Kinetic mechanism of 1-N6-etheno-2-aza-ATP hydrolysis by bovine ventricular myosin subfragment 1 and actomyosin subfragment 1. The nucleotide binding steps.

The large change in fluorescence emission of 1-N6-etheno-2-aza-ATP (epsilon-aza-ATP) has been used to investigate the kinetic mechanism of etheno-aza nucleotide binding to bovine cardiac myosin subfragment 1 (myosin-S1) and actomyosin subfragment 1 (actomyosin-S1). The time course of nucleotide fluorescence enhancement observed during epsilon-aza-ATP hydrolysis is qualitatively similar to the time course of tryptophan fluorescence enhancement observed during ATP hydrolysis. In single turnover experiments, the nucleotide fluorescence rapidly increases to a maximum level, then decreases with a rate constant of 0.045 s-1 to a final level, which is about 30% of the maximal enhancement; a similar fluorescence enhancement is obtained by adding epsilon-aza-ADP to cardiac myosin-S1 or actomyosin-S1 under the same conditions (100 mM KCl, 10 mM 4-morpholinepropanesulfonic acid, 5 mM MgCl2, 0.1 mM dithiothreitol, pH 7.0, 15 degrees C). The kinetic data are consistent with a mechanism in which there are two sequential (acto)myosin-S1 nucleotide complexes with enhanced nucleotide fluorescence following epsilon-aza-ATP binding. The apparent second order rate constants of epsilon-aza-ATP binding to cardiac myosin subfragment 1 and actomyosin subfragment 1 are 2-12 times slower than those for ATP. Actin increases the rate of epsilon-aza-ADP dissociation from bovine cardiac myosin-S1 from 1.9 to 110 s-1 at 15 degrees C which can be compared to 0.3 and 65 s-1 for ADP dissociation under similar conditions. Although there are quantitative differences between the rate and equilibrium constants of epsilon-aza- and adenosine nucleotides to cardiac actomyosin-S1 and myosin-S1, the basic features of the nucleotide binding steps of the mechanism are unchanged.

etheno-2-aza-ATP (t-aza-ATP) has been used to investigate the kinetic mechanism of etheno-aza nucleotide binding to bovine cardiac myosin subfragment 1 (myosin-S1) and actomyosin subfragment 1 (actomyosin-Sl). The time course of nucleotide fluorescence enhancement observed during 6-aza-ATP hydrolysis is qualitatively similar to the time course of tryptophan fluorescence enhancement observed during ATP hydrolysis. In single turnover experiments, the nucleotide fluorescence rapidly increases to a maximum level, then decreases with a rate constant of 0.045 s-l to a final level, which is about 30% of the maximal enhancement; a similar fluorescence enhancement is obtained by adding c-aza-ADP to cardiac myosin-S1 or actomyosin-S1 under the same conditions (100 mM KCl, 10 mM 4-morpholinepropanesulfonic acid, 5 mM MgClz, 0.1 mM dithiothreitol, pH 7.0, 15 "C). The kinetic data are consistent with a mechanism in which there are two sequential (acto)myosin-S1 nucleotide complexes with enhanced nucleotide fluorescence following t-aza-ATP binding. The apparent second order rate constants of t-aza-ATP binding to cardiac myosin subfragment 1 and actomyosin subfragment 1 are 2-12 times slower than those for ATP. Actin increases the rate of e-aza-ADP dissociation from bovine cardiac myosin-S1 from 1.9 to 110 s-l at 15 "C which can be compared to 0.3 and 65 s-' for ADP dissociation under similar conditions.
Although there are quantitative differences between the rate and equilibrium constants of t-aza-and adenosine nucleotides to cardiac actomyosin-S1 and myosin-S1, the basic features of the nucleotide binding steps of the mechanism are unchanged.
Biochemical studies of the actomyosin ATPase cycle have provided a description of how the chemical reaction of ATP hydrolysis is coupled to the interaction of actin and myosin in solution (1-3). A minimal mechanism of hydrolysis of ATP by myosin and actomyosin is shown in Equation 1, KAT KAH where M represents a myosin head; A, actin; T, ATP; D, ADP; and P, phosphate.' The essential features of the scheme are: (i) myosin and M-D bind tightly to and dissociate very slowly from actin (5). In contrast, the states M-T and M-D-P are weakly bound to actin and dissociate rapidly. As a result the kinetic pathway alternates between tightly and weakly bound states (3). (ii) In the absence of actin, the rate-limiting step in the pathway occurs after product formation (1,2). (iii) In the presence of actin, product release is accelerated at least 200-fold. The mechanism of hydrolysis of ATP ,by cardiac myosin-Sl' and actomyosin-Sl i s basically similar to that shown in Equation 1 for the skeletal proteins with differences in the rates of some of the steps (6-8). For example, the rate constant of ADP dissociation from rabbit skeletal acto-myosin31 is too rapid to be measured by stopped-flow methods, whereas that for cardiac actomyosin-S1 is readily measured (8). Therefore, we have chosen to study the mechanism of cardiac rather than skeletal actomyosin-S1 ATP hydrolysis pathway, so that a more complete kinetic analysis can be made. In addition, such a study provides the background necessary to understand the molecular basis of the changes in cardiac contractility induced by adaptive or pathological conditions.
