Properties of extensor digitorum longus muscle and skinned fibers from adult and aged male and female Actn3 knockout mice

PROPERTIES OF EXTENSOR DIGITORUM LONGUS MUSCLE ANDSKINNED FIBERS FROM ADULT AND AGED MALEAND FEMALE ACTN3 KNOCKOUT MICESTEPHEN CHAN, BSc,1 JANE T. SETO, PhD,2,3 PETER J. HOUWELING, PhD,2 NAN YANG, PhD,2,3

KATHRYN N. NORTH, PhD, MD,2,3 and STEWART I. HEAD, PhD1

1 School of Medical Sciences, University of New South Wales, Sydney, New South Wales, Australia2 Institute for Neuroscience and Muscle Research, The Children’s Hospital at Westmead, Sydney, New South Wales, Australia3Discipline of Paediatrics and Child Health, Faculty of Medicine, University of Sydney, Sydney, 2006 New South Wales, Australia

Accepted 27 April 2010

ABSTRACT: Absence of a-actinin-3, encoded by the ACTN3‘‘speed gene,’’ is associated with poorer sprinting performance inathletes and a slowing of relaxation in fast-twitch muscles ofActn3 knockout (KO) mice. Our first aim was to investigate, atthe individual-fiber level, possible mechanisms for this slowedrelaxation. Our second aim was to characterize the contractileproperties of whole extensor digitorum longus (EDL) musclesfrom KO mice by age and gender. We examined caffeine-induced Ca2þ release in mechanically skinned EDL fibers fromKO mice, and measured isolated whole EDL contractile proper-ties. The sarcoplasmic reticulum of KO muscle fibers loadedCa2þ more slowly than that of wild-types (WTs). Whole KO EDLmuscles had longer twitch and tetanus relaxation times thanWTs, and reduced mass and cross-sectional area. These effectsoccurred in both male and female mice, but they diminished withage. These changes in KO muscles and fibers help to explainthe effects of a-actinin-3 deficiency observed in athletes.

Muscle Nerve 43: 37–48, 2011

The a-actinins are a group of actin-binding pro-teins found in the Z-disks of skeletal muscle sarco-meres. Two isoforms are found in the Z-disk: a-acti-nin-2, which occurs in all muscle fibers; anda-actinin-3, which is restricted mainly to fast glyco-lytic fibers.1

An estimated 18% of individuals worldwidecompletely lack a-actinin-3, due to homozygosityfor a common polymorphism in the ACTN3 gene.2

This gene has become known colloquially as the‘‘gene for speed.’’ a-Actinin-3 deficiency is not asso-ciated with any disease phenotype. However, elitesprint and power athletes have a reduced fre-quency of a-actinin-3 deficiency,3–6 whereas thereis a tendency toward a higher frequency of a-acti-nin-3 deficiency among elite endurance athletes.3,7

In non-athlete populations, a-actinin-3 deficiencyhas been associated with reduced muscle strength8

and poorer sprinting performance.9 Taken overall,these data indicate that a lack of a-actinin-3 is det-rimental to sprint and power performance butbeneficial for endurance activities.

We have generated an Actn3 knockout mouse,2

which displays a phenotype consistent with thatobserved in humans. Knockout mice were able torun 33% further than wild-type mice on a motor-ized treadmill endurance test,2 but had lower gripstrength and lower muscle weights than the wild-types.10 The Actn3 knockout mouse is thus anappropriate model for examining the effects of a-actinin-3 deficiency.

Our earlier studies on the Actn3 knockoutmouse provide insights into the mechanisms bywhich a-actinin-3 deficiency might lead to reducedsprinting performance and enhanced enduranceperformance. We have found that knockoutmuscles display characteristics that are more slowtwitch and oxidative than wild-type muscles. Com-pared with wild-type (WT) muscles, knockout(KO) muscles have slower twitch half-relaxationtimes, smaller fast-fiber diameters, higher activityof oxidative enzymes, lower activity of anaerobicenzymes, and quicker recovery of force followingfatiguing stimulation.10,11 Such changes would be adisadvantage in activities that require repeatedrapid contractions, such as sprinting,12 but theywould be beneficial for activities that depend onaerobic metabolism, such as endurance sports.

These findings suggest that a-actinin-3 may playa role in the development of fast-twitch, glycolyticproperties in a muscle fiber and that, in its ab-sence, slower twitch, more oxidative propertiesmay develop. At a subcellular level many factorscontribute to the characterization of a fiber as fasttwitch or slow twitch. These include: (i) the sensi-tivity of the contractile apparatus to Ca2þ; and (ii)the rate of release and reuptake of Ca2þ by the sar-coplasmic reticulum (SR).13 Our first aim in thisstudy was to find out what changes occur at thelevel of the contractile proteins and SR whena-actinin-3 is absent. Using individual fibers fromthe extensor digitorum longus (EDL) muscles ofActn3 KO and WT mice, we employed the skinnedfiber technique to examine: (i) the Ca2þ loadingproperties of the SR; and (ii) the Ca2þ and Sr2þ

activation characteristics of the contractile filaments.The skinned fiber technique is used widely for

Abbreviations: ACTN3, a-actinin-3 gene; ANOVA, analysis of variance;EDL, extensor digitorum longus; EGTA, ethylene-glycol-bis[b-aminoethylether] N,N,N0,N0-tetraacetic acid; HDTA, hexamethylenediamine-N,N,N0,N0-tetraacetic acid; HEPES, N-[2-hydroxyethyl] piperazine-N0-[2-ethanesulfonic acid]; KO, knockout; Lo, optimum length; SR, sarcoplasmicreticulum; WT, wild-type

Correspondence to: S.I. Head; e-mail: [email protected]

VC 2010 Wiley Periodicals, Inc.Published online 30 September 2010 in Wiley Online Library(wileyonlinelibrary.com). DOI 10.1002/mus.21778

Key words: a-actinin-3, exercise, sarcoplasmic reticulum, skeletal muscle,skinned fiber

Properties of Actn3 KO Muscle MUSCLE & NERVE January 2011 37

elucidating mechanisms at the level of the con-tractile proteins and SR.14

As the Actn3 KO mouse is a newly developedmodel of a-actinin-3 deficiency, it is useful to char-acterize the contractile properties of whole fast-twitch muscles from this mouse. This was previ-ously done in an adult male cohort11; in this studywe were also interested in determining whetherthese characteristics vary with age and gender.Thus, our second aim was to characterize the con-tractile properties of whole EDL muscles frommale and female Actn3 KO mice in two age groups:adult (2–6 months old) and aged (19–22 monthsold).

