1843U89 is a potent inhibitor (Ki = 0.09 nM) of thymidylate synthase (TS; EC 2.1.1.45) that is in clinical trial for the treatment of solid tumors. Although it is an excellent substrate for the folate anabolizing enzyme folylpolyglutamate synthetase (FPGS), 1843U89 differs from other folate-based inhibitors of TS (e.g., CBC3717, D1694, and LY231514), in that the parent compound is as potent an enzyme inhibitor as its polyglutamated analogues. As reported (D.S. Duch et at., Cancer Res., 53:810-818, 1993), 1843U89 is 10-80-fold more cytotoxic than the close structural analogue 1031U89, which is an equipotent inhibitor of TS but is a less efficient substrate for FPGS. This correlation between substrate efficiency for FPGS and cytotoxicity suggests that polyglutamation of 1843U89 contributes to its cytotoxicity. In the current study, we measured intracellular levels of polyglutamated anabolites of 1031U89, 1843U89, and three other benzoquinazoline inhibitors of TS as well as anabolites of D1694 in HCT-8 ileocecal carcinoma cells. Each TS inhibitor was anabolized to polyglutamated analogues with one to five added glutamyl residues after exposure for 24 h to IC90 concentrations (those that inhibit growth by 90% after 72 h of constant exposure). D1694, which requires polyglutamation for potent enzyme inhibition as well as for cytotoxicity, was anabolized mostly to penta- and hexaglutamates, whereas approximately 80% of intracellular 1843U89 was the diglutamate analogue. The substrate efficiency of the benzoquinazolines for FPGS was predictive of the extent of intracellular anabolism. The diglutamate analogue of 1843U89 was only 1/100 as efficient a substrate for further glutamation as was 1843U89 itself. The efficient anabolism to the diglutamate analogue and the lack of dependence on further polyglutamation for enzyme inhibition or cytotoxicity provide a rationale for the reported 1843U89 sensitivity of cells with impaired FPGS activity. As part of an investigation of the effects of polyglutamation, we measured the retention of intracellular 1843U89 and D1694 anabolites after 24 h of exposure to 20 nM of each compound. After 48 h in drug-free medium, 7% of intracellular 1843U89 (mostly diglutamate analogue) and 36% of D1694 (mostly penta- and hexaglutamates) remained in the cells. Because prolonged retention (associated with tissue storage of polyglutamates) can contribute to clinical toxicities, 1843U89 may present fewer long-term toxicities than D1694.