Molecular dynamics simulations allowed us to understand the binding modes of different imidazolium-activated dinucleotide intermediates in nonenzymatic RNA template copying process. We demonstrate that the rate and fidelity of the reaction depends on the conformational space of the activated dinucleotide and its preference to maintain a stacked A-RNA-like structure, and not solely on the number of hydrogen bonds in the Watson–Crick interaction.

Abstract

Nonenzymatic self-replication is considered as one of the most primordial functions of RNA, which likely preceded the emergence of more complex ribozymes. Among different possible scenarios, nucleotide activation with imidazole derivatives attracted substantial attention over the last years. However, despite the progress in proposing plausible variants of nonenzymatic RNA template copying with phosphoroimidazolides, mechanistic aspects of this process still remain obscure. Furthermore, efficient RNA self-replication involving activated uridine and adenosine still remains a challenge. Here, we employed classical molecular dynamics simulations to evaluate the binding specificity of different imidazolium-bridged dinucleotide intermediates, which was suggested to control the yield and fidelity of the reaction. In particular, RMSD-based clustering of the MD trajectories revealed previously unknown structural arrangements of activated dinucleotide intermediates that may play a critical role in nonenzymatic primer extension. Most importantly, our results indicate that yield and fidelity of nonenzymatic RNA template copying cannot be simply associated with the number of Watson–Crick hydrogen bonds between the activated dinucleotides and the templating strand. Instead, the efficiency of the reaction correlates with the preference for the formation of the canonically stacked form of the activated dinucleotide intermediate, which can then selectively bind to the template and participate in the primer extension reaction.