Altered Regulatory Properties of Human Cardiac Troponin I Mutants That Cause Hypertrophic Cardiomyopathy*

Cardiac troponin I (cTnI) is the inhibitory component of the troponin complex and is involved in the calcium control of heart muscle contraction. Recently, specific missense mutations of the cTnI gene (TNNI3) have been shown to cause familial hypertrophic cardiomyopathy (HCM). We have analyzed the functional effects of two HCM mutations (R145G and R162W) using purified recombinant cTnI. Both mutations gave reduced inhibition of actin-tropomyosin-activated myosin ATPase, both in experiments using cTnI alone and in those using reconstituted human cardiac troponin under relaxing conditions. Both mutant troponin complexes also conferred increased calcium sensitivity of ATPase regulation. Studies on wild type/R145G mutant mixtures showed that the wild type phenotype was dominant in that the inhibition and the calcium sensitivity conferred by a 50:50 mixture was more similar to wild type than expected. Surface plasmon resonance-based assays showed that R162W mutant had an increased affinity for troponin C in the presence of calcium. This observation may contribute to the increased calcium sensitivity found with this mutant and also corroborates recent structural data. The observed decreased inhibition and increased calcium sensitivity suggest that these mutations may cause HCM via impaired relaxation rather than the impaired contraction seen with some other classes of HCM mutants.

Cardiac troponin I (cTnI) is the inhibitory component of the troponin complex and is involved in the calcium control of heart muscle contraction. Recently, specific missense mutations of the cTnI gene (TNNI3) have been shown to cause familial hypertrophic cardiomyopathy (HCM). We have analyzed the functional effects of two HCM mutations (R145G and R162W) using purified recombinant cTnI. Both mutations gave reduced inhibition of actin-tropomyosin-activated myosin ATPase, both in experiments using cTnI alone and in those using reconstituted human cardiac troponin under relaxing conditions. Both mutant troponin complexes also conferred increased calcium sensitivity of ATPase regulation. Studies on wild type/R145G mutant mixtures showed that the wild type phenotype was dominant in that the inhibition and the calcium sensitivity conferred by a 50:50 mixture was more similar to wild type than expected. Surface plasmon resonance-based assays showed that R162W mutant had an increased affinity for troponin C in the presence of calcium. This observation may contribute to the increased calcium sensitivity found with this mutant and also corroborates recent structural data. The observed decreased inhibition and increased calcium sensitivity suggest that these mutations may cause HCM via impaired relaxation rather than the impaired contraction seen with some other classes of HCM mutants.
Familial hypertrophic cardiomyopathy (HCM) 1 is an autosomal dominant disease of the myocardium. Patients present with a wide range of symptoms including chest pain, dyspnea, syncope, stroke, and sudden death (1). The affected heart shows increased left ventricular mass, myocyte hypertrophy and disarray, and fibrosis. The first HCM-causing mutation to be reported was the R403Q missense mutation in the ␤-myosin heavy chain gene (2); further mutations have since been found in this gene and in 8 other genes all of which encode sarcomeric proteins (3,4). Investigations into the functional abnormalities of the HCM mutant proteins have suggested that the end stage disease is the result of numerous different effects: some mutations result in diminished contractility (for example, mutations at Arg 403 in ␤ myosin heavy chain), some in decreased calcium sensitivity while others enhance contractility or calcium sensitivity (4).
In 1997, Kimura et al. (5) were the first to report HCMcausing mutations in the cardiac troponin I (cTnI) gene (TNNI3). Five missense mutations (R145G, R145Q, R162W, G203S, and K206Q) were reported along with a mutation causing the deletion of one codon (Lys 183 ). Troponin I (TnI) is the inhibitory component of the troponin complex and is involved in the Ca 2ϩ regulation of muscle contraction in both skeletal and cardiac muscle. The protein forms a 1:1:1 complex with troponin C (TnC) and troponin T (TnT) which binds to the thin filament via interactions with both actin and tropomyosin (TM). The binding of Ca 2ϩ by TnC causes a conformational change in the complex which results in a shift in the position of TM with respect to the actin helix, allowing strong force-generating interactions between actin and myosin (6,7). The precise structure of the complex has yet to be elucidated, although the structure of isolated TnC is known (8). Low resolution studies and biochemical data have shown that TnI and TnC bind strongly to form an antiparallel dimer with multiple interaction sites (9,10).
