Regulatory Properties of the NH2- and COOH-terminal Domains of Troponin T

The contraction of skeletal muscle is regulated by Ca2+ binding to troponin C, which results in an internal reorganization of the interactions within the troponin-tropomyosin complex. Troponin T is necessary for Ca2+-dependent inhibition and activation of actomyosin. Troponin T consists of an extended NH2-terminal domain that interacts with tropomyosin and a globular COOH-terminal domain that interacts with tropomyosin, troponin I, and troponin C. In this study we used recombinant troponin T and troponin I fragments to delimit further the structural and regulatory interactions with the thin filament. Our results show the following: (i) the NH2-terminal region of troponin T activates the actomyosin ATPase in the presence of tropomyosin; (ii) the interaction of the globular domain of troponin T with the thin filament blocks ATPase activation in the absence of Ca2+; and (iii) the COOH-terminal region of the globular domain anchors the troponin C-troponin I binary complex to troponin T through a direct Ca2+-independent interaction with the NH2-terminal region of troponin I. This interaction is required for Ca2+-dependent activation of the actomyosin ATPase activity. Based on these results we propose a refined model for the troponin complex and its interaction with the thin filament.

Troponin (Tn) 1 and tropomyosin (Tm) mediate the Ca 2ϩ -dependent regulation of skeletal muscle contraction through control of conformational features of the actin-based thin filament (1)(2)(3)(4)(5). Actin, Tm, and Tn are present in the thin filaments in a molar ratio of 7:1:1 (6 -8). The binary troponin C-troponin I (TnC⅐TnI) complex confers little Ca 2ϩ sensitivity to the actomyosin ATPase activity at physiological ratios of actin to Tn, is not stably assembled on the actin filament, and is not capable of activating the ATPase at high Ca 2ϩ levels. Troponin T (TnT) is necessary for full Ca 2ϩ sensitivity of the actomyosin ATPase (9 -13) and for restoration of full velocity in sliding filament assays (14).
The mechanism through which TnT exerts its role in the regulatory function of the troponin complex is not fully understood. TnT interacts with TnC, TnI, and Tm and holds the TnC/TnI dimer in the thin filament irrespective of the Ca 2ϩ concentration (9,(15)(16)(17).
Each Tm dimer spans seven actins, strongly suggesting that the regulatory function of the troponin complex is transduced through Tm to the actin molecules. TnT is the troponin subunit that most strongly binds to Tm (1,2). Two separate sites of attachment for TnT on Tm have been identified. The first site is near the head-to-tail overlap of sequential Tm molecules along the filament (16) and interacts with the T 1 fragment of TnT (Fig. 1). This interaction is independent of Ca 2ϩ binding to TnC (18). The second site is within fragment T 2 of TnT ( Fig. 1) which binds near residues 150 -180 of the Tm molecule (16,19). The T 2 fragment of TnT also interacts with TnC and TnI in vitro and to Tm in a Ca 2ϩ -sensitive manner in the presence of TnC (18).
In this work we used different combinations of TnT and TnI fragments to map the regions involved in both structural and regulatory interactions present in the trimeric troponin complex within the thin filament in the presence and in the absence of calcium. (20) and myosin (21) were prepared from the pectoralis major muscle of adult chickens. ␣-Tm (22) was prepared from adult chicken heart muscle. Recombinant TnI and TnC were isolated as described (23,24). Deletion mutants of TnI were prepared as described (13).

