Targeted deletion of titin N2B region leadsto diastolic dysfunction and cardiac atrophyMichael H. Radke*, Jun Peng†, Yiming Wu†, Mark McNabb†, O. Lynne Nelson‡, Henk Granzier†,and Michael Gotthardt*†§

*Department of Neuromuscular and Cardiovascular Cell Biology, Max-Delbruck-Center for Molecular Medicine, D-13122 Berlin-Buch, Germany;and †Department of Veterinary and Comparative Anatomy, Pharmacology, and Physiology, and ‡Department of Veterinary Clinical Sciences,Washington State University, Pullman, WA 99164

Edited by Christine E. Seidman, Harvard Medical School, Boston, MA, and approved December 20, 2006 (received for review September 27, 2006)

Titin is a giant protein that is in charge of the assembly and passivemechanical properties of the sarcomere. Cardiac titin contains aunique N2B region, which has been proposed to modulate elas-ticity of the titin filament and to be important for hypertrophysignaling and the ischemic stress response through its bindingproteins FHL2 and �B-crystallin, respectively. To study the role ofthe titin N2B region in systole and diastole of the heart, wegenerated a knockout (KO) mouse deleting only the N2B exon 49and leaving the remainder of the titin gene intact. The resultingmice survived to adulthood and were fertile. Although KO heartswere small, they produced normal ejection volumes because of anincreased ejection fraction. FHL2 protein levels were significantlyreduced in the KO mice, a finding consistent with the reduced sizeof KO hearts. Ultrastructural analysis revealed an increased exten-sion of the remaining spring elements of titin (tandem Ig segmentsand the PEVK region), which, together with the reduced sarcomerelength and increased passive tension derived from skinned cardi-omyocyte experiments, translates to diastolic dysfunction as doc-umented by echocardiography. We conclude from our work thatthe titin N2B region is dispensable for cardiac development andsystolic properties but is important to integrate trophic and elasticfunctions of the heart. The N2B-KO mouse is the first titin-basedmodel of diastolic dysfunction and, considering the high preva-lence of diastolic heart failure, it could provide future mechanisticinsights into the disease process.

cardiac muscle � hypertrophy � mechanics � cardiology � disease

T itin forms a continuous filament along the myofibril thatdetermines the elastic properties of cardiac myocytes (for

review, see ref. 1). The extensible region of titin is found in theI-band region of the sarcomere and comprises tandemly ar-ranged Ig-like domains and the so-called PEVK region (2). Inaddition, cardiac titin contains a third extensible region, the N2Belement (2), which is absent in skeletal muscle. The N2B regionextends greatly toward the upper limit of the physiologicalsarcomere length of cardiac muscle (3, 4). It has been suggestedthat this extension reduces the steepness of the passive force–sarcomere length relation, decreasing the likelihood of theunfolding of Ig domains (3). Mutations in the N2B region canlead to dilated or hypertrophic cardiomyopathy, apparentlythrough altered affinity to FHL2, a heart-specific member of theLIM domain gene family (5). To understand the role of the titinN2B region in cardiac function and disease, we have eliminatedexon 49, which encodes the N2B region, and investigated itseffect on the mechanical and trophical properties of the knock-out (KO) heart.

ResultsN2B-Deficient Titin Integrates Properly into the Sarcomere. Usinghomologous recombination we replaced titin exon 49 with an Flprecombinase target (FRT)-flanked neomycin resistance cassettethat was subsequently removed by germ line expression of theFlp recombinase (Fig. 1A). Homozygous KO mice survive to

adulthood and are fertile, with no obvious abnormalities. PCR,Southern blotting, and protein gels confirmed the deletion of theN2B region (Fig. 1 B–D). We studied titin expression in knock-out mice by Western blotting and found that upon excision of theN2B region, the reading frame is maintained and that theC-terminal M-line region is included in the truncated protein(Fig. 2A). The N2B-deficient protein is integrated properly intothe sarcomere as shown by immunofluorescence labeling withthe anti-M-line-titin antibody (Fig. 2 B and C). Overall, there wasno phenotypic change in the initial molecular characterization ofthe N2B-KO aside from a minor but significant shift in titinisoform expression from the smaller (stiffer) N2B to the larger(more compliant) N2BA isoform [see supporting information(SI) Fig. 7].

