Beta subunit Glu-185 of Escherichia coli H(+)-ATPase (ATP synthase) is an essential residue for cooperative catalysis.

Glu-β185 of the Escherichia coli H+-ATPase (ATP synthase) β subunit was replaced by 19 different amino acid residues. The rates of multisite (steady state) catalysis of all the mutant membrane ATPases except Asp-β185 were less than 0.2% of the wild type one; the Asp-β185 enzyme exhibited 15% (purified) and 16% (membrane-bound) ATPase activity. The purified inactive Cys-β185 F1-ATPase recovered substantial activity after treatment with iodoacetate in the presence of MgCl2; maximal activity was obtained upon the introduction of about 3 mol of carboxymethyl residues/mol of F1. The divalent cation dependences of the S-carboxymethyl-β185 and Asp-β185 ATPase activities were altered from that of the wild type. The Asp-β185, Cys-β185, S-carboxymethyl-β185, and Gln-β185 enzymes showed about 130, 60, 20, and 50% of the wild type unisite catalysis rates, respectively. The S-carboxymethyl-β185 and Asp-β185 enzymes showed altered divalent cation sensitivities, and the S-carboxymethyl-β185 enzyme showed no Mg2+ inhibition. Unlike the wild type, the two mutant enzymes showed low sensitivities to azide, which stabilizes the enzyme Mg•ADP complex. These results suggest that Glu-β185 may form a Mg2+ binding site, and its carboxyl moiety is essential for catalytic cooperativity. Consistent with this model, the bovine glutamate residue corresponding to Glu-β185 is located close to the catalytic site in the higher order structure (Abrahams, J. P., Leslie, A. G. W., Lutter, R., and Walker, J. E.(1994) Nature 370, 621-628).

The H ϩ -ATPase (ATP synthase) of Escherichia coli synthesizes ATP similar to those of mitochondria or chloroplasts (see Refs. 1-4 for reviews). The catalytic site of the enzyme is in the ␤ subunit of the membrane extrinsic F 1 sector. Studies on mutant enzymes indicated that Lys-␤155 and Thr-␤156 in the ␤ subunit phosphate loop or conserved glycine-rich sequence (Gly-Gly-Ala-Gly-Val-Gly-Lys-Thr, residues 149 -156; conserved residues underlined) and Glu-␤181 and Arg-␤182 in the conserved Gly-Glu-Arg sequence (residues 180 -182) are essential catalytic residues (5)(6)(7). Affinity labeling with ATP analogues indicated that Lys-␤155 bound the ␤ and ␥ phosphate moiety of ATP (8). The crystal structure (9) of the bovine F 1 sector reported recently is essentially consistent with these results.
The purified F 1 (␣ 3 ␤ 3 ␥␦⑀ or F 1 -ATPase) hydrolyzes ATP through unisite (single site) or multisite (steady state) catalysis. The multisite rate is 10 5 -10 6 -fold faster than the unisite one due to the cooperativity of the multiple catalytic sites (10,11). Conformational transmission for cooperativity may be initiated from a specific region(s) or residue(s) in the single catalytic site of the ␤ subunit. Mutations near catalytic site residues often dramatically lower the multisite rate without changing unisite catalysis (6,11,12), possibly due to the defective conformational transmission between catalytic sites essential for the catalytic cooperativity. However, the role of a specific residue or region for the cooperativity has been questionable because all mutations so far introduced at a certain position did not always have the same effects on multisite catalysis. A typical example is the result of mutations at Gly-␤149 of the phosphate loop; the Ala-␤149 or Ser-␤149 enzyme exhibited similar ATPase activity to the wild type, whereas the Cys-␤149 enzyme had only 8% of the wild type ATPase activity (13), indicating that Gly-␤149 is not an essential residue for conformational transmission.
In this study, we were interested in conserved Glu-␤185, which is near essential catalytic residues (Glu-␤181 and Arg-␤182) described above and substituted it with 19 different residues. Surprisingly, all the mutants except Asp-␤185 exhibited no multisite catalysis (less than 0.2% of the wild type activity); Asp-␤185 had about 16% of the wild type membrane ATPase activity. Purified F 1 -ATPases with Asp-␤185, Gln-␤185, and Cys-␤185 residues showed unisite catalysis with rates of a similar order of magnitude to that of the wild type. The Cys-␤185 enzyme showed substantial multisite catalysis upon chemical modification with sodium iodoacetate (IAA). 1 These results clearly indicate that Glu-␤185 is the first residue identified as being absolutely essential for multisite catalysis. The roles of Glu-␤185 are discussed on the basis of the properties of the mutant enzymes.

