NH2-terminal truncation of skeletal muscle troponin T does not alter the Ca2+ sensitivity of thin filament assembly.

To investigate how Ca2+ binding to troponin C regulates muscle contraction, the Ca2+-sensitive properties of thin filament assembly were studied as the tropomyosin binding, NH2-terminal region of troponin T was progressively shortened. Troponin complexes were prepared that contained skeletal muscle troponin C, troponin I, and either intact troponin T (TnT) (residues 1-259) or fragment TnT-(70-259), TnT-(151-259), or TnT-(159-259). In the absence of Ca2+ their respective affinities for pyrene-labeled tropomyosin were 2.3 × 107M−1, 1.2 × 107M−1, 1.9 × 105M−1, and 1.9 × 105M−1. Ca2+ had only a small effect on these affinities: 1.1 × 107M−1 for whole troponin, 2 × 105M−1 for troponin-(151-259), and 2.8 × 105M−1 for troponin-(159-259). Forms of troponin that bound weakly to tropomyosin in the absence of actin increased the actin affinity of tropomyosin only 2-3-fold, even in the absence of Ca2+; weak binding of troponin to tropomyosin correlated with weak effects on tropomyosin-actin binding. In contrast, whole troponin had an approximately 500-fold effect on tropomyosin binding to actin, regardless of whether Ca2+ was present. The small effect of Ca2+ on the energetics of thin filament assembly is not attributable to the amino-terminal region of troponin T. The results suggest that Ca2+ causes the interaction between actin and the globular region of troponin to switch between two energetically similar states.

The regulation of muscle contraction is accomplished by the reversible binding of Ca 2ϩ to the thin filament protein troponin. Regulation is proposed to involve a Ca 2ϩ -induced dissociation of one region of troponin from actin and from tropomyosin (Hitchcock, et al., 1973;Margossian and Cohen, 1973;Potter and Gergely, 1974;Pearlstone and Smillie, 1983;Tanokura and Ohtsuki, 1984;Ishii and Lehrer, 1991), thereby facilitating repositioning of tropomyosin on the actin filament (Lehman, et al., 1994). In particular, Ca 2ϩ weakens the actin binding of the troponin subunit, TnI, 1 and the tropomyosin binding of the carboxyl-terminal region of TnT. In this widely held scheme of thin filament function, troponin remains anchored to the thin filament by relatively Ca 2ϩ -insensitive binding of the aminoterminal region of TnT to the carboxyl-terminal region of tropomyosin (Pato, et al., 1981;Mak and Smillie, 1981;Tanokura, et al., 1983;Cho and Hitchcock-DeGregori, 1990;Ishii and Lehrer, 1991). Indeed, rotary-shadowed electron micrographs of troponin and of TnT (Flicker, et al., 1982) show a highly extended molecule with a narrow tail mostly attributable to the amino-terminal portion of TnT (White et al., 1987) and a distinctly more globular head region that is likely to include portions of all three subunits.
Consistent with the existence of a Ca 2ϩ -insensitive anchor, we recently reported that the affinity of troponin for actintropomyosin remains tight (Ͼ10 8 M Ϫ1 ) regardless of whether Ca 2ϩ is present (Dahiya, et al. 1994). Perhaps more surprisingly, the troponin-tropomyosin complex bound much more tightly to actin than did tropomyosin alone, even in the presence of Ca 2ϩ or in the absence of TnI. Parallel results were found for the binding of troponin, troponin-Ca 2ϩ , or TnT to actin-tropomyosin. These data would appear to complicate models of regulation in which troponin has two sites of interaction with the thin filament: a Ca 2ϩ -insensitive site primarily between tropomyosin and the elongated subunit TnT and a Ca 2ϩ -reversible site that primarily involves actin-TnI binding and tropomyosin-TnT binding. Rather, although Ca 2ϩ may cause conformational changes in troponin that result in substantially changed binding to the thin filament, the net energetics of this binding are not significantly altered by Ca 2ϩ .
This recent report was not the first to examine the effect of Ca 2ϩ on troponin binding to the thin filament. The new aspects included an attempt to exclude the contributions of cooperative effects in thin filament assembly by the application of a linear lattice model. Also, equilibrium linkage relationships were used to calculate the affinity of troponin (ϮCa 2ϩ ) for actintropomyosin. In general this affinity has been too tight to directly measure, with or without a linear lattice approach. Thus, the conclusions of this report are dependent upon two caveats, that the assumptions of the linear lattice model are valid for calculating and excluding cooperative aspects of assembly and that the described equilibrium linkage relationships are correct.
