Team:Heidelberg/Model/Pumilio PPR

Modular RBP modelling
Assembling RBPs on demand

Introduction

Overview

Rosetta is a macromolecular modeling suite for biomolecular structure prediction and design Das2008MacromolecularMW. Rosetta was originally developed for modeling proteins. Rosetta offers ab initio structure prediction as well as protein design ad docking. Methods in structure biology, such as NMR or X-ray analysis can be integrated into Rosetta structure prediction as well. Rosetta offers tools mainly for proteins, but also integrates nucleic acids.

The modeling and design in Rosetta depends on an energy score function. This score function depends on terms such as van der Waals forces, solvation and hydrogen bonding and. These forces yield from the conformation of a protein, nucleic acid and ligand and should the situation arise the docking.

An overview about the tools and protocols of Rosetta is given here.To start modelling it is also possible to make use of PyRosetta, which is a Python-interface for Rosetta Chaudhury2010PyRosettaAS . Pyrosetta is an easy starting point for Rosetta and documented comprehensively here. The ROSIE and Robetta server offer several Rosetta scripts and tools free and easy-to-use Lyskov2013ServerificationOM Kim2004ProteinSP

Protein-RNA complex modeling

Overview about RNA modeling

For RNA modeling Rosetta can also be used, although the number of applications are more limited than for protein modeling Das2010AtomicAI . Applications include de novo prediction of RNA and Protein-RNA complex modeling Kappel2019SamplingNS Watkins2019FARFAR2ID . Some of the applications may depend on ViennaRNA Lorenz2011ViennaRNAP2 . Beauchene2016BindingSI Beauchne2016FragmentbasedMO

Protein-RNA complex modeling.

3DOC enables us to concatenate fusion proteins or protein domains. We applied 3DOC on pumby as well as ppr modules Adamala2016ProgrammableRP Coquille2014AnAP. 3DOC generates using PyRosetta a PDB for these proteins. An example for a generated fusion pumby module can be seen below:

Figure 1: The generated pumby protein binding the RNA-motif "AC".
Here two pumby module binding the RNA-motif "AC" designed via "3DOC" and it's integrated PyRosetta tools is shown. For pumby sequences of the right length, as proposed by Adamala et al., the right relative conformation appears after relaxing the structure.


For further modeling, we are also offering two scripts (rna_denovo_preparation.py and rnp_structure_prediction_preparation.py), which generate the input for the RNA denovo (FARFAR) and RNP (Ribonucleicprotein) - Complex modeling protocols in Rosetta. With these the corresponding RNAs and Protein-RNA complexes can be modeled.

References