The Myosin Cardiac Loop Participates Functionally in the Actomyosin Interaction*

The motor protein myosin in association with actin transduces chemical free energy in ATP into work in the form of actin translation against an opposing force. Me-diating the actomyosin interaction in myosin is an actin binding site distributed among several peptides on the myosin surface including surface loops contributing to affinity and actin regulation of myosin ATPase. A structured surface loop on (cid:1) -cardiac myosin, the cardiac or C-loop, was recently demonstrated to affect myosin ATPase and was indirectly implicated in the actomyosin interaction. The C-loop is a conserved feature of all myosin isoforms with crystal structures, suggesting that it is an essential part of the core energy transduction ma-chinery. It is shown here that proteolytic digestion of the C-loop in (cid:1) -cardiac myosin eliminates actin-acti-vated myosin ATPase and reduces actomyosin affinity in rigor more than 100-fold. Studies of C-loop function in smooth muscle myosin were also undertaken using site-directed mutagenesis. Mutagenesis of a single charged residue in the C-loop of smooth muscle myosin alters actomyosin affinity and doubles myosin in vitro motility and actin-activated ATPase velocities, thereby involv-ing a charged region of the loop in the actomyosin interaction. It appears likely that the C-loop is an essential electrostatic in

The motor protein myosin in association with actin transduces chemical free energy in ATP into work in the form of actin translation against an opposing force. Mediating the actomyosin interaction in myosin is an actin binding site distributed among several peptides on the myosin surface including surface loops contributing to affinity and actin regulation of myosin ATPase. A structured surface loop on ␤-cardiac myosin, the cardiac or C-loop, was recently demonstrated to affect myosin ATPase and was indirectly implicated in the actomyosin interaction. The C-loop is a conserved feature of all myosin isoforms with crystal structures, suggesting that it is an essential part of the core energy transduction machinery. It is shown here that proteolytic digestion of the C-loop in ␤-cardiac myosin eliminates actin-activated myosin ATPase and reduces actomyosin affinity in rigor more than 100-fold. Studies of C-loop function in smooth muscle myosin were also undertaken using sitedirected mutagenesis. Mutagenesis of a single charged residue in the C-loop of smooth muscle myosin alters actomyosin affinity and doubles myosin in vitro motility and actin-activated ATPase velocities, thereby involving a charged region of the loop in the actomyosin interaction. It appears likely that the C-loop is an essential electrostatic binding site for actin involved in modulation of actomyosin affinity and regulation of actomyosin ATPase velocity.
The motor protein myosin is an ATPase and an actin-binding protein transducing ATP free energy into work. In muscle, myosin, actin, and ATP constitute the unitary work producer. The cyclical interaction of these components, a sequence of states each characterized as a static relation between the proteins and an intermediate in the degradation of ATP, is a contraction cycle. In a contraction cycle, myosin hydrolyzes ATP in its active site and forms a weak association with actin at a separate actin binding site. Release of phosphate from the active site initiates strong binding to actin and a large conformation change in myosin that produces work in the form of actin translation against a load. Several solved crystal structures represent intermediates in ATP hydrolysis and define myosin conformation changes in the absence of actin (1)(2)(3)(4)(5)(6). Presently, we lack a detailed understanding of actomyosin structure and in particular how the elements making up the actomyosin interface contribute to mutual affinity, actin regu-lation of phosphate release, and the ability of myosin-bound nucleotide to modulate actomyosin affinity.
