Dual regulation of mammalian myosin VI motor function.

Myosin VI is expressed in a variety of cell types and is thought to play a role in membrane trafficking and endocytosis, yet its motor function and regulation are not understood. The present study clarified mammalian myosin VI motor function and regulation at a molecular level. Myosin VI ATPase activity was highly activated by actin with K(actin) of 9 microm. A predominant amount of myosin VI bound to actin in the presence of ATP unlike conventional myosins. K(ATP) was much higher than those of other known myosins, suggesting that myosin VI has a weak affinity or slow binding for ATP. On the other hand, ADP markedly inhibited the actin-activated ATPase activity, suggesting a high affinity for ADP. These results suggested that myosin VI is predominantly in a strong actin binding state during the ATPase cycle. p21-activated kinase 3 phosphorylated myosin VI, and the site was identified as Thr(406). The phosphorylation of myosin VI significantly facilitated the actin-translocating activity of myosin VI. On the other hand, Ca(2+) diminished the actin-translocating activity of myosin VI although the actin-activated ATPase activity was not affected by Ca(2+). Calmodulin was not dissociated from the heavy chain at high Ca(2+), suggesting that a conformational change of calmodulin upon Ca(2+) binding, but not its physical dissociation, determines the inhibition of the motility activity. The present results revealed the dual regulation of myosin VI by phosphorylation and Ca(2+) binding to calmodulin light chain.

Myosin, a motor protein that translocates actin filaments upon hydrolysis of ATP, constitutes a superfamily with 18 classes based upon phylogenetic sequence comparisons of the motor domains (1)(2)(3)(4). Class VI myosins were first identified in Drosophila melanogaster (5), and subsequently found in mammals (6 -8). While myosin VI is found in various tissues, its physiological significance may be felt most prominently in auditory function, where it is found that the mutation of the myosin VI gene results in auditory malfunction (7,8). Actually, myosin VI is found in the neuroepithelium of the cochlea of the inner ear in both the inner and outer hair cells. In these cells, myosin VI is concentrated at actin-rich stereocilia and may play a role in the rigidity of anchoring of the stereocilia (7,8); therefore, it has been suggested that myosin VI plays a role in the mechanical function of stereocilia (7,8). In Drosophila, myosin VI is associated with particles that move in a cell cycle-dependent manner (9,10). It has also been suggested that myosin VI plays a role in membrane trafficking and endocytosis (11).
The structure of myosin VI has been predicted based upon its amino acid sequence to be composed of a head domain, a coiledcoil domain, and a globular tail domain. The head domain is itself divided into a globular motor domain and a neck domain containing a light chain binding region. The sequence at the neck region contains a single IQ motif that is implicated as a calmodulin or myosin light chain binding consensus motif as found in a variety of calmodulin-binding proteins and myosins (1)(2)(3)(4). The coiled-coil domain is present at the C-terminal side of the neck region, so it is predicted that myosin VI is a twoheaded myosin. Finally, the globular tail domain is hypothesized to be a targeting domain that determines the cellular binding counterpart.
There are two unique inserts in the head domain, one at the surface in the upper 50-kDa domain and the other at the junction between the converter domain and the IQ domain. Until quite recently, all myosin motors were characterized as moving toward the barbed end of actin filaments. Since unconventional myosins play a role in translocating cellular organelles along actin cables, a myosin having opposite moving directionality would be expected to have unique cellular functions. Of the 18 different classes of myosins, it was found that myosin VI moves toward the minus end of F-actin filaments (12). While the mechanism underlying the reverse movement of myosin VI is unclear, the unique large insertion in the myosin VI head domain between the motor domain and the light chain binding domain (lever arm) has led to the postulation (12) that this alters the angle of the lever arm switch movement, thus changing the direction of motility.