Proteins-Actin
Construction of TnT Deletion Mutants-Chicken skeletal muscle TnT-3 cDNA (GenBank TM accession number M22156) (25) was used as a template for site-directed mutagenesis (26). An NdeI site was inserted at codon Met 1 (13). M13mp18-TnT-3 was mutated with the oligonucleotide 5Ј-GTTATACCAGTAGCAGACTGA-3Ј to change the codon Leu 217 into a stop codon (underlined), with the oligonucleotide 5Ј-GGCGCAAG-TAATTGAACATTGA-3Ј to change the Pro 192 codon into a stop codon, and with the oligonucleotide 5Ј-AGGCGCAAGCATATGAACATTGA-CC-3Ј to introduce an NdeI site (underlined) at codons 192 and 193. The NdeI-EcoRI fragments of the M13mp18-TnT-3 containing a stop codon at positions 192 or 217 were cloned into the same restriction sites of the pET-3a (27). These vectors express, respectively, the fragments TnT 1-216 (the first 216 amino acids of TnT) and TnT  (the first 191 amino acids of TnT). The small NdeI-EcoRI fragment (produced by the second NdeI site, inserted at positions 192 and 193) was subcloned in the same sites of pET-3a, producing a vector for the expression of the fragment TnT 194 -263 (amino acids 194 -263). For the production of the fragment TnT 157-263 , the NcoI-BamHI fragment of the expression vector pET-TnT* (11) was subcloned in the same sites of pET-3d (27).
Reconstitution of the Troponin Complex-Tn subunits (20 M final of each) were combined in a 1:1:1 molar ratio in 6 M urea, 1 M KCl, 50 mM CaCl 2 , 20 mM imidazole, pH 7.5, 1 mM DTT. Successive dialysis (4°C, 12 h each) against the same buffer containing 4.6 M urea, 2 M urea, no urea, 100 mM KCl, and no KCl were used to gradually reduce the urea and salt concentrations. After dialysis the reconstituted complexes were centrifuged (12,000 ϫ g, 10 min), and the supernatant was aliquoted and stored at Ϫ70°C.
Co-sedimentation Assays-Tn binding to actin/Tm was analyzed using an ultracentrifugation assay. Before each assay, Tn subunits and complexes were centrifuged to remove insoluble material. Actin (20 M), Tm (2.86 M), and Tn (2.86 M) were combined in 20 mM imidazole, pH 7.0, 60 mM NaCl, 3 mM MgCl 2 , 0.5 mM CaCl 2 , 2 mM ␤-mercaptoethanol (ϩCa 2ϩ ), or in the same buffer with 0.5 mM EGTA replacing the CaCl 2 (ϪCa 2ϩ ). The mixtures were centrifuged at 315,000 ϫ g for 10 min at 4°C in a Beckman Optima TLX ultracentrifuge. The pellets were rinsed and resuspended in the original volume. Equivalent volumes of the mixture before centrifugation and of the supernatants and pellets after centrifugation were analyzed by 15% SDS-PAGE or 12.5% Tricine/SDS-PAGE (30) for the small fragments TnT 194 -263 and TnT 157-263 . Densitometric quantification was performed using a dual wavelength scanner (Shimadzu CS-9000) at 550 nm. Since TnT and TnI and their fragments have a low solubility at low ionic strength, control experiments in the absence of actin-Tm were performed to ensure that at these relatively low protein concentrations, all the Tn components were soluble.

RESULTS
TnT Deletion Mutants-A schematic representation of the TnT mutants is shown in Fig. 1. The precise sites for the insertion of stop codons were chosen so that the predicted ␣-helices would not be interrupted (Fig. 1). The mutants were expressed as non-fusion proteins and purified to homogeneity.
Whole Tn complexes containing wt-TnC, wt-TnI, and wt-TnT or the different TnT fragments were reconstituted from the isolated components. After reconstitution, the TnT mutants became soluble at low ionic strength at high protein concentrations, indicating that they were incorporated into stable complexes. When analyzed by gel filtration in the presence of Ca 2ϩ , all complexes eluted as a single peak (data not shown). In the absence of Ca 2ϩ , TnT 1-216 was dissociated from the ternary complex, which suggests that residues 217-263 in the COOHterminal region of TnT are necessary for the incorporation of TnT into the trimeric complex in the absence of Ca 2ϩ (see below).
