Structural dynamics of troponin during activation ofskeletal muscleLuca Fusi1, Elisabetta Brunello1,2, Ivanka R. Sevrieva, Yin-Biao Sun, and Malcolm Irving3

Randall Division of Cell and Molecular Biophysics, King’s College London, London SE1 1UL, United Kingdom

Edited by Yale E. Goldman, University of Pennsylvania/Pennsylvania Muscle Institute, Philadelphia, PA, and accepted by the Editorial Board February 7, 2014(received for review November 22, 2013)

Time-resolved changes in the conformation of troponin in the thinfilaments of skeletal muscle were followed during activation insitu by photolysis of caged calcium using bifunctional fluorescentprobes in the regulatory and the coiled-coil (IT arm) domains oftroponin. Three sequential steps in the activation mechanism wereidentified. The fastest step (1,100 s−1) matches the rate of Ca2+

binding to the regulatory domain but also dominates the motionof the IT arm. The second step (120 s−1) coincides with the azi-muthal motion of tropomyosin around the thin filament. The thirdstep (15 s−1) was shown by three independent approaches to trackmyosin head binding to the thin filament, but is absent in theregulatory head. The results lead to a four-state structural kineticmodel that describes the molecular mechanism of muscle activa-tion in the thin filament–myosin head complex under physiologicalconditions.

muscle regulation | excitation–contraction coupling | muscle signaling

Contraction of skeletal and cardiac muscle is initiated bya transient increase in the concentration of intracellular Ca2+

ions, which bind to troponin in the thin filaments of the musclesarcomere. This leads to azimuthal movement of tropomyosinaround the thin filament, which uncovers the myosin bindingsites on actin and allows the head domain of myosin from thethick filaments to bind to actin and generate force (1, 2). In vitrostudies using isolated protein components showed that myosinhead binding can produce a further motion of tropomyosin, atleast in low [ATP] or rigor-like conditions (2–4), but the func-tional significance of this effect in physiological conditions andintact sarcomeres is not clear.To elucidate the molecular structural basis of muscle regula-

tion and the role of myosin binding in situ, we introduced bi-functional fluorescent probes into the calcium-binding subunit oftroponin, troponin C (TnC) (Fig. 1, yellow), in demembranatedfibers from skeletal muscle (5–7). One probe cross-linked a pairof cysteines introduced into the C helix of TnC (Fig. 1, green),close to the regulatory Ca2+ binding sites (Fig. 1, black spheres)in its N-terminal lobe, and reports the rotation and opening ofthis lobe on binding Ca2+ (5). The N-lobe opening is associatedwith binding of the switch peptide of troponin I (TnI) (Fig. 1,blue) to a hydrophobic pocket on its surface, and this is a keystep in the signaling pathway of calcium regulation (8, 9).A second probe was attached to the E helix of TnC (Fig. 1,

magenta) in its C-terminal lobe, which contains a pair of divalentcation binding sites (Fig. 1, gray spheres) that can bind Mg2+ aswell as Ca2+. The C lobe of TnC is clasped between two longhelices of TnI, one of which forms a coiled coil with part of thetropomyosin-binding component of troponin, troponin T (TnT)(Fig. 1, orange). The C lobe of TnC and these long TnI and TnThelices form a well-defined structural domain called the “ITarm” (9, 10). Although the C-lobe E helix of TnC is continuouswith the N-lobe D helix in the Ca2+-bound crystal structureshown in Fig. 1, the D/E helix is broken in situ, as it is in thecrystal structures of the Ca2+-bound cardiac isoform and the apostate of the skeletal isoform (9, 10). Thus, the C- and E-helixprobes give independent information about the orientations ofthe TnC N lobe and the IT arm, respectively, in a muscle fiber.

We separated the structural effects of Ca2+ and myosinbinding during activation of demembranated muscle fibers inphysiological conditions kinetically, using rapid jumps in in-tracellular [Ca2+] produced by photolysis of nitrophenyl-EGTA(NP-EGTA or caged Ca). Binding of Ca2+ to the regulatory sitesof troponin is at least 10–20 times faster than myosin binding inthe conditions used here, so we were able to resolve the kineticsof intermediate structural changes in the troponin signalingpathway, and relate them to those of calcium binding to tropo-nin, myosin binding to actin, and force generation. We usedthree additional protocols to assess the role of myosin binding inmuscle regulation in physiological conditions: we (i) imposedrapid ramp shortening on active muscle fibers to drive myosindetachment, (ii) abolished active force generation with a smallmolecule inhibitor, and (iii) stretched the muscle fibers to removethe overlap between the thick and thin filaments. The results leadto a four-state model that describes the sequence of structuralchanges in troponin and the thin filament during muscle activation.

ResultsStructural Changes in TnC Following a Ca Jump Are Much Faster thanForce Generation. Photolysis of caged Ca in a demembranatedfiber from rabbit psoas muscle held at fixed length leads to rapidforce generation with a half-time of 54 ± 12 ms (mean ± SD; n =7) at 12 °C, corresponding to a rate constant of ∼20 s−1 (Fig. 2,top trace). Ca2+ ions are liberated from caged Ca with a rateconstant of ca. 68,000 s−1 under similar conditions (11), threeorders of magnitude faster than force generation. In the con-ditions used here, Ca2+ is expected to bind to the regulatory sites

Significance

Muscle contraction is controlled by structural changes in thethin filament of the muscle sarcomere triggered by calciumbinding to troponin. It has long been suspected that myosinbinding has an additional effect in switching on the thin fila-ment, but the biological function of this effect was unknown.We have elucidated the in situ sequence of calcium-inducedstructural changes of troponin and identified a kinetic com-ponent tracking myosin binding to the thin filament. We pro-pose a model of muscle regulation with kinetics determined bycoordinated changes in the structures of both thick and thinfilaments in response to mechanical conditions rather than, asin the conventional view, solely by the calcium transient andstructural changes in the thin filament.