Myosin consists of a globular head domain containing the enzymatic portion of the molecule and a tail involved in filament assembly. The globular head separated from its tail portion is called subfragment 1 (S1). 1 S1 interacts with filamentous actin (F-actin) consisting of polymerized actin monomers. An atomic model for F-actin built from monomeric actin crystal structures (7-9) satisfies x-ray diffraction data restraints (10) and is consistent with electron microscopic imagery (11). In acto-S1, S1 contacts two adjacent actin monomers in a filament (12). Protein-protein contacts in actomyosin were identified by experimental structural studies combined with simulated docking of the myosin and actin crystal structures (13)(14)(15)(16) and by detecting changes in actin binding strength, actin-activated myosin ATPase, and in vitro motility caused by the mutation of small peptide segments or individual residues in myosin (17)(18)(19)(20)(21)(22)(23)(24)(25). Primary actin contacts on S1, aa 529 -558 2 and aa 647-659, are helices maintaining hydrophobic interactions with actin. The unstructured surface Loop 2 (aa 626 -647) is another primary actin interaction site maintaining ionic interactions with the actin N terminus. Secondary sites are an unstructured surface loop (aa 567-578) and the structured myopathy loop (aa 404 -415) (13) also on the S1 surface.
Our previous work comparing tertiary structure of the skeletal and ␤-ventricular cardiac myosin isoforms identified a structured loop on the cardiac S1 (␤S1) surface (aa 362-376), called the cardiac loop or C-loop, having significant influence on energy transduction and possibly an interaction with actin (26). We characterized C-loop influence on energy transduction and actin binding using proteolytic digestion. Limited proteolysis of skeletal S1 cleaves the heavy chain at Loop 1 (aa 202-217) and Loop 2, producing 25-, 50-, and 20-kDa molecular mass fragments (27). These loops are involved in the regulation of substrate release (Loop 1) and in actin binding and regulation of actin-activated ATPase (Loop 2) (19,22,28,29). Limited proteolysis of ␤S1 cleaves the heavy chain at equivalent points and at the C-loop within the 50-kDa fragment (26). We found that C-loop cleavage dramatically affects ␤S1 Mg 2ϩ -ATPase, suggesting that the C-loop participates in energy transduction (26). Actin binding protects Loop 2 from proteolysis in skeletal S1 indicating Loop 2 involvement in actin binding (30,31). Actin binding to ␤S1 fails to inhibit Loop 2 cleavage (26, 32) but does inhibit C-loop cleavage (26). These observations indicate that ␤S1 and skeletal S1 differ in conformation and identify the C-loop as a possible actin binding site. The C-loop has recently been proposed as a site of actin binding in skeletal S1 (33), in myosin V (34), and in a class-I myosin, where it is called Loop 4 (35).
In studies described here, we characterize the significance of the C-loop/actin interaction by perturbing C-loop structure. We find that tryptic cleavage of the C-loop in ␤S1 eliminates actinactivated myosin ATPase and lowers actomyosin affinity. We also introduced mutations into the C-loop utilizing a smooth muscle myosin baculovirus-based eukaryotic expression system (36). The C-loop is a recognizable feature of smooth muscle myosin (5) and in all of the known S1 crystallographic structures, suggesting that its functional properties observed with ␤S1 are conserved in other myosins. Two different single residue mutants of smooth muscle heavy meromyosin (HMM) were selected to disrupt a hypothetical electrostatic interaction between actin and the C-loop and to perturb overall C-loop conformation. Each mutation caused distinctive changes in acto-HMM properties supporting the finding from the proteolysis studies that the C-loop interacts functionally with actin.
Preparation of Cardiac Myosin, ␤S1, and Skeletal Actin-␤-Cardiac myosin was prepared from bovine heart ventriculi and ␤S1 was prepared by digesting ␤-cardiac myosin filaments with ␣-chymotrypsin all as described previously (26). Skeletal G-actin was prepared from rabbit skeletal muscle acetone powder according to Pardee and Spudich (38). Protein concentrations were obtained with absorbance using an A (1%) at 280 nm of 5.33 for cardiac myosin and 7.45 for ␤S1, 6.5 for smooth muscle HMM, and A (1%) at 290 nm of 6.4 for actin. Protein concentrations were also estimated using the Bradford assay with bovine serum albumin as calibration. Molecular masses were assumed to be 500, 330, 115, and 42 kDa for ␤-cardiac myosin, smooth muscle HMM, ␤S1, and actin.
