Induction of Nonpolymerizable Tropomyosin Binding to F-actin by Troponin and Its Components*

in which 11 resi- dues have been quantitatively cleaved from the COOH terminus of muscle tropomyosin (TM) by enzymic digestion, does not bind to F-actin. Binding is restored in the presence of troponin (Tn) and absence of Ca2+. The binding is stronger than for intact TM alone and shows residual cooperativity. In the presence of Ca2+, the binding is at least 10-fold weaker and cooperativity is not observed. Tn-T alone is more effective than Tn-I alone in inducing nonpolymerizable TM binding. Tn-T plus Tn-I induce binding as effectively as whole Tn (without Ca2+). In the absence of Ca2+, Tn-T + Tn-C and Tn-I + Tn-C are more effective in promoting binding than in the presence of Ca2+. These observations emphasize the importance of the head to tail overlap region of TM in the cooperative interactions of the thin filament assembly. The effects of Ca2+ are largely understandable in terms of its known effects on the strength of interactions between Tn-I and TM + actin and between Tn-T and TM. The residual cooperativity observed in nonpolymerizable TM binding in the presence of Tn (without Ca2+) may indicate that the T1 fragment region (residues 1-158) of Tn-T spans the head to tail overlap gap between the neighboring nonpolymerizable

Nonpolymerizable tropomyosin, in which 11 residues have been quantitatively cleaved from the COOH terminus of muscle tropomyosin (TM) by enzymic digestion, does not bind to F-actin. Binding is restored in the presence of troponin (Tn) and absence of Ca2+. The binding is stronger than for intact TM alone and shows residual cooperativity. In the presence of Ca2+, the binding is at least 10-fold weaker and cooperativity is not observed. Tn-T alone is more effective than Tn-I alone in inducing nonpolymerizable TM binding. Tn-T plus Tn-I induce binding as effectively as whole Tn (without Ca2+). In the absence of Ca2+, Tn-T + Tn-C and Tn-I + Tn-C are more effective in promoting binding than in the presence of Ca2+. These observations emphasize the importance of the head to tail overlap region of TM in the cooperative interactions of the thin filament assembly. The effects of Ca2+ are largely understandable in terms of its known effects on the strength of interactions between Tn-I and TM + actin and between Tn-T and TM. The residual cooperativity observed in nonpolymerizable TM binding in the presence of Tn (without Ca2+) may indicate that the T1 fragment region (residues 1-158) of Tn-T spans the head to tail overlap gap between the neighboring nonpolymerizable TM molecules. Alternatively, or in addition, the cooperativity may arise from conformational changes transmitted through actin from one nonpolymerizable TM-Tn binding site to others.
Tropomyosin together with the troponin complex, troponins I, T, and C, constitute the calcium-dependent regulatory elements of the contractile mechanism in striated muscle tissues (1,2). Each TM' molecule interacts with seven (or 14) actin monomers on one (or both) of the two strands of Factin and is in contact with neighboring T M molecules through a head to tail overlap of 8-9 amino acid residues. The T n complex, located a t regular intervals corresponding to the axial length of the TM molecule along the thin filament assembly, is linked to TM and actin through its Tn-T and Tn-I components, both of which are in intimate contact with each other and with Tn-C (for reviews, see Refs. 3 and 4). Recent evidence strongly supports the view that, while Tn-C and Tn-I are only moderately asymmetric (5,6), Tn-T is a highly extended molecule which interacts with T M over an extensive region of the COOH-terminal half of its structure. Thus, the fragments CB1 and T1 of Tn-T (residues 1-151 and 1-158, respectively), which promote head to tail aggregation of TM and its fragments (7,8), are believed to interact close to or at the COOH-terminal end of TM, possibly involving the head to tail overlap region (8)(9)(10)(11). Fragment T 2 of Tn-T (residues 159-259) has been shown to interact in the region of cysteine 190 of TM, a distance of some 14 nm from its COOH-terminal end (12-16).
