Computer simulations provide a potentially powerful complement to conventional experimental techniques in elucidating the structures, dynamics and interactions of macromolecules. In this thesis, I present three applications of computer simulations to investigate important biomolecules with sizes ranging from two-residue peptides, to proteins, and to whole chromosome structures.
First, I describe the results of 441 independent explicit-solvent molecular dynamics (MD) simulations of all possible two-residue peptides that contain the 20 standard amino acids with neutral and protonated histidine. 3JHNHα coupling constants and δHα chemical shifts calculated from the MD simulations correlated quite well with recently published experimental measurements for a corresponding set of two-residue peptides. Neighboring residue effects (NREs) on the average 3JHNHα and δHα values of adjacent residues were also reasonably well reproduced. The intrinsic conformational preferences of each residue, and their NREs on the conformational preferences of adjacent residues, were analyzed. Finally, these NREs were compared with corresponding effects observed in a coil library and the average β-turn preferences of all residue types were determined.
Second, I compare the abilities of three derivatives of the Amber ff99SB force field to reproduce a recent report of 3JHNHα scalar coupling constants for hundreds of two-residue peptides. All-atom MD simulations of 256 two-residue peptides were performed and the results showed that a recently-developed force field (RSFF2) produced a dramatic improvement in the agreement with experimental 3JHNHα coupling constants. I further show that RSFF2 also improved modestly agreement with experimental 3JHNHα coupling constants of five model proteins. However, an analysis of NREs on the 3JHNHα coupling constants of the two-residue peptides indicated little difference between the force fields’ abilities to reproduce experimental NREs. I speculate that this might indicate limitations in the force fields’ descriptions of nonbonded interactions between adjacent side chains or with terminal capping groups.
Finally, coarse-grained (CG) models and multi-scale modeling methods are used to develop structural models of entire E. coli chromosomes confined within the experimentally-determined volume of the nucleoid. The final resolution of the chromosome structures built here was one-nucleotide-per-bead (1 NTB), which represents a significant increase in resolution relative to previously published CG chromosome models, in which one bead corresponds to hundreds or even thousands of basepairs. Based on the high-resolution final 1 NTB structures, important physical properties such as major and minor groove widths, distributions of local DNA bending angles, and topological parameters (Linking Number (Lk), Twist (Tw) and Writhe (Wr)) were accurately computed and compared with experimental measurements or predictions from a worm-like chain (WLC) model. All these analyses indicated that the chromosome models built in this study are reasonable at a microscopic level. This chromosome model provides a significant step toward the goal of building a whole-cell model of a bacterial cell.