The importance of light chains in the regulation of myosin motor function is seen in both vertebrate smooth muscle/nonmuscle myosin and invertebrate myosin, in which the phosphorylation of the regulatory light chain (13)(14)(15) and Ca 2ϩ binding to the essential light chain (16), respectively, trigger the activation of motor activity.
The role of the IQ motif and bound calmodulin serving as a regulatory component of unconventional myosins was first studied for mammalian myosins I. For both brush border myosin I (17,18) and myosin I␤ (19 -21), high Ca 2ϩ inhibits motor activity due to its 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 (17,18). However, since virtually no calmodulin dissociation is observed at pCa 6, where the motility activity is completely abolished, this view has been questioned (20,21). Quite recently, a similar finding was made for myosin V (22). While calmodulin can be dissociated from its heavy chain at high Ca 2ϩ , the motility activity of myosin V is abolished at pCa 6, where no calmodulin dissociation takes place, suggesting that a Ca 2ϩ binding-induced conformational change of the bound cal-modulin, but not a physical dissociation, is critical for regulation. However, most calmodulin-binding myosins contain multiple calmodulin molecules at their neck domains, and it has been shown that neck-deleted myosin V containing a single bound calmodulin is unregulated by Ca 2ϩ (22,23). Therefore, a remaining question is whether or not the motor activity of myosin VI, which contains a single calmodulin at its neck domain, is regulated by Ca 2ϩ .
Myosin VI contains a potential phosphorylation site in a loop near the tip of the head that is in a homologous position to the amoeba myosin I phosphorylation site (24). Since the phosphorylation of the heavy chain of amoeba myosin I at this site is necessary for the activation of both actin-activated ATPase activity and motility activity (25), a hypothesis is raised that myosin VI motor function might be regulated by the phosphorylation at this position by protein kinases, although no supporting data are available to date.
The aim of the present study is to clarify the motor function and regulation of myosin VI. To achieve the goal, we expressed myosin VIHMM in Sf9 cells, purified and examined for motor function. The obtained results reveal a dual regulation of myosin VI motor function by Ca 2ϩ binding and p21-activated kinase 3-catalyzed phosphorylation.

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 (26). Recombinant calmodulin from Xenopus oocyte (27) was expressed in Escherichia coli as described (28).
Generation of the Expression Vectors for Myosin VI Constructs-Mouse cDNA clones containing Ϫ150 to 2565 (B10) and 1460 -3708 (B2) in pBluescript were kindly provided from Dr. K. Avraham (Tel Aviv University). B10 clone was digested by KpnI/PflM1, and the cDNA fragment of myosin VI (residues 1679 -3309) obtained from the B2 clone by KpnI/PflM1 digestion was inserted into the B10 clone (B2/B10). A unique NheI site was created at the 5Ј side of the initiation codon, and then the myosin VI cDNA cut by NheI/KpnI digestion was inserted into pBluebac 4 (Invitrogen, Carlsbad, CA) baculovirus transfer vector at the polylinker region. A hexahistidine tag sequence with a stop codon was introduced at the 3Ј side of the KpnI site. This construct (M6HMM), containing the entire coiled-coil domain, was used to produce recombinant baculovirus expressing myosin VI as described previously (29). Thr 406 of M6HMM was mutated to Ala by site-directed mutagenesis (30), and the mutation was confirmed by direct sequence analysis.
Generation of PAK3 1 Construct-Rat PAK3 cDNA in a pGEX4T-2 vector was kindly provided by Dr. G. Cote (Queen's University, Canada). The clone was digested with BamHI/XhoI, and the obtained fragment containing the entire coding region was in-frame ligated into pFastBacHTb (Life Technologies, Inc.) baculovirus transfer vector. To produce a constitutively active form, a BamHI site was created at codon 210, and the nucleotides encoding the N-terminal 210 amino acid residues were excised by BamHI digestion. The produced construct was used to express PAK3 kinase domain (PAK3KD). The plasmids were transformed into DH10BAC E. coli. Recombinant bacmids (recombinant virus DNA) were isolated and used to transfect Sf9 insect cells.