METHODS

Animals. Use of animals was approved by the Ani-mal Care and Ethics Committees of the Children’sMedical Research Institute and Children’s Hospitalat Westmead and the University of New SouthWales. Mice used in the whole muscle experimentswere in two age groups: adult mice (2–6 monthsold) and aged mice (19–22 months old). Adultmice consisted of 24 males (12 WTs, 12 KOs) and15 females (8 WTs, 7 KOs). Aged mice consistedof 14 males (7 WTs, 7 KOs) and 12 females(6 WTs, 6 KOs). Nine adult male mice (4 WTs, 5KOs) were used for the skinned fiber experiments.

Fiber Preparation—Skinned Fiber Experiments. Allanimals were anesthetized with halothane andkilled by cervical dislocation. The EDL muscle wasdissected from the hindlimb. Individual fibers wereisolated and mechanically skinned in paraffin oiland mounted on a force transducer to monitor iso-metric force. The length and diameter of eachfiber were measured at slack length, and the fiberwas then stretched by 20% to maximize the forceresponses.

Ca21 Loading of SR. The fiber was first depleted ofCa2þ by exposure for 2 min to a potassium–hexam-ethylenediamine-N,N,N0,N0-tetraacetic acid (Kþ-HDTA) solution containing low Mg2þ (0.25 mM)and 30 mM caffeine, to maximally release Ca2þ

from the SR, and 0.25 mM ethylene-glycol-bis[b-aminoethyl ether] N,N,N0,N0-tetraacetic acid(EGTA), to chelate all released Ca2þ and preventSR Ca2þ reaccumulation. The fiber was thenreloaded with Ca2þ for predetermined periods oftime by exposure to a highly buffered Ca2þ solu-tion (pCa 6.57) made by combining solutions Aand B at a ratio of 0.85:1.15 (see Contractile Appa-ratus subsection for compositions of solutions Aand B). Loading was rapidly terminated at the endof each loading period by brief exposure ( �1–2 s)to solution A, after which the fiber was washed

(for 6 s) in a Kþ-HDTA solution to remove excessEGTA. The fiber was then re-exposed to the caf-feine solution (see earlier), and the force responsewas measured. The area under the force-responsecurve was used as a measure of the amount ofCa2þ released, and hence of the amount of Ca2þ

loaded by the SR during the loading period.14,15

The amount of Ca2þ loading for each loading pe-riod was expressed as a percentage of maximum,and an exponential curve was fitted to these data(one-phase association model; GraphPad Prism sta-tistical software). Before exposure to the caffeinesolution, the fiber was incubated for 30 s in a Kþ-HDTA solution containing 0.25 mM EGTA to allowtime for the EGTA to equilibrate within the fiber.

Contractile Apparatus. After the Ca2þ loadingexperiments, the fiber was placed for 10 min in so-lution A (see later) with 2% Triton X-100 added tochemically skin all remaining membranous cell ele-ments. The fiber was then exposed to a series ofsolutions of different free Ca2þ concentrations.The strongly buffered Ca2þ solutions were pre-pared by mixing specific proportions of EGTA-con-taining solution (solution A) and Ca-EGTA–con-taining solution (solution B). Solution A contained117 mM Kþ, 36 mM Naþ, 8 mM adenosine triphos-phate (ATP, total), 1 mM free Mg2þ, 10 mM crea-tine phosphate, 50 mM EGTA (total), 60 mM N-[2-hydroxyethyl] piperazine-N0-[2-ethanesulfonic acid](HEPES), and 1 mM NaN3 (pH 7.10). Solution Bwas similar to solution A, with the exception thatthe EGTA and Ca-EGTA concentrations of solutionB were 0.3 and 49.7 mM, respectively. The freeCa2þ concentrations of the solutions were calcu-lated using a Kapparent for EGTA of 4.78 � 106

M�1.16 Maximal force was determined by exposureto solution B, containing a free Ca2þ concentra-tion of 3.5 � 10�5 M. Force was returned to base-line after maximal activation by exposure to solu-tion A. The plateaus of the force responses elicitedby exposure to solutions of increasing free Ca2þ

concentration were expressed as a percentage ofmaximum Ca2þ-activated force and plotted as afunction of pCa. The fibers were then similarlyactivated by exposure to a series of Sr2þ solutionsmade by combining solution A with a solution con-taining 9.5 mM EGTA and 40.0 mM Sr-EGTA. Theresulting forces were similarly plotted against pSr.The force–pCa data were fitted with Hill curvesusing the following equation:

F ¼ 1

1þ 10nCaðpCa� pCa50Þwhere F is the force (as a proportion of maximumforce) generated at a particular pCa, nCa is the Hillcoefficient for Ca2þ, and pCa50 is the pCa value at

38 Properties of Actn3 KO Muscle MUSCLE & NERVE January 2011

which force is half-maximum. An analogous equa-tion was used to fit the force–pSr data. In examin-ing the properties of the contractile apparatus, thefollowing parameters were of interest: (i) nCa andnSr, indicating the cooperativity of Ca2þ and Sr2þ

binding sites; (ii) pCa50 and pSr50, indicating thesensitivity to Ca2þ and Sr2þ; and (iii) pCa50–pSr50,indicating the relative sensitivity to Ca2þ and Sr2þ.

Muscle Preparation—Whole Muscle Experiments. TheEDL muscle was dissected as described previously,then tied by its tendons to a force transducer(World Precision Instruments, Fort 10) at one endand a fixed metal hook at the other using silksuture (Deknatel 6.0). It was placed in a bath con-tinuously superfused with Krebs solution with acomposition of 4.75 mM KCl, 118 mM NaCl, 1.18mM KH2PO4, 1.18 mM MgSO4, 24.8 mM NaHCO3,2.5 mM CaCl2, and 10 mM glucose, with 0.1% fetalcalf serum and continuously bubbled with 95%O2–5% CO2 to maintain pH at 7.4. The musclewas stimulated by delivering a supramaximal cur-rent between two parallel platinum electrodes usingan electrical stimulator (A-M Systems). At the startof the experiment, the muscle was set to the opti-mum length, Lo, that produced maximum twitchforce. All experiments were conducted at room tem-perature ( �22� to 24�C).

Twitch Relaxation. The muscle was stimulated witha supramaximal pulse of 1-ms duration, and theresulting twitch was recorded. The twitch datawere smoothed by averaging the raw data over 2.5-ms intervals, and from the resulting smoothed datathe half-relaxation time (time taken to relax tohalf of peak twitch force) was obtained. Somemuscles were also subjected to a strenuous fatigueprotocol, after which another twitch was recordedand measured as described previously. The fatigueprotocol consisted of a 1-s, 100-HZ tetanus every2 s over a period of 30 s.

