The effects of smooth muscle calponin on the strong and weak myosin binding sites of F-actin.

We have investigated the mechanism of inhibition of the actomyosin MgATPase by the smooth muscle protein calponin. We have shown previously the specific interaction of calponin with Glu334 of actin (EL-Mezgueldi, M., Fattoum, A., Derancourt, J., and Kassab, R. (1992) J. Biol. Chem. 267, 15943-15951). This residue is within the sequence 332-334, which has been proposed to be an important part of the strong myosin binding site (Rayment, I., Holden, H. M., Whittaker, M., Yohn, C. B., Lorenz, M., Holmes, K. C., and Milligan, R. A. (1993) Science 261, 58-65). Therefore, we suggested that calponin will affect the strong binding actin-myosin interaction. To test this hypothesis we have investigated the effect of calponin on the strong binding of S-1·MgAMP-PNP (5′-adenylyl imidodiphosphate) and on the weak binding of S-1·MgADP·Pi to actin. We found that an inhibitory concentration of calponin decreased the binding of S-1·MgAMP-PNP to actin but had no effect on the binding of S-1·MgADP·Pi. Similar results were obtained with skeletal muscle and smooth muscle S-1. In competition experiments calponin was found to displace S-1·MgAMP-PNP and S-1·MgADP but not S-1·MgADP·Pi from the actin filament. S-1 displaced calponin from actin in the rigor state, in the presence of MgADP, and in the presence of MgAMP-PNP. We conclude that calponin inhibits the actin activated S-1 ATPase by blocking a strong S-1 binding site on actin and does not block the weak binding site.

Three actin-binding proteins, troponin I, caldesmon, and calponin, have been identified as capable of inhibiting the activation of myosin ATPase by muscle thin filaments (1)(2)(3). Although the mechanism of inhibition of the actomyosin ATPase by troponin I and caldesmon has been studied extensively (4 -9), fewer studies have approached the mechanism of inhibition by calponin (10,11). Calponin is a 34-kDa smooth muscle protein that binds actin (3), myosin (12), Ca 2ϩ -binding proteins (13,14), and tropomyosin (15). In vitro calponin inhibits the actomyosin ATPase activity upon binding to actin, and this inhibition is reversed by its phosphorylation (16) or by its interaction with Ca 2ϩ -binding proteins such as calmodulin or caltropin (13,14). Consequently, it has been suggested that calponin may have a role in controlling smooth muscle contraction. Although the actin binding and inhibitory activity of troponin I and caldesmon are affected strongly by tropomyosin (4,8,17), calponin binding to actin and its inhibitory activity are not affected by tropomyosin (18,19). This suggests that calponin inhibits the actomyosin ATPase activity at the molecular level by a mechanism different from that of caldesmon and troponin I.
We have investigated previously the calponin-actin interface using proteolysis, peptide synthesis, and recombinant technology (20 -22). Those studies led us to build up a detailed picture of the inhibitory domain of calponin. The actin-calponin interface seems to be a multiple contact site involving the calponin sequence extended from Val 142 to Tyr 182 . One contact site involving the calponin amino acids VKYAEK at position 142-147 seems to be directly responsible for ATPase inhibition (21). On the actin monomer, three sequences have been implicated in calponin binding: (i) the sequence 1-226, a large fragment corresponding mainly to actin subdomains 1, 2, and 4 (23); (ii) the sequence 326 -355, which spans parts of subdomain 1 and subdomain 3 (20); (iii) the COOH-terminal three amino acids (24). The sequence 326 -355 is of particular interest since this region contains the segment 332-334, which has been proposed to be part of the strong binding site of myosin to actin (25). Moreover, calponin has been cross-linked to Glu 334 within this segment using the zero-length cross-linker carbodiimide (20). In contrast Lys 61 and the NH 2 terminus of actin, proposed to be part of the weak binding site of the myosin head (S-1), 1 have been excluded from calponin binding (11). From these structural studies (summarized in Fig. 7) it is reasonable to suggest that calponin would inhibit the strong binding actin-myosin interaction and that calponin and the S-1 myosin head would compete for a common binding site on actin.