An alternative approach is to examine the properties of the Ca 2ϩ -sensitive region of troponin (TnI, TnC, and the carboxyl terminus of TnT), in the absence of the apparently more Ca 2ϩinsensitive tail region (i.e. the elongated amino-terminal portion of TnT). Because it lacks the Ca 2ϩ -insensitive, anchoring portion of TnT, the effect of this truncated troponin complex on thin filament assembly may be very Ca 2ϩ -sensitive. This can be tested by measuring the Ca 2ϩ dependence of the affinity constants involved in assembling thin filaments containing either troponin or truncated troponin. To accomplish this, troponin molecules were prepared containing intact TnT or one of several COOH-terminal fragments of TnT. Troponin binding to pyrene-modified tropomyosin was weaker when TnT was truncated but was no more sensitive to Ca 2ϩ . Whereas whole troponin promoted tropomyosin binding to actin about 500-fold, regardless of whether Ca 2ϩ was present, truncated troponin had a minimal effect on tropomyosin-actin binding, again regardless of whether Ca 2ϩ was present. These results have significance for how Ca 2ϩ regulates muscle contraction and for the participation of different regions of troponin in thin filament assembly and regulation.

EXPERIMENTAL PROCEDURES
Protein Preparation-Rabbit fast skeletal muscle actin, troponin, and troponin subunits were prepared as described previously . Troponin was also reconstituted, and the ternary complex was purified (Tobacman and Lee, 1987) from purified rabbit TnI, TnC, and rat fast skeletal muscle  or . These TnT fragments are numbered according to the rabbit sequence, and respectively include either Met-151 to the COOH terminus or Met-70 to the COOH terminus. They were purified from DE3 (BL21) cells transformed with the pET8c expression vector (Studier, et al., 1990) with the TnT cDNA inserted at the NcoI/BamHI site . Construction of the TnT70 -259 expression plasmid, expression of the protein, and the protein purification procedure were described previously  To insert the cDNA encoding TnT residues 151-259 into the pET3d vector, a polymerase chain reaction fragment was generated using a vector-sequence primer and a primer overlapping the codon for Met-151 but including an NcoI site at this position. The restricted fragment was inserted into pET8c at the NcoI/BamHI sites. The complete coding sequence was confirmed by dideoxynucleotide sequencing (Sanger et al., 1977). Another form of truncated troponin was obtained by controlled chymotryptic digestion of whole rabbit troponin followed by purification of the ternary complex of TnC, TnI, and TnT-(159 -259) (Morris and Lehrer, 1984;Hill et al., 1992). Cardiac tropomyosin was purified as described previously (Tobacman and Adelstein, 1986) and stoichiometrically labeled at Cys-190 either by carboxymethylation with [ 3 H]iodoacetic acid (Amersham Corp.)  or by reaction with N-(1-pyrene)iodoacetamide (Molecular Probes) (Morris and Lehrer, 1984;Dahiya et al., 1994). Bovine cardiac tropomyosin is approximately 90% the ␣ isoform, diminishing the heterogeneity that results when ␣/␤ skeletal muscle tropomyosin is denatured for labeling and then renatured to a mixture of ␣/␤, ␣/␣, and ␤/␤ forms. However, we cannot exclude the possibility that slightly different results would have been obtained using skeletal muscle (␣/␤) tropomyosin.
Binding Assays-Binding of troponin or truncated troponin to pyrene-tropomyosin was performed using an SLM 8000 spectrofluorometer. Water jacketed 2-ml samples were incubated in silenized cuvettes at 25°C in the presence of 10 mM Tris-HCl (pH 7.5), 3 mM MgCl 2 , 0.1 mM dithiothreitol, either 0.5 mM EGTA or 0.1 mM CaCl 2 , and either 60 mM or 300 mM KCl. Sequential aliquots of troponin (or truncated troponin) in the same buffer were added (maximum no more than 15% of initial volume) and the fluorescence intensity was monitored. Excitation was at 340 nm, and emission was at 405 nm. The effects of dilution on all concentrations were calculated. The best fit of each data set to simple 1:1 equilibrium binding was calculated as described previously (Dahiya, et al., 1994) using the curve-fitting program MINSQ (MicroMath).