Force–Frequency Curve. A force–frequency curvewas obtained by delivering 500-ms stimuli of differ-ent frequencies (2, 15, 25, 37.5, 50, 75, 100, 125,and 150 HZ), and measuring the force producedat each frequency of stimulation. A 30-s rest wasallowed between each frequency. A curve relatingthe muscle force, P, to the stimulation frequency,f, was fitted to these data. The curve had the fol-lowing equation17:

P ¼ Pmin þ Pmax � Pmin

1þ Kf

f

� �h

The values of r2 for the fitting procedure werenever lower than 99.3%. From the fitted parame-ters of the curve, the following contractile proper-

ties were obtained: maximum force (Pmax), half-fre-quency (Kf), Hill coefficient (h), and twitch-to-tetanus ratio (Pmin/Pmax). Some muscles were alsosubjected to a strenuous fatigue protocol (asdescribed in the Twitch Relaxation subsection),after which a second force–frequency curve wasobtained.

Tetanus Relaxation. Tetanus relaxation consists ofa slow linear phase followed by a fast exponentialphase. The linear phase is the easier to interpret,as relaxation during this phase is homogeneousalong the fiber, whereas in the exponential phasesome parts of the fiber are lengthening and othersare shortening.18 In our measurements, the startof the linear phase was defined to be the point atwhich force began to fall following cessation ofstimulation. Linear regression was then performedbetween this point and all subsequent points. Thepoint at which the linear regression began to yieldr2 < 98.5% was defined to be the end of the linearphase. We used the duration of this linear phaseand the rate of force decline over this phase asmeasures of the rate of relaxation following a teta-nus. The tetanus analyzed was the 125-HZ tetanusfrom the force–frequency curve.

Mass and Cross-Sectional Area. At the end of thewhole muscle experiment, the muscle was removedfrom the bath. The tendons were trimmed, andthe muscle was lightly blotted on filter paper andweighed. An estimate of the cross-sectional areawas obtained by dividing the muscle’s mass by theproduct of its Lo and the density of mammalianmuscle (1.06 mg/mm3).

Statistical Analyses. Data are presented as mean6 SEM. For the skinned fiber data, we used two-tailed t-tests at a significance level of 5%. For thewhole muscle data, we compared Actn3 KO vs. WTin adult and aged animals. Hence, for all compari-sons we used two-way analysis of variance (ANOVA)in which genotype (WT or KO) and age (adult oraged) were the two factors. Comparison of WT vs.KO in each age group was made using Bonferro-ni’s posttests with an overall significance level of5%. All statistical tests and curve fitting were per-formed using standard statistical software (Graph-Pad Prism, version 5 for Windows; GraphPad Soft-ware, San Diego, California).

RESULTS

Ca21 Loading of the Sarcoplasmic Reticulum. In pre-vious studies of the Actn3 KO mouse we observedthat whole fast-twitch muscles from the knockoutmouse showed a shift toward having more slow-twitch characteristics.10,11 At the individual fiberlevel, one distinguishing feature between fast-twitch and slow-twitch fibers is the rate of Ca2þ


uptake by the SR.13 Hence, we compared SR Ca2þ

uptake in Actn3 KO and WT fibers to determinewhether any differences might be contributing tothe slower twitch phenotype of whole KO muscle.We used an established procedure for investigatingSR Ca2þ loading, where the SR of mechanicallyskinned fibers is loaded with Ca2þ for a predeter-mined period of time and then depleted of allreleasable Ca2þ with caffeine.14,15

Figure 1A shows an example of the process. Atthe outset, we depleted the SR of all its endoge-nous Ca2þ by exposing the fiber to a caffeine–EGTA solution. The first curve shows the forceresponse during this exposure. It can be seen thatthe caffeine–EGTA solution fully empties the SR ofall Ca2þ, because re-exposure of the fiber to thecaffeine–EGTA solution without any interveningexposure to the Ca2þ-loading solution results in noforce response. After we depleted the SR of allendogenous Ca2þ, we carried out repeating cyclesof load and release, where the SR was loaded withCa2þ by exposing the fiber to a low-[Ca2þ] solutionfor a known length of time, then fully depleted ofCa2þ by exposing the fiber to the caffeine–EGTAsolution. The subsequent curves show the forceresponses during exposure to the caffeine–EGTAsolution. The area under the curve of the forceresponse was used as a measure of the amount ofCa2þ released and hence of the amount of Ca2þ

loaded into the SR during the time of exposure tothe low-[Ca2þ] solution.

Figure 1B shows the results for our sample offibers. Each point on the graph shows the amountof Ca2þ loaded into the SR during a particularloading duration, expressed as a percentage ofmaximum Ca2þ loaded. It can be seen that maxi-mum loading is achieved by 10 s in both WT and

Actn3 KO fibers. For any particular loading timeshorter than this, the SR in KO fibers is able toload less Ca2þ than the SR in WT fibers, indicatingthat Ca2þ uptake by the SR in a-actinin-3–deficientfibers is slower than in WT fibers.

To obtain a quantitative indication of this dif-ference in rates of SR Ca2þ uptake, exponentialcurves were fitted to these data. The time constant(tau) was 2.3 s in WT fibers and 7.1 s in KO fibers.

Another factor that could affect how quickly afiber relaxes is the rate of dissociation of Ca2þ

from troponin C. As an estimate of the relativerates at which Ca2þ dissociates from the myofila-ments in KO and WT fibers, we compared thehalf-relaxation times of their force-response curvesfollowing maximum loading. In terms of Figure1A, this is the half-relaxation time of the force-response curve after loading for 10 s. The rationalehere is that the EGTA in the caffeine solutionwould chelate all the Ca2þ released from the SR,and the rate of relaxation would reflect the rate ofCa2þ removal from the myofilaments. We foundthat, on average, the half-relaxation times in Actn3KO fibers were 40% longer than in WT fibers, sug-gesting that, in a-actinin-3–deficient fibers, thekinetics of Ca2þ dissociation from the myofila-ments is slower than in WT fibers.

