Mycalolide B, a Novel Actin Depolymerizing Agent*

We investigated the effects of a novel marine toxin, mycalolide B, on actin polymerization and actin-acti- vated myosin Me-ATPase activity using purified actin and myosin from rabbit skeletal muscle. The results were compared with cytochalasin D which inhibits actin polymerization by binding to the barbed end of F-actin. By monitoring fluorescent intensity of pyrenyl-actin, mycalolide B did not accelerate actin polymerization but quickly depolymerized F-actin, whereas cytochala- sin D accelerated actin nucleation and depolymerized F-actin at slower rate. The kinetics of depolymerization suggest that mycalolide B severs F-actin. The relationship between the concentration of total actin and F-ac-tin at different concentration of mycalolide B suggests that mycalolide B forms 1:l complex with G-actin. Vis-cometry and electron microscopic observation further suggest that actin filament was depolymerized by mycalolide B. Unlike cytochalasin D, furthermore, mycal- olide B suppressed actin-activated myosin Me-ATPase activity. We concluded that mycalolide B severs F-actin and sequesters G-actin and may serve as a novel pharmacological tool for analyzing actin-mediated cell functions.

W e investigated the effects of a novel marine toxin, mycalolide B, on actin polymerization and actin-activated myosin Me-ATPase activity using purified actin and myosin from rabbit skeletal muscle. The results were compared with cytochalasin D which inhibits actin polymerization by binding to the barbed end of F-actin. By monitoring fluorescent intensity of pyrenyl-actin, mycalolide B did not accelerate actin polymerization but quickly depolymerized F-actin, whereas cytochalasin D accelerated actin nucleation and depolymerized F-actin at slower rate. The kinetics of depolymerization suggest that mycalolide B severs F-actin. The relationship between the concentration of total actin and F-actin at different concentration of mycalolide B suggests that mycalolide B forms 1:l complex with G-actin. Viscometry and electron microscopic observation further suggest that actin filament was depolymerized by mycalolide B. Unlike cytochalasin D, furthermore, mycalolide B suppressed actin-activated myosin Me-ATPase activity. W e concluded that mycalolide B severs F-actin and sequesters G-actin and may serve as a novel pharmacological tool for analyzing actin-mediated cell functions.
Mycalolide B was isolated from the marine sponge Mycale sp. from the Bay of Gokasho, Kii Peninsula, Japan (Fusetani et al., 1989). Its chemical structure is notable, possessing trisoxazole in macrolide ring (M, = 1027) (Fusetani et al., 1989). Mycalolide B has antifungal and cytotoxic effects (Fusetani et al., 1989) and also inhibits Mg2"ATPase activity of native actomyosin prepared from chicken gizzard smooth muscle and inhibits smooth muscle contraction (Hori et al., 1993), suggesting that mycalolide B acts directly on either actin or myosin molecules.
Actin is one of the most abundant and common components of the cytoskeleton. Actin also regulates various cell functions, such as muscle contraction, cell motility and cell division. Cytochalasins, a group of fungal metabolites, serve as actin capping compounds which bind to the barbed end of actin filaments and shift the polymerization-depolymerization equilibrium toward net depolymerization of F-actin (Cooper, 1987).
Here we report a novel marine toxin, mycalolide B, that binds to actin in a 1:l molar ratio, and actively, selectively, and completely depolymerizes F-actin to G-actin. As a result mycalolide B, unlike cytochalasins, suppresses actin-activated myosin Mg2"ATPase purified from rabbit skeletal muscle.
from the Ministry of Education, Science, and Culture of Japan. The * This work was supported by a grant-in-aid for Scientific Research costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
$ To whom correspondence should be addressed: Tel.: 81-3-3812-2111 MATERIALS AND METHODS Chemicals-Mycalolide B was isolated from marine sponge Mycale sp. by the method described previously (Fusetani et al., 1989). Cytochalasin D and N-(1-pyrene)iodoacetamide were obtained from Sigma and Molecular Probes (Eugene, OR), respectively. [Y-~~PIATP was purchased from Amersham (Backinghamshire, UK) Preparation of Proteins-Actin was purified from rabbit skeletal muscle (Spudich and Watt, 1971) using buffer G containing 0.2 m CaCl,, 0.2 m~ ATP, 0.5 m P-mercaptoethanol, and 2 m M Tris (pH 8.0 with HCl). Myosin was purified from rabbit skeletal muscle according to the method of Perry (1955). Myosin subfragment-1 61)' was prepared from purified myosin by the method of Weeds and Taylor (1975).
