Ca2+-dependent Regulation of the Motor Activity of Myosin V*

Mouse myosin V constructs were produced that consisted of the myosin motor domain plus either one IQ motif (M5IQ1), two IQ motifs (M5IQ2), a complete set of six IQ motifs (SHM5), or the complete IQ motifs plus the coiled-coil domain (thus permitting formation of a double-headed structure, DHM5) and expressed in Sf9 cells. The actin-activated ATPase activity of all constructs except M5IQ1 was inhibited above pCa 5, but this inhibition was completely reversed by addition of exogenous calmodulin. At the same Ca2+ concentration, 2 mol of calmodulin from SHM5 and DHM5 or 1 mol of calmodulin from M5IQ2 were dissociated, suggesting that the inhibition of the ATPase activity is due to dissociation of calmodulin from the heavy chain. However, the motility activity of DHM5 and M5IQ2 was completely inhibited at pCa 6, where no dissociation of calmodulin was detected. Inhibition of the motility activity was not reversed by the addition of exogenous calmodulin. These results indicate that inhibition of the motility is due to conformational changes of calmodulin upon the Ca2+ binding to the high affinity site but is not due to dissociation of calmodulin from the heavy chain.

Myosins are motor proteins that translocate actin filaments upon hydrolysis of ATP, and thus they play a critical role in diverse forms of cell contractility and motility. During the last decade a number of myosin-like proteins have been found, and the myosins are currently organized into 15 classes based upon phylogenetic sequence comparisons of the motor domains (1)(2)(3)(4)(5). Class V myosin was originally identified in brain as a calmodulin-binding protein that had actin-dependent ATPase activity (6). Myosin V is a member of the myosin superfamily that is expressed in variety of cell types and is involved in a variety of membrane trafficking and organelle transport functions (1)(2)(3)(4)(5). Myosin V has two heads that are connected with a long coiledcoil domain; however, in contrast to conventional myosin, it contains a globular C-terminal domain and does not form thick filaments (7). The head domain is composed of a globular motor domain and an elongated neck domain that is associated with a number of light chains. The sequence at the neck region contains six IQ motifs that have been implicated as calmodulin or myosin light chain binding consensus motifs as found in a variety of calmodulin-binding proteins and myosins (7). Since light chains play a critical role in the regulation of various conventional myosins, it has been proposed that the IQ domain serves as a regulatory component of myosin V. The role of the IQ motif and bound calmodulin serving as a regulatory component of unconventional myosins is best studied for mammalian myosin Is. For both brush border myosin I (8 -10) and myosin I␤ (11)(12)(13), high Ca 2ϩ inhibits motor activity due to Ca 2ϩ binding to the calmodulin light chain. Since 1 mol of bound calmodulin dissociates from myosin I at high Ca 2ϩ , it was originally thought that this dissociation of calmodulin was responsible for the inhibition of myosin I motor activity. However, since virtually no calmodulin dissociation is observed at pCa 6 where the motility activity is completely abolished, this view has been questioned (13). For naturally isolated myosin V, motility activity is inhibited at high Ca 2ϩ , whereas actin-activated ATPase activity markedly increases in the presence of Ca 2ϩ (7). Recently it was shown (14) that a truncated recombinant monomeric myosin V with two IQ motifs had motility activity that was inhibited at high Ca 2ϩ , but only in the absence of exogenous calmodulin, suggesting that the inhibition is via the physical dissociation of calmodulin. Interestingly, the truncated myosin V showed inhibition of actin-activated ATPase activity by Ca 2ϩ rather than activation, as is found for naturally isolated myosin V (14). This apparent discrepancy is not understood, but there are several possible explanations. Since the two-headed structure is critical for the regulation of both conventional smooth muscle and non-muscle myosin motor function, it is plausible that the two-headed structure may play some role in the regulatory mechanisms of myosin V. Alternatively, a complete neck domain may be required for the proper regulation of myosin V.
