We have analyzed ligand-receptor complexes resulting from (i) the incubation of canine hepatic plasma membranes with [12'I5iodoglucagon and (ii) subsequent gentle solubilization of receptor-bound ligand with digitonin. The complexes (molecular weight 500,000) retain the radiolabeled ligand during gel filtration and subsequent manipulation at 40C in the absence of covalent crosslinking. Affinity chromatography of the glucagon-receptor complexes on columns of wheat germ lectin linked to agarose resulted in two fractions, one of which was not bound by the column and the other of which was specifically eluted by N-acetylglucosamine. The presence of GTP during the incubation of plasma membranes with [125I]iodoglucagon caused about a 50% decrease in total ligand binding but affected only the ligand-receptor complexes that bound to wheat germ lectin. Moreover, it was found that the proportion of the two forms of ligand-receptor complexes identified by chromatography on wheat germ lectin depended on the degree of saturation of the membrane receptor. Thus, both the inhibition by glucagon of radiolabeled glucagon binding to membranes and the concomitantly decreased extent of association of the radiolabeled ligand with solubilized receptor complexes could be modeled in terms of two noninteracting receptor populations (having dissociation constants of about 0.35 and 4.94 x 10-9 M). We conclude that (i) glucagon-receptor complexes formed on canine hepatic plasma membranes exist in two forms that differ after solubilization by digitonin in their avidities for wheat germ lectin, (a) the highand low-affinity binding of glucagon characteristic of hepatic plasma membranes arises from distinct receptor populations that probably differ in glycosylation, and (Uii) the effect of GTP to decrease binding of glucagon to membranes arises from interactions of the nucleotide with the receptor complex that binds to wheat germ lectin.Although the binding ofligand to plasma membrane receptors is known to initiate the mechanism of peptide hormone action, many details ofligand-receptor interactions remain to be clarified. Such interactions often fail to conform to those expected for single binding equilibria, and it has been noted that this deviation from the simple case is particularly acute in the binding of glucagon to hepatic plasma membranes and to surface receptors of isolated hepatocytes (1-6): as much as a 50,000-fold increase in the concentration of hormone is required to cause a reduction from 90% to 10% of maximal binding of [125I]iodoglucagon, whereas theory predicts that a 100-fold increase should be sufficient for a single binding equilibrium. In fact, the complex binding of glucagon both to rat hepatocytes and to canine hepatocytes has been successfully modeled in terms of two noninteracting populations of receptors having different affinities for the ligand (1, 3, 5). In the case of canine hepatocytes, cells that appear to bind glucagon with a notable degree of complexity, high-and low-affinity receptor populations were...