Production of High Value Drug Metabolites Using Engineered Cytochromes P450

Production of high value drug metabolites using engineered cytochromes P450

A thesis submitted to The University of Manchester for the degree of Doctor of Philosophy in the Faculty of Life Sciences

2013

Christopher Butler

Table of Contents

Title Page 1 Table of Contents 2 List of Figures 8 List of Tables 15 Abbreviations 17 Abstract 20 Declaration and Copyright statement 21 Acknowledgements 22 Preface to the alternate format thesis 23

Chapter 1 – Introduction 26

1.1 Drug metabolites 26 1.1.1 What are they? 26 1.1.2 How are they produced? 26 1.1.3 Why are they important? 27 1.1.4 The enzymes involved 28 1.1.5 Regulatory guidance 29 1.1.6 Reasons for regulation 30 1.1.7 Genetics 30 1.1.8 Current methods of metabolite synthesis 31

1.2 Cytochromes P450 32 1.2.1 Overview 32 1.2.2 Physiological roles 33 1.2.3 Heme proteins 36 1.2.4 The P450 catalytic cycle 37 1.2.5 Active species in the P450 catalytic cycle 38 1.2.6 The P450 heme iron spin state 41 1.2.7 Cytochrome P450 structures 42 1.2.8 Redox partners 43 1.2.9 P450 electron transfer reactions 46

1.3 P450 BM3 from Bacillus megaterium 51 1.3.1 Overview 51 1.3.2 P450 BM3 structure 53

1.4 Biotechnology 57

1.5 References 62

Chapter 2 – Title page 83 “Key mutations alter the cytochrome P450 BM3 conformational landscape and remove inherent substrate bias”

2.1 Summary 84

2.2 Introduction 85

2.3 Experimental Procedures 87 2.3.1 Generation, expression and purification of WT and 87 variant P450 BM3 proteins 2.3.2 Quantification of P450 BM3 enzymes and 88 determination of their substrate affinity and steady- state kinetic properties 2.3.3 Omeprazole and 5-OH omeprazole turnover and 89 analysis by LC-MS 2.3.4 Omeprazole turnover and analysis by NMR 90 2.3.5 Examination of hemoprotein stability by differential 90 scanning calorimetry 2.3.6 Crystallization of P450 BM3 heme domains and 91 determination of protein structures

2.4 Materials 92

2.5 Results 93 2.5.1 Characterization of omeprazole binding properties of 93 BM3 variants 2.5.2 Steady-state kinetics of P450 BM3 variants with 96 omeprazole 2.5.3 Oxidation of omeprazole by WT and variant P450 96 BM3 enzymes 2.5.4 Structural analysis of omeprazole-binding P450 BM3 100 variants

2.6 Discussion 112

2.7 Supplemental Data 116 2.7.1 Characterization of substrate and oxidized products 116 by NMR spectroscopy 2.7.2 Differential Scanning Calorimetry (DSC) 121

2.8 References 124

2.9 Footnotes 130

Chapter 3 – Title Page 131 “Human P450-like oxidative transformations of proton pump inhibitor drugs by a P450 BM3 variant that induces conformational reconfiguration of the enzyme”

3.1 Abstract 132

3.2 Introduction 133

3.3 Experimental Procedures 135 3.3.1 Mutagenesis and expression of WT and variant P450 135 BM3 enzymes 3.3.2 Purification of WT and variant intact P450 BM3 and 135 heme domains

3.3.3 P450 quantification 3.3.4 Substrate binding and kinetic properties of WT and 136 variant intact BM3 with fatty acids and PPI substrates 136 3.3.5 EPR spectroscopy 3.3.6 Enzymatic oxidation of substrates and product 137 characterization 137 3.3.7 Crystallization of BM3 heme domains and determination of protein structures 139

3.4 Materials 140 3.5 Results 3.5.1 Spectral binding studies 141 3.5.2 Steady state kinetics 147 3.5.3 EPR spectroscopy 151 3.5.4 Oxidation of PPI’s by WT and variant P450 BM3 157 enzymes using Liquid Chromatography Mass Spectrometry (LCMS) 3.5.5 Oxidation of PPI’s by WT and Variant P450 BM3 165 enzymes using NMR 3.5.6 X-ray crystallography 180

3.6 Discussion 184

3.7 References 187

Chapter 4 – Title Page 193 “Oxidation of diverse drug molecules by P450 BM3 gatekeeper variants”

4.1 Abstract 194

4.2 Introduction 195

4.3 Materials and methods 198 4.3.1 Mutagenesis and expression of WT and variant P450 198 BM3 enzymes 4.3.2 Purification of WT and variant P450 BM3 and heme 198 domains 4.3.3 P450 quantification 199 4.3.4 Thermofluor binding assay 199 4.3.5 EPR Spectroscopy 200 4.3.6 Analysis of the kinetics of substrate-dependent 200 NADPH oxidation by WT and variant intact BM3 enzymes 4.3.7 Drug turnover and analysis by LCMS 201

4.4 Materials 202

4.5 Results 203 4.5.1 Thermal unfolding of WT and variant P450 BM3 203 heme domains using a Thermofluor assay 4.5.2 Electron Paramagnetic Resonance spectrometry 205 (EPR) 4.5.3 Steady state kinetic analysis of drug-dependent 209 NADPH oxidation 4.5.4 Product identification from P450 BM3 turnover of drug 213 substrates by LCMS

4.6 Discussion 222

4.7 References 229

Chapter 5 – Summary, conclusions and 235 further work

5.1 Summary 235

5.2 Conclusions 245

5.3 Further Work 246

5.4 References 247

Appendix 251

List of Figures

Chapter 1

Figure 1.1 - Simplified reaction pathway for 33 cytochromes P450

Figure 1.2 - Common P450 reaction types 34

Figure 1.3 - The major forms of heme prosthetic groups 36

Figure 1.4 - The P450 catalytic cycle 38

Figure 1.5 - The P450 radical rebound mechanism 39

Figure 1.6 - The ferric heme iron d-orbital electron state 41 in low spin and high spin configurations

Figure 1.7 - General P450 structure 42

Figure 1.8 - Structures of the P450 reductase enzyme 44 cofactors

Figure 1.9 - Schematic representation of the 3 major 45 P450 classes

Figure 1.10 - Diagrammatic representation of flavin 46 oxidation states

Figure 1.11 - The electron transfer process in type I 47 P450 systems

Figure 1.12 - Electron transfer in type II P450 redox 49 systems that use CPR

Figure 1.13 - P450 electron transfer in type III redox 50 systems

Figure 1.14 - The crystal structure of the heme domain 54 of P450 BM3 with N-palmitoylglycine (NPG) bound (PDB code 1PJZ)

Chapter 2

Figure 2.1 - The structure of omeprazole 93

Figure 2.2 - Binding and oxidation of omeprazole by 95 P450 BM3 variants

Figure 2.3 - LC-MS analysis of products derived from 97 omeprazole oxidation by the P450 BM3 F87V/A82F (DM) double mutant enzyme

Figure 2.4 - Optical binding titration for the BM3 DM with 99 5-OH OMP

Figure 2.5 - Time course of substrate oxidation and 100 product formation in the reaction of the P450 BM3 DM enzyme with omeprazole

Figure 2.6 - Structures of P450 BM3 enzymes and their 101 omeprazole binding sites

Figure 2.7 - Interactions of omeprazole in the active site 105 of the A82F BM3 heme domain

Figure 2.8 - Stereoviews of structural overlays of 107 substrate-bound forms of the BM3 A82F- containing variant heme domains with WT BM3

Figure 2.9 - Structural overlay of the omeprazole-bound 109 A82F variant with the WT BM3 heme domain

Figure 2.10 - Conformational equilibria and the 111 relationship with structural stability in P450 BM3

Chapter 2 – Supplemental Data

Figure 2.S1 - 1H NMR spectrum of omeprazole 117

Figure 2.S2 - HMBC spectrum of omeprazole 118

Figure 2.S3 - 1H NMR spectrum of turnover products 119 from omeprazole oxidation

Figure 2.S4 - HMBC spectra of turnover products from 120 omeprazole oxidation

Figure 2.S5 - Graphical overlay of DSC data for WT and 123 variant BM3 heme domains

Chapter 3

Figure 3.1 - Proton pump inhibitor (PPI) drug structure 141 and functional groups

Figure 3.2 - Esomeprazole binding to the P450 BM3 DM 142 enzyme

Figure 3.3 - Pantoprazole binding to the P450 BM3 DM 143 enzyme

Figure 3.4 - Lansoprazole and rabeprazole binding to 143 the DM P450 BM3 enzyme

Figure 3.5 - Esomeprazole binding to the P450 BM3 144 A82F enzyme

Figure 3.6 - Pantoprazole binding to the P450 BM3 145 A82F enzyme

Figure 3.7 - Lansoprazole and rabeprazole binding to 145 the A82F P450 BM3 enzyme

Figure 3.8 - PPI-dependent steady state kinetic analysis 148 for the A82F P450 BM3 variant

Figure 3.9 - PPI-dependent steady state kinetic analysis 149 for the F87V P450 BM3 variant

Figure 3.10 - PPI-dependent steady state kinetic 150 analysis for the F87V/A82F (DM) P450 BM3

Figure 3.11 - PPI-dependent steady state kinetic 151 analysis for the WT P450 BM3 with lansoprazole

Figure 3.12 - EPR analysis of WT P450 BM3 heme 153 domain

Figure 3.13 - EPR analysis of the A82F P450 BM3 154 heme domain

Figure 3.14 - EPR analysis of the F87V BM3 heme 155 domain

Figure 3.15 - EPR analysis of the DM P450 BM3 heme 156 domain

Figure 3.16 - Reactions schemes outlining pathways of 157 P450 metabolism of PPI drugs

Figure 3.17 - LC-MS traces of esomeprazole turnover 158 by the P450 BM3 DM enzyme

Figure 3.18 - LC-MS traces for lansoprazole turnover by 160 the P450 BM3 DM enzyme

Figure 3.19 - LC-MS traces showing pantoprazole 161 turnover from the P450 BM3 DM enzyme

Figure 3.20 - LC-MS traces for rabeprazole turnover by 163 BM3 enzymes

Figure 3.21 - Proportions of PPI turnover products 165 identified by LCMS

Figure 3.22 - 1H NMR spectrum of esomeprazole 167

Figure 3.23 - HMBC spectrum of esomeprazole 168

Figure 3.24 - 1H NMR spectrum of turnover products 169 from esomeprazole oxidation

Figure 3.25 - HMBC spectra of turnover products from 170 esomeprazole oxidation

Figure 3.26 - 1H NMR spectrum of lansoprazole 172

Figure 3.27 - 1H NMR spectrum of the lansoprazole 173 sulfone standard

Figure 3.28 - 1H NMR spectrum of turnover products 174 from lansoprazole oxidation

Figure 3.29 - 1H NMR spectrum of rabeprazole 176

Figure 3.30 - COSY NMR spectrum of rabeprazole 177

Figure 3.31 - HMBC spectrum of rabeprazole 178

Figure 3.32 - 1H NMR spectrum of turnover products 179 from rabeprazole (100 µM) oxidation

Figure 3.33 - 1H NMR spectrum of turnover products 180 from rabeprazole (50 µM) oxidation

Figure 3.34 - Overlay of the DM BM3 heme domain 182 esomeprazole crystal structure with the previously reported DM omeprazole structure (4KEY)

Figure 3.35 - Figure 3.35 Density overlay of the 183 pantoprazole-bound DM heme domain crystal structure with the esomeprazole-bound DM heme domain structure.

Chapter 4

Figure 4.1 - P450 enzyme stability using the Thermofluor™ 205 method

Figure 4.2 - EPR spectra of ligand-free and drug-bound 208 WT and variant P450 BM3 heme domains

Figure 4.3 - Steady state kinetics of drug-dependent 211 NADPH oxidation for BM3 variants

Figure 4.4 - Oxidation of drugs by P450 BM3 variants 213

Figure 4.5 - LCMS traces for P450 BM3-catalysed DM 216 nifedipine (NIF) turnover

Figure 4.6 - LCMS traces for P450 BM3 DM catalysed 217 amodiaquine (AMO) turnover

Figure 4.7 - LCMS traces for P450 BM3 A82F variant 218 catalysed diclofenac (DCF) turnover

Figure 4.8 - LCMS traces for P450 BM3 DM catalysed 219 dextromethorphan (DEX) turnover

Figure 4.9 - LCMS traces of P450 BM3 F87V catalysed 220 diazepam (DIA) turnover

Figure 4.10 - Product formation from P450 BM3 221 oxidation of human drugs

List of Tables

Chapter 1

Table 1.1 - Common drug metabolism reactions 28

Chapter 2

Table 2.1 - Substrate binding and turnover data for 94 P450 BM3 enzymes

Table 2.2 - Binding and kinetics of oxidation of 5-OH 99 OMP by WT and variant P450 BM3 enzymes

Table 2.3 - Data reduction and final structural 103 refinement statistics for P450 BM3 variants and their OMP-substrate complexes

Table 2.S1 - DSC data for thermal unfolding of WT and 122 variant P450 BM3 heme domains in ligand-free and substrate-bound forms

Chapter 3

Table 3.1 - Binding and steady state kinetics for WT and 146 variant P450 BM3 enzymes with PPI substrates.

Chapter 4

Table 4.1 - Transition midpoint (Tm) values for WT and 204 variant BM3 heme domains using thermal unfolding assay

Table 4.2 - NADPH dependent drug oxidation kinetics 210 for WT and variant P450 BM3 enzymes

Abbreviations

5-COOH OMP 5-Carboxy omeprazole 5-OH OMP 5-Hydroxy omeprazole Å Angstrom A Absorbance A82F A82F variant of P450 BM3 ACN Acetonitrile AMO Amodiaquine amu Atomic mass unit BM3 Cytochrome P450 BM3 (CYP102A1)Correlation COSY spectroscopy CPR Cytochrome P450 reductase CYP Cytochrome P450 DCF Diclofenac DEAE Diethyl aminoethyl cellulose DEX Dextromethorphan DIA Diazepam dH2O Distilled deionised water DM Double Mutant (F87V/A82F Variant) of P450 BM3 DMSO Dimethyl sulfoxide DNA Deoxyribonucleic acid DSC Differential scanning calorimetry DSF Differential scanning fluorimetry EDTA Ethylenediaminetetraacetic acid EPR Electron paramagnetic resonance ESO Esomeprazole (S-isomer of omeprazole) F87V F87V variant of P450 BM3 FAD Flavin adenine dinucleotide FDA Food and Drug Administration FMN Flavin mononucleotide ΔHcal Calorimetric enthalpy HD Heme domain HMBC Heteronuclear Multiple Bond Correlation HMQC Heteronuclear Multiple Quantum Correlation

HQ Hydroquinone HS High spin

ΔHVH Van’t Hoff enthalpy

Kd Dissociation constant kDa Kilo Dalton KPi Potassium phosphate buffer LAN Lansoprazole LB Luria Bertani bacterial growth medium LCMS Liquid chromatography mass spectrometry LS Low spin MeOH Methanol

Mr Molecular mass MS Mass spectrometry NAD (NAD+) Nicotinamide adenine dinucleotide NADH NAD (reduced form) NADP (NADP+) Nicotinamide adenine dinucleotide phosphate NADPH NADP (reduced form) ND Not determinable NIF Nifedipine NMR Nuclear magnetic resonance NPG N-palmitoylglycine OD Optical density O/N Overnight OMP Omeprazole P450 Cytochrome P450 PAN Pantoprazole PCR Polymerase chain reaction PDB Protein Data Bank PEG Polyethylene glycol pKa Acid dissociation constant PPI Proton pump inhibitor RAB Rabeprazole RD Reductase domain RPM Revolutions per minute RT Retention time SDS-PAGE Sodium dodecyl sulfate polyacrylamide gel electrophoresis

SQ Semiquinone TAE Tris/acetate/EDTA TB Terrific Broth TE Tris/EDTA

Tm Transition midpoint (or melting temperature) Tris Tris(hydroxymethyl)aminomethane TMS Tetramethylsilane UV Ultraviolet v/v Volume to volume WT Wild Type P450 BM3 w/v Weight to volume

Abstract Production of drug metabolites is an area of biotechnology where manufacture by cytochromes P450 (P450s) offers significant advantages over synthetic methods. P450 BM3 from Bacillus megaterium is an ideal candidate for metabolite synthesis. It is a soluble fusion of a P450 to its partner (cytochrome P450 reductase) with the fastest reaction rate for any P450 in oxidising fatty acids. BM3 is structurally very well understood, and has been studied extensively by random mutagenesis, with the aim of generating high activity oxidase variants for biotechnological uses. Several mutations occur frequently in such variants, but there is little understanding of how common mutations underpin major alterations in BM3 substrate selectivity and activity. This thesis provides new information for BM3 Ala82 and Phe87 variants, two of the most commonly mutated residues, to explain their impact on defining novel oxidase function using a combination of structural and solution data. It is shown that two common conformations in BM3 P450 structures – the substrate free (SF) and substrate bound (SB) states – are in an equilibrium (regardless of binding of substrates), and that thermodynamic stability of the enzyme is key to their energetic separation. The A82F mutation destabilises SF, allowing easier access to SB, and generating a more malleable P450 that binds diverse compounds. The F87V mutation removes steric bulk from the active site, allowing easier access for non-natural substrates to bind in a less strained conformation above the P450 heme. In isolation and combination, these mutations induce major changes in BM3 catalytic properties. The gastric proton pump inhibitor (PPI) drug omeprazole binds tightly to BM3 A82F variants. X-ray crystal structures show that the 5-methyl group of the PPIs omeprazole/esomeprazole is close to the heme, providing the first BM3 structure in an “active” conformation for substrate binding position. These PPIs are oxidized on the 5-methyl group. The main human PPI metabolising enzymes are CYP2C19 and CYP3A4. Our data show that the CYP2C19 metabolites of omeprazole, esomeprazole and rabeprazole are also those produced by BM3 variants, and with lansoprazole the major metabolite is also that for CYP3A4. Expanding these studies to other human P450 marker drug substrates revealed that, by differential scanning fluorimetry (DSF) and electron paramagnetic resonance (EPR), multiple binding interactions occur between variant BM3 P450s and drugs such as diclofenac and dextromethorphan. Although drug substrates used are markers for diverse human P450s, product analysis showed that BM3 variants produced mainly a human drug metabolite in each case. This suggests specific oxidation reactions are favoured for P450s, and occur preferentially when substrates are mobile in P450 active sites. Our work shows that certain “gatekeeper” mutations in BM3 create conformational flexibility by destabilising the SF versus the SB state. This provides novel routes to screen for altered activity by using stability, rather than the more constrained single activity screening techniques.

Declaration

No portion of the work referred to in the thesis has been submitted in support of an application for another degree or qualification of this or any other university or other institute of learning. However, this thesis is written in the University of Manchester’s alternative format which allows the inclusion of work already published.

Copyright statements i. The author of this thesis (including any appendices and/or schedules to this thesis) owns certain copyright or related rights in it (the “Copyright”) and s/he has given The University of Manchester certain rights to use such Copyright, including for administrative purposes. ii. Copies of this thesis, either in full or in extracts and whether in hard or electronic copy, may be made only in accordance with the Copyright, Designs and Patents Act 1988 (as amended) and regulations issued under it or, where appropriate, in accordance with licensing agreements which the University has from time to time. This page must form part of any such copies made. iii. The ownership of certain Copyright, patents, designs, trade marks and other intellectual property (the “Intellectual Property”) and any reproductions of copyright works in the thesis, for example graphs and tables (“Reproductions”), which may be described in this thesis, may not be owned by the author and may be owned by third parties. Such Intellectual Property and Reproductions cannot and must not be made available for use without the prior written permission of the owner(s) of the relevant Intellectual Property and/or Reproductions. iv. Further information on the conditions under which disclosure, publication and commercialisation of this thesis, the Copyright and any Intellectual Property and/or Reproductions described in it may take place is available in the University IP Policy (see http://www.campus.manchester.ac.uk/medialibrary/policies/intellectual-property.pdf), in any relevant Thesis restriction declarations deposited in the University Library, The University Library’s regulations (see http://www.manchester.ac.uk/library/aboutus/regulations) and in The University’s policy on presentation of Theses.

Acknowledgements

First and foremost, I would like to thank my supervisor Andrew Munro for the opportunity to undertake a PhD, and his support and guidance throughout the process. I would like to thank Michael Voice and Caroline Peet and all the team at Cypex for the opportunity they have afforded me. The project would not have developed as it has without the facilities I was able to use and the expert guidance from Cal at Cypex. So many others contributed their knowledge and ideas it would be impossible to acknowledge them all.

I couldn’t have asked for a better group of people than the Munro lab, my time there has left me with firm friendships which will stay with me forever. Special thanks goes to Dr Amy Mason, for teaching me all I know about biology, making work fun and always having ideas for every problem we encountered. Also to Dominika, Soi, Shannon and Linus for listening to all my rants, giving me someone to work (and Climb) with on all those Saturdays and late nights, and reminding me that work isn’t everything in life.

Lastly but by no means least, I thank my family for their support, encouraging me to achieve all I have and giving me every opportunity to pursue my chosen path.

Preface to the alternative format thesis

This thesis in presented in the University of Manchester alternative style of PhD thesis. This allows the inclusion of data already published and prepared for publication in peer reviewed journals. The structure of each chapter therefore follows that of the journal in which it is prepared for publications, meaning a separate methods chapter is not included as they are in the individual results chapters, and each chapters heading labels may differ. Therefore the layout of the thesis is as follows; an abstract, introduction, three results chapters (papers), and a summary and conclusions chapter, each with self-contained references. The text formatting and page numbering has been altered to keep it consistent throughout the thesis.

The alternative format thesis has been introduced to be more relevant to the processes involved in scientific research, with particular importance to the reporting of such research and publication of papers. The nature of scientific publications is collaborative; in order to better achieve a scientific goal. For this reason, as part of the alternative format, the contributions of each co-author to the papers must be established.

Papers included as results chapters

Chapter 2

Butler CF, Peet C, Mason AE, Voice MW, Leys D, and Munro AW (2013). “Key mutations alter the cytochrome P450 BM3 conformational landscape and remove inherent substrate bias.” J Biol Chem Epub ahead of print.

Chapter 3

Butler CF, Peet C, Mason AE, Mclean KJ, Fisher K, Rigby SE, Voice MW, Leys D and Munro AW (2013). “Human P450-like oxidative transformations of proton pump inhibitor drugs by a P450 BM3 variant that induces conformational reconfiguration of the enzyme.” Biochem J Submitted

Chapter 4

Butler CF, Peet C, Mason AE, Fisher K, Rigby SE, Voice MW, Leys D and Munro AW (2013). “Oxidation of diverse drug molecules by P450 BM3 gatekeeper variants.” (2013) This paper is currently an unpublished manuscript. It is hoped that once additional data is collected it will be published in a biochemical journal such as the Journal of Biological Chemistry.

Contributions from authors

As PhD supervisor, Andrew W. Munro contributed data analysis and manuscript preparation for all papers.

Chapter 2 Peet C, Aided with turnover method development and LCMS method development and analysis of products by LCMS.

Mason AE, performed molecular biology to design the various P450 BM3 constructs used and inset mutations.

Voice MW, Provided equipment and laboratory space at Cypex to undertake experiments as well as project funding.

Leys D, performed X-ray diffraction of the P450 BM3 crystals at the Diamond syncatron source, solved the structures by molecular replacement and prepared the figures for publication.

Chapter 3 Peet C, Aided with turnover method development and LCMS method development and analysis of products by LCMS.

Mason AE, performed molecular biology to design the various P450 BM3 constructs used and inset mutations.

McLean KJ, undertook analysis of the pantoprazole turnover product by LCMS at Agilent, Cheadle, UK.

Rigby, SE. and Fisher, K are responsible for the operation of the EPR facility at the University of Manchester and aided with analysis of EPR data.

Voice MW, Provided equipment and laboratory space at Cypex to undertake experiments as well as project funding.

Leys D, performed X-ray diffraction of the P450 BM3 crystals at the Diamond syncatron source, solved the structures by molecular replacement and prepared the figures for publication.

Chapter 4 Peet C, Aided with turnover method development and LCMS method development and analysis of products by LCMS.

Mason AE, performed molecular biology to design the various P450 BM3 constructs used and inset mutations.

Rigby, SE and Fisher K are responsible for the operation of the EPR facility at the University of Manchester and aided with analysis of EPR data.

Voice MW, Provided equipment and laboratory space at Cypex to undertake experiments as well as project funding.

1. Introduction

1.1 Drug metabolites

1.1.1 What are they?

Xenobiotics are chemicals that are outside an organism’s normal metabolic processes. Thus while testosterone is a natural steroid, dexamethasone is a xenobiotic steroid (1). Almost all known drugs fit into this xenobiotic classification as they are generally derived from synthetic sources (e.g. the benzodiazepine drug diazepam) or are extrinsic natural products (e.g. the opiate analgesic drug morphine). The metabolism of drugs and other xenobiotics can be classified into phases, where phase I involves reactions that increase the polarity of the compounds to increase their clearance, and phase II involves mainly conjugation reactions for removal of reactive groups and to achieve increased active transport to facilitate excretion (2).

1.1.2 How are they produced?

The life cycle of a drug in the body can be expressed in four ways (3).

1. Elimination unchanged 2. Retention unchanged 3. Spontaneous chemical transformation 4. Enzymatic transformation

There is considerable scientific interest in the study of drugs and their enzymatic transformations (pharmacology and pharmacokinetics). Many such transformations are carried out by microsomal oxidase enzymes in the liver, known as cytochromes P450 (P450s or CYPs). These enzymes are classed as microsomal, though this is a biochemical characterization based on the fractionation of homogenised cells, and is not indicative of a particular cellular location (4).

As indicated above, enzymatic transformation of xenobiotics consists of two major pathway classes known as phase I and phase II metabolism. Phase I enzymes are the primary metabolisers and are responsible for several oxidation, reduction and hydrolysis reactions. In

contrast, phase II enzymes are responsible for conjugation reactions that often involve the phase I metabolites. The phase I enzymes are predominantly found in the liver, although they can be located all over the body with other major sites including the brain (5), kidney (6), and lungs (7). This concentration of metabolising enzymes explains the predominance of drug side effects caused by toxic metabolites in the liver and kidneys, as high concentrations of the metabolites are produced here and can affect the surrounding tissue (8). An example of this is liver necrosis caused by cytochrome P450 activation of acetaminophen, primarily through formation of N-acetyl-p-benzo-quinone imine (NAPQI) that reacts irreversibly with glutathione sulfhydryl groups (9).

Reduction reactions also occur in phase I metabolism (10), though they are relatively rare and the primary pathways are oxidative. This is understandable, as increasing renal clearance is often easiest to achieve by adding polar groups.

Microorganisms can be both the best producers of xenobiotics and the most efficient metabolisers of them. Many xenobiotics are metabolised by microorganisms both in soil and water, and these are often cited as sources of bioremediation (11). Sometimes problems with using human only models for drug metabolism testing occur when metabolites produced by the gut flora, and not by human enzymes, turn out to be toxic (12). An example is the conversion of arbutin to hydroquinone in gut microflora. This mutagenic and carcinogenic compound is not seen in models using only human enzymes (13).

1.1.3 Why are they important?

There are many metabolic reactions involved in drug metabolism (Table 1.1), although the main area of importance for biotechnological application is that involving phase I metabolites generated through enzymatic transformations. These are the reactions primarily caused by the oxidoreductase class of enzymes, largely CYP enzymes, but also with involvement of some flavin-containing and other enzymes (e.g. flavin monooxygenases).

Reaction Enzyme Class Hydroxylation CYP (Cytochrome P450) Dealkylation CYP Epoxidation CYP Desulfuration CYP, FMO (Flavin monooxygenase) Glutathione conjugation GST (Glutathione S-transferase) Sulfation SULT (Sulfotransferase) Sulfoxidation CYP Acetylation Acetyl transferase Ester Cleavage CYP, Esterase N-Oxidation CYP, FMO

Table 1.1. Common drug metabolism reactions and the enzyme classes responsible for the reactions. Adapted from Croom (2012) (14).

As we can see from Table 1.1 the P450s are the enzymes with primary responsibility for these type I metabolism reactions (reactions excluding conjugations). The P450s are important as they catalyse a broad range of reactions from hydroxylation to aromatic dehalogenation (15), performing essential metabolic reactions in the clearance of toxic compounds.

1.1.4 The Enzymes involved

Phase I enzymes, e.g. P450s, prepare very lipophilic compounds for conjugation by adding a reactive functional group, e.g. hydroxyl (OH). Phase II enzymes conjugate a highly water soluble compound, such as acetate or a glucuronosyl group, to produce a compound more readily excreted in urine or bile (16). Drug interaction issues are nearly always at phase I, and are commonly due to P450s. Phase I reactions are the rate limiting step in drug metabolism as phase II reactions are driven by a large excess of available conjugates. The phase I enzymes are mostly found in endoplasmic reticulum in the liver, and the main drug metabolising P450s are CYP1A2, the CYP3A family, CYP2C9, CYP2C19 and, CYP2D6 (17). The nomenclature of P450s is based on genetic relationship, and not on function, e.g. CYP2D6: CYP = cytochrome P450, 2 = gene family, D = gene subfamily, 6 = specific gene. A P450 family member (by current rules) indicates that the relevant P450s have ≥40% amino

acid sequence identity (18). Subfamily members share at least 55% amino acid sequence identity. Importantly, small changes in the structure of a drug can have major effects on its action. Also, small changes in the amino acid sequence of an enzyme can totally alter its activity. For example, the gastric proton pump inhibitor drug omeprazole is metabolised by CYP2C19, but not by the genetically similar CYP2C9 with which it shares 91% sequence identity, with only 42 different amino acids over 490 residues. A study showed that 16 of these residues were important for substrate selectivity (many distant from the active site). Mutation of these residues to those found in CYP2C19 generated a variant capable of metabolising CYP2C19 substrates such as omeprazole (19,20).

1.1.5 Regulatory guidance

The International Conference on Harmonization of Technical Requirements for Registration of Pharmaceuticals for Human Use (ICH) set guidelines for metabolite identification. These have been agreed by the US Food and Drug Administration (FDA), European Medicines Agency (EMEA) and the Japanese ministry of health, labour and welfare. Though there are no harmonized tests for metabolites, it is agreed that at stage III trials all metabolites require identification and testing as set out in the guidelines. However, it is noted that many companies set out to identify metabolites before this, as failure of a late stage compound due to metabolite toxicity/reactivity is much more costly than early identification and possible reformulation (21).

The accepted guidance states that metabolites should be tested when:

1. They are found only in human and not in animal subjects 2. They are found in large amounts (greater than 10% of the administered drug) 3. When the pharmacological and toxicological activities of the metabolites is considered significant (22).

The US FDA guidance is very similar and states that metabolites, especially those that differ from animal to human models, should be identified as early as possible in the drug development cycle (23). Metabolites that are a safety concern are those formed at greater than 10% of the drug systemic exposure at steady state. The choice of a level of greater than 10% for characterization of drug metabolites reflects consistency with FDA and Environmental Protection Agency guidance (24).

1.1.6 Reasons for regulation

The focus of drug companies and regulators on the effects of drug metabolites follows multiple instances of toxicity and efficacy relating to their production. Examples include drugs seen as safe, such as acetaminophen (paracetamol), which causes liver necrosis if the dose is high enough (25). The antihistamine terfenadine and its carboxylic acid metabolite fexofenadine are an example of the advantages of a metabolite over the parent drug. Fexofenadine retains all the antihistamine properties of terfenadine but is a much less potent potassium channel blocker, and therefore far fewer adverse drug reactions are associated with its use (26). There also is the problem that many drugs are prodrugs, where the prodrug relies on activation by metabolising enzyme(s) to form the relevant active species. An example is the activation of codeine to morphine by CYP2D6. Co-administration of a drug or candidate that is an inhibitor or inducer of these activating enzymes can also have an adverse effect, such as reduced pain relief from codeine when it is administered along with the 2D6 inhibitor quinidine (27). Food/drug interactions are also possible, For example, the bioavailability of felodipine is increased when a patient drinks grapefruit juice, as grapefruit contains a number of inhibitors, including bergamottin and furanocoumarins, that inhibit CYP3A4 activity, preventing metabolism of felodipine and other drugs (28). Medical regulatory agencies constantly watch out for new interactions, but early identification of metabolites and their effect on patients and interaction with current drugs reduces risk.

1.1.7 Genetics

Genetic factors have a prominent role to play in drug metabolism, as many P450 enzymes have polymorphisms. In addition, different populations may have either multiple copies of a gene, or none at all. CYP2D6, CYP2C19 and CYP2C9 all exhibit polymorphisms.

CYP2D6 is absent in 7% of Caucasians and 1-2% of non-Caucasians, yet is present in multiple genetic copies (up to 13) in east Africans (29). It is this population of ultra-rapid metabolisers that is problematic to doctors, as drugs such as beta-blockers and narcotic analgesics such as codeine, metabolised by 2D6, get a poor therapeutic response (30). CYP2C19 is absent in 20-30% of Asians and 3-5% of Caucasians. The absence in such a high proportion of Asians is important as they metabolise omeprazole and diazepam poorly (31,32). The dose of omeprazole is recommended to be reduced in patients where CYP2C19 is absent (33), though this means that Asians have a higher cure rate for H. pylori

infection (34,35). Omeprazole also acts as an inhibitor of CYP2C19, causing an interaction with diazepam that leads to its slower metabolism and increased plasma concentration (36).

CYP2C9 is absent in 1% of Caucasians and African Americans, and is the P450 that metabolises most Non-Steroidal Anti-inflammatory Drugs (NSAIDs) including cyclooxygenase-2 (COX-2) inhibitors, and S-warfarin (the active isomer). Almost all variability of warfarin treatment is explained by differences in CYP2C9 activity (37). The azole drug fluconazole abolishes CYP2C9 activity at conventional doses and increases warfarin blood levels by twofold (38).

1.1.8 Current Methods of Metabolite Synthesis

Most metabolites are currently produced using conventional synthesis methods. As with most fine chemicals, these molecules offer the greatest benefit from enzymatic synthesis, due to the small quantity required and the relatively high cost - often $100s/g, as compared to $1s/g for bulk chemicals (39). For small MS/MS identification and CYP inhibition studies, the use of recombinant human P450s has been a success for companies such as Cypex, though when larger quantities are required for NMR or other high concentration studies, alternative lower cost pathways (such as bioreactors) become preferable (40,41). A step toward this was made by Codexis with a panel of Bacillus megaterium P450 BM3 (CYP102A1, BM3) and other P450 variants and chimeras designed to produce a range of metabolites from a single drug candidate (42). When considering the advantages conferred by enzymatic synthesis, there are many, including simplification of reaction pathways, removing the need for protecting groups due to selectivity of enzymes, lowering the temperatures and pressures required, and reducing the solvent content of reactions. These factors combine to make this synthesis route very environmentally beneficial (43). The economics of large scale production have been the biggest constraint to widespread use of enzymes. Although BASF (enantiomerically pure amines) (44) and DSM (aspartame) (45) both produce chemicals this way, they largely rely on immobilised enzymes not requiring cofactors and having a high solvent tolerance. The expense of producing compounds biologically largely falls on the production costs of the enzymes themselves (up to 40% of total overheads), though a 10% total enzyme cost is allowable for economic pharmaceutical or fine chemical synthesis (46). This is where BM3 confers an advantage over other P450s, as the partner reductase is on the same polypeptide chain, reducing the need for separate enzyme production. In addition, the ease of BM3 production in high yields confers significant economic advantages.

Cytochromes P450s

1.2.1 Overview

As noted in the drug metabolites section, the primary drug and xenobiotic metabolising enzymes are the cytochromes P450. These are microsomal enzymes primarily found in the mammalian liver, although also found in most tissues and across all major life forms (18). Professor David Nelson runs the P450 nomenclature website that currently has over 20,000 named P450s, ranging over plants, animals, bacteria and fungi, and indicating their ubiquity and importance in all forms of life (47). P450s are a class of redox enzyme that catalyses the scission of molecular oxygen and the insertion of one atom of oxygen into an un-activated bond. The first discovered oxygenase enzyme was pyrocatechase by Osamu Hayaishi, which was shown to oxidatively cleave catechol using molecular oxygen (48), rather than by previously described mechanisms involving dehydrogenation. (49) There are many natural activators of molecular oxygen, a large number of which are metalloenzymes such as nitric oxide synthase, lipoxygenase and dioxygenases (50).

Cytochromes P450 are characterised by the presence of a heme cofactor with a cysteine thiolate as the proximal ligand. When the heme is reduced in the presence of CO this gives a characteristic major (Soret) heme absorption peak at ~450 nm (51). P450s have 5 key spectral UV-Vis features the alpha (α), beta (β), gamma (γ, the Soret) and, delta (δ) bands, and the protein signal. In the ferric, low spin state, the P450 α and β bands (together referred to as the Q-bands) are seen at 500-600 nm, with the Soret at ~418 nm, the δ band at ~360 nm and the protein signal at ~280 nm (mainly from aromatic amino acids). Upon substrate binding and a high spin shift in heme iron spin-state equilibrium, the Q-band signals are altered, and the Soret shifts towards 390 nm. Upon inhibitor binding directly to the heme iron (e.g. by a coordinating nitrogen, as would be the case for inhibition by an azole drug) the α and β peaks are again altered and the Soret band shifts to ~424 nm. This behaviour allows for relatively simple analysis of the heme ligation and spin state in aqueous solution. Since their discovery, P450 enzymes have been implicated in a number of important biological reactions, most notably the metabolism of drugs and xenobiotics in human liver microsomes, but also in the synthesis of steroid hormones and vitamins (51). The biotechnological interest in P450s arises from their ability to activate otherwise inert chemical bonds, often in a stereo- and enantio-selective manner. Human P450s are notoriously difficult to work with due to problems with insolubility and membrane bound domains (52,53). Consequently, this led to the search for bacterial models, as bacterial P450s are cytoplasmic as opposed to

memberane-associated and are therefore soluble and much easier to manipulate. The ease with which bacterial P450s are over-expressed, purified and crystallised, and the fact that their tertiary architecture is similar to eukaryotic P450s has led to them being used as models for all P450 systems, with the camphor hydroxylase P450cam (CYP101A1) and the fatty acid hydroxylase P450 BM3 being the enzymes of choice (54).

1.2.2 Physiological roles

Cytochrome P450s are implicated in a plethora of biological processes, both biosynthetic and degradative.

The catalytic mechanism is discussed in detail later, but simply put the enzymes utilise an iron ligated heme cofactor to reductively split molecular oxygen, using two electrons (from NADPH) and two protons from active site solvent to facilitate insertion of one oxygen atom into the substrate, according to the following scheme.

Figure 1.1. Simplified reaction pathway for cytochromes P450. The scheme shows a typical mono-oxygenation reaction (hydroxylation). RH denotes an organic substrate.

P450s can be classified by their reaction types, as follows (55,56):

1. Carbon hydroxylation – alcohol formation 2. Heteroatom release – by hydroxylation of an adjacent carbon and then collapse of an unstable intermediate 3. Heteroatom oxygenation – e.g. forming N and S oxides 4. Epoxidation – addition of oxygen across an unsaturated C-C bond 5. Oxidative group transfer – e.g. 1,2 C shift to form a carbonyl at C1 6. Alkyne/Alkene suicide destruction – inactivation of the heme of the P450 by the enzymatic product

Figure 1.2. Common P450 reaction types. Reactions shown are aromatic hydroxylation (1), heteroatom release (2), epoxide formation (3), 1,2 group transfer (4) and suicide destruction (of the P450 heme) (5).

Other more complex reactions have been shown to be possible by P450s, such as nitration, cyclopropanation and N-oxide reduction, though these are largely outliers rather than major classes of P450 reactions (57-59).

P450s have a wide range of natural substrates, someP450s being specific to a single or small group of substrates, whereas other P450s, for example from human liver microsomes,

accept a diverse range of substrates (51). These substrates tend to be hydrophobic molecules which have poor water solubility, and for which e.g. hydroxylation or epoxidation leads to more soluble products. The major functions of the liver are to store energy and to metabolise fatty acids (many P450s are very effective fatty acid hydroxylases). For this reason they have high lipid contents, meaning they attract lipophilic compounds at higher than blood concentrations. As drugs are often designed to be lipophilic in order to cross cell membranes, they also accumulate in the liver in high concentrations. Hydrophilic compounds tend to have rapid kidney clearance and behave like phase I metabolites. Just a single human P450 (CYP3A4) can metabolise about 50% of all pharmaceuticals (60), which shows the plastic nature of human P450s which have evolved to act on myriad xenobiotics entering the body. Cytochromes P450 have been a highly active area of research due to their involvement in such varied reaction pathways as drug metabolism, xenobiotic oxygenation, steroid synthesis, lipid signalling mediator synthesis, polyketide antibiotic synthesis, and various applications in biotechnology (15,61-65). Plants tend to have more P450 genes than animals, which they often use for the generation of colored compounds and defensive compounds (66). In mammals, P450s are widely involved in synthesis of natural products, such as steroids and lipid signalling molecules, and they have also recently been shown to be involved in metabolism of psychoactive drugs in the brain (67). The drug metabolising CYPs are small in number but their importance in terms of research and development of new drugs is paramount. Rapid metabolism of a compound by P450s can quickly reduce its bioavailability, which often leads to the failure of the compound in drug trials. Another problem can be activation of a drug by a P450 to produce a compound that binds irreversibly to, or otherwise interacts with, a particular macromolecule to cause toxic effects. For instance polyaromatic hydrocarbons (PAHs) such as benzo[a]pyrene are activated by CYP1A1 and CYP1B1 to produce potent DNA intercalators (68). Even suicide inhibition of CYPs is seen, which may cause an autoimmune response to the inactivated CYP and can also increase the concentration of other drugs metabolised by the same enzyme in a so called drug-drug interaction (69,70).

Research into the human drug metabolising P450s has found they often have a broad selectivity, but can accept compounds based mostly on size, shape and polarity (71). The success (or failure) of lead drug candidates can often be attributed to metabolic instability and toxicity, and therefore drug companies invest in simulation software such as Simcyp which combines in vivo data with modelling algorithms to simulate the metabolism of lead compounds by major human CYPs and to look at possible drug-drug interactions early in development (72).

1.2.3 Heme proteins

Heme proteins have long been studied for their vital roles in oxygen binding and transport. Though they are a large class of enzymes, most fall into the oxygen sensing or metabolism roles. Heme proteins contain a heme cofactor which consists of an iron atom at its centre bound to a protoporphyrin ring by four nitrogen ligands. There are many variations based on the porphyrin structure, the most common are shown in Figure 1.3. Cytochromes P450 along with hemoglobin and myoglobin all contain the heme b cofactor (Figure 1.3), others include the covalently bound derivatives of heme b, such as hemes l and m from peroxidase enzymes. In hemoglobin and myoglobin the iron is ligated to a histidine nitrogen, though in P450s it is ligated to a proximal cysteine thiolate (73). The final axial ligand is provided by either water in the substrate-free form of P450s or by dioxygen in substrate-bound enzymes, once reduced to the ferrous state (73).

Figure 1.3. The major forms of heme prosthetic groups. Heme a is found in cytochrome c oxidase, while c-type hemes are found in respiratory chains and are covalently linked to the protein backbone via the two thiol groups shown, that are usually cysteine residues donated by the protein. The heme b cofactor is found in cytochromes P450 and hemoglobin/myoglobin. All hemes comprise a protoporphyrin ring ligated to an iron atom at the centre. The P450s have a conserved proximal cysteine thiolate ligand.

1.2.4 The P450 catalytic cycle

The catalytic cycle of P450 enzymes has been studied in detail and a model catalytic cycle proposed with many intermediates characterised (74). Due to the short lived nature of the highly reactive iron-oxo compounds I and 0, they have proved exceptionally difficult to characterise, with compound I (the iron-oxo intermediate considered to be the major active molecule in P450 catalysis) not being definitively identified until Rittle and Green’s study in 2010 (75). Figure 1.4 provides a model of the cycle as an 8 step process. Step 1 – Substrate binds to the enzyme displacing the distal water ligand to the heme iron and resulting in a low to high spin shift (S=1/2 to S=5/2) in the ferric iron (FeIII). Step 2 – The iron centre is reduced from ferric (FeIII) to ferrous (FeII) with an electron delivered from a redox partner (usually derived from NAD(P)H). This is highly dependent on the presence of substrate as binding causes an increase in the redox potential of the heme which drives electron donation (76). Step 3 – Molecular oxygen then binds to the heme iron to form a nucleophilic ferrous Fe(II)- oxo intermediate (isoelectronic with the ferric-superoxo form). Step 4 – Next a ferric (FeIII) peroxy) complex is formed by delivery of a second electron from the redox partner. In this two electron reduced species one electron is on the oxygen while to other is delocalised over the cysteine ligand (77). Step 5 – Protonation of molecular oxygen forms the ferric (FeIII hydroperoxo) intermediate compound 0. (78) Step 6 – A second protonation of oxygen causes scission of oxygen and loss of water to produce the ferryl (FeIV)-oxo complex compound I (75). Step 7 – Compound I is highly reactive and is thought to abstract hydrogen from the substrate before the transient species reacts with the substrate radical to form the oxygenated product, the so called radical rebound mechanism (75,79). Step 8 – Water displaces the product reforming the resting form of the enzyme in its low spin ferric (FeIII) resting state.

Figure 1.4. The P450 catalytic cycle. The proposed P450 catalytic cycle (54) is shown with steps numbered, as explained in the text. The cysteine ligated iron porphyrin is shown as a red circle with Fe and S as the iron and sulfur. Substrate is shown in blue and is denoted as RH, while the product is R-OH, comprising one atom of molecular oxygen. The uncoupled pathways are also shown, and explained below. Key species in the cycle are 6. Ferric (FeIII hydroperoxo) intermediate compound 0, and 7. Ferryl (FeIV)-oxo complex compound I,

1.2.5 Active species in the P450 catalytic cycle

The accepted mechanism of substrate oxidation was first postulated by Groves in 1978 (Figure 1.5), and is known as the radical oxygen rebound mechanism (79). This proceeds with the highly reactive compound I attacking the un-activated CH bond, abstracting hydrogen to leave a substrate radical and forming the ferryl hydroxyl species compound II. Compound II is subsequently attacked by the substrate radical forming the hydroxylated

product and ferric heme (75,79-81). Further computational studies have indicated two transition state pathways, one low-spin doublet and the high spin quartet state which is the stable radical state (82). The large kinetic isotope effect seen in compound I substrate hydroxylation is further proof of a hydrogen abstraction model of the reaction (75).

Figure 1.5. The P450 radical rebound mechanism. Compound I (left) is shown with the porphyrin described as a circle and Fe bound to thiolate sulfur below, and to oxygen above the plane of the heme. The heme iron oxidation state shown is in Roman numerals, and the organic substrate RH is shown in blue. Compound II (centre) follows a similar notation, with the substrate radical shown in blue. Following completion of the reaction, the ferric heme (right) is shown with the hydroxylated organic product (ROH) in blue/black. The reaction is described in detail in the accompanying text.

There has been some controversy about the formation of compound I and its seemingly transient nature. This is largely due to the long wait for confirmation of the compound with many groups applying differing techniques and therefore submitting competing claims to have identified the species. The compound I oxygen rebound mechanism was supported by the mechanism of horseradish peroxidase which was structurally proven to have an FeIV-oxo intermediate by a multi-crystal analysis technique that allowed elucidation of the entire catalytic pathway by electron induced reduction of the sample followed by microspectrometry (83). Also, the closest relative of P450, chloroperoxidase (CPO), which has a proximal cysteine thiolate ligand just as P450s, has been extensively studied by EPR, Endor and Mössbaur. When freeze quenched with peroxyacetic acid an FeIV-OH species is seen with a spin = 1 and spin ½ radical on the porphyrin (84-86). These provided good standards with which to compare P450 compound I data. Thinking that very complex capture methods would be required, researchers focused on using fluorinated compounds to slow the rate of hydroxylation of substrate in the hope that compound I would build up. Flash quenching was used to try and produce high concentrations of compound I by supplying a single electron to the FeIII-OH compound II, and the resulting species was reported as compound I (87,88).

Subsequently, the spectra were suggested as being more likely to result from nitrosyl complexes (89,90). Rittle and Green required improved enzyme purification for their definitive studies on compound I with the CYP119A1 enzyme from the thermophilic archaeon Sulfolobus solfataricus, as they found that small molecules (C12-C18 fatty acids) were bound to active site of enzyme and reacted with compound I before it could accumulate. Once highly purified and reacted with peroxide donor compounds, they were able to trap compound I and provide convincing spectroscopic evidence for its formation.

There is evidence that not all P450 reactions proceed via the compound I mechanism shown above (Figure 1.5), such as the sterol 14α-demethylase (CYP51) which instead likely uses the preceding species in the cycle, the FeII-peroxo compound 0, for the last of three oxidations in the oxidative demethylation of sterols, e.g. in the final deformulation reaction on a doubly hydroxylated derivative of lanosterol catalysed by fungal CYP51 (producing formate). This is due to the substrate acting a barrier to delivery of the second proton, therefore making the compound I reaction energetically unfavourable (91). Aldehyde formation from cyclohexane carboxaldehyde to produce formic acid and cyclohexene is catalysed by mammalian CYP2B4, and this is thought to proceed via a peroxygenase route (92,93). Work on the stereochemical outcomes of BM3 sulfoxidation of thioethers led to a postulated mechanism of direct oxygen transfer via compound 0, which was suggested may also be involved in amine oxidation (94). Horseradish peroxidase (HRP) and chloroperoxidase (CPO) both utilise peroxide for oxidation of organic compounds (95). It has since been discovered that some P450s do utilise peroxide as oxidants in their catalytic cycle, such as OleT from a Jeotgalicoccus species, which is a fatty acid decarboxylase and terminal olefin synthesising P450 (96). Other peroxide dependent P450s are P450SPα from

Sphingomonas paucimobilis and P450BSβ (CYP152A1) from Bacillus subtilis both fatty acid hydroxylases (97,98). It has been shown that many P450s (including BM3) and P450cam can act as peroxygenases and generate oxidised products, although this is typically inefficient in cases where the P450s have not evolved specifically for this function. This peroxide shunt pathway (Figure 1.4) can be driven by hydrogen peroxide or by other organic peroxides. The main problem with the pathway is that oxidative destruction of the heme, caused by non-productive interactions with the peroxide, can occur. This vastly reduces the total turnover numbers capable of the P450 enzymes due to their inactivation (95,99).

1.2.6 The P450 heme iron spin state

As shown in the catalytic cycle, the first step involves the displacement of water as the 6th ligand to the heme iron. This substrate induced displacement elevates the heme iron potential and enables the transfer of an electron to the heme iron (from a redox partner protein), converting it to the ferrous state (FeII). As water is a weak ligand for this state, the iron becomes five coordinate and the d-orbital configuration of the heme iron undergoes rearrangement from a low spin to a high spin state (Figure 1.6). This shift and the associated changes in optical, EPR and Mössbauer spectroscopic properties offer detailed characterisations of the heme spin state, allowing analysis of substrate- or conformationally- induced spin state change (100,101). The spin state shift is a thermodynamic equilibrium which can be affected by both the extent of substrate saturation of the active site, and by temperature (76).

Figure 1.6. The ferric heme iron d-orbital electron state in low spin and high spin configurations. In the low spin state all electrons are in the t2g orbital and the overall spin =

½. In the high spin state 3 electrons are in t2g and 2 in eg, with overall spin = 5/2. The energy difference between the orbitals is denoted Δoct.

The cytochrome P450 heme differs from that in related proteins such as myoglobin, as it has a much lower redox potential (e.g. the P450cam redox potential in the substrate-free form is ~-300 mV vs the standard hydrogen electrode (SHE) compared to myoglobin at ~60 mV)(102,103), due to the cysteine thiolate having a “push” effect that increases electron density at the porphyrin (104). This “push” effect also causes the ferric hydroperoxo species (6) to undergo protonation preferentially over reduction.

1.2.7 Cytochrome P450 Structures

In the past few years many crystal structures of P450 enzymes have been solved (105-107). These data have shown that, although the amino acid sequence is varied across the superfamily, the overall fold and topology of the enzyme is highly conserved. The P450 fold is unique to these enzymes and is not shared with any non-P450 structures, and so is likely evolved for the particular oxygen activation chemistry that these enzymes provide (108). The presence of a cysteine thiolate, essential for catalytic activity, is also conserved across the superfamily (109).

Figure 1.7. General P450 structure. A typical P450 structure is shown – the heme (P450) domain of Bacillus megaterium flavocytochrome P450 BM3 (CYP102A1, PDB code 1JPZ) (110) with α-helices represented as tubes (green), β-sheets as flattened arrows (green) and the heme cofactor in the centre (red). Helices are labelled A-L following the standard nomenclature (111). The substrate NPG (N-palmitoylglycine) is shown in black.

The P450 BM3 structure (Figure 1.7) shows the general topology of P450 enzymes. Although the amino acid sequences differ greatly across different structures, this overall fold is conserved, so allowing a standard nomenclature across the family (111). As can be seen, the structure consists of two domains. The first (on the right) is the β-domain as it is

composed mostly of β-sheets, and the second larger domain (above and to the left of the heme cofactor) is the α-domain as it is comprised mostly of α-helices. The active site region is crossed by the distal I helix which makes many of the substrate contacts, and by the proximal L helix (112). The active site heme is well buried in the protein and not exposed to the bulk solvent (111). The most highly conserved regions of the structure are centred around the heme cofactor, as this is responsible for the oxygen activation (108). All P450s contain the conserved cysteine thiolate ligand bound proximal to the heme. The heme is surrounded by peptide NH groups which provide good H-bond interactions and which increase the redox potential of the system by reducing the negative charge on the sulfur (112). This conserved cysteine and peptide bonding arrangement is also present in other similar monooxygenase proteins such as nitric oxide synthase (NOS) and chloroperoxidase (CPO) (113,114). There is a conserved Phe (Phe 393 in BM3) seven residues from the cysteine which acts to modulate the redox potential of the heme iron, mutagenesis of which increased electron transfer from CPR and reduced rates of fatty acid hydroxylation, due to stabilisation of FeIII-oxo intermediate (115,116). Another conserved region is the I-helix, the portion of this helix that crosses the heme pocket is kinked due to a rearrangement in the helix H-bonding structure. The near-completely conserved Thr or Ser residue (Thr252 in P450cam) is essential in the H-bond network for proton delivery to the heme for oxygen scission. Upon substrate binding, an ordered water molecule enters the active site, binding to the hydroxyl of the Thr/Ser, which is then prevented from hydrogen bonding to the I-helix residues, leading to the opening and kinking of the I-helix (117). The structure around the Thr/Ser and its adjacent acidic amino acid (a Thr/Asp pair in P450cam) are conformationally altered, causing a flip in the aspartate carbonyl oxygen and generating a hydrogen bonding network toward the heme iron for efficient delivery of protons in the cycle. The release of water required for catalysis is associated with the kinking of the I-helix, suggesting that rearrangements that occur on displacement of the axial water are responsible for this event. This kinking phenomenon is thought to allow more efficient proton transfer from the bulk solvent to the heme, by opening space to allow in new water molecules upon oxygen binding (118).

1.2.8 Redox partners

All P450 oxygenases require two electrons to catalyse the scission of molecular oxygen, and these are almost exclusively provided by NAD(P)H and transferred with the aid of protein- bound cofactors (54,119). Class I P450s are mainly prokaryotic enzymes and deliver electrons from NAD(P)H via ferredoxins (2Fe-2S/3Fe-4S and occasionally 4Fe-4S) or

flavodoxins (FMN-binding proteins) from a FAD cofactor-containing reductase (54). Class II P450s are generally eukaryotic membrane bound enzymes using a cytochrome P450 reductase enzyme (CPR), which delivers electrons from NADPH via FAD then FMN cofactors bound in the CPR (54).

Figure 1.8. Structures of the P450 reductase enzyme cofactors. Flavin adenine dinucleotide (FAD, left); and flavin mononucleotide (FMN, right). A 2Fe-2S cluster as found in certain bacterial ferredoxins (e.g. putidaredoxin, the P450cam partner) is shown in the centre.

More recently, a class of prokaryotic (and lower eukaryotic) P450s with a heme (P450) domain and CPR fused has been characterised (Type III, Figure 1.9), and this arrangement typically gives high catalytic efficiency (120). These enzymes have possibly evolved due to the advantage in accelerating the CPR-to-P450 electron transfer rate, and so catalytic efficiency, provided by the close proximity of the heme and reductase domains (74). P450 BM3 (CYP102A1, BM3) from Bacillus megaterium was the first discovered P450 fusion enzyme (121) and will be discussed in detail later. Numerous prokaryotic and eukaryotic redox fusion enzymes have now been identified (74). P450foxy, a fatty acid hydroxylase from Fusarium oxysporum (a fungal pathogen of rice), is another P450-CPR fusion protein (122-124). The fatty acid hydroxylases CYP102A2 and CYP102A3 from Bacillus subtilis were also found to be BM3 homologues and fatty acid hydroxylases (125). Data suggest that

redox partner interactions may do more than just transfer electrons, and their interaction can also mediate catalysis. This is seen with binding of P450cam to its putidaredoxin redox partner, which causes a shift to the “open” form (with respect to active site access) and rearranges the I-helix, allowing H-bond mediated proton-coupled electron transfer to take place (126).

Figure 1.9. Schematic representation of the 3 major P450 classes. The type I bacterial, Type II mammalian, and Type III redox partner fusion systems are shown in a schematic diagram. The P450 is represented in the resting (ferric) water bound state, with a central Fe atom and with the porphyrin ring as a red circle, and with the cysteinate ligand as S. Redox partners are represented in dark blue boxes with their labelled cofactors shown in orange as either a triple ring (illustrating the isoalloxazine ring of flavins) or a cube (illustrating an iron- sulfur cluster). The relevant cellular membrane is shown as a brown box. The diagram was created in Adobe Illustrator.

1.2.9 P450 electron transfer reactions

Pivotal to the electron transfer processes that drive the P450 reaction cycle are flavoproteins. The flavins in the schemes below (Figures 1.11-1.13) are versatile redox molecules, being able to exist in three oxidation states, neutral, semiquinone and hydroquinone. The semi- and hydroquinone forms can then exist in neutral and anionic states, dependent on the protonation state of the flavin (127). This makes them very effective single electron donors of the sort required for P450 redox processes. The existence of these states allows for multiple forms of spectroscopic analysis, e.g. by UV-visible spectroscopy, fluorescence and EPR (128-130).

Figure 1.10. Diagrammatic representation of flavin oxidation states. The atoms are numbered in purple on the isoalloxazine ring, and the standard reduction potential for each electron addition is shown, along with the pKa for the interconversion of the neutral and negative semiquinone and hydroquinone forms (for FMN). Adapted from (127).

As can be seen from the P450 catalytic cycle (Figure 1.4) the two single electron transfers must occur during the activation of oxygen. The electron transfer process differs for each of the systems mentioned above (Figure 1.9).

Class I systems, of which P450cam from Pseudomonas putida is the model system, consist of a relay system shuttling two electrons from NAD(P)H to putidaredoxin reductase (Pdr) to form a two electron reduced flavin adenine dinucleotide (FAD) hydroquinone. Single electrons are then transferred to the Fe2S2 centre of putidaredoxin (Pdx), which in turn binds to the P450 and provides the first and second reducing electrons in the cycle in two consecutive steps (Figure 1.11). The protein doesn’t just act as an electron transfer protein, but also mediates structural rearrangement of the P450 to allow the coupled proton- mediated electron transfer to take place (126). Electron transfer between NAD(P)H and Pdr is fast, with the Pdx to P450 electron transfer steps being rate limiting in catalysis. No FAD semiquinone is seen during the reaction mechanism and the Pdr is readily re-reduced by NADPH as single electrons are transferred to Pdx (131). Semiquinone formation is, however, possible by laser flash photolysis using 5-deazariboflavin semiquinone, and a blue neutral FAD semiquinone is then seen on Pdr. This is unstable and rapidly forms either the hydroquinone or the oxidised FAD, indicating it is likely a transient species in the mechanism (132).

Figure 1.11. The electron transfer process in type I P450 systems. An example is shown for the Pseudomonas putida P450cam. Figure adapted from (133).

The eukaryotic microsomal (Type II) systems contain a diflavin reductase (CPR) system which is highly conserved and essential for life (134). This is a two domain protein consisting of a NADPH and flavin adenine dinucleotide (FAD) binding domain (itself likely originating from evolutionary fusion of two smaller domaimns), and a flavin mononucleotide (FMN) binding domain. These are joined by a peptide linker region which is essential for catalytic activity, and in aiding docking to the P450. The electron transfer proceeds from NADPH to FAD to FMN to P450 (Figure 1.12). The process is controlled by Trp676 in human CPR, which sits over the re-face of the FAD, preventing NADPH hydride transfer until a structural rearrangement occurs to displace the tryptophan side chain. Reaction transients from NADPH-dependent hydride transfer reductions can be followed by fluorescence spectrometry of the tryptophan – and these likely follow both the displacement and return of the Trp676 side chain during the CPR reduction process (135). The electron transfer reaction is controlled by the redox potentials of the various flavin species in the system. Interestingly, the initial NADPH to FAD transfer is moderately thermodynamically unfavourable in some enzymes, although at equilibrium the partial reduction of the CPR FAD would occur (108). Reduction of the FAD by NADPH is a multi-step process (with a fast initial reduction of CPR, and a slower second reduction process on CPR that already bears electron). Charge transfer complexes are formed formed between the oxidised FAD and + NADPH (or reduced FADH2 and NADP ) until the two hydride transfers are complete, with the process gated by NADP+ release (108). The single electron transfer from CPR to P450 is moderated by the protein stabilising the neutral blue semiquinone forms (FADH• and FMNH•) (136). Initially on reduction by an equivalent of NADPH, the FAD is in a two electron reduced state. Eukaryotic CPR enzymes are considered to exist naturally in a stable blue semiquinone state (on the FMN cofactor). Thus, FAD reduction converts CPR to a 3-electron reduced state. The FMN hydroquinone is considered to be the relevant electron donor to the heme iron. Thus, electron transfer from FAD to FMN occurs, producing a FAD semiquinone (SQ)/FMN hydroquinone (HQ) state. The FMN HQ reduces the P450, and is then regenerated by electron transfer from the FAD SQ – producing an FAD oxidised/FMN HQ form. The FMN HQ then passes a second electron to the P450, so returning to the resting state in a 1-3-2-1 cycle, where the numbers indicate the total number of electron on the CPR flavins at different steps in the reaction (108). The formation of individual species and the electron transfer can be followed spectroscopically as they have distinct absorbances. Fully oxidised rat CPR has flavin absorbance peaks at approximately 455 nm, 382nm, with a shoulder at 484 nm, formation of a blue semiquinone leads to a broad peak at 587 nm with a shoulder at 630 nm (137). The hydroquinone forms of the flavins are largely featureless in the range from ~300 nm upwards.

Figure 1.12. Electron transfer in type II P450 redox systems that use CPR. The figure is adapted from (138). The reactions shown are typical of those catalysed by mammalian CPR enzymes. It should be noted that mammalian CPRs likely occupy a stable FMN semiquinone state in vivo. In this case, reduction would start from an FMN radical form, with NADPH reducing the FAD to FADH2, and then electron transfer from FAD to FMN producing FMNH2, which donates the first electron to the P450 heme iron. A second electron transfer from the

FAD radical then regenerates FMNH2. Finally, FMNH2 provdes the second electron to the heme iron to complete the catalytic cycle and restore the FMN radical “resting state”.

The P450 BM3 reductase is also a diflavin reductase system, though this time linked directly to the P450 domain via a peptide linker. The catalytic process is different to the major eukaryotic CPRs, with the major exceptions being that the FMN neutral semiquinone is now disfavoured and a red anionic semiquinone forms transiently instead in BM3 during catalysis, and in that the resting state of the BM3 reductase is the fully oxidised form. Therefore the electron transfer progresses from a single equivalent of NADPH to the FAD to form the hydroquinone, with an internal transfer to the FMN then forming a blue FAD semiquinone and a red FMN semiquinone. The red SQ is unstable, and can be further reduced to the HQ (by electron transfer from the FAD SQ). However, the FMN HQ has a more negative redox potential and is an inefficient reductant of the P450 heme. Thus, a rapid first electron transfer occurs from the anionic FMN SQ to the heme iron, with the FMN SQ then replenished by electron transfer from the FAD SQ. The anionic FMN SQ then transfers the second electron

to the P450 to enable completion of the catalytic cycle. Thus BM3 CPR undergoes a 0-2-1-0 electron cycle (108). The four electron reduced form of FAD and FMN hydroquinones is disfavoured as the redox potential is slightly more negative than that of the NADPH/NADP+ redox couple, although a proportion of this species is likely formed with excess NADPH (138). The non-catalytic pathway that leads to accumulation of the FMN HQ can occur if the P450 is not substrate bound or if NADPH is in excess, and explains the rapid drop in the rate of BM3 activity after the initial burst of NADPH oxidation in substrate limited conditions.

Figure 1.13. P450 electron transfer in type III redox systems, as exemplified for the P450 BM3 enzymes. Figure adapted from reference (138).

1.2 P450 BM3 from Bacillus megaterium

1.3.1 Overview

Cytochrome P450 BM3 (CYP102A1) is a 119 kDa catalytically self-sufficient P450 from Bacillus megaterium, discovered by the Fulco lab at UCLA.(121) It was the first single polypeptide P450 shown to comprise both the heme and reductase (CPR) domains and thus only require NADPH, substrate and O2 for activity. The self sufficient nature of the protein along with its stability, solubility and ease of production led to it (along with P450cam) becoming a model P450 system, and it has been extensively studied by spectroscopic and structural techniques, including UV-Vis spectroscopy, NMR, resonance Raman, EPR, MCD and X-ray crystallography (139-144). The CYP102A subfamily it spawned now consists of over 50 proteins, though BM3 was the first and most effectively studied of the class. The reductase domain is fused to the C-terminus of the heme subunit. BM3 was shown to catalyse predominantly ω-2 hydroxylation of long chain saturated fatty acids and hydroxylation and epoxidation of unsaturated fatty acids, with C15-C16 substrate chain lengths being optimal. The hydroxylation to epoxidation ratio for unsaturated compounds was shown to be pH dependent (145). In its natural host, BM3 was shown to be induced by phenobarbital and a range of other barbiturates (121), though this form of induction for protein production is no longer required for high level expression once the CYP102A1 gene is cloned into a plasmid vector with a different inducible promoter system, and expressed in E. coli. (146,147). Other inducers are polyunsaturated fatty acids, though these are toxic to B. megaterium, making them potential physiological substrates associated with a detoxification role for BM3 (148,149). However, more plausible endogenous substrates are branched chain fatty acids, which make up the majority of B. megaterium’s natural fatty acid content (150). BM3 was an important discovery for the P450 field as it had an activity much greater than any P450 oxidase previously reported (121), due to both the heme and reductase domains being contained within the same polypeptide chain and inter-domain electron transport being efficient. BM3 consists of a 55 kDa heme domain and a 65 kDa reductase domain containing FAD and FMN cofactors (151). The reductase domain can be further dissected into two domains, a bacterial flavodoxin-like FAD domain, and a spinach ferredoxin reductase-like FMN-binding domain (152). This heme-reductase domain combination is more mammalian-like in construction than its bacterial counterparts, with the heme domains having structural and functional similarity to eukaryotic fatty acid hydroxylases of the CYP4A family (109). This means that BM3 provides a useful model for

mammalian (class II) redox systems, with the advantage of its being soluble and easily expressed in E. coli. (109). This fusion system confers another advantage to BM3 in that it has the highest activity of any known P450 oxidase with an activity of ~17,000 min-1 with arachidonic acid (153). This activity is so high due to the efficiency of electron transfer from NADPH to the heme via the reductase domain. The linker region is particularly important for this fast electron flow and activity, and if shortened or broken the activity is substantially reduced or lost near-completely (154,155). It has been shown that the active form of BM3 is a dimer with electrons transferred from the CPR FMN of one monomer to the heme of the other monomer (74). The Km of BM3 for fatty acids is similar to eukaryotic fatty acid hydroxylases (CYP4 enzymes), but its kcat is much larger, resulting in a higher kcat/Km ratio – which indicates a more efficient enzyme. The fast rate of flavin-to-heme electron transfer, 223 s-1 for the first electron transfer, is much higher than is observed in eukaryotic P450s and a major factor in the high catalytic rate of P450 BM3 (74). The single electron transfer properties results from the ability of the BM3 CPR to form semiquinones (as described in section 1.2.8), the red negative semiquinone FMN is formed in BM3 instead of the blue neutral form seen in eukaryotic CPR, and this was shown to be due to a missing glycine residue between tyrosine 536 and asparagine 537 in BM3, introduction of which partially restores eukaryotic CPR-like behaviour (156). Over-incubation of BM3 with NADPH causes over-reduction of its CPR, leading to formation of FMN hydroquinone – which is a thermodynamically less inefficient reductant of the P450 heme domain than is the anionic semiquinone form. Excess NADPH (in absence of a P450 substrate) also leads to futile cycling of electrons in both BM3 reductase and eukaryotic CPR – resulting in production of reactive oxygen species (superoxide and hydrogen peroxide) and eventually to inactivation of the enzymes (Figure 1.4) (157).

Extensive work has been carried out on BM3 to identify intermediate species in its catalytic process. To date, these relate mainly to states of the CPR domain, along with early steps in the iron-oxo species cycle. The separated heme and reductase domains were initially reported to catalyse the oxidation of fatty acids efficiently. However, since the dissected domain have limited affinity for one another, the fatty acid oxidation rate is very slow, and associated with extensive uncoupling of NADPH oxidation from fatty acid oxidation (158).

Both NMR and P450 crystal data show that fatty acids are too far from the heme iron for functional catalysis in the oxidized form of the enzyme. This may be due to the low temperatures used in these studies, but could also reflect the natural state of the enzyme, in that the substrate may be predominantly located in a thermodynamically more favourable binding mode (with its carboxylate group able to interact with the side chains of Arg47 and

Tyr51 near the mouth of the active site), but also able to migrate deeper into the active site and explore other binding modes (159,160). This may explain the lack of strict regioselectivity in the position of fatty acid hydroxylation, since BM3 typically oxidizes saturated fatty acids in the ω-1, ω-2 and ω-3 positions. Low spin to high spin proportions of the ferric heme iron were also shown to be affected by pressure (161), substrate concentration (162) and ionic strength (163), and excessively tight binding substrates may also be detrimental to catalysis, likely through low rates of product dissociation (149). The relationship of heme iron spin state to activity can be weak, with small changes in heme iron spin state (towards high spin) not always associated with low activity, despite the more positive potential of the heme iron in the high spin state, which should favour electron transfer from the reductase (164,165). The redox potential becomes considerably more oxidising in P450 BM3 on addition of substrate (by up to ~140 mV) and the first electron transfer rate constant from the reductase increases from ~5 s-1 to 226s-1 (166). The reactions follow Marcus theory, whereby the energy of reorganisation of solvent spheres, and not simply redox potential or transition states, governs activation energy (167).

