Characterization of the functional domains on the C-terminal region of caldesmon using full-length and mutant caldesmon molecules.

A series of C-terminal deletion mutants of chicken gizzard smooth muscle caldesmon (CaD) were made using a polymerase chain reaction cloning strategy and a baculovirus expression system, and the precise locations of the functional domains of CaD involved in the regulation of actomyosin ATPase and the binding of actin, tropomyosin, and calmodulin were analyzed. Our results reveal a high affinity calmodulin-binding domain that consists of at least three calmodulin-binding determinants localized in residues 690-717, 658-689, and 628-657. The residues between positions 718 and 756 and positions 598 and 627 have no detectable calmodulin-binding site. A high affinity tropomyosin-binding domain is located between residues 718 and 756. The 159 residues at the C terminus of CaD contain multiple actin-binding determinants; the major ones are localized in the regions between residues 718 and 756 and residues 690 and 717. The amino acid residues between positions 718 and 756 contain the major determinant involved in the inhibition of the actin activation of smooth muscle myosin ATPase since CaD-(1-717) caused only 30% of the inhibition produced by the full-length CaD. Further deletion between residues 690 and 717 (CaD-(1-689) revealed a low level (10% of that seen for full-length CaD) of inhibition of the actomyosin ATPase. These data clearly demonstrate that the region of the last 66 amino acid residues at the CaD C terminus contains two or more major actin-binding motifs, one tropomyosin-binding domain, one high affinity calmodulin-binding determinant, and the domain that is responsible for the inhibition of the actin-activated ATPase of myosin.

ated regulation that complements the myosin-mediated regulation of smooth muscle myosin via myosin light chain phosphorylation and dephosphorylation (Refs. 8 -10; for review, see Ref. 11).
Structurally, CaD is a long asymmetric monomeric (80 ϫ 2 nm) molecule that is localized along the actin filament, covering 7-14 actin monomers (12)(13)(14). While the CaD C terminus binds to actin, the N-terminal region of CaD binds to the S-2 regions of the myosin (15,16). An understanding of the structural and functional domains on the CaD molecule, responsible for its interaction with myosin, actin, calmodulin, and tropomyosin, is essential to elucidate the mechanism by which CaD down-regulates actomyosin ATPase activity. Previous studies (17)(18)(19) have demonstrated an actin-and a calmodulin-binding region on the C-terminal proteolytic fragment of ϳ38 kDa (as determined by SDS-PAGE). The 38-kDa fragment possesses the ability to inhibit actomyosin ATPase activity, but it fails to cross-link myosin to actin (19,20). Using a bacterial expression system, Redwood and Marston (21) made several CaD fragments containing the sequences for actin-and calmodulinbinding sites and for the region responsible for the inhibition of the actin-activated ATPase activity of myosin. Since CaD fragments, either derived from proteolysis or produced by a bacterial expression system, lack a major portion of the molecule, they do not produce the structural association between actin and caldesmon that is likely to be associated with the thin filament-mediated regulation in smooth muscle.
Recombinant CaD produced using the baculovirus expression system is functionally and structurally similar to the native molecule (22). In this study, we utilize caldesmon mutants made in a baculovirus expression system to characterize the actin-, tropomyosin-, and calmodulin-binding domains and the ATPase inhibitory domain on the CaD C terminus. Deletion of specific sequences from the C-terminal region, without affecting the major portion of the molecule responsible for the structural association with the actin filament, enabled us to delineate the calmodulin-, actin-, and tropomyosin-binding domains and to study the effect of these domains on the actin-activated ATPase activity of smooth muscle myosin. We demonstrate that the sequence that is present in the proteolytic CaD fragment of 7.3 kDa (residues between Leu 597 and Phe 665 ), thought to contain the region responsible for the inhibition of the actinactivated ATPase of myosin S-1 (12), is not important for the inhibition of the actin activation of smooth muscle myosin ATPase in either the presence or absence of tropomyosin. The region between residues 690 and 756 contains the high affinity actin-binding site(s) and the domain that is important for the inhibition of actomyosin ATPase activity. We also show that deletion of a sequence (residues 718 -725) previously reported to be important for calmodulin binding to CaD, based on experiments using a bacterially expressed CaD fragment (23,24), has no effect on calmodulin binding.