EXPERIMENTAL PROCEDURES
E. coli and Growth Conditions-Strain DK8 (⌬unc B-C, ilv::Tn10, thi) (14) lacking the unc operon was used as a host for recombinant plasmids. A rich medium (with or without 50 g/ml ampicillin) supplemented with 50 g/ml thymine and a minimal medium containing 50 g/ml thymine, 2 g/ml thiamine, 50 g/ml isoleucine, 50 g/ml valine, and 5 mM glucose (or 15 mM succinate) were used (15). Minimal medium with 0.5% glycerol was used for preparing membranes.
Construction of Recombinant Plasmids Carrying the unc Operon with Mutations at Position 185 of the ␤ Subunit-Recombinant plasmids carrying mutations at position 185 of the ␤ subunit were constructed using pUDSE709 (13). The following mutations were introduced by replacing the XhoI-ClaI segment of pUDSE709 with the desired synthetic double stranded DNA: Gly (GGT), Ala (GCG), Ser (TCC), Thr (ACC), Asp (GAC), Asn (AAC), Gln (CAG), His (CAC), Lys (AAA), Arg (CGT), Cys (TGT), Tyr (TAC), Phe (TTC), Leu (CTG), Ile (ATC), Val * This work was supported by grants from the Japanese Ministry of Education, Science, and Culture and the Human Frontier Science Program. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
‡ Supported by a Postdoctoral Fellowship from the Japan Society for the Promotion of Science. § Present address: Laboratory of Biochemistry, Faculty of Pharmaceutical Sciences, Osaka University, Suita, Osaka 567, Japan.
Modification of Cys-␤185 F 1 -ATPase with IAA-Purified Cys-␤185 F 1 -ATPase was passed through a centrifuge column (Sephadex G-50, 0.4 ϫ 6 cm) (17) equilibrated with 10 mM HEPES-NaOH, pH 8.0, to remove dithiothreitol and ATP included in the purified enzyme solution and then incubated with 100 M IAA in 50 mM HEPES-NaOH, pH 8.0, and 20 mM MgCl 2 for 2 h at 30°C in the dark. The reaction was terminated by 200-fold dilution with a buffer (2 mM Tris-HCl, pH 8.0, 2 mM dithiothreitol, and 2 g/ml bovine serum albumin) or by removal of excess IAA using a centrifuge column.
Other Procedures-Membrane vesicles were prepared as described elsewhere (18) using 10 mM Tris-HCl buffer, pH 8.0, containing 140 mM KCl, 0.5 mM dithiothreitol, 10% glycerol, 0.5 mM phenylmethanesulfonyl fluoride, 5 g/ml leupeptin, and 5 g/ml pepstatin A. The Asp-␤185 and wild type F 1 -ATPases were purified as described previously (19). The Gln-␤185 and Cys-␤185 F 1 -ATPases were purified by the same procedure as for the wild type except that all column chromatography was carried out at room temperature. The enzyme contained about 0.5 mol of the ␦ subunit/mol of protein. ATPase activities were assayed at 37°C in 20 mM Tris-HCl, pH 8.0, 4 mM ATP, and 2 mM MgCl 2 (19) unless otherwise specified. One unit of the enzyme was defined as the amount hydrolyzing 1 mol of ATP/min at 37°C under the above conditions. When indicated, varying concentrations of MgCl 2 or CaCl 2 were included in the reaction mixture.
Unisite catalysis was assayed using 0.25 M [␥-32 P]ATP and 0.5 or 10 mM MgSO 4 at 25°C (20 -23). The ATP binding rate (k ϩ1 ) was measured as the decrease of ATP in the medium using hexokinase and glucose (5). The rate of ATP synthesis was assayed at 25°C by the published method (16). Protein concentration measurement (24) using bovine serum albumin as a standard and polyacrylamide gel electrophoresis (25) were described previously.

