Xenopus Actin-interacting Protein 1 (XAip1) Enhances Cofilin Fragmentation of Filaments by Capping Filament Ends*

Transcription

1 THE JOURNAL OF BIOLOGICAL CHEMISTRY Vol. 277, No. 45, Issue of November 8, pp , by The American Society for Biochemistry and Molecular Biology, Inc. Printed in U.S.A. Xenopus Actin-interacting Protein 1 (XAip1) Enhances Cofilin Fragmentation of Filaments by Capping Filament Ends* Received for publication, April 1, 2002, and in revised form, June 2, 2002 Published, JBC Papers in Press, June 7, 2002, DOI /jbc.M Kyoko Okada, Laurent Blanchoin, Hiroshi Abe **, Hui Chen, Thomas D. Pollard, and James R. Bamburg From the Department of Biochemistry and Molecular Biology, Colorado State University, Fort Collins, Colorado 80523, the Department of Biology, Faculty of Science, Chiba University, Chiba , Japan, the Structural Biology Laboratory, The Salk Institute for Biological Studies, La Jolla, California 92037, and the **Department of Cell Biology, Natural Institute for Basic Biology, Okazaki 444, Japan Xenopus actin-interacting protein 1 (XAip1) is thought to promote fragmentation of actin filaments by cofilin. To examine the mechanism of XAip1, we measured polymer lengths by fluorescence microscopy and the concentration of filament ends with an elongation assay. Cofilin creates ends by severing actin filaments. XAip1 alone does not sever actin filaments or prevent annealing/ redistribution of mechanically severed filaments and has no effect on the concentration of ends available for subunit addition. In the presence of XAip1, the apparent filament fragmentation by cofilin is enhanced, but XAip1 reduces rather than increases the concentration of ends capable of adding subunits. Electron microscopy with gold-labeled antibodies showed that a low concentration of XAip1 bound preferentially to one end of the filament. A high concentration of XAip1 bound along the length of the filament. In the presence of gelsolin-actin to cap filament barbed ends, XAip1 does not enhance cofilin activity. We conclude that XAip1 caps the barbed end of filaments severed by cofilin. This capping blocks annealing and depolymerization and allows more extensive severing by cofilin. Actin dynamics underlie various cellular activities including cell motility, cytokinesis, endocytosis, and translocation of some intracellular organelles. Spatial and temporal regulation of actin assembly and disassembly as well as three-dimensional filament organization are required for these phenomena. Numerous actin-binding proteins coordinate these processes. Analysis of their mechanisms is required to understand complex processes such as actin assembly and filament turnover * This work was supported by grants from the National Institutes of Health (GM35126 and GM54004 to J. R. B. and GM26338 to T. D. P.), the Japanese Ministry of Science (to H. A.), and a Japanese Society for Promotion of Science (JSPS) Research Fellowship for Young Scientists (to K. O.). 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. Present address: Laboratoire de Physiologie Cellulaire Végétale, Batiment C2, piece 435 Département de Réponse et Dynamique Cellulaires, CEA Grenoble, 17 rue des Martyrs, Grenoble cedex 9, France. Present address: Dept. of Biological Chemistry, The Johns Hopkins University School of Medicine, Baltimore, MD Present address: Dept. of Molecular, Cellular, and Developmental Biology, Yale University, New Haven, CT To whom correspondence should be addressed: Dept. of Biochemistry and Molecular Biology, Colorado State University, Fort Collins, CO Tel.: ; Fax: ; jbamburg@lamar.colostate.edu. This paper is available on line at during extension of lamellipodia at the leading edge of migrating cells and the movement of intracellular bacterial pathogens (reviewed in Refs. 1 and 2). In vitro reconstitution of the actinbased motility of pathogenic bacteria demonstrated that five essential factors were required for sustained motility: actin, Arp2/3 complex activated by a bacterial surface protein or N-WASP, barbed end heterodimeric capping protein, and an ADF 1 /cofilin protein (3). Motility was enhanced by adding profilin, -actinin, and VASP. Detailed analyses demonstrated how interactions of these proteins affect the magnitude of the oriented polymerization and the stability of the actin filament network (4, 5). Proteins of the ADF/cofilin family are ubiquitously expressed and highly conserved in eukaryotes (reviewed in Refs. 6 and 