1. The effect of Tityus serrulatus scorpion venom and its toxin components on the rabbit isolated corpus cavernosum was investigated by use of a bioassay cascade. 2. Tityus serrulatus venom (3-100 microg), acetylcholine (ACh; 0.3-30 nmol) and glyceryl trinitrate (GTN; 0.5-10 nmol) dose-dependently relaxed rabbit isolated corpus cavernosum preparations precontracted with noradrenaline (3 microM). The selective soluble guanylate cyclase inhibitor 1H-[1,2,4] oxadiazolo [4,3,-alquinoxalin-1-one] (ODQ; 30 microM) increased the basal tone of the rabbit isolated corpus cavernosum and abolished the relaxations induced by the agents mentioned above. Methylene blue (30 microM) also inhibited the relaxations induced by Tityus serrulatus venom but, in contrast to ODQ, the inhibition was irreversible. 3. The non-selective NO synthase (NOS) inhibitors Nomega-nitro-L-arginine methyl ester (L-NAME; 10 microM) and NG-iminoethyl-L-ornithine (L-NIO; 30 microM) also increased the tone of the rabbit isolated corpus cavernosum and markedly reduced both ACh- and Tityus serrulatus venom-induced relaxations without affecting those evoked by GTN. The inhibitory effect was reversed by infusion of L-arginine (300 microM), but not D-arginine (300 microM). The neuronal NOS inhibitor 1-(2-trifluoromethylphenyl) imidazole (TRIM, 100 microM) did not affect either the tone of the rabbit isolated corpus cavernosum or the relaxations induced by ACh, bradykinin (Bk), Tityus serrulatus venom and GTN. TRIM was approximately 1,000 times less potent than L-NAME in inhibiting rabbit cerebellar NOS in vitro, as measured by the conversion of [3H]-L-arginine to [3H]-L-citrulline. 4. The protease inhibitor aprotinin (Trasylol; 10 microg ml[-1]) and the bradykinin B2 receptor antagonist Hoe 140 (D-Arg-[Hyp3,Thi5,D-Tic7, Oic8]-BK; 50 nM) did not affect the rabbit isolated corpus cavernosum relaxations induced by Tityus serrulatus venom. The ATP-dependent K+ channel antagonist glibenclamide (10 microm) and the Ca2+-activated K+ channel antagonists apamin (0.1 microM) and charybdotoxin (0.1 microM) also failed to affect the venom-induced relaxations. Similarly, the K+ channel blocker tetraethylammonium (TEA; 10 microM) had no effect on the venom-induced relaxations. 5. Capsaicin (3 and 10 nmol) relaxed the rabbit isolated corpus cavernosum in a dose-dependent and non-tachyphylactic manner. Ruthenium red (30 microM), an inhibitor of capsaicin-induced responses, markedly reduced the relaxations caused by capsaicin, but failed to affect those induced by Tityus serrulatus venom. L-NAME (10 microM) had no effect on the capsaicin-induced relaxations of the rabbit isolated corpus cavernosum. 6. The sodium channel blocker tetrodotoxin (TTX; 1 microM) abolished the relaxations of the rabbit isolated corpus cavernosum induced by Tityus serrulatus venom without affecting those evoked by capsaicin, ACh and GTN. Tetrodotoxin (1 microM) also promptly reversed the response to the venom when infused during the relaxation phase. 7. The bioassay cascade of the toxin components purified from Tityus serrulatus venom revealed that only fractions X, XI and XII caused dose-dependent relaxations of the rabbit isolated corpus cavernosum and these were markedly reduced by either TTX (1 microM) or L-NAME (10 microM). 8. Our results indicate that Tityus serrulatus scorpion venom (and the active fractions X, XI and XII) relaxes rabbit corpus cavernosum via the release of NO. This release is specifically triggered by the activation of capsaicin-insensitive cavernosal non-adrenergic non-cholinergic (NANC) fibres, that may possibly be nitrergic neurones. Tityus serrulatus venom may therefore provide an important tool for understanding further the mechanism of NANC nitrergic nerve activation.