Alternative Conformations of lncRNAs Identified Through Structural Deconvolution of SHAPE- and DMS-MaP Datasets

The biological function of many classes of RNAs depend on their structures, which can exist as structural ensembles, rather than a single minimum free energy fold. In the past decade, long noncoding RNAs (lncRNAs) have emerged as functional transcripts in gene regulation that behave through their primary sequences and the structures they adopt. Chemical probing experiments, like selective 2-hydroxyl acylation analyzed by primer extension and mutational profiling (SHAPE-MaP), and dimethyl sulfate-MaP (DMS-MaP), facilitate the characterization of RNA secondary structure both inside and outside the cell. But chemical probing experiments yield an average reactivity profile, representative of all the structures a particular RNA transcript adopts at the time of chemical probing, weighted by their relative populations. Chemical probing experiments often struggle to identify coexisting conformations a lncRNA might sample. Computational methods (DRACO, DREEM, DANCE-MaP) have been developed to identify alternate conformations of RNAs by deconvoluting chemical probing data. In this work, we investigate the propensity for lncRNAs to sample multiple structured states, and find each of the studied lncRNAs possess coexisting folds. We discuss the implications of lncRNAs harboring multiple structures and how it may contribute to their multifunctionality in regulating biological processes.