We have previously demonstrated that both endomorphin-1 (EM-1) and endomorphin-2 (EM-2) at high doses (1.75-35 nmol) given intrathecally (i.t.) or intracerebroventricularly produce antinociception by stimulation of -opioid receptors. Now, we report that EM-2 at small doses (0.05-1.75 nmol), which injected alone did not produce antinociception, produces antianalgesia against opioid agonist-induced antinociception. The tail-flick (TF) response was used to test the antinociception in male CD-1 mice. Intrathecal pretreatment with EM-2 (0.02-1.75 nmol) 45 min before i.t. morphine (3.0 nmol) injection dose dependently attenuated morphine-induced TF inhibition. On the other hand, a similar dose of EM-1 (1.64 nmol) failed to produce any antianalgesic effect. The EM-2 (1.75 nmol)-produced anti-analgesia against morphine-induced TF inhibition was blocked by i.t. pretreatment with the -opioid antagonist naloxone or 3-methoxynaltrexone, but not ␦-opioid receptor antagonist naltrindole, -opioid receptor antagonist nor-binal- Endogenous opioid tetrapeptides endomorphin-1 (EM-1) and endomorphin-2 (EM-2) have been found to be highly selective for -opioid receptors (Zadina et al., 1997). Both EM-1 and EM-2 potently compete with -opioid receptor binding with no appreciable affinity with ␦-and -opioid receptors and selectively activate -opioid receptor-mediated G-proteins with [35 S]guanosine 5Ј-O-(3-thio)triphosphate binding (Goldberg et al., 1998;Narita et al., 1998Narita et al., , 2000Monory et al., 2000). The increases of the [ 35 S]guanosine 5Ј-O-(3-thio)triphosphate binding and antinociceptive effects induced by both EM-1 and EM-2 are selectively blocked by the pretreatment with selective -opioid receptor antagonist -funaltrexamine or D-Phe-Cys-Tyr-D-Trp-Orn-Thr-Pen-Thr-NH 2 , indicating that the effects are mediated by the stimulation of -opioid receptors.However, recent studies indicate that the antinociceptive effects induced by EM-1 and EM-2 are mediated by the stimulation of different subtypes of -opioid receptors. Pretreatment with -opioid receptor antagonist 3-methoxynaltrexone selectively attenuated EM-2-but not EM-1-induced antinociception, whereas -funaltrexamine inhibits both (Sakurada et al., 2000). There is an asymmetric cross-tolerance between EM-1 and EM-2, where mice made acutely tolerant to EM-1 are not cross-tolerant to EM-2, but mice made acutely tolerant to EM-2 cause partial cross-tolerance to EM-1 (Wu et al., 2001). Intrathecal pretreatment with antisense oligodeoxynucleotides against exon-8 of -opioid receptor clone inhibits the antinociception induced by EM-1