After SDS-PAGE analysis, fractions containing myosin VI were pooled and dialyzed against 30 mM KCl, 20 mM MOPS, pH 7.5, 1 mM EGTA, 1 mM MgCl 2 , and 1 mM dithiothreitol. The purified myosin VI was stored on ice and used within 2 days.
Production of Antibody-A peptide (12-mer) containing phospho-Thr 406 with N-terminal Cys was chemically synthesized as antigen and bound to the carrier protein, keyhole limpet hemocyanin at the Nterminal cysteine residue. Antibodies were prepared by injecting two rabbits with keyhole limpet hemocyanin-coupled peptide. Antibodies against phospho-Thr 406 were affinity-purified from the obtained serum by affinity chromatography with the peptide-conjugated resin. The purified antibodies were subjected to absorption on unphosphorylated peptide-coupled resin.
Gel Electrophoresis and ATPase Assay-SDS-polyacrylamide gel electrophoresis was carried out on a 7.5-20% polyacrylamide gradient slab gel using the discontinuous buffer system of Laemmli (31). Molecular mass markers used were smooth muscle myosin heavy chain (204 kDa, ␤-galactosidase (116 kDa), phosphorylase b (97.4 kDa), bovine serum albumin (66 kDa), ovalbumin (45 kDa), carbonic anhydrase (29 kDa), myosin regulatory light chain (20 kDa), and ␣-lactalbumin (14.2 kDa). The amount of the myosin VI heavy chain and calmodulin was determined by densitometry as described previously (20). The steadystate ATPase activity was determined by measuring liberated P i at 25°C as described previously. (32). The ATPase activity was also measured in the presence of 20 units/ml pyruvate kinase, 3 mM phosphoenolpyruvate as a ATP regeneration system. The liberated pyruvate was determined as described (33).
In Vitro Motility Assay-The in vitro motility assay was performed as described previously (34). Myosin VI was attached to a 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.
Phosphorylation of Myosin VI by PAK3-Prior to the phosphorylation reactions, PAK3KD was incubated in 75 mM NaCl, 30 mM imidazole-HCl (pH 7.5), 4 mM MgCl 2 , 1 mM ATP, 2 mM dithiothreitol, and 0.2 M microcystin for 30 min at 25°C (35). The phosphorylation reaction was then carried out in a buffer containing 50 mM NaCl, 20 mM imidazole-HCl (pH 7.5), 2 mM MgCl 2 , 0.1 M microcystin, 1 mM dithiothreitol, 0.5 mM [␥-32 P]ATP (300 Ci/mol) by the addition of 4 g of myosin VI. The reaction mixtures were incubated at 25°C for the indicated time periods. The reaction was stopped by adding 5% trichloroacetic acid and analyzed by SDS-PAGE and autoradiography. To quantify phosphate incorporation, the myosin VI heavy chain bands of the SDS-PAGE gels were excised, and the radioactivity was quantitated by a scintillation counter. Phosphorylation of myosin VI at Thr 406 was determined by Western blotting using anti-phospho-Thr 406 -specific antibodies. Western blotting was carried out as described previously (36).
Actin Co-sedimentation Assay-The binding of calmodulin to M6HMM heavy chain was determined by actin co-sedimentation assay. M6HMM was incubated in buffer containing 25 mM imidazole-HCl, pH 7.5, 25 mM KCl, 5 mM MgCl 2 , 1 mM EGTA, 0.025 mg/ml F-actin, and various concentrations of CaCl 2 at 25°C for 15 min. The sample was ultracentrifuged at 100,000 ϫ g for 30 min, and the pellets were analyzed by SDS-polyacrylamide gel electrophoresis. The amounts of the co-sedimented M6HMM heavy chain and calmodulin were determined by densitometry as described previously (20).
