A novel regio-specific cyclosporin hydroxylase gene revealed through the genome mining of <Emphasis Type="Italic">Pseudonocardia autotrophica</Emphasis>

The regio-specific hydroxylation at the 4th N-methyl leucine of the immunosuppressive agent cyclosporin A (CsA) was previously proposed to be mediated by a unique cytochrome P450 hydroxylase (CYP), CYP-sb21 from the rare actinomycetes Sebekia benihana. Interestingly, a different rare actinomycetes species, Pseudonocardia autotrophica, was found to possess a different regio-selectivity, the preferential hydroxylation at the 9th N-methyl leucine of CsA. Through an in silico analysis of the whole genome of P. autotrophica, we describe here the classification of 31 total CYPs in P. autotrophica. Three putative CsA CYP genes, showing the highest sequence homologies with CYP-sb21, were successfully inactivated using PCR-targeted gene disruption. Only one knock-out mutant, ΔCYP-pa1, failed to convert CsA to its hydroxylated forms. The hydroxylation activity of CsA by CYP-pa1 was confirmed by CYP-pa1 gene complementation as well as heterologous expression in the CsA non-hydroxylating Streptomyces coelicolor. Moreover, the cyclosporine regio-selectivity of CYP-pa1 expressed in the ?CYP-sb21 S. benihana mutant strain was also confirmed unchanged through cross complementation. These results show that preferential regio-specific hydroxylation at the 9th N-methyl leucine of CsA is carried out by a specific P450 hydroxylase gene in P. autotrophica, CYP-pa1, setting the stage for the biotechnological application of CsA regio-selective hydroxylation.     

In vitro reconstitution of the cyclosporine specific P450 hydroxylases using heterologous redox partner proteins

The cytochrome P450 enzymes (CYPs) CYP-sb21 from Sebekia benihana and CYP-pa1 from Pseudonocardia autotrophica are able to hydroxylate the immunosuppressant cyclosporin A (CsA) in a regioselective manner, giving rise to the production of two hair-stimulating agents (with dramatically attenuated immunosuppressant activity), γ-hydroxy-N-methyl-l-Leu4-CsA (CsA-4-OH) and γ-hydroxy-N-methyl-l-Leu9-CsA (CsA-9-OH). Recently, the in vitro activity of CYP-sb21 was identified using several surrogate redox partner proteins. Herein, we reconstituted the in vitro activity of CYP-pa1 for the first time via a similar strategy. Moreover, the supporting activities of a set of ferredoxin (Fdx)/ferredoxin reductase (FdR) pairs from the cyanobacterium Synechococcus elongatus PCC 7942 were comparatively analyzed to identify the optimal redox systems for these two CsA hydroxylases. The results suggest the great value of cyanobacterial redox partner proteins for both academic research and industrial application of P450 biocatalysts.     

Identification of a vitamin D3-specific hydroxylase genes through actinomycetes genome mining

We previously completed whole-genome sequencing of a rare actinomycete named Sebekia benihana, and identified the complete S. benihana cytochrome P450 complement (CYPome), including 21 cytochrome P450 hydroxylase (CYP), seven ferredoxin (FD), and four ferredoxin reductase (FDR) genes. Through targeted CYPome disruption, a total of 32 S. benihana CYPome mutants were obtained. Subsequently, a novel cyclosporine A region-specific hydroxylase was successfully determined to be encoded by a CYP-sb21 gene by screening the S. benihana CYPome mutants. Here, we report that S. benihana is also able to mediate vitamin D₃ (VD₃) hydroxylation. Among the 32 S. benihana CYPome mutants tested, only a single S. benihana CYP mutant, ΔCYP-sb3a, failed to show regio-specific hydroxylation of VD₃ to 25-hydroxyvitamin D₃ and 1α,25-dihydroxyvitamin D₃. Moreover, the VD₃ hydroxylation activity in the ΔCYP-sb3a mutant was restored by CYP-sb3a gene complementation. Since all S. benihana FD and FDR disruption mutants maintained VD₃ hydroxylation activity, we conclude that CYP-sb3a, a member of the bacterial CYP107 family, is the only essential component of the in vivo regio-specific VD₃ hydroxylation process in S. benihana. Expression of the CYP-sb3a gene exhibited VD₃ hydroxylation in the VD₃ non-hydroxylating Streptomyces coelicolor, implying that the regio-specific hydroxylation of VD₃ is carried out by a specific P450 hydroxylase in S. benihana.     

