Novel Insights into the Chemical Mechanism of ATP Synthase

The chemical mechanism by which the F1 moiety of ATP synthase hydrolyzes and synthesizes ATP remains unknown. For this reason, we have carried out studies with orthovanadate (Vi), a phosphate analog which has the potential of “locking” an ATPase, in its transition state by forming a MgADP·Vi complex, and also the potential, in a photochemical reaction resulting in peptide bond cleavage, of identifying an amino acid very near the γ-phosphate of ATP. Upon incubating purified rat liver F1 with MgADP and Vi for 2 h to promote formation of a MgADP·Vi-F1 complex, the ATPase activity of the enzyme was markedly inhibited in a reversible manner. When the resultant complex was formed in the presence of ultraviolet light inhibition could not be reversed, and SDS-polyacrylamide gel electrophoresis revealed, in addition to the five known subunit bands characteristic of F1 (i.e. α, β, γ, δ, and ε), two new electrophoretic species of 17 and 34 kDa. Western blot and N-terminal sequencing analyses identified both bands as arising from the β subunit with the site of peptide bond cleavage occurring at alanine 158, a conserved residue within F1-ATPases and the third residue within the nucleotide binding consensus GX 4GK(T/S) (P-loop). Quantification of the amount of ADP bound within the MgADP·Vi-F1 complex revealed about 1.0 mol/mol F1, while quantification of the peptide cleavage products revealed that no more than one β subunit had been cleaved. Consistent with the cleavage reaction involving oxidation of the methyl group of alanine was the finding that [3H] from NaB[3H]4 incorporates into MgADP·Vi-F1 complex following treatment with ultraviolet light. These novel findings provide information about the transition state involved in the hydrolysis of ATP by a single β subunit within F1-ATPases and implicate alanine 158 as residing very near the γ-phosphate of ATP during catalysis. When considered with earlier studies on myosin and adenylate kinase, these studies also implicate a special role for the third residue within the GX 4GK(T/S) sequence of many other nucleotide-binding proteins.

Despite our extensive knowledge about the structure and function of the F 1 moiety of ATP synthases (1)(2)(3)(4)(5), sufficient information is not available to write a chemical mechanism by which ATP is hydrolyzed and synthesized. In contrast, myosin-ATPase has been successfully studied using orthovanadate, V i , 1 a phosphate analog which in the presence of MgADP forms a transition state MgADP⅐V i -myosin inhibitory complex (6 -8).
Irradiation of this complex with uv light results in the modification of the single serine within the nucleotide binding consensus GESGAGKT followed by peptide bond cleavage at this site ( Fig. 1A) (8,9). These studies strongly implicated this serine as contacting directly the ␥-phosphate of ATP in the transition state, and were recently confirmed by x-ray structural analysis (10). The reaction pathway of F 1 -ATPase is believed to be quite similar to that of myosin-ATPase (11). In addition, within the catalytic sites of both enzymes resides the nucleotide binding consensus GX 4 GKT (P-loop) (12), which in the ␤-subunit of F 1 (GGAGVGKT), contains alanine in place of the internal V i -sensitive serine characteristic of the myosin consensus (GESGAGKT). As this serine in myosin is known to contact the ␥-phosphate of ATP in the transition state (10), it seemed reasonable to assume that in F 1 a nearby serine residing outside the consensus region or the terminal threonine within the consensus region may serve this role in the transition state. Alternatively, the third position within the consensus region of F 1 , although containing an alanine, may play an important role in the transition state of F 1 as does the serine in the same position in myosin. Studies described below were carried out both to define optimal conditions for trapping F 1 -ATPase in a MgADP⅐V i -F 1 inhibitory transition state complex, and to establish in this state the identity of an amino acid residue near the ␥-phosphate of ATP. Both goals were accomplished and provide novel insights into the chemical mechanism of ATP synthases.

