Three-dimensional structure of actin filaments and of an actin gel made with actin-binding protein

Purified muscle actin and mixtures of actin and actin-binding protein were examined in the transmission electron microscope after fixation, critical point drying, and rotary shadowing. The three-dimensional structure of the protein assemblies was analyzed by a computer-assisted graphic analysis applicable to generalized filament networks. This analysis yielded information concerning the frequency of filament intersections, the filament length between these intersections, the angle at which filaments branch at these intersections, and the concentration of filaments within a defined volume. Purified actin at a concentration of 1 mg/ml assembled into a uniform mass of long filaments which overlap at random angles between 0 degrees and 90 degrees. Actin in the presence of macrophage actin-binding protein assembled into short, straight filaments, organized in a perpendicular branching network. The distance between branch points was inversely related to the molar ratio of actin-binding protein to actin. This distance was what would be predicted if actin filaments grew at right angles off of nucleation sites on the two ends of actin-binding protein dimers, and then annealed. The results suggest that actin in combination with actin-binding protein self-assembles to form a three- dimensional network resembling the peripheral cytoskeleton of motile cells.     

Cytochalasin B and the structure of actin gels

We analyzed the structure of gels formed when macrophage actin-binding protein crosslinks skeletal muscle actin polymers and the effect of the fungal metabolite cytochalasin B on this structure. Measurement of the actin filament length distribution permitted calculation of the critical concentration of crosslinker theoretically required for gelation of actin polymer networks. The experimentally determined critical concentration of actin-binding protein agreed sufficiently with the theoretical to conclude that F-actin-actin-binding protein gels are networks composed of isotropically oriented filaments crosslinked at intervals. The effects of cytochalasin B on these actin networks fits this model. Cytochalasin B (1) bound to F-actin (but not to actin-binding protein), (2) decreased the length of actin filaments without increasing the quantity of monomeric actin, (3) decreased the rigidity of actin networks both in the presence and absence of crosslinking proteins and (4) increased the critical concentration of actin-binding protein required for incipient gelation by a magnitude predicted from network theory if filaments were divided and shortened by the extents observed. The effects of cytochalasin B on gelation were highly dependent on actin concentration and were inhibited by the actin-stabilizing agent phalloidin. Therefore, cytochalasin B diminishes actin gel structure by severing actin filaments at limited sites. The demonstration of gel-sol transformations in actin networks caused by limited actin filament cleavage suggests a new mechanism for the control of cytoplasmic structure.     

Mechanisms of actin rearrangements mediating platelet activation.

The detergent-insoluble cytoskeleton of the resting human blood platelet contains approximately 2,000 actin filaments approximately 1 micron in length crosslinked at high angles by actin-binding protein and which bind to a spectrin-rich submembrane lamina (Fox, J., J. Boyles, M. Berndt, P. Steffen, and L. Anderson. 1988. J. Cell Biol. 106:1525-1538; Hartwig, J., and M. DeSisto. 1991. J. Cell Biol. 112:407-425). Activation of the platelets by contact with glass results within 30 s in a doubling of the polymerized actin content of the cytoskeleton and the appearance of two distinct new actin structures: bundles of long filaments within filopodia that end at the filopodial tips (filopodial bundles) and a circumferential zone of orthogonally arrayed short filaments within lamellipodia (lamellipodial network). Neither of these structures appears in cells exposed to glass with cytochalasin B present; instead the cytoskeletons have numerous 0.1-0.3-microns-long actin filament fragments attached to the membrane lamina. With the same time course as the glass-induced morphological changes, cytochalasin-sensitive actin nucleating activity, initially low in cytoskeletons of resting platelets, increases 10-fold in cytoskeletons of thrombin-activated platelets. This activity decays with a time course consistent with depolymerization of 0.1-0.3-microns-long actin filaments, and phalloidin inhibits this decay. Cytochalasin-insensitive and calcium-dependent nucleation activity also increases markedly in platelet extracts after thrombin activation of the cells. Prevention of the rise in cytosolic Ca2+ normally associated with platelet activation with the permeant Ca2+ chelator, Quin-2, inhibits formation of lamellipodial networks but not filopodial bundles after glass contact and reduces the cytochalasin B-sensitive nucleation activity by 60% after thrombin treatment. The filopodial bundles, however, are abnormal in that they do not end at the filopodial tips but form loops and return to the cell body. Addition of calcium to chelated cells restores lamellipodial networks, and calcium plus A23187 results in cytoskeletons with highly fragmented actin filaments within seconds. Immunogold labeling with antibodies against gelsolin reveals gelsolin molecules at the ends of filaments attached to the submembrane lamina of resting cytoskeletons and at the ends of some filaments in the lamellipodial networks and filopodial bundles of activated cytoskeletons. Addition of monomeric actin to myosin subfragment 1-labeled activated cytoskeletons leads to new (undecorated) filament growth off the ends of filaments in the filopodial bundles and the lamellipodial network. The simplest explanation for these findings is that gelsolin caps the barbed ends of the filaments in the resting platelet. Uncapping some of these filaments after activation leads to filopodial bundles.(ABSTRACT TRUNCATED AT 400 WORDS)     

