Molecular Biology of the Cell - Volume 2 - Alberts B., Bray D., Lewis J., Raff M., Roberts K., Watson J. 1993

Cytoskeleton
Actin filaments and the cell cortex

In many Eukaryotic Cells, Actin is present in large quantities, often accounting for 5% or more of the total cellular protein. Although distributed throughout the Cytoplasm, most animal cells feature a particularly dense network of actin filaments and associated Proteins immediately beneath Cell/30.html">The Plasma Membrane. This network forms The Cell cortex, which imparts mechanical strength to the cell's surface layer and enables it to change shape and move. The Structure of the cortex can vary among different cells, and even within different Regions of the same cell. Sometimes it is a dense, three-dimensional network of cross-linked actin filaments that excludes Organelles and other large particles from the adjacent layers of cytoplasm (Fig. 11-29); in other cases, the cortex is markedly thinner and more akin to a two-dimensional structure. In certain regions of animal cells, small bundles of actin filaments extending from the outer face of the cortex fill surface protrusions, whereas in others, actin filaments draw the membrane inward. The plasma membrane is so intimately associated with the cortical actin layer that for some purposes it is best considered a single functional entity.

In most animal cells, approximately half of all actin molecules exist in an unpolymerized form—as free monomers or small complexes with other proteins. A dynamic equilibrium exists between this actin pool and the actin filaments, facilitating cell surface movements. In this section, we discuss how actin-binding proteins regulate the assembly of actin filaments, bundle or cross-link them into networks, and determine their arrangement, length, and other properties.

11.2.1. Actin-Binding Proteins Cross-Link Actin Filaments into Extensive Networks [18]

Actin filaments are frequently linked into rigid three-dimensional networks by specialized cross-linking proteins. The most common of these is filamin; its long, flexible molecule consists of two identical polypeptide chains joined HEAD-to-head, with actin filament-binding sites located at the tail ends (Fig. 11-30). In many cells, proteins of this type constitute nearly 1% of the total protein (roughly 1 dimer per 50 actin monomers).

Class="center">

Fig. 11-29. Electron micrograph of a thin section of a leukocyte showing the actin cortex. Although various types of granules fill the cytoplasm, they are excluded from the thin layer directly beneath the plasma membrane (the cortex). The cortex contains a network of actin filaments and associated proteins that drive cell surface movements. (Courtesy of Dorothy Bainton.)

Fig. 11-30. By forming flexible cross-links between adjacent actin filaments, filamin weaves them into a three-dimensional network with gel-like mechanical properties. Each filamin dimer, when fully extended, is about 160 nm long.

Gels formed in vitro by actin filaments and cross-linking proteins exhibit fascinating mechanical properties: they maintain their shape when subjected to a sudden, brief stress, yet readily deform under a weaker, sustained force. The putative molecular mechanism underlying this behavior—which is also characteristic of the cortical cytoplasm—is illustrated in Fig. 11-31. Such actin networks presumably enable the cell to rapidly recover its shape through elasticity following minor jolts, while still undergoing significant deformations under the prolonged action of weak forces.

Fig. 11-31. Mechanical properties of an actin filament gel formed by actin-cross-linking proteins. The gel resists abrupt deformations (A) because the cross-linking proteins lack sufficient time to dissociate from the actin filaments. Resistance to slow deformations is much weaker, as cross-linking proteins have adequate time to dissociate and rebind to actin filaments in new positions. (From M. Sato, W.H. Schwartz, T.D. Pollard, Nature 325: 828–830, 1987.)

11.2.2. Ca²⁺-Activated Gelsolin Induces Actin Filament Fragmentation [19]

Extracts derived from many Types of Animal cells form a gel when supplemented with ATP and warmed to 37 °C. This process involves the interaction of actin filaments and a cross-linking protein such as filamin, but The behavior of such a gel is more complex than that of a simple mixture of filamin and actin filaments. For instance, when the Ca²⁺ concentration is raised above 10^-7 M, the actin gel begins to liquefy. Under the microscope, regions undergoing this gel-to-sol transition display vigorous local streaming. Clearly, extracts must contain additional components besides actin filaments and filamin through which Ca²⁺ ions trigger the gel-sol transformation and fluid movement. These components are likely responsible for the cytoplasmic streaming observed in certain large cells, where it is essential for the uniform distribution of metabolites and other substances. These Intracellular Movements appear to be coupled to rapid, localized changes in cytoplasmic consistency—namely, gel/sol transitions.

Several proteins capable of liquefying actin gels in the presence of Ca²⁺ have been isolated from cell extracts. Among them, gelsolin is the best characterized—a compact protein with a molecular mass of approximately 90,000. Upon binding Ca²⁺ ions, gelsolin is activated; it severs actin filaments and caps the newly exposed plus ends, leading to The breakdown of the cross-linked actin network. Similar proteins are found in the cortex of many vertebrate cells. These severing proteins are activated at Ca²⁺ concentrations (around 10^-6 M) that are achieved in the Cytosol only transiently; they are thought to mediate cellular cortex responses to external signals. For example, when a phagocytic leukocyte contacts a microorganism, the actin filament network in that region of the cortex disassembles, allowing the surface cytoplasm to engulf and internalize the microbial cell. We will return to the mechanisms underlying such movements shortly.

