Molecular Dynamics for Synthetic Biology

Abstract

Synthetic biology is the field concerned with the design, engineering, and construction of organisms and biomolecules. Biomolecules such as proteins are nature's nano-bots, and provide both a shortcut to the construction of nano-scale tools and insight into the design of abiotic nanotechnology. A fundamental technique in protein engineering is protein fusion, the concatenation of two proteins so that they form domains of a new protein. The resulting fusion protein generally retains both functions, especially when a linker sequence is introduced between the two domains to allow them to fold independently. Fusion proteins can have features absent from all of their components; for example, FRET biosensors are fusion proteins of two fluorescent proteins with a binding domain. When the binding domain forms a complex with a ligand, its dynamics translate the concentration of the ligand to the ratio of fluorescence intensities via FRET.

Despite these successes, protein engineering remains laborious and expensive. Computer modelling has the potential to improve the situation by enabling some design work to occur virtually. Synthetic biologists commonly use fast, heuristic structure prediction tools like ROSETTA, I-TASSER and FoldX, despite their inaccuracy. By contrast, molecular dynamics with modern force fields has proven itself accurate, but sampling sufficiently to solve problems accurately and quickly enough to be relevant to experimenters remains challenging.

In this thesis, I introduce molecular dynamics to a structural biology audience, and discuss the challenges and theory behind the technique. With this knowledge, I introduce synthetic biology through a review of fluorescent sensors. I then develop a simple computational tool, Rangefinder, for the design of one variety of these sensors, and demonstrate its ability to predict sensor performance experimentally. I demonstrate the importance of the choice of linker with yet another sensor whose performance depends critically thereon. In chapter 6, I investigate the structure of a conserved, repeating linker sequence connecting two domains of the malaria circumsporozoite protein. Finally, I develop a multi-scale enhanced sampling molecular dynamics approach to predicting the structure and dynamics of fusion proteins. It is my hope that this work contributes to the structural biology community's understanding of molecular dynamics and inspires new techniques developed for protein engineering.