Amplitude of the actomyosin power stroke dependsstrongly on the isoform of the myosin essentiallight chainPiyali Guhathakurta, Ewa Prochniewicz, and David D. Thomas1

Department of Biochemistry, Molecular Biology, and Biophysics, University of Minnesota, Minneapolis, MN 55455

Edited by James A. Spudich, Stanford University School of Medicine, Stanford, CA, and approved March 5, 2015 (received for review October 20, 2014)

We have used time-resolved fluorescence resonance energy trans-fer (TR-FRET) to determine the role of myosin essential light chains(ELCs) in structural transitions within the actomyosin complex.Skeletal muscle myosins have two ELC isoforms, A1 and A2, whichdiffer by an additional 40–45 residues at the N terminus of A1, andsubfragment 1 (S1) containing A1 (S1A1) has higher catalytic effi-ciency and higher affinity for actin than S1A2. ELC’s location at thejunction between the catalytic and light-chain domains gives it thepotential to play a central role in the force-generating powerstroke. Therefore, we measured site-directed TR-FRET between adonor on actin and an acceptor near the C terminus of ELC, detect-ing directly the rotation of the light-chain domain (lever arm) rel-ative to actin (power stroke), induced by the interaction of ATP-bound myosin with actin. TR-FRET resolved the weakly bound (W)and strongly bound (S) states of actomyosin during the W-to-Stransition (power stroke). We found that the W states are essen-tially the same for the two isoenzymes, but the S states are quitedifferent, indicating a much larger movement of S1A1. FRET fromactin to a probe on the N-terminal extension of A1 showed closeproximity to actin. We conclude that the N-terminal extension ofA1-ELC modulates the W-to-S structural transition of acto-S1, sothat the light-chain domain undergoes a much larger power strokein S1A1 than in S1A2. These results have profound implications forunderstanding the contractile function of actomyosin, as neededin therapeutic design for muscle disorders.

muscle | actomyosin | light chains | fluorescence resonance energy transfer

Acentral challenge in the biophysics of motility is to understandthe molecular mechanisms of structural transitions in theactomyosin ATPase (enzyme code 3.6.4.1) cycle, in which myosinalters its affinity for actin in a nucleotide-dependent manner fromweak (W) actin-binding states (with ATP or ADP-Pi bound to themyosin active site) to strong (S) actin-binding states (with ADP orno nucleotide, i.e., rigor), resulting in the generation of force. It isparticularly important to obtain direct information about thestructural dynamics of the whole actomyosin complex, becausethere is evidence that each protein, when studied separately,undergoes changes in structure and dynamics during transitionsfrom theW to S states (1–5). S states are relatively stable and easyto study, but theW actomyosin complex is less accessible to study,due to its transient and structurally dynamic nature.The myosin heavy chain starts with the N-terminal catalytic

domain (CD), which contains the nucleotide-binding pocket andthe actin-binding site. A small converter region connects the CDto an extended α-helical segment that forms the core of the light-chain domain (LCD) (6, 7). The LCD contains two IQ motifs thatserve as binding sites for one essential light chain (ELC) and oneregulatory light chain (RLC) and is proposed to function as a le-ver arm that amplifies small structural changes in the CD into thelarger motions by which myosin exerts force on actin. In mam-malian muscle, RLC modulates the interaction of the myosinhead with actin through phosphorylation of residues near theN terminus (8, 9), whereas the functional role of ELC is not as well

understood. However, important insight into ELC function hasbeen gained by studying its distinct isoforms. Fast skeletal musclehas two ELC isoforms, A1 and A2, which differ primarily at theirN-terminal regions, where A1 contains an additional domaincontaining 40–45 aa (10). In contrast, slow skeletal muscle andheart contain only the longer A1 isoform (11). In cardiac muscle,both ventricular and atrial ELC isoforms have N-terminal exten-sions of approximately the same length as their A1 counterparts inskeletal muscle and are important regulators of cardiac contrac-tility (12, 13). In smooth muscle and nonmuscle myosins, only theshort (A2) isoforms are expressed (14, 15). Functional differencesbetween the A1 and A2 isoforms are observed only in the pres-ence of actin (16), resulting in much higher catalytic efficiency forthe A1 isoform (17) and higher in vitro actin motility for A2 (18).Cryo-EM of skeletal muscle acto-S1 (19) suggests that the

N terminus of A1 is near the C terminus of actin in the absenceof nucleotide (Fig. 1). Removal of ELC from skeletal musclemyosin strongly inhibits the in vitro motility of actin and actomyosinforce while retaining ∼50% of ATPase activity (12, 20–22). MyosinS1 containing A1 (S1A1) showed higher apparent affinity for actin(lower KATPase) than S1A2, suggesting that intermolecular contactsbetween A1 and actin result in stronger and slower interaction ofmyosin cross-bridges with actin (18, 23, 24). Differences have beenobserved even between the two long isoforms of cardiac A1, in-dicating different cycling kinetics of cross-bridges containing dif-ferent ELCs (25). In smooth muscle myosin, A2 modulates theactomyosin interaction through phosphorylation of RLC (26). It isclear that the mechanism of regulation of actomyosin by ELCwarrants further study.We hypothesize that functional differences associated with the

two ELC isoforms in skeletal muscle involve isoform-specific

Significance

Myosin is an enzyme that uses energy from ATP to exert force onanother protein, actin, resulting in muscle contraction. Force isgenerated in the power stroke, when the actin–myosin complextransitions from a weak-binding structure (ATP bound to myosin)to a strong-binding structure (ADP). We detected the amplitude ofthe power stroke directly by measuring fluorescence resonanceenergy transfer from actin to myosin, with high time resolutionfollowing excitation with a pulsed laser. We found that the am-plitude of the power stroke is strongly dependent on the form ofthe essential light chain that is bound to myosin. This provides astructural explanation for previous functional observations, withimplications for muscle pathophysiology and therapeutic design.