To determine the binding of myosin VI with actin in the presence of ATP, M6HMM (0.1 mg/ml) was incubated in buffer containing 20 mM imidazole, pH 7.5, 50 mM KCl, 2 mM MgCl 2 , 2 mM ATP, and various concentrations of actin in the presence of ATP regeneration system (20 units/ml pyruvate kinase and 3 mM phosphoenolpyruvate) at 25°C for 30 min. The samples were ultracentrifuged at 100,000 ϫ g for 30 min, and the pellets were analyzed by SDS-polyacrylamide gel electrophoresis. The amount of the co-sedimented heavy chain was determined by densitometry as described previously (20).

Expression and Purification of Mammalian Myosin VI-
Mouse myosin VI construct was produced and expressed in Sf9 insect cells. The construct (M6HMM) contains the entire coiledcoil domain and the complete head domain along with a Cterminal hexahistidine tag to aid in purification (Fig. 1). Histidine tagging at the C-terminal end of the molecule has been performed on conventional (37) as well as unconventional myosins, 2 and no influence on motor function has been observed. The cells were co-infected with an appropriate ratio of myosin VI-expressing virus and calmodulin-expressing virus. It should be noted that functional myosin VI was only obtained with co-infection of calmodulin virus, in contrast to myosin V in which functional protein can be obtained without calmodulin co-infection (23). The purification process was basically two steps (i.e. F-actin co-precipitation followed by ATP-induced dissociation from F-actin and Ni 2ϩ -agarose affinity chromatography using the hexahistidine tag (see "Experimental Procedures"). The former step selects the functionally active molecules, and the second step eliminates the endogenous Sf9 cell myosin and F-actin. Fig. 2 shows SDS-PAGE appearance of the purified myosin VI. The purified M6HMM wild type construct was composed of a high molecular mass band and a low molecular mass band and was free from 200-kDa Sf9 conventional myosin and actin. The high molecular mass band (121 kDa) was consistent with the calculated molecular mass of M6HMM and was recognized by anti-His antibodies (Santa Cruz Biotechnology, Inc., Santa Cruz, CA), indicating that the high molecular mass band is the expressed myosin VI heavy chain (not shown). The small subunits showed a mobility shift with a change in [Ca 2ϩ ] that is characteristic of calmodulin, suggesting that the small subunits are indeed calmodulin (not shown). The identification of the small subunit was also confirmed using anti-calmodulin antibodies (not shown). The stoichiometry of calmodulin versus myosin VI heavy chain was 1.2, consistent with the single IQ motif in M6HMM. Fig. 2 also shows the SDS-PAGE appearance of the T406A mutant of M6HMM, showing that it has the same peptide composition.
Phosphorylation of Myosin VI by PAK3-Based upon the sequence alignment of myosin VI with other myosins, it was noticed that mammalian myosin VI contains a potential phosphorylation site (Thr 406 ) in a loop near the tip of the head. It has been shown that the phosphorylation at a homologous position of the amoeba myosin I catalyzed by a PAK family kinase is required for its motor activity (i.e. its actin-activated ATPase activity and motility activity) (24,25). We therefore investigated whether myosin VI is phosphorylated at Thr 406 . Fig. 3 shows the phosphorylation of myosin VI by mammalian PAK3KD. Myosin VI heavy chain was stoichiometrically phosphorylated (Fig. 3A). To identify the site phosphorylated by PAK3KD, the phosphorylated myosin VI was subjected to Western blot analysis using anti-phosphothreonine 406-specific antibodies (Fig. 3B). The antibodies recognized the myosin VI heavy chain only after PAK3KD-induced phosphorylation but not before the phosphorylation. The signal strength of the Western blot increased over time with PAK3KD-induced phosphorylation, and the time course of the signal increase was similar to that of the extent of 32 P incorporation. Furthermore, the phosphorylation (incorporation of 32 P) by PAK3KD was markedly attenuated when the T406A mutant was used as a 2 M. Ikebe, unpublished observation. substrate (data not shown). These results clearly indicate that myosin VI heavy chain is phosphorylated by PAK3KD at Thr 406 .
