Document Type



Doctor of Philosophy



First Adviser

Vavylonis, Dimitrios


Actin proteins polymerize into many different filamentous structures within individual cells. These actin structures coexist, each playing a significant role in the function of cells. The biophysical basis of this competition however remains an area in need of further investigation. In fission yeast actin patches (nucleated by the Arp2/3 protein complex) and actin cables (polymerized by formin proteins) coexist and regulate endocytosis and cell tip growth, respectively. The available quantitative data and the existence of only two distinct actin structures offer the possibility of using fission yeast as model system to develop quantitative mathematical models to study the interdependence of actin cytoskeleton structures in cells. Recent experimental studies have shown that actin patches and actin cables compete for the same pool of monomeric actin under the regulation of many proteins such as profilin, fimbrin, cofilin, and tropomyosin. To quantify this competition, we developed a mathematical model using a set of differential equations. The model incorporates the most important regulatory factors revealed by prior experiments while using a minimal set of parameter values. In the model actin can be distributed in three pools: patches, cables and cytoplasm. The Arp2/3 complex contributes to patch nucleation and is consumed in patches. Fimbrin and cofilin incorporate in patches and cables and regulate patch and cable lifetime. Profilin binds to actin monomers in the cytoplasm and regulates the elongation rate of actin filaments in cables.The model captured the main qualitative and quantitative trends in several prior experimental studies, such as the observed increase in ectopic actin filament bundles upon treatment with the drug CK-666 that disassembles actin patches. It can also capture the change in actin patches and actin cables upon underexpression/overexpression of actin , in combination with CK-666, as well as the increase in actin patch number in cofilin and formin mutants. The model can also describe the change in patch number in experiments of profilin overexpression. The model provides predictions that can be tested in future experiments and illustrates the degree of complexity of mutual dependencies among actin cytoskeletal structures.The development of actin networks of different structure and morphology depends on many proteins that regulate the dynamics of actin filaments, such as their length, lifetime and binding interactions. In particular, several actin filament side-binding proteins can sever, stabilize or bundle actin filaments. In this study we focused on three of these proteins, namely tropomyosin, cofilin and fimbrin, which are found in many actin cytoskeletal structures. Recent in vitro studies have shown that their actin side-binding dynamics are affected in the presence of each other. In order to study the kinetics and organization of these competing binders along actin filaments, we use stochastic simulations. In the model the actin filament is represented as two independent lattices, representing the two protofilaments of the actin filament double helix. For simplicity, we neglected the mutual dependence between the bound proteins of one protofilament to the other. In accordance with prior in vitro experiments, in our model the binding of a protein to one or more (for the case of tropomyosin) lattice units excludes the binding of proteins of different or same type to these lattice units. Taking into account their actin binding cooperativity properties, we parametrized the model by fitting prior experimental data and using parameters from previous models. The model reveals the range of concentrations where one protein dominates against the other from the start of the simulation until equilibrium but also areas of concentrations where there is a shift of the dominating protein between early times and equilibrium We find concentration ranges where initially tropomyosin occupies a large portion of the lattice but then either cofilin or fimbrin dominate the equilibrium state. In these cases, we find that while initially cofilin or fimbrin bind in smaller numbers than tropomyosin, they create boundaries that don’t allow for long stable tropomyosin chains, so in time tropomyosin is being removed by the lattice. Simulations of actin polymerization that includes tropomyosin, cofilin and fimbrin showed that fimbrin inhibits the elongation of tropomyosin chains on early times, allowing the binding of cofilin on sites where the actin filament hasn’t released Pi.

Included in

Physics Commons