Equilibrium distribution of skeletal actin–tropomyosin– troponin states, determined by pyrene–tropomyosin fluorescence Boris Gafurov 1 and Joseph M. Chalovich 2 1 Uniformed Services University of the Health Sciences, Department of Pharmacology, Bethesda, MD, USA 2 Department of Biochemistry and Molecular Biology, Brody School of Medicine at East Carolina University, Greenville, NC, USA The ATPase activity of striated muscle myosin is low unless it is bound to actin. Actin activation is inhibited by the regulatory proteins tropomyosin, troponin T, troponin I and troponin C, which bind along actin fila- ments of skeletal and cardiac muscles. Activation of striated muscle contraction occurs when Ca 2+ binds to troponin C, or in a Ca 2+ -independent manner when rigor-type myosin binds to actin [1–3]. Myosin is both the target enzyme that hydrolyzes ATP and a potential allosteric activator. Much current work is devoted to understanding the structural and functional changes that occur in the large co-operative system consisting of myosin, actin, troponin and tropomyosin. Structural changes in troponin [4] and tropomyosin [5], in response to either Ca 2+ or myosin subfragment 1 (S1) binding, have been documented. The structure of actin is plastic [6] and could also change in response to the regulatory proteins. Keywords parallel pathway model; pyrene iodoacetamide; regulation of contraction; tropomyosin; troponin Correspondence Joseph M. Chalovich, Department of Biochemistry and Molecular Biology, Brody School of Medicine at East Carolina University, 5E-122 Brody Bldg, Greenville, NC 27834, USA Fax: +1 252 7443383 Tel: +1 252 7442973 E-mail: chalovichj@ecu.edu Website: http://www.ecu.edu/biochemistry/ Chalov.htm (Received 11 December 2006, revised 10 February 2007, accepted 1 March 2007) doi:10.1111/j.1742-4658.2007.05765.x Actin–tropomyosin–troponin has three structural states, but the functional properties of regulation can be explained with models having two func- tional states. As a step towards assigning functional properties to all the structural states, we examined fluorescent probes that monitor changes in troponin and tropomyosin. Tropomyosin labeled with pyrene–iodoacetamide is thought to reflect the transition to the most active state, where- as N-((2-iodoacetoxy)ethyl)-N-methyl)-amino-7-nitrobenz-2-oxa-1,3-diazole- labeled troponin I is thought to monitor the transition to any state other than the inactive state. The fraction of actin in an active state determined from pyrene excimer fluoresecence agreed with that calculated from light- scattering measurements of myosin subfragment 1 (S1)–ADP to regulated actin in both the presence and absence of Ca 2+ over a range of ionic strength conditions. The only exceptions were conditions where the binding of S1–ADP to actin was too strong to measure accurately. Pyrene–tropo- myosin excimer fluorescence was Ca 2+ dependent and so reflected the change in population caused by both Ca 2+ binding and S1–ADP binding. Pyrene labeling of tropomyosin did not cause a large perturbation of the transition among states of regulated actin. Using pyrene–tropomyosin fluorescence we were able to extend the ionic strength dependence of the parameters describing the co-operativity of binding of S1–ADP to actin as low as 0.1 m. The probes on tropomyosin and troponin I had different responses to Ca 2+ and S1–ADP binding. These different sensitivities can be explained by an intermediate between the inactive and active states of regulated actin. Abbreviations IANBD, N-(((2-iodoacetoxy)ethyl)-N-methyl)-amino-7-nitrobenz-2-oxa-1,3-diazole; S1, myosin subfragment 1. FEBS Journal 274 (2007) 2287–2299 ª 2007 The Authors Journal compilation ª 2007 FEBS 2287 Tropomyosin occupies three positions on actin (Fig. 1), depending on the amount of Ca 2+ bound to troponin and to the amount of S1 bound to actin [7–10]. These three structural states are in rapid equi- librium with each other [11–13], so that in each condi- tion there is a distribution of states [8]. Some models of regulation are constructed around the assumption that each structural state has a unique function. Other models use the minimum number of states required to simulate function. An ongoing question is what are the properties of these three states and how do they relate to regulation. Two types of regulatory models are shown in Fig. 2. In parallel pathway models (Fig. 2A,B), actin exists in two or three states, with discrete abilities to serve as cofactors for myosin-catalyzed ATP hydrolysis. The relative populations of these actin states are deter- mined by Ca 2+ and bound S1–ADP. The overall activ- ity of the system at any condition is defined by the fraction of time that myosin is bound to each of these actin states. More detailed schemes of a parallel path- way model, showing some steps in ATP hydrolysis, have been published previously [14,15]. The formalism for a parallel pathway model was originally defined for two functional states of actin, for simplicity [14,16]. Despite early concerns that a two-state model could not explain the binding kinetics, it has been shown to simulate equilibrium binding, binding kinetics and regulation of ATPase activity correctly [15]. Tropomy- osin is a switch, in the parallel model, that changes the structure of actin in some way that alters its ability to stimulate myosin ATPase activity [17]. Because the two-state model is able to explain many features of regulation, the properties of any intermediate state that may be present are undefined. The potential to define the intermediate state does exist if it can be observed in real time. In sequential models of regulation, actin passes from state A B (blocked) to A C (closed) to A O (open). In sequential models, one cannot define the activity of an individual state. Only state A O supports myosin activ- ity, so it is necessary to go stepwise from the blocked to the closed to the open states. The model shown in Fig. 2C is from McKillop & Geeves [18] and is based on the multiple-step binding of myosin to actin. Another model, that of Butters & Tobacman, has three states of actin with different orientations of tropomyo- sin that are in equilibrium with each other and with a fourth state, in which actin has undergone a conforma- tional change to an active state with a structure similar to that stabilized by the binding of myosin [19]. That