Properties of 19 Mutants at Position 185 (Glu, wild type) of the ␤ Subunit-Systematic mutagenesis between Thr-␤156 and
Lys-␤201 of the ␤ subunit indicated that Glu-␤181 and Arg-␤182 are essential for catalysis (6). We were interested in the conserved Glu-␤185 residue located near these residues and replaced it with 19 different residues including Gln. All the mutants except Asp-␤185 could not grow on succinate through oxidative phosphorylation, although they exhibited substantial F 0 F 1 assemblies in membranes (Table I). The Asp-␤185 mutant showed essentially the same growth yield as the wild type. Eight mutants (Tyr-␤185, Asn-␤185, Thr-␤185, Arg-␤185, His-␤185, Cys-␤185, Val-␤185, and Ile-␤185) exhibited about 50% of the wild type assembly, whereas the others exhibited essentially similar assembly to that in the case of the wild type. The low degrees of assemblies in the eight mutants may suggest that Glu-␤185 is located in the critical region for interaction of the ␤ subunit with other subunit(s). The importance of Glu-␤185 for subunit assembly was suggested previously by the fact that the isolated Gln-␤185 and Lys-␤185 ␤ subunits could not form an ␣ 3 ␤ 3 ␥ complex in vitro (26).
Asp-␤185 exhibited about 16% of the wild type membrane ATPase activity, whereas other mutants exhibited no membrane ATPase activity (less than 0.2% of wild type multisite catalysis). Membrane ATPase of the Cys-␤185 mutant became detectable after incubation with IAA but not with the same concentration of iodoacetoamide; activity of about 0.05 units/mg protein became detectable after incubation of Cys-␤185 membranes with 100 M IAA in 50 mM Tris-HCl, pH 8.0, at room temperature for 10 min (ATPase activity of Cys-␤185 F 1 -ATPase without IAA treatment, about 0.01-0.02 units/mg). This result suggests that the ␤S-carboxymethylated enzyme has activity. Detailed studies of the IAA effects were then carried out below using purified Cys-␤185 F 1 -ATPase. Consistent with the low membrane ATPase activity and negative growth by oxidative phosphorylation, mutant membranes (Gln-␤185 or Cys-␤185) did not show significant ATP synthesis (Table II, right column). Membranes treated with IAA also did not show ATP synthesis, because the mutant membranes lost respiration-driven proton transport after IAA treatment (data not shown). A similar IAA effect was observed for wild type membranes.
Properties of Purified Mutant Enzymes-Three mutant F 1 -ATPases (Asp-␤185, Gln-␤185, and Cys-␤185) were purified using a procedure developed for the wild type; they behaved similarly during column chromatographies and showed essentially the same recoveries as that of the wild type (about 50% from the EDTA extract). The Gln-␤185 and Cys-␤185 enzymes showed no multisite catalysis with ATP (Յ0.1% of the wild type level), ITP, or GTP as a substrate (with MgCl 2 or CaCl 2 as a divalent cation). Both enzymes did not show ATPase activity after incubation with pyruvate kinase and phosphoenolpyruvate to remove endogenous exchangeable ADP (Ref. 27 and data not shown), suggesting that the enzymes are not in the highly inhibited state with Mg⅐ADP. On the other hand, the Asp-␤185 enzyme showed about 15% of the wild type rate, similar to the membrane enzyme.
The mutant enzymes showed unisite catalysis with initial rates of about 50 (Gln-␤185), 60 (Cys-␤185), and 130% (Asp-␤185) of that of the wild type, and Asp-␤185 F 1 -ATPase exhibited a k ϩ1 value (rate of ATP binding) of a similar order of magnitude as that of the wild type (Table II). The k ϩ1 values for Gln-␤185 and Cys-␤185 were slightly lower than that of the wild type. The wild type and all the mutant enzymes except for Cys-␤185 F 1 -ATPase showed cold chase in unisite catalysis, consistent with the partial release of the ␦ subunit from F 1 during purification (23). These results clearly indicate that the major defect of the mutant enzymes is not in the catalytic reaction itself but in the catalytic cooperativity required for multisite catalysis.
