Abstract:
The length and time scales involved in biomolecular systems, including scenarios in which multiple proteins interact at the same time, are often too large for an efficient calculation via conventional Molecular Dynamics (MD) simulations. In such cases, coarse-graining (CG) provides a valuable approach to explore these otherwise inaccessible regimes. CG is not only born out of necessity due to insufficient computing power, but can provide a simpler, functional description of the molecular system at hand --- one that is difficult to extract from atomic chaos. However, care is needed in the process: the challenge is to select which details to preserve and which to omit.
The present thesis addresses this aforementioned issue in its first part, among other things. There, a flexible, low-resolution CG model of bovine serum albumin (BSA, a globular protein frequently investigated in the literature) is developed, and a systematic method to benchmark different realizations of a CG model against its atomistic counterpart is presented. In a second step, the effects of molecular flexibility and anisotropy on diffusive properties and solution structure in self-crowded CG systems are investigated. For this purpose, a CG model of the antibody protein immunoglobulin G (IgG), acting as an anisotropic reference molecule, is implemented.
In the second part of this work, we attempt to rationalize our newly obtained BSA model in a more general, colloidal picture and seek to find similarities to patchy particle systems based on mutual thermodynamic properties. This is of particular interest, since in experiments with globular proteins solution stability is often predicted in terms of isotropic reference systems from colloidal theory. As our model exhibits explicit shape anisotropy, we investigate to which extent our CG model can be effectively considered isotropic or anisotropic, with a particular focus on the effects of different electrostatic charge states of the molecule.
It is well established that most internal structural fluctuation of proteins is confined to a limited number of dominant modes, as revealed by principal component analysis (PCA) of atomistic simulation trajectories. However, the impact of crowding on these fluctuations remains largely unclear. The third part of this thesis addresses this topic by analyzing CG systems of self-crowded BSA and IgG. For the latter, we demonstrate that PCA fails to fully capture the relevant motion patterns, whereas time-lagged independent component analysis (TICA) provides a more accurate description of the dominant dynamical modes.
Proteins carry out their function in densely crowded environments, inside biological cells. This macromolecular crowding amplifies the importance of hydrodynamic interactions (HIs), which in turn affect the speed of the diffusive search process which the proteins undergo to fulfill their purpose. In the last part of this thesis, we present an analysis of the influence of HIs on reaction speed in a cellular environment, in which the intracellular molecular composition of E. Coli is represented by a polydisperse mixture of spheres.