The aim of the present study is to clarify the regulation of myosin V motor function by Ca 2ϩ . To address these questions, we have produced a two-headed myosin V construct, a singleheaded construct having the entire neck domain (six IQ motifs), and a truncated single-headed construct containing only two IQ motifs. These myosin V constructs were expressed, purified, and examined for motor function. It was found that although calmodulin dissociation is responsible for inhibition of actinactivated ATPase activity at high Ca 2ϩ , inhibition of motility occurs at lower Ca 2ϩ where calmodulin is not dissociated from myosin V heavy chain.

EXPERIMENTAL PROCEDURES
Materials-Restriction enzymes and modifying enzymes were purchased from New England Biolabs (Beverly, MA). Actin was prepared from rabbit skeletal muscle acetone powder according to Spudich and Watt (15). Recombinant calmodulin from Xenopus oocyte (16) was expressed in Escherichia coli as described (17).
Generation of the Expression Vectors for Myosin V Mutants-A baculovirus transfer vector for mouse myosin V variants in pBluebac4 (Invitrogen, CA) was produced as follows. Mouse myosin V cDNA clones containing Ϫ35-1549 and 1549 -3928 in pBluescript were kindly supplied by Dr. N. Jenkins (NCI, National Institutes of Health). A cDNA fragment-(1549 -3928) flanked with EcoRI sites was ligated to the vector containing a cDNA fragment-(Ϫ35-1549) at the unique EcoRI site at nucleotide 1549. A unique NheI site was created at the 5Ј side of the initiation codon; a unique KpnI site at nucleotide 2125 was deleted without changing the resulting amino acid residue, and a new KpnI site was created at nucleotide 3578. A cDNA fragment was excised with NheI/KpnI digestion and in-frame ligated to a PBluebac4His baculovirus transfer vector containing a hexahistidine tag sequence with a stop codon at the 3Ј side of the KpnI site. This construct (DHM5), containing the entire coiled-coil domain, was used to express double-headed myosin V. To obtain single-headed myosin V with complete IQ motifs, a KpnI site was created at nucleotide 2744 of DHM5. The vector was digested with KpnI and the nucleotide fragment-(2744 -3578) was removed. The vector was self-ligated and used as a construct expressing a single-headed myosin V with complete IQ motifs (SHM5). A KpnI site was also created at nucleotide 2468 or 2393 of DHM5. Vectors containing nucleotides 1-2468 and 1-2393 of myosin V were used to express single-headed myosin V that contained two IQ motifs (M5IQ2) and one IQ motif (M5IQ1), respectively. For M5IQ2, nucleotides encoding a Myc tag sequence (EQKLISEEDL) were inserted at 5Ј side of the hexahistidine tag.
Preparation of Recombinant Myosin V-To express recombinant myosin V, 200 ml of Sf9 cells (about 1 ϫ 10 9 ) were coinfected with two separate viruses expressing the myosin V heavy chain and calmodulin, respectively. The cells were cultured at 28°C in 175-cm 2 flasks and harvested after 3 days. Cells were lysed with sonication in 30 ml of Lysis Buffer (0.15 M NaCl, 30 mM KP i , pH 8.0, 5 mM MgCl 2 , 2 mM ␤-mercaptoethanol, 2 mM phenylmethanesulfonyl fluoride, 0.2 mM N-ptosyl-L-phenylalanine chloromethyl ketone, 0.2 mM N-p-tosyl-L-lysine chloromethyl ketone, 0.01 mg/ml leupeptin, 0.1 mg/ml trypsin inhibitor, 10% glycerol, and 2 mM ATP). After centrifugation at 140,000 ϫ g for 15 min, the supernatant was incubated with 50 mM glucose and 20 units/ml hexokinase at 4°C for 30 min to hydrolyze completely residual ATP. F-actin (0.2 mg/ml) was added to the sample and centrifuged (140,000 ϫ g for 20 min) to coprecipitate the expressed myosin V. The pellets were washed once with buffer A (0.3 M NaCl, 30 mM KP i , pH 8.0, 0.1 mM EGTA, 10 mM ␤-mercaptoethanol) and then resuspended with buffer A with 5 mM MgCl 2 and 5 mM ATP to release myosin V from F-actin. The supernatant was mixed with 0.2 ml of nickel-nitrilotriacetic acid-agarose (Qiagen, Hilden, Germany) 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 10-fold volume of buffer B (0.3 M KCl, 20 mM imidazole, pH 7.5, 0.1 mM EGTA, and 10 mM ␤-mercaptoethanol). Myosin V was eluted with buffer C (0.1 M KCl, 0.1 mM EGTA, 10 mM ␤-mercaptoethanol, and 0.2 M imidazole, pH 7.5).
