Functional Analysis of a Troponin I (R145G) Mutation Associated with Familial Hypertrophic Cardiomyopathy*

Familial hypertrophic cardiomyopathy has been associated with several mutations in the gene encoding human cardiac troponin I (HCTnI). A missense mutation in the inhibitory region of TnI replaces an arginine residue at position 145 with a glycine and cosegregates with the disease. Results from several assays indicate that the inhibitory function of HCTnIR145G is significantly reduced. When HCTnIR145G was incorporated into whole troponin, TnR145G(HCTnT·HCTnIR145G·HCTnC), only partial inhibition of the actin-tropomyosin-myosin ATPase activity was observed in the absence of Ca2+ compared with wild type Tn (HCTnT·HCTnI·HCTnC). Maximal activation of actin-tropomyosin-myosin ATPase in the presence of Ca2+ was also decreased in TnR145G when compared with Tn. Using skinned cardiac muscle fibers, we determined that in comparison with the wild type complex 1) the complex containing HCTnIR145G only inhibited 84% of Ca2+-unregulated force, 2) the recovery of Ca2+-activated force was decreased, and 3) there was a significant increase in the Ca2+ sensitivity of force development. Computer modeling of troponin C and I variables predicts that the primary defect in TnI caused by these mutations would lead to diastolic dysfunction. These results suggest that severe diastolic dysfunction and somewhat decreased contractility would be prominent clinical features and that hypertrophy could arise as a compensatory mechanism.

Familial hypertrophic cardiomyopathy (FHC) 1 has been linked to mutations in genes of nine different sarcomeric proteins. These mutations have been found in the genes for ␣-myosin heavy chain (1), cardiac myosin essential light chain and cardiac myosin regulatory light chain (2), ␣-tropomyosin, cardiac troponin T (TnT) (3), cardiac myosin-binding protein C (4,5), troponin I (TnI) (6), ␣-actin (7), as well as titin (8), and possibly troponin C (TnC) (9). This disease has recently gained significant attention due to several highly publicized reports of sudden death and fainting spells in young athletes who were asymptomatic and otherwise healthy individuals. In general, patients with FHC demonstrate an increase in heart muscle mass and sometimes an irregular echocardiogram (10). Kimura et al. (6) reported five missense mutations in TnI, R145G, R145Q, R162W, G203S, and K206Q, that cosegregate with FHC (Fig. 1). Three other TnI FHC mutants (S199N, Lys-183 deletion, and an exon 8 deletion mutant encompassing the stop codon of the cardiac TnI gene) have recently been discovered (11,12). Functionally TnI is the inhibitory subunit of the troponin (Tn) complex that controls the interaction between actin and myosin in a Ca 2ϩ -dependent manner (13)(14)(15). Studies using proteolytic fragments of fast skeletal TnI identified the central TnI sequence (residues 96 -116) as being responsible for its inhibitory activity. Residues 104 -115 of fast skeletal TnI (comparable to residues 136 -147 in cardiac TnI) formed the minimum sequence necessary for inhibition of muscle contraction (16 -19). Two of these mutations occur within the inhibitory region of TnI at a highly conserved amino acid residue (R145G and R145Q).
The HCTnI R145G mutation has been investigated previously by Elliot et al. (20) and shown to reduce the inhibition of the actin-Tm-activated myosin ATPase. Additional biochemical studies of the HCTnI R145G mutation and other FHC mutations are necessary to understand the mechanism by which this mutation and other hypertrophic cardiomyopathy mutations cause cardiac hypertrophy and sudden death. Currently the mechanism by which this mutation and other TnI FHC mutations cause cardiac hypertrophy and sudden death is uncertain but may involve primarily defects in relaxation (21,22).
In the present investigation we have determined how this arginine to glycine change within the inhibitory region of TnI alters the inhibitory function and how this mutation may cause deviations in the normal mechanisms of contraction as an explanation for its association with FHC. This article is the first report of the R145G mutation impairing force development, reducing maximal force, and reducing muscle relaxation. This is the first report of the R145G mutation being used in skinned cardiac fibers and to show that R145G impairs force development and relaxation. Mathematical simulations of intact cardiac fibers, in which the Ca 2ϩ affinity of troponin C and the effectiveness of troponin I as an inhibitor were altered, predict a lower contractility and an increased resting tension in HCTnI R145G myocardium compared with the normal myocardium. These intrinsic contractile changes will likely result in diastolic dysfunction in vivo.

Mutation, Expression, and Purification of HCTnI and HCTnI R145G -
The HCTnI mutants (R145G and A86T,R145G) were formed by overlapping PCR using HCTnI cDNA previously cloned in our laboratory from human cardiac tissue (23). The sequence of the TnI mutants was verified by sequencing prior to expression and purification.
HCTnI, HCTnI R145G , and HCTnI A86T,R145G were purified via conventional methods. Briefly, crude bacterial supernatants were purified by column chromatography on an S-Sepharose column at 4°C and eluted with a linear KCl gradient of 0 -0.5 M in a Tris-HCl buffer containing 6 M urea. Semipure HCTnI and HCTnI mutants were dialyzed against a solution containing 50 mM Tris-HCl, pH 7.5, 1 M KCl, 1 M urea, 1 mM dithiothreitol, and 2 mM CaCl 2 and loaded onto an affinity column having covalently bound HCTnC. Pure HCTnI and HCTnI mutants were eluted with a gradient of 0 -3 mM EDTA and 1-6 M urea. The purity of the TnI proteins was determined by SDS-PAGE (Fig. 1B).
