The FASEB Journal • Research Communication

Mutations in fast skeletal troponin I, troponin T, and�-tropomyosin that cause distal arthrogryposis allincrease contractile function

Paul Robinson,* Simon Lipscomb,† Laura C. Preston,*,† Elissa Altin,* Hugh Watkins,*Christopher C. Ashley,† and Charles S. Redwood*,1

Departments of *Cardiovascular Medicine and †Physiology, University of Oxford, Oxford, UK

ABSTRACT Distal arthrogryposes (DAs) are a groupof disorders characterized by congenital contracturesof distal limbs without overt neurological or muscledisease. Unexpectedly, mutations in genes encodingthe fast skeletal muscle regulatory proteins troponin T(TnT), troponin I (TnI), and �-tropomyosin (�-TM)have been shown to cause autosomal dominant DA. Wetested how these mutations affect contractile functionby comparing wild-type (WT) and mutant proteins inactomyosin ATPase assays and in troponin-replacedrabbit psoas fibers. We have analyzed all four reportedmutants: Arg63His TnT, Arg91Gly �-TM, Arg174GlnTnI, and a TnI truncation mutant (Arg156ter). Thinfilaments, reconstituted using actin and WT troponinand �-TM, activated myosin subfragment-1 ATPase in acalcium-dependent, cooperative manner. Thin fila-ments containing either a troponin or �-TM DA mutantproduced significantly enhanced ATPase rates at allcalcium concentrations without alternating calcium-sen-sitivity or cooperativity. In troponin-exchanged skinnedfibers, each mutant caused a significant increase inCa2� sensitivity, and Arg156ter TnI generated signifi-cantly higher maximum force. Arg91Gly �-TM wasfound to have a lower actin affinity than WT and forma less stable coiled coil. We propose the mutationscause increased contractility of developing fast-twitchskeletal muscles, thus causing muscle contractures andthe development of the observed limb deformities.—Robinson, P., Lipscomb, S., Preston, L. C., Altin, E.,Watkins, H., Ashley, C. C., Redwood, C. S. Mutations infast skeletal troponin I, troponin T, and �-tropomyosinthat cause distal arthrogryposis all increase contractilefunction. FASEB J. 21, 896–905 (2007)

Key Words: contractility � muscle disease

Distal arthrogryposes (das) are a group of disor-ders that are characterized by multiple congenitalcontractures of the distal limb joints producing defectssuch as club foot and camptodactyly, a flexion defor-mity of the finger, which results in a bent digit thatcannot be completely straightened. The contracturesare nonprogressive and occur frequently in the pres-ence of other morphological abnormalities but are not

associated with clinically apparent primary neurologicaland/or muscle disease that affects limb function (1).These syndromes are transmitted by autosomal domi-nant inheritance. Various attempts have been made toclassify the different forms of DAs based on the pres-ence or absence of associated abnormalities. Initially,six forms, labeled I and IIA-IIE were defined (1)whereas more recently ten different variants (1, 2A, 2B,3–9) have been identified (2, 3). Of relevance to thisstudy, distal arthrogryposis type 1 (DA1) occurs in1:10,000 to 1:50,000 births and is a common cause ofinherited clubfoot and vertical talus (4). Affected indi-viduals also have clenched hands at birth with ulnardeviation and digitotalar dysmorphism, and the shoul-ders and hips may also be affected. Type 2B (DA2B) issimilar to Freeman-Sheldon or whistling face syndrome(now classified as DA2A) and is characterized by dis-tinctive facial abnormalities along with campodactylyand vertical talus (5).

The precise cause of these malformations has beenunclear. It has been postulated that decreased fetalmovement brought about by a variety of means maylead to contractures at birth via the accumulation ofadditional connective tissue around the joints (6). Thereduced movement could be due to primary muscularor neuronal disorders, connective tissue abnormalities,or problems within the uterus, such as intrauterinevascular compromise (1, 6, 7). Recent studies havesought to identify the relevant disease genes. Bam-shad’s group identified the pericentromeric region ofchromosome 9 to be a disease locus for DA1 (8) andhas since shown that a missense mutation (Arg91Gly) inthe TPM2 gene within this region can cause this disease(9). TPM2 encodes the muscle regulatory protein �-tro-pomyosin, which is expressed in all skeletal muscles.Molecular genetic work has also shown that DA2B canbe caused by two mutations in TNNI2, a Arg174Glnmissense mutation and a C3T nonsense mutation,resulting in the loss of the C-terminal 26 amino acids

1 Correspondence: Department of Cardiovascular Medi-cine, University of Oxford, Wellcome Trust Centre ofHuman Genetics, Oxford OX3 7BN, UK. E-mail:credwood@well.ox.ac.uk

doi: 10.1096/fj.06-6899com

896 0892-6638/07/0021-0896 © FASEB

(Arg156ter), (9) and a Arg63His missense mutation inTNNT3 (10). These genes encode the fast skeletalisoforms of troponin I and troponin T, respectively,and hence it appears that alteration of the Ca2�-regulation of fast skeletal muscle can be the underlyingcause of these disorders.

Skeletal muscle contraction is principally controlledby the concentration of Ca2� ions surrounding themyofilaments. The Ca2� concentration signal is sensedby the trimeric troponin complex (subunits C, I and T)and transmitted to the �-helical coiled coil tropomyosindimer that lies along the actin filament (11, 12). Bothtroponin and the tropomyosin dimer are present atratios of 1:7 with actin. On activation of the muscle,troponin C binds Ca2� at its two low-affinity, regulatorysites; the subsequent conformational changes withinthe troponin complex result in both the inhibitoryeffect of troponin I being removed and a change in theposition of the tropomyosin, allowing productive myo-sin head interaction with actins not directly in contactwith troponin I and hence cooperative switch on of thethin filament. Kinetic measurements have suggestedthat this mechanism involves three states: blocked (inthe absence of Ca2�, tropomyosin prevents productivemyosin head binding), closed (the tropomyosin posi-tion moves in response to Ca2� binding to allow partialmyosin attachment), and myosin-induced (in whichmyosin head binding has shifted tropomyosin to exposefully the interaction sites on actin) (13–17). This hasbeen strongly supported by electron microscopy (EM)reconstruction data of thin filaments from differentmuscles in the presence and absence of Ca2� (18–20).

