No / NAD(P)H:quinone oxidoreductase-1 (NQO1) is a potential target for therapeutic intervention but attempts to exploit NQO1 using quinone-based bioreductive prodrugs have been largely compromised by toxicity to organs that inherently express high levels of NQO1. In an attempt to circumvent this problem, this study describes the development of a tripartite quinone-based drug delivery system, the ultimate objective of which is to release a targeted therapeutic agent following the reduction of a quinone "trigger" by NQO1. Molecular modeling of drug/NQO1 interactions were conducted prior to the synthesis of N-{4-[bis-(2-chloroethyl)-amino]-phenyl}-beta,beta,2,4,5-pentamethyl-3,6-dioxo-1,4-cyclohexadiene-1-propanamide (prodrug 1). Prodrug 1 is a good substrate for purified NQO1 (V(max) and K(m) values of 11.86 +/- 3.09 micromol/min/mg and 2.70 +/- 1.14 micromol/L, respectively) and liquid chromatography-mass spectrometry analysis of the metabolites generated showed that lactone 3 and aniline mustard 4 were generated in a time- and NQO1-dependent manner. Chemosensitivity studies showed that prodrug 1 is selectively toxic to cells that overexpress NQO1 under aerobic conditions, and comet assay analysis confirmed the presence of elevated interstrand cross-links in NQO1-rich compared with NQO1-deficient cells. Hypoxic sensitization (hypoxic cytotoxicity ratio = 15.8) was observed in T47D cells that overexpress cytochrome P450 reductase. In conclusion, the results of this study provide mechanistic proof of principle that a tripartite benzoquinone drug delivery system is enzymatically reduced to release an active therapeutic agent. Further development of this concept to fine-tune substrate specificity for specific reductases and/or the inclusion of alternative therapeutic agents is warranted.