Several hydrolytic and reductive bacterial enzymes (beta-glucuronidase, GN; beta-glucosidase, GS; arylsulphatase, AS; azoreductase, AR; nitroreductase, NR) involved in production of mutagenic or genotoxic metabolites were measured in human colonic contents. Cell-associated AS and extracellular GS were approximately twice as high in the distal colon compared with the proximal bowel, while AR changed little throughout the gut. Measurements of these enzymes in faeces from seven healthy donors confirmed that the majority were cell-associated, and demonstrated high levels of inter-individual variability. NR decreased four-fold between the proximal and distal colon while extracellular GN was reduced by 50%. Most probable number (MPN) analysis on faeces obtained from six healthy donors showed that counts of intestinal bacteria producing GS and AR were c. 10(10) and 10(11)/g, respectively, in all samples tested. Numbers of GN- and AS-forming organisms were between two and three orders of magnitude lower. Inter-individual carriage rates of bacterial populations synthesising NR were highly variable. Screening of 20 pure cultures of intestinal bacteria, belonging to six different genera, showed that Bacteroides ovatus, in particular, synthesised large amounts of GS, whereas B. fragilis, B. vulgatus and Bifidobacterium pseudolongum formed the highest cell-associated levels of GN. In general, bifidobacteria and Lactobacillus acidophilus did not produce significant amounts of AR. All five clostridia studied (Clostridium bifermentans, C. septicum, C. perfringens, C. sporogenes and C. butyricum) produced NR and AR, as did the bacteroides (B. fragilis, B. ovatus and B. vulgatus). Escherichia coli and C. perfringens formed large amounts of NR. Levels of AS production were invariably low and few of the organisms screened synthesised this enzyme. In-vitro studies investigating the effect of intestinal transit time on enzyme production, in a three-stage (V1-V3) continuous culture model of the colon operated at system retention times (R) of either 31.1 or 68.4 h, showed that specific activities of GS were up to four-fold higher (V3) at R = 31.1 h. Bacteriological analysis demonstrated that representative populations of colonic micro-organisms were maintained in the fermentation system, and indicated that changes in GS activity were not related to numbers of the predominant anaerobic or facultative anaerobic species within the model, but were explainable on the basis of substrate-induced modulation of bacterial metabolism.