Human P-glycoprotein pseudoreceptor modeling: 3D-QSAR study on thioxanthene type multidrug resistance modulators

Citation
Ik. Pajeva et M. Wiese, Human P-glycoprotein pseudoreceptor modeling: 3D-QSAR study on thioxanthene type multidrug resistance modulators, QSAR, 20(2), 2001, pp. 130-138
Citations number
22
Categorie Soggetti
Chemistry & Analysis
Journal title
QUANTITATIVE STRUCTURE-ACTIVITY RELATIONSHIPS
ISSN journal
09318771 → ACNP
Volume
20
Issue
2
Year of publication
2001
Pages
130 - 138
Database
ISI
SICI code
0931-8771(200108)20:2<130:HPPM3S>2.0.ZU;2-W
Abstract
P-glycoprotein (P-gp) is the main multidrug resistance (MDR) bioregulator r esponsible for the active outward transport of antitumor drugs in resistant tumor cells. The primary sequence of human P-gp is known and its low-resol ution structure has recently been determined. Basing on hydropathy plots a secondary structure with 12 membrane spanning helixes is assumed. However, no direct correlation between any structural feature observed and primary s tructure of P-gp has yet been identified. Mutation and labeling studies sug gest that P-gp interaction sites are located in the transmembrane domains ( TMs) and TM6 and TM12 are actively involved in the interactions with many M DR modulators. The purpose of this study is structure-based fitting to the putative P-gp 3D receptor of TM6 and TM12 using pseudoreceptor modeling app roach. It is based on the "six plus six" topology of P-gp with TM6 neighbor ing TM12. Data on 24 MDR ligands of thioxanthene and phenothiazine type and the program PrGen were used. A stepwise strategy was applied: (i) single d omain models (TM6 or TM12) to decide on the appropriate orientation of the ligands towards the single domain; (ii) combined domain models (TM6 and TM1 2) - to simulate the behavior of the ligands in the protein environment of interest; (iii) extended models - to validate the models. Training was done on thioxanthenes only and test - on thioxanthenes and phenothiazines. A nu mber of pseudoreceptor models were derived. Scramble tests were performed w ith the extended models. The best model involved TM6 and TM12 acting togeth er and yielding 1.6 factor of uncertainty in the binding constants of the t est ligands. Among the mostly contributing AAs were those proven by mutatio n to influence P-gp functioning. TM12 was suggested to be involved through Gln990 in specific H-bond interactions with Phe983 and Gln990 being mostly contributing to the model. TM6 was suggested to perform hydrophobic and sha pe regulating functions through its bulky and hydrophobic AAs with Phe335 a nd Va1338 mainly involved. As this is the first attempt to model the intera ction between P-gp and MDR modulators, the obtained results have to be furt her proven on larger and structurally diverse data sets.