Ala Scanning of the Inhibitory Region of Cardiac Troponin I*

In skeletal and cardiac muscles, troponin (Tn), which resides on the thin filament, senses a change in intracellular Ca2+ concentration. Tn is composed of TnC, TnI, and TnT. Ca2+ binding to the regulatory domain of TnC removes the inhibitory effect by TnI on the contraction. The inhibitory region of cardiac TnI spans from residue 138 to 149. Upon Ca2+ activation, the inhibitory region is believed to be released from actin, thus triggering actin-activation of myosin ATPase. In this study, we created a series of Ala-substitution mutants of cTnI to delineate the functional contribution of each amino acid in the inhibitory region to myofilament regulation. We found that most of the point mutations in the inhibitory region reduced the ATPase activity in the presence of Ca2+, which suggests the same region also acts as an activator of the ATPase. The thin filaments can also be activated by strong myosin head (S1)-actin interactions. The binding of N-ethylmaleimide-treated myosin subfragment 1 (NEM-S1) to actin filaments mimics such strong interactions. Interestingly, in the absence of Ca2+ NEM-S1-induced activation of S1 ATPase was significantly less with the thin filaments containing TnI(T144A) than that with the wild-type TnI. However, in the presence of Ca2+, there was little difference in the activation of ATPase activity between these preparations.

In skeletal and cardiac muscles, troponin (Tn), which resides on the thin filament, senses a change in intracellular Ca 2؉ concentration. Tn is composed of TnC, TnI, and TnT. Ca 2؉ binding to the regulatory domain of TnC removes the inhibitory effect by TnI on the contraction. The inhibitory region of cardiac TnI spans from residue 138 to 149. Upon Ca 2؉ activation, the inhibitory region is believed to be released from actin, thus triggering actin-activation of myosin ATPase. In this study, we created a series of Ala-substitution mutants of cTnI to delineate the functional contribution of each amino acid in the inhibitory region to myofilament regulation. We found that most of the point mutations in the inhibitory region reduced the ATPase activity in the presence of Ca 2؉ , which suggests the same region also acts as an activator of the ATPase. The thin filaments can also be activated by strong myosin head (S1)-actin interactions. The binding of N-ethylmaleimide-treated myosin subfragment 1 (NEM-S1) to actin filaments mimics such strong interactions. Interestingly, in the absence of Ca 2؉ NEM-S1-induced activation of S1 ATPase was significantly less with the thin filaments containing TnI(T144A) than that with the wild-type TnI. However, in the presence of Ca 2؉ , there was little difference in the activation of ATPase activity between these preparations.
Striated muscle thin filaments exist in equilibrium among multiple states. Ca 2ϩ binding to the regulatory domain of troponin C (TnC) 2 along the thin filaments and strong crossbridge interactions with thick filaments are thought to shift the equilibrium. Ca 2ϩ binds to the regulatory domain of TnC, which regulates the interaction of troponin I (TnI) with actintropomyosin (Tm) and TnC (1)(2)(3). In the thin filaments, the inhibitory region of TnI (residues 104 -115 of rabbit fast skeletal TnI (fsTnI) or 138 -149 of mouse cardiac TnI (cTnI)) undergoes a structural transition depending on the Ca 2ϩ state of TnC (4,5). In the absence of Ca 2ϩ at the regulatory site(s) of TnC, the inhibitory region interacts with actin to prevent activation of myosin ATPase activity. When Ca 2ϩ binds to the regulatory site(s) of TnC, the switch region of TnI, which is located at the C terminus of the inhibitory region, interacts with the newly exposed hydrophobic patch of the N-terminal regulatory domain of TnC (6 -8). This interaction causes the removal of the inhibitory region and the second actin-Tm binding region of TnI from the actin surface and allows actin to interact with myosin. In the presence of Ca 2ϩ at the regulatory sites of TnC, the inhibitory region and the central helical region of TnC are mutually stabilized, according to the recent x-ray crystal structure of the core domain of the fsTn complex (9). The sequence variations in the N-terminal and the C-terminal regions of TnT, another component of the Tn complex, are known to alter the Ca 2ϩ sensitivity of myofilament activity (10,11). In addition, TnT is involved in the Ca 2ϩ -dependent interaction of the Tn complex with actin-Tm (12). However, the molecular mechanism whereby TnT participates in the Ca 2ϩ regulation has not been established.