Available evidence (13,14) strongly suggests that Tn-C and Tn-I are also located in the region of cysteine 190 of TM, a conclusion consistent with the observation that these components interact with Tn-T fragment T2 but not T1 (14,(17)(18)(19). Strong support for this asymmetric nature of T n on the muscle thin filament has recently been obtained by an electron microscopic study of rotary shadowed T n complex and Tn-T (20). The entire complex has both a globular (mostly Tn-I and Tn-C) and a rod-like dom+n with the tail (mostly Tn-T) having a length of 160 f 35 A.
An alternative view, that Tn is a more compact structure and interacts with T M over a more restricted region, is held by Nagano et al. (21), Ohtsuki (22), Nagano and Ohtsuki (23), and Ohtsuki and Nagano (24). However, their conclusions are based in part on predictive model building and neglect the accumulating chemical and structural evidence for the involvement of the COOH-terminal region of T M in its interaction with the T1 fragment region of Tn-T.
While the mechanism by which this assembly controls the contractile events is not fully understood, it is clear that the association of these thin filament proteins is strongly interdependent. Although T M interacts stoichiometrically (molar ratio 1:7) with F-actin under optimal salt conditions, nonpolymerizable TM, prepared by the removal of 11 amino acid residues from its COOH-terminal end, fails to bind (25). This is consistent with the observation that the binding curve of TM to F-actin is highly cooperative (26) and that the low binding constant for an isolated T M molecule (-lo3 M-') is dramatically increased by the interaction of contiguous T M molecules linked through their head to tail contacts (27,28). Under conditions (low free [Mg2+]) in which rabbit skeletal TM and platelet T M fail to bind to F-actin, binding can be induced by Tn-I and the S-1 fragment of myosin (29-31).
The interactions among the thin filament proteins are markedly dependent on Ca2+ concentration. Thus, the binding of Ca2+ to Tn-C leads to a strengthening of the interactions between Tn-C and both Tn-I and the T2 fragment region of Tn-T (17, [32][33][34][35], to a change in the interactions between Nonpolymerizable Tropomyosin Binding to F-actin Tn-I and Tn-T (36), and to a weakening of the interaction between Tn-I and TM-F-actin (37)(38)(39) and between the T2 fragment of Tn-T and TM (10,11). The association between the T1 fragment of Tn-T and the COOH-terminal region of T M is relatively independent of Ca2+ but may be affected by transmitted conformational changes induced by changes in the interaction of Tn-I and the fragment T2 region with T M in the region of cysteine 190 (11).
T o provide further information on the structural and functional relationship of this complex system and, in particular, of the role of the head to tail overlap region of T M molecules, we have previously described the preparation of nonpolymerizable T M and reported its inability to bind to F-actin and the CB1 and T1 fragments (residues 1-151 and 1-158, respectively) of Tn-T (9,25). In the present investigation, we report the effectiveness of Tn and of the Tn components, alone or in combination, in the presence and absence of Caz+ for the induction of nonpolymerizable T M binding to F-actin.

MATERIALS AND METHODS
Preparation of TM, Nonpolymerizable TM and Actin-Rabbit a-TM was from cardiac tissue and is identical with that from skeletal muscle (40). Nonpolymerizable TM was prepared as previously described (25). For some experiments, the nonpolymerizable TM was further purified by treatment with F-actin, which binds intact TM but not nonpolymerizable TM, using the same conditions as described below for the actin binding experiments. The TM-F-actin complex was pelleted a t 97,000 X g for 90 min. The nonpolymerizable TM in the supernatant was purified from actin by chromatography on CM32 using the same conditions as described previously for the fractionation of a-and 0-TMs (41). No differences were observed between the two nonpolymerizable TM preparations in the experiments described in this paper. Actin was extracted and purified from acetone powders of rabbit skeletal muscle (42). The G-actin was stored at 4 "C and used within 2 weeks of preparation.
