-terminal domain. Here, the "DELSEED-loop," a helixturn-helix structure containing the conserved DELSEED motif, is in an "up" position when the catalytic site on the respective ␤ subunit is filled with nucleotide and in a "down" position when the site is empty (Fig. 1A). When all three catalytic sites are occupied by nucleotide, the previously open ␤ E subunit assumes an intermediate, half-closed (␤ HC ) conformation. It cannot close completely because of steric clashes with ␥ (11).
The DELSEED-loop of each of the three ␤ subunits makes contact with the ␥ subunit. In some cases, these contacts consist of hydrogen bonds or salt bridges between the negatively charged residues of the DELSEED motif and positively charged residues on ␥. The interactions of the DELSEED-loop with ␥, its movement during catalysis, the conservation of the DELSEED motif (see Table 1), and a number of mutagenesis experiments led to the assumption that the DELSEED-loop might play an essential role in coupling between catalysis and rotation of ␥ (12)(13)(14). Thus, the finding that an AALSAAA mutant in the ␣ 3 ␤ 3 ␥ complex of ATP synthase from the thermophilic Bacillus PS3, where several hydrogen bonds/salt bridges to ␥ are removed simultaneously, could drive rotation of ␥ with the same torque as the wild-type enzyme (14) came as a surprise. On the other hand, it seems possible that it is the bulk of the DELSEED-loop, more so than individual interactions, that drives rotation of ␥. According to a model favored by several authors (6,15,16) (see also , binding of ATP (or, more precisely, MgATP) to the low affinity catalytic site on ␤ E and the subsequent closure of this site, accompanied by its conversion into the high affinity site, are responsible for driving the large (80 -90°) rotation substep during ATP hydrolysis, with the DELSEED-loop acting as a "pushrod." A recent molecular dynamics (20) study supports this model and implicates mainly the region around several hydrophobic residues upstream of the DELSEED motif (specifically ␤I386 and ␤L387) 3 as being responsible for making contact with ␥ during the large rotation substep.
In the present study, we investigated the function of the DEL-SEED-loop using an approach less focused on individual residues, by deleting stretches of 3-7 amino acids between positions ␤380 and ␤402 of ATP synthase from the thermophilic Bacillus PS3. We analyzed the functional properties of the deletion mutants after expression in Escherichia coli. The mutants showed ATPase activities, which were in some cases surprisingly high, severalfold higher than the activity of the wild-type control. On the other hand, in all cases where ATP synthesis could be measured, the rates where below or equal to those of the wild-type enzyme. In Arrhenius plots, the hydrolysis rates of the mutants were less temperature-dependent than those of wild-type ATP synthase. In those cases where nucleotide binding to the catalytic sites could be tested, the deletion mutants had a much reduced affinity for MgATP at high affinity site 1. The functional role of the DELSEED-loop will be discussed in light of the new information.
The first seven deletion mutations were also introduced in plasmid pNM1. Plasmid pNM1 is a derivative of plasmid pKAGB1 (23). pKAGB1 is used to express a Cys-and Trp-less form of the ␣ 3 ␤ 3 ␥ subcomplex of PS3. pNM1 contains an additional mutation to generate an ␣ 3 (␤Y341W) 3 ␥ subcomplex, which allows monitoring of nucleotide binding to the three catalytic sites. For expression, pNM1 and the derived deletion mutants were transformed into E. coli strain JM103 ⌬(uncB-uncD).
Isolation of Inverted Membrane Vesicles and Determination of F 1 F 0 Content in E. coli Membranes-E. coli strain DK8 harboring wild-type or mutated pTR19-ASDS plasmids was aerobically cultivated at 37°C for 18 h in 2ϫ YT medium (24) containing 100 g/ml ampicillin. Inverted membrane vesicles from E. coli cells expressing thermophilic F 1 F 0 were prepared as described (21,25). The amount of wild-type F 1 F 0 in E. coli membrane preparations was determined by SDS-PAGE as visualized by staining with Coomassie Brilliant Blue (21). The relative amount of mutant F 1 F 0 in the membranes was estimated via Western blots, using an anti-␤ antibody (Agrisera, Vännäs, Sweden) or an anti-␣/anti-␤ antibody (kindly provided by Dr. Bill Brusilow, Wayne State University). The staining intensity was quantified using a Photodyne imaging system and ImageJ acquisition software (National Institutes of Health).
Functional Analysis of Mutant Strains and Enzymes-Growth of strains expressing wild-type or mutant PS3 ATP synthase on succinate plates and growth in limiting glucose was determined as described previously (26).