Regulation of Actomyosin ATPase-We analyzed the regulatory properties of complexes containing the different TnT mutants ( Fig. 2A). Under physiological molar ratios of actin, Tm, and Tn (7:1:1), we confirmed that the control Tn complex containing wt-TnT confers full Ca 2ϩ sensitivity to actomyosin ATPase, i.e. it inhibits the ATPase activity in the absence of Ca 2ϩ and activates the ATPase activity in the presence of Ca 2ϩ (10) (activation is defined as the ability of troponin, in the presence of Ca 2ϩ , to increase the actomyosin/tropomyosin ATPase activity to levels above its activity in the absence of troponin). We also confirmed that the binary TnI⅐TnC complex was not able to inhibit or to activate the ATPase activity at physiological ratios of actin to troponin ( Fig. 2A) (10,31).
Inhibitory Function of the Mutant TnT Complexes-Deletions of the COOH-terminal region of TnT or deletion of the T 1 region reduces the inhibitory function of the troponin complex ( Fig. 2A): TnT 1-216 and TnT 157-263 complexes inhibit the ATPase activity to about 60% and TnT  complex to about 75%, whereas the wt⅐TnT complex inhibits to about 35%. The complex containing TnT 194 -263 conferred no Ca 2ϩ sensitivity to the actomyosin ATPase, a behavior identical to the TnI⅐TnC binary complex ( Fig. 2A). Since the partial inhibition observed for the TnT mutant complexes could be explained by a lower affinity for the thin filament, we performed ATPase experiments using increasing molar ratios of the troponin complex to actin. We observed full inhibitory activity of complexes containing TnT 1-191 , TnT 1-216 , and TnT 157-263 at actin:troponin molar ratios of 4:7, 3:7, and 3:7 respectively (data not shown). These results imply that the region shared by these mutants, namely amino acids 157-216, bind to the thin filament and may have an inhibitory role.
Activation Function of the Mutant TnT Complexes-The complexes containing TnT 1-216 , TnT 157-263 , and TnT 194 -263 were not capable of activating the ATPase activity ( Fig. 2A) even when increasing ratios of Tn to actin were used (data not shown). The complex containing TnT  was the only one capable of activating the ATPase activity like wt-Tn in the presence of Tm ( Fig. 2A). This suggests that the NH 2 -terminal end of TnT (residues 1-191) contains the region responsible for the activation of the ATPase activity. The evidence presented so far suggests that regions 1-191, 157-216, and 217-263 of TnT participate, respectively, in activation, inhibition/thin filament binding, and TnI/TnC binding. The experiments described below were designed to test and essentially confirm these hypotheses.
Effects of Isolated TnT Fragments on the Actomyosin ATPase-The direct effects of the different regions of TnT on the inhibition and activation of the ATPase activity was analyzed in the presence of Tm but in the absence of TnI and TnC (Fig. 2B). TnT  is able to activate the ATPase activity, whereas wt-TnT and the mutants TnT 157-263 and TnT 194 -263 are not. TnT 1-216 , which is not able to activate the ATPase activity in the context of the ternary complex ( Fig. 2A), is able to activate it in the absence of TnC and TnI, although to a lesser extent than TnT  . These results indicate that residues 191-263 of TnT are blocking the ability of region 1-191 to activate the ATPase activity and may explain why no activation is observed with isolated full-length TnT.
