Microfluidic perfusion shows intersarcomere dynamicswithin single skeletal muscle myofibrilsFelipe de Souza Leitea, Fabio C. Minozzoa, David Altmanb, and Dilson E. Rassiera,c,d,1

aDepartment of Kinesiology and Physical Education, McGill University, Montreal, QC, Canada, H2W 1S4; bDepartment of Physics, Willamette University,Salem, OR 97301; cDepartment of Physiology, McGill University, Montreal, QC, Canada, H3G 1Y6; and dDepartment of Physics, McGill University, Montreal,QC, Canada H3A 2T8

Edited by James A. Spudich, Stanford University School of Medicine, Stanford, CA, and approved July 5, 2017 (received for review January 30, 2017)

The sarcomere is the smallest functional unit of myofibrils instriated muscles. Sarcomeres are connected in series through anetwork of elastic and structural proteins. During myofibril activa-tion, sarcomeres develop forces that are regulated through complexdynamics among their structures. The mechanisms that regulateintersarcomere dynamics are unclear, which limits our understand-ing of fundamental muscle features. Such dynamics are associatedwith the loss in forces caused by mechanical instability encounteredin muscle diseases and cardiomyopathy and may underlie potentialtarget treatments for such conditions. In this study, we developed amicrofluidic perfusion system to control one sarcomere within amyofibril, while measuring the individual behavior of all sarco-meres. We found that the force from one sarcomere leads toadjustments of adjacent sarcomeres in a mechanism that is de-pendent on the sarcomere length and the myofibril stiffness. Weconcluded that the cooperative work of the contractile and theelastic elements within a myofibril rules the intersarcomere dynam-ics, with important consequences for muscle contraction.

muscle contraction | sarcomere | intersarcomere dynamics |force development | microfluidic perfusion

The smallest contractile unit of animal striated muscles is thesarcomere, which is formed from a bipolar array of thick and

thin filaments composed mostly of myosin and actin proteins, re-spectively. The cyclic interaction between myosin and actin driven byATP hydrolysis drives sarcomere shortening and ultimately, producesforce (1, 2). In addition, the sarcomeres are structurally inter-connected through the Z disks to form myofibrils (1, 3). Therefore,individual sarcomeres are continuously interacting with each other.The sarcomeres contain titin, an elastic protein that spans the half-sarcomere (4, 5). Titin molecules link all of the different areas of asingle sarcomere as well as adjoining sarcomeres, creating an elasticnetwork throughout the length of a myofibril (6, 7). Because thesarcomeres are connected in series in a myofibril, changes in thelength of one sarcomere on activation may affect the length of ad-jacent sarcomeres. Such phenomenon is hereupon referred to asintersarcomere dynamics, which may affect force in ways that aredifficult to predict based solely on the sliding filament theory.The classic force–sarcomere length (FSL) relationship is widely

used to predict force from the overlap between thick and thin fil-aments in a sarcomere during Ca2+ activation, because differentsarcomere lengths (SLs) lead to different degrees of filamentoverlap. However, the presence of intersarcomere dynamics andnonuniformity of SLs (8) provide additional complexity to the sys-tem during myofibril activation. It has been known that the lengthsof different sarcomeres change and differ significantly during acti-vation, with consequences for force production. Studies in-vestigating spontaneous oscillatory contractions (9) and the residualforce enhancement observed after myofibrils are stretched duringactivation (10) suggest that changes in individual SLs are the resultof links between the contractile apparatus and intermediate fila-ments composed mostly of titin molecules. Accordingly, sarcomerenonuniformity during activation may lead to the extension of somesarcomeres with the consequent engagement of titin molecules for

passive force production, which enables the myofibrils to stabilize ina given activation condition. Because of technical limitations, themechanisms governing the interaction of sarcomeres in a myofibriland its consequences for force production remain unclear (11).In this study, we tested the hypotheses that the mechanical work

of one sarcomere effectively communicates with other sarcomeresin series through the passive work of titin and that myofibril me-chanics are largely governed by intersarcomere dynamics. To testthese hypotheses, we used microfluidic perfusions to locally controlone sarcomere or a predetermined group of sarcomeres within anisolated myofibril. In this way, we could manipulate the activationand relaxation of a target sarcomere while measuring the behaviorof the other sarcomeres in a myofibril. Our data show that inter-sarcomere dynamics are regulated through a mechanism thatcombines the work of the contractile apparatus and intermediatefilament systems along myofibrils.

ResultsForce Produced by Myofibrils After Point Activation of One Sarcomere.We tested isolated single sarcomeres and groups of sarcomereswithin a rabbit myofibril in three different ranges of initial sarco-mere length (SLi), which were chosen so as to induce differentinitial passive tension: 2.4–2.65 μm (very short; called here “slack”),2.65–2.9 μm (short), and >2.9 μm (long). A local flow of activationsolution, which surrounded ∼1 μm of the preparation, resulted inthe contraction of one sarcomere, whereas the other sarcomeres inseries were maintained in a relaxed state (Fig. 1A and Movies S1and S2). The striation pattern of the nonactivated sarcomeresremained clear over the repeated activation/relaxation cycles of the

Significance

The sarcomere contains a variety of molecules responsible forforce generation in striated muscles. During muscle contrac-tion, many sarcomeres work cooperatively to produce force.The mechanisms behind the interaction among sarcomeresduring muscle activation have puzzled scientists for decades.To investigate intersarcomere dynamics, we used microfluidicperfusions to activate a single sarcomere within isolatedmyofibrils. Using computational sarcomere tracking, we ob-served that the contraction of one sarcomere affects othersarcomeres in series. Studying the interactions between sar-comeres is crucial, because sarcomere nonuniformity has beenlong associated with several phenomena in muscle contractionthat cannot be easily understood.

