An unbiased conditioned place preference (CPP) paradigm was used to evaluate the reward effects of endogenous mu-opioid receptor ligands endomorphin-1 (EM-1) and endomorphin-2 (EM-2) from the mesolimbic posterior nucleus accumbens (Acb) shell and the ventral tegmental area (VTA) in CD rats. EM-1 (1.6-8.1 nmol) microinjected into posterior Acb shell produced CPP, whereas EM-2 (8.7-17.5 nmol) given into the same Acb shell produced conditioned place aversion (CPA). EM-1 (1.6-16.3 nmol) microinjected into the VTA produced CPP, whereas EM-2 (8.7 and 17.5 nmol) given into the same VTA site did not produce any effect, but at a high dose (35 nmol) produced CPP. EM-1 (3.3 nmol) or EM-2 (17.5 nmol) microinjected into the nigrostriatal substantia nigra was not significantly different from vehicle-injected groups. D-Phe-Cys-Tyr-D-Trp-Orn-Thr-Pen-Thr-NH(2) (CTOP) at 94.13 pmol or 3-methoxynaltrexone at 0.64 pmol microinjected into the posterior Acb shell blocked EM-1-induced CPP and EM-2-induced CPA. At a higher dose, CTOP (941.3 pmol) and 3-methoxynaltrexone (6.4 pmol) produced CPA and CPP, respectively. Coadministration with antiserum against dynorphin A(1-17) (Dyn) (10 microg) microinjected into the posterior Acb shell blocked EM-2-induced CPA. However, it did not affect EM-1-induced CPP. It is concluded that EM-1 and EM-2 produce site-dependent CPP and CPA, respectively, by stimulation of different subtypes of mu-opioid-receptors; stimulation of one subtype of mu-opioid-receptor at the posterior Acb shell and VTA by EM-1 induces CPP, whereas stimulation of another subtype of mu-opioid receptor at the posterior Acb shell, but not the VTA, by EM-2 induces the release of Dyn to produce CPA.