Author contributions: L.F., E.B., and M.I. designed research; L.F. and E.B. performed re-search; L.F., E.B., I.R.S., and Y.-B.S. contributed new reagents/analytic tools; L.F. and E.B.analyzed data; and L.F., E.B., and M.I. wrote the paper.

The authors declare no conflict of interest.

This article is a PNAS Direct Submission. Y.E.G. is a guest editor invited by theEditorial Board.1L.F. and E.B. contributed equally to this work.2Present address: Laboratorio di Fisiologia, Dipartimento di Biologia, Università di Fire-nze, 50019 Sesto Fiorentino, Italy.

3To whom correspondence should be addressed. E-mail: malcolm.irving@kcl.ac.uk.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1321868111/-/DCSupplemental.

4626–4631 | PNAS | March 25, 2014 | vol. 111 | no. 12 www.pnas.org/cgi/doi/10.1073/pnas.1321868111

Dow

nloa

ded

by g

uest

on

Mar

ch 1

, 202

0

with a rate constant of ca. 1,000 s−1 based on measurements onisolated TnC in solution (12, 13).The changes in orientation of the C and E helices of TnC

during activation by photolysis of caged Ca were described interms of the order parameters <P2> and <P4> obtained from themeasured polarized fluorescence intensities. <P2> is a measureof how parallel the probe dipole is to the filament axis, varyingfrom +1 for perfectly parallel alignment to −1 for a perpendic-ular orientation. <P4> is a higher-order harmonic term, givinghigher resolution angular information (14). <P2> for both theC-helix probe (Fig. 2, green) and the E-helix probe (Fig. 2,magenta) decreased rapidly after photolysis of NP-EGTA, in-dicating that both probe dipoles became more perpendicular tothe filament axis, as reported previously for steady-state activa-tion (5). <P4> decreased for the C-helix probe but increased forthe E-helix probe, reflecting a greater increase in angular dis-order of the E helix on activation, as also reported previously (5).Most of the change in orientation of the C- and E-helix probes

occurred in the first few milliseconds after photolysis, as shownmore clearly in Fig. 2 Insets, and was much faster than forcedevelopment. In both cases the orientation changes also hadslower components on the 10- and 100-ms timescales, and thesewere more prominent for the E helix (Fig. 2, magenta).

The Orientation of the TnC E Helix, but Not That of the C Helix, TracksDetachment of Myosin Heads from the Thin Filaments During RapidShortening. About 320 ms after photolysis, when isometric forcedevelopment was almost complete, the length of the fiber (L0)was rapidly decreased by 4%, allowing 10–20 ms of unloadedshortening followed by isometric force redevelopment (Fig. 2).Unloaded shortening induced partial recovery of <P2> for theE-helix probe toward its relaxed value (Fig. 2, magenta), witha smaller effect on the C-helix probe (Fig. 2, green). Unloadedshortening leads to detachment of myosin heads from the thinfilaments whereas Ca2+ remains bound to the troponin regula-tory sites. Thus, these results show that myosin head binding

affects the orientation of the E helix of TnC more than that ofthe C helix.This phenomenon was investigated in more detail in a sepa-

rate series of experiments (Fig. 3) in which unloaded shorteningwas imposed during maximal activation at pCa 4.5, 12 °C, bytemperature jump. As in the caged Ca protocol, 4% L0 unloadedshortening produced little or no change in the orientation of theC helix (Fig. 3A), but a reproducible change in the orientation ofthe E helix (Fig. 3B). <P2> for the E-helix probe increased by0.036 ± 0.002 (mean ± SE; n = 8) during 4% L0 unloadedshortening, corresponding to 15.4 ± 0.7% of the difference be-tween relaxation and maximal activation (Table S1). The timecourse of the increase in <P2> was well fitted by a single expo-nential with rate constant of 320 ± 34 s−1. This orientation changeis clearly slower than the force decrease during shortening, whichhad a rate constant of ca. 2,000 s−1, and its time course matchesthat of myosin detachment from actin (∼300 s−1) during unloadedshortening in the same conditions (15). After the end of unloadedshortening, <P2> recovered with a rate constant of 14.5 ± 0.7 s−1,very similar to that of force redevelopment (12.1 ± 0.2 s−1), whichis expected to closely follow that of reattachment of myosin toactin (15–17).

Fig. 1. Troponin core complex with bifunctional probes on the muscle thinfilament. On the left, the structure of the core complex of troponin fromskeletal muscle in the Ca2+-saturated form (10) is shown, containing TnC(yellow) and parts of TnI (blue) and TnT (orange). BR probes cross-linkedcysteines 56 and 63 (red spheres) along the C helix (green) in the TnC N lobeor cysteines 96 and 103 along the E helix (magenta) in the TnC C lobe. Blackspheres indicate the Ca2+ regulatory sites and gray spheres the Ca2+/Mg2+

sites. On the right is the schematic representation of the troponin corecomplex oriented on the actin filament (light gray) according to the modelof Knowles et al. (7). Each complex is anchored to one tropomyosin strand(dark gray) through the N terminus of TnT (orange arrow). Orange and bluesticks form the IT arm. The C terminus of TnI contains the two actin-bindingregions (small blue rectangles), the switch peptide (blue triangle) and theremainder of the mobile domain (large blue rectangle). The position of theprobes labeling C and E helices of TnC is marked by green and magentarectangles respectively.