Reduced Heart Size and Impaired Diastolic Function in the Absence ofthe Titin N2B Region. Functional analysis of the N2B-knockoutanimal with echocardiography revealed that KO mice hadsignificantly reduced ventricular dimensions (reduced internaldiameter, diastolic volume, and calculated LV/body weightratio); see Table 1. Interestingly, fractional shortening wasincreased in the KO mice (by 22% relative to littermatecontrols), resulting in a stroke volume that was not signifi-cantly different from that of wild-type mice (WT). Thereduced heart to body weight ratio calculated from echo datawas consistent with actual weight measurements (SI Fig. 8).

Diastolic function was evaluated by using Doppler imaging ofmitral inflow (Table 2). In KO animals, we found a significantreduction in deceleration time (MV DT) and an increased in theE/A ratio, indicating a restrictive filling pattern.

Diastolic Wall Stress Is Increased in N2B-KO Hearts. To study dia-stolic and systolic function under controlled conditions, weperformed isolated heart experiments. Representative wallstress–volume diagrams from WT and KO animals are pro-vided in Fig. 3A. Although developed wall stress does notchange significantly (circles), the diastolic wall stress (squares)is increased in the KO, especially at high volumes. Statisticalanalysis demonstrates a robust increase in diastolic wall stress

Author contributions: M.H.R., J.P., H.G., and M.G. designed research; M.H.R., J.P., Y.W.,M.M., and O.L.N. performed research; M.H.R., J.P., Y.W., M.M., O.L.N., H.G., and M.G.analyzed data; and H.G. and M.G. wrote the paper.

The authors declare no conflict of interest.

This article is a PNAS direct submission.

Abbreviations: ANP, atrial natriuretic peptide; FRT, Flp recombinase target; KO, knockout;LV, left ventricular; MV DT, mitral valve deceleration time.

§To whom correspondence should be addressed at: Department of Veterinary and Com-parative Anatomy, Pharmacology, and Physiology, Washington State University, WegnerHall, Room 205, Pullman, WA 99164-6520 or Max-Delbruck-Center for Molecular MedicineBerlin-Buch, Robert Rossle Strasse 10, 13122 Berlin, Germany. E-mail: gotthard@vetmed.wsu.edu or gotthardt@mdc-berlin.de.

This article contains supporting information online at www.pnas.org/cgi/content/full/0608543104/DC1.

© 2007 by The National Academy of Sciences of the USA

3444–3449 � PNAS � February 27, 2007 � vol. 104 � no. 9 www.pnas.org�cgi�doi�10.1073�pnas.0608543104

Dow

nloa

ded

by g

uest

on

Mar

ch 2

9, 2

020

in KO versus WT mice (P � 0.006, n � 18) (Fig. 3B Left).Similar results were found in the presence of dobutamine orpropranolol (Fig. 3B Center and Right). Systolic function wasnormal in KO mice, and no significant differences were foundin developed wall stress of WT and KO mice at all volumestested (Fig. 3 C and D). Dobutamine increased developed wallstress of hearts of all genotypes with a trend toward a largerincrease in the KO (Fig. 3D; P � 0.08, n � 18).

Altered Structure and Mechanical Properties of the N2B-DeficientSarcomere. To investigate the molecular mechanism underlyingaltered diastolic function in N2B-KO animals, we studied themechanical and structural properties of skinned cardiomyocytes(Fig. 4). Although the myocyte width, length, cross-sectional

area, and maximal active tension were unchanged in myocytes ofN2B-KO mice, slack sarcomere length was reduced significantly(Fig. 4E). Furthermore, passive tension of cardiac myocytes,which is known to be largely the result of titin (6), was increasedsignificantly (Fig. 4F).

To understand the structural basis of increased titin-basedpassive tension, we investigated how the absence of the N2Belement affects the extensibility of the remaining springelements of titin (Fig. 5). We demarcated those elements withantibodies by using immunoelectron microscopy (compare Fig.5B). The distance of titin epitopes from the Z-disk was plottedagainst sarcomere length (Fig. 5C) to deduce the extension ofthe proximal tandem Ig segment (T-12 to UN), the N2B-USsequence only present in WT mice (UC to UN), the PEVK(UC-I84), and the distal tandem Ig segment (I84 to MIR). Thesegment lengths at a sarcomere length of 2.3 �m revealed thatexcision of the N2B element results in an increased extensionof the tandem Ig segments (both proximal and distal) as wellas the PEVK region (Fig. 5D). The PEVK region, althoughshortest, extends the most in response to the loss of N2Belasticity, followed by the proximal Ig region and the distal Ig(Fig. 5E).

