High Affinity Ca2+ Binding Sites of Calmodulin Are Critical for the Regulation of Myosin Iβ Motor Function*

We coexpressed myosin Iβ heavy chain with three different calmodulin mutants in which the two Ca2+-binding sites of the two N-terminal domain (E12Q), C-terminal domain (E34Q), or all four sites (E1234Q) are mutated in order to define the importance of these Ca2+binding sites to the regulation of myosin Iβ. The calmodulin mutated at the two Ca2+ binding sites in N-terminal domain and C-terminal domain lost its lower affinity Ca2+ binding site and higher affinity Ca2+ binding site, respectively. We found that, based upon the change in the actin-activated ATPase activities and actin translocating activities, myosin Iβ with E12Q calmodulin has the regulatory characteristics similar to myosin Iβ containing wild-type calmodulin, while myosin Iβ with E34Q or E1234Q calmodulin lose all Ca2+ regulation. While the increase in myosin Iβ ATPase activity paralleled the dissociation of 1 mol of calmodulin from myosin Iβ heavy chain for both wild type (abovepCa 5) and E12Q calmodulin (above pCa 6), the Ca2+ level required for the inhibition of actin-translocating activity of myosin Iβ was lower than that required for dissociation of calmodulin, suggesting that the conformational change induced by the binding of Ca2+ at the high affinity site but not the dissociation of calmodulin is critical for the inhibition of the motor activity. Our results suggest that the regulation of unconventional myosins by Ca2+ is directly mediated by the Ca2+ binding to calmodulin, and that the C-terminal pair of Ca2+-binding sites are critical for this regulation.

tissues with the highest expression levels in heart, lung, adrenal gland, esophagus, and stomach (6,7,9,10). Myosin I␤ localizes to actin-rich peripheral structures, such as filopodia and lamellipodia of culture cells (9), and it is thought to play a role in cytoskeleton rearrangement. Interestingly, myosin I␤ is also found in hair bundles purified from the bullfrog sacculus, suggesting that myosin I␤ may function as an adaptation motor which regulates the tip link-associated cation selective channels (11,12).
Studies from both naturally isolated and recombinant myosins I␤ have shown that calmodulin is associated with myosin I␤ heavy chain (5,13). In contrast to most calmodulindependent enzymes, the association of calmodulin with myosin I does not require Ca 2ϩ binding to calmodulin. Thus, calmodulin functions as a light chain subunit. Myosin I␤, like all the other vertebrate unconventional myosins, has several repeats of a 24 -30-amino acid sequence called the IQ motif at the neck region between the myosin head motor domain and the tail domain. This motif has been suggested to provide the binding site for EF-hand family proteins such as calmodulin (1,14). All vertebrate unconventional myosins that have been characterized so far contain calmodulin as light chains. For some unconventional myosins, other small proteins besides calmodulin have also been found to function as light chains (15). A common property of calmodulin targets that contain IQ motif, such as the unconventional myosins and neuromodulin, is that they have a higher affinity for the Ca 2ϩ -free form of calmodulin (16).
Calmodulin is one of the major intracellular Ca 2ϩ -sensor proteins, containing four EF-hand type Ca 2ϩ -binding loops. The N-terminal pair are linked to the C-terminal pair by a central flexible linker. Among the four Ca 2ϩ -binding sites, the C-terminal pair of Ca 2ϩ -binding sites have a higher Ca 2ϩbinding affinity than those of the N terminus (17,18). Ca 2ϩ binding induces structural changes in calmodulin, and it is believed that these Ca 2ϩ -induced conformational changes allow calmodulin to activate target enzymes when the cytosolic Ca 2ϩ concentration is elevated (reviewed in Ref. 19). Mutagenesis studies have shown that a conserved glutamic acid residue at the 12th position of each Ca 2ϩ -binding loop is critical for Ca 2ϩ binding, and substitution of this conserved glutamic acid with glutamine in each Ca 2ϩ -binding site abolishes its Ca 2ϩ binding ability (17,20).
