Enzymatic characterization and functional domain mapping of brain myosin-V.

The actin binding and ATPase properties, as well as the functional domain structure of chick brain myosin-V, a two-headed, unconventional myosin, is reported here. Compared to conventional myosin from skeletal muscle, brain myosin-V exhibits low K-EDTA- and Ca-ATPase activities (1.8 and 0.8 ATP/s per head). The physiologically relevant Mg-ATPase is also low (∼0.3 ATP/s), unless activated by the presence of both F-actin and Ca2+ (Vmax of 27 ATP/s). Ca2+ stimulates the actin-activated Mg-ATPase over a narrow concentration range between 1 and 3 μM. In the presence of saturating Ca2+ and 75 mM KCl, surprisingly low concentrations of F-actin activate the Mg-ATPase in a hyperbolic manner (KATPase of 1.3 μM). Brain myosin-V also binds with relatively high affinity (compared to other known myosins) to F-actin in the presence of ATP, as assayed by cosedimentation. Digestion of brain myosin-V with calpain yielded a 65-kDa head domain fragment that cosediments with actin in an ATP-sensitive manner and a 80-kDa tail fragment that does not interact with F-actin. The 80-kDa fragment results from cleavage one residue beyond the proline-, glutamate-, serine-, threonine-rich region. Our data indicate that the Mg-ATPase cycle of brain myosin-V is tightly regulated by Ca2+, probably via direct binding to the calmodulin light chains in the neck domain, which like brush border myosin-I, results in partial (∼30%) dissociation of the calmodulin associated with brain myosin-V. The effect of Ca2+ binding, which appears to relieve suppression by the neck domain, can be mimicked by calpain cleavage near the head/neck junction.

Although the conventional myosins (myosins-II) responsible for processes such as muscle contraction and cytokinesis have been intensively studied, relatively little is known about the properties of the unconventional myosins. One widely distributed group of unconventional myosins, the class V myosins, have been suggested to play a role in organelle transport or membrane targeting (reviewed in Refs. 1 and 2). Brain myosin-V (BM-V 1 ), which was originally identified as a calmodulin (CaM)-binding protein in vertebrate brain, is the only member of this class of unconventional myosins to be purified and characterized biochemically (3)(4)(5)(6)(7). The class V myosins share a common domain structure consisting of an N-terminal head domain containing the actin-binding and ATP hydrolysis sites, an extended neck domain containing six IQ motifs (which form binding sites for CaM and/or related light chains), and a tail domain consisting of a region predicted to form coiled coils attached to a globular region of unknown function (7,8). The hypothetical functions of the class V myosins and their novel domain structure, particularly the extended neck domain, raise the obvious question of how the basic biochemical properties of the class V myosins compare with those of other types of myosins.
Among the critically important properties of a molecular motor are its steady-state ATPase activity and the factors which regulate this activity. Known myosins are regulated either indirectly via actin-associated proteins such as troponin/ tropomyosin (as in vertebrate skeletal muscle myosin) or directly via the subunits of the myosin molecule itself. In the latter type of direct "myosin-linked" regulation, the light chains associated with the neck domain often function as the regulating subunits. The myosin light chains are all members of the CaM superfamily of proteins, although not all of them have retained the ability to bind to Ca 2ϩ . Vertebrate non-muscle and smooth muscle myosins-II are "turned on" by phosphorylation of their regulatory light chains by Ca 2ϩ /CaM-dependent myosin light chain kinase (9), whereas molluscan myosin-II and vertebrate brush border (BB) myosin-I are regulated by the direct binding of Ca 2ϩ to their light chains (10,11). Like BB myosin-I, BM-V has multiple CaM light chains that remain bound in the absence of Ca 2ϩ (3,5,7). The neck domain of BM-V has been shown to be the precise site of CaM binding (8), suggesting that this myosin might also be directly regulated by Ca 2ϩ binding. The initial reports of BM-V ATPase activity also indicated that this protein is activated by Ca 2ϩ (5,7), although the K ATPase for Ca 2ϩ and F-actin were not determined.
