Putative muscarinic receptors (mAChRs) were characterized in the CNS of larval Manduca sexta through the binding of 3H-quinuclidyl benzilate ([3H]-QNB). Specific binding isotherms revealed a high affinity binding site in both crude homogenates (Ka = 3.22 +/- 0.62 nM-1 (Kd 0.311 nM), Bmax = 65.4 +/- 9.8 fmoles/mg protein) and in purified membrane preparations (Ka = 7.61 +/- 1.78 nM-1 (Kd 0.130 nM), Bmax = 22.8 +/- 2.15 fmoles/mg protein). In purified membrane preparations the binding was complex, consisting of at least two sites (Hill coefficient = 0.514 +/- 0.041). Because of the high proportion of non-specific binding at QNB concentrations greater than 10 nM, the binding parameters for the additional low affinity site could not be determined accurately but the Kd was estimated to be greater than 8 nM. Complex binding was also exhibited in the kinetics of [3H]-QNB association and dissociation in purified membranes. Analysis resolved two high affinity sites (dissociation rate constants = 80.4 and 1.2 min, Kds = 0.272 and 0.909 nM, consisting of 65.9 +/- 3.7 and 26 +/- 4.9% of the sites respectively). Presumably, the high affinity site identified in saturation studies consists of these two components. The primary high affinity site was isolated using a dissociative method and its pharmacology determined in competition studies. [3H]-QNB binding at this site was displaced by pirenzepine and methoctramine (Kis = 248 and 707 nM) closely matching the pharmacology of the cloned Drosophila mAChR. Competition data for 4-DAMP were better fitted by a two site model (Kis = 37 and 547 nM). These results demonstrate unexpected complexity in muscarinic ligand binding and are consistent with mAChR heterogeneity in the CNS of Manduca.