The irreversible mu-opioid antagonists beta-funaltrexamine (beta-FNA) and beta-chlornaltrexamine (beta-CNA) are important pharmacological tools but have a kappa-agonist activity and, in the latter case, low selectivity. This work examines whether clocinnamox (C-CAM) and the newer analog, methocinnamox (M-CAM), represent improved long-lasting antagonists for examining mu-opioid-mediated effects in vivo. beta-FNA, beta-CNA, C-CAM, and M-CAM were compared after systemic administration in mice and in vitro. beta-FNA and beta-CNA were effective agonists in the writhing assay, reversible by the kappa-antagonist norbinaltorphimine. Neither C-CAM nor M-CAM had agonist activity in vivo. M-CAM was devoid of agonist action at cloned opioid receptors. All four compounds depressed the dose-effect curve for the mu-agonist morphine in the warm-water tail-withdrawal test 1 h after administration; at 48 h, recovery was evident. In the writhing assay, the dose-effect curve for morphine was shifted in a parallel fashion in the order M-CAM >> C-CAM > beta-CNA > or = beta-FNA. In comparison with their ability to shift the dose-effect curve for bremazocine (kappa) and BW373U86 (delta), beta-CNA was the least mu-selective, followed by C-CAM < beta-FNA < M-CAM. M-CAM (1.8 mg/kg) produced a 74-fold increase in the ED(50) of morphine while showing no effect on bremazocine or BW373U86 dose-response curves. In binding assays, C-CAM and M-CAM were 8-fold selective for mu- over kappa-receptors, whereas beta-FNA and beta-CNA were mu/delta-, but not mu/kappa, selective. However, ex vivo binding assays confirmed the mu-receptor selectivity of M-CAM. M-CAM is thus a potent, long-lasting, and specific antagonist at mu-receptors in vivo that lacks confounding agonist actions.