Báo cáo khoa học: Equilibrium distribution of skeletal actin–tropomyosin– troponin states, determined by pyrene–tropomyosin fluorescence potx

13 284 0
  • Loading ...
    Loading ...
    Loading ...

Tài liệu hạn chế xem trước, để xem đầy đủ mời bạn chọn Tải xuống

Tài liệu liên quan

Thông tin tài liệu

Ngày đăng: 07/03/2014, 09:20

Equilibrium distribution of skeletal actin–tropomyosin–troponin states, determined by pyrene–tropomyosinfluorescenceBoris Gafurov1and Joseph M. Chalovich21 Uniformed Services University of the Health Sciences, Department of Pharmacology, Bethesda, MD, USA2 Department of Biochemistry and Molecular Biology, Brody School of Medicine at East Carolina University, Greenville, NC, USAThe ATPase activity of striated muscle myosin is lowunless it is bound to actin. Actin activation is inhibitedby the regulatory proteins tropomyosin, troponin T,troponin I and troponin C, which bind along actin fila-ments of skeletal and cardiac muscles. Activation ofstriated muscle contraction occurs when Ca2+binds totroponin C, or in a Ca2+-independent manner whenrigor-type myosin binds to actin [1–3]. Myosin is boththe target enzyme that hydrolyzes ATP and a potentialallosteric activator. Much current work is devoted tounderstanding the structural and functional changesthat occur in the large co-operative system consistingof myosin, actin, troponin and tropomyosin. Structuralchanges in troponin [4] and tropomyosin [5], inresponse to either Ca2+or myosin subfragment 1 (S1)binding, have been documented. The structure of actinis plastic [6] and could also change in response to theregulatory proteins.Keywordsparallel pathway model; pyreneiodoacetamide; regulation of contraction;tropomyosin; troponinCorrespondenceJoseph M. Chalovich, Department ofBiochemistry and Molecular Biology, BrodySchool of Medicine at East CarolinaUniversity, 5E-122 Brody Bldg, Greenville,NC 27834, USAFax: +1 252 7443383Tel: +1 252 7442973E-mail: chalovichj@ecu.eduWebsite: http://www.ecu.edu/biochemistry/Chalov.htm(Received 11 December 2006, revised 10February 2007, accepted 1 March 2007)doi:10.1111/j.1742-4658.2007.05765.xActin–tropomyosin–troponin has three structural states, but the functionalproperties of regulation can be explained with models having two func-tional states. As a step towards assigning functional properties to all thestructural states, we examined fluorescent probes that monitor changes introponin and tropomyosin. Tropomyosin labeled with pyrene–iodoacetamideis 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 otherthan the inactive state. The fraction of actin in an active state determinedfrom pyrene excimer fluoresecence agreed with that calculated from light-scattering measurements of myosin subfragment 1 (S1)–ADP to regulatedactin in both the presence and absence of Ca2+over a range of ionicstrength conditions. The only exceptions were conditions where the bindingof S1–ADP to actin was too strong to measure accurately. Pyrene–tropo-myosin excimer fluorescence was Ca2+dependent and so reflected thechange in population caused by both Ca2+binding and S1–ADP binding.Pyrene labeling of tropomyosin did not cause a large perturbation of thetransition among states of regulated actin. Using pyrene–tropomyosinfluorescence we were able to extend the ionic strength dependence of theparameters describing the co-operativity of binding of S1–ADP to actin aslow as 0.1 m. The probes on tropomyosin and troponin I had differentresponses to Ca2+and S1–ADP binding. These different sensitivities canbe explained by an intermediate between the inactive and active states ofregulated actin.AbbreviationsIANBD, 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 2287Tropomyosin occupies three positions on actin(Fig. 1), depending on the amount of Ca2+bound totroponin 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 modelsof regulation are constructed around the assumptionthat each structural state has a unique function. Othermodels use the minimum number of states required tosimulate function. An ongoing question is what are theproperties of these three states and how do they relateto regulation.Two types of regulatory models are shown in Fig. 2.In parallel pathway models (Fig. 2A,B), actin exists intwo or three states, with discrete abilities to serve ascofactors for myosin-catalyzed ATP hydrolysis. Therelative populations of these actin states are deter-mined by Ca2+and bound S1–ADP. The overall activ-ity of the system at any condition is defined by thefraction of time that myosin is bound to each of theseactin states. More detailed schemes of a parallel path-way model, showing some steps in ATP hydrolysis,have been published previously [14,15]. The formalismfor a parallel pathway model was originally defined fortwo functional states of actin, for simplicity [14,16].Despite early concerns that a two-state model couldnot explain the binding kinetics, it has been shown tosimulate equilibrium binding, binding kinetics andregulation of ATPase activity correctly [15]. Tropomy-osin is a switch, in the parallel model, that changes thestructure of actin in some way that alters its ability tostimulate myosin ATPase activity [17]. Because thetwo-state model is able to explain many features ofregulation, the properties of any intermediate state thatmay be present are undefined. The potential to definethe intermediate state does exist if it can be observedin real time.In sequential models of regulation, actin passes fromstate AB(blocked) to AC(closed) to AO(open). Insequential models, one cannot define the activity of anindividual state. Only state AOsupports myosin activ-ity, so it is necessary to go stepwise from the blockedto the closed to the open states. The model shown inFig. 