11.2.3. Cytoplasmic Streaming Can Be Driven by Myosin [20]

Although artificial mixtures of actin filaments, filamin, and gelsolin exhibit Ca²⁺-dependent gel-sol transitions, these mixtures are incapable of contraction and do not generate the streaming movements seen in actin-rich gels derived from cells. Myosin appears to be essential for this process: crude actin extracts from which myosin has been selectively depleted lose both contractility and streaming capability. It is therefore inferred that the force for these movements stems from some interaction between Actin and myosin.

Because actin filaments in the initial extract presumably form a disordered three-dimensional network, the question arises: how can actin and myosin produce coordinated movement? Actin filaments, as we know, possess distinct polarity, and myosin heads can bind to and slide along them only when properly oriented with respect to this polarity. Small bipolar complexes of nonmuscle myosin molecules (see Fig. 11-26) could, to some extent, “order” actin filaments in solution simply by pulling one group relative to another, even if the myosin complexes and actin filaments were initially poorly aligned (Fig. 11-32).

Fig. 11-32. Bipolar aggregates of nonmuscle myosin molecules (see Fig. 11-26) drive the sliding of two oppositely polarized actin filaments (as in Muscle). Through this mechanism, myosin can generate contraction even within a randomly oriented network of actin filaments.

11.2.4. Cytoplasmic Streaming in Giant Algal Cells Is Mediated by Actin and Myosin [21]

Plant cells can grow much larger than animal cells because they possess a rigid Cell wall and contain a vast central vacuole (see Chapter 20, Section 20.4.9). Diffusion is insufficiently efficient over such distances, meaning that very large cells require robust cytoplasmic streaming to mix their cytoplasm. The giant multinucleate green Algae *Chara* and *Nitella* reach lengths of 2–5 cm, and it would take weeks for a protein molecule to diffuse from one end of the cell to the other. Unsurprisingly, these organisms provide some of the most striking Examples of cytoplasmic streaming. A continuous, narrow stream of cytoplasm moves along the cell in an endless helical ribbon, circumventing the massive central vacuole (Fig. 11-33A). This stream carries along internal membrane structures, Mitochondria, and nuclei at speeds of up to 75 µm/s.

The system generating streaming in these cells is located between the moving cytoplasmic layer and a stationary monolayer of METABOLISM/14.html">Chloroplasts situated directly beneath the plasma membrane (Fig. 11-33B). Cytology/cytology/93.html">ELECTRON MICROGRAPHS OF this transitional region reveal large bundles of actin filaments of uniform polarity. This polarity is such that the directed movement of myosin molecules along them (from the minus end to the plus end) coincides with the direction of the observed cytoplasmic currents. Furthermore, a giant *Nitella* cell can be "opened up" to expose the actin bundles On the surface of the chloroplast layer; if latex beads coated with myosin are then added to the preparation, these beads move along the actin network in the presence of ATP, with a motion remarkably similar to the translocation of organelles in an intact cell. Thus, cytoplasmic streaming is very likely driven by myosin-based mechanisms.

For this type of cytoplasmic streaming, as seen in *Nitella*, bipolar myosin filaments are not required—since the process involves continuous unidirectional movement rather than contraction. In addition to filamentous nonmuscle myosin (see Fig. 11-26), many cells contain relatively small molecules known as mini-myosins. These molecules were first isolated from large amoebae (*Acanthamoeba*); they consist of a single globular head attached to a flexible tail that, instead of forming a coiled-coil with another molecule's tail, binds to membranes—either directly or via another protein. Purified mini-myosins associate with membrane-bound organelles in vitro and can transport them along oriented bundles of actin filaments. This mechanism appears to account for The Active Transport of many organelles within cells.

Fig. 11-33. Diagram of cytoplasmic streaming in the giant algal cell *Nitella*. A. The pathway of cytoplasmic flow within the cylindrical cell. For clarity, the cell diameter is exaggerated relative to its length. B. Longitudinal section through a region of such cell, illustrating the Organization of stationary and moving cytoplasmic layers. The stationary cortical layer contains chloroplasts associated with underlying actin filament bundles; beneath these bundles lies the moving cytoplasmic layer, which houses nuclei, mitochondria, and other organelles. In reality, the relative dimensions of the vacuole are much larger than depicted.

11.2.5. Cortex Organization Is Determined by the Interaction of Actin Filaments with the Plasma Membrane

Up to this point, we have sidestepped a question central to understanding the Structure and function of the actin cortex: What is The Nature of the linkage between actin filaments and the plasma membrane? It is believed that specialized Membrane Proteins serve as organizing centers for the actin network. The forces generated within the cortical layer of actin filaments, which drive cell surface movements, must be transmitted to the membrane through these or other membrane proteins. Little is known about the identities of these proteins or how they interact with actin. It is evident, however, that there are at least three functional types of actin attachment to the plasma membrane: the first primarily imparts mechanical strength to the membrane and determines its shape; the second enables actin filaments to pull portions of the membrane inward; and finally, the third type involves actin filaments driving rapid outward protrusions of the membrane. Let us examine each of these types in turn.