The human cTnI gene encodes a protein of 210 amino acids (11) which is the sole TnI isoform expressed in adult heart. cTnI has considerable homology to the skeletal muscle isoform, the principal difference between the two being a 30-residue N-terminal extension of the cardiac protein which contains 2 serines, the phosphorylation of which by cAMP-dependent protein kinase is known to regulate cTnI function. The most important functional sequence of cTnI is a short central stretch of residues referred to as the inhibitory region, required for the inhibition of actin-TM-activated myosin ATPase. This sequence also binds to TnC in the presence of Ca 2ϩ and as an isolated peptide is capable of the inhibition of actin-TM-activated myosin ATPase and the relaxation of skinned muscle fibers, both of which can be relieved by the addition of TnC in the presence of activating concentrations of Ca 2ϩ (12,13). The minimum inhibitory sequence consists of residues 137 to 148 in the human cardiac protein (104 to 115 in rabbit skeletal muscle TnI). Further studies on skeletal TnI peptides have revealed a "second actin-TM-binding site" (161 to 181 in human cTnI), necessary for enhanced inhibition compared with that achieved with the minimum sequence, and a region critical for the interaction with TnC (149 to 164, a putative "second TnC-binding site") (14). Analysis of truncation mutants has shown that in the case of cTnI, additional C-terminal residues (amino acids 187-198 in the human cardiac sequence) influence actin binding and inhibition (15).
Perturbation of TnI function has been linked to myocardial stunning in which reversible cardiac dysfunction takes place. This phenomenon has been shown to be caused by proteolysis of TnI, initially producing a fragment missing the C-terminal 17 amino acids, and concomitant alteration in thin filament regulation (16). A recent report has shown that transgenic mice expressing this 1-193 truncated protein at 9 -17% of the level of endogenous TnI, developed ventricular dilatation, diminished myocyte contractility, and reduced myofilament calcium responsiveness (17). The vital importance of the cardiac isoform has also been emphasized by the severe phenotype of a cTnI knockout mouse (18).
In order to understand how subtle alterations in troponin I cause disease, we have used purified recombinant human cTnI to compare the in vitro function of wild type TnI with two HCM TnI mutants, one containing the R145G mutation, the other R162W. R145G is one of the two reported HCM mutations that change residue Arg 145 in the inhibitory region and the R162W mutation causes a charge change to a residue located in the overlap between the second TnC-binding site and second actin-TM-binding site as defined by Tripet et al. (14). The functional abnormalities revealed, namely reduced ability to inhibit actin-TM-activated myosin ATPase and an increase in Ca 2ϩ sensitivity of ATPase regulation, suggest that these mutations may result in relaxation abnormalities in vivo. In addition, the finding that the R162W mutant has an increased affinity for TnC supports recent structural data concerning the interaction of these two proteins.

Preparation of Recombinant Human Cardiac Troponin Subunits-
Human cTnI cDNA (11), obtained as a gift from Dr. P. Barton (Imperial College, London, United Kingdom), was amplified by polymerase chain reaction to introduce the silent mutations in codons 2 and 4 suggested by al Hillawi et al. (19). The product was cloned into the bacterial expression vector pMW172 (20) using NdeI and HindIII. The TnI sequence was altered using a two-step polymerase chain reaction-based oligonucleotide-directed mutagenesis protocol to produce the R145G and R162W HCM mutations. The TnI constructs were overexpressed in Escherichia coli strain BL21(DE3)pLysS (21), and the cells lysed after freeze-thawing in buffer containing 5 M urea. Both wild type and mutant TnI were purified from the extract using cationic exchange and hydroxyapatite chromatography.
Human cardiac TnT (22) and TnC (19) were expressed and purified as described previously. The pET11c construct encoding human cardiac TnC was obtained as a gift from Prof. I. Trayer (Birmingham, UK).