Preparation of Tn and Its Components-The Tn complex was prepared from rabbit skeletal muscle as described (43). The method of Hartshorne and Mueller (44) was modified to fractionate the T n into Tn-C and a mixture of Tn-I and Tn-T. To 1 g of Tn complex dissolved in 1 liter of 0.05 M HCl, 10 mM P-mercaptoethanol at room temperature, 120 ml of 10% perchloric acid a t room temperature were added slowly. After gentle stirring for 30 min, Tn-C was collected by centrifugation a t room temperature at 13,700 X g for 15 min. The Tn-C pellet was dissolved in 5 mM p-mercaptoethanol, adjusted to pH 7.0, dialyzed exhaustively at 4 'C against 2 mM P-mercaptoethanol, and lyophilized. It was further purified on a DEAE-Sephadex A-25 column as described by Byers and Kay (45). Tn-I and Tn-T in the supernatant were collected by 70% (NH&S04 saturation at pH 7.0 and centrifugation at 13,700 X g for 10 min. The Tn-I and Tn-T were separated on a CM-cellulose column as described by Wilkinson (46).
Binding of TM and Nonpolymerizable TM to F-actin-The preparative ultracentrifugation method of Eaton et al. (29) was used. TM and nonpolymerizable TM were labeled with ' ' ' 1 (New England Nuclear) as reported previously (9). Stock solutions of Tn (-15 p M ) , Tn-I (10-50 p~) Tn-T ( " 8 p~) , and Tn-C (6-26 p M ) in binding buffer (3 mM imidazole, 0.15 M KC1, 5.5 mM MgClz, 0.01% NaN3, 1 mM dithiothreitol, pH 7.0, and either 0.1 mM CaC12 or 1 mM EGTA were stored at 4 "C. Because of the limited solubility of Tn-I and Tn-T in the binding buffer, these proteins were initially dissolved in 1% formic acid and dialyzed exhaustively against binding buffer. Normally, concentrations up to 1.0 and 0.3 mg/ml could be obtained for Tn-I and Tn-T, respectively. G-actin was polymerized by dialyzing against the appropriate binding buffer (0.1 mM CaC12 or 1 mM EGTA) with the inclusion of 1 mM ATP at 4 "C for 16 h prior to the binding experiments. All stock solutions with the exception of F-actin were centrifuged at 97,000 X g at 4 "C for 60 min to remove any protein aggregates before the actin binding experiments. For stock solutions of Tn-I, nonpolymerizable TM, and TM, aliquots of 1 M dithiothreitol were added to make a final concentration of 1 mM just before this step. The final concentration of F-actin was kept constant a t 7 p~. For studies of the cooperativity of binding of nonpolymerizable TM or TM to F-actin, the concentration of nonpolymerizable TM or TM was varied from 0 to 5 p~ while the T n complex concentration was kept constant at 2 p~ such that the molar ratio of Tn to actin was 0.28. For studying the effects of the T n complex and its components on the binding of nonpolymerizable TM to F-actin, the concentration of nonpolymerizable TM was kept constant. at 2 p~ and the concentration of T n or its components was varied. All protein concentrations were determined by amino acid analyses of the stock solutions.

RESULTS
Binding of Intact TM and Nonpolymerizable TM to Factin-The results of binding experiments in which the actin concentration was maintained at 7 PM and the TM or nonpolymerizable T M concentration was increased from 0 to 3 PM or higher are shown in Fig. 1A. The sigmoidal binding curve of T M with actin indicates a highly cooperative process. This is also demonstrated clearly in the curved Scatchard plot in Fig. 1B. The apparent constant for the binding of TM to actin, K , expressed as the reciprocal of the free TM at halfsaturation of the actin filaments, was 1.7 X lo6 "'. These results are in good agreement with similar experiments previously carried out by Yang et al. (26).
When the interaction of nonpolymerizable T M with Factin was examined under identical conditions, negligible binding was observed (Fig. U). This observation is consistent with the data of Wegner (27) and Walsh and Wegner (28) who from light scattering measurements calculated that the equilibrium constant of binding for an isolated T M molecule on F-actin was -lo3 M-'. This binding was increased -10"- 105-fold when the interaction of contiguous T M molecules linked head to tail was considered. Since nonpolymerizable T M lacks the capacity to form head to tail contacts (25), its interaction with actin is equivalent to that of a single isolated T M molecule and with the protein concentrations (1-10 @M) used in these experiments, negligible binding is expected.
