The covalent modification of myosin's proteolytic fragments by a purine disulfide analog of adenosine triphosphate. Reaction at a binding site other than the active site.

A purine disulfide analog of ATP, 6,6'-dithiobis(inosinyl imidodiphosphate), forms mixed disulfide bonds between the 6 thiol group on the purine ring and certain key cysteines on myosin, heavy meromyosin, and subfragment one. The EDTA ATPase activities of myosin and heavy meromyosin were completely inactivated when 4 mol of thiopurine nucleotide was bound. When similarly inactivated, subfragment one, depending on its method of preparation, incorporated either 1 or 2 mol of thiopurine nucleotide. Modification of a single cysteine on subfragment one resulted in an inhibition of both the Ca2+ and the EDTA ATPase activities, but the latter always to a greater extent. Modification of two cysteines per head of heavy meromyosin had the same effect suggesting that the active sites were not blocked by the thiopurine nucleotides. Direct evidence for this suggestion was provided by equilibrium dialysis experiments. Heavy meromyosin and subfragment one bound 1.9 and 0.8 mol of [8-3H]adenylyl imidodiphosphate per mol of enzyme, respectively, with an average dissociation constant of 5 X 10(-7) M. Heavy meromyosin with four thiopurine nucleotides bound or subfragment one with two thiopurine nucleotides bound retained 65-80% of these tight adenylyl imidodiphosphate binding sites confirming the above suggestion. Thus previous work assuming reaction of thiopurine nucleotide analogs at the active site of myosin must be reevaluated. Ultracentrifugation studies showed that heavy meromyosin which had incorporated four thiopurine nucleotides did not bind to F-actin while subfragment one with one thiopurine nucleotide bound interacted only very weakly with F-actin. Thus reaction of 6,6'-dithiobis(inosinyl imidodiphosphate) at nucleotide binding sites other than the active sites reduces the rate of ATP hydrolysis and inhibits actin binding. It is suggested that these second sites may function as regulatory sites on myosin.     

Subunit location of sulfhydryl groups of myosin labeled with a purine disulfide analog of adenosine triphosphate.

A purine disulfide analog of ATP, 6,6'-dithiobis(inosinyl imidodiphosphate), forms mixed disulfides with cysteine residues at what are believed to be ATP regulatory sites of myosin. Blocking these sites causes inactivation of the ATPase activity at the active sites. Two cysteine residues per head are specifically modifed by this disulfide analog. The thiopurine nucleotides can be stoichiometrically displaced from myosin by [14-C]cyanide to give a more stable thiocyanato derivative of the enzyme. [14-C]Thiocyanatomyosin (3.7 14-CN/myosin) was dissociated in 4 M urea and the individual subunits were isolated. The heavy chains each had 0.78 14-CN bound per 200,000 molecular weight unit. The light chain with molecular weight of 20,700 had 1.00 14-CN bound and the 16,500 molecular weight light chain had 0.65 14-CN bound. The two 19,000 molecular weight light chains were not labeled. The two labeled light chains have only a single cysteine which is stoichiometrically modified. These two light chains show a high degree of homology and presumably perform identical functions in myosin. Their specific modification by the purine disulfide analog and their other known properties suggest that they contribute directly to the ATP regulatory sites and may, in fact, function as regulatory subunits.     

Stoichiometry of labeling of myosin's proteolytic fragments by a purine disulfide analog of adenosine triphosphate.

A site-specific analog of ATP, 6,6'-dithiobis (inosinyl imidodiphosphate (S2P-PNP), inactivates the ATPase activities of myosin's proteolytic fragments, heavy meromyosin (HMM) and subfragment one (SF1), by formation of mixed disulfides between the 6 position of the purine ring and certain key cysteines. The stoichiometry of the reaction was determined by quantitatively displacing the thiopurine nucleotides from the labeled enzymes with sodium[14-C]cyanide. The thiocyanatoenzyme formed regained 25 percent of the original activity showing that the cysteines modified were not essential for catalysis. The rate of uptake of label paralleled the rate of inactivation. HMM was completely inactivated when 4 mol of thiopurine nucleotide was bound. SF1 made by a papain digestion of myosin incorporarted 2 mol of thiopurine nucleotide when completely inactivated. Having adenylyl imidodiphosphate, areversible competitive inhibitor of myosin's ATPase, present during the inactivation of HMM by S2P-PNP demonstrated that only one cysteine per head needed to be blocked to inactivate the enzyme. Moreover, SF1 made by a trypsin digest of HMM was completely inactivated when only 1.1 mol of the thiopurine nucleotide bound again indicating that blocking only a single cysteine per head was sufficient to cause inactivation. This sulfhydryl is thought to be at an ATP binding site distinct from the ATPase site. The properties of this second ATP binding site are consistent with it being an ATP regulatory site.     