The tail domains of myosins are unique to each class in the myosin superfamily and probably reflect the specific functions * This work was supported by grants from Fundacao de Amparo a Pesquisa do Estado de Sã o Paulo and Conselho Nacional de Desenvolvimento Cientifico e Tecnologico (to R. E. L.) and by the following grants (to M. S. M.): a basic research grant from the Muscular Dystrophy Association, National Institutes of Health Program Project Grant DK38979, and National Institutes of Health Grant DK25387. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18  and/or targets for each individual myosin (1). Although the proximal region of the BM-V tail appears to lead to dimer formation via coiled coil segments, BM-V does not form myosin filaments (7). The medial region of the BM-V tail contains a PEST (proline-, glutamate-, serine-, threonine-rich) region (8), a sequence motif which has been hypothesized to be associated with intracellular proteolysis by calpain (12,13). BM-V is, in fact, a substrate for limited proteolysis by calpain (5,6), which results in a stable 65-kDa head fragment and a stable 80-kDa tail fragment. The function of the globular C-terminal tail domain that characterizes the class V myosins is unknown, although this domain shares sequence homology with the human protein AF-6 and the related Drosophila protein, canoe (14).
In the present study, we have determined the ATPase properties of highly purified BM-V from freshly hatched chicks and demonstrate that the actin-activated Mg-ATPase activity is tightly regulated by physiologically relevant concentrations of Ca 2ϩ . The maximal actin-activated Mg-ATPase of BM-V is quite high and is comparable to that of skeletal muscle myosin. Surprisingly, however, myosin-V is activated by much lower concentrations of F-actin than is skeletal muscle myosin. We have taken advantage of the cleavage of BM-V by calpain to obtain information on the functions and interactions of the domains of BM-V. Our data indicate that, similar to other regulated myosins, the neck domain of BM-V has a suppressive effect on the ATP-hydrolysis/actin-binding cycle of the head domain. This suppression can be relieved by binding to Ca 2ϩ or by specific cleavage near the head/neck junction. We have also localized one of the two major calpain cleavage sites in BM-V to a site in the tail just one amino acid downstream from the PEST region. The 80-kDa tail fragment generated by this cleavage includes the globular domain of unknown function. This domain did not bind to actin in cosedimentation assays and does not appear to contain a second actin-binding site.
Proteins-Myosin-V was purified from chick brain essentially as described by Cheney et al. (7), except for a change in the initial centrifugations from 18,000 to 40,000 ϫ g and the addition of the protease inhibitors Pefabloc (1 mM), benzamidine (1 mM), and aprotinin (2 g/ ml). These modifications and improved assay conditions resulted in higher yields and higher maximum ATPase activity than previously obtained. Actin was purified from chicken pectoralis muscle by the method of Spudich and Watt (15). Affinity-purified antibodies to the myosin-V tail ("tail antibody"; Ref. 8) and a monoclonal antibody, MaB CX-1, made to BB myosin-I that reacts with the N-terminal third of the BB myosin-I head as well as with BM-V ("head antibody"; Ref. 16) were used for immunoblot analysis.
ATPase Assays-ATPase activity was assayed at 37°C using the colorimetric method of Taussky and Shorr (17). The K-EDTA-, Ca-, and Mg-ATPase activities of 0.06 M BM-V were measured over 15 min; actin-activated Mg-ATPase activities of 0.02 M BM-V were determined for 5-min reactions. The hydrolysis of ATP, expressed as ATP/s per myosin head, was calculated by considering the molecular mass of native myosin-V dimers to be 640 kDa (7). The final conditions used in the assays were as follows: K-EDTA-ATPase, 20 mM imidazole, pH 7. Ca 2ϩ on the actin-activated Mg-ATPase activity was assessed by the addition of 4 mM Ca-EGTA buffers to the standard assay, covering a range of free calcium from 10 Ϫ7 to 10 Ϫ4 M. Free calcium was calculated by the Mcalc program of Fabiato (18), using the Schwartzenbach dissociation constants for Ca-EGTA. The dependence of Mg-ATPase activity on actin concentration was determined by addition of F-actin over a range of concentrations from 0.1-60 M. Phalloidin (1:1 molar ratio with actin) was added to solutions containing 0.1 to 10 M actin to stabilize the actin as filaments. The ATPase activity of F-actin alone was measured at each concentration under the same reaction conditions, and its contribution was subtracted from the corresponding actinactivated activity of BM-V.