Actin Labeling-Pyrene labeling was performed by the method described by Wendel and Dancker (1986). Briefly, the extracted actin was polymerized with 50 mM KC1 for 1 h at 25 "C and diluted to 48 1.1~. N-(l-pyrene)iodoacetamide was then added at final concentration of 48 p. After pyrene labeling for 30 min at 25 "C, the sample was stored at 0 "C for overnight. The labeled actin was then centrifuged (Hitachi rotor 100AT4) at 48,000 rpm for 3 h, and the pellet was dialyzed for 3 days. Before using the pyrene-labeled actin, the sample was centrifuged again to remove degraded actin.
Measurement of Polymerization-Actin polymerization was started by the addition of 1 m~ MgCl, to buffer G. The time course of polymerization (or depolymerization) was continuously monitored at 25 "C with fluorometer (FP-2060, Japan Spectroscopic) (excitation at 365 nm and emission at 407 nm).
Viscometry-Viscosity of actin solutions was determined with Ostwald-type viscometers with flow time for water of about 30 s. Specific viscosity (qJ is defined as (qrell), where qrel is the flow time for the actin solution divided by the flow time for the corresponding buffer.
with 4% uranyl acetate and observed using a JEOL JEM-1200EX elec-Electron Microscopy-Pyrenyl-F-actin (12 p) was negatively stained tron microscope at an accelerating voltage of 60 kV.

RESULTS
The effect of mycalolide B on actin polymerization was monitored by measuring fluorescence intensity of pyrenyl-actin.
Actin is polymerized in the presence of M e and K'. As demonstrated in Fig. l A , addition of Mg2' polymerized G-actin after a lag phase. Others have suggested that the lag phase is due to formation of actin dimers or trimers (nucleation), and this is the rate-limiting step for actin polymerization (Nishida and Sakai, 1983). Cytochalasin D decreased the lag phase (acceleration of nucleation) (Howard and Lin, 1979), and it inhibited the final level of polymerization. In contrast, mycalolide B completely inhibited polymerization without accelerating nucleation ( The abbreviation used are: S1, myosin subfragment-1; acto-S1 ATPase activity, actin-activated S1 Mg2"ATPase activity. cence under polymerizable condition is normalized with control. Note tion of approximately 70% in this condition, Le. 30% were still in polymerized state. Fig. 2 shows the rates of depolymerization of F-actin elicited by mycalolide B (3 PM) and cytochalasin D (30 p~). The initial rate of depolymerization caused by cytochalasin D ( t , was >13 min) was slower than the depolymerization caused by mycalolide B ( t , was within 1 min). Pipetting the solutions containing F-actin for several times with a capillary tube increased the rate of depolymerization caused by cytochalasin D, due to the severing action on actin filaments. These results suggest that mycalolide B may actively depolymerize F-actin.
To know whether the fast depolymerization elicited by mycalolide B is attributable to the severing activity, we compared the number of F-actin before and after treatment with mycalolide B by the dilution method as described by Maciver et al. (1991). The decay of fluorescence after the dilution of F-actin is considered to be due to the net actin depolymerization from both ends of F-actin. Therefore, the rate of depolymerization should be proportional to the number of F-actin (Maciver et al., 1991). As shown in Fig. 3, when mycalolide B was applied to the F-actin simultaneously with the dilution, the initial depolymerization rate was greatly increased. Furthermore, pretreatment of F-actin with mycalolide B for 30 min before the dilution (fluorescent intensity was decreased by approximately 50% before dilution) also accelerated the subsequent depolymerization by dilution. These results suggest that mycalolide B severs F-actin which results in an increasing its number concentration.
