An R111C polymorphism in wild turkey cardiac troponin I accompanying the dilated cardiomyopathy-related abnormal splicing variant of cardiac troponin T with potentially compensatory effects.

Cardiac muscle contraction is regulated by Ca(2+) through the troponin complex consisting of three subunits: troponin C (TnC), troponin T (TnT), and troponin I (TnI). We reported previously that the abnormal splicing of cardiac TnT in turkeys with dilated cardiomyopathy resulted in a greater binding affinity to TnI. In the present study, we characterized a polymorphism of cardiac TnI in the heart of wild turkeys. cDNA cloning and sequencing of the novel turkey cardiac TnI revealed a single amino acid substitution, R111C. Arg(111) in avian cardiac TnI corresponds to a Lys in mammals. This residue is conserved in cardiac and skeletal muscle TnIs across the vertebrate phylum, implying a functional importance. In the partial crystal structure of cardiac troponin, this amino acid resides in an alpha-helix that directly contacts with TnT. Structural modeling indicates that the substitution of Cys for Arg or Lys at this position would not disrupt the global structure of troponin. To evaluate the functional significance of the different size and charge between the Arg and Cys side chains, protein-binding assays using purified turkey cardiac TnI expressed in Escherichia coli were performed. The results show that the R111C substitution lowered binding affinity to TnT, which is potentially compensatory to the increased TnI-binding affinity of the cardiomyopathy-related cardiac TnT splicing variant. Therefore, the fixation of the cardiac TnI Cys(111) allele in the wild turkey population and the corresponding functional effect reflect an increased fitness value, suggesting a novel target for the treatment of TnT myopathies.

Contraction of vertebrate striated muscle is powered by actomyosin ATPase and is regulated by intracellular Ca 2ϩ through the thin filament-based troponin-tropomyosin system. The troponin complex consists of three protein subunits: the Ca 2ϩ -binding subunit troponin C (TnC) 1 , the tropomyosinbinding subunit troponin T (TnT), and the inhibitory subunit troponin I (TnI). During muscle contraction, cytoplasmic Ca 2ϩ increases, binds to TnC, and evokes a series of conformational changes in the thin filament (1)(2)(3). These thin filament protein interactions translate the cytosolic [Ca 2ϩ ] signal into an activation of actomyosin ATPase, resulting in the development of force. Among the three troponin subunits, TnC belongs to a family of Ca 2ϩ -signaling proteins that includes calmodulin and myosin light chains (4), whereas TnI and TnT are striated muscle-specific proteins encoded by closely linked genes (5). The interaction between TnT and TnI is an essential component of the Ca 2ϩ -regulated allosteric regulation of muscle contraction (6 -8). The co-evolutionary relationship between TnT and TnI (5) indicates their interrelated structure and function.
We reported previously that turkeys with inherited dilated cardiomyopathy (DCM) and heart failure express an unusually low M r cardiac TnT resulting from the splice-out of a normally conserved 11 amino acid N-terminal segment encoded by exon 8 (9). The deletion of the exon 8-encoded segment from cardiac TnT alters binding affinities to TnI and tropomyosin and increases the calcium sensitivity of actomyosin ATPase. Expression of the exon 8-deleted cardiac TnT occurs prior to the development of DCM in turkeys, indicating a role in pathogenesis. We also demonstrated similar aberrant splicing of cardiac TnT in mammals that exhibit a high incidence of DCM (10). Exclusion of the same exon as the turkey exon 8 was found in canine cardiac TnT. Transgenic expression of the aberrantly spliced cardiac TnT variant in the mouse heart resulted in altered cardiomyocyte contraction (10). These findings lead to the hypothesis that changes in thin filament Ca 2ϩ -sensitivity resulting from the expression of abnormally spliced cardiac TnT variants may affect myocardial function and contribute to the pathogenesis and pathophysiology of cardiomyopathy and heart failure.
Alternative splicing of the N-terminal region of TnT modulates the overall molecular conformation. One of the primary consequences of physiological or pathological alterations in TnT structure is the effect on its interaction with TnI (11)(12)(13). The structure-function relationship of TnT and TnI has been studied extensively. Protein binding studies (6,7), functional assays using various TnT (6) and TnI (7) fragments, and the newly published crystallography structure of the partial human cardiac troponin complex (8) have identified the TnT-TnI contact sites. Of specific interest is the TnT-TnI double ␣-helical structure corresponding to amino acid residues E 226 -K 276 of cardiac TnT and F 90 -R 136 in cardiac TnI. Determination of the TnT-TnI interface structure provides valuable guidance in understanding the structure-function relationship of the troponin subunits.
In the present study, we characterize a unique polymorphism of cardiac TnI observed in wild turkey hearts. cDNA cloning and sequencing revealed a single amino acid substitution of Cys for Arg 111 in the novel turkey cardiac TnI. Arg 111 in avian cardiac TnI corresponds to Lys 117 in human cardiac TnI and is conserved in cardiac and skeletal muscle TnIs across the vertebrate phylum. This amino acid resides in the TnT-TnI contacting ␣-helix (8). Structural modeling indicates that the substitution of Cys for Arg or Lys at this position would not disrupt the overall structure of TnI. Protein-binding assays using purified turkey cardiac TnI expressed in Escherichia coli were carried out to investigate the potential effect of the size and charge differences between Arg 111 and Cys 111 side chains on the interaction between TnI and TnT. The results demonstrate that the R111C substitution decreased the binding affinity of TnI for TnT, which is potentially compensatory to the increased TnI-binding affinity of the DCM-related cardiac TnT splicing variant. These results suggest that the fixation of the cardiac TnI Cys 111 allele in the wild turkey population reflects a selection value and indicate a novel target for the treatment of TnT myopathies.