Overall, the differences in sarcomere slack length, extensibil-ity, and increase in passive tension of N2B-deficient cardiomy-ocytes provide the biomechanical framework to explain theimpaired diastolic function and increased diastolic wall stress inthe KO animals.

Reduced Cardiac Growth Is Associated with Decreased FHL2 ProteinLevels. To address the molecular basis of the reduced size of KOhearts, we determined protein levels for the N2B-bindingproteins �B-crystallin and FHL2. Only the latter has beenlinked to atrophy/hypertrophy signaling (5, 7), and as expected,we found �B-crystallin unchanged but FHL2 protein down-regulated in the KO (Fig. 6 A–C). Although RNA levels weresimilar in KO and WT animals (Fig. 6D), FHL2 protein isreduced in N2B-KO mice. Finally, because trophic changes inthe FHL2-KO are accompanied by an increased mRNA levelof the hypertrophy marker atrial natriuretic peptide (ANP)(7), we also measured ANP expression in N2B-KO mice.

Fig. 2. The exon 49-KO affects exclusively the N2B region with properintegration of the truncated protein into the sarcomere. (A) Western blot.India ink staining shows equal loading for WT (�/�) and KO (�/�). Z1-Z2antibody recognizes the N terminus of titin; the N2B-US antibody, the uniqueN2B sequence; and the M8/M9 antibody, the C terminus of titin. In KO mice,only the N2B region is deleted, whereas the N- and C- terminal parts of the titinprotein are unaffected. (B) Confocal microscopy of double-immunofluores-cence-stained cardiomyocytes. Anti-�-actinin (red), labels the Z-disk, and anti-N2B (green) localizes at the I-band close to the Z-disk in WT animals (�/�). InKO mice, the N2B staining is absent. (C) Costaining with anti-�-actinin (red)and the anti-M8/M9 (green) antibody directed against the M-band region oftitin shows proper integration of N2B-deficient titin into the sarcomere(alternating actinin/M-band titin staining). (Scale bar, 5 �m.)

Fig. 1. Generation of titin N2B-deficient animals. (A) Targeting strategy. The neomycin resistance cassette in the targeting vector replaces the N2B region (exon49) in the WT allele. The exon encodes three Ig domains and a unique sequence. The Flp recombinase was used to remove the neomycin cassette. The Southernprobe is shown as a black box. Restriction sites for Southern blot analysis are EcoRI (E) and BamHI (B). The FRT sites are depicted as gray arrowheads, andprimer-binding sites for PCR genotyping are shown as black arrows. (B) PCR genotyping. PCR products derived from the WT allele (�) are 376 bp; targeted allele(neo), 210 bp; KO allele (�), 252 bp. (C) Genomic Southern blot of genomic DNA from the same mice analyzed in B. Digestion of DNA from WT (�/�), heterozygous(neo/� or �/�), and homozygous (�/�) KO animals with EcoRI and BamHI produces bands of the expected sizes (�, 5.2 kb; neo, 8.2 kb; �, 6.4 kb). (D) Coomassieblue-stained 2% agarose gel with ventricular lysates from �/�, �/�, and �/� animals. N2BA, N2B isoforms, and T2 degradation products are indicated. In theheterozygote, the WT and the mutant proteins (higher mobility) are expressed. The KO expresses only truncated N2B and N2BA titin (tN2BA and tN2B).

Radke et al. PNAS � February 27, 2007 � vol. 104 � no. 9 � 3445

MED

ICA

LSC

IEN

CES

Dow

nloa

ded

by g

uest

on

Mar

ch 2

9, 2

020

Indeed, we found that ANP was also up-regulated in theN2B-KO (Fig. 6E).

DiscussionThe cardiac-specific N2B element integrates mechanical andsignaling functions through its large extensible region and itsinteraction with �B-crystallin and FHL2. Analysis of theN2B-KO mice allowed us to test the hypothesis that the N2Bunique sequence eliminates the necessity for unfolding of Igdomains under physiological conditions (3). In absence of theN2B element, we found that the extension of PEVK and tandemIg segments is increased (Fig. 5). At the upper limit of thephysiological sarcomere length range, the increased extension isthe highest for the PEVK region, suggesting that at this sarco-mere length the PEVK is most compliant. Interestingly, the extraextension of the tandem Ig segment is distributed unevenlybetween the proximal and distal tandem Ig segment, highest inthe proximal Ig segment (Fig. 5D), which leads to the conclusionthat the proximal tandem Ig segment is more compliant than thedistal tandem Ig, consistent with in vitro studies (3). Thus, ourwork supports the view that the extensibility provided by the N2Bunique sequence obviates the need for the unfolding of Igdomains toward the upper limit of the physiological sarcomerelength range.

