Interaction of Cytochalasin D with Actin Filaments in the Presence of ADP and ATP*

Transcription

1 THE JOURNAL OF BIOLOGICAL CHEMISTRY Vol. 261, No. 5, Issue of February 15, pp ,1986 Printed in U.S.A. Interaction of Cytochalasin D with Actin Filaments in the Presence of ADP and ATP* (Received for publication, July 16,1985) Marie-France CarlierS, Patrik Criquet, Dominique PantaloniSt, and Edward D. Korn From the Laboratory of Cell Biology, National Heart, Lung, and Blood Institute, National Institutes of Health, Bethesda, Maryland Cytochalasin D strongly inhibits the faster components in the reactions of actin filament depolymerization and elongation in the presence 10 of mm Tris-C1-, ph 7.8, 0.2 mm dithiothreitol, 1 mm MgC12, 0.1 mm CaC12, and 0.2 mm ATP or ADP. Assuming exclusive an and total capping of the barbed end by the drug, the kinetic parameters derived at saturation by cytochalasin D refer to the pointed end and are fold lower than at thebarbedend.inatp,the critical concentration increases with cytochalasin D up to 12- fold its value when both ends are free; as a result of the lowering of the free energy of nucleation by cytochalasin D, short oligomers of F-actin exist just above and below the critical concentration. Cytochalasin D interacts strongly with the barbed ends independently It is well known that the rate and extent of polymerization and the length distribution of actin filaments in nonmuscle cells are regulated by a class of proteins called capping proteins that generally bind to the barbed end (rapidly growing end) of the filaments (for a review see Ref. 1). A family of drugs called cytochalasins has also been described as having properties very similar to those of the physiological capping proteins: the cytochalasins bind to and block elon- gation at the barbed end of actin filaments (2-6) and enhance the nucleation process (7-9). These similarities, as pointed out by Tellam and Frieden (9), emphasize the importance of understanding the mechanism of interaction of the cytochalasins with actin filament ends as a possible model for the * 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 U.S.C. Section 1734 solely to indicate this fact. $ Permanent address: Laboratoire d Enzymologie, Centre National de la Recherche Scientifique, Gif sur Yvette, France. Supported in part by the Centre National de la Recherche Scientifique and the Ligue Nationale Contre le Cancer. The two ends of the actin filament are referred to as barbed and pointed because of the electron microscopic appearance of filaments decorated with heavy meromyosin. action of the physiologically important capping proteins. A more fundamental point that is addressed by the present study concerns the mechanism of polymerization of actin itself and, more specifically, the possibly different dynamics at the two ends of the polymer. Indeed, as shown by Wegner (lo), the fact that hydrolysis of ATP accompanies actin polymerization makes possible the existence of different critical concentrations at the two ends of actin filaments. Blocking the barbed end by different capping proteins in the presence of ATP results in an increase of the critical concentra- tion, indicating that the pointed end has a higher critical concentration than the barbed end (11-13). Recent kinetic studies of the correlation between actin polymerization and ATP hydrolysis (14-16) have led to the conclusion that ATP of the ADP-G-actin concentration (K = 0.5 n ). In is hydrolyzed on F-actin in a reaction subsequent to polymcontrast, the affinity of cytochalasin D decreases co- erization; as a consequence, a cap of ATP-subunits forms operatively with increasing ATP-G-actin concentra- at the ends of the growing actin filament and is maintained tion. These data are equally well accounted for by two at steady state, at least at the barbed end. Additional evidence different models: either cytochalasin D binds very for the existence of an ATP cap is provided by the different poorly to ATP-capped filament ends whose proportion kinetic characteristics of the filament ends depending on increases with actin concentration, or cytochalasin D whether ATP or ADP is bound. This was demonstrated by binds equally well to ATP-ends and ADP-ends and also the nonlinearity of the actin concentration dependence of the binds to actin dimers in ATP but not in ADP. A linear rate of filament elongation in ATP (16-18). In contrast, a actin concentration dependence of the rate of growth linear plot was obtained in ADP (16), and a linear plot would was found at the pointed end, consistent with the vir- also have been obtained in ATP, if ATP hydrolysis were tual absence of an ATP cap at that end. tightly coupled to polymerization.2 Although in the above experiments the sum of events taking place at both filament ends was measured, the data reflect mainly events at the 2041 barbed end because both the association and dissociation rates are much faster at the barbed end than at the pointed end (19). In connection with these recent results, two interesting questions can be addressed with the use of capping agents. 1) What are the kinetic parameters at the pointed end when the barbed end is totally blocked? Specifically, is there an ATP cap at the pointed end which could allow the dynamics of the pointed end to change with the concentration of G-actin, as observed for the barbed end? 2) Do the barbed ends react with capping agents in the same way irrespective of whether ATP or ADP is bound to the terminal actin subunits? In order to answer these questions, we have used cytochalasin D, the cytochalasin species known to interact most strongly with F-actin (4, 7, 20-22), and actin labeled with a fluorescent probe. The critical concentration and the inhibition of both filament elongation and dissociation in ATP and Recall that in the case of tight coupling between these two reactions only ADP-actin is present at the tip of the filament at all concentrations of ATP-G-actin because the rate of ATP hydrolysis is faster than the rate of association of ATP-actin. Therefore, only the association reactions of ATP-actin (kt) and dissociation of ADPactin (k!) are expected to occur in the whole range of ATP-G-actin, leading to a linear plot.

