Titin strain contributes to the Frank Starling law of the heart …...2016/02/04  · Titin strain contributes to the Frank–Starling law of the heart by structural rearrangements - [PDF Document] (2024)

Titin strain contributes to the Frank–Starling law of theheartby structural rearrangements of both thin- andthick-filamentproteinsYounss Ait-Moua,1,2, Karen Hsua,b,1,3, Gerrie P. Farmana,4,Mohit Kumara, Marion L. Greaserc, Thomas C. Irvingb,and Pieter P.de Tombea,5

aDepartment of Cell and Molecular Physiology, Loyola UniversityChicago, Stritch School of Medicine, Maywood, IL 60153; bDepartmentof Biological andChemical Sciences, Illinois Institute ofTechnology, Chicago, IL 60616; and cDepartment of Animal Sciences,Muscle Biology Laboratory, University ofWisconsin–Madison, Madison,WI 53706

Edited by J. G. Seidman, Harvard Medical School, Boston, MA, andapproved January 12, 2016 (received for review August 21, 2015)

The Frank–Starling mechanism of the heart is due, in part,tomodulation of myofilament Ca2+ sensitivity by sarcomere length(SL)[length-dependent activation (LDA)]. The molecularmechanism(s)that underlie LDA are unknown. Recent evidence hasimplicated thegiant protein titin in this cellular process,possibly by positioning themyosin head closer to actin. To clarifythe role of titin strain in LDA,we isolated myocardium from eitherWT or hom*ozygous mutant(HM) rats that express a giant spliceisoform of titin, and subjectedthe muscles to stretch from 2.0 to2.4 μm of SL. Upon stretch, HMcompared with WT muscles displayedreduced passive force, twitchforce, and myofilament LDA.Time-resolved small-angle X-ray diffrac-tion measurements of WTtwitching muscles during diastole revealedstretch-induced increasesin the intensity of myosin (M2 and M6) andtroponin (Tn3)reflections, as well as a reduction in cross-bridge radialspacing.Independent fluorescent probe analyses in relaxed permea-bilizedmyocytes corroborated these findings. X-ray electrondensityreconstruction revealed increased mass/ordering in boththick andthin filaments. The SL-dependent changes in structureobserved inWT myocardium were absent in HM myocardium. Overall, ourre-sults reveal a correlation between titin strain and theFrank–Star-ling mechanism. The molecular basis underlying thisphenomenonappears not to involve interfilament spacing or movementof myo-sin toward actin but, rather, sarcomere stretch-inducedsimulta-neous structural rearrangements within both thin andthickfilaments that correlate with titin strain and myofilamentLDA.

myofilament length-dependent activation | small-angle X-raydiffraction |rat | passive force | fluorescent probes

The Frank–Starling law of the heart describes a cardiacreg-ulatory control mechanism that operates on a beat-to-beatbasis(1). There is a unique relationship between ventricularend-systolic volume and end-systolic pressure that is determinedbycardiac contractility. As a result, ventricular stroke volume isdi-rectly proportional to the extent of diastolic filling. Inconjunctionwith heart rate and contractility, the Frank–Starlingmechanismconstitutes a major determinant of cardiac output.Although theFrank–Starling mechanism has been well established forwell overa century, the molecular mechanisms underlying thisphenomenonare not resolved (1). At the cellular level, an increasein sarcomerelength (SL) results in an immediate increase in twitchforce devel-opment. Existing data, mostly derived frompermeabilized isolatedmyocardium, strongly support the notion thatthis phenomenon isdue to an increase in the Ca2+ responsiveness ofthe cardiac con-tractile apparatus, a phenomenon termed“myofilament length-dependent activation” (LDA) (1).The mechanismby which the mechanical strain signal is trans-

duced by the cardiac sarcomere is not known. We haverecentlydemonstrated that LDA develops within a few millisecondsfol-lowing a change in SL (2), a finding suggestive of amolecularmechanism caused by strain-dependent mechanicalrearrange-ment of contractile proteins. Moreover, although LDA isa

general property of striated muscle, it manifests itself to amuchgreater extent in cardiac muscle compared withslow-twitchskeletal muscle (3). Cardiac LDA has been shown to bemodu-lated by contractile protein phosphorylation (4–7), as well asbycardiac disease-associated mutations within variouscontractileproteins (6). In addition, evidence has emerged that thepassiveforce originating from the giant elastic sarcomeric proteintitindirectly acts to modulate myofilament Ca2+ responsiveness(8,9). Of note, the titin molecule spans the entirehalf-sarcomerefrom the Z-disk to the center of the thick filament,and is thuswell positioned within the contractile apparatus torelay themechanical SL input signal (8). The mechanismsunderlyingthe impact of titin strain on myofilament LDA, however,areincompletely understood.A unifying theory has been advancedwhereby the distance

between the thin and thick filaments constituting themuscle’ssarcomeres is proposed to modulate myofilament Ca2+respon-siveness by affecting the probability of cross-bridgeformation.

Significance

The Frank–Starling law of the heart represents afundamentalregulatory mechanism whereby cardiac pump performanceisdirectly modulated by the extent of diastolic ventricularfillingon a beat-to-beat basis. It is now well established thatsarco-mere length (SL)-induced changes in cardiac contractileproteinresponsiveness to activating calcium ions play a major roleinthis phenomenon. However, the molecular mechanisms thatunderliethis SL-sensing property are not known. Here, we showby small-angleX-ray diffraction and fluorescent probe techniquesthat the giantprotein titin likely transmits the length signal toinducestructural alterations in both thin- and thick-filamentcon-tractile proteins. These findings provide insights into themolecularbasis of the Frank–Starling regulatory mechanism.

Author contributions: Y.A.-M., K.H., T.C.I., and P.P.d.T.designed research; Y.A.-M., K.H.,M.K., and T.C.I. performedresearch; M.L.G. contributed new reagents/analytic tools;Y.A.-M.,K.H., G.P.F., M.K., T.C.I., and P.P.d.T. analyzed data; andY.A.-M., K.H., T.C.I., and P.P.d.T.wrote the paper.