Both the enzymatic and mechanical activities of vertebrate myosin I have been shown to be regulated by Ca 2ϩ (5,13,(21)(22)(23), and this is also true for other unconventional myosins, such as myosin V (15). A member of another myosin I subclass, intestinal brush-border myosin I (BBMI), 1 has been extensively characterized in terms of Ca 2ϩ effects. It was found that BBMI moved actin filaments although the velocity was quite low (ϳ0.05 m/s), and the activity was abolished at Ca 2ϩ concentrations above 5 M. On the other hand, actin-activated Mg 2ϩ -ATPase activity of BBMI increased with increasing Ca 2ϩ concentrations (24). Interestingly, partial dissociation of calmodulin from BBMI was observed at a Ca 2ϩ concentration of 10 M. Similarly, Zhu et al. (13) have shown that one of the three calmodulin molecules bound to recombinant myosin I␤ dissociated from the heavy chain at a Ca 2ϩ concentration of 10 M. While Mg 2ϩ -ATPase activity increased above pCa 6, actin sliding velocity of myosin I␤ was abolished at pCa 6 (13). These results suggest that Ca 2ϩ binding to the calmodulin light chains is critical for the regulation of vertebrate myosin I motor function. However, it is unclear whether the dissociation of calmodulin is necessary to stop myosin I motor activity since the Ca 2ϩ concentration required for the dissociation of one calmodulin from myosin I␤ heavy chain seems to be higher than the Ca 2ϩ concentration for inhibition of the actin translocating activity (13).
In the present study, we have examined the mechanism by which Ca 2ϩ and calmodulin regulate myosin I␤ motor function by coexpressing various calmodulin mutants defective in Ca 2ϩ binding with myosin I␤ heavy chain and analyzing the actinactivated ATPase activity and motor function of the expressed myosin I␤.

Expression of Bovine Myosin I␤ Together with Mutant Calmodulin in
Sf9 Cells-Expression of bovine myosin I␤ cDNA with calmodulin was performed as described previously (13). cDNA for wild-type calmodulin and calmodulin with the two N-terminal, two C-terminal, or all four Ca 2ϩ -binding sites mutated, termed E12Q, E34Q, and E1234Q, respectively (25), were subcloned into pBlueBacM baculovirus transfer vector at the EcoRI site in the polylinker region. Orientation and accuracy of the subcloning were examined by DNA sequencing (Sequenase 2.0, U. S. Biochemical Corp.). Recombinant baculoviruses containing these mutant calmodulin cDNAs were obtained by blue plaque selection and subsequent steps of purification and amplification as described in the manual from Invitrogen, MaxBac Baculovirus Expression System. Sf9 cells were coinfected with the recombinant viruses of myosin I␤ heavy chain and each calmodulin mutant.
Purification of Recombinant Myosin I␤ with Mutant Calmodulin-The purification of recombinant myosin I␤ was performed as previously with slight modifications (13). Briefly, cells were harvested after 3 days of culture at 28°C and lysed in the presence of ATP, Triton X-100, Nonidet P-40, and various protease inhibitors. The supernatant (150,000 ϫ g for 30 min) of lysed cells was incubated with 10 mM glucose and 20 units/ml hexokinase at 0°C for 30 min to completely hydrolyze residual ATP. F-actin (1 mg/ml) was added to coprecipitate the expressed myosin I␤. The pellet was resuspended with buffer containing 5 mM MgCl 2 , 100 mM KCl, 1 mM EGTA, and 25 mM Tris-HCl, pH 7.5, and 1 mM ATP was added to release myosin I␤ from the myosin I-actin complex. The sample was ultracentrifuged at 150,000 ϫ g for 30 min to remove F-actin, and the supernatant containing the expressed myosin I␤ heavy chain and calmodulin was subject to a DE52 column (1 ϫ 10 cm). The protein was eluted with a linear gradient (12 ml-12 ml) of 50 -250 mM KCl. Approximately 100 g of myosin I␤ can be obtained from 800 ml of culture.
