In this work, we evaluate the structural differences of transmembrane helix 3 in rhodopsin and the 5-hydroxytryptamine 1A (5-HT 1A ) receptor caused by their different amino acid sequence. Molecular dynamics simulations of helix 3 in the 5-HT 1A receptor tends to bend toward helix 5, in sharp contrast to helix 3 in rhodopsin, which is properly located within the position observed in the crystal structure. The relocation of the central helix 3 in the helical bundle facilitates the experimentally derived interactions between the neurotransmitters and the Asp residue in helix 3 and the Ser/Thr residues in helix 5. The different amino acid sequence that forms helix 3 in rhodopsin (basically the conserved Gly 3.36 Glu 3.37 motif in the opsin family) and the 5-HT 1A receptor (the conserved Cys 3.36 Thr 3.37 motif in the neurotransmitter family) produces these structural divergences. These structural differences were experimentally checked by designing and testing ligands that contain comparable functional groups but at different interatomic distance. We have estimated the position of helix 3 relative to the other helices by systematically changing the distance between the functional groups of the ligands (1 and 2) that interact with the residues in the receptor. Thus, ligand 1 optimally interacts with a model of the 5-HT 1A receptor that matches rhodopsin template, whereas ligand 2 optimally interacts with a model that possesses the proposed conformation of helix 3. The lack of affinity of 1 (K i Ͼ 10,000 nM) and the high affinity of 2 (K i ϭ 24 nM) for the 5-HT 1A receptor binding sites, provide experimental support to the proposed structural divergences of helix 3 between the 5-HT 1A receptor and rhodopsin.

In the present paper, we report the synthesis and the binding profile on 5-HT1A, alpha1 and D2 receptors of a new series of 1-[omega-(4-arylpiperazin-1-yl)alkyl]-3-(diphenylmethylene)- 2, 5-pyrrolidinediones (III) (1-4) and -3-(9H-fluoren-9-ylidene)-2, 5-pyrrolidinediones (IV) (1-4), in which the alkyl linker contains 1-4 methylenes and the aryl group is variously substituted. The results obtained are compared to those previously reported for bicyclohydantoin (I) and the related bicyclic amine (II) series. A considerable part of the tested compounds 1-4 demonstrated moderate to high affinity for 5-HT1A and alpha1 receptor binding sites but had no affinity for D2 receptors. The study of the length of the alkyl chain and the imide substructure has allowed us to suggest some differences between the 5-HT1A and the alpha1-adrenergic receptors: (i) for III and IV, affinity for the 5-HT1A receptor as a function of the length of the methylene linker decreases in the order 4 > 1 >> 3 approximately 2, while for the alpha1 receptor affinity decreases in the order 3 approximately 4 > 1 approximately 2; (ii) the no-pharmacophoric steric pocket (receptor zone which does not hold the pharmacophore of the ligand but holds a nonessential fragment of the molecule) in the 5-HT1A receptor has less restriction than the corresponding pocket in the alpha1 receptor. Compounds 3a,e, which are highly selective for alpha1-adrenergic receptors, displayed antagonist activity. On the other hand, the best compromise between affinity and selectivity for 5-HT1A receptors is reached in these new series with n = 1, which is in agreement with our previous results for the bicyclohydantoin derivatives I. Two selected compounds (1d and 4e) retain agonist properties at postsynaptic 5-HT1A receptors. The same 5-HT1A agonist profile found in these compounds suggests the existence of two different no-pharmacophoric steric pockets in this receptor and a different interaction of compounds with n = 1 and n = 4. The information obtained from the interpretation of the energy minimization and 2D-NOESY experiments of compounds 1-4 together with the synthesis and binding data of new conformationally restrained analogues 4k-m is in good agreement with this working hypothesis.