As part of a major study to evaluate the mutagenicity of chemicals produced during the cooking of foods, we examined the responses of bacteria and cultured Chinese hamster cells to the compounds Trp-P-2 (3-amino-1-methyl-5H-pyrido[4,3-b]indole) and IQ (2-amino-3-methylimidazo[4,5-f]quinoline), constituents identified in cooked beef and fish. In the Ames/Salmonella tester strain TA1538, both compounds were confirmed to be extremely potent mutagens that were active at levels below 1 ng/plate in the presence of hamster-liver S9 microsomal fraction. 50-fold higher doses of both compounds were required for mutagenicity in the uvr+ tester strain TA1978. Trp-P-2 also behaved as a strong mutagen in CHO cells using the standard exogenous activation with hamster-liver S9 fraction. At concentrations below 1 microgram/ml it produced dose-dependent increases in cell killing, mutations at the hprt and aprt loci, sister-chromatid exchanges, and chromosomal aberrations. An excision-repair-deficient strain was about 2-fold more sensitive than the normal CHO cells with respect to these genotoxic effects of Trp-P-2. IQ had unexpectedly weak activity for all genetic endpoints in the CHO cells, and it produced clear-cut responses only in the repair-deficient cells and only above a concentration of 10 micrograms/ml. The toxicity that was observed with IQ was not affected by the repair capacity of the cells and was not associated with chromosomal aberrations, indicating that damage to cellular structures other than nuclear DNA was likely the predominant pathway for cell killing. Because the culture conditions normally used for CHO cell exposure were shown to be competent in producing bacterial mutagenicity with IQ, it was concluded that the active metabolite of IQ was present in the medium but was somehow ineffective in reaching the DNA of CHO cells and/or reacting with it. These results suggest that the relative mutagenic potency of compounds in Salmonella may bear no direct relationship to relative mutagenicity in CHO cells, emphasizing precaution in attempting to extrapolate microbial data to mammalian somatic cells. This study illustrates the use and merits of a multi-endpoint assay for genetic damage in CHO cells, the utility of using CHO cells that are defective in excision repair of DNA, and the importance of comparative testing between bacterial and mammalian systems.