To test this hypothesis we have investigated the effect of calponin on the binding of S-1 to actin in the absence or the presence of MgAMP-PNP, MgADP, or MgATP. We found that an inhibitory concentration of calponin decreased the strong binding of S-1⅐MgAMP-PNP to actin. However, calponin had no effect on the weak binding of S-1⅐MgADP⅐P i to actin. In competition experiments calponin was found to displace S-1⅐MgAMP-PNP and S-1⅐MgADP from actin but did not interfere with the binding of S-1⅐MgADP⅐P i . On the other hand S-1 displaced calponin from actin in the rigor state and in the presence of MgADP and MgAMP-PNP.
We conclude that calponin binding to actin affects the strong binding of myosin to actin but has no effect on the weak binding.

MATERIALS AND METHODS
Muscle Proteins-Chicken gizzard calponin and its chymotryptic fragment CH22, spanning residues 7-182, were produced as described previously (20). Rabbit skeletal muscle F-actin and S-1 were prepared * This research was supported by a grant from the Wellcome Trust. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
‡ To whom correspondence and reprint requests should be addressed.  (26,27). Chicken gizzard smooth muscle S-1 was prepared according to Marston and Taylor (28). S-1 was labeled with [ 14 C]iodoacetamide according to Marston and Weber (29). The protein concentrations were measured by the Lowry method.
Competition Experiments-Mixtures of actin, S-1, and calponin were sedimented at 25°C in 10 mM PIPES, pH 7.0, 10 mM KCl, 2.5 mM MgCl 2 , and 0.1 mM DTT, in the absence or presence of nucleotides. The supernatants were removed carefully, and the pellets were dissolved in Laemmli buffer. The proteins were separated by electrophoresis on 8 -18% polyacrylamide gradient 1% sodium dodecyl sulfate gels (Pharmacia Biotech Inc. Excel Gel). The amount of bound calponin or S-1 was determined by scanning densitometry of the gels. Densitometric scanning of the protein bands was performed with a Howtek Scanmaster 3 scanner and PDI system I version 2.2 software. A calibration curve was generated by scanning known quantities of calponin or S-1 loaded on the same gel. All points were within the linear range. The quantity of protein bound in the actin pellet was normalized by the formula [(area S-1 heavy chain ϩ area S-1 light chain)/area actin] for S-1 and [area calponin/area actin] for calponin.
Binding Measurements-S-1 binding to actin was determined by cosedimenting actin with S-1 bound to actin as described by Marston et al. (9) and assaying the decrease of the S-1 concentration in the supernatant. For strong binding measurements, 0 -16 M [ 14 C]iodoacetamide-labeled S-1 was mixed with 10 M actin or 10 M actin and 20 M calponin in a volume of 100 l of 10 mM PIPES, pH 7.0, 10 mM KCl, 2.5 mM MgCl 2 , and 0.1 mM DTT. Mixtures contained 2 mM MgAMP-PNP plus 5 mM glucose and 4 units/ml hexokinase. These components were made as 10 ϫ concentrated stock solutions and incubated for 30 min at room temperature to ensure the removal of all traces of ATP before being added to the assay solution. 25-l samples of supernatants were taken from the mixture before and after sedimentation (90 min at 50,000 ϫ g, Sorvall Sm24 rotor, 25°C), and [ 14 C]iodoacetamide-labeled S-1 was assayed by scintillation counting. Binding was calculated as follows. [Free S-1] ϭ S-1 in supernatant after sedimentation; [bound S-1] ϭ S-1 before sedimentation Ϫ S-1 in supernatant after sedimentation. Corrections were made for S-1 that sedimented in absence of actin and S-1 that never sedimented (29,30).
To measure the binding of S-1⅐MgADP⅐P i to actin or actin-calponin, 100 l of unlabeled S-1 was sedimented in the presence of 5 mM MgATP at 4°C for 20 min at 300,000 ϫ g. S-1 in the supernatant was determined by assay of the EDTA-ATPase activity according to (8,31).
ATPase Measurements-The activation of the skeletal and smooth muscle S-1 MgATPase by skeletal muscle actin was determined in 10 mM PIPES, 10 mM KCl, 2.5 mM MgCl 2 , 0.1 mM DTT, pH 7.0, at 30°C. The reaction was started by adding 5 mM MgATP and terminated after 3 or 25 min for skeletal muscle S-1 and 30 min for smooth muscle S-1 by adding 0.5 ml of 10% trichloroacetic acid. P i was determined colorimetrically by the method of Taussky and Schorr (32).