7). ADF/cofilins accelerate actin filament turnover either by severing (8, 9) and/or promoting subunit dissociation from filament ends (10). ADF/cofilins also enhance P i release from ADP-P i actin filaments, which leads to debranching of actin filament networks (5). Severing activity of ADF/cofilin proteins also contributes to the formation of new actin filament barbed ends in lamellipodia (11, 12). Although they have little effect on the concentration of polymerized actin, the combination of profilin and cofilin increases the rate of subunit flux from polymer to monomer pool and back to filaments (8, 13). Phosphorylation inhibits the ability of vertebrate, protozoan and plant ADF/ cofilins (see Ref. 6 for review) to bind actin monomers and filaments (14 17). Rapid cycling of phosphorylation and dephosphorylation is another mechanism proposed to augment ADF/cofilins to release actin subunits and recycle for further filament disassembly (18) (reviewed in Ref. 19). Actin-interacting protein 1 (Aip1) was first identified as a novel actin-binding protein using the yeast two-hybrid system (20). Aip1 is conserved in a variety of organisms including Schizosaccharomyces pombe, Physarum polycephalum (21), Dictyostelium discoideum (22, 23), Caenorhabditis elegans (24), Xenopus laevis (25), and birds and mammals (26). Little Aip1 binds to pure actin filaments, whereas added cofilin creates more binding sites (22, 25, 27). Thus, Aip1 is an unusual, cofilin-dependent actin-binding protein. Two-hybrid analysis showed Aip1 interacts with both actin (20) and cofilin (27). In vivo Aip1 is a multicopy suppressor of a cofilin temperature-sensitive mutant (28). An Aip1 null mutant of Dictyostelium had prolonged cytokinesis, delayed phagocytosis and macropinocytosis, and inhibited motility, although cofilin 1 The abbreviations used are: ADF, actin-depolymerizing factor; Aip1, actin-interacting protein 1; Xaip1, Xenopus Aip1; DBP, vitamin D- binding protein; DTT, dithiothreitol; XAC, Xenopus ADF/cofilin; PIPES, 1,4-piperazinediethanesulfonic acid.

2 43012 Aip1 and Cofilin Sever and Cap Actin Filaments distribution was normal (23). Microinjection of excess Aip1 into Xenopus embryos disturbed the cortical localization of both actin and cofilin, suggesting that Aip1 is an activator of ADF/ cofilin, especially in cells where phosphorylation does not regulate ADF/cofilin. In multicellular organisms expressing multiple ADF/cofilins, Aip1 might cooperate with certain isoforms of ADF/cofilin that have weaker severing/depolymerizing activities (29). We used fluorescence microscopy to measure polymer lengths and an actin filament elongation assay to measure the concentration of polymer ends produced by treating filaments with XAip1 and cofilin. We found that Xenopus Aip1 (XAip1) promotes actin filament severing by cofilin but inhibits both elongation and depolymerization of the new barbed ends. Electron micrographs of samples treated with gold-labeled antibodies revealed that XAip1 binds preferentially to one end of cofilin-actin filaments. These results suggest that XAip1 caps filaments severed by cofilin, preventing their annealing and elongation. MATERIALS AND METHODS Proteins XAip1 was purified from Xenopus ovary as described (25) with slight modification. Stepwise ammonium sulfate cuts at 45, 60, and 80% saturation were used. XAip1 is found in the 80% saturated ammonium sulfate precipitate. Recombinant chicken cofilin (30), Xenopus ADF/cofilin (XAC) (31), gelsolin (32), and gelsolin-actin (1:1) complex (33) were purified as described. Actin was purified from rabbit skeletal muscle acetone powder (34) and monomeric Ca-ATP-actin was isolated by Sephacryl S-300 chromatography (35) and labeled on Cys- 374 with pyrene iodoacetamide (36, 37). Actin was polymerized by adding 1:9 (v/v) 10 KME buffer (500 mm KCl, 10 mm MgCl 2,10mM EGTA, and 100 mm PIPES, ph 6.8, or 100 mm Tris-HCl, ph 8.0). Vitamin D-binding protein (DBP), also called Gc-globulin, was from Calbiochem-Novabiochem. Drs. James Casella and Susan Craig of Johns Hopkins Medical School kindly provided spectrin-actin-protein 4.1 complex. Sedimentation Experiments Effects of XAip1 and/or cofilin on the concentration of polymerized actin were obtained by sedimentation of actin filaments at 436,000 g for 20 min in a TLA100 rotor in a Beckman TL100 centrifuge. The total supernatant was removed, and the pellet was resuspended in SDS-containing buffer (0.25 M Tris, ph 6.8, 10% glycerol, 