model is not shown here, but it may be imagined as a funnel in which three states of regulated actin funnel to an active state that supports contraction. The models in Fig. 2 share the idea of multiple forms of regulated actin with different activities in equilibrium with each other. Changes in the distribu- 1 2 3 4 EGTA Ca 2+ rigor S1 bound Fig. 1. Cross-sections of actin–tropomyosin–troponin showing the structural states identified in the absence of Ca 2+ , with saturating Ca 2+ and with bound rigor-type myosin subfragment 1 (S1). Tropo- myosin is shown in black. The cross-section of an actin filament is shown in outline and the orientation of the four subdomains is shown. The dashed line is for reference. The figure is based on Craig & Lehman [51]. A B A C A O MA C MA O MA R K 1 K 2 K 1 K B K T C MA i MA a K 1 K 2 A A i A i(Ca) A a B α β A i A a Fig. 2. Models of regulation of striated muscle contraction. Actin is represented by the letter A with a subscript to designate its state; myosin is represented by the letter M. The large differences in interactions among different myosin-nucleotide states is not shown. Panels A and B show two-state and three-state parallel pathway models. In the two-state version, myosin binds to actin that is either in the inactive (A i ) or active (A a ) state. The distribution between A i and A a is determined by the fraction of troponin C (TnC) sites with bound Ca 2+ and the fraction of actin sites with bound rigor-type cross-bridges. Rapid ATP hydrolysis occurs when actin is in state A a . The three-state model shown in (B) considers the possibility that regulated actin that has bound Ca 2+ , but no rigor-type cross-bridges, has an intermediate level of activity. For simplicity, the binding to myosin is not shown for this case. In this model, state A a is active and state A i is inactive, but the properties of state A i(Ca) , are uncertain. Panel C shows a sequential model in which there are three states of actin namely blocked (A B ), closed (A C ) and open (A O ). Actin makes sequential transitions to the open state, A O , which is competent for accelerating ATP hydrolysis and proceeding into the force-producing state MA R . Distribution of actin–tropomyosin–troponin states B. S. Gafurov and J. M. Chalovich 2288 FEBS Journal 274 (2007) 2287–2299 ª 2007 The Authors Journal compilation ª 2007 FEBS tion of regulated actin states determine the activity of actin–myosin, regardless of whether that activity chan- ges as a normal regulatory response [14,20], or as a result of some mutations in troponin [21,22] or in experimentally produced mutations in tropomyosin [23]. Therefore, it is important to have reliable meth- ods of determining the state of actin in real time. This manuscript explores, in detail, a well-known method of monitoring the state of regulated actin. The fraction of actin in the active state can be meas- ured in real time by fluorescence changes of probes on troponin and tropomyosin. Probes on troponin I respond to both Ca 2+ binding and to S1–ADP binding. These probes give a good estimate of the changes in dis- tribution of regulated actin as S1 or S1–ADP binds to actin [11,12,24,25]. Resonance energy transfer measure- ments between probes on actin and troponin I [26] or troponin T [13] have also proven to be valuable for measuring the state of the actin filament. Changes in pyrene–tropomyosin fluorescence have been shown to be a measure of the fraction of actin in the active state [27]. Pyrene–tropomyosin excimer fluorescence was sensitive to activation by S1, but Ca 2+ had little effect [27,28]. Pyrene–tropomyosin excimer fluorescence did give the predicted change in regulated actin distribution as the amount of S1–ADP was altered, but its usefulness was only demonstrated at relatively high ionic strength. The response of pyrene–tropomyosin fluorescence to S1 binding led to the idea that this probe measures entry into the most active state of actin, but is insensitive to transitions to states of intermediate activity. We report here a comparison of pyrene–tropomyo- sin excimer fluorescence to predicted changes in the actin state that occur in response to Ca 2+ and S1–ADP binding under conditions ranging from 100 to 240 mm ionic strength. We also compare changes in pyrene excimer fluorescence with N-(((2-iodoacetoxy)- ethyl)-N-methyl)-amino-7-nitro benz-2-oxa-1,3-diazole (IANBD)-labeled troponin fluorescence when both probes are present on the same actin filament. The results can be readily explained by the presence of an intermediate between the inactive and fully active states of regulated actin. Pyrene excimer formation did not appreciably affect the distribution of actin states. Furthermore, pyrene excimer fluorescence gave reason- able estimates of the distribution of actin states at ionic strengths as low as 0.1 m, where it may be possi- ble to correlate these changes with ATPase activities. Results Regulated actin is predominantly in the inactive state in the absence of Ca 2+ and bound S1. Both Ca 2+ and S1–ADP bind more tightly to the active state of actin than to the inactive state, and stabilize the active state. Increasing concentrations of free S1–ADP results in a co-operative binding curve, indicating a transition from a lower affinity to a higher affinity state of actin– tropomyosin–troponin. This change in affinity is read- ily seen in the absence of Ca 2+ as sigmoidal increases in theta with increasing free S1–ADP concentrations (Fig. 3A–D). Changes in pyrene–actin fluorescence are often used to measure the binding of S1 to actin (Fig. 3; solid squares). In order to compare changes in pyrene–tropomyosin excimer fluorescence with changes in occupancy of actin with S1, we utilized light scatter- ing to measure binding (open circles). Light scattering measurements gave binding patterns that were similar to previous measurements using pyrene–actin fluores- cence (compare circles with solid squares). Theoretical curves, describing the relationship between theta and free S1–ADP, were produced by fitting the Hill model to the data at the four ionic strength conditions shown in Fig. 3. This fitting procedure produced values of K 1 ,K 2 ,L¢ and Y. Those parameters were used to pro- duce theoretical curves for p2, the fraction of actin in the active state shown by solid curves in Fig. 3. Figure 3A–D also shows that changes in pyrene– tropomyosin excimer