Actin-activated ATPase Activity- Fig. 4 shows the actin dependence of M6HMM ATPase activity. The ATPase activity was markedly activated by actin. The actin concentration dependence of the activity showed a single saturation curve, and a K actin of 9 M was obtained. The value is significantly lower than those of nonprocessive conventional myosins. The actinactivated ATPase activity of M6HMM was also measured as a function of ATP concentration. As shown in Fig. 5, an extremely high concentration was required for the saturation of the steady state ATPase activity, and a K ATP value of 150 M was obtained. The effect of the phosphorylation at Thr 406 on the actin-activated ATPase activity was also studied (Figs. 4 and 5). M6HMM was phosphorylated by PAK3KD to stoichiometric level prior to the ATPase assay. The ATPase activity was not significantly affected by the phosphorylation, and similar values of V m , K ATP , and K actin were obtained for the phosphorylated and dephosphorylated M6HMM (Table I).
The actin-activated ATPase activity was measured as a function of free Ca 2ϩ (Fig. 6). Ca 2ϩ had no effect on the actinactivated ATPase activity in contrast to other calmodulin binding unconventional myosins so far reported such as myosin V and myosin I (17)(18)(19)(20)(21)(22)(23). Furthermore, Ca 2ϩ changed neither K actin nor K ATP of the actin-activated ATPase activity of myosin VI (Figs. 4 and 5; Table I).
Since the change in the actin-activated ATPase activity of mammalian myosin V and myosin I as a function of Ca 2ϩ has been attributed to the dissociation of bound calmodulin (17,18,23), we examined the dissociation of calmodulin from myosin VI. The myosin VI construct was mixed with F-actin in various free Ca 2ϩ concentrations and then ultracentrifuged to determine bound calmodulin. F-actin co-precipitated myosin VI with bound calmodulin was analyzed by SDS-PAGE, and the calmodulin band was quantitated by densitometry and normalized to myosin VI heavy chain. As a control, free calmodulin was centrifuged with F-actin, but no calmodulin co-precipita-tion was detected. The myosin VI construct was also tested for precipitation, but no myosin VI was precipitated in the absence of F-actin. As shown in Fig. 7, the amount of calmodulin bound to myosin VI was unchanged with Ca 2ϩ increase. Consistently, the addition of exogenous calmodulin did not change the actinactivated ATPase activity (Fig. 6). These results suggest that calmodulin does not dissociate from the myosin VI heavy chain when its binding of Ca 2ϩ induces a conformational change.
Inhibition of Actomyosin VI ATPase Activity by ADP-It has been shown that low concentration of ADP significantly inhibits the actin-activated ATPase activity of myosin V, and this is  Table I.  Table I. related to the strong binding of ADP to myosin V and the myosin V⅐ADP as a stable steady state intermediate (38,39). Fig. 8 shows the time course of the actin-activated ATPase reaction of M6HMM in the presence and absence of ATP regeneration system. In the absence of the ATP regeneration system, the rate of P i release was markedly inhibited.

Binding of Myosin VI and F-actin in the Presence of ATP-
Binding of M6HMM to F-actin was directly determined by an F-actin co-precipitation assay. Myosin VI was mixed with Factin in the presence of Mg 2ϩ -ATP, and the amounts of the bound myosin VI and the dissociated myosin VI were determined after centrifugation (see "Experimental Procedures"). A majority of myosin VI was co-precipitated with F-actin in the presence of ATP (Fig. 9). The dissociation constant was estimated from the actin concentration dependence of the binding to be K d of 1.9 M. It should be noted that M6HMM was readily dissociated from actin in the presence of Mg 2ϩ -ATP at high ionic strength (0.5 M KCl). The result suggests that the stable steady state intermediate of the myosin VI ATPase cycle is the strong actin-binding form.
