Mechanisms of DNA hybridization: Transition path analysis of a simulation-informed Markov model

Complementary deoxyribonucleic acid (DNA) strands in solution reliably hybridize to form stable duplexes. We study the kinetics of the hybridization process and the mechanisms by which two initially isolated strands come together to form a stable double helix. We adopt a multi-step computational approach. First, we perform a large number of Brownian dynamics simulations of the hybridization process using the coarse-grained oxDNA2 model. Second, we use these simulations to construct a Markov state model of DNA dynamics that uses a state decomposition based on the inter-strand hydrogen bonding pattern. Third, we take advantage of transition path theory to obtain quantitative information about the thermodynamic and dynamic properties of the hybridization process. We find that while there is a large ensemble of possible hybridization pathways, there is a single dominant mechanism in which an initial base pair forms close to either end of the nascent double helix, and the remaining bases pair sequentially in a zipper-like fashion. We also show that the number of formed base pairs by itself is insufficient to describe the transition state of the hybridization process.