fluorescence (triangles) followed the predicted changes in the fraction of actin in the active state. The agreement between the theoretical curves and the measurements was particularly good at higher ionic strengths where the measurements were most accurate. Deviations between the predicted values of p2 (solid curve) and the measured value (triangles) were apparent at 0.1 m ionic strength. Whereas exci- mer fluorescence (triangles) was low at zero free S1–ADP, the solid curve predicted from equilibrium binding data (circles) predicts an excess of 20% of the actin to be present in the active state. Values of p2 near zero would be consistent with known activities. That is, the p2 values determined from tropomyosin fluorescence are probably more reliable than those calculated from binding studies at low ionic strength. Values of equilibrium binding parameters, deter- mined in the absence of Ca 2+ as a function of ionic strength, are shown in Fig. 4A–C. The open symbols show the present results of binding of S1–ADP to actin containing troponin and pyrene-labeled tropomy- osin. Equilibrium binding parameters were calculated by fitting the Hill formalism to light scattering alone (circles), or to pyrene–tropomyosin fluorescence alone (triangles). The values of K 2 shown in Fig. 4A were independent of the type of fitting used and they agreed very well with earlier values determined from pyrene– actin fluorescence shown as solid squares. The model B. S. Gafurov and J. M. Chalovich Distribution of actin–tropomyosin–troponin states FEBS Journal 274 (2007) 2287–2299 ª 2007 The Authors Journal compilation ª 2007 FEBS 2289 is not particularly sensitive to values of K 1 , so these values are not shown. Figure 4B,C shows the parameters Y and L¢.Y decreased with increasing ionic strength, indicating a decreased tendency of adjacent regulatory units to exist in the same functional state. Values of Y, calculated from light scattering, were similar to those calculated from pyrene–tropomyosin fluorescence. However, val- ues of Y tended to be slightly lower for the pyrene– tropomyosin system than for the pyrene actin system examined earlier, shown as solid squares. It is unclear if this difference is a result of the different probes used. Values of L¢ tended to increase with increasing ionic strength. Therefore, high ionic strength stabilized the Fig. 3. Changes in light scattering (circles) and pyrene–tropomyosin fluorescence (triangles) as a function of free myosin subfragment 1 (S1)–ADP concentration in the absence (A–D) and presence (E–H) of Ca 2+ . Measurements were made at 0.1 (A, E), 0.12 (B, F), 0.18 (C, G) and 0.24 (D, H) molar ionic strengths. The curves shown with a dashed line are fits of the Hill model to the fraction of actin with bound S1, determined by light scattering. Curve fitting was per- formed simultaneously with paired data sets, in the presence and absence of Ca 2+ , to constrain the variables. Fractions of actin in the active state, p2, were calculated from the equilibrium binding param- eters (solid curves). Estimates of p2 determined from pyrene–tropo- myosin fluorescence (triangles) are also shown. Solid squares are from a previous study with pyrene actin [15] to show that similar values of theta are obtained by light scattering measurements and earlier pyrene-actin measurements. All measurements were made using skeletal troponin and tropomyosin under the following condi- tions: 0.3 l M phalloidin actin, 0.06 lM pyrene-labeled tropomyosin, 0.06 l M troponin, 25 °C, in a buffer containing 20 mM Mops, pH 7.0, 5m M MgCl 2 ,1mM dithiothreitol, 2 mM ADP, 0.2 mgÆmL )1 bovine serum albumin, sufficient KCl to reach the target ionic strength and either 1 m M EGTA (A–D) or 0.1 mM CaCl 2 (E–H). Fig. 4. Effect of ionic strength on equilibrium binding parameters in the absence (A–C) and presence (D–E) of Ca 2+ . Values of K 2 (A, D), Y (B, E) and L¢ (C, F), determined by light scattering (circles) and pyrene-excimer fluorescence (triangles), are compared with earlier values determined from pyrene–actin fluorescence (solid squares). Values obtained from light scattering were obtained by a global fit of the model to data obtained at zero and saturating Ca 2+ . Earlier values from pyrene–actin fluorescence were the result of a global fit of data from six different free Ca 2+ concentrations but the same ionic strength [15]. The conditions were the same as for Fig. 3, with 1 m M EGTA used in the experiments with results shown in panels A–C and 0.1 m M CaCl 2 used in the experiments with results shown in panels D–F. Distribution of actin–tropomyosin–troponin states B. S. Gafurov and J. M. Chalovich 2290 FEBS Journal 274 (2007) 2287–2299 ª 2007 The Authors Journal compilation ª 2007 FEBS inactive state of regulated actin relative to the active state when no rigor type S1 was bound to actin. L¢ values were similar when determined by S1–ADP bind- ing or by tropomyosin–pyrene excimer fluorescence, and the results were in general agreement with earlier pyrene–actin fluorescence measurements. To determine the relationship of K 2 , Y and L¢ to ionic strength in Ca 2+ , we first determined the effect of Ca 2+ on fluorescence so that the initial point of p2 could be defined. Figure 5 shows pyrene–tropomyosin fluorescence measurements of regulated actin as a function of Ca 2+ concentration at 180 mm ionic strength. In 0.1 mm EGTA, the free Ca 2+ was below that required for activation (open circles). The pyrene fluorescence intensity increased to a maximum value when Ca 2+ exceeded the EGTA concentration. A con- trol experiment was performed in the absence of EGTA (solid circles). As expected, there was no change in fluorescence with the addition of Ca 2+ because the initial Ca 2+ concentration was already high enough to give the full effect. We performed another control by comparing the effects of Ca 2+ on probes on both tropomyosin and troponin. Actin filaments were reconstituted with pyrene-labeled tropomyosin and troponin containing IANBD-labeled troponin I. Figure 6A shows that the addition of excess Ca 2+ to an EGTA-containing solu- tion caused 40% of the maximum possible change in pyrene–tropomyosin fluorescence, but, on average, 92% of the maximum in IANBD–troponin I fluores- cence. The complete change of pyrene–tropomyosin required the addition of nucleotide-free S1. Figure 6B compares the effect of both probes to the