Effect of Ionic Strength on Calponin Inhibitory Activity-
Before investigating the mechanism of the acto-S-1 ATPase inhibition by calponin we explored the effect of ionic strength on the inhibitory activity of calponin. Fig. 1 shows the effect of increasing the concentration of calponin on the activation of the ATPase activity of smooth and skeletal muscle myosin S-1 by F-actin at two different KCl concentrations. At 10 mM KCl, increasing concentrations of calponin led to a progressive loss of acto-S-1 MgATPase activity, and about 70% inhibition of the ATPase activity is obtained at a calponin concentration that is equimolar to actin for both smooth and skeletal muscle S-1. At this ionic strength, tropomyosin had no effect on the inhibitory activity of calponin (data not shown). This is in agreement with the previously described pattern of ATPase inhibition by calponin at low ionic strength (18,20,33). Increasing the KCl concentration led to a reduction in the calponin ATPase inhibitory activity. 40 M calponin was needed to reach 60% inhibition at 60 mM KCl and gave only 30% inhibition at 110 mM KCl (data not shown). Raising the KCl concentration to 160 mM KCl abolished the calponin ATPase inhibitory activity even at the highest concentration used with both smooth and skeletal muscle S-1 (Fig. 1). When we measured the binding of calponin to actin we found that increasing the ionic strength from 10 to 160 mM KCl had only a slight effect on calponin binding to actin ( Fig. 2) in agreement with previous reports (19,34). Plotting inhibition of acto-S-1 ATPase activity against bound calponin clearly shows that calponin binding at a stoichiometry of about 1/actin is related to inhibition only at low ionic strength (Fig. 2, inset). This pattern of results was obtained in four sets of experiments. There is no difference in the pattern of inhibition of calponin between smooth and skeletal S-1 at both low and high ionic strength ( Fig. 1), and this indicates that the calponin inhibition mechanism involves only the thin filament.
Displacement of Calponin from Actin by Strongly Bound S-1 Complexes-We investigated the competition between calponin and skeletal muscle S-1 for actin binding by cosedimentation. Mixtures of 10 M actin and 10 M calponin, which give approximately 65% inhibition in acto-S-1 ATPase assays, were sedimented with increasing concentrations of S-1 in the presence of 2 mM MgAMP-PNP (Fig. 3A), 2 mM MgADP (Fig. 3B), or in the absence of nucleotides (Fig. 3C). The amount of calponin bound to actin in the pellet was assayed. The presented curves clearly show the ability of all three of these S-1 strong binding complexes to displace calponin from actin. In the presence of MgAMP-PNP and MgADP, about 50% of bound calponin was displaced by an equimolar concentration of S-1 to actin (Fig. 3,  A and B), whereas the same S-1 concentration displaced about 80% of bound calponin in the rigor state (Fig. 3C). We note that displacement of calponin does not exceed 80% even at high concentrations of S-1 rigor complexes which have high affinity for actin. This suggests that part of the calponin binding site on actin does not overlap with the S-1 binding site.
Displacement of S-1 from Actin by Calponin-We then tested the ability of calponin to displace strongly and weakly bound S-1 complexes from actin. Fig. 4 shows that increasing concentrations of calponin led to a progressive displacement of S-1⅐MgAMP-PNP ( MgATPase assay, only 30% of both S-1 strong complexes were displaced. A maximum of 60% displacement was observed at the highest concentration of calponin with no indication of a plateau being reached. The observed competition between calponin and S-1 strong binding complexes for actin binding is compatible with the structural studies, indicating partial overlap between the calponin binding region and the S-1 binding region on the actin sequence (20,25).