Formation of the Troponin Complex-Formation of the human cardiac troponin complexes containing recombinant TnT, TnC, and TnI was carried out as recently described by Szczesna et al. (24). Proper stoichiometry was verified by SDS-PAGE. Although we do not routinely analyze reconstituted Tns formed by our method by chromatography, gel filtration of these Tn complexes showed that this reconstitution method resulted in a single species.
Actin-Tm-Activated Myosin ATPase Assay-Porcine cardiac myosin, rabbit skeletal F-actin, porcine cardiac tropomyosin, and recombinant human cardiac TnC were prepared as described previously (25). The ATPase inhibitory assay was performed in a 1-ml reaction mixture of 100 mM KCl, 4 mM MgCl 2 , 1.0 mM EGTA, 2.5 mM ATP, 0.1 mM dithiothreitol, 10 mM MOPS, pH 7. The ATPase activation assay was carried out in the same 1-ml reaction mixture with 1 mM EDTA replaced with 0.5 M CaCl 2 . F-actin (3.5 M), myosin (0.6 M), and Tm (1 M) were homogenized (24) and added to the reaction tube after the addition of buffer and either wild type human cardiac Tn (WTHCTn), human cardiac Tn containing TnI R145G (HCTn R145G ), or human cardiac Tn containing TnI A86T,R145G (HCTn A86T,R145G ) to the assay tube. The ATPase reaction was initiated with the addition of ATP and stopped after 20 min with 50% trichloroacetic acid. After sedimenting the precipitate, the inorganic phosphate concentration in the supernatant was determined according to the method of Fiske and Subbarow (26). The ATPase rates, measured by a single time point, were predetermined to be linear with time.
Preparation of Porcine Skinned Cardiac Muscle Bundles-Cardiac skinned muscle fibers were prepared following a common laboratory procedure published by Zhang et al. (23). Freshly isolated porcine hearts were incubated in an O 2 -saturated solution containing 140 mM NaCl, 4 mM KCl, 1.8 mM CaCl 2 , 1 mM MgCl 2 , 1.8 mM NaHPO 4 , 5.5 mM glucose, and 5 mM HEPES, pH 7.4. The hearts were obtained within 1 h of slaughter from a local slaughterhouse near the University of Miami. Cardiac muscle bundles were dissected from the left ventricle of the porcine hearts and were chemically skinned by incubating with 50% glycerol and 1% Triton X-100 in the relaxing solution (pCa 8.0) containing 10 Ϫ8 M Ca 2ϩ , 5 mM Mg 2ϩ , 7 mM EGTA, 20 mM MgATP, 20 mM creatine phosphate, and 15 units/ml creatine phosphokinase, pH 7.0, at an ionic strength of 150 mM at 4°C for 24 h. These skinned muscle preparations were dissected into small bundles (1-2 cm in length, 2-3 mm in diameter) and were stored at Ϫ20°C in the same solution without Triton X-100 before use.

Displacement of the Endogenous Tn Complex in Porcine Skinned Cardiac Muscle Preparations with HCTnT and Reconstitution with the Tn Complex Containing HCTnI R145G : Steady-state Force and the Ca 2ϩ
Sensitivity of Force Development-The skinned fiber preparation was mounted with stainless steel clips on a force transducer and was immersed in the contracting solution to measure initial force before treatment. The contraction solution (pCa 4) had the same composition as the relaxation buffer except for the increased Ca 2ϩ concentration (10 Ϫ4 M). To determine the Ca 2ϩ dependence of force development, the contraction of the skinned fibers was tested in solutions of intermediate concentrations of Ca 2ϩ . The Ca 2ϩ dependence of force was determined before and after performing the displacement and reconstitution protocols that are described below. To remove the endogenous Tn complex, the TnT displacement method was used (24,27). The cardiac fiber was incubated with HCTnT (0.8 mg/ml) for a total of 2.5 h with an intermediate buffer change containing fresh HCTnT. After displacement of the endogenous complex by HCTnT, the level of unregulated force development was observed by measuring the level of force reached by skinned fibers in both the pCa 8 solution and the pCa 4 solution. To determine the Ca 2ϩ dependence of force development, the contraction of skinned fibers was tested in solutions with intermediate concentrations of Ca 2ϩ (from pCa 8 to pCa 4). The Ca 2ϩ dependence was determined before and after treatment of the skinned fibers with displacement and reconstitution solutions. The Ca 2ϩ dependence data were fit to the Hill equation with SIGMAPLOT (Jandel Scientific): relative force (%) ϭ [Ca 2ϩ ] n /([Ca 2ϩ ] n ϩ [Ca 50 ] n ), where pK is the midpoint (pCa 50 ) and n is the Hill coefficient. Control experiments carried out on porcine fibers treated the same way as described above but without any protein showed that the average rundown for these fibers was ϳ12-18%. Taking this average rundown of the fibers into account the average recovered force for these skinned fiber studies is ϳ70% of the original force before treatment with TnT and reconstitution with TnI-TnC. The average force at the end of these experiments is ϳ60% of the initial force. All data are presented as mean Ϯ S.D.