After SDS-polyacrylamide gel electrophoresis analysis, fractions containing myosin V were pooled and dialyzed against 150 mM KCl, 20 mM MOPS, 1 pH 7.0, and 1 mM dithiothreitol. The purified myosin V was stored on ice and used within 2 days. Typically, 1 mg of isolated myosin V was obtained.
In Vitro Motility Assay-The in vitro motility assay was performed as described previously (20). Myosin V was attached to the coverslip. For M5IQ2, a coverslip was first coated with anti-Myc antibodies (Santa Cruz Biotechnology, Santa Cruz, CA), and then the surface of the coverslip was blocked with bovine serum albumin. Myosin V was then applied to the coverslip. 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.

Expression of Myosin V Constructs-
In order to study the regulatory mechanism of myosin V motor function, various myosin V constructs were produced and expressed in Sf9 insect cells. The DHM5 construct contains the entire coiled-coil domain in addition to the complete head domain, so it is anticipated that it has a double-headed structure (Fig. 1). SHM5 has an entire head domain with complete IQ motifs but no coiledcoil domain, thus it is expected to be a single-headed molecule ( Fig. 1). On the other hand, M5IQ2 contains a motor domain but only 2 N-terminal IQ motifs out of the six IQ motifs typically present. All constructs contain hexahistidine tags to aid in purification. The histidine tags were introduced at the C-terminal end of the molecule rather than its N-terminal end to avoid possible misfolding of the protein and any effects on motor function. Addition of a hexahistidine tag at the C-terminal end of myosins has been performed with conventional (21) as well as unconventional myosin, 2 and no influence on motor function was observed. The cells were coinfected with myosin V-expressing virus and calmodulin-expressing virus. The ratio of the two viruses to achieve the best myosin V expression was empirically determined. It should be noted that myosin V with bound calmodulin can be obtained without the coinfection of calmodulin-expressing virus but that the fraction of soluble myosin V increases significantly with calmodulin coexpression. This tendency was more prominent for constructs having an entire IQ domain. Fig. 2 shows SDS-polyacrylamide gel electrophoresis of the purified constructs. Each myosin V construct was composed of a high molecular mass band and a low molecular mass band. The molecular mass of each slow mobility band was consistent with the calculated molecular mass of each myosin construct, i.e. 130, 100, and 90 kDa for DHM5, SHM5, and M5IQ2, respectively. These bands were recognized by anti-His antibodies (Santa Cruz Biotechnology, Santa Cruz, CA) indicating that these high molecular mass bands are the expressed myosin V heavy chains (not shown). The small sub- 1 The abbreviation used is: MOPS, 4-morpholinepropanesulfonic acid. 2  units showed a mobility shift with a change in [Ca 2ϩ ] that is characteristic of calmodulin, suggesting that the small subunits are indeed calmodulin. The identification of the small subunit was also confirmed using anti-calmodulin antibodies (not shown). The stoichiometries of calmodulin versus myosin V heavy chain for DHM5, SHM5, and M5IQ2 were 5.7 Ϯ 0.7, 5.3 Ϯ 0.7, and 2.1 Ϯ 0.5, respectively. This is consistent with the number of IQ motifs in each construct.