2C is from McKillop & Geeves [18] and is basedon the multiple-step binding of myosin to actin.Another model, that of Butters & Tobacman, has threestates of actin with different orientations of tropomyo-sin that are in equilibrium with each other and with afourth state, in which actin has undergone a conforma-tional change to an active state with a structure similarto that stabilized by the binding of myosin [19]. Thatmodel is not shown here, but it may be imagined as afunnel in which three states of regulated actin funnelto an active state that supports contraction.The models in Fig. 2 share the idea of multipleforms of regulated actin with different activities inequilibrium with each other. Changes in the distribu-1234EGTA Ca2+rigor S1 boundFig. 1. Cross-sections of actin–tropomyosin–troponin showing thestructural states identified in the absence of Ca2+, with saturatingCa2+and with bound rigor-type myosin subfragment 1 (S1). Tropo-myosin is shown in black. The cross-section of an actin filament isshown in outline and the orientation of the four subdomains isshown. The dashed line is for reference. The figure is based onCraig & Lehman [51].AB ACAOMAC MAOMARK1K2K1KBKTCMAiMAaK1K2AAiAi(Ca) AaBαβAiAaFig. 2. Models of regulation of striated muscle contraction. Actin isrepresented by the letter A with a subscript to designate its state;myosin is represented by the letter M. The large differences ininteractions among different myosin-nucleotide states is notshown. Panels A and B show two-state and three-state parallelpathway models. In the two-state version, myosin binds to actinthat is either in the inactive (Ai) or active (Aa) state. The distributionbetween Aiand Aais determined by the fraction of troponin C(TnC) sites with bound Ca2+and the fraction of actin sites withbound rigor-type cross-bridges. Rapid ATP hydrolysis occurs whenactin is in state Aa. The three-state model shown in (B) considersthe possibility that regulated actin that has bound Ca2+, but norigor-type cross-bridges, has an intermediate level of activity. Forsimplicity, the binding to myosin is not shown for this case. In thismodel, state Aais active and state Aiis inactive, but the propertiesof state Ai(Ca), are uncertain. Panel C shows a sequential model inwhich there are three states of actin namely blocked (AB), closed(AC) and open (AO). Actin makes sequential transitions to the openstate, AO, which is competent for accelerating ATP hydrolysis andproceeding into the force-producing state MAR.Distribution of actin–tropomyosin–troponin states B. S. Gafurov and J. M. Chalovich2288 FEBS Journal 274 (2007) 2287–2299 ª 2007 The Authors Journal compilation ª 2007 FEBStion of regulated actin states determine the activity ofactin–myosin, regardless of whether that activity chan-ges as a normal regulatory response [14,20], or as aresult of some mutations in troponin [21,22] or inexperimentally produced mutations in tropomyosin[23]. Therefore, it is important to have reliable meth-ods of determining the state of actin in real time. Thismanuscript explores, in detail, a well-known method ofmonitoring 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 ontroponin and tropomyosin. Probes on troponin Irespond to both Ca2+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 toactin [11,12,24,25]. Resonance energy transfer measure-ments between probes on actin and troponin I [26] ortroponin T [13] have also proven to be valuable formeasuring the state of the actin filament. Changes inpyrene–tropomyosin fluorescence have been shown tobe a measure of the fraction of actin in the active state[27]. Pyrene–tropomyosin excimer fluorescence wassensitive to activation by S1, but Ca2+had little effect[27,28]. Pyrene–tropomyosin excimer fluorescence didgive the predicted change in regulated actin distributionas the amount of S1–ADP was altered, but its usefulnesswas only demonstrated at relatively high ionic strength.The response of pyrene–tropomyosin fluorescence to S1binding led to the idea that this probe measures entryinto the most active state of actin, but is insensitive totransitions to states of intermediate activity.We report here a comparison of pyrene–tropomyo-sin excimer fluorescence to predicted changes in theactin state that occur in response to Ca2+andS1–ADP binding under conditions ranging from 100to 240 mm ionic strength. We also compare changes inpyrene excimer fluorescence with N-(((2-iodoacetoxy)-ethyl)-N-methyl)-amino-7-nitro benz-2-oxa-1,3-diazole(IANBD)-labeled troponin fluorescence when bothprobes are present on the same actin filament. Theresults can be readily explained by the presence of anintermediate between the inactive and fully activestates of regulated actin. Pyrene excimer formation didnot appreciably affect the distribution of actin states.Furthermore, pyrene excimer fluorescence gave reason-able estimates of the distribution of actin states ationic strengths as low as 0.1 m, where it may be possi-ble to correlate these changes with ATPase activities.ResultsRegulated actin is predominantly in the inactive statein the absence of Ca2+and bound S1. Both Ca2+andS1–ADP bind more tightly to the active state of actinthan to the inactive