Smooth Muscle HMM Expression and Site-directed Mutagenesis-HMM wild type (WT) and mutant isoforms were expressed as described previously (39). Smooth muscle myosin cDNA encoding Met 1 -Ser 1110 flanked with SpeI/SalI was introduced into the polylinker region of pFastbac1 baculovirus transfer vector. A hexahistidine tag was added at the C-terminal side of Ser 1110 in aid of purification.
To express smooth muscle HMM, 200 ml of sf9 cells (about 1 ϫ 10 9 ) were co-infected with three separate viruses expressing the smooth muscle HMM heavy chain, essential light chain (ELC), and regulatory light chain (RLC). The cells were cultured at 28°C in 175-cm 3 flasks and harvested after 72 h. Cells were lysed with sonication in 40 ml of lysis buffer (30 mM Tris-HCl, pH 7.5, 0.3 M KCl, 5 mM MgCl 2 , 5 mM ATP, 5 mM 2-mercaptoethanol, 1 mM phenylmethanesulfonyl fluoride, 0.01 mg/ml leupeptin, 0.002 mg/ml pepstatin A, and 1 g/ml Sigma type IIS trypsin inhibitor). After centrifugation at 100,000 ϫ g for 30 min, the supernatant was mixed with 1.0 ml of Ni 2ϩ -nitrilotriacetic acid-agarose (Qiagen) in a 50-ml conical tube on a rotating wheel for 30 min at 4°C. The resin suspension was then loaded on a column (1 ϫ 10 cm) and was washed with a 10-fold volume of buffer containing 20 mM imidazole, pH 7.5, 0.3 M KCl, 0.1 mM EGTA, and 5 mM 2-mercaptoethanol. HMM was eluted with buffer containing 200 mM imidazole, pH 7.5, 0.3 M KCl, 0.1 mM EGTA, and 5 mM 2-mercaptoethanol. Fractions containing smooth muscle HMM were pooled and dialyzed against 20 mM Tris-HCl, pH 7.5, 30 mM KCl, and 5 mM 2-mercaptoethanol. The purified smooth muscle HMM was stored on ice or in a Ϫ80°C freezer after quick freezing with liquid nitrogen.
Proteolytic Digestion of ␤S1-␤S1 was digested with trypsin according to Bá lint et al. (27) and Muhlrad et al. (40). For the tryptic digestion, the reaction mixture included 60 mM KCl and 50 mM Tris-HCl, pH 7.0, with a final ␤S1 concentration of 2 mg/ml. All reactions were carried out at 25°C. The weight ratios of trypsin to ␤S1 varied as 1:50, 1:100, or 1:200 depending on the experiment. In experiments with actin, the ␤S1/actin molar ratio was 1:3. The proteolysis reaction was followed in time by removing aliquots from the digestion mixture at intervals after trypsin addition. Soybean trypsin inhibitor addition at a 2:1 (w/w) ratio with trypsin stopped the digestion. Quantitative analysis of digestion products was described previously (26).
ATPase and Actin-activated ATPase Assays-K ϩ EDTA-ATPase measurements were made on samples at 25°C (␤S1) or 37°C (HMM) from 0.5-ml aliquots containing 20 -40 g of ␤S1 or smooth muscle HMM, 2 mM ATP, 0.6 M KCl, 6 mM EDTA, and 25 mM Tris-HCl at pH 8.0. Inorganic phosphate production from the ATPase reaction was measured by the method of Fiske and Subbarow (41).