Actin-activated ATPase Activity-The actin-activated ATPase activity was measured as a function of free Ca 2ϩ (Fig.  3). For most constructs, i.e. DHM5, SHM5, and M5IQ2, the ATPase activity decreased markedly above pCa 6. In contrast, M5IQ1 showed no pCa dependence (not shown). Interestingly, inhibition of ATPase activity was reversed by addition of 12 M calmodulin in assay solution (Fig. 3). The effect of exogenous calmodulin on the ATPase activity was common for all three constructs. Fig. 4 shows the calmodulin concentration dependence of the actin-activated ATPase activity of the myosin V constructs. The calmodulin dependence on the activation of the ATPase activity for the three constructs was similar, and a half-maximal activation was observed at 10 Ϫ6 M calmodulin concentration. There was no difference in calmodulin dependence of ATPase activity between pCa 5 and 4, suggesting that the Ca 2ϩ effect is saturated at pCa 5. Fig. 5 shows the F-actin dependence of the actin-activated ATPase activities of the three constructs. The actin concentration dependence was sim-ilar to each other for the three constructs. K actin estimated from the actin dependence of the ATPase activity was 0.95, 0.80, and 0.60 M for DHM5, SHM5 and M5IQ2, respectively, in the presence of EGTA. On the other hand, the actin concentration dependence at high [Ca 2ϩ ] showed two phases. At low F-actin concentration, activity increased with actin concentration, and it was not dramatically different from that in the presence of EGTA. However, the activity decreased at higher actin concentrations. This result suggests that the Ca 2ϩ -dependent inhibition of ATPase activity is not due to a change in the affinity for F-actin. The binding of myosin V constructs to F-actin was examined by cosedimentation analysis in the presence of Mg 2ϩ -ATP to test further the effect of Ca 2ϩ on the affinity between myosin V and F-actin. For all three constructs a significant amount of myosin V cosedimented with F-actin at low ionic strength (more than 50%), but no significant effect of Ca 2ϩ on the binding was observed (not shown). The difference in F-actin dependence of the ATPase activities of myosin V at high and low Ca 2ϩ conditions is discussed under "Discussion." Dissociation of Calmodulin from Myosin V-The reversal of the ATPase activity inhibition by high concentrations of exogenous calmodulin suggests that it is due to a decrease in the affinity of calmodulin for myosin V when calmodulin binds Ca 2ϩ . To address this notion, the dissociation of calmodulin from myosin V was examined. Myosin V constructs were mixed with F-actin in various free Ca 2ϩ concentrations and then ultracentrifuged to determine bound calmodulin. F-actin coprecipitated myosin V with bound calmodulin was analyzed by SDS-polyacrylamide gel electrophoresis, and the calmodulin band was quantitated by densitometry normalized with myosin V heavy chain. As a control free calmodulin was centrifuged with F-actin, but no calmodulin coprecipitation was detected.
Myosin V constructs were also tested for precipitation, but no myosin V was precipitated in the absence of F-actin. As shown in Fig. 6, the calmodulin bound to myosin V decreased significantly between pCa 6 and 5 for all three constructs. Approximately 55% of bound calmodulin was dissociated from M5IQ2 heavy chain, indicating that 1 mol of calmodulin is dissociated from M5IQ2. On the other hand, approximately 35% of the bound calmodulin was dissociated from the DHM5 and SHM5 constructs, indicating that 2 mol of bound calmodulin out of 6 total are dissociated.
