We have reported that noncytotoxic concentrations of 3'-azido-3'-deoxythymidine (AZT) increase the cytotoxicity of ICI D1694, a folate-based thymidylate synthase (TS) inhibitor, with increasing AZT incorporation into DNA. We postulated that the inhibition of TS by ICI D1694 would decrease 5'-deoxythymidine triphosphate (dTTP) pools, which compete with AZT triphosphate (AZT-TP) as a substrate for DNA polymerase. Furthermore, the inhibition of TS would increase the activity of both thymidine kinase (TK) and thymidylate kinase (TdK). Each of these consequences of TS inhibition would favor more incorporation of AZT into DNA. Thus, we reasoned that other TS inhibitors should also result in increased AZT incorporation into DNA and, perhaps, in increased cytotoxicity. N6-[4-(Morpholinosulfonyl)benzyl]-N6-methyl-2,6-diaminobenz[ cd]indole glucuronate (AG-331) differs from ICI D1694 in that it is a de novo designed lipophilic TS inhibitor, it does not require a specific carrier for cellular uptake, and it does not undergo intracellular polyglutamation. This potent TS inhibitor causes minimal cytotoxicity in MGH-U1 human bladder cancer cells. A 24-h exposure to 5 microM AG-331 causes almost complete TS inhibition but only 35% cell kill. The combination of AZT and AG-331 in MGH-U1 cells resulted in an enhanced antitumor effect relative to that of each agent alone; 50 microM AZT, noncytotoxic alone, increased the cell kill of induced by AG-331 from 35% to 50%. Biochemical studies of this combination revealed that simultaneous treatment with 5 microM AG-331 plus 1.8 microM [3H]-AZT produced as much as a 68% +/- 7% increase in AZT incorporation into DNA. This observation was associated with an increase in DNA single-strand breaks, measured as comet tail moment, of up to 6.6-fold. These studies support our original premise that TS inhibition favors increased incorporation of AZT into DNA and that the combination causes more cell kill than either drug alone in MGH-U1 cells.