10% 2-mercaptoethanol and 1% SDS). An aliquot of the supernatant was prepared in the same buffer. Fractions from supernatant and pellet were subjected to SDS-PAGE (15% T, 2.67% C polyacrylamide gels), and amounts of the proteins in each fraction were quantified by densitometry following staining with Coomassie Blue R-250 or chemifluorescence detection following staining with Syproorange. In some experiments with XAip1 alone, filaments were prepared in the presence of ADP as well as ATP and were mechanically sheared by 50 passages through a yellow pipette tip. Actin Filament Elongation Experiments Actin filaments (2 M) were incubated in 1 KME buffer with 2 M cofilin and various concentrations of XAip1 at room temperature for 20 min. The mixtures were diluted 3-fold in 2 M actin monomers (10% pyrenyl-labeled) and KME buffer to initiate polymerization. Polymerization of actin was measured as the fluorescence change (excitation at 366 nm and emission at 387 nm) with an Alphascan spectrofluorometer (Photon Technology International, South Brunswick, NJ). Under these conditions, the rate of elongation from the added filaments is much faster than polymerization from spontaneously formed nuclei. Concentrations of actin filaments were calculated from the initial rate of polymerization (V 0 ) according to Equation 1, V 0 k A N k N (Eq. 1) where k is the barbed end association rate constant for ATP-actin (11.4 M 1 s 1 ), [A] is the concentration of actin monomer, [N] isthe concentration of free barbed ends and k is the barbed end dissociation rate constant for ATP-actin (1.4 s 1 ) (38). Except where we capped the barbed ends with gelsolin, we disregarded the pointed end elongation rate, because it contributes little to elongation under these conditions. Actin Filament Depolymerization Experiments The effect of XAip1 on filament depolymerization was measured in the presence of XAC according to the method of Moriyama and Yahara (33). Briefly, recombinant gelsolin at concentrations from 4 to 18 nm was used to nucleate assembly of 3.3 M actin (5% pyrene-labeled) in buffer F (100 mm KCl, 2mM MgCl 2, 0.2 mm ATP, 0.2 mm EGTA, 0.5 mm DTT, and 10 mm Tris at ph 7.8). When the assembly reached steady state (at about 4 h), 0.12 M XAC (with or without 60 nm XAip1) and 150 nm gelsolin-actin (1:1) complex (to cap free barbed ends) were added to the actin and allowed to incubate 10 min. In control experiments the XAC was omitted. Changes in fluorescence intensity were monitored at 18 C with an AVIV ATF 105 spectrofluorometer with excitation at 365 nm and emission at 404 nm. DBP, a potent actin monomer-sequestering protein (K D 1nM (39)), was added in excess to sequester monomers and thus induce depolymerization from pointed ends of the filaments by maintaining the actin concentration below the critical concentration. Initial depolymerization rates were calculated from the fit of the data to the exponential rate equation (V i k[n]), where N equals the number of free pointed ends. In control experiments to demonstrate complete capping by the gelsolin-actin (1:1) complex, filaments were nucleated from spectrin-actin-protein 4.1 seeds and were capped with gelsolinactin prior to treatment with DBP. Fluorescence Microscopy of Actin Filaments Actin filaments were labeled with 2 M rhodamine-phalloidin for 2 min at room temperature, diluted 50-fold with 1 KME buffer, and then diluted into fluorescence buffer (50 mm KCl, 1 mm MgCl 2, 0.1 M DTT, 20 g/ml catalase, 100 g/ml glucose oxidase, 3 mg/ml glucose, 0.5% methyl cellulose, 10 mm imidazole, ph 7.0). An aliquot (2 l) of the sample was applied to a nitrocellulose-coated coverslip, made by applying a 0.1% solution of nitrocellulose in amylacetate. Fluorescence images were observed with an Olympus IX-70 microscope using a mercury illumination source. Images were captured with a Hamamatsu digital camera. Filament lengths were measured manually with IPLab software. Immuno-electron Microscopy Protein solutions were placed on carbon-coated Formvar grids, fixed with 2% formaldehyde and 0.1% glutaraldehyde in 60 mm KCl,2mM MgCl 2,1mM DTT, and 50 mm PIPES, ph 6.8, for 30 min, and washed with the same buffer without formaldehyde. The grids were treated with phosphate-buffered saline containing 1% bovine serum albumin for 1 h, incubated with anti- XAip1 monoclonal antibody, and treated with 10 nm gold-conjugated goat anti-mouse IgG (Jackson Immunoresearch, West Grove, PA) in 1% bovine serum