Author contributions: P.G., E.P., and D.D.T. designed research; P.G. and E.P. performedresearch; P.G., E.P., and D.D.T. analyzed data; and P.G., E.P., and D.D.T. wrote the paper.

The authors declare no conflict of interest.

This article is a PNAS Direct Submission.1To whom correspondence should be addressed. Email: ddt@umn.edu.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1420101112/-/DCSupplemental.

4660–4665 | PNAS | April 14, 2015 | vol. 112 | no. 15 www.pnas.org/cgi/doi/10.1073/pnas.1420101112

Dow

nloa

ded

by g

uest

on

June

20,

202

1

http://crossmark.crossref.org/dialog/?doi=10.1073/pnas.1420101112&domain=pdfmailto:ddt@umn.eduhttp://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1420101112/-/DCSupplementalhttp://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1420101112/-/DCSupplementalwww.pnas.org/cgi/doi/10.1073/pnas.1420101112

effects on the structural states of both actin and myosin in theactomyosin ATPase cycle. We have previously shown that S1A1is more effective than S1A2 in restricting the amplitude of actin’sintrafilament rotational dynamics, particularly in the presence ofsaturating ATP (27), suggesting that complexes of S1A1 andS1A2 with actin are structurally different during the ATPasecycle. To test this hypothesis, we have expressed and purified A1and a truncated A1 with 44 residues deleted from the N terminus(referred to below as A2) and prepared homogeneous samples ofS1A1 and S1A2. We labeled actin at C374 with a donor probeAlexa Fluor 568 (AF568), and we labeled each ELC at its nativeC-terminal cysteine (C180 in A1, C136 in A2) with an acceptorprobe Alexa Fluor 647 (AF647) (Fig. 1). For a better un-derstanding of the role of A1’s N-terminal extension in actin–myosin interaction, we engineered a cysteine at position 16(A16C) of the A1 (Fig. 1) and labeled it with DABCYL mal-eimide, to serve as the acceptor for fluorescence resonance en-ergy transfer (FRET) from IAEDANS (donor) attached to actinC374. Structural transitions in the labeled acto-S1 complexesduring transitions from weakly to strongly bound states weredetected using time-resolved FRET (TR-FRET). A key advan-tage of TR-FRET is its ability to resolve and measure distanceswithin a bound complex directly, without interference from un-bound proteins (28, 29). This is particularly important formeasurements in the presence of ATP, which decreases the af-finity of myosin for actin, so the bound complex is only a fractionof the total protein population. The results indicate that theW-to-S structural transition in the actomyosin complex duringthe ATPase cycle is strongly modulated by ELC isoforms, withprofound implications for contractile function.

ResultsEnzymatic Properties of Labeled Isoenzymes.We have examined theeffects of labeling and exchange of A1 and A2 into myosin S1 bymeasuring functional properties of the two isoenzymes, S1A1and S1A2. The interaction of these isoenzymes with actin is

reflected in their enzymatic activities (SI Methods, Fig. S1A, andTable 1). Vmax and KATPase of S1A1 are independent of theposition of the probe (C180 or C16). Vmax of S1A1 is ∼40% thatof S1A2, whereas the apparent actin affinity (1/KATPase) is morethan seven times greater (Table 1), so the catalytic efficiency(Vmax/KATPase) of acto-S1A1 is nearly four times that of acto-S1A2. These results are similar to those reported previously forunlabeled proteins (23, 30), indicating that labeling and ELCexchange did not perturb the enzymatic properties of the twoisoenzymes. This is consistent with a previous report that the rateof product release is the same for both isoenzymes, but forS1A1 other processes become rate limiting (17). Kd measured bycosedimentation (Fig. S1B) confirmed the much greater actinaffinity of S1A1 compared with S1A2 (Table 1), consistent withprevious reports for unlabeled proteins (31).

Strongly Bound Complex. In the absence of ATP, actin affinity(measured by cosedimentation) and quenching of pyrene-labeledactin were essentially the same for both isoenzymes, showing thatboth isoenzymes, S1A1 and S1A2, bind actin stoichiometricallywith submicromolar affinity. Time-resolved fluorescence showsthat binding of S1 (labeled in the C-terminal lobe of ELC with theAF647 acceptor) to actin (labeled at C374 with the AF568 donor)leads to a faster decay (shorter lifetime) of the donor in bothisoenzymes (Fig. 2A, black to green), indicating FRET. Analysis ofFRET decays (FD+A) (SI Methods) of strongly bound complexes(Fig. 2B, green) showed that the mean distance RS between donorand acceptor probes in the strongly bound complex is independentof the fraction of bound acceptor-labeled S1, at substoichiometricas well as excess concentrations of added S1 (Fig. 3A), showingthat FRET is not affected by the proximity of multiple acceptorsand that the unbound S1 does not contribute to FRET. The dis-tance RS was quite dependent on the ELC isoform (Figs. 2B,green, and 3A), but the width of the distribution (FWHM = Γ) wasnot. The shorter RS value in S1A1 (than in S1A2) is consistent withour previous results suggesting that interaction of the N terminusof A1 with actin (Fig. 1) results in increased rigidity of actin (27).

