EMS-induced mutant frequency and mutation spectrum as well as background mutant frequency have been characterized fur bone marrow of the D6-2 transgenic mice. ThelacI genes carried on pSPORT1 vectors were recovered from the treated or untreated mouse genomic DNA by excision and circularization, and analyzedin vitro for mutations that occurred in the mouse bone marrow, lacI(-) mutants were positively selected with the M9/L media. The 6 lacI(-) mutants were identified out of 11 935 vectors recovered from genomic DNA of the treated mice (mutant frequency was 50 x 10(5)), while no mutant was found in 11 649 vectors Imm untreated mice (the background mutant frequency wan lower than 8.6 x 10(-5)). Two regions oflacI for each mutant, in which the majority of sensitive sites for inactivation of thelacI gene product have been located, were sequenced and 16 mutation events were identified. The predominant mutations (14/16 or 87.5%) were base substitutions, whereas the remaining 2 mutations were single base deletions (12.5%). Of these base substitutions, transversions made up 9/14 or 64%, and transitions cornprised 5/14 or 36%, These findings were markedly different from the spontaneous spectra characterized by using Big-Blue system, as well as from the EMS-induced mutation spectra obtained within vitro assay systems, where the EMS-induced predominant mutations are CG --> AT transitions. In addition, 45% of mutations analyzed occurred at CpG dinucleotides, which was in accordance with previous studies with other systems. These data show that: (i) the D6-2 transgenic mouse lineage is a suitable mdel for studying mutagenesisin vivo; (ii) a fundamental difference in mutagenesis for EMS betweenin nitro andin vivo assay systems may exist, but more extensive sequence analyses are required to determine the possible differences in mutation spectra.