Materials
Rats (Harlan Sprague-Dawley, white males) were obtained from Charles River Breeding Laboratories. ATP, ADP, MgCl 2 , MOPS, CAPS, sodium orthovanadate, phosphoenolpyruvate, pyruvate kinase, and lactic dehydrogenase were obtained from Sigma. SDS, acrylamide, and bisacrylamide were from Bio-Rad. Ammonium sulfate and potassium phosphate were from J. T. Baker Chemical Co. Tuberculin syringes and Sephadex G-50 used in nucleotide binding assays were from Becton-Dickinson Co. and Pharmacia Biotech Inc., respectively. PVDF membranes were obtained from Millipore and Western blot reagents from Amersham. A polyclonal antibody against the rat F 1 ␤-subunit was raised in rabbits using the synthetic peptide KIGLFGGAGVGKCT.

Methods
Purification of Rat Liver F 1 -ATPase-The enzyme was purified by the procedure of Catterall and Pedersen (13) with the modification described by Pedersen et al. (14). The purified enzyme, in 250 mM KP i and 5.0 mM EDTA, was divided into 100-l aliquots, lyophilized to dryness, and stored at Ϫ20°C. Prior to use, aliquots (150 -250 g) of lyophilized F 1 were dissolved at 25°C in 100 l of water and precipitated twice with 3 M ammonium sulfate, 5 mM EDTA, redissolving between precipitations in 200 mM K 2 SO 4 , 10 mM Tris-Cl, pH 7.5, or 50 mM MOPS, pH 8.0.
Preparation of Orthovanadate (V i ) Solution-To minimize the presence of polymeric species the following protocol was followed: Na 3 VO 4 powder was dissolved in water and the pH adjusted with HCl to pH 10 (orange color). The solution was boiled for 2 min at which time the solution became clear. The pH was readjusted to pH 10 and the previous boiling repeated 2 times. After the optical density was determined at 265 nm, the V i concentration was calculated using the molar extinction coefficient of 2925 M Ϫ1 cm Ϫ1 . The stock solutions used in this study were 155 mM and were covered with aluminum foil and stored until use at Ϫ80°C.
Prior Treatment of F 1 with V i -F 1 (50 g) was priorly incubated in a 100-or 200-l system containing 50 mM MOPS, pH 8.5, 10% glycerol (v/v) and, where indicated, also V i , V i ϩ ADP, V i ϩ ADP ϩ MgCl 2 , V i ϩ ATP ϩ MgCl 2 , or MgCl 2 , all at concentrations indicated in legends. Prior incubations were carried out at 25°C for the indicated times. In experiments where photoactivation of vanadate was induced with uv light (320 nm), the incubation mixture in an open Eppendorf tube was placed under a 100 watt, long wavelength mercury spot lamp (BLAK-RAY (Model B 100A, 115 V, 2.5 amperes; San Gabriel, CA) at a distance of 7.8 cm. The time in the presence of the light source varied as indicated in the figure legends.
Assay for ATPase Activity-The spectrophotometric procedure was used in which ADP formed was coupled to the pyruvate kinase and lactic dehydrogenase reactions (13). The reaction mixture contained the following in a volume of 1 ml at pH 7.5 and 25°C: 0.2 mM ATP, 65 mM Tris-Cl, 4.8 mM MgCl 2 , 2.5 mM KP i , 0.40 mM NADH, 0.60 mM phosphoenolpyruvic acid, 5 mM KCN, 1 unit of lactic dehydrogenase, 1 unit of pyruvate kinase, and 1.5 g of F 1 .
Assay for ADP Binding-Binding assays were carried out at 25°C by incubating F 1 , or fractions derived therefrom, for 20 min in a final volume of 100 l, containing concentrations of F 1 , [ 3 H]ADP, MgCl 2 , V i , and buffer as indicated in the legend to Fig. 4. The entire reaction mixture was loaded onto a Sephadex G-50 "fine" column (1-cm 3 tuberculin syringe with a filter at the bottom), which had been pre-equilibrated with 50 mM Tris-Cl, pH 7.6, and priorly centrifuged for 1.5 min at 2,500 rpm in an IEC model HN-SII clinical centrifuge. Centrifugation of the reaction mixture was carried out for 1.5 min at 2,500 rpm to separate nucleotide bound to F 1 from free nucleotide (15,16).