There is evidence supporting the idea that each amino acid residue in the inhibitory region of TnI contributes differently and to a different degree to myofilament activities. One example is genetic mutations and phosphorylation of amino acid residues in the inhibitory region of cardiac TnI that cause the modification of myofilament activities. In hypertrophic or restrictive cardiomyopathy-linked mutations found in the inhibitory region, such as R142Q, L145Q, and R146G/Q/W mutations (mouse cTnI sequence number), induce Ca 2ϩ sensitization of myofilament activities and an increase in ATPase/ tension at low [Ca 2ϩ ] (13,14). Recently we reported that thin filaments reconstituted with R146G or R146W mutant cTnI bind Ca 2ϩ tighter than those with cTnI(wt) (15). The Ca 2ϩ sensitization may occur as a result of the destabilization of the off-state of the thin filaments due to the mutation introduced into the actin-Tm-interacting residue, i.e. Arg-146, of cTnI. On the other hand, Thr-144 is phosphorylated by protein kinase C (PKC) specifically, although the consequence of the PKC-dependent phosphorylation of Thr-144 has not yet been clearly defined. Pseudophosphorylation of Thr-144 was shown to cause Ca 2ϩ desensitization in in vitro motility assays (16), whereas there is a report that indicates phosphorylation of Thr-144 sensitizes skinned cardiomyocytes to Ca 2ϩ (17). Furthermore, Tachampa et al. reported that Thr-144 of cTnI is important for length-dependent activation of skinned cardiac muscle (18). Thus in each case presented above, a specific change in a single amino acid in the inhibitory region of TnI induced different and divergent effects on myofilament activities.
Our aim of this study is to assess the functional contributions of the individual amino acid residues in the inhibitory region to the regulatory function. To assess the functional roles of the individual amino acid residues systematically, we used Ala scanning (19,20). Ala substitution deletes all the interactions made by atoms beyond ␤-C yet does not alter the peptide back-bone conformation, unless it is applied to Gly or Pro. Ala is one of the most abundant amino acids and is found in both buried and exposed positions. We found that almost the entire minimum inhibitory region of cTnI we investigated ( Fig. 1) is important for both the inhibition and activation. Our data also indicate that the C-terminal part of the inhibitory region destabilizes the active state of the thin filaments. We also found that Thr-144 is involved in NEM-S1-dependent activation of ATPase activity in the absence of Ca 2ϩ .

MATERIALS AND METHODS
Proteins-Recombinant human wild-type and mutant cTnCs, cTnC(C35S/T53C/C84S) and cTnC(C35S), in pET3d vector were expressed using BL21(DE3) cells. cTnC was extracted with 5% sucrose, 1 mM EDTA, 50 mM Tris-HCl, pH 8.0 and the protease inhibitors (AEBSF, E-64, and pepstatin A). After centrifugation, the supernatant fraction was collected, and CaCl 2 and MgCl 2 were added to final concentrations of 1 mM and 5 mM, respectively, followed by the addition of ammonium sulfate to 60% saturation. After centrifugation, the supernatant fraction was applied to a phenyl-Sepharose column equilibrated with 1 M NaCl, 1 mM CaCl 2 , and 50 mM Tris-HCl, pH 8.0. After washing with the same solution, the column was washed further with the same solution except that 0.2 mM CaCl 2 was used instead of 1 mM. Finally cTnC was eluted with 5 mM EDTA and 20 mM Tris-HCl, pH 8.0. The cTnC-containing fraction was dialyzed against 1 mM EDTA, 20 mM Tris-HCl, pH 8.0, and 1 mM DTT. Solid urea was added to the protein solution to 6 M, and cTnC was separated on a QAE fast flow-Sepharose column equilibrated with 6 M urea, 1 mM EDTA, and 20 mM Tris-HCl, pH 8.0. cTnC was eluted with a linear gradient of 0 -0.5 M NaCl. Recombinant mouse cTnIs were expressed and purified as described previously (21). Recombinant mouse cTnT with a Myc tag at the N terminus was expressed and purified with a combination of ammonium sulfate fractionation and DEAE-Sepharose column chromatography as described (22). Previously we showed that the Myc tag at the N terminus of cTnT has no affect on myofilament activity (19). Tropomyosin was extracted from bovine left ventricular ether powder with 1 M KCl, 1 mM EGTA, 1 mM DTT, and 50 mM Tris-HCl, pH 8.0 followed by isoelectric point precipitation. The precipitation fractions were resuspended with 1 M KCl, 10 mM MOPS, pH 7.0, and ammonium sulfate fractionation was carried out. The 45-60% saturation fraction was collected, and Tm was further purified on a QAE fast flow-Sepharose column. Actin was prepared from rabbit skeletal muscle acetone powder as described by Spudich and Watt (23). After purification, F-actin solution was stored in the presence of phalloidin (1:1 molar ratio). Myosin subfragment-1 (S1) was prepared by chymotryptic digestion of rabbit psoas muscle myosin and purified on a SP-Sephadex column as described previously (24).
