Investigation of multiscale biological phenomenon using kinetic modeling and molecular dynamics simulations

Abstract

Biological systems inherently function at multiple spatial and temporal scales by integrating information across these scales. In this thesis, we are interested in investigating two classes of biological phenomena that occur at different time and length scales. The first phenomenon is concerned with the development of a mechanistic kinetic model to capture mitochondria and microtubule interaction and predict mitochondrial evolution dynamics in fission yeast cells. The second phenomenon that we investigate occurs primarily at the molecular level, where we are interested in analyzing membrane-binding proteins implicated in bacterial infections.

Mitochondrial populations in cells are maintained by cycles of fission and fusion events, and recent experiments with fission yeast cells [Mehta et al., J. Biol. Chem., 2019,294,3385], illustrate the intricate coupling between mitochondria and the dynamic population of microtubules within the cell. In order to understand this coupling, we carried out kinetic Monte Carlo simulations to predict the evolution of mitochondrial size distributions for different cases; wild-type cells, cells with short and long microtubules, and cells without microtubules. Comparisons are made with mitochondrial distributions reported in experiments with fission yeast cells. The model provides greater physical insight into the temporal evolution of mitochondrial populations in different microtubule environments, allowing one to study both the short-time evolution as observed in the experiments (<5 minutes) as well as their transition towards a steady state (>15 minutes).

In the second part of the thesis, we used molecular dynamics simulations to study pore-forming toxins, which are a class of proteins secreted by various bacterial strains to mediate infections. We examine the effect of the 25- hydroxycholesterol (oxysterol) in the pore formation pathway of ClyA toxin expressed by E. coli, using molecular dynamics (MD) simulations. Previous reports from our laboratory elucidated the role of cholesterols in assisting the pore formation activity of ClyA. On the contrary, oxidized sterol derivatives such as 25-hydroxycholesterol are thought to provide immunity against bacterial infections by altering the accessible cholesterol content in the cell. Using all-atom MD simulations, we report higher structural deviations in the N-terminus membrane binding domain of ClyA in the presence of 10 % oxysterols in the membrane. The free energy for the transition, obtained using enhanced path based sampling methods, between the monomer and membrane bound states for the key membrane binding motifs show a barrier of ∼ 10 kT for the transition. These simulations provide insights into the loss of pore formation with oxysterols.

The last part of the thesis is concerned with the mechanism of pore formation of a two-component α-helical bacterial pore-forming toxin, YaxAB, using molecular dynamics (MD) simulations. YaxAB belongs to the ClyA family of toxins expressed by Yersinia enterocolitica and forms pores with the ensuing dimerization of two proteins, YaxA and YaxB. We report that a single protomer consisting of the YaxA-YaxB heterodimer subunit possesses the ability to permeabilize the membrane and form a stable water channel. We show for the first time that cholesterol molecules can stabilize the YaxA- YaxB subunit through strong interaction with various residues of the membrane inserted foot domain. Our simulations shed light on the putative role of YaxA acting as an anchor to help form the initial dimer that assembles into a pore complex of the bipartite YaxAB toxin.