The CAT-Tox (L) assay has recently been developed and validated for detecting and quantifying the specific molecular mechanisms that underlie toxicity of various xenobotic chemicals. We performed this assay to measure the transcriptional responses associated with 2,4,6-trinitrotoluene (TNT) and 2 of its byproducts [2,4 and 2,6-dinitotoluenes (DNTs)] to 13 different recombinant cell lines generated from human liver carcinoma cells (HepG2) by creating stable transfectants of mammalian promoter chloramphenicol acetyltransferase (CAT) gene fusions. Cytoxicity test with the parental HepG2 cells, using the MTT [3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide]-based assay for cell viability, yielded LC50 values of 105 +/- 6 mg/mL for TNT in 1% dimethyl sulfoxide (DMSO), and > 300 mg/mL for DNTs, upon 48 h of exposure. TNT appeared to be more toxic than 2,4-DNT, which also showed a higher toxicity compared to 2,6-DNT. Of the 13 recombinant constructs evaluated, 8 (CYP 1A1, GST Ya, XRE, HMTIIA, c-fos, HSP70, GADD153, and GADD45), 5 (c-fos, HSP70, GADD153, GADD45, and GRP78), and none showed inductions to significant levels (p < 0.05), for TNT, 2,4-DNT, and 2,6-DNT, respectively. For most constructs, the induction of stress genes was concentration-dependent. These results show the potential for TNT and 2,4-DNT to cause protein damage and/or perturbations of protein biosynthesis (HSP70 and GRP78), alterations in DNA sequence or its helical structure (c-fos, GADD153, GADD45), and the potential involvement of TNT in the biotransformation process (CYP 1A1, GST Ya, XRE), and in the toxicokinetics of metal ions (HMTIIA). Within the range of concentrations tested (0-300 mg TNT or DNT/mL in 1% DMSO), no significant inductions (p > 0.05) of NFKBRE, p53RE, CRE, and RARE were found.