Labeling of Proteins-Single Cys residues of mutant cTnC(C35S/T53C/C84S) and cTnC(C35S) were labeled with 2-fold amounts of IAANS over Cys in the presence of 0.1 M NaCl, 1 mM EDTA, and 20 mM HEPES, pH 7.5. The reaction was quenched by the addition of DTT, and the excess fluorescence dye was removed by dialysis and a desalting column. The labeling yield was determined using ⑀ 326 nm ϭ 27,000 M Ϫ1 for IAANS (15).
Reconstitution of the Tn Complex and the Thin Filament-Equimolar amounts of Tn components were combined in a solution containing 6 M urea, 1 M NaCl, 5 mM MgCl 2 , 0.1 mM CaCl 2 , 1 mM DTT, and 20 mM Tris-HCl, pH 8.0. The protein mixture was dialyzed against the same solution without urea. The NaCl concentration was next reduced to 0.3 M and finally to 0.1 M. After dialysis, the protein mixture was clarified by centrifugation, and then the resulting supernatant fractions were applied to a Resource-Q (1 ml, GE Healthcare) column equilibrated with 0.1 M NaCl, 5 mM MgCl 2 , and 20 mM Tris-HCl, pH 8.0. The Tn complex was eluted with a linear gradient of 0.1 to 0.5 M NaCl in the same solution.
To reconstitute thin filaments, we first mixed actin and Tm, followed by Tn. For the Ca 2ϩ binding measurements, we first mixed actin and Tm in a 8.5:1 molar ratio followed by the addition of IAANS-labeled Tn to 8.5:1:1 (actin:Tm:Tn) molar ratio with cTnC(C35S/T53C/C84S). A slightly high molar ratio of actin to Tn-Tm was necessary to minimize free Tn in the thin filament preparations. Because free Tn shows an opposite direction of fluorescence change upon binding Ca 2ϩ (25), the presence of excess free Tn may interfere with the measurements of Ca 2ϩ binding to thin filaments. To confirm that a slightly lower molar ratio of Tn and excess amounts of Tn produce the same results, we also carried out Ca 2ϩ binding measurements using an excess amount of Tn with IAANS-labeled cTnC(C35S) over actin and Tm.
Ca 2ϩ Binding Measurements-Steady-state fluorescence measurements were carried out using a Model 2000-4 spectrofluorometer equipped with two 814 PMT photon-counting detectors (Photon Technology International) with a cell holder containing a thermostat and a magnetic stirrer. The Ca 2ϩ binding was monitored by fluorescence emission of IAANS attached at Cys-53 of cTnC(C35S/T53C/C84S) or Cys-84 of cTnC(C35S). As mentioned above, the latter was used to make sure that the data obtained with a subsaturating amount of Tn in the thin filaments gave the same affinity as that of thin filaments with saturated amounts of Tn. The fluorescence emission intensity change observed was assumed to be due to direct Ca 2ϩ binding to the regulatory site of cTnC in the protein complexes. The solution conditions were 100 mM NaCl, 5 mM MgCl 2 , 1 mM EGTA, and 20 mM MOPS, pH 7.0. The titration was carried out at 25°C, and the free Ca 2ϩ concentration was calculated using the WEBMAXC Standard Program (26). The titration curves were analyzed as described previously (15,25). As a measure of Ca 2ϩ sensitivity, pCa 50 values were then calculated as ϪlogK d from the apparent dissociation constant, K d . The apparent dissociation constant, K d and pCa 50 are expressed as a mean Ϯ S.E. from 4 -6 experiments. The apparent coupling energy of Tn with Ca 2ϩ binding and actin-Tm interaction was calculated as in Equation 1, where K d CaTn and K d CaThin are the Ca 2ϩ dissociation constants for the Tn complex and the thin filaments, respectively. Error propagation was calculated by Equation 2, where ⌽ is a function of measurable quantities 1 and 2 , and , 1 , and 2 are errors associated with ⌽, 1 , and 2 , respectively. Acto-S1 ATPase Measurements-The reaction conditions were 5.0 M actin, 0.2 M myosin S1, 1.0 M Tm, and 2.0 M Tn in 35 mM NaCl, 5 mM MgCl 2 , 20 mM MOPS, pH 7.0, and either 0.1 mM CaCl 2 or 2 mM EGTA at 25°C. A reaction was initiated by the addition of a final concentration of 1 mM ATP. ATPase activity was determined from the time course of inorganic phosphate liberation using a malachite green assay (27). Steadystate ATPase activity was determined from 5 to 6 time points. In the case of NEM-S1-activated acto-S1 ATPase activity, free actin concentrations were held at 5.0 M as described (28,29), and ATPase rates were expressed per unmodified S1 concentration. In all figures and tables, the rate for S1 alone has been subtracted from the measured rates, and the rate is expressed as a mean Ϯ S.E. from 4 (without NEM-S1) or 8 -12 (with NEM-S1) measurements.