Fig. 2. Time courses of orientation changes in TnC C and E helices duringmuscle activation by photolysis of caged Ca. The vertical arrow indicates thetime of the UV flash (time 0). (Top and Upper Middle) Force and fiber lengthchange (black lines). For simplicity only the force trace recorded during anexperiment with the E-helix probe is shown. (Lower Middle and Bottom)Order parameter (<P2> and <P4>) transients recorded during a single acti-vation by photolysis. Green dots represent the TnC C helix and magenta dotsthe TnC E helix. Data points were collected at 0.12-ms intervals. Insets showthe order parameter transients on an expanded time scale. Sarcomerelength = 2.50 μm; T = 12 °C. Black and cyan lines are the multiexponential fitson the C- and E-helix transients, respectively.

Fusi et al. PNAS | March 25, 2014 | vol. 111 | no. 12 | 4627

PHYS

IOLO

GY

Dow

nloa

ded

by g

uest

on

Mar

ch 1

, 202

0

Kinetics of Structural Changes in TnC in the Absence of Force-GeneratingMyosin Heads. To further characterize the influence of myosinbinding on the kinetics of structural changes in the thin filament,we activated fibers by photolysis of caged Ca in conditions inwhich such binding is eliminated. We used two independentapproaches to study the kinetics of thin filament activation inthe absence of force-generating myosin heads. First, we usedN-benzyl-p-toluenesulphonamide (BTS), which binds to the myosinhead domain and inhibits its strong binding to actin (18, 19). In-cubation of rabbit psoas muscle fibers in 0.1 mM BTS reducedforce in normal activating solution to less than 5% of control.Force inhibition by BTS had no effect on the difference between<P2> values for the C helix in activating and relaxing conditions(Fig. 4A, dark green bar and Table S1), or on <P4> (Fig. 4A, lightgreen bar), in agreement with previous results (6). In contrast, thedifference between <P2> values for the E helix in activating andrelaxing conditions was reduced to 71 ± 7% (mean ± SD; n = 4)(Fig. 4A, dark magenta bar and Table S1) of the change duringactivation in control conditions, similar to that for <P4> (77 ±16%) (Fig. 4A, light magenta bar), indicating that part of the ori-entation change in the E helix on activation is associated withmyosin binding to actin in force generating states (6). No force wasgenerated when Ca2+ was released from caged Ca in the presenceof 0.1 mM BTS (Fig. 4C), but large changes in <P2> and <P4>from the C-helix probe (Fig. 4C, green) and E-helix probe (Fig. 4C,magenta) still occurred during the first 5 ms after photolysis.Because myosin may still be able to bind weakly to actin in the

presence of BTS, and such weak binding might have an effect ontroponin conformation, we also activated fibers by photolysis ofcaged Ca in conditions in which all myosin–actin interaction isremoved, by stretching fibers to sarcomere length 3.9 μm beforephotolysis. The results (Fig. 4 B and D and Table S1) were es-sentially the same as those obtained with force inhibition by BTS(Fig. 4 A and C). The difference between <P2> (Fig. 4B, darkgreen bar) or <P4> (Fig. 4B, light green bar) values for the Chelix in active and relaxing conditions was the same at sarcomerelength 3.9 μm as in control conditions at 2.5 μm (5), but thecorresponding parameters for the E helix were reduced to 76 ±8% for <P2> and 89 ± 10% for <P4> (mean ± SD; n = 3) (Fig.4B, dark and light magenta bars). Resting force was high atsarcomere length 3.9 μm, and no additional force was generatedwhen Ca2+ was released from caged Ca (Fig. 4D), but the largefast changes in <P2> and <P4> from the C helix probe (Fig. 4D,green) and E-helix probe (Fig. 4D, magenta) were still present.

Structural Changes in TnC in Response to Ca2+ Binding in Situ HaveFour Kinetic Components. For quantitative analysis of the timecourses of orientation changes of the C and E helices of TnCdescribed above, we first consider the simpler case in whichstrong binding of myosin to actin and active force generation areabsent (Fig. 4). In this case three exponential components gavea good fit to the time courses of the <P2> and <P4> tran-sients, either in the presence of BTS (Fig. 4C, where fits areblack and cyan lines) or at sarcomere length 3.9 μm (Fig. 4D).Three exponential components gave a better description of thedata than two, as determined by the Akaike Information Crite-rion (SI Methods). The fitted parameters for the two protocolswere statistically indistinguishable (Table S2), and were there-fore combined (“No active force” in Table 1).The fastest and largest component (phase 1) has a rate con-

stant of ca. 1,100 s−1 (Table 1), matching that expected for Ca2+

binding (12, 13). Phase 2, with a rate constant of ca. 120 s−1 and thesecond largest amplitude, has a time course similar to that of theazimuthal motion of tropomyosin around the thin filament (20–22).Phase 3, the smallest and slowest component with a time constantof ca. 5 s−1, is likely to be associated with Ca2+/Mg2+ exchangeon troponin, which has a rate constant of 10 s−1 for isolatedTnC at 25 °C (23) (SI Text, section 1 and Fig. S1). The relativeamplitudes of phases 1–3 were similar for the C- and E-helix probes.Three exponentials also gave a good fit to the order parame-

ters describing the orientation change of the C helix followingCa2+ release in the presence of actin-bound myosin heads (Fig.2, black lines through <P2> and <P4> traces), with essentiallythe same rate constants and relative amplitudes for phases 1–3