Productive and non-productive complexes in cytochrome P450-containing systems

The equilibrium dissociation constants K_D, the complex association / dissociation rate constants (k _on /k _off) and lifetimes of the complexes of redox partners were measured for three cytochrome P450-containing monooxygenase systems (P450cam, P450scc, and P450 2B4) under hydroxylation conditions. The Q parameter representing the ratio of protein-protein complex lifetime (τ _lT ) to time required for a single hydroxylation cycle (τ_turnover) was introduced for estimation of productivity of complexes formed within the systems studied. The Q parameter was insignificantly changed upon transition from the oxidation to hydroxylation conditions. Lifetimes (τ _lT ) for the binary complexes formed within the P450cam and the P450scc systems obligatory requiring an intermediate electron transfer protein between the reductase and cytochrome P450 could not realize hydroxylation reactions for substrates with known τ_turnover and so they were non-productive while the binary complexes formed within the P450 2B4 system, not requiring such intermediate electron-transfer protein, appeared to be productive. Formation of ternary complexes was demonstrated under hydroxylation conditions in all three systems. Analysis of Q values led to the conclusion that the ternary complexes formed within the P450cam and the P450scc systems were productive. In the case of the P450 2B4 system, more than half (about 60%) ternary complexes were also found to be productive.     

Insights into the functional properties of the marneral oxidase CYP71A16 from Arabidopsis thaliana

The Arabidopsis thaliana gene encoding CYP71A16 is part of the gene cluster for the biosynthesis and modification of the triterpenoid marneral. Previous investigations of A. thaliana have revealed that CYP71A16 catalyzes marneral oxidation, while it also can accept marnerol as substrate. The aim of the present study was to investigate functional properties of CYP71A16 in vitro. For this purpose, heterologous expression of a N-terminally modified version of CYP71A16 was established in Escherichia coli, which yielded up to 50 mg L^− 1 recombinant enzyme. The enzyme was purified and activity was reconstituted in vitro with different redox partners. A heterologous bacterial redox partner system consisting of the flavodoxin YkuN from Bacillus subtilis and the flavodoxin reductase Fpr from E. coli clearly outperformed the cytochrome P450 reductase ATR2 from A. thaliana in supporting the CYP71A16-mediated hydroxylation of marnerol. Substrate binding experiments with CYP71A16 revealed a dissociation constant K_D of 225 μM for marnerol. CYP71A16 catalyzed the hydroxylation of marnerol to 23-hydroxymarnerol with a K_M of 142 μM and a k_cat of 3.9 min^− 1. Furthermore, GC/MS analysis revealed an as of yet unidentified overoxidation product of this in vitro reaction. This article is part of a Special Issue entitled: Cytochrome P450 biodiversity and biotechnology, edited by Erika Plettner, Gianfranco Gilardi, Luet Wong, Vlada Urlacher, Jared Goldstone.     

Hydroxylation of daidzein by CYP107H1 from Bacillus subtilis 168

CYP107H1, from Bacillus subtilis 168 known as fatty acid hydroxylase, showed the ortho-specific hydroxylation activity to daidzein, when coupled to the putidaredoxin reductase (camA) and putidaredoxin (camB) from Pseudomonas putida as the redox partners. The electron transfer system of the three proteins was constructed in Escherichia coli BL21 (DE3) system using the two plasmids containing different selection markers. The daidzein hydroxylation was demonstrated with recombinant whole cell and in vitro system using the artificial redox partner for electron transfer. The identification of the hydroxylation reaction yielding 7,3′,4′-trihydroxyisoflavone was elucidated using gas chromatography mass spectrometry (GC–MS). This oxidizing activity of CYP107H1 towards daidzein represents the new hydroxylation of aromatic compound as substrate.     