state, and stabilize the active state.Increasing concentrations of free S1–ADP results in aco-operative binding curve, indicating a transitionfrom a lower affinity to a higher affinity state of actin–tropomyosin–troponin. This change in affinity is read-ily seen in the absence of Ca2+as sigmoidal increasesin theta with increasing free S1–ADP concentrations(Fig. 3A–D). Changes in pyrene–actin fluorescence areoften used to measure the binding of S1 to actin(Fig. 3; solid squares). In order to compare changes inpyrene–tropomyosin excimer fluorescence with changesin occupancy of actin with S1, we utilized light scatter-ing to measure binding (open circles). Light scatteringmeasurements gave binding patterns that were similarto previous measurements using pyrene–actin fluores-cence (compare circles with solid squares). Theoreticalcurves, describing the relationship between theta andfree S1–ADP, were produced by fitting the Hill modelto the data at the four ionic strength conditions shownin Fig. 3. This fitting procedure produced values ofK1,K2,L¢ and Y. Those parameters were used to pro-duce theoretical curves for p2, the fraction of actin inthe active state shown by solid curves in Fig. 3.Figure 3A–D also shows that changes in pyrene–tropomyosin excimer fluorescence (triangles) followedthe predicted changes in the fraction of actin in theactive state. The agreement between the theoreticalcurves and the measurements was particularly good athigher ionic strengths where the measurements weremost accurate. Deviations between the predicted valuesof 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 freeS1–ADP, the solid curve predicted from equilibriumbinding data (circles) predicts an excess of 20% of theactin to be present in the active state. Values of p2near zero would be consistent with known activities.That is, the p2 values determined from tropomyosinfluorescence are probably more reliable than thosecalculated from binding studies at low ionic strength.Values of equilibrium binding parameters, deter-mined in the absence of Ca2+as a function of ionicstrength, are shown in Fig. 4A–C. The open symbolsshow the present results of binding of S1–ADP toactin containing troponin and pyrene-labeled tropomy-osin. Equilibrium binding parameters were calculatedby fitting the Hill formalism to light scattering alone(circles), or to pyrene–tropomyosin fluorescence alone(triangles). The values of K2shown in Fig. 4A wereindependent of the type of fitting used and they agreedvery well with earlier values determined from pyrene–actin fluorescence shown as solid squares. The modelB. S. Gafurov and J. M. Chalovich Distribution of actin–tropomyosin–troponin statesFEBS Journal 274 (2007) 2287–2299 ª 2007 The Authors Journal compilation ª 2007 FEBS 2289is not particularly sensitive to values of K1, so thesevalues are not shown.Figure 4B,C shows the parameters Y and L¢.Ydecreased with increasing ionic strength, indicating adecreased tendency of adjacent regulatory units to existin the same functional state. Values of Y, calculatedfrom light scattering, were similar to those calculatedfrom pyrene–tropomyosin fluorescence. However, val-ues of Y tended to be slightly lower for the pyrene–tropomyosin system than for the pyrene actin systemexamined earlier, shown as solid squares. It is unclearif this difference is a result of the different probesused.Values of L¢ tended to increase with increasing ionicstrength. Therefore, high ionic strength stabilized theFig. 3. Changes in light scattering (circles) and pyrene–tropomyosinfluorescence (triangles) as a function of free myosin subfragment 1(S1)–ADP concentration in the absence (A–D) and presence (E–H) ofCa2+. 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 witha dashed line are fits of the Hill model to the fraction of actin withbound S1, determined by light scattering. Curve fitting was per-formed simultaneously with paired data sets, in the presence andabsence of Ca2+, to constrain the variables. Fractions of actin in theactive 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 arefrom a previous study with pyrene actin [15] to show that similarvalues of theta are obtained by light scattering measurements andearlier pyrene-actin measurements. All measurements were madeusing skeletal troponin and tropomyosin under the following condi-tions: 0.3 lM phalloidin actin, 0.06 lM pyrene-labeled tropomyosin,0.06 lM troponin, 25 °C, in a buffer containing 20 mM Mops, pH 7.0,5mM MgCl2,1mM dithiothreitol, 2 mM ADP, 0.2 mgÆmL)1bovineserum albumin, sufficient KCl to reach the target ionic strength andeither 1 mM EGTA (A–D) or 0.1 mM CaCl2(E–H).Fig. 4. Effect of ionic strength on equilibrium binding parameters inthe absence (A–C) and presence (D–E) of Ca2+. Values of K2(A, D),Y (B, E) and L¢ (C, F), determined by light scattering (circles) andpyrene-excimer fluorescence (triangles), are compared with earliervalues determined from pyrene–actin fluorescence (solid squares).Values obtained from light scattering were obtained by a global fitof the model to data obtained at zero and saturating Ca2+. Earliervalues from pyrene–actin fluorescence were the result of a globalfit of data from six different free Ca2+concentrations but the sameionic strength [15]. The conditions were the same as for Fig. 3,with 1 mM EGTA used in the experiments with results shown inpanels A–C and 0.1 mM CaCl2used in the experiments with resultsshown in panels D–F.Distribution of