Aromatase (CYP19) catalyzes three consecutive hydroxylation reactions converting C19 androgens to aromatic C18 estrogenic steroids. In this study, five human aromatase mutants (E302D, S478A, S478T, H480K, and H480Q) were prepared using a mammalian cell expression system. These mutants were evaluated by enzyme kinetic analysis, inhibitory profile studies, and reaction intermediate measurements. Three steroidal inhibitors [4-hydroxyandrostenedione (4-OHA), 7alpha-(4'-amino)phenylthio-1,4-androstandiene-3,17-dione (7alpha-APTADD), and bridge (2,19-methyleneoxy) androstene-3,17-dione (MDL 101003)], and four nonsteroidal inhibitors [aminoglutethimide (AG), CGS 20267, ICI D1033, and vorozole (R83842)] were used in the inhibitory profile studies. Our computer model of aromatase suggests that Glu302 is situated in the conserved I-helix region and located near the C-19 position of the steroid substrate. The model was supported by significant changes in kinetic parameters and a sevenfold increase in the Ki value of MDL 101,003 for the mutant E302D. As S478A was found to have kinetic properties similar to the wild-type enzyme and a much higher activity than S478T, Ser478 is thought to be situated in a rather restricted environment. There was a 10-fold increase in the Ki value of 7alpha-APTADD for S478T over that for the wild-type enzyme, suggesting that Ser478 might be near the C-7 position of the substrate. The reaction intermediate analysis revealed that significantly more 19-ol intermediate was generated by both S478A and S478T than the wild-type enzyme. These results would support a hypothesis that Ser478 plays a role in the first and second hydroxylation reactions. A positive charged amino acid is preferred at position 480 as shown by the fact that H480K has a significantly higher activity than H480Q. The Ki value of 4-OHA for H480Q was found to be three times that of the wild-type enzyme. In addition, significantly more 19-ol and 19-al intermediates were detected for both mutants H480K and H480Q than for the wild-type enzyme. Evaluation of the two mutations at His480 allows us to propose that this residue may participate in the aromatization reaction (the third step) by acting as a hydrogen bond donor for the C-3 keto group of the substrate. Furthermore, new products were generated when the enzyme was mutated at Ser478 and His480. Thus, these two residues must play an important role in the catalysis and are likely closer to the substrate binding site than previously predicted.