Whole-cell biotransformation with recombinant cytochrome P450 for the selective oxidation of Grundmann’s ketone

25-Hydroxy-Grundmann’s ketone is a key building block in the chemical synthesis of vitamin D₃ and its derivatives through convergent routes. Generally, the chemical synthesis of this compound involves tedious procedures and results in a mixture of several products. Recently, the selective hydroxylation of Grundmann’s ketone at position C25 by cytochrome P450 (CYP) 154E1 from Thermobifida fusca YX was described. In this study a recombinant whole-cell biocatalyst was developed and applied for hydroxylation of Grundmann’s ketone. Biotransformation was performed by Escherichia coli cells expressing CYP154E1 along with two redox partner systems, Pdx/PdR and YkuN/FdR. The system comprising CYP154E1/Pdx/PdR showed the highest production of 25-hydroxy-Grundmann’s ketone and resulted in 1.1 mM (300 mg L⁻¹) product concentration.     

Comparative functional characterization of a novel benzoate hydroxylase cytochrome P450 of Fusarium oxysporum

FoCYP53A19, a novel cytochrome P450 capable of performing benzoate hydroxylation, was identified and characterized from the ascomycete Fusarium oxysporum f.sp. lycopersici. Comparative functional analysis of FoCYP53A19 with the heterologous and homologous cytochrome P450 reductases (CPR) such as Saccharomyces cerevisiae (ScCPR), Candida albicans (CaCPR) and F. oxysporum (FoCPR) revealed novel catalytic properties. The catalytic efficiency and substrate specificity of FoCYP53A19 were significantly influenced and altered by the source of the reductase employed. The yeast reconstitution system of FoCYP53A19 with ScCPR performed the hydroxylation of benzoic acid (BA) and demethylation of 3-methoxybenzoic acid (3-MBA); but when reconstituted with CaCPR, FoCYP53A19 performed only the essential hydroxylation of fungal benzoate catabolism. Remarkably, FoCYP53A19 with its homologous reductase FoCPR, not only demonstrated the improved conversion rates of BA and 3-MBA, but also exhibited activity toward the hydroxylation of 3-hydroxybenzoic acid. The electron transfer compatibility and the coupling efficiency between the homologous FoCYP-FoCPR system are significant and it favored enhanced monooxygenase activity with broader substrate specificity.     

Bacterial Cytochrome P450 System Catabolizing the Fusarium Toxin Deoxynivalenol

Deoxynivalenol (DON) is a natural toxin of fungi that cause Fusarium head blight disease of wheat and other small-grain cereals. DON accumulates in infected grains and promotes the spread of the infection on wheat, posing serious problems to grain production. The elucidation of DON-catabolic genes and enzymes in DON-degrading microbes will provide new approaches to decrease DON contamination. Here, we report a cytochrome P450 system capable of catabolizing DON in Sphingomonas sp. strain KSM1, a DON-utilizing bacterium newly isolated from lake water. The P450 gene ddnA was cloned through an activity-based screening of a KSM1 genomic library. The genes of its redox partner candidates (flavin adenine dinucleotide [FAD]-dependent ferredoxin reductase and mitochondrial-type [2Fe-2S] ferredoxin) were not found adjacent to ddnA; the redox partner candidates were further cloned separately based on conserved motifs. The DON-catabolic activity was reconstituted in vitro in an electron transfer chain comprising the three enzymes and NADH, with a catalytic efficiency (k_cat/K_m) of 6.4 mM⁻¹ s⁻¹. The reaction product was identified as 16-hydroxy-deoxynivalenol. A bioassay using wheat seedlings revealed that the hydroxylation dramatically reduced the toxicity of DON to wheat. The enzyme system showed similar catalytic efficiencies toward nivalenol and 3-acetyl deoxynivalenol, toxins that frequently cooccur with DON. These findings identify an enzyme system that catabolizes DON, leading to reduced phytotoxicity to wheat.     

