Résumé : Since the late eighties five muscarinic receptor subtypes (m1 - m5) have been cloned and four of them (M1 - M4) have also been pharmacologically characterized. However, there is still a lack of potent muscarinic agonists and antagonists, which are highly selective for one muscarinic receptor subtype over all other subtypes. For the treatment of Alzheimer's disease, M1-selective agonists capable of penetrating into the CNS are needed. It is hypothezised that such substances would not only improve memory and cognitive ability, but also delay the progression of the disease. In our laboratory, the functionally M1-selective quaternary ammonium compound McN-A-343 has been used as a starting point for the design of such CNS active muscarinic ligands. Structure-activity relationship studies led to the tertiary amine 4-(4-fluorophenylcarbamoyloxy)-2-butynylpyrrolidine (4-F-PyMcN), which was found to stimulate M1 receptors with some functional selectivity. In order to increase the potency and selectivity of 4-F-PyMcN several new derivatives were synthezised and pharmacologically characterized in different functional assays as well as in binding and biochemical (PI turnover) studies. The most promising results were obtained with (S)-4-(4-fluorophenylcarbamoyloxy)-1-methyl-2-butynylpyrrolidine (4-F-MePyMcN). Due to its potent partial agonistic activity at M1 receptors and its M2-antagonistic properties leading to an increase of acetylcholine release by blockade of M2 autoreceptors, this compound may be considered as an important tool for future drug research of cognitive disorders. M2 receptor antagonists may also be used for the treatment of Alzheimer's disease, furthermore in the therapy of supraventricular bradycardia and for quantifying M2 receptors in the CNS with PET imaging. In the search for antagonists which clearly differentiate M2 from other muscarinic receptors, we investigated the two enantiomers of the widely used H1-antihistaminic drug dimethindene. (S)-Dimethindene proved to be a potent M2-selective antagonist with lower affinities for the M1, M3 and M4 receptors. In addition, the (S)-enantiomer was more potent than the (R)-enantiomer in all muscarinic assays. Interestingly, the stereoselectivity was inverse at histamine H1 receptors, the (R)-enantiomer being the eutomer. M3 receptor antagonists may be useful in the treatment of spastic disorders of the gastrointestinal, urogenital and respiratory tract as well as for the relief of glandular hypersecretion. In previous studies, hexahydro-difenidol (HHD) and its sila-analogue, hexahydro-sila-difenidol (HHSiD), as well as the antiparkinsonian drug trihexyphenidyl (THP) were found to be valuable tools for the discrimination of M3 and M2 receptors. In order to further assess the structural requirements (including stereochemical aspects) of the above-mentioned compounds for potency and selectivity, a series of HHD and THP analogues as well as of the corresponding silicon and germanium derivatives (sila- and germa-substitution) were studied. The (R)-enantiomers displayed higher affinities and selectivities than the corresponding (S)-isomers. The enantioselectivity of some of these analogues is best explained by the concept of the four-binding-subsite model suggesting that the differences in affinity of the (R)- and (S)-enantiomers at muscarinic receptors are due to opposite binding of the phenyl and the cyclohexyl ring to the preferring subsites. Surprisingly, there was no significant difference between the Si and Ge analogues indicating a strongly pronounced Si/Ge bioisosterism in this series of compounds. The related carbon derivatives, however, showed higher receptor affinities as well as greater stereoselectivities at all muscarinic receptors studied compared with the silicon and germanium analogues. © 1996 Elsevier B.V. All rights reserved.