Comparative receptor mapping of serotoninergic 5-HT3 and 5-HT4 binding sites.

The clinical use of currently available drugs acting at the 5-HT4 receptor has been hampered by their lack of selectivity over 5-HT3 binding sites. For this reason, there is considerable interest in the medicinal chemistry of these serotonin receptor subtypes, and significant effort has been made towards the discovery of potent and selective ligands. Computer-aided conformational analysis was used to characterize serotoninergic 5-HT3 and 5-HT4 receptor recognition. On the basis of the generally accepted model of the 5-HT3 antagonist pharmacophore, we have performed a receptor mapping of this receptor binding site, following the active analog approach (AAA) defined by Marshall. The receptor excluded volume was calculated as the union of the van der Waals density maps of nine active ligands (pKi > or = 8.9), superimposed in pharmacophoric conformations. Six inactive analogs (pKi < 7.0) were subsequently used to define the essential volume, which in its turn can be used to define the regions of steric intolerance of the 5-HT3 receptor. Five active ligands (pKi > or = 9.3) at 5-HT4 receptors were used to construct an antagonist pharmacophore for this receptor, and to determine its excluded volume by superimposition of pharmacophoric conformations. The volume defined by the superimposition of five inactive 5-HT4 receptor analogs that possess the pharmacophoric elements (pKi < or = 6.6) did not exceed the excluded volume calculated for this receptor. In this case, the inactivity may be due to the lack of positive interaction of the amino moiety with a hypothetical hydrophobic pocket, which would interact with the voluminous substituents of the basic nitrogen of active ligands. The difference between the excluded volumes of both receptors has confirmed that the main difference is indeed in the basic moiety. Thus, the 5-HT3 receptor can only accommodate small substituents in the position of the nitrogen atom, whereas the 5-HT4 receptor requires more voluminous groups. Also, the basic nitrogen is located at ca. 8.0 A from the aromatic moiety in the 5-HT4 antagonist pharmacophore, whereas this distance is ca. 7.5 A in the 5-HT3 antagonist model. The comparative mapping of both serotoninergic receptors has allowed us to confirm the three-component pharmacophore accepted for the 5-HT3 receptor, as well as to propose a steric model for the 5-HT4 receptor binding site. This study offers structural insights to aid the design of new selective ligands, and the resulting models have received some support from the synthesis of two new active and selective ligands: 24 (Ki(5-HT3) = 3.7 nM; Ki(5-HT4) > 1000 nM) and 25 (Ki(5-HT4) = 13.7 nM; Ki(5-HT3) > 10,000 nM).