A computational method to estimate affinities between flexible proteins and small ligands

Author(s)

Ariane Ferreira Nunes Alves

• Document type: Master's Dissertation
• Press: São Paulo.
• Institution: Universidade de São Paulo (USP). Conjunto das Químicas (IQ e FCF) (CQ/DBDCQ)
• Defense date: 2013-05-06

Examining board members

Guilherme Menegon Arantes; Luis Gustavo Dias; Roberto Kopke Salinas

• Advisor: Guilherme Menegon Arantes

Abstract

Computational methods are used to generate protein-ligand complex structures and estimate their binding affinities. This work investigated how different representations of protein flexibility affect poses obtained by molecular docking and the affinities attributed to these poses. T4 lysozyme mutants L99A and L99A/M102Q were chosen as model systems. A descriptor for prediction of affinities based on linear interaction energy (LIE) approximation was parametrized specifically to lysozyme ligands and was used to estimate affinities. The protein was represented as a group of crystal structures or as structures from a molecular dynamics trajectory. OPLS-AA force field was used to model protein and ligands and the Generalized Born approximation was used to model solvent. The parametrized affinity descriptor resulted in average deviations between experimental and calculated affinities of 1.8 kcal/mol for a test set. Descriptor performance was satisfactory in the separation between crystal poses and false-positive ones and in the identification of false-positive poses. Clustering of complexes was tried out to reduce computational cost to estimate affinities, but results were poor. The best approximations to the implicit ligand theory proposed here in order to estimate affinities consider groups of receptor structures with the same weight. Ligand configurations also have the same weight or are dominated by only one configuration. The representation of protein flexibility requires an adequate statistical treatment when used to estimate affinities. Here, the linking between LIE and the implicit ligand theory proved itself useful. (AU)

• FAPESP's process: 11/04354-6 - Computational study on the role of protein flexibility in the recognition of small molecules
• Grantee: Ariane Ferreira Nunes Alves
• Support Opportunities: Scholarships in Brazil - Master