ATPase activities were assayed in a buffer containing 50 mM Tris/H 2 SO 4 , 10 mM ATP, and 4 mM MgSO 4 , pH 8.0. For the ␣ 3 ␤ 3 ␥ subcomplex, the assay was performed at 60°C, for membrane vesicles at 42°C. The reaction was started by the addition of 10 -20 g/ml ␣ 3 ␤ 3 ␥ or 50 -150 g/ml membrane vesicles and stopped after 1, 2, or 10 min (depending on the activity) by the addition of sodium dodecyl sulfate (final concentration 5% (w/v)). The released P i was measured as described (27). 1 unit of enzymatic activity corresponds to 1 mol ATP hydrolyzed (equivalent to 1 mol of P i produced) or synthesized per min. Turnover numbers (k cat ) in membrane vesicles were calculated based on the F 1 F 0 content measured as described above. The temperature dependence of ATPase activities of membrane vesicles was measured in the buffer given above. The temperature was varied between 20 and 50°C. The reaction mixture was preincubated for 10 min at the desired temperature. Activation energies, and entropic and enthalpic components were calculated from Arrhenius plots as described (28).
ATP synthesis activity in vitro was measured as follows. Inverted membrane vesicles were suspended in a solution containing 20 mM HEPES/KOH, 100 mM KCl, 5 mM MgCl 2 , 1 mM ADP, 5 mM KP i , 10% glycerol, pH 7.5, and incubated at 42°C. The reaction was initiated by the addition of 2 mM NADH. After 10, 40, and 70 s, aliquots of the reaction mixture (each containing 20 g of membrane protein) were transferred into boiling buffer of 100 mM Tris/H 2 SO 4 , 4 mM EDTA, pH 7.75, for heat denaturation. The samples were incubated at 100°C for 2 min, cooled on ice, and centrifuged for 1 min at 1000 ϫ g. The amount of ATP was determined by the luciferin/luciferase method (CLS II ATP bioluminescence kit, Roche Applied Sci-ence). Light emission was measured at 562 nm in a Fluorolog 3 spectrofluorometer (HORIBA Jobin Yvon, Edison, NY).
NADH-and ATP-driven H ϩ -pumping in membrane vesicles was measured via fluorescence quenching of ACMA at 42°C. To a buffer of 10 mM HEPES/KOH, 100 mM KCl, and 5 mM MgCl 2 , pH 7.5, 0.5 mg/ml membrane vesicles and 0.3 g/ml ACMA were added. Proton pumping was initiated by adding NADH or ATP to a final concentration of 1 mM and terminated by adding CCCP (final concentration 1 M). The excitation wavelength was 410 nm and the emission wavelength 480 nm.
Binding of MgATP to the catalytic sites of the purified ␣ 3 ␤ 3 ␥ subcomplex was measured using the fluorescence of the inserted Trp residue, ␤W341 (8). Before use, the ␣ 3 ␤ 3 ␥ ammonium sulfate precipitate was pelleted by centrifugation and redissolved in a buffer containing 50 mM Tris/HCl, 10 mM CDTA, pH 8.0. After 1 h of incubation at 23°C, the ␣ 3 ␤ 3 ␥ subcomplex was passed through two subsequent centrifuge columns containing Sephadex G-50, equilibrated with 50 mM Tris/ HCl and 0.1 mM EDTA, pH 8.0. After this treatment, the enzyme subcomplex is essentially nucleotide-free (29). Fluorescence titrations were performed in a buffer containing 50 mM Tris/H 2 SO 4 , 2.5 mM MgSO 4 , pH 8.0, with ATP added in the appropriate concentrations. Each protein sample was used to acquire maximally two data points. K d values were determined by fitting of theoretical curves to the experimental data points by nonlinear least-squares analysis, assuming a model with three different independent sites (8).
Protein concentrations of membrane vesicles were determined by the Lowry method (30) and those of purified ␣ 3 ␤ 3 ␥ subcomplex by the Bradford method (31). Both assays used bovine serum albumin as standard.
Modeling-Homology modeling including energy minimization refinement was performed using the program PRIME (Schroedinger Inc.). Templates were the structures of bovine mitochondrial ␤ subunits in closed (␤ TP and ␤ DP ), half-closed (␤ HC ), and empty (␤ E ) conformations, taken from Protein Data Bank files 1h8e (11) and 1e79 (32). The ␥ subunit was included during the energy minimization step. Fig. 1B shows the helix-turn-helix structure known as the DELSEEDloop in the C-terminal domain of the ␤ subunit of ATP synthase. As can be seen from Table 1, not only the DELSEED motif itself but the whole tip of the loop is strongly conserved. To assess the function of the loop, we designed deletion mutants that had 3-4 contiguous residues removed. The exact length of the deletion (3 or 4) was selected to make a "reconnection" of both ends of the protein chain as easy as possible as judged by eye. Originally, all mutations were made in E. coli ATP synthase; however, none of the deletion mutants assembled properly. Thus, the mutations were repeated in ATP synthase of the thermophilic bacterium Bacillus PS3 expressed in E. coli. In this system, all of the mutant enzymes with deletions of 3-4 residues removed could be obtained in membranebound form. The deletions were: ⌬ 380 LQDI 383 , ⌬ 384 IAIL 387 , ⌬ 388 GMDE 391 , ⌬ 392 LSD 394 , ⌬ 395 EDKL 398 , and ⌬ 399 VVHR 402 (color-coded in Fig. 1B). It should be noted that DELSEED is actually 390 DELSDED 396 in the thermophilic enzyme. In a second mutagenesis round, several longer deletions of 7 residues were generated, mostly by combining