Efficient hydroxylation of 1,8-cineole with monoterpenoid-resistant recombinant <Emphasis Type="Italic">Pseudomonas putida</Emphasis> GS1

In this work, monoterpenoid hydroxylation with Pseudomonas putida GS1 and KT2440 were investigated as host strains, and the cytochrome P450 monooxygenase CYP176A1 (P450_cin) and its native redox partner cindoxin (CinC) from Citrobacter braakii were introduced in P. putida to catalyze the stereoselective hydroxylation of 1,8-cineole to (1R)-6β-hydroxy-1,8-cineole. Growth experiments in the presence of 1,8-cineole confirmed pseudomonads’ superior resilience compared to E. coli. Whole-cell P. putida harboring P450_cin with and without CinC were capable of hydroxylating 1,8-cineole, whereas coexpression of CinC has been shown to accelerate this bioconversion. Under the same conditions, P. putida GS1 produced more than twice the amount of heterologous P450_cin and bioconversion product than P. putida KT2440. A concentration of 1.1 ± 0.1 g/L (1R)-6β-hydroxy-1,8-cineole was obtained within 55 h in shake flasks and 13.3 ± 1.9 g/L in 89 h in a bioreactor, the latter of which corresponds to a yield Y_P/S of 79 %. To the authors’ knowledge, this is the highest product titer for a P450 based whole-cell monoterpene oxyfunctionalization reported so far. These results show that solvent-tolerant P. putida GS1 can be used as a highly efficient recombinant whole-cell biocatalyst for a P450 monooxygenase-based valorization of monoterpenoids.     

A recombinant Escherichia coli whole cell biocatalyst harboring a cytochrome P450cam monooxygenase system coupled with enzymatic cofactor regeneration

A cytochrome P450cam monooxygenase (P450cam) system from the soil bacterium Pseudomonas putida requires electron transfer among three different proteins and a cofactor, nicotinamide adenine dinucleotide (NADH), for oxygenation of its natural substrate, camphor. Herein, we report a facile way to significantly enhance the catalytic efficiency of the P450cam system by the coupling of its native electron transfer system with enzymatic NADH regeneration catalyzed by glycerol dehydrogenase (GLD) in Escherichia coli whole cell biocatalysts. Recombinant E. coli harboring the P450cam system, but lacking GLD, exhibited little activity for camphor hydroxylation. In contrast, coexpression of GLD with the proteinaceous electron transfer components of P450cam resulted in about tenfold improvement in the substrate conversion, implying that the whole cell biocatalyst utilized molecular oxygen, endogenous NADH, and glycerol in the cell for catalysis. The addition of glycerol to the reaction media further promoted camphor hydroxylation, suggesting that exogenous glycerol is also available for GLD in the host cell and actively participates in the catalytic cycle. These results clearly show the utility of GLD towards functional reconstruction of the native P450cam system. The present approach may also be useful for E. coli whole cell biocatalysts with the other NADH-dependent oxygenases and oxidoreductases.     

P450BM3 fused to phosphite dehydrogenase allows phosphite-driven selective oxidations

To facilitate the wider application of the NADPH-dependent P450_BM3, we fused the monooxygenase with a phosphite dehydrogenase (PTDH). The resulting monooxygenase-dehydrogenase fusion enzyme acts as a self-sufficient bifunctional catalyst, accepting phosphite as a cheap electron donor for the regeneration of NADPH.

