Structural and Functional Evaluation of Steroidogenic Cytochrome P450 Enzymes

Cytochromes P450 (CYP450) are heme-containing monooxygenase enzymes that perform a variety of functions in humans, including xenobiotic metabolism and the production of endogenous signaling molecules. Six isoforms of cytochrome P450 including cytochrome P450 17A1 (CYP17A1) and cytochrome P450 21A2 (CYP21A2), are responsible for the generation of steroid hormones, making these enzymes crucial for development and homeostasis. However in certain pathological conditions, inhibition of steroidogenic cytochrome P450 enzymes can be therapeutically useful. For prostate cancers dependent on androgens such as dihydrotestosterone for tumor growth and proliferation, inhibition of CYP17A1, a required enzyme in androgen biosynthesis, is a promising treatment strategy. Although inhibition of CYP17A1 has been successful clinically, its additional roles in the biosynthesis of other steroid hormones and its similarity to other steroidogenic P450 enzymes can bring about off-target effects. An understanding of the structure and function of CYP17A1, as well as other steroidogenic countertarget enzymes such as CYP21A2, could therefore be fundamental to developing improved inhibitors of this enzyme for the treatment of prostate cancer. CYP17A1 performs two different reactions in the same active site to generate androgens, a hydroxylation reaction followed by a carbon-carbon bond cleavage (or lyase) reaction. While the second, carbon-carbon bond cleavage reaction commits a steroid to the androgen pathway, the initial hydroxylation reaction is also necessary for the production of glucocorticoids. Human CYP17A1 performs the hydroxylation reaction on two structurally similar substrates, Δ4-progesterone and Δ5-pregnenolone, which only exhibit differences on the A ring on the opposite end of the steroid from the carbon which is hydroxylated by CYP17A1. Following hydroxylation, Δ5-17α-hydroxypregnenolone will undergo the second lyase reaction, but the Δ4-17α-hydroxyprogesterone does not. To elucidate a structural basis for preferential lyase turnover of Δ5 steroids by human CYP17A1, the crystal structure of the enzyme containing a background mutation was determined in the presence of both hydroxylase substrates and both lyase substrates. These structures reveal some similarities in binding among all four substrates but also differences in positions relative to the heme iron between hydroxylase substrates, the poor lyase substrate 17α-hydroxyprogesterone, and the efficient lyase substrate 17α-hydroxypregnenolone. Observed differences in distances between 17α-hydroxyprogesterone or 17α-hydroxypregnenolone and the heme iron may reflect differential stabilization of the proposed intermediate for the 17,20-lyase reaction. In addition to substrates, steroidal inhibitors of CYP17A1 can have different configurations of the A ring. X-ray crystal structures of a series of inhibitors with such modifications were also determined with CYP17A1 to compare active site interactions among A-ring modified steroidal inhibitors. Modifications to the A ring of steroidal inhibitor abiraterone did not alter direct contacts with CYP17A1, but did change indirect contacts with the enzyme through active site water networks. The structure of CYP17A1 with one A-ring modified inhibitor was resolved to 2.0 Å, making it the highest resolution crystal structure of CYP17A1 to date, and revealed an additional steroid binding site in the periphery of enzyme that had not been fully appreciated in previous CYP17A1 structures. CYP17A1 and redox partner proteins were recombinantly expressed in E. coli and purified, providing a well-defined and well-controlled system for evaluation of CYP17A1 function and inhibition. This strategy was employed to compare hydroxylase and lyase deficiencies among clinically-reported mutants of the enzyme, as well as steroidal and non-steroidal clinical inhibitors for selective inhibition of the lyase reaction compared to the hydroxylase reaction. Most clinical inhibitors of CYP17A1 demonstrated 1- to 3- fold selectivity for 17,20-lyase inhibition over 17α-hydroxylase inhibition. Only one of the four inhibitors to reach clinical trials, S-orteronel, demonstrated 3- to 5- fold selectivity for 17,20-lyase inhibition compared to inhibition of progesterone and pregnenolone 17α-hydroxylase reactions. However, its enantiomer, R-orteronel demonstrated 8- to 11-fold selectivity. X-ray crystal structures of CYP17A1 with non-steroidal inhibitors reported to selectively inhibit the lyase reaction were also determined to investigate the structural basis for lyase selectivity. Finally, some CYP17A1 inhibitors have also been shown to interact with another P450 responsible for steroid biosynthesis, CYP21A2. The physiological consequences of off-target CYP21A2 inhibition by compounds developed to target CYP17A1 are complex. The crystal structure of human CYP21A2 was determined to provide a structural comparison to CYP17A1 and potentially aid in the design of inhibitors more selective for CYP17A1 over CYP21A2. Some structural features of the CYP17A1 active site, including a hydrophobic pocket over the I helix, are not conserved in CYP21A2. Exploitation of this pocket is a potential strategy for the development of inhibitors with reduced affinity for CYP21A2. In aggregate, the studies described herein use structural information coupled with functional analysis to better understand steroidogenic cytochromes P450. These enzymes act as targets and countertargets in the treatment of hormone dependent diseases including prostate cancer. More detailed knowledge of how these enzymes interact with both substrates and inhibitors could inform the development of better prostate cancer therapeutics