Twenty-four genetic polymorphisms in the CYP2D6 gene were analysed in liver DNA samples of 39 Japanese and 44 Caucasians and compared with CYP2D6 protein levels and bufuralol 1'- and 6-hydroxylation activities in liver microsomes of these human samples. We detected 13 types of CYP2D6 genetic polymorphisms and classified these into 20 genotypes; nine types were found in Japanese and 14 types in Caucasian samples. CYP2D6*10B, but not CYP2D6*10A, was the most frequent (34.6%) in Japanese. In Caucasians, several CYP2D6 polymorphisms including CYP2D6*4, *4D, *4E, *4L, *3, *9, *5 and *2E (frequencies of 6.8, 3.4, 4.5, 9.1, 1.1, 2.3, 2.3 and 4.5%, respectively) were detected. A Caucasian having CYP2D6*3/*5 had a protein with slower gel mobility (immunoblotting with anti-CYP2D6 antibody) and very low activity for bufuralol 1'-hydroxylation. Five Caucasian samples (CYP2D6*4/*4, *4/*4L, or *4D/*4L) had no measurable CYP2D6 protein and very low bufuralol 1'-hydroxylation activities. Seven Japanese subjects with CYP2D6*10B/*10B had CYP2D6 protein at levels of approximately 20% of those present in humans with CYP2D6*1 and *2 and catalysed bufuralol 1'-hydroxylation at low rates. Kinetic analysis of bufuralol 1'- and 6-hydroxylation indicates that (i) the Km values for 1'-hydroxylation were lower in individuals with CYP2D6*1/*1, *1/*2, *1/*2X2, and *2/*2 than those with CYP2D6*4/*4, *4/*4L, *4D/*4L, or *10B/*10B and Vmax values tended to be higher in the former groups (*1, *2), and (ii) individuals with heterozygous CYP2D6*1/*4D, *1/*4L, and *1/*5 had relatively high Vmax/Km ratios, whereas individuals with heterozygous CYP2D6*1/*9, *2/4D, *2/*5, *2/*10B, *2E/*4E, *3/*5, *4L/*9, and *10B/*39 had lower Vmax/Km ratios for bufuralol 1'-hydroxylation. Quinidine inhibited bufuralol 1'-hydroxylation in liver microsomes, particularly at low substrate concentrations, in individuals with CYP2D6*1/*1, and 1/1*2, but not those with CYP2D6*4/*4 and very slightly in individuals with CYP2D6*10B/*10B. The latter two groups were found to be more sensitive to alpha-naphthoflavone than the former groups, indicative of the contribution of CYP1A2. These results support the view that CYP2D6*3, *4, *4D, and *4L are major genotypes producing poor metabolizer phenotypes in CYP2D6 in Caucasians, whereas CYP2D6*10B is a major factor in decreased CYP2D6 protein expression and catalytic activities in Japanese.