2 2042 Interaction of Cytochalasin D Filaments Actin with ADP have been studied as a function of cytochalasin D concentration. Cytochalasin D behaves as a potent capping agent for the barbed end. Assuming total, and exclusive, capping of the barbed ends, the experiments allow the determination of the rate parameters and the critical concentration for both the barbed and the pointed ends of actin filaments in ATP and ADP. MATERIALS AND METHODS Actin was purified from rabbit muscle according to Spudich and Watt (23) with the modification of Eisenberg and Kielley (24) followed by chromatography of monomeric actin on Sephadex G-200 equilibrated in buffer G (5 mm Tris-C1-, ph 7.8,0.2 mm dithiothreitol), 0.1 mm CaC12, 0.1% NaN3, and 0.2 mm ATP). Only column fractions with a protein concentration of at least % of the actin concentration in the peak fraction were pooled, to avoid all contamination of monomeric actin by short oligomers. Actin was kept on ice in buffer G and used within 2-3 weeks. Actin concentration was determined spectrophotometrically using an extinction coefficient of mg" cm2 (25). Pyrenyl-labeled actin was prepared according to Kouyama and Mihashi (26). ADP-G-actin was prepared, as described previously (16), by polymerization of the 1:l ATP-G-actin complex freed from unbound ATP by Dowex 1 treatment (27,28); exhaustion of all ATP was attained by repeated sonications (usually five times for 10 s each). The ADP-F-actin was then depolymerized by 5-fold dilution in buffer G, which reduced the Mg+ concentration to 0.1 mm, followed by concentration of the actin in an Amicon ultrafiltration cell; ADP-G-actin was isolated by centrifugation at 200,000 X g for 2 h and used within a few hours due to its lability. 10 PM Ap5A3 (29) was included in all buffers in experiments with ADP-actin to inhibit the myokinase activity that contaminates actin preparations (30). Cytochalasin D (Aldrich) was dissolved in (CH&SO. Kinetic and steady-state measurements of actin polymerization were made by monitoring the increase in fluorescence of the pyrenyllabeled actin that accompanies polymerization (26). The polymerization buffer was buffer G supplemented with 1 mm MgC1,. Typically, actin solutions containing 5-10% pyrenyl-labeled actin were used. Fluorescence measurements were made at 25 "C using a Spex fluorimeter equipped with a Datamate, or an SLM 4000 fluorimeter. Excitation and emission wavelengths were 366 and 386 nm, respectively. The rates of elongation or depolymerization of actin filaments were determined as previously described (16). A small volume (5% of the total volume) of a 20 PM F-actin solution was added at time 0 to a solution of G-actin to which cytochalasin D (if needed) and 1 mm MgCl, had been added prior to the addition of seeds. Serial dilutions of cytochalasin D were made so that the same volume of (CH&SO, representing no more than 0.7% of the total volume, was added to each sample. Controls showed that the presence of (CH&SO at this concentration did not affect the observations. The solutions of seeds and G-actin contained the same proportion of labeled actin so that the same specific fluorescence change was associated with the polymerization and depolymerization reactions. The fluorescence changes could be converted to concentration of F-actin subunits by sedimentation of labeled F-actin solutions at steady state and determination of the concentration of G-actin in the supernatant by the method of Lowry et al. (31). The preceding procedure was generally used to measure the rate of filament elongation (J) as a function of G-actin concentration (c) above and below the critical concentration, but it could not be used to obtain the J(c) plot for ATP-G-actin above its critical concentration in the presence of cytochalasin D because of the rapid rate of nucleation at the high concentration of cytochalasin D (500 nm) that was necessary to be sure that all barbed filament ends were blocked. A new method was therefore devised in which the rates of filament elongation were determined as the derivative of spontaneous polymerization curves with respect to time. The J(c) plot in the presence of cytochalasin D was obtained from the slopes of the change in fluorescence at different times in the polymerization process after the number concentration of filaments became constant. In the absence of cytochalasin D, the nucleation process overlapped a large part of The abbreviations used are: Ap5A, P1,P5-di(adenosine-5')-pentaphosphate; c, actin concentration; c,, critical concentration; pyrenyllabeled actin, actin labeled at cysteine-374 by reaction with N- pyrenyliodoacetamide. the elongation process, as previously found (33, 34); in contrast, in the presence of cytochalasin D the steady-state number concentration of filaments was established very early in the polymerization process, that is, at a time when 80-90% of actin was still in the monomeric form, and remained constant thereafter. Therefore, in this latter case, differentiation of the major part of the polymerization curve with respect to time allowed an easy determination of the J(c) plot over a large range of G-actin concentrations above the critical concentration. The positive branch of the J(c) plot (when c > cc), determined in this way, was connected to the negative branch (when c < cc), determined as described above, by equating the number of filaments present in the two solutions. The absolute values for pointed end elongation rates were obtained by comparing this plot to one obtained simultaneously in the absence of cytochalasin D, which gives the elongation rate for both ends (for which absolute values had previously been determined (16, 32)). Measurement of the evolution of the number concentration of filaments during spontaneous polymerization was made using the elongation assay as follows. Aliquots of the reaction solution (which contained pm actin) were removed at different times during the polymerization process and diluted 15-fold in an assay cuvette containing 4 p~ G-actin in polymerization buffer. The initial rate of growth at constant G-actin concentration is proportional to the number concentration of filaments. In order to account for the variable amounts of G-actin (0-0.8 p ~ that ) were transferred with the filaments from the original polymerization solution at the different times of the assay, a calibration curve was made for the rate of growth uersw G-actin concentration in the range needed, i.e. 4-5 p ~, using as seeds a solution of F-actin that had reached steady state under the same conditions. This method enabled comparison of the number of filaments at each time of the polymerization process with the final number of filaments reached at steady state in the same solution. RESULTS Cytochalasin D Has No Effect on the Critical Concentration of Actin in ADP-Since the polymerization of ADP-actin takes place under