Actin Binding Assays-The binding of BM-V to F-actin in the presence of ATP was assessed using assay conditions identical to those used for the Mg-ATPase assays, except that the Mg-ATP concentration was 10 mM and the pH was 7.1. BM-V (0.03 M) was mixed with phalloidinstabilized F-actin at 0, 0.5, 1.0, 5.0, and 10 M, in either 4 mM EGTA or 4 mM Ca 2ϩ /EGTA buffer (4 mM CaCl 2 , 4 mM EGTA; ϳ11 M free Ca 2ϩ ), and then centrifuged for 1 h at 100,000 ϫ g. The resulting pellet and supernatant fractions were analyzed by SDS-PAGE, and the percentage of BM-V cosedimented with F-actin was determined by densitometry. Identification of the Primary Calpain Cleavage Site on BM-V-The 80-kDa fragment, generated by calpain cleavage of 0.5 mg of BM-V, was purified by anion exchange chromatography on a 1-ml Q-Sepharose column equilibrated in 10 mM imidazole, pH 7.4, containing 75 mM KCl, 2.5 mM MgCl 2 , 0.1 mM EGTA, and 2 mM DTT. Fractions (0.5 ml) were collected during elution of the column with 30 ml of 75-1000 mM KCl gradient in column buffer, at a flow rate of 10 ml/h, and analyzed by SDS-PAGE. Those containing the pure 80-kDa fragment were pooled, precipitated with ice-cold 5% trichloroacetic acid, and submitted to N-terminal microsequencing at the Yale University W. M. Keck Foundation Biotechnology Resource Laboratory.
Amino Acid Analysis of 65-and 80-kDa Fragments of BM-V-The 65and 80-kDa fragments, generated by calpain cleavage of BM-V, were purified by reverse phase high-performance liquid chromatography on a C 4 column (Vydac, Hesperia, CA) using a 12-80% acetonitrile gradient in 0.1% trifluoroacetic acid. Approximately 860 ng of the 65-kDa and 970 ng of the 80-kDa fragments were vacuum dried, hydrolyzed in 6 N HCl, derivatized with phenyl isothiocyanate, and analyzed according to Bidlingmeyer and Tarvin (20) on a model MP3000E binary system from Laboratory Data Control (Milton Roy Co., Riviera Beach, FL).
Other Methods-SDS-PAGE was performed using 5-16% linear gradient minigels. Immunoblot analysis was performed as described by  Table I are strikingly low, especially when compared with the Mg-ATPase observed in the presence of both F-actin and Ca 2ϩ (Fig. 1). BM-V exhibits little or no actin-activated Mg-ATPase activity in the absence of Ca 2ϩ . Even in the presence of Ca 2ϩ , BM-V by itself does not manifest substantial Mg-ATPase activity. Only if both F-actin and Ca 2ϩ are present is ATP rapidly hydrolyzed. These data show that BM-V is tightly regulated by Ca 2ϩ in a manner reminiscent of molluscan myosin-II (26) and vertebrate myosin-I (19,23,27,28).
The actin-activated Mg-ATPase activity of BM-V is further stimulated by the addition of bovine brain CaM (Fig. 1). The degree of this stimulation varies somewhat from preparation to preparation, perhaps due to loss of CaM during purification or storage. Since the addition of higher concentrations of CaM did not lead to further increases in ATPase activity (data not shown), 5.9 M CaM was included in all subsequent Mg-ATPase assays to guarantee maximal activity.