Fig. 3. N2B-deficient hearts have increased diastolic wall stress. (A) Examplesof � � V relation in WT (Left) and KO mice (Right). Diastolic stress is shown bysquares and developed stress by circles. (B) Mean � SEM of WT (light gray) andKO (dark gray) mice (six animals each). Results are shown at Veq � 15%, underbaseline conditions (BL) and in the presence of either dobutamine (Dob) orpropranolol (Prop). Under all conditions, diastolic wall stress is increased in KOmice (P � 0.01). (C and D) Developed wall stress at Veq and Veq � 15%.Developed wall stress is largely unchanged between the genotypes. There isa trend toward higher developed � in the presence of dobutamine, which didnot reach statistical significance (P � 0.08). The corresponding pressures areprovided in SI Fig. 9.

Table 2. Doppler of N2B-KO hearts

MV AT, ms MV DT, msIV RT (LV),

ns ET (LV), ms MV E:A

N2B�/� 14.4 � 1.0 23.8 � 2.2 14.4 � 1.0 45.4 � 2.2 1.37 � 0.14N2B�/� 15.5 � 1.5 17.4 � 1.2* 12.4 � 0.8 44.8 � 1.4 1.90 � 0.15*

MV AT, mitral valve acceleration time; MV DT, mitral valve decelerationtime; IV RT (LV), left ventricular isovolumic relaxation time; ET (LV), leftventricular ejection time; MV E:A, mitral valve E-wave to A-wave ratio (�/�,n � 8; �/�, n � 9). *, P � 0.05 versus N2B�/� value.

Tab

le1.

Ech

oca

rdio

gra

ph

yo

fN

2B-K

Oh

eart

s

BW

,gH

R,b

pm

IVS,

d,m

mIV

S,s,

mm

LVPW

,d,

mm

LVPW

,s,

mm

LVID

,d,m

mLV

ID,s

,mm

Vo

l,d,�

lV

ol,s

,�l

StrV

ol,

�l

LVEF

,%LV

FS,%

LVW

c/B

W,

mg

/g

N2B

�/�

27.1

�0.

948

0�

130.

78�

0.03

1.18

�0.

050.

83�

0.03

1.13

�0.

044.

05�

0.14

2.88

�0.

1765

.2�

4.7

29.6

�3.

435

.6�

2.9

56.0

�3.

329

.2�

2.1

3.6

�0.

2N

2B�

/�28

.8�

0.9

466

�15

0.79

�0.

031.

24�

0.04

0.77

�0.

041.

09�

0.05

3.72

�0.

08*

2.40

�0.

09*

51.5

�2.

1*18

.5�

1.4*

32.9

�1.

665

.5�

2.6*

35.6

�1.

9*2.

8�

0.1*

BW

,bo

dy

wei

gh

t;H

R,h

eart

rate

;IV

S,d

,in

terv

entr

icu

lar

sep

tum

ind

iast

ole

;IV

S,s,

inte

rven

tric

ula

rse

ptu

min

syst

ole

;LV

PW,d

,lef

tve

ntr

icu

lar

po

ster

ior

wal

lin

dia

sto

le;L

VPW

,s,l

eft

ven

tric

ula

rp

ost

erio

rw

alli

nsy

sto

le;L

VID

,d,l

eftv

entr

icu

lari

nn

er-d

ista

nce

dia

sto

le;L

VID

,s,l

eftv

entr

icu

lari

nn

er-d

ista

nce

syst

ole

;Vo

l,d,c

ham

ber

volu

me

ind

iast

ole

;Vo

l,s,c

ham

ber

volu

me

insy

sto

le;S

trV

ol,

stro

kevo

lum

e;LV

EF,l

eftv

entr

icu

lar

ejec

tio

nfr

acti

on

;LV

FS,l

eft

ven

tric

ula

rfr

acti

on

alsh

ort

enin

g;L

VW

c/B

W,c

alcu

late

dle

ftve

ntr

icu

lar

wei

gh

tp

erb

od

yw

eig

ht

(n�

9). *

,P�

0.05

vers

us

N2B

�/�

valu

e.