conditions of micro-reversibility, both ends of the polymer should have the same critical concentration; therefore, no change in critical concentration is expected when one end is blocked. This expectation was tested in two experiments. First, the critical concentration of ADP-actin was directly determined from fluorescence measurements of solutions of F-actin at different concentrations in the presence and absence of 0.5 /IM cytochalasin D. Fig. 1 shows that the two critical concentration plots merged at high actin concentration into a single straight line which extrapolated to a critical concentration of 8.0 f 0.05 pm. This value of the critical concentration of ADP-actin is consistent with the one previously found (30). Fig. 1 also shows that the critical concentration curves in the presence of cytochalasin D deviate from linearity in the region of the critical concentration. We believe that the curvature at 20 h after dilution (Fig. 1, dashed line) cannot be attributed to incomplete depolymerization at the pointed end because 85% of the decrease in fluorescence had occurred within the first 5 h after dilution (Fig. 1, dotted line) and because equilibria obviously had been reached at actin concentrations much higher and much lower than the critical concentration. A similar observation, indicating the presence of polymerized actin at and below the critical concentration, was also made with ATP-actin, as will be discussed in greater detail later. In a second experiment, ADP-G-actin was polymerized under sonication in the presence of different concentrations of cytochalasin D. The time course of polymerization under sonication of ADP-actin has previously been analyzed and mathematically described (17) by a simple model of polymerization of filaments of constant length m, with the number of elongating sites generated by fragmentation increasing in proportion to the mass concentration of polymer formed. In

3 Interaction of Cytochalasin D Filaments Actin with [ADP-actin], pm FIG. 1. Critical concentration of ADP-actin in the presence and absence of cytochalasin D. Serial dilutions of ADP-F-actin (29.4 p ~ 5%, pyrenyl-labeled) at equilibrium in ADP were made into polymerizing buffer, in the absence (0) or presence (0) of 0.5 pm cytochalasin D. The dilutions were made in spectrophotometer cuvettes, and the solutions were sonicated for 30 s at 0, 1, and 2 h to accelerate depolymerization to the new equilibria. The fluorescence intensities were monitored over a 20-h period. The symbols refer to fluorescence readings 20 h after dilution. The final fluorescence was obtained after only 5 h in the absence of cytochalasin D in the presence of cytochalasin D, the fluorescence measured after 5 h is indicated by the dotted line. The light straight line going through the origin is the fluorescence at time 0 after dilution of the F-actin, before any dissociation had occurred. the presence of increasing amounts of cytochalasin D, a decrease in the rate of polymerization is expected as the barbed ends generated by fragmentation are capped. When the concentration of cytochalasin D is high enough to block all the barbed ends, only the pointed ends can grow. The same extent of polymer formation is expected to take place as in the absence of cytochalasin D, but at the slower rate of polymerization characteristic of the pointed end. The data are shown in Fig. 2. The rate of polymerization was partially inhibited by 10 nm cytochalasin D and maximally inhibited by 100 and 250 nm cytochalasin D, indicating that this limit curve is the polymerization curve characteristic of the pointed end alone. The same maximum fluorescence change was observed in the presence or absence of cytochalasin D (not shown for 100 and 200 nm cytochalasin D), confirming that the critical concentration of ADP-actin is the same at the pointed and barbed ends. The effect of cytochalasin D on the polymerization kinetics of ADP-actin was then examined in detail. Theoretical Effects of Cytochalasin D on Actin Polymerization Kinetics-The rate of elongation of actin filaments, V, is the sum of the rates at the barbed end, VB, and at the pointed end, Vp: v = VB + v p (1) In the presence of a barbed end capping agent X, we have: ~Blmtnl = [Bl + [BXI = [B1(1 + [XI/KI) where [B], [BX] and [X] are the concentrations of free and capped barbed ends and free capping agent, respectively, and (2) 0 1 ' 1 ' , 1. [ loo 1m Time, min FIG. 2. Polymerization of ADP-actin under sonication in the absence and presence of cytochalasin D. ADP-G-actin (5% pyrenyl-labeled) at a concentration of 20.7 p~ was polymerized under sonication by adding MgClz to buffer G in the presence of the following concentrations of cytochalasin D: 0,lO nm, 100 nm and 250 nm. The sonicator tip was immersed in the cuvette and set at the lowest power. Sonication was for 0.2 s every 0.8 s, to minimize protein damage, using a time controller attached to the ultrasound generator (18). KI is the equilibrium dissociation constant of X with the barbed ends. Assuming that the capping agent blocks all association and dissociation processes at the barbed end ( VBX = 0), the rate of filament elongation is expected to vary with the concentration of capping agent, as follows: v = VB(1 - X) + v p (3) where is the proportion of capped barbed ends X = [Xl/([Xl + Kr) (4) In the absence of capping agent, Vo = VB + Vp. At infinite concentration of capping agent, V, = V,. Equation (3) can be written: which is the Lineweaver-Burk representation of the binding of X to the barbed ends. Alternatively, a Dixon type representation could be used as follows: Therefore, both the relative values of the kinetic parameters of the barbed and pointed ends and the affinity of the capping agent for the barbed end can be determined from the dependence of the rate of elongation of actin filaments on the concentration of capping agent. Of course, if the capping agent does not totally block the barbed end, i.e. is VBX # 0, the analysis above is still valid except that V, = Vp + VBx. In other words, the values of the rate parameters found in the presence of saturating amounts of capping agent cannot unambiguously be attributed to the pointed end alone unless it is known that the barbed end is capped absolutely. Effect of Cytochalasin D on Actin Polymerization in ADP- Inhibition by cytochalasin D of association and dissociation reactions at the barbed ends of actin filaments was first assayed in the presence of ADP; that is, in the simple case where actin polymerizes reversibly. The dissociation and association rates of ADP-actin were measured by dilution of