Fig. 11-34. This hypothetical scheme illustrates how the Cytoskeleton underlying the mammalian erythrocyte membrane might derive from a more conventional actin-dominated cortex. A nucleated precursor cell possesses a membrane-associated cortex consisting of actin filaments held together by spectrin tetramers and other cross-linking proteins. As this cell matures into an erythrocyte, most of its actin is lost. Consequently, the actin filaments depolymerize and the cross-linking proteins largely dissociate from them, leaving behind a thin, two-dimensional network of spectrin and short actin filament "segments" linked to the plasma membrane.

11.2.6. Membrane-Associated Cytoskeletal Networks Provide a Mechanical "Framework" for the Plasma Membrane [22]

In Chapter 6 (Section 6.2.4), we already discussed spectrin and ankyrin—proteins first discovered in mammalian erythrocytes as essential Components of the membrane-associated cytoskeleton. In mature mammalian erythrocytes, The Nucleus and internal membranes are lost, leaving the plasma membrane as the sole remaining membrane. It is supported by a two-dimensional network of spectrin tetramers whose ends are joined by very short actin filaments; these tetramers are linked via ankyrin bridges to the transmembrane protein Band 3 (see Fig. 6-26). Proteins highly homologous to both spectrin and ankyrin are found in the cortex of many vertebrate cells. It has been suggested that the remarkably thin (only about 20 nm thick) cytoskeletal network underlying the plasma membrane of a mature erythrocyte forms from the thicker cortex of a nucleated precursor cell through the gradual depolymerization of actin filaments (Fig. 11-34). In this view, the Spatial Organization of cortical proteins in the erythrocyte can serve as a model for the actin-based cytoskeletal networks that support the plasma membrane in all other animal cells.

11-10

11.2.7. Microvilli Provide an Example of How Bundles of Cross-Linked Actin Filaments Can Stabilize Local Protrusions of the Plasma Membrane [23]

Microvilli are finger-like projections of the plasma membrane on The surface of many animal cells. They are particularly abundant in epithelial cells that require a large surface area for efficient physiological function. For example, in the human Small Intestine, a single such cell bears several thousand microvilli on its apical surface. Each microvillus is roughly 1 µm long and 0.08 µm wide, increasing the absorptive surface area of the cell by a factor of 20 compared to a smooth surface. The highly specialized plasma membrane of microvilli is coated externally with a dense layer of Polysaccharides and digestive Enzymes.

Inside each intestinal microvillus lies a rigid bundle of 20–30 parallel actin filaments extending from its tip to the base, where they insert into the cell cortex. All filaments in the bundle are oriented with their plus-ends pointing outward (away from the cell) and are held at uniform intervals from one another by several actin-binding proteins, notably fimbrin and fascin (Fig. 11-35). Unlike filamin and other actin-cross-linking proteins, whose flexible molecules organize filaments into a loose network, these actin-binding proteins are relatively small and compact, with the polypeptide chain forming two distinct actin-binding sites. As a result, the actin filaments are bundled into tight arrays where they are held firmly in parallel at a spacing of approximately 10 nm.

Fig. 11-35. The core of a microvillus is formed by a bundle of parallel actin filaments held together by actin-cross-linking proteins. The periphery of this bundle is connected to the surrounding plasma membrane by lateral "bridges". The plus-ends of all actin filaments are located at the tip of the microvillus, where they are embedded in an amorphous, intensely staining material of unknown composition.

Fig. 11-36. An intestinal epithelial cell showing the terminal web underlying the apical plasma membrane (electron micrograph, quick-freeze deep-etch technique). Bundles of actin filaments that form the microvillar core extend down into the terminal web, where they are interconnected primarily by spectrin. A layer of Intermediate filaments lies beneath the terminal web. (N. Hirokawa, J. E. Heuser, J. Cell Biol. 91: 399-409, 1981. With permission from the Rockefeller University Press.)

The lower end of the actin bundle in a microvillus is anchored in a specialized cortex at the apical pole of the cell. This cortex, known as the terminal web, contains a dense meshwork of spectrin molecules overlying a layer of intermediate filaments (Fig. 11-36); the terminal web appears to anchor the actin bundles—and thus the microvilli—perpendicular to the apical surface of the cell.

How is the actin filament bundle of a microvillus attached to the surrounding plasma membrane? Electron Microscopy reveals two distinct types of linkage: side-arms that contact the membrane at regular intervals along the length of the bundle, much like the steps of a spiral staircase, and a cap of amorphous, densely staining material at the tip of the microvillus that connects the plus ends of the actin filaments to the membrane (see Fig. 11-35).

If the plasma membrane of an intestinal epithelial cell is solubilized with a non-ionic detergent, the side-arms of the actin bundles remain attached to the exposed cytoskeleton. However, they can be released by adding ATP to the preparation, which reveals that each side-arm consists of a mini-myosin molecule firmly bound to the calcium-binding protein calmodulin. Through its ATP-dependent head domain, mini-myosin binds to the actin filaments of the microvillar core, while its short tail domain attaches to the plasma membrane. It remains unclear why a motor protein is used to form this linkage; however, because myosin molecules move along actin filaments toward the plus end, one would expect them to pull membrane components toward the tip of the microvillus. This mechanism may facilitate the continuous shedding of plasma membrane from the microvillar tip into the intestinal lumen, where membrane-bound digestive enzymes continue to function.

