作者
Qi Ye,Li Zl,Ying Cui,Hongjia Wei,Y. Victoria Zhang,Fan Zhao,Fuju Wang,Yongbo Song,Weizhuo Xu,Jinghai Zhang
摘要
Cytochrome P450BM3 (CYP102A1) is a versatile biocatalyst capable of selective C–H hydroxylation, yet achieving precise regio- and stereoselectivity toward steroid substrates remains challenging. Here, an integrated computational–experimental framework combining Rosetta-based enzyme design, molecular dynamics (MD) simulations, and QM/MM analysis has been developed to rationally reprogram P450BM3 for site-selective steroid hydroxylation. Using LG23 as the parental enzyme, targeted mutations at key sites were introduced to reshape the substrate channel and modulate active-site polarity, improving catalytic activity and selectivity for six structurally diverse substrates, including testosterone (1), nandrolone (2), androstenedione (3), androsta-1,4-diene-3,17-dione (4), 22-hydroxy-23,24-bisnorchol-4-ene-3-one (5), and progesterone (6). The resulting variants, particularly S88T/Q188L and G87A/S88L/W330L, achieved near-quantitative conversions (>95%) and >90% 7β-selectivity toward compact steroids 1–4. For bulky C17-substituted steroids 5 and 6, the L82T/S88L/W330L variant enhanced conversion from <6% to >80% with strict 15β-selectivity. Under optimized biotransformation conditions (20 °C, pH 8.0) with hydroxypropyl-β-cyclodextrin (HP-β-CD), the titer of 7β-hydroxylated 4 reached 645.4 mg/L, which is among the highest reported titers for this transformation to date. Mechanistic analyses revealed that S88 remotely controls regioselectivity via a structurally conserved water-mediated hydrogen-bond network, confirmed by QM/MM calculations, whereas W330L modulates channel gating to accommodate bulky C17 substituents. This work establishes a potentially generalizable paradigm for the rational engineering of regioselectivity in P450 enzymes, enabling high-yield production of hydroxylated steroids and providing transferable design principles for selective C–H hydroxylation of complex steroidal scaffolds.