Reconstitution of Troponin Complexes-Recombinant human cardiac troponin was reconstituted by mixing a 1.5:1:1 molar ratio of recombinant troponins C, I, and T. The protein was diluted to 0.75 mg/ml using buffer containing 10 mM imidazole, pH 7.0, 1 mM DTT, 0.01% sodium azide, 50 M CaCl 2 , 6 M urea, and 1 M KCl. The concentrations of urea and KCl were decreased using a stepwise dialysis protocol and the proteins were finally dialyzed into 5 mM PIPES, pH 7.0, 3.87 mM MgCl 2 , 1 mM DTT ("ATPase buffer"). Aggregated protein was removed by centrifugation at 12,300 ϫ g for 10 min and the troponin was concentrated using Centricon 10 concentrators. The composition of the reconstituted troponin complex was determined by SDS-PAGE, followed by Coomassie Blue staining and scanning densitometry (22). Wild type complex was found to be close to 1:1:1 with respect to the three subunits (TnT:TnI:TnC was 1:0.95:1.12 (n ϭ 3)) and the mutant complexes were shown to be indistinguishable from wild type. Although not routinely used, gel filtration chromatography was used to show that this reconstitution method using either wild type or mutant TnI resulted in a single species (the 1:1:1 complex).
Actin-TM-activated Myosin S-1 ATPase Assay-Actin-TM-activated myosin S-1 ATPase rates were measured using 0.5 M rabbit skeletal myosin S-1, 7 M rabbit skeletal actin, and 1 M rabbit ␣-TM in ATPase buffer containing 3 mM ATP at 37°C in the presence of 0 -4 M TnI or 0 -2 M troponin complex. For the titration of troponin under activating or relaxing conditions, assays were carried out at pCa 4.5 or 8.5. For the study of Ca 2ϩ dependence of ATPase, assays were carried out in the presence of 2 M troponin at a range of pCa values between 4.5 and 8.5. The free Ca 2ϩ concentration was set using 1 mM EGTA and the appropriate concentration of CaCl 2 as determined using the WINMAXC program (27). The reaction was stopped by the addition of 5% trichloroacetic acid and the amount of phosphate released measured by the method of Taussky and Schorr (28). Each assay was carried out in triplicate and on at least three separate preparations of TnI.
Actin-TM Sedimentation Assays-7 M rabbit skeletal actin and 1 M rabbit ␣-TM were mixed in ATPase buffer in the presence of 0 -2 M TnI or 0 -2 M troponin complex and incubated at 25°C for 15 min. The mixture was centrifuged at 313,000 ϫ g for 10 min at 25°C (Beckman TLA-100 rotor) to pellet the actin-TM filaments. Samples of the mixture prior to centrifugation and the supernatant and pellet were analyzed by SDS-PAGE using a 15% polyacrylamide gel; protein was stained using Coomassie Blue. Under these conditions Ͼ95% of the actin was found in the pellet and Ͻ10% of the TnI or troponin complex sedimented alone.
Surface Plasmon Resonance-based Assay of TnI-TnC Binding-TnI-TnC binding was investigated using a BIAcore-X in an assay based on the method of Reiffert et al. (29). TnC was labeled with biotin by incubation for 2 h at 4°C with a 3-fold molar excess of 6-(biotinamidocaproylamido)-caproic acid N-hydroxy-succinimide ester (Sigma) in 20 mM NaHCO 3 . Excess uncomplexed biotin was removed by extensive dialysis against 20 mM MOPS, pH 7.0, 0.3 M KCl, and 1 mM DTT. Biotin incorporation was determined using the protocol supplied with HABA/ Avidin reagent (Sigma). Biotinylated TnC was coupled to one flow cell of a BIAcore SA chip as previously detailed and sensorgrams showing the specific interaction of TnI with TnC were generated by subtraction of a second flow cell response (lacking captured TnC) from this test cell response. 100-s injections of 25-500 nM TnI in 20 mM MOPS, pH 8.0, 0.5 M KCl, 5 mM MgCl 2 , and 2 mM DTT ("Binding buffer"), with either 0.1 mM CaCl 2 or 1 mM EGTA, were made at a flow rate of 30 l/min. The dissociation phase was allowed to proceed for a further 100 s. The chip surface was regenerated with 60-s injections of 20 mM MOPS, pH 8.0, 6 M urea, 1 M KCl, 10 mM EDTA, and 2 mM EGTA and 2 mM DTT. Using the BIAEvaluation software, portions of the dissociation and association phases equivalent to those used by Reiffert et al. (29) were analyzed to derive estimates of the dissociation rate constant (k d ) and the association rate constant (k a ) assuming simple 1:1 kinetics. The association equilibrium constant (K a ) was then calculated (K a ϭ k a /k d ).
