Transfection of individual opioid receptors in Chinese hamster ovary (CHO) cells provides a pure, homogeneous population of receptors for screening drug candidates, and an alternative to the use of selective ligands. To evaluate the potential of this system, we chose a series of (-)-5,9 alpha-dimethyl-2-hydroxy-N-substituted-6,7-benzomorphans, for which the receptor selectivity and in vivo activity had been characterized recently, and tested them in CHO cells stably transfected with either the rat delta-opioid receptor or the mouse mu-opioid receptor. [3H]Diprenorphine was used to measure opioid receptors in P2 membrane preparations. A Bmax of 7.58 +/- 0.8 pmol/mg protein and a Kd of 0.42 +/- 0.04 nM was obtained in the mu-opioid receptor expressing cell line used in these studies. In addition, [3H]naltrindole was used to confirm the delta-specificity of the cloned receptor. Both compounds gave a Bmax of 1.2 pmol/mg in the CHO cells expressing the rat delta-opioid receptor. Displacement assays were performed with eleven (-)-N-alkyl-benzomorphans in the absence and presence of 150 mM NaCl, as well as known delta- and mu-selective agonists. Sodium reduced agonist affinity in the transfected cell lines. The benzomorphan compounds displayed a range of affinities in the mu- and delta-opioid receptor expressing cell lines. Good correlations were found between their affinities at the cloned mu- and delta-opioid receptors and those in rat brain and monkey cortex (r2 from 0.73 to 0.89, P < 0.001). Comparative analysis of Ki values with in vivo potency in the mouse tail flick test indicated a high degree of correlation between antinociception and affinity in the mu-opioid receptor cell line (r2 = 0.83, p < 0.0001). Lesser correlations were found between antinociception in the mouse and affinity at the rat mu-opioid receptor (r2 = 0.6610) and at the monkey mu-opioid receptor (r2 = 0.695). In sum, these studies indicate that the cell lines expressing the cloned mu- and delta-opioid receptors are appropriate models for determining the binding affinities of this class of opioid compounds. The diminishing correlations found between species when comparing in vitro and in vivo activity suggest that caution should be taken when extrapolating binding data to pharmacological activity among species.