Native and digested ␤S1 actin-activated Mg 2ϩ ATPase activity was tested in a buffer containing 0.5-1.0 M ␤S1, 2 mM ATP, 2 mM MgCl 2 , 0.2 mM CaCl 2 , 60 mM KCl, 1 mM DTT, 20 mM TES buffer, pH 7.3, at 25°C. WT and mutant smooth muscle HMM actin activated Mg 2ϩ ATPase activity was tested in 300-l volumes containing 0.08 mg/ml HMM, 30 mM KCl, 2 mM MgCl 2 , 0.2 mM CaCl 2 , 0.5 mM DTT, 1 mM ATP, 30 mM TES, pH 7.6, at 25°C. HMM was phosphorylated by adding 15 g/ml chicken gizzard myosin light chain kinase and 10 g/ml bovine brain calmodulin to the assay medium 15 min prior to ATPase activity measurement. Unphosphorylated HMM actin-activated ATPase activity was measured in the presence of 1 mM EGTA replacing CaCl 2 , myosin light chain kinase, and calmodulin. Unphosphorylated WT and mutant smooth muscle HMM showed no ATPase activity activation in the presence of 190 M actin. Phosphorylated WT HMM showed Ͼ30fold maximal actin activation in agreement with previous results (42,43). We did not directly measure the extent of RLC phosphorylation. Inorganic phosphate production from actin-activated ATPase reactions in cardiac and smooth muscle myosin were measured by using the malachite green assay (44).
Actin Gliding Assay-The motile activities were measured by in vitro motility assay. A coverslip was first coated with nitrocellulose, and then anti-histidine tag monoclonal antibody (Qiagen) was applied to the coverslip. After blocking with 0.5 mg/ml bovine serum albumin, smooth muscle HMM was applied to the coverslip. Actin filament motility was observed in 25 mM imidazole (pH 7.4), 25 mM KCl, 2 mM MgCl 2 , 1 mM EGTA, 1 mM DTT, 18 g/ml catalase, 0.1 mg/ml glucose oxidase, 3.0 mg/ml glucose, 0.5% methylcellulose, and 4 mM Mg 2ϩ -ATP with an ATP regeneration system (20 units/ml pyruvate kinase and 3 mM phosphoenol pyruvate). Actin filament velocity was calculated from the movement distance and the elapsed time in successive snapshots. Student's t test was used for statistical comparison of mean values. A value of p Ͻ 0.01 was considered to be significant.
Pyrene-labeled Actin Experiments-Actin was labeled with pyrene iodoacetamide and treated with phalloidin (45,46). The actin binding assay was performed in rigor or in the presence of MgADP (MgADP state) as described previously (46 -48).
Actin binding experiments with ␤S1 were performed with 0.5 M actin, of which 64% was labeled with pyrene. The acto-␤S1 interaction was observed in 0.1 M KCl, 2 mM MgCl 2 , 1 mM DTT, and 20 mM HEPES, pH 7.3, at 20°C. Binding constants from the titration curves were surmised using a bimolecular reaction scheme modified to accommodate (i) distinguishable ␤S1 affinities for native and pyrene labeled actin and (ii) distinguishable actin affinities for ␤S1 fragmented by proteolytic digestion. These cases involved coupled reactions determining the equilibrium concentrations of distinguishable bound and dissociated species.
Actin binding experiments with smooth muscle HMM were performed with 0.2 M actin, of which 90% was labeled with pyrene. The acto-HMM interaction was observed in 0.1 M KCl, 2 mM MgCl 2 , 1 mM DTT, and 20 mM HEPES, pH 7, at 20°C in a volume of 150 l. The ϩADP conditions also contained 1 mM ADP pretreated with 20 M hexokinase and 2 mM glucose to remove ATP. Binding constants from the titration curves were surmised using Scheme 1 (49), where A 2 represents two adjoining actin monomers in F-actin, A 2 .HMM has one actomyosin bond, A 2 :HMM has two actomyosin bonds, K 0 is the actin binding constant for S1, and [c] is the equivalent concentration of S1 needed to give the binding constant for formation of the second actomyosin interaction within one HMM molecule. The overall equilibrium acto-HMM binding constant for Scheme 1, K H , is given by the following (47,49).
