Modeling protein structure based on constraints obtained from chemical cross-linking and mass spectrometry

Author(s)

Állan Jhonathan Ramos Ferrari

• Document type: Doctoral Thesis
• Press: Campinas, SP.
• Institution: Universidade Estadual de Campinas (UNICAMP). Instituto de Química
• Defense date: 2019-09-26

Examining board members

Fábio Cesar Gozzo; Tatiana de Arruda Campos Brasil de Souza; Adriana Franco Paes Leme; Denize Cristina Favaro

• Advisor: Fábio Cesar Gozzo; Leandro Martínez

Abstract

Chemical cross-linking/mass spectrometry is an experimental method that allows one to obtain distance constraints between amino acid residues. These constraints, in turn, can be used to investigate the tertiary and quaternary structure of biomolecules. In principle, these constraints provide only an upper limit along the surface of the biomolecule. Although there is great success on the use of this technique for structural characterization of protein complexes, attempts to use such constraints to determine the tertiary structure of proteins have not been successful. This indicates the need of specifically designed strategies for the representation of these constraints within modeling algorithms. In this thesis, we develop TopoLink, a package for structural model evaluation with cross-linking data. TopoLink shows superior results compared to previous software described in the literature and is made freely available at http://m3g.iqm.unicamp.br/topolink as source code with a user-friendly graphical interface for Windows. TopoLink was used to compute the probability of satisfying topological distance of a cross-linking specie in high-resolution structures as a function of the Euclidean distance between the residues involved. These probability distributions are then converted in a set of potential energy functions dependent on the cross-linker length and the amino acid residue pairs, generating the first cross-linking force field (XLFF). As the potential is described in terms of Euclidean distance, it can be easily incorporated in most current methods and software available. The force field was implemented, and it is distributed, to be used with Rosetta ab initio protocol. The strategy developed shows that the upper limits of distance constraints should be shorter than it is usually used in the literature. The benchmark test of 19 protein targets of various sizes and topologies shows that the complete force field expressively improves the quality of models obtained in comparison with previous heuristic strategies of representation. We also demonstrate the improvement associated with considering the experimental constraints from sampling the conformation neighborhoods of the crystallographic structure. These results bring to reality the possibility of modeling from XLMS constraints the tertiary structures of proteins, especially for those which other structural data is not available or is insufficient to characterizing the protein fold (AU)

• FAPESP's process: 16/13195-2 - Modeling of protein structure and protein complexes using mass spectrometry data
• Grantee: Allan Jhonathan Ramos Ferrari
• Support Opportunities: Scholarships in Brazil - Doctorate