Reaction of cardiac myosin with a purine disulfide analog of adenosine triphosphate. II. Stoichiometry and subunit location of labeling.

The reaction of bovine cardiac myosin with the site-specific purine disulfide analog of ATP, 6,6'-dithiobis (inosinyl imidodiphosphate), was studied to determine the stoichiometry of labeling and subunit location of the reactive cysteines. The analog inactivates myosin by forming a mixed disulfide between the thiopurine nucleotide and certain key cysteines. The thiopurine nucleotide was displaced quantitatively by 14CN to form the more stable thiocyanato-enzyme derivatives. In cardiac myosin, the reactive cysteines could be categorized into three classes, nonessential, critical, and noncritical. The modification of the critical cysteines (two per myosin) inactivated the EDTA and Ca2+ ATPase activities, the latter to a lesser extent. The nonessential cysteines (two to three per myosin) and the noncritical cysteines (two per myosin), differentiated by their rates of reaction, had no effect on the ATPase activities after modification. Thiocyanato-modified myosin was analyzed by sodium dodecyl sulfate gel electrophoresis to determine the distribution of 14CN in the subunits. The critical cysteines were found on the 21,000-dalton light chain (LC1) and the noncritical cysteines on the heavy chains. More specifically, the critical cysteine modified in cardiac LC1 (determined from the products after cyclization and chain cleavage at the thiocyanatoalanyl residues) was shown to be the thiol residue whose surrounding amino acid sequence is homologous to that of the single cysteine of the skeletal myosin alkali light chains, confirming the likely similar structure and function of these light chains in the two different muscle types.     

Modification of the alkali light chains of skeletal myosin inhibits actin binding and adenosine triphosphate cleavage.

Heavy meromyosin treated with the ATP analog, 6,6'-dithiobis(inosinyl-5'-yl imidodiphosphate), (slppNHp)2, in the presence of adenyl-5'-yl imidodiphosphate at 0 degrees loses its EDTA-ATPase activity and actin binding ability in a parallel manner. Studies with myosin show that under the above conditions (slppNHp)2 reacts preferentially with the single cysteines of the alkali light chains (Mr = 20,700 and 16,500) suggesting a role for these subunits in regulating actin-myosin interaction and ATP cleavage.     

Observations on the kinetics, subunit composition, and sulfhydryl reactivity of myosin from Physarum polycephalum.

A highly purified preparation of myosin from Physarum polycephalum has been shown by sodium dodecyl sulfate polyacrylamide gel electrophoresis to contain heavy chains and only one molecular weight class of light chains, of approx. 15 000 daltons. Kinetic investigations of the Ca2+-ATPase and Mg2+-ATPase (ATP phosphohydrolases, EC 3.6.1.3) at pH 8.0 gave Km and V values of 17.3 muM and 1.25 mumol Pi/min per mg, and 2.4 muM and 0.12 mumol Pi/min per mg, respectively. Adenylyl imidodiphosphate, a beta-gamma-imido ATP analog, inhibited the ATPase activity of Physarum myosin competitively with Ki values equal to 350 and 12 muM in the presence of Ca2+ and Mg2+, respectively. The ATPase activity of Physarum myosin was inhibited at a very low rate (t1/2 = 24 h) by the ATP analog, 6,6'-dithiobis(inosinyl imidodiphosphate), with concentrations of inhibitor previously shown to inactivate (t1/2 approximately 10 min) skeletal and cardiac myosins rapidly by reacting with key cysteines.     