SDS-PAGE-This was carried out in a Bio-Rad Mini-Protean dual slab cell in 15% acrylamide according to the method of Laemmli (17), or in cylindrical glass tubes in 5% acrylamide according to the method of Weber and Osborn (18). Where indicated, densitometric analysis of the Coomassie-stained bands was carried out using a Fujifilm Bas-1500 PhosphorImager and MacBas (V 2.31) software.
Western Blot Analysis-After conducting SDS-PAGE, the proteins on the gel were transferred electrophoretically onto a PVDF membrane (1 h at 100 volts and 0.2 amp at 4°C in 10 mM CAPS, 10% methanol transfer buffer, pH 11). The product was then blocked for 1 h with 2% bovine serum albumin plus 5% nonfat dry milk in PBS-T (80 mM Na 2 HPO 4 , 20 mM NaH 2 PO 4 , 100 mM NaCl, 0.1% Tween 20, pH 7.5), incubated for 1 h at 23°C with a rat liver F 1 -␤ subunit polyclonal antibody (see "Experimental Procedures"), and then further incubated for 1 h at 23°C with the secondary antibody (horseradish peroxidaseconjugated anti-mouse IgG). The immunoreactive bands were detected by the enhanced chemiluminescene (ECL) system of Amersham Life Sciences.
N-terminal Sequence Analysis-F 1 ␤-subunit peptide fragments were transferred from SDS-PAGE gels onto PVDF membranes by electroblotting. Transfer conditions were for 1 h in 10 mM CAPS buffer, 10% methanol, pH 11, at 4°C in the case of the 17-kDa F 1 -␤ subunit fragment, and 2 h in the same buffer at 25°C in the case of the 34-kDa fragment. Only under the latter conditions could an N-terminal sequence be obtained with the 34-kDa fragment. The peptides were then excised and subjected to N-terminal sequencing (19) using an Applied Biosystems 475A Protein Sequencing System (20).
Determination of Protein-Protein was determined by the method of Lowry et al. (21) after first precipitating with 5% trichloroacetic acid.

F 1 Forms an Inhibitory Complex in the Presence of MgCl 2 , ADP, and V i -To promote formation of an inhibitory
MgADP⅐V i -F 1 transition state complex, several precautions were taken. First, the V i stock solution was priorly treated exactly as described under "Experimental Procedures" and maintained at pH 10 in the dark at Ϫ80°C until use to prevent formation of polymeric species. Second, the zwitterionic buffer MOPS was used in all experiments as it both stabilizes F 1 and prevents or minimizes inhibition by MgCl 2 and the product MgADP (or ADP). Third, a pH of 8.5 was used in all experiments, as this pH is near optimal for the hydrolysis of ATP catalyzed by rat liver F 1 . Finally, equimolar amounts of MgCl 2 , ADP, and V i were used to promote formation of MgADP⅐V i at the active site of F 1 . Fig. 2A summarizes results of an experiment where F 1 was priorly incubated in the presence of 200 M each of MgCl 2 , ADP, and V i for the indicated times and then assayed for ATPase activity as described under "Experimental Procedures." Here it is clear that under these conditions, which are optimal for forming a MgADP⅐V i -F 1 transition state complex, F 1 -ATPase activity is markedly inhibited. Half-maximal inhibition is reached in about 45 min, and maximal inhibition (ϳ80%) is reached in about 2 h. In control experiments with F 1 alone, F 1 ϩ V i , F 1 ϩ ADP, and F 1 ϩ MgADP, inhibition at 2 h is, respectively, Ͻ1, 10, 20, and 20% under conditions described in the figure legend. In an experiment not presented, F 1 prior incubated with MgCl 2 alone, was not inhibited after 2 h. Finally, Fig. 2B shows that when F 1 is prior incubated with 200 M each of MgCl 2 , ADP, and V i , but in the presence of light (320 nm) and atmospheric oxygen, an essentially identical inhibi-FIG. 1. A, simplified scheme depicting how orthovanadate (V i ) interacts with nearby serine (or threonine) residues within proteins in an oxidation reaction activated by uv light. The dotted boxes represent a serine within a protein prior to the reaction (left), and the products of the reaction (middle and right). Following the uv and oxygen-dependent reactions, which result in peptide bond cleavage at two places in the serine residue, the amino group of the serine is covalently attached to the preceding amino acid in the protein, the side chain is oxidized to formic acid, and the ␣-carbon and carbonyl group are oxidized to the oxalyl moiety which remains at the N terminus of the subsequent amino acid. (See Ref. 27 for a more detailed description of the chemistry.) B, amino acid sequence of the rat liver F 1 -␤ subunit deduced from cDNA showing the Walker A and B nucleotide binding consensus sequences (boxes). Threonine 163 which terminates the A consensus is depicted with an asterisk. The predicted catalytic base, glutamate 188, which lies between the A and B consensus regions, is underlined. The first Nterminal amino acid detected in the ␤ subunit of isolated rat liver F 1 is alanine. (See underlined region.) tion profile is observed.