We have set out to understand how the subtlemutations that cause DA alter the control of skeletalmuscle contractility and how this may lead to thedisease state. The ability of recombinant mutant tropo-nin/tropomyosin to regulate reconstituted actomyosinATPase has been compared to WT, and the function ofDA mutant troponins has been tested in exchanged fastskeletal muscle fibers. Our data suggest that all fourreported mutations cause increased contractility andare likely to produce a “hypercontractile” muscle invivo. We suggest that this aberrant thin filament regu-lation in developing skeletal muscles causes contrac-tures, which may be responsible for the development ofthe observed limb defects.

MATERIALS AND METHODS

Expression and purification of human fast skeletal troponinand tropomyosin

Mammalian Gene Collection clones for TNNC2 (IMAGE#3934188), TNNI2 (IMAGE #5020339) and TPM2 (IMAGE#3640927) were obtained from the Medical Research CouncilGene Service (UK). Polymerase chain reaction (PCR) prod-ucts containing each complete coding sequence were sub-cloned using NdeI and HindIII into the bacterial expressionplasmid pMW172 (21). Nine base pairs encoding Met-Ala-Serwere inserted 5� to the initiator methionine codon of the

�-tropomyosin construct. A clone corresponding to TNNT3was generated by PCR using a skeletal muscle cDNA mixture(Clontech) as template and also cloned into pMW172 usingNdeI and HindIII. pMW172 expression constructs encodingeach of the reported DA mutations in troponin T, troponin I,and �-tropomyosin were made from the WT constructs usinga two-step PCR protocol for site-directed mutagenesis.

The pMW172 constructs were used to transform the E. colistrain BL21(DE3)pLysS and large-scale cultures were grownand overexpression was induced according to standard meth-ods (22). Human recombinant fast skeletal troponin T andtroponin C were purified, according to our established pro-tocols for the cardiac isoforms of these proteins (23). Inclu-sion bodies containing recombinant troponin I were isolatedusing a proprietary extraction buffer (Novagen), and thecrude protein (CP) was solubilized in 6M urea, 1 mM EDTA,20 mM MOPS pH 6.5 and 1 mM 2-mercaptoethanol. Tropo-nin I was subsequently purified using sequential cation ex-change and hydroxyapatite chromatography. Bacterial celllysates containing recombinant human Ala-Ser �-tropomyo-sin were heated to 95°C, before clarification by centrifugationat 33,200 g for 10 min. The resulting supernatant was frac-tionated by reducing the pH to 4.8 and recombinant proteinpurified by anion exchange chromatography. Each mutantwas purified by the same method as used for the correspond-ing WT protein.

Actin-tropomyosin-activated myosin ATPase assays

Whole troponin complexes were reconstituted from theindividual recombinant subunits by stepwise dialysis and gelfiltration using a Superdex 200 column (GE Healthcare), asdescribed previously for human cardiac troponin (24). Rabbitskeletal muscle actin (25) and myosin subfragment-1 (S-1)(26) were prepared using standard methods. Actin-activatedmyosin ATPase assays were carried out as described previouslyusing 0.5 �M myosin S1, 3.5 �M actin, 1 �M �-tropomyosin,and 0.5 �M reconstituted troponin in 5 mM PIPES pH 7.0,3.87 mM MgCl2, 1 mM dithiothreitol at 37°C (23, 27). Thefree calcium concentration was set using 1 mM EGTA, and anappropriate concentration of CaCl2 was calculated by theWinMAXC program (28) to give a pCa range from 9.0 to 5.5.The data were fitted to the Hill equation using KaleidaGraph(Synergy Software) and the mean pCa50 and nH calculatedfrom the derived pCa50 and nH values from 4 independentexperiments.

Exchange of human troponin into rabbit psoas fibers

Human fast skeletal troponin was used to replace the endog-enous complex in rabbit-skinned psoas fibers according tothe whole troponin replacement method of Brenner (29)with our previously described modifications (30). Briefly,small bundles of glycerinated psoas fibers (100 �m � 4 mm)were attached at one end to an Akers 801 piezoelectric forcetransducer and at the other to a static wire hook. The fiberswere then skinned in relaxing solution (pCa9) containing 1%vol/vol Triton X-100 for 2 min. Fibers were brought to rigorby washes in prerigor solution followed by rigor solution toensure complete removal of ATP (see (30) for details ofsolutions). The fibers were then incubated in 5 mg/ml wholetroponin solution (either WT or mutant) for 2 h at 20°C.Following this, the fibers were washed in exchange buffer andrelaxing solution to remove excess protein. SDS-PAGE anal-ysis of exchanged fibers and immunoblotting showed that WTtroponin accounted for 55 � 3% of total troponin; theexchange of mutant troponins was not significantly different.For WT troponin-replaced fibers, the recovered maximum

897CONTRACTILE EFFECTS OF ARTHROGRYPOSIS MUTATIONS

force after exchange was 88 � 19% of maximum force beforereplacement. For pCa-force measurements, initial sarcomerelength was set to 2.5 �m, and fibers were cycled betweenrelaxing solution and solution of increasing Ca2� concentra-tion by means of a rotating bath mechanism (31).

Solution compositions were calculated according to acomputer program (32) using the affinity constants of Smithand Martell (33). All contained a final concentration of free1 mM Mg2�, 5 mM MgATP, 10 mM creatine phosphate, 7 mMEGTA, 10 mM imidazole, pH 7.0. Ionic strength was made upto 0.18 mM using potassium propionate. Activating solutionswere made by adding CaCl2 to give final pCa values from 6.4to 4.0. Relaxing solution was the same, with no addedcalcium.

Measurement of the affinity of tropomyosin for actin

The binding of tropomyosin to actin was measured by cosedi-mentation. 0.25–12 �M of recombinant Ala-Ser �-tropomyo-sin and 21 �M of native rabbit skeletal actin were cosedi-mented by centrifugation at 436,000 g in a Beckman TLA100rotor at 20°C in 200 mM NaCl, 10 mM Tris HCl pH 7.5, 3.87mM MgCl2, and 0.5 mM dithiothretol. Pellets and superna-tants were analyzed using Coomassie blue-stained 10.5%SDS-polyacrylamide gels and quantified using an FC8800geldoc chemiluminescence densitometry program (AlphaInotech). All calculated tropomyosin concentrations wereadjusted by percentage pellet recovery, which was indepen-dently calculated using the same densitometry system. Appar-ent binding constants Kapp were determined by fitting thedata to the Hill equation using KaleidaGraph (Synergy Soft-ware).