Transfection and Isolation of Recombinant Baculovirus-Transfection of Spodoptera frugiperda (Sf9) insect cells and isolation of recombinant baculovirus were carried out as described (25,26). Immunoblotting assay using polyclonal antibody against chicken gizzard smooth muscle CaD was performed to verify the CaD deletion mutants (22).
Purification of Recombinant Caldesmon and Other Proteins-Recombinant CaD was prepared according to Wang et al. (22). Smooth muscle actin, CaD, myosin (fully phosphorylated), and tropomyosin were purified from chicken gizzard (27,28). Bovine brain calmodulin was prepared according to Dedman and Kaetzel (29) serum albumin as a standard. The binding of CaD mutants to calmodulin was determined using calmodulin-coupled agarose (Sigma). Before the binding assay, calmodulin-coupled agarose was equilibrated with washing buffer containing 20 mM Tris-HCl (pH 7.2), 0.5 mM CaCl 2 , 1 mM DTT, 100 mM NaCl, and 0.1 mM NaN 3 . An aliquot (50 l) of the equilibrated calmodulin-coupled agarose was incubated with equimolar amounts (4.5 M) of full-length CaD or each CaD mutant for 2 h at 4°C. After being washed three times with washing buffer, the agarose was incubated with elution buffer (washing buffer containing 2 mM EDTA) for 30 min at 4°C. The agarose suspension was centrifuged at 12,000 ϫ g for 5 min, and the supernatants from washing and elution were subjected to SDS-PAGE (31) followed by Coomassie Blue staining. BSA and tropomyosin were treated as described above and used as negative controls. The amounts of bound and unbound full-length CaD or CaD mutants were determined by scanning densitometry after SDS-PAGE. Calmodulin binding to caldesmon was also determined using full-length CaD and CaD mutants labeled with [ 14 C]iodoacetamide as described (5). Calmodulin binding to caldesmon was measured by incubation of 50 l of calmodulin-agarose with increasing concentrations of [ 14 C]iodoacetamide-labeled full-length CaD or CaD mutants. The calmodulin binding assay conditions are described in the figure legends. After elution, the pellets and supernatants were counted in a liquid scintillation counter (Beckman Instruments). Nonspecific binding was estimated by incubation of uncoupled agarose with increasing concentrations of 14 C-labeled full-length CaD or CaD mutants, and these values were subtracted from each point on the binding curve. All calmodulin binding assays were carried out in triplicate.
Fluorescence Measurement for Interaction of CaD Mutants with Smooth Muscle Tropomyosin-The tropomyosin purified from chicken gizzard was reduced with DTT (5 mM) and labeled with N-(1-pyrenyl)iodoacetamide (Polyscience) as described by Ishii and Lehrer (32) under denaturing conditions. Fluorescence intensity of pyrene-labeled tropomyosin was measured using increasing concentrations of each of the purified CaD mutants at a constant concentration of pyrene-labeled tropomyosin using a fluorescence spectrophotometer (Photon Technology International Inc.) at 25°C (33). The value ⌬F/F (F and ⌬F represent initial fluorescence intensity and change in fluorescence intensity after addition of CaD, respectively) was plotted against the concentration of caldesmon. The level of labeling for pyrene-labeled tropomyosin, determined from the absorbance at 344 nm using an extinction coefficient of 2.3 ϫ 10 4 M Ϫ1 cm Ϫ1 (33), was found to be 1.9.
Binding Assays for Smooth Muscle Myosin, Actin, and Tropomyosin-Actin-The binding of CaD mutants to myosin, actin, and tropomyosinactin was determined by cosedimentation using an Airfuge (Beckman Instruments) as described (22,33). The bound and unbound CaD mutants were estimated by scanning densitometry after SDS-PAGE (22,31). The 14 C-labeled full-length CaD and CaD mutants were also used FIG. 1. Schematic representation of CaD C-terminal mutant proteins. The CaD mutant proteins that have deletions at the C terminus of CaD were generated using the baculovirus expression system. The numbers indicate the amino acid residue numbers. The hatched areas represent the sequence homology between CaD and troponin T characterized previously (35,37). The two cysteine residues on CaD molecule are presented as SH, and the central repetitive region consisting of a 13-amino acid sequence repeated eight times is indicated by the open bars (37).
to determine actin binding in the presence or absence of tropomyosin as described before (5). All determinations of the binding assays were done in triplicate.