Regulation of the Actin-activated ATPase Occurs over a Narrow Range of Ca 2ϩ Concentration-The effect of Ca 2ϩ on the actin-activated Mg-ATPase activity of BM-V is shown in Fig. 2. The activation occurs over a very narrow range of Ca 2ϩ concentration, with no significant effect detected up to about 1 M and a near maximum effect at about 3 M Ca 2ϩ (50% activation at 1.8 M Ca 2ϩ ). Although there are few data points at the critical inflection region, a Hill plot of the available data yielded an n value of 13, suggesting a high degree of cooperativity. Even at concentrations of Ca 2ϩ up to 1 mM, BM-V maintained the high rate of ATP turnover (data not shown). To determine whether this range of Ca 2ϩ results in dissociation of CaM from BM-V, as has been observed for BB myosin-I (11), two different preparations of BM-V were used to examine the effect of elevated Ca 2ϩ on CaM light chain content. Using actin cosedimentation to separate heavy chain bound and free CaM light chains (19), we observed partial dissociation of CaM (27-32%; 1-1.3 CaM/heavy chain) at Ca 2ϩ concentrations above 5 M, with no further increase even at 1 mM Ca 2ϩ .
BM-V Is Activated at Low Actin Concentrations and Hydrol-yses ATP at a Fast Rate-We determined the effect of the actin concentration on the Mg-ATPase activity in the presence of saturating Ca 2ϩ . F-actin activates the Mg-ATPase activity in a simple hyperbolic manner (Fig. 3A). The dependence of the Mg-ATPase activity on actin concentration was also assayed at various myosin concentrations (data not shown). We observed no evidence of the triphasic activation seen in the Acanthamoeba myosin-I, which is due to the presence of a second actin-binding site in the tail domain (25). A double reciprocal  plot of the BM-V data (Fig. 3B) yielded a V max of 27 ATP/s per myosin head and a K ATPase of 1.3 M. This value for maximal actin-activated Mg-ATPase activity puts BM-V amongst the enzymatically fastest myosins, including skeletal muscle myosin-II (24), molluscan myosin-II (26), and amoeba myosin-I (25,29,30). Interestingly, however, the K ATPase value suggests that BM-V binds much more tightly to F-actin than these other myosins. Consistent with this finding, BM-V exhibits relatively high levels (compared to other known myosins) of binding to actin in the presence of ATP, particularly in the presence of Ca 2ϩ (Fig. 4), as assessed by cosedimentation with F-actin under solution conditions comparable to those used in the Mg-ATPase assay, except that the Mg-ATP concentration was 10 mM (to insure that ATP depletion did not occur). Comparable results to those summarized in Fig. 4 were seen in the presence of exogenously added CaM (results not shown). This binding is actin concentration-dependent, whereas no such dependence can be detected for binding under rigorous conditions, where Ͼ90% pelleting of BM-V is observed at all actin concentrations tested (results not shown).
Calpain Cleaves BM-V at the PEST Site in the Tail and near the Head/Neck Junction-We have previously shown that the calcium-dependent protease calpain cleaves BM-V in a limited manner in vitro and generates two stable polypeptides with apparent molecular masses of 65 and 80 kDa, which correspond to regions of the head and tail of BM-V, respectively (5,6). In the present study, the time course of proteolysis of purified BM-V by calpain was determined (Fig. 5A). BM-V was rapidly cleaved by calpain generating 80-and 130-kDa fragments. The 130-kDa fragment was further proteolyzed to a 65-kDa fragment. Both the 130-and 65-kDa fragments were recognized by head antibodies (Fig. 5B), whereas the 80-kDa fragment was recognized only by tail antibodies (Fig. 5C). A diagram of the putative localization of these sites is shown in Fig. 5D.