Inhibition of F 1 -ATPase Activity by MgCl 2 ϩ ADP ϩ V i Is Reversible in the Absence of UV Light but Not in Its Presence-If formation of a MgADP⅐V i -F 1 transition state complex is responsible for the inhibition observed in the presence of MgCl 2 ϩ ADP ϩ V i , this state should be reversible. Results presented in Fig. 3A show that this is the case. Thus, following maximal inhibition of F 1 -ATPase activity in the presence of 200 M each of MgCl 2 , ADP, and V I (Fig. 3A, left panel), an aliquot was removed and subjected to two dilution/wash cycles in the presence of 50 mM MOPS, 10% glycerol, pH 8.5. This resulted after two such cycles in the restoration of the original activity to a level near 90% (Fig. 3A, right panel). Significantly, reversal of ATPase activity could not be achieved (Fig. 3B, right panel) following inhibition of F 1 -ATPase activity under identical conditions but in the presence of uv light (Fig. 3B, left panel) and with longer incubation times. The reason for this becomes clear in the description of other experiments described below.
V i Induces Inhibition of F 1 -ATPase Activity under Turnover Conditions-It is expected that if a MgADP⅐V i -F 1 transition state is formed in the presence of MgCl 2 , ADP, and V i as implicated from the above studies, it would be formed also under turnover conditions, i.e. when ATP, the substrate for ATP hydrolysis is present. For this reason, F 1 was incubated exactly as described above but with ATP replacing ADP in the prior incubation mixture. Thus, the final prior incubation mixture contained 200 M each of MgCl 2 , ATP, and V i . After 1 h under these turnover conditions, aliquots were removed and assayed for ATPase activity. Results presented in Fig. 4 show that F 1 -ATPase activity is inhibited about 50% in the absence of uv light and about 70% in its presence. Although after 1 h the degree of inhibition is not as great as that achieved when prior incubation is carried out with ADP rather than ATP in the prior incubation mixture, this is to be expected. Thus, under the latter conditions ATP must first undergo hydrolysis before ADP is available for formation of the MgADP⅐V i -F 1 complex and MgATP competes with ADP for binding to F 1 -ATPase active sites.
Polypeptide Chain Cleavage Occurs within the MgADP⅐V i -F 1 Complex in the Presence of UV Light-As indicated earlier,  , see arrows), a 0.1-ml aliquot was withdrawn and diluted 6-fold in 50 mM MOPS, 10% glycerol, pH 8.5. The diluted solution was filtered through Amicon's Microcon 100 Filtration Unit at 25°C by centrifugation at 500 ϫ g for 15 min. The filtrate was discarded and the retentate was diluted to 0.1 ml and assayed for F 1 -ATPase activity as described under "Experimental Procedures" (right panels, Cycle 1). The dilution, washing, assay procedures were then repeated (right panels, Cycle 2). The entire experiment was repeated with essentially identical results.