Effect of Ca 2ϩ on the Motility Activity of Myosin VI-To evaluate the motor activity of myosin VI more directly, the purified myosin VI was subjected to in vitro motility assay. Fig.  10 shows the motility activity of M6HMM at various calcium ion concentrations. The motility activity was completely inhibited at Ca 2ϩ concentrations higher than pCa 6. The inhibition of the motility was not reversed by the addition of exogenous calmodulin (up to 12 M). The recovery of motility inhibited at high Ca 2ϩ was only achieved by reducing Ca 2ϩ concentration. The results are consistent with the fact that high Ca 2ϩ does not induce the dissociation of calmodulin from myosin VI heavy chain (Fig. 7) and suggest that the binding of Ca 2ϩ to the myosin VI-bound calmodulin light chain triggers the inhibition of the motility activity. While the reason for the apparent uncoupling between the ATPase activity and the motility activity at high Ca 2ϩ is unclear, it would be plausible that Ca 2ϩ might alter the affinity of myosin VI for actin. To address this possibility, the effect of Ca 2ϩ on actin-activated ATPase activity was measured as a function of actin concentration, but no significant effect of Ca 2ϩ was observed on K actin of the actinactivated ATPase activity of M6HMM (Fig. 4).
Effect of Thr 406 Phosphorylation on the Motility Activity of Myosin VI-The effect of Thr 406 phosphorylation on the myosin VI motor activity was more directly studied by measuring in vitro actin sliding activity. Fig. 11A shows the effect of Thr 406 phosphorylation by PAK3KD on the motility activity of myosin VI. There was little motility activity of M6HMM before incubation with PAK3KD, while a significant amount of F-actin filaments moved with the velocity of 0.4 Ϯ 0.1 m/s after incubation of M6HMM with PAK3KD. The addition of stauro-  sporine, a nonselective protein kinase inhibitor, significantly attenuated the PAK3KD-induced activation of the motility activity. The result suggested that PAK3KD-induced phosphorylation activated the motility activity of M6HMM. To evaluate this view, the phosphorylation level of M6HMM at Thr 406 was monitored by Western blot using anti-phosphorylated Thr 406specific antibodies as probes (Fig. 11B). There was a trace amount of the phosphorylated M6HMM observed before incubation with PAK3, while a significant amount of phosphorylation was observed after incubation of M6HMM with PAK3KD. The addition of staurosporine significantly diminished the phosphorylation, although it did not completely abolish the phosphorylation. These results clearly indicated that the phosphorylation of M6HMM at Thr 406 activates the motility activity of myosin VI. Interestingly, T406A mutant as well as T406E mutant showed in vitro motility activity similar to that of the phosphorylated wild type M6HMM (not shown). It has been shown that the mutation of the phosphorylatable Ser to Ala mimics the phosphorylated state of the protein rather than the dephosphorylated state, suggesting that the hydroxyl side chain of Ser plays a critical role in the dephosphorylated state (40). The present study provides another example of the Ala mutation mimicking the phosphorylated state.
It should be mentioned that a few F-actin filaments moved on M6HMM even before incubation with PAK3KD with similar velocity to the PAK3KD-phosphorylated M6HMM. But this is likely to be due to the presence of some prephosphorylated myosin VI, and there was a trace amount of phosphorylated M6HMM before incubation with PAK3KD (Fig. 11B). Consistently, it was found that the M6HMM preparation contains some trace myosin VI kinase activity (not shown).