Author contributions: F.d.S.L. and D.E.R. designed research; F.d.S.L. and F.C.M. performedresearch; F.d.S.L. and F.C.M. analyzed data; F.d.S.L. developed an experimental system;D.A. created a mathematical model requested by the reviewer; and F.d.S.L., D.A., andD.E.R. wrote the paper.

The authors declare no conflict of interest.

This article is a PNAS Direct Submission.

Freely available online through the PNAS open access option.1To whom correspondence should be addressed. Email: dilson.rassier@mcgill.ca.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1700615114/-/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.1700615114 PNAS Early Edition | 1 of 6

BIOPH

YSICSAND

COMPU

TATIONALBIOLO

GY

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 7,

202

0

target sarcomeres. The isometric forces produced by isolated singlesarcomeres (Fig. 1B) were between ∼10 and 60 nN/μm2 within therange previously published in the literature (12).At all SLi tested, the targeted sarcomere could overcome the

overall myofibril passive tension. As long as there was overlapbetween the thick and thin filaments (SL up to ∼3.6 μm), allsarcomeres reached a similar final sarcomere length (SLf) onactivation (Fig. S1B and Movie S3). The force produced by anisolated sarcomere was higher than the force produced by onetarget sarcomere that was activated within a myofibril (Fig. 1B).Furthermore, when we compared preparations with the samenumber of sarcomeres in series, the force was higher at longerSLi, suggesting a SL dependence for the force development (Fig.1 B and E). For the same SLi group, force transmitted to the

precalibrated needles decreased with increasing numbers ofsarcomeres in series. These results were reproduced in our me-chanical model, in which the elements that compose a myofibrilare modeled as either linear or nonlinear spring elements. Weinitiated the model calculations using the same average SLi thatwe used for the experiments with the myofibrils. When we acti-vated one sarcomere, the force developed by a myofibril de-creased nonlinearly as a function of the number of adjacentinactive sarcomeres and was higher at longer SLi (Fig. 1 C andF). At the end of the contraction, the forces developed in themodel were also similar to the experimental forces.In our system, the measured force is a function of the displace-

ment and stiffness of the microneedles used during mechanicaltesting. Thus, assuming that the stiffness of sarcomeres within asingle myofibril is similar, our data suggest that the measured forceis dependent on the magnitude of shortening of the target sarco-mere, which is greater at longer SLi (Fig. 1D) (P < 0.001). The forcesensed by the precalibrated needles results from the active and thepassive forces components in the myofibrils. Based on the charac-teristics of titin, it is likely that, at the beginning of activation, in-dividual sarcomere shortening results in an increased strain over thetitin molecules of the nonactivated sarcomeres. Because passiveforce develops in a nonlinear manner (13), contractions starting atlonger SLi will lead to a greater force development than contrac-tions produced at shorter lengths (for a given change in length).Furthermore, the higher passive tension may cause the activatedsarcomere to cease shortening at slightly longer SLf. Although ourdata show a lack of statistical difference (Fig. 1C) in the range ofSLf below 2.00 μm, there is a steep region of the ascending limb ofthe FSL relation, and thus, even very small differences in SLf (thatwe may have failed to detect) may strongly affect the force. Weobserved that ≥40 sarcomeres in a myofibril can adjust their lengthafter a point activation, a result that was confirmed with ourmechanical model.

Nonactivated Sarcomeres Adjust Their Lengths After Target Activation ofOne Sarcomere. To evaluate (i) how an activated target sarcomeredynamically interacts with adjacent sarcomeres and (ii) how thepassive forces in adjacent sarcomeres are adjusted along the myo-fibril, we measured the SLs (SLs; Z disk to Z disk distances) (Fig.S1A) and tracked the A-band displacements of all sarcomeres inresponse to the activation of a sarcomere in the center of themyofibril (Fig. 2 A and B). After activation, the SLfs were not dif-ferent among nonactivated sarcomeres (Fig. 2C); all of these sar-comeres responded by slightly stretching (Fig. 2D and Fig. S1) (i.e.,there is a mechanical linkage among the sarcomeres). To furtherinvestigate the processes leading to these SL adjustments, wetracked the A bands of the myofibrils (Fig. 2B and Fig. S2). Theabsolute displacement of the A bands was larger in nonactivatedsarcomeres located closer to the activation point than in sarcomereslocated farther from the activation point. The absolute A-banddisplacement was dependent on SLi and therefore, the initial pas-sive tension of myofibrils (Fig. 2B). When the A-band displacementswere normalized, this dependence of SLi was not observed (Fig.S3A). These experimental results were well-captured in our me-chanical model, which shows a decrease in the displacement of Abands when they are located in sarcomeres away from the activatedsarcomere (Fig. S3B).It has been shown that the elastic network of intermediate fila-

ments can link all force-bearing structures of the sarcomere (14).This mechanism is mainly governed by titin molecules, which arethe proteins responsible for passive force development in myofibrils.During our experiments, the local increase in Ca2+ leads the targetsarcomere to contract and stretch the adjacent sarcomeres, likelyenhancing the strain over titin filaments. This strain is passivelytransmitted across the myofibril via the intermediate filament sys-tem, affecting several sarcomeres in series.

