Accelerated simulations of RNA phase separation: a systematic study of non-redundant tandem repeats

Abstract

RNA promotes phase separation of proteins to form biomolecular condensates. Notably, pure RNA systems composed of RNA repeat sequences can phase separate and form liquid-like droplets. Significantly, such RNA tandem repeats have been implicated in several neurological and neuromuscular disorders. Thus, investigating the principles governing RNA repeat condensation can provide insights into RNA repeat expansion disorders. However, it is still challenging to characterize the microscopic structures and dynamics of RNA condensates. To address this, we have implemented and optimized a coarse-grained model to study the phase behavior of RNA. Our simulations reproduce key experimental findings on RNA–RNA phase separation. Importantly, we achieve multifold speed up in the simulation time compared to previous work. Leveraging this efficiency, we study the phase separation propensity of all 20 non-redundant trinucleotide repeat sequences. Our study aligns with findings fromin vivoexperiments, emphasizing that canonical base pairing and G–U wobble pairs play a dominant role in influencing the condensation of RNA repeat sequences and regulating the composition of ensuing condensates. Our implementation of this RNA model should make it feasible to efficiently probe sequence–phase behavior relationships for RNAs, and, when combined with coarse-grained protein models, elucidate the rules governing composition and dynamics of normal and aberrant biomolecular condensates.