Constraining the 1,2,5,6-tetrahydropyridine ring of arecoline 1a and the related muscarinic agonists 1b-e by replacing the N-methyl group with an ethano bridge between the nitrogen and the C5-position afforded compounds 2a-e. These compounds have enhanced affinities for muscarinic receptors, which supports our proposed binding conformation (1a)ax of arecoline. There has been considerable recent interest in the potential of cholinomimetics as cognition enhancers for the treatment of senile dementia of the Alzheimer type (SDAT). Clinical evaluation of the classical muscarinic agonist arecoline 1a produced small improvements in patients suffering from SDAT. However, the potential benefits of the drug are compromised by its poor metabolic stability and peripheral side effects. This has led to the development of muscarinic agonists related to arecoline in which the ester group has been replaced by a metabolically stable bioisostere such as a heteroaromatic ring or an oxime ether group. We recently reported that the combination of N-methoxy imidoyl halide or nitrile moieties with the 1,2,5,6-tetrahydropyridine ring system afforded potent muscarinic agonists 1b-e. In addition to metabolic stability, some of these compounds showed favourable functional selectivity which we related to their partial agonist nature. In an earlier report we hypothesised that the muscarinic affinity of arecoline is associated with the conformer in which the N-methyl group adopts an axial orientation (1a)ax. Although the calculated energy of the axial conformer (1a)ax is higher than that of the equatorial conformer (1a)eq, the difference is only 1.87 kcal/mol and the Boltzmann population distribution at 37°C indicates a significant contribution from (1a)ax. This reasoning led us to prepare azabicyclic esters in which a six-membered piperidine ring is constrained in either a boat or chair conformation by a bridging group which effectively fulfills the requirement for an axially oriented nitrogen substituent. We then went on to develop a wide range of ester bioisosteres some of which gave compounds of extremely high affinities for muscarinic receptors with a range of efficacies. The success of this approach has encouraged us to test our hypothesis further and this paper reports the preparation of the ester 2a in which the 1,2,5,6-tetrahydropyridine ring has been constrained in a half-chair conformation by an ethano bridge between the nitrogen and the 5-position. As with the azabicyclic compounds, the nitrogen substituent is tied back in the proposed 'axial' binding conformation. In addition, bioisosteric replacements of the ester group by aldoxime ether, N-methoxy imidoyl halide and N-methoxy imidoyl nitrile moieties have been investigated in this novel ring system. The affinity and efficacy of the compounds were assessed by a two-stage binding assay designed to measure both parameters. The ability of the compounds to inhibit the binding of the muscarinic agonist [³H]oxotremorine-M (OXO-M) provided a measure of affinity for the high-affinity agonist state of the receptor. The ratio of the IC50 values for inhibition of binding of [³H]quinuclidinyl benzilate (QNB), a muscarinic antagonist, and OXO-M was used to predict efficacy. Ratios greater than 100 are associated with full agonists; antagonists give ratios close to unity and intermediate values indicate partial agonists. The results are shown in the Table together with those of the N-methyl-1,2,5,6-tetrahydropyridine analogues 1a-1e for comparison. The azabicyclic ester 2a shows an affinity for muscarinic receptors about four times that of arecoline 1a with an efficacy ratio approximately 2/3 that of 1a, but still predictive of full agonism. Similarly, the azabicyclic oximes 2b-e display higher affinity than the monocyclic analogues 1b-e, although the enhancements in affinity are modest. The efficacy ratios are also generally lowered (with the exception of 1d vs 2d) and are predictive of partial agonists. This decrease in efficacy is probably related to the increased steric demands of the azabicyclic ring and is consistent with our earlier findings that marked changes in efficacy can be effected by relatively subtle structural changes. The increased affinities of 2a-2e over 1a-1e provides additional support for the hypothesis that the N-methyl group of arecoline is axial in the binding conformation. Recent efforts by others to mimic alternative binding conformations of arecoline have been less successful. The enantiomers of 3-carbomethoxy-2-methyl-2-azabicyclo[2.2.1]heptane were prepared as rigid analogues of the X-ray conformation of arecoline and found to be only weak partial agonists. Interestingly, their lack of muscarinic activity was attributed to an unfavourable equatorial orientation of the N-methyl group. A related study reported that bridging the C-2 and C-5 positions of the potent muscarinic agonist 3-(3-amino-1,2,4-oxadiazol-5-yl)-1-methyl-1,2,5,6-tetrahydropyridine by an ethano bridge also led to loss of affinity for muscarinic receptors.