Filamin A, the Arp2/3 complex, and the morphology and function of cortical actin filaments in human melanoma cells.

The Arp2/3 complex and filamin A (FLNa) branch actin filaments. To define the role of these actin-binding proteins in cellular actin architecture, we compared the morphology of FLNa-deficient human melanoma (M2) cells and three stable derivatives of these cells expressing normal FLNa concentrations. All the cell lines contain similar amounts of the Arp2/3 complex. Serum addition causes serum-starved M2 cells to extend flat protrusions transiently; thereafter, the protrusions turn into spherical blebs and the cells do not crawl. The short-lived lamellae of M2 cells contain a dense mat of long actin filaments in contrast to a more three-dimensional orthogonal network of shorter actin filaments in lamellae of identically treated FLNa-expressing cells capable of translational locomotion. FLNa-specific antibodies localize throughout the leading lamellae of these cells at junctions between orthogonally intersecting actin filaments. Arp2/3 complex-specific antibodies stain diffusely and label a few, although not the same, actin filament overlap sites as FLNa antibody. We conclude that FLNa is essential in cells that express it for stabilizing orthogonal actin networks suitable for locomotion. Contrary to some proposals, Arp2/3 complex-mediated branching of actin alone is insufficient for establishing an orthogonal actin organization or maintaining mechanical stability at the leading edge.     

Identification of a membrane skeleton in platelets

Platelets have previously been shown to contain actin filaments that are linked, through actin-binding protein, to the glycoprotein (GP) Ib-IX complex, GP Ia, GP IIa, and an unidentified GP of Mr 250,000 on the plasma membrane. The objective of the present study was to use a morphological approach to examine the distribution of these membrane-bound filaments within platelets. Preliminary experiments showed that the Triton X-100 lysis buffers used previously to solubilize platelets completely disrupt the three-dimensional organization of the cytoskeletons. Conditions were established that minimized these postlysis changes. The cytoskeletons remained as platelet-shaped structures. These structures consisted of a network of long actin filaments and a more amorphous layer that outlined the periphery. When Ca2+ was present, the long actin filaments were lost but the amorphous layer at the periphery remained; conditions were established in which this amorphous layer retained the outline of the platelet from which it originated. Immunocytochemical experiments showed that the GP Ib-IX complex and actin-binding protein were associated with the amorphous layer. Analysis of the amorphous material on SDS-polyacrylamide gels showed that it contained actin, actin-binding protein, and all actin-bound GP Ib-IX. Although actin filaments could not be visualized in thin section, the actin presumably was in a filamentous form because it was solubilized by DNase I and bound phalloidin. These studies show that platelets contain a membrane skeleton and suggest that it is distinct from the network of cytoplasmic actin filaments. This membrane skeleton exists as a submembranous lining that, by analogy to the erythrocyte membrane skeleton, may stabilize the plasma membrane and contribute to determining its shape.     

