1. The ability of a range of substituted imidazole compounds to inhibit mouse cerebellar neuronal nitric oxide synthase (nNOS), bovine aortic endothelial NOS (eNOS) and inducible NOS (iNOS) from lungs of endotoxin-pretreated rats was investigated. In each case the substrate (L-arginine) concentration employed was 120 nM. 2. 1-(2-Trifluoromethylphenyl) imidazole (TRIM) was a relatively potent inhibitor of nNOS and iNOS (IC50S of 28.2 microM and 27.0 microM respectively) but was a relatively weak inhibitor of eNOS (IC50, 1057.5 microM). The parent compound, imidazole, was a weak inhibitor of all three NOS isoforms (IC50S: nNOS, 290.6 microM; eNOS, 101.3 microM; iNOS, 616.0 microM). Substitution of imidazole with a phenyl group to yield I-phenylimidazole (PI) resulted in an isoform non-selective increase in inhibitory potency (IC50S: nNOS, 72.1 microM; eNOS, 86.9 microM; iNOS, 53.9 microM). Further substitution of the attached phenyl group resulted in an increase in nNOS and a decrease in eNOS inhibitory potency as in TRIM, 1-chlorophenylimidazole (CPI; IC50S: nNOS, 43.4 microM; eNOS, 392.3 microM; iNOS, 786.5 microM) and 1-(2,3,5,6-tetrafluorophenyl) imidazole (TETRA-FPI; IC50S; nNOS, 56.3 microM; eNOS, 559.6 microM; iNOS, 202.4 microM). 3. The ability of TRIM to inhibit mouse cerebellar nNOS activity in vitro was influenced by the concentration of L-arginine (0.12-10.0 microM) in the incubation medium. When mouse cerebellar nNOS was used as enzyme source a double reciprocal (Lineweaver-Burk) plot in the presence/absence of TRIM (50 microM) revealed a competitive inhibitory profile. The K(m) for L-arginine and the Ki for TRIM calculated from these data were 2.4 microM and 21.7 microM, respectively. The ability of TRIM to inhibit mouse cerebellar nNOS activity in vitro was unaffected by varying the time of exposure of the enzyme to TRIM from 0-60 min at 0 degree C. 4. TRIM exhibits potent antinociceptive activity in the mouse as evidenced by inhibition of acetic acid induced abdominal constrictions. The ED50 for TRIM following i.p. administration was 20 mg kg-1 (94.5 mumol kg-1). The antinociceptive effect of TRIM was reversed by pretreatment of animals with L-arginine (50 mg kg-1, i.p.) and was not accompanied by sedation, motor ataxia or behavioural changes (rearing, crossing, circling, dipping) as assessed by use of a box maze procedure. 5. L-NG nitro arginine methyl ester (L-NAME, 20 mg kg-1, i.v.) but not TRIM (0.5-20 mg kg-1, i.v.) increased mean arterial blood pressure (MAP) in the urethane-anaesthetized rat. 6. L-NAME (100 microM) potentiated the contractile response of the rabbit isolated aorta to phenylephrine (ED50; 0.084 +/- 0.01 microM in the presence and 0.25 +/- 0.05 microM in the absence of L-NAME; maximum response, 7.7 +/- 0.4 g in the presence and 5.6 +/- 0.5 g in the absence of L-NAME, n = 6, (P < 0.05) whilst TRIM (1-100 microM) was without effect. L-NAME (100 microM) but not TRIM (1-100 microM) also reduced carbachol-induced relaxation of the phenylephrine-precontracted rabbit aorta preparation. 7. L-NAME (50 microM) potentiated the vasoconstrictor effect of bolus-injected noradrenaline (10-1000 nmol) and reduced the vasodilator effect of carbachol (10 microM) added to the Krebs reservoir in the rat perfused mesentery preparation. L-NAME (50 microM) also reduced nitric oxide (NO) release (measured by chemiluminescence of nitrite in the Krebs perfusate) in response to noradrenaline (100 nmol; 53.8 +/- 4.0 pmol ml-1 in the presence and 84.8 +/- 8.0 pmol ml-1 in the absence of L-NAME, n = 15, P < 0.05) and carbachol (10 microM; 63.9 +/- 5.0 pmol ml-1 in the presence and 154.0 +/- 9.0 pmol ml-1 in the absence of L-NAME, n = 15, P < 0.05). TRIM (50 microM) did not affect either the vasoconstrictor response to noradrenaline or the vasodilator response to carbachol or the accompanying release of NO from the perfused rat mesentery.