EXPERIMENTAL PROCEDURES
Turkey Cardiac Muscle Samples-Domestic turkey samples were obtained from commercial sources and included both Giant White and Broad Breasted Bronze strains. The heart samples were frozen immediately after the birds were sacrificed and stored at Ϫ80°C until use. Skeletal muscle samples were collected in the same manner for use as controls. Wild turkey hearts were collected in Northeast Ohio during the spring hunting season. The hearts were harvested within 1 h postmortem, kept in ice during transportation, and then stored at Ϫ80°C until use. We have demonstrated previously that no detectable cardiac TnI degradation occurred when mouse corpses were stored at room temperature for 8 h (14). The anticipated integrity of cardiac TnI in the wild turkey heart samples was confirmed by later sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and Western blot analysis. Triton X-100-washed cardiac myofibrils were prepared from the wild turkey ventricular muscle, as described previously (9).
SDS-PAGE and Western Blotting-Ventricular muscle tissues or myofibrils were homogenized in SDS-PAGE sample buffer containing 1% SDS, heated to 80°C for 5 min, and clarified by centrifugation. Total protein extracts were resolved by 14% Laemmli gel with an acrylamide: bisacrylamide ratio of 180:1 or by 12% Laemmli gel with an acrylamide: bisacrylamide ratio of 29:1. Resulting gels were stained with Coomassie Blue R250 to reveal the resolved protein bands, and duplicate gels were electrically blotted to nitrocellulose membranes, as described previously (11). After blocking in 1% bovine serum albumin (BSA), the nitrocellulose membranes were incubated with an anti-TnI monoclonal antibody (mAb) (TnI-1) (15) or an anti-cardiac and slow TnT mAb (CT3) (16). The membranes were then washed with high stringency using Tris-buffered saline containing 0.5% Triton X-100 and 0.05% SDS, incubated with alkaline phosphatase-labeled anti-mouse IgG second antibody (Sigma), washed again, and developed in 5-bromo-4-chloro-3indolylphosphate/nitro blue tetrazolium substrate solution, as described previously (11).
cDNA Cloning-Reverse transcription-PCR (RT-PCR) cloning of cDNAs encoding the turkey cardiac TnI variants was used to determine the protein primary structure. The turkey belongs to the avian order of Galliformes that also includes the domestic chicken (17). To obtain primer sequences appropriate for the cloning of turkey cardiac TnI cDNA, we first extended the published partial chicken cardiac TnI cDNA sequence (18) to full-length. A cDNA library was constructed from poly(A) ϩ mRNA extracted from 5-week-old White Leghorn chicken hearts, as described previously (19). Briefly, total RNA was prepared from the ventricular muscle with TRIzol reagent (Invitrogen) according to the manufacturer's protocol. Poly(A) ϩ RNA was isolated by using a biotinylated-oligo(dT) selection method (the poly(A) tract system II from Promega) as described by the manufacturer. Using reagents from Stratagene, double-stranded cDNA was synthesized from 5 g of poly(A) ϩ RNA with EcoRI-XhoI unidirectional primer-linkers in which the first strand DNA synthesis was carried out using equal molars of oligo dT and random hexamer primers. The cDNA synthesized was unidirectionally ligated to the EcoRI/XhoI-digested ZAPII phage arms, packaged in vitro using the Gigapack-Gold phage extracts and propagated with XL1-Blue MRFЈ E. coli host cells.
From the chicken cardiac cDNA library, the exon 5-8 region of chicken cardiac TnI cDNA was amplified by PCR using a pair of forward and reverse primers designed according to the published partial chicken cardiac TnI cDNA sequence (18). The 5Ј-region of chicken cardiac TnI cDNA was amplified from the library in a separate PCR using a reverse primer in the exon 7 region paired with T3 primer (in this phage library, a forward binding site for the T3 primer is present flanking the 5Ј end of the cDNA insert). The PCR-amplified cDNA fragments were cloned into the pCR4-TOPO vector using the TOPO TA cloning kit (Invitrogen) following the manufacturer's instructions and sequenced using the dideoxy chain-termination method. Full-length chicken cardiac TnI cDNA was constructed by pasting restriction fragments of the 5Ј and 3Ј cDNAs.
cDNA encoding the turkey cardiac TnI was then cloned from domestic and wild turkey hearts by RT-PCR using primers synthesized according to the chicken cardiac TnI cDNA sequence. As described previously (9), turkey ventricular muscle RNA was isolated using the TRIzol reagent. 2 g of the total cardiac RNA was used to synthesize cDNA from all poly(A)ϩ RNA by reverse transcription using an anchored oligo dT primer (5Ј-TTTTTTTTTTTTTTTTTTTV-3Ј, where V ϭ A, C, or G). Double-stranded cDNA encoding turkey cardiac TnI was then amplified by PCR with a forward primer corresponding to the region of the translation initiation codon (5Ј-GGTGCATATGGCTGAGGAGGAGGAGC-C-3Ј) and a reverse primer (5Ј-ATGCCACAACGCTGCCCTTAAAGGT-GA-3Ј) synthesized corresponding to a 3Ј untranslated sequence 97-124 nucleotides downstream of the stop codon. Resultant PCR products were cloned into the pCR4-TOPO vector as above. Recombinant plasmid DNA was purified, and the cDNA insert was sequenced.
