Modeling Small Noncanonical RNA Motifs with the Rosetta FARFAR Server

Abstract

Noncanonical RNA motifs help define the vast complexity of RNA structure and function, and in many cases, these loops and junctions are on the order of only ten nucleotides in size. Unfortunately, despite their small size, there is no reliable method to determine the ensemble of lowest energy structures of junctions and loops at atomic accuracy. This chapter outlines straightforward protocols using a webserver for Rosetta Fragment Assembly of RNA with Full Atom Refinement (FARFAR) ( http://rosie.rosettacommons.org/rna_denovo/submit ) to model the 3D structure of small noncanonical RNA motifs for use in visualizing motifs and for further refinement or filtering with experimental data such as NMR chemical shifts.

Keywords: RNA 3D structure prediction; RNA Motifs.

Fig. 1

(Left) secondary structure of GCAA tetraloop. (Right) 3D structure of GCAA tetraloop (PDB: 1ZIH)

Fig. 2

Main page of the FARFAR (RNA De Novo) webserver. Here the user can enter a sequence and secondary to submit a job to generation an all atom model of their construct

Fig. 3

Example chemical shift data. Column description is as follows. (1) Atom entry number. (2) Residue author sequence code. (3) Residue sequence code. (4) Residue label. (5) Atom name. (6) Atom type. (7) Chemical shift value. (8) Chemical shift value error. (9) Chemical shift ambiguity code

Fig. 4

The status page for a submitted FARFAR (RNA De Novo) job

Fig. 5

Results page for a RNA De Novo job

Fig. 6

(a) GCAA tetraloop (1ZIH): RNA De Novo (through fragment assembly of RNA with full atom refinement, FARFAR) gives lowest energy models displaying structural convergence. (b) Pseudoknot (1L2X) [27], less converged then tetraloop–but also a larger RNA–gives models that are still within 3 Å heavy-atom RMSD for top model. (c) 4 × 4 internal loop solved by NMR at PDB ID 2L8F [28], converges despite presenting four noncanonical base pairs. (d) Tandem GA (1MIS) [26] without application of ¹H chemical shifts. (e) Tandem GA with ¹H chemical shifts, demonstrates the improved convergence with the addition of ¹H chemical shift