The authors declare no conflict of interest.

This article is a PNAS Direct Submission.

Freely available online through the PNAS open accessoption.1Y.A.-M. and K.H. contributed equally to this work.2Presentaddress: Department of Cardiovascular Research, Sidra Medical andResearchCenter, Doha, Qatar.

3Present address: Department of Biology, San Diego StateUniversity, San Diego,CA 92182-4614.

4Present address: Department of Biological Sciences, Universityof Massachusetts at Low-ell, Lowell, MA 01854; and Department ofPhysiology and Biophysics, Boston University,Boston, MA 02118.

5To whom correspondence should be addressed. Email:[emailprotected].

This article contains supporting information online atwww.pnas.org/lookup/suppl/doi:10.1073/pnas.1516732113/-/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.1516732113 PNAS Early Edition| 1 of 6

PHYS

IOLO

GY

Dow

nloa

ded

by g

uest

on

June

21,

202

1

http://crossmark.crossref.org/dialog/?doi=10.1073/pnas.1516732113&domain=pdf&date_stamp=2016-02-06mailto:[emailprotected]://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1516732113/-/DCSupplementalhttp://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1516732113/-/DCSupplementalwww.pnas.org/cgi/doi/10.1073/pnas.1516732113

Consistent with this notion, analyses of X-ray diffractionpatternsobtained from both isolated cardiac and skeletal musclerevealan inverse relationship between myofilament latticespacingand SL (10). However, in a multitude of experimentalmodels,we could not find a consistent correlation betweenmyofila-ment lattice spacing and myofilament Ca2+ responsiveness,ren-dering interfilament spacing a less likely candidate forthemolecular mechanism underlying LDA (1). Instead, weobtainedexperimental evidence suggesting a direct impact of SL onthespread of cooperative activation along the thin filament(11),potentially by modulation of the ordering of myosin heads inre-laxed muscle, that is, before electrical activation (12).However,the primary molecular mechanism by which the strain signalistransmitted to the contractile apparatus could not bedeterminedby those studies.Here, we use a rat model in which anaturally occurring mu-

tation within the splicing factor RBM20 disrupts titinmRNAsplicing. One result of this mutation is the cardiac expressionof agiant titin isoform in hom*ozygous mutant (HM) animals atallages (13). The presence of the giant titin isoform in HMmyo-cardium was associated with reduced cardiac passive forceuponstretch, as well as a blunted Frank–Starling response andre-duced myofilament LDA. Time-resolved small-angle X-raydif-fraction revealed stretch-induced conformational structuralchangesin both thin- and thick-filament contractile proteins duringdi-astole in WT, but not HM, muscles. Our results suggest aprom-inent contribution of titin strain to the cardiacFrank–Starlingmechanism. The mechanism underlying this phenomenonappearsnot to involve interfilament spacing or movement of myosinheadstoward actin in the relaxed muscle but, rather,stretch-inducedstructural rearrangements in both the thin and thickfilamentsthat is likely directly mediated by titin strain.

ResultsImpact on Muscle Function. Fig. 1A shows originalrecordings oftwitch force obtained in WT and HM rat myocardium.Muscleswere electrically stimulated (Fig. 1A, arrowheads) at eithershortSL (2.0 μm, red) or following stretch to long SL (2.4 μm,green).Just before electrical stimulation, muscles were exposed toabrief X-ray pulse as indicated by the blue bar (Fig. 1A).Thisprotocol was repeated 30 times every 10th twitch while aCCD-based X-ray detector recorded the 2D X-ray diffractionpattern(Fig. 2). In both WT and HM rat myocardium, stretchinducedan increase in both passive and active twitch force, but thein-crease was significantly blunted in HM muscles. Fig. 1Bsum-marizes the average normalized twitch force increaseuponsarcomere stretch. Stretch induced ∼320% of baselinetwitchforce in WT muscles (Fig. 1B, open bar), compared with∼134%

in HM muscles (Fig. 1B, solid bar), demonstrating thatreducedtitin strain is associated with a significant blunting ofthe myo-cardial Frank–Starling mechanism.

Impact on 2D X-Ray Meridional Reflections. Fig. 2A showstypical2D X-ray diffraction patterns recorded at short (Top) andlong(Bottom) SL in WT myocardium. The meridional reflections,whichrun horizontal in Fig. 2A, arise from both thin- and thick-filamentproteins, notably myosin (M1–M6), troponin (Tn1–Tn3),andreflections arising from myosin-binding protein C (Fig. S1).Themyosin-binding protein C reflections appear as a series ofdoublets,possibly the result of interference between the twohalf-sarcomeres,with each pair (C1, C2, and C4 are visible inour patterns) indexingon an ∼44-nm repeat. Stretch induced anapparent increase in some,but not all, meridional reflections, asindicated by the arrows inFig. 2A (Bottom). Fig. 2B shows theaverage meridional projectionsrecorded in WT (Top, n = 11)and HM (Bottom, n = 10) muscles atshort (red) and long (green)SL. Although significant changes insarcomere structure, asreported by the meridional reflections, wereobserved in WTmuscles, no changes were recorded in HM muscles uponstretch.All measured meridional intensities and periodicitiesobtainedfrom both groups are summarized in Table S1. On average,theintensity of the second order of myosin-binding protein C(C2,2),the second and sixth orders of myosin (M2 and M6), and thethirdorder of troponin (Tn3) increased significantly in intensityuponstretch in WT, but not HM, muscles. Moreover, upon stretch,theperiodicity of the M2 and M6 myosin reflection, as well as theTn3reflection, significantly increased in WT muscles, but notablynotin HM muscles.

