Proc. Nail. Acad. Sci. USAVol. 88, pp. 7101-7105, August 1991Biophysics

Regulation of skeletal muscle stiffness and elasticity by titinisoforms: A test of the segmental extension model ofresting tension

(cardiac muscle/connectin/elastic rdaments/elastic limit/sarcomere matrix)

KUAN WANG*t4, ROGER MCCARTER§, JOHN WRIGHT*, JENNATE BEVERLY§, AND RUBEN RAMIREZ-MITCHELLt*Clayton Foundation Biochemical Institute, Department of Chemistry and Biochemistry, and tCell Research Institute, The University of Texas at Austin,Austin, TX 78712; and §Department of Physiology, University of Texas Health Science Center at San Antonio, TX 78284

Communicated by Lester J. Reed, May 10, 1991

ABSTRACT To explore the role of titin filaments in muscleelasticity, we measured the resting tension-sarcomere lengthcurves of six rabbit skeletal muscles that express three sizeclasses of titin isoform. The stress-strain curves of the splitfibers of these muscles displayed a similar multiphasic shape,with an exponential increase in tension at low sarcomere strainfollowed by a leveling of tension and a decrease in stiffness atand beyond an elastic limit (yield point) at higher sarcomerestrain. Significantly, positive correlations exist between the sizeof the expressed titin isoform, the sarcomere length at the onsetof exponential resting tension, and the yield point of eachmuscle. Immunoelectron microscopic studies of an epitope inthe extensible segment of titin revealed a transition in the elasticbehavior of the titin ifaments near the yield point sarcomerelength of these muscles, providing direct evidence of titin'sinvolvement in the genesis ofresting tension. Our data led to theformulation of a segmental extension model of resting tensionthat recognizes the interplay of three major factors in shapingthe stress-strain curves: the net contour length of an extensiblesegment of titin filaments (between the Z line and the ends ofthe thick filaments), the intrinsic molecular elasticity of titin,and the strength of titin thick filament anchorage. Our datafurther suggest that skeletal muscle cells may control andmodulate stiffness and elastic limit coordinately by selectiveexpression of specific titin isoforms.

A quiescent skeletal muscle is remarkably elastic. It extendsand develops tension when it is stretched and then snaps backto restore its original length when released. As quiescentmuscle is activated, it shortens and develops tension and thenrelengthens to its original length when activation ceases (1).The structural and molecular basis ofthe long-range elasticityremains obscure. Earlier physiological studies have generallymodeled elasticity as elastic elements either in series or inparallel with a contractile unit without specifying their ana-tomical origin. Recent studies clearly indicate that within thephysiological range of muscle length change, myofibrillarstructures are the major source of elasticity and that thesarcolemma and extracellular connective tissues begin tocontribute significantly only in highly extended muscles (refs.2 and 3 and references therein). Since neither actin normyosin filaments of the sarcomere display long-range elas-ticity, a prime candidate is the newly recognized sarcomerematrix, which contains elastic titin filaments that connectmyosin filaments, along their length, from the M line to theZ line (4-10).A reduction of resting tension of stretched muscle when

titin was preferentially destroyed by radiation (11) or bycontrolled proteolysis (12) implicated titin in muscle elastic-

ity. The positional stability ofthick filaments during isometriccontraction is also thought to be a manifestation of titinelasticity (13). We have taken a more direct approach bymeasuring resting tension as skeletal muscle sarcomeres arestretched stepwise to well beyond its elastic limit and inter-pret these resting tension-sarcomere length (SL) curves(stress-strain curves) based on the extensibility of titin fila-ments as determined by epitope translocation studies. Ouranalysis is greatly facilitated by a comparative study of sixrabbit skeletal muscles that express three size classes of titinisoform and thus offer an excellent opportunity to definestructure-function correlations (14, 15). Our data suggest thatthe characteristic shape of stress-strain curves can be un-derstood by the tension generated by the reversible extensionof a segment of titin between the Z line and the ends of thickfilaments as well as the strength of titin thick filamentanchorage. Furthermore, our data suggest that skeletal mus-cle cells may control and modulate elasticity and complianceand the elastic limit of the sarcomere by selective expressionof specific titin size isoforms.