The effects of troponin or truncated troponin on the binding of tropomyosin to actin were measured by cosedimentation of radiolabeled tropomyosin with actin . Conditions were as follows: 25°C, 10 mM Tris-HCl (pH 7.5), 3 mM MgCl 2 , 0.1 mM dithiothreitol, 3-5 M F-actin, 60 or 300 mM KCl, and either 0.5 mM EGTA or 0.1 mM CaCl 2 . The concentrations of troponin and tropomyosin varied for each experiment as indicated below. The actin-bound tropomyosin was calculated from the difference between the total and the supernatant 3 H-tropomyosin after sedimentation for 30 min at 35,000 rpm using a TLA100 rotor in the TL100 ultracentrifuge. (In the absence of actin, there was no sedimentation.) Troponin was added in a constant molar excess compared with the tropomyosin concentration, and the concentration of the troponin-tropomyosin complex was calculated as described (Dahiya et al., 1994), using the affinity constants determined under the same conditions in the fluorescence experiments below. The binding of troponin-tropomyosin to actin, as a function of the troponintropomyosin concentration, was analyzed as a linear lattice problem in which the ligand spans seven equivalent sites on the lattice (McGhee and Von Hippel, 1974;Tsuchiya and Szabo, 1982;Willadsen et al., 1992). Binding data are fit to obtain three parameters: the affinity of the ligand for an isolated site for the lattice, the -fold increase in affinity when binding involves cooperative interactions with one adjacent bound ligand, and the concentration of bound ligand at saturation. Fig. 1 demonstrates a saturable, concentration-dependent increase in fluorescence intensity when troponin was added to pyrene-tropomyosin in the presence of 60 mM KCl, 3 mM MgCl 2 . The curves are best fit binding isotherms for this representative data set and illustrate the effect of Ca 2ϩ concentration on troponin-tropomyosin affinity. A 1:1 troponin:tropomyosin binding stoichiometry was suggested by similar titrations (not shown) performed in the presence of much higher tropomyosin concentrations. Under conditions of low tropomyosin concentration, which facilitates measurement of tight binding, the averages of several experiments as shown in Fig. 1 imply that removal of Ca 2ϩ from troponin strengthens its affinity for tropomyosin about 2-fold, from 1.1 Ϯ 0.4 ϫ 10 7 M Ϫ1 in the presence of 50 mM CaCl 2 to 2.3 Ϯ 0.3 ϫ 10 7 M Ϫ1 in the presence of EGTA (Table I). This small, Ca 2ϩ -dependent change is consistent with similar results using cardiac troponin (Dahiya, et al., 1994), with a 1.7-fold increase in skeletal muscle troponin-tropomyosin photocross-linking (via TnT) in the absence of Ca 2ϩ (Chong and Hodges, 1982) and is in the direction suggested by the standard model for troponin function, referenced earlier. However, a 2-fold change is quite small for a major regulatory switch. Furthermore, Ca 2ϩ had no effect on the binding of skeletal muscle troponin to immobilized tropomyosin (Pearlstone and Smillie, 1983), no effect on troponin-tropomyosin affinity in previous fluorescence studies ( Morris and Lehrer, 1984;Ingraham and Swenson, 1985), and no effect on troponin-tropomyosin photocross-linking as detected by others (Tao et al. 1986). It seems evident that Ca 2ϩ has no more than a 2-fold effect on the binding affinity of whole troponin for tropomyosin. The remaining question is whether a Ca 2ϩ -insensitive interaction between troponin and the carboxyl-terminal region of tropomyosin is masking a larger and important effect of Ca 2ϩ on the interaction between troponin and tropomyosin near  Effect of TnT Truncation on Troponin Binding to Tropomyosin-As shown in Table I, removal of 69 amino acids from the amino terminus of TnT, producing a reconstituted troponin FIG. 1. Affinity of skeletal muscle troponin for tropomyosin in the presence and in the absence of Ca 2ϩ . Tropomyosin was labeled on Cys-190 with N-(1-pyrene)iodoacetamide, and the effect of troponin on the pyrene monomer fluorescence is shown in this representative experiment. ex ϭ 340 nm and em ϭ 405 nm. Conditions were as follows: 25°C, 3 mM MgCl 2 , 60 mM KCl, 5 mM Tris HCl (pH 7.5), 0.1 mM dithiothreitol, 0.1 mM pyrene-tropomyosin, and either 0.5 mM EGTA (q) or 0.1 mM CaCl 2 (E). The solid lines are best fit theoretical curves corresponding to affinities of 2.3 ϫ 10 7 M Ϫ1 (EGTA) or 6.9 ϫ 10 6 M Ϫ1 (CaCl 2 ). The abscissa shows the total added troponin concentration based upon the initial volume, but the theoretical curves are calculated with dilution (Ͻ6%) of both troponin and tropomyosin taken into consideration. The units of fluorescence intensity are arbitrary. designated troponin-(70 -259), weakened its affinity for tropomyosin by approximately one-half. This modest effect suggests that these residues make little contribution to troponin-tropomyosin binding. This is consistent with evidence that CB3 (comprised of troponin T residues 1-70) does not bind to tropomyosin (Pearlstone and Smillie, 1982;White et al., 1987). In fact, deletion of TnT residues 1-45 causes troponin binding to immobilized tropomyosin to be slightly stronger rather than weaker (Pan et al., 1991).