Circular dichroism measurements.

Thermal stability measurements were made by following themolar ellipticity of Ala-Ser �-tropomyosin at 222 nm as afunction of temperature in buffer containing 0.5 M NaCl, 10mM sodium phosphate pH 7.5, 1 mM EDTA and 0.5 mMdithiothretol using a Jasco J720 spectropolarimeter equippedwith variable temperature bath (cell path length 1 cm). Datawere obtained at 1.0°C intervals from 10 to 90°C using aprotein concentration of 0.1 mg/ml. The molar ellipticity wasnormalized (10°C�1, 90°C�0) to give fraction folded andfitted to a double exponential equation:

Fraction folded � ε1 �e(H1/RT)T/Tm1�

1 � e(H1/RT)T/Tm1�� ε2 �

eH2/RT)T/Tm2�

1�e(H2/RT)T/Tm2�

where ε corresponds to the proportion of each state ofunfolding, H is the enthalpy change of denaturation foreach transition, while the Tm gives the median temperature(in K) for each transition. These parameters were used tocalculate the free energy (G in kcal/mol) of the tropomyo-sin helix at 20°C using the following equation: G � ε1(H1–293 S1) � ε2 (H2–293 S2) where S is the entropycalculated at the Tm of each transition (34).

RESULTS

Recombinant WT and DA mutant forms of each humanfast skeletal troponin subunit and �-tropomyosin wereexpressed and purified, using similar protocols to thosewe have developed for the human cardiac isoforms (23,24). �-tropomyosin was expressed with a Met-Ala-SerN-terminal leader sequence, reduced to Ala-Ser in the

mature protein. This modification has been used forrecombinant �-tropomyosin and shown to mimic theN-terminal acetylation of the native protein, which isnecessary for normal head-to-tail interactions and bind-ing to actin (35). SDS-PAGE analysis of the recombi-nant proteins is shown in Fig. 1. Our sequential dialysisand gel filtration protocol for troponin reconstitutionreproducibly resulted in the purification of complexesof 1:1:1 subunit stoichiometry, this being true for WTand all mutant troponins.

Effect of DA mutations on the Ca2�-regulation ofactin -activated myosin ATPase

The functional effects of mutant troponin I, troponinT, and Ala-Ser �-tropomyosin were compared to WT byassay of thin-filament activation of skeletal muscle my-osin S-1 ATPase activity. WT and mutant troponincomplexes were reconstituted from recombinant hu-man fast skeletal troponin subunits; no difference inthe efficacy of reconstitution was noted between WTand any mutant troponin. Thin filaments were assem-bled using rabbit skeletal actin and myosin S-1, WT ormutant Ala-Ser �-tropomyosin and WT or mutant tro-ponin. Actin cosedimentation assays were carried out,as described previously (23) and showed that bindingof WT and mutant troponin were indistinguishable(data not shown). WT troponin and Ala-Ser �-tropomy-osin regulated ATPase activity in a Ca2�-dependent,cooperative manner with pCa50 of 7.29 � 0.02 (n�4)and Hill coefficient (nH), a measure of cooperativity,2.58 � 0.28 (n�4) (Fig. 2). Each mutant troponin ortropomyosin also conferred Ca2�-dependent, coopera-tive regulation with no statistically significant differ-ences in either pCa50 or nH; however, each mutant didhave the striking effect of increasing ATPase activitythroughout the pCa range (Fig. 2). Thus, we havedocumented this by comparing the maximally activatedATP activities at pCa5.5 for WT and each mutant and

Figure 1. Recombinant human fast skeletal proteins. A Coo-massie blue-stained 12% polyacrylamide gel showing thepurified recombinant proteins used in this study. Lanes: 1,molecular weight markers (sizes indicated on the left); 2,human fast skeletal troponin T; 3, human fast skeletal tropo-nin I; 4, Arg156ter mutant human fast skeletal troponin I; 5,human fast skeletal troponin C; 6, human fast skeletal tropo-nin complex reconstituted from recombinant subunits; 7,human AS-� tropomyosin. All mutants were obtained by thesame purification protocol as their wild-type (WT) counter-parts and at equivalent purity.

898 Vol. 21 March 2007 ROBINSON ET AL.The FASEB Journal

Figure 2. Ca2� regulation of actin-activated myosin S-1 ATPase by WT and DA mutant troponin/tropomyosin. Actin-activatedmyosin ATPase assays were carried out using 0.5 �M myosin S1, 3.5 �M actin, 1 �M Ala-Ser �-tropomyosin, and 0.5 �Mreconstituted troponin in 5 mM 1,4-piperazinebis(ethane sulfonic acid) pH 7.0, 3.87 mM MgCl2, 1 mM dithiothreitol at 37°Cat free-calcium concentrations between pCa 9.0 and 5.5. WT data (F) is compared with data generated using a troponin or a�-tropomyosin DA mutant and with that observed using a 1:1 mix of DA mutant and the corresponding WT component. Eachpoint represents the mean rate � sem measured in 4 independent experiments. A) Arg156ter troponin I: 100% mutant (E),50% mutant/50% WT (i), B) Arg174Gln troponin I: 100% mutant (E), 50% mutant/50% WT (i), C) Arg63His troponin T:100% mutant (E), 50% mutant/50% WT (i), D) Arg91Gly �-tropomyosin: 100%mutant (E), 50% mutant/50% WT (i).

Figure 3. Effect of DA mutant troponin and tropomyosin on maximally activated and inhibited ATPase activities. Actin-activatedmyosin ATPase assay was carried out as described in Fig. 2. A) ATPase rates at pCa5.5 for thin filament containing DA mutants(open bars) and DA muants in 1:1 ratio with the corresponding WT component (hatched bars) are compared with WT. B)ATPase rates at pCa9.0 for thin filament containing DA mutants (open bars) and DA muants in 1:1 ratio with the correspondingWT component (hatched bars) are compared with WT. Each point represents the mean rate � sem measured in independentexperiments (n�12 WT, n�8 each mutant and 1:1 WT/mutant).