addition of S1 in the absence of Ca 2+ . Although the changes are in opposite directions, the sensitivities to S1 concentra- tion were similar. Knowing the value of p2 to be 0.4, in the absence of S1–ADP we were able to examine the relationship between predicted values of p2 and pyrene excimer fluorescence in the presence of Ca 2+ . Figure 3E–H shows light scattering and pyrene excimer fluorescence at four ionic strength conditions at saturating Ca 2+ . Values of p2 reached their maximum values at subsat- urating concentrations of S1–ADP in all cases. The Fig. 5. The fluorescence of actin filaments reconstituted with pyrene-labeled tropomyosin is Ca 2+ dependent at 180 mM ionic strength. Pyrene–tropomyosin fluorescence was measured in the presence (open circles) or absence (closed circles) of 0.1 m M EGTA. The curve obtained in the presence of EGTA shows the increase in fluorescence as the total Ca 2+ concentration was increased. The conditions were the same as for Fig. 3. Fig. 6. Fluorescence changes in pyrene-labeled tropomyosin (cir- cles, solid lines) and N-(((2-iodoacetoxy)ethyl)-N-methyl)-amino- 7-nitrobenz-2-oxa-1,3-diazole (IANBD)-labeled troponin I (squares, broken lines) upon titration of actin–tropomyosin–troponin with Ca 2+ and myosin subfragment 1 (S1). Both fluorescent probes were present in the actin filament at the same time and the fluor- escence changes of each probe were measured about 10 min after each addition of S1. (A) Effect of adding 1.2 m M Ca 2+ to the EGTA-containing solution and then subsequently adding S1. The response to Ca 2+ was more extreme for IANBD–troponin I than for pyrene-labeled tropomyosin. Multiple lines are from emission measurements made at 10 nm wavelength increments. (B) Titra- tion of regulated actin containing both probes with S1 in the absence of Ca 2+ . The conditions were the same as for Fig. 3, with 150 m M KCl. B. S. Gafurov and J. M. Chalovich Distribution of actin–tropomyosin–troponin states FEBS Journal 274 (2007) 2287–2299 ª 2007 The Authors Journal compilation ª 2007 FEBS 2291 tropomyosin transition measured by pyrene fluores- cence was not co-operative in the presence of Ca 2+ . The dashed lines are fits of the Hill model to the val- ues of light scattering data, and the predicted curves for p2 are shown as solid lines. The measured values of p2 were similar to the predicted values. Poor fits to light scattering data, as in Fig. 3H, were, in part, a result of the fact that these were not best fits to a sin- gle data set, but were global fits to data in the presence and absence of Ca 2+ . The ionic strength dependencies of K 2 , Y and L¢, determined by fitting the Hill model to the data of Fig. 3E–H, are shown in Fig. 4D–F. The agreement of values of K 2 , Y and L¢ was good between light scatter- ing (circles) and pyrene–tropomyosin fluorescence (tri- angles) measured on the same proteins. Values of K 2 were similar to those measured in the absence of Ca 2+ . Values of Y were near 1 at low ionic strength and decreased slightly as the ionic strength was raised. If Y was constrained to be greater than 1, the value of Y would be 1 over the ionic strength range (data not shown). Values of L¢, determined by both methods, increased with increasing ionic strength as they did in the absence of Ca 2+ . Values of Y and L¢ were substantially different for actin filaments containing pyrene-labeled tropomyosin compared with those with pyrene on the actin. Fitting was generally more difficult in the presence of Ca 2+ because of the lack of features in those curves. Estima- tions of L¢ and Y are problematic because changes in the value of Y can be compensated, to some extent, for changes in L¢. The parameter, p2, defines the activity of the actin filament in parallel pathway models. Under conditions where all of the S1-ATP is bound to actin, the ATPase activity is approximately equal to p2*r max + (1 ) p2)*r min , where r max and r min are the rates for the active and inactive actin species, respectively. A correc- tion to this equation can be made for the small differ- ence in affinity of S1-ATP for actin in states 1 and 2. Values of r max and r min can be determined from the k cat for actin in the active and inactive states, respect- ively. Although these ATPase parameters have not been determined at the conditions used for the binding experiments, relative changes in ATPase activity can be approximated by observation of changes in p2. Figure 7 shows how p2 would change if actin filaments were activated by the attachment of an activating form of S1, such as S1–ADP. The inset shows values of p2 as a function of the square root of the ionic strength. The difference between the EGTA and Ca 2+ rates are expected to be approximately constant over the range of ionic strengths examined. Discussion Transitions between the inactive and active states of regulated actin are important determinants of the regu- lation of striated muscle contraction. The distribution of these states determines the ATPase activity, whereas the rates of transitions among the states may affect the rate of force redevelopment [11]. Some disease-causing mutations in troponin T change in the distribution between the states of regulated actin [21,22]. The abil- ity to measure state transitions rapidly and relate them to function will be helpful in studying such defects. Fluorescent probes on troponin and tropomyosin have the potential to measure the distribution of states in real time. Ishii & Lehrer reported that probes on tropomyosin reflect changes in the fraction of actin in the active state resulting from S1 binding [27]. Acrylodan-labeled tropomyosin was useful for actin–tropomyosin, but the signal was too small in the presence of troponin [29]. Pyrene-labeled tropomyosin was the prefered probe for actin–tropomyosin and actin–tropomyosin–troponin [27,28]. Pyrene–iodoacetamide labeling was preferred over pyrene–maleimide labeling because of the rapid response of its excimer fluorescence to S1 binding [27]. The S1-induced increase in excimer fluorescence is caused by an increase in the fraction of pyrene mole- cules forming excimers. Pyrene–iodoacetamide-labeled tropomyosin excimer fluorescence exhibited a small change with Ca 2+ at low ionic strength. Because of these considerations, we have examined more closely Fig. 7. Calculated probabilities of actin–tropomyosin–troponin in the active state (p2) in the presence (solid lines and solid circles) and absence (dashed line and open circles) of Ca 2+ . Simulations were made from equilibrium binding parameters determined at 120 m M ionic strength. The inset shows how values of p2 in the absence of added myosin subfragment 1 (S1) change as a function of the square root of the ionic strength. Distribution of actin–tropomyosin–troponin states B. S. Gafurov and J. M. Chalovich 2292 FEBS Journal 274 (2007) 2287–2299 ª 2007 The Authors Journal compilation ª 2007 FEBS the suitability of pyrene–tropomyosin excimer fluores- cence as a measure of regulated actin state changes. We studied tropomyosin excimer fluorescence over a range of ionic strength conditions because ATPase measurements and S1–ADP binding cannot be readily measured under the same conditions and an extrapola- tion of parameters is necessary. Furthermore, examin- ing the behavior at different conditions increases the reliability of parameters obtained by curve fitting [15,21]. Values of the fraction of actin in the active state, p2, were calculated from S1–ADP binding (light scattering). Values of p2 were also directly measured by pyrene excimer fluorescence. Pyrene excimer fluores- cence generally agreed with the predicted values of p2. Deviations occurred when S1–ADP binding became too strong to measure accurately. In those cases, exci- mer fluorescence was a more reliable measure of p2. To determine if the energetics of formation of tropo- myosin–pyrene excimers biased the distribution of actin states, we compared the present results with ear- lier studies where binding was measured with pyrene- labeled actin and unlabeled tropomyosin. Values of L¢ obtained from light scattering measurements with pyrene-labeled tropomyosin in the absence of Ca 2+ are in reasonable agreement with earlier values where there was no excimer formation (Fig. 4). Pyrene probes on tropomyosin did not significantly alter the values of K2, L¢ or Y at any ionic strength examined. Further- more, when troponin containing an IANBD probe on troponin I was reconstituted with N-(1-pyrene)iodo- acetamide (pyrene–iodoacetamide)-labeled tropomyosin and actin, the IANBD probe retained its typical responses to changes in Ca 2+ and S1 binding (Fig. 6). Fitting binding curves to obtain binding parameters is difficult in the case of Ca 2+ because the curves are featureless hyperbolas. Although we observed only small differences in binding curves measured with pyrene–actin and pyrene–tropomyosin in Ca 2+ (Fig. 3G–H), there was poor agreement between the values of L¢ calculated in the two cases. We also noted that at low ionic strength the values of Y tended to be greater in the presence of Ca 2+ , but this was not observed in the present case with unlabeled actin. It is also worth pointing out that the parameters determined in our earlier study with pyrene–actin resulted from global fits of the data. This change in fitting may contribute to differences in the final values of the parameters. The parameters K2, L¢ and Y varied with ionic strength, in agreement with our earlier observations [15,21]. High ionic strength decreased the fraction of regulated actin in active states (increased L¢). This result is consistent with in vitro motility assays where higher Ca 2+ is required for full activation at high ionic strength [30]. This trend has now been observed from 0.1 to 0.24 m ionic strength. The extension of this result to the lower ionic strength range is useful for extrapolation of the values for future simulation of ATPase activities under conditions where they can be readily measured. Tropomyosin excimer fluorescence was Ca 2+ dependent, but it did not directly track Ca 2+ binding. Rather, the change was consistent with a state change, such as partial transition, to the most active state of actin or a total transition to an intermediate state. Ca 2+ binding resulted in % 40% of the maximum observed change of excimer fluorescence obtained with full activation by rigor-type myosin binding. This agrees with the observation of Williams et al., that Ca 2+ alone provides 40% of the maximum value of k cat [31]. In vitro motility assays support the view that Ca 2+ alone does not provide full activation of regulated actin. High levels of loading of filaments with myosin produced about a doubling of the rate at saturating Ca 2+ [32] and a velocity 1.8 times higher than that of unregulated actin [33]. Activities that exceed actin alone are probably the result of partial stabilization of the most active state of regulated actin. In the case of cardiac troponin–tropomyosin, this extra activation was only evident for some disease-causing mutations of troponin [34]. Under those conditions, the velocity was increased 1.6-fold over full activation of the wild- type cardiac troponin. Some mutations have the effect of partially stabilizing the fully active state [21], so this 1.6-fold increase is probably an underestimate of the maximum level of activation. These results suggest that in the motility assay, Ca 2+ alone produces 50–55% of the maximum activation. The results could be closer to the 40% activation seen in solution for Ca 2+ alone if the actin filaments in the in vitro studies were not max- imally activated. The ability of S1–ADP and rigor S1 to activate actin filaments raises the question of how an active muscle can relax once the free Ca 2+ concentration is decreased. A larger fraction of strongly bound cross- bridges is required for activation in EGTA than in Ca 2+ . However, in EGTA at 0.18 m ionic strength, 30% saturation of the actin produces thin filaments that are 50% active (Fig. 3C). A 90% relaxation would require less than 5% of the actin to contain strongly bound cross-bridges. However, muscle may not behave identically to the proteins in solution. Geometrical considerations, and the presence of other protein components or small molecules, could result in a considerable shift of the curves shown in Fig. 3. B. S. Gafurov and J. M. Chalovich Distribution of actin–tropomyosin–troponin states FEBS Journal 274 (2007) 2287–2299 ª 2007 The Authors Journal compilation ª 2007 FEBS 2293 Probes on troponin I report changes in the state of regulated actin caused by S1 binding to actin [12] and also respond directly to changes in Ca 2+ [12,24,25]. Different sensitivities of fluorescent probes to Ca 2+ have been used in the past to argue for