A minimal CYP102A1 mutant library of only 24 variants plus wild type was constructed by combining five hydrophobic amino acids (alanine, valine, phenylalanine, leucine and isoleucine) in two positions. Both positions are located close to the centre of the haem group. The first, position 87, has been shown to mediate substrate specificity and regioselectivity in CYP102A1. The second hotspot, position 328, was predicted to interact with all substrates during oxidation and has previously been identified by systematic analysis of 31 crystal structures and 6300 sequences of cytochrome P450 monooxygenases. By systematically altering the size of the side chains, a broad range of binding site shapes was generated. All variants were functionally expressed in E. coli. The library was screened with four terpene substrates geranylacetone, nerylacetone, (4R)-limonene and (+)-valencene. Only three variants showed no activity towards all four terpenes, while eleven variants demonstrated either a strong shift or improved regio- or stereoselectivity during oxidation of at least one substrate as compared to CYP102A1 wild type.

CYP153A from Marinobacter aquaeolei has been identified as a fatty acid ω-hydroxylase with a broad substrate range. Two hotspots predicted to influence substrate specificity and selectivity were exchanged. Mutant G307A is 2- to 20-fold more active towards fatty acids than the wild-type. Residue L354 is determinant for the enzyme ω-regioselectivity.

To examine the molecular basis of activity and regioselectivity of the clinically important human microsomal cytochrome P450 (CYP) monooxygenase 2C9 toward its substrate warfarin, 22 molecular dynamics simulations (3-5 ns each) were performed in the presence and absence of warfarin. The resulting trajectories revealed a stable protein core and mobile surface elements. This mobility leads to the formation of two surface channels in the region between F-G loop, B' helix/B-B' loop, beta(1)-sheet, and between helices F and I and the turn in the C-terminal antiparallel beta-sheet in the presence of warfarin. Besides the nonproductive state of the CYP2C9 warfarin complex captured in the crystal structure, three additional states were observed. These states differ in the shape of the substrate binding cavity and the position of the warfarin molecule relative to heme. In one of these states, the 7- and 6-positions of warfarin contact the heme with a marked geometrical preference for position 7 over position 6. This modeling result is consistent with experimentally determined regioselectivity (71 and 22% hydroxylation in positions 7 and 6, respectively). Access to the heme group is limited by the core amino acids Ala297, Leu362, Leu366, and Thr301, which therefore are expected to have a major impact on regioselectivity. In addition, modeling predicts that autoactivation of warfarin is sterically hindered. Our study demonstrates how the combination of mobile surface and rigid core leads to interesting properties: a broad substrate profile and simultaneously a high regioselectivity.

The large and diverse family of cytochrome P450 monooxygenases (CYPs) was systematically analyzed to identify selectivity- and specificity-determining residues in the substrate recognition site 5, which is located in close vicinity to the heme center. A positively charged heme-interacting residue was identified in the structures of 29 monooxygenases and in 97.7% of the 6379 CYP sequences investigated here. This heme-interacting residue restricts the conformation of the substrate recognition site 5 and is preferentially located at position 10 or 11 after the conserved ExxR motif (in 94.4% of the sequences), in 3.3% of the sequences at position 9 or 12. As a result, a classification by the position of the heme-interacting residue allows to predict residues that are closest to the heme center and restrict its accessibility. In 98.4% of all CYP sequences a preferentially hydrophobic residue is located at position 5 after the ExxR motif that is predicted to point close to the heme center. Replacing this residue by hydrophobic residues of different size has been shown to change substrate specificity and regioselectivity for CYPs of different superfamilies. Twenty-seven percent of all CYPs are predicted to contain a second selectivity-determining residue at position 9 after the ExxR motif that can be identified by the pattern EXXR-X(7)-{P}-x-P-[HKR].

A minimal enriched P450 BM3 library was screened for the ability to oxidize inert cyclic and acyclic alkanes. The F87A/A328V mutant was found to effectively hydroxylate cyclooctane, cyclodecane and cyclododecane. F87V/A328F with high activity towards cyclooctane hydroxylated acyclic n-octane to 2-(R)-octanol (46% ee) with high regioselectivity (92%).