actin–tropomyosin–troponin states B. S. Gafurov and J. M. Chalovich2290 FEBS Journal 274 (2007) 2287–2299 ª 2007 The Authors Journal compilation ª 2007 FEBSinactive state of regulated actin relative to the activestate 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 earlierpyrene–actin fluorescence measurements.To determine the relationship of K2, Y and L¢ toionic strength in Ca2+, we first determined the effectof Ca2+on fluorescence so that the initial point of p2could be defined. Figure 5 shows pyrene–tropomyosinfluorescence measurements of regulated actin as afunction of Ca2+concentration at 180 mm ionicstrength. In 0.1 mm EGTA, the free Ca2+was belowthat required for activation (open circles). The pyrenefluorescence intensity increased to a maximum valuewhen Ca2+exceeded the EGTA concentration. A con-trol experiment was performed in the absence ofEGTA (solid circles). As expected, there was nochange in fluorescence with the addition of Ca2+because the initial Ca2+concentration was alreadyhigh enough to give the full effect.We performed another control by comparing theeffects of Ca2+on probes on both tropomyosin andtroponin. Actin filaments were reconstituted withpyrene-labeled tropomyosin and troponin containingIANBD-labeled troponin I. Figure 6A shows that theaddition of excess Ca2+to an EGTA-containing solu-tion caused 40% of the maximum possible change inpyrene–tropomyosin fluorescence, but, on average,92% of the maximum in IANBD–troponin I fluores-cence. The complete change of pyrene–tropomyosinrequired the addition of nucleotide-free S1. Figure 6Bcompares the effect of both probes to the addition ofS1 in the absence of Ca2+. Although the changes arein opposite directions, the sensitivities to S1 concentra-tion were similar.Knowing the value of p2 to be 0.4, in the absence ofS1–ADP we were able to examine the relationshipbetween predicted values of p2 and pyrene excimerfluorescence in the presence of Ca2+. Figure 3E–Hshows light scattering and pyrene excimer fluorescenceat four ionic strength conditions at saturating Ca2+.Values of p2 reached their maximum values at subsat-urating concentrations of S1–ADP in all cases. TheFig. 5. The fluorescence of actin filaments reconstituted withpyrene-labeled tropomyosin is Ca2+dependent at 180 mM ionicstrength. Pyrene–tropomyosin fluorescence was measured in thepresence (open circles) or absence (closed circles) of 0.1 mMEGTA. The curve obtained in the presence of EGTA shows theincrease in fluorescence as the total Ca2+concentration wasincreased. 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 withCa2+and myosin subfragment 1 (S1). Both fluorescent probeswere present in the actin filament at the same time and the fluor-escence changes of each probe were measured about 10 minafter each addition of S1. (A) Effect of adding 1.2 mM Ca2+to theEGTA-containing solution and then subsequently adding S1. Theresponse to Ca2+was more extreme for IANBD–troponin I thanfor pyrene-labeled tropomyosin. Multiple lines are from emissionmeasurements made at 10 nm wavelength increments. (B) Titra-tion of regulated actin containing both probes with S1 in theabsence of Ca2+. The conditions were the same as for Fig. 3, with150 mM KCl.B. S. Gafurov and J. M. Chalovich Distribution of actin–tropomyosin–troponin statesFEBS Journal 274 (2007) 2287–2299 ª 2007 The Authors Journal compilation ª 2007 FEBS 2291tropomyosin transition measured by pyrene fluores-cence was not co-operative in the presence of Ca2+.The dashed lines are fits of the Hill model to the val-ues of light scattering data, and the predicted curvesfor p2 are shown as solid lines. The measured valuesof p2 were similar to the predicted values. Poor fits tolight scattering data, as in Fig. 3H, were, in part, aresult of the fact that these were not best fits to a sin-gle data set, but were global fits to data in the presenceand absence of Ca2+.The ionic strength dependencies of K2, Y and L¢,determined by fitting the Hill model to the data ofFig. 3E–H, are shown in Fig. 4D–F. The agreement ofvalues of K2, Y and L¢ was good between light scatter-ing (circles) and pyrene–tropomyosin fluorescence (tri-angles) measured on the same proteins. Values of K2were similar to those measured in the absence ofCa2+. Values of Y were near 1 at low ionic strengthand decreased slightly as the ionic strength was raised.If Y was constrained to be greater than 1, the value ofY would be 1 over the ionic strength range (data notshown). Values of L¢, determined by both methods,increased with increasing ionic strength as they did inthe absence of Ca2+.Values of Y and L¢ were substantially different foractin filaments containing pyrene-labeled tropomyosincompared with those with pyrene on the actin. Fittingwas generally more difficult in the presence of Ca2+because of the lack of features in those curves. Estima-tions of L¢ and Y are problematic because changes inthe value of Y can be compensated, to some extent,for changes in L¢.The parameter, p2, defines the activity of the actinfilament in parallel pathway models. Under conditionswhere all of the S1-ATP is bound to actin, the ATPaseactivity is approximately equal to p2*rmax+(1 ) p2)*rmin, where rmaxand rminare the rates for theactive 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 rmaxand rmincan be determined from thekcatfor actin in the active and inactive states, respect-ively. Although these ATPase parameters have notbeen determined at the conditions used for the