TnC/TnI Regions Involved in the Activation of the ATPase-Since full-length TnT alone does not activate the ATPase, but does so in the presence of TnI/TnC, we determined which regions within TnC/TnI are required for activation. We used TnI 103-182 , which binds to TnC-Ca 2ϩ and is capable of inhibiting the ATPase activity but does not bind to TnT, and the TnI 1-98 mutant, which binds to TnT and TnC but shows no inhibitory ability (13). The regulatory properties of the ternary complexes containing TnT, TnC, and the different TnI frag-ments are shown in Fig. 3A. Although the complex containing TnI 103-182 inhibited the ATPase activity as well as the complex containing TnI, it was not able to activate the ATPase activity. The complex containing TnI 1-98 activates the ATPase to the same extent as the complex containing wt-TnI (Fig. 3A). This  (1,2). Two highly conserved regions present in TnTs from different species (from Drosophila melanogaster to Homo sapiens) are represented by shaded boxes. The first of these regions corresponds closely to CB2, which is predicted to be highly helical and the site of interaction of the T1 fragment with Tm. The second, within T2, coincides with a region containing a series of heptad repeats, typical of coiled-coil structures (2,5,36,54). The exons present in the cDNA of isoform 3 (25) are numbered. The predicted ␣-helices (55) are represented by boxes below wt-TnT. Two COOH-terminal deletion mutants and two NH 2 -terminal deletion mutants were constructed. TnT 157-263 corresponds closely to the T2 proteolytic fragment (TnT 157-216 is four amino acids longer at the NH 2 terminus than T2). activation was independent of the Ca 2ϩ concentration, although slightly higher levels of activation were produced in the absence of Ca 2ϩ as observed previously (13). These results indicate that the NH 2 -terminal region of TnI but not the COOH-terminal region is necessary for the activation of the ATPase activity.
Next we analyzed if the NH 2 -terminal half of TnI is sufficient to activate the ATPase activity or if TnC is also required. The ATPase activity in the presence of filaments containing the different TnT mutants was titrated with increasing amounts of TnI 1-98 in the absence of TnC (Fig. 3B). TnI 1-98 in combination with wt-TnT does not activate the ATPase in the absence of TnC. TnI 1-98 in combination with TnT 157-263 was also not able to activate. The effects of TnT  and TnT 1-216 on the ATPase activity were not altered by the addition of TnI 1-98 . These results indicate that TnI 1-98 requires TnC to produce activation with wt-TnT.
In conclusion, our ATPase results show that residues 1-191 of TnT contain an activation domain, whereas the region between residues 191 and 263 has an inhibitory effect. Thus activation can only be observed with intact TnT in the presence of TnC/TnI and Ca 2ϩ . Furthermore, the NH 2 -terminal regions of TnI in conjunction with TnC are necessary for activation by the ternary complex.
Incorporation of the Troponin Complexes into the Thin Fila-ment-The interactions between TnT fragments, TnI and TnC, were analyzed in an actin/tropomyosin co-sedimentation assay. The TnC⅐TnI complex is partially retained in the thin filament in the absence of Ca 2ϩ but remains soluble if Ca 2ϩ is present (Fig. 4A). In the presence of wt-TnT, all three components of the Tn complex remained in the thin filament, irrespective of the Ca 2ϩ concentration (Fig. 4B) (15,31). If trimeric complexes containing TnT 1-216 and TnT  are incubated with actin/Tm, these two TnT mutants remained associated with the thin filament in the presence or absence of Ca 2ϩ but were not able to retain the TnC/TnI dimer in the thin filament in the pres- ence of Ca 2ϩ (Fig. 4, C and D). These controls confirm the previous observations that TnT is required to anchor the TnC⅐TnI complex to the thin filament in the presence of Ca 2ϩ (15,31) and that only the T 2 region of TnT interacts with the TnC/TnI dimer (16,32). These results specifically implicate the COOH-terminal half of T 2 (residues 217-263) in this interaction.
Interaction between the COOH-terminal Region of TnT and NH 2 -terminal Region of TnI-It has been previously demonstrated that the NH 2 -terminal region of TnI (TnI 1-98 ) is required for the incorporation of TnT into the ternary complex (12,13). Fig. 4E shows that the binary TnC⅐TnI 1-98 complex does not associate with the thin filament in the absence of TnT irrespective of the Ca 2ϩ concentration. In contrast, complexes containing TnC⅐TnI 1-98 ⅐wt-TnT are always associated with the thin filament (Fig. 4F). Since TnT 1-216 and TnT 1-191 lack residues 216 -263 which interact with the TnC⅐TnI binary complex, they were not expected to, and indeed did not, retain TnI 1-98 /TnC bound to the thin filament (Fig. 4, G and H).