The intrinsic tryptophan fluorescence of the myosin head is enhanced during hydrolysis of ATP approximately 25% as denoted by the asterisks in Equation 1. This has allowed the kinetics of the hydrolytic pathway to be investigated. However, only kinetic steps that involve changes in the environment of the tryptophan chromaphore can be observed. More detailed information about the mechanism of the reaction can be obtained by monitoring some spectroscopic property of the nucleotide during the reaction (for example, using 31P NMR (9) or the fluorescence of ATP analogues (10-12)). The reactions of the related fluorescent nucleotide 1-N6-etheno-ATP with myosin-S1 and actomyosin-S1 from skeletal and smooth muscle have been studied in detail (12). There are some interesting similarities and differences between the rates and apparent binding constants of this analogue and those for by skeletal myosin-S1 is about 6 times faster than that of ATP, while the apparent second order rate constant for 1-N6etheno-ATP binding to skeletal myosin-Sl is similar to that of ATP. However, the maximum steady state rate of hydrolysis of l-N'-etheno-ATP by skeletal actomyosin-S1 is about 10 times less than that for ATP (12). The chemical and functional properties of the related ATP analogue, e-aza-ATP, make it particularly useful to study the kinetics of nucleotide hydrolysis and to investigate the cross-bridge cycle in muscle fibers (13), since it produces 75% of the isometric tension obtained with ATP. The fluorescence emission maximum is enhanced approximately %fold and shifted from 485 to 460 nm during steady state hydrolysis by skeletal heavy meromyosin (14). The shift of the fluorescence emission spectrum of t-aza-ATP on binding to myosin-SI enables kinetic measurements to be made without the necessity of quenching the fluorescence of the unbound nucleotide with reagents such as acrylamide, as is required for etheno-ATP (12, 14). We show here that the large change in fluorescence emission at 410-450 nm can be used to study the nucleotide binding steps of the hydrolytic pathway of bovine cardiac myosin-Sl and actomyosin-S1. Experiments were carried out at 15 "C, so that the kinetic data could be compared with that for ATP and ADP (6-8); measurements were also made at low temperature, 0 "C, to provide a set of kinetic parameters for the hydrolysis of the nucleotide under conditions where the reactions were likely to be slow enough to be measured in structured systems. The hydrolytic pathways of ATP and t-aza-ATP are qualitatively similar although there are significant differences (2-12-fold) in the values of some of the rate and equilibrium constants. Moreover, the maximum steady state rates of hydrolysis of ATP or t-aza-ATP by skeletal actomyosin-S1 (13) are the same. Thus, t-aza-ATP is suitable for use as a probe to compare the rate of certain steps in the hydrolytic pathway for bovine cardiac myosin-S1, actomyosin-Sl, and myofibrils. The results of these experiments are described in the accompanying paper (16).

MATERIALS AND METHODS
All solutions were prepared using glass distilled water. Ammonium sulfate was absolute grade (Research Plus Laboratories). The following chemicals were obtained from Sigma: ATP (Sigma grade vanadium free, used for kinetic experiments and grade I1 used for preparative procedures), ADP dicyclohexylammonium salt, A,,A, chymotrypsin type 1-S, dithiothreitol, and lima bean trypsin inhibitor. MOPS was Ultrol grade from Calbiochem-Behring. Cardiac actin, myosin, and myosin-S1 were purified from the left ventricles of bovine hearts as described by Siemankowski and White (8). The concentration of bovine ventricular myosin and myosin-Sl were determined from the absorption at 280 nm using extinction coefficients (&I%, w/v, 1 cm) of 0.55 and 0.64 (17). The concentration of bovine ventricular actin was determined using an extinction coefficient at 280 nm of 1.15 (0.1%, w/v, 1 cm).
Preparation of t-Aza-ATP-l-N'-Etheno-ATP was prepared from ATP (18) and was used to synthesise t-aza-ATP (19). The synthesis and purification were monitored using thin layer and paper chromatography (19). The crude material was purified by ion exchange chromatography on DEAE 52-cellulose in 0.1 M NaCl, 10 mM Tris, pH 8.0, with a linear gradient to 0.5 M NaCl. Fractions which to ligands already bound. If there is more than one intermediate with the same stoichiometry then the intermediates and corresponding rate and equilibrium constants are numbered sequentially beginning with 0 for the collision complex. For example, klAT is the rate constant for the transition AM-a-ATP0 to AM-a-ATP,*. The abbreviations used are: SI, subfragment 1 of myosin; MOPS, 4-morpholinepropanesulfonic acid; t-aza-ATP, 1-N6-etheno-2-azaadenosine-5' triphosphate; c-aza-ADP, l-N6-etheno-2-aza-adenosine-5' diphosphate; Ap5A, P',P'-di(adenosine 5')-pentaphosphate. fluoresced strongly green, when irradiated with a long wave ultraviolet lamp (UVSL- 25), and had the expected fluorescence emission spectra were pooled and re-chromatographed repeatedly on DEAE 52 as described above and finally on DEAE Bio-Gel in 0.05 M NaC1,5 mM MES, pH 6.5, with a linear gradient to 0.4 M NaC1. From each column separation, fractions for which the ratio of ultraviolet absorption at 265 nm to that at 242 nm was less than 0.4 were analyzed using isotachophoresis. The leading electrolyte was 18 mM @-alanine, 5 mM HCl, 0.5% hydroxypropyl methyl cellulose, and the terminating electrolyte was 5 mM a-caproic acid. The nucleotide was detected using an ultraviolet monitor at 280 nm. The purity of the final product was 90-95% with a single major impurity as judged by comparison of the ratio of the areas under the peaks of the isotachophoresis record (20). The nucleotide migrated as a single compound on paper chromatography and thin layer chromatography (19).
