A series of position 4-substituted endomorphin-2 (Tyr-Pro-Phe-Phe-NH2) analogs containing 3-(1-naphthyl)-alanine (1-Nal) or 3-(2-naphthyl)-alanine (2-Nal) in L- or D-configuration, was synthesized. The opioid activity profiles of these peptides were determined in the mu-opioid receptor representative binding assay and in the Guinea-Pig Ileum assay/Mouse Vas Deferens assay (GPI/MVD) bioassays in vitro, as well as in the mouse hot-plate test of analgesia in vivo. In the binding assay the affinity of all new analogs for the mu-opioid receptor was reduced compared with endomorphin-2. The two most potent analogs were [D-1-Nal(4)]- and [D-2-Nal4]endomorphin-2, with IC50 values 14 +/- 1.25 and 19 +/- 2.1 nM, respectively, compared with 1.9 +/- 0.21 nM for endomorphin-2. In the GPI assay these analogs were found to be weak antagonists and they were inactive in the MVD assay. The in vitro GPI assay results were in agreement with those obtained in the in vivo hot-plate test. Antinociception induced by endomorphin-2 was reversed by concomitant intracerebroventricula (i.c.v.) administration of [D-1-Nal4]- and [D-2-Nal4]-endomorphin-2, indicating that these analogs were mu-opioid antagonists. Their antagonist activity was compared with that of naloxone. At a dose 5 microg per animal naloxone almost completely inhibited antinociceptive action of endomorphin-2, while [D-1-Nal4]endomorphin-2 in about 46%.