This work is aimed at further exploring the concept that phenytoin-related compounds might present with an anti-HIV potential. We screened for anti-HIV activity, selected compounds whose structural design rests on pharmacophores successfully shown to convey phenytoinergic anticonvulsant activity. We determined the corresponding anticonvulsant protective doses in mice via the i.p. route of administration using the maximal electroshock seizure test (a test in which the anticonvulsant activity of phenytoin is well expressed). Firstly, 4-aminophthalimide pharmacophores were utilized with either N-(2,6-dimethyl)phenyl or N-(1-adamantyl) substitutions. While the former was found to be highly potent, the latter was devoid of significant activity. Secondly, the pharmacophores N-(2,6-dimethylphenyl)phthalimide and N-(1-adamantyl)phthalimide were compared for antiviral (antiHIV-1 and antiHIV-2) properties in CEM (human T-lymphocyte) cells infected with HIV-1 or HIV-2 strains. Various phthalimide C4-substitutions (H, NO2, NH2, Cl, CH3, OCH3, COOH) of these pharmacophores were studied. From this set of experiments, 4-amino-N-(1-adamantyl)phthalimide emerged with EC50 (effective concentration-50) values of 16 and 27 microM against HIV-1 and HIV-2, respectively. The CC50 (cytostatic concentration-50) of this compound was 30 microM. Thirdly, the N-(2,6-dimethylphenyl) and N-(1-adamantyl) substitutions of the 4-aminobenzamide pharmacophore (another known phenytoinergic anticonvulsant platform) were shown to be devoid of anti-HIV activities. A similar negative result was obtained for amantadine. Taken as a whole, the present data indicate that both the 4-aminophthalimide pharmacophore and N-(1-adamantyl) substitutions are required for anti-HIV properties. Molecular modeling studies further provide clues for this dual requirement.