Expression and Purification of Calmodulin Mutants-Wild-type calmodulin and mutant proteins were expressed by infecting Sf9 cells with recombinant virus. Cells were homogenized with 5 volumes of buffer A (30 mM Tris-HCl, pH 7.5, 8 M urea, 5 mM dithiothreitol, 10 g/ml leupeptin) for 5 min. After centrifugation at 35,000 ϫ g for 15 min, 5% trichloroacetic acid was added to the supernatant. The pellets were collected and suspended with 8 M urea, and the pH was adjusted to be neutral. The suspension was dialyzed against buffer B (30 mM Tris-HCl, pH 7.5, 1 mM dithiothreitol) for 2 h. After insoluble material was removed by centrifugation, the sample was loaded to a DE52 ion exchange column (24 ϫ 3 cm) equilibrated with buffer B. Calmodulin was eluted with a 0 -0.4 M NaCl linear gradient. The fractions containing calmodulin were combined, concentrated, and loaded onto a Sephacryl S200 gel filtration column (100 ϫ 3 cm). The fractions containing calmodulin were identified by SDS-polyacrylamide gel electrophroesis analysis. They were combined and concentrated. The concentrated cal-modulin was dialyzed against buffer C (30 mM Tris-HCl, pH 7.5, 30 mM NaCl, 1 mM dithiothreitol) and stored in Ϫ80°C. Bacterially expressed wild type and mutant calmodulins were prepared as described previously (25).
Fluorimeter Titrations-Ca 2ϩ -induced conformational changes in calmodulins were determined by Ca 2ϩ titrations in the presence of 9-anthroylcholine (9-AC) according to Beckingham with some modifications (20). Purified calmodulin was dialyzed extensively against the titration buffer (0.5 mM CaCl 2 , 100 mM KCl, and 10 mM MOPS, pH 7.0). 2-ml titration buffer containing 5 M of such prepared calmodulin and 20 M of 9-AC bromide (Molecular Probes, Eugene, OR) were transferred to a quartz cuvette, and fluorescence measurements were made as sequential addition of small volume (1 to several microliters) EGTA solutions (20 mM, 200 mM, or 1 M). The amount of EGTA was added so that the concentrations of free Ca 2ϩ in the cuvette were the desired ones (26). The fluorescence intensity was measured by using a Spex Fluorolog fluorimeter (Spex Inc., Edison, NJ) with excitation at 366 nm and emission at 418 nm. The pH of the samples was examined both before and after titration.
ATPase Assays and Other Biochemical Procedures-The effect of Ca 2ϩ on actin-activated ATPase activity was assayed in 30 mM KCl, 2 mM MgCl 2 , 20 mM imidazole-HCl, pH 7.0, with or without 50 M actin, in the presence of 1 mM EGTA or various concentrations of Ca 2ϩ (26). All assays were initiated by adding 100 M [␥-32 P]ATP (Amersham Corp.) to the reaction mix. The liberated 32 P was measured as described previously (27) to determine ATPase activity.
The effect of Ca 2ϩ on the binding of calmodulin to myosin I␤ heavy chain was determined as follows: myosin I␤ was dialyzed against different Ca 2ϩ buffers in the absence of ATP, and the sample was ultracentrifuged at 150,000 ϫ g for 30 min in the presence of 1 mg/ml F-actin. Then the pellet was analyzed by SDS-polyacrylamide gel electrophroesis. The amount of the cosedimented myosin I␤ and calmodulin was determined by densitometry, using a Macintosh computer equipped with a frame grabber (LG-3, Scion Corp., Walkersville, MD) connected to a video camera (Pulnix, TM-745, Motion Analysis, Inc., Eugene, OR), and the software used was NIH Image 1.56b18 (Bethesda, MD).
