To further investigate our hypothesis on the structure of the platelet-activating factor (PAF) receptor, 35 compounds derived from 1,4-bis(3',4',5'-trimethoxybenzoyl)piperazine were synthesized and their in vitro antagonistic effect was measured. Substitution of the compounds in position 2, by ester or carbamate groups, giving increased steric hindrance and hydrophobicity, increased the platelet aggregation inhibitory activity from 2 microM (without substitution, compound 2) to 0.07 microM (compound 1h) and gave a maximum displacement of [3H]PAF from platelet membrane of 0.05 microM (compound 1k). It appears that the PAF antagonistic effect is only weakly enantiospecific, as observed in many cases including antagonists structurally related or not to PAF. 3D electrostatic potential maps (calculated at -10 kcal/mol) of such compounds revealed a double "Cache-oreilles" (ear-muffs) system. One of these systems has been previously described (distance between atoms generating negative wells, 11-14 A). The second shorter "Cache-oreilles" (6-7 A) system appears to be required for increased PAF antagonistic activity. This short distance between groups generating the negative wells is present in the gingkolides, a series of naturally occurring PAF antagonists. The present study indicates that the structure of the PAF receptor may be more complicated than our initial hypothesis and may be a tetrapolarized structure, with alternants of electropositive and hydrophobic areas. This modified hypothesis is in agreement with recent publications concerning PAF antagonists bearing a cationic moiety.