two consecutive shorter deletions, ⌬ 381 QDIIAIL 387 , ⌬ 388 GMDELSD 394 , and ⌬ 392 LSD-EDKL 396 , plus ⌬ 390 DELSDED 396 where the DELSEED motif itself is removed. With the exception of the ⌬ 381 QDIIAIL 387 deletion mutant, in all other cases the amount of F 1 F 0 found in membrane vesicles was not sufficient for inclusion in this study FIGURE 1. The ␤DELSEED-loop. A, interaction of the ␤ TP and ␤ E subunits with the ␥ subunit. ␤ subunits are shown in yellow and ␥ in blue. The DELSEED-loop (shown in orange, with the DELSEED motif itself in green) of ␤ TP interacts with the C-terminal helix ␥ and the short helix that runs nearly perpendicular to the rotation axis. The DELSEED-loop of ␤ E makes contact with the convex portion of ␥, formed mainly by the N-terminal helix. A nucleotide molecule (shown in stick representation) occupies the catalytic site of ␤ TP , and the subunit is in the closed conformation. The catalytic site on ␤ E is empty, and the subunit is in the open conformation. This figure is based on Protein Data Bank file 1e79 (32). B, deletions in the ␤DELSEED-loop. The loop was "mutated" in silico to represent the PS3 ATP synthase. The 3-4-residue segments that are removed in the deletion mutants are color-coded as follows: 380 LQDI 383 , pink; 384 IAIL 387 , green; 388 GMDE 391 , yellow; 392 LSD 394 , cyan; 395 EDKL 398 , orange; 399 VVHR 402 , blue. Residues that are the most involved in contacts with ␥ are labeled. All figures were generated using the program PyMOL (DeLano Scientific, San Carlos, CA).

␤DELSEED-loop of ATP Synthase
(Ͻ2% of wild type). In the case of the ⌬ 390 DELSDED 396 deletion mutant, this finding reflects an earlier result (14) where no ␣ 3 ␤ 3 ␥ subcomplex of thermophilic F 1 containing this deletion could be obtained.
Oxidative Phosphorylation in Vivo-As described previously (21), wild-type PS3 F 1 F 0 expressed in E. coli (strain pTR19-ASDS/DK8) allowed growth on plates containing succinate as the sole carbon source, demonstrating that the PS3 enzyme is able to catalyze ATP synthesis in vivo. The growth of strain pTR19-ASDS/DK8 was about 75% as strong as that of pBWU13.4/DK8, which expresses wild-type E. coli ATP synthase. In contrast, none of the deletion mutants showed detectable growth on succinate plates after overnight incubation at 37°C. Prolonged incubation (3 days) resulted in the growth of some of the mutants, especially ⌬ 388 GMDE 391 and ⌬ 392 LSD 394 . However, as there was also growth, albeit very weak, even on the negative control plate containing strain pUC118/DK8 (presumably due to remnants of the LB medium used to streak out the bacteria), this line of experiments was aborted.
In growth yield assays in limiting glucose, strain pTR19-ASDS/DK8 expressing wild-type PS3 F 1 F 0 grew to a turbidity (measured as absorbance at 595 nm) of 62% of that of the control strain pBWU13.4/DK8 expressing wild-type E. coli ATP synthase. The negative (unc Ϫ ) control, strain pUC118/DK8, reached 38% of the value for strain pBWU13.4/DK8 and 61% of the value for strain pTR19-ASDS/DK8. Three of the deletion mutants, ⌬ 380 LQDI 383 , ⌬ 388 GMDE 391 , and ⌬ 392 LSD 394 , showed growth yields significantly higher than the negative control but clearly below the value for the wild-type PS3 enzyme (Table 2). Deletion mutants ⌬ 395 EDKL 398 and ⌬ 399 VVHR 402 exhibited growth yields similar to those of the negative control, whereas ⌬ 384 IAIL 387 and ⌬ 381 QDIIAIL 387 grew less well. The latter result may be indicative of a highly uncoupled enzyme or of severe proton leaks through the membranes due to incorrectly or incompletely inserted proteins.
ATP Synthesis Activity of Membrane Preparations-In addition to the in vivo growth assays to monitor oxidative phosphorylation, we measured NADH-driven ATP synthesis in vitro. At 42°C, the wild-type PS3 enzyme in E. coli membrane vesicles showed an ATP synthesis activity of 71 milliunits/mg membrane protein. Except for ⌬ 395 EDKL 398 and ⌬ 399 VVHR 402 , the other deletion mutants also exhibited some ATP synthesis activity. As well as being a direct consequence of a mutation, lack of activity can be due to lack of expression or to oligomeric instability of the enzyme; thus, it was necessary to quantify the amount of enzyme on the membranes. The amount of wildtype PS3 F 1 F 0 in E. coli membrane preparations was found to be ϳ20% of the total membrane protein (this study and Ref. 21). Membrane preparations containing PS3 ATP synthase with the deletions ⌬ 380 LQDI 383 , ⌬ 384 IAIL 387 , ⌬ 388 GMDE 391 , ⌬ 392 LSD 394 , ⌬ 395 EDKL 398 , ⌬ 399 VVHR 402 , and ⌬ 381 QDIIAIL 387 had between 10 and 100% of the amount of the wild-type enzyme ( Table 2). The lack of activity of the ⌬ 395 EDKL 398 and ⌬ 399 VVHR 402 mutants could reflect the low amount of enzyme in the membrane preparations. On the other hand, the ⌬ 381 QDIIAIL 387 mutant ATP synthase, present in similarly low quantity, showed significant synthesis activity. After correction for the different amounts of enzyme, the turnover rates of the deletion mutants (except for ⌬ 395 EDKL 398 and ⌬ 399 VVHR 402 ) were between 60 and 100% of that of the wild-type enzyme.