Preparation of c-Aza-ADP-c-aza-ADP was prepared by enzymatic hydrolysis of the triphosphate: 1 mM 6-aza-ATP was incubated at 20 "C with 4 ~L M skeletal myosin-S1, 5 p M skeletal actin, 10 p M A,& (10 mM KCl, 10 mM MOPS, 5 mM MgClZ, 0.1 mM dithiothreitol, pH 7.0); when the hydrolysis was complete as monitored by enhancement of fluorescence at 440 nm, the mixture was centrifuged at 100,000 X g for 2 h to remove the actomyosin-S1. The supernatant containing the crude t-aza-ADP had 0.5-2.5% t-aza-ATP, which could be removed by chromatography on DEAE 52 as described above.
The concentration of the nucleotides was determined from the extinction coefficient at 354 nm of 1530 M" cm-I (14).
Kinetic Measurements-Single turnover fluorescence measurements were made using a Spex Fluorolog spectrofluorometer in the photon counting mode. Transient kinetic measurements on myosin-S1 and actomyosin-S1 were made in a stopped flow fluorometer with a 2-cm path length cell as described by Siemankowski and White (8). Excitation wavelengths 365 nm for nucleotide fluorescence measurements and 340 nm for light scattering were obtained using interference filters (Oriel Corp.) to select appropriate wavelength lines from a mercury arc (Osram 100/W2). Fluorescence emission was measured using a 450-nm interference filter with a 40-nm band width, or a 417nm interference filter with a 10-nm band width (Oriel Corp.) or a combination of a sharp cut off high pass filter (Schott KV-418) and a low pass interference filter (Corion LS450) to maximize the contribution of the fluorescence signal of the bound nucleotide relative to that of the free nucleotide. The combination of the low and high pass filters gave a band pass from approximately 415 to 445 nm, which is the region of maximum fluorescence enhancement during steady state c-aza-ATP hydrolysis by myosin-S1. These filters were used for most experiments in which the final nucleotide concentration exceeded 100 p~ because they gave the best signal to noise ratio.
The output of the photomultiplier tube was smoothed using an active filter circuit, in which the reciprocal time constant was always at least 20 times the rate constant of the fastest component of the reaction being studied. Data were recorded at 1024 equally spaced time points using a Nicolet Explorer I11 digital oscilloscope. To improve the signal to noise ratio, three to six data traces for each experimental set of conditions were summed before fitting; however, the fit rate constants for the summed and single data traces did not differ significantly. Rate and amplitude information were obtained by fitting the data to either one or two exponential equations with adjustable end points (Equations 2, a and b).
where Z ( t ) is the fluorescence signal at time t, Zl and Iz are the amplitude coefficients of reactions with rate constant kbsl and kobsz.
Where the data was fit to two exponentials there are 5 adjustable parameters. Therefore, several fitted curves could sometimes be obtained with similar root mean square deviations from the experimental data, each described by a different set of parameter values. This effect is especially apparent for data in which the signs of the amplitude terms are the same and the rate constants differ by a factor of 2-4.
To test whether the parameters obtained by the double exponential fits were dependent upon the fitting procedure used, data were fit by three methods; a manual analogue fitting procedure, and computer fitting either by the Method of Moments (21) or using the algorithm of Foss (22). Data were initially fit manually using an analogue signal from a calibrated RC circuit which was a single exponential or the sum or difference of two exponentials. The signal to Cardiac Myosin-S1 and Actomyosin-SI was adjusted until the difference between the data and simulation were a minimum. Computer fits of the data were made using a DEC Rainbow computer. Data were routinely fit to single or double exponentials by the Method of Moments using a modified version of a program written by Dr. Ken Johnson, Dept. of Biochemistry, Pennsylvania State University, which required approximately 3 min to fit 1024 points to two exponentials. Data were considered to be adequately fit by a single exponential (Equation 2a) if upon visual inspection the difference between the fit curve and the data did not contain long runs of positive or negative points or if the root mean square deviation of the data from the fit curve was not improved by at least a factor of two upon fitting to a double exponential equation. In some cases, data were also fitted by the method of Foss (22); data were averaged two at a time prior to fitting using a modified version of a program written by Duane Flamig in order to avoid memory overflow and reduce the amount of computer time to -2 h for 512 points. The variability of the fits of the slow rate constant between the different fitting procedures was generally less than 5%. There was considerably more variability in fitting the fast rate constant, generally -20%.