DISCUSSION
During the last decade, a number of new myosins have been found that lack thick filament forming ability and are thought to play a role in diverse cellular contractile/motile functions. A key issue is how these myosins determine their specific role in a particular cellular function. One aspect of this issue involves the characteristics of motor activity such as processivity, actin translocating velocity, force, and directionality of movement. In this regard, myosin VI is a unique motor that moves on actin filaments toward the minus end (12) in a manner opposite to other known myosin family members. As a second factor, there may be a myosin-specific targeting site that determines each myosin's cellular binding partner and subcellular localization. The tail portion of each myosin is a good candidate region for FIG. 9. Binding of acto-M6HMM in the presence of ATP. The co-sedimentation assay was performed as follows. M6HMM (0.1 mg/ml) was incubated in buffer containing 20 mM imidazole, pH 7.5, 50 mM KCl, 2 mM ATP, 3 mM MgCl 2 , and various concentrations of F-actin in the presence of an ATP regeneration system (20 units/ml pyruvate kinase and 3 mM phosphoenolpyruvate) at 25°C for 30 min. The samples were ultracentrifuged at 100,000 ϫ g for 30 min, and the pellets were analyzed by SDS-polyacrylamide gel electrophoresis. The amount of the co-sedimented heavy chain was determined by densitometry as described previously (20). achieving this specificity, since the structure of this domain is quite diverse. The third factor determining cellular specificity of myosin function is that the regulatory cascades that turn each myosin's activity on could also be distinct, thus allowing separate activation of the different types of myosin. However, little is known for the regulatory mechanism of these unconventional myosins. The aim of this study is to clarify the motor function and regulation of mammalian class VI myosin.
The present study revealed a dual regulation of myosin VI by PAK3-induced phosphorylation and by Ca 2ϩ binding to calmodulin. The former is required for the activation of myosin VI motility activity, and the latter inhibits the motility activity. The phosphorylation of myosin VI was previously observed with a bacteria-expressed glutathione S-transferase fusion fragment (amino acids 308 -631) of myosin VI (41), but the present study is the first to identify the phosphorylation site as Thr 406 , because 1) the anti-phospho-Thr 406 -specific antibodies recognized myosin VI phosphorylated by PAK3 and 2) 32 P incorporation was markedly attenuated by the mutation of Thr 406 to Ala.
We concluded that the phosphorylation of myosin VI is required for the activation of the motility activity of myosin VI, based upon the following results: 1) the incubation of myosin VI with PAK3 significantly increased the number of moving actin filaments, and this is accompanied by the increase in Thr 406 phosphorylation, and 2) a protein kinase inhibitor, staurosporine, significantly attenuated the number of sliding actin filaments, and this is correlated with the decrease in the phosphorylation at Thr 406 . It should be mentioned that a few actin filaments moved without prephosphorylation with PAK3, but this is likely to be due to the presence of a low amount of phosphorylated myosin VI. The amount of prephosphorylated myosin VI varied from preparation to preparation, and we have prepared M6HMM that is entirely inactive in actin sliding activity without phosphorylation by PAK3. The results indicate that the phosphorylation of myosin VI is critical for the activation of the motility activity.
The phosphorylation of an analogous site has been reported for amoeba myosin I (24). In the case of amoeba myosin I, the phosphorylation is required for the actin activation of the ATPase activity. This is quite different from the regulation of myosin VI, in which the phosphorylation is necessary for the motility activity but not the actin-activated ATPase activity. The apparent mechanism for the uncoupling between the ATP hydrolysis and the motility is unknown. It has been suggested that a change in the rigidity of the lever arm domain of myosin could uncouple the motility and ATPase activity of myosin (42); however, this is less likely, since the location of Thr 406 is near the tip of the head of myosin. Another possibility is a change in the fraction of the strong and weak binding states induced by phosphorylation that is in turn related to the "processivity" of myosin. It has been shown recently that vertebrate myosin V is a processive motor (43,44) and that the processivity is closely related to the identity of the stable intermediate of the myosin V ATPase reaction as myosin⅐ADP, a strong actin binding intermediate (38). This is also less likely, since during the ATPase reaction the apparent affinities (K actin ) for actin of phosphorylated and dephosphorylated myosin VI are similar to each other. The mechanism by which phosphorylation activates the motility activity of myosin VI requires further investigation.