Activation of the Purified Cys-␤185 F 1 -ATPase with IAA-Multisite catalysis of the purified Cys-␤185 F 1 -ATPase was very low but became detectable when it was incubated with IAA (Table II). The activation was dependent on the IAA concentration; about 2 mol/mg⅐min ATPase activity/mg protein  (Fig. 1a,  closed circles). The activation increased dramatically with the addition of MgCl 2 ; maximal activity (10 mol/mg⅐min protein) was obtained with 100 M IAA and 20 mM MgCl 2 (Fig. 1a, open circles; Fig. 1b). CaCl 2 had less effect on the activation by IAA; the maximal activity obtained was about 4 mol/mg⅐min protein (Fig. 1b, diamonds). These results indicate that the Cys-␤185 residue became more reactive to IAA upon the addition of MgCl 2 . The maximal activity obtained corresponds to about 30% of that of the wild type. Similar activation was not observed with other sulfhydryl reagents (1 mM), such as iodoacetamide, dithiobis(2-nitrobenzoic acid), 4-chloro-7-sulfobenzofurazan, and N-ethylmaleimide, indicating that the carboxyl moiety introduced at position 185 after incubation with IAA is essential for enzyme activation. The initial rate and k ϩ1 of unisite catalysis by the S-carboxymethyl-␤185 enzyme were about 1 ⁄4 and 1 ⁄10 of those of the wild type enzyme, respectively (Table II). These results clearly indicate that the increased multisite catalysis described above was not due to the increased rate of unisite catalysis. Properties of S-Carboxymethyl-␤185 F 1 -ATPase-As shown in Fig. 2, the [ 14 C]carboxymethyl moiety was incorporated into the Cys-␤185 enzyme when it was incubated with sodium [2-14 C]iodoacetate, whereas no radioactivity was incorporated into the wild type ␤ subunit. These results indicate that the [ 14 C]carboxymethyl moiety was incorporated into position 185 of the mutant. The amounts of carboxymethyl residue incorporated into the Cys-␤185 enzyme were 2.7 and 2.1 mol/mol of F 1 for the mutant F 1 incubated with 100 M IAA in the presence and absence of 20 mM MgCl 2 , respectively. These results suggest that F 1 -ATPase with S-carboxymethyl-␤185 in three ␤ subunits exhibited dramatically increased multisite catalysis activity, and the low activity observed after incubation with IAA in the absence of MgCl 2 (Fig. 1a) may be due to the small amount of F 1 incorporated into the three carboxymethyl moieties.
The Asp-␤185 and S-carboxymethyl-␤185 enzymes showed altered requirements for MgCl 2 (Fig. 3). The wild type enzyme showed the highest activity with 2 mM MgCl 2 and 37% maximal activity with 10 mM MgCl 2 when assayed in the presence of 4 mM ATP, confirming previous results (28,29). On the other hand, the S-carboxymethyl-␤185 enzyme showed maximal activity with 10 mM and only 6% activity with 2 mM MgCl 2 . It was of interest that S-carboxymethyl-␤185 enzyme did not show Mg 2ϩ inhibition, although the wild type enzyme was inhibited with a higher Mg 2ϩ concentration. On the other hand, unisite catalysis of the S-carboxymethyl-␤185 enzyme showed similar Mg 2ϩ dependence to that of the wild type (slightly lower rate in 10 mM MgSO 4 ; data not shown), suggesting that Mg 2ϩ had different effects on the catalytic cooperativities of the wild type and S-carboxymethyl-␤185 enzymes. The Asp-␤185 enzyme showed maximal activity with 6 mM MgCl 2 and was slightly inhibited with a higher concentration. It is noteworthy that the S-carboxymethyl-␤185 and Asp-␤185 enzymes had very low ATPase activities, which were dependent on Ca 2ϩ ; the Ca 2ϩdependent activity of the mutant enzymes was only 6 -9% of the Mg 2ϩ -dependent activity when 10 -20 mM of the divalent cations were used. These results suggested that Glu-␤185 or its vicinity is closely related to the Mg 2ϩ binding required for multisite catalysis and that the length of the side chains for the carboxyl moiety affected the divalent cation dependence of the catalysis.