Construction of N-terminal Truncated Turkey Cardiac TnI-Polymerase chain reaction was used to generate a truncated cDNA encoding turkey cardiac TnI without the cardiac specific N-terminal region. A forward primer (5Ј-GGGCATATGGCTGTGGAGCCCCA-3Ј) was synthesized containing an NdeI restriction enzyme site (underlined) and incorporating a Met codon prior to the turkey cardiac TnI codon for amino acid 23. This mutagenesis primer was used together with the T7 primer flanking the 3Ј end of the cloned turkey cardiac TnI cDNA in the pCR4-TOPO vector sequence to generate the truncated cDNA template by PCR. The PCR products were digested with the restriction enzymes NdeI and EcoRI and ligated into the compatibly cut pAED4 vector. Recombinant plasmid DNA was purified, and the cDNA insert was sequenced as above to verify the construction.
Expression and Purification of N-terminal Truncated Turkey Cardiac TnI-The N-terminal truncated turkey cardiac TnI (TcTnI-ND) was expressed in E. coli from the cDNA construct for functional characterization. As described previously (9), BL21(DE3)pLysS E. coli cells were transformed with the recombinant pAED4 plasmid and cultured in liquid media containing ampicillin and chloramphenicol. The culture was induced at mid-log phase with isopropyl-1-thiol-␤-D-galactoside. After 3 additional h of culture, the bacterial cells were harvested and lysed by three passes through a French cell press in 5 mM EDTA, 15 mM ␤-mercaptoethanol, 0.5 mM phenylmethylsulfonyl fluoride, 50 mM Tris-HCl, pH 8.0. The protein extracts were fractionated by ammonium sulfate precipitation at 0°C. The fraction between 20 -40% saturation was dialyzed against 0.5 mM EDTA containing 6 mM ␤-mercaptoethanol. After dialysis, the fraction was brought to 6 M urea, 0.1 mM EDTA, 6 mM ␤-mercaptoethanol, 0.1 mM phenylmethylsulfonyl fluoride, 10 mM sodium acetate, pH 6.6, clarified by centrifugation, and chromatographed on a CM52 cellulose cation-exchange column equilibrated in the same buffer. The column was eluted with a linear KCl gradient (0 -500 mM), and the protein peaks were analyzed by SDS-PAGE. The fractions containing TcTnI-ND were verified by Western blotting using the anti-TnI mAb TnI-1, dialyzed against de-ionized water containing 0.1% formic acid and 6 mM ␤-mercaptoethanol, and concentrated by lyophilization. The N-terminal truncated cardiac TnI CM52 fractions were further purified to homogeneity by Sephadex G-75 gel filtration chromatography in 0.5 M KCl, 6 M urea, 0.1 mM EDTA, 6 mM ␤-mercaptoethanol, 10 mM imidazole, pH 7.0, as described previously (9). The pure TcTnI-ND peak was identified by SDS-PAGE, dialyzed against 0.5% formic acid, and lyophilized. All of the dialysis and chromatography steps were carried out at 4°C.
Preparation of Other Proteins-Intact chicken cardiac TnI was purified from adult ventricular muscle as described previously (9). The wild-type (WT) turkey cardiac TnT and exon 8-excluded (⌬E8) turkey cardiac TnT were expressed from cloned cDNA and purified as described previously (9).
Protein Conformation Assays-Enzyme-linked immunosorbent assay (ELISA) epitope analysis (11) was carried out to examine whether the deletion of the N-terminal 22 amino acids from the turkey cardiac TnI (TcTnI-ND) affects the global conformation in comparison to that of the intact chicken cardiac TnI. The mAb TnI-1 against an epitope in the C-terminal domain of TnI (15) was used to monitor overall conformational changes that alter the antibody binding affinity. Similar to that described previously (9), purified TcTnI-ND and intact chicken cardiac TnI were dissolved in buffer A (0.1 M KCl, 3 mM MgCl 2 , 10 mM PIPES, pH 7.0) and coated on microtiter plates by incubation at 4°C overnight. Unbound cardiac TnI was washed away with buffer A containing 0.05% Tween 20 (buffer T), and the remaining plastic surface was blocked with 1% BSA in buffer T. The immobilized cardiac TnI was incubated with serial dilutions of the TnI-1 mAb at room temperature for 2 h. After washes with buffer T to remove unbound first antibody, the plates were further incubated with horseradish peroxidase-conjugated anti-mouse immunoglobulin second antibody (Sigma) at room temperature for 45 min. The unbound second antibody was washed away with buffer T and H 2 O 2 /2,2Ј-azinobis-(3-ethylbenzthiazolinesulfonic acid) was added for substrate reaction. A 405 of each assay well was recorded at a series of time points by an automated microplate reader (Bio-Rad Benchmark). The A 405 values in the linear course of the color development were used to plot the TnI-1 mAb titration curves, allowing the quantification of binding affinity to the epitope on cardiac TnI. All experiments were done in triplicate.