albumin/phosphate-buffered saline. The antibody incubations were performed for 1hatroomtemperature followed by washing thoroughly with phosphate-buffered saline. The specimens were negatively stained with 1.5% uranyl acetate and observed with a JOEL JEM 100CX electron microscope at an accelerating voltage of 80 kv. RESULTS Effects of Cofilin and XAip1 on Actin Filament Length Treatment of 2 M polymerized actin with 2 M cofilin reduced the mean length of the filaments from 5.39 (n 354) to 1.68 m (n 93) at ph 6.8 and to 2.19 m (n 158) at ph 8.0 (Fig. 1). Treatment of actin filaments with XAip1 alone had no effect on polymer length (mean 5.38 m, n 160), but treatment with cofilin and XAip1 reduced the filament lengths beyond the effect of cofilin alone (Fig. 1A, c and d). Effects of Cofilin and XAip1 on the Concentration of Polymerized Actin Samples of F-actin at ph 6.8 or 8.0 were treated with equal molar cofilin and different concentrations of XAip1. After high-speed centrifugation to separate out the filamentous actin, actin in the supernatant and pellet fractions was quantified by SDS-PAGE. At ph 6.8, the combination of the two proteins had no effect on the concentration of actin that pelleted by ultracentrifugation (Fig. 2). At ph 8.0, cofilin plus XAip1 reduced the pelleted polymer concentration only slightly. The difficulty of pelleting short filaments was strong evidence that the polymer concentration did not change. Both with and without cofilin, the concentration of actin in the supernatant was about 0.15 M, close to the critical concentration under these conditions (0.10 M). Effects of XAip1 on the Ability of Cofilin-fragmented Actin to Nucleate Actin Assembly When used as seeds for elongation by actin monomers, the cofilin-treated filaments elongated 2.5- fold faster than untreated actin filaments (Fig. 3A). Because the polymerization rate is proportional to the concentration of

3 Aip1 and Cofilin Sever and Cap Actin Filaments FIG. 1.Effects of cofilin and XAip1 on the length of actin filaments observed by fluorescence microscopy after labeling with rhodamine-phalloidin. Samples of 2 M polymerized actin 2 M recombinant chicken cofilin 20 or 200 nm XAip1 were incubated at room temperature for 20 min in 50 mm KCl, 1 mm MgCl 2,1mM EGTA, and 10 mm buffer and then labeled with 2 M rhodamine-phalloidin for 2 min at room temperature. They were diluted 625-fold and viewed by fluorescence microscopy. A, at ph 6.8: a, actin filaments alone; b, actin with 200 nm XAip1; c, actin with cofilin; d, actin with cofilin and 20 nm XAip1; e, actin with cofilin and 200 nm XAip1. B, at ph 8.0: a, actin with 200 nm XAip1; b, actin with cofilin; c, actin with cofilin and 200 nm XAip1. C (ph 6.8) and D (ph 8.0), length distribution plots showing the percentage of filaments longer than X versus length X. Filled diamonds, actin alone; filled squares, actin with 200 nm XAip1; filled triangles, actin with cofilin; open circles, actin with cofilin and 20 nm XAip1; filled circles, actin with cofilin and 200 nm XAip1. barbed ends, treatment with cofilin increased the concentration of ends 2.5-fold, in reasonable agreement with the 3.1-fold reduction in length at constant actin polymer concentration. At ph 8.0 cofilin increased the initial elongation rate 4-fold and reduced the length 2.6-fold. The elongation rate is more likely than microscopy to give an accurate measurement of polymer ends in samples with cofilin, because short filaments are more difficult to detect than long ones. These observations confirm the severing activity of cofilin observed previously (reviewed in Refs. 6 and 40). Treatment of actin filaments with XAip1 had no effect on the rate of elongation when the filaments were used as seeds at either ph 6.8 (Fig. 3A, h) or 8.0 (data not shown). Combined with our observation above that XAip1 had no effect on polymer length, these results confirm that XAip1 alone neither severs actin filaments nor caps their barbed ends. Although treatment of filaments with cofilin plus XAip1 reduced their lengths beyond the effect of cofilin alone (see Fig. 1), paradoxically the combination of cofilin and XAip1 also reduced the concentration of barbed ends capable of elongation (Fig. 3). At ph 6.8 with 2 M cofilin, mean filament lengths were 0.6 m(n 241) with 20 nm XAip1 and 0.53 m(n 483) with 200 nm XAip1. The concentration of ends produced by cofilin depended on the concentration of XAip1 (Fig. 3). At ph 6.8 with 2 M XAC, the minimum