Circular Dichroism (CD) of HCTnI and Its Mutants-CD spectra of the HCTnI and two mutants were recorded on a Jasco J-720 spectropolarimeter using a cell pathlength of 0.1 cm at ambient temperature (20°C) in a 10 mM sodium phosphate, pH 7.0, 500 mM NaCl solution. Spectra were recorded at 200 -250 nm with a bandwidth of 1 nm at a speed of 50 nm/min and a resolution of 0.2 nm. Analysis and processing of data was done using the Jasco system software (Windows Standard Analysis, version 1.20). Ten scans were averaged, base lines were subtracted, and no numerical smoothing was applied.
Mean residue ellipticity ([] MRE , in degrees⅐cm 2 /dmol) for the spectra were calculated (using the Jasco system software) using the following equation where [] is the measured ellipticity in millidegrees, Cr is the mean residue molar concentration, and l is the path length in cm. The ␣-helical content for each protein was calculated using the standard equation for [] at 222 nm (28): [] 222 ϭ Ϫ30,300f H Ϫ 2,340, where f H is the fraction of ␣-helical content (f H ϫ 100, expressed in %). Spectra are presented as the mean residue ellipticity.
Simulation of Twitch Force in Intact Fibers with HCTnI R145G -To assess the impact of changes incurred in force regulation in cardiac skinned fibers by HCTnI R145G, we used a quantitative computer model that integrates Ca 2ϩ binding to various buffers in cardiac cells to predict the amplitude and time course of the intracellular Ca 2ϩ transient, of Ca 2ϩ bound to each of various Ca 2ϩ buffers (TnC, calmodulin, and sarcoplasmic reticulum uptake), and of force. This method quantifies buffering of intracellular Ca 2ϩ , sarcoplasmic reticulum Ca 2ϩ release and uptake, and free [Ca 2ϩ ] i based on an initially assumed intracellular free Ca 2ϩ transient and published rate constants (29,30). The model is similar to that of Robertson et al. (31) and in addition incorporates the following processes: (a) a variation of the Ca 2ϩ affinity of TnC with developed force (k off ϭ k off.rest e Ϫgain ϫ relative force ) (33,34), (b) a modulation of TnI effectiveness as an inhibitor dependent primarily on the concentration of the Ca 2ϩ ⅐TnC complex (TnI effect ϭ [Tn] ϫ (1 Ϫ TiAmp ϫ e ϪTiPower ϫ [CaT] ), where TiAmp and TiPower are amplitude and gain variables of the effectiveness of TnI to inhibit actomyosin interactions, and (c) a two-state cross-bridge model (32). Extensive details of this quantitative computer model can be found in the methods and appendix sections of Miller et al. (35) along with the additional features mentioned above. The terms used in this computer model are briefly described below.
Force Dependence of Affinity of TnC for Ca 2ϩ -The rate of release of Ca 2ϩ from TnC is slowed by the presence of cross-bridges. k off(TnC⅐Ca) therefore becomes smaller as cross-bridges form and force develops (34,37,38). The exponential dependence of k off(TnC⅐Ca) on force is based on experimental observations in cardiac and skeletal muscle (32)(33)(34) and is quantified by the equation k off.TnC⅐Ca ϭ k off.TnC⅐Ca.rest ϫ e Ϫgain ϫ relative force , where k off.TnC⅐Ca.rest is the off-rate of Ca 2ϩ from TnC in the absence of force development. For the simulations used in this study, gain was set at 8.  (32), f is the attachment (on) rate constant of detached cross-bridges, g is the detachment (off) rate of attached cross-bridges, and TnI eff is the effective TnI concentration permitting actomyosin interactions consequent to Ca 2ϩ occupancy of TnC. We assumed constant values for f ϭ 400,000 mol⅐kg Ϫ1 ⅐s Ϫ1 and g ϭ 10 s Ϫ1 . Finally force is displayed as normalized to maximal force that could theoretically be obtained, i.e. as the ratio [CB]/[total CB].
Simulation Procedures-Simulations for reproducing experimental results of skinned fibers were carried out for equilibrium conditions in pCa values of 8 -4 in 0.1-pCa steps. Values of k off.TnC⅐Ca.rest , gain (TnC-Ca 2ϩ binding dependence on force), and TiAmp and TiPower (amplitude and gain of TnI effectiveness as an inhibitor) were varied until a unique combination of values produced pCa-force relations with identical pCa 50 and Hill n coefficients as the wild type. The same procedure was carried out to simulate pCa-force relations of the HCTnI R145F mutant for which a unique convergent solution was found. Cross-bridge variables were not changed.
Subsequently variables found from simulations in skinned fibers were fed into the simulation programs that generate twitch contractions of intact cardiac fibers. In the initial control conditions, once sarcoplasmic reticulum Ca 2ϩ release was obtained, steady-state control conditions were obtained by repeating the simulation over several successive twitch contractions by setting the initial values of [Ca 2ϩ ], [CaT], [CaC], and [CB] of a given twitch to their values obtained at the end of the previous twitch (where [CaC] is calmodulin occupied with Ca 2ϩ and [CaT] is troponin occupied with Ca 2ϩ ). Simulations carried out for a change in one or more rate constants or variables were also allowed to reach steady-state conditions over several twitches. In most instances, steady-state conditions were reached in 3-5 contractions at frequencies corresponding to heart rates of 150 min Ϫ1 .