Protein Binding Assays-The ELISA-based solid-phase protein binding assay (11) was used to investigate the interactions of cardiac TnI with wild-type and exon 8-excluded turkey cardiac TnT.
Standard ELISA (11) was first carried out on purified wild-type and exon 8-excluded turkey cardiac TnT to validate the equal coating of the two TnT variants onto the microtiter plate. Turkey cardiac TnTs were dissolved in buffer A and incubated at varied concentrations in the microtiter plates. Cardiac TnT bound to the plate was assessed by incubation with a saturating concentration of the anti-TnT mAb CT3 and quantified by standard ELISA, as described above. The A 405 values for the maximally bound CT3 mAb were used as a quantitative indicator of the amount of turkey cardiac TnT immobilized in the well of the microtiter plates.
To assess TnI-TnT interactions by solid-phase binding experiments, purified cardiac TnT or BSA control was dissolved at 5 g/ml in buffer A and coated onto triplicate wells of the microtiter plates by incubation at 4°C overnight. After washes with buffer T to remove the unbound protein, the plates were blocked with buffer T containing 1% BSA. After washes with buffer T, the plates were then incubated with serial dilutions of cardiac TnI in buffer T containing 0.1% BSA at room temperature for 2 h. After washes with buffer T, the cardiac TnI bound to immobilized cardiac TnT was quantified by means of the anti-TnI mAb TnI-1 (15) and standard ELISA procedure.
Computer Modeling-To investigate the potential structural effect of R111C substitution in turkey cTnI, we calculated the lowest energy homology models of the troponin complex, including these variants. The 52-kDa human cardiac troponin crystal structure (8) was retrieved from the NCBI website and cTnI residues 114 -117 were changed from the original human cardiac to that of the turkey. Structures containing either Arg or Cys at residue 117 (equivalent to avian residue 111 and corresponding to the two turkey cTnI variants) were compared by analysis in the low mode search of Macromodel 7.2 (Schrodinger, Inc.) using an OPLS-AA force field and GBSA solvent model.
Data Analysis-The DNA and protein sequence analyses were done using computer programs from DNAStar. Statistical analysis of the ELISA antibody titration and protein binding assays was done by a Student's t test. All values are given as mean Ϯ S.D.

A Novel Cardiac TnI in Wild Turkey Hearts with Unique Gel
Mobility-Normally the adult mammalian and avian hearts express only one cardiac TnT and one cardiac TnI. Previously, we reported the domestic turkey expresses two cardiac TnT, with the low M r variant resulting from abnormal RNA splicing to the exclusion of the exon 8 segment. On cardiac muscle homogenates, SDS-PAGE (Fig. 1A) and Western blot analysis using the anti-cardiac and slow TnT mAb CT3 demonstrate that the wild turkey heart expresses the additional low M r cardiac TnT variant as previously demonstrated in the domestic turkey heart ( Fig. 1B; Ref. 9). The use of the anti-TnI mAb TnI-1 Western blot of domestic turkey heart identified a single cardiac TnI of identical SDS-gel mobility to that of the chicken cardiac TnI (Fig. 1C). Interestingly, an additional slow migrating cardiac TnI was detected in the heart of the wild turkeys. The heart of wild turkey #1 (shown in Fig. 1C) expressed both the normal and a slower migrating variant of cardiac TnI, whereas wild turkey #2 expressed only the slower migrating cardiac TnI (Fig. 1C). The novel, slow migrating TnI detected in FIG. 1. A novel cardiac TnI in the heart of wild turkeys. A, SDS-PAGE of total muscle homogenate resolved on 14% Laemmli gel with an acrylamide:bisacrylamide ratio of 180:1 demonstrates the integrity of the cardiac muscle samples. B, Western blot using the anticardiac and slow TnT mAb CT3 demonstrates that, unlike the single cardiac TnT band found in adult human, mouse, and chicken hearts, the adult wild turkey hearts express an additional low molecular mass cardiac TnT (arrow) identical to that of the DCM-related ⌬E8 cardiac TnT reported previously in the domestic turkey. The ⌬E8 cardiac TnT is distinguishable from slow skeletal muscle TnT detected by CT3 in turkey mixed fiber. C, Western blot of cardiac muscle homogenates resolved on 12% Laemmli gel with an acrylamide:bisacrylamide ratio of 29:1 using the anti-TnI mAb TnI-1 demonstrates that all mammalian and avian species examined express a single cardiac TnI in the adult heart. In contrast to the similarly migrating single cardiac TnI in the chicken and domestic turkey hearts, an additional cardiac TnI of slower SDS-gel mobility was detected in the heart of wild turkey #1, whereas the heart of wild turkey #2 expressed only the slow migrating cardiac TnI (arrow). This slow migrating cardiac TnI is also distinct from the fast and slow skeletal muscle TnIs. D, the slow migrating cardiac TnI was proportionally incorporated into the wild turkey cardiac myofilaments, as shown by the mAb TnI-1 Western blot on extensively washed myofibrils. E, expression of cloned cDNA encoding chicken cardiac TnI. Western blot performed as in C on total protein extracted from E. coli transformed with the chicken cardiac TnI cDNA-expressing plasmid confirmed that the protein encoded is recognized by the anti-TnI mAb TnI-1 and is of identical size to that of the cardiac TnI in adult chicken heart. Low and high molecular mass bands in the E. coli extract were also recognized by mAb TnI-1, indicating degradation and aggregation, respectively, of chicken cardiac TnI in bacterial cells. the wild turkey hearts is distinct from fast or slow skeletal muscle TnI, as its SDS-gel mobility was different from that seen in a mixed-fiber skeletal muscle of a newly hatched domestic turkey and the fast skeletal muscle of a 15-day-old domestic turkey (Fig. 1C). The slow migrating cardiac TnI was proportionally incorporated into the myofilaments as shown by the Western blot of extensively washed wild turkey cardiac myofibrils (Fig. 1D).