Even less is understood about the Nature of the amorphous cap at the tip of the microvillus. As we shall see, there is reason to suspect that proteins localized in this region regulate the thickness and length of the actin filament bundle, thereby determining the dimensions of the microvillus (Section 11.6.10). If so, the interaction between the actin filaments and the plasma membrane at the microvillar tip must be remarkably complex.

11.2.8. Through focal contacts, actin filaments can exert a pulling force on the substrate [24]

Bundles of actin filaments frequently attach to the plasma membrane in a manner that enables them to transmit pulling forces to the substrate, whether it is the Extracellular matrix or another cell. We have already mentioned such modes of attachment for actin filament bundles in fibroblasts and smooth muscle cells (Sections 11.14 and 11.17). Transmembrane linker Glycoproteins participate in joining the end of the bundle to the plasma membrane. The attachment sites of cultured fibroblasts to the extracellular matrix have been studied most thoroughly. When fibroblasts grow in a culture dish, the major part of their substrate-facing surface is separated from it by a gap of more than 50 nm. However, in certain regions known as focal contacts or adhesion plaques, this gap narrows to 10–15 nm. In reflection Interference microscopy, which collects only light reflected from the lower surface of the cell, these regions appear as dark patches. By staining the cell with Antibodies against actin, it can be demonstrated that the ends of stress fibers terminate at the plasma membrane precisely at these sites (Fig. 11-37). At a focal contact, the cell cortex is linked via transmembrane linker proteins to components of the extracellular matrix (specifically Fibronectin; see Section 14.2.13) adsorbed onto the dish. A typical transmembrane linker is the fibronectin receptor, a glycoprotein composed of two chains belonging to the integrin family (Section 14.2.17). Its extracellular domain binds to fibronectin, while its cytoplasmic domain is connected to the actin filaments within stress fibers (Fig. 11-38, A). This connection is indirect, involving at least four additional attachment proteins, including talin and vinculin. Talin binds to both the cytoplasmic domain of the fibronectin receptor and to vinculin. Neither of these proteins, however, is directly connected to the actin filaments; instead, direct attachment appears to be mediated by other proteins: an actin-capping protein that binds to their plus ends, and a-actinin, which in muscle is attached to actin filaments near the plus end in the Z-disc region (Section 11.1.13). A putative arrangement of all these proteins within a focal contact is shown in Fig. 11-38, B.

Fig. 11-37. Relationship between focal contacts and stress fibers in a fibroblast in vitro. Focal contacts are best visualized in a living cell using reflection interference microscopy (A). Here, light is reflected from the lower surface of the cell attached to the coverslip, and focal contacts appear as dark spots. In photo B, the same cell is stained (after fixation) with antibodies against actin; it can be seen that the majority of its actin filament bundles (or stress fibers) terminate at or in the immediate vicinity of focal contacts. (Courtesy of Grenham Ireland.)

Fig. 11-38. Model illustrating how transmembrane linker glycoproteins of the plasma membrane can connect intracellular actin filaments to the extracellular matrix at focal contacts. Focal contacts are formed when the binding of matrix glycoproteins to the cell surface leads to the clustering of transmembrane linkers in the contact region (A). On the right (B), the putative arrangement of intracellular attachment proteins mediating The Link Between transmembrane linker glycoproteins (such as the fibronectin receptor) and actin filaments is shown.

Very similar to focal contacts are the sites of actin filament attachment to the plasma membrane in smooth muscle cells (Section 11.1.14). Another related (though less so) structure is the zonula adherens (adhesion belt), which joins epithelial cells into sheets and enables contractile bundles of actin filaments to interact across two adjacent Plasma Membranes (Section 14.1.3). These Intercellular junctions contain vinculin and a-actinin, but lack talin, meaning that the mode of attachment of actin filaments to the plasma membrane must differ somewhat from that in focal contacts.

So far, we have discussed examples of actin filament attachment to the membrane where the filaments either play a structural role (such as in microvilli) or are capable of exerting a pulling force on the membrane (as in focal adhesions). In both cases, we are dealing with permanent structures—products of a relatively stable association of actin filaments. Rapid changes in cell surface Morphology (for example, during cell locomotion), however, frequently rely on short-term, regulated actin polymerization. But before examining The Mechanism of rapid plasma membrane protrusion driven by actin polymerization, we will make a brief digression to review what is known about these polymerization processes in vitro.

Fig. 11-39. Kinetics of in vitro actin polymerization. Polymerization is initiated by increasing the Ionic strength of the actin solution and typically begins after a short lag phase.