METHODSTension and SL Measurements. Single fibers of rabbit

muscle tissues were removed and mechanically skinned bysplitting longitudinally into myofibrillar bundles in a relaxingsolution as described (15). The split fiber was attachedbetween a force transducer (Cambridge Technology 400A)and a muscle ergometer (Cambridge Technology 300S)mounted on a computer-controlled micromanipulator. Thefibers were slowly stretched stepwise (a 10% step in 30 sec)at 24°C. At the end of each stretch, fiber length was heldconstant for 150 sec to allow tension to decay from peak (peaktension) to near plateau (plateau tension), and the SL wasmonitored by the first-order diffraction of a He/Ne laser.Gel Electrophoresis and Immunoelectron Microscopy. Titin

size isoforms were identified by SDS gel electrophoresis asdescribed (15). Titin epitope translocation was monitored byimmunoelectron microscopy as described (15). Details aregiven in the figure legends.

RESULTSTitin Size Isoforms in Skeletal Muscles. To evaluate expres-

sion of titin size isoforms in various muscle tissues of adultrabbit, a group of six skeletal muscles-adductor magnus(AM), psoas (PS), longissimus dorsi (LD), sartorius (SA),

Abbreviations: AM, adductor magnus; PS, psoas; LD, longissimusdorsi; SA, sartorius; SO, soleus; ST, semitendinosus; CA, cardiac;SL, sarcomere length.1To whom reprint requests should be addressed at: Department ofChemistry and Biochemistry, University of Texas at Austin, Aus-tin, TX 78712.

7101

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement"in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Dow

nloa

ded

by g

uest

on

May

6, 2

020

Proc. Natl. Acad. Sci. USA 88 (1991)

soleus (SO), and semitendinosus (ST)-and cardiac (CA)muscle were analyzed by a high-resolution gel system opti-mized for resolving megadalton proteins (15). The gel pat-terns (Fig. 1) revealed that certain tissues (SO and ST)displayed a single titin band, whereas others contained adoublet (AM, PS, LD, and SA) or even a triplet (CA). Theslowest titin bands in skeletal muscles, presumably the intacttitin (Ti), decrease in mobility in the following order: AM andPS < LD and SA < SO and ST. In contrast, the faster titinbands, thought to be a degraded fragment (T2), have identicalmobility in different tissues. To estimate the relative molec-ular mass of T1 in these tissues, PS sample was added as aninternal standard and the mixture was electrophoresed on along-format gel to enhance mobility differences. Plots ofmolecular mass vs. mobility, assuming 2.8 and 2.4 MDa forT1 and T2 of rabbit PS, respectively, yielded the following setof values for T1: AM and PS, 2.8 MDa; LD and SA, 2.88MDa; SO and ST, 2.94 MDa (16, 17). It should be noted,however, since accurate molecular mass standards in thissize range are not yet available, these values are tentative andare useful mainly as a measure of the relative difference inmolecular mass. These data were reproducible from muscletissues of five adult rabbits and confirmed and extended ourearlier analysis (15).

Stress-Strain Curves of Split Muscle Fibers in RelaxingSolutions. The stress-strain curves of mechanically skinnedsplit fibers from each of the six skeletal muscles weremeasured in a relaxing solution and compared (Fig. 2). It wasobserved that upon each step of stretching, the restingtension of the split fiber quickly rose to a peak and thendecayed slowly in an exponential manner toward a steady-state plateau value. This stress-relaxation took place withoutdetectable changes in SL and its relaxation rate varied withthe magnitude and speed oflength change as well as the initialSL (data not shown). We have operationally defined peaktension as the maximum of the tension immediately after thestretch and plateau tension as the tension after 2.5 min ofrelaxation, which, for the experiments reported here, iswithin 90%o of the tension measured after 40-60 min of stressrelaxation.

Plots of either peak tension or plateau tension of split fibersstretched from 2.2 to 6 pum in SL displayed multiphasiccurves similar in character for all six skeletal muscles (Fig. 2).