ATPase Activity of Membrane Preparations-As seen in Table 3, all deletion mutants showed significant ATPase activity. After correction for the different amount of enzyme on the membranes, the turnover rates (at 42°C) reached 10 -15% of the wild type for ⌬ 395 EDKL 398 and ⌬ 399 VVHR 402 to astonishingly high values of 350 and 500% of wild type in ⌬ 381 QDIIAIL 387 and ⌬ 384 IAIL 387 , respectively. Inclusion of the

Conservation of residues in the DELSEED-loop
Amino acids found in selected species in the turn region of the DELSEED-loop. Listed are all positions subjected to deletions in the present study. Residue numbers refer to the PS3 enzyme. Consensus annotation: p, polar residue; s, small residue; h, hydrophobic residue; ؊, negatively charged residue; ϩ, positively charged residue.

TABLE 2 Oxidative phosphorylation in vivo and ATP synthesis activities in vitro
Oxidative phosphorylation in vivo and ATP synthesis activities of membrane preparations of deletion mutants and controls were measured as described under "Experimental Procedures." The positive control was strain pTR19-ASDS/DK8, which expresses Bacillus PS3 ATP synthase in E. coli; this strain served as background for the deletions. The negative control was strain pUC118/DK8, which expresses neither PS3 ATP synthase nor the endogenous E. coli enzyme. The amount of F 1 F 0 in the membrane preparations was measured by quantitative immunoblot analysis as described under "Experimental Procedures." Turnover rates were calculated using a molecular mass of 531 kDa for the holoenzyme, taking into account the differing amounts of ATP synthase in the individual membrane preparations. ND, no activity detectable beyond background.

Strain/mutation
Growth yields in limiting glucose Just taken by itself, this finding could be attributed to the failure to build up a significant proton gradient or to the inability of the proton gradient to exert sufficient back-pressure in this specific mutant, i.e. uncoupling. To obtain a more quantitative description of coupling efficiency, we calculated the ratio of the rates for ATP synthesis and hydrolysis (Table 3). Coupling appears to be impaired in all mutants except ⌬ 388 GMDE 391 , which has a synthesis/hydrolysis ratio close to that of the wild type. NADH-and ATP-induced Proton Pumping-NADH-induced ACMA quenching was not reduced in membrane vesicles containing deletion mutants. The quenching reached values around 80 -85% just as with wild-type PS3 enzyme, indicating that the mutations do not prevent the buildup of a considerable proton gradient. Thus, the deletions did not cause stability problems which increased the "leakiness" of the membranes. It should be noted, however, that reduced stability does not always manifest itself in enhanced proton leak rates. With the possible exception of ⌬ 395 EDKL 398 and ⌬ 399 VVHR 402 , for the other deletion mutants this result was expected based on their normal or close-to-normal ATP synthesis rates.
Upon ATP hydrolysis, wild-type PS3 ATP synthase in E. coli membrane vesicles forms a proton gradient that is only slightly smaller than that produced by the "native" E. coli enzyme. ATPinduced ACMA quenching reached values of ϳ75% with the PS3 enzyme and 85% with wild-type E. coli ATP synthase. In the former case, but not in the latter, quenching seemed to occur in two phases, a fast phase, which was completed after 10 -15 s and was associated with 35-40% quenching, followed by a slow phase of Ͼ100 s (Fig. 2). Similar behavior had been observed previously (see e.g. Fig. 3A in Ref. 21 and Fig. 2C in Ref. 33) but was not addressed. As the assay contains an excess of Mg 2ϩ over ATP, it might be speculated that during the slow phase only a fraction of the enzyme molecules is active, whereas the majority is in an MgADP-inhibited form. This view is supported by the finding that the addition of external 1 mM P i , which has been shown to suppress MgADP inhibition (34), reduces the second phase or even abolishes it completely (data not shown).