the presence of an intermediate state of regulated actin. Because the probes can affect the behavior of the regulatory com- plex, it is difficult to compare directly the results of probes on separate regulatory complexes. We have now utilized IANBD on troponin I and pyrene on tropomyosin within the same regulatory complex. Both probes responded to S1 binding in a similar way (Fig. 6A), but exhibited different responses to Ca 2+ (Fig. 6B). This result is consistent with the existence of an intermediate structural state [7]. We used the two-state parallel pathway model of Hill et al. for predicting the fraction of actin in the act- ive state. That model is consistent with the measured effects of Ca 2+ on binding in the presence of ATP [35,36], equilibrium binding in the presence of ADP [16], binding kinetics [15,37] and the general features of ATPase activities [14]. Furthermore, in our view, the functionally indistinguishable state is not the first state of a series, but rather the state corresponding to bound Ca 2+ and no bound rigor S1 (A i(Ca) in Fig. 2B). That intermediate may resemble the inactive (A i(EGTA) ) or fully active (A a ) states in terms of key functional properties. Although our results can be explained with two functional states, there is evidence for three structural states of regulated actin. Pirani et al. estimated the dis- tributions of structural states by image reconstruction of electron micrographs following dilution of the pro- teins to low ionic strength [8]. They predicted 22% of the actin to be in the closed state in the absence of Ca 2+ [8]. Because the actin filament is has little activ- ity in EGTA [31], the closed state must be inactive. Pirani et al. predicted the distribution in Ca 2+ to be 20% blocked, 68% closed and 12% M state (active state). The 40% activation, predicted in the present study, from tropomyosin fluorescence does not agree with this distribution. This could be an indication that there is not a simple correlation between observed struc- tural states and functional states of regulated actin. We also evaluated our results in terms of the three- state sequential model of regulation proposed by McKillop & Geeves [18], as shown in Fig. 2C. The increased rate of binding of S1–ADP to regulated actin in Ca 2+ compared with EGTA was interpreted, in that model, as 75% of actin sites being blocked in the absence of Ca 2+ . We have an alternative explanation for that effect [37]. However, for the present exercise we forced the fit to populate the blocked state in EGTA in accordance with their model. We used most of the constraints set by McKillop & Geeves, namely, 0 < KB < 10, 0 < KT < 20, 0 < N < 7, 10 3 <K1<10 6 and K2 ¼ 200. We did not constrain the values of ‘n’ and we consequently obtained a dif- ferent pattern of changes in this parameter. The simu- lations shown in Fig. 8 demonstrate that populations of both the blocked and closed states decreased with increasing amounts of bound S1 in both the absence and presence of Ca 2+ . The population of the open state was much higher in Ca 2+ than in EGTA in the absence of bound S1. Regulated actin was almost exclusively in the open state at saturating S1, irrespect- ive of the Ca 2+ concentration. Whereas the population of the open state does not correlate directly with our predicted p2, they do follow the same trend. Tropomyosin–pyrene excimer fluorescence gives a good estimate of the fraction of actin in the active state over a range of conditions. Simultaneous mea- surements of probes on tropomyosin and troponin give evidence for an intermediate state. By taking further advantage of this system, it may be possible to determine the role of this intermediate in regulation. Fig. 8. Distribution of the blocked (circles), closed (triangles) and open (squares) states in the course of myosin subfragment 1 (S1) binding. (A) The predicted occupancy of the states at 0.18 M ionic strength in the pres- ence of 0.1 m M Ca 2+ . The diamonds are the p2 parameter that represents the transition of the actin filament into the active state in Hill’s model. (B) The same parameters in the Ca 2+ -free case. Distribution of actin–tropomyosin–troponin states B. S. Gafurov and J. M. Chalovich 2294 FEBS Journal 274 (2007) 2287–2299 ª 2007 The Authors Journal compilation ª 2007 FEBS This is particularly important for the study of disor- ders of the regulatory system. Experimental procedures Protein preparation Actin [38,39], myosin [40], troponin and tropomyosin [41] were isolated from rabbit back muscle. Myosin S1 was made by digestion of myosin with chymotrypsin [42]. Pro- tein concentrations were determined by light absorbance at 280 nm, corrected for scattering, at 340 nm, using the fol- lowing extinction coefficients (e 0.1% ) for 280 nm: actin, )1.15; myosin-S1, )0.75; tropomyosin, )0.33; and troponin, )0.37. The molecular masses assumed for the key proteins were: actin, )42 000 Da; myosin-S1, )120 000 Da; tropo- myosin, )68 000 Da; troponin, )71 000 Da. Actin was stored as a 40 lm stock in 4 mm imidazole (pH 7.0), 1 mm dithiothreitol, 2 mm MgCl 2 and 40 lm phalloidin. Actin was sometimes labeled with N-(1-pyrenyl) iodoacetamide [43]. Tropomyosin was labeled with N-(1-pyrene)iodoacetamide (pyrene–iodoacetamide) [27]. In some cases, troponin I was labeled with IANBD [12]. The extents of labeling were % 60% and 35% for tropomyosin and troponin, respectively. Reconstituted actin was pre- pared by mixing actin with troponin and pyrene-labeled tropomyosin in a 3 : 1 : 1 molar ratio to ensure saturation of actin at the low concentrations used for binding studies. Equilibrium fluorescence measurements Equilibrium fluorescence measurements were made on an Aminco Bowman II Luminescence Spectrometer (Thermo Electron Corp., Madison, WI, USA), having the cell com- partment maintained at 25 °C with a circulating water bath. For light scattering measurements, the excitation and emis- sion monochrometers were set at the same wavelength. Exci- tation and emission wavelengths used were 340 and 480 nm, respectively, for tropomyosin–pyrene excimer fluorescence and 492 and 536 nm, respectively, for IANBD–troponin fluorescence. Equilibrium titrations of actin with S1–ADP were per- formed by observing the light scattering, pyrene–tropomyo- sin fluorescence [44] and by quenching of pyrene–actin fluorescence [43,45,46]. Details of the binding measurements are