ATitin T1[E

T2 -

E KC

Nebulin-

BTitinT1[ .*Ti (

T 2-

Nebulin-

AM PS LD SA SO ST CA

FIG. 1. Titin size isoforms of rabbit skeletal and cardiac muscles.(A) Muscle tissues from the same adult rabbit were snap frozen,pulverized in liquid nitrogen, solubilized in hot SDS, and analyzed ona 2-12% gradient gel. (B) To reveal subtle mobility differences, PSsample was added as an internal standard and the mixture wassubjected to electrophoresis. Only the top portion of each gel isshown. Of the skeletal muscles, isoforms of intact titin (T1) increasein size in the following order: AM and PS < LD and SA < SO andST. CA displays the smallest titin. Note also that two size classes ofskeletal muscle nebulin are expressed, with SO and ST displaying thelarger ones. No nebulin was found in cardiac muscle.

As the split fiber was stretched stepwise beyond slack SL(SL.) around 2.2 um, little or no tension developed until acharacteristic length (SLe) was reached beyond which restingtension increased exponentially. The tension peaked at ayield point sarcomere length (SLy), which varies between 3.7and 5.0,m. Beyond the yield point, the tension leveled offor decreased slightly. Tension began to increase near 5.5 Umuntil the fiber was either torn apart or off the hooks. It isapparent from examining Fig. 2 that the yield point is morepronounced and easier to define with the peak tension curvesthan with the plateau tension curves. This is especiallycritical for the pair of muscles (SO and ST) for which the yieldpoints were clearly discerned only in peak tension curves(Fig. 2).

Correlation of Resting Tension Characteristics and Size ofTitin Isoform. Closer examination of these curves revealedseveral interesting trends. First, the SLe, which marks theonset of exponential increase of resting tension, varies withmuscle tissues. Quantitatively, SLe was determined as the xintercept of the linear portion of a plot of log(peak or plateautension) vs. SL. SL. values of this set of muscles increase inthe following order: AM and PS (2.8 pm) < LD and SA (3.0jLm) < SO and ST (3.6 gum). It is interesting that this orderfollows closely the size ranking of titin isoforms in thesemuscles. Thus, PS developed tension beyond 2.8 aum,whereas ST can be stretched readily up to 3.6 j.m withoutappreciable tension. Second, SLY, which marks the yieldpoint, also varies with muscle tissues. Quantitatively, SLYwas determined by the local maximum in tension that wasidentified by a plot of the first derivative of either peak orplateau tension vs. SL (data not shown). The yield point ischaracteristic of each tissue and very little variation wasdetected when a large number of fibers were examined.Reproducible yield points, however, were observed onlywhen fresh fibers were used without extended storage (<6hr). In aged fibers, yield points became less pronounced orundetectable. Sly values of this set of muscles increase in thefollowing order: AM and PS (3.8 ttm) < LD and SA (4.2 ,um)< SO and ST (4.8 ,um). Again, this order is strikingly similarto that of the titin isoforms expressed in these muscles. Thus,the muscle tissues that express longer titin tend to havelonger SL, and SLY values.Taken together, a positive correlation of two independent

parameters ofthe stress-strain curves of resting muscles withthe size of titin isoforms strongly implicates titin in thegenesis of resting tension. Moreover, the length and size oftitin appears to be an important factor in determining whensarcomeres will develop resting tension upon stretch andwhere the sarcomere will yield under stress.

Titin Segment Strain: Biphasic Epitope TranslocationCurves of Titin by Immunoelectron Microscopy. Previousstructural studies indicate that each titin filament in thesarcomere consists of two segments, with the segment be-tween the Z line and the edge of the A band being compliantand extensible and the remaining segment that overlaps withthick filaments being stiff and inextensible (4, 5, 7-10). It isthus reasonable to propose that the reversible length changeof the extensible segment of titin must underlie the genesis ofresting tension. To explore this segmental extension hypoth-esis further, we measured the titin segment strain ofrabbit PSand ST by the epitope translocation technique. The stretchdependence of a titin epitope recognized by monoclonalantibody RT13 was monitored by measuring the center tocenter distance of labeled epitope zone to the Z line (Ab-Z)as the split fiber was stretched to various lengths beforeantibody labeling and electron microscopy (Fig. 3A). Theresulting epitope translocation curves (plots ofAb-Z vs. SL)are biphasic, consisting of two linear segments (Fig. 3B). Insarcomeres stretched to a moderate degree, the RT13 epitopeshifts its position with a slope of 0.24 for PS and 0.33 for ST.