Since the concentration of Mg2+ has been shown to significantly affect the interaction of T M with actin (26,27,31), the Mg2+ concentration was varied from 2 to 8 mM under otherwise identical conditions. No effect on the binding of nonpolymerizable TM to actin was observed (data not shown), indicating that the enhancing effects of Mg2+ in promoting binding are not adequate to overcome the absence of cooperativity resulting from the lack of head to tail interaction in nonpolymerizable TM. binding of nonpolymerizable TM to F-actin, the actin and nonpolymerizable T M concentrations were kept constant at 7 and 2 p~, respectively, while the concentration of T n was varied from 0 to 7 p~. As shown in Fig. 2, T n ( -ea2+) restored the stoichiometric binding of nonpolymerizable TM to actin when the molar ratio of Tn to actin was in excess of 0.3. On the other hand, T n (+Ca2+) only partially restored the binding even at higher Tn to actin ratios.

Effects of Tn-T, Tn-I, and Tn-T + Tn-I on Binding of
Nonpolymerizable TM to F-actin-As shown in Fig. 3, Tn-T alone can substantially enhance the binding of nonpolymerizable TM to actin, reaching a value of 85% saturation at a molar ratio of Tn-T to actin of 0.8. Tn-I, on the other hand, increased the affinity of nonpolymerizable TM and actin to a lesser extent, reaching a level of about 50% saturation at high molar ratios of Tn-I to actin. Tn-C alone had no effects on the binding of nonpolymerizable TM to actin (data not shown). When the effects of Tn-T + Tn-I were tested, full induction of nonpolymerizable T M binding to actin was observed and the binding curve was similar to that observed with whole T n (-Ca"); compare Figs. 2 and 3. These results show that the effects of T n (-Ca") on the binding of nonpolymerizable TM to actin are mediated largely through Tn-T but that Tn-I also participates in this phenomenon and potentiates the effect of Tn-T. The presence of Tn-C in the absence of Ca2+ would appear to have little effect.
Effect of Tn-T + Tn-C (kCa") and Tn-I + Tn-C (kea'+) on the Binding of Nonpolymerizable T M to Actin-As shown in Fig. 4, Tn-C (-Ca") has little effect on the binding induced by Tn-T of nonpolymerizable TM to actin (compare Figs. 3  and 4). In the presence of Ca2+, however, Tn-C significantly reduces the effect of Tn-T. The presence of Tn-C on the binding induced by Tn-I significantly reduces the latter's effect (compare Figs. 3 and 4) to a level where an effect of Ca2+ on the induction was not experimentally significant.

DISCUSSION
In the present work, a preparation of nonpolymerizable T M has been employed which differs from intact T M in that residues 274-284 have been quantitatively removed by treatment with carboxypeptidase A (9,25). Previous studies (25) have demonstrated that this preparation has completely lost its ability to polymerize in a head to tail manner and has a M , of 66,000 in solution even a t low ionic strength. It retains, however, its coiled coil structure with only minimal changes in its stability properties (25). Interestingly, this preparation of nonpolymerizable TM no longer binds to F-actin under conditions known to be optimal for the binding of intact TM, demonstrating the importance of the head to tail interaction of contiguous T M molecules in the cooperative interaction of T M with F-actin. These observations are in excellent agreement with the studies of Wegner (27) and Walsh and Wegner (28) which indicate that the equilibrium constant for the binding of TM to singly contiguous sites on the actin filaments is 600-1200 times greater, depending on the Mg2+ concentration, than that for the binding of TM to isolated sites. In addition, it has been observed that platelet TM, which displays a significantly weaker tendency to aggregate in a head to tail manner, also binds more weakly to F-actin relative to intact skeletal T M under similar experimental conditions. In the case of the platelet TM, however, the binding to F-actin can be induced to stoichiometric levels by increasing the Mg2+ concentration to 8-10 mM (31). In the present work with nonpolymerizable TM, this was not observed. We conclude that the effects of increasing Mg2+ concentration in inducing the binding of intact skeletal TM and platelet TM to F-actin are inadequate to promote such binding in the complete absence of head to tail polymerization, as is the case with nonpolymerizable TM.