bindingexperiments, relative changes in ATPase activitycan be approximated by observation of changes in p2.Figure 7 shows how p2 would change if actin filamentswere activated by the attachment of an activating formof S1, such as S1–ADP. The inset shows values of p2as a function of the square root of the ionic strength.The difference between the EGTA and Ca2+rates areexpected to be approximately constant over the rangeof ionic strengths examined.DiscussionTransitions between the inactive and active states ofregulated actin are important determinants of the regu-lation of striated muscle contraction. The distributionof these states determines the ATPase activity, whereasthe rates of transitions among the states may affect therate of force redevelopment [11]. Some disease-causingmutations in troponin T change in the distributionbetween the states of regulated actin [21,22]. The abil-ity to measure state transitions rapidly and relate themto function will be helpful in studying such defects.Fluorescent probes on troponin and tropomyosin havethe potential to measure the distribution of states inreal time.Ishii & Lehrer reported that probes on tropomyosinreflect changes in the fraction of actin in the activestate resulting from S1 binding [27]. Acrylodan-labeledtropomyosin was useful for actin–tropomyosin, but thesignal was too small in the presence of troponin [29].Pyrene-labeled tropomyosin was the prefered probe foractin–tropomyosin and actin–tropomyosin–troponin[27,28]. Pyrene–iodoacetamide labeling was preferredover pyrene–maleimide labeling because of the rapidresponse of its excimer fluorescence to S1 binding [27].The S1-induced increase in excimer fluorescence iscaused by an increase in the fraction of pyrene mole-cules forming excimers. Pyrene–iodoacetamide-labeledtropomyosin excimer fluorescence exhibited a smallchange with Ca2+at low ionic strength. Because ofthese considerations, we have examined more closelyFig. 7. Calculated probabilities of actin–tropomyosin–troponin in theactive state (p2) in the presence (solid lines and solid circles) andabsence (dashed line and open circles) of Ca2+. Simulations weremade from equilibrium binding parameters determined at 120 mMionic strength. The inset shows how values of p2 in the absence ofadded myosin subfragment 1 (S1) change as a function of thesquare root of the ionic strength.Distribution of actin–tropomyosin–troponin states B. S. Gafurov and J. M. Chalovich2292 FEBS Journal 274 (2007) 2287–2299 ª 2007 The Authors Journal compilation ª 2007 FEBSthe suitability of pyrene–tropomyosin excimer fluores-cence as a measure of regulated actin state changes.We studied tropomyosin excimer fluorescence over arange of ionic strength conditions because ATPasemeasurements and S1–ADP binding cannot be readilymeasured under the same conditions and an extrapola-tion of parameters is necessary. Furthermore, examin-ing the behavior at different conditions increases thereliability of parameters obtained by curve fitting[15,21]. Values of the fraction of actin in the activestate, p2, were calculated from S1–ADP binding (lightscattering). Values of p2 were also directly measuredby pyrene excimer fluorescence. Pyrene excimer fluores-cence generally agreed with the predicted values of p2.Deviations occurred when S1–ADP binding becametoo 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 ofactin 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 withpyrene-labeled tropomyosin in the absence of Ca2+are in reasonable agreement with earlier values wherethere was no excimer formation (Fig. 4). Pyrene probeson tropomyosin did not significantly alter the values ofK2, L¢ or Y at any ionic strength examined. Further-more, when troponin containing an IANBD probe ontroponin I was reconstituted with N-(1-pyrene)iodo-acetamide (pyrene–iodoacetamide)-labeled tropomyosinand actin, the IANBD probe retained its typicalresponses to changes in Ca2+and S1 binding (Fig. 6).Fitting binding curves to obtain binding parametersis difficult in the case of Ca2+because the curves arefeatureless hyperbolas. Although we observed onlysmall differences in binding curves measured withpyrene–actin and pyrene–tropomyosin in Ca2+(Fig.3G–H), there was poor agreement between the values ofL¢ calculated in the two cases. We also noted that atlow ionic strength the values of Y tended to be greaterin the presence of Ca2+, but this was not observed inthe present case with unlabeled actin. It is also worthpointing out that the parameters determined in ourearlier study with pyrene–actin resulted from global fitsof the data. This change in fitting may contribute todifferences in the final values of the parameters.The parameters K2, L¢ and Y varied with ionicstrength, in agreement with our earlier observations[15,21]. High ionic strength decreased the fraction ofregulated actin in active states (increased L¢). Thisresult is consistent with in vitro motility assays wherehigher Ca2+is required for full activation at high ionicstrength [30]. This trend has now been observed from0.1 to 0.24 m ionic strength. The extension of thisresult to the lower ionic strength range is useful forextrapolation of the values for future simulation ofATPase activities under conditions where they can bereadily measured.Tropomyosin excimer fluorescence was Ca2+dependent, but it did not directly track Ca2+binding.Rather, the change was consistent with a state change,such as partial transition, to the most active state