Molecular evolution of a steroid hydroxylating cytochrome P450 using a versatile steroid detection system for screening

CYP154C5 Regioselectivity in Steroid Hydroxylation Explored by Substrate Modifications and Protein Engineering*

CYP154C5 from Nocardia farcinica is a P450 monooxygenase able to hydroxylate a range of steroids with high regio- and stereoselectivity at the 16α-position. Using protein engineering and substrate modifications based on the crystal structure of CYP154C5, an altered regioselectivity of the enzyme in steroid hydroxylation had been achieved. Thus, conversion of progesterone by mutant CYP154C5 F92A resulted in formation of the corresponding 21-hydroxylated product 11-deoxycorticosterone in addition to 16α-hydroxylation. Using MD simulation, this altered regioselectivity appeared to result from an alternative binding mode of the steroid in the active site of mutant F92A. MD simulation further suggested that the entrance of water to the active site caused higher uncoupling in this mutant. Moreover, exclusive 15α-hydroxylation was observed for wild-type CYP154C5 in the conversion of 5α-androstan-3-one, lacking an oxy-functional group at C17. Overall, our data give valuable insight into the structure-function relationship of this cytochrome P450 monooxygenase for steroid hydroxylation.

Bacterial cytochrome P450-catalyzed regio- and stereoselective steroid hydroxylation enabled by directed evolution and rational design

Steroids are the most widely marketed products by the pharmaceutical industry after antibiotics. Steroid hydroxylation is one of the most important functionalizations because their derivatives enable a higher biological activity compared to their less polar non-hydroxylated analogs. Bacterial cytochrome P450s constitute promising biocatalysts for steroid hydroxylation due to their high expression level in common workhorses like Escherichia coli. However, they often suffer from wrong or insufficient regio- and/or stereoselectivity, low activity, narrow substrate range as well as insufficient thermostability, which hampers their industrial application. Fortunately, these problems can be generally solved by protein engineering based on directed evolution and rational design. In this work, an overview of recent developments on the engineering of bacterial cytochrome P450s for steroid hydroxylation is presented.

Directed Evolution of P450 BM3 towards Functionalization of Aromatic O-Heterocycles

The O-heterocycles, benzo-1,4-dioxane, phthalan, isochroman, 2,3-dihydrobenzofuran, benzofuran, and dibenzofuran are important building blocks with considerable medical application for the production of pharmaceuticals. Cytochrome P450 monooxygenase (P450) Bacillus megaterium 3 (BM3) wild type (WT) from Bacillus megaterium has low to no conversion of the six O-heterocycles. Screening of in-house libraries for active variants yielded P450 BM3 CM1 (R255P/P329H), which was subjected to directed evolution and site saturation mutagenesis of four positions. The latter led to the identification of position R255, which when introduced in the P450 BM3 WT, outperformed all other variants. The initial oxidation rate of nicotinamide adenine dinucleotide phosphate (NADPH) consumption increased ≈140-fold (WT: 8.3 ± 1.3 min-1; R255L: 1168 ± 163 min-1), total turnover number (TTN) increased ≈21-fold (WT: 40 ± 3; R255L: 860 ± 15), and coupling efficiency, ≈2.9-fold (WT: 8.8 ± 0.1%; R255L: 25.7 ± 1.0%). Computational analysis showed that substitution R255L (distant from the heme-cofactor) does not have the salt bridge formed with D217 in WT, which introduces flexibility into the I-helix and leads to a heme rearrangement allowing for efficient hydroxylation.

Binding modes of CYP106A2 redox partners determine differences in progesterone hydroxylation product patterns

Natural redox partners of bacterial cytochrome P450s (P450s) are mostly unknown. Therefore, substrate conversions are performed with heterologous redox partners; in the case of CYP106A2 from Bacillus megaterium ATCC 13368, bovine adrenodoxin (Adx) and adrenodoxin reductase (AdR). Our aim was to optimize the redox system for CYP106A2 for improved product formation by testing 11 different combinations of redox partners. We found that electron transfer protein 1(516–618) showed the highest yield of the main product, 15β-hydroxyprogesterone, and, furthermore, produced a reduced amount of unwanted polyhydroxylated side products. Molecular protein–protein docking indicated that this is caused by subtle structural changes leading to alternative binding modes of both redox enzymes. Stopped-flow measurements analyzing the CYP106A2 reduction and showing substantial differences in the apparent rate constants supported this conclusion. The study provides for the first time to our knowledge rational explanations for differences in product patterns of a cytochrome P450 caused by difference in the binding mode of the redox partners.