0 5 10 15 20 25 30 35 400

10

20

30

40

50

60

70

Number of Sarcomeres

Forc

e (n

N/

m2 ) Short SLi

Slack SLi

B C

Slack Short Long Slack Short Long1.5

2.0

2.5

3.0

3.5

Sarc

omer

e Le

ngth

(µm

)

Inital Length Final Length

***

***

***

p = 0.05

0 5 10 15 200

10

20

30

40

Forc

e (n

N/

m2 )

Time (s)

Sarcomere

Microperfusion Standard Perfusion

A-band

Z-disks

A

D

0 10 20 30 40 50 60 70 800

10

20

30

40

50

60

70

Number of Sarcomeres

Forc

e (n

N/

m2 )

E

0.0 0.5 1.0 1.5 2.0 2.50

10

20

30

40

Time (s)

Forc

e (n

N/

m2 )

F

Fig. 1. Experimental system used to control myofibril subareas and forcemeasured by precalibrated flexible needles in response to the activation.(A, Upper) Scheme of the experimental system showing the microfluidicperfusion tip (blue), which is built to be smaller than the sarcomere A-bandlength. Microperfusion was controlled by a pressure system that allowedlocal flow of solutions (black dashed arrow). The red dashed arrows repre-sent the flow from the regular perfusion system coming from a double-barreled pipette, which allows rapid changes of solutions. (A, Lower) Atypical myofibril suspended between two flexible needles. The contrast andbrightness of the image were adjusted to improve visibility and clarity. Be-cause the experiments were performed using phase contrast microscopy, someareas around the needles appear brighter than other areas in the myofibrils.(Scale bar: 5 μm.) (B) Force produced by experiments with single sarcomeres(colored circles at 1 in x axis) and experiments in which activation of one sar-comere was performed within myofibrils with different numbers of sarcomeres(colored circles at 5–40 on the x axis) at three different SLi: slack (n = 34; red),short (n = 73; green) and long (n = 71; blue) lengths. All data were fit using asecond-order polynomial. (C) Model simulation of the experiments shown in B,showing the final force produced by activation of a single sarcomere as afunction of the number of adjacent inactive sarcomeres. SLi was set at 2.5, 2.8,and 3.15 μm. (D) Average SLi and SLf of the target sarcomere for the prepara-tions tested at slack, short, and long SLi. (E) Typical traces obtained from amyofibril tested at different SLs. (F) Simulation of the first seconds of force de-velopment after activation of a single sarcomere adjacent to 10 inactive sarco-meres. Error bars represent SEM. ***P < 0.001.

2 of 6 | www.pnas.org/cgi/doi/10.1073/pnas.1700615114 de Souza Leite et al.

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 7,

202

0

To further test the myofibril adjustment to local activation, werepeated the previous experiments using inflexible needles, sothat the total length of the myofibrils would not change signifi-cantly as a result of activation, removing partially the effects ofend compliance. Myofibrils were tested at long SLi (2.9–3.3 μm),very long SLi (3.4–3.6 μm), and long SLi in rigor solution (i.e., inthe absence of ATP and Ca2+), a condition that substantiallyincreases the stiffness of the preparation, because it induces alarge number of cross-bridges. For all of these conditions,myofibrils remained able to contract in response to the pointactivation, suggesting that sarcomeres in series can adapt theirlengths in response to a local increase in Ca2+ concentrationwithout significant changes in total myofibril length (Fig. 2E).There was some shortening observed at each end of the myofi-brils as indicated by A-band displacements. However, our math-ematical model that incorporates the end compliances was ableto predict sarcomere contraction and shortening even withprobes of very high stiffness and myofibrils in rigor state (Fig. S3

C and D), strengthening the interpretation that intersarcomeredynamics are not necessarily dependent on changes in totalmyofibril length.

Half-Sarcomeres Are Displaced During Activation. During the ex-periments described above, we noticed a small displacement ofthe target A band in the activated sarcomere, which may havebeen caused by half-sarcomere nonuniformity during activation.Moreover, previous studies have suggested that the stabilizationof the A band in the center of sarcomeres during muscle con-traction is regulated by titin (15, 16). Consequently, we evaluatedA-band displacements in sarcomeres by changing the position ofthe point activation to measure how it affected intersarcomeredynamics. Point activation with the microperfusion systemplaced at the Z disk led to the activation of two adjoining half-sarcomeres, because two A bands were collapsed together at theend of the activation (Movie S4).Fig. 3B shows that activation of two half-sarcomeres led to a

relatively large displacement from the zero position of at leastone of the A bands compared with one target sarcomere acti-vation. However, in this situation, the adjoining sarcomeresdisplaced less than in conditions where one entire sarcomere wasactivated. This change in position of the thick filament has beenseen before at the beginning of myofibril contraction and is at-tributed to half-sarcomere mechanics (17), leading us to in-vestigate how the myofibril responds to point activation of justone half-sarcomere. We hypothesized that the asymmetricalbehavior results from differential half-sarcomere mechanics. Wetested this hypothesis by activating half-sarcomeres using mi-cropipettes with diameters of 0.5 μm to reach a small flow area

-0.75-0.50-0.250.00 0.25 0.50 0.75 1.000

25

50

75

Sarcomere Length ( m)

Fre

qu

ency

(%

)

0.0

-12 -9 -6 -3 Act. 3 6 9 121.5

2.0

2.5

3.0

3.5

Sarcomere PositionSarc

omer

e Le

ngth

(µm

)

-12 -9 -6 -3 3 6 9 12Activation0.00

0.25

0.50

0.75

1.00

1.25

1.50

Sarcomere Position

A-B

and

Dis

plac

emen

t (m

)

Slack SLi

Short SLi

A

B

E

(+)Displacement Direction

(-) Displacement Direction

Act. ±1 ±2 ±3 ±4 ±5 ±6 ±7 ±8 ±9±10±1

1±1

20.00

0.25

0.50

0.75

1.00

1.25

± Sarcomeres PositionA-B

and

Dis

plac

emen

t (m

)