Of the deletion mutants, ⌬ 395 EDKL 398 and ⌬ 399 VVHR 402 did not produce measurable ATP-induced ACMA quenching; the membrane ATPase activity was apparently not high enough to overcome the natural proton leak rate of the membranes. In contrast, all other deletion mutants could use ATP hydrolysis to form a proton gradient, demonstrating that the ATPase activity was not completely uncoupled. Like wild-type PS3 ATP synthase, the ⌬ 388 GMDE 391 mutant showed biphasic kinetics, reaching a total value of about 60% quenching (Fig. 2). Of the other deletion mutants, ⌬ 380 LQDI 383 and ⌬ 392 LSD 394 did not show an obvious second quenching phase. Both reached their respective final quenching values of ϳ45 and ϳ 80% nearly instantaneously. With the ⌬ 384 IAIL 387 and ⌬ 381 QDIIAIL 387 deletion mutants, only the initial fast phase resulted in quenching. In the slower phase, ATP-driven H ϩ -pumping appeared to be reduced to such an extent that it could no longer compensate for the natural leak rate of the membranes. It should be noted that in all cases after the addition of 1 M CCCP to dissipate the proton gradient, ACMA fluorescence returned to within 3% of its original value.
Arrhenius Analysis-Turnover rates at saturating ATP concentrations were measured as a function of temperature; because of their low membrane ATPase activities, the ⌬ 395 EDKL 398 and ⌬ 399 VVHR 402 deletion mutants were not included in this analysis. The results for the other deletion mutants and the wild-type enzyme are shown in Fig. 3

ATPase activities of membrane vesicles of deletion mutants
ATPase activities of membrane preparations of deletion mutants and controls were measured as described under "Experimental Procedures." Except where indicated, the assay medium did not contain CCCP. The positive control was strain pTR19-ASDS/DK8, which expresses Bacillus PS3 ATP synthase in E. coli; this strain served as background for the deletions. The negative control was strain pUC118/DK8, which expresses neither PS3 ATP synthase nor the endogenous E. coli enzyme. The values for the amount of F 1 F 0 in the membrane preparations were taken from Table  2. Turnover rates were calculated using a molecular mass of 531 kDa for the holoenzyme, taking into account the differing amounts of ATP synthase in the individual membrane preparations. The synthesis/hydrolysis ratio was obtained by dividing ATP synthesis activities from Table 2

␤DELSEED-loop of ATP Synthase
form of Arrhenius plots (ln k cat versus 1/T). From the slope of the regression lines, the activation energy (E a ) and changes of enthalpy (⌬H ‡ ), entropy (⌬S ‡ ), and Gibbs free energy of activation (⌬G ‡ ) for the rate-limiting step of the overall reaction can be calculated (see Table 4 for values at 50°C). First, it should be noted that the k cat for wild-type ATP synthase steadily increases, as expected, by a factor of Ͼ2 for each 10°C increase in temperature over the entire observation range. This is in contrast to a recent report (35) using the ␣ 3 ␤ 3 ␥ subcomplex of PS3 ATP synthase, which found, surprisingly, a virtually temperature-independent steady-state ATPase activity for the range between 25 and 45°C. The reason for this difference in behavior is not that, as in the case described here working with the membrane-embedded holoenzyme, a proton gradient is built up, which is obviously not possible using an ␣ 3 ␤ 3 ␥ subcomplex. The addition of CCCP to our experimental setup, to dissipate the proton gradient, did not alter the results significantly ( Table 4). As to the deletion mutants, their ⌬G ‡ values are close to the wild-type value. In contrast, there are profound differences in E a and ⌬H ‡ , and consequently, because ⌬G ‡ ϭ ⌬H ‡ Ϫ T⌬S ‡ , in T⌬S ‡ . The E a and ⌬H ‡ values for the mutants with deletions of 3-4 amino acids were all rather similar, ϳ70% of the respective wild-type value. The E a and ⌬H ‡ values for the 7-amino acid deletion mutant, ⌬ 381 QDIIAIL 387 , was even lower, ϳ50% of the wild-type values. Fig. 4 summarizes the data in the form of a free energy diagram, plotting ⌬H ‡ as a function of T⌬S ‡ . The results indicate that in the rate-limiting step of the overall reaction in the mutants less interactions have to be formed or broken than in the wild-type enzyme. In the enzyme with the longer deletion of 7 residues the number of remaining energetically relevant interactions was even smaller than in the mutants with the shorter deletions of 3-4 amino acid residues.