ofactin or a total transition to an intermediate state.Ca2+binding resulted in % 40% of the maximumobserved change of excimer fluorescence obtained withfull activation by rigor-type myosin binding. Thisagrees with the observation of Williams et al., thatCa2+alone provides 40% of the maximum value ofkcat[31].In vitro motility assays support the view that Ca2+alone does not provide full activation of regulatedactin. High levels of loading of filaments with myosinproduced about a doubling of the rate at saturatingCa2+[32] and a velocity 1.8 times higher than that ofunregulated actin [33]. Activities that exceed actinalone are probably the result of partial stabilization ofthe most active state of regulated actin. In the case ofcardiac troponin–tropomyosin, this extra activationwas only evident for some disease-causing mutationsof troponin [34]. Under those conditions, the velocitywas increased 1.6-fold over full activation of the wild-type cardiac troponin. Some mutations have the effectof partially stabilizing the fully active state [21], so this1.6-fold increase is probably an underestimate of themaximum level of activation. These results suggest thatin the motility assay, Ca2+alone produces 50–55% ofthe maximum activation. The results could be closer tothe 40% activation seen in solution for Ca2+alone ifthe actin filaments in the in vitro studies were not max-imally activated.The ability of S1–ADP and rigor S1 to activateactin filaments raises the question of how an activemuscle can relax once the free Ca2+concentration isdecreased. A larger fraction of strongly bound cross-bridges is required for activation in EGTA than inCa2+. However, in EGTA at 0.18 m ionic strength,30% saturation of the actin produces thin filamentsthat are 50% active (Fig. 3C). A 90% relaxationwould require less than 5% of the actin to containstrongly bound cross-bridges. However, muscle maynot behave identically to the proteins in solution.Geometrical considerations, and the presence ofother protein components or small molecules, couldresult in a considerable shift of the curves shown inFig. 3.B. S. Gafurov and J. M. Chalovich Distribution of actin–tropomyosin–troponin statesFEBS Journal 274 (2007) 2287–2299 ª 2007 The Authors Journal compilation ª 2007 FEBS 2293Probes on troponin I report changes in the state ofregulated actin caused by S1 binding to actin [12] andalso respond directly to changes in Ca2+[12,24,25].Different sensitivities of fluorescent probes to Ca2+have been used in the past to argue for the presence ofan intermediate state of regulated actin. Because theprobes can affect the behavior of the regulatory com-plex, it is difficult to compare directly the results ofprobes on separate regulatory complexes. We havenow utilized IANBD on troponin I and pyrene ontropomyosin within the same regulatory complex. Bothprobes responded to S1 binding in a similar way(Fig. 6A), but exhibited different responses to Ca2+(Fig. 6B). This result is consistent with the existence ofan intermediate structural state [7].We used the two-state parallel pathway model ofHill et al. for predicting the fraction of actin in the act-ive state. That model is consistent with the measuredeffects of Ca2+on binding in the presence of ATP[35,36], equilibrium binding in the presence of ADP[16], binding kinetics [15,37] and the general featuresof ATPase activities [14]. Furthermore, in our view,the functionally indistinguishable state is not the firststate of a series, but rather the state corresponding tobound Ca2+and no bound rigor S1 (Ai(Ca)in Fig. 2B).That intermediate may resemble the inactive (Ai(EGTA))or fully active (Aa) states in terms of key functionalproperties.Although our results can be explained with twofunctional states, there is evidence for three structuralstates of regulated actin. Pirani et al. estimated the dis-tributions of structural states by image reconstructionof electron micrographs following dilution of the pro-teins to low ionic strength [8]. They predicted 22% ofthe actin to be in the closed state in the absence ofCa2+[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 Ca2+to be20% blocked, 68% closed and 12% M state (activestate). The 40% activation, predicted in the presentstudy, from tropomyosin fluorescence does not agreewith this distribution. This could be an indication thatthere 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 byMcKillop & Geeves [18], as shown in Fig. 2C. Theincreased rate of binding of S1–ADP to regulated actinin Ca2+compared with EGTA was interpreted, in thatmodel, as 75% of actin sites being blocked in theabsence of Ca2+. We have an alternative explanationfor that effect [37]. However, for the present exercisewe forced the fit to populate the blocked state inEGTA in accordance with their model. We usedmost of the constraints set by McKillop & Geeves,namely, 0 < KB < 10, 0 < KT < 20, 0 < N < 7,103<K1<106and K2 ¼ 200. We did not constrainthe 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 populationsof both the blocked and closed states decreased withincreasing amounts of bound S1 in both the absenceand presence of Ca2+. The population of the openstate was much higher in Ca2+than in EGTA in theabsence of bound S1. Regulated actin was almostexclusively in the open state at saturating S1, irrespect-ive of the Ca2+concentration. Whereas the populationof the open state does not correlate directly with ourpredicted p2, they do follow the same trend.Tropomyosin–pyrene