Correcting Pervasive Errors in RNA Crystallography with Rosetta

Stanford University, Stanford CA

Investigators

Linked publications & trials

Abstract

DESCRIPTION (provided by applicant): The continuing discoveries of non-coding RNAs and their critical roles in cellular and viral machinery are inspiring novel antibacterial, antitumor, nd antiviral therapies based on disrupting or manipulating the RNAs involved. Most of RNA's biological functions depend on the formation of intricate 3D structure and binding to ligands and proteins. Unfortunately, crystallographic models, our richest sources of RNA structural information, contain pervasive errors due to ambiguities in manually fitting RNA backbones into experimental density maps. We have recently brought Rosetta high- resolution RNA structure prediction together with PHENIX diffraction-based refinement and MolProbity validation, to create Enumerative Real-space Refinement ASsisted by Electron density under Rosetta. The ERRASER method corrects the majority of identifiable sugar pucker errors, steric clashes, suspicious backbone rotamers, and incorrect bond lengths/angles in a benchmark of RNA data sets, including a ribosomal subunit. Furthermore, the method, on average, improves Rfree factors to rigorously set- aside data. In this exploratory grant, we first aim to expand ERRASER to resolve ambiguities at RNA/ligand, RNA/protein, and RNA crystal contacts, as will be necessary for correcting RNA enzyme active sites, ligand binding sites, and ribonucleoprotein machines. Second, we aim to make ERRASER available as a fully automated server that will both refine all extant PDB-deposited RNA and ribonucleoprotein models and enable crystallographers to rapidly correct errors in their future data sets. By rapidly and systematicall disambiguating RNA model fitting, ERRASER will enable RNA crystallography with significantly fewer errors.