Effect of Ca 2ϩ on the Motility Activity of Myosin V- Fig. 7 shows the motility activity of DHM5 and M5IQ2 at various calcium ion concentrations. Consistent with the actin-activated ATPase activity, the motility activity was completely inhibited at high Ca 2ϩ concentrations. An important finding is that the inhibition of motility was achieved at much lower Ca 2ϩ concentrations than the inhibition of the ATPase activity. Furthermore, in contrast to its effects on the inhibition of ATPase activity, addition of exogenous calmodulin did not reverse the motility inhibition (Fig. 7, C and D). We added exogenous calmodulin up to 59 M at pCa 4, but the motility activity was not recovered in contrast to the ATPase activity. A similar result was also obtained at pCa 6. The recovery of motility inhibited at high Ca 2ϩ was only achieved by reducing Ca 2ϩ concentration (Fig. 7, A and B).
Inhibition of the Actin-activated ATPase Activity by ADP-It was quite recently shown that the ADP release step is the rate-limiting step for the actin-activated ATPase cycle of truncated myosin V (22). Therefore, ADP inhibition of the ATPase activity of myosin V and the effect of Ca 2ϩ on this inhibition were studied. As shown in Fig. 8, ADP strongly inhibited the actin-activated ATPase activity both in the presence and absence of Ca 2ϩ . By using a K ATP of 1.4 M (22), a K i of 1.47 M was obtained from the ADP dependence of the ATPase activity under EGTA conditions, whereas a K i of 1.80 M was found under pCa 4 conditions. This value was significantly lower than for conventional myosins (23)(24). Interestingly, Ca 2ϩ did not affect the ADP inhibition indicating that the affinity of myosin V for ADP is not influenced by Ca 2ϩ . DISCUSSION Except for striated muscle myosin II, the biochemically well characterized myosins have self-regulated motor function or contain regulatory components within the myosin molecule itself. The motor function of conventional myosin from vertebrate smooth muscle and non-muscle tissues is regulated by the phosphorylation of its regulatory light chain (25)(26)(27). Molluscan conventional myosin is regulated by direct Ca 2ϩ binding to the essential light chain as supported by the regulatory light chain (28). Interestingly, the regulation requires a two-headed structure and the involvement of inter-head interaction in the regulatory mechanism (20, 29 -30). Myosin I regulation is the most well studied of the unconventional myosins from vertebrates. These vertebrate myosin Is contain calmodulin as their light chains, and Ca 2ϩ binding to this calmodulin light chain regulates their motor activity (8 -13). These results raise several questions concerning the regulation of myosin V. Does Ca 2ϩ regulate myosin V motor function? If so, does calmodulin play a critical role in the regulation? Is the number of calmodulins in the myosin V molecule important for the regulation? Finally, is the two-headed structure critical for regulation? In this study, we produced three different constructs of myosin V to address these questions concerning regulation, i.e. doubleheaded myosin V, single-headed myosin V with complete IQ motifs, and single-head myosin V with two IQ motifs.
It was reported that the essential light chain of conventional myosin copurified with myosin V from chicken tissue (31) suggesting that it participates as a subunit of myosin V. However, the essential light chain was not found in a myosin V preparation from mouse. Moreover, it was quite recently reported that myosin V expressed in Sf9 cells that were coinfected with calmodulin and essential light chain contains calmodulin exclusively and not the essential light chain (32). Therefore, the myosin V variants produced in the present study that have only calmodulin as light chains represent the likely subunit composition of physiologically relevant myosin V.
Quite recently, double-headed and single-headed myosin V constructs similar to those produced in the present study were reported (32). It was reported that the double-headed myosin V was quite unstable and degraded during cell culture by proteolysis at the IQ domain (32). In contrast, we have not seen any significant degradation of our double-headed myosin V sample. Although the difference is unclear, since the degradation was found at the IQ domain it is likely that the degradation occurred due to a lack of calmodulin binding there. Consistent with this notion, the SHM5 isolated by Wang et al. (31) increased in ATPase activity with exogenous calmodulin even while in EGTA, suggesting that their SHM5 was not saturated with calmodulin. The myosin V constructs in this study were not further activated by exogenous calmodulin under EGTA conditions, suggesting that the sites are saturated with bound calmodulin consistent with the determined stoichiometry of the bound calmodulin.