Fig. 3. Effect of unloaded shortening and force redevelopment on TnChelix orientation for the C-helix (A) and E-helix (B) probes. (Top and UpperMiddle) Force and fiber length change. Ramp shortening (4% of initial fiberlength L0 in 3 ms) was applied at the plateau of an isometric contraction.(Lower Middle and Bottom) <P2> and <P4> time courses. <P2> and <P4>data are the average of three repeats. Vertical dashed and continuous lines,marking the ramp start and the start of force redevelopment, respectively,indicate the duration of unloaded shortening. Cyan lines in B are singleexponential fits on order parameter time courses during unloaded short-ening and force redevelopment.

Fig. 4. Kinetics of TnC orientation changes in the absence of active forcegeneration. Active force was reduced to less than 5% of control using 0.1 mMBTS at sarcomere length 2.5 μm (A and C) or by increasing the resting sar-comere length to 3.9 μm (nonoverlap, B and D). T = 12 °C. (A and B) Orderparameter change (ΔPx) during activation in presence of BTS (A) or at non-overlap (B) with respect to the change during normal activation (ΔPx0; T =12 °C; sarcomere length = 2.5 μm). Dark columns represent <P2> and lightcolumns <P4>. Horizontal dashed line indicates no change with respect tothe control values. Green columns represent the C-helix probe and the ma-genta columns the E-helix probe. (C and D) Activation of skinned musclefibers by photolysis of caged Ca in the presence of BTS (C) and at nonoverlap(D). (Top) Force response after photolysis (arrow, time 0). At sarcomerelength 3.9 μm (D) the passive force produced by the increase in sarcomerelength is ∼90% of the isometric force at full overlap. (Middle and Bottom)<P2> and <P4> transients recorded during a single maximal activation. Samecolor code as in A and B. Black and cyan lines are the three-exponential fitson C and E helix probe transients, respectively.

4628 | www.pnas.org/cgi/doi/10.1073/pnas.1321868111 Fusi et al.

Dow

nloa

ded

by g

uest

on

Mar

ch 1

, 202

0

(“Normal activation” in Table 1), consistent with the previousconclusion that the orientation of the C helix is independent ofthe presence of actin-bound myosin heads. However, for theE-helix probe it was necessary to introduce a fourth term, whichwe refer to as the “actomyosin” or “AM” phase, to describe thecontribution of the myosin heads. The rate constants of phases1–3 and the amplitude of phase 3 were constrained to be thesame as those in the no-active-force case, whereas the otherparameters were free, so that the rate constant and the ampli-tude of the AM phase were determined uniquely from the fit.With these constraints, the time courses of <P2> and <P4> forthe E-helix probe following Ca2+ release in the presence of actin-bound myosin heads were accurately reproduced (Fig. 2, cyanlines). The relative amplitude of the AM phase was 15 ± 7% for<P2> and 10± 4% for <P4> and its rate constant was 13.1 ± 6.9 s−1.These values are similar to those associated with force redevel-opment after ramp shortening (Fig. 3B), but smaller than thosereported in some previous studies (SI Text, section 2). Essentiallythe same rate constants for phases 1–4 were obtained by fittingthe <P2> and <P4> values for each time point with a Gaussiandistribution of probe orientations (Fig. S2), then applying multi-exponential fitting to the time courses of the Gaussian parameters(SI Text, section 3 and Table S3).

DiscussionUsing bifunctional rhodamine (BR) probes attached along the Chelix in the N lobe or regulatory domain of TnC, and along the Ehelix of its C lobe, in the IT arm of troponin (Fig. 1), we showedthat the dynamics of structural changes in these two troponindomains following Ca2+ binding in the intact muscle sarcomerein physiological conditions can be described by four exponentialcomponents, with rate constants of ca. 1,100 s−1 (phase 1), 120 s−1

(phase 2), 5 s−1 (phase 3), and 15 s−1 (AM phase). Because phase3 is slower than force generation, and its functional role is unclear(SI Text, section 1), it will not be considered further here.Phase 1, with a rate constant of ca. 1,100 s−1, is consistent with

previous estimates of the rate of Ca2+ binding to the regulatorysites of troponin C from skeletal muscle in isolated TnC (12, 13,24), in whole troponin and reconstituted thin filaments (25, 26),and in myofibrils (26). The largest component of the change inorientation of the E helix in the C lobe of TnC also has a rate ofca. 1,100 s−1, consistent with earlier evidence from isolated pro-teins that the structural change in the N lobe of TnC induced byCa2+ binding is rapidly transmitted to the C lobe (23). Thus, theC lobe responds equally rapidly to Ca2+ binding during physio-logical activation in the intact muscle sarcomere, and the wholetroponin core complex (Fig. 1) is reoriented on the thin filamentwith a rate constant of ca. 1,000 s−1. This conclusion is supportedby time-resolved X-ray diffraction studies on intact muscles; theintensity of the 38 nm X-ray reflection from the axial periodicityof troponin along the thin filaments increases very rapidly after

