Although oxazepam (Serax), a widely used benzodiazepine anxiolytic, does not induce gene mutations in vitro or chromosomal aberrations in vivo, it was found to be a hepatocarcinogen in a 2 year bioassay in B6C3F1 mice. Thus, it was of interest to determine whether this carcinogen is mutagenic in vivo. Male B6C3F1 Big Blue transgenic mice were fed 2500 p.p.m. oxazepam or control diet alone for 180 days and killed on the next day. The mutant frequency (MF) of lacI in control mice was 5.02 +/- 2.4x10(5), whereas the MF in the oxazepam-treated mice was 9.17 +/- 4.82x10(-5), a significant increase (P < 0.05). Correction of the mutant frequency of lacI from the oxazepam-treated mice for clonality resulted in a decrease in the mean mutant frequency to 8.15 +/- 2. 54x10(-5). Although the mutant frequency difference was small, sequencing of a random collection of the mutants from each oxazepam-exposed mouse showed a significant difference (P < 0.015) in the mutation spectrum compared with that from control mice. In the oxazepam-exposed mice, an increase in G:C-->T:A and G:C-->C:G transversions and a concomitant decrease in G:C-->A:T transitions were observed. Clonal expansion of mutations at guanines in 5'-CpG-3' sequencing contexts at three sites was noted. It is postulated that some of the mutations found in the oxazepam-derived spectrum were due to oxidative damage elicited by induction of CYP2B isozymes as the result of chronic oxazepam administration. This study demonstrates that the in vivo Big Blue transgenic rodent mutation assay can detect mutations derived from a carcinogen that did not induce gene mutations in vitro or micronuclei in mouse bone marrow. Moreover, the sequencing of the recovered mutants can distinguish between the mutation spectrum from treated mice compared with that from control mice, thereby confirming the genotoxic consequences.