Binding of ADP and ATP analogs to cross-linked and non-cross-linked acto X S-1

We previously determined the binding constants of ADP, adenylyl imidodiphosphate (AMP-PNP), and inorganic pyrophosphate (PPi) to acto . myosin subfragment 1 (acto X S-1) by measuring the dissociation of acto X S-1 as a function of ATP analog concentration (Greene, L.E., and Eisenberg, E. (1980) J. Biol. Chem. 255, 543-548). In the present study, we reinvestigated this question by measuring the extent to which these ATP analogs inhibit the acto X S-1 ATPase activity using both cross-linked actin X S-1 and non-cross-linked proteins. No significant difference was found between the cross-linked and non-cross-linked acto X S-1 complexes in their affinity for either ADP or AMP-PNP. The binding constant of ADP to acto X S-1 determined by the inhibition method was in excellent agreement with that obtained previously by the dissociation method, both techniques giving values of about 7 X 10(3) M-1. However, this was not the case for AMP-PNP and PPi, with the inhibition method giving about 10-fold weaker binding constants than those determined previously by the dissociation method. Upon redoing our dissociation experiments over a wider range of actin concentrations than we used previously, we now find that the dissociation method gives much weaker values for the binding constants of PPi and AMP-PNP to acto X S-1, i.e. values on the order of 4 X 10(2) M-1. The very weak binding of these ATP analogs to acto X S-1 makes it difficult to obtain these values with great accuracy. Nevertheless, they seem to be in good agreement with the binding constants determined by the inhibition method. The weak binding of AMP-PNP and PPi to acto X S-1 is consistent with the recent fiber studies of Pate and Cooke (Pate, E., and Cooke, R. (1985) Biophys. J. 47, 773-780) and Schoenberg and Eisenberg (Schoenberg, M., and Eisenberg, E. (1986) Biophys. J. 48, 863-872).     

Dissociation of the actin.subfragment 1 complex by adenyl-5'-yl imidodiphosphate, ADP, and PPi

The ability of adenyl-5'-yl imidodiphosphate (AMP-PNP), ADP, and PPi to dissociate the actin.myosin subfragment 1 (S-1) complex was studied using an analytical ultracentrifuge with UV optics, which enabled the direct determination of the dissociated S-1. At mu = 0.22 M, pH 7.0, 22 degrees C, with saturating nucleotide present, ADP weakens the binding of S-1 to actin about 40-fold (K congruent to 10(5) M-1), while both AMP-PNP and PPi weakens the binding about 400-fold (K congruent to 10(4) M-1). This 10-fold stronger dissociating effect of AMP-PNP and PPi compared to ADP correlates with our data showing that the binding of AMP-PNP and PPi to S-1 is about 10-fold stronger than the binding of ADP. In contrast, the binding constants of ADP, AMP-PNP, and PPi to acto.S-1 are nearly identical (K congruent to 5 x 10(3) M-1). At 4 degrees C, AMP-PNP has only a 3-fold stronger dissociating effect than ADP and, similarly, our data suggest that the binding of AMP-PNP and ADP to S-1 is quite similar at 4 degrees C. AMP-PNP and PPi are, therefore, somewhat better dissociating agents than ADP, but the difference among these three ligands is quite small. These data also show that actin and nucleotide bind to separate but interacting sites on S-1 and that the S-1 molecules bind independently along the F-actin filament with a binding constant of about 1 x 10(7) M-1 at 22 degrees C and physiological ionic strength.     

Effect of ethylene glycol and Ca2+ on the binding of Mg2+ x adenyl-5'-yl imidodiphosphate to rabbit skeletal myofibrils

The binding of Mg2+ X adenyl-5'-yl imidodiphosphate (Mg2+ X AMP-PNP) to rabbit skeletal myofibrils has been measured in aqueous solution and in 50% ethylene glycol in the presence and absence of Ca2+. In water, the observed binding was weak with less than half the calculated myosin active sites filled even at 1 mM Mg2+ X AMP-PNP. In 50% ethylene glycol, the binding is at least 100-fold tighter and extrapolates to the expected number of binding sites. This is contrasted to the small change seen for Mg2+ X ADP binding between the same sets of conditions. This difference between Mg2+ X AMP-PNP and Mg2+ X ADP is attributed to the strong coupling of Mg2+ X AMP-PNP binding to dissociation of myosin cross-bridges. The Ca2+ sensitivity of Mg2+ X AMP-PNP binding in 50% ethylene glycol is taken as further evidence of the thermodynamic coupling of Mg2+ X AMP-PNP binding to cross-bridge dissociation. In addition, the binding of Mg2+ X AMP-PNP in 50% ethylene glycol is biphasic while Mg2+ X ADP binding under the same conditions is not. The biphasic Mg2+ X AMP-PNP binding could be caused by either the presence of two or more classes of cross-bridges or by negative cooperativity, but the presence of only a single class of Mg2+ X ADP-binding sites implies that if multiple classes of sites are involved, they do not simply differ in steric hindrance or accessibility of the binding site as a whole. The importance of using purified AMP-PNP in the study of actomyosin X AMP-PNP complexes is discussed.     