Weakly Bound Complex.Resolution of the structural states of the Wcomplex required acquisition of TR-FRET decays during the briefsteady state of the actomyosin ATPase cycle. AF647-labeled S1(5–10 μM final concentration) was added to AF568-labeled actin(1 μM final concentration), followed by mixing with a saturatingconcentration of ATP (1–3 mM). ATP decreased the rate of de-cay, indicating a decrease in FRET (Fig. 2A, red) and thus anincrease in interprobe distance (Fig. 2B). We recorded a series ofTR-FRET waveforms at intervals of several seconds and analyzedthem globally (SI Methods) to determine the distance distributioncorresponding to the W state (Fig. 2B, red), the mole fraction ofactin having bound myosin (Fig. 2C, XB), and the mole fractions inthe W and S complexes (Fig. 2C, XW + XS = XB). For both S1A1and S1A2, the steady state of ATP hydrolysis persisted ∼1 minafter addition of ATP, as reflected in the constant values of boundmole fractions (XB) of S1A1 and S1A2 (Fig. 2C). Acquisition ofhigh signal/noise waveforms at intervals of several seconds, asrequired to accurately capture the steady-state signal (Fig. 2, red)was made possible by our direct waveform recording technology

Fig. 1. Model of acto-S1 complex with S1A1 and S1A2 (heavy chain, blue;ELC, green) bound strongly to actin (yellow). Spheres show labeling sites onactin (C374) and myosin ELC (C180, C136, and C16). Data from ref. 38.

Table 1. Interaction of labeled S1 isoforms with labeled actin inthe presence of saturating ATP

S1 Vmax, s−1 KATPase, μM Kd, μM

S1A1 9.2 ± 0.3 4.9 ± 0.7 12.3 ± 1.0S1A1(A16C) 8.8 ± 0.6 4.9 ± 1.7 6.7 ± 1.1S1A2 24.8 ± 4.7 36.7 ± 13.4 47.2 ± 5.0

SEs were calculated by the fitting program.

Guhathakurta et al. PNAS | April 14, 2015 | vol. 112 | no. 15 | 4661

BIOPH

YSICSAND

COMPU

TATIONALBIOLO

GY

Dow

nloa

ded

by g

uest

on

June

20,

202

1

http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1420101112/-/DCSupplemental/pnas.201420101SI.pdf?targetid=nameddest=STXThttp://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1420101112/-/DCSupplemental/pnas.201420101SI.pdf?targetid=nameddest=SF1http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1420101112/-/DCSupplemental/pnas.201420101SI.pdf?targetid=nameddest=SF1http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1420101112/-/DCSupplemental/pnas.201420101SI.pdf?targetid=nameddest=STXThttp://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1420101112/-/DCSupplemental/pnas.201420101SI.pdf?targetid=nameddest=STXT

(32). The steady-state value of XB was 0.64 ± 0.02 for S1A1 and0.38 ± 0.01 for S1A2. With increasing time of data acquisition,leading to exhaustion of ATP, in all cases XB gradually increasedup to the level of the strongly bound complex (Fig. 2C). Controlexperiments (Fig. S2) showed that the dissociation of the weaklybound acto-S1 complex by adding 0.1 M KCl completely elimi-nated FRET, indicating that all S1 is capable of binding ATP, thusweakening the interaction with actin.For acto-S1A1 with AF647 at C180, the W and S states are

clearly distinct in structure (Fig. 2B, Top): the mean distancebetween probes on actin and C180 of A1 decreases from 9.6 ±0.2 nm in W to 7.9 ± 0.1 nm in S. The widths ΓW and ΓS are notsignificantly different. For acto-S1A2 with AF647 at C136, the Wand S states show only a small difference in the actin–C136 dis-tance (Fig. 2B, Bottom): RW = 9.7 ± 0.2 nm and RS = 9.3 ±0.1 nm. However, the width of the distance distribution in the Wstate (ΓW = 6.0 ± 1.2 nm) is nearly double that in the S state (ΓS =3.6 ± 0.5 nm). Thus, the W-to-S transition in acto-S1A2 is asso-ciated primarily with a disorder-to-order transition. The distanceof separation (RW) is independent of the fraction of bound my-osin for both isoforms (Fig. 3B), indicating that, as in the FRETanalysis of strongly bound complexes (Fig. 3A), the mean dis-tances obtained in FRET analysis of weakly bound complexes arenot affected by the presence of multiple acceptors and that theunbound S1 does not contribute significantly to FRET. Weconclude that the W-to-S structural transition in actomyosin, asdetected by FRET between actin and ELC, is strongly modulatedby the ELC isoform, and the extent of the transition is greater forS1A1 than for S1A2.

FRET from Actin to Probe Near the N Terminus of A1. To test the hy-pothesis that the N terminus of the A1-ELC interacts with actin(Fig. 1), we performed TR-FRET from actin 374, labeled with theIAEDANS donor, to S1A1, labeled at A16C of ELC with the

DABCYL acceptor (Fig. 4). Analysis (Fig. 4, Inset) shows that inthe S state, the distance between IAEDANS-labeled actin and theN terminus of A1 is RS = 2.9 ± 0.2 nm, whereas in the W state thisdistance increases to RW = 3.6 ± 0.2 nm. These distances areclearly short enough to support the proximity (suggesting tethering)of the N terminus of A1-ELC to actin, and the distance change is inthe same direction as that observed for the probe in the C lobe(Fig. 2B, Top).