Statistical Analysis-Data sets were analyzed to give mean Ϯ S.E. Statistical analysis of the data was performed using the GraphPad InStat software. The Newman-Keuls post-test was used to compare pairs of group means. Values were defined as statistically significantly different with p values Ͻ0.05.

Effect of R145G and R162W Mutations on TnI Inhibition of
Actin-TM-activated Myosin ATPase Activity-The inhibition of actin-TM-activated myosin ATPase conferred by wild type and mutant TnI was compared. Addition of wild type TnI to myosinactin-␣-TM caused a decrease of 70% in the ATPase rate at 37°C with saturation of inhibition at 2 M TnI and an EC 50 of inhibition of 0.8 M. Both R145G and R162W TnI mutants were less effective inhibitors, with EC 50 of inhibition of 1.4 and 1.5 M, respectively, although the saturating levels of inhibition for both were similar to wild type (Fig. 1A). We analyzed whether these differences in inhibition were due to reduced binding of the mutant TnI to actin-TM. In sedimentation assays performed under ATPase conditions we found that wild type TnI and both mutants bound tightly to actin-TM; Fig. 1B shows that for wild type, R145G and R162W, each at 2 M, Ͼ95% of the TnI was present in the pellet. This result shows that the large differences in the potency of inhibition conferred by wild type and mutant TnI in this assay are not caused by weaker binding of the TnI to actin-TM.
Effect of R145G and R162W Mutations in TnI on the Function of the Troponin Complex-Troponin complexes were reconstituted as described using purified recombinant human cardiac isoforms of TnC, TnT, and either wild type or mutant TnI. The final subunit compositions of each reconstituted troponin determined by SDS-PAGE were identical. The Ca 2ϩ regulation of actin-TM-activated myosin ATPase conferred by wild type and mutant complexes was compared. In activating Ca 2ϩ conditions (pCa 4.5) addition of wild type complex to myosin-actin-␣-TM caused an increase of over 100% in the ATPase rate at 37°C; addition of troponin containing R145G or R162W mu-tant TnI was found to give indistinguishable activation ( Fig.  2A). At pCa 8.5, wild type troponin inhibited the actin-TMactivated myosin ATPase by 60%; however, under these conditions, it was found that the R162W mutant complex gave only 25% inhibition compared with wild type and the R145G mutant complex gave little or no inhibition ( Fig. 2A). As with TnI alone, sedimentation assays were used to show that the differences in inhibition were not caused by a difference in binding of the wild type and mutant complexes to actin-TM (Fig. 2IB). Further experiments using thin filament concentrations 2-fold lower (3.5 M) and 2-fold higher (14 M) confirmed that the percentage inhibition conferred by both wild type and mutant troponin at pCa 8.5 was independent of thin filament concentration.
Analysis of the Function of Troponin Mixtures Containing Both Wild Type and Mutant TnI-HCM is an autosomal dominant disease and affected individuals have one wild type copy and one mutant copy of a particular contractile protein gene. It is likely that in the hearts of those carrying a HCM cTnI mutation both wild type and mutant TnI proteins are expressed and incorporated into troponin complex. There are no reports of the determination of the in vivo proportion of cTnI protein that is mutant but it is unlikely to be more than 50% of the total. Recent analysis of the function of a cardiac TnT HCM mutant in the in vitro motility assay has suggested that the behavior of troponin mixtures containing 50% or less mutant complex cannot necessarily be predicted from the study of 100% mutant and 100% wild type preparations (22). Thus we have investigated the effect of troponin complex containing different wild type:mutant ratios on actin-TM-activated myosin ATPase at pCa 8.5 (Fig. 3). The response to increasing the proportion of R145G TnI troponin containing from 0 to 100% was clearly not linear: the transition from 0 to 50% mutant was accompanied by an increase of the rate from 38 to 51% of actin-TM-activated ATPase whereas the ATPase rose by almost 4 times as much (from 51 to 102%) as the proportion of mutant rose from 50 to 100%. The data obtained using mixtures of wild type and R162W TnI containing complexes showed a less clear biphasic relationship between ATPase and proportion of mutant troponin and could also be adequately fitted to a straight line.