All experiments were carried out at 25°C, and all the results are represented by mean Ϯ S.E.
In Vitro Motility Assay-In vitro motility assays were performed as described previously (28). A larger nitrocellulose-coated coverslip (24 ϫ 30 mm) was covered by smaller coverslip (18 ϫ 18 mm), each edge of which was applied with 0.1 g of silicon grease (Dow Corning, MI), to create a fluid-filled flow cell. After 60 l of myosin I-containing solution was applied to the flow cell, unbound myosin I was washed away with 180 l of buffer A (400 mM KCl, 25 mM HEPES, pH 7.5, 4 mM MgCl 2 , and 10 mM dithiothreitol). Then the unoccupied nitrocellulose surface was coated with 0.5 mg/ml bovine serum albumin in buffer B (30 mM KCl, 20 mM HEPES, 1 mM EGTA, pH 7.5). The flow cell was washed with buffer B, then rhodamine-phalloidin (Molecular Probes Inc., Eugene, OR) labeled actin filaments in the motility buffer (50 mM KCl, 5 mM MgCl 2 , 25 mM imidazole, 1 mM EGTA, 1% 2-mercaptoethanol, 0.5% methylcellulose, 4.5 mg/ml glucose, 216 g/ml glucose oxidase, 36 g/ml catalase) with various concentrations of Ca 2ϩ were introduced onto the myosin-coated coverslip. After 120 l of motility buffer were perfused to wash out unbound actin filaments, motility buffer containing Mg 2ϩ -ATP was applied to initiate the reaction. Movements of fluorescent actin filaments were observed using an inverted fluorescence microscope (Diaphot, Nikon) with a SIT camera (VE 1000 SIT, DAGE MTI) and a video cassette recorder. The actin sliding velocity was determined as described previously (13).

RESULTS
Three different calmodulin Ca 2ϩ binding site mutants, termed E12Q, E34Q, and E1234Q (25), were used in this study. In each protein, particular Ca 2ϩ -binding sites were inactivated by mutation of a conserved glutamate residue (at position 12 of the Ca 2ϩ -binding loops) to glutamine. In E12Q the two Nterminal sites were mutated, in E34Q the two C-terminal sites were mutated, and in E1234Q, all four sites carried the mutation.
Wild-type calmodulin and each of the calmodulin mutants was coexpressed with myosin I␤ heavy chain in insect Sf9 cells, and the expressed myosin I␤ was isolated. All of the calmodulin mutants copurified with myosin I␤ heavy chain, suggesting that the mutations do not affect the binding of calmodulin to myosin I␤ heavy chain. The ability of Ca 2ϩ to increase electrophoretic mobility, a characteristic of wild-type calmodulin, was examined for each of these three mutants (Fig. 1). Wild type calmodulin migrated with an apparent molecular mass of 16 kDa in 5 mM Ca 2ϩ but with an apparent molecular mass of 21 kDa in the presence of 1 mM EGTA. E12Q calmodulin mutant migrated at 22 kDa in EGTA but at 19 kDa in the presence Ca 2ϩ . On the other hand, E34Q migrated at 21 and 18 kDa in the absence and presence of Ca 2ϩ , respectively. The mobility shift by Ca 2ϩ was abolished with E1234Q calmodulin mutant, which migrated at 21 kDa under both conditions. These results confirm that the calmodulin light chain associated with the myosin I␤ heavy chain in each preparation is indeed the expressed recombinant calmodulin mutant, and not endogenous calmodulin. They also suggest that the effect of mutating the two N-terminal Ca 2ϩ binding sites on the conformational change of calmodulin is different from that of mutating the two C-terminal sites.