Affinity of Troponin for Tropomyosin-
On the other hand, the present data demonstrate much weaker tropomyosin binding by a further truncation of troponin, troponin-(151-259), which includes rabbit TnC, rabbit TnI, and the carboxyl-terminal portion of (recombinant rat) TnT. The affinity of troponin-(151-259) for tropomyosin was 2 orders of magnitude less than the affinity of whole troponin for tropomyosin (Table I, Fig. 2). Because the binding was so weak, it could not be determined whether the fluorescence transition was attributable to a 1:1 association between tropomyosin and troponin-(151-259). This weak binding (in comparison to whole, nontruncated troponin) was not due to the few amino acid differences between rat and rabbit TnT, nor to the presence versus the absence of reconstitution from subunits: indistinguishable results (Table I) were obtained using troponin-(159 -259), produced by controlled chymotryptic digestion of rabbit whole troponin, which selectively removes TnT residues 1-158 from the ternary troponin complex.
One possibility from these data is that the interaction between troponin and the region of tropomyosin near Cys-190 (Ohtsuki, 1979;Pearlstone and Smillie, 1982;Chong and Hodges, 1982;Morris and Lehrer, 1984) is quite weak when whole troponin is present. Several pertinent observations support this interpretation. Electron microscopic images of troponin-tropomyosin polymers show that the globular head region of troponin is often distinctly separated from the tropomyosin filament (Flicker et al., 1982). Co-crystals of tropomyosin and troponin show that troponin is most ordered near the carboxyl terminus of tropomyosin and less ordered near tropomyosin residues 190 -235 (White et al., 1987;White, 1988). Also, troponin has little effect on the local thermal unfolding of tropomyosin near Cys-190 (Ishii and Lehrer, 1991). Furthermore, although troponin causes prolongation of the fluorescence decay of IAEDANS attached to tropomyosin Cys-190, this prolonged component is only 20% of the total amplitude (Lamkin et al., 1983). Finally, the binding of the TnI⅐TnT-(159 -259) complex to immobilized tropomyosin is weakened substantially by the addition of TnC, even in the absence of Ca 2ϩ (Pearlstone and Smillie, 1983).
Comparisons among the progressively truncated forms of troponin (Table I) show that removal of TnT residues 70 -150 had a 60-fold effect on the binding constant, whereas removal of TnT residues 1-69 or 151-158 had little effect. This suggests that TnT residues 70 -150 include the major region within the TnT "tail" that interacts with tropomyosin. These data show that ternary troponin complexes behave similarly to isolated troponin T fragments. The CB1 fragment of TnT (residues 71-151) binds readily to immobilized tropomyosin (Pearlstone and Smillie, 1977) and also binds to glutaraldehyde-fixed Bailey crystals of tropomyosin (White et al., 1987).
Notably, there is no detectable Ca 2ϩ sensitivity to the relatively weak binding of tropomyosin to troponin-(159 -259) or to troponin-(151-259) ( Table I). In the absence of the extended amino-terminal region of TnT, which binds to tropomyosin in a largely Ca 2ϩ -insensitive manner, the binding of the remaining portion of troponin to tropomyosin is much reduced, yet it does not become more dependent upon the presence or absence of Ca 2ϩ . This does not prove that Ca 2ϩ has no effect on troponintropomyosin interactions. However, it does imply that the net energetics of this interaction are not significantly altered by Ca 2ϩ , even when the elongated tail region of TnT is absent. A possible explanation for this result is that, whether Ca 2ϩ is present or not, there is only a weak interaction between tropomyosin and this region of troponin.
