Recently, mutagenic activity on several strains of Salmonella typhimurium has been found in many heat-processed foodstuffs. The previously reported direct-acting mutagenic activity of coffee in Salmonella typhimurium TA100 (Ames assay) was confirmed in our study. In addition to TA100, a mutagenic effect of coffee was also found by using the newly developed strain TA102. The mutagenic activity was abolished by the addition of rat-liver homogenate. 10% S9 mix completely eliminated the mutagenic activity of 30 mg of coffee per plate. The addition of reduced glutathione to active S9 further decreased the mutagenic activity and also reduced the mutagenicity together with inactivated S9. The compound or compounds responsible for this inactivation are heat-labile and seem to be located in the cytosol fraction of the S9. Part of the mutagenicity of coffee was also lost spontaneously upon incubation at temperatures between 0 degrees and 50 degrees C. The loss of activity was dependent on temperature, being more pronounced at 50 degrees C compared to 0 degrees C (at 50 degrees C approximately 50% of the mutagenic activity was lost after 6 h). As anaerobic conditions prevented this loss of mutagenicity almost totally, oxidative processes are probably responsible for the inactivation. The stability of the mutagen was not influenced by incubation at low pH values (pH 1-3), with or without the addition of pepsinogen. The mutagenic properties of methylglyoxal, which to some extent could be responsible for the mutagenic activity of coffee, were compared with those of coffee. Methylglyoxal was strongly mutagenic towards Salmonella typhimurium TA100 and TA102. Its mutagenic activity was partially inactivated by the addition of 10% S9. Glyoxalase I and II together with reduced glutathione abolished the mutagenic activity of methylglyoxal but reduced the mutagenicity of coffee by only 80%. Since these enzymes occur in mammalian cells, the mutagenic compound(s) of coffee could also be degraded in vivo. This conclusion is supported by the fact that a long-term carcinogenicity study with rats was negative. These results clearly demonstrate that the effects observed in vitro do not necessarily also occur in vivo, but that in vitro experiments may contribute to the understanding of fundamental mechanisms of chemical carcinogenesis.