We simulated force during a twitch that would occur if one or more of the following would be affected by the presence of HCTnI R145G on the thin filament: (a) a decrease in the resting (zero force) off-rate Ca 2ϩ from the Ca 2ϩ -specific site of TnC (k off(TnC⅐Ca) ) from a control value of 700 s Ϫ1 and/or (b) a decrease in the amplitude (TiAmp) and gain (TiPower) of troponin I to inhibit actomyosin interactions. We chose not to alter cross-bridge kinetic variables as there is no a priori reason to assume that the mutation would directly affect actomyosin kinetic behavior. Based on each possible effect of HCTnI R145G (see "Results") this theoretical analysis predicts muscle contraction and relaxation as would be encountered in vivo and provides for a series of hypotheses that could be tested in the hearts of HCTnI R145G transgenic animals. The calculations were programmed in Microsoft Quickbasic 4.0 and run on a personal computer, and the results (saved as ASCII files) were replotted with SigmaPlot 5.0 (SPSS, Inc., Richmond, CA).  (41)) and later HCTnI sequences that coded for Ala at residue 86 (42,43). Fig. 1B shows an SDS-polyacrylamide gel of purified wild type and mutant TnIs used in these studies.

Fig
Effect of the R145G Mutation on TnI Inhibition of Actin-Tm-Activated Myosin ATPase Activity-We examined the ability of HCTnI R145G and WTHCTnI to inhibit F-actin-Tm-activated myosin ATPase activity in reconstituted thin filament systems to determine whether the inhibitory activity of the HCTnI R145G mutant of TnI was affected by the missense mutation within the inhibitory region of the protein. We discovered that while the wild type HCTnI was able to inhibit ATPase activity nearly fully (approximately 90% at 3 M HCTnI concentration), HCTnI R145G was less effective at inhibiting actomyosin ATPase activity even at higher protein concentrations (Ͻ60% at 4 M HCTnI concentration, Fig. 2).
Regulation of Actin-Tm-Myosin ATPase Activity by Tn and Tn Containing HCTnI R145G -When Tn R145G was reconstituted with actin, Tm, and myosin in the presence of EGTA, the ability of the mutant complex to inhibit ATPase activity was significantly less (20 -25% lower at Tn concentrations between 1-2 M) than wild type Tn (Fig. 3A). In the presence of EGTA wild type Tn (2 M) inhibited the actin-TM-activated myosin ATPase activity by 82 Ϯ 4%. Under the same conditions used for the wild type Tn, HCTn R145G inhibited the actin-TM-activated myosin ATPase activity by 61 Ϯ 4%. The maximum ATPase activity for Tn R145G in the presence of Ca 2ϩ was also less (25% lower at Tn concentrations between 1-2 M) than wild type Tn (Fig. 3B). In these experiments the amount of Tn required for maximal ATPase activation was 1-1.5 M, which is consistent with a ratio of Tn:Tm of 1:1(which is considered to be the physiological ratio of these components in intact muscle fibers). The amount of Tn required for maximal ATPase inhibition (in the absence of Ca 2ϩ ) was 1.5-2 M. Force Development and the Ca 2ϩ Dependence of Force Development-We used a well established method in our laboratory (24) to displace the endogenous Tn complex from skinned porcine skinned cardiac muscle preparations. In these experiments, after displacing the endogenous porcine Tn complex with HCTnT (0.8 mg/ml) either wild type HCTnI⅐HCTnC, HCTnI R145G ⅐HCTnC, or HCTnI A86T,R145G ⅐HCTnC complexes were used to reconstitute the HCTnT-replaced skinned fibers. After determining the level of unregulated force in skinned fibers, they were incubated with either wild type or mutant HCTnI⅐HCTnC complexes in low calcium buffer (pCa 8). This allowed us to determine whether the TnI proteins could fully inhibit Ca 2ϩ -unregulated force established after treatment with HCTnT and also to determine whether the proteins were able to fully reconstitute the skinned fibers by forming a functional Tn complex. Wild type HCTnI⅐CTnC complex resulted in complete inhibition of Ca 2ϩ -unregulated force. However, when the skinned fibers were incubated with the HCTnI R145G ⅐HCTnC complex, only 84% inhibition of unregulated force was observed (Figs. 4 and 5B). Full inhibition of unregulated force was not achieved with the complex containing the mutant TnI even after an extended incubation time of 2.5 h (data not shown). The data depicted in the absence of Ca 2ϩ represents the level of force remaining after reconstituting the fibers with the appropriate TnI⅐TnC complex (Figs. 4 and 5B). This level of force is equivalent to the percentage of Ca 2ϩ -unregulated force that was not inhibited by the Tn complex. The fibers reconstituted with Tn R145G were able to inhibit only 84% of unregulated force compared with 100% inhibition by wild type Tn. Recovered force is equivalent to the level of force developed in fibers after reconstituting the fibers with the appropriate Tn complex and treating the fiber with pCa 4 solution. Recovered force is represented in Fig. 4 as a percentage of unregulated force in the presence of Ca 2ϩ . Fibers reconstituted with the mutant complex did not recover full force. The level of force recovered with the mutant complex (81.9 Ϯ 4.5%) was significantly different from force recovered by the fibers reconstituted with the wild type complex (97.4 Ϯ 3.2%). These results correlated with the ATPase results indicating that this mutant may somehow hinder strong myosin head binding to the actin filament in the skinned cardiac muscle system.
