Hands-off Linear Interaction Energy Approach to Binding Mode and Affinity Estimation of Estrogens

• Published in: Journal of chemical information and modeling Vol. 53; no. 10; pp. 2681 - 2688
• Main Authors: Durmaz, Vedat, Schmidt, Sebastian, Sabri, Peggy, Piechotta, Christian, Weber, Marcus
• Format: Journal Article
• Language: English
• Published: Washington, DC American Chemical Society 28.10.2013
• Subjects: Analytical, structural and metabolic biochemistry; Applied sciences; Benzhydryl Compounds - chemistry; Binding Sites; Biological and medical sciences; Computer science; control theory; systems; Correlation analysis; Crystallography; Crystallography, X-Ray; Data processing. List processing. Character string processing; Databases, Protein; Entropy; Estradiol - chemistry; Estrogen Receptor alpha - chemistry; Estrogens; Exact sciences and technology; Fundamental and applied biological sciences. Psychology; General aspects, investigation methods; Genistein - chemistry; Humans; Kinetics; Ligands; Memory organisation. Data processing; Models, Molecular; Molecules; Monte Carlo Method; Monte Carlo simulation; Phenols - chemistry; Protein Binding; Proteins; Software; Tamoxifen - analogs & derivatives; Tamoxifen - chemistry; Thermodynamics; Energy method; Estrogen receptor; Crystallographic direction; Ligand; Cross validation; Experimental study; Network interfaces; Modeling; Free energy; Protein; Thermodynamics; Model matching; Binding energy; Classification; System identification; Bioinformatics; Interaction energy
• ISSN: 1549-9596; 1549-960X
• DOI: 10.1021/ci400392p

With this work we target the development of a predictictive model for the identification of small molecules which bind to the estrogen receptor alpha and, thus, may act as endocrine disruptors. We propose a combined thermodynamic approach for the estimation of preferential binding modes along with corresponding free energy differences using a linear interaction energy (LIE) ansatz. The LIE model is extended by a Monte Carlo approach for the computation of conformational entropies as recently developed by our group. Incorporating the entropy contribution substantially increased the correlation with experimental affinity values. Both squared coefficients for the fitted data as well as the more meaningful leave-one-out cross-validation of predicted energies were elevated up to r Fit 2 = 0.87 and q LOO 2 = 0.82, respectively. All calculations have been performed on a set of 31 highly diverse ligands regarding their structural properties and affinities to the estrogen receptor alpha. Comparison of predicted ligand orientations with crystallographic data retrieved from the Protein database pdb.org revealed remarkable binding mode predictions.