While nonpolymerizable T M by itself fails to bind to Factin, the present studies demonstrate that the addition of T n complex to the system in the absence of Ca'+ ions restores stoichiometric binding. The apparent constant for this binding is in fact almost 2-fold greater than for intact muscle T M alone under the same conditions. Interestingly, the binding curve and Scatchard plot demonstrate a significant degree of residual cooperativity in this interaction, In the presence of Ca2+, the Tn complex is much less effective in inducing nonpolymerizable T M binding to actin, with an apparent binding constant estimated to be at least lo-fold lower than in the absence of Ca2+. No cooperativity in this binding could be detected in the present experiments.
TO further investigate the molecular basis of these obser-vations with whole Tn, we examined the effects of individual Tn components either alone or in combination for their ability to induce nonpolymerizable T M binding to F-actin. Tn-T was most effective, while Tn-I produced a significant but lesser induction of binding (see Fig. 3). The addition of both Tn-T and Tn-I induced nonpolymerizable T M binding to a level equivalent to that observed with whole T n in the absence of Ca2+. Thus, Tn-C would appear to have little effect on the binding when Ca'+ is absent. While the induction of binding by Tn-I is understandable since it is known to interact with both TM and actin, that of Tn-T is less obvious since it interacts only with TM and reportedly not with actin (38, 39). The interaction of Tn-T with intact T M is now known to involve two regions on each of these two proteins. Thus, fragment T2 (residues 159-259) of Tn-T binds to TM in the region of cysteine 190, about one-third of the molecular distance from its COOH-terminal end (12-16). With Ca'+, this interaction is disrupted in the presence of Tn-I and Tn-C to which this fragment also binds (10, 11). On the other hand, fragment T1 (residues 1-158) of Tn-T binds close to or at the COOH-terminal end of the TM molecule, possibly involving the head to tail overlap region (9). This interaction is insensitive to the binding of Ca2+ to Tn-C, although it may be affected by conformational changes transmitted from the fragment T2 binding region through the T M molecule (11). In the case of nonpolymerizable TM, this interaction with the T1 (or CB1) region (residues 1-158 (or 1-151)) has been shown to be significantly weakened but perhaps not eliminated (9,10). Thus, the significant effects of Tn-T on the induction of binding of nonpolymerizable TM to actin may be explicable in terms of a bridging of the gap between adjacent nonpolymerizable T M molecules created by the removal of the COOH-terminal 11 residues to produce nonpolymerizable TM. Such an interpretation would be consistent with the residual cooperativity of binding of nonpolymerizable T M observed in the presence of whole T n in the absence of Ca2+ (see Fig. 1, A and B).
The effects of the addition of Ca2+ in reducing the induction of binding of nonpolymerizable T M by Tn-T + Tn-C, by Tn-I + Tn-C, and by whole Tn are not unexpected, since the interactions between Tn-I and F-actin + TM and between the T2 fragment region of Tn-T and TM have been shown to be weakened by Ca'+ in a system also containing Tn-C (11). In the case of the strength of interaction of fragment T1 of Tn-T with the head to tail overlap region of TM, no direct effect of Ca2+ is anticipated, nor has one been observed since this segment of Tn-T does not interact with Tn-C. However, we have shown previously (11) that the binding of Tn-I to T M leads to an increased affinity of interaction between the Tn-T fragment T1 and TM, presumably through a conformational change transmitted through TM from the region of its Cys-190 residue to its head to tail overlap. If one accepts the view that in the case of nonpolymerizable TM, the T1 region of Tn-T may bridge the gap between adjacent nonpolymerizable T M molecules, this interaction could thus be further indirectly weakened by the binding of Ca2+ to Tn-C. This could lead to a reduction in the binding of nonpolymerizable T M induced by T n in the presence of Ca2+ and to a loss of cooperativity in that binding, consistent with the observations in the present study. Alternatively, or in addition, Tn in the absence of Ca2+ may induce cooperative binding of nonpolymerizable T M by transmission of conformational changes through the F-actin structure. In the case of Tn-T alone, which reportedly does not bind to F-actin, such induced binding could only occur by a series of conformational changes transmitted from nonpolymerizable T M by guest on March 22, 2020 http://www.jbc.org/ Downloaded from through actin. At the present level of understanding of the system, it is not possible to distinguish among these possibilities nor to assess their contributions to the observed effects.