Tanja Sagadin et al. show that different redox systems can be used to tune the rate selectivity and yield of progesterone conversion by the cytochrome P450 CYP106A2. They screen 11 redox partner combinations and identify specific combinations that may be used to improve biotechnological production of mono- and polyhydroxylated products.

Characterization of CYP154F1 from Thermobifida fusca YX and Extension of Its Substrate Spectrum by Site-Directed Mutagenesis

Previous studies on cytochrome P450 monooxygenases (CYP) from family 154 reported their substrate promiscuity and high activity. Hence, herein, the uncharacterized family member CYP154F1 is described. Screening of more than 100 organic compounds revealed that CYP154F1 preferably accepts small linear molecules with a carbon chain length of 8-10 atoms. In contrast to thoroughly characterized CYP154E1, CYP154F1 has a much narrower substrate spectrum and lower activity. A structural alignment of homology models of CYP154F1 and CYP154E1 revealed few differences in the active sites of both family members. By gradual mutagenesis of the CYP154F1 active site towards those of CYP154E1, a key residue accounting for the different activities of both enzymes was identified at position 234. Substitution of T234 for large hydrophobic amino acids led to up to tenfold higher conversion rates of small substrates, such as geraniol. Replacement of T234 by small hydrophobic amino acids, valine or alanine, resulted in mutants with extended substrate spectra. These mutants are able to convert some of the larger substrates of CYP154E1, such as (E)-stilbene and (+)-nootkatone.

CYP106A2-A versatile biocatalyst with high potential for biotechnological production of selectively hydroxylated steroid and terpenoid compounds

CYP106A2 from Bacillus megaterium ATCC13368, was identified in the 1970s as one of the first bacterial steroid hydroxylases responsible for the conversion of progesterone to 15β-hydroxyprogesterone. Later on it has been proven to be a potent hydroxylase of numerous 3-oxo-Δ⁴ as well as 3-hydroxy-Δ⁵-steroids and has recently also been characterized as a regioselective allylic bacterial diterpene hydroxylase. The main hydroxylation position of CYP106A2 is thought to be influenced by the functional groups at C3 position in the steroid core leading to a favored 15β-hydroxylation of 3-oxo-Δ⁴-steroids and 7β-hydroxylation of 3-hydroxy-Δ⁵-steroids. However, in some cases the hydroxylation is not strictly selective, resulting in the formation of undesired side-products. To overcome the unspecific hydroxylations or, on the contrary, to gain more of these products in case they are of industrial interest, rational protein design and directed evolution have been successfully performed to shift the stereoselectivity of hydroxylation by CYP106A2. The subsequently obtained hydroxylated steroid and terpene derivatives are especially useful as drug metabolites and drug precursors for the pharmaceutical industry, due to their diverse biological properties and hardship of their chemical synthesis. As a soluble prokaryotic P450 with broad substrate spectrum and hydroxylating capacity, CYP106A2 is an outstanding candidate to establish bioconversion processes. It has been expressed with respectable yields in Escherichia coli and Bacillus megaterium and was applied for the preparative hydroxylation of several steroids and terpenes. Recently, the application of the enzyme was assessed under process conditions as well, depicting a successfully optimized process development and getting us closer to industrial scale process requirements and a future large scale application. 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.