11-11

11.2.9. Actin Filament Growth Occurs Primarily at the Plus End [25]

As already mentioned, in vitro actin monomers can spontaneously polymerize in the presence of ATP to form filaments. This process is typically triggered by increasing the ionic strength of the actin solution, and it is most easily monitored by measuring the increase in solution light scattering or the fluorescence of a covalently bound dye. Actin polymerization proceeds non-linearly: it begins with a lag phase that reflects the initial stage—the slowest and most unfavorable step—in which three actin monomers must assemble into a specific geometric configuration (Fig. 11-39). This stage is known as nucleation. Once nucleation has occurred, the subsequent addition of new actin monomers to the filament ends proceeds rapidly. Filament assembly is reversible, and the monomer concentration eventually drops to a level where the rates of monomer association and dissociation become equal. This concentration is termed the critical concentration. If an actin solution is rapidly diluted so that the monomer concentration falls below the critical concentration, actin filaments begin to break down, and depolymerization continues until the critical monomer concentration is restored. As we will see, both nucleation and actin polymerization within the cell are governed by various actin-binding proteins, which thereby determine the spatial distribution and length of actin filaments. Actin polymerization exhibits another key feature that helps explain the directionality of filament growth during various cellular movements. As we are well aware, actin filaments are polarized, and an important consequence of this polarity is the distinct polymerization kinetics at the plus and minus ends. This difference can be revealed by decorating a short segment of an actin filament with myosin heads and then adding actin monomers under polymerization-favorable conditions. When the growing actin filaments are fixed after a brief interval and observed under an Electron microscope, their plus ends are seen to have grown much more extensively than their minus ends (Fig. 11-40). By conducting a series of such experiments, one can measure the growth rates of both filament ends at various actin monomer concentrations. Under ionic conditions mimicking the intracellular environment, purified actin filaments elongate at their plus ends 5 to 10 times faster than at their minus ends. Evidently, in vivo growth almost exclusively occurs at the plus end; consequently, by anchoring the plus end of a filament in a specific orientation, a cell can dictate both the rate and direction of filament elongation.

Fig. 11-40. Asymmetric growth of an actin filament. When short fragments of actin filaments decorated with myosin heads are used as "seeds" for actin polymerization, assembly is observed to occur much more rapidly at the plus end of the initial fragment than at the minus end. (Courtesy of M.S. Runge and T.D. Pollard.)

11-12

11.2.10. Actin Filaments Undergo Subunit Treadmilling [25]

In vitro experiments (such as the one shown in Fig. 11-40) demonstrate that the actin concentration at which filament growth ceases—that is, the critical concentration—differs for the two ends of the filament. Therefore, when actin polymerizes in a test tube, a steady state is reached in which monomer addition occurs predominantly at the plus end, while monomer loss occurs at the minus end. The rates of these two processes at steady state are equal, and thus the concentration of free monomers remains constant. Although the total polymer length does not change, individual actin molecules continuously migrate from one end of the filament to the other (Fig. 11-41). This process, known as treadmilling, can be likened to people moving forward in a queue.

Fig. 11-41. Growth rates of actin filaments at opposite ends as a function of actin monomer concentration. The graph demonstrates that the two ends grow at different rates and have distinct critical concentrations. Consequently, There is a range of free actin concentrations (shaded region) where the plus end of the actin filament polymerizes while the minus end depolymerizes. When free actin is in equilibrium with actin filaments in vitro, no net growth occurs because The rate of subunit dissociation from the minus end equals the rate of association at the plus end. At this free actin subunit concentration (intermediate between the critical concentrations for the two ends, as indicated by the colored arrow), the filament length remains constant, but the actin molecules within it continuously move from one end of the polymer to the other. This process is called treadmilling (see Scheme 11-1).

Fig. 11-42. Lamellipodia and microspikes at the leading edge of a human fibroblast migrating in culture (scanning electron micrograph). The arrow indicates the direction of movement. As the cell advances, lamellipodia and microspikes retreat along its dorsal surface, producing a "ruffling" appearance. (Courtesy of Julian Heath.)

Treadmilling requires an input of energy; otherwise, it could be used to build a perpetual motion machine—a device capable of performing work without consuming energy, thereby violating the Laws of Thermodynamics. In the case of treadmilling, energy is supplied by ATP Hydrolysis: each actin monomer binds an ATP molecule, which it rapidly hydrolyzes to ADP and inorganic phosphate shortly after incorporating into the filament. Scheme 11-1 explains how ATP hydrolysis drives treadmilling. As we will soon see (Section 11.2.11), treadmilling may serve as one of the mechanisms by which actin filaments and their associated components perform mechanical work within the cell.

11.2.11. Many Cells Form Motile, Actin-Containing Structures on Their Surface: Microspikes and Lamellipodia [26]

Dynamic cell surface protrusions containing actin filaments are a common feature of animal cells, particularly those actively migrating or changing their shape. Cultured cells, for instance, frequently generate numerous thin, rigid projections roughly 0.1 µm thick and 5–10 µm long, termed microspikes, which contain loose bundles of approximately 20 actin filaments oriented with their plus ends pointing outward. The growing tip of a nerve cell axon—the growth cone—extends even longer microspikes known as filopodia, which can reach lengths of up to 50 µm (Section 19.7.7). These surface protrusions are highly dynamic, appearing and disappearing with great rapidity. They appear to act like tentacles with which the cell probes its environment; microspikes that firmly attach to a substrate guide the moving cell toward that more adhesive region, whereas those that fail to anchor are swept backward along the DORSAL SIDE OF the cell and retracted.

In addition to microspikes, crawling cells and growth cones periodically extend thin sheet-like projections from their actively advancing ("leading") edge, known as lamellipodia. Like microspikes, some lamellipodia successfully attach to the substrate.