excimer fluorescence gives agood estimate of the fraction of actin in the activestate over a range of conditions. Simultaneous mea-surements of probes on tropomyosin and troponingive evidence for an intermediate state. By takingfurther advantage of this system, it may be possible todetermine the role of this intermediate in regulation.Fig. 8. Distribution of the blocked (circles),closed (triangles) and open (squares) statesin the course of myosin subfragment 1 (S1)binding. (A) The predicted occupancy of thestates at 0.18M ionic strength in the pres-ence of 0.1 mM Ca2+. The diamonds are thep2 parameter that represents the transitionof the actin filament into the active state inHill’s model. (B) The same parameters inthe Ca2+-free case.Distribution of actin–tropomyosin–troponin states B. S. Gafurov and J. M. Chalovich2294 FEBS Journal 274 (2007) 2287–2299 ª 2007 The Authors Journal compilation ª 2007 FEBSThis is particularly important for the study of disor-ders of the regulatory system.Experimental proceduresProtein preparationActin [38,39], myosin [40], troponin and tropomyosin [41]were isolated from rabbit back muscle. Myosin S1 wasmade by digestion of myosin with chymotrypsin [42]. Pro-tein concentrations were determined by light absorbance at280 nm, corrected for scattering, at 340 nm, using the fol-lowing extinction coefficients (e0.1%) for 280 nm: actin,)1.15; myosin-S1, )0.75; tropomyosin, )0.33; and troponin,)0.37. The molecular masses assumed for the key proteinswere: 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 MgCl2and 40 lmphalloidin. Actin was sometimes labeled with N-(1-pyrenyl)iodoacetamide [43]. Tropomyosin was labeled withN-(1-pyrene)iodoacetamide (pyrene–iodoacetamide) [27]. Insome cases, troponin I was labeled with IANBD [12]. Theextents of labeling were % 60% and 35% for tropomyosinand troponin, respectively. Reconstituted actin was pre-pared by mixing actin with troponin and pyrene-labeledtropomyosin in a 3 : 1 : 1 molar ratio to ensure saturationof actin at the low concentrations used for binding studies.Equilibrium fluorescence measurementsEquilibrium fluorescence measurements were made on anAminco Bowman II Luminescence Spectrometer (ThermoElectron 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 fluorescenceand 492 and 536 nm, respectively, for IANBD–troponinfluorescence.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–actinfluorescence [43,45,46]. Details of the binding measurementsare described elsewhere [15] and are similar to those des-cribed by others [46,47]. Our binding solutions contained20 mm Mops, pH 7.0, 5 mm MgCl2,1mm dithiothreitol,2mm ADP, 0.2 mgÆmL)1bovine serum albumin, sufficientKCl to reach the target ionic strength and 0.1 mm CaCl2or1mm EGTA. The actin concentration in binding experi-ments was 0.3 lm. Solutions also contained 14 unitsÆmL ofhexokinase and 1 mm glucose to scavenge ATP and 20 lmAp5A to inhibit ATP formation through the myokinase reac-tion. Titrations were carried out by the stepwise addition ofS1 to a 2 mL volume of pyrene-labeled actin–tropomyosin–troponin at 5 min intervals. This time interval was importantto ensure equilibrium at each step. Fluorescence intensitiesand protein concentrations were corrected for the volumechange (< 10%) caused by adding S1. Rabbit skeletal tropo-nin and tropomyosin were used in this study for comparisonwith our existing data for those regulatory proteins.Values of theta (S1 bound to the actin total ratio) andthe free S1 concentration from fluorescence or light scatter-ing measurements were calculated using the equations:h ¼FiÀ FminFmaxÀ Fmin½FreeS1¼½S1totalÀ h ýActintotalð1ÞWhere Fiis the fluorescence or scattering intensity at a totalS1 concentration of i (lm); and Fmaxand Fminare the maxi-mum and minimum values of intensity, respectively.Modeling experimental resultsLight scattering was used to measure the binding of S1–ADP to actin and tropomyosin. Pyrene excimer fluorescencewas used to monitor the fraction of actin in the active state.Equilibrium-binding parameters were extracted from lightscattering data by using the co-operative binding model ofHill et al. [16] or by the model of McKillop & Geeves [18].Fitting to the parallel pathway model of Hill was describedin detail earlier [15]. Briefly, the relationship between thefraction of actin with bound S1 and the free S1 concentra-tion can be described by the following equations [16]:h ¼ p1h1þ p2h2hi¼KiC1 þ KiCp1¼ 1 À p2p2¼2aYffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffið1 À aÞ2þ 4aYq1 À a þffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffið1 À aÞ2þ 4aYqa ¼ð1 þ K2CÞnð1 þ K1CÞnY22ðqÞLY11ðqÞL0¼LY11ðqÞY22ðqÞY ¼Y11ðqÞY22ðqÞY12ðqÞY21ðqÞYijðqÞ¼xijþ 2kaqyijþ k2bq2zij9>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>=>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>;ð2Þp1and p2are fractions of actin units in the inactive andactive states, respectively; hIand h2are fractions of actincontaining bound S1 in the inactive and active states,respectively; K1and K2are S1-binding constants to theB. S. Gafurov and J. M. Chalovich Distribution of actin–tropomyosin–troponin statesFEBS Journal 274 (2007) 2287–2299 ª 2007 The Authors Journal compilation ª 2007 FEBS 2295inactive and active states of actin, respectively; C is the freeS1 concentration; q is the free Ca2+concentration; n isthe number of actin monomers in one actin–tropomyosin–troponin unit (assumed to be seven); L is the equilibriumconstant for