Cloning of the 5Ј Region of Chicken Cardiac TnI cDNA-Based on the 3Ј partial sequence of chicken cardiac TnI cDNA reported by Hastings et al. (18), we cloned a full-length chicken cardiac TnI cDNA by PCR from a unidirectional chicken cardiac cDNA library. Expression of the cloned cDNA in E. coli yielded a protein recognized by the anti-TnI mAb TnI-1 with identical size to cardiac TnI in the chicken heart (Fig. 1E). Sequence of the full-length chicken cardiac TnI cDNA has been submitted to GenBank TM /EBI with the accession number AY463242. The physical properties of the protein deduced from the cDNA sequence demonstrate that the chicken cardiac TnI is of similar molecular weight to that of human and mouse cardiac TnI. The fact that the chicken cardiac TnI (208 amino acids, M r ϭ 23,627, isoelectric point (pI) ϭ 9.96) exhibits a significantly faster SDS-gel migration rate than that of the similar-sized human (210 amino acids, M r ϭ 24,037, pI ϭ 9.87) and mouse (211 amino acids, M r ϭ 24,258, pI ϭ 9.57) cardiac TnI (Fig. 1C) demonstrates that SDS-gel mobility is a sensitive indicator of differences in TnI primary structure.
A Single Nucleotide Mutation Results in a Cys Substitution for Arg 111 in Wild Turkey Cardiac TnI-Based upon the fulllength chicken cardiac TnI cDNA sequence, we were able to use RT-PCR to clone cDNAs encoding the fast and slow migrating turkey cardiac TnIs. Sequencing analysis revealed that chicken, domestic turkey, and the faster migrating wild turkey cardiac TnIs are identical in amino acid sequence. Therefore, the domestic turkey cardiac TnI sequence is considered the wild type. Consistent with the expression patterns of the fast migrating (chicken-like, wild-type) and/or slow migrating cardiac TnI in the wild turkey hearts (Fig. 1C), wild-type cardiac TnI cDNA clones were obtained from wild turkey #1 but not from wild turkey #2. cDNA clones isolated from wild turkey #2 were different from the wild-type cardiac TnI by a single G 3 T transversion of nucleotide 331 located in the exon 6 sequence. This nucleotide transversion resulted in the mis-sense mutation of codon 111, which corresponds to a change of Arg 111 into Cys (R111C) (Fig. 2A). The cDNA cloning data demonstrate that the two turkey cardiac TnI bands of different SDS-gel mobility (Fig. 1C) represent a polymorphism caused by two cardiac TnI alleles. The cDNA sequences of the Arg 111 and Cys 111 turkey cardiac TnI alleles have been submitted to Gen-Bank TM /EBI with the accession numbers AY463243 and AY463244, respectively. Whereas the domestic turkey is a homozygote of the wild-type Arg 111 allele, wild turkey #1 is a heterozygote of Arg 111 /Cys 111 alleles, and wild turkey #2 is a homozygote of the Cys 111 allele. Because substitution of the Cys for the wild-type Arg 111 results in a decreased protein mass and slightly acidic shift of pI (23,574 and 9.78 versus 23,627 and 9.96, respectively), the slower SDS-gel mobility of the turkey cardiac TnI Cys 111 was not due to a change in molecular mass but rather indicates a change in protein conformation. The substitution of a neutral side chain for a positively charged side chain may confer structural and functional alterations in the wild turkey cardiac TnI.
Alignment of published TnI amino acid sequences demonstrates the conservation of a positively charged residue (Arg or Lys) at the Arg 111 position across the vertebrate phylum (Fig.  2B). The sequences flanking this residue are also highly con-served. Based upon the partial human cardiac troponin crystal structure recently published by Takeda et al. (8), the Arg 111 in avian cardiac TnI (equivalent to Lys 117 in human cardiac TnI) is located in an ␣-helix that directly contacts TnT (Fig. 3A) within the coiled-coil IT arm structure in the troponin complex. The substitution of Arg for Lys 117 in human cardiac TnI does not result in a structural change in the lowest energy homology model (Fig. 3B). This result is consistent with the fact that both Lys and Arg are normally found at this position in TnI structure (Fig. 2B). Similarly, substitution of Cys for Lys 117 does not result in an overall structure change in this TnI-TnT interface. In addition to changing the side chain size and charge, modeling the substitution of Cys for Lys 117 predicts an additional hydrogen bond between TnI Cys 117 and Ile 114 (Met 108 in turkey cardiac TnI) in the same helix (Fig. 3C) although not in the lowest energy homology model. The data suggest that the substitution of Cys for Arg 111 in wild turkey cardiac TnI does not disrupt the structure of TnI or the TnI-TnT interface. This is in agreement with the observation that the cardiac TnI Cys 111 allele is fixed in the wild turkey population without an apparent negative impact. Nonetheless, the notable alteration in the TnI-TnT interface due to the R111C substitution suggests functional modifications.