described elsewhere [15] and are similar to those des- cribed by others [46,47]. Our binding solutions contained 20 mm Mops, pH 7.0, 5 mm MgCl 2 ,1mm dithiothreitol, 2mm ADP, 0.2 mgÆmL )1 bovine serum albumin, sufficient KCl to reach the target ionic strength and 0.1 mm CaCl 2 or 1mm EGTA. The actin concentration in binding experi- ments was 0.3 lm. Solutions also contained 14 unitsÆmL of hexokinase and 1 mm glucose to scavenge ATP and 20 lm Ap5A to inhibit ATP formation through the myokinase reac- tion. Titrations were carried out by the stepwise addition of S1 to a 2 mL volume of pyrene-labeled actin–tropomyosin– troponin at 5 min intervals. This time interval was important to ensure equilibrium at each step. Fluorescence intensities and protein concentrations were corrected for the volume change (< 10%) caused by adding S1. Rabbit skeletal tropo- nin and tropomyosin were used in this study for comparison with our existing data for those regulatory proteins. Values of theta (S1 bound to the actin total ratio) and the free S1 concentration from fluorescence or light scatter- ing measurements were calculated using the equations: h ¼ F i À F min F max À F min         ½FreeS1¼½S1 total À h ýActin total ð1Þ Where F i is the fluorescence or scattering intensity at a total S1 concentration of i (lm); and F max and F min are the maxi- mum and minimum values of intensity, respectively. Modeling experimental results Light scattering was used to measure the binding of S1– ADP to actin and tropomyosin. Pyrene excimer fluorescence was used to monitor the fraction of actin in the active state. Equilibrium-binding parameters were extracted from light scattering data by using the co-operative binding model of Hill et al. [16] or by the model of McKillop & Geeves [18]. Fitting to the parallel pathway model of Hill was described in detail earlier [15]. Briefly, the relationship between the fraction of actin with bound S1 and the free S1 concentra- tion can be described by the following equations [16]: h ¼ p 1 h 1 þ p 2 h 2 h i ¼ K i C 1 þ K i C p 1 ¼ 1 À p 2 p 2 ¼ 2a Y ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ð1 À aÞ 2 þ 4 a Y q 1 À a þ ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ð1 À aÞ 2 þ 4 a Y q  a ¼ ð1 þ K 2 CÞ n ð1 þ K 1 CÞ n Y 22 ðqÞ LY 11 ðqÞ L 0 ¼ LY 11 ðqÞ Y 22 ðqÞ Y ¼ Y 11 ðqÞY 22 ðqÞ Y 12 ðqÞY 21 ðqÞ Y ij ðqÞ¼x ij þ 2k a qy ij þ k 2 b q 2 z ij 9 > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > = > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > ; ð2Þ p 1 and p 2 are fractions of actin units in the inactive and active states, respectively; h I and h 2 are fractions of actin containing bound S1 in the inactive and active states, respectively; K 1 and K 2 are S1-binding constants to the B. S. Gafurov and J. M. Chalovich Distribution of actin–tropomyosin–troponin states FEBS Journal 274 (2007) 2287–2299 ª 2007 The Authors Journal compilation ª 2007 FEBS 2295 inactive and active states of actin, respectively; C is the free S1 concentration; q is the free Ca 2+ concentration; n is the number of actin monomers in one actin–tropomyosin– troponin unit (assumed to be seven); L is the equilibrium constant for transition of an isolated actin–tropomyosin– troponin unit with no neighbors, no bound Ca 2+ and no bound S1 from state 2 to state 1; L¢ is the equilibrium con- stant defining the transition from the active state to the inactive state for the entire actin filament, but without S1; Y describes the co-operativity between adjacent regulatory units of seven actin monomers; Y is the overall co-operativ- ity parameter; Y ij are individual co-operative interactions between units in states i and j (we assumed that Y ij ¼ Y ji ); x ij , y ij and z ij represent the free energies of nearest neighbor tropomyosin interactions (W ij ) in exponential form e –Wij ⁄ kT [16]; and k a and k b are affinities of troponin in states 1 and 2 for Ca 2+ with values of % 10 5 and 10 6 Æm )1 , respectively [48]. We assumed that the values of k a,b did not change over the ionic strength range in this study. The simulated curves were not very sensitive to the value of K 1 , so simula- tions were normally carried out with the assumption that K 1 ¼ K 2 ⁄ 8 [49]. All measurements were carried out in both Ca 2+ -free and in Ca 2+ -saturated conditions. Binding data obtained at high and low Ca 2+ , but at the same ionic strength, were analyzed using a global fit procedure [15]. The global fit helped to constrain the parameters. Values of L¢, K 2 and Y obtained from the fits were used to simulate p 2 , the fraction of actin in the active state. We also fitted theoretical values of p 2 to the tropomyosin fluorescence to obtain L¢, K 2 and Y. From those values we were able to calculate curves of h as a function of free S1–ADP. Tropomyosin fluorescence was normalized from 0 to 1 in the absence of Ca 2+ because we assumed that essentially none of the actin was in the active state in the absence of Ca 2+ and bound S1. This assumption is reasonable based on ATPase activity measurements. The flux is proportional to the amount of S1 bound to each state multiplied by the k cat associated with that state. Ca 2+ increases the k cat by % 22-fold [20], whereas the binding of NEM-S1 increases the k cat by a further 2.5-fold [31]. This means that the fraction in the active state in EGTA is 1.8%. Binding studies were car- ried out at higher ionic strength conditions than the ATPase measurements. Because the fraction of actin in the active state decreases with increasing ionic strength [15], the value of 1.8% is an upper limit. The ATPase rates also predict that in the presence of Ca 2+ alone, 40% of the regulated actin is in the active state. Again, this fraction is also likely to be an upper limit because of ionic strength considerations. In order to define the fraction of actin in the active state in the presence of Ca 2+ , but in the absence of bound S1, we observed the changes in fluorescence that occurred dur- ing Ca 2+ titrations. With measured values of the initial value in EGTA, the change that occurred with the addition of Ca 2+ and the further change that occurred with satur- ating S1–ADP, we were able to calculate the initial p2 in Ca 2+ . The fluorescence data in Ca 2+ were normalized from this initial value to 1.0 for the maximum fluorescence observed in the presence of both Ca 2+ and saturating S1–ADP. Although the