Human urine is known to contain substances that strongly inhibit bacterial mutagenicity of aromatic and heterocyclic amines in vitro. The biological relevance of these anti-mutagens was examined by comparing levels of tobacco-related DNA adducts in exfoliated urothelial cells from smokers with the anti-mutagenic activity in corresponding 24-h urine samples. An inverse relationship was found between the inhibition of 2-amino-1-methyl-6-phenylimidazo[4,5-b]pyridine (PhIP)-mutagenicity by urine extracts in vitro and two DNA adduct measurements: the level of the putatively identified N-(deoxyguanosine-8-yl)-4-aminobiphenyl adduct and the total level of all tobacco-smoke-related carcinogen adducts including those probably derived from PhIP. Urinary anti-mutagenicity in vitro appears thus to be a good indicator of the anti-genotoxicity exerted by substances excreted in urine, that protect the bladder mucosal cells (and possibly other cells) against DNA damage. These substances appear to be dietary phenolics and/or their metabolites because (i) the anti-mutagenic activity of urine extracts (n = 18) was linearly related to their content in phenolics; (ii) the concentration ranges of these substances in urine extracts were similar to those of various plant phenols (quercetin, isorhamnetin and naringenin) for which an inhibitory effect on the liver S9-mediated mutagenicity of PhIP was obtained; (iii) treatment of urines with beta-glucuronidase and arylsulfatase enhanced both anti-mutagenicity and the levels of phenolics in urinary extracts; (iv) urinary extracts inhibited noncompetitively the liver S9-mediated mutagenicity of PhIP as did quercetin, used as a model phenolics. Several structural features of the flavonoids were identified as necessary for the inhibition of PhIP and 2-amino-3,8-dimethylimidazo[4,5-f]quinoxiline mutagenicity. Fractionation by reverse-phase HPLC and subsequent analysis of two urinary extracts, showed the presence of several anti-mutagenic substances and phenolics; more lipophilic phenolics displayed the highest specific inhibitory activity. This suggests that enzymatic conversion of dietary flavonoids into their more lipophilic and anti-mutagenic O-methylcatechol derivatives, as noted for quercetin, may occur in vivo in man. Onion, lettuce, apples and red wine are important sources of dietary flavonoids which are probably responsible for the anti-mutagenicity associated with foods and beverages. After HPLC fractionation of urinary extracts, the distribution profile of anti-mutagenic activity corresponded roughly to that of onion and wine extract combined. Our study strongly suggests that smokers ingesting dietary phenolics, probably flavonoids, are partially protected against the harmful effects by tobacco carcinogens within their bladder mucosal cells. This protective effect of dietary phenolics against the cancer of the bladder (and possibly other sites) should be verified and explored as a part of a chemoprevention strategy.