electrical stimulation, and a comparison with the free [Ca2+]transient recorded in the same preparation showed that thechange in troponin conformation lags Ca2+ binding to tro-ponin by only about 1 ms in amphibian skeletal muscle at 16 °C(27, 28). The change in troponin orientation in response to Cabinding appears to be much slower in cardiac muscle (SI Text,section 4).Phase 2, with a rate constant of 120 s−1 coincides with the

smaller, slower component of signals from probes on troponinobserved in some previous studies using isolated proteins(146 s−1 at 25 °C) (23) and myofibrils (150 s−1 at 10 °C) (26). Phase2 also has a similar time course as the azimuthal motion oftropomyosin, as reported by the changes in the intensity of thesecond actin layer line measured in X-ray diffraction studies ofactivation in intact amphibian muscle (20, 29). In a protocolsimilar to that used here, when bundles of demembranatedrabbit psoas fibers were activated by photolysis of caged Ca at4 °C (22), the increase in the intensity of the second actin layerline had two exponential components. The faster component,accounting for about 60% of the signal amplitude, had a rateconstant of 113 s−1, similar to the rate of phase 2 reported here.The slower component had a rate constant of 16 s−1, close to that ofactive force development. Thus, phase 2 coincides with the fastcomponent of tropomyosin movement.The AM phase, with a rate constant of 15 s−1 in the conditions

of the present experiments, was only observed with the E helixprobe and coincides with myosin head binding to actin. Whensuch binding was prevented, either chemically using BTS or byprestretching muscle fibers to a sarcomere length at which overlapbetween the myosin and actin filaments is removed (Fig. 4), theAM phase was absent. When rapid shortening was imposed atfull activation (Fig. 3), the orientation of the E helix, but not thatof the C helix, changed toward the relaxed orientation with a rateof ca. 300 s−1, slower than the force decrease but equal to theexpected rate of myosin head detachment (15). This experimentshows that this component of the change in orientation of theE helix coincides with myosin head binding to actin rather thanto active force per se, although during the initial activation followingCa2+ release force generation follows myosin head binding witha lag of less than 1 ms (30).The amplitude of the change in the orientation of the E helix

associated with rapid shortening, about 15% of the differencebetween the orientations corresponding to relaxation and fullactivation, was equal to the amplitude of the AM phase followingactivation by Ca jump (Table 1). During force redevelopment atthe end of rapid shortening, the orientation of the E helix re-covered with a rate constant of ca. 14 s−1, very close to that offorce redevelopment, 12 s−1, which closely tracks rebinding ofmyosin heads to actin on this timescale. The similarity of thesetwo rates, and that of the ca. 300 s−1 orientation change at thestart of rapid shortening with the myosin head detachment rate,

Table 1. Multiexponential fit parameters for TnC C- and E-helix probes

C helix, TnC-BR56–63 E helix, TnC-BR96–103

Protocol i r, s−1 <P2>, Ar <P4>, Ar r, s−1 <P2>, Ar <P4>, Ar

No activeforce

1 1,183 ± 374 0.75 ± 0.12 0.80 ± 0.19 1,098 ± 486 0.67 ± 0.12 0.61 ± 0.082 137 ± 41 0.17 ± 0.08 0.10 ± 0.15 114 ± 36 0.23 ± 0.11 0.30 ± 0.093 3.8 ± 1.2 0.08 ± 0.05 0.10 ± 0.07 5.9 ± 2.4 0.10 ± 0.02 0.09 ± 0.03

n = 5 n = 5 n = 4 n = 5 n = 5 n = 5Normal

activation1 1,096 ± 392 0.80 ± 0.05 0.82 ± 0.15 1,098 0.41 ± 0.14 0.55 ± 0.112 111 ± 31 0.13 ± 0.05 0.11 ± 0.12 114 0.38 ± 0.09 0.29 ± 0.073 7.1 ± 1.8 0.07 ± 0.03 0.07 ± 0.03 5.9 0.06 0.06

AM — — — 13.1 ± 6.9 0.15 ± 0.07 0.10 ± 0.04n = 4 n = 4 n = 4 n = 3 n = 3 n = 3

Rates and relative amplitude of each exponential component (i) of <P2> and <P4> during activation in the presence of BTS or at sarcomere length 3.9 μm(averaged as no active force) and during normal activation (mean ± SD, n = number of fibers). Values used as constraints for the fitting procedure are inboldface.

Fusi et al. PNAS | March 25, 2014 | vol. 111 | no. 12 | 4629

PHYS

IOLO

GY

Dow

nloa

ded

by g

uest

on

Mar

ch 1

, 202

0

shows that the steps between myosin head binding/unbinding andthe conformation of the IT arm are relatively fast. This generalconclusion was first deduced from the kinetics of changes in thefluorescence amplitude from a probe on Cys-133 of troponin Iwhen rapid length changes were imposed on active skeletalmuscle fibers in conditions of partial activation (31).In vitro studies of calcium regulation in isolated thin filaments

and their interaction with proteolytic myosin head fragments led tothe concept of three regulatory states of the thin filament, corre-sponding to three azimuthal positions of tropomyosin (2–4, 32):blocked, closed, and open. In the blocked state, populated at lowfree [Ca2+], tropomyosin covers the myosin binding sites on actin; inthe closed state, with Ca2+ bound to the regulatory sites on tropo-nin, tropomyosin has moved azimuthally to uncover weak myosinbinding sites on actin; in the open state a further azimuthal move-ment of tropomyosin is associated with strong binding of myosinheads in the absence of ATP. However, the structural and func-tional interaction between actin and myosin in the regularlattice of interdigitating filaments in an intact sarcomere atphysiological [ATP] may be fundamentally different from that inthe solution studies. The present results allowed this possibilityto be tested.We found that the kinetics of muscle activation of intact sar-