Ca21 and Sr21 Activation Characteristics of Individual

Skinned Fibers. Another distinguishing featurebetween fast-twitch and slow-twitch fibers is theCa2þ sensitivity of the contractile proteins.13 Toinvestigate the possibility that a-actinin-3 deficiencymay change the properties of the contractile pro-teins and contribute to a slower twitch phenotypein whole muscle, we examined the Ca2þ and Sr2þ

FIGURE 1. Ca2þ loading of the SR. (A) Representative force responses produced in a KO fiber when it is exposed to a caffeine–

EGTA solution to empty the SR of all releasable Ca2þ. The SR was first emptied of all endogenous Ca2þ (‘‘endogenous release’’

curve). The subsequent curves show the force responses after loading the SR with Ca2þ for different periods of time. The area under

each curve gives an indication of the amount of Ca2þ released, and hence of the amount loaded during the loading period. (B) Amount

of Ca2þ loaded by the SR over different loading times, expressed as a percentage of maximum loading. The data points have been fit-

ted with exponential curves. It is apparent that the SR in KOs loads Ca2þ more slowly than the SR in WTs.


activation characteristics of individual skinnedfibers.

After the SR loading experiments, muscle fiberswere chemically skinned with Triton X-100 andthen exposed to a series of highly buffered Ca2þ

solutions of increasing concentrations. The forceproduced by the fiber in this series of solutions isshown in the left of Figure 2A for a representativefiber. Forces were then expressed as a percentageof maximum and a force–pCa curve was fitted tothe data points. The force–pCa curves for WTs andKOs are the left pair of curves shown in Figure 2B.Force–pSr curves were also produced for each fiberby exposure to a series of Sr2þ solutions. The forceproduced in this series of Sr2þ solutions is shownat the right of Figure 2A for a representative fiber,and the resulting force–pSr curves for WTs andKOs are the right pair of curves shown in Figure2B. From the fitted force–pCa and force–pSrcurves, it was possible to determine the followingparameters: (i) nCa and nSr, the cooperativity ofCa2þ and Sr2þ binding sites; (ii) pCa50 and pSr50,the sensitivity to Ca2þ and Sr2þ; and (iii) pCa50–pSr50, the relative sensitivity to Ca2þ and Sr2þ.

The values of these parameters in WT and KOEDL fibers are shown in Figure 2C. KO fibers had

significantly higher nCa and nSr values than WTs,indicative of a higher cooperativity of the contract-ile filaments in Actn3 KOs. There was no differencebetween WTs and KOs in their sensitivity to Ca2þ,as measured by the pCa50. There was a statisticallysignificant difference in pSr50 between WTs andKOs, but this difference was extremely small.There was no difference between WTs and KOs intheir pCa50–pSr50, which indicates the lack of agross difference in myosin heavy chain isoformcompositions.

Specific forces upon maximal Ca2þ activationwere 302 6 48 mN/mm2 in Actn3 KO fibers and314 6 50 mN/mm2 in WT fibers. These valueswere not significantly different, and they are con-sistent with maximum specific forces of whole EDLmuscles in our study (327 6 10 mN/mm2 for KOs,302 6 7.6 mN/mm2 for WTs, not significantlydifferent).

Frequency of Myofibrillar Oscillations. At submaxi-mal levels of activation by Ca2þ or Sr2þ, skinnedfibers exhibit characteristic oscillations in force.19–21

We also observed these oscillations in our skinnedfibers during the generation of the force–pCa andforce–pSr curves. These oscillations appear to be a

FIGURE 2. Ca2þ and Sr2þ activation characteristics of individual skinned fibers. (A) Representative force responses in an individual

skinned fiber when it is exposed to a series of solutions of increasing [Ca2þ] (left) and increasing [Sr2þ] (right). [Ca2þ] and [Sr2þ] areindicated under the curves (expressed as pCa and pSr). The force–pCa curves and force–pSr curves for WTs and KOs are shown in

(B). The left pair of curves is for pCa, and the right pair is for pSr. The values of parameters derived from these curves are shown in

(C). KO fibers had significantly higher nCa and nSr values than WTs. There was also a statistically significant difference in their pSr50values, but this difference was extremely small. There was no difference between WTs and KOs in their pCa50–pSr50 (sample sizes:

17 WT fibers, 18 KO fibers; $$0.001 < P < 0.01, $0.01 < P < 0.05).


property of the contractile apparatus and reflect thenature of the interaction between actin filamentsand the myosin heads.22 Hence, it is possible thatsubtle alterations to actin–myosin affinity in the ab-sence of a-actinin-3 might show up as changes inthe frequency of myofibrillar oscillations.

We thus analyzed all myofibrillar oscillationsobserved in our skinned fiber preparations whenthey were activated by Ca2þ and Sr2þ. Figure 3Ashows examples of oscillations observed in a WTfiber and in a KO fiber. In Figure 3B, the fre-quency of oscillation is plotted against the percent-age of maximally activated force at which the oscil-lation occurred. Oscillations typically occurred at<40% of maximum activation, although someoccurred at up to 60% of maximum activation. Itcan be seen that higher frequency oscillations gen-erally occurred in only the KO fibers. Of thosecases where oscillation frequency exceeded 1.25HZ in the Ca2þ solutions, only 2 were in WT fibers,whereas 12 were in KO fibers. In the Sr2þ solu-tions, only KO fibers displayed oscillation frequen-cies exceeding 1.25 HZ. These data suggest thatfibers in Actn3 KO muscles may have altered actin–myosin affinity, giving rise to higher oscillation fre-quencies that do not occur in the fibers of WTmuscles.

Mass and Cross-Sectional Area. As the Actn3 KOmouse is a newly developed model of a-actinin-3deficiency, we sought to characterize the contractile

properties of whole EDL muscle from both maleand female mice in two age groups: adult (2–6months old) and aged (19–22 months old). Fig-ure 4 shows the mass and cross-sectional area ofEDL muscles from WT and Actn3 KO animals. Themass and cross-sectional area in adult male micewas 9% lower in KOs than in WTs (0.001 < P <0.01). In adult female mice, KO muscles had 11%lower mass and 13% smaller cross-sectional areathan WTs (0.001 < P < 0.01). In aged animals,however, these differences were no longer appa-rent. There were no significant differencesbetween Actn3 KOs and WTs in either aged malemice or aged female mice. Hence, a-actinin-3 defi-ciency is associated with a reduction in EDL mus-cle mass and cross-sectional area in both malesand females, but only in adult animals.