Expression, Purification, and Characterization of N-terminal Truncated Turkey Cardiac TnI-Due to the extremely low level expression of the intact avian cardiac TnI in E. coli (Fig. 1E), we constructed cDNAs encoding N-terminal truncated turkey

FIG. 2. Primary structure of turkey cardiac TnI and a G 3 T transversion in wild turkey cardiac TnI gene, resulting in a substitution of Cys for Arg 111 .
A, cDNA cloning and sequencing revealed the primary structures of the slow and fast migrating turkey cardiac TnIs. The deduced amino acid sequence of the wild-type turkey cardiac TnI was found to be identical to the chicken cardiac TnI, with Arg at position 111. A class of cDNA cloned from both wild turkey #1 and #2 corresponding to the slow migrating cardiac TnI differs from the wild type by a single G 3 T nucleotide transversion, resulting in the substitution of Cys for Arg 111 (outlined by the box). The initiation site of TcTnI-ND is indicated in the sequence. B, the protein primary structural map of cardiac TnI and sequence alignment for the region flanking Arg/Cys 111 demonstrates that this position is conserved (Arg or Lys) in all TnI sequenced to date (26 -35). The highlighted residues in the human cardiac TnI sequence are residues that contact with TnT in the crystal structure of troponin (8).
cardiac TnI (TcTnI-ND) containing either Arg or Cys at position 111 for functional characterization. The N-terminal truncation was designed according to a naturally occurring cardiac TnI proteolytic fragment originally found in a rat model of simulated microgravity (14). The TcTnI-ND construct consists of amino acids 23-208 ( Fig. 2A) and preserves the core structure of TnI, as observed in fast and slow skeletal muscle TnI (Fig. 4A). The embryonic heart exclusively expresses slow skeletal muscle TnI lacking the cardiac-specific N-terminal domain (20,21). Therefore, TnI without the N-terminal region should be able to sustain cardiac muscle contraction.
The calculated molecular mass and pI are 21,070 Da and 10.05 for TcTnT-ND-Arg 111 and 21,017 Da and 9.83 for TcTnT-ND-Cys 111 . Expression of the TcTnI-ND cDNA constructs in E. coli yielded a high level expression of proteins that were readily detectable by the TnI-1 mAb (Fig. 4B). The TcTnI-ND-Arg 111 and TcTnI-ND-Cys 111 proteins also clearly exhibit the fast and slow SDS-gel migration rates observed in the intact turkey cardiac TnI variants (Fig. 4B). Large scale expression of TcTnI-ND-Arg 111 and TcTnI-ND-Cys 111 in E. coli yielded good quantities of protein, demonstrating that the removal of N-terminal 22 amino acids significantly improved the compatibility of avian cardiac TnI to the bacterial host cell. The TnI purification methods developed in this study are highly effective in obtaining high quality materials for functional characterization (Fig. 5).

FIG. 3. R111C at the TnI-TnT interface in troponin complex.
The TnI-TnT interface structure corresponding to the region flanking the R111C polymorphism is modeled in this figure. The helices are depicted as ribbons: green for TnI, and red for TnT. Hydrogens are not shown for clarity. Backbone atoms are colored green, and side chain atoms are colored according to their element type (carbon, gray; oxygen, red; nitrogen, blue; sulfur, yellow). The human cTnI crystallographic structure is shown in A. Residues 114 -117 were then changed from the original human cardiac to the turkey cTnI and a normal mode analysis was performed with the low mode search of Macromodel 7-2 (Schrodinger, Inc.) using an OPLS-AA force field and a GBSA solvent model. Only the four residues illustrated were allowed to vary during the conformational search. All other residues were frozen in the crystallographic position. The lowest energy homology models are shown in B (Arg 117 ) and C (Cys 117 ). Backbone carbonyls to which hydrogen bonding was observed during the normal mode vibrations are colored purple based on 794 structures from the Arg 117 simulation and 973 structures from the Cys 117 simulation. Note that hydrogen bonding of Cys 117 to Ile 114 (Met 108 in turkey cardiac TnI) was observed but not in the lowest energy structure. The difference between the alpha carbons of the homology models and the original crystal structure was 0.4 Å RMSD.

FIG. 4. Construction and expression of the N-terminal truncated turkey cardiac TnI.
A, as illustrated in the outline comparison of cardiac, slow skeletal, and fast skeletal muscle TnI primary structures, the N-terminal domain of cardiac TnI is a unique extension that is not present in skeletal muscle TnI and can be removed proteolytically during physiological adaptation (14). cDNAs were constructed accordingly, encoding amino acids 23-208 and containing Arg or Cys at position 111 (TcTnI-ND-Arg 111 and TcTnI-ND-Cys 111 ). B, 12% Laemmli gel with an acrylamide:bisacrylamide ratio of 29:1 and mAb TnI-1 Western blot of turkey heart homogenates and E. coli-expressed TcTnI-ND-Arg 111 and TcTnI-ND-Cys 111 demonstrate similar shifts in gel migration rate because of the presence of wild-type Arg 111 or the Cys 111 substitution in both intact and N-terminal-truncated cardiac TnIs.