Statistical Evaluation-Statistical evaluation was carried out by one-way ANOVA followed by the Scheffe test as a post-hoc multiple comparison test; p Ͻ 0.05 was considered significant. Note that the Scheffe test is one of the most conservative evaluations and least likely produces "Type-1" (false-positive) error (30).

RESULTS
Ala-scanning Mutations-A segment cTnI, residues from 143 to 149 (Fig. 1), was subjected to Ala scanning to assess the impact of Ala substitution at each position. It should be mentioned that the Ala substitution of each of these residues seems to perturb neither the flexible structure of the inhibitory region nor the ␣-helical structure of the switch region, which follows the inhibitory region, based upon the secondary structural prediction by the AGADIR computer program (31).
Ca 2ϩ Binding Measurements and Ca 2ϩ -dependent Actin-Tm Interaction-We measured Ca 2ϩ binding to the Tn complex with Ala mutation of cTnI in the inhibitory region residues 143-149 either alone or in reconstituted thin filaments. The Ca 2ϩ binding was reported by fluorescence emission intensity change of IAANS attached to Cys-53 of the cTnC mutant. As we reported recently, IAANS attached to Cys-53 of the mutant cTnC(C35S/T53C/C84S) shows Ca 2ϩ binding affinity as well as Ca 2ϩ dissociation kinetics from the regulatory site of cTnC and of the cTn complex similar to those measured directly with Quin-2 using unlabeled wild-type cTnC and cTn (25). Furthermore, IAANS at Cys-53 of cTnC reports Ca 2ϩ binding to the reconstituted thin filaments with almost the same affinity as expected from previous reports (15,32,33). Thus, this labeling can be used not only for the Ca 2ϩ binding measurements for the cTn complex but also for the thin filaments as well. This allowed us to compare results obtained from different regulatory complexes. Ca 2ϩ binding to the Tn complex induces a decrease of the IAANS emission intensity, whereas Ca 2ϩ binding to the thin filaments enhances the fluorescence intensities. This indicates that the microenvironments of IAANS attached to Cys-53 are different in the Tn complex and the thin filaments in the presence of Ca 2ϩ . In the thin filaments, IAANS may face toward the actin-Tm surface so that it experiences a less exposed environment upon Ca 2ϩ binding. For the reconstituted thin filaments, we also conducted the Ca 2ϩ binding measurements using cTnC with IAANS attached to Cys-84 as mentioned under "Materials and Methods." The Ca 2ϩ binding properties of the Tn complexes and the reconstituted thin filaments are summarized in Table 1. The changes in the apparent pCa 50 of the regulatory complexes with mutant cTnI from that with wild type (⌬pCa 50 ) were FIGURE 1. Inhibitory region of TnI. A, sequence comparison of the minimum inhibitory region from various vertebrates. The amino acid residues that are different from fsTnI are colored green in cardiac sequences. Note the amino acid sequence of the inhibitory region is highly conserved. Also the amino acid sequences of the minimum inhibitory region of the mutants we investigated in this study are shown. B, crystal structure of the inhibitory region and its surrounding region in chicken fsTn complex in the Ca 2ϩ -bound form (PDB: 1YTZ). TnC, pink; TnT, light blue; TnI, gray. The segment, corresponding to residues 143-149 of mouse cTnI, is colored red. calculated from the apparent dissociation constants (K d ) and are shown in Fig. 2. These Ca 2ϩ binding properties were used to evaluate the perturbation of the interaction of the inhibitory region with other thin filament components by Ala substitution at each position in the segment residues from 143 to 149 (Fig. 3).
The effects of Ala substitution of one of these residues on Ca 2ϩ binding to the Tn complex are summarized in Table 1 and Fig. 2A: P143A, T144A, and L145A mutations did not perturb Ca 2ϩ binding properties of the Tn complex. Compared with the residues from segment 143-145 of cTnI, Ala substitution of the amino acid residues in the segment 146 -149 showed decreases in Ca 2ϩ affinity of the Tn complex. The R147A mutation demonstrated the largest decrease (⌬pCa 50 ϭ 0.18 Ϯ 0.01; p Ͻ 0.05 versus each of the rest of the mutations investigated). Nonetheless, there were relatively small perturbations of Ca 2ϩ binding properties of the Tn complex caused by Ala substitution of individual amino acid residues in the minimum inhibitory region.