Effect of Truncated Troponin on the Binding of Tropomyosin to Actin-Although truncated troponin binds weakly to tropomyosin, regardless of the Ca 2ϩ concentration, this interaction might be important when actin is also present. For whole cardiac troponin, we have recently shown there is only a 2-fold effect of Ca 2ϩ on troponin binding to actin-tropomyosin. To see if this effect is much larger for truncated troponin, which lacks the region believed to act as a Ca 2ϩ -insensitive anchor, we investigated the coincident binding of tropomyosin and truncated troponin to actin. SDS-polyacrylamide gel electrophoresis analysis of sedimentation data obtained under the same conditions as in Fig. 1 (not shown) indicated that the amount of truncated troponin that cosedimented with actin-tropomyosin did not saturate at a 1:1:7 ratio of truncated troponin:tropomyosin:actin. Instead, the amount of pelleted truncated tropo-  2. Tropomyosin binds weakly to troponin lacking the amino-terminal region of TnT. Troponin was reconstituted from skeletal muscle TnC, TnI, and a recombinant, carboxyl-terminal fragment of skeletal muscle TnT lacking 150 NH 2 -terminal residues. Binding to pyrene-tropomyosin was studied under the same conditions as in Fig. 1. Two representative curves are shown, either in the presence of EGTA (q) or in the presence of CaCl 2 (E). The fluorescence intensity increased as increasing concentrations of troponin-(151-259) were added. The solid lines are best fit theoretical curves, corresponding to affinities of 2.2 ϫ 10 5 M Ϫ1 (EGTA) and 1.8 ϫ 10 5 M Ϫ1 (CaCl 2 ). CaCl 2 had little effect on the affinity constant. Binding of troponin-(151-259) to tropomyosin was much weaker than binding of whole troponin to tropomyosin, but in neither case did CaCl 2 significantly weaken the affinity (see also Table  I). The units of fluorescence intensity are arbitrary. nin was severalfold higher than the pelleted tropomyosin. This result implied that the truncated troponin was not binding to the thin filaments with the same stoichiometry as whole troponin and is consistent with the weak tropomyosin-truncated troponin demonstrated in Fig. 2. Furthermore (Fig. 3A), truncated troponin had little effect on the binding of tropomyosin to actin, even in the absence of Ca 2ϩ . The largest effect seen in the experiment shown in Fig. 3A  Because of the weak affinity of troponin-(159 -259) for tropomyosin (Table I), it was necessary to determine whether the results in Fig. 3A were attributable to inadequate concentrations of truncated troponin. Fig. 3B shows that this was not the case, since little change was seen at higher concentrations of truncated troponin. Fig. 3, A and B, shows a trend for truncated troponin to have a greater effect on tropomyosin-actin binding in the absence of Ca 2ϩ . Since the free actin concentration is nearly constant, one can estimate the effect of Ca 2ϩ on the affinity of tropomyosin-truncated troponin for actin from the Ca 2ϩ -induced change in the ratio of bound to free tropomyosin in the presence of 6 M troponin-(151-259) or troponin-(159 -259). (This calculation uses the last data point of each curve in Fig. 3B.) Removal of Ca 2ϩ increased this ratio by a factor of three in each case, suggesting a 3-fold effect of Ca 2ϩ on tropomyosin-truncated troponin binding to actin.
