?Computational models applied to the prediction of the affinity of compounds to P-Glycoprotein?.

PALESTRO PABLO; PEREZ ANDRES; GAVERNET LUCIANA; BRUNO BLANCH LUIS E.

Buenos Aires

Congreso; Biowaivers; 2015

P-Glycoprotein (P-gp) is involved in the transport of xenobiotic compounds and it is responsible for the decrease of the drug accumulation in multi drug-resistant cells. Here we present a computational approach, known as molecular docking, to predict the affinities of drugs and new compounds to the P-gp. The final purpose is to have in silico filters for the early recognition of substrates of efflux transporters in the drug development projects.Materials and methods:Docking is a computer based method frequently employed to predict and quantify the binding affinity of a small molecule (a drug or a new compound) to one biological target (such as P-gp). It calculates the preferred orientation of the small molecule into the target active site, and measures the strength of association through the calculation of the binding energies.To perform this methodology we first constructed a tridimensional structure of human P-gp. We employed the experimental data from mouse P-gp structure (protein data bank code: 3G61) as a template, and I-TASSER server and PSVS server to design and validate the 3D human model.Then we analyzed the capacity of different docking programs to discriminate 26 known P-gp substrates and 26 non-substrates taken from literature: An optimal docking program will assign higher affinities to the substrates than non-substrates (i. e. more negative binding energies). Several conditions and software were used and the performances of these models were analyzed with the receiver operating characteristic curves (ROC curves). Finally we re-evaluate the best model with an extended database composed by 666 P-gp known inhibitors and 609 non-inhibitors.