MgATP Binding to the Catalytic Sites in the ␣ 3 ␤ 3 ␥ Subcomplex-To measure nucleotide binding to the catalytic sites of the deletion mutants, we inserted the same set of depletions into strain pNM1/JM103⌬(uncB-uncD), which expresses an ␣ 3 (␤Y341W) 3 ␥ subcomplex of PS3 ATP synthase. Unfortunately, only two of the deletion mutants, ⌬ 388 GMDE 391 and ⌬ 392 LSD 394 , could be obtained as an ␣ 3 ␤ 3 ␥ subcomplex in sufficient yield and purity to allow fluorescence-based nucleotide binding studies. As expected from the membrane vesicle experiments, both mutant subcomplexes are active ATPases. The following activities were measured at 60°C: ␣ 3 (␤Y341W) 3 ␥, 31.4 units/mg; ⌬ 388 GMDE 391 , 7.9 units/mg; ⌬ 392 LSD 394 , 11.3 units/mg. Assuming a molecular mass of 352 kDa for the subcomplex, this corresponds to turnover rates of 184 s Ϫ1 , 46 s Ϫ1 , FIGURE 3. Arrhenius analysis of the ATPase activity of the deletion mutants. ATPase activities were measured at different temperatures (25-50°C) and were calculated as turnover rates (k cat , in s Ϫ1 ). From the negative slope of the regression lines, the activation energy, E a , of the rate-limiting step of the overall reaction can be determined (see Table 3

TABLE 4 Arrhenius analysis of ATPase activities of deletion mutants
ATPase activities were measured at temperatures between 25 and 50°C as described under "Experimental Procedures." Except where indicated, the assay medium did not contain CCCP. From the resulting Arrhenius plots (ln k cat versus 1/T), the activation energy, E a , and changes of enthalpy, ⌬H ‡ , entropy, ⌬S ‡ (expressed as T⌬S ‡ ), and Gibbs free energy of activation, ⌬G ‡ , for the rate-limiting step of the overall reaction at 50°C were calculated. ⌬⌬ values give the difference of the respective parameter between wild-type and mutant enzymes (or, for wild-type enzyme, between the activities in the absence and presence of CCCP). and 66 s Ϫ1 , respectively. Interestingly, the activities of both mutant subcomplexes were substantially below those for the parental ␣ 3 (␤Y341W) 3 ␥, whereas as F 1 F 0 holoenzyme in membranes the same mutants had approximately the same (⌬ 388 GMDE 391 ) or even a higher activity (⌬ 392 LSD 394 ) than the wild-type control (Table 3). Part, although probably not all, of this discrepancy can be related to the higher assay temperature for the ␣ 3 ␤ 3 ␥ subcomplexes (60 versus 42°C for the membranebound ATPase), considering the stronger temperature dependence of the activity of the wild-type enzyme as compared with that of the mutants (Fig. 3). MgATP binding was measured using the fluorescence of the inserted Trp ␤W341 as signal. The results are shown in Fig. 5, and K d values are given in Table 5. The ␣ 3 (␤Y341W) 3 ␥ subcomplex exhibited a binding pattern resembling that determined for E. coli F 1 (7,8,36) and F 1 F 0 (37), with three sites of different binding affinity for the Mg 2ϩ -nucleotide. The K d2 value for binding to the medium affinity catalytic site was somewhat higher than determined previously for the ␣ 3 ␤ 3 ␥ subcomplex of PS3 (38 -40), which might be because of differences in the buffer composition. The deletion mutants showed a pronounced change in the binding behavior. The binding affinity of the medium affinity site 2 decreased to the level of the low affinity site 3. More importantly, the affinity of the high affinity site 1 decreased by at least 2 orders of magnitude. Thus, the two deletion mutants no longer have a true "high affinity" site.

Strain/mutation
Modeling the Deletion Mutants-To understand the functional consequences of the deletions in terms of the enzyme mechanism on a residue level, we modeled the deletions into the available structures of the ␤ subunit (␤ TP , ␤ DP , ␤ HC , and ␤ E ). In the wild-type enzyme, in all three ␤ subunits the region of the DELSEED-loop investigated here (residues 380 -402 in the PS3 enzyme) makes contact with ␥ via hydrogen bonds/salt bridges as well as van der Waals and hydrophobic interactions (Fig. 6A). In all three ␤ subunits, ␤E391 is involved in these contacts. In ␤ TP , the carboxylate groups of ␤D390 and ␤E391 can hydrogen bond to the peptide bond nitrogen of ␥G86, whereas the carboxylate group of ␤D394 can form a hydrogen bond/salt bridge with ␥R94 (Interestingly, the ␤D394/␥R94 pair in the Bacillus enzyme is replaced by a Glu/Lys pair in ATP synthases from E. coli and bovine mitochondria; both amino acid pairs can cover approximately the same distance). Both ␥G86 and ␥R94 are located in a short helix that runs nearly perpendicular to the long N-and C-terminal helices of ␥. Furthermore, a hydrogen bond is possible between the peptide bond oxygen of ␤I386 and the side chain of ␥N247 in the C-terminal helix. In ␤ DP , the carboxylate group of ␤E391 can interact with the guanidino group of ␥R82 and the peptide bond nitrogen of ␥L84. In addition, a hydrogen bond/salt bridge can be formed between the carboxylate function of ␤D382 and the guanidino group of ␥R9 in the N-terminal helix. In the third ␤ subunit, at least in its half-closed form (11) (i.e. when a nucleotide is bound to the catalytic site), the side chains of ␤E391 and ␥R33 interact. In all three ␤ subunits, residues ␤I386 and ␤L387 are involved in van der Waals and hydrophobic interactions, and residue ␤I383 makes contact with ␥ in ␤ DP and ␤ HC .