vanadate is photoreactive and has been shown in the case of the MgADP⅐V i -myosin complex to modify an active site serine residue within contact distance of V i and, in the presence of light and atmospheric oxygen, to induce peptide bond cleavage at this site (8,9) (Fig. 1A). For this reason F 1 was priorly incubated in the presence of 200 M each of MgCl 2 , ADP, and V i in the absence and presence of uv light exactly as described for Figs. 2 and 3 and then subjected to SDS-PAGE. When the resultant SDS-PAGE profiles (Fig. 5, A and B) of the two incubation mixtures (absence and presence of light) are compared, it is clear that only in the latter case has polypeptide bond cleavage occurred. Thus, in addition to Coomassie-stained bands corresponding to the known F 1 subunits ␣, ␤, ␥, ␦, and ⑀, bands distinct from these subunits with apparent molecular masses of 17 and 34 kDa appear (Fig. 5B, lane 6). Polypeptide bond cleavage within the F 1 molecule is highly specific, and is not observed in the absence of uv light under any condition tested (Fig. 5A), and is observed in the presence of uv light with MgCl 2 ϩ ADP ϩ V i (Fig. 5B, lane 6). Fig. 5C shows that the appearance of the 17-and 34-kDa peptide fragments increases with time as expected and levels off after about 2 h (lane 9). These findings are consistent with polypeptide bond cleavage within the nucleotide-binding consensus region (GGAGVGKT) of the 51.5-kDa ␤-subunit, as this would give rise to two fragments with molecular masses near the experimentally determined values of the 17-and 34-kDa bands (Fig. 5C). Fig. 6A, lanes 1-5, show that, under conditions which result in cleavage of a polypeptide chain within the MgADP⅐V i -F 1 complex in the presence of uv light and atmospheric oxygen, the oligomeric state of the F 1 molecule remains intact. In fact, the F 1 molecule remains intact in uv light under all conditions tested (i.e. alone or with V i , MgCl 2 ϩ ADP, ADP ϩ V i , or MgCl 2 ϩ ADP ϩ V i ; Fig. 6A, lanes  1-5, respectively). Specifically, as it applies to the uv light and V i -dependent cleavage of a polypeptide chain within the MgADP⅐V i -F 1 complex (Fig. 5B, lane 6), these results are consistent with a very localized reaction which is otherwise without deleterious effect on the remaining part of the F 1 molecule. Identification of the ␤ or "catalytic" subunit within F 1 as the source of the 17-and 34-kDa bands was derived from two separate experiments. In the first, SDS-PAGE gels of the uv light-treated MgADPV i ⅐F 1 complex, after transfer to PVDF membranes, were probed with a polyclonal, antibody raised against the synthetic peptide KIGLFGGAGVGKCT, containing the GX 4 GKT consensus region of the rat liver ␤ subunit. As shown in Fig. 6B, lane 6, only the intact ␤ subunit and the 17-kDa band cross-react with the antibody. This is the expected result if polypeptide bond cleavage occurs within the nucleotide binding consensus region of the ␤ subunit, as the epitope reactive with the antibody would be largely retained at the C terminus of the 17-kDa fragment, but missing from the 34-kDa fragment. In the second experiment, N-terminal sequence analysis (Fig. 6C) which identified the 7-amino acid stretch APK-AGTA confirmed the ␤ subunit (Fig. 1B) as the origin of the 17-kDa fragment. (The first six amino acids at the N terminus of the rat liver ␤ subunit predicted from c-DNA cloning (Fig.  1B) are not present in the isolated protein.) N-terminal sequence analysis of the 34-kDa fragment also proved possible by carrying out the transfer from SDS-PAGE to PVDF membranes for 2 rather than 1 h, and at 25°C rather than 4°C (see "Experimental Procedures"). This modification in the transfer procedure was done to promote removal of an oxalyl group (HO-CO-CO-) predicted to be at the N terminus of the 34-kDa fragment following the V i -dependent cleavage reaction (Fig.  1A). Significantly, the N-terminal sequence obtained, GVGK-TVLIMELINN (Fig. 6D), not only confirmed the 34-kDa frag- FIG. 4. Demonstration that V i induces inhibition of F 1 -ATPase activity under turnover conditions. Prior incubation was carried out at 25°C with F 1 -ATPase in the absence or presence of 200 M each of MgCl 2 , ATP, and V i . After 1 h, 3-l aliquots (1.5 mg F 1 ) were withdrawn and assayed for ATPase activity as described under "Experimental Procedures." Where indicated a control was also carried out under identical conditions but with ADP rather than ATP in the prior incubation mixture. Dark bars, absence of light; shaded bars, presence of light. The data presented are averages of duplicate determinations. ment as being derived from the ␤ subunit, but placed the site of cleavage at alanine 158 within the GGAGVGKT consensus region.