To define whether the shared site on actin corresponds only to the strong myosin binding site or also includes the weak binding site we investigated the relationship between calponin inhibition and the binding of S-1⅐MgADP⅐P i to actin. Fig. 4C shows that up to 20 M calponin did not displace the weak binding S-1⅐MgADP⅐P i complex from actin even though this concentration range gave 80% inhibition of the MgATPase. Thus calponin does not interfere with the weak binding site at concentrations that gave full inhibition of MgATPase activity and caused 50% displacement of S-1⅐MgADP and S-1⅐MgAMP-PNP bound to actin. At higher concentrations of calponin there was displacement of S-1⅐MgADP⅐P i from actin, but displacement did not correlate with more inhibition and is probably due to actin aggregation usually observed at a high calponin/actin ratio. We conclude that calponin inhibition does not require displacement of S-1⅐MgADP⅐P i from actin.
Effect of Calponin on the Weak and the Strong Binding Actin-Myosin Complexes-We measured the effect of calponin on the strong binding of S-1 to actin by cosedimenting actin and [ 14 C]iodoacetamide-labeled S-1 in the presence of 2 mM MgAMP-PNP and in the absence or presence of an inhibitory concentration of calponin. The same method has been used previously to investigate the effect of caldesmon-tropomyosin and troponin-tropomyosin on the strong binding of S-1 to actin (9,35). Actin-S-1⅐MgAMP-PNP is considered to be in a strong binding cross-bridge state since MgAMP-PNP does not relax muscle mechanically, and its attached structure is similar to that of the rigor state (35,36).
In the presence of calponin, at a concentration that gives 90% inhibition of the acto-S-1 MgATPase, the binding of skeletal muscle S-1⅐MgAMP-PNP to actin was decreased strongly (Fig.  5A). From the initial slope we estimated the affinity of skeletal muscle S-1⅐MgAMP-PNP for actin in the absence of calponin to be 10 times higher than in the presence of calponin. This pattern of results was obtained in five separate experiments. In all of the experiments we noted a slight cooperativity in skeletal muscle S-1⅐MgAMP-PNP binding to actin in the presence of calponin. We also tested the effect of calponin on the binding of smooth muscle S-1⅐AMP-PNP to actin. The results are presented in Fig. 5B. Smooth muscle S-1⅐AMP-PNP bound to actin with an affinity of 6.2 ϫ 10 5 M Ϫ1 . At a concentration of calponin which gives 85% inhibition in ATPase assay the binding of smooth muscle S-1⅐AMP-PNP to actin was reduced significantly (the affinity of smooth muscle S-1⅐AMP-PNP for actin is seven times higher than for actin-calponin). We also investigated the effect of calponin on the binding of skeletal muscle S-1⅐MgADP to actin at low ionic strength. Skeletal muscle S-1⅐MgADP bound actin with a significantly higher affinity, estimated to be Ͼ10 6 M Ϫ1 . Binding was substantially reduced by inhibitory concentrations of calponin (data not shown). These results show directly that calponin decreases the strong binding of both smooth and skeletal muscle S-1 to actin.
Since calponin weakened the binding of S-1⅐MgAMP-PNP to actin, it seemed possible that calponin might also affect the binding of the weak complex S-1⅐MgADP⅐P i to actin. Previously, assay of the reduction in the EDTA-ATPase activity of S-1 in the supernatant when S-1⅐MgADP⅐P i is cosedimented with actin has been used to measure the weak binding of myosin to actin (37). Skeletal muscle S-1⅐MgADP⅐P i binding was measured over a range of actin concentrations from 0 to 50 M. Calponin was added at a fixed molar ratio to actin; at the higher actin/calponin concentrations significant aggregation of filaments occurred which led to irregular measurements of S-1⅐MgADP⅐P i bound. Experiments with 2:1 calponin/actin molar ratio, giving approximately 90% inhibition of acto-S-1 MgATPase activity, gave very irregular data but did not suggest a decrease in S-1⅐MgADP⅐P i binding compared with actin alone in the low concentration range where there was no aggregation. Experiments with a calponin/actin molar ratio of 1:1, giving around 65% inhibition, avoided aggregation, and binding could be measured accurately (Fig. 6). Under these conditions the binding of skeletal muscle S-1⅐MgADP⅐P i to the calponin-inhibited actin was indistinguishable from pure actin.
We also did binding experiments using actin inhibited by the chymotryptic calponin fragment CH22 (amino acids 7-182), which does not aggregate filaments like whole calponin. We found that at a CH22/actin molar ratio of 2:1, which gives 60% inhibition of acto-S-1⅐MgATPase activity, CH22 had no effect on the binding of S-1⅐MgADP⅐P i to actin (data not shown).
