Previous studies measuring opioid inhibition of cyclic adenosine monophosphate in SH-SY5Y cells supported the hypothesis that continuous agonist stimulation causes a gradual conversion of the mu opioid receptor to a sensitized or constitutively active state termed mu*. Conversion to mu* was prevented by the kinase inhibitor H7, but not its close analog H8. Naloxone was proposed to act as a negative antagonist (inverse agonist) blocking mu* activity, whereas D-Phe-Cys-Tyr-D-Trp-Arg-Thr-Pen-Thr-NH2 (CTAP) appeared to act as a neutral antagonist having no effect on mu* activity. Initial in vivo results indicated that mu* activity may play a role in narcotic tolerance and dependence (Wang et al., Life Sci. 54: PL339-PL350 1994). Our study explores the pharmacology of H7 and H8, naloxone and CTAP in mice after induction of acute tolerance and dependence induced by a single s.c. dose of morphine (100 mg/kg). Physical dependence was defined by withdrawal jumping induced by i.p. naloxone injections 4 hr after the morphine dose, the time of maximal physical dependence. Neither H7 nor H8 (50 nmol or less) induced jumping, affected morphine antinociception or produced significant behavioral effects, when injected by the intracerebroventricular (i.c.v.) or intrathecal (i.th.) routes. When given 30 min before the naloxone challenge, H7, but not H8, significantly reduced naloxone jumping by i.c.v. injection. Administration of naloxone into the central nervous system, rather than by i.p. administration, required coinjection by both i.c.v. and i.th. routes to elicit full withdrawal jumping (30 nmol at each site). In contrast, the putative neutral antagonist CTAP caused little withdrawal jumping when coinjected i.c.v. and i.th., as expected if modulation of mu* activity played a role in dependence. However, CTAP was capable of partially reversing naloxone (i.p.) induced jumping when given either i.c.v. or i.th., indicating that CTAP competes with naloxone at mu*. Moreover, these results demonstrate that both spinal and supraspinal sites are required for full opioid withdrawal jumping in mice. Antinociceptive tolerance was also evaluated by determining the response to morphine in the 55 degrees C warm-water tail-flick test. Morphine pretreatment (100 mg/kg, s.c., -5 hr) produced antinociceptive tolerance as shown by a 2.7-fold increase in the calculated morphine A50 value. Tolerance was reversed by H7, but not H8, treatment (50 nmol, i.c.v., -30 min). These results are consistent with the hypothesis that a sensitized or constitutively active mu* state plays a role in narcotic tolerance and dependence.