Polypeptide Chain Cleavage within the MgADP⅐V i -F 1 Complex Induced by UV Light Does Not Alter the Oligomeric State of F 1 and Occurs Only within the ␤-Subunit-Results of native PAGE experiments presented in
The Ratio of the Number of ␤ Subunits Cleaved/F 1 to the Number of ␤ Subunits Binding ADP Is Near 1-To determine the extent of involvement of the 3 ␤ subunits of F 1 in the formation of the MgADP⅐V i -F 1 transition state complex, and in the formation of the 17-and 34-kDa cleavage products, both ADP-binding studies and densitometric analysis of Coomassiestained bands were carried out. Fig. 6E shows that under the conditions used in this study, rat liver F 1 binds only about 1 mol of ADP/mol of F 1 and that V i and MgCl 2 have little or no affect on this stoichiometric ratio. When MgCl 2 , ADP, and V i are added together, each at a concentration of 200 M to favor formation of the MgADP⅐V i -F 1 complex ( Fig. 2A), the stoichiometry of ADP binding remains constant. Thus, formation of the transition state complex appears to be restricted predominantly to the involvement of 1 ␤ subunit. This correlates well with the estimated number of ␤ subunits involved in formation of the 17-and 34-kDa cleavage products when the transition state complex is formed in the presence of uv light. Here, a loss of 33% of the total F 1 ␤ subunit staining intensity results which is fully recovered by the sum of the staining intensities of the 17-and 34-kDa products (Fig. 6F).
Studies with Sodium Borohydride Provide Further Evidence That an Alanine Residue Is Oxidized when the MgADP⅐V i -F 1 Complex Is Treated with UV Light-Studies described above provide rather compelling evidence that, in the presence of atmospheric oxygen, uv light-induced cleavage of a single ␤ subunit within the MgADP⅐V i -F 1 complex occurs at alanine 158. If the ␤-methyl group of this alanine residue is oxidized, it is expected to proceed through a series of reactions resulting first in the formation of serine, then an aldehyde (Fig. 7A), and finally other intermediates before peptide bond cleavage finally occurs (see Fig. 1A and Refs. 22 and 27). Therefore, it should be possible, via reduction of the aldehyde, to demonstrate a uv light-dependent incorporation of tritium [ 3 H] from 3 H-labeled sodium borohydride into F 1 -ATPase following the enzyme's inactivation by MgADP⅐V i . Results presented in Fig. 7B show that such an incorporation does occur in very significant amounts over control levels obtained in the absence of MgADP⅐V i . Incorporation reaches a maximal level in about 7 min, and then declines as expected because the aldehyde is a transient intermediate in the overall oxidation process.
