RNA Structure modeling GDR MASIM Paris, 16-17 th November 2017 Bruno Sargueil, CNRS UMR 8015 Faculté de pharmacie – PARIS Bruno.sargueil@parisdescartes.fr
RNA structure • Primary structure 5’ aaaaagcaaaaatgtgatcttgcttgtaaatacaattttgagaggttaataaattacaagtagtgcta tttttgtatttag gttagctatttagctttacgttccagg atgcctagtg gcagccccac aatatccagg aagccctctctgcggttttt 3’ • Secondary structure • Tertiary structure
Nucleotide interactions
Some examples of tertiary motifs
RNA structure modeling • Secondary structure modeling is a limiting step • Structure modeling software (Mfold, RNAfold …) are based on : • Thermodynamic – experimental data have defined a free energy for a bp in a given context (nearest neighbour theory) • Probability (Boltzmann statistics) • Such modeling is often inexact if the RNA is over > 50(ish) nucleotide long • Thermodynamic model is incomplete • Does not predict non canonical base pairs – pseudoknots • Does not take into account folding kinetics • A single RNA may adopt several foldings • Yields several models – how to choose?
Generating folding constraints Goal: Experimentally define nucleotides that are in single strand conformation • Single stand RNAse : T1, A, S1 etc … • Small molecules: DMS, CMCT, SHAPE reagents
Probing RNA structure The reactivity map is used as (soft) constrains by the modeling software (Bonus/penalty)
Is this enough ? Modeling using probing structure data X-Ray structure Modeling without constraint Not predicting the tertiary structure impairs the 2D prediction
Multiple probes Each molecular probe brings different information The experimental process has been entirely automated Use a multiprobing approach to improve modeling
Multiprobing approach
Multiprobing approach
Multiprobing approach
Multiprobing approach
Developpement of a new model that takes into account all the probing results Currently validating the approach on a « benchmark » RNA
Detecting the tertiary structure Reactivity low medium high Probing the structure in presence/absence of Mg2+ can reveal tertiary contacts We are currently developping approaches to predict pseudoknots using such data
Combining Probing and NGS
Naive probing of multiple mutants • • •
Loop IIId is crucial for HCV IRES/40S binding H BK Loop IIId B Adapted from Hashem et al. 2014 (CryoEM envelop) Base pairing (kissing complex) between loop IIId (HCV IRES) and ES7 (18S rRNA) favours the 40S recruitment, and is required for an efficient translation
IRES structure and structural rearrangment SHAPE reactivity > 0.7 N Less N 0,5 - 0,7 More Reactive in presence of 40S N 0,2 - 0,5 N 0 - 0,2 WT N Undetermined Loop IIId is protected from modification Domain IV unfolds
IRES footprinting on the18S rRNA ES7 ES7 rRNA loop is protected Pattern modifications are observed in other sites
Fitting the atoms in the envelop Coordinates from Hashem et al. 2013 Model by Benoit Masquida
3D model of the IRES-40S CryoEM envelop Hashem et al. 2013 Angulo et al. 2016 Model by Benoit Masquida
People involved LIX – Ecole Polytechnique Nathalie Chamond (CR CNRS) Yann Ponty (CR CNRS) Christelle Vasnier (AI CNRS) Afaf Saaidi (PhD Student) Delphine Allouche (PhD Student) Grégoire de Bisschop (M2 student) CIRI Pontificia Universidad Católica ENS Lyon de Chile T.Ohlmann M.Lopez-Lastra S. De Breyne M.Vallejos C. Herbreteau J. Angulo F. Carvajal CNRS UMRR7156 Strasbourg B. Masquida
Recommend
More recommend
