Receptor-based design of novel dihydrofolate reductase inhibitors: benzimidazole and indole derivatives

Although many thousands of inhibitors of the enzyme dihydrofolate reductase (DHFR) have been synthesized, all of the very active compounds have been 2,4-diaminopyrimidines or very close analogues. This paper describes 2,4-diamino-6-benzylbenzimidazole (3b) and the corresponding indole (4), as well as more complex tri- and tetracyclic derivatives (5 and 6). These were designed on the basis of molecular modeling to the known X-ray structure of Escherichia coli DHFR, in an effort to determine whether one could drastically alter the diamino configuration by placing one amino substituent in a 5-membered nitrogen-containing ring and the second in the ortho position of a fused ring and still inhibit DHFR significantly. Although the electronics and bond angles are quite different from that of a 2,4-diaminopyrimidine, the pKa values are in an appropriate range, and hydrogen-bond distances appear to be quite reasonable. The most active compound, 4, was very unstable and active only in the 10(-4) M range. Dihydroindenoimidazole derivatives such as 6 showed quite a good fit to the enzyme by modeling studies, but had low activity. Since the most active compound made was 2 orders of magnitude weaker as an inhibitor of bacterial DHFR than the unsubstituted 5-benzyl-2,4-diaminopyrimidine, we concluded that such a ring system was unlikely to produce the high inhibitory potency of trimethoprim (1), even with greatly improved hydrophobic contacts. Thus the 2,4-diaminopyrimidine system remains unparalleled to date for the competitive inhibition of this enzyme.