The well-expressed fusion enzyme was purified and analyzed in comparison to the parent enzymes. Using lauric acid as substrate for P450_BM3, it was found that the fusion enzyme had similar substrate affinity and hydroxylation selectivity while it displayed a significantly higher activity than the non-fused monooxygenase. Phosphite-driven conversions of lauric acid at restricted NADPH concentrations confirmed multiple turnovers of the cofactor. Interestingly, both the fusion enzyme and the native P450_BM3 displayed enzyme concentration dependent activity and the fused enzyme reached optimal activity at a lower enzyme concentration. This suggests that the fusion enzyme has an improved tendency to form functional oligomers.

To explore the constructed phosphite-driven P450_BM3 as a biocatalyst, conversions of the drug compounds omeprazole and rosiglitazone were performed. PTDH-P450_BM3 driven by phosphite was found to be more efficient in terms of total turnover when compared with P450_BM3 driven by NADPH. The results suggest that PTDH-P450_BM3 is an attractive system for use in biocatalytic and drug metabolism studies.

     

Engineering cytochrome P450 monooxygenase CYP 116B3 for high dealkylation activity

Cytochrome P450 monooxygenase CYP116B3 from Rhodococcus ruber catalyzes the dealkylation of 7-ethoxycoumarin and the hydroxylation of substituted and unsubstituted aromatics. However, since activities were quite low, a combination of site-specific mutagenesis and directed evolution was applied to produce 7800 variants of CYP116B3, which were screened via a newly developed high-throughput screening system based on the dealkylation of 7-ethoxycoumarin catalyzed by recombinant E. coli. The best mutant was found after four rounds of directed evolution and had a 240-fold increased deethylation activity toward 7-ethoxycoumarin (223 nmol product/nmol P450.min) and a 10-fold increased demethylation activity toward 7-methoxycoumarin (9 nmol product/nmol P450.min).     

Synthesis of ω-hydroxy dodecanoic acid based on an engineered CYP153A fusion construct

A bacterial P450 monooxygenase-based whole cell biocatalyst using Escherichia coli has been applied in the production of ω-hydroxy dodecanoic acid from dodecanoic acid (C12-FA) or the corresponding methyl ester. We have constructed and purified a chimeric protein where the fusion of the monooxygenase CYP153A from Marinobacter aquaeloei to the reductase domain of P450 BM3 from Bacillus megaterium ensures optimal protein expression and efficient electron coupling. The chimera was demonstrated to be functional and three times more efficient than other sets of redox components evaluated. The established fusion protein (CYP153A_{M. aq.}-CPR) was used for the hydroxylation of C12-FA in in vivo studies. These experiments yielded 1.2 g l^–1 ω-hydroxy dodecanoic from 10 g l^–1 C12-FA with high regioselectivity (> 95%) for the terminal position. As a second strategy, we utilized C12-FA methyl ester as substrate in a two-phase system (5:1 aqueous/organic phase) configuration to overcome low substrate solubility and product toxicity by continuous extraction. The biocatalytic system was further improved with the coexpression of an additional outer membrane transport system (AlkL) to increase the substrate transfer into the cell, resulting in the production of 4 g l^–1 ω-hydroxy dodecanoic acid. We further summarized the most important aspects of the whole-cell process and thereupon discuss the limits of the applied oxygenation reactions referring to hydrogen peroxide, acetate and P450 concentrations that impact the efficiency of the production host negatively.     