Ca 2ϩ -induced conformational changes in the calmodulin mutants were further monitored as a function of Ca 2ϩ concentration by use of the reporter molecule 9-AC bromide. The Ca 2ϩinduced appearance of hydrophobic sites on calmodulin is revealed by the enhanced fluorescence of 9-AC upon binding to these sites, and this technique has been used previously to examine the Ca 2ϩ binding and conformational properties of Ca 2ϩ binding site mutants of calmodulin (20). The 9-AC fluorescence enhancement for wild type calmodulin as a function of Ca 2ϩ concentration is shown in Fig. 2. A single transition is detected, with midpoint at a Ca 2ϩ concentration (10 Ϫ7 ) that is lower than the dissociation constant for the high affinity sites on calmodulin. This finding probably reflects increased overall affinity for Ca 2ϩ and increased cooperativity of Ca 2ϩ binding induced by 9-AC. The hydrophobic reporter 1-anilino-8-naphtalene sulfonate has been shown previously to increase the affinity of calmodulin for Ca 2ϩ (29). The curve for the E12Q (see Fig. 2) is very similar to the wild type curve at low Ca 2ϩ concentrations but shows no increase in fluorescence enhancement at Ca 2ϩ concentrations above pCa 6. This is consistent with induction of a conformational change as a result of Ca 2ϩ binding to the two intact C-terminal sites present on this protein followed by absence of Ca 2ϩ binding and the associated conformational change in the N-terminal domain. In contrast, the E34Q mutant shows no fluorescence enhancement at low Ca 2ϩ and relatively minor enhancement of fluorescence with a midpoint at about pCa 6 as Ca 2ϩ levels are increased. Thus a major conformational change normally associated with C-terminal high affinity sites is lost in the E34Q mutant leaving a smaller conformational change associated Ca 2ϩ binding in the intact N-terminal domain. The sum of the fluorescence changes for E12Q and E34Q equals the changes for the wild type calmodulin (Fig. 2). The E1234Q mutant showed no change in 9-AC fluorescence throughout the entire pCa range tested (data not shown).
The actin-activated Mg 2ϩ -ATPase activity of myosin is coupled to actomyosin cross-bridge turnover. In order to examine the effect of Ca 2ϩ binding at the N-and C-terminal sites of calmodulin on myosin I␤ mechanoenzymatic function, actinactivated Mg 2ϩ -ATPase activity of myosin I␤ containing the mutant calmodulins was measured as a function of Ca 2ϩ (Fig.  3). For all assays, the timecourse of P i liberation was determined and the activity was estimated from the slope of the P i release timecourse. Unlike myosin I␤-containing wild-type calmodulin, for which Mg 2ϩ -ATPase activity increased in both the absence and presence of F-actin with Ca 2ϩ concentration above pCa 6 (Fig. 3A), myosin I␤-containing mutant calmodulin showed different Ca 2ϩ dependencies. The ATPase activity of E12Q myosin I␤ started to increase below pCa 6 and reached maximum at pCa 6 ( Fig. 3B). On the other hand, the ATPase activities of myosin I␤ containing both E34Q and E1234Q mutants did not show any Ca 2ϩ dependence (Fig. 3, C and D). The activity of myosin I␤ containing these two mutant calmodulins was similar to that of myosin I␤ with wild-type calmodulin in EGTA, suggesting that Ca 2ϩ binding to calmodulin in the Cterminal domain induced the enhancement of ATPase activity. These data suggest that the Ca 2ϩ -dependence of actin-activated ATPase activity of myosin I␤ is mediated through Ca 2ϩ binding to calmodulin and that the two C-terminal Ca 2ϩ -binding sites (high affinity sites) have a more important role in the Ca 2ϩ sensitivity of myosin I␤ ATPase activity.