1.3.2 P450 BM3 structure

Although no full length crystal structure of Intact P450 BM3 exists, the heme domain has been crystallised in both substrate-free and substrate-bound forms (168,169). In some cases the BM3 heme domain forms a crystallographic dimer (PDB files; 2HPD, 1JPZ, 1SMI), even though it appears almost completely monomeric in solution state. Distinct structures of the substrate-free and substrate-bound forms of the BM3 heme domain show two very different conformations, one ‘open’ and the other ‘closed’ with respect to active site access (170). As stated above, BM3 functions in solution as a dimer of the full length protein. There is some disagreement as to whether electron transfer is from FAD of one monomer to the FMN of the other, or from FMN of one dimer to the heme of the second, but there is consensus that the protein is a functional dimer (171-173). This may explain why two molecules of the heme domain occur in asymmetric units in certain crystal forms of the heme domain, although other factors (e.g. crystal packing effects) may be the main determinant here.

The conformational changes undergone by the BM3 heme domain have been analysed by molecular dynamics (MD) simulations (174). When atomic resolution structures are achieved, these also show that the heme porphyrin ring can be bound in two conformations related by a 180 degree “flip” of the prosthetic group (175). Separate structures of the FAD binding domain and the FMN binding domain of the reductase have also been solved,

meaning that structural data for all individual domains of BM3 are now available. However, a full length P450 BM3 structure has proven elusive to date (176,177). A truncated form of P450 BM3 with the heme domain fused to the FMN portion of the reductase (through the natural linker region) domains has also been crystallised and structurally resolved (178). This structure shows the FMN domain to be positioned near the proximal face of the heme, with the FMN cofactor orientated toward the cysteine ligand. However, the fact that the FMN domain is proteolytically separated from the heme domain in the crystal structure and present in a non-stoichiometric ratio (1:2, FMN:heme) (178) does raise questions as to the accuracy of an extended electron transfer pathway proposed to explain catalysis in the intact P450 BM3.

Structural and mutagenesis studies have also identified key residues involved in substrate binding and reactivity in BM3.

Figure 1.14 The crystal structure of the heme domain of P450 BM3 with N- palmitoylglycine (NPG) bound (PDB code 1PJZ) (168). The diagram was created in Pymol, with substrate recognition sequences (SRS) colored: SRS 1 blue, SRS 2/3 yellow, SRS 4 purple and SRS 5 orange. The SRS regions are peptide sequences identified by Gotoh as parts of the P450 lining the active site and affecting substrate recognition in all P450s (179).

The structure of the BM3 P450 domain resembles eukaryotic P450s, with the heme protein made up of α and β domains. The α domain contains four α helices (D, L and I being parallel and E anti-parallel) in a four helix bundle; and above this bundle are five α helices and three

310 helices, with three more α helices and two β sheets underneath (169). This α domain comprises 70 percent of the heme domain structure. The β domain is comprised of two β sheets inside three α helices and a 310 helix. This domain represents 22 percent of the overall structure (169). The heme is in the α domain between the I and L helices and shielded from the surrounding solvent. When substrate-free the heme-iron is 6-coordinate with the axial ligands being C400 and a water molecule (169). The substrate binding pocket is a long thin chamber 8-10 Å wide, bordered with non-aromatic hydrophobic residues (169). Upon substrate binding the water is displaced. This is an entropically favourable displacement, leaving the active site free of solvent so that efficient catalysis can occur (168). Substrate is locked in place over the active site by a conformational shift of the two domains closing over it. The entrance to the binding pocket contains a poorly defined Arg47 residue which is likely of catalytic importance in binding substrate carboxylate, and BM3 activity is reduced if Arg47 is mutated (169). Substrate binding causes a conformational shift in the I helix region, causing displacement of the water ligand to the heme iron, which results in a low to high spin shift in the iron d-electron configuration (168). Thr268 in the I helix has an important role in catalysis. It is located next to the heme distal face, and hydrogen bonding in the I helix is disturbed by the presence of a water molecule, causing the helix groove to expand, which leaves a kink common to all P450s (168). Upon substrate binding the I helix undergoes a shift from its kinked geometry to a straighter one, the kink is reduced from 13o to 5o, caused by displacement of a water molecule bound between Ile263-Glu267 and Ala264-Thr268, which has the knock on effect of displacing the water molecule that provides the 6th ligand to the heme iron (168). This involves the acid alcohol pair Glu267/Thr268, which are conserved residues across the oxidase P450s, and which participate in activation of P450s by facilitating protonation of iron-oxo intermediates in the catalytic cycle. Ala284 interacts with a water molecule, both when the water is heme iron- bound as the 6th ligand, and when it is displaced to the secondary site upon substrate binding. An A264E mutation was made with the aim of covalently linking the heme porphyrin to the protein, given that Ala264 is replaced by a glutamate in BM3-related eukaryotic CYP4 enzymes that undergo catalysis-dependent covalent linkage of a heme methyl group to the glutamate. However, spectroscopic and structural analysis of the BM3 A264E variant showed that the Glu264 carboxylate oxygen directly ligated the heme iron in place of the water 6th ligand in a proportion of the enzyme. Addition of substrate induced structural rearrangement and led to complete coordination of the heme iron by Glu264 (180,181). Phe393 is proximal to the porphyrin and its cysteinate ligand and controls heme iron

reduction potential, thus altering the heme reduction rate, as demonstrated in studies of F393A, F393W and other variants (182). A phenylalanine is conserved in this position in almost all P450 oxidases, but is absent in various P450s that perform reactions such as isomerisation, and thus do not directly activate oxygen.

A number of other important residues have been identified by rational mutagenesis of P450 BM3. Residue Phe87 has been shown to be an important active site amino acid, lying above the heme and near the binding position of the ω-terminus of the fatty acid substrate. Phe87 rotates upon substrate binding anchoring the terminal methyl group (153). Replacement of Phe87 with non-aromatic amino acids allows the substrate to move further into the binding pocket and so moves the position of hydroxylation further from the ω-terminus (183). A binding pocket by the active site can be filled by the mutations A82F/A82W, moving the ω terminus of the fatty acid further down the active site, which also increases the apparent substrate binding affinity by approximately 800-fold (184). These mutations increase the extent of low- to high-spin shift in the heme iron on fatty acid binding and also improve catalytic efficiency. With laurate substrate, A82F favoured ω-2 hydroxylation, but A82W proved unstable with low coupling efficiency of NADPH to substrate oxidation, and the enzyme appeared to become inactivated by enhanced production of ROS (184). Substrate specificity was also altered, as A82F and A82W variants were shown to produce indigo from the indole present in the growth medium (184). Residue Arg47 has been shown to move considerably on fatty acid binding (185) and also has a major effect on fatty acid binding (109), with the double variant R47L/Y51F showing increased ability to oxidise polycyclic aromatic hydrocarbons (186). As also discussed above, Phe393 is almost completely conserved across the P450 family, and its aromatic side chain stacks with the iron cysteine bond, in this way controlling heme iron reduction potential in order to help regulate electron transfer from the reductase partner (187,188).

In the reductase domain, Gly570, Tyr536 and Trp574 are important residues for FMN binding, with Trp574 also possibly involved in reductase to heme electron transfer (189). In the FAD binding site, Trp1046 regulates access of the nicotinamide group of NADPH to the FAD isoalloxazine ring for electron transfer. The W1046A mutation resulted in the enzyme discriminating only 1.5-fold in favour of NADPH over NADH, compared to 8571-fold in the wild-type reductase (190). As also mentioned above, Thr268 in the heme domain regulates the binding of oxygen and its activation, with a T268A mutation greatly reducing substrate hydroxylation, likely through direct effects on protonation steps in the cycle (191).

1.4 Biotechnology

The biotechnology industry only accounts for a small proportion of chemical and pharmaceuticals synthesised, but its growth far outstrips that of the industry as a whole. This growth is driven partly by the desire to reduce the use of volatile and often toxic chemicals. However, most of the drive is economic, and reactions performed by enzymes at ambient temperature and atmospheric pressure are far more viable than the high temperature and high pressure chemical alternatives. Another advantage of enzyme synthesis is the ability to perform reactions not possible by conventional synthesis, or which would require many synthesis steps as opposed to a single enzymatic reaction (192). Numerous enzymatic reactions are used by industry today, many involving proprietary enzyme systems. An example is the synthesis of cyanohydrins (193) in the production of vitamin B5. Lipases are used to produce esters as they offer favourable alternatives to acid catalysis (194). Styrene oxygenases are used for selective epoxidation reactions providing enantiomerically pure styrene oxide (195,196). Metalloproteins catalyse a wide range of biological reactions and have therefore been of key interest to the biotechnology industry (197).

The ability of cytochromes P450 to activate CH bonds and perform complex hydroxylations and epoxidations make them an attractive target for chemical synthesis (15). Stereoselective oxidation reactions are difficult chemically and require powerful oxidising agents and protecting groups (198). P450 BM3 is the obvious choice for P450 biotechnology due to its high activity, solubility, mammalian-like CPR domain and ease of expression (109). Directed evolution, a method of random mutagenesis, recombination and screening (199), has often been used in the biotechnology industry to produce enzymes with increased activity and altered substrate selectivity (200). Chimeras of the reductase domain of BM3 fused to human P450s were generated to try and create soluble self sufficient eukaryotic P450s, though the catalytic rates achieved are much slower than those of BM3 itself, likely due to inefficient reductase to eukaryotic P450 heme electron transfer (201).

P450 BM3 has proved to be a versatile biocatalyst, although the WT enzyme is primarily a fatty acid hydroxylase and often doesn’t oxidise unnatural substrates, or else does so slowly. However, it has proven amenable to mutagenesis in order to dramatically widen its range of substrates (149). BM3 heme domain active site mutations predominate in such variants, as though oxidation should be at the most active centre in the substrate molecule, for BM3 (as with other P450s) the way the substrate fits in the active site is more important than the reactivity of the carbon centre. With BM3 there are a small number of residues in contact with the substrate making selecting mutations more difficult, and often making a substrate

specific to a particular structure rather than to a broader class of related molecules. Some P450 residues are more important than others in dictating substrate selectivity, as can been seen in the number of times they appear in random variants. Phe87 is the most commonly mutated BM3 residue and most variants generated by directed evolution contain a substitution at either Ala82, Phe87 or Ala328 (149).

Although the WT BM3 shows limited regioselectivity, aside from an aversion to terminal hydroxylation of fatty acids, variants can show a high degree of regio- and stereo-selectivity (202,203). Initial work focused on molecules that were similar to the natural substrates, and so alkanes were a particular focus as the alcohols produced would be higher value and beneficial for biofuel production or other fine chemical uses. Rational redesign of the BM3 active site to place the Arg47/Tyr51 binding site closer to the heme changed activity in favour of short chain fatty acids, proving the important role of this motif in substrate recognition (204). In contrast, V78A/F87A/I263G and S72Y/V78A/F87A variants pushed hydroxylation positions of fatty acids further up the chain to ω-4 to ω-6 positions (205). Directed evolution was used to produce variants capable of producing ethanol from ethane, though as with many such directed evolution studies, the high number of mutations now involved makes it hard to pinpoint crucial residues (206). A domain based strategy, where the individual domains were mutated separately then recombined, was used to produce an alkane oxidizing variant that was used to convert propane to propanol in whole cells (207). More recently, the WT BM3 enzyme was shown to hydroxylate short chain alkanes by use of perflourocarboxylic acids that were able to bind to BM3 and to force gases into the BM3 active site for hydroxylation of propane and butane (208). The I401P mutation in the heme domain also increases the oxidation rate of both lauric acid and a range of non-natural compounds (209). A crystal structure of the I401P variant revealed some of the basis for this change in properties, and showed that the substrate-free I401P structure was more similar to the fatty acid bound (SB) form of WT BM3, with the distal water displaced and a redox potential more oxidising than the SB WT enzyme (210).

The ability of BM3 variants to hydroxylate indole in culture medium was first noted as they produced indigo, turning the cells blue. This has subsequently been used as an effective screen for altered activity against multiple compounds, such as oxidation of various drug (211). Other screening techniques include a continuous fluorescence based screening assay using p-nitrophenoxycarboxylic acids (pNCA) that are converted to the yellow chromophore p-nitrophenolate upon ω-hydroxylation. This assay proved much more effective with the F87A variant, likely since the increased space around the heme allowed the chromophore

better access (212). The use of fluorescent substrates such as resorufins, which fluoresce when O-dealkylated, has been shown to be effective at screening BM3 for pharmaceutical activity, both in batch and continuous flow assays. This screening method highlighted the R47L/F87V/L188Q variant as effective at dealkylation reactions (213,214). An end point assay for oxidation based on NADPH depletion by alkali, done by treating the sample to produce a florescent product from NADP+, had the advantage of being applicable in a whole cell assay (215,216). In another assay, the protein was labelled with the fluorescent probe N,N'-dimethyl-N-(iodoacetyl)-N'-(7-nitrobenz-2-oxa-1,3-diazol-4-Yl)-ethylenediamine (IANBD) to analyse binding of substrates that don’t cause a spin shift in the heme iron upon binding. The fluorescence change upon substrates coming into close proximity to the heme were used to calculate dissociation constants for the molecules (such as propranolol), giving a rationale to the oxidation of various drugs by BM3 (217).

Multiple BM3 variants have been developed that metabolise substrates of the human cytochromes P450, and many of these have been shown to be more active against substrates of human CYP2E1 and CYP3A4 (218). An F87V/A328F variant was shown to oxidise cyclooctane, cyclodecane and cyclododecane, converting n-octane to 2-(R)-octanol at 46% ee and 92% regioselectivity (219). The R47L/Y51F/F87A/A264G variant showed ability to metabolise polyaromatic hydrocarbons (220). Changing Phe87 to Val, Gly or Ala changed the enantioselectivity of the reaction with propylbenzene and 3-chlorostyrene. (221) The F87A/I263A variant metabolised valencene to nootkatone (a grapefruit flavouring) in whole cell biotransformations (222). The D251G/Q307H variant, with only surface mutations, was shown to be effective at metabolism of the drugs diclofenac, ibuprofen and tolbutamide, and even showed spin state shifts on substrate binding, indicating that active site mutations are not essential for altered activity (223). Steroids were also targeted, and the A82W variant showed ability to hydroxylate steroids testosterone and norethisterone at the 16β position (224). Also, the S72I mutation was shown to invert stereochemistry of testosterone hydroxylation in variants capable of the reaction (225). A novel route to the antimalarial drug artemisinin was developed using a semi-synthetic pathway involving BM3 in one of the steps (226). A bioreactor experiment shown that Bacillus megaterium could catalyse the demethylation of the anti-cancer drug colchicine, likely through activity of P450 BM3 (227). The F87V variant of BM3 was shown to hydroxylate flavonoids, potent anti-oxidants (228). Enantioselectivity for styrene epoxidation was shown to invert on mutation of Ala184 to Lys/Arg, with structural data showing that the substrate binding mode is affected (229). For larger scale turnovers by BM3, biphasic systems have been trialled to remove the problem of toxic metabolite build up (230). Large scale batch production of BM3 in E. coli using fermenters has also been achieved for potential biotechnological applications (231). The

versatility and stability of BM3 for biotechnological uses has been shown with the production of a panel of BM3 variants lyophilised on a 96 well plate in order to produce a wide range of metabolites from a single compound (232).

Focus has been placed on improving the BM3 enzyme for biocatalysis by targeting properties that affect its usefulness in an industrial setting, including increasing its solvent stability (233). Extensive work has been carried out to produce thermostable variants such as the 5H6 variant, which has a half-life at 57 degrees that is 250 times that of the WT enzyme (234). Chimeragenesis has also been employed in an attempt to generate more thermostable BM3 constructs (235). Alternative redox pathways have been looked at, as NADPH is expensive for large scale production. One such method involves putting the peroxide uncoupling pathway into reverse by directly adding hydrogen peroxide to BM3 heme domain samples (the so-called peroxide shunt pathway). The F87A variant proved effective at epoxidising styrene this way, though total turnover numbers were low due to oxidative destruction of the heme (236). Another method involved fusing the BM3 heme domain with a ruthenium catalyst to allow light activated reduction of the P450. This approach enabled the catalysis of lauric acid oxidation at rates exceeding those of the peroxide shunt (237-239). Many regeneration systems are used to convert the cheaper NADP+ to NADPH and thus to regenerate NADPH to allow longer reaction times. The glucose-6-phosphate dehydrogenase system is widely used for this purpose (56,240). The BM3 NADPH binding site has also been targeted by mutagenesis in order to allowing it to accept NADH as a far cheaper electron donor than NADPH. As mentioned above, the W1046A mutation in the reductase domain achieves this desired switch in coenzyme selectivity (173). It was found that the effective concentration of NADPH becomes rate limiting even in whole cell turnover systems, and thus introducing a NADPH regeneration system was found to be effective at increasing product yields (241).

More unusual reactions than simple hydroxylations and epoxidations have also proven possible with BM3 variants. For instance, the F87G variant catalysed the dehalogenation of 4-fluorophenol and 4-chlorophenol, and then reduced the semiquinone to the hydroquinone. The reaction showed a cooperative effect on the addition of long chain aldehydes, with catalytic rates of dehalogenation increasing (242). The F87G variant has been shown to form heme adducts with alkenes, explained by the fact it performs terminal oxidation to an aldehyde, which then links to the γ-meso carbon of the heme via a radical mechanism (243). F87G BM3 is also inactivated by terminal alkynes, this occurring via direct attack of compound I oxygen to the acetylene bond, followed by attack of the pyrrole nitrogen on an acetylene carbon (244). Recent work has involved mutating the heme domain of BM3 as a

magnetic resonance imaging (MRI) contrast agent for the detection of dopamine and serotonin, which act as inhibitors by directly ligating the heme iron through their nitrogen groups (245-247). The same team showed that BM3 was capable of catalysing cyclopropanation reactions via carbene transfer by a reductive mechanism in anaerobic conditions, a reaction that hasn’t previously been shown to be possible enzymatically (58). An F87A/T268A variant was also shown to catalyse the conversion of aldehydes to alcohols by direct hydride transfer from NADPH (248).

The combination of the high catalytic rate of BM3, coupled with its ability to be evolved to metabolise unnatural substrates make it an excellent choice for the production of high value pharmaceutical metabolites.

1.5 References

1. Bunim, J. J., Black, R. L., Lutwak, L., Peterson, R. E., and Whedon, G. D. (1958) Studies on dexamethasone, a new synthetic steroid, in rheumatoid arthritis—a preliminary report. Arthritis & Rheumatism 1, 313-331 2. Rushmore, T. H., and Tony Kong, A. N. (2002) Pharmacogenomics, Regulation and Signaling Pathways of Phase I and II Drug Metabolizing Enzymes. Curr. Drug Metab. 3, 481-490 3. Caldwell, J., Gardner, I., and Swales, N. (1995) An Introduction to Drug Disposition: The Basic Principles of Absorption, Distribution, Metabolism, and Excretion. Toxicologic Pathology 23, 102-114 4. Gonzalez, R. J., and Tarloff, J. B. (2001) Evaluation of hepatic subcellular fractions for Alamar blue and MTT reductase activity. Toxicol. in Vitro 15, 257-259 5. Agarwal, V., Kommaddi, R. P., Valli, K., Ryder, D., Hyde, T. M., Kleinman, J. E., Strobel, H. W., and Ravindranath, V. (2008) Drug Metabolism in Human Brain: High Levels of Cytochrome P4503A43 in Brain and Metabolism of Anti-Anxiety Drug Alprazolam to Its Active Metabolite. PLoS ONE 3, e2337 6. Omura, T. (2006) Mitochondrial P450s. Chem.-Biol. Interact. 163, 86-93 7. DeVore, N. M., and Scott, E. E. (2012) Nicotine and 4-(Methylnitrosamino)-1-(3- pyridyl)-1-butanone Binding and Access Channel in Human Cytochrome P450 2A6 and 2A13 Enzymes. J. Biol. Chem. 287, 26576-26585 8. Guengerich, F. P. (2006) Cytochrome P450s and other enzymes in drug metabolism and toxicity. AAPS J 8, E101-E111 9. Dong, H., Haining, R. L., Thummel, K. E., Rettie, A. E., and Nelson, S. D. (2000) Involvement of Human Cytochrome P450 2D6 in the Bioactivation of Acetaminophen. Drug Metab. Dispos. 28, 1397-1400 10. Amunom, I., Dieter, L. J., Tamasi, V., Cai, J., Conklin, D. J., Srivastava, S., Martin, M. V., Guengerich, F. P., and Prough, R. A. (2011) Cytochromes P450 Catalyze the Reduction of α,β-Unsaturated Aldehydes. Chem. Res. Toxicol. 24, 1223-1230 11. Nikolopoulou, M., Eickenbusch, P., Pasadakis, N., Venieri, D., and Kalogerakis, N. Microcosm evaluation of autochthonous bioaugmentation to combat marine oil spills. New Biotechnology 12. Jeong, H. G., Kang, M. J., Kim, H. G., Oh, D. G., Kim, J. S., Lee, S. K., and Jeong, T. C. (2013) Role of intestinal microflora in xenobiotic-induced toxicity. Molecular Nutrition & Food Research 57, 84-99

13. Blaut, M., Braune, A., Wunderlich, S., Sauer, P., Schneider, H., and Glatt, H. (2006) Mutagenicity of arbutin in mammalian cells after activation by human intestinal bacteria. Food Chem. Toxicol. 44, 1940-1947 14. Croom, E. (2012) Chapter Three - Metabolism of Xenobiotics of Human Environments. in Progress in Molecular Biology and Translational Science (Ernest, H. ed.), Academic Press. pp 31-88 15. Isin, E. M., and Guengerich, F. P. (2007) Complex reactions catalyzed by cytochrome P450 enzymes. Biochim. Biophys. Acta (BBA) - General Subjects 1770, 314-329 16. (2005) Drug Metabolism in Context. in Drug Metabolism (Ionescu, C., and Caira, M. eds.), Springer Netherlands. pp 1-40 17. (1992) The Human Hepatic Cytochromes P450 Involved in Drug Metabolism. Critical Reviews in Toxicology 22, 1-21 18. Nelson, D. R., Koymans, L., Kamataki, T., Stegeman, J. J., Feyereisen, R., Waxman, D. J., Waterman, M. R., Gotoh, O., Coon, M. J., Estabrook, R. W., Gunsalus, I. C., and Nebert, D. W. (1996) P450 superfamily: update on new sequences, gene mapping, accession numbers and nomenclature. Pharmacogenetics and Genomics 6, 1-42 19. Lasker, J. M., Wester, M. R., Aramsombatdee, E., and Raucy, J. L. (1998) Characterization of CYP2C19 and CYP2C9 from Human Liver: Respective Roles in Microsomal Tolbutamide,S-Mephenytoin, and Omeprazole Hydroxylations. Arch. Biochem. Biophys. 353, 16-28 20. Wada, Y., Mitsuda, M., Ishihara, Y., Watanabe, M., Iwasaki, M., and Asahi, S. (2008) Important Amino Acid Residues that Confer CYP2C19 Selective Activity to CYP2C9. J. Biochem. 144, 323-333 21. Minagawa, T., Nakano, K., Furuta, S., Iwasa, T., Takekawa, K., Minato, K., Koga, T., Sato, T., Kawashima, K., Kurahashi, Y., Onodera, H., Naito, S., and Nakamura, K. (2012) Perspectives on non-clinical safety evaluation of drug metabolites through the JSOT workshop. The Journal of Toxicological Sciences 37, 667-673 22. Ohno, Y. (1991) Toxicity testing: Regulatory Perspectives. in Proceedings of the First International Conference on Harmonisation (Daph, DWG. ed., The Queen's University of Belfast, Brussels, Belgium 23. Baillie, T. A., Cayen, M. N., Fouda, H., Gerson, R. J., Green, J. D., Grossman, S. J., Klunk, L. J., LeBlanc, B., Perkins, D. G., and Shipley, L. A. (2002) Drug Metabolites in Safety Testing. Toxicol. Appl. Pharmacol. 182, 188-196 24. Agency, U. S. E. P. (1998) Health Effects Test Guidelines, Metabolism and Pharmacokinetics. in OPPTS 870.7485, U. S. E. P. A ed., USA

25. Boyd, E. M., and Bereczky, G. M. (1966) Liver necrosis from paracetamol. Br. J. Pharmacol. Chemother. 26, 606-614 26. Woosley, R. L., Chen, Y., Freiman, J. P., and Gillis, R. A. (1993) Mechanism of the cardiotoxic actions of terfenadine. JAMA 269, 1532-1536 27. Tirkkonen, T., and Laine, K. (2004) Drug interactions with the potential to prevent prodrug activation as a common source of irrational prescribing in hospital inpatients[ast]. Clin. Pharmacol. Ther. 76, 639-647 28. Lown, K. S., Bailey, D. G., Fontana, R. J., Janardan, S. K., Adair, C. H., Fortlage, L. A., Brown, M. B., Guo, W., and Watkins, P. B. (1997) Grapefruit juice increases felodipine oral availability in humans by decreasing intestinal CYP3A protein expression. J. Clin. Invest. 99, 2545-2553 29. Aklillu, E., Persson, I., Bertilsson, L., Johansson, I., Rodrigues, F., and Ingelman- Sundberg, M. (1996) Frequent distribution of ultrarapid metabolizers of debrisoquine in an ethiopian population carrying duplicated and multiduplicated functional CYP2D6 alleles. Journal of Pharmacology and Experimental Therapeutics 278, 441-446 30. Caraco, Y., Sheller, J., and Wood, A. J. (1996) Pharmacogenetic determination of the effects of codeine and prediction of drug interactions. Journal of Pharmacology and Experimental Therapeutics 278, 1165-1174 31. Rosenblat, R., and Tang, S. W. (1987) Do Oriental psychiatric patients receive different dosages of psychotropic medication when compared with occidentals. Can. J. Psychiatry 32, 270-274 32. Desta, Z., Zhao, X., Shin, J.-G., and Flockhart, D. (2002) Clinical Significance of the Cytochrome P450 2C19 Genetic Polymorphism. Clin. Pharmacokinet. 41, 913-958 33. AstraZeneca. (2012) Prilosec® (omeprazole) delayed-release capsules and Prilosec® (omeprazole magnesium) for delayed-release oral suspension prescribing information. (AstraZeneca ed., Wilmington, USA 34. Furuta, T., Ohashi, K., Kamata, T., Takashima, M., Kosuge, K., Kawasaki, T., Hanai, H., Kubota, T., Ishizaki, T., and Kaneko, E. (1998) Effect of genetic differences in omeprazole metabolism on cure rates for Helicobacter pylori infection and peptic ulcer. Ann. Intern. Med. 129, 1027-1030 35. Furuta, T., Sugimoto, M., Shirai, N., and Ishizaki, T. (2007) CYP2C19 pharmacogenomics associated with therapy of Helicobacter pylori infection and gastro-esophageal reflux diseases with a proton pump inhibitor. Pharmacogenomics 8, 1199-1210 36. Ishizaki, T., Chiba, K., Manabe, K., Koyama, E., Hayashi, M., Yasuda, S., Horai, Y., Tomono, Y., Yamato, C., and Toyoki, T. (1995) Comparison of the interaction potential of a new proton pump inhibitor, E3810, versus omeprazole with diazepam in

extensive and poor metabolizers of S-mephenytoin 4[prime]-hydroxylation[ast]. Clin. Pharmacol. Ther. 58, 155-164 37. Loebstein, R., Yonath, H., Peleg, D., Almog, S., Rotenberg, M., Lubetsky, A., Roitelman, J., Harats, D., Halkin, H., and Ezra, D. (2001) Interindividual variability in sensitivity to warfarin-Nature or nurture? Clin. Pharmacol. Ther. 70, 159-164 38. Black, D. J., Kunze, K. L., Wienkers, L. C., Gidal, B. E., Seaton, T. L., McDonnell, N. D., Evans, J. S., Bauwens, J. E., and Trager, W. F. (1996) Warfarin-fluconazole. II. A metabolically based drug interaction: in vivo studies. Drug Metab. Dispos. 24, 422- 428 39. Duetz, W. A., Van Beilen, J. B., and Witholt, B. (2001) Using proteins in their natural environment: potential and limitations of microbial whole-cell hydroxylations in applied biocatalysis. Curr. Opin. Biotechnol. 12, 419-425 40. Martinez, C. A., and Rupashinghe, S. G. (2013) Cytochrome p450 bioreactors in the pharmaceutical industry: challenges and opportunities. Curr. Top. Med. Chem. 13, 1470-1490 41. Caswell, J. M., O’Neill, M., Taylor, S. J. C., and Moody, T. S. (2013) Engineering and application of P450 monooxygenases in pharmaceutical and metabolite synthesis. Curr. Opin. Chem. Biol. 17, 271-275 42. Sawayama, A. M., Chen, M. M. Y., Kulanthaivel, P., Kuo, M.-S., Hemmerle, H., and Arnold, F. H. (2009) A Panel of Cytochrome P450 BM3 Variants to Produce Drug Metabolites and Diversify Lead Compounds. Chemistry – A European Journal 15, 11723-11729 43. Schmid, A., Dordick, J. S., Hauer, B., Kiener, A., Wubbolts, M., and Witholt, B. (2001) Industrial biocatalysis today and tomorrow. Nature 409, 258-268 44. Schulze, B., and Wubbolts, M. G. (1999) Biocatalysis for industrial production of fine chemicals. Curr. Opin. Biotechnol. 10, 609-615 45. Oyama, K., Irino, S., and Hagi, N. (1987) [46] Production of aspartame by immobilized thermoase. in Methods Enzymol. (Klaus, M. ed.), Academic Press. pp 503-516 46. Tufvesson, P. r., Lima-Ramos, J., Nordblad, M., and Woodley, J. M. (2010) Guidelines and Cost Analysis for Catalyst Production in Biocatalytic Processes. Org. Process Res. Dev. 15, 266-274 47. Nelson, D. (2009) The Cytochrome P450 Homepage. Human Genomics 4, 59 - 65 48. Hayaishi, O., Katagiri, M., and Rothberg, S. (1955) Mechanism of the pyrocatechase reaction. J. Am. Chem. Soc. 77, 5450-5451 49. Heinrich, W., and Hans, L. (1912) Die Dissoziation der Tetraarylhydrazine und der Diarylnitrosamine (XII). Justus Liebig's Annalen der Chemie 392, 156-169

50. Klinman, J. P. (2007) How Do Enzymes Activate Oxygen without Inactivating Themselves? Acc. Chem. Res. 40, 325-333 51. Ilia, G. D., Thomas, M. M., Stephen, G. S., and Ilme, S. (2005) Structure and Chemistry of Cytochrome P450. ChemInform 36 52. Muller, J. r. J., Lapko, A., Bourenkov, G., Ruckpaul, K., and Heinemann, U. (2001) Adrenodoxin Reductase-Adrenodoxin Complex Structure Suggests Electron Transfer Path in Steroid Biosynthesis. J. Biol. Chem. 276, 2786-2789 53. Murataliev, M. B., Feyereisen, R., and Walker, F. A. (2004) Electron transfer by diflavin reductases. Biochimica et Biophysica Acta (BBA) - Proteins & Proteomics 1698, 1-26 54. Munro, A. W., Girvan, H. M., and McLean, K. J. (2007) Variations on a (t)heme) - novel mechanisms, redox partners and catalytic functions in the cytochrome P450 superfamily. Nat. Prod. Rep. 24, 585-609 55. Guengerich, F. P. (2001) Common and uncommon cytochrome P450 reactions related to metabolism and chemical toxicity. Chem. Res. Toxicol. 14, 611-650 56. Waltham, T. N., Girvan, H. M., Butler, C. F., Rigby, S. R., Dunford, A. J., Holt, R. A., and Munro, A. W. (2011) Analysis of the oxidation of short chain alkynes by flavocytochrome P450 BM3. Metallomics 3, 369-378 57. Guengerich, F. P., and Munro, A. W. (2013) Unusual Cytochrome P450 Enzymes and Reactions. J. Biol. Chem.(in press) 58. Coelho, P. S., Brustad, E. M., Kannan, A., and Arnold, F. H. (2013) Olefin Cyclopropanation via Carbene Transfer Catalyzed by Engineered Cytochrome P450 Enzymes. Science 339, 307-310 59. Nishida, C. R., Lee, M., and de Montellano, P. R. O. (2010) Efficient Hypoxic Activation of the Anticancer Agent AQ4N by CYP2S1 and CYP2W1. Mol. Pharmacol. 78, 497-502 60. Zhou, S. F. (2008) Drugs behave as substrates, inhibitors and inducers of human cytochrome P450 3A4. Curr. Drug. Metab. 9, 310-322 61. Wrighton, S. A., and Stevens, J. C. (1992) The Human Hepatic Cytochromes P450 Involved in Drug Metabolism. Critical Reviews in Toxicology 22, 1-21 62. Nebert, D. W., and Russell, D. W. (2002) Clinical importance of the cytochromes P450. The Lancet 360, 1155-1162 63. Sligar, S. G., Makris, T. M., and Denisov, I. G. (2005) Thirty years of microbial P450 monooxygenase research: Peroxo-heme intermediates--The central bus station in heme oxygenase catalysis. Biochem. Biophys. Res. Commun. 338, 346-354 64. Xue, Y., and Sherman, D. H. (2001) Biosynthesis and Combinatorial Biosynthesis of Pikromycin-Related Macrolides in Streptomyces venezuelae. Metab. Eng.3, 15-26

65. Urlacher, V. B., Lutz-Wahl, S., and Schmid, R. D. (2004) Microbial P450 enzymes in biotechnology. Appl. Microbiol. Biotechnol. 64, 317-325 66. Tanaka, Y., and Brugliera, F. (2013) Flower colour and cytochromes P450. Philos. Trans. R. Soc. Lond. B. Biol. Sci. 368, 20120432 67. Ferguson, C. S., and Tyndale, R. F. (2011) Cytochrome P450 enzymes in the brain: emerging evidence of biological significance. Trends Pharmacol. Sci. 32, 708-714 68. Phillips, D. H. (1983) Fifty years of benzo(a)pyrene. Nature 303, 468-472 69. Beaune, P., Dansette, P. M., Mansuy, D., Kiffel, L., Finck, M., Amar, C., Leroux, J. P., and Homberg, J. C. (1987) Human anti-endoplasmic reticulum autoantibodies appearing in a drug-induced hepatitis are directed against a human liver cytochrome P-450 that hydroxylates the drug. Proc. Natl. Acad. Sci. U.S.A 84, 551-555 70. Beaune, P., Pessayre, D., Dansette, P., Mansuy, D., and Manns, M. (1994) Autoantibodies against cytochromes P450: role in human diseases. Adv. Pharmacol. 30, 199-245 71. Lewis, D. F. V. (1999) Homology modelling of human cytochromes P450 involved in xenobiotic metabolism and rationalization of substrate selectivity. Experimental and Toxicologic Pathology 51, 369-374 72. (2009) The Simcyp® Population-based ADME Simulator. Expert Opinion on Drug Metabolism & Toxicology 5, 211-223 73. Woggon, W.-D. (ed) (2007) Structural and functional mimics of cytochromes P450, Vol. 3, Wiley, Chichester 74. McLean, K. J., Dunford, A. J., Neeli, R., Driscoll, M. D., and Munro, A. W. (2007) Structure, function and drug targeting in Mycobacterium tuberculosis cytochrome P450 systems. Arch. Biochem. Biophys. 464, 228-240 75. Rittle, J., and Green, M. T. (2010) Cytochrome P450 Compound I: Capture, Characterization, and C-H Bond Activation Kinetics. Science 330, 933-937 76. Denisov, I. G., Makris, T. M., Sligar, S. G., and Schlichting, I. (2005) Structure and chemistry of cytochrome P450. Chem. Rev. 105, 2253-2277 77. Meunier, B., and Bernadou, J. (2000) Active Iron-Oxo and Iron-Peroxo Species in Cytochromes P450 and Peroxidases; Oxo-Hydroxo Tautomerism with Water-Soluble Metalloporphyrins. in Metal-Oxo and Metal-Peroxo Species in Catalytic Oxidations (Meunier, B. ed.), Springer Berlin Heidelberg. pp 1-35 78. Denisov, I. G., Mak, P. J., Makris, T. M., Sligar, S. G., and Kincaid, J. R. (2008) Resonance Raman Characterization of the Peroxo and Hydroperoxo Intermediates in Cytochrome P450†. J. Phys. Chem. A 112, 13172-13179

79. Groves, J. T., McClusky, G. A., White, R. E., and Coon, M. J. (1978) Aliphatic hydroxylation by highly purified liver microsomal cytochrome P-450. Evidence for a carbon radical intermediate. Biochem. Biophys. Res. Commun. 81, 154-160 80. Green, M. T., Dawson, J. H., and Gray, H. B. (2004) Oxoiron(IV) in Chloroperoxidase Compound II Is Basic: Implications for P450 Chemistry. Science 304, 1653-1656 81. Newcomb, M., Halgrimson, J. A., Horner, J. H., Wasinger, E. C., Chen, L. X., and Sligar, S. G. (2008) X-ray absorption spectroscopic characterization of a cytochrome P450 compound II derivative. Proc. Natl. Acad. Sci. U. S. A. 105, 8179-8184 82. Ogliaro, F., Harris, N., Cohen, S., Filatov, M., de Visser, S. P., and Shaik, S. (2000) A Model “Rebound” Mechanism of Hydroxylation by Cytochrome P450: Stepwise and Effectively Concerted Pathways, and Their Reactivity Patterns. J. Am. Chem. Soc. 122, 8977-8989 83. Berglund, G. I., Carlsson, G. H., Smith, A. T., Szoke, H., Henriksen, A., and Hajdu, J. (2002) The catalytic pathway of horseradish peroxidase at high resolution. Nature 417, 463-468 84. Rutter, R., and Hager, L. P. (1982) The detection of two electron paramagnetic resonance radical signals associated with chloroperoxidase compound I. J.Biol. Chem. 257, 7958-7961 85. Rutter, R., Hager, L. P., Dhonau, H., Hendrich, M., Valentine, M., and Debrunner, P. (1984) Chloroperoxidase compound I: Electron paramagnetic resonance and Mössbauer studies. Biochemistry 23, 6809-6816 86. Kim, S. H., Perera, R., Hager, L. P., Dawson, J. H., and Hoffman, B. M. (2006) Rapid freeze-quench ENDOR study of chloroperoxidase compound I: The site of the radical. J. Am. Chem. Soc. 128, 5598-5599 87. Newcomb, M., Zhang, R., Chandrasena, R. E. P., Halgrimson, J. A., Horner, J. H., Makris, T. M., and Sligar, S. G. (2006) Cytochrome P450 compound I. J. Am. Chem. Soc. 128, 4580-4581 88. Yuan, X., Wang, Q., Horner, J. H., Sheng, X., and Newcomb, M. (2009) Kinetics and activation parameters for oxidations of styrene by Compounds I from the cytochrome P450BM-3 (CYP102A1) heme domain and from CYP119. Biochemistry 48, 9140- 9146 89. Behan, R. K., Hoffart, L. M., Stone, K. L., Krebs, C., and Green, M. T. (2007) Reaction of cytochrome P450BM3 and peroxynitrite yields nitrosyl complex. J. Am. Chem. Soc. 129, 5855-5859 90. Krest, C. M., Onderko, E. L., Yosca, T. H., Calixto, J. C., Karp, R. F., Livada, J., Rittle, J., and Green, M. T. (2013) Reactive Intermediates in Cytochrome P450 Catalysis. J. Biol. Chem. 288, 17074-17081

91. Sen, K., and Hackett, J. C. (2010) Peroxo−Iron Mediated Deformylation in Sterol 14α-Demethylase Catalysis. J. Am. Chem. Soc. 132, 10293-10305 92. Vaz, A. D., Pernecky, S. J., Raner, G. M., and Coon, M. J. (1996) Peroxo-iron and oxenoid-iron species as alternative oxygenating agents in cytochrome P450- catalyzed reactions: switching by threonine-302 to alanine mutagenesis of cytochrome P450 2B4. Proc. Natl. Acad. Sci. U.S.A 93, 4644-4648 93. Roberts, E. S., Vaz, A. D., and Coon, M. J. (1991) Catalysis by cytochrome P-450 of an oxidative reaction in xenobiotic aldehyde metabolism: deformylation with olefin formation. Proc. Natl. Acad. Sci. U.S.A 88, 8963-8966 94. Cryle, M. J., and De Voss, J. J. (2006) Is the Ferric Hydroperoxy Species Responsible for Sulfur Oxidation in Cytochrome P450s? Angew. Chem., Int. Ed. 45, 8221-8223 95. Hrycay, E. G., and Bandiera, S. M. (2012) The monooxygenase, peroxidase, and peroxygenase properties of cytochrome P450. Arch. Biochem. Biophys. 522, 71-89 96. Rude, M. A., Baron, T. S., Brubaker, S., Alibhai, M., Del Cardayre, S. B., and Schirmer, A. (2011) Terminal olefin (1-alkene) biosynthesis by a novel p450 fatty acid decarboxylase from Jeotgalicoccus species. Appl. Environ. Microbiol. 77, 1718-1727 97. Imai, Y., Matsunaga, I., Kusunose, E., and Ichihara, K. (2000) Unique heme environment at the putative distal region of hydrogen peroxide-dependent fatty acid alpha-hydroxylase from Sphingomonas paucimobilis (peroxygenase P450(SPalpha). J. Biochem. 128, 189-194 98. Lee, D.-S., Yamada, A., Sugimoto, H., Matsunaga, I., Ogura, H., Ichihara, K., Adachi, S.-i., Park, S.-Y., and Shiro, Y. (2003) Substrate Recognition and Molecular Mechanism of Fatty Acid Hydroxylation by Cytochrome P450 from Bacillus subtilis : Crystallographic, spectroscopic, and mutational studies. J. Biol. Chem. 278, 9761- 9767 99. Nordblom, G. D., White, R. E., and Coon, M. J. (1976) Studies on hydroperoxide- dependent substrate hydroxylation by purified liver microsomal cytochrome P-450. Arch. Biochem. Biophys. 175, 524-533 100. Tsai, R., Yu, C. A., Gunsalus, I. C., Peisach, J., Blumberg, W., Orme-Johnson, W. H., and Beinert, H. (1970) Spin-state changes in cytochrome P-450cam on binding of specific substrates. Proc. Natl. Acad. Sci. U.S.A 66, 1157-1163 101. Luthra, A., Denisov, I. G., and Sligar, S. G. (2011) Spectroscopic features of cytochrome P450 reaction intermediates. Arch. Biochem. Biophys. 507, 26-35 102. Fisher, M. T., and Sligar, S. G. (1985) Control of heme protein redox potential and reduction rate: linear free energy relation between potential and ferric spin state equilibrium. J. Am. Chem. Soc. 107, 5018-5019

103. Varadarajan, R., Zewert, T., Gray, H., and Boxer, S. (1989) Effects of buried ionizable amino acids on the reduction potential of recombinant myoglobin. Science 243, 69-72 104. Ogliaro, F., de Visser, S. P., and Shaik, S. (2002) The ‘push’ effect of the thiolate ligand in cytochrome P450: a theoretical gauging. J. Inorg. Biochem. 91, 554-567 105. Poulos, T. L. (2003) The past and present of P450cam structural biology. Biochem. Biophys. Res. Commun. 312, 35-39 106. Williams, P. A., Cosme, J., Vinkovic, D. M., Ward, A., Angove, H. C., Day, P. J., Vonrhein, C., Tickle, I. J., and Jhoti, H. (2004) Crystal Structures of Human Cytochrome P450 3A4 Bound to Metyrapone and Progesterone. Science 305, 683- 686 107. Williams, P. A., Cosme, J., Ward, A., Angove, H. C., Matak Vinkovic, D., and Jhoti, H. (2003) Crystal structure of human cytochrome P450 2C9 with bound warfarin. Nature 424, 464-468 108. Ortiz de Montellano, P. R. (2005) Cytochrome P450: Structure, Mechanism, and Biochemistry, 3 ed., Kluwer Academic/Plenum Publishers, New York 109. Munro, A. W., Leys, D. G., McLean, K. J., Marshall, K. R., Ost, T. W. B., Daff, S., Miles, C. S., Chapman, S. K., Lysek, D. A., Moser, C. C., Page, C. C., and Dutton, P. L. (2002) P450 BM3: the very model of a modern flavocytochrome. Trends Biochem. Sci. 27, 250-257 110. Haines, D. C., Tomchick, D. R., Machius, M., and Peterson, J. A. (2001) Pivotal role of water in the mechanism of P450BM-3. Biochemistry 40, 13456-13465 111. Poulos, T. L., Finzel, B. C., and Howard, A. J. (1987) High-resolution crystal structure of cytochrome P450cam. J. Mol. Biol. 195, 687-700 112. Meharenna, T. L. P. a. Y. T. (ed) (2007) Structure of P450 Proteins and Their Molecular Phylogeny, Vol. 3, Wiley, Chichester 113. Sono, M., Stuehr, D. J., Ikeda-Saito, M., and Dawson, J. H. (1995) Identification of Nitric Oxide Synthase as a Thiolate-ligated Heme Protein Using Magnetic Circular Dichroism Spectroscopy: Comparison with cytochrome P-450-Cam and chloroperoxidase. J. Biol. Chem. 270, 19943-19948 114. Fann, Y.-C., Gerber, N. C., Osmulski, P. A., Hager, L. P., Sligar, S. G., and Hoffman, B. M. (1994) ENDOR Determination of Heme Ligation in Chloroperoxidase and Comparison with Cytochrome P-450Cam. J. Am. Chem. Soc. 116, 5989-5990 115. Ost, T. W. B., Munro, A. W., Mowat, C. G., Taylor, P. R., Pesseguiero, A., Fulco, A. J., Cho, A. K., Cheesman, M. A., Walkinshaw, M. D., and Chapman, S. K. (2001) Structural and Spectroscopic Analysis of the F393H Mutant of Flavocytochrome P450 BM3. Biochemistry 40, 13430-13438

116. Ost, T. W. B., Miles, C. S., Munro, A. W., Murdoch, J., Reid, G. A., and Chapman, S. K. (2001) Phenylalanine 393 Exerts Thermodynamic Control over the Heme of Flavocytochrome P450 BM3†. Biochemistry 40, 13421-13429 117. Schlichting, I., Berendzen, J., Chu, K., Stock, A. M., Maves, S. A., Benson, D. E., Sweet, R. M., Ringe, D., Petsko, G. A., and Sligar, S. G. (2000) The Catalytic Pathway of Cytochrome P450cam at Atomic Resolution. Science 287, 1615-1622 118. Poulos, T. L., and Madrona, Y. (2013) Oxygen activation and redox partner binding in cytochromes P450. Biotechnol. Appl. Biochem. 60, 128-133 119. Cirino, P. C., and Arnold, F. H. (2002) Protein engineering of oxygenases for biocatalysis. Curr. Opin. Chem. Biol. 6, 130-135 120. Munro, A. W., Girvan, H. M., and McLean, K. J. (2007) Cytochrome P450-redox partner fusion enzymes. Biochimica et Biophysica Acta (BBA) - General Subjects 1770, 345-359 121. Narhi, L. O., and Fulco, A. J. (1986) Characterization of a catalytically self-sufficient 119,000-dalton cytochrome P-450 monooxygenase induced by barbiturates in Bacillus megaterium. J. Biol. Chem. 261, 7160-7169 122. Kitazume, T., Takaya, N., Nakayama, N., and Shoun, H. (2000) Fusarium oxysporum Fatty-acid Subterminal Hydroxylase (CYP505) Is a Membrane-bound Eukaryotic Counterpart of Bacillus megaterium Cytochrome P450BM3. J. Biol. Chem. 275, 39734-39740 123. Nakayama, N., Takemae, A., and Shoun, H. (1996) Cytochrome P450foxy, a Catalytically Self-Sufficient Fatty Acid Hydroxylase of the Fungus Fusarium oxysporum. J. Biochem. 119, 435-440 124. Shoun, H., Sudo, Y., and Beppu, T. (1985) Subterminal Hydroxylation of Fatty Acids by a Cytochrome P-450-Dependent Enzyme System from a Fungus, Fusarium oxysporum. J. Biochem. 97, 755-763 125. Gustafsson, M. C. U., Roitel, O., Marshall, K. R., Noble, M. A., Chapman, S. K., Pessegueiro, A., Fulco, A. J., Cheesman, M. R., von Wachenfeldt, C., and Munro, A. W. (2004) Expression, Purification, and Characterization of Bacillus subtilis Cytochromes P450 CYP102A2 and CYP102A3: Flavocytochrome Homologues of P450 BM3 from Bacillus megaterium. Biochemistry 43, 5474-5487 126. Tripathi, S., Li, H., and Poulos, T. L. (2013) Structural Basis for Effector Control and Redox Partner Recognition in Cytochrome P450. Science 340, 1227-1230 127. Simondsen, R. P., and Tollin, G. (1980) Structure-function relations in flavodoxins. Mol. Cell Biochem. 33, 13-24 128. Stanley, R. J. (2001) Advances in flavin and flavoprotein optical spectroscopy. Antioxid. Redox Signal 3, 847-866

129. Huang, S., Heikal, A. A., and Webb, W. W. (2002) Two-photon fluorescence spectroscopy and microscopy of NAD(P)H and flavoprotein. Biophys. J. 82, 2811- 2825 130. Pudney, C. R., Heyes, D. J., Khara, B., Hay, S., Rigby, S. E., and Scrutton, N. S. (2012) Kinetic and spectroscopic probes of motions and catalysis in the cytochrome P450 reductase family of enzymes. FEBS J. 279, 1534-1544 131. Roome, P. W., Jr., Philley, J. C., and Peterson, J. A. (1983) Purification and properties of putidaredoxin reductase. J. Biol. Chem. 258, 2593-2598 132. Sevrioukova, I. F., Hazzard, J. T., Tollin, G., and Poulos, T. L. (2001) Laser flash induced electron transfer in P450cam monooxygenase: putidaredoxin reductase- putidaredoxin interaction. Biochemistry 40, 10592-10600 133. Gunsalus, I. C., and Sligar, S. G. (1976) Redox regulation of cytochrome P450cam mixed function oxidation by putidaredoxin and camphor ligation. Biochimie 58, 143- 147 134. Wang, M., Roberts, D. L., Paschke, R., Shea, T. M., Masters, B. S. S., and Kim, J.-J. P. (1997) Three-dimensional structure of NADPH–cytochrome P450 reductase: Prototype for FMN- and FAD-containing enzymes. Proc. Natl. Acad. Sci. U. S. A. 94, 8411-8416 135. Gutierrez, A., Grunau, A., Paine, M., Munro, A. W., Wolf, C. R., Roberts, G. C., and Scrutton, N. S. (2003) Electron transfer in human cytochrome P450 reductase. Biochem. Soc. Trans. 31, 497-501 136. Iyanagi, T., and Mason, H. S. (1973) Some properties of hepatic reduced nicotinamide adenine dinucleotide phosphate-cytochrome c reductase. Biochemistry 12, 2297-2308 137. Das, A., and Sligar, S. G. (2009) Modulation of the Cytochrome P450 Reductase Redox Potential by the Phospholipid Bilayer. Biochemistry 48, 12104-12112 138. Sevrioukova, I. F., and Peterson, J. A. (1995) NADPH-P-450 reductase: structural and functional comparisons of the eukaryotic and prokaryotic isoforms. Biochimie 77, 562-572 139. Miles, J. S., Munro, A. W., Rospendowski, B. N., Smith, W. E., McKnight, J., and Thomson, A. J. (1992) Domains of the catalytically self-sufficient cytochrome P-450 BM-3. Genetic construction, overexpression, purification and spectroscopic characterization. Biochem. J. 288 ( Pt 2), 503-509 140. Munro, A. W., Gordon Lindsay, J., Coggins, J. R., Kelly, S. M., and Price, N. C. (1996) Analysis of the structural stability of the multidomain enzyme flavocytochrome P-450 BM3. Biochimica et Biophysica Acta (BBA) - Protein Structure and Molecular Enzymology 1296, 127-137

141. Hudeèek, J., Baumruk, V., Anzenbacher, P., and Munro, A. W. (1998) Catalytically Self-Sufficient P450 CYP102 (Cytochrome P450 BM-3): Resonance Raman Spectral Characterization of the Heme Domain and of the Holoenzyme. Biochem. Biophys. Res. Commun. 243, 811-815 142. Aldo Gutierrez, A. W. M., Alex Grunau, C. Roland Wolf, Nigel S. Scrutton, Gordon C. K. Roberts,. (2003) Interflavin electron transfer in human cytochrome P450 reductase is enhanced by coenzyme binding. Eur. J. Biochem. 270, 2612-2621 143. Girvan, H. M., Seward, H. E., Toogood, H. S., Cheesman, M. R., Leys, D., and Munro, A. W. (2007) Structural and Spectroscopic Characterization of P450 BM3 Mutants with Unprecedented P450 Heme Iron Ligand Sets: New heme ligation states influence conformational equilibria in P450 BM3. J. Biol. Chem. 282, 564-572 144. Ravichandran, K., Boddupalli, S., Hasermann, C., Peterson, J., and Deisenhofer, J. (1993) Crystal structure of hemoprotein domain of P450BM-3, a prototype for microsomal P450's. Science 261, 731-736 145. Ruettinger, R. T., and Fulco, A. J. (1981) Epoxidation of unsaturated fatty acids by a soluble cytochrome P-450-dependent system from Bacillus megaterium. J. Biol. Chem. 256, 5728-5734 146. Wen, L. P., and Fulco, A. J. (1987) Cloning of the gene encoding a catalytically self- sufficient cytochrome P-450 fatty acid monooxygenase induced by barbiturates in Bacillus megaterium and its functional expression and regulation in heterologous (Escherichia coli) and homologous (Bacillus megaterium) hosts. J. Biol. Chem. 262, 6676-6682 147. Narhi, L. O., Wen, L.-P., and Fulco, A. J. (1988) Characterization of the protein expressed in Escherichia coli by a recombinant plasmid containing the Bacillus megaterium cytochrome P-450BM-3 gene. Mol. Cell. Biochem. 79, 63-71 148. English, N., Hughes, V., and Wolf, C. R. (1994) Common pathways of cytochrome P450 gene regulation by peroxisome proliferators and barbiturates in Bacillus megaterium ATCC14581. J. Biol. Chem. 269, 26836-26841 149. Whitehouse, C. J. C., Bell, S. G., and Wong, L.-L. (2012) P450BM3 (CYP102A1): connecting the dots. Chem. Soc. Rev. 41, 1218-1260 150. Kaneda, T. (1991) Iso- and anteiso-fatty acids in bacteria: biosynthesis, function, and taxonomic significance. Microbiol. Rev. 55, 288-302 151. Narhi, L. O., and Fulco, A. J. (1987) Identification and characterization of two functional domains in cytochrome P-450BM-3, a catalytically self-sufficient monooxygenase induced by barbiturates in Bacillus megaterium. J. Biol. Chem. 262, 6683-6690

152. Govindaraj, S., and Poulos, T. L. (1997) The domain architecture of cytochrome P450BM-3. J. Biol. Chem. 272, 7915-7921 153. Warman, A. J., Roitel, O., Neeli, R., Girvan, H. M., Seward, H. E., Murray, S. A., McLean, K. J., Joyce, M. G., Toogood, H., Holt, R. A., Leys, D., Scrutton, N. S., and Munro, A. W. (2005) Flavocytochrome P450 BM3: an update on structure and mechanism of a biotechnologically important enzyme. Biochem. Soc. Trans. 33, 747- 753 154. Boddupalli, S. S., Estabrook, R. W., and Peterson, J. A. (1990) Fatty acid monooxygenation by cytochrome P-450BM-3. J. Biol. Chem. 265, 4233-4239 155. Govindaraj, S., and Poulos, T. L. (1995) Role of the linker region connecting the reductase and heme domains in cytochrome P450BM-3. Biochemistry 34, 11221- 11226 156. Sevrioukova, I. F., Li, H., Zhang, H., Peterson, J. A., and Poulos, T. L. (1999) Structure of a cytochrome P450–redox partner electron-transfer complex. Proc. Nat. Acad. Sci.U.S.A 96, 1863-1868 157. Murataliev, M. B., and Feyereisen, R. (1996) Functional interactions in cytochrome P450BM3. Fatty acid substrate binding alters electron-transfer properties of the flavoprotein domain. Biochemistry 35, 15029-15037 158. Boddupalli, S. S., Oster, T., Estabrook, R. W., and Peterson, J. A. (1992) Reconstitution of the fatty acid hydroxylation function of cytochrome P-450BM-3 utilizing its individual recombinant hemo- and flavoprotein domains. J. Biol. Chem. 267, 10375-10380 159. Modi, S., Primrose, W. U., Boyle, J. M., Gibson, C. F., Lian, L. Y., and Roberts, G. C. (1995) NMR studies of substrate binding to cytochrome P450 BM3: comparisons to cytochrome P450 cam. Biochemistry 34, 8982-8988 160. Brenner, S., Hay, S., Girvan, H. M., Munro, A. W., and Scrutton, N. S. (2007) Conformational Dynamics of the Cytochrome P450 BM3/N-Palmitoylglycine Complex: The Proposed Proximal Distal Transition Probed by Temperature-Jump Spectroscopy. J. Phys. Chem. B 111, 7879-7886 161. Eva Anzenbacherová, Nicole Bec, Pavel Anzenbacher, Jiří Hudeček, Pavel Souček, Christiane Jung, Andrew W. Munro, Reinhard Lange. (2000) Flexibility and stability of the structure of cytochromes P4503A4 and BM-3. Eur. J. Biochem. 267, 2916-2920 162. Perera, R., Sono, M., Raner, G. M., and Dawson, J. H. (2005) Subzero-temperature stabilization and spectroscopic characterization of homogeneous oxyferrous complexes of the cytochrome P450 BM3 (CYP102) oxygenase domain and holoenzyme. Biochem. Biophys. Res. Commun. 338, 365-371

163. Rowlatt, B., Yorke, J. A., Strong, A. J., Whitehouse, C. J., Bell, S. G., and Wong, L. L. (2011) Chain length-dependent cooperativity in fatty acid binding and oxidation by cytochrome P450BM3 (CYP102A1). Protein Cell 2, 656-671 164. Whitehouse, C. J., Bell, S. G., Tufton, H. G., Kenny, R. J., Ogilvie, L. C., and Wong, L. L. (2008) Evolved CYP102A1 (P450BM3) variants oxidise a range of non-natural substrates and offer new selectivity options. Chem. Commun. (Camb), 966-968 165. Farinas, E. T., Alcalde, M., and Arnold, F. (2004) Alkene epoxidation catalyzed by cytochrome P450 BM-3 139-3. Tetrahedron 60, 525-528 166. Whitehouse, C. J., Yang, W., Yorke, J. A., Tufton, H. G., Ogilvie, L. C., Bell, S. G., Zhou, W., Bartlam, M., Rao, Z., and Wong, L. L. (2011) Structure, electronic properties and catalytic behaviour of an activity-enhancing CYP102A1 (P450(BM3)) variant. Dalton Trans. 40, 10383-10396 167. Honeychurch, M. J., Hill, A. O., and Wong, L. L. (1999) The thermodynamics and kinetics of electron transfer in the cytochrome P450cam enzyme system. FEBS Lett. 451, 351-353 168. Haines, D. C., Tomchick, D. R., Machius, M., and Peterson, J. A. (2001) Pivotal Role of Water in the Mechanism of P450BM-3. Biochemistry 40, 13456-13465 169. Ravichandran, K. G., Boddupalli, S. S., Hasermann, C. A., Peterson, J. A., and Deisenhofer, J. (1993) Crystal structure of hemoprotein domain of P450BM-3, a prototype for microsomal P450's. Science 261, 731-736 170. Boddupalli, S. S., Hasemann, C. A., Ravichandran, K. G., Lu, J. Y., Goldsmith, E. J., Deisenhofer, J., and Peterson, J. A. (1992) Crystallization and preliminary x-ray diffraction analysis of P450terp and the hemoprotein domain of P450BM-3, enzymes belonging to two distinct classes of the cytochrome P450 superfamily. Proc. Natl. Acad. Sci. U.S.A 89, 5567-5571 171. Kitazume, T., Haines, D. C., Estabrook, R. W., Chen, B., and Peterson, J. A. (2007) Obligatory intermolecular electron-transfer from FAD to FMN in dimeric P450BM-3. Biochemistry 46, 11892-11901 172. Neeli, R., Girvan, H. M., Lawrence, A., Warren, M. J., Leys, D., Scrutton, N. S., and Munro, A. W. (2005) The dimeric form of flavocytochrome P450 BM3 is catalytically functional as a fatty acid hydroxylase. FEBS Lett. 579, 5582-5588 173. Girvan, H. M., Dunford, A. J., Neeli, R., Ekanem, I. S., Waltham, T. N., Joyce, M. G., Leys, D., Curtis, R. A., Williams, P., Fisher, K., Voice, M. W., and Munro, A. W. (2011) Flavocytochrome P450 BM3 mutant W1046A is a NADH-dependent fatty acid hydroxylase: implications for the mechanism of electron transfer in the P450 BM3 dimer. Arch. Biochem. Biophys. 507, 75-85

174. Paulsen, M. D., and Ornstein, R. L. (1995) Dramatic differences in the motions of the mouth of open and closed cytochrome P450BM-3 by molecular dynamics simulations. Proteins 21, 237-243 175. Girvan, H. M., Seward, H. E., Toogood, H. S., Cheesman, M. R., Leys, D., and Munro, A. W. (2007) Structural and spectroscopic characterization of P450 BM3 mutants with unprecedented P450 heme iron ligand sets. New heme ligation states influence conformational equilibria in P450 BM3. J. Biol. Chem. 282, 564-572 176. Warman, A. J., Roitel, O., Neeli, R., Girvan, H. M., Seward, H. E., Murray, S. A., McLean, K. J., Joyce, M. G., Toogood, H., Holt, R. A., Leys, D., Scrutton, N. S., and Munro, A. W. (2005) Flavocytochrome P450 BM3: an update on structure and mechanism of a biotechnologically important enzyme. Biochem. Soc. Trans. 33, 747- 753 177. Joyce, M. G., Ekanem, I. S., Roitel, O., Dunford, A. J., Neeli, R., Girvan, H. M., Baker, G. J., Curtis, R. A., Munro, A. W., and Leys, D. (2012) The crystal structure of the FAD/NADPH-binding domain of flavocytochrome P450 BM3. FEBS J. 279, 1694- 1706 178. Sevrioukova, I. F., Li, H., Zhang, H., Peterson, J. A., and Poulos, T. L. (1999) Structure of a cytochrome P450 redox partner electron-transfer complex. Proc. Natl. Acad. Sci. U.S.A 96, 1863-1868 179. Gotoh, O. (1992) Substrate recognition sites in cytochrome P450 family 2 (CYP2) proteins inferred from comparative analyses of amino acid and coding nucleotide sequences. J. Biol. Chem. 267, 83-90 180. Joyce, M. G., Girvan, H. M., Munro, A. W., and Leys, D. (2004) A single mutation in cytochrome P450 BM3 induces the conformational rearrangement seen upon substrate binding in the wild-type enzyme. J. Biol. Chem. 279, 23287-23293 181. Girvan, H. M., Marshall, K. R., Lawson, R. J., Leys, D., Joyce, M. G., Clarkson, J., Smith, W. E., Cheesman, M. R., and Munro, A. W. (2004) Flavocytochrome P450 BM3 mutant A264E undergoes substrate-dependent formation of a novel heme iron ligand set. J. Biol. Chem. 279, 23274-23286 182. Chen, Z., Ost, T. W., and Schelvis, J. P. (2004) Phe393 mutants of cytochrome P450 BM3 with modified heme redox potentials have altered heme vinyl and propionate conformations. Biochemistry 43, 1798-1808 183. Graham-Lorence, S., Truan, G., Peterson, J. A., Falck, J. R., Wei, S., Helvig, C., and Capdevila, J. H. (1997) An Active Site Substitution, F87V, Converts Cytochrome P450 BM-3 into a Regio- and Stereoselective (14S,15R)-Arachidonic Acid Epoxygenase. J. Biol. Chem. 272, 1127-1135

184. Huang, W.-C., Westlake, A. C. G., Maréchal, J.-D., Joyce, M. G., Moody, P. C. E., and Roberts, G. C. K. (2007) Filling a Hole in Cytochrome P450 BM3 Improves Substrate Binding and Catalytic Efficiency. J. Mol. Biol. 373, 633-651 185. Li, H., and Poulos, T. L. (1997) The structure of the cytochrome p450BM-3 haem domain complexed with the fatty acid substrate, palmitoleic acid. Nat. Struct. Mol. Biol. 4, 140-146 186. Angus, B. C., and Luet-Lok, W. (2001) Protein engineering of Bacillus megaterium CYP102. Eur. J. Biochem. 268, 3117-3125 187. Ost, T. W. B., Miles, C. S., Munro, A. W., Murdoch, J., Reid, G. A., and Chapman, S. K. (2001) Phenylalanine 393 Exerts Thermodynamic Control over the Heme of Flavocytochrome P450 BM3. Biochemistry 40, 13421-13429 188. Ost, T. W. B., Munro, A. W., Mowat, C. G., Taylor, P. R., Pesseguiero, A., Fulco, A. J., Cho, A. K., Cheesman, M. A., Walkinshaw, M. D., and Chapman, S. K. (2001) Structural and Spectroscopic Analysis of the F393H Mutant of Flavocytochrome P450 BM3 Biochemistry 40, 13430-13438 189. Klein, M. L., and Fulco, A. J. (1993) Critical residues involved in FMN binding and catalytic activity in cytochrome P450BM-3. J. Biol. Chem. 268, 7553-7561 190. Neeli, R., Roitel, O., Scrutton, N. S., and Munro, A. W. (2005) Switching Pyridine Nucleotide Specificity in P450 BM3: Mechanistic analysis of the W1046H and W1046A enzymes. J. Biol. Chem. 280, 17634-17644 191. Yeom, H., Sligar, S. G., Li, H., Poulos, T. L., and Fulco, A. J. (1995) The role of Thr268 in oxygen activation of cytochrome P450BM-3. Biochemistry 34, 14733- 14740 192. Guengerich, F. P. (2002) Cytochrome P450 enzymes in the generation of commercial products. Nat. Rev. Drug. Discov. 1, 359-366 193. Beate, P., Zhibin, L., Richard, G., Manuela, A., Wolfgang, S., Karl, G., Herfried, G., and Anton, G. (2008) Efficient Biocatalytic Synthesis of (R)-Pantolactone. Adv. Synth. Catal. 350, 1943-1948 194. Swarts, J. W., Vossenberg, P., Meerman, M. H., Janssen, A. E. M., and Boom, R. M. (2008) Comparison of two-phase lipase-catalyzed esterification on micro and bench scale. Biotechnol. Bioeng. 99, 855-861 195. Panke, S., Witholt, B., Schmid, A., and Wubbolts, M. G. (1998) Towards a Biocatalyst for (S)-Styrene Oxide Production: Characterization of the Styrene Degradation Pathway of Pseudomonas sp. Strain VLB120. Appl. Environ. Microbiol. 64, 2032- 2043

196. van Hellemond, E. W., Janssen, D. B., and Fraaije, M. W. (2007) Discovery of a Novel Styrene Monooxygenase Originating from the Metagenome. Appl. Environ. Microbiol. 73, 5832-5839 197. Lu, Y., Yeung, N., Sieracki, N., and Marshall, N. M. (2009) Design of functional metalloproteins. Nature 460, 855-862 198. Loughlin, W. A. (2000) Biotransformations in organic synthesis. Bioresour. Technol. 74, 49-62 199. Edgardo , T. F., Ulrich, S., Anton, G., and Frances , H. A. (2001) Directed Evolution of a Cytochrome P450 Monooxygenase for Alkane Oxidation. Adv. Synth. Catal. 343, 601-606 200. Turner, N. J. (2009) Directed evolution drives the next generation of biocatalysts. Nat. Chem. Biol. 5, 568-574 201. Dodhia, V. R., Fantuzzi, A., and Gilardi, G. (2006) Engineering human cytochrome P450 enzymes into catalytically self-sufficient chimeras using molecular Lego. J. Biol. Inorg. Chem. 11, 903-916 202. Dennig, A., Lulsdorf, N., Liu, H., and Schwaneberg, U. (2013) Regioselective o- Hydroxylation of Monosubstituted Benzenes by P450 BM3. Angew. Chem. Int. Ed. Engl. (in press) 203. Agudo, R., Roiban, G. D., and Reetz, M. T. (2012) Achieving regio- and enantioselectivity of P450-catalyzed oxidative CH activation of small functionalized molecules by structure-guided directed evolution. ChemBioChem 13, 1465-1473 204. Ost, T. W., Miles, C. S., Murdoch, J., Cheung, Y., Reid, G. A., Chapman, S. K., and Munro, A. W. (2000) Rational re-design of the substrate binding site of flavocytochrome P450 BM3. FEBS Lett. 486, 173-177 205. Dietrich, M., Do, T. A., Schmid, R. D., Pleiss, J., and Urlacher, V. B. (2009) Altering the regioselectivity of the subterminal fatty acid hydroxylase P450 BM-3 towards gamma- and delta-positions. J. Biotechnol. 139, 115-117 206. Meinhold, P., Peters, M. W., Chen, M. M. Y., Takahashi, K., and Arnold, F. H. (2005) Direct Conversion of Ethane to Ethanol by Engineered Cytochrome P450 BM3. ChemBioChem 6, 1765-1768 207. Fasan, R., Chen, M. M., Crook, N. C., and Arnold, F. H. (2007) Engineered Alkane- Hydroxylating Cytochrome P450BM3 Exhibiting Nativelike Catalytic Properties. Angew. Chem., Int. Ed. 46, 8414-8418 208. Kawakami, N., Shoji, O., and Watanabe, Y. (2011) Use of Perfluorocarboxylic Acids To Trick Cytochrome P450BM3 into Initiating the Hydroxylation of Gaseous Alkanes. Angew. Chem., Int. Ed. 50, 5315-5318