While the phosphorylation is necessary for the activation of the motility activity of myosin VI, Ca 2ϩ inhibited the motility activity. The loss of motility at high Ca 2ϩ was not due to the denaturation of myosin VI, since perfusion of the flow cell with low Ca 2ϩ buffer (below pCa 7) completely restored the motility (Fig. 10). We examined the effect of various exogenous calmodulin concentrations up to 12 M; therefore, it is unlikely that the loss of motility is due to unsaturation of bound calmodulin. The Ca 2ϩ -dependent inhibition of mammalian unconventional myosin was first demonstrated with myosin I (17)(18)(19)(20)(21) and then subsequently with myosin V (22,23). It was thought originally that the inhibition of motility was due to the dissociation of calmodulin from the heavy chain at high Ca 2ϩ (17,18). However, recent studies have revealed that the inhibition does not require the dissociation of calmodulin, but rather Ca 2ϩ binding at the high affinity sites of calmodulin triggers the inhibition presumably due to a large conformational change of calmodulin (20 -22). For myosin V, Ca 2ϩ -dependent regulation requires at least two calmodulin binding sites (22,23), and a truncated mutant having a single calmodulin was constitutively active (22,23). Interestingly, the present study revealed that myosin VI, also binding a single calmodulin, is regulated by Ca 2ϩ . The bound calmodulin was not dissociated from the heavy chain at high Ca 2ϩ ; nevertheless, myosin VI motility was inhibited. This is consistent with the most recent results with myosin I␤ and myosin V and supports the idea that calmodulin dissociation is not the inhibitory mechanism of calmodulin binding myosins. The reason why the IQ/calmodulin of myosin VI, but not the corresponding first IQ/calmodulin of myosin V, plays a role in the regulation of motility is unclear. But it is plausible that the position of the first calmodulin of myosin VI resembles the second calmodulin site for myosin V, since a unique large insert is present between the converter and the IQ domain of myosin VI.
The inhibition occurs between pCa 7 and pCa 6 where cytoplasmic Ca 2ϩ concentration is regulated in most cell types, and therefore the observed inhibition is physiologically relevant. 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 abolition of the C-terminal Ca 2ϩ binding sites abolishes inhibition of the motility of myosin I␤ (21).
There is an apparent decoupling between the ATP hydrolysis cycle and mechanical events at higher Ca 2ϩ . One possibility is a change in the "processivity" of myosin VI by Ca 2ϩ , but this is not the case, since Ca 2ϩ did not affect the K actin of the actinactivated ATPase reaction. This is different from the regulation of conventional myosins in which the regulatory domain regulates both the ATPase and mechanical activities (13)(14)(15). 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 VI, thus inhibiting motility.
A question is whether myosin VI is a processive motor. As shown in Fig. 9, myosin VI strongly binds to actin in the presence of Mg 2ϩ -ATP. It is also revealed that the actin-activated ATPase activity is markedly decreased with time, and this is due to the ADP produced during the course of myosin VI ATPase reaction, because the decrease in the activity with time is completely eliminated in the presence of an ATP regeneration system. The result suggests that myosin VI has a strong ADP binding relative to the ATP binding. Consistently, K ATP was extremely high for the myosin VI ATPase reaction (150 M), and a much lower K ADP value (3 M) is estimated. This result is consistent with the strong actin binding nature of myosin VI in the presence of Mg 2ϩ -ATP and suggests that the ADP-bound form (i.e. a strong actin binding state) is the pre-dominant steady state intermediate of myosin VI ATPase cycle. It should be noted that even at low phsphorylation level, a few actin filaments moved with a velocity similar to that of the highly phosphorylated myosin VI. The result suggests that the active form of myosin VI can support the actin filament sliding at very low surface density of the active myosin molecules. These properties of myosin VI are similar to those of myosin V, a processive motor (38,43,44). Therefore, it is anticipated that myosin VI has a processive nature in its motor function, although more direct evidence is required to determine the processivity of myosin VI.