To access the effects of mutating the N-and C-terminal Ca 2ϩ -binding sites of calmodulin on myosin I␤ motor activity, actin sliding velocity was measured by the in vitro motility assay system (Table I) with the one reported in the previous study (13). While switching motility buffer from 1 mM EGTA to 1 or 10 M Ca 2ϩ abolished the actin filament movement for wild-type and E12Q myosin I␤, it had little effect on the motility activity of E34Q and E1234Q myosin I␤. These results suggest that the Ca 2ϩ regulation of motor activity of myosin I␤ is also mediated through the binding of Ca 2ϩ to calmodulin. They further demonstrate that the two C-terminal Ca 2ϩ -binding sites but not the two N-terminal sites are critical for this regulation.
The effects of Ca 2ϩ on the binding of the various calmodulins to myosin I␤ heavy chain were also examined (Fig. 4). Purified myosin I␤ containing each calmodulin mutant was coprecipitated with F-actin at various Ca 2ϩ concentrations, and the precipitated myosin I␤ and calmodulin were subjected to SDSpolyacrylamide gel electrophroesis followed by densitometry analysis to quantify the stoichiometry of the bound calmodulin. It is known that 3 mol of calmodulin bind to 1 mol of myosin I␤ heavy chain (13). For wild-type calmodulin, one of the three molecules of bound calmodulin was dissociated from the heavy chain above pCa 5 (Fig. 4A). On the other hand, for E12Q myosin I␤, the dissociation of calmodulin was observed at lower Ca 2ϩ (i.e. pCa 6) (Fig. 4B). In contrast, Ca 2ϩ had no effect on the binding of E34Q calmodulin, i.e. all three calmodulin molecules were associated with the heavy chain even at pCa 4 (Fig.   4C). As expected, the binding of E1234Q mutant calmodulin to the heavy chain showed no Ca 2ϩ sensitivity (Fig. 4D).

DISCUSSION
In this study, we have coexpressed myosin I␤ heavy chain with three different calmodulin mutants, in which the conserved critical glutamic acid residue at the 12th position of the two N-terminal, two C-terminal, or all four of the Ca 2ϩ binding loops were substituted by glutamine. The importance of the carbonyl side chain of this glutamic acid to the Ca 2ϩ coordination system was revealed by crystallographic studies of calmodulin (30). As expected, the Ca 2ϩ -binding abilities of the mutated sites are completely abolished, based upon the Ca 2ϩ binding induced conformational changes probed by the fluorescent hydrophobic reporter molecule, 9-AC. The results are consistent with the earlier study in which the conserved glutamic acid residue in each individual Ca 2ϩ -binding site was mutated (20).
The fluorescence titration results show that the two C-terminal Ca 2ϩ -binding sites have a higher affinity for Ca 2ϩ than the two N-terminal sites. The different electrophoretic mobility of calmodulin in the presence of EGTA as compared with Ca 2ϩ also reflects conformational changes upon Ca 2ϩ occupation. The obtained results suggest that the conformational change caused by Ca 2ϩ binding to the two N-terminal sites (E34Q) is different from that caused by Ca 2ϩ binding to the two Cterminal sites (E12Q); occupation of the two C-terminal Ca 2ϩbinding sites has a greater impact on the overall conformational change. This is consistent with the previous findings that the Ca 2ϩ binding at the C-terminal sites of calmodulin induces a larger conformational change (31). Recent structural studies suggest that the C-terminal domain of calmodulin exists in a semi-open conformation in contrast to the close conformation of the N-terminal domain in the absence of Ca 2ϩ , and it changes to an open conformation upon Ca 2ϩ binding (32).   concentration are mediated through calmodulin, and that its two high affinity C-terminal Ca 2ϩ -binding sites are critical for the regulatory effect of Ca 2ϩ .
