We report mutations of the type II 3 beta-hydroxysteroid dehydrogenase (3 beta HSD) gene in two siblings, male and female, with congenital adrenal hyperplasia caused by classical nonsalt-losing 3 beta HSD deficiency. During childhood, the male sibling, born with ambiguous genitalia, and the female sibling, born with normal genitalia, both manifested symptoms of mild androgen excess; both apparently had normal zona glomerulosa function. Gonadal dynamic study at puberty showed the presence of partial gonadal 3 beta HSD deficiency in both siblings despite their spontaneous pubertal maturation. The 5'-region as well as exons I-II, III, and IV and portions of the adjacent introns of the type II 3 beta HSD gene were amplified by polymerase chain reaction and sequenced. In both siblings and their mother, an identical single nucleotide substitution mutation in intron III, six bases up-stream from exon IV, was identified in one allele. This mutation, G to A at nucleotide 6651, may create a new splicing junction and affect the normal splicing of the messenger ribonucleic acid. In the other allele of both siblings, a missense mutation from GGG (Gly) to AGG (Arg) at codon 129 (G129R) in exon IV was found. We assessed the effect of the G129R missense mutation on enzymatic activity by in vitro analysis of the mutant recombinant enzyme generated by site-directed mutagenesis after its transient expression in COS-1 cells. Using homogenates from transfected cells, the G129R 3 beta HSD enzyme showed a Km value for pregnenolone of 10 +/- 2 mumol/L compared with 1.00 +/- 0.03 mumol/L for the wild-type type II 3 beta HSD enzyme. When dehydroepiandrosterone was used as substrate, the Km value for G129R3 beta HSD was 14 +/- 2 mumol/L compared with 2.1 +/- 0.2 mumol/L for the wild-type II 3 beta HSD enzyme. In addition to an apparent decrease in affinity, the G129R mutation caused a marked decrease in the apparent relative specific activity, thus leading to apparent relative specific efficiencies (relative specific activity/Km) of 2.0% and 4.7% that of the normal type II 3 beta HSD using pregnenolone or dehydroepiandrosterone as substrate, respectively. It appears likely that this low level of activity is sufficient to prevent salt loss, but it is also possible that part of the enzymatic activity comes from the putative remaining percentage of correctly spliced n6651 allele in these patients.