GROWTH RATES DIFFER AT THE PLUS AND MINUS ENDS

Assembly (polymerization) and disassembly (depolymerization) of actin filaments occur through the addition and loss of monomers at the filament ends. During assembly, a filament grows faster at one end than at the other; the rapidly growing end is designated as the plus end, and the slowly growing end as the minus end. The difference in growth rates between opposite ends arises because each subunit undergoes a conformational change upon incorporating into the polymer.

This conformational change preferentially affects subunit addition at one of the ends.

CRITICAL SUBUNIT CONCENTRATION

The number of subunits joining a polymer per second is proportional to the concentration of free subunits (k_acc [C]); at the same time, subunits dissociate from the polymer ends at a constant rate (k_дисс) independent of [C]. As the polymer grows, free subunits are consumed and [C] drops until it reaches a constant value—the critical concentration (C_c)—at which the rate of subunit addition to the polymer equals the rate of their dissociation from it. At this equilibrium state

Although The values of k_асс and k_дисс for the plus and minus ends will differ, their ratio k_дисс/k_асс, and hence C_c, must be identical at both ends. This is because the dissociation of a subunit from either end of the polymer breaks identical bonds, and the final states of the subunits are likewise identical. Thus, The change in Free energy ΔG upon subunit dissociation—which determines the Equilibrium Constant for subunit association with an end (see Table 3-3)—is the same at both ends; if the plus end grows four times faster than the minus end, it must also shorten four times faster.

Thus, when [C] < C_c, both ends grow;

when [C] > C_c, both ends shorten (see Fig. 11-41).

ATP HYDROLYSIS ALTERS THE EQUILIBRIUM CONSTANT AT EACH END

Each actin molecule carries a tightly bound ATP molecule. Soon after the conformational change of an actin subunit upon its addition to the polymer, this ATP is hydrolyzed to ADP, which remains tightly bound to the actin.

at either end. Typically, the form attaches to the filament, whereas the form leaves it (which then converts into in solution), so The process of subunit detachment from the polymer is not simply the reverse of the addition reaction.

Let us consider the events occurring at the plus end:

As before, the polymer will grow until [C] drops to C_c. For clarity, we can assume the values of k_асс and k_дисс are small; polymer growth will cease at This is not a true equilibrium, but a steady state, since the hydrolyzed ATP must be replenished by nucleotide exchange on the free subunit

TREADMILLING

Since k_acc and k_дисс refer to different reactions, their ratio k_дисс/k_асс is not necessarily the same at different ends of the polymer. Indeed, ATP hydrolysis has been shown to create distinct critical concentrations for the opposite ends, where

C_c (for minus end) > C_c (for plus end).

If both ends of the polymer are "open," polymerization will predominate until [C] reaches a level above C_c for the plus end and below C_c for the minus end (see Fig. 11-14). In this steady state, subunits will assemble into the polymer at the plus end and disassemble from it at the minus end, with the rates of these processes being equal. Consequently, the polymer length remains unchanged despite the directional flow of subunits through the polymer—a process known as treadmilling:

ATP hydrolysis is not required for filament assembly. Under appropriate conditions, actin monomers carrying a non-hydrolyzable ATP analog also polymerize. However, the resulting filaments are incapable of treadmilling—this phenomenon relies entirely on the ATP hydrolysis that accompanies polymerization. ATP hydrolysis decreases the affinity of a subunit for its neighbors, thereby increasing its probability of dissociation. Meanwhile, other subunits "slip" and likewise move backward in a wave-like manner along the upper surface of the cell, a process termed "ruffling" (Fig. 11-42).

Fig. 11-43. Actin filaments at the leading edge of a cultured fibroblast. A. Electron micrograph of the whole leading edge of a cell extracted with a non-ionic detergent to remove the plasma membrane and the bulk of soluble proteins. Note the oriented network of actin filaments in the lamellipodium in which the microspike is embedded. B. Diagram of the arrangement of lamellipodial actin filaments. (A—J.V. Small, J. Cell Biol. 91: 695-705, 1981. Reproduced by permission of the Rockefeller University Press.)

The lamellipodium can be viewed as a two-dimensional variant of a microspike; its leading edge frequently features a row of short microspikes. When a cell is gently fixed and stained for electron microscopy, the actin filaments in the lamellipodia of a migrating cell appear much more organized than those in other cortical regions; ordered bundles of filaments extend toward the leading edge, where their plus ends abut the plasma membrane (Fig. 11-43).

Studies of migrating cells in culture show that monomer treadmilling in actin filaments at the leading edge of the cell is significantly faster than in filaments formed in vitro. The movement of actin molecules within a cell can be detected using fluorescence recovery after photobleaching (FRAP) experiments. When fluorescently labeled actin molecules are microinjected into a migrating cell, they rapidly incorporate into all actin filaments. If the fluorescent label in a small spot at the thin leading edge of the cell is subsequently destroyed by a laser beam, this dark spot can be seen gradually moving inward from the cell margin at a rate of about 0.8 µm/min. This indicates rapid treadmilling of actin molecules within the lamellipodial and microspike filaments (Fig. 11-44). Evidently, actin monomers continuously add to the plus ends near the plasma membrane and dissociate from the minus ends deeper within the cell.