The precise cleavage site that generates the 80-kDa fragment was determined by microsequencing of the N-terminal of the purified 80-kDa polypeptide (Fig. 5E) and is located between Arg 1140 and Met 1141 , just one amino acid downstream from the PEST site. Surprisingly, analysis of the microsequence data revealed the presence of two sequences in about equal amounts differing only by a single amino acid, the insert of a Gln as the third amino acid in one of the peptides (Fig. 5E). This same difference in a single amino acid appears in the two published primary sequences of chicken BM-V based on cDNA ( Fig. 5F; Refs. 8 and 31) and thus may reflect different alleles or a single amino acid splicing variant.
Since the polypeptide corresponding to the C-terminal side of the second cleavage site has not been isolated, the exact site that generates the 65-kDa fragment has not yet been determined by microsequencing. We have, however, determined the approximate location of this cleavage site based on the apparent molecular mass in SDS-PAGE and a comparison of the amino acid composition of the purified 65-kDa fragment with the deduced amino acid composition of BM-V. This analysis, which is described in the legends of Table II and Fig. 6, suggests that the second cleavage site occurs near amino acid 629 Ϯ 2 amino acids of BM-V (Fig. 6). Based on sequence alignment with skeletal myosin-II, this would place the cleavage site near the 50-kDa/20-kDa head domain junction in loop 2 of the actin-binding site (1,8,32).
The 65-kDa Head Fragment of BM-V, but not the 80-kDa Tail Fragment, Binds to F-Actin-The actin binding properties of the head and tail fragments of BM-V were assessed by actin cosedimentation in the absence and presence of ATP (Fig. 7). The 65-kDa head fragment exhibited ATP-dependent cosedimentation with F-actin. The 80-kDa fragment, on the other hand, remains in the supernatant under both conditions, indicating that it does not bind to actin filaments.
Calpain Proteolysis Deregulates the Actin-activated ATPase Activity of BM-V-After various incubation times with calpain, the actin-activated Mg-ATPase activity of BM-V was determined in the presence of 0.5 mM Ca 2ϩ or 5 mM EGTA (Fig. 8).
After proteolysis for 30 -60 min, when intact BM-V was no longer detected, 65% of the actin-activated ATPase, originally Ca 2ϩ -dependent, was retained. On the other hand, the Ca 2ϩindependent, actin-activated ATPase activity increased with time of proteolysis, eventually attaining rates of ATP turnover equal to those in the presence of Ca 2ϩ . The low levels of ATPase activity observed in the absence of actin were not affected by proteolysis (data not shown). Thus, as seen in the insert to Fig.  8, cleavage by calpain leads to a loss of the Ca 2ϩ -regulated ATPase activity while the actin-activated Mg-ATPase remains high. This "deregulation" is similar to that reported for molluscan myosin-II and BB myosin-I after proteolytic cleavage near their neck domains (10,19). DISCUSSION The purification of myosin-V from vertebrate brain has allowed for the characterization of the mechanochemical properties and functional domain organization of now a third class amongst the 10 -11 classes of the myosin superfamily of molecular motors (reviewed in Ref. 1). Thus, in addition to myosin-II and myosin-I, which have been extensively studied as native proteins, there is now enough biochemical data on BM-V to allow for an initial description of the domain functions and mechanochemical properties of a class V myosin, as well as a comparison with specific class I and class II myosins. Sequence analysis has shown that, although the different classes of myosins have strikingly different tail domains, their head domains are conserved in general structure and organization (1,33,34). Purified myosins, however, exhibit quantitative differences in mechanochemical properties, such as ATPase activity and sliding filament velocity, as well as substantial differences in regulation. It is now realized that all myosins contain a distinct neck domain responsible for light chain binding, which can have a regulatory role and perhaps also function as a lever for the generation of movement (33,(35)(36)(37)(38). Because the neck domain of the class V myosins is much larger than the neck domain of the conventional myosins, there is much interest in comparing the mechanochemical properties of these myosins.