Control complexes containing TnI 103-182 (which does not bind TnT) were analyzed. TnI 103-182 remained associated with the filament in the absence of Ca 2ϩ and was removed by TnC in the presence of Ca 2ϩ , both in the presence and absence of TnT (13, data not shown). These findings demonstrate that the NH 2 -terminal region of TnI interacts with the region between residues 216 and 263 of TnT and that this interaction is the major point of calcium-independent anchoring of the TnC/TnI dimer to TnT.
Since isolated TnC has been shown to interact with TnT in the presence of Ca 2ϩ (reviewed in Refs. 1, 2, and 5), we analyzed the binding of the binary TnC⅐TnT complex (without TnI) to actin/Tm. The amount of TnC associated with the thin filaments was measured as a function of total TnC added in the presence or absence of TnI (Fig. 5). In the absence of TnI, a 10-fold excess TnC (30 M) had to be used in the sedimentation assay to obtain approximately a 1:1 ratio of TnC:TnT bound to the filament (Fig. 5). In the presence of TnI, a 2-fold excess of TnC (6 M) was sufficient. This suggests that there are two different sites of interaction between TnC and the thin filament in the presence of Ca 2ϩ . The first, through binding to TnI which in turn binds TnT, is a strong binding site (reviewed in Refs. 1, 2, and 5). The second, through direct interaction of TnC with TnT, is much weaker.

The Interaction of the Region between Residues 157 and 263 of TnT with the Thin Filament-In the presence and absence of
Ca 2ϩ the affinity of the TnT 194 -263 ⅐TnC⅐TnI complex for the thin filament is reduced (Fig. 6B) when compared with the affinity of the TnT 157-263 ⅐TnI⅐TnC complex (Fig. 6A). This result suggests that a binding site to actin/Tm may reside in the NH 2 -terminal region of TnT 157-263 (residues 157-191) and that this binding is influenced by Ca 2ϩ binding to TnC. When the control TnT 157-263 ⅐TnC complex was incubated with actin/Tm in the absence of Ca 2ϩ , TnT 157-263 was incorporated into the filament while TnC remained in solution (Fig. 6C). This confirms previous observations that TnT 157-263 itself presents a binding site to actin/Tm (16,32). In the presence of Ca 2ϩ , both subunits remained in the supernatant indicating that TnC is able to remove TnT 157-263 (TnT 2 ) from the thin filament in the presence of Ca 2ϩ (Fig. 6C).
Even though TnT 2 fragment alone is able to interact with Tm (18) and with actin/Tm, it is still not clear how strong this interaction is in the context of the ternary troponin complex. Our results do not rule out the possibility that the interaction of TnT 157-263 with the thin filament in the absence of Ca 2ϩ is mainly mediated by TnI, which itself binds to the thin filament. To address this question we investigated how the interaction of TnT 157-263 with the thin filament is modulated by the TnI fragments (Fig. 6, D and E).
In the presence of Ca 2ϩ the three subunits remained in solution (Fig. 6E). These results indicate that although Ca 2ϩ binding to TnC was able to remove both TnT 157-263 and TnI 103-182 from the filament, the interaction of TnI 1-98 /TnC with TnT reduces the affinity of TnT 157-263 for the thin filament irrespective of the Ca 2ϩ concentration. Therefore, we conclude the following. (i) TnI 1-98 seems to predispose the dissociation of TnT 157-263 from the thin filament independent of Ca 2ϩ concentration. Therefore, in the context of the ternary complex, the direct interaction between TnT 2 and the thin filament is likely to be diminished. This observation needs to be confirmed with direct evidence obtained from equilibrium binding experiments. This explains the Ca 2ϩ -independent activation by TnT/TnI 1-98 /TnC. (ii) The main function of Ca 2ϩ binding to TnC is the dissociation of TnI 103-182 from the thin filament.