initial raw fluorescence values were higher in Ca 2+ than in EGTA, the values at saturating S1 were about the same in both cases. Fitting parameters and constraints were similar to the ones used in our earlier work [15]. Global fitting was per- formed in the mlab Modeling System (Civilized Software, Bethesda, MD, USA) and always produced reasonable fits with correlation coefficients R 2 > 0.85. Analysis using the model of McKillop & Geeves Because the original two-state parallel pathway model of Hill was able to account for the present data, the model was not expanded to include a third state. We did, however, ana- lyze some of these data with the three-state sequential model of McKillop & Geeves [18], shown in Fig. 2C. We fitted the model expressed in Eqn (3) to our binding isotherms and obtained key binding parameters K 1 , K 2 , K B , K T and n for each ionic strength and Ca 2+ concentration used: h ¼ K 1 cðK T ð1 þ K 2 ÞP nÀ1 ÞþQ nÀ1 K T P n þ Q n þ 1=K B ð3Þ P ¼ 1 þ K 1 cð1 þ K 2 Þ Q ¼ 1 þ K 1 c where K 1 and K 2 are S1-binding constants, K B is the equi- librium constant for proceeding from the blocked to the closed state, K T is the equilibrium constant for proceeding from the closed state to the open state, and n is a number of actin monomers forming a co-operative unit. We used constraints similar to those described elsewhere [15,50]. We determined the occupancy of the various states as a function of S1 bound by using differential equations to des- cribe the probability for each state [37]. Curve fitting was carried out to our binding isotherms at 180 mm ionic strength, measured with or without Ca 2+ . The 3 · 3 scheme of the kinetic reactions, which take place when n ¼ 1, is shown below, as derived previously [37]: 1 2 4 3 56 a B a -B ck 1 k -1 a T a -T ck 1 k -1 a T a -T k 2 k -2 Scheme 1. Distribution of actin–tropomyosin–troponin states B. S. Gafurov and J. M. Chalovich 2296 FEBS Journal 274 (2007) 2287–2299 ª 2007 The Authors Journal compilation ª 2007 FEBS [...]... for a period of time sufcient for the state to reach equilibrium, as dened by a lack of change in any of the states with time At equilibrium the fractions of the blocked, closed and open states are given as follows, respectively: Blocked ẳ p1, Closed ẳ p2 + p3, Open ẳ p4 + p5 + p6 We solved these equations at a series of S1 concentrations to obtain the occupancy of the states as a function of S1 bound... rates of switching movement of troponin T between three states of skeletal muscle thin laments determined by uorescence resonance energy transfer J Biol Chem 280, 26132619 Hill TL, Eisenberg E & Chalovich JM (1981) Theoretical models for cooperative steady-state ATPase activity FEBS Journal 274 (2007) 22872299 ê 2007 The Authors Journal compilation ê 2007 FEBS 2297 Distribution of actintropomyosintroponin... mutations in troponin I (K183D, G203S, K206Q) enhance lament sliding Physiol Genomics 14, 117128 Chalovich JM, Chock PB & Eisenberg E (1981) Mechanism of action of troponin- tropomyosin: inhibition of actomyosin ATPase activity without inhibition of myosin binding to actin J Biol Chem 256, 575578 Tobacman LS & Adelstein RS (1986) Mechanism of regulation of cardiac actin-myosin subfragment 1 by troponin- tropomyosin... measured by uorescence changes of troponin I modied with different uorophores J Biol Chem 261, 12791285 Brenner B, Kraft T, Yu LC & Chalovich JM (1999) Thin lament activation probed by uorescence of N-(2-(Iodoacetoxy) ethyl)-N-methyl) amino-7-nitrobenz2-oxa-1,3-diazole-labeled troponin I incorporated into skinned bers of rabbit psoas muscle Biophys J 77, 26772691 Miki M & Iio T (1993) Kinetics of structural... Chalovich JM (1999) Kinetics of thin lament activation probed by uorescence of N-(2-(Iodoacetoxy) ethyl)-N-methyl) amino-7-nitrobenz-2-oxa-1, 3-diazole-labeled troponin I incorporated into skinned bers of rabbit psoas muscle: implications for regulation of muscle contraction Biophys J 77, 26922708 Trybus KM & Taylor EW (1980) Kinetic studies of the cooperative binding of subfragment 1 to regulated... studies of the state of tropomyosin in reconstituted muscle thin laments Biochemistry 26, 49224924 Dancker P, Low I, Hasselbach W & Wieland T (1975) Interaction of actin with phalloidin: polymerization and stabilization of F-actin Biochim Biophys Acta 400, 407414 Criddle AH, Geeves MA & Jeffries T (1985) The use of actin labelled with N-(1-pyrenyl) iodoacetamide to study Distribution of actintropomyosintroponin... Values of LÂ, K2 and Y obtained from the ts were used to simulate p2, the fraction of actin in the active state We also tted theoretical values of p2 to the tropomyosin uorescence to obtain LÂ, K2 and Y From those values we were able to calculate curves of h as a function of free S1ADP Tropomyosin uorescence was normalized from 0 to 1 in the absence of Ca2+ because we assumed that essentially none of the... Chalovich JM & Eisenberg E (1982) Inhibition of actomyosin ATPase activity by troponin- tropomyosin without blocking the binding of myosin to actin J Biol Chem 257, 24322437 Gafurov B, Fredricksen S, Cai A, Brenner B, Chase PB & Chalovich JM (2004) The Delta14 mutation of human cardiac troponin T enhances ATPase activity and alters the cooperative binding of S1-ADP to regulated actin Biochemistry 43,... constants, KB is the equilibrium constant for proceeding from the blocked to the closed state, KT is the equilibrium constant for proceeding from the closed state to the open state, and n is a number of actin monomers forming a co-operative unit We used constraints similar to those described elsewhere [15,50] We determined the occupancy of the various states as a function of S1 bound by using differential... changes of reconstituted skeletal muscle thin laments observed by uorescence resonance energy transfer J Biol Chem 268, 71017106 Ishii Y & Lehrer SS (1990) Excimer uorescence of pyrenyliodoacetamide-labeled tropomyosin: a probe of the 2298 28 29 30 31 32 33 34 35 36 37 38 39 40 41 state of tropomyosin in reconstituted muscle thin laments Biochemistry 29, 11601166 Geeves MA & Lehrer SS (1994) Dynamics of . Equilibrium distribution of skeletal actin–tropomyosin– troponin states, determined by pyrene–tropomyosin fluorescence Boris Gafurov 1 and. the absence of added myosin subfragment 1 (S1) change as a function of the square root of the ionic strength. Distribution of actin–tropomyosin troponin states