Involvement of the Cytochrome P450 System EthBAD in the N-Deethoxymethylation of Acetochlor by Rhodococcus sp. Strain T3-1

Acetochlor [2-chloro-N-(ethoxymethyl)-N-(2-ethyl-6-methylphenyl)-acetamide] is a widely applied herbicide with potential carcinogenic properties. N-Deethoxymethylation is the key step in acetochlor biodegradation. N-Deethoxymethylase is a multicomponent enzyme that catalyzes the conversion of acetochlor to 2′-methyl-6′-ethyl-2-chloroacetanilide (CMEPA). Fast detection of CMEPA by a two-enzyme (N-deethoxymethylase–amide hydrolase) system was established in this research. Based on the fast detection method, a three-component enzyme was purified from Rhodococcus sp. strain T3-1 using ammonium sulfate precipitation and hydrophobic interaction chromatography. The molecular masses of the components of the purified enzyme were estimated to be 45, 43, and 11 kDa by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). Based on the results of peptide mass fingerprint analysis, acetochlor N-deethoxymethylase was identified as a cytochrome P450 system, composed of a cytochrome P450 oxygenase (43-kDa component; EthB), a ferredoxin (45 kDa; EthA), and a reductase (11 kDa; EthD), that is involved in the degradation of methyl tert-butyl ether. The gene cluster ethABCD was cloned by PCR amplification and expressed in Escherichia coli BL21(DE3). Resting cells of a recombinant E. coli strain showed deethoxymethylation activity against acetochlor. Subcloning of ethABCD showed that ethABD expressed in E. coli BL21(DE3) has the activity of acetochlor N-deethoxymethylase and is capable of converting acetochlor to CMEPA.     

Cloning, expression, and characterization of a self-sufficient cytochrome P450 monooxygenase from Rhodococcus ruber DSM 44319

A new member of class IV of cytochrome P450 monooxygenases was identified in Rhodococcus ruber strain DSM 44319. As the genome of R. ruber has not been sequenced, a P450-like gene fragment was amplified using degenerated primers. The flanking regions of the P450-like DNA fragment were identified by directional genome walking using polymerase chain reaction. The primary protein structure suggests a natural self-sufficient fusion protein consisting of ferredoxin, flavin-containing reductase, and P450 monooxygenase. The only flavin found within the enzyme was riboflavin 5′-monophosphate. The enzyme was successfully expressed in Escherichia coli, purified and characterized. In the presence of NADPH, the P450 monooxygenase showed hydroxylation activity towards polycyclic aromatic hydrocarbons naphthalene, indene, acenaphthene, toluene, fluorene, m-xylene, and ethyl benzene. The conversion of naphthalene, acenaphthene, and fluorene resulted in respective ring monohydroxylated metabolites. Alkyl aromatics like toluene, m-xylene, and ethyl benzene were hydroxylated exclusively at the side chains. The new enzyme’s ability to oxidize such compounds makes it a potential candidate for biodegradation of pollutants and an attractive biocatalyst for synthesis.     

Cytochrome P450–redox partner fusion enzymes

The cytochromes P450 (P450s) are a broad class of heme b-containing mono-oxygenase enzymes. The vast majority of P450s catalyse reductive scission of molecular oxygen using electrons usually derived from coenzymes (NADH and NADPH) and delivered from redox partner proteins. Evolutionary advantages may be gained by fusion of one or more redox partners to the P450 enzyme in terms of e.g. catalytic efficiency. This route was taken by the well characterized flavocytochrome P450_BM3 system (CYP102A1) from Bacillus megaterium, in which soluble P450 and cytochrome P450 reductase enzymes are covalently linked to produce a highly efficient electron transport system for oxygenation of fatty acids and related molecules. However, genome analysis and ongoing enzyme characterization has revealed that there are a number of other novel classes of P450–redox partner fusion enzymes distributed widely in prokaryotes and eukaryotes. This review examines our current state of knowledge of the diversity of these fusion proteins and explores their structural composition and evolutionary origins.     