Twitch Relaxation. Figure 5A shows the timecourse of relaxation during an isometric twitch,both in unfatigued EDL muscles and in musclesfatigued by repetitive stimulation (see Methods forfatigue protocol). Twitch-relaxation data for indi-vidual EDL muscles were expressed as a percent-age of peak twitch force and aggregated into singlecurves for each age/gender group. It can be seenthat, in all cases, except for unfatigued musclesfrom adult female mice, the relaxation curvesfor Actn3 KOs lie to the right of those for WTs,suggesting that Actn3 KO muscles relax at aslower rate. An example of this slowing in a KO

FIGURE 3. Myofibrillar oscillations in Ca2þ and Sr2þ solutions. (A) Examples of oscillations occurring in a WT fiber (left) and a KO

fiber (right). Oscillations occurred at 27% of maximum force in the WT fiber and at 31% of maximum force in the KO fiber. The space

between the arrows is where the chart recorder ran at maximum speed. (B) Frequency of oscillation plotted against the percentage of

maximally activated force at which the oscillation occurred. Each circle represents one instance of oscillations occurring. The left dia-

gram shows instances of oscillations in the Ca2þ solutions, and the right diagram shows instances of oscillations in the Sr2þ solutions.

Open circles: WT fibers; filled circles: KO fibers. The line drawn at 1.25 HZ shows that higher frequency oscillations generally occurred

only in KO fibers, as can be seen from the points that lie above this line.


muscle compared with a WT muscle is shown inFigure 5B.

In unfatigued muscles from adult male mice,the half-relaxation time (time taken to fall to 50%of peak twitch force) was significantly longer inKOs than in WTs (2.2 ms longer, P < 0.001). In

fatigued muscles from the same group of mice,there was an even greater difference. The half-relaxation time of KO muscles was 7.7 ms longerthan that of WT muscles (P < 0.001). In addition,in the muscles of adult female mice, the half-relax-ation time for KOs was 3.6 ms longer than for WTs(0.01 < P < 0.05). In contrast to the adult animals,

FIGURE 4. Mass and cross-sectional area. Mass is shown in

(A) for males and (B) for females. Cross-sectional area is

shown in (C) for males and (D) for females. In adult male and

adult female mice, EDL muscles had significantly lower mass

and cross-sectional area than WTs. However, no differences

were apparent in aged muscles (>19 months old) (sample

sizes: adult males—17 WTs, 22 KOs; aged males—11 WTs, 9

KOs; adult females—13 WTs, 10 KOs; aged females—6 WTs,

6 KOs; $$0.001 < P < 0.01).

FIGURE 5. Twitch relaxation. The time course of relaxation dur-

ing an isometric twitch is shown in (A). Each line shows the

time taken to relax to 75%, 50%, and 25% of peak twitch force.

In all cases, except for unfatigued muscles from adult females,

the curves for KOs are shifted to the right of those for WTs,

suggesting a slowing of twitch relaxation in KOs. This was stat-

istically significant in unfatigued muscles from adult males,

fatigued muscles from adult males, and fatigued muscles from

adult females (using half-relaxation time as a measure of relax-

ation rate). (B) Example of slowed relaxation in a twitch from a

KO EDL muscle compared with a twitch from a WT muscle

(sample sizes: for unfatigued muscles: adult males—17 WTs,

22 KOs; aged males—11 WTs, 9 KOs; adult females—13 WTs,

10 KOs; aged females—5 WTs, 6 KOs; for fatigued muscles:

adult males—6 WTs, 8 KOs; aged males—5 WTs, 4 KOs; adult

females—6 WTs, 5 KOs; aged females—5 WTs, 6 KOs;$$P < 0.001, $0.01 < P < 0.05).


there were no statistically significant differencesbetween WTs and KOs in aged male or agedfemale mice.

Tetanus Relaxation. Relaxation after an isometrictetanic contraction can be divided into two phases:a slow linear phase, in which relaxation is homoge-neous along the muscle fiber, followed by a fast ex-ponential phase, in which some parts of the fibershorten and others lengthen.18 Figure 6A showsthe decline in force over time after a maximal te-tanic contraction in an EDL muscle. We definedthe linear phase to be that part of the curve wherea linear regression yielded an r2 of >98.5% (seeMethods). We measured the duration of this linearphase (Fig. 6B) and the rate of force decline dur-ing this linear phase (Fig. 6C), both in unfatiguedmuscles and in muscles that had been fatigued byrepetitive stimulation.

The results in Figure 6B and C show that Actn3KO muscles have a longer duration of the linearphase, and a slower rate of force decline, than WTmuscles, suggesting that KO muscles relax moreslowly than WTs after a tetanus. The longer dura-tion in KOs was statistically significant in unfa-tigued muscles of adult males (2.1 ms longer than

WTs, 0.01 < P < 0.05) and unfatigued muscles ofadult females (3.3 ms longer, 0.001 < P < 0.01).The slower rate of force decline in KOs was statisti-cally significant in unfatigued muscles of adultmales (0.16 mN/ms slower than WTs, P < 0.001),unfatigued muscles of adult females (0.15 mN/msslower, P < 0.001), and fatigued muscles of adultfemales (0.10 mN/ms slower, 0.01 < P < 0.05).However, no statistically significant differenceswere found in aged animals.

Force–Frequency Characteristics. Figure 7 showsthe force–frequency curves for males and femalesby age. Force–frequency data for individual EDLmuscles were expressed as a percentage of maxi-mum force and aggregated into single curves foreach age/gender group.

In all groups, the rapidly rising portion of thecurve ( �37.5 to 75 HZ) is slightly steeper for KOsthan for WTs as a result of a higher Hill coefficientin Actn3 KO EDL muscles. The higher Hill coeffi-cient in KOs was statistically significant in adultmales (Hill coefficient of 5.60 6 0.17 for KOs vs.5.01 6 0.07 for WTs, 0.01 < P < 0.05) and agedfemales (5.15 6 0.23 for KOs vs. 4.46 6 0.23 forWTs, 0.01 < P < 0.05).

FIGURE 6. Tetanus relaxation. (A) Two phases of relaxation after a tetanus: a linear phase (r2 > 98.5%), followed by an exponential

phase. We measured the duration (B) and rate of force decline (C) of this linear phase. The duration of this linear phase is greater,

and the rate of force decline is lower in KOs than in WTs. This suggests a slowing of relaxation in KOs compared with WTs after a te-

tanic contraction. Statistically significant differences between KOs and WTs are indicated (sample sizes—for unfatigued muscles: adult

males—6 WTs, 8 KOs; aged males—4 WTs, 3 KOs; adult females—6 WTs, 5 KOs; aged females—6 WTs, 6 KOs; for fatigued

muscles: adult males—6 WTs, 8 KOs; aged males—5 WTs, 4 KOs; adult females—6 WTs, 5 KOs; aged females—6 WTs, 6 KOs;$$$P < 0.001, $$0.001 < P < 0.01, $0.01 < P < 0.05).


