alpha 2-Adrenergic agonist preincubation resulted in a leftward shift in the subsequent concentration-response curve to forskolin-stimulated adenylate cyclase activity in membranes from HT29 cells, a human colon adenocarcinoma cell line. This effect was much less pronounced than the effect seen in the intact cell cyclic AMP production assays. Removal of GTP from the assay caused a further slight leftward shift in the concentration-response curve. In [3H]forskolin binding experiments, alpha 2-adrenergic agonist preincubation caused a doubling of the maximal number of binding sites (80 vs 31 fmol/mg protein) compared to control. The addition of MgCl2 and NaF to the assay buffer increased control binding 5-fold. With agonist preincubation, there was a further increase in binding in the presence of MgCl2 and NaF which was not significantly different from the appropriate control. Pertussis toxin pretreatment blocked both the leftward shift in the forskolin concentration-response curve and the increase in maximal number of binding sites, indicating that a pertussis toxin sensitive protein is involved in these changes. Activation of cyclic AMP production in the intact cell by cholera toxin followed by norepinephrine preincubation and then stimulation by forskolin resulted in a degree of sensitization similar to that seen in the membrane adenylate cyclase and binding assays. Pertussis toxin also blocked this sensitization. It appears that if the cyclase system is highly activated, then the degree of sensitization is similar in the membrane and intact cell assay.