Amidohydrolase Process: Expanding the use of l-N-carbamoylase/N-succinyl-amino acid racemase tandem for the production of different optically pure l-amino acids

A bienzymatic system comprising an N-succinylamino acid racemase from Geobacillus kaustophilus CECT4264 (GkNSAAR) and an enantiospecific l-N-carbamoylase from Geobacillus stearothermophilus CECT43 (BsLcar) has been developed. This biocatalyst has been able to produce optically pure natural and non-natural l-amino acids starting from racemic mixtures of N-acetyl-, N-formyl- and N-carbamoyl-amino acids by dynamic kinetic resolution. The fastest conversion rate was found with N-formyl-amino acids, followed by N-carbamoyl- and N-acetyl-amino acids, and GkNSAAR proved to be the limiting step of the system due to its lower specific activity. Metal ion cobalt was essential for the activity of the biocatalyst and the system was optimally active when Co²⁺ was added directly to the reaction mixture. The optimum pH for the biocatalyst proved to be 8.0, for both N-formyl- and N-carbamoyl-amino acid substrates, whereas optimum temperature ranges were 45–55 °C for N-formyl-amino acids and 55–70 °C for N-carbamoyl-derivatives. The bienzymatic system was equally efficient in converting aromatic and aliphatic substrates. Total conversion was also achieved using high substrate concentrations (100 and 500 mM) with no noticeable inhibition. This “Amidohydrolase Process” enables the production of both natural and non-natural l-amino acids from a broad substrate spectrum with yields of over 95%.     

Engineering of recombinant E. coli cells co‐expressing P450pyrTM monooxygenase and glucose dehydrogenase for highly regio‐ and stereoselective hydroxylation of alicycles with cofactor recycling

E. coli (P450pyrTM‐GDH) with dual plasmids, pETDuet containing P450pyr triple mutant I83H/M305Q/A77S (P450pyrTM) and ferredoxin reductase (FdR) genes and pRSFDuet containing glucose dehydrogenase (GDH) and ferredoxin (Fdx) genes, was engineered to show a high activity (12.7 U g^?1 cdw) for the biohydroxylation of N‐benzylpyrrolidine 1 and a GDH activity of 106 U g^?1 protein. The E. coli cells were used as efficient biocatalysts for highly regio‐ and stereoselective hydroxylation of alicyclic substrates at non‐activated carbon atom with enhanced productivity via intracellular recycling of NAD(P)H. Hydroxylation of N‐benzylpyrrolidine 1 with resting cells in the presence of glucose showed excellent regio‐ and stereoselectivity, giving (S)‐N‐benzyl‐3‐hydroxypyrrolidine 2 in 98% ee as the sole product in 9.8 mM. The productivity is much higher than that of the same biohydroxylation using E. coli (P450pyrTM)b without expressing GDH. E. coli (P450pyrTM‐GDH) was found to be highly regio‐ and stereoselective for the hydroxylation of N‐benzylpyrrolidin‐2‐one 3 , improving the regioselectivity from 90% of the wild‐type P450pyr to 100% and giving (S)‐N‐benzyl‐4‐hydroxylpyrrolidin‐2‐one 4 in 99% ee as the sole product. A high activity of 15.5 U g^?1 cdw was achieved and (S)‐ 4 was obtained in 19.4 mM. E. coli (P450pyrTM‐GDH) was also found to be highly regio‐ and stereoselective for the hydroxylation of N‐benzylpiperidin‐2‐one 5 , increasing the ee of the product (S)‐N‐benzyl‐4‐hydroxy‐piperidin‐2‐one 6 to 94% from 33% of the wild‐type P450pyr. A high activity of 15.8 U g^?1 cdw was obtained and (S)‐ 6 was produced in 3.3 mM as the sole product. E. coli (P450pyrTM‐GDH) represents the most productive system known thus far for P450‐catalyzed hydroxylations with cofactor recycling, and the hydroxylations with E. coli (P450pyrTM‐GDH) provide with simple and useful syntheses of (S)‐ 2 , (S)‐ 4 , and (S)‐ 6 that are valuable pharmaceutical intermediates and difficult to prepare. Biotechnol. Bioeng. 2013; 110: 363–373. © 2012 Wiley Periodicals, Inc.