4 2044 Interaction of Cytochalasin D with Filaments Actin filaments in polymerization buffer containing ADP-G-actin below and above its critical concentration. In the presence of saturating amounts of cytochalasin D, both rates were inhibited to only about 6-10% of the rates in the absence of cytochalasin D (Fig. 3). These low, limiting rates correspond to the dissociation and association rates at the pointed end, if we assume that cytochalasin D totally blocked the barbed end. The experiment described in Fig. 3 was repeated using several different concentrations of ADP-actin. Fig. 4 summarizes the data plotted according to Equation 5. The data for all actin concentrations fell on the same straight line. Since the total concentration of filament ends in thesexperiments was less than 0.5 nm and the concentration of total cytochalasin D varied between 0.5 and 50 nm, we considered that the concentration of cytochalasin D bound to filaments A I I I I l 0 I w 150 CYTOCHALASIN D CONCENTRATION, nm FIG. 3. Inhibition by cytochalasin D of the association and dissociation reactions of ADP-G-actin at the ends of actin filaments. An aliquot of a solution of F-actin (14 FM, 10% pyrenylactin) at equilibrium in the presence of ADP was diluted 16-fold into polymerization buffer without (0) or with 15.8 p~ ADP-G-actin (A) in the presence of different concentrations of cytochalasin D. The initial rates of depolymerization (0) or polymerization (A) were measured from the change in fluorescence and plotted as per cent of the maximum rate versus cytochalasin D concentration. was negligible compared to the concentration of the unbound ligand. Fig. 4, inset, shows that the KI value was constant at nm at all concentrations of ADP-G-actin. Change in the Critical Concentration for ATP-Actin Polym- erization upon Addition of Cytochalasin D-Fig. 5 shows the critical concentration plots that were obtained at different concentrations of cytochalasin D in ATP. All of the plots are linear at sufficiently high actin concentrations. The critical concentration at a given concentration of cytochalasin D is the intercept of the extrapolation of the linear segment with the linear curve of the fluorescence of G-actin as a function of concentration. In contrast to the constant critical concentration of ADP-actin (Figs. 1 and 2), the critical concentration of ATP-actin increased with cytochalasin D from 0.35 PM in the absence of cytochalasin D to a maximum of PM (Fig. 5, inset). We believe this limit value is the critical concentration of the pointed end which is reached when all barbed ends are capped by cytochalasin D. The curvature in the critical concentration curves in the presence of cytochalasin D, which contrasts with the sharp monomer-polymer transition observed in its absence, was more accentuated at high concentrations of cytochalasin D. The curvature of these plots (which was also seen with ADPactin (Fig. 1)) indicates that some actin species with fluores- cence higher than that of G-actin was present below the critical concentration when the ratio of cytochalasin D to total actin concentration was higher than %o. This observation is consistent with Oosawa's (35) theoretical prediction that equilibrium oligomers should exist below the critical concentration. The concentration, ch, of monomers incorporated into these helical oligomers is: where c1 is the monomer concentration, K is the elongation association constant, and u is the unitless constant associated with the free energy change, 6, for nucleation, as defined by Oosawa (35). u = exp (- $) I 1 I I I I FIG. 4. Binding of cytochalasin D to ADP-actin filament ends as a function of the ADP-actin concentration. The experiments described in Fig. 3 were repeated at several concentrations of ADP-G-actin. The data are plotted according to Equation 5. Open and closed symbols refer to experiments done the same day at the following ADP- G-actin concentrations in pm: 0.5, A, 1.9, 0, 15.8,0, 17.5, A. Inset, the equilibrium dissociation constant for the interaction of cytochalasin D and filament ends as a function of the concentration of ADP- G-actin. 5 b 4- z 2 8 > t3 3- L 6- - f 5- s 1- f I I I I llCytochalasin Dl, nm-1

5 I [ATP-Acthi, pm FIG. 5. Critical concentration of ATP-actin in the presence of different concentrations of cytochalasin D. Actin (10% pyrenyl-labeled), 20 p~ in buffer G, was polymerized by addition of 1 mm MgClz in the presence of different concentrations of cytochalasin D. After all samples reached steady state, serial dilutions of each sample were made into polymerization buffer containing the same concentration of cytochalasin D as in the original sample. The dilutions were made in spectrophotometer cuvettes that were briskly inverted to break filaments in order to accelerate depolymerization to the new steady states. The fluorescence intensities were measured after 4 h. The fluorescence levels were the same after 12 h when the concentration of cytochalasin D was 0.5 pm or less. Above 0.5 pm cytochalasin D, the fluorescence readings were slighlty lower after 12 h; this was attributed to a significant increase in the ADP/ATP ratio in solution, due to the high ATPase activity of actin in the presence of cytochalasin D (7), which causes a shift of the critical concentration toward the higher critical concentration of ADP-actin. The critical concentrations were estimated from the intercepts of the linear dependence of fluorescence on the concentrations of G-actin (dashed line) and F-actin (solid lines), in the presence of the following con- Interaction of Cytochalasin D with Actin Filaments 2045 I A, 0.005; v, 0.010; *,., centrations of cytochalasin D (in pm): 0, 0.025; 0, 0.05; m, 0.1; A, 0.25; V, 0.5; 0, 1.0; 0, 1.5. Inset shows the evolution of the critical concentration of ATP-actin with the concentration of cytochalasin D. Where k is the Boltzman constant and T the absolute temperature. As recently demonstrated by Newman et al. (36), the mass concentration, ch, of oligomers is negligible below the critical concentration only in the case where the free energy is large (a small). However, when the free energy of nucleation is lowered, i.e. when nucleation is facilitated, a increases and an appreciable concentration of polymerized subunits, ch, can be measured below the critical concentration. We calculated the value of the parameter u at different cytochalasin D concentrations using the following equation given by Newman et al. (36): KC^)^ - (2 + KC~)(KC,)~ + (1 + 2Kc0 - Kco = 0 (9) where K is the elongation association constant (reciprocal of the critical concentration), co the total actin concentration, and c1 the concentration of actin monomer. Assuming the steady-state actin fluorescence to be linearly related to the concentration of polymerized subunits, the amount of polymer and monomer present at the critical concentration (where co = 1/K) at different concentrations of cytochalasin D was estimated from Fig. 5, and Equation 9 was solved for u. Table I shows that u increases with cytochalasin D in agreement with the theory. The average number of subunits (i) of these oligomers present at and below the critical concentration