209. Whitehouse, C. J. C., Bell, S. G., Yang, W., Yorke, J. A., Blanford, C. F., Strong, A. J. F., Morse, E. J., Bartlam, M., Rao, Z., and Wong, L.-L. (2009) A Highly Active Single- Mutation Variant of P450BM3 (CYP102A1). ChemBioChem 10, 1654-1656 210. Whitehouse, C. J., Yang, W., Yorke, J. A., Rowlatt, B. C., Strong, A. J., Blanford, C. F., Bell, S. G., Bartlam, M., Wong, L. L., and Rao, Z. (2010) Structural basis for the properties of two single-site proline mutants of CYP102A1 (P450BM3). ChemBioChem 11, 2549-2556 211. Park, S. H., Kim, D. H., Kim, D., Jung, H. C., Pan, J. G., Ahn, T., and Yun, C. H. (2010) Engineering bacterial cytochrome P450 (P450) BM3 into a prototype with human P450 enzyme activity using indigo formation. Drug Metab. Dispos. 38, 732- 739 212. Schwaneberg, U., Schmidt-Dannert, C., Schmitt, J., and Schmid, R. D. (1999) A continuous spectrophotometric assay for P450 BM-3, a fatty acid hydroxylating enzyme, and its mutant F87A. Anal. Biochem. 269, 359-366 213. Lussenburg, B. M., Babel, L. C., Vermeulen, N. P., and Commandeur, J. N. (2005) Evaluation of alkoxyresorufins as fluorescent substrates for cytochrome P450 BM3 and site-directed mutants. Anal. Biochem. 341, 148-155 214. Reinen, J., Ferman, S., Vottero, E., Vermeulen, N. P., and Commandeur, J. N. (2011) Application of a fluorescence-based continuous-flow bioassay to screen for diversity of cytochrome P450 BM3 mutant libraries. J. Biomol. Screen. 16, 239-250 215. Tsotsou, G. E., Cass, A. E. G., and Gilardi, G. (2002) High throughput assay for cytochrome P450 BM3 for screening libraries of substrates and combinatorial mutants. Biosensors and Bioelectronics 17, 119-131 216. Tsotsou, G. E., Di Nardo, G., Sadeghi, S. J., Fruttero, R., Lazzarato, L., Bertinaria, M., and Gilardi, G. (2013) A rapid screening for cytochrome P450 catalysis on new chemical entities: cytochrome P450 BM3 and 1,2,5-oxadiazole derivatives. J. Biomol. Screen. 18, 211-218 217. Ferrero, V. E., Di Nardo, G., Catucci, G., Sadeghi, S. J., and Gilardi, G. (2012) Fluorescence detection of ligand binding to labeled cytochrome P450 BM3. Dalton Trans. 41, 2018-2025 218. Di Nardo, G., Fantuzzi, A., Sideri, A., Panicco, P., Sassone, C., Giunta, C., and Gilardi, G. (2007) Wild-type CYP102A1 as a biocatalyst: turnover of drugs usually metabolised by human liver enzymes. J. Biol. Inorg. Chem. 12, 313-323 219. Weber, E., Seifert, A., Antonovici, M., Geinitz, C., Pleiss, J., and Urlacher, V. B. (2011) Screening of a minimal enriched P450 BM3 mutant library for hydroxylation of cyclic and acyclic alkanes. Chem. Commun. (Camb) 47, 944-946

220. Carmichael, A. B., and Wong, L.-L. (2001) Protein engineering of Bacillus megaterium CYP102. Eur. J. Biochem. 268, 3117-3125 221. Li, Q. S., Ogawa, J., Schmid, R. D., and Shimizu, S. (2001) Residue size at position 87 of cytochrome P450 BM-3 determines its stereoselectivity in propylbenzene and 3-chlorostyrene oxidation. FEBS Lett. 508, 249-252 222. Sowden, R. J., Yasmin, S., Rees, N. H., Bell, S. G., and Wong, L. L. (2005) Biotransformation of the sesquiterpene (+)-valencene by cytochrome P450cam and P450BM-3. Org. Biomol. Chem. 3, 57-64 223. Tsotsou, G. E., Sideri, A., Goyal, A., Nardo, G. D., and Gilardi, G. (2012) Identification of Mutant Asp251Gly/Gln307His of Cytochrome P450 BM3 for the Generation of Metabolites of Diclofenac, Ibuprofen and Tolbutamide. Chem. Eur. J. 224. Rea, V., Kolkman, A. J., Vottero, E., Stronks, E. J., Ampt, K. A., Honing, M., Vermeulen, N. P., Wijmenga, S. S., and Commandeur, J. N. (2012) Active site substitution A82W improves the regioselectivity of steroid hydroxylation by cytochrome P450 BM3 mutants as rationalized by spin relaxation nuclear magnetic resonance studies. Biochemistry 51, 750-760 225. Venkataraman, H., Beer, S. B., Bergen, L. A., Essen, N., Geerke, D. P., Vermeulen, N. P., and Commandeur, J. N. (2012) A single active site mutation inverts stereoselectivity of 16-hydroxylation of testosterone catalyzed by engineered cytochrome P450 BM3. ChemBioChem 13, 520-523 226. Dietrich, J. A., Yoshikuni, Y., Fisher, K. J., Woolard, F. X., Ockey, D., McPhee, D. J., Renninger, N. S., Chang, M. C. Y., Baker, D., and Keasling, J. D. (2009) A Novel Semi-biosynthetic Route for Artemisinin Production Using Engineered Substrate- Promiscuous P450BM3. ACS Chemical Biology 4, 261-267 227. Dubey, K. K., and Behera, B. K. (2011) Statistical optimization of process variables for the production of an anticancer drug (colchicine derivatives) through fermentation: at scale-up level. N. Biotechnol. 28, 79-85 228. Kitamura, E., Otomatsu, T., Maeda, C., Aoki, Y., Ota, C., Misawa, N., and Shindo, K. (2013) Production of Hydroxlated Flavonoids with Cytochrome P450 BM3 Variant F87V and Their Antioxidative Activities. Biosci. Biotechnol. Biochem. 77, 1340-1343 229. Shehzad, A., Panneerselvam, S., Linow, M., Bocola, M., Roccatano, D., Mueller- Dieckmann, J., Wilmanns, M., and Schwaneberg, U. (2013) P450 BM3 crystal structures reveal the role of the charged surface residue Lys/Arg184 in inversion of enantioselective styrene epoxidation. Chem Commun. (Camb) 49, 4694-4696 230. Maurer, S. C., Kühnel, K., Kaysser, L. A., Eiben, S., Schmid, R. D., and Urlacher, V. B. (2005) Catalytic Hydroxylation in Biphasic Systems using CYP102A1 Mutants. Adv. Synth. Catal. 347, 1090-1098

231. Pflug, S., Richter, S. M., and Urlacher, V. B. (2007) Development of a fed-batch process for the production of the cytochrome P450 monooxygenase CYP102A1 from Bacillus megaterium in E. coli. J. Biotechnol. 129, 481-488 232. Andrew , M. S., Michael , M. Y. C., Palaniappan, K., Ming-Shang, K., Horst, H., and Frances , H. A. (2009) A Panel of Cytochrome P450 BM3 Variants to Produce Drug Metabolites and Diversify Lead Compounds. Chem. Eur. J. 15, 11723-11729 233. Wong, T. S., Arnold, F. H., and Schwaneberg, U. (2004) Laboratory evolution of cytochrome p450 BM-3 monooxygenase for organic cosolvents. Biotechnol. Bioeng. 85, 351-358 234. Salazar, O., Cirino, P. C., and Arnold, F. H. (2003) Thermostabilization of a Cytochrome P450 Peroxygenase. ChemBioChem 4, 891-893 235. Li, Y., Drummond, D. A., Sawayama, A. M., Snow, C. D., Bloom, J. D., and Arnold, F. H. (2007) A diverse family of thermostable cytochrome P450s created by recombination of stabilizing fragments. Nat. Biotechnol. 25, 1051-1056 236. Cirino, P. C., and Arnold, F. H. (2003) A Self-Sufficient Peroxide-Driven Hydroxylation Biocatalyst. Angew. Chem., Int. Ed. 42, 3299-3301 237. Tran, N. H., Huynh, N., Chavez, G., Nguyen, A., Dwaraknath, S., Nguyen, T. A., Nguyen, M., and Cheruzel, L. (2012) A series of hybrid P450 BM3 enzymes with different catalytic activity in the light-initiated hydroxylation of lauric acid. J. Inorg. Biochem. 115, 50-56 238. Tran, N. H., Huynh, N., Bui, T., Nguyen, Y., Huynh, P., Cooper, M. E., and Cheruzel, L. E. (2011) Light-initiated hydroxylation of lauric acid using hybrid P450 BM3 enzymes. Chem. Commun. (Camb) 47, 11936-11938 239. Ener, M. E., Lee, Y. T., Winkler, J. R., Gray, H. B., and Cheruzel, L. (2010) Photooxidation of cytochrome P450-BM3. Proc. Natl. Acad. Sci. U.S.A 107, 18783- 18786 240. Celik, A., Sperandio, D., Speight, R. E., and Turner, N. J. (2005) Enantioselective epoxidation of linolenic acid catalysed by cytochrome P450(BM3) from Bacillus megaterium. Org. Biomol. Chem. 3, 2688-2690 241. Schewe, H., Kaup, B. A., and Schrader, J. (2008) Improvement of P450(BM-3) whole-cell biocatalysis by integrating heterologous cofactor regeneration combining glucose facilitator and dehydrogenase in E. coli. Appl. Microbiol. Biotechnol. 78, 55- 65 242. Harkey, A., Kim, H. J., Kandagatla, S., and Raner, G. M. (2012) Defluorination of 4- fluorophenol by cytochrome P450(BM(3))-F87G: activation by long chain fatty aldehydes. Biotechnol. Lett. 34, 1725-1731

243. Raner, G. M., Hatchell, A. J., Morton, P. E., Ballou, D. P., and Coon, M. J. (2000) Stopped-flow spectrophotometric analysis of intermediates in the peroxo-dependent inactivation of cytochrome P450 by aldehydes. J. Inorg. Biochem. 81, 153-160 244. Waltham, T. N., Girvan, H. M., Butler, C. F., Rigby, S. R., Dunford, A. J., Holt, R. A., and Munro, A. W. (2011) Analysis of the oxidation of short chain alkynes by flavocytochrome P450 BM3. Metallomics 3, 369-378 245. Shapiro, M. G., Westmeyer, G. G., Romero, P. A., Szablowski, J. O., Kuster, B., Shah, A., Otey, C. R., Langer, R., Arnold, F. H., and Jasanoff, A. (2010) Directed evolution of a magnetic resonance imaging contrast agent for noninvasive imaging of dopamine. Nat. Biotechnol. 28, 264-270 246. Lelyveld, V. S., Brustad, E., Arnold, F. H., and Jasanoff, A. (2011) Metal-substituted protein MRI contrast agents engineered for enhanced relaxivity and ligand sensitivity. J. Am. Chem. Soc. 133, 649-651 247. Brustad, E. M., Lelyveld, V. S., Snow, C. D., Crook, N., Jung, S. T., Martinez, F. M., Scholl, T. J., Jasanoff, A., and Arnold, F. H. (2012) Structure-guided directed evolution of highly selective p450-based magnetic resonance imaging sensors for dopamine and serotonin. J. Mol. Biol. 422, 245-262 248. Kaspera, R., Sahele, T., Lakatos, K., and Totah, R. A. (2012) Cytochrome P450BM-3 reduces aldehydes to alcohols through a direct hydride transfer. Biochem. Biophys. Res. Commun. 418, 464-468

2.0 Key mutations alter the cytochrome P450 BM3 conformational landscape and remove inherent substrate bias

1Christopher F. Butler, 2Caroline Peet, 1Amy E. Mason, 2Michael W. Voice, 1David Leys and 1Andrew W. Munro*

1Manchester Institute of Biotechnology, Faculty of Life Sciences, University of Manchester, 131 Princess Street, Manchester M1 7DN, UK. 2Cypex Ltd, 6 Tom McDonald Avenue, Dundee DD2 1NH, UK.

Running title: Diversification of P450 BM3 substrate selectivity

Keywords: cytochrome P450, omeprazole, conformational destabilization, enzyme engineering Background: P450 BM3 is a high activity enzyme with biotechnological potential. Results: Mutations perturbing P450 BM3’s conformational state and active site facilitate human P450-like oxidation of the drug omeprazole. Conclusion: Conformational destabilization enables P450 BM3 to explore novel conformations and accept diverse substrates. Significance: “Gatekeeper” mutations that decrease the energetic barrier for transition to the substrate-bound state can reconfigure P450 BM3 specificity and reactivity

2.1 SUMMARY

Cytochrome P450 monooxygenases (P450s) have enormous potential in production of oxychemicals, due to their unparalleled regio- and stereoselectivity. The Bacillus megaterium P450 BM3 enzyme is a key model system, with several variants (many distant from the active site) reported to alter substrate selectivity. It has the highest reported monooxygenase activity of the P450 enzymes and this catalytic efficiency has inspired protein engineering to enable its exploitation for biotechnologically relevant oxidations with structurally diverse substrates. However, a structural rationale is lacking to explain how these mutations have such effects in absence of direct change to the active site architecture. Here we provide the first crystal structures of BM3 variants in complex with a human drug substrate: the proton pump inhibitor omeprazole. Supported by solution data, these structures reveal how mutation alters the conformational landscape and decreases the free energy barrier for transition to the substrate-bound state. Our data point to the importance of such “gatekeeper” mutations in enabling major changes in substrate recognition. We further demonstrate that these variants catalyze the same 5-hydroxylation reaction as performed by human CYP2C19, the major human omeprazole metabolizing P450 enzyme.

2.2 Introduction

The cytochrome P450 monooxygenases (P450s) are hemoproteins that catalyse a huge range of biochemical transformations, including key reactions involved in the biosynthesis of steroids, antibiotics and signaling lipids (1,2). Recent years have seen intensive efforts to exploit the P450s’ ability to catalyze regio- and stereoselective oxidations of substrates in order to make biotechnologically useful products. Examples include the rational engineering of the Pseudomonas putida camphor hydroxylase P450cam (CYP101A1) for improved binding and oxidation of environmentally recalcitrant polychlorinated benzenes, and the use of directed evolution for the conversion of specificity of the Bacillus megaterium P450 BM3 (CYP102A1, BM3) from long chain fatty acids towards short chain hydrocarbons (3,4). BM3 has proven a particularly versatile and popular model system for use in engineering studies, benefiting from the fact that it is a natural fusion of a P450 to its mammalian-like diflavin reductase redox partner (enabling efficient electron transfer that underpins its high monooxygenase activity (5,6)), and that the structure of its P450 domain and roles of many active site amino acids are well understood (7-10).

Recent work to diversify BM3’s substrate selectivity and reactivity has produced variants that catalyze olefin cyclopropanation by carbene transfer and oxidation of e.g. testosterone, polycyclic aromatic hydrocarbons and pharmaceuticals (11-14). In the latter case, an aim is to engineer BM3 to generate high levels of metabolites typical of those formed by human P450s. Approaches to engineering BM3 have included directed evolution, chimeragenesis (with homologs from B. subtilis) and CASTing, as well as structure-led mutagenesis guided by X-ray crystal structures of BM3’s P450 (heme) domain in substrate-free (SF), fatty acid- bound (SB) and various variant forms (e.g. 7,8,15,16). Many studies identified common residues that help facilitate substrate diversification. These include Phe87, Ala82, Val78 and Arg47; the first three of which are internal residues, while Arg47 interacts with the fatty acid carboxylate at the protein surface (6,10). Phe87 interacts with the ω-end of fatty acids and prevents oxidation at this position. Phe87 mutations alter regioselectivity of fatty acid oxidation, with positions of hydroxylation of lauric acid moving from ω-1/ω-2/ω-3 (WT BM3) towards ω-5 in F87A/S/G variants, and even further from the ω-position in a F87A/V78A double mutant (6,17). The F87A mutation also activated BM3 in hydroxylation of testosterone (12).

While alterations that influence local structure in the BM3 active site cavity clearly have potential to alter binding position of fatty acids and to enable docking of novel substrates, other BM3 mutations were shown to have a more profound effects on the P450’s conformational landscape (18,19). One of the more common activity-altering BM3 variants

(A82F) shows greatly enhanced affinity for fatty acid substrates. While the crystal structure of palmitate-bound A82F heme domain was solved in the SB-type conformation (20), this did not provide a clear rationale for its much improved affinity for fatty acids. Also, despite several studies involving mutagenesis of BM3 to improve its capacity to oxidize human pharmaceuticals, there are no structural data to provide insight into how improved variants enable binding of the new substrates, or to guide subsequent engineering to further enhance binding and desired activities.

In this study we characterize the structural and biochemical/catalytic properties of the A82F, F87V and F87V/A82F variants of BM3, demonstrating a novel activity in oxidation of the widely used gastric proton pump inhibitor (PPI) omeprazole (Figure 1), and generating products typical of those formed by the major human metabolizing P450 enzyme CYP2C19 (21). Structural data for substrate-free and omeprazole-bound forms of the A82F and double mutant BM3 heme domains provide clear evidence for the “gatekeeper” nature of the A82F mutation, which produces a major structural rearrangement of the BM3 heme domain that leads to novel molecular selectivity. These first structural data for BM3 in complex with a human drug substrate highlight how combinations of conformational effector mutations (e.g. A82F) with secondary mutations that make local structural changes to alter binding in the heme vicinity (e.g. F87A/V) can be combined to cause dramatic changes in P450 substrate selectivity for biotechnological applications.

2.3 Experimental Procedures

2.3.1 Generation, expression and purification of WT and variant P450 BM3 proteins

Variants of the intact P450 BM3 and its heme domain were generated by oligonucleotide- directed mutagenesis. Intact WT P450 BM3 in pET15b was used for mutagenesis to create A82F, F87V and F87V/A82F (DM) variants. Mutations (positions underlined) were generated using the QuikChange Lightning site directed mutagenesis kit (Stratagene-Agilent UK). Primers used were A82F 5’-CTTAAATTTGTACGTGATTTTTTCGGAGACGGGTTA-3’, F87V 5’-TTGCAGGAGACGGGTTAGTTACAAGCTGGACGCATG-3’, F87V in A82F background 5’-TTTTCGGAGACGGGTTAGTTACAAGCTGGACGCATG-3’ (and their reverse complements). These intact BM3 enzymes were expressed as N-terminal hexahistidine tagged enzymes either using the pET15b (F87V, DM) constructs directly, or following cloning of the WT and A82F genes into pET14b using NdeI/BamHI sites. WT and variant heme domain genes were generated using the relevant pET14b/15b constructs. To generate the heme domain constructs, a stop codon pair (underlined) was introduced after residue 473 by PCR using the same mutagenesis kit, and with primers StopF 5'- CAGTCTGCTAAAAAAGTACGCAAATAGTAGGAAAACGCTCATAATACGCCGCTG-3' and StopR 5'- CAGCGGCGTATTATGAGCGTTTTCCTACTATTTGCGTACTTTTTTAGCAGACTG-3'. The heme domain genes (amino acids 1-473) were transferred as NdeI/BamHI fragments to pET20b to enable heme domain production in absence of a N-terminal His-tag for improved crystallization. Genes were sequenced to ensure presence of desired mutation(s) and absence of other mutations. The WT and A82F intact BM3, and the WT and all variant P450 BM3 heme domain proteins were expressed in BL21-Gold (DE3) E. coli cells (Stratagene- Agilent UK) using TB medium with cell growth at 37 °C and with agitation at 200 RPM in an orbital incubator. The F87V and DM intact BM3 proteins were grown in autoinduction TB medium (Melford Ltd, Ipswich UK). Typically, 4 L bacterial cultures were used for protein production with cell growth for 24-36 hours.

Following cell growth, bacterial cells were recovered by centrifugation at 4 °C (6000 g, 10 min) and resuspended in ice-cold buffer B (50 mM KPi, 250 mM NaCl, 10% (v/v) glycerol, pH 7.0) containing protease inhibitors (EDTA Free CompleteTM tablets, Roche, Germany). Protease inhibitors were maintained in all subsequent buffers used for protein purification. Cells were lysed by sonication on ice using a Bandelin Sonopuls sonicator (40% power, 50 pulses for 5 s with 25 s between pulses). The supernatant containing soluble intact BM3 and heme domain proteins was separated from cell debris by centrifugation (20,000 g, 40 min, 4

°C) and ammonium sulfate was added to 30% saturation on ice with slow stirring for ~4 h. Centrifugation (20,000 g, 40 min, 4 °C) was done to fractionate soluble BM3 and heme domain proteins from insoluble material. Intact BM3 WT and variant proteins in the supernatant were purified using His-tag affinity by mixing with Ni-IDA resin (Qiagen, UK) overnight at 4 °C in buffer B with 5 mM imidazole, prior to elution with 200 mM imidazole in buffer B. Isolated proteins were dialyzed into buffer A (50 mM Tris, 1 mM EDTA, pH 7.2) and further purified by size exclusion chromatography using a Sephacryl S-200 column (GE Healthcare, 26 x 60 cm on an AKTA purifier system). BM3 fractions were checked for purity by SDS-PAGE, concentrated by ultrafiltration (Vivaspin, Vivaproducts, USA) and frozen in buffer A plus 50% glycerol at -80 °C). For the non-tagged BM3 heme domains, supernatants subsequent to the 30% ammonium sulfate fractionation were subjected to a second ammonium sulfate fractionation step at 60% salt saturation. The pellet was resuspended in buffer A and extensively dialyzed into the same buffer to desalt, prior to loading onto a Q- Sepharose anion exchange column (16 x 10 cm, on an AKTA) and eluting using a gradient of 0-500 mM KCl in buffer A. Heme domain-containing fractions were desalted by using a desalting column (GE Healthcare, 26 x 10 cm on an AKTA) into 25 mM KPi pH 7.0, and then loaded onto a hydroxyapatite column (Bio-Rad, USA, 16 x 11 cm) and eluted in a linear gradient of 25 mM – 500 mM KPi, pH 7.0 (200 mL). Pure heme domain fractions were concentrated by ultrafiltration (Vivaspin) and used immediately for crystallography, or flash frozen in liquid nitrogen and stored at -80 °C. Due to the enhanced affinity for fatty acids in BM3 and heme domain proteins carrying the A82F mutation, all enzymes were passed through a Lipidex 1000 column (Perkin Elmer, UK) in 25 mM KPi, pH 7.0 to remove any fatty acid bound during purification, and prior to use for crystallization (heme domains), or for binding or turnover studies (heme domain and intact BM3 enzymes).

2.3.2 Quantification of P450 BM3 enzymes and determination of their substrate affinity and steady-state kinetic properties

Concentrations of the low-spin forms of the WT and variant forms of intact P450 BM3 and its heme domain were determined using extinction coefficients of ɛ = 105 and 95 mM-1 cm-1, respectively, at the Soret maximum (418 or 419 nm), as described previously (6). Thiolate coordination of the heme iron in all samples was established by formation of the Fe(II)CO complex at ~450 nm by bubbling of sodium dithionite-reduced WT and variant BM3 and heme domain samples (ca 2-4 µM) with carbon monoxide gas, as described by Omura and Sato (22). All samples showed near complete conversion to the P450 form, with negligible formation of an ~420 nm peak relating to the thiol-coordinated P420 form (23,24).

Dissociation constants (Kd values) for binding of the substrates N-palmitoylglycine (NPG) and omeprazole (OMP) to WT and variant BM3 heme domains were determined by UV- visible absorption titrations using ~1-4 µM protein in 100 mM KPi (pH 7.0) at 25 °C (assay buffer) in 1 cm pathlength quartz cuvettes, and as in our previous studies (6,16,25). Spectra were recorded for substrate-free enzymes, and following addition of ligands during substrate titrations (typically 800-250 nm). Titrations were recorded until no further spectral changes were observed in the P450s. Difference spectra were generated by subtraction of the spectrum for ligand-free protein from spectra recorded after each addition of substrate. Maxima and minima in difference spectra were identified (using the same wavelength pair in each titration) and the overall absorbance changes Δ(Apeak minus Atrough) were plotted versus [substrate]. Data were fitted using either a standard (Michaelis-Menten) hyperbolic function or (for tight binding substrates where the Kd value is ≤ 5x the P450 concentration) by using the Morrison equation (as described previously) to determine Kd values (26,27). UV-Vis spectroscopy was carried out on a Cary 50 UV-Vis spectrometer (Cary-Agilent, UK), with data analysis and fitting done using Origin Pro software (OriginLab, Massachusetts USA).

Steady-state kinetic parameters for WT and F87V, A82F and DM variants of intact BM3 were analyzed on a Cary 50 UV-Visible spectrophotometer using substrate (OMP and NPG)- -1 -1 dependent NADPH oxidation at 340 nm (Δɛ340 = 6210 M cm ) across a range of substrate concentrations, and using pure BM3 enzymes (25-125 nM) in a total of 1 mL assay buffer at 25 °C. NADPH was maintained at a saturating concentration (200 µM). Data points were collected in at least triplicate. Rate constants for substrate-dependent NADPH oxidation were plotted versus [substrate] and data were fitted using the Michaelis-Menten function

(using Origin Pro) to obtain kcat and KM parameters in each case. Vmax values for WT and variant BM3 enzymes with 5-OH OMP were measured similarly using a saturating concentration of the metabolite.

2.3.3 Omeprazole and 5-OH omeprazole turnover and analysis by LC-MS

Turnover reactions for oxidation of OMP were carried out in deep well blocks at 37 °C with shaking for 30 min. Reaction mixtures contained purified WT or variant (F87V, A82F or DM BM3) enzymes (0.1 µM), substrate (10 µM), NADPH regeneration system (glucose-6- phosphate, 7.76 mM; NADP+, 0.6 mM; and glucose-6-phosphate dehydrogenase, 0.75 U/ml) in turnover buffer (50 mM KPi, 5 mM CaCl2, pH 7.4) in a final volume of 500 µL. Following completion of the reaction, protein was precipitated by addition of an equal volume of acetonitrile (ACN) containing 1 µg/ml fluconazole by shaking the mixed samples at 800 rpm for 10 min. The precipitated protein was filtered through protein precipitation plates

(Phenomenex, Macclesfield UK) into mass spectrometry vials (FluidX, Nether Alderley UK) and clarified by centrifugation (4000 g, 25 min, 10 oC). Analysis was carried out on a Thermo Exactive LC-MS with a CTC PAL auto sampler (Thermo Scientific, UK) with a Kinetex 2.6U XB – C18 100A column (Phenomenex). A gradient of 0.1 % formic acid to ACN was used to resolve products. Reaction of the DM BM3 enzyme with 5-OH OMP and subsequent product analysis was done in the same way as for the OMP turnovers.

2.3.4 Omeprazole turnover and analysis by NMR

Turnover reactions with OMP were carried out in a 100 mL flask at 37 °C with shaking of reagents at 100 rpm for 30 min. Reaction mixtures contained purified WT or variant (F87V, A82F or DM intact BM3) enzymes (1 µM), substrate (100 µM), NADPH regeneration system (glucose-6-phosphate 7.76 mM, NADP+ 0.6 mM, and glucose-6-phosphate dehydrogenase 0.75 U/ml) in 60 mL assay buffer (100 mM KPi, pH 7.0). Products were extracted using

Strata-X SPE columns (Phenomenex), dried under vacuum and eluted in CDCl3. Analysis was carried out on a Bruker Avance 400 MHz NMR (Bruker, Coventry UK). 1H spectra were collected at 400 MHz and 13C spectra at 101 MHz. Spectra were baseline corrected and referenced to tetramethylsilane (TMS) standard by the residual non-deuterated solvent in the sample. δ values are in ppm, J values are in Hz. Full assignments were made by COSY, HMBC and HMQC methods. Signal splittings were recorded as singlet (s), doublet (d), doublet of doublets (dd) alpha beta system (AB) and multiplet (m). Processing was carried out using MestReNova Lite (Mestrelab Research, Santiago de Compostela, Spain) and ACD NMR Processor (Advanced Chemistry Development, Inc. [ACD Labs], Toronto, Canada).

2.3.5 Examination of hemoprotein stability by differential scanning calorimetry

DSC was carried out on a Microcal VP-DSC instrument. Data analysis was done using Microcal Origin software. The parameters used were: 20-80 oC temperature gradient, 90 oC/hour scan rate, 10 min prescan thermostat. Background scans were carried out with degassed assay buffer, saturated with OMP or NPG for the substrate-bound samples. Protein samples were prepared in assay buffer by extensive buffer exchange and dialysis. All samples were run using 20 µM protein and saturating substrate. Once two overlapping baseline scans were achieved, a degassed protein sample was run.

2.3.6 Crystallization of P450 BM3 heme domains and determination of protein structures

Crystallography was performed using the sitting drop method using a seeding protocol at 4 °C. Crystals obtained during initial screens for each variant were used to create microcrystal screen stocks and consecutive screens (Molecular Dimensions) were made with drops that consisted of 150 nL of WT or variant heme domain proteins (230 µM), 50 µL of seed stock and 200 nL well solution using a Mosquito liquid handling robot (TTP LabTech Ltd, Melbourn UK). For the OMP variant heme domain complex structures, proteins were saturated with racemic OMP ligand prior to crystallization. Ligands were titrated into heme domain samples until no further change in heme iron spin-state (towards high-spin) was observed. Thereafter, samples were concentrated by ultrafiltration in the presence of saturating ligand. Micro seeding was also used to produce diffraction quality crystals. Crystals were obtained under a range of conditions and flash-cooled in liquid nitrogen prior to data collection. The mother liquor was supplemented with 10% PEG 200 where an additional cryo-protectant was required. Data were collected at Diamond synchrotron beamlines and reduced and scaled using XDS (28). Structures were solved by molecular replacement with previously solved BM3 heme domain structures (PDB 1JPZ) using PHASER (29). Structures were refined using Refmac5 (29) and Coot (30). PDB codes for the new heme domain structures are A82F: 4KF0; F87V/A82F (DM, imidazole-bound): 4KF2; A82F (OMP-bound): 4KEW; and DM (OMP-bound): 4KEY.

2.4 Materials

Oligonucleotide primers were from Eurofins MWG Operon (Ebersberg, Germany). Omeprazole was from Cypex Ltd (Dundee, UK). 5-OH omeprazole was from Santa Cruz Biotechnology, Inc. (Dallas, USA). The bacterial growth medium (Terrific Broth, TB) was from Melford Ltd (Ipswich, UK). Unless otherwise stated, other chemicals were from Sigma- Aldrich and of the highest purity available.

2.5 Results

2.5.1 Characterization of omeprazole binding properties of BM3 variants

Preceding studies identified Ala82 and Phe87 as important in controlling molecular selectivity and regioselectivity of BM3 substrate oxidation (16,20,31). WT, F87V, A82F and the F87V/A82F double mutant (DM) forms of intact P450 BM3 and of its heme domain (residues 1-473) were expressed and purified as described in the Experimental Procedures. The BM3 ferric heme iron undergoes a shift from low-spin (LS) towards high-spin (HS) on binding substrates that displace its 6th ligand (a water molecule), accompanied by a shift of Soret absorption maximum from ~418 nm to ~392 nm (32). Binding of the substrate N- palmitoylglycine (NPG) was done for WT, F87V, A82F and the DM BM3 proteins. WT and all variants bound NPG tightly, with Kd values < 1 µM (Table 2.1).

Figure 2.1. The structure of (S)-omeprazole. The chemical structure of the proton pump inhibitor omeprazole (OMP) is shown. The pyridine ring shows the accepted numbering (hydroxylation occurs at the 5-methyl position). Also shown is the characteristic MS fragmentation position that gives the methoxybenzimidazole and 4-methoxy-3,5- dimethylpyridin-2-yl (pyridinyl) fragments. Hydroxylation on the 5-methyl group is performed by engineered variants of P450 BM3 described in this study. 5-hydroxylation is also the primary reaction catalyzed by the major human OMP-metabolizing enzyme CYP2C19. Omeprazole is chiral around the central sulfur atom. The S-isomer is shown, as a drug preparation, omeprazole is a racemate of two isomers.

The A82F variant has particularly high affinity for lipids and was purified from E. coli with fatty acid bound, as was the DM. Separation of the bound lipid by Lipidex chromatography

was done prior to analysis. Binding studies with the PPI omeprazole (OMP) revealed no evidence for its association with WT BM3, but partial spin-state conversion was observed for

F87V BM3, indicative of ability of the drug to access the active site (Kd = 49 µM). Much tighter binding and more extensive HS heme accumulation was observed with both A82F and the DM (Kd values = 1.67 and 0.21 µM, respectively), indicating that the mutations had synergistic effects in enhancing OMP affinity, with A82F having the major role (Figure 2.2A, Table 2.1).

-1 Protein Substrate Kd (µM) kcat (min ) Km (µM) WT NPG 0.082 ± 0.011 4770 ± 160 13.9 ± 2.7 A82F NPG 0.297 ± 0.069 5130 ± 570 26.3 ± 5.3 F87V NPG 0.204 ± 0.045 4970 ± 240 14.9 ± 2.8 DM NPG 0.004 ± 0.003 4050 ± 250 1.91 ± 0.31 WT OMP N/A 238 ± 12 124 ± 20 A82F OMP 1.67 ± 0.05 1460 ± 30 18.7 ± 0.8 F87V OMP 49.0 ± 2.7 2180 ± 20 38.7 ± 4.6 DM OMP 0.212 ± 0.014 1500 ± 45 1.93 ± 0.22

Table 2.1. Substrate binding and turnover data for BM3 enzymes. The table shows data for the binding of substrates (N-palmitoylglycine [NPG]; and omeprazole [OMP]), to WT and

A82F, F87V and DM P450 BM3 enzymes (Kd values from optical titrations) and for the kinetics of substrate-dependent NADPH oxidation (kcat, Km values). All data points were collected in triplicate, errors are SEM. N/A indicates that no evidence of binding of OMP to WT BM3 was found in optical titrations.

Figure 2.2: Binding and oxidation of omeprazole by P450 BM3 variants. Panel A. Binding titration for F87V/A82F (DM) intact P450 BM3 (1 µM) with omeprazole. Main panel:

Plot of the induced heme Soret absorption change (ΔA389 – ΔA419) versus [OMP] with data fitted to yield a Kd = 0.212 ± 0.014 µM; Inset: Selected OMP-induced absorption difference spectra from titration at OMP concentrations: 0.05 µM (green), 0.15 µM (blue), 0.30 µM (magenta), 0.40 µM (purple) and 1.0 µM (red). Panel B. Turnover data for OMP with WT (black column), A82F (red), F87V (blue) and F87V/A82F (DM, orange) P450 BM3 enzymes. Assays were done for 30 minutes. Products (5-hydroxy omeprazole [5-OH OMP] and 5- carboxy omeprazole [5-COOH OMP]) are shown as a percentage of the initial OMP concentration used in the assay, with data corrected for an internal standard (fluconazole) and for enzyme-independent degradation of OMP substrate.

2.5.2 Steady-state kinetics of P450 BM3 variants with omeprazole

Steady-state kinetic analysis was done for WT and each of the intact BM3 variants by following NADPH oxidation on addition of various concentrations of NPG or OMP (24). Table -1 1 details high kcat values for all BM3 enzymes with NPG (4050-5130 min ), with DM BM3 having the lowest Km (1.91 µM). With OMP, some substrate-stimulated NADPH oxidation -1 was seen for WT BM3 (kcat = 238 min ), but kinetics were substantially improved in all the variants (e.g. 1500 min-1 for DM), and apparent OMP affinity for the DM was ~65-fold greater than for WT BM3 (Km values of 1.9 µM vs. 124 µM), resulting in a >400-fold improvement in catalytic efficiency (kcat/Km ratio) for the DM over WT BM3.

2.5.3 Oxidation of omeprazole by WT and variant P450 BM3 enzymes

To validate novel omeprazole oxidase activity in variant BM3 enzymes, in vitro turnover studies were done as described in detail in the Experimental Procedures, using LCMS and NMR to characterize products formed. WT BM3 exhibited very small amounts of oxidation of OMP (< 1% of the starting material), but the low amounts obtained precluded determination of position(s) of oxidation. However, each of the three variants extensively oxidized the drug.

LC-MS revealed a 362.1162 amu species in all the variant turnover reactions, corresponding to introduction of an oxygen atom into OMP (Figure 2.3). A natural fragmentation of the OMP molecule occurs in the MS, with bond breakage between the sulfur and the methoxybenzimidazole moiety (Figure 1). A +16 increase in mass of the larger fragment indicated that oxidation occurs on this portion of OMP. Further fragmentation of the oxidized molecule showed that the position of oxidation was likely on one of the two 4-methoxy-3,5- dimethylpyridin-2-yl (hereafter termed pyridinyl) methyl groups (Figure 2.3). This was confirmed by 1H NMR spectroscopy after larger scale reactions of variant enzymes with OMP, and the absolute position of oxidation was confirmed using 2D NMR – showing that a specific hydroxylation occurred on the pyridinyl 5-methyl group (forming 5-OH OMP) (Supplemental Data Figures 2.S1-2.S4). This same reaction is also catalyzed by CYP2C19, the major metabolizing P450 for OMP and other PPIs in humans (33).

Figure 2.3. LC-MS analysis of products derived from omeprazole oxidation by the P450 BM3 F87V/A82F (DM) double mutant enzyme. The figures show data from LC-MS studies of OMP before and after its enzymatic turnover by the P450 BM3 DM (F87V/A82F) enzyme. These data demonstrate hydroxylation and subsequent oxidation of OMP, and also the fragmentation of the OMP (and its oxidized products) that occurs during MS analysis. Panel A (retention time [RT] = 5.39 min) shows data for OMP prior to addition of enzyme and initiation of its oxidation by the BM3 DM enzyme. Peaks at m/z 346.1212 and 198.0581 (circled) correspond to the fragmentation of OMP at the sulfone group (between the sulfur and the methoxybenzimidazole moiety), with the smaller species representing the sulfur-

containing fragment. Panel B (RT = 5.26 min) is following an enzymatic reaction for 30 minutes. The m/z peaks at 362.1162 and 214.0530 (circled) are for the 5-OH OMP and its hydroxylated fragment. Panel C (RT = 5.32 min) is following an enzymatic reaction for 30 minutes. The m/z peaks at 376.0955 and 228.0323 (circled) are for the 5-COOH OMP and its carboxylated fragment.

In OMP turnovers done over 30 min, the total oxidative turnover of OMP was greatest for the DM (~55%), followed by the F87V (~50%) and the A82F (~20%) variants (Figure 2.2B). LCMS demonstrates the formation of considerable amount of a +32 species (~10%) in the case of the DM, which was shown to be the carboxylic acid product at the pyridinyl 5-methyl group (5-COOH OMP), resulting from further P450-mediated oxidation at the same position. By comparing the amount of NADPH oxidized with the quantity of OMP oxidized over the first two minutes of reactions, enzymatic coupling was estimated as 68% for both the A82F and DM BM3 enzymes, and 39% for the F87V variant. These data suggest that structural changes induced by the A82F mutation are the major determinant of the productive binding mode for OMP, and thus the coupling efficiency.

Consistent with the model of further BM3 variant-catalyzed oxidation of the primary product,

5-OH OMP showed tight binding to both A82F and the DM (Kd values of 15.2 ± 0.9 µM and 2.62 ± 0.19 µM, respectively) (Table 2.2, Figure 2.4), indicating that oxidized products do bind the enzyme. The further oxidation of 5-OH OMP to 5-COOH OMP was confirmed by

LCMS studies with the BM3 DM enzyme (data not shown). Vmax values determined for 5-OH OMP-stimulated NADPH oxidation in the BM3 variants were 1050 ± 50 min-1 (A82F), 645 ± 30 min-1 (F87V) and 785 ± 35 min-1 (DM), but NADPH oxidation was not stimulated significantly above background in the WT BM3 (Table 2.2). A time course for the oxidation of OMP into 5-OH OMP and 5-COOH OMP products by the BM3 DM enzyme is shown in Figure 5. Approximately 30% of the 5-OH OMP is further oxidized to 5-COOH OMP by the DM BM3 over 30 minutes.

-1 Protein Substrate Kd (µM) Vmax (min )

WT 5-OH OMP ND ND

A82F 5-OH OMP 15.2 ± 0.9 1050 ± 50

F87V 5-OH OMP ND 645 ± 30

DM 5-OH OMP 2.62 ± 0.19 785 ± 35

Table 2.2. Binding and kinetics of oxidation of 5-OH OMP by WT and variant P450 BM3 enzymes. The table shows Kd values (where determinable) for the binding of the oxidized product 5-OH OMP to WT and variant intact P450 BM3 enzymes. The Kd values were derived from optical titrations. Also shown are Vmax values for 5-OH OMP-dependent NADPH oxidation catalyzed by WT, F87V, A82F and DM BM3 enzymes. All data points were collected in triplicate, errors are SEM. 5-OH OMP is the primary metabolite of OMP generated by human CYP2C19 and by the BM3 variants (most efficiently by the F87V/A82F DM). Data were collected as described in the Experimental Procedures section.

Figure 2.4. Optical binding titration for the BM3 DM with 5-OH OMP. Panel A shows UV- visible binding spectra for a titration of intact DM BM3 (~1.0 µM, red spectrum) and following addition of 2 µM (blue), 6 µM (magenta), 16 µM (orange) and 40 µM (black) 5-OH OMP. The Soret band shifts from 418 nm to 393 nm on binding 5-OH OMP. The inset shows difference spectra obtained by subtraction of the substrate-free spectrum from each of the shown 5- OH-bound spectra (color coding remains same). Panel B shows a plot of induced Soret absorbance change (ΔA389 nm minus ΔA421 nm) versus the relevant [5-OH OMP], with data fitted using the Morrison equation to give a Kd value of 2.62 ± 0.19 µM (26). 5-OH OMP binds both A82F and DM heme domains to induce a substrate-like shift in heme iron spin-

state equilibrium towards the HS state. Full Kd and Vmax data for 5-OH OMP binding/turnover with WT and variant BM3 enzymes are given in Table 2.2.

Figure 2.5. Time course of substrate oxidation and product formation in the reaction of the P450 BM3 DM enzyme with omeprazole. OMP substrate is shown in black squares, and the products 5-OH OMP and 5-carboxy OMP (5-COOH OMP) are shown in open circles and open triangles, respectively. Reactions were done as described in the Experimental Procedures section. The reactions reach completion in ~10-15 minutes, with most substrate oxidation (and 5-OH OMP formation) occurring in the first 2.5 minutes.

2.5.4 Structural analysis of omeprazole-binding P450 BM3 variants

The effect of the F87V mutation in enabling OMP binding may be explained using available structural information for the BM3 heme domain, as space vacated by the F87V substitution directly above the heme plane is likely to allow binding of the bulky substrate (6). The marked effects of the A82F mutation, however, are not intuitive given the peripheral position of A82 to the substrate binding pocket.

To gain detailed understanding of the effects of the A82F mutation in promoting OMP binding, both the A82F and the F87V/A82F double variant were co-crystallized with OMP, and structures solved by molecular replacement using the NPG-bound heme domain structure (PDB code 1JPZ) (34). Crystal structure PDB codes are 4KEW for the OMP-bound A82F heme domain, and 4KEY for DM heme domain. In both cases, the global conformation

100

of the heme domain is remarkably similar to previously determined fatty-acid bound structures (0.3 Å over 450 Cα atoms) (8,34) (Figure 2.6A, Table 2.3).

Figure 2.6. Structures of P450 BM3 enzymes and their omeprazole binding sites. A. Comparison of WT and variant BM3 heme domain structures. The FG helices are colored, while the remainder of the protein structures is depicted in grayscale. The A82F mutation is shown in spheres (where present) and substrate molecules are shown in atom-colored spheres. The heme is shown as red sticks. Panel 1: The F87V/A82F (DM) P450 BM3 variant heme domain in complex with OMP. Panel 2: DM heme domain in the ligand-free form. Panel 3: WT heme domain complex with NPG (PDB code 1JPZ) (34). Panel 4: WT heme domain in the ligand-free form (PDB code 1BU7) (57). B. The mode of binding of OMP is

101

shown for the DM (left panel) and A82F (right panel) variant BM3 heme domain active sites. Due to weak electron density, the labile sulfone oxygen is omitted from the models shown. Key residues contacting the ligand are shown as sticks and water molecules hydrogen bonding to the OMP are in red. In the right panel, the OMP from the A82F heme domain structure is overlaid with that from the DM heme domain. In the right panel the Phe82 residues are shown in green for the DM and in cyan for the A82F variant. The DM Val87 is in green and the A82F Phe87 is in cyan. The distance between the P450 heme iron and the OMP 5-methyl group is 3.9 Å in the A82F heme domain, and 4.1 Å in the DM heme domain.

102

A82F DM:imidazole A82F:OMP DM:OMP

(4KF0) (4KF2) (4KEW) (4KEY)

Space group P21 P21 P212121 P212121

a=59.1Å a=59.3Å a=59.4Å a=59.3Å Cell b=147.2Å b=151.7Å b=129.5Å b=130.7Å parameters c=64.0Å c=60.8Å c=145.1Å c=146.0Å beta=97.5˚ beta=95.9˚

47-1.82 (1.86- 65-1.89 55-2.05 Resolution (Å) 64-1.45 (1.5-1.45) 1.82) (1.94-1.89) (2.1-2.05)

Rmerge (%) 7.8 (35.5) 6.8 (47.2) 8.4 (43.8) 8.5 (39.2)

I/sigI 16.2 (2.1) 14.4 (1.9) 11.8 (3.2) 11.4 (3.0)

14.5/17.8 18.2/22.2 18.8/23.6 18.7/23.5 R/Rfree (%) (28.1/32.3) (27.6/31.3) (28.5/33.6) (24.2/29.2)

Average B (Å2) 16.8 22.2 19.7 30.5

Rmsd 0.025Å/2.04˚ 0.024Å/1.91˚ 0.022Å/1.88˚ 0.022Å/1.79˚ bonds/angles

25% 15% PEG20K, 15% PEG4K, 15% PEG4K, PEG2000MME, 15% 0.2 M MgCl2, 0.2 M MgCl2, Crystallization 0.2 M MgCl2, PEG550MME pH 6.5 pH 6.5 conditions pH 6.5 (0.1 M 0.06 M MgCl2, (0.1 M sodium (0.1 M sodium sodium pH 6.5 (0.1 M cacodylate) cacodylate) cacodylate) imidazole/MES)

Table 2.3. Data reduction and final structural refinement statistics for P450 BM3 variants and their OMP-substrate complexes.

The PPI ligand is clearly identified from the electron density occupying the substrate binding channel for the OMP complex. In both cases, electron density corresponding to the oxygen atom of the central sulfinyl groups is weak, and it is possible both stereoisomers are present from the racemic mixture. OMP was also shown to be relatively labile, with loss of oxygen from the sulfone noted in aqueous solution (35) (Figure 6B). We thus modeled OMP without

103

the sulfone oxygen. The pyridinyl moiety is placed in a near-perpendicular orientation directly above the heme plane. The close distance to the heme of the OMP 5-methyl group, in particular, leads to displacement of the water 6th ligand, and this conformation is consistent with the observed 5-OH OMP product. The heme iron to OMP 5-methyl group distances are 3.9 Å and 4.1 Å in the A82F and DM heme domains, respectively. A single direct hydrogen bond is made between the backbone carbonyl of Leu437 and one of the methoxybenzimidazole nitrogens. Water molecules are present in close proximity to the other OMP polar groups, and these also mediate a network of hydrogen bonds between protein and ligand. Specific interactions are made via a bridging water molecule between (i) the Ala74 backbone nitrogen and the methoxy oxygen of the methoxybenzimidazole; and (ii) the hydroxyl oxygen of Ser72 and the second methoxybenzimidazole nitrogen. While there are no substantial differences in the orientation of the methoxybenzimidazole group for both variants, close contacts between Phe87 and the pyridinyl moiety in the A82F variant structure leads to small reorientation of the substrate in comparison to the DM structure. In the latter, the reduced bulk of the Val87 appears to provide sufficient space for the substrate to adopt a less strained conformation. This likely explains the difference in affinity observed between the A82F and DM variants. Other protein-OMP interactions in the A82F and DM complex structures involve hydrophobic interactions between residues near both the pyridinyl (e.g. Thr438, Ile263, Ala328) and methoxybenzimidazole (e.g. Val26, Leu188) ends of the OMP substrate (Figure 2.7).

104

Figure 2.7. Interactions of omeprazole in the active site of the A82F BM3 heme domain. The diagram shows the binding site of omeprazole (OMP, without the labile sulfinyl oxygen) in the A82F variant BM3 heme domain. For OMP, carbon atoms are shown in black, oxygens in red, sulfur in yellow, nitrogens in blue and the oxygens of water molecules in cyan. Bonds in the OMP substrate are shown in purple, and bonds in selected amino acids are in brown. Hydrogen bonds are shown (with their lengths) as green dashed lines. Amino acids making hydrophobic interactions with the OMP are shown as red arcs with radiating lines. OMP atoms involved in these hydrophobic interactions are shown with radiating red lines. A direct hydrogen bond interaction is made between the backbone carbonyl of Leu437 and one of the OMP benzimidazole group nitrogens (NE1). A further bridging hydrogen bond occurs from the Ser72 hydroxyl group through a water molecule (water 781) to the other benzimidazole nitrogen (NV). A final bridging hydrogen bond interaction occurs between the backbone nitrogen of Ala74 and the benzimidazole methoxy oxygen (O3) via another water molecule (water 761). A number of hydrophobic protein-OMP interactions are seen. These include interactions with Leu188 at the benzimidazole methoxy group (C4), and with Ala328 at the pyridinyl 5-methyl group (C1). The diagram was produced using Ligplot+ using the structure of the A82F heme domain-omeprazole complex solved in this study (PDB code 4KEW) (58).

105

A clear rationale for the marked difference in OMP affinity between A82F containing variants and the WT BM3 is not provided by the variant-OMP structures. Figure 2.8A shows the active sites in detail for the overlaid A82F OMP-bound BM3 heme domain and the NPG- bound WT heme domain. The additional bulk of the Phe82 side chain only interacts edge-on with the pyridinyl moiety of OMP, and the only significant change in the positions of nearby key residues occurs for Phe87. The Phe87 side chain is seen in different conformations in the A82F/OMP structure, compared to the WT BM3/NPG heme domain. Figure 2.8B shows an alternative view of the active site for the DM/OMP-bound heme domain and the WT/NPG- bound P450, looking along the I helix, with the F/G helices highlighted. This reinforces the strong structural similarity between the WT and DM variant substrate-bound structures. All these structures occupy the substrate-bound (SB) conformation. However, the ligand-free A82F and DM heme domains also have the SB conformation (Figure 2.6A) However, the WT BM3 heme domain occupies a distinct conformation when substrate-free (SF), and thus the A82F mutation is key to shifting the equilibrium between SF and SB conformations. Previous studies also showed that the structural state of the heme domain can be significantly affected by single mutations at key positions (36).

106

Figure 2.8. Stereoviews of structural overlays of substrate-bound forms of the BM3 A82F-containing variant heme domains with WT BM3. Panel A shows a stereoview of the A82F-OMP heme domain active site (in red) with that of the WT-NPG structure (1JPZ, in blue) (34). Key amino acids are shown in lines while the bound ligands are shown in atom colored sticks (OMP with magenta carbons; NPG with light blue carbons). Besides the nature of the ligand itself, and the obvious difference of the A82F mutation, there are very few structural differences between the structures, and these are mainly limited to Phe87 occupying multiple conformations in the A82F-OMP structure. Panel B shows an alternative view of the BM3 double mutant (DM, F87V/A82F) OMP-bound heme domain structure overlaid with the NPG substrate-bound structure of the WT BM3 heme domain. The F/G helix region is colored in red for the BM3 DM and in blue for the substrate-bound WT BM3. Key amino acid residues and the respective ligands are shown as sticks. As also seen for panel A, surprisingly little change can be observed in the DM protein structure compared to WT BM3, despite the distinct nature of the ligand and the introduction of two mutations.

107

We also determined the crystal structures of the ligand-free and imidazole-bound A82F (PDB code 4KF0) and DM (4KF2) heme domains. In the latter case, imidazole present in the crystallization buffer ligated to the heme iron. These structures share a similar conformation that is again distinct from previously observed WT BM3 heme domain SF structures (Figure 2.6A, Figure 2.9). These data again point to the A82F substitution producing dramatic changes in the structure and behavior of ligand-free A82F-containing variants. The most significant changes occur in the positioning of the FG helices, which adopt a conformation that places the FG-loop further away from the protein core, leading to increased mobility. This altered position is a consequence of a reorganization of the hydrophobic contact between the FG-helices and the I-helix. In the WT SF conformation, the N-terminal region of the I-helix adopts a slightly different orientation compared to the SB conformation. In the case of the A82F-containing variants, this motion leads to a close contact between Ile263 and Phe82 (Figure 2.9). This results in a minor reorientation of the I-helix and the Ile263 side chain, which directly affects the position of the G-helix residue Met177. The repositioning of the G-helix is accompanied by reorientation of hydrophobic residues on the B- and F-helices.

108

Figure 2.9. Structural overlay of the omeprazole-bound A82F variant with the WT BM3 heme domain. A stereoview is shown for a structural overlay of the A82F BM3 heme domain with the WT heme domain (1BU7). Color coding is as in Figure 2.6, with F/G helices in green for the substrate-free A82F variant and in yellow for the substrate-free WT BM3 heme domain.

While the F87V variant enables binding of OMP detectable by heme absorbance shift, the Kd for OMP binding to F87V BM3 heme domain is ~30-fold and 230-fold weaker than that observed for the A82F and DM heme domains, respectively. The improved ligand binding properties of the A82F variants can be understood in view of the large changes introduced by the A82F mutation in the ligand-free, but not ligand-bound, structures. The shift in the BM3 conformational equilibrium from SF to SB (as induced by ligand binding) is dependent on the free energy difference between both conformations, and the free energy associated with ligand binding. The A82F-containing variant crystal structures suggest that the mutation leads to a substantially altered free energy difference between both conformations. The fact OMP only appears to bind with measurable affinity to A82F-containing variants suggests that the A82F mutation destabilizes the substrate-free conformation.

To confirm this hypothesis, we performed DSC analysis of the substrate-free and OMP- bound forms of WT and all variant heme domains; and compared data with NPG-bound

109

forms. The results showed that both F87V and A82F mutations diminished thermal stability of the BM3 heme domain. The WT heme domain has two unfolding transitions (Tm values) at

65.7 °C (Tm1, major) and 59.0 °C (Tm2, minor). The Tm values are not altered significantly by

OMP, but NPG stabilizes the P450 (Tm1 = 70.4 °C). Consistent with our model, the A82F

Tm’s are 59.7 °C and 49.7 °C – substantially lower than WT BM3. F87V was also destabilized, albeit to a lesser extent, with the minor transition no longer seen (Tm1 = 61.3

°C). The DM showed greatest destabilization (Tm’s = 50.9 °C and 58.0 °C). Binding of OMP to A82F-containing variants resulted in single unfolding transitions, with negligible change in

Tm1 for A82F (59.6 °C) and an ~2 °C stabilization for the DM (59.9 °C). A similar effect was seen for F87V heme domain (Tm = 63.1 °C) (Supplemental Data Table 2.S1). Thus, A82F has the major effect on stability of BM3, but F87V also destabilizes the protein and there is an additive effect on combining the mutations (Figure 2.10, Supplemental Data Figure 2.S5, Table 2.S1).

110

Figure 2.10. Conformational equilibria and the relationship with structural stability in P450 BM3. A. DSC data for the WT and DM P450 BM3 heme domains in substrate-free, OMP- and NPG-bound forms. B. A schematic overview of the conformational equilibria proposed for the BM3 WT and DM variant heme domains. Individual conformational states (as represented by crystal structures) are depicted as rectangles, gray shaded when largely unpopulated, and in color (color coded to match panel A) when significantly populated. The y-axis indicates the relative Tm values for the unfolding of these proteins.

111

2.6 Discussion

The development of efficient biocatalysts that generate human drug metabolites has become an area of great interest. Human P450s are responsible for most phase I xenobiotic metabolism, producing numerous oxidized and other metabolites from human drugs (37). For omeprazole, the principal metabolic pathway is 5-hydroxylation catalyzed by CYP2C19, the product of which can then be converted to the sulfone by CYP3A4. Oxidized metabolites of human drugs are required for pharmaceutical compliance, for metabolite safety testing, and for molecular interaction studies as required by FDA guidelines. Use of human P450s to produce small quantities of specific metabolites is a possibility, but caveats include slow reaction rates, requirement for a separate redox partner, and enzyme instability. Synthetic chemistry to make oxidized metabolites is an alternative, but controlling regioselectivity of oxidation, maintaining stability of compounds, and requirements for several steps giving low yield are major issues. An attractive alternative is to engineer high activity microbial P450s for specific oxidation of drugs.

A system of choice is P450 BM3, due to its catalytically self sufficient nature, high turnover rates and availability of excellent structural data to guide protein engineering. Crystal structures of substrate-free and fatty acid-bound forms of the wild-type BM3 heme domain drove early protein engineering studies on BM3 (7,8). This enabled identification of residues important in binding the substrate carboxylate at the mouth of the active site (Arg47 and Tyr51), as well as amino acids crucial for regulating heme iron redox potential (Phe393) and the coupling of electron transfer to substrate oxidation (Thr268) (6,38-40). Parallel studies documented its fast rates of fatty acid substrate oxidation, resulting from rapid electron transfer from NADPH through its fused cytochrome P450 reductase (CPR) partner (5,41), and more recent studies have confirmed that the enzyme is functional as a dimer (42,43), and that electron transfer to drive catalysis occurs between the CPR module of one monomer and the heme domain of the other in the BM3 dimer (44).

The application of random mutagenesis, recombination and directed evolution approaches (pioneered by Frances Arnold’s group) brought a new dimension to research on BM3, demonstrating that radical changes in its substrate selectivity could be engineered. A key development was the production of the 139-3 BM3 variant (containing 11 mutations), which enabled the hydroxylation of a range of alkanes from octane through to propane, including cyclohexane (45). However, only one mutation (V78A) occurred at a position known to be in contact with fatty acid from preceding structural data (8). While mutations at Ala82 and Phe87 were not present in the 139-3 variant, the addition of an A82L mutation to 139-3

112

resulted in increased coupling of NADPH oxidation to the hydroxylation of propane and octane, indicating improvements to the productive binding of these substrates. A related engineered BM3 variant (named 9/10A and containing 13 mutations, 8 of which are shared with 139-3) was inactive in ethane hydroxylation, but was shown to develop ethane hydroxylase activity and to improve propane hydroxylation when an A82S mutation was incorporated, together with either two or four further mutations (46). A related propane hydroxylase variant (35E11) containing 16 mutations (mostly accumulated from the 139-3 and 9/10A progenitors) also has the A82S mutation, although propane hydroxylation rate and coupling were improved on introduction of further groups of mutations that included L188P (47). The later structural data for the 139-3 variant heme domain in complex with NPG provided insights into structural changes that could promote binding and oxidation of short chain alkanes, while retaining a conformational state (SB) similar to the WT-NPG complex structure (34). In particular, active site mutation V78A enlarged a hydrophobic pocket close to the heme, while A184V enabled interactions with Leu437 across the active site channel, likely enabling van der Waals contacts with the terminal carbon of octane and protecting the substrate from solvent (48). Predictive modelling based on the 139-3 structure indicated that A82S and other mutations altered substrate channel organization, but that the destabilizing L188P mutation was a major determinant in promoting efficient propane oxidation, and likely acted by inducing structural disruption at the end of the F-helix, favouring a conformational change towards a catalytically productive form (48). Given the conservative nature of the A82S mutation, it is unclear whether this mutation alone might favour the SB conformation (as does the A82F mutation in our work). However, the SB structural conformation might be expected in any case for the NPG-bound enzyme.

The fact that the structurally destabilizing L188P mutation has profound effects on BM3 binding/turnover of propane has parallels in recent studies by Luet Wong’s group – in which (i) a substrate-free I401P variant (the mutated residue being adjacent to the heme proximal ligand Cys400) was crystallized in a SB-like conformation, and (ii) an A330P variant in an inter-helical beta sheet part of the P450 caused structural disruption that reshaped the active site cavity. For both these variants, large increases in catalytic activity with non-natural substrates such as toluene, propylbenzene and 3-methylpentane were observed (18,49).

While some structural data are available for BM3 variants promoting short chain/alkane binding and oxidation, there is also much interest in the application of BM3 variants for the production of metabolites of human drugs. BM3 variants have been shown to oxidize drugs such as diclofenac, ibuprofen and acetaminophen, but no structural data were obtained (50,51). In this study, we present the first structure of a BM3 P450 bound to a human drug substrate (OMP), and show that a single active site mutation (A82F) is sufficient to facilitate

113

tight binding of this PPI drug, enabling the same oxidative reaction (hydroxylation at the pyridinyl 5-methyl group) as catalyzed by the main human metabolizing P450 CYP2C19. OMP binding is further enhanced by introduction of the F87V mutation in the immediate vicinity of the heme, avoiding close contacts between Phe87 and the OMP. Despite the obvious differences between the natural fatty acid substrates and OMP, the conformation of the OMP-bound variants is near identical to previously determined WT fatty acid complexes. This suggests BM3 active site architecture is largely determined by the overall protein conformation (SB vs. SF) rather than the exact nature of the ligand. Studies of human urinary metabolites of OMP showed that the 5-OH OMP and 5-COOH OMP were the major derivatives detected, with the former predominant (35,52). The relatively tight binding of the primary metabolite (5-OH OMP) to both A82F and (particularly) the DM BM3 reinforces that the oxidized derivative retains ability to bind in a catalytically competent mode to A82F- containing variants, consistent with its further oxidation to 5-COOH OMP by the DM BM3 (Figure 2.4, Table 2.2). The BM3 DM’s ability to catalyze successive OMP oxidations to generate 5-COOH OMP points to the possibility that CYP2C19 and/or other human CYPs might further oxidize 5-OH OMP to 5-COOH OMP.

Early structural studies of BM3’s heme domain identified Phe87 as a key active site residue, mutations of which altered binding and regioselectivity of fatty acid oxidation (6,10,17), and both Phe87 and Ala82 mutations are frequently contained in BM3 variants generated by directed evolution and other strategies for altered substrate specificity (e.g. 13,46). Other studies on Phe87 variants have further highlighted its importance in controlling access to the heme center and in altering reactivity. Notable recent studies have highlighted that an F87V BM3 variant catalyzes hydroxylation of both testosterone and progesterone. In the case of testosterone, 2β-hydroxytestosterone and 16β-hydroxytestosterone were formed in roughly equal amounts by F87V BM3. However, triple variants that also contained the A82F mutation (F87V, A82F, V78L/T/I) showed 3-4 fold greater catalytic efficiency and produced the 16β- hydroxytestosterone product at ~90%, suggesting an important influence of the A82F mutation on steroid binding mode (12). Other work has pointed to the importance of Phe87 in tolerance to solvents used for substrate solubilization. Crystal structure data showed that DMSO was able to compete with the heme’s 6th ligand water molecule for coordination to the iron in the F87A heme domain variant, but not in the WT heme domain at the same DMSO concentration (53).

While the influence of Phe87 mutations on accessibility to the heme center is easily rationalized from crystal structure data, the mechanism by which Ala82 mutations diversify BM3 substrate selectivity has been relatively poorly understood. Huang et al. solved the crystal structure of a palmitate-bound A82F heme domain, after purifying the variant from E.

114

coli in the lipid-bound form (20). The structure was similar to that of the NPG-bound WT heme domain (34). Our studies of the substrate-free A82F structure reveal that the mutation induces a change in the protein conformational landscape, leading to a distinct and previously unobserved position of the FG helices. These structural changes have profound consequences for substrate binding in the A82F variants, and point to a role for this residue as a “gatekeeper” for preferred substrate access as a consequence of its regulation of the conformational state and/or dynamics of structural change in the P450. Thus, although the mutation is removed from the immediate vicinity of the heme iron, it induces structural reconfiguration that enables the binding of OMP, and evidently that of other molecules. Our DSC studies show that the A82F mutation (in particular) alters the thermodynamic stability of

BM3, decreasing the Tm for protein unfolding. This suggests that its new conformational “flexibility” is a key facet underpinning its diversification of substrate selectivity, at least in part due to removal of inherent substrate bias towards long chain fatty acids by decreasing the free energy barrier corresponding to the transition to the SB-state (Figure 2.10B). These data are consistent with conclusions drawn by Bloom et al. with respect to the capacity of protein destabilizing mutations to impart novel substrate selectivity/reactivity on an enzyme (54).

In conclusion, our data point to new lessons to be learned from the outcomes of previous studies of BM3 (and other enzymes) in that amino acid changes that affect the dynamics of proteins in unpredictable ways may lead to mutations “distant” from the active site (or otherwise appearing not to impact significantly on substrate binding/affinity) having profound effects on enzyme catalysis and substrate specificity. The “gatekeeper” hypothesis thus points to Ala82 as a crucial target residue in research to enable further engineering of BM3 for diverse functions. Our findings show that, contrary to previous approaches focusing on increasing enzyme stability as a path to biotechnologically relevant enzymes (55,56), enzyme conformational destabilization is key to reducing the thermodynamic barrier to substrate binding and therefore to altered enzymatic activities, which could enable rapid identification of P450 (and other enzyme) variants with biotechnologically important activities.

115

2.7 Supplemental Data

2.7.1 Characterization of substrate and oxidized products by NMR spectroscopy

Nuclear Magnetic Resonance (NMR) Spectroscopy was used for the analysis of the OMP substrate (Figures 2.S1 and 2.S2) and the main monohydroxylated products from OMP turnover. Confirmation of the position of oxidation of OMP at the 5-methyl group was obtained from a combination of 1H, 13C and 2D NMR (Figures 2.S3 and 2.S4).

1H, 13C, HMQC (Heteronuclear Multiple Quantum Correlation) and HMBC (Heteronuclear Multiple Bond Correlation) experiments were used to characterize the starting materials and products. The 1H NMR omeprazole spectrum is shown below in Figure 2.S1 and was assigned as: (CDCl3 400MHz) δ = 8.15 (S, 1H), 7.45 (Broad S, 1H), 7.19 (Broad S, 1H), 6.88 (dd, J = 8.84, 1.53 Hz, 1H), 4.66 (AB, J = 13.64, 8.72 Hz, 2H), 3.78 (S, 3H), 3.57 (S, 3H), 13 2.16 (S, 3H), 2.08 (S, 3H). C NMR (CDCl3 101MHz) δ = 164.5, 149.7, 148.7, 127.1, 126.3, + 60.6, 59.8, 55.7, 13.3, 11.5. LCMS [M+H] 346.1213 (C17H20N3O3S). Assignment of the omeprazole spectrum was made using the HMBC spectrum (Figure 2.S2). Upon enzymatic turnover with the BM3 F87V and DM enzymes (less products were formed with the A82F variant, although LC-MS clearly showed formation of the 5-OH OMP product at the same retention time), a mixture of starting material and monohydroxylated product was identified. The product was identified as 5-OH OMP and the additional peaks seen in Figure 2.S3 were 1 assigned: H NMR (CDCl3 400MHz) δ = 8.27 (S, 1H), 7.45 (Broad S, 1H), 6.88 (d, J = 2.40 Hz, 1H), 6.85 (dd, J = 8.84, 2.27 Hz, 1H), 4.65 (AB, J = 3.92, 1.26 Hz, 2H), 4.59 (S, 2H), 13 3.75 (S, 3H), 3.56 (S, 3H), 2.02 (S, 3H). C NMR (CDCl3 101MHz) δ = 164.2, 150.4, 148.2, + 129.3, 127.2, 61.2, 60.2, 58.6, 11.5. LCMS [M+H] 362.1162 (C17H20N3O4S). Assignment of the 5-OH OMP spectrum was again made using the HMBC spectrum (Figure 2.S4).

116

Figure 2.S1. 1H NMR spectrum of omeprazole. The 1H spectrum for the OMP starting material is shown with peaks labelled and integrated. The overlay shows the S-Omeprazole (actual drug is racemic mixture) structure with accepted (non IUPAC) numbering, with peaks labelled accordingly. Data were collected on a 400 MHz NMR in CDCl3, corrected to TMS by residual non-deuterated solvent.

117

Figure 2.S2. HMBC spectrum of omeprazole. The spectrum shows the long range coupling of 1H to 13C nuclei. Coupling is observed between the 3-methyl group and the AB system; and between the 5-methyl group and the pyridinyl methoxy and pyridinyl hydrogen.

Data were collected on a 400 MHz NMR in CDCl3, corrected to TMS by residual non- deuterated solvent.

118

Figure 2.S3. 1H NMR spectrum of turnover products from omeprazole oxidation. Products were generated in the reaction of the F87V BM3 with OMP substrate. The overlay shows the 5-OH S-Omeprazole (drug is racemic mixture) structure with accepted (non IUPAC) numbering, and with peaks labelled accordingly. Only the product peaks are labelled and integrated for clarity. The spectrum shows the generation of a methoxy peak at δ 4.59 and a new methyl peak at δ 2.02, indicative of hydroxylation at one of the OMP methyl groups. The downfield shift of the pyridinyl hydrogen signal and the lack of shift in the AB system indicates hydroxylation at the 5 position. Data were collected on a 400 MHz NMR in

CDCl3, corrected to TMS by residual non-deuterated solvent.

119

Figure 2.S4. HMBC spectra of turnover products from omeprazole oxidation. Products were generated in the reaction of the F87V BM3 with OMP substrate. The data show the long range coupling of 1H to 13C nuclei. Additional peaks generated (by comparison with panel B above) show coupling of the new methoxy peak (δ 4.59) to the pyridinyl methoxy (δ 3.54), the pyridinyl hydrogen (δ 8.27), and to the new methyl peak (δ 2.02). Lack of coupling to the AB system confirms that hydroxylation occurs on the 5-methyl (and not the 3-methyl) group. Data were collected on a 400 MHz NMR in CDCl3, corrected to TMS by residual non- deuterated solvent.