For myosin I␤ associated with E12Q, the effects of Ca 2ϩ on actin-activated ATPase activity, motor activity, and the calmodulin dissociation are similar to those for myosin I␤ associated with wild-type calmodulin. On the other hand, the disruption of the Ca 2ϩ -binding sites in the C-terminal domain of calmodulin abolishes the Ca 2ϩ dependence of the ATPase activity, the motor activity of myosin I␤, and the dissociation of 1 mol of calmodulin from the heavy chain at above pCa 6. These results suggest that Ca 2ϩ binding to the C-terminal domain of calmodulin, i.e. high affinity Ca 2ϩ binding sites, is responsible for the dissociation of one calmodulin molecule from myosin I␤ heavy chain. The increase in ATPase activity parallels the dissociation of one molecule of calmodulin according to the Ca 2ϩ dependence data (Fig. 4), thus it is reasonable to conclude that the dissociation of calmodulin increases myosin I␤ ATPase activity. It should be noted, however, actin independent ATPase activity but not actin dependent activity increases with Ca 2ϩ . Similar result has also been found for conventional myosin in which the dissociation of regulatory light chain increases basal myosin ATPase activity (33). It should be noted that while the deletion of the Ca 2ϩ binding at the N-terminal domain of calmodulin did not prevent the dissociation of 1 mol of calmodulin from the heavy chain, it shifted the Ca 2ϩ required for calmodulin dissociation to lower concentration. This result suggests that there is a cross-talk between the N-terminal and C-terminal domains and that deletion of the Ca 2ϩ binding ability at the N-terminal domain of calmodulin affects the conformational change at the C-terminal domain of calmodulin. This is consistent with the earlier finding that conformation of the Tyr-138 in the C-terminal domain of calmodulin is significantly influenced by a change in the Ca 2ϩ binding properties of the N-terminal domain (25).
Clearly, the binding of Ca 2ϩ to calmodulin at the C-terminal sites is critical for the inhibition of actin translocating activity of myosin I␤ by Ca 2ϩ , since this Ca 2ϩ -induced inhibition of the motility was not observed with myosin I␤ containing E34Q or E1234Q calmodulin (Table I). However, it is more complicated to determine whether or not the dissociation of the calmodulin molecule from myosin I␤ heavy chain is critical for the inhibition of the motor activity. Thus, although the wild-type myosin I␤ still binds all three calmodulin light chains at pCa 6, its motor activity is completely inhibited at this Ca 2ϩ concentration. One possible explanation is that although the C-terminal domain of the one calmodulin molecule is dissociated from myosin I␤ heavy chain at pCa 6, the N-terminal domain is still associated with the heavy chain at this Ca 2ϩ concentration, and further conformational change induced by Ca 2ϩ binding at the N-terminal low affinity sites of this molecule is necessary for the complete dissociation (Fig. 5). Presumably, this incomplete association of calmodulin with myosin I␤ is no longer able to support motor activity. For the E12Q mutant, the conformational change induced by the binding of Ca 2ϩ to the higher affinity sites may be sufficient to dissociate calmodulin from myosin I␤ heavy chain. It should be noted, however, Ca 2ϩ binding to calmodulin dissociates only one of the three bound calmodulin from myosin I␤ heavy chain. According to the amino acid sequence, myosin I␤ has three IQ motifs, one of which is not a completely matched IQ motif, IQXXXRGXXXR (one-letter amino acid code; X is any amino acid residue) (6). It is plausible that the calmodulin bound to the incomplete IQ motif is dissociated from myosin I␤ when Ca 2ϩ binds to the C-terminal domain. Alternatively, the conformational change in all three calmodulin upon Ca 2ϩ binding to the C-terminal domain results in the inhibition of motility and the additional conformational change upon the Ca 2ϩ binding at the N-terminal low affinity sites destabilizes the association of one of the bound calmodulin to the heavy chain presumably due to steric hindrance. Further studies are needed to clarify the reason why only one molecule of calmodulin is dissociated from myosin I␤.
Studies reported here with mutant calmodulin show that motor function of myosin I␤ is regulated by Ca 2ϩ binding to the high affinity sites of calmodulin light chains. This regulatory mechanism may also apply to those of other unconventional myosins which contain calmodulin as their light chains.