All of these myosin V constructs, i.e. DHM5, SHM5, and M5IQ2, showed the same pCa dependence in their actin-activated ATPase activity. The activity was significantly reduced above pCa 6. It can be concluded that this decrease in the ATPase activity is due to the dissociation of calmodulin molecule from the heavy chain because: 1) myosin V-bound calmodulin decreased at pCa 5 and higher, and approximately 1 mol of calmodulin from M5IQ2 or 1-2 mol of calmodulin from DHM5 and SHM5 were dissociated from the heavy chain; and 2) the decrease in ATPase activity observed at pCa 5 and higher was completely reversed by addition of exogenous calmodulin. Since Ca 2ϩ binds to calmodulin within this range of free Ca 2ϩ , the scenario would be that Ca 2ϩ binding to the calmodulin bound to myosin V decreases the affinity of calmodulin for myosin V. However, since only 1 or 2 mol of the bound calmodulin disso-ciates from myosin V, the apparent decrease in affinity should be the site-specific. The nature of the binding of the IQ peptide with calmodulin has been modeled (33). In the absence of Ca 2ϩ , the N-terminal lobe of calmodulin adopts a closed conformation, whereas the C-terminal lobe adopts a semi-open conformation that interacts with the N-terminal side of the IQ peptide via a number of H bonds. Upon Ca 2ϩ binding, calmodulin adopts the open conformation, thus the C-terminal lobe cannot bind to the IQ peptide. While the interaction between apocalmodulin and the IQ peptide may diminish, Ca 2ϩ /calmodulin may now bind in a distinct manner to the IQ peptide since in general it retains the amphipathic character also found in various calmodulin target peptides (34). It is plausible, therefore, that the dissociation of calmodulin at high Ca 2ϩ occurs at specific IQ sites that do not allow calmodulin to rebind at high Ca 2ϩ presumably due to a lack of amphipathic character. The present result suggests that 1-2 mol of calmodulin are dissociated from myosin V at high Ca 2ϩ . One such Ca 2ϩ -dependent site would be the second IQ region, since M5IQ2 loses 1 mol of calmodulin at high Ca 2ϩ but M5IQ1 does not. This result is consistent with the recent report (14). Another possible site is the last IQ region, which is least conserved and where the second consensus Arg that interacts via three H bonds with the FIG. 7. Ca 2؉ and calmodulin dependence of actin filament sliding velocity of myosin V constructs. A, Ca 2ϩ dependence of motility activity of DHM5; B, Ca 2ϩ dependence notility activity of M5IQ2. The flow cells with myosin V constructs filled with pCa 6 buffer was refilled with pCa 7 buffer for restoration of the motility. C and D show the effect of exogenous calmodulin on the motility activity of DHM5 and M5IQ2, respectively. Exogenous calmodulin failed to rescue the motility activity of myosin V at pCa 4. The same result was also obtained at pCa 6 condition. Actin filament motility was observed in 30 mM KCl, 1 mM MgCl 2 , 12 M calmodulin, 20 mM MOPS, pH 7.0, 4.5 g/ml glucose, 216 g/ml glucose oxidase, 36 g/ml catalase, and 2 mM ATP at 25°C. Measurements were made with three independent preparations, and 11-20 actin filaments were measured to obtain an average velocity for preparation. All values are mean velocity Ϯ S.D. In the conditions where actin filaments were moved, nearly all actin filaments observed moved continuously. N-terminal lobe of calmodulin (according to the model (33)) is replaced by Lys.