Long SLi +RIGORVery Long SLiLong SLi

C

D

Fig. 2. Response of sarcomeres after point activation. (A, Upper) A typicalview of a myofibril (100× magnification) before, during, and after pointactivation. For clarity, the side to the right of the activated sarcomere isdefined as the positive side, whereas the side to the left of the activatedsarcomere is defined as the negative side. Vertical arrows represent the di-rection of the solution flow (red arrows indicate the large flow; the blackarrow indicates the flow delivered via microperfusion). When the micro-perfusion flow is applied to the target sarcomere, it contracts and producesforce. (A, Lower) A-band tracking used for data analysis. Each white circlerepresents one A band tracked, and each color overlapping the A bandsrepresents the displacement (x, y coordinates for every frame). (Scale bar:5 μm.) (B) Absolute displacement of A bands in the myofibrils after activationof one sarcomere. The activated sarcomere is represented at the center[slack (n = 31), short (n = 47), and long (n = 67) lengths] of the figure (dottedline). The A bands of the adjacent sarcomeres are displaced toward the ac-tivated sarcomere. Activation of the target sarcomere creates a quasisym-metrical displacement of the nonactivated sarcomeres in the plus and minussides. (C) SL changes after activation of one sarcomere. Starting from dif-ferent SLi (circles), the final SLf (diamonds) of the nonactivated sarcomeresdo not change substantially after activation. (D) Histogram showing thechange in SL (ΔSL) of the nonactivated sarcomeres caused by point activa-tion of one sarcomere. The orange bars represent control measurementsthat we performed by taking two images from the same myofibril in a re-laxed state (∼5 s apart). The continuous line shows the Gaussian fit for thesecontrol experiments. The dotted line shows the Gaussian fit when threeexperimental groups are merged into the same analysis. Note that the cure isslightly skewed to the right, suggesting a small stretch of the nonactivatedsarcomeres. (E) A-band displacements of myofibrils tested in three condi-tions: long SLi (3–3.3 μm; n = 26), long SLi in rigor solution (n = 26), and verylong SLi (3.35 to ∼3.6 μm; n = 22). Error bars represent SEM. Act., activation.

-12 -9 -6 -3 3 6 9 12Act.0.00

0.25

0.50

0.75

1.00

1.25

1.50

Sarcomere Position

A-B

and

Dis

plac

emen

t (m

)

Two Half-Sarcomeresactivation

One Sarcomere - Short SLi

-12 -9 -6 -3 3 6 9 12Act.0.00

0.25

0.50

0.75

1.00

1.25

1.50

Sarcomere Position

A-B

and

Dis

plac

emen

t (m

)

A-Band remains at zero postion

A-Band shifts from zero postion

-12 -9 -6 -3 3 6 9 12Act.0.00

0.25

0.50

0.75

1.00

1.25

1.50

Sarcomere Position

A-B

and

Dis

plac

emen

t (m

)

*** ** *

A B

C D

Half-Sarcomere Two Half-Sarcomeres

A-bandZ-disks

Fig. 3. Half-sarcomere activation. A, Left shows a sarcomere in three dif-ferent conditions: (Top) before point activation, (Middle) half-sarcomereactivated at the positive side, and (Bottom) half-sarcomere activated atthe negative side. A, Top Right shows two half-sarcomeres activated to-gether. A, Middle Right shows the start of the activation; note that one half-sarcomere was contracted. A, Bottom Right shows that, after the secondhalf-sarcomere was contracted, both half-sarcomeres come close together.(Scale bar: 5 μm.) (B) A-band displacement of all sarcomeres in a myofibril inresponse to activation of two half-sarcomeres by targeting the micro-perfusion flow to a Z disk (n = 30). (C) A-band displacement in response toactivation of one half-sarcomere (n = 37). (D) Response of adjacent sarco-meres to the activation of one half-sarcomere (Z-disk displacement, n = 20;A-band displacement, n = 17). The data were adjusted based on the A-banddisplacement (displacements of the target A band were significantly different;P < 0.0001). The diagrams display the behavior observed by the A bands. Errorbars represent SEM. *P < 0.05; **P < 0.01; ***P < 0.001. Act., activation.

de Souza Leite et al. PNAS Early Edition | 3 of 6

BIOPH

YSICSAND

COMPU

TATIONALBIOLO

GY

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 7,

202

0

(Movie S5). A-band tracking showed that one side of the myo-fibril was displaced more than the other side (Fig. 3C). Sur-prisingly, the side that was displaced more was not always theone adjacent to the activated half-sarcomere. Instead, two be-haviors were observed: either (i) the A band did not move (andthe Z disk was pulled toward the A band) or (ii) the A band wasdisplaced toward the Z disk (Fig. 3D), explaining the asymmetryobserved when two half-sarcomeres are activated together. Forsimplicity, we will consider one sarcomere that has one half-sarcomere activated. If the A band is displaced from its zeroposition toward the Z disk of this half-sarcomere, it drags theother half-sarcomere (Fig. 3D, blue). If the A band does notchange its position, it pulls the Z disk and the half-sarcomeres(Fig. 3D, orange). In practice, both situations happen in ourexperiments. Note that these movements do not result in dif-ferential SLf between the two sides of the myofibril.

Sarcomeres and Myofibrils May Produce Different Forces. Sarco-meres comprising a myofibril have a similar structure and cross-section. Thus, if all sarcomeres were able to reach the same SLf,they should hypothetically produce similar active forces. To testthis hypothesis, we point-activated all sarcomeres within amyofibril (one at a time) and used the average displacement ofthe ±1 and ±2 A bands in relation to the target sarcomere to es-timate the individual sarcomere forces. Although an indirect mea-surement of the force in this situation, this procedure was based onour observation that the displacement of the A bands surroundingan activated sarcomere is very similar to the displacement of theneedles when we activated one sarcomere (mechanically isolatedfrom the myofibril). In general, we found forces varying from 8 to45 nN/μm2 when comparing all sarcomeres tested in differentmyofibrils (merged data in Fig. 4A), similar to the forces producedby isolated sarcomeres (dotted line in Fig. 1A). We observed a smallrange of forces among the different sarcomeres within a singlemyofibril caused by small variations in SLi (Fig. 4A). When myofi-brils were analyzed individually, it was possible to observe differ-ences in forces produced by sarcomeres in some myofibrils. Thesedifferences are likely related to different changes in length causedby nonuniform SLi or SLf.According to our data, sarcomeres reach similar SLf when