The converse experiment is shown in Fig. 4C. At a constant actin and S-1⅐MgADP⅐P i concentration, addition of calponin inhibited actin activation but did not change the amount of S-1⅐MgADP⅐P i bound.
We conclude from these experiments that calponin inhibits the actin-activated S-1⅐MgATPase without affecting the weak binding of S-1⅐MgADP⅐P i to actin.

DISCUSSION
Regulation of cross-bridge cycling in smooth and striated muscles involves regulatory proteins that act either on the thick or the thin filaments. Despite the diversity of regulatory proteins it is generally observed that they control cross-bridge cycling by a common mechanism that involves regulating the transition from the weak (actin-myosin⅐ADP⅐P i ) to the strong (actin-myosin⅐ADP) complex (38). This may be a direct effect, as in the regulation by myosin light chain phosphorylation (37), or an indirect mechanism in which the state of actin-tropomyosin is controlled and the block of the strong site is due to tropomyosin (1). Troponin and caldesmon are examples of this kind of regulation (4,9). Calponin is a component of smooth muscles associated with actin filaments which is an inhibitor of actomyosin cross-bridge cycling in vitro and which has therefore been proposed as a potential muscle regulatory protein (3). Our goal in carrying out this work was to determine the mechanism by which calponin inhibits the acto-S-1 ATPase at the molecular level.
From published work it is evident that calponin inhibition of actin filament activity involves a mechanism that is different from striated muscle troponin or smooth muscle caldesmon, since inhibition is not mediated by tropomyosin (18,19). Nevertheless, calponin could act upon the weak to strong transition directly. To determine how calponin inhibits the actin-activated myosin MgATPase we used binding methods that have been used to study the effect of troponin-I-tropomyosin (4) and caldesmon-tropomyosin (9) on the weak and strong binding of S-1 to actin.
We found that calponin inhibition was maximal at low ionic strength and was diminished rapidly as ionic strength was increased, irrespective of whether measurements were made using skeletal or smooth muscle myosin subfragments. This made it necessary to do our studies under low ionic strength conditions. The observation that regulation of acto-S-1 by calponin is the same with the smooth and skeletal muscle S-1, despite the 50-fold difference in their V max and the strikingly different ionic strength dependences of the acto-S-1 interactions (39), indicates that calponin acts only upon the properties of the actin filament. Therefore we are justified in using smooth and skeletal muscle S-1 interchangeably; it should be noted that the majority of previously published experiments on calponin have used skeletal muscle S-1 and assumed that results would be valid for smooth muscle S-1.
A direct measurement of the affinity of the strong binding complex S-1⅐AMP-PNP for actin showed that inhibitory concentrations of calponin strongly reduced the binding affinity, showing that calponin did indeed inhibit the weak to strong transition. The binding experiment was backed up by competition experiments showing that calponin and strong binding myosin complexes such as S-1⅐AMP-PNP compete for a common binding site on actin. In contrast, neither competition nor binding experiments showed any evidence for an effect of calponin on the weak binding of S-1⅐MgADP⅐P i to actin. Together, these results suggest that calponin inhibits the acto-S-1 ATPase by blocking an actin-S-1 contact site exclusively involved in the strong binding of S-1 to actin and crucial for cross-bridge cycling.
This conclusion is consistent with previously published works. Horiuchi and Chacko (10) reported from kinetic studies of smooth muscle acto-S-1 MgATPase that calponin inhibition corresponds mainly to a decrease of the velocity (V max ) with only a slight effect on the affinity of S-1 for actin (K m ). This suggests that calponin has only a slight effect on the binding of the myosin head to actin; instead it affects primarily the ratelimiting step (presumably corresponding to the weak to strong transition). Furthermore, calponin inhibited maximal shortening velocity with only a minor effect on force in skinned Taenia coli smooth muscle fibers (40). In the in vitro motility assay, calponin decreased the fraction of motile actin filaments (34,41), which was also observed with troponin and caldesmon, both known to inhibit by blocking the strong binding state (42,43).