ATPase Assays-ATP assays were carried out at 25°C as described (34). The assay conditions are described in the figure legends.

Expression and Purification of CaD Deletion Mutants-Sf9
cells in suspension culture in the log phase of growth were infected with each recombinant baculovirus, and 36 h postinfection, cells were harvested by centrifugation. The expression of full-length CaD and CaD mutant proteins in Sf9 cells was estimated. The proteins present in the extracts of cells infected with recombinant baculovirus containing the cDNA for various deletion mutants are shown in Fig. 2A. All deletion mutants were overexpressed, although at different levels. Sf9 cells infected with wild-type Autographa californica multiple nuclear polyhidrosis virus DNA did not reveal protein bands with molecular mass similar to that of CaD ( Fig. 2A, lane 3). The expression level of each of these mutants was determined by SDS-PAGE (31) of the heat-stable fractions from the Sf9 cell extract and densitometric scanning of the Coomassie Bluestained gels. The yields for 1-ml cell cultures (10 6 cells) varied from 30 to 50 g for the recombinant caldesmons; the yield for full-length CaD was higher than the yields for the truncated molecules. A 32-kDa protein present in the extract of transfected Sf9 cells, although similar in molecular size to polyhedrin protein, is shown to be a breakdown product of caldesmon (22).

Binding of CaD Mutants to Smooth Muscle Myosin-
The effect of the C-terminal portion of CaD on the binding of CaD to smooth muscle myosin was determined by cosedimentation assay. Increasing concentrations of each CaD mutant were mixed with a constant concentration of purified chicken gizzard smooth muscle myosin in the phosphorylated form. As expected, all the CaD mutants bound strongly to myosin (Fig. 3). The differences in myosin binding between full-length CaD and CaD mutants were not remarkable. At a CaD/myosin molar ratio of 0.8, the myosin binding of CaD-(1-717), CaD-(1-627), and CaD-(1-597) was 6, 11, and 13% lower than that of fulllength CaD, respectively. The myosin binding of CaD-(1-717) and CaD-(1-689) was identical at a CaD/myosin molar ratio of 1.4:1. The myosin binding of native CaD and baculovirus-produced full-length CaD was indistinguishable, as reported (22).
Binding of CaD Mutants to Calmodulin-The ability of CaD mutants to interact with calmodulin was determined using a rapid and convenient assay in which CaD was cosedimented with calmodulin-agarose. An equimolar amount (4.5 M) of each of the purified CaD mutants was mixed with a fixed amount of calmodulin-agarose equilibrated with assay buffer containing 10 mM imidazole HCl (pH 7.0), 20 mM KCl, 0.1 mM CaCl 2 , 2 mM DTT, 2 mM MgCl 2 , and 2 mM ATP. Unbound imidazole HCl (pH 7.0) for 10 min at room temperature, and the mixtures were cosedimented in an Airfuge as described previously (28). The proteins in the supernatants and pellets were analyzed by SDS-PAGE and quantitated by scanning densitometry. caldesmon was washed off the calmodulin-agarose by low speed centrifugation. Bound CaD was eluted from the sedimented agarose using a buffer containing 10 mM imidazole HCl (pH 7.0), 20 mM KCl, 2 mM DTT, and 2 mM EDTA. The proteins in the washing and elution solutions were precipitated using 10% trichloroacetic acid and analyzed by SDS-PAGE (Fig. 4A). After washing, most of the full-length CaD and CaD-(1-717) remained bound to the calmodulin-coupled agarose, whereas only small amounts of CaD-(1-597), CaD-(1-627), and CaD-(1-657) bound to the calmodulin-coupled agarose as indicated by the presence of these proteins in the wash (Fig. 4, A-C). BSA and smooth muscle tropomyosin were treated under conditions identical to those of the CaD mutants and used as negative controls for nonspecific binding to calmodulin-agarose; almost all of these proteins were present in the wash (Fig. 4C). Nonspecific binding was considered as the background. Quantitations of various CaD mutants bound to calmodulin-agarose by scanning densitometry revealed that CaD-(1-717), which lacks 39 amino acid residues at the C terminus, bound to calmodulin as efficiently as full-length CaD did (Fig. 4C). With further deletions at the C terminus of CaD, the calmodulin binding affinities decreased significantly. The binding of CaD-(1-689), which lacks the last 67 amino acid residues, to calmodulin was only 40% of that exhibited by full-length CaD or CaD-(1-717). The calmodulin binding of CaD-(1-657), which lacks 99 amino acid residues at the C terminus, was ϳ10-fold weaker than that of full-length CaD (Fig. 4A). The relative amounts of full-length CaD and CaD mutants bound or unbound to calmodulin-agarose are shown in Fig. 4D.