Engineering of CYP106A2 for steroid 9α- and 6β-hydroxylation

CYP 106A2 from Bacillus megaterium ATCC 13368 has been described as a 15β-hydroxylase showing also minor 11α-, 9α- and 6β-hydroxylase activity for progesterone conversion. Previously, mutant proteins with a changed selectivity towards 11α-OH-progesterone have already been produced. The challenge of this work was to create mutant proteins with a higher regioselectivity towards hydroxylation at positions 9 and 6 of the steroid molecule. 9α-hydroxyprogesterone exhibits pharmaceutical importance, because it is a useful intermediate in the production of physiologically active substances which possess progestational activity. Sixteen mutant proteins were selected from a library containing mutated proteins created by a combination of site-directed and saturation mutagenesis of active site residues. Four mutant proteins out of these catalyzed the conversion of progesterone to 9α-OH-progesterone as a main product. For further optimization site-directed mutagenesis was performed. The introduction of seven mutations (D217V, A243V, A106T, F165L, T89N, T247V or T247W) into these four mutant proteins led to 28 new variants, which were also used for an in vivo conversion of progesterone. The best mutant protein, F165L/A395E/G397V, showed a ten-fold increase in the selectivity towards progesterone 9α-hydroxylation compared with the wild type CYP106A2. Also 6β-OH-progesterone is a pharmaceutically important compound, especially as intermediate for the production of drugs against breast cancer. For the rational design of mutant proteins with 6β-selectivity, docking of the 3D-structure of CYP106A2 with progesterone was performed. The introduction of three mutations (T247A, A243S, F173A) led to seven new mutant proteins. Clone A243S showed the greatest improvement in 6β-selectivity being more than ten-fold. Finally, an in vivo conversion of 11-deoxycorticosterone (DOC), testosterone and cortisol with the best five mutant proteins displaying 9α- or 6β-hydroxylation, respectively, of progesterone was performed to investigate whether the introduced mutations also effected the conversion of other substrates.

A recombinant CYP11B1 dependent Escherichia coli biocatalyst for selective cortisol production and optimization towards a preparative scale

BACKGROUND

Human mitochondrial CYP11B1 catalyzes a one-step regio- and stereoselective 11β-hydroxylation of 11-deoxycortisol yielding cortisol which constitutes not only the major human stress hormone but also represents a commercially relevant therapeutic drug due to its anti-inflammatory and immunosuppressive properties. Moreover, it is an important intermediate in the industrial production of synthetic pharmaceutical glucocorticoids. CYP11B1 thus offers a great potential for biotechnological application in large-scale synthesis of cortisol. Because of its nature as external monooxygenase, CYP11B1-dependent steroid hydroxylation requires reducing equivalents which are provided from NADPH via a redox chain, consisting of adrenodoxin reductase (AdR) and adrenodoxin (Adx).

RESULTS

We established an Escherichia coli based whole-cell system for selective cortisol production from 11-deoxycortisol by recombinant co-expression of the demanded 3 proteins. For the subsequent optimization of the whole-cell activity 3 different approaches were pursued: Firstly, CYP11B1 expression was enhanced 3.3-fold to 257 nmol∗L(-1) by site-directed mutagenesis of position 23 from glycine to arginine, which was accompanied by a 2.6-fold increase in cortisol yield. Secondly, the electron transfer chain was engineered in a quantitative manner by introducing additional copies of the Adx cDNA in order to enhance Adx expression on transcriptional level. In the presence of 2 and 3 copies the initial linear conversion rate was greatly accelerated and the final product concentration was improved 1.4-fold. Thirdly, we developed a screening system for directed evolution of CYP11B1 towards higher hydroxylation activity. A culture down-scale to microtiter plates was performed and a robot-assisted, fluorescence-based conversion assay was applied for the selection of more efficient mutants from a random library.

CONCLUSIONS

Under optimized conditions a maximum productivity of 0.84 g cortisol∗L(-1)∗d(-1) was achieved, which clearly shows the potential of the developed system for application in the pharmaceutical industry.

Terpene hydroxylation with microbial cytochrome P450 monooxygenases

Terpenoids comprise a highly diverse group of natural products. In addition to their basic carbon skeleton, they differ from one another in their functional groups. Functional groups attached to the carbon skeleton are the basis of the terpenoids' diverse properties. Further modifications of terpene olefins include the introduction of acyl-, aryl-, or sugar moieties and usually start with oxidations catalyzed by cytochrome P450 monooxygenases (P450s, CYPs). P450s are ubiquitously distributed throughout nature, involved in essential biological pathways such as terpenoid biosynthesis as well as the tailoring of terpenoids and other natural products. Their ability to introduce oxygen into nonactivated C-H bonds is unique and makes P450s very attractive for applications in biotechnology. Especially in the field of terpene oxidation, biotransformation methods emerge as an attractive alternative to classical chemical synthesis. For this reason, microbial P450s depict a highly interesting target for protein engineering approaches in order to increase selectivity and activity, respectively. Microbial P450s have been described to convert industrial and pharmaceutically interesting terpenoids such as ionones, limone, valencene, resin acids, and triterpenes (including steroids) as well as vitamin D3. Highly selective and active mutants have been evolved by applying classical site-directed mutagenesis as well as directed evolution of proteins. As P450s usually depend on electron transfer proteins, mutagenesis has also been applied to improve the interactions between P450s and their respective redox partners. This chapter provides an overview of terpenoid hydroxylation reactions catalyzed by bacterial P450s and highlights the achievements made by protein engineering to establish productive hydroxylation processes.