Mg 2ϩ -ATPase Activity of BM-V-The results presented here indicate that the actin-activated Mg-ATPase of BM-V is comparable in rate to that of skeletal muscle myosin-II. However, a striking difference between skeletal muscle myosin and BM-V is that BM-V is activated by very low concentrations of F-actin. Since the ability of skeletal muscle myosin to bind to actin is greatly inhibited as the salt concentration increases, the magnitude of this difference may be even higher at truly physiological salt concentrations. Thus, BM-V appears to bind much more tightly to actin in the presence of ATP than does skeletal muscle myosin. Indeed, this binding was directly observed by cosedimentation of BM-V with F-actin under conditions comparable to those used in the ATPase assays (Fig. 4). Although Ca 2ϩ enhances the binding of BM-V to actin in the presence of ATP, significant binding of BM-V to actin is seen in the absence of Ca 2ϩ as well. Thus, the actin binding is much less stringent in its requirement for Ca 2ϩ than the ATPase activation. This raises the possibility that the binding of BM-V to actin in the presence of ATP is mediated by a second, ATP-insensitive site, as observed in the tail domain of amoeboid myosins-I (25). This possibility cannot be ruled out by our present data. However, two lines of evidence suggest that this is not the case. (a) The actin activation of BM-V Mg-ATPase does not exhibit the triphasic behavior characteristic of amoeboid myosins-I. (b) Assessment of the actin binding properties of the calpain cleavage products of BM-V indicates that the 80-kDa tail fragment does not bind F-actin.
The activation of BM-V by low concentrations of actin and its relatively high affinity for actin in the presence of ATP both suggest that this myosin could function in regions of low actin filament concentration. This might be the case if, as has been suggested, BM-V functions as a motor for organelle transport.
BM-V Is Tightly Regulated by Ca 2ϩ via a Myosin-linked System-The enzymatic studies presented here clearly show that BM-V is a regulated myosin. Maximum activation of Mg-ATPase activity occurred in the presence of about 20 M purified skeletal muscle F-actin and 3 M Ca 2ϩ . Removing Ca 2ϩ at the same actin concentration resulted in a greater than 90% inhibition of the ATPase activity. In fact, in most preparations, the degree of inhibition was closer to 99% (Figs. 1, 2, and 8). The concentration range of Ca 2ϩ required to stimulate the ATPase is within a physiologically relevant realm, which suggests that Ca 2ϩ may regulate the activity of this motor in vivo. Although we have previously shown that the tail domain of BM-V is a substrate for Ca 2ϩ /CaM-dependent protein kinase (6), protein kinase activity was not detected in the purified BM-V used in the present studies. 2 Therefore, the Ca 2ϩ regulation of BM-V is similar to the myosin-linked mechanisms observed for the Ca 2ϩ regulation of molluscan myosin-II and vertebrate myosin-I. Regulation in molluscan myosin-II involves light chain mediated Ca 2ϩ binding (38), which results in tight control over the phosphate release step of the ATPase cycle (39). In BB myosin-I, multiple copies of CaM serve as Ca 2ϩ -binding light chains, which stimulate Mg-ATPase activity in the presence of Ca 2ϩ (11). Although BM-V is similar to BB myosin-I in having multiple CaMs bound to the neck region, the actin-activated Mg-ATPase of BM-V is much more tightly regulated by Ca 2ϩ . It is also important to note that in addition to the 4 CaM light chains, per heavy chain, BM-V also has 1 each of two distinct light chains of 17-and 23-kDa, which may also interact with the neck domain and participate in regulation (7).