comeres at physiological [ATP] can be described in terms ofa sequence of four structural states of the thin filament (Fig. 5).The core complex of troponin is represented here in cartoon

form in the orientation deduced from polarized fluorescenceintensities from sets of bifunctional probes on the N and C lobesof TnC (Fig. 5, yellow) (5, 7). In the blocked state, which is pop-ulated in relaxed muscle fibers at low [Ca2+], the tropomyosinstrand (Fig. 5, dark gray) at the front right of the filament is shownfor graphical clarity as held in the blocking position by the mobiledomain (residues 131–182) (Fig. 5, large blue rectangle) at the Cterminus of TnI (33). However, it is also possible that this tropo-myosin position is stabilized by the right-hand troponin corecomplex via the N-terminal region of TnT (Fig. 5, orange arrow).In any case, in the blocked state the N lobe of TnC (Fig. 5, yellow)containing the C-helix probe (Fig. 5, green) is closed, and the in-hibitory regions of TnI (Fig. 5, small blue rectangles) are boundto actin.When Ca2+ is released and binds to the N lobe of TnC,

a state that we will refer to as the “blocked+Ca2+” state isformed at a rate of ca. 1,100 s−1 (phase 1). The N lobe of TnCis now open, creating a hydrophobic pocket to which the TnIswitch peptide (Fig. 5, blue triangle) binds and leading toa change in the orientation of both the N and C lobes of TnC(Fig. 5), but tropomyosin has not yet moved. Formation ofthe closed state with a rate constant of ca. 120 s−1 (phase 2)is coupled to dissociation of the TnI mobile and inhibitoryregions from actin, allowing azimuthal motion of tropomyosin.Finally, attachment of myosin heads (Fig. 5, black) to actininduces further azimuthal motion of tropomyosin with a rate of15 s−1 (AM phase) to form the open state. The IT arm, car-rying the E-helix probe (Fig. 5, magenta), undergoes furtherrotation during the AM phase, but the N lobe of TnC with the

C-helix probe (Fig. 5, green) does not move. Because themobile domain of TnI, the TnI switch peptide, and the N lobe ofTnC are uncoupled from the right strand of tropomyosin duringthe AM phase, this motion of the IT arm is likely to be mediatedthrough the TnT connection of the right-hand troponin, orthrough a direct effect of myosin binding on troponin.The main structural changes in the thin filaments of skeletal

muscle associated with activation (phases 1 and 2) are fast, asexpected from a tissue that has evolved for rapid control ofmovement. However, myosin binding to actin is much slowerthan thin filament activation; its rate constant is about 8× slowerthan that of phase 2, and about 75× slower than phase 1. Thus,the original concept of the steric blocking model of muscle regu-lation, in which azimuthal movement of tropomyosin allowsmyosin head binding (1, 34) must be modified to the extent thatthe rate of myosin head binding is not limited by the rate ofactivation of the thin filament; myosin head binding is relativelyslow, and this limits the rate of force generation in physiologicalconditions. The physiological rate of myosin head binding maybe limited by structural changes in the thick rather than in thethin filaments (35).The additional slow changes in the thin filaments linked to

myosin head binding in the AM phase have generally been as-sumed to be associated with positive feedback of myosin bindingon thin filament activation, as suggested by early in vitro studiesin rigor conditions (2, 4). However, in physiological conditionsthe steep cooperativity of thin filament activation is not de-pendent on myosin binding in the steady state (6, 36, 37). Thepresent results extend this conclusion to the time domain; thedominant components of thin filament activation (phases 1 and2) are much faster than myosin head binding, and are not af-fected by its presence (Figs. 2 and 4 and Table 1).What then is the functional role of the additional change in

thin filament structure induced by myosin head binding in theAM phase? One possibility suggested by the present results isthat it provides a reverse signaling pathway that allows the thinfilament to sense myosin head binding and therefore the me-chanical state of the muscle. This reverse signaling pathway isfast, as shown by the kinetics of the changes in the E-helix probeassociated with a period of rapid shortening (Fig. 3). Partialdeactivation of the thin filament during an extended period ofrapid shortening would increase muscle efficiency in conditionsin which work output is low. However, the most important func-tion of such a reverse signaling pathway may be in normal re-laxation. Mechanical relaxation in isometric conditions is muchslower than the fall of the Ca2+ transient (38–40), but can be ac-celerated by stretching the muscle (41, 42), such as might occur bythe action of an antagonist muscle in vivo. Thus, the presentresults lead to a model of muscle regulation in which the rates ofboth activation and relaxation are not determined solely by thecalcium transient and structural changes in the thin filament, butby coordinated changes in the structures of both thick and thinfilaments in response to mechanical conditions.

Fig. 5. Signaling pathway of thin filament activation. See Discussion for details.

4630 | www.pnas.org/cgi/doi/10.1073/pnas.1321868111 Fusi et al.

Dow

nloa

ded

by g

uest

on

Mar

ch 1

, 202

0

MethodsApproximately 6-mm-long segments of demembranated fibers dissectedfrom rabbit psoas muscle were mounted in the experimental trough be-tween the lever arms of a force transducer and a loudspeaker motor. Fiberextremities were fixed using 5% (vol/vol) glutaraldehyde in rigor solutionand glued to the clips with shellac dissolved in ethanol. The sarcomerelength was set at 2.50 μm and the average cross-sectional area (CSA) was5,244 ± 1,586 μm2 (mean ± SD; n = 27 fibers). The solution composition isdescribed in SI Methods.

