First, the antinociception with the tail-flick test of D-Pro(2)-endomorphin-1 and D-Pro(2)-endomorphin-2 given i.t. was compared with that produced by endomorphin-1 and -2 in male CD-1 mice. High doses of D-Pro(2)-endomorphin-1 (0.2-0.4 pmol) and D-Pro(2)-endomorphin-2 (300-800 pmol) given i.t. produced antinociception with low intrinsic activity [about 25% maximum possible effect (MPE)] compared with that of endomorphin-1 (16.4 nmol) and endomorphin-2 (35 nmol) (>90% MPE). Second, coadministration of a low dose of D-Pro(2)-endomorphin-1 (0.1 pmol), which given alone did not affect the tail-flick latencies, markedly attenuated the antinociception induced by endomorphin-1 (16.4 nmol) but not by endomorphin-2 (35 nmol). Similarly, coadministration of a low dose of D-Pro(2)-endomorphin-2 (200 pmol), which given alone did not affect the tail-flick latencies, significantly attenuated the antinociception induced by endomorphin-2 (35 nmol) and, to a much lesser extent, endomorphin-1 (16.4 nmol). It is concluded that D-Pro(2)-endomorphin-1 and D-Pro(2)-endomorphin-2 at high doses were partial opioid receptor agonists to produce antinociception, and at low doses were opioid receptor antagonists to block selectively the antinociception induced by endomorphin-1 and endomorphin-2, respectively. Furthermore, our results are consistent with the view that the antinociception induced by endomorphin-1 and endomorphin-2 is mediated by the stimulation of different subtypes of mu-opioid receptors.