RESULTS
Emission Spectra of the Steady State Complex during t-Aza-ATP Hydrolysis-The fluorescence emission maximum of eaza-ATP is shifted from 485 to 460 nm and is enhanced (approximately 5-fold in the 415-445 nm region for ratios of e-aza-ATP/myosin-Sl<l) during the steady state hydrolysis by cardiac myosin-SI. The fluorescence emission spectrum of the myosin-S1-e-aza-ATP complex can be obtained from the maximum fluorescence at each wavelength observed on addition of substoichiometric amounts of e-aza-ATP to myosin-S1. Such a spectrum for cardiac myosin-S1 is shown in Fig.   1. The spectrum of the nucleotide bound to cardiac myosin-S1 is similar to that observed for the complex of the nucleotide with skeletal heavy meromyosin (14). The time course of enhancement of fluorescence at 440 nm observed upon adding e-aza-ATP to cardiac myosin-S1 is shown in Fig. 2. Addition of amounts of nucleotide substoichiometric to the number of myosin heads produces a rapid increase in fluorescence that is followed by a slow decay. kcat of e-aza-ATP hydrolysis obtained from the half-time of the exponential decrease in fluorescence enhancement is 0.02 s-l at 0 "C. The fluorescence 1 400 440 480 520 560 600 WAVELENGTH (nm) Fac. 1. Fluorescence emission spectra of c-ma-ATP in the presence and absence of cardiac myosin-S1. Both in the presence (dashed line) and absence (solid line) of cardiac myosin-S1 (16 p~) , the concentration of e-aza-ATP was 9 p~. The time course of fluorescence enhancement at each wavelength was monitored on addition of substoichiometric amounts of e-aza-ATP cardiac myosin-SI. An example of such data is shown in Fig. 2. The maximum fluorescence at each wavelength was plotted to give the spectrum. Excitation at 364 nm; band width; 10 nm. Conditions: 100 mM KC1, remains enhanced to about 30% of maximum even after hydrolysis has occurred; a similar fluorescence enhancement can be obtained by adding e-aza-ADP to myosin-Sl under the same conditions, suggesting that this complex, which has 30% of the maximal fluorescence enhancement obtained with taza-ATP, corresponds to M-a-ADP. On addition of ATP, the fluorescence intensity decreases to approximately that of the myosin-S1 plus that of the unbound nucleotide which was added (shown on the right). The rate at which t-aza-ADP is displaced from the nucleotide-myosin-S1 complex (M-a-ADP*) by ATP is 0.05 s-l at 0 "C. The time course of fluorescence enhancement observed upon addition of t-aza-ATP to cardiac myosin-S1 at 15 "C is similar to that at 0 "C (data not shown); the steady state rate of eaza-ATP hydrolysis which can be determined from the rate of decay of the fluorescence enhancement in single turnover experiments is 0.045 s-l.
Kinetics of Binding t-Aza-ATP to Cardiac Myosin-SI-The rate of the rapid transient enhancement of fluorescence of eaza-ATP when mixed with cardiac myosin-S1 can be measured in a stopped flow fluorometer. Fig. 3 shows examples of such data at 15 "C. The fluorescence transient can be fit to a single exponential, which is shown by the solid line through the trace (Fig. 3A). However, from 0 to 100 p~ e-aza-ATP, the fluorescent transients are better fit by two exponentials, as judged by the root mean square deviations of the data from the fit which are approximately 50% lower for the fit to two exponentials (Equation 2b) than for the single exponential fit (Equation 2a). In this concentration range, the fitted rate constants (kohl and k&2) differ by less than a factor of three making unique determination of the rate and amplitude data difficult; this can be seen by comparison of two possible combinations of values for kObs and amplitude coefficients for the data shown in Fig. 3B. Therefore, where t-aza-ATP < 100 p~, k b b S was obtained from single exponential fits of this data. At t-aza-ATP concentrations greater than 100 p~ (Fig. 3C), the values for the rate constants obtained by double exponential fits of the data are unambiguous. The dependence of k& I C-azo-ATP 1 t """""""""_"""""-  upon the concentration of t-aza-ATP is shown in Fig. 4. A rapid phase of the reaction corresponding to 40-50% of the fluorescence enhancement increases linearly with t-aza-ATP Concentration and is described by a apparent second order rate constant of -2 X lo5 s-' at 15 "c. The slower component of the fluorescence enhancement corresponds to 50-60% of the total amplitude. The relationship between kobs for this slow component and the concentration of e-aza-ATP can be described by a hyperbola with a maximum rate of 13.3 s-l as estimated by linear least squares analysis of the data.
The fluorescence transients observed on mixing t-aza-ATP with cardiac myosin-S1 were also monitored at 0 "C. At 0 "C, as at 15 "C, the transients were better fit by two exponential terms, although only the data at concentrations of t-aza-ATP > 100 PM could be unambiguously analyzed (data not shown).