The Ca 2ϩ dependence of force development was measured before treating the fiber with HCTnT and after incubating the fiber with either the wild type HCTnI⅐HCTnC or HCTnI R145G ⅐HCTnC complexes to determine whether HCTnI R145G altered the Ca 2ϩ dependence of force development in skinned cardiac fibers. As seen in Fig. 5, a significant change in the Ca 2ϩ dependence of force was observed between the fibers reconstituted with the wild type HCTnI⅐HCTnC complex (pCa 50 ϭ 5.66, n ϭ 2.44) and those reconstituted with the HCTnI R145G ⅐HCTnC complex (pCa 50 ϭ 5.79, n ϭ 1.4). The inset in Fig. 5A shows the Ca 2ϩ dependence of force of porcine fibers

FIG. 3. Inhibition of actin-Tm-activated myosin ATPase activity by either WTHCTn or HCTn R145G in the absence of Ca 2؉ (A) and activation of actin-Tm-activated myosin ATPase activity by either WTHCTn or HCTn R145G in the presence of Ca 2؉ (B).
The assay included actin (3.5 M), myosin (0.6 M), and Tm (1 M) and was reconstituted with either HCTn (q) or HCTn R145G (OE) at the concentrations shown on the x axis. Each data point represents the average of three to four separate experiments each performed in duplicate and is expressed as mean Ϯ S.D. The maximal ATPase activity of actin-Tmactivated myosin ATPase in the presence of WTHCTn was 1.14 Ϯ 0.042 mol of P i /mol of myosin/s, while reconstituted thin filaments containing HCTn R145G had a maximal activity of 0.95 Ϯ 0.049 mol of P i /mol of myosin/s in the presence of Ca 2ϩ . The ATPase activity values for HCTn R145G are significantly different from WTHCTn both in the presence and absence of Ca 2ϩ (p Ͻ 0.05).

FIG. 4. Effect of HCTnI R145G on inhibition of unregulated force and recovered force in skinned cardiac fiber preparations.
Each fiber was treated with HCTnT (0.8 mg/ml) for 2 h at room temperature and subsequently reconstituted with either the WTHCTnI⅐HCTnC complex or the HCTnI R145G ⅐HCTnC complex. The black bar represents the percentage of unregulated force remaining after reconstituting the skinned fibers with either the WTHCTnI⅐HCTnC or HCTnI R145G ⅐HCTnC complex. The % unregulated force remaining ϭ 100% unregulated force Ϫ % inhibition of unregulated force. Inhibition of unregulated force is 102.0 Ϯ 2.4% (mean Ϯ S.D., n ϭ 4) for the wild type complex and 84 Ϯ 3.3% (n ϭ 4) for HCTnI R145G . The gray bar represents the percentage of force recovered after reconstituting the fiber with the WTHCTnI⅐HCTnC complex (97.4 Ϯ 3.2%) or HCTnI R145G ⅐HCTnC complex (81.9 Ϯ 4.5%) and treating the fiber with pCa 4 buffer. Each bar represents four separate experiments. *, p Ͻ 0.05; **, p Ͻ 0.01. before displacement with TnT and after reconstitution of the wild type TnI⅐TnC complex. As noted in a previous article (24) on the Ca 2ϩ dependence of force, porcine fibers reconstituted with human CTn are less sensitive to Ca 2ϩ than the intact untreated fibers. Fig. 5B shows the force obtained after TnT treatment, and HCTnI⅐HCTnC reconstitution normalized to the maximal force in the fibers after treatment with wild type HCTnT. Fig. 5C shows a representative 15% SDS-PAGE of wild type Tn-and Tn R145G -displaced porcine cardiac muscle preparations. The TnT-displaced fibers (Fig. 5C, lane 3) lacked the endogenous TnI that was displaced during the treatment. These gels shows that the reconstitution of HCTnI⅐HCTnC and HCTnI R145G ⅐HCTnC were similar.