DiscussionThe distance between probes on actin and myosin dependsstrongly on the ELC isoform and the presence of ATP (Figs. 2–4),suggesting the model in Fig. 5. In the weak-binding state W, themean distance RW between probes on actin and the C lobe of ELC(C180 in A1 and C136 in A2) was essentially the same for acto-S1A1 (9.6 ± 0.2 nm) and acto-S1A2 (9.7 ± 0.2 nm), but the widthΓW was much greater in acto-S1A2 (Fig. 2B), indicating increaseddisorder (Fig. 5, Top Right). In the strong-binding state S, sub-stantial ELC-dependent differences were detected. In acto-S1A1,the mean distance RS between actin and C180 was 7.9 ± 0.1 nm,indicating a decrease by 1.7 ± 0.3 nm during theW-to-S transition.A significant decrease in distance was also observed with the ac-ceptor near the N terminus of A1, from 3.6 ± 0.2 nm (W) to 2.9 ±0.2 nm (S) (Fig. 4). In acto-S1A2, the mean distance RS betweenactin and C136 was significantly longer than in acto-S1A1, 9.3 ±0.1 nm, so the change during the W-to-S transition for the A2complex was much smaller (0.4 ± 0.3 nm). Assuming that theactin-ELC distance depends on the orientation of the LCD (leverarm) relative to actin, these results indicate that the LCD in S1A1undergoes substantial rotation during theW-to-S transition (Fig. 5,Left), whereas the extent of this rotation in S1A2 is much less andis primarily an order–disorder transition (Fig. 5, Right). Previoussteady-state FRET measurements using acto-S1A1 also detected a

Fig. 2. Structural states of the actomyosin complex detected by TR-FRET. (A) Fluorescence decay of 1 μM AF568 actin (FD, black) with 5 μM AF647-labeledS1A1 and 10 μM S1A2, respectively, in the absence (FD+A, green) and presence of saturating ATP (FD+A + ATP, red), acquired during the steady state. Fasterdecay in the presence of acceptor indicates FRET. (B) Interprobe distance distribution (best fit to a Gaussian function) corresponding to the bound acto-S1complex, determined from data in A for S (green) and W (red) complexes. (C) Time dependence of FRET-detected mole fractions of the structural states in B,after addition of ATP (at time 0) to a mixture of donor-labeled actin and acceptor-labeled myosin S1. XB (□, black) is the fraction of donor that has boundacceptor. XB = XW + XS, where XW (red) and XS (green) are the mole fractions of W and S complexes.

4662 | www.pnas.org/cgi/doi/10.1073/pnas.1420101112 Guhathakurta et al.

Dow

nloa

ded

by g

uest

on

June

20,

202

1

http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1420101112/-/DCSupplemental/pnas.201420101SI.pdf?targetid=nameddest=SF2www.pnas.org/cgi/doi/10.1073/pnas.1420101112

change in the LCD position relative to actin during the W-to-Stransition (33), but the lack of time resolution prevented accu-rate analysis of the bound W complex, and no studies were doneon S1A2.The detected distance RS = 9.3 ± 0.1 nm between probes on

C374 of actin and C136 of A2 in the S acto-S1A2 complex is ingood agreement with the predicted 9.0-nm distance betweenα-carbons of these residues in the currently available structuralmodel of skeletal acto-S1, based on the crystal structure of S1A2(6, 34). Starting from the post-power-stroke position, we modeledthe W-to-S transition in acto-S1A2, assuming upward rotations ofLCD for three experimentally reported angles 70° (smooth musclemyosin) (7), 18° (skeletal muscle myosin) (35), and 36° (scallopmyosin) (36) (SI Methods,Molecular Modeling of the Power Stroke,Fig. S3, and Tables S1 and S2). Of these three models the 18°power stroke observed for skeletal muscle myosin agrees best withour FRET data (Fig. S3 and Tables S1 and S2). Our FRET dataclearly do not agree with a much larger LCD movement (70°)proposed on the basis of crystallographic studies on smoothmuscle myosin (7). This suggests that the extent of the powerstroke is myosin specific, and changes predicted for one myosinisoform cannot be easily applied to another myosin isoform.

Role of the N-Terminal Extension of A1 in the Structural States of theActomyosin Complex. The only published structural model foracto-S1A1 (Fig. 1) was generated by computational modelingof the N-terminal extension of A1 into the acto-S1A2 complex,

assuming that the position of the C terminus of ELC and of thewhole LCD is the same for both, S1A1 and S1A2. Our data in-dicate that this assumption is not correct. In the S complex ofS1A1, the distances from actin 374 to probes on A1-ELC aresubstantially shorter than computed by the model in Fig. 1: C180was observed at 7.9 ± 0.1 nm and C16 at 2.9 ± 0.2 nm (Figs. 2B, 3,and 4), whereas the model in Fig. 1 predicted 9.0 nm and 3.6 nm,respectively. The most probable explanation for the shorter ob-served distances in the strongly bound acto-S1A1 complex is thatthe binding of the N-terminal extension of A1 to actin shifts theLCD position toward the actin filament (Fig. 5). This result isconsistent with the previously published results of electron mi-croscopy (19), functional mutations (37), cross-linking (30), andmolecular modeling (38). In contrast to the ELC-dependent dis-tances in the strongly bound complex, in the weakly bound com-plex the distance within acto-S1A1 (9.6 ± 0.2 nm) was essentiallythe same as in acto-S1A2 (9.7 ± 0.2 nm). Thus, the modeling ofthe W position of LCD in acto-S1A2 probably applies to acto-S1A1. Despite similar mean distances, the structural states ofweakly bound acto-S1A1 and acto-S1A2 complexes are clearlydifferent, as the width of the distance distribution is greater forS1A2 (6.0 ± 1.2 nm) than for S1A1 (3.6 ± 0.5 nm) (Fig. 2B). Thiscan be explained by stabilization of the position of LCD by theactin-bound N terminus in the W state of S1A1 (Fig. 4).Our findings lead to a structural model of theW-to-S transition

in which the acto-myosin power stroke, proposed as a swing of theLCD domain in response to binding and hydrolysis of ATP in theCD (6), is regulated by the interaction of the N-terminal exten-sion of A1 with actin (Fig. 5). In this model, theW-to-S transitionin acto-S1A1 is associated with a shift of the N-terminal extensionon actin, which facilitates the downward swing of the LCD andstabilizes the post-power-stroke state of LCD. In the absence ofthe N-terminal extension (S1A2), the power stroke has muchsmaller amplitude and is characterized primarily as an order-to-disorder transition. This model is consistent with our previousstudies focused on actin dynamics, where the W-to-S transitionwas much greater for acto-S1A1 than for acto-S1A2 (27).Previous spectroscopic studies on skeletal muscle fibers

detected the power stroke of the LCD, but effects of ELCisoforms were not evident. These studies (EPR, fluorescence