Presence of calmodulin in human platelet cytoskeletons and its concentration change upon activation of platelets

We found that a small, reproducible amount of calmodulin is present in the cytoskeleton of human platelets. Triton-insoluble materials (cytoskeletons), which were prepared by cetrifugation at 1000 × g for 10 min of platelets after lysis by Triton X-100, stimulated cyclic AMP phosphodiesterase activity in the presence of Ca²⁺ but not in the presence of the calcium chelator, EGTA, or the calmodulin antagonist, trifluoperazine. The activation of the enzyme was also obtained after heating Triton-insoluble materials. An alkaline glycerol polyacrylamide gel electrophoresis of fractions obtained after gel fitration of solubilized Triton residues showed a protein band which had a faster electrophoretic mobility in the absence than in the presence of Ca²⁺. Upon thrombin activation of platelets, calmodulin in the Triton-insoluble cytoskeletons increased rapidly parallel to actin, actin-binding protein and myosin. With other stimulants such as collagen, epinephrine and ADP, similar results were obtained but with slower association of these proteins with cytoskeletons. However, after treatment with the Ca²⁺-inophore A23187, calmodulin, actin and actin-binding protein in Triton residues decreased rapidly, whereas the association of myosin increased. Thus, calmodulin seems to be associated with actin filaments rather than myosin filaments, and may be involved in the generation of contractile force in the cell.     

Regulation of water flow by actin-binding protein-induced actin gelatin.

Actin filaments inhibit osmotically driven water flow (Ito, T., K.S. Zaner, and T.P. Stossel. 1987. Biophys. J. 51: 745-753). Here we show that the actin gelation protein, actin-binding protein (ABP), impedes both osmotic shrinkage and swelling of an actin filament solution and reduces markedly the concentration of actin filaments required for this inhibition. These effects depend on actin filament immobilization, because the ABP concentration that causes initial impairment of water flow by actin filaments corresponds to the gel point measured viscometrically and because gelsolin, which noncovalently severs actin filaments, solates actin gels and restores water flow in a solution of actin cross-linked by ABP. Since ABP gels actin filaments in the periphery of many eukaryotic cells, such actin networks may contribute to physiological cell volume regulation.     

Isolation of actin-binding protein and villin from toad oocytes

Two actin-modulating proteins have been purified from toad oocytes. A high-molecular weight protein, similar in structure and function to macrophage actin-binding protein, accounts for the isotropic actin-crosslinking activity in oocyte homogenates. A calcium-dependent activity in toad oocyte homogenates which shortens actin filaments is accounted for by a 95,000-dalton protein which resembles villin, an actin-severing and -bundling protein of avian epithelial brush borders. In the presence of high (? μM) calcium, this protein shortens actin filaments in a concentration-dependent fashion and stimulates filament assembly when added to monomeric actin. In the absence of calcium the protein promotes the formation of actin filament bundles. Therefore, in the toad oocyte actin can be crosslinked into a network by actin-binding protein. Calcium regulation of the actin network may be mediated by villin. These results are different from those reported in echinoderm eggs.     

Dynamic polymorphism of single actin molecules in the actin filament

Actin filament dynamics are critical in cell motility. The structure of actin filament changes spontaneously and can also be regulated by actin-binding proteins, allowing actin to readily function in response to external stimuli. The interaction with the motor protein myosin changes the dynamic nature of actin filaments. However, the molecular bases for the dynamic processes of actin filaments are not well understood. Here, we observed the dynamics of rabbit skeletal-muscle actin conformation by monitoring individual molecules in the actin filaments using single-molecule fluorescence resonance energy transfer (FRET) imaging with total internal reflection fluorescence microscopy (TIRFM). The time trajectories of FRET show that actin switches between low- and high-FRET efficiency states on a timescale of seconds. If actin filaments are chemically cross-linked, a state that inhibits myosin motility, the equilibrium shifts to the low-FRET conformation, whereas when the actin filament is interacting with myosin, the high-FRET conformation is favored. This dynamic equilibrium suggests that actin can switch between active and inactive conformations in response to external signals.     

