Probing the binding mechanism of substituted pyridine derivatives as effective and selective lysine-specific demethylase 1 inhibitors using 3D-QSAR, molecular docking and molecular dynamics simulations

Lysine-specific demethylase 1 (LSD1) was regarded as a promising anticancer target for the novel drug discovery. In this work, we carried out a molecular modeling study on the substituted pyridine derivatives as LSD1 inhibitors using three-dimensional quantitative structure-activity relationship (3D-QSAR), molecular docking and molecular dynamics (MD) simulations. Molecular docking studies predicted the probable binding mode of ligands, and suggested Lys661 and Asp555 might be key residues. Our 3D-QSAR models exhibited satisfactory internal and external predicted capacity. For the comparative molecular field analysis (CoMFA) model, its training set had q2 of 0.595 and r2 of 0.959, while test set had q2 of 0.512 and r2 of 0.846. For the best comparative molecular similarity indices analysis (CoMSIA) model, its training set had q2 of 0.733 and r2 of 0.982, while test set had q2 of 0.695 and r2 of 0.922. MD simulations result revealed the detailed binding process and found an important conserved water-bridge motif between ligands and protein. The binding free energies calculation using Molecular Mechanics Poisson–Boltzmann Surface Area (MM-PBSA) approach coincided well with the experimental bioactivity and demonstrated that the electrostatic interaction was the major driving force for binding. The energy decomposition pointed out some significant residues (Asp555, Lys661, Trp695, Tyr761 and FAD) for the LSD1 potency increase. Based on these results, five new inhibitors were designed, and their activities were predicted using our 3D-QSAR models. Communicated by Ramaswamy H. Sarma Abbreviations3D-QSARthree-dimensional quantitative structure-activity relationshipADapplicability domainAMLacute myeloid leukemiaCoMFAComparative molecular field analysisCoMSIAComparative molecular similarity indices analysisDNMTsDNA methyltransferasesE2F1E2F transcription factor 1ESPElectrostatic potentialsFADFlavin adenine dinucleotideGAFFGeneral Amber force fieldH3K4Histone 3 lysine 4H3K9Histone 3 lysine 9LOOLeave-one-outLSD1Lysine-specific demethylase 1MAO-AMonoamine oxidase AMAO-BMonoamine oxidase BMAEMean absolute errorMDMolecular dynamicsMLLmixed lineage leukemiaMM-PBSAMolecular Mechanics Poisson–Boltzmann Surface AreaPLSPartial least squarePMEParticle Mesh EwaldRMSDRoot-mean square deviationRMSERoot-mean square errorRMSFRoot-mean-square fluctuationsSASASolvent accessible surface areaSEEStandard error of estimateTCPTranylcypromine three-dimensional quantitative structure-activity relationship applicability domain acute myeloid leukemia Comparative molecular field analysis Comparative molecular similarity indices analysis DNA methyltransferases E2F transcription factor 1 Electrostatic potentials Flavin adenine dinucleotide General Amber force field Histone 3 lysine 4 Histone 3 lysine 9 Leave-one-out Lysine-specific demethylase 1 Monoamine oxidase A Monoamine oxidase B Mean absolute error Molecular dynamics mixed lineage leukemia Molecular Mechanics Poisson–Boltzmann Surface Area Partial least square Particle Mesh Ewald Root-mean square deviation Root-mean square error Root-mean-square fluctuations Solvent accessible surface area Standard error of estimate Tranylcypromine