concentration of ends reached FIG. 2. Effect of cofilin and XAip1 on the concentration of polymerized actin. Samples of 2 M actin filaments in 50 mm KCl, 1 mm MgCl 2,1mM EGTA, and 10 mm buffer were treated with 2 M recombinant chicken cofilin and a range of concentrations of XAip1 and then centrifuged at 436,000 g for 20 min to pellet actin filaments. Proteins in the supernatant (s) and pellet (p) were subjected to SDS- PAGE and stained with Sypro-orange. Upper panel (ph 6.8) and lower panel (ph 8.0), from left to right in the presence of 0, 10, 20, 50, 100, 200 nm XAip1. The graph shows the concentrations of pelleted actin at ph 6.8 (filled circles) and ph 8.0 (open squares) over the range of XAip1 concentrations. FIG. 3.Effect of XAip1 and cofilin on elongation of actin filament seeds. A and B, time course of actin polymerization from preexisting filament ends. Conditions: 2 M polymerized actin was mixed with 2 M recombinant chicken cofilin and a range of concentrations of XAip1 either at ph 6.8 (A) or 8.0 (B) and incubated at room temperature for 20 min. These filaments were diluted 3-fold into 2 M pyrenyl Mg-ATP-actin monomers with polymerization buffer to initiate elongation. In A, XAip1 concentrations were: a, 0; b, 5 nm; c, 10 nm; d, 20 nm; e, 50nM; f, 100 nm; g, 200 nm; h, actin filaments plus 200 nm XAip1; i, actin filaments alone. In B, XAip1 concentrations were: a, 0 nm; b, 10 nm; c, 20 nm; d, 50 nm; e, actin filaments alone; f, actin filaments with cofilin 100 nm XAip1; g, actin filaments with cofilin and 200 nm XAip1. C and D, dependence of filament length on the concentration of XAip1. Concentration of filament ends is calculated from the elongation rate using Equation 1 (open circles). Filament lengths are calculated from the filament end concentration (open squares). Average filament length is measured by fluorescence microscopy from Fig. 1, A and B (filled squares).

4 43014 Aip1 and Cofilin Sever and Cap Actin Filaments FIG. 4.XAip1 does not enhance severing or depolymerization of XAC-treated F-actin when barbed ends are capped with gelsolin-actin. A, filaments were assembled in 100 mm KCl, 2 mm MgCl 2, 0.2 mm EGTA, and 10 mm buffer, ph 7.8, from 3.3 M (5% pyrenyl) actin initiated by 60 nm spectrin-actin seeds (to block pointed ends) or 21 nm gelsolin (to block barbed ends). Both produced 1.8 nm filaments (as determined by assembly rate assays to quantify filament ends). One sample of actin assembled from spectrin-actin seeds was treated with 200 nm gel-filtered gelsolin-actin (1:1) complex (33). These filaments were treated with 3.5 M DBP to bind monomers and initiate depolymerization, and the time course of the pyrene fluorescence was recorded. a, filaments seeded by spectrin-actin and treated with gelfiltered 200 nm gelsolin-actin (both ends blocked); b, filaments nucleated from gelsolin (21 nm) with 200 nm gelsolin-actin added (barbed ends blocked, depolymerization from pointed ends); c, no treatment (pointed ends blocked, depolymerization from barbed ends). The inset table shows the calculated depolymerization rates in fluorescence units/min for the three curves. GA, gelsolin-actin. B, filaments were assembled as in A using different concentrations of gelsolin to nucleate assembly. After assembly was complete (about 4 h), filaments were incubated with 150 nm gelsolin-actin (1:1) complex alone (triangles), with the complex along with 0.12 M XAC (diamonds), or with 0.12 mm XAC plus 60 nm XAip1 (squares). After 10 min, 3.5 mm DBP was added, and the depolymerization was followed as described in A. Initial depolymerization rates were measured by a computer fit to the exponential fluorescence decrease; these rates were plotted versus the nucleating gelsolin concentration, which is assumed to be proportional to the initial filament concentration (33). XAip1 did not enhance the depolymerizing activity of XAC when filament barbed ends were capped. a plateau at concentrations of XAip1 in the range of 100 nm, far below the concentration of 2 M polymerized actin (Fig. 3C). At ph 8.0 XAip1 and cofilin reduced the concentration of ends even further, below that of untreated actin filaments (Fig. 3D). XAip1 and tissue-derived XAC gave similar results at ph 6.8. Given that the filaments were shorter but had fewer ends capable of elongation, the combination of cofilin and XAip1 must cap most barbed ends. XAip1 Does Not Enhance F-actin Depolymerization by Cofilin FIG. 5.Localization of XAip1 