Impact on 2D X-Ray Myosin Layer Lines. Fig. 3A showsrepresen-tative first-order myosin layer line projections recordedatshort (red) and long (green) SL in WT (Top) and HM(Bottom)muscles. Fig. 3A (Inset) also shows a typical 2D X-raypattern inwhich the first myosin layer line is delineated by therectangle(yellow arrow). The radial position of the center of massofmyosin heads can be estimated directly from the position offirst-intensity maxima along the layer line; that is, assuming athree-stranded thick filament with helical symmetry, the peakpositionof the first myosin layer line should correspond to thefirst maxi-mum of a J3 Bessel function with the argument 2 • π • r• R,where r is a radial reciprocal lattice coordinate and R is theradiusto the center of mass of the myosin heads around the thickfila-ment backbone (14). The radial distances of the myosin headto

Fig. 1. Impact of titin length on cardiac muscle function.Cardiac muscleswere isolated from WT or HM rats and electricallystimulated (arrowheads).(A) Force and SL recordings; SL in thediastolic phase was either maintainedat SL = 2.0 μm (red) orincreased transiently to SL = 2.4 μm (green). (B) Averagepercentageincrease of twitch force upon stretch (*P < 0.05 WT vs. HM).

Fig. 2. Two-dimensional X-ray diffraction and meridionalanalysis.(A) Representative 2D X-ray diffraction patterns in WT atshort and longSL. Stretch-induced distinct alterations in themeridional reflections (yellowarrows) are shown. (B) Averagemeridional projections at short (red) and long(green) SL. Averageintensities and periodicities are summarized in Table S1.

2 of 6 | www.pnas.org/cgi/doi/10.1073/pnas.1516732113 Ait-Mou etal.

Dow

nloa

ded

by g

uest

on

June

21,

202

1

http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1516732113/-/DCSupplemental/pnas.201516732SI.pdf?targetid=nameddest=SF1http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1516732113/-/DCSupplemental/pnas.201516732SI.pdf?targetid=nameddest=ST1http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1516732113/-/DCSupplemental/pnas.201516732SI.pdf?targetid=nameddest=ST1www.pnas.org/cgi/doi/10.1073/pnas.1516732113

the thick-filament backbone calculated from these data aresum-marized in Fig. 3B. Overall, the average radial cross-bridgepositionwas ∼19% further away from the thick filament in WTcomparedwith HM muscles. Moreover, sarcomere stretch in WTmusclespositioned the myosin head ∼8% closer to the thickfilament,whereas no such movement was recorded in the HMmuscles.

Impact on Equatorial Reflections. Fig. 4A shows averagefirst-orderequatorial projections recorded at short (red) and long(green)SL in WT (Top) and HM (Bottom) muscles. Fig. 4Bsummarizesthe average lattice spacing calculated from thesereflections(Top) and the intensity ratio between the 1,1 and 1,0planes ofsymmetry (Bottom). Stretch resulted in a reduction inlatticespacing, as reported previously (1), in both groups.Moreover,the average lattice spacing was slightly (∼2%), butsignificantly,compressed in the HM compared with WT muscles.Likewise, asreported previously (1), the I1,1/I1,0 intensity ratiodecreasedupon stretch in both muscle groups. Moreover, this ratiowassignificantly (∼12%) smaller in HM compared with WTmuscles.Traditionally, the intensity ratio has been interpreted toindicatemass movement of myosin toward actin (12). However, thecurrentdata suggest that interpretation may need to be revisited(see Dis-cussion). Finally, although the functional contractileresponses andmyofilament length-dependent properties of WT and HMmyo-cardium are very different (see Fig. 1, Fig. S2, and Table S2),theresponses of both lattice spacing and intensity ratio uponstretch arequite similar overall, indicating that neither parametercorrelateswith cardiac LDA, as we have reported previously (1, 12).Theaverage equatorial intensities, normalized to the 1,1 intensity(seeMaterials and Methods) recorded in both muscle groups aresum-marized in Table S3.

Impact on Electron Density Maps. By using phase informationesti-mated from sarcomere structural models (15, 16), werecon-structed radial projection electron density (ED) maps fromthefirst five equatorial reflections of the 2D X-ray diffractionpattern(Table S3; note, a typical 2D X-ray pattern obtained fromaWTmuscle is shown in Fig. 5, Bottom Left). The average EDmapsforshort (red) and long (green) SL and the difference map be-tween theshort and long SL (heat maps; Fig. 5, Right), calculatedfor the WTand HM muscles, are shown in Fig. 5. In WT muscle,stretch induced asignificant (P < 0.01) increase in the ED of boththe thick (7%)and thin (6%) filaments, presumably due to anincrease in orderingof both thin and thick filaments around their

lattice positions, whereas no significant changes were observedinHM muscles. Moreover, upon stretch, an unidentified ED(peakexcess density ∼25% of the peak excess thick-filamentbackbonedensity) was observed between the thin (A) and thick (M)fila-ments in WT muscles, as highlighted by the yellow arrow inthemagnified difference map in Fig. 5 (Bottom Right). Of note,cal-culations based on alternative phase assumptions yieldedsimilarrelative changes in thick- and thin-filament densities uponstretch:8% and 6% for thick and thin filaments, respectively, inWT; nosignificant change in HM; and the appearance of a linkingdensityin WT muscles at long SL that is absent in HM muscle (Fig.S3).