The apparent second order rate constant for the enhancement of e-aza-ATP fluorescence is 5 X lo4 M-' s-l and the maximum rate of the slow phase of the fluorescence enhancement is 4-5 s-l. The apparent second order rate constants for the binding of t-aza-ATP to cardiac myosin-S1 at 0 and 15 "C are 2-10 times slower than the values for ATP obtained from the enhancement of tryptophan fluorescence under the same conditions, 1.1 X lo5 "' s-l and 2 X lo6 "' s-', respectively  Kinetics of the Dissociation of Cardiac Myosin-S1-t-aza-ADP by MgATP-The rate of dissociation of t-aza-ADP from cardiac myosin-S1 can be measured from the decrease in nucleotide fluorescence enhancement observed on mixing excess ATP with the myosin-S1-E-ma-ADP complex as shown in Equation is more rapid than the dissociation of t-aza-ADP from the complex. Fig. 3 0 shows an example of the data from such an experiment. e-Aza-ADP is displaced from the complex at a rate of 1.9 s-l, at 15 "C. This is -6 times faster than the rate constant measured for ADP dissociation from bovine cardiac myosin-SI (7).
Kinetics of the Fluorescence Enhancement on Binding t-Aza-ATP to Cardiac Actomyosin-SI-t-Aza-ATP binding to actomyosin-S1 may be measured from the enhancement of nucleotide fluorescence. Fig. 5 shows the increase in fluorescence that occurs upon mixing cardiac actomyosin-S1 with Eaza-ATP. The fluorescence data at 15 "C could be fit reasonably by a single exponential for concentrations of €-aza-ATP <lo p~ (Fig. 5A) and 40-80 p~ (Fig. 5C). In the range of 10-40 p~ 6-aza-ATP (Fig. 5B), the data were best fit by two exponential terms with amplitude coefficients of opposite sign. At concentrations of e-aza-ATP >lo0 pM, the fluorescence enhancement was biphasic (Fig. 5 0 ) and was best fit by two exponential terms with amplitude coefficients of the same sign. The excellent signal to noise ratio of the data, from approximately 0.5% root mean square at low eaza-ATP concentrations to 4% at the highest concentrations, enabled us to make reliable double exponential fits of the fluorescence data. At concentrations of t-aza-ATP >400 p~, the signal to noise ratio was not good enough to obtain reliable double exponential fits of the data. A maximum rate can be estimated to be -100 s-l from the curvature of the mdiac Myosin-S1 and Actomyosin-SI dependence of the rapid component upon t-aza-ATP concentration. The apparent second order rate constant of t-aza-ATP binding to actomyosin-S1 can be estimated from the slope of the dependence of kobs on t-aza-ATP concentration to be 5 x lo5 M-l s-' .
The kinetic data for the fluorescence enhancement observed on eaza-ATP binding to bovine cardiac actomyosin-S1 were modeled by the three-step mechanism shown in Equation 4a. The measured rate constants of kinetic steps associated with myosin intermediates that are in rapid equilibrium with actin may represent combinations of several rate constants: are fractional amplitude coefficients described in the legend to Fig. 3. Excitation was at 365 nm and emission from 415 to 445 nm. Experimental conditions are the same as those described in Fig. 3.
Experimentally measured rate constants of kinetic steps associated with myosin intermediates that are in rapid equilibrium with actin may represent combinations of several rate constants. For example, the kobs of the first fluorescence transition, which occurs after e-aza-ATP binds to actomyosin-Sl of the slow and fast processes upon nucleotide concentration. The value of klkz/k-l was estimated from the second order rate constant obtained from the experimental data at low nucleotide concentrations. A systematic investigation of the dependence of the observed rate constants and amplitude coefficients upon the microscopic rate constants defined by Equation 4a resulted in the following conclusions for the general model. 1) The initial assumption that the maximum rate of the slow component of the reaction at high nucleotide concentration is k3 + k-3 is correct. If k3 > k-3, the model predicts that observed rate constants and amplitude coefficients are essentially independent of kP3. Solutions in which k-3 > ks are not ruled out on kinetic grounds, but the intermediate (A)M-a-ATP** occurs indicating that the equilibrium constant for its formation (K3) cannot be unfavorable. 2 ) The initial assumption that the maximum rate of the fast component at high nucleotide concentration is & + k-2 is correct.
During the modeling, the rate constants of the fast processes were found to be insensitive to the value of k-z. In contrast, the concentration of c-aza-ATP at which the slow component of the fluorescence enhancement reached one-half of its max-imum rate was found to depend on the value of k--2. The dependence of the relative amplitude coefficients of the slow and fast components upon t-aza-ATP concentration is also very insensitive to the values chosen for h-2. No attempt was made to obtain the best fit of the rate constants by covariance methods. The solid lines drawn through the data in Fig. 6 were calculated for the values kl = 10' "' s-', k-1 = 2 X IO4 S-l, kz = 100 s-l, k+ = 0.2 s-', k3 + k-3 = 15 s-l and the fluorescence enhancements relative to t-aza-ATP (zero) were AM-a-ATP = 0, AM-a-ATP* = 0.5, and AM-a-ATP** = 1.0. The dependence of the calculated rates and amplitude coefficients is essentially independent of the values of kl and k-1 if kl > IO7 M" s-l and the ratio kl/k-l is kept constant (5 X IO3 " l ) .
