A biologically supported global RNA architecture of cucumber mosaic virus satellite RNA
Listed in
This article is not in any list yet, why not save it to one of your lists.Abstract
Satellite RNAs (satRNAs) are parasitic subviral agents whose biological activities are mediated largely by specific sequence determinants and structured RNA elements. Establishing biologically supported global RNA structures is therefore essential for understanding how RNA architecture underlies satRNA functions. Cucumber mosaic virus (CMV) satRNA is one of the best-characterized models for investigating satRNA structure-function relationships; however, a biologically supported global RNA architecture of CMV satRNA has yet to be established. Here, we applied AlphaFold3 modeling to predict the global structure of CMV satRNA T1 (sat-T1). Initial full-length structure modeling generated multiple long-distance interactions that lacked biological support. We therefore used fragment-based modeling, combined with chemical probing, evolutionary covariation, and compensatory mutagenesis, to derive a biologically supported global secondary structure. To determine whether structurally distant regions could interact in the context of the full-length RNA, we engineered a structure-guided T1-ZD mutant that preserved the supported secondary structure while reducing alternative base-pairing possibilities. Full-length AlphaFold3 modeling of T1-ZD largely recapitulated the proposed architecture, while one predicted model revealed a long-distance interaction that was subsequently supported by compensatory mutagenesis analysis. These findings suggest that the 3′ terminus of sat-T1 may undergo conformational switching between alternative structural states. Together, our work establishes a biologically supported global RNA architecture for CMV sat-T1 and provides a structural framework for investigating the molecular basis of satRNA function.
AUTHOR SUMMARY
Satellite RNAs are small RNA molecules that depend on viruses for replication, yet they can strongly influence virus infection and disease symptoms in plants. Their biological activities are often determined by how the RNA folds into specific structures. However, understanding the overall structure of a long RNA remains challenging. Here, we established a biologically supported global structure for a satellite RNA associated with cucumber mosaic virus. Initial prediction of the complete RNA sequence using AlphaFold3 generated long-distance interactions that were not supported by biological evidence. Thus, we adopted the fragment-based strategy to generate an integrated global RNA model, which was supported by multiple experimental evidence. Finally, we engineered a mutant guided by the predicted structure and identified a biologically relevant long-distance interaction in the prediction models of the full-length mutant RNA. Our study provides a structural basis for understanding satellite RNA function and illustrates how experimentally supported RNA structures can guide artificial intelligence-assisted prediction of long RNA structures.