Effect of Azide on the S-Carboxymethyl-␤185 and Asp-␤185 Enzymes-Azide is known to inhibit multisite catalysis of F 1 -ATPase by stabilizing F 1 ⅐Mg⅐ADP complex (30), and its inhibition is dependent on the Mg 2ϩ concentration, although azide has no effect on unisite catalysis (22). Thus it was reasonable to assume that multisite catalyses with the S-carboxymethyl-␤185 and Asp-␤185 enzymes may exhibit different azide sensitivities from that of the wild type because the two enzymes exhibited altered Mg 2ϩ inhibition (Fig. 3). As expected, the S-carboxymethyl-␤185 and Asp-␤185 enzymes retained most of their activities even in the presence of 1 mM sodium azide, whereas the wild type activity was completely inhibited (Fig.  4). These results indicate that the mutant enzymes became 100-1000-fold less sensitive to azide, which is consistent with the notion that the Glu-␤185 residue is closely related to the Mg 2ϩ site.
The S-carboxymethyl-␤185 and Asp-␤185 enzyme became highly sensitive to salts such as LiCl, NaCl, KCl, or Na 2 SO 4 . About 70 and 90% of the activities of S-carboxymethyl-␤185 and Asp-␤185 enzymes were inhibited, respectively, by 150 mM LiCl (Fig. 5), whereas less than 10% of the wild type enzyme was inhibited. Similar results were obtained with other salts. Therefore, the inhibition of the mutant ATPase activities with high NaN 3 might be due to the effect of the high salt concentrations.

DISCUSSION
Extensive mutagenesis studies on F 1 -ATPase showed that the Lys-␤155 and Thr-␤156 residues of the phosphate loop (5) and Glu-␤181 (6, 7) and Arg-␤182 (6) of the conserved Gly-Glu-Arg (positions 180 -182) sequence are essential residues for uni-and multisite catalysis. Thus, the roles of other residues near the phosphate loop and the Gly-Glu-Arg sequence are of interest. We were interested in the Glu-␤185 residue, which is TABLE II Properties of mutant enzymes at position 185 of the ␤ subunit Mutant or wild type F 1 -ATPases were purified, and their catalytic activities were assayed at 25°C. The Cys-␤185 enzyme was incubated for 30 min with 100 M IAA in the presence of 20 mM MgCl 2 , and the S-carboxymethyl-␤185 enzyme (about 3 mol of carboxymethyl residues/mol of F 1 ) was obtained after removal of excess IAA on a centrifuge column. The reaction mixture for multisite catalysis contained 50 mM Tris-SO 4 , pH 8.0, 10 mM ATP, and 5 mM MgSO 4 . As the mutant and wild type enzymes showed different sensitivities to Mg 2ϩ (Fig. 3), the mutant/wild type ratios of multisite catalysis with high MgCl 2 were different; the rates of multisite catalysis with 20 mM MgCl 2 are shown in parentheses. Membrane ATP synthesis was assayed at 30°C. The rates of unisite catalysis obtained in the presence of 0.5 mM MgSO 4 are shown. The initial rates and k ϩ1 values were slightly low (k ϩ1 values, about 60, 70, and 40% for Cys-␤185, S-carboxymethyl-␤185, and the wild type, respectively) when assayed with 10 mM MgSO 4 . conserved in all the ␤ subunits so far sequenced (57 different species; SWISS PROT Release 30). It was surprising to find that all the mutants except Asp-␤185 were unable to grow by oxidative phosphorylation and exhibited no functional multisite catalysis. The purified Gln-␤185 and Cys-␤185 F 1 -ATPases also exhibited no multisite catalysis. Cross and co-workers (27) showed recently that E. coli F 1 -ATPase, similar to chloroplast or mitochondrial F 1 (30), is inhibited by the catalytic site-bound Mg⅐ADP. They proposed that the effect of Mg⅐ADP should be considered before kinetic results are interpreted. However, we think that the possibility of highly increased Mg⅐ADP inhibition of mutant enzymes is low because phosphoenolpyruvate and pyruvate kinase (treatment to release Mg⅐ADP) did not increase the activities of the Gln-␤185 and Cys-␤185 F 1 -ATPases. Furthermore, the S-carboxymethyl-␤185 and Asp-␤185 enzymes were not inhibited by Mg 2ϩ , as discussed below.