120

2.7.2 Differential Scanning Calorimetry (DSC)

DSC was carried out as detailed in the Experimental Procedures section of the main paper. Data analysis was carried out using Microcal Origin software. Prior to data analysis, a reference baseline trace (using either the buffer alone for substrate-free samples, or buffer plus substrate at the working concentration for substrate-bound samples) was subtracted from the relevant BM3 heme domain data set, and then the data were normalized for protein concentration. Fitting was carried out using a standard non 2-state function. Transition midpoint (Tm), calorimetric enthalpy (ΔHcal) and Van’t Hoff enthalpy (ΔHVH) data were collected and are tabulated below (Table 2.S1). The WT BM3 and variant heme domains either had a single unfolding transition, or a main unfolding transition preceded by a minor transition at lower temperature. For comparison purposes, the Tm of the main unfolding event is used in all cases. The Tm value for the substrate (NPG)-bound F87V heme domain was determined from an incomplete data set, as exothermic aggregation took place prior to the complete unfolding of the protein in this case. The data show that WT BM3 heme domain is extensively stabilized by NPG, though omeprazole has little effect. The A82F heme domain is stabilized a little by NPG, but omeprazole also has little effect. The F87V heme domain is stabilized a little by omeprazole, but more by NPG. The DM heme domain is stabilized to an almost equal extent by both substrates. All DSC data sets are overlaid in Figure 2.S5.

121

Protein ΔH1 (kcal ΔH 1 (kcal T 2 ΔH2 (kcal ΔH 2 (kcal T 1 (oC) VH m VH (ligand) m mol-1) mol-1) (oC) mol-1) mol-1) 65.73 1.17 E5 1.77 E5 58.95 5.12 E4 1.41 E5 WT (SF) (0.03) (1.61 E3) (2.73 E3) (0.08) (1.69 E3) (4.93 E3) WT 70.39 1.31 E5 1.80 E5 64.36 7.47 E4 1.13 E5 (NPG) (0.03) (2.95 E3) (3.14 E3) (0.14) (3.13 E3) (3.74 E3) WT 65.44 8.91 E4 1.88 E5 59.28 3.76 E4 1.24 E5 (OMP) (0.07) (4.06 E3) (8.04 E3) (0.34) (4.37 E3) (1.37 E4) A82F 59.66 1.68 E5 1.72 E5 49.72 3.07 E4 9.53 E4 (SF) (0.02) (1.28 E3 (1.55 E3) (0.19) (1.59 E3) (6.16 E3) A82F 62.26 1.89 E5 1.68 E5 ------(NPG) (0.04) (3.11 E3) (3.46 E3) A82F 59.58 1.07 E5 1.93 E5 ------(OMP) (0.02) (8.26 E2) (1.87 E3) F87V 61.33 1.12 E5 1.67 E5 ------(SF) (0.02) (8.17 E2) (1.53 E3) F87V 65.15 1.98 E5 1.68 E5 ------(NPG)* (0.10) (8.46 E3) (8.85 E3) F87V 63.10 8.8 E4 (8.48 1.55 E5 ------(OMP) (0.03) E2) (1.87 E3) 58.04 1.88 E5 1.88E5 (5.83 50.86 1.12 E4 1.31 E5 DM (SF) (0.01) (4.99 E2) E2) (0.13) (5.58 E2) (7.68 E3) DM 60.35 1.95 E5 1.60 E5 ------(NPG) (0.04) (3.14 E3) (3.22 E3) DM 59.86 1.40 E5 1.96 E5 ------(OMP) (0.02) (1.19 E3) (2.08 E3)

Table 2.S1. DSC data for thermal unfolding of WT and variant BM3 heme domains in ligand-free and substrate-bound forms. The thermal Transition midpoint (Tm), calorimetric enthalpy (ΔH) and Van’t Hoff enthalpy (ΔHVH) data for WT and variant BM3 heme domains in substrate-free and OMP- and NPG-bound forms are shown. *Indicates that values presented were calculated using data up to the point at which exothermic aggregation was observed for the F87V-NPG complex (close to the end of its thermal unfolding transition). The data reveal thermal destabilization of the variant forms by comparison to WT BM3 heme domain. The BM3 DM heme domain is stabilized by OMP to a similar extent as it is by NPG.

122

Figure 2.S5. Graphical overlay of DSC data for WT and variant BM3 heme domains. The thermal unfolding profiles are shown for the various substrate-free, NPG- and OMP- bound forms of WT and variant (F87V, A82F and DM) BM3 heme domains. The relevant Tm values for the transitions observed (and accompanying thermodynamic data) are shown in Table 2.S1.

123

2.8 References

1. Guengerich, F. P. (2005) Human cytochrome P450 enzymes. Pp. 377-530 in Cytochrome P450: Structure, Mechanism and Biochemistry (3rd edition). Ed. Ortiz de Montellano, P. R., Kluwer Academic/Plenum Press, New York

2. Munro, A. W., Girvan, H. M., and McLean, K. J. (2007) Variations on a (t)heme – novel mechanisms, redox partners and catalytic functions in the cytochrome P450 superfamily. Nat. Prod. Rep. 24, 585-609

3. Xu, F., Bell, S. G., Rao, Z., and Wong, L. L. (2007) Structure-activity correlations in pentachlorobenzene oxidation by engineered cytochrome P450cam. Prot. Des. Sel. 20, 473-480

4. Peters, M. W., Meinhold, P., Glieder, A., and Arnold, F. H. (2003) Regio- and enantioselective alkane hydroxylation with engineered cytochromes P450 BM-3. J. Am. Chem. Soc. 125, 13342-13450

5. Munro, A. W., Daff, S., Coggins, J. R., Lindsay, J. G., and Chapman, S. K. (1996) Probing electron transfer in flavocytochrome P-450 BM3 and its component domains. Eur. J. Biochem. 239, 403-409

6. Noble, M. A., Miles, C. S., Chapman, S. K., Lysek, D. A., MacKay, A. C., Reid, G. A., Hanzlik, R. P., and Munro, A. W. (1999) Roles of key active-site residues in flavocytochrome P450 BM3. Biochem. J. 339, 371-379

7. Ravichandran, K. G., Boddupalli, S. S., Hasemann, C. A., Peterson, J. A., and Deisenhofer, J. (1993) Crystal structure of hemoprotein domain of P450BM-3, a prototype for microsomal P450’s. Science 261, 731-736

8. Li, H., and Poulos, T. L. (1997) The structure of the cytochrome P450BM-3 haem domain complexed with the fatty acid substrate, palmitoleic acid. Nat. Struct. Biol. 4, 140-146

9. Munro, A. W., Leys, D., McLean, K. J., Marshall, K. R., Ost, T. W., Daff, S., Miles, C. S., Chapman, S. K., Lysek, D. A., Moser, C. C., Page, C. C., and Dutton, P. L. (2002) P450 BM3: The very model of a modern flavocytochrome. Trends Biochem. Sci. 27, 250-257

10. Whitehouse, C. J., Bell, S. G., and Wong, L. L. (2012) P450BM3 (CYP102A1): connecting the dots. Chem. Soc. Rev. 41, 1218-1260

124

11. Coelho, P. S., Brustad, E., Kannan, A., and Arnold, F. H. (2013) Olefin cyclopropanation via carbene transfer catalyzed by engineered cytochrome P450 enzymes. Science 339, 307-310

12. Kille, S., Zilly, F. E., Acevedo, J. P., and Reetz, M. T. (2011) Regio- and stereoselectivity of P450-catalysed hydroxylation of steroids controlled by laboratory evolution. Nat. Chem. 3, 738-743

13. Carmichael, A. B., and Wong, L. L. (2001) Protein engineering of Bacillus megaterium CYP102. The oxidation of polycyclic aromatic hydrocarbons. Eur. J. Biochem. 268, 3117-3125

14. Sawayama, A. M., Chen, M. M., Kulanthaivel, P., Kuo, M. S., and Arnold, F. H. (2009) A panel of cytochrome P450 BM3 variants to produce drug metabolites and diversify lead compounds. Chemistry 15, 11723-11729

15. Reetz, M. T. (2011) Laboratory evolution of stereoselective enzymes: a prolific source of catalysts for asymmetric reactions. Angew. Chem. Int. Ed. 50, 138- 174

16. Ost. T. W., Miles, C. S., Murdoch, J., Cheung, Y., Reid, G. A., Chapman, S. K., and Munro, A. W. (2000) Rational re-design of the substrate binding site of flavocytochrome P450 BM3. FEBS Lett. 486, 173-177

17. Dietrich, M., Do, T. A., Schmid, R. D., Pleiss, J., and Urlacher, V. B. (2009) Altering the regioselectivity of the subterminal fatty acid hydroxylase P450 BM-3 towards δ and ɛ positions. J. Biotechnol. 139, 115-117

18. Whitehouse, C. J., Yang, W., Yorke, J. A., Rowlatt, B. C., Strong, C. F., Bell, S. G., Bartlam, M., Wong, L. L., and Rao, Z. (2010) Structural basis for the properties of two

single-site proline mutants of CYP102A1 (P450BM3). Chembiochem 11, 2549-2556

19. Joyce, J. G., Girvan, H. M., Munro, A. W., and Leys, D. (2004) A single mutation in cytochrome P450 BM3 induces the conformational rearrangement seen upon substrate binding in the wild-type enzyme. J. Biol. Chem. 279, 23287-23293

20. Huang, W.-C., Westlake, A. C., Maréchal, J. D., Joyce, M. G., Moody, P. C., and Roberts, G. C. (2007) Filling a hole in cytochrome P450 BM3 improves substrate binding and catalytic efficiency. J. Mol. Biol. 373, 633-651

21. Ibeanu, G. C., Ghanayem, B. I., Linko, P., Li, L., Pederson, L. G., and Goldstein, J. A. (1996) Identification of residues 99, 220 and 221 of human cytochrome P450 2C19 as key determinants of omeprazole activity. J. Biol. Chem. 271, 12496- 12501

125

22. Omura, T., and Sato, R. (1964) The carbon monoxide-binding pigment of liver microsomes. II. Solubilization, purification, and properties. J. Biol. Chem. 239, 2379- 2385

23. Perera, R., Sono, M., Sigman, J. A., Pfister, T. D., Lu, Y., and Dawson, J. H. (2003) Neutral thiol as a proximal ligand to ferrous heme iron: implications for heme proteins that lose cysteine thiolate on reduction. Proc. Natl. Acad. Sci. U. S. A. 100, 3641- 3646

24. Dunford, A. J., McLean, K. J., Sabri, M., Seward, H. E., Heyes, D. J., Scrutton, N. S., and Munro, A. W. (2007) Rapid P450 heme iron by laser photoexcitation of Mycobacterium tuberculosis CYP121 and CYP51B1. Analysis of CO complexation reactions and reversibility of the P450/P420 equilibrium. J. Biol. Chem. 282, 24816- 24824

25. Driscoll, M. D., McLean, K. J., Levy, C., Mast, N., Pikuleva, I. A., Lafite, P., Rigby, S. E., Leys, D., and Munro, A. W. (2010) Structural and biochemical characterization of Mycobacterium tuberculosis CYP142: Evidence for multiple cholesterol 27- hydroxylase activities in a human pathogen. J. Biol. Chem. 285, 38270-38382

26. Morrison, J. F. (1969) Kinetics of the reversible inhibition of enzyme-catalysed reactions by tight-binding inhibitors. Biochim. Biophys Acta 185, 269-286

27. Bui, S. H., McLean, K. J., Cheesman, M. R., Bradley, J. M., Rigby, S. E., Levy, C. W., Leys, D., and Munro, A. W. (2012) Unusual spectroscopic and ligand binding properties of the cytochrome P450-flavodoxin fusion enzyme XplA. J. Biol. Chem. 287, 19699-19714

28. Kabsch, W. J. (1993) Automatic processing of rotation diffraction data from crystals of initially unknown symmetry and cell constants. J. Appl. Crystallogr. 26, 795-800

29. Collaborative Computation Project No. 4. (1994) Acta Crystallogr. D Biol. Crystallogr. 50, 760-763

30. Emsley, P., and Cowtan, K. (2004) Coot: model-building tools for molecular graphics. Acta Crystallogr. D Biol. Crystallogr. 60, 2126-2132

31. Venkataraman, H., Beer, S. B., Bergen, L. A., Essen, N. V., Geerke, D. P., Vermeulen, N. P., and Commandeur, J. N. (2012) A single active site mutation inverts stereoselectivity of 16-hydroxylation of testosterone catalyzed by engineered cytochrome P450 BM3. Chembiochem 13, 520-523

126

32. Miles, J. S., Munro, A. W., Rospendowski, B. N., Smith, W. E., McKnight, J. E., and Thomson, A. J. (1992) Domains of the catalytically self-sufficient cytochrome P-450 BM-3. Genetic construction, overexpression, purification and spectroscopic characterization. Biochem. J. 288, 503-509

33. Desta, Z., Zhao, X., Shin, J. G., and Flockhart, D. A. (2002) Clinical significance of the P4502C19 genetic polymorphism. Clin. Pharmacokinet. 41, 913-958

34. Haines, D. C., Tomchick, D. R., Machius, M., and Peterson, J. A. (2001) Pivotal role of water in the mechanism of P450BM-3. Biochemistry 40, 13456-13465

35. Renberg, L., Simonsson, R., and Hoffmann, K.-J. (1989) Identification of two main urinary metabolites of [14C]omeprazole in humans. Drug Metab. Dispos. 17, 69-76

36. Girvan, H. M., Seward, H. E., Toogood, H. S., Cheesman, M. R., Leys, D., and Munro, A. W. (2007) Structural and spectroscopic characterization of P450 BM3 mutants with unprecedented P450 heme iron ligand sets. New heme ligation states influence conformational equilibria in P450 BM3. J. Biol. Chem. 282, 564-572

37. Guengerich, F. P. (2007) Cytochrome P450 and chemical toxicology. Chem. Res. Toxicol. 21, 70-83

38. Li, H., and Poulos, T. L. (1999) Fatty acid metabolism, conformational change, and electron transfer in cytochrome P450 BM-3. Biochim. Biophys. Acta 1441, 141-149

39. Ost, T. W., Miles, C. S., Munro, A. W., Murdoch, J., Reid, G. A., and Chapman, S. K. (2001) Phenylalanine 393 exerts thermodynamic control over the heme of flavocytochrome P450 BM3. Biochemistry 40, 13421-13429

40. Yeom, H., Sligar, S. G., Li, H., Poulos, T. L., and Fulco, A. J. (1995) The role of Thr268 in oxygen activation of cytochrome P450 BM-3. Biochemistry 34, 14733- 14740

41. Narhi. L. O., and Fulco, A. J. (1986) Characterization of a catalytically self-sufficient 119,000-dalton cytochrome P-450 monooxygenase induced by barbiturates in Bacillus megaterium. J. Biol. Chem. 261, 7160-7169

42. Neeli, R., Girvan, H. M., Lawrence, A., Warren, M. J., Leys, D., Scrutton, N. S., and Munro, A. W. (2005) The dimeric form of flavocytochrome P450 BM3 is catalytically functional as a fatty acid hydroxylase. FEBS Lett. 579, 5582-5588

127

43. Kitazume T., Haines, D. C., Estabrook, R. W., Chen, B., and Peterson, J. A. (2007) Obligatory intramolecular electron-transfer from FAD to FMN in dimeric P450BM-3. Biochemistry 46, 11892-11901

44. Girvan, H. M., Dunford, A. J., Neeli, R., Ekanem, I. S., Waltham, T. N., Joyce, M. G., Leys, D., Curtis, R. A., Williams, P., Fisher, K., Voice, M. W., and Munro, A. W. (2011) Flavocytochrome P450 BM3 mutant W1046A is a NADH-dependent fatty acid hydroxylase: implications for the mechanism of electron transfer in the P450 BM3 dimer. Arch. Biochem. Biophys. 507, 75-85

45. Glieder, A., Farinas, E. T., and Arnold, F. H. (2002) Laboratory evolution of a soluble, self-sufficient, highly active alkane hydroxylase. Nat. Biotechnol. 20, 1135-1139

46. Meinhold, P., Peters, M. W., Chen, M. M. Y., Takahashi, K., and Arnold, F. H. (2005) Direct conversion of ethane to ethanol by engineered cytochrome P450 BM3. Chembiochem 6, 1765-1768

47. Fasan, R., Chen, M. M., Crook, N. C., and Arnold, F. H. (2007) Engineered alkane- hydroxylating cytochrome P450BM3 exhibiting nativelike catalytic properties. Angew. Chem. Int. Ed. Engl. 46, 8414-8418

48. Fasan, R., Meharenna, Y. T., Snow, C. D., Poulos, T. L., and Arnold, F. H. (2008) evolutionary history of a specialized P450 propane monooxygenase. J. Mol. Biol. 383, 1069-1080

49. Whitehouse, C. J. C., Bell, S. G., Yang, W., Yorke, J. A., Blanford, C. F., Strong, A. J. F., Morse, E. J., Bartlam, M., Rao, Z., and Wong, L.-L. (2009) A highly active single-mutation variant of P450BM3 (CYP102A1). Chembiochem 10, 1654-1656.

50. Tsotsou, G. E., Sideri, A., Goyal, A., DiNardo, G., and Gilardi, G. (2012) Identification of mutant Asp251Gly/Gln307His of cytochrome P450 BM3 for the generation of metabolites of diclofenac, ibuprofen and tolbutamide. Chemistry 18, 3582-3588

51. Damsten, M. C., van Vugt-Lussenburg, B. M., Zeldenthuis, T., de Vlieger, J. S., Commandeur, J. N., and Vermeulen, N. P. (2008) Application of drug metabolising mutants of cytochrome P450 BM3 (CYP102A1) as biocatalysts for the generation of reactive metabolites. Chem. Biol. Interact. 171, 96-107

52. Äbelö, A., Andersson, T. B., Antonsson, M., Naudot, A. K., Skånberg, I., and Weidorf, L. (2000) Stereoselective metabolism of omeprazole by human cytochrome P450 enzymes. Drug Metab. Dispos. 28, 966-972

128

53. Kuper, J., Tee, K. L., Wilmanns, M., Roccatano, D., Schwaneberg, U., and Wong, T. S. (2012) The role of active-site Phe87 in modulating the organic co-solvent tolerance of cytochrome P450 BM3 monooxygenase. Acta Crystallogr. Sect. F Struct Biol. Cryst. Commun. 68, 1013-1017

54. Bloom, J. D., Labthavikul, S. T., Otey, C. R., and Arnold, F. H. (2006) Protein stability promotes evolvability. Proc. Natl. Acad. Sci. U.S.A. 103, 5869-5874

55. Romero, P. A., Krause, A., and Arnold, F. H. (2013) Navigating the protein fitness landscape with Gaussian processes. Proc. Natl. Acad. Sci. U.S.A. 110, 193-201

56. Li, Y., Drummond, D. A., Sawayama, A. M., Snow, C. D., and Bloom, J. D. (2007) A diverse family of thermostable cytochrome P450s created by recombination of stabilizing fragments. Nat. Biotechnol. 25, 1051-1056

57. Sevrioukova, I. F., Li, H., Zhang, H., Peterson, J. A., and Poulos, T. L. (1999) Structure of a cytochrome P450 redox partner electron transfer complex. Proc. Natl. Acad. Sci. U.S.A. 96, 1863-1868

58. Laskowski R. A., and Swindells M. B. (2011) Ligplot+: multiple ligand-protein interaction diagrams for drug discovery. J. Chem. Inf. Model. 51, 2778-2786

129

FOOTNOTES

The authors acknowledge funding from the UK Biotechnology and Biological Sciences Research Council (BBSRC: research grant BB/F00883X1 to AWM/DL supporting AEM; Industrial CASE studentship BB/G01698/1 with Cypex Ltd to AWM/MWV, supporting CFB). The authors acknowledge the following University of Manchester staff: Dr. Colin Levy for assistance with synchrotron X-ray data collection, Dr. Tom Jowitt for assistance with DSC studies, and Dr. Robert Šardzík for helpful discussions on NMR analysis.

The abbreviations used are: BM3, flavocytochrome P450 BM3; CPR, NADPH-cytochrome P450 reductase; CYP, cytochrome P450; DM, BM3 F87V/A82F double mutant; DSC, differential scanning calorimetry; HS, high-spin ferric heme iron; LS, low-spin ferric heme iron; NPG, N-palmitoylglycine; OMP, omeprazole; PPI, proton pump inhibitor; SB, substrate bound; SF, substrate free; 5-OH OMP, 5-hydroxy omeprazole; 5-COOH OMP, 5-carboxy omeprazole.

130

3.0 Human P450-like oxidative transformations of proton pump inhibitor drugs by a P450 BM3 variant that induces conformational reconfiguration of the enzyme.

1Christopher F. Butler, 2Caroline Peet, 1Amy E. Mason, 1Kirsty J. Mclean, 1Karl Fisher, 2Michael W. Voice, 1Steve E Rigby, 1David Leys and 1Andrew W. Munro*

1Manchester Institute of Biotechnology, Faculty of Life Sciences, University of Manchester, 131 Princess Street, Manchester M1 7DN, UK. 2Cypex Ltd, 6 Tom McDonald Avenue, Dundee DD2 1NH, UK.

131

3.1 Abstract

The chemical production of drug metabolites is an area where enzyme chemistry has significant advantages over synthetic chemistry methods. These high value products are often complex to synthesise and offer a niche market for biotechnology companies. The vast majority of drugs are metabolised by the cytochrome P450 monooxygenases (P450s), with the reactions catalysed usually being highly regio- and stereoselective. The class of compounds known as proton pump inhibitors (PPIs) are a group of drugs that are extensively metabolised by the human P450s, producing a diverse array of metabolites according to the specific substrate. Here we show that a single mutation (A82F) to P450 BM3 from Bacillus megaterium causes a major alteration in its substrate selectivity such that the set of PPI molecules become good substrates in A82F and F87V/A82F variants. The substrate specificity switch is analysed by drug binding, enzyme kinetic and organic product analysis data to confirm these altered activities, and X-ray crystallography is used to provide a structural basis for the binding of esomeprazole to the F87V/A82F enzyme. These studies confirm that single “gatekeeper” mutations (exemplified by A82F) in P450 BM3 can produce major perturbations to P450 enzyme substrate selectivity and enable novel reactions that are typical of those performed by the human P450s. Efficient transformation of a range of PPI drugs to human-like metabolites by BM3 variants provides a new route to production of these compounds.

132

3.2 Introduction

The interest in replacing traditional synthetic routes to complex molecules with enzyme catalysed reactions results from the many advantages offered. These include lower energy requirements, fewer steps in the synthesis, and the ability to activate centres that are not possible by traditional chemical synthesis methods. The most promising area for this research is the production of fine chemicals (1,2) and drug metabolites (3,4), as these are complex molecules to produce and expensive to prepare chemically, and therefore ideal candidates for the application of enzymatic synthesis given their high value as standards and as reagents in drug testing (5) for the biotechnological and pharmaceutical industries. Cytochromes P450 (P450s) are a class of monooxygenases that are appealing candidates for biotechnological use, due to their ability to activate oxygen and to insert an oxygen atom into unactivated CH bonds. P450s are responsible for numerous natural reactions and are increasingly exploited for unnatural reactions with potential uses in e.g. medicine, bioremediation and pharmacology. These include hydroxylation of fatty acids/steroids, epoxidation of styrene, dealkylation of various drugs, and

applications in catabolic (e.g. breakdown of vitamin D3 by CYP24A1) and anabolic (e.g. cholesterol synthesis by CYP51A1) processes (6-8). Human P450s are responsible for most of the primary (phase I) xenobiotic metabolism in humans, producing oxidized metabolites that are more readily excreted, or further modified by phase II enzymes. P450s are also responsible for the generation of active drugs from their prodrug forms (9). P450s have been a target for biotechnological uses and subject to diversification of substrate selectivity/reactivity by engineering methods such as chimeragenesis (10,11) and directed evolution (12), leading to novel product formation by fermentation and bioreactor synthesis (13). The most promising candidate for a biotechnologically relevant P450 is the Bacillus megaterium cytochrome P450/cytochrome P450reductase (CPR) fusion enzyme P450 BM3 (CYP102A1), due to its high expression levels, soluble nature (compared to membrane-bound eukaryotic P450s), and convenient single component nature (P450 and CPR on the same polypeptide chain). The BM3 fusion structure and the fast electron transfer system (both within the CPR domain and between the CPR and the P450 domain) results in BM3 having the highest reported catalytic rate of the P450s (e.g. 17,000 min-1 with arachidonic acid) (14). BM3 has been targeted for many different reactions, including recent work to create an olefin cyclopropanation catalyst (15), production of oxidized steroids (16), and hydroxylation of short chain

133

alkenes (17), thus showing its versatility as an enzyme for engineering and biotechnological applications. Most of the mutagenesis approaches taken to alter substrate selectivity have relied on use of random mutagenesis, mostly by error prone PCR. Although this has identified a number of interesting variants, it is a laborious technique and a random approach to novel catalyst generation (18). Reported BM3 variants generated by directed evolution approaches typically have numerous mutations dispersed across the P450 domain (19). Through reviewing data for past evolution studies on BM3, it becomes clear that certain P450 domain amino acids are mutated quite frequently in variants with novel activities. Among the most frequent are those at positions Ala82 and Phe87 (20). Previously, we showed from X-ray crystallographic analysis that the A82F mutation causes structural destabilization of the heme domain, resulting in a lower P450 melting temperature, but also to an altered substrate specificity profile. The substrate-free A82F variant also occupied a conformation different to that of the substrate-free WT P450 (21). The human gastric proton pump inhibitor (PPI) drug omeprazole became a good substrate for the A82F, and our structural data revealed the productive mode of binding for this drug that enables its oxidation by the BM3 A82F variant at the same position as that catalysed by the major human metabolizing P450 CYP2C19 (21). Binding of omeprazole was further improved by the additional F87V mutation, where removal of the aromatic bulk of Phe87 frees space immediately in the vicinity of the BM3 heme iron.

In this study, we extend our previous studies to show that variant BM3 P450s containing Ala82 and Phe87 mutations also facilitate binding and oxidation of a range of other PPI drugs (esomeprazole, lansoprazole, pantoprazole and rabeprazole). The products of oxidation of these drugs are diverse from those of omeprazole, but in most cases mimic those produced by human cytochrome P450s in their metabolism of these drugs. Our data provide further evidence that mutations at a small number of residues that effect large scale structural reorganisation of P450 BM3 (“gatekeeper” mutants) are sufficient to induce major alterations in substrate selectivity and to enable BM3 to produce human-like metabolites of pharmaceuticals with potential uses as standards and reagents for pharmaceutical testing and FDA compliance.

134

3.3 Experimental Procedures

3.3.1 Mutagenesis and expression of WT and variant P450 BM3 enzymes. The gene encoding intact WT flavocytochrome P450 BM3 in pET15b was used for mutagenesis to create A82F, F87V and F87V/A82F (double mutant, DM) variants, as described in previous studies (21). Intact BM3 enzymes were expressed as N- terminal hexahistidine tagged enzymes either from pET15b (F87V, DM) constructs directly, or after cloning the WT and A82F genes into pET14b using NdeI/BamHI sites. WT and variant heme domain genes were generated using the relevant pET14b/15b constructs, as described previously (21). The heme domain genes (amino acids 1-473 of the 1048 amino acid flavocytochrome) were transferred as NdeI/BamHI fragments into pET20b to enable heme domain production in absence of a N-terminal His-tag, and to enable improved protein crystallization. All genes were fully sequenced to confirm relevant mutations and to ensure no exogenous mutations were incorporated. The WT and A82F intact BM3, and WT and all variant P450 BM3 heme domains were expressed in BL21 Gold (DE3) E. coli cells (Stratagene-Agilent UK) in TB medium with cells grown at 37 oC, and with agitation at 200 RPM in an orbital incubator. F87V and DM intact BM3 proteins were grown using autoinduction TB medium (Melford Ltd, Ipswich UK) from 4 L transformant cultures and with cell growth for 24-36 hours.

3.3.2 Purification of WT and variant intact P450 BM3 and heme domains.

Intact WT and variant P450 BM3 (BM3) and heme domains were purified essentially as described previously. (21) Briefly, cells were collected by centrifugation at 4 °C (6000 g, 10 min) and resuspended in ice-cold buffer B (50 mM KPi, 250 mM NaCl, 10% (v/v) glycerol, pH 7.0). Protease inhibitors (EDTA Free CompleteTM tablets, Roche, Germany) were maintained in all buffers used during purification. Cells were lysed by sonication, and supernatant containing soluble intact BM3/heme domain proteins was collected after high speed centrifugation (20,000 g, 40 min, 4 °C). The supernatant was collected again after a 30% ammonium sulfate cut on ice. P450 proteins were purified by Ni-IDA chromatography (Qiagen, UK), with bound proteins washed extensively at 4 °C in buffer B plus 5 mM imidazole, then eluted with 200 mM imidazole in buffer B. Proteins thus purified were transferred into buffer A (50 mM Tris, 1 mM EDTA, pH 7.2) and passed down a Sephacryl S-200 SEC column (GE Healthcare, 26 x 60 cm on an AKTA purifier system). Pure BM3 fractions

135

(checked by SDS-PAGE) were concentrated by ultrafiltration (Vivaspin, Vivaproducts, USA) and stored in buffer A plus 50% glycerol at -80 °C. For non- tagged BM3 heme domains, a further 30-60% ammonium sulfate cut was applied. The P450-containing pellet was resuspended in buffer A and dialyzed into the same buffer to desalt, then further purified by anion exchange chromatography on an AKTA, using a Q-Sepharose anion exchange column (16 x 10 cm), with elution in a gradient of 0-500 mM KCl in buffer A. Heme domain fractions were desalted (GE Healthcare column, 26 x 10 cm on an AKTA) into 25 mM KPi pH 7.0, loaded onto a hydroxyapatite column (Bio-Rad, USA, 16 x 11 cm) and eluted in a 200 mL gradient of 25–500 mM KPi, pH 7.0. Pure heme domains were concentrated by ultrafiltration (Vivaspin) and used immediately for crystallography, or flash frozen in liquid nitrogen and stored at -80 °C. As described previously, intact BM3 and heme domain proteins with the A82F mutation were passed through a Lipidex 1000 column (Perkin Elmer, UK) in 25 mM KPi, pH 7.0 to remove fatty acid retained during purification (21).

3.3.3 P450 quantification.

Concentrations of the low spin forms of WT and variant forms of intact P450 BM3 and heme domain were determined using extinction coefficients of ɛ418 = 105 and -1 -1 ɛ419 = 95 mM cm , respectively, at the Soret maximum (21,22). Fe(II)CO complexes were formed by bubbling sodium dithionite-reduced WT/variant intact BM3 and heme domains (ca 2-4 µM) with CO gas (23). WT and all variants showed almost complete formation of the P450 (thiolate-coordinated) state, with little P420 (likely cysteine thiol-coordinated) formed in any case (24,25).

3.3.4 Substrate binding and kinetic properties of WT and variant intact BM3 with fatty acids and PPI substrates.

The dissociation constants (Kd values) for WT/variant intact BM3 enzymes binding to N-palmitoylglycine (NPG), esomeprazole (ESO), lansoprazole (LAN), pantoprazole (PAN) and rabeprazole (RAB) were determined by absorption titrations (~1-4 µM protein) in 100 mM KPi (pH 7.0) at 25 °C (assay buffer) in 1 cm pathlength quartz cuvettes, as described previously (21,26). Titrations were continued until no further spectral changes occurred in the P450 heme. Difference spectra (produced by subtraction of each ligand-bound spectrum from that of the ligand-free enzyme in each case) were generated in each titration. Maxima and

136

minima in each set of difference spectra were identified (using the same wavelength pair in each titration) and the overall absorbance changes Δ(Apeak minus Atrough) were plotted versus [substrate]. Data were fitted using a (Michaelis-Menten, equation 1) hyperbolic function or (where the Kd value is ≤ 5x the P450 concentration) using the Morrison (quadratic) equation (equation 2, as described previously) for tight binding ligands, in order to determine Kd values (27,28). UV-Vis spectroscopy was carried out on a Cary 50 UV-Vis spectrometer (Agilent, UK). Data analysis and fitting was done using Origin Pro (OriginLab, Massachusetts USA). Steady-state kinetics. Kinetic studies were done on a Cary 50 UV-Vis spectrophotometer. Substrate (ESO, LAN, PAN, RAB)-dependent oxidation of NADPH was determined at 340 nm. BM3 concentration was kept constant (in range 25-150 nM) with substrate concentration varied, and near-saturating NADPH used (200 µM). Assays were done at 25 °C in assay buffer with a 1 cm pathlength quartz cuvette. Enzyme rate constants for substrate-induced NADPH oxidation were determined in triplicate at each substrate -1 -1 concentrations at 340 nm, using Δɛ340 = 6.21 mM cm . Rate constants were plotted versus substrate concentration. Data were fitted to equation 1 to define the kcat and Km parameters for substrate-dependent NADPH oxidation reported in Table 3.1.

3.3.5 EPR Spectroscopy

EPR spectroscopy (for ligand-free and drug substrate-bound WT and variant BM3 enzymes) was done on a Bruker ER-300D series electromagnet with microwave source interfaced with a Bruker EMX control unit and fitted with an ESR-9 liquid helium flow cryostat (Oxford Instruments), and a dual-mode microwave cavity from Bruker (ER-4116DM). Spectra were recorded at 10 K with a microwave power of 2.08 mW and modulation amplitude of 1 mT. Protein samples (200 μM) in KPi buffer (100 mM pH 7.0) were prepared with 1.8 µl methanol (MeOH)/DMSO and 400 µM PPI drug in 1.8 µl MeOH/DMSO.

3.3.6 Enzymatic oxidation of substrates and product characterization. Esomeprazole, Lansoprazole, Pantoprazole and Rabeprazole turnover and analysis by LC-MS. Turnover reactions for oxidation of ESO, LAN, PAN and RAB were done in deep well blocks at 37 oC with shaking for 30 min. Reaction mixtures contained purified WT or variant (F87V, A82F or DM BM3) enzymes (1 µM), substrate (10 µM), NADPH regeneration system (glucose-6-phosphate 7.76 mM, NADP+ 0.6 mM,

137

and glucose-6-phosphate dehydrogenase 0.75 U/ml) in turnover buffer (50 mM KPi,

5 mM CaCl2, pH 7.4) in a final volume of 500 µl. On completion of the reaction, protein was mixed with an equal volume of acetonitrile (ACN) containing 1 µg/ml fluconazole internal standard (IS) by shaking the mixed samples at 800 rpm for 10 min. Precipitated protein was filtered through protein precipitation plates (Phenomenex, Macclesfield UK) into mass spectrometry vials (FluidX, Nether Alderley UK) and clarified by centrifugation (4000 g, 25 min, 10 oC). Analysis was done on a Thermo Exactive LC-MS with a CTC PAL auto sampler (Thermo Scientific, UK) with a Kinetex 2.6U XB – C18 100A column (Phenomenex). A gradient of 0.1 % formic acid to ACN was used to resolve products. Drugs and metabolites were run in positive mode with the molecular ion as M[H]+. All high intensity peaks were selected from the total ion chromatogram and analyzed using Thermo Xcalibur quantification software, along with the fluconazole internal standard (IS). This software then gave total ion readings for both the IS and the metabolites formed. The total ion data were then corrected for the IS and for degradation of product that occurs non-enzymatically.

Esomeprazole, Lansoprazole and Rabeprazole turnover and analysis by NMR. Turnover reactions with ESO, LAN and RAB were done in 100-500 ml flasks at 37 oC with shaking of reagents at 100 rpm for 2 hours. Reaction mixtures contained purified WT or variant (F87V, A82F or DM) intact BM3 enzymes (1 µM), substrate (10-100 µM), NADPH regeneration system (as in preceding section) in 60-500 mL assay buffer. Products were extracted using Strata-X SPE columns (Phenomenex), dried under vacuum and eluted in 50/50 acetonitrile (ACN)/MeOH, followed by drying under nitrogen and water removal by freeze drying. Analysis was done on a Bruker Avance 400 MHz NMR (Bruker, Coventry UK). 1H spectra were collected at 400 MHz and 13C spectra at 101 MHz. Spectra were baseline corrected and referenced to tetramethylsilane (TMS) standard by the residual non-deuterated solvent in the sample. δ values are in ppm, J values are in Hz. Full assignments were made by COSY, HMBC and HMQC methods. Signal splittings were recorded as singlet (s), doublet (d), doublet of doublets (dd) alpha beta system (AB) and multiplet (m). Processing was carried out using MestReNova Lite (Mestrelab Research, Santiago de Compostela, Spain) and ACD NMR Processor (Advanced Chemistry Development, Inc. [ACD Labs], Toronto, Canada).

138

3.3.7 Crystallization of BM3 heme domains and determination of protein structures. Crystallography was performed using the sitting drop method using a seeding protocol at 4 °C. Crystals obtained during initial screens for each variant were used to create microcrystal screen stocks and consecutive screens (Molecular Dimensions) were made with drops that consisted of 150 nL of WT or variant heme domain proteins (230 µM), 50 nL of seed stock and 200 nL well solution using a Mosquito liquid handling robot (TTP LabTech Ltd, Melbourn UK). For the ESO and PAN variant heme domain complex structures, proteins were saturated with ESO/PAN ligand prior to crystallization. Ligands were titrated into heme domain samples until no further change in heme iron spin-state (towards high-spin) was observed. Thereafter, samples were concentrated by ultrafiltration in the presence of saturating ligand. Micro seeding was also used to produce diffraction quality crystals. Crystals were obtained under a range of conditions and flash-cooled in liquid nitrogen prior to data collection. The mother liquor was supplemented with 10% PEG 200 where an additional cryo-protectant was required. Data were collected at Diamond synchrotron beamlines and reduced and scaled using XDS (29). Structures were solved by molecular replacement with previously solved BM3 heme domain structure (PDB 1JPZ) using PHASER (30). Structures were refined using Refmac5 (31) and Coot (32).

139

3.4 Materials

Oligonucleotide primers for mutagenesis were from Eurofins MWG Operon (Ebersberg, Germany). Esomeprazole and lansoprazole were from Sigma-Aldrich (Poole, UK), Pantoprazole, rabeprazole, and all standards were from Santa Cruz (Dallas, USA). Bacterial growth medium (Terrific Broth, TB) was from Melford Ltd (Ipswich, UK). Unless specified, other chemicals were from Sigma-Aldrich and of the highest purity available.

140

3.5 Results

Figure 3.1 Proton pump inhibitor (PPI) drug structure and functional groups. Backbone PPI structure is shown with variations at the R groups for each compound indicated in the table. Abbreviations are esomeprazole (ESO), lansoprazole (LAN), pantoprazole (PAN), and rabeprazole (RAB). Esomeprazole is an S-isomer all other PPIs shown are racemic mixtures.

3.5.1 Spectral Binding Studies In view of the conversion of specificity of BM3 to an omeprazole binding/oxidizing P450 through A82F and F87V mutations (21), we examined whether these mutations also facilitated the binding of other PPI drugs in clinical use. We selected the major PPI drugs esomeprazole (the pharmacologically active S-enantiomer of omeprazole), lansoprazole, pantoprazole and rabeprazole. P450 spectral binding studies were done to ascertain whether these PPIs were also able to bind to variant (and WT) BM3 P450s, and if binding induced a low-spin (LS) to high-spin (HS) shift in heme iron spin-state equilibrium typical of BM3 substrate association. This is usually observed as a heme Soret band maximum shift from ~418 to ~392 nm in the case of BM3. Binding was done previously with NPG, a tight binding fatty acid derived substrate for BM3 (33), for the WT, A82F, F87V and F87V/A82F (DM) proteins (21), and these data showed that all P450s gave a near-complete heme iron HS shift with Kd values less than 1 µM (Table 3.1). Binding studies with the selected PPIs showed no spectral perturbation with WT BM3, suggesting negligible productive binding to the native enzyme. The F87V variant also showed no

141

discernible spectral shifts upon PPI titration. However, both the A82F and DM variants bound all four PPIs with substantial HS shifts, thus indicating that the A82F mutation is the primary cause of altered substrate selectivity for these drugs. In each case, there was an additive effect on PPI affinity on inclusion of the F87V mutation, with the DM showing a greater proportion of HS shift than the A82F variant, likely due to the effects of the F87V mutation in increasing the size of the active site cavity, and allowing the drugs to move further towards the heme.

Esomeprazole shows the highest affinity, with an A82F Kd of 23.9 µM and a DM Kd of 2.89 µM, associated with a near-complete HS conversion. Lansoprazole gives an

~60% shift to the HS form, with Kd values of 141 µM for A82F and 58.6 µM for the

DM. Pantoprazole gives up to ~50% shift, with an A82F Kd of 25.6 µM and DM Kd of 8.5 µM. Rabeprazole showed relatively little HS shift with A82F (~10%) and

comparatively weak binding (Kd of 159 µM). However, rabeprazole binding was

greatly improved in the DM with ~40% HS accumulated at saturation, and a Kd of 43.9 µM (Table 3.1). These data indicated that diverse PPI class compounds bind the BM3 gatekeeper variants in proximity to the heme catalytic site to displace the heme iron’s distal (6th) water ligand, and thus in a productive mode typical of fatty acid substrates for P450 BM3.

Figure 3.2. Esomeprazole binding to the BM3 DM enzyme. Panel A shows UV- visible spectra for the F87V/A82F (DM) intact BM3 variant (~1 µM), with the red spectrum that for the ligand-free form. Other selected spectra are shown, and were collected at various points during the titration with esomeprazole (ESO). UV-Vis spectra are shown at ESO concentrations of 0.9 µM (blue), 1.8 µM (pink), 5.4 µM (orange), and 36 µM (black). The inset shows overlaid difference spectra for ESO- bound forms at the same [ESO] as color coded in the main panel. Panel B shows a

142

plot of the Soret absorption change (ΔA389 – ΔA421) versus [ESO] with data fitted using a hyperbolic function (equation 1) to yield a Kd = 2.56 ± 0.33 µM.

Figure 3.3. Pantoprazole binding to the BM3 DM enzyme. Panel A shows UV- visible spectra for the F87V/A82F (DM) intact BM3 variant (~1 µM). The red spectrum is that for the ligand-free P450, with other selected spectra collected during titration with pantoprazole (PAN) also shown at PAN concentrations of 3.75 µM (blue), 10 µM (pink), 20 µM (orange), and 35 µM (black). The inset shows overlaid difference spectra for PAN-bound forms at the same [PAN] as color coded

in the main panel. Panel B shows a plot of the Soret absorption change (ΔA389 –

ΔA421) versus [PAN] with data fitted using a hyperbolic function (equation 1) to yield a Kd = 7.34 ± 0.44 µM.

Figure 3.4. Lansoprazole and rabeprazole binding to the DM P450 BM3 enzyme. Panel A shows spectral data from a lansoprazole (LAN) binding titration with 0.9 µM DM intact enzyme (~1 µM) (red), and following LAN ligand additions to a final concentration of 10 µM (blue), 20 µM (pink), 30 µM (orange), 50 µM (green)

and 70 µM (black). The inset shows a plot of Soret absorption change (ΔA388 –

143

ΔA421 from difference spectra generated as illustrated in Figure 3.2) versus [LAN], with data fitted using equation 1 to yield a Kd of 55.0 ± 4.5 µM. Panel B shows a rabeprazole (RAB) binding titration with DM intact BM3 (~1.8 µM) (red) and following addition of RAB to final concentrations of 10 µM (blue), 30 µM (pink), 40 µM (orange), 80 µM (green), and 200 µM (black). The inset shows a plot of Soret absorption change (ΔA388 – ΔA421) versus [RAB], with data fitted using equation 1 to yield a Kd of 46.5 ± 1.9 µM.

Figure 3.5. Esomeprazole binding to the BM3 A82F enzyme. Panel A shows UV- visible spectra for the A82F intact BM3 variant (~1.2 µM). The red spectrum is for the ligand-free form, and is shown with other selected spectra collected during titration with esomeprazole (ESO). UV-Vis spectra are shown at ESO concentrations of 3.59 µM (blue), 14.4 µM (pink), 53.7 µM (orange), and 178 µM (black). The inset shows overlaid difference spectra for ESO-bound forms at the same [ESO] as color coded in the main panel. Panel B shows a plot of the Soret absorption change (ΔA389 – ΔA421) versus [ESO] with data fitted using equation 1 to

yield a Kd of 22.6 ± 1.95 µM.

144

Figure 3.6 Pantoprazole binding to the BM3 A82F enzyme. Panel A shows UV- visible spectra for the A82F intact BM3 variant (~1 µM). The red spectrum is for the ligand-free form, and other selected spectra collected during titration with pantoprazole (PAN) are also shown. UV-Vis spectra are shown at PAN concentrations of 15.0 µM (blue), 30.0 µM (pink), 55.0 µM (orange), and 80.0 µM (black). The inset shows overlaid difference spectra for PAN-bound forms at the same [PAN] as color coded in the main panel. Panel B shows a plot of the Soret absorption change (ΔA389 – ΔA421) versus [PAN] with data fitted using equation 1 to

yield a Kd of 28.0 ± 3.33 µM.

Figure 3.7 Lansoprazole and rabeprazole binding to the A82F P450 BM3 enzyme. Panel A shows spectral data from a lansoprazole (LAN) binding titration with A82F intact enzyme (~1 µM) (red spectrum), and following LAN ligand additions to final concentrations of 39.9 µM (blue), 59.8 µM (pink), 79.8 µM (orange), and 149 µM (black). The inset shows a plot of Soret absorption change

(ΔA388 – ΔA422) from difference spectra generated (as described in Figure 3.2) versus [LAN], with data fitted using equation 1 to yield a Kd of 145 ± 20 µM. Panel B shows a rabeprazole (RAB) binding titration with A82F intact BM3 (~1.8 µM) (red

145

spectrum) and following addition of RAB to final concentrations of 60 µM (blue), 150 µM (pink), 299 µM (orange), and 990 µM (black). The inset shows a plot of Soret

absorption change (ΔA382 – ΔA422) versus [RAB], with data fitted using equation 1 to

yield a Kd = 159 ± 16 µM.

-1 kcat/Km Protein Substrate % HS Kd (µM) kcat (min ) Km (µM) µM min-1 WT NPG ~100 0.082 ± 0.011 4770 ± 160 13.9 ± 2.7 343 A82F NPG ~100 0.297 ± 0.069 5130 ± 570 26.3 ± 5.3 195 F87V NPG ~100 0.204 ± 0.045 4970 ± 240 14.9 ± 2.8 334 DM NPG ~100 0.004 ± 0.003 4050 ± 250 1.91 ± 0.31 2120 WT Eso ND ND ND ND ND AF Eso ~30 23.9 ± 3.2 1950 ± 55 45.7 ± 3.3 42.7 FV Eso ND ND 1380 ± 20 26.8 ± 2.6 51.5 DM Eso ~90 2.89 ± 0.23 2090 ± 60 3.26 ± 0.50 641 WT Lan ND ND 481 ± 9.0 61.8 ± 3.7 7.78 A82F Lan 40-70% 140 ± 23 1230 ± 32 31.7 ± 1.8 38.8 F87V Lan ND ND 1560 ± 41 61.3 ± 3.9 25.4 DM Lan 60-80 58.6 ± 5.4 435 ± 15 6.85 ± 0.63 63.5 WT PAN ND ND ND ND ND A82F PAN ~30 25.6 ± 2.7 2130 ± 110 51.1 ± 5.7 41.7 F87V PAN ND ND 1410 ± 39 44.2 ± 2.9 31.9 DM PAN ~50 8.50 ± 0.5 2290 ± 80 5.3 ± 0.5 432 WT Rab ND ND ND ND ND A82F Rab ~10 159 ± 15 1360 ± 33 126 ± 9.0 10.8 F87V Rab ND ND 1110 ± 57 101 ± 15 11.0 DM Rab ~40 43.9 ± 2.8 2890 ± 81 34.5 ± 3.2 83.8

Table 3.1. Binding and steady state kinetics for WT and variant BM3 enzymes with PPI substrates. ND indicates that the value was not determinable.

146

3.5.2 Steady state Kinetics Steady state kinetics studies (in presence of near-saturating NADPH) were performed on intact WT and BM3 variants using NPG and each of the PPIs at a range of substrate concentrations. As shown in Table 3.1 the natural NPG substrate -1 has high kcat values (4050-5130 min ) and low Km values (26.3-1.9 µM), leading to high catalytic efficiency in all cases. With WT BM3, only lansoprazole showed -1 significant substrate-stimulated NADPH oxidation (kcat = 480 min ) albeit with a relatively high Km (62 µM) (Figure 3.11). In contrast to data for the WT BM3, all the BM3 variants showed substrate stimulated oxidation with the four PPIs. The DM gave by far the greatest catalytic efficiencies. With ESO, the DM had the greatest -1 kcat (2090 min ) and the lowest Km (3.26 µM), giving the highest catalytic efficiency. -1 The F87V variant showed the highest kcat with LAN (1560 min ), though with a

modest Km of 61.3 µM. The DM is a slightly more efficient enzyme with a kcat of 435 -1 -1 min and a Km of 6.85 µM. With PAN, the DM had the highest kcat (2290 min ) and -1 lowest Km (5.35 µM). The DM had the highest kcat for RAB (2890 min ) with a Km of 34.5 µM (Table 3.1). Thus, the data clearly show that the mutations greatly enhance the ability of BM3 to bind PPIs and to catalyze PPI-stimulated NADPH oxidation. With the WT BM3, PPI-stimulated NADPH oxidation is only clearly observed with lansoprazole. However, the WT LAN steady state kinetics are somewhat at odds with the turnover data (Figure 3.21). This may be due to uncoupling of the P450 cycle, and therefore to non-productive NADPH oxidation. If the substrate binds weakly, then non-productive reduction of oxygen may lead to formation of hydrogen peroxide (peroxide uncoupling), water (oxidase uncoupling), or superoxide (superoxide uncoupling), according to the particular stage in the catalytic cycle at which a reactive iron-oxo intermediate breaks down non-productively (34).

147

Figure 3.8 PPI-dependent steady state kinetic analysis for the A82F BM3 variant. PPI-stimulated NADPH oxidation was determined using intact A82F BM3 (50-62.5 nM) across a range of PPI substrate concentrations. Data were collected as in the Methods section. Data points were collected in triplicate at different [PPI]; with error bars showing the SEM. Data were fitted using the Michaelis-Menten equation (equation 1) (red line); A. ESO B. LAN C. PAN D. RAB. Data are reported in Table 3.1

148

Figure 3.9 PPI-dependent steady state kinetic analysis for the F87V BM3 variant. Kinetic data were collected for the intact F87V BM3 enzyme (50-62.5 nM) with the different PPIs, as described in the Methods section and in Figure 3.8. Data points were collected in triplicate, with error bars showing the SEM. Data were fitted using the Michaelis-Menten equation (equation 1) (red); A. ESO B. LAN C. PAN D. RAB. Data are reported in Table 3.1.

149

Figure 3.10. PPI-dependent steady state kinetic analysis for the F87V/A82F (DM) BM3. Kinetic data were collected for the intact DM BM3 enzyme (50-62.5 nM) with the different PPIs. Data were collected as described in the Methods section. Points were collected in triplicate with error bars showing the SEM. Data were fitted using the Michaelis-Menten equation (equation) 1) (red); A. ESO B. LAN C. PAN D. RAB. Data are reported in Table 3.1.

150

Figure 3.11. PPI-dependent steady state kinetic analysis for the WT BM3 with lansoprazole. Data are shown for LAN-stimulated NADPH oxidation with intact WT BM3 (125 nM) across a range of LAN concentrations. Data were collected as described in the Methods section. Data points were collected in triplicate, with error bars showing the SEM. Data were fitted using the Michaelis-Menten equation -1 (equation 1) The kcat and Km values are 481 ± 9.0 min and 61.8 ± 3.7 µM, respectively.

3.5.3 EPR Spectroscopy

The EPR data for WT P450 BM3 (heme domain) shows a single set of low spin ferric heme g-values of gz = 2.45, gy = 2.25 and gx = 1.92 (2.45/2.25/1.92). These are largely consistent with previously reported data for WT BM3 (2.42/2.25/1.92)

(though with slight variation in the gz values) and indicative of a cysteine thiolate- coordinated P450 enzyme. (35) There is no significant signal corresponding to a high spin ferric state. However, HS EPR signals are often greatly diminished in intensity at the low temperatures (10 K in this case) required for heme EPR.

Addition of solvents perturbs the EPR spectra. Addition of methanol (MeOH) shifts the LS g-values to 2.40/2.25/1.92, while DMSO produces two sets of LS species at 2.45/2.25/1.91 and 2.41/2.25/1.92. These data indicate that the solvents affect the environment around the heme iron, likely influencing the solvation state within the active site. For DMSO, the second set of g-values may be associated with direct coordination of the heme iron by the solvent (36,37).

151

None of the PPI substrates showed any significant extent of high spin heme iron by EPR spectroscopy. However, given that fatty acid substrates also show little or no HS heme iron signal at 10 K, this was not unexpected. WT BM3 showed new species with both LAN and RAB that were different from those induced by the DMSO solvent in which they are dissolved (Figure 3.12). These g-values are at 2.40/2.25/1.92 in both cases. A82F BM3 shows three LS states with g-values at 2.47/2.25/1.90, 2.42/2.25/1.92 and 2.39/2.25/1.92, and a HS species with g-values at 8.04/3.44/1.65 in the resting form (Figure 3.13). The spectral complexity is reduced to two LS and no HS species on addition of MeOH (2.44/2.25/1.91 and 2.42/2.25/1.90), and to a single LS species with DMSO (2.44/2.25/1.91), reinforcing the important effects of even small volumes of solvent on the P450 heme environment. For A82F, it is plausible that solvent displaces a ligand from the P450 that is tightly bound in a small proportion of the protein in order to remove the HS signal. For the PPIs, the most profound effects are seen with LAN and RAB. The F87V variant shows the greatest substrate effect on EPR spectra with ESO, LAN and RAB all showing new species (Figure 3.14). Three LS states are observed in the native form (2.48/2.25/1.87, 2.45/2.25/1.90 and 2.41/2.25/1.92), and these are almost unchanged with addition of MeOH (2.48/2.25/1.90, 2.43/2.25/1.91, and 2.40/2.25/1.92), although they are shifted and reduced to two states on addition of DMSO (2.52/2.25/1.87, and 2.44/2.25/1.91). Addition of the three PPI substrates returns the EPR spectra to three states, but with altered g-values compared to the resting forms. In contrast, EPR spectra for the DM are unchanged with any of the PPI substrates compared to effects of the solvent alone. The fact the DM shows an unchanged EPR signal is interesting considering it has the tightest Kd values for each drug, and is kinetically the most active of the BM3 variants (Figure 3.15).

152

Figure 3.12 EPR analysis of WT BM3 heme domain. EPR spectra are shown for the WT BM3 heme domain, with applied magnetic field strength (Gauss) indicated on the x-axis. Optical spectra at room temperature show extensively low spin proteins, consistent with the EPR data shown. Spectra are overlaid with g-values marked, and labelled as follows: no additive (black), ESO in MeOH (red), LAN in DMSO (blue), RAB in DMSO (pink), MeOH (green), and DMSO (orange).

153

Figure 3.13. EPR analysis of the A82F BM3 heme domain. EPR spectra for the A82F BM3 heme domain are show with magnetic field strength (Gauss) indicated on the x-axis. Optical spectra at RT show low spin proteins in all cases, consistent with the EPR data Spectra are overlaid with g-values marked, and labelled as follows: no additive (black), ESO in MeOH (red), LAN in DMSO (blue), RAB in DMSO (pink), MeOH (green), DMSO (orange).

154

Figure 3.14. EPR analysis of the F87V BM3 heme domain. EPR spectra are show for the F87V BM3 heme domain with magnetic field strength (Gauss) indicated on the x-axis. Optical spectra at RT show low spin protein in all cases, consistent with the EPR data. Spectra are overlaid with g-values marked, and labelled as follows: no additive (black), ESO in MeOH (red), LAN in DMSO (blue), RAB in DMSO (pink), MeOH (green), and DMSO (orange).

155

Figure 3.15. EPR analysis of the DM BM3 heme domain. EPR spectra of the DM BM3 heme domain are shown with magnetic field strength (Gauss) indicated on the x-axis. Optical spectra at RT show low spin protein, consistent with the EPR data. Spectra are overlaid with g-values marked, and labelled as follows: no additive (black), ESO in MeOH (red), LAN in DMSO (blue), RAB in DMSO (pink), MeOH (green), and DMSO (orange).

156

3.5.4 Oxidation of PPI’s by WT and Variant BM3 enzymes using Liquid Chromatography Mass Spectrometry (LCMS)

To analyse the oxidase activity of the WT and variant forms of BM3, in vitro turnover of each PPI was performed with intact BM3 enzymes and products analysed by LCMS. Turnover studies were carried out as described in the oxidation and product identification part of the Methods (Section 3.36), with LCMS analysis done of the products to identify any reaction taking place.

Figure 3.16. Reactions schemes outlining pathways of P450 metabolism of PPI drugs. P450 BM3 pathways are shown in blue. Human P450 pathways are shown in red along with the human CYP isoform to which they correspond. Pantoprazole shown as S-isomer, as a drug it is a racemate as is lansoprazole and rabeprazole, esomeprazole is the active S-isomer of the PPI omeprazole.

The LCMS of esomeprazole showed a M[H+] of 146.1224 for the starting material, with the natural sulfone fragmentation occurring between the sulfur and the

157

methoxybenzimidazole moiety to give the corresponding 198.0587 species. Upon enzymatic turnover we see a +16 increase in these m/z peaks to 362.1134 and 214.0536, indicating insertion of oxygen into the pyridinyl fragment, generating oxidised esomeprazole (Figure 3.17). The product quantities for ESO turnover were very high with ~90% of a single monohydroxylated product obtained. This is superior to A82F, which shows ~40% of the same monohydroxylated product (Figure 3.21). It is interesting to note that, in contrast to omeprazole (21), turnover of esomeprazole shows negligible formation of the carboxylated ESO, which was found to occur for OMP through three successive oxidation reactions at the 5- position. This indicates a different metabolism for the R- and S-isomers of omeprazole by BM3 variants carrying the A82F mutation, as also seen with human CYP2C19 (38,39).

Figure 3.17 LC-MS traces of esomeprazole turnover by the P450 BM3 DM enzyme. Panel A (retention time [RT] = 5.30 min) shows data for ESO prior to addition of enzyme and the initiation of its oxidation by the BM3 DM enzyme. Peaks

158

at m/z 346.1222 and 198.0587 correspond to fragmentation of ESO at the sulfone group (between the sulfur and the methoxybenzimidazole moiety), with the smaller species representing the sulfur-containing fragment. Panel B (RT = 4.92 min) is following an enzymatic reaction for 30 minutes. The m/z peaks at 362.1171 and 214.0536 are for the 5-OH ESO and its hydroxylated fragment.

The LCMS for the lansoprazole substrate shows the substrate as M[H+] 370.0823 with a natural fragmentation between the benzimidazole and pyridinyl moieties, with the pyridinyl fragment seen as 252.0297. The main oxidised product is seen as M[H+] 386.0777 without fragmentation or loss of the pyridinyl moiety, indicating a change in the structure of the central sulfone to render it less labile. With LAN substrate no significant turnover occurs with WT BM3 (<1%), and the A82F variant only shows ~2.5% turnover to two oxidised products. The F87V and DM enzymes show much more turnover of LAN, with 16% and 33% product formed, respectively. The DM BM3 is not only more active, but also more selective with almost all turnover producing a single oxidised product (lansoprazole sulfone). The BM3 F87V variant produced a mixture of a main product, the same as that for the DM BM3 product, and also a secondary oxidised product at ~2.8% of total product (Figure 3.21). The second oxidative metabolite was in too small a concentration for NMR analysis, though the LCMS showed a +16 product peak at M[H+] 386.0774 and an oxidised pyridinyl fragment at 268.0247, indicating hydroxylation on the pyridinyl portion of the molecule. With five possible sites of oxidation on this fragment and no standards available, identification based purely on LCMS was inconclusive.

159

Figure 3.18 LC-MS traces for lansoprazole turnover by the P450 BM3 DM enzyme. Panel A (retention time [RT] = 5.65 min) shows data for LAN prior to addition of the BM3 DM enzyme and initiation of its oxidation. Peaks at m/z 370.0830 and 252.0302 correspond to fragmentation of LAN at the sulfone group (between the sulfur and the benzimidazole moiety), with the smaller species representing the sulfur-containing fragment. Panel B (RT = 6.02 min) is following an enzymatic reaction for 30 minutes. The m/z peaks at 386.0782 and 163.1330 are for the LAN sulfone and the benzimidazole moiety.

The LCMS for the pantoprazole substrate shows the substrate as M[H+] 384.0819 with a natural fragmentation between the benzimidazole and pyridinyl moieties, and with the pyridinyl fragment seen as 200.0375 (Figure 3.19). The main oxidised product is seen as M[H+] 400.0774 without fragmentation and loss of the pyidinyl moiety. Product(s) from PAN turnover were difficult to reconcile with observed MS data, as the oxidised product also appeared in the starting material and in control

160

samples. This led us to conclude that it was likely a synthesis by-product. We screened various by-products and found a match with pantoprazole N-oxide. Pantoprazole shows no discernible turnover to the N-oxide (above that seen in the starting material) with WT BM3, and A82F showed little conversion (~1.5%) and F87V only a little more (~4%). The greatest amount of PAN turnover was with the DM at ~11%. These values were corrected for the proportion of the N-oxide seen in the staring material (which was ~2% of total starting material identified). With such low turnover levels, it was not possible to confirm products by NMR, although comparisons with the available standard provided conclusive data.

Figure 3.19 LC-MS traces showing pantoprazole turnover from the P450 BM3 DM enzyme. Panel A (retention time [RT] = 5.82 min) shows data for PAN prior to addition of the BM3 DM enzyme and initiation of its oxidation. Peaks at m/z 384.0819 and 200.0375 correspond to fragmentation of PAN at the sulfone group (between the sulfur and the benzimidazole moiety), with the smaller species representing the sulfur-containing fragment. Panel B (RT = 6.23 min) is following an enzymatic reaction for 30 minutes. The m/z peak at 400.0774 is for the pantoprazole N-oxide, and no fragmentation is seen with this product.

161

It is clear from the LCMS of the rabeprazole starting material that there is a M[H+] peak at 360.1368, and that fragmentation at the sulfone occurs with a pyridinyl fragment at 242.0843. RAB turnover by LCMS shows three major metabolites, which are products of oxidative dealkylation reactions. Some non-enzymatic degradation of RAB occurs at ~25% with the control samples, largely to the thioether. This is a naturally occurring process (40), though exacerbated by the presence of the NADPH regeneration system, likely due to availability of electrons to reduce the sulfone. The three metabolites observed are the desmethyl RAB, the demethylated RAB thioether, and the thioether product with loss of the entire pyridinyl alkyl chain. The desmethyl RAB shows this same characteristic fragmentation pattern as the RAB, but with minus 14 for the methyl loss, with M[H+] 346.1220 and the pyridinyl fragment at 228.0692 (Figure 3.20). The second metabolite, the desmethyl RAB thioether, shows a M[H+] 330.1269 with fragmentation now at the benzimidazole ring, and showing a 163.1329 species for the fragment with the methyl group and a 149.0234 for the demethylated species. This altered pattern is likely due to the differing stability of the pyrdinyl moiety following the loss of the oxygen. The third and major metabolite was identified as the dealkyl metabolite of the RAB thioether, involving loss of oxygen and then loss of the entire pyridinyl alkyl chain. The fragmentation pattern is the same as that seen for the other thioether with fragmentation at the benzimidazole moiety, and with a 163.1330 species for fragmentation at the pyridinyl group, and a 149.0235 species for fragmentation at the sulfur. This change in fragmentation is likely due to a change in stability of the pyridinyl group and the sulfur upon the degradative sulfone oxygen loss.

162

163

Figure 3.20 LC-MS traces for rabeprazole turnover by BM3 enzymes. Panel A (retention time [RT] = 5.30 min) shows data for RAB prior to addition of BM3 enzymes and initiation of its oxidation. Peaks at m/z 360.1369 and 242.0844 correspond to fragmentation of RAB at the sulfone group (between the sulfur and the benzimidazole moiety), with the smaller species representing the sulfur- containing fragment. Panel B (RT = 2.76 min) is following an enzymatic reaction (with A82F BM3) for 30 minutes. The m/z peaks at 346.1220 and 228.0692 are for the desmethyl RAB and for its demethylated fragment. Panel C (RT = 3.91 min) is following an enzymatic reaction for 30 minutes (with A82F BM3). The m/z peaks at 330.1269 and 149.0234 are for the desmethyl RAB thioether and the benzimidazole thioether fragment. Panel D (RT = 4.20 min) is following an enzymatic reaction for 30 minutes (with DM BM3). The m/z peaks at 272.0854 and 149.0235 are for the dealkyl RAB thioether and the benzimidazole thioether fragment.

Rabeprazole was extensively metabolised by WT and by each of the variant BM3 enzymes, with WT BM3 generating 30% of the major metabolite dealkyl RAB thioether. The A82F variant showed ~70% conversion and a very different profile, with the primary metabolite being the desmethyl RAB thioether at ~45%, and with the desmethyl RAB at 20% and the dealkyl thioether only comprising ~10%. The F87V variant again shows the dealkyl thioether metabolite as the primary product with ~60% conversion, and with the desmethyl thioether as the secondary product (~20%), and the desmethyl RAB at ~15%. Finally, the DM BM3 showed the greatest total turnover at more than ~95%, with ~70% dealkyl thioether, ~20% desmethyl thioether and ~10% desmethyl RAB (Figure 3.21).

The major human metabolites of rabeprazole are the non-enzymatically produced thioether and the human CYP-mediated metabolites desmethyl RAB and RAB sulfone, produced by CYP2C19 and CYP3A4 respectively (41-43).

164

Figure 3.21. Proportions of PPI turnover products identified by LCMS. The charts show proportions of products from different PPI drugs seen from LCMS analysis after 30 min incubation, corrected for an internal standard (fluconazole), and reflecting data averages of two repeats with error bars showing SEM. WT (black), A82F (red), F87V (blue), DM (orange). A. Esomeprazole showing proportions of ESO, 5-hydroxy ESO (ESO-OH) and 5-carboxyl ESO (ESO-COOH) B. Lansoprazole showing proportions of LAN, LAN sulfone (LAN SUL) and hydroxylated LAN (LAN-OH) C. Pantoprazole showing proportions of PAN and PAN N-oxide (PAN N-OX) D. Rabeprazole showing proportions of RAB, dealkyl RAB thioether (RAB TE DA), demethyl RAB (RAB DM) and demethyl RAB thioether (RAB TE DM).

3.5.5 Oxidation of PPI’s by WT and Variant BM3 enzymes using NMR

Nuclear Magnetic Resonance (NMR) Spectroscopy was used for the analysis of the substrates and the main monohydroxylated products from ESO, LAN, and RAB turnover. Large scale enzymatic turnovers were carried out for each of the

165

substrates in order to use NMR to identify each of the products seen in our LCMS analysis.

Confirmation of the position of oxidation of ESO at the 5-methyl group was obtained from a combination of 1H, 13C and 2D NMR. Esomeprazole starting material was 1 assigned as: H NMR (CDCl3 400 MHz) δ = 8.13 (S, 1H), 7.46 (d, J = 8.84), 6.95 (Broad S, 1H), 6.87 (dd, J = 8.97, 2.40 Hz, 1H), 4.69 (AB, J = 13.52, 11.12 Hz, 2H), 13 3.77 (S, 3H), 3.63 (S, 3H), 2.16 (S, 3H), 2.15 (S, 3H). C NMR (CDCl3 101 MHz) δ = 164.5, 149.7, 148.7, 127.1, 126.5, 60.8, 60.0, 55.8, 13.4, 11.6. LCMS [M+H]+

346.1224 (C17H20N3O3S).

As can be seen from the ESO NMR (Figure 3.22-3.25) the product peaks show generation of a single new methyl singlet at 2.14, slight shifts in each methoxy group, a new singlet δ 4.69 corresponding to a methoxy group, and a shift in the pyridinyl hydrogen from δ 8.13 to δ 8.37. This pattern is indicative of hydroxylation at one of the pyridinyl methyl groups to form a methoxy group. The coupling seen in the 2D NMR combined with the downfield shift of the pyridinyl hydrogen signal and the lack of shift in the AB system indicates hydroxylation at the 5-position, the same as that catalyzed by the human CYP2C19 (44).

166

Figure 3.22 1H NMR spectrum of esomeprazole. The 1H spectrum for the ESO starting material is shown with peaks labelled and integrated. The overlay shows the esomeprazole structure with accepted (non IUPAC) numbering, and peaks are labelled accordingly. Data were collected on a 400 MHz NMR in CDCl3, corrected to TMS by residual non-deuterated solvent.

167

Figure 3.23 HMBC spectrum of esomeprazole. The spectrum shows the long range coupling of 1H to 13C nuclei. Coupling is observed between the 3-methyl group and the AB system; and between the 5-methyl group and the pyridinyl methoxy and pyridinyl hydrogen. Data were collected on a 400 MHz NMR in CDCl3, corrected to TMS by residual non-deuterated solvent.

Upon enzymatic oxidation of ESO, a mixture of starting material and monohydroxylated product were identified. The product was assigned to 5-OH ESO 1 and the additional peaks seen were: H NMR (CDCl3 400 MHz) δ = 8.37 (S, 1H), 7.56 (Broad S, 1H), 6.96 (d, J = 2.40 Hz, 1H), 6.93 (dd, J = 8.84, 2.40 Hz, 1H), 4.76

168

(AB, J = 13.64, 11.65 Hz, 2H), 4.69 (S, 2H), 3.82 (S, 3H), 3.67 (S, 3H), 2.14 (S, 3H). 13 C NMR (CDCl3 101 MHz) δ = 164.5, 149.6, 148.2, 127.1, 126.6, 61.2, 58.4, 55.8, + 13.4, 11.6. LCMS [M+H] 362.1134 (C17H20N3O4S).

Figure 3.24 1H NMR spectrum of turnover products from esomeprazole oxidation. Products were generated in the reaction of the F87V BM3 variant with ESO substrate. Only the product peaks are labelled and integrated for clarity. The overlay shows esomeprazole structure with accepted (non IUPAC) numbering, and peaks are labelled accordingly. The spectrum shows the generation of a methoxy peak at δ 4.69 and a new methyl peak at δ 2.14, indicative of hydroxylation at one of the OMP methyl groups. The downfield shift of the pyridinyl hydrogen signal and the lack of shift in the AB system indicates hydroxylation at the 5 position. Data were

collected on a 400 MHz NMR in CDCl3, corrected to TMS by residual non- deuterated solvent.

169

Figure 3.25 HMBC spectra of turnover products from esomeprazole oxidation. Products were generated in the reaction of the F87V BM3 with ESO substrate. The data show the long range coupling of 1H to 13C nuclei. Additional peaks generated (by comparison with Figure 3.23 above) show coupling of the new methoxy peak (δ 4.69) to the pyridinyl methoxy (δ 3.67), the pyridinyl hydrogen (δ 8.37), and to the new methyl peak (δ 2.14). Lack of coupling to the AB system confirms that hydroxylation occurs on the 5-methyl (and not the 3-methyl) group. Data were

collected on a 400 MHz NMR in CDCl3, corrected to TMS by residual non- deuterated solvent.

