Actin Filaments

An actin filament (also called F-actin, for filamentous actin) is built from a single kind of subunit, G-actin (globular actin), a ~42 kDa protein with a deep nucleotide-binding cleft that holds ATP or ADP plus a divalent cation (Mg²⁺ or Ca²⁺ in vivo). Thousands of G-actin monomers polymerize non-covalently into a helical filament roughly 7–9 nm wide, which can be described equivalently as a two-start helix or as a single long-pitch helix of dimers repeating every 36 nm. Because every monomer is added in the same orientation, the filament has a built-in structural asymmetry — a polarity — that is not decoration but the mechanistic key to nearly everything actin does. The two ends of a filament are functionally different. The barbed end (also called the plus end) grows quickly; the pointed end (minus end) grows slowly and, under cellular conditions, tends to shrink. These names come from the arrowhead appearance filaments acquire when decorated with myosin S1 fragments and viewed by electron microscopy — the myosin heads all point one way, giving a barbed/pointed "arrowhead" look that visually reveals polarity even though it is an artifact of decoration, not of the bare filament's shape. Polarity matters because it means monomer addition and loss are not symmetric: the cell can bias net growth toward one end, which is the mechanistic basis of directional filament elongation and of the phenomenon called treadmilling, in which a filament maintains constant length while individual subunits flow through it from barbed to pointed end. Actin is one of the most conserved and abundant proteins in eukaryotic cells (often several percent of total cell protein by mass), existing in a dynamic equilibrium between monomeric G-actin and polymeric F-actin. That equilibrium is governed by nucleotide state (ATP- versus ADP-actin), ionic conditions, and a large cast of actin-binding proteins that the cell uses to nucleate new filaments, cap existing ones, sever old ones, crosslink them into networks, or transport monomers to where they are needed. This regulatory layer is what turns a simple self-assembling rod into the adaptable machinery underlying lamellipodia, filopodia, the contractile ring, stress fibers, and the cortical meshwork beneath the plasma membrane.

Step 1 — Monomer pool and nucleotide state. In the cytosol, G-actin is bound predominantly to ATP because ATP is present at much higher concentration than ADP and actin's affinity for ATP is comparable. ATP-G-actin is the preferred substrate for polymerization. Small proteins such as profilin bind G-actin and, rather than simply sequestering it, actually promote nucleotide exchange (ADP→ATP) and channel the monomer toward barbed ends of growing filaments (profilin blocks the pointed-end-binding surface, so profilin-actin adds only at barbed ends). Thymosin-β4 binds G-actin and sequesters it, buffering the free monomer pool and keeping much of cellular actin polymerization-ready but held in reserve — critical because free G-actin concentration in cells (tens of micromolar) sits close to the critical concentrations for polymerization, so small shifts in the buffering system can trigger rapid filament assembly on demand.

Step 2 — Nucleation. Spontaneous formation of a new filament is kinetically difficult: actin dimers and trimers are unstable and fall apart faster than they grow, so nucleation is the rate-limiting step of actin polymerization in a test tube (this produces the characteristic lag phase seen in polymerization assays). Cells overcome this barrier with dedicated nucleators:

• The Arp2/3 complex (Actin-related proteins 2 and 3, plus five other subunits) binds the side of an existing "mother" filament and nucleates a new "daughter" filament at a characteristic 70° branch angle, building dendritic, branched networks typical of the lamellipodium at a crawling cell's leading edge. Arp2/3 is itself activated by nucleation-promoting factors such as WASP/WAVE family proteins, which couple it to upstream signals (e.g., Rho-family GTPases like Cdc42 and Rac1).
• Formins are dimeric proteins that use a doughnut-shaped FH2 domain to processively cap and track the barbed end of a growing filament while remaining attached, adding profilin-actin monomers and producing long, unbranched, bundled filaments typical of filopodia, stress fibers, and the cytokinetic contractile ring.

Step 3 — Elongation and the barbed/pointed asymmetry. Once a filament nucleus exists, monomers add reversibly at both ends, but far faster at the barbed end. Each end has its own critical concentration (Cc), the free monomer concentration at which addition and loss are balanced (net growth = 0). For ATP-actin under typical ionic conditions:

Because cellular free ATP-G-actin concentration typically sits between these two values, filaments simultaneously add subunits at the barbed end and lose subunits at the pointed end — the essence of treadmilling.

