Molecular modeling, mutational analysis and steroid specificity of the ligand binding pocket of mPRα (PAQR7): Shared ligand binding with AdipoR1 and its structural basis

Abstract

The 7-transmembrane architecture of adiponectin receptors (AdipoRs), determined from their X-ray crystal structures, was used for homology modeling of another progesterone and adipoQ receptor (PAQR) family member, membrane progesterone receptor alpha (mPRα). The mPRα model identified excess positively charged residues on the cytosolic side, suggesting it has the same membrane orientation as AdipoRs with an intracellular N-terminus. The homology model showed identical amino acid residues to those forming the zinc binding pocket in AdipoRs, which strongly implies that zinc is also present in mPRα. The homology model showed a critical H-bond interaction between the glutamine (Q) residue at 206 in the binding pocket and the 20-carbonyl of progesterone. Mutational analysis showed no progesterone binding to the arginine (R) 206 mutant and modeling predicted this was due to the strong positive charge of arginine stabilizing the presence of an oleic acid (C18:1) molecule in the binding pocket, as observed in the X-rays of AdipoRs. High Zn2+ concentrations are predicted to form a salt with the carboxylate group of the oleic acid, thereby eliminating its binding to the free fatty acid (FFA) binding pocket, and allowing progesterone to bind. This is supported by experiments showing 100 µM Zn2+ addition restored [3H]-progesterone binding of the Q206R mutant to levels in WT mPRα and increased [3H]-progesterone binding to mPRγ and AdipoR1 which have arginine residues in this region. The model predicts hydrophobic interactions of progesterone with amino acid residues surrounding the binding pocket, including valine 146 in TM3, which when mutated into a polar serine resulted in a complete loss of [3H]-progesterone binding. The mPRα model showed there is no hydrogen bond donor in the vicinity of the 3-keto group of progesterone and ligand structure-activity studies with 3-deoxy steroids revealed that, unlike the nuclear progesterone receptor, the 3-carbonyl oxygen is not essential for binding to mPRα. Interestingly, the small synthetic AdipoR agonist, AdipoRon, displayed binding affinity for mPRα and mimicked progesterone signaling, whereas D-e-MAPP, a ceramidase inhibitor, blocked progesterone signaling. Thus, critical residues around the binding pocket and steroid structures that bind mPRα, as well as similarities with AdipoRs, can be predicted from the homology model.