Circular Dichroism Spectra of HCTnI, HCTnI R145G , and HCTnI A86T,R145G -CD spectroscopy was carried out to determine the secondary structure of HCTnI, HCTnI R145G , and HCTnI A86T,R145G (Fig. 6). The average mean residue molar ellipticity ([]) for HCTnI at 222 nm was Ϫ10,587 degrees⅐cm 2 / dmol (corresponding to an ␣-helical content of 27.2%). HCTnI R145G had a [] at 222 nm of Ϫ11,421 degrees⅐cm 2 /dmol (corresponding to an ␣-helical content of 30%), while HCTnI A86T,R145G had a [] at 222 nm of Ϫ11,217 degrees⅐cm 2 / dmol (corresponding to an ␣-helical content of 29.3%). These results suggest that HCTnI R145G and HCTnI A86T,R145G may have a slightly higher ␣-helical content than HCTnI. However this difference is small, suggesting that the overall secondary structure of the different TnIs is very similar. This slightly higher ␣-helical content in HCTnI R145G may be functionally important for the observed differences between HCTnI R145G and wild type HCTnI. However, which region(s) is structurally affected by the R145G mutation is not known. Fig.  7A illustrates the strong dependence of the dissociation rate of Ca 2ϩ from TnC as a function of force, while Fig. 7B illustrates the relationship between [CaT] and effective [TnI]. Variables pertaining to the affinity of TnC for Ca 2ϩ and to the effectiveness of TnI to transduce the TnC Ca 2ϩ binding to relief of inhibition of actomyosin cross-bridge formation were changed stepwise in isolation and in combination until a unique set of variables was identified that reproduced the pCa-force relations of skinned fibers that contained WTHCTnI or HCTnI R145G , respectively (Fig. 8). Starting from a resting (noforce) k off.TnC⅐Ca.rest of 700 s Ϫ1 (WT), a TnI effectiveness of 1 (maximal inhibition in the absence of Ca 2ϩ ) (WT) and a TnI gain of 10,000 in the WT fibers, the HCTnI R145G mutation was reliably reproduced by the combination of 1) a decrease in the TnC Ca 2ϩ off-rate (to 420 s Ϫ1 ) and 2) a decrease in TnI effectiveness (TiAmp ϭ 0.93) and gain (TiPower ϭ 5,500). The WT and R145G values for pCa 50 (5.762 in WT and 5.665 in R145G) and Hill coefficient n (2.427 in WT and 1.584 in R145G) are very close to those observed experimentally (compare simulation results of Fig. 8 with experimental findings of Fig. 5). Furthermore, Ca 2ϩ -independent force of the R145G mutant was 20.0% of the WT peak force, and peak force of the R145G developed by each fiber after TnT treatment and HCTnI⅐HCTnC reconstitution. The inset shows the force development in porcine skinned fibers before Tn displacement and after displacement with TnT and reconstitution with HCTnI⅐HCTnC. B, forces obtained after TnT treatment and HCTnI⅐HCTnC reconstitution were normalized to the maximal force in the fibers after treatment with wild type TnT.   5. The effect of the HCTnI R145G mutation on the Ca 2؉ sensitivity of force development. Each skinned muscle preparation was treated with HCTnT (0.8 mg/ml) to displace the endogenous Tn complex. The level of Ca 2ϩ -unregulated force following the TnT treatment was equal to Ϸ60% of the force developed by the untreated skinned fiber preparations. For clarity this level of Ca 2ϩ -unregulated force was set to 100% to determine the percentage of force activation and inhibition following the treatment. After displacement of endogenous Tn, the preparation was reconstituted with either wild type HCTnI⅐HCTnC (q, pCa 50 ϭ 5.66, n ϭ 2.44) or HCTnI R145G ⅐HCTnC (OE, pCa 50 ϭ 5.79, n ϭ 1.4) complexes (25 M for 1.5 h). After switching to the pCa 4 solution, the fibers developed force that was expressed as a percentage of the maximal force (100%) measured after TnT treatment. For more details see Szczesna et al. (24). The Ca 2ϩ dependence of force was measured in each preparation after reconstituting whole HCTn. Each point is the average of three to four experiments and represents the mean Ϯ S.D. A, forces were normalized to the maximum force mutant was 80.4% of the WT peak force, both in accordance with experimental observations in these fibers (Fig. 5).

Simulation of pCa-Force Relations of Skinned Fibers-
Simulation of Twitch Force in Intact Fibers with HCTnI R145G -In an effort to predict the behavior of intact ventricular myocardium afflicted with the HCTnI R145G mutation, we inserted the values of TnC Ca 2ϩ affinity and of TnI amplitude (TiAmp) and gain (TiPower) in the simulation programs that predict the time course of force in a continuous series of twitch contractions in steady-state conditions. Fig. 9A show that the changes in TnC Ca 2ϩ affinity and in TnI inhibitory effectiveness as found in skinned fibers cause a substantial increase in resting force in R145G myocardium. When twitch force was normalized to that of the WT, it appears that the R145G mutation decreases peak developed force (Fig. 9B), increases time to peak force, and causes a slight delay in isometric relaxation, all illustrated in Fig. 9B. If heart rate is not allowed to change, the residual force at the end of one twitch is present at the beginning of the next. Since HCTnI R145G is a poor inhibitor of contractile activity compared with the wild type, resting force or diastolic tension is elevated. These mathematical simulations of intact cardiac fibers predict a decreased contractility and an increased resting tension in HCTnI R145G myocardium compared with the normal myocardium.

DISCUSSION
In this study, we have investigated the effects of the arginine to glycine missense mutation in the inhibitory region of HCTnI on its inhibitory properties. This current work supports the theory that impaired contractility and diastolic dysfunction, a consequence of cardiac hypertrophy, may be due to a regulatory  . Developed force is less in the R145G TnI mutant than in the WT, and peak force is attained later. Isometric relaxation in the twitch with mutant R145G TnI is slightly delayed. AU, arbitrary units. perturbation within the interactions of the Tn subunits. The HCTnI R145G mutation occurs at a highly conserved position across many species and causes a change from a basic arginine residue to a nonpolar glycine residue at the 145 position (Fig.  1). Also, results from our laboratory using mutants of skeletal TnI suggest that residues 105-115 of the inhibitory region (comparable to residues 137-147 of HCTnI) are much more critical for the biological activity of TnI than residues 96 -106 (39). From this standpoint, a significant amino acid change in this region could definitely affect mechanisms of cardiac muscle contraction.