If the fluorescence of actin filaments is photobleached in a small region anywhere else in the cell, the fluorescence recovers within a few minutes without any visible displacement: this indicates that the organization or behavior of actin at the leading edge possesses unique features. It has been suggested that continuous actin polymerization directly beneath the plasma membrane can be harnessed to push out the leading edge, thereby driving cell locomotion (see Section 11.6.5). Investigations into the motility mechanism of certain invertebrate sperm have indeed confirmed the involvement of actin polymerization in this process.

Fig. 11-44. Actin treadmilling in cultured fibroblasts. Fluorescent actin molecules were microinjected into a cell, where they incorporated into actin filaments. A small spot on an actin filament at the cell's leading edge was "bleached" by a laser beam. The cell was then photographed at regular intervals using a fluorescence microscope equipped with a video system (Section 4.1.6). The rapid movement of the bleached spot backward demonstrates that actin molecules continuously flow along the filament away from the leading edge.

11-13

11.2.12. Explosive actin polymerization promotes The formation of the acrosomal process in the sperm of certain invertebrates [27]

Although the appearance of many cell surface protrusions is thought to depend on the polymerization of actin filaments, the mechanisms triggering this process remain poorly understood. One exception is the acrosomal reaction in the sperm of certain invertebrates (Section 15.4.1). In sea cucumber sperm, for example, the acrosomal region (in the head) is packed with unpolymerized actin bound to profilin, a small protein abundant in most animal cells. Profilin forms an equimolar complex with actin molecules, thereby blocking the nucleation step of spontaneous actin filament polymerization, though it has no significant effect on the elongation of the plus ends of pre-existing filaments. When a sperm contacts the outer coat of an egg, its cytoplasmic pH rises and actin dissociates from profilin. An explosive polymerization of actin ensues, rapidly forming a membrane-bounded, long, thin acrosomal process that punctures the egg coat like a harpoon, allowing the egg and sperm membranes to fuse (Fig. 11-45). In addition to actin polymerization, the elongation of the acrosomal process is apparently facilitated by the influx of Water into the sperm head, which generates hydrostatic pressure within it.

The Formation of the acrosomal process indicates that actin polymerization can likely drive rapid plasma membrane protrusion without compromising its continuity. Although it is unlikely that lamellipodia and microspikes are formed by Changes in the pool of actin monomers available for polymerization, their protrusion presumably occurs in much the same way—through the localized growth of actin filaments directly beneath the plasma membrane.

11.2.13. Actin assembly is controlled by the plasma membrane [28]

The cell must somehow regulate the formation of actin-rich surface protrusions so that they appear only where and when they are needed. For example, a macrophage extends pseudopodia to engulf a bacterium only at the site of contact (Section 6.5.15), and the actin polymerization that helps thrust out the acrosomal process in activated sea cucumber sperm occurs only when the sperm head touches the jelly coat on the egg surface.

Fig. 11-45. Successive stages of acrosomal process elongation in sea cucumber sperm. Light micrographs: the first was taken 2 s after artificial stimulation of the acrosomal reaction, and the remaining ones at 0.75 s intervals. The arc to the right of the sperm head is part of its curved tail entering the field of view. (L. G. Tilney, S. Inoue, J. Cell Biol. 93: 820-827, 1982. Reprinted by permission of Rockefeller University Press.)

The "time and place" for such actin-dependent responses are apparently regulated by controlling nucleation sites for actin filament growth near the plasma membrane. A case in point is the chemotactic response of neutrophils. During a localized bacterial infection, these phagocytic leukocytes cross Blood capillary walls in large numbers and crawl through Tissues toward the site of infection. They are guided to their target by small, freely diffusing molecules released in the infected area. This process is known as chemotaxis: cells "sense" a chemical gradient and respond by altering their direction of movement. Neutrophils possess surface receptor proteins that enable them to detect very low concentrations (~10^-10 M) of specific chemoattractants. Under favorable conditions, neutrophils can detect a concentration difference of a chemoattractant peptide across opposite sides of The Cell as small as 1% and appropriately reorient their leading edge to move up the gradient.

Some light has been shed on the link between receptor activation and reorientation by experiments studying the response of neutrophils to a sudden increase in chemoattractant peptide concentration. Upon stimulation, neutrophils begin extending microspikes and lamellipodia all over their surface, and the proportion of polymerized actin in the cell increases from 30% to 60% within 20 seconds. This rapid actin polymerization most likely results from an increase in the number of nucleation sites in the plasma membrane: when such chemoattractant-treated cells are lysed, the lysate exhibits an enhanced ability to nucleate actin filament growth in vitro; furthermore, each activated cell surface receptor induces the growth of a substantial number of actin filaments. We will return to The Role of controlled actin polymerization in directed cell migration later when discussing cell motility as an integrated whole-cell response (Section 11.6.3).

11.11.