Steroid conversion with CYP106A2 - production of pharmaceutically interesting DHEA metabolites

Background

Steroids are lipophilic compounds with a gonane skeleton and play an important role in higher organisms. Due to different functionalizations - mainly hydroxylations - at the steroid molecule, they vary highly in their mode of action. The pharmaceutical industry is, therefore, interested in hydroxysteroids as therapeutic agents. The insertion of hydroxyl groups into a steroid core, however, is hardly accomplishable by classical chemical means; that is because microbial steroid hydroxylations are investigated and applied since decades. CYP106A2 is a cytochrome P450 monooxygenase from Bacillus megaterium ATCC 13368, which was first described in the late 1970s and which is capable to hydroxylate a variety of 3-oxo-delta4 steroids at position 15beta. CYP106A2 is a soluble protein, easy to express and to purify in high amounts, which makes this enzyme an interesting target for biotechnological purposes.

Results

In this work a focused steroid library was screened in vitro for new CYP106A2 substrates using a reconstituted enzyme assay. Five new substrates were identified, including dehydroepiandrosterone and pregnenolone. NMR-spectroscopy revealed that both steroids are mainly hydroxylated at position 7beta. In order to establish a biotechnological system for the preparative scale production of 7beta-hydroxylated dehydroepiandrosterone, whole-cell conversions with growing and resting cells of B. megaterium ATCC1336 the native host of CYP1062 and also with resting cells of a recombinant B. megaterium MS941 strain overexpressing CYP106A2 have been conducted and conversion rates of 400 muM/h (115 mg/l/h) were obtained. Using the B. megaterium MS941 overexpression strain, the selectivity of the reaction was improved from 0.7 to 0.9 for 7beta-OH-DHEA.

Conclusions

In this work we describe CYP106A2 for the first time as a regio-selective hydroxylase for 3-hydroxy-delta5 steroids. DHEA was shown to be converted to 7beta-OH-DHEA which is a highly interesting human metabolite, supposed to act as neuroprotective, anti-inflammatory and immune-modulatory agent. Optimization of the whole-cell system using different B. megaterium strains lead to a conversion of DHEA with B. megaterium showing high selectivity and conversion rates and displaying a volumetric yield of 103 mg/l/h 7beta-OH-DHEA.

Cytochromes P450 as promising catalysts for biotechnological application: chances and limitations

Cytochromes P450 (CYPs) belong to the superfamily of heme b containing monooxygenases with currently more than 21,000 members. These enzymes accept a vast range of organic molecules and catalyze diverse reactions. These extraordinary capabilities of CYP systems that are unmet by other enzymes make them attractive for biotechnology. However, the complexity of these systems due to the need of electron transfer from nicotinamide adenine dinucleotide (phosphate) (NAD(P)H) via redox partner proteins for the initial hydroxylation step limits a broader technical implementation of CYP enzymes. There have been several reviews during the past years tackling the potential CYPs for synthetic application. The aim of this review is to give a critical overview about possibilities and chances for application of these interesting catalysts as well as to discuss drawbacks and problems related to their use. Solutions to overcome these limitations will be demonstrated, and several selected examples of successful CYP applications under industrial conditions will be reviewed.