binding sites on actin filaments by electron microscopy of negatively stained specimens. The anti-xaip1 monoclonal antibody was localized with a secondary antibody conjugated to 10 nm gold. A, 4.7 M actin was polymerized with 5.2 M recombinant chicken cofilin and then incubated with 120 nm XAip1 for 5 min. B, 4.8 M actin was polymerized with equimolar XAC and then incubated with 960 nm XAip1 for 15 min. Bar in A 0.2 m; bar in B 0.1 m. FIG. 6.XAip1 alone does not bind F-actin nor does it prevent the annealing/redistribution of mechanically sheared actin filaments. F-actin (5 M), alone or in the presence of 1 M XAip1, and in the presence of either ATP or ADP (1 mm), was mechanically sheared by 50 passages through a yellow tip. Samples were centrifuged immediately (lanes 13 16) or allowed to stand for 40 min before sedimentation at 436,000 g for 20 min, and the proteins in the supernatant and pellet fractions were analyzed by SDS-PAGE. The amount of XAip1 in the pellet fraction was the same for all treatments. Filament annealing/ redistribution was rapid in ATP such that samples centrifuged immediately after shearing (lanes 14 and 15) showed no increase in supernatant actin over samples allowed to recover (lanes 5 6 and 9 10). Samples sheared in ADP and centrifuged immediately (lanes 15 and 16) showed a significant fraction (about 30%) of the actin in the supernatant, whereas after 40 min recovery (lanes 7 8 and 11 12), the actin distribution was the same as for the ATP samples. The presence of XAip1 did not prevent full recovery of the actin distribution as found in the control (lanes 1 and 2). We measured the effect of cofilin in the presence and absence of XAip1 on the rate of depolymerization from filament pointed ends at ph 7.8 using DBP to sequester actin monomers as they dissociated spontaneously from polymer ends. Filaments with free barbed ends (grown from spectrin-actin-protein 4.1 seeds (41)) depolymerize rapidly upon addition of DBP but do not depolymerize when treated prior to addition of DBP with gelsolin-actin (1:1) complex (Fig. 4A), demonstrating that the gelsolin-actin complex effectively caps the filament barbed ends. Actin filaments nucleated with different amounts of gelsolin depolymerized slowly in the presence of DBP, but the rate was greatly enhanced by the addition of 0.12 M XAC (Fig. 4B). The addition of 60 nm XAip1 to the XAC did not alter the depolymerization kinetics, indicating that it did not increase the severing activity of XAC to generate more free pointed ends, nor did it alter the off-rates at the pointed ends already present. Localization of XAip1 on Actin Filaments We used immuno-electron microscopy to visualize XAip1 binding sites on cofilin-decorated actin filaments. At 120 nm XAip1 was localized with a higher frequency at one end than along the side of actin filaments (Fig. 5A). Gold beads were observed at one end of 82% of filaments, at both ends of 0% of filaments, and along the sides of 38% of filaments (n 133). At 960 nm XAip1 gold-labeled antibodies bound along the length of most filaments (Fig. 5B). XAip1 Alone Does Not Inhibit Annealing/Redistribution of Mechanically Sheared F-actin To determine whether XAip1 alone was able to prevent annealing/redistribution of F-actin, we subjected samples of 5 M F-actin in the presence or absence of 1 M XAip1 to mechanical shear (50 passages through a yellow pipette tip) and sedimented the sample immediately at 436,000 g for 20 min or allowed it to recover for 40 min before centrifugation. Supernatant and pellet fractions were then analyzed by SDS-PAGE (Fig. 6). Filaments were slow to anneal after shearing in the presence of ADP (compare lanes 11 and 12 with 15 and 16 in Fig. 6), as there was a larger fraction of actin in the supernatant of these samples. But even in the presence

