Androgen Receptor

Endocrine disruption: Estimation of the binding affinity of small molecules towards the Androgen receptor.

Based on the three-dimensional structure of the androgen receptor with bound dihydrotestosterone (PDB code = 1I37), we have docked 122 compounds comprising eleven molecular classes and covering a range of six orders of magnitude in IC50 (1.6 nM-46 mM). Each complex was fully optimized using the Yeti software and the binding modes were identified by Monte-Carlo sampling (software Yeti) and molecular-dynamical simulations (software AMBER). The resulting orientations and conformations were then composed into a 3D pharmacophore and used as input for the multi-dimensional QSAR software Raptor which is based on a novel scoring function including hydrophobicity, hydrogen bonding, entropy and the quantification of explicitly simulated induced fit. The philosophy underpinning the new concept is a dual-shell representation of the binding-site surrogate, allowing to simulate anisotropic effects.

Receptor model as generated by Raptor 1.2 (dual-shell QSAR) with ligand Mibolerone shown.

The simulation reached a classic r² of 0.867 and yielded a predictive r² of 0.754. The maximal deviation from the experiment corresponds to a factor 13 in IC50 for the ligands of the training set and to a factor 17 for the ligands of the test set. As for the estrogen receptor, we are now in a position to scan any hypothetical or real compound for potential activity towards the AR.

Comparison of experimental and predicted IC50 values for the androgen receptor (Quasar simulation).
(training set = black, test set = red)

Reference: J. Med. Chem. 2005, 48, 5666–5674. View abstract