The binding affinities of the CaD mutants for calmodulin were determined using [ 14

Binding of CaD Mutants to Smooth Muscle Tropomyosin-
The tropomyosin-binding domain of CaD was believed to be localized at a highly homologous region (residues 508 -565) corresponding to the tropomyosin-binding site on troponin T from skeletal muscle (35,37). However, it was recently reported that multiple tropomyosin-binding sites were also found on other regions of CaD, but not between residues 658 and 756 (21). To investigate the precise tropomyosin-binding sites at the C terminus, the tropomyosin binding of the CaD mutants was tested by measuring the excimer fluorescence emission from pyrene-labeled tropomyosin. As reported previously, the monomer peaks displayed between 385 and 405 nm, and the excimer peak was observed at 485 nm (33). When CaD bound to pyrene-labeled tropomyosin, the monomer fluorescence of pyrene-labeled tropomyosin was increased, and the excimer fluorescence of pyrene-labeled tropomyosin was decreased. Because the change in the excimer fluorescence induced by CaD was higher and more consistent than that in the monomer fluorescence, the intensity of the excimer fluorescence (485 nm) was measured to determine the interaction of CaD or CaD mutants with pyrene-labeled tropomyosin. As shown in Fig. 6, the intensity of the excimer fluorescence of pyrene-labeled tropomyosin in the presence of recombinant full-length CaD was ϳ75% of that in its absence (a 25% reduction). When the CaD mutant lacking amino acid residues between positions 718 and 756 (CaD-(1-717)) at the C terminus was added to pyrenelabeled tropomyosin, the decrease in the excimer fluorescence was only ϳ35% of that observed for full-length CaD (Fig. 6). The change in the excimer fluorescence upon addition of pyrene-labeled tropomyosin to CaD mutants with deletion of amino acid residues between positions 598 and 717 was similar to that for CaD- (1-717). Therefore, the region between amino acid residues 718 and 756 on CaD contains the major tropomyosin-binding domain, in contrast to the observation made from experiments using CaD fragments (21).