7102 Biophysics: Wang et al.

Dow

nloa

ded

by g

uest

on

May

6, 2

020

Proc. Natl. Acad. Sci. USA 88 (1991) 7103

2 3 4 5 6

LD SLy

0 0

0 00

0.0.0

0

-.0 VSLOV

SA

cr

I.s.

~~~~~~~~0

.02

1 5-

SLy 000

00.+ Wtlt90.

-

VALov 2 3 4 5-

23 4 5 6Sarcomere Length (gm)

So

00

0

0

ST SLy40

0000

0

0 0

0*0o00

an10 ovV

3 4 5 6

FIG. 2. Stress-strain curves of split fibers of rabbit skeletal muscles in a relaxing solution. Single muscle fibers from each tissue weremechanically skinned and split into myofibrillar bundles in a relaxing solution. Each split fiber was mounted to a force transducer and thenstretched stepwise. Peak tension (tension generated immediately after each stretch) (o) and plateau tension (steady-state tension after 2.5 minof stress-relaxation) (e) were plotted as a function of SL. The stress-strain curves of all split fibers displayed similar multiphasic profiles: theresting tension increased exponentially, followed by a gradual leveling before another increase in tension at a very high degree of stretch. Threecharacteristic SLs were identified: SL2 (onset of exponential tension), SLy (the yield point), and SLO, (thick and thin filament nonoverlap).(Insets) Plots of log(tension) vs. SL used to define SL,. It was noted that both SLe and SLy for each muscle increase in the following order:AM and PS < LD and SA < SO and ST.

Since these values are intermediate between 0 and 0.5, valuesthat are expected for an epitope that remains stationaryrelative to the Z line or to the A band, respectively, the RT13epitope shifted elastically upon stretch. At higher degrees ofstretch, however, the epitope zone broadened significantly(Fig. 3C) and the epitope shifted with a diminished slope (0.12for PS and 0.15 for ST) (Fig. 3B). The transition occurredaround a characteristic SL (SLRT13) for each muscle: 3.8 ,gmfor PS and 4.8 ,m for ST. Thus, the titin extensible segmentchanges its extensibility at this transition point. Interestingly,the transition point for each muscle corresponds closely to

40

A 35 ST

(A 30SLy

25 000° X~~~000

is15

1 0

F-

5- 8_

B PS ST

SL 'P*

4 5 67RT3

N

2 3 4 5 6 7 2 3 4

Sarcomere Length (gm) Sarcom

~ t'

ierem":U,

the yield point of its stress-strain curve (Fig. 2). The closecorrelation of the change in extensibility of titin filaments andthe mechanical responses of the sarcomeres upon stretchprovides additional and strong support for the notion that titinsegment strain is a major, if not exclusive, contributing factorin the genesis of resting tension.

Yield Point Tension per Thick Filament. When fibers werestretched beyond the yield point, resting tension leveled offand the fiber became more compliant and elongated easilywith additional stretch. This behavior suggests that at theyield point the sarcomere has reached its elastic limit. To

0

0

0

0 FIG. 3. Biphasic titin epitope translo-

cation curves of rabbit PS and ST. Splitfibers of rabbit PS and ST muscles were

- stretched to various lengths in a relaxingsolution mounted on a single slot foldgrid, fixed with formaldehyde, and la-beled with a monoclonal anti-titin anti-

hbody (RT13; 20 Ag/ml in phosphate-

buffered saline for 15 hr at 40C). This wasiT13 followed with rabbit anti-mouse immuno-

.- 0. globulin and colloidal gold-protein Aprior to plastic embedding and sectioning(15). RT13 epitopes were localized to theI band in both muscles (C). In longersarcomeres, RT13 epitopes broadenedand were translocated away from the Zline. The midpoint of each epitope zonethat is demarcated by gold conjugates is

, . , . used to measure the average distance of6 7 8 epitopes to the center of the adjacent Z

Length (gm) line (C, arrows). Plots ofAb-Z vs. SL (B)are biphasic and consist of two linearsegments points (SLRT13)

that coincide with the yield point SLy of

the same muscle (A). The striking corre-

spondence between the yield point of

;Z.-.S*: - resting tension and a structural transitionof titin filaments strongly suggests that

the genesis of resting tension and thean.5>_ *m yield phenomenon is titin mediated.

C/)c)

0

cnc)H1Rae

eE E N A_ ~~~~~~II ..,. ....

Biophysics: Wang et al.

- -41-1-..:.;Wm;;-:..I'"W"