Geometrical arguments suggest that [c] is Ͻ100 M (49) and a value of ϳ90 M was measured for skeletal HMM (47). We found [c] of 90 M to be consistent with our observations. Correction to distinguish HMM interactions with native from pyrene-labeled actin is unnecessary due to the high degree of actin labeling used in these experiments. ADP binding to smooth muscle HMM was shown previously to induce asymmetry between the myosin heads for their actin binding affinity (50). In our experiments, ADP did not influence the total amplitude for pyrene-actin fluorescence change due to strong HMM binding, suggesting no measurable asymmetry between the heads, in agreement with Ellison et al. (51).

RESULTS
Tryptic Digestion of the C-loop in ␤S1-We showed previously that the tryptic digestion of skeletal S1 and ␤S1 splits the heavy chains at two points, Loops 1 and 2, producing the characteristic 25-, 50-, and 20-kDa fragments (32, 52) (Fig. 1). The 25-and 20-kDa fragments from ␤S1 are indistinguishable from the corresponding skeletal S1 fragments. Two new fragments were detected in the ␤S1 digest having molecular masses of ϳ30 and ϳ20 kDa. They were shown to be products of the ␤ 50-kDa fragment cleavage at Arg 371 in the C-loop. Tryptic digestion of ␤S1 with bound nucleotide (ATP or ADP) or trapped nucleotide analog ADP.vanadate (ADP.Vi) accelerates 30-kDa fragment production, indicating identical acceleration of C-loop cleavage. Digestion with bound or trapped nucleotide does not change the overall fragmentation pattern. These data show that the active site can influence C-loop conformation detected by its accessibility to proteolysis (26).  Fig. 3 (26). ELC in ␤S1 is involved in actin binding, and this peptide cleavage might affect acto-␤S1 affinity (53). ELC and Loop 2 cleavage occur simultaneously, so when referring to the effect of Loop 2 cleavage on acto-␤S1 affinity, summation of effects of both cuts is implied.
Effect of C-loop Cleavage on Actin-activated ATPase-Actin activates myosin ATPase by ϳ10 -100-fold compared with basal activity. Measured product release velocity, V, versus actin concentration, [A], are related by Michaelis-Menten kinetics, giving V ϭ V max [A]/(K m ϩ [A]) from which V max and K m are determined. Actin-activated myosin ATPase is a functional test of the efficiency of the actomyosin complex to generate force (54).
Effect of C-loop Cleavage on Strong Actin Binding Affinity-We investigated the effect of C-loop cleavage on actomyosin affinity using phalloidin-stabilized (46), pyrene-labeled F-actin. Competition between native actin and lower affinity pyrene-actin for binding ␤S1 produces sigmoidal binding curves (55). The latter effect was observed and accounted for quantitatively with different binding constants for the native and pyrene-labeled actin.
Site-directed Mutagenesis of Smooth Muscle Myosin-We investigated the influence of the C-loop influence on the actomyosin interaction using WT and two mutants of smooth muscle HMM. Table II indicates myosin sequences including the highly conserved glycine (Gly 362 ) and an arginine found at the tip of the C-loop in the skeletal (Arg 371 ) and smooth muscle (SmR370) S1 crystal structures (1,5). These residues are interesting candidates for study for different reasons. The glycine is proposed to facilitate backbone-mediated communication between the active site and the C-loop. A G362A mutant will have restricted backbone flexibility because of the alanine side chain. The arginine at SmR370 is proposed to bind actin electrostatically. An SmR370E mutant reverses the sign of a charged residue potentially interfering with electrostatic attraction to actin.
