1. Site-directed mutagenesis of human plasma butyrylcholinesterase has led to novel hydrolases that rapidly destroy cocaine. We are investigating whether viral gene transfer of such enzymes might reduce addiction liability by blocking cocaine from its sites of action. 2. As groundwork for a possible gene therapy, we previously studied adenoviral transduction of cocaine hydrolases in the rat. Systemically injected vectors raised plasma cocaine hydrolase activity greatly, reduced pressor responses to cocaine, and lowered cocaine's tissue burden. 3. In the present study, to reduce cocaine's brain access still further, vectors were injected directly into the nucleus accumbens. Six days later, medium sized neurons gained dramatic butyrylcholinesterase activity. Species-selective immunohistochemistry proved that the transgene accounted for this activity. 4. Since the transgene product is so efficient with cocaine as a substrate, it is now reasonable to begin testing gene therapy in rodent models of cocaine addiction.