170

NMR analysis of lansoprazole along with the standard LAN showed the sulfone to be the product (formed when a second oxygen atom is introduced to the central sulfur). The LAN sulfone is the major human metabolite formed by the human CYP 3A4 (45). COSY and HMBC spectra were used to make full assignments of LAN and LAN sulfone. Lansoprazole assignment (DMSOd6 400 MHz) δ 13.59 (S, 1H), 8.29 (d, 1H, J = 5.68 Hz), 7.69 (Broad d, 2H), 7.31 (Broad d, 2H), 7.09 (d, 1H, J = 5.68 Hz), 4.91 (q, 2H, J = 8.72 Hz), 4.80 (AB, 2H, J = 13.77 Hz), 2.18 (S, 3H). 13C NMR (DMSOd6 101 MHz) 161.25, 154.14, 150.90, 148.10, 125.13, 122.37, 122.02, 107.00, 64.59 (q), 59.94, 10.52.

As can be seen from the lansoprazole NMR turnover reaction (Figure 3.26-3.28), there is a mixture of the starting material LAN and the product LAN sulfone. The combination of LCMS and NMR analysis showed that although there was a +16 increase in the product, there was no discernible hydroxylation on a LAN carbon. The only change observed was the loss of the AB system quartet to give a singlet at δ 5.09. The peak positions and coupling pattern in this product was equal to that seen in the lansoprazole sulfone standard (Santa Cruz, USA), giving us conclusive proof that our product is the sulfone.

171

Figure 3.26 1H NMR spectrum of lansoprazole. The 1H spectrum for the LAN starting material is shown with peaks labelled and integrated. The overlay shows the S-lansoprazole (drug is racemic mixture) structure with accepted (non IUPAC) numbering, and with peaks labelled accordingly. Data were collected on a 400 MHz NMR in DMSO d6, corrected to TMS by residual non-deuterated solvent.

Lansoprazole sulfone standard. (DMSO d6 400 MHz) δ 13.79 (Broad S, 1H), 8.11 (d, 1H, J = 5.68 Hz), 7.69 (Broad S, 2H), 7.39 (m, 2H), 7.07 (d, 2H, J = 5.68 Hz), 5.12 (S, 2H), 4.91 (q, 2H, J = 8.72 Hz), 2.23 (S, 3H). 13C NMR (DMSOd6 101 MHz) 161.43, 148.07, 147.86, 147.76, 123.26, 122.35, 107.47, 64.78, 64.44, 60.37, 10.91.

172

Figure 3.27 1H NMR spectrum of the lansoprazole sulfone standard. The 1H spectrum for the LAN sulfone standard is shown with peaks labelled and integrated. The overlay shows the S-lansoprazole sulfone (metabolite is racemic mixture) structure with accepted (non IUPAC) numbering, and with peaks labelled accordingly. Data were collected on a 400 MHz NMR in DMSO d6, corrected to TMS by residual non-deuterated solvent.

DM lansoprazole turnover product (DMSOd6 400 MHz) δ 13.62 (Broad S, 1H), 8.13 (d, 1H, J = 5.68 Hz), 7.66 (Broad S, 2H), 7.35 (m, 2H), 7.09 (app. t, 1H, J = 5.68 Hz), 5.09 (S, 2H), 4.91 (q, 2H, J = 8.72 Hz), 2.22 (S, 3H).

173

Figure 3.28 1H NMR spectrum of turnover products from lansoprazole oxidation. Products were generated in the reaction of the F87V/A82F (DM) BM3 with LAN substrate. Only the product peaks are labelled and integrated for clarity. The overlay shows the S-lansoprazole sulfone (metabolite is racemic mixture) structure with accepted (non IUPAC) numbering, and with peaks labelled accordingly. The spectrum shows the generation of a methoxy peak at δ 5.09 and a new methyl peak at δ 2.22, indicative of generation of the LAN sulfone. The aromatic region is largely unchanged, though with overlapping product and starting material peaks increasing the integrated values. Data were collected on a 400 MHz NMR in DMSO d6, corrected to TMS by residual non-deuterated solvent.

The NMR analysis of rabeprazole metabolism with the BM3 DM found its metabolism to be substrate concentration dependent, with the main products being the thioether at 100 µM substrate, the desmethyl thioether at 50 µM substrate, and the dealkyl thioether at 10 µM substrate. These data were collected using LCMS, since an unfeasibly large reaction volume would be required to produce sufficient amounts of products for analysis by NMR. This concentration dependency is likely due to substrate inhibition, as is often seen in P450 enzymes (46,47), though not

174

reported with BM3 previously. Assignments of rabeprazole were made by 1H, COSY and HMBC NMR, and assessments of missing peaks were made with reference to the 1H product spectra.

Rabeprazole standard (DMSOd6 400 MHz) δ 8.28 (d, 1H, J = 5.56 Hz), 7.46 (m, 2H), 6.93 (d, 1H, J = 5.68 Hz), 6.88 (m, 2H), 4.57 (AB, 2H, J = 12.88 Hz), 4.10 (t, 2H, J = 6.19 Hz), 3.49 (t, 2H, J = 6.32 Hz), 3.25 (S, 3H), 2.17 (S, 3H), 1.98 (quin., 2H, J = 6.19 Hz) 13C NMR (DMSOd6, 101 MHz) δ 162.61, 152.41, 147.95, 146.62, 121.74, 118.20, 117.32, 105.92, 68.30, 64.92, 59.64, 57.96, 28.67, 10.83.

The NMR analysis of rabeprazole turnover showed a high substrate concentration dependency on products formed. At high substrate concentration (100 µM) the product is primarily the RAB thioether, shown by loss of the AB system to form a singlet at δ 4.78 (Figure 3.32). At lower concentration (50 µM), analysis shows that the thioether is also formed, but now the singlet at δ 3.26 is also lost. This singlet corresponds to the methyl group on the 3-methoxypropoxy group, showing that a demethylation reaction occurred (Figure 3.33). LCMS analysis indicated this was further broken down to the alcohol with loss of the entire 3-methoxypropoxy group, though NMR analysis was again impractical here due to the low concentration of the products and thus the large reaction volumes required.

175

Figure 3.29 1H NMR spectrum of rabeprazole. The 1H spectrum for the RAB starting material is shown with peaks labelled and integrated. The overlay shows the S-rabeprazole (drug is racemic mixture) structure with accepted (non IUPAC) numbering, and with peaks labelled accordingly. Data were collected on a 400 MHz NMR in DMSO d6, corrected to TMS by residual non-deuterated solvent.

176

Figure 3.30 COSY NMR spectrum of rabeprazole. The COSY spectrum for the RAB starting material is shown. This was used to assign the methoxypropyl carbon chain δ 1.98, 3.49 and 4.10. Data were collected on a 400 MHz NMR in DMSO d6, corrected to TMS by residual non-deuterated solvent.

177

Figure 3.31 HMBC spectrum of rabeprazole. The spectrum shows the long range coupling of 1H to 13C nuclei. Coupling is observed between the 3-methyl group and the AB system; and between the varios hydrogen on the alkyl chain. Data were collected on a 400 MHz NMR in DMSO d6, corrected to TMS by residual non- deuterated solvent.

Rabeprazole (100 µM) product 1H NMR (DMSO d6, 400 MHz) δ 12.63 (Broad S, 1H), 8.23 (d, 1H, J = 5.56 Hz), 7.45 (Broad m, 2H), 7.12 (m, 2H), 6.95 (d, 1H, J = 5.68 Hz), 4.69 (S, 2H), 4.10 (t, 2H, J = 6.19 Hz), 3.48 (t, 2H, J = 6.19 Hz), 3.24 (S, 3H), 2.21 (S, 3H), 1.98 (quin. 2H, J = 6.19 Hz). 13C NMR (DMSOd6, 101 MHz)

178

162.64, 154.66, 150.22, 147.76, 119.69, 106.29, 68.26, 65.05, 57.95, 36.22, 28.64, 10.37.

Figure 3.32 1H NMR spectrum of turnover products from rabeprazole oxidation. Products were generated in the reaction of the F87V/A82F (DM) BM3 with RAB substrate at 100 µM. The spectrum shows the total loss of the AB system to produce a singlet at δ 4.69. As this integrates for both protons, this indicates the loss of the sulfone oxygen to produce a thioether. The rest of the spectrum is unchanged, indicating that no further reaction took place. Data were collected on a 400 MHz NMR in DMSO d6, corrected to TMS by residual non-deuterated solvent.

Rabeprazole (50 µM) product 1H NMR (DMSO d6, 400 MHz) δ 12.72 (Broad S, 1H), 8.23 (d, 1H, J = 5.68 Hz), 7.45 (m, 2H), 7.12 (m, 2H), 6.95 (d, 1H, J = 5.81 Hz), 4.69 (S, 2H), 4.12 (t, 2H, 6.19 Hz), 3.58 (t, 2H, J = 6.19), 2.21 (S, 3H), 1.89 (quin., 2H, J = 6.06 Hz). 13C NMR (DMSOd6, 101 MHz) 162.76, 154.61, 150.25, 147.74, 121.31, 119.68, 106.30, 64.99, 57.04, 36.16, 31.80, 10.37.

179

Figure 3.33 1H NMR spectrum of turnover products from rabeprazole oxidation. Products were generated in the reaction of the F87V/A82F (DM) BM3 with RAB substrate at 50 µM. The spectrum shows the total loss of the AB system to produce a singlet at δ 4.69. As this integrates for both protons, this indicates the loss of the sulfone oxygen to produce a thioether. There is also total loss of the methyl end of the methoxypropyl group, indicating oxidative demethylation has taken place. The demethylation causes small shifts in the adjacent propyl chain signals, but otherwise the rest of the spectrum is unchanged, indicating that no further reaction took place. Data were collected on a 400 MHz NMR in DMSO d6, corrected to TMS by residual non-deuterated solvent.

3.5.6 X-ray crystallography

The structure of ESO bound to the DM BM3 heme domain was solved to 1.83 Å resolution, using molecular replacement with our previous DM heme domain omeprazole-bound structure (4KEY). Preliminary refinement of the structure showed that, within error, it is identical to that for DM omeprazole-bound form. An overlay of the DM ESO-bound active site ligand density with the OMP-bound DM structure

180

(4KEY) is shown in Figure 3.35, indicating no significant differences between the omeprazole and esomeprazole DM heme domain structures. The higher resolution in the ESO-bound DM structure over the OMP-bound one allows us to confirm that the sulfone oxygen has been lost, as there is no electron density around the central sulfur, either from the breakdown of the omeprazole in the aqueous environment (48) or from the action of the synchrotron X-ray irradiation. This means that in both DM heme domain structures omeprazole/esomeprazole are bound as the thioether forms and are therefore non-chiral. This prevents us from gaining a clear structural rationale as to why omeprazole binds tighter yet is metabolised less selectively than esomeprazole. The DM BM3 heme domain is in the substrate bound (SB) conformation seen previously for fatty acid bound-WT BM3 Heme domain (49), strengthening the argument that structural destabilisation is the cause of altered binding in this variant, and so even the weaker binding ESO molecule is seen in this conformation.

181

Figure 3.34 Overlay of the DM heme domain esomeprazole-bound crystal structure with the previously reported DM omeprazole-bound heme domain structure (4KEY). The overlay shows the electron density of the ESO ligand, revealing it does not have any oxygen on the central sulfur. The protein structure and ligand conformation is identical to that for the omeprazole-bound structure. The structural resolution is at 1.83 Å, solved using the original DM OMP-bound structure (4KEY) and refined to 20.2/22.9 R/Rfree without any model rebuilding.

The structure of PAN bound to the DM BM3 heme domain was solved to 2.04 Å resolution, using molecular replacement with our previous DM heme domain omeprazole-bound structure (4KEY). Preliminary refinement of the structure showed that much of the ligand density was missing. We can see from the overlay with esomeprazole (Figure 3.35) that the density around the central sulfur and benzimidazole ring overlay well, but the pyridinyl region is distinctly lacking in density. The lack of density could mean that the ligand has degraded in the

182

aqueous solution or during crystallization, or that it isn’t tightly enough bound to be in a single conformation and therefore may exist in multiple conformations in the active site. However, what is also clear is that the PAN molecule is bound such that the pyridinyl part of the structure would be orientated towards the P450 heme. This would be consistent with the observe oxidation of PAN on the pyridinyl nitrogen atom

Figure 3.35. Density overlay of the pantoprazole-bound DM heme domain crystal structure with the esomeprazole-bound DM heme domain structure. The overlay shows the electron density of the PAN ligand, revealing that it overlays well with the ESO at the central sulfur and benzimidazole ring regions. Very little density is seen for PAN at the pyridinyl ring, and the density shows that a 6th ligand (likely the oxygen atom of a water molecule) is directly ligating the heme iron. The structural resolution is at 2.04 Å, solved using the original DM OMP-bound structure (4KEY).

183

3.6 Discussion

The production of drug metabolites is a relatively recent area of interest for the biocatalysts industry, as these oxidised compounds are often expensive and difficult to produce using conventional chemical methods (1). Most human drug metabolites are formed directly or indirectly by the cytochrome P450 enzymes, which are responsible for the majority of phase I xenobiotic metabolism. Understanding the pharmacological activities and toxicity of these metabolites is important in developing a full understanding of the mode of action of the parent drug, as well as in identifying any serious health risks posed by its use. Moreover, the FDA now demands that any drug metabolite formed to significant levels in the body should be subject to similar rigorous testing as is the parent drug itself (50). These regulations pose challenges for production of sufficient amounts of such metabolites, particularly in cases where regio-and/or stereoselective oxidation of the drug occurs in formation of the metabolite. Use of recombinant human P450s for generation of small amounts of such drug metabolites is an option, but may be impractical on a larger scale since amounts of products formed may be limited by stability of the membranous P450s, their rather slow catalytic rates, as well as their requirement for an additional CPR redox partner.

These issues have led researchers to consider using instead the Bacillus megaterium P450-CPR fusion enzyme P450 BM3, which has the highest reported rates of substrate oxidation across the P450 superfamily, and which has other attractive properties such as its catalytic self-sufficiency and soluble nature. BM3 has also been a test bed for P450 mutagenesis, and its activity profile has been altered both by structure-guided engineering and by directed evolution approaches (4,51,52). Our previous studies showed that single mutations within the BM3 heme domain could dramatically alter substrate selectivity from fatty acids to the proton pump inhibitor omeprazole, and a model that related altered structural stability of A82F-containing variants (“gatekeeper” variants) to substrate specificity diversification was developed and validated (21). In this study, four further members of the PPI class of drugs were analysed for their ability to bind to and be oxidized by WT, A82F, F87V and F87V/A82F (DM) BM3 enzymes. It was shown that esomeprazole, lansoprazole, pantoprazole and rabeprazole all bind the BM3 variants, and show a similar pattern of improved binding (lower Kd values) for the F87V/A82F (DM) over the A82F point variant. The

184

PPIs have structural similarities, which likely in part explains their ability to bind the BM3 variants. However, the altered conformational “flexibility” of the A82F- containing BM3 variants is in itself a major determinant of the altered substrate selectivity in the gatekeeper variants, and different modes of binding of the tested PPIs is evident from the various positions at which oxidation of the different PPI substrates occurs (Table 3.1, Figure 3.16).

The binding and efficient turnover of the diverse PPIs (with ESO showing tightest binding and greatest turnover) by the BM3 gatekeeper variants provides further clear evidence of the ability to radically transform substrate selectivity by a semi- rational strategy that first introduces a structurally destabilizing mutation into the P450. Here, the A82F mutation is key to enabling the productive binding of a range of PPI drugs, and their transformation to diverse products also indicates not only that there are important differences in specific binding modes of the different PPIs, but also that the products formed arise from oxidation of these drugs at different positions, and that in many cases replicates the major metabolites formed by their human P450 counterparts with the same drugs.

More specifically, esomeprazole is transformed by BM3 (A82F, F87V and DM) to the same 5-hydroxylated product as the major human metabolite, produced by CYP2C19 (53). In comparison to our initial studies on the oxidation of the racemate omeprazole by BM3 variants, there is negligible formation of the 5-carboxylated ESO product. This suggests strongly that the 5-hydroxylated ESO (5-OH S-OMP) no longer binds productively in the BM3 A82F-containing variants, whereas the 5- OH R-OMP from the racemate may be able to do so to enable formation of the 5- COOH OMP product. These data indicate that, as with the human P450s, omeprazole metabolism is specific for each enantiomer (38,39). The major metabolites seen in humans for lansoprazole are the sulfone and the 5-hydroxy LAN, produced by CYP3A4 and CYP2C19, respectively. (54) Our data show the primary product with our variant BM3 enzymes (with up to 33% conversion by the DM) is the lansoprazole sulfone, which is also the major human metabolite from CYP3A4. The major pathway for rabeprazole metabolism is its non-enzymatic (reductive) conversion to the thioether, with up to 50% conversion in aqueous solution in one hour (55,56). This is also observed in our reactions, and control reactions showed that BM3 enzymes had little effect on this reductive reaction, but inclusion of the NADPH cofactor regeneration system more than doubled the amount of this product formed. As the thioether reaction is reductive, it is possible

185

the regeneration system provides rabeprazole with the electrons required to speed up this reaction above that seen in cofactor-free aqueous solution. The other two major human metabolites of RAB are the desmethyl and sulfone forms, as produced by CYP2C19 and CYP3A4, respectively (56). Our results show that we can preferentially produce the thioether or desmethyl RAB by varying the substrate concentration in our turnover reactions (using the DM BM3 enzyme). Also, a novel metabolite can be produced that involves the loss of the entire ether chain, producing a dealkyl thioether of rabeprazole. This is observed if the substrate concentration is lowered still further. This type of high substrate concentration dependency is also seen in other human P450 reactions (e.g. the CYP2C9 catalyzed methyl-hydroxylation of the non-steroidal anti-inflammatory drug celecoxib and the CYP2D6 catalyzed O-demethylation of the antitussive dextromethorphan) (46,47).

186

3.7 References

1. Girhard, M., and Urlacher, V. B. (2010) Biooxidation with Cytochrome P450 Monooxygenases. in Modern Oxidation Methods, Wiley-VCH Verlag GmbH & Co. KGaA. pp 421-450 2. Axarli, I., Prigipaki, A., and Labrou, N. E. (2005) Engineering the substrate specificity of cytochrome P450 CYP102A2 by directed evolution: production of an efficient enzyme for bioconversion of fine chemicals. Biomol. Eng. 22, 81-88 3. Geier, M., Braun, A., Emmerstorfer, A., Pichler, H., and Glieder, A. (2012) Production of human cytochrome P450 2D6 drug metabolites with recombinant microbes – a comparative study. Biotechnology Journal 7, 1346-1358 4. Sawayama, A. M., Chen, M. M. Y., Kulanthaivel, P., Kuo, M.-S., Hemmerle, H., and Arnold, F. H. (2009) A Panel of Cytochrome P450 BM3 Variants to Produce Drug Metabolites and Diversify Lead Compounds. Chem. Eur. J. 15, 11723-11729 5. Bornscheuer, U. T., Huisman, G. W., Kazlauskas, R. J., Lutz, S., Moore, J. C., and Robins, K. (2012) Engineering the third wave of biocatalysis. Nature 485, 185-194 6. Huang, W.-C., Cullis, P. M., Raven, E. L., and Roberts, G. C. K. (2011) Control of the stereo-selectivity of styrene epoxidation by cytochrome P450 BM3 using structure-based mutagenesis. Metallomics 3, 410-416 7. Meunier, B., de Visser, S. P., and Shaik, S. (2004) Mechanism of Oxidation Reactions Catalyzed by Cytochrome P450 Enzymes. Chem. Rev. (Washington, DC) 104, 3947-3980 8. Nebert, D. W., and Russell, D. W. (2002) Clinical importance of the cytochromes P450. The Lancet 360, 1155-1162 9. Ortiz de Montellano, P. R. (2013) Cytochrome P450-activated prodrugs. Future Medicinal Chemistry 5, 213-228 10. Eiben, S., Bartelmäs, H., and Urlacher, V. (2007) Construction of a thermostable cytochrome P450 chimera derived from self-sufficient mesophilic parents. Appl. Microbiol. Biotechnol. 75, 1055-1061 11. Otey, C. R., Silberg, J. J., Voigt, C. A., Endelman, J. B., Bandara, G., and Arnold, F. H. (2004) Functional Evolution and Structural Conservation in

187

Chimeric Cytochromes P450: Calibrating a Structure-Guided Approach. Chem. Biol. 11, 309-318 12. Sideri, A., Goyal, A., Di Nardo, G., Tsotsou, G. E., and Gilardi, G. (2013) Hydroxylation of non-substituted polycyclic aromatic hydrocarbons by cytochrome P450 BM3 engineered by directed evolution. J. Inorg. Biochem. 120, 1-7 13. Martinez, C. A., and Rupashinghe, S. G. (2013) Cytochrome p450 bioreactors in the pharmaceutical industry: challenges and opportunities. Curr. Top. Med. Chem. 13, 1470-1490 14. Munro, A. W., Girvan, H. M., and McLean, K. J. (2007) Variations on a (t)heme--novel mechanisms, redox partners and catalytic functions in the cytochrome P450 superfamily. Nat. prod. Rep. 24, 585-609 15. Coelho, P. S., Brustad, E. M., Kannan, A., and Arnold, F. H. (2013) Olefin Cyclopropanation via Carbene Transfer Catalyzed by Engineered Cytochrome P450 Enzymes. Science 339, 307-310 16. Venkataraman, H., de Beer, S. B. A., van Bergen, L. A. H., van Essen, N., Geerke, D. P., Vermeulen, N. P. E., and Commandeur, J. N. M. (2012) A Single Active Site Mutation Inverts Stereoselectivity of 16-Hydroxylation of Testosterone Catalyzed by Engineered Cytochrome P450 BM3. ChemBioChem 13, 520-523 17. Ost, T. W. B., Miles, C. S., Murdoch, J., Cheung, Y.-F., Reid, G. A., Chapman, S. K., and Munro, A. W. (2000) Rational re-design of the substrate binding site of flavocytochrome P450 BM3. FEBS Lett. 486, 173- 177 18. van Vugt-Lussenburg, B. M. A., Stjernschantz, E., Lastdrager, J., Oostenbrink, C., Vermeulen, P. E., and Commandeur, J. N. M. (2007) Identification of Critical Residues in Novel Drug Metabolizing Mutants of Cytochrome P450 BM3 Using Random Mutagenesis. J. Med. Chem. 50, 455-461 19. Chen, M. M. Y., Snow, C. D., Vizcarra, C. L., Mayo, S. L., and Arnold, F. H. (2012) Comparison of random mutagenesis and semi-rational designed libraries for improved cytochrome P450 BM3-catalyzed hydroxylation of small alkanes. Protein Engineering Design and Selection 25, 171-178 20. Whitehouse, C. J. C., Bell, S. G., and Wong, L.-L. (2012) P450BM3 (CYP102A1): connecting the dots. Chem. Soc. Rev. 41, 1218-1260

188

21. Butler, C. F., Peet, C., Mason, A. E., Voice, M. W., Leys, D., and Munro, A. W. (2013) Key mutations alter the cytochrome P450 BM3 conformational landscape and remove inherent substrate bias. J. Biol. Chem. (in press) 22. Noble, M. A., Miles, C. S., Chapman, S. K., Lysek, D. A., MacKay, A. C., Reid, G. A., Hanzlik, R. P., and Munro, A. W. (1999) Roles of key active-site residues in flavocytochrome P450 BM3. Biochem. J. 339, 371-379 23. Omura, T., and Sato, R. (1964) The carbon monoxide-binding pigment of liver microsomes II. Solubilization, Purification, and properties. J. Biol. Chem. 239, 2379-2385 24. Perera, R., Sono, M., Sigman, J. A., Pfister, T. D., Lu, Y., and Dawson, J. H. (2003) Neutral thiol as a proximal ligand to ferrous heme iron: implications for heme proteins that lose cysteine thiolate ligation on reduction. Proc Natl Acad. Sci. U.S.A 100, 3641-3646 25. Dunford, A. J., McLean, K. J., Sabri, M., Seward, H. E., Heyes, D. J., Scrutton, N. S., and Munro, A. W. (2007) Rapid P450 Heme Iron Reduction by Laser Photoexcitation of Mycobacterium tuberculosis CYP121 and CYP51B1: Analysis of CO complexation reactions and reversibility of the P450/P420 equilibrium. J. Biol. Chem. 282, 24816-24824 26. Girvan, H. M., Seward, H. E., Toogood, H. S., Cheesman, M. R., Leys, D., and Munro, A. W. (2007) Structural and spectroscopic characterization of P450 BM3 mutants with unprecedented P450 heme iron ligand sets. New heme ligation states influence conformational equilibria in P450 BM3. J. Biol. Chem. 282, 564-572 27. Morrison, J. F. (1969) Kinetics of the reversible inhibition of enzyme- catalysed reactions by tight-binding inhibitors. Biochim. Biophys. Acta. 185, 269-286 28. Bui, S. H., McLean, K. J., Cheesman, M. R., Bradley, J. M., Rigby, S. E., Levy, C. W., Leys, D., and Munro, A. W. (2012) Unusual spectroscopic and ligand binding properties of the cytochrome P450-flavodoxin fusion enzyme XplA. J. Biol. Chem. 287, 19699-19714 29. Kabsch, W. (1993) Automatic processing of rotation diffraction data from crystals of initially unknown symmetry and cell constants. J. Appl. Crystallogr. 26, 795-800 30. McCoy, A. J., Grosse-Kunstleve, R. W., Adams, P. D., Winn, M. D., Storoni, L. C., and Read, R. J. (2007) Phaser crystallographic software. J. Appl. Crystallogr. 40, 658-674

189

31. Murshudov, G. N., Skubak, P., Lebedev, A. A., Pannu, N. S., Steiner, R. A., Nicholls, R. A., Winn, M. D., Long, F., and Vagin, A. A. (2011) REFMAC5 for the refinement of macromolecular crystal structures. Acta Crystallogr., Sect. D: Biol. Crystallogr. 67, 355-367 32. Emsley, P., and Cowtan, K. (2004) Coot: model-building tools for molecular graphics. Acta Crystallogr., Sect. D: Biol. Crystallogr. 60, 2126-2132 33. Haines, D. C., Tomchick, D. R., Machius, M., and Peterson, J. A. (2001) Pivotal role of water in the mechanism of P450BM-3. Biochemistry 40, 13456-13465 34. Loida, P. J., and Sligar, S. G. (1993) Molecular recognition in cytochrome P- 450: mechanism for the control of uncoupling reactions. Biochemistry 32, 11530-11538 35. Miles, J. S., Munro, A. W., Rospendowski, B. N., Smith, W. E., McKnight, J., and Thomson, A. J. (1992) Domains of the catalytically self-sufficient cytochrome P-450 BM-3. Genetic construction, overexpression, purification and spectroscopic characterization. Biochem. J. 288 ( Pt 2), 503-509 36. Kuper, J., Tee, K. L., Wilmanns, M., Roccatano, D., Schwaneberg, U., and Wong, T. S. (2012) The role of active-site Phe87 in modulating the organic co-solvent tolerance of cytochrome P450 BM3 monooxygenase. Acta Crystallogr. F-Struct. Biol. Cryst. Commun. 68, 1013-1017 37. Wong, T. S., Arnold, F. H., and Schwaneberg, U. (2004) Laboratory evolution of cytochrome p450 BM-3 monooxygenase for organic cosolvents. Biotechnol. Bioeng. 85, 351-358 38. Äbelö, A., Andersson, T. B., Antonsson, M., Naudot, A. K., Skånberg, I., and Weidolf, L. (2000) Stereoselective Metabolism of Omeprazole by Human Cytochrome P450 Enzymes. Drug Metab. Dispos. 28, 966-972 39. Shiohira, H., Yasui-Furukori, N., Yamada, S., Tateishi, T., Akamine, Y., and Uno, T. (2012) Hydroxylation of R(+)- and S(−)-Omeprazole after Racemic Dosing are Different among the CYP2C19 Genotypes. Pharm. Res. 29, 2310-2316 40. Turck, D., Berard, H., Fretault, N., and Lecomte, J. M. (1999) Comparison of racecadotril and loperamide in children with acute diarrhoea. Alimentary Pharmacology & Therapeutics 13, 27-32 41. Yasuda, S., Horai, Y., Tomono, Y., Nakai, H., Yamato, C., Manabe, K., Kobayashi, K., Chiba, K., and Ishizaki, T. (1995) Comparison of the kinetic disposition and metabolism of E3810, a new proton pump inhibitor, and

190

omeprazole in relation to S-mephenytoin 4[prime]-hydroxylation status. Clin. Pharmacol. Ther. 58, 143-154 42. Nakai, H., Shimamura, Y., Kanazawa, T., Yasuda, S., and Kayano, M. (1994) Determination of a new H+—K+ ATPase inhibitor (E3810) and its four metabolites in human plasma by high-performance liquid chromatography. J. Chromatogr., B: Anal. Technol. Biomed. Life Sci. 660, 211-220 43. VandenBranden, M., Ring, B. J., Binkley, S. N., and Wrighton, S. A. (1996) Interaction of human liver cytochromes P450 in vitro with LY307640, a gastric proton pump inhibitor. Pharmacogenetics and Genomics 6, 81-91 44. Renberg, L., Simonsson, R., and Hoffmann, K. J. (1989) Identification of two main urinary metabolites of [14C]omeprazole in humans. Drug Metab. Dispos. 17, 69-76 45. Pearce, R. E., Rodrigues, A. D., Goldstein, J. A., and Parkinson, A. (1996) Identification of the human P450 enzymes involved in lansoprazole metabolism. J. Pharmacol. Exp. Ther. 277, 805-816 46. Lin, Y., Lu, P., Tang, C., Mei, Q., Sandig, G., Rodrigues, A. D., Rushmore, T. H., and Shou, M. (2001) Substrate Inhibition Kinetics for Cytochrome P450- Catalyzed Reactions. Drug Metab. Dispos. 29, 368-374 47. Korzekwa, K. R., Krishnamachary, N., Shou, M., Ogai, A., Parise, R. A., Rettie, A. E., Gonzalez, F. J., and Tracy, T. S. (1998) Evaluation of Atypical Cytochrome P450 Kinetics with Two-Substrate Models: Evidence That Multiple Substrates Can Simultaneously Bind to Cytochrome P450 Active Sites†. Biochemistry 37, 4137-4147 48. DellaGreca, M., Iesce, M. R., Previtera, L., Rubino, M., Temussi, F., and Brigante, M. (2006) Degradation of lansoprazole and omeprazole in the aquatic environment. Chemosphere 63, 1087-1093 49. Li, H., and Poulos, T. L. (1997) The structure of the cytochrome p450BM-3 haem domain complexed with the fatty acid substrate, palmitoleic acid. Nat. Struct. Mol. Biol. 4, 140-146 50. Robison, T. W., and Jacobs, A. (2009) Metabolites in safety testing. Bioanalysis 1, 1193-1200 51. Christopher R. Otey, Geethani Bandara, James Lalonde, Katsuyuki Takahashi, and Frances H. Arnold. (2006) Preparation of human metabolites of propranolol using laboratory-evolved bacterial cytochromes P450. Biotechnol. Bioeng. 93, 494-499 52. Tsotsou, G. E., Sideri, A., Goyal, A., Nardo, G. D., and Gilardi, G. (2012) Identification of Mutant Asp251Gly/Gln307His of Cytochrome P450 BM3 for

191

the Generation of Metabolites of Diclofenac, Ibuprofen and Tolbutamide. Chem. Eur. J. 53. Andersson, T., RÖhss, K., Bredberg, E., and Hassan-Alin, M. (2001) Pharmacokinetics and pharmacodynamics of esomeprazole, the S-isomer of omeprazole. Alimentary Pharmacology & Therapeutics 15, 1563-1569 54. Pichard, L., Curi-Pedrosa, R., Bonfils, C., Jacqz-Aigrain, E., Domergue, J., Joyeux, H., Cosme, J., Guengerich, F. P., and Maurel, P. (1995) Oxidative metabolism of lansoprazole by human liver cytochromes P450. Mol. Pharmacol. 47, 410-418 55. Ren, S., Park, M.-J., Sah, H., and Lee, B.-J. (2008) Effect of pharmaceutical excipients on aqueous stability of rabeprazole sodium. Int. J. Pharm. 350, 197-204 56. Ward, R. M., and Kearns, G. L. (2013) Proton pump inhibitors in pediatrics : mechanism of action, pharmacokinetics, pharmacogenetics, and pharmacodynamics. Paediatr. Drugs 15, 119-131

192

4.0 Oxidation of diverse drug molecules by P450 BM3 gatekeeper variants

1Christopher F. Butler, 2Caroline Peet, 1Amy E. Mason, 1Karl Fisher, 2Michael W. Voice, 1Steve E Rigby, 1David Leys and 1Andrew W. Munro*

1Manchester Institute of Biotechnology, Faculty of Life Sciences, University of Manchester, 131 Princess Street, Manchester M1 7DN, UK. 2Cypex Ltd, 6 Tom McDonald Avenue, Dundee DD2 1NH, UK.

193

4.1 Abstract

The effect of protein stability on evolution is a contested area of research. Some say protein stability is necessary for activity altering mutations to occur (1), while others say it is the destabilisation caused by activity altering mutations that prevents evolution reversing(2). Research presented in this paper is consistent with the latter model, whereby activity altering P450 BM3 mutations result in the stability of the enzyme being reduced while variant activities towards non-natural substrates increases dramatically. Here we use a fluorescence based protein stability assay to examine WT and variant Bacillus megaterium P450 BM3 enzymes, and to obtain accurate Tm values for thermal unfolding. Interactions are shown between P450 BM3 variants and various drug substrates that are markers for major human P450s, using fluorescence thermal unfolding and EPR methods. Product formation from P450 reactions are also analysed by both steady-state kinetics and turnover analysis by LCMS. Our data provide confirmation that single mutations in BM3 can cause structural destabilization to produce more plastic, flexible enzymes that can accept diverse substrate molecules and that can metabolise compounds that are substrates for different human P450 isoforms. These data provide further proof that thermodynamic destabilisation of the P450 enzyme is the key factor in allowing altered activity, by providing a lower energy barrier to altered substrate binding and oxidation. The diversification of substrate selectivity is suggested to occur by the lowering of the transition energy, enabling the relevant “gatekeeper” BM3 variant enzymes to become more promiscuous. In this study, the BM3 substrate specificity range is widened from fatty acids (the favoured substrates for WT BM3) to drugs such as amodiaquine, nifedipine and dextromethorphan. Human metabolites also predominate in the BM3 gatekeeper variant catalysed oxidation of these drugs, suggesting that these reactions are more energetically favourable for P450 enzymes, and that these reactions will occur preferentially, perhaps even if the binding modes of these compounds enable exposure of a variety of potential target sites to reactive iron-oxo species in the P450 active site.

194

4.2 Introduction

Cytochromes P450 (CYPs, P450s) are responsible for most drug and xenobiotic metabolism (3). They catalyse an array of oxygenase reactions, including steroid metabolism (4), activation of carcinogens (5), and biosynthesis of antibiotics (6). They also catalyse many unusual reactions, such as the nitration of tryptophan in the synthesis of thaxtomin by Streptomyces turgidiscabies and the cyclopropanation of styrene with diazoacetate by P450 BM3 (BM3) (7,8). These reactions are often carried out in highly enantio- and stereo-specific manner. These properties makes P450s an attractive target for producers of niche and fine chemical products, as complex reactions requiring multi-step processes and costly reactions can often be done in a simple, single step manner by P450 enzymes. Many studies have identified BM3 variants that can oxidise human drug molecules and other chemical structures dissimilar to the natural fatty acid substrates for the WT P450 BM3 (9-11). However, a key element lacking is often a clear explanation as to why some or all of the random mutations in BM3 variants generated by directed evolution give rise to major alterations in activity. Recent studies have started to identify the structural basis by which altered activity is conferred in some cases (12,13). Screening for altered activity is a complex process and many assays have been generated to look directly at novel oxidase activity. For example, the purpald assay for de-alkylation reactions (14,15), the γ-(4-nitrobenzyl)pyridine (NBP) and p-nitrothiophenolate (pNTP) epoxide detection assays (16,17), and 4-aminoantipyrine (4-AAP) for aromatic hydroxylations (18). These assays have the combined problem that “you get what you screen for”, i.e. that such screens are often highly specific to a single drug or compound of interest. Companies rarely want a single enzyme to carry out a single reaction, as the cost and time constraints of evolving a new enzyme for each individual reaction becomes prohibitive. Often, companies require a ‘toolkit’ whereby a class of compound or reaction can be targeted by selecting known mutations with predictable effects, thus reducing the time and expense of finding a productive enzyme for a required reaction. This may be the holy grail of enzyme engineering, and could be particularly difficult with P450s given their natural diversity of substrate selectivity. However, by a rational approach we may make strides toward achieving this outcome. In steps toward this overall goal, P450 substrate recognition sequences (SRS) were identified as regions in the general P450 structure that are highly diverse across families (while located around the substrate binding cavity), and thus likely to affect substrate recognition and binding. This is the reason they

195

are often targeted with the aim of producing activity altering variants (19). Here, we have decided to focus on a limited set of mutations found commonly in other BM3 variants produced by direct evolution strategies, order to identify key positions responsible for altered activity, and thus to develop a clearer understanding of how they confer altered BM3 activity with key human P450 marker substrates (20).

Stability has a key place in enzyme redesign, with increased thermal stability often seen as a key measure of an enzyme’s use to industry (1,21,22). However, it is often the case that key productive mutations have the opposite effect, i.e. destabilise the protein (13,23). This observation allows us to pose the question “is destabilisation key to generating an enzyme that will accept altered substrates?” This would follow Dollo’s Law that evolution is irreversible (24), and therefore that mutations conferring altered ability destabilise the native protein, while subsequent mutations return it to stability, preventing a reversal to the native form without crossing this enthalpic barrier (2). This raises the likely possibility that a few key mutations are responsible for altered activity, and that other mutations formed in random mutagenesis/directed evolution strategies either stabilise this “new” protein, or are responsible for “locking” it in a new conformation that disfavours the old (natural) substrate activity. This model leads us to infer that instability is the key and stability is the lock in protein evolution.

Production of drug metabolites is a recent area of academic and industrial interest due to the fact they are needed for pharmaceutical research, compliance testing and for drug-drug and drug-protein interaction studies (25,26). They make an excellent target for biotechnology as synthetic means of their production are costly for such small scale requirements. P450 BM3 from Bacillus megaterium is an excellent tool in the arsenal of protein engineers, with high activity due to its P450-

cytochrome P450 reductase (CPR) redox partner fusion nature (e.g. a kcat of ~17,000 min-1 with arachidonic acid) (27), and since it is soluble, easy to produce, backed by a wealth of structural data to provide insights into mechanism and mode of substrate binding (28-31), and comes with a history of successful mutagenesis studies that have diversified its substrate range (32-34). Current production of drug metabolites frequently uses recombinant human P450s, but introducing BM3 would reduce costs and simplify the process, saving time and money (35). However, this would require engineering of BM3 to introduce desired activities. With this in mind we set out to try and explain why certain mutations confer novel substrate specificity on BM3, and to seek reasons for how A82F and F87V variants exhibit new

196

activities, and whether these are confined to the oxidation of gastric proton pump inhibitor drugs (PPIs) (13), or extend to a wider range of human P450 marker substrates.

To this end, we have utilised a differential scanning fluorimetry technique (DSF) (36) to assess the relative stability of different BM3 variants in the substrate-free state and when bound to the human P450 marker drugs diazepam (DIA), diclofenac (DCF), nifedipine (NIF), amodiaquine (AMO), and dextromethorphan (DEX). We show that these mutations confer altered conformational stability leading to new activities, and provide EPR and kinetic data to assess binding modes and activity, along with product identification by LCMS.

197

4.3 Materials and Methods

4.3.1 Mutagenesis and expression of WT and variant P450 BM3 enzymes. The gene encoding intact WT flavocytochrome P450 BM3 in pET15b was mutagenized to create A82F, F87V and F87V/A82F (DM) variants, as described in previous studies (13). Intact BM3 enzymes were expressed as N-terminal hexahistidine tagged enzymes either from pET15b (F87V, DM) constructs directly, or after cloning the WT and A82F genes into pET14b using NdeI/BamHI sites. WT and variant heme domain genes were generated using the relevant pET14b/15b constructs, as described previously (13). Genes encoding the heme domains (amino acids 1-473 of the 1048 amino acid P450 BM3) were transferred as NdeI/BamHI fragments to pET20b for heme domain production without the N- terminal His-tag, and for improved crystallization. Whole gene sequencing ensured only the correct mutations had been introduced. The WT and A82F intact BM3, and WT and all variant P450 BM3 heme domains were expressed in BL21-Gold(DE3) E. coli cells (Stratagene-Agilent UK) in Terrific Broth with cells grown at 37ºC, and with shaking at 200 RPM in an orbital incubator. F87V and DM intact BM3 proteins were grown using autoinduction TB medium (Melford Ltd, Ipswich UK) from 4 L transformant cultures with cell growth for 24-36 hours.

4.3.2 Purification of WT and variant P450 BM3 and heme domains. Intact WT and variant P450 BM3 (BM3) and heme domains were purified as previously (13). Cells were collected by centrifugation at 4 °C (6000 g, 10 min) and resuspended in ice-cold buffer B (50 mM KPi, 250 mM NaCl, 10% glycerol, pH 7.0). Protease inhibitors (EDTA Free CompleteTM tablets, Roche, Germany) were kept in all buffers during purification. Cells were lysed by sonication, and supernatant containing soluble intact BM3/heme domain proteins was collected after centrifugation (20,000 g, 40 min, 4 °C), The supernatant was again retained after a 30% ammonium sulfate cut on ice. P450 proteins were purified by Ni-IDA chromatography (Qiagen, UK). Bound proteins were washed extensively at 4 °C in buffer B plus 5 mM imidazole, then eluted with 200 mM imidazole in buffer B. Proteins purified in this way were transferred into buffer A (50 mM Tris, 1 mM EDTA, pH 7.2) and passed through a Sephacryl S-200 SEC column (GE Healthcare, 26 x 60 cm on an AKTA purifier). Pure BM3 fractions (verified by SDS- PAGE) were concentrated by ultrafiltration (Vivaspin, Vivaproducts, USA) and stored in buffer A plus 50% glycerol at -80 °C. For non-tagged heme domains, a

198

further 30-60% ammonium sulfate cut was applied. The P450-containing pellet was resuspended in buffer A and dialyzed into the same buffer to desalt, then protein was further purified by ion exchange chromatography using a Q-Sepharose column (16 x 10 cm), with elution in a gradient of 0-500 mM KCl in buffer A. Heme domain fractions were desalted (GE Healthcare column, 26 x 10 cm on an AKTA) into 25 mM KPi pH 7.0, loaded onto a hydroxyapatite column (Bio-Rad, USA, 16 x 11 cm) and eluted in a 200 mL gradient of 25–500 mM KPi, pH 7.0. Pure heme domains were concentrated by ultrafiltration (Vivaspin) and used immediately for crystallography, or flash frozen in liquid nitrogen and stored at -80 °C. Intact BM3 and heme domain proteins with the A82F mutation were passed through a Lipidex 1000 column (Perkin Elmer, UK) in 25 mM KPi, pH 7.0 to remove fatty acid retained during purification (13).

4.3.3 P450 quantification.

Concentrations of WT/variant forms of intact BM3 and heme domain proteins were -1 -1 defined using extinction coefficients of ɛ418 = 105 and ɛ419 = 95 mM cm , respectively, at the Soret maximum (13,37). Fe(II)CO complexes were formed by bubbling sodium dithionite-reduced WT/variant BM3 and heme domains (ca 2-4 µM) with CO gas (38). WT and all variant enzymes showed almost complete formation of the P450 (thiolate-coordinated) state, with little P420 (likely cysteine thiol- coordinated) formed in any case (39,40).

4.3.4 Thermofluor binding assay.

Thermofluor binding was carried out on a Bio-Rad CFX96 real-time system C100 thermal cycler, using 96-well RT-PCR plates (Bio-Rad Laboratories, Hercules, CA, USA). Excitation and emission wavelengths of 490 and 575 nm, respectively, were used for the measurements. A temperature gradient of 15 to 95 °C was used with a fluorescence measurement after every 0.2 °C increase and with a 5 s delay for signal stabilisation. Fluorescence is measured and plotted (as a first derivative)

against temperature, enabling determination of the melting temperature (Tm).

Experiments were repeated twice, and the reported Tm values are the mean values determined from the peaks of the derivatives of the experimental data. In the drug screening experiment, a final protein concentration of 228 μM (heme domain) was mixed with 7.5 μL of 300x SYPRO orange dye (12.5% DMSO in dH2O) (90x final), 12.5 μL of 100 mM KPi buffer (50 mM final concentration), with 2 µL methanol or 2

199

µL drug in methanol, directly into the plate wells (on ice). The mixture was then diluted with distilled, deionised H2O to a final volume of 25 μL. Initially, a range of protein and dye concentrations were tested to find optimal conditions (i.e. a good signal-to-noise ratio). Studies were done using the BM3 heme domains, due to fluorescence from bound and free (in case of FMN) flavins giving overlapping signals with those from the SYPRO fluorophore.

4.3.5 EPR Spectroscopy

EPR spectra for the ligand-free and drug-bound WT and variant BM3 heme domain enzymes were recorded on a Bruker ER-300D series electromagnet with microwave source interfaced with a Bruker EMX control unit and fitted with an ESR-9 liquid helium flow cryostat (Oxford Instruments), and a dual-mode microwave cavity from Bruker (ER-4116DM). Spectra were recorded at 10 K with a microwave power of 2.08 mW and modulation amplitude of 1 mT. Samples consisting of protein (200 μM) in KPi buffer (100 mM, pH 7.0) were prepared, with 1.8 µl methanol and 400 µM drug in 1.8 µl methanol added for consistency. EPR spectra were collected by Dr Stephen Rigby (University of Manchester).

4.3.6 Analysis of the kinetics of substrate-dependent NADPH oxidation by WT and variant intact BM3 enzymes.

Kinetic studies were performed on a Cary 50 UV-Visible spectrophotometer by

following substrate (drug)-dependent oxidation of NADPH at 340 nm (∆ε340 = 6.21 mM-1 cm-1). BM3 enzyme concentration was kept constant (in range from 50-100 nM) with substrate concentration varied and NADPH at a near-saturating concentration (200 µM). Assays were done at 25 °C in assay buffer using a 1 cm pathlength quartz cuvette. Initial rates were determined in triplicate at all substrate concentrations at 340 nm Rate constants were plotted versus substrate concentration and data were fitted using the Michaelis-Menten (or Hill function) using Origin Pro software to produce the kcat and Km (or KH in for the Hill function – the apparent midpoint concentration from a sigmoidal fit) parameters for substrate- dependent NADPH oxidation, as reported in Table 2.

200

4.3.7 Drug turnover and analysis by LCMS

Turnover reactions for oxidation of all drugs tested were done in deep well blocks at 37 oC with shaking for 30 min. Reaction mixtures contained purified WT or variant (F87V, A82F or F87V/A82F [DM] BM3) enzymes (1 µM), substrate amodiaquine (AMO), dextromethorphan (DEX), diazepam (DIA), diclofenac (DCF), or nifedipine (NIF), (10 µM), NADPH regeneration system (glucose-6-phosphate 7.76 mM, NADP+ 0.6 mM, and glucose-6-phosphate dehydrogenase 0.75 U/ml) in turnover buffer (50 mM KPi, 5 mM CaCl2, pH 7.4) in a final volume of 500 µl. On completion of the reaction, protein was mixed with an equal volume of acetonitrile (ACN) containing 1 µg/ml fluconazole internal standard (IS) by shaking the mixed samples at 800 rpm for 10 min. Precipitated protein was filtered through protein precipitation plates (Phenomenex, Macclesfield UK) into mass spectrometry vials (FluidX, Nether Alderley UK) and clarified by centrifugation (4000 g, 25 min, 10 oC). Analysis was done on a Thermo Exactive LC-MS with a CTC PAL auto sampler (Thermo Scientific, UK) with a Kinetex 2.6U XB – C18 100A column (Phenomenex). A gradient of 0.1% formic acid to ACN was used to resolve products. Drugs and metabolites were run in positive mode with the molecular ion as M[H]+. All high intensity peaks were selected from the total ion chromatogram and analyzed using Thermo Xcalibur quantification software, along with the fluconazole internal standard (IS). This software then gave total ion readings for both the IS and the metabolites formed. The total ion data were then corrected for the IS.

201

4.4 Materials

Oligonucleotide primers for mutagenesis were from Eurofins MWG Operon (Ebersberg, Germany). All drugs and standards were provided by Cypex Ltd (Dundee, UK), Sigma-Aldrich (Gillingham, UK) or Santa Cruz (Dallas, USA). Bacterial growth medium (Terrific Broth, TB) was from Melford Ltd (Ipswich, UK). Unless specified, other chemicals were from Sigma-Aldrich and of the highest purity available.

202

4.5 Results

4.5.1 Thermal unfolding of WT and variant P450 BM3 heme domains using a Thermofluor assay

In previous work, the binding of omeprazole and other PPI drugs was shown to induce the development of HS heme iron in the P450 BM3 F87V, A82F and F87V/A82F (DM) variants. However, binding studies using a panel of human drugs as potential substrates/ligands did not induce significant heme optical shift to enable

Kd determination by this method. As our tested substrates (diazepam [DIA], diclofenac [DCF], nifedipine [NIF], amodiaquine [AMO] and dextromethorphan [DEX]) proved unresponsive to P450 optical binding assays, we instead decided to screen for binding based on a thermal unfolding assay. In previous studies, we demonstrated that the A82F and DM variants were considerably thermodynamically destabilized and had lower thermal melting temperatures (Tm values) than the WT by differential scanning calorimetry (DSC), while the F87V variant was much less affected. However, fatty acid substrate (N-palmitoylglycine, NPG) binding increased

the Tm for WT and variant BM3 enzymes, while omeprazole (OMP) substrate

binding also stabilized the F87V and DM P450s, with ~2 °C increases in their Tm values. In studies with the same variant heme domains, we used different scanning fluorimetry (DSF – using the thermofluorTM method) and showed that the drug compounds tested interacted with the BM3 proteins, as evidenced by their causing a shift in their thermal transition midpoint (Tm). As can be seen in Table 1 the WT heme domain is unaffected by either the addition of solvent (MeOH) or by the addition of any of the drugs. By DSF, the A82F and DM had greatly reduced thermal stability compared to the WT and F87V enzymes, consistent with previously reported DSC data for these proteins (13). Both the A82F and DM enzymes showed significant (>1 °C) further destabilisation by the solvent (MeOH), whereas the F87V variant was little affected. These data suggest that the destabilizing mutation A82F (which causes altered P450 BM3 confirmation) also reduces its stability to MeOH solvent.

203

Substrate/Solvent WT Tm (°C) AF Tm (°C) FV Tm (°C) DM Tm (°C)

None 59.9 ± 0.2 53.1 ± 0.1 59.6 ± 0.1 54.6 ± 0.2

MeOH 59.7 ± 0.1 51.8 ± 0.1 59.3 ± 0.1 52.9 ± 0.1

DIA 59.3 ± 0.1 48.5 ± 0.4 56.7 ± 0.1 50.4 ± 0.1

DCF 59.7 ± 0.1 51.5 ± 0.1 59.5 ± 0.1 52.2 ± 0.1

NIF 59.9 ± 0.2 50.4 ± 0.1 56.9 ± 0.2 52.3 ± 0.2

AMO 59.7 ± 0.1 51.8 ± 0.1 59.5 ± 0.1 51.2 ± 0.2

DEX 59.2 ± 0.1 49.5 ± 0.1 57.1 ± 0.1 51.6 ± 0.1

Table 4.1 Transition midpoint (Tm) values for WT and variant BM3 heme domains using a thermal unfolding assay. Temperatures are in °C, and are the average of two repeats for the WT and variant BM3 heme domains with drug substrates. Diazepam (DIA), diclofenac (DCF), nifedipine (NIF), amodiaquine (AMO) and dextromethorphan (DEX). Errors shown are SEM.

In previous studies using OMP, this novel substrate for the BM3 gatekeeper variants

bound strongly to these proteins (Kd values of 1.67 µM and 0.21 µM for the A82F and DM variants, respectively) and induced increases in the Tm value. However, since none of our panel of drugs induced any significant level of HS heme iron development, it was unclear at this stage if the lack of optical binding evidence indicated failure to access the P450 active sites, or else reflected weak binding and/or interactions at distinct positions within the P450 active sites. As can been seen from the data (Table 4.1, Figure 4.1) the A82F variant actually showed significant thermodynamic destabilisation with DIA, NIF and DEX. This suggests that there is likely to be weak binding, and possibly interactions at diverse protein sites with the ligands, and either in folded or partially unfolded states of the P450. The F87V variant showed significant destabilisation with DIA, NIF and DEX, also indicating binding of these ligands to this protein. The DM also showed destabilisation with DIA, AMO and DEX, likewise indicating interactions of the protein with these ligands.

204

Figure 4.1. P450 enzyme stability using the ThermofluorTM method. The DSF traces shown are the first derivatives of fluorescence intensity (associated with the thermal unfolding of the BM3 heme domain proteins) against temperature (°C). Panel A: WT; Panel B: A82F; Panel C: F87V: Panel D: DM BM3. Traces shows are for no additive (black), and with addition of methanol (red), diazepam (blue), diclofenac (pink), nifedipine (green), amodiaquine (navy), and dextromethorphan (purple). The figure was produced in Origin Pro and Adobe Photoshop.

4.5.2 Electron Paramagnetic Resonance spectrometry (EPR)

The interaction of the various drugs with the WT and variant P450 BM3 enzymes was further probed using EPR spectrometry with the heme domains. EPR relies on the paramagnetic state of the ferric heme iron giving a defined signal, allowing the identification of both the spin state and any changes in ligation to the heme iron. It is interesting to note the effect that the drug solvent has on the EPR signal. Although <1% (v/v) solvent was used in these experiments, there was a specific effect on the heme EPR signal for both the WT BM3 heme domain and each of the variant

205

enzymes. The WT BM3 spectrum shows a shift in the low spin g-values from gz =

2.41, gy = 2.25, gx = 1.92 to 2.40/2.25/1.92 on addition of MeOH solvent. This very small gz shift (Data not shown) is repeated with each of the drug molecules dissolved in MeOH. This indicates that the observed shift is caused by a solvent interaction with the heme iron or by change in protein solvation state, and likely not by the drugs’ action. These solvent effects were also noted in studies of the other BM3 heme domain variants.

The EPR data for the A82F heme domain are distinct from those of the WT protein, as there is a proportion of HS ferric heme in the substrate/solvent free state (g- values = 8.04/3.44/1.45). This indicates that the A82F protein is in a mixed spin state with a small proportion of HS heme iron, even though the UV-visible spectrum is dominated by the LS form. As with WT heme domain, there is a shift in the EPR spectrum on addition of MeOH, primarily by loss of the HS species, but also through an apparent loss of one of the LS species (2.47/2.25/1.90), with two LS species remaining at 2.44/2.25/1.91 and 2.42/2.25/1.90 (Figure 4.2). The DEX, DCF and NIF ligands had no obvious effect on the EPR spectra over that seen from MeOH alone. The greatest effect is that seen with DIA, where there are four observed LS EPR species, with three of these being distinct from those seen for the A82F spectum collected with MeOH alone. These are seen at 2.52/2.25/1.90, 2.44/2.25/1.91 and 2.39/2.25/1.92. These data indicate that DIA likely affects the

environment of H2O ligand in the 6-coordinate water bound species, and possibly

(for the gz = 2.52 form) interacts directly (or indirectly via the water ligand) with the heme iron via the DIA nitrogen atom. These EPR data confirm that DIA enters the active site of A82F BM3, and is consistent with optical spectroscopy data, suggesting that no displacement of the water ligand may occur. The other ligand that has an effect on the A82F heme domain is AMO. This drug also generates additional LS EPR species. Two of the three species are novel, at 2.52/2.25/1.90, and 2.44/2.25/1.91. These g-values are similar to those species formed by DIA, but different to those seen for MeOH alone. As with DIA, these data indicate that AMO likely interacts with the 6th ligand water in at least one of these species, and that one of the AMO nitrogen atoms may interact with the

heme iron for the 2.52 gz species.

For the F87V variant there are three LS species observed, two of approximately equal intensity and one of higher intensity in the substrate-free enzyme at 2.48/2.25/1.87, 2.45/2.25/1.90 and 2.41/2.25/1.92, respectively. Upon MeOH

206

addition, the triplet of EPR signals shows small shifts to 2.48/2.25/1.90,

2.43/2.25/1.91 and 2.40/2.25/1.92. The gz = 2.43 species is now dominant, with the others of lower intensity. As for the A82F variant, only two of the drugs have notable effects on the EPR spectra that are distinct from those of MeOH. The first (and most striking) is again with DIA, which causes small shifts in the g values and an apparent increase in intensity of the central species (2.42/2.25/1.92), while diminishing the proportions of the other two LS EPR signals at 2.47/2.25/1.90 and 2.39/2.25/1.93. The second compound that gives obvious effects on the F87V spectrum is DCF, which does not alter the g-values of the species observed (2.48/2.25/1.90, 2.43/2.25/1.91 and 2.40/2.25/1.92), but again alters their relative quantities, with the first species becoming more prominent and the central one diminished. As mentioned previously, these data indicate an alteration of the heme water ligation state caused by the interacting ligands.

Finally, the DM heme domain shows four different LS heme ligation states when substrate-free at 2.53/2.25/1.87, 2.48/2.25/1.90, 2.44/2.25/1.91 and 2.40/2.25/1.92, indicating multiple low spin confirmations. When MeOH is added, three species are seen at 2.52/2.25/1.87, 2.47/2.25/1.90 and 2.43/2.25/1.91. Again the most interesting ligand is DIA, with four species produced at 2.52/2.25/1.86, 2.47/2.25/1.90, 2.42/2.25/1.92 and 2.39/2.25/1.93. These species are all of similar intensity. The other most notable shift in the EPR spectra came from DCF, which decreased the complexity of the EPR signal from four species in the drug-free form to only two species in the DCF-bound DM heme domain. The two DCF-bound EPR species are at 2.47/2.25/1.90 and 2.43/2.25/1.91. These sets of g-values are also distinct from any obtained for the ligand-free DM protein. The decreased complexity of the DCF-bound signal suggests that there are only two major heme coordination states, and that the water ligand is retained on the heme iron in both, albeit in different environments and affected by the DCF ligand.

The fact that DIA gives rise to multiple species in the EPR spectra of the BM3 variants is likely consistent with the DSF data for the DIA-bound forms, where DIA destabilises the proteins. This may be due to several different binding positions for the drug, at least some of which occur in less stable conformations of the protein.

207

Figure 4.2. EPR spectra of ligand-free and drug-bound WT and variant P450 BM3 heme domains. A: A82F BM3 (black), A82F with MeOH (red), A82F with DIA (blue), A82F with AMO (pink). B: F87V BM3 (black), F87V with MeOH (red), F87V with DIA (blue) F87V DCF (pink). C: DM BM3 (black), DM with MeOH (red), DM with AMO (blue), DM with DCF (pink). D: DIA-bound spectral overlay for WT (black), AF (red), FV (blue), DM (pink). All heme domains are at 200 µM concentration, MeOH was added at 1.8 µL, and drugs at 400 µM final concentration in 1.8 µL MeOH – added to a final volume of 225 µL in each sample. Spectra were recorded at 10 K with a microwave power of 2.08 mW and modulation amplitude of 1 mT.

These EPR data show that even small amounts of MeOH solvent have effects on the heme iron ligation state, even if this is not observed through significant heme absorbance changes in the UV-visible spectra. These perturbations likely from changes in the solvation state of the active site pocket. With respect to the drugs themselves, influences of some of the potential substrates on the heme is clear from EPR spectroscopy, while not being obvious in UV-visible spectra. In some cases,

208

the EPR spectra are indicative of the retention of the 6th ligand water molecule in presence of the drug, suggesting that the drug can affect the orientation of the water ligand, and that several conformations may be seen even at the cryogenic

temperatures use for heme EPR. For cases where the gz is >2.5, there is also the possibility that a nitrogen ligand from the relevant drug coordinates to the heme iron and displaces the water ligand.

4.5.3 Steady state kinetic analysis of drug-dependent NADPH oxidation

Steady-state kinetic analysis of substrate-stimulated NADPH oxidation was carried out at various drug concentrations to define the Km (i.e. apparent substrate Kd under

turnover conditions) and kcat values. This was not possible for AMO and NIF, as these are highly colored compounds that give intense signals in the 300-400 nm range, and mask the NADPH absorption change. Therefore, kinetic studies were feasible only for DIA, DCF and DEX. Consistent with the lack of effects observed using UV-visible and EPR spectroscopy, the WT BM3 showed no increase in NADPH oxidation over the background rate with any of the drugs tested. The A82F variant showed enhanced oxidation with both DIA and DCF (Table 4.2, Figure 4.3), -1 -1 with kcat values of 427 min and 238 min , respectively. While much lower than values typical for activity of WT BM3 with fatty acids (e.g. ~5000 min-1 with lauric acid), these turnover numbers are substantially higher than kcat values for human P450s with these drugs – which are of the order of ~10 min-1 with their CPR partner.

The Km values were 83 µM for DIA and 106 µM for DCF, suggesting that the single A82F “gatekeeper” mutation in BM3 enables development of significant affinity for these drugs. The F87V variant also showed substrate-stimulated NADPH oxidation kinetics with -1 -1 DIA and DCF, with kcat values of 185 min and 365 min , respectively. The relative

Km values were 427 and 326 µM. Finally, the DM produced sets of kinetic data for -1 DCF and DEX. The DM BM3 had the fastest kcat of all, with DCF at 636 min , and

the lowest Km at 54.6 µM. With DEX, enzyme-catalyzed NADPH oxidation showed a sigmoidal dependence on drug concentration. Therefore data were fitted using both the Michaelis Menten and Hill (sigmoidal) functions. The Michaelis Menten fitting -1 gave a kcat of 188 min and a Km of 108 µM, while fitting using the Hill function gave -1 broadly similar data with a kcat of 155 min and an apparent Km (KH) of 98.4 µM.

However, the statistical error on the KH value is much lower than that for the Km, indicating that fitting with the Hill function is superior (Table 4.2). The Hill coefficient

209

was 2.43, perhaps suggesting that cooperativity in binding of DEX substrate molecules occurs.

-1 Enzyme Substrate kcat (min ) Km (µM) KH (µM) A82F DCF 427± 20 83.2 ± 12.1 ---- A82F DIA 238 ± 10 106 ± 18 ---- F87V DCF 185 ± 11 427 ± 101 ---- F87V DIA 365 ± 33 326 ± 58 ---- DM DCF 636 ± 52 54.6 ± 12.1 ---- DM DEX 188 ± 18 108 ± 39 ---- DM DEX (Hill) 155 ± 10 98.4 ± 13.7 2.43 ± 0.94

Table 4.2. NADPH dependent drug oxidation kinetics for WT and variant P450

BM3 enzymes. The kcat and Km data are derived from the kinetics of WT and variant P450 BM3-dependent NADPH oxidation, stimulated by the binding of the drug substrates diazepam (DIA), diclofenac (DCF) and dextromethorphan (DEX). All data points were collected in triplicate. Kinetic data were fitted to the Michaelis Menten function for all data sets, with the exception of experiments for the BM3 DM with DEX, where the dependence of NADPH oxidation rate on [DEX] was sigmoidal. In this case, the data were fitted using the Hill function to determine the apparent limiting rate (kcat) and the substrate concentration equal to 0.5 x kcat (KH).

210

Figure 4.3. Steady state kinetics of drug-dependent NADPH oxidation for BM3 variants. A: A82F BM3 with DCF. B: A82F BM3 with DIA. C: F87V BM3 with DCF. D: F87V BM3 with DIA. E: DM BM3 with DCF. F: DM BM3 with DEX. All assays were performed with intact P450 BM3 enzymes (50-100 nM) across a range of drug substrate concentrations. Assays were performed in triplicate at each substrate concentration, with error bars showing the SEM. Data are fitted to the Michaelis- Menten equation in A-E and the insert panel in F. The main section of panel F

211

shows the fitting of the DM DEX data using the Hill function. The individual panels were generated in Origin Pro, and the overall figure compiled in Adobe Photoshop.

212

4.5.4. Product identification from P450 BM3 turnover of drug substrates by LCMS

Figure 4.4. Oxidation of drugs by P450 BM3 variants. The reaction schemes show the conversion of drugs to the relevant metabolite by WT P450 BM3 and its

213

A82F, F87V and DM variants (blue), and by the major human P450 isoform (red). A: amodiaquine to N-desethylamodiaquine (CYP2C8), B: dextromethorphan to dextrorphan (CYP2D6), C: diazepam to N-desmethyldiazepam (CYP2C19/CYP3A4), D: diclofenac to 5-hydroxydiclofenac (CYP3A4), E: nifedipine to oxidised nifedipine (CYP3A4). The figure was drawn in Chemdraw and compiled in Adobe Photoshop.

To validate drug oxidation by the BM3 enzymes, in vitro turnover studies were carried out with product identification by LCMS by the use of MS fragmentation and product standards. The WT enzyme was ineffective against the majority of substrates, though ~10% conversion of nifedipine to the oxidised nifedipine metabolite was identified. Products were identified by MS/MS fragmentation and LC retention time in comparison to metabolite standards. It is interesting to note that the major product produced in each case was a human metabolite produced by one of the main human drug-metabolizing P450 (Figure 4.4). We postulate that this likely occurs due to these being the lowest energy or most favourable P450 reactions, and therefore the primary targets for an enzyme able to bind these drugs in different conformations. Nifedipine was by far the best substrate for our engineered enzymes, with between 10-75% conversion to the oxidised metabolite according to the variant used. This is an interesting reaction as it is a 2 electron oxidation to form an aromatic ring, and appears unusual for a P450 to perform, though mechanistically very simple (41). In the LCMS the nifedipine starting material is seen at a retention time (RT) of 3.02 min, with both M[H+] m/z 347.1233 and M[K+] 385.0790, a fragment is also seen showing loss of a methoxy group at 315.0972. The oxidised nifedipine metabolite was seen at RT 2.80 min with M[H+] 345.1075. This corresponds to the RT and mass of the oxidised nifedipine standard (Cypex Ltd, Dundee, UK). Amodiaquine was also highly metabolised by the DM, showing the greatest conversion at ~30%. The WT BM3 showed little turnover of this drug, although the A82F and F87V variants gave ~15% and ~25% conversion respectively. Amodiaquine was seen at RT 2.11 min with an m/z M[H+] 356.1524 and the phenol fragment at 178.5797. Desethyl amodiaquine was seen at RT 2.08 min with an M[H+] of 328.1208, and with a phenol fragment at 164.5643. This mass and fragmentation pattern clearly indicates desethyl amodiaquine. Diclofenac was the next highest metabolised drug, though only to ~10% by the A82F and DM enzymes. Diclofenac was seen at RT 6.76 min in the negative mode mass spectrum at m/z M[-H] 294.0093 with a characteristic Cl splitting pattern, and fragmentation

214

from the loss of the carboxyl to give 250.0197, also with the Cl splitting pattern. The 5-OH diclofenac was seen at RT 6.39 and a M[-H] 310.0041 with the decarboxyl fragment shifted to 266.0144, indicating insertion of oxygen. This was compared with a standard of 5-OH diclofenac and showed the same RT and splitting pattern, conclusively identifying the metabolite as the 5-OH DCF (as generated by CYP3A4) and not the 4-OH DCF (from CYP2C9) (42). Dextromethorphan was oxidised to a similar extent (~10%) by the DM, though to lower extents by the other BM3 variants. Dextromethorphan was seen at RT 5.70 min with an m/z M[H+] 272.2001, and the metabolite eluted close to the parent drug at RT 5.68 min with an M[H+] 258.1852, and was identified as the desmethyl metabolite dextrorphan, as produced by human CYP2D6. The final drug tested, diazepam, showed the least turnover, although there was enough to identify the product by LCMS and standards. Diazepam was seen at RT 4.40 min and an m/z M[H+] 285.0787, and the N-demethylated nordiazepam was identified at RT 3.85 min and 271.0634 (corresponding to the standard RT and mass).

215

Figure 4.5. LCMS traces for P450 BM3-catalysed DM nifedipine (NIF) turnover. Panel A (retention time [RT] = 3.02 min) shows data for NIF prior to addition of enzyme and initiation of its oxidation by the BM3 DM enzyme. Peaks at m/z 347.1233 and 385.0790 correspond to the M[H+] and M[K+] molecular ions, with 315.0972 being a fragment caused by loss of a methoxy group. Panel B (RT = 2.80 min) is following an enzymatic reaction for 30 minutes with the BM3 DM. The m/z peak at 345.1075 corresponds to oxidised nifedipine.

216

Figure 4.6. LCMS traces for P450 BM3 DM-catalysed amodiaquine (AMO) turnover. Panel A (retention time [RT] = 2.11 min) shows data for AMO prior to addition of enzyme and initiation of its oxidation by the BM3 DM enzyme. Peaks at m/z 356.1524 and 178.5797 correspond to the amodiaquine M[H+] and phenol fragments, respectively. The fragmentation is caused by cleavage of the central amino group to form chloroquinoline and amino-phenol fragments. Panel B (RT = 2.08 min) is following an enzymatic reaction for 30 minutes. The m/z peaks at 328.1208 and 164.5643 correspond to desethyl amodiaquine and to the desethyl phenol fragment, respectively.

217

Figure 4.7. LCMS traces for P450 BM3 A82F variant-catalysed diclofenac (DCF) turnover. Panel A (retention time [RT] = 6.76 min) shows data for DCF prior to addition of enzyme and initiation of its oxidation by the BM3 A82F P450. Peaks at m/z 294.0093 and 250.0197 correspond to the M[-H] and the decarboxyl fragment, respectively. The fragmentation is caused by loss of the terminal carboxyl group. Panel B (RT = 6.39 min) is following an enzymatic reaction for 30 minutes. The m/z peaks at 310.0041 and 266.0144 correspond to M[-H] 5-OH diclofenac and to the decarboxylated fragment, respectively.