5 Aip1 and Cofilin Sever and Cap Actin Filaments of 1 M XAip1, filament distribution between supernatant and pellet fractions recovered completely after 40 min. DISCUSSION The paradoxical effects of cofilin plus XAip1 in reducing both the length and number of filament ends capable of elongation under conditions in which polymer mass is constant is explained most simply by these proteins capping many of the barbed ends produced by severing. We conclude from the strong inhibition of elongation and the inhibition of depolymerization that barbed ends are capped, because capping pointed ends would have a minimal effect on either the elongation or depolymerization rates. Capping by XAip1 appears to depend on cofilin in two ways. First, severing by cofilin creates ends for capping by XAip1. Second, capping itself may also depend on the presence of cofilin, as we detected no capping of pure actin filaments by adding XAip1 at either ph 6.8 or 8.0. Thus, the observed capping was not due to contamination with animal capping protein or gelsolin, both of which cap F-actin alone. Inhibition of annealing secondary to capping may explain why XAip1 reduces the length of fragments produced by cofilin. Other barbed end capping proteins inhibit not only elongation but also annealing of the fragments produced by severing or fragmentation. The steady state polymer length in the presence of ADF/cofilin depends on the balance between ongoing severing and annealing (8, 9), and therefore inhibition of annealing with ongoing severing would produce shorter filaments without the necessity of invoking an increased rate of severing. Many details remain to be investigated. For example, we do not yet know the affinity of XAip1 for barbed ends or the effectiveness of the capping. The XAip1 concentration dependence of the reduction in free ends (Fig. 3, C and D) suggests that the affinity is high, with a K d in the 25 nm range, but this experiment is not a definitive method to measure affinity. The residual elongation of filaments saturated with XAip1 may be due to pointed end growth or to incomplete capping of barbed ends. Incomplete capping would explain why the combination of cofilin and XAip1 does not reduce the polymer concentration. Complete capping of barbed ends is required to increase the critical concentration of Mg-ATP actin from 0.1 to 0.7 M. The pioneering work on Aip1 (22, 25, 27) concluded that Aip1 enhances the disassembly of actin filaments by cofilin. However, these conclusions were based upon data from sedimentation experiments in which it is difficult to separate short filaments from monomers. Indeed, Okada et al. (25) did examine the nonsedimentable pool of actin monomer using the DNase I inhibition assay and found no increase in G-actin when XAip1 was added to cofilin-actin mixtures. In retrospect all of the published data are consistent with severing and capping, which can reduce light scattering and pelleting of short actin filaments. These papers also concluded that Aip1 has a low affinity for actin filaments but could not distinguish between low affinity and a low concentration of binding sites, as would occur if the high affinity binding site were at the end of a filament. In fact, the published data are consistent with high affinity binding to a limited number of sites. For example, Rodal et al. (27) detected no binding of nm yeast Aip1 to 3.75 M actin filaments (about 2 nm ends, assuming a mean polymer length of 5 m) but did detect binding of a low (but unspecified) concentration in radioactive in vitro translated Aip1, which was enhanced by cofilin. Okada et al. (25) demonstrated additional lower affinity binding sites by carrying out sedimentation binding studies to XAip1 concentrations of 6 M, which likely represent the lower affinity lateral binding observed here by electron microscopy. The exchange of nucleotide at the barbed end of actin filaments is very rapid (42). When ADF/cofilins bind to G-actin they inhibit nucleotide exchange (43 45) and would likely do so as well at the barbed ends of filaments. Thus the barbed ends of unsevered filaments will consist of ATP-actin, whereas the barbed ends of ADF/cofilin severed filaments will consist of ADP-actin. Nucleotide exchange at these ends will be slow. Thus, our results suggest that XAip1 may have a specific affinity for the ADF/cofilin-ADP-actin complex at the barbed end of newly severed filaments and little or no affinity for the ATP-actin barbed ends of unsevered filaments or mechanically severed filaments in which rapid nucleotide exchange at the barbed end will occur. In vivo studies on Aip1 point to its role in regulating actin dynamics and remodeling. Aip1 (Unc78) mutants in C. elegans show abnormal accumulation of actin aggregates in muscle (24), whereas in Aip1 null yeast, cofilin localizes abnormally to actin cables, which turn over inefficiently (27). Aip1 null mutants of Dictyostelium were defective in cytokinesis, phagocytosis, and motility (23). Overexpression of Aip1 in Dictyostelium inhibited phagocytosis similarly to that of the actin monomer binding compound, latrunculin A, suggesting