DISCUSSION
Troponin inhibits the actomyosin-tropomyosin Mg 2ϩ -ATPase in the absence of Ca 2ϩ and activates the ATPase in the presence of Ca 2ϩ . The original model of two-site binding between troponin and actin-Tm (32,33) distinguished the Ca 2ϩ -independent interaction between the NH 2 -terminal domain of TnT (TnT 1 ) and Tm-actin from the Ca 2ϩ -dependent interaction of the globular troponin domain (TnT 2 -TnI-TnC) with actin-Tm. In the two-site binding model, Ca 2ϩ -induced dissociation of the globular domain removes inhibition by TnI; this model does not address how activation is achieved. In this paper we have identified a TnT fragment that possesses intrinsic activation activity. We also showed that the anchoring of the TnC/TnT dimer to TnT occurs via an interaction between the aminoterminal domain of TnI and the last 50 residues of TnT. These two new aspects of the troponin complex have been incorporated into a refined model of thin filament regulation.
The Amino-terminal Region of TnT Activates Actomyosin ATPase-Ca 2ϩ -induced activation of actomyosin ATPase activity is a well known property of muscle troponin (10, 34) and recombinant troponin (11). In the absence of the inhibitory domain of TnI, the troponin complex activates the ATPase independently of the Ca 2ϩ concentration (Ref. 13 and this work). Although TnT is required for activation in the context of the whole complex, very few studies have analyzed the regulatory properties of isolated TnT and TnT fragments. Studies of the effects of isolated TnT on the ATPase have varied from no significant effect (this work), to a slight inhibitory effect ascribed to contaminating TnI (10), to a significant inhibitory effect (35). In this report we demonstrate that two aminoterminal fragments of TnT (TnT  and TnT 1-216 ) can activate the ATPase in the presence of Tm to levels observed for the whole troponin complex in the presence of Ca 2ϩ . Although this effect is probably mediated by Tm, to our knowledge, this is the first report that activation of the actomyosin ATPase can be obtained with a fragment derived from a single troponin subunit.
In the context of the whole troponin complex, only full-length TnT and TnT 1-191 activated the ATPase, whereas a larger NH 2 -terminal fragment (TnT 1-216 ) or a COOH-terminal fragment (TnT 156 -263 ) failed to activate. Troponin complexes reconstituted with these TnT fragments bound to the thin filament and did inhibit in the absence of Ca 2ϩ , although full inhibition was achieved only at Tn:actin ratios higher than those required by the wild-type complex. Troponin reconstituted with the smallest COOH-terminal TnT fragment (TnT 194 -263 ) did not inhibit or activate and showed very low affinity for the filament, suggesting that residues 156 -193 are involved in either binding or inhibition (also see "Discussion" in Ref. 36). Other workers have analyzed the regulatory properties of TnT fragments incorporated into troponin complexes. Ohtsuki and coworkers (37,38) have presented the only other studies of the regulatory properties of an amino-terminal fragment of TnT, namely TnT 1 (residues 1-158; see Fig. 1) obtained by chymotryptic digestion of rabbit skeletal muscle TnT. The behavior of TnT 1 was significantly different from what we observed for TnT  . Troponin reconstituted with TnT 1 inhibited the ATPase at low Ca 2ϩ concentrations and did not activate the ATPase to levels above those observed in the absence of TnT. A direct comparison of the results obtained with rabbit TnT 1 and chicken TnT  is complicated by their different lengths (our chicken TnT  corresponds to residues 1-186 of the rabbit TnT sequence, Ref. 39). Their different sequences and the fact that the rabbit TnT fragments presumably contain a mixture of isoforms with significant heterogeneity (25,40) must also be taken into consideration. The region between residues 162 and 191 alone is not responsible for activation since TnT 157-263 did not activate on its own or in reconstituted troponin. Analysis of the regulatory properties of smaller amino-terminal fragments of TnT isoforms is necessary to localize the precise sequences responsible for this activation.