The highly cooperative nature of the Ca 2ϩ activation curve in the range of 1-3 M (Fig. 2) is reminiscent of the Ca 2ϩ activation curves of other CaM-regulated enzymes (40) and is consistent with known affinity constants for Ca 2ϩ binding to CaM determined under ionic conditions similar to those used here (41). Thus our data support the hypothesis that Ca 2ϩ is acting on BM-V via binding to its CaM light chains. Based on studies of BB myosin-I, two, not necessarily mutually exclusive, modes of CaM light chain regulation have been proposed (reviewed in Ref. 11). One is through Ca 2ϩ -dependent dissociation of CaM from the heavy chain, and we report here that Ca 2ϩ in the micromolar range does indeed result in partial (ϳ30%) CaM light chain dissociation. The other mode is through allosteric regulation induced by Ca 2ϩ binding to neck-bound CaM light chains.
A number of key questions remain. What is the nature and location of the putative non-CaM light chains? What role could Ca 2ϩ binding to distal CaMs play in the regulation of ATP turnover in the head, if any? Is the structural basis for regulation via light chains basically the same for all myosins, no matter how large their neck domain?
Calpain Cleavage of the BM-V Tail-The Ca 2ϩ -dependent protease calpain is known to cleave a number of proteins into stable domains, and in several cases cleavage by calpain has been suggested to play a physiological role (13). The localization of a calpain cleavage site one amino acid away from the PEST region is an intriguing finding since PEST regions have previously been hypothesized to be associated with cleavage by calpain (13). Our results support the idea that PEST regions form specific calpain recognition sites or form flexible interdo-2 F. Espindola, unpublished data. performed as described under "Experimental Procedures." At the times indicated, aliquots of the reaction mix were prepared for SDS-PAGE and immunoblots. A, Coomassie Blue-stained gel of these samples; B, Western blot of the same samples probed with the "head" antibody as described under "Experimental Procedures"; C, Western blot of the same samples probed with the myosin-V "tail" antibody. hc, the intact BM-V heavy chain. Polypeptides of 130, 80, and 65 kDa are indicated to the right of the figure, and positions of molecular mass standards are indicated to the left. A diagram of the fragments generated by calpain on BM-V is shown in D. The prominent 2 amino acids of each of 10 cycles of microsequence data from the N-terminal of the 80-kDa fragment is shown in E. Where only one amino acid is given, that amino acid was more than three times more abundant than any other amino acid at that cycle. pmol values are given in parentheses. The corresponding sequences derived from cDNA sequencing from two laboratories is given in F. Calpain Cleavage of BM-V Leads to a Loss of Ca 2ϩ Regulation-Although the majority of actin-activated Mg-ATPase (65%) of BM-V is retained after limited proteolytic digestion by calpain, this ATPase is no longer regulated by Ca 2ϩ (Fig. 8).
The 65-kDa head domain resulting from calpain cleavage also cosediments with actin in an ATP-sensitive manner (Fig. 7). Thus, the abilities to hydrolyze ATP and to bind actin are retained in spite of calpain cleavage. We have used indirect methods to determine that the cleavage site producing the 65-kDa fragment is located near residue 629, at a position a The molar ratio of amino acids to polypeptide was calculated from the pmol values determined by amino acid analysis of the polypeptide, as described in "Experimental Procedures," using a molecular weight value calculated from the known primary sequency by arbitrarily choosing a possible cleavage site. For example, the values expressed in the Table are based on: (a) for the polypeptide with an apparent molecular mass of 65 kDa, -a cleavage site 630 amino acids from the N-terminal, which gives a molecular weight of 71,793; and (b) for the polypeptide with an apparent molecular weight of 80 kDa, a cleavage site at Arg 1140 , giving a molecular weight of 79,319 for the C-terminal polypeptide.
b Once the cleavage site is choosen, the exact molar ratio of amino acids is determined from the known sequence. c The difference between the calculated molar ratio based on amino acid analysis and the exact molar ratio based on sequence is expressed as percentage of the exact molar ratio. d Met, Cys and Trp were not determined by the amino acid analysis. Therefore, absolute values were introduced into the calculations based on the sequence data.