Equilibrium Linkage Analysis of Troponin and Tropomyosin Binding to Actin-The relevant comparisons for the effects of truncated troponin in Fig. 3 are the analagous effects of whole skeletal muscle troponin. More specifically, we sought to compare the effects of whole troponin and truncated troponin on thin filament assembly and to compare the effects of Ca 2ϩ in each case. However, the properties of the two forms of troponin could not be studied under identical conditions. Truncated troponin has little interaction with actin or tropomyosin unless the ionic strength is relatively low (e.g. 60 mM KCl as in Fig. 2). The effects of whole troponin on thin filament assembly, on the other hand, require experimental conditions of higher ionic strength (300 mM KCl) to prevent troponin-tropomyosin polymerization . Fig. 4 shows a representative experiment to determine the affinity of troponin for pyrene-tropomyosin under these higher ionic strength conditions. In comparison to these data with relatively weak binding, the association constant is 25-30-fold higher in the presence of 60 mM KCl (summarized in Table I). The effect of Ca 2ϩ is similar at both KCl concentrations. Fig. 5 shows a representative experiment of troponin-tropomyosin binding to actin. Troponin was added in a constant molar excess of the tropomyosin concentration. The tropomyosin was radioactively tagged, permitting measurement of its FIG. 3. Effect of truncated troponin on binding to tropomyosin to actin. The binding of 3 H-tropomyosin to 3.5 mM F-actin was studied by cosedimentation using a tabletop ultracentrifuge, and conditions were as in Fig. 1. A, the concentrations of tropomyosin and troponin-(159 -259) were varied in parallel so that the total concentration of truncated troponin was always 2 mM in excess of the total tropomyosin concentration. Open symbols, EGTA; filled symbols, CaCl 2 ; circles, tropomyosin alone; triangles, tropomyosin plus troponin-(159 -259). In the absence of truncated troponin, there is no effect of CaCl 2 on the free (i.e. non-actin-bound) tropomyosin concentration required for half saturation of the actin. Truncated troponin had a small effect on the apparent affinity, most notably in the absence of CaCl 2 , when the apparent K d was shifted to about 0.1 mM, as compared with 0.2-0.3 mM for tropomyosin alone. B, increasing concentrations of truncated troponin, either troponin-(159 -259) (diamonds) or troponin-(151-259) (squares), were added to samples containing 0.3 mM 3 H-tropomyosin and 5 mM F-actin. A concentration of approximately 2 mM truncated troponin was sufficient to approach saturation of its effect on tropomyosin-actin binding, regardless of whether CaCl 2 (filled symbols) was present. This is roughly 3-fold less than would be predicted from the data in Table I, perhaps because of a difference between pyrene-tropomyosin and the carboxymethylated 3 H-tropomyosin.
FIG. 4. Binding of skeletal muscle troponin to tropomyosin in the presence of high ionic strength. Increasing concentrations of rabbit fast skeletal muscle troponin were added to 0.1 mM pyrenetropomyosin, and a representative experiment demonstrating the resultant fluorescence change is shown. Conditions were as in Fig. 1, except the KCl concentration was 300 mM instead of 60 mM. The solid lines correspond to best fit values of troponin-tropomyosin binding constants of 8.6 ϫ 10 5 M Ϫ1 (Ⅺ; EGTA) and 4.2 ϫ 10 5 M Ϫ1 (*; CaCl 2 ). See Table I for average values of several such determinations. Fluorescence intensity units are arbitrary. free and actin-bound concentrations. Comparing the filled and open symbols in Fig. 5, it can be seen that removal of Ca 2ϩ strengthens binding of troponin-tropomyosin to actin about 2-fold, similar to the results with truncated troponin (Fig. 3). (The Fig. 3 results can only be analyzed qualitatively because of uncertainty in the stoichiometry of truncated troponin binding to tropomyosin and to the thin filament.) The analysis in Fig. 5 differs from previous results using skeletal muscle troponin , because the concentration of the troponin-tropomyosin complex is calculated based upon the affinity constants in Table I, rather than assumed to be the same as the non-actin-bound tropomyosin concentration. (We recently published a similar analysis for cardiac troponin (Dahiya et al., 1994) but now present skeletal muscle troponin data so that it may be related to the truncated skeletal muscle troponin data.) The advantage of this analysis is that it permits calculation of the affinity of troponin for actin-tropomyosin, using the thermodynamic linkage relationships schematically shown in Fig.  6. It should be noted that there is a potential error in this calculation, because pyrene-tropomyosin may have different properties than 3 H-tropomyosin. Fluorescence competition studies (Dahiya et al., 1994) suggest there are no major differences between them, at least at low ionic strength.
Significantly, whereas truncated troponin has a weak, difficult to quantify effect on the binding of tropomyosin to actin, whole skeletal muscle troponin has a very large effect (Fig. 6). The figure shows obligate equilibrium linkage relationships in the associations of tropomyosin and troponin for an isolated site on F-actin. Note that the actin binding constants do not depend upon cooperative effects (this is required for this scheme to be valid and is one of the measurements that results from the linear lattice analysis of actin binding data) and therefore differ from apparent binding constants (1/K D ) that partially depend upon cooperative binding of tropomyosin or troponin-tropomyosin to actin. The average affinity of troponintropomyosin for an isolated site on F-actin (determined from several experiments similar to those in Fig. 5) is 3.4 Ϯ 1.0 ϫ 10 5 M Ϫ1 in the absence of Ca 2ϩ and 2.6 Ϯ 0.8 ϫ 10 5 M Ϫ1 in the presence of Ca 2ϩ . Regardless of whether Ca 2ϩ is present, troponin binding to tropomyosin causes its affinity for actin to increase about 500-fold. This is very different from the small effects of truncated troponin on tropomyosin-actin binding seen in Fig. 3, despite concentrations of truncated troponin that would be expected to produce, based upon Table I and in the absence of actin, significant formation of the tropomyosin-truncated troponin complex. Fig. 6 also shows the affinity of troponin for an isolated actin-tropomyosin site, as indirectly calculated from the equilibrium linkage relationships. This affinity is too tight to measure directly, whether Ca 2ϩ is present (2 ϫ 10 8 M Ϫ1 ) or absent (4 ϫ 10 8 M Ϫ1 ).