Impact on Recombinant Troponin C Fluorescence. To obtaininde-pendent information regarding the structural rearrangementoftroponin upon stretch, we used fluorescent probe analysisinchemically permeabilized single myocytes isolated from WT andHMmyocardium. Recombinant rat troponin C (TnC) was labeledwith thefluorescent probe 5-iodoacetamido-fluorescein (IAF) andpartiallyexchanged for endogenous TnC. In addition, actin waslabeled withAlexa-680 phalloidin (Life Technologies, ThermoScientific) tocontrol for motion artifacts. Fig. 6A shows atypical confocalrecording of a mechanically attached myocyte;images are shown forthe transmission channel (Top), redphalloidin (Middle), and greenTnC (Bottom); magnified falsered/green color images are shown(Right), together with the red/green overlay demonstratingcolocalization of the labeled TnCwith actin. Fig. 6B shows totalmyocyte TnC fluorescence nor-malized to the short SL condition as afunction of [Ca2+]. Thesedata were obtained in the presence of ahigh concentration of2,3-butanedione monoxime (50 mM), an agentthat blocksstrongly bound, force-generating actin–myosininteractions (17).In WT, but not HM, myocytes, stretch induced asignificant in-crease in TnC fluorescence at all [Ca2+]. Incontrast, increasing[Ca2+] induced a sigmoidal decrease in TnCfluorescence, as hasbeen reported previously for fluorescent probesconjugated tothis residue on TnC (18). Of note, the apparent Ca2+sensitivity,as indexed by the EC50 parameter, was not affected bysarcomerestretch (Fig. S4); the apparent level of cooperativity, asindexedby the Hill coefficient (19), which was ∼1.0, consistentwithbiochemical results obtained from isolated TnC or troponin(20),was also not affected. A similar sigmoidal relationship,albeitwith a slightly higher Ca2+ sensitivity, was observed in theHMmyocytes, despite the absence of an impact of stretch in thisgroup.Thus, stretch in WT myocytes induced a conformational

Fig. 3. Myosin layer line analysis. (A) Myosin layer lineprojections in WTand HM muscle at short and long SL scaled toradial spacing r (in nm−1); theblack arrow highlights the smallerradial spacing in the WT upon stretch.(Inset) Myosin layer lineposition (yellow arrow). (B) Average calculatedcross-bridge radialspacing (#P < 0.05 long vs. short; *P < 0.05 WT vs. HM).

Fig. 4. Equatorial analysis. (A) Average equatorial projectionsscaled tolattice spacing S (in nm−1). (B) Average calculatedlattice spacings and first-order intensity ratios (#P < 0.05long vs. short; *P < 0.05 WT vs. HM).

Ait-Mou et al. PNAS Early Edition | 3 of 6

PHYS

IOLO

GY

Dow

nloa

ded

by g

uest

on

June

21,

202

1

http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1516732113/-/DCSupplemental/pnas.201516732SI.pdf?targetid=nameddest=SF2http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1516732113/-/DCSupplemental/pnas.201516732SI.pdf?targetid=nameddest=ST2http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1516732113/-/DCSupplemental/pnas.201516732SI.pdf?targetid=nameddest=ST3http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1516732113/-/DCSupplemental/pnas.201516732SI.pdf?targetid=nameddest=ST3http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1516732113/-/DCSupplemental/pnas.201516732SI.pdf?targetid=nameddest=SF3http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1516732113/-/DCSupplemental/pnas.201516732SI.pdf?targetid=nameddest=SF4

rearrangement of troponin, as measured by a florescentprobepositioned on TnC, that was distinct from rearrangementoftroponin induced by Ca2+ ions and that was absent in theHMmyocytes.

DiscussionMyofilament Ca2+ sensitivity is modulated by SL via anunknownmolecular mechanism. Recent evidence has implicatedmechanicalstrain of the giant protein titin in this cellularprocess (8, 9). Here,we studied isolated cardiac muscle from WT andHM ratsexpressing a giant titin isoform associated with reducedpassivetension, blunted myofilament LDA properties (21), andbluntedFrank–Starling responses (9). In WT, X-ray diffractionmea-surements showed stretch-induced increases in intensityandperiodicity of several myosin and troponin reflections, as wellasa reduced cross-bridge radius. Moreover, EDreconstructionrevealed stretch-induced increased ED at both thethick- andthin-filament positions, and the appearance of anunidentifiedED spanning the space between these filaments. Thelength-dependent changes in structure and function were absent inHMmyocardium.The goal of these studies was to use small-angle X-raydif-

fraction of intact twitching myocardium from WT and titinmu-tant rats to investigate the role of changes in sarcomerestructurein modulating myofilament LDA. When we initiated thesestudies,our hypothesis was that analysis of the equator and layerlineswould reveal a radially outward movement of myosin headsatlonger length indicative of a higher degree of actomyosin

interaction in diastole in twitching muscle that we couldinvokeas part of the molecular explanation for LDA. Our currentdataconclusively rule out this possibility.The ratio of theintensities of the 1,1 equatorial reflection to

the intensities of the 1,0 reflection, I11/I10, is often used asameasure of the degree of association of cross-bridge mass withthethin filament, usually assumed to be due to a radial shift ofmassaway from the thick-filament backbone toward the thinfilament.Here, we show that the I11/I10 ratio is reduced uponsarcomerestretch in both WT and HM myocardium, contrary toour expectation.Interestingly, the I11/I10 ratio was lower at bothSL in HM comparedwith WT muscles (Fig. 4B), even though,upon stretch, it shows asimilar relative change as the change inthe WT muscles. Variousauthors (15, 22) have noted that adecrease in the I1,1/I1,0 ratiocan be due to a reduced radial or-dering of the thin filamentrelative to the thick filament, suchthat a decrease in theI1,1/I1,0 does not necessarily require radialmovement of thecross-bridges, or, alternatively, that it could bedue merely to theremoval of actin out of the bare zone in thecenter of the thickfilament upon stretch (23). The 2D ED mapscalculated from thehigher order equatorial data (Fig. 5) indeedshow lower density inthe thin- and thick-filament positions in HMcompared with WTmuscles, supporting this notion. The first my-osin layer line datacan be used to inform on this issue, becausethey can be moredirectly interpreted in terms of the radial positionof the centerof mass of the myosin heads relative to the thickfilament. Fig. 3Bshows that the radius to the center of the my-osin heads at thelong SL in WT muscles is smaller than at theshort SL, consistentwith the I1,1/I1,0 ratio data shown in Fig. 4. InHM myocardium,however, although the I1,1/I1,0 ratio is reducedupon stretch, thecentroids of the layer line maxima are un-changed. Moreover, thecalculated cross-bridge radius in the HMmyocardium is smaller atboth short SL and long SL comparedwith the cross-bridge radiusobserved in WT myocardium. Oneshould use caution, therefore, ininterpreting the I11/I10 ratio pa-rameter solely in terms of radialmotion of myosin heads. Theseresults, along with the evidence fromED maps, suggest that alarge part of the lower I11/I10 ratio atshort SL in both WT andHM myocardium is due to a higher“temperature factor”-typelattice disorder (where the myofilamentsoccupy a distribution ofpositions around the expected latticepoints), and not to radialmovements of myosin heads per se (14). Itmay also be that thepresence of the larger HM titin isoform, andits consequent loweramount of titin-based passive force, reducesthe ability of the sar-comere to maintain the myofilaments in theexpected lattice po-sitions. Such a phenomenon, by itself, mayreduce the probabilityof productive actomyosin interaction. Thishypothesis would beconsistent with reported data obtained in animalmodels, where thepresence of a longer than normal isoform of titinin the sarcomereis associated with reduced calcium saturatedmaximummyofilamentforce (9, 13, 21, 24), an observation that weconfirmed in the present

