Actin Based Cell Motility

Actin is central to the function of the cell and plays many important roles. Structural biology has provided us with a wealth of actin and actin-related protein structures in recent years, and using these structures along with a wide array of computational techniques we are able to look at the dynamics and interactions of actin at the molecular level. Using such methods our past work has helped in the understanding of basic processes such as actin nucleation and polymerization, including an explanation of the difference between the barbed and pointed end polymerization rates. Our current work includes many different projects looking at the interaction and regulation of actin and actin associated proteins. A few of these projects are summarized below.

Arp 2/3 Complex

The Arp2/3 Complex plays two important roles within the cell: it nucleates the formation of new actin filaments following activation, and it also forms branches off of existing actin filaments leading to the dendritic network observed at the leading edge of motile cells. It had been determined that p20 and p34, two proteins in the Arp2/3 complex, make the principle interactions with the filament, however the details of the binding were not known. To investigate this interaction we performed molecular dynamics simulations on an actin filament as well as the Arp2/3 complex. We selected p20 and p34 structures from this simulation and docked them to the filament structures resulting in a model for the Arp2/3 complex branch.

We find excellent interaction between p20-p34 and F-actin (see Figure above), and many of the key interacting residues we have identified are currently being tested (and now verified) in mutagenesis studies (Matt Welch, Berkeley). Further analysis indicates that p16 may play a role in helping to activate the Arp2/3 complex, and this work is ongoing. Other efforts are ongoing in the lab to determine how the Arp2/3 complex is activated by VCA peptides, what the structure of the activated Arp2/3 complex looks like, and what the roles of the individual subunits in the Arp2/3 complex are.

Capping Protein

Capping protein binds to the barbed end of actin filaments and is required to regulate the extent of actin polymerization, but its function is likewise regulated by other proteins.

Phosphatidylinositol-4,5-bisphosphate (PIP2) can bind and inhibit capping protein leading to the polymerization of actin at the leading edge of the cell. Through a series of molecular dynamics and small molecule docking simulations we identified the binding site of PIP2 on capping protein as shown in the figure above. Three positively-charged residues in CP at the base of the α-C-terminus were identified, and subsequent mutagenesis work from John Cooper's lab (Washington University School of Medicine) has confirmed our predictions. In addition to PIP2, we are looking at the interaction of capping protein with Twinfilin, CARMIL, CD2AP, CKIP-1 and Myotrophin/V-1. Our lab has developed a novel method for predicting protein-peptide and protein-protein interactions and we are using these methods to study each of these systems, as well as the basic interaction between capping protein and the barbed end of the actin filament.

Multi-scale Modeling

The functional form of actin is truly the actin filament, but structural information about the filament is very difficult to come by and must be interpreted from averaged cryo-EM structures, cross-linking studies and other techniques. We have performed long time-scale molecular dynamics trajectories on actin monomers and short filaments in order to look at the dynamics and interactions within the filament. These simulations have allowed us to look at details such as the hydrophobic plug and the conformation/interactions of subdomain 2 and the DNase I loop. These simulations are being extended to examine the differences between ADP and ATP actin, as well as the intrinsic differences between yeast and muscle actin. Apart from molecular simulations, we have also developed a multiscale model for an actin filament that is able to capture many of the observed macroscopic properties. In our model, the four subdomains of the actin protein are treated as beads connected by simple springs. The spring constants and equilibrium lengths are determined from long time scale molecular dynamics simulations of the monomer. Similarly, molecular dynamics simulations of a short filament are used to parameterize springs between monomers within the filament. Using a Brownian dynamics approach, we are able to both reproduce the coarse-grained dynamics of the monomer and simulate filaments of hundreds to thousands of monomers. Without adding any additional factors we reproduce macroscopic filament properties such as the persistence length and the extensibility of filaments based on single molecule experiments. This method appears to be an ideal approach to studying such a system and will serve as an excellent starting point for studying the interaction of actin binding proteins and other large protein complexes.

F-Actin Binding Proteins

Having simulations of F-actin has subsequently allowed us to study a large number of other problems. Shown below are docking predictions for the binding of jasplakinolide (panel B) and phalloidin (panel C) to an actin filament.

These results are in excellent agreement with cryo-EM and mutagenesis data and confirm that these drugs are competitive inhibitors. We have also started work to look at the binding of ADF/cofilin to F-actin. These proteins bind cooperatively to F-actin, twist and sever actin filaments. Through binding and dynamics simulations we plan to determine how these proteins disrupt the filament structure and lead to the depolymerization of filaments. This work is being performed with Emil Reisler (UCLA), Peter Rubenstein (University of Iowa) and David Sibley (Washington University School of Medicine).

There are numerous other actin related projects currently going on in our group including work with formins (Bruce Goode, Brandeis), tropomodulin (Velia Fowler, Scripps), profilin, Aip1p (David Amberg , SUNY Upstate Medical University), fascin (Anders Carlsson, Washington University) and parasitic actins (David Sibley, Washington University School of Medicine).