L----j4.

Dow

nloa

ded

by g

uest

on

May

6, 2

020

7104 Biophysics: Wang et al.

characterize the yield point further, we calculated yield pointtension per thick filaments by counting the total number ofthick filaments per split fiber in the electron micrographs ofcross-sections of a set of PS and ST fibers after stress-strainmeasurements. It was calculated that the plateau tension perthick filament at the yield point is 1.5 x 10-5 (PS) and 5.8 x10-6 dynes (ST). Since there are an average of six titinpolypeptides per thick filament per half sarcomere in thesemuscles (4, 5), the average force-bearing capacity for eachextensible titin segment is 2.5 x 10-6 (PS) and 0.9 x 10-7dynes (ST). Since the yield point tension represents a localmaximum for the exponential phase of the resting tensionstress-strain curve, it follows that PS is a much stiffer musclethan ST (i.e., elastic module is twice as high). These dataraise the intriguing possibility that the expression ofa specifictitin isoform may also dictate both the stiffness and themaximum resting tension that each muscle can exert.

DISCUSSIONExpression of Titin Isoforms and Correlation with Tissue-

Specific Resting Tension Characteristics. A systematic studyof the stress-strain curves of a larger number of split fibersfrom various rabbit tissues, each of which expresses one ofthe three titin size isoforms (Fig. 1), has revealed a strikingcorrelation of isoform expression and tissue-specific restingtension characteristics. Several trends have emerged. (i)Muscles that express larger titin isoforms tend to initiatetension at a longer SL (Fig. 4). (ii) Muscles that express largertitin isoforms reach their elastic limit at higher SL (Fig. 4). (iii)Muscles that express the longest titin (ST) also develop thelowest tension (see Results). Our data revealed that theelastic characteristics of skeletal muscles are intimatelylinked to, or perhaps are dictated by, the specific expressionof a particular titin size isoform. It should be noted that thesecorrelations hold true irrespective of the absolute molecularmass values of intact titin, insofar as gel mobility is a validmeasure of titin size.

Segmental Extension of Titin Filament as the StructuralBasis of Stress-Strain Curves. The fundamental features ofstress-strain curves of resting split fibers can be understoodon the notion that the reversible extension of a nonanchoredor nonconstrained segment of the titin filament is the deter-mining factor for the functional forms ofmechanical responseupon stretch. Our current working hypothesis is shown inFig. 5. The stress-strain curves are divided into four regions(I-IV), each with a distinct structural interpretation.Region I: From SLO (resting length) to SLe (the onset of

exponential increase in tension). No significant tension is

ea

a7)

.:,, Si:~of S[-

8

A!r--;' P

-,,-

;'s1

F~~ ~ ~

SA S 0 S

FIG. 4. Correlation of resting tension characteristics and molec-ular properties of titin isoforms. The molecular mass of titin isoform(MDa) of each tissue displays a positive correlation with the SL atwhich tension increases exponentially (SLe) with the SL at whichtension levels off at yield point (SLy) as well as with the SL at whichtitin epitopes in the I band begin to translocate differently (SLRT13).

Proc. Natl. Acad. Sci. USA 88 (1991)

generated upon stretch. This could result from a linearextension of the titin segment (as demonstrated by epitopetranslocation curves of Fig. 3) in a non-force-producingmode.Region II: From SLe (exponential tension increase) to SLy

(yield point). Tension increases exponentially and subse-quently deviates and reaches a local maximum. In this region,titin extends its contour length linearly, perhaps by unfoldingprotein chains in a force-generating mode.Region III: Around SLy (yield point). Tension reaches a

local maximum at yield point. This arises at least in part fromthe dislodging of titin thick filament anchorage by the sub-stantial resting tension. Alternatively, the unfolding andextension of the titin segment begins to generate less addi-tional tension (i.e., with a smaller modulus).Region IV: Beyond SLy (post-yield point). Tension levels