The relationship between the total actin concentration and filamentous actin was examined at different concentrations of mycalolide B (Fig. 4A). The lines shifted in a parallel manner as a function of mycalolide B concentration, suggesting increase of the critical concentration. Similar findings have been described for Acanthamoeba profilin (La1 and Korn, 1985) which sequesters G-actin and inhibits polymerization. Fig. 4A also suggests that mycalolide B binds to G-actin in a 1:l molar ratio, preventing the incorporation of G-actin into filaments.
From these results, Kd value was calculated, by assuming a 1:l complex of actin with mycalolide B, to be 13-20 nM (see legend for Fig. 4). Since mycalolide B showed a similar concentration dependence in its ability to inhibit G-actin polymerization and to cause F-actin depolymerization, we plotted the amount of increase of the critical concentration where the reaction reached a steady state level at a given concentration of mycalolide B (Fig. 4B). This figure also shows that mycalolide B binds to G-actin in a 1:l molar ratio.
To confirm if mycalolide B depolymerizes F-actin, viscosity was measured with a n Ostwald-type viscometer. As shown in Fig. 5, mycalolide B depolymerized F-actin in a similar manner as the result obtained with pyrenyl-actin (Fig. 2). In this case, however, the rate of depolymerization caused by cytochalasin D was relatively higher than that shown in Fig. 2. This may be because passage through the capillary of viscometer showed similar effect to pipetting in Fig. 2 and thus increased the initial rate of depolymerization. Fig. 6 shows electron micrograph of negatively stained F-actin in the absence (Fig. 6 A ) or presence of mycalolide B (Fig.  6B)  Myosin possesses a low level of Mg2"ATPase activity, and this enzymatic activity is enhanced by adding F-actin (Offer et al., 1972) (Fig. 7A). Mycalolide B (30 p~) had no effect on the basal myosin Me-ATPase activity in the absence of F-actin, but it completely inhibited the actin-activated myosin M e -ATPase activity. In contrast, cytochalasin D (30 p~) inhibited neither the basal nor the actin-activated myosin Me-ATPase activity. Fig. 7B shows that mycalolide B inhibited acto-S1 ATPase activity in a concentration-dependent manner, and the maximum inhibition was obtained when its molar ratio to actin was higher than 1. As shown in Fig. 7C, mycalolide B (10 p~; molar ratio to actin = 1.4) decreased viscosity of F-actin in the presence of S1. These results suggest that mycalolide B depolymerizes actin in the presence of myosin and this may be the mechanism of inhibition of actin activation of myosin ATPase activity.

FIG. 4. Effects of mycalolide B on fluorescence intensity at different concentrations of actin (4.4% of the total actin had been labeled with pyrene) (A) and on the increase of the critical concentration of actin ( B ) . In
Finally, we examined the effect of mycalolide B in the presence of phalloidin, which binds to F-actin in a 1:l stoichiometry and inhibits dissociation of G-actin from both ends of F-actin (Sampath and Pollard, 1991). Phalloidin (12 PM) almost completely inhibited the depolymerization induced by 10 V M mycalolide B (Fig. 8). F-actin solution in the presence of S1. Actin (7.1 p) and S1 (3 PM) was incubated in the buffer overnight under 0 "C before measurement.
Mycalolide B (10 w) was applied at the time 0.

DISCUSSION
The present results showed that mycalolide B is the depolymerizing agent, and this mechanism is quite different from that of cytochalasin D. Stossel (1989) classified actin-binding proteins with their mode of action. "Severing" proteins, such as gelsolin or villin, block the barbed ends of actin filaments and promote nucleation. Depactin, actophorin, destrin, and actokinin are categorized as "nibbling" proteins, which sever F-actin and bind to G-actin in stoichiometry with no capping activity nor promotion of nucleation. However, the nibbling proteins with one molecule in the presence of Mp"' (Goddette and Frieden, 1986) Cytochalasin D binds to G-actin and forms the complex of two G-actins (nucleation t ). Mycalolide B binds to F-actin and actively depolymerizes it, forming the G-actin-mycalolide B complex with a 1:l molar ratio (severing or nibbling). When mycalolide B binds to G-actin, this complex never polymerizes even in the presence of Mg2' (monomer sequestration). accelerate the rate of polymerization since they sever F-actin into short filaments and increase the number of nuclei for polymerization (Maciver et d . , 1991;Mabuchi, 1983;Nishida et al., 1984). Profilin, which controls monomer polymerizability (Stossel, 19891, binds only to G-actin and decreases polymerization rate by sequestering G-actin (Pollard and Cooper, 1984).
