A series of 2,4-diamino-5-methyl-6-(anilinomethyl)pyrido[2,3-d]pyrimidines 4-9 were synthesized as 5-deaza nonclassical antifolates containing trimethoxy, dichloro-, or trichlorophenyl substitutions and a N-H, N-CH3, or N-CHO at the 10-position. The compounds were evaluated as inhibitors of dihydrofolate reductases (DHFR) from Pneumocystis carinii (P. carinii), Toxoplasma gondii (T. gondii), rat liver (RL), and Lactobacillus casei (L. casei); as inhibitors of T. gondii and P. carinii cell growth in culture; and as antitumor agents. The compounds were prepared by modifications of procedures for classical 5-deaza folates. 2,4-Diamino-5-methyl-6-[(3',4',5'-trimethoxy-N- methylanilino)methyl]pyrido[2,3-d]pyrimidine (5a) exhibited high potency as well as selectivity (compared to RL DHFR) for P. carinii and T. gondii DHFR. Compound 5a is one of the most potent and selective nonclassical folate inhibitors of T. gondii DHFR known. The N-10 formyl analogue 2,4-diamino-5-methyl-6-[(N-formyl-3',4',5'-trimethoxyanilino) methyl]pyrido-[2,3-d]pyrimidine (6a) had decreased potency, but it maintained high selectivity for T. gondii DHFR. The corresponding chloro-substituted analogues maintained potency or had decreased potency; N-10 substitution did not increase potency or selectivity to the extent observed in the 3',4',5'-trimethoxy series. Partial reduction of the B ring to afford the dihydro analogue 2,4-diamino-5-methyl-6-[(N-formyl-3',4',5'-trimethoxyanilino) methyl]-5,8-dihydropyrido[2,3-d]pyrimidine (7), its 5,6,7,8-tetrahydropyrido[2,3-d]pyrimidine analogue 8, and 2,4-diamino-5-methyl-6-[(3',4',5'-trimethoxyanilino)methyl]-5,6,7, 8- tetrahydropyrido[2,3-d]pyrimidine (9) resulted in a significant decrease in potency. In T. gondii cell culture inhibitory studies, 2,4-diamino-5-methyl-6-[(3',4',5'- trimethoxyanilino)methyl]pyrido[2,3-d]pyrimidine (4a), 5a, and 6a were less potent compared to their DHFR inhibitory potencies. Against P. carinii cells in culture, 4a and 5a at 10 micrograms/mL were as effective as the clinically used combination of trimethoprim/sulfamethoxazole (50/250 micrograms/mL). With the exception of the B ring reduced analogues 7-9, all of the compounds were significantly cytotoxic to leukemia CCRF-CEM cells in culture. The chloro-substituted analogues, in general, were more potent against a variety of other tumor cells in culture than the trimethoxy analogues. These results were corroborated by the preclinical tumor screening program at the National Cancer Institute where the most potent compound 2,4-diamino-5-methyl-6-[(3',4'-dichloroanilino)methyl]pyrido[2,3- d]pyrimidine (4b) was found to inhibit the growth of 26 tumor cell lines at an IG50 < 1.00 x 10(-8) M.