can also be calculated, according to Oosawa (35), as: ic. (i) = 2 = 2 (provided c1 <1/K) (10) ~i 1 - KC, K (the reciprocal of the critical concentration) is known, c1 can be calculated.as the difference between the total actin concentration, co, and the measured concentration of polymerized subunits, ch, and, therefore, (i) can be calculated from Equation 10. From the data in Fig. 5, an average length of 5-6 subunits/oligomer at the extrapolated critical concentration (co = 1/K) was calculated. Effect of Cytochalasin D on the Rate of Filament Elongation and Depolymerization in ATP-Inhibition of filament depolymerization and growth in ATP was next assayed. As in ADP, cytochalasin D inhibited the depolymerization process upon extensive dilution, with a ratio of dissociation rate constants in the presence and absence of cytochalasin D of 10 to 15 in several independent experiments (data not shown). The data plotted according to Equation 6 yielded a linear Dixon plot with an equilibrium dissociation constant of 2 k 0.3 nm (Fig. 6, upper-most curve). Inhibition by cytochalasin D of the rate of filament elongation was also observed in the presence of ATP. As expected from the data in Fig. 5, depolymerization of filaments was observed in the presence of saturating amounts of cytochalasin D (V- < 0) at concentrations of ATP-G-actin between 0.5 and 4 PM, while a slow rate of polymerization (V, > 0) was observed at concentrations of ATP-G-actin above 4 PM (the critical concentration of the pointed end in ATP). Because cytochalasin D greatly enhances the rate of nucleation of ATP-actin (6-8), polymerization time courses in the pres- ence of cytochalasin D must be corrected for the rate of spontaneous polymerization observed in the absence of added filaments. The amounts of filaments added as seeds were adjusted so that the correction for spontaneous polymerization represented less than 20% of the rate measured in the presence of seeds at the highest concentration of cytochalasin D. The results are shown in Fig. 6. The striking observation is that while the inhibition data, plotted according to Equation 6, gave linear plots at all ATP-G-actin concentrations, the equilibrium dissociation constant of cytochalasin D increased TABLE I Decrease in the free energy of nucleation of actin in the presence of cytochalasin D The concentrations of polymerized subunits at the critical concentration in the presence of different concentrations of cytochalasin D were derived from the data in Fig. 5, and the value of the constant c was calculated at each concentration of cytochalasin D using Equation 9. Cytochalasin D U PM 0 <<

6 2046 Interaction of Cytochalasin D Filaments Actin with filaments depolymerize and J(c) < 0; above the critical concentration of the pointed end, the filaments grow, i.e. J(c) > 0. In the simple case of a reversible polymerization, according to Oosawa s model, J(c) is a straight line described by the following equation: I I I I [Cytochalasin Dl, nm FIG. 6. Inhibition by cytochalasin D of the association and dissociation reactions of G-actin at the ends of the filaments in the presence of ATP. The experiments were conducted as described., in Fig. 4, using G-actin and F-actin solutions prepared in the presence of ATP. The data are plotted according to Equation 6. The following concentrations of G-actin (in FM) were used in the assays, from left to right: 0, 0.02; 0, 2; 0, 3; 0, 4; A, 4.5; A, 5.1; U, 6.5; 8. The filament seeds depolymerized when diluted into buffer to give 0.02 PM ATP-G-actin so the rate of depolymerization was measured (0); in all of the other experiments the rate of elongation was measured. To derive the straight line shown for the experiment with 8 NM ATP-G-actin, points other than those that could be shown on this scale were used. Inset, evolution of the equilibrium dissociation constant of cytochalasin D with ATP-G-actin concentration. with the concentration of ATP-G-actin in a cooperative fashion (Fig. 6, inset). For instance, 50% inhibition of the rate of growth of filaments required 4 and 20 nm cytochalasin D in the presence of 2 and 6 pm G-actin, respectively. This last result is at variance with the constant value of KI obtained in the presence of ADP (Fig. 4). An identical plot of Kr uersus actin concentration was obtained when the experiments were done using Mg-actin, i.e. G-actin that had been incubated for 3 min with 0.2 mm EGTA and 50 ~ LMgClz before addition of 0.95 mmmgc12 and F-actin seeds. Possible models to account for these results are proposed in the discussion. G-Actin Concentration Dependence of the Rate of Filament Elongation at the Pointed End in the Presence of ATP-In order to get a better understanding of the dynamics of the pointed end of actin filaments, the information described in the preceding section was used to determine the evolution of the rate of growth with G-actin concentration, a J(c) plot, under conditions where the barbed end was capped by cytochalasin D. The plot of the rate of growth uersw G-actin concentration provides a direct measure of the critical concentration, the actin concentration where J(c) = 0. Below the critical concentration of the pointed end, barbed-end capped J(c) = k+[f]c, - k[f] (11) where [F] is the filament concentration and k+ and k- are the association and dissociation rate constants. The values of k, and k- can be calculated from the slope of the line and its intercept at zero actin. The data in Fig. 6 show that 0.5 p~ cytochalasin D is sufficient to inhibit at least 90% of the growth at the barbed end up to 9 p~ ATP-G-actin. Because of the high rate of spontaneous nucleation observed above 5-7 p~ actin, however, a large correction is necessary when the experiment is done the usual way, i.e. by fold dilution of a solution of 20 pm F-actin at steady state into solutions of G-actin at given concentrations and measuring the initial rate of fluorescence change. Therefore, this method was used only to obtain the negative branch of the J(c) plot; that is at G-actin concentrations lower than 5 p ~ To. obtain data at higher concentration of G-actin, a new method (described under Material and Methods ) was designed, making use of the ability of cytochalasin D to accelerate nucleation. The results are shown in Fig. 7. From the intercept with the line J(c) = 0, a value of 4 p~ is found for the critical concentration of the pointed end, in agreement with the steady-state data (Fig. 5). Most interestingly, the J(c) plot does not show a downward bend in the region of the critical concentration, similar to the one previ- ously observed for the barbed end (16). In contrast, in the presence of cytochalasin D, the plot is almost linear between c = 0 and c = 5 p ~ however, ; the slope of the plot increases above the critical concentration, giving an upward curvature to the positive branch. The method used to derive this plot