Actin filament uncapping localizes to ruffling lamellae and rocketing vesicles

Regulated actin filament assembly is critical for eukaryotic cell physiology. Actin filaments are polar structures, and those with free high affinity or barbed ends are crucial for actin dynamics and cell motility. Actin filament barbed-end-capping proteins inhibit filament elongation after binding, and their regulated disassociation is proposed to provide a source of free filament ends to drive processes dependent on actin polymerization. To examine whether dissociation of actin filament capping proteins occurs with the correct spatio-temporal constraints to contribute to regulated actin assembly in live cells, I measured the dissociation of an actin capping protein, gelsolin, from actin in cells using a variation of fluorescence resonance energy transfer (FRET). Uncapping was found to occur in cells at sites of active actin assembly, including protruding lamellae and rocketing vesicles, with the correct spatio-temporal properties to provide sites of actin filament polymerization during protrusion. These observations are consistent with models where uncapping of existing filaments provides sites of actin filament elongation.     

Actin-binding protein amplifies actomyosin contraction,and gelsolin confers calcium control on the direction of contraction

Cross-linking of muscle actin filaments by low concentrations of actin-binding protein reduces the concentration of muscle myosin required for contraction of actin. Gelsolin, a macrophage protein that divides actin filaments in the presence of calcium, inhibits the amplifying effect of actin-binding protein on contraction of actomyosin. In a calcium gradient, the actomyosin gel moves from high to low calcium concentrations, indicating that calcium-controlled lattice formation can impart directionality to the movement of an isotropic actin network.     

Structure of macrophage actin-binding protein molecules in solution and interacting with actin filaments

We have examined the structure of actin-binding molecules in solution and interacting with actin filaments. At physiological ionic strength, actin-binding protein has a M_r value of 540 × 10³ as determined by direct and indirect hydrodynamic measurements. It is an asymmetrical dimer composed of 270 × 10³ dalton subunits. Viewed in the electron microscope after negative staining or low angle shadowing, actin-binding protein molecules assume a broad range of conformations varying from closed circular structures to fully extended strands 162 nm in contour length. All configurations are apparently derived from the same structure which consists of two monomer chains connected end-to-end. The radius of gyration determined from the electron microscopic images was 21.3 nm in agreement with the value of 17.6 nm calculated from hydrodynamic assays. The average axial ratio from hydrodynamic measurements was 17:1, whereas fully extended dimer molecules in the electron microscope would have an axial ratio of 54:1. All of these observations indicate that actin-binding protein dimers are extremely flexible. The flexibility parameter λ (Landau &; Lifshits, 1958) for actinbinding protein is 0.18 nm^?1.As determined by sedimentation, actin-binding protein binds to actin filaments with a K_a value of 2 × 10⁶m^?1 and a capacity of one dimer to 14 actin monomers in filaments. After incubation of high concentrations (molar ratio to actin ≥ 1:10) of actin-binding protein with actin filaments, long filament bundles are visible in the electron microscope. Under these conditions, actin-binding protein molecules decorate the actin filaments in the bundles at regular 40 nm intervals or once every 15 monomers, approximately equivalent to the binding capacity measured by sedimentation. Low concentrations of actin-binding protein (molar ratio to actin ≥ 1:50) which promote the gelation of actin filaments in solution, did not detectably alter the isotropy of the actin filaments. Direct visualization of actinbinding protein molecules between actin filaments in the electron microscope showed that dimers are sufficient for crossbridging of actin filaments and that actinbinding protein dimers are bipolar, composed of monomers connected head-to-head and having actin-binding sites located on the free tails.We conclude that actin-binding protein is a dimer at physiological ionic strength. Each dimer has two actin filament binding sites and is therefore sufficient to gel actin filaments in solution. The length and flexibility of the actin-binding protein subunits render this molecule structurally suited for the crosslinking of large helical filaments into isotropic networks.     