Fig. 3. (A) Distance of separation (RS) between probes on AF568-labeledactin and AF647-labeled S1 isoenzymes in the absence of ATP. (B) Distance ofseparation (Rw) in the presence of ATP. Total concentration of added myosinis 5 μM and 10 μM for S1A2 (left to right) and 1 μM, 2 μM, 5 μM, and 10 μM forS1A1 (left to right). Values are mean ± SEM for n = 4–6 measurements.

Fig. 4. Fluorescence decay of 1 μM actin, labeled at C374 with IAEDANS(donor), in the absence (FD, black) and presence of 5 μM S1A1, labeled at A16Cwith DABCYL (acceptor), in the absence (FD+A, green) and presence (FD+A +ATP, red) of saturating ATP, acquired during the steady state. Inset shows theinterprobe distance distribution (best fit to a Gaussian function) correspond-ing to the bound acto-S1 complex, determined for S and W complexes.

Guhathakurta et al. PNAS | April 14, 2015 | vol. 112 | no. 15 | 4663

BIOPH

YSICSAND

COMPU

TATIONALBIOLO

GY

Dow

nloa

ded

by g

uest

on

June

20,

202

1

http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1420101112/-/DCSupplemental/pnas.201420101SI.pdf?targetid=nameddest=STXThttp://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1420101112/-/DCSupplemental/pnas.201420101SI.pdf?targetid=nameddest=SF3http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1420101112/-/DCSupplemental/pnas.201420101SI.pdf?targetid=nameddest=ST1http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1420101112/-/DCSupplemental/pnas.201420101SI.pdf?targetid=nameddest=ST2http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1420101112/-/DCSupplemental/pnas.201420101SI.pdf?targetid=nameddest=SF3http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1420101112/-/DCSupplemental/pnas.201420101SI.pdf?targetid=nameddest=ST1http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1420101112/-/DCSupplemental/pnas.201420101SI.pdf?targetid=nameddest=ST2

polarization, and fluorescence anisotropy) (35, 39, 40) requiredincorporation of labeled light chains (RLC as well as ELC) intofibers, and the effects of specific ELC isoforms were probablyobscured by (i) incomplete replacement with labeled ELC,(ii) unknown isoform composition in myosin heterodimers, and(iii) the inability to distinguish between free and actin-boundheads. In the present study, we have overcome these limitationsby measuring TR-FRET on actin-bound heads containing ex-clusively one ELC isoform.The importance of the N-terminal extension of A1 for the ex-

tent of the power stroke is most evident in cardiac myosin, whereA1 is the only isoform of ELC and truncation of its 43 N-terminalresidues negatively affects cardiac contractility (41). However, thelack of the N-terminal extension is not necessarily detrimental formuscle function (42). Absence of the N-terminal extension gen-erates lower force (41), probably causing a less prominent powerstroke in S1A2, and is not related to abnormal cardiac function.On the other hand, in the skeletal muscle of aged rats, re-placement of A1 in a fraction of heads by A2 improved contrac-tility by increasing the rate of shortening (43). This is consistentwith the hypothesis that the N-terminal extension enhances iso-metric force while slowing the speed of unloaded shortening.These effects show that ELC isoforms regulate muscle contrac-tility, but also indicate that this regulation is muscle specific.

Relationship to Myosin Crystal Structures. The conventional modelof the myosin power stroke was based on the crystal structure ofskeletal muscle S1 containing a truncated form of A1, which isequivalent to S1A2 (6), so this model was unable to predict theeffect of the ELC isoform on LCD movement. Subsequent EMand crystallographic studies on smooth muscle (7), scallop (44), ornonmuscle myosins (45) with short, A2-like ELC indicated that theN-terminal extension is not required for a power stroke. However,these myosins are structurally different from skeletal muscle my-osin. The structure of the converter region and the interactionsbetween ELC and the CD show substantial differences betweenskeletal, smooth (7), and scallop myosins (44). Furthermore,although all myosins go through the actomyosin ATPase cycle, the

equilibria and kinetics within the cycle are myosin specific. Theisoform-specific transmission of nucleotide-induced structuralchanges from the CD to the LCD, which is the essence of thepower stroke, is probably necessary to accommodate diversefunctional roles of actomyosins in muscle and nonmuscle cells.

ConclusionThe N-terminal extension in the A1 light chain of myosin ELCenables S1A1 to form a complex with actin that is structurallydifferent from that of S1A2 in both pre- and post-power-strokestates. The result provides a compelling explanation for the de-pendence of muscle function on the ELC isoform, with poten-tially profound implications for the function and regulation ofstriated muscle. Future FRET studies are needed to test andrefine these models (Fig. 5).

