BACKGROUND The formation of salsolinol (SAL) has been hypothesized to be a factor contributing to alcoholism and alcohol abuse. If SAL is formed under chronic alcohol-drinking conditions, then it may contribute to alcohol addiction by being rewarding itself. Because SAL can be formed by the nonenzymatic condensation of acetaldehyde with dopamine, the reinforcing effects of SAL were tested in the nucleus accumbens shell, a dopamine-rich site considered to be involved in regulating alcohol-drinking behavior. METHODS The intracranial self-administration technique was used to test the reinforcing properties of SAL. Adult, female alcohol-preferring (P) rats were stereotaxically implanted with guide cannulae aimed at the nucleus accumbens shell. After 7 to 10 days to allow recovery from surgery, P rats were attached to the electrolytic microinfusion transducer system, placed in two-lever experimental chambers, and allowed to respond for the self-infusion of 100 nl of modified artificial cerebrospinal fluid (aCSF) or 0.03, 0.3, 3.0, or 12.5 microM SAL (3-1250 fmol/100 nl). Sessions were 4 hr in duration and were conducted in the dark cycle every 48 hr. The effects of coinfusing 10 to 400 microM sulpiride (given in sessions 5 and 6 after four acquisition sessions) on the intracranial self-administration of 3.0 microM SAL were tested in a separate experiment. RESULTS P rats given 0.3 to 12.5 microM SAL received significantly more infusions per session than did the group given aCSF alone (e.g., 50 infusions for 3.0 microM SAL versus 10 or fewer infusions for the aCSF group) and responded significantly more on the active than inactive lever. Coinfusion of 100 or 400 microM sulpiride reduced the responding on the active lever (80-100 responses/session without sulpiride) to levels observed for the inactive lever (fewer than 10 responses/session with sulpiride). This effect was reversible because giving SAL alone in session 7 reinstated responding on the active lever. CONCLUSIONS SAL is reinforcing in the nucleus accumbens shell of P rats at concentrations that are pharmacologically possible, and these reinforcing actions are mediated in part by D2/D3-like receptors.