Actin-Binding Proteins in Plant Cells

Abstract: Actinoccurs in all plant cells, as monomers, filaments and filament assemblies. In interphase, actin filaments form a cortical network, co-align with cortical microtubules, and extend throughout the cytoplasm functioning in cytoplasmic streaming. During mitosis, they co-align with microtubules in the preprophase band and phragmoplast and are indispensa ble for cell division. Actin filaments continually polymerise and depolymerise from a pool of monomers, and signal transduction pathways affecting cell morphogenesis modify the actin cytoskeleton. The interactions of actin monomers and filaments with actin-binding proteins (ABP5) control actin dynamics. By binding to actin monomers, ABPs, such as profilin, regulate the pool of monomers available for polymerisation. By breaking filaments or capping filament ends, ABPs, such as actin depoly-merising factor (ADF), prevent actin filament elongation or loss of monomers from filament ends. By bivalent cross-linking to actin filaments, ABPs, such as fimbrin and other members of the spectrin family, produce a variety of higher order assemblies, from bundles to networks. The motor protein ABPs,. which are not covered in this review, move organelles along ac tin filaments. The large variety of ABPs share a number of functional modules. A plant representative of ABPs with particular modules, and therefore particular functions, is treated in this review.     

Ca2+ control of actin gelation. Interaction of gelsolin with actin filaments and regulation of actin gelation

We elucidated the mechanism by which gelsolin, a Ca2+-dependent regulatory protein from lung macrophages, controls the network structure of actin filaments. In the presence of micromolar Ca2+, gelsolin bound Ca2+. The Ca2+-gelsolin complex reduced the apparent viscosity and flow birefringence of F-actin and the lengths of actin filaments viewed in the electron microscope. However, concentrations of gelsolin causing these alterations did not effect proportionate changes in the turbidity of actin filament solutions or in the quantity of nonsedimentable actin as determined by a radioassay. From these findings, we conclude that gelsolin shortens actin filaments without net depolymerization. Such an effect on the distribution of actin filament lengths led to the prediction that low concentrations of gelsolin would increase the critical concentration of actin-binding protein required for incipient gelation of actin filaments in the presence of Ca2+, providing an efficient mechanism for controlling actin network structure. We verified the prediction experimentally, and we estimated that the Ca2+-gelsolin complex effectively breaks the bond between actin monomers in filaments with a stoichiometry of 1:1. The effect of Ca2+-gelsolin complex on actin solation was rapid, independent of temperature between 0 degrees and 37 degrees C, and reversed by reducing the free Ca2+ concentration.     

The interaction of cofilin with the actin filament

Cofilin is a key actin-binding protein that is critical for controlling the assembly of actin within the cell. Here, we present the results of molecular docking and dynamics studies using a muscle actin filament and human cofilin I. Guided by extensive mutagenesis results and other biophysical and structural studies, we arrive at a model for cofilin bound to the actin filament. This predicted structure agrees very well with electron microscopy results for cofilin-decorated filaments, provides molecular insight into how the known F- and G-actin sites on cofilin interact with the filament, and also suggests new interaction sites that may play a role in cofilin binding. The resulting atomic-scale model also helps us understand the molecular function and regulation of cofilin and provides testable data for future experimental and simulation work.     