eliminates the possibility that this curvature might be due to an increase in the number of filaments at increasing G-actin concentration. However, because growth at the barbed end is much faster than at the pointed end, a small percentage of uncapped barbed ends would make a significant contribution to the observed rate of elongation in the presence of 0.5 ~ L M cytochalasin D. Since the effectiveness of the capping of barbed ends by cytochalasin D decreases when the concentration of G-actin increases (Fig. 6) so does the percentage of capped barbed ends. Therefore, the data of Fig. 6, inset, were used to calculate the proportion, iz, of capped barbed ends at different G-actin concentrations according to Equation 4, using the value of the apparent KI found for cytochalasin D at each G-actin concentration. x was found to vary from 0.98 at 5 pm G-actin to 0.91 at 8.5 p~ G-actin. The contribution to the barbed end in the elongation in the presence of cytochalasin D was then calculated at different G-actin concentrations, according to the equation: v, = (1 - X)(q - C!)ks, (12) where c1 is the concentration of G-actin, and and k: are the critical concentration and association rate constant, respectively, at the barbed end ($ = 0.3 pm; k: = 1.4 p s- ). The evolution of VB due to the remaining uncapped barbed ends with ATP-G-actin concentration had the same upward curvature as the original measured J(c) plot, and the contribution of VB represented a large proportion of the overall measured rate (Fig. 7, dashed line). Once this correction was made, the upper branch of the J(c) plot at the pointed end

7 Interaction of Cytochalasin D Filaments Actin with 2047 FIG. 7. ATP-G-actin concentration dependence of the rate of elongation in the presence of 0.5 p~ cytochalasin D. At concentrations of ATP-G-actin lower than 5 FM, measurements of the initial rates of depolymerization or elongation of F-actin seeds were made as described previously (16). At concentrations of ATP-G-actin above 5 p ~ the, rates of elongation were determined as the derivatives with respect to time of the spontaneous polymerization curves as described under Materials and Methods. All solutions contained 0.5 PM cytochalasin D. Different symbols (B, A, 0, 0) refer to four different experiments. The dashed line (X) represents the calculated contribution of uncapped barbed ends to the measured elongation rates (see text). The straight light line is the difference between the experimental curve and the calculated curve and represents the calculated elongation rates at the pointed end alone. was consistent with a linear extension of the linear lower branch. DISCUSSION The data presented in this paper show that cytochalasin D is a potent inhibitor of the faster component of the reactions of filament elongation and depolymerization. In view of previous electron microscopy observations (6), the most likely interpretation of the present results is that the barbed end is efficiently blocked by cytochalasin D, and the kinetic data obtained in its presence are those of the pointed end. However, as mentioned previously, if cytochalasin D at saturation only partially inhibits barbed end growth or if cytochalasin D also binds to pointed ends and partially inhibits their growth at saturation, as suggested by a recent observation (37), the kinetic parameters found at saturating concentration of cytochalasin D would not be the true values for the pointed end in the absence of the drug. We think this is not very likely since the critical concentration of ADP-actin is unchanged in the presence of cytochalasin D, which would then require that the ratio of the dissociation and association rate constants of the partially inhibited ends be the same as the ratio of the rate constants in the absence of cytochalasin D. We have, therefore, assumed specific capping of the barbed ends by cytochalasin D and total inhibition of elongation and depolymerization at the barbed ends at saturating concentrations of cytochalasin D. This assumption is further justified by the fact that the rate constants and critical concentration found for the pointed end in these experiments are essentially the same as those found using plasma gelsolin to block the barbed end (13). The observation that oligomers exist at and below the critical concentration in the presence of cytochalasin D is not an exception to the theory of nucleated helical polymerization. According to theory, such oligomers exist even in the absence of a capping agent, but their number concentration is so small under the usual conditions (Cis = lo- M) that the amount of polymerized subunits is undetectable by the methods commonly used to monitor polymerization. The presence of a capping agent results in a large increase in the number concentration of polymer (reflected by the increase in the u constant) which amplifies the phenomenon and creates a [ATP-G-ActinL FM measurable amount of oligomers. At actin concentrations far above the critical concentration, the polymers grow indefinitely until the concentration of monomers decreases to the critical concentration. Therefore, at steady state the mass concentration of polymer is much larger than the concentration of monomer, the length distribution is shifted toward very long filaments, and oligomers are no longer present. Even in the presence of cytochalasin D, then, there is no deviation from the classical monomer-polymer system when the total actin concentration is much larger than the critical concentration. In the elongation assays in the presence of cytochalasin D, the solutions of filaments used as seeds would have consisted of long polymers and monomeric actin only. Therefore, the measurements of initial rates of growth refer to the effect of cytochalasin D on the addition or loss of monomers to and from the ends of these filaments. Of course, when the total actin concentration is in the region of the critical concentration, the length distribution of filaments will evolve toward a final state where short oligomers will be present. Our measurements of the initial rates of filament growth, however, would have been completed before, and therefore be unaffected by, any change in filament length distribution. The results obtained for ADP-actin are clear and indicate that the rate constants for both ADP-actin association to (k,) and dissociation from (k-) filaments are fold higher at the barbed end than at the pointed end. With this information, numerical values for the kinetic parameters of ADPactin at the barbed and pointed ends could be derived from the previously measured values for the sums of the values at both ends (16, 32) (Table 11). As expected for an equilibrium polymer, for which the critical concentration should be the same at both ends of the filaments, the critical concentration of ADP-actin, c, = 8 pm, was unchanged by the addition of cytochalasin D. In contrast to the results for, ADP-actin, hydrolysis of ATP allows the two ends to have different critical concentrations when ATP-actin is polymerized (10). In fact, in the presence of saturating amounts of cytochalasin D, there was a 12-fold increase in the critical concentration to a value of p~ which most likely represents the critical concentration of ATP-actin for the pointed ends. A high concentration of cytochalasin D is necessary to reach the