""]

Fig. 6. TnC fluorescence in permeabilized myocytes.(A, Left)Attached rat permeabilized cardiac myocyte.Transmitted light image(Top), Alexa-680 phalloidinimage (Middle), andTnC-5-iodoacetamido-fluorescein(IAF) image (Bottom) are shown. (A,Right) Expandedscale images also show the red/green mergedimage.(B) TnC fluorescence-[Ca2+] relationships; averageEC50parameters are summarized in Fig. S4.

Fig. 5. ED maps. Average radial projection ED calculated fromthe first fiveequatorial reflections. A, thin filament;M, thickfilament. (Calibration bar, 50 nm.)

4 of 6 | www.pnas.org/cgi/doi/10.1073/pnas.1516732113 Ait-Mou etal.

Dow

nloa

ded

by g

uest

on

June

21,

202

1

http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1516732113/-/DCSupplemental/pnas.201516732SI.pdf?targetid=nameddest=SF4www.pnas.org/cgi/doi/10.1073/pnas.1516732113

study (Fig. S2 and Table S2). In any case, it is clear thatmyofilamentLDA is not a result of a radial outward movement ofmyosin headstoward the thin filament upon stretch of the sarcomerein diastole, aswe suggested earlier (12).The M2 meridionalreflection is one of the so-called “forbidden”

reflections from the myosin-containing thick filament, becauseitshould not be observed if the thick filament exhibitedstrictthreefold helical symmetry. The existence of forbiddenreflectionshas been attributed to the existence of localizedregions on thethick filament with helical tracks containing myosinheads that aredistorted from their helical positions. It ispremature to attempta detailed modeling of the structural changeupon stretch (andhigher titin-based passive tension), but anincrease in the M2 re-flection intensity clearly reflects a greaterdistortion of the helicalarrangement of myosin heads, due directlyto increased thick-filament strain, or possibly due to interactionsbetween myosinand titin or involvement of myosin-binding proteinC.Stretch induced a significant reduction in the myosin radial

spacing in WT, but not HM, muscles. Moreover, in the HMmuscles,this parameter was significantly smaller and not affectedby stretch(Fig. 3). These results imply that cross-bridges movetoward thethick-filament backbone upon diastolic stretch in WTmuscles,whereas in HM muscles, cross-bridges are already closerto thethick-filament backbone and, moreover, do not relocateupon stretch.These findings are inconsistent with the notion thatmyofilamentCa2+ sensitivity is regulated by a closer approach ofmyosin towardactin. The reason why myosin radial spacing isreduced in HM musclescannot be determined from our study. Itis unlikely related tointerfilament spacing, because the relativechanges in thisparameter with changes in SL were similarbetween the WT and HMmuscles (Fig. 4) and, moreover, notcorrelated to myofilament Ca2+sensitivity at either SL in theHM muscles (Fig. S2 and TableS2).Stretch of the sarcomere in WT myocardium induced a signif-

icant increase in the periodicities of several of the myosinreflec-tions (∼0.1–0.3%), and an even larger increase in theapparentperiodicity of the Tn3 reflection (∼1.0%). Selective andvariablestretch-induced lengthening of the periodicity of some, butnot all,X-ray reflections may be an indication that this increasein peri-odicity is not due to a simple elongation of the underlyingsarco-mere structure. Rather, it is more likely that thisphenomenon isthe result of a stretch-induced change in thedistribution betweenvarious structural states sampled in time bymyosin and troponin;that is, stretch may cause a reduction in themobility of these twoprotein domains. Such a narrowing of thestructural substratedistributions could result in both an increasein the reflection in-tensity and, simultaneously, a change in itsapparent periodicity(Table S1). Of note, a recent X-ray diffractionstructural study ontetanized frog skeletal muscle (25) revealed astress-dependentalteration in myosin periodicity, which wasinterpreted by thoseinvestigators to indicate a redistribution ofmyosin heads from afolded “OFF” conformation toward an extended“ON” conformation.The structural rearrangement in troponin uponstretch as re-

ported by the fluorescent probe was distinct from andindepen-dent of the structural rearrangement induced by Ca2+activation.It is known that Ca2+ binding induces an opening of ahydrophobicpatch on TnC with affinity toward the switch peptidedomain ofTnI, ultimately resulting in release of TnI from theactin-bindingsite, freeing up the myosin site so as to initiatemuscle con-traction (26). Hence, it appears that the structuralrearrangementof troponin upon stretch is different from thestructural re-arrangement of troponin induced by Ca2+ binding toTnC. Ofinterest, the apparent binding affinity of TnC for Ca2+ wasnotaffected by SL, and there was no indication of cooperativity(19)in this process (Fig. 6 and Fig. S4), in contrast to there-lationship between active force development and [Ca2+] (Fig.S2and Table S2). The implication of this result is thatmyofilamentLDA and the steep cooperativity of myofilament Ca2+activationfor force development must be the result of molecularpro-cesses downstream of Ca2+ binding to TnC. Of interest, stretchofisolated HM permeabilized myocytes to a far greater SL (2.9μm),

