FRET-guided integrative modelling of non-coding ribonucleic acids

Richard Börner

The functional diversity of RNA is encoded in their innate conformational heterogeneity. The combination of single-molecule fluorescence spectroscopy and computational modeling offers new opportunities to map structural transitions within ribonucleic acid ensembles. We developed a streamlined workflow, which integrates de novo structure prediction through Rosetta paired with MD simulation using GROMACS followed by in silico FRET predictions with FRETraj [1] as a FRET-guided integrative modelling approach to capture RNA structural dynamics and to gain RNA folding pathways [2,3]. FRETraj enables to compute the accessible-contact volume (ACV) of the fluorescent dyes along MD trajectories or de novo generated structural ensembles (multi-ACV) without the need of explicit dye labeling in silico. FRETraj uses experimental measures such as fluorescence quantum yield, detection efficiency and intensity burst size distribution for FRET predictions yielding an exceptional agreement with the experimental derived FRET distribution. We use multi-ACVs and thus FRET as a post-hoc scoring method for fragment-assembly in Rosetta FarFar2 and we demonstrate that FRET effectively refines de novo RNA structure prediction.

We benchmark our FRET-assisted modeling approach on double-labeled DNA strands and validate it against an intrinsically dynamic Mn(II)-binding riboswitch [4] and a Mg(II)-sensitive ribosomal RNA tertiary contact [5]. We show that already one FRET coordinate, i. e., describing the assembly of a fourway junction and the GAAA binding to a kissing loop, allows to recapitulate the global fold of both, the riboswitch and the tertiary contact, and to significantly reduce the de novo generated structure ensemble.

We conclude that computational fluorescence spectroscopy facilitates the interpretability of dynamic structural ensembles and improves the mechanistic understanding of nucleic acid interactions.

References:

[1] F. D. Steffen et al., Bioinformatics, 37, 21, 3953-3955 (2021), https://doi.org/10.1093/bioinformatics/btab615.
[2] F. Erichson et al., Hochschule Mittweida, 2, 230–233 (2021), https://doi.org/10.48446/opus-12283.
[3] F. D. Steffen et al., Biorxiv, (2023), https://doi.org/10.1101/2023.08.07.552238.
[4] K. C Suddala et al., Nat. Commun., 10, 4304 (2019), https://doi.org/10.1038/s41467-019-12230-5.
[5] S. Gerhardy et al., Nat. Commun., 12, 4696 (2021), https://doi.org/10.1038/s41467-021-24964-2.