Tying the Knot: Unraveling the Intricacies of the Coronavirus Frameshift Pseudoknot

Abstract

Understanding and targeting functional RNA structures towards treatment of coronavirus infection can help us to prepare for novel variants of SARS-CoV-2 (the virus causing COVID-19), and any other coronaviruses that could emerge via human-to-human transmission or potential zoonotic (inter-species) events. Leveraging the fact that all coronaviruses use a mechanism known as -1 programmed ribosomal frameshifting (-1 PRF) to replicate, we apply algorithms to predict the most energetically favourable secondary structures (each nucleotide involved in at most one pairing) that may be involved in regulating the -1 PRF event in coronaviruses, especially SARS-CoV-2. We compute previously unknown most stable structure predictions for the frameshift site of coronaviruses via hierarchical folding, a biologically motivated framework where initial non-crossing structure folds first, followed by subsequent, possibly crossing (pseudoknotted), structures. Using mutual information from 181 coronavirus sequences, in conjunction with the algorithm KnotAli, we compute secondary structure predictions for the frameshift site of different coronaviruses. We then utilize the Shapify algorithm to obtain most stable SARS-CoV-2 secondary structure predictions guided by frameshift sequence-specific and genome-wide experimental data. We build on our previous secondary structure investigation of the singular SARS-CoV-2 68 nt frameshift element sequence, by using Shapify to obtain predictions for 132 extended sequences and including covariation information. Previous investigations have not applied hierarchical folding to extended length SARS-CoV-2 frameshift sequences. By doing so, we simulate the effects of ribosome interaction with the frameshift site, providing insight to biological function. We contribute in-depth discussion to contextualize secondary structure dual-graph motifs for SARS-CoV-2, highlighting the energetic stability of the previously identified 3 8 motif alongside the known dominant 3 3 and 3 6 (native-type) -1 PRF structures. Integrating experimental data within minimum free energy (MFE) hierarchical folding algorithms provides novel structure predictions to distill the relationship between RNA structure and function. In particular, fully categorizing most stable secondary structure predictions via hierarchical folding supports our identification of motif transitions and critical site targets for future therapeutic research. Author summaryFinding evolutionary connections between coronaviruses frameshift element RNA structures is a worthwhile goal in contributing to treatment development for afflicted human and animal populations. Predicting the most energetically favourable RNA secondary structures, and how they may form via the hierarchical folding hypothesis, is an efficient use of computational resources to shed light on RNA structure-function. We used the KnotAli algorithm to obtain mutual information from 181 coronaviruses frameshift RNA sequences. Guided by this evolutionary information, we computed secondary structure predictions to allow comparison of marked similarities and subtle differences between SARS-CoV-2 and other coronaviruses frameshift element RNA structures. In addition, we applied the Shapify algorithm to predict secondary structures for extended SARS-CoV-2 frameshift element sequences informed by SHAPE reactivity data. Here we critically expand the known landscape of most stable -1 PRF secondary structure conformations, isolating the location of key secondary structure motif transitions that can improve site targeting of viral therapeutics. Our application of hierarchical folding algorithms contributes novel predictions of functional RNA structures, enhancing discussion of how secondary structures unfold or refold to regulate frameshifting in coronaviruses.