Mapping of the Actin-binding Domains on the C-terminal Region of CaD-To delineate the region responsible for the binding of CaD to actin, purified 14 C-labeled CaD mutant proteins were tested for their ability to bind to smooth muscle actin using the cosedimentation assay. As shown in Fig. 7A, deletion of residues between positions 718 and 756 (CaD-(1-717)) at the C terminus of CaD resulted in an ϳ31% loss in the ability of CaD to bind to actin. The actin binding of CaD was weakened further (37% of that for full-length CaD) when the residues between positions 690 and 756 were absent (CaD-(1-689)). The actin binding of the CaD mutant that lacks the residues between positions 658 and 756 (CaD-(1-657)) was only 29% of that of full-length CaD. CaD-(1-627) showed weaker actin binding, indistinguishable from that of CaD-(1-597) (Fig. 7A) CaD mutants to the reconstituted tropomyosin-actin filament. As seen in Fig. 7 (A and B), the presence of tropomyosin increased the binding affinities of full-length CaD, CaD-(1-717), and CaD-(1-689) for actin. In the presence of tropomyosin, the apparent K d value of actin binding for full-length CaD (0.28 Ϯ 0.015 ϫ 10 Ϫ6 M) was 2-fold lower than the K d value measured in the absence of tropomyosin. Deletion of residues 718 -756 caused a significant increase in the K d value to The effect of Ca 2ϩ -calmodulin on the reversal of actin binding was also determined. These results are shown in Fig. 8 (A  and B). The actin binding of all CaD mutants decreased upon raising the concentration of calmodulin at a constant 1:7 molar ratio of CaD to actin. At a 1:1 molar ratio of calmodulin to actin (CaD/CaM ratio of 1:7), CaD released from actin was ϳ76% for full-length CaD and CaD-(1-717), 60% for CaD-(1-689), 56% for CaD-(1-657), and 20% for both CaD-(1-627) and CaD- . The release of all CaD mutants from actin by Ca 2ϩcalmodulin was less in the presence of tropomyosin than in its absence (Fig. 8B). These data suggest that tropomyosin increases the affinity of CaD for actin due to an interaction between CaD and tropomyosin.

Effects of CaD Mutants on Actin-activated ATP Hydrolysis by Smooth Muscle
Myosin-The region of the CaD molecule responsible for the inhibition of the activation of myosin ATPase activity by actin or tropomyosin-actin has been shown to reside in the C-terminal portion of CaD (18,21). To precisely localize the region responsible for the inhibition of actin-activated myosin ATPase activity and to determine if the binding of CaD mutants to actin correlates with the inhibition of the actin activation of myosin ATPase, actin-activated ATP hydrolysis by phosphorylated myosin was measured in the presence of CaD mutants. As illustrated in Fig. 9A, the actin-activated ATPase activity was inhibited upon raising the concentration of full-length CaD, CaD-(1-717), or CaD-(1-689), whereas CaD-(1-657), CaD-(1-627), and CaD-(1-597) failed to inhibit the actin-activated ATPase activity, even at a CaD/actin molar ratio of 1:1 (data not shown). At an ϳ1:5 molar ratio of CaD to actin, the inhibition observed with full-length CaD, CaD-(1-717), and CaD-(1-689) was 60, 15, and 6%, respectively (Fig.  9A). Interestingly, the inhibition caused by CaD-(1-717), as well as that caused by full-length CaD, was augmented in the presence of tropomyosin. However, the inhibition of the actinactivated ATPase activity produced by CaD-(1-689) was not enhanced by tropomyosin (Fig. 9B), indicating that the functional domains required for the tropomyosin-dependent inhibition reside in the regions between amino acids 690 and 717 and amino acids 718 and 756. It appears that the sequence between residues 658 and 689 is responsible for a weak tropomyosinindependent inhibition of the actin-activated ATPase activity. The inhibition of the actin-activated ATPase activity induced by both CaD-(1-756) and CaD-(1-717) was completely abolished when the molar ratio of CaM to CaD (or CaD-(1-717)) reached 10:1. However, at a CaM/CaD molar ratio of 3:1, the inhibition by CaD-(1-717) was reversed slightly more than that by full-length CaD.
To determine whether the inhibition of ATPase described above correlated with the binding of CaD to actin either in the presence or absence of smooth muscle tropomyosin, the binding assays and ATPase assays were carried out simultaneously. As shown in Fig. 10 (A and B), the inhibition caused by full-length CaD and CaD-(1-717) was ϳ40 and 15%, respectively, when ϳ0.06 mol of CaD was bound per mol of actin. However, in the presence of tropomyosin under the same condition (0.06 mol of CaD bound per mol of actin), the inhibition shown by fulllength CaD and CaD-(1-717) was ϳ52 and 25%, respectively. These data suggest that tropomyosin enhances the inhibition of the actin-activated myosin ATPase activity without additional binding of CaD to actin. DISCUSSION Our data utilizing these truncated CaD proteins reveal some important differences from the data obtained from studies using bacterially expressed or proteolytic CaD fragments, presumably due to the lack of structural association present in protein-protein interaction. The major myosin-binding site is located in the N terminus of CaD (15,39), and a recent study shows that the residues between positions 235 and 531 of rat nonmuscle CaD coprecipitate with smooth muscle myosin in an in vitro binding assay (40). This finding is further supported by Marston and co-workers (41), who showed that a bacterially produced fragment composed of the 288 residues at the C-FIG. 8. Effect of Ca 2؉ -calmodulin on binding of CaD mutants to actin or tropomyosin-actin. Conditions are the same as described for Fig. 6 except that the concentrations of CaD and actin were kept constant (0.15 mol of CaD/mol of actin). A, reversal of binding of CaD mutants to actin; B, reversal of binding of CaD mutants to tropomyosinactin. Symbols are as described for Fig. 3. Note that the release of mutants CaD-(1-597), CaD-(1-627), and CaD-(1-657) is not remarkable since the binding of these mutants to actin is also very low (see Fig.  7, A and B).
