This study examines the role of NAD(P)H:quinone acceptor oxidoreductase (NQOR) (EC 1.6.99.2) in the metabolism of aziridinylbenzoquinones and the ensuing formation of reactive oxygen species in the induction of the cell cycle inhibitor p21 (WAF1, Cip1, or sdi1) in human colon carcinoma cells. The aziridinylbenzoquinones used were 2,5-diaziridinyl-1,4-benzoquinone (DZQ) and 2,5-bis(carboethoxyamino)-3,6-diaziridinyl-1,4-benzoquinone (AZQ). The cell lines used in this study, BE and HT29 human colon carcinoma cell lines, are devoid of and overexpress NQOR activity, respectively. The rate of reduction of the above quinones in BE cells proceeded at similar rates (approximately 170 nmol/min/ mg protein) and, expectedly, it was not affected by the NQOR inhibitor, dicumarol. The metabolism of DZQ in HT29 cells was largely accomplished by NQOR (approximately 94%), whereas that of AZQ was accomplished by dicumarol-insensitive reductases. The metabolism of DZQ in HT29 cells was accompanied by H2O2 formation, which was approximately 10-fold higher than that ensuing from the activation of AZQ. In agreement with these data, the production of H2O2 during the activation of DZQ by purified NQOR was approximately 10-fold higher than that of AZQ. The formation of H2O2 during the metabolism of aziridinylbenzoquinones in BE cells was 24- to 57-fold lower than that in HT29 cells. At variance with HT29 cells, H2O2 formation by BE cells was insensitive to the catalase inhibitor sodium azide. The bioactivation of AZQ and DZQ in BE cells yielded O2.- and HO. as detected by spin trapping/EPR, the intensity of the former adduct being approximately 2-fold higher than that of the latter. These signals were insensitive to dicumarol. The metabolism of DZQ in HT29 cells yielded mainly HO. and a modest contribution of O2.- (ratio HO./O2.- approximately 10), whereas that of AZQ yielded a HO./O2.- approximately 2. The effect of dicumarol on the free radical pattern obtained during DZQ metabolism resulted in a strong inhibition (80%) of HO. production and a substantial increase of O2.- generation. The metabolism of DZQ and AZQ in BE cells was associated with a significant increase of p21 mRNA levels; the former quinone was approximately 2-fold more efficient than the latter. DZQ metabolism in HT29 cells led to an increase of p21 mRNA levels 15-fold higher than that observed with AZQ activation. Dicumarol did not inhibit p21 induction associated with the metabolism of DZQ in the NQOR-deficient BE cells, whereas the inhibitor decreased p21 induction in HT29 cells by approximately 30%. This modest inhibition is likely due to the low concentration of dicumarol used, which did not affect p21 constitutive levels in control experiments carried out in the absence of the quinone. p21 induction in HT29 cells was also inhibited by DTPA, a metal chelator, and N-acetylcysteine, a potent cellular anti-oxidant, suggesting that HO. may serve as an ultimate mediator for the induction. It may be surmised that the higher efficiency of DZQ in p21 induction may be related to its efficient metabolism by NQOR in HT29 cells and the associated high level of reactive oxygen species. The role of reactive oxygen species in p21 induction was further assessed upon supplementation of cells with H2O2:p21 induction in BE cells was 4-fold higher than that in HT29 cells. These findings suggest that assessment of the role of NQOR and reactive oxygen species in p21 induction requires careful consideration of the cell genotype.