that Aip1 increased the monomer pool (23). An increase in the disassembly of actin upon microinjection of Aip1 into Xenopus oocytes was measured directly (25). All of these effects can be explained by the in vitro activity of XAip1 reported here. Blocking the annealing and barbed end growth of ADF/cofilin fragmented filaments will enhance the number of pointed ends from which ADF/cofilins can depolymerize actin. It is not yet known how XAip1 affects the ability of ADF/cofilins to depolymerize the F-actin from the pointed end. However, the fact that XAip1 caps barbed filament ends and inhibits elongation may shift the state of cofilin-decorated actin filaments to disassembly, thus amplifying the ADF/cofilin function of increasing actin monomers that can be recycled rapidly for localized actin polymerization. REFERENCES 1. Borisy, G. G., and Svitkina, T. M. (1999) Curr. Opin. Cell Biol. 12, Pollard, T. D., Blanchoin, L., and Mullins, R. D. (2000) Annu. Rev. Biophys. Biomol. Struct. 29, Loisel, T. P., Boujemaa, R., Pantaloni, D., and Carlier, M. F. (1999) Nature 401, Blanchoin, L., Amann, K. J., Higgs, H. N., Marchand, J. B., Kaiser, D. D., and Pollard T. D. (2000) Nature 404, Blanchoin, L., Pollard, T. D., and Mullins, R. D. (2000) Curr. Biol. 10, Bamburg, J. R. (1999) Annu. Rev. Cell Dev. Biol. 15, Chen, H., Bernstein, B. W., and Bamburg, J. R. (2000) Trends. Biochem. Sci. 25, Blanchoin, L., and Pollard, T. D. (1998) J. Biol. Chem. 273, Maciver, S. K., Pope, B, J., Whytock, S., and Weeds, A. G. (1998) Eur. J. Biochem. 56, Carlier, M.-F., Laurent, V., Santolini, J., Melki, R., Didry, D., Xia, G., Hong, Y., Chua, N., and Pantaloni, D. (1997) J. Cell Biol. 136, Zebda, N., Bernard, O., Bailly, M., Welti, S., Lawrence, D. S., and Condeelis, J. S. (2000) J. Cell Biol. 151, Ichetovkin, I., Grant, W., and Condeelis, J. (2002) Curr. Biol. 12, Didry, D., Carlier, F. F., and Pantaloni, D. (1998) J. Biol. Chem. 273, Morgan, T. E., Lockerbie, R. O., Minamide L. S., Browning M. D., and Bamburg, J. R. J. Cell Biol. 122, Agnew, B. J., Minamide, L. S., and Bamburg, J. R. (1995) J. Biol. Chem. 270, Moriyama, K., Iida, K., and Yahara, I. (1996) Genes Cells 1, Blanchoin, L., Robinson. R. C., Choe, S., and Pollard, T. D. (2000) J. Mol. Biol. 295, Meberg, P. J., Ono, S., Minamide, L. S., Takahashi, M, and Bamburg, J. R. (1998) Cell Motil. Cytoskeleton 39, Kuhn, T. B., Meberg, P. J., Brown, M. D., Bernstein, B. W., Minamide, L. S., Jensen, J. R., Okada, K., Soda, E. A., and Bamburg, J. R. (2000) J. Neurobiol. 44, Amberg, D. C., Basart, E., and Botstein, D. (1995) Struct. Biol. 2, Matsumoto, S., Ogawa, M., Kasakura, T., Shimada, Y., Mitsui, M., Maruya, M., Isohata, M., Yahara, I., and Murakami-Murofushi, K. (1998) J. Biochem. 124, Aizawa, H., Katadae, M., Maruya, M., Sameshima, M., Murakami-Murofushi, K., and Yahara, I. (1999) Genes Cells 4, Konzok, A., Weber, I., Simmeth, E., Hacker, U., Maniak, M., and Müller- Taubenberger, A. (1999) J. Cell Biol. 146, Ono, S. (2001) J. Cell Biol. 152,

6 43016 Aip1 and Cofilin Sever and Cap Actin Filaments 25. Okada, K., Obinata, T., and Abe, H. (1999) J. Cell Sci. 112, Adler, H. J., Winnicki, R. S., Gong, T. L., and Lomax, M. I. (1999) Genomics 56, Rodal, A. A., Tetreault, J. W., Lappalainen, P., Drubin, D. G., and Amberg, D. C. (1999) J. Cell Biol. 145, Iida, K., and Yahara, I. (1999) Genes Cells 4, Bernstein, B. W, Painter, W. B., Chen, H., Minamide, L. S., Abe, H., and Bamburg, J. R. (2000) Cell Motil. Cytoskeleton 47, Abe, H., Endo, T., Yamamoto, K., and Obinata, T. (1990) Biochemistry 29, Abe, H., Obinata, T., Minamide, L. S., and Bamburg, J. R. (1996) J. Cell Biol. 132, Pope, B. J., Gooch, J. T., and Weeds, A, G. (1997) Biochemistry 36, Moriyama, K., and Yahara, I. (1999) EMBO J. 18, Spudich, J. A., and Watt, S. (1971) J. Biol. Chem. 246, MacLean-Fletcher, S., and Pollard, T. D. (1980) Cell 20, Kouyama, T., and Mihashi, K. (1981) Eur. J. Biochem. 114, Pollard, T. D. (1984) J. Cell Biol. 99, Pollard, T. D. (1986) J. Cell Biol. 103, McLeod, J. F., Kowalski, M. A., and Haddad, J. G., Jr. (1989) J. Biol. Chem. 264, Maciver, S. K. (1998) Curr. Opin. Cell Biol. 10, Casella, J. F., Maack, D. J., and Lin, S. (1986) J. Biol. Chem. 261, Teubner, A., and Wegner, A. (1998) Biochemistry 37, Nishida, E. (1985) Biochemistry 24, Hawkins, M., Pope, B., Maciver, S. K., and Weeds, A. G. (1993) Biochemistry 32, Hayden, S. M., Miller, P. S., Brauweiler, A., and Bamburg, J. R. (1993) Biochemistry 32,