Forty percent of the native TnC in muscle fibers was replaced by one of themutant TnCs labelled with BR (43). Polarized fluorescence intensities wererecorded using a new setup (Fig. S3) and used to calculate the second- andfourth-rank order parameters of the orientation distribution of the dipole,<P2> and <P4> respectively, and to assess the amplitude of the fast motionof the probe, characterized by <P2d> (14). Fibers were maximally activatedby photolysis of caged calcium using a ∼60-mJ pulse from a frequencydoubled ruby laser. Before the photolysis the fiber was incubated in pre-activating solution for ∼5 min and then transferred to the caged calciumsolution for another 5 min before UV flash at 12 °C. Fiber free [Ca2+] before

and after the flash was estimated from the known relationship with <P2>and <P4> for the TnC probes (6). Before the flash, pCa was set to 6.8 for eachfiber by adding to the caged calcium solution the amount of Ca2+ needed toinduce a 5–10% change toward the values of the order parameters duringmaximal activation. The activation level produced by the flash was estimatedby transferring the fiber to activating solution (T = 12 °C; pCa = 4.5) 0.6 safter the flash; the activation by photolysis was maximal when this changeof solution did not induce a further change of the order parameters (Fig. S4).Only fibers which produced maximal activation were included in the anal-ysis; for each fiber only the first activation by photolysis was used for theanalysis. Further details are in SI Methods.

ACKNOWLEDGMENTS. We thank Marty Rajaratnam, Mario Dolfi, and KevinEason for mechanical and electronic engineering support. We also thankB. D. Sykes and D. R. Trentham for comments on the manuscript. This studywas supported by the Medical Research Council and the Wellcome Trust,United Kingdom. E.B. acknowledges a grant from Fondo per gli Investimentidella Ricerca di Base–Futuro in Ricerca Project RBFR08JAMZ. Y.-B.S. helda British Heart Foundation (BHF) fellowship and I.R.S. a BHF studentship.

1. Huxley HE (1973) Structural changes in the actin- and myosin-containing filamentsduring contraction. Cold Spring Harb Symp Quant Biol 37:361–376.

2. Gordon AM, Homsher E, Regnier M (2000) Regulation of contraction in striatedmuscle. Physiol Rev 80(2):853–924.

3. McKillop DF, Geeves MA (1993) Regulation of the interaction between actin andmyosin subfragment 1: Evidence for three states of the thin filament. Biophys J 65(2):693–701.

4. Bremel RD, Weber A (1972) Cooperation within actin filament in vertebrate skeletalmuscle. Nat New Biol 238(82):97–101.

5. Ferguson RE, et al. (2003) In situ orientations of protein domains: Troponin C inskeletal muscle fibers. Mol Cell 11(4):865–874.

6. Sun YB, Brandmeier B, Irving M (2006) Structural changes in troponin in response toCa2+ and myosin binding to thin filaments during activation of skeletal muscle. ProcNatl Acad Sci USA 103(47):17771–17776.

7. Knowles AC, Irving M, Sun YB (2012) Conformation of the troponin core complex inthe thin filaments of skeletal muscle during relaxation and active contraction. J MolBiol 421(1):125–137.

8. Gagné SM, Tsuda S, Li MX, Smillie LB, Sykes BD (1995) Structures of the troponin Cregulatory domains in the apo and calcium-saturated states. Nat Struct Biol 2(9):784–789.

9. Takeda S, Yamashita A, Maeda K, Maéda Y (2003) Structure of the core domain ofhuman cardiac troponin in the Ca(2+)-saturated form. Nature 424(6944):35–41.

10. Vinogradova MV, et al. (2005) Ca(2+)-regulated structural changes in troponin. ProcNatl Acad Sci USA 102(14):5038–5043.

11. Ellis-Davies GC, Kaplan JH, Barsotti RJ (1996) Laser photolysis of caged calcium: Ratesof calcium release by nitrophenyl-EGTA and DM-nitrophen. Biophys J 70(2):1006–1016.

12. Johnson JD, Charlton SC, Potter JD (1979) A fluorescence stopped flow analysis of Ca2+exchange with troponin C. J Biol Chem 254(9):3497–3502.

13. Johnson JD, Nakkula RJ, Vasulka C, Smillie LB (1994) Modulation of Ca2+ exchange withthe Ca(2+)-specific regulatory sites of troponin C. J Biol Chem 269(12):8919–8923.

14. Dale RE, et al. (1999) Model-independent analysis of the orientation of fluorescentprobes with restricted mobility in muscle fibers. Biophys J 76(3):1606–1618.

15. Caremani M, Melli L, Dolfi M, Lombardi V, Linari M (2013) The working stroke of themyosin II motor in muscle is not tightly coupled to release of orthophosphate from itsactive site. J Physiol 591(Pt 20):5187–5205.

16. Linari M, Caremani M, Lombardi V (2010) A kinetic model that explains the effect ofinorganic phosphate on the mechanics and energetics of isometric contraction of fastskeletal muscle. Proc Biol Sci 277(1678):19–27.

17. Piazzesi G, et al. (2007) Skeletal muscle performance determined by modulation ofnumber of myosin motors rather than motor force or stroke size. Cell 131(4):784–795.