off or drops slightly. This reduction may be caused by theextension of an additional segment newly recruited from theA-band region after the dislodging of anchorage. Alterna-tively, it may be caused by a less effective force-generatingmode of extension or partial breakage of titin. The observa-tion that, in post-yield point sarcomeres, titin epitope (RT13)translocated with a smaller Ab-Z/SL ratio (Fig. 3) can bemore easily accounted for by the recruitment of an extralength of titin on the A-band side of the epitope.

This working hypothesis proposes that the length change ofan extensible titin segment is a simple and powerful concept

SLoI slo, noOfoCe

Si. .|eX~te~llisill. f'orlce.-genleratinig

/dlislodg, ancl>liehor!age, force-lve.X 1119

recruited A-segmeE t

2 3 - - (I.tortd erni

!I extensih.ie segniliit strain TTE ,IF

11 ITI:V

/v

SLo SLU SLYSaircornere Ie t h1 iI -11

FIG. 5. Segmental extension of titin filaments as the structuralbasis of stress-strain curves of resting muscles. The postulatedstructural events of the titin filaments that underlie the stress-straincurves are shown. The titin filament that extends from the Z line to theM line consists of two mechanically distinct segments: an extensiblesegment in the I band (open symbols) and an inextensible segment thatis constrained by interaction with thick filaments (solid symbols). Atslack sarcomere length (SL0) titin may be flaccid and stretching to SL,causes no net change in contour length and generates no significantforce. Beyond SLe, a linear extension of titin segment generates anexponential increase in tension. At SLY, the extensible segmentbecomes longer by recruiting previously inextensible titin when itsanchorage to thick filaments starts to slip or when distal ends of thickfilaments become distorted, resulting in the leveling of tension. Thenet contour length of the extensible segment of titin is thought to belonger in a sarcomere expressing a larger titin isoform. As a conse-quence, SL, and SLY increase and the stress-strain curves of variousmuscles can be plotted as a function of extensible segmental strain oftitin (TE/TE.). Note that for all six muscles studied here, SLe occursat a segmental strain around 2 and SLY occurs around 4.

... ....... _ . .. _ _ _ _ _ . _ . .D

ownl

oade

d by

gue

st o

n M

ay 6

, 202

0

Biophysics: Wang et al.

to understand the stress-strain curves, including the shape,its tissue and isoform specificity, as well as the phenomenonof yielding that underlies the reduction of elastic modulus athigh sarcomere strain.The analysis described above assumes that the contour

length of the extensible segment is the longest for the muscleexpressing the longest titin isoform, because the length oftheanchored segment, which is thought to be the same as the halfA-band width, does not vary among these muscles. Thedegree of extension of the titin segment (extensible segmentstrain) can be estimated by the ratio of segment length at agiven SL (TE) to that of the resting sarcomere (TEO), whereTE = 0.5 (SL - A-band width) and TEO = contour length oftitin - 0.5 A-band width. Since the linear mass of titin is =2.5MDa/Am (17), the contour length of titin isoforms is =1.1 &m(AM and PS), =1.13 Am (LD and SA), and =1.16Am (SO andST). It follows that the contour length of the extensiblesegment (TEO) is 0.3 ,um (AM and PS), 0.33 pm (LD and SA),and 0.37 pm (SO and ST). It is worth noting that a smallrelative difference in total length of titin is amplified >2-foldinto a larger difference in the contour length of the extensiblesegments. This difference, in turn, is manifested as propor-tionally longer SLe and SLY values of the stress-strain curve.It is significant that for all muscles in this group, TE/TE. is2 at SLe and 4 at SLY. Thus, these curves can be scaled to ageneric form by plotting resting tension vs. extension ratio(TE/TEO) as shown in Fig. 5.The same notion of segmental extension is applicable to the