3446 � www.pnas.org�cgi�doi�10.1073�pnas.0608543104 Radke et al.

Dow

nloa

ded

by g

uest

on

Mar

ch 2

9, 2

020

The passive force of titin is entropic in nature, with forceincreasing with the titin fractional extension (8). Thus, theincreased extension of the tandem Ig segment and PEVK regionof the N2B-KO mice will result in a larger fractional extensionand hence a higher passive force, consistent with our measure-ments made on cardiac myocytes (Fig. 4). Considering thatwithin the physiological sarcomere length range titin is the maincontributor to passive tension of the myocardium (9), increasedpassive tension of cardiac myocytes is a likely explanation of theelevated diastolic LV wall stress that was found in the isolatedheart experiments (Fig. 3). The reduced slack length of cardiacmyocytes of KO mice might also relate to the excision of one ofthe spring elements of titin. The mean square end-to-enddistance of a flexible chain at zero external force (titin in slacksarcomeres) is a function of the contour length of the chain.Thus, a reduction in contour length (N2B-KO) will result in areduction of the end-to-end distance and hence a reduction inslack sarcomere length. It has been speculated for more than adecade that titin is a determinant of slack length (10), and ourwork provides direct evidence thereof.

Increased diastolic wall stress explains the diastolic dysfunc-tion of KO mice that was revealed by echocardiography. KOmice had a reduced deceleration time of the E wave (early rapidfilling phase) and a restrictive filling pattern as revealed by thereduced ratio of the peak of E wave and peak of A wave (latefilling because of atrial systole). Because MV DT is inverselycorrelated with LV stiffness in both animals (11) and humans(12), the reduction in MV DT of N2B-KO mice can be ascribedto their increased LV stiffness. Atrial contraction will contributeless to the filling of a stiff ventricle explaining the reduced E:Aratio of KO mice. Thus, echocardiography reveals that N2B-KOmice have diastolic dysfunction as a result of a stiff ventricle. The

slight increase in expression of N2BA titin N2B-KO mice (SI Fig.7) may be viewed as an attempt to improve diastolic function [theN2BA isoform is less stiff than N2B titin (ref. 13)]. However, theensuing reduction in passive tension is small (�5%) and appar-ently insufficient to offset the large increase in stiffness thatresults from the excision of the N2B element.

In the N2B-KO mouse we did not find major changes insystolic function. Maximal active tension of skinned cardiacmyocytes was not different in KO mice (Fig. 4D), and theisolated heart experiments did not reveal significant systolicchanges (Fig. 3 C and D). There was a trend toward an increaseddobutamine response in the N2B-KO (Fig. 3D), but it did notreach statistical significance (P � 0.08). Interestingly, the echo-cardiography study revealed an increased fractional shorteningin KO mice (Table 1). Titin-based passive tension has beensuggested to increase calcium sensitivity (1), but this effect onlytakes place at long sarcomere lengths and, thus, is unlikely to beinvolved in prolonging the ejection phase (where sarcomeres areshort). Future studies are required to establish the molecularmechanism underlying increased fractional shortening in KOmice. Assuming that the intrinsic myocardial tension duringsystole is indeed not altered in the KO mice, their smallerventricles could result in a higher ventricular pressure, whichmight help explain their increased ejection fraction, althoughother factors could be involved as well.

Previous work on the N2B region in zebrafish by using botha genetic mutant (pickm171) and a morpholino approach todisrupt expression of the N2B exon resulted in sarcomeredisassembly and impaired systolic function (14). This outcomeis in contrast to the N2B-KO mice, which survive to adulthoodwith no detectable defect in sarcomere assembly or systolicfunction. Because the pickm171 mutation generates a stopcodon, the resulting truncated protein lacks not only a func-tional N2B region but also the M-line region. Our previouswork on titin in sarcomere assembly indicated a critical role forthe M-line region of titin (15, 16). Thus, the lack of the M-lineregion in pickm171 might explain their severe phenotype. Themorpholino approach resulted in the same phenotype as thepickm171 mutant. It remains to be examined whether thisoutcome is because of a lack of exon specificity (i.e., whetherthe M-line region is absent) or whether there is a true speciesdifference. In the mouse, the N2B region does not appear toplay a major role in systolic function.

