The statistical mechanics of free and protein-bound DNA by Monte Carlo simulation

There are many challenges involved in the simulation of DNA. In this work, novel Monte Carlo techniques are developed and applied to understanding the biophysical properties of DNA. A coarse-grained model is applied to feasibly simulate long DNA chains of hundreds to thousands of base pairs, using a reduced base-pair step representation of the DNA. Using this model, a canonical Monte Carlo simulation of DNA is developed to characterize the structure and flexibility of double-helical DNA. By applying a unique algorithm for generating uncorrelated DNA conformations a priori, limitations of the original Metropolis Monte Carlo algorithm are avoided.Furthermore, there is developing experimental evidence that non-specifically associating proteins that induce DNA bending modulate the in-vivo flexibility of DNA. To investigate the effect of these proteins, a grand canonical Monte Carlo simulation technique is developed, extending the model of free DNA to incorporate non-specific protein-DNA interactions. In this technique, DNA chains are simulated with varying numbers of bound proteins. Ubiquitous DNA architectural proteins such as the prokaroytic nucleoid protein HU and the eukaryotic HMG-box proteins are investigated with this technique. By incorporating structural information from the protein-DNA complexes currently available in the Nucleic Acid Database, models of these DNA-binding proteins are constructed and used in this method.The results predict an enhancement of DNA flexibility due to non-specific binding of these proteins, and calculations of the cyclization (ring-closure) properties and force-extension responses of protein-bound DNA chains compared to free DNA chains are presented. In addition, the effects of these proteins on the topological properties of closed circular DNA and on the looping properties of DNA constrained by binding to the Lac repressor protein assembly are characterized in large-scale parallel simulations. Coordination of protein binding on circular and looped DNA and induction of negative supercoiling of DNA by DNA architectural proteins is predicted, with important biological implications for chromosome organization and transcription regulation.