where passive force is comparable to passive force seen inWTmyocytes at SL = 2.4 μm, induced increased myofilamentcalciumsensitivity (Fig. S2 and Table S2). These data support thenotionthat titin mechanical strain contributes to myofilament LDA;fur-thermore, they demonstrate that myofilament LDA is indeedop-erational in HM myocardium, albeit only at an extremely longSL,where passive force starts to develop in these myocytes. Itshould benoted that such an experiment is not feasible in theintact multi-cellular preparations used here for X-ray diffraction;presence ofextracellular elastic structures in those preparations,such as colla-gen, would resist stretch to such an extreme SL.Moreover, thoseextracellular structures would likely bear most ofthe passiveforce at such extreme SL, placing most of the mechanicalstrainexternal to the cardiac sarcomere (8, 9).In the presentstudy, we compared WT to HM myocardium

that was isolated from a rat strain with a spontaneousmutationin the splicing factor RBM20. Although this mutationpromi-nently affects the splicing of titin, it should be noted thatthismutation is also known to regulate the splicing of >50differentcardiac muscle proteins (13), including proteins involvedin cardiaccalcium homeostasis. However, for the purpose of thepresentstudy, which focused on diastolic intact multicellularmuscles andpermeabilized single myocytes, we assumed that the RBM20mu-tation only affects titin length within the cardiacsarcomere.Changes in troponin structure with changes in SL, assuggested

by the X-ray results here, point to a mechanism wherebythetitin-based strain transmitted by the putativethick/thin-filamentbridging structures directly affects theregulatory apparatus pro-moting productive actomyosin interaction,and hence more force atlonger SL. Such a mechanism is alsosupported by our fluorescentprobe findings (Fig. 6), wherebystretch in WT myocytes induceda conformational rearrangement withintroponin that was dis-tinct from the conformational rearrangementinduced by Ca2+

ions and that was, moreover, absent in HM myocytes. Inaddi-tion, extensive stretch of HM myocytes to an SL, wherepassiveforce matched the passive force recorded in the WTmyocyte,induced increased myofilament Ca2+ sensitivity (Fig. S2andTable S2), supporting the notion that myofilament LDA iscausedby titin-mediated strain and subsequent structural rear-rangementswithin the relaxed thin and thick filaments. Whatmay constitute themolecular mechanism(s) underlying thisphenomenon? Our study clearlyeliminates closer positioning ofmyosin heads toward the thinfilament (Fig. 3) and decreasedinterfilament spacing (Fig. 4).Instead, titin strain may be trans-mitted directly to the thickfilament via interactions within theA-band (8, 27) or,alternatively, via interactions between titin andthe thin filamentwithin the I-band (28). However, because wefound that the troponinand myosin structures are both strain-dependent, our resultsrequire that both molecular mechanismswould have to operatesimultaneously upon stretch.An alternative mechanism may be relatedto the ED we ob-

served bridging the thick and thin filaments upon stretch inWTmyocardium (Fig. 5). Because of the low resolution of there-construction (∼13 nm), it is not possible to determinethechemical identity of this bridging structure directly fromourdata, and caution should be exercised to not overinterpretthisdensity. The bridging density could simply be due to some ofthesmall number of cycling cross-bridges, known to exist evenindiastole (29), special “troponin bridges” linking the thickfila-ment directly to the troponin complex (30, 31), or perhapstomyosin-binding protein C. Myosin-binding protein C is emergingasan important regulator of muscle contraction (27, 32–34).Recentevidence indicates that myosin-binding protein C mayactivate thethin filament via a direct interaction between its N′domain andactin and/or tropomyosin (35, 36) that may be strain-dependent.Such a mechanism would explain, in part, the prom-inent myofilamentLDA property of the WT cardiac sarcomere,and the blunting of thisproperty in case of low titin strain (8, 9, 37),the absence ofmyosin-binding protein C (38, 39), or phosphorylationby proteinkinase A (4–7). Although it is not possible, at thistime, toidentify conclusively the conduits for titin-based strain to

Ait-Mou et al. PNAS Early Edition | 5 of 6

PHYS

IOLO

GY

Dow

nloa

ded

by g

uest

on

June

21,

202

1

http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1516732113/-/DCSupplemental/pnas.201516732SI.pdf?targetid=nameddest=SF2http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1516732113/-/DCSupplemental/pnas.201516732SI.pdf?targetid=nameddest=ST2http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1516732113/-/DCSupplemental/pnas.201516732SI.pdf?targetid=nameddest=SF2http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1516732113/-/DCSupplemental/pnas.201516732SI.pdf?targetid=nameddest=ST2http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1516732113/-/DCSupplemental/pnas.201516732SI.pdf?targetid=nameddest=ST1http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1516732113/-/DCSupplemental/pnas.201516732SI.pdf?targetid=nameddest=SF4http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1516732113/-/DCSupplemental/pnas.201516732SI.pdf?targetid=nameddest=SF2http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1516732113/-/DCSupplemental/pnas.201516732SI.pdf?targetid=nameddest=ST2http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1516732113/-/DCSupplemental/pnas.201516732SI.pdf?targetid=nameddest=SF2http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1516732113/-/DCSupplemental/pnas.201516732SI.pdf?targetid=nameddest=ST2http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1516732113/-/DCSupplemental/pnas.201516732SI.pdf?targetid=nameddest=SF2http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1516732113/-/DCSupplemental/pnas.201516732SI.pdf?targetid=nameddest=ST2

the myofilaments, it is clear that such conduits are necessarytoexplain the simultaneous thick- and thin-filamentstructuralrearrangements we observed upon stretch, rearrangementsthatare clearly correlated with increased myofilament calciumsen-sitivity. It is also clear from our current results that wecanconclusively rule out altered interfilament lattice spacing aswellas a closer approach of myosin heads toward actin at longerSLas being responsible for myofilament LDA.