In general, Ala substitution had more impact on Ca 2ϩ binding properties of the thin filaments (Table 1 and Fig. 2B) than those of the cTn complexes. Mutations that affected Ca 2ϩ binding to the thin filaments resulted in sensitization of the system to Ca 2ϩ , as shown by positive values of ⌬pCa 50 in Fig. 2B. The V148A mutation significantly increased the apparent Ca 2ϩ binding affinity of the thin filaments by ⌬pCa 50 ϭ 0.34 Ϯ 0.03. R146A and R149A, as well as T144A, also demonstrated sensitization of the thin filaments to Ca 2ϩ , but these were not significantly different from cTnI(wt) with ANOVA analysis (they were significant versus cTnI(wt) with the Student's t test). The P143A mutation did not alter the Ca 2ϩ binding property of the thin filaments. The thin filaments with the L145A mutation did not demonstrate a fluorescence intensity change when titrated with Ca 2ϩ . Therefore, we could not measure Ca 2ϩ binding to the thin filaments with L145A mutation.
Ca 2ϩ binding to the Tn complex and actin-Tm binding are negatively coupled. That is, Ca 2ϩ binding to the Tn complex   reduces its affinity for actin-Tm. The apparent Ca 2ϩ binding constants for the Tn complexes and the thin filaments were used to calculate coupling energies between Ca 2ϩ and actin-Tm (⌬⌬G coupl ) for each Tn complex (Fig. 3). The apparent coupling energy, ⌬⌬G coupl , is a measure of apparent Ca 2ϩdependence of the interaction of the Tn complex with actin-Tm (Reaction Scheme 1) and was calculated as described under "Materials and Methods." It should be mentioned that the thin filaments exist in equilibrium between at least two states, which possess different affinities for Ca 2ϩ ; Ca 2ϩ binding itself is capable of shifting the thin filaments state. Thus the scheme could be more complicated. Ca 2ϩ -dependent interactions of the Tn complex with actin-Tm can also be assessed using a co-sedimentation assay. However, it may require much higher salt concentrations as demonstrated previously (34). As evident from Fig. 3, the Ca 2ϩ -dependent interaction of the Tn complex with actin-Tm was affected by Ala mutations in the C-terminal-half of the inhibitory region (p Ͻ 0.05; R146A, R147A, V148A, and R149 versus wt). The V148A mutation showed a 0.55 Ϯ 0.04 kcal/mol (⌬⌬G coupl (wt) Ϫ ⌬⌬G coupl (V148A)) decrease in the Ca 2ϩ -dependent interaction. Because we could not determine the apparent Ca 2ϩ binding affinity of the thin filaments with the L145A mutation, we could not determine the ⌬⌬G coupl for the L145A mutation.
Actin-activated Mysoin-S1 ATPase with Ala-substitution Mutants-A summary of actin-activated myosin S1-ATPase activities is shown in Fig. 4. The addition of the wild-type Tn to the complex of myosin S1, actin, and Tm increased the ATPase activity to 0.9 s Ϫ1 from 0.28 s Ϫ1 in the presence of Ca 2ϩ (Fig. 4A), whereas it decreased the activity to 0.05 s Ϫ1 in the absence of Ca 2ϩ (Fig.  4B). Thus, the ATPase activity was increased ϳ18-fold by Ca 2ϩ in the presence of the Tn complex. In the presence of Ca 2ϩ , all but the T144A mutation examined in this study resulted in a decrease of ATPase activity compared with cTnI(wt) (Fig. 4A). These data clearly demonstrated the involvement of the inhibitory region in Ca 2ϩ activation. L145A, R147A, and V148A mutations greatly impaired Ca 2ϩ activation. ATPase activities with one of these mutations were 0.36 -0.40 s Ϫ1 . It is noteworthy that the P143A mutation impaired Ca 2ϩ activation without affecting the Ca 2ϩ binding property. In general, the shift of the equilibrium of the thin filament states seems to be accompanied by the alteration of the Ca 2ϩ binding property. Therefore, this observation for the P143A mutation deserves further investigation. In the absence of Ca 2ϩ , most of the Ala mutations impaired the inhibitory action (Fig. 4B). The L145A mutation showed the largest effect (p Ͻ 0.05 versus every cTnI examined in this study, including cTnI(wt)): it decreased the ATPase activity only to 0.17 s Ϫ1 , resulting in only ϳ2-fold Ca 2ϩ sensitivity. P143A, T144A, and V148A mutations showed no significant effect on the ATPase activity in the absence of Ca 2ϩ .