899CONTRACTILE EFFECTS OF ARTHROGRYPOSIS MUTATIONS

also the maximally inhibited rates at pCa9 (Fig. 3). WTtroponin/tropomyosin gave a maximally activated rateof 4.50 � 0.01 s�1 (n�12); each DA mutant gaveenhanced activation of ATPase activity, the mostmarked being produced by the �-tropomyosinArg91Gly mutant (6.40�0.08 s�1, n�6, P 0.001) withthe troponin mutants giving a more moderate butsignificant (P 0.01, n�8), increase (TnI Arg156ter6.04�0.1 s�1, TnI Arg174Gln � 5.02�0.08 s�1, TnTArg63His�5.02�0.07 s�1) (Fig. 3A). Under relaxingconditions at pCa 9.0, WT troponin and Ala-Ser �-tro-pomyosin gave an inhibited activity of 1.43 � 0.03 s�1

(n�12) and significant (P 0.01) increases in this ratewere observed using 3 of the DA mutant proteins (TnIArg156ter 2.39�0.08 s�1, TnT Arg63His�2.08�0.08s�1, �TM Arg91Gly 3.14�0.01 s�1). TnI Arg174Glngave a small, nonsignificant (P�0.05) increase (Fig.3B).

The disease caused by the DA mutations is autosomaldominant, and it would be anticipated that the fastskeletal thin filaments of affected individuals containequal proportions of the relevant WT and mutantprotein. There have been no direct biopsy analyses ofWT and DA mutant protein ratios reported; in the fewcases in which studies have been carried out on con-tractile protein mutants that cause cardiomyopathy,close to 1:1 ratios have usually been reported; forexample, in the case of missense �-tropomyosin hyper-trophic cardiomyopathy (HCM)-causing mutants (36).ATPase experiments were therefore repeated using WTand mutant protein in a 1:1 mix in an attempt to reflectmore accurately the in vivo ratio. The Ca2�-regulationof ATPase activity by 1:1 mixes showed no significantdifferences from WT in either pCa50 or nH but gaveincreases in rate of ATPase at all Ca2� concentrationsintermediate to the 100% mutant and WT levels (Fig.2). In activating conditions, all mutants in a 1:1 mixgave significantly enhanced ATPase rates comparedwith WT. At pCa9, all except the TnI Arg174Glnmutant produced significantly higher inhibited activitycompared with WT; this mutant gave a small butnonsignificant increase (1.47�0.06 s�1) (Fig. 3B).

The effect of DA mutations on the Ca2� regulationof force generation in chemically skinned rabbitpsoas fibers

To investigate the effect of the DA troponin mutants onforce generation, reconstituted human skeletal tropo-nin was used to displace the endogenous troponincomplex in rabbit-skinned psoas fibers using themethod of Brenner (29). Using this technique, weincorporated WT human skeletal troponin at 55 � 3%of total troponin in these fibers, and exchange of themutant troponins was not significantly different fromWT. A section of a force trace from a typical experi-ment (in this case using troponin containingArg174Gln TnI) is shown in Fig. 4. The maximumCa2�-activated force produced by fibers containingWT human troponin was 0.079 � 0.010 N mm�2 and

that were produced by fibers containing eitherArg174Gln TnI or Arg63His TnT was not significantlydifferent (Fig. 5). However, fibers containing thetruncation mutant of TnI, Arg156ter, gave markedlyenhanced maximum force (0.154�0.025 N mm�2;P 0.05). In all three cases, incorporation of mutanttroponin did not result in a change in resting tensionat pCa9 (illustrated for Arg174Gln TnI in Fig. 4). TheCa2�-sensitivity of force generation was also tested(Fig. 6). WT-exchanged fibers had a pCa50 of forcegeneration of 5.26 � 0.11 with Hill coefficient, nH, of2.27. All three mutants tested showed enhancedCa2�-sensitivity with pCa50s of 5.69 � 0.05 (Arg156terTnI), 5.62 � 0.08 (Arg174Gln TnI), and 5.49 � 0.05(Arg63His TnT).

Analysis of the effect of the Arg91Gly mutation on �tropomyosin structure and actin binding

We carried out analysis of the effect of the mutation onthe coiled-coil structure of tropomyosin and its actinbinding properties, alterations in which may help toexplain the alteration in regulatory function. The affin-ity of Ala-Ser �-tropomyosin for actin was measured bycosedimentation. Both WT and mutant displayed coop-erative binding, and the data were fitted to the Hillequation (Fig. 7A). WT bound with a calculated disso-ciation constant Kapp of 1.89 � 10�6 M and nH of 2.58.The affinity of the DA mutant was significantly lower:Kapp was 7.72 � 10�6 M and nH of 3.47.

Figure 4. A section of the force trace from a typical experi-ment in which troponin containing Arg174Gln TnI wasexchanged into skinned rabbit psoas fibers. Chemicallyskinned rabbit psoas fibers were mounted, and control con-tractions at pCa5 were performed (3 min each). The fiberswere put into rigor, and then the fibers were incubated withtroponin solution for 20 min. After the washing with ex-change buffer and returning to relaxing solution, a series ofcontractions (3 min each) were performed (pCa5 and 6.4illustrated). Note there is no increase in resting tension afterthe troponin exchange. Bathing solutions are indicated abovethe trace (shaded boxes); clear boxes indicate relaxing solu-tion (pCa9). Force/cross-sectional area scale is indicated onthe left.

900 Vol. 21 March 2007 ROBINSON ET AL.The FASEB Journal

The impact of the Arg91Gly mutation on the coiled-coil structure of Ala-Ser �-tropomyosin was examinedby monitoring the molar ellipticity at 222 nm as afunction of temperature (Fig. 7B). The transition of theArg91Gly protein from coiled-coil to separate strandstook place at a markedly lower temperature range thanWT, indicating that the mutation had a significantimpact on the coiled-coil stability. The Tm of thecomplete transition was 45.5°C for WT and 38.5°C forthe DA mutant. Previous analyses of tropomyosin un-folding have fitted the data to either two or threetransitions (34, 37); both our WT and mutant datafitted well to a two-step model (Fig. 7B). Using theconstants derived from these fits, the calculated G offolding at 20°C were calculated as �5.04 and �4.10kcalmol�1 for WT and Arg91Gly �-TM, respectively.