8 2048 Interaction of Cytochalasin D with Actin Filaments TABLE I1 Kineticparameters of the barbed (B) andpointed ends (P) of actin filaments in the presence of ATP and ADP Polymerization conditions were: 10 mm Tris Cl-, ph 7.8, 0.2 mm dithiothreitol, 0.01% NaN3, 0.1 mm CaCL, 1 mm MgCL, and either 0.2 mm ADP or 0.2 mm ADP. k+ k- B P B P wm" s-' S-1 ADP-actin f f 0.1 ATP-actin f " "According to our model (38), the rate of dissociation of ATPactin from ATP-capped filaments is dependent on the number (1, 2, or 23) of terminal ATP subunits. The highest value is that calculated for filaments with a long stretch of ATP subunits and the lowest value for a short cap of perhaps only 2 ATP subunits. The pointed end will also have an ATP cap at a sufficiently high concentration of ATP-G-actin, but those data have not yet been obtained because of the experimental difficulties encountered using cytochalasin D at high concentrations of ATP-G-actin (Fig. 7 and text). critical concentration of the pointed end because, as explained by Walsh et al. (12), the presence of a very small percentage of uncapped barbed ends has a very large effect due to the much faster association-dissociation reactions at the barbed end. Direct information on the dynamics of the pointed end in ATP was provided by the ATP-G-actin concentration dependence of the rate of elongation (Fig. 7); the rate constants are summarized in Table 11. Furthermore, the linear plot obtained in Fig. 7, after correction for the contribution of the few per cent of residual uncapped barbed ends, indicates either that there is virtually no ATP cap at the pointed end over the entire range of actin concentration studied or, if there is an ATP cap, that the kinetic parameters of the terminal subunit at the pointed end are very similar whether this subunit has bound ATP or ADP.We consider it morelikely that, in contrast to the barbed end, there is no ATP cap at the pointed end and that, when both ends are free, the pointed end has a terminal ADP-subunit up to approximately 8 pm ATP-Gactin, and the barbed end has a terminal ATP-subunit at ATP-G-actin concentrations equal to or greater than 0.35 phi. At some ATP-G-actin concentration, however, an ATP cap must also exist at the pointed end because the critical concentration found previously (18) for the ATP-F-actin equilibrium polymer must be the same at both ends. Because of the lack of a high affinity pointed-end capper, it is difficult to obtain kinetic data for the barbed end alone. However, the J(c) plot for the barbed end can be calculated by subtracting the J(c) plot for the pointed end (Fig. 7) from the J(c) plot previously obtained when both ends were free (16). This is shown in Fig. 8. Because the rate of elongation at the barbed end varies with the ATP-actin concentration, the plot is not strictly linear and obeys a cubic equation with respect to c (38). Therefore, this graphic difference method is the simplest way to obtain the critical concentration of ATPactin at the barbed end. At the intercept of the two experimental curves, the rate of elongation of both ends together is equal to the rate of elongation of the pointed end alone, so this defines the critical concentration of the barbed end (actin concentration for zero rate of elongation of the barbed end). The determined value, 0.3 p ~ is, only 20% less than the critical concentration for both ends together (0.35 pm). Comparison of the kinetics of the inhibition of elongation by cytochalasin D in ATP and ADP may provide additional interesting information on the properties of the barbed end. Cytochalasin D interacts with the filament ends in ADP with I I I [ATP-G-actin], pm FIG. 8. Derivation of the ATP-G-actin dependence of the rate of elongation at the barbed end. The data obtained for the pointed end alone (Fig. 7) are shown in open symbols and are subtracted from the data previously obtained (16) when both ends were free (closed symbols). The difference curve (dashed line) is the calculated rate of elongation at the barbed end alone. a very high affinity, 0.5 k 0.1 n"', independent of the G- actin concentration, as measured by the inhibition of elongation and dissociation (Fig. 4, inset). The inhibition pattern (Fig. 4) is indicative of a single class of binding of sites. The same high affinity of 0.5 nm-' was also found in ATP for the inhibition by cytochalasin D of depolymerization upon dilu- tion (Fig. 6, inset, lowest actin concentration), conditions under which ADP-actin is known to dissociate from filaments and essentially only ADP subunits are present at the tip of the barbed end (16). However, the affinity of cytochalasin D for filaments in ATP, as measured by inhibition of elongation, appeared to decrease cooperatively with the concentration of ATP-G-actin (Fig. 6, inset). Our results, in this regard, corroborate and make more quantitative the initial report by Brown and Spudich (4) who observed that cytochalasin D is a more potent inhibitor of filament elongation in ADP than in ATP. We have considered two possible explanations for these results. (i) The first possibility is that cytochalasin D binds very poorly to an ATP end and, as predicted by our model for actin polymerization in the presence of ATP (18,38)', the probability for an ATP subunit to be present at the barbed-end of the filament increases strongly with the concentration of ATP- G-actin. The model proposed (18, 38) assumes that a steady state exists between different species of filaments, F, which differ by the number, i, of terminal ATP subunits. Fo filaments consist of ADP subunits only with no ATP-terminal subunit. The respective proportions fi of these filaments change with the ATP-G-actin concentration, while their sum is constant; that is: f0+fi+f2+f3+...= 1. (13) The analytical formulation of the model (38) showed that: fo = (I + K ~c + [KlKzc2/(1 - K'c)])-' (14) where Kl and Kz and K' are equilibrium association constants of ATP-G-actin with Fo, F,, and Fi ends, respectively.