Modeling of the deletion mutants indicated that the deletions leave the overall shape of the DELSEED-loop mostly intact (see Fig. 6B for selected cases). Even in the case of the 7-residue deletion, in all three ␤ subunits the loop still extends far enough to make contact with ␥. This is achieved mainly by "unraveling" of the helices flanking the turn or the short helical segment within the turn. Instead of a helical conformation, the respective amino acid stretches assume a straighter, more linear conformation, which can cover a longer distance. As of the residues investigated here, ␤E391 is involved in the largest number of ␤/␥ hydrogen bonds/salt bridges, it is not surprising that this number is most reduced in the ⌬ 388 GMDE 391 deletion mutant, from 7 to 8 in the wild-type enzyme to about 3 in the mutant. ⌬ 392 LSD 394 and ⌬ 381 QDIIAIL 387 have ϳ4 residual ␤/␥ hydrogen bonds/salt bridges. In ⌬ 392 LSD 394 , the deletion is just downstream of ␤E391 and changes the conformation of the glutamate side chain. ⌬ 381 QDIIAIL 387 experiences the most pronounced conformational rearrangements overall of all the mutants investigated here, due to the size of the deletion. As most hydrophobic interactions involve residues ␤I386, ␤L387,  Table 5.

DISCUSSION
The goal of the present study was to gain a better understanding of the role of the ␤DELSEED-loop in the coupling of catal-ysis and subunit rotation of ATP synthase. Thus far, mutational analysis of this region has focused on the negatively charged loop residues, such as ␤D382 (41), and those of the 390 DEL-SEED 396 motif itself, especially ␤E391 (12,14,42). To use a less restrictive approach, we constructed deletions of 3, 4, or 7 residues in the turn region of the loop of ATP synthase from the thermophilic bacterium Bacillus PS3. All deletions of 3-4 residues plus one of 7 residues could be expressed in E. coli and were incorporated into the membrane in sufficient amounts to allow functional analysis. The results show that the ␤DEL-SEED-loop plays an important role in the catalytic mechanism. It is not possible, however, to single out individual ␤/␥ interactions that might be essential for function. In contrast, there appears to be a multitude of relevant contacts, each of which makes a rather limited contribution. Even the ⌬ 381 QDIIAIL 387 mutant, where the majority of residues involved in contacts to ␥ is missing (such as ␤D382, ␤I383, ␤I386, and ␤L387, all of which are strictly conserved), is still functional to such an extent that one has to assume that it operates by the normal rotational mechanism.
All deletion mutants that we analyzed displayed the ability to hydrolyze ATP, and all except ⌬ 395 EDKL 398 and ⌬ 399 VVHR 402 showed ATP synthesis activity in vitro. ⌬ 395 EDKL 398 and ⌬ 399 VVHR 402 were also the only two mutants in which ATP hydrolysis did not result in the formation of a proton gradient. These deficiencies could be due to a combination of low abundance of enzyme on the membranes and low turnover rate, e.g. in ATP hydrolysis the membrane ATPase activity might be too low to compensate for the natural proton leak rate of the membranes. Alternatively, it is possible that these two mutants are uncoupled, although when NADH-driven proton pumping was measured using ACMA fluorescence, both mutants were able to generate a proton gradient of a magnitude comparable with that of the other mutants and the wild-type enzyme. In general, the deletion mutants seem to be less well coupled than the wild type, as evidenced by their less pronounced acceleration of membrane ATPase activities upon addition of an uncoupler and by their (with the exception of ⌬ 388 GMDE 391 ) reduced ATP synthesis/hydrolysis ratio. Further indication of possible uncoupling was found for the ⌬ 384 IAIL 387 and ⌬ 381 QDIIAIL 387 mutants upon studying ATP-driven H ϩ -pumping and growth yields in limiting glucose. Interestingly, the defect that abolished oxidative phosphorylation in vivo still allowed ATP synthesis in vitro.
The ATPase activities of the individual deletion mutants differ widely. ⌬ 395 EDKL 398 and ⌬ 399 VVHR 402 have the lowest activities. It should be noted that these two deletions are the only ones that do not eliminate residues that, according to the available crystal structures, make contact with ␥. Either mutation abolishes a turn in the C-terminal helix of the DEL-SEED-loop downstream of the ␥ interaction site. It seems possible that the deletions might affect the orientation of the side chain of residue ␤R404, which is part of a functionally important hydrogen-bonding network at the ␤ DP /␣ DP interface (43). It should be mentioned, however, that this hypothesis is not supported by the modeling results, which show the side chain of ␤R404 in the same position in the mutants and in the wild-type enzyme. The DELSEED-loops of the three ␤ subunits are shown in yellow, and ␥ is shown in blue. Amino acid residues involved in hydrogen bonds/salt bridges are shown in stick representation and labeled using PS3 residue numbers. Dots indicate the van der Waals surface of the side chains of residues ␤I383, ␤I386, and ␤L387 involved in contacts with ␥. This figure is based on Protein Data Bank file 1h8e (11), modified to represent the Bacillus PS3 ATP synthase using PRIME. B, models of the ␤DELSEED-loop of the deletion mutants. The main chain of the respective deletion mutant is shown as the thicker line in cartoon representation and the main chain of the wild-type enzyme as the thinner line. Side chains (in stick representation) are given only for the deletion mutants. Color-coding is as described in the legend for Fig. 1B. APRIL 24, 2009 • VOLUME 284 • NUMBER 17

JOURNAL OF BIOLOGICAL CHEMISTRY 11343
At the other extreme, as shown in Table 3, some of the mutants have a significantly higher ATPase activity than the wild-type enzyme (although if one extrapolates the Arrhenius plots, at the optimal growth temperature of Bacillus PS3, 65°C, these differences will be less pronounced). The increased activity could be due to perturbation of the interaction with the ⑀ subunit. ⑀ exists in two different conformations, "up" and "down." In the up conformation it acts as intrinsic inhibitor of ATPase activity. In the wild-type enzyme, the up conformation of ⑀ interacts with the ␤DELSEED-loop (44 -46). If the relevant interactions are reduced in the deletion mutants, this might shift the equilibrium to the non-inhibitory down conformation. Obviously, increased ATPase activities without concomitant increase in ATP synthesis activity, as observed here, results in reduced synthesis/hydrolysis ratios and therefore in decreased apparent coupling efficiencies. If the increased ATPase activities in the mutants should be due to changes in interaction with ⑀, as suggested here, this would support the notion of a role of ⑀ in coupling, as proposed for the PS3 as well as the E. coli enzyme (47,48).