The carboxy terminus of AFAP-110 modulates direct interactions with actin filaments and regulates its ability to alter actin filament integrity and induce lamellipodia formation

The actin filament-associated protein AFAP-110 is an SH2/SH3 binding partner for Src. AFAP-110 contains several protein-binding motifs in its amino terminus and has been hypothesized to function as an adaptor molecule that could link signaling proteins to actin filaments. Recent studies using deletional mutagenesis demonstrated that AFAP-110 can alter actin filament integrity in SV40 transformed Cos-1 cells. Thus, AFAP-110 may be positioned to modulate the effects of Src upon actin filaments. In this report, we sought to determine whether (a) AFAP-110 could interact with actin filaments directly and (b) deletion mutants could affect actin filament integrity and cell shape in untransformed fibroblast cells. The data demonstrate that the carboxy terminus of AFAP-110 is both necessary and sufficient for actin filament association, in vivo and in vitro. Analysis of the carboxy terminus revealed a mean 40% similarity with other known actin-binding motifs, indicating a mechanism for binding to actin filaments. AFAP-110 can also induce lamellipodia formation. Contiguous with the alpha-helical, actin-binding motif is an alpha-helical, leucine zipper motif. Deletion of the leucine zipper motif (AFAP(Deltalzip)) followed by cellular expression enabled AFAP(Deltalzip) to alter actin filament integrity and cell shape in untransformed cells as evidenced by the induction of lamellipodia formation. We hypothesize that AFAP-110 may be an important signaling protein that can directly modulate changes in actin filament integrity and induce lamellipodia formation.     

Action of cytochalasin D on cytoskeletal networks

Extraction of SC-1 cells (African green monkey kidney) with the detergent Triton X-100 in combination with stereo high-voltage electron microscopy of whole mount preparations has been used as an approach to determine the mode of action of cytochalasin D on cells. The cytoskeleton of extracted BSC-1 cells consists of substrate-associated filament bundles (stress fibers) and a highly cross-linked network of four major filament types extending throughout the cell body; 10-nm filaments, actin microfilaments, microtubules, and 2- to 3-nm filaments. Actin filaments and 2- to 3-nm filaments form numerous end- to-side contacts with other cytoskeletal filaments. Cytochalasin D treatment severely disrupts network organization, increases the number of actin filament ends, and leads to the formation of filamentous aggregates or foci composed mainly of actin filaments. Metabolic inhibitors prevent filament redistribution, foci formation, and cell arborization, but not disorganization of the three-dimensional filament network. In cells first extracted and then treated with cytochalasin D, network organization is disrupted, and the number of free filament ends is increased. Supernates of preparations treated in this way contain both short actin filaments and network fragments (i.e., actin filaments in end-to-side contact with other actin filaments). It is proposed that the dramatic effects of cytochalasin D on cells result from both a direct interaction of the drug with the actin filament component of cytoskeletal networks and a secondary cellular response. The former leads to an immediate disruption of the ordered cytoskeletal network that appears to involve breaking of actin filaments, rather than inhibition of actin filament-filament interactions (i.e., disruption of end-to-side contacts). The latter engages network fragments in an energy-dependent (contractile) event that leads to the formation of filament foci.     

Differences in the organization of actin in the growth cones compared with the neurites of cultured neurons from chick embryos

Sensory neurons from chick embryos were cultured on substrata that support neurite growth, and were fixed and prepared for both cytochemical localization of actin and electron microscopic observation of actin filaments in whole-mounted specimens. Samples of cells were treated with the detergent Triton X-100 before, during, or after fixation with glutaraldehyde to determine the organization of actin in simpler preparations of extracted cytoskeletons. Antibodies to actin and a fluorescent derivative of phallacidin bound strongly to the leading margins of growth cones, but in neurites the binding of these markers for actin was very weak. This was true in all cases of Triton X- 100 treatment, even when cells were extracted for 4 min before fixation. In whole-mounted cytoskeletons there were bundles and networks of 6-7-nm filaments in leading edges of growth cones but very few 6-7-n filaments were present among the microtubules and neurofilaments in the cytoskeletons of neurites. These filaments, which are prominent in growth cones, were identified as actin because they were stabilized against detergent extraction by the presence of phallacidin or the heavy meromyosin and S1 fragments of myosin. In addition, heavy meromyosin and S1 decorated these filaments as expected for binding to F-actin. Microtubules extended into growth cone margins and terminated within the network of actin filaments and bundles. Interactions between microtubule ends and these actin filaments may account for the frequently observed alignment of microtubules with filopodia at the growth cone margins.     