Another feature of the force–frequency curvesis that, with the exception of aged females, thecurve for Actn3 KOs starts at a lower level than thecurve for WTs as a result of a lower twitch-to-teta-

nus ratio in KO EDL muscles. The lower twitch-to-tetanus ratio was statistically significant in adultmales (twitch-to-tetanus ratio of 18.4 6 0.4% forKOs vs. 20.3 6 0.4% for WTs, 0.01 < P < 0.05).

DISCUSSION

Muscles that lack a-actinin-3 display a shift fromfast-twitch toward slower twitch characteristics.10,11

The location of a-actinin-3 within the Z-disk of fast-twitch muscle fibers, where it interacts with a mul-titude of structural, signaling, and metabolic pro-teins,1,23 suggests that, at a subcellular level, it mayinfluence the properties that distinguish fast-twitchand slow-twitch fibers. At a subcellular level, two im-portant factors that characterize a fiber as fast-twitchor slow-twitch are: (i) the properties of the contract-ile filaments; and (ii) the rate at which the SR rese-questers the Ca2þ released during excitation.13

Our first aim in this study was to use theskinned fiber technique to determine whetherthere were any differences in these two factorsbetween wild-types and knockouts that might con-tribute to the slower twitch characteristics of Actn3KO muscles. The skinned fiber technique has beenused widely to characterize the properties of thecontractile filaments and SR in skeletal muscle,and it is the most sensitive probe with which toelucidate mechanisms at the level of contractileproteins and SR.14

To examine the rate of Ca2þ uptake by the SR,we looked at the force responses generated by caf-feine-induced Ca2þ release from the SR ofmechanically skinned fibers after predeterminedloading periods. This is an established techniquefor investigating Ca2þ uptake by the SR14,15 andhas the advantage of being able to directly probeSR reuptake mechanisms in the absence of any sar-coplasmic contents or differing phosphorylationlevels, which could confound the investigation ofthe basic mechanisms of SR Ca2þ uptake. Our majorfinding from these studies is that the SR in KOfibers loaded Ca2þ more slowly than the SR in WT

FIGURE 7. Force–frequency characteristics. Force–frequency

data from individual EDL muscles have been aggregated to

form single curves for each age/gender group. The curve for

KOs (dashed line) rises more steeply than the curve for WTs

(solid line) in all groups, indicative of a higher Hill coefficient in

KOs. The curve for KOs starts at a lower level than the curve

for WTs in all groups except aged females, indicative of a lower

twitch-to-tetanus ratio in KO EDL muscles. In all cases these

differences were not statistically significant. A star on the steep

portion of the curve indicates a statistically significant difference

in Hill coefficient and a star under the start of the curve indi-

cates a significant difference in twitch-to-tetanus ratio (sample

sizes: adult males—11 WTs, 14 KOs; aged males—6 WTs, 4

KOs; adult females—7 WTs, 5 KOs; aged females—6 WTs, 6

KOs; $0.01 < P < 0.05).


fibers. For any particular duration of loading, theamount of Ca2þ loaded by the SR, expressed as apercentage of the maximum amount it could load,was lower in KO than in WT fibers (see Fig. 1B).

This alteration in SR Ca2þ uptake is evidenceof a shift at a subcellular level toward slower twitchcharacteristics in the absence of a-actinin-3. It iswell established that the uptake of Ca2þ by the SRis considerably slower in slow-twitch than in fast-twitch fibers. This is due to slow-twitch fibers hav-ing a smaller surface area of SR membrane and atwo- to threefold lower density of Ca2þ-ATPasepumps in the SR membrane.24 In addition, themaximum Ca2þ capacity of the SR in slow-twitchfibers is only about one-third of that in fast-twitchfibers due to smaller amounts of calsequestrinbeing available to buffer rises in [Ca2þ]. Thisinhibits the ability of the Ca2þ-ATPase to pumpCa2þ back into the SR lumen.25 It is possible thatfactors such as these underlie the slower Ca2þ

uptake by the SR in Actn3 KO fibers and that a-acti-nin-3 deficiency may be associated with smaller SRmembrane surface area, lower Ca2þ-ATPase pumpdensity, or lower SR Ca2þ capacity and calseques-trin levels. It should also be noted that anincreased rate of Ca2þ leakage from the SR wouldalso reduce the rate of Ca2þ loading,14 and we can-not rule out the possibility that Ca2þ leakage isincreased in the SR of knockout fibers.

Trinh and Lamb,15 in their study on the prop-erties of the SR, found that fast fibers from the rattook a comparatively long time (approximately 25–90 s) to reload the SR to half-maximum capacity.Our fast fibers loaded significantly faster (approxi-mately 2–5 s to reach 50%; see Fig. 1B), which ismost likely a reflection of the different [Ca2þ] inthe loading solutions. Our loading solution con-tained 50 mM EGTA, compared with 1 mM EGTAin the loading solution used by these other investi-gators. Although the low [EGTA] used by Trinhand Lamb was useful in indicating the maximumabsolute Ca2þ capacity of the SR, our high [EGTA]loading solution was designed to measure the rateat which the SR could pump Ca2þ.

The characterization of a fiber as fast twitch orslow twitch is also influenced by the properties ofthe contractile filaments. To investigate the effectsof a-actinin-3 deficiency on the properties of thecontractile filaments, we chemically skinned thefibers after the SR loading experiments andexposed them to solutions of varying [Ca2þ] and[Sr2þ] to generate force–pCa and force–pSr curves.The value of pCa50 � pSr50 was virtually identicalin KOs and WTs (Fig. 2C), suggesting an absenceof any gross shift in myosin heavy chain isoformcomposition in a-actinin-3–deficient fibers. Thisresult is consistent with our earlier immunohisto-

chemical studies showing that the proportion offibers that stain for fast or slow MHC isoforms wasvirtually the same in WT and KO EDL muscles.10

The force–pCa and force–pSr curves weresteeper for KOs than for WTs, as indicated by thehigher nCa and nSr values in KO fibers (Fig. 2C).The nCa and nSr values are indicators of the degreeof cooperativity between subunits of the thick andthin filaments in producing tension,26 and suggestthat there is a greater cooperativity of the contract-ile filaments in the absence of a-actinin-3.