The N2B region of titin not only determines the elasticproperties of the sarcomere, it also relates to the ischemic stressand cardiac hypertrophy responses through its binding proteins�B-crystallin (17) and FHL2 (18), respectively. We did not findchanges in �B-crystallin expression in KO mice, suggesting thatinteraction between �B-crystallin and the N2B region of titin isnot required for normal cardiac function.

The previously published FHL2-KO (7) and our N2B-KOshare the impaired balance of hypertrophy and atrophy. Wefound that FHL2 protein was down-regulated upon loss of itsbinding site within the N2B region of titin. This down-regulation occurs at the posttranslational level because mRNAlevels are unchanged (Fig. 6). Both altered FHL2 protein levelsand the relation of FHL2 and N2B to cardiac growth make ittempting to speculate that the interaction of N2B and FHL2provides a novel regulatory pathway to control the balance ofhypertrophy and atrophy in the heart. Consistent with thisproposal is the recent finding that patients with a mutation inthe titin N2B region (S3799Y) develop hypertrophic cardio-myopathy associated with an increased binding of FHL2 totitin (5).

In summary, we used a KO approach to establish the N2Bregion as a critical determinant of the elastic properties of theheart and cardiac growth. The N2B-KO mouse holds the po-tential for studying interaction between titin and FHL2 and as a

Fig. 4. Slack sarcomere length (SL) is reduced, and passive tension is in-creased in skinned cardiomyocytes of N2B-KO mice. Myocyte length (A), width(B), cross-sectional area (C), and maximal force produced (D; tested at twosarcomere lengths) are unchanged in N2B-KO (�/�) versus WT (�/�) controlanimals. (E) Slack sarcomere length is significantly reduced in KO versus WTcardiomyocytes (n � 30 each). (F) Passive tension is increased in KO versus WTat sarcomere length �2 �m. Myocytes were derived from age-matched KO(n � 11) and littermate control WT animals (n � 10) at 6 months of age.

Radke et al. PNAS � February 27, 2007 � vol. 104 � no. 9 � 3447

MED

ICA

LSC

IEN

CES

Dow

nloa

ded

by g

uest

on

Mar

ch 2

9, 2

020

tool to develop therapeutic strategies to combat diastolicdysfunction.

MethodsGeneration of Titin N2B-KO Mice (Titin N2B�/�). A targeting con-struct was assembled to replace exon 49 (N2B exon) with aFRT-flanked neomycin expression cassette. Titin N2B (�/neo)animals were generated from two independent ES cell cloneswith subsequent removal of the neomycin cassette by using theFlp deleter stain (19). For more details on generation andgenotyping of KO mice, see SI Methods.

All experiments involving animals were carried out followinginstitutional and National Institutes of Health guidelines (20).

Echocardiography. For echocardiography we used the Vevo 770system (VisualSonics Inc., Toronto, ON, Canada) with a45-MHz transducer mounted on an integrated rail system.Standard imaging planes, M-mode, Doppler, and functionalcalculations were obtained according to American Society ofEchocardiography guidelines. The LV parasternal long-axisfour-chamber view was used to derive fractional shortening,ejection fraction, and ventricular dimensions and volumes. Thesubcostal long-axis view from the left apex was used forDoppler imaging of mitral inf low and aortic ejection profiles.An extended description of the procedure is available in SIMethods.

Isolated Heart Experiments. The developed and passive wall stressto volume relationship (� � V) was determined by using theisolated heart preparation with a single-beat Frank–Starlingprotocol. The Frank–Starling protocol was run first in normalTyrode solution, then the response to �-adrenergic stimulation(0.2 �M dobutamine) and �-adrenergic blockade (0.1 �M

propranolol) was determined. Details are available in SIMethods.

Myocyte Mechanics. Cardiac myocytes were isolated from mouseLV as described previously (6). Passive tension–sarcomerelength curves were obtained as detailed in ref. 21. Brief ly,myocytes were added to a temperature-controlled f low-through chamber mounted on the stage of a phase-contrastmicroscope. One end of a single cell was glued to a motor. Thefree end was then bent with a micromanipulator so that themyocyte axis aligned with the microscope optical axis, and themyocyte cross-sectional area was measured. Finally, the freeend of the myocyte was glued to a force transducer. Sarcomerelength was measured by laser diffraction. Cells were activatedwith a pCa 4.5 (pCa adjusted with CaCO3) activating solutionto verify cell quality, and then passive tension was measured inrelaxing solution (6). Experiments were conducted at roomtemperature.