WT, SmR370E, and G362A HMM Properties in the Absence of Actin-The ATP-sensitive tryptophan fluorescence increments and K ϩ -EDTA ATPases for the unphosphorylated WT and mutant HMMs are compared in Table III. The ATP-sensitive tryptophan fluorescence increments are similar for each isoform and, for the WT, agree with previous measurements  Table III. Table III shows that phosphorylation increases V max for the WT HMM by ϳ36-fold, in agreement with published results (42,43). In the phosphorylated species, V max is identical for the WT and G362A species but increases ϳ2-fold in the SmR370E mutant. K m indicates weakened actin binding in the presence of ATP for both mutants compared with WT. The SmR370E mutant shows a much higher actin-activated ATPase sensitivity to phosphorylation of RLC and a larger V max , indicating accelerated transitioning from weak to strong actin binding (i.e. the force producing transition in a muscle fiber). These findings suggests a fundamental role for the Cloop in energy transduction. It appears that C-loop charge (perturbed by SmR370E) is critically important for determining V max .
b Proteolytic cleavage in the presence of actin produces a longer 20-kDa fragment (20.75 kDa in Fig. 2) designated 2Ј. Proteolytic cleavage in the absence of actin produces two lighter fragments (20.25 and 20 kDa in Fig. 2) but none of the 20.75-kDa peptide. We found no evidence that actin activated ATPase, and the strong binding of actin distinguishes digested ␤S1 formed from the various 20-kDa peptides.

TABLE II Myosin sequence comparison in the C-loop
Smooth muscle myosin mutants studied include G362A (Gly 362 indicated in boldface type) and SmR370E (Arg 371 or SmR370 indicated in italic type). Highly conserved residues are in boldface type.  Table III). Because SmR370E mutant showed 2-fold higher velocity and V max , it is anticipated that the mutation increases the rate-limiting phosphate release step in the cross-bridge cycle and the rate of the power stroke presumably coupled with the ADP off step (57). WT, SmR370E, and G362A HMM Strong Actin Binding-Strong actin binding to unphosphorylated HMM was measured using phalloidin stabilized, pyrene-labeled actin assay. Pyrene fluorescence versus myosin head concentration curves were fitted using Scheme 1 and provide the acto-HMM binding constant, K H (Equation 1). Computed values for K H are listed in Table IV for rigor and MgADP states of WT, SmR370E, and G362A smooth muscle HMM isoforms.
The highest acto-HMM affinity, K H ϭ 5.3 ϫ 10 8 M Ϫ1 , was observed for WT HMM in rigor. MgADP binding to the complex reduces K H by ϳ4-fold. Both values are in rough agreement with earlier work under comparable conditions (58). The large error associated with the rigor value is from the relative insensitivity of the binding curve to affinity constants Ͼ Ͼ10 8 M Ϫ1 in our experimental conditions. The C-loop HMM mutants have ϳ2-fold lower affinity for actin. MgADP binding to SmR370E and G362A reduce actin affinity by 2-and 3-fold, respectively. In the MgADP state, all isoforms have about equal actin affinity. DISCUSSION The discovery and the first functional analysis of myosin surface loops utilized their extreme sensitivity to limited proteolysis. The two proteolytic cutting points in skeletal S1 are within unstructured surface loops involved in the rate of substrate release and in regulation of actin activated ATPase (17)(18)(19). Lately, structured surface loops have figured in myosin motor function. The myopathy loop is a structured surface loop containing an unusual clustering of heart disease-implicated mutations, implying a functional significance for the loop (20,21,59). The myopathy loop is probably an actin binding site (13,16); however, the role of its disease-implicated point mutations in motor functions remains unclear (20,21,25,60). The C-loop is the second structured surface loop to be implicated in cardiac myosin motor function. Its significance was discovered by proteolysis of the cardiac myosin isoform, where it appears that the unique structure of the C-loop in ␤S1 permits its cleavage by trypsin (26). C-loop cleavage impacts myosin ATPases, suggesting that it is involved in energy transduction. Until now, no disease-implicated mutations were shown to reside within the C-loop sequence defined in Table II. The C-loop is located on the 50-kDa peptide of S1 adjacent to the myopathy loop and is a recognizable feature of the S1 surface in crystal structures for various myosin isoforms (Fig.  2). Fig. 7 shows the schematic layout of the myopathy and C-loops. C-loop sequence comparisons for several myosins with crystal structures (except ␤S1) shown in Table II indicate the a ATPases are per myosin head. b ATPases were measured from phosphorylated (P) or unphosphorylated (U) HMM. The P/U ratio compares V max for phosphorylated and unphosphorylated species. Errors are S.D. and when not indicated are ϳ10%.  highly conserved residues of the myosin family. The C-loop begins at a highly conserved glycine (Gly 362 ) on the end of a helix-turn-helix segment originating from the vicinity of the ATP binding site. Stabilizing C-loop structure are peptide backbone hydrogen bonds spanning the neck of the loop usually located near the highly conserved Phe 366 . In skeletal and ␤S1, a salt bridge between Lys 369 and Glu 372 also stabilizes loop structure. The ␤S1 C-loop is unique for its sensitivity to proteolysis. The crystallized myosin isoforms have differing detailed C-loop structure, although all are qualitatively similar in appearance except for myosin-IE, where the C-loop is much larger (35). The C-loop connects to the myopathy loop via a small helix-turn-helix segment (only one helix segment is shown in Fig. 7). The myopathy loop/C-loop joining segment contains several mutations implicated in heart disease (59).
