A Structure Based Approach for Defining Ligand Interaction Requirements with ATP-Binding Cassette Proteins

Abstract

Membrane proteins belonging to the ATP-Binding Cassette (ABC) superfamily are responsible for transporting substrates into or out of cells. ABC transporters, such as P-glycoprotein (Pgp/ABCB1) defend cells by the active efflux of many chemically diverse cytotoxic compounds. The ABC protein, CFTR/ABCC7, is a passive chloride ion channel that provides fluid homeostasis. Pgp and CFTR of are high clinical interest since mutations in CFTR can result in Cystic Fibrosis (CF) and upregulation of Pgp in tumor cells can result in multidrug resistance (MDR). There has been a major effort to develop clinical drugs to treat CF by either repairing defects in CFTR (correctors) or holding the chloride channel in an open state for more time (potentiators). Efforts have also focused on developing inhibitors to Pgp to thwart MDR. Now that atomic structures of both CFTR and Pgp are available in the presence of multiple ligands, structure-activity relationships between ligand and protein can be probed in greater detail. I test a hypothesis that a universal amide linker present in the CFTR potentiator VX-770, and the CFTR corrector VX-809, when replaced with a triazole moiety, will maintain chemical and structural features that preserve modulation of CFTR function. To explore the hypothesis, I have solved sub-atomic x-ray crystal structures of eleven VX-770 and VX-809 analogs synthesized by collaborators and compared to structures I determined of their parent molecules. Triazole analogs of VX-770 were tested in CFTR potentiation assays, with at least one compound showing good efficacy. Drawbacks also became evident, and are discussed. To probe the mechanism of polyspecific ligand-recognition in Pgp, I tested the hypothesis that aromatic amino acids are a major determinant in the compensatory ligand-recognition mechanism intrinsic to Pgp. I determined the crystal structures of six strategic point-mutations (aromatic residue to alanine) of Pgp in the presence of a known inhibitor of ATPase activity. The inhibitor still bound to Pgp but occupied completely novel binding sites depending on the mutation, and the ligand still inhibited ATPase in all cases. My findings reveal that aromatic residues in the drug-binding pocket of Pgp are critical for defining polyspecific ligand recognition by Pgp.