Gel Electrophoresis and Western Blot Analysis. Protein samplesfrom LVs were prepared by homogenization in liquid N2 andlysis in 8 M urea/50% (vol/vol) glycerol/DTT (80 mM final)/protease inhibitors (leupeptin, E-64, PMSF). Titin isoformswere separated by using an SDS/agarose gel electrophoresissystem followed by Coomassie blue staining (22). FHL2 and�B-crystallin were separated on 12% SDS/polyacrylamidegels. For Western blot analysis, samples were transferred tonitrocellulose membranes (GE Healthcare, Piscataway, NJ)for titin or Hybond P PVDF membranes (GE Healthcare).Membranes were blocked with 5% skim milk in PBS-Tween 20(PBS-T) followed by incubation with antibodies: anti-Z1-Z2,UN, N2B Us, UC, and M8-M9 (all generous gifts from S.Labeit), and Vectastain (Vector Laboratories, Burlingame,CA) avidin/biotin–AP complex 1:10,000 was used for detec-

Fig. 5. Compensation for loss of the N2B element by the PEVK and tandem Ig regions as determined by immunoelectron microscopy. (A) WT (�/�) and KO(�/�) LVs were colabeled with anti-N2B and anti-MIR antibody. The N2B labeling is only observed in the �/� and absent in the �/� animal. (B) Antibody-bindingsites are shown along the titin I-band region. (C) Measurement of epitope distance from the Z-line at various sarcomere lengths tested on �/� and �/� musclestrips. (D) Segment extension at sarcomere length 2.3 �m. Tandem Ig and PEVK segments are increased in length in KO mice. (E) Increased extension of tandemIg and PEVK segments in KO mice as a percentage of extension of I-band titin. Loss of the N2B element results in the largest extension of PEVK segment followedby the proximal tandem Ig and the distal tandem Ig region.

3448 � www.pnas.org�cgi�doi�10.1073�pnas.0608543104 Radke et al.

Dow

nloa

ded

by g

uest

on

Mar

ch 2

9, 2

020

tion. For loading, control membranes were stained with Indiaink. Commercial antibodies against FHL2 (mouse monoclonal;MBL, Naga-Ku, Nagoya, Japan) or �B-crystallin (rabbit poly-

clonal; Calbiochem, San Diego, CA) were used according tothe manufacturer’s instructions and detected by using horse-radish peroxidase-conjugated secondary antibodies andchemiluminescence staining with ECL (Supersignal West PicoChemiluminescent Substrate; Pierce, Rockford, IL).

Real-Time PCR. RNA from three individual ventricles was isolatedand amplified with TaqMan probes for ANP, FHL2, and 18SRNA (for normalization) as described previously (16).

Immunofluorescence Staining and Immunoelectron Microscopy. Car-diomyocytes were seeded to laminin-coated glass coverslips (4h) and fixed with cold methanol. Primary antibodies were:mouse monoclonal anti-sarcomeric �-actinin (1:500; EA53;Sigma–Aldrich, St. Louis, MO) and the rabbit polyclonalantibodies anti-titin N2B (1:200) and anti-titin M8/M9 (1:200)(both gifts from Siegfried Labeit, Universitatsklinikum Mann-heim, Germany). Secondary f luorescent-conjugated second-ary antibodies were: Alexa Fluor 488 goat anti-rabbit(Molecular Probes Jackson ImmunoResearch, West Grove,PA). Stained cells were analyzed on a confocal scanning lasermicroscope (LSM 5 Pascal version 3.0 SP2; Karl Zeiss, Jena,Germany) with a PLAN-NEOFLUAR 100� lens (1.3 NA)(Karl Zeiss). Immunoelectron microscopy to localize titinepitopes at different sarcomere lengths has been describedpreviously (3). SI Methods contains a detailed version includ-ing references to the antibodies used.

Statistical Analysis. For statistical analysis, SPSS 11.0 software wasused. All results are expressed as means � SEM. An unpairedtwo-tailed t test was performed to assess differences between twogroups.

We thank Beate Goldbrich, Gemaine Wright, Joe Popper, and MartinTaube for expert technical assistance and Arnd Heuser for supportwith the Vevo System. This work was supported by American HeartAssociation Grant GIA 0450195Z (to M.G.), Deutsche Forschungs-gemeinschaft Grant GO 865/3 (to M.G.), the Sofja–Kovalevskayaprogram of the Alexander von Humboldt Foundation (to M.G.), andby National Institutes of Health Grant HL61487/HL62881 (to H.G.).