We have characterized C-loop interactions with actin and its participation in energy transduction with biochemical and molecular biological approaches. The biochemical approach makes use of the sensitivity of the C-loop in ␤S1 to limited proteoloysis. Earlier work showed that C-loop cleavage is accelerated by saturation of the active site with ATP or its analogues (26). Fig.  3 shows how actin binding completely inhibits C-loop cleavage. Finally, C-loop cleavage eliminates actin-activated ATPase ( Fig. 4 and Table I) and dramatically attenuates actin affinity (Table I). These results strongly suggest the C-loop structure is influenced by the substrate in the active site and that it is involved in actin binding and regulation of actin-activated ATPase. However, we were unable to cleave the C-loop without cutting other trypsin-sensitive points in ␤S1 (Loops 1 and 2 and the ELC), making it impossible to be certain the demonstrated effects are due solely to disruption of the C-loop. Fortunately, the molecular biological approach of site-directed mutagenesis permits selective perturbation of the C-loop.
Glycine residues facilitate larger secondary structure changes due to their ability to freely swivel (61). In myosin, the highly conserved Gly 699 was shown to be a pivot vital to functionality by using the Gly-Ala substitution where alanine inhibits swiveling by the steric clash of its side chain with the peptide backbone (62). When the peptide backbone serves as a line of communication between distant sites, intervening highly conserved glycines are logical targets for Gly-Ala mutagenesis. The highly conserved Gly 362 residue fits the profile for facilitating communication between the active site and the C-loop, suggesting the G362A substitution might disrupt active site/C-loop communication. A trypsin cleavage point in Loop 2 is in a cluster of charged residues facilitating electrostatic actin binding. The tip of the C-loop at Arg 371 (Table II) is the trypsinsensitive point in ␤S1 also in a cluster of charged residues potentially serving as an electrostatic actin binding site. The replacement of the positively charged Arg 371 (SmR370) with a negatively charged glutamic acid (SmR370E) might then disrupt an electrostatic interaction with actin.
Chimeric loop substitution into a host myosin proved useful for detecting myosin isoform differences in Loop 2 function (17). The fundamental assumption of myosin structure/function research is that the functional significance of key structural elements of myosin are preserved across myosin isoforms, suggesting that chimeric substitutions into dicty or smooth muscle myosin hosts are useful model systems when expression for the original myosin (in this case ␤S1) is problematic (19,25,63). We apply the fundamental assumption to the C-loop and the smooth muscle myosin isoform model, although we acknowledge that in some cases mutagenized ␤S1 behaves unlike its mutagenized model (20,21,25). Recently, various arguments have been advanced to rationalize these differences while maintaining the validity of the fundamental assumption for the myopathy loop (60).
