Steroidal C7β alcohols and their respective esters have shown significant promise as neuroprotective and anti‐inflammatory agents to treat chronic neuronal damage like stroke, brain trauma, and cerebral ischemia. Since C7 is spatially far away from any functional groups that could direct C−H activation, these transformations are not readily accessible using modern synthetic organic techniques. Reported here are P450‐BM3 mutants that catalyze the oxidative hydroxylation of six different steroids with pronounced C7 regioselectivities and β stereoselectivities, as well as high activities. These challenging transformations were achieved by a focused mutagenesis strategy and application of a novel technology for protein library construction based on DNA assembly and USER (Uracil‐Specific Excision Reagent) cloning. Upscaling reactions enabled the purification of the respective steroidal alcohols in moderate to excellent yields. The high‐resolution X‐ray structure and molecular dynamics simulations of the best mutant unveil the origin of regio‐ and stereoselectivity.

The synthesis of steroid drugs by multistage modifications of the steroidal core is challenging since site-specific and selective modification is essentially required, which is often difficult or complicated for chemocatalysis. For example, the synthesis of Trenbolone (3), a versatile anabolic−androgenic steroid, relies on a four-step chemical procedure on its core modifications of estra-4,9-diene-3,17-dione (1). Here, we have designed a two-step chemoenzymatic strategy that includes a biocatalytic one-pot C11-hydroxylation/17β-ketoreduction of 1 with a computationally designed P450 monooxygenase and an appropriate 17-ketosteroid reductase to generate 11α-OH-9(10)dehydronandrolone (2a) as an intermediate followed by chemical dehydration to introduce the double bond at carbons 11 and 12 with the formation of Trenbolone (3). To obtain a highly active and C11-selective enzyme, molecular dynamics simulations were performed, uncovering a crucial role of water molecules for substrate recognition and targeted hydroxylation of steroids. Moreover, Trenbolone is further subjected to esterification to produce Trenbolone acetate (9) that has been widely used in veterinary medicine. Finally, our approach enables the regio-and stereoselective synthesis of both steroid drugs 3 and 9 on a (nearly) gram scale with 83−91% isolated yields, showing great potential for industrial applications.

No abstract

Many steroids are important pharmaceutically active compounds, while cytochrome P450 monooxygenases (CYPs) are attractive enzymes for applications in steroidal drug synthesis. However, the catalytic efficiency of existing P450s is not routinely high enough, as well as the molecular basis for selectivity control is unclear, which severely restrict their real applications. Here, a 16β steroid-hydroxylase CYP109B4 from Bacillus sonorensis is identified with excellent selectivity and activity. The crystallization and structural analysis of CYP109B4 reveal potential three "hotspot" residues (V84, V292, and S387) responsible for selectivity control. Then, guided by the sequence−function relationships revealed from the mutability landscape construction on the three residues, focused rational iterative site-specific mutagenesis (FRISM) and limited iterative saturation mutagenesis were performed, which provide variant B4-M7 (L240V/S387F/V84L/V292S/I291T/M290F/F294I) with completely switched regioselectivity from 16β to 15β. The subsequent computational analysis uncovers insights into the substrate binding modes in CYP109B4 and its variants, which further confirms the critical role of the "hotspot" residues for selectivity control. Finally, the generality of conserved-"hotspots"mediated selectivity control is demonstrated by performing scaffold sampling between a panel of CYP109B members. Overall, in addition to the present chemical results, our study provides guidance in rationally designing more excellent P450 biocatalysts for potential practical (industrial) applications.