Experiments were also carried out using 1 M total troponin to ensure that the nonlinear relationship was not due to the displacement of mutant troponin complex by wild type troponin complex. Under these conditions Ͼ95% of the troponin was bound to actin-TM (Fig. 2B). The biphasic relationship in the wild type/R145G troponin titration was still clearly seen (data not shown).

Effect of R145G and R162W Mutations on the Ca 2ϩ Dependence of Troponin Regulation of Actin-TM-activated Myosin
ATPase-The Ca 2ϩ regulation of actin-TM-activated myosin ATPase by wild type troponin, mutant troponin, and mixtures of 50% wild type, 50% mutant troponin were compared using a constant concentration of troponin in buffers of pCa from 4.5 to 8.5. Indistinguishable ⌬pCa 50 values were obtained using ei-  (Fig. 4) total troponin. The Ca 2ϩ regulation conferred by wild type troponin was shown to have a pCa 50 of 6.26 Ϯ 0.03. Troponin containing R145G TnI was found to have greatly increased Ca 2ϩ sensitivity of regulation (⌬pCa 50 ϭ ϩ0.56 Ϯ 0.03; p Ͻ 0.001); the 50:50 mixture of wild type and R145G troponin also showed an increased calcium sensitivity with a ⌬pCa 50 ϭ ϩ0.17 Ϯ 0.03 (p Ͻ 0.001).
An increased calcium sensitivity of a smaller magnitude (⌬pCa 50 ϭ ϩ0.13 Ϯ 0.03; p Ͻ 0.05) was shown by troponin complex containing R162W TnI. Again, the 50:50 mixture of wild type and mutant showed a Ca 2ϩ sensitivity lying between that of 100% wild type and 100% mutant (⌬pCa 50 ϭ ϩ0.08 Ϯ 0.04), although statistical analysis showed that this increased sensitivity was not significantly different from either 100% wild type or 100% R162W (p Ͼ 0.05).
Effect of R145G and R162W Mutations on the Interaction of TnI with TnC-The interaction of TnI with immobilized TnC was analyzed in real time using a surface plasmon resonance based assay (29). Purified recombinant TnC was biotinylated and bound to a BIAcore SA chip. The binding of wild type TnI and each mutant was measured over a range of concentrations in buffer containing either 0.1 mM CaCl 2 or 1 mM EGTA. A set of binding curves for wild TnI in 0.1 mM CaCl 2 is shown in Fig.   5A. By analysis of the association and dissociation phases of the binding curves, estimates of the association (k a ) and dissociation (k d ) rate constants and the association equilibrium constant (K a ) were obtained. The K a values obtained in the presence and absence of Ca 2ϩ using wild type TnI (1.61 ϫ 10 8 M Ϫ1 and 7.34 ϫ 10 7 M Ϫ1 , respectively) were very similar to those obtained in a solution binding assay using fluorescently labeled TnC (1.27 ϫ 10 8 M Ϫ1 and 4.17 ϫ 10 7 M Ϫ1 , respectively) (30). The binding of R145G TnI to TnC under both conditions was shown to be indistinguishable from that of wild type (Fig. 5B and Table I). However, the binding of R162W to TnC in 0.1 mM CaCl 2 was found to be significantly stronger (K a increased 2.6 times) due to a markedly slower dissociation rate although the k a was found to be the same as wild type (Fig. 5C and Table I). A small, but not significant, difference in binding of R162W TnI to TnC in 1 mM EGTA was found (K a increased 1.3 times) ( Table I). DISCUSSION This study of the in vitro properties of two human cTnI mutants that cause HCM has yielded information on the structure-function of TnI while also revealing functional differences between mutant and wild type protein that may underlie the pathogenesis of the disease.
