Palindrome Presence, Location, and Structural Context Are Predictive of RNA Self-Association

Abstract

RNA self-association is a key elementary event behind many cellular processes from ribosome assembly, to regulation of gene expression, to assembly of biomolecular condensates. While RNAs self-interact in many ways, one direct and unambiguous sequence-encoded pattern for RNA self-recognition is a palindrome, or palindromic motif. In isolation palindromic motifs tend to fold back on themselves, favoring stable stem-loop secondary structures. However, palindromes have been implicated in some cases of RNA self-recognition, segregation and phase separation, raising the question of which palindromes in a given RNA are actually available for intermolecular engagement. Toward answering this question, we define three classes of palindromic regions based on local secondary-structure context, and degree of availability for interactions via their nucleobases, sorting into three classes. We then apply this framework across known structures of RNA chains and find that the majority of palindromic regions are bound up in secondary structures. We then designed a minimal set of short stem-loop RNA sequences to evaluate the self-interaction stability of each defined palindrome class using coarse-grained molecular dynamics simulations. Overall, this work provides a residue-level, structure-aware framework that connects RNA sequence, secondary-structure context, and intermolecular self-association for two interacting RNA chains.