Actin–myosin crossbridge interactions can influ-ence cooperativity. The binding of myosin headsto actin facilitates further binding and stabilizesthe thin filament in a state of higher Ca2þ sensitiv-ity.13 This is relevant to our finding that KO fibersexhibited myofibrillar oscillations of a higher fre-quency than WT fibers when they were stimulatedsubmaximally by Ca2þ and Sr2þ (Fig. 3B). Oscilla-tions of myofibrillar origin occur in skinned skele-tal and cardiac muscle fibers at submaximal levelsof activation and when they are rapidly activatedunder specific conditions of temperature, ioniccomposition of the myoplasmic environment, andsarcomere length.19,27–29 Myofibrillar oscillationsare a reflection of crossbridge interactions; thehigher frequency in our KO fibers could beexplained if the properties of myosin had changedin such a way that either the rate of detachment ofthe myosin heads had increased or the rate ofattachment had increased.22 The altered myosin–actin interaction suggested by the change in fre-quency of myofibrillar oscillations could be a factorcontributing to the altered cooperativity observedin Actn3 KO fibers.

The Actn3 KO mouse is a newly generatedmodel of a-actinin-3 deficiency,2 so it is importantto characterize the contractile properties of wholefast-twitch muscle from this mouse model. This hasso far been done only in an adult male cohort ofmice,11 and it is not known whether these proper-ties vary with age and gender. This informationwould be a valuable database for future studiesinto the mechanisms that produce the altered mus-cle characteristics of Actn3 KO fast-twitch muscle.Hence, our second aim in this study was to exam-ine the contractile properties of whole Actn3 KOEDL muscle from males and females in two agegroups: adult (2–6 months old) and aged (19– 22months old).

In the muscles of adult males and adultfemales, we found a significant slowing of relaxa-tion in Actn3 KOs for both a twitch (Fig. 5A) and atetanus (Fig. 6A and B). In light of our skinnedfiber results, this is most likely due to the sloweruptake of Ca2þ by the SR in a-actinin-3–deficientfibers. Another contributing factor could be the


slower kinetics of dissociation of Ca2þ from thecontractile filaments that we found in a-actinin-3–deficient fibers. The effect of a steeper force–pCacurve in Actn3 KO fibers is more difficult to inter-pret. A steeper curve would in fact cause fasterrelaxation, because a given drop in [Ca2þ] will beaccompanied by a larger drop in force. The factthat relaxation is actually slower in KO musclessuggests that the effect of slower Ca2þ uptake bythe SR outweighs the effect of a steeper force–pCarelationship. In fact, from the exponential curvesfitted to our SR Ca2þ loading data (Fig. 1B), wecalculated that, for a loading time of 20 ms (corre-sponding to a stimulation frequency of around 50HZ), the percentage of maximum Ca2þ loadingwas 2.5 times greater in WT fibers than in KOfibers. The magnitude of this difference suggeststhat the reduced rate of Ca2þ uptake by the SR ina-actinin-3–deficient fibers is likely to be the majorcontributor to the slower relaxation rates of Actn3KO muscles.

Another factor that could contribute to theslowing of twitch and tetanus relaxation is a reduc-tion in the levels of myoplasmic Ca2þ buffers suchas parvalbumin.30 Our skinned fiber experimentswere designed to specifically investigate the proper-ties of the contractile proteins and SR without theconfounding effects of factors such as parvalbuminthat would be present in an intact fiber prepara-tion. Hence, in addition to the contractile proteinand SR properties that we have examined in thisstudy, it must be noted that, if there is a slightreduction in parvalbumin levels in a-actinin-3–defi-cient fibers, this would also contribute to slowerrelaxation times. We cannot rule out a contribu-tion from this source.

We also found changes in the force–frequencycurves of whole Actn3 KO EDL muscles (Fig. 7). Inadult males and aged females the force–frequencycurves were significantly steeper for KOs than forWTs. This parallels the steeper force–pCa andforce–pSr curves of skinned Actn3 KO fibers and inthe same way could reflect a greater cooperativityof the contractile proteins in producing tension. Itmay also be a consequence of longer twitch-relaxa-tion times in KO muscles; all other factors beingequal, longer relaxation times would lead to twitchsummation at earlier frequencies of stimulation anda steeper force–frequency relationship. We alsofound that the twitch-to-tetanus ratio of KO muscleswas significantly lower than that of WTs in adultmales. This is consistent with a shift toward slowertwitch properties in a-actinin-3–deficient muscle, asslow-twitch motor units have lower twitch-to-tetanusratios than fast-twitch motor units.31

The mass and cross-sectional areas of wholeActn3 KO EDL muscles also showed a shift toward

slow-twitch characteristics. Slow-twitch fibers havesmaller cross-sectional areas than fast-twitch fibers,and we have shown that there is, in fact, a reduc-tion in fast fiber diameter in muscles from theActn3 KO mouse.10 This would explain the signifi-cantly smaller cross-sectional areas of Actn3 KOEDL muscles from adult males and adult females.

Overall, there was little difference betweenmales and females regarding the effects of a-acti-nin-3 deficiency on whole EDL muscle. Changes inmass, cross-sectional area, twitch-relaxation, tetanus-relaxation, and force–frequency characteristics inthe absence of a-actinin-3 were found in both malesand females. However, there was some reduction inthe effects of a-actinin-3 deficiency with age. Therehas been a shift in the whole muscle phenotype ofthe Actn3 KO mouse with age, with the statisticallysignificant differences found in adult animals nolonger being found in aged animals.

In conclusion, our data provide further insightinto the mechanisms that underlie changes in skel-etal muscle performance associated with a-actinin-3deficiency. Our skinned fiber data show that aslowing of Ca2þ reuptake by the SR in a-actinin-3–deficient fibers is one factor that contributes to theslower relaxation of whole KO muscles after both atwitch and a tetanus. In whole KO EDL muscle, wefound changes in mass, cross-sectional area, twitch-relaxation, tetanus-relaxation, and force–frequencycharacteristics that are consistent with a shift to-ward a slower twitch profile in fast-twitch musclesthat lack a-actinin-3. There is a clear ameliorationof the Actn3 KO phenotype with age. In contrast,gender did not affect the Actn3 KO phenotype.These changes to the properties of individualfibers and whole muscles seen in the Actn3 KOmouse are detrimental to sprint and power activ-ities, as has been observed in both athlete andnon-athlete human cohorts.

The authors acknowledge M.B. Arber for her expert technical as-sistance and for analyzing the skinned fiber data. We thank Profes-sor George Stephenson for his evaluation of the oscillationsobtained in the skinned fiber study.