Both the R145G and R162W mutants were shown to be less potent inhibitors of actin-TM-activated myosin ATPase than wild type TnI, with the EC 50 of inhibition of both mutants approximately double that of wild type TnI (Fig. 1A). This was not a consequence of weaker binding of the mutant TnI and hence is a primary effect of both mutations on TnI function. The equivalent of the R145G mutation in a 12-mer skeletal muscle TnI inhibitory peptide caused a much more profound decrease in the potency of inhibition (31). However, two single amino acid substitutions within the inhibitory region of fulllength rabbit skeletal muscle TnI were both found to give similar reductions in the potency of TnI inhibition to that reported here for the two HCM mutants (32). The discrepancy between the results obtained with peptide and full-length protein probably reflects the influence of residues outside of this region in stabilizing the binding of the minimal inhibitory sequence. Decreased inhibitory properties of a full-length TnI with a single amino acid substitution outside of the inhibitory region have not been previously reported. However, in a peptide equivalent to residues 129 to 181 in the cardiac protein, the triple mutation of the equivalent of Lys 174 , Lys 177 , and Lys 178 to Ala produced a similar reduced potency of inhibition as observed in this study with the R162W mutant (14).
Titration of actin-TM-activated myosin ATPase activity with whole troponin revealed activation at pCa 4.5 similar to wild type but significant differences in regulation between wild type and mutant troponin at pCa 8.5. R145G TnI gave 0% inhibition at saturation compared with 60% for wild type whereas the R162W mutant inhibited by 25% (Fig. 2). These data contrast with the inhibition obtained with TnI alone in that the differences between wild type and mutant are considerably more profound and they are present at the saturated levels of inhibition. The characteristics of complex containing R145G TnI (normal activation at pCa 4.5 and the lack of inhibition at pCa 8.5) are very similar to the properties we find of troponin containing a human cardiac TnT HCM mutant which lacks the C-terminal 28 amino acids (22).
The great majority of in vitro studies of the function of HCM mutant contractile proteins have compared the properties of 100% mutant with 100% wild type preparations. In those cases in which admixtures have been studied, however, the influence of mutant on wild type function has not been the straightforward result predicted from the pure wild type and mutant data (22,33). We have recently compared the function of wild type troponin with complex containing a truncated HCM mutant human cardiac TnT in the in vitro motility assay. It was found that under relaxing conditions the fraction of filaments motile parameter showed a biphasic transition from 100% wild type troponin (inhibited) to 100% mutant troponin (uninhibited) with troponin mixtures containing less than 50% mutant even causing further decrease of fraction filaments motile (22). In this present study the relationship between inhibition by troponin of actin-TM-activated myosin ATPase at pCa 8.5 and the proportion of R145G mutant TnI in the complex was found to be distinctly non-linear; at 50% mutant, troponin inhibited 51% compared with 70% predicted from the mean of the inhibition caused by pure wild type (38%) and pure mutant (102%) troponin. This implies that the propagation of inhibition brought about by wild type troponin at pCa 8.5 is sufficient to compensate to some extent for the poor inhibitory ability of troponin containing R145G TnI. This is in accordance with the HCM phenotype in that individuals do not normally present with symptoms until late puberty/early adulthood or later and thus the in vivo functional differences are expected to be quite subtle.
Using troponin containing 100% R145G or 100% R162W TnI the Ca 2ϩ sensitivity of actin-TM-activated myosin ATPase was found to be increased compared with wild type complex. The R145G mutation caused a large shift in pCa 50 compared with wild type (⌬pCa 50 ϭ ϩ0.56 Ϯ 0.03) and a significant shift was also seen with troponin containing the R162W mutant (⌬pCa 50 ϭ ϩ0.13 Ϯ 0.03). The shifts in pCa 50 in 50% mutant troponin admixtures were intermediate between the 100% wild type and 100% mutant values for both mutants. As found for the inhibition of actin-TM-activated myosin ATPase at pCa 8.5, 50% R145G mutant behaves more like pure wild type than pure mutant although still maintaining a significant functional difference. Significant increases in Ca 2ϩ sensitivity have previously been shown to be caused by a R105G mutation within the inhibitory region of whole skeletal TnI (equivalent to R138G in the cardiac protein) both in actin-TM-activated myosin ATPase regulation (32) and in reconstitution experiments in skinned cardiac fibers (34) (⌬pCa 50 approximately ϩ0.2 in both cases). Using the recombinant proteins described here, replacement of endogenous cTnI in skinned guinea pig cardiac fibers by R145G human cTnI gives a greater Ca 2ϩ sensitivity of force generation compared with replacement by wild type human cTnI. 2 Leftward shifts in the Ca 2ϩ sensitivity of regulation may be caused by a weaker interaction of the residues responsible for inhibition with actin-TM, a failure of the mutant troponin to fully shift the thin filament into an inhibitory conformation, or a stronger interaction with TnC within the complex. Altered thin filament interactions may occur with both R145G and R162W mutants whereas only the R162W mutant showed stronger binding to TnC (Fig. 5 and Table I; discussed below). Furthermore, the HCM mutations in TnI may cause a change in the 2 D. Burton and C. Ashley, unpublished data. Ca 2ϩ affinity of TnC directly leading to a shift in Ca 2ϩ sensitivity (35). It is likely that a full explanation of the mechanism of the Ca 2ϩ sensitivity shift will rely on a combination of these effects. HCM mutations in other contractile proteins, for example, myosin regulatory light chain (36), ␣-TM (37), and cardiac TnT (38 -40), have also been shown to increase the Ca 2ϩ sensitivity of regulation. In vivo, it would be predicted that these mutations would bring about increased force at submaximal Ca 2ϩ concentrations and lead to a hypercontractile state and abnormalities in relaxation (4). Interestingly, the expression in mice of a truncated TnI which is associated with myocardial stunning, led to a reduction in Ca 2ϩ sensitivity. Thus the aberrant protein linked with stunned myocardium may produce opposite functional anomalies compared with the HCM TnI mutants (17), as might be expected for the contrasting clinical features of these conditions.
The surface plasmon resonance-based studies of the binding of TnI and TnC revealed that the R145G mutation, within the inhibitory region of TnI, did not significantly perturb this interaction ( Fig. 5B and Table I). However, the R162W mutation, within the so called second TnC-binding region (14), was shown to reduce the dissociation rate constant (k d ) of the TnI-TnC interaction in the presence of Ca 2ϩ and hence increase the association equilibrium constant (K a ) ( Fig. 5C and Table I). The magnitude of change (K a of R162W is 2.6 times that of wild type) is similar to that observed between bisphosphorylated and dephosphorylated cTnI in a similar assay; however, the K a difference between the phosphorylated states is caused by a decreased k a (29). The binding of TnI and C is mediated via multiple interaction sites, with the N terminus of TnI binding to the C-terminal structural domain of TnC and the C terminus of TnI interacting with the N-terminal regulatory domain of TnC (9,10). The regulatory domain of cardiac TnC, in contrast to the skeletal muscle isoform, binds only one Ca 2ϩ ion via site II (site I is inactive) and this is sufficient for activation of the thin filament. Studies on skeletal TnI have shown that the regulatory domain of TnC interacts in a Ca 2ϩ -dependent manner with the inhibitory region (equivalent to residues 137-148 in cTnI) and the second TnC-binding region (149 -164 in cTnI) (14). This latter region binds to a hydrophobic pocket in TnC which is revealed by the "closed to open" conformational change which TnC undergoes upon binding Ca 2ϩ . The extent of exposure of the hydrophobic pocket in cardiac TnC is considerably reduced compared with that of the skeletal muscle protein (41). However, recent NMR spectroscopy analysis has confirmed that a peptide composed of residues 148 to 164 of human cTnI binds to a hydrophobic patch on the cardiac isoform of TnC (42) and indeed it has further been shown that, unlike the case with skeletal TnC, the binding of cTnI is necessary for the closed to open transition to occur (43). The R162W HCM mutation in cTnI causes a basic to hydrophobic charge change and the observed increased binding may be explained by an energetically more favorable interaction of the residue 162 side chain with the hydrophobic patch on TnC. The observation that the significantly increased binding is only seen in the presence of Ca 2ϩ supports this notion.
In summary, the two HCM TnI mutations both result in reduced inhibition of actin-TM-activated myosin ATPase by the troponin complex under relaxing conditions and increased Ca 2ϩ sensitivity of actin-TM-activated myosin ATPase regulation. If these functional differences manifest themselves in vivo there will be impairment of relaxation of cardiac muscle and this altered contractility may provide a hypertrophic stimulus leading to the disease state. The functional changes brought about by the HCM TnI mutants are qualitatively opposite to those brought about by truncated cTnI known to occur in stunned myocardium (17).