Implications for the Chemical Mechanism of ATP Hydrolysis by F 1 -Studies reported here strongly implicate alanine 158 in the third position of the GGAGVGKT consensus of the rat liver F 1 ␤-subunit as residing very near the ␥-phosphate group of ATP in the transition state. Significantly, alanine in this position is conserved in all F 1 -ATPases (23), and in Escherichia coli, F 1 mutations in this position to valine or proline result, respectively, in a Ͼ90% loss or a 2-fold activation of catalytic activity (24). Moreover, the comparable "third position" residue, serine 181, in Dictyostelium myosin subfragment 1, has been shown to be within contact distance (2.6 Å) of the V i oxygen atoms in the x-ray structure of the MgADP⅐V i -myosin S1 complex (10). With these facts, and the extensive data presented in this paper in mind, a tentative pathway for ATP hydrolysis at the active site of F 1 is proposed. Here, formation of the transition state (Fig. 8A, center panel) from the pre-ATP hydrolysis state (Fig. 8A, left panel) is considered to align the C ␤ atom of alanine in close proximity to the ␥-phosphorus atom of ATP. In the proposed trigonal bipyramidal transition state complex, the planar ␥-phosphorus atom, a water molecule, and a potential catalytic base, "B," are considered to be sufficiently close to facilitate ATP hydrolysis. As only one of the three ␤ subunits forms a transition state complex in the presence of MgADP⅐V i (Fig. 6, E and F), and F 1 is believed to function by involving sequential participation of the ␤ subunits in catalysis via ␥,␤-subunit interactions (5), it is not unreasonable to suggest also that these interactions may help stabilize the transition state.
The mechanistic role of the third position serine in the GX 4 GK(T/S) sequence in myosin (8,9) involves a much greater capacity to form hydrogen bonds to the oxygen of the ␥-phosphate of ATP than does the alanine in the same position in F 1 -ATPase. Therefore, the third position serine in myosin may play a different role in the mechanism of this enzyme than does the third position alanine in F 1 -ATPases. Nevertheless, in both cases, this third position residue may also be critical in determining the polarity, size, and/or rate of formation of the binding pocket in which the ␥-phosphate group of ATP is contained in the transition state. Significantly, differences in these parameters among different ATP-dependent enzymes may help "set" the catalytic rate at a value most compatible with an enzyme's physiological role, while determining in part substrate specificity (see also Refs. 24 and 25). Along these lines, it is interesting to note that V i in the presence of uv light and oxygen also cleaves adenylate kinase at the third position (proline 17) within the nucleotide binding consensus GGPGSGKGT (26). Thus, in support of the view proposed here, three different enzymes, myosin, F 1 -ATPase, and adenylate kinase are all cleaved at the same third position despite the fact that the amino acid occupying this position is very different in all cases (Fig. 8B), but conserved within its specific enzyme class.
Finally, it should be noted that results presented here, which have focused on alanine 158 of rat liver F 1 , do not preclude other amino acids near the ␥-phosphate of ATP in the transition state. Significantly, in a recent intriguing paper Senior and colleagues (28) have summarized the possible roles of three catalytic site residues in the E. coli F 1 -ATPase. One of these, lysine 155 (lysine 162 in rat liver F 1 ) is considered to be involved in the major functional interaction with the ␥-phosphate of MgATP in the substrate bound or "ground state," but to FIG. 8. Scheme proposed for the initial events in the V i -dependent oxidation of alanine 158 within the nucleotide binding consensus region of F 1 -ATPase. A, scheme depicting the relative positions of the C ␤ atom of alanine 158 and the ␥-phosphorus atom of ATP in the pre-hydrolysis state (left panel), the transition state (center panel), and post-hydrolysis state (right panel). Distances in the left and right panels were obtained from the coordinates of the x-ray structure of bovine heart F 1 (5) kindly provided by Dr. J. E. Walker. :B represents an unknown base, involved either in the abstraction of a proton from water (5) or in sterochemically orienting and polarizing the attacking water without net proton abstraction (28). The hatched line between the C B carbon of alanine 158 and the ␥-P group of ATP in the transition state indicates that they are very near one another, not that there is a direct chemical interaction, although this possibility cannot be excluded. B, comparison of the amino acid sequences of the Walker A motif within myosin (rabbit muscle), adenylate kinase (chicken muscle), and F 1 -ATPase (rat liver). See Refs. 8 and 26, respectively, for myosin and adenylate kinase and refer to Fig. 1B of this paper for F 1 -ATPase. undergo conformational repositioning during catalysis. Therefore, when taken together with the novel findings from studies reported here, it will be of considerable interest to visualize the precise location and orientation of lysine 162 and alanine 158 in the transition state when the x-ray structure of the MgADP⅐V i -F 1 complex is elucidated.