A search for sources of drug resistance by the 4D-QSAR analysis of a set of antimalarial dihydrofolate reductase inhibitors

A set of 18 structurally diverse antifolates including pyrimethamine, cyclo guanil, methotrexate, aminopterin and trimethoprim, and 13 pyrrolo[2,3-d]py rimidines were studied using four-dimensional quantitative structure-activi ty relationship (4D-QSAR) analysis. The corresponding biological activities of these compounds include IC50 inhibition constants for both the wild typ e, and a specific mutant type of Plasmodium falciparum dihydrofolate reduct ase (DHFR). Two thousand conformations of each analog were sampled to gener ate a conformational ensemble profile (CEP) from a molecular dynamics simul ation (MDS) of 100,000 conformer trajectory states. Each sampled conformati on was placed in a 1 Angstrom cubic grid cell lattice for each of five tria l alignments. The frequency of occupation of each grid cell was computed fo r each of six types of pharmacophore groups of atoms of each compound. Thes e grid cell occupancy descriptors (GCODs) were then used as a descriptor po ol to construct 4D-QSAR models. Models for inhibition of both the 'wild' ty pe and the mutant enzyme were generated which provide detailed spatial phar macophore requirements for inhibition in terms of atom types and their corr esponding relative locations in space. The 4D-QSAR models indicate some str uctural features perhaps relevant to the mechanism of resistance of the Pla smodium falciparum DHFR to current antimalarials. One feature identified is a slightly different binding alignment of the ligands to the mutant form o f the enzyme as compared to the wild type.