point-activated, but this activation affects other sarcomeres inseries. Therefore, during full myofibril activation, the dynamic workof sarcomeres in series and the differences in SLi among sarcomeresmay influence how a myofibril develops force to reach a steadystate. In such a state, myofibril internal motion is close to equilib-rium. If, during myofibril activation, every sarcomere shortened tothe same SLf, the force would be a function of the number ofsarcomeres in series and the SLi of each sarcomere. We, therefore,tested the hypothesis that a greater number of sarcomeres in seriesallows a greater shortening magnitude and hence, higher forces. Weperformed full activation of myofibrils with different numbers ofsarcomeres in series at an SLi of ∼2.8 μm. We first activatedmyofibrils with ∼25 sarcomeres, and after the first activation, wereduced the number of sarcomeres in series in the preparation byone-half by positioning the holding needle close to the center of themyofibril. We found that myofibrils with a greater number of sar-comeres produced higher forces, regardless of the average SLf (Fig.S4 A and D). These results may be attributed to cumulative short-ening when more sarcomeres are activated (i.e., a greater shorten-ing of the myofibril may be possible when more sarcomeres areactivated) (Fig. S4 B and C).To further test our hypothesis that more sarcomeres in series

allow for a greater shortening and to remove the potentialconfounding effect of myofibril damage during manipulation, wemeasured the forces produced by the first contraction of95 myofibrils and compared them with single sarcomeres. Weconfirmed that myofibrils with more sarcomeres in series short-ened more than myofibrils with fewer sarcomeres and produced

higher forces (Fig. 4 B and D). However, the increase in force inlonger myofibrils was not linearly related to the number of sar-comeres in series (Fig. 4B). When the force was normalized bythe number of sarcomeres in the myofibrils, they produced lessforce per sarcomere than single sarcomeres (Fig. 4C).To evaluate whether the final SL reached during the contractions

was in the zone where maximal force is produced, we performedseparate experiments to derive an FSL relation for the myofibrilsused in this study (Movie S6). We observed an FSL relation similarto that reported previously in other preparations (18), with a pla-teau between SLfs of ∼2.1 and 2.7 μm (Fig. 4E). Based on thesemeasurements, we evaluated if the SLf of the myofibrils reached

1 10 20 30 40 500

50

100

150

200

250

Number of Sarcomeres

Forc

e (n

N/

m2 )

1 10 20 30 40 500

20

40

60

Number of Sarcomeres

Forc

e pe

r Sar

com

ere

1 10 20 30 40 500

5

10

15

20

Number of Sarcomeres

Shor

teni

ng (

m)

1.0 1.5 2.0 2.5 3.0 3.5 4.0

8101112171825273540

Sarcomere Length ( m)

Num

ber o

f sar

com

eres

1.0 1.5 2.0 2.5 3.0 3.5 4.0Sarcomere Length ( m)

Nor

mal

ized

For

ce (%

)100

75

50

25

0

B C D

E F G

0 10 20 30 40 501.0

1.5

2.0

2.5

3.0

3.5

4.0

Number of Sarcomeres

Sarc

omer

e Le

ngth

(µm

)

Merged Data

0

10

20

30

40

50

Myofibril

Forc

e (n

N/

m2 )

Number of Sarcomeres Tested

A

Fig. 4. Force of several sarcomeres within a myofibril and full activation ofmyofibrils with different numbers of sarcomeres in series. The average dis-placement of ±1 and ±2 sarcomeres was multiplied by the needle’s stiffnessto estimate the force of the majority of sarcomeres in a myofibril. (A) Bluebox and whiskers plots represent the 25th percentile, the median (line), andthe 75th percentile of different myofibrils, and the error bars show theminimum and maximum values. The dots represent the force produced bysarcomeres within one myofibril. The dotted line shows the average forceconsidering all sarcomeres tested. The plot in Right merges the force of allpreparations. (B) Isometric force of myofibrils ranging from 4 to 48 sarco-meres in series compared with forces produced by one sarcomere. (C) Forcenormalized by the number of sarcomeres in series in a myofibril. (D) Re-lationship between the shortening magnitude during myofibril contractionand the number of sarcomeres in series in a myofibril. (E) FSL relationshipproduced during full activation of isolated myofibrils (n = 12). The force wasnormalized by the highest value for a given myofibril. Data were fit using athird-order polynomial regression (r2 = 0.83). The blue area represents the pla-teau of the FSL relationship. The green and yellow areas represent the ascendingand descending limbs of the FSL relationship, respectively. Note that the greenarea covers lengths in which filament overlap is between 70 and 90% of theoptimum overlap. The dotted line was created based on theoretical predictions(20) and previous experimental data (18). (F) Average SLf of all myofibrils plottedas a function of the number of sarcomeres and the amount (percentage) ofoptimum overlap between myosin and actin filaments according to the FSL re-lationship (n = 95). (G) Length of all sarcomeres (excluding the ones at the edges)of 10 randomly chosen myofibrils that contracted to an SLf within the plateauregions (blue) of the FSL relationship. Green and yellow areas represent theascending and descending limbs of the FSL, respectively. Note that the greenarea covers lengths in which filament overlap is 70–90% of the optimum. Graybox and whiskers plots represent a different myofibril with the 25th percentile,the median (line), and the 75th percentile; error bars show the minimum andmaximum values.

4 of 6 | www.pnas.org/cgi/doi/10.1073/pnas.1700615114 de Souza Leite et al.

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 7,

202

0

optimal filament overlap. In fact, the majority of the myofibrilscontracted to the plateau of the FSL relationship or in an areawithin 70% of the optimum filament overlap (Fig. 4E). There weresome sarcomeres in these myofibrils that reached a length longerthan those at the plateau of the FSL, likely because of an increasedSL nonuniformity (Fig. 4F). Based on these findings, we decided totest the hypothesis that a larger degree of nonuniformity in SLsalong the descending limb of the FSL leads to a decrease in thetotal force produced by the myofibrils.