NEM-S1 Activation of ATPase Activities with Ala-substitution Mutants-The thin filaments can be activated through a strong cross-bridge interaction. We used NEM-S1 to mimic such an interaction. NEM-treated S1 has little ATPase activity by itself, binds to actin strongly even in the presence of ATP, and stabilizes the active state of the thin filaments. Thus, addition of NEM-S1 to the ATPase reaction mixture activates the thin filaments and enhances ATPase activity. Fig. 5 shows the comparisons of NEM-S1 activation of ATPase rates with Ala mutations versus those with cTnI(wt). As NEM-S1 concentration increased, the ATPase rate increased. In the case of cTnI(wt), the ATPase rate increased to nearly 2.0 s Ϫ1 with 1 M NEM-S1 in the presence of Ca 2ϩ (white bars in Fig. 5, A-G) and to 1.4 s Ϫ1 in the absence of Ca 2ϩ (dark gray bars in Fig. 5, A-G). With higher concentrations of NEM-S1, ATPase rates would be indistinguishable in the presence and absence of Ca 2ϩ . As reported above, the T144A mutation did not affect Ca 2ϩ -dependent actin-activated S1-ATPase activity (Fig. 4). The addition of NEM-S1-activated S1-ATPase activity with T144A mutation as well as that with wild-type Tn in the presence of Ca 2ϩ . Surprisingly, in the absence of Ca 2ϩ , ATPase activity with T144A was not activated as greatly as ATPase with wildtype TnI by NEM-S1 (Fig. 5B). At least up to 1 M NEM-S1, ATPase activities were always suppressed. With 3 M NEM-S1, The rate for S1 alone has been subtracted from the measured rates, and the rate is expressed as a mean Ϯ S.E. from four measurements. In A, * indicates p Ͻ 0.05 from cTnI(wt). # indicates p Ͻ 0.05 from cTnI(wt), cTnI(T143A), cTnI(T144A), cTnI(R146A), and cTnI(R149A). In B, * indicates p Ͻ 0.05 from cTnI(wt). # indicates p Ͻ 0.05 from the rest of cTnIs. Note that scales for the y axis are different in A and B.
ATPase activity was indistinguishable from that with wild-type TnI in the absence of Ca 2ϩ (data not shown), indicative of the lower affinity of NEM-S1 for the thin filaments with T144A. In the case of L145A mutation, which introduced a reduced ATPase activity in the presence of Ca 2ϩ and an increased activity in the absence of Ca 2ϩ compared with wild type (Fig.  4), an addition of 0.3 M NEM-S1 enhanced ATPase activity to almost the same level as that with wild type in the presence of Ca 2ϩ (Fig. 5C). On the other hand, in the absence of Ca 2ϩ , ATPase activities with NEM-S1 were higher or the same as those with wild type. Taken together, these results indicate the stabilization of the intermediate state of the thin filaments by the L145A mutation. In the case of the Ala mutations of the C-terminal-half of the inhibitory region (R146A to R149A mutation), ATPase rates were almost the same as those with cTnI(wt) with NEM-S1 in the absence of Ca 2ϩ (Fig. 5, D-G). In the presence of Ca 2ϩ and NEM-S1, however, ATPase rates were constantly lower than those with cTnI(wt). This indicates the destabilization of active state of the thin filaments by these Ala mutations and is also consistent with the data for the actin-activated S1-ATPase shown in Fig. 4.

DISCUSSION
Experiments reported here are the first to investigate the effects of the systematic Ala replacement of the inhibitory region of cTnI on the Ca 2ϩ -dependent and the strong cross-bridge-dependent myofilament activities using fully reconstituted systems. We found: 1) the C-terminal-half of the inhibitory region is responsible for the Ca 2ϩdependent interaction of the Tn complex with actin-Tm, 2) almost the entire segment that we investigated is responsible for the cTnI inhibitory action, 3) almost the entire segment is responsible for Ca 2ϩ activation, and most notably 4) Thr-144 is involved in strong cross-bridge-dependent activation of the thin filaments in the presence of EGTA, although T144A mutation did not impair inhibitory action in the absence of Ca 2ϩ nor activation in the presence of Ca 2ϩ without NEM-S1.