It is of interest to identify how Ca 2ϩ decreases the actinactivated ATPase activity. Actin binding of myosin V in the presence of ATP was unchanged with Ca 2ϩ ; therefore, Ca 2ϩ does not alter its affinity for F-actin. It was reported quite recently (21) that the rate-limiting step of the unregulated truncated myosin V construct, i.e. (M5IQ1), is the ADP off step. We examined the effect of ADP on the inhibition of the ATPase activity, but the apparent K i value was not influenced by Ca 2ϩ . Interestingly, an inhibition of ATPase activity was observed at high, but not low, F-actin concentration in the presence of Ca 2ϩ . For conventional myosin it is known that the inhibition of ATPase activity by high F-actin concentrations is due to the inhibition of the ATP hydrolysis step by F-actin (35). Therefore, it is plausible that the inhibition of the ATPase activity by Ca 2ϩ for myosin V is due to an inhibition of ATP hydrolysis step.
The inhibition of the actin-activated ATPase activity was completely reversed by the presence of micromolar levels of calmodulin. Since a concentration of free calmodulin of several micromolar is present in cells, the inhibition of ATPase activity may not be operating in vivo. An important finding is that the motility of myosin V (DHM5) is completely abolished below pCa 6 where no apparent dissociation of calmodulin takes place. Furthermore, addition of a high concentration of exogenous calmodulin failed to rescue this inhibition of motility. These results indicate that the inhibition of the motility of myosin V is not due to dissociation of calmodulin from the heavy chain. The inhibition occurs between pCa 7 and 6 where cytoplasmic Ca 2ϩ concentration is regulated in most cell types; therefore, the observed inhibition is physiologically relevant. Similar results were reported for mammalian myosin I␤ (12,13). Since this range of Ca 2ϩ concentrations corresponds to the high affinity C-terminal sites of calmodulin, it is likely that Ca 2ϩ binding to the C-terminal lobe of calmodulin and consequent conformational changes are responsible for the inhibition of motility. Supporting this notion, it was shown that deletion of the C-terminal Ca 2ϩ -binding sites abolishes inhibition of the motility of myosin I␤ (13).
Quite recently, it was reported that truncated myosin V containing 2 IQ motifs showed motility activity even at high Ca 2ϩ when high exogenous calmodulin was present although the velocity of the motility is reduced (14). At present the reason of this apparent discrepancy between the present result and that by Trybus et al. (14) is obscure. The loss of the motility at high Ca 2ϩ was not due to the denaturation of myosin V since the perfusion of the flow cell with low Ca 2ϩ buffer (pCa 7 or less) completely restored the motility (Fig. 7). We examined various exogenous calmodulin concentrations up to 59 M that is 5 times higher than that used by Trybus et al. (14); therefore, it is unlikely that the loss of the motility is due to the unsaturation of the bound calmodulin. Actually, the inhibition of the motility is observed at pCa 6, where virtually no calmodulin is dissociated from myosin V heavy chain. The present results suggest that true regulation of motility could require two Nterminal IQ domains, in which the conformational change of calmodulin, but not the physical dissociation of calmodulin, would trigger the inhibition of motility. There is an apparent de-coupling between the ATP hydrolysis cycle and mechanical events at higher Ca 2ϩ . This is different from the regulation of conventional myosins in which the regulatory domain regulates both ATPase and mechanical activities (25)(26)(27)(28). It is plausible that the change in calmodulin conformation alters the rigidity of the "lever arm" and thus decouples the chemical and mechanical events. Alternatively, the conformational change of calmodulin could alter the interaction between calmodulin and the "converter" domain of myosin V, thus inhibiting motility.
The present results clearly indicate that the two-headed structure is not critical for regulation of mechanoenzymatic activity of myosin V. This is distinct from the regulation of conventional myosin, in which an interaction between the two heads is involved in the regulation (20 -21, 28 -30). Mammalian myosin I, a single-headed unconventional myosin having calmodulin as light chain subunits, has been shown to have Ca 2ϩ dependence similar to that shown for myosin V in the present study (8 -13). Therefore, it is plausible that there is a common motor activity regulatory mechanism in unconventional myosins carrying calmodulin light chains. However, more detailed biophysical and biochemical information is required for further understanding of the molecular mechanism.