DISCUSSION
The data in this report suggest the following. 1) Troponin greatly promotes the binding of individual tropomyosin molecules to actin, regardless of the Ca 2ϩ concentration. This effect depends upon the presence of TnT residues 1-150. More narrowly, the important region for this effect consists of residues 70 -150, since troponin-(70 -259) promotes tropomyosin-actin binding almost as well as does whole troponin .
2) The amino-terminal region of TnT, especially residues 70 -150, is required for tight association of troponin to tropomyosin. These results are consistent with observations involving TnT fragments, as opposed to the ternary troponin complexes in the present work. Also, Morris and Lehrer (1984) found that troponin-(159 -259) bound more weakly to tropomyosin than did intact troponin. In a separate study of TnT fragments, Ishii and Lehrer (1991) concluded that it was the NH 2 -terminal rather than the COOH-terminal region of TnT that binds most FIG. 5. Binding of the troponin-tropomyosin complex to actin, analyzed by the linear lattice binding equation. Experimental conditions were the same as in in Fig. 4. The concentrations of troponin and tropomyosin were varied in parallel so that the concentration of total troponin was always in excess of the total tropomyosin concentration. Two representative data sets are shown, either in the presence of EGTA (E) or in the presence of CaCl 2 (q). The concentration of free troponin-tropomyosin was calculated from the measured, non-actinbound concentration of 3 H-tropomyosin and the binding constants in Table I. The theoretical curves correspond to K 0 ϭ 1.8 ϫ 10 5 M Ϫ1 and y ϭ 29 in the presence of CaCl 2 and to K 0 ϭ 2.2 ϫ 10 5 M Ϫ1 and y ϭ 44 in the presence of EGTA. K 0 is the affinity of troponin-tropomyosin (or troponin-tropomyosin-Ca 2ϩ ) for an isolated site on F-actin in the absence of any interaction with other troponin-tropomyosin complexes. y is a cooperativity parameter, and the product K 0 y is approximately equal to the apparent overall binding constant.
FIG. 6. Equilibrium linkage analysis of troponin and tropomyosin binding to actin. The figure shows association constants, in units of mM Ϫ1 for several steps involved in thin filament assembly. Conditions are as in Figs. 4 and 5,i.e. in the presence of 300 mM KCl. These results are for binding events at an isolated site on the actin filament, a small portion of which is schematically shown. Analysis of isolated site binding is essential for the equilibrium linkage scheme to be valid. Troponin binds tightly (K Ͼ 10 8 M Ϫ1 ) to actin-tropomyosin, regardless of whether Ca 2ϩ is present. Similarly, troponin has a profound effect on tropomyosin binding to actin, increasing its association constant about 500-fold, regardless of whether Ca 2ϩ is present. This is in contrast to the 2-or 3-fold apparent effect of truncated troponin (Fig. 3). Replacement of Ca 2ϩ by Mg 2ϩ causes an approximately 2-fold increase in the affinity of troponin either for tropomyosin or for actin-tropomyosin. The affinity of tropomyosin for an isolated site on F-actin, 600 M Ϫ1 , is from Hill et al. (1992). The affinity of troponin for actin-tropomyosin (4 ϫ 10 8 and 2 ϫ 10 8 M Ϫ1 in the absence and presence of Ca 2ϩ , respectively) was not directly measured but was calculated from the linkage relationships to the other values in the figure. tightly to actin. Our results more narrowly suggest the importance of the TnT-(70 -150) region but do not by themselves allow any conclusion comparing this region with the troponin-(159 -259) region. 3) There is a direct relationship between the strength of troponin binding to tropomyosin (as the aminoterminal region of TnT is progressively deleted or included), and the ability of troponin to enhance tropomyosin-actin binding. Binding of troponin-(151-259) or troponin-(159 -259) to tropomyosin had only a small effect on the affinity of the resultant complex for actin, regardless of the Ca 2ϩ concentration. One possible explanation is that, unlike whole troponin, these truncated troponins do not interact simultaneously with both actin and tropomyosin; troponin-(159 -259) is primarily an actin-binding portion of troponin and TnT-(1-150) is the primary tropomyosin-binding portion. Alternatively, it is possible that the weak effect of truncated troponin on thin filament assembly occurs because deletion of TnT residues 1-150 has a detrimental effect on the stability and function of the remainder of the troponin complex.