To validate the use of the N-terminal truncated turkey cardiac TnI for the functional characterization of the R111C substitution, we compared the TcTnI-ND-Arg 111 and TcTnI-ND-Cys 111 to the intact chicken cardiac TnI by ELISA epitope conformation analysis. The results in Fig. 6A demonstrate that the binding affinity and maximal binding (inset) of mAb TnI-1, with its epitope in the C-terminal region of TnI, to TcTnI-ND-Arg 111 and TcTnI-ND-Cys 111 were identical to that of the intact chicken cardiac TnI. These results indicate that the removal of the N-terminal 22 amino acids does not destroy the global structure of cardiac TnI.
To further verify that the N-terminal truncated cardiac TnI is representative for the intact cardiac TnI in binding to cardiac TnT, we compared TcTnI-ND-Arg 111 and the intact chicken cardiac TnI in ELISA solid-phase protein binding experiments. The results in Fig. 6B demonstrate that there is no detectable difference between the intact chicken cardiac TnI (Arg 111 ) and the TcTnI-ND-Arg 111 in binding affinity or maximal binding (inset) to the wild-type turkey cardiac TnT. Therefore, under the experimental conditions, removal of the N-terminal domain of the turkey cardiac TnI did not affect the binding to TnT. The results justify the use of N-terminal truncated cardiac TnI in comparing the Arg 111 and Cys 111 variants for TnT interaction.
Substitution of Cys for Arg 111 in Turkey Cardiac TnI Alters the Binding to Cardiac TnT-The TcTnI-ND-Arg 111 and TcTnI-ND-Cys 111 proteins were examined for the effect of the R111C substitution within the TnI-TnT interface on the binding of TnI to the turkey cardiac TnT variants. To validate the solid-phase TnT binding assay, we first demonstrated that saturated amounts of the anti-TnT mAb CT3 bound to the WT and ⌬E8 TcTnT (Fig. 5B) coated on the microtiter plates were identical A, ELISA epitope affinity titration curves and maximum binding levels (the A 405 reading shown in inset) for mAb TnI-1 against a C-terminal epitope on TnI show no significant differences among the intact chicken cardiac TnI, TcTnI-ND-Arg 111 , and TcTnT-ND-Cys 111 immobilized on a microtiter plate. The results demonstrate that deletion of the N-terminal domain from turkey cardiac TnI does not disrupt the global structure of TnI. B, solid-phase protein binding curves demonstrate that there is no significant difference between intact chicken cardiac TnI (Arg 111 ) and TcTnI-ND-Arg 111 in binding affinity to wild-type turkey cardiac TnT immobilized on a microtiter plate, as defined by the TnI concentration required for 50% of maximum binding. The maximal binding levels of intact chicken cardiac TnI and TcTnI-ND-Arg 111 to wild-type turkey cardiac TnT were also similar (the A 405 reading shown in inset). The results indicate that deletion of the N-terminal domain from turkey cardiac TnI did not affect binding to cardiac TnT under the assay conditions. (Fig. 7A). These results indicate that similar amounts of the WT and ⌬E8 TcTnT were immobilized on the microtiter plates.
Using the solid-phase protein binding assay, we demonstrated that the binding affinity of TcTnI-ND-Cys 111 to TcTnT-WT was significantly lower than that of TcTnI-ND-Arg 111 (Fig. 7B). The lowered binding affinity was reflected by the significantly higher concentration of TcTnI-ND-Cys 111 required to reach 50% of maximum binding compared with that in TcTnI-ND-Arg 111 (0.056 M Ϯ 0.006 versus 0.022 M Ϯ 0.004, p Ͻ 0.01), demonstrating a higher K d during equilibrium binding. Likewise, TcTnI-ND-Cys 111 exhibited decreased binding affinity to TcTnT-⌬E8, as compared with that of TcTnI-ND-Arg 111 (Fig. 7C; the concentrations required for 50% of maximum binding were 0.044 M Ϯ 0.004 and 0.013 M Ϯ 0.001, respectively; p Ͻ 0.01). Fig. 7D summarizes the effects of the R111C substitution in cardiac TnI on binding affinities for the WT versus ⌬E8 TcTnT. We showed previously that the DCMrelated TcTnT-⌬E8 had a higher binding affinity to cardiac TnI (wild type) compared with the TcTnT-WT, correlating to an increased Ca 2ϩ sensitivity in the activation of myofilament ATPase (9). This cardiac TnT abnormality was also detected in the TcTnI-ND binding experiments (Fig. 7D). By lowering the abnormally high binding affinity of TcTnT-⌬E8 to TnI, the cardiac TnI Cys 111 polymorphism in wild turkey hearts may be compensatory to the DCM-related TcTnT-⌬E8 abnormality.

Sensitivity of TnI SDS-gel Mobility to Structural Modifications-
The turkey cardiac TnI R111C polymorphism was originally identified as a gel mobility shift in SDS-PAGE (Fig. 1C). SDS-PAGE is one of the standard methods used to determine the apparent molecular weight of proteins because charge and shape variations are diminished when proteins are saturated with SDS (22). However, not all proteins show a linear rela-tionship between their molecular mass and SDS-gel migration rate (23). Although mammalian and avian cardiac TnI are of similar molecular weight and amino acid composition, both chicken and turkey cardiac TnI migrate significantly faster than that of the mammalian cardiac TnI (Fig. 1C). This observation suggests that the molecular conformation and/or flexibility of TnI, even in the presence of SDS, is sensitive to minor structural alterations. A smaller SDS-gel mobility shift has also been observed in mouse cardiac TnI with a single amino acid substitution. 2 This hypothesis is consistent with the nature of TnI as an allosteric regulatory protein.