1. Granzier HL, Labeit S (2004) Circ Res 94:284–295.2. Labeit S, Kolmerer B (1995) Science 270:293–296.3. Trombitas K, Freiburg A, Centner T, Labeit S, Granzier H (1999) Biophys J

77:3189–3196.4. Linke WA, Rudy DE, Centner T, Gautel M, Witt C, Labeit S, Gregorio CC

(1999) J Cell Biol 146:631–644.5. Matsumoto Y, Hayashi T, Inagaki N, Takahashi M, Hiroi S, Nakamura T,

Arimura T, Nakamura K, Ashizawa N, Yasunami M, et al. (2005) J Muscle ResCell Motil 26:367–374.

6. Granzier HL, Irving TC (1995) Biophys J 68:1027–1044.7. Kong Y, Shelton JM, Rothermel B, Li X, Richardson JA, Bassel-Duby R,

Williams RS (2001) Circulation 103:2731–2738.8. Granzier H, Labeit S (2002) J Physiol 541:335–342.9. Wu Y, Cazorla O, Labeit D, Labeit S, Granzier H (2000) J Mol Cell Cardiol

32:2151–2162.10. Labeit S, Kolmerer B, Linke WA (1997) Circ Res 80:290–294.11. Ohno M, Cheng CP, Little WC (1994) Circulation 89:2241–2250.12. Garcia MJ, Firstenberg MS, Greenberg NL, Smedira N, Rodriguez L, Prior D,

Thomas JD (2001) Am J Physiol 280:H554–H561.13. Trombitas K, Redkar A, Centner T, Wu Y, Labeit S, Granzier H (2000) Biophys

J 79:3226–3234.

14. Xu X, Meiler SE, Zhong TP, Mohideen M, Crossley DA, Burggren WW,Fishman MC (2002) Nat Genet 30:205–209.

15. Gotthardt M, Hammer RE, Hubner N, Monti J, Witt CC, McNabb M,Richardson JA, Granzier H, Labeit S, Herz J (2003) J Biol Chem 278:6059–6065.

16. Weinert S, Bergmann N, Luo X, Erdmann B, Gotthardt M (2006) J Cell Biol173:559–570.

17. Bullard B, Ferguson C, Minajeva A, Leake MC, Gautel M, Labeit D, Ding L,Labeit S, Horwitz J, Leonard KR, et al. (2004) J Biol Chem 279:7917–7924.

18. Lange S, Auerbach D, McLoughlin P, Perriard E, Schafer BW, Perriard JC,Ehler E (2002) J Cell Sci 115:4925–4936.

19. Rodriguez CI, Buchholz F, Galloway J, Sequerra R, Kasper J, Ayala R, StewartAF, Dymecki SM (2000) Nat Genet 25:139–140.

20. National Institutes of Health Animal Advisory Committee (1999) UsingAnimals in Intramural Research: Guidelines for Investigators and Guidelines forAnimal Users (Natl Inst Health, Bethesda).

21. Helmes M, Trombitas K, Centner T, Kellermayer M, Labeit S, Linke WA,Granzier H (1999) Circ Res 84:1339–1352.

22. Warren CM, Krzesinski PR, Greaser ML (2003) Electrophoresis 24:1695–1702.

Fig. 6. FHL2 protein level is decreased in the N2B-KO, �B-crystallin is un-changed, and ANP expression is increased. (A) The titin N2B-binding proteinsFHL2 and �B-crystallin were quantified by Western blot analysis of WT (�/�) andKO (�/�) LVs (n � 3 each). Although �B-crystallin is unchanged, FHL2 levels arereduced significantly in KOs (�/�). (B and C) Quantification of FHL2 and �B-crystallin normalized to actin levels and (�/�) set to 100%. The difference issignificant at P � 0.01 (n � 6 per genotype). (D and E) FHL2 RNA levels asdetermined by real-time quantitative RT-PCR from left ventricle are unchanged(D), but the hypertrophy marker ANP (E) is up-regulated in KO animals (n � 3 pergenotype).

Radke et al. PNAS � February 27, 2007 � vol. 104 � no. 9 � 3449

MED

ICA

LSC

IEN

CES

Dow

nloa

ded

by g

uest

on

Mar

ch 2

9, 2

020