218

Figure 4.8. LCMS traces for P450 BM3 DM-catalysed dextromethorphan (DEX) turnover. Panel A (retention time [RT] = 5.70 min) shows data for DEX prior to addition of enzyme and initiation of its oxidation by the BM3 DM P450. The peak at m/z 272.2001 corresponds to the M[H+] ion of DEX. Panel B (RT = 5.68 min) is following an enzymatic reaction for 30 minutes. The m/z peak at 258.1852 corresponds to the M[H+] of the demethylated product dextrorphan.

219

Figure 4.9. LCMS traces of P450 BM3 F87V-catalysed diazepam (DIA) turnover. Panel A (retention time [RT] = 4.40 min) shows data for DIA prior to addition of enzyme and initiation of its oxidation by the BM3 F87V enzyme. The peak at m/z 285.0787 corresponds to the M[H+] ion of DIA. Panel B (RT = 3.85 min) is following an enzymatic reaction for 240 minutes. The m/z peak at 271.0634 corresponds to the M[H+] of the desmethyl product nordiazepam.

220

Figure 4.10. Product formation from P450 BM3 oxidation of human drugs. The bar charts show the proportions of products formed by P450 BM3 variants, as identified from LCMS after 30 min (DIA 240 min) incubations and corrected for an internal standard (fluconazole). The data shown are the average of two repeats, with error bars showing the SEM. WT BM3 (black), A82F (red), F87V (blue), DM (orange). A: amodiaquine (AMO) and N-desethylamodiaquine (DeEt). B: diclofenac (DCF) and 5-hydroxydiclofenac (5OH). C: dextromethorphan (DEX) and dextrorphan (DeMe). D: diazepam (DIA) and nordiazepam (NOR). E: nifedipine (NIF) and oxidised nifedipine (Ox NIF).

221

4.6 Discussion

Our previous studies revealed that the A82F and F87V variants of P450 BM3 introduced ability to bind and oxidise the PPI drug omeprazole, and that the combined mutations (the BM3 DM) further stimulated this activity. (13) The A82F mutation was termed a “gatekeeper” in view of its profound effect on the structural and conformational properties of the BM3 heme domain, which enable the enzyme to adopt novel substrate specificity. In previous work, DSC revealed that thermodynamic destabilization was a feature of the gatekeeper variants (evidenced by lowered Tm values), and likely reflected conformational flexibility that enhanced their affinity for omeprazole. In this study, the data from DSF analysis are confirmatory that the structurally destabilising A82F mutation pushes down the Tm values for the A82F and DM BM3 heme domains considerably (Figure 4.1, Table 4.1), but also show that these mutations lead to an increase in the drug metabolite generating ability of these enzymes. It was found previously that DMSO binds in the active site of the F87V variant, and interacts with the heme iron, likely assisting stability (43). For this reason we avoided DMSO as a drug solvent in these studies. As with previous data for binding of fatty acids and omeprazole (13), we would expect tight binding ligands/substrates to give an increased Tm for protein unfolding. However, the effects of weaker binding ligands on P450 stability is less certain, particularly since they may bind to different conformers of the enzyme, some of which may be less thermodynamically stable, and therefore could have both overall

stabilising and destabilising effects on the enzyme. A decrease in the Tm induced by a ligand can be caused by the ligand having a single binding mode to the active protein, but multiple binding modes to unfolded or less stable conformational state(s). Even ligands shown to bind to a protein of interest can cause destabilising effects, as seen when ATP binds to the ATP binding site of the glucose transporter

GLUT1, lowering the Tm by as much as 10 ºC (44). Therefore, although a number of the drugs tested actually had apparent destabilizing effects (i.e. decreased the P450

Tm, in fact none of the drugs tested gave an increase in P450 Tm), this can still be a useful screen for detecting new variants with altered substrate selectivity, as evidenced by capacity of our BM3 variants to oxidise the various drugs studied here. The thermostability assay design has a relatively high solvent content (3.75% DMSO in all reactions from the dye solution, and an additional 8% MeOH in the MeOH controls and drug samples for the rather hydrophobic drugs used), giving 11.75% total solvent. This is relatively high for P450 assay conditions, but

222

nonetheless served to provide clear evidence of structural destabilisation induced on binding of the tested drugs to the BM3 varients.

High co-solvent tolerant BM3 variants have been generated which retain activity at up to 25% DMSO, and with significant amounts of other solvents (45). The addition of MeOH solvent (2 µL) has a measurable effect on the Tm of the A82F-containing variants, but not on the WT and F87V enzymes. The turnover assays used were endpoint assays, and thus no enzyme specific activities are inferred (though these data are presented in drug-stimulated NADPH oxidation assays). However, it is clear that the solvent effects are more profound in the A82F-containing variants, and therefore that this is a destabilising mutation.

Further analyses of the tested drug ligand effects were carried out by electron paramagnetic resonance (EPR) spectroscopy. This technique exploits the fact that unpaired electrons pulsed with microwaves release a photon equivalent in energy to the splitting of its spin half separation. This allows the environment of different unpaired electrons to be analysed and compared. In cytochromes P450 the oxidised (ferric) protein has either 1 or 5 unpaired electron on the heme iron (d-electrons) in the low and high spin forms, respectively. This allows for EPR determination of the P450 heme iron spin state, as well as for any change in heme iron ligation, or for th alterations of the heme active site that perturb the resting (H2O 6 ligand) state of the enzyme. P450 BM3 heme been well characterised, and the LS ferric form of the heme domain was reported to have low-spin g-values at 2.42, 2.26 and 1.92 from X- band EPR studies (46), which are consistent with the WT heme domain data reported here. Our data show that there are profound solvent effects with both the WT and variant BM3 forms. Although <1% MeOH or drug in methanol was added to each enzyme at 200 µM protein concentration, shifts in LS g-values are seen even with the WT heme domain. EPR studies with the WT heme domain indicate that there is no apparent direct effect of the drugs on the heme, since the same EPR shifts are observed with the solvent alone, and these data are as expected from the thermofluor (DSF) analysis. The WT and A82F EPR data are interesting in that in substrate- and MeOH-free forms a proportion of high spin heme iron is detected, although there is limited evidence for high spin heme iron from UV-visible spectroscopy, where the spectrum is dominated by a low spin Soret feature. This high spin state is easier to explain for the A82F variant as energetically it is closer to the substrate bound form, and so at the cryogenic temperatures used for EPR it could more easily be locked in a HS confirmation. The high spin state in WT BM3

223

heme domain is more difficult to explain, though for both WT and A82F BM3, fatty acid affinity is high and residues on EPR tubes could be a factor. The HS heme iron signal is lost on addition of MeOH, which may reflect alterations in the solvation state of the active site, either reinforcing water-ligated form(s), or possibly resulting in the displacement of residual fatty acid from the active site, even at these low solvent concentrations. The A82F BM3 also shows three LS species in the native form (Figure 4.2), suggesting multiple conformational states and/or altered configurations of the 6th water ligand. For A82F there is again a change in the EPR spectrum on the addition of MeOH, leading to two LS states and no detectable HS form. This is a profound effect from the solvent which is mirrored in all of the drug- bound heme domain spectra. There is no high spin species formed on addition of any of the drugs to the BM3 proteins, as might be expected with non-natural substrates that bind weakly and potentially in multiple conformations. EPR also shows that the heme environment is perturbed in the DIA-bound A82F spectrum with four low spin forms, three of which are new. Again, this likely indicates that DIA has multiple binding modes in the heme environment.

Similarly, amodiaquine showed new EPR binding modes not seen in the MeOH spectra. The AMO data allow us to see binding traits not observable by UV-visible spectroscopy due to the highly colored nature of the substrate, and provide an indication that AMO may associate directly with the heme iron, as well as th influencing the environment of the H2O 6 ligand in a distinct binding mode. The

F87V variant, which showed very little change in Tm compared to WT BM3, also provided limited alterations to EPR on addition of drugs. The only major change was with the DIA substrate, which gave small changes in the relative populations of the existing three sets of g-values. This different type of EPR response likely occurs since F87V does not structurally destabilise BM3, and since its increased activity with non-natural substrates is most likely explained by an enlarged size of the active site cavity in proximity to the heme. The DM enzyme also showed four LS species (as with the A82F point mutant heme domain), though no HS component was observed. A similar pattern of spectral change from four to three low spin species was observed on addition of MeOH. Again, DIA had the most profound effect on the DM heme, indicating that this molecule likely has multiple binding modes with all the variant enzymes, although this isn’t translated into high levels of product generation (possibly since many of the modes are not productive). DCF was the other major binder to the DM BM3, decreasing the number of observed LS species from four to two. With a single oxidised product produced for DCF predominantly with the DM, it

224

is likely that DCF makes close approaches to the DM heme in a single productive binding mode, even if this approach is not close enough to induce formation of HS heme iron. It is also of interest to note that, in terms of total product formed, DIA is the least efficiently transformed drug, and yet it is the most EPR active compound among those tested. This is likely explained due to its exhibiting multiple non- productive binding modes, which therefore limit product formation. Collectively, these EPR data complement those from the thermofluor screening studies, and indicate that tight binding is not seen with any of the drugs used. Most likely, competing binding effects (i.e. in different active site regions) mask any stabilising effects from these drugs. The presence of multiple non-productive weak binding modes compared one or few productive binding modes, likely leads to multiple EPR signatures and explains why the net effect is a lowering of the P450 Tm value (47). Kinetic analysis of drug substrate–dependent NADPH oxidation was achieved through steady state analysis of the variant enzymes with the various drugs. Ironically, the two best substrates (NIF and AMO) were highly colored and overlapped spectrally to interfere with changes associated with NADPH oxidation. This prevented meaningful determination of kinetic constants for these substrates. Of the substrates that were possible to analyse, DCF was shown to be the most

effective at stimulating NADPH oxidation, with the rank order of kcat values being DM (640 min-1), A82F (430 min-1) and F87V (190 min-1). While these values are lower than those for WT BM3 with fatty acids, the observed rate constants are much greater than those for the human P450 enzyme catalysts, largely as a consequence of the much lower efficiency of electron transfer from the membranous CPR redox partner in the human enzymes (48,49). Thus, the rate constants for the variants are relatively high given that the drug substrates used are non-natural, and do not show HS binding modes (that would result in increased heme iron reduction potential and faster rates of electron transfer to the heme iron). Moreover, the mutations that give rise to these novel drug oxidation activities are from rational design, as opposed to selecting for desired activities through a directed evolution approach. The same rank order of rate constants also follows for the 5-OH diclofenac product formation, with the DM and A82F BM3 enzymes giving the highest, and F87V the lowest rate

constants. Even though the DSF data indicate unchanged TM values on P450 variant binding this drug, there is a significant change in the EPR signals for the DM, which may indicate one or more productive binding modes. DIA also promotes BM3- dependent NADPH oxidation, with the highest rate constant seen with the F87V -1 BM3 variant at 370 min . This is linked to a Km value of 330 µM, which fits well with the DSF and EPR data that both indicate multiple binding modes and possibly a

225

preference for non-functional/productive modes over the functional one(s) (50). The -1 A82F BM3 variant also showed activity with DIA (kcat = 240 min ; Km = 110 µM). Since turnover studies indicated that this was the second most productive substrate, the lower Km here may likely point to a more specific and productive binding mode. The A82F heme domain bound to DIA also showd a high degree of destabilisation

in the thermofluor assay, with a 3 ºC reduction in the Tm. However, the DM BM3 didn’t show significant stimulation of NADPH oxidation with DIA, consistent with this being the least metabolised of the drugs tested. The DEX drug only stimulates NADPH oxidation with the DM BM3 enzyme. This finding concurs with the turnover data, as this variant gave the highest product conversion rate at 10 % in the assay -1 period. The relatively low kcat (190 min ) and high Km (98 µM) for DM DEX oxidation are indicative that this is not one of the best substrates for the BM3 variants. However, an interesting aspect of the steady-state kinetic behaviour in this case is that there is an apparent sigmoidal dependence of NADPH oxidation rate on DEX concentration. This suggests that there may be two (or more) cooperative binding sites for DEX. Similar BM3 behaviour has been observed previously, e.g. in the case of testosterone metabolising variants of P450 BM3 (that also contain the F87V mutation) (51).

In general, the turnover data collected are consistent with the other thermodynamic, kinetic and spectroscopic data presented. Thermal unfolding (DSF) assays suggest that none of the drugs are stabilizing for the P450s, which might indicate multiple binding modes, some of which are non-productive and probably destabilizing. EPR showed no obvious HS heme iron formation for any of the drugs, suggesting that binding did not occur in modes that led to displacement of the heme 6th water ligand, consistent with UV-visible absorbance titration data at ambient temperature. The rate constants for BM3 variant drug-dependent NADPH oxidation were generally fast for P450 enzymes (rates in the 100’s min-1), albeit much slower than those for WT BM3 with its typical fatty acid substrates. The relatively low amounts of products (in some cases) may be explained by drugs having both productive and non-productive binding modes, with the latter perhaps giving rise to NADPH oxidation that is not coupled to drug oxidation. The NIF data were unexpected, with a high level of product formed from an unusual type of P450 reaction. There was also some NIF product formation observed with the WT BM3 enzyme, suggesting that this substrate can also bind productively to the WT BM3, while the A82F and/or F87V mutations are required for activity with the other drugs.

226

These data clearly show that the A82F and F87V point mutations (and their combination in the DM) have a profound effect on the substrate selectivity of the P450 BM3 enzyme, converting it (certainly in the case of A82F-containing variants) to a more unstable form that can more readily accept non-natural substrates, and generate human-like metabolites from a range of drugs. This work helps to build a model to explain how single mutations, which may be distant from the active site, can have far reaching effects on substrate selectivity. The thermodynamic destabilisation of P450 BM3 (or any enzyme) may provide the explanation in many cases, as the energy barrier between substrate free and substrate bound conformations is reduced, enhancing the binding of substrates and enabling their turnover. In the case of BM3, the “gatekeeper” mutations were shown to facilitate binding and oxidation of omeprazole and other PPIs, and here are shown to gain affinity for a range of human drugs. As is often the case with error prone or direct evolution techniques, the impact of a single mutation can be lost in the sum of several mutations, making it hard to specify the cause of the altered activity (52). Using high throughput techniques (such as the 96 well thermal fluorescence assay we have employed) could allow us to start defining these individual effects and to develop improved methods of rational design of enzymes, in combination with structural and spectroscopic techniques. Thus, we would use this technique to identify key destabilising mutations which allow substrate binding flexibility (such as A82F) and then attempt to introduce structurally stabilising mutations, and mutations that increase coupling (e.g. A264G, A82W) (53,54) and reaction rates with specific substrates (e.g. the M5 multiple mutant, L75T/L181K) (55,56) in order that we can, in effect, streamline the process of making oxidase enzymes with enhanced activities with desired substrates. This would involve developing a BM3 “toolkit” of known mutations to be selected for particular purposes, thus reducing the reliance on screening against random mutants in the hope of finding a match for a particular substrate (32). Current enzyme “toolkits” rely on panels of random mutant screens to be tested against selected compounds for the desired product outcome, with little rationale to the design (57).

At the outset of these studies, we could not predict the generation of a human P450 metabolite in each case from studies with the BM3 variants. However, this likely says more about the human P450s and the chemical nature of the substrates than it does about the BM3 variants themselves. It is likely the case that the reactions catalysed by these P450s are those that offer the lowest energy barrier and are therefore preferred, as seen previously with BM3-catalysed pinene oxidation

227

following the route of lowest CH bond energy (58). This is perhaps predictable in the thermodynamically destabilized BM3 gatekeeper variants, where the drug substrates are predicted to have multiple binding modes, enabling conformational selection for low barrier oxidations. As the human drug-metabolizing P450s are rarely very substrate specific, with each responsible for the metabolism of large numbers of xenobiotics, the binding modes and oxidation kinetics are likely regulated by energetics rather than by strict binding interactions. For example, the mammalian CYP2D6 enzyme has been seen in a number of structural conformations, leading to the idea that conformational flexibility (as with our evolved enzymes) is key to its catalytic promiscuity (59). Much has been made of the notion that stability is the key to evolvability (1). However, if we look at true evolution, it is instability that provides the driving force for this process (2). The data presented in this chapter provides clear evidence that P450 BM3 destabilization provides the basis for novel drug substrate selection. These and other data are starting to give a strong indication that in order to derive new functions, enzymes must first pass through an instability “barrier” that allows them the conformational flexibility to make new molecular interactions.

228

4.7 References

1. Bloom, J. D., Labthavikul, S. T., Otey, C. R., and Arnold, F. H. (2006) Protein stability promotes evolvability. Proc. Natl. Acad. Sci. U.S.A 103, 5869-5874 2. Bridgham, J. T., Ortlund, E. A., and Thornton, J. W. (2009) An epistatic ratchet constrains the direction of glucocorticoid receptor evolution. Nature 461, 515-519 3. Ortiz de Montellano, P. R. (2005) Cytochrome P450: Structure, Mechanism, and Biochemistry, 3 ed., Kluwer Academic/Plenum Publishers, New York 4. Cheng, Q., Sohl, C. D., Yoshimoto, F. K., and Guengerich, F. P. (2012) Oxidation of Dihydrotestosterone by Human Cytochrome P450 19A1 and 3A4. J. Biol. Chem. 287, 29554-29567 5. Wang, K., and Guengerich, F. P. (2013) Bioactivation of fluorinated 2-aryl- benzothiazole antitumor molecules by human cytochrome P450s 1A1 and 2W1 and deactivation by cytochrome P450 2S1. Chemical Research in Toxicology 6. Tian, Z., Cheng, Q., Yoshimoto, F. K., Lei, L., Lamb, D. C., and Guengerich, F. P. (2013) Cytochrome P450 107U1 is required for sporulation and antibiotic production in Streptomyces coelicolor. Arch. Biochem. Biophys. 530, 101-107 7. Guengerich, F. P., and Munro, A. W. (2013) Unusual cytochrome p450 enzymes and reactions. Journal of Biological Chemistry 8. Coelho, P. S., Brustad, E. M., Kannan, A., and Arnold, F. H. (2013) Olefin cyclopropanation via carbene transfer catalyzed by engineered cytochrome P450 enzymes. Science 339, 307-310 9. Venkataraman, H., de Beer, S. B. A., van Bergen, L. A. H., van Essen, N., Geerke, D. P., Vermeulen, N. P. E., and Commandeur, J. N. M. (2012) A single active site mutation inverts stereoselectivity of 16-hydroxylation of testosterone catalyzed by engineered cytochrome P450 BM3. ChemBioChem 13, 520-523 10. Di Nardo, G., and Gilardi, G. (2012) Optimization of the Bacterial Cytochrome P450 BM3 System for the Production of Human Drug Metabolites. International Journal of Molecular Sciences 13, 15901-15924 11. Waltham, T. N., Girvan, H. M., Butler, C. F., Rigby, S. R., Dunford, A. J., Holt, R. A., and Munro, A. W. (2011) Analysis of the oxidation of short chain alkynes by flavocytochrome P450 BM3. Metallomics 3, 369-378

229

12. Shehzad, A., Panneerselvam, S., Linow, M., Bocola, M., Roccatano, D., Mueller-Dieckmann, J., Wilmanns, M., and Schwaneberg, U. (2013) P450 BM3 crystal structures reveal the role of the charged surface residue Lys/Arg184 in inversion of enantioselective styrene epoxidation. Chem. Commun. (Camb) 49, 4694-4696 13. Butler, C. F., Peet, C., Mason, A. E., Voice, M. W., Leys, D., and Munro, A. W. (2013) Key mutations alter the cytochrome P450 BM3 conformational landscape and remove inherent substrate bias. J. Biol. Chem. (in press) 14. Dickinson, R. G., and Jacobsen, N. W. (1970) A new sensitive and specific test for the detection of aldehydes: formation of 6-mercapto-3-substituted-s- triazolo[4,3-b]-s-tetrazines. J. Chem. Soc. D 0, 1719-1720 15. Lee, C.H., and Frasch, C. E. (2001) Quantification of bacterial polysaccharides by the purpald assay: measurement of periodate-generated formaldehyde from glycol in the repeating unit. Anal. Biochem. 296, 73-82 16. Tee, K. L., Dmytrenko, O., Otto, K., Schmid, A., and Schwaneberg, U. (2008) Transversion-enriched sequence saturation mutagenesis (SeSaM- Tv+): a random mutagenesis method with consecutive nucleotide exchanges that complements the bias of error-prone PCR. J. Mol. Catal. B: Enzym. 50, 121-127 17. Alcalde, M., Farinas, E. T., and Arnold, F. H. (2004) Colorimetric high- throughput assay for alkene epoxidation catalyzed by cytochrome P450 BM- 3 variant 139-3. J. Biomol.Screen. 9, 141-146 18. Wong, T. S., Wu, N., Roccatano, D., Zacharias, M., and Schwaneberg, U. (2005) Sensitive assay for laboratory evolution of hydroxylases toward aromatic and heterocyclic compounds. J. Biomol.Screen. 10, 246-252 19. Gotoh, O. (1992) Substrate recognition sites in cytochrome P450 family 2 (CYP2) proteins inferred from comparative analyses of amino acid and coding nucleotide sequences. J. Biol. Chem. 267, 83-90 20. Lin, Y., Lu, P., Tang, C., Mei, Q., Sandig, G., Rodrigues, A. D., Rushmore, T. H., and Shou, M. (2001) Substrate inhibition kinetics for cytochrome P450- catalyzed reactions. Drug Metab. Dispos. 29, 368-374 21. Romero, P. A., Krause, A., and Arnold, F. H. (2013) Navigating the protein fitness landscape with Gaussian processes. Proc. Natl. Acad. Sci. 110, E193–E201 22. Kumar, S., Zhao, Y., Sun, L., Negi, S. S., Halpert, J. R., and Muralidhara, B. K. (2007) Rational engineering of human cytochrome P450 2B6 for

230

enhanced expression and stability: importance of a Leu264->Phe substitution. Mol. Pharmacol. 72, 1191-1199 23. Arendse, L., Blundell, T. L., and Blackburn, J. (2013) Combining in silico protein stability calculations with structure-function relationships to explore the effect of polymorphic variation on cytochrome P450 drug metabolism. Curr. Drug. Metab. 14, 745-63. 24. Gould, S. (1970) Dollo on Dollo's law: irreversibility and the status of evolutionary laws. J. Hist. Biol. 3, 189-212 25. Baillie, T. A., Cayen, M. N., Fouda, H., Gerson, R. J., Green, J. D., Grossman, S. J., Klunk, L. J., LeBlanc, B., Perkins, D. G., and Shipley, L. A. (2002) Drug metabolites in safety testing. Toxicol. Appl. Pharmacol. 182, 188-196 26. Robison, T. W., and Jacobs, A. (2009) Metabolites in safety testing. Bioanalysis 1, 1193-1200 27. Munro, A. W., Girvan, H. M., and McLean, K. J. (2007) Variations on a (t)heme--novel mechanisms, redox partners and catalytic functions in the cytochrome P450 superfamily. Nat. Prod. Rep. 24, 585-609 28. Ravichandran, K., Boddupalli, S., Hasermann, C., Peterson, J., and Deisenhofer, J. (1993) Crystal structure of hemoprotein domain of P450BM- 3, a prototype for microsomal P450's. Science 261, 731-736 29. Joyce, M. G., Ekanem, I. S., Roitel, O., Dunford, A. J., Neeli, R., Girvan, H. M., Baker, G. J., Curtis, R. A., Munro, A. W., and Leys, D. (2012) The crystal structure of the FAD/NADPH-binding domain of flavocytochrome P450 BM3. FEBS J. 279, 1694-1706 30. Li, H., and Poulos, T. L. (1997) The structure of the cytochrome p450BM-3 haem domain complexed with the fatty acid substrate, palmitoleic acid. Nat. Struct. Biol. 4, 140-146 31. Sevrioukova, I. F., Li, H., Zhang, H., Peterson, J. A., and Poulos, T. L. (1999) Structure of a cytochrome P450-redox partner electron-transfer complex. Proc. Natl. Acad. Sci. 96, 1863-1868 32. Sideri, A., Goyal, A., Di Nardo, G., Tsotsou, G. E., and Gilardi, G. (2013) Hydroxylation of non-substituted polycyclic aromatic hydrocarbons by cytochrome P450 BM3 engineered by directed evolution. J. Inorg. Biochem. 120, 1-7 33. Tran, N.-H., Huynh, N., Chavez, G., Nguyen, A., Dwaraknath, S., Nguyen, T.-A., Nguyen, M., and Cheruzel, L. (2012) A series of hybrid P450 BM3

231

enzymes with different catalytic activity in the light-initiated hydroxylation of lauric acid. J. Inorg. Biochem. 115, 50-56 34. Chen, M. M. Y., Snow, C. D., Vizcarra, C. L., Mayo, S. L., and Arnold, F. H. (2012) Comparison of random mutagenesis and semi-rational designed libraries for improved cytochrome P450 BM3-catalyzed hydroxylation of small alkanes. Protein Eng. Des. Sel. 25, 171-178 35. Caswell, J. M., O’Neill, M., Taylor, S. J. C., and Moody, T. S. (2013) Engineering and application of P450 monooxygenases in pharmaceutical and metabolite synthesis. Curr. Opin. Chem. Biol. 17, 271-275 36. Niesen, F. H., Berglund, H., and Vedadi, M. (2007) The use of differential scanning fluorimetry to detect ligand interactions that promote protein stability. Nat. Protocols 2, 2212-2221 37. Noble, M. A., Miles, C. S., Chapman, S. K., Lysek, D. A., MacKay, A. C., Reid, G. A., Hanzlik, R. P., and Munro, A. W. (1999) Roles of key active-site residues in flavocytochrome P450 BM3. Biochem. J. 339, 371-379 38. Omura, T., and Sato, R. (1964) The carbon monoxide-binding pigment of liver microsomes. II. solubilization, purification, and properties. J. Biol. Chem. 239, 2379-2385 39. Perera, R., Sono, M., Sigman, J. A., Pfister, T. D., Lu, Y., and Dawson, J. H. (2003) Neutral thiol as a proximal ligand to ferrous heme iron: implications for heme proteins that lose cysteine thiolate ligation on reduction. Proc. Natl. Acad. Sci. U.S.A 100, 3641-3646 40. Dunford, A. J., McLean, K. J., Sabri, M., Seward, H. E., Heyes, D. J., Scrutton, N. S., and Munro, A. W. (2007) Rapid P450 heme iron reduction by laser photoexcitation of Mycobacterium tuberculosis CYP121 and CYP51B1. Analysis of CO complexation reactions and reversibility of the P450/P420 equilibrium. J. Biol. Chem. 282, 24816-24824 41. Guengerich, F. P. (1990) Low kinetic hydrogen isotope effects in the dehydrogenation of 1,4-dihydro-2,6-dimethyl-4-(2-nitrophenyl)-3,5- pyridinedicarboxylic acid dimethyl ester (nifedipine) by cytochrome P-450 enzymes are consistent with an electron/proton/electron transfer mechanism. Chemical Research in Toxicology 3, 21-26 42. Tang, W. (2003) The metabolism of diclofenac--enzymology and toxicology perspectives. Curr. Drug. Metab. 4, 319-329 43. Kuper, J., Tee, K. L., Wilmanns, M., Roccatano, D., Schwaneberg, U., and Wong, T. S. (2012) The role of active-site Phe87 in modulating the organic

232

co-solvent tolerance of cytochrome P450 BM3 monooxygenase. Acta Crystallogr. F-Struct. Biol. Cryst. Commun. 68, 1013-1017 44. Epand, R. F., Epand, R. M., and Jung, C. Y. (2001) Ligand-modulation of the stability of the glucose transporter GLUT 1. Protein Sci. 10, 1363-1369 45. Wong, T. S., Arnold, F. H., and Schwaneberg, U. (2004) Laboratory evolution of cytochrome p450 BM-3 monooxygenase for organic cosolvents. Biotechnol. Bioeng. 85, 351-358 46. Miles, J. S., Munro, A. W., Rospendowski, B. N., Smith, W. E., McKnight, J., and Thomson, A. J. (1992) Domains of the catalytically self-sufficient cytochrome P-450 BM-3. Genetic construction, overexpression, purification and spectroscopic characterization. Biochem. J. 288 ( Pt 2), 503-509 47. Holdgate, G. A., and Ward, W. H. J. (2005) Measurements of binding thermodynamics in drug discovery. Drug Discovery Today 10, 1543-1550 48. Ost, T. W. B., Clark, J., Mowat, C. G., Miles, C. S., Walkinshaw, M. D., Reid, G. A., Chapman, S. K., and Daff, S. (2003) Oxygen activation and electron transfer in flavocytochrome P450 BM3. J. Am. Chem. Soc. 125, 15010- 15020 49. Whitehouse, C. J. C., Bell, S. G., and Wong, L.-L. (2012) P450(BM3) (CYP102A1): connecting the dots. Chem. Soc. Rev. 41, 1218-1260 50. Waldron, T. T., and Murphy, K. P. (2003) Stabilization of proteins by ligand binding: application to drug screening and determination of unfolding energetics. Biochemistry 42, 5058-5064 51. van Vugt-Lussenburg, B. M. A., Damsten, M. C., Maasdijk, D. M., Vermeulen, N. P. E., and Commandeur, J. N. M. (2006) Heterotropic and homotropic cooperativity by a drug-metabolising mutant of cytochrome P450 BM3. Biochem. Biophys. Res. Commun. 346, 810-818 52. Chen, R. (2001) Enzyme engineering: rational redesign versus directed evolution. Trends Biotechnol. 19, 13-14 53. Carmichael, A. B., and Wong, L.-L. (2001) Protein engineering of Bacillus megaterium CYP102. The oxidation of polycyclic aromatic hydrocarbons. Eur. J. Biochem. 268, 3117-3125 54. Haines, D. C., Hegde, A., Chen, B., Zhao, W., Bondlela, M., Humphreys, J. M., Mullin, D. A., Tomchick, D. R., Machius, M., and Peterson, J. A. (2011) A single active-site mutation of P450BM-3 dramatically enhances substrate binding and rate of product formation. Biochemistry 50, 8333-8341 55. Nazor, J., Dannenmann, S., Adjei, R. O., Fordjour, Y. B., Ghampson, I. T., Blanusa, M., Roccatano, D., and Schwaneberg, U. (2008) Laboratory

233

evolution of P450 BM3 for mediated electron transfer yielding an activity- improved and reductase-independent variant. Protein Eng. Des. Sel. 21, 29- 35 56. Ost, T. W. B., Miles, C. S., Murdoch, J., Cheung, Y.-F., Reid, G. A., Chapman, S. K., and Munro, A. W. (2000) Rational re-design of the substrate binding site of flavocytochrome P450 BM3. FEBS Lett. 486, 173- 177 57. Sawayama, A. M., Chen, M. M. Y., Kulanthaivel, P., Kuo, M.-S., Hemmerle, H., and Arnold, F. H. (2009) A panel of cytochrome P450 BM3 variants to produce drug metabolites and diversify lead compounds. Chem. Eur. J. 15, 11723-11729 58. Branco, R. J. F., Seifert, A., Budde, M., Urlacher, V. B., Ramos, M. J., and Pleiss, J. (2008) Anchoring effects in a wide binding pocket: the molecular basis of regioselectivity in engineered cytochrome P450 monooxygenase from B. megaterium. Proteins: Struct., Funct., Bioinf. 73, 597-607 59. Ekroos, M., and Sjögren, T. (2006) Structural basis for ligand promiscuity in cytochrome P450 3A4. Proc. Natl. Acad. Sci. 103, 13682-13687

234

235

5.0 Summary, conclusions and further work

5.1 Summary

Discovering new synthetic methods for the production of high value chemicals such as pharmaceuticals and their metabolites is a challenge for today’s chemists. The huge growth in the biotechnology industry is driven by the hope that we can make products using less energy, increased selectivity and with less waste than traditional methods. Oxygenase enzymes are ideal targets for this type of catalysis, due to the difficulty of activating CH bonds chemically, leading to multi step synthesis and generation of unwanted by-products. Pharmaceutical companies and regulatory agencies are increasingly aware of the need to test drug metabolites for numerous reasons. They may be as straightforward as direct toxic effects of the metabolites, as is the case for acetaminophen (1), or more subtle, such as fluconazole increasing the circulating concentration of warfarin through its interaction with CYP2C19, and thus retarding metabolism of the anticoagulant (2). The demand for these metabolite compounds means that they command a high price, much more so than the compounds they are derived from. This provides the basis for the market to access enzymatic synthesis of metabolites, which can be more cost competitive in providing small quantities of higher value chemicals (3). Human production of xenobiotic metabolites is primarily performed in the liver by the microsomal P450 enzymes. This class of enzymes is known to accept a wide range of substrates, and yet can still catalyse their oxidation with a high degree of stereo and regioselectivity. Current enzymatic synthesis methods for drug metabolites relies on recombinant human P450s, but these enzymes are difficult to handle; require extensive gene engineering to facilitate their utilization and the purification of the redox partner CPR; give low total turnover numbers; and are expensive to produce (4). The focus of much research has been to find soluble bacterial enzymes that will produce these metabolites more cheaply, and which are easier and more efficient to use than the human P450s, so that greater metabolite yields can be achieved (5). Much of this focus has fallen on the fatty acid hydroxylase P450 BM3 from Bacillus megaterium, due to its soluble, self-sufficient nature - comprising both P450 and redox partner CPR on the same polypeptide chain. The fusion of the P450 enzyme and redox partner gives a high reaction rate, (up to 17, 000 min-1 with arachidonic acid) which, along with BM3’s solubility and high expression levels in E. coli, make this enzyme an ideal biotechnology candidate (6). Much research has been focused

236

on the mutation of this enzyme for altered function, most of which has been done by directed evolution using random error-prone PCR (7,8). Our approach was more rational and involved finding the key mutations which most often gave altered activity to BM3 in directed evolution studies, and to try to rationalise the reasons by which such mutations result in modified activity.

We selected soluble cytochrome P450 BM3 for these studies in view of the body of preceding work providing evidence for its amenability to engineering, and since it has the highest oxidase activity of any known P450 (9). The BM3 gene was cloned into various pET expression vectors and silent restriction enzyme cloning sites were introduced into the heme domain to allow for multiple mutations to be combined in a single gene. Expression and purification of the proteins proved relatively simple requiring Ni-NTA and gel filtration steps for isolation of the full length BM3 enzymes, and Q-Sepharose, hydroxyapatite and gel filtration steps for the non-tagged heme domain proteins. Initially we screened several single mutations which were common in activity-altering BM3 variants reported in the literature, and in order to discover what effects they had individually, and initially with the expectation that we would need to combine these individual mutations with others from focused library mutant libraries to have any great effect on switching substrate selectivity. We found that, two individual mutations had separate and complementary effects: A82F on the binding of unnatural substrates, and F87V on activity with the unnatural substrates. Though mutations R47L and L188Q, both common BM3 literature mutations (10), were also screened, they were found to have no effect against the group of pharmaceuticals we trialled (unpublished data). As P450s have a major UV-Visible spectral absorbance feature (the Soret band) which shifts from ~418 nm to ~390 nm on substrate binding (due to a low-spin to high-spin shift in the P450 heme iron), this method was used to screen an initial batch of compounds which were marker substrates used by the industrial partner Cypex to assess the activity of the human P450s that they produce. A major hit came with omeprazole, a proton pump inhibitor (PPI) drug used in the treatment of stomach ulcers and gastric reflux (11). This drug exhibited very tight binding for a non-natural BM3 substrate (Kd in the µM range), showing us it had a strong interactions with A82F-containing variants of BM3. We undertook steady-state kinetic studies involving drug substrate-dependent NADPH oxidation in order to determine if this binding also translated into stimulation of electron transfer from the CPR partner. This proved to be the case, as high catalytic rates were obtained with both the A82F- and F87V-containing variants. Further study of the products of this BM3-dependent catalysis by LCMS and NMR

237

determined that the reaction catalysed was a 5-hydroxylation of omeprazole, which is the same reaction catalysed by the human P450 CYP2C19, for which omeprazole is a marker substrate. The DM containing both the A82F and F87V mutations produced a side product generated by further oxidation of the 5-hydroxy OMP to the 5-carboxylated product. This “over-oxidation” is likely due to the tight binding of both omeprazole and the 5-hydroxy OMP to DM BM3 allowing further oxidation to take place before the product is released. Using the WT BM3 crystal structure (12) it was easy to rationalise why F87V containing variants were able to oxidise omeprazole, since the removal of the phenylalanine residue from directly above the heme provides space for the larger, non-natural substrate to bind and be oxidised. The effects of the A82F mutation, which is outside the active site pocket with no conceivable direct interaction with the substrate, were more difficult to reconcile. We crystallised the heme domain of BM3 in complex with omeprazole to gain insight into the novel OMP substrate binding mode in the variants, and to resolve why the A82F mutation is key to altered binding. Initially we generated the heme domains by inserting two stop codons into the full length BM3 gene and expressed them with an N terminal His tag. This was not highly successful initially and resulted in many problems. The first was the propensity of the A82F-containing variants to co-purify with long chain fatty acids from E. coli (predominantly palmitic acid) which couldn’t be removed by standard methods (13). Extensive trials were carried out to find a solution to this problem, and we found that Lipidex resin (PerkinElmer, Massachusetts, USA) was an effective way of removing bound lipids from the proteins. The second problem was the very tight binding of imidazole to the variants, leading to inhibitor-bound protein from which imidazole couldn’t readily be removed by dialysis. In efforts to resolve this problem, we replaced imidazole with L-histidine for the elution of protein from Ni-NTA resin, but found that the protein didn’t crystallise following this treatment. We considered that the lack of crystals might likely be due to the inability of the protein to crystallise with a N-terminal His tag, and to solve this issue we re-cloned the gene into vector pET20b, which encodes a C- terminal tag, and inserted stop codons at the end of the heme domain gene in order to prevent His-tag expression in these constructs. Finally, substrate-free crystals were obtained, although substrate-bound P450 protein wells contained black spots, which were the results of omeprazole breakdown products. Analysis of omeprazole stability in buffered solution showed us that it breaks down rapidly overnight, likely due to the labile sulfone group (unpublished data) as is seen in aqueous solution (14). Alongside these studies, we analysed the stability of omeprazole when fully bound to BM3 enzymes, and noted that there was no omeprazole breakdown and

238

that the complex was stable at room temperature for days. Therefore, for crystallography studies, an altered method for substrate addition was designed, whereby a large scale titration of omeprazole (in MeOH) into dilute P450 BM3 heme domain solution (30 ml) was done until it was fully bound to the drug. The relevant protein was then concentrated by ultrafiltration in a Vivaspin device to remove any excess omeprazole. This proved successful and OMP-bound crystals were obtained for both A82F and DM heme domains at resolutions of 1.9 and 2.1 Å, respectively. Both the substrate -free and substrate-bound crystals required micro seeding to produce diffraction quality crystals. The seed crystals for the DM BM3 heme domain were taken from crystallization conditions containing imidazole, and so this molecule can also be seen in the DM heme domain substrate-free structure. The OMP-bound crystal structures were almost identical in conformation to previous fatty acid-bound BM3 heme domain structures, and with just a small change in the omeprazole orientation to a less strained conformation induced by removal of the Phe87 side chain in the F87A/A82F DM heme domain structure. The less strained structure explains why the DM enzyme exhibits tighter binding of omeprazole compared to the A82F point mutant, and the 4 Å distance from the heme iron to the omeprazole 5-methyl group is consistent with our product analysis data, that showed that this was the position of hydroxylation/carboxylation of omeprazole. The substrate-free data for the A82F-containing variants revealed novel conformations for BM3, with a reorientation of the F and G helices caused by disruption in the I-helix hydrogen bonding. This led us to propose that the stability of the substrate-free structure is compromised, making the substrate-bound conformation preferential. Looking further into the causes of this phenomenon, we undertook differential scanning calorimetry (DSC) analysis of the proteins in the ligand-free state, with the tight binding fatty acid derivative NPG substrate, and with the novel omeprazole substrate. We discovered that the A82F mutation (which is key for novel substrate binding) causes a dramatic reduction in stability for the ligand-free form as compared to the WT enzyme. This could be expected, as mutations have been shown to both stabilise and destabilise various proteins. The more interesting finding comes with the substrate-bound P450 data, which showed that the WT BM3 heme domain is greatly stabilised when bound to NPG substrate, but that omeprazole binding has negligible effect. The impact on P450 thermostability is greatly changed upon introduction of the A82F mutation, either in isolation or in combination with F87V. Rather than providing an increase in stability with omeprazole, there is a loss of stability with this substrate, which has the effect of altering the SF to SB conformational equilibrium, allowing a lower energy transition

239

to the SB state. It has been shown that P450 BM3 requires a shift to the SB conformation to aid in catalysis, accompanied by changes such as the displacement of the heme iron water ligand and the rearranging of the I-helix for efficient proton delivery (15). This hypothesis is backed up by structural data that were collected, which shows that the SF structures for the “gatekeeper” variants are different from any seen before, with increased flexibility in the FG helices and altered hydrophobic contacts between these helices and the I-helix, bringing Ile263 and Phe 82 into close contact.

This work was further expanded to include studies of the P450 BM3 gatekeeper variants with other members of the PPI drug class, namely esomeprazole, lansoprazole, pantoprazole and rabeprazole. As esomeprazole is the active S- isomer of omeprazole, we had expected its binding to the BM3 variants to be the same as that for the racemate omeprazole. However, though esomeprazole followed a similar binding pattern with improved affinity for the DM over the A82F variant, its apparent Kd for esomeprazole was approximately an order of magnitude weaker than that for omeprazole, indicating that the R-isomer is tighter binding. Lansoprazole, pantoprazole and rabeprazole both bound to the A82F and DM BM3 enzymes, albeit with lower affinity than that for the other PPI substrates. These drugs also showed a preference for binding to the DM, indicating that the added space created around the heme by the F87V mutation allows these compounds to enter further into the active site cavity and to displace the heme axial water ligand more efficiently. We took advantage of the strong BM3 EPR signal in order to look for additional information on binding interactions with the PPI drugs using X-band EPR spectroscopy. The findings provided further interesting results. We found that the ostensibly substrate-free (lipidex-treated) A82F enzyme still retained a proportion of high spin heme iron. However, when solvent was added (either DMSO or MeOH) this high spin EPR species is lost. Similarly, addition of any substrate added in solvent also removed the high spin component. This could mean that some lipid remains bound to the enzyme after lipidex treatment, and that the small energy barrier between the HS and LS forms in the A82F variant means that a HS species can still be formed. However, when solvent is added the lipid may have preference for the solvent-containing buffer solution over the protein active site, and therefore become LS. Alternatively, this phenomenon may also be caused by a change in the solvation state around the heme, with the HS form being a consequence of the altered structure leading to the displacement of the axial water ligand, and addition of solvent recovering the water as the 6th ligand. Though the

240

PPIs displace the water in UV-visible absorbance binding experiments, it is noted from EPR that at low temperatures P450 BM3 reverts to the LS state, even in the substrate-bound state (16). Several features were observed in the EPR spectra that were indicative of drug substrate binding, and of particular interest were those with the F87V variant, which didn’t give evidence of drug association from optical binding assays. These data revealed that multiple ligand protein interactions occurred in the low spin spectra. Even the WT BM3 heme domain spectra showed a new LS EPR species with both lansoprazole and rabeprazole. This may help to explain why lansoprazole stimulates NADPH oxidation in WT BM3 and why rabeprazole is oxidized by WT BM3, yet neither show evidence of binding from optical totration studies. Once again, kinetic data from PPI-dependent NADPH oxidase studies showed fast reaction rates for the four PPIs, with most NADPH oxidation rates being at more than a thousand turnovers per minute. The products from BM3 enzyme- dependent metabolism of the PPIs were analysed by LCMS and NMR, which provided some further interesting results. Again it was found that esomeprazole metabolism differed from that of racemic omeprazole, and we found that this isomer was much more active than the racemate, with far more 5-hydroxy esomeprazole product generation. This allowed us to conclude that there was a different reaction pathway for each of the omeprazole isomers, as is also seen with the human metabolism of the omeprazole isomers (17). We also no longer observed the carboxylated product with esomeprazole, leading us to conclude that it is the R- isomer which forms this product, due to its very tight binding nature and its allowing iterative oxidations to take place before product release. Metabolite characterisation was possible by both LCMS and NMR as high amounts of product were generated. With lansoprazole, the sulfone is generated by direct oxidation of the central sulfur, a reaction seen with other P450s and the same reaction catalysed by human CYP3A4 (18,19). Rabeprazole metabolism is more complicated. There is a non- enzymatic pathway which is also seen in human metabolism to form the thioether by loss of the central oxygen (20). Both rabeprazole and the thioether were metabolised to the desmethyl derivative, as seen in human CYP2C19 metabolism of these molecules (20). Only small amounts of the desmethyl rabeprazole were generated, and the A82F BM3 variant favoured the desmethyl thioether, with 50% of the product seen as this compound. A further oxidation pathway was also seen that was different from that seen with human enzymes. This pathway involved the loss of the entire pyridinyl alkyl chain of the thioether. This reaction was most prevalent in the WT, F87V and DM BM3 enzymes, with over 70% of the product from the BM3 DM identified as this molecule. All of the metabolites generated were due to

241

reactions on the pyridinyl ring or the sulfone sulfur, leading us to infer that it is the benzimidazole interactions in the omeprazole structures that direct reactivity in this class of compounds. The sulfone generated with lansoprazole shows the diversity in reaction types that is feasible in these reactions, with a direct heteroatom oxidation such as that which is seen with other P450s, and also previously with T268A BM3. However it is more interesting that we observe oxidation on a sulfur already oxidized to a sulfone, rather than on a terminal sulfur as observed in the previous BM3 studies (21). In observing hydroxylation, dealkylation and sulfoxidation with three very structurally similar compounds, the data indicate that their binding modes in the BM3 active sites may change substantially, even with small structural differences in PPI structure. With pantoprazole, we found that determining the product formed was arduous as the NMR data were inconclusive due to the low product concentration and since the LCMS showed the product was also present to some extent in the starting material. It wasn’t until we were able to use Agilent’s metabolite identification library that we were able to screen against synthesis by-products. With this we saw the product again to result from heteroatom oxidation, this time forming the N-oxide. N-oxidation is one of the core P450 reactions (22), but in this case the interest is that instead of forming a human metabolite we instead produce a natural by-product of pantoprazole synthesis (23). We showed that, in many cases, we produced the human metabolite in these oxidative reactions with the BM3 variants. However, with rabeprazole we also generated significant quantities of a new metabolite and with pantoprazole we produced a synthesis by-product. The high similarity in reactivity with the human enzymes suggests that the novel substrate molecules have preferential sites of oxidation for either steric or energetic reasons, and that in easing substrate entry to BM3 through mutagenesis, these are also the preferential metabolism sites observed in the bacterial enzyme. The large amounts of products produced and the high regioselectivity in metabolite generation achieved with these drugs makes them possible targets for scaling up these reactions to whole bacterial cell cultures. Heme domain crystal structure data were essentially the same for the esomeprazole and the omeprazole complexes, and we concluded this was due to the lack of the PPI’s sulfone oxygen seen in the crystal structure, meaning that it had either broken down in solution, or lost as a consequence of irradiation by the X-ray beam. This meant that the compound was no longer chiral, explaining why the resulting derivatives from omeprazole and esomeprazole bind in exactly the same manner. The data showed that the addition of the F87V variant to A82F in the DM heme domain removed a close contact between the substrate and the phenylalanine residue, allowing for omeprazole binding in a less strained

242

conformation. The less strained binding likely explains the lower dissociation constant for the DM over the A82F point mutant. In each of the variant heme domain structures there were remarkably few direct interactions between the ligand and protein, with just a single hydrogen bond from a benzimidazole nitrogen to the carbonyl of leucine 437. The other benzimidazole nitrogen forms a bridging hydrogen bond through a water molecule to serine 72, and a final bridging hydrogen bond occurs from the backbone nitrogen of alanine 74 to the oxygen of the benzimidazole methoxy group. Most interactions are hydrophobic and can be expected for BM3’s large hydrophobic active site. Most interesting is the position of the 5-methyl group directly over the heme and just 3.9-4.1 Ångstroms from the heme iron. This is the first BM3 structure in which the substrate is bound in an active conformation, i.e. with the substrate close enough for a reaction to take place. All preceding fatty acid-bound structures show the substrates too far from the heme iron for direct oxidation. This finding indicates that these substrates are likely mobile in the active site, likely allowing for oxidation to take place at multiple positions on the chain (as is observed for fatty acids ω with ω-1, ω-2 and ω-3 oxidation predominating). The well-defined conformations observed for omeprazole and esomeprazole explains the high regioselectivity with which the oxidative reactions take place in the BM3 gatekeeper variants. Attempts were also made to crystallise lansoprazole- and rabeprazole-bound forms of the P450 BM3 variants in the same manner, but these compounds were found to be much less stable than omeprazole or esomeprazole, and they started to break down even during the process of concentrating the P450s in the presence of an excess of these PPI drugs. Some crystals were obtained, but the diffraction quality was too poor for meaningful structural determination, and thus further work in this area is needed in e.g. analysis of appropriate buffering conditions that might stabilize the PPI drugs, or perhaps the use of less hydrophobic and/or more stable analogues of these PPIs. Pantoprazole crystallisation was more promising, with good quality crystals produced that diffracted to 2.4 Å, but pantoprazole could not be fully resolved in the active site, likely due to there being multiple binding modes. This would be the most promising target for further crystallography, and using the known BM3 omeprazole structures we could model binding and target new residues for mutagenesis. Our work progressed further to look at different ways of screening for altered activity in these BM3 variants through use of a range of human P450 marker substrate compounds supplied by the industrial supervisor. Optical binding data proved inconclusive, as results showed only small absorbance changes and were not reproducible (data not shown) once protein was lipidex treated to remove any

243

endogenous lipid. It was concluded that BM3 P450 absorbance shifts observed in non-lipidex treated enzymes were likely due to either the compounds displacing bound lipid to push the protein more low spin, or instead bind in the presence of the lipid to enhance its binding and/or push it further into the active site to drive the P450 more high spin. Thus, the preliminary absorbance data collected indicated that the drug compounds used did penetrate the BM3 variant active sites. In subsequent work, we used a fluorescence stability assay (created to screen buffer and additive conditions for crystallography) which relies on sypro-orange dye that preferentially binds to unfolded protein. This was done specifically to provide an indication of the stability of our variant P450 enzymes and of the influence of drug binding on P450 stability (24). As we had previously carried out DSC analysis on the WT and variant BM3 enzymes, information was already available on their relative stability of the heme domains and the influence of substrate binding – which provided an important comparison with the thermofluor (differential scanning fluorimetry, DSF) assay results. The DSF assay results mirrored those from the DSC for all but the F87V variant, which appeared to be approximately as stable as the WT heme domain by DSF, rather than slightly less stable as found by DSC. The presence of a high level of DMSO in the DSF assay was a potential explanation for this difference, as Phe87 variants have been shown to ligate DMSO directly to the heme iron, and can be stabilised as a result of this ligand interaction (25). The DSF assay did not show any increase in WT or variant BM3 P450 thermostability for the drugs trialled. However, decreases in stability provided further evidence of interaction between the drugs and the proteins. Some of these drugs produced quite substantial decreases in P450 thermostability on addition of ligand. The loss of stability can potentially be explained by their having multiple binding modes to the protein, some of which are stabilising, and others of which are not. Another putative explanation is that the binding is tighter to one or more unfolded or less stable states of the protein, and destabilizes the P450s by driving the equilibrium in that direction (26). EPR analysis was also done for the P450 proteins both with and without drug ligands in order to see if any perturbations of the ferric heme iron signals were observed. As with the PPI EPR data, we observed significant solvent effects which are also mirrored in the drug samples, allowing us to discount these from actual ligand effects. By far the most interesting effect was that shown by diazepam which produced up to four different species in the LS EPR. We also noted a correlation between the decreased stability caused by drug binding in the DSF assay and an increase in the number of species in the EPR spectrum, adding weight to the theory that multiple binding modes may underpin the loss of stability. Steady-state kinetic analysis of drug-

244

dependent NADPH oxidation showed that turnover was stimulated by some of the compounds analysed, and was associated with kcat values in the order of hundreds per minute (rather than the thousands per minute observed with fatty acids). However, these are still rather fast rate constants as compared to those for human enzymes (27). Diazepam, diclofenac and dextromethorphan were each shown to stimulate BM3 NADPH oxidation kinetics. At this stage, oxidation of amodiaquine and nifedipine could not be ruled either, but gathering data by this method was not possible as these drugs are highly colored and obscure changes in the NADPH absorbance. Attempts were made to analyse their steady-state turnover by dual beam spectroscopy, but changes in the NADPH signal at 340 nm could not be reliably determined. Time-courses were subsequently done with each of the drugs and for the WT and variant BM3 enzymes, analysing any metabolites produced using LCMS. Product formation levels varied from 2-80% with the DM being the most effective metaboliser. Nifedipine was the highest metabolised substrate. This may be due to the reaction being highly favourable as an aromatisation of a ring. This oxidation reaction is also fast when catalysed by the prototypical human P450 CYP3A4 (28). The fact that all of these structurally diverse drugs, known to be metabolised by a range of human P450s, were metabolised by single point variants and the DM variant of BM3 is interesting in its own right, but when we compared products formed with known standards, it was found that most BM3 variant reactions produced the major human metabolite and all produced at least one of the human metabolites. All these data combined indicate that the structural flexibility introduced in the BM3 enzyme by these mutations creates a malleable enzyme that can accept multiple substrates and bind them in multiple conformations. The products formed and why they are so similar to the major human metabolites is likely due to a combination of factors. The first of these is that, in the absence of a tight and highly specific binding mode, the energetically most favourable reaction(s) will likely take place, and so the aromatisation of nifedipine or the de-ethylation of amodiaquine are favoured over other more energetically challenging reactions. Another factor is that the hydrophobic nature of the BM3 active site seems to direct each compound towards the heme by their more polar end. Finally, the majority of the compounds tested have nitrogens on or near the site of reaction. Coordination of heme iron by nitrogen ligands is well known in P450 inhibition, and it is possible that this also occurs transiently for some of the drugs tested, potentially allowing for productive binding modes and for oxidation of the drugs once inhibitory interactions are lost and the heme iron can be reduced and bind dioxygen to initiate the catalytic cycle.

245

5.2 Conclusions

Work presented in this thesis has added much to our knowledge about why certain mutations confer altered selectivity to P450s and (specifically for P450 BM3) why mutations are two sites (Ala82 and Phe87) are so prevalent in activity altering BM3 variants generated by random mutagenesis studies. The research has shed light on how little was understood about the reasons why two of the most highly mutated residues within BM3 confer altered activities. This research also adds important new data to the debate about stability as a factor in protein evolvability, and whether increased or decreased stability is the essential factor. We have also made strides towards a wider goal (in collaboration with Cypex) of finding bacterial P450s capable of generating human drug metabolites more cheaply and in higher quantities than their human counterparts. The structural rationale for BM3 binding to and activity with a non-natural drug substrate has been established for the first time. We have also published the first structure of the P450 BM3 heme domain in an active conformation with the heme iron and substrate in close enough proximity for a productive reaction to occur. We have also demonstrated that it is possible to identify variants that transform a particular marker drug substrate (such as omeprazole) and then extrapolate to demonstrate their effectiveness with a class of compounds (in this case the PPIs), which may be an effective method when using other BM3 variants to examine oxidation of different classes of compounds, or for specific oxidative reactions. We have shown that chirality in a compound of interest can affect its route of metabolism and the eventual product formed. Preliminary efforts have also been made to understand why a range of unrelated compounds are metabolised by these variants (i.e. drugs that are prototypical substrates for the major human drug-metabolizing P450s), and the ground has been prepared for further studies in this area. The preliminary data collected provide compelling evidence for the role of P450 protein destabilization in evolution towards a selection of novel substrates.

246

5.3 Further Work

Additional work I believe would be useful in the research area of this thesis would involve further steady-state kinetic analysis to be carried out on the gatekeeper variants with the nifedipine and amodiaquine substrates. These were effective substrates, but attempts to follow substrate-dependent NADPH oxidation by dual beam UV-Vis spectroscopy proved unsuccessful due to the strong color of the substrates. Instead, a fluorescence-based NADPH depletion assay could be developed to obtain kcat and Km parameters. Fast reaction kinetics (stopped-flow methods) could also be used to determine microscopic rate constants for steps in the catalytic cycle – such as binding of PPI and other drug substrates to the P450 (with induced heme spin-state shift) and substrate-dependent electron transfer to the P450 heme (ferrous heme could be “trapped” by carbon monoxide and the kinetics of electron transfer followed by development of the P450 Fe(II)CO complex at 450 nm). In this way, more detailed insights into the novel catalytic reactions performed by the BM3 variants could be obtained. A substantial amount of data was collected the proton pump inhibitor, pantoprazole, including binding, pantoprazole-dependent NADPH oxidation kinetics and product analysis. However, although crystal structural studies led to a DM heme domain structure with clear evidence for the binding of pantoprazole in the active site, we could not resolve the binding mode of the drug. This suggests that pantoprazole can occupy different binding poses in the DM active site. In efforts to resolve the binding mode of pantoprazole, further active site mutagenesis could be done (probably with reference to the omeprazole-bound DM structure) with the aim of restricting the active site binding mode of pantoprazole to decrease its mobility and enable its structure to be resolved. Ultimately, further engineering studies of this type should be done in order to produce more specific BM3 variants that are able to generate fewer and more specific oxidized products from PPIs and other drugs. According to predictions made in this thesis, we would expect such successful second phase variants to have greater thermostability (by DSC and/or DSF) than do their less specific/selective progenitors. Preparative work was carried out to generate libraries of mutations in the BM3 gene at the substrate recognition sequence (SRS) regions, which are areas and residues known to affect substrate selectivity, and which are found in all of the P450s (29). This work should also be extended, and further research in this area would include ligation of these individual SRS library “cassettes” into our gatekeeper variant

247

enzymes in an attempt to further diversify their activities. I would suggest screening these libraries using the stability assay both to identify residues that either increase or decrease stability of the BM3 heme domain. Variants selected should be examined for activities towards a wider panel of human drug substrates. Given the success with the gatekeeper variants, it would be particularly interesting to establish if any substrates identified lead to increased stability of the variant enzymes, as none of the non-PPI drugs investigated to date were seen to increase the Tm values of the gatekeeper variants. This work could also be extrapolated to whole cell LCMS screening by using small scale cultures from entire positional libraries and looking for altered activity to particular drugs in a high-throughput manner. Only when a desired metabolite or high level of enzymatic turnover is identified would we deconvolute the library to discover the mutations responsible for a novel activity. This approach should speed up the screening process for finding variants with improved activity towards targeted drugs, saving time and money. The further mutagenesis studies could also show whether the promiscuity seen in these gatekeeper BM3 point variants is reversible, driving the enzyme to be highly specific to another compound and away from the natural fatty acid substrates. To fulfil the requirements of the partner company Cypex, further work aimed at production of high value pharmaceutical metabolites should be done, with efforts focused on compounds in highest demand due to accumulated evidence for their pharmacological effects and difficulty in their chemical synthesis. A range of these variants should be tested for such biotechnologically important activities. This could be done by adding WT BM3 and variants to established screening trials for generation of various metabolites, which in many cases involve identifying fluorescence changes using marker substrates for particular human drug- metabolising P450 isoforms. This approach could again streamline the process of identifying productive BM3 variants and enable these variants to be used for reactions to make useful amounts of the desired human metabolite(s). This scale up would be achieved by generation of E. coli spheroplast cells to allow entry of the target drugs into the bacteria, which would then provide the required NADPH from endogenous sources and remove the expense of having to add the cofactor and any regeneration system (30). This would be followed by extraction and purification of the key metabolites produced.

248

References

1. Dong, H., Haining, R. L., Thummel, K. E., Rettie, A. E., and Nelson, S. D. (2000) Involvement of human cytochrome P450 2D6 in the bioactivation of acetaminophen. Drug Metab. Dispos. 28, 1397-1400 2. Black, D. J., Kunze, K. L., Wienkers, L. C., Gidal, B. E., Seaton, T. L., McDonnell, N. D., Evans, J. S., Bauwens, J. E., and Trager, W. F. (1996) Warfarin-fluconazole. II. A metabolically based drug interaction: in vivo studies. Drug Metab. Dispos. 24, 422-428 3. Tufvesson, P. r., Lima-Ramos, J., Nordblad, M., and Woodley, J. M. (2010) Process considerations for the asymmetric synthesis of chiral amines using transaminases.Org. Process Res. Dev. 15, 266-274 4. Gillam, E. M. J., and Guengerich, F. P. (2001) Exploiting the versatility of human cytochrome P450 enzymes: the promise of blue roses from biotechnology. IUBMB Life 52, 271-277 5. Caswell, J. M., O’Neill, M., Taylor, S. J. C., and Moody, T. S. (2013) Engineering and application of P450 monooxygenases in pharmaceutical and metabolite synthesis. Curr. Opin. Chem. Biol. 17, 271-275 6. Urlacher, V. B., and Girhard, M. (2012) Cytochrome P450 monooxygenases: an update on perspectives for synthetic application. Trends in Biotechnology 30, 26-36 7. van Vugt-Lussenburg, B. M. A., Stjernschantz, E., Lastdrager, J., Oostenbrink, C., Vermeulen, P. E., and Commandeur, J. N. M. (2007) Identification of critical residues in novel drug metabolizing mutants of cytochrome P450 BM3 using random mutagenesis. J. Med. Chem. 50, 455- 461 8. Shapiro, M. G., Westmeyer, G. G., Romero, P. A., Szablowski, J. O., Kuster, B., Shah, A., Otey, C. R., Langer, R., Arnold, F. H., and Jasanoff, A. (2010) Directed evolution of a magnetic resonance imaging contrast agent for noninvasive imaging of dopamine. Nat. Biotechnol. 28, 264-270 9. Warman, A. J., Roitel, O., Neeli, R., Girvan, H. M., Seward, H. E., Murray, S. A., McLean, K. J., Joyce, M. G., Toogood, H., Holt, R. A., Leys, D., Scrutton, N. S., and Munro, A. W. (2005) Flavocytochrome P450 BM3: an update on structure and mechanism of a biotechnologically important enzyme. Biochem. Soc. Trans. 33, 747-753

249

10. Whitehouse, C. J. C., Bell, S. G., and Wong, L.-L. (2012) P450(BM3) (CYP102A1): connecting the dots. Chem. Soc. Rev. 41, 1218-1260 11. Lind, T., van Zanten, S. V., Unge, P., Spiller, R., Bayerdörffer, E., O'Morain, C., Bardhan, K. D., Bradette, M., Chiba, N., Wrangstadh, M., Cederberg, C., and Idström, J.-P. (1996) Eradication of Helicobacter pylori using one-week triple therapies combining omeprazole with two antimicrobials: the MACH I Study. Helicobacter 1, 138-144 12. Ravichandran, K., Boddupalli, S., Hasermann, C., Peterson, J., and Deisenhofer, J. (1993) Crystal structure of hemoprotein domain of P450BM- 3, a prototype for microsomal P450's. Science 261, 731-736 13. Huang, W.-C., Westlake, A. C. G., Maréchal, J.-D., Joyce, M. G., Moody, P. C. E., and Roberts, G. C. K. (2007) Filling a hole in cytochrome P450 BM3 improves substrate binding and catalytic efficiency. J. Mol. Bio. 373, 633-651 14. DellaGreca, M., Iesce, M. R., Previtera, L., Rubino, M., Temussi, F., and Brigante, M. (2006) Degradation of lansoprazole and omeprazole in the aquatic environment. Chemosphere 63, 1087-1093 15. Ortiz de Montellano, P. R. (2005) Cytochrome P450: Structure, Mechanism, and Biochemistry, 3 ed., Kluwer Academic/Plenum Publishers, New York 16. Jovanovic, T., Farid, R., Friesner, R. A., and McDermott, A. E. (2005) Thermal equilibrium of high- and low-spin forms of cytochrome P450 BM-3: repositioning of the substrate? J. Am. Chem. Soc. 127, 13548-13552 17. Abelo, A., Andersson, T. B., Antonsson, M., Naudot, A. K., Skanberg, I., and Weidolf, L. (2000) Stereoselective metabolism of omeprazole by human cytochrome P450 enzymes. Drug Metab. Dispos. 28, 966-972 18. Usmani, K. A., Karoly, E. D., Hodgson, E., and Rose, R. L. (2004) In vitro sulfoxidation of thioether compounds by human cytochrome P450 and flavin- containing monooxygenase isoforms with particular reference to the CYP2C subfamily. Drug Metab. Dispos. 32, 333-339 19. Kumar, D., de Visser, S. P., Sharma, P. K., Hirao, H., and Shaik, S. (2005) Sulfoxidation mechanisms catalyzed by cytochrome P450 and horseradish peroxidase models: spin selection induced by the ligand. Biochemistry 44, 8148-8158 20. Ward, R. M., and Kearns, G. L. (2013) Proton pump inhibitors in pediatrics : mechanism of action, pharmacokinetics, pharmacogenetics, and pharmacodynamics. Paediatr. Drugs 15, 119-131

250

21. Volz, T. J., Rock, D. A., and Jones, J. P. (2002) Evidence for two different active oxygen species in cytochrome P450 BM3 mediated sulfoxidation and N-dealkylation reactions. J. Am. Chem. Soc. 124, 9724-9725 22. Rydberg, P., Jørgensen, M. S., Jacobsen, T. A., Jacobsen, A.-M., Madsen, K. G., and Olsen, L. (2013) Nitrogen inversion barriers affect the N-oxidation of tertiary alkylamines by cytochromes P450. Angew. Chem., Int. Ed. 52, 993-997 23. Reddy, G. M., Bhaskar, B. V., Reddy, P. P., Ashok, S., Sudhakar, P., Babu, J. M., Vyas, K., and Mukkanti, K. (2007) Structural identification and characterization of potential impurities of pantoprazole sodium. J. Pharm. Biomed. Anal. 45, 201-210 24. Niesen, F. H., Berglund, H., and Vedadi, M. (2007) The use of differential scanning fluorimetry to detect ligand interactions that promote protein stability. Nat. Protocols 2, 2212-2221 25. Kuper, J., Tee, K. L., Wilmanns, M., Roccatano, D., Schwaneberg, U., and Wong, T. S. (2012) The role of active-site Phe87 in modulating the organic co-solvent tolerance of cytochrome P450 BM3 monooxygenase. Acta Crystallogr. F-Struct. Biol. Cryst. Commun. 68, 1013-1017 26. Waldron, T. T., and Murphy, K. P. (2003) Stabilization of proteins by ligand binding: application to drug screening and determination of unfolding energetics. Biochemistry 42, 5058-5064 27. Guengerich, F. P., Parikh, A., Yun, C. H., Kim, D., Nakamura, K., Notley, L. M., and Gillam, E. M. (2000) What makes P450s work? Searches for answers with known and new P450s. Drug Metab. Rev. 32, 267-281 28. Guengerich, F. P. (1990) Low kinetic hydrogen isotope effects in the dehydrogenation of 1,4-dihydro-2,6-dimethyl-4-(2-nitrophenyl)-3,5- pyridinedicarboxylic acid dimethyl ester (nifedipine) by cytochrome P-450 enzymes are consistent with an electron/proton/electron transfer mechanism. Chem. Res. Toxicol. 3, 21-26 29. Gotoh, O. (1992) Substrate recognition sites in cytochrome P450 family 2 (CYP2) proteins inferred from comparative analyses of amino acid and coding nucleotide sequences. J. Biol. Chem. 267, 83-90 30. McGinnity, D. F., Griffin, S. J., Moody, G. C., Voice, M., Hanlon, S., Friedberg, T., and Riley, R. J. (1999) Rapid characterization of the major drug-metabolizing human hepatic cytochrome P-450 enzymes expressed in Escherichia coli. Drug metab. Dispos. 27, 1017-1023

251

Enzymology: Key Mutations Alter the Cytochrome P450 BM3 Conformational Landscape and Remove Inherent Substrate Bias

Christopher F. Butler, Caroline Peet, Amy E. Mason, Michael W. Voice, David Leys and Andrew W. Munro J. Biol. Chem. 2013, 288:25387-25399. doi: 10.1074/jbc.M113.479717 originally published online July 3, 2013

Access the most updated version of this article at doi: 10.1074/jbc.M113.479717

Find articles, minireviews, Reflections and Classics on similar topics on the JBC Affinity Sites.

Alerts: • When this article is cited • When a correction for this article is posted

Click here to choose from all of JBC's e-mail alerts

Supplemental material: http://www.jbc.org/content/suppl/2013/07/03/M113.479717.DC1.html

This article cites 57 references, 17 of which can be accessed free at http://www.jbc.org/content/288/35/25387.full.html#ref-list-1

Downloaded from http://www.jbc.org/ at The University of Manchester Library on September 2, 2013 THE JOURNAL OF BIOLOGICAL CHEMISTRY VOL. 288, NO. 35, pp. 25387–25399, August 30, 2013 © 2013 by The American Society for Biochemistry and Molecular Biology, Inc. Published in the U.S.A.

Key Mutations Alter the Cytochrome P450 BM3 Conformational Landscape and Remove Inherent Substrate Bias*□S Received for publication, May 7, 2013, and in revised form, July 2, 2013 Published, JBC Papers in Press, July 3, 2013, DOI 10.1074/jbc.M113.479717 Christopher F. Butler‡, Caroline Peet§, Amy E. Mason‡, Michael W. Voice§, David Leys‡, and Andrew W. Munro‡1 From the ‡Manchester Institute of Biotechnology, Faculty of Life Sciences, University of Manchester, 131 Princess Street, Manchester M1 7DN, United Kingdom and §Cypex Ltd., 6 Tom McDonald Avenue, Dundee DD2 1NH, Scotland, United Kingdom

Background: P450 BM3 is a high activity enzyme with biotechnological potential. Results: Mutations perturbing P450 BM3’s conformational state and active site facilitate human P450-like oxidation of the drug omeprazole. Conclusion: Conformational destabilization enables P450 BM3 to explore novel conformations and accept diverse substrates. Significance: “Gatekeeper” mutations that decrease the energetic barrier for transition to the substrate-bound state can reconfigure P450 BM3 specificity and reactivity.