terminal region of human nonmuscle CaD bound to myosin. But the ability of this fragment to bind to myosin was not compared with that of CaD containing the central helical region and parts of the C terminus. In the present study, a slight decrease (6, 11, and 13% for CaD-(1-717), CaD-(1-627), and CaD-(1-597), respectively) in the binding of caldesmon mutants to myosin was observed when compared with the fulllength molecule. This is probably caused by the conformational change in the rest of the CaD molecule. Maximal binding of CaD to myosin seems to require the C-terminal region of CaD to be intact.
Studies using chymotryptic peptides of CaD have demonstrated that the C-terminal region of CaD contains a calmodulin-binding site, presumably in the region between residues 629 and 666 (37). Zhan et al. (42) showed that a synthetic peptide with an amino acid sequence homologous to CaD between residues 658 and 666 bound to calmodulin. A recent report showed that a synthetic peptide spanning the residues between positions 675 and 695 bound to calmodulin-Sepharose and was eluted with EGTA (43). These data indicate the possibility that another calmodulin-binding site exists in the Cterminal end of CaD. Data from the present study reveal that the C-terminal high affinity calmodulin-binding site of chicken gizzard smooth muscle CaD is located in the last 129 amino acids, probably involving two or more discontinuous epitopes. The residues between positions 690 and 717 and positions 658 and 689 play a major role in maintaining high affinity binding to calmodulin since deletion of either of these regions results in a significant decrease in calmodulin binding when compared with that of full-length CaD. These data further confirmed the results from analyses of synthetic peptides (42,43), which determined that the last 99 amino acid residues contained two calmodulin-binding determinants. It has been proposed that the core regions of the two calmodulin-binding determinants more likely involve residues 658 -666 (calmodulin-binding motif 1) (42) and residues 675-695 (calmodulin-binding motif 2) (43), respectively. Based on the information from the data in the present study and on comparison with data from published reports (43), the core region of calmodulin-binding motif 2 can be further defined to a 5-amino acid stretch from Asn 690 to Lys 695 (Fig. 11). Our results also show that the residues between positions 628 and 657, part of the sequence included in the region indicated by Wang et al. (37) as the partial sequence for calmodulin binding on the CaD molecule, do not contain a region responsible for the high affinity binding of CaD to calmodulin. However, CaD-(1-657) binds weakly to calmodulin, but CaD-(1-627) does not (Figs. 4 and 5), indicating that the region between residues 628 and 657 contains a sequence responsible for weak calmodulin binding. In addition, our data rule out a calmodulin-binding site in the region between residues 718 and 756, a region suggested by Marston and coworkers (23,24) to be important for calmodulin binding. The difference in the conformation between the truncated proteins containing the N-terminal and central helical regions used in the present study and the caldesmon fragments used in previous studies may account for this discrepancy. Calmodulin binding is similar for truncated CaD-(1-717) and full-length CaD (Fig. 5). However, in the presence of CaM, the dissociation of full-length CaD and CaD-(1-717) from actin is the same, despite the weak binding of the latter to actin. This suggests that calmodulin dissociates the bound CaD from actin by inducing a similar conformational change in the regions between residues 690 and 717 and residues 718 and 756.
