Periodic Reporting for period 1 - HoogsCG (Development of a multiscale modeling strategy to decipher how hybrid DNA/RNA triplexes and G-quadruplexes affect gene expression regulation)
Reporting period: 2015-05-01 to 2017-04-30
This project deals with developing and applying computational tools to acurately describe complex biomolecular systems. Accurate, reproducible and well-converged computational results are still challenging, specially for systems containing large number of particles. This is crucially important to elucidate the interplay between structure and dynamics of large complex molecules in solution at atomistic level, for instance those responsible for maintaining genomic integrity.
This project is important for society as it allows robust and accurate predictions on the behaviour of key players in fundamental biological processes, with direct implications with both basic research and potential biomedical applications.
The overall objective was (1) to create a refined modelling strategy for describing nucleic acids, in the shape of new so-called force-field potentials, and incorporating computational strategies to enhance the description of such systems. Once this was developed, we used the refined set of potentials to (2) study the conformational landscape of DNA-protein complexes, as well as the folding and dynamics of quadruplex DNA. Finally, we (3) developed an accurate predictor for the stability of RNA-DNA2 triplexes, and applied it to predict the formation of such triplexes in vivo.
As a final conclusion, we have successfully aided the development of a refined modelling strategy for nucleic acids (see publication entitled ""PARMBSC1: A refined force-field for DNA simulations"", and ""How accurate are accurate force-fields for B-DNA?""), co-developed a multi-scale approach to describe how ""Chromatin unfolding by epigenetic modifications (is) explained by dramatic impairment of internucleosome interactions"", as well as help determine the structure and dynamics of DNA-repair inducing protein RNF169, which recognises ubiquitylated chromatin."
During the initial phases of the project, Dr. Portella has collaborated with the group of Prof. Orozco in Barcelona to aid the refinement of an already very successful force-field for nucleic acids, the newest iteration being Parmbsc1 (resulting in publication doi:10.1038/nmeth.3658). Thanks to the expertise gained in the host organisation, specifically in the lab of Prof. Vendruscolo, Dr Portella performed extensive validations on the new set of proposed force-field parameters by means of comparison with NMR observables. These calculations were at times enhanced by special sampling techniques developed at the lab of Prof. Vendruscolo in Cambridge, notably Bias-exchange metadynamics with experimental residual dipolar couplings restraints. We continued the collaboration to produce a comprehensive comparison of competing so-called accurate nucleic acid force-fields, a computational study that was thoroughly validated against newly generated NMR data for a duplex DNA (resulting in publication doi:10.1093/nar/gkw1355). Also in cooperation with Prof. Orozco, Dr Portella developed a sequence, concentration and pH dependent predictor for the stability of RNA-DNA2 triplexes, which was applied to discover gene-regulation roles of mRNA - genomic DNA interaction in vivo.
We also applied the combination of RDC-enhanced bias-exchange metadynamics, along with the new Parmbsc1 parameters, to describe simultaneously the structure and dynamics of DNA G-quadruplexs. This work is under way, as we are currently preparing a manuscript for its publication.
During the duration of the action, Dr Portella co-authored a manuscript that details how acetylation of certain histone tails, particularly H4 lys-16, leads to a decompaction of the chromatin fiber (publication doi: 10.1021/jacs.5b04086). This work is important to understand from first principles, how changes in chromatin organization linked to regulation of fundamental template-directed functions such as DNA transcription, replication, and repair. In our work, we have used a novel combination of coarse-grain models to describe long chromatin fibers, all-atoms multi microsecond simulations of dinucleosome systems and chemical-shift restrained enhanced simulations of histone tails. Finally, similar methodologies were applied to the last manuscript published so far during the fellowship, a collaborative work between experimental and computation biophysicists to elucidate the mechanism of action of protein RNF169, which binds ubiquitinated nucleosomes during a pathway of DNA repair. It was found that RNF169 binds to both ubiquitin on H2A and an acidic patch on the nucleosome to ensure that this protein only promotes DNA repair only when required.
Crucial to all these works were the methodologies developed in the lab of Prof. Vendruscolo, instrumental in enhancing the repertoire of modelling tools of Dr Portella and further advancing the researcher’s career. Likewise, constant interactions with researchers from the department of Chemistry in Cambridge University, e.g. the group of Prof. Wales, Prof. Frenkel and Prof. Balasubramanian, has been extremely beneficial for the development of Dr Portella’s skills.
Dissemination of the results of this project has been very effectively carried out by means of publications in major peer-reviewed journals. Both the work on Parmbsc1 and H4K16-ac mediated decompaction of nucleosomes were disseminated via “popular science” news outlets. The set of newly derived nucleic acids parameters is distributed freely, the link is available from Dr. Portella’s projects public website.