D(2)-like dopamine receptors mediate functional changes via activation of inhibitory G proteins, including those that affect adenylate cyclase activity, and potassium and calcium channels. Although it is assumed that the binding of a drug to a single isoform of a D(2)-like receptor will cause similar changes in all receptor-mediated functions, it has been demonstrated in brain that the dopamine agonists dihydrexidine (DHX) and N-n-propyl-DHX are "functionally selective". The current study explores the underlying mechanism using transfected MN9D cells and D(2)-producing anterior pituitary lactotrophs. Both dopamine and DHX inhibited adenylate cyclase activity in a concentration-dependent manner in both systems, effects blocked by D(2), but not D(1), antagonists. In the MN9D cells, quinpirole and R-(-)-N-propylnorapomorphine (NPA) also inhibited the K(+)-stimulated release of [(3)H]dopamine in a concentration-responsive, antagonist-reversible manner. Conversely, neither DHX, nor its analogs, inhibited K(+)-stimulated [(3)H]dopamine release, although they antagonized the effects of quinpirole. S-(+)-NPA actually had the reverse functional selectivity profile from DHX (i.e., it was a full agonist at D(2L) receptors coupled to inhibition of dopamine release, but a weak partial agonist at D(2L) receptor-mediated inhibition of adenylate cyclase). In lactotrophs, DHX had little intrinsic activity at D(2) receptors coupled to G protein-coupled inwardly rectifying potassium channels, and actually antagonized the effects of dopamine at these D(2) receptors. Together, these findings provide compelling evidence for agonist-induced functional selectivity with the D(2L) receptor. Although the underlying molecular mechanism is controversial (e.g., "conformational induction" versus "drug-active state selection"), such data are irreconcilable with the widely held view that drugs have "intrinsic efficacy".