One of the observed rate constants, which corresponds to nucleotide binding and the first fluorescence transition (AM + T c, A-M-T -A-M-T*) is shown by the solid line segments in Fig. 6. The other observed rate constant corresponding to the nucleotide binding steps (AM-T* c, AM-T**) is shown by the dashed line segments at kobs -15 s-l in Fig. 6. At nucleotide concentrations -40 PM, shown by the dotted line segments in Fig. 6, the two observed rate constants are nearly the same and assignment of either to a single physical process cannot be made.
The dependence of the fractional amplitude coefficient Ifaet/ ItOt on t-aza-ATP concentration agreed well with that predicted by the three-step sequential mechanism calculated using these parameters. In particular the model predicts that the fractional amplitude coefficient, Ifast/ltot is less than zero for concentrations of t-aza-ATP in the range 10-40 PM. It should be noted that an exponential term with a negative coefficient does not necessarily indicate that there is a decrease in fluorescence intensity associated with one step of the mechanism; rather, a negative amplitude coefficient can also occur for a two-step reaction in which the fluorescence enhancement of one step of the reaction is intermediate between the initial and final fluorescence intensities. At nucleotide concentrations above 100 p~ the mechanism predicts biphasic fluorescence data with the fast component accounting for 40-50% of the fluorescence signal at e-aza-ATP >200 p~. At concentrations 4 0 and 40-80 p~ the data are predicted to be essentially single exponential curves either because the amplitude Coefficient of the fast component is small (e-aza-ATP <lo /IM) or because the rate constants of the two components differ by less than 50% (40-80 PM) such that the two components cannot be resolved.
The fluorescence data at 0 "C can be fitted by the same model. The rate constants at 0 "C used were kl = 1 0 ' M" s-', k-l = 3 x lo4 s-l, kz = 50 s-', k-z = 0.1 s-' , and (k3 + k-3) = 5 s-l. Again, the amplitude coefficients of the two components and the rate constants measured at high nucleotide concentration predict the negative amplitude coefficient of the fast exponential term at intermediate nucleotide concentration.
The solution for a branched pathway mechanism, shown in Equation 5, is the sum of three exponentials if the nucleotide binding step is in rapid equilibrium, k-l > > & + kPz + k'2 + k'+. We have evaluated the observed rate constants and amplitude coefficients for a model in which the fluorescence transitions have the same rates but occur in opposite order in the upper and lower pathways. The dependence of two of the observed rate constants and amplitude coefficients obtained for the branched pathway mechanism were nearly the same as those previously obtained for the linear mechanism. The additional term in the branched pathway mechanism has an observed rate constant of kl3 + kL3, 100.2 s-l, that is independent of nucleotide concentration. The amplitude coefficient increases linearly from <0.01 at low nucleotide concentration to 0.1 at 400 p~. Such a small additional component would be too small to measure. Thus the linear and branched mechanisms are essentially indistinguishable. The data, which we have presented.here for 6-aza-ATP binding to bovine cardiac actomyosin-S1 can be explained by the simpler linear mechanism, although a more complex branched pathway mechanism of the type suggested by Rosenfeld and Taylor (12) for the mechanism of 1-N6-etheno-ATP binding to rabbit skeletal and chicken smooth muscle myosin-S1 cannot be excluded. Another possible class of mechanisms is illustrated by Equation 6. There are two species of actomyosin-S1, AM1 and AM2, that do not interconvert or interconvert slowly on the time scale of nucleotide binding.

AMl
AMl-a-ATP AMl-a-ATP* AM2 AM2-a-ATP e AM2-a-ATP* Such mechanisms would account for the dependence of the rates of the slow and fast processes upon nucleotide concentration but predict that either single exponential or biphasic kinetics would occur at all nucleotide concentrations. The type of kinetic behavior observed in Fig. 5B, in which there is a rapid component with a negative amplitude coefficient would not occur for this class of mechanism. Therefore the simplest mechanism that will explain the data of Fig. 6 for eaza-ATP binding to cardiac actomyosin-S1 is a linear mechanism of the class described in Equation 4a.
Kinetics of the Dissociation of Cardiac Actomyosin-S1 by 6-Aza-ATP-The decrease in light scattering intensity observed upon mixing cardiac actomyosin-S1 with e-aza-ATP at 15 "C is shown in Fig. 7. The dissociation of the actomyosin-S1 complex by 6-aza-ATP was complete as judged by comparison of the extent of the light scattering change accompanying the reaction with that observed for ATP. At low concentrations of 6-aza-ATP, the light scattering changes are fit reasonably well by a single exponential equation as shown in Fig. 7, A-C. The second order rate constant measured from the light scattering at 6-aza-ATP 4 0 0 pM, 5 X lo5 M-' s-' , is the same as that determined for the fluorescence enhancement. At nucleotide concentrations <60 p~? the decrease in light scattering intensity can be fit to a single exponential giving values for hobs similar to those of the fast fluorescence transient. However, at higher nucleotide concentrations better fits of the data could be obtained by fitting to two exponentials. For example, at 400 p~ e-aza-ATP, the more rapid component of the light scattering signal is 172 s-' and accounts for 55% of the amplitude of the light scattering change (Fig. 70). The slower component of the light scattering has a value of 38 s-l and contributes 45% of the signal.
