Just like complex vertebrates, cells are supported by their own kind of skeleton. In fact, the cell’s skeleton, called the cytoskeleton, is very dynamic—the constituent “bones” can assemble and disassemble as required. One key player in the cytoskeleton is the actin filament, which is composed of actin subunits. The mechanism of actin filament assembly has been studied for many decades because it is involved in a range of crucial cellular processes, from cell division to muscle contraction.
Thomas Pollard, Sterling Professor of Molecular, Cellular and Developmental Biology at Yale, has dedicated much of his career to studying the cytoskeleton. In 1975, one of his students discovered that actin filaments grow much faster at one end—the “barbed” end— than the “pointed” end. More than forty years later, Pollard and Steven Chou, an associate research scientist, recently published a study that unveiled new molecular insights into how exactly actin polymerizes. They were able to obtain three high-resolution crystal structures of actin filaments bound to three different nucleotides—the individual subunits of DNA and RNA—allowing them to directly visualize key moments in actin polymerization. “These three new structures of actin filaments are a dream come true for me. They answer questions that arose but were never answered over the past seventy years,” Pollard said.
Cryogenic electron microscopy
For a long time, molecular biologists have studied the structure of biological molecules and their atomic-level details by examining their crystal structures. However, the assembly of actin filaments does not fit any crystal packings, so crystallization is not feasible for observing actin filament structure. “To crystallize, the proteins have to be very ordered—the flexible parts may never be able to crystallize,” Chou said.
Cryogenic electron microscopy (cryo-EM) is a modern technique that allows researchers to determine the structures of macromolecules without having to crystallize them. Typically, an aqueous solution of the molecule is laid on a grid, a circular disk made with a metal mesh and a thin layer of carbon membrane on one side. This grid is then plunged into liquid ethane at negative 128 degrees Fahrenheit, capturing the molecule in a frozen lattice of water. “It’s exciting that we have more advanced detector technology that lets us routinely obtain images of these structures at near-atomic resolution,” Chou said. The new study used cryo-EM to examine structures of actin at various stages of molecular processes.
Obtaining the structures
Obtaining high-resolution images using cryo- EM has its own challenges. Pollard and Chou first tried to freeze the sample with Vitrobot, a robot plunger. One crucial aspect during the freezing is controlling humidity, which is important for imaging quality. Vitrobot does this by sandwiching the cryo-EM grid between filter paper. In this case, however, the filter paper was found to damage the actin filaments. Manually removing the extra buffer from the edge of the grid did not work because a manual procedure introduced variability in temperature and humidity. They settled for an approach using Vitrobot whereby the filter paper only touched the edge of the grid to avoid crushing the filaments, while also controlling for temperature and humidity. This approach allowed for the filaments to remain intact and for the experiments to be reproducible.
According to Chou, the preparation of the sample requires an extremely thin layer of ice, since thick ice deteriorates the image contrast and causes large filament tilt, leading to low resolution when reconstructing the final structure. Actin “tilting” is best understood as the way a straw in a glass of water might slant to its side. Even within the same sample, each actin filament would have a different tilt angle. To ensure optimal visualization, Pollard and Chou used software called AutoPick that detected the filament with minimal tilt angle for optimal filament selection. With all these optimizations, Pollard and Chou were able to achieve exceptional, high-resolution images.
A key aspect of imaging actin was capturing actin bound to nucleotides, which is crucial for its biological activity. Actin is typically complexed with the molecular energy currency adenosine triphosphate (ATP), a structure consisting of an adenosine nucleotide, a pentose sugar, and three phosphate groups. ATP is the molecular energy currency of the cell, as its breakdown releases energy that drives other important cellular processes. In this case, ATP-actin complexes add onto actin filaments. Over time, the ATP molecule is hydrolyzed—or broken down by water—to give adenosine diphosphate (ADP), releasing an inorganic phosphate molecule.
Pollard and Chou produced visualizations of actin bound to AMP-PNP, ADP, and ADP-Pi, each of which offered unique insights into different aspects of actin polymerization depending on the identity of the bound nucleotide.
Tracking conformational changes
Through the high-resolution images of the nucleotide-bound actin, answers to many questions became clear. For instance, the cryo-EM images provided a clear explanation as to why actin filaments grow faster at the barbed end. When each additional actin unit binds, the filament undergoes a change in conformation that creates a better binding site for the next molecule to add because of a change in the angle at which actin subunits come together, known as a “dihedral angle.” This explains how the barbed end, to which more units add, grows quicker. In this way, the conformational changes affect the directionality of binding.
Another key piece of information obtained by the researchers was the timing for conformational changes: The major conformational, structural changes in the filament occur when actin molecules are incorporated into the filaments, rather than during ATP hydrolysis and phosphate release.
Moreover, these EM structures explain why actin molecules are inactive as monomers and active in filaments, meaning that biological activity and motility are only present in filaments. The rate of ATP hydrolysis is accelerated by 42,000-fold when monomers are assembled into filaments. When an actin molecule is incorporated into the filament, the catalytic residues targeted during ATP hydrolysis are re-oriented so that water may more easily attack the terminal phosphate groups of ATP on the actin filament.
Phosphate release
Phosphate release is an important step in ATP hydrolysis. These structures suggest a mechanism for how the terminal phosphate group is released from actin filaments once ATP is hydrolyzed into ADP bound to inorganic phosphate (Pi), denoted ADP-Pi. The process is akin to opening a backdoor to allow the phosphate molecule to leave.
After ATP is hydrolyzed into ADP and Pi, the phosphate is trapped in the filaments for several minutes before being released to the solution. The “backdoor” for phosphate release is closed in ADP-Pi-actin filaments. Chou and Pollard found that there is a large cavity near the phosphate into which the phosphate can move and change the interaction network. The interaction network change opens the backdoor for phosphate release. In ADP-actin filaments, the backdoor is open, allowing the phosphate in the solution to come into the filaments.
It is also known that phosphate release increases the filament’s binding affinity for cofilin, a molecule that disrupts the interactions of subunits in the filaments and promotes disassembly of the filament. Phosphate release changes the conformation of the C-terminus of actin subunits and increases the filament flexibility. This change in flexibility explains why cofilin can bind to ADP-actin filament but not the ADP-Pi-actin filament.
Cofilin binds to ADP-actin filaments close to the D-loop of each subunit. In the ADP-Pi-actin filament, the D-loop is rigid and in a conformation that prevents the attachment of cofilin to the filament. After phosphate release, the D-loop becomes flexible and allows the cofilin attachment. The binding of cofilin to the actin filament twists the actin subunit from the flattened conformation to the twist conformation. Therefore, cofilin can sever long ADP-actin filaments into shorter pieces.
The future of actin
By providing a comprehensive atomic-level view of the major molecular processes in actin polymerization, this study represents a tour de force of cryo-EM used in the context of elucidating molecular mechanisms in biological processes. The power of this method for studying a system as dynamic as actin is quite marvelous—it is likely that studies on similar systems will benefit from this approach. Given the accurate characterization of a quintessential cellular element, there is no question that current and future work related to actin will benefit from these insights.