The inhibitory region was shown to interact with actin in the absence of Ca 2ϩ and inhibit the strong interaction between actin and myosin. One possible molecular mechanism for the inhibitory action was proposed by Patchell et al. (35). Based on the competition for the actin surface between a peptide derived from the inhibitory region of TnI and the actin-binding peptide derived from myosin, they concluded that the inhibitory region interacts with actin so that the actin monomer undergoes a conformational change that stabilizes the actin in a way to produce a surface less suitable for stable complex formation with myosin. On the other hand, threedimensional reconstruction of the thin filaments from electron microscope images suggests steric blocking of myosin binding to actin by Tm with the interaction between the inhibitory closed, mutant cTnI in EGTA. The rate is expressed as a mean Ϯ S.E. from 8 -12 measurements per unmodified S1. H, ATPase activities with the mutant cTnIs expressed relative to that with cTnI(wt). For each mutant labeled at the bottom, S1 ATPase activities without NEM-S1 (in the presence (white bars) and the absence (light gray bars) of Ca 2ϩ ) and with 0.5 M NEM-S1 (in the presence (dark gray bars) and the absence (black bars) of Ca 2ϩ ) are shown. Those with a significant difference (p Ͻ 0.05 by ANOVA) when compared with cTnI(wt) are marked with an asterisk. region and actin holding Tm onto the myosin-binding site of actin to prevent actin from interaction with myosin (36 -38). Regardless of the molecular mechanism underlying the inhibitory action by the inhibitory region of TnI, the equilibrium constant between the turned-off state of the thin filaments and the Ca 2ϩ -induced state of the thin filaments seems to be relatively small (39,40). We found that apparent free energy changes of Ca 2ϩ -dependent interaction of Tn to actin-Tm caused by Ala substitution of amino acid residues in the inhibitory region are only up to 0.55 kcal/mol (Fig. 3). However, ATPase activities were substantially affected by Ala substitution. Assuming that the ⌬pCa 50 values observed with the Ca 2ϩ binding experiments of the Tn complex and those with the thin filament reflect the mutational effects on the Ca 2ϩ -bound state and Ca 2ϩ -free state of the thin filaments, respectively, the changes in Ca 2ϩ -free state are more responsible for the loss of the apparent Ca 2ϩ -dependent interaction energy. This observation is consistent with other findings that demonstrated hypertrophic cardiomyopathy-linked mutations found in the inhibitory region do not reduce the affinity of the Tn complex to actin-Tm significantly, whereas myofilament activity in the absence of Ca 2ϩ is greatly disturbed (41)(42)(43). Also, this mutational effect of the inhibitory region is in striking contrast to the case with the mutation in the second actin-Tm binding domain. The second actin-Tm binding site of TnI is located in the C-terminal domain of TnI. The C-terminal mobile domain appears to form a globular structure when it binds to actin-Tm in the absence of Ca 2ϩ both in fsTnI and cTnI (44,45). Murakami et al. (44) determined the solution structure of the C-terminal domain of fsTnI and docked the structure of the C-terminal mobile domain to electron microscope images of the thin filaments. In their structural model, the second actin-Tm binding site spans residues ϳ138 -175 of fsTnI and ϳ172-207 of mouse cTnI. The 18-residue truncation from the C terminus of cTnI, as most of the Ala mutations of the inhibitory region observed in this study, resulted in a sensitization to Ca 2ϩ of the system, but, unlike Ala mutations, it did not affect the basal level of ATPase activity (46,47). This is true for some of the HCM-linked mutations. Most notably in in vitro motility assay, HCM-linked mutation from the inhibitory region, R146G (mouse sequence), caused a complete loss of Ca 2ϩ -dependent control, whereas HCM-linked mutations from the C-terminal domain, G204S, ⌬K184, and K207Q, retained Ca 2ϩ -dependent regulation over filament sliding velocity (48,49). Assuming the C-terminal mobile domain of cTnI forms the same structure as that of fsTnI proposed by Murakami et al. (44), Lys-184 and Lys-207 of cTnI are suggested to be involved in the direct actin binding. Thus these observations may reflect the functional difference between the inhibitory region and the second actin-Tm binding site of TnI. Interestingly the calculated apparent free energy changes of Ca 2ϩ -dependent interaction of Tn to actin-Tm shown in Fig. 3 do not always reflect the degree of impairment of the inhibitory action of ATPase activity in the absence of Ca 2ϩ shown in Fig.  4B. For example, the mutation V148A, that perturbs the Ca 2ϩdependent interaction of Tn with actin-Tm, did not impair the inhibitory activity. Data presented in Fig. 4 also demonstrate that the full inhibition of the acto-S1 ATPase activity at low [Ca 2ϩ ] requires the nearly entire minimum inhibitory region of TnI, although the degree of the effects of Ala mutations on the inhibitory action differs from residue to residue. Previously, van Eyk and Hodges (50) conducted Gly-substitution experiments for the inhibitory region of fsTnI. Although their experiments were carried out with short 12-residue peptides corresponding to the minimum inhibitory region and their system for the ATPase measurements did not include TnT and TnC, they also found a broad distribution of the amino acid residues that affect the basal level of ATPase activity.