Our approach to determine how R145G affected the function of the thin filament was to observe the mutant in both reconstituted thin filaments as well as in skinned cardiac muscle systems. We observed that HCTnI R145G alone in a reconstituted system inhibited actin-Tm-myosin ATPase activity to a lesser extent than WTHCTnI (Fig. 2). This lends to the idea that the interaction between actin-Tm and HCTnI R145G is disrupted by the missense mutation in HCTnI.
Complexation of HCTnI R145G with HCTnC and HCTnT in the reconstituted thin filament system resulted in reduced inhibitory function of Tn R145G (Fig. 3). This reduced inhibition could occur if the mutant HCTnI only partially holds Tm in the closed position and partially blocks the binding of strong myosin heads to the thin filament; in this way, the ATPase activity remains elevated at low [Ca 2ϩ ]. Interestingly this loss in inhibitory function in the absence of Ca 2ϩ was also observed in skinned fiber preparations in that HCTnI R145G was not able to fully inhibit Ca 2ϩ -unregulated force. Notably the maximum level of ATPase activity was also changed due to HCTnI R145G in the reconstituted thin filament. This result suggests that HCTnI R145G affects the ability of HCTnC to remove the inhibition of ATPase activity by HCTnI in the presence of Ca 2ϩ . We also observed a decrease in the maximal recovered force with HCTnI R145G (Fig. 4). These results suggest that in the skinned cardiac muscle system an alteration in the HCTnC-HCTnI R145G interaction exists and that the myosin crossbridge interactions in these fibers are altered in such a way that maximal force cannot be obtained. Thus, HCTnI R145G appears defective not only in its inhibitory ability but also confers some change on the regulatory system that impairs maximal ATPase and force activation. This could form altered TnI-TnC or TnI-TnT interactions, which result in other changes in the regulatory mechanism, e.g. altered TnT-Tm or actin interactions. In addition, a moderate shift in the Ca 2ϩ dependence of force development due to the mutation was observed. This could result from either a change in the TnC-TnI interaction that would lead to direct changes in the affinity of the single Ca 2ϩ -specific regulatory site on TnC or to an indirect effect via alterations in cross-bridge interactions, which would feed back and alter TnC Ca 2ϩ affinity. These results suggest that altered interactions exist between HCTnI R145G , actin, and Tm and/or between HCTnI R145G , TnC, and TnT, which affect the inhibitory and regulatory functions of HCTnI R145G . However, using a semiquantitative method, actin-tropomyosin sedimentation assays, we found no difference in the binding of HCTnI or HCTnI R145G to actin-tropomyosin over a wide TnI concentration range (data not shown) similar to what was found by Elliott et al. (20). Hence, the results obtained appear not to be due to a weaker binding of TnI to actin-tropomyosin. We also know that the affinity between the wild type and mutant TnI and TnC are not significantly changed. The new results presented in Fig. 2 suggest that the loss in inhibition by the mutant TnI is not due to alterations in its affinity for actin or to an effect of TnC or TnT on its physiological function.
It is possible that this mutation causes a structural pertur-bation in the HCTnI subunit and as a result in whole Tn. To determine whether this mutation caused a change in secondary structure of TnI, CD spectroscopy was carried out on WTTnI and the TnI mutants (Fig. 6). The TnI mutants (R145G and A86T,R145G) showed a small difference in secondary structure when compared with HCTnI. The average ␣-helical content of wild type HCTnI was 27.2%, while the ␣-helical content of HCTnI R145G and HCTnI A86T,R145G were 30 and 29.3%, respectively. Using reconstituted filaments, Elliot et al. (20) found that R145G reduced the inhibition of actin-Tm-activated myosin ATPase activity and increased the Ca 2ϩ sensitivity of ATPase regulation. However, they found no difference in the activation properties of Tn containing wild type TnI and Tn containing the TnI R145G mutation. We found that the maximal ATPase level in Tn containing the TnI R145G mutation was significantly decreased when compared with wild type Tn, consistent with the lower level of force seen in the reconstituted fibers (Fig. 4). In agreement with our present results, Takahashi-Yanaga et al. (40) using a porcine cardiac myofibrillar preparation also found that the maximal level of ATPase activity with the mutant TnI R145G was lower than that for wild type TnI. These results suggest that the other components of the reconstituted filaments besides Tn are important for ATPase activity variations. Elliot et al. (20) used rabbit skeletal actin, rabbit cardiac Tm, and rabbit skeletal myosin S1, while we used rabbit skeletal actin, porcine cardiac Tm, and porcine cardiac myosin.
Since the human cardiac TnI (which was obtained from Dr. P. Barton (Imperial College London)) used by Elliot et al. (20) might be different from the TnI used in this study and by Takahashi-Yanaga et al. (40) we decided to determine whether a possible difference between these two TnIs (a single amino acid change, A86T) could account for the difference in the ATPase activatory properties of TnI R145G . When Barton (41) first cloned HCTnI using PCR the translated protein contained Thr at residue number 86 (EMBL accession no. X54163). However, other researchers as well as Barton's group later obtained HCTnI cDNA that coded for Ala at residue number 86 (42,43). This HCTnI containing Ala at residue number 86 is the sequence in Swiss-Prot (accession no. P19429). A TnI mutant containing A86T and R145G (TnI A86T,R145G ) was expressed, purified, reconstituted with HCTnT and HCTnC and used for ATPase assays. No significant difference between the HCTnI R145G and HCTnI A86T,R145G mutants were observed with respect to ATPase activity in the presence or absence of Ca 2ϩ .