Inhibition of gill (Na+ + K+)-ATPase in rainbow trout (Salmo gairdneri) by a purinedisulfide analog of adenosine triphosphate.

The effect of the adenosine triphosphate analog, 6,6'-dithiobis(inosinyl imidodiphosphate), (sIMP-PNP)2, was tested on the ouabain-sensitive (Na+ + K+)-ATPase (ATP phosphohydrolase, EC 3.6.1.3) and the ouabain-insensitive Mg2+ - ATPase in microsomes prepared from gill tissue of sea water-adapted rainbow trout, Salmo gairdneri. The (Na+ + K+)-ATPase was completely inhibited by low concentrations of (sIMP-PNP)2 (6 micrometer) but the Mg2+ - ATPase was unaffected by the inhibitor at concentrations as high as 28 micrometer, supporting the suggestion that the two activities represent separate enzymes. The specificity of inactivation could be demonstrated both at a physiological temperature (13 degrees C) and at 37 degrees C. The rates of inactivation were similar at both temperatures. Inactivation of the (Na+ + K+)-ATPase by (sIMP-PNP)2 was reversed by dithiothreitol, suggesting that the inhibitor forms a mixed disulfide with sulfhydryl groups on the enzyme. The inability of substrate (either ATP or its analog, adenyl-5'-yl imidodiphosphate) to protect against inactivation suggests that (sIMP-PNP)2 is reacting with sulfhydryl groups which are not associated with the active site.     

Exploring sites on mitochondrial ATPase for catalysis,regulation, and inhibition

Evidence is presented that mitochondrial ATPase has two types of sites that bind adenine nucleotides. The catalytic site, C, binds the substrates ATP, GTP, or ITP and the inhibitor guanylyl imidodiphosphate (GMP-PNP). A second type of site, R, binds ATP, ADP, adenylyl imidodiphosphate (AMP-PNP), and the chromium complexes of ATP or ADP. All of these substances binding to the R site inhibit the hydrolysis of ATP in a competitive manner; their inhibition of hydrolysis of ITP and GTP is noncompetitive. GMP-PNP inhibits oxidative phosphorylation in submitochondrial particles but AMP-PNP does not. The localization on mitochondrial membranes of sites for the binding of various antibiotics that inhibit oxidative phosphorylation is discussed.     

The binding constant of ATP to myosin S1 fragment.

The extent of ATP synthesis from ADP and Pi at the active centre of myosin subfragment 1 has been reinvestigated. The results have been interpreted using a treatment which is not dependent on the number or nature of the intermediates in the ATPase mechanism. An average value for the binding constant of ATP of (3.25 +/- 0.96) X 10(11) M-1 at pH 8.0 23 degrees C and ionic strength 0.12 M was obtained. Additional evidence is given to confirm that synthesis at the active site has been investigated.     

Effects of adenylyl imidodiphosphate, a nonhydrolyzable adenosine triphosphate analog, on reactivated and rigor wave sea urchin sperm

A nonhydrolyzable ATP analog, adenylyl imidodiphosphate (AMP-PNP), has been used to study the role of ATP binding in flagellar motility. Sea urchin sperm of Lytechinus pictus were demembranated, reactivated, and locked in "rigor waves" by a modification of the method of Gibbons and Gibbons (11). Rigor wave sperm relaxed within 2 min after addition of 4 micrometer ATP, and reactivated upon addition of 10-12 micrometer ATP. The beat frequency of the reactivated sperm varied with ATP concentration according to Michaelis-Menten kinetics ("Km" = 0.24 mM; "Vmax" = 44 Hz) and was competitively inhibited by AMP-PNP (Ki" approximately to 8.1 mM). Rigor wave sperm were completely relaxed (straightened) within 2 min by AMP-PNP at concentrations of 2-4 mM. The possibilities that relaxation in AMP-PNP was a result of ATP contamination, AMP-PNP hydrolysis, or lowering of the free Mg++ concentration were conclusively ruled out. The results suggest that dynein cross-bridge release is dependent upon ATP binding but not hydrolysis.     