A number of studies have analyzed the regulatory properties of COOH-terminal fragments (or small NH 2 -terminal deletions) of TnT in the context of the whole complex. Ohtsuki and co-workers (37,38,41) analyzed the regulatory properties of TnT 2a (residues 159 -259; see Fig. 1) derived from chymotryptic digestion of rabbit skeletal muscle TnT. In the two earlier studies (37,41), troponin complexes reconstituted with TnT 2a inhibited the ATPase but did not activate above levels observed in the absence of troponin, in agreement with our observations for complexes reconstituted with TnT 156 -263 . In a later study (38), both activation and inhibition was observed at levels comparable to that observed for wild-type troponin. These differences were attributed to contamination by TnT 2b fragments (residues 159 -242) in the earlier reports. Recent reports from other laboratories have concentrated on larger COOH-terminal fragments. Fragments corresponding to residues 39 -284 of bovine cardiac TnT and residues 46 -259 of rabbit skeletal TnT both activated and inhibited the acto-S1 ATPase, but in both cases the level of activation was significantly reduced (see Fig.  3 in Ref. 42 and Fig. 6 in Ref. 43). These results are consistent with the conclusion that the amino-terminal region of TnT is involved in the activation activity of TnT.
The Carboxyl Terminus of TnT Anchors the TnC/TnI Dimer by Way of the Amino Terminus of TnI-It is well established that the TnC⅐TnI binary complex binds to the COOH-terminal globular domain of TnT, although the exact sequences involved in the interactions with TnC and TnI are not fully elucidated (reviewed in Refs. 2 and 44). Although TnC can bind to TnT on its own (45)(46)(47)(48)(49), comparison of the free energies of formation of troponin binary and ternary complexes shows that the TnT-TnC interaction within troponin may be weaker than that observed in the binary complexes (47,48). We also found that the TnC-TnT interaction is relatively weak when TnT is bound to the thin filament and that stoichiometric binding at micro-molar protein concentrations only occurs in the presence of TnI.
The amino-terminal domain of TnI interacts with TnT (13, 36, 50 -52) and is involved in the Ca 2ϩ -independent interaction with TnC (13,53). In this work we presented the following evidence supporting the conclusion that residues 216 -263 of TnT are involved in the interaction with TnI 1-98 : (i) deletion of residues 216 -263 of TnT abolishes the binding of TnI 1-98 -TnC to the thin filament; and (ii) the deletion of residues 1-102 of TnI abolishes the binding of TnC/TnI 103-182 to actin-Tm-TnT in the presence of calcium. We conclude that the three subunits are held together by a core of structural, Ca 2ϩ -independent interactions between the COOH-terminal domain of TnC, the amino-terminal domain of TnI (TnI 1-98 ), and the last 50 residues of TnT (Fig. 7). Consistent with this finding, Jha et al. (36) reported that a deletion of residues 202-258 of human fast skeletal TnT reduces its interaction with rabbit TnI-TnC and that an amino-terminal fragment of TnI (TnI 1-120 ) forms a stable complex with TnT.