To determine the possible impact of the HCTnI R145G mutation on contraction and relaxation of intact ventricular myocardium, we used a simple mathematical model (35) of intracellular Ca 2ϩ buffering and of force generation based on Huxley's two-state cross-bridge model to predict the time course and amplitude of free [Ca 2ϩ ], force, and various Ca 2ϩ buffers that would be observed if native TnI were replaced with mutant HCTnI R145G . In view of our findings that the loss of inhibition of HCTnI R145G is a direct effect, a mathematical step describing TnI inhibitory effectiveness was inserted between the formation of the Ca 2ϩ ⅐TnC complex and actin-myosin cross-bridge formation. Specifically the term TiAmp describes the inhibitory function of TnI. The value of this factor is 1 by default for WT and was decreased to 0.93 for HCTnI R145G . The Ca 2ϩ -dependent function of TnI was mathematically approximated by a "gain" factor amplifying the effect of Ca 2ϩ bound to TnC. Its default value was set at 10,000. Since the HCTnI R145G mutant also confers some change in the regulatory system that impairs maximal ATPase and force, the Ca 2ϩ -dependent gain (TiPower) was found to decrease to 5,500. The increased Ca 2ϩ sensitivity of force conferred by the HCTnI R145G mutation in skinned cardiac fibers was simulated as a decrease in the off-rate of Ca 2ϩ from the Ca 2ϩ -specific site II of cardiac TnC. Results in skinned fibers, both WT and HCTnI R145G , were reproduced by simulation that changed only the apparent affinity of TnC for Ca 2ϩ and the TnI-specific inhibitory amplitude (decrease by 7%) and Ca 2ϩ -dependent gain (decrease by 45%). No direct changes in cross-bridge kinetics are postulated, although they may exist as secondary effects.
The changes in three troponin variables were introduced in the simulations for steady-state twitch contractions of cardiac muscle. The twitch contraction (in cardiac muscle) operates from a pCa range of ϳ7.67 (at rest) to 5.66, and in this range steady-state force in skinned fibers containing HCTnI R145G is higher than force of skinned fibers with WTHCTnI. However, equilibrium conditions are not achieved during a twitch, and the pCa-force relationship is much steeper in intact than in skinned fibers (44) most likely because of the loss of natively present constituents and of surface membrane regulation in the latter. Therefore, one cannot with confidence extrapolate equilibrium conditions from skinned fibers to dynamic conditions in intact fibers, and simulations specific for intact fibers need to be carried out. Simulations of twitch contractions in intact fibers demonstrated that the HCTnI R145G mutation increased diastolic tone and decreased contractility with minor effects on isometric relaxation. The most prominent finding was the vastly increased diastolic tone that translates clinically into severe diastolic dysfunction. In summary, these studies suggest that the R145G change within the TnI subunit alters interactions with the Tn complex that affects the Ca 2ϩ dependence of cross-bridge formation in the muscle system.
Elliot et al. (20) using surface plasmon resonance-based assays found that the TnI R145G mutant had a similar affinity for TnC as wild type TnI in the presence of Ca 2ϩ . We also carried out surface plasmon resonance-based assays and found similar results (data not shown). Since an increase in the Ca 2ϩ sensitivity of force must be due to some change in the affinity of HCTnC for Ca 2ϩ these results suggest that the TnI mutation is exerting its effect on another thin filament protein (such as HCTnT or actin), which then affects TnC.
James et al. (45) recently reported a transgenic model of the CTnI R145G FHC mutation. The transgenic mice that were generated contained either mouse CTnI or a mouse CTnI containing the R146G (corresponding to R145G in the human sequence) mutation. Transgenic mice containing the CTnI R145G mutation showed cardiomyocyte disarray and interstitial fibrosis and suffered premature death. Skinned fiber studies on these transgenic mice fibers containing the TnI FHC mutation showed increased sensitivity to Ca 2ϩ , consistent with what is presented in this article. These transgenic mice fibers also showed a decreased maximal tension when compared with wild type and nontransgenic mice similar to what is seen in our skinned porcine fiber studies. Transgenic mice containing the TnI FHC mutation also showed enhanced systolic function with diastolic dysfunction in whole heart studies. This impaired relaxation (diastolic dysfunction) seen in these latter mice is predicted by the computer modeling presented in this article. However, since the mouse CTnI is not identical to the human CTnI the differences that exist between these proteins may be enough to give results that may be different from what occurs in the human heart.
Changes in the phosphorylation states of a few cardiac regulatory proteins (such as TnI) seem to be the basis for cardiac dysfunction in several disease states recently reported. A recent report by Deng et al. (46) showed that the TnI R145G mutation renders the actomyosin system at maximum Ca 2ϩ activation insensitive to protein kinase A phosphorylation suggesting that muscles containing this TnI mutation might have an altered response to this intracellular phosphorylation pathway.
In summary, it is likely that the human ventricle containing HCTnI R145G exhibits an increased Ca 2ϩ sensitivity of force and impaired inhibitory effectiveness of troponin I. This would manifest itself clinically with impaired isovolumetric relaxation, impaired rapid early diastolic filling, and increased left ventricular peak pressure. More clinical and biochemical studies are needed to lend more insight into the possible physiological effects of this mutation and all other FHC-associated mutations on cardiac muscle contraction and its regulatory processes.