It is well known that tropomyosin, a coiled-coil thin filamentassociated protein, can directly bind to CaD with a stoichiometry of 1-2 mol of tropomyosin/mol of CaD (13,14,44). Electron microscopic observations of reconstituted thin filament revealed that CaD fixes tropomyosin in a particular position on actin filaments (45). However, there is very little information on the tropomyosin-binding sites on CaD. Based upon sequence alignment analysis with the tropomyosin-binding site on troponin T, the tropomyosin-binding site on chicken gizzard smooth muscle CaD is proposed to lie between residues 508 and 565 (35,37). Our data with truncated CaD show that the deletion of 39 residues at the C-terminal end of CaD remarkably reduces the binding of CaD to tropomyosin (Fig. 6), in contrast to the report by Redwood and Marston (21) showing that there is no tropomyosin-binding site between residues 658 and 756. Further removal of up to 120 residues from Asn 598 to Gly 717 (as in CaD-(1-597), CaD-(1-627), CaD-(1-657), and CaD-(1-689)) caused only a slight decrease in the binding to tropomyosin. These results suggest strongly that CaD contains two or more tropomyosin-binding sites; the major tropomyosinbinding site apparently resides in the C-terminal 39 amino acids. It appears that the region between amino acids 627 and 717 of CaD is not involved in the binding of tropomyosin since only a small difference is observed between CaD-(1-627) and CaD-(1-717) in tropomyosin binding.
This study is in agreement with our previous finding that tropomyosin can enhance the binding of CaD to actin via a triggering of conformational changes on these thin filamentassociated proteins (33). The binding of full-length CaD and CaD-(1-717) or CaD-(1-689) to actin was higher in the presence of tropomyosin than in its absence (Fig. 7A). It should be noted that the actin binding of CaD-(1-657), CaD-(1-627), and CaD-(1-597) can be slightly enhanced by tropomyosin (Fig.  7B), indicating that a tropomyosin-binding site present in the region between residues 1 and 597, perhaps in the region between residues 230 and 419 of CaD shown to bind to tropomyosin (21), is also responsible for the tropomyosin-enhanced binding of CaD to actin.
The major actin-binding site(s) on the CaD molecule has been traced to the C-terminal region of CaD, perhaps between residues 659 and 665 (12) or involving two independent binding motifs that are separated by residues 629 -666 (37). Our cosedimentation binding assays reveal that there are a number of actin-binding motifs present within the last 159 amino acids at the C terminus of chicken gizzard CaD. CaD-(1-717), which lacks 39 amino acids, shows diminished binding to actin, suggesting that the residues between positions 718 and 756 contributed to the high affinity actin-binding site as postulated by Wang et al. (37). However, another crucial actin-binding determinant has been found to lie between residues 690 and 717 (Fig. 7A). It is possible that these sequences are components of one binding site with more than one determinant. The observation that CaD-(1-657) weakly bound to actin rules out the possibility that the region between residues 597 and 629 (37) is attributable to the high affinity actin binding of CaD. As described in our previous report (22), CaD-(1-597), which lacks 156 residues at the C terminus of CaD, weakly binds to actin. However, the apparent K d value for CaD-(1-597) was 5-fold higher than that for full-length CaD, indicating that the region FIG. 10. Relationship between CaD bound to actin and inhibition of actin-or tropomyosin-actin-activated myosin ATPase activity. Actin-activated ATP hydrolysis of phosphorylated smooth muscle myosin was determined in the absence and presence of tropomyosin. The binding of purified full-length CaD (q) and the CaD mutant CaD-(1-717) (E) to actin (A) and tropomyosin-actin (B) was determined in parallel experiments. Conditions of the assays are as described for Fig. 9. A, relationship between the inhibition of actinactivated ATPase activity and the actin binding of full-length CaD and the CaD mutant CaD-(1-717) in the absence of tropomyosin; B, relationship between the inhibition of actin-activated ATPase activity and the actin binding of full-length CaD and the CaD mutant CaD-  in the presence of tropomyosin.