Calponin has been shown to inhibit the ATPase activity of the carbodiimide cross-linked skeletal acto-S-1 complex at low ionic strength (11). In this complex S-1 is attached covalently to the NH 2 terminus of actin and is permanently bound to actin, so it is impossible to change the initial weak binding of S-1 to actin. Hence any inhibition observed must be due to inhibition of other steps in the ATPase cycle. We have observed that smooth muscle S-1 cross-linked to actin is also inhibited by calponin at low ionic strength and that this inhibition is abolished at higher ionic strength (data not shown) as it was for the reversible complex (Fig. 1). Caldesmon-tropomyosin and troponin-tropomyosin can also inhibit the carbodiimide cross-linked acto-S-1; once again this experiment emphasizes that calponin inhibits the same step in cross-bridge cycling as caldesmon and troponin.
Previously we have shown the specific interaction of calponin with actin Glu 334 (20). In their model of the actin-myosin interaction Rayment and co-workers (25,44,45) predicted the actin segment 332-334 as an important part of the actinmyosin interface and proposed this contact site to be part of the strong binding site in the actin-myosin complex (see Fig. 7). Our finding that calponin decreases exclusively the strong binding of S-1 to actin confirms the participation of this actin-S-1 contact site exclusively in the strong binding state. It is noteworthy that calponin is the only actin-binding protein that has been cross-linked to the putative strong binding site on actin. This seems to be fortuitous as only calponin has the right amino acid at the interface with actin to permit this cross-linking.
Our findings are also compatible with previous evidence indicating that the calponin binding region on actin does not overlap functionally with the weak binding sites of S-1 (see Fig.  7). Cleavage of actin at Met 47 -Gly 48 yields actin split into 9-and 36-kDa fragments and reduces the affinity of actin for myosin (46) but has no effect on the binding of calponin to actin (20). This supports the lack of influence of the NH 2 -terminal actin segment on the interaction of actin with calponin. In addition, modification of the actin NH 2 terminus by ethylenediamine decreased the binding of S-1⅐ADP⅐P i to actin (47) but had no effect on the binding of calponin to actin (11).
These results confirm and complete our view of the calponinactin interface (21). There are clearly two types of interactions; the inhibitory interaction is strongly salt-dependent and is competitive with myosin strong binding, whereas the other interaction is salt-independent. Our previous work (21) suggests that the calponin sequence VKYAEK at position 142-147 is an essential part of this interaction, but the sequence 148 -182 also binds to actin and could be responsible for calponin binding at higher salt concentrations. Although the actin sequence around Glu 334 probably binds the calponin inhibitory sequence, it is not clear where on actin the noninhibitory sequence of calponin binds; it cannot be in a position that blocks the weak binding site.
Although calponin inhibition seems to control the same step in the actomyosin pathway as troponin and caldesmon, it involves a different mechanism. Full inhibition of the acto-S-1 ATPase activity is achieved at a molar ratio to actin of 1:7 for troponin and 1:14 for caldesmon in the presence of tropomyosin (8), whereas the same inhibition is obtained at an equimolar ratio of calponin to actin (Fig. 2). Troponin and caldesmon bind to subdomain 1 on actin and block the weak binding site of myosin on the actin monomer involved in their attachment. However, their effect on the strong binding site of the 14 monomers for caldesmon and the 7 monomers for troponin is mediated and propagated by tropomyosin (6,8,9). Calponin acts directly upon the strong binding of S-1 at a ratio to actin of 1:1 and appears to be unique in that it specifically blocks only the strong site on actin. This property could make calponin a useful reagent to probe the cross-bridge cycle and its regulation.
It is not clear whether calponin plays a significant role in regulating smooth muscle actomyosin in vivo since its location, quantity, and phosphorylation pattern do not seem to be compatible with a physiological role (48 -51). In the present study we also show that at physiological ionic strength calponin has no inhibitory activity, and this is observed with both smooth and skeletal muscle myosin. It is possible that in smooth muscle the concentration of proteins and the presence of other regulatory proteins may modulate calponin's effect on the actomyosin ATPase. Nevertheless, our study of the mechanism of ATPase inhibition by calponin confirms the general principle of regulation in actomyosin systems, namely that the regulatory molecules control the weak to the strong transition in the cross-bridge cycle (38).