Repeated forced-swim stress (FSS) produced analgesia, immobility and potentiation of cocaine-conditioned place preference (CPP) in wild-type C57Bl/6 mice, but not in littermates lacking the kappa opioid receptor (KOR) gene. These results were surprising because kappa agonists are known to produce conditioned place aversion and to suppress cocaine-CPP when coadministered with cocaine. The possibility that disruption of the kappa system blocked the stress response by adversely affecting the hypothalamic-pituitary axis was examined by measuring plasma corticosterone levels. However, disruption of the dynorphin/kappa system by gene deletion or receptor antagonism did not reduce the FSS-induced elevation of plasma corticosterone levels. A second explanation for the difference is that kappa receptor activation caused by FSS occurred prior to cocaine conditioning rather than contemporaneously. To test this hypothesis, we measured the effects of the kappa agonist (trans)-3,4-dichloro-N-methyl-N-[2-(1-pyrrolidinyl)-cyclohexyl]benzeneacetamide (U50,488) administered to mice at various intervals preceding cocaine conditioning. The results showed that the interaction between the kappa system and cocaine reinforcement depended on the timing of the drug pairing. Mice given U50,488 60 min prior to cocaine showed a robust, nor-BNI-sensitive potentiation of cocaine-CPP, whereas administration 15 min before cocaine significantly suppressed cocaine-CPP. In the absence of cocaine, U50,488 given 60 min prior to saline conditioning produced no place preference, whereas administration 15 min before saline conditioning produced significant place aversion. The results of this study suggest that kappa receptor activation induced by FSS prior to the cocaine-conditioning session may be both necessary and sufficient for potentiation of the reinforcing actions of cocaine.