Cytochrome P450 monooxygenases (P450s) have enormous tions, including key reactions involved in the biosynthesis of potential in the production of oxychemicals, due to their unpar- steroids, antibiotics, and signaling lipids (1, 2). Recent years alleled regio- and stereoselectivity. The Bacillus megaterium have seen intensive efforts to exploit the P450s’ ability to cata- P450 BM3 enzyme is a key model system, with several mutants lyze regio- and stereoselective oxidations of substrates to make (many distant from the active site) reported to alter substrate biotechnologically useful products. Examples include the ratio- selectivity. It has the highest reported monooxygenase activity nal engineering of the Pseudomonas putida camphor hydroxy- of the P450 enzymes, and this catalytic efficiency has inspired lase P450cam (CYP101A1) for improved binding and oxidation protein engineering to enable its exploitation for biotechnologi- of environmentally recalcitrant polychlorinated benzenes, and cally relevant oxidations with structurally diverse substrates. the use of directed evolution for the conversion of specificity of However, a structural rationale is lacking to explain how these the Bacillus megaterium P450 BM3 (CYP102A1, BM3) from mutations have such effects in the absence of direct change to long chain fatty acids toward short chain hydrocarbons (3, 4). the active site architecture. Here, we provide the first crystal BM3 has proven to be a particularly versatile and popular structures of BM3 mutants in complex with a human drug sub- model system for use in engineering studies, benefiting from strate, the proton pump inhibitor omeprazole. Supported by the fact that it is a natural fusion of a P450 to its mammalian- solution data, these structures reveal how mutation alters the like diflavin reductase redox partner (enabling efficient elec- conformational landscape and decreases the free energy barrier tron transfer that underpins its high monooxygenase activity (5, for transition to the substrate-bound state. Our data point to the 6)) and that the structure of its P450 domain and roles of many importance of such “gatekeeper” mutations in enabling major active site amino acids are well understood (7–10). changes in substrate recognition. We further demonstrate that Recent work to diversify BM3’s substrate selectivity and reac- these mutants catalyze the same 5-hydroxylation reaction as tivity has produced variants that catalyze olefin cyclopropana- performed by human CYP2C19, the major human omeprazole- tion by carbene transfer and oxidation of e.g. testosterone, poly- metabolizing P450 enzyme. cyclic aromatic hydrocarbons, and pharmaceuticals (11–14). In the latter case, an aim is to engineer BM3 to generate high levels of metabolites typical of those formed by human P450s. The cytochrome P450 monooxygenases (P450s)2 are hemo- Approaches to engineering BM3 have included directed evolu- proteins that catalyze a huge range of biochemical transformation, chimeragenesis (with homologs from Bacillus subtilis) and CASTing, as well as structure-led mutagenesis guided by * This work was supported by United Kingdom Biotechnology and Biological Sciences Research Council Research Grant BB/F00883X1 (to A. W. M. and x-ray crystal structures of BM3’s P450 (heme) domain in sub- D. L. supporting A. E. M.) and Industrial CASE Studentship BB/G01698/1 strate-free (SF), fatty acid-bound (SB), and various mutant from Cypex Ltd. (to A. W. M. and M. W. V. supporting C. F. B.). forms (7, 8, 15, 16). Many studies identified common residues □S This article contains supplemental Results, Figs. S1–S5, and Table S1. Theatomiccoordinatesandstructurefactors(codes4KEW,4KEY,4KF0,and4KF2) that help facilitate substrate diversification. These include Phe- have been deposited in the Protein Data Bank (http://wwpdb.org/). 87, Ala-82, Val-78, and Arg-47, the first three of which are 1 To whom correspondence should be addressed. Tel.: 44-161-3065151; Fax: internal residues, whereas Arg-47 interacts with the fatty acid 44-161-3068918; E-mail: [email protected]. 2 The abbreviations used are: P450, cytochrome P450 monooxygenase; BM3, carboxylate at the protein surface (6, 10). Phe-87 interacts with flavocytochrome P450 BM3; DM, BM3 F87V/A82F double mutant; DSC, dif- the ␻-end of fatty acids and prevents oxidation at this position. ferential scanning calorimetry; NPG, N-palmitoylglycine; OMP, omepra- Phe-87 mutations alter regioselectivity of fatty acid oxidation, zole; PPI, proton pump inhibitor; SB, substrate bound; SF, substrate free; ␻ 5-OH OMP, 5-hydroxy omeprazole; 5-COOH OMP, 5-carboxy omeprazole; with positions of hydroxylation of lauric acid moving from -1/ PDB, Protein Data Bank. ␻-2/␻-3 (WT BM3) toward ␻-5 in F87A, F87S, and F87G point

AUGUST 30, 2013•VOLUMEDownloaded 288•NUMBER from 35http://www.jbc.org/ at The University of Manchester LibraryJOURNAL on September OF BIOLOGICAL 2, 2013 CHEMISTRY 25387 Diversification of P450 BM3 Substrate Selectivity

mutants. Mutations (positions underlined) were generated using the QuikChange Lightning site-directed mutagenesis kit (Stratagene-Agilent UK). Primers used were as follows: A82F, 5Ј-CTTAAATTTGTACGTGATTTTTTCGGAGACGGGT- TA-3Ј; F87V, 5Ј-TTGCAGGAGACGGGTTAGTTACAAGC- TGGACGCATG-3Ј; F87V in A82F background, 5Ј-TTTTCG- FIGURE 1. Structure of omeprazole. The chemical structure of the proton GAGACGGGTTAGTTACAAGCTGGACGCATG-3Ј (and their pump inhibitor OMP is shown. The pyridine ring shows the accepted num- reverse complements). These intact BM3 enzymes were bering (hydroxylation occurs at the 5-methyl position). Also shown is the characteristic MS fragmentation position that gives the methoxybenzimida- expressed as N-terminal hexahistidine-tagged enzymes either zole and 4-methoxy-3,5-dimethylpyridin-2-yl (pyridinyl) fragments. Hydroxy- using the pET15b (F87V, DM) constructs directly or following lation on the 5-methyl group is performed by engineered variants of P450 cloning of the WT and A82F genes into pET14b using NdeI/ BM3 described in this study. 5-Hydroxylation is also the primary reaction catalyzed by the major human OMP-metabolizing enzyme CYP2C19. Omepra- BamHI sites. WT and mutant heme domain genes were gener- zole is chiral around the central sulfur atom. As a drug preparation, omepra- ated using the relevant pET14b/15b constructs. To generate the zole is a racemate of two isomers. heme domain constructs, a stop codon pair (underlined) was introduced after residue 473 by PCR using the same mutagen- mutants, and even further from the ␻-position in a F87A/V78A esis kit, and with primers StopF 5Ј-CAGTCTGCTAAAAAAG- double mutant (6, 17). The F87A mutation also activated BM3 TACGCAAATAGTAGGAAAACGCTCATAATACGCCGC- in hydroxylation of testosterone (12). TG-3Ј and StopR 5Ј-CAGCGGCGTATTATGAGCGTTTTC- Although alterations that influence local structure in the CTACTATTTGCGTACTTTTTTAGCAGACTG-3Ј. The heme BM3 active site cavity clearly have potential to alter the binding domain genes (amino acids 1–473) were transferred as NdeI/ position of fatty acids and to enable docking of novel substrates, BamHI fragments to pET20b to enable heme domain produc- other BM3 mutations were shown to have more profound tion in the absence of an N-terminal His tag for improved crys- effects on the P450’s conformational landscape (18, 19). One of tallization. Genes were sequenced to ensure the presence of the more common activity-altering BM3 mutants (A82F) desired mutation(s) and the absence of other mutations. The shows greatly enhanced affinity for fatty acid substrates. WT and A82F intact BM3 and the WT and all mutant P450 Although the crystal structure of palmitate-bound A82F heme BM3 heme domain proteins were expressed in BL21-Gold domain was solved in the SB-type conformation (20), this did (DE3) Escherichia coli cells (Stratagene-Agilent UK) using TB not provide a clear rationale for its much improved affinity for medium with cell growth at 37 °C and with agitation at 200 rpm fatty acids. Also, despite several studies involving mutagenesis in an orbital incubator. The F87V and DM intact BM3 proteins of BM3 to improve its capacity to oxidize human pharmaceuti- were grown in autoinduction TB medium (Melford Ltd., Ips- cals, there are no structural data to provide insight into how wich, UK). Typically, 4-liter bacterial cultures were used for improved mutants enable binding of the new substrates or to protein production with cell growth for 24–36 h. guide subsequent engineering to further enhance binding and Following cell growth, bacterial cells were recovered by cen- desired activities. trifugation at 4 °C (6000 ϫ g, 10 min) and resuspended in ice-

In this study we characterize the structural and biochemical/ cold buffer B (50 mM KPi, 250 mM NaCl, 10% (v/v) glycerol, pH catalytic properties of the A82F, F87V, and F87V/A82F 7.0) containing protease inhibitors (EDTA-free CompleteTM mutants of BM3, demonstrating a novel activity in oxidation of tablets, Roche Applied Science). Protease inhibitors were main- the widely used gastric proton pump inhibitor (PPI) omepra- tained in all subsequent buffers used for protein purification. zole (Fig. 1) and generating products typical of those formed by Cells were lysed by sonication on ice using a Bandelin Sonopuls the major human metabolizing P450 enzyme CYP2C19 (21). sonicator (40% power, 50 pulses for 5 s with 25 s between Structural data for substrate-free and omeprazole-bound forms pulses). The supernatant containing soluble intact BM3 and of the A82F and double mutant BM3 heme domains provide heme domain proteins was separated from cell debris by cen- clear evidence for the gatekeeper nature of the A82F mutation, trifugation (20,000 ϫ g, 40 min, 4 °C), and ammonium sulfate which produces a major structural rearrangement of the BM3 was added to 30% saturation on ice with slow stirring for ϳ4h. heme domain that leads to novel molecular selectivity. These Centrifugation (20,000 ϫ g, 40 min, 4 °C) was done to fraction- first structural data for BM3 in complex with a human drug ate soluble BM3 and heme domain proteins from insoluble substrate highlight how combinations of conformational effec- material. Intact BM3 WT and mutant proteins in the superna- tor mutations (e.g. A82F) with secondary mutations that make tant were purified using His tag affinity by mixing with nickel- local structural changes to alter binding in the heme vicinity iminodiacetic acid (Ni-IDA) resin (Qiagen, UK) overnight at (e.g. F87A/V) can be combined to cause dramatic changes in 4 °C in buffer B with 5 mM imidazole, prior to elution with 200 P450 substrate selectivity for biotechnological applications. mM imidazole in buffer B. Isolated proteins were dialyzed into buffer A (50 mM Tris, 1 mM EDTA, pH 7.2) and further purified EXPERIMENTAL PROCEDURES by size exclusion chromatography using a Sephacryl S-200 col- Generation, Expression, and Purification of WT and Mutant umn (GE Healthcare, 26 ϫ 60 cm on an AKTA purifier system). P450 BM3 Proteins—Mutants of the intact P450 BM3 and its BM3 fractions were checked for purity by SDS-PAGE, concen- heme domain were generated by oligonucleotide-directed trated by ultrafiltration (Vivaspin, Vivaproducts), and frozen in mutagenesis. Intact WT P450 BM3 in pET15b was used for buffer A plus 50% glycerol at Ϫ80 °C. For the nontagged BM3 mutagenesis to create A82F, F87V, and F87V/A82F (DM) heme domains, supernatants subsequent to the 30% ammo-

25388 JOURNAL OF BIOLOGICALDownloaded CHEMISTRY from http://www.jbc.org/ at The University of Manchester Library onVOLUME September 288•NUMBER 2, 2013 35•AUGUST 30, 2013 Diversification of P450 BM3 Substrate Selectivity

nium sulfate fractionation were subjected to a second ammo- visible spectrophotometer using substrate (OMP and NPG)- ⌬⑀ ϭ Ϫ1 nium sulfate fractionation step at 60% salt saturation. The pellet dependent NADPH oxidation at 340 nm ( 340 6210 M was resuspended in buffer A and extensively dialyzed into the cmϪ1) across a range of substrate concentrations and using same buffer to desalt, prior to loading onto a Q-Sepharose pure BM3 enzymes (25–125 nM) in a total of 1 ml of assay buffer anion exchange column (16 ϫ 10 cm, on an AKTA) and eluting at 25 °C. NADPH was maintained at a saturating concentration using a gradient of 0–500 mM KCl in buffer A. Heme domain- (200 ␮M). Data points were collected in at least triplicate. Rate containing fractions were desalted by using a desalting column constants for substrate-dependent NADPH oxidation were ϫ (GE Healthcare, 26 10 cm on an AKTA) into 25 mM KPi,pH plotted versus [substrate], and data were fitted using the 7.0, and then loaded onto a hydroxyapatite column (Bio-Rad, Michaelis-Menten function (using Origin Pro) to obtain kcat ϫ 16 11 cm) and eluted in a linear gradient of 25–500 mM KPi, and Km parameters in each case. Vmax values for WT and pH 7.0 (200 ml). Pure heme domain fractions were concen- mutant BM3 enzymes with 5-OH OMP were measured simi- trated by ultrafiltration (Vivaspin) and used immediately for larly using a saturating concentration of the metabolite. crystallography or flash-frozen in liquid nitrogen and stored at Omeprazole and 5-OH Omeprazole Turnover and Analysis Ϫ80 °C. Because of the enhanced affinity for fatty acids in BM3 by LC-MS—Turnover reactions for oxidation of OMP were car- and heme domain proteins carrying the A82F mutation, all ried out in deep well blocks at 37 °C with shaking for 30 min. enzymes were passed through a Lipidex 1000 column Reaction mixtures contained purified WT or mutant (F87V, ␮ ␮ (PerkinElmer Life Sciences) in 25 mM KPi, pH 7.0, to remove A82F, or DM BM3) enzymes (0.1 M), substrate (10 M), and any fatty acid bound during purification and prior to use for NADPH regeneration system (glucose 6-phosphate 7.76 mM, ϩ crystallization (heme domains) or for binding or turnover stud- NADP 0.6 mM, and glucose-6-phosphate dehydrogenase, 0.75

ies (heme domain and intact BM3 enzymes). units/ml) in turnover buffer (50 mM KPi,5mM CaCl2, pH 7.4) in Quantification of P450 BM3 Enzymes and Determination of a final volume of 500 ␮l. Following completion of the reaction, Their Substrate Affinity and Steady-state Kinetic Properties— protein was precipitated by addition of an equal volume of ace- Concentrations of the low spin forms of the WT and mutant tonitrile containing 1 ␮g/ml fluconazole by shaking the mixed forms of intact P450 BM3 and its heme domain were deter- samples at 800 rpm for 10 min. The precipitated protein was Ϫ1 mined using extinction coefficients of ⑀ ϭ 105 and 95 mM filtered through protein precipitation plates (Phenomenex, cmϪ1, respectively, at the Soret maximum (418 or 419 nm), as Macclesfield, UK) into mass spectrometry vials (FluidX, Nether described previously (6). Thiolate coordination of the heme Alderley, UK) and clarified by centrifugation (4000 ϫ g, 25 min, iron in all samples was established by formation of the Fe(II)CO 10 °C). Analysis was carried out on a Thermo Exactive LC-MS complex at ϳ450 nm by bubbling of sodium dithionite-reduced with a CTC PAL autosampler (Thermo Scientific, UK) with a WT and mutant BM3 and heme domain samples (ϳ2–4 ␮M) Kinetex 2.6U XB-C18 100A column (Phenomenex). A gradient with carbon monoxide gas, as described by Omura and Sato of 0.1% formic acid to acetonitrile was used to resolve products. (22). All samples showed near complete conversion to the P450 Reaction of the DM BM3 enzyme with 5-OH OMP and subse- form, with negligible formation of an ϳ420-nm peak relating to quent product analysis were done in the same way as for the the thiol-coordinated P420 form (23, 24). OMP turnovers.

Dissociation constants (Kd values) for binding of the sub- Omeprazole Turnover and Analysis by NMR—Turnover strates N-palmitoylglycine (NPG) and omeprazole (OMP) to reactions with OMP were carried out in a 100-ml flask at 37 °C WT and mutant BM3 heme domains were determined by UV- with shaking of reagents at 100 rpm for 30 min. Reaction mix- visible absorption titrations using ϳ1–4 ␮M protein in 100 mM tures contained purified WT or mutant (F87V, A82F, or DM ␮ ␮ KPi, pH 7.0, at 25 °C (assay buffer) in 1-cm path length quartz intact BM3) enzymes (1 M), substrate (100 M), NADPH ϩ cuvettes and as in our previous studies (6, 16, 25). Spectra were regeneration system (glucose 6-phosphate 7.76 mM, NADP recorded for substrate-free enzymes and following addition of 0.6 mM, and glucose-6-phosphate dehydrogenase 0.75 units/

ligands during substrate titrations (typically 800 to 250 nm). ml) in 60 ml of assay buffer (100 mM KPi, pH 7.0). Products were Titrations were recorded until no further spectral changes were extracted using Strata-X SPE columns (Phenomenex), dried

observed in the P450s. Difference spectra were generated by under vacuum, and eluted in CDCl3. Analysis was carried out subtraction of the spectrum for ligand-free protein from spec- on a Bruker Avance 400 MHz NMR (Bruker, Coventry, UK). 1H tra recorded after each addition of substrate. Maxima and min- spectra were collected at 400 MHz and 13C spectra at 101 MHz. ima in difference spectra were identified (using the same wave- Spectra were base-line corrected and referenced to tetrameth- length pair in each titration), and the overall absorbance ylsilane standard by the residual nondeuterated solvent in the ⌬ ␦ changes (Apeak minus Atrough) were plotted versus [substrate]. sample. values are in ppm;, J values are in Hz. Full assignments Data were fitted using either a standard (Michaelis-Menten) were made by COSY, HMBC, and HMQC methods. Signal hyperbolic function or (for tight binding substrates where the splittings were recorded as singlet (s), doublet (d), doublet of Յ ϫ ␣␤ Kd value is 5 the P450 concentration) by using the Morrison doublets (dd), system (AB), and multiplet (m). Processing equation (as described previously) to determine Kd values (26, was carried out using MestReNova Lite (Mestrelab Research, 27). UV-visible spectroscopy was carried out on a Cary 50 UV- Santiago de Compostela, Spain) and an NMR processor visible spectrometer (Cary-Agilent, UK), with data analysis and (Advanced Chemistry Development, Inc., Toronto, Canada). fitting done using Origin Pro software (OriginLab). Examination of Hemoprotein Stability by Differential Scan- Steady-state kinetic parameters for WT and F87V, A82F, and ning Calorimetry—DSC was carried out on a Microcal VP-DSC DM mutants of intact BM3 were analyzed on a Cary 50 UV- instrument. Data analysis was done using Microcal Origin soft-

AUGUST 30, 2013•VOLUMEDownloaded 288•NUMBER from 35http://www.jbc.org/ at The University of Manchester LibraryJOURNAL on September OF BIOLOGICAL 2, 2013 CHEMISTRY 25389 Diversification of P450 BM3 Substrate Selectivity

ware. The parameters used were as follows: 20–80 °C temper- TABLE 1 ature gradient, 90 °C/h scan rate, 10-min prescan thermostat. Substrate binding and turnover data for BM3 enzymes Background scans were carried out with degassed assay buffer Table shows data for the binding of substrates (NPG and OMP) to WT and A82F, F87V, and DM P450 BM3 enzymes (Kd values from optical titrations) and for the and saturated with OMP or NPG for the substrate-bound sam- kinetics of substrate-dependent NADPH oxidation (kcat and Km values). All data ples. Protein samples were prepared in assay buffer by extensive points were collected in triplicate; errors are S.E. NA indicates that no evidence of binding of OMP to WT BM3 was found in optical titrations. buffer exchange and dialysis. All samples were run using 20 ␮M Protein Substrate Kd kcat Km protein and saturating substrate. Once two overlapping base- Ϫ1 ␮M min ␮M line scans were achieved, a degassed protein sample was run. WT NPG 0.082 Ϯ 0.011 4770 Ϯ 160 13.9 Ϯ 2.7 Crystallization of P450 BM3 Heme Domains and Determina- A82F NPG 0.297 Ϯ 0.069 5130 Ϯ 570 26.3 Ϯ 5.3 F87V NPG 0.204 Ϯ 0.045 4970 Ϯ 240 14.9 Ϯ 2.8 tion of Protein Structures—Crystallography was performed DM NPG 0.004 Ϯ 0.003 4050 Ϯ 250 1.91 Ϯ 0.31 using the sitting drop method using a seeding protocol at 4 °C. WT OMP NA 238 Ϯ 12 124 Ϯ 20 A82F OMP 1.67 Ϯ 0.05 1460 Ϯ 30 18.7 Ϯ 0.8 Crystals obtained during initial screens for each mutant were F87V OMP 49.0 Ϯ 2.7 2180 Ϯ 20 38.7 Ϯ 4.6 used to create microcrystal screen stocks, and consecutive DM OMP 0.212 Ϯ 0.014 1500 Ϯ 45 1.93 Ϯ 0.22 screens (Molecular Dimensions) were made with drops that Ͻ ␮ consisted of 150 nl of WT or mutant heme domain proteins and all mutants bound NPG tightly, with Kd values 1 M (230 ␮M), 50 ␮l of seed stock, and 200 nl of well solution using a (Table 1). Mosquito liquid handling robot (TTP LabTech Ltd., Melbourn, The A82F mutant has particularly high affinity for lipids and UK). For the OMP mutant heme domain complex structures, was purified from E. coli with fatty acid bound, as was the DM. proteins were saturated with racemic OMP ligand prior to crys- Separation of the bound lipid by Lipidex chromatography was tallization. Ligands were titrated into heme domain samples done prior to analysis. Binding studies with the PPI omeprazole until no further change in heme iron spin state (toward high (OMP) revealed no evidence for its association with WT BM3, spin) was observed. Thereafter, samples were concentrated by but partial spin-state conversion was observed for F87V BM3, ϭ ultrafiltration in the presence of saturating ligand. Micro seed- indicative of the ability of the drug to access the active site (Kd ␮ ing was also used to produce diffraction quality crystals. Crys- 49 M). Much tighter binding and more extensive high spin tals were obtained under a range of conditions and flash-cooled heme accumulation was observed with both A82F and the DM ϭ ␮ in liquid nitrogen prior to data collection. The mother liquor (Kd values 1.67 and 0.21 M, respectively), indicating that the was supplemented with 10% PEG 200 where an additional cryo- mutations had synergistic effects in enhancing OMP affinity, protectant was required. Data were collected at Diamond syn- with A82F having the major role (Fig. 2A and Table 1). chrotron beamlines and reduced and scaled using XDS (28). Steady-state Kinetics of P450 BM3 Mutants with Omeprazole— Structures were solved by molecular replacement with previ- Steady-state kinetic analysis was done for WT and each of the ously solved BM3 heme domain structures (PDB 1JPZ) using intact BM3 mutants by measuring NADPH oxidation following PHASER (29). Structures were refined using Refmac5 (29) and the addition of various concentrations of NPG or OMP (24). Table 1 details high k values for all BM3 enzymes with NPG Coot (30). PDB codes for the new heme domain structures are cat (4050–5130 minϪ1), with DM BM3 having the lowest K value as follows: A82F, 4KF0; F87V/A82F (DM, imidazole-bound), m (1.91 ␮M). With OMP, some substrate-stimulated NADPH oxi- 4KF2; A82F (OMP-bound), 4KEW; and DM (OMP-bound), dation was seen for WT BM3 (k ϭ 238 minϪ1), but kinetics 4KEY. cat were substantially improved in all the mutants (e.g. 1500 minϪ1 Oligonucleotide primers were from Eurofins MWG Operon for DM), and apparent OMP affinity for the DM was ϳ65-fold (Ebersberg, Germany). Omeprazole was from Cypex Ltd. greater than for WT BM3 (K values of 1.9 ␮M versus 124 ␮M), (Dundee, Scotland, UK). 5-OH omeprazole was from Santa m resulting in a Ͼ400-fold improvement in catalytic efficiency Cruz Biotechnology, Inc. (Dallas). The bacterial growth (k /K ratio) for the DM over WT BM3. medium (Terrific Broth) was from Melford Ltd (Ipswich, UK). cat m Oxidation of Omeprazole by WT and Mutant P450 BM3 Unless otherwise stated, other chemicals were from Sigma and Enzymes—To validate novel omeprazole oxidase activity in of the highest purity available. mutant BM3 enzymes, in vitro turnover studies were done as described in detail under “Experimental Procedures,” using RESULTS LC-MS and NMR to characterize products formed. WT BM3 Characterization of Omeprazole Binding Properties of BM3 exhibited very small amounts of oxidation of OMP (Ͻ1% of the Mutants—Preceding studies identified Ala-82 and Phe-87 as starting material), but the low amounts obtained precluded important in controlling molecular selectivity and regioselec- determination of position(s) of oxidation. However, each of the tivity of BM3 substrate oxidation (16, 20, 31). WT, F87V, A82F, three mutants extensively oxidized the drug. and the F87V/A82F DM forms of intact P450 BM3 and of its LC-MS revealed a 362.1162 atomic mass unit species in all heme domain (residues 1–473) were expressed and purified as the mutant turnover reactions, corresponding to introduction described under “Experimental Procedures.” The BM3 ferric of an oxygen atom into OMP (Fig. 3). A natural fragmentation heme iron undergoes a shift from low spin toward high spin on of the OMP molecule occurs in the MS, with bond breakage binding substrates that displace its 6th ligand (a water mole- between the sulfur and the methoxybenzimidazole moiety (Fig. cule), accompanied by a shift of Soret absorption maximum 1). A ϩ16 increase in mass of the larger fragment indicated that from ϳ418 nm to ϳ392 nm (32). Binding of the substrate NPG oxidation occurs on this portion of OMP. Further fragmenta- was done for WT, F87V, A82F, and the DM BM3 proteins. WT tion of the oxidized molecule showed that the position of oxi-

25390 JOURNAL OF BIOLOGICALDownloaded CHEMISTRY from http://www.jbc.org/ at The University of Manchester Library onVOLUME September 288•NUMBER 2, 2013 35•AUGUST 30, 2013 Diversification of P450 BM3 Substrate Selectivity

FIGURE 2. Binding and oxidation of omeprazole by P450 BM3 mutants. A, binding titration for F87V/A82F (DM) intact P450 BM3 (1 ␮M) with omepra- ⌬ zole. Main panel, plot of the induced heme Soret absorption change ( A389– ⌬ ϭ Ϯ ␮ A419) versus [OMP] with data fitted to yield a Kd 0.212 0.014 M. Inset, selected OMP-induced absorption difference spectra from titration at OMP concentrations: 0.05 ␮M (green), 0.15 ␮M (blue), 0.30 ␮M (magenta), 0.40 ␮M (purple), and 1.0 ␮M (red). B, turnover data for OMP with WT (black column), A82F (red), F87V (blue), and F87V/A82F (DM, orange) P450 BM3 enzymes. Assays were done for 30 min. Products (5-OH OMP and 5-COOH OMP) are shown as a percentage of the initial OMP concentration used in the assay, with data corrected for an internal standard (fluconazole) and for enzyme- independent degradation of OMP substrate. dation was likely on one of the two 4-methoxy-3,5-dimethylpyridin-2-yl (hereafter termed pyridinyl) methyl groups (Fig. 3). This was confirmed by 1H NMR spectroscopy after larger scale reactions of mutant enzymes with OMP, and the absolute position of oxidation was confirmed using two-dimensional NMR, FIGURE 3. LC-MS analysis of products derived from omeprazole oxidation showing that a specific hydroxylation occurred on the pyridinyl by the P450 BM3 F87V/A82F (DM) double mutant enzyme. The figures show data from LC-MS studies of OMP before and after its enzymatic turnover by the 5-methyl group (forming 5-OH OMP) (supplemental Figs. P450 BM3 DM (F87V/A82F) enzyme. These data demonstrate hydroxylation and S1–S4). This same reaction is also catalyzed by CYP2C19, the subsequent oxidation of OMP, and also the fragmentation of the OMP (and its ϭ major metabolizing P450 for OMP and other PPIs in humans oxidized products) that occurs during MS analysis. A (retention time 5.39 min) shows data for OMP prior to addition of enzyme and initiation of its oxidation by (33). the BM3 DM enzyme. Peaks at m/z346.1212 and 198.0581 (circled) correspond to In OMP turnovers done over 30 min, the total oxidative turn- the fragmentation of OMP at the sulfone group (between the sulfur and the ϳ methoxybenzimidazole moiety), with the smaller species representing the sul- over of OMP was greatest for the DM ( 55%), followed by the fur-containing fragment. B (retention time ϭ 5.26 min) is following an enzymatic F87V (ϳ50%) and the A82F (ϳ20%) mutants (Fig. 2B). LC-MS reaction for 30 min. The m/z peaks at 362.1162 and 214.0530 (circled) are for the demonstrates the formation of a considerable amount of a ϩ32 5-OH OMP and its hydroxylated fragment. C, (retention time ϭ 5.32 min) is following an enzymatic reaction for 30 min. The m/z peaks at 376.0955 and species (ϳ10%) in the case of the DM, which was shown to be 228.0323 (circled) are for the 5-COOH OMP and its carboxylated fragment. the carboxylic acid product at the pyridinyl 5-methyl group (5-COOH OMP), resulting from further P450-mediated oxida- the A82F and DM BM3 enzymes and 39% for the F87V variant. tion at the same position. By comparing the amount of NADPH These data suggest that structural changes induced by the A82F oxidized with the quantity of OMP oxidized over the first 2 min mutation are the major determinant of the productive binding of reactions, enzymatic coupling was estimated as 68% for both mode for OMP, and thus the coupling efficiency.

AUGUST 30, 2013•VOLUMEDownloaded 288•NUMBER from 35http://www.jbc.org/ at The University of Manchester LibraryJOURNAL on September OF BIOLOGICAL 2, 2013 CHEMISTRY 25391 Diversification of P450 BM3 Substrate Selectivity

Consistent with the model of further BM3 mutant-catalyzed A82F mutation, however, are not intuitive given the peripheral oxidation of the primary product, 5-OH OMP showed tight position of Ala-82 to the substrate binding pocket. Ϯ binding to both A82F and the DM (Kd values of 15.2 0.9 and To gain detailed understanding of the effects of the A82F 2.62 Ϯ 0.19 ␮M, respectively) (Table 2 and Fig. 4), indicating mutation in promoting OMP binding, both the A82F and the that oxidized products do bind the enzyme. The further oxida- F87V/A82F double mutant were co-crystallized with OMP, and tion of 5-OH OMP to 5-COOH OMP was confirmed by LC-MS structures solved by molecular replacement using the NPG- studies with the BM3 DM enzyme (data not shown). Vmax val- bound heme domain structure (PDB code 1JPZ) (34). Crystal ues determined for 5-OH OMP-stimulated NADPH oxidation structure PDB codes are 4KEW for the OMP-bound A82F in the BM3 mutants were 1050 Ϯ 50 minϪ1 (A82F), 645 Ϯ 30 heme domain, and 4KEY for the OMP-bound DM heme minϪ1 (F87V), and 785 Ϯ 35 minϪ1 (DM), but NADPH oxida- domain. In both cases, the global conformation of the heme tion was not stimulated significantly above background in the domain is remarkably similar to previously determined fatty- WT BM3 (Table 2). A time course for the oxidation of OMP acid bound structures (0.3 Å over 450 C␣ atoms) (Fig. 6A, Table into 5-OH OMP and 5-COOH OMP products by the BM3 DM 3) (8, 34). enzyme is shown in Fig. 5. Approximately 30% of the 5-OH The PPI ligand is clearly identified from the electron density OMP is further oxidized to 5-COOH OMP by the DM BM3 occupying the substrate binding channel for the OMP complex. over 30 min. Structural Analysis of Omeprazole-binding P450 BM3 Mutants—The effect of the F87V mutation in enabling OMP binding may be explained using available structural information for the BM3 heme domain, as space vacated by the F87V substitution directly above the heme plane is likely to allow binding of the bulky substrate (6). The marked effects of the

TABLE 2 Binding and kinetics of oxidation of 5-OH OMP by WT and mutant P450 BM3 enzymes

The table shows Kd values (where determinable) for the binding of the oxidized product 5-OH OMP to WT and mutant intact P450 BM3 enzymes. The Kd values were derived from optical titrations. Also shown are Vmax values for 5-OH OMP- dependent NADPH oxidation catalyzed by WT, F87V, A82F, and DM BM3 enzymes. All data points were collected in triplicate; errors are S.E. 5-OH OMP is the primary metabolite of OMP generated by human CYP2C19 and by the BM3 mutants (most efficiently by the F87V/A82F DM). Data were collected as described under “Experimental Procedures.” ND indicates not determinable. FIGURE 5. Time course of substrate oxidation and product formation in Protein Substrate K V d max the reaction of the P450 BM3 DM enzyme with omeprazole. OMP sub- Ϫ1 ␮M min strate is shown in black squares, and the products 5-OH OMP and 5-COOH WT 5-OH OMP ND ND OMP are shown in open circles and open triangles, respectively. Reactions were A82F 5-OH OMP 15.2 Ϯ 0.9 1050 Ϯ 50 done as described under “Experimental Procedures.” The reactions reach F87V 5-OH OMP ND 645 Ϯ 30 completion in ϳ10–15 min, with most substrate oxidation (and 5-OH OMP DM 5-OH OMP 2.62 Ϯ 0.19 785 Ϯ 35 formation) occurring in the first 2.5 min.

FIGURE 4. Optical binding titration for the BM3 DM heme domain with 5-OH OMP. A shows UV-visible binding spectra for a titration of intact DM BM3 (ϳ1.0 ␮M, red spectrum) and following addition of 2 ␮M (blue), 6 ␮M (magenta), 16 ␮M (orange), and 40 ␮M (black) 5-OH OMP. The Soret band shifts from 418 to 393 nm on binding 5-OH OMP. The inset shows difference spectra obtained by subtraction of the substrate-free spectrum from each of the shown 5-OH-bound spectra ⌬ Ϫ⌬ (color coding remains same). B shows a plot of induced Soret absorbance change ( A389 nm A421 nm) versus the relevant [5-OH OMP], with data fitted using Ϯ ␮ the Morrison equation to give a Kd value of 2.62 0.19 M (26). 5-OH OMP binds both A82F and DM heme domains to induce a substrate-like shift in heme iron spin state equilibrium toward the high spin state. Full Kd and Vmax data for 5-OH OMP binding/turnover with WT and mutant BM3 enzymes are given in Table 2.

25392 JOURNAL OF BIOLOGICALDownloaded CHEMISTRY from http://www.jbc.org/ at The University of Manchester Library onVOLUME September 288•NUMBER 2, 2013 35•AUGUST 30, 2013 Diversification of P450 BM3 Substrate Selectivity

FIGURE 6. Structures of P450 BM3 enzymes and their omeprazole-binding sites. A, comparison of WT and mutant BM3 heme domain structures. The FG-helices are in color, and the remainder of the protein structures is depicted in grayscale. The A82F mutation is shown in spheres (where present) and substrate molecules are shown in atom-colored spheres. The heme is shown as red sticks. Panel 1, F87V/A82F (DM) P450 BM3 mutant heme domain in complex with OMP. Panel 2, DM heme domain in the ligand-free form. Panel 3, WT heme domain complex with NPG (PDB code 1JPZ) (34). Panel 4, WT heme domain in the ligand-free form (PDB code 1BU7) (57). B, mode of binding of OMP is shown for the DM (left panel) and A82F (right panel) mutant BM3 heme domain active sites. Because of weak electron density, the labile sulfone oxygen is omitted from the models shown. Key residues contacting the ligand are shown as sticks, and water molecules hydrogen bonding to the OMP are in red. Right panel, OMP from the A82F heme domain structure is overlaid with that from the DM heme domain. Right panel, Phe-82 residues are shown in green for the DM and in cyan for the A82F mutant. The DM Val-87 is in green, and the A82F Phe-87 is in cyan. The distance between the P450 heme iron and the OMP 5-methyl group is 3.9 Å in the A82F heme domain and 4.1 Å in the DM heme domain.

TABLE 3 Data reduction and final structural refinement statistics for P450 BM3 mutants and their OMP-substrate complexes A82F (4KF0), DM-imidazole (4KF2), A82F-OMP (4KEW), DM-OMP (4KEY),

space group P21 space group P21 space group P212121 space group P212121 Cell parameters a ϭ 59.1 Å, b ϭ 147.2 Å, a ϭ 59.3 Å, b ϭ 151.7 Å, a ϭ 59.4 Å, b ϭ 129.5 Å, a ϭ 59.3 Å, b ϭ 130.7 Å, c ϭ 64.0 Å, ␤ ϭ 97.5° c ϭ 60.8 Å, ␤ ϭ 95.9° c ϭ 145.1 Å c ϭ 146.0 Å Resolution 64 to 1.45 Å (1.5 to 1.45 Å) 47 to 1.82 Å (1.86 to 1.82) 65 to 1.89 (1.94 to 1.89) 55 to 2.05 (2.1 to 2.05) Rmerge 7.8% (35.5%) 6.8% (47.2%) 8.4% (43.8%) 8.5% (39.2%) I/␴I 16.2 (2.1) 14.4 (1.9) 11.8 (3.2) 11.4 (3.0)

R/Rfree 14.5/17.8% (28.1/32.3%) 18.2/22.2% (27.6/31.3%) 18.8/23.6% (28.5/33.6%) 18.7/23.5% (24.2/29.2%) Average B 16.8 Å2 22.2 Å2 19.7 Å2 30.5 Å2 Root mean square 0.025 Å/2.04° 0.024 Å/1.91° 0.022 Å/1.88° 0.022 Å/1.79° deviation bonds/angles

Crystallization conditions 25% PEG2000MME, 15% PEG20K, 15% PEG4K, 0.2 M MgCl2, 15% PEG4K, 0.2 M MgCl2, 0.2 M MgCl2,pH6.5 15% PEG550MME 0.06 M MgCl2, pH 6.5 (0.1 M sodium pH 6.5 (0.1 M sodium (0.1 M sodium cacodylate) pH 6.5 (0.1 M imidazole/MES) cacodylate) cacodylate)

In both cases, electron density corresponding to the oxygen above the heme plane. The close distance of the heme to the atom of the central sulfinyl groups is weak, and it is possible OMP 5-methyl group, in particular, leads to displacement of both stereoisomers are present from the racemic mixture. the water 6th ligand, and this conformation is consistent with OMP was also shown to be relatively labile, with loss of oxygen the observed 5-OH OMP product. The heme iron to OMP from the sulfone noted in aqueous solution (Fig. 6B) (35). We 5-methyl group distances are 3.9 and 4.1 Å in the A82F and DM thus modeled OMP without the sulfone oxygen. The pyridinyl heme domains, respectively. A single direct hydrogen bond is moiety is placed in a near-perpendicular orientation directly made between the backbone carbonyl of Leu-437 and one of the

AUGUST 30, 2013•VOLUMEDownloaded 288•NUMBER from 35http://www.jbc.org/ at The University of Manchester LibraryJOURNAL on September OF BIOLOGICAL 2, 2013 CHEMISTRY 25393 Diversification of P450 BM3 Substrate Selectivity methoxybenzimidazole nitrogens. Water molecules are present in close proximity to the other OMP polar groups, and these also mediate a network of hydrogen bonds between protein and ligand. Specific interactions are made via a bridging water molecule between (i) the Ala-74 backbone nitrogen and the methoxy oxygen of the methoxybenzimidazole and (ii) the hydroxyl oxygen of Ser-72 and the second methoxybenzimidazole nitrogen. Although there are no substantial differences in the orientation of the methoxybenzimidazole group for both mutants, close contacts between Phe-87 and the pyridinyl moiety in the A82F mutant structure lead to small reorientation of the substrate in comparison with the DM structure. In the latter, the reduced bulk of the Val-87 appears to provide sufficient space for the substrate to adopt a less strained conformation. This likely explains the difference in affinity observed between the A82F and DM mutants. Other protein-OMP interactions in the A82F and DM complex structures involve hydrophobic interactions between residues near both the pyridinyl (e.g. Thr- 438, Ile-263, and Ala-328) and methoxybenzimidazole (e.g. Val-26 and Leu-188) ends of the OMP substrate (Fig. 7). A clear rationale for the marked difference in OMP affinity between A82F containing mutants and the WT BM3 is not provided by the mutant-OMP structures. Fig. 8A shows the active sites in detail for the overlaid A82F OMP-bound BM3 heme domain and the NPG-bound WT heme domain. The additional bulk of the Phe-82 side chain only interacts edge-on with the pyridinyl moiety of OMP, and the only significant change in the positions of nearby key residues occurs for Phe- 87. The Phe-87 side chain is seen in different conformations in the A82F/OMP structure, compared with the WT BM3/NPG FIGURE 7. Interactions of omeprazole in the active site of the A82F BM3 heme domain. Fig. 8B shows an alternative view of the active heme domain. The diagram shows the binding site of OMP (without the labile sulfinyl oxygen) in the A82F mutant BM3 heme domain. For OMP, car- site for the DM/OMP-bound heme domain and the WT/NPG- bon atoms are shown in black, oxygens in red, sulfur in yellow, nitrogens in bound P450, looking along the I-helix, with the F/G-helices blue, and the oxygens of water molecules in cyan. Bonds in the OMP substrate highlighted. This reinforces the strong structural similarity are shown in purple, and bonds in selected amino acids are in brown. Hydro- gen bonds are shown (with their lengths) as green dashed lines. Amino acids between the WT and DM mutant substrate-bound structures. making hydrophobic interactions with the OMP are shown as red arcs with All these structures occupy the SB conformation. However, the radiating lines. OMP atoms involved in these hydrophobic interactions are shown with radiating red lines. A direct hydrogen bond interaction is made ligand-free A82F and DM heme domains also have the SB con- between the backbone carbonyl of Leu-437 and one of the OMP benzimida- formation (Fig. 6A). However, the WT BM3 heme domain zole group nitrogens (NE1). A further bridging hydrogen bond occurs from occupies a distinct conformation when SF, and thus the A82F the Ser-72 hydroxyl group through a water molecule (water 781) to the other benzimidazole nitrogen (NV). A final bridging hydrogen bond interaction mutation is key to shifting the equilibrium between SF and SB occurs between the backbone nitrogen of Ala-74 and the benzimidazole conformations. Previous studies also showed that the structural methoxy oxygen (O3) via another water molecule (water 761). A number of state of the heme domain can be significantly affected by single hydrophobic protein-OMP interactions are seen. These include interactions with Leu-188 at the benzimidazole methoxy group (C4) and with Ala-328 at mutations at key positions (36). the pyridinyl 5-methyl group (C1). The diagram was produced using Ligplotϩ We also determined the crystal structures of the ligand-free using the structure of the A82F heme domain-omeprazole complex solved in and imidazole-bound A82F (PDB code 4KF0) and DM (4KF2) this study (PDB code 4KEW) (58). heme domains. In the latter case, imidazole present in the crystallization buffer ligated to the heme iron. These structures different orientation compared with the SB conformation. In share a similar conformation that is again distinct from previ- the case of the A82F-containing mutants, this motion leads to a ously observed WT BM3 heme domain SF structures (Figs. 6A close contact between Ile-263 and Phe-82 (Fig. 9). This results and 9). These data again point to the A82F substitution produc- in a minor reorientation of the I-helix and the Ile-263 side ing dramatic changes in the structure and behavior of ligand- chain, which directly affects the position of the G-helix residue free A82F-containing mutants. The most significant changes Met-177. The repositioning of the G-helix is accompanied by occur in the positioning of the FG-helices, which adopt a con- reorientation of hydrophobic residues on the B- and F-helices. formation that places the FG-loop further away from the pro- Although the F87V mutant enables binding of OMP detect- tein core, leading to increased mobility. This altered position is able by heme absorbance shift, the Kd value for OMP binding to a consequence of a reorganization of the hydrophobic contact F87V BM3 heme domain is ϳ30- and 230-fold weaker than that between the FG-helices and the I-helix. In the WT SF confor- observed for the A82F and DM heme domains, respectively. mation, the N-terminal region of the I-helix adopts a slightly The improved ligand binding properties of the A82F mutants

25394 JOURNAL OF BIOLOGICALDownloaded CHEMISTRY from http://www.jbc.org/ at The University of Manchester Library onVOLUME September 288•NUMBER 2, 2013 35•AUGUST 30, 2013 Diversification of P450 BM3 Substrate Selectivity

FIGURE 8. Stereoviews of structural overlays of substrate-bound forms of the BM3 A82F-containing mutant heme domains with WT BM3. A shows a stereoview of the A82F-OMP heme domain active site (in red) with that of the WT-NPG structure (PDB 1JPZ in blue) (34). Key amino acids are shown in lines, and the bound ligands are shown in atom-colored sticks (OMP with magenta carbons; NPG with light blue carbons). Besides the nature of the ligand itself, and the obvious difference of the A82F mutation, there are very few differences between the structures, and these are mainly limited to Phe-87 occupying multiple conformations in the A82F-OMP structure. B shows an alternative view of the BM3 double mutant (DM, F87V/A82F) OMP-bound heme domain structure overlaid with the NPG substrate-bound structure of the WT BM3 heme domain. The F/G-helix region is colored in red for the BM3 DM and in blue for the substrate-bound WT BM3. Key amino acid residues and the respective ligands are shown as sticks. As also seen for A, surprisingly little change can be observed in the DM protein structure compared with WT BM3, despite the distinct nature of the ligand and the introduction of two mutations.

the free energy difference between both conformations and the free energy associated with ligand binding. The A82F-containing mutant crystal structures suggest that the mutation leads to a substantially altered free energy difference between both conformations. The fact OMP only appears to bind with measurable affinity to A82F-containing mutants suggests that the A82F mutation destabilizes the substrate-free conformation. To confirm this hypothesis, we performed DSC analysis of the substrate-free and OMP-bound forms of WT and all mutant heme domains, and we compared data with NPG- bound forms. The results showed that both F87V and A82F mutations diminished thermal stability of the BM3 heme domain. The WT heme domain has two unfolding transitions

(Tm values) at 65.7 °C (Tm1, major) and 59.0 °C (Tm2, minor). The Tm values are not altered significantly by OMP, but NPG FIGURE 9. Structural overlay of the omeprazole-bound A82F mutant with ϭ the WT BM3 heme domain. A stereoview is shown for a structural overlay of stabilizes the P450 (Tm1 70.4 °C). Consistent with our model, the A82F BM3 heme domain with the WT heme domain (PDB 1BU7). Color the A82F T values are 59.7 and 49.7 °C, substantially lower coding is as in Fig. 6, with F/G-helices in green for the substrate-free A82F m mutant and in yellow for the substrate-free WT BM3 heme domain. than WT BM3. F87V was also destabilized, albeit to a lesser ϭ extent, with the minor transition no longer seen (Tm1 ϭ can be understood in view of the large changes introduced by 61.3 °C). The DM showed greatest destabilization (Tm the A82F mutation in the ligand-free but not ligand-bound 50.9 °C and 58.0 °C). Binding of OMP to A82F-containing structures. The shift in the BM3 conformational equilibrium mutants resulted in single unfolding transitions, with negligible ϳ from SF to SB (as induced by ligand binding) is dependent on change in Tm1 for A82F (59.6 °C) and an 2 °C stabilization for

AUGUST 30, 2013•VOLUMEDownloaded 288•NUMBER from 35http://www.jbc.org/ at The University of Manchester LibraryJOURNAL on September OF BIOLOGICAL 2, 2013 CHEMISTRY 25395 Diversification of P450 BM3 Substrate Selectivity

major issues. An attractive alternative is to engineer high activity microbial P450s for specific oxidation of drugs. A system of choice is P450 BM3, due to its catalytically self-sufficient nature, high turnover rates, and availability of excellent structural data to guide protein engineering. Crys- tal structures of substrate-free and fatty acid-bound forms of the wild-type BM3 heme domain drove early protein engineering studies on BM3 (7, 8). This enabled identification of residues important in binding the substrate carboxylate at the mouth of the active site (Arg-47 and Tyr-51), as well as amino acids crucial for regulating heme iron redox potential (Phe-393) and the coupling of electron transfer to substrate oxidation (Thr-268) (6, 38–40). Parallel studies documented its fast rates of fatty acid substrate oxidation, resulting from rapid electron transfer from NADPH through its fused cytochrome P450 reductase partner (5, 41), and more recent studies have confirmed that the enzyme is functional as a dimer (42, 43) and that electron transfer to drive catalysis occurs between the NADPH-cytochrome P450 reductase module of one monomer and the heme domain of the other in the BM3 dimer (44). The application of random mutagenesis, recombination, and directed evolution approaches (pioneered by Arnold and co- workers) brought a new dimension to research on BM3, demonstrating that radical changes in its substrate selectivity could be engineered. A key development was the production of the 139-3 BM3 mutant (containing 11 mutations), which enabled FIGURE 10. Conformational equilibria and the relationship with struc- the hydroxylation of a range of alkanes from octane through to tural stability in P450 BM3. A, DSC data for the WT and DM P450 BM3 heme domains in substrate-free, OMP-, and NPG-bound forms. B, schematic over- propane, including cyclohexane (45). However, only one muta- view of the conformational equilibria proposed for the BM3 WT and DM tion (V78A) occurred at a position known to be in contact with mutant heme domains. Individual conformational states (as represented by fatty acid from preceding structural data (8). Although muta- crystal structures) are depicted as gray-shaded rectangles when largely unpopulated and color rectangles (color-coded to match A) when significantly tions at Ala-82 and Phe-87 were not present in the 139-3 populated. The y axis indicates the relative Tm values for the unfolding of mutant, the addition of an A82L mutation to 139-3 resulted in these proteins. increased coupling of NADPH oxidation to the hydroxylation of propane and octane, indicating improvements to the produc- the DM (59.9 °C). A similar effect was seen for F87V heme tive binding of these substrates. A related engineered BM3 var- ϭ domain (Tm 63.1 °C) (supplemental Table S1). Thus, A82F iant (named 9/10A and containing 13 mutations, 8 of which are has the major effect on stability of BM3, but F87V also destabi- shared with 139-3) was inactive in ethane hydroxylation but lizes the protein, and there is an additive effect on combining was shown to develop ethane hydroxylase activity and to the mutations (Fig. 10 and supplemental Fig. S5 and Table S1). improve propane hydroxylation when an A82S mutation was incorporated, together with either two or four further muta- DISCUSSION tions (46). A related propane hydroxylase variant (35E11) con- The development of efficient biocatalysts that generate taining 16 mutations (mostly accumulated from the 139-3 and human drug metabolites has become an area of great interest. 9/10A progenitors) also has the A82S mutation, although pro- Human P450s are responsible for most phase I xenobiotic pane hydroxylation rates and coupling were improved on intro- metabolism, producing numerous oxidized and other metabo- duction of further groups of mutations that included L188P lites from human drugs (37). For omeprazole, the principal (47). The later structural data for the 139-3 mutant heme metabolic pathway is 5-hydroxylation catalyzed by CYP2C19, domain in complex with NPG provided insights into structural the product of which can then be converted to the sulfone by changes that could promote binding and oxidation of short CYP3A4. Oxidized metabolites of human drugs are required for chain alkanes, while retaining a conformational state (SB) sim- pharmaceutical compliance, for metabolite safety testing, and ilar to the WT-NPG complex structure (34). In particular, for molecular interaction studies as required by Food and Drug active site mutation V78A enlarged a hydrophobic pocket close Administration guidelines. Use of human P450s to produce to the heme, whereas A184V enabled interactions with Leu-437 small quantities of specific metabolites is a possibility, but cave- across the active site channel, likely enabling van der Waals ats include slow reaction rates, requirement for a separate contacts with the terminal carbon of octane and protecting the redox partner, and enzyme instability. Synthetic chemistry to substrate from solvent (48). Predictive modeling based on the make oxidized metabolites is an alternative, but controlling 139-3 structure indicated that A82S and other mutations regioselectivity of oxidation, maintaining stability of com- altered substrate channel organization but that the destabiliz- pounds, and requirements for several steps giving low yield are ing L188P mutation was a major determinant in promoting

25396 JOURNAL OF BIOLOGICALDownloaded CHEMISTRY from http://www.jbc.org/ at The University of Manchester Library onVOLUME September 288•NUMBER 2, 2013 35•AUGUST 30, 2013 Diversification of P450 BM3 Substrate Selectivity

efficient propane oxidation, and likely acted by inducing struc- studies on Phe-87 mutants have further highlighted its impor- tural disruption at the end of the F-helix, favoring a conforma- tance in controlling access to the heme center and in altering tional change toward a catalytically productive form (48). Given reactivity. Notable recent studies have highlighted that an F87V the conservative nature of the A82S mutation, it is unclear BM3 mutant catalyzes hydroxylation of both testosterone and whether this mutation alone might favor the SB conformation progesterone. In the case of testosterone, 2␤-hydroxytestoster- (as does the A82F mutation in our work). However, the SB one and 16␤-hydroxytestosterone were formed in roughly structural conformation might be expected in any case for the equal amounts by F87V BM3. However, triple mutants that also NPG-bound enzyme. contained the A82F mutation (F87V, A82F, and V78L/V78T/ The fact that the structurally destabilizing L188P mutation V78I) showed 3–4-fold greater catalytic efficiency and pro- has profound effects on BM3 binding/turnover of propane has duced the 16␤-hydroxytestosterone product at ϳ90%, suggest- parallels in recent studies by Wong and co-workers, in which (i) ing an important influence of the A82F mutation on steroid a substrate-free I401P mutant (the mutated residue being adja- binding mode (12). Other work has pointed to the importance cent to the heme proximal ligand Cys-400) was crystallized in a of Phe-87 in tolerance to solvents used for substrate solubiliza- SB-like conformation, and (ii) an A330P mutant in an inter- tion. Crystal structure data showed that DMSO was able to helical ␤-sheet part of the P450 caused structural disruption compete with the heme’s 6th ligand water molecule for coordi- that reshaped the active site cavity. For both these mutants, nation to the iron in the F87A heme domain mutant but not in large increases in catalytic activity with non-natural substrates the WT heme domain at the same DMSO concentration (53). such as toluene, propylbenzene, and 3-methylpentane were Although the influence of Phe-87 mutations on accessibility observed (18, 49). to the heme center is easily rationalized from crystal structure Although some structural data are available for BM3 variants data, the mechanism by which Ala-82 mutations diversify BM3 promoting short chain/alkane binding and oxidation, there is substrate selectivity has been relatively poorly understood. also much interest in the application of BM3 mutants for the Huang et al. (20) solved the crystal structure of a palmitate- production of metabolites of human drugs. BM3 variants have bound A82F heme domain, after purifying the mutant from been shown to oxidize drugs such as diclofenac, ibuprofen, and E. coli in the lipid-bound form. The structure was similar to that acetaminophen, but no structural data were obtained (50, 51). of the NPG-bound WT heme domain (34). Our studies of the In this study, we present the first structure of a BM3 P450 substrate-free A82F structure reveal that the mutation induces bound to a human drug substrate (OMP), and we show that a a change in the protein conformational landscape, leading to a single active site mutation (A82F) is sufficient to facilitate tight distinct and previously unobserved position of the FG-helices. binding of this PPI drug, enabling the same oxidative reaction These structural changes have profound consequences for sub- (hydroxylation at the pyridinyl 5-methyl group) as catalyzed by strate binding in the A82F mutants and point to a role for this the main human metabolizing P450 CYP2C19. OMP binding is residue as a gatekeeper for preferred substrate access as a con- further enhanced by introduction of the F87V mutation in the sequence of its regulation of the conformational state and/or immediate vicinity of the heme, avoiding close contacts dynamics of structural change in the P450. Thus, although the between Phe-87 and the OMP. Despite the obvious differences mutation is removed from the immediate vicinity of the heme between the natural fatty acid substrates and OMP, the confor- iron, it induces structural reconfiguration that enables the mation of the OMP-bound mutants is nearly identical to previ- binding of OMP and evidently that of other molecules. Our ously determined WT fatty acid complexes. This suggests BM3 DSC studies show that the A82F mutation (in particular) alters active site architecture is largely determined by the overall pro- the thermodynamic stability of BM3, decreasing the Tm for protein conformation (SB versus SF) rather than the exact nature of tein unfolding. This suggests that its new conformational “flex- the ligand. Studies of human urinary metabolites of OMP ibility” is a key facet underpinning its diversification of sub- showed that the 5-OH OMP and 5-COOH OMP were the strate selectivity, at least in part due to removal of inherent major derivatives detected, with the former predominant (35, substrate bias toward long chain fatty acids by decreasing the 52). The relatively tight binding of the primary metabolite free energy barrier corresponding to the transition to the SB (5-OH OMP) to both A82F and (particularly) the DM BM3 state (Fig. 10B). These data are consistent with conclusions reinforces that the oxidized derivative retains the ability to bind drawn by Bloom et al. (54) with respect to the capacity of pro- in a catalytically competent mode to A82F-containing mutants, tein-destabilizing mutations to impart novel substrate selectiv- consistent with its further oxidation to 5-COOH OMP by the ity/reactivity on an enzyme. DM BM3 (Fig. 4 and Table 2). The BM3 DM’s ability to catalyze In conclusion, our data point to new lessons to be learned successive OMP oxidations to generate 5-COOH OMP points from the outcomes of previous studies of BM3 (and other to the possibility that CYP2C19 and/or other human cyto- enzymes) in that amino acid changes that affect the dynamics of chrome P450s might further oxidize 5-OH OMP to 5-COOH proteins in unpredictable ways may lead to mutations “distant” OMP. from the active site (or otherwise appearing not to impact sig- Early structural studies of BM3’s heme domain identified nificantly on substrate binding/affinity) having profound Phe-87 as a key active site residue, mutations of which altered effects on enzyme catalysis and substrate specificity. The gate- binding and regioselectivity of fatty acid oxidation (6, 10, 17), keeper hypothesis thus points to Ala-82 as a crucial target res- and both Phe-87 and Ala-82 mutations are frequently con- idue in research to enable further engineering of BM3 for tained in BM3 mutants generated by directed evolution and diverse functions. Our findings show that, contrary to previous other strategies for altered substrate specificity (13, 46). Other approaches focusing on increasing enzyme stability as a path to

AUGUST 30, 2013•VOLUMEDownloaded 288•NUMBER from 35http://www.jbc.org/ at The University of Manchester LibraryJOURNAL on September OF BIOLOGICAL 2, 2013 CHEMISTRY 25397 Diversification of P450 BM3 Substrate Selectivity biotechnologically relevant enzymes (55, 56), enzyme confor- 18. Whitehouse, C. J., Yang, W., Yorke, J. A., Rowlatt, B. C., Strong, A. J., mational destabilization is key to reducing the thermodynamic Blanford, C. F., Bell, S. G., Bartlam, M., Wong, L. L., and Rao, Z. (2010) barrier to substrate binding and therefore to altered enzymatic Structural basis for the properties of two single-site proline mutants of CYP102A1 (P450BM3). ChemBioChem 11, 2549–2556 activities, which could enable rapid identification of P450 (and 19. Joyce, M. G., Girvan, H. M., Munro, A. W., and Leys, D. (2004) A single other enzyme) variants with biotechnologically important mutation in cytochrome P450 BM3 induces the conformational rear- activities. rangement seen upon substrate binding in the wild-type enzyme. J. Biol. Chem. 279, 23287–23293 Acknowledgments—We acknowledge the following University of 20. Huang, W.-C., Westlake, A. C., Maréchal, J. D., Joyce, M. G., Moody, P. C., Manchester staff: Dr. Colin Levy for assistance with synchrotron x-ray and Roberts, G. C. (2007) Filling a hole in cytochrome P450 BM3 improves substrate binding and catalytic efficiency. J. Mol. Biol. 373, 633–651 data collection, Dr. Tom Jowitt for assistance with DSC studies, and ˇ 21. Ibeanu, G. C., Ghanayem, B. I., Linko, P., Li, L., Pederson, L. G., and Dr. Robert Sardzík for helpful discussions on NMR analysis. Goldstein, J. A. (1996) Identification of residues 99, 220, and 221 of human cytochrome P450 2C19 as key determinants of omeprazole activity. J. Biol. Chem. 271, 12496–12501 REFERENCES 22. Omura, T., and Sato, R. (1964) The carbon monoxide-binding pigment of 1. Guengerich, F. P. (2005) in Cytochrome P450: Structure, Mechanism and liver microsomes. II. Solubilization, purification, and properties. J. Biol. Biochemistry (Ortiz de Montellano, P. R., ed) 3rd Ed., pp. 377–530, Kluwer Chem. 239, 2379–2385 Academic/Plenum Press, New York 23. Perera, R., Sono, M., Sigman, J. A., Pfister, T. D., Lu, Y., and Dawson, J. H. 2. Munro, A. W., Girvan, H. M., and McLean, K. J. (2007) Variations on a (2003) Neutral thiol as a proximal ligand to ferrous heme iron: implica- (t)heme–novel mechanisms, redox partners, and catalytic functions in the tions for heme proteins that lose cysteine thiolate on reduction. Proc. Natl. cytochrome P450 superfamily. Nat. Prod. Rep. 24, 585–609 Acad. Sci. U.S.A. 100, 3641–3646 3. Xu, F., Bell, S. G., Rao, Z., and Wong, L. L. (2007) Structure-activity cor- 24. Dunford, A. J., McLean, K. J., Sabri, M., Seward, H. E., Heyes, D. J., Scrut- relations in pentachlorobenzene oxidation by engineered cytochrome ton, N. S., and Munro, A. W. (2007) Rapid P450 heme iron by laser pho- P450cam. Protein Eng. Des. Sel. 20, 473–480 toexcitation of Mycobacterium tuberculosis CYP121 and CYP51B1. Anal- 4. Peters, M. W., Meinhold, P., Glieder, A., and Arnold, F. H. (2003) Regio- ysis of CO complexation reactions and reversibility of the P450/P420 and enantioselective alkane hydroxylation with engineered cytochromes equilibrium. J. Biol. Chem. 282, 24816–24824 P450 BM-3. J. Am. Chem. Soc. 125, 13442–13450 25. Driscoll, M. D., McLean, K. J., Levy, C., Mast, N., Pikuleva, I. A., Lafite, P., 5. Munro, A. W., Daff, S., Coggins, J. R., Lindsay, J. G., and Chapman, S. K. Rigby, S. E., Leys, D., and Munro, A. W. (2010) Structural and biochemical (1996) Probing electron transfer in flavocytochrome P-450 BM3 and its characterization of Mycobacterium tuberculosis CYP142: Evidence for component domains. Eur. J. Biochem. 239, 403–409 multiple cholesterol 27-hydroxylase activities in a human pathogen. 6. Noble, M. A., Miles, C. S., Chapman, S. K., Lysek, D. A., MacKay, A. C., J. Biol. Chem. 285, 38270–38382 Reid, G. A., Hanzlik, R. P., and Munro, A. W. (1999) Roles of key active-site 26. Morrison, J. F. (1969) Kinetics of the reversible inhibition of enzyme- residues in flavocytochrome P450 BM3. Biochem. J. 339, 371–379 catalysed reactions by tight-binding inhibitors. Biochim. Biophys. Acta 7. Ravichandran, K. G., Boddupalli, S. S., Hasermann, C. A., Peterson, J. A., 185, 269–286 and Deisenhofer, J. (1993) Crystal structure of hemoprotein domain of 27. Bui, S. H., McLean, K. J., Cheesman, M. R., Bradley, J. M., Rigby, S. E., Levy, P450BM-3, a prototype for microsomal P450s. Science 261, 731–736 C. W., Leys, D., and Munro, A. W. (2012) Unusual spectroscopic and 8. Li, H., and Poulos, T. L. (1997) The structure of the cytochrome ligand binding properties of the cytochrome P450-flavodoxin fusion en- P450BM-3 haem domain complexed with the fatty acid substrate, palmit- zyme XplA. J. Biol. Chem. 287, 19699–19714 oleic acid. Nat. Struct. Biol. 4, 140–146 28. Kabsch, W. (1993) Automatic processing of rotation diffraction data from 9. Munro, A. W., Leys, D. G., McLean, K. J., Marshall, K. R., Ost, T. W., Daff, crystals of initially unknown symmetry and cell constants. J. Appl. Crys- S., Miles, C. S., Chapman, S. K., Lysek, D. A., Moser, C. C., Page, C. C., and tallogr. 26, 795–800 Dutton, P. L. (2002) P450 BM3: The very model of a modern flavocyto- 29. Collaborative Computation Project No. 4 (1994) The CCP4 suite: pro- chrome. Trends Biochem. Sci. 27, 250–257 grams for protein crystallography. Acta Crystallogr. D Biol. Crystallogr.

10. Whitehouse, C. J., Bell, S. G., and Wong, L. L. (2012) P450BM3 50, 760–763 (CYP102A1): connecting the dots. Chem. Soc. Rev. 41, 1218–1260 30. Emsley, P., and Cowtan, K. (2004) Coot: model-building tools for molec- 11. Coelho, P. S., Brustad, E. M., Kannan, A., and Arnold, F. H. (2013) Olefin ular graphics. Acta Crystallogr. D Biol. Crystallogr. 60, 2126–2132 cyclopropanation via carbene transfer catalyzed by engineered cyto- 31. Venkataraman, H., Beer, S. B., Bergen, L. A., Essen Nv, Geerke, D. P., chrome P450 enzymes. Science 339, 307–310 Vermeulen, N. P., and Commandeur, J. N. (2012) A single active site mu- 12. Kille, S., Zilly, F. E., Acevedo, J. P., and Reetz, M. T. (2011) Regio- and tation inverts stereoselectivity of 16-hydroxylation of testosterone cata- stereoselectivity of P450-catalysed hydroxylation of steroids controlled by lyzed by engineered cytochrome P450 BM3. ChemBioChem 13, 520–523 laboratory evolution. Nat. Chem. 3, 738–743 32. Miles, J. S., Munro, A. W., Rospendowski, B. N., Smith, W. E., McKnight, 13. Carmichael, A. B., and Wong, L. L. (2001) Protein engineering of Bacillus J., and Thomson, A. J. (1992) Domains of the catalytically self-sufficient megaterium CYP102. The oxidation of polycyclic aromatic hydrocarbons. cytochrome P-450 BM-3. Genetic construction, overexpression, purifica- Eur. J. Biochem. 268, 3117–3125 tion and spectroscopic characterization. Biochem. J. 288, 503–509 14. Sawayama, A. M., Chen, M. M., Kulanthaivel, P., Kuo, M. S., Hemmerle, 33. Desta, Z., Zhao, X., Shin, J. G., and Flockhart, D. A. (2002) Clinical signif- H., and Arnold, F. H. (2009) A panel of cytochrome P450 BM3 variants to icance of the P4502C19 genetic polymorphism. Clin. Pharmacokinet. 41, produce drug metabolites and diversify lead compounds. Chemistry 15, 913–958 11723–11729 34. Haines, D. C., Tomchick, D. R., Machius, M., and Peterson, J. A. (2001) 15. Reetz, M. T. (2011) Laboratory evolution of stereoselective enzymes: a Pivotal role of water in the mechanism of P450BM-3. Biochemistry 40, prolific source of catalysts for asymmetric reactions. Angew. Chem. Int. Ed. 13456–13465 Engl. 50, 138–174 35. Renberg, L., Simonsson, R., and Hoffmann, K.-J. (1989) Identification of 16. Ost, T. W., Miles, C. S., Murdoch, J., Cheung, Y., Reid, G. A., Chapman, two main urinary metabolites of [14C]omeprazole in humans. Drug S. K., and Munro, A. W. (2000) Rational re-design of the substrate binding Metab. Dispos. 17, 69–76 site of flavocytochrome P450 BM3. FEBS Lett. 486, 173–177 36. Girvan, H. M., Seward, H. E., Toogood, H. S., Cheesman, M. R., Leys, D., 17. Dietrich, M., Do, T. A., Schmid, R. D., Pleiss, J., and Urlacher, V. B. (2009) and Munro, A. W. (2007) Structural and spectroscopic characterization of Altering the regioselectivity of the subterminal fatty acid hydroxylase P450 P450 BM3 mutants with unprecedented P450 heme iron ligand sets. New BM-3 toward␥ and ␦ positions. J. Biotechnol. 139, 115–117 heme ligation states influence conformational equilibria in P450 BM3.