The light scattering data at 0 "C are qualitatively similar.
The apparent second order rate constant estimated from the dependence of kobs on e-aza-ATP concentration (at values <40 p~) is 2 X IO5 M-' s-l. The data are also biphasic at higher nucleotide concentrations. For example, at 400 p~ 6-aza-ATP the light scattering transients are fit by two exponentials; the faster component at 92 s-l accounting for 45% of the signal and the slower component at 18 s-l accounting for 55% of the signal. Under identical experimental conditions, at 15 "C, the decrease in light scattering observed upon mixing cardiac actomyosin-S1 with ATP is single exponential at concentrations of nucleotide where this reaction is slow enough to be measureable, <400 s-l (Ref. 8 and confirmed here). However, if the decrease in light scattering on mixing ATP with actomyosin-S1 is monitored at 0 "C, the transients are biphasic at higher concentrations of ATP (>lo0 p~) ; for example, at 220 p~ ATP (fiial concentration), 54% of the total amplitude is contributed by a fast reaction 225 s-l and the remainder by a slower component of 20 s -~.~ The dissociation of actomyosin-S1 from chicken gizzard and heart by ATP at 20 "C (25) has also been shown to be biphasic.
We have considered several models to explain the fluorescence and light scattering data. For actomyosin-S1 the kinetic pathway shown in Equation 7 is consistent with the data obtained in the fluorescence studies and the light scattering data at low nucleotide concentrations. Rapid dissociation of M-a-ATP* from actin occurs between the fast and slow fluorescence transitions. The values of the rate constants indicated are those for 15 "C.

Ethem-aza Nucleotide Binding to
Cardiac Myosin-S1 and Actomyosin-S1 15153 diates into the mechanism. More complicated models, which can explain the fast component of the biphasic light scattering data require a fast dissociation step to occur prior to any fluorescence change. One such mechanism is given in Equation 8. The collision complex (AM-a-ATP) transforms to an additional AM-a-ATP intermediate that dissociates before the first fluorescence change. Kinetics of the Dissociation of Cardiac Actomyosin-S1-6-Aza-ADP by MgATP-The rate constant of t-aza-ADP dissociation from cardiac actomyosin-S1 can be measured from the rate at which 6-aza-ADP is displaced.from the complex by ATP. The reaction can be observed by monitoring the decrease in light scattering as the actomyosin-S1-t-aza-ADP complex dissociates upon mixing with ATP in a stop flow. At 0 "C, the light scattering data can be fairly well fitted to a single exponential at concentrations of ATP <50 p~. The dependence of the rate of dissociation of cardiac actomyosin-S1 upon the concentration of ATP, in the absence or presence of E-aza-ADP (50 @M final concentration) is shown in Fig. 8.
In the absence of t-aza-ADP, the apparent second order rate constant for the dissociation of cardiac actomyosin-S1 by ATP is 8 X IO5 M-' s-' at 0 "C. In the presence of t-aza-ADP (data shown by crosses), the apparent second order rate constant for dissociation is reduced to 2.5 x lo5 M-' s-' . At higher concentrations of ATP (>200 p~) , the rate constant for dissociation of the actomyosin-SI-6-aza-ADP complex reaches a plateau at 20 s-'. This behavior is similar to that of the cardiac actomyosin-S1-ADP complex (8) and can be quantitatively described by a model shown in Equation 9, in which t-aza-ADP acts as a competitive inhibitor of ATP binding at the active site of actomyosin-SI. The association constant of e-aza-ADP for cardiac actomyosin-S1, KAD, calculated from Equation 10 is 4.4 X lo4 "' at 0 "C. The dissociation of 6-aza-ADP from cardiac actomyosin-S1 could also be followed by monitoring the decrease in enhancement of fluorescence of t-aza-ADP on mixing the complex with ATP in the stop flow as indicated by the data shown by closed circles in Fig. 8. The smooth curves drawn through both sets of data are for a second order rate constant of 8 X lo5 M" s" in the absence of e-aza-ADP and rate and equilibrium constants for t-aza-ADP dissociation of 20 s-l and 25 pM. Solvent conditions were identical to those in Fig. 3. Excitation was at 365 nm. Light scattering was measured with a 365nm emission filter and fluorescence with a 450-nm filter with a 40nm band width emission filter. slow rates of approximately the same amplitude as shown by the boxed data points. More complete solutions to Equation 10, which are multi-exponential (8) predict that the reaction is biphasic over this range of ATP concentrations.