DISCUSSION

We have performed the first study of the in vitro and insitu functional effects caused by the mutations in thin-filament protein genes reported to cause the limbdeformities that compose DA. All four reported muta-tions in the three disease genes have been examined.Recombinant forms of each mutant protein have beencompared with WT in actin-activated myosin ATPaseassays, and each troponin mutant analyzed in troponin-exchanged rabbit psoas fibers. In the biochemicalstudies, all four mutations significantly enhancedATPase activity, not only in experiments comparing WTand mutant, but also in the case of 1:1 ratios of WT andmutant, the predicted proportion of the mutant pro-tein in vivo. The fiber work on the troponin mutations,carried out at close to 1:1 endogenous WT:humansubunit, showed that each mutation caused increased

Ca2� sensitivity of force regulation. These functionalassays show that all of the DA mutants increase contrac-tile function in vitro. Furthermore, analysis of the

Figure 5. Effect of DA troponin mutants on maximum forcegeneration in human troponin-exchanged skinned psoasfibers. Maximum force (mM mm�2), measured at pCa4, wasdetermined using fibers containing either wild human fastskeletal troponin or troponin containing either Arg156 terTnI, Arg174Gln TnI, or Arg63His TnT.

Figure 6. Effect of DA troponin mutants on pCa-force regu-lation in human troponin-exchanged skinned psoas fibers.Force generation at Ca2� concentrations between pCa 6.4and 4 was determined using fibers containing wild human fastskeletal troponin and compared with that obtained fromfibers containing either Arg156ter TnI (A), Arg174Gln TnI(B), or Arg63His TnT (C). Tension was normalized by forceat pCa6.4 to 0% and force at pCa4 to 100%.

901CONTRACTILE EFFECTS OF ARTHROGRYPOSIS MUTATIONS

�-tropomyosin DA mutant showed that the Arg91Glymutation significantly disrupts both actin binding andcoiled-coil stability.

Each of the four mutants had similar effects on the

regulation of ATPase activity and the three troponinmutants all increased the Ca2�-sensitivity of force gen-eration, despite the mutations altering different partsof troponin-tropomyosin. The two TnI mutations affectthe C-terminal region of the protein with the Arg156termutant missing the C-terminal 27 amino acids. Thisregion has previously been shown to be required forfull inhibitory activity of TnI in studies of C-terminaltruncations of both the fast skeletal protein (38) andthe cardiac isoform (39). The Arg63His TnT mutationaffects an amino acid within the extended N-terminalT1 domain of the protein, which does not participate indirect interaction with either TnI and TnC but liesalong the tropomyosin coiled coil. Residues 92–110 ofthe cardiac TnT sequence have been suggested to bindto the overlap region of tropomyosin (40); these cor-respond to 61–79 in fast skeletal TnT, and hence Arg63is likely to interact with tropomyosin in this structurallyimportant region. The Arg91 residue of �-tropomyosinis located within the third of seven pseudorepeat re-gions, each of which has been proposed to contributeto the binding to actin (41). We hypothesize that theDA mutations in TnT and �-tropomyosin causechanges in the equilibria governing the blocked7closed7 myosin-induced transitions described by thethree-state steric blocking model (14), in favor of themyosin-induced open state. The TnI mutations mayalso have an impact on these equilibria via alteredinteractions of the troponin head with tropomyosin butalso are likely to have a direct effect on the inhibitoryrole of TnI. In the measurements of thin-filament-activated myosin S-1 ATPase activity, these thin-filamentperturbations are expressed as an increase in rate; inthe fiber studies, the troponin mutations result insignificant increases in Ca2� sensitivity. The likely mainreason is the oriented actin-myosin lattice array presentin the intact sarcomeres which gives an added elementof cooperativity, as attached myosin heads can increasethe transition of neighboring troponin-tropomyosincomplexes from closed to the myosin-induced openstate, effectively causing increased Ca2� sensitivity. Fur-thermore, crossbridge strain affects the affinity of my-osin for adenine nucleotides, and this would havesubtle effects on ATPase rates in the fibers (12, 42, 43).The stoichiometry of incorporation of human troponininto the fibers was close to the expected 1:1 WT/mutant ratio, although it should be borne in mind thatthe exchange of exogenous subunits does not alwaysoccur evenly across the thin filaments (44), a factor thatmay have some effect on the degree of the observedchanges.

There have been few previous reports on the bio-chemistry of specifically the � isoform of tropomyosinbecause of difficulty in obtaining a pure preparation ofthe protein from native �/� mixtures. Here, we reportthe overexpression and purification of human �-tropo-myosin with an N-terminal Met-Ala-Ser leader, which isprocessed to give an Ala-Ser tag in bacteria. The abilityof unmodified recombinant tropomyosin to undergonormal head-to-tail interactions and to bind actin is

Figure 7. Effect of the Arg91Gly mutation in �-tropomyosinon actin binding and coiled-coil thermostability. A) Cooper-ative binding of WT and Arg91Gly �-tropomyosin to actin.Cosedimentation experiments were carried out at 20°C inbuffer containing 200 mM NaCl, 10 mM Tris HCL pH7.5,3.87mM MgCl2, 0.5 mM DTT. Lines of best fit to the Hillequation were calculated using Kaleidagraph (Synergy Soft-ware). WT (E) and Arg91Gly (�) �-tropomyosin. B) Thetemperature dependence of unfolding of WT and Arg91Gly�-tropomyosin. Fraction folded as measured by the relativeellipticity at 222 nm as a function of temperature, in 500 mMNaCl, 10 mM NaPO4 pH 7.0, 1 mM EDTA, and 0.5 mM DTT.WT (E) and Arg91Gly (�) �-tropomyosin. The data werefitted to an equation (see Methods) showing two exponentialtransitions, as described previously (34). The derived con-stants were fraction in first transition � 0.55 (WT), 0.71(Arg91Gly); H of first transition � �28.3 kcalmol�1 (WT),�45.7 kcalmol�1 (Arg91Gly); Tm of first transition � 39.8°C(WT), 35.0°C (Arg91Gly); fraction in second transition �0.45 (WT), 0.29 (Arg91Gly); H of second transition � �99.4kcalmol�1 (WT), �89.7 kcalmol�1 (Arg91Gly); Tm of secondtransition � 49.2°C (WT), 51.7°C (Arg91Gly).