Induced Nonuniformity/Deactivation Reduces Total Myofibril Force.Over the years, SL nonuniformity has been observed to happennaturally during muscle contraction. To test how nonuniformityof SLs affects the final force reached by the myofibrils, we usedthe microperfusion system to induce nonuniformity in activatedmyofibrils. We performed five different sets of experiments.First, we measured the myofibril force when three to five sar-comeres were activated before full myofibril activation (pre-activation group) and compared the results with those of regularfull activation (control) (Movie S7). Two segments of the samemyofibril were tested for preactivation. The protocol consisted ofthree cycles of full myofibril activation followed by two cycles ofpreactivation and a control contraction. When preactivated, themyofibril generated a small increase in force before a largerincrease caused by full activation (dashed square in Fig. 5A).Interestingly, we did not observe significant differences in thefinal forces measured, showing that initial nonuniformities did

not affect force development (Fig. 5A). Second, in another seriesof experiments, nonuniformity was induced either after con-traction or before activation by using a low-Ca2+ solution to relaxone sarcomere in a fully activated myofibril (Movie S8). Wenoticed a sudden reduction in the total myofibril force (Fig. 5B)that was reversed after removal of the microperfusion pipette.We then repeated this experiment but used a high-Ca2+ solutionwith 40 mM 2,3-butanedione monoxime (BDM), which led to asimilar reduction in force (Fig. 5C). A reduction in the totalmyofibril force was also observed by delaying the activation of asmall portion of the myofibril (approximately one sarcomere)during a contraction (Fig. 5D). For all of these situations, thetarget sarcomere was able to contract, although both sides of themyofibril were inducing sarcomere stretching. This finding sug-gests that one sarcomere is able to pull several activated sarco-meres and produce a force that is close to its maximum level.Finally, we used a high-strength ionic solution that depletes the

preparation from thick myosin filaments (19) (Movie S9). Extrac-tion of the A band results in a sarcomere with only the intermediatefilaments that is able to communicate with other sarcomeresthrough passive forces. However, A-band extraction may also leadto more compliant titin because of the loss of the interaction be-tween titin and myosin filaments. We compared forces developed bythe myofibrils before and after the extraction of the A band of onesarcomere. We found a significant reduction in the isometric forcecompared with experiments in which the A band was intact (Fig.5E), supporting the idea that the number of sarcomeres is an im-portant factor for total force development. The extraction of one Aband had a smaller effect on the total force in myofibrils with a largenumber of sarcomeres than in myofibrils with a small number ofsarcomeres (Fig. 5F). In longer myofibrils, force dropped after theadditional extraction of A bands. We performed a few experimentsin which we applied a small stretch after extraction of the A bandfrom myofibrils to evaluate if force could be recovered back to theoriginal values. However, the force remained lower than it wasbefore extraction of the A band (Fig. S5 and Movie S10).Overall, our data suggest that, during muscle contraction and force

development, events that create nonuniform distribution of SL alongmyofibrils have an impact on the final force that they produce.

DiscussionIn this study, we assessed intersarcomere dynamics in isolatedmyofibrils using microfluidic perfusion, a technique that allowsthe point activation of one (half-) sarcomere within a myofibrilwhile measuring the response of all other sarcomeres. We triggeredcross-bridge formation and contraction of target sarcomeres by lo-cally increasing the Ca2+ concentration. We showed that theshortening of an activated sarcomere leads to an internal regulationin the remaining sarcomeres in the myofibril through passive forces.This result shows that intersarcomere dynamics arise from the co-operative work of the contractile and elastic systems, enabling theaction of one sarcomere to be transmitted to the others.We found that single sarcomeres are able to contract as long as

there is myofilament overlap (20), reaching similar SLf independentof the SLi (with small differences when SLi is ∼3.6 μm). This be-havior is not observed during full myofibril activation (Figs. 1D and4G), where the SLf is highly nonuniform during activation (17).Although passive and active forces have been investigated sepa-rately (21, 22), we showed that the passive force component is re-sponsible for the interaction of sarcomeres in series, facilitating thetransmission of the force of one sarcomere to others.Recent studies have shown that titin controls important as-

pects of thick filament structure, playing an important role inmuscle activation (23, 24). In the future, it will be important toconsider how the activation of one sarcomere may be sensed by athick filament from the neighboring sarcomeres. Moreover, ithas been suggested that titin can assist muscle contraction atphysiological SLs using the stored elastic energy from unfolded

0

50

100

150

Forc

e (n

N/

m2 ) ***

***

Control Clamp Re-act.0

50

100

150

Forc

e (n

N/

m2 )

Control BDM React.

** **

0

50

100

150

Forc

e (n

N/

m2 )

Control One Extraction

**

0 5 60 70Time (s)

One Extraction

Control

0 20 40 60 80 100Time (s)

Control

One Sarcomere BDM

Re-Activation

0 20 40 60 80Time (s)

Control

One Sarcomere Relaxing (Deact.)

Re-Activation

0 10 20 30Time (s)

One Sarcomere Clamping

Re-Activation

A B

D

E

0 10 20 30 40 50

Time (s)

ControlPre-Activation

Control

Pre-Act

Control

Pre-Act

Control0

50

100

150

Forc

e (n

N/

m2 )

C

F

11 12 13 18 18 20 22 22 23 24 24 31 33 36 420.5

0.6

0.7

0.8

0.9

1.0

1.1

Number of Sarcomeres

% I

som

etric

For

ce

1 2 3 4 6 7 A-Bands Extracted

0

50

100

150

Forc

e (n

N/

m2 )

Control Deact. Re-act.