Functional characterization of steroid hydroxylase CYP106A1 derived from Bacillus megaterium

In this study, we examined the catalytic activity of CYP106A1 from the Bacillus megaterium American Type Culture Collection 14581 strain. The CYP106A1 gene was cloned from B. megaterium, heterologously expressed in Escherichia coli, and purified. Potential electron partners and possible bacterial CYP106A1 substrates were identified by examining the oxidative activity toward a set of steroids in the presence of several reductase systems. The activities of CYP106A1 in a reconstituted system could not be achieved using rat NADPH-P450 reductase or a putidaredoxin reductase-putidaredoxin pair. However, the spinach redox proteins, a ferredoxin reductase-ferredoxin pair, were found to be efficient redox partners for CYP106A1. CYP106A1 catalyzes the hydroxylation of a set of steroids including testosterone, progesterone, 17α-hydroxyprogesterone, 11-deoxycorticosterone, corticosterone, and 11-deoxycortisol to produce monohydroxylated products as the major metabolites. These results suggest that CYP106A1 would be useful for the bioconversion of steroid hormones to hydroxylated products that can be used for industrial applications.

Cytochromes P450: exploiting diversity and enabling application as biocatalysts

The remarkable chemical reactivity and substrate range displayed by cytochromes P450 (P450s) renders them attractive as potential catalysts for a host of challenging chemical reactions in industry. The opportunities afforded by these biocatalysts are increased by the availability of greater diversity provided by the genomic resource and the variant libraries of well-known P450s produced by rational and random engineering techniques. The exploitation of this enormous diversity will require novel tools in screening, to identify enzyme reactions of interest, and also in the enabling of these valuable activities through protein engineering and bioprocess optimisation.

Regioselective hydroxylation of norisoprenoids by CYP109D1 from Sorangium cellulosum So ce56

Sesquiterpenes are particularly interesting as flavorings and fragrances or as pharmaceuticals. Regio- or stereoselective functionalizations of terpenes are one of the main goals of synthetic organic chemistry, which are possible through radical reactions but are not selective enough to introduce the desired chiral alcohol function into those compounds. Cytochrome P450 monooxygenases are versatile biocatalysts and are capable of performing selective oxidations of organic molecules. We were able to demonstrate that CYP109D1 from Sorangium cellulosum So ce56 functions as a biocatalyst for the highly regioselective hydroxylation of norisoprenoids, alpha- and beta-ionone, which are important aroma compounds of floral scents. The substrates alpha- and beta-ionone were regioselectively hydroxylated to 3-hydroxy-alpha-ionone and 4-hydroxy-beta-ionone, respectively, which was confirmed by (1)H NMR and (13)C NMR. The results of docking alpha- and beta-ionone into a homology model of CYP109D1 gave a rational explanation for the regio-selectivity of the hydroxylation. Kinetic studies revealed that alpha- and beta-ionone can be hydroxylated with nearly identical V (max) and K (m) values. This is the first comprehensive investigation of the regioselective hydroxylation of norisoprenoids by CYP109D1.

Adrenodoxin: the archetype of vertebrate-type [2Fe-2S] cluster ferredoxins

Adrenodoxin is probably the best characterized member of the vertebrate-type [2Fe-2S]-cluster ferredoxins. It has been in the spotlight of scientific interest for many years due to its essential role in mammalian steroid hormone biosynthesis, where it acts as electron mediator between the NADPH-dependent adrenodoxin reductase and several mitochondrial cytochromes P450. In this review we will focus on the present knowledge about protein-protein recognition in the mitochondrial cytochrome P450 system and the modulation of the electron transfer between Adx and its redox partners, AdR and CYP(s). We also intend to point out the potential biotechnological applications of Adx as a versatile electron donor to different cytochromes P450, both in vitro and in vivo. Finally we will address the comparison between the mammalian cytochrome P450-associated adrenodoxin and ferredoxins involved in iron-sulfur-cluster biosynthesis. Despite their different functions, these proteins display an amazing similarity regarding their primary sequence, tertiary structure and biophysical features.

Protein engineering of microbial enzymes

Protein engineering has emerged as an important tool to overcome the limitations of natural enzymes as biocatalysts. Recent advances have mainly focused on applying directed evolution to enzymes, especially important for organic synthesis, such as monooxygenases, ketoreductases, lipases or aldolases in order to improve their activity, enantioselectivity, and stability. The combination of directed evolution and rational protein design using computational tools is becoming increasingly important in order to explore enzyme sequence-space and to create improved or novel enzymes. These developments should allow to further expand the application of microbial enzymes in industry.