18. Linari M, Piazzesi G, Lombardi V (2009) The effect of myofilament compliance onkinetics of force generation by myosin motors in muscle. Biophys J 96(2):583–592.

19. Cheung A, et al. (2002) A small-molecule inhibitor of skeletal muscle myosin II. NatCell Biol 4(1):83–88.

20. Kress M, Huxley HE, Faruqi AR, Hendrix J (1986) Structural changes during activationof frog muscle studied by time-resolved X-ray diffraction. J Mol Biol 188(3):325–342.

21. Yagi N (2003) An x-ray diffraction study on early structural changes in skeletal musclecontraction. Biophys J 84(2 Pt 1):1093–1102.

22. Tamura T, Wakayama J, Inoue K, Yagi N, Iwamoto H (2009) Dynamics of thin-filamentactivation in rabbit skeletal muscle fibers examined by time-resolved x-ray diffraction.Biophys J 96(3):1045–1055.

23. Grabarek Z, Leavis PC, Gergely J (1986) Calcium binding to the low affinity sites introponin C induces conformational changes in the high affinity domain. A possibleroute of information transfer in activation of muscle contraction. J Biol Chem 261(2):608–613.

24. Rosenfeld SS, Taylor EW (1985) Kinetic studies of calcium and magnesium binding totroponin C. J Biol Chem 260(1):242–251.

25. Rosenfeld SS, Taylor EW (1985) Kinetic studies of calcium binding to regulatorycomplexes from skeletal muscle. J Biol Chem 260(1):252–261.

26. Lopez-Davila AJ, et al. (2012) Kinetic mechanism of Ca²⁺-controlled changes of skel-etal troponin I in psoas myofibrils. Biophys J 103(6):1254–1264.

27. Matsuo T, Iwamoto H, Yagi N (2010) Monitoring the structural behavior of troponinand myoplasmic free Ca2+ concentration during twitch of frog skeletal muscle. Biophys J99(1):193–200.

28. Yagi N (2011) Mechanism of latency relaxation in frog skeletal muscle. Prog BiophysMol Biol 105(3):180–186.

29. Matsuo T, Yagi N (2008) Structural changes in the muscle thin filament during con-tractions caused by single and double electrical pulses. J Mol Biol 383(5):1019–1036.

30. Brunello E, et al. (2006) Structural changes in the myosin filament and cross-bridgesduring active force development in single intact frog muscle fibres: Stiffness andX-ray diffraction measurements. J Physiol 577(Pt 3):971–984.

31. Brenner B, Chalovich JM (1999) Kinetics of thin filament activation probed by fluo-rescence of N-((2-(iodoacetoxy)ethyl)-N-methyl)amino-7-nitrobenz-2-oxa-1,3-diazole-labeled troponin I incorporated into skinned fibers of rabbit psoas muscle: Im-plications for regulation of muscle contraction. Biophys J 77(5):2692–2708.

32. Lehman W, Craig R, Vibert P (1994) Ca(2+)-induced tropomyosin movement in Limulusthin filaments revealed by three-dimensional reconstruction. Nature 368(6466):65–67.

33. Pirani A, et al. (2006) An atomic model of the thin filament in the relaxed andCa2+-activated states. J Mol Biol 357(3):707–717.

34. Parry DA, Squire JM (1973) Structural role of tropomyosin in muscle regulation:Analysis of the x-ray diffraction patterns from relaxed and contracting muscles. J MolBiol 75(1):33–55.

35. Reconditi M, et al. (2011) Motion of myosin head domains during activation and forcedevelopment in skeletal muscle. Proc Natl Acad Sci USA 108(17):7236–7240.

36. Sun YB, Irving M (2010) The molecular basis of the steep force-calcium relation inheart muscle. J Mol Cell Cardiol 48(5):859–865.

37. Moreno-Gonzalez A, Fredlund J, Regnier M (2005) Cardiac troponin C (TnC) and a siteI skeletal TnC mutant alter Ca2+ versus crossbridge contribution to force in rabbitskeletal fibres. J Physiol 562(Pt 3):873–884.

38. Sun YB, Lou F, Edman KA (1996) The relationship between the intracellular Ca2+transient and the isometric twitch force in frog muscle fibres. Exp Physiol 81(5):711–724.

39. Blinks JR, Rüdel R, Taylor SR (1978) Calcium transients in isolated amphibian skeletalmuscle fibres: Detection with aequorin. J Physiol 277:291–323.

40. Brunello E, et al. (2009) Structural changes in myosin motors and filaments duringrelaxation of skeletal muscle. J Physiol 587(Pt 18):4509–4521.

41. Lombardi V, Piazzesi G, Goldman YE (1990) Following tetanic stimulation, relaxationis accelerated by quick stretches of frog single muscle fibres. J Physiol 426:38P.

42. Caputo C, Edman KA, Lou F, Sun YB (1994) Variation in myoplasmic Ca2+ concen-tration during contraction and relaxation studied by the indicator fluo-3 in frogmuscle fibres. J Physiol 478(Pt 1):137–148.

43. Corrie JE, Craik JS, Munasinghe VR (1998) A homobifunctional rhodamine for labelingproteins with defined orientations of a fluorophore. Bioconjug Chem 9(2):160–167.

Fusi et al. PNAS | March 25, 2014 | vol. 111 | no. 12 | 4631

PHYS

IOLO

GY

Dow

nloa

ded

by g

uest

on

Mar

ch 1

, 202

0