The purine nucleoside analogue NBMPR {nitrobenzylthioinosine or 6-[(4-nitrobenzyl)thio]-9--D-ribofuranosylpurine} was selectively phosphorylated to its nucleoside 5'-monophosphate by Toxoplasma gondii but not mammalian adenosine kinase (EC 2.7.1.20). NBMPR was also cleaved in toxoplasma to its nucleobase, nitrobenzylmercaptopurine. However, nitrobenzylmercaptopurine was not a substrate for either adenosine kinase or hypoxanthine-guanine-xanthine phosphoribosyltransferase (EC 2.4.2.8). Because of this unique and previously unknown metabolism of NBMPR by the parasite, the effect of NBMPR as an antitoxoplasmic agent was tested. NBMPR killed T. gondii grown in human fibroblasts in a dose-dependent manner, with a 50% inhibitory concentration of approximately 10 µM and without apparent toxicity to host cells. Doses of up to 100 µM had no significant toxic effect on uninfected host cells. The promising antitoxoplasmic effect of NBMPR led to the testing of other 6-substituted 9--D-ribofuranosylpurines, which were shown to be good ligands of the parasite adenosine kinase (M. H. Iltzsch, S. S. Uber, K. O. Tankersley, and M. H. el Kouni, Biochem. Pharmacol. 49:1501-1512, 1995), as antitoxoplasmic agents. Among the analogues tested, 6-benzylthioinosine, p-nitrobenzyl-6-selenopurine riboside, N⁶-(p-azidobenzyl)adenosine, and N⁶-(p-nitrobenzyl)adenosine, like NBMPR, were selectively toxic to parasite-infected cells. Thus, it appears that the unique characteristics of purine metabolism in T. gondii render certain 6-substituted 9--D-ribofuranosylpurines promising antitoxoplasmic drugs.