902 Vol. 21 March 2007 ROBINSON ET AL.The FASEB Journal

much reduced because of the absence of the usualacetylation of the N-terminal methionine found ineukaryotic native tropomyosins (45). The presence ofthe dipeptide tag has been reported to mimic theposttranslational modification in the case of recombi-nant �-tropomyosin, and has been shown to restore itsnear-normal actin binding and regulatory properties(35, 46, 47). Human Ala-Ser-�-tropomyosin bound toF-actin cooperatively (nH�2.6) with Kapp of �1.9 � 10�

(6) M. This is somewhat weaker than the affinityreported for Ala-Ser-�-tropomyosin (8.3�10�7 M(35)), which may reflect that the binding by �� isweaker than that of �� dimers or that the N-terminaltag is a less effective mimic of N-terminal acetylation for�-tropomyosin. The Arg91Gly mutation caused an ap-proximately four-fold decrease in actin affinity, anunexpectedly large effect given that only a single resi-due in a region outside of the overlapping termini wasmutated. The residue is located within the third ofseven pseudorepeat regions and the deletion of thisentire repeat (residues 89–123) in rabbit �-tropomyo-sin resulted in a 26-fold drop in actin affinity (48). Themutation from Arg to Gly also causes a pronounceddestabilization of the coiled-coil (Fig. 7B). Glycine iswell known to reduce stability of helices, and further-more, the removal of the positively charged argininefrom the “g” position of the heptad repeat will result inthe loss of favorable interactions with nearby glutamateresidues (e.g., Glu96 in the “e” position of the opposingstrand). Recent studies by Hitchcock-DeGregori andcoworkers have examined the effect on the stability ofthe �-tropomyosin coiled-coil of mutation of severalalanine residues thought to confer molecular flexibilityand to enable efficient actin binding. It was found thatstabilizing the tropomyosin structure at either the N- orC-terminus drastically reduced the affinity for actin(49–51). These data contrast with our finding that theArg91Gly mutation in �-tropomyosin destabilizes thecoiled coil but also reduces actin binding, thereforeindicating that the reported relationship between thesetwo parameters is not always observed.

The Arg91Gly mutation in �-tropomyosin that causesDA type 1A gives a greater increase in ATPase activitythan mutations in the troponin complex, which causethe phenotypically more severe DA type 2B (2). Thisparadoxical observation can be explained by the anal-ysis of the tissue distribution of the disease causingproteins involved. Fast skeletal troponin is the predom-inant isoform of the distal fast-twitch skeletal musclefibers (52). Fast-twitch tropomyosin, however, is madeup of both �-tropomyosin and �-tropomyosin, with the� isoform present at no more than 50% of totaltropomyosin; moreover, studies have suggested that theisoforms preferentially dimerize such that a consider-able majority of the tropomyosin regulating the thinfilament in vivo is an �-�-heterodimer (53, 54). Thismeans that the Arg91Gly mutant �-tropomyosin wouldonly constitute at most 25% of the total tropomyosin invivo, and hence, the severe structural and functionaleffects caused by the Arg91Gly mutation measured in

this study would be diluted. A direct analysis of theeffect of Arg91Gly mutant � -tropomyosin on �-�-heterodimer function was not attempted in this studybecause of the practical difficulties in preparing het-erodimers and in quantitating the �/�, �/�, and �/�distribution of the formed dimers. We also note that itis conceivable that small changes in the effects of thetroponin mutations may be observed if they were to bereconstituted in thin filaments with �/� instead of�/�-tropomyosin.

Our laboratories and others have previously analyzedthe effects of mutations in cardiac thin-filament pro-teins, which cause either HCM or dilated cardiomyop-athy (DCM) (55). A distinct pattern has emerged fromthese studies: the HCM mutations in troponin and�-tropomyosin have been largely shown to increaseCa2� sensitivity, whereas DCM-causing mutations in thesame proteins reduce Ca2� sensitivity (24, 47, 56).Hence, it is likely that the HCM mutations initially giverise to increased contractility and DCM to lower, andthese changes provide separate stimuli for generationof the two different diseases of cardiac muscle. Thequalitative effects of the DA mutant proteins moreclosely resemble those of the HCM mutants, althoughthe DA mutants show Ca2� sensitivity changes only inthe fiber experiments. Interestingly, the Arg63His mu-tation in fast skeletal TnT occurs at the equivalentposition to a known HCM-causing mutation in thecardiac sequence. This mutation Arg94Leu is somewhatunusual in that it produces the myocyte disarray char-acteristic of HCM and causes sudden death but doesnot generate significant ventricular hypertrophy (57).The magnitude of some of the effects seen with the DAmutants is greater than that seen with HCM mutations,suggesting a more severe disturbance of skeletal musclefunction in DA than likely could be compatible with lifeif cardiac muscle were affected to a similar extent.

We propose that if the observed functional changescaused by the DA mutant proteins are manifest in vivoin affected individuals, the markedly altered contractileregulation will cause increased tension in the develop-ing muscles. The chronic effect of this will be thecontractures and limb defects observed in these syn-dromes. Thus the development of the skeletal abnor-malities is an active process, rather than a passive one,as had been previously proposed (7). This is in keepingwith the demonstration that the contractures developin utero in the second trimester in previously straightlimbs at a time when fast-twitch muscle isoforms arebeing expressed at a high level (9). These disorders aregenerally not progressive after birth, and indeed, someof the symptoms can be treated with simple physicaltherapy (1). The proximal limb skeletal muscle perfor-mance in affected adults is generally normal, as ismuscle histology (e.g., of the quadriceps muscle). Thelikely interpretation of these findings is that it is onlythe distal muscles that are predominantly fast-twitchand so are susceptible to contracture, whereas thealtered regulation at the myofilament level may becompensated for in muscles with mixed fast and slow

903CONTRACTILE EFFECTS OF ARTHROGRYPOSIS MUTATIONS

troponin and tropomyosin isoforms, for example, bychanges in Ca2� handling.

This work has been supported by the British Heart Foun-dation.

REFERENCES

1. Hall, J. G., Reed, S. D., and Greene, G. (1982) The distalarthrogryposes: delineation of new entities—review and noso-logic discussion. Am. J. Med. Genet. 11, 185–239

2. Bamshad, M., Jorde, L. B., and Carey, J. C. (1996) A revised andextended classification of the distal arthrogryposes. Am. J. Med.Genet. 65, 277–281

3. Beals, R. K. (2005) The distal arthrogryposes: a new classifica-tion of peripheral contractures. Clin. Orthop. Relat. Res. 203–210

4. Bamshad, M., Bohnsack, J. F., Jorde, L. B., and Carey, J. C.(1996) Distal arthrogryposis type 1: clinical analysis of a largekindred. Am. J. Med. Genet. 65, 282–285

5. Krakowiak, P. A., Bohnsack, J. F., Carey, J. C., and Bamshad, M.(1998) Clinical analysis of a variant of Freeman-Sheldon syn-drome (DA2B). Am. J. Med. Genet. 76, 93–98