MethodsProtein Preparations and Labeling. Actin was prepared from rabbit skeletalmuscle by extracting acetone powder in cold water, as described before (3).A total of 140 μMAlexa Fluor 568 C5maleimide and/or 700 μM1-5 IAEDANS, usedas donor, freshly dissolved in dimethylformamide, was added to 2.5 mL of 70 μMF-actin and the sample was incubated for 18 h at 4 °C. Labeling was terminatedby adding 10 mMDTT, and actin was ultracentrifuged for 30 min at 350,000 × g.The F-actin pellet was suspended in G-Mg buffer (5 mM Tris, 0.5 mM ATP,0.2 mMMgCl2, pH 7.5) followed by clarification at 300,000 × g for 10 min. Actinwas again polymerized for 30 min at 25 °C in the presence of 3 mM MgCl2 andultracentrifuged, and then the pellet was suspended in F-Mg buffer (3 mMMgCl2, 10 mM Tris, pH 7.5) containing 0.2 mM ATP. The labeled F-actin wasimmediately stabilized against depolymerization and denaturation by adding1 Meq of phalloidin.

S1 was prepared by α-chymotryptic digestion of rabbit skeletal musclemyosin. A pQE60 vector containing recombinant A1 ELC was a generous giftfrom J. Borejdo (University of North Texas, Fort Worth, TX). The A2 ELC wasprepared from A1, using PCR, by deleting 44 aa from the N terminus of A1.The PCR 5′ primer with the BsaI restriction site was designed to select E45 ofthe A1 as the start site of A2 and the 3′ primer was designed to incorporatethe XbaI site to the stop codon at the C terminus. The PCR product wassubcloned to TOPO, digested with BsaI and XbaI restriction enzymes, andthen ligated into the BsaI site of the pE-SUMOstar vector containing His andSUMO tags. Engineering of A16C was done by quick change in cys-lite A1.TheA1 and A2 sequences were confirmed by DNA sequencing. Native A1 wastransformed into Escherichia coliM15 and A16C and A2 was transformed intoE. coli BL21A1-competent cells. Transformed cells were grown for large-scalepreparation in LB media supplemented with 100 μg/mL ampicillin. A1 ex-pression was induced by adding 1 mM IPTG. A16C and A2 expression wasinduced by 1 mM IPTG and 0.2% L-arabinose. His-tagged ELCs were puri-fied from bacterial cell pellets, using Talon (Clontech) affinity resin; con-centrated; dialyzed to 10 mM Tris, pH 7.5; and stored frozen at −80 °C in150 mM sucrose.

Before labeling, ELCs were reduced with 5 mM DTT for 60 min on ice.Excess DTT was removed using Zeba spin columns (Pierce) in 50 mM Tris,pH 7.5. C180 of A1 and C136 of A2 (250 μM) were labeled with fivefold molarexcess of Alexa Fluor 647 C2 Maleimide (Invitrogen) in 6 M GuHCl, 5 mMEDTA, and 1 mM Tris(2-carboxyethyl)phosphine for 18 h at 23 °C. Unbounddye was removed by absorbing ELCs to phenyl Sepharose in PS buffer[50 mM Tris and 45% (wt/vol) saturated ammonium sulfate, pH 7.5] and thenwashing the resin with PS buffer. Labeled ELC was eluted in 50 mM Tris, pH7.5; concentrated; and exchanged to 10 mM imidazole, pH 7.0. C16 was la-beled with 10-fold molar excess of DABCYL plus maleimide (Anaspec), usingthe same procedure. AF647 labeling of both A1 and A2 and DABCYL la-beling of C16 were very efficient (>95%) as confirmed using electrosprayionization mass spectrometry. Labeled A1 and A2 were exchanged into S1 asdescribed previously (23) and as described in SI Methods.

Spectroscopic Data Acquisition and FRET Analysis. Fluorescence waveforms wereacquired using a high-performance time-resolved fluorescence spectrometerconstructed in this laboratory, which uses direct waveform recording ratherthan the conventional method of time-correlated single-photon counting (32).As shown previously, when identical samples are studied, this direct waveformrecording instrument offers 105 times higher throughput than time-correlatedsingle-photon counting while providing at least comparable performance insignal/noise, accuracy, and resolution of distinct components. AF568 actin andIAEDANS actin were excited using a passively Q-switched microchip YAG laser

Fig. 5. Proposed structural model of actomyosin in pre (red)- and post (green)-power-stroke states, based on TR-FRET measured from actin C374 to the Clobe and the N terminus of ELC (gray arrows). S1A1 (Left) undergoes a sub-stantial W-to-S transition relative to actin, whereas this movement is muchsmaller for S1A2. The N-terminal extension of A1 is shown as a curved blue line.

4664 | www.pnas.org/cgi/doi/10.1073/pnas.1420101112 Guhathakurta et al.

Dow

nloa

ded

by g

uest

on

June

20,

202

1

http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1420101112/-/DCSupplemental/pnas.201420101SI.pdf?targetid=nameddest=STXTwww.pnas.org/cgi/doi/10.1073/pnas.1420101112

(NanoUV-532; JDS Uniphase) at 532 nm and 355 nm with a pulse repetitionfrequency of 10 kHz. The high-energy (1 μJ/pulse), narrow (∼1 ns full width athalf maximum) laser pulses are highly uniform in shape and intensity. Emittedphotons were passed through a polarizer set at a magic angle 54.7°, followedby an interference band-pass filter (Semrock 600/15 nm and 470/20 nm), anddetected with a photomultiplier tube module (H5773-20; Hamamatsu) anddigitizer (Acqiris DC252; time resolution, 0.125 ns). All FRET experiments wereperformed at 25 °C in F-Mg buffer. Acto-S1 complexes were formed by adding0.5–10 μM of labeled S1A1 and S1A2 (as needed to obtain significant boundcomplex during the steady state) to 1 μM donor-labeled actin. Fluorescenceintensity decays were acquired first on the S complexes. Then ATP was addedat a final concentration of 1–3 mM (as needed to obtain a reliable steady state,ensuring saturation), and samples were mixed quickly and measurementsstarted typically within 5 s after addition of nucleotide. Consecutive decayswere measured every ∼5 s for 10 min to ensure that the steady state of theactomyosin ATPase was achieved. For S1A2, several successive decays wereaveraged, as needed to obtain sufficient signal/noise for the bound complex.