25398 JOURNAL OF BIOLOGICALDownloaded CHEMISTRY from http://www.jbc.org/ at The University of Manchester Library onVOLUME September 288•NUMBER 2, 2013 35•AUGUST 30, 2013 Diversification of P450 BM3 Substrate Selectivity

J. Biol. Chem. 282, 564–572 48. Fasan, R., Meharenna, Y. T., Snow, C. D., Poulos, T. L., and Arnold, F. H. 37. Guengerich, F. P. (2008) Cytochrome P450 and chemical toxicology. (2008) evolutionary history of a specialized P450 propane monooxyge- Chem. Res. Toxicol. 21, 70–83 nase. J. Mol. Biol. 383, 1069–1080 38. Li, H., and Poulos, T. L. (1999) Fatty acid metabolism, conformational 49. Whitehouse, C. J., Bell, S. G., Yang, W., Yorke, J. A., Blanford, C. F., Strong, change, and electron transfer in cytochrome P450 BM-3. Biochim. Bio- A. J., Morse, E. J., Bartlam, M., Rao, Z., and Wong, L.-L. (2009) A highly phys. Acta 1441, 141–149 active single-mutation variant of P450BM3 (CYP102A1). ChemBioChem 39. Ost, T. W., Miles, C. S., Munro, A. W., Murdoch, J., Reid, G. A., and 10, 1654–1656 Chapman, S. K. (2001) Phenylalanine 393 exerts thermodynamic control 50. Tsotsou, G. E., Sideri, A., Goyal, A., Di Nardo, G., and Gilardi, G. (2012) over the heme of flavocytochrome P450 BM3. Biochemistry 40, Identification of mutant Asp251Gly/Gln307His of cytochrome P450 BM3 13421–13429 for the generation of metabolites of diclofenac, ibuprofen, and tolbuta- 40. Yeom, H., Sligar, S. G., Li, H., Poulos, T. L., and Fulco, A. J. (1995) The role mide. Chemistry 18, 3582–3588 of Thr268 in oxygen activation of cytochrome P450 BM-3. Biochemistry 51. Damsten, M. C., van Vugt-Lussenburg, B. M., Zeldenthuis, T., de Vlieger, 34, 14733–14740 J. S., Commandeur, J. N., and Vermeulen, N. P. (2008) Application of drug 41. Narhi, L. O., and Fulco, A. J. (1986) Characterization of a catalytically metabolising mutants of cytochrome P450 BM3 (CYP102A1) as biocata- self-sufficient 119,000-dalton cytochrome P-450 monooxygenase induced lysts for the generation of reactive metabolites. Chem. Biol. Interact. 171, by barbiturates in Bacillus megaterium. J. Biol. Chem. 261, 7160–7169 96–107 42. Neeli, R., Girvan, H. M., Lawrence, A., Warren, M. J., Leys, D., Scrutton, 52. Abelö, A., Andersson, T. B., Antonsson, M., Naudot, A. K., Skånberg, I., N. S., and Munro, A. W. (2005) The dimeric form of flavocytochrome and Weidolf, L. (2000) Stereoselective metabolism of omeprazole by hu- P450 BM3 is catalytically functional as a fatty acid hydroxylase. FEBS Lett. man cytochrome P450 enzymes. Drug Metab. Dispos. 28, 966–972 579, 5582–5588 53. Kuper, J., Tee, K. L., Wilmanns, M., Roccatano, D., Schwaneberg, U., and 43. Kitazume T., Haines, D. C., Estabrook, R. W., Chen, B., and Peterson, J. A. Wong, T. S. (2012) The role of active-site Phe-87 in modulating the or- (2007) Obligatory intramolecular electron-transfer from FAD to FMN in ganic co-solvent tolerance of cytochrome P450 BM3 monooxygenase. dimeric P450BM-3. Biochemistry 46, 11892–11901 Acta Crystallogr. Sect. F Struct. Biol. Cryst. Commun. 68, 1013–1017 44. Girvan, H. M., Dunford, A. J., Neeli, R., Ekanem, I. S., Waltham, T. N., 54. Bloom, J. D., Labthavikul, S. T., Otey, C. R., and Arnold, F. H. (2006) Joyce, M. G., Leys, D., Curtis, R. A., Williams, P., Fisher, K., Voice, M. W., Protein stability promotes evolvability. Proc. Natl. Acad. Sci. U.S.A. 103, and Munro, A. W. (2011) Flavocytochrome P450 BM3 mutant W1046A is 5869–5874 a NADH-dependent fatty acid hydroxylase: implications for the mecha- 55. Romero, P. A., Krause, A., and Arnold, F. H. (2013) Navigating the protein nism of electron transfer in the P450 BM3 dimer. Arch. Biochem. Biophys. fitness landscape with Gaussian processes. Proc. Natl. Acad. Sci. U.S.A. 507, 75–85 110, E193–E201 45. Glieder, A., Farinas, E. T., and Arnold, F. H. (2002) Laboratory evolution of 56. Li, Y., Drummond, D. A., Sawayama, A. M., Snow, C. D., and Bloom, J. D. a soluble, self-sufficient, highly active alkane hydroxylase. Nat. Biotechnol. (2007) A diverse family of thermostable cytochrome P450s created by 20, 1135–1139 recombination of stabilizing fragments. Nat. Biotechnol. 25, 1051–1056 46. Meinhold, P., Peters, M. W., Chen, M. M., Takahashi, K., and Arnold, F. H. 57. Sevrioukova, I. F., Li, H., Zhang, H., Peterson, J. A., and Poulos, T. L. (1999) (2005) Direct conversion of ethane to ethanol by engineered cytochrome Structure of a cytochrome P450 redox partner electron transfer complex. P450 BM3. ChemBioChem 6, 1765–1768 Proc. Natl. Acad. Sci. U.S.A. 96, 1863–1868 47. Fasan, R., Chen, M. M., Crook, N. C., and Arnold, F. H. (2007) Engineered 58. Laskowski R. A., and Swindells M. B. (2011) Ligplotϩ: multiple ligand- alkane-hydroxylating cytochrome P450BM3 exhibiting native-like cata- protein interaction diagrams for drug discovery. J. Chem. Inf. Model. 51, lytic properties. Angew. Chem. Int. Ed. Engl. 46, 8414–8418 2778–2786

AUGUST 30, 2013•VOLUMEDownloaded 288•NUMBER from 35http://www.jbc.org/ at The University of Manchester LibraryJOURNAL on September OF BIOLOGICAL 2, 2013 CHEMISTRY 25399 Diversification of P450 BM3 substrate selectivity

SUPPLEMENTAL DATA

Key mutations alter the cytochrome P450 BM3 conformational landscape and remove inherent substrate bias

1Christopher F. Butler, 2Caroline Peet, 1Amy E. Mason, 2Michael W. Voice, 1David Leys and 1Andrew W. Munro*

1Manchester Institute of Biotechnology, Faculty of Life Sciences, University of Manchester, 131 Princess Street, Manchester M1 7DN, UK. 2Cypex Ltd, 6 Tom McDonald Avenue, Dundee DD2 1NH, UK.

Running title: Diversification of P450 BM3 substrate selectivity

*To whom correspondence should be addressed: A. W. Munro Tel: +44 161 3065151; Fax: +44 161 3068918; E-mail: [email protected]

Supplemental Results

Characterization of substrate and oxidized products by NMR spectroscopy Nuclear Magnetic Resonance (NMR) Spectroscopy was used for the analysis of the OMP substrate (Figures S1 and S2) and the main monohydroxylated products from OMP turnover. Confirmation of the position of oxidation of OMP at the 5-methyl group was obtained from a combination of 1H, 13C and 2D NMR (Figures S3 and S4).

1H, 13C, HMQC (Heteronuclear Multiple Quantum Correlation) and HMBC (Heteronuclear Multiple Bond Correlation) experiments were used to characterize the starting materials and products. The 1H NMR omeprazole spectrum is shown below in Figure S1 and was assigned as: (CDCl3 400MHz) δ = 8.15 (S, 1H), 7.45 (Broad S, 1H), 7.19 (Broad S, 1H), 6.88 (dd, J = 8.84, 1.53 Hz, 1H), 4.66 (AB, J = 13.64, 13 8.72 Hz, 2H), 3.78 (S, 3H), 3.57 (S, 3H), 2.16 (S, 3H), 2.08 (S, 3H). C NMR (CDCl3 101MHz) δ = + 164.5, 149.7, 148.7, 127.1, 126.3, 60.6, 59.8, 55.7, 13.3, 11.5. LCMS [M+H] 346.1213 (C17H20N3O3S). Assignment of the omeprazole spectrum was made using the HMBC spectrum (Figure S2). Upon enzymatic turnover with the BM3 F87V and DM enzymes (less products were formed with the A82F mutant, although LC-MS clearly showed formation of the 5-OH OMP product at the same retention time), a mixture of starting material and monohydroxylated product was identified. The product was identified 1 as 5-OH OMP and the additional peaks seen in Figure S3 were assigned: H NMR (CDCl3 400MHz) δ = 8.27 (S, 1H), 7.45 (Broad S, 1H), 6.88 (d, J = 2.40 Hz, 1H), 6.85 (dd, J = 8.84, 2.27 Hz, 1H), 4.65 (AB, J 13 = 3.92, 1.26 Hz, 2H), 4.59 (S, 2H), 3.75 (S, 3H), 3.56 (S, 3H), 2.02 (S, 3H). C NMR (CDCl3 101MHz) δ + = 164.2, 150.4, 148.2, 129.3, 127.2, 61.2, 60.2, 58.6, 11.5. LCMS [M+H] 362.1162 (C17H20N3O4S). Assignment of the 5-OH OMP spectrum was again made using the HMBC spectrum (Figure S4).

1 Downloaded from http://www.jbc.org/ at The University of Manchester Library on September 2, 2013 Diversification of P450 BM3 substrate selectivity

Figure S1. 1H NMR spectrum of omeprazole. The 1H spectrum for the OMP starting material is shown with peaks labelled and integrated. The overlay shows the omeprazole structure with accepted (non IUPAC) numbering, with peaks labelled accordingly. Data were collected on a 400 MHz NMR in CDCl3, corrected to TMS by residual non-deuterated solvent.

2 Downloaded from http://www.jbc.org/ at The University of Manchester Library on September 2, 2013 Diversification of P450 BM3 substrate selectivity

Figure S2. HMBC spectrum of omeprazole. The spectrum shows the long range coupling of 1H to 13C nuclei. Coupling is observed between the 3-methyl group and the AB system; and between the 5-methyl group and the pyridinyl methoxy and pyridinyl hydrogen.

3 Downloaded from http://www.jbc.org/ at The University of Manchester Library on September 2, 2013 Diversification of P450 BM3 substrate selectivity

Figure S3. 1H NMR spectrum of turnover products from omeprazole oxidation. Products were generated in the reaction of the F87V BM3 with OMP substrate. The overlay shows the 5-OH omeprazole structure with accepted (non IUPAC) numbering, and with peaks labelled accordingly. Only the product peaks are labelled and integrated for clarity. The spectrum shows the generation of a methoxy peak at δ 4.59 and a new methyl peak at δ 2.02, indicative of hydroxylation at one of the OMP methyl groups. The downfield shift of the pyridinyl hydrogen signal and the lack of shift in the AB system indicates hydroxylation at the 5 position. Data were collected on a 400 MHz NMR in CDCl3, corrected to TMS by residual non-deuterated solvent.

4 Downloaded from http://www.jbc.org/ at The University of Manchester Library on September 2, 2013 Diversification of P450 BM3 substrate selectivity

Figure S4. HMBC spectra of turnover products from omeprazole oxidation. Products were generated in the reaction of the F87V BM3 with OMP substrate. The data show the long range coupling of 1H to 13C nuclei. Additional peaks generated (by comparison with panel B above) show coupling of the new methoxy peak (δ 4.59) to the pyridinyl methoxy (δ 3.54), the pyridinyl hydrogen (δ 8.27), and to the new methyl peak (δ 2.02). Lack of coupling to the AB system confirms that hydroxylation occurs on the 5- methyl (and not the 3-methyl) group.

5 Downloaded from http://www.jbc.org/ at The University of Manchester Library on September 2, 2013 Diversification of P450 BM3 substrate selectivity

Differential Scanning Calorimetry (DSC). DSC was carried out as detailed in the Experimental Procedures section of the main paper. Data analysis was carried out using Microcal Origin software. Prior to data analysis, a reference baseline trace (using either the buffer alone for substrate-free samples, or buffer plus substrate at the working concentration for substrate-bound samples) was subtracted from the relevant BM3 heme domain data set, and then the data were normalized for protein concentration. Fitting was carried out using a standard non 2-state function. Transition midpoint (Tm), calorimetric enthalpy (∆Hcal) and Van’t Hoff enthalpy (∆HVH) data were collected and are tabulated below (Table S1). The WT BM3 and mutant heme domains either had a single unfolding transition, or a main unfolding transition preceded by a minor transition at lower temperature. For comparison purposes, the Tm of the main unfolding event is used in all cases. The Tm value for the substrate (NPG)-bound F87V heme domain was determined from an incomplete data set, as exothermic aggregation took place prior to the complete unfolding of the protein in this case. The data show that WT BM3 heme domain is extensively stabilized by NPG, though omeprazole has little effect. The A82F heme domain is stabilized a little by NPG, but omeprazole also has little effect. The F87V heme domain is stabilized a little by omeprazole, but more by NPG. The DM heme domain is stabilized to an almost equal extent by both substrates. All DSC data sets are overlaid in Figure S5.

6 Downloaded from http://www.jbc.org/ at The University of Manchester Library on September 2, 2013 Diversification of P450 BM3 substrate selectivity

Protein ∆H1 (kcal ∆H 1 (kcal ∆H2 (kcal ∆H 2 (kcal T 1 (oC) VH T 2 (oC) VH (ligand) m mol-1) mol-1) m mol-1) mol-1) 65.73 1.17 E5 (1.61 1.77 E5 (2.73 58.95 5.12 E4 (1.69 1.41 E5 (4.93 WT (SF) (0.03) E3) E3) (0.08) E3) E3) 70.39 1.31 E5 (2.95 1.80 E5 (3.14 64.36 7.47 E4 (3.13 1.13 E5 (3.74 WT (NPG) (0.03) E3) E3) (0.14) E3) E3) 65.44 8.91 E4 (4.06 1.88 E5 (8.04 59.28 3.76 E4 (4.37 1.24 E5 (1.37 WT (OMP) (0.07) E3) E3) (0.34) E3) E4) 59.66 1.68 E5 (1.28 1.72 E5 (1.55 49.72 3.07 E4 (1.59 9.53 E4 (6.16 A82F (SF) (0.02) E3 E3) (0.19) E3) E3) A82F 62.26 1.89 E5 (3.11 1.68 E5 (3.46 ------(NPG) (0.04) E3) E3) A82F 59.58 1.07 E5 (8.26 1.93 E5 (1.87 ------(OMP) (0.02) E2) E3) 61.33 1.12 E5 (8.17 1.67 E5 (1.53 F87V (SF) ------(0.02) E2) E3) F87V 65.15 1.98 E5 (8.46 1.68 E5 (8.85 ------(NPG)* (0.10) E3) E3) F87V 63.10 8.8 E4 (8.48 1.55 E5 (1.87 ------(OMP) (0.03) E2) E3) 58.04 1.88 E5 (4.99 1.88E5 (5.83 50.86 1.12 E4 (5.58 1.31 E5 (7.68 DM (SF) (0.01) E2) E2) (0.13) E2) E3) 60.35 1.95 E5 (3.14 1.60 E5 (3.22 DM (NPG) ------(0.04) E3) E3) DM 59.86 1.40 E5 (1.19 1.96 E5 (2.08 ------(OMP) (0.02) E3) E3)

Table S1. DSC data for thermal unfolding of WT and mutant BM3 heme domains in ligand-free and substrate-bound forms. The thermal Transition midpoint (Tm), calorimetric enthalpy (∆H) and Van’t Hoff enthalpy (∆HVH) data for WT and mutant BM3 heme domains in substrate-free and OMP- and NPG-bound forms are shown. *Indicates that values presented were calculated using data up to the point at which exothermic aggregation was observed for the F87V-NPG complex (close to the end of its thermal unfolding transition). The data reveal thermal destabilization of the mutant forms by comparison to WT BM3 heme domain. The BM3 DM heme domain is stabilized by OMP to a similar extent as it is by NPG.

7 Downloaded from http://www.jbc.org/ at The University of Manchester Library on September 2, 2013 Diversification of P450 BM3 substrate selectivity

Figure S5. Graphical overlay of DSC data for WT and mutant BM3 heme domains. The thermal unfolding profiles are shown for the various substrate-free, NPG- and OMP-bound forms of WT and mutant (F87V, A82F and DM) BM3 heme domains. The relevant Tm values for the transitions observed (and accompanying thermodynamic data) are shown in Table S1.

8 Downloaded from http://www.jbc.org/ at The University of Manchester Library on September 2, 2013 View Article Online / Journal Homepage / Table of Contents for this issue Metallomics Dynamic Article Links

Cite this: Metallomics, 2011, 3, 369–378

www.rsc.org/metallomics PAPER

Analysis of the oxidation of short chain alkynes by ﬂavocytochrome P450 BM3wz

Timothy N. Waltham,a Hazel M. Girvan,a Christopher F. Butler,a Stuart R. Rigby,b Adrian J. Dunford,a Robert A. Holtb and Andrew W. Munro*a

Received 11th January 2011, Accepted 3rd March 2011 DOI: 10.1039/c1mt00004g

Bacillus megaterium flavocytochrome P450 BM3 (BM3) is a high activity fatty acid hydroxylase, formed by the fusion of soluble cytochrome P450 and cytochrome P450 reductase modules. Short chain (C6, C8) alkynes were shown to be substrates for BM3, with productive outcomes (i.e. alkyne hydroxylation) dependent on position of the carbon–carbon triple bond in the molecule. Wild-type P450 BM3 catalyses o-3 hydroxylation of both 1-hexyne and 1-octyne, but is suicidally inactivated in NADPH-dependent turnover with non-terminal alkynes. A F87G mutant of P450 BM3 also undergoes turnover-dependent heme destruction with the terminal alkynes, pointing to a key role for Phe87 in controlling regioselectivity of alkyne oxidation. The terminal alkynes access the BM3 heme active site led by the acetylene functional group, since hydroxylated products are not observed near the opposite end of the molecules. For both 1-hexyne and 1-octyne, the predominant enantiomeric product formed (up to B90%) is the (S)-()-1-alkyn-3-ol form. Wild-type P450 BM3 is shown to be an effective oxidase catalyst of terminal alkynes, with strict regioselectivity of oxidation and potential biotechnological applications. The absence of measurable octanoic or hexanoic acid products from oxidation of the relevant 1-alkynes is also consistent with previous studies suggesting that removal of the phenyl group in the F87G mutant does not lead to significant levels of o-oxidation of alkyl chain substrates.

Introduction iron ﬁrst to the ferrous form (allowing dioxygen binding) and then further reduces this oxy complex to a ferric peroxo The cytochromes P450 (P450s or CYPs) are a superfamily of

Published on 23 March 2011. Downloaded by The University of Manchester Library 02/09/2013 22:32:22. state. This is followed by protonation reactions that heme b-containing enzymes found throughout the kingdoms were proposed to generate a ferryl-oxo state (compound I) of life, and which reductively activate molecular oxygen bound as the highly reactive form that is ultimately responsible to their heme iron. This leads (for the vast majority of for oxygen transfer to the substrate.1,4 Only very recently P450-catalysed reactions) to the insertion of an atom of has the first compelling spectroscopic evidence been dioxygen into a substrate bound close to the heme iron, and presented for the transient formation of a catalytically the production of a molecule of water from the other oxygen active compound I species: in CYP119 from the thermophile 1–3 atom. Essential for the reactivity of the P450s is the Sulfolobus solfataricus.5 presence of a phylogenetically conserved cysteine residue Well known microbial P450-catalysed reactions include the that provides a thiolate ligand to the heme iron trans to the 5-exo hydroxylation of camphor by the Pseudomonas putida dioxygen. The resting form of a P450 has heme iron in CYP101A1 (P450cam, leading to camphor catabolism as an the ferric form. However, two successive single electron energy source), and the 3-step oxidation of the 14a-methyl transfers from a redox partner to the heme iron reduces the group of the fungal sterol lanosterol to generate the demethylated sterol ergosterol essential for membrane integrity.6,7 a Manchester Interdisciplinary Biocentre, Faculty of Life Sciences, The latter reaction is catalyzed by fungal CYP51 family University of Manchester, 131 Princess Street, Manchester M1 enzymes (and bacterial homologues) and is the target reaction 7DN, UK. E-mail: [email protected], for fluconazole and other azole class drugs.8,9 A recent review http://www.manchester.ac.uk/research/Andrew.Munro/; Tel: +44-141-306-5151 indicated that close to 20 000 P450s are found in the genomes b Piramal Healthcare, Biocatalysis Division, The Wilton Centre, of organisms from bacteria and archaea through to man, with Redcar, Cleveland TS10 4RF functions including antibiotic synthesis, steroid metabolism w This article is published as part of a themed issue on Cytochromes, and drug/xenobiotic detoxification. The majority of these have Guest Edited by Norbert Jakubowski and Peter Roos. z Electronic supplementary information (ESI) available. See DOI: been classified into defined families and sub-families within the 10.1039/c1mt00004g P450 enzyme superfamily.10,11

This journal is c The Royal Society of Chemistry 2011 Metallomics, 2011, 3, 369–378 369 View Article Online

eﬃcient electron transport system to the P450 heme iron.14 BM3 catalyses hydroxylation of several fatty acids of chain

length BC10–C20, and has the highest reported activity for a P450 monooxygenase (B285 s1 with arachidonic acid).15 BM3 typically hydroxylates lipid substrates at o-1, o-2 and o-3 positions, with the omega terminal carbon unmodified and likely protected by interactions of the terminal methyl group of the substrate with the side chain of phenylalanine 87 in the BM3 active site16 (Fig. 1). Although omega hydroxylation of fatty acids is energetically disfavoured (due to the greater bond dissociation energies of the primary methyl C–H bonds compared to those of adjacent secondary or tertiary C–H bonds), it appears that BM3 has adapted its structure to further decrease the likelihood of fatty acid o-hydroxylation through strategic positioning of the Phe87 side chain.17 The heme domain of P450 BM3 is related to various members of the CYP4 family of fatty acid hydroxylases in eukaryotes, many of Fig. 1 Active site structure in P450 BM3. The structure shows key which hydroxylate their lipid substrates preferentially at the 18 residues in the active site of the heme domain of WT P450 BM3 in omega methyl group (o-hydroxylation). While this is an complex with the C16 monounsaturated fatty acid palmitoleic acid energetically more challenging reaction for the P450s, (PDB code 1FAG).16 Key amino acids are shown in atom coloured structural control over regioselectivity by the relevant P450s sticks (carbons in light green), while the heme cofactor is shown in red is required for specific generation of metabolites with distinc- sticks, with its central iron atom shown in red spacefill. The substrate tive cellular functions. For instance, o-1 hydroxylation of palmitoleic acid is shown in stick presentation with carbons in cyan arachidonic acid produces a vasodilatory product, while and oxygen atoms in red. The palmitoleic acid carboxylate oxygens o-hydroxylation produces a vasoconstrictor.18,19 Interestingly, interact with the side chains of arginine 47 and tyrosine 51 near the several eukaryotic CYP4 enzymes have been shown to undergo mouth of the active site, and the fatty acid alkyl chain extends down the BM3 active site, with the terminal (omega) methyl group interacting turnover dependent covalent attachment of their heme macro- with the aromatic side chain of phenylalanine 87 above the plane of the cycle to the protein backbone through ester linkage of a heme heme. Cysteine 400 provides the proximal axial ligand to the heme iron methyl to the carboxylate group of an active site amino acid 20 and is seen below the plane of the heme. Distances (closest approaches) (usually a conserved glutamate). It is plausible that this between the substrate o-2 carbon and heme iron, and between the covalent linkage of the heme assists in promoting fatty acid substrate o-carbon and Phe87 are shown. o-hydroxylation over o-1 (or o-2) hydroxylation, and this would be consistent with data such as a migration from An important member of the P450 enzyme superfamily is o-hydroxylation towards o-1 and o-2 hydroxylation in flavocytochrome P450 BM3 (CYP102A1, BM3) from the soil E310A/G mutants of rabbit CYP4B1 that remove the relevant bacterium Bacillus megaterium. Most P450s source electrons glutamic acid residue.21,22 However, studies of the A264E

Published on 23 March 2011. Downloaded by The University of Manchester Library 02/09/2013 22:32:22. for the reductive activation of dioxygen from one or more mutant of BM3 (in which a glutamate was introduced at the NAD(P)H-dependent redox partner systems.4,12 P450 BM3 relevant position in the P450 structure) did not result in was the ﬁrst characterized example of a P450 enzyme fused to covalent heme attachment, but instead in the binding of the its redox partner, in this case a eukaryotic-like FAD- and E264 carboxylate as the 6th ligand to the heme iron.23,24 FMN-containing NADPH-cytochrome P450 reductase Due in the main to its catalytic eﬃciency and convenient, (CPR).13 For BM3, this fusion arrangement facilitates a highly single component nature, P450 BM3 has been widely used in

Fig. 2 Mechanism of BM3 heme alkylation following P450 compound I-mediated attack on an acetylene. The cysteinate ligated ferriprotoporphyrin IX (heme) group of P450 BM3 is represented as 4 interconnected nitrogens. Oxidative attack by the BM3 compound I [(FeQO)3+] at an internal acetylenic carbon leads to covalent bonding between the other acetylenic carbon and a pyrrole nitrogen. The enol initially formed by oxygen addition then tautomerises to a ketone. An alternative pathway exists when oxidative attack occurs at the external acetylenic carbon of a terminal alkyne. In such cases, migration of the terminal hydrogen to the vicinal carbon occurs simultaneously with oxygen transfer to the carbon. The ketene product formed may then either modify active site residues (amines, alcohols or cysteine thiol), or else be hydrolysed to a carboxylic acid.44

370 Metallomics, 2011, 3, 369–378 This journal is c The Royal Society of Chemistry 2011 View Article Online

protein engineering studies aimed at diversifying its substrate In this paper, we explore the activity of P450 BM3 with hexyne selectivity and enabling production of biotechnologically and octyne substrates with triple bonds at different positions, useful oxygenated molecules.25,26 A prominent example of examining catalytic rates, oxy-product formation and the influence success in this area was the use of directed evolution and then of the phenylalanine 87 residue on the catalytic outcome in a site-directed mutagenesis to produce BM3 variants with F87G mutant. The results point to the wild-type enzyme as being alkane hydroxylase activities. Arnold and co-workers reported a good catalyst for hydroxylation of terminal alkynes, with the production of several such mutants which catalysed potential applications in generation of synthetic intermediates. hydroxylation of various linear alkanes (from propane This activity is lost in the F87G mutant, with important ramifica- through to decane) at sub-terminal positions, and with tions for the understanding of the role of this amino acid in the activities substantially increased compared to the wild-type control of regioselectivity of substrate oxidation in P450 BM3. (WT) BM3.27,28 While WT and various mutants of BM3 are known to oxidize various alkanes and alkenes, to date the catalytic capacity of P450 BM3 to oxidize short chain alkyne Materials and methods substrates has not been examined. Alkynols have multiple uses Materials in organic synthesis, including in production of substituted alkenes, while 1-octyn-3-ol, for instance, is known to prevent Octyne and hexyne substrates (1-octyne, 2-octyne, 3-octyne, corrosion of iron and also to act as an attractant for 4-octyne, 1-hexyne, 2-hexyne and 3-hexyne, along with mosquitoes.29,30 However, studies by Ortiz de Montellano product standards (the (R)-(+)-1-octyn-3-ol and (S)-()-1- and co-workers have demonstrated that P450-mediated octyn-3-ol enantiomers, and the racemic 1-octyn-3-ol and oxidation at terminal carbon–carbon triple (acetylenic) bonds 1-hexyn-3-ol samples) were from Sigma Aldrich (Poole, leads to generation of reactive ketenes if oxidative attack UK). Bacterial growth media (tryptone, yeast extract) and is at the terminal end of the triple bond, and these may antibiotics for E. coli transformation selection (ampicillin) go on to covalently modify proximal nucleophilic amino acid were from Formedium (Hunstanton, UK). Unless otherwise side chains unless they are first hydrolysed to a carboxylic stated, all other chemicals used were from Sigma Aldrich, and acid. Moreover, covalent modification of the heme cofactor were of the highest purity available. itself (with destruction of the chromophore) readily occurs Expression and purification of wild-type and F87G P450 BM3 if P450-mediated oxygen addition is to the internal enzymes acetylenic carbon, with resultant alkylation of a heme pyrrole nitrogen by its attachment to the terminal acetylenic carbon The WT flavocytochrome P450 BM3 enzyme and its F87G (Fig. 2).31,32 mutant were expressed and purified as described in our Published on 23 March 2011. Downloaded by The University of Manchester Library 02/09/2013 22:32:22.

Fig. 3 Alkynes substrates/inhibitors and reaction products from studies with WT and F87G P450 BM3 enzymes. Structures A-D are 1-, 2-, 3-, and 4-octyne, respectively. Structures E-G are 1-octyn-3-ol, 1-octyn-4-ol and 1-octyn-5-ol products, respectively, from WT BM3-mediated oxidation of 1-octyne. Structures H-J are 1-, 2-, and 3-hexyne, respectively. Structure K is 1-hexyn-3-ol product from WT BM3-mediated oxidation of 1-hexyne.

This journal is c The Royal Society of Chemistry 2011 Metallomics, 2011, 3, 369–378 371 View Article Online

previous studies.23,33 Details are provided in the Supplementary and glucose-6-phosphate (4 mM) were added to recycle Data section.z Final protein purity was verified by SDS-PAGE, NADPH for BM3 function. Reactions were done overnight and protein concentration calculated by the method of at 25 1C in tinted glass vials. Products and unconverted Omura and Sato, based on absorption difference of the BM3 alkyne/alkane substrates were extracted into methyl tertiary ferrous-CO complex by comparison to that of its ferrous form butyl ether (MTBE) solvent. Approximately 5 g of sodium 1 1 34 (De450–490 = 91 000 M cm ). Both flavocytochromes were sulfate were added to each reaction vial (as drying agent) and replete with flavin cofactors, but the heme content of the F87G then the mixture was vortexed extensively. After settling, a mutant flavocytochrome was consistently in the 50–80% range small amount of the MTBE layer was transferred to a separate compared to WT BM3. The WT P450 BM3 heme domain was glass vial (pre-washed in the solvent) for further characteriza- purified similarly, and as described previously.23,33 tion using GC-MS. An Agilent 7890 GC and Agilent 5975C Inert XL MSD with Spectroscopic analysis of BM3 enzymes and their interactions triple axis detector were used in all GC-MS studies. 1 ml with alkynes samples of standards or extracted products were injected onto All UV-visible spectroscopic measurements and kinetic studies a capillary column (model Agilent DB-5MS) with the oven 1 were carried out on a Cary UV-50 spectrophotometer starting temperature at 40 C. A temperature gradient was run 1 (Varian, UK) using a 1 cm pathlength quartz cuvette. Analysis over 44.5 min (with the temperature held at 40 C for the first 1 of spectral properties of the P450 BM3 enzymes in their 6 min then increased 10 C per minute) with helium gas to a 1 oxidized, NADPH- and sodium dithionite-reduced forms, final temperature of 325 C. Standards of 1-octyn-3-ol and and in their reduced carbon/monoxide bound complexes were 1-hexyn-3-ol (Sigma-Aldrich) were also diluted in MTBE done as described previously.15,33 The effects of incubations of and run on the GC-MS to compare their retention time WT and F87G BM3 enzymes with alkynes (see Fig. 3 for and chromatograms with the products generated from alkynes used) on the UV-visible absorption spectrum of the BM3-mediated turnover of 1-octyne and 1-hexyne. For those proteins was followed at 25 1C in assay buffer (100 mM alkyne reactions with BM3 enzymes that generated products, potassium phosphate pH 7.0). Alkyne (5 mM) was added to these reaction samples were also run through a chiral column WT or F87G flavocytochrome P450 BM3 (3 to 5 mM) and an to identify the relevant enantiomers, using a Chirasil-DEX CB initial spectrum collected. Spectra were then collected every 15 s column on an Agilent 6890 GC. A ‘‘spiking’’ method was used until no further changes were observed. Measurements were whereby the reaction sample was run through the chiral made both with and without addition of NADPH (300 mM) to column in isolation, or with either extra R-orS-enantiomer the enzyme sample. stock (all from Sigma-Aldrich) added to the sample. In this way, either a second chiral peak appeared on the chromato- Kinetic analysis of alkyne oxidation by P450 BM3 gram, or the existing reaction product peak increased in intensity, allowing chirality of the original reaction product Assays of the substrate-dependent NADPH oxidation to be deduced. This approach was taken using both S- and catalysed by WT and F87G mutants of P450 BM3 were done R-enantiomers of 1-octyne, but since only a racemic mixture of in assay buffer (100 mM potassium phosphate pH 7.0) in a 1-hexyne enantiomers was commercially available, chirality of 1 1 cm pathlength quartz cuvette at 25 C on a Cary UV-50 products from BM3-mediated oxidation of 1-hexyne was spectrophotometer, typically using up to 200 nM enzyme in

Published on 23 March 2011. Downloaded by The University of Manchester Library 02/09/2013 22:32:22. inferred from the order of elution of the enantiomers of the each assay, and with NADPH maintained at a near-saturating 1-hexyne racemic standard by comparison with those for the concentration (200 mM, the Km value is approximately 1-octyne enantiomers. 1 ml samples of the 1-octyne or 1-hexyne 35 7.2 mM). To determine the kcat and Km parameters for the reaction products were injected onto the chiral column. The substrates, the alkyne substrate concentration was varied initial oven temperature was 60 1C for 5 min, followed by a and the initial rate of NADPH oxidation was determined gradient over 10 min to 200 1C. The process was repeated with (typically over the first 15–30 s of the reaction) using samples of either (R)-(+)-1-octyn-3-ol or (S)-( )-1-octyn-3-ol, 1 1 De340 = 6.21 mM cm . At least three replicates were done and for a racemic 1-octynl-3-ol sample. The process was then at each substrate concentration, and plots of initial rate versus repeated for the 1-octyne reaction product spiked with either substrate concentration data were fitted using the Michaelis– the (R)-(+)-1-octyn-3-ol or (S)-()-1-octyn-3-ol enantiomer. Menten function to produce the relevant kinetic constants. In similar fashion, the reaction products of 1-hexyne reactions, Control reactions were also done in the absence of added the racemic 1-hexyn-3-ol sample and combinations thereof alkynes, in presence of hexanoic and octanoic acids, octane were run on the chiral column. Agilent GC-MS system soft- and hexane, and also using arachidonic acid (a good substrate ware was used for identification of specific P450-generated 15 for P450 BM3). products by their comparison to molecules in a GC-MS fragment library. Analysis of products from P450 BM3 turnover of alkynes Turnover reactions of WT and F87G mutants of P450 BM3 with alkyne substrates (and also with hexane and octane) were Results done using 1 mM purified enzyme, 100 mM NADPH and 5 mM Interaction of alkynes with oxidized P450 BM3 alkyne/alkane substrate (dissolved in DMSO) in 10 ml assay buffer. DMSO solvent did not exceed 5% of the overall Typical substrates for flavocytochrome P450 BM3 are long reaction volume. Glucose-6-phosphate dehydrogenase (200 mU) chain fatty acids, interactions with which result in a shift in the

372 Metallomics, 2011, 3, 369–378 This journal is c The Royal Society of Chemistry 2011 View Article Online

Fig. 4 UV-visible absorption spectra collected during incubations of P450 BM3 with alkynes in the presence of NADPH. Panels A and B show UV-visible absorption spectra collected for WT P450 BM3 under turnover conditions with 2-octyne (B3.5 mM enzyme) and 1-octyne (B4 mM enzyme), respectively. In each case, reactions were initiated by addition of 5 mM alkyne to a 1 ml enzyme sample in assay buﬀer immediately following addition of NADPH (300 mM). Spectra were acquired every 15 s until no further changes were observed. Panel A shows selected spectra indicating progressive and substantial destruction of the heme chromophore by 2-octyne concomitant with NADPH oxidation, seen both in the region of the Soret band (B418 nm) and for the smaller heme alpha/beta bands between 500–600 nm. The thick solid line is immediately before addition of 2-octyne, and the dotted line is the ﬁrst spectrum collected after 2-octyne addition, Subsequent spectra are shown as

thin solid lines, demonstrating continuous depletion of the heme chromophore as NADPH is oxidized (decreasing A340). On exhaustion of NADPH, spectral increases in the B440–490 nm region demonstrate the reoxidation of the BM3 ﬂavin cofactors (FAD and FMN), with the ﬁnal spectrum (collected at 90 min) shown as a dashed line. Panel B shows a similar reaction with 1-octyne. In this case, enzyme mediated oxidation of

NADPH (decreasing A340) occurs without substantial destruction of the heme chromophore, and consistent with the productive oxidation of the substrate. The final spectrum shown was collected at 90 min. For the 1-octyne reaction, more extensive aerobic incubation of the 1-octyne reaction also led to reoxidation of the BM3 flavin cofactors once NADPH was exhausted. In both panels, arrows indicate directions of absorption change (as appropriate) in regions of NADPH, heme and flavins, and following the addition of alkynes to start the reactions.

1 ferric heme iron spin-state from low-spin (S = 2) towards high had limited solubility in water. Thus, determination of aﬃnity 5 spin (S = 2), with a concomitant shift of the major heme could not be achieved by optical titration of the P450s absorption band (the Soret) from B418 nm to B390 nm with any of these molecules. In absence of added NADPH (a so-called type 1 P450 optical shift). This property can be (the reducing coenzyme for BM3), only very minor decreases

Published on 23 March 2011. Downloaded by The University of Manchester Library 02/09/2013 22:32:22. used to enable determination of the apparent binding constant in the heme spectrum were observed over several hours

(Kd) for fatty acids binding to P450 BM3. In the case of the incubation with various alkynes, possibly indicative of small polyunsaturated C20 substrate arachidonic acid, the Kd was amounts of heme loss or modification by alkynes. Parallel reported at B3.5 mM,23,36 and for the monounsaturated C16 control studies of the isolated heme domain of BM3 indicated 37 substrate palmitoleic acid the reported Kd was 7 mM. Both similar minor decreases in heme absorption over time. substrates induce substantial optical shifts consistent with However, substantial destruction of the heme chromophore extensive development of high-spin heme iron. X-ray crystallo- was observed when NADPH and a NADPH regenerating graphic studies revealed a binding mode for palmitoleic acid in system was added to WT and F87G flavocytochrome BM3 which the carboxylate group is tethered near the mouth of the enzymes in the presence of various alkynes. P450 active site through interactions with the side chains of residues of arginine 47 and tyrosine 51.16 The other end of the NADPH-dependent BM3 heme modification by non-terminal substrate extends down the active site towards the heme iron, hexynes and octynes and a likely catalytically relevant mode is one in which the terminal methyl group interacts with the side chain of Having shown that no significant changes to heme spectra are phenylalanine 87, protecting it from oxidative attack by observed with addition of alkynes to oxidized BM3 enzymes, BM3 compound I, but exposing adjacent carbons to enable we next sought to establish whether the alkynes interacted hydroxylation at o-1, o-2 and o-3 positions (Fig. 1). with the P450 heme under turnover conditions, i.e. on the In preliminary work, we investigated whether the addition addition of NADPH. Potentially, oxidative attack at the of a series of short chain alkynes (1-octyne, 2-octyne, 3-octyne, acetylenic bonds of the alkynes could lead to modification of 4-octyne, 1-hexyne, 2-hexyne and 3-hexyne) to WT and F87G the P450 heme by heme alkylation, with perturbation of the P450 BM3 enzymes induced a type 1 shift of the heme Soret heme spectrum. The UV-visible spectroscopic studies carried band in either enzyme. Only very minor optical changes were out above were thus repeated in the presence of 300 mM observed on addition of any of these molecules, all of which NADPH, and demonstrated clear destruction of the heme

This journal is c The Royal Society of Chemistry 2011 Metallomics, 2011, 3, 369–378 373 View Article Online

chromophore in the majority of cases examined. For the WT was also shown that NADPH-dependent turnover of F87A

BM3 enzyme, the rate of NADPH oxidation (DA340) P450 BM3 actually resulted in production of o-4, o-5 and o-6 was considerably enhanced in the presence of the alkynes hydroxylated myristic acid (as well as some o-1 to o-3 (compared to in their absence), with the least effective alkynes products), but not the o-hydroxy myristic acid.40 Studies on in this regard being 2-hexyne and 3-hexyne. This suggests the F87V BM3 mutant also indicated that the preferred that the alkynes can bind BM3 in a substrate-like manner, oxidation site moved away from the o-terminal for the displacing a water molecule that occupies the distal heme iron substrate arachidonic acid, forming (14S,15R)-epoxyeico- ligand position in the resting state of the enzyme and enabling satrienoic acid as 99% of the total products.41 Our own electron transfer from the reductase to the P450 heme to previous studies indicated the retention of high catalytic facilitate oxidative attack on the alkynes.23 Concomitant with activity towards various fatty acid substrates in a F87G the stimulated NADPH oxidation, substantial destruction of mutant.15 Thus, in light of uncertainties of the role of Phe87 the heme chromophore was also observed during reactions of in regioselectivity of substrate oxidation, we re-investigated WT P450 BM3 with all alkynes tested, apart from 1-hexyne the influence of the panel of alkynes on the F87G BM3 heme and 1-octyne. Fig. 4A shows data from the reaction of WT spectrum in the presence of NADPH. For this mutant, BM3 with 2-octyne in the presence of NADPH, clearly incubation with all alkynes (including 1-hexyne and 1-octyne) demonstrating the progressive diminution of the heme resulted in heme destruction, suggesting that (unlike the case absorption concomitant with NADPH consumption. This for WT BM3) the F87G mutation enabled access of the P450’s result indicated that the stimulated NADPH oxidation compound I intermediate to the terminal acetylenic group of observed in the WT BM3 reactions with the non-terminal the alkynes. However, for the F87G BM3 mutant, the alkyne- alkynes was associated with the oxidative attack of the BM3 dependent enhancements of NADPH oxidation rate appeared ferryl-oxo (compound I) intermediate on the carbon–carbon less than for the WT BM3 with most of the alkynes tested. To triple bond to generate a reactive species that alkylates the explore further the catalytic capacities of the WT and F87G heme at the pyrrole nitrogen (see Fig. 2). As also seen in BM3 enzymes with the various alkynes, we analysed catalytic Fig. 4A, following the complete exhaustion of the NADPH parameters and investigated products formed. (to form NADP+) there is an increase in absorbance in the 440–500 nm region. This reflects the reoxidation of the flavin Steady-state kinetic analysis of alkyne-dependent NADPH (FAD and FMN) cofactors in P450 BM3, and indicates that oxidation in P450 BM3 these cofactors are not inactivated by the alkynes. Fig. 4B shows the reaction of WT BM3 with 1-octyne. In As described in the Materials and Methods section, the apparent

this case, there is negligible change in the spectrum of the kcat and Km parameters for WT and F87G BM3 mutants in protein over the course of the reaction, with the only significant reactions with various alkynes were determined by measuring changes observed reflecting NADPH oxidation at 340 nm. substrate-dependent NADPH oxidation spectrophotometri- A similar spectral set was obtained for the reaction of WT cally. Initial control studies with arachidonic acid indicated BM3 with 1-hexyne. It was also clear from these reactions that limiting rates in excess of 5000 min1 for WT and F87G (despite lack of heme chromophore loss) NADPH oxidation enzymes, confirming their integrity and expected catalytic by WT BM3 was completed more quickly in the presence of behaviour. Data collected are presented in Table 1. For

Published on 23 March 2011. Downloaded by The University of Manchester Library 02/09/2013 22:32:22. 1-hexyne and 1-octyne than in the presence of the other comparison, catalytic constants were also determined for C6 non-terminal alkynes. The lack of heme destruction observed (hexanoic) and C8 (octanoic) acids. For both WT and F87G

for the terminal alkynes suggested that these compounds BM3 enzymes, superior kcat/Km values were obtained for might be oxidized productively by WT P450 BM3 with octanoic acid compared to hexanoic acid (e.g. 2115 min1/ hydroxylation at one or more positions away from the 12.3 mM versus 1180 min1/86.9 mM for WT BM3). The WT

carbon–carbon triple bond. enzyme also exhibited higher kcat values than F87G BM3 1 In previous work, mutation of BM3 at residue Phe87 was for octanoate/hexanoate (F87G kcat values are 766 min / reported to result in near-complete o-oxidation of lauric and 575 min1). The apparent turnover numbers for 1-hexyne and myristic acids by a F87A mutant (from NMR analysis of 1-octyne were lower for WT BM3 than with the respective products), compared with o-1 to o-3 oxidation of these acids (372 min1 and 1465 min1, respectively), but again substrates catalyzed by the WT BM3.38 This was consistent higher than those for F87G BM3 with 1-hexyne/1-octyne with results from structural analysis of the BM3 heme domain, (144 min1/204 min1). By comparison, apparent turnover where Phe87 was seen to interact with the terminal methyl rates for 2-hexyne and 3-hexyne were much slower, and only group of palmitoleic acid16 (Fig. 1), assuming that Phe87 slightly higher than the background rate of NADPH oxidation mutation results in exposure of the substrate terminal methyl by WT/F87G enzymes through the reductase domain 1 group as the favoured site of oxidative attack by the enzyme. (B5 min ). The kcat/Km parameters could not be determined However, other studies have provided varying catalytic results accurately in these cases, or for the F87G BM3 mutant with 2-, from studies of mutants at this position. Weber et al. reported 3- or 4-octyne substrates. For WT BM3, there was clearly that a F87A/A328F double mutant produced predominantly stimulation of NADPH oxidation with the non-terminal (92%) 2-octanol from octane substrate, with 2-octanol alkynes, although maximal rates were o80 min1 for all these production also enhanced (compared to WT BM3) for the alkynes. Further studies using hexane and octane also showed F87A mutant. Formation of the o-hydroxylated product some stimulation of WT and F87G-dependent NADPH (1-octanol) was not reported by Weber et al.39 In addition, it oxidation, although limiting rates were o50 min1 in all cases

374 Metallomics, 2011, 3, 369–378 This journal is c The Royal Society of Chemistry 2011 View Article Online

Table 1 Catalytic parameters of alkyne-dependent NADPH oxidation and for WT/F87G with other alkynes tested, may instead in P450 BM3. The apparent kinetic parameters were determined based indicate that enzyme inactivation by heme modification is the on alkyne substrate-dependent oxidation of NADPH as described in the Materials and Methods section. ND1 indicates that in the case of predominant pathway observed in other cases. Attempts to WT and F87G BM3 enzymes with 2- and 3-hexyne as substrates, and identify products from WT- and F87G BM3-mediated oxida- for the F87G BM3 enzyme with 2-, 3- and 4-octyne as substrates, the tion of the alkynes (and hexane/octane) were next made using NADPH oxidation rate was not stimulated sufficiently over the back- GC-MS methods. ground rate of alkyne-independent NADPH oxidation (B5 min1)to 2 enable accurate analysis of the kcat and Km parameters. ND indicates that in the case of the WT BM3 enzyme with hexane, octane and with Analysis of product formation from P450 BM3-mediated 2-, 3- and 4-octyne as substrates (and for F87G BM3 with hexane and oxidation of alkynes octane), low apparent limiting rates of alkyne-dependent NADPH 1 oxidation of o80 min in all cases (labelled as kcat) could be To establish whether oxidized products were formed from the determined, but a clear dependence on alkyne concentration could various alkyne substrates, reactions were set up for each not be established to enable accurate K estimation m alkyne with both WT and F87G mutant P450 BM3 enzymes, WT P450 BM3 F87G P450 BM3 incorporating both NADPH and a NADPH-regeneration system. Reactions were allowed to progress for B14 h, and Substrate k (min1) K (mM) k (min1) K (mM) cat m cat m samples analysed by GC-MS as described in the Materials and octanoic acid 2115 130 12.3 2.1 766 51 7.7 1.4 Methods section. Significant amounts of product were detected hexanoic acid 1180 175 86.9 32.6 575 70 56.0 16.5 only in the reactions of WT BM3 with 1-hexyne and 1-octyne. 1-hexyne 372 43 30.7 8.1 144 8 2.0 0.5 2-hexyne ND1 ND1 ND1 ND1 With 1-octyne (and a NADPH regeneration system), 3-hexyne ND1 ND1 ND1 ND1 approximately 96.8% of the substrate was converted to 1-octyne 1465 162 12.7 3.5 204 14 1.3 0.5 product following the turnover reaction with WT BM3, with 2-octyne 70 6ND2 ND1 ND1 3-octyne 62 7ND2 ND1 ND1 a very small peak of unconverted substrate still remaining 4-octyne 54 4ND2 ND1 ND1 (B3.2%) with retention time 7 min 42 s. The major product hexane 38 7ND2 48 8ND2 (57.7%) was identified with a gas chromatogram retention 2 2 octane 23 5ND 19 4ND time of 11 min 54 s, as can be seen in Fig. 5. Automated comparison with the GC-MS fragment library selected (Table 1). Collectively, these data again suggest that 1-hexyne 1-octyn-3-ol as the most likely product, and this was and 1-octyne may be competent substrates for WT P450 BM3, confirmed by comparison to a commercially available but the much lower rates for F87G with both these substrates, standard. The two other most prominent peaks had retention Published on 23 March 2011. Downloaded by The University of Manchester Library 02/09/2013 22:32:22.

Fig. 5 GC-MS of 1-octyne turnover by ﬂavocytochrome P450 BM3. Panel A show the gas chromatogram from a 1-octyne turnover reaction. 1-octyne was incubated with WT P450 BM3, NADPH and the glucose-6-phosphate/glucose-6-phosphate dehydrogenase NADPH coenzyme regeneration system. Reaction details are given in the Materials and Methods section. Products and unconverted substrate were extracted from the aqueous phase using MTBE and resolved by gas chromatography. Peak a is assigned to unreacted substrate (1-octyne) and elutes after 7 min 42 s. DMSO (used as substrate solvent) is assigned to peak b (8 min). The principal product, which elutes after 11 min 54 s, is assigned to 1-octyn-3-ol (peak c). Peaks d and e, which elute after 12 min 8 s and 12 min 23 s, are assigned to 1-octyn-5-ol and 1-octyn-4-ol, respectively. Peak f (retention time of 12 min 32 s) is putatively assigned to a ketone derivative of one of the earlier eluting products. Panel B shows the mass spectrum corresponding to the GC observed for peak c. Library searches allowed assignment of the product as 1-octyn-3-ol, as also conﬁrmed by comparison to a commercially obtained 1-octyn-3-ol standard. Indicated is the principal fragmentation yielding the 70 and 55 amu fragment ions.

This journal is c The Royal Society of Chemistry 2011 Metallomics, 2011, 3, 369–378 375 View Article Online

times of 12 min 8 s and 12 min 23 s, corresponding to 1-octyn- superfamily through its fusion of a P450 to a eukaryotic-like 5-ol (12.6%) and 1-octyn-4-ol (20.8%). A further minor CPR module and its very high activity in hydroxylation and product at retention time 12 min 32 (5.7%) was tentatively epoxidation of a wide range of fatty acids.42 Studies of BM3 assigned to a ketone derivative of one of the aforementioned structure and mechanism have enabled assignment of roles of products. In studies of 1-hexyne only one major peak was specific amino acids in functions such as substrate binding, identified at 6 min 54 s, which was identified as 1-hexyn-3-ol, proton relay to heme iron and regulation of heme iron again both by library search and by comparison to a commer- potential, and have also highlighted the rapid electron transfer cially purchased standard. There were no obvious secondary between the fused CPR and heme (P450) domains as a key products (e.g. hexyn-4-ol) observed. No unreacted substrate factor underlying the high oxygenase activity of this enzyme 1 14–16,43–46 was identified in the case of 1-hexyne, but it appears likely that (kcat is B285 s with arachidonic acid). In recent 1-hexyne co-eluted in the solvent phase from comparison with years there have been important studies aimed at diversifying substrate-only control studies. Following assignment of the the substrate specificity range of BM3 by rational mutagenesis major turnover products as 1-octyn-3-ol and 1-hexyn-3-ol, and directed evolution approaches, including work that has their chirality was established by injecting onto a chiral GC enabled the production of BM3 variants capable of e.g. column as described in the Materials and Methods. Both oxygenating short chain alkanes and alkanoic acids, and (S)-()-1-octyn-3-ol and (R)-(+)-1-octyn-3-ol standards were generating human drug metabolites (e.g. ref. 26,34,47). run, and these eluted with retention times of 12 min 54 s and To date, there have been no reports of the catalytic properties 13 min 6 s, respectively. Comparison with the 1-octyn-3-ol of BM3 towards short chain alkynes, although turnover of turnover product indicated that the major enantiomeric 17-octadecynoic acid (17-ODYA an acetylenic derivative of product formed could be assigned as (S)-()-1-octyn-3-ol stearic acid and a substrate for BM3) showed that this (B90%), with B10% of product assigned to the (R)-(+)-1- compound was inhibitory and led to heme alkylation.48 For octyn-3-ol enantiomer (Supplementary Data Fig. S1z). In 17-ODYA, the o-2 hydroxylated product was observed studies of 1-hexyn-3-ol products, we could not obtain (16-hydroxy-17-ODYA) in addition to inactivated enzyme, commercially the relevant enantiomers in either pure R or S and further turnover studies using 16-hydroxy-17-ODYA also forms. However, given the similar nature of these compounds resulted in heme chromophore destruction. These data are to the 1-octyn-3-ol enantiomers, a common elution pattern consistent with WT BM3-mediated oxidation of 17-ODYA could be inferred. Two product peaks were identified with at both o-1 (leading to heme alkylation) and o-2 (generating elution times of 10 min 0 s and 10 min 18 s, which we assign to 16-hydroxy-17-ODYA product) positions.48 Other alkynes are (S)-()-1-hexyn-3-ol (B90%) and (R)-(+)-1-hexyn-3-ol also well known inactivators of P450 function, and this has (10%), respectively. Thus, substantial amounts of alkynol been exploited in clinically used acetylenic sterol-derived drugs products were generated only in turnover of 1-hexyne and such as Danazol (pregna-2,4-dien-20-yno[2,3-d]isoxazol-17-ol) 1-octyne with WT BM3, forming predominantly the respective and Norgestrel (13-ethyl-17-hydroxyl-18, 19-dinor-17 alpha- (S)-()-1-alkyn-3-ol enantiomer in both cases. The absence of pregn-4-en-20-yn-3-one) used mainly in endometriosis treat- any notable oxygenated products with other alkynes is ment and as a contraceptive agent, respectively.49 We decided consistent with their inducing heme destruction in the WT to investigate the capacity of BM3 to catalyse oxidation of C6 BM3 enzyme. The lack of products in the case of F87G BM3 and C8 alkynes in order to explore further the ability of the

Published on 23 March 2011. Downloaded by The University of Manchester Library 02/09/2013 22:32:22. with any of the alkyne substrates is also consistent with the enzyme to oxidize such molecules, and to establish the regio- heme destruction noted in these reactions, indicating the selectivity of the enzyme and the influence of the position of accessibility of the terminal acetylene group for oxidative the carbon–carbon triple bond on the relative abilities of these attack from the P450 compound I in this mutant. molecules to act as substrates or inhibitors of P450 function. A For comparison, the WT BM3-mediated turnover of octane F87G BM3 mutant was also used in these studies of alkyne and hexane was also analysed. In the case of octane, there was oxidation, in view of previous conflicting reports of exactly a much poorer conversion of the substrate (possibly due to its how Phe87 controls regioselectivity of oxidation of alkyl chains. limited solubility in water), with B60% of octane unconverted Our studies demonstrate firstly that WT BM3 is clearly after extended incubation with enzyme. Two major products capable of oxidative attack on all of the alkynes tested were observed with retention times of 11 min 54 s and 12 min 6 s, (1-, 2- and 3-hexynes, 1-, 2-, 3- and 4-octynes), but also show these being identified as octan-4-ol (B24%) and octan-3-ol that all but the terminal alkynes (1-hexyne and 1-octyne) act as (B14%), respectively. A further small peak at 11 min 30 s suicide substrates, with compound I attack resulting in loss (B2%) was assigned as a ketone derivative, putatively octan- of the chromophore through heme alkylation. However, 4-one (Supplementary Data Fig. S2z). In the case of hexane, 1-hexyne and 1-octyne are instead competent substrates for hexan-3-ol (B98%) and hexan-2-ol (B2%) products were WT BM3, with negligible chromophore loss during turnover, identified, although residual hexane substrate was not resolved and with 1-hexyn-3-ol and 1-octyn-3-ol being the major from the solvent system used, preventing establishment of products formed. Only in the case of 1-octyne were additional reaction efficiency (data not shown). 1-octyn-4-ol and 1-octyn-5-ol products detected (and a minor putative ketone species). For 1-hexyne, the 1-hexyn-3-ol was Discussion the only detectable product. Other minor peaks seen in GC-MS studies did not correlate with the relevant C6 or C8 The Bacillus megaterium flavocytochrome P450 BM3 enzyme acids, indicating it unlikely that a pathway occurs in which has become a paradigm system in the P450 enzyme o-oxidation of 1-hexyne and 1-octyne by WT BM3 produces

376 Metallomics, 2011, 3, 369–378 This journal is c The Royal Society of Chemistry 2011 View Article Online

reactive ketenes that are then hydrolysed to the respective energetically favoured oxidation of the acetylene groups carboxylic acids. Analysis of the retention time and comparisons being removed. with standards indicated that the major enantiomers formed Given the apparent ‘‘productive’’ binding mode of the were (S)-()-1-octyn-3-ol (B90%) and (S)-()-1-hexyn-3-ol 1-alkynes with WT BM3 (i.e. orientated with the acetylenic (B90%). The data indicate that WT BM3 effectively protects group interacting with Phe87) and the lack of evidence for both the o- and o-1 acetylenic carbons in 1-hexyne and hydroxylation(s) near the opposite ends of these substrate 1-octyne, favouring product formation by o-2 oxidation. In molecules, the WT BM3 enzyme clearly has biotechnological the F87G mutant, there were not any notable products potential in oxidation of terminal alkynes. Assuming that this obtained following extended turnover of alkynes, and heme substrate binding property extends to the binding of other chromophore destruction occurred in all cases. This indicates terminal alkynes (and possibly alkenes) of differing lengths, that the absence of the Phe87 phenyl side chain enables the this could clearly provide the basis for the exploitation mutant enzyme’s compound I to attack the o-1 carbon, of the WT BM3 enzyme for production of 1-alkyn-3-ols leading to heme alkylation in the dominant pathway, and (and possibly 1-alken-3-ols) from different chain length 1-alkynes without any significant o-oxidation that could generate (1-alkenes), and likely with a high enantiomeric excess of carboxylic acid products. The data presented here for WT product. A potential issue here is the efficiency with which BM3-dependent oxidation of terminal alkynes are consistent these BM3-catalysed reactions occur. In the case of 1-hexyne with preceding data for BM3-dependent oxidation of the and 1-octyne as substrates, the extent of coupling of NADPH acetylenic fatty acid 17-ODYA insofar as o-2 hydroxylation oxidation to alkyne hydroxylation is difficult to assess, given is observed in all cases. However, heme chromophore that a NADPH regeneration system is required to enable destruction in reactions with 17-ODYA (but not to any substantial amounts of hydroxylated product. However, it is considerable extent with the 1-alkynes) suggests that o-1 unlikely that the coupling efficiency is greater than 20% in hydroxylation is more favoured with this substrate.43 This these reactions. Moreover, the conversion of P450 BM3 from a may reflect differing binding modes of these substrates, with ‘‘fast’’ to a ‘‘slow’’ oxidase form is known to occur with poor the 1-alkynes lacking a carboxylate group capable of making substrates, and this is almost certainly due to the accumulation stabilizing binding interactions with the Arg47/Tyr51 motif at of the 2-electron reduced (hydroquinone) form of the FMN the mouth of the active site (Fig. 1). Also of note is the fact cofactor in the reductase module of the enzyme. The 1-alkynes that the products formed by WT BM3’s turnover of 1-octyne do not induce any considerable perturbation of the BM3 heme and 1-hexyne (predominantly the respective 1-alkyn-3-ols, iron spin-state (towards high-spin), unlike several long chain but also 1-octyn-4-ol and 1-octyn-5-ol in the case of fatty acid substrates.13 Fatty acid-induced heme iron spin-state 1-octyne substrate) almost certainly result from the substrates change results in a positive shift of the potential of the ferric progressing into the BM3 active site led by the acetylenic end iron by B130–140 mV, stimulating electron transfer from of the molecules. In the WT BM3 enzyme, this is possibly due reductase FMN-to-heme iron and the efficient oxidation of to stabilizing interactions between the aromatic side chain P450 substrate.14,36 The favoured electron donor to the heme of Phe87 and the p electrons of the acetylene triple bond, iron is the FMN semiquinone, but in absence of efficient resulting in a major productive binding mode exposing electron transfer to the heme, further NADPH-dependent adjacent positions for oxidation by the P450 compound I50 reduction of the reductase domain generates FMN 52

Published on 23 March 2011. Downloaded by The University of Manchester Library 02/09/2013 22:32:22. (predominantly o-2 for 1-hexyne, o-2/o-3 for 1-octyne). hydroquinone. The FMN hydroquinone is an inefficient Further studies of the turnover of octane and hexane by WT electron donor to the heme iron,53 an effect that would be BM3 also indicate that once oxidation occurs near one end of compounded by the lack of high-spin heme iron in 1-alkyne the molecule (e.g. o-3 hydroxylation of octane), subsequent bound enzyme. Routes to improving the binding of 1-alkynes oxidation near the other end does not occur (at least with any would include rational mutagenesis or directed evolution measurable efficiency). Simply from an energetic perspective, approaches, possibly creating a constrained environment for it might have been expected that BM3 would oxidise the the substrate in the BM3 heme vicinity to enable more terminal acetylene group of the 1-alkynes to generate reactive effective substrate-dependent heme iron spin-state conversion ketenes/carboxylic acids or to result in heme alkylation, (and thus faster electron transfer from FMN semiquinone) according to whether oxygen addition occurred at the o-or while retaining regioselectivity of oxidation. o-1 carbon, respectively. However, it is known that BM3 In conclusion, studies of various C6 and C8 alkynes as suppresses o-oxidation of fatty acids by sequestration of the substrates for the WT and F87G mutant of flavocytochrome terminal methyl group of substrates in a ‘‘pocket’’ adjacent to P450 BM3 indicated that only the WT BM3 evaded heme Phe87.45,48 Bond energies for o-2 secondary propargylic C–H chromophore destruction by heme alkylation through prefer- bonds (o94 kcal mol1) are slightly weaker than those at o-3 ential 3-hydroxylation of 1-hexyne and 1-octyne. In all other (B98 kcal mol1), and o-2 oxidation is indeed predominant in cases, oxidative attack at the carbon–carbon triple bond WT BM3-dependent oxidation of 1-hexyne and 1-octyne.51 resulted in heme alkylation and loss of enzyme chromophore However, it appears that structural constraints dominate the and activity. The ability of WT BM3 to catalyse production of regioselectivity of oxidation of these molecules by WT BM3 1-alkyn-3-ols might prove useful in generation of building (as is also the case for fatty acid substrates), and the impor- blocks for synthesis of other chemicals (e.g. ref. 54), or for tance of Phe87 in this selectivity becomes evident in the production of the alkynols in their own right. For instance, F87G mutant, where alkynes at all positions become suicide 1-octyn-3-ol has applications in prevention of corrosion of inhibitors of the enzyme due to steric restrictions to the iron,55 and is also an attractant for mosquitoes.56

This journal is c The Royal Society of Chemistry 2011 Metallomics, 2011, 3, 369–378 377 View Article Online

References 31 K. L. Kunze, B. L. Mangold, C. Wheeler, H. S. Beilan and P. R. Ortiz de Montellano, J. Biol. Chem., 1983, 258, 4202–4207. 1 I. G. Denisov, T. M. Makris, S. G. Sligar and I. Schlichting, Chem. 32 W. K. Chan, Z. Sui and P. R. Ortiz de Montellano, Chem. Res. Rev., 2005, 105, 2253–2277. Toxicol., 1993, 6, 38–45. 2 A. W. Munro, H. M. Girvan and K. J. McLean, Nat. Prod. Rep., 33 J. S. Miles, A. W. Munro, B. N. Rospendowski, W. E. 2007, 24, 585–609. Smith, J. McKnight and A. J. Thomson, Biochem. J., 1992, 288, 3 F. P. Guengerich, Chem. Res. Toxicol., 2008, 21, 70–83. 503–509. 4 A. W. Munro, H. M. Girvan and K. J. McLean, Biochim. Biophys. 34 T. Omura and R. Sato, J. Biol. Chem., 1964, 239, 2370–2378. Acta, Gen. Subj., 2007, 1770, 345–359. 35 O. Roitel, N. S. Scrutton and A. W. Munro, Biochemistry, 2003, 5 J. Rittle and M. T. Green, Science, 2010, 330, 924–925. 42, 10809–10821. 6 B. P. Unger, I. C. Gunsalus and S. G. Sligar, J. Biol. Chem., 1986, 36 S. N. Daff, S. K. Chapman, K. L. Turner, R. A. Holt, 261, 1158–1163. S. Govindaraj, T. L. Poulos and A. W. Munro, Biochemistry, 7 G. I. Lepesheva and M. R. Waterman, Biochim. Biophys. Acta, 1997, 36, 13816–13823. Gen. Subj., 2007, 1770, 467–477. 37 L. A. Cowart, J. R. Falck and J. H. Capdevila, Arch. Biochem. 8 S. L. Kelly, D. C. Lamb, M. Cannieux, D. Greetham, Biophys., 2001, 387, 117–124. C. J. Jackson, T. Marczylo, C. Ugochukwu and D. E. Kelly, 38 C. F. Oliver, S. Modi, M. J. Sutcliffe, W. U. Primrose, L. Y. Lian Biochem. Soc. Trans., 2001, 29, 122–128. and G. C. Roberts, Biochemistry, 1997, 36, 1567–1572. 9 K. J. McLean, A. J. Dunford, M. Sabri, R. Neeli, H. M. Girvan, 39 E. Weber, A. Seifert, M. Antonovici, C. Geinitz, J. Pleiss and P. R. Balding, D. Leys, H. E. Seward, K. R. Marshall and V. B. Urlacher, Chem. Commun., 2011, 47, 944–946. A. W. Munro, Biochem. Soc. Trans., 2006, 34, 1178–1182. 40 P. C. Cirino and F. H. Arnold, Adv. Synth. Catal., 2002, 344, 10 D. R. Nelson, Biochim. Biophys. Acta, Proteins Proteomics, 2011, 932–937. 1814, 14–18. 41 S. Graham-Lorence, G. Truan, J. A. Peterson, J. R. Falck, S. Wei, 11 D. R. Nelson, L. Koymans, T. Kamataki, J. J. Stegeman, C. Helvig and J. H. Capdevila, J. Biol. Chem., 1997, 272, R. Feyereisen, D. J. Waxman, M. R. Waterman, O. Gotoh, M. J. 1127–1135. Coon, R. W. Estabrook, I. C. Gunsalus and D. W. Nebert, 42 A. W. Munro, D. G. Leys, K. J. McLean, K. R. Marshall, Pharmacogenetics, 1996, 6, 1–42. T. W. Ost, S. Daff, C. S. Miles, S. K. Chapman, D. A. Lysek, 12 F. Hannemann, A. Bichet, K. M. Ewen and R. Bernhardt, C. C. Moser, C. C. Page and P. L. Dutton, Trends Biochem. Sci., Biochim. Biophys. Acta, Gen. Subj., 2007, 1770, 330–344. 2002, 27, 250–257. 13 L. O. Narhi and A. J. Fulco, J. Biol. Chem., 1986, 261, 7160–7169. 43 T. W. Ost, C. S. Miles, A. W. Munro, J. Murdoch, G. A. Reid and 14 A. W. Munro, S. Daff, J. R. Coggins, J. G. Lindsay and S. K. Chapman, Phenylalanine 393 exerts thermodynamic control S. K. Chapman, Eur. J. Biochem., 1996, 239, 403–409. over the heme of flavocytochrome P450 BM3, Biochemistry, 2001, 15 M. A. Noble, C. S. Miles, S. K. Chapman, D. A. Lysek, 40, 13421–13429. A. C. Mackay, G. A. Reid, R. P. Hanzlik and A. W. Munro, 44 H. Yeom, S. G. Sligar, H. Li, T. L. Poulos and A. J. Fulco, The Biochem. J., 1999, 339, 371–379. role of Thr268 in oxygen activation of cytochrome P450BM-3, 16 H. Li and T. L. Poulos, Nat. Struct. Biol., 1997, 4, 140–146. Biochemistry, 1995, 34, 14733–14740. 17 J. B. Johnston, H. Ouellet, L. M. Podust and P. R. Ortiz de 45 K. G. Ravichandran, S. S. Boddupalli, C. A. Hasermann, Montellano, Arch. Biochem. Biophys., 2011, 507, 86–94. J. A. Peterson and J. Deisenhofer, Science, 1991, 261, 731–736. 18 P. R. Ortiz de Montellano, Drug Metab. Rev., 2008, 40, 405–426. 46 A. W. Munro, J. G. Lindsay, J. R. Coggins, S. M. Kelly and 19 J. D. Imig, A. P. Zhou, D. E. Stec, D. R. Harder, J. R. Falck and N. C. Price, FEBS Lett., 1994, 343, 70–74. R. J. Roman, Am. J. Physiol., 1996, 270, R217–227. 47 A. M. Sawayama, M. M. Chen, P. Kulanthaivel, M. S. Kuo, 20 U. Hoch, Z. Zhang, D. L. Kroetz and P. R. Ortiz de Montellano, H. Hemmerle and F. H. Arnold, Chem.–Eur. J., 2009, 15, Arch. Biochem. Biophys., 2000, 373, 63–71. 11723–11729. 21 U. Hoch and P. R. Ortiz de Montellano, J. Biol. Chem., 2001, 276, 48 N. Shirane, J. A. Peterson and P. R. Ortiz de Montellano, 11339–11346. Biochemistry, 1993, 32, 13732–13741. 22 Y. M. Zheng, B. R. Baer, M. B. Kneller, K. R. Henne, K. L. Kunze 49 P. R. Ortiz de Montellano and M. A. Correia, Inhibition of and A. E. Rettie, Biochemistry, 2003, 42, 4601–4606. cytochrome P450 enzyme,inCytochrome P450: Structure,

Published on 23 March 2011. Downloaded by The University of Manchester Library 02/09/2013 22:32:22. 23 H. M. Girvan, K. R. Marshall, R. J. Lawson, D. Leys, Mechanism and Biochemistry, ed. P.R. Ortiz de Montellano, M. G. Joyce, J. Clarkson, W. E. Smith, M. R. Cheesman and 2005, Kluwer Academic/Plenum Publishers, NewYork, USA, A. W. Munro, J. Biol. Chem., 2004, 279, 23274–23286. pp. 247–322. 24 M. G. Joyce, H. M. Girvan, A. W. Munro and D. Leys, J. Biol. 50 F. Meyer-Wegner, H.-W. Lerner and M. Bolte, Acta Crystallogr., Chem., 2004, 279, 23287–23293. Sect. C: Cryst. Struct. Commun., 2010, 66, o182–o184. 25 T. W. Ost, C. S. Miles, J. Murdoch, Y. Cheung, G. A. Reid, 51 K. W. Egger and A. T. Cocks, Helv. Chim. Acta, 1973, 56, S. K. Chapman and A. W. Munro, FEBS Lett.,2000,486, 173–177. 1516–1536. 26 R. J. Sowden, S. Yasmin, N. H. Rees, S. G. Bell and L. L. Wong, 52 M. B. Murataliev, M. Klein, A. J. Fulco and R. Feyereisen, Org. Biomol. Chem., 2005, 3, 57–64. Biochemistry, 1997, 36, 8401–8412. 27 A. Glieder, E. T. Farinas and F. H. Arnold, Nat. Biotechnol., 2002, 53 S. C. Hanley and S. Daﬀ, Biochem. Biophys. Res. Commun., 2004, 20, 1135–1139. 325, 1418–1423. 28 M. W. Peters, P. Meinhold, A. Glieder and F. H. Arnold, J. Am. 54 M. Tiecco, L. Testaferri, A. Temperini, R. Terlizzi, L. Bagnoli, Chem. Soc., 2003, 125, 13442–13450. F. Marini and C. Santi, Tetrahedron Lett., 2007, 48, 4343–4345. 29 C. Fiaud, A. Harch, D. Mallouh and M. Tzinmann, Corros. Sci., 55 C. Fiaud, A. Harch, D. Mallouh and M. Tzinmann, Corros. Sci., 1993, 35, 1437–1444. 1993, 35, 1437–1444. 30 D. L. Kline, S. A. Allan, U. R. Bernier and C. H. Welch, Medical 56 D. L. Kline, S. A. Allan, U. R. Bernier and C. H. Welch, Medical and Veterinary Entomology, 2007, 21, 323–331. and Veterinary Entomology, 2007, 21, 323–331.

378 Metallomics, 2011, 3, 369–378 This journal is c The Royal Society of Chemistry 2011