6. Gordon, N. (1998) Arthrogryposis multiplex congenita. BrainDev. 20, 507–511

7. Hall, J. G. (1997) Arthrogryposis multiplex congenita: etiology,genetics, classification, diagnostic approach, and general as-pects. J. Pediatr. Orthop. B. 6, 159–166

8. Bamshad, M., Watkins, W. S., Zenger, R. K., Bohnsack, J. F.,Carey, J. C., Otterud, B., Krakowiak, P. A., Robertson, M., andJorde, L. B. (1994) A gene for distal arthrogryposis type I mapsto the pericentromeric region of chromosome 9. Am. J. Hum.Genet. 55, 1153–1158

9. Sung, S. S., Brassington, A. M., Grannatt, K., Rutherford, A.,Whitby, F. G., Krakowiak, P. A., Jorde, L. B., Carey, J. C., andBamshad, M. (2003) Mutations in genes encoding fast-twitchcontractile proteins cause distal arthrogryposis syndromes.Am. J. Hum. Genet. 72, 681–690

10. Sung, S. S., Brassington, A. M., Krakowiak, P. A., Carey, J. C.,Jorde, L. B., and Bamshad, M. (2003) Mutations in TNNT3cause multiple congenital contractures: a second locus for distalarthrogryposis type 2B. Am. J. Hum. Genet. 73, 212–214

11. Squire, J. M., and Morris, E. P. (1998) A new look at thinfilament regulation in vertebrate skeletal muscle. FASEB J. 12,761–771

12. Gordon, A. M., Homsher, E., and Regnier, M. (2000) Regula-tion of contraction in striated muscle. Physiol. Rev. 80, 853–924

13. Lehrer, S. S., and Morris, E. P. (1982) Dual effects of tropomy-osin and troponin-tropomyosin on actomyosin subfragment 1ATPase. J. Biol. Chem. 257, 8073–8080

14. McKillop, D. F., and Geeves, M. A. (1993) Regulation of theinteraction between actin and myosin subfragment 1: evidencefor three states of the thin filament. Biophys. J. 65, 693–701

15. Maytum, R., Lehrer, S. S., and Geeves, M. A. (1999) Coopera-tivity and switching within the three-state model of muscleregulation. Biochemistry 38, 1102–1110

16. Smith, D. A., and Geeves, M. A. (2003) Cooperative regulationof myosin-actin interactions by a continuous flexible chain II:actin-tropomyosin-troponin and regulation by calcium. Biophys.J. 84, 3168–3180

17. Smith, D. A., Maytum, R., and Geeves, M. A. (2003) Cooperativeregulation of myosin-actin interactions by a continuous flexiblechain I: actin-tropomyosin systems. Biophys. J. 84, 3155–3167

18. Vibert, P., Craig, R., and Lehman, W. (1997) Steric-model foractivation of muscle thin filaments. J. Mol. Biol. 266, 8–14

19. Craig, R., and Lehman, W. (2001) Crossbridge and tropomyosinpositions observed in native, interacting thick and thin fila-ments. J. Mol. Biol. 311, 1027–1036

20. Lehman, W., Hatch, V., Korman, V., Rosol, M., Thomas, L.,Maytum, R., Geeves, M. A., Van Eyk, J. E., Tobacman, L. S., andCraig, R. (2000) Tropomyosin and actin isoforms modulate thelocalization of tropomyosin strands on actin filaments. J. Mol.Biol. 302, 593–606

21. Way, M., Pope, B., Gooch, J., Hawkins, M., and Weeds, A. G.(1990) Identification of a region in segment 1 of gelsolin criticalfor actin binding. EMBO J. 9, 4103–4109

22. Studier, F. W., Rosenberg, A. H., Dunn, J. J., and Dubendorff,J. W. (1990) Use of T7 RNA polymerase to direct the expressionof cloned genes. Methods Enzymol. 185, 60–89

23. Redwood, C. S., Lohmann, K., Bing, W., Esposito, G. M., Elliott,K., Abdulrazzak, H., Knott, A., Purcell, I., Marston, S. B., andWatkins, H. (2000) Investigation of a truncated cardiac tropo-nin T that causes familial hypertrophic cardiomyopathy: Ca2�

regulatory properties of reconstituted thin filaments depend onthe ratio of mutant to wild-type peptide. Circ. Res. 86, 1146–1152

24. Robinson, P., Mirza, M., Knott, A., Abdulrazzak, H., Willott, R.,Marston, S., Watkins, H., and Redwood, C. (2002) Alterations inthin filament regulation induced by a human cardiac troponinT mutant that causes dilated cardiomyopathy are distinct fromthose induced by troponin T mutants that cause hypertrophiccardiomyopathy. J. Biol. Chem. 277, 40710–40716

25. Pardee, J. D., and Spudich, J. A. (1982) Purification of muscleactin. Methods Enzymol. 85, 164–181

26. Margossian, S. S., and Lowey, S. (1982) Preparation of myosinand its subfragments from skeletal muscle. Methods Enzymol. 85,55–71

27. Elliott, K., Watkins, H., and Redwood, C. S. (2000) Alteredregulatory properties of human cardiac troponin I mutants thatcause hypertrophic cardiomyopathy. J. Biol. Chem. 275, 22069–22074

28. Bers, D. M., Patton, C. W., and Nuccitelli, R. (1994) A practicalguide to the preparation of Ca2� buffers. Methods. Cell Biol. 40,3–29

29. Brenner, B., Kraft, T., Yu, L. C., and Chalovich, J. M. (1999)Thin-filament activation probed by fluorescence of N-((2-(Iodoacetoxy)ethyl)-N-methyl)amino-7-nitrobenz-2-oxa-1, 3-dia-zole-labeled troponin I incorporated into skinned fibers ofrabbit psoas muscle. Biophys. J. 77, 2677–2691

30. Lipscomb, S., Preston, L. C., Robinson, P., Redwood, C. S.,Mulligan, I. P., and Ashley, C. C. (2005) Effects of troponin Cisoform on the action of the cardiotonic agent EMD 57033.Biochem. J. 388, 905–912

31. Griffiths, P. J., and Jones, A. (1994) A simple device for transferof single muscle fibres by rotation between 70-�l chamberswhile making optical measurements. J. Physiol. 438, 146P