Despite the absence of multisite catalysis, the purified mutant F 1 -ATPases (Gln-␤185 and Cys-␤185) retained substantial unisite catalysis. Furthermore, multisite catalysis of the Cys-␤185 enzyme was recovered on the introduction of a carboxymethyl group after treatment with IAA, whereas the same treatment did not increase the unisite catalysis of the enzyme. Taken together with the observation of Asp-␤185 mutant, these results indicate that the carboxyl moiety at position 185 is required for catalytic cooperativity. It is noteworthy that Glu-␤185 is the first residue found to be essential for multisite catalysis. Similar residues were not identified previously because multisite catalysis was lost to varying degrees depending on the residues substituted (16,31).
MgCl 2 had a dramatic effect on the activation of Cys-␤185 F 1 -ATPase with IAA; ATPase activity obtained with MgCl 2 was about 5-fold higher than that on incubation without it. About 3 and 2 mol of S-carboxymethyl residues were incorporated into the mutant enzyme, respectively, on incubation with and without MgCl 2 , respectively. Thus, all three Cys-␤185 residues bound carboxymethyl moieties in the presence of Mg 2ϩ and became fully active, consistent with the requirement of three active ␤ subunits for multisite activity (11).
The S-carboxymethyl-␤185 and Asp-␤185 enzymes had in- teresting properties. Their ATPase activities showed divalent cation dependences different from those of the wild type: the two mutant enzymes required more MgCl 2 for maximal multisite catalysis than the wild type and exhibited very low CaCl 2dependent activity. Interestingly, S-carboxymethyl-␤185 enzyme activity is accelerated by excess MgCl 2 (4 mM ATP and 10 mM MgCl 2 ), suggesting the importance of free Mg 2ϩ ion. On the other hand, the divalent cation requirements of the mutant enzymes for unisite catalysis were similar to those of the wild type (data not shown). Thus, a change in the side chain length of the carboxyl moiety (Asp, Glu, and S-carboxymethyl) at position 185 affected the divalent cation requirement for multisite catalysis. These results suggest that the carboxyl group of the Glu-␤185 residue may be close to Mg 2ϩ at the catalytic site or forming the Mg 2ϩ binding site. The bovine glutamate (position 192) residue corresponding to E. coli Glu-␤185 is actually located in the catalytic site close to the Mg 2ϩ ion in the x-ray structure of bovine F 1 -ATPase (9). Thus, we propose that the Glu-␤185 residue contributes to the catalytic cooperativity through Mg 2ϩ binding. In this regard, Weber and co-workers reported that the cooperativity for ATP binding is dependent on Mg 2ϩ (32).
In contrast to the strong inhibition of the wild type enzyme by MgCl 2 at higher than 3 mM (about 60% inhibition with 5 mM MgCl 2 ), excess MgCl 2 did not inhibit the multisite catalysis of S-carboxymethyl-␤185 and only slightly inhibited the Asp-␤185 enzyme (about 10% inhibition with 10 mM MgCl 2 ). The Mg 2ϩ inhibition of the ATPase activity of the wild type enzyme was shown to be due to the Mg⅐ADP binding to the catalytic site (27). Similar to wild type enzyme, S-carboxymethyl-␤185 and Asp-␤185 enzyme retained about five bound nucleotides detected after passing through a centrifuge column (data not shown). Thus, the low Mg 2ϩ inhibition of the mutant enzymes suggests that the affinity of the Mg 2ϩ ion to the catalytic site bound ADP was lower in the mutant than in the wild type. In addition, the azide sensitivities of the S-carboxymethyl-␤185 and Asp-␤185 enzymes were decreased by more than 2 orders of magnitude. Azide inhibits F 1 -ATPase by stabilizing the enzyme-Mg⅐ADP complex (27,30), suggesting that the low azide sensitivity of Asp-␤185 or S-carboxymethyl-␤185 is because the mutant enzyme-Mg⅐ADP complex is not stabilized by azide. Previously, we reported that azide did not inhibit unisite catalysis (22). Thus, azide may change the environment around the Mg 2ϩ ion binding site including the Glu-␤185 residue, resulting in strong inhibition of the catalytic cooperativity through stabilization of the enzyme-Mg-ADP complex. Asp-␤185 (squares)) and wild type (circles) F 1 -ATPases were assayed with varying concentrations of LiCl in the presence of 4 mM ATP and 10 mM MgCl 2 . The results are expressed as relative rates of percentage of control (without LiCl). The control rates were given in the legend to Fig. 4.