Materials and MethodsRight ventricular trabeculae or smallpapillary muscles were dissected fromWTor HM titin mutant rats andmounted in an experimental chamber equippedwith a lengthcontroller, force transducer, and real-time SL detector.X-raydiffraction experiments were conducted at the BioCAT beamline18 ID at the

Advanced Photon Source, Argonne National Laboratory. Myocyteswere pre-pared from frozen tissue by mechanical hom*ogenization andattached tomicroneedles situated on an inverted confocallaser-scanning microscope.Detailed methods are provided in SIMaterials and Methods.

All experimental procedures involving live rats were performedaccordingto institutional guidelines concerning the care and use ofexperimental ani-mals, and the Institutional Animal Care and UseCommittee of the LoyolaUniversity Stritch School of Medicineapproved all protocols.

ACKNOWLEDGMENTS. We thank Peter Schemmel for assistance withX-raydata analysis. This work was supported, in part, by NIH GrantsHL075494 andGM103622. Use of the Advanced Photon Source, an Officeof Science UserFacility operated for the US Department of Energy(DOE) Office of Science byArgonne National Laboratory, wassupported by the US DOE under ContractDE-AC02-06CH11357.

1. de Tombe PP, et al. (2010) Myofilament length dependentactivation. J Mol CellCardiol 48(5):851–858.

2. Mateja RD, de Tombe PP (2012) Myofilament length-dependentactivation developswithin 5 ms in guinea-pig myocardium. Biophys J103(1):L13–L15.

3. Konhilas JP, Irving TC, de Tombe PP (2002) Length-dependentactivation in threestriated muscle types of the rat. J Physiol544(Pt 1):225–236.

4. Konhilas JP, et al. (2003) Troponin I in the murinemyocardium: Influence on length-dependent activation andinterfilament spacing. J Physiol 547(Pt 3):951–961.

5. Hanft LM, Biesiadecki BJ, McDonald KS (2013) Lengthdependence of striated muscleforce generation is controlled byphosphorylation of cTnI at serines 23/24. J Physiol591(Pt18):4535–4547.

6. Sequeira V, et al. (2013) Perturbed length-dependentactivation in human hypertrophiccardiomyopathy with missensesarcomeric gene mutations. Circ Res 112(11):1491–1505.

7. Kumar M, et al. (2015) Cardiac Myosin-binding Protein C andTroponin-I Phosphory-lation Independently Modulate MyofilamentLength-dependent Activation. J BiolChem 290(49):29241–29249.

8. Hidalgo C, Granzier H (2013) Tuning the molecular giant titinthrough phosphoryla-tion: Role in health and disease. TrendsCardiovasc Med 23(5):165–171.

9. Methawasin M, et al. (2014) Experimentally increasing titincompliance in a novelmouse model attenuates the Frank-Starlingmechanism but has a beneficial effect ondiastole. Circulation129(19):1924–1936.

10. Irving TC, Konhilas J, Perry D, Fischetti R, de Tombe PP(2000) Myofilament latticespacing as a function of sarcomere lengthin isolated rat myocardium. Am J PhysiolHeart Circ Physiol279(5):H2568–H2573.

11. Farman GP, Allen EJ, Schoenfelt KQ, Backx PH, de Tombe PP(2010) The role of thinfilament cooperativity in cardiaclength-dependent calcium activation. Biophys J99(9):2978–2986.

12. Farman GP, et al. (2011) Myosin head orientation: Astructural determinant for theFrank-Starling relationship. Am JPhysiol Heart Circ Physiol 300(6):H2155–H2160.

13. Guo W, et al. (2012) RBM20, a gene for hereditarycardiomyopathy, regulates titinsplicing. Nat Med 18(5):766–773.

14. Malinchik S, Xu S, Yu LC (1997) Temperature-inducedstructural changes in the myosinthick filament of skinned rabbitpsoas muscle. Biophys J 73(5):2304–2312.

15. Irving TC, Millman BM (1989) Changes in thick filamentstructure during compressionof the filament lattice in relaxed frogsartorius muscle. J Muscle Res Cell Motil 10(5):385–394.

16. Fujiwara S, Takezawa Y, Sugimoto Y, Wakabayashi K (2009)Neutron diffractionmeasurements of skeletal muscle using thecontrast variation technique: Analysis ofthe equatorial diffractionpatterns. J Struct Biol 167(1):25–35.

17. Farman GP, et al. (2008) Blebbistatin: Use as inhibitor ofmuscle contraction. PflugersArch 455(6):995–1005.

18. Putkey JAJ, et al. (1997) Fluorescent probes attached to Cys35 or Cys 84 in cardiactroponin C are differentially sensitive toCa(2+)-dependent events in vitro and in situ.Biochemistry36(4):970–978.

19. Rice JJ, Wang F, Bers DM, de Tombe PP (2008) Approximatemodel of cooperativeactivation and crossbridge cycling in cardiacmuscle using ordinary differential equa-tions. Biophys J95(5):2368–2390.

20. Tachampa K, et al. (2008) Increased cross-bridge cyclingkinetics after exchange ofC-terminal truncated troponin I inskinned rat cardiac muscle. J Biol Chem 283(22):15114–15121.

21. Patel JR, Pleitner JM, Moss RL, Greaser ML (2012) Magnitudeof length-dependentchanges in contractile properties varies withtitin isoform in rat ventricles. Am JPhysiol Heart Circ Physiol302(3):H697–H708.

22. Yu LC, Steven AC, Naylor GR, Gamble RC, Podolsky RJ (1985)Distribution of mass inrelaxed frog skeletal muscle and itsredistribution upon activation. Biophys J 47(3):311–321.