The mutagenesis data are summarized in Figs. 5 and 6 and Tables III and IV. In the absence of actin, the mutant constructs are similar to the WT protein (Table III). The K ϩ -EDTA ATPase is mildly inhibited in the G362A mutant, suggesting a slight perturbation of structure. The ATP-sensitive tryptophan fluorescence follows a similar trend to the K ϩ -EDTA ATPase, with the G362A mutant most affected, although the data indicate that all isoforms have robust fluorescence enhancement upon ATP binding. Differences between mutant and WT proteins are amplified by introduction of actin. Actinactivated myosin ATPase ( Fig. 5 and Table III) indicates both smooth muscle HMM constructs bind actin weaker than the WT protein in the presence of ATP. Only the SmR370E construct has a V max different (elevated) from the WT, showing that the substitution affects an additional aspect of actin regulation of ATPase. In vitro motility (Fig. 6) likewise identifies the SmR370E construct as unique for its ability to move actin at twice the velocity of WT and G362A. In the motility assay, WT and G362A behave identically. The SmR370E construct more rapidly converts from the weak to strong actin binding states when compared with either the WT or G362A isoforms.
SmR370E and G362A construct actin binding affinities in rigor (Table IV) are reduced ϳ2-fold compared with WT. All HMM species bind actin with equal affinity in the presence of MgADP. The MgADP state actin affinity is diminished compared with rigor, and thus the mutants show diminished MgADP modulation of actin binding affinity. The latter suggests native C-loop involvement in the ability of MgADP to modulate strong binding affinity in myosin. Related to this ability to modulate strong binding is an intact active site/C-loop communication pathway. We proposed that the G362A substitution might preferentially disrupt active site/C-loop communication; however, strong actin binding sensitivity to active site-bound MgADP is about the same for each mutant. Thus by every test, utilizing both proteolysis and site-directed mutagenesis, perturbation of C-loop structure leads to a reduction in FIG. 7. Schematic two-dimensional structure of the ␤S1 backbone in the region of the C-loop and myopathy loop adapted from the skeletal myosin structure (1). Gly 362 and Phe 366 are highly conserved residues in the myosin family. Arg 371 is proposed to be involved in an electrostatic interaction with actin, and Glu 372 and Lys 369 form a salt bridge stabilizing C-loop structure. actomyosin affinity. In addition to this general affect, the SmR370E replacement disrupts the normal motility and actin activation of myosin ATPase, implying the charge reversal of the Arg/Glu substitution is the more structurally specific effect. It appears likely that the C-loop is an electrostatic binding site for actin, a regulator of strong actin binding affinity in the presence of nucleotide, and a regulator of actin-activated myosin ATPase in the manner of Loop 2. The suggestion that the G362A substitution selectively disrupts active site/C-loop communication is not substantiated by our findings. We now propose that the ability of the G362A construct to reduce myosin's affinity for actin is a general effect connected to disruption of native C-loop structure rather than to a residuespecific effect.
The active site/actin-binding site communication in myosin is bidirectional, with nucleotide binding able to modulate actin affinity and actin binding able to accelerate product release from the active site. Bound ATP or trapped ADP.Vi accelerates C-loop proteolysis in ␤S1, demonstrating active site influence on C-loop structure (26). Conversely, the doubled V max and motility velocity in the SmR370E isoform demonstrates C-loop influence on the active site. Changes in V max and motility velocity in the SmR370E isoform suggest the C-loop is an actin contact detector communicating actin proximity to the active site. The enhanced actin activation in the SmR370E isoform was not duplicated in the G362A isoform, suggesting that communication of actin proximity to the active site does not flow via the shortest through-backbone route (from the C-loop, through Gly 362 , toward the N terminus). Instead we propose that actin proximity detected at the C-loop is transmitted toward the C terminus to the vicinity of the Glu 468 side of the "back door" salt bridge known to facilitate phosphate release (64,65). This communication pathway includes residue Arg 403 in the myopathy loop. Arg 403 is another residue known to up-or down-regulate V max and motility velocity (60). These neighboring structured loops may work in concert to regulate actinactivated ATPase.