FIG. 11. Summary of domain structure at C terminus of CaD. The region between residues 508 and 565 is homologous to the tropomyosin-binding site of troponin T and is considered to be the tropomyosin-binding site on the CaD molecule (37). High affinity calmodulin (CaM)-binding site 1 (MWEKGNVFS) is derived from Wang and co-workers (42), and high affinity calmodulin-binding site 2 is deduced from our present data and the published data of Mezgueldi et al. (43). The tropomyosin-binding domain was localized in the region between residues 718 and 727 (Ref. 47). between amino acids 1 and 597 contains an additional low affinity actin-binding site. Interestingly, an actin-binding site in residues 483-508 has been proposed by Wang et al. (37), and a recent electron microscopic image (46) also provides direct evidence that both ends of CaD interact with actin filaments.
It has been well established that CaD inhibits the actinactivated or tropomyosin-actin-activated ATPase activity of both myosin I and myosin II (1)(2)(3)(4)(5)38). Several lines of evidence show that the inhibitory region is located in the last 99 amino acids at the C terminus of CaD (19,21). Our present study reveals that there are at least two ATPase inhibitory determinants in the C-terminal region of chicken gizzard smooth muscle CaD. These determinants are between residues 718 and 756 and residues 690 and 717 (Fig. 9, A and B). The residues between positions 658 and 689, which contain most of the amino acid sequence of the 7.3-kDa chymotryptic fragment (12), cause only a very weak ATPase inhibition (see inhibition by CaD-(1-689) and CaD-(1-657) in Fig. 9). In contrast, the domain between residues 718 and 756 contains the critical determinants for the ATPase inhibition since CaD-(1-717), lacking the residues between positions 718 and 756, displays 30% of the ATPase inhibition produced by full-length CaD (Fig.  9). As expected, both CaD-(1-627) and CaD-(1-597) did not elicit any effect on the actin-activated ATPase activity, consistent with previous reports (21,22,37). These data indicate that the sequences between residues 690 and 717 and residues 718 and 756 contain the essential regulatory motifs for the actinactivated ATPase activity, and these regulatory motifs are functionally independent or separate. Our present results are in agreement with the observation of Wang et al. (37) that the last 39 amino acid residues of CaD are mainly responsible for the inhibition of the actin activation of myosin ATPase. These findings conflict with the observation of Chalovich et al. (12) that the residues between positions 597 and 665 encompass the major inhibitory sequences for actomyosin ATPase. The reason for this discrepancy is not clear; interestingly, this proteolytic fragment bound to actin at a 1:1 molar ratio, and the inhibition of actomyosin ATPase required a very high CaD fragment (7.3 kDa) to actin molar ratio, although the intact CaD molecule covers between 7 and 14 actin monomers.
Tropomyosin-dependent ATPase inhibition was also tested in our study. The ATPase inhibition caused by CaD-(1-717), but not by CaD-(1-689), was potentiated by tropomyosin, suggesting that the region between residues 690 and 756 is related to the tropomyosin-dependent ATPase inhibition. At the same molar ratio of CaD to actin, the percent inhibition in the presence of tropomyosin is higher than that in its absence (Fig. 10,  A and B).
In conclusion, as shown in Fig. 11, the last 66 amino acid residues at the C terminus of CaD contain two or more major motifs for actin binding, a tropomyosin-binding determinant, and one high affinity calmodulin-binding determinant, of which the last 39 amino acid residues are involved in both actin binding and tropomyosin binding. The tropomyosin-binding motif in the region consisting of the last 39 amino acids can be further defined to a 10-amino acid stretch between residues 718 -727 using a C-terminal deletion mutant that lacks residues between positions 728 and 756 (47). This region is also responsible for the major inhibition of the actin-activated ATPase activity of myosin. The 28-amino acid stretch (residues 690 -717) encompasses both low affinity actin binding and high affinity calmodulin binding. These results suggest that the calmodulin-, actin-, and tropomyosin-binding motifs on the CaD molecule are overlapping or are closely juxtaposed with each other. The localization of these important motifs at the C terminus of CaD helps add to the basic understanding of the structural and functional domains on the CaD molecule involved in the regulation of actin-activated ATP hydrolysis by myosin.