Similar experiments were carried out at 15 "C, monitoring the decrease in light scattering on mixing cardiac actomyosin-S1-e-aza-ADP with ATP. At this temperature, the binding of t-aza-ADP to actomyosin-S1 is about 3 times weaker than that at 0 "C, so that the fluorescence signal observed on dissociation of the actomyosin-S1-t-aza-ADP complex by ATP is too small to measure accurately; however, values of &b obtained from the light scattering transients also gave good fits to the fluorescence data at a given concentration of ATP. At high concentrations of ATP, the rate constant for dissociation reached a plateau at 110 s-'. The association constant of t-aza-ADP for cardiac actomyosin-Sl, KAD, calculated from Equation 10 is 1.4 x lo4 M-'. The second order rate constant for 6-aza-ADP binding to cardiac actomyosin-S1 (kAD) estimated from this data is 1.5 x lo6 "' s-'. Under similar conditions, ADP binds to cardiac actomyosin-S1 with an apparent second order rate constant 1 X lo7 M" s-' (8). Table I summarises the data obtained in these experiments for the binding of E-aza-ATP to, and the dissociation of 6-aza-ADP from, cardiac myosin-Sl and actomyosin-S1 and compares the values of equilibrium and rate constants with those for ATP and ADP. The rate constants of t-aza-ATP binding to cardiac myosin-S1 and actomyosin-S1 are 5-10-fold less  than the corresponding values for ATP binding. Actin increases the rate of 6-aza-ADP dissociation from cardiac myosin-Sl -60-fold at 15 "C and 400-fold at 0 "C. A similar increase in the rate of dissociation of ADP from myosin-S1 has been observed in the presence of actin (5,8).

DISCUSSION
t-Aza-ATP binding to bovine cardiac myosin-S1 and acto-myosin431 is a useful system for studying the mechanism of nucleotide triphosphate binding. The large fluorescence enhancement provides data with extremely good signal to noise ratio that are required to make reliable double exponential fits. In addition, the maximum rates of the reactions are sufficiently slow <IO0 s-l that amplitude data can be obtained without making large corrections for signal loss due to the dead time of the stopped flow. In contrast, when the intrinsic tryptophan fluorescence of the myosin head is used to monitor ATP binding, the relatively small signal size and large rate of the rapid phase of the reaction make such detailed analysis of the data difficult. The mechanism of t-aza-ATP binding and hydrolysis by cardiac actomyosin-S1 is kinetically complex. Work presented here indicates the existence of 2 sets of transitions between nucleotide bound actomyosin-S1, and between the corresponding myosin-S1 states. This is the first indication of this phenomenon with cardiac proteins, although similar observations have been made for the hydrolysis of 1-N6-etheno-ATP by skeletal and smooth muscle proteins (12). Several common features are apparent for the binding and hydrolysis of these two analogues, although a detailed comparison of the kinetic mechanism of the hydrolytic pathways of 1-N6-etheno-ATP by skeletal and gizzard actomyosin-Sl with the hydrolytic pathway of e-aza-ATP hydrolysis by car-diac actomyosin-Sl is not justified. For both nucleotides, complex kinetics are observed, which require mechanisms more complicated than a simple two-step binding mechanism in which a collision intermediate is followed by a conversion to a more tightly bound intermediate. These mechanisms may result in fluorescence data that are obviously the sum or difference of two exponentials (e-aza-ATP binding to cardiac myosin-S1 and actomyosin-S1; l-N'-etheno-ATP binding to skeletal and smooth muscle myosin-S1 and smooth muscle actomyosin-Sl) or data which are approximately fit by a single exponential equation (l-N'-etheno-ATP binding to skeletal actomyosin-Sl). The extent of deviation of the observed data from single exponential kinetics will be determined by the rate constants of each step of the mechanism and the fluorescence amplitudes of the transitions. Complex kinetics are observed for various different nucleotides under a variety of different experimental conditions (11, 24, 26). This suggests that such mechanisms, which are not of the simple two-step type, are a general feature of nucleotide binding to myosin-S1 and actomyosin-S1, although they may not be obvious under a particular set of experimental conditions. For example, a rapid initial phase of tryptophan fluorescence enhancement has been observed during ATP binding to skeletal (26) and cardiac myosin-Sl: but only at temperatures below 10 "C.
The biochemical and physiological properties of E-aza-ATP make it very useful as a substrate with which to investigate the cross-bridge cycle in muscle. A complete description of the kinetics of hydrolysis of t-aza-ATP in uitro should facilitate interpretation of the data from such studies on whole muscles or muscle fibers (13). Another approach to investi-gating the actomyosin ATP hydrolysis mechanism in vivo is to adapt biochemical techniques to study simple structured systems (27). The similarity of the overall hydrolysis mechanisms for t-aza-ATP and ATP makes this analogue useful to compare the rates of some of the steps of the nucleotide hydrolysis mechanism for myosin-SI, actomyosin-SI, and myofibrils. The results of such studies on cardiac contractile proteins are described in the accompanying paper (16) and provide a direct test of the rationale that the biochemical intermediates of nucleotide triphosphate hydrolysis which occur in solution are related to those in a structured system.