Whereas the physiological importance of the interaction of the inhibitory region with actin at low [Ca 2ϩ ] is obvious, the interaction with TnC at high [Ca 2ϩ ] has been puzzling and controversial. The C-terminal part of the inhibitory region, rather than the inhibitory region itself, was identified as the Ca 2ϩ -dependent switch region, which interacts with the regulatory domain of TnC in a Ca 2ϩ -dependent manner (6 -8). In the case of fast skeletal Tn, most of the Ca 2ϩ -dependent binding energy between TnC and TnI appears to stem from the interaction between the switch region of TnI and the N-terminal regulatory domain of TnC. For the inhibitory region, on the other hand, the crystal structure of the core domain of fsTn complex in the presence of Ca 2ϩ showed that it interacts with the central helical region (D/E helix) of TnC (9). For cardiac Tn, the D/E helix of TnC was disordered, and the inhibitory region was not visible in the crystal structure of the core domain of cTn complex (51), whereas small angle x-ray diffractions of the cardiac thin filaments indicate an extended structure of the D/E helix region of TnC (52). These observations indicate flexibility in the linker region of the Tn complex and suggest a relatively weak interaction between the inhibitory region of TnI and the central helix of TnC. However, Ala substitution in the inhibitory region greatly suppresses actin-activated S1 ATPase activity in the presence of Ca 2ϩ . From a study employing a wide range of NEM-S1 concentrations to activate the thin filaments, we recently concluded that cardiomyopathy-linked mutations found in the inhibitory region of cTnI, R146G and R146W (mouse sequence number) stabilize the functional intermediate state of the thin filaments (51). The discrepancy between the Ca 2ϩ affinity and the ATPase rate found in this study also indicates an intermediate state with a unique activity. In our Ala-scanning experiments, the C-terminal-half part of the inhibitory region destabilized the active state of the thin filaments when mutated to Ala. Whether the extended structure of the central D/E helix is involved in the activation process as originally proposed by Herzberg and James (53) or Ala mutation of the C-terminal-half of the inhibitory region affects the structural opening of the N-terminal regulatory domain of cTnC, which may be a primary determinant of myofilament activity (54), remains to be solved.
In this study, we could not determine the ⌬⌬G coupl with the Leu-145 mutation, because the thin filaments with the L145A mutation did not show the fluorescence intensity change when titrated with Ca 2ϩ . However, as shown in Fig. 4B, the L145A mutation increased the actin-activated S1-ATPase activity in the absence of Ca 2ϩ , suggesting this mutation impaired the interaction between the inhibitory region of cTnI and actin in the absence of Ca 2ϩ . Also the L145A mutation resulted in a decrease in the actin-activated S1-ATPase activity in the presence of Ca 2ϩ . As we discussed in detail elsewhere (43), a mutation that causes a decrease and an increase of the ATPase activity in the presence and the absence of Ca 2ϩ , respectively, is likely to stabilize the functional intermediate state of the thin filaments. This is consistent with our data on the NEM-S1 activation of ATPase activity shown in Fig. 5C, which clearly illustrate a stabilization of the intermediate state of the thin filaments as mentioned above. The thin filaments with a mutation that stabilizes the intermediate state show a higher affinity for Ca 2ϩ (15), although the Ca 2ϩ binding properties of the each state of the thin filaments have not been characterized. Therefore, it is plausible that the L145A mutation impairs the Ca 2ϩdependent interaction with actin-Tm and decreases the ⌬⌬G coupl value.
The T144A mutation showed little effects on Ca 2ϩ -dependent interaction of Tn with actin-Tm (Fig. 3), and no effects on actoS1-ATPase activities in the presence and absence of Ca 2ϩ (Fig. 4). However, the T144A mutation impaired NEM-S1-dependent activation and S1 binding of the thin filaments in the absence of Ca 2ϩ (Fig. 5, B and H). This is most likely due to a reduced affinity of S1 for the off-state of the thin filaments with T144A. Thus modification of Thr-144, such as phosphorylation or replacement with another amino acid, could result in an alteration of Ca 2ϩ sensitivity and/or cooperativity of myofilament activity, because strong cross-bridge interaction can affect these parameters.
In summary, our investigation of the nearly complete minimum inhibitory region of cTnI indicates that it is important for both the inhibition and activation. This observation strongly suggests the presence of an intermediate state of the thin filaments that possesses the intermediate activity. We also found that Thr-144, which showed no effect on Ca 2ϩ -dependent ATPase activity when mutated to Ala, is involved in strong cross-bridge-dependent activation of ATPase in the absence of Ca 2ϩ .