9 Interaction of Cytochalasin D withactin Filaments 2049 We have already shown that cytochalasin D binds strongly to ADP-ends (Fo). If we assume that it does not bind at all to ATP-ends, i.e. to Fi(i > 0), the apparent equilibrium inhibition constant is proportional to the reciprocal of the proportion of Fo filaments: Because K2 > K,, at actin concentration c < 10 pm (Equation 14), reduces to: fo 1 + KlKzcz (15) elongation at the barbed end is: VB = k+ifl(1 "X)(IAl - c?), ([AI = cd (20) where x is the proportion of capped barbed ends. The equations of mass conservation are: [Xltot = [XI + [FXI + [&XI (21) [Altot = [A] + ~[Az] + ~[AzX] (22) Since [FX] << [XItot and [A21 + [A,XJ << [A],,, then [A] = [A hot, and and &app E KA1 + K IK~) (16) where KI is the inhibition constant for Fo ends alone (fo = 1). The data in Fig. 6, inset, give the predicted straight line when the observed KI (which is KIapp in Equation 16) is plotted versus the square of the ATP-G-actin concentration (Fig. 9). Therefore, this model can explain the results. (ii) The other possible explanation for the increase of Kl with ATP-actin concentration would be that cytochalasin D, X, binds not only to barbed ends but also to a dimeric species. A2, of ATP-actin. In this case, a part of the cytochalasin D would be sequestered in the A2X complex, and the concentration of A2X would increase directly with the square of the ATP-actin concentration. The equations describing the system are: where KI, KO, and K2 are the equilibrium dissociation constant for the binding of cytochalasin D to filaments, for dimerization of ATP-G-actin, and for binding of cytochalasin D to the actin dimer, respectively. The rate of filament [ATP-ACTIN]?,MI2 FIG. 9. Actin concentration dependence of the apparent inhibition constant, for the inhibition ofilament elongation by cytochalasin D in the presence of ATP. The data shown in Fig. 6, inset, are plotted uers'sus the square of the concentration of G-actin with the assumption that the observed K1 is KIspp according to the equation (16). Equation 23 shows that, at a given ATP-actin concentration, the concentrations of free and total cytochalasin D are kept within a constant ratio. Therefore, the Lineweaver-Burk and Dixon plots of the inhibition of filament elongation by cytochalasin D would still be linear in this case. However, the apparent KI that is measured would be the sum of the concentrations of free and dimer-bound cytochalasin D at the point where 50% of the barbed ends are capped; that is: Equation 24 shows that within this scheme, too, KIapp varies linearly with the square of the ATP-G-actin concentration. The slope of the plot in Fig. 8 would give a value of 4 p~~ for KDK2. The concentration of cytochalasin D bound-actin dimer, A2X, would be 12.5 nm at 5 p~ G-actin, and 50 nm at 10 p~ G-actin. Assuming that cytochalasin D binds to the dimer with the same affinity as to the filament ends (Kz = 2 nm), then the value of the dimerization constant KO would be 2 mm. While this second model accounts for our data as well as the first one, it raises two puzzling questions. First, if the A2X species were present in solution, it would be expected to act as a pointed end nucleus for filament growth. Considering the high concentration of these nuclei as compared to the concentration of added filaments (10-50 verszs 0.5 nm) in the assay, the rate of polymerization in the absence of seeds should have been much larger than was observed. In fact, the addition of cytochalasin D should have resulted in an apparent increase in the rate of elongation of filaments. Therefore, if the A2X species exists, it must not behave as a true pointed end nucleus. This peculiarity would make the interaction of cytochalasin D with actin similar to that of gelsolin (13). The second puzzling point raised by the second model is that the A2X species could not be formed in the presence of ADP, even though cytochalasin D binds very strongly to barbed ends of ADP filaments and to short oligomers of ADPactin because KI does not change with ADP-actin concentration (Fig. 4). This contradiction remains unresolved. As dis- cussed in detail at the beginning of this section, the existence of short cytochalasin-d-capped actin oligomers when the total actin concentration is close to the critical concentration does not mean that cytochalasin D binds to monomeric or dimeric actin. Further investigation is, therefore, needed to answer the above questions and to discriminate between the two models. Acknowledgment-We thank Victoria Putprush for her careful preparation of the manuscript.

10 2050 Interaction of Cytochalasin D Filaments Actin with REFERENCES Korn, E. D. (1982) Physiol. Rev. 62, Brenner, S. L., and Korn, E. D. (1979) J. Biol. Chem. 254, Flanagan, M. D., and Lin, S. (1980) J. Biol. Chem. 255, Brown, S. S., and Spudich, J. A. (1979) J. Cell Biol. 83, Lin, D. C., Tobin, K. D., Grumet, M., and Lin, S. (1980) J. Cell Biol. 84, McLean-Fletcher, S., and Pollard, T. D. (1980) Cell 20, Brenner, S. L., and Korn, E. D. (1980) J. Bio2. Chem. 255, Selden, L. A., Gershman, L. C., and Estes, J. E. (1980) Biochem. Biophys. Res. Commun. 95, Tellam, R., and Frieden, C. (1982) Biochemistry 21, Wegner, A. (1976) J. Mol. Biol. 108, Wegner, A., and Isenberg, G. (1983) Proc. Natl. Acad. Sci. U. S. A. 80, Walsh, T. P., Weber, A., Higgins, J., Bonder, E. M., and Mooseker, M. S. (1984) Biochemistry 23, Cou6, M., and Korn, E. D. (1985) J. Bid. Chem. 260, Pardee, J., and Spudich, J. A. (1982) J. Cell Biol. 98, Pollard, T. D., and Weeds, A. G. (1984) FEBS Lett. 170, Carlier, M.-F., Pantaloni, D., and Korn, E. D. (1984) J. Riol. Chem. 259, Carlier, M.-F., Pantaloni, D., and Korn, E. D. (1985) J. Biol. Chem. 260, Pantaloni, D., Carlier; M.-F., and Korn, E. D. (1985) J. Biol. Chen. 260, Pollard, T. D., and Mooseker, M. S. (1981) J. Cell Biol. 88, Carter, S. B. (1967) Nature 213, Lin, S., Lin, D. C., and Flanagan, M. D. (1978) Proc. Natl. Acad. SC~. U. S. A. 76, Miranda, A. F., Godman, G. C., Deitch, A. D., and Tanenbaum, S. W. (1974) J. Cell Biol. 61, Spudich, J. A., and Watt, S. (1971) J. Biol. Chem. 246, Eisenbere. E.. and Kiellev. ", W. W. (1974) ~, J. Bwl. Chem fi48 ' 25. Gordon, D. J.. Yaw, Y.-2.. and Korn. E. D. (1976)., J. Biol. Chem. 251,'7474-?479" 26. Kouyama, T., and Mihashi, K. (1981) Eur. J. Bwchem. 114, Strzelecka-Golaszewska, H., Nagy, B., and Gergely, J. (1974) Arch. Biochem. Biophys. 161, Mockrin, S. C., and Korn, E. D. (1980) Biochemistry 19, Lienhard, G. E., and Secemski, I. I. (1983) J. Biol. Chem. 248, Pantaloni, D., Carlier, M.-E., Coub, M., Lal, A. A., Brenner, S. L., and Korn, E. D. (1984) J. Biol. Chem. 259, Lowry, 0. H., Rosebrough, N. J., Farr, A. L., and Randall, R. J. (1951) J. Biol. Chem. 193, Lal, A. A., Korn, E. D., and Brenner, S. L. (1984) J. Biol. Chem. 259, Tobacman, L. S., and Korn, E. D. (1983) J. Biol. Chem. 258, Cooper, J. A., Buhle, E. C., Walker, S. B., Tsong, T. Y., and Pollard, T. D. (1983) Biochemistry 22, Oosawa, F. (1983) Muscle and Nonmuscle Motility (Stracher, A., ed) Vol. I, p , Academic Press, New York 36. Newman, J., Estes, J. E., Selden, L. A., and Gershman, L. C. (1985) Biochemistrv Maruyama, K., Yamada, N., and Mabuchi, I. (1984) J. Biochem. (Tokyo) 96, Pantaloni. D.. Hill. T. L.. Carlier, M.-F.. and Korn, E. D. (1985) Proc.'Natl. Acad.'Sci. U. S. A. 82,