The present study investigates the possibility that imidazoline receptors mediate modulation of cholinergic motor functions of the guinea-pig ileum. For this purpose, the effects of a series of compounds with known affinity for alpha2-adrenoceptors and/or imidazoline recognition sites were examined on the cholinergic twitch contractions evoked by electrical field stimulation (0.1 Hz) of longitudinal muscle-myenteric plexus preparations. Additional experiments were carried out on ileal strips preincubated with [3H]choline, superfused with physiological salt solution containing hemicholinium-3, and subjected to electrical field stimulation (1 Hz). The stimulation-induced outflow of radioactivity was taken as an index of endogenous acetylcholine release. Alpha-methyl-noradrenaline, noradrenaline, clonidine, medetomidine, oxymetazoline and xylazine caused a concentration-dependent inhibition of twitch responses (IC50 from 0.13 to 1.05 microM; Emax from 85.9 to 92.5%). Rilmenidine and agmatine were less potent in reducing the twitch activity, and the latter compound acted also with low intrinsic activity (IC50=44.9 microM; Emax=35.5%). In interaction experiments, the inhibitory action of clonidine on twitch responses was competitively antagonized by RX 821002 (2-(2-methoxy-1,4-benzodioxan-2-yl)-2-imidazoline), idazoxan, rauwolscine, yohimbine and BRL 44408 (2-[2H-(1-methyl-1,3-dihydroisoindole)-methyl] -4,5-dihydroimidazoline), whereas prazosin (10 microM), ARC 239 (2-(2,4-(O-methoxy-phenyl)-piperazin-1-yl)-ethyl-4,4-dimethyl- 1,3-(2H,4H)-isoquinolindione; 10 microM) and BRL 41992 (1,2-dimethyl-2,3,9,13b-tetrahydro-1H-dibenzo[c,f]imidazol[1,5-a]a zepine; 10 microM) were without effect. Rauwolscine antagonized the inhibitory effects of various agonists on ileal twitch activity in a competitive manner and with similar potency. Agmatine and idazoxan did not significantly modify the twitch contractions when tested in the presence of alpha2-adrenoceptor blockade by rauwolscine (3 microM) or RX 821002 (1 microM). Linear regression analysis showed that the affinity values of antagonists correlated with their affinity at the alpha2A and alpha2D binding sites as well as at previously classified alpha2A/D adrenoceptor subtypes, whereas no significant correlation was obtained when comparing the potency estimates of agonists and antagonists with the affinity at I1 or I2 binding sites. When tested on the electrically induced outflow of tritium, alpha-methyl-noradrenaline, noradrenaline, clonidine, medetomidine, oxymetazoline, xylazine and rilmenidine yielded inhibitory concentration-response curves which were shifted rightward to a similar extent in the presence of rauwolscine (3 microM). In the absence of further drugs, agmatine significantly reduced the evoked tritium outflow at the highest concentrations tested (10 and 100 microM), whereas idazoxan (up to 100 microM) was without effect. When RX 821002 (1 microM) was added to the superfusion medium, neither agmatine nor idazoxan modified the evoked outflow of radioactivity. The results argue against modulation by imidazoline receptors of acetylcholine release from myenteric plexus nerve terminals. They provide evidence that compounds endowed with imidazoline-like structures affect the cholinergic motor activity of the guinea-pig ileum by interacting with presynaptic alpha2-adrenoceptors belonging to the alpha2D subtype.