Docking to homology models highlights the molecular determinants of ligand binding to the AhR

Abstract Molecular docking studies can shed light into the molecular determinants of ligand binding. When no experimental structures are available, docking can also be applied to homology models. Given that the model quality, especially in the binding site, greatly affects the accuracy of docking predictions [1], development of strategies able to include protein flexibility [2] may help in the effective use of docking to homology models. In this work, ligand binding of structurally different ligands to the mouse Aryl hydrocarbon Receptor (mAhR) homology model is analyzed. Introduction AhR is a ligand-dependent transcription factor that responds to exogenous and endogenous chemicals with the induction of gene expression and production of diverse biological and toxic effects [3]. The mechanism is initiated by ligand binding to the AhR, which is present in the cytosol as a multiprotein complex, and the PAS-B domain acts as ligand binding domain (LBD). The most “classical” AhR ligands are the exogenous halogenated aromatic hydrocarbons (HAHs) (including the most toxic 2,3,7,8-tetrachlorodibenzo-p-dioxin, TCDD), polycyclic aromatic hydrocarbons (PAHs), and PAH-like chemicals; but also “non-classical” natural, endogenous and synthetic AhR ligands with diverse structure and physico-chemical characteristics have been identified [3]. Given that AhR shows different biological responses for different ligands, understanding the molecular determinants of binding could help in elucidating the mechanism of action associated to each ligand. Methods No experimental information is available for the AhR LBD but since last years the X-ray structures of many homologous PAS domains have been determined, in particular the Hypoxia-Inducible Factor 2α (HIF2α) shows nearly 30% of sequence identity with the mAhR LBD. 10 homology models of the mAhR LBD were developed using MODELLER, each one based on a different HIF2α template structure in order to describe the flexibility of the binding cavity. All the ligands and solvent molecules in the HIF2α internal cavity were maintained during the modeling and refinement stages. Docking was performed using Glide XP, that treats ligands as flexible. An ensemble docking approach based on ligand docking to the 10 LBD models was used to include receptor flexibility. Selected binding poses were subjected to Molecular Dynamics simulation (10ns–20ns). The binding free energy (ΔGbind) was calculated by the MM-GBSA approach implemented in AMBER and the most stabilizing contributions were obtained by per-residue decomposition analysis. Results & Conclusions The set of 12 ligands here analyzed is composed by both classical and non-classical AhR ligands with diverse structures and physico-chemical characteristics. Protein flexibility and plasticity were introduced in both the ensemble docking to multiple homology models and the post-docking MD simulations to obtain an accurate description of binding of such diverse chemicals within the binding cavity of the receptor. The ΔGbind analysis, performed to identify the most relevant residues for each pose stabilization, allowed to discriminate the molecular determinants for binding of the different ligands. Three main arrangements within the binding cavity were identified: hydrophobic molecules interact mainly with residues located at the bottom of the cavity (TCDD); others occupy the whole cavity and in some cases are stabilized by hydrogen bonds with residues in the middle of the cavity (FICZ); the smallest and polar ligands stay at the entrance of the cavity (leflunomide). These poses will be validated by a mutagenesis study focused on the key stabilizing residues predicted for each ligand by the computational protocol here presented 1. Bordogna A, Pandini A & Bonati L, J Comput Chem 32, 81 (2011) 2. Fan H, Irwin J & Webb B, J Chem Inf Model 2512 (2009) 3. Denison MS, Soshilov AA, He G, DeGroot DE & Zhao B, Toxicol Sci 124, 1 (2011)