REFERENCES

1. MacArthur DG, North KN. A gene for speed? The evolution andfunction of a-actinin-3. Bioessays 2004;26:786–795.

2. MacArthur DG, Seto JT, Raftery JM, Quinlan KG, Huttley GA, HookJW, et al. Loss of ACTN3 gene function alters mouse muscle metabo-lism and shows evidence of positive selection in humans. Nat Genet2007;39:1261–1265.

3. Yang N, MacArthur DG, Gulbin JP, Hahn AG, Beggs AH, Easteal S,et al. ACTN3 genotype is associated with human elite athletic per-formance. Am J Hum Genet 2003;73:627–631.

4. Niemi AK, Majamaa K. Mitochondrial DNA and ACTN3 genotypesin Finnish elite endurance and sprint athletes. Eur J Hum Genet2005;13:965–969.

5. Papadimitriou ID, Papadopoulos C, Kouvatsi A, Triantaphyllidis C.The ACTN3 gene in elite Greek track and field athletes. Int J SportsMed 2008;29:352–355.

6. Roth SM, Walsh S, Liu D, Metter EJ, Ferrucci L, Hurley BF. TheACTN3 R577X nonsense allele is under-represented in elite-levelstrength athletes. Eur J Hum Genet 2008;16:391–394.


7. Eynon N, Duarte JA, Oliveira J, Sagiv M, Yamin C, Meckel Y, et al.ACTN3 R577X polymorphism and Israeli top-level athletes. Int JSports Med 2009;30:695–698.

8. Clarkson PM, Devaney JM, Gordish-Dressman H, Thompson PD,Hubal MJ, Urso M, et al. ACTN3 genotype is associated withincreases in muscle strength in response to resistance training inwomen. J Appl Physiol 2005;99:154–163.

9. Moran CN, Yang N, Bailey ME, Tsiokanos A, Jamurtas A, MacArthurDG, et al. Association analysis of the ACTN3 R577X polymorphismand complex quantitative body composition and performance phe-notypes in adolescent Greeks. Eur J Hum Genet 2007;15:88–93.

10. MacArthur DG, Seto JT, Chan S, Quinlan KG, Raftery JM, Turner N,et al. An Actn3 knockout mouse provides mechanistic insights intothe association between a-actinin-3 deficiency and human athleticperformance. Hum Mol Genet 2008;17:1076–1086.

11. Chan S, Seto JT, MacArthur DG, Yang N, North KN, Head SI. Agene for speed: contractile properties of isolated whole EDL musclefrom an a-actinin-3 knockout mouse. Am J Physiol Cell Physiol 2008;295:C897–C904.

12. Allen DG, Lannergren J, Westerblad H. Muscle cell function duringprolonged activity: cellular mechanisms of fatigue. Exp Physiol 1995;80:497–527.

13. Bottinelli R, Reggiani C. Human skeletal muscle fibers: molecularand functional diversity. Prog Biophys Mol Biol 2000;73:195–262.

14. Bakker AJ, Head SI, Wareham AC, Stephenson DG. Effect of clenbu-terol on sarcoplasmic reticulum function in single skinned mammalianskeletal muscle fibers. Am J Physiol Cell Physiol 1998;274:C1718–C1726.

15. Trinh HH, Lamb GD. Matching of sarcoplasmic reticulum and con-tractile properties in rat fast- and slow-twitch muscle fibers. Clin ExpPharmacol Physiol 2006;33:591–600.

16. Head SI, Stephenson DG, Williams DA. Properties of enzymaticallyisolated skeletal fibers from mice with muscular dystrophy. J Physiol1990;422:351–367.

17. Motulsky HJ, Christopoulos A. Fitting models to biological datausing linear and nonlinear regression. A practical guide to curve fit-ting. San Diego, CA: GraphPad Software; 2003.

18. Westerblad H, Lannergren J. Slowing of relaxation during fatigue insingle mouse muscle fibers. J Physiol 1991;434:323–336.

19. Stephenson DG, Williams DA. Effects of sarcomere length on theforce–pCa relation in fast- and slow-twitch skinned muscle fibersfrom the rat. J Physiol 1982;333:637–653.

20. Fink RHA, Stephenson DG, Williams DA. Calcium and strontiumactivation of single skinned muscle fibers of normal and dystrophicmice. J Physiol 1986;373:513–525.

21. Williams DA, Head SI, Lynch GS, Stephenson DG. Contractile prop-erties of skinned muscle fibers from young and adult normal anddystrophic (mdx) mice. J Physiol 1993;460:51–67.

22. Smith DA, Stephenson DG. The mechanism of spontaneous oscilla-tory contractions in skeletal muscle. Biophys J 2009;96:3682–3691.

23. Frey N, Frank D, Lippl S, Kuhn C, Kogler H, Barrientos T, et al. Cal-sarcin-2 deficiency increases exercise capacity in mice through calci-neurin/NFAT activation. J Clin Invest 2008;118:3598–3608.

24. Pette D, Staron RS. Cellular and molecular diversities of mammalianskeletal muscle fibers. Rev Physiol Biochem Pharmacol 1990;116:1–76.

25. Fryer MW, Stephenson DG. Total and sarcoplasmic reticulum con-tents of skinned fibers from rat skeletal muscle. J Physiol 1996;493:357–370.

26. Moss RL, Diffee GM, Greaser ML. Contractile properties of skeletalmuscle fibers in relation to myofibrillar protein isoforms. Rev PhysiolBiochem Pharmacol 1995;126:1–63.

27. Stephenson DG, Williams DA. Calcium-activated force responses infast- and slow-twitch skinned muscle fibers of the rat at different tem-peratures. J Physiol 1981;317:281–302.

28. Okamura N, Ishiwata S. Spontaneous oscillatory contraction of sarco-meres in skeletal myofibrils. J Muscle Res Cell Motil 1988;9:111–119.

29. Shimamoto Y, Suzuki M, Ishiwata S. Length-dependent activationand auto-oscillation in skeletal myofibrils at partial activation byCa2þ. Biochem Biophys Res Commun 2008;366:233–238.

30. Stephenson DG, Lamb GD, Stephenson GMM. Events of the excita-tion–contraction–relaxation (E-C-R) cycle in fast- and slow-twitchmammalian muscle fibers relevant to muscle fatigue. Acta PhysiolScand 1998;162:229–245.

31. Burke RE, Levine DN, Tsairis P, Zajac FE. Physiological types andhistochemical profiles in motor units of the cat gastrocnemius.J Physiol 1973;234:723–748.


Documents

Properties of extensor digitorum longus muscle and skinned fibers from adult and aged male and female Actn3 knockout mice