FIG. 6. Prediction of the calpain cleavage site on BM-V that generates the 65-kDa polypeptide. The sum of the differences between the calculated molar ratios based on amino acid analysis versus molar ratios based on the known sequence was determined as described in Table II for several chosen amino acid positions (⌺ analysis-sequence) and plotted as a function of that position. The rationale was that the closer the fit between the molar ratio based on amino acid composition and the known sequence, the closer one would be to the cleavage site. This analysis shows a clear minimum, which points to a possible cleavage site near amino acid 629 Ϯ 2 amino acids. To test this rationale, we performed the same procedure on the 80-kDa polypeptide, the exact cleavage site of which has been determined by microsequencing to be at Arg 1140 . This analysis also showed a clear minimum (amino acid 1136 Ϯ 5 amino acids) that included the known site. As an additional control, a 629-amino acid sequence from the C-terminal of BM-V was compared by this method to the 65-kDa fragment. The sum of the differences in this case was 614, a value well above the range of values obtained near the putative site.  1 and 2) and BM-V calpain fragments cosedimented in the absence (lanes 3 and 4) or presence (lanes 1, 2, 5, and 6) of 9 M F-actin in the absence (lanes 1, 3, and 5) and presence (lanes 2, 4, and 6) of 2 mM ATP. Positions of actin (A), the intact BM-V heavy chain (hc), and the 80-and 65-kDa fragments are indicated. equivalent to the proteolytically sensitive 50-kDa/20-kDa junction found in conventional myosins (reviewed in Ref. 42). On a gel filtration column, the 65-kDa fragment, together with CaM, elutes well before the volume expected for a globular 65-kDa protein. 3 These data suggest that, in spite of the cleavage, non-covalent domain interactions are maintained between the 65-kDa segment and at least a part of the neck domain with its CaM.
Other light chain-regulated myosins have suppressive neck domains that, in the absence of Ca 2ϩ , are believed to prevent the interdomain rearrangements necessary for the ATPase cycle to occur. Experimental means of relieving the neck inhibition of regulated myosins include limited proteolytic digestion near the neck (10,19). Biochemical removal of light chains also unregulates molluscan myosin (10) and smooth muscle myosin (42). Our data suggest that the neck domain of BM-V is similar to the other known regulated myosins in that it has a suppressive effect on the head domain. This effect can be relieved by proteolytic cleavage upstream of the head/neck junction. The major effect of calpain cleavage is thus to unregulate BM-V while leaving its enzymatic and actin binding properties relatively unimpaired.
Future Directions-A number of intriguing questions about the mechanochemistry of this unconventional myosin remain unanswered. A basic question is whether Ca 2ϩ stimulates or inhibits BM-V motor activity in the cell. Even though Ca 2ϩ greatly stimulates activity in Mg-ATPase assays, it leads to a reduction in the rate of actin filament movement in motility assays (7). Although the basis for this is behavior is unknown, there are many other examples where ATPase activity and motility assay behavior do not correlate well (43,44). In the case of brain myosin-V, it is possible that absorption to nitrocellulose and other surfaces activates motor activity, perhaps due to interactions between the large neck domain of myosin-V and the surface (discussed in Ref. 11). Another question is why BM-V, which has a maximal ATPase similar to that of skeletal muscle myosin, moves more slowly in motility assays (ϳ0.4 m/s versus ϳ7 m/s; Refs. 7 and 37). A final question is raised by the recent model that the neck domain of the conventional myosins, which consists of the two light chains wrapped around an extended ␣-helix (36,38), might act as a lever to amplify a small conformational change in the head into a much larger power stroke of ϳ10 nm (reviewed in Ref. 32). This model suggests that BM-V, which has a neck domain three times larger than conventional myosins, should exhibit a step size that is also three times larger. On the other hand, extended necks (i.e. more than two IQ motifs) may have other functions besides ATPase regulation and mechanical transduction (for an example, see Ref. 45). Hopefully, the characterization and comparison of myosins from different classes will contribute to a better understanding of how the generic myosin motor works, as well as the specific cellular functions of each class.