Two 14beta-p-nitrocinnamoyl derivatives of dihydrocodeinone, 14beta-(p-nitrocinnamoylamino)-7,8-dihydrocodeinone (CACO) and N-cyclopropylmethylnor-14beta-(p-nitrocinnamoylamino)- 7, 8-dihydrocodeinone (N-CPM-CACO), and the corresponding chlorocinnamoylamino analogs, 14beta-(p-chlorocinnamoylamino)-7, 8-dihydrocodeinone (CAM) and N-cyclopropylmethylnor-14beta-(p-chlorocinnamoylamino) -7, 8-dihydrocodeinone (MC-CAM), were tested in opioid receptor binding assays and the mouse tail-flick test to characterize the opioid affinity, selectivity, and antinociceptive properties of these compounds. In competition binding assays, all four compounds bound to the mu opioid receptor with high affinity. When bovine striatal membranes were incubated with any of the four dihydrocodeinones, binding to the mu receptor was inhibited in a concentration-dependent, wash-resistant manner. Saturation binding experiments demonstrated that the wash-resistant inhibition of mu binding was due to a decrease in the Bmax value for the binding of the mu-selective peptide [3H][D-Ala2, MePhe4,Gly(ol)5] enkephalin and not a change in the Kd value, suggesting an irreversible interaction of the compounds with the mu receptor. In the mouse 55 degrees C warm water tail-flick test, both CACO and N-CPM-CACO acted as short-term mu-selective agonists when administered by i. c.v. injection, whereas CAM and MC-CAM produced no measurable antinociception at doses up to 30 nmol. Pretreatment of mice for 24 h with any of the four dihydrocodeinone derivatives produced a dose-dependent antagonism of antinociception mediated by the mu but not the delta or kappa receptors. Long-term antagonism of morphine-induced antinociception lasted for at least 48 h after i.c. v. administration. Finally, shifts in the morphine dose-response lines after 24-h pretreatment with the four dihydrocodeinone compounds suggest that the nitrocinnamoylamino derivatives may produce a greater magnitude long-term antagonism of morphine-induced antinociception than the chlorocinnamoylamino analogs.