*** ***

Fig. 5. Controlling nonuniformity. (A) Preactivation of a short area of themyofibril does not affect the total force. A subarea of the myofibril wasactivated before full activation using microperfusion. In a typical experi-ment, the first contraction induced full myofibril contraction, and the secondcontraction was induced after preactivation. The increase in force caused bypreactivation is shown in the dashed square. Note that the total forces weresimilar in both contractions. (B) One sarcomere is deactivated during fullmyofibril contraction using relaxing solution (pCa2+ 9.0) within the micro-perfusion system (n = 21). (C) The same protocol used in A, but this time, thesolution (pCa2+ 4.5) used to deactivate the sarcomere contained 40 mM BDM(n = 8). (D) A small area of a myofibril was kept relaxed, whereas the rest ofthe myofibril was activated and contracted (n = 12). (E, Left) Myofibril forcebefore and after one A-band extraction using high-strength ionic solution inthe microperfusion system (n = 15). E, Right show typical experimentaltraces. (F) Extraction of one A band had a smaller effect or a lack of effect inlonger myofibrils. We performed additional A-band extractions in threemyofibrils (with 31, 33, and 36 sarcomeres), which reduced the force sig-nificantly. The straight line represents the normalized isometric force beforeA-band extraction. Error bars represent SEM. **P < 0.01; ***P < 0.001.

de Souza Leite et al. PNAS Early Edition | 5 of 6

BIOPH

YSICSAND

COMPU

TATIONALBIOLO

GY

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 7,

202

0

Ig domains (25). The importance of the elastic system was shownin our experiments, in which a few sarcomeres were activatedbefore full myofibril activation (Fig. 5A). There was not a dif-ference in force compared with myofibrils that were not pre-activated, showing that the passive force components hold thenonactivated area of the myofibril at the optimum length forcontraction.Our microfluidic perfusion system allowed the control of a

long known phenomenon of muscle contraction, the nonuniformlengths of sarcomeres during activation (26, 27). From the pro-posal of the sliding filament theory (1, 2), nonuniform behaviorof sarcomeres and its effects in force production during con-traction have puzzled scientists. We found that, in short myofi-brils, enabling the relaxation of one sarcomere either during orbefore contraction leads to a reduction in the total force (Fig. 5 B–D). This result can be explained by the reduction in the totalshortening caused by the relaxed sarcomeres. The same reductionwas observed when we extracted the A bands from myofibrils (Fig.5E). However, this reduction in force was attenuated in myofibrilswith more sarcomeres in series (Fig. 5F). It is possible that the lackof one sarcomere was compensated for by an increase in shorteningof other sarcomeres. Such a mechanism may have important im-plications, because the structure of the muscle cell may be designedto balance the presence of weaker or damaged sarcomeres alongthe myofiber.This study uses a microfluidic perfusion to show the impor-

tance of intersarcomere dynamics and the cooperative work ofthe contractile and elastic proteins in the myofibril, openingpossibilities for new venues of investigation in muscle biophysics.

MethodsExperimental Solutions and Experimental System. Myofibril preparation andimaging (SI Text) were done as described before (28). The protocol wasapproved by the McGill University Animal Care Committee and compliedwith the guidelines of the Canadian Council on Animal Care. Rigor solutionwas 50 mM Tris, 100 mM KCl, 4 mM MgCl2, and 10 mM EGTA, pH 7.0.Relaxing solution was 70 mM KCl, 20 mM imidazole, 5 mMMgCl2, 5 mM ATP,14.5 mM creatine phosphate, 7 mM EGTA, and pCa2+ (equal to –log[Ca2+]) 9.0,pH 7.0. Activation solution was 50 mM KCl, 20 mM imidazole, 5 mM MgCl2,5 mM ATP, 14.5 mM creatine phosphate, 7.2 mM EGTA, and pCa2+ 4.5, pH

7.0. We used two precalibrated needles to pierce the myofibrils parallel tothe Z disks (28). The stiffness of all needles (43.40 ± 0.76 nN/μm) was mea-sured using the same atomic force cantilever. All procedures were imagedusing a Hamamatsu Orca-ER digital camera. Movies were recorded at43.3 frames per 1 s. Myofibrils were pierced and lifted from the coverslip,and the SLi was adjusted. Sarcomeres within myofibrils were point-perfusedby a microperfusion system, in which glass micropipettes filled with thedesired solution were controlled by a pressure system and micromanipula-tors (NT88-V3; Nikon). To avoid diffusion of the microperfusion solution toundesired areas, myofibrils were constantly perfused by a larger perfusionflow that controlled the remaining sarcomeres in the preparation. Micro-pipettes and needles were pulled using a capillary puller (Kopf Model 720;David Kopf Instruments). Micropipette tip diameter was adjusted to 1 μm orless (<0.5 μm for half-sarcomeres) to keep it smaller than the sarcomere Abands. All experimental procedures were recorded and analyzed using theTrackMate plugin for ImageJ software (NIH). Absolute sarcomere displace-ment was defined as the difference between the zero position (relaxedstate) and the position of the tracked A band during the experiment. Theexperiments comparing the active force development of myofibrils beforeand after reduction in the number of sarcomeres in series were performedusing atomic force microscopy as previously described (28). Because theforces produced by myofibrils are proportional to their cross-sectional areasand because we wanted to compare different conditions irrespective of thisvariable, all measured forces were normalized and reported in units ofstress, assuming a circular geometry for the myofibril (nanonewtonsper micrometer2).

Statistical Analyses. We used a two-way ANOVA for repeated measures tocompare displacement of A bands. For nonuniformity experiments, we usedone-way ANOVA for repeated measures. We used a Student–Newman–Keulstest when significant changes were observed. A level of significance of P ≤0.05 was set for all analyses. Values are presented as means ± SEM.

Mechanical Model. We developed a mechanical model of a half-myofibril tobetter interpret our experimental data (SI Text, Figs. S6–S8, Table S1, andMovie S11) using the framework of Campbell (29).