understanding ofthe behavior ofpost-yield point sarcomeres.In this region, the extensible segment is now longer, corre-sponding to the segment of pre-yield point sarcomeres plus aportion of the originally anchored segment. Consequently,the extension ratio of titin in a post-yield point sarcomeresegment is smaller than that of a pre-yield point sarcomere atthe same SL. If the intrinsic extensibility of titin filamentalong its length is similar, then the resulting tension would belower than expected or even level off (depending on theelastic modulus and the net length of the recruited segment).In addition, this would lead to a reduction in stiffness and analtered stress-strain profile when the post-yield point fiber isreleased and then stretched again. Both predicted conse-quences were indeed observed (unpublished observations).While the extension ratio ofthe titin segment appears to be

a determining factor in shaping the stress-strain curve, it isthe intrinsic extensibility and force-generating capability(i.e., elastic modulus) of each titin isoform that determinesthe magnitude of resting tension. This is demonstrated by-ourobservation that yield point tension per thick filament andelastic modulus (stiffness) for rabbit PS is twice that of ST.Thus, titin isoforms, at least this pair, differ in- intrinsicmolecular elasticity. This observation also explains why it isso much easier to stretch ST than PS to nonoverlap; ST

Proc. Natl. Acad. Sci. USA 88 (1991) 7105

displays a lower resting tension and reaches its elastic limitonly at a length significantly beyond thick and/or thin fila-ment overlap (SLy > SLO,; Fig. 4), whereas PS has a hightension and reaches yield point near the nonoverlap SL (SLy= SLO,) where thick filament distortion ensues (unpublishedobservations). A corollary of this observation is that if yieldpoint tension indeed represents the strength of titin thickfilament anchorage, then anchorage strength is weaker forST. Curiously, despite the difference, the anchorage strengthis just sufficient to allow yield point to occur at the sameextension ratio of 4 for both muscles.

It is interesting that rabbit CA titin has a molecular mass of2.4-2.5 MDa (Fig. 1) and, if it behaves in a fashion analogousto its skeletal muscle counterpart, the CA sarcomere wouldhave a predicted yield point around 2.6-3.0 gm. We notedthat for rat CA myocytes, an elastic limit around 2.4-2.6 Amhas been reported (18, 19). Therefore, the segmental exten-sion model may be applicable to cardiac muscle as well.

We thank Gustavo Gutierrez for gel analysis of muscle proteinsand Henk Granzier for critical reading of the manuscript. This workwas supported in part by a grant from the Foundation for Research(to K.W. and R.M.) and by National Institutes of Health GrantDK20270 (to K.W.).

1. Cooke, R. (1986) CRC Crit. Rev. Biochem. 21, 53-118.2. Magid, A. & Law, D. J. (1985) Science 230, 1280-1282.3. Higuchi, H. & Umazume, Y. (1985) Biophys. J. 48, 137-147.4. Wang, K. (1985) Cell Muscle Motil. 6, 315-369.5. Maruyama, K. (1986) Int. Rev. Cytol. 104, 81-114.6. Wang, K., Wright, J. & Ramirez-Mitchell, R. (1984) J. Cell

Biol. 99, 435 (abstr.).7. Furst, D. 0, Osborn, M., Nave, R. & Weber, K. (1988) J. Cell

Biol. 106, 1563-1572.8. Itoh, Y., Suzuki, J., Kimura, S., Ohashi, K., Higuchi, H.,

Sarvada, H. & Shimizu, T. (1988) J. Biochem. 104, 504-508.9. Whiting, A., Wardale, J. & Trinick, J. (1989) J. Mol. Biol. 208,

263-267.10. Salvaiati, G., Betto, R., Ceoldo, S. & Pierobon-Bormiolo, S.

(1990) Am. J. Physiol. 56, 144-149.11. Horowits, R., Kempner, E. S., Bishert, M. E. & Podolsky,

R. J. (1986) Nature (London) 323, 160-164.12. Yoshioka, T., Higuchi, H., Kimura, S., Ohashi, K., Umazume,

Y. & Maruyama, Y. (1986) Biomed. Res. 7, 181-186.13. Horowits, R. & Podolsky, R. J. (1988) Biophys. J. 54, 165-171.14. Hu, D. H., Kimura, S. & Maruyama, K. (1985) J. Biochem. 99,

1485-1492.15. Wang, K. & Wright, J. (1988) J. Cell Biol. 107, 2199-2212.16. Kurzban, G. P. & Wang, K. (1988) Biochem. Biophys. Res.

Commun. 150, 1155-1161.17. Hainfeld, J. F., Wall, J. S. & Wang, K. (1988) FEBS Lett. 234,

145-148.18. Iwazumi, T. (1989) in Cardiac Mechanics and Function in the

Normal and Diseased Heart, eds. Homi, M., Suga, H., Baan,J.-& Yellin, E. L. (Springer-Verlag, Tokyo), pp. 13-22.

19. Roos, K. P. & Brady, A. J. (1991) Biophys. J. 59, 422 (abstr.).

Dow

nloa

ded

by g

uest

on

May

6, 2

020