Nonclassical folate analogs as inhibitors of dihydrofolate reductase and thymidylate synthase

The principal goal of this work has been to design exceptionally potent inhibitors of the folate dependent enzymes dihydrofolate reductase (DHFR) and thymidylate synthase (TS) which are at the same time selective enough to have reduced host toxicity. Toward this end, the synthesis and evaluation of a series of five novel 2,4-diamino-pyrido[3,2-d]pyrimidine inhibitors of dihydrofolate reductase (DHFR) and of two novel 10-deaza flexible folate" inhibitors of thymidylate synthase (TS) have been accomplished. The five 2,4-diaminopyrido (3,2-d) pyrimidine inhibitors displayed potent nM level inhibition of DHFRs from P. carinii, T. gondii, human and rat liver. Inhibition of DHFR from T. gondii was approximately equipotent with that observed for the clinically approved agent trimetrexate. Though potent inhibitors, these new agents offered no significant improvement in selectivity for the P. carinii or T. gondii DHFR. Cocrystallization of a previously characterized "flexible folate" multisubstrate analog inhibitor (MAI) in the TS active site and resolution of the structure by difference Fourier analysis demonstrated that linkage of the MAI's folate and nucleotide moieties by means of a methylene bridge to an $sp\sp2$ hybridized carbon at position C-5 of the nucleotide's pyrimidine ring prevents irreversible inhibition of TS. The crystallographic and molecular modeling studies served as a basis for the subsequent design of two novel 10-deaza "flexible folate" inhibitors of TS. No inhibitory activity was observed for either compound against TS enzymes from P. carinii. T. gondii, human rat liver or Trypanosoma cruzii. The lack of biological activity of these inhibitors emphasizes the important contribution of entropic factors to the free energy of ligand binding. For conformationally flexible inhibitors which lack functional groups capable of participating in hydrogen bonding networks and/or forming strong electrostatic interactions with the enzyme, entropic factors may predominate in the energetics of ligand binding and must be given careful consideration. The findings presented herein reemphasize the necessity of adopting an integrated approach to structure-based drug design in which successive cycles of crystallographic resolution of enzyme-ligand complexes, ligand design, synthesis and biological evaluation are required for optimization of inhibitor potency and selectivity.