The human arylamine N-acetyltransferases NAT1 and NAT2 catalyse the acetyl-CoA-dependent N- and O-acetylation of primary arylamine and hydrazine xenobiotics and their N-hydroxylated metabolites. We previously used a panel of recombinant NAT1/NAT2 chimaeric proteins to identify linear amino acid segments that have roles in imparting the distinct catalytic specificities to these proteins [Dupret, Goodfellow, Janezic and Grant (1994) J. Biol. Chem. 269, 26830-26835]. These studies indicated that a conserved central region (residues 112-210) distinct from that containing the active-site cysteine residue Cys(68) was important in determining NAT substrate selectivity. In the present study we have refined our analysis through further chimaera generation of this conserved region and by subsequent site-directed mutagenesis of individual amino acids. Enzyme-kinetic analysis of these mutant proteins with the NAT1-selective and NAT2-selective substrates p-aminosalicylic acid (PAS) and sulphamethazine (SMZ) respectively suggests that residues 125, 127 and 129 are important determinants of NAT1-type and NAT2-type substrate selectivity. Modification of Arg(127) had the greatest effect on specificity for PAS, whereas changing Phe(125) had the greatest effect on specificity for SMZ. Selected NAT mutants exhibited K(m) values for acetyl-CoA that were comparable with those of the wild-type NATs, implying that the mutations affected acceptor substrate specificity rather than cofactor binding affinity. Taken together with previous observations, these results suggest that residues 125, 127 and 129 might contribute to the formation of the active-site pocket surrounding Cys(68) and function as important determinants of NAT substrate selectivity.