Present results demonstrated that mycalolide B binds G-actin with 1:l (monomer sequestering activity) (Fig. 4) with no promotion of nucleation (Fig. 1). However, its depolymerizing effect cannot be explained by simple monomer sequestering activity as shown in Fig. 3 that mycalolide B increased filament number. Janmey et al. (1985) showed the different time course of actin depolymerization between actin-severing protein (gelsolin) and G-actin-sequestering protein (vitamin D-binding protein) after dilution (Janmey et al., 1985), and we also conclude that mycalolide B possesses severing activity. In this respect, mycalolide B seems to act on actin like nibbling proteins. Unlike the nibbling proteins, however, mycalolide B does not accelerate the rate of polymerization. The differences in the mode of action between mycalolide B and nibbling proteins may be explained by that mycalolide B would be chelated by G-actin because of the strong monomer binding activity. Alternatively, the severing activity of mycalolide B is weaker than nibbling proteins. On the other hand, it is known that cytochalasin D caps the barbed end of F-actin (Sampath and Pollard, 1991) and also promotes nucleation (Howard and Lin, 1979). Therefore cytochalasin D acts on actin like capping proteins such as p-actinin.
Mycalolide B inhibited actin-activated myosin Mg2"ATPase activity. In contrast, cytochalasin D, even at 30 p~, did not inhibit actin-activated myosin Mg2"ATPase activity (Fig. 8 A ) . The lack of the inhibitory effect of cytochalasin D on actinactivated myosin-ATPase activity may be attributable to the incomplete depolymerization of F-actin by cytochalasin D (Howard and Lin, 1979) (see Fig. 1B). F-actin, which has been shortened by cytochalasin D, may still activate myosin M e -ATPase activity.
The present results further demonstrated that the mycalolide B-induced depolymerization was completely suppressed by phalloidin (Fig. 6). Phalloidin has been suggested to bind F-actin and block the dissociation of actin monomer from both ends of F-actin (Sampath and Pollard, 1991). Stoichiometry of its binding for filaments is one phalloidin to one actin protomer in filaments and the binding causes conformational change of F-actin (Sampath and Pollard, 1991). Further investigations are needed to know whether phalloidin inhibits the binding of mycalolide B or blocks the cleavage of actin filament.
Mycalolide B had no effect on the polymerization of tubulin isolated from swine brain (data not shown), the Mg2"ATPase of skeletal muscle myosin alone (Fig. a), the phosphorylation of smooth muscle myosin regulatory light chain, or cytosolic Ca2+ concentrations in smooth muscles (Hori et al., 19931, suggesting that mycalolide B affects the contractile apparatus in a rather specific manner, i.e. via depolymerization of F-actin. Recently, two natural products were identified and found to act on actin. Patterson et al. (1993) demonstrated that tolytoxin, a macrolide isolated from cyanobacteria with resembled chemical structure with mycalolide B (Ishibashi et al., 1986;Fusetani, 19871, inhibits actin polymerization and disrupts cytoskeletal conformation in L1210 cells. However, the precise mechanism of its action on actin has not yet been revealed. Furukawa et al. (1993) demonstrated that goniodomin A, another macrolide isolated from dinoflagellate, activates (at lower concentration) or inhibits (at higher concentration) actomyosin Me-ATPase activity. Higher concentration of goniodomin A reduces fluorescence of pyrene labeled actin. However, the mode of action of goniodomin A seems to be different from mycalolide B, since goniodomin A does not change the length of F-actin.
In conclusion, (i) mycalolide B binds to G-actin with a 1:l molecular ratio, (ii) depolymerizes actin filaments at rates that exceed the maximal rate of depolymerization achieved by cytochalasin D, and (iii) inhibits polymerization of G-actin. These effects are different from the capping effect of cytochalasin D, which only shifts the polymerization-depolymerization equilibrium to increase G-actin (see Fig. 9). Mycalolide B is a "depolymerizing" agent, and it is a novel and potentially important tool for studies of actin-mediated cell functions.