Structure-function relationships in ABCG2: insights from molecular dynamics simulations and molecular docking studies

Abstract

Efflux pumps of the ATP-binding cassette transporters superfamily (ABC transporters) are frequently involved in the multidrug-resistance (MDR) phenomenon in cancer cells. Herein, we describe a new atomistic model for the MDR-related ABCG2 efflux pump, also named breast cancer resistance protein (BCRP), based on the recently published crystallographic structure of the ABCG5/G8 heterodimer sterol transporter, a member of the ABCG family involved in cholesterol homeostasis. By means of molecular dynamics simulations and molecular docking, a far-reaching characterization of the ABCG2 homodimer was obtained. The role of important residues and motifs in the structural stability of the transporter was comprehensively studied and was found to be in good agreement with the available experimental data published in literature. Moreover, structural motifs potentially involved in signal transmission were identified, along with two symmetrical drug-binding sites that are herein described for the first time, in a rational attempt to better understand how drug binding and recognition occurs in ABCG2 homodimeric transporters.

Aberrant DNA Methylation of ABC Transporters in Cancer

“…ABCG2 is a half transporter, consisting of only one NBD and one TMD. It has been found to form homodimers or even higher order oligomers to become fully functional [ 66 , 67 ].…”

Residues contributing to drug transport by ABCG2 are localised to multiple drug-binding pockets

“…Data regarding the effects of mutation on the transport of multiple ABCG2 substrates can be rationalised against structural data to provide a molecular explanation for the effects observed. Recently, many structural models of ABCG2 have become available based upon either homology to the ABCG5/G8 crystal structure, or determined directly from cryo-EM data [ 27 – 30 ]. We performed docking studies of MX onto the homology model of ABCG2 described by Stockner and Kuchler [ 29 ], preferring this to the cryo-EM structure [ 27 ] for two reasons.…”

“…This would require solvating the ABCG2 model in a representative lipid bilayer, and would also require us to generate equivalent models of the single alanine mutations prior to any docking studies. Such work is beyond the scope of the current investigation, but our docking with MX in vacuo does permit comparison to other recent studies describing MX binding sites in ABCG2 [ 28 – 30 ]. In their homology modelling paper on ABCG2, László et al describe four possible binding sites, two of which (referred to as site 2 and 3 in [ 28 ]) contain several of our investigated residues.…”

“…Low-resolution electron microscopy of the protein was, until very recently, the only source of structural information we had for the protein [ 23 – 25 ]. Since the commencement of this study, our structural knowledge of ABCG2 has been bolstered by publications of an X-ray crystallographic structure of the related ABCG5/G8 heterodimeric transporter [ 26 ], a medium resolution cryoelectron microscopy structure of ABCG2 complexed with an inhibitory antibody [ 27 ], and homology models of ABCG2 built from these structural templates [ 28 – 30 ]. Such models provide a valuable framework for data interpretation, but currently do not provide an unambiguous understanding of drug binding and release sites on ABCG2.…”

Does human P-glycoprotein efflux involve transmembrane alpha helix breakage?

“…Under certain cryo-EM conditions, the internal lipid dynamics provided by the nanodisc may render unfavorable conditions for the structural stability of the transporters’ architecture, leading to severe distortions of TM4 and TM10 helices. This also upholds the importance of a careful choice of the experimental conditions and protocol to be used to obtain reliable models of membrane proteins since these cryo-EM structures are establishing new and improved starting points for improved computer simulations [21–23].…”