32. Fabiato, A., and Fabiato, F. (1979) Calculator programs used forcomputing the composition of the solutions containing multi-ple metals and ligands used for experiements iin skinnedmuscle cells. J. Physiol. (Paris) 75, 463–505

33. Smith, R. M., and Martell, A. E. (1974) Aminocarboxylic acids.In Critical Stability Constants (Smith, R. M., ed) pp. 1–65, PlenumPress, New York

34. Greenfield, N. J., and Hitchcock-DeGregori, S. E. (1995) Thestability of tropomyosin, a two-stranded coiled-coil protein, isprimarily a function of the hydrophobicity of residues at thehelix-helix interface. Biochemistry 34, 16797–16805

35. Monteiro, P. B., Lataro, R. C., Ferro, J. A., and Reinach, F.(1994) Functional alpha-tropomyosin produced in Escherichiacoli. A dipeptide extension can substitute the amino-terminalacetyl group. J. Biol. Chem. 269, 10461–10466

36. Bottinelli, R., Coviello, D. A., Redwood, C. S., Pellegrino, M. A.,Maron, B. J., Spirito, P., Watkins, H., and Reggiani, C. (1998) Amutant tropomyosin that causes hypertrophic cardiomyopathyis expressed in vivo and associated with an increased calciumsensitivity. Circ. Res. 82, 106–115

37. Heller, M. J., Nili, M., Homsher, E., and Tobacman, L. S. (2003)Cardiomyopathic tropomyosin mutations that increase thinfilament Ca2� sensitivity and tropomyosin N-domain flexibility.J. Biol. Chem. 278, 41742–41748

38. Ramos, C. H. (1999) Mapping subdomains in the C-terminalregion of troponin I involved in its binding to troponin C and tothin filament. J. Biol. Chem. 274, 18189–18195

39. Rarick, H. M., Tu, X.-H., Solaro, R. J., and Martin, A. F. (1997)The C terminus of cardiac troponin I is essential for fullinhibitory activity snd Ca2� sensitivity of rat myofibrils. J. Biol.Chem. 272, 26887–26892

40. Palm, T., Graboski, S., Hitchcock-DeGregori, S. E., and Green-field, N. J. (2001) Disease-causing mutations in cardiac troponinT: identification of a critical tropomyosin-binding region. Bio-phys. J. 81, 2827–2837

41. McLachlan, A. D., and Stewart, M. (1976) The 14-fold period-icity in alpha-tropomyosin and the interaction with actin. J. Mol.Biol. 103, 271–298

904 Vol. 21 March 2007 ROBINSON ET AL.The FASEB Journal

42. Ashley, C. C., Lea, T. J., Mulligan, I. P., Palmer, R. E., andSimnett, S. J. (1993) Activation and relaxation mechanisms insingle muscle fibres. Adv. Exp. Med. Biol. 332, 97–114; discussion114–115

43. Smith, D. A., and Geeves, M. A. (1995) Strain-dependentcross-bridge cycle for muscle. Biophys. J. 69, 524–537

44. She, M., Trimble, D., Yu, L. C., and Chalovich, J. M. (2000)Factors contributing to troponin exchange in myofibrils and insolution. J. Muscle Res. Cell Motil. 21, 737–745

45. Hitchcock-DeGregori, S. E., and Heald, R. W. (1987) Alteredactin and troponin binding of amino-terminal variants ofchicken striated muscle alpha-tropomyosin expressed in Esche-richia coli. J. Biol. Chem. 262, 9730–9735

46. Bing, W., Redwood, C. S., Purcell, I. F., Esposito, G., Watkins, H.,and Marston, S. B. (1997) Effects of two hypertrophic cardio-myopathy mutations in �- tropomyosin, Asp175Asn andGlu180Gly, on Ca2� regulation of thin-filament motility. Bio-chem. Biophys. Res. Commun. 236, 760–764

47. Mirza, M., Marston, S., Willott, R., Ashley, C., Mogensen, J.,McKenna, W., Robinson, P., Redwood, C., and Watkins, H.(2005) Dilated cardiomyopathy mutations in three thin-filamentregulatory proteins result in a common functional phenotype.J. Biol. Chem. 280, 28498–28506

48. Hitchcock-DeGregori, S. E., Song, Y., and Greenfield, N. J.(2002) Functions of tropomyosin’s periodic repeats. Biochemistry41, 15036–15044

49. Brown, J. H., Kim, K. H., Jun, G., Greenfield, N. J., Dominguez,R., Volkmann, N., Hitchcock-DeGregori, S. E., and Cohen, C.(2001) Deciphering the design of the tropomyosin molecule.Proc. Natl. Acad. Sci. U. S. A 98, 8496–8501

50. Singh, A., and Hitchcock-DeGregori, S. E. (2006) Dual require-

ment for flexibility and specificity for binding of the coiled-coiltropomyosin to its target, actin. Structure 14, 43–50

51. Singh, A., and Hitchcock-DeGregori, S. E. (2003) Local desta-bilization of the tropomyosin coiled coil gives the molecularflexibility required for actin binding. Biochemistry 42, 14114–14121

52. Perry, S. V. (1985) Properties of the muscle proteins—a com-parative approach. J. Exp. Biol. 115, 31–42

53. Bronson, D. D., and Schachat, F. H. (1982) Heterogeneity ofcontractile proteins. Differences in tropomyosin in fast, mixed,and slow skeletal muscles of the rabbit. J. Biol. Chem. 257,3937–3944

54. Lehrer, S. S., and Stafford, W. F., 3rd (1991) Preferentialassembly of the tropomyosin heterodimer: equilibrium studies.Biochemistry 30, 5682–5688

55. Seidman, J. G., and Seidman, C. (2001) The genetic basis forcardiomyopathy: from mutation identification to mechanisticparadigms. Cell 104, 557–567

56. Hernandez, O. M., Housmans, P. R., and Potter, J. D. (2001)Invited Review: pathophysiology of cardiac muscle contractionand relaxation as a result of alterations in thin filament regula-tion. J. Applied Physiol. 90, 1125–1136

57. Varnava, A., Baboonian, C., Davison, F., de Cruz, L., Elliott,P. M., Davies, M. J., and McKenna, W. J. (1999) A new mutationof the cardiac troponin T gene causing familial hypertrophiccardiomyopathy without left ventricular hypertrophy. Heart 82,621–624

Received for publication July 25, 2006.Accepted for publication September 29, 2006.

905CONTRACTILE EFFECTS OF ARTHROGRYPOSIS MUTATIONS