We have focused our efforts to elucidate: i) the catalytic cycle and transport pathway of human P-gp; ii) the molecular basis of the polyspecificity of P-gp; iii) the interaction of clinically relevant tyrosine kinase inhibitors (TKIs) with P-gp and ABCG2; iv) determination of the binding site of nilotinib, a second generation TKI, on P-gp; v) pharmacophore features required for binding of nilotinib to P-gp and ABCG2; and vi) the use of single particle cryo-EM for the analysis of the conformational landscape of human P-gp during its catalytic cycle. We have been employing cell-based, biochemical, biophysical, pharmacological, and physiological techniques along with molecular biology and molecular modeling approaches to extend our understanding of the mechanistic aspects and the structure-function relationships of ABC drug transporters. In addition, we have devoted considerable effort to the screening and development of TKIs and small molecule modulators for P-gp and ABCG2. We found that several tyrosine kinase inhibitors, which are used in the clinic for treatment of various types of cancers, are either transport substrates or inhibitors of P-gp and/or ABCG2. 1. Elucidation of the catalytic cycle of ATP hydrolysis and transport pathway of P-gp and the role of conserved motifs in the ATP-binding cassette: We are continuing our studies on the catalytic cycle and transport pathway of P-gp. We are using molecular modeling and mutagenesis approaches to elucidate on a molecular level how this transporter recognizes and transports a wide variety of structurally dissimilar compounds. We have begun to use tmFRET, which is a novel biophysical method developed to determine short range (5 - 25 angstrom) distances within different locations of the protein at very low concentrations. Using this sensitive fluorescence-based method, we have begun to determine the changes in distance associated with the apo and the closed (ATP/Vi trapped) conformations of P-gp. With tmFRET, preliminary results show that there is a significant change in the distance of the two NBDs between the apo and closed conformations ( 20 angstrom). Similarly, results of disulfide crosslinking studies with the oxidant copper phenanthroline and bi-functional sulfhydryl group reagents indicate that human P-gp is a very flexible molecule and its NBDs are much closer to each other in the apo form. The distance between the C431 and C1074 residues in the Walker A motif of NBDs ranges from 5 to 25 angstroms in the apo conformation (in the absence of ATP and drug-substrate). In collaboration with Dr. Di Pietro's group at the CNRS, University of Lyon, France, we found that the linker region of human ABCG2 harbors another C- signature motif and point mutations of the first two residues (L and S) of this motif completely abolished the transport-coupled ATPase activity, and resulted in complete reversal of cell resistance to mitoxantrone. These results suggest that the C2-signature motif in the linker region of ABCG2 plays an important mechanistic role in ATP binding and/or hydrolysis of this transporter. 2. Mechanism of the drug-mediated inhibition of P-gp ATPase activity: Most of the substrates or modulators of P-gp stimulate its basal ATPase activity, and only a few drugs have been found to inhibit it. Zosuquidar, tariquidar and elacridar, high affinity inhibitors of transport function, also inhibit Pgp ATPase activity, while a variety of substrates including verapamil, paclitaxel and vinblastine stimulate ATP hydrolysis. The molecular mechanisms that are in play, in either case (stimulation or inhibition), remain elusive. The development of an effective P-gp inhibitor certainly would benefit from the understanding of drug-mediated inhibition of ATP hydrolysis. Using directed mutagenesis, we identified a pair of phenylalanine-tyrosine structural motifs of P-gp that are critical for the inhibition of ATP hydrolysis by high-affinity modulators. These structural motifs are located in the drug-binding pocket of Pgp. We found that drugs that inhibit the ATPase activity switch to stimulating the ATPase activity when any of these residues are mutated. For instance, zosuquidar inhibits the basal ATP hydrolysis of cysless WT P-gp with high affinity (concentration required for 50 percent inhibition = 10-20 nM). The inhibition is completely lost upon mutation of Y953 to alanine and is switched to stimulation when three polar residues are mutated (Y307A/Q725A/Y953A). Molecular modeling revealed that the phenylalanine residues F978 and F728 interact with the tyrosine residues Y953 and Y310, respectively, in an edge-to-face conformation, helping the tyrosine residues to adopt the proper orientation to effectively establish hydrogen-bond contact with the inhibitors. Biochemical investigations along with transport studies in intact cells showed that the inhibitors bind at a high affinity site to produce inhibition of ATP hydrolysis and transport. Upon mutation, they bind at lower affinity sites that lead to stimulation of ATP hydrolysis and a poor inhibition of transport. These studies also demonstrated that screening chemical compounds for their ability to inhibit the basal ATPase activity can be a reliable tool to identify modulators with high affinity for P-gp. 3. Resolution of the three-dimensional structure of human Pgp: The resolution of the three-dimensional structure of P-gp is an ongoing project and for this we have developed a purification scheme that has yielded total protein of 7.5-10.0 mg of 99% homogeneously pure Pgp. Due to the flexible nature of human P-gp and the difficulty of generating crystals of good diffraction quality, we are also using single particle analysis by the cryo-electron microscopy technique. The current studies indicate that the structural features of human Pgp in the presence and absence of a Fab of conformation-sensitive monoclonal antibody can be observed at 9.5 to 13-angstrom resolution. We surprisingly found that in the absence of the transport substrate and nucleotides, human P-gp can exist in both open (NBDs apart; inward-facing) and closed (NBDs close; outward facing) conformations. These studies further demonstrated that the NBDs of P-gp in the presence of vanadate and ATP (ADP-vanadate trapped) were found to be close to each other. Whereas in the presence of ADP alone, the NBDs were separated from each other to varying degrees. We are optimizing conditions to reach a sub-nanometer resolution to obtain the structure of human P-gp in at least three different (apo, ADP-vanadate trapped and Fab-bound) conformations. The cryo-EM studies are carried out in collaboration with Dr. Sriram Subramanian. 4. Substrate and modulator-binding sites on ABCG2: There is very limited information available on the substrate and modulator binding sites on ABCG2. We have begun by using a mutagenesis and molecular modeling approach to identify substrate and modulator binding sites. As ABCG2 is a homodimer, it is possible that the substrate and modulator binding sites are located at the interphase of two monomers. We found that the position of the ten-histidine tag at the N-terminal or the C-terminal end has an effect on the function of the transporter. Although the ABCG2 with the ten-histidine tag is expressed well at the cell surface, it's transport function is completely abolished. On the other hand, the ten-histidine tag at the N-terminal has no effect on the expression level or transport function, which is similar to the wild-type protein without any tag. It appears that the histidine tag at the C-terminal interferes with the catalytic cycle of ABCG2.
Showing the most recent 10 out of 89 publications