Fine structure of gels prepared from an actin-binding protein and actin: comparison to cytoplasmic extracts and cortical cytoplasm in amoeboid cells of cortical cytoplasm in amoeboid cells of Dictyostelium discoideum

We have identified the three-dimensional ultrastructure of actin gels that are formed in well-characterized cell extracts and mixtures of purified actin and the 120K actin-binding protein and compared these to the ultrastructure of the cytoplasmic matrix in regions of nonextracted Dictyostelium amoebae that are rich in actin and 120K. This ultrastructural characterization was achieved by using critical-point-dried whole-mount preparations. All three preparations--gelled extracts, purified proteins, and cortical cytoplasm--are composed of filament networks. The basic morphological feature of these networks is the presence of contacts between convergent filaments resulting in "T" or "X" shaped contacts. The finding that actin-containing gels are composed of filament networks, where the primary interaction occurs between convergent filaments, reconciles the known requirement of F actin for gelation with the amorphous appearance of these gels in thin sections. Increasing the molar ratio of 120K dimer to actin monomer increases the number of contacts between filaments per unit volume and decreases the lengths of filaments between contacts. This indicates that 120K stabilizes interactions between filaments and is consistent with biochemical evidence that 120K crosslinks actin filaments. The cortical network in situ resembles more closely networks formed in 120K-rich extracts than networks assembled in mixtures of purified 120K and actin. The heterogeneity of filament diameters and variation of network density are properties shared by extracts and the cytomatrix in situ while networks found in purified 120K-actin gels have filament diameters and densities that are more uniform. These differences are certainly due to the more complex composition of cell extracts and cortical cytoplasm as compared to that of purified 120K-actin gels.     

Actin filament organization in the fish keratocyte lamellipodium

From recent studies of locomoting fish keratocytes it was proposed that the dynamic turnover of actin filaments takes place by a nucleation- release mechanism, which predicts the existence of short (less than 0.5 microns) filaments throughout the lamellipodium (Theriot, J. A., and T. J. Mitchison. 1991. Nature (Lond.). 352:126-131). We have tested this model by investigating the structure of whole mount keratocyte cytoskeletons in the electron microscope and phalloidin-labeled cells, after various fixations, in the light microscope. Micrographs of negatively stained keratocyte cytoskeletons produced by Triton extraction showed that the actin filaments of the lamellipodium are organized to a first approximation in a two-dimensional orthogonal network with the filaments subtending an angle of around 45 degrees to the cell front. Actin filament fringes grown onto the front edge of keratocyte cytoskeletons by the addition of exogenous actin showed a uniform polarity when decorated with myosin subfragment-1, consistent with the fast growing ends of the actin filaments abutting the anterior edge. A steady drop in filament density was observed from the mid- region of the lamellipodium to the perinuclear zone and in images of the more posterior regions of lower filament density many of the actin filaments could be seen to be at least several microns in length. Quantitative analysis of the intensity distribution of fluorescent phalloidin staining across the lamellipodium revealed that the gradient of filament density as well as the absolute content of F-actin was dependent on the fixation method. In cells first fixed and then extracted with Triton, a steep gradient of phalloidin staining was observed from the front to the rear of the lamellipodium. With the protocol required to obtain the electron microscope images, namely Triton extraction followed by fixation, phalloidin staining was, significantly and preferentially reduced in the anterior part of the lamellipodium. This resulted in a lower gradient of filament density, consistent with that seen in the electron microscope, and indicated a loss of around 45% of the filamentous actin during Triton extraction. We conclude, first that the filament organization and length distribution does not support a nucleation release model, but is more consistent with a treadmilling-type mechanism of locomotion featuring actin filaments of graded length. Second, we suggest that two layers of filaments make up the lamellipodium; a lower, stabilized layer associated with the ventral membrane and an upper layer associated with the dorsal membrane that is composed of filaments of a shorter range of lengths than the lower layer and which is mainly lost in Triton.