ACKNOWLEDGMENTS. We thank Dr. Kenneth S. Campbell for helpful discus-sions during the writing of this paper. This study was funded by the CanadianInstitutes for Health Research and the Natural Science and Engineering ResearchCouncil of Canada. F.d.S.L. was a recipient of a scholarship from the NationalCounsel of Technological and Scientific Development (CNPq, Brazil). D.E.R. is aCanada Research Chair (Tier I) in Muscle Biophysics.

1. Huxley AF, Niedergerke R (1954) Structural changes in muscle during contraction;interference microscopy of living muscle fibres. Nature 173:971–973.

2. Huxley H, Hanson J (1954) Changes in the cross-striations of muscle during contractionand stretch and their structural interpretation. Nature 173:973–976.

3. Gilev VP (1962) A study of myofibril sarcomere structure during contraction. J Cell Biol12:135–147.

4. Maruyama K, Kimura S, Kuroda M, Handa S (1977) Connectin, an elastic protein ofmuscle. Its abundance in cardiac myofibrils. J Biochem 82:347–350.

5. Wang K, McClure J, Tu A (1979) Titin: Major myofibrillar components of striatedmuscle. Proc Natl Acad Sci USA 76:3698–3702.

6. Agarkova I, Perriard JC (2005) The M-band: An elastic web that crosslinks thick fila-ments in the center of the sarcomere. Trends Cell Biol 15:477–485.

7. Gregorio CC, et al. (1998) The NH2 terminus of titin spans the Z-disc: Its interactionwith a novel 19-kD ligand (T-cap) is required for sarcomeric integrity. J Cell Biol 143:1013–1027.

8. Telley IA, et al. (2006) Dynamic behaviour of half-sarcomeres during and after stretchin activated rabbit psoas myofibrils: Sarcomere asymmetry but no ‘sarcomere pop-ping’. J Physiol 573:173–185.

9. Shimamoto Y, Suzuki M, Mikhailenko SV, Yasuda K, Ishiwata S (2009) Inter-sarcomerecoordination in muscle revealed through individual sarcomere response to quickstretch. Proc Natl Acad Sci USA 106:11954–11959.

10. Rassier DE, Herzog W, Pollack GH (2003) Dynamics of individual sarcomeres duringand after stretch in activated single myofibrils. Proc Biol Sci 270:1735–1740.

11. Telley IA, Denoth J, Ranatunga KW (2003) Inter-sarcomere dynamics in muscle fibres.A neglected subject? Adv Exp Med Biol 538:481–500, discussion 500.

12. Rassier DE, Pavlov I (2012) Force produced by isolated sarcomeres and half-sarcomeresafter an imposed stretch. Am J Physiol Cell Physiol 302:C240–C248.

13. Horowits R (1992) Passive force generation and titin isoforms in mammalian skeletalmuscle. Biophys J 61:392–398.

14. Wang K, Ramirez-Mitchell R (1983) A network of transverse and longitudinal in-termediate filaments is associated with sarcomeres of adult vertebrate skeletalmuscle. J Cell Biol 96:562–570.

15. Tskhovrebova L, Trinick J (2003) Titin: Properties and family relationships. Nat Rev MolCell Biol 4:679–689.

16. Horowits R, Kempner ES, Bisher ME, Podolsky RJ (1986) A physiological role for titin

and nebulin in skeletal muscle. Nature 323:160–164.17. Telley IA, Denoth J, Stüssi E, Pfitzer G, Stehle R (2006) Half-sarcomere dynamics in

myofibrils during activation and relaxation studied by tracking fluorescent markers.

Biophys J 90:514–530.18. Pavlov I, Novinger R, Rassier DE (2009) The mechanical behavior of individual sarco-

meres of myofibrils isolated from rabbit psoas muscle. Am J Physiol Cell Physiol 297:

C1211–C1219.19. Maruyama K, Natori R, Nonomura Y (1976) New elastic protein from muscle. Nature

262:58–60.20. Sosa H, Popp D, Ouyang G, Huxley HE (1994) Ultrastructure of skeletal muscle fibers

studied by a plunge quick freezing method: Myofilament lengths. Biophys J 67:

283–292.21. Linke WA, Ivemeyer M, Mundel P, Stockmeier MR, Kolmerer B (1998) Nature of PEVK-

titin elasticity in skeletal muscle. Proc Natl Acad Sci USA 95:8052–8057.22. Gordon AM, Huxley AF, Julian FJ (1966) The variation in isometric tension with sar-

comere length in vertebrate muscle fibres. J Physiol 184:170–192.23. Linari M, et al. (2015) Force generation by skeletal muscle is controlled by mecha-

nosensing in myosin filaments. Nature 528:276–279.24. Fusi L, Brunello E, Yan Z, Irving M (2016) Thick filament mechano-sensing is a calcium-

independent regulatory mechanism in skeletal muscle. Nat Commun 7:13281.25. Rivas-Pardo JA, et al. (2016) Work done by titin protein folding assists muscle con-

traction. Cell Reports 14:1339–1347.26. ter Keurs HE, Iwazumi T, Pollack GH (1978) The sarcomere length-tension relation in

skeletal muscle. J Gen Physiol 72:565–592.27. Granzier HL, Pollack GH (1990) The descending limb of the force-sarcomere length

relation of the frog revisited. J Physiol 421:595–615.28. Leite FS, et al. (2016) Reduced passive force in skeletal muscles lacking protein argi-

nylation. Am J Physiol Cell Physiol 310:C127–C135.29. Campbell KS (2009) Interactions between connected half-sarcomeres produce emer-

gent mechanical behavior in a mathematical model of muscle. PLOS Comput Biol 5:

e1000560.

6 of 6 | www.pnas.org/cgi/doi/10.1073/pnas.1700615114 de Souza Leite et al.

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 7,

202

0