23. Matsubara I, Elliott GF (1972) X-ray diffraction studies onskinned single fibres of frogskeletal muscle. J Mol Biol72(3):657–669.

24. Mateja RD, Greaser ML, de Tombe PP (2013) Impact of titinisoform on length de-pendent activation and cross-bridge cyclingkinetics in rat skeletal muscle. BiochimBiophys Acta1833(4):804–811.

25. Linari M, et al. (2015) Force generation by skeletal muscleis controlled by mecha-nosensing in myosin filaments. Nature528(7581):276–279.

26. Solaro RJ, Kobayashi T (2011) Protein phosphorylation andsignal transduction incardiac thin filaments. J Biol Chem286(12):9935–9940.

27. Pfuhl M, Gautel M (2012) Structure, interactions andfunction of the N-terminus ofcardiac myosin binding protein C(MyBP-C): Who does what, with what, and towhom? J Muscle Res CellMotil 33(1):83–94.

28. Kulke M, et al. (2001) Interaction between PEVK-titin andactin filaments: Origin of aviscous force component in cardiacmyofibrils. Circ Res 89(10):874–881.

29. Brenner B, Schoenberg M, Chalovich JM, Greene LE, EisenbergE (1982) Evidence forcross-bridge attachment in relaxed muscle atlow ionic strength. Proc Natl Acad SciUSA 79(23):7288–7291.

30. Agianian B, et al. (2004) A troponin switch that regulatesmuscle contraction bystretch instead of calcium. EMBO J23(4):772–779.

31. Perz-Edwards RJ, et al. (2011) X-ray diffraction evidencefor myosin-troponin con-nections and tropomyosin movement duringstretch activation of insect flight muscle.Proc Natl Acad Sci USA108(1):120–125.

32. Moss RL, Fitzsimons DP, Ralphe JC (2015) Cardiac MyBP-Cregulates the rate and forceof contraction in mammalian myocardium.Circ Res 116(1):183–192.

33. Previs MJ, Beck Previs S, Gulick J, Robbins J, Warshaw DM(2012) Molecular mechanicsof cardiac myosin-binding protein C innative thick filaments. Science 337(6099):1215–1218.

34. Sadayappan S, de Tombe PP (2014) Cardiac myosin bindingprotein-C as a centraltarget of cardiac sarcomere signaling: Aspecial mini review series. Pflugers Arch466(2):195–200.

35. Kampourakis T, Yan Z, Gautel M, Sun Y-B, Irving M (2014)Myosin binding protein-Cactivates thin filaments and inhibits thickfilaments in heart muscle cells. Proc NatlAcad Sci USA111(52):18763–18768.

36. Mun JY, et al. (2014) Myosin-binding protein C displacestropomyosin to activatecardiac thin filaments and governs theirspeed by an independent mechanism. ProcNatl Acad Sci USA111(6):2170–2175.

37. Cazorla O, Wu Y, Irving TC, Granzier H (2001) Titin-basedmodulation of calciumsensitivity of active tension in mouse skinnedcardiac myocytes. Circ Res 88(10):1028–1035.

38. Cazorla O, et al. (2006) Length and protein kinase Amodulations of myocytes incardiac myosin binding proteinC-deficient mice. Cardiovasc Res 69(2):370–380.

39. Mamidi R, Gresham KS, Stelzer JE (2014) Length-dependentchanges in contractiledynamics are blunted due to cardiac myosinbinding protein-C ablation. Front Physiol5:461.

40. Farman GP, Allen EJ, Gore D, Irving TC, de Tombe PP (2007)Interfilament spacing ispreserved during sarcomere length isometriccontractions in rat cardiac trabeculae.Biophys J 92(9):L73–L75.

41. Cingolani HE, Pérez NG, Cingolani OH, Ennis IL (2013) TheAnrep effect: 100 yearslater. Am J Physiol Heart Circ Physiol304(2):H175–H182.

42. Hammersley A (1998) FIT2D Reference Manual, Version 3.1(European SynchrotronRadiation Facility, Grenoble, France).

43. Yu LC (1989) Analysis of equatorial x-ray diffractionpatterns from skeletal muscle.Biophys J 55(3):433–440.

44. Wojdyr M (2010) Fityk: A general-purpose peak fittingprogram. J Appl Crytstallogr43:1126–1128.

45. Haselgrove JC, Huxley HE (1973) X-ray evidence for radialcross-bridge movement andfor the sliding filament model in activelycontracting skeletal muscle. J Mol Biol 77(4):549–568.

46. Haselgrove JC, Stewart M, Huxley HE (1976) Cross-bridgemovement during musclecontraction. Nature 261(5561):606–608.

47. Ait Mou Y, le Guennec J-Y, Mosca E, de Tombe PP, Cazorla O(2008) Differentialcontribution of cardiac sarcomeric proteins inthe myofibrillar force response tostretch. Pflugers Arch457(1):25–36.

48. Biesiadecki BJ, Kobayashi T, Walker JS, Solaro RJ, de TombePP (2007) The troponinC G159D mutation blunts myofilamentdesensitization induced by troponin I Ser23/24phosphorylation. CircRes 100(10):1486–1493.

49. Fan D, Wannenburg T, de Tombe PP (1997) Decreased myocytetension developmentand calcium responsiveness in rat rightventricular pressure overload. Circulation95(9):2312–2317.

50. Hawkins CJ, Bennett PM (1995) Evaluation of freezesubstitution in rabbit skeletalmuscle. Comparison of electronmicroscopy to X-ray diffraction. J Muscle Res CellMotil16(3):303–318.

51. Malinchik S, Yu LC (1995) Analysis of equatorial x-raydiffraction patterns frommusclefibers: factors that affect theintensities. Biophys J 68(5):2023–2031.

6 of 6 | www.pnas.org/cgi/doi/10.1073/pnas.1516732113 Ait-Mou etal.

Dow

nloa

ded

by g

uest

on

June

21,

202

1