TR-FRETwaveforms were analyzed globally as described previously (28, 46),as elaborated in SI Methods. In the S state (no ATP), a two-state model wassufficient to fit the data, a mole fraction (1 − XB) of unbound donor anda single-Gaussian distance distribution, characterized by a center distance RSand a full width at half maximum ΓS (Figs. 2B and 4, green). After addition ofATP, a three-state model was necessary and sufficient (SI Methods and Figs.S4 and S5) to fit the data: a mole fraction (1 − XB) of unbound donor, a molefraction XS of the S state characterized by a center distance RS and a full widthat half maximum ΓS, and a mole fraction XW (= XB − XS) of the W statecharacterized by a center distance RW and a full width at half maximum ΓW(Fig. 2B, red). In Fig. 2C, these mole fractions are plotted vs. time after ad-dition of ATP.

ACKNOWLEDGMENTS. We thank Joseph M. Muretta, Bengt Svensson, andKurt C. Petersen for technical advice. We thank Octavian Cornea for assistancewith preparation of the manuscript. Fluorescence experiments were per-formed at the Biophysical Spectroscopy Center, University of Minnesota. Thisstudy was supported by NIH Grants AR032961 and AG26160 (to D.D.T.).

1. Thomas DD, Ramachandran S, Roopnarine O, Hayden DW, Ostap EM (1995) Themechanism of force generation in myosin: A disorder-to-order transition, coupled tointernal structural changes. Biophys J 68(4, Suppl):135S–141S.

2. Thomas DD, Prochniewicz E, Roopnarine O (2002) Changes in actin and myosinstructural dynamics due to their weak and strong interactions. Results Probl CellDiffer 36:7–19.

3. Prochniewicz E, Walseth TF, Thomas DD (2004) Structural dynamics of actin duringactive interaction with myosin: Different effects of weakly and strongly bound myosinheads. Biochemistry 43(33):10642–10652.

4. Thomas DD, Kast D, Korman VL (2009) Site-directed spectroscopic probes of acto-myosin structural dynamics. Annu Rev Biophys 38:347–369.

5. Thomas DD, Muretta JM, Colson BA, Mello RN, Kast D (2012) Spectroscopic probes ofmuscle proteins. Comprehensive Biophysics, ed Egelman EH (Elsevier, Amsterdam), pp226–250.

6. Rayment I, et al. (1993) Structure of the actin-myosin complex and its implications formuscle contraction. Science 261(5117):58–65.

7. Dominguez R, Freyzon Y, Trybus KM, Cohen C (1998) Crystal structure of a vertebratesmooth muscle myosin motor domain and its complex with the essential light chain:Visualization of the pre-power stroke state. Cell 94(5):559–571.

8. Sweeney HL, Bowman BF, Stull JT (1993) Myosin light chain phosphorylation in ver-tebrate striated muscle: Regulation and function. Am J Physiol 264(5 Pt 1):C1085–C1095.

9. Sanbe A, et al. (1999) Abnormal cardiac structure and function in mice expressingnonphosphorylatable cardiac regulatory myosin light chain 2. J Biol Chem 274(30):21085–21094.

10. Frank G, Weeds AG (1974) The amino-acid sequence of the alkali light chains of rabbitskeletal-muscle myosin. Eur J Biochem 44(2):317–334.

11. Lowey S, Risby D (1971) Light chains from fast and slow muscle myosins. Nature234(5324):81–85.

12. Morano I, et al. (1995) Myosin light chain-actin interaction regulates cardiac con-tractility. Circ Res 76(5):720–725.

13. Morano I (1999) Tuning the human heart molecular motors by myosin light chains.J Mol Med 77(7):544–555.

14. Trybus KM (1994) Regulation of expressed truncated smooth muscle myosins. Role ofthe essential light chain and tail length. J Biol Chem 269(33):20819–20822.

15. Chen TL, Kowalczyk PA, Ho G, Chisholm RL (1995) Targeted disruption of the Dic-tyostelium myosin essential light chain gene produces cells defective in cytokinesisand morphogenesis. J Cell Sci 108(Pt 10):3207–3218.

16. Taylor RS, Weeds AG (1977) Transient-phase of ATP hydrolysis by myosin sub-frag-ment-1 isoenzymes. FEBS Lett 75(1):55–60.

17. Wagner PD, Slater CS, Pope B, Weeds AG (1979) Studies on the actin activation ofmyosin subfragment-1 isoezymes and the role of myosin light chains. Eur J Biochem99(2):385–394.

18. Lowey S, Waller GS, Trybus KM (1993) Function of skeletal muscle myosin heavy andlight chain isoforms by an in vitro motility assay. J Biol Chem 268(27):20414–20418.

19. Milligan RA, Whittaker M, Safer D (1990) Molecular structure of F-actin and locationof surface binding sites. Nature 348(6298):217–221.

20. Timson DJ (2003) Fine tuning the myosin motor: The role of the essential light chain instriated muscle myosin. Biochimie 85(7):639–645.