The Cardiac TnI Cys 111 Allele in Wild Turkey Population-Arg 111 in avian cardiac TnI is a conserved residue in all TnIs sequenced to date (Fig. 2B). In contrast, the Cys 111 allele was readily detectable in the wild turkey population. Although the Cys 111 allele can coexist with the Arg 111 allele (wild turkey #1 is an Arg 111 /Cys 111 heterozygote, as shown by both Western blot (Fig. 1C) and cDNA cloning/sequencing), a homozygous Cys 111 individual was also found (wild turkey #2). Therefore, the Cys 111 allele seems to have a high frequency in the wild turkeys. The fixation of this unique polymorphism in the turkey but not other species indicates a specific selection value. It is worth noting that the chicken and turkey are evolutionarily closely related species and that their wild-type cardiac TnI has identical amino acid sequences. According to the avian constraint hypothesis, functional constraint on avian proteins causes the reduction of genetic divergence and, therefore, avian protein structure generally has a lower tolerance to amino acid substitutions compared with that of other species (24). Therefore, the fixation and spread of the cardiac TnI Cys 111 allele in wild turkeys indicates a strongly favored selection value.   7. R111C substitution in turkey cardiac TnI alters binding affinity to WT and ⌬E8 turkey cardiac TnT. A, to verify the comparable coating of turkey cardiac TnT to the microtiter plates of the protein binding assays, ELISA titration showed no significant difference between the levels of saturated binding of mAb CT3 to the wild-type (WT) and exon 8-excluded (⌬E8) turkey cardiac TnT immobilized to the plate. B and C, solid-phase protein binding assays demonstrate the binding affinities of TcTnI-ND-Cys 111 to WT and ⌬E8 TcTnT were significantly lower than those of TcTnI-ND-Arg 111 , as shown by the higher concentrations required for 50% of maximum binding. D, although the binding curves confirm the increased TnI binding affinity of TcTnT-⌬E8 as compared with that of TcTnT-WT (9), the decreased binding affinity of TcTnI-ND-Cys 111 for TnT suggests a compensatory effect. domestic turkey has a high instance of DCM that causes round heart disease with considerable mortality (25). In contrast to the short inactive life span of the domestic turkey, the wild turkey population is under stringent natural selection. We reported previously that the abnormally spliced cardiac TnT variant in the domestic turkey may contribute to the pathogenesis of DCM (9). As shown in Fig. 1B, wild turkeys also express the exon 8-deleted cardiac TnT in cardiac muscle identical to that of domestic turkeys. The presence of this DCM-related cardiac TnT abnormality in the wild turkey would impose a specific selection pressure on functionally related genes. Therefore, it is logical to consider that the cardiac TnI Cys 111 allele is of compensatory value to improve survival of the wild turkey population that otherwise would suffer from a high instance of DCM because of the presence of abnormal cardiac TnT.
Potential Compensatory Effect of the Turkey Cardiac TnI Cys 111 Allele on the DCM-related TnT Abnormality-Protein binding experiments revealed that the R111C substitution lowered the binding affinity of turkey cardiac TnI to TnT, compensatory to the increased TnI-binding affinity of the DCM-related exon 8-deletion in turkey cardiac TnT (Fig. 7). On the other hand, the incorporation of the Cys 111 cardiac TnI into the cardiac myofilaments was not affected, as shown by its proportional incorporation into the wild turkey cardiac myofibrils (Fig. 1D). This finding is consistent with the notion that the R111C substitution in cardiac TnI does not produce drastic effects on troponin structure (Fig. 3).
Supporting the decreased binding affinity of cardiac TnI to TnT, modeling the substitution of Cys for Lys 117 in human troponin structure adds a hydrogen bond between Cys 117 and Ile 114 in the TnI-TnT coiled-coil structure (8, Fig. 3C). This might reduce the flexibility of the TnI structure that interfaces with TnT in the functionally important IT arm of troponin (8).
We have demonstrated previously that the increased TnI-binding affinity is one of the primary functional changes produced by the DCM-related cardiac TnT splicing variant (9, 10). The reduced TnT-binding affinity of the wild turkey cardiac TnI Cys 111 polymorphism is proposed, therefore, to obtain its selection value by compensating for the TnT-originated abnormality in TnT-TnI interaction and Ca 2ϩ -regulation of cardiac muscle contraction. It is worth noting that the binding affinities of cardiac TnI Cys 111 for both wild type and ⌬E8 cardiac TnT are reduced to levels lower than that between cardiac TnI Arg 111 and wild-type cardiac TnT, which is the normal reference point (Fig. 7D). However, the significantly weakened binding between cardiac TnI Cys 111 and TnT may minimize the effect of cardiac TnT abnormality on myocardial contraction. The bio-chemical and physiological mechanisms deserve further investigation.
In summary, as an example of the combined power of protein polymorphism analysis, molecular cloning, and protein structure-function characterization, the finding of the cardiac TnI Cys 111 allele in wild turkeys with potentially compensatory effect on TnT abnormality demonstrates the closely related structure-function relationship between TnI and TnT and suggests a novel target for the treatment of TnT myopathies.