The Arrhenius analysis reflects the reduction of energetically meaningful interactions between ␤ and ␥ in the deletion mutants. In the tested cases of mutants with deletions of 3-4 amino acids (excluding ⌬ 395 EDKL 398 and ⌬ 399 VVHR 402 ), the activation energy, E a , and consequently of ⌬H ‡ , was reduced by 18 -22 kJ/mol, in the mutant with 7 deleted residues by 36 kJ/mol. The results indicate a decreased number of bond rearrangements that contribute to the rate-limiting step of the overall reaction, strongly supporting the notion of a pivotal role for the DELSEED-loop in coupling catalysis and ␥ rotation. Although this interpretation seems straightforward, it is remarkable how well the decrease in E a and ⌬H ‡ correlates with the length of the deletion, seemingly independently of the character of the amino acids that were removed.
Although only two of the deletion mutants could be isolated in the form of an ␣ 3 ␤ 3 ␥ subcomplex suitable for a Trp fluorescence-based nucleotide binding assay, these studies gave a remarkable result. In both mutants, ⌬ 388 GMDE 391 and ⌬ 392 LSD 394 , the affinity of the high affinity site was drastically decreased. Either deletion removes a residue (␤E391 and ␤D394, respectively) that in the ␤ TP subunit, which carries the high affinity binding site, interacts with ␥. Interestingly, a similarly pronounced reduction in affinity of this site was recently described for a ␤D390C point mutation (42). Nevertheless, at this point in time it seems premature to ascribe the decrease in affinity to loss of a specific interaction between the ␤ TP DEL-SEED-loop and ␥. A 390 AALSAAA 396 quintuple point mutant, where all the negative charges of the DELSEED motif are removed (including those contained in ⌬ 388 GMDE 391 , ⌬ 392 LSD 394 , and ␤D390), could still bind MgATP and MgADP with high affinity. 4 It appears possible that the deletion mutants ⌬ 388 GMDE 391 and ⌬ 392 LSD 394 as well as the ␤D390C point mutation cause larger scale conformational rearrangements that affect nucleotide binding to the high affinity catalytic site. Modeling of the deletion mutants did not reveal a specific pattern of conformational changes common to both mutants.
The reduction in affinity of the high affinity site had been observed previously, but mostly for mutations of residues involved directly in substrate binding and/or turnover (49 -51). These mutants had very low residual enzymatic activity, and it is not clear whether they hydrolyze ATP by the normal rotational mechanism. In contrast, both deletion mutants described here, as well as the ␤D390C point mutation, display normal enzymatic activity using rotation, as evidenced by the observation of oxidative phosphorylation in vivo, ATP synthesis in vitro, and ATP-driven proton pumping. These findings seem to contradict the assumption contained in several models of the catalytic mechanism (see the Introduction) that tightening of the binding site around the newly bound MgATP, from "low affinity" to "high affinity," might be responsible for driving the 80°rotation substep. Using the data given in Table 5, the increase in MgATP binding energy upon transition from low affinity (K d3 ) to high affinity (K d1 ) drops from 20.3 kJ/mol in the wild-type PS3 enzyme to less than half that value in the deletion mutants (5.0 kJ/mol for ⌬ 388 GMDE 391 , 8.3 kJ/mol for ⌬ 392 LSD 394 ). A comparison of 1/K 1 observed under uni-site conditions (corresponding to K d1 measured here) and K m for multi-site catalysis (corresponding to K d3 ) in wild-type and mutant E. coli enzyme (42) gives a similar decrease for the ␤D390C mutant. However, before it can be concluded that the mechanistic models are incorrect in this aspect, significantly more information is needed about the changes of ligand binding affinities of the catalytic sites as a function of the rotational angle of ␥.