transition of an isolated actin–tropomyosin–troponin unit with no neighbors, no bound Ca2+and nobound S1 from state 2 to state 1; L¢ is the equilibrium con-stant defining the transition from the active state to theinactive state for the entire actin filament, but without S1;Y describes the co-operativity between adjacent regulatoryunits of seven actin monomers; Y is the overall co-operativ-ity parameter; Yijare individual co-operative interactionsbetween units in states i and j (we assumed that Yij¼ Yji);xij, yijand zijrepresent the free energies of nearest neighbortropomyosin interactions (Wij) in exponential form e–Wij ⁄ kT[16]; and kaand kbare affinities of troponin in states 1 and2 for Ca2+with values of % 105and 106Æm)1, respectively[48]. We assumed that the values of ka,bdid not changeover the ionic strength range in this study. The simulatedcurves were not very sensitive to the value of K1, so simula-tions were normally carried out with the assumption thatK1¼ K2⁄ 8 [49].All measurements were carried out in both Ca2+-freeand in Ca2+-saturated conditions. Binding data obtained athigh and low Ca2+, but at the same ionic strength, wereanalyzed using a global fit procedure [15]. The global fithelped to constrain the parameters. Values of L¢, K2and Yobtained from the fits were used to simulate p2, the fractionof actin in the active state. We also fitted theoretical valuesof p2to the tropomyosin fluorescence to obtain L¢, K2andY. From those values we were able to calculate curves of has a function of free S1–ADP.Tropomyosin fluorescence was normalized from 0 to 1 inthe absence of Ca2+because we assumed that essentiallynone of the actin was in the active state in the absence ofCa2+and bound S1. This assumption is reasonable based onATPase activity measurements. The flux is proportional tothe amount of S1 bound to each state multiplied by the kcatassociated with that state. Ca2+increases the kcatby% 22-fold [20], whereas the binding of NEM-S1 increases thekcatby a further 2.5-fold [31]. This means that the fraction inthe active state in EGTA is 1.8%. Binding studies were car-ried out at higher ionic strength conditions than the ATPasemeasurements. Because the fraction of actin in the activestate decreases with increasing ionic strength [15], the valueof 1.8% is an upper limit. The ATPase rates also predict thatin the presence of Ca2+alone, 40% of the regulated actin isin the active state. Again, this fraction is also likely to be anupper limit because of ionic strength considerations.In order to define the fraction of actin in the active statein the presence of Ca2+, but in the absence of bound S1,we observed the changes in fluorescence that occurred dur-ing Ca2+titrations. With measured values of the initialvalue in EGTA, the change that occurred with the additionof Ca2+and the further change that occurred with satur-ating S1–ADP, we were able to calculate the initial p2 inCa2+. The fluorescence data in Ca2+were normalized fromthis initial value to 1.0 for the maximum fluorescenceobserved in the presence of both Ca2+and saturatingS1–ADP. Although the initial raw fluorescence values werehigher in Ca2+than in EGTA, the values at saturating S1were about the same in both cases.Fitting parameters and constraints were similar to theones 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 fitswith correlation coefficients R2> 0.85.Analysis using the model of McKillop & GeevesBecause the original two-state parallel pathway model ofHill was able to account for the present data, the model wasnot expanded to include a third state. We did, however, ana-lyze some of these data with the three-state sequential modelof McKillop & Geeves [18], shown in Fig. 2C. We fitted themodel expressed in Eqn (3) to our binding isotherms andobtained key binding parameters K1, K2, KB, KTand n foreach ionic strength and Ca2+concentration used:h ¼K1cðKTð1 þ K2ÞPnÀ1ÞþQnÀ1KTPnþ Qnþ 1=KBð3ÞP ¼ 1 þ K1cð1 þ K2ÞQ ¼ 1 þ K1cwhere K1and K2are S1-binding constants, KBis the equi-librium constant for proceeding from the blocked to theclosed state, KTis the equilibrium constant for proceedingfrom the closed state to the open state, and n is a numberof actin monomers forming a co-operative unit. We usedconstraints similar to those described elsewhere [15,50].We determined the occupancy of the various states as afunction of S1 bound by using differential equations to des-cribe the probability for each state [37]. Curve fitting wascarried out to our binding isotherms at 180 mm ionicstrength, measured with or without Ca2+. The 3 · 3 schemeof the kinetic reactions, which take place when n ¼ 1, isshown below, as derived previously [37]:124356aB a-Bck1 k-1aTa-Tck1 k-1aTa-Tk2k-2Scheme 1.Distribution of actin–tropomyosin–troponin states B. S. Gafurov and J. M. Chalovich2296 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 Gafurov1and. the absence of added myosin subfragment 1 (S1) change as a function of thesquare root of the ionic strength. Distribution of actin–tropomyosin troponin states
- Xem thêm -

Xem thêm: Báo cáo khoa học: Equilibrium distribution of skeletal actin–tropomyosin– troponin states, determined by pyrene–tropomyosin fluorescence potx, Báo cáo khoa học: Equilibrium distribution of skeletal actin–tropomyosin– troponin states, determined by pyrene–tropomyosin fluorescence potx, Báo cáo khoa học: Equilibrium distribution of skeletal actin–tropomyosin– troponin states, determined by pyrene–tropomyosin fluorescence potx