Human male sexual differentiation requires production of fetal testicu
lar testosterone, whose biosynthesis requires steroid 17,20-lyase acti
vity(1,2). Patients with putative isolated 17,20-lyase deficiency have
been reported(3,4). The existence of true isolated 17,20-lyase defici
ency, however, has been questioned because 17 alpha-hydroxylase and 17
,20-lyase activities are catalyzed by a single enzyme(5-8), microsomal
cytochrome P450c17, and because the index case of apparent isolated 1
7,20-lyase deficiency had combined deficiencies of both activities(9,1
0). We studied two patients with clinical and hormonal findings sugges
tive of isolated 17,20-lyase deficiency. We found two patients homozyg
ous for substitution mutations in CYP17, the gene encoding P450c17. Wh
en expressed in COS-1 cells, the mutants retained 17 alpha-hydroxylase
activity but had minimal 17,20-lyase activity. Substrate competition
experiments suggested that the mutations did not alter the enzyme's su
bstrate-binding capacity, but co-transfection of cells with P450 oxido
reductase, the electron donor used by P450c17, indicated that the muta
nts had a diminished ability to interact with redox partners. Computer
-graphic modelling of P450c17 suggests that both mutations lie in or n
ear the redox-partner binding site, on the opposite side of the haem f
rom the substrate-binding pocket. These mutations alter electrostatic
charge distribution in the redox-partner binding site, so that electro
n transfer for the 17,20-lyase reaction is selectively lost or diverte
d to uncoupling reactions. These are the first proven cases of isolate
d 17,20-lyase deficiency, and they demonstrate a novel mechanism for l
oss of enzymatic activity.