Reaction of 5'-p-fluorosulfonylbenzoyl-1,N6-ethenoadenosine with histidine and cysteine residues in the active site of rabbit muscle pyruvate kinase

The inactivation of rabbit muscle pyruvate kinase by 0.3 mM 5'-p-fluorosulfonylbenzoyl-1,N6-ethenoadenosine at pH 7.8 is biphasic. The first phase proceeds rapidly to yield a partially active enzyme (46% residual activity) followed by a slower rate which leads to total inactivation. The inactivation of the first phase can be reversed by addition of 20 mM dithiothreitol, whereas the second phase is unaffected. These two phases have second-order rate constants of 250 M-1 X min-1 (dithiothreitol-sensitive reaction) and 52 M-1 X min-1 (dithiothreitol-insensitive reaction), respectively. Marked protection against inactivation is afforded by phosphoenolpyruvate and by metal-nucleotide complexes in the presence of free metal, indicating that reaction occurs in the region of the active site. Loss of approximately two sulfhydryls per enzyme subunit correlates well with the dithiothreitol-sensitive inactivation, suggesting that this phase of the inactivation may be attributable to disulfide formation. Incorporation of about one mole of fluorescent reagent per enzyme subunit correlates closely with the dithiothreitol-insensitive phase of inactivation, yielding a modified histidine residue. The quantum yield of the fluorescent sulfonylbenzoyl-1,N6-ethenoadenosine-pyruvate kinase is only 0.007, as compared to 0.54 for the parent nucleoside 1,N6-ethenoadenosine. The quenched fluorescence is consistent with stacking of the sulfonylbenzoyl moiety on the purine ring in the modified enzyme, which suggests that the altered histidine may be located in the adenine region of the metal-nucleotide binding site.     

Sarcoplasmic reticulum ATPase catalyzes hydrolysis of adenyl-5'-yl imidodiphosphate

Sarcoplasmic reticulum ATPase has been found to cleave the ATP analog adenyl-5'-yl imidodiphosphate in a calcium-dependent reaction. The reaction products were determined by 31P NMR to be inorganic phosphate and adenyl-5'-yl phosphoramidate (AMP-PN). AMP-PNP hydrolysis, like ATP hydrolysis, drives active Ca2+ accumulation by sarcoplasmic reticulum vesicles.     

Interactions of the actin and nucleotide binding sites on myosin subfragment 1.

The effects of selected nucleotides (N) on the binding of myosin subfragment 1 (S-1) and pure F-actin (A) were measured by time-resolved fluorescence depolarization for 0.15 M KCl, pH 7.0 at 4 degrees. The association constants K'A, KN, and K'N in the scheme (see article), were determined for the magnesium salts of ADP, adenyl-5'-yl imidodiphosphate AMP-P(NH)P, and PPi. The nucleotide binding site on S-1 was "mapped" with respect to its interaction on the actin binding site. The subsites were the beta- and gamma-phosphoryl groups of ATP bind had the largest effects. A quantitative measure of the interaction, the interaction free energy, was defined as -RT ln (KA/K'A). For ADP, K'A was 2.7 X 10(5) M-1 and the interaction free energy was -4.67 kJ M-1. For AMP-P(NH)P and PPi it was much larger. A ternary complex was shown to exist for ADP, S-1, and actin in the presence of Mg2+ and evidence from AMP-P(NH)P and PPi measurements indicated that ATP also likely forms a ternary complex. The mechanism of (S-1)-actin dissociation is discussed in light of these results.     

Characterization of myosin V binding to brain vesicles

Myosin II and V are important for the generation and segregation of subcellular compartments. We observed that vesicular myosin II and V were associated with the protein scaffolding of a common subset of vesicles by density sedimentation, electron microscopy, and immunofluorescence. Solubilization of either myosin II or V was caused by polyphosphates with the following efficacy at 10 mM: for myosin II ATP-Mg(2+) = ATP = AMP-PNP (5'-adenylyl imidodiphosphate) > pyrophosphate = tripolyphosphate > tetrapolyphosphate = ADP > cAMP = Mg(2+); and for myosin V pyrophosphate = tripolyphosphate > ATP-Mg(2+) = ATP = AMP-PNP > ADP = tetrapolyphosphate > cAMP = Mg(2+). Consequently, we suggest solubilization was not an effect of phosphorylation, hydrolysis, or disassociation of myosin from actin filaments. Scatchard analysis of myosin V binding to stripped dense vesicles showed saturable binding with a K(m) of 10 nM. Analysis of native vesicles indicates that these sites are fully occupied. Together, these data show there are over 100 myosin Vs/vesicle (100-nm radius). We propose that polyphosphate anions bind to myosin II and V and induce a conformational change that disrupts binding to a receptor.     