Step 4 — ATP hydrolysis as a molecular clock. Shortly after an ATP-actin monomer is incorporated into a filament, the actin subunit's intrinsic ATPase activity hydrolyzes the bound ATP to ADP-Pi, and phosphate is subsequently released more slowly, leaving ADP-actin within the filament core. This creates an age gradient along the filament: newly added subunits near the barbed end are ATP- or ADP-Pi-bound, while older subunits toward the pointed end are ADP-bound. ADP-actin subunits pack less favorably in the filament lattice and dissociate from the pointed end far more readily than ATP-actin — so hydrolysis is not just bookkeeping, it is the thermodynamic engine that drives net disassembly specifically at the pointed end, sustaining treadmilling and giving the cell a way to recycle actin monomers rapidly.

Step 5 — Regulatory proteins sculpt the network.

• Capping proteins (e.g., CapZ) bind barbed ends and block further addition or loss, stabilizing filament length and forcing new growth to depend on fresh nucleation rather than elongation of old filaments — essential for maintaining a steep, responsive lamellipodial network.
• Cofilin/ADF (actin-depolymerizing factor) preferentially binds ADP-actin subunits, destabilizing the filament lattice; it severs older, ADP-rich filament segments and promotes pointed-end disassembly, dramatically accelerating monomer recycling — treadmilling in cells is far faster with cofilin than the bare kinetics alone would allow.
• Crosslinkers and bundlers (α-actinin, fimbrin/plastin, filamin, spectrin) organize filaments into parallel bundles (filopodia, stereocilia, microvilli) or orthogonal gels (cortex), determining whether an actin structure behaves as a stiff spike or a soft, isotropic meshwork.
• Motors (myosins) walk along filaments toward the barbed end (for most myosins) using ATP hydrolysis, generating contractile force in muscle sarcomeres, the cytokinetic ring, and cortical tension generally — this motor mechanochemistry is developed fully elsewhere.

Step 6 — From dynamics to shape. Because nucleation, elongation, capping, severing, and crosslinking are all independently tunable, a cell can build strikingly different actin architectures from the same monomer pool: a branched Arp2/3 meshwork pushing the plasma membrane forward in a lamellipodium; parallel formin-nucleated bundles supporting rigid filopodial spikes; a contractile ring of antiparallel filaments and myosin II that constricts to divide the cell; and a dense cortical actin gel beneath the plasma membrane that gives the cell surface its mechanical resilience. Local signaling (Rho-family GTPases, phosphoinositides, calcium) determines which nucleators are switched on where, translating extracellular cues (a chemoattractant gradient, a mechanical stimulus) into a specific spatial pattern of polymerization.

The polarity of actin filaments was first revealed not by biochemistry but by a classic decoration trick. In the 1960s, H. E. Huxley showed that treating actin filaments with the proteolytic myosin fragment heavy meromyosin (HMM), or its subfragment S1, causes myosin heads to bind actin at a fixed angle along the whole filament. Under electron microscopy this decoration produces a repeating arrowhead pattern, and crucially, all arrowheads on a given filament point the same direction. This directly demonstrated that the filament is structurally polarized — every monomer is oriented the same way — providing the physical basis for calling one end "barbed" (blunt arrowhead) and the other "pointed." The kinetic and thermodynamic asymmetry between the two ends — that they have different critical concentrations and different growth rates — was established through in vitro polymerization kinetics, notably by Terrell Hill and Elisha Kondo's theoretical framework and experimental confirmation from groups including Marie-France Carlier, Dominique Pantaloni, and colleagues in the late 1970s and 1980s, alongside earlier polarity kinetics work by Alexander Wegner (1976). Wegner demonstrated that actin filaments seeded and observed under polymerizing conditions grow at different rates from their two ends, formally establishing the kinetic (not just structural) polarity of the filament and predicting that a filament could, in principle, add subunits at one end while losing them from the other — the theoretical prediction of treadmilling. The direct, quantitative proof of treadmilling came from pyrene-actin fluorescence assays. Actin monomers can be labeled with the fluorophore pyrene at a specific cysteine residue; pyrene-actin fluoresces roughly 20-fold more intensely when incorporated into filaments than as free monomer, because polymerization restricts pyrene's rotational freedom and quenching. By mixing pyrene-labeled monomers with pre-formed, unlabeled filament seeds and monitoring fluorescence over time, Carlier, Pantaloni, and coworkers (key papers 1983–1985) could measure elongation rates at each end separately (using seeds capped at one end, or gel-filtered filament fragments enriched for one polarity) and showed:

• Barbed-end elongation rate constants for ATP-actin association are roughly 10-fold higher than at the pointed end.
• At steady state, with free monomer concentration held between the two critical concentrations, filaments showed constant total polymer mass while individual pyrene-labeled subunits could be shown (using pulse-chase-like protocols and later single-filament imaging) to flux through the polymer from barbed to pointed end — the direct observation of treadmilling.
• ATP hydrolysis was shown to be required for this behavior: filaments polymerized from a slowly hydrolyzable ATP analog, or artificially loaded with ADP rather than ATP, lost the pointed-end-biased disassembly and the characteristic asymmetry in critical concentrations collapsed, confirming that nucleotide hydrolysis — not just binding — drives the asymmetry.

A complementary and famous demonstration came from total internal reflection fluorescence (TIRF) microscopy of single actin filaments (pioneered in reconstituted systems in the 1990s–2000s), which allowed researchers to directly watch individual filaments elongate at their barbed ends in real time while shrinking at their pointed ends, visually confirming decades of bulk kinetic inference with single-molecule imaging. Together, decoration electron microscopy (structural polarity), pyrene fluorescence kinetics (functional/kinetic polarity and treadmilling), and single-filament TIRF imaging (direct visualization) form the converging, mutually reinforcing evidence base for the modern picture of dynamic, polarized actin assembly. A second landmark line of evidence concerns nucleation. Purified Arp2/3 complex alone nucleates filaments very inefficiently; Thomas Pollard and colleagues (late 1990s–2000s) showed using pyrene assays and electron microscopy that Arp2/3 requires activation by nucleation-promoting factors (WASP family proteins) and an existing "mother" filament to generate the characteristic branched networks, and that these reconstituted branched networks, when combined with capping protein and profilin-actin in vitro, reproduce the treadmilling, pushing behavior of a lamellipodium — the celebrated in vitro reconstitution of actin-based "comet tail" motility (originally characterized using the intracellular pathogen Listeria monocytogenes, which hijacks Arp2/3 machinery to propel itself through host cytoplasm, work led by Julie Theriot, Tim Mitchison, and colleagues in the 1990s) proved that the branched-network nucleation-elongation-capping cycle alone, with no additional cellular machinery, is sufficient to generate directional force.

Example 1 — Calculating net filament behavior from critical concentrations. Suppose an in vitro reaction contains free ATP-G-actin at 0.3 µM, with a barbed-end critical concentration of 0.10 µM and a pointed-end critical concentration of 0.70 µM (typical published values for pure ATP-actin under physiological ionic strength).

• At the barbed end: free monomer (0.3 µM) > Cc (0.10 µM), so the barbed end grows net.
• At the pointed end: free monomer (0.3 µM) < Cc (0.70 µM), so the pointed end shrinks net.
• Because the barbed end gains subunits while the pointed end loses them, and (in a closed system at steady state) the rate of loss from the pointed end comes to match the rate of gain at the barbed end, total filament mass stays constant while the filament treadmills — subunits enter at the barbed end and exit at the pointed end.
• If a drug (say, a capping-protein mimic) blocks the barbed end selectively, only pointed-end loss continues unopposed: the filament will now shrink net overall, since there is no compensating barbed-end addition. This is mechanistically what cytochalasins do experimentally, and it is why capping protein is essential in cells to prevent uncontrolled barbed-end growth from consuming the entire monomer pool.

This calculation illustrates why a single free monomer concentration can simultaneously support growth at one end and loss at the other — the qualitative signature of treadmilling — and why measuring only "total polymer" in a bulk assay can look like nothing is happening even while subunits are churning through the filament.

Example 2 — Diagnosing a motility defect. A researcher engineers a cultured fibroblast cell line expressing a mutant cofilin that cannot bind ADP-actin (loss-of-severing-function). Predict the phenotype at the leading edge, and reason it through.