21. VanBuren P, et al. (1994) The essential light chain is required for full force productionby skeletal muscle myosin. Proc Natl Acad Sci USA 91(26):12403–12407.

22. Lowey S, Waller GS, Trybus KM (1993) Skeletal muscle myosin light chains are es-sential for physiological speeds of shortening. Nature 365(6445):454–456.

23. Wagner PD, Weeds AG (1977) Studies on the role of myosin alkali light chains. Re-combination and hybridization of light chains and heavy chains in subfragment-1preparations. J Mol Biol 109(3):455–470.

24. Weeds AG, Taylor RS (1975) Separation of subfragment-1 isoenzymes from rabbitskeletal muscle myosin. Nature 257(5521):54–56.

25. Morano I, Haase H (1997) Different actin affinities of human cardiac essential myosinlight chain isoforms. FEBS Lett 408(1):71–74.

26. Taylor KA, et al. (2014) Role of the essential light chain in the activation of smoothmuscle myosin by regulatory light chain phosphorylation. J Struct Biol 185(3):375–382.

27. Prochniewicz E, Guhathakurta P, Thomas DD (2013) The structural dynamics of actinduring active interaction with myosin depends on the isoform of the essential lightchain. Biochemistry 52(9):1622–1630.

28. Li J, James ZM, Dong X, Karim CB, Thomas DD (2012) Structural and functional dy-namics of an integral membrane protein complex modulated by lipid headgroupcharge. J Mol Biol 418(5):379–389.

29. Dong X, Thomas DD (2014) Time-resolved FRET reveals the structural mechanism ofSERCA-PLB regulation. Biochem Biophys Res Commun 449(2):196–201.

30. Timson DJ, Trayer HR, Trayer IP (1998) The N-terminus of A1-type myosin essentiallight chains binds actin and modulates myosin motor function. Eur J Biochem 255(3):654–662.

31. Chalovich JM, Stein LA, Greene LE, Eisenberg E (1984) Interaction of isozymes ofmyosin subfragment 1 with actin: Effect of ionic strength and nucleotide. Bio-chemistry 23(21):4885–4889.

32. Muretta JM, et al. (2010) High-performance time-resolved fluorescence by directwaveform recording. Rev Sci Instrum 81(10):103101.

33. Trayer HR, Trayer IP (1988) Fluorescence resonance energy transfer within the com-plex formed by actin and myosin subfragment 1. Comparison between weakly andstrongly attached states. Biochemistry 27(15):5718–5727.

34. Mendelson R, Morris EP (1997) The structure of the acto-myosin subfragment 1complex: Results of searches using data from electron microscopy and x-ray crystal-lography. Proc Natl Acad Sci USA 94(16):8533–8538.

35. Mello RN, Thomas DD (2012) Three distinct actin-attached structural states of myosinin muscle fibers. Biophys J 102(5):1088–1096.

36. Baker JE, Brust-Mascher I, Ramachandran S, LaConte LE, Thomas DD (1998) A largeand distinct rotation of the myosin light chain domain occurs upon muscle contrac-tion. Proc Natl Acad Sci USA 95(6):2944–2949.

37. Lowey S, Saraswat LD, Liu H, Volkmann N, Hanein D (2007) Evidence for an in-teraction between the SH3 domain and the N-terminal extension of the essential lightchain in class II myosins. J Mol Biol 371(4):902–913.

38. Aydt EM, Wolff G, Morano I (2007) Molecular modeling of the myosin-S1(A1) isoform.J Struct Biol 159(1):158–163.

39. Midde K, et al. (2011) Evidence for pre- and post-power stroke of cross-bridges ofcontracting skeletal myofibrils. Biophys J 100(4):1024–1033.

40. Knowles AC, et al. (2008) Orientation of the essential light chain region of myosin inrelaxed, active, and rigor muscle. Biophys J 95(8):3882–3891.

41. Muthu P, et al. (2011) Structural and functional aspects of the myosin essential lightchain in cardiac muscle contraction. FASEB J 25(12):4394–4405.

42. Kazmierczak K, et al. (2014) Remodeling of the heart in hypertrophy in animal modelswith myosin essential light chain mutations. Front Physiol 5:353.

43. Kim JH, et al. (2012) Myosin light chain 3f attenuates age-induced decline in con-tractile velocity in MHC type II single muscle fibers. Aging Cell 11(2):203–212.

44. Houdusse A, Szent-Gyorgyi AG, Cohen C (2000) Three conformational states of scallopmyosin S1. Proc Natl Acad Sci USA 97(21):11238–11243.

45. Volkmann N, et al. (2005) The structural basis of myosin V processive movement asrevealed by electron cryomicroscopy. Mol Cell 19(5):595–605.

46. Muretta JM, Petersen KJ, Thomas DD (2013) Direct real-time detection of the actin-activated power stroke within the myosin catalytic domain. Proc Natl Acad Sci USA110(18):7211–7216.

Guhathakurta et al. PNAS | April 14, 2015 | vol. 112 | no. 15 | 4665

BIOPH

YSICSAND

COMPU

TATIONALBIOLO

GY

Dow

nloa

ded

by g

uest

on

June

20,

202

1

http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1420101112/-/DCSupplemental/pnas.201420101SI.pdf?targetid=nameddest=STXThttp://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1420101112/-/DCSupplemental/pnas.201420101SI.pdf?targetid=nameddest=STXThttp://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1420101112/-/DCSupplemental/pnas.201420101SI.pdf?targetid=nameddest=SF4http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1420101112/-/DCSupplemental/pnas.201420101SI.pdf?targetid=nameddest=SF4http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1420101112/-/DCSupplemental/pnas.201420101SI.pdf?targetid=nameddest=SF5