The influence of Mg2+ on anion binding to sarcoplasmic reticulum membranes as detected by 35Cl-NMR.

35Cl-NMR spectroscopy has been used to study the competition between anions, including nucleotides, on skeletal muscle sarcoplasmic reticulum membranes. Different chloride binding sites can be distinguished according to their Mg2+ sensitivity. Phosphate binding is enhanced by Mg2+ whereas the anion transport inhibitor pyridoxalphosphate-6-azophenyl-2'-sulfonic acid (PPAPS) binding is not. The affinity of the enzyme for the Mg-adenylyl imidodiphosphate (MgAMP-PNP) complex is decreased whereas that for MgATP is increased. Three sets of binding sites can be discriminated from which chloride is displaced by different anions with varying efficiency. High affinity binding of AMP-PNP and PPAPS occurs at the same site, that can also be occupied by phosphate. Low-affinity binding of PPAPS and AMP-PNP also coincides, but in a site where phosphate binding is negligible. ATP and ADP bind to both sites. In the presence of Mg2+ a third anion binding site can be occupied by phosphate but neither by AMP-PNP nor PPAPS.     

Actin mediated release of ATP from a myosin-ATP complex.

The apparent second-order rate constant, ka-2, of actin binding to a myosin-ATP state (M*.ATP) and releasing ATP to the medium has been determined by two methods. The first was the measurement of the amount of ATP released when actin was added to the intermediate state, M*.ATP; the second was the measurement of oxygen exchange between ATP and HOH. A quantitative treatment of ATP in equilibrium HOH exchange is given to allow extraction of elementary rate constants from the data. Agreement between the two methods was good and at low ionic strength and 23 degrees C, ka-2 is 6 X 10(5) M-1 s-1 which is about one-third the value of the apparent second-order rate constant, ka4, of actin binding to the myosin product state (M**.ADP.Pi). The determination of ka-2 allows a lower limit of 6 s-1 to be placed upon the first-order rate of ATP release from AM.ATP. This is to be compared with a value of less than or equal to 1.5 X 10(-4) s-1 for the equivalent steps of the myosin scheme; thus actin enhances the rate by a factor of 4 X 10(4) or more. A greater proportion of the bound ATP is released to the medium as ATP with increasing actin concentration. This reflects the contribution to rate limitation at saturating actin concentration of steps between myosin states dissociated from actin.     

Adenine nucleotide binding sites on beef heart F1-ATPase. Specificity of cooperative interactions between catalytic sites

Cooperative interactions between nucleotide binding sites on beef heart mitochondrial F1-ATPase have been studied by measuring substrate-promoted release of 5'adenylyl-beta,gamma-imidodiphosphate (AMP-PNP) from a single high affinity site. The site is initially loaded by incubating F1 with an equimolar amount of the nonhydrolyzable ATP analog. When unbound [3H]AMP-PNP is removed and the complex diluted to a concentration below the Kd, release of ligand shows an apparent absolute requirement for medium ADP. Release is biphasic with the extent of release during the initial rapid phase dependent on the concentration of medium ADP. Although phosphate alone has no effect, it enhances the rapid phase of ADP-promoted release over 2-fold with a half-maximal effect at 60 micrometers P1. The binding of efrapeptin (A23871) to the F1.AMP-PNP complex completely prevents ADP-promoted dissociation. Although AMP-PNP release also occurs in the presence of medium ATP, the F1.AMP-PNP complex does not dissociate if an ATP-regenerating system of sufficient capacity to prevent accumulation of medium ADP is added. Consistent with an inability of nucleoside triphosphate to promote release is the failure of medium, nonradioactive AMP-PNP to affect retention of the 3H-labeled ligand. The stability of F1.AMP-PNP complex in the absence of medium nucleotide and the highly specific ability of ADP plus P1 to promote rapid release of the ATP analog are interpreted as support for an ATP synthesis mechanism that requires substrate binding at one catalytic site for product release from an adjacent interacting site.