• Cofilin's normal jobs are to (1) sever aged, ADP-rich filament segments and (2) accelerate pointed-end depolymerization, both of which recycle actin monomers back into the free G-actin pool for reuse at the barbed end.
• Without functional cofilin, filaments persist longer: pointed-end depolymerization slows, so the flux of monomers back into the cytosolic pool decreases.
• Because Arp2/3-driven barbed-end growth at the leading edge depends on a steady supply of free ATP-actin monomer (often delivered via profilin), a shrinking free-monomer pool will eventually slow lamellipodial protrusion, even though existing filaments are, if anything, more stable.
• Experimentally, this predicts (and matches observations from cofilin knockdown/knockout studies) a cell with a larger standing mass of filamentous actin at the cortex, but reduced rates of leading-edge protrusion and reduced cell migration speed — because what limits productive motility is not simply "more filament" but rapid turnover, which depends on cofilin-mediated recycling.
• This example generalizes an important principle: actin-based motility is rate-limited by monomer recycling, not by the intrinsic polymerization reaction alone — a fast treadmill needs a fast return belt.

Much of the quantitative framework for actin dynamics traces to Marie-France Carlier, a French biophysical chemist whose laboratory, working closely with Dominique Pantaloni at the CNRS in Gif-sur-Yvette, developed and applied the pyrene-actin fluorescence assay that turned actin polymerization from a qualitative electron-microscopy phenomenon into a precisely measurable chemical kinetic system. Their work in the 1980s established the differential critical concentrations at the two filament ends, quantified the coupling between ATP hydrolysis and subunit stability, and provided the first direct biochemical demonstration of treadmilling as a real, ATP-dependent phenomenon rather than a theoretical curiosity. This tradition of reconstituting cytoskeletal dynamics from purified components in a test tube — rather than only observing cells — became a defining methodology of the field, later extended by Thomas Pollard (whose lab characterized Arp2/3-mediated nucleation and branching in exquisite mechanistic detail) and by single-molecule imaging pioneers who watched individual filaments grow and shrink in real time.

Actin dynamics are a direct target of clinically important drugs and a central theme in modern cell biology research. Cytochalasins cap actin filament barbed ends and are used experimentally to dissect the role of actin polymerization in cell processes; latrunculins sequester G-actin monomers and are widely used pharmacological tools in cell biology labs to acutely disrupt the actin cytoskeleton and test causality in live-cell experiments. Jasplakinolide, by contrast, stabilizes filaments and is studied both as a research tool and, in modified forms, for potential anticancer activity, since it can lock the dynamic actin network of rapidly dividing or migrating tumor cells into a nonfunctional, over-stabilized state. In cancer biology, metastatic tumor cells frequently hijack the same Arp2/3- and formin-driven protrusive machinery used by immune cells and fibroblasts, and invasive structures called invadopodia rely on localized actin polymerization to degrade and penetrate the extracellular matrix; drugs targeting Arp2/3 or upstream Rho-GTPase signaling are active areas of oncology research aimed at blocking invasion and metastasis rather than only cell division. In immunology, actin-based motility underlies immune cell chemotaxis toward infection sites, and mutations in actin-regulatory genes (for example, WAS, encoding Wiskott-Aldrich syndrome protein, a WASP-family activator of Arp2/3) cause Wiskott-Aldrich syndrome, a human immunodeficiency with defective platelet and lymphocyte actin dynamics, illustrating how a single node in this pathway, when broken, produces a defined clinical syndrome. In muscle biology, mutations in sarcomeric actin genes (ACTA1, ACTC1) cause inherited myopathies and cardiomyopathies by disrupting the ordered actin-myosin arrays needed for contraction. Beyond medicine, understanding actin dynamics underlies agricultural and plant biology research into cell wall deposition and pollen tube growth (both actin-dependent), and informs synthetic biology efforts to build minimal, actin-based motile or shape-changing synthetic cells — an engineering ambition that depends entirely on correctly harnessing the polarity and treadmilling principles first established in these classic experiments.

Ask the tutor: How does Rho-GTPase signaling decide whether a cell builds a lamellipodium versus a filopodium at a given location?