Ca 2ϩ binding to the thin filament results in a structural change generally considered to include a repositioning of tropomyosin on actin (Haselgrove, 1972;Huxley, 1972;Parry and Squire, 1973;Lehman et al., 1994). This event rapidly follows Ca 2ϩ release, even in the absence of cross-bridge binding (Kress et al., 1986). Although cross-bridges also affect thin filament conformation and cooperative cross-bridge effects may be important in muscle activation (Lehrer, 1994), it is clear that Ca 2ϩ alone has a major effect on thin filament structure. Despite progress on the Ca 2ϩ -induced conformational change in TnC at high resolution (Herzberg et al., 1986;Gagnè et al., 1994), the absence of an atomic strucure for whole troponin has hampered determination of how this change in TnC is transmitted to the remainder of the thin filament. At present, the best supported idea (reviewed in Leavis and Gergely, 1984;Zot and Potter, 1987;Chalovich, 1992) is that TnI interacts with actin in the presence of low Ca 2ϩ concentrations and, because TnI is part of a ternary troponin complex that is bound to tropomyosin, the position of tropomyosin on the actin filament is thereby constrained. Ca 2ϩ binding to TnC releases this constraint by facilitating the interaction between TnI and TnC, which disrupts the TnI-actin binding. Indeed, fluorescence resonance energy transfer data support a Ca 2ϩ -induced movement of TnI relative to actin within the thin filament (Tao et al., 1990). Furthermore, a specific TnI peptide has been shown to interact with either Ca 2ϩ -TnC or with actin and to be inhibitory of actin-myosin interactions (Syska et al., 1976).
It is worthwhile to consider how the present work relates to this model for regulation. Simply put, the small effect of Ca 2ϩ summarized in Fig. 6 is inconsistent with the idea that the action of Ca 2ϩ can be understood as a release of TnI from actin. This small energetic change in troponin-thin filament binding is not plausibly consistent with disruption of an actin-TnI protein-protein interface. Rather, a Ca 2ϩ -induced release of TnI-actin binding can only be occurring if some other interactions(s) between troponin and the thin filament are strengthened by Ca 2ϩ . A recent publication (Dahiya et al., 1994) reached a similar conclusion regarding cardiac thin filament assembly and suggested TnT as one possibility for a Ca 2ϩstrengthened interaction with the thin filament, perhaps via the direct TnT-actin binding that was demonstrated by Heeley and Smillie (1988). Also, it was shown that an increased cooperative interaction between adjacent cardiac troponin-tropomyosin complexes did not occur on Ca 2ϩ binding, so one should look primarily within each troponin-tropomyosin-7 actin region for the thermodynamic changes accompanying Ca 2ϩ binding to TnC. The present work shows similar behavior for skeletal muscle troponin, which is the protein used for most of the data supporting the model of TnI-mediated regulation. More significantly, the present data suggest that the small effect of Ca 2ϩ on thin filament assembly (Fig. 6) is probably not attributable to compensatory effects involving the amino-terminal region of TnT. Removal of this region weakened troponin's interaction with tropomyosin, but did not increase the Ca 2ϩ sensitivity of thin filament assembly.
The above arguments suggest that, within the intact thin filament, Ca 2ϩ has little effect on the energetics of the interaction between actin-tropomyosin and the TnC ⅐ TnI ⅐ TnT-(159 -259) region of troponin. However, it is hard to imagine how regulation could be accomplished without Ca 2ϩ causing major alterations in the interface between this region of troponin and actin-tropomyosin. It is likely that many specific interactions are altered, even if the net energetics are little changed, and that different actin and troponin residues are involved in these interactions in the presence as opposed to the absence of Ca 2ϩ . The identification of these residues by mutagenesis of actin and troponin may prove a useful avenue for future studies.