Transmission of the Ca 2ϩ Binding Signal to TnT-Although not explicitly stated in the two-site binding model (32,33), an activation function of the TnT 1 fragment is implied since it would be the only fragment interacting with actin-Tm in the presence of Ca 2ϩ . In this work we have explicitly shown that a TnT fragment containing the TnT 1 region can activate the ATPase on its own. In the context of the two-site binding model, Ca 2ϩ -induced dissociation of the globular domain would liberate the activation function of the NH 2 -terminal domain of TnT. What interactions are responsible for this liberation? In other words, is the Ca 2ϩ binding signal transmitted to TnT directly through TnC-TnT interactions, through the TnI 1-98 -TnT interaction, or both? It has been suggested that the activation of the ATPase comes solely from a specific Ca 2ϩ -dependent interaction between TnT and TnC (12). This was based on ATPase studies using a TnI deletion mutant (TnI d57 , deletion of residues 1-57) which interacts with TnC but not with TnT in binary complexes (12,52,53). On the other hand, a role for TnI in activation is suggested by our observation that TnT-TnI 1-98 -TnC activates the ATPase whereas TnC-TnT does not. It is noteworthy that both the COOH-terminal region of TnT and the NH 2 -terminal region of TnI have conserved hydrophobic heptad repeats (54). The TnI heptad is present in both TnI d57 and TnI 1-98 . If these repeats are involved in the TnT-TnI interaction as has been suggested (Ref. 54 and reviewed in Refs. 2, 5, and 36), this interaction would be present in troponin complexes containing TnI d57 and TnI 1-98 . This could explain why troponin ternary complexes containing TnI d57 or TnI 1-98 are able to activate the ATPase (Fig. 3A and Ref. 12). Finally, as neither TnC⅐TnT nor TnI 1-98 ⅐TnT complexes have been observed to activate on their own, a synergistic activation effect by TnI 1-98 -TnC is suggested. Consistent with this view is our observation that the affinity of TnT 157-263 for the thin filament is reduced by its association with TnI 1-98 -TnC (ϮCa ϩ2 ).
The intersubunit interactions in troponin can be divided into two classes (Fig. 7). The first class consists of interactions that are not significantly changed with the alterations of the Ca 2ϩ concentrations and are "structural" in nature. These core interactions, best represented by the last 50 residues of TnT (residues 216 -263), the NH 2 -terminal domain of TnI (residues 1-98), and the COOH-terminal domain of TnC, are responsible for maintaining the structure of the troponin complex and its interaction with the actin-tropomyosin filament in both the "on" and "off" states (Fig. 7). The second class of interactions are dependent on the presence of Ca 2ϩ and are responsible for the regulatory function of the troponin complex; the domains involved in these interactions switch their preferred interaction from one domain to another as a function of Ca 2ϩ binding to the regulatory domains of TnC. In the absence of Ca 2ϩ , the inhibitory/carboxyl region of TnI interacts strongly with the thin filament and weakly with TnC. In addition, the COOHterminal domain of TnT may be interacting with the thin filament blocking the activation function of the amino-terminal domain of TnT. In this situation, the ATPase activity is inhibited. In the presence of Ca 2ϩ , the inhibitory domain of TnI interacts strongly with the amino-terminal domain of TnC and weakly with the thin filament, thereby removing inhibition. The COOH-terminal region of TnT then interacts more strongly with TnI 1-98 ⅐TnC than with the thin filament, thereby liberating the activation function of amino-terminal domain of TnT (Fig. 7). This activation may be mediated by a change in TnT 1-191 -Tm interactions (Fig. 7) which would affect the equilibrium between two or more actin-Tm binding states (3,56). In the absence of Ca 2ϩ , the inhibitory region of TnI (present within C-TnI, amino acids 103-182) is strongly bound to the thin filament resulting in inhibition of the actomyosin ATPase. It is not clear whether the weak interaction between the NH 2 -terminal half of TnT 2 (residues 157-216) and the thin filament in the absence of Ca 2ϩ is involved in this inhibitory function. In the presence of Ca 2ϩ , C-TnI switches to a strong interaction with N-TnC and there is also a weak interaction (represented by thin black lines) between TnC and TnT. The precise regions of TnC or TnT involved in this interaction in the presence of Ca 2ϩ were not determined. The Ca 2ϩ -induced changes result in the removal of the inhibitory region of TnI and the region between residues 157-216 of TnT from the thin filament. This leads to activation of the actomyosin ATPase activity by the activation domain of TnT (located within residues . The attachment of the NH 2 -terminal half of TnI (N-TnI, residues 1-98) to residues 216 -263 of TnT and to TnC are necessary for the release of this activation function present within the amino-terminal domain of TnT.