A Single Amino Acid in ABCG2 is Essential for Clamping Substrates and Inhibitors into the Binding Pocket

Abstract

ABCG2 is an ATP‐binding cassette transporter that, by exporting diverse structurally unrelated anticancer drugs, was identified as a strong contributor to cancer cell multidrug resistance. Subsequently, it was revealed to have important physiological roles in restricting the penetration of multiple xeno‐ and endobiotics into cells and playing a protective role in vivo against cytotoxins for susceptible host tissues. Although ligand binding is the obligate initial step in the transport cycle, it is unknown if any one residue has an indispensable role in binding either substrates and/or inhibitors. The ABCG2 structural studies indicate that substrates interact with fewer residues in the binding pocket than do inhibitors. This finding suggests that a single residue that strongly impacts the binding of both substrates and inhibitors (we will refer to as ligands) would not be found.To evaluate ABCG2 ligand binding, we performed a thermal shift assay using isolated plasma membranes. Among the tyrosine kinase inhibitors (ABCG2 ligands) tested, lapatinib showed the strongest stabilization of ABCG2 against thermal denaturation. Structure‐based docking identified potential lapatinib binding residues. With this knowledge as a guide, mutagenesis and biochemical studies (thermal shift, viability, and transport assays), revealed that a single highly conserved phenylalanine residue was critical for binding, not just lapatinib, but all tested substrates and inhibitors. The alanine substitution for phenylalanine resulted in the loss of ABCG2 interaction with all ligands tested. Conversion of the phenylalanine to amino acids with similar planar, aromatic side‐chain rings, i.e., tryptophan or tyrosine, restored binding and transport activity comparable to wildtype ABCG2. Importantly, amino acid substitution at other individual binding pocket residues did not result in a complete loss of ABCG2 binding and function. Structural modeling using ABCG2 (PDB: 6ETI) showed that the π‐π interactions from the phenylalanine of each monomer were capable of mediating the binding of a surprisingly diverse array of structurally unrelated therapeutic substrates and inhibitors. Apparently, this symmetrical π‐π interaction “clamps” the ligand into the binding pocket. From the molecular features of ABCG2 ligands that use the π‐π clamp, along with structural studies, we created a “global” ABCG2 pharmacophore model that will be invaluable in predicting ABCG2 interacting drugs and xenobiotics.These findings have important therapeutic implications by revealing a key structural feature that ABCG2 uses to engage and initiate transport of multiple drugs and xenobiotics.Support or Funding InformationThis work was supported by NIH and by ALSAC.