11.2.14. Certain substances affect cell behavior by altering the state of actin polymerization [29]

One might assume that many of the movements produced by the cell cortex, such as phagocytosis or locomotion, depend on the dynamic equilibrium between free (non-polymeric) actin and actin filaments. However, compared to the "explosive" changes triggered by sperm activation, alterations in actin polymerization during these movements are usually too subtle and transient to be easily detected. Nevertheless, the crucial role of actin polymerization and depolymerization in such movements is highlighted by the effects of various substances that prevent changes in the state of actin, thereby disrupting its motile function. For instance, cytochalasins (Fig. 11-46)—a family of metabolites produced by various Molds—suppress many forms of vertebrate cell motility, including locomotion, phagocytosis, cytokinesis, the formation of lamellipodia and microspikes, and the rolling of epithelial sheets into tubes. At the same time, these agents do not affect chromosome segregation during mitosis, which relies primarily on spindle microtubule function, nor do they affect Muscle contraction, which involves stable actin filaments that do not undergo assembly and disassembly. The main "target" of cytochalasins is the rapidly growing plus ends of actin filaments, to which they specifically bind and block The addition of new actin monomers.

Fig. 11-46. Chemical Structure of cytochalasin B.

Phalloidin—a highly toxic alkaloid from the death cap mushroom (Amanita phalloides)—unlike cytochalasins, stabilizes actin filaments by inhibiting their depolymerization. However, it crosses the plasma membrane very poorly and must be introduced into the cell by microinjection to exert its effect. Introduced in this manner, phalloidin blocks the migration of both amoebae and cultured vertebrate cells, indicating a pivotal role for the dynamic assembly and disassembly of actin filaments in this type of motility. Because phalloidin binds specifically to actin filaments, its fluorescent derivatives are frequently used instead of actin antibodies to visualize actin filaments within cells.

11.2.15. The properties of the cell cortex depend on the balance of cooperative and competitive interactions among a diverse group of actin-binding proteins

Much more is currently known about actin-binding proteins than about proteins interacting with the other two major filament systems of the cytoskeleton: intermediate filaments and microtubules. Our knowledge of the actin-binding proteins from non-muscle vertebrate cells discussed in this chapter is summarized in Fig. 11-47. This picture is far from complete; each functional category shown here includes several members with slightly different properties. Furthermore, some actin-binding proteins have no obvious Functions, and undoubtedly there are still undiscovered actin-binding proteins waiting to be found. In particular, cells must somehow distinguish the opposite ends of actin filaments and control their orientation and localization within the cytoplasm. Presumably, this discrimination involves "capping proteins" that selectively bind to the plus ends of actin filaments, anchor them near the plasma membrane, and regulate the addition of actin monomers. By contrast, the minus ends, which are relatively inert regarding both polymerization and depolymerization, may often remain "uncapped." Much remains to be learned about the nature of membrane-associated actin-organizing centers; for example, almost nothing is known about the molecular nature of the electron-dense material at the tips of microvilli, which appears responsible for organizing the actin core of these projections and controlling their growth and regeneration (Section 11.2.7).

Fig. 11-47. Some Major Classes of actin-binding proteins found in most vertebrate cells.

Fig. 11-48. Examples of competitive and cooperative interactions between actin-binding proteins. Tropomyosin and filamin bind tightly to actin filaments but compete with each other. Because tropomyosin binds cooperatively to actin filaments, extensive regions of the actin network will be dominated by either tropomyosin or filamin. Other actin-binding proteins, such as a-actinin or myosin, can be competitively displaced from specific sites; for instance, a-actinin binds along the entire length of purified actin filaments in vitro, but binds relatively weakly to such filaments in the cell, where it is localized predominantly near the plus ends due to competition with other proteins. Conversely, cooperative interactions can enhance binding; for example, tropomyosin appears to promote myosin binding. A multitude of such interactions among the actin-binding proteins shown in Fig. 11-47 (and several others) is thought to account for the extraordinary diversity of actin structures found in all eukaryotic cells.

Even after all the components of actin networks are identified, a more difficult challenge will remain: understanding the functional consequences of their numerous interactions. The proteins illustrated in Fig. 11-47 do not bind to actin simultaneously or at random; rather, they cooperate and/or compete with one another, generating the well-ordered structural relationships so vital to the cell (Fig. 11-48). In addition, network interactions are modulated by local changes in ion concentrations and by mechanical forces that stretch or compress the network, thereby shifting its components relative to one another. Consequently, it is clear that our understanding of this key cytoskeletal compartment is still in its infancy.

Conclusion

Eukaryotic cells feature a specialized cortical layer of actin filaments located directly beneath the plasma membrane. Overall, it forms a uniform, three-dimensional, cross-linked network with gel-like properties. At the same time, cortical actin filaments form a variety of specialized structures. For example, bundles of actin filaments complexed with myosin attach to the plasma membrane, equipping the cell with contractile structures. In other regions, controlled polymerization of actin filaments at their plus ends can push the plasma membrane outward, generating motile surface protrusions. The Diversity of cortical structures and their functions depends on a broad array of actin-binding proteins that cross-link actin filaments into a loose gel, bundle them into rigid structures, move along them to generate mechanical force, or anchor them to